Dejong 2005 Phdthesis

Nitrogen compounds in pressurised fluidised bed gasification of biomass and fossil fuels

Wiebren de Jong

Nitrogen compounds in pressurised fluidised bed gasification of biomass and fossil fuels

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft, op gezag van de Rector Magnificus prof. dr. ir. J.T. Fokkema, voorzitter van het College voor Promoties, in het openbaar te verdedigen op maandag 7 februari 2005 om 15:30 uur

door Wiebren DE JONG scheikundig ingenieur geboren te Rotsterhaule (Haskerland)

Dit proefschrift is goedgekeurd door de promotoren: Prof. Dr. –Ing. K.R.G. Hein Prof. dr. J.A. Moulijn Samenstelling promotiecommissie: Rector Magnificus, Prof. Dr. –Ing. K.R.G. Hein, Prof. dr. J.A. Moulijn Prof. Dr. –Ing. H. Spliethoff Prof. dr. ir. P.J. Jansens Prof. dr. ir. C. Daey Ouwens Prof. dr. M. Hupa Dr. ir. P.D.J. Hoppesteyn

voorzitter Technische Universiteit Delft, promotor Technische Universiteit Delft, co-promotor Technische Universiteit Delft Technische Universiteit Delft Technische Universiteit Eindhoven Åbo Akademi University, Finland Corus, IJmuiden

Drs. J. Andries heeft als begeleider in belangrijke mate aan de totstandkoming van het proefschrift bijgedragen.

Published and distributed by: Optima Grafische Communicatie Pearl Buckplaats 37 3009 CC Rotterdam The Netherlands Telephone: +31 10 220 11 49 Telefax: +31 10 456 63 54 E-Mail: [email protected] ISBN 90-8559-028-0 Copyright © 2005 by W. de Jong All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the publisher. Printed in the Netherlands

Alles heeft zijn uur en ieder ding onder de hemel zijn tijd. Prediker 3:1. There is a time for everything, and a season for every activity under heaven. Ecclesiastes 3:1.

Voor: Heit en mem en mijn echtgenote Klarine

Nitrogen compounds in pressurised fluidised bed gasification of biomass and fossil fuels

Summary This PhD thesis assesses the experimental and theoretical work which was performed to study the behaviour of nitrogen compounds during airblown pressurised stationary fluidised bed gasification of biomass and brown coal, followed by high temperature ceramic gas filtration. Fossil fuels have dominated the energy supply in modern societies and will continue to do so in the 21st century. The resources, however, are depleting, especially of oil and natural gas. Therefore, other energy sources are to be exploited further within this century. Biomass is one of the almost CO2 neutral, renewable contributors to the future energy production. Nowadays many modern, high efficiency (combined) power and heat producing systems using biomass are or become commercially available. One promising route to efficient power and heat supply is the Integrated Gasification Combined Cycle. This cycle is particularly of interest for medium to larger scale installations. Pressurised operation of the gasifier offers the advantage of smaller process equipment, including that for the necessary downstream gas cleaning. Also, the work needed to compress the gas for gasturbine use will be much smaller or not needed at all. High temperature gas filtration offers the benefit of increased overall efficiencies of the power and heat producing cycle. This integrated gasification technology, however, is still in a stage of development and demonstration. When instead of absorption gas cleaning, high temperature, dry gas filtration is applied, nitrogen compounds, like ammonia (NH3) and hydrogen cyanide (HCN), are not dissolved in the absorption liquid. As a result NOx emissions in gas turbine combustors are produced. NOx is known for its negative effects on the health of humans and animals and acidification of soil and water. Therefore, increasingly stringent emission restrictions are imposed on this component. Both coal and biomass contain nitrogen in their chemical structure and in gasification processes this so-called fuel bound nitrogen is converted to a large extent into NOx precursors. Although woody biomass contains low amounts of nitrogen (only a few tens of mass percentage on dry fuel basis), there is still a high emission potential based on the fuel's energy content, because of the low calorific value of the fuel, as compared to coal. Thus biomass causes significant NOx emissions when no further measures are taken. An introduction regarding the use of biomass in energy production, the potential of NOx emissions of a range of young and old fuels and open research questions is given in chapter 1. A literature overview is presented in chapter 2. A state-of-the-art review of the fluidised bed gasification activities is given. Also, an overview of the modelling of fluidised bed gasification on large and small scale, with emphasis on the emission of NOx precursors is presented. The influence of fuels, additives and process parameters on the release of these compounds is addressed. Both the primary and later stages of the conversion of solid fuels are considered. In chapter 3 the experimental facilities used to study the thermal conversion behaviour of fuel bound nitrogen are presented. These facilities can be divided into two categories: 1) Pilot scale test rigs; a 1.5 MWthermal Pressurised Fluidised Bed Gasifier situated at the section Energy Technology of Delft University of Technology and a similar 50 kWthermal test rig available at the Institute of Process Engineering and Power Plant Technology of Stuttgart University (Germany).

vii

2) Characterisation equipment to study solid fuel reactivity, especially in the early stages of conversion in gasification processes, namely pyrolysis. Here, two techniques are used representing slow and rapid heating conditions, respectively: TG-FTIR (thermogravimetric analysis coupled to FTIR), situated at AFR inc. (Hartford, CT, USA) and heated grid reactor equipped with in-situ IR diode laser diagnostics, available at the Technical Physics department of Eindhoven University of Technology. In chapter 4 the experimental results obtained by using the facilities described above are presented and discussed. In the pressurised fluidised bed gasification tests performed at Delft university, miscanthus and wood pellets have been used. Brown coal has been chosen because it is an older, but still comparatively high-volatile, fossil fuel. No significant radial gradients in the concentration of the main and minor gaseous product constituents were observed. The concentrations of the main gasification product gas components were comparable to the limited open literature data, available from other pressurised fluidised bed test rigs at comparable air stoichiometry values. Axial gradients in the gas concentrations during the pressurised fluidised bed gasification tests could be clearly observed for acetylene, which is related to reactions involving tar and soot precursor formation and destruction. Under the pressurised fluidised bed gasification conditions studied, the main fuel bound nitrogen component produced is NH3, whereas HCN is formed to a minor extent (only a few percent of the fuel bound nitrogen). This was comparable with results from other bottom-fed FB gasifiers. On the other hand, comparatively low values have been found for a top-fed pressurised FB. HNCO and NO were never detected by means of even a high resolution FTIR spectrophotometer under the pressurised gasification test conditions studied. Tests with Ca-containing dolomite and a Ca-less additive (MinPhyl, or Pyrophyllite) under otherwise comparable process conditions showed that an increased Ca inventory in the gasifier increases the NH3/HCN ratio significantly. To obtain basic model input data, flash pyrolysis experiments with miscanthus were conducted using the heated grid reactor set up. This research was focused on measuring the yield of CO, CO2 and NH3 at a heating rate of 280-320 K/s and final temperatures in the range of 1050-1400K. Unfortunately, NH3 could not be detected, due to condensation or the limited frequency range that could be achieved with the tuneable laser. CO and CO2 yields have been measured in-situ and were compared with the FG-DVC biomass pyrolysis model. For all fuels used, kinetic parameters for this pyrolysis model have been determined by application of the Tmax method, using a TG-FTIR system with heating rates of 10, 30 and 100 K/min. However, this model does not predict the pyrolysis product yield satisfactorily at high heating rates, based on the kinetics determined by low heating rates. It gave a reasonable quantitative yield prediction for CO but a substantial under-prediction for CO2. The competition between the evolution of primary products like primary tar fragments and carboxylic acids on one side and light gases like CO, CO2 and H2O on the other side is probably the reason. This competition is expected to be heating rate dependent. Apparently the primary pyrolysis products like tars and carboxylic acids, which contain precursor groups for the formation of CO and CO2, are quickly removed from the reaction zone and quenched immediately. Thus, no time is available for further primary tar and carboxylic acid decomposition into CO2 and CO (to a minor extent), which results in low yields. According to this hypothesis the yields of primary tar and carboxylic acids must be significant. This is confirmed by the observation that under pyrolysis conditions the tar yield increases at increasing heating rates. In chapter 5 the modelling of pressurised fluidised bed gasification is described and the influence of process parameters is assessed. The model is a steady state plug flow-in-series model with detailed reaction kinetics. Heterogeneous char oxidation and gasification (by H2O, CO2 and H2), heterogeneously catalysed HCN hydrolysis and homogeneous reactions (including nitrogen molecular and radical species and a simplified tar cracking reaction) are taken into account.

viii

The possibility to increase the conversion of NH3 into N2 by varying the process conditions or by adding specific compounds to the gasifier has been studied theoretically. However, NH3 is found to be a very stable compound, which is hardly converted in pressurised fluidised bed gasifiers. Decomposition into N2 is slightly increased by increasing temperature, but this option is limited due to the risk of bed sintering for alkali containing biomass fuels. The NH3 conversion is only slightly dependent on reactor pressure. A minimum conversion was obtained around 2.5 bar. At higher pressures (10 bar) the NH3 conversion slightly increased. Gas residence time in the reactor did not affect the fuel-N conversion. The destruction of NH3 is only taking place in the presence of O, H and OH radicals, which are consumed very fast in the initial part of the bed. Larger NH3 conversions can be reached by injecting NO or NO2 into the bed. However, HCN is formed and NO conversion is limited, which leads to undesired emissions. Addition of O2 favours NH3 conversion. However, the main nitrogen species formed was NO rather than N2. In the most favourable case, at near-stoichiometric conditions, NH3 is converted for 50% into NO and 50% into N2. Moreover, the addition of secondary air decreases the already low LCV gas heating value. The presence of a high concentration of CH4 in the bed part of the gasifier reduces the NH3 conversion, probably due to the competition for radicals between CH4, its intermediates (mainly CH3 radicals) and NH3. H2O2 and H2O (steam) addition into the bed did not affect at all NH3 conversion. A comparison between the model based simulations and the experiments on the 1.5 MWth Delft University pressurised fluidised bed gasification scale and the 50 kWth Stuttgart University rig is presented in chapter 6. The agreement for N-species prediction and measurements is quite good for the fuels and more in particular for the main fuel bound nitrogen component in the product gas: NH3. For HCN the concentrations are often underpredicted, probably due to the heterogeneous hydrolysis reaction, taking place at the char surface, which can have slower kinetics than assumed in literature. The model predicts the formation of super-ppmv HNCO and NO concentrations, but they have never been detected by FTIR analysis in our work. Probably, catalytic hydrolysis converts HNCO into NH3 and CO and this reaction has not been considered in the model. For NO, catalytic reduction by ash constituents probably plays a role at the gasification temperatures. It is also possible that neglection of S and Cl chemistry causes deviations between model and experimental results for the minor species. This is less the case for wood. Also, simplification of tar- and char-nitrogen reactions can be significant: the model assumes that the nitrogen which is not available as gas species will be released initially in the form of HCN. The agreement between model and experimental results for the main product gases is reasonably good. The differences between the calculated and measured values can be attributed to pyrolysis yields of H2. These were obtained from a correlation of literature data and as such used as input in the biomass pyrolysis sub-model. Also, the use of simplified tar cracking kinetics, in terms of possible reactions, product yields and rates probably plays a major role in the deviations observed for CO and CO2 concentrations. The deviations are the highest for wood, and less for miscanthus and brown coal. This is in-line with the hypothesis that the simplified tar cracking kinetics plays a major role, as of the tested fuels wood pyrolysis shows the highest initial tar yields. Differences in the heterogeneous combustion and gasification reactions are not expected to have a major impact, as the carbon conversions are quite well predicted. Finally, chapter 7 gives an overview of the conclusions and the recommendations specified for pressurised fluidised bed experimental pilot scale research, fuel characterisation and modelling.

Wiebren de Jong

ix

Stikstofcomponenten in drukwervelbedvergassing van biomassa en fossiele brandstoffen

Samenvatting Dit proefschrift behandelt het experimentele en theoretische werk dat is uitgevoerd in het kader van een studie naar het lot van de stikstof componenten tijdens drukwervelbedvergassing in een stationair wervelbed in combinatie met hoge temperatuur filtratie, gebruikmakend van keramische filters. Fossiele brandstoffen hebben de energievoorziening van de moderne maatschappij bepaald en zullen dit blijven doen in de 21e eeuw. De voorraden raken echter uitgeput, in het bijzonder die van olie en aardgas. Daarom moeten andere energiebronnen in deze eeuw worden aangeboord. Biomassa is een van de bijna CO2 neutrale, hernieuwbare brandstoffen voor de toekomstige energievoorziening. Tegenwoordig zijn er al commerciële systemen op de markt voor (gecombineerde) warmte en kracht productie met een hoog rendement op basis van biomassa. Een veelbelovende optie voor efficiënte elektriciteit- en warmtevoorziening is het gecombineerde vergasser – STEG systeem. Dit systeem is in het bijzonder van belang voor middelgrote tot grote installaties. Het onder druk bedrijven van de vergasser biedt het voordeel van kleinere proces apparatuur, inclusief de downstream gasreiniging. Compressie van het geproduceerde gas, hetgeen nodig is voor het bedrijven van moderne gasturbines, zal in dit geval niet of in mindere mate nodig zijn. Hoge temperatuur gasreiniging door middel van bijvoorbeeld keramische filters biedt het voordeel van hogere efficiënties van het warmte en elektriciteit producerende systeem. Deze geïntegreerde vergassingstechnologie, echter, bevindt zich nog steeds in het stadium van ontwikkeling en demonstratie. Bij toepassing van hoge-temperatuur en dus droge gas filtratie in plaats van natte absorptie technieken zullen stikstofcomponenten, zoals ammoniak (NH3) en waterstofcyanide (HCN), niet oplossen in de absorptievloeistof en zullen NOx vormen in gasturbine verbrandingskamers. NOx heeft negatieve effecten op de gezondheid van mens en dier en veroorzaakt verzuring van de grond en het oppervlaktewater. Daarom worden wereldwijd steeds stringentere emissie eisen opgelegd voor deze emissiecomponent. Zowel kolen als biomassa bevatten chemisch gebonden stikstof en het is deze brandstofgebonden stikstof die in vergassingsprocessen voor een groot deel wordt omgezet in NOx precursors. Hoewel houtachtige biomassa lage gehaltes aan stikstof vertoont (slechts enkele tiendes massaprocenten op droge basis), is er in vergelijking met kolen toch een hoog emissiepotentieel als de lage stookwaarde van de brandstof in ogenschouw wordt genomen. Daarom draagt biomassa significant bij tot NOx emissies wanneer er geen proces gerelateerde maatregelen worden genomen. In hoofdstuk 1 van dit proefschrift wordt een inleiding gegeven in de toepassing van biomassa in de energievoorziening, potentiële NOx emissies van een reeks jonge tot oude brandstoffen, alsmede open onderzoeksvragen op dit gebied. In hoofdstuk 2 wordt een literatuuroverzicht geboden. Hierin wordt een “state-of-the-art” overzicht gegeven van de wervelbedvergassingsactiviteiten op zowel kleine als grote schaal en wordt een overzicht gepresenteerd van de modellering van wervelbedvergassing. De nadruk wordt gelegd op de emissie van NOx precursors, zowel in de primaire als ook de latere stadia van de conversie van vaste brandstoffen. De invloed van verschillende brandstoffen, additieven en proces gerelateerde parameters op het vrijkomen van deze componenten wordt behandeld. Hoofdstuk 3 geeft een overzicht van de toegepaste experimentele technieken de gebruikt zijn voor de studie naar het thermische conversiegedrag van brandstofgebonden stikstof. Deze kunnen worden onderverdeeld in twee categorieën:

xi

1) Proefopstellingen op pilot schaal: de 1.5 MWthermisch drukwervelbedvergassingsopstelling in het laboratorium van de sectie Energy Technology van de Technische Universiteit Delft, en de 50 kWthermisch opstelling die beschikbaar is bij het “Institut für Verfahrenstechnik und Dampfkesselwesen (IVD)” van de Universiteit Stuttgart (Duitsland). 2) Fundamentele karakteriseringsapparatuur voor bestudering van de vaste brandstofreaktiviteit, in het bijzonder in de vroege stadia van conversie in het vergassingsproces, namelijk de pyrolyse. Hier worden twee technieken besproken die langzame en snelle verhittingscondities vertegenwoordigen: respectievelijk TG-FTIR (thermogravimetrische analysie gekoppeld met FTIR), bij AFR Inc. (Hartford, CT, USA) en de heated grid reactor voorzien van in-situ IR diode laser diagnostiek, bij de afdeling Technische Natuurkunde van de Technische Universiteit Eindhoven. In hoofdstuk 4 worden de experimentele resultaten gepresenteerd en besproken. In de Delftse drukwervelbed experimenten zijn miscanthus- en houtpellets als biobrandstoffen gebruikt. Bruinkool is geselecteerd omdat het een oudere, maar toch nog hoog-vluchtige fossiele brandstof is. Er is geen significante gradiënt in het radiaal concentratieprofiel van hoofd- en sporecomponenten waargenomen in de vergasser. De concentraties van de hoofdcomponenten van het productgas waren vergelijkbaar met de beperkte gemeten open literatuurdata van andere drukwervelbed opstellingen bij vergelijkbare stoichiometrie. In de axiale gas concentratie profielen tijdens de Delftse drukwervelbed experimenten werden duidelijke gradiënten waargenomen voor acetyleen, een component welke gerelateerd is aan reakties die teer en roet-precursor vorming en afbraak betreffen. Bij de toegepaste drukwervelbed vergassingscondities is de belangrijkste brandstofgebonden stikstof component NH3, terwijl HCN in mindere mate wordt gevormd (enkele procenten van de vaste brandstofgebonden stikstof). De conversie naar NH3 en HCN was vergelijkbaar met andere FB vergassers met bodemvoeding, in contrast met de relatief lage waardes die werden waargenomen voor een drukwervelbed met topvoeding. HNCO and NO zijn nooit gedetecteerd, zelfs niet met een hogeresolutie FTIR spectrofotometer, onder de bestudeerde drukvergassingscondities. Experimenten uitgevoerd met Ca-houdend dolomiet en een Ca-loos additief (MinPhyl, of Pyrophylliet) onder vergelijkbare procescondities toonden aan dat een verhoogd Ca aanbod in de vergasser leidt tot een significante toename in de NH3/HCN verhouding. Ter verkrijging van de basis model inputgegevens, zijn heated grid flash pyrolyse experimenten uitgevoerd met miscanthus. Dit deelonderzoek was gericht op het bepalen van opbrengst van CO, CO2 and NH3 als functie van de temperatuur in de range 1050-1400K en opwarmsnelheden van 280320K/s. NH3 kon helaas niet worden gedetecteerd, hetgeen te wijten was aan ofwel condensatie of het beperkte frequentie gebied dat kon worden ingesteld met de tuneable laser. De CO en CO2 opbrengsten zijn in-situ bepaald en vergeleken met de FG-DVC biomassa pyrolyse model uitkomsten. Voor alle toegepaste brandstoffen zijn de kinetische parameters bepaald voor dit pyrolyse model door toepassing van de Tmax methode, uitgaande van TG-FTIR metingen bij verhittingssnelheden van 10, 30 en 100 K/min. De extrapolatie die dit model gebruikt om de pyrolyse productopbrengst bij hoge opwarmsnelheden te voorspellen gebaseerd op de kinetiek parameters die verkregen zijn bij lage opwarmsnelheden, levert een wisselend beeld op. Het resulteert in een redelijk correcte kwantitatieve opbrengstvoorspelling voor CO en een duidelijk te lage voorspelling voor CO2. De competitie tussen het vrijkomen van primaire producten als primaire teerfragmenten en carboxylzuren enerzijds en lichte gassen als CO, CO2 en H2O anderzijds is waarschijnlijk de oorzaak. Deze competitie is waarschijnlijk afhankelijk van de opwarmsnelheid. De primaire pyrolyse producten als teren en carboxylzuren, die precursor groepen bevatten voor CO en CO2 vorming, worden blijkbaar snel uit de reaktiezone verwijderd en direct afgekoeld. Er is dan dus geen tijd beschikbaar voor verdere ontleding in CO2 en (in mindere mate) CO, wat resulteert in lage productopbrengsten.

xii

Volgens deze hypothese moeten de opbrengsten van primaire teren en carboxylzuren significant zijn. Dit wordt bevestigd door de waarneming dat voor pyrolyse bij toenemende opwarmsnelheden de teer opbrengst toeneemt. In hoofdstuk 5 wordt de modellering van drukwervelbedvergassing behandeld. Hier wordt het model beschreven en de invloed van proces parameters bestudeerd. Het model is een stationair propstroomreactor-in-serie model met gedetaileerde reactiekinetiek. Heterogene kool-residue oxidatie en vergassing (door H2O, CO2 en H2), heterogeen gekatalyseerde HCN hydrolyse en homogene reacties (met inbegrip van stikstofhoudende moleculen en radicalen en een vereenvoudigde teer kraakreactie) worden in het model meegenomen. De mogelijkheid om NH3 conversie in N2 te vergroten door procescondities te variëren of door toevoeging van specifieke componenten aan de vergasser is theoretisch bestudeerd. Het blijkt dat NH3 een erg stabiele component is, welke nauwelijks kan worden omgezet in drukwervelbedvergassers. Decompositie in N2 neemt iets toe door verhoging van de temperatuur, hoewel deze optie beperkt is door het risico van bed sintering voor alkali-houdende biobrandstoffen. De NH3 conversie is slechts in beperkte mate afhankelijk van de druk. Een minimum in de conversie is gevonden bij 2.5 bar. Bij hogere drukken (10 bar) neemt de NH3 conversie licht toe. De gas verblijftijd in de reaktor heeft praktisch geen invloed op de brandstof-N conversie. Afbraak van NH3 vindt alleen plaats in de aanwezigheid van O, H en OH radicalen, die erg snel reageren in de initiële bedzone. Hogere NH3 conversies kunnen worden gerealiseerd door injectie van NO of NO2 in het bed. Anderzijds wordt dan HCN gevormd en wordt ongereageerd NO voorspeld, die beide ongewenste emissiecomponenten zijn. Additie van O2 bevordert de NH3 conversie. De belangrijkste stikstofcomponent wordt dan echter NO en in het meest ideale geval, bij praktisch stoichiometrische condities, wordt de brandstofgebonden N voor 50% in NO omgezet en voor 50% in N2. Toevoegen van secundaire lucht verlaagt de toch al lage LCV gas stookwaarde. Aanwezigheid van een hoge concentratie CH4 in de bedzone van de vergasser reduceert de NH3 conversie, waarschijnlijk door de competitie voor radikalen tussen CH4, haar intermediaire omzettingsproducten (voornamelijk CH3 radikalen) en NH3. H2O2 and H2O (stoom) additie in het bed heeft geen invloed op de NH3 conversie. Een vergelijking tussen de met behulp van het in hoofdstuk 5 beschreven model uitgevoerde simulaties en de experimenten uitgevoerd op de 1.5 MWth drukwervelbed schaal en de 50 kWth schaal wordt in hoofdstuk 6 gepresenteerd. De overeenkomst tussen berekende en gemeten waarden is voor de stikstofcomponenten tamelijk goed voor de brandstoffen en in het bijzonder voor de belangrijkste brandstofgebonden stikstofcomponent: NH3. Voor HCN wordt de concentratie vaak te laag voorspeld, mogelijk door de heterogeen gekatalyseerd hydrolyse reaktie, welke plaatsvindt op het kool-residue oppervlak en die een langzamere reaktiekinetiek zou kunnen hebben dan aangenomen is in de literatuur. Het model voorspelt ruim 10 ppmv HNCO en NO, maar deze stoffen zijn nooit door middel van FTIR analyse gedetecteerd. Waarschijnlijk wordt HNCO katalytisch in NH3 en CO omgezet en deze reaktie is niet in het model meegenomen. Voor NO speelt katalytische reductie door as-elementen waarschijnlijk een rol bij de heersende vergassingstemperaturen. Het is ook mogelijk dat verwaarlozing van de S en Cl chemie voor de trace componenten de afwijkingen tussen model en experimenteel resultaten veroorzaakt. Dit is in mindere mate het geval voor hout. De vereenvoudiging van de teer- en koolresiduegebonden stikstof kan ook van betekenis zijn: in het model wordt aangenomen dat de stikstof die niet vrijkomt als gas component bij de pyrolyse, initieel als HCN vrijkomt. Voor de hoofdcomponenten van het productgas is de overeenkomst tussen de model- en experimentele resultaten tamelijk goed. De verschillen tussen de berekeningen en de metingen kunnen worden toegeschreven aan de op basis van de literatuur aangenomen pyrolyse opbrengst van waterstof in het biomassa pyrolyse sub-model. De onzekerheid aangaande de teerontledingskinetiek, in termen van mogelijke reacties, product opbrengsten en snelheid, speelt waarschijnlijk ook een prominente rol in de waargenomen verschillen voor CO en CO2.

xiii

De afwijkingen zijn het hoogst voor hout en minder voor miscanthus en bruinkool. Dit komt overeen met de hypothese dat onnauwkeurigheden in de teerontledingskinetiek hier een belangrijke rol in spelen, omdat hout de hoogste initiële teeropbrengst vertoont in de snelle pyrolyse stap. Verschillen in de heterogene verbrandings- en vergassingsreactiekinetiek hebben naar verwachting een veel minder grote invloed op de voorspelde concentraties, aangezien de koolstofconversie in het algemeen tamelijk goed wordt voorspeld. Tenslotte wordt in hoofdstuk 7 een overzicht gegeven van de getrokken conclusies en aanbevelingen gedaan voor verder onderzoek, gespecificeerd naar experimentele drukwervelbed pilot schaal onderzoek, brandstofkarakaterisering en modellering.

Wiebren de Jong

xiv

Table of Contents Summary

vii

Samenvatting

xi

Notation

xix

Chapter 1: Introduction 1.1 1.2 1.3

1.4 1.5 1.6

Towards a renewable energy based world Biomass as part of renewable power generation Technology for biomass utilisation for heat and power generation 1.3.1 Small scale versus large scale processes 1.3.2 Technological options for large scale biomass based heat and power generation 1.3.2.1 Combustion 1.3.2.2 Gasification Power production from biomass gasification, open research questions The fate of fuel bound nitrogen Outline of this thesis

1 3 3 4 5 5 6 8 9 11

Chapter 2: Fluidised bed solid fuel gasification processes, overview and analysis of experimental research and modelling 2.1 2.2 2.3 2.4

2.5

2.6 2.7

The fluidised bed reactor applied for solid fuel gasification Industrial fluidised bed gasification systems An overview of recent research, development and small scale demonstration activities Experimental findings regarding the fate of fuel nitrogen during fluidised bed gasification 2.4.1 Influence of fuel type 2.4.2 Influence of air stoichiometry 2.4.3 Influence of temperature 2.4.4 Influence of pressure 2.4.5 Influence of additives 2.4.6 Influence of particle diameter 2.4.7 Influence of steam 2.4.8 Influence of feed location Fluidised bed gasifier modelling 2.5.1 General overview 2.5.2 Drying and flash pyrolysis, initial steps in the process 2.5.2.1 Experimental techniques and findings 2.5.2.1.1 Main components and hydrocarbons 2.5.2.1.2 Nitrogen components 2.5.2.2 Modelling approaches 2.5.3 Heterogeneous char-gas reactions 2.5.3.1 Main carbon based reactions 2.5.3.2 Heterogeneous and heterogeneously catalysed homogeneous nitrogen reactions 2.5.4 Homogeneous gas phase reaction mechanisms, including nitrogen chemistry Potential primary measures for fuel_NOx emission reduction Conclusions and research requirements

13 13 17 20 20 23 24 25 26 26 27 27 28 28 32 32 33 53 60 62 62 65 66 68 70

xv

Chapter 3: Experimental set-ups and measurement techniques 3.1 3.2

Introduction The Delft Pressurised Fluidised Bed Gasification (PFBG) test rig 3.2.1 Description of the rig 3.2.2 Analysis and sampling techniques 3.2.2.1 FTIR spectrophotometer 3.2.2.2 Gas Chromatography 3.2.2.3 On-line Non Dispersive Infrared/UV, colorimetric and paramagnetism based analysers 3.2.2.4 Sampling probes and analysis of tar compounds

71 71 71 75 75 81

3.3

The 50 kW(thermal) IVD Pressurised Fluidised Bed Gasifier (DWSA) 3.3.1 Description of dimensions and operation 3.3.2 Analysis techniques applied

89 89 90

3.4 3.5

The TG-FTIR set-up at Advanced Fuel Research Company Inc. (USA) The heated grid reactor at Eindhoven University of Technology

91 92

83 86

Chapter 4: Experimental results 4.1

4.2

4.3

4.4

4.5

4.6

xvi

Choice of fuels, bed materials and additives 4.1.1 Fuel choice 4.1.2 Fuel composition and related chemical properties 4.1.3 Physical property characterisation of fuels and bed materials Experimental results of PFBG gasification tests 4.2.1 Experimental data representation and definitions of relevant parameters 4.2.2 Background information on the measurement campaigns 4.2.3 Miscanthus gasification 4.2.4 Wood gasification 4.2.5 Brown coal gasification Experimental results of DWSA gasification tests 4.3.1 Overview of the DWSA measurement programme 4.3.2 Wood gasification 4.3.3 Brown coal gasification Experimental results of TG-FTIR pyrolysis tests 4.4.1 Overview of the TG-FTIR experimental programme 4.4.2 Kinetic analysis approach 4.4.3 TG-FTIR analysis results and derived kinetic parameters for miscanthus 4.4.4 TG-FTIR analysis results and derived kinetic parameters for wood (Labee “A quality” energy pellets) 4.4.5 TG-FTIR analysis results and derived kinetic parameters for brown coal 4.4.6 TG-FTIR analysis discussion Experimental results of heated grid pyrolysis tests 4.5.1 The heated grid experimental programme 4.5.2 Miscanthus pyrolysis results 4.5.3 Discussion of the results Conclusions and recommendations 4.6.1 Conclusions and recommendations related to PFB gasification 4.6.2 Conclusions and recommendations for fuel characterisation

99 99 100 101 104 104 105 106 117 124 129 129 130 131 133 133 133 135 140 144 147 149 149 149 153 155 155 156

Chapter 5: Modelling bubbling fluidised bed gasification, focussed on nitrogen compounds 5.1 5.2

Modelling approach Description of the model 5.2.1 Idealised reactor approach Simulation results using the idealised reactor modelling approach Conclusions

5.3

5.4

159 159 159 169 182

Chapter 6: Comparison of modelling results and experiments 6.1 6.2

Choices made for the comparison between model and experiments Gasification experiments compared with model results 6.2.1 TUD PFBG experimental and simulation results 6.2.2 IVD DWSA experimental and simulation results Discussion of the results Conclusions and recommendations

6.3 6.4

183 183 183 192 196 197

Chapter 7: Conclusions and recommendations for further research 7.1 7.2

Conclusions Recommendations

Bibliography

199 201 203

Appendices 1

Details of analytic measurements 1.1 Spectra, spectral windows and calibration curves used for quantitative species analysis with FT-IR 1.2 Calibration curves used for quantitative species analysis with gas chromatography

2

Relevant chemical & physical properties of the gasification product gas components

227 227 241

2.1 Gas phase viscosity 2.2 Diffusion coefficients of gas phase components 2.3 Gas Phase Thermal Conductivity 2.4 Thermodynamic data

249 249 250 251 252

3

Detailed homogeneous reaction scheme “Kilpinen 97”

255

4

Results of TG-FTIR measurements and comparison of FG-DVC model results with experiments at different heating rates

259

Calculation of axial Péclet numbers for PFBG and DWSA tests simulated

275

List of publications

277

5 6

Dankwoord

281

Curriculum Vitae

283

xvii

Notation Abbreviations ABFB ACFB AFR ar or a.r. BC BHF BTX CFD daf db FBN FG-DVC FID FTIR GC Gtoe HCV or HHV HGR HRSG IGCC LCV or LHV mf MSW Mtoe NDIR NDUV ODE PFB PBFB PCFB PFBG RDF RPS SCADA SCO SCR SPA SPE SRC SS TCD TGA TUD WP

Atmospheric Bubbling Fluidised Bed Atmospheric Circulating Fluidised Bed Advanced Fuel Research inc. As received Brown Coal Baghouse Filter Benzene, Toluene, Xylene Computational Fluid Dynamics Dry and ash-free basis Dry basis Fuel Bound Nitrogen Functional Group-Depolymerisation Vaporisation Condensation Flame Ionisation Detector Fourier Transform InfraRed Gas Chromatograph Gigaton oil equivalent Higher Calorific Value Heated Grid Reactor Heat Recovery Steam Generator Integrated Gasification Combined Cycle Low Calorific Value Moisture free Municipal Solid Waste Megaton oil equivalents Non Dispersive InfraRed Non Dispersive UltraViolet Ordinary Differential Equation Pressurised Fluidised Bed Pressurised Bubbling Fluidised Bed Pressurised Circulating Fluidised Bed Pressurised Fluidised Bed Gasifier Refuse-derived Fuel Rotating Particle Separator Supervision Control and Data Acquisition Selective Catalytic Oxidation Selective Catalytic Reduction Solid Phase Adsorption technique Solid Phase Extraction Short Rotation Coppice Stainless Steel Thermal Conductivity Detector Thermogravimetric Analysis (or Analyser) Delft University of Technology Wood Pellets

Latin Symbols A A0

Area of gasifier reactor part Area of nozzle holes

[m2] [m2]

xix

A(υ) ai’(υ) Bo b b ci’ cp CC d Dax DT Di,m E f g Ga H H I ki ki0 mi M MWi OF OM P Pe R Re t T u v x X Yi

Absorbance at wave number υ Absorption coefficient at wave number υ of species i' Bodenstein number Path length through the sample Temperature coefficient in the Arrhenius equation Concentration of species i' in the sample Specific heat Carbon Conversion Diameter Axial dispersion coefficient Diameter of gasifier reactor part Diffusion coefficient of component i in mixture Activation energy Radiation frequency Acceleration of gravity (= 9.81) Galilei number Enthalpy Height of bed Intensity of radiation Rate constant of component i Frequency factor (or pre-exponential factor) Mass fraction of i in fuel Heating rate Mole mass of component i Distance between beamsplitter and fixed mirror Distance between beamsplitter and movable mirror Pressure Péclet number Universal gas constant (=8314.3) Reynolds number Time Temperature Velocity Velocity Fraction of reacted material Mass fraction of char remaining Mass fraction of component or functional group i

[-] [m-1] [-] [m] [-] [-] [J.kmol-1.K] [%] [m] [m2.s-1] [m] [m2.s-1] [J.kmol-1] [s-1] [m.s-2] [-] [J.kmol-1] [m] [*] [s-1] [s-1] [-] [K.s-1] [kg.kmol-1] [m] [m] [Pa] [-] [J.kmol-1.K-1] [-] [s] [K] [m.s-1] [m.s-1] [-] [-] [-]

ν ρ ρ σ φ Φm

Retardation, or difference in optical path length Volume fraction Dynamic viscosity Efficiency Wavelength Wave number Kinematic viscosity Density Parameter in the empirical relation for σ Gaussian distribution of the activation energies Sphericity factor bed material Μass flow

[m] [-] [Pa.s] [%] [m] [cm-1] [m2.s-1] [kg.m-3] [−] [J.kmol-1] [-] [kg.s-1]

ωj

Reaction rate of component j

[kmol.m-3.s-1]

Greek Symbols δ ε η η λ

υ

.

xx

Indices A ax b bed C diff e f g m max mf mix p r r s SMD Stoich th x

Absolute Axial direction Bubble Bed (section) Carbon (char) Diffusion Electrical Forward reaction Gas (phase) Mass Maximum At minimum fluidisation condition Mixture (gas) Particle Residence Backward reaction Solid (phase) Sauter Mean Diameter Stoichiometric Thermal In axial direction

xxi

Chapter 1 Introduction 1.1 Towards a renewable energy based world Up till now, fossil fuel utilisation has made a large contribution to the energy systems of our modern age. However, most of the fossil fuel reserves for power and heat supply, transportation and chemicals manufacturing decrease and need to be substituted by alternatives in order to maintain the present way of life. The 20th century was based on these conventional energy carriers, but the 21st century will have to adjust to the decline of the oil and natural gas based economy and the increasing public awareness of the negative effects of environmental pollution from fossil fuel utilisation.

Coal 606 Gtoe

51 Gtoe

Uranium

Uranium

57 Gtoe

Figures 1.1a through 1.1d give an overview of recent data concerning fossil fuel and uranium reserves, as well as the expected cumulative demand over the period 1990-2050.

Oil 343 Gtoe

411 Gtoe

Coal 710 Gtoe

Natural Gas 333 Gtoe

Coal

Natural Gas 237 Gtoe

0.65 Gtoe

Figure 1.1b Global primary Energy Reserves Commercially and technically exploitable, World Economic Outlook, [IEA, 1998].

Uranium

40 Gtoe Uranium

Figure 1.1a Global primary Energy Reserves, Commercially and technically exploitable, Global Energy Perspective, [IIASA/WEC, 1998].

Oil

Oil 261 Gtoe

Coal 2.13 Gtoe

273 Gtoe Natural Gas

Natural Gas

211 Gtoe

2.06 Gtoe

Figure 1.1c Cumulative primary energy demand (1990-2050) Energy Reserves, [IIASA/WEC, 1998].

Oil 3.46 Gtoe

Figure 1.1d Primary energy utilisation in 1999, [BP Amoco, 2001].

Figures 1.1a-c show that reserves of oil and natural gas are (just) sufficient for the indicated period. Based on previous experience, however, it is expected that more reserves of natural gas will be found in e.g. hydrate fields (clathrates) in deep-sea area’s, but the quantities still have to be investigated.

1

The exploitation of these fields, though, will be comparatively expensive. Also for oil there are probably unknown reserves, which are by now not yet commercially or technically exploitable. Coal as conventional fuel is expected to be still available for several centuries because of its huge reserves and its broad global reserve distribution. The figures further indicate that the reserves of Uranium, the fuel for power generation based on nuclear fission, are of the same order of magnitude as the expected demand in the indicated period. This situation is not positive for the long term use of this fuel, unless technology is shifted to a more efficient use of fuel sources, like e.g. in fast breeder systems [Gardiner, 1990]. However, radioactive waste disposal and removal problems are still preventing the widespread using of nuclear technology. Also, the danger of proliferation of nuclear energy knowledge and implementation on a global scale might lead to increased nuclear arms proliferation. Furthermore, the world is faced with a growing human population and an increasing level of industrialisation, giving rise to an increasing demand for personal comfort and energy demand per capita. Long term scenario’s, like the one mentioned in figure 1.1 c, differ in the way the energy use per capita increases, but in all scenario’s 50-60% of the world’s energy supply will rely on fossil sources as primary energy supply. The use of fossil fuels as primary energy source gives rise to an increasing emission of environmentally hazardous species, which is becoming a growing problem. The rising worldwide awareness of environmental constraints leads to legislative actions in most industrialised countries in order to restrict local and regional emissions of acid rain or smog precursor gases (e.g. SOx and NOx) and dust. Furthermore, several industrialised countries are heading for a reduction of the human contribution to the greenhouse effect, by reducing emission of these so-called greenhouse gases, like in particular CO2, CH4 and N2O. Of these gases, CO2 is the most important contributor to the absorption of infrared radiation emitted from the earth’s surface. The increase of the CO2 concentration in the atmosphere is supposed to contribute significantly to the enhanced greenhouse effect (from 55% [Wójtowicz et al., 1993] to 63.5% [Sloss, 2002]). CH4 also absorbs infrared radiation and contributes for approximately 15% -20.5% to the enhanced greenhouse effect. Finally, N2O has a contribution of about 6.5% to the enhanced greenhouse effect. In the Kyoto-conference [UNFCCC, 1997], held in December 1997, it was agreed to reduce worldwide greenhouse gas emissions of the industrialised nations by 2012 to a 5.2% lower level compared to the emissions in 1990. For the Dutch situation this agreement implies that in the period 2008-2010 compared to the 1990 level 6% less greenhouse gases will be allowed to be emitted according to the “Uitvoeringsnota Klimaatbeleid”[Duurzame energie in uitvoering, 1999]. This will have to be realised under higher economic grow rates and accompanied with a higher consumption level, which potentially causes increasing greenhouse gas emission levels [Energierapport, 1999]. Therefore, a reduction of the use of fossil fuels, which are the major source of CO2 emission, is required. In order to reach the goal of decreasing CO2 emissions, growing attention is drawn toward two directions. The first is to improve the efficiency of existing fossil fuel conversion processes. The second is the use of sources that are practically CO2 emission free, like wind, solar, geothermal, hydropower, biomass and advanced nuclear sources [Gardiner, 1990]. Biomass is ‘almost CO2 neutral’ and renewable, because CO2 is taken up from the atmosphere during the growth of biomass and released again during combustion in a relatively short cycle time, as compared to fossil fuels. An increased use of renewable energy sources can support governments to achieve a wide range of policy goals: e.g. improved energy security and diversity, enhanced levels of technology export to countries which are less developed and reduced emissions of greenhouse gases and other pollutants, such as sulphur oxides, nitrogen oxides, particulates and trace metals. In the world, the USA is the country with the highest energy demand per capita. According to the International Energy Outlook [IEO, 2000], the share of renewables in the US electricity market is projected to be about 20% in 2020. The majority of this renewable power production will be hydroelectrity. Biomass will provide a smaller contribution to the US power supply, which is currently approximately 4% of the primary energy [Costello & Chum, 1998]. The US department of Energy (DOE) has published a figure of 1% use of grid-connected biomass power capacity, which is about 7 GW. A goal set by the US government is to implement an additional 17 GW biomass power the next two decades [DOE, 1996]. 2

The European Union recently published a White Paper on Renewable Energy, stating that Europe could double its use of renewable fuels from 6 % in 1997 to 12% by the year 2010. Currently, in Europe energy from biomass sources accounts for about 45 Mtoe (approximately 3% of the total consumption), while the European Commission proposes that biomass in total will contribute an additional 90 Mtoe per year by 2010 with an increased share in the total energy consumption [EC, 1997]. The Dutch government has agreed to a goal of 10% energy from renewable sources for the year 2020, with biomass as the main contributor [Weterings et al., 1999]. This implies about 120 PJ of bioenergy. Recent expectations are such that in ten years’ time biomass for energy supply could increase from a current level of approximately 13 PJ to a level of 80 PJ [Energierapport, 1999]. For this purpose, the Dutch government uses several instruments in a liberalising electricity market, such as offering fiscal advantages, to stimulate the demand of green electricity as well as the production of energy from renewable sources [Kwant & Leenders, 1999]. 1.2 Biomass as part of renewable power generation Since the dawn of mankind, biomass has been used as food for life and energy for heating and cooking by combustion. It is a form of solar energy stored in organic form. Nowadays it is recognised that the use of biomass for supply of energy offers the advantage of reducing the net CO2 emission. Power generating facilities based on fossil fuels, of which coal is the most important component, cause comparatively high specific emissions of the greenhouse gas CO2. Today around 40% of the world’s electricity generation is based on coal and 20% of the CO2 emissions are caused by coal-fired power plants [Campbell et al., 2000]. Biomass can nowadays be utilised in quite efficient processes for power and heat production, which have been developed to industrial scale in the last decades. For this purpose, biomass from energy crop cultivation, set-aside land can be used in an economic way in certain countries. This helps to restructure the agricultural situation with the benefit of job creation along with the agro-economic activities for this purpose. A valuable bio-diversity can be created along with erosion protection, depending on the crop used. Drought tolerance and low fertiliser and/or pesticide requirements are important requirements [Hall et al., 1993], [DOE, 1996], [Faaij, 1997]. Also, (industrial) waste streams of biological origin can be used, thereby reducing disposal problems or incineration, which is inefficient compared to electricity (co-) generation. The current commercial and non-commercial biomass use for energy production is estimated to be between 6 and 17% of world primary energy, most of this is used in developing countries, where biomass accounts for up to one third of energy needs [Bauen & Kaltschmitt, 2001]. By contrast, biomass provides at most 3% of energy in industrialised countries [Gross et al., 2003]. The utilisation of solid fossil together with biomass for heat and power production offers additional advantages compared to the use of biomass as a single fuel [Rüdiger et al., 1996]:

• variations in the availability of biomass can be met by changing coal-biomass ratio’s; • it enables a wider range of system sizes, increasing optimisation possibilities; • it can show synergistic effects that reduce operational problems and emissions. 1.3 Technology for biomass utilisation for heat and power generation An important aspect for the choice of the technology to convert (biomass) fuels is the cycle efficiency for power and heat generation. Efficiency improvements decrease specific fuel consumption and emissions. Available energy conversion systems for solid fuels have thermodynamic limitations with respect to attainable efficiencies. Power generation in conventional systems is based on either steam turbine or gas turbine technology. The application of advanced efficient systems such as fuel cells is interesting

3

for the longer term but these are not considered in this study as they are presently comparatively expensive and not technically feasible for large scale biomass/waste based conversion processes. The first systems used at commercial scale were based on a steam turbine cycle with direct combustion of the fuel. In this conversion process heat is transferred from a high (combustion) temperature to a level determined by the maximum allowable steam conditions (pressure and temperature). These conditions are mainly determined by material strength constraints. Due to the relatively large difference in the temperatures mentioned (combustion and steam temperature), the process efficiency is limited thermodynamically by the upper temperature level of the steam cycle. A gas turbine can be used efficiently for generating power using high temperature working fluids, like flue gas from solid fuel or gasification-derived gas combustion. One of the limitations in power generating efficiency in a gas turbine (or Brayton) cycle is in the lower temperature level to which heat can be released. A combination of a gas turbine and a steam turbine, a so-called combined cycle, leads to potentially higher efficiencies as compared to the separate steam and gasturbine cycles. In the following sub paragraph, a comparison is made between small and large-scale systems for power and heat production, as these have their own advantages and disadvantages. 1.3.1 Small scale versus large scale processes Electric power from biomass can be generated in a decentral or central way. Large scale centralised power production offers the following main advantages (with specific biomass related points indicated in italic): + the availability of electricity is well secured; + the economics of scale cause comparatively low power production costs; + emissions can be controlled adequately, due to the gas cleaning equipment being relatively well known on large scale and operated by skilled personnel. Disadvantages are: − users are dependent on a grid, which is sometimes not reliable in developing countries; − significant losses in the distribution and transportation of electricity, although in a densely populated country, like e.g. the Netherlands, this loss effect can be ignored; − transportation distances of fuel are relatively long, with accompanied increased costs and CO2 emission; − the availability of especially biofuels can be insecure. In the decentralised option, which was also characteristic for traditional biomass use, electricity is produced by small scale units in local communities, or farms located far away from densely populated areas. This has the following advantages:

+ + +

users are not dependent on the electric grid, which especially in countries in the developing world is not always secured, a critical issue for e.g. hospitals; transportation and distribution losses of electricity are relatively small; the average distance of transportation of biomass fuels is short.

Disadvantages are: − emissions of small local power producing equipment cannot be controlled as adequate as those of large scale power plants; − maintenance of the units requires special skills and is possibly less secured as compared to large-scale power production; − fuel conversion, gas cleaning (both for protection of downstream equipment and emission reduction) and prime mover equipment is relatively expensive due to the negative economics of scale;

4

−

availability is not always secured, as skilled personnel is not always there and also because there can be problems in biomass supply.

In the Dutch situation the advantages of large-scale systems outweigh the disadvantages and therefore the focus of this study will be on large-scale systems. 1.3.2 Technological options for large scale biomass based heat and power generation For large scale heat and electricity generation using biomass, the main technological options are combustion or gasification. Pyrolysis, or liquefaction in combination with a combined cycle e.g. is still in a conceptual stage [Siemons, 2002]. These techniques are more interesting for smaller scale decentralized heat and power production (scale-up is still under development) or for producing biofuels for transportation. 1.3.2.1 Combustion Co-combustion of biomass in large-scale utility boilers with, as additional option, co-production of electricity and heat is already a commercial common practice [Spliethoff, 2000]. The energy conversion systems based on combustion basically consist of a primary combustion section, a boiler section and a steam turbine. The combustion section ensures conversion of the chemical energy bound in the solid fuel into heat in the presence of an overall over-stoichiometric amount of air. This heat is transferred by radiation and convection to the water and the steam in the boiler and sensible heat of the flue gases; with conduction playing a minor role as heat transfer mechanism. The steam is expanded in a steam turbine to drive the generator for electricity production. Part of the energy content in the steam that cannot be converted into power, generates heat on a relatively low temperature level and can be used for e.g. district heating provided that a consumer net is commercially possible. Often the heat produced cannot be used to its full extent. The amount of power generated divided by the amount of heat produced of most biomass-fired power stations amounts to values less than 0.5 [Van den Heuvel & Stassen, 1994]. Depending on the local requirements variable amounts of heat can be utilized, but comparatively high electric power generation efficiency is the driving force in most situations. The electric efficiency of common steam cycles is determined by:

• the temperature of steam at the turbine inlet; the higher the temperature (accompanied with a higher • • • •

pressure) the higher the efficiency; the maximum temperature is determined by the material properties of the steam generator; the steam pressure at the turbine outlet (condenser pressure), which is determined by the temperature to which the outlet steam can be cooled down; the lower this temperature, the higher the overall efficiency; thermal losses of the boiler system; combustion efficiency; generator losses.

There are several commercial combustion system configurations available for both biomass and solid fossil fuels, with the following reactors as primary combustor types: • fixed bed combustors (practically not applied anymore for solid fossil fuels), like e.g.: fixed flat or inclined grate firing units; moving flat or sloped grate ovens; combustors with bottom screw feeding; • fluidised bed combustion reactors, such as: bubbling beds; circulating beds; • other reactor systems, with as examples:

5

pulverised solid fuel burners (entrained flow reactors); cyclone burners. When compared to other combustion technologies fluidised beds offer the advantages of: • very high flexibility with respect to fuel properties like size distribution, density, moisture and ash content; • good heat transfer leading to installations with relatively small specific reactor volumes; • in-situ primary gas cleaning by addition of sulphur binding compounds, like e.g. limestone, so that downstream cleaning is less extensive; • relatively low reactor temperatures, leading to less corrosion, deposition problems and potentially lower NOx and SOx emissions than high temperature systems. Problems of these fluidised bed combustion systems are fouling and slagging of heat exchanging equipment in the system (boiler) and sintering tendencies of ash and bed material especially with fuels containing alkali metals and chlorine. Also, for fluidised bed combustion of coal comparatively high N2O emissions have been observed. 1.3.2.2 Gasification An alternative to direct combustion of biomass and/or coal for conversion of fuel into power and heat is conversion by thermal gasification. In principle there are three basic gasification reactor configurations: the fixed bed gasifier (dry ash or slagging), which can be distinguished into three subtypes: • the co-current down flow reactor; • the counter-current reactor, and • the cross-current gasifier; the fluidised bed gasifier (non slagging operation), which can be divided into two sub configurations: • the bubbling fluidised bed and • the circulating fluidised bed; the entrained flow gasifier (slagging). The process can be configured as autothermal or allothermal. In the first, most commonly found configuration, the heat necessary to drive the endothermic gasification reactions is provided in-situ by partial combustion of the fuel. For the latter configuration, the heat needed for the endothermic reactions is provided externally. Application of oxygen/steam as gasification medium is common for coal gasification, because smaller equipment can be applied when compared to air blown gasifiers. Gas cleaning is a critical issue for the successful implementation of gasification technology for electricity generation. The combination of the gasifier, down stream followed by high temperature dry gas cleanup processes offers the advantage of higher overall efficiencies than systems using wet, low temperature, raw gas cleaning techniques. This holds especially for downstream applications that are using hot product gas. Relatively high power efficiencies can be attained by application of Integrated Gasification & Combined Cycle (IGCC) technology. High efficiency, natural gas fired combined cycles are applied on a large scale throughout the world with natural gas as fuel. These units require comparatively low investments. The IGCC process, characterised by a relatively high power to heat production, has been demonstrated already for coal as solid fuel on large scale and seems to be also especially attractive in the medium-size power production range of 20-150 MWe [Kurkela et al., 1993b]. Examples of technically successful coal-based IGCC demonstration projects in Europe are "Buggenum" (253 MWe, see figure 1.2 a and b) [Ploeg, 2000] and "Puertollano" (300 MWe) [Schellberg, 2000]. In the USA, coal-based IGCC projects are: "Polk Country" (250 MWe, 4.5 year demonstration project ended in 2001, now continuing to operate commercially), "Wabash River" (262 MWe, 5 year demonstration project ended in 1999, now continuing to operate commercially) and 6

"Piñon Pine" (100 MWe), see e.g. [Campbell et al., 2000]. The first two mentioned can be characterised as technically successful, with availability data as high as 70-80%; the last one has ended less successfully [Henderson, 2003]. Furthermore, also in Japan and Australia IGCC projects based on coal have been launched [Henderson, 2003].

Figure 1.2 a. Picture of the Buggenum IGCC power station; [Scheibner&Wolters, 2002].

b. Process scheme of the plant.

The IGCC process comprises a gasifier with appropriate product gas cleanup and both a gas turbine and a steam turbine for electricity generation. The competitiveness of the process alternatives has been based on the overall price of electricity, discarding the environmental benefits related to IGCC technology. In the future, the environmental factors and sustainability aspects may probably be of higher importance and competitiveness of IGCC systems based on a variety of solid fuels will then be significantly improved. However, the competitiveness of IGCC cannot be based only on environmental superiority. The overall feasibility of IGCC technology also has to be improved. This can be attained by developing and improving equipment components of IGCC units. In addition, the fuel flexibility has to be further improved – one can consider all kinds of waste material in this respect - accompanied by reduction of operational costs. Fluidised bed technology is an attractive process for gasification in comparison to the fixed bed or entrained flow alternatives, because of: + the well-established principles of operation; + the fact that the process can be scaled successfully to the medium-size power range ; + its flexibility with respect to feedstock characteristics (type, particle size); + the favourable heat and mass transfer properties; + the energy-efficiency of the process?. A main disadvantage is that the fluidised bed process is limited in the temperature range to be applied, due to possibility of bed material sintering, which causes fuel conversion to be several percent points less than 100%. This disadvantage holds most for relatively unreactive fuels, such as bituminous coals, but not so for reactive biomass and brown coal. The most attractive gasification medium in the fluidised bed gasification process is air since no expensive air separation unit is necessary. A pressurised gasification process has as a major advantage that compression of the product gas, necessary for combustion in (advanced) gas turbines, is not required to such an extent as for atmospheric gasification. Compression of air for the gasifier is in this sense cheaper and easier to accomplish than compression of the fuel gas. Finally, reduced size and, hence, less costly equipment is sufficient for the gasifier and gas cleanup section as compared to atmospheric gasification processes.

7

As pressure resistant materials are more costly this argument holds for scales in a range from 20-30 MW electric and higher. As an example the process scheme of the 18 MWth pressurised fluidised bed IGCC demonstration plant at Värnamo is given in figure 1.3. The commercialisation of combined cycle power generation based on the gasification of solid biomass fuels has been slower than expected about one decade ago. An overview of these processes is given in chapter 2.

Figure 1.3 Process scheme of the Värnamo pressurised fluidised bed IGCC demonstration unit [Ståhl, 2001] 1.4 Power production from biomass gasification, open research questions There are still partly unanswered questions, which have to be answered by means of measurement and process modelling. These questions are related to the fate of species harmful to equipment and environment based on biomass gasification with fluidised beds as primary solid fuel converters. They are summarised below:

• What is the mechanism of the formation and destruction of higher hydrocarbon species (tars), which can clog process equipment parts like fuel control valves and gas analysis equipment and can cause a decreased downstream gas turbine combustion efficiency and soot emission? • What is the fate of trace elements; will they be emitted to the environment in solid form as ashes (from bed and fly ash) or in gaseous form (air toxics) in the exhaust gas? • How is fuel bound sulphur partitioned in the gasifier into mainly H2S and COS and subsequently SOx during gas turbine combustion? • What is the fate of nitrogen species (like NH3 and HCN) in NOx formation in processes including gas turbine combustors downstream of the gasifier, when dry, high temperature gas cleaning processes are applied [Hoppesteyn, 1999]? The first open research question is not dealt with in this thesis. The second is the subject of another thesis from our institute [Ünal, 2005]. Regarding the remaining questions a further consideration is given below. Table 1.1 gives an overview of typical elementary compositions of a selection of coal and biomass species. The biomass examples are from agricultural waste origin (straw types), energy crop cultivation (miscanthus) and forestry.

8

Coals presented in the table below are brown coal (from the German Hambach open mining) and black bituminous coals from the UK (Daw Mill coal) and the USA (Pittsburgh and Utah coal). Data for peat (from eastern Finland), a fuel intermediate between biomass and coal, is also given. Table 1.1 Fuel composition of several solid fuel sources (on dry basis) related to their nitrogen and sulphur species emission potential. Biomass Rice straw 1) Miscanthus 2) Sawdust 3) Wood pellets (‘clean’)2) Peat and Coal Finnish peat 4) Hambach Brown coal 2) Daw Mill coal Pittsburgh coal 1) Utah coal 1) 1)

C

H

O

N

S

HHV

N

S

(mass%, dry)

(mass%, dry)

(mass%, dry)

(mass%, dry)

(mass%, dry)

(MJ/kg)

(kg/GJ)

(kg/GJ)

39.2 48.2 48.5 51.4

5.1 5.4 5.1 6.0

35.8 42.8 46.0 42.2

0.6 0.58 0.03 0.17

0.1 0.16 0.03 0.11

15.4 19.3 19.2 20.3

0.4 0.3 0.02 0.08

0.07 0.08 0.02 0.05

54.5 64.5 69.4 75.5 77.9

5.6 4.4 4.4 5.0 6.0

33.6 26.1 10.0 4.9 9.9

1.80 0.64 1.2 1.2 1.5

0.25 0.39 1.6 3.1 0.6

21.8 25.2 30.0 31.8 33.0

0.8 0.3 0.6 0.4 0.5

0.1 0.2 0.8 1.0 0.2

[Reed,1981] 2) this study 3) [Zhou,1998] 4) [Kurkela et al., 1992]

Coal is a sedimentary rock composed of both organic and inorganic constituents and formed through partial decomposition of plant debris under the action of heat, pressure and time, [Bend, 1992], [Van Krevelen, 1993]. The main difference between coal and biomass is the coal’s higher heating value, which is closely related to the oxygen content in the fuels. During carbonisation, oxygen and hydrogen are reduced and coals of increasing rank from lignite to anthracite are formed, see [Speight, 1983]. Here, coal rank is defined as an indicator for the stage of alteration, or degree of coalification, attained by a particular coal. The greater the alteration, the higher the rank of the coal. Related to this difference in elementary composition is the much higher reactivity of biomass as compared to coal in thermochemical conversion processes, because less stable oxygen containing structures are present in the fuel. Absolute sulphur content is higher in the coals presented than in biomass types shown. This is also the case, although somewhat less pronounced, when the sulphur contents are compared on an energy basis. As sulphur species emission for biomass gasification is not a major problem for a broad range of these fuels, this issue is not further studied in this context. Although the absolute values of the nitrogen content on mass basis in biomass materials in most cases is significantly lower than coal, the nitrogen quantity on an energy basis is comparable for certain biomass species, like miscanthus and straw. This indicates the relatively high emission potential of significant nitrogen containing species, like NH3 and HCN as NOx precursors. The chemical structures in which the nitrogen is bound, however, are of importance too for the still not completely understood partitioning behaviour. This is the reason why the fate of nitrogen species will be the focus of this thesis. 1.5 The fate of fuel bound nitrogen The emission of NOx, a collective noun for NO and NO2, is a worldwide occurring regional problem as its effects are both on the environment (plants, buildings) as well as human and animal health. One of its effects is contribution to photochemical smog formation, in which NOx, CH3COO2NO2 (peroxylacetyl nitrate, PAN) and O3 play a role [Leighton, 1961]. Smog affects health negatively and reduces visibility. NOx also participates in the formation of O3 in the troposphere causing problems for vegetation. It is also a contributing agent in the greenhouse effect. Nitrous oxide, N2O, is known as “laughing gas”. When it is inhaled, it can give uncontrollable laughter, further inhalations lead to enjoyable hallucinations and longer exposure leads to anaesthesia [Hayhurst & Lawrence, 1992]. N2O has a relatively long lifetime in the troposphere of approximately 150 - 170 years [Badr & Probert,

9

1992] and [Hayhurst & Lawrence, 1992]. It can therefore be transported to the higher regions of the stratosphere, where it is destroyed by photolysis and a reaction with O atoms to NO as important product. NOx herein acts as a catalyst in stratospheric ozone destruction via these reactions [Badr & Probert, 1993a]. Finally, NOx contributes also to eutrophication and to acid rain or deposition, which negatively influences plant growth, aquatic life and fertile soil, as well as promotes corrosion and erosion of man-made constructions. It is estimated that NOx contributes about 30% of the acidity of rain while SO2 accounts for the rest [Badr & Probert, 1993b]. The main source for NOx in the environment nowadays is the combustion of fossil fuels in stationary (power plants, industry, and to a lesser extent households) and mobile (traffic) sources. Natural sources are the formation from living microbiological species and lightning. NO formed in combustion processes is generally discharged to the atmosphere and subsequently converted to NO2. The air oxidation of NO to NO2 at room temperature in is a relatively slow process. N2O emissions by human activity arise from chemical industry (e.g. nitric acid production). Also, fluidised bed combustion of higher rank coals cause significant N2O emissions. When air blown gasification of biomass and/or coal is applied as thermal conversion technology, a product gas is generated with heating values in the range of 3-6 MJ/kg, which contains relatively high concentrations of fuel-bound NOx precursor species [Hoppesteyn, 1999]. These are NH3, HCN and nitrogen-containing hydrocarbons, as has been measured by several investigators, e.g. [Chen, 1998], [Hämäläinen, 1996], [Kurkela et al., 1996], [Kurkela et al., 1995], [Kurkela et al., 1993b], [Kurkela and Ståhlberg, 1992], [Zhou, 1998]. The NOx emission from gas turbines integrated in an IGCC process is mainly formed from these fuel-nitrogen species. The formation of thermal and prompt NOx during combustion of LCV gas is negligible compared to above-mentioned fuel nitrogen derived NOx [Hoppesteyn, 1999]. N2O emissions have not been reported from gas turbine combustion processes and are not expected due to the relatively high combustion temperature range. There are several potential ways to reduce NOx emissions:

• use of advanced low-NOx burners in the combustion stage; • conventional gas cleaning (wet scrubbing); • cleanup of the flue gas of gas turbine combustors with e.g. selective catalytic NOx reduction (SCR);

• cleaning of the gas produced by the gasifier by e.g. selective catalytic N-species reduction (SCR), or selective catalytic oxidation (SCO);

• use of fuel with low nitrogen content; • increase the conversion of NH3 and HCN into N2 in the primary gasification process.

The use of low fuel-NOx burners for gas turbines is an attractive option, which is also still subject of study, see e.g. [Hoppesteyn, 1999], [Nakata, 1996].

Conventional gas cleaning, or wet scrubbing, to remove NH3 from the product gas, is an option but this has disadvantages like lower overall power generation efficiencies, as substantial product gas cooling is necessary; see [Woudstra & Woudstra, 1995] for an exergy analysis of IGCC systems with wet and dry hot gas cleaning. Also a large spectrum of organic (tars) and inorganic chemical species (like e.g. NH3, HCN, H2S etc.) has to be removed from the scrubber water before it can be recycled or released to a water treatment facility. The use of an “end-of-pipe” solution such as catalytic NOx removal downstream of the gas turbine, e.g. SCR (selective catalytic removal by NH3 injection), is limited due to catalyst poisoning by the product gas, particulates or temperature limitations. This solution is also less attractive because of the large gas flows, which have to be handled due to the high amount of excess air added to the gas turbine, which requires relatively big equipment. When SCR is applied under non-optimised process conditions, some of the injected, unreacted NH3 is vented to the atmosphere, thus generating a new and not acceptable pollution problem. Application of SCR upstream of the combustor by means of injection of NO, necessary to reduce NH3, can lead to catalyst poisoning by sulphur species, heavy metals and solid deposits. 10

This imposes additional requirements on the gas cleaning system and this process is not yet applied at commercial scale. The option of using low nitrogen containing fuels is not always available. When agricultural residues or waste material from biological origin are applied, there is an amount of nitrogen in the fuel on an energy basis comparable to fossil fuels, as illustrated by the figures in table 1.2. Wood seems to be a good alternative in this respect, though. An increase in the conversion of fuel bound nitrogen into harmless molecular nitrogen upstream of the combustion and gas cleaning section is an attractive and promising option. The fuel bound nitrogen species concentrations in the product gas of the gasifier are decreased by measures taken within the gasifier, which leads to less extensive and thus less expensive gas cleaning before the gas turbine combustion. For the high temperature, dry gas cleanup, the argument is quite clear. With respect to wet scrubbing using water, it is necessary to emphasize that the bound nitrogen species (especially NH3) have to be removed from the scrubber water. Therefore acid chemicals are needed to neutralise the water or the ammonia has to be removed by stripping processes. The use of neutralizing agents contributes negatively to the sustainability of the biomass gasification process, due to high consumption of chemicals that are produced using fossil fuels. The ammonia stripping contributes to lower overall process efficiencies because the concentrated ammonia needs to be dealt with. Therefore, selective primary conversion of bound nitrogen into N2 in the gasifier seems to be the better solution. However, there is still no clear picture how the fuel bound nitrogen in biomass is converted into NH3, HCN, nitrogen-containing hydrocarbons, nitrogen oxides and N2 during the gasification process. In particular for pressurised (fluidised bed) gasification no model including bound nitrogen conversion has been set up and tested against experimental data so far. There are also no extensive concentration profile measurements of nitrogen species before and downstream of high temperature gas cleaning units available in the open literature. This is partly due to measurement difficulties and partly caused by confidentiality policies of gasifier manufacturers. The aim of this thesis is to set up a reactor model that is capable of predicting the main product gas composition and fuel bound nitrogen speciation using experimental data from fuel characterisation experiments and validate it with measurements carried out at pilot scale (50 - 1500 kWthermal) pressurised air-blown fluidised bed reactors. 1.6

Outline of this thesis

This thesis deals with experimental and modelling work concerning biomass/coal pressurised airblown bubbling fluidised bed gasification. Emphasis is put on the formation and destruction of nitrogen compounds from solid fuel bound nitrogen, which to the knowledge of the author has not been considered in models for such gasifiers so far. The present study is limited to steady state performance. Start-up, shutdown or other dynamic phenomena are not studied. The influence of the fuel sources and additives as well as process conditions like pressure, temperature and reaction stoichiometry on the main gas constituents and gaseous nitrogen species are investigated. The experimental validation of the developed model is performed on two scales of thermal input. In chapter 2 an overview is given of biomass based commercial fluidised bed gasifiers and also of experimental fluidised bed gasification research and development conducted throughout the world. Furthermore, special attention is paid to literature concerning modelling of the fundamental stages of fluidised bed gasification and to experiments performed to determine the fate of nitrogen species in gasification processes. Potentially attractive primary measures for fuel-NOx emission reduction are highlighted. Finally, the experimental and theoretical objectives of this study are presented. Chapter 3 describes the pressurised fluidised bed test rigs of Delft University of Technology and Stuttgart University that have been used for the experimental validation experiments at two scales. Furthermore, a TG-FTIR set up is presented which has been used for the characterisation of the pyrolysis process of fuels used in the gasification experiments.

11

Finally the pressurised heated grid reactor at Eindhoven University of Technology is described, which has been used for flash pyrolysis experiments of fuels used in the experimental programme of this research. Chapter 4 presents the results of the gasification tests conducted in both pressurised fluidised bed test installations. Also, pyrolysis experiments performed with the TG-FTIR set-up are presented and discussed in relation with the gasification experiments. Finally experimental results from heated grid flash pyrolysis are presented and analysed. Chapter 5 deals with the modelling of the pressurised fluidised bed gasification process with emphasis on the fate of fuel bound nitrogen species. Here, simulation cases are presented with the aim to test the model for various process conditions and to study the expected effect of primary measures to reduce fuel bound nitrogen content in the gasification product gas. In chapter 6, modelling and experimental results for the pressurised fluidised bed tests carried out at two scales of thermal input are compared and analysed. Finally, in chapter 7 conclusions of the experimental work, the modelling study, the comparison of model and experiments are summarised. In addition, recommendations for further research are presented.

12

Chapter 2

Fluidised bed solid fuel gasification processes, overview and analysis of experimental research and modelling 2.1

The fluidised bed reactor applied for solid fuel gasification

The fluidised bed reactor has been one of the workhorses for large scale coal conversion, since it was first patented by Fritz Winkler in 1922 and commercialised in 1926, see e.g. [Howard, 1983] and [Kunii & Levenspiel, 1991]. The reactor scaling up from bench scale units via pilot plants to large units for combustion or gasification has been mostly carried out empirically. Emission studies from fluidised bed thermal conversion processes have often also been empirical. Modelling studies, however, concerning fuel conversion and main product gas components generation for gasification have been ongoing since a few decades. In this chapter an overview is given of biomass based commercial fluidised bed gasifiers. Hereafter recent experimental fluidised bed gasification research and development conducted worldwide both on small (<50 kWth) and larger scale is given. Special attention is then paid to literature concerning the fate of nitrogen species in fluidised bed gasification processes, because the literature regarding fuel bound nitrogen (FBN) conversion to NOx precursors in this type of gasifiers is scarce. Subsequently the pyrolysis sub-process taking place during fluidised bed gasification is dealt with. Emphasis is put on the FBN chemistry. Furthermore, potentially attractive primary measures for fuel_NOx emission reduction are discussed. Finally, research requirements with an outlook on this thesis work are presented. 2.2

Industrial fluidised bed gasification systems

Nowadays there are worldwide several biomass based fluidised bed gasifiers using wood, agricultural residues, energy crops and (at least for a part) waste, like municipal solid waste (MSW), or sludges as feedstock. These gasifiers are in the design phase, or already operating on a demonstration or (semi-) commercial scale (see e.g. [Venendaal & Van Haren, 1999]). These installations (see table 2.1) are from companies, briefly described as follows: Battelle (FERCO corp.) In the late 1970’s Battelle Columbus developed indirectly heated gasification technology to produce a medium heating value gas (15-17 MJ/m3n, HHV basis) from biomass [Paisley et al., 1997], [Paisley and Farris, 1995]. The process features two interconnected atmospheric pressure circulating fluidised bed (CFB) reactors for steam gasification in one reactor and residual char oxidation with air in the second one while solids are exchanged between the two reactors. Several biomass fuels were tested, including pine poplar and switch grass, which are potential energy plantation crop species suitable for moderate climates. The process has been successfully demonstrated. Ebara corporation This company has developed a chemical process, called TwinRec, to process e.g. automotive shredder residue on a large scale, using atmospheric fluidised bed gasification at relatively low temperatures (500-600 °C) as the first thermal conversion process step followed by an ash melting furnace, operating at 1350-1450 °C. The process aims to process waste together with the recovery of (precious) metals. By now, 14 process lines have started operation and the technology looks promising for waste processing on a commercial scale.

13

Energy Products of Idaho (EPI) EPI is a US company (Coeur d’Alene, Idaho), manufacturing airblown atmospheric biomass/waste fluidised bed gasifiers. Their AFB can process fuel with moisture contents up to 55% and high ash contents over 25%. A wide range of fuels such as wood waste, bark, wood chips, RDF (Refuse Derived Fuel), hogged fuel, agricultural waste, urban wood waste, coal, PET and polyvinylbutyryl are mentioned. Most of the activities are in (fluidised bed) combustion. In the eighties and nineties of last century also airblown fluidised bed gasifiers were delivered for application in clay calcining kilns, steam generating boilers and for an office heating system in California. The company still manufactures all types of bubbling fluidised bed based energy systems [energyproducts, 2004], [Murphy, 2001]. Foster Wheeler Energy International Inc. Foster Wheeler owns several gasification patents and technologies. The first commercial atmospheric circulating fluidised bed (ACFBG) application replaced fuel oil in a lime kiln at Wisaforest Oy (Pietarsaari, Finland). This installation was delivered in 1983. Since then, similar gasifiers have been installed at two pulp mills in Sweden and one mill in Portugal, in a range of 17-35 MWth [Raskin et al., 2001]. In Lahti, Foster Wheeler Energia Oy designed and constructed a CFB gasifier of which the low calorific product gas is burnt in a pulverised coal-fired boiler, co-fired with natural gas. The 350 MWth Kymijärvi power plant belongs to the Finnish power company Lahden Lämpövoima Oy. The fuels applied are saw dust, wet and dry wood residues and recycled fuel (like plastics, paper, cardboard and wood). The gasifier has a capacity in the range of 40-70 MWth [Nieminen & Kivelä, 1998]. The project has been co-financed by the EU in the framework of the Thermie program with the following partners: Foster Wheeler Energia Oy (Finland), Lahden Lämpövoima Oy (Finland), VTT (Finland), Elkraft Power Company Ltd. (Denmark) and Plibrico Ab (Sweden). A joint venture company (Bioflow Ltd.) was temporarily formed in 1991 to market the pressurised CFB technology developed by the Swedish Company Sydkraft AB and Foster Wheeler International Inc. (initially companies of the Ahlström Group) [Ståhl, 2001], [Ståhl and Neergaard, 1998], [Ståhl et al., 1997]. The demonstration programme, in which also EdF (France) and Energi E2 (Denmark) participated, was co-financed by the EU, the Swedish Electric Utilities’ R&D Company and the Swedish National Energy Administration. The 18 MWth demonstration plant for co-generation was operated from 1996 until 1999. It comprised a pressurised air-blown CFB gasifier operating at 950-1000 °C and approximately 1.8 MPa. The gasifier was integrated with a fuel gas cooler, a hot gas filtration unit (initially equipped with Schumacher ceramic candle filters, later with metal filters) and a combined cycle of gasturbine (a 4.2 MWe Alstom Power Typhoon) and a steam turbine (1.8 MWe). The fuels tested were bark, logging residues, waste wood, wood chips, sawdust, Salix (short rotation forestry), straw and RDF. Figure 2.1 shows a birds-eye picture of the plant.

Figure 2.1 Picture of the Värnamo Demonstration Plant in Sweden.

14

Institute of Gas Technology (former IGT nowadays called GTI) / Carbona Inc. ®

From the end of the seventies last century, IGT developed the RENUGAS process with a pressurised bubbling fluidised bed gasifier (BFB) which utilises air and steam as oxidizer, see [Lau & Carty, 1994] and [Lau, 1998]. A 15 MWth pilot plant using this process was constructed in Tampere (Finland) by Enviropower Inc., a subsidiary of Tampella Power [Salo, 1998b]. The installation was commissioned in 1993 as a demonstration plant and has been operated for more than 2000 hours on paper mill waste, straw and coal mixtures, alfalfa stems and a variety of wood fuels, in total more than 5000 tonnes. The pressure at which the gasifier operates is approximately 2 MPa. Carbona, a licensed Finnish company, has been involved in international biomass gasification projects, among which a recent 17 MWth BIG GE (biomass based integrated gasification with gas engine) installation in Skive (Denmark) [Patel, 2004]. At the Biomass Gasifier Facility at the HC&S sugar factory in Paia, Hawaii (Maui) a ® RENUGAS gasifier was built and operated in the second half of the 1990’s. The capacity of the installation was 50 dry tonnes/day and the operation pressure was around 2 MPa. Westinghouse Electric Corporation (WEC) used this demonstration unit to test their high temperature ceramic filter technology under pressurised gasification conditions, see [Lau, 1998]. The demonstration programme was ended by 1998 because of technical problems. Lurgi Energy und Umwelt Inc. The German company Lurgi is involved as designer/constructor of CFB biomass gasifiers. Since 1983 the CFB reactor in the Lurgi AG Research and Development Centre has been gasifying more than 6000 hours during test runs. This test rig has a capacity of 1.7 MWthermal [Greil & Vierrath, 2000]. The company designed and constructed CFB gasifiers for use in the cement industry in Austria (Pöls, 27 MWth using tree bark fuel, now shut down) and in Germany (Rüdersdorf, 100 MWth, operating on clean/waste wood, RDF, lignite waste and rubber waste). Currently, two European projects on biomass gasification for power generation are based on Lurgi technology. In Italy (Pisa) the EU co-finances a project “Energy Farm”. The project comprises a CFB air-blown gasifier integrated with a combined cycle of a 10.9 MWe single shaft Nuovo Pignone PGT10 B/1 gas turbine and a heat recovery steam generator (HRSG) of 5 MWe on rated power basis. The design fuel consists of a mixture of wood chips from short rotation coppice (SRC), forest and agricultural residues, including olive-stones and grape-seed flour. The commissioning phase of the project was envisaged to commence during autumn 2000 [DeLange & Barbucci, 1998], but was delayed. The second European project for power production from biomass in which Lurgi contributed to the construction of the gasifier is the Amer power station operated by Essent in the Netherlands. The main fuel there to be processed will be wood. The installation consists of an 80 MWth air-blown CFB gasifier originally equipped with a gas cooler, baghouse filter and scrubber. In this concept, ammonia recovered by a scrubber/stripper process combination is redirected to the gasifier and the product gas leaving the scrubber unit is co-fired in a pulverised coal combustor [Willeboer, 1998]. Start up and initial testing has been performed. After this phase, changes have been made in the gas-cleaning unit, such that the raw product gas is now directly fired into the coal combustor. Termiska Processor Sweden AB. (TPS) TPS is a privately owned research and development company based in Sweden. The focus of the company is mainly on air-blown CFB technology for biomass and RDF. Nowadays the company is involved in four main biomass and RDF gasification projects. In Italy (Greve-in-Chianti) TPS technology was licensed to Ansaldo Aerimpianti SpA for use in a waste gasification plant [Barducci, 1999], [Rensfelt, 1997]. The facility processes 200 tonnes/day of pelletised RDF, which is fed to two CFB units, with a total capacity of 30 MWth. The resulting producer gas is burnt in a boiler to generate steam for a 6.7 MWe condensing steam turbine. In Brazil two biopower projects were defined [Arrieta&Sanchez, 1999]. First there was the Brazilian wood-fired BIG GT demonstration project. Main goal of this project was to prove commercial viability of BIG GT technology using wood from eucalyptus plantations.

15

The net electric power output was projected to be 32 MWe. TPS is also involved in a project in cooperation with the Brazilian sugar industry, Copersucar. The fuel for the air-blown CFB gasification process in this case will be bagasse and cane trash. Both projects seem to have been merged at the moment of writing of this thesis, see [Maniatis et al., 2003]. In Europe TPS air-blown gasification technology was applied in the EU co-funded ARBRE project in the United Kingdom. Construction of the unit in Eggorough in North Yorkshire was partly performed by the Dutch company NEM/Schelde. The fuel foreseen was willow and poplar from dedicated SRC. The power plant consists of an IGCC unit, including an Alstom Typhoon gas turbine. The net electric power output of the plant was projected to be 8 MWe. Startup was planned at the beginning of the year 2000; see [Pitcher et al., 1998]. A delay was encountered and first gasification runs were performed in 2001, resulting in a test with low calorific product gas fired in the gas turbine early 2002 [Pitcher et al., 2002]. The installation is now waiting for a restart in a new European project [Maniatis et al., 2003]. Table 2.1 gives an overview of the main demonstration projects with thermal capacities higher than 10 MWth. The overview given indicates that biomass/waste gasification using fluidised bed reactors for the medium scale of thermal input has potential commercial interest. Indirect co-firing concepts have been shown to be successful in longer-term operation (Aomori, Lahti, Varkaus, Zeltweg) and are already applied in commercial operation. For the Lahti boiler by co-firing biomass the NOx emission decreased approximately by 10 mg/MJ, equalling a 5-10% decrease [Anttikoski, 2002]. The status of several projects, however, also shows that at the moment more advanced concepts with potentially higher overall electrical efficiencies, like fluidised bed based IGCC, and also biofuel production still suffer severe competition from fossil fuel based processes.

16

Table 2.1 Current (semi-) commercial large scale (>5 MWth) fluidised bed biomass / waste gasification projects, an overview classified according to thermal input scale. Project

Gasifier

Use of product gas

FICFB process Güssing (Austria)

Austrian Energy & TU Wien, CFB

Gas engine 4.5MWth, 2 MWel

BioCoComb, Zeltweg (Austria)

Austrian Energy & Environment CFB

Co-firing in pulverised coal combustor, 10 MWth biomass / 330 MWth coal

Skive Fjernvarme (Denmark)

Carbona Pressurised BFB

BIG gas engine, 11.5 MWth, 5.5 MWe

Enviropower, Tampere (Finland)

Carbona Pressurised BFB IGT RENUGAS® Pressurised BFB

Biomass Gasifier Facility (BGF), Hawaii (USA) Bioflow Värnamo (Sweden)

15 MWth Gasification testing for electricity and methanol production (50/50), 90 tonne/day (15 MWth,max)

Current status

References

In operation

[Hofbauer et al., 2003]

In operation from 1997-2001, shut down after successful operation Contract signed, construction fall 2004 Shut down after demonstration Shut down after demonstration in 1998 Shut down after successful demonstration, 1999 Start-up phase ended, now mothballed

[Anderl et al., 1998] [Fernando, 2002] [Patel, 2004] [Salo & Patel, 1997] [Lau, 1998], [Wiant et al., 1998], [Lau & Carty, 1994]

Foster Wheeler Pressurised CFB

IGCC, 18 MWth, 6 MWe

TPS Atmospheric CFB

IGCC, 8 MWe

Ebara corporation ABFB

Boiler for power generation 2 x 40 MW

In operation

[Selinger et al., 2003]

Battelle Columbus interconnected CFB’s

Initially in steam cycle power plant, later in IGCC 15 MWe

Demonstration

[Paisley et al., 1997]

Pressurised BFB, Lurgi licensed

Demonstration unit for IGCC, approx. 35 MWth

Shut down after demonstration in period 1989-1992

[Adlhoch et al., 1992a,b]

Corenso Varkaus (Finland)

Foster Wheeler/ Corenso Ltd. Atmospheric BFB

Use of gas in boiler 40 MWth

In operation

[Opet, 2001]

THERMIE Energy Farm, Pisa (Italy)

Lurgi Atmospheric CFB

IGCC, 16 MWe

Construction procedure pending

[DeLange & Barbucci, 1998]

Electrabel Ruien (Belgium)

Foster Wheeler Atmospheric CFB

Under construction

[Anttikoski, 2002]

In operation

[Raskin et al., 2001], [Nieminen & Kivelä, 1998]

ARBRE Aire Valley (United Kingdom) Aomori Automotive Shredder Plant (Japan) Battelle process (FERCO), Burlington (Vermont, USA) HTW Rheinbraun, Wesseling (Germany)

Co-firing in Pulverised coal/natural gas boiler, 50 MWth Co-firing in pulverised coal/natural gas boiler, 40-70 MWth

[Ståhl, 1998] [Ståhl, 2001] [Pitcher et al., 1998] [Pitcher et al., 2002]

Kymijärvi power plant Lahti (Finland) Amer power station (Essent), unit 9, Geertruidenberg (NL) Aerimpianti, Greve-in-Chianti (Italy)

Lurgi Atmospheric CFB

Co-firing in Pulverised coal combustor, 80 MWth

Start-up phase

[Willeboer, 1998] [Greil et al., 2000] [Fernando, 2002]

TPS Atmospheric CFB

Combustion of LCV gas, power generation in steam cycle

In operation

[Rensfelt, 1997] [Barbucci, 1999]

BIG-GIT Bahia (Brasil)

TPS Atmospheric CFB

IGCC, 32 MWe

Definition, progress unclear

[Arrieta & Sanchez, 1999] [Maniatis et al., 2003]

HTW Rheinbraun, Berrenrath (Germany)

Pressurised BFB, Lurgi licensed

Demonstration unit for methanol production, 140 MWth

Shut down after demonstration in period 1986 - 1997

[Adlhoch et al., 1992a]

2.3

Foster Wheeler Atmospheric CFB

An overview of recent research, development and small scale demonstration activities

In the past two decades quite a number of fluidised bed gasifiers for biomass/coal research and process development have been operated in thermal input range up to 3 MW. Table 2.2 presents an overview of the process development installations having a solid fuel input capacity of 50 kWthermal and higher, which have been relatively well documented in the literature.

17

Table 2.2 Overview of fluidised bed biomass (+ solid fossil fuel) gasification process development test rigs (>50 kWth) and small scale (semi-) commercial plants (<5 MWth); an overview classified according to thermal input scale. Institute University Company IVD University Stuttgart (Germany) DTU (Denmark) University of Saragossa (Spain) VTT Energy (Finland) KTH University Stockholm (Sweden)

Gasifier Type PBFB

Therm. Cap.

D Bed

(kWth) 50

(m)

L Freeboard (m)

Feed Location from bottom (m)

0.18

1.0

3.0

0.1

L Bed

(m)

D Freeboard (m)

0.10

Gas Cleanup Technology Cyclone + candle filter Cyclone

Literature [Nagel et al., 1998] [Nagel, 2002] [Stoholm et al., 2000] [Caballero et al., 2000], [Gil et al., 1999a,b and 1997] [Leppälahti & Kurkela, 1991] [Brage et al., 2000], [Sjöström et al., 1999], [Chen, 1998]

ACFB

50

-

-

-

3.0*

-

ABFB

65

0.15

-

-

3.2*

bottom

ACFB

75

0.16

0.16

<0.8

3.65*

bottom

2 Cyclones

PBFB

80

0.14

0.20

0.6

1.0

0.6

Metal filter

LUND university(Sweden)

PBFB

90

0.10

0.10

-

3.3*

0.30

CUTEC (Germany) IPE, Brno (Czech Rep.)

ABFB ABFB

100 100

-

-

0.30 -

>2.1*

0.30 -

Ceramic candle filter Cyclone Cyclone

RWTH Aachen (Germany)

ACFB

100

0.25

-

-

-

-

2 cyclones

TU Delft (NL) TU Magdeburg (Germany) TU Wien (Austria) Universityof Minnesota (USA) Paul Scherrer Institute Villigen (Switzerland)

ACFB ABFB ACFB

100 100 100

0.08 0.4 0.3

0.08 0.6 -

-

0.2 -

Cyclone Cyclone

ABFB

100

0.16

0.221

0.69

5.0* 3.0 4.25* 0.43 0.58

0.08

Cyclone

[Jiang,1991]

-

Cyclone

[De Sousa, 2001], [De Sousa & Stucki, 1997] [Buffinga, 2002] [Lewis et al., 2002]

ABFB

120

0.21

0.21

BTG/Duys Bladel (NL)

ABFB

135

-

-

Oklahoma State University (USA)

ABFB

170

0.25

-

0.95

4.0

-

-

-

-

-

-

2 Cyclones

1.0*

0.13

3 Cyclones

[Van den Aarsen, 1985]

3.0

0.3

Cyclone Cyclone + fibre filter

[Van den Enden & Silva, 2004]

0.21.0 2.0

TU Twente (NL)

ABFB

200

0.3

0.3

ABFB

245

0.57

0.75

ACFB

250

0.1

0.1

-

6.5

0.95

0.84

0.69 - 2.0

1.585 ††

0.26

Ciemat Madrid (Spain) Enerkem Technologies Sherbrooke (Canada) IPT Sao Paolo (Brazil)

ABFB ACFB

280

0.42

300

0.20

0.20

ABFB

400

0.31

0.46

ABFB

430

0.5

-

0.60 † -

6.5* 2.9*

5.0*

Cyclone

0.37

Cyclone

0.29 (+0.46) 0.3

2 Cyclones, scrubbers,filt. Cyclone

ECN (NL)

ACFB

500

0.20

0.20

-

6.0*

1.0

2 cyclones

IER Laboratory Serpong (Indonesia)

ABFB

500

0.40

0.40

0.60

3.7*

bottom

2 cyclones

Umsicht Oberhausen (D)

ACFB

500

0.31

-

-

8.0*

-

VTT Energy (Finland)

PBFB

500

0.15

0.25

-

4.2*

0.15

Iowa State University Ames (USA)

ABFB

800

0.46

0.46

0.60

2.44*

bottom

Putian Huaguang Miye Ltd. (China)

ACFB

1000

1.25

8

Sanya Timber Factory, Hainan Island (China)

ACFB

1200

1.8

2 Cyclones+filt. 2 Cyclones+ 5 ceramic filters 2 cyclones

Cyclone,Vent uri, 2 scrubbers 2 cyclones, 4 scrubbers

-

-

-

1.8

-

8.0*

1.7

4.5 ††

-

Ceramic filter

-

-

-

TU Delft (NL)

PBFB

1500

0.38

0.49

2.0 †

Lurgi Umwelt GmbH Frankfurt (Germany)

ACFB

1700

-

-

-

Cratech, Tahoka (USA)

PBFB

ca. 2000

0.61

0.61

0.70

-

-

4 metal filters

CTDD Cheltenham (UK)

PBFB

2000

0.3

0.45

4.0

-

0

IGT, Chicago (USA)

PBFB

2000

0.28

-

-

6.4*

bottom

Zhanjiang wood factory (Cn)

ACFB

2000

0.41

0.41

-

4.0*

Bottom

Foster Wheeler Karhula (Finland)

ACFB

3000

0.60

0.60

-

10*

3.0

3 Cyclones Cyclone+ filter Cyclone+hot baghouse filter+scrubbe r

Free University of Brussels (Belgium)

ABFB

3000

0.80

1.20

0.6

2.0

-

* total reactor height ** static bedheight

18

† maximum

[Arvelakis et al., 2002] [Zdenek et al., 2002] [Gudenau et al., 1993] [Gudenau & Hahn, 1993] [Chen et al., 2003] [Hamel, 2001] [Hofbauer & Rauch, 2001]

Cyclone, RPS, tar cracker

UNIFEI-NEST (Bra) University of British Columbia (Canada) UNICAMP Campinas (Bra)

[Padban, 2000], [Padban & Odenbrand, 1999]

†† minimum

Cyclone

[Li et al., 2004] [Gómez et al., 1999] [García-Ibañez et al., 2001] [Garcia et al., 2003] [Bilodeau,1993], [Reed & Gaur, 1999], [Abatzoglou, 2003] [Van den Enden, 2000] [Kersten, 2002], [v.d. Drift et al., 2001]

[Hartiniati et al., 1989] [Ising et al.,1998] [Kurkela et al., 1993b] [Smeenk et al. 1998,1999]

[Yin et al., 2002]

[Zhang et al., 2002]

[this thesis]

[Greil, 1998]

[Craig, 1996],[Purvis&Craig, 1998] [Paterson, 1997]

[Knight, 2000]

[Bingyan et al., 1994]

[Asikainen et al., 2002] [Lee et al., 2003]

[El Ashri, 2000] [Maniatis et al., 1989]

The overview in table 2.2 shows that in the end of the last and the beginning of this century several research and small-scale development programmes were directed towards fluidised bed gasification. Attention is given to both circulating and stationary bed systems. The majority of these were used in the framework of biomass to electricity routes. Most of the systems described aim at atmospheric gasification combined with low temperature gas cleaning. Systems with pressurised gasification and hot gas cleaning form a minority despite their potential high efficiencies. For those systems operating with a hot gas cleaning unit not much research attention was paid to the fate of nitrogen species, with a few exceptions in the work of VTT, KTH, Lund University and in this thesis. Table 2.3 gives an overview of the bench scale rigs for biomass (and fossil) solid fuels, most of them used in the past decade for fundamental and applied research. The installations are generally situated at universities or research foundations. Emphasis of research topics is for the majority of the facilities on fuel conversion behaviour into main gas components, carbon conversion and tar formation. The behaviour of minor nitrogen species has not received much attention with the exception of some studies by KTH, VTT, TPS and the university of Hawaii. Table 2.3 Overview of bench scale fluidised bed biomass (+ solid fossil fuel) gasification installations (< 50 kWth) ); an overview classified according to thermal input scale. Institute University Company

Gasifier Type

Therm. Cap. (kWth)

D bed (m)

D Freeboar d (m)

L Bed (m)

L Freeboard (m)

Feed Location from bottom (m)

Gas cleanup

Literature

VTT Energy Espoo (Finland) University of Saragossa (Spain) Seikei university Tokyo (Japan)

PBFB

Batch operation

0.031

-

-

-

-

filter

ABFB

0.4

-

-

-

-

-

cyclone

ABFB

0.6

0.05

0.099

0.24†

0.74*

0.35

filter

[Kojima et al., 1993]

IC, London (UK)

PBFB

1.5

0.028

0.028

-

0.5*

0

Tar trap + Filters

ABFB

2***

0.07

0.145

-

0.5*

top

Cyclone

[Collot et al., 1999] [Paterson et al., 2002] [Franco et al. 1998] [Pinto et al., 2002]

ABFB

2***

0.05

0.104

0.15 0.30

0.45

2 positions (top/bottom)

2 hot gas filters

[Vriesman et al., 2000], [Vriesman, 1999]

ABFB

3

0.05

0.10

-

-

-

-

[Moilanen & Kurkela, 1998]

University of Madrid (Spain)

ABFB

4***

0.06

-

-

-

-

Filter + catalytic tar cleaning

[Corella et al.,1999]

ECN, Petten (NL)

ABFB

5

0.074

0.108

0.5

0.6

0.03

ABFB

6.5

0.043

0.114

1.025

0.5

0.082

2 Cyclones+fi lter

[Pan et al., 2000 and 1999]

ABFB

12

0.15

0.15

-

1.36*

top

Cyclone + filter

[Guo et al.,2001]

ABFB

13

0.089

0.152

0.70**

2.50*

0.25

Metal filter

[Turn et al., 1998a], [Zhou,2000] and [Zhou et al.,1998]

ABFB

25

0.11

0.135

0.8-0.9

2.8*

0.10

Cyclone + filter

[Miccio et al., 1999 a and b], [Mörsch et al., 2000]

ABFB

26***

0.2

0.2

-

3.2*

above bed

Cyclone+ scrubber

[André et al., 2000]

ABFB

30***

0.255

0.395

0.1950.315

2.7*

bottom

ABFB

30

0.20

0.27

0.59

1.2-1.4

ABFB

-

-

-

-

ABFB

-

0.155

-

ACFB

-

0.12

0.15

INETI Lisbon (Portugal) KTH University Stockholm (Sweden) VTT Energy (Finland)

Universitat Politècnica de Catalunya (Spain) Tsinghua university Beijing (China) University Of Hawaii (USA) IVD University Stuttgart (Germany) INETI Lisbon (Portugal) Dalhousie University Halifax (Canada) TPS Studsvik (Sweden) Agricultural University of Athens Forschungszentrum Karlsruhe (Germany) Chinese Academy of Sciences Guangzhou(China)

* total reactor height ** static bedheight

*** estimated

Cyclone

[Moilanen & Kurkela, 1998] [Garcia et al., 2001], [Esperanza et al., 1999]

[Van der Drift&Olsen, 1999] [Van Paasen et al., 2002]

Cyclone

[Mansaray et al., 2000a,b] [Mansaray et al., 1999]

0.10

Cyclone + filter

[Berg et al., 2001], [Berg et al, 1999 a and b]

-

-

Cyclone

[Abeliotis et al., 2000]

0.4

0.7*

-

-

[Henrich et al.,1999] [Rumpel, 2000]

-

2.0*

-

-

[Liu et al., 2000]

† maximum

†† minimum

19

2.4 Experimental findings regarding the fate of fuel nitrogen during fluidised bed gasification Only for few of the fluidised bed gasification projects mentioned in tables 2.1-2.3, data of nitrogen compound concentrations or N-conversions have been published in the open literature. In this section, general trends observed for the influence of several process- and fuel related parameters on the fuel bound nitrogen conversion during fluidised bed gasification are described. 2.4.1 Influence of fuel type Especially relevant for this study, which is concentrated on pressurised fluidised bed gasification of biomass and fossil fuels, are the experimental findings of [Goldschmidt et al., 2001]. These authors showed data from the 18 MWth Värnamo pressurised CFB gasification IGCC demonstration plant. The fuels used in this demonstration programme were cellulose chips/sawdust, bark and forest residue, bark only, willow, 15% RDF in bark, 25-30% RDF in sawdust and wheat straw. They cover a broad range of nitrogen contents, from 0.1% in sawdust to approximately 1% in wheat straw. The gasifier is bottom-fed and air-blown. The NH3 content of the LCV product gas from the gasifier was analysed downstream of the hot gas filter (operating at a temperature of around 350 °C) using a continuous on-line infrared spectroscopic instrument (OPSIS). Also, NH3 was analysed by researchers of Lund University, using a Fourier transform infrared spectrometer (FTIR). Additionally, HCN and NH3 were sampled and analysed by VTT, using (wet chemical) methods described by [Ståhlberg et al., 1998]. The results of the three analysis methods for NH3 agreed well with each other. The conversion of fuel bound nitrogen to NH3 in the gasifier was reported to be 60-65% for all fuels, independent of fuel type and nitrogen content. The concentration levels of HCN were about 1-2% of the NH3 concentrations. At VTT (Espoo, Finland), impressive experimental experience was gathered during the past decades in the field of (pressurised) fluidised bed gasification with small to process demonstration rigs. The gasifiers are all reactors equipped with bottom feeding. NH3 and HCN were reported to be the main fixed nitrogen species present in the product gas, see e.g. [Kurkela & Ståhlberg, 1992], [Kurkela et al., 1993b], [Leppälahti, 1993] and [Kurkela et al., 1996]. As fuels, different types of coal, lignite, peat, woody material and (agricultural) waste residues were applied. The research carried out using the 500 kWth PFB at 0.4-1.0 MPa and freeboard temperatures of circa 1070-1270 K is the closest related to the work described in this thesis. In coal gasification, the conversion of fuel bound nitrogen into NH3 varied from 19% with bituminous coal to 93% with high volatile coal. In peat and wood gasification, the conversion of fuel bound nitrogen into NH3 was 5595% and 72-97%, respectively. Fuel-N conversion to HCN was highest (ca. 5%) for wood. The conversion of fuel nitrogen into tar bound nitrogen was the highest for peat (0.5%) and pyridine was the main tar-N species. For sawdust gasification relatively low concentrations of pyridine were found at low freeboard temperatures, but for brown coal gasification all the nitrogen-containing tar components were negligible in the whole temperature range studied. Table 2.4 shows the straw gasification results at 0.5 MPa, reported by [Kurkela et al., 1996]. The gas concentrations were measured downstream of the ceramic filter unit. The particle size of straw varied between 0 and 5 mm. Nitrogen contents varied between 0.5-0.6 wt% on a dry fuel basis. Fuel nitrogen conversions to NH3 were in the range of 51.8-70.7% and 4.5-12% for HCN. Fuel nitrogen remaining in tar was relatively small (0.8-1.8%). Unreleased char nitrogen was in the range of 3.8-16.7%. As additives Al2O3 and dolomite were applied. Bed temperatures were kept below 800 °C, as the fuel is known to cause bed agglomeration.

20

Table 2.4 Overview of straw gasification in VTT’s 500 kWth PFB [Kurkela et al., 1996]. Experiment number

E1

E2

E3

E4

E5A

E5B

E6

E10

additive P (MPa) λ T bed (°C) T fb (°C) Steam/air mass ratio (-) Additive/fuel mass ratio (-) Wet gas composition H2O (vol.%) (calc.) CO (vol.%) CO2 (vol.%) H2 (vol.%) N2 (+Ar) CH4 (vol.%) C2 hydrocarbons (vol.%) H2S (ppmv) COS (ppmv) NH3 (ppmv) HCN (ppmv) C-conversion to gas & tar (%) Fuel-N to NH3 (%) Fuel-N to HCN (%) Fuel-N to tar-N (%) Fuel-N to char-N (%)

Al2O3 0.5 0.26 720 815 0.18 0.012

Al2O3 0.5 0.28 670 830 0.20 0.012

Al2O3 0.5 0.32 690 860 0.20 0.013

Al2O3 0.5 0.34 750 885 0.24 0.017

dolomite 0.5 0.31 750 885 0.12 0.031

dolomite 0.5 0.28 790 800 0.24 0.027

Al2O3 0.5 0.29 765 805 0.23 0.025

Al2O3 0.5 0.31 770 850 0.17 0.017

22.8 9.03 12.27 6.56 44.45 3.78 0.86 147 8 2146 154 84.4 67.3 4.8 1.2 16.7

25.8 8.90 10.76 3.78 46.36 3.34 0.84 119 15 1788 260 85.5 60.2 8.8 1.5 13.2

24.5 8.68 10.87 4.30 47.58 3.25 0.60 113 8 1721 370 93.9 65.6 14.2 1.6 9.1

25.2 7.70 10.92 4.26 48.40 2.84 0.49 112 7 1399 307 97.2 58.5 12.9 0.8 3.8

19.5 8.21 13.60 6.36 47.35 3.94 0.78 137 8 2165 137 92.3 70.7 4.5 0.8 8.0

25.9 7.19 12.30 5.11 44.63 3.56 1.10 141 4 1741 222 91.5 60.4 7.7 1.8 6.4

25.8 7.57 11.65 4.60 45.86 3.34 0.96 134 7 1766 230 89.8 58.5 7.5 1.7 7.7

23.8 8.99 11.51 3.96 47.18 3.58 0.78 145 8 1501 343 93.8 51.8 12.0 1.8 5.0

Table 2.5 shows the results of PFB gasification tests using sawdust as fuel, reported by [Kurkela et al., 1993b]. The gas composition was determined downstream of the ceramic filter. The fuel particle sizes were in the range of 0-5 mm. The fuel bound nitrogen content was varying between 0.08 and 0.11 wt% (db) with ash contents being very low: between 0.08 and 0.34 wt% (db). Lower NH3 and HCN concentrations in the product gas are observed as compared to the straw gasification experiments, reflecting the lower fuel bound nitrogen contents. No values of fuel nitrogen conversion to NH3 and HCN were reported for the experiments given, although an indication was reported that these values were in the range of 72-97% for NH3 and 1-5% for HCN. Table 2.5 Overview of sawdust gasification in VTT’s 500 kWth PFB [Kurkela et al., 1993b]. Experiment number

SD1

SD2

SD 3

SD 4

SD 5

SD6

SD7

SD8

SD9

SD10

Additive

Sand

Sand

Sand

Al2O3

P (MPa) λ T bed (°C) T fb (°C) Steam/air mass ratio (-) Additive/fuel mass ratio (-) Wet gas composition H2O (vol.%) (calc.) CO (vol.%) CO2 (vol.%) H2 (vol.%) N2 (+Ar) (vol.%) CH4 (vol.%) C2 comp. (vol.%) NH3 (ppmv) HCN (ppmv) C-conversion to gas and tar (%)

0.4 0.28 745 900 0.050

0.4 0.39 840 960 0.12

0.4 0.43 750 1020 0.091

0.4 0.32 735 945 0.088

Dolo mite 0.4 0.30 770 940 0.088

Dolo mite 0.4 0.34 770 1010 0.082

Dolo mite 0.4 0.31 780 940 0.089

Dolo mite 0.5 0.31 835 935 0.072

Dolo mite 0.5 0.39 830 970 0.085

Dolo mite 0.5 0.39 860 1000 0.073

0

0

0

0

0.030

0.032

0.034

0.033

0.065

0.043

15.5

19.1

17.0

13.8

13.7

14.4

13.8

13.2

17.2

12.8

13.5 13.2 9.0 42.8 5.2 0.72 6.8 x102 n.i.

11.3 12.2 6.8 46.9 3.3 0.23 8.1 x102 n.i.

11.0 11.2 8.3 49.9 2.4 0.20 2.5 x102 n.i.

13.6 11.6 8.4 47.1 4.9 0.51 3.4 x102 n.i.

13.5 12.6 9.1 45.5 4.9 0.53 3.5 x102 n.i.

13.4 12.2 9.3 46.8 3.6 0.21 4.3 x102 n.i.

12.2 13.5 9.4 45.7 4.9 0.38 4.3 x102 n.i.

11.4 14.0 9.3 46.8 4.8 0.57 5.2 x102 n.i.

8.9 13.8 7.8 48.3 3.8 0.15 2.5 x102 n.i.

10.4 13.9 9.7 49.3 3.7 0.17 4.4 x102 n.i.

95.0

97.9

98.1

96.7

94.6

94.5

95.1

98.2

98.6

98.6

n.i.: no indication of values per experiment, only range of 5-30 ppmv in dry gas reported.

[Leppälahti, 1998] concluded that the chemical functionality of the fuel bound nitrogen is important to explain the yields of NH3 and HCN from different fuels and at different process conditions. 21

He suggested that relatively young fuels such as peat and wood, contain mostly amino type nitrogen. These release nitrogen in primary reactions mainly as NH3 and to a lesser extent HCN through secondary reactions from heterocyclic ring structures, when compared to coal. HCN is then converted to NH3 in the presence of H2 by reactions inside the pores of char particles. At Lund University, a 90 kWth PFB facility was used for gasification of a range of biomass and fossil fuels [Padban, 2000]. Different sawdust types, bark, carton, textile waste, plastic waste and pulverised Polish coal were used as fuels. The NH3 concentrations observed in the product gas ranged from 100 to 5500 ppmv, depending on fuel, temperature and air stoichiometry applied. HCN concentrations were close to zero but with a large uncertainty level. The fuel-nitrogen content of the fuel influences the NH3 gas phase concentration to a major extent. A weak positive relationship between conversion of fuel bound nitrogen to NH3 and N content of the fuel was found, although the scatter in the data was quite large, due to variation of the air stoichiometry and temperature. [Padban, 2000] reported fuel-N conversions to NH3 in a range of 10-80 %, with most values in the range of 40-60%. Enviropower (Finland), nowadays Carbona Inc. [Patel, 2004], used their single stage 15 MWth pressurised fluidised air-blown bubbling bed gasifier for long term gasification tests with wood and paper mill residues, and also straw together with coal. The main test variables were fuel type, moisture content, freeboard temperature, pressure, filter temperature, fluidisation velocity, gas residence time, bed material feed rate, bed material type and bed particle size, temperature of filter pulsing gas, gasifier controls and gasifier flow arrangements. The carbon conversion was in the range of 97-99%. The concentrations of NH3 and HCN were in the range of 1450-2810 ppmv (dry) and 10-30 ppmv (dry), respectively. The conversion of fuel bound nitrogen to NH3 was roughly 55% for paper mill waste and 40% for wood. The NH3 concentration was mainly dependent on the fuel nitrogen content [Salo & Patel, 1997]. Foster Wheeler company (Karhula, Finland) published measurement data for their 3 MWth ACFB PDU with straw gasification. The N content of the fuel, both in pelletised and loose form, was in the range of 0.6 and 1.1 mass%. Only fuel compositions and resulting gas qualities were reported for 5 test runs [Asikainen et al., 2002]. The NH3 content in the product gas was 2870-3230 ppmv in dry gas. Unfortunately, no data were published regarding other bound nitrogen species or of N speciation. The company TPS used a 30 kWth ABFB reactor for biomass gasification experiments. In part of the experimental research programme emphasis was put on the fate of fuel-bound nitrogen, see [Berg et al., 2001] and [Berg et al., 1999a and b]. These researchers concluded that in their experiments with miscanthus and sawdust pellets NH3 was the main carrier of bound nitrogen. [Berg et al., 1999a] carried out fluidised bed gasification experiments to investigate the influence of nitrogen functionality in the fuel on the NH3 conversion. Therefore, pelletised wood (10 mm in diameter) was enriched in nitrogen using two different nitrogen carriers, melamin (50% heterocyclic-N) and urea (amino-N). These artificial fuels were gasified in a dolomite bed. They found very similar nitrogen conversions to NH3 (about 60%), irrespective of the nitrogen functionality. However, 30% of the “naturally bound” nitrogen of the wood fuel was converted to NH3 in reference tests, using no N-additive to the fuel. These results indicate that the fraction of the fuel bound nitrogen in the additives that is converted to N2 did not depend on the type of nitrogen binding. At the university of Hawaii at Manoa gasification and pyrolysis experiments have been carried out using a small-scale air blown atmospheric bubbling fluidised bed [Zhou, 1998]. Experiments using feedstock with varying N-content (0.08-2.51% db) suggest that NH3 content in the product gas increases linearly with fuel bound nitrogen content for the same fuel type. This was also reported for tests with bagasse and different types of banagrass by [Turn et al., 1998b] at the same university. The conversion efficiency of fuel-N into NH3, which is the main nitrogen-containing compound, decreased with increasing fuel_N content from 80 – 40 % for N contents in the range of 0.12-0.44 mass% (db). NO was also detected in the product gas with values of 10-100 ppmv, the highest concentrations observed for sawdust containing low nitrogen amounts [Zhou, 1998]. This NO concentration was decreasing with increasing N content in the fuel. The fuel-N to NO conversion was found to be related to the NO3-:NH4+ molar ratio in the fuel. The higher this value, the higher the NO production.

22

For wood with low N content, the NO concentration in the gas was even higher than the HCN concentration in the range of air stoichiometries studied. This was not the case for the fuels with higher N content. A 500 kWth ACFB gasifier has been used by ECN for a broad range of biomass fuels and sewage sludge. The conversion of fuel nitrogen to NH3 during air blown experiments ranged from 27-83%, with 62% as average value. No values of fuel-N to HCN conversion were reported [van der Drift et al., 2001. Peat, lignite and sub-bituminous coal have been gasified at North Carolina State University using steam/oxygen in a pilot-scale fluidised bed reactor at 0.8 MPa [Zand et al., 1985]. Approximately 60% of fuel bound nitrogen was converted into NH3 for both coal types and less than 40% was converted in the case of peat. Roughly 5% of fuel-N was converted into tar-N, 10-15% into cyanate species (from scrubber water analysis, which could be oxidised HCN), 1% in thiocyanate and 0.5% cyanide and 917% in solid-bound N. The lowest conversion efficiency gen was obtained in the sub-bituminous coal char and the lowest in the peat char. The gasification conditions had no measurable effect on distribution the fuel-N between nitrogeneous products. In the KRW (Kellog-Rust-Westinghouse) PDU lignite, bituminous and sub-bituminous coals were gasified using steam/oxygen or air/steam at 1.5 MPa (see [Mann et al., 1985]; [Haldipur et al., 1989]). Lignite and sub-bituminous coals were gasified at lower temperatures (1093 K) than bituminous coals (1255-1373 K). With lignite and sub-bituminous coal, about 60% of fuel-N was converted into NH3. Small amounts of HCN were also observed. With bituminous coal the NH3 conversions were significantly lower (0.6-15%). Summarizing the influence of fuel type it can be noticed that increased nitrogen contents of the solid fuel result in higher concentrations of fuel bound nitrogen gas species, mainly consisting of NH3. The fuel bound nitrogen conversion trends are much more unclear, due to masking effects of the process parameters air stoichiometry and the -often resulting- bed and freeboard temperature. Measurements by researchers of Hawaii University, in which temperature and air stoichiometry were varied independently, though, indicate that at increased N content of the same fuel the conversion to NH3 decreases. Conversion into tar-N species is very low compared to nitrogen speciated into gas species and char bound nitrogen. 2.4.2 Influence of air stoichiometry Gasification results from the Lund 90 kWth PBFB gasifier using sawdust, bark as a fuel showed a tendency toward higher fuel nitrogen conversion to NH3 for increasing air stoichiometry, λ, and higher temperatures, see figure 2.2 [Padban, 2000]. For the other fuels indicated this tendency is not clear or even opposite. The air stoichiometry was varied between 0.16 and 0.46. The temperature in these single fuel experiments varied from about 800-905 °C for the bed section and from about 770-925 °C for the freeboard. All experiments were carried out at a pressure of 12 bar. The total fuel conversion increased at higher air stoichiometry levels and higher temperatures. At low air stoichiometry and temperatures only the easy-to-volatilize N is released from the fuel as NH3 and the rest of the fuel-bound nitrogen remains inside the structure of the char residue. A part of the fuel-N can also be converted into tarry compounds that subsequently release NH3. The maximum conversion of fuel bound nitrogen to tars was about 8%.

23

λ (-) Figure 2.2 Conversion of fuel bound nitrogen to NH3 for Lund PFB gasifier [Padban,2000]; SA, SB: sawdust;AT: agglomerated textile; P: plastic waste and TD: textile dust. [Berg et al., 2001] report on atmospheric miscanthus gasification using fluidised beds at two scales. In these experiments λ was varied between 0.16 and 0.29. Bed temperatures ranged from about 740 to 810 °C. The sawdust gasification experiments resulted in a somewhat lower fuel-N conversion to NH3, compared to miscanthus, see figure 2.3. Uncertainties in the fuel-N amount determination could explain the differences observed. The conversion of fuel-N to NH3 increased with increasing air stoichiometry.

λ (-) Figure 2.3 Relative conversion of fuel nitrogen to NH3 under atmospheric conditions as a function of the air stoichiometry factor for miscanthus in the KTH 0.05m internal diameter rig (y) and for Miscanthus ({) and sawdust ( ) in the 30 kWth TPS test rig. An increase of the NH3 concentration with air stoichiometry was also reported for atmospheric downdraft fixed bed gasification of wood and almond shells at circa 850 °C, while HCN concentrations decreased with air stoichiometry [de Bari et al., 2000]. Fuel bound nitrogen conversions were not reported, unfortunately, and could not be calculated from the data presented. Variation of the air stoichiometry in atmospheric BFB tests while keeping the temperature constant by external heating showed that the influence of this parameter was much less important than temperature [Zhou, 1998]. 2.4.3 Influence of temperature Experiments with peat as fuel (0-4 mm diameter) in a small atmospheric FB test facility showed that about five times as much NH3 as HCN was formed in the temperature range 810-940 °C: [NH3] varied between 3170 and 4600 ppm and [HCN] was in the range of 520-930 ppm. An increasing fuel nitrogen conversion to both NH3 and HCN was observed with increasing freeboard temperature.

24

The increased fuel-N conversion to HCN with increasing temperature was attributed partly to tar-N cracking [Leppälahti & Kurkela, 1991]. The fuel nitrogen conversion to NH3 for these atmospheric tests was ca. 30% and to HCN ca. 5-6%. By adding secondary air to the freeboard of their 500 kWth pressurised gasifier and so increasing freeboard temperatures, VTT researchers observed a decrease of the fuel nitrogen conversion to NH3, see [Kurkela&Ståhlberg, 1992]. [Leppälahti&Koljonen, 1995] showed that the measured reduction was lower than could be explained with homogeneous decomposition kinetics of NH3. They indicated that particles and other surfaces in the freeboard of the gasifier may act as catalysts in this decomposition, or that the reduction can be related to reduced NH3 formation at high-temperature pyrolysis, an aspect discussed later. [Rosén et al., 1997] presented results for their top-fed PFB at KTH and concluded that the conversion of fuel nitrogen into NH3 increased with increasing temperature. Values in the range of 5-57% were measured. HCN was only a minor nitrogen compound, with yields in the range of 0.07-2.5%. Also, NO was a minor nitrogen species with fuel N to NO values ranging from 0.95-11%. The conversion to NO decreased with increasing temperature. The nitrogen converted to char bound nitrogen was comparatively low: 0.54-6.3%; values of tar bound nitrogen were not given. [Berg et al., 199b] reported for their ABFB gasifier that secondary air addition to the freeboard led to only a minor decrease of NH3 conversion. This result, however, could not be studied separately from a change in temperature profile over the entire reactor. For a 6 MWth atmospheric CFB gasifier at Sanya Timber Factory (STF), [Zhang et al., 2002] report fuel-N conversion to NH3 ranging from 6 – 70%. The highest values were obtained at the highest operating temperatures, which was 820 °C. Also a clear increase with reactor height was observed. This nitrogen release behaviour was attributed to a devolatilisation process assumed to be taking place throughout the whole reactor, where, compared to bubbling beds, relatively high velocities are applied. Also, in this case very small fuel particles were applied (>50% of particles < 0.28 mm). Another explanation of the observed increase in fuel-N conversion to NH3 with temperature given by the authors was the lower char content with higher temperatures, leading to a decreased char-NO reduction to N2. A different trend was observed in a bench scale ABFB gasifier by [Zhou, 1998]. Increasing the temperature, while keeping the air stoichiometry the same, lead to a drastic decrease in the fuel-N to NH3 conversion and an increase in the conversion to N2 for several biomass fuels in an atmospheric ABFB. For Leucaena, a high N containing fuel, fuel-N conversion to NH3 decreased from about 64 to 18% with temperature increasing in the range of 750°C to 950 °C. For HCN the values decreased from 0.10 to 0.07%. Tar-N was not observed, though measured, and char-N decreased from 7.7% to 1.2%. In the same range fuel-N to N2 increased from about 39 to 86%. In his experiments the N balance closure was between 105 and 116%. In general the observations are such that fuel nitrogen conversions to NH3 and HCN can be expected to decrease with temperature. The exceptions reported by [Zhang et al., 2002] and [Rosén et al., 1997] could be attributed to low fuel conversions, so that at increasing temperature the solid fuel conversion increases and thereby the conversion to NH3 as well. 2.4.4 Influence of pressure The Finnish VTT group investigated the effect of pressure during bottom-fed fluidized bed gasification of peat and wood. The results indicate that increased pressure favours the formation of NH3 and the NH3/HCN ratio is increased (see [Leppälahti & Kurkela, 1991]). On the other hand, researchers at KTH found a slightly decreasing trend of the fuel nitrogen conversion to NH3 with increasing pressure in their top-fed fluidised bed gasifier for birch gasification [Rosén et al., 1997].

25

[Nichols et al., 1987] found that the NH3 concentration increased, whereas the HCN concentration decreased with increasing pressure in an entrained flow coal gasifier. [Zand et al., 1985] reported that pressure had no measurable effect on the production rates of the nitrogenous compounds for fluidised bed peat and lignite gasification, although their applied pressure range was limited (770-830 kPa). The reported influence of pressure on fuel nitrogen partitioning is such that at this stage no conclusive statements can be made. In this respect, the opposing observations by KTH and VTT can possibly be related to their different feed location configurations; KTH operated a top fed gasifier, versus VTT and most of the mentioned references a bottom fed gasifier. 2.4.5 Influence of additives Additives in the fluidised bed gasification of solid fuels are used for several purposes: • reducing gaseous sulphur emissions, effective for relatively high sulphur containing fuels, see e.g. [Thambimuthu, 1993] and [Kurkela et al., 1996]; • cracking of tars, see e.g. [Milne et al., 1998]; • prevention of agglomeration, see for instance [Moilanen & Kurkela, 1998] and [Turn et al, 1998b]; • catalysis of the process for e.g. hydrogen yield optimisation, see e.g. [Rapagna et al., 2000 & 1998]. Especially for the first three purposes, dolomite or limestone find widespread application. The use of (calcined) dolomite and/or limestone as additive in FB gasification resulted in decreasing HCN and increasing NH3 concentrations, see e.g. [Leppälahti & Kurkela, 1991]; [Leppälahti & Koljonen, 1995]. This effect was reported in literature, where CaO was suggested to react with HCN, forming a relatively instable CaCN2 compound, see [Leppälahti et al., 1991]. The reaction of CaCN2 with water (vapour) back to CaO, forming CO and NH3, is thermodynamically favoured, see [Jensen, 1996]. Also, [Berg et al., 2001] cited confidential reports on biomass gasification ([Waldheim et al., 1999] and [Blackadder et al., 1995]), showing that dolomite and limestone addition favoured NH3 formation at the expense of HCN. [Paterson et al., 2002] experimentally found that in their pressurised spouted bed gasifier addition of limestone to Daw Mill Coal led to increased NH3 concentration in the exit gas. Iron in the fuel was reported to decrease the fuel nitrogen conversion into NH3 significantly, see [Leppälahti, 1998] and [Leppälahti et al., 1991]. 2.4.6 Influence of particle diameter From their CFB gasification observations using differently sized wood particles as feedstock, [Van der Drift & Van Doorn, 2001] report that sawdust with a maximum diameter of 2 mm shows lower fuel nitrogen conversions into NH3 than increasingly larger cylindric willow wood particles of 10x10 and 10x40 mm. When small sized sawdust was mixed with the larger cilindric wood particles lower NH3 conversion values were observed. The authors did not explain this effect for the fuel nitrogen behaviour, but reported that the riser showed significantly different temperature profiles. With fine sawdust large temperature differences are created over the reactor, accompanied with relatively high maximum temperatures, whereas with the larger particles a more even temperature distribution was observed with lower maximum temperature. Possibly the higher temperatures obtained by feeding fine sawdust have contributed to NH3 destruction.

26

2.4.7 Influence of steam The use of steam prevents peak temperatures that can lead to ash sintering and melting. Steamflow is therefore a process parameter used to avoid operational problems caused by sintering [Moilanen&Kurkela, 1998]. Addition of steam in air/N2 coal gasification of Daw Mill Coal in a small scale spouted bed reactor at a pressure of 1.3 MPa and a temperature range of 770 (upper bed)-900 °C (spout) lead to almost complete conversion of the fuel bound nitrogen into NH3, see [Paterson et al., 2002] and [Zhuo et al., 2002]. Significantly lower fuel-N conversions into NH3 were reported when no steam was used as (almost a factor 10 lower). The authors attributed the observed effect to the reaction of steam or H2, produced by steam gasification, with char bound nitrogen. The observation confirms experimental results with steam gasification of a range of coals in a combined fluidised bed – fixed bed reactor system by [Chang et al., 2003]. 2.4.8 Influence of feed location At KTH (Stockholm, Sweden) fluidised bed biomass gasification research is carried out since about two decades. Pressurised gasification is performed in a top-fed 80 kWth process demonstration unit. Relatively low conversion values of fuel bound nitrogen into NH3 and HCN are reported as compared to other fluidised bed units, which are basically all bottom fed reactors. For birch wood gasification at 0.4 MPa and approximately 900 °C, fuel-N conversion to NH3 amounted to 12-21%. For conversion into HCN these values were in the range 0-0.7% and virtually no char-N was left. For experiments at the same pressure, but at temperatures of appoximately 700 °C, the fuel-nitrogen conversion to NH3 was as low as 0-1.6%, conversion into HCN was 0-0.1%, and char-N amounted to 2.6-13.6%. NO formation was observed in the range 0-11.1%. The particle size of the wood feedstock was 1-3 mm. Nitrogen content of the fuel was ca. 0.1 wt% (db). Oxygen enriched air was used as gasifying medium [Chen, 1998]. In a small scale AFB gasifier, tests were performed to elucidate the difference in the fuel bound nitrogen behaviour for top feeding versus bottom feeding, see [Vriesman et al., 2000]. These authors varied feed location in a small scale AFB for gasification of miscanthus. Top feeding was shown to result into lower fuel bound nitrogen conversions into NH3 compared to bottom feeding. The difference was explained by the observed higher CO concentrations during top feeding, which leads to a higher yield of N2 from the reaction of NO and char and therefore a lower NH3 yield would result. The mentioned explanation looks plausible, although it can also be attributed to the different environment in which the initial fuel conversion step, fast pyrolysis, takes place. For top feeding this environment is reducing, whereas it is for bottom feeding oxidizing. Also, particles face different heating rates for both configurations, which can be a source of differing initial nitrogen species yields.

27

2.5

Fluidised bed gasifier modelling

2.5.1 General overview Gasification models are essential tools for designing a gasifier, predicting the operating behaviour and emissions and describing dynamic behaviour during start-up, shutdown, fuel and load changes. For all types of modelling different details of the process have to be studied. Experiments, especially carried out at large scale, are often too expensive or complicated. On the other hand, models are always a simplification of the complex reality and improvements are needed. The thermal conversion of a solid fuel in fluidised bed gasifier proceeds along the following steps, which in practice occur partly sequentially and partly in parallel: • fast particle heating up, • swelling (coals), • drying, • pyrolysis, • volatile combustion, • char combustion, • char gasification, • char fragmentation, • char elutriation. During drying of the solid particle, its temperature remains practically constant until the moisture has been completely released. A further increase in temperature with the high heating rates prevailing in a fluidised bed initiates socalled fast pyrolysis. Heating rates can vary between 103 – 106 K/s, depending mainly on fuel particle size [Elliot, 1981]. As an illustration that biomass pyrolysis in a fluidised bed is a fast phenomenon, calculations by [Di Felice et al., 1999] show that for particle diameters of 1-10 mm the conversion time was in the range of 5-22 s. [Rumpel, 2000] experimentally demonstrated that 5 mm cubic wood particles showed pyrolysis conversion times of ca. 20s. Fuel pyrolysis can be considered to occur in two steps: primary and secondary pyrolysis [Chen & Niksa, 1992]. During primary pyrolysis, both gases and (condensable) tar are formed (all together called ‘volatiles’), leaving a solid carbonaceous and mineral matter behind, called ‘char’. Subsequently, during secondary pyrolysis, tar may crack forming gases and soot. Also, further release of gases from primary char may take place from more stable structures as compared to the initial gas release. Four main groups of pyrolysis products therefore result from this process: a rather porous char, gases, tar and soot. Figure 2.4 shows a schematic of the process [Jensen, 1999].

28

LIGHT GASES

CHAR

SOLID FUEL

: Primary conversion : Secondary conversion

TAR

SOOT

Figure 2.4 Schematic representation of solid fuel pyrolysis The gaseous volatiles released during pyrolysis consist of the components CO, CO2, CH4, H2, H2O, NH3, HCN, N2, NO, N2O, H2S, SO2 and other hydrocarbons. The pyrolysis chemistry and the ratio of the pyrolysis time to the residence time of the particles in the bed determine the distribution of volatiles over the bed. After the pyrolysis process the volatiles are burnt in a diffusion flame surrounding the particle or transported away before burning [Tullin et al., 1993]. Subsequently, these species can react homogeneously, or heterogeneously in the bubbles or in the more dense interstitial space between bed material particles (emulsion phase) and in the freeboard above the bed. Due to the large particle concentration in the emulsion phase, radical quenching may decrease the rate of homogeneous reactions, in which radicals play an essential role. Also, char ignition/oxidation takes place. Char combustion is a relatively fast gas-solid reaction compared to heterogeneous gasification reactions and occurs at the gas-solid interface. During pyrolysis, the pore structure of the particle develops, resulting in a highly increased porosity and an internal surface, which is much larger than the outer particle surface area. For fast heterogeneous reactions, like char oxidation, the particle conversion rate is almost completely determined by external mass transfer limitation, especially for larger particles, and not so much by the intrinsic chemical reaction rate. When the oxygen is almost used up, slower heterogeneous gasification reactions start to play a more prominent role (char-H2O, char-CO2 and to a significantly smaller extent char-H2). The development of the pore structure for these heterogeneous reactions is more important than the fast oxidation, as it determines the available char surface and therefore the overall solid consumption rate. In reality all pyrolysis-, combustion- and gasification reactions can take place in parallel to each other, depending on the particle size and fuel type [Padban, 2000]. For large particles a temperature gradient develops during heat-up, resulting in thermal stress within the particle. Devolatilisation, swelling, thermal stress and attrition can result in fuel particle fragmentation, which reduce the particle size significantly. For coal it was found experimentally that, dependent on the type, no fragmentation takes place for particles with diameters smaller than 2-3 mm [Brem, 1990]. Attrition in the freeboard can be neglected because of the much lower particle concentrations in that part of the fluidised bed. Models for fluidised bed gasification can be divided into: thermodynamic equilibrium and kinetic rate models. It is known that thermodynamic equilibrium is not achieved during fluidised bed gasification because of the relatively low operation temperatures (product gas temperatures at the outlet of gasifiers are generally between 750 and 1000 °C) and the relatively short residence times in FB gasifiers. An “approach temperature” analysis could be used, where a temperature for the equilibrium calculation is chosen that lies between the reactor and sampling probe temperature that reflects an approach of the main gas composition, see e.g. [Kersten, 2002]. This method, however, doesn’t seem to be generally applicable, especially in predicting minor emission components as they show a broad

29

range of thermodynamic stability. Nevertheless, gasification models based on equilibrium calculations have been widely applied because they can be useful for initial reactor design calculations and determination of the influence of operating parameters on main product yields and composition. Several gasification models can be found in literature, see e.g. for coal: [Gumz, 1952], [Denn et al., 1979], [Kosky & Floess, 1980], [Kovacik et al., 1990], [Watkinson et al., 1991] and for biomass: [Cousins, 1978], [Desrosiers, 1979], [Shand & Bridgwater, 1984], [Double & Bridgwater, 1985], [Bacon et al., 1985], [Double et al., 1989], [Shesh & Sunavala, 1990], [Kinoshita et al., 1991], [Kilpinen et al., 1991], [Buekens & Schoeters, 1994], [Ergüdenler & Ghaly, 1997], [Nordin et al., 1997], [Mansaray et al., 2000a], [Kersten, 2002] and [Sadaka et al., 2002a,b,c]. Table 2.6 presents an overview of the mathematical gasifier models based on kinetic models presented in the literature. Here, an extension is given to the overview presented by [Hamel, 2001] with respect to nitrogen speciation. Table 2.6 Overview of gasifier models with chemistry details taken into account R type FixB FixB FixB FixB

Heterogeneous reactions R1 R2 R3 R4 C K K K K C K K K K C K K K K CK K K K

R5 CC

FixB FixB FixB EFR EFR EFR EFR TC SB SB CFB CFB CFB

C C B C C C C B C CK C/B B B

K K CC K K K K K K K K K K

K K K K K K K K K K K K K

K K K K K K K K K K K K K

K K K K K K K K

C1 CC CC K K CC CC CC K K K

C2 C1 CC CC K EQ CC CC CC K K K

C1 CC K EQ CC CC CC K K*

K K EQ K K EQ EQ EQ K K K K

FB FB FB FB FB FB FB FB FB FB FB FB FB FB FB

C C C C C CK CK CK CK C/B B B B B B

K CC K K CC CC K K CC CC K K

K K K K K K K K K K K K K K K

K K K K K K K K K K K

K K K K K K K K K -

K K K CC K CC CC K K

K K K CC K CC CC K K

K CC K CC CC K K

K K EQ K ? EQ K EQ EQ K K EQ K

F

Homogeneous reactions R6 R7 R8 R9 K K EQ CC CC EQ -

R10 C2 -

N-Sp. -

K K EQ K K K

CC K K K K

EQ K

EQ K K

K K K K

K

* formation of CO instead of CO2

Explanation of symbols used in table 2.6: R type = Reactor type: EFR = Entrained Flow Reactor FB = Fluidised Bed CFB = Circulating Fluidised Bed FixB = Fixed Bed SB = Spouted Bed TC = Turbulence Chamber F = Fuel type:

30

B = Biomass C = Coal CK = Charcoal

Dry

Pyr

P

literature

1

1

C C2 -

C C2 -

P P P P

C2 C2 C2 K 2 C 2 C C2 C2 C2 C1 C2 C2 C2 C2

K K

K K K K C2 K K K K

P P A P P P A A A P A/P A A

[Biba et al., 1976 a,b;1978] [Schlich, 1977] [Yoon et al., 1978] [Arri & Amundson, 1978] [Amundson & Arri, 1978] [Mengis, 1983] [Guntermann & Franke, 1985] [Groeneveld, 1980] [Wen & Chaung, 1979] [Govind & Shah,1984] [Gockel, 1994] [Chen et al. 1999, 2000 a,b] [Pfab, 2001] [Lucas et al., 1998] [Bi et al., 1997] [Hamel, 2001] [Jennen, 2000] [Liu & Gibbs, 2003]

C2 K C2 C2 C2 K K C2 C2 C2 C2

A A P A A A P A P P A A A A/P A

[Weimer & Clough, 1981] [Neogi et al., 1986] [Rhinehart et al., 1987] [Yan et al., 1998, 2000] [Kim et al.,2000] [Caram&Amundson, 1979a,b] [Purdy et al., 1981] [Sett&Bhattacharya,1988] [Sowa, 1991] [De Souza-Santos, 1989, 1994] [Raman et al., 1981] [v.d. Aarsen, 1985] [Jiang&Morey, 1992] [Bettagli et al., 1995] [El Asri et al., 2001]

→ CO2 (g) C (s) + O2 (g) C (s) + ½ O2 (g) → CO (g) R CO (g) + H2 (g) R2: C (s) + H2O (g) R3: C (s) + CO2 (g) R 2 CO (g) R CH4 (g) R4: C (s) + 2 H2 (g) → CO2 (g) R5: CO (g) + ½ O2 (g) → H2O (g) R6: H2 (g) + ½ O2 (g) R7: CH4(g) + 2 O2 (g) → CO2 (g) + 2 H2O (g) R CO2 (g) + H2 (g) R8: CO (g) + H2O (g) R9: CH4 (g) + H2O (g) R CO (g) + 3 H2 (g) R10: tar oxidation or reforming N-Sp.: nitrogen species speciation considered Dry.: fuel drying taken into account Pyr.: fuel pyrolysis taken into account P: gasifier pressure with A = atmospheric R1:

or

In table 2.6 the following symbols are used:

P = pressurised

CC = complete stoichiometric conversion K = kinetic reaction rate calculation EQ = chemical equilibrium calculation C1 = constant reaction/devolatilisation rate C2 = instantaneously converted after feeding

Several authors (e.g. [Norman et al., 1997], [Leppälahti, 1998], [Zhou, 1998], [Padban, 2000]) concluded that for industrial-scale gasification of coal and biomass in fluidised bed gasifiers, the fate of nitrogen species is controlled by chemical kinetics. Thermodynamic equilibrium models predict that N2 is the most stable compound under typical fluidised bed conditions and thus these cannot be used to predict nitrogeneous emissions from fluidised bed gasifiers. Concerning nitrogen speciation in gasifier modelling, NOx precursor formation and reduction mechanism classification is an important premise for a systematic approach to understanding which reactions are effective and how these can be affected. The formation and destruction reactions usually occur simultaneously and are intimately coupled. NOx and NOx precursor formation mechanisms can be classified in two groups [Jensen, 1999]: 1) N2 based mechanisms (occurring at high temperatures): • Thermal NOx formation (see e.g. [Bowman, 1992]) • Prompt NOx ([Fenimore, 1972]) 2) Fuel_N mechanisms (preceded by pyrolysis): • Gas phase reactions i. Homogeneous gas phase ii. Heterogeneously catalysed gas phase iii. Tar_N based • Solid-gas reactions i. Char_N ii. Soot_N (to a much minor extent compared to char_N) Thermal NOx formation, resulting from reactions of N2 and N radicals derived from N2 with O2, O and (to a much smaller extend) OH radicals, is low in fluidised bed gasifiers as this mechanism only becomes important from ca. 1300 °C and higher. Prompt NOx formation, by reaction of N2 and hydrocarbon radicals, will be very small due to the low hydrocarbon content of product gas from the solid fuel and due to the low temperature, resulting in low CH-radical concentrations, see [Miller & Bowman, 1989]. For atmospheric fluidised bed combustion it was found that concentrations of oxides of nitrogen did not change when changing gas

31

atmosphere from air to argon/oxygen and it can be concluded that abovementioned mechanisms don’t play a significant role [Oude Lohuis et al., 1992]. Therefore, under the process conditions prevailing in fluidised bed gasification, thermal and prompt NOx formation are considered to be negligible. In this study, no further discussion will be devoted to these mechanisms. On the other hand fuel_N mechanisms are important for this work. Attention will be paid mainly to gas phase reaction mechanisms and to a lesser extent to heterogeneous reactions, because biomass is a highly volatile fuel where the main contribution to emissions of NOx precursors will be from gas phase reactions. The following subparagraphs are dedicated to the provision of experimental information that is needed as model input for the gasification subprocesses. 2.5.2 Drying and flash pyrolysis, initial steps in the process 2.5.2.1 Experimental techniques and findings There have been numerous studies concerning pyrolysis of older fuels like brown coal and coal, see e.g. [Howard, 1981], [Gavalas, 1982], [Wall, 1987] and [Solomon et al., 1992]. In this thesis, emphasis is put on biomass. Pyrolysis of biomass, especially the fate of its main constituents, cellulose, hemi-cellulose and lignin [Gaur & Reed, 1998], has been studied experimentally to determine kinetic conversion characteristics, using several different techniques. These techniques can be divided into two groups based on the heating rate: slow and fast heating. Very common in fuel characterisation is the application of the TGA, characterised by low heating rates till about 100 K/min under inert or reactive atmosphere. For higher heating rates, electrically heated wire mesh (also called heated grids or screen heaters), drop tube reactors/entrained flow reactors and Curie-Point reactors are used. See table 2.7 for an overview of pyrolysis characterization equipment. Table 2.7. Overview of experimental equipment for pyrolysis research. [Jahns, 1990] Thermo Gravimetric Analyser (TGA)

Heated grid / Screen heater/ Wire-mesh reactor

Convection

Electrical

MS, FTIR 1 –100 K/min 300-1500 K 1-100 bar N2, H2, He, H2O 1.5 g 0.8-1.0 mm

MS, IR abs. FTIR 100-1000 K/s 300-3000 K 0.1-100 bar He, H2, N2, CO2 1-10 mg 0.05-2.0 mm

Curie-Point reactor

Entrained flow reactor/ Drop tube furnace

Equipment

Heating method Gas analysis Heating rate Final temperature Pressure Gas atmosphere Sample size Particle size

Electro Magnetic Induction MS, FTIR 500-104 K/s 700-1300 K 0.1-300 bar N2, H2 5 mg 0.05-0.8 mm

Convection / radiation MS, FTIR 5x104 K/s 300-2000K 1-200 bar N2, H2, CO, CO2, H2O 1-10 kg/h <0.8 mm

The heating rate, holding time, (volatile) residence time, and the peak temperature all affect the fuel pyrolysis. During the primary pyrolysis reactions, volatiles are formed. These can further crack during secondary and tertiary reactions. The secondary cracking, which is not always avoidable at higher temperatures, can be minimised by using a reactor where the produced gas is quickly removed and cooled. To determine the flash pyrolysis reactions and the yields, only the primary reactions and cracking are important for this research. Gas residence time therefore needs to be kept as short as possible. This means that as soon as volatiles and other pyrolysis products are formed, the temperature should be

32

decreased with a high cooling rate and brought to ambient temperature: quenching. However, it is not always specified in literature whether the residence time portrays the solids or gas residence time, and whether the process conditions make secondary and tertiary cracking possible. 2.5.2.1.1 Main components and hydrocarbons Light pyrolysis gases produced during flash pyrolysis mainly consist of CO, H2, H2O, CO2 and light hydrocarbons. The composition depends on fuel type, temperature, pressure, gas environment, particle size and heating rate. For coals the structure of the heavy hydrocarbons generated is quite similar to that of the parent fuel, though somewhat enriched in hydrogen. Char formed during pyrolysis is carbon rich, whereas H and O content are depleted compared to the parent fuel. The main constituents of biomass are: cellulose, hemi-cellulose and lignin. Table 2.8 shows overage values of the amount of these compounds for some biomass types. Table 2.8 Main biomass constituent composition (mass%) [ECN-phyllis, 2003] Untreated wood Straw

Lignin 26 16

Cellulose 41 39

Hemi-cellulose 24 23

Ash & components 9 22

Cellulose is a polysaccharide (C6H10O5)x of glucose units that constitute the main part of the cell walls of plants. It occurs naturally in such fibrous products as cotton and kapok and is the raw material of many manufactured goods (as paper, rayon, and cellophane). Most biomass material consists of about 40-50 wt% cellulose. The cellulose structure consists of upto 14000 linearly coupled cellulose molecule (D-glucopyranoside) units connected by β-glycosidic linkages in 1:4 fashion [Solomons, 1984]. These linkages are commonly known as weak bonds. They are easily broken and help initiate the degradation of the cellulose molecule. Figure 2.5 shows a mechanistic view of the pyrolytic cracking of cellulose, see [Lewellen et al., 1976].

β-glycosidic bond

Figure 2.5 Pyrolysis kinetics of cellulose. Hemi-celluloses are polysaccharides that are related to the cellulose found in the cell walls of plants (hexosans, pentosans). They serve the plants as a frame and as reserve substances, providing glucose and other sugars during hydrolysis. Hemi-celluloses consist of small molecules containing 50 to 200 monosaccharide residues and contain linkages different from those in celluloses. They are classified by the kind of main monosaccharide in their structure. The types are D-xylans, L-arabino- D-xylans, D-mannans, D-galacto-D-mannans, D-gluco-D-mannans, and L-arabino-D-galactans, which can be grouped into xylans, glucomannans, and arabinogalactans. The basic structure of xylans is found as the linear backbone of 1,4-anhydro-D-xylopyranose in plants. The glucomannans in wood have small molecular weights and are primarily linear in structure. Arabinogalactans are water-soluble polysaccharides and are diverse regarding physicochemical properties. They are highly branched.

33

The distribution of hemi-cellulose in plant tissue varies with different species. The fractions present in hardwoods are more soluble compared to softwoods. The soluble fractions of hardwood contain glucose, xylose, mannose, and galactose, while those of softwood are arabinoglucoronoxylan.

a

b

Figure 2.6 Typical structure of hemi-cellulose (a), with different monosaccharides (b) Lignin is an amorphous polymer consisting of a three-dimensional network of aromatic sub-structures that provides rigidity and, together with cellulose, forms the woody cell walls of plants and the cementing material between them. For a schematic of the structure see figure 2.7. Most biomass consists of 20 to 30 wt% lignin, which is deposited between the space of the cells, called lignification. As can be seen in table 2.8, straw contains less lignin than wood, which results in a much more pronounced rigidity of the latter. The three classes of plant lignins are gynosperm (softwood), angiosperm (hardwood), and grass lignins. Lignin is primarily aromatic in nature, which can be seen by looking at the chemical structure: phenylpropane polymer (upper right branch of the Guaiacyglycerol-β-aryl ether structure).

34

Guaiacyglycerolβ-aryl ether structure

Figure 2.7 Partial molecular structure of lignin from birchwood [Struschka, 1993] For the research work described in this thesis a selection of consistent pyrolysis data obtained at high heating rates (flash pyrolysis) is given. Results of the pyrolysis of cellulose, a model component for biomass, are presented here. Furthermore, flash pyrolysis data for wood, straw and miscanthus will be dealt with in this section, as they are closely related to the experimental research described in this thesis. The effects of peak temperature, heating rate, fuel size, and pressure (when available) are analysed. Pyrolysis product yields of char, gas, tar, and water as well as the composition of the gas products are given. The equipment and fuels relevant for this study used by the various authors for whom data are presented, are listed in table 2.9. All use N2 as inert gas, with the exception of [Hajaligol et al., 1982], [Nunn et al., 1985a] and [Drummond&Drummond, 1996], who use helium.

35

Table 2.9: Pyrolysis equipment, process conditions and fuel data Institute

Temp range Pressure Heating rate Residence [°C ] [MPa] [K/s] Time [s]

Author

Equipment

[El Asri et al.,1999]

Fluidised bed

600-900

0.1

n.i.

[Van den Aarsen, 1985]

Fluidised bed

700-900

0.1

n.i.

University of Stuttgart, (Germany)

[Storm et al., 1999] [Rüdiger et al.,1996]

Entrained flow pyrolysis reactor

800-1200

0.1

n.i.

KTH, Stockholm, (Sweden)

[Zanzi et al., 1998] [Chen, 1998]

Free-fall tubular reactor

750-1100

0.1-5

n.i.

Free University of Brussels (Belgium) Twente University (The Netherlands)

Université Paul Sabatier, Toulouse (France)

Fuel type

Fuel size [mm]

n.i.

Straw Sawdust

0.15-0.5 0.15-0.5

n.i.

Beech Wood

1-2

Straw Miscanthus

0.75-4 1.5-6

Birch Wood

0.5-1.3 chips

Straw

0.5-1.3 pellets and chopped

2-5

n.i.

[Corté et al., 1987]

Vertical electrical furnaces

700-900

0.1

250-300

0.1-1

Beech wood Cellulose

2 2

MIT, Cambridge, (USA) University of Connecticut, Storrs (USA)

[Hajaligol et al., 1982] [Nunn et al., 1985a]

Electrical screen heater

300-1200

0.13

≤100-15000

0-30

Cellulose Sweet GumHardwood

1 sheets 100 layers of 0.045-0.088

[Avni et al., 1985]

Electrical screen heater

300-900

<0.1-0.1

≤600

?

Lignin

?

Imperial College, London (UK)

[Drummond &Drummond, 1996]

Electrical screen heater

300-900

0.1

0.1-1000

?

Bagasse, Silver Birch

?

University of New South Wales (Australia)

[Stubington&Aiman, 1994]

Electrical s creen heater

300-1100

0.1

200-10000

0-30

Bagasse

0.064-0.422

n.i.: not indicated

Cellulose [Hajaligol et al., 1982] studied the effect of various process conditions on the yields and compositions from the flash pyrolysis of cellulose using a heated grid reactor. The cellulose samples used in this work were approximately 100 mg in the form of thin strips of filter paper of dimensions 2x6x0.01 cm, and elemental composition: C, 43.96 wt%, H, 6.23 wt%, O, 49.82 wt%. To ensure the rapid dilution and quenching of the volatiles upon exiting the hot zone, the gas within the reactor was kept close to room temperature. The results are shown in fig 2.8a and b. These figures show the effects at a heating rate of 1000 K/s. Here it is shown that measurable decomposition occurs between 300 and 400°C, which increases with temperature until almost 95 mass% of the fuel is converted at 750 °C. Also it can be seen that most of the weight loss occurs between 500 and 700 °C. At temperatures above 800 °C, the char yield increases slightly, probably as a result of carbon deposition arising from secondary cracking of tar and light oxygenated volatiles. This explains the decrease in tar yield at temperatures above 700 °C as well. Secondary reactions of some volatiles are possible because sufficiently high temperatures can be reached in the immediate neighbourhood of the heated grid and cellulose sample, despite the short residence time.

36

800 tar

700 600 500 400 300

char gas incl water

200 100 0 200

400

600 800 temperature [°C]

1000

hydrocarbon yields [g/kg of cellulose

yields [g/kg of cellulose]

900 20

15

10

C2H4 CH3CHO CH4 CH3OH Ac+Fu C3H6 C2H6 CO CO2 H2O

250

200

150

100

5

0 300

50

500

700 peak temperature

900

CO CO2 H2O yields [g/kg of cellulose]

25

1000

0 1100

a b Figure 2.8 Cellulose product yields (a) and gas species yield (b) (P=0.135 MPa, dp=0.1 mm, dT/dt=1000 K/s) [Hajaligol et al., 1982] The gas consists mainly of CO, CO2, and H2O; hydrocarbon yields are much lower as well as hydrogen (not indicated in the graph). The gas yield increases steadily until asymptotes are reached at peak temperatures ranging from about 700 °C for H2O to about 900-950 °C for H2, CH4, and C2H4. The other gases reach the asymptote at about 800 °C. The yields of H2, CO, CH4, C2H4, and C2H6 increase markedly above the 650-750°C temperature range where tar yield has peaked. The authors suggest that a significant, and perhaps the dominant, contribution to the yield of these products is caused by tar decomposition rather than by primary degradation of the parent cellulose. The solids residence time, or holding time, is defined by [Hajaligol et al., 1982] as the time the sample and heated grid are maintained at a certain temperature before the start of cooling. The effect on char yield is plotted in figure 2.9a. The effects on tar and gas (including water) yield are plotted in figure 2.9b and c.

37

0s

1000

1000

2s 0s

1000 800

2 5s

1000 800

] ] tar yield [g/kg cellulose tar yield [g/kg cellulose

] ] charchar yieldyield [g/kg cellulose [g/kg cellulose

a a

5 10s s

800 600

10 30 s

30 s

600 400 400 200 200 0 0200 200

600 400 400 200

1000

1200

400 Peak600 800 [C] 1000 Temperature

1200

200

600

800

] ] gas gas yieldyield [g/kg cellulose [g/kg cellulose

Peak Temperature [C] 600 600 500 500 400 400 300 300 200

0s 2s 0s 5 ss 2 10ss 5 30 ss 10 30 s

800 600

200 0 0200

400

b b

400

600

800

400Peak Temperature 600 800 [C]

1000

1000

Peak Temperature [C] 0s 2s 0s 5s 2s 10 s 5s 30 s 10 s 30 s

c c

200 100 100 0 0200 200

400

600

800

400 Peak 600 800 Temperature [C]

1000

1000

Peak Temperature [C]

Figure 2.9 Effect of holding time on char (a), tar (b) and gas, including water (c) yields in [g/kg] of cellulose versus peak temperature (P=0.135 MPa, dp=0.1 mm, dT/dt=1000 K/s) [Hajaligol et al., 1982] Below 800 °C the tar and gas yields increase with increasing holding time, while char yields decrease. Above 800 °C, the holding time has little or no influence on the yields. These effects are believed to be due to the incomplete decomposition at lower temperatures and short holding times, while at higher temperatures the effect of holding time is small because the reactions occur rapidly enough for maximum decomposition. The effect of holding time on the water yield is different from the effect on other volatiles. Water reaches an asymptote at a temperature of 600 °C and 2-3 s of 51 [g/kg], while the other gases reach an asymptote at around 1000 °C. The effect of heating rate on char, tar and gas yields as found by [Hajaligol et al., 1982] are shown in figures 2.10 a through c. Interesting is that at a given peak temperature, the volatile conversion (sum of tar and gas yield) increases as the heating rate decreases. The same behaviour is found for the gas and tar yields below about 700 and 750 °C, respectively. This is explained by the fact that at the lower heating rates sufficient time is available for conversion during the heat-up period. Tar yields at higher heating rates show a maximum. The tar yields at lower heating rates are high and do not decrease above a certain peak temperature. This is because the secondary reactions in this case are minimal. The conditions allow sufficient time during the sample heat-up period for most of the tar to be generated and to escape from the immediate neighbourhood of the heated grid before that region becomes hot enough for significant cracking.

38

K/s 10-15000 °C/s 1000 °C/s K/s 350 °C/s K/s 100 °C/s K/s

1000 800 600 400 200 0 200

400

600

800

1000

b

1000 tar yield [g/kg cellulose]

char yield [g/kg cellulose]

a

800 600 400 200 0 200

1200

400

Peak Temperature [°C]

gas yield [g/kg cellulose]

600 500

K/s 10-15000 °C/s 1000 °C/s K/s 350 °C/s K/s 100 °C/s K/s

600

800

1000

Peak Temperature [°C] K/s 10-15000 °C/s 1000 °C/s K/s 350 °C/s K/s 100 °C/s K/s

c

400 300 200 100 0 200

400

600

800

1000

Peak Temperature [°C]

Figure 2.10 Effect of heating rate for cellulose pyrolysis on char (a), tar (b) and gas (c) yield versus peak temperature (p = 0.345 bar, dp = 0.0010 mm, dT/dt=100-15000 K/s) [Hajaligol et al., 1982] [Hajaligol et al., 1993] also investigated the effects of pressure on tar release. The cooling profile is slightly influenced by pressure. Cooling of the installation under vacuum was established by radiation, but at higher pressures by radiation and free convection. The cooling rate at 1.3 atm was about 200 K/s. This was higher at higher pressures due to enhanced free convection. Figure 2.11 a-c shows the yields at vacuum, 1.3 bar and 69 bar for various temperatures up to 1100°C. Noteworthy is the maximum in tar yield at each pressure, and the substantial reductions in tar yields with increasing pressure at temperatures above 650 °C – 750 °C. These observations suggest that above a certain temperature, the measured tar yields are the result of competing processes, at least one of which is pressure dependent. Plausible mechanisms are tar generation by cellulose thermal decomposition, tar destruction by secondary reactions, and tar release by mass transfer. The tar yields, including the magnitude and location of the maximum at each pressure, are determined by the kinetics of these processes and by the available time. The lower tar yields at increasing pressure are ascribed to pressure effects on tar transport and tar secondary reactions. The char yield at temperatures above about 650 °C increases with the increase of pressure from 1.3 to 69 bar, but decreases when pressure is increased from vacuum to 1.3 bar. Decreasing pressure is expected to diminish the opportunity for secondary reactions of the volatiles, by accelerating their escape from the hot-zone. Thus to the extent that char results from tar secondary reactions, lower char yields would be expected as pressure decreases. This is observed above 800 °C when the pressure is decreased from 69 to 1.3 bar but not when pressure is further decreased from 1.3 to vacuum. This suggests that pressure effects on char yield in the heated grid reactor are complex, though limited.

39

vacuum 1.3 bar 69 bar

600 400 200 0 200

400 600 800 1000 Peak Temperature [C]

1000

vacuum

b

1.3 bar 800

69 bar

600 400 200 0 200

400 600 800 1000 Peak Temperature [C]

800 gas yield [g/kg cellulose]

800

a

tar yield [g/kg cellulose]

char yield [g/kg cellulose]

1000

vacuum

c

1.3 bar 600

69 bar

400 200 0 200 400 600 800 1000 1200 Peak Temperature [C]

Figure 2.11 Effect of pressure on char (a), tar (b), and gas (c) yields for cellulose pyrolysis versus peak temperature (dp = 0.01 cm, dT = 1000 K/s) [Hajaligol et al. 1982]

Wood Experimental results from fluidised bed wood pyrolysis by [El Asri et al., 1999] are presented in figure 2.12 a, which shows the product yields of char, tar, dry gas and water as a function of temperature for the temperature range 600-900 °C. The results indicate that: • the gas yield increases from 30% at 600 °C to nearly 80% at 900 °C on mass basis; • the water yield decreases slightly from 240 to 150 [g/kg]; • tar yield shows a strong decrease from 280 to 20 [g/kg]; • char yield decreases to a nearly constant value of 30 [g/kg]. Figure 2.12 b shows the gas composition in [g/kg] of total feed. The effects observed are: • the gas consists primarily of CO, CO2 and CH4 and the CO yield tends to increase sharply until 700 °C, continuing to increase with a lower increment, though care should be taken to draw hard conclusions, as no indication of accuracies was given; • the yield of CO2 reaches a maximum between 700 and 800 °C; • the CH4 and C2H4 product yields increase sharply until 800 °C after which they remain fairly constant; • the H2 gas starts forming significantly around 600 °C and shows a sharp increase from 800 to 900 °C; • the yield of C2H2 decreases sharply until 700 °C after which the yield remains fairly constant. • C2H6 shows a maximum yield at 700 °C.

40

200

600

1000

CO

600

gas [g/kg of the total feed] .

yields [g/kg total feed]

800 tar

400

dry gas

200

400 C02 100 CH4

C2H2

H2

char 0 600

700 800 Temperature [°C]

a

900

200

C2H4

C2H6

0 600

CO [g/kg of the total feed] .

water

800 Temperature [°C]

0

1000

b

Figure 2.12 Sawdust: pyrolysis yields versus bed temperatur; char, tar and gas (a) and gas species(b) (p = 0.1 MPa, dp = 0.15-0.5 mm) [El Asri et al., 1999] [Nunn et al., 1985a] investigated Sweet Gum Hardwood over a broader temperature range using a heated grid reactor. They analysed all main products except H2, which allowed the determination of material and elemental balances. Secondary reactions were minimised, though not eliminated, by removing the formed gasses rapidly from the hot (reaction) region. The values are given on a dry basis. The pyrolysis product yields are shown in figure 2.13 a. Yields of CO, CO2, H2O and some hydrocarbons are given in figure 2.13 b. When the temperature is raised to values above 700 °C, most of the reactive tar is converted to light volatiles or relatively stable tars. Effects observed from these figures are: • both gases and tars begin to evolve around 300 °C peak temperature and increase steadily while the char yield decreases; • most of the wood weight loss occurs between 400 and 600 °C; • above 700 °C, the tar yield decreases, and the gas yield increases, while the char yield remains constant, indicating secondary tar cracking; • the CO yield increases significantly between 500 and 800 °C; • the yield of most of the gases reaches an asymptote around 700 °C; • CH4 and C2H4 yields continue increasing, but at a lower rate, above 700 °C. Significant additional amounts of CO are produced at temperatures above 600 °C where the weight loss, and thus the primary decomposition of the sample, has essentially ceased. Nunn concluded that CO is also a major product of secondary tar cracking at elevated temperatures. The CH4 yield increases with 50 [g/kgdry] from about 900 to about 1250 °C. [Nunn et al., 1985a] claim that the experimental values of the tar yield in this temperature range have an uncertainty range of about 50 [g/kgdry]. The authors conclude that this implies that secondary tar cracking continues up to about 1250°C, which is the highest temperature studied. They do not give product yield data for C2H2. [El Asri et al., 1999] show this component to decrease with increasing temperature, which can serve as an explanation for the increase in C2H4 yield.

41

1000

500

gas incl water

char

0 227

60

CO CO2

50

H2O

40

140 120 80

CH4 CH2O C2H4 C3H6 C2H6

20 10

627 1027 temperature [°C]

160

100

30

0 300

180

500

700

900

1100

60 40

CO yield [g/kg dry]

tar

gas yields [g/kg dry]

yields [g/kg dry]

70

20 0

1300

Peak Temperature [°C]

a

b

Figure 2.13 Sweet Gum Hardwood: dry gas composition yields [g/kgdaf] versus bed temperature (p = 1.35 bar, dp = 45-88 µm, dT/dt = 1000 K/s) [Nunn et al., 1985a] [Corté et al., 1987] studied the influence of temperature on fast pyrolysis of beech wood in an electrically heated furnace. The data reported covered only the gas yields: no tar or char yields were presented. The ultimate and proximate analyses were not published either. These authors found that the optimum temperature range for producing fuel gas is about 800 to 900 °C; above this temperature, both the hydrocarbon yield and the heating value decrease. The higher temperature range of 900 to 1000°C, however, is an optimum for the production of synthesis gas that is mainly composed of a mixture of H2 and CO (which is easier to see in a figure on vol% gas basis). From figure 2.14, summarizing their gas species yield results, one can conclude that: the CO yield doubles when the temperature increases from 700 to 1000 °C; the product gas H2 yield nearly triples in the temperature range of 700 to 1000 °C; the yield of CO2 is hardly influenced by temperature and varies between 140 and 170 [g/kgdaf]; the amount of CH4 produced reaches an asymptotic value of 83 [g/kgdaf] around 900 °C; the C2+ hydrocarbon yield reaches a maximum around 900 °C. 200

CO

CO2

150 Gas [g/kg daf]

600

400 CO [g/kg daf]

• • • • •

100 CH4 200 C2+

50

H2 0

0 700

800

900

1000

Peak Temperature [C]

Figure 2.14 Beech wood: gas yields for dry Beech wood pyrolysis versus temperature (dp = 2 mm, dT/dt = 250-300 K/s, N2 flow 1.00 l/min) [Corté et al., 1987]

42

[Van den Aarsen, 1985] studied the effect of pyrolysis on beech wood using a bench scale fluidised bed reactor. The author states that in fluidised bed pyrolysis, char is produced in small quantities (less than 10%) and thus pyrolysis in the fluidised bed applied at heating rates of ca. 500 K/s can be placed between slow pyrolysis processes (1 K/s) and flash pyrolysis processes (2.104 K/s) where 25% and 1% char are produced, respectively. Pyrolysis yields in his experiments add up to a perfect 1000 [g/kg] as the yields were corrected by the author to close the mass balance. The gas composition was also corrected to satisfy the elemental balance for C, H, and O.

600

200 water tar

500

600 400

gas

150

CO

300

100 CH4 50

char 815 Peak Temperature [°C]

915

0 700

200

H2 100

200 0 715

400

C2H4

CO yield [g/kg wet]

800

CO2

yield [g/kg wet]

yields [g/kg wet]

1000

0

800 900 Peak Temperature [C]

Figure 2.14 Beech wood: pyrolysis yields as a function of temperature for fluidised bed pyrolysis (p = 1bar, dp = 1-2 mm) [van den Aarsen, 1985] The size of the particle affects the heating pattern. The outside of the particle is heated first and quickest, while the inside of the particle is heated at a lower pace. The smaller the particle, the more uniform the heating pattern. Table 2.10, summarizing wood pyrolysis data from a free fall reactor, shows a decrease in gas yield with increasing particle size, while the tar yield remains unchanged and the char yield increases. Water and losses (sum of components that are not measured), calculated by subtracting the gas, tar and char yields from 1000 [g/kg], increases with particle size. [Zanzi et al. 1996, 1997 & 1998] concluded that the produced gas leaves the smaller particles faster than the large particles, resulting in a lower char yield for the smaller particles. This results in a longer gas residence time outside the particle and consequently cracking in the gas phase is favoured. The smaller sizes affect the composition of the gas in such a way that the cracking of hydrocarbons is favoured with an increase of hydrogen yield. Table 2.10 also shows, on a volume basis, that H2, and CO2 product yields decrease and that CH4, C2H2, C2H4 and CO yields increase, while the amounts of C2H6 and benzene formed remain constant. The influence remains limited on the gas composition.

43

Table 2.10 Birch wood: influence of particle size (pmax = 50 bar, dp = 0.5-1.0 mm) [Zanzi et al. 1998] Effect of particle size at 800 °C Particle size [mm] 0.5-0.8 0.8-1.0 Residence time [s] n.i. n.i. Gas yield [g/kgdaf] 811 777 Tar yield [g/kgdaf] 11 11 Char yield [g/kgdaf] 58 72 water and losses 120 140 Gas species vol% vol% dry dry gas gas H2 17.3 16.8 CH4 15.7 16.2 C2H2.C2H4 5.8 6.2 C2H6 0.3 0.3 Benzene 1.2 1.2 Toluene CO2 9.6 8.3 CO 50 50.7 total 99.9 99.7 * vol% nitrogen and water-free basis

[Zanzi et al., 1996 & 1998] investigated the influence of residence time. They varied the residence time of particles in the hot zone of a free fall reactor by changing the number of electrical heaters used. Thus an increase in particle residence time was accompanied by an increase in gas residence time. Temperature has a larger impact on the yields than residence time. This can be observed from the experimental pyrolysis results summarized in table 2.11. Increasing the residence time, increases the char yield at higher temperatures, and decreases the tar yield at lower temperatures. The other values remain fairly constant. Apparently, at longer residence times and at high temperatures the tar is converted into char. Note the increase in H2. At 750 °C, the influence of residence time is less drastic. The gas composition becomes constant at higher residence times, indicating saturation of secondary reactions at this temperature. Table 2.11 Fluidised bed wood pyrolysis: influence of residence time (pmax = 50 bar, dp = 0.5-0.7 mm, T = 750 and 900°C) [Zanzi et al., 1996]

90% birch and 10% aspen 7.8wt% moisture and 0.42 wt%dry ash Temperature 750 900 pressure [kPa] 280 270 300 260 270 Particle size [mm] 0.5-0.7 0.5-0.7 Particle residence time [s] 1.0 1.6 2.7 0.6 0.8 Tar yield [g/kg] 17 12 11 12 11 Char yield [g/kg] 72 72 72 52 55 Gas yield [g/kg] 732 730 735 813 813 Gas residence time [s] 5.6 8.2 11.9 2.6 3.7 Composition of gaseous products [vol% dry] H2 10.1 13.4 13.8 14.2 18.1 CH4 19.8 16.8 17.2 16.2 16.0 C2H2.C2H4 6.5 6.2 6.2 6.5 6.2 C2H6 2.3 1.5 1.3 1.0 0.4 butane 0.3 0.3 0.1 0.2 pentane 0.1 0.1 0.1 Benzene 0.4 0.4 0.5 0.3 0.3 Toluene 0.1 0.1 0.2 0.1 0.1 CO2 13.4 9.4 9.5 8.9 8.1 CO 47.0 51.8 51.2 52.6 50.8

44

270 1.7 11 59 811 6.8 21.0 16.0 4.4 0.2 0.4 8.2 49.8

As can be seen in a summary of the main pyrolysis yields in table 2.12, the gas yields from the various authors vary somewhat. [El Asri et al., 1999], [Zanzi et al., 1998], and [Van den Aarsen, 1985] show a gas yield between 700 and 800 g/kg at 800 °C. [El Asri et al., 1999] show a stronger increase in gas yield when raising the peak temperature. These authors report a lower yield at 700 °C but a higher at 900 °C. [Zanzi et al., 1998], using a free fall reactor, shows the highest yield. So, the differences are fairly insignificant taking into account that this is on a daf basis. [Nunn et al., 1985a] present a much lower gas yield because they use a heated grid reactor where the volatiles are quickly removed from the hot-zone, minimising the secondary reactions. They also report a higher tar yield than both [El Asri et al., 1999] and [Van den Aarsen, 1985]. [Van den Aarsen, 1985] is the only literature source that does not show a clear decrease in tar yield with an increase in temperature. [Zanzi et al., 1998] indicate the lowest tar yield. The water yield decreases with temperature and reaches about 130 [g/kg] at 900 °C. [El Asri et al., 1999] show the largest decrease, as compared to [Van den Aarsen, 1985]. Table 2.12 Wood: fast pyrolysis yields

char

[El Asri et al.,1999] [Van den Aarsen,1985] [Nunn et al., 1985a] [Zanzi et al.. 1998] sawdust beech wood gum hardwood birch 600°C 700°C 800°C 900°C 715°C 815°C 915°C 627°C 827°C 1270°C 800 1000 [g/kg] [g/kg] [g/kg] [g/kg] [g/kgwet] [g/kgwet] [g/kgwet] [g/kgdry] [g/kgdry] [g/kgdry] [g/kgdaf] [g/kgdaf] 130 120 30 30 92 57 51 130 70 70 72 56

dry gas

290

580

720

800

662

759

765

290

400

420

777

870

tar

270

125

20

20

60

60

60

550

490

460

11

2

water

240

220

170

130

186

125

124

In all cases, the produced gas consists mostly of CO. The CO2 concentration is also fairly high. The rest of the dry gas consists of H2 and CH4 with other hydrocarbons.. The CO concentration varies between 400 and 500 [g/kg] at 800 °C. The CH4 yield increases with temperature and varies between 60 and 75 [g/kg] at 800 °C. The CO yield increases sharply until 700 °C after which it continues to increase but at a lower rate, see [El Asri et al., 1999]. These authors found that the CO2 concentration reaches a maximum between 700 and 800 °C. The results given by [Zanzi et al., 1996] show that the gas yield decreases with increasing particle size at 800 °C. This is because the larger sizes have a less homogeneous heating profile: the outside is heated quickly, whereas the inside is heated slower. The tar yield remains the same at this temperature, and the char increases. Zanzi et al. 1996, 1997 & 1998) and [Rüdiger et al., 1995] give the same explanation of these effects. In smaller particles, the produced gas leaves the particles faster, lowering the char yield. The produced gas then has a longer residence time outside the particle while still at peak temperature and this increases the secondary reactions of the hydrocarbons and thus the H2 yield. [Zanzi et al. 1996 & 1998] investigated the influence of particle- and gas residence time; both were increased simultaneously. They found that the residence time has a smaller impact on the yields than the temperature. Increasing the residence time, increases the char yield at 900 °C, and decreases the tar yield at lower temperatures. The other values remain fairly constant. Apparently the tar reacts at longer residence times and at high temperatures to char and H2. At 750 °C, the influence of residence time is less drastic. The gas composition stabilises at higher residence times, indicating a saturation of secondary reactions at this temperature. The type of installation used for the pyrolysis influences the yields. The results obtained using a heated grid reactor show the strongest difference with the other installations [Nunn et al., 1985a]. This reactor makes it possible to quench the volatiles right after they are produced, which minimises the secondary reactions in which the tar is further cracked to produce gases. These authors provide evidence for this theory by showing that with increasing temperatures, the tar yield decreases and gas yield increases while the char yield remains constant.

45

Installations where the particle and gas residence time differ significantly, such as the fluidised bed reactor, and installations where the particle and gas residence times are practically equal, such as the free fall reactor, do not minimise the secondary reactions. Therefore, these installations produce higher gas yields. Straw Figures 2.16, 2.17 and 2.18 present the effects of peak temperature on the pyrolysis yields from straw according to the findings of [El Asri et al., 1999] and [Storm et al., 1999]. The figures showing the flash pyrolysis yields are in the form of stacked areas, giving an indication of the total mass balance. Figure 2.16a shows the influence of temperature on the pyrolysis yields of char, dry gas, tar, and water using a fluidised bed [El Asri et al., 1999]. Char-, tar-, and water yields all decrease while the dry gas yield increases with temperature, especially above 800°C. Figure 2.16b shows the yields of the main components of the gas CO-, CH4,-, and C2H4 yields increase with increasing temperatures. The yields of CO2 and H2 show a relatively strong increase above 800°C. The C2H2 yield decreases after reaching a maximum yield at 700°C. The yield of C2H6 is hardly influenced by the temperature. 100

600

water

tar

400 dry gas 200 char 0 600

600

90 CO

80 70 60 50

CH4 C2H2

C2H4

900

Figure 2.16 a Straw: pyrolysis yields versus bed temperature (p = 1 bar, dp = 0.15-0.5 mm) [El Asri et al., 1999]

400 300

40 30

CO2

20

H2

10

200 100

C2H6

0

700 800 Temperature [°C]

500

CO, CO2 [g/kg total feed] .

800

CH4, C2H4, C2h2, H2, C2H6 [g/kg total feed]

yields [g/kg total feed]

1000

0 600

700 800 900 Temperature [°C]

Figure 2.26 b Straw: gas composition yields (p = 1 bar, dp = 0.15-0.5 mm) [El Asri et al., 1999]

[Storm et al., 1999] and [Rüdiger et al., 1996] used an entrained flow pyrolysis reactor and determined the effect of temperature on the main gas yields (figure 2.17) and on light and heavy tars (figure 2.18). Water and char yields were not published because the papers concentrated on the gas and tar production during co-pyrolysis of biomass and coal mixtures. Figure 2.17 shows the yield of the main components of gas yields of straw: CO, CO2, CmHn, and H2, where CO is presented on the right axis. The CO evolution starts at 400°C and the H2 evolution at 600°C and both rapidly increase at higher temperatures. Hydrocarbon product yields reach a maximum around 800°C but remain below the CO yield. The CO2 yield is hardly influenced by temperature. The results are not entirely consistent with those of [El Asri et al., 1999], the main difference being the effects of temperature on CO2 and H2 yields. The CO2 yield remains fairly constant and the H2 yield slowly increases as opposed to the sharp increase around 800 °C indicated in figure 2.16 b.

46

CmHn

dry gas yields [g/kg daf]

600

CO2

180

H2

160

500

CO

140

400

120 100

300

80

200

60 40

CO yields [g/kg daf]

200

100

20 0

0 400

600

800

1000

1200

Temperature [°C]

Figure 2.17 Straw pyrolysis: main gas yields of straw (p =1 bar, dp=0.75-4 mm, 8.103

pyrolysis tars [g/kgdaf]

140 120 100 80 60

(El Asri '99)

40 20

heavy tar (Storm '99) (Zanzi '98)

0 400

600

800 1000 Temperature [°C]

1200

Figure 2.18 Straw pyrolysis: tar yields by [Storm et al., 1999] (p=1 bar, dp=0.75-4 mm, ts=40-50 ms) compared to [El Asri et al., 1999] (p=1 bar, dp=0.15-0.5 mm, T=600-900°C) and [Zanzi et al., 1998]. (p = not specified, dp=0.5-0.9 mm) The data on tar evolution by [El Asri et al., 1999] are not specified for different tars. However, considering the measured total yields of tar [Storm et al., 1999], they arrive at a similar conclusion: tar yields decrease with temperature. [Zanzi et al., 1998] experimented with chopped straw and pelletised straw at 800 °C and 1000 °C. For both types of straw the tar yield decreased from approximately 9 to 1 [g/kgdaf fuel]. [El Asri et al., 1999] and [Zanzi et al., 1998] show a lower tar yield. An explanation for the different results may be that the tar definitions used differ: there is disagreement on which components, such as benzene, should be included as tar and which ones should not. [Rüdiger et al., 1995] mentioned that tar quantification is not easy to carry out and imply a certain measurement uncertainty.

47

Table 2.13 shows the effect of temperature on the pyrolysis of pelletised wheat straw according to [Zanzi et al., 1998]. This information, with the exception of tar, is not presented in a graph because the measurements are done only at two temperatures. The tar and char yields decrease and the gas yield increases agreeing with the data presented earlier. Table 2.13 Straw pellets: pyrolysis yields at 800 and 1000 °C [Zanzi et al., 1997] (p = not specified, dp=0.5-0.9 mm) Temperature [°C] Gas [g/kgdaf] Tar [g/kgdaf] Char [g/kgdaf] water and losses [g/kgdaf] Total [g/kgdaf]

800 754 8 136 102 1000

1000 855 1 105 39 964.9

Gas component H2 CH4 C2H2.C2H4 C2H6 Benzene CO2 CO Total

[g/kg daf] 15.2 81.2 79.8 4.7 6.6 266.2 300.3 754

[g/kg daf] 37.6 61.9 2.6 bdl. bdl. 100.2 652.7 855

bdl = below detection limit [Rüdiger et al., 1995] used grounded straw with diameters between 1.5 and 4.0 mm. The influence of particle size on the gas and CO and CmHn yield is minimal. Figure 2.19 presents the total gas yields and the CO and CmHn concentrations for the two particle sizes studied. This shows that the gas yield at 400 °C and 1000 °C is similar in magnitude, while at 600, 800, and 1000 °C the larger particles show a lower yield. The lower heating rate of bigger particles was given as a reason for this behaviour. The longer residence time of the primary pyrolysis components in the bigger particles was given as a second reason, because this influences the primary reactions within the particle and the secondary reactions in the gaseous phase. This last conclusion, however, would imply that with a longer residence time more instead of less gas would be produced. The CO yield is lower at all temperatures except at 400 °C, while the CmHn product yield is lower at all temperature below 1000 °C. No significant influence on tar yield was observed in the range of particle sizes investigated. 600

1200 1.5mm

800

4.0mm

600 400

500

Gas Yield [g/kg daf]

Gas Yield [g/kg daf]

1000

CmHn 1.5 mm CmHn 4.0 mm CO 1.5 mm CO 4.0 mm

400 300 200 100

200 0 400

600 800 1000 1200 Peak Temperature [C]

0 200

400

600 800 1000 Peak Temperature [C]

1200

1400

Figure 2.19 Straw pyrolysis: gas yields (p = 1 bar, dp = 1.5 and 4 mm, ts = 40-50 ms) [Rüdiger et al, 1995] [Zanzi et al., 1997] used both chopped and pelletised straw. The fuel was milled and sieved to obtain a fraction with a uniform particle size of 0.5-0.9 mm. Table 2.14 shows the yield of products during rapid pyrolysis in a free fall reactor. The difference in yields is not very large. The residence times were not specified, except that the first step (flash pyrolysis) was completed within seconds.

48

Pelletised straw shows slightly smaller gas- and tar yields when compared to chopped straw, while the char yield is slightly larger. Pelletised straw has a lower H2 and CO2 yield, while the hydrocarbon- and CO yields are higher. Table 2.15 presents the gas composition of the pyrolysis products of chopped and pelletised straw in [vol%] and [g/kgdaf]. Table 2.14 Straw pyrolysis in a free fall reactor: influence of size and shape with temperature (pmax = 50 bar, dp = 0.5-0.9 mm, T = 800-1000°C) [Zanzi et al., 1997] Fuel form temperature Gas [g/kgdaf] Tar [g/kgdaf] Char [g/kgdaf] water and losses

Chopped 800 °C 758 9 132 101

1000 °C 860 1 108 31

Pelletised 800 °C 754 8 136 102

1000 °C 855 1 105 39

Table 2.15 Straw pellets: gas composition (vol%. nitrogen and water-free basis) (pmax = 50 bar, dp = 0.5-0.9 mm, dT = ? K/s) [Zanzi et al., 1997] composition temperature H2 CH4 C2H2.C2H4 C2H6 Benzene CO2 CO Total

Chopped [vol%] 800 °C 35 9.5 3.1 0.1 0.6 23.7 28 100

1000 °C 43.9 4.8 0.1 5 46.2 100

Pelletised Chopped [vol%] [g/kgdaf] 800 °C 1000 °C 800 °C 24.2 38.8 23.8 16.2 8.0 51.5 4.7 0.1 57.0 0.5 bdl 1.0 0.7 bdl 6.1 19.3 4.7 353.9 34.4 48.4 264.6 100 100 758 bdl = below detection limit

1000 °C 45.2 39.4 1.5 113.0 660.8 860

Pelletised [g/kgdaf] 800 °C 15.2 81.2 79.8 4.7 6.6 266.2 300.3 754

1000 °C 37.6 61.9 2.6 100.2 652.7 855

Increasing the temperature increases the gas yield, and decreases the char, tar and water yields. The char, tar and gas yields vary by author. This can be due to the reactor used. The yield differences at a temperature of 800 °C are as shown in table 2.16. Table 2.16 Straw pyrolysis: summary of yields at 800 °C. source

installation

flow

Char yield

Gas yield

Tar yield

[Zanzi et al.., 1997] [El Asri et al., 1999] [Storm et al., 1999] [Rüdiger et al., 1995]

Free fall reactor Fluid bed Entrained flow reactor

laminar turbulent

140 [g/kgdaf] 30 [g/kg]

754 [g/kgdaf] 570 [g/kg]

laminar

-

750 [g/kgdaf]

8 [g/kgdaf] 10 [g/kg] light: 135 [g/kgdaf] heavy: 40 [g/kgdaf]

Size [mm] 0.5-1.3 0.15-0.5 1.5

[Zanzi et al., 1997] reported a very high char- and gas yield when compared to [El Asri et al., 1999], while the tar yields are about the same. [Storm et al., 1999] and [Rüdiger et al., 1995] found the highest tar yields. One explanation of the different yields found by [El Asri et al., 1999] is that these authors used wet fuel which influences the water and gas yields. Another explanation is that they used a fluidised bed reactor where the flow is turbulent and the particle residence time fairly long. This provides enough time and opportunity to break down the particles giving a low char yield. Figure 2.20 shows that only the work of [El Asri et al., 1999] deviates strongly from the other authors at 800 °C.

49

gas yield [g/kg]

1000 800 600 400 200 0 200

El Asri [g/kg] Storm [g/kg daf] Zanzi pellets [g/kg daf] Zanzi chopped [g/kg daf] 400 600 800 1000 1200 Peak Temperature [C]

Figure 2.20 Straw pyrolysis: gas yields as reported by several authors Both [Zanzi et al., 1997] and [Storm et al., 1999] free fall/entrained flow installations characterised by laminar flow and both present the results on a daf basis. The gas yields are almost the same at 800 °C, while the tar yields differ significantly. This could be due to different tar definitions. [Rüdiger et al., 1995] defines light tars as aliphatic or aromatic species with a mole mass up to approximately 120 g/mol. The heavy tars are defined as aromatic species with two and more rings and include naphthalene, fluorene, dibenzofuran, phenanthrene, anthracene, fluoranthene, pyrene and chrysene. The high tar yield found by [Storm et al., 1999] and [Rüdiger et al, 1995] in combination with a high gas yield is unexpected. The results do not entirely contradict the secondary cracking theory as the heavy tar yield indeed decreases and the light tar yield increases. Above 800 °C, the light tar yield decreases in favour of the gas yield. Drawing conclusions is complicated by the fact that the char and water yields were not published. In all cases, the produced gas consists mostly of CO. The CO2 concentration, though, is also fairly high. The rest of the dry gas consists of H2 and hydrocarbons of which CH4 concentration is the highest. In summary, the influence of temperature on straw pyrolysis product yield was found to have trends as follows: • The CO concentration is about 300 [g/kg] at 800 °C; • [El Asri et al., 1999] show that the CH4 concentration increases with temperature. [Zanzi et al., 1997] shows a decrease with temperature; • [El Asri et al., 1999] and [Storm et al., 1999] show an increase in CO2 concentration with temperature, while [Zanzi et al., 1997] show a decrease. The yields at 800 °C are 135 [g/kg], 248 [g/kgdaf], and 266 [g/kgdaf], respectively. The influence of particle form on the pyrolysis yields as found by [Zanzi et al., 1997] is as follows: • pelletised straw gives almost the same gas and tar yields compared to chopped straw; • the char yield is slightly higher for pelletised straw; • pelletised straw at 800 °C shows lower H2 and CO2 product yields, while the hydrocarbons and CO yields are higher. The slightly higher char yield for pelletised straw could be explained by its slightly higher ash content: 3.9 versus 3.2 [wt%dry] for chopped straw. The ash amount and its composition (not given) is probably also the background for the observed difference in gas yields, as tars can crack into gases via secondary reactions in the char pores (see figure 2.4). The influence of particle size on the pyrolysis process as found by [Rüdiger et al., 1995] with particles with a diameter of 1.5 and 4.0 mm is as follows:

50

• • •

the gas yield at 400 and 1000 °C is about the same, while at 600, 800, and 1000 °C the larger particles show a lower yield; there is no significant influence on tar yield found in the range of particle sizes investigated; the influence on gas yield and on CO and CmHn yields is minimal. The CO product yield is lower at all temperatures except at 400 °C, while the CmHn yield is lower at temperatures below 1000 °C.

One reason for the lower gas yield could be the lower heating rate of bigger particles. Another reason may be that the primary pyrolysis components have a longer residence time in larger particles, which influences primary reactions in the particle and secondary reactions in the gaseous phase. Miscanthus [Rüdiger et al., 1996] and [Storm et al., 1999] used an entrained flow reactor for pyrolysis experiments. The maximum temperature was 1200°C and the residence time was between 2 and 5s. The preheating of the inert gas to the reactor temperature insured high heating rates. The fuel was dried before the pyrolysis to water contents between 2.5 and 10%. The char yield was determined with an ash balance of the raw material and the residual char. Figure 2.21 shows that all product yields decrease with the exception of gas yield.

1000 yields [g/kg daf]

water 800

tar

600 gas

400 200 0 600

char 700

800 900 temperature [°C]

1000

1050

Figure 2.21 Miscanthus: pyrolysis yields vs. reactor temperature (p=1 bar, dp=1.5-6 mm, 8.103

51

light tar Miscanthus heavy tar Miscanthus

100

Figure 2.22

52

100

50

50

0 200

150 yield [g/kg]

yield [g/kg]

150

light tar straw heary tar straw

400

600 800 1000 Peak Temperature [C]

1200

0 200

400

600 800 1000 Peak Temperature [C]

1200

Miscanthus and straw pyrolysis: light and heavy tar yields versus reactor temperature (p = 1 bar, dp = 1.5 mm) [Rüdiger et al., 1996].

2.5.2.1.2 Nitrogen components During pyrolysis of solid fuels, nitrogen bound in the organic matter of the solid fuel is converted into NOx precursors such as NH3, HCN, HNCO as well as nitrogen containing tar and char, see e.g. [Tan and Li, 2000a], [Li et al., 1996], [Nelson et al., 1996], [Pels et al., 1995], [Varey et al., 1996], [Baumann and Möller, 1991], [Cai et al., 1993], [Kambara et al., 1993], [Leppälahti, 1995], [Solomon et al., 1982], [Bassilakis et al., 1993], [Leppälahti and Koljonen, 1995] and [Nelson et al., 1996]. The release of oxidised nitrogen species during pyrolysis does not play an important role [Rüdiger, 1997]. The main nitrogen species during the overall pyrolysis process of the fuel is expected to be HCN when the fuel is present in pyridinic (6-membered ring) or pyrrolic (5-membered ring) aromatic structures, mainly present in older fossil fuels, but also in e.g. chlorophyll in biomass [Van Smeerdijk & Boon, 1987]. Pyrolysis is a complicated combination of heat and mass transfer phenomena and complex chemical reactions (free radical chain reactions, substitution reactions etc.) [Solomon et al., 1992]. The mechanisms and factors influencing nitrogen partitioning during this process have therefore only partly been understood also due to the heterogeneous nature of fuel and complex mixture of pyrolysis products. Influence of type of fuel rank and nitrogen bonding Figure 2.23 shows a variety of typical forms of nitrogen bondings present in the fuel structure [Hämämäläinen, 1994].

Figure 2.23 Functional forms of nitrogen [Hämäläinen, 1994].

53

A large variation in the fuel bound nitrogen partitioning exists, resulting in different HCN / NH3 ratio’s in the gas phase for various fuels during pyrolysis. For (low volatile) coal it is reported that during primary pyrolysis the volatile nitrogen compounds formed are present as aromatically bound nitrogen species, see [Rüdiger, 1997], [Solomon et al., 1992], [Chen&Niksa, 1992a] and [Freihaut et al., 1982]. During secondary pyrolysis these species are converted into gaseous compounds, mainly HCN and to a significantly smaller extent NH3, by ring rupture see [Axworthy et al, 1978], [Haussmann & Kruger, 1990], [Johnsson, 1991], [Johnsson & Jensen, 1994] and [Alzueta et al., 2002]. Typical conversion efficiencies of pyridine, a model aromatic nitrogen compound, into HCN for temperatures in the range 1233 K– 1373 K are approximately 40% - 100%, respectively, and residence times longer than 1s, as determined by [Axworthy et al, 1978] in a quartz capillary flow reactor. [Houser et al., 1980], confirmed these values. [Bruinsma et al., 1988a] showed the onset of pyrolysis of pyridine at circa 1073K and large conversion values at temperatures between 1173 and 1273 K. in a diffusion cell with residence time of 5s. Substantial fractions of the volatile-nitrogen from primary devolatilisation can be incorporated into a carbonaceous soot matrix as has been reported for a range of coal ranks [Chen&Niksa, 1992]. Nitrogen partioning is often explained in terms of differences in the way N is bound in the organic solid fuel substrate. Aromatic structures were found experimentally by X-ray photoelectron spectroscopy (XPS) studies on coal, with most of the fuel-N being in pyrrolic form, see [Perry & Grint, 1983], [Bartle et al., 1987], [Burchill & Welch, 1989], [Wallace et al., 1989], [Nelson et al., 1992], [Kambara et al., 1993], [Aho et al., 1993a], [Buckley, 1994], [Kelemen et al., 1994], [Pels, 1995], [Wójtowicz et al., 1995] and [Thomas, 1997]. Also, by application of XANES, [Mullins et al., 1993] and [Mitra-Kirtley et al., 1993] showed that pyrrolic-N was most abundant in coals. For the lowest rank of coals studied, the pyridine fraction declined to very low values while the pyridone fraction increased considerably. [Burchill, 1987] showed that the relative amount of pyridinic nitrogen augmented and that of pyrrolic nitrogen decreased when the carbon content of the coal increased from 85-90 mass%. Of the total nitrogen in coal, 50-75% is present as pyridine quinoline derivatives [Meyers, 1982]. [Kambara et al., 1993], however, showed different results: pyrrole-N content was the highest with higher ranks and clearly decreased as the carbon content of coal decreased. Also, by these authors a trend was shown that the proportion of both quaterny-N with an unclear nature [Molina et al., 2000] (also indicated to be present by [Nelson et al., 1992], [Kambara et al., 1993], [Buckley, 1994], [Kelemen et al., 1994], [Pels, 1995] and [Kelemen et al., 1998]) and pyridine-N decreased with increasing rank. As a result, in combustion processes low volatile coals are reported to produce higher N2O and lower NO emission levels than high volatile coals, as HCN is seen as the major N2O precursor, see e.g. [Kramlich et al. 1989], [Kilpinen & Hupa, 1991], [Wallman et al., 1993]. [De Soete, 1990] and [Moritomi et al., 1991], though, attribute the N2O emission to char bound nitrogen to a large extent. More NH3 evolves from younger, lower volatile fuels, in which the nitrogen is bound in quaternary nitrogen, amino-acidic, proteinic, nucleic acidic, porphyrin and amine groups, according to several researchers [Aho et al., 1993], [Baumann & Möller, 1991], [Bose et al., 1988], [Chen & Niksa, 1992 b], [Chen et al., 1982], [Hansson et al., 2003], [Hayhurst&Lawrence, 1992], [Hiltunen et al., 1991], [Leckner & Åmand, 1992], [Van Krevelen, 1993], [Leppälahti, 1995], [Nelson et al., 1991], [Niksa&Cho, 1996], [Phong-Anant et al., 1985], [Tian et al., 2002]. Recently, [Saastamoinen & Hämäläinen, 1999] and [Winter et al., 1999] found that pyridone structures in wood are important in the formation of NH3. [Mullins et al., 1993] found that a low pyridine content in low rank (highoxygen) coals correlates with a large pyridone content and they observed aromatic amine fractions in coal of 6-10%, which is comparable to the amount of quaternary nitrogen identified in earlier studies [Van Krevelen, 1993]. [Rüdiger, 1997] carried out entrained flow pyrolysis experiments with a wide range of coals, biomass and sewage sludge. Significantly higher values of fuel bound nitrogen conversion into NH3 and HCN were observed for biomass compared to coals. The release of NH3 was of the same order of magnitude as HCN in their experiments. The authors indicate that NH3 can be converted to HCN by a secondary reaction with carbon (see reaction (R2.43) in paragraph 2.5.3.2), referring to [Tabasaran et al., 1977].

54

[Baumann & Möller, 1991] performed fluidised bed pyrolysis experiments with a large range of coals. They concluded that HCN is the primary pyrolysis product and that NH3 is formed by hydrogenation of HCN. Since the thermal stability of amino groups is lower than that of aromatic nitrogen species, NH3 should be released before HCN. Nevertheless they found that the conversion of NH3, HCN and H2 increase with temperature and that in all cases, the evolution of HCN started at lower temperatures (400-450 °C) than that of NH3. This was consistent with results obtained from pulverised coal combustion by [Ghani & Wendt, 1990]. In most cases, NH3 was formed simultaneously with H2. This was seen as plausible, as HCN is the primary product of cleavage of N-aromatic compounds. [Björkman et al., 1997], however, concluded that formation of NH3 from HCN was not plausible. They performed experiments at 800 and 900 °C, where HCN in a mixture with phenol, glycole and ethanol was introduced over an Fe doped CaO bed and observed practically no NH3 formation. Younger coals were found to form NH3 at considerably lower temperatures than H2. Obviously these high volatile coals can release NH3 also in a direct way by cleavage of amino groups or amides. Also, [Rüdiger, 1997] made plausible that hydrogenation of HCN should not play an important role. He based this on the observed trend of decreasing HCN/NH3 ratio with increasing oxygen content of the flaming pyrolysis environment. In an atmospheric pyrolysis study in a temperature range of 750-900 °C with aniline (a simple aromatic amine), pyridine and propylamine as model compounds in a 1:1 volume ratio with ethanol, different bed materials were applied, quartz and olivine sand, as well as CaO doped with several metallic compounds [Björkman&Larsson, 1996]. The authors found that the chemical nature of the fuel bound nitrogen was probably of minor importance, as aniline and pyridine showed similar behaviour and produced comparable amounts of NH3 and HCN. In contrast to the abovementioned findings of [Björkman&Larsson, 1996], [Hansson et al., 2003] found for fluidised bed pyrolysis of proteins that the protein’s amino acid composition had a marked impact on the composition of the pyrolysate. At 800 °C the ratio of HCN to NH3 during pyrolysis of poly-L-Leucine was approximately 2.2, whereas at the same fluidised bed temperature for poly-Lproline a ratio close to 10 was found. Recently, [Tan & Li, 2000b] and [Li & Tan, 2000] proposed that the formation of NH3 can be attributed to direct hydrogenation of the nitrogen in the pyrolysing fuel/char particles. The active hydrogen (containing) radicals required for this hydrogenation process would be generated from the thermal cracking reactions taking place inside the particles. H radicals hereby seem to be much more active than other species in initiating the opening of hetero-aromatic ring systems, see e.g. [Li et al., 1998]. Thermally unstable N-containing structures are mainly responsible for the formation of HCN, whereas the thermally more stable N-containing structures can be converted more slowly to NH3 according to [Li & Tan, 2000], [Tian et al., 2002]. [Xie et al., 2001] from the same research group also found that in addition to coal rank, the petrographic composition and/or geographic origin of coal are important factors influencing the formation of HCN and NH3 during pyrolysis. Inertite-rich Chinese coals tend to release more NH3 during pyrolysis than Australian coals of comparable carbon contents. The authors believe that the structure of inertites favours the formation of H radicals in the pyrolysing solid in the same temperature range as the activation and subsequent hydrogenation of the Ncontaining ring systems for NH3 formation. The formation of HCN as the precursor to NH3 formation from the char-N is also thought to be unlikely by [Paterson et al., 2002], as the volatiles (which are the source of HCN) will have been destroyed during the formation of the char-N. During fluidised bed pyrolysis of coals the largest portion of nitrogen remains in the char after pyrolysis. With coal ranks ranging from lignite to anthracite, about 60% to over 90% of the nitrogen was retained in the char at temperatures between 970 and 1370 K [Nelson et al., 1992], [Baumann & Möller, 1991], [Pohl & Sarofim, 1976]. This was also found during heated grid pyrolysis experiments [Baumann et al., 1989]. For entrained flow pyrolysis this trend was also observed for black coal. For brown coal lower char-N contents were observed, especially for temperatures higher than 900 °C. [Solomon & Colket, 1978], however, found for a high volatile coal a similar distribution of nitrogen over char and tar. Somewhat higher values of nitrogen release to the gas/volatile phase were reported by [Albrecht, 1992] for coals ranging from brown coal to old black coal. XPS studies of different coals

55

and their chars by [Wójtowicz et al., 1995] showed that at least part of the pyrrolic-N is converted to thermally more stable pyridinic-N during fluidised bed pyrolysis at 1170 K. The initial enrichment of nitrogen in char is larger for low rank coals than for higher rank coals [Baxter et al., 1996] and [Pohl&Sarofim, 1976]. This is attributed to low temperature cross-linking, whereby low rank coals release mainly H2O and CO2, thus non-N containing species. Recent studies show that, in addition to NH3 and HCN, significant amounts of N2 are being formed during pyrolysis of low rank coals and lignite [Stanczyk&Boudou, 1994], [Ohtsuka et al., 1994], [Wu&Ohtsuka, 1996], [Ohtsuka et al., 1997] and [Kidena et al., 2000]. Sewage sludge pyrolysis in a fluidised bed lead to relatively high conversions of fuel bound nitrogen into molecular N2 at ca. 1103K [Augustin, 1986]. For sewage sludge pyrolysis performed in an entrained flow reactor, [Rüdiger, 1997] found that nitrogen was depleted from the remaining char over the whole range of temperatures studied (973-1273 K) with probable N2 formation to a large extent. Also for miscanthus pyrolysis in this reactor the char-N conversion was well below 20%. Influence of bound oxygen The fuel bound oxygen content of the fuel has a significant impact on the conversion of fuel bound nitrogen. At first the fraction of nitrogen released with volatiles increases with oxygen content in the parent fuel [Bassilakis et al., 1993]. Also, a low O/N ratio of the fuel results in a high HCN/NH3 ratio [Aho et al., 1993a and b], confirmed by [Winter et al., 1999] for flaming pyrolysis conditions. Especially phenolic OH-groups in the molecular structure of the fuel matrix are considered to increase the conversion of nitrogen to NH3 by radical OH formation, reacting with HCN inside a fuel particle (see e.g. [Hämäläinen et al., 1994] and [Hämäläinen & Aho, 1995]). [Redlich et al., 1989] found a broad increase of phenolic oxygen content with increasing oxygen content. At least with high heating rate and at high temperatures, phenolic compounds were found to produce OH radicals during pyrolysis [Bruinsma et al., 1988b]. The results were explained by the gas phase reaction sequence [Peck et al., 1991]: +OH +H +H,H2O HCN → HNCO → NH2→ NH3 [Miller & Bowman, 1989], though, found that conversion of HCN to NH3 with OH radicals occurs only at relatively low oxygen concentrations. [Aho, 1998] reported that in this respect the presence of phenolic OH-groups and the O/N characteristics are more important than fuel rank with its accompanying nitrogen functionality. Other O-containing groups did not have a marked effect on the conversions of model N compounds to HCN/NH3. Influence of particle size Variation of particle size (dp in the range 63-250 µm) did not affect fuel-N conversion to N2O and NO under combustion conditions [Aho&Rantanen, 1989], reflecting the aforegoing pyrolytic conversion to HCN and NH3 respectively. Yet larger dp values should favour the reaction between OH radicals and HCN due to pyrolysis products flowing out of the particle surface, reducing O2 penetration into the particle in a flaming pyrolysis zone with oxygen present [Aho et al., 1993a]. [Baumann et al., 1989] in heated grid research showed that larger particles led to decreased nitrogen release due to secondary reactions within the particle. Influence of mineral matter The presence of Ca containing carbonates, like limestone and dolomite influences the nitrogen conversion significantly. [Chambers et al., 1996] experimentally showed that CaO enhanced NH3 decomposition for temperatures higher than 1073 K. Also, CaO enhanced the conversion of NH3 and NO into N2 in an inert atmosphere. NH3 decomposition activity of CaO decreased when exposed to an H2-CO-CO2 atmosphere. This resembles the main pyrolysis/gasification product gas without H2O and N2. In an experimental study by [Abul-Milh and Steenari, 2001] in a fixed bed reactor, NH3 was

56

introduced into a bed of quartz sand and different carbonates. They found that under atmospheric conditions in which the carbonates were calcined (600 °C and higher temperatures), NH3 was decomposed to a large extent while forming HCN, HNCO and N2. Release of HCN and HNCO was also found during comparable experiments performed by [Acke&Lindqvist, 1997]. These authors observed that the presence of CO had a pronounced increasing effect on HCN formation and caused the disappearance of HNCO. The experiments were done in a water free environment. [Abul-Milh and Steenari, 2001] observed that the presence of water vapour inhibits NO formation. This inhibition was suggested to occur because of the adsorption of H2O to the same type of active sites as those active in the NH3 oxidation. Calcium, as well as potassium, can promote formation of NH3 and N2 and suppress HCN/tar bound N formation [Ohtsuka et al., 1997], [Tsubouchi et al., 2001]. At temperatures of 850 °C and higher, N2 formation was enhanced at increasing temperatures for pyrolysis of polyacrylonitril derived char and low rank coals. The formation of Ca nitrides, such as CaCN2 and Ca3N2 is probably acting as intermediates in N2 formation. This, however, could not be confirmed by XRD study. [Björkman&Larsson, 1996] found that Ni catalyst and Fe doped CaO significantly reduced the conversion of model components into NH3 and HCN. Decreasing the contact time increased the amount of NH3, HCN and NO at 900 °C, independently from the O2 concentration, whereas at 800 °C the trend was inverted. An explanation given, was that the bed material both affected the formation and decomposition reactions, but the decomposition dominated at 900 °C. A bed containing quartz and olivine sand produced higher HCN concentrations than a bed containing CaO. Iron influences the removal of char bound nitrogen positively [Ohtsuka et al., 1993] and [Friebel&Köpsel, 1999]. It was indicated to promote N2 formation at the expense of N in light gases, tar and char [Mori et al., 1996] and [Wu & Ohtsuka, 1997]. Precipitation of Fe from an FeCl3 solution on browncoal enhanced the formation of N2 (fuel N conversion to N2 in the range of 50-60%) and decreased markedly the nitrogen retainment by tar and char [Ohtsuka et al., 1994]. The temperature of N2 release was decreased by about 100 °C both during gasification and pyrolysis. This is explained by the fact that at temperatures of approximately 900 °C Fe is available in reduced form and present in ultrafine (20-50 nm) particles. Physically mixing of ultrafine Fe particles with coal was less effective in decreasing N2 formation from fuel bound nitrogen for pyrolysis [Ohtsuka&Furimsky, 1995]. Influence of heating rate In pyrolysis experiments using Illinois No. 6 coal in a heated grid reactor [Cai et al., 1993] found that the conversion to char-N decreased slightly from 60% (at 5 K/s) to 55% (at 5000 K/s) for end temperatures of 1170 K. They also found that increasing heating rate enhanced the conversion of volatilised nitrogen into tar nitrogen. Under rapid pyrolysis conditions, the main nitrogeneous component in the gas fraction is HCN. [Axworthy et al., 1978] measured HCN yields of 18-20% of fuel-N and the sum of NH3/N2 was 8-10% for bituminous coal at a temperature of 1239 K. In high temperature pyrolysis experiments (1270-1670) by [Blair et al., 1976] using bituminous and subbituminous coals, the maximum HCN yield was 15% of the fuel bound N whereas the NH3 yield was only ca. 4%. Results of [Nelson et al., 1992], [Kambara et al., 1993], [Bassilakis et al., 1993], [Wahlers, 1989] and [Hill, 1945] support the theory that high heating rates (and high temperatures) increase the ratio HCN/NH3 in pyrolysis, as stated by [Bassilakis et al., 1993]. [Kidena et al., 2000] compared fast (Curie Point analysis) and slow pyrolysis (Infrared Image Furnace) of seven coals and observed that HCN formation was favoured by increased heating rates. At 1040 °C coal-N conversion to HCN ranged from 11-23% with the highest values for low rank coal. Fluidised bed pyrolysis experiments show significantly higher NH3 yields than HCN for different coals, see [Baumann & Möller, 1991], [Hirama et al., 1983] and [Åmand&Leckner, 1991]. A possible explanation in terms of a homogeneous reduction of HCN with OH and H radicals is suggested by [Miller & Bowman, 1989]. [Lumbreras et al., 2001] showed for slow pyrolysis that HCN yields increased with heating rate and char bound nitrogen decreased at the same time.

57

Influence of pressure At higher pressures, the release of nitrogen to gas phase species during coal pyrolysis was reported to increase, whereas the tar bound nitrogen decreases [Okumura et al., 2002]. The authors attributed this behaviour primarily to large amounts of metaplast remaining in the coal. Specifically, the recombination and peripheral group eliminations in the intermediate fragment group are activated by the pressure increase, consequently resulting in more fuel bound nitrogen gas formation. The fuelnitrogen conversion to NH3 increased and HCN formation showed an opposite trend. It was also suggested by these authors that high pressure prevents release of Fe and Ca compounds, which enhance release of nitrogen gas. Also, the influence of alkyl and H radicals would be larger at high pressure. For Illinois No.6 pyrolysis in a heated grid reactor [Cai et al., 1993] reported that at 700 °C and atmospheric pressure, about 70% of the nitrogen is released as tar bound nitrogen, whereas at 7.0 MPa the conversion falls to about 50%. [Solomon&Colket, 1978] studied rapid pyrolysis in vacuum using a heated grid reactor. They found that less than 10 % of the total amount of gaseous nitrogen was released at temperatures below 970 K. It was concluded that most of the volatile nitrogen is in the tar fraction under these conditions. At higher temperatures, the conversion of fuel bound nitrogen into gaseous species is increased. Influence of temperature Increase in temperature increases the pyrolysis rate and consequently the flow through the particle surface. In a flaming oxygen-containing pyrolysis zone, the O2 content near the particle is then reduced, which could favour conversion of HCN to NH3 [Aho et al., 1993a]. For fluidised bed pyrolysis of brown, sub-bituminous and bituminous coals it was reported that conversion of fuel bound nitrogen into NH3 ceases to increase with temperature at approximately 800 °C, whereas the conversion into HCN continued to increase above that temperature, see [Nelson et al., 1992]. For German coals, nitrogen evolution was usually less than that of higher rank coals at temperatures between 970 and 1270 K, except for char from anthracite whose N/C ratio was almost the same as that of the parent coal [Baumann & Möller, 1991]. The preferential retention of nitrogen in char at lower temperatures is greater for low rank coals and biomass than for coals of higher rank [Pohl & Sarofim, 1976] and [Baxter et al., 1996]. This preferential retention in char at low temperatures was also indicated by [Blair et al., 1976], [Solomon & Colket, 1978], [Phong-Anant et al., 1985], [Bruinsma et al., 1988b], [Haussmann & Kruger, 1990], [Kambara et al., 1993], [Man et al., 1993], and [Takagi et al., 1999]. At temperatures above 1270 K and longer residence times, a significant reduction of char-N was observed and the normalised N/C ratio was lower than one, which is consistent with results obtained by [Slaughter et al., 1988]. [Blair et al., 1976] found for three different coals a nitrogen release of 2040% at 1173 K, whereas at 1973 K about 80% of the nitrogen was released as volatile. The shift from enrichment to depletion of nitrogen during pyrolysis seems to occur at lower temperatures for biomass than for most coals [Glarborg et al., 2001]. Formation of HCN and NH3 during coal pyrolysis has been shown to follow the production of tar and it is likely that a significant fraction of these light N-containing gases comes from cracking reactions of tar [Nelson et al., 1992]. Unfortunately, the knowledge about kinetics and product distribution of the thermal decomposition of tars is limited. Thermal cracking of tar derived from coal at temperatures below ca. 1100 K yields relatively small amounts of gas phase nitrogen compounds but these increase with higher temperatures, due to an increase of HCN yields. NH3 and HNCO yields increase till about 1123 K and decrease at higher temperatures [Li et al., 1996], [Ledesma et al., 1998]. The thermal stability of nitrogen species in tar appears to obey the following order: pyrrolic
58

Results of rice husk pyrolysis in a fixed bed reactor set up with heating rates between 5-15 K/min (slow heating) indicate that the increasing end temperature of the process increases HCN yields and lowers the final char bound nitrogen yields [Lumbreras et al., 2001]. An increase of the HCN/NH3 ratio with increasing temperature during fluidised bed pyrolysis was reported by [Hansson et al., 2003] for proteins. For poly-L-Leucine they reported an increase of this ratio from 1.9 at 700°C to 2.2 at 800 °C. In this study HNCO was detected as released species, but it could not be quantified. An analogue study was made for bark, whey, soya beans and shea [Hansson et al., 2004]. Here, the authors found an increase of the molar HCN/NH3 from about 0.3 to 1.5 in the temperature range of 700-1000 °C. On the other hand, [Zhou, 1998] reported that for pyrolysis of a diversity of biomass feedstock in a bubbling AFB, NH3 was by far the major nitrogen containing species. Over a temperature range between 750 and 950 °C the NH3 yield was observed to decrease by a factor 5, whereas the much lower yields of NO and HCN were practically constant with temperature. Influence of process environment Decreasing oxygen concentrations under flaming pyrolysis conditions were found to lead to increased HCN/NH3 ratios, see [Björkman&Larsson, 1996], [Baumann&Möller, 1991]. [Rüdiger, 1997] also observed increasing HCN/NH3 ratios with lower air stoichiometric values for Göttelborn black coal pyrolysis. Trends of fuel-N conversion to HCN, NH3, NO and N2O by [Winter et al., 1999] showed a more diverse picture for several young fuels. They found increasing NH3 yields with decreasing oxygen content for all these fuels, but HCN yields increased only for peat and malt waste. For spruce wood they observed decreasing HCN yields at lower oxygen contents, whereas for alder wood the HCN yield was practically zero. With increasing oxygen content in the pyrolysis atmosphere and slow heating rates, maximum nitrogen release from coal fuel was shifted to lower temperatures, which was attributed by [Klein, 1973] to easier breakup of the single bond in C=N-C molecular groups under oxygen influence. A higher oxygen concentration in the pyrolysis environment leads to increased N/C ratios in tar released, according to [Baumann&Möller, 1991]. This is caused by increased carbon oxidation compared to nitrogen release rates. In the presence of CO2, the formation of HCN, and to a lesser extent also NH3, was suppressed [Ohtsuka&Furimsky, 1995]. These results were confirmed by [Chang et al., 2003] performing pyrolysis of a range of coals in a combined drop tube/fixed bed reactor, with Ar and CO2 as gas flow media. The authors explain the decreased formation of bound nitrogen species (HCN and NH3) accompanied with CO2 addition with the hypothesis that CO2 is chemisorped at N-sites. It blocks these sites for reaction with H radicals necessary for HCN/NH3 formation. Also, they indicate that CO2 can be adsorbed on other (H-rich) sites. Here it consumes freshly generated H radicals. In the absence of H radicals, the N-sites could be converted into mobile C(N) species [Jones et al., 1995] which during gasification would lead to N2 formation through recombination. NO from the oxidation of the N-site by CO2 may also react with active C(N) char species to form N2, particularly at high temperatures, as suggested by [Ohtsuka & Wu, 1999]. The effect of diminishing bound nitrogen species by CO2 is especially the case for thermal cracking of chars that generate relatively large amounts of H radicals. When this is not the case, CO2 can have a slightly increasing effect on fuel nitrogen release in the form of NH3 and HCN. Steam has a positive effect on the formation of NH3 (and to a much smaller extent to the formation of HCN) from different coals, as studied by [Chang et al., 2003]. These authors indicate that introduction of H2O increases the H radical formation potential in the system, which is in their hypothesis the precursor for NH3/HCN formation from nascent char N-sites.

59

2.5.2.2 Modelling approaches In order to model the flash pyrolysis sub-process of the gasification process, two theoretical approaches can be adopted. The data required for the description of pyrolysis must be carefully selected because they should be both easily computable and readily obtainable, while maintaining the ability to account for solid fuel heterogeneity. The first approach involves the setting up of a detailed model that accounts for the decomposition of the solid fuel matrix during particle heat up. This approach requires detailed analytical data to be retrieved from the specific fuel to be modelled. This includes the initial fuel matrix structure and the release of both gases and tars (condensable aromatic hydrocarbon species) during pyrolysis [Williams et al., 2000]. This approach has been applied to coal pyrolysis by several researchers, leading to models called FLASHCHAIN (see e.g. [Niksa, 1996]), CPD [Smith et al., 1994] and FG-DVC [Solomon et al., 1993]. The second approach involves modeling the devolatilisation process using generalised expressions based on a limited set of chemical reactions. Commonly a two-step mechanism is used with cellulose as biomass model component. In the model of [Kilzer & Broido, 1995] cellulose decomposes by means of two competitive reactions. One of these reactions results in the formation of dehydrocellulose, which in a consecutive reaction is converted to char and gases; the other reaction results in char formation. Another approach is another two-step model given by [Antal & Varhegyi, 1995]. According to this model, the primary products of pyrolysis are volatiles. If these are not removed from the system, the secondary vapour-solid reaction will lead to char formation. [Bradbury et al., 1979] presented a three-step model scheme in which biomass is first converted into an active intermediate species which is converted by two parallel reactions into char and gas on the one hand and volatiles on the other hand. These volatiles can be further converted into gas in a consecutive step. This two or three step reaction approach is more commonly adopted within CFD codes, which would become too slow if detailed solid fuel matrix pyrolysis programs are utilised in the main body of the computation. Ideally, by inputting solid fuel properties it would be useful to predict, by means of a pre-processing computer subroutine, the solid fuel pyrolysis behaviour. This is the first mentioned approach, which will be followed in this work. Recently, the FG-DVC model has been adapted to describe biomass flash pyrolysis [Chen et al., 1997]. The biomass fuels characterised was cellulose, poplar wood and corn stalks. The first versions of the FG-DVC model, focussing on coal/lignite pyrolysis from the US Argonne Premium Coal Samples Bank, have been developed in the eighties by the American research company Advanced Fuel Research, see e.g. [Solomon et al., 1988], [Solomon et al., 1990], [Solomon et al., 1993]. The formation of gases, tars and carbonaceous solids during the flash pyrolysis or devolatilisation process is described by a Functional Group (FG) model for gas evolution, combined with a statistical depolymerisation, vaporisation and cross-linking (DVC) model for tar and char formation. The functional group part of the program is based on the premise that a fraction of the total gases evolved is produced when specific functional groups break away from the macromolecular network. Other functional groups stay attached to the main solid fuel structure, eventually being released as part of a light tar molecule. The fraction of the functional groups that are released as gases is determined by the heating rate, the final temperature, the number of specific functional groups, and the structural properties of the original solid fuel. Small groups will produce gases; larger molecular fragments (or groups) will give tars. The second (DVC) part of the program models the thermal breakdown of remaining macromolecular networks. It starts with a two-dimensional simplification of this network, including all the major structural attributes; molecular weights of the aromatic ring clusters, cross-link density, and the potential number of labile bridges. The decomposition of this network is modelled with a Monte Carlo simulation. The smaller fragments are converted into tars and gases. The remaining macromolecule forms the char. The pyrolysis of biomass and its main constituents has been the subject of numerous studies and is summarised in several reviews ([Antal, 1985], [Bridgwater & Cottam, 1992], [Antal & Varhegyi, 1995]). One of the main differences between the pyrolysis of coal and biomass is the fact that coal consists predominantly of aromatic material, whereas the aromatic part of biomass (lignin) is relatively 60

small. Biomass generally contains much more oxygen than coal does. The oxygen is present in ether, hydroxyl, carboxyl, aldehyde and ketone functionalities, which decompose during pyrolysis and form oxygenated gases (e.g. CO, CO2 and H2O). The yields of these compounds are similar to those found during pyrolysis of low-rank coals (5-10 dry mass% for CO2 and H2O and 5-15 dry mass% for CO). Biomass pyrolysis, however, gives much higher tar (liquid) yields compared to low-rank coals (40-50 mass% versus 10-20 mass% on a dry basis). The increased tar yield evolves primarily at the expense of char, the yield of which is much lower for biomass than for low-rank coals (<10 mass% versus 4050 mass%). Biomass depolymerisation is the predominant pyrolytic reaction [Shafizadeh, 1984], whereas for coal, depolymerisation reactions compete with cross-linking, which enhances char formation [Solomon et al., 1993]. Most of the char formed from biomass is derived from the lignin component, which is close to low-rank coal in its chemical composition. The yield and distribution of products from pyrolysis depend on other variables in addition to the final temperature and holding time. These include heating rate, total pressure, ambient gas composition and the presence of mineral catalysts [Shafizadeh, 1984]. One approach to the quantitative modelling of biomass pyrolysis is based on the approximation that the three main components of biomass (cellulose, hemi cellulose and lignin) behave independently during pyrolysis (see e.g. Nunn et al., in [Overend et al., 1985]). Consequently, yields can be predicted base on knowledge of the pure component behaviour. The shortcoming of this technique is that it does not account for possible interactions between the biomass components. The FG-DVC model approach was successfully used by [Serio et al., 1994] for lignin pyrolysis. For reasons discussed earlier, the cross-linked aromatic network features of the DVC model were deemphasized for biomass in favour of the original functional group (FG) description of gas and tar evolution [Solomon et al., 1985]. The FG model permits detailed prediction of the composition of volatile species and char. It yields gas production, tar yield and tar functional group composition, as well as char elemental and functional group composition. The model is described by solid fuel rank dependent rates for the decomposition of individual assumed functional groups in the solid fuel and char to produce gas species. The parent solid fuel’s functional group composition determines the ultimate yields of each gas species. Tar development is a process occurring in parallel competing for all the functional groups in the parent solid fuel. The basic solid transformation processes are described as follows by the FG-DVC model: GAS FORMATION The formation of each gas species from a related specific functional group considered is assumed to be a first order reaction process:

d Wi ( gas) dt

= k i Wi (char ) = k i XYi

(2.1)

with d Wi/dt the rate of evolution of species i into the gas phase, ki is a distributed rate constant (not one single exact value) for species i and Wi(char) is the functional group source remaining in the char. X is the mass fraction of the remaining char and Yi the specific functional group mass fraction. The rate constant ki is given by an Arrhenius expression:

⎡ − (E i ± σ i ) ⎤ k i = k 0 exp ⎢ ⎥ i RT ⎢⎣ ⎥⎦

(2.2)

With σi characteristic for a Gaussian distribution of the activation energies Ei.

⎡ Wi (E i ) = ⎢ ⎢⎣

⎡ − ⎡E − E0 ⎤ ⎤ i i ⎦ ⎥ exp ⎢ ⎣ 2 [σi 2π] ⎥⎦ ⎣⎢ 2σi Wi0

2

⎤ ⎥ ⎥ ⎦

(2.3)

61

Where E0i is the average activation energy and σi is the width of the Gaussian distribution. TAR FORMATION

The tar evolution kinetics is described by summing the functional group contributions evolved with the tar. The rate of tar species evolution is described by: d Wi (tar ) ⎡d X ⎤ = −⎢ ⎥ Yi dt ⎣ dt ⎦

(2.4)

where dWi (tar)/dt is the rate of evolution of each functional group component with the tar. CHAR FORMATION

The change in the i-th char pool, Wi (char), is calculated by summing the losses to the gas and tar: d W (char ) d W ( gas ) d W (tar ) i i i =− − (2.5) dt dt dt In order to determine the FG model parameters, like the functional group compositions and the kinetic species evolution rates, Thermo gravimetric Analysis combined with Fourier Transform Infrared Spectroscopy is used (TG-FTIR). The model is fitted to the TG-FTIR data at three heating rates (3, 30 and 100 K min-1). When there are multiple sources for a given species and the sources have overlapping peaks, the determination of parameters is not unique and some assumptions must be made. Based on chemical arguments, k0i is restricted between 1012 and 1015 s-1. Also the preexponential factor for a given species pool is assumed to be rank-invariant. This assumption is based on the observed rank variation of the evolution curves. With increasing coal rank, the leasing edges and the early peaks (extra-loose or loose pools) shift to higher temperatures, while the trailing edges (tight or extra-tight pools) remain at the same temperature. For biomass pyrolysis, a smaller number of functional groups is sufficient to describe the process, as the evolution of each species can usually be represented by the decomposition of a single functional group. In other words, each gas species released during biomass pyrolysis in TG-FTIR analysis evolves in the form of a single peak. This is in contrast to coal or lignin pyrolysis where several peaks can be observed for an individual species. Methane is an exception, in the sense that it requires two functional groups in the case of coals and woody biomass but only a single peak for herbaceous biomass. 2.5.3

Heterogeneous char-gas reactions

Heterogeneous reactions in coal/biomass gasification include the reactions of char, the solid residue of pyrolysis, with gas phase species (mainly O2, H2O, CO2 and H2). In this paragraph the main reactions are presented and kinetic modelling approaches are described. 2.5.3.1 Main carbon based reactions The main heterogeneous reactions playing a role in gasification can be represented as: C (s) C (s) C (s) C (s)

+ (1/φ) O2 (g) + H2O (g) + CO2 (g) + 2 H2 (g)

(2 φ−2)/φ CO (g) + (2-φ)/φ CO2 CO (g) + H2 2 CO (g) CH4 (g)

(R2.1) (R2.2) (R2.3) (R2.4)

The value of φ in reaction (2.1) is subject to some controversy in the literature. Some researchers suggest that this value cannot be found for the complex interacting phenomena taking place during gasification [Denn et al., 1979]. According to [Smith, 1982], CO is the primary product (φ = 2) for coal. [Laurendeau, 1978] and [Martens, 1984] concluded that both CO and CO2 are primary reaction

62

products of this combustion reaction. High temperature as well as low pressure favours CO. [Jensen et al., 1995] and [Johnsson & Jensen, 1994] show that CO is the main primary product at PFBC conditions, but relative low levels of (10-30%) of primary CO2 formation are possible. [Arthur, 1951] found for graphite and coal char, under atmospheric conditions and suppressing further oxidation of gaseous CO using POCl3, the following relation for the mole ratio CO/CO2: CO/CO2 = 103.4.exp(-12400 / (R T)) At temperatures typical for fluidised bed operation, 700-1000 °C, 4.1< CO/CO2 <18.8 so CO will be the major component of initial carbon oxidation. [Monson et al., 1995] gives a relation for the molar ratio CO/CO2 under pressurised bituminous coal char combustion (verified in a drop tube reactor with a temperature range of 1000-1500K and O2 concentrations in the range of 5-21%): CO/CO2 = 3.108.exp(-30178 / Ts) At temperatures typical for fluidised bed operation, 700-1000 °C, 1.02.10-5< CO/CO2 <1.52.10-2 so the coefficients in (R2.1) for CO and CO2 are practically 0 and 1, respectively. In table 2.17 an overview of literature data for the rate of char gasification reactions with CO2 and H2O is given. Data for (hard) coal are not shown here, as this fuel is not discussed in this study. The gasification reaction of char with H2 is much slower than the char-H2O/CO2 reaction [Kosky & Floess, 1980] and is not further dealt with in this section. Char gasification reaction with CO2 can be represented by the following mechanism, as pointed out by [Barrio et al., 2001]: k1f Cf (s) + CO2 C(O) + CO (2.5) k1b C(O) → CO + Cf (s) (2.6) k3 In these expressions, Cf represents an available active carbon site and C(O) an occupied site, or carbon-oxygen complex/transitional surface oxide. CO can have an inhibiting effect on the reaction rate, which consists of lowering the steady state concentration of C(O) complexes by increased backwards reaction 1b. The rate expression is of the Langmuir-Hinshelwood type: k 1f P CO 2 (2.7) R = k 1b k 1f 1+ + P P k CO 2 k CO 3

3

Often a further simplification to nth order kinetics is applied:

R = k Pn CO 2

(2.8)

The char gasification reaction with H2O is more complex than the char-CO2 reaction, because more molecules are involved. [Hüttinger & Merdes, 1992] presented a comprehensive description of the models proposed in the literature for this reaction. Basically, there are two models of the reaction mechanism: the oxygen exchange model and the hydrogen inhibition model, as summarized by [Barrio et al., 2001]: Cf (s) + H2O C(O) Cf (s) + H2 Cf (s) + ½ H2

k1f k1b → k3 k4f k4b k5f

C(O) + H2

(2.9)

CO + Cf (s)

(2.10)

C(H)2

(2.11)

C(H)

(2.12)

k5b

63

The oxygen exchange model is based on the first two equations. The traditional hydrogen inhibition model is an extension of the oxygen exchange model (reactions (2.9) and (2.10), with (2.9) being irreversible) with reaction (2.11). A second version of the hydrogen inhibition model consists of the first two and the last reaction equation. The rate expression relating to this mechanism is again of Langmuir-Hinshelwood type:

R =

k 1f P H2O k3

with:

(2.13)

( )

1+ k 1f P H2O+ f P H2

f (PH2) = (k1b/k3) * PH2 [Oxygen exchange model] f (PH2) = (k4f/k4b) * PH2 [Hydrogen inhibition model, traditional] f (PH2) = (k1b/k3) * √PH2 [Hydrogen inhibition model, second version]

Often a further simplification to nth order kinetics is applied:

R = k Pn H2O

(2.14)

Table 2.17 Overview of char gasification reaction rate data. Reference

Fuel

Medium

CC [%]

[Roll,1994]

Misc. char

CO2

30

[Roll,1994]

Misc. char

H2O

30

CO2

30

[Gudenau & Hahn., 1993] [van den Aarsen, 1985]

[Kojima, 1993] [Rensfelt, 1978] [Rensfelt, 1978] [Rensfelt, 1978]

[Nandi, 1985] [Groeneveld, 1980]

[Gudenau & Hahn, 1993]

Misc. char Beech wood char Saw dust Poplar char

Straw Bark Pine char Wood char Brown Coal char

CO2

H2O

H2O

H2O H2O H2O

n.i.

45

45 45 40

CO2

[Hansen et al., 1997]

H2O,H2

[Barrio et al., 2001]

Birch wood char

H2O,H2

[Barrio et al., 2001]

Beech wood char

H2O,H2

[Barrio and Hustad, 2001]

Birch wood char

CO2

R = ki cin mchar [mol/s]

R = ki cin mchar [mol/s]

R = ki cin mchar [mol/s]

R = ki ci

n

2 [mol/(m ..s]

n

R = ki Pi mchar

R=

[kg/s] ki Pin

mchar

[kg/s]

R = ki Pin mchar [kg/s]

R = ki Pin mchar [kg/s]

?? R = ki cin

H2O, CO2

Wheat Straw char

64

n.i.

Reaction rate equation*

2 [mol/(m ..s]

30

n.i.

R = ki cin mchar [mol/s]

R = mchar k1 PH2O / (1+k2 PH2+ k3 PH2O) [kg/s]

R = mchar k1 PH2O / (1+k2 PH2+ k3 PH2O) [kg/s]

R = mchar k1 PH2O / (1+k2 PH2+ k3 PH2O) [kg/s]

R = mchar k1 PCO2 / (1+k2 PCO2+ k3 PCO) [kg/s]

n

Ea [kJ/mol]

ki,0

dp [mm]

T [°C]

P [MPa]

1

365

3.6.1014 [m3/(kg.s)]

n.i.

750890

0.1

1

266

5.3.1010 [m3/(kg.s)]

n.i.

700830

0.1

1

160.3

6.40.104 [m3/(kg.s)]

n.i.

800900

0.1

0.83

166

7.2 [(mol/m3)1-n(m/s)]

1-2

720885

0.1

0.41

179

1773 [s-1]

1-2

182

6

850950 650800 650800 650800 760815

0.073 (PH2O) 0.073 (PH2O) 0.073 (PH2O)

1

1 1

182 178

-1

2. 10 [s ]

<1.5

5

-1

<1.5

6

-1

<1.5

6

-1

9.83. 10 [s ] 1.52. 10 [s ]

0.1

1

164

2.14.10 [s ]

1.5

0.7

217

106-107 [(mol/m3)1-n(m/s)]

?

?

0.1

1

455.5

1.57.1019 [m3/(kg.s)]

n.i.

800900

0.1

-

Ea,1 = 149 Ea,2 = -117.9 Ea,3 = -108

<0.15

750925

-

Ea,1 = 214 Ea,2 = 11 Ea,3 = -59

0.0450.060

750950

0.1

-

Ea,1 = 199 Ea,2 = -79 Ea,3 = -26

0.0450.060

750950

0.1

0.0320.045

750950

0.1

Ea,1 = 165 Ea,2 = -71 Ea,3 = -215.2

k1,0 = 5.13.104 [1/(s.bar)] k2,0 = 8.04.10-5 [1/bar] k3,0 = 9.10.10-6 [1/bar] k1,0 = 7.6.107 [1/(s.bar)] k2,0 = 1.31.102 [1/bar] k3,0 = 4.75.10-3 [1/bar] k1,0 = 2.0.107 [1/(s.bar)] k2,0 = 2.14.10-2 [1/bar] k3,0 = 2.38.10-1 [1/bar] k1,0 = 1.3.105 [1/(s.bar)] k2,0 = 4.02.10-3 [1/bar] k3,0 = 1.11.10-8 [1/bar]

0.17

1

2.5.3.2 Heterogeneous and heterogeneously catalysed homogeneous nitrogen reactions. Of less importance compared to R2.1-R2.4 under gasification conditions, but strongly related to the study of bound nitrogen behaviour, is the heterogeneous reaction between char and NO (R2.15 + 16). C (s) + NO (g) C (s) + 2 NO (g)

½ N2 (g) + CO (g) N2 (g) + CO2 (g)

(R2.15) (R2.16)

Compared to R2.1-R2.4 this reaction has been much less studied. Studies to determine reaction kinetics of this reaction were reviewed by [Aarna & Suuberg, 1997 and 1999]. These reactions can play a role in combination with the following heterogeneously catalysed gas phase reaction: CO (g) + NO (g)

½ N2 (g) + CO2 (g)

(R2.17)

[Aarna & Suuberg, 1999] reported enhancement of NO reduction reactions by CO for various chars (coal, phenolic resin-derived, graphite). This was attributed to reaction R2.17. Catalytic surfaces for this reaction can be: • quartz ([Wittler et al., 1988]; [Berger&Rotzoll, 1995]) • impure sands ([Schoderböck et al., 1996]) • calcined limestone and dolomite ([Tsujimura et al., 1983];[Olanders&Strömberg, 1995]; [Shimizu et al., 1992]; [Hansen et al., 1992]) • sulphided limestone under fuel rich conditions ([Furusawa et al., 1985]) • coal ash and CFBC bed ash ([Johnsson, 1994]) Additionally, [Wójtowicz et al., 1993] and [De Soete, 1990] indicate the role of the following reactions in heterogeneous combustion systems (with in brackets: solid bound atoms): (2C) + NO (g) (CN) + NO (g)

(CN) + (CO) N2O (g) + (C)

The following heterogeneous reaction with N2O [Jensen, 1996]: C (s) + N2O (g) N2 (g) + N2 (g) + C (s) + N2O (g) N2 (g) + (CO) + N2O (g)

(R2.18) (R2.19)

is presented by [Leppälahti & Koljonen, 1995] and CO (g) (CO) CO2 (g) + C (s)

(R2.20) (R2.21) (R2.22)

Direct oxidation of char bound nitrogen is possible via: → 2 NO (g) O2 + 2 char-N → 2 N2O (g) O2 + 4 char-N O2 + C + (CN) → (CO) + (CNO)

(R2.23) (R2.24) (R2.25)

NO formation: (CNO) (CN) + (CO)

(R2.26) (R2.27)

N2O formation: (CNO) + (CN) 2(CNO)

→ →

→

NO (g) + C NO (g) + 2 C

→ N2O (g) + 2 C (CO) + C + N2O

(R2.28) (R2.29)

Also, hydrogen analogue to CO can reduce NO with char acting as catalyst ([Furusawa et al., 1982];[Chan et al., 1983]): ½ N2 (g) + H2O (g) (R2.30) H2 (g) + NO (g) In gasifiers, though, the oxygen concentrations are relatively low, as compared to CO2 and H2O for which components the following reactions were given by [Leppälahti&Koljonen, 1995]:

65

(CN) + CO2(g) (CN) + H2O(g)

→ →

(CNO) + CO(g) (CNO) + H2(g)

(R2.31) (R2.32)

Also, hydrogen is present in significant amounts in gasification reactors, for which the following reactions were given: → (CH) + (CHN) (R2.33) H2 + C + (CN) (CH) + (CN) → HCN (g) + C (R2.34) → HCN (g) + (*) (R2.35) (CHN) → NH3 (g) + C (R2.36) H2 + (CHN) H2 + (CO) → H2O(g) + C (R2.37) Ammonia, either directly formed during pyrolysis, or by reaction of hydrogen with char bound nitrogen can further react. Catalytic ammonia decomposition can take place according to (R2.38): N2 (g) + 3 H2 (g) 2 NH3 (g)

(R2.38)

Heterogeneously catalysed homogeneous reactions of ammonia are: 4 NO (g) + 6 H2O (g) 4 NH3 (g) + 5 O2 (R2.39) [Duo, 1990], [Jensen et al., 1995], [Johnsson&Dam-Johansen, 1991] and [Minchener & Kelsall, 1990]. 4 NH3 (g) + 3 O2 2 N2 (g) + 6 H2O (g) (R2.40) [Jensen et al., 1995], [Johnsson & Dam-Johansen, 1991] Under conditions where both NO and NH3 are present, catalytic reactions can take place according to (R2.41), see e.g. [Jensen et al., 1995], [Johnsson & Dam-Johansen, 1991] and [Minchener & Kelsall, 1990]. 4 NH3 (g) + 6 NO

5 N2 (g) + 6 H2O (g)

or, when oxygen is present: 4 N2 (g) + 6 H2O (g) 4 NH3 (g) + 4 NO + O2 [Duo, 1990], [Jensen et al., 1995] and [Minchener&Kelsall, 1990].

(R2.41) (R2.42)

Reactions (R2.39) - (R2.41) can be catalysed by char, CaO and bed material [Jensen et al., 1995]. Also a heterogeneous reaction between carbon and ammonia is indicated by [Tabasaran et al., 1977]: HCN (g) + H2 (g) C + NH3 (R2.43) 2.4.4

Homogeneous gas phase reaction mechanisms, including nitrogen chemistry

Overall conversion rates of gaseous compounds are determined by a combination of the rates of mixing of the reacting gases and of chemical reaction kinetics. In this section only chemical reaction rates are analysed. Two main approaches are typically applied for describing reaction rates of volatile constituents [Baxter et al., 1996]: • global reaction mechanisms • elementary reaction mechanisms Global reaction chemistry mechanisms do not require detailed knowledge about the actual reactions taking place and have the advantage of being mathematically simple to use. However, they do not represent the actual reactions taking place, as elementary reactions do. Intermediate species that

66

couple (seemingly unrelated) reactions are therefore not accounted for by global reaction mechanisms. These reactions can be important for the formation and reduction of NOx precursors in the gas phase but their importance for the overall combustion rate might only be minor. Elementary reaction mechanism descriptions encounter two main problems. Firstly a detailed understanding and description of reactions is only available for simple C, H, O and N containing species incorporating less than about 3 carbon atoms per molecule. This means that higher hydrocarbons and tars, which typically are among the main primary pyrolysis products from the solid fuels, cannot be directly accounted for. As a result, an empirical approach for the understanding of the formation and destruction of these species is necessary. Secondly, homogeneous gas phase reactions may be catalysed heterogeneously, e.g. by char or additives (CaO). CaO catalyses CO oxidation for example, see [Dam-Johansen et al., 1993]. For the description of combustion/gasification processes, several elementary reaction schemes have been developed. Well known is the GRI-mech. 3.0, the most recent and comprehensive collection of mechanisms developed at the Gas Research Institute. It is publicly available on the internet [Smith et al., 2000], including a thermodynamic data list. The mechanism consists of 325 elementary reactions including 53 species and developed for natural gas ignition and flame propagation phenomena, including NO formation and reduction. The current mechanism version includes propane and C2 oxidation products, as well as new paths for formation of formaldehyde, NO and reburn purpose. It does not include chemistry related to selective non-catalytic reduction of NO. [Coda Zabetta & Kilpinen, 2001] discourage the use of this model for fuels different from natural gas and methane, including the byproducts originating from combustion of natural gas, i.e. methanol, propane, ethylene and acetylene, as different radical intermediates and their reactions play an important role. GRI-mech. 3.0 has been optimised for premixed combustion systems with temperatures in the range of 1000-2500 K, pressures between 10 Torr and 10 atm and equivalence ratio’s varying between 0.1 and 5. [Dagaut et al., 2000] have developed a chemical mechanism for modelling reburn by natural gas blends to reduce NOx. A dedicated thermodynamic data library for 112 species accompanies this model. The mechanism includes 871 elementary reactions, including reactions involved in the selective oxidation of ammonia. This mechanism, abbreviated as ‘Dagaut-00’ has been validated against atmospheric experiments, like gas mixtures containing ca. 3000 ppmv propene, 750-1000 ppmv of NO, oxidised at fuel/air equivalence ratio’s in the range of 0.5-2 and temperatures between 1100 and 1450 K. [Glarborg et al., 1998] presented a detailed chemical rate scheme, also with the purpose to describe reburn processes for NOx reduction. This scheme is composed of 438 reactions between 62 molecular and radical species, including nitrogen compounds. At Åbo Akademi University much research work has been performed in the field of combustion and gasification processes involving biomass, coals and waste. In this work [Kilpinen, 1992] has contributed to the development of elementary mechanisms. The first generation of these models is called “Kilpinen-92” [Kilpinen et al., 1999] and it is based on the work of Glarborg and co-workers ([Glarborg et al., 1986;1993;1994;1995a and b] and [Glarborg & Hadvig, 1991] and [Miller & Bowman, 1989]. The model Kilpinen-92 involves 253 reactions between 49 species. The first 234 reactions were taken from [Glarborg & Hadvig, 1991], work that is based on [Glarborg et al., 1986] and [Miller & Bowman, 1989] but with new kinetic data, especially for the submechanism of CH4 oxidation and hydrocarbon-nitrogen interactions. The additional reaction kinetic data were directly taken from [Miller & Bowman, 1989]. It includes the oxidation sub-mechanisms for C1-C2 hydrocarbons, HCN, NH3, and the sub-mechanisms describing the interactions between hydrocarbon (CHi, HCCO) and nitrogen (NO, NHi, N2) compounds. This mechanism was adopted in reduced form (109 reactions) by [Zhou, 1998] to simulate the fate of fuel nitrogen in an atmospheric BFB. A further developed model based on Kilpinen-92 is the Kilpinen-97 mechanism. This mechanism comprises 353 reactions between 57 species. This upgraded mechanism involves updates of many kinetic rate constants, as well as the introduction of new reactions involving the species N2H3, N2H4, HONO, NO3, H2NO, NCN, C2N2 and HNNO [Coda Zabetta et al., 2000a,2001]. Kilpinen-97 was reported to allow for simulations of processes at high pressure, for which a number of modifications was made, as discussed by [Kilpinen et al., 1997]. The mechanism has been validated against

67

experimental data from [Rota et al., 1998; 1997] and shown to describe nitrogen and hydrocarbon combustion kinetics satisfactorily at low pressure in a wide range of temperatures and air to fuel ratios. A limited number of comparisons [Kilpinen & Hupa, 1998] and [Leppälahti et al., 1998] made at high pressure (upto 20 bar) and in the temperature range 700-1000 °C, however, suggest that the model may also be applicable for predictions at high pressure. Figure 2.24 shows the nitrogen species involved in the mechanism as well as the reaction pathways between the compounds. The direct route from N2 to NO is not favoured under typical fluidised bed gasification conditions. The same holds for the Fenimore route where CHi radicals react with N2 at high temperatures to HCN (prompt NOx), as temperatures are too low. Also, N2O formation is not favoured under fluidised bed gasification conditions. The reactions, however, are included in the mechanism.

Figure 2.24

2.6

Simplified presentation of the kinetic scheme of the major gas-phase reactions for nitrogen species [Coda Zabetta et al., 2000a].

Potential primary measures for fuel_NOx emission reduction

The addition of oxidizing agents, like oxygen, air, NO, NO2 and partly recycled flue gas from gas turbine combustion to gasification product gas upstream of a gas cleaning unit in potential can potentially reduce NOx precursor concentrations, especially NH3. This can be done also in combination with the in-bed addition of e.g. carbonate rocks (like dolomite or limestone). In combustion installations, injection of NH3 into high temperature flue gases is applied to reduce NO to N2 (the so-called SNCR process, see e.g. [Lyon, 1975]; [Lyon&Benn, 1978]). Abovementioned addition of oxidizing agents would reflect the reverse process. Optimum temperatures for SNCR are reported to be in a range from 830 °C where significant NO reduction begins, gradually increasing to 960 °C [Duo, 1990]. Other investigations on SNCR in combustion processes showed an optimum temperature of circa 950 °C ([Banna&Branch, 1981]; [Seidle&Branch, 1983]), a range of 880 – 980 °C [Lucas&Brown, 1982], or a window from about 870 to 110 °C and preferably from 930 to 1040 °C [Lyon, 1975]. A key role in this process is attributed mainly to the OH radical concentration, see e.g. [Kimball-Linne&Hanson, 1986]. Low temperature leads to relatively low radical concentrations and thereby the reduction process is negatively influenced.

68

Some combustible additives (like hydrocarbons or hydrogen) may ignite at lower temperatures than NH3, so more radicals will then be produced, whereby the breakdown of NH3 via NH2 is promoted. An increase of the NH3/NO ratio (from 0.5-10) does not significantly change the optimum temperature but widens the temperature window and increases the maximum NO reduction. [Duo, 1990] Decomposition of NH3 over dolomite as well as over CaO and MgO has been reported to take place readily in several gas mixtures [Björkman & Sjöström, 1991]. These authors also report that the presence of light hydrocarbons (CH4, C2H4) can inhibit the decomposition reaction by carbon formation on the dolomite sites; the presence of H2 and H2O decreased the reaction rate by influencing the equilibrium negatively. NH3 decomposition has been demonstrated under gasification conditions at high temperature with Fe- and Ni-containing catalysts, see e.g. [Leppälahti et al., 1991] and [Krishnan et al., 1988]. Catalytic processes can also be applied downstream of hot gas cleaning units for removal of solids to a large extend and also trace elements that can harm the catalyst in shorter or longer term. These are Selective Catalytic Decomposition (SCD) where NH3 is decomposed into harmless N2 and H2, carried out at high temperatures (window 800-950 °C) and Selective Catalytic Oxidation (SCO), where (optimally ca. 400-500 °C) NH3 is oxidized with O2, NO, NO2 or mixtures at lower temperatures, resulting in N2 and water. The SCD process has been studied by [Simell et al., 1996;1997] and [Mojtahedi et al., 1995]. The relatively high temperature window in which this process is optimal is not always attainable and requires a good thermal stability of the catalyst. Catalysts that possess good activity are usually based on transition or noble metals (e.g. nickel, iron, ruthenium). Ni-based catalysts have been used to decompose ammonia ina simulated gasifier gas with excellent results [Mojtahedi and Abbasasian, 1995] and [Krishnan et al., 1988]. Close to 85% ammonia decomposition was achieved, corresponding to almost 99% approach of the equilibrium ammonia concentration [Mojtahedi and Abbasasian, 1995]. The catalysts used in all the tests were monolith structures; metallic monoliths coated with a porous washcoat which contained catalytically active metals. The catalysts had a high specific area, excellent thermal durability, low pressure drop and also high activity at high pressure. These materials are very sensitive for sulphur poisoning. [Gangwal et al., 1996] reported that HTSR-1 catalyst, a proprietary Ni-based catalyst on a thermostable Haldor-Topsoe carrier exhibited excellent activity for NH3 decomposition in simulated Texaco gas without H2S at 725 °C. With H2S this catalyst was poisoned but the activity could be restored at 800 °C even in the presence of H2S. MoS2-based catalysts showed low activity for NH3 decomposition and surface area stabilisation with ZrO2 was necessary for these catalysts to have any activity at all. The catalysts containing Ni, Co, Mo and W on high surface TiO2 support showed modererate activity (10-20% decomposition) for NH3 decomposition at 725 °C. The TiO2 showed sintering at these temperatures and required stabilisation. Mixing the catalytic material with zinc titanate sorbent allowed the catalysts to function longer. As sorbent got loaded with H2S, the exit H2S level increased accompanied by activity decrease for NH3 decomposition. SCO can be used at lower temperatures than SCD and with cheaper catalyst material, e.g. Al2O3. The lower temperature has the advantage that mainly the NH3 destruction reactions are catalysed whereas otherwise the gas system remains far from global equilibrium. This means that lower NH3 levels are attainable in SCO than in high-temperature catalytic decomposition [Leppälahti et al., 1998]. A mixture of NO2 and O2 was shown to reduce NH3 levels in simulated gasification product gas to a large extend below 500 °C. At higher temperatures the destruction efficiency was lower, due to the phenomenon that NO2 decomposes too fast to be available for NH3 at the catalyst surface. Also, at the higher temperatures other species like H2, CH4 and CO start to consume oxidizer speices. The use of NO/O2 as oxidizer mixture gives good NH3 removal results at slightly higher temperatures. Addition of O2 alone also reduces NH3 significantly over an Al2O3 catalyst, but at higher temperatures (upto 700 °C).

69

2.7

Conclusions and research requirements

There is a need for more detailed experimental studies on the effect of scale, pressure, temperature (resulting from applied stoichiometry), additives, steam addition and hot gas filtration using ceramic filters on gas composition for pressurised fluidised bed biomass gasification based on air/steam as oxidizers, especially regarding nitrogen compounds. There is a scarcity of accurate experimental data from pressurised fluidised bed pilot scale test rigs, especially from positions within the primary reactor. A better understanding of fuel nitrogen release during pyrolysis of biomass species is also needed because this is the initial conversion step in gasification and combustion in practical fluidised bed systems. In order to be able to predict formation of NOx precursors, a model with adequate detail in nitrogen chemical reaction rates should be developed and validated using the abovementioned experimental data.

70

Chapter 3 Experimental set-ups and measurement techniques 3.1 Introduction In order to validate and develop models which describe the formation of the main LCV gas components and the fate of fuel bound nitrogen during pressurised fluidised bed gasification of biomass and/or coal, experiments are necessary which enable the determination of the distribution of the relevant chemical compounds in the gas and solid phases. For this purpose gasification experiments were performed on two scales, using the 1.5 MWth Delft Pressurised Fluidised Bed (PFBG) test rig (described in section 3.2) and a smaller 50 kWth pressurised fluidised bed test rig at the Institut für Verfahrenstechnik und Dampfkesselwesen (IVD) at Stuttgart University (DWSA, description in paragraph 3.3). Supporting experiments using TG-FTIR were performed in cooperation with Advanced Fuel Research Company Inc. (AFR, USA) to determine kinetic properties of the initial pyrolysis behaviour of the fuels involved in the experimental programme. Details of this analysis technique are presented in section 3.4. At the Gas Dynamics laboratory of the Technical Physics department of Eindhoven Technical University a small heated grid reactor was used to determine the yields of the main gaseous components and the partitioning of fuel bound nitrogen during the flash pyrolysis process, the initial process of gasification. Details of the heated grid reactor are given in paragraph 3.5. 3.2 The Delft Pressurised Fluidised Bed Gasification (PFBG) test rig 3.2.1 Description of the rig The test rig at the Laboratory for Thermal Power Engineering of Delft University of Technology, which is used to conduct the large scale gasification experiments, consists of an air blown pressurised fluidised bed gasifier (PFBG), a ceramic wall-flow filter and a pressurised combustor for the Low Calorific Value (LCV) gasification fuel gas (Figure 3.1). The main design data of the fluidised bed reactor are presented in Table 3.1. The gasifier consists of a cylindrical AISI310 stainless steel vessel, which is placed inside a pressure vessel. Compressed air supplied by two diesel-driven Ingersoll-Rand screw compressors, each with a capacity of 1000 mn3 enters the bottom of the fluidised bed through a distributor plate, after having been preheated in the annular space between inner and outer vessel. Pressurised steam can be introduced into the reactor through a central nozzle in the air distributor plate with 72 holes with a diameter of 4 mm each. The outer pressure vessel wall is water-cooled over the maximum bed height. Start-up of the test rig is accomplished by atmospheric combustion of natural gas; hereby the flow control is dictated by CO emission concentration measurement. After the bed material has reached a temperature of 800 °C, the in-bed feeding of solid fuel is gradually started, the test rig is pressurised and natural gas flow is stopped. Fuel(s) and optionally solid additive (e.g. limestone or dolomite) are fed from three big bags onto a conveyor belt by means of three independently controlled screw feeders and are transported into a double-valve lock hopper system, from where it is fed by an adjustable-speed screw into an intermediate vessel. The feed rate of the fuel mix is determined from calibration data of the three screw feeders and the lock hopper filling cycle time. Bed material can be fed into this intermediate vessel from a separate (pressurised) storage vessel. From there, the material is fed pneumatically with a part of the fluidisation air into the bed through a single feed point in the distributor plate.

71

A cap construction is directing the flow towards the top part of the central nozzle. The cap is placed on top of the opening of the feeding point in order to disperse the supplied fuel and air in lateral direction. Gasification is attained by increasing the solid fuel flow (at a constant air flow) to a value corresponding to the stoichiometry aimed at. The bed contents can be kept constant by using an automatic solids removal system, with nitrogen purge, at the bottom part of the reactor, which is controlled by the bed pressure drop. The bed section further contains a vertical probe with 6 (K-type Chromel-Alumel) thermocouples mounted on a rod which is mounted on the air distributor plate. The thermocouples are located at a distance of 100 mm, 200 mm, 400 mm, 600 mm, 800 mm and 1000 mm, respectively from the distributor plate. This type of thermocouples is used because of the good chemical and mechanical properties, good stability and a relatively high electrical sensitivity of 40 µV/°C. The freeboard is well insulated and can be considered to be adiabatic. It is free of internals, except for three radial probes inserted through ports (P1.1, P2.1 and P3.1 in figure 3.1). Temperatures are determined by the abovementioned type of thermocouples at probe tip positions, one separately inserted through the freeboard wall at the position of probe P2.1 and one situated in the gasifier top. The reactor pressure is measured at two positions in the freeboard. The pressure is controlled by the outlet control valve downstream of the pressurised combustor. A wall-flow ceramic filter is used with high solid removal efficiency. The filter unit, manufactured by Cerafilter, a daughter company of Foster Wheeler Inc., consists of three honeycomb like elements consisting of β-cordierite, Al3(Mg,Fe)2[Si5AlO18]. These are coated with a ceramic membrane with pore sizes (ca. 0.3 µm) considerably smaller than that of the support matrix. The filter elements are cleaned by frequent electrically preheated (circa 200 °C) nitrogen pulsing, to maintain constant base-line pressure drop. The pulsing action is performed in carrousel mode, one at a time. The valves for this pulsing action are fast acting (100 ms time basis) solenoid valves. The dislodged ash, containing significant amounts of non-converted carbonaceous solids, falls into a lock hopper and transported in an inert nitrogen atmosphere to a (relatively) cold filter where solids are collected in big bag, which is weighed on-line. 8026

P6.1: G.A./S.A.

Gas Turbine Combustor

6385

1250

1000

P3.1: G.A. 485

Preheated Nitrogen Pulse Gas (3 x) P7.1: S.A.

1000

2985

P8.1: G.A.

P2.1: G.A. P1.1: G.A.

1000

E10.1: G.A.

37

Cooling Air

2550

E11.2: G.A.

380

V-cone flowmeter

585

0 level

Steam

Gasifier RevNo Revision note

Primary Air

Date

Signature

Air+fuel Checked

Bed removal

S.A.

Ceramic Filter

E13.1:G.A. G.A.: Gas Analysis S.A.: Solid Analysis Itemref

Quantity

Designed by

Title/Name, designation, material, dimension etc

Checked by

Article No./Reference

Approved by - date

Filename

WERV.DWG

M.TJON-A-SAN Owner

TU-DELFT

Scale

1:24

19-3-97

PFBC/G INSTALLATIE

800 1124 32.0 64.0

Figure 3.1 Schematic of the PFBG test rig; dimensions in [mm]; G.A/S.A.: probe positions.

72

Date

Title/Name

Drawing number

Edition

Sheet

Table 3.1. Main design data of the PFBG gasifier. General Maximum pressure (MPa) Maximum thermal capacity (MW) Fluidised bed reactor Bed diameter (m) Central nozzle ext. diameter (mm) Central nozzle height (mm) Number of holes in central nozzle (-) Diameter per hole (mm) Maximum bed height (m) Freeboard diameter (m) Minimum freeboard height (m) fluidisation velocity (m/s)

1 1.5 0.38 76 72 72 4 2 0.485 4.5 0.5

Figure 3.2 a and b show the vicinity of the test installation.

Figure 3.2

a. PFBG view from combustor side

b. view of the PFBG freeboard & probes

Figure 3.3 illustrates the working principle of the ceramic wall-flow filter.

Figure 3.3 Wall-flow filter concept with alternate plugged cells.

73

A criterion for the transition from the combustion to the gasification mode was established in the course of this research work. At first, the transition was made when the fuel gas inlet temperature at the LCV gas combustor was high enough (i.e. 600-650 °C) to ensure reliable fuel gas ignition. Later, it was discovered that filter cake ignition takes place at filter temperatures of ca. 350 °C, see [Andries et al., 2000] and it was decided to make the transition from the combustion to the gasification mode at a filter temperature of approximately 300 °C. When the transition to gasification is made, ignition of the gas in the combustor is started. Two types of pressurised combustors have been applied during the investigations: 1) a down scaled ALSTOM Typhoon combustion chamber with swirlers for air and LCV fuel gas and 2) a non-swirl flow based combustor with a bluff body designed and built at Delft university. A detailed description of these combustors is given by [Hoppesteyn, 1999]. An electrical air preheater is used to preheat the combustion/cooling air to ca. 350 °C. The exhaust gas of the pressurised combustor is cooled in a water-cooled double pipe heat exchanger before the pressure control valve at the exit of the installation. The exhaust gas is then led through an atmospheric afterburner, operated on natural gas, to remove the remaining combustible components. The installation has been equipped with an advanced Sattcon-2000 Programming Logic Analogue Control (PLAC) system supplied by Alfa Laval Automation AB with DOX10 version 3.4 as operation software. A Supervision Control And Data Acquisition (SCADA) package is used for process visualisation and data acquisition. Every 100 ms all digital and analogue in- and outputs are sampled and stored. The measured process data, available as integer values and with scaling factors converted to real physical data, are stored in a database with a frequency of 1/10 Hz. For operation, control and measurement purposes the Sattcon2000 control programme has available 128 analogue inputs, 24 analogue outputs, as well as 96 digital inputs and 96 digital outputs. Entering gas flows and the cooled down flue gas from the pressurised combustor are measured by vortex volume flow measurement instruments (Endress & Hauser). The measurement is based on the frequency of sound generated by the gas flow around a wing-shaped body. With turbulent flow, the frequency is proportional to the gas velocity. The hot and pressurised LCV product gas flow behind the ceramic filter is determined using a V-cone device, manufactured by McCrometer. This flow measurement technique is based on the pressure drop created by the insertion of a conically shaped body in the gas stream. The product gas flow is also determined by evaluating the nitrogen balance over gasifier-filter subsystem. During normal gasification operation the following process parameters are controlled: gasifier pressure, fluidisation velocity (based on actual average bed temperature, pressure and inlet gas flows), inlet gas flows (air, steam, nitrogen) and the temperature of the cooling water from the heat exchanger downstream of the combustor. When the installation is shut down (planned or in cases of emergency) the fuel supply and the air flows to combustor and gasifier are stopped, blanketing nitrogen is entering the gasifier and the pressure control valve is set to a fixed position, which can be changed by manual control, to relieve the pressure in a controlled way. After the bed has cooled down to approximately 200 °C, the bed inventory is removed from the bottom part of the reactor and samples are taken for further analysis (like proximate and ultimate analysis). As the devolatilisation of the last fresh biomass particles proceeds much faster than decreasing the reactor temperature, it can be assumed safely that no more than 10% of the total amount of carbonaceous material in the bed inventory is formed after switching off air and fuel supply [Kersten, 2002].

74

3.2.2 Analysis and sampling techniques Gas sampling is performed at the sample points indicated as G.A. in Figure 3.1. For all experiments sampling was performed after the ceramic filter unit. In some of the experiments, sampling was also done just before the ceramic filter and at the three axially centred positions in the freeboard. Details of the sampling probes are given in paragraph 3.2.2.4. The gas analysis instrumentation consists of an off-line Fourier Transformed Infrared (FTIR) spectrophotometer, an off-line multi-component and an on-line single component gas chromatograph and several single component on-line analysers of Non-Dispersive Infrared (NDIR) Non-Dispersive Ultraviolet (NDUV) and Para magnetism based type. The equipment is described below. The on-line analysers are calibrated before each test using certified calibration gases. 3.2.2.1 FTIR spectrophotometer A Fourier Transformed Infrared (FTIR) spectrophotometer (Nicolet 560 with OMNIC software package) which uses a heated gas cell (kept at 150 °C) with an optical path length of 2 m is used for identification and quantification of H2O, NH3, HCN, N2O, NO, NO2, CO, CO2, CH4, C2H4, C2H2, HCl, SO2 and COS. FTIR-spectroscopy can be used for the simultaneous measurement of compounds of interest, which are absorbing in the mid-infrared region, such as carbon oxides, hydrocarbons, sulphur oxides, nitrogen oxides and other gases. Infrared spectroscopy is based on the interaction between (gas) molecules and infrared radiation with a wavelength between 0.7 and 1000 micrometers (or wave number between 10 and. 15000 cm-1). The type of molecules mentioned, absorb infrared radiation and then show transitions in states of vibrations, like symmetric or asymmetric stretching, scissoring, bending, twisting, wagging, rocking and deformation. Gases without a dipole moment (e.g. gases consisting of homonuclear diatomic molecules consisting of atoms of the same kind, like N2, O2 or H2 and noble gases, like He or Ar) cannot be analysed by infrared spectroscopy. When infrared light passes through a gaseous sample consisting of molecules with a dipole moment, part of the irradiative energy will be absorbed at a specific wave number. The amount of energy absorbed is a function of the path length through the gas and the concentration of the absorbing species. The determination of the concentration of the absorbing gas species is based on LambertBeer’s Law:

A(υ ) = b ∑ a i'(υ ) c i' . i' where: A(υ) = absorbance at wave number υ ai’(υ) = absorption coefficient at wave number υ of species i' b = path length through the sample ci’ = concentration of species i' in the sample

(3.1) [-] [m-1] [m] [-]

υ = λ-1 = f v-1 λ v

= wavelength = radiation velocity

(3.2) [m] [m.s-1]

A(υ) = log { I0(υ)/I(υ) } I0(υ) I(υ)

= intensity of the incident radiation = intensity of transmitted radiation

(3.3) [*] [*]

Figure 3.4 shows a schematic of an FTIR, including the main elements of a Michelson interferometer, which forms the heart of the FTIR. This device divides an infrared light beam from a ceramic light source into two bundles by using a beam splitter, which splits the IR beam approximately in ideally

75

two identical bundles, and recombines the bundles after reflection into two mirrors. The two beams are reflected by a fixed mirror and a mobile (scanning) mirror and returning back to the beam splitter where the two beams are recombined. In figure 3.5 the resulting interferogram is shown when either monochromatic radiation or white light is traversing the interferometer.

Figure 3.4 Elements of a Michelson Interferometer

Figure 3.5 FTIR spectrometry. (a) basic components (b) working principle of the interferometer: single frequency source (central panel,left) is modulated to a cosine wave signal observed by the detector (central panel,right). A white-light source (e.g. emitted from a globar) is transformed to the interferogram (lower panel) [Naumann, 2000].

Due to interference, the intensity of the recombined beam leaving the beamsplitter is a function of the difference in optical path length, or retardation (δ), of the two beams (figure 3.5). The retardation is a function of time caused by the movement of the mirror. δ = 2 (OM – OF) (3.4) with: OM = distance between beamsplitter and movable mirror [m] OF = distance between beamsplitter and fixed mirror [m]

76

A laser beam (shown in figure 3.4), which is traversing the same interferometer, is being used to produce a time-dependent signal that is used to trigger the A/D conversion of the time dependent infrared signal (figure 3.5). The trigger frequency is determined by the frequency of the laser light and enables a very fast and accurate sampling of the detector signal. Fourier transformation is used to calculate the intensities of the single frequency signals, based on the time-dependent signal, which combined together, determine the interference pattern measured at the detector. In this way a spectrum (a plot of intensity, or absorbance, versus wave number, of the infrared light falling on the detector is determined. According to [Griffiths & De Haseth, 1986] the interferogram measured with an ideal interferometer is given by:

(

I (δ ) = ∫ 0.5 I (υ ) cos ( 2πυδ )

)

dυ

`

(3.5)

Equation (3.5) describes a cosine Fourier transform, relating the time domain to the frequency domain. As the beamsplitter is not ideal, as well as detector response and electronic component (e.g. A/D converter, amplifier and low pass filter) characteristics, the result is non-ideality of the interferometer. This phenomenon can be described with a wave number dependent correction factor, H(υ):

I (δ ) =

∫ ( 0.5 H (υ ) I (υ ) cos ( 2πυδ ) )

+∞

dυ

(3.6)

−∞

equalling the factor 0.5 H(υ) I(υ) to B(υ), leads to: I (δ ) =

+∞

∫ ( B (υ ) cos ( 2πυδ ) )

dυ

(3.7)

−∞

The spectral range is not infinite and the retardation of the instrument is limited. This limited retardation is the cause of the restricted resolution of the instrument: the longer the retardation, the higher the spectral resolution. The distance scanned by the moving mirror is approximately inversely proportional to the resolution [Griffiths & De Haseth, 1986]. B(υ) also can be written as Fourier transform equation related to I(δ): B (υ ) =

+∞

∫ ( I (δ ) cos ( 2πυδ ) )

dδ

(3.8)

−∞

As the interferometer has a limited and not infinite retardation, the interferogram has to be multiplied with a so called truncation function, D(δ): (3.9a) D(δ) = 1 for –∆ < δ < ∆ D(δ) = 0 for δ > ∆ (3.9b) with ∆ being the maximum retardation. Equations 3.9 a and b represent the so-called boxcar truncation function. The spectrum that results from the interferometer with a maximum retardation can be described by: B (υ ) = ∫ ( I (δ ) D (δ ) cos ( 2πυδ ) ) d δ

(3.10)

The Fourier Transform of the product of two functions is the convolution of the Fourier Transform of each function. The Fourier Transform of I(δ) is the true spectrum, B(υ), while the Fourier Transform of D(δ), f(υ), is represented by: f(υ) = 2∆ sin (2πυ∆) / (2πυ∆) = 2 ∆ sinc (2πυ∆) The convolution, G(υ) of both B(υ) and f(υ) can be described by:

(3.11)

77

(( )(

G (υ ) = ∫ B υ ' f υ − υ '

)) dυ '

(3.12)

Figure 3.6 shows a plot of several applied truncation functions and their Fourier Transform. As can be seen from this figure, the boxcar-derived function f(υ) is a symmetric function with a strong peak in the centre and slowly damping side peaks. These oscillations in f(υ) lead to disturbances in the edges of spectral absorbance bands after the convolution operation. By using another type of truncation function, this problem can be diminished. The technique is called apodisation.

Figure 3.6 Apodisation functions; a) boxcar truncation; b) trapezoidal; c) triangular; d) triangular squared.

Another apodisation function is the Happ-Genzel function. This truncation function is applied in our FTIR instrument and is given by: ⎛ δ⎞ D (δ ) = 0.54 + 0.46cos ⎜ π ⎟ ⎝ ∆⎠ D(δ) = 0

for –∆ < δ < ∆

(3.13a)

for δ > ∆

(3.13b)

Ideally, an interferogram is perfectly symmetrical around the zero path difference point. In practice, the interferogram is somewhat asymmetrical due to optical (beamsplitter characteristics) and electronic (filters to remove high frequency noise) effects. The asymmetry of the interferogram can be compensated by adding a negative correction term for the phase angle, (2πυδ), in equation (3.5). This technique is known as phase correction. The most common phase correction is the so-called Phase Apodisation, which is described in detail by [Griffiths & De Haseth, 1986]. Placing a gas cell between the beamsplitter and the detector enables the measurement of an absorption spectrum. After measuring a background spectrum, by flushing the gas cell with an IR inert gas like nitrogen, a real spectrum can be recorded using the gas to be analysed. Subtracting the background spectrum from the measured spectrum gives ‘net’ spectrum.

78

Lambert Beer’s law (3.1) together with calibration spectra then are used to determine the concentration of components quantitatively. In the instrument used for our research, the infrared light is produced by a Nickel-Chrome glower and is in the range of 50 to 9600 cm-1. The beam splitter is made of KBr plates with a germanium (Ge) coating. The light is transferred through a sample holder. The sample holder is a gas cell, which has a volume of 200 ml to contain the sample gas and is kept at a temperature of 150 °C. Mirrors reflect the light resulting in an optical length of 2 meters. The mirrors in the gas cell are plated gold. The windows of the gas cell are made of Zinc-Selenide (ZnSe). A Mercury Cadmium Telluride (MCT) detector is used. This type of detector is based on semiconductor technology and needs to be kept at a working temperature of -196 oC. Therefore it is cooled with liquid nitrogen. The infrared radiation detected first passes a diaphragm. This optical element controls the amount of infrared light that reaches the detector. During the measurements of both the background and the sample gas spectra, the diaphragm needs to be the same. A photograph and a schematic of the FTIR system are shown in figure 3.7. The gas transfer lines are traced to 150 oC and consist of stainless steel. The gas from the PFBG test rig is led to this section by a flexible heated Teflon transfer line. The instrument is calibrated using certificated calibration gases. A range of calibration concentrations is established by adding pure nitrogen to the calibration gas by using (Brooks 6800 type) mass flow controllers, and mixing them a vessel with glass beads.

Figure 3.7 FTIR equipment and calibration schematic.

Table 3.2 gives the main characteristics of the FTIR applied in the research. Calibration ranges and corresponding signal to noise ratios of the main species are presented in Table 3.3. These ratios have been determined for the lowest and highest calibrated concentrations; as worst case the signal to noise ratio of the lowest concentrations are given. The number of scans has been varied. A value of 15 has been chosen as a trade-off between increasing accuracy with increasing number and analysis time [Bosma, 1997]. The signal level was determined at IR absorption regions specific for the species presented in appendix 1.1.

79

Table 3.2 Characteristics of the FTIR instrument. FTIR- spectrometer Spectral range Resolution Number of scans per measurement Measurement time Infrared source Detector Gas cell Volume gas cell Operation temperature of gas cell Operation pressure of gas cell Gas flow through sample cell Optical pathlength through gas cell Material of mirrors in gas cell Material of windows in gas cell Beamsplitter

Nicolet 560 400-4000 cm-1 0.125 cm-1 15 70 s Heated Nichrome glower Liquid nitrogen cooled MCT Stainless stell vessel with electrical heating 200 ml 423 K 13 mbarg Approximately 10 ln/min 2m Au ZnSe KBr plates with Ge coating

Table 3.3 Gas compounds analysed by FTIR, calibration ranges and signal to noise ratios Species CO CO2 CH4 C2H4 NH3 HCN NO NO2

Calibration Range [ppm] 40 - 207000 20000 -200000 5000 - 50000 90 - 29000 300 – 3000 50 – 500 20 – 200 20 – 200

Signal/Noise Ratio for lowest conc. 11 - 80 12 - 170 7 - 70 10 - 66 8 - 80 8 - 70 2 -26 6 - 110

Species N2O H2O HCl COS SO2

Cal. Range [ppm] 20 - 200 2000 - 20000 9.1 – 91.2 5 - 500 10 - 1000

Signal/Noise Ratio for lowest conc. 10 - 110 10 - 80 not determined not determined not determined

A typical FTIR analysis of sample gas takes about 1 minute and the gas cell is flushed with the gas to be analysed during at least 5 minutes. An example of a spectrum of biomass gasification product gas (non-condensed) is given in figure 3.8

80

CO2

H2O

HCN

C2H2

C2H4 CH4

NH3 CO

Figure 3.8 A typical raw biomass gasification product gas spectrum.

3.2.2.2 Gas Chromatography A gas chromatograph of Chrompack 9001 (‘Refinery Gas Analyser’, RGA) type is operated off-line for analysis of C1-C5 aliphatic hydrocarbons, H2 and the following gases: CO2, CO, C2H4, CH4, N2, O2, Ar and H2S. Figure 3.9 shows a schematic of the RGA. Before the sample enters the RGA, the gas is dried by leading the gas through a set consisting of a water cooler and an ice cooler.

Figure 3.9 Schematic of the RGA gas chromatograph

Gas Chromatography is based on the separation of a gas mixture into its components. The separation takes place in a (series of) tubes (columns), filled with separation material (Stationary Phase). A gas sample is injected into a carrier gas flow (helium, argon or nitrogen), which transports the gas sample through the column. Inside the column the separation process takes place. The system consists of two

81

phases, a mobile, moving phase and a stationary solid phase. The mobile phase can be either a liquid or gas. If the components in the mobile phase partition differently over the two phases in the column, they are separated. A component which stays longer in the stationary phase will be transported less

quickly and a component which stays in the mobile phase will be transported faster and have a shorter residence time. The time needed to exit the column is called: the retention time. An example of a separation process is shown in figure 3.10. Figure 3.10 Separation principle gas chromatography. Downstream of the column(s) a detector is placed, which gives a time-dependent electrical signal, which is a function of the concentration and the retention time of the gaseous components in the gas sample. The detectors applied in our research are of the Flame Ionisation (FID) and Thermal Conductivity (TCD) type. An FID measures the electrical conductivity of a flame, basically a Hydrogen flame. The flame conductivity increases when organic molecules from the gas sample enter the FID flame, because carbon containing ions and electrons are formed. These carbon containing ions and the electrons are formed from prevailing radical species, see e.g. [Sternberg et al., 1962]. Two electrodes, located inside the FID, measure the increased conductivity. The TCD detector measures the heat conductivity of a gas, which depends on the gas composition. The working principle of the two detectors is illustrated in figures 3.11 A and B:

Figure 3.11 A: Thermal Conductivity Detector. B: Flame Ionisation Detector Thermal Conductivity Detector 1 column outlet 2 reference channel 3 measuring cell 4 reference cell 5 hot wires

82

Flame Ionisation Detector 1 column outlet 2 hydrogen 3 air 4 cathode 5 flame 6 anode 7 outlet

For the analysis of the light aliphatic hydrocarbons, three columns in series are applied: a fused silica column with CP-SIL 5 CB stationary phase material, 12.5 m length and 0.32 mm inside diameter [ID], a fused silica column with Al2O3/Na2SO4 stationary phase material of 50 m length and 0.53 mm ID and finally a 3 m, 0.32 mm ID fused silica column with deactivated stationary phase. On this column set-up Helium is used as carrier gas. The detector in this case is a Flame Ionisation Detector (FID). Hydrogen is separated and analysed applying two columns in series i.e. a Stainless Steel column with 80-100 mesh Hayesep Q support of 1.0 m length and 2.0 mm ID and a 1.0 m 2.0 mm ID Stainless steel column with 80-100 mesh Molsieve 5A support. For this part, nitrogen is used as carrier gas and a Thermal Conductivity Detector (TCD) is applied for the quantification of the H2 concentration in the gas. The permanent gases, CO2, CO, C2H4, CH4, N2, O2, Ar and H2S, are separated and analysed applying three columns in series: a 0.5 m 2.0 mm ID Ni column with 80-100 mesh Hayesep T support, a Ni column with 80-100 mesh size Hayesep Q stationary phase of 0.5 m and 2.0 mm ID and a Stainless steel column of 1.5 m length, 2.0 mm ID with Molsieve 13X support. Here, Helium is used as carrier gas. A TCD detector is used in the apparatus for component analysis. A typical RGA analysis takes about 40 minutes. An on-line micro-gc of Chrompack 2002 type is applied for H2 analysis. Here a column packed with molsieve 5A material is applied. Carrier gas is Nitrogen. The applied detector in this device is of the TCD type. The detector temperature is 160 °C. A measurement takes approximately 45 s. 3.2.2.3 On-line Non Dispersive Infrared / UV, colorimetric and paramagnetism based analysers For quantitative analysis of CO, CO2, SO2 Non Dispersive Infrared (NDIR) single component analysers are applied in continuous operation. The term Non Dispersive refers to the fact that all the light passes through the gas sample and is only filtered immediately before the detector. Dispersive IR detectors use a grating or prism to pre-select the desired wavelength of light and pass only this through the gas sample to the detector. The analysers for these gas species have been manufactured by Elsag & Bailey and are of the following type: URAS 3G (CO and SO2), URAS 3 (CO2), URAS 10P (CO and CO2). Figure 3.12 shows a schematic of an NDIR analyser. The measurement technique is like FTIR based on the principle that polyatomic, nonelemental gases having a dipole moment absorb radiation in the midinfrared region of the spectrum. The infrared beam emitted by the source (St 1) is divided and mechanically modulated. The beam falls alternating into the analysis and reference chambers (M 1 and M 2) of the sample cell (M) and into the sample and reference chambers of the detector (E). The sample gas is directed into the sample cell. The reference chamber is filled with a gas (N2), which does not absorb infrared light. The four chambers of the detector contain the measuring components. The front chambers (E 1.1, E 2.1) and the rear chambers (E 1.2, E 2.2) are connected to each other by channels. Depending upon changes in the concentration of the component of interest in the measuring chamber of the sample cell, the infrared beam reaches the detector chamber in a more or less weakened state. The beam passing through the reference chamber of the sample cell reaches the reference cells of the detector without being influenced. Thus, an energy difference is produced, the result of synchronous pressure fluctuations between the separate detector chambers caused by the modulation. This pressure difference is detected by a diaphragm capacitor and converted into an electrical signal proportional to the concentration of the component of interest. The signal is then amplified, linearised and provided as a continuous, standardised output signal for further processing.

83

AT: ET: E: E1.1: E 1.2: E 2.1: E 2.2: E 3: E 4: E 5: M: M 1: M 2: St 1: St 2: St 4: St 9:

Analyser section Electronics section Detector (4-chamber detector Front measuring chamber Rear measuring chamber Front reference chamber Rear reference chamber Partly transparent window Metal Diaphragm Counter Electrode Sample cell Analysis chamber Reference chamber Radiator coil (IR source) Beam splitter Chopper wheel Light barrier

Figure 3.12 Schematic of the NDIR analyser (example Uras 3).

The most important sulphur-containing compound, H2S, is measured on-line, using a Maihak Monocolor 1N instrument, which operates semi-continuously based on a colorimetric analysis principle. Quanitification of H2S by FTIR appeared not to be possible due to too strong overlap with CO2 [Bosma, 1997]. The H2S concentration is measured with a dry reaction on a test paper strip that is saturated with lead acetate, a chemically selective colour indicator. The colour change is evaluated photometrically. The paper strip is fed stepwise through a block with two geometrically identical chambers. The paper in these chambers is exposed to a light through a prism. The intensity of the reflected light is detected by two photo elements. During the paper feed intervals, the sample gas flows through a test paper section and causes staining of the indicator paper. The staining is proportional to the concentration and the flow rate. The reflected light of this paper section is compared with the unchanged reference section. The correlation between the intensity difference and the H2S concentration is determined quantitatively by calibration with a test gas of known composition (calibration). The principle of operation is shown in figure 3.13.

Figure 3.13 Operation principle of the H2S analyser. A Sample gas inlet B Sample gas outlet C Window / Filter D Test paper strip E Photo element F Instrument

84

G Lamp H Reference cell I Measuring cell J Transport roll K Press-roll L Wind-up roll

Analysers for continuous, on-line O2 measurement are of Elsag & Bailey’s MAGNOS 3, MAGNOS 6G and SERVOMEX OA. 137 type. Figure 3.14 gives a schematic of this instrument. The measurement is based on the paramagnetic behaviour of O2. The measuring compartment, through which the sample gas flows, is placed in a very inhomogeneous magnetic field, which is produced by two permanent magnets. A light displaceable body (dumb-bell) is suspended in the inhomogeneous magnetic field using torsion strips in such a way that it can rotate. If the sample gas contains O2, these molecules experience a force, which draws them into the magnetic field. The strongly inhomogeneous magnetic field thereby establishes an O2 partial pressure gradient. The partial pressure is greatest at places where the magnetic field strength is large. The partial pressure gradient exerts a torque on the dumb-bell. This torque and the corresponding rotary displacement of the dumb-bell are proportional to the O2 concentration in the samples gas. The primary torque due to the mentioned gradient is compensated automatically by a control loop. A photocell senses the dumb-bell position by means of infrared light beam reflected via a mirror. The resulting compensation current, which is proportional to the dumb-bell position, produces an electromagnetic torque in the conductor loops of the dumb-bell. This opposes the primary torque, so that the dumb-bell is restored to the original position. The compensation current is proportional to the oxygen concentration. After conversion and amplification, a continuous standard output signal is available for further processing. The measuring compartment assembly is housed in a thermostatic casing so that the reading is largely independent of ambient temperature fluctuations. ME: MA: MK: H: K: S: MS: F: OA: IR: E: RF: DA: BT: DVE: LT: NT:

Sample gas inlet Sample gas outlet Measuring compartment Dumb-bell Compensation loop Mirror Magnet gap Window Optical pickup Infrared diode Detector Feedback Digital display Operator keypad Digital processing electronics Power section Power supply

Figure 3.14 Schematic of an oxygen analyser (example Magnos 6 G).

For continuous NOx analysis, an Elsag & Bailey manufactured RADAS1 Non Dispersive Ultraviolet analyser is applied. A schematic of this device is depicted in figure 3.15. The measured effect is the absorption of the gas component in the UV spectral region. A hollow cathode lamp produces the UV radiation. A chopper wheel (B) causes UV radiation to be transferred through the system intermittently and the beam splitter (S) divides the radiation into two separated beams. The measuring beam passing through the sample cell (MK) reaches detector (E). The unchanged reference beam reaches the correction detector (KE). The electronic processing of these four signals eliminates the disturbing effects of nonselective absorptions, e.g. those due to dirt on the sample cell, and ageing effects of the radiation source and detectors. A linearised continuous standardised output signal is available for further processing. The signal is proportional to the volumetric concentration of the measured component.

85

AT: L: B: K: F: S: MK: E: KE: KK: ET: D1,D2: DV: SH:

Analyser section Lamp Rotating chopper wheel with gas filter Collimator Optical filter Beam splitter Sample cell Detector Correction detector Calibration cell Electronics section Quotient generator Difference amplifier Sample and hold circuit

Figure 3.15 Schematic of an NDUV analyser (example Radas 1 G).

3.2.2.4 Sampling probes and analysis of tar components For the analysis of aromatic tar compounds several sampling techniques have been used. A novel sampling method was developed at KTH Stockholm, Sweden [Brage et al., 1997]. The method is called Solid Phase Adsorption, or simply SPA method. This method was our standard tar measurement technique. The method requires that a hot gas sample (ca. 250-300 °C) is collected by adsorption and condensation at room temperature on a commercially obtained Bakerbond Solid Phase Extraction (SPE) tube containing 500 mg (1.3 o.d.x7.5 cm 7088-03) of amino-propylsilane phase (surface 400-600 m2/g) bonded to silica gel (40µm APD, 60 A). A steel needle is fixed to the tip of the SPE column and a gas tight syringe (100 ml volume) to the head of the column via a small length of silicone tubing. See figure 3.16 for a schematic of the sampling method. Prior to sampling, the SPE column is conditioned with ca. 0.5 ml Dichloromethane (DCM, ‘pro analyse’ quality) using pure nitrogen purging and then drying for ca. 10 min. at 105 °C. In order to prevent contact with contaminated surroundings, the SPE column was placed in a likewise conditioned test tube and tightly capped with a stopper before and after sampling. Tar samples are collected in circa 1 min. by manually pulling product gases through the column. The column is subsequently disconnected, placed in a test tube, capped with a polyethylene stopper and put in a freezer until a GC analysis was performed. This GC analysis was performed at KTH and the analytic procedures and GC configuration have been described by [Brage et al., 1997]. For that purpose, the column is eluted with DCM and DCM/ IsoPropylAlcohol (IPA) for Polyaromatic Hydrocarbons (PAH’s) and Phenolic compounds, respectively.

86

Syringe

Amino-phase Needle

Septum Septum holder

LCV gas

Figure 3.16 Schematic of tar sampling, according to the SPA technique.

Tar components were also analysed by absorption in liquid organic compounds, like Dichloromethane and 1-methoxy-2-propanol, with subsequent analysis by GC. Since 1998 activities (initially by IEA network) are going on to standardise sampling and quantification of tar levels in gasifier producer gas. As a result of this process, recently a Tar Sampling and Analysis Guideline was developed and published on the internet [Neeft et al., 2002]. As tar was not the focus of this PhD thesis, application of this Guideline was deemphasised during the gasification experiments. Additionally, a novel on-line method was applied, which was developed at Stuttgart University (IVD), see [Mörsch et al., 2000] and [Mörsch, 2000]. This technique is based on the use of an FID detector for total carbon quantification. A schematic of the method is shown in figure 3.17.

Figure 3.17 Working principle of the IVD on-line tar analyser.

Figure 3.18 shows a schematic of the gas- and tar sampling probes in the freeboard and in the tube section before the ceramic filter unit. Fly ash particles in the sample gas are removed by a small cyclone, with an inlet area of 8x4 mm and an internal height of 10 cm. There are three of these probes in the freeboard at 3.5 (P1.1), 4.5 (P2.1) and 5.5 m (P3.1) from the bottom plate of the gasifier; probe P1.1 is radially traversable, the other ones are fixed. The probe downstream of the ceramic filter (at position P7.1, figure 3.1) doesn’t contain a cyclone, due to the very low particulate concentrations in the gas at that position. The sample lines downstream of the probes are kept at a temperature between 200 and 300 °C in order to prevent the condensation of water, sulphuric species (see [Verloop, 1998]) and higher aromatic compounds. Higher temperatures are not possible due to material constraints. Sampling port positions were already presented in figure 3.1. To protect the gas analysers an additional quartz fibre fine filter operating at 200 – 300 C, is used, supplied by Schleicher & Schuell.

87

Downstream of this filter the gas can be cooled down for the on-line analysers, which require dry gas or kept above the condensation temperature in the heated transfer lines (largely consisting of PTFE lining and kept at ca. 150 °C) for the analyse of the wet gas by FTIR. When the probes are not in use, they are purged with a small inert N2 flow.

Figure 3.18 Schematic of a gas/tar sampling probe.

Figure 3.19 Detailed overview of a gas/tar sampling probe.

88

3.3 The 50 kW(thermal) IVD Pressurised Fluidised Bed Gasifier (DWSA)

3.3.1 Description of dimensions and operation Figure 3.20 shows the schematic of the pressurised fluidised bed test installation (DWSA) at IVD, University of Stuttgart. The DWSA has been used for several years for research in the field of conversion of solid fossil fuels under well-defined reproductive process conditions and on a small scale (50 kWth maximal), see [Nagel et al., 1998],[Nagel,2002]. A picture of the test rig is shown in figure 3.21. Table 3.4 presents the range within which the DWSA test installation process variables can be varied. In table 3.5 the main dimensions are given.

Figure 3.20 Schematic of the DWSA test rig.

Figure 3.21 Picture of the DWSA test rig, situated at IVD, University of Stuttgart.

89

Table 3.4 Operating range (gasification) of the DWSA test rig Variable

Range DWSA

Pressure (MPa) Temperature (°C) Air Stoichiometry, λ (-) Fluidisation velocity (m/s) Fuel

0.12 – 1.6 750 – 1000 0.3-1.0 0.1 – 1.0 Coal, Brown Coal, Biomass

Table 3.5 Main dimensions of the DWSA gasifier Bed diameter (m) Max. bed height (m) Number of nozzles Number of holes per nozzle (-) Nozzle hole diameter (mm) Freeboard diameter (m) Freeboard height (m)

0.10 1.0 4 6 2 0.177 3.0

The DWSA installation can be divided into two main parts. The first part consists of an air preheater, fluidised bed reactor, solid fuel dosing vessel with on-line mass determination system and a hot gas cleaning section with a cyclone and a single ceramic candle filter (Schumacher type). The presence of a cyclone is an important difference with the Delft PFBG rig. In the fluidised bed reactor the solid fuel is gasified with air to produce a low calorific value (LCV) gas, which is cleaned of fly ash and unreacted solid carbonaceous material. Air, which can be preheated, and nitrogen are introduced into the gasification reactor through four nozzles with a square pitch just above the distributor plate. These gas flows are measured using thermal conductivity based mass flow controllers. The reactor is electrically heated in order to maintain a constant temperature in the bed and the freeboard section. Also this is different from the Delft test rig, which is only well insulated. The solid fuel, of which the mass flow is determined by the decrease of mass in the bin in time, is fed into the bottom part the bed section just above the distributor by a horizontal screw feeder. The hot gas cleaning section is operated at temperatures of approximately 500 °C. The second part of the test rig consists of a combustion air preheater, a specially designed LCV gas burner, a flue gas cooler and a pressure control valve. The LCV gas combustor is situated in a watercooled pressure vessel. This swirl-diffusion combustor is centred in the ceramic combustion chamber. The combustion air is entering the combustion chamber through 2 concentric, annular channels. At the beginning of an experiment the gasifier is heated to the desired temperature and solid fuel is fed into the reactor and combusted (well above 600 °C). Hereafter the reactor is pressurised by closing and (manually) controlling the outlet valve situated downstream of the combustor. Then the transition to gasification is made by increasing the fuel flow and the electrical ignition of the combustor is started simultaneously adding preheated combustion air. The flame generated is observed through a video camera system. 3.3.2 Analysis techniques applied The analysis of the produced LCV gas is performed directly downstream of the Schumacher ceramic candle filter by means of continuous on-line O2 (paramagnetic), CO- and CO2 (NDIR) analysers. In addition, H2, CO, CH4 and N2 concentrations are measured off-line by means of a gas chromatograph. An FTIR of the same type as the one applied for the Delft PFBG experiments and described in section 3.2.2.1, is used for quantitative analysis of NH3, HCN, CO, CO2, CH4, C2H4, C2H2, HCl, COS and H2O.

90

The analysis of the combustion gases directly behind the combustor is performed by means of continuous on-line analysis for O2, CO, CO2, NOx, N2O and SO2. The operating principles of these analysers have been described in paragraph 3.2.2.3. The on-line analysers are calibrated before each test using certified zero and span gases. The solids from the bed, the cyclone and the filter have been sampled and weighed after each experiment. The solids are taken from the bed by removing the bottom section of the bed and collecting the bed content. No tar sampling and analysis has been performed in this test rig during the tests described in this work. Proximate and ultimate analyses have been carried out and the heating value has been determined. 3.4 The TG-FTIR set-up at Advanced Fuel Research Company Inc. (USA)

In order to characterize the devolatilisation behaviour of the fuels applied in this research and to determine the kinetic parameters needed as input for the FG-DVC model (see paragraph 2.4.1), a Thermogravimetric Analyser combined with an FTIR to detect evolved gases, has been applied. These experiments were performed by Advanced Fuel Research Inc. [Wojtowicz et al., 2001]. The TG was a DuPont 951 type and the FTIR was a Bomem M-100 FTIR spectrometer. See paragraph 3.2.2.1 for a background on the FTIR quantitative analysis technique. The TG-FTIR instrument is shown in figure 3.22 and consists of a sampleholder connected with a balance in a gas stream within a furnace. As the sample is heated, the evolving volatile products are carried out of the furnace directly into a 5.1 cm diameter gas cell. The cell is kept at ca. 155 °C to prevent condensation of water and higher hydrocarbon species. Here, the gases evolved are analysed. The FTIR spectrometer collects spectra every 30-40 s to determine quantitatively the evolution rate and composition of several compounds. The system allows the sample to be heated on a preprogrammed temperature profile, at rates between 3 and 100 K/min in a temperature window of 20 to 1100 °C. Isothermal steps with a specified hold time are also possible. The system continuously monitors: (1) the time-dependent evolution of volatile species; (2) the heavy liquid (in gaseous form, tar) evolution rate and its infrared spectrum with identifiable bands from the functional groups (a feature which is still under development and (3) weight of the remaining non-volatile material (char residue). Helium carrier gas is passed through an oxygen trap to ensure an oxygen-free environment during pyrolysis. The flow rate of carrier gas through the TGA system was about 390 ml/min, and the total gas flow through the gas cell was 978 ml/min. Hereby secondary reactions were minimized. The optical path length of the gas cell was 16 passes x 27.2 cm (cell length) = 453.2 cm.

Figure 3.22 Schematic of the TG-FTIR set up at AFR.

The sample size was 13-16 mg in the case of biomass and 30-35 mg in the case of coal. Pelletised fuel, wood and Miscanthus, were ground using an A-10 Tekmar analytical mill, and subsequently sieved to pass a 35 mesh sieve, resulting in particle sizes below 500 µm. 91

Biomass samples were heated in Helium at 10 K/min, initially to 80 °C to dry the sample for 20 minutes, and then up to 900 °C for pyrolysis. For coal samples drying was performed at 150 °C for 3 minutes. Upon reaching 900 °C and holding the temperature at this level for 3 minutes, the sample was immediately cooled to 250 °C over a period of 20 minutes. After this cooling period, a small Oxygen flow was added to the Helium sweep gas and the temperature was ramped to 900 °C at 30 K/min to combust the remaining char. This heating profile was repeated at 30 K/min and 100 K/min for all samples. In all experiments, FTIR absorbance spectra were obtained every 30 s. Concentrations of all volatile species with a dipole moment, except for tar, were obtained using quantification routines obtained from calibration runs performed with pure compounds. Tar evolution patterns and yields were determined by difference using the sum of gases quantified by FTIR and the balance curve obtained thermogravimetrically. The tar evolution therefore also includes evolved species not detectable by FTIR, like H2 and N2. 3.5 The heated grid reactor at Eindhoven University of Technology

In order to characterize fast pyrolysis of solid fuel particles, which is considered to be the initial fast fuel conversion step in fluidised bed gasification, a heated grid reactor is applied. This type of smallscale reactor is widely used for fast pyrolysis experiments, see e.g. [Cai et al., 1996] and [Man et al., 1997a and b]. An experimental set up equipped with laser diagnostics is situated at the Gas Dynamics laboratory of the department of Technical Physics of Eindhoven University of Technology. In the past, this small-scale experimental facility was used for coal pyrolysis, gasification and combustion kinetics measurement [Moors, 1998]. More recently, also biomass char gasification experiments were conducted [Guo, 2004]. Fast pyrolysis of solid fuel particles, with heating rates up to 9000 K/s (values characteristic for fluidised bed thermal conversion processes), can be attained in this device. Figure 3.23 shows the experimental set up and the main characteristic design data are presented in table 3.6.

a)

c)

b)

Figure 3.23 Schematic of the pressurised heated grid reactor at TU Eindhoven; a: view from aside; b: view from the top; c: side view with overview of analysis equipment Table 3.6 Main dimensions and characteristics of the heated grid set up. Cilinder diameter (mm) Cilinder length (mm) Reactor volume (cm3) Grid dimensions (mm x mm) Wire metal (2 types) Wire diameter (mm) Maximum pressure (MPa) Maximum temperature (K) Maximum sample amount (mg)

92

15 224 45.48 6x2 Platinum(90%)-Rhodium(10%), Stainless steel 0.076 (Pt-Rh), 0.14 (SS) 2.5 <2045 approx. 1

The 0.31 mm mesh grid (for Pt-Rh; 0.32 mm for SS) has a U-shape of approximately 6 mm long and 2 mm wide, with the opening to the side. The pieces of biomass or coal are sandwiched between the folded grid and heated from top and bottom. An external power supply heats up the grid. The temperature can be varied by manipulating the electrical current. Also, different heating rates of the grid can be achieved. The reactor is designed to withstand an internal pressure of 25 bar. The pressure is measured with a Kistler piezo-resistive transducer. Gas enters the reactor at the two sides of the cylinder. Sintered porous material in the gas inlets reduces the gas velocity to prevent that the biomass sample is blown away. The reactor is evacuated with a vacuum pump and refilled with nitrogen in order to create inert conditions before pyrolysis experiments. In the middle of the reactor there are two windows, made of ordinary glass (BK7), making it possible to observe the sample on the grid during the experiment. The temperature of the grid can be measured through one of these windows with a manual colour pyrometer. A chopper interrupts the beam intermittently because the photo-diode cannot operate continuously. The temperature of the grid is measured with a thermocouple. At grid temperatures above 800°C, the final grid temperature is measured with a manual colour pyrometer as well. This is done with an empty grid before an experiment starts. It is assumed that the particles on the grid are so small that they have the same temperature as the grid. With the steady-state end temperature and the signal of the photodiode the heating rate of the grid is reconstructed. On both sides of the reactor windows are mounted made of the ceramic material CaF2 with good IR transmission properties. These windows are not parallel but tilted at a small angle in order to avoid light interference effects between the two windows. The beam of the tuneable diode laser enters and leaves through these windows, scanning the full cross-sectional area of the cylinder for absorbing gases. Only a small volume is not scanned. For this part of the reactor volume (approximately 8 %) is compensated in the calculations of the results.

Figure 3.24 Open grid reactor.

Figure 3.25 Grid with thermocouple.

The heated grid reactor is used to measure the release of different gases from solid fuel particles insitu at high heating rates. For this purpose IR absorption spectrometry is applied, which makes use of the infrared absorption spectrum of molecules to detect them. In the experimental work the focus is on CO, CO2 and NH3. The tuneable IR laser used, is able to make scans of 2-4 cm-1. The laser diode (lead salt chip in a gold-plated copper package) in the set-up is cooled by liquid nitrogen and is tuneable by modulating the current and adjusting the temperature. The current through the diode determines the wavelength range of the infrared radiation.

93

The modulation current periodically changes the injection current from a value below threshold to another, thus the wavelength of the emitted radiation varies periodically. In this way a time-resolved spectrum can be obtained. The centre wavelength of the chosen laser depends on the temperature. Tuning the temperature will produce another absorption line [Guo, 2004]. The current is modulated to produce a saw-tooth signal with a frequency up to 20 kHz (20 scans per millisecond), as is shown schematically in figure 3.26, see [Will, 1999] and [Moors, 1998]. This signal is tuned to get the zero emission level and the absorption peak in one scan. If the specific gas is released in the grid reactor, part of the laser light is absorbed at the specific wavelength and the detector detects the decrease in light intensity. To calculate the absorption (related to the species concentration and yield) a computer program was written in Matlab 6.0. The program is described in [Slabbekoorn, 2002].

zero level

zero level

intensity on detector

intensity on detector

absorption peak Time, wavelength

Figure 3.26 Schematic of the measured signal as a function of time (also wavelength).

A reference cell (see figure 3.23) is filled with the gas to be analysed, in order to determine at what conditions a suitable absorption peak occurs, and in order to be sure the right gas is detected in the grid reactor. The laser beam is split by means of a coated thin grating as beamsplitter and sent through the reference cell and the grid reactor to two nitrogen cooled HgCdTe pn detectors. The laser beam covers the full cross sectional area of the reactor and scans practically the whole reactor volume. Between the grid reactor and the detector a monochromator or interferometer can be used if necessary. The first one is to select just one laser mode out of the signal. The interferometer is used to check whether one or more modes are visible and to observe the wavelength interval of the mode. A digital oscilloscope was used to record and to verify the quality of the signals from both IR detectors and the photodiode. An external clock controlled the oscilloscope. The clock was set in such way that the memory of the oscilloscope was filled periodically with short blocks of data in high resolution. The oscilloscope saves the infrared radiation intensity data of the laser on the detector as a function of time in a file. This file comprises 60,000 data points. The function of the data analysis program is abstracting useful information from this file to produce the absorption of infrared light as function of time. The first 4 seconds of an experiment the grid was not heated, in order to record a laser scan without absorption peaks. Using ‘Labview’ this signal was subtracted from the signal with absorption peaks. Theoretically the result is a signal with the absorption peaks only (figure 3.28). In practice the transition of the laser from one period to the next adds a lot of noise to the signal. The frequency of the noise is of the same order of the signal. Filtering out the noise was therefore found to be not possible and the data was analysed manually from the ‘Labview’ output.

94

Figure 3.27 Total signal with absorption peaks recorded with the oscilloscope.

Figure 3.28 Subtracted signal with transition noise produced with the old data analysis method.

The total intensity of the detected laser light decreases as the grid is heated, as can be seen in figure 3.27. A light intensity decrease over the whole spectrum is caused by a thermal effect: the hot grid creates a local thermal gradient in the gas surrounding the grid and this gradient changes the breakingindex for light. The local hot gas spot could be compared to an optical lens that scatters the beam of the laser. The effect is strong just after the grid is turned on and decreases with time. The decrease of this effect could be caused by convection. [Bruinsma, 2001] has investigated the phenomenon with so-called “zero” experiments, experiments without a sample on the grid. The experiments are repeated during this project. Since the light intensity decreases over the total range of wavelengths, and the absorption is a proportion of the intensity, the absorption is not influenced. The pressure in the reactor influences not only the peak width by line broadening (also called collisional broadening, as it increases linearly with molecular collision rate and thus pressure), but also the maximum light intensity of the detected signal (also in figure 3.27). The pressure increases during an experiment as a result of the released gas and as a result of increasing temperature. This pressure effect is clearly seen if the reactor is slowly filled with gas, and the grid is not heated. This phenomenon causes a systematic error in the analysis with the ‘Labview’ based analysis programme, as it is based on peak height evaluation rather than peak surface area analysis. The new analysis program was developed in Matlab 6.0 to automate the data analysis and obtain higher accuracy. The first step is application of a non-causal moving average filter of 5 data points to smooth the signal. The developed tool makes use of the periodical behaviour of the signal. The oscilloscope registers a block of 500 data points a time. With a total memory of 60000 points and two absorption peaks per data block, in total 240 peaks are recorded every experiment. The 8-bit oscilloscope was set to a frequency of 100 kHz. This means that a block of 500 data points is filled in 5 milliseconds. The interval between two data blocks determines the total length of the experiment. In this project most experiments lasted for 6 seconds. Then the sampling interval is 50 milliseconds. It is assumed is that the height of an absorption peak at the central absorption wavelength is proportional to the concentration of the measured gas in the grid reactor. This is illustrated in figure 3.29. The central wavelength method is discussed below.

95

Light intensity (V)

a.

b.

g.

c. d.

h. e. f. Datapoints

Figure 3.29 Working principle of the developed data analysis tool. This is the filtered signal of two CO2 absorption peaks in a data block of approximately 500 data points.

Figure 3.30 Result of the data analysis tool: The absorption curve of a CO2 experiment with a curve fit of function (3.15).

The average value for the plateau between a. and b. in figure 3.29 is the zero light level, Ia,b. Point g. is the central wavelength peak height, Ig. A local polynomial is used to calculate a value for point h at the central wavelength, Ih. Point h should have been the amount of light detected if the light was not absorbed at this wavelength by the gas. The polynomial is fitted on data between measured points on the absorption curve c. and d. and between e. and f. A linear polynomial gave the best results. The absorption is defined as:

absorption =

I h - Ig

(3.14)

I h - Ia,b

where

I h = Light intensity at point h (V) Ig = Light intensity at point g (V) Ia,b = Average light intensity between point a and b (V) This routine is repeated every 500 data points. Only one of the two peaks in a data block is used since the interval between the two peaks in a data block (2,5 ms) is small compared to the interval between two blocks (50-100 ms). Using all 240 peaks instead of 120 will barely increase the accuracy. Figure 3.30 shows the result of the analysis tool: the absorption curve as a function of time. The tool is extended with a routine to make a non linear curve-fit on the absorption curve. Equation (3.3) is used in the curve fit of figure 3.30. Note that the equation has been derived from the well-known Arrhenius equation. t ⎛ c3 ⎜ Ia ( t ) = c1 + c 2 × 1 - e ⎜ ⎝

⎞ ⎟ ⎟ ⎠

where Ia = absorbed radiation fraction at a certain time c1 = constant

c2 = absorbed radiation fraction at end of experiment (final absorption) c3 = constant

96

(3.15)

The curve fit is used to gather a variety of process variables, like release time, maximum absorption. The Matlab code of the data analysis program can be found in [Slabbekoorn, 2002]. The program can be further improved by include a subroutine that determines the reaction rate, numerically or by computation of the derivative of the curve fit. There are a few disadvantages on using the central wavelength absorption. An absorption peak has a Gaussian shape. The height of the peak is determined by the concentration and temperature of the gas, the pressure in the reactor determines the width of the Gaussian distribution. Hence a better method would be to determine numerically the area between the polynomial c-d-e-f and the absorption peak in figure 3.29. This area is then used to determine the concentration of the gas with a computer program like Molspec of [Laser Photonics, 1992], or by an experimental calibration. This will increase the accuracy, especially at elevated pressures. With the central wavelength method this is not possible. A gas of a certain concentration will give at different pressures the same peak height at the central wavelength. Only the volume of the peaks differs. Therefore measurements with a reference gas of known concentration are necessary to calibrate the absorption peak height at a certain pressure. These measurements can be used to determine the gas concentration that belongs to other absorption peak heights, at the same pressure. Then, if the initial mass of the sample, the volume of the reactor, and the pressure and temperature in the reactor are known, it is possible to determine the weight percent of the specific gas that is released during pyrolysis of the sample. The end yield on a dry and ash free basis of gaseous components from the pyrolysed fuel samples was determined by the following relation, assuming ideal gas:

Yieldi =

PVreactor Mi yi

(

RTg mfuel,wet 1- mash,wet - m moisture,wet

)

(3.16)

The temperature of the gas is taken as room temperature, as confirmed by thermocouple measurements taken at a distance from the grid. The volume fractions of the gas are derived from the IR absorption values (see equation 3.2). Therefore, measurements with calibration gases were performed, see [Slabbekoorn, 2002].

97

Chapter 4 Experimental results 4.1 Choice of fuels, bed materials and additives 4.1.1

Fuel choice

This thesis aims at setting up a model for pressurised fluidised bed gasification of biomass and fossil fuels that includes the fate of nitrogen compounds and validating that model with experimental data. Therefore, a choice was made for application of the following solid fuels: Miscanthus Sinensis Giganteus pellets (further called: miscanthus); Wood pellets: Labee A-quality energy pellets; Brown coal: German, from the Rhenish Hambach open mine. Miscanthus was selected because it is a representative agricultural crop species for energy supply based on comparatively high yield energy crops. This perennial plant, which uses water very efficiently, was imported for the first time in Europe (Denmark) from Japan in 1935 [Gudenau & Hahn, 1993], [Lewandowski et al., 2000]. It is a so-called C4 plant, which refers to the pathway along which CO2 is absorbed in plants to form hexose (C6H12O6, see also figure 2.6). C4 plants have a relative growth advantage in a hot and humid environment under high sun illumination conditions, which accounts for their prevalence in the tropics. C3 plants, their counterpart, consume 18 ATP molecules per hexose molecule formed in the absence of photorespiration compared to 30 ATP for a C4 plant. This makes that C3 plants are more efficient at temperatures of less than 28 °C, and they predominate therefore in temperate environments [Stryer, 1988]. In the moderately mild climate of e.g. Germany, the yields can be 10 - 15 ton/ha from the third harvest onwards, as field tests have shown, see e.g. [Röhricht & Beier, 1998]. Yields even higher than 20 ton dry matter ha-1.year-1 have been reported [Lewandowski et al., 2000]. Taking into account the yields achieved with beetroot and silage maize in the Netherlands, 15 tonnes/ha appears to be a quite reasonable long-term expectation for miscanthus cultivation in Dutch climate and agricultural practices. In terms of yields and costs, miscanthus is a representative example of an energy crop producing dry ligno-cellulosic fuels [Siemons, 2002]. Miscanthus pellets manufactured from chopped plants (harvest April 1997) were purchased from Agromiscanthus (Ter Apel, The Netherlands). Nitrogen contents are comparable to straw, so that in this sense miscanthus can act as a model component for this agricultural residue. Figure 4.1 shows a picture of the harvest of miscanthus (April 1997) from a field in the northern province of Groningen in The Netherlands.

Figure 4.1 Harvest of miscanthus on a field in the Netherlands.

99

Wood pellets, also called “energy pellets – A quality”, which are sold commercially e.g. to Scandinavian countries mainly for domestic heating purpose by the Dutch company Labee (Moerdijk), were selected as woody species. This is a typical representative for clean wood, as can be seen in the relatively low total ash content and main ash composition, see table 4.1. Table 4.1 also shows that these pellets have a comparatively low nitrogen content, which covers the minimum fuel bound nitrogen release potential for our gasification tests. The main advantage of the use of pelletised biomass fuel is the high volumetric energy density, which makes transport costs and investment costs for fuel storage and process feeding smaller than in the case of non-pelletised fuel. In pelletised form the fuels show much less tendency to bridge in feeding systems compared to e.g. chips or dust, which makes complete automatic feeding feasible. The successful biomass feeding is one of the main points of attention in application of solid biofuels. An early study of [Reed & Bryant, 1978] showed that densification of wood to pellets at a scale of 300 tons/day by a Californian company requires about 7% of the energy (LHV basis) contained in the initial feedstock. Lower values were reported by [Van den Heuvel, 1995] for straw pelletisation: a value of 300 MJ/tonne was assumed for the process, which is approximately 1.5% of the LHV of the pellets. For pelletisation of switchgrass, [Samson & Duxbury, 2000] report a value of 108 MJ/tonne, which is approximately 0.6% of the switchgrass' LHV. As older fossil fuel representative, Rhenish brown coal was selected. It is a coal that is still relatively volatile, compared to older, black coals. The brown coal has been produced in the open mine “Hambach” in the Western part of Germany. It is obtained from RWE-Rheinbraun, where it is used for large-scale power (and heat) generation in the German Ruhr area. 4.1.2 Fuel composition and related chemical properties Table 4.1 shows the average composition of the fuels used. They have been analysed by IVD (University of Stuttgart, Germany), KTH (Stockholm, Sweden) and ECN (the Netherlands). The proximate analysis of the fuels was done by TGA (e.g. Leco TGA-500 analyser at IVD) and the elementary analyses of C, H, N and S were done using commercially available analysers (e.g. the Vario el, “Elementar”). Chlorine analysis was performed by combustion of the fuel sample in an oxygen bomb with subsequent potentiometric titration using an aqueous AgNO3 solution. The main ash components have been analysed by X-Ray Fluorescence Spectrometry using a Philips PW1480 at IVD. The heating value was determined using an IKA C4000 calorimeter at IVD. For brown coal the values presented are averages of 6 samples, for wood of 8 values and miscanthus 10. The indicated uncertainty range is the standard deviation based of these analyses. The main differences between the older brown coal fuel and the biomass samples can be seen in the proximate analysis (lower volatile and higher fixed carbon content for brown coal), the elemental analysis (higher carbon and lower oxygen content for brown coal), and in the heating value (as a result of the higher carbon and lower oxygen contents a higher value for brown coal). The biomass species applied differ significantly in N-content, and also in ash and Cl contents. Wood is the low N content species in this research work and it shows a relatively high scatter of the N content. The brown coal used, has a nitrogen content comparable to miscanthus. The composition of miscanthus is comparable to data presented by [Gudenau & Hahn, 1993], the composition of “high K miscanthus” reported by [Hallgren et al., 1999], data presented by [Lewandowski et al., 2000] and in the ECN Phyllis database [Phyllis, 2003]. The wood pellet composition is close to data that can be found in the ECN Phyllis database [Phyllis, 2003] as well as to data for pinewood given by [Kurkela et al., 1992 and 1993a]. The brown coal composition is comparable with values given by [Gudenau & Hahn, 1993] and [Kurkela et al., 1992,1993a and 1995].

100

Table 4.1 Proximate and ultimate analyses and heating values of the applied fuels, as received Brown coal (Hambach) Proximate analysis: Fixed carbon(mass%) Volatiles (mass%) Moisture (mass%) Ash (mass%) Ultimate analysis: C (mass%) O (mass%) H (mass%) N (mass%) S (mass%) Cl (mass%) Ash composition (mass% dry ash): SiO2 Al2O3 Fe2O3 CaO MgO Mn Na2O K2O P2O5 SO3 Heating value: HHV (MJ/kg) n.d.: not detectable (<0.03 mass%)

Crushed wood (Labee-A pellets)

Miscanthus giganteus (pellets)

37.4+1.52 45.8+1.92 12.6+3.42 4.2+0.24

16.5+0.97 74.9+0.84 8.4+0.45 0.16+0.08

15.5+3.56 73.5+3.18 8.5+1.39 2.5+0.54

56.4+2.14 33.1+2.32 5.3+0.47 0.56+0.06 0.34+0.18 0.03+0.01

47.0+0.46 46.1+0.55 6.5+0.34 0.15+0.10 0.10+0.01 n.d.

44.1+1.01 46.7+0.96 5.9+0.39 0.53+0.14 0.14+0.052 0.17+0.052

2.6+0.47 3.7+1.13 27.8+6.48 35.5+5.35 14.6+1.32 0.6+0.20 1.3+1.14 0.7+0.43 0.08+0.067 12.5+0.71

260* 99* 150* 990* 130* 110* 40* 340* 42* n.m.

34.5+2.40 2.3+0.40 1.9+0.44 6.0+0.92 4.3+0.35 0.2+0.022 1.3+0.17 33.9+1.52 5.9+0.69 5.7+0.38

22.1+1.05

18.6+0.11

17.7+0.30

*: element analysis on (mg/kg dry fuel basis of the elements, ECN-phyllis databank)

4.1.3 Physical property characterisation of fuels and bed materials The gasification experiments comprised the use of several different solid materials. As mentioned, two types of biomass were applied and one older fuel: brown coal. As bed material sand was chosen, because this is a cheap material and it can be easily fluidised. As additive to the biomass fuels for the purpose of tar cracking and sulphur capture, dolomite (CaCO3.MgCO3) was chosen. In brown coal already a comparatively high content of Ca is present that can capture sulphur and tar formation for this fuel is much less a problem. Therefore no additive was considered for this fuel. Dolomite was also selected as additive to the biomass fuels in order to study the influence of the Ca inventory in the reactor on the yield of NH3 and HCN. As a natural clay additive known for alkali metal capture (to prevent sintering), Minphyl (mainly consisting of Pyrophyllite, Al2Si4O10(OH)2 and SiO2) was selected. It was mined in South-West Europe. As additive it was applied for miscanthus, a fuel with relatively high K content in the ash. Minphyl also represents a non-Ca containing additive, so that the influence on fuel nitrogen portioning can be investigated. Table 4.2 presents the particle size distribution (PSD), average diameter and density of sand and additives. Figure 4.2 shows the data in graphical form. The sand used shows a single maximum distribution, as it has been sieved to this specific diameter window. The dolomite shows a distribution with more maxima, as the sieving of the raw material as delivered was not in a very small diameter window. The Minphyl material was obtained as a relatively fine material, corresponding to finer fractions of the dolomite used. Bed materials and additives have comparable densities.

101

Table 4.3 shows the PSD of the fuels applied, with a graphical presentation in figure 4.3. All particle size distributions were determined by sieving a sample of the material using a Sonic Shifter with standardised sieves. The two biomass fuels show quite corresponding PSD’s, although miscanthus has a large fraction (about 20 mass%) having a diameter > 5.6 mm. This is attributed to the more difficult grindability as it is relatively fibrous material. The brown coal was delivered with a broad fuel size distribution, less strictly sieved as necessary for pulverised fuel firing for which smaller particles sizes are needed. The particle density was determined by simple picnometry, using demineralised water as liquid. The densities of the core fuel materials are quite comparable. Table 4.2 Particle size distributions and particle densities of bed material and additive. dp range (µm) 0 53 53 90 90 - 125 125 - 180 180 - 250 250 - 300 300 - 425 425 - 500 500 - 600 600 - 710 710 - 850 850 - 1200 1200 - 2360 2360 - 3150 3150 - 4000 4000 - 5000 5000 + dp, average (µm) dp,SMD *** average (µm) ρparticle (g.cm-3) *

> 200 µm

**

Sand (Mass %) 0.04 0.0 0.0 0.0 0.02 0.02 1.44 19.15 73.41 5.08 0.84 0.0 0.0 0.0 0.0 0.0 0.0 537 528 2.63

dp average (µm) 26.5 71.5 107.5 152.5 215 275 362.5 462.5 550 655 780 1025 1780 2755 3575 4500 5500

as indicated by the producer of the material

***

Dolomite (Mass %) 10.71 16.23 19.70 11.73 11.21 3.72 8.51 4.96 4.26 8.97 0.0 0.0 0.0 0.0 0.0 0.0 0.0 224 99 2.77

Minphyl (Mass %) **

< 18.0 * < 5.0 < 0.5 < 0.1

2.8

Sauter Mean Diameter

80 Sand Dolomite

70

Mass Percentage [%]

60

50

40

30

20

10

0-53

53-90

90-125

125-180

180-250

250-300

300-425

425-500

500-600

600-710

710-850

850-1200

1200-2360

2360-3150

3150-4000

4000-5000

>5000

0

Particle size [µm]

Figure 4.2 Size distribution of bed material and additive used in the Delft PFBG tests. The main bed material, sand, belongs to group B according to the classification of [Geldart, 1973], although its classification is near group D. Sand-like particles belong to group B within the particle

102

diameter range of approximately 40 – 500 µm and densities of 1.4-4 g/cm3, whereas for group D an indication is given that particles sizes are larger than 600 µm. In beds belonging to the B group, bubbles form as soon as the gas velocity exceeds the minimum fluidisation velocity, umf. At higher gas velocities, the bed behaves as follows [Kunii & Levenspiel, 1991]: • small bubbles form at the distributor and grow and coalesce as they rise through the bed; • the bubble size increases roughly linearly with distance above the distributor and excess gas velocity, u0 - umf; • the bubble size is roughly independent of mean particle size; • vigorous bubbling enhances the circulation of solids. Table 4.3 Fuel particle size distributions and particle densities for the Delft PFBG tests. dp range (µm) 0 53 53 90 90 - 125 125 - 180 180 - 250 250 - 300 300 - 425 425 - 500 500 - 600 600 - 710 710 - 850 850 - 1400 1400 - 2000 2000 - 3150 3150 - 4000 4000 - 4750 4750 - 5600 5600 + dp,average (µm) dp,SMD * (µm) ρparticle (g.cm-3) ρbulk (g.cm-3) n.m.: not measured.

Miscanthus (Mass %) 0.25 0.22 0.27 0.61 0.72 0.81 1.41 1.05 0.95 1.66 2.04 7.47 7.04 15.60 16.38 13.72 10.54 19.26 3544 1470 1.30 n.m. * Sauter Mean Diameter

Wood (Mass %) 0.11 0.33 0.64 1.38 1.61 0.63 3.04 1.60 2.49 1.72 2.51 8.24 7.84 18.18 18.54 19.59 9.96 1.61 2897 1155 1.49 0.36

dp average (µm) 26.5 71.5 107.5 152.5 215.0 275.0 362.5 462.5 550.0 655.0 780.0 1125 1700 2575 3575 4375 5175 6025

Brown Coal (Mass %) 7.93 3.02 7.33 5.38 6.72 3.96 12.57 6.92 8.55 8.05 8.32 18.43 1.93 0.35 0.29 0.14 0.08 0.02 559 168 1.47 0.67

25 crushed Labee wood pellets crushed Miscanthus Giganteus pellets Hambach Brown Coal

Mass Percentage [%]

20

15

10

5

< 53

53 - 90

90 - 125

125 - 180

180 - 250

250 - 300

300 - 425

425 - 500

500 - 600

600 - 710

710 - 850

850 - 1400

1400 - 2000

2000 - 3150

3150 - 4000

4000 - 4750

4750 - 5600

> 5600

0

Particle size [µ m]

Figure 4.3 Size distribution of fuels used in the Delft PFBG tests. 103

4.2 Experimental results of PFBG gasification tests 4.2.1 Experimental data representation and definitions of relevant parameters Before reporting the details of the operating data, in this paragraph some definitions will be given for process parameters that characterise the gasification conditions. These parameters will be used in tables and graphs throughout this work. The air stoichiometry ratio, λ, (in some literature also called “equivalence ratio” or “air factor”) is defined by equations (4.1): ⎡ Φ m , a ir ⎤ ⎢ ⎥ ⎢⎣ Φ m , fu e l (d a f) ⎥⎦ A c tu a l λ= ⎡ ⎤ Φ m , a ir ⎢ ⎥ ⎢⎣ Φ m , fu e l (d a f) ⎥⎦ S to ic h

(4.1)

where ⎡ ⎤ Φ m,air ⎢ ⎥ ⎢ ⎥ Φ ⎢⎣ m, fuel (daf) ⎥⎦ Stoich

⎧⎛ ⎫ ⎞ ⎛m ⎞ ⎞ ⎛m ⎞ ⎛m ⎪ m C,daf ⎪ H,daf N,daf S,daf MW O ⎟ + ⎜ MW O ⎟ + ⎜ MW O ⎟ + ⎜ MW O ⎟ -m O,daf ⎬ ⎨⎜ 2 ⎟ ⎜ 4 MW 2 ⎟ ⎜ 2 MW 2 ⎟ ⎜ MW 2⎟ ⎪⎜⎝ MW C ⎪⎭ H N ⎠ ⎝ ⎠ ⎝ S ⎠ ⎝ ⎠ =⎩ m O ,air 2

Another important parameter is the carbon conversion, which can be calculated in two ways, based on solids catch and gas+tar yield, respectively: ⎡ ⎤ m Φ (4.2a) C C s o lid s = ⎢1 - C , s o lid s o u t m , s o lid s o u t ⎥ × 1 0 0 % m C , fu e l (d a f) Φ m , fu e l(d a f) ⎥⎦ ⎢⎣ ⎡ m C , g a s + ta r Φ m , g a s + ta r ⎤ ⎥ × 100% ⎢⎣ m C , fu e l (d a f) Φ m , fu e l(d a f) ⎥⎦

C C g a s /ta r = ⎢

(4.2b)

The difference between the two calculated conversions is used, next to main elemental balances and energy balance to judge if the measurements done during a specific experiment are successful in terms of mass/energy balance closure. The cold gas efficiency applied in this work is based on the higher heating value (HHV) of the producer gas and dry ash free fuel, respectively and is defined as: ⎡

⎤

H H V g as Φ m , g as ⎥ × 100% η co ld g as = ⎢ ⎢ ⎥ H H V Φ fu el (d af) m , fu e l (d af) ⎥⎦ ⎢⎣

(4.3)

The HHV is calculated according to the iso-6976 norm [iso, 1995]. For the calculation, the heating values of CO, H2, CH4, C2H2, C2H4 and C2H6 are taken into account. Also, the gas density is calculated according to this norm, taking into account the species mentioned above. The element balances over the gasifier and filtration unit are calculated according to:

⎛ ∑ mijΦ m, j ⎞ ⎜ ⎟ Element closure, element i = ⎜1 − out (4.4) ⎟ × 100% m Φ ⎜ ∑ ij m, j ⎟ in ⎝ ⎠ in which mij is the mass fraction of element I in molecule type j and Φm,j is the mass flow of molecule type j. 104

4.2.2. Background information on the measurement campaigns Gasification experiments using biomass and coal as fuels have been performed from 1996 on. The first experiments were carried out to investigate the operating envelope of the test rig (see [de Jong et al., 1998]). The PFBG installation was first operated with one hot cyclone, but without a ceramic filter unit. The operation of the pressurised combustor downstream of the cyclone, though, appeared to be difficult due to the relatively high dust loads. In 1997 the hot gas cleaning filter unit was installed and the hot cyclone removed. All experiments from 1998 on have been performed using this gas cleanup configuration. The test rig schematic and further details on the installation are given in §3.2.1. The miscanthus gasification experiments described in table 4.4a have been carried out using the first set of three ceramic filters. This set of filters showed visually observable cracks due to the fact that between the experiments 981007_2 and 981019 a probe situated in the LCV product gas combustor had been removed when the filter elements were still too hot. As air was leaking into the test rig, spontaneous ignition and combustion took place within the filter blocks. Due to the heat released, which in this case was not removed by forced gas flow, the filters cracked [Andries et al., 2001]. A second set of the same type of ceramic filters was then installed and used for the miscanthus gasification experiments 981214 – 990126, shown in table 4.4b. Due to an interruption of the fuel feed caused by a malfunctioning of the level indicator for biomass feeding in a following experiment, a fast transition from gasification to combustion conditions in the gasifier took place, which resulted in burning of the carbonaceous dust-cake and cracking of the second set of filters [Andries et al., 2001]. A third set of these filters was installed and used in the miscanthus gasification experiments 020429-020513 and in the experiments using wood (see table 4.5) and brown coal (table 4.6). During the wood gasification experiments without addition of steam, filter blocking took place within approximately 8 hours of operation, probably due to soot formation in the dust cake and filter pores. This is comparable to the ceramic candle filter plugging described by [Kurkela et al., 1993a] for similar filtration temperatures. The filters could be regenerated, though, during the subsequent gasifier start up, operating near the stoichiometric point and using clean wood pellets as fuel with steam and air as oxidizer [de Jong et al., 2002]. The use of steam in the subsequent wood gasification tests prevented filter plugging. Brown coal gasification, without steam addition, did not lead to significant pressure drop increases. This can be attributed to the type and amount of tar formed in this processes, which is different from that formed during wood gasification. More information on hot gas filtration, ash and trace metal behaviour is given by [Ünal, 2005]. The pressure during the experiments was varied between 0.35 MPa and 0.7 MPa. Lower pressures could not be applied due to process control restrictions. Due to material strength constraints no pressures higher than approximately 1 MPa could be applied. Higher pressures than those reported here were not applied, because the fuel feeding capacity was limited in such a way that relevant gasification conditions could not be achieved anymore. Care was taken to keep the fluidised bed temperatures below approximately 830 °C during miscanthus gasification. Mainly for this purpose steam was added, because the fuel has a potential risk of sintering near and above this temperature. Also, in one test alkali getter material was used to prevent sintering. As an example, [Gudenau & Hahn, 1993] reported a sintering point of 820 °C for this fuel in their 100 kWth ACFB gasification test rig. [Hallgren et al., 1999] performed tests of miscanthus with a relatively high K content in a 20 kWth BFB gasifier. They observed sintering tendencies at 800 °C with Olivine sand as bed material and at 850 °C with dolomite as bed material. [Roll, 1994] presented a weakening point for the ash of this fuel of 900 °C, indicating that the mineral matter in the fuel shows sintering tendencies in the temperature window where fluidised beds are usually operated. [Kurkela et al., 1996] indicated that for straw, which is comparable with respect to ash composition and therefore sintering risk, no sintering occurred, except for one straw type in one experiment, was observed for temperatures in the freeboard in the range of 790 – 830 °C. They observed, however, severe sintering problems when raising the temperature to values above 850 °C. 105

The product gas heating value is somewhat at the low side when steam is being used. This was the result of the bed temperature limitation for miscanthus to ensure long-term operation without agglomeration problems. For the other fuels there was no need to keep this bed temperature limit due to their different ash composition with much lower agglomeration risks. The fluidisation velocity based on inlet gas flows has been kept at 0.5 m/s, meaning that the actual fluidisation velocity based on product gas is in the range of 0.6-0.8 m/s. Under these conditions, the bed is in bubbling mode, cf. [Kunii & Levenspiel, 1991]. The minimum fluidisation velocity, umf, for the sand has been determined in an atmospherically operated small quartz fluidised bed and has been found to be 0.16+0.04 m/s, see [Bos, 1998] and [Schot & Laro, 2000]. Figure 4.4 depicts the related pressure drop behaviour versus the superficial velocity from which umf has been determined for the sand that has been used in the PFBG tests. The ratio between the actual and the minimum fluidisation velocity (u0/umf) is in the range of 5 – 6, satisfying the criterion of u0/umf > 2.5, which ensures good solids mixing in the bed [Maniatis et al., 1988].

umf

Figure 4.4 Pressure drop versus superficial velocity for sand applied in TUD PFBG experiments.

4.2.3 Miscanthus gasification The majority of the gasification tests were carried out using miscanthus as fuel. Table 4.4a and b present the results of the miscanthus gasification experiments performed with the Delft PFBG test installation. The data show the composition of the product gas downstream of the ceramic filter unit, because this was the standard analysis position during all experiments performed. Figure 4.5 and 4.6 show the on-line gas analysis results during a typical measurement day for the main gas phase concentrations and some reactor temperatures, respectively. An experiment consists of three phases: start-up in combustion mode, gasification and shutdown (see also §3.2.1 for details). During start-up of miscanthus gasification experiments, the procedure consists of an initial natural gas combustion period, followed by a transition to clean wood combustion, as there is a risk for sintering using miscanthus under high temperature combustion conditions. This phase is continued until the temperatures in the downstream gas combustor are high enough to ensure fast ignition of LCV gas. Before the transition to gasification is made by an increase of the fuel flow, miscanthus is introduced into the feeding system. The transition to gasification is characterised by a steep increase of CO and H2 concentrations and a decrease in the CO2 concentration. This is accompanied by a sharp temperature decrease due to endothermic heterogeneous gasification reactions and also due to physical cooling due to increased fuel flow. The shutdown is accomplished by a simultaneous stop of the fuel and gasification medium feeding and ‘blanketing’ by N2. A sharp temperature decline is the result.

106

Figure 4.5 A typical measurement day: main gas concentrations for miscanthus gasification experiment number 020429.

Figure 4.6 A typical measurement day: gasifier temperatures for miscanthus gasification experiment number 020429

Figure 4.7 shows a slightly increasing trend of the CO and H2 concentrations with decreasing air stoichiometry, except for two experiments with higher concentrations, where no steam was added. Increased oxygen availability at the higher λ values leads to increased formation of CO2 from CO and H2O from H2. The values measured are comparable to those presented by [Kurkela et al., 1996] for PFB straw gasification, although these authors used higher steam/air ratios. 107

Figure 4.8 presents an increase of the concentrations of methane, representative for light hydrocarbon species, with decreasing air stoichiometry. These results agree with the negative correlation coefficients for the volume percentages of CH4 with applied air stoichiometry found experimentally during atmospheric bubbling FB gasification of biomass (pine sawdust) by [Narvaéz et al, 1996]. Figure 4.9 shows the yields of main carbon containing LCV gas components expressed as the fuel bound carbon conversion into these species. This shows the fuel conversion behaviour better than the resulting gas phase concentrations only. The increased conversion of C into CO2 with increasing air stoichiometry is due to increased availability of reactive oxygen, leading to more complete oxidation. For CO the trend is less clear in the range measured. The light hydrocarbons CH4 and C2H4 are characterised by a relative constancy as compared to the behaviour of CO and CO2. The background of this behaviour can be explained by their formation from initial fast fuel devolatilisation and subsequent tar cracking accompanied with their relative stability with respect to further reactions with oxygen and steam for example. The values indicated in figure 4.9 show a quite good agreement with a range of conversion values regarding all carbon containing components presented by [Padban, 2000] for pressurised BFB gasification of biomass (sawdust), performed at 1.2 MPa and at a 90 kWth scale. Figure 4.10 depicts the increasing trend of the higher heating value of the produced gas with decreasing air stoichiometry applied, which is a result of the increasing CO, H2, CH4 and other hydrocarbon combustible concentrations, as indicated above. This trend is in-line with those reported for pressurised bubbling fluidised bed gasification of biomass by e.g. [Kurkela et al., 1996] and [Padban, 2000]. Also, for atmospheric CFB gasification such a relation has been observed [Van der Drift et al., 2001]. The higher heating value of the gas was sufficient for stable pressurised combustion in the downscaled ALSTOM Typhoon gas turbine combustor, see [Hoppesteyn, 1999]. In figure 4.11 the carbon conversion based on solid catch and the cold gas efficiency (on HHV basis) are depicted. Carbon conversion was lowest for experiments where low air stoichiometry values were applied. Due to lower oxygen availability at lower λ values, less heat is generated in the heterogeneous char oxidation and volatiles combustion. Also endothermic reactions begin to play a more important role, leading to a further decrease in temperature. A similar trend was observed for PFB gasification of straw, a fuel quite comparable to miscanthus, by [Kurkela et al., 1996]. The carbon conversion values they measured were slightly higher than our values. This can be attributed to differences in the fuel composition and related char reactivity concerning the main heterogeneous gasification reactions. Also, the particle size distribution range of the fuel was somewhat higher than the values reported by [Kurkela et al., 1996], see table 2.4. The influence of this PSD is mainly on the extent of the (partial) combustion of the char particles in such a sense that lower conversion values are expected, cf. [Kersten, 2002]. Also differences in heat losses could contribute to differences observed in the carbon conversion. Most experiments were performed at somewhat higher λ values, where carbon conversions of approximately 90% could be attained. The cold gas efficiency varied between 43 and 61%. There is not a clear trend, although it appears that there is an optimum range for λ values between 0.4-0.45. This can be explained by the fact that although the heating value of the product gas increases with decreasing λ, the carbon conversion also decreases. The values are relatively low compared to large-scale installations, due to the comparatively higher heat losses, but the values are quite comparable to other test rigs. Tar is a major problematic group of organic compounds in the gas produced by biomass gasification. It contributes to fouling of equipment (e.g. gas engine, turbines and gas analysis transfer lines) and to emissions in gas cleaning (when water scrubbers are used) and combustion processes (CO, soot), see e.g. [Neeft, 2000] and [Milne et al., 1998]. Tars have been defined as organic aromatic species with a molecular weight higher than benzene, see e.g. [Simell et al., 2000]. Figure 4.12 shows an increase of the specific tar concentrations of polyaromatic hydrocarbons (PAH) and phenols in the produced LCV gas for decreasing air stoichiometry values. These compounds were quantified by means of the novel solid phase adsorption (SPA) technique, developed at KTH Sweden, see [Brage et al., 1997]. Especially the contribution of phenols appears to be important at lower λ values, accompanied with temperatures lower than 800°C. For gasification using steam as (co-)gasifying medium the 108

contribution of phenols is reported to be significant by [Milne et al., 1998]. Reproducible measurement of benzene, toluene and xylenes with the SPA sampling technique were not possible in this work, probably due to evaporation of these species during the time between sampling, sample sending and subsequent elution and GC analysis. The PAH’s analysed range from indene, naphtalene to pyrene, with naphtalene being the major species. The measured values are typical for pressurised bubbling fluidised bed gasification when compared to [Sjöström et al., 1998], [Kurkela et al., 1993], [Padban, 2000] and [Evans et al., 1985] who performed FB experiments at various elevated pressures. Figures 4.13 and 4.14 present the experimental results with respect to formation of ammonia and hydrogen cyanide and the conversion of fuel-bound nitrogen into these nitrogen components. No gaseous HNCO was detected in our FTIR analyses of the producer gas. These species are known precursors of NOx under e.g. gas turbine combustion conditions, which is a problem when dry, high temperature gas cleaning is applied, see e.g. [Hoppesteyn, 1999]. From the results it can be concluded that a major part of the fuel-bound nitrogen is converted to ammonia. The NH3 concentrations measured are in-range with values observed by [Wang & Olofsson, 2002] and [Padban, 2000] for pressurised gasification in the 90 kWth Lund BFB gasifier of a variety of biomass and waste fuels with an N-content of ca. 0.15 – 1.5% (db). The fuel-bound nitrogen to NH3 conversions are also comparable to values reported by VTT, where a slightly smaller scale pressurised fluidised bed is operated (ca. 500 kWth), for experiments with straw. This fuel is in many respects comparable to miscanthus, see table 2.5 [Kurkela et al., 1996]. During gasification of straw with dolomite as additive, they observed values in the range of 60 - 71% fuel-nitrogen conversion to NH3, at air stoichiometry values between 0.28 and 0.31. The use of dolomite is a factor reported to promote the formation of NH3 by catalysis of HCN conversion to NH3, see e.g. [Berg et al., 2001]. [Chen, 1998] observed significantly lower values of fuel-nitrogen conversion to NH3, where wood (and coal) was gasified in a pressurised fluidised bed reactor equipped with top feeding. [Vriesman et al., 2000] performed experiments with miscanthus as fuel in a small scale atmospheric bubbling fluidised bed and observed a difference in top feeding as compared to bottom feeding. Top feeding led to a significantly lower conversion fuel-nitrogen to NH3. Differences were attributed to the environment in which the initial fuel flash pyrolysis takes place, i.e. reducing for top feeding versus oxidizing for bottom feeding. Relatively high HCN values were observed when miscanthus was gasified with the addition of Minphyl instead of dolomite (exp. 020513 versus 020429, see also figure 4.15) under otherwise the same process conditions. This can be attributed to the effect of the Ca inventory on the enhanced conversion of HCN into NH3, because Minphyl does not contain any Ca. In figure 4.15 an overview is given of the measurements at three axial positions in the freeboard of the gasifier, at 3.5, 4.5 and 5.5 m, measured from the bottom plate. The concentration profile of acetylene, C2H2, is one of the more interesting trends. For the two experiments without steam addition, 020429 and 020513, the highest C2H2 concentrations can be seen. The difference in these experiments was the additive used: dolomite versus Minphyl. For the experiment with Minphyl addition, we see the highest C2H2 concentration resulting. With addition of steam, values of about a factor 3 lower are observed and a slightly decreasing profile compared to the abovementioned two experiments. This effect is much larger than the gas dilution by water, as the water content in the gasification experiments varies only in a limited range, see figure 4.15. C2H2 plays a role in the build up of tar and soot, via the well known (C2+C4) route (reaction of C2H2 with 1-3 butadiene[-derivative]) or through a C3 channel route (via formation of C3H3) as pointed out by e.g. [Dagaut & Cathonnet, 1998]. Figure 4.16 and 4.17 present the results of radial measurements at probe position P1.1 (3.5m from the bottom plate, see figure 3.1) in the freeboard at two pressures and comparable air stoichiometry values. The radial position 0 cm corresponds to the centreline. From these figures it can be seen that a significant radial concentration profile for main and minor species is not present under the conditions prevailing in this pressurised BFB. This in contrast to measurements in an atmospheric CFBG test rig performed at ECN, the Netherlands, where a distinct radial profile was measured [Kersten, 2002]. This is attributed to the different velocity profiles leading to comparatively high radial Péclet numbers in circulating beds, characteristic for worse radial mixing. With respect to these analysis results, a plug flow regime assumption would be adequate. Therefore, extensive measurement of radial profiles was not repeated for subsequent experiments. 109

110

CO H2 CH4 C2H4 C2H6 C3H6 CO2 H2O N2 Ar C2H2 H2S COS NH3 HCN

[vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [ppmv] [ppmv] [ppmv] [ppmv] [ppmv] [mg/ mn3,dry] [mg/ mn3,dry] [mg/ mn3,dry]

[-]

[MPa] [K] [K] [K] [kg/h] [kg/h] [-] [-] [kg/h] [kg/h] [kg/h] [kg]

6.19 5.84 3.33 0.53 0.31 0.05 14.73 22.77 45.51 0.46 3) 59 n.m. 3 2683 74 0.7x103 1.2x103 3.9x103

0.32

0.50 1009 997 904 188 303.0 0.088 0.036 36.8 0.5 1.9 145

980708

110

[MJ/ mn3] 3.44 Higher Heating Value (wet gas) [%] 91.7 Carbon conversion(solids basis) [%] 48.6 Cold gas efficiency (HHV basis) [%] +10 C-balance closure gasifier + filter [%] -8 O-balance closure gasifier + filter [%] 0 N-balance closure gasifier + filter [%] -15 H-balance closure gasifier + filter [%] +32 Ash-balance closure gasifier + filter [%] 79.1 Fuel_N to NH3 conversion [%] 2.2 Fuel_N to HCN conversion 1) determined by spa 2) MinPhyl used instead of dolomite 3) Calculated from Ar balance

BTX (dry basis) 1) PAH’s (dry basis) 1) Phenols (dry basis) 1)

Composition LCV gas after filter (wet basis)

λ gasifier

Pressure (top of gasifier) Bed temperature Freeboard temperature Temperature behind the filter unit Mass flow fuel Mass flow gasification air Steam:air ratio (mass basis) 5) Dolomite:fuel ratio (mass basis) N2 flow fuel feeding N2 purge flow probes N2 purge flow ceramic filter Bed mass

Experiment

5.04 5.72 2.38 0.39 0.13 0.01 14.78 16.66 54.16 0.57 3) 127 n.m. 3 1514 40 0.1x103 0.5x103 0.3x103

0.45

0.50 1082 991 895 135.4 306.4 0.087 0.036 26.7 1.9 1.9 162

981002_2

5.82 6.21 3.38 0.52 0.29 0.11 13.78 20.30 48.91 0.50 3) 237 n.m. 3 1676 72 0.2x103 0.4x103 2.2 x103

0.31

0.40 978 946 861 163.7 253.9 0.079 0.036 26.6 1.9 2.9 138

981007_1

5.29 6.16 3.23 0.36 0.27 0.01 14.46 20.21 49.32 0.49 3) 37 n.m. 6 1987 37 0.3x103 0.6x103 0.3x103

0.38

0.70 1081 1049 955 203.9 392.8 0.11 0.035 56.9 1.9 2.9 138

981007_2

2.66 3.48 3.16 87.6 82.6 84.4 42.8 43.8 50.8 +6 +14 0 +10 +9 -1 0 0 0 +9 +1 -8 +7 +36 +40 50.9 44.0 66.7 1.4 1.9 1.3 5) steam @ saturation temperature; air @ ambient temperature

5.49 6.15 3.40 0.51 0.28 0.09 13.95 18.68 50.72 0.51 3) 107 n.m. 4 2067 79 0.7x103 0.6x103 2.2x103

0.31

0.50 995 964 871 195 308.7 0.086 0.037 40.7 1.9 2.9 162

981002_1

3.58 3.42 82.4 74.9 43.3 43.5 +11 +6 -8 +12 0 0 -12 +7 +29 +26 65.0 54.8 1.9 2.1 4) Calculated by difference

6.21 5.85 3.57 0.57 0.32 0.07 14.59 25.47 42.65 0.43 3) 69 n.m. 8 2579 73 1.5x103 1.5x103 5.7x103

0.26

0.50 995 996 905 240 309.7 0.086 0.031 36.8 0.5 3.2 161

980715

Table 4.4a Experimental data for miscanthus gasification applying the Delft PFBG test rig.

2.99 90.2 48.6 +6 +1 0 -13 +26 64.4 0.9 n.m.: not measured

5.53 6.18 2.79 0.45 0.16 0.01 14.78 16.69 52.68 0.55 3) 60 n.m. 6 1898 27 n.m. n.m. n.m.

0.44

0.50 1050 1021 917 142.4 313.6 0.038 0.049 29.9 1.9 1.4 209

981019

111

CO H2 CH4 C2H4 C2H6 C3H6 CO2 H2O N2 4) Ar C2H2 H2S COS NH3 HCN

[vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [ppmv] [ppmv] [ppmv] [ppmv] [ppmv] [mg/ mn3,dry] [mg/ mn3,dry] [mg/ mn3,dry]

[-]

[MPa] [K] [K] [K] [kg/h] [kg/h] [-] [-] [kg/h] [kg/h] [kg/h] [kg]

5.70 5.33 2.71 0.51 0.18 0.037 14.71 16.43 53.67 0.58 3) 71 n.m. 3 1712 38 n.m. n.m. n.m.

0.44

0.40 1011 958 859 107.6 240.6 0.094 0.036 15.4 1.9 0.96 155

981214

5.07 7.11 3.40 0.33 0.26 0.005 15.69 19.82 47.66 0.48 17 n.m. 5 1681 17 0.9x103 2.3x103 0.3x103

0.38

0.70 1065 1034 965 199.7 391.9 0.11 0.053 50.4 1.9 1.3 159

981223

4.63 5.76 2.89 0.29 0.20 0.002 14.60 20.51 50.42 0.50 26 n.m. 6 1969 21 0.5x103 2.6x103 0.1x103

0.45

0.70 1082 1041 969 167.6 384 0.11 0.053 46.5 1.9 1.3 149

4.79 5.36 2.61 0.44 0.12 0.008 15.20 18.70 52.02 0.54 91 n.m. 2 1948 35 0.5x103 3.2x103 0.2x103

0.45

0.40 1020 993 905 103 238.2 0.094 0.052 15.4 1.9 0.80 149

990107_2

5.04 6.41 3.13 0.29 0.21 0.0003 14.86 21.22 48.14 0.53 21 n.m. 4 1615 21 n.m. n.m. n.m.

0.38

0.70 1062 1024 978 201.4 392.3 0.11 0.053 46.3 1.9 1.3 160

990121

5.43 7.04 3.35 0.34 0.28 0.004 15.63 18.88 48.37 0.50 3) 20 n.m. 5 1741 18 0.7x103 2.5x103 0.1 x103

0.43

0.70 1071 1060 963 174.4 388.4 0.11 0.056 44.4 1.9 1.3 149

990126

9.09 6.64 3.57 0.59 0.13 0.004 15.40 14.36 49.41 0.55 3) 121 145 8 2191 47 0.2 x103 3.7x103 0.1 x103

0.33

0.35 1084 1108 952 127.2 226.9 0 0.033 16.0 1.9 1.0 147

020429

2.69 3.03 3.33 3.89 90.3 90.5 90.7 89.7 44.9 48.8 60.7 51.1 +4 +8 -10 +7 +6 -2 -5 -3 0 0 0 +14 +2 -10 -17 -2 -24 +6 +15 +11 67.4 54.0 65.9 97.0 1.3 0.7 0.7 2.1 5) steam @ saturation temperature; air @ ambient temperature

990107_1

3.29 2.78 87.4 90.4 54.6 50.5 -3 0 -5 0 0 -14 -11 +18 +6 57.8 74.1 0.6 1.7 4) Calculated by difference

4.56 4.19 2.64 0.39 0.14 0.006 15.14 19.01 53.17 0.56 65 n.m. 2 1771 32 0.8x103 1.7x103 1.8x103

0.40

0.50 1024 1018 922 144.6 293.3 0.10 0.050 25.7 1.9 0.96 177

981216

[MJ/ mn3] 2.93 2.51 Higher Heating Value (wet gas) [%] 88.3 91.6 Carbon conversion(solids basis) [%] 45.8 36.8 Cold gas efficiency (HHV basis) [%] +7 +17 C-balance closure gasifier + filter [%] +14 +12 O-balance closure gasifier + filter [%] 0 0 N-balance closure gasifier + filter [%] +13 +14 H-balance closure gasifier + filter [%] -2 +44 Ash-balance closure gasifier + filter [%] 55.5 54.0 Fuel_N to NH3 [%] 1.3 1.0 Fuel_N to HCN 1) determined by spa 2) MinPhyl used instead of dolomite 3) Calculated from Ar balance

BTX (dry basis) 1) PAH’s (dry basis) 1) Phenols (dry basis) 1)

Composition LCV gas after filter (wet basis)

λ gasifier

Pressure (top of gasifier) Bed temperature Freeboard temperature Temperature behind the filter unit Mass flow fuel Mass flow gasification air Steam:air ratio (mass basis) 5) Dolomite:fuel ratio (mass basis) N2 flow fuel feeding N2 purge flow probes N2 purge flow ceramic filter Bed mass

Experiment

Table 4.4b Experimental data for miscanthus gasification applying the Delft PFBG test rig.

3.46 92.3 43.5 +13 -1 +8 -2 +34 51.1 4.2 n.m.: not measured

8.96 5.15 3.29 0.55 n.m. n.m. 14.97 15.71 50.62 0.55 3) 224 154 9 1636 131 0.4x103 6.4x103 0.04x103

0.33

0.35 1103 1129 952 130 222.7 0 0.0282) 18.6 1.9 1.0 150

020513

111

[CO],[H2] (vol.%,wet, purge N2 corrected)

12

10

[CO] @ 0.35 MPa [CO] @ 0.4 MPa [CO] @ 0.5 MPa

8

[CO] @ 0.7 MPa [H2] @ 0.35 MPa [H2] @ 0.4 MPa

6

[H2] @ 0.5 MPa [H2] @ 0.7 MPa

4

2

0 0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

λ (−)

Figure 4.7 CO and H2 concentrations in the Delft PFBG miscanthus gasification tests. [CH4] (vol.%,wet,N2 purge corrected)

5.0 4.5

[CH4] @ 0.35 MPa [CH4] @ 0.4 MPa

4.0

[CH4] @ 0.5 MPa [CH4] @ 0.7 MPa

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0.60

λ (−)

Figure 4.8 CH4 concentrations in the Delft PFBG miscanthus gasification tests.

Fuel_C into gaseous species (%)

70

%Fuel_C to CO @0.7MPa %Fuel_C to CO @0.5MPa %Fuel_C to CO @0.4MPa %Fuel_C to CO @0.35MPa %Fuel_C to CO2 @0.7MPa %Fuel_C to CO2 @0.5MPa %Fuel_C to CO2 @0.4MPa %Fuel_C to CO2 @0.35MPa %Fuel_C to CH4 @0.7MPa %Fuel_C to CH4 @0.5MPa %Fuel_C to CH4 @0.4MPa %Fuel_C to CH4 @0.35MPa %Fuel_C to C2H4 @0.7MPa %Fuel_C to C2H4 @0.5MPa %Fuel_C to C2H4 @0.4MPa %Fuel_C to C2H4 @0.35MPa

60 50 40 30 20 10 0 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (−)

Figure 4.9 Fuel bound C conversion into gas for the Delft PFBG miscanthus gasification tests.

112

6 HHV @ 0.35 MPa HHV @ 0.4 MPa HHV @ 0.5 MPa HHV @ 0.7 MPa

4

3

HHV (MJ/mn ,wet gas)

5

3 2 1 0 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ(−)

Figure 4.10 Higher heating value of miscanthus pellet derived product gas behind the filter.

Carbon Conversion @ 0.35 MPa Carbon Conversion @ 0.4 MPa Carbon Conversion @ 0.5 MPa Carbon Conversion @ 0.7 MPa Cold Gas Efficiency @ 0.35 MPa Cold Gas Efficiency @ 0.4 MPa Cold Gas Efficiency @ 0.5 MPa Cold Gas Efficiency @ 0.7 MPa

Carbon conversion (%)

90 80 70 60

65 60 55 50 45

50

40

40

35

30

Cold Gas Efficiency on HHV basis (%)

70

100

30 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (−)

Tar concentration (mg/mn3, dry purge N2 free)

Figure 4.11 Carbon conversion and cold gas efficiency versus λ for miscanthus gasification. 1.0E+04 [PAH] @ 0.35 MPa [Phenols] @ 0.35 MPa [PAH] @ 0.4 MPa [Phenols] @ 0.4 MPa [PAH] @ 0.5 MPa [Phenols] @ 0.5 MPa [PAH] @ 0.7 MPa [PAH] @ 0.7 MPa

1.0E+03

1.0E+02

1.0E+01

1.0E+00 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ ( −)

Figure 4.12 Tar concentrations as measured behind the filter versus λ for miscanthus gasification.

113

3500

700 [HCN] (ppmv,wet,purge N2 corrected)

[NH3] (ppmv,wet,purge N2 corrected)

[NH3] @ 0.35 MPa

3000

600

[NH3] @ 0.4 MPa [NH3] @ 0.5 MPa [NH3] @ 0.7 MPa

2500

[HCN] @ 0.35 MPa

500

[HCN] @ 0.4 MPa

2000

400

[HCN] @ 0.5 MPa [HCN] @ 0.7 MPa

1500

300 Minphyl addition

1000

200

500

100

0 0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0 0.60

λ (−)

Figure 4.13 NH3 and HCN concentration versus air stoichiometry for miscanthus gasification.

25

Fuel_N conversion to NH 3 (%)

90

Fuel_N to NH3 @ 0.35 MPa

80

20

Fuel_N to NH3 @ 0.4 MPa Fuel_N to NH3 @ 0.5 MPa

70

Fuel_N to NH3 @ 0.7 MPa

60

Fuel_N to HCN @ 0.35 MPa

50

Fuel_N to HCN @ 0.5 MPa

Fuel_N to HCN @ 0.4 MPa Fuel_N to HCN @ 0.7 MPa

40

MinPhyl addition

30 20

15

10

5

Fuel_N conversion to HCN (%)

100

10 0 0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

0 0.60

λ (−)

Figure 4.14 Fuel bound nitrogen conversion into NH3 and HCN versus air stoichiometry for miscanthus gasification.

114

7.0

12.0

6.0

10.0 "020429"

[H2] (vol.%,wet)

[CO] (vol.%,wet)

5.0

"020513"

8.0

"990107_2" "981216"

"981223"

6.0

"020429" "020513"

"990107_1"

4.0

"990107_2" "981216"

4.0

"981223" "990107_1"

3.0 2.0

2.0

1.0 0.0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

2

7

2.5

3

3.5

4

5.5

6

6.5

7

"020429"

16.0

3.5

"990107_2" "981216" "981223"

2.0

"990107_2"

"990107_1"

1.5

"981216" "981223"

12.0

[CO2] (vol.%,wet)

"020513"

2.5

"020513"

14.0

"020429"

3.0 [CH4] (vol.%,wet)

5

18.0

4.0

"990107_1"

10.0 8.0 6.0

1.0

4.0

0.5

2.0 0.0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

2

7

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

Reactor height (m)

Reactor height (m)

700

25.0

600

"020429"

20.0

"020429"

"020513"

[C2H2] (ppmv,wet)

"981216"

15.0

"020513"

500

"990107_2"

[H2O] (vol.%,wet)

4.5

Reactor height (m)

Reactor height (m)

"981223" "990107_1"

10.0

"990107_2" "981216"

400

"981223" "990107_1"

300 200

5.0

100 0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

2

7

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

Reactor height (m)

Reactor height (m) 400

2500

350

2000

"020429"

[HCN] (ppmv,wet)

[NH3] (ppmv,wet)

"990107_2" "981216"

1500

"020429"

300

"020513"

"981223" "990107_1"

1000

"020513"

"990107_2"

250

"981216" "981223"

200

"990107_1"

150 100

500

50 0

0 2

2.5

3

3.5

4

4.5

5

Reactor height (m)

5.5

6

6.5

7

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

Reactor height (m)

Figure 4.15 Measured gas concentrations in the PFBG freeboard for miscanthus gasification.

115

7

[CO2] (vol%,wet)

18

2500

9

2400

8

2300

120

2200

110

2100

100

2000

90

17

7

16

6

15

5

14

4

13

3

12

2

Wall

Centre

11

1

10 5

10

15

20

Wall

Centre

1900

130

80

1800

70

1700

60

1600

50

1500

0 0

140 NH3 (ppmv, wet) HCN (ppmv, wet)

40 0

25

HCN concentration (ppmv,wet)

19

10

[H2], [CO] (vol%, wet)

CO2 (vol%,wet) H2 (vol%,wet) CO (vol%,wet)

NH3 concentration (ppmv,wet)

20

5

10

15

20

25

Radial position (cm)

Radial Position (cm)

Figure 4.16 Measured gas concentrations at radial positions in the PFBG freeboard for miscanthus gasification, experiment number 990107_1, probe position P1.1 (P=0.7 MPa).

10 CO2 (vol%,wet) H2 (vol%,wet) CO (vol%,wet)

[CO2] (vol%,wet)

18

2500

9 8

17

7

16

6

15

5

14

4

13

3

12

140

NH3 (ppmv, wet) HCN (ppmv, wet)

2400 2300 [H2], [CO] (vol%, wet) [NH3] (ppmv,wet)

19

130 120

Wall

Centre

2200

110

2100

100

2000

90

1900

80

1800

70

2

1700

60

11

1

1600

50

10

0

1500

Wall

Centre 0

5

10

15

Radial Position (cm)

20

25

40 0

5

10

15

20

25

Radial position (cm)

Figure 4.17 Measured gas concentrations at radial positions in the PFBG freeboard for miscanthus gasification, experiment number 990107_2, probe P1.1 (P=0.4 MPa).

116

[HCN] (ppmv,wet)

20

4.2.3 Wood gasification Table 4.5 presents the experimental PFBG process conditions and measured data with crushed wood pellets as fuel. In total 8 experiments were performed with this fuel. Representative graphs of the main experimental results have been grouped at the end of this paragraph. Compared to the miscanthus gasification experiments, higher temperatures in bed and freeboard section could be allowed due to the much lower sintering risk. The use of steam to moderate temperatures therefore was not necessary. For comparable air stoichiometry values, higher temperatures were obtained due to the higher heating values of wood compared to miscanthus (see table 4.1). Table 4.5 Experimental data for wood gasification from the Delft PFBG test rig. Experiment

011030 011127 020111 020129

Pressure (top of gasifier) Bed temperature Freeboard temperature Temperature behind the filter unit Mass flow fuel Mass flow gasification air Steam:air ratio (mass basis) 4) Dolomite:fuel ratio (mass basis) N2 flow fuel feeding N2 purge flow probes N2 purge flow ceramic filter Bed mass λ gasifier Composition LCV gas after filter (wet basis)

BTX (dry basis) 1) PAH’s (spa, dry basis) 1) Phenols (spa, dry basis) 1) Higher Heating Value (wet gas) Carbon conversion(solids basis) Cold gas efficiency (HHV basis) C-balance closure gasifier + filter O-balance closure gasifier + filter N-balance closure gasifier + filter H-balance closure gasifier + filter Ash-balance closure gasifier + filter Fuel_N to NH3 Fuel_N to HCN

CO H2 CH4 C2H4 C2H6 CO2 H2O O2 N23) Ar2) C2H2 H2S COS NH3 HCN

[MPa] [K] [K] [K] [kg/h] [kg/h] [-] [-] [kg/h] [kg/h] [kg/h] [kg]

0.35 1160 1057 887 129.1 253.5 0 0 12.7 1.9 1.9 146

[-] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [ppmv] [ppmv] [ppmv] [ppmv] [ppmv] [mg/mn3] [mg/mn3] [mg/mn3]

0.36 9.74 6.82 3.91 0.80 n.m 14.89 11.09 0 52.10 0.50 141 n.m. 0 1271 100 n.m. n.m. n.m.

[MJ/mn3] [%] [%] [%] [%] [%] [%] [%] [%] [%]

4.17 98.2 64.1 +2 -12 -4 -22 -73 74.2 6.0

1) determined by spa 2) Calculated from Ar balance n.m.: not measured

3)

0.35 1131 1004 886 111.6 217.2 0 0 11.2 1.9 3.2 130

0.35 1175 1037 891 88.2 209.0 0 0.036 8.7 1.9 3.2 143

0.32 0.39 11.30 9.13 7.27 5.24 3.97 3.26 0.86 0.57 0.20 0.09 14.82 14.81 11.77 19.12 0 0 49.12 47.12 0.53 0.60 277 221 n.m. n.m. 0 0 1051 483 144 41 1.0*103 1.9*102 2.6*103 2.7*103 2.3*102 0 4.64 98.1 65.5 -1 -10 +3 +7 +56 96.2 13.5

Calculated by difference

3.55 97.3 54.1 +4 -17 +13 -9 +30 47.8 4.2 4)

020205

020212

020220

020226

0.50 1214 1092 919 126.3 288.3 0 0.036 18.0 1.9 3.2 151

0.50 1207 1090 946 128.1 290.0 0 0.036 18.1 1.9 3.2 146

0.50 1167 1103 940 108.9 258.2 0.10 0.036 15.5 1.9 3.2 147

0.35 1087 1033 871 87.7 194.8 0.099 0 8.9 1.9 3.2 136

0.35 1117 1078 881 68.4 188.8 0.10 0 6.8 1.9 0.4 142

0.38 10.67 6.39 3.91 0.33 0.07 14.90 12.21 0 50.90 0.57 78 n.m. 0 381 27 n.m. n.m. n.m.

0.37 9.62 6.37 3.87 0.30 0.086 15.37 13.86 0 49.83 0.57 44 41 0 1086 63 1.6*103 3.0*103 0

0.39 7.16 6.09 3.26 0.34 0.11 15.29 18.54 0 48.64 0.54 59 n.m. 0 223 15 4.8*102 1.9*103 0

0.37 7.42 6.18 3.01 0.63 0.13 14.87 18.70 0 48.49 0.54 115 69 0 257 19 7.1*101 2.0*103 2.0*102

0.46 5.78 5.38 2.48 0.45 0.065 15.12 22.20 0 47.91 0.56 86 n.m. 0 428 27 6.0*101 1.8*103 1.6*101

3.98 96.6 61.5 -2 -10 +5 +4 +1 68.7 4.9

3.82 97.3 58.3 -1 -11 +8 +4 +34 62.1 3.7

3.27 98.5 54.8 +3 -4 +7 +10 +17 44.0 3.1

3.41 97.9 55.4 +4 -4 +6 +11 -49 49.1 3.8

2.73 98.7 51.8 -1 -10 +7 0 -62 94.6 6.2

steam @ saturation temperature; air @ ambient temperature

Figure 4.18 shows the concentrations of CO2, CO, H2 and CH4, which are main gas species, as a function of the air stoichiometry, λ. The values are decreasing in the sequence mentioned above. With decreasing value of λ the concentration of combustible gases is increasing. The measured concentrations of these compounds compare well to values reported by [Kurkela et al., 1993b] for sawdust gasification at 0.4 – 0.5 MPa in a 300 kWth bubbling PFB, see also table 2.5. 117

The results for these main components also agree quite well with the atmospheric bubbling FB results given by [Gil et al., 1999b] and [Narvaez et al., 1996] for small pine wood particles gasified with air. The conversion of carbon into the main carbon containing gaseous components is shown in figure 4.19. The conversion of carbon into CO2 shows a similar trend as the one observed for miscanthus gasification (see figure 4.9). The CO yields are higher than those of miscanthus gasification. The carbon conversion is also significantly higher. Two factors contribute to the higher CO yields for wood gasification: higher temperatures, which are caused by the absence of steam, and smaller particles which lead to better conversion of carbon in heterogeneous partial combustion, an important factor in the overall carbon conversion (see [Kersten, 2002]). The conversion of carbon into the main carbonaceous gas constituents is comparable to those reported by [Padban, 2000] for sawdust gasification in a 100 kWth PFB test rig. For gasification of woody biomass in an ACFB test rig, qualitatively the same trends as observed in this work are shown [Kersten, 2002]. In his work the CCO2 conversion is slightly lower, whereas the C-CO conversion is somewhat higher. In our case the CCH4 conversion is higher. At this stage, no conclusion can be drawn regarding this difference. As a result of the abovementioned effect, the higher heating value of the gasification product gas is increasing with decreasing λ values, as is seen in figure 4.20. The HHV values are in a range which is characteristic for low calorific gases, because air, which has a high N2 content, is used as oxidizer in the gasification process. The carbon conversion observed in the wood gasification experiments is presented in figure 4.21, together with cold gas efficiency. The carbon conversion is well above 95% for the wood gasification conditions applied in our PFBG (λ values from 0.32 to 0.46). The carbon conversion values are quite well comparable to those reported by [Kurkela et al., 1993b] for the VTT 500 kWth pressurised FB. As a result of the increasing heating value of the gas produced with decreasing λ and almost constant carbon conversion values, the cold gas efficiency is increasing as λ values decline. Such a trend was also observed under ACFB conditions for wood, see [Van der Drift et al., 2002]. As compared to miscanthus higher carbon conversions were obtained for comparable air stoichiometry values. This can be attributed to differences in particle size (wood tests were performed with somewhat finer particles) and higher reactivity of the wood fuel, which is also observed under flash pyrolysis conditions when agricultural residues were compared with woody biomass. In paragraph 4.4.3 and 4.4.4 these differences can be observed in higher volatile matter yields for wood pyrolysis compared to miscanthus. Char yields were higher for the agricultural residues. This was attributed to both higher lignin contents and higher amounts of ash, favouring charring reactions [Zanzi, 2001]. Tar concentrations behind the filter unit as determined by the SPA technique are given in figure 4.22 as a function of λ. Slightly increasing values of light polyaromatic species (MW<202 kg/kmol) concentrations were observed with declining air stoichiometry values. The phenolic compound concentrations remained well below 1 [g/mn3], whereas PAH concentrations in the range of a few [g/mn3] were measured. These values are typical for pressurised wood gasification in bubbling fluidised beds, see e.g. [Kurkela et al., 1993a], [Milne et al., 1998], [Mörsch, 2000], [Neeft, 2000], [Padban, 2000], though measured by different, non-standardised techniques. Somewhat higher values for light PAH species were reported by [Brage et al., 2000] for top-fed bubbling PFB gasification performed at 0.4 MPa. Already in earlier studies this was also found in atmospheric fluidised bed gasification, see [Delgado et al., 1997] and [Corella et al., 1988]. Figure 4.23 and 4.24 show the concentrations of the gaseous nitrogen compounds, NH3 and HCN, in the producer gas. No indication was found by means of FTIR analysis for the presence of HNCO. Significantly lower values of NH3 concentrations were observed as compared to the miscanthus gasification experiments. This is largely contributed to the lower fuel bound nitrogen content, as can be observed in table 4.1. The main part of the fuel bound nitrogen is converted into bound gaseous species, as also already observed for miscanthus gasification. The values for the fuel bound N conversion into NH3 are comparable to those reported by [Padban, 2000] and [Wang & Olofsson, 2002] for sawdust and bark in bottom-fed 90 kWth bubbling PFB gasification tests. 118

The NH3 concentrations were in range with VTT 500 kWth PFB sawdust gasification tests, published by [Kurkela et al., 1993], cf. table 2.5. The fuel-N to NH3 conversions reported by [Chen, 1998] for a wood-fired top-fed PFB, however, were significantly lower. In a study with a variable feed location, [Vriesman et al., 2000] showed that the location has a significant influence on the fuel nitrogen behaviour: top feeding resulted in lower fuel-N to NH3 conversions. This was attributed to the (pyrolysis) atmosphere in which the particles are pyrolysed, for top feeding this is a reducing environment with relatively high carbon contents, whereas for bottom feeding the atmosphere is a flaming pyrolysis environment where initially oxygen is available and which is characterised by comparatively low carbon contents. In figure 4.25 the axial concentration profiles of several main and minor product gas compounds over the freeboard length are depicted. Unfortunately, due to practical reasons, measurements at all three probe positions could not be carried out for all experiments. For most components the axial profiles show practically constant concentrations, as can be seen in this figure. An exception is the clear decrease in acetylene (C2H2) concentrations. This is probably due to the involvement of this component in tar and soot formation processes and requires further research, which is beyond the scope of this thesis. Relatively low sulphur compound levels were measured, as can be seen in table 4.5. COS could not be detected quantitatively by FTIR and the experiments during which H2S was measured indicate a concentration well below 100 ppmv. This is attributed to the very low sulphur content of the Labee wood pellets and this is an advantage when this fuel type is used in practical thermal conversion systems.

119

[CO],[H2],[CO2],[CH4] (vol%,wet,purge N2 free)

18 16 [CO] wood gas @ 0.5 MPa [CO] wood gas @ 0.35 MPa [H2] wood gas @ 0.5 MPa [H2] wood gas @ 0.35 MPa

14 12

[CO2] wood gas @ 0.5 MPa [CO2] wood gas @ 0.35 MPa [CH4] wood gas @ 0.5 MPa [CH4] wood gas @ 0.35 MPa

10 8 6 4 2 0 0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

Figure 4.18 Main gas phase concentrations in the Delft PFBG wood gasification tests.

Fuel_C into gaseous species (%)

70

%Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected] %Fuel_C to [email protected]

60 50 40 30 20 10 0 0.3

0.4

0.5

0.6

λ (−)

Figure 4.19 Main gas phase concentrations in the Delft PFBG wood gasification tests. 6

H H V @ 0 .5 M P a H H V @ 0 .3 5 M P a

4

3

HHV (MJ/mn )

5

3 2 1 0 0 .2 5

0 .3

0 .3 5

0 .4

0 .4 5 λ (-)

0 .5

0 .5 5

0 .6

Figure 4.20 Higher heating value of wood pellet derived product gas behind the filter.

120

70

Carbon Conversion (%)

95

65 60

90

55

85

50 45

80 Carbon Conversion @ 0.35 MPa Carbon Conversion @ 0.5 MPa Cold Gas Efficiency @ 0.35 MPa Cold Gas Efficiency @ 0.5 MPa

75 70

40 35

Cold Gas Efficiency on HHV basis (%)

75

100

30 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

Figure 4.21 Carbon conversion and cold gas efficiency versus air stoichiometry for wood gasification.

1000

3

tar concentration (mg/mn ,dry,purge free)

10000

100

PAH @ 0.5 MPa PAH @ 0.35 MPa Phenols @ 0.5 MPa Phenols @ 0.35 MPa

10 0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

Figure 4.22 Tar concentrations as measured behind the filter versus air stoichiometry for wood gasification

121

[NH3],[HCN] (ppmv,wet,purge N2 corrected)

1400 1200

[NH3] @ 0.5 MPa

[NH3] @ 0.35 MPa

1000

[HCN] @ 0.5 MPa [HCN] @ 0.35 MPa

800 600 400 200 0 0.25

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

50

90

45

80

40

70

35

60

30

50

25

40

20

Fuel_N to NH 3 (%)

100

30 20

10

Fuel_N to HCN @ 0.5 MPa Fuel_N to HCN @ 0.35 MPa

10 0 0.25

15

Fuel_N to NH3 @ 0.5 MPa Fuel_N to NH3 @ 0.35 MPa

Fuel_N to HCN (%)

Figure 4.23 NH3 and HCN concentrations versus air stoichiometry for wood gasification.

5 0

0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

Figure 4.24 Fuel-N conversion into gaseous species versus air stoichiometry for wood gasification.

122

14.0

8.0

12.0

7.0

"011030" "011127"

"020111" "020220"

8.0

"020226"

6.0

"020212"

[H2] (vol%,wet)

[CO] (vol%,wet)

10.0

"011030" "011127" "020111" "020220" "020226" "020205" "020212"

6.0

"020205"

4.0

5.0 4.0 3.0 2.0

2.0

1.0

0.0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

2

2.5

3

3.5

Reactor height (m)

4.5

4.5

5

5.5

6

6.5

7

20.0 18.0

4.0 "011030"

3.5

"020111"

3.0

"020220" "020226"

2.5

"020205" "020212"

2.0

"011030"

16.0

"011127"

[CO2] (vol.%,wet)

[CH4] (vol.%,wet)

4

Reactor height (m)

1.5 1.0

"011127"

14.0

"020111"

"020220"

12.0

"020226" "020205"

10.0

"020212"

8.0 6.0 4.0

0.5

2.0

0.0

0.0

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

2

2.5

3

3.5

Reactor height (m)

4

4.5

5

5.5

6

6.5

7

Reactor height (m)

700

25.0

600 "011030"

20.0

500 [C2H2] (ppmv,wet)

[H2O] (vol.%,wet)

"011127" "020111" "020220"

15.0

"020226" "020205" "020212"

10.0

"011030" "011127" "020111" "020220" "020226" "020205" "020212"

400 300 200

5.0

100 0.0

0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

2

2.5

3

3.5

4

Reactor height (m)

4.5

5

5.5

6

6.5

7

Reactor height (m)

1400

200 180

1200

"011127"

[HCN] (ppmv,wet)

[NH3] (ppmv,wet)

160 "011030"

1000

"020111"

800

"020220" "020226"

600

"020205" "020212"

400

"011030"

140

"011127"

120

"020111"

100

"020220" "020226" "020205"

80

"020212"

60 40

200

20

0

0 2

2.5

3

3.5

4

4.5

5

Reactor height (m)

5.5

6

6.5

7

2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

7

Reactor height (m)

Figure 4.25 Measured gas concentrations in the PFBG freeboard at different axial positions for wood gasification.

123

4.2.4 Brown coal gasification Table 4.6 gives the main results of the Delft PFBG experiments with brown coal as fuel. The gas concentrations and conversions are measured downstream of the ceramic filter unit. The fuel doesn’t show the high sintering risk as high alkali biomass, so also in this case higher temperatures were allowed. No additive was applied in these experiments, as no sintering was expected and comparatively low tar contents in the produced gas were expected. Additive use for sulphur capture was also not considered as the Ca content of the fuel was considered to be sufficient for this purpose. Table 4.6 Experimental data for brown coal gasification from the Delft PFBG test rig. 020306 020319 020409 020416

Experiment [MPa] [K] [K] [K] [kg/h] [kg/h] [-] [-] [kg/h] [kg/h] [kg/h] [kg]

0.35 1036 1110 886 56.1 206.1 0.23 0 2.9 1.9 1.6 214

0.50 1131 1171 976 103.0 309.9 0 0 7.4 1.9 3.2 168

0.50 1119 1119 928 86.4 268.1 0.10 0 6.6 1.9 1.0 128

0.35 1181 1140 908 61.4 207.9 0 0 2.9 1.9 0.5 149

BTX (dry basis) PAH’s (spa, dry basis) Phenols (spa, dry basis)

[-] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [vol%] [ppmv] [ppmv] [ppmv] [ppmv] [ppmv] [mg/mn3] [mg/mn3] [mg/mn3]

0.52 4.04 7.56 0.88 0.10 n.m. 15.22 24.0 0 47.54 0.55 0 n.m. 6 1136 6 n.m. n.m. n.m.

0.40 14.09 9.24 1.35 0.02 0.021 11.73 7.24 0 55.50 0.66 0 149 16 1296 23 n.m. n.m. n.m.

0.43 9.49 10.13 1.54 0.095 0.05 14.16 14.52 0 49.25 0.58 0 132 11 1690 9 n.m. n.m. n.m.

0.47 10.91 7.52 0.85 0.047 0.016 14.28 7.09 0 58.47 0.70 0 82 16 1016 9 n.m. n.m. n.m.

Higher Heating Value (wet gas) Carbon conversion(solids basis) Cold gas efficiency (HHV basis) C-balance closure gasifier + filter O-balance closure gasifier + filter N-balance closure gasifier + filter H-balance closure gasifier + filter Ash-balance closure gasifier + filter Fuel_N to NH3 Fuel_N to HCN

[MJ/mn3] [%] [%] [%] [%] [%] [%] [%] [%] [%]

1.88 95.4 43.2 +5 -4 +9 +1 -22 58.8 0.3

3.52 94.9 55.4 +13 -5 +13 +4 +18 45.9 0.8

3.19 90.4 52.8 +2 -8 +16 -9 +56 74.5 0.4

2.71 91.3 41.6 +7 -4 +10 +1 +15 47.9 0.4

Pressure (top of gasifier) Bed temperature Freeboard temperature Temperature behind the filter unit Mass flow fuel Mass flow gasification air Steam:air ratio (mass basis) 3) Dolomite:fuel ratio (mass basis) N2 flow fuel feeding N2 purge flow probes N2 purge flow ceramic filter Bed mass λ gasifier Composition LCV gas after filter (wet basis)

1)

calculated by difference n.m.: not measured

2)

CO H2 CH4 C2H4 C2H6 CO2 H2O O2 N21 Ar2 C2H2 H2S COS NH3 HCN

calculated from Ar balance

3)

steam @ saturation temperature; air @ ambient temperature

In figure 4.26 the concentration of the main fuel product gas constituents, CO and H2, are depicted for different stoichiometries. The concentrations of the main compounds are increasing with decreasing air stoichiometry, which is due to the decreased availability of O2 for CO2 and H2O formation from CO and H2, respectively. This trend is confirmed by figure 4.27, where the carbon conversion into these species is shown. The values for the conversion of C into CO are slightly lower than those reported by [Kurkela et al., 1995] for a quite comparable fuel and λ range. The difference can possibly be attributed to the differences in air distribution in the reactor; the authors use a relatively large amount of secondary air.

124

Figure 4.28 shows increasing product gas HHV values with decreasing λ value. Gasification of brown coal results in lower hydrocarbon concentrations and C to light hydrocarbon species conversion, whereas CO and H2 concentrations are high compared to the wood gasification experiments. Biomass, compared to brown coal, is a young fuel with high oxygen content, which promotes reactivity. During fast biomass pyrolysis, the initial gasification process, relatively high amounts of tar and oxygenates are produced, which are converted into smaller hydrocarbons, during secondary reactions. Conversion of biomass into H2 is low compared to brown coal, which is attributed to its fuel structure, of which H atoms can more easily react with fuel-O to form water. Figure 4.29 and table 4.6 show carbon conversions in excess of 90% and cold gas efficiencies between circa 43 and 55%. These carbon conversion values are comparable to extrapolated values obtained by [Nagel, 2002] in the range of 0.5 < λ < 0.9 for Garzweiler and Hambach brown coal. These values are somewhat lower than for woody biomass, which is probably a consequence of the lower heterogeneous fuel reactivities observed for the fossil fuel. Figure 4.30 and 4.31 show the NH3 and HCN concentrations and the relative conversion values. Low HCN concentrations –when compared to biomass- were measured during the brown coal gasification experiments. This can be attributed to the relatively high Ca content of the fuel ash, which causes the catalysed conversion of fuel bound nitrogen into NH3. Similar data of fuel-N conversion to NH3 and HCN were obtained by [Kurkela et al., 1992] for Rhenish brown coal gasification in their PFB reactor. The data given in table 4.6 indicate that steam addition results in higher fuel-N conversion into NH3, which is in agreement with the findings of [Paterson et al., 2002] and [Zhuo et al., 2002] for a coalfuelled pressurised spouted bed. No HNCO could be detected just as in the case of our experiments with wood and miscanthus as fuel. The very low HCN/NH3 ratio of the product gas in the tests was also found by [Paterson et al., 1997], for the 2 MWth CTDD pressurised fluidised bed gasifier with a comparable Rheinbraun brown coal. Unfortunately, these authors only indicate the concentrations of minor species. The average carbon conversion they present is 79%, which is much lower than our values. Low fuel nitrogen conversions to HCN of only a few percent were also found by [Kurkela et al., 1992], which is somewhat higher than in our experiments for this fuel; these authors reported somewhat higher values of NH3 concentrations in the product gas (approximately 2000 ppmv in dry gas), which can also be attributed to their reported nitrogen content of the brown coal fuel (being 0.8 versus 0.58 mass% dry basis in our case). Figure 4.32 shows the axial gas profiles in the freeboard. Some increase in CO2 and a decrease in H2O concentrations can be seen for experiment 020409. Most experiments show relative constant gas concentrations. HCN concentrations show a slight decrease for experiments 020416 and 020319 which could be attributed to hydrolysis, which is slower for these experiments possibly due to the lower water concentrations compared to experiment number 020409. Again, as in the case of biomass gasification, acetylene (C2H2) shows a clear decrease over the freeboard height. For this fuel, the values are much lower than the values for biomass, i.e. one order of magnitude lower. This is accompanied with lower tar concentrations. Unfortunately, tar concentrations could not be determined by the SPA technique. An on-line tar analyser was applied for experiment 020416 to quantify the total condensable tar content of the gas. A value of 6.8+0.4 .102 mg/mn3 (wet) was determined behind the filter and 4.9+0.3 .102 mg/mn3 at probe position P2.1. These values are clearly lower than for biomass for similar λ values. This is in accordance with literature data [Kurkela et al., 1993a]. [Kurkela et al., 1995] present a value of total tar 743 mg/mn3 including benzene for Rhenish brown coal gasification at λ=0.44 and a pressure of 0.5 MPa, comparing well to the order of magnitude found in our study using the IVD on-line tar analyser. Because the German brown coal used was low in sulphur content and the Ca content in the fuel was comparatively high, relatively low contents of H2S and COS were measured, as can be seen in table 4.6. The Ca:S molar ratio in the fuel was ca. 2.5. The COS concentration was 10-20% of the H2S concentration. The sulphur gas concentrations, H2S and COS, were somewhat lower than those reported by [Kurkela et al., 1995] despite their higher Ca:S molar ratio of 3.7-3.8. The difference is attributed to the smaller particles used in our study.

125

[CO],[H2],[CO2],[CH4] (vol%,wet,purge corrected)

18 [CO] @ 0.5 MPa

16

[CO] @ 0.35 MPa

14

[H2] @ 0.5 MPa

12

[CO2] @ 0.5 MPa

[H2] @ 0.35 MPa [CO2] @ 0.35 MPa

10

[CH4] @ 0.5 MPa [CH4] @ 0.35 MPa

8 6 4 2 0 0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

λ (-)

Figure 4.26 Main components concentrations for brown coal gasification versus air stoichiometry.

Fuel_C to main gaseous species (%)

100 %Fuel_C to CO @0.5 MPa

90

%Fuel_C to CO @0.35 MPa

80

%Fuel_C to CO2 @0.35 MPa

%Fuel_C to CO2 @0.5 MPa %Fuel_C to CH4 @0.5 MPa

70

%Fuel_C to CH4 @0.35 MPa %Fuel_C to C2H4 @0.5 MPa

60

%Fuel_C to C2H4 @0.35 MPa

50 40 30 20 10 0 0.3

0.4

0.6

0.5 λ (-)

Figure 4.27 Fuel bound carbon conversion into main gaseous fuel components for brown coal. 4 HHV @ 0.5 M pa

3

Higher heating value (wet, MJ/mn )

3.5

HHV @ 0.35 M Pa

3 2.5 2 1.5 1 0.5 0 0.3

0.35

0.4

0.45

0.5

0.55

0.6

λ (-)

Figure 4.28 HHV of brown coal derived product gas behind the filter versus air stoichiometry.

126

65

95

Carbon conversion (%)

60 Carbon conversion @ 0.5 MPa Carbon conversion @ 0.35 MPa Cold gas efficiency @ 0.5 MPa Cold gas efficiency @ 0.35 MPa

90

55 50

85

45

80

40 75

35

70

Cold Gas Efficiency on HHV basis (%)

70

100

30 0.2

0.25

0.3

0.35

0.4

0.45 λ (-)

0.5

0.55

0.6

0.65

0.7

Figure 4.29 C-conversion and η cold gas versus air stoichiometry for brown coal gasification. 100

2500 [NH3] @ 0.35 MPa

2000

80

[HCN] @ 0.5 MPa [HCN] @ 0.35 MPa

1500

60

1000

40

500

20

0

[HCN] (ppmv,wet,purge free)

[NH3] (ppmv,wet,purge free)

[NH3] @ 0.5 MPa

0 0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

λ (-) 100

50

90

45

80

40

70

35

60

30

50

25

40

20

30

15

Fuel_N to NH3 @ 0.5 MPa Fuel_N to NH3 @ 0.35 MPa

20

10

Fuel_N to HCN @ 0.5 MPa

10

Fuel_N to HCN (%)

Fuel_N to NH3 (%)

Figure 4.30 NH3 and HCN concentrations versus λ for brown coal gasification.

5

Fuel_N to HCN @ 0.35 MPa

0

0 0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

λ (-)

Figure 4.31 Fuel bound nitrogen conversion into NH3 and HCN versus λ for brown coal gasification.

127

12.0

16.0 14.0

10.0

"020416"

"020416"

"020319"

"020409"

[H2] (vol.%,wet)

[CO] (vol.%,wet)

12.0 10.0 8.0 6.0

8.0

"020319"

"020409"

6.0

4.0

4.0

2.0

2.0 0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

0.0

7

2

2.5

3

3.5

Reactor height (m) 1.8

4.5

5

5.5

6

6.5

7

16.0

1.6

14.0 "020416"

1.4

12.0

"020416" "020319"

1.2

[CO2] (vol.%,wet)

[CH4] (vol.%,wet)

4

Reactor height (m)

"020409"

1.0 0.8 0.6

"020319" "020409"

10.0 8.0 6.0 4.0

0.4

2.0

0.2

0.0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

2

7

2.5

3

3.5

Reactor height (m)

4

4.5

5

5.5

6

6.5

7

Reactor height (m) 70

18.0 16.0

60 50

12.0

[C2H2] (ppmv,wet)

[H2O] (vol.%,wet)

14.0 "020416" "020319" "020409"

10.0 8.0 6.0

"020416"

40

"020319" "020409"

30 20

4.0

10

2.0

0

0.0 2

2.5

3

3.5

4

4.5

5

5.5

6

6.5

2

7

2.5

3

3.5

2000

100

1800

90

1600

4.5

5

5.5

6

6.5

7

80

"020416"

1400

"020319"

1200

"020409"

[HCN] (ppmv,wet)

[NH3] (ppmv,wet)

4

Reactor height (m)

Reactor height (m)

1000 800 600

70

50 40 30

400

20

200

10

0

"020416" "020319" "020409"

60

0

2

2.5

3

3.5

4

4.5

5

Reactor height (m)

5.5

6

6.5

7

2

2.5

3

3.5

4

4.5

5

5.5

6

Reactor height (m)

Figure 4.32 Measured gas concentrations in the PFBG freeboard for brown coal gasification.

128

6.5

7

4.3 Experimental results of DWSA gasification tests

4.3.1 Overview of the DWSA measurement programme Six experiments were performed using the IVD DWSA installation. The average duration of stable operation of the test rig was between 3.5 to 4.5 hours. Almost 1.5 hours were needed to reach steady state operation conditions. Table 4.7 presents an overview of the main process conditions of the experiments using the DWSA test rig. The main parameters were fuel type and air stoichiometry. Air was introduced into the reactor at ambient temperature. The fuels used in the gasification tests were: German brown coal (BC) from the Hambach open mine as a relatively young fossil fuel and crushed wood pellets from Labee A type wood (PW) as biomass species. For these experiments the air-stoichiometry was the most important process variable. For the wood experiments the pressure was also varied. Table 4.7 Experimental programme overview for the DWSA gasification test rig. Experiment

991103

991115

991118

991202

991206

991208

BC 0.51 791 3.7 7.3 2.5 12.8 0.32

BC 0.51 802 2.5 8.4 2.8 13.6 0.51

BC 0.51 858 2.2 9.2 2.7 14.1 0.66

PW 0.51 782 3.3 5.8 7.2 16.3 0.30

PW 0.51 824 3.5 9.0 3.0 14.9 0.48

PW 0.15 792 1.6 4.6 1.4 7.6 0.52

Fuel Pressure; P (MPa) Bed temperature; Tbed (°C) Fuel flow (raw, kg/h) Primary air flow (kg/h) N2 flow to gasifier; φm,N2(kg/h) LCV gas flow; φm,LCV (kg/h) Primary air stoichiometry; λ (-)

In table 4.8 the particle size distributions are given for bed material and for the fuels. Somewhat smaller sand particle sizes were applied, so that lower fluidisation velocities could be applied than the Delft PFBG experiments. In this way comparable air stoichiometry values could be applied with the existing fuel feeding capacity. The fuel particle size distributions differ also from those used in the Delft PFBG experiments. For the biomass fuels these are shifted to smaller values due to differences in the screw feeder dimensions (screw speed and space between screw and wall). Figure 4.33 shows a graphical representation of these data. Table 4.8 Particle size distributions of bed material and fuels used. dp range (µm) 0 53 53 90 90 - 200 200 - 400 400 - 600 600 - 800 800 - 1000 1000 – 1400 1400 – 2000 2000 – 3150 3150 – 4000 4000 – 5000 5000 +

Dp,average (µm) Dp,SMD (µm) ρparticle (g.cm-3) ρbulk (g.cm-3)

dp average (µm) 26.5 71.5 145 300 500 700 900 1200 1700 2575 3575 4500 5500

Sand (Mass %)

Wood (Mass %)

Brown Coal (Mass %)

0.05 0.20 1.62 51.72 46.29 0.10 0.01 0.0 0.0 0.0 0.0 0.0 0.0 390 356

0.01 0.78 3.51 8.28 12.30 10.84 10.62 19.18 22.38 11.41 0.68 0.0 0.0 1192 672

0.0 0.21 4.56 18.91 27.11 38.08 2.08 8.69 0.31 0.03 0.03 0.0 0.0 595 463

2.58 1.45

1.49 0.36

1.47 0.67

129

60

Crushed Wood Pellets Hambach Brown Coal Sand

Mass Percentage (%)

50

40

30

20

10

0-53

53-90

90-200

200-400

400-600

600-800

800-1000

1000-1400

1400-2000

2000-3150

3150-4000

4000-5000

>5000

0

Particle size per class (µm)

Figure 4.33 Size distribution of fuels and bed material used in the IVD DWSA tests.

4.3.2 Wood gasification Table 4.9 presents the results of steady state averaged measurements of gas concentrations downstream of the hot ceramic candle filter and data related to solid fuel conversion. The input data are given in table 4.7. Table 4.9 LCV wood product gas composition downstream of the filter and conversion data. Experiment

991202

991206

991208

7.76 5.19 2.83 0.53 9.83 8.15 65.39 0.31 126 n.m. 6 176 3.11 100 65.7 35.5 45.0 12.9 4.9 100 0** +2 -15 +4 -16

5.33 3.83 1.88 0.33 13.63 11.28 63.18 0.53 78 n.m. 6 149 2.12 100 40.8 22.2 56.8 7.8 2.8 80 2** +10 -4 0 -4

7.16 5.32 1.93 0.73 13.28 10.60 60.44 0.54 144 n.m. 5 176 2.81 100 57.0 34.3 58.6 8.5 6.4 99.8 2** -8 -8 +5 -17

LCV gas composition

CO (vol%, wet) H2 (vol%, wet) CH4 (vol%,wet) C2H4 (vol%, wet) CO2 (vol%, wet) H2O (vol%, wet) N2 (vol%, wet) Ar* (vol%, wet) C2H2 (ppmv, wet) H2S (ppmv, wet) COS (ppmv, wet) NH3 (vol%, wet) Higher heating value; HHV (MJ/mn3) Carbon conversion (%) (solid catch basis) Cold gas efficiency (%) Fuel-C conversion to CO (%) Fuel-C conversion to CO2 (%) Fuel-C conversion to CH4 (%) Fuel-C conversion to C2H4 (%) Fuel-N conversion to NH3 (%) Fuel-N conversion to HCN (%) C-balance closure gasifier + filter O-balance closure gasifier + filter N-balance closure gasifier + filter H-balance closure gasifier + filter

* calculated from Ar balance ** HCN could be detected by FTIR in single digit values, but these were lower than the measurement error

130

The values determined for fuel bound carbon conversion into main gas phase carbonaceous species during the experiments 991202 and 991206, both performed at ca. 0.51 MPa, are comparable to the values observed for gasification experiments using the Delft PFBG test rig. The values for carbon conversion into CO and CH4 are somewhat higher for experiment 991208 in comparison to the PFBG experiments shown in figure 4.19. The lower pressure in this experiment could be the reason for this observed deviation. The observed carbon conversions are relatively high and practically 100%. This almost complete conversion observed for wood can be attributed to its relatively high reactivity as compared to e.g. brown coal. The conversions were even slightly higher than in the larger scale PFBG experiments, which can be attributed to the smaller particle size applied in the DWSA tests. The experiments regarding wood gasification show that the fuel bound nitrogen is almost completely converted into NH3. Conversion of nitrogen into HCN and solid char bound nitrogen is small to negligible. This is characteristic for bottom-fed FB systems, as discussed before. HNCO could not be observed in the FTIR spectra obtained during these experiments. The NH3 and HCN concentrations are significantly lower than for the PFBG tests with comparable air stoichiometry values. This can be attributed to the comparatively lower fuel bound nitrogen content of the fuel used in these tests, as compared to the wood batches applied in the PFBG tests. 4.3.3 Brown coal gasification Table 4.10 shows steady state averaged gas concentrations downstream of the filter and data related to solid fuel conversion. The input data are given in table 4.7. Table 4.10 LCV Brown coal product gas composition downstream of the filter and conversion characteristics. Experiment

991103

991115

991118

14.00 10.42 1.99 0.09 11.19 5.20 56.40 0.48 0 487 25 2290 3.95 94 56.0 42.6 34.1 6.1 0.5 73 0.6 6 +11 -11 +2 -17

8.83 7.00 1.39 0.12 12.10 5.29 64.59 0.54 0 n.m. 27 1339 2.64 98 55.7 39.3 53.9 6.2 1.1 62 0.4 2.5 -3 -7 +3 -31

5.74 4.49 0.58 0.08 12.83 6.43 69.18 0.59 3 n.m. 27 770 1.58 99 37.9 28.7 64.1 2.9 0.8 40 0.3 0.6 +3 -2 +1 -19

LCV gas composition

CO (vol%, wet) H2 (vol%, wet) CH4 (vol%,wet) C2H4 (vol%, wet) CO2 (vol%, wet) H2O (vol%, wet) N2 (vol%, wet) Ar* (vol%, wet) C2H2 (ppmv, wet) H2S (ppmv, wet) COS (ppmv, wet) NH3 (vol%, wet) Higher heating value; HHV (MJ/mn3) Carbon conversion (%) (solid catch basis) Cold gas efficiency (%) Fuel-C conversion to CO (%) Fuel-C conversion to CO2 (%) Fuel-C conversion to CH4 (%) Fuel-C conversion to C2H4 (%) Fuel-N conversion to NH3 (%) Fuel-N conversion to HCN (%) Fuel-N conversion to char_N (%) C-balance closure gasifier + filter O-balance closure gasifier + filter N-balance closure gasifier + filter H-balance closure gasifier + filter

* calculated value 131

In table 4.10 and 4.9 it can be seen that the concentration of the light aliphatic hydrocarbons and the accompanying specific carbon conversion values are lower than for the wood based gasification experiments. The CO concentration and conversion of C to CO, however, are significantly higher for brown coal gasification under practically identical process conditions. The differences in composition of the main LCV gas can be attributed to the structure in which C, H and O are bound in the fuels. Biomass consists basically of cellulose, hemi-cellulose and lignin. In older fossil fuels like brown coal, however, more aromatic molecular structures (pyridinic, pyrrolic) are present. This leads to different behaviour during flash pyrolysis, which is together with drying the initial step in the fluidised bed gasification process. This different behaviour results in a different yield of initial products. The conversion of fuel bound nitrogen into NH3 with the brown coal experiments is comparable to values measured at the TU Delft PFBG test rig for similar λ values, although the process conditions could not be exactly matched. The conversion to HCN is below 1%, which has also been observed in the TU Delft PFBG measurements. The fuel bound nitrogen conversion to NH3 is significantly lower when compared to wood gasification (see also Table 4.9). This has been reported in the literature for a bottom-fed PFB, see e.g. [Kurkela, 1996]. The background of this different fuel-nitrogen release behaviour can be explained to a certain extent by the different nature of the chemical bonding of nitrogen, see for example [Leppälahti and Koljonen, 1995] and [Zhou, 1998]. In biomass, nitrogen is mainly present in the form of peptide bounds (in e.g. amino-acids and proteins). The oxygen content, more specifically the fuel bound O: fuel bound N ratio in the fuel, has been reported by [Hämäläinen and Aho 1995] also to influence the NH3/HCN ratio released during flash pyrolysis in such a way that at higher ratios relatively more NH3 is formed. Background of this phenomenon is probably that phenolic OH groups in the fuel structure enhance the reaction of OH radicals with HCN to finally form NH3 at the charring fuel surface. In older fuels, like brown coal and sub-bituminous coal, the nitrogen is mainly bound in pyridinic and pyrrolic structures. There are also investigations which show significantly lower conversions of fuelnitrogen into NH3 with biomass, see e.g. [Chen, 1998]. This was observed with a top fed fluidised bed, which is differing from bottom fed systems. The understanding of the devolatilisation mechanisms, however, is still not complete. Differences can possibly also be attributed to the environment in which the primary fast pyrolysis takes place: either oxidizing or reducing.

132

4.4 Experimental results of TG-FTIR pyrolysis tests

4.4.1 Overview of the TG-FTIR experimental programme TG-FTIR characterisation was performed for all fuels applied in this research work: Labee ‘A quality’ wood pellets, miscanthus Giganteus and German Rhenish brown coal. The experimental method and equipment were described in chapter 3 of this thesis. This characterisation work was done in order to obtain information on the kinetics of decomposition of the fuel under pyrolysis conditions in an inert environment. To derive the kinetic parameters, as will be shown below, experiments were performed at three heating rates: 10, 20 and 100 K/min. 4.4.2 Kinetic analysis approach It is assumed that the decomposition reaction rates of the fuels follow first-order kinetics, which is a reasonable starting assumption, because when an increasing amount of fuel is available, the higher the chance that this is decomposed without the reaction being retarded by itself. While isothermal techniques are useful to determine kinetic parameters, their implementation is time-consuming. Nonisothermal techniques provide faster means to obtain the necessary kinetic information. The most commonly applied non-isothermal technique is the so-called Friedman method [Friedman, 1963], where the logarithm of the rate constant, k, is plotted as a function of the inverse temperature. The rate constant, k, is calculated from the equation

dw dt

= k ( w f − w(t) )

(4.5)

where w is the sample mass at time t, and wf is the final sample mass. The rate constant, k, can be given by the Arrhenius equation: k = k 0 exp(E/RT)

(4.6)

The parameters k0 (pre-exponential factor) and E (activation energy) can be determined from the linear region(s) of the plot of ln(k) versus 1/T. A drawback of this method is the fact that it introduces a bias in the values of k0 and E when the reaction is characterised by a distribution of activation energies [Braun and Burnham, 1987]. In such a case, the Friedman method is not able to differentiate between the effect of the distribution and the effect of the magnitude of the mean activation energy, and gives an erroneous value for the mean value of E, E0. This is usually lower than the “real” value. Another non-isothermal method of determining the kinetic parameters involves the measurement of the temperature at which the rate of volatile evolution is maximal, Tmax (see [Teng et al., 1995] and [Van Heek and Juentgen, 1968]). This technique is used to process TG-FTIR data. The method has been shown to be applicable to the determination of E0 and an approximate value of k0 even for wide distributions of activation energies [Braun and Burnham, 1987]. In a typical sequence of experiments, thermal-decomposition rates are measured at different heating rates. The relationship between the heating rate M and the value of Tmax is then given by the following equation, for evolution of a species specified as peak i:

⎛ M ⎞ ⎛ k i0approx R ⎞ E0 ln ⎜ 2 ⎟ = ln ⎜ ⎟⎟ ⎜ ⎜ ⎟ E ⎝ ⎠ ( RTmax ) ⎝ T max ⎠ in which

(4.7)

133

M = dT/dt

(4.8)

from which the kinetic parameters ki0 and E can be determined graphically (‘Kissinger plot’) [Teng et al., 1995]. While the value of E is accurately determined even for wide distributions, the value of ki0 usually requires a slight adjustment, which is typically within a factor of 2. The width of the distribution, σ, can then be determined from the width of the peak representing the rate of weight loss, as follows by an empirical relation:

σ =−

1.1 0.66 − 30 + 2.88 ρ − 1.12 3

ρ

(4.9)

ρ

where

⎛ 52.4 ⎞⎛ ∆Tσ ⎞ (4.10) ⎟⎜ ⎟ ⎝ E 0 ⎠⎝ ∆T0 ⎠ ∆Tσ is directly obtained from the measured reaction rate profile as full width at half height. ∆T0 can be calculated from the approximate rate equation by iterative solution for the two values of T at which

ρ =⎜

dx i 1 ⎛ dx i ⎞ (4.11) = 2 ⎜⎝ dt ⎟⎠ dt Tmax in which xi is the fraction unreacted of a species from a precursor pool, released as a single peak and −

0 2 ⎡ -E dx i ⎧ -E ⎫ k RT ⎤ ≈ k i0 exp ⎢ 0 − exp ⎨ 0 ⎬ i ⎥ dt ⎩ RT ⎭ ME 0 ⎦ ⎣ RT

(4.12)

The adjustment of the pre-exponential factor taking into account σ is given as empirical relation:

{

(

)}

⎤ k i0 = k i0 approx ⎡1 − 0.4 1 − exp −σ 2.5 ⎥⎦ ⎢⎣

(4.13)

In principle, the ratio ∆Tσ/ ∆T0 depends on heating rate, but this dependency is comparatively small over the range of most laboratory experiments. Equation (4.9) was developed from experiments performed under heating rate conditions of 1 K/h, 1 K/min and 10 K/min. The value of s will deviate from the real value by about 10% for each 50-fold variation in heating rate [Braun & Burnham, 1987]. In the Tmax method, some difficulties can be encountered when peaks are not well resolved; in such cases, substantial shifts in Tmax can occur. However, the same problem arises when using the Friedman method, unless deconvolution of the peaks is attempted [Kim et al., 1995]. Another limitation is associated with the presence of small, multiple maxima superimposed on a broader peak, i.e., when the assumption of the first-order kinetics is not fully supported. In this case, again, the applicability of both the Tmax and Friedman methods is reduced. The exact value of Tmax may also be difficult to determine for large broad peaks. Once the values of ki0, E and σ have been determined, the sizes of precursor pools, which is the yield of species i at a released peak, for individual species are determined by adjusting the simulated peak heights so that they best fit the TG-FTIR data. If necessary, minor adjustments to E, k0,i, and σ are made to improve the fit to the data. At the end of the abovementioned procedure, the following information is available for the evolution of each precursor pool (peak): ki0, E, σ and the pool size, i.e., the concentration of the precursor species. In general, each volatile species (light gases, tars etc.) may evolve as one or more peaks, and, accordingly, a single pool or multiple precursor pools are used in the simulations. Now basically the model input for the FG-DVC model, as described in chapter 2, with de-emphasized DVC part [Chen et al., 1998] can be determined. In the present study, an attempt was made to employ the Tmax method, but limited data resolution, especially at the highest applied heating rate, 100 K/min, caused a significant uncertainty in the determination of Tmax. For this reason, k0,i, E, and σ values were fitted into experimental data using a trial-and-error approach. Care was taken to ensure that pre-exponential factors were consistent with the well-known transition-state theory (ki0 should have a value in the range: 1011 – 1016 s-1).

134

4.4.3 TG-FTIR analysis results and derived kinetic parameters for miscanthus Table 4.11 and figure 4.34 show the yields of compounds quantitatively analysed by FTIR under TG pyrolysis conditions with different heating rates. Details of the method and conditions were described in chapter 3. An analysis and discussion of the results is given in §4.4.6, where results for all fuels are compared. Table 4.11 Measured yields (mass%) of compounds during TG-FTIR of miscanthus. Heating rate [K/min] Ash (ar) Moisture (ar) Volatile Matter (daf) Char (daf) Tar (daf)* CH4 (daf) H2O (pyr.) (daf) CO (daf) CO2 (daf) C2H4 (daf) HCN (daf) NH3 (daf) Isocyanic Acid (HNCO) (daf) COS (daf) SO2 (daf) Formaldehyde (CH2O) (daf) Acetaldehyde (CH3CHO) (daf) CH3OH (daf) Formic Acid (HCOOH) (daf) Acetic Acid (CH3COOH) (daf) Phenol (C6H5OH) (daf) Acetone (CH3OCH3) (daf) *

10

30

100

1.5 5.0 80.2 19.8 17.72 1.18 17.65 8.56 10.41 0.52 0.22 0.13 0.20 0.31 0.00 1.21 10.21 1.42 2.12 4.25 2.31 1.80

3.9 2.5 79.4 20.6 22.13 1.14 18.57 7.15 9.43 0.27 0.14 0.10 0.11 0.16 0.00 1.12 9.02 1.13 1.66 3.59 1.82 1.86

0.0 6.0 81.9 18.1 27.10 1.16 20.64 5.65 9.11 0.13 0.11 0.11 0.07 0.11 0.00 1.16 7.70 1.07 0.37 3.35 1.99 2.10

part of the volatile matter, so calculated by difference

Yield (mass%,daf)

30

25

10 K/min 30 K/min

20

100 K/min

15

10

5

M

Ta r( by

di

C ffe har r e e W tha nc Ca ate ne e) ( r rb on (H CH4 2O ) M Ca rb ono , py on r x . D ide ) io ( C x H yd Eth ide O) ro ge ylen (CO n Cy e (C 2) an 2H Is Am ide 4) oc m (HC y Ca anic onia N) rb on Aci (NH d yl ( H 3) Su Sulp N CO lf h ) Fo ur D ide rm io (CO A x ce alde ide S) ta ld hyd (SO eh e( 2) y C M de ( H2 et C O Fo ha H3 ) rm nol CH i A ce c A (CH O) tic ci 3 A d (H OH ci CO ) d Ph (CH OH e 3 A ce nol CO ) to ne (C6 OH (C H5 ) O H 3C H) O CH 3)

0

Figure 4.34 Product yields during pyrolysis of miscanthus at different heating rates.

135

Table 4.12 Kinetic parameters and precursor pool size (yields) for main and N-species for pelletised Dutch miscanthus giganteus. Species *

ki0 (s-1)

E/R (K)

σ/R (K)

1-CO2 2-CO2 3-CO2 1-CO 2-CO 3-CO 1-H2O 2-H2O 1-CH4 2-CH4 3-CH4 1-Tars 1-HCN 2-HCN 1-HNCO 2-HNCO 1-NH3 2-NH3 1-C2H4 1-CH3OH 2-CH3OH 1-CH2O 2-CH2O 1- CH3CHO 1- HCOOH 1- CH3COOH 1- C6H5OH 1- CH3OCH3

5.20E+11 5.20E+11 2.80E+12 5.20E+11 6.50E+12 2.80E+12 5.20+E11 4.40E+12 6.50E+11 6.50E+11 3.10E+12 5.20E+11 2.80E+12 2.80E+12 2.80E+12 2.80E+12 5.20E+11 5.20E+11 2.80E+12 5.20E+11 2.30E+11 5.20E+11 5.20E+11 5.20E+11 5.20E+12 5.20E+11 7.40E+18 5.20E+12

16300 19500 24830 19550 26300 35930 19100 24000 19500 22300 26900 19700 20530 28000 16930 25930 18600 25000 24930 17500 19930 17100 19500 19500 20500 18700 29100 20500

900 500 900 400 3200 3300 400 2500 600 500 1600 500 1900 3100 900 2800 800 2500 800 500 400 500 200 300 500 900 2100 800

Yield (wt. % daf) 0.98 7.50 0.93 3.90 1.85 1.40 14.14 4.50 0.15 0.23 0.76 21.34 0.07 0.07 0.02 0.09 0.07 0.03 0.21 0.20 0.95 0.33 0.82 9.02 1.66 3.59 2.03 1.86

* Numbers before species names indicate precursor-material sources (pools), for which the numbering increases with increased stability. Some species evolve as a single peak (one source), the evolution of other species is bimodal or trimodal (two or three sources).

A set of TG-FTIR data as well as FG-DVC model fits to these data using the input files for miscanthus mentioned in table 4.12 are presented in figures 4.35a – c. These data are for a heating rate of 30 K/min, measurements at 10 and 100 K/min are presented in Appendix 4. For all the experiments, the balance curve is plotted as a function of time, with a thermocouple measured temperature given on the primary axis (left). Evolution and weight curves are presented for water, carbon containing gas species and nitrogen compounds evolving from the samples. The evolution curves are given in “weight percent per minute” (primary axis, left), while product yield curves, which are the integral of the evolution curves and show the total amount of the species evolved, are presented in the units “weight percent” (secondary axis, right).

136

Water

600

60

400

40

200

20

Temp. (C)

80

0

0 0

50

7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

20.00 15.00 10.00 5.00 0.00 25

100

35

Time (min)

4.00 2.00 0.00

0.10 0.05 0.00 25

35

45

Time (min)

Carbon Monoxide

Ethylene

0.50 0.00 35

45

Time (min)

55

55

0.30

0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0.25 0.20 0.15 0.10 0.05

Yield (wt%)

1.00

Yield (wt%)

1.50

Rate (%/min)

Time (min)

8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00 25

0.15

55

2.00

Rate (%/min)

Rate (%/min)

6.00

1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00

0.20

Yield (wt%)

8.00

Yield (wt%)

Rate (%/min)

10.00

45

55

Methane

3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 35

45

Time (min)

Carbon Dioxide

25

Yield (wt%)

100

800

Weight Loss %

1000

Rate (%/min)

Temperature & Weight Loss

0.00 25

35

45

55

Time (min)

Figure 4.35a. Weight loss curve and main components; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus Giganteus pellets. Heating rate: 30 K/min. Lines without markers: FGDVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

137

Methanol

1.4

0.8

0.2

0.6 0.4

0.1

0.2

0

0 35

45

Time (min)

Acetaldehyde

6

25

55

10

6

3

4

2

2

1 0 35

45

1.5

0.6

1

0.4

0.5

0.2 0

55

0 25

35

Time (min)

0.6 0.4 0.2 0 45

Rate (%/min)

Rate (%/min)

0.8

0.8

1.5

0.6

1

0.4

0.5

0.2 0

0 25

55

35

45

55

Time (min)

Time (min)

Phenol

0.6

2.5

0.5

2

0.4

1.5

0.3

1

0.2

0.5

0.1 0

Yield (wt%)

Rate (%/min)

2

0 25

35

45

55

Time (min)

Figure 4.35b. Oxygenated hydrocarbons; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus Giganteus pellets. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

138

Yield (wt%)

1

35

55

Acetone

1

Yield (wt%)

4 3.5 3 2.5 2 1.5 1 0.5 0

1.2

25

45

Time (min)

Acetic Acid

1.4

2

0.8

0 25

55

Yield (wt%)

8

4

45

Formic Acid

1

Yield (wt%)

5

35

1.4 1.2 1 0.8 0.6 0.4 0.2 0

Time (min)

Rate (%/min)

25

Rate (%/min)

Rate (%/min)

1

0.3

Formaldehyde

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Yield (wt%)

1.2

0.4

Yield (wt%)

Rate (%/min)

0.5

0.010 0.005 0.000 25

35

45

0.014 0.012 0.010 0.008 0.006 0.004 0.002 0.000

0.12 0.10 0.08 0.06 0.04 0.02 0.00 25

55

Time (min)

Yield (wt%)

0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00

0.015

Rate (%/min)

Isocyanic Acid (HNCO) Yield (wt%)

Rate (%/min)

Hydrogen Cyanide

35

45

55

Time (min)

0.025

0.12

0.020

0.10 0.08

0.015

0.06

0.010

0.04

0.005

0.02

0.000

Yield (wt%)

Rate (%/min)

Ammonia

0.00 25

35

45

55

Time (min)

Figure 4.35c. Nitrogen species; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus Giganteus pellets. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

139

4.4.4 TG-FTIR analysis results and derived kinetic parameters for wood (Labee “A quality” energy pellets) Table 4.13 and figure 4.36 show the yields of compounds quantitatively analysed by FTIR under TG pyrolysis conditions with different heating rates. Details of the method and conditions were described in paragraph 3.4. An analysis and discussion of the results is given in §4.4.6, where results for all fuels are compared. Table 4.13 Measured yields of compounds during TG-FTIR of crushed wood pellets Heating rate [K/min] Ash (ar) Moisture (ar) Volatile Matter (daf) Char (daf) Tar (daf)* CH4 (daf) H2O (pyr.) (daf) CO (daf) CO2 (daf) C2H4 (daf) HCN (daf) NH3 (daf) Isocyanic Acid (HNCO) (daf) COS (daf) SO2 (daf) Formaldehyde (CH2O) (daf) Acetaldehyde (CH3CHO) (daf) Methanol (CH3OH) (daf) Formic Acid (HCOOH) (daf) Acetic Acid (CH3COOH) (daf) Phenol (C6H5OH) (daf) Acetone (CH3OCH3) (daf) *

10 0.0 5.6 86.2 13.8 37.39 1.27 10.59 7.48 5.90 0.15 0.08 0.00 0.11 0.18 0.17 3.67 9.28 0.99 2.68 2.78 1.46 2.06

30 0.0 6.3 86.2 13.8 37.81 1.30 16.01 6.77 5.08 0.23 0.09 0.02 0.02 0.11 0.12 2.93 8.11 0.74 1.73 2.37 0.65 2.15

100 0.0 6.6 86.2 13.8 47.53 1.15 13.64 5.10 4.88 0.09 0.05 0.03 0.05 0.00 0.40 3.15 3.97 0.64 0.82 2.28 0.73 1.63

part of the volatile matter, so calculated by difference

50 45

Yield (mass%,daf)

40

10 K/min

35

30 K/min

30

100 K/min

25 20 15 10 5

Ta r(

by

di Ch f M fere ar et n W han ce) e( Ca ate C r rb on (H2 H4 ) O M ,p Ca o rb nox yr. on ) i D de ( io C xi O H ) de yd Et (C ro hy O ge len 2 n Cy e ( C ) an 2H id 4) Is Am e (H oc m C y N o ) Ca anic nia (N rb A ci on d H3) yl S u ( HN lp Su h i CO lf ) d Fo ur D e (C rm io O x S) a A ce lde ide hy (S ta ld eh de ( O2) CH yd M e (C 2O et ) H 3 Fo han o l CHO rm (C ic A H ) A ce 3O ci tic d A (H H) ci C d ( C OO H H Ph 3 en CO ) A o ce to l (C OH ne ) 6 (C H5O H H 3C ) O CH 3)

0

Figure 4.36 Product yield during pyrolysis of Labee wood pellets at different heating rates.

140

Table 4.14 Kinetic parameters and precursor pool size (yields) for main and N-species for Labee wood pellets. Species* 1-CO2 2-CO2 3-CO2 1-CO 2-CO 3-CO 1-H2O 2-H2O 1-CH4 2-CH4 3-CH4 1-Tars 1-HCN 2-HCN 1-HNCO 2-HNCO 1-C2H4 1-CH3OH 2-CH3OH 1-CH2O 2-CH2O 1- CH3CHO 1- HCOOH 1- CH3COOH 1- C6H5OH 1- CH3OCH3

k0,i (s-1)

E/R (K)

σ/R (K)

5.20E+11 5.20E+11 2.80E+12 5.20E+11 6.50E+12 2.80E+12 5.20+E11 4.40E+12 6.50E+11 6.50E+11 3.10E+12 5.20E+11 2.80E+12 2.80E+12 2.80E+12 2.80E+12 2.80E+12 5.20E+11 2.30E+11 5.20E+11 5.20E+11 5.20E+11 5.20E+12 5.20E+11 7.40E+18 5.20E+12

18300 20500 24830 20300 26300 35930 19900 26000 19200 22500 27700 20500 21230 27000 17930 28930 25930 17800 21030 18400 20500 20400 20800 19300 32500 20900

900 500 900 500 3200 3300 1200 2500 800 800 2100 500 1900 2500 900 3800 800 800 700 500 200 500 800 900 2500 800

Yield (wt. % daf) 0.95 3.10 0.90 1.90 3.05 1.80 11.94 3.00 0.05 0.60 0.60 38.00 0.05 0.04 0.02 0.02 0.20 0.40 0.39 1.20 2.10 7.00 1.70 2.30 0.75 1.90

* Numbers before species names indicate precursor-material sources (pools), for which the numbering increases with increased stability. Some species evolve as a single peak (one source), the evolution of other species is bimodal or trimodal (two or three sources).

A set of TG-FTIR data as well as FG-DVC model fits to these data using the input files for Labee wood pellets mentioned in table 4.14 are presented in figures 4.37a – c. These data are for a heating rate of 30 K/min; measurements at 10 and 100 K/min are presented in Appendix 4.

141

Temperature & Weight Loss

60

400

40

200

20

0

0 50

100

25

Carbon Dioxide

Methane 6.00 5.00 4.00 3.00 2.00 1.00

35

45

0.15 0.10 0.05 0.00

55

25

55

0.60 0.40 0.20 0.00 55

Rate (%/min)

0.80

0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0.25 0.20 0.15 0.10 0.05

Yield (wt%)

8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00 0.00

1.00

Yield (wt%)

Rate (%/min)

45

Ethylene

1.20

Time (min)

35

Time (min)

Carbon Monoxide

45

1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00

55

0.20

Time (min)

35

18.00 16.00 14.00 12.00 10.00 8.00 6.00 4.00 2.00 0.00

0.25

0.00

25

45

Time (min)

1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 25

35

Time (min)

Yield (wt%)

Rate (%/min)

0

Rate (%/min)

Temp. (C)

600

3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00

Yield (wt%)

80

Yield (wt%)

100

800

Weight Loss %

1000

Rate (%/min)

Water

0.00 25

35

45

55

Time (min)

Figure 4.37a. Weight loss curve and main components; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

142

35

45

0

Rate (%/min) Rate (%/min)

3 2 1.5 1 0.5 0

35

45

0 35

45

Time (min)

55

Formic Acid

2

0.6

1.5

0.5 0.4

1

0.3 0.2

0.5

0.1

0 35

Time (min)

55

Acetone

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

2.5 2 1.5 1 0.5 0

25

55

45

Yield (wt%)

2.5

25

0.5

25

Yield (wt%)

Rate (%/min)

Acetic Acid

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

1

0.2

0

55

Rate (%/min)

Rate (%/min)

1

Time (min)

1.5

0.4

Yield (wt%)

2

45

2

0.6

0.7

Yield (wt%)

9 8 7 6 5 4 3 2 1 0

3

35

2.5

0.8

25

4

25

3

0

55

Acetaldehyde

3.5

1

Time (min)

5

Formaldehyde

1.2

Yield (wt%)

25

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Yields (wt%)

Rate (%/min)

Methanol

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

35

45

55

Time (min)

Time (min)

Phenol

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0.15 0.1 0.05 0 25

35

45

Yield (wt%)

Rate (%/min)

0.2

55

Time (min)

Figure 4.37b. Oxygenated hydrocarbons; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TGFTIR measurements for yields (secondary axis) and rates (primary axis).

143

Isocyanic Acid (HNCO) 0.006

0.08

0.005

0.008

0.06

0.006

0.04

0.004 0.002

0.02

0.000

0.00 25

35

45

Rate (%/min)

0.10

0.010

0.040 0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000

0.004 0.003 0.002 0.001 0.000

55

25

35

Time (min)

45

Yield (wt%)

0.012

Yield (wt%)

Rate (%/min)

Hydrogen Cyanide

55

Time (min)

Ammonia 0.020

0.005

0.015

0.004 0.003

0.010

0.002

0.005

0.001 0

Yield (wt%)

Rate (%/min)

0.006

0.000 25

35

45

55

Time (min)

Figure 4.37c. Nitrogen species; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

4.4.5 TG-FTIR analysis results and derived kinetic parameters for brown coal Table 4.15 and figure 4.38 show the yields of compounds quantitatively analysed by FTIR under TG pyrolysis conditions with different heating rates. Details of the method and conditions were described in paragraph 3.4. An analysis and discussion of the results is given in §4.4.6, where results for all fuels are compared. Table 4.15 Measured yields of compounds during TG-FTIR of Hambach brown coal. Heating rate [K/min] Ash (ar) Moisture (ar) Volatile Matter (daf) Char (daf) Tar (daf) CH4 (daf) H2O (pyr.) (daf) CO (daf) CO2 (daf) C2H4 (daf) HCN (daf) NH3 (daf) COS (daf) SO2 (daf) *

144

10 2.6 11.1 53.1 46.9 11.59 1.71 11.12 16.05 11.77 0.25 0.23 0.07 0.21 0.06

30 4.2 10.6 53.3 46.7 13.76 1.67 12.79 12.56 11.80 0.31 0.16 0.08 0.16 0.00

part of the volatile matter, so calculated by difference

100 4.2 10.6 53.3 46.7 16.80 1.55 12.32 10.05 11.92 0.32 0.07 0.09 0.15 0.01

50 45

Yield (mass%,daf)

40

10 K/min

35

30 K/min

30

100 K/min

25 20 15 10 5

di ffe re nc e) M et ha ne (C H W 4) at er (H 2O Ca rb ,p on yr .) M on ox id Ca e( rb CO on ) D io xi de (C O 2) Et hy l en H yd e( ro C2 ge H n 4) Cy an id e( H CN A ) m m on Ca ia rb (N on H yl 3) Su lp hi de (C Su O lfu S) rD io xi de (S O 2)

Ta r( by

Ch ar

0

Figure 4.38 Product yield during pyrolysis of Hambach brown coal at different heating rates. Table 4.16 Kinetic parameters and precursor pool size (yields) for main and N-species for Hambach brown coal. Species*

k0,i (s-1)

E/R (K)

σ/R (K)

Yield (wt. % daf)

1-CO2 2-CO2 3-CO2 1-CO 2-CO 3-CO 4-CO 1-H2O 1-CH4 2-CH4 3-CH4 1-Tars 1-C2H4 1-HCN 1-NH3 2-NH3 3-NH3

5.20E+11 5.20E+11 2.80E+12 5.20E+11 6.50E+12 2.80E+12 2.80E+12 4.40+E12 6.50E+11 6.50E+11 3.10E+12 5.20E+11 2.80E+12 1.80E+12 5.20E+11 5.20E+11 2.80E+11

16300 21100 28830 22300 31200 34330 39930 24200 23000 26100 32700 22100 25130 23230 22400 24900 27230

900 2100 1900 2500 1800 400 2200 3500 800 1100 1800 1200 1200 4100 900 900 200

0.40 10.60 1.70 3.30 5.65 2.50 1.20 12.30 0.44 1.10 0.11 14.00 0.31 0.17 0.02 0.02 0.03

* Numbers before species names indicate precursor-material sources (pools), for which the numbering increases with increased stability. Some species evolve as a single peak (one source), the evolution of other species is bimodal or trimodal (two or three sources).

A set of TG-FTIR data as well as FG-DVC model fits to these data using the input files for Hambach brown coal mentioned in Table 4.16 are presented in Figure 4.39. These data are for a heating rate of 30 K/min, measurements at 10 and 100 K/min are presented in Appendix 4.

145

Temperature & Weight Loss

T (°C)

80

600

60

400

40

200

20

0 0

20

60

Time (min)

80

Carbon Dioxide

2

14 10 8

1

6 4

0.5

2 0 15

20

25

30

6 4 2 0 15

0.1 0.05 0 15

0.6

6

0.4

4

0.2

2

0

0 30

15

20

25

Time (min)

30

35

30

0.35 0.3 0.25 0.2

0.04

0.15 0.1

0.02

0.05

0

0 10

0.18 0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 35

Ethylene

15

20

25

Time (min)

30

35

Ammonia

0.014

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0.012 0.01 0.008 0.006 0.004 0.002 0 10

15

20

25

30

35

Time (min)

Figure 4.39. Weight loss curve and selected species; comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of Hambach brown coal. Heating rate: 30 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

146

Yield (wt%)

10

25

0.06

35

Hydrogen Cyanide

0.018 0.016 0.014 0.012 0.01 0.008 0.006 0.004 0.002 0

Rate (%/min)

0.08

Rate (%/min)

Rate (%/min)

12

Yield (wt%)

Rate (%/min)

0.1

8

Time (min)

20

Yield (wt%)

14

Yield (wt%)

0.8

25

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

0.2

10

10

20

35

Time (min)

1

15

Time (min)

30

0.15

35

Carbon Monoxide

10

25

Methane

Time (min)

1.2

20

0.25

0 10

8

Yield (wt%)

1.5

10

0.3

12

14 12

10

Yield (wt%)

Rate (%/min)

40

0 100

Water

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Yield (wt%)

800

Char mass(%)

100

Rate (%/min)

120

Rate (%/min)

1000

4.4.6 TG-FTIR analysis discussion From figures 4.34, 4.36 and 4.38 it can be observed that char yields for all fuels studied are practically constant at different heating rates. This relatively constant char yield with increasing heating rates applied, implies that cross-linking reactions are relatively unimportant, under the process conditions used in this experimental study. It is also possible that the major factors determining char formation, i.e. bond breaking and cross-linking, are in balance with each other [de Jong et al., 2003]. Tar yields for brown coal (ca. 12-17 wt.%) are observed to be generally lower than those for biomass (17-50 wt.%). Also it can be seen from these figures that tar yields increase as the heating rate increases. For the biomass samples, this happens at the expense of the yields of lighter gaseous species mostly CO, CO2, acetaldehyde, methanol, formic acid, acetic acid and ethylene. For brown coal, the increase in tar yields associated with increasing heating rate is accompanied by a corresponding decrease in the yield of CO. The observed tar yield increase can be explained by differences in kinetic effects associated with the release of tar and light species. Apparently, at higher heating rates, there is not enough time available for the evolution of light species and thus, tar fragments leaving the biomass carry with them precursor material that could potentially have resulted in light gas compound formation. This reaction behaviour during pyrolysis causes the tar compounds to possess a comparatively high molecular weight, which is manifested in higher tar yields. The heating rate does not have a significant effect on the yields of methane, formaldehyde and acetone. This behaviour suggests that the species mentioned may have different precursors than CO, CO2 and acetaldehyde. Besides an increase in tar yield an increase in water release is generally also observed, though it is less pronounced for wood and brown coal, compared to miscanthus. This could indicate that during the release of tar fragments water is formed, which stabilizes the char left. The heating rate does not have a significant effect on the yields of methane (CH4), formaldehyde (CH2O) and acetone (CH3OCH3). This behaviour suggests that these species may have different precursors than CO, CO2 and acetaldehyde. Furthermore, a potential relation between char formation and release of CH4, CH2O and CH3OCH3 can be supposed since their observed relatively constant yields when varying the heating rates. These suppositions, however, have to be further investigated. The effect of the heating rate is more complex (e.g., a maximum or a minimum is observed) with regard to ethylene (in the case of wood pellets), phenol and water (for wood pellets). Pyrolysis of miscanthus produces higher yields of char and lower yields of volatile matter as compared to wood. This different behaviour may be explained by the different biochemical composition of the two biomass types. In biomass, lignin shows generally a relatively large contribution to the production of fixed carbon than other main constituents such as cellulose or hemi-cellulose. The lignin content of wood pellets (pine wood) is about 23 wt.% [Miller & Bellan, 1997]. [Klass, 1998] presents values between 10 and 40 wt.% for various herbaceous species such as bagasse and straw, while a value of 21 wt.% is given in the ECN on-line database Phyllis [Phyllis, 2003] for miscanthus. A value of ca. 20 wt.% is reported by [Roll, 1994]. However, these chemical analysis data have not been determined for our miscanthus giganteus, thus making it unsure whether or not the differences in char yields can be related unequivocally to the diverse lignin content of the biomass types. Differences in mineral constituents with respect to amount and composition could also be the reason for this different char forming behaviour. Of the nitrogen species, the HCN and HNCO yields decrease with increasing heating rate, whereas NH3 shows a less pronounced decrease. This can be due to a stronger primary relation with tar released by HCN and HNCO. Comparable product yields of HCN, HNCO and NH3 are observed, although comparison of the data presented in tables 4.11, 4.13 and 4.15 does not reveal a consistent pattern. Studies accomplished for biomass conversion are presented by [Leppälahti & Koljonen, 1995], [Li & Tan, 2000b] and [Glarborg et al., 2001], showing that at low heating rates NH3 is generally the dominant N-product.

147

The pyrolysis experiments done in this study were performed at low heating rates, obtaining results which are in disagreement with the mentioned references, because HCN yields were found to be mostly higher than those of NH3 (see tables 4.11, 4.13 and 4.15 for nitrogen compound yields for miscanthus, wood and brown coal, respectively). According to [Li & Tan, 2000b], HCN formation may go to completion much more rapidly than that of NH3, which is found to be the main N-product when secondary reactions are considered in conversion processes characterised by longer residence times. This could partly explain the high concentrations of HCN found in this study. In fact, TG-FTIR experiments were carried in such a way to minimise secondary reactions. From the TG-FTIR pyrolysis curves, presented in the figures 4.35, 4.37 and 4.39, it can be observed that all nitrogen-containing species generally evolve over a wide range of temperatures (100-900 °C). In contrast, methanol, formaldehyde, acetaldehyde, formic and acetic acids, as well as a number of minor species, evolve at relatively low temperature (100-400 °C). CH4, C2H4 and for a large part also CO, CO2 and water form a class of compounds which evolve at moderate temperature. It can be observed in the graphs mentioned that the majority of the evolved species showed multi-modal evolution patterns. This indicates the presence of different precursors – functional groups – within the fuel samples. Relatively few species evolved as a single peak (e.g. acetaldehyde) The release curves of the low molecular weight (non-tar) products from TG-pyrolysis of miscanthus, shown in figure 4.35 a-c have a striking similarity with the release curves of wheat straw [Bassilakis et al., 2001], which implies that miscanthus giganteus used in this work is a good model component for an agricultural waste material as straw. The TG-FTIR analysis reveals great complexity in the evolution patterns for the different fuels investigated here. The interpretation is by no means straightforward. This technique provides valuable input for predictive modelling of biomass thermal conversion processes. The kinetic parameters for the FG-DVC biomass model derived from this analysis therefore will be used in model simulations for experimental pyrolysis at higher heating rates to validate this model. This will be subject of the next paragraph.

148

4.5 Experimental results of heated grid pyrolysis tests

4.5.1 The heated grid experimental programme The fast pyrolysis experiments using a heated grid were focused on determining the pyrolysis product yield of CO, CO2 and NH3 as a function of temperature at a high heating rate. Because of time limitations only atmospheric experiments with miscanthus were carried out. The main focus of the experiments was the determination of the CO and CO2 release as a function of the end temperature. Very limited experience had been gained in the past with an NH3 diode laser. Attempts of [Beuken, 2001] to detect NH3 release from miscanthus failed. The reason for this was probably the presence of condensing water in the reactor which absorbs the NH3. Attempts to detect NH3 in our work by introduction of a calibration gas into the reactor have failed since no absorption peak could be found. The reason for this could be that the concentration of NH3 in the reference cell was too low or that incorrect laser settings were used. Miscanthus pellets were milled and subsequently sieved. The fraction with sieve size 38-63 µm was kept at 20 °C and 60% relative humidity. The mass of biomass between the folded screen needed to be determined accurately to be able to quantify yields of gaseous species released. Therefore, batches of approximately 50 mg of the miscanthus powder were pressed at 1130 MPa under vacuum to tablets of 13.0 mm diameter and ca. 0.5 mm thickness. The tablets were cut into very small pieces with a mass between 0.30 and 1.00 mg. The mass of the pieces was measured with a balance of an accuracy of 10-3 mg. Experiments with both the finely grounded powder and the small cut tablets were performed. 4.5.2 Miscanthus pyrolysis results Figure 4.40 shows that the yield of CO is strongly correlated with the final temperature in the range 1050-1400K, for a heating rate in the range of 250-320 K/s and a grid hold time of 6 s. The mass yield of CO produced by the first miscanthus sample varies from approximately 4.0% at 1050K to 13.0% at 1350K. The estimated error made in the experiments is within -16% and +21%. 1.00

16.00 12.00 8.00

R2 = 0.87

4.00 0.00 1000

1100

1200

1300

1400

Temperature (K) Miscanthus Linear (Miscanthus)

Figure 4.40 The CO yield (wt%) of pyrolysed miscanthus is strongly correlated to the final temperature in the range 1050-1400K, ambient pressure, heating rate 250-320K/s.

Wt% CO2 released (daf)

Wt% CO released (daf)

20.00

0.80 0.60 2

R = 0.00

0.40 0.20 0.00 1000

1100

1200 1300 1400 Temperature (K) Miscanthus Linear (Miscanthus)

Figure 4.41 The CO2 yield (wt%) of pyrolysed miscanthus is not correlated to the final temperature in the range 10501400K, ambient pressure, heating rate 250-320K/s.

Contrary to CO, figure 4.41 shows that the yield of CO2 is practically not correlated with temperature in the same range and for the same experimental conditions. The average CO2 mass yield was 0.59wt%.

149

The 95% pyrolysis time is defined as the time required for release of 95% of the final gaseous product yield. Figure 4.42 shows a plot of this parameter versus temperature. A clear increase in time is shown with decreasing temperature. Apparently, CO is released faster than CO2. The release of both components starts at practically the same time, as can be seen from figure 4.43. CO is released slightly earlier from powder than from pellets as can be observed in this figure. Probably a heat transfer effect plays a role in this observation. 3.0

2.0

Begin time of pyrolysis (s)

95% Pyrolysis time (s)

3.0

2

R = 0.78 2

R = 0.67 1.0 2

R = 0.78 0.0 1000

1100

CO pellets

1200 1300 Temperature (K) CO powder

1400

2.0

1.0

0.0 1000

CO2 pellets

Figure 4.42 The 95% pyrolysis-time is correlated with temperature. Deviations are caused by variation in sample mass.

1100

1200 1300 1400 Temperature (K) CO Pellets CO powder CO2 pellets

Figure 4.43 Start moment of devolatilisation of CO and CO2 from miscanthus as function of temperature.

1.5

ln(1/(95% Pyrolysis time (s))

1.0 0.5

y = -5.01x + 3.98 R2 = 0.76

y = -4.86x + 4.57 R2 = 0.83

0.0 -0.5

y = -2.90x + 1.84 R2 = 0.84

-1.0 -1.5 0.70

0.75

0.80 CO pellets

1000 / T(K)

0.85

CO powder

0.90

0.95

CO2 pellets

Figure 4.44 Arrhenius plot for the devolatilisation of CO from miscanthus pellets and powder and from CO2 pellets. The dashed lines are the least square linear fits. An Arrhenius plot is shown in figure 4.44 to determine the activation energies of the reactions. The slope of the dashed lines in this plot represents the activation energy of the decarbonylation (release of CO) and decarboxylation (release of CO2) reactions. Obviously the slopes of the two trend lines for CO release during pyrolysis from miscanthus pellets and powder are practically equal to each other since the reaction mechanism -decarbonylation- is similar. The interception of the line figure 4.44 with the vertical axis represents the value of the pre-exponential in the Arrhenius equation for devolatilisation. The observed activation energy for both reactions is shown in table 4.17.

150

Table 4.17 Activation energy calculated from Arrhenius plot for carbonylation and carboxylation reactions.

Species CO CO2

Reaction

Ea/R (K)

Decarbonylation Decarboxylation

4.94·103 2.90·103

Activation energy Ea (kJ/mol) 41.0 24.1

These values for the activation energies determined at a heating rate of ca. 300 K/s, are significantly lower than the values found for the TG-FTIR pyrolysis experiments (at 10-100 K/min) used to obtain the kinetic parameters for the FG-DVC model. The Ea/R values of the TG-FTIR experiments found in [de Jong et al., 2003] and also described in §4.4.3 are 19.6·103 up to 35.9·103 K for CO (163 <Ea< 298 kJ/mol) and 16.3·103 up to 24.8·103 K for CO2 (136 <Ea< 206 kJ/mol). This would imply that decarbonylation and decarboxylation reactions at high heating rates are initiated at lower temperatures. [Stubington & Aiman, 1994] obtained activation energy values for CO release of 59.5-72 kJ/mol, whereas for CO2 values in the range of 46.5-48.6 kJ/mol were measured for wire-mesh pyrolysis at 1000 K/s of bagasse. These values are comparatively close to the ones measured in this research work. For bagasse pyrolysis in a heated grid reactor, [Drummond & Drummond, 1996] also mention lower values of total fuel decomposition as compared to slow heating rate equipment, like a TGA. The experimental results obtained in this research are compared to simulations with the FG-DVC biomass pyrolysis model at a heating rate of 300 K/s using the kinetic rate parameters determined by TG-FTIR analysis of miscanthus (see §4.4.3) and to results of the limited well comparable experiments published in accessible literature. The comparison is shown in figures 4.45 and 4.46 for CO and CO2 yields, respectively. Most experimental research focuses on pyrolysis at low heating rates (up to 100 K/min) and/or low temperature (up to 900 K), e.g. liquefaction. A series of systematic flash pyrolysis experiments were carried out with sweetgum hardwood, wood lignin en cellulose at high heating rates (100-15000 K/s) in a heated grid reactor in the period 1980-1985 at MIT (Cambridge, USA), see [Hajaligol et al., 1982], [Nunn et al., 1985 a and b]. [Chen et al., 1998] have used these experimental results in the first attempt to validate the FG-DVC biomass model at high heating rates. This comparison is also shown in figure 4.45 and 4.46. A discussion follows in §4.5.3.

10.0 CO2 yield (wt% daf)

CO yield (wt% daf)

20.0 15.0 10.0 5.0 0.0 500

700

900 1100 Temperature (K)

Miscanthus 300K/s (1) Hardwood 1000K/s (2) Wood Lignin 1000K/s (3)

1300

1500

Misc. simulation 300K/s (4) Hardwood sim. 1000K/s (5)

Figure 4.45 Comparison of heated grid pyrolysis experiments and FG-DVC model simulations. CO yield in wt% (daf basis) as function of final (or peak) pyrolysis temperature.

8.0 6.0 4.0 2.0 0.0 500

700

900 1100 1300 Temperature (K)

Miscanthus 300K/s (1) Hardwood 1000K/s (2) Wood Lignin 1000K/s (3)

1500

Misc. simulation 300K/s (4) Hardwood sim. 1000K/s (3)

Figure 4.46 Comparison of heated grid pyrolysis experiments and FG-DVC model simulations. CO2 yield in wt% (daf basis) as function of final (or peak) pyrolysis temperature.

151

Remarks for figure 4.45 and 4.46: (1)

Experimental results of this work. Milled Sweetgum hardwood pyrolysis experiments from [Nunn et al, 1985a]. (3) Milled wood lignin pyrolysis experiments from [Nunn et al, 1985b]. (4) FG-DVC model simulations of pyrolysis of miscanthus (this work). (5) FG-DVC model simulations of pyrolysis of hardwood [Chen et al., 1998]. Note that the type of hardwood used in (2) differs slightly from (5). (2)

At a final pyrolysis temperature of 900°C the main product yield is compared at different heating rates and biomass feedstocks, as shown in table 4.18. In §4.5.3 the results are discussed. Table 4.18 Comparison of the pyrolysis product yields of the main species at 900°C end temperature as a function of heating rate (HR), for different biomass materials. Material

Wood pellets (1)

Method

TG-FTIR

HR (K/s)

0.17

0.5

Miscanthus pellets (1)

Lignin (2)

SG-Hard wood (2)

Cellulose (3)

1000

1000

Heated grid reactor 1.67

0.17

0.5

1.67

300

1000

CO (daf)

7.5

6.8

5.1

8.6

7.2

5.7

7.2

16.0

16.0

20.0

CO2 (daf)

5.9

5.1

4.9

10.4

9.4

9.1

0.6

3.8

5.5

3.3

H2O (daf)

10.6

16.0

13.6

17.7

18.5

20.6

-

3.8

5.0

9.2

CH4 (daf)

1.3

1.3

1.2

1.2

1.1

1.2

-

2.8

1.6

2.6

Tar (daf)

37.2

38.8

48.5

18.1

21.9

28.3

-

46

45

50

Char (daf)

13.8

13.8

13.8

19.8

20.6

18.1

-

15

8

4

Volatiles (daf)

86.2

86.2

86.2

80.2

79.4

81.9

-

-

-

-

Ash (ar)

0.0

0.0

0.0

1.5

3.8

3.8

-

-

-

-

Moisture (ar)

5.6

6.3

6.6

5.0

2.5

6.0

-

-

-

-

(1)

This research. (2) Wood lignin and Sweetgum Hardwood from [Nunn et al., 1985a,b]. (3) Cellulose from [Hajaligol,1982].

Formation of Metal Carbonyl species All infrared absorption curves of CO for the miscanthus pyrolysis experiments in the heated grid reactor show a release pattern as depicted in figure 4.47. No such observation was made for the CO2 release. Therefore, most likely this phenomenon is caused by formation of transition metal carbonyls on the stainless steel reactor wall [Slabbekoorn, 2002] and [Guo, 2004]. In average 21 wt% of the released CO disappears within 1-2 seconds after being released. The fraction of CO that disappears is correlated to the total CO yield based on the initial mass of the miscanthus samples.

152

45.0

CO disappeared (% of yield)

40.0 35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0 0.00

2.00

4.00

6.00

8.00

CO Yield (wt% of initial sample)

Figure 4.47 Typical IR absorption curve for the release of CO from miscanthus pellets. Final temperature 1240 K, heating rate 280 K/s, atmospheric pressure.

Figure 4.48 Disappeared part of the CO yield (in % of the maximum yield) as function of the total CO yield. miscanthus, 250-320 K/s, 1050-1400 K, atmospheric pressure.

Figure 4.48 shows that at a low partial pressure of CO in the reactor, the absorbed amount of CO is correlated with the partial pressure of CO in the reactor. The curve suggests that a maximum amount of CO molecules can be absorbed at higher partial pressure of CO in the reactor. This supports the hypothesis that the reactor wall absorbs parts of the CO yield. An additional set of experiments with a larger sample could clarify whether the data fits the Langmuir absorption isotherm. From the CO yield results shown in figure 4.40 and especially 4.45 it is concluded that the results are hardly affected by the absorption phenomenon, since the absorption peak value of the experiments, as e.g. shown in figure 4.47, was used to construct these figures and the values are comparable or higher than the model predictions. The absorption phenomenon is just slow enough to make detection of the final CO yield possible. The experiments with the CO reference gas are most likely not affected by the absorption phenomenon since the grid reactor was flushed with the reference gas several times before measurements were done. 4.5.3 Discussion of the results The CO2 yield of flash pyrolysis (at approximately 300 K/s, or 18000 K/min) of miscanthus is a factor 16 lower than the yield observed for slow pyrolysis (10-100 K/min) at a temperature of 900°C, as shown in table 4.18. Also, the CO2 yield is a factor 6-10 lower than found in [Nunn et al., 1985a and b] and [Hajaligol et al., 1982 and 1993] using lignin, hard wood and cellulose. For heated grid pyrolysis of dry bagasse at 1000 K/s in a dry N2 atmosphere, [Stubington & Aiman, 1994] measured a CO2 yield of approximately 3 %, which is slightly closer to the values observed in this work. [Avni et al., 1985] reported low CO2 yields, ranging from ca. 0.4 to 1%, for vacuum flash pyrolysis (600 K/s) of aspen wood (treated by steam explosion) in a heated grid. Their grid and fuel characteristics, though, were not published. Assuming that differences in biomass materials like wood, bagasse and miscanthus will result in a maximum CO2 yield difference of a factor 2, questions are raised with regard to the correctness of our results. Experiments with a stainless steel grid versus a platinum grid, chemical equilibrium calculations and thoroughly re-examination of the experimental method learn that it is highly unlikely that the set-up material or method affected the results [Slabbekoorn, 2002]. Probably the pyrolysis conditions in the grid reactor of this research have been different from the conditions in the grid reactor used by especially [Nunn et al., 1985a and b] and [Hajaligol et al., 1982 and 1993] causing such a low CO2 yield.

153

Reviewing the TG-FTIR results with wood and miscanthus, shown before in table 4.12 and 4.18, leads to the conclusion that the tar yield increases at the expense of the light gas species CO, CO2 and acetaldehyde if the heating rate increases from 10 to 100 K/min. This can be explained by differences in kinetic properties of the release of tar and light species. Tar fragments leave the reaction zone carrying with them precursor material for formation of light gases, with the accompanying effect that the tar molecules are larger and heavier. The heating rate does not significantly affect the yield of char, methane, formaldehyde and acetone, according to the TG-FTIR results alone. However, the results of [Hajaligol, 1982 and 1993] and [Nunn et al., 1985 a,b] performed at higher heating rate experiments of about 1000 K/s show no significant decrease in CO2 yield, and a factor two increase in CO yield (see table 4.18). The associated tar yield, however, is almost doubled. Apparently the increasing amount of (heavier) tar fragments, do not carry with them precursor material for CO2, the carboxyl groups. The specific precursor material for carboxyl groups must have another origin in the pyrolysis process. Considering the abovementioned, the specific precursor groups for CO2 could be formed during the primary stage of flash pyrolysis. In the experiments of [Hajaligol, 1982 and 1993] and [Nunn et al., 1985 a,b], and to a lesser extent for the experiments of [Stubington & Aiman, 1994] these precursor groups most likely further decomposed under formation of CO2. In the experiments performed in this research they most probably didn’t. These specific groups are likely to be carboxylic acids, like e.g. formic acid (HCOOH) and acetic acid (CH3COOH), released in the primary stage of flash pyrolysis from lignin and (hemi-)cellulose compounds of the biomass. Formic acid itself proceeds from carboxylic groups of uronic acid, whereas acetic acid is released by elimination of acetyl groups originally linked to the xylose unit [Demirbas, 2000]. In [Hajaligol, 1982 and 1993] and [Nunn et al., 1985 a,b] large samples of 100 mg fine powder (4588µm) were suspended on a 325-mesh folded grid of 60 x 20 mm. This corresponds with 0.083 mg/mm2. [Stubington & Aiman, 1994] applied ca. 20-35 mg fine powder (64-422 µm), resulting in an average grid load of 0.035 mg/mm2. In our research small samples of 0.3-0.8 mg were folded in a 76mesh grid of 6 x 2 mm, corresponding to an average of 0.046 mg/mm2. The fine structure of the grid described in [Hajaligol, 1982 and 1993], [Nunn et al., 1985 a,b] and [Stubington & Aiman, 1994] (325 wires per linear inch, 12.8 per mm) combined with a large sample applied by [Hajaligol, 1982 and 1993] and [Nunn et al., 1985 a,b] can explain the comparatively high CO2 yield of their experiments. Figure 4.49 illustrates the situation difference. The residence time of the carboxylic acids in the hot reaction zone in their set-up is just long enough to decompose into CO2. In our research the comparatively coarse-meshed grid allows the carboxylic acids to leave the reaction zone faster from the small samples. The relatively cold nitrogen environment ensures immediate quenching, making secondary reactions to form CO2 impossible. More grid reactor experiments at high heating rates focussing on a broader range of pyrolysis products, including carboxylic acids, are necessary to further establish this hypothesis.

Figure 4.49 Sketch of fine grid – high pulverised fuel mass vs. coarse grid – low fuel mass.

A similar hypothesis can explain the higher CO yield of [Hajaligol, 1982 and 1993], [Nunn et al., 1985 a,b] for similar temperatures compared to the results of this work. During the pyrolysis process in this research the primary tar fragments are formed and removed quickly from the reaction zone, carrying precursor material, for example ether-links, carbonyl and to a minor extent carboxyl groups, for CO with them. This results in a lower CO yield. The transition process of large primary tar fragments (e.g. levoglucosan) into smaller secondary tars (more aromatic compounds), which involves cracking of ether-links at higher temperatures accompanied by formation of CO, is expected to have contributed

154

significantly to the comparatively high CO yield in the experiments in [Hajaligol, 1982 and 1993] and [Nunn et al., 1985 a,b]. The relatively small difference in CO yield (factor 1.5-2.5 at 900°C) could be ascribed to differences in biomass material properties as well. In figure 4.45 and 4.46 the results of a FG-DVC flash pyrolysis simulation at a heating rate of 300K/s, a total residence time of 6 seconds and final temperature, are compared to the experimental results obtained in this research. The figures also show the FG-DVC simulation of wood pyrolysis [Chen et al., 1998] with experimental data of [Nunn et al., 1985a,b]. The qualitative trend is predicted correctly in all cases. However, it is difficult to predict the yield quantitatively correctly, mainly for the higher temperature range. Unfortunately, no experiments could be conducted at lower temperatures, where secondary reactions are expected to play an even smaller role. This was due to the minimum power supply requirement of the heated grid. The extrapolation that the model uses to predict the evolution of gases at high heating rates based on kinetic constants obtained from low heating rate experiments is considered to be risky. It is believed that the differences between model and experiments can be caused by the occurrence of secondary pyrolysis reactions. It is very difficult to exclude the occurrence of these reactions during the TG-FTIR experiments at low heating rates (10-100 K/min) to determine the kinetic constants, as it is almost impossible to ensure immediate quenching of the pyrolysis products. The large differences between the CO and CO2 yields (factor 1.5-2.5 and 6-10 respectively) measured during flash pyrolysis experiments at high heating rates of 300 and 1000 K/s using different experimental set ups could be explained by secondary reactions of tars as well. This shows the sensitivity of the pyrolysis reaction mechanisms to the experimental equipment in terms of not perfectly controlled time-temperature profile of the sample and the fate of pyrolysis products during the experiment. The implications for a one-stage multi-reaction model as the FG-DVC biomass model are obvious. Tars and organic carboxylic acids in the basic FG-DVC model are modelled to have one functional group for each, as shown in table 4.12 in the case of miscanthus pyrolysis. No relations are included between formation or destruction of these groups and the formation of light gases (CO, CO2, H2O, CH4, etc.) The statistical network submodel for Depolymerization, Vaporization and Crosslinking that was applied in the coal version of the FG-DVC model is de-emphasised in the biomass version because of the entirely different and incomparable chemical structure of biomass. Currently, new subroutines are being developed at AFR Inc. for the FG-DVC biomass model taking into account secondary reaction possibilities and the relation between formation of primary tars and light gases, in order to improve the performance of the model. 4.6 Conclusions and recommendations

4.6.1 Conclusions and recommendations related to PFB gasification research An experimental measurement programme involving pressurised bubbling fluidised bed gasification of miscanthus, wood and brown coal was performed at two scales of thermal input, at the 50 kWth (max) DWSA unit at IVD (University of Stuttgart) and at the 1.5 MWth (max) PFBG test rig at Delft University of Technology. No significant radial concentration profile of main and minor gaseous product constituents was observed in pressurised fluidised bed freeboard measurements performed at the large scale PFBG unit in contrast to profiles found in circulating fluidised bed gasifiers. In this respect the assumption of operation in a plug flow regime, characterised by flat concentration and velocity profiles, prevailing in the bubbling bed gasifiers appears to be sufficient.

155

The concentrations of the main gasification product gas components were comparable to the limited public literature data from other pressurised fluidised bed test rigs for the air stoichiometry values applied. Axial profiles in gas concentrations during the PFBG tests could be clearly observed for acetylene, which is related to reactions involving tar and soot precursor formation and destruction. Under the PFB gasification conditions studied at two scales of thermal input, the main bound nitrogen component produced was NH3, whereas HCN was formed to a minor extent of only a few percent of the fuel bound nitrogen content. HNCO, a component present under (C)FBC conditions, was never detected by means of even a high resolution FTIR spectrophotometer under the pressurised gasification test conditions studied. Conversion to NH3 and HCN was found to be comparable with other bottom-fed FB gasifiers, whereas comparatively low values were found for a top-fed pressurised FB. This difference is attributed to different pyrolysis conditions with faster heating rates and the presence of oxygen in bottom-fed reactors, leading to more bound nitrogen gas. In top-fed systems more tar and char bound nitrogen is expected. An increased Ca inventory in the gasifier by application of additive supply or from the fuel’s inorganic constituents tends to increase the NH3/HCN ratio significantly. The following recommendations can be formulated: •

•

Further study the background of the influence of the feeding location on the fuel bound nitrogen speciation behaviour. This could be accomplished by performing fast pyrolysis experiments in different oxidizing media (variable O2 and H2O contents) and also by studying the nitrogen partitioning behaviour for particles with different moisture content. The observation of a profound axial acetylene profile and the influence of steam addition could be studied into more detail in relation to the fate of tars and soot, as it is related to proper functioning of high temperature dry gas filtration (prevention of blockage by fine carbonaceous material).

4.6.2 Conclusions and recommendations for fuel characterisation Flash pyrolysis experiments with miscanthus were conducted using a heated grid reactor equipped with in-situ infrared absorption spectroscopy with a tuneable laser at TU Eindhoven. This research was focused on measuring the pyrolysis yield of CO, CO2 and NH3 at a heating rate of 280-320K/s and a final temperature of 1050-1400K. Qualitative trends of CO and CO2 formation were quite well predicted by the FG-DVC biomass pyrolysis model, developed by AFR (USA). However, the extrapolation that the model uses to predict the pyrolysis product yield at high heating rates, based on a kinetic input file determined by applying low heating rate TG-FTIR analysis experiments, was found to be precarious. This resulted in a reasonable quantitative yield prediction for CO and a firm under-prediction for CO2. The competition between the evolution of primary products like primary tar fragments and carboxylic acids on one side and light gases like CO, CO2 an H2O on the other side was believed to be the reason. This competition is expected to be a heating rate dependent phenomenon that makes yield predictions by means of extrapolation from low to high heating rate pyrolysis precarious. Apparently the primary pyrolysis products like tars and carboxylic acids, containing precursor groups for formation of CO and CO2, are quickly removed from the reaction zone and quenched immediately in the experiments carried out. Therefore, no time is available for further decomposition of the primary tars and carboxylic acids into CO and CO2, respectively, resulting in low yields. Carboxylic acids are oxygen rich acids and contain the precursor carboxyl group for CO2 formation. Primary tars contain ether links, precursors for CO formation during cracking at higher temperatures. According to this hypothesis the yields of primary tar and acetic acids must be significant. This is confirmed by the observations that at high heating rates the tar yield increases for biomass pyrolysis, as has been reported in several prominent scientific articles.

156

It is observed that due to the formation of transition metal carbonyls on the grid reactor wall about 20 mass% of the CO yield disappears within 2 seconds after being released. This phenomenon did not influence the yield results, since just enough time was available to measure the maximum CO yield before the CO disappeared. An attempt to detect NH3 in-situ during flash pyrolysis experiments in the grid reactor has failed. No absorption peak could be found for the NH3 reference gas in the reactor. The reason could be condensation together with water on relatively cold walls. Also, it can be the case that the frequency range of absorption was shifted, so that in fact another laser would be necessary as the applied laser can only be tuned in a very narrow range of frequencies. There were, however, no time and money resources left in the framework of this thesis to further study this item. The following recommendations can be formulated: •

For better predictive capabilities of the FG-DVC biomass model, a mechanism should be developed that accounts for the competition between primary pyrolysis products like primary tars and carboxylic acids on one side and light gases on the other side. Most likely the new version of the FG-DVC biomass model, to be released in the near future, will already include a mechanism for the competition between tars and gases.

•

More experimental research must be performed on pyrolysis of different biomass species at high heating rates for validation of the model, preferably starting at lower temperatures and at elevated pressures. This should include the whole range of major and minor product species and their evolution mechanisms. Combining FTIR analysis or multi-laser IR absorption spectrometry with a heated grid reactor and thermographic equipment suitable for high heating rates could provide the necessary experimental research equipment for this research.

•

The data analysis program should be optimised by numerically determining the area of an absorption peak, instead of using the central wavelength method. This will increase the accuracy, especially at elevated pressures, when pressure broadening affects the absorption peaks. A subroutine should be included that determines the reaction rate, numerically or by computation of the derivative of the curve fit. The last option is already included in the program but is currently not used due to convergence problems in the calculation procedure.

•

The thermocouple must be replaced by one with a junction diameter that is at least a factor 5 smaller than the current junction. Combined with a fast feedback control loop on the power supply and continuous registration of the thermocouple temperature this will significantly improve temperature measurement during grid reactor experiments.

•

The hypothesis that approximately 20% of the CO yield disappears due to formation of transition metal carbonyls at the reactor wall during flash pyrolysis experiments should be investigated. A coating on the stainless steel reactor shell inside the grid reactor can eliminate formation of such metal carbonyls.

•

The possibility to detect selected carboxylic acids formed during pyrolysis, by means of tuneable diode laser infrared absorption spectroscopy should be studied. Detection of large carboxylic acid concentrations released during flash pyrolysis experiments in the heated grid reactor used in this thesis could proof the hypothesis for the low CO2 yield.

•

The in-situ detection of NH3, HCN or HNCO in biomass pyrolysis experiments should be pursued so as to be able to measure the product end yields. Therefore, an investigation of the wavelengths of the absorption peaks to be used is needed. For this purpose, the minimum NH3 concentration in the reactor necessary to get absorption peaks large enough to distinguish from the signal-noise should be determined.

157

Chapter 5 Modelling bubbling fluidised bed gasification, focussed on nitrogen compounds 5.1 Modelling approach This chapter presents the pressurised fluidised bed gasification reactor model, which has been developed. The model describes the conversion of the fuel, taking into account the chemistry of formation and destruction of main gaseous components in the product gas as well as the formation and destruction of fuel bound nitrogen (FBN) species, which are precursors for NOx formation. The gasification reactor is hydrodynamically described as a series of ideal plug flow reactors, which is a relatively simple approach. Furthermore, a sensitivity study is presented to show the model’s response to the addition of various compounds, and changes in the temperature, pressure, fluidisation velocity and air stoichiometry. 5.2 Description of the model 5.2.1 Idealised reactor approach For the modelling of the gasifier using complex nitrogen chemistry kinetics, a series of idealised reactors can be used for a comparatively simple description of the hydrodynamic behaviour of the fluidised bed. A fluidised bed reactor can be considered as a chemical reactor in which continuously homogeneous gas-gas and heterogeneously catalysed gas-gas as well as gas-solid reactions are occurring. The solids are supposed to be well mixed (ideally stirred tank reactor) whereas the behaviour of the gas flow is intermediate between that in a continuously ideally stirred tank reactor and a plug flow reactor. In the continuously operated ideally mixed tank reactor, the composition of the reaction mixture is assumed to be uniform and equal to the composition at the outlet. For an ideal plug flow reactor there is no mixing in the axial direction of the conduit. In steady state, the conditions at any point in the reactor are independent of time, and, in particular, the linear velocity u of the reacting mixture is the same at every point in a cross section perpendicular to the direction of the flow.

Reactants

Figure 5.1a Plug flow reactor.

Products

Figure 5.1b Continuous well stirred tank reactor.

In order to investigate the fluidised bed reactor regime, it is possible to follow an approach presented in [Westerterp et al., 1984]. If the Péclet number for longitudinal dispersion for the gasifier is higher than 100, a plug flow regime can be assumed. The axial Péclet number, Peax, is a dimensionless number defined as: Pe ax =

u ⋅L D ax

(5.1)

where u is the average actual fluid velocity, L is the length of the reactor (bed height or freeboard length in this case) and Dax is the coefficient of longitudinal (axial) dispersion.

159

The Peax number can be considered as the ratio between the transport rate by convection and the transport rate by axial dispersion. Two extreme cases can be pointed out [Westerterp et al., 1984]: in case Peax is infinite the dispersion rate is negligible compared to the convection rate. This is plug flow; for the case that Peax approaches zero the convection rate is much slower than the dispersion rate so that the flow region is completely mixed. It was calculated that both for the Delft and Stuttgart gasifiers Peax is higher than 100, so that a plug flow regime for the reactor models can be assumed. This is in agreement with [Kunii & Levenspiel, 1990] who stated that the gas flow in both the bubble and emulsion phase can be considered as plug flow. This flow pattern is also assumed in the slugging bed model of [Hovmand & Davidson, 1968]. Figure 5.2 and 5.3 give a schematic of the structure of the model applied to the IVD and the TUD pressurised fluidised bed gasifiers, respectively. The schemes differ due to different reactor configurations. The TUD gasifier is equipped with probes in the freeboard with their own nitrogen purging, therefore the freeboard sections between the probes are modelled as separate sequential reactors. The IVD gasifier is operated in such a way that bed contents reach the maximum bed height, which is not the case for the TUD gasifier. That is why for the TUD gasifier the bed section is divided into two reactors and the IVD gasifier bed section not.

BIOMASS

PYROZONE

Modelled bed section Modelled freeboard section

RYIELD AIR PULSN2 STEAM

MIXER1 MIXER

BED

FB1 GASOUT

RPLUG

RPLUG

Figure 5.2: Model schematic of the IVD fluidised bed reactor as a series of ideal plug flow reactors

160

Modelled bed section

BIOMASS

Modelled freeboard section

PYROZONE

RYIELD STEAM

AIR

N2FEED PROBE1

MIXER1 MIXER

BED

BED2

FB1

RPLUG

RPLUG

RPLUG

PROBE3

PROBE2

MIX1

MIX2

MIXER

MIXER

FB2

FB3

RPLUG

RPLUG

CERFN2

PROBE4

MIX3

MIX4

MIX5

MIXER

MIXER

MIXER

GASOUT

Figure 5.3: Model schematic of the TUD fluidised bed reactor as a series of ideal plug flow reactors. A plug flow reactor represents an idealised reactor with the following properties: 1) steady flow; 2) no mixing in the axial direction, implying that molecular and/or turbulent mass diffusion is negligible in the flow direction; 3) uniform properties in the direction perpendicular to the flow, i.e., one-dimensional flow, meaning that at any cross-section, a single velocity, temperature, pressure, composition completely characterizes the flow; 4) ideal frictionless flow; an assumption allowing the use of the simple Euler equation to relate pressure and velocity. The ideal plug flow reactor modelling is easily derived from the schematic overview given in figure 5.4. The following conservation equations hold: Mass conservation:

d ( ρ v x A) =0 ⇔ dx

1 dρ ρdx

1 d vx

+

vx

dx

+

1 dA A dx

=0

(5.2)

X-momentum conservation:

d vx dP + ρ vx =0 dx dx

(5.3)

Energy conservation:

d ( h + v 2x / 2)

dx

=0

⇔

dh

dx The enthalpy function, h, can be expressed as:

h = h (T, Yi )

+ vx

d vx dx

=0

i = 1,2,...,N so exploiting the chain rule to relate dh/dx and dT/dx, yields:

(5.4)

(5.5)

161

dh dT =c + dx p dx

d Yi

N

∑h

i

(5.6)

dx

i =1

Species conservation: .

dYi ω i MWi =0 dx ρ vx

(5.7)

with: N

.

ω i = ∑ν ji qi

(5.8)

i=1

where N

N

j=1

j=1

∑ν 'ji X j ⇔ ∑ν ''ji X j and

ν ji

=

(ν ''ji

−ν

' ji

for i = 1,2,...L

(5.9)

)

(5.10)

and q =k i

N

ν 'ji

∏ ⎡X j ⎤⎦

f,i j=1 ⎣

− k

N

ν 'ji'

∏ ⎡ X j ⎤⎦

(5.11)

r,i j=1 ⎣

The ideal gas equation of state holds: ρRT

P=

(5.12)

MWmix

so that dMWmix 1 dP 1 dρ 1 dT 1 = + − P dx ρ dx T dx MWmix dx

(5.13)

The average molar mass of the mixture, MWmix, can be expressed as: MWmix

⎡N Y ⎤ = ⎢∑ i ⎥ ⎣⎢ i=1 MWi ⎦⎥

−1

(5.14)

and N 1 dY dMWmix 2 i = − MWmix ∑ dx i=1 MWi dx

(5.15)

The set of conservation equations can now be expressed in terms of dρ/dx, dT/dx and dYi/dx thereby reduced to three ODE’s by substitution: (1 − dρ = dx

v2 P (1 + x ) - ρv 2x cp T

. N d T v 2x d ρ v 2x 1 d A 1 = + ( )− ∑ h i ω i MWi d x ρ cp d x cp A d x v x ρ cp i =1

162

.

.

MWmix R 1 dA ρR )ρ2 v 2x ( )+ c p T) ∑ MWi ω i ( h i − c MWmix A dx v x c p MWmix i =1 MWi p N

(5.16)

(5.17)

.

ω i MWi = dx ρ vx

dYi

(5.18)

In equations (5.4) and (5.16) it is already assumed that the wall heat flux function is zero. Furthermore, the residence time, tr, is introduced as: dt r 1 (5.19) = dx v x Finally, in order to solve the set of ODE’s, intial conditions are defined for (5.15) to (5.18): T(0) = T0

ρ (0) = ρ 0

(5.20 a) (5.20 b)

Yi (0) = Yi0

for i = 1,2,...,N

t r ( 0) = 0

(5.20 c) (5.20 d)

Figure 5.4 The ideal plug flow reactor

For the calculation of the bed height of the fluidised bed reactor, an approach is taken based on Davidson&Harrison’s two phase theory as described by [Jiang, 1991]. In this approach a relation is given with the height under minimum fluidisation conditions, Hmf. H

H − H mf = ε b (Z) dZ

∫0

H mf =

m bed ρ s A(1 − ε mf )

(5.21)

(5.22)

163

ε

mf

=[

ε b (Z) =

1

]

14 φ

1 3

(5.23)

u b (Z)

(5.24)

u bA (Z)

u b (Z) = u 0 − u mf = u excess

(5.25)

u bA (Z) = u 0 − u mf + 0.711 g db

(5.26)

The minimum fluidisation velocity is calculated according fluidised beds, given by [Rowe, 1984]: ⎧ ⎪ ⎡ 0.0003112 Ga ε 3 η ⎪ mf u mf = 42.9 1 − ε mf ) ⎨ ⎢1 + ( ⎢ 2 ρgas d p (1 − ε ) ⎪⎢ ⎣ mf ⎩⎪

to an empirical relation for pressurised 1 ⎫ ⎤ 2 ⎪ ⎪ ⎥ − 1⎬ ⎥ ⎪ ⎦⎥ ⎭⎪

(5.27)

with Ga being the Galilei number, which is defined as: Ga =

ρgas ( ρ s − ρ gas ) gd p

3

(5.28)

η2

The average particle diameter of the bed material is taken to be the Sauter mean diameter: 1 dp = ` N⎛ Y ⎞ ∑⎜ ⎟ i=1 ⎜ d p ⎟ ⎝ ⎠i with Yi: mass fraction of solid bed material. H − H mf =

H

∫u 0

u excess excess

+ 0.711 g db ( Z )

Z*

= u excess { ∫ 0

(5.29)

dZ

1 (u excess + 0.711 g db ( Z ))

dZ +

H

∫

Z* (u excess

1 + 0.711 g db ( Z * ))

dZ}

(5.30)

with

(

db (Z) = 0.54 u 0.4 Z − 4 A0 excess

)

0.8 −0.2

g

(5.31)

and

with

and so that

164

⎛ A ⎞ 0 ⎟ Z* = 3.5D ⎜ 1 − T⎜ A ⎟ ⎝ ⎠ 2 A = 1 π DT 2 2 A 0 = N orifice 1 π Dorifice 2

(5.32)

(5.33) (5.34)

H = H mf

⎡ ⎤ ⎢ Z* ⎥ * 1 (H − Z ) ⎥ (5.35) + u excess ⎢ ∫ dZ + ⎢0 ⎧ 0.8 * ⎫⎥ 0.8 0.4 ⎨u excess + 0.711 g d b (Z ) ⎬ ⎥ ⎢ (u excess + 0.711 0.54 g u excess {Z −4 A0 } ⎩ ⎭⎦ ⎣

The first term between brackets is numerically integrated using a Romberg scheme with Richardson extrapolation using a very small step size of h/210. In our approach to model the chemistry of a fluidised bed gasification reactor including nitrogen species, the gasifier is considered to consist of two main sections: bed and freeboard. The fluidised bed gasification process modelling is based on several assumptions, which are discussed below. General assumptions The reactor shows steady state operation; no transient effects are taken into account. The gas phase is described by the ideal gas law. This assumption is very reasonable, as the pressures and temperatures prevailing in the reactor are not excessively high. Using simple mixing rules for critical temperature and pressure (Tc, Pc) values of 178.6 – 284.6 K and 50.7 – 87.1 bar are calculated. The relative temperature and pressure for produced gas downstream of the filter unit for the experimental set points applied are then in the range of 3.13 – 5.31 for Tr and 0.042 – 0.099 for Pr. The compressibility factor, z, applied to indicate deviation from ideality (see e.g. [Smith et al., 2001]), is 1 for the indicated Tr and Pr range within 2% error. The gasifier shows isothermal behaviour and intraparticle temperature gradients are negligible. This assumption is confirmed by [Srinivas & Amundson,1980] and [Bliek et al., 1986] who estimated temperature difference between the particle’s surface and centre to be small using the relationship of [Prater,1958]. [Weimer & Clough, 1980] concluded that internal fuel particle temperature differences are negligible when the particle is smaller than 7 mm. As the temperature is considered to be constant, no enthalpy balance has to be solved, but this balance is checked and the higher heating value (HHV) of the fuel is varied to close it. A check is made whether the obtained HHV value is realistic. Ash conversion is not considered, as only the conversion of the organic part of the fuel is modelled. Assumptions for the bed zone of the gasifier Instantaneous particle drying & devolatilisation in the initial bed entrance zone, with uniform distribution of volatiles over this reactor part. Immediate mixing of air, steam and N2 in the initial bed entrance zone. The bed behaves as an ideal plug flow reactor with respect to the gas phase. Ideal mixing of solid char and bed material. Char is considered to consist only of pure C. For biomass it is assumed that char conversion by oxidation and reduction reactions can be modelled by the pore tree model (equation (5.45)), see [Simons & Finnson, 1979]. An overview of various pore models is presented by [Weeda, 1995] and [Weeda et al., 1993]. The fit parameters used in this model can be determined a priori. Also, this model shows a maximum in the reaction rate as a function of the conversion, attributed to coalescence of growing pores [Janse et al., 1998]. For brown coal the volume model was reported to be well applicable over the whole range of fuel conversion and is applied here for that fuel [Hamel, 2001]. Particles fed are assumed to be spherical and uniform with respect to their diameter. Abrasion (attrition), agglomeration, fragmentation and entrainment of solid particles is neglected as it is considered not be important for the gas phase emission predictions.

165

Assumptions for the bed zone of the gasifier The freeboard is characterised by plug flow for the gas phase, heterogeneous reactions in this section are neglected, due to the very low solids hold up as compared to the bed section. Abrasion (attrition), agglomeration, fragmentation and entrainment of solid particles is neglected in this section as well, as it is considered not be important for the gas phase emission predictions. With respect to the instantaneous flash pyrolysis yields assumed in the initial bed zone, two approaches can be followed. Yields can be obtained directly from (fluidised bed) flash pyrolysis experiments, or they can be calculated from characterization experiments carried out at slower heating rates, using a model for which the kinetic parameters are determined by using these characterization techniques. For the first approach, a complete data set of released gases, char and tar is necessary for the temperature and pressure conditions that are relevant for the gasification conditions to be simulated. Also, the flash pyrolysis yield data need to be consistent with respect to closed mass- and element balances. This type of data often lacks completeness, although some can be used. In our view the data from fluidised bed pyrolysis experiments by [Van den Aarsen, 1985] can be used, although they still are incomplete. The experiments described by this author have been determined by fluidised bed operation at temperatures between 700 and 900 °C, under atmospheric conditions. The second approach uses the FG-DVC model to determine instantaneous flash pyrolysis yields, based on the fuel characterization data, which are described in chapter 4. A heterogeneous char oxidation gas-solid reaction, is considered to be limited by external mass transfer limitation of O2. This assumption is checked for all simulations. In summary, all the heterogeneous reactions considered are: C + O2 → CO2 C + H2O ' CO + H2 C + CO2 ' 2 CO C + 2 H2 ' CH4

(R5.1) (R5.2) (R5.3) (R5.4)

In this work, CO2 is considered to be the only product of the heterogeneous combustion reaction (R5.1). This assumption is often taken for particles of a size larger than 1 mm, see e.g. [Chakrabourty & Howard, 1981]. For the intrinsic kinetics of solid char gasification with CO2 and H2O, a rate expression published by [Van den Aarsen, 1985] for beech wood was used for the crushed wood pellet gasification experiments. For (R5.1) the following relations represent the char conversion rate to CO2, assuming fast surface oxidation reaction (external diffusion limitation): 6 (5.36) rc,R1 = k diff CO dp 2 with k diff = and

Sh DO

2

dp

Sh = 2 εg,mf + 0.69 Re0.5 Sc0.33

(5.37) (5.38)

a relation proposed by [Chakrabourty & Howard, 1981], which is also applied in the IEA-CFBC model [Hannes, 1996] The rate of diffusion to the char surface (depending on the carbon conversion, X) was compared to the rate of oxidation for biomass, as presented by [Janse et al., 1998]: ⎛ -125.103 ⎞ 0.53 rc,R1 =5.3 . 105exp ⎜ ⎟ .P (1−X )0.49 ⎜ RT ⎟ O2 ⎝ ⎠

166

(5.39)

The char oxidation rates with the assumption of oxygen diffusion limitation, calculated according to (5.36), were much lower than those calculated for reaction limitation by equation (5.39) for the temperature and oxygen levels relevant in the biomass based simulations. For brown coal as fuel the char oxidation rate is expressed as: ⎡ 6 ⎢ 1 1 + rc,R5.1 = ⎢ dp ⎢ ⎡ E ⎤ k diff ⎢ k 0,R5.1T exp ⎢- RT ⎥ ⎣ ⎦ ⎣

−1

⎤ ⎥ ⎤ ⎥ ⎡O ⎥ ⎣ 2,∞ ⎦ ⎥ ⎦

(5.40)

with k0,R5.1=10.4 m/(s.K) and E/R = 10300 K ([Hobbs et al., 1992] for SUBC coal, high volatile lignite type of coal). The kinetics of the heterogeneous gasification reactions (R5.2) and (R5.3) for biomass, beech wood, are given by [v.d. Aarsen, 1985]: ⎡ - Ea,R5.2 ⎤ -0.17 ⎥ c0.83 rc,R5.2 =14.4 S exp ⎢ . (1000 ) εc H O ⎢⎣ R.T ⎥⎦ 2 ⎡- E ⎤ -0.17 rc,R5.3 =7.2 S exp ⎢ a,R5.3 ⎥ c0.83 1000 ) εc ( CO 2 ⎢⎣ R.T ⎥⎦

(5.41) (5.42)

with Ea,R2 = Ea,R3 = 166156 J/mol. In the expressions (5.41) and (5.42) for biomass, the specific particle surface area (S), which depends on the carbon conversion (X), is calculated using the pore tree model [Simons & Finnson, 1979]: S(X) S0

⎪⎧ ⎡ X ⎤ ⎪⎫ = (1-X) ⎨ ⎢ ⎥ + (1-X) ⎬ ⎢ ⎥ ⎩⎪ ⎣ ε 0 ⎦ ⎭⎪

1

2

(5.43)

Values for S0 and εo are taken from [Van den Aarsen, 1985] for beech wood: S0 = 105.106 m2/m3 and ε0 = 0.69. [Zygourakis, 1988] found for lignite that fuel particle porosity and surface area are only slightly enlarged when the heating rate was increased from 0.1 to 1000 K/s. Similarly, [McDonald et al., 1992] concluded that the surface area is not so much affected by the heating rate or pyrolysis (end) temperature. For Rhenish brown coal, [Hamel, 2001] determined the following kinetics for gasification with CO2 (R5.3), for 1073 < T < 1223 K and 8 < [CO2] < 40 vol.%: ⎡ - Ea,R5.3 ⎤ ρ 0.45 1-X 0.57 p,0 ε ⎥ PCO ( ) MC c RT ⎦⎥ 2 ⎣⎢

rc,R5.3 =33.22 exp ⎢

(5.44)

with Ea,R3=131 kJ/mol and P in [Pa]. For the reaction with H2O a factor of 2 was applied to the preexponential coefficient. The reaction rate expression for the hydro gasification reaction R5.4 was derived from [Biba et al., 1978] and applied for both biomass and brown coal. This reaction proceeds much slower when compared to the other gasification reactions, R5.2 and R5.3. ⎡ -E ⎤ (5.45) rc,R5.4 =2.7.107 S exp ⎢ a,R5.4 ⎥ cH εc ⎢⎣ RT ⎥⎦ 2 with Ea,R4 = 230274 J/mol.

167

HCN is assumed to be converted by hydrolysis with water to NH3 and CO. This reaction is considered to be catalysed on a char surface and to be fast with respect to external mass transfer of HCN. See also [Shimizu et al., 1993]. HCN + H2O → NH3 + CO

(R5.5)

Based on this assumption, the following rate expression can be given for reaction (R5.5): 6 rc,R5.5 = k diff CHCN dp

(5.46)

with kdiff defined in an analogue way to (5.37). Tar conversion is modelled as a single decomposition step [Van den Aarsen, 1985]: CxHyOz → z CO + ¼ y CH4 + (x - z - ¼ y) C

(R5.6)

with the accompanying kinetic rate expression: ⎡ -Ea,R5.6 ⎤ ⎥ cTar ⎢⎣ RT ⎥⎦

rc,R5.6 = 3.7.107 exp ⎢

(5.47)

with Ea,R5 = 118900 J/mol. For the gas phase a homogeneous reaction kinetics scheme published by [Coda Zabetta et al., 2000 a and b] was applied, consisting of 353 reactions between 58 radical and molecular species. For simplification, detailed reactions of higher hydrocarbons (C4+) are not considered in this model. The rate coefficients, k, have been derived from the rate constants as by: ⎡ -E a ⎤ ⎥ ⎣ RT ⎦

k = k 0 T b exp ⎢

(5.48)

The pre-exponential factor values are usually expressed in units mole-cm-s-K, and the activation energy, Ea, in cal/mol. For ordinary homogeneous and third body enhanced reactions the following reaction rate expression holds (with the second factor being 1 for ordinary reactions): ri

⎛

J

⎜ ⎝

j=1

= ⎜ k i,f ∏ [C j ]

ν' ji

J

-

k i,r

[C j ] ∏ j=1

ν'' ji

⎞ ⎛ ⎟×⎜ ⎟ ⎜ ⎠ ⎝

J

⎞

α ji ×[C j ] ⎟ ∑ ⎟ j=1

(5.49)

⎠

The rate constant for backward reactions, ki,r, is calculated from the rate data given for the forward reaction and the equilibrium constant, according to: Ki =

k i,f

(5.50)

k i,r

K i,p = exp (

-∆G i RT

(5.51)

)

with Ki =

K i,p ∑ v"-∑ v' j

RT ⎛ ⎞j ⎜ ⎟ 5 ⎝ 1.013×10 Pa ⎠

(5.52)

For pressure dependent reactions the Arrhenius parameters are required for both the high- (subscript ∞) and low-pressure limiting case (subscript 0). In this work the Lindemann approach is followed to describe the intermediate pressure range to yield a pressure dependent rate constant k = k(P) using these two limiting cases. For the low-pressure limit the rate constant is:

168

-E b k 0 = k 00T 0 exp ( a0 ) RT and for the high pressure limit: -E 0 b∞ k∞ = k∞ T exp ( a∞ ) RT At any pressure within the pressure limits k becomes: p k = k∞ ( r ) 1 + pr with k [C] pr = ( 0 ) k∞ in which [C] is the total concentration of the third body compounds (enhanced or normal).

(5.53)

(5.54)

(5.55)

(5.56)

Table A3.1 in appendix 3 gives an overview of the reaction rate data used for the homogeneous gas phase reactions. The values are given by [Coda Zabetta et al., 2000 a and b] and the numbers before each reaction correspond to the reaction number in this reference. Thermodynamic data for the selected species involved in this research were selected from the Sandia Thermodynamic Database [Kee et al., 1990]. Modelling of radical quenching by formation of stable molecules on solid surfaces, as for gaseous combustion in fluidised beds is recently reported by [Loeffler&Hofbauer, 2002], has not been taken into account. The mass, energy and species balances in the plug flow reactor model, set up using an ASPEN PLUS® flowsheeting package from ASPEN Technology Inc. with user-defined subroutines for the reaction kinetics, lead to a system of ordinary differential equations. This system is numerically solved using the GEARS algorithm. The bed section of the gasifier is integrated to a length, which is calculated based on equation (5.35). 5.3

Simulation results and discussion

The main aim of this work is to study the evolution of the NOx-precursors in a biomass-fed pressurised fluidised bed gasifier by investigating how a variation in chosen process parameters might affect their fate within the gasifier reactor. By using the model implemented in ASPEN PLUS software, a number of simulations have been carried out. Relevant parameters have been varied and additional streams have been added to the original model layouts. The output results have been exported into a spreadsheet program, which is used for a parametric analysis. In the first part of this paragraph, the parameters that have been varied are described and their variation range, while in the second part the species evolution results are discussed and analysed. The model developed for the IVD gasifier, shown in figure 5.2, has been used in this part of the work. Pyrolysis input data for fluidised bed beech wood pyrolysis, atmospheric with temperatures between 715 and 915 °C by [Van den Aarsen, 1985], with initial 100% fuel_N conversion into NH3, are applied. Table 5.1 shows the yield and composition correlations. The input data (except the single parameter which is varied each simulation) were kept constant in this simulation campaign and a parametric analysis was performed.

169

Table 5.1 Correlations for product yields fluidised bed pyrolysis of beech wood derived from the work of [Van den Aarsen, 1985] as model input. Component

Mass yield (-)*

X**

Y**

Z**

Char Tar Water Dry Gas Gas species

1.187-2.569.10-3T+1.45.10-6T2 0.06 2.37-5.2.10-3T+3.0.10-6T2 -2.683+7.932.10-3T-4.55.10-6T2 Volume yield (%)***

1 1 -

10.027-2.432.10-2T+1.55.10-5T2 7.237-1.8.10-2T+1.15.10-5T2 -

3.906-9.115.10-3T+5.5.10-6T2 1.448-3.26.10-3T+2.0.10-6T2 -

CO H2 CH4 CO2 C2H4

-195.985+5.908.10-1T-3.6.10-4T2 151.132-2.914.10-1T+1.8.10-4T2 -40.738+1.304.10-1T-8.0.10-5T2 91.768-1.996.10-1T+1.2.10-4T2 6.06-4.0.10-3T+2.572.10-11T2

* ** ***

Yields are normalised to 1; T in (°C) In overall molecular formula CxHyOz; T in (°C) Yields are normalised when mass yields are calculated, together with fuel_N species yield from assumed distribution between NH3, HCN or N2; T in (°C)

A first set of simulations has been carried out for a fixed model layout and varying the main operating conditions: pressure, temperature, fluidisation velocity and the air to fuel ratio λ (defined by equation 4.2) were changed over an extensive range as shown in table 5.2. A second group of simulations has been performed to investigate the effect on the fuel gas composition (especially on the N-species) when mixing particular compounds to the bed and/or to the freeboard, thus, changing slightly the model layout by adding a new input stream. The reagent addition in the gas will increase the concentration of the H, OH, and O radicals needed for H, + OH, + O decomposition of NH3 to amino species according to NH 3 ⎯+⎯ ⎯⎯ ⎯→ NH i . In these simulations the input data and operating conditions correspond to the IVD gasification experiment 991202, the so-called base case. In table 5.3 the species added and their amount are summarised. Table 5.2 Overview of the simulations: different process conditions.

Base case Pressure variation

170

Pressure [bar] 5.1 1 2.6 7.6 10.2

Temperature [K] 1055

Fluidisation velocity [m/s] 0.36

Air-fuel ratio λ [-] 0.30

1055

0.36

0.30

0.36

0.30

1030 1073 1098 1123

Temperature variation

5.1

Fluid. velocity variation

5.1

1055

Air-fuel ratio variation

5.1

1055

0.27 0.45 0.54 0.72 0.36

0.30 0.25 0.42 0.63 0.97

Table 5.3 Overview of the simulations campaign: addition of reactive components. Added species Secondary Air NO NO and O2 H2O (steam) NO2 H2O2 CH4

Added amount 20-30% of primary air (mass basis) 1:1 NH3 from pyrolysis yield (molar basis) 1:1:1 NH3 from pyrolysis yield (molar basis) 10% of primary air (mass basis) 1:1 NH3 from pyrolysis yield (molar basis) 1:1 NH3 from pyrolysis yield (molar basis) Up to 1:10 CH4 from pyrolysis (mass basis)

Stream added to Freeboard Bed Bed Bed Bed Bed Bed/Freeboard

All gas species compositions (main and minor compounds) are given as a function of the residence time of the gas in the reactor for each single simulation. In figure 5.4 the concentrations of a selected number of main species are shown as a function of the residence time for the base case. These species are the most interesting main gas compounds. It can be seen that the O2 is consumed within approximately 0.5 s, which corresponds to approximately 10 cm. This is confirmed by e.g. [Kilpinen et al., 2002]. [Yoon et al., 1978] have shown that for a fixed bed coal gasifier the combustion and devolatilisation zones are also physically small compared to gasification/reduction zone. 0.1

Freeboard

Bed

CO CO2 H2O

Molar fraction (-)

0.08

0.06

H2

0.04

CH4

0.02

H2 O2 H2O CO CH4 CO2

O2

0 0

5

10

15

20

25

30

Residence time (s)

Figure 5.4 Example of main species behaviour within the gasifier predicted by the model for the base case: 991202. 0.01

Bed

0.001

Freeboard NH3

Molar fraction

0.0001 1E-05

NO HCN HNCO

1E-06 1E-07

N2O

1E-08

NO NO2 N2O H3N CHN HNO HNCO HCNO

HNCO

1E-09

NO2

1E-10 0

5

HCNO 10

15

20

25

30

Residence time (s)

Figure 5.5 Example of nitrogen species behaviour within the gasifier predicted by the model for base case: 991202.

171

In order to study how the N-species evolve when one of the operating conditions or an input stream is varied, fuel nitrogen mass balances over the bed and the freeboard have been calculated from the exported ASPEN simulation output data in a spreadsheet program. For this purpose, the partitioning of the fuel-N over all the N-components which are considered has been depicted on a percentage scale, where 100% of input fuel-N is released in the form of NH3 after flash pyrolysis. Other species are assumed to release according to the pyrolysis data as determined from the yield relations given in table 5.1, with bed temperature as measured being the input. In some simulation cases the N mass balance reported a 3%max error due to the limited number of digital outputs for the reported N2 mass flow from ASPEN (N2 flow was about 8 orders of magnitude bigger than the other minor N species). The error, computed by difference between the total N-output and the total N-input, has been subsequently charged on N2, thus closing the balance correctly. In figure 5.6 the N mass balance is shown for the base case.

% converted from N-fuel

4

bed freeboard

3

2

1

0 N2

NO

NO2

N2O

HCN

HNO

HNCO HOCN HCNO NO3

N-species

Figure 5.6 Fuel-N partitioning model prediction for the base case among main N-species within the gasifier (Fuel-N input only NH3); outlet of bed and freeboard are considered and remaining is NH3..

The data in the abovementioned diagrams refer to the outlet conditions of the bed and freeboard showing the partitioning of nitrogen into N-species in the bed and freeboard blocks respectively. In figure 5.6 it is shown how NH3 is a very stable compound, since only ca. 4% of the nitrogen present in the input fuel-N (for these simulations considered to consist only of NH3) is converted into other Nfixed species at the freeboard outlet (1% to N2, about 2% to NO and 0.5% to HCN and HNCO). By coupling the data for each simulation group, a parametric description of the behaviour of main nitrogen species into the gasifier has been performed. In the following subsections each simulation group is discussed.

Pressure simulations In the simulations with pressure as variable to be studied, the aim is to keep the air to fuel ratio (alternatively stated, λ), residence time and bed height constant, so that the total input mass flows of biomass and air are adjusted. Pressure has been varied from 1 to 10.2 bar corresponding to the following new mass flows, which have been calculated following the ideal gas law. With the new mass flows the residence time in the gasifier did not change more than approximately 6%.

172

Table 5.4 Mass flows into the bed (in bold: base case). Pressure (bar) 1 2.6 5.1 7.6 10.2

Total mass flow (kg/hr) 3.2 8.3 16.3 24.3 32.6

Air mass fow (kg/hr) 1.3 3.0 5.8 8.6 11.6

Biomass mass flow (kg/hr) 0.7 1.7 3.3 5.0 6.7

Nitrogen mass flow (kg/hr) 1.4 3.7 7.2 10.7 14.3

Fuel-N input (kgNH3/s)·107 0.9 2.4 4.7 7.1 9.5

An increase in pressure has resulted in our model into faster kinetics. At higher pressures species evolution becomes quicker even though the final molar concentration does not differ significantly from lower pressure cases. The main reason for increased kinetic rates with higher pressure can be explained by the pressure-dependent reactions, in particular the recombination fall-off reaction type, of which the rate increases with increasing pressure. In this type of reaction e.g. two radicals combine to one molecule; in the low-pressure limit, a third-body collision is required for the reaction to proceed The main N-species concentrations have been tabulated in table 5.5. Table 5.5 Main N-species concentrations at the gasifier outlet at different pressures (in bold: base case). Pressure (bar) 1 2.6 5.1 7.6 10.2

[NH3] (ppmv) 152 161 161 157 155

[HCN ] (ppmv) 1 0 0 1 1

[NO] (ppmv) 4 3 3 3 3

[HNCO] (ppmv) 0 0 0 0 0

4

98

3

96

2

94

1

92

0

90 1

2.6

5.1

7.6

Unconverted NH3 (%)

Fuel-N conversion (%)

From table 5.5, it is clear that NH3 conversion is very slightly favoured at low pressure (1 bar) or rather when increasing the reactor pressure (above 8 bar). In particular, N is converted from NH3 to N2 as shown in figure 5.7, which shows how the model predicts the fuel-N partitioning into gas N-species on a mass basis (fuel-N input is assumed 100% in the form of NH3 right after flash pyrolysis).

NO HCN N2 NH3

10.2

Pressure (bar)

Figure 5.7 Fuel-N conversion pressure dependence over the entire reactor.

173

Temperature simulations The temperature has been varied from 1030 to 1123 K in steps of 25 K. In these simulations the mass flow has not been changed. The overall residence time changed less than 2.5 s, when varying the temperature, which is comparable to that of the pressure simulations. An increasing temperature resulted in enhanced reaction rates both in the bed and in the freeboard section. The NH3 conversion was slightly higher at higher temperature. The overall ammonia decomposition reaction is given by: 2NH3 ↔ N2 + 3H2 (R5.7) and is favoured towards the products at higher temperature. However, within the interval of about 100 K the NH3 concentration was only reduced with a few ppm (6 ppmv) confirming the relatively high stability of the NH3 compound in this temperature range, which is typical for fluidised bed gasification of biomass. The main N compound yields are presented in table 5.6. Table 5.6 Main N-species concentrations at the gasifier outlet at different temperatures (in bold: base case). Temperature (K) 1030 1055 1073 1098 1123

[NH3] (ppmv) 161 161 158 156 155

[HCN] (ppmv) 0 0 1 3 3

[NO] (ppmv) 3 3 3 4 3

[HNCO] (ppmv) 0 0 0 0 0

3

98

2

96

1

94

0

92 1030

1055

1073

1098

Unconverted NH3 (%)

Fuel-N conversion (%)

At higher temperatures ammonia conversion is increasing, being converted into N2 and HCN, as depicted in figure 5.8.

1123

Temperature (K)

Figure 5.8 Simulation results for fuel-N conversion: temperature dependence.

174

NO HCN N2 NH3

Fluidisation velocity simulations

Fluidisation velocity simulations have been carried out to investigate if residence time does affect species evolution. The fluidisation velocity has been varied by decreasing the total mass flow into the gasifier and keeping the mass flow ratio of input streams unchanged. Table 5.7 Fluidisation velocity and related residence time (in bold: base case). Fluidisation velocity (m/s) Bed residence time (s) Freeboard residence time (s)

0.27 3.7 34.1

0.36 3.0 25.6

0.45 2.6 20.5

0.54 2.3 17.1

0.72 1.9 12.8

Residence time only affects to a little extent the outlet fuel gas composition. The main oxidation reactions in the bed, such as H2 oxidation, are so fast that they are not depending on the residence time. The reaction rates in the freeboard are instead rather slow and not significantly affecting the outlet gas fractions. NH3 conversion occurs almost only in the bed section and independent from residence time and fluidisation velocity.

Air to fuel ratio simulations The air to fuel ratio λ has been varied in order to investigate the predictions of the model when approaching combustion conditions. In this work the definition used for λ is given by equation 4.2. Table 5.8 summarises the different λ values used in the simulations and the corresponding mass flows. Table 5.8 Mass flows into the bed (in bold: base case) Air stoichiometry λ [-] 0.25 0.31 0.42 0.63 0.97

Total mass flow (kg/h) 16.29 16.29 16.29 16.29 16.29

Air mass flow (kg/h) 5.79 5.79 5.79 5.79 5.79

Biomass mass flow (kg/h) 4.17 3.34 2.50 1.67 1.10

N2 mass flow (kg/h) 6.32 7.16 7.99 8.83 9.39

The λ value has been changed by varying the biomass mass flow while the air mass flow has been kept constant. The N2 mass flow has been subsequently varied to maintain the same fluidisation velocity in the gasifier, making the results of this simulation group comparable with those of other different simulations. When λ = 0.97, which is almost the stoichiometric amount of oxygen, the main species are converted mainly into CO2 and H2O and no CH4 or H2 are found in the gasifier outlet. The fuel-N partitioning into the main N-species is presented in figure 5.9.

175

100

40

75

30

NO HCN N2 NH3

20

50

25

10 0

Unconverted NH3 (%)

Fuel-N conversion (%)

50

0

0.25

0.31

0.42

0.63

0.97

λ (-)

Figure 5.9 Model predicted fuel-N conversion versus air stoichiometry, λ..

It can be observed from figure 5.9 that the fuel-N (100% NH3) is converted into NO and N2 when approaching combustion conditions following the partial and complete ammonia combustion reaction paths as sketched in figure 5.10. It is worthwhile to remark that no N2O was predicted by the model at any λ value. Gasification

Combustion

NH3 HCN

N-fuel (NH3)

NO

NO

N2O N2

N2

Figure 5.10 Simplified paths for NH3 conversion in gasification and combustion.

From figure 5.9, it is also evident that between λ = 0.63 and λ = 0.97, thus, when oxygen approaches the stoichiometric amount, also HCN is combusted, and hence converted primarily into NO and N2. Table 5.9 Main N-species concentrations at the gasifier outlet at different λ (in bold: base case). Air stoichiometry λ (-) 0.25 0.31 0.42 0.63 0.97

[NH3] (ppmv) 199 161 114 53 0

[HCN] (ppmv) 0 0 1 3 0

[NO] (ppmv) 2 3 8 18 28

[N2O] (ppmv) 0 0 0 0 0

Simulations with addition of reactive components The model has been used to simulate the behaviour of the N species when specific compounds are added to the bed and/or to the freeboard. The objective of this part is to investigate whether a conversion of NH3 into N2 would be possible by adding small amounts of oxygen and/or nitric oxide or other compounds to the gasification gases.

176

The basic idea is that the added component in the gas will increase the concentration of the H, OH, and O radicals needed for decomposition of NH3 to amino species according to: H, + OH, + O NH 3 ⎯+⎯ ⎯⎯ ⎯→ NH i .

New streams have been added to the IVD model layout. In the following subsections the conversion of the main components and the fuel-N are presented and discussed. Secondary air

Secondary air in the amount of 20 and 30 mass% of the primary air has been added to the freeboard. Most of the added O2 reacts relatively fast with the H2, and slower with CO and CH4. The NH3 conversion is also affected by the presence of O2, mainly according to the overall ammonia combustion reaction NH3 + 2O2 → NO + 3/2 H2O. Very little ammonia is converted into molecular nitrogen (less than 3%). Table 5.10 Main N-species concentrations at the gasifier outlet with secondary air. Secondary air % of prim air 0 20 30

[NH3] (ppmv) 154 140 133

[HCN] (ppmv) 2 2 2

[NO] (ppmv) 6 9 11

In figure 5.11, NH3 conversion has been quantified as in the previous simulations. It is clear that more NH3 is converted when the amount of air (i.e. of O2) is increased.

94

NH3 unconverted (%)

92 Bed Freeboard

90

88

86

84 0%

20%

30%

% sec. air (% of prim. air, mass basis)

Figure 5.11 Model predicted NH3 conversion with and without secondary air.

NO addition

The effect of NO addition on the reduction of NH3 has been studied. NO has been added to the bed in a molar ratio to the NH3 of 1:1.

177

NH3 unconverted (%)

100 90 NH3 conversion with NO addition NH3 conversion without NO addition

80 70 60 1

2

3

1 = input fuel-N; 2 = bed outlet; = freeboard outlet

3

Figure 5.12 Simulation result for NH3 conversion with and without NO addition into the bed (1:1 on a molar basis)

NO addition enhances the NH3 conversion into N2 according to: H, + OH, + O NO NH 3 ⎯+⎯ ⎯⎯ ⎯→ NH i ⎯+⎯ ⎯ → N2

The same reactions describe the reduction of NO to N2 by addition of NH3 in combustion gases [Kilpinen et al., 1999]. In the following figure 5.13, input of nitrogen species has been supposed to be partitioned into 50% NH3 and 50% NO on a mass basis. NO is broken down to 34% and NH3 to 36%. N is converted mostly in N2 and to a less but significant extent into HCN.

N-species input and conversion (%)

60

N input N conversion outlet bed N conversion outlet freeboard

50 40 30 20 10 0 N2

NO

NO2

N2O

NH3

HCN

HNO

HNCO

HCNO

N-species

Figure 5.13 Simulation result for nitrogen partitioning among main N-species within the gasifiers (nitrogen species input = 50% NH3 and 50% NO)

178

The concentrations of these components are presented tabled in table 5.11. Table 5.11 Main N-species at the gasifier outlet with and without NO addition (*) NO/NH3 ratio [NH3] [HCN] [NO] [N2O] molar basis (ppmv) (ppmv) (ppmv) (ppmv) 0 154 2 6 0 1 121 31 114 1 (*) The reaction mechanism used did not involve HCN hydrolysis nor the complete char combustion reaction

NO and O2 addition

In the previous subsections, it has been shown that the addition of O2 and NO when added separately gave rise to a significant conversion of NH3 into N2 and NO. The objective of this section is to see whether simultaneous additions of O2 and NO into the bed will result in synergistic effects. The study has been made for a constant molar ratio of additive to incoming NH3, thus 1:1:1 to NH3 from pyrolysis yield (molar basis). O2 addition has not shown to be very effective when added together with NO, see table 5.12. In fact, NH3 is converted up to 26% (with only NO the NH3 conversion was 25%). Also from this table it can be seen that the nitrogen fixed species concentrations do not differ significantly from the NO addition case. [Coda Zabetta et al., 2001] pointed out that no significant synergy can be expected of the simultaneous addition of both oxidisers. When the time which is available for oxidation exceeds approximately 1s, NO becomes more effective as NH3 destruction oxidiser than O2. Table 5.12 Main N-species concentrations at the gasifier outlet with and without NO + O2 addition (*) NO/NH3/O2 molar NH3 ppmv HCN ppmv NO ppmv N2O ppmv basis 0/1/0 154 2 6 1 1/1/1 118 29 114 1 (*) The reaction mechanism used did not involve HCN hydrolysis nor complete char combustion reaction

H2O (steam) addition

Steam has been added to the bed in the amount of 10% (mass basis) of primary air. No significant increase in the concentration of the H, OH, and O radicals needed for the decomposition of NH3 into fixed-N species has been found, thus no increase in fuel-N conversion has been observed. The results of this simulation in terms of nitrogen partitioning do not differ from the base case. NO2 addition

The effects of NO2 on fuel-N conversion has been investigated by adding NO2 to the bed in the amount of 1:1 on molar basis to NH3 from pyrolysis yield.

179

NH3 unconverted (%)

100

90 NH3 conversion with NO2 addition NH3 conversion without NO2 addition

80

70

60 1

2

3

1 = input fuel-N; 2 = bed outlet; = freeboard outlet

3

Figure 5.14 Model predicted NH3 conversion with and without NO2 addition into the bed

It turned out that NO2 (as NO in the previous subsection) also favours NH3 conversion up to 26%, but on the other hand, when supposing the input nitrogen species to consist of 50% NO2 and 50% NH3 some HCN (10%) and NO (34%) are formed (figure 5.15).

N-species input and conversion (%)

60

Fuel-N input N conversion outlet bed N conversion outlet freeboard

50 40 30 20 10 0 N2

NO

NO2

N2O

NH3

HCN

HNO

HNCO

HCNO

N-species

Figure 5.15 Simulation result for fuel-N partitioning among main N-species within the gasifier (nitrogen species input = 50% NH3 and 50% NO2)

The concentrations of the most relevant N-components are shown in table 5.13. Table 5.13 Main N-species concentrations at the gasifier outlet with and without NO2 addition (*) [NH3] [HCN] [NO] [N2O] NO2/NH3 molar basis (ppmv) (ppmv) (ppmv) (ppmv) 0 154 2 6 0 1 121 31 114 1 (*) The reaction mechanism used did not involve HCN hydrolysis nor the complete char combustion reaction

180

The basic reactions of NH3 conversion into N2 are the same as presented in the NO addition simulation case. In fact, NO2 is readily converted into NO (see figure 5.15) which further converts NH3 into N2 according to H, + OH, + O NO NH 3 ⎯+⎯ ⎯⎯ ⎯→ NH i ⎯+⎯ ⎯ → N2 .

H2O2 addition

H2O2 has been found to be an ammonia decomposition agent [Azuhata et al., 1982]. H2O2 decomposes into OH radicals via the reaction H2O2 + M → 2OH + M offering the possibility to increase NH3 conversion into other N-fixed species. Hence, the effect of H2O2 has been investigated by adding it into the bed corresponding to a 1:1 molar ratio to NH3 from the pyrolysis gas yield. Table 5.14 Main N-species concentrations at the gasifier outlet with and without H2O2 addition (*) H2O2/NH3 [NH3] [HCN] [NO] molar basis (ppmv) (ppmv) (ppmv) 0 154 2 6 1 152 2 7 (*) The reaction mechanism used did not involve HCN hydrolysis nor the complete char combustion reaction

H2O2 addition has no significant effect on fuel-N conversion (see table 5.14). Only approximately 1% fuel-N conversion increase is the result, compared to the base case. CH4 addition

CH4 has been added to the bed and freeboard in order to check whether the HCN is formed directly from conversion of NH3 via CH3 radicals according to the following path (Kilpinen et al., 1999): CH 3 OH, + H NH 3 ⎯+⎯ ⎯ ⎯→ NH i ⎯+⎯ ⎯→ HCN

CH4 addition quantities have been summarised in table 5.15 Table 5.15 CH4 amounts added Stream added to Bed Bed Freeboard Freeboard

Amount of added CH4 (kg/s) Base case x 2 Base case x 10 Base case x 5 Base case x 10

A significant decrease in the NH3 conversion has been obtained when CH4 was added into the bed, while no appreciable changing has been reported when adding CH4 into the freeboard. The large amount of CH4 added into the bed probably caused the competition of radicals between CH3 radicals and NH3, thus decreasing the total ammonia reduction percentage. However, in any case no extra HCN has been formed.

181

100 Bed Freeboard

NH3 unconverted (%)

98 96 94 92 90 88 0

x2

x10

Figure 5.16 Simulated NH3 conversion with and without CH4 addition into the bed. 5.4

Conclusions

The possibility to convert NH3 to N2 by varying process conditions or by adding amounts of specific compounds to gasification gases has been studied theoretically using a plug flow model study with detailed nitrogen chemistry for woody biomass, assuming fast pyrolysis. The product yields of this fast pyrolysis subproces were obtained from fluidised bed pyrolysis experiments reported in literature. A sensitivity analysis has been carried out by changing only one parameter for each simulation group (pressure, residence time, NO addition, etc.). Besides, an N-mass balance over the reactor sections (bed and freeboard) has been accomplished, obtaining a clear N-partitioning between the fixed nitrogen compounds. The results can be summarised as follows:

• • • •

•

• • •

182

NH3, which represents the major fuel-NOx precursor in biomass gasification, is a very stable compound, which is hardly converted in PFB gasifiers. The conversion of NH3 into N2 can only be slightly favoured by increasing the temperatures up to 1200 K. However, above 1150 K sintering of bed particles might occur with alkali containing biomass fuels, posing a limit to higher reactor temperatures. The NH3 conversion is only slightly dependent on reactor pressure. A minimum in ammonia conversion was obtained around 2.5 bar. At higher pressures (10 bar) the NH3 conversion slightly increased (just 6 ppm less when varying the pressure from 5.1 to 10.2 bar). More relevant NH3 conversions can be reached when adding NO or NO2 into the bed. The ammonia concentration decreased from 154 to 121 ppm when adding NO or NO2 in 1:1 ratio to NH3 on molar basis. On the other hand, HCN was formed (30 ppm) and unreacted NO was also predicted among the undesired emissions (114 ppm). Addition of O2 (as secondary air or as primary air via an increased air to fuel ratio, or λ) favours the ammonia conversion significantly. However, the main nitrogen species was NO rather than N2, or in the most favourable case (λ → 1, thus stoichiometric combustion) is converted for 50% into NO and 50% into N2. The addition of secondary air lowers the already low energy content of the produced LCV-gas. The presence of a high concentration of CH4 in the bed part of the gasifier reduces the NH3 conversion, probably due to the competition for radicals between CH4, its intermediates (mainly CH3 radicals) and NH3. Gas residence time in the reactor practically did not affect the fuel-N conversion. Destruction of NH3 is only taking place in the presence of O, H and OH radicals, which react very fast in the initial part of the bed. H2O2 and H2O (steam) addition into the bed did not affect at all NH3 conversion.

Chapter 6 Comparison of modelling results and experiments 6.1 Choices made for the comparison between model and experiments In this chapter, a comparison is made between the results obtained from pressurised fluidised bed gasification experiments, carried out using the 1.5 MWth test rig at Delft University of Technology and the 50 kWth installation at Stuttgart university, described in chapter 3, and simulation results using the gasifier model presented in chapter 5. The FG-DVC model (biomass version, [Chen et al., 1998]), as described in paragraph 2.4.2.2, has been used as the sub-model for the (flash) pyrolysis sub-process in this chapter. TG-FTIR experiments, described in paragraph 4.4, have been performed for the fuels actually applied in this study to produce the input data for the FG-DVC pyrolysis sub-model. As no kinetic data were available for H2 production during pyrolysis, because of limitations of the applied infrared spectroscopy analysis method (FTIR), the H2 yield as a function of the temperature under fluidised bed pyrolysis conditions was taken from literature data. The H2 yields of miscanthus and wood were taken from data published by [Van den Aarsen, 1985], see also table 5.2. For brown coal experimental fluidised bed H2 devolatilisation yield data published by [Rüdiger, 1997] were applied. No data were available for the nitrogen (N2) evolution during fast pyrolysis and the yield of this component was neglected, which is a reasonable assumption. 6.2 Gasification experiments compared with model results In this section both experiments performed at comparatively large scale, applying the 1.5 MWth Delft PFBG test rig and small scale, using the 50 kWth Stuttgart DWSA test rig, are compared with simulations. For the PFBG gasification simulations, experiments have been selected in which the axial species concentration profiles in the freeboard have been determined. This was not the case for the DWSA tests, as there was no possibility to measure in the reactor but only in a position downstream of the ceramic filter unit. 6.2.1 TUD PFBG experimental and simulation results Figure 6.1 a presents the results of the measurements and simulation for the PFB miscanthus gasification experiment 990107_2 performed at 0.4 MPa with respect to the major gaseous products. More details and the results of other miscanthus gasification experiments and accompanying simulations can be found in table 6.1. The volumetric concentrations are expressed as a function of the residence time as calculated by the model, the bed and freeboard sections are indicated. As can be seen, oxygen is consumed in the lower bed section. This is confirmed by e.g. [Kilpinen et al., 2002]. [Yoon et al., 1978] have shown that for a fixed bed coal gasifier the combustion and devolatilisation zones are also physically small compared to gasification/reduction zone. The reactions that play a major role in this combustion sub-process are homogeneous hydrogen combustion and heterogeneous carbon combustion. The carbon combustion reaction is governed by mass transfer limitation, as the particle size is such that diffusion still governs the conversion for the fast carbon combustion process. The main gasification reaction is the heterogeneous water-gas reaction, as can be seen in the figure, where the H2O concentration is decreasing in the bed zone. As a main species CO2 is initially increasing in concentration, due to the heterogeneous carbon oxidation, which is assumed to be complete. Higher in the bed section (at longer residence times), the CO2 concentration is slightly decreasing as a consequence of the heterogeneous Boudouard reaction. The CH4 concentration in the bed zone is increasing due to the cracking of tar, which in the model is assumed to result in CO, CH4 and solid carbon. The carbon-hydrogen reaction leading to CH4 plays a minor role in the gasification process.

183

0.25

H2

Freeboard

0.25

CO

0.2 Volume fraction (-)

O2 H2O CH4 CO2

H2O model

H2Omodel

H2 exp.

0.2

H2O exp. CO exp. CH4 exp. CO2 exp.

0.15

0.15

CO2 model CO2 model

0.1

0.1

CO model COmodel

0.05

0

0

H2 model

O2 model

0.05

O2 model CH4 model H2 model

CH4 model

Volume fraction (-)

Bed

Outlet filter

0.01

0.1

1

Residence time (s)

0

2

4

6 8 10 Residence time (s)

12

14

16

a

b

Figure 6.1 Miscanthus gasification - main species behaviour predicted by the model for entire reactor (a) and lower bed zone (b) and comparison with experimental set point 990107_2 (a); Tbed = 1020 K, Tfreeboard = 993 K, p= 0.4 MPa, λ=0.45, other relevant experimental conditions see table 4.4b (chapter 4). The agreement between the main gas constituents' model predictions and experimental data is quite good, although CO is underpredicted and CO2 slightly overpredicted by the model. This is probably mainly due to the fact that oxygen bound in tar, which is a major constituent of the initial pyrolysis product spectrum, is assumed to be converted into CO and not into CO2. Also, in the model it is assumed that the oxygen-containing light hydrocarbon species, like carboxylic acids, which are formed in the pyrolysis and predicted by the FG-DVC pyrolysis model, are treated in the same way as tar components, following the same cracking reaction. It is known, however, that these carboxylic acids tend to generate CO2 via decarboxylation at already comparatively low temperatures. The observed deviation between simulation and measurements regarding the hydrogen concentration can also be due to the fact that this species yield is based on experimental pyrolysis yields for woody biomass, obtained from literature, and not on miscanthus. In the freeboard no significant changes of the measured and calculated concentrations occur, as is clearly seen in figure 6.1 a. Heterogeneous reactions don’t play a role anymore due to the negligible carbon hold up in this reactor zone. In figure 6.1 b the simulation results for the lower bed section are given. 2 Bed

4

6

8

10

Freeboard

NH3

0.001

HNCO NO

0.0001

Volume fraction (-)

12

1E-05 N2O

1E-06

HCN

14

16

Outlet filter NH3 exp.

Residence time (s) 0.01

0.1

1 1.E-02

NH3

HCN exp.

1.E-03

HCN HNCO NO2

1.E-04

NO

1.E-05

H2NO

N2O

1.E-06

HNO

1E-07

HCNO HNO HOCN C2N2

1E-08 1E-09

NO2

1E-10

NH2

N2H3 HOCN NH2 N2H4 HCNO C2N2

1.E-07 1.E-08

N2H2 NO3

1.E-09

H2CN

1.E-10

NH

1E-11 1E-12 1E-13

Volume fraction (-)

0

0.01

1.E-11 NO3 NH

N2H4

N2H2 N2H3

HNNO CN

1.E-12 1.E-13

Residence time (s)

a

b

Figure 6.2 Nitrogen species behaviour predicted by the model for entire reactor (a) and lower bed zone (b) and comparison with experimental set point 990107_2 (a); Tbed = 1020 K, Tfreeboard = 993 K, p= 0.4 MPa, λ=0.45, other relevant experimental conditions see table 4.4b (chapter 4).

184

In figure 6.2 the results of experiments and simulations with respect to the nitrogen species for miscanthus gasification at 0.4 MPa are shown. The FG-DVC pyrolysis model applied for this case yields comparable amounts of HCN, HNCO and NH3, which is the input for further reaction kinetic calculations in bed and freeboard zone. The HCN concentration in the initial part is predicted to decrease, hereby forming NH3. This is mainly due to the reaction with water, which is heterogeneously catalysed by char. The predicted NH3 concentration agrees quite well with experimental data, whereas the calculated HCN concentration is somewhat lower than the measured value. NH3 is in both cases the main bound nitrogen carrier. HNCO was not detected experimentally by FTIR analysis, although it is predicted to exist. It is possible that this species is catalytically hydrolysed with water to NH3. NO was also predicted by the model but was not found experimentally. Bed

Freeboard

H2 O2 H2O CO CH4 CO2

Outlet Filter

Volume fraction (-)

0.2

0.15

0.18 0.16

H2 exp. H2O exp. CO exp. CH4 exp. CO2 exp.

H2O model

0.12

CO model

0.1

CO model CO2 model

0.1

0.14

H2O model

O2 model

0.08

CH4 model

0.06

CH4 model

0.05

CO2 model

0.04

H2 model

0.02

H2 model

0 0.01

0.1

O2 model

0 0

2

4

6

8

10

Volume fraction (-)

0.25

1

Residence time (s)

12

14

Residence time (s)

a

b

Figure 6.3 Wood gasification - main species behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison (a) with exp. 020111; Tbed = 1175 K, Tfreeboard = 1037 K, p= 0.35 MPa, λ=0.39, other exp. conditions see table 4.5 (chapter 4).

0.001

2

4

6

8

10

Freeboard

Bed

16 NO NO2 N2O

NH3

NH3

NH3 exp. HCN exp.

HNCO NO HCN

1E-05 1E-06

NH

N2O

1E-08

HNO HOCN NH2

1E-10 H2NO NO2

1 1.0E-01

NH3 HCN CN NH

NH2

HNO

N2H2

HNCO

H2CN

HOCN

CNO

HNO HNCO

H2NO

HOCN

HONO

HCNO

NO3

H2NO

C2N2

1.0E-02

NO2 N2O

NH2

HCNO

1E-09

0.1 NO

N2H2 CNO

HCNO

0.01

HCN CN

H2CN

1E-07

1E-11

14

Outlet Filter

0.0001 Volume fraction (-)

12

HONO NO3

1.0E-03

HCN NH3 HNCO

1.0E-04

NO2

1.0E-05

NO

1.0E-06

H2NO HOCN N2O HNO HCNO C2N2 NH2

1.0E-07 1.0E-08

HONO CNO NO3 H2CN NH

1.0E-09

1.0E-11

N2H4

C2N2

N2H2

N2H3

N2H4

CN

NH3 exp.

N2H3

1.0E-10

1.0E-12

HCN exp.

HONO C2N2 CNO

1E-12 1E-13

Volume fraction (-)

0 0.01

1.0E-13

Residence time (s)

Residence time (s)

a

b

Figure 6.4 Wood gasification – model predicted nitrogen species behaviour for the entire reactor (a) and the lower bed zone (b) and comparison with exp. 020111(a); Tbed = 1175 K, Tfreeboard = 1037 K, p= 0.35 MPa, λ=0.39, other exp. conditions see table 4.5 (chapter 4). For wood gasification applying the Delft PFBG test installation, the main results of both simulation and experiments are given in figure 6.3 and 6.4 for the most abundant gas constituents and the nitrogen species. More details and the results of other wood pellet gasification experiments and related 185

simulations can be found in table 6.2. The agreement between the simulated and measured concentrations of H2 and CH4 is quite good, but CO and CO2 are less well predicted. The agreement for these components is much less than for the miscanthus gasification case presented in figures 6.1 and 6.2. This is probably due to the tar cracking kinetics used: only CO, CH4 and C are defined as products, whereas CO2 could also be a product. For the case of wood the pyrolysis tar yield resulting from the FG-DVC biomass pyrolysis model is significantly higher than for miscanthus. That can be the reason why predictions are less accurate, compared to the miscanthus gasification experiments. The order of magnitude of the predicted NH3 concentration is still good. The HCN concentration is predicted to be lower than the measured values in the freeboard and downstream of the filter. It can be that the heterogeneous HCN hydrolysis reaction in the bed section is slower than assumed in the model. Both NO and HNCO are predicted to exist in the freeboard, but these components were not detected by FTIR analysis. The predicted levels are above 10 ppm, which should be possible to detect. Possibly these species are catalytically reacting, whereby NO is reduced to N2 and HNCO is most probably hydrolysed to NH3. 0.18

CO model

0.14 Volume fraction (-)

H2 O2 H2O CO CH4 CO2 H2 exp H2O exp. CO exp. CH4 exp. CO2 exp.

Outlet Filter

0.12 CO2 model

0.1

H2O model

0.08

H2 model

0.06

0.18

H2O model

0.16 0.14 0.12 0.1 CO2 model

0.08

O2 model

0.04

0.06 0.04

H2 model

CO model

0.02

CH4 model

0.02

0.001

CH4 model O2 model

0 0

2

4

6

8

10

12

Volume fraction (-)

Freeboard

Bed

0.16

0

0.01

0.1

1

Residence time (s)

14

16

Residence time (s)

a b Figure 6.5 Brown coal - main species behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with exp. set point 020416 (a); Tbed = 1181 K, Tfreeboard = 1140 K, p= 0.35 MPa, λ=0.47, other exp. conditions see table 4.6 (chapter 4). 2

4

6

8

10

12

14

Volume fraction (-)

1.E-03

Bed

16

Outlet Filter

NH3 exp.

Freeboard

1.E-04

NO model

1.E-05

HCN model

NO

NO2 N2O

NH3 model

H3N

CHN N

HNCO model

HCN exp.

H3N

NH2

N

1.E-07

H2CN

NH2 model HCNO model HOCN model NO2 model

1.E-09 1.E-10 1.E-11

HONO model

1.E-12

C2N2 model H2CN model NNH model

1.E-13

N2O

CHN

N2H2

H2NO model

NO2

CN

1.E-06

1.E-08

0.01

CN NH NH2

NNH

CNO

N2H2

HNO

H2CN

HNCO

HOCN HCNO

CNO HNO HNCO

HOCN HCNO

H2CN

H2NO

HONO

HNNO

CN

C2N2

HONO

C2N2

N2H4

N2H4

N2H3

N2H3

1.E-03 1.E-04 1.E-05 1.E-06 1.E-07 1.E-08

CNO NO3 NH

HNNO

NO3

1.E-09 1.E-10

NNH HNNO N

1.E-11 1.E-12 1.E-13

NH3 exp. HCN exp.

1 1.E-02

NH3 NO HNCO NO2 N2O HCN HCNO H2NO HNO HOCN NH2 N2H3 HONO C2N2 N2H4 N2H2

H2NO

NO3

0.1

NO

NH

NNH

HNO model N2O model

0.001

Volume fraction (-)

0 1.E-02

Residence time (s)

Residence time (s)

a

b

Figure 6.6 Brown coal - nitrogen species behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with exp. 020416 (a); Tbed = 1181 K, Tfreeboard = 1140 K, p= 0.35 MPa, λ=0.47, other exp. conditions see table 4.6 (chapter 4).

186

For the low rank Rhenish brown coal, figures 6.5 and 6.6 show the calculated and measured results for the PFBG experiment 020416, performed at 0.35 MPa. More details and the results of other brown coal gasification experiments and related simulations are given in table 6.3. As can be seen in figure 6.5b, the oxygen is reacting relatively fast. This is a result of the smaller particles applied, as compared to the biomass used in the PFBG experiments, which leads to a faster heterogeneous oxidation process in which reaction kinetics and mass transfer limitation are both taken into account. The agreement is quite good for the main gas constituents, although larger differences occur between the simulated and the measured concentrations of CO and CO2. This can be attributed again to the assumed tar cracking mechanism with only CO, CH4 and C as products. The tar yield of the pyrolysis step, however, is low and char formation is more favoured when compared to pyrolysis of biomass and especially wood: tar yield wood > miscanthus > brown coal. This is probably the main reason why these brown coal simulation results regarding the main gas constituents are in better agreement with the measured values. The agreement for the nitrogen compounds is quite well, although no NO could be detected in the experiments whereas this species is predicted to occur in the order of 100 ppmv. Probably catalytic effects of ash reduce the NO. HNCO is predicted in the shown case with a concentration of less than 10 ppmv and for all simulated cases in the range of 5-15 ppmv, it was however not be detected by FTIR. This can be due to the concentration level predicted, which is near the detection limit for this type of minor species. Catalytic effects of ash could also promote hydrolysis of HNCO into NH3.

187

188

Experiment/ Simulation Probe position Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P8.1 Model P8.1 Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1 Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1 Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1

188

n.d.: not detectable

020513

020429

990107_2

990107_1

981216

Set point

[H2] (vol.%, wet) 5.8 3.36 5.4 3.36 4.2 3.34 4.6 3.22 4.5 3.24 5.0 3.27 5.8 3.26 5.2 2.01 5.3 2.01 5.1 2.01 5.4 1.99 6.0 5.97 5.9 5.97 5.9 5.97 6.6 5.94 6.5 6.67 6.6 6.68 5.2 6.66

n.m.: not measured

[CO] (vol.%, wet) 6.0 11.55 6.0 11.54 4.6 11.49 6.0 10.92 5.7 10.88 5.8 10.85 4.6 10.82 5.1 9.25 5.1 9.24 5.0 9.22 4.8 9.19 10.7 16.98 10.2 16.96 10.3 16.93 9.1 16.87 9.7 18.92 9.9 18.88 9.0 18.81

[CH4] (vol.%, wet) 3.2 4.76 3.3 4.76 2.6 4.74 3.3 3.24 3.1 3.24 3.2 3.24 2.9 3.23 2.5 4.35 2.6 4.35 2.7 4.34 2.6 4.33 3.7 5.90 3.6 5.90 3.7 5.88 3.6 5.86 3.4 4.53 3.5 4.53 3.3 4.51

[C2H4] (vol.%, wet) 0.6 0.05 0.6 0.05 0.4 0.04 0.5 0.09 0.5 0.09 0.5 0.09 0.3 0.09 0.5 0.05 0.5 0.05 0.5 0.05 0.4 0.05 0.8 0.09 0.8 0.09 0.8 0.09 0.6 0.09 0.7 0.12 0.7 0.12 0.6 0.12

[C2H6] (vol.%, wet) 0.2 0.01 0.2 0.01 0.1 0.01 0.1 0.02 0.1 0.02 0.1 0.02 0.2 0.02 0.1 0.02 0.1 0.02 0.1 0.01 0.1 0.01 0.1 0.02 0.07 0.02 0.07 0.02 0.1 0.02 n.m. 0.02 0.04 0.02 n.m. 0.02

[CO2] (vol.%, wet) 16.5 12.38 16.2 12.38 15.1 12.33 14.9 11.20 14.7 11.22 15.1 11.24 14.6 11.21 14.5 13.32 14.9 13.31 14.6 13.29 15.2 13.25 14.9 11.28 14.9 11.27 14.4 11.26 15.4 11.22 15.5 10.33 14.9 10.32 15.0 10.29

[H2O] (vol.%, wet) 19.4 18.84 21.0 18.82 19.0 18.75 16.4 20.15 20.6 20.11 18.4 20.07 20.5 20.01 19.2 19.70 18.9 19.68 20.6 19.65 18.7 19.59 11.7 10.36 12.8 10.34 13.6 10.32 14.4 10.28 14.8 10.81 13.6 10.78 15.7 10.75

[N2] (vol.%, wet) 47.5 48.23 46.6 48.28 53.2 48.47 53.5 50.40 49.8 50.44 51.1 50.47 50.4 50.60 51.9 50.46 51.8 50.52 50.6 50.58 52.0 50.74 51.7 48.65 51.5 48.71 51.0 48.77 49.4 48.96 49.2* 47.79 50.5* 47.85 50.6 48.04

[Ar] (vol.%, wet) 0.7 0.51 0.7 0.51 0.6 0.51 0.6 0.52 0.6 0.52 0.5 0.52 0.5 0.51 0.7 0.55 0.6 0.55 0.7 0.55 0.5 0.55 n.m. 0.53 n.m. 0.53 n.m. 0.53 0.6 0.53 n.m. 0.51 n.m. 0.51 0.6 0.51

[C2H2] (ppmv, wet) 211 7 181 7 65 7 128 32 117 32 91 31 21 31 190 15 189 15 184 15 91 15 534 37 489 37 444 36 121 36 634 59 609 58 224 58

[NH3] (ppmv, wet) 1684 2770 1692 2780 1771 2770 2020 2220 1986 2220 1977 2220 1969 2210 1938 2560 1917 2560 1848 2560 1948 2550 1686 2000 1994 2020 2085 2030 2191 2020 2139 2790 2220 2790 1636 2780

[HCN] (ppmv, wet) 94 0.2 88 0.2 32 0.2 63 4 72 6 81 7 21 7 63 0.4 75 0.4 67 0.5 35 0.5 157 0.7 173 1 174 1 47 1 332 4 372 4 131 4

[HNCO] (ppmv, wet) n.d. 159 n.d. 152 n.d. 145 n.d. 75 n.d. 57 n.d. 44 n.d. 44 n.d. 156 n.d. 151 n.d. 147 n.d. 147 n.d. 109 n.d. 88 n.d. 71 n.d. 71 n.d. 49 n.d. 37 n.d. 37

Table 6.1 Overview of the PFBG gasification results with miscanthus as fuel; comparison of model and experimental results

[NO] (ppmv, wet) 0 40 0 39 0 37 0 148 0 129 0 113 0 113 0 71 0 69 0 67 0 67 0 44 0 42 0 40 0 39 0 58 0 52 0 52

[NO2] (ppmv, wet) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

[N2O] (ppmv, wet) 0 2 0 1 0 1 0 2 0 0.8 0 0.4 0 0.4 0 4 0 4 0 3 0 3 0 0.9 0 0.6 0 0.4 0 0.4 0 0.7 0 0.4 0 0.4

92.3 99.0

Carbon Conv. (%) 91.6 99.56 90.4 99.3 90.3 99.2 89.7 98.8

189

Experiment P2.1 Model P2.1 I Model P2.1 II Experiment P8.1 Model P8.1 I Model P8.1.II Experiment P1.1 Model P1.1 I Model P1.1 II Experiment P2.1 Model P2.1 I Model P2.1 II Experiment P8.1 Model P8.1 I Model P8.1 II Experiment P8.1 Model P8.1 I Model P8.1 I Experiment P1.1 Model P1.1 I Model P1.1 II Experiment P3.1 Model P3.1 I Model P3.1 II Experiment P8.1 Model P8.1 I Model P8.1 II Experiment P2.1 Model P2.1 I Model P2.1 II Experiment P8.1 Model P8.1 I Model P8.1 II

011127

7.2 5.04 7.07 7.3 5.05 7.14 5.8 3.02 6.24 5.8 3.09 6.44 5.2 3.12 6.56 6.4 5.63 9.55 5.6 4.54 7.22 6.0 5.17 8.21 6.4 5.13 8.15 6.3 3.92 6.56 6.1 4.18 6.99

[H2] (vol.%, wet)

n.m.: not measured

11.7 6.79 18.59 11.3 16.52 18.22 10.2 12.34 17.02 9.7 12.25 16.78 9.1 12.02 15.76 10.7 14.07 18.04 10.4 13.41 15.92 10.3 12.60 14.72 9.6 12.50 14.60 8.9 10.33 10.64 7.2 9.89 9.98

[CO] (vol.%, wet) 4.1 3.94 1.61 4.0 3.91 1.61 3.5 5.51 2.50 3.5 5.50 2.52 3.3 5.43 2.51 3.9 4.30 2.43 3.9 4.01 2.30 4.0 4.16 2.47 3.9 4.13 2.45 3.7 3.18 1.38 3.3 3.20 1.45

[CH4] (vol.%, wet) 0.9 0.39 0.15 0.9 0.37 0.14 0.7 0.42 0.18 0.6 0.43 0.17 0.6 0.42 0.16 0.3 0.46 0.09 0.5 0.60 0.21 0.4 0.55 0.14 0.3 0.54 0.14 0.4 0.53 0.17 0.3 0.51 0.13

[C2H4] (vol.%, wet) 0.2 0.04 0.02 0.2 0.04 0.02 0.07 0.01 0.01 0.08 0.01 0.01 0.09 0.01 0.01 0.07 0.02 0.006 n.m. 0.03 0.01 n.m. 0.02 0.009 0.09 0.02 0.009 n.m. 0.03 0.01 0.1 0.03 0.01

[C2H6] (vol.%, wet) 14.5 8.87 6.16 14.8 8.85 6.26 16.3 10.98 6.52 16.4 11.04 6.73 14.8 10.97 6.86 14.9 10.13 7.35 14.9 9.17 6.79 15.2 9.96 7.96 15.4 9.88 7.90 17.2 9.24 8.03 15.3 9.50 8.53

[CO2] (vol.%, wet)

I: H2 pyrolysis yield based on [El Asri, 1999]; II H2 yield [Van den Aarsen ,1985]

n.d.: not detectable

020212

020205

020129

020111

Experiment/ Simulation Probe position

Set point 12.8 15.01 17.58 11.8 14.76 17.22 9.9 14.21 16.41 11.3 14.12 16.17 19.1 13.87 15.76 12.2 12.54 12.09 15.7 15.72 16.65 12.2 14.90 15.45 13.9 14.78 15.33 16.3 23.77 25.05 18.5 23.21 24.26

[H2O] (vol.%, wet) 47.9 49.23 48.18 48.9 49.81 48.76 52.5 52.64 50.42 51.8 52.71 50.50 47.0 53.32 51.12 50.8 52.10 49.85 48.9 51.59 50.16 51.8 51.70 50.31 49.7 52.11 50.69 47.2 48.31 47.56 48.5 48.81 48.07

[N2] (vol.%, wet) 0.7 0.55 0.53 0.8 0.54 0.53 0.8 0.59 0.57 0.8 0.59 0.57 0.7 0.58 0.56 0.7 0.56 0.54 n.m. 0.57 0.55 n.m. 0.57 0.55 0.7 0.56 0.55 n.m. 0.53 0.52 0.7 0.53 0.52

[Ar] (vol.%, wet) 429 417 124 277 405 117 593 1590 343 413 1560 328 221 1510 310 78 1500 197 449 1970 476 208 1810 349 44 1790 346 167 1110 253 59 1050 207

[C2H2] (ppmv, wet) 1022 1040 955 1051 1030 941 459 945 770 438 947 762 483 937 747 381 512 403 990 1610 1320 1116 1600 1310 1086 1590 1300 225 462 437 223 457 432

[NH3] (ppmv, wet) 160 6 9 144 7 10 57 3 11 53 4 15 41 5 17 27 8 18 131 11 20 90 14 24 63 14 24 32 12 13 15 13 13

[HCN] (ppmv, wet)

Table 6.2a Overview of the PFBG gasification results with wood pellets as fuel; comparison of model and experimental results

n.d. 12 7 n.d. 10 7 n.d. 23 8 n.d. 18 7 n.d. 15 7 n.d. 6 4 n.d. 20 12 n.d. 17 11 n.d. 17 11 n.d. 4 3 n.d. 4 2

[HNCO] (ppmv, wet) 0 32 37 0 28 30 0 18 67 0 15 56 0 13 46 0 4 17 0 12 28 0 2 7 0 2 6 0 9 10 0 7 7

[NO] (ppmv, wet) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

[NO2] (ppmv, wet) 0 0.02 0.05 0 0.01 0.03 0 0 0.02 0 0 0.01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

[N2O] (ppmv, wet)

98.1 89.2 79.8 97.3 98.4 90.6 96.6 99.0 97.5 97.3 93.4 86.9 98.5 87.7 75.7

Carbon Conv. (%)

190

Experiment P1.1 Model P1.1 I Model P1.1 II Experiment P2.1 Model P2.1 I Model P2.1 II Experiment P3.1 Model P3.1 I Model P3.1 II Experiment P8.1 Model P8.1 I Model P8.1 II

Experiment P1.1 Model P1.1 I Model P1.1 II Experiment P2.1 Model P2.1 I Model P2.1 II Experiment P3.1 Model P3.1 I Model P3.1 II Experiment P8.1 Model P8.1 I Model P8.1 II

020220

020226

5.2 2.65 3.87 5.5 2.73 4.04 4.9 2.80 4.22 5.4 2.79 4.20

5.8 3.76 6.59 5.9 3.78 6.63 5.9 3.79 6.66 6.2 3.75 6.59

[H2] (vol.%, wet)

n.m.: not measured

7.4 9.88 10.69 7.1 9.78 10.47 6.6 9.68 10.25 5.8 9.66 10.23

8.0 11.72 14.31 7.9 11.68 14.24 7.5 11.65 14.18 7.4 11.53 14.03

[CO] (vol.%, wet)

2.8 2.71 0.98 2.8 2.71 1.00 2.6 2.72 1.02 2.5 2.71 1.02

3.1 5.65 3.07 3.1 5.64 3.07 3.2 5.63 3.07 3.0 5.57 3.04

[CH4] (vol.%, wet)

0.6 0.39 0.23 0.6 0.39 0.22 0.5 0.39 0.21 0.5 0.39 0.21

0.7 0.18 0.15 0.7 0.18 0.15 0.7 0.18 0.15 0.6 0.18 0.15

[C2H4] (vol.%, wet)

0.04 0.02 0.02 0.04 0.03 0.02 0.05 0.03 0.02 0.07 0.03 0.02

0.1 0.03 0.04 0.1 0.03 0.04 0.1 0.03 0.04 0.1 0.03 0.04

[C2H6] (vol.%, wet)

16.0 9.27 7.45 16.6 9.34 7.64 14.7 9.41 7.83 15.1 9.39 7.81

14.7 10.27 7.55 15.1 10.27 7.58 14.5 10.27 7.62 14.9 10.16 7.54

[CO2] (vol.%, wet)

I: H2 pyrolysis yield based on [El Asri, 1999]; II H2 yield [Van den Aarsen ,1985]

n.d.: not detectable

Experiment/ Simulation Probe position

Set point

13.1 24.80 27.04 11.4 24.64 26.80 21.4 24.52 26.56 22.2 24.45 26.49

21.0 21.04 22.33 20.2 20.99 22.26 19.1 20.95 22.18 18.7 20.72 21.95

[H2O] (vol.%, wet)

54.1 49.63 49.10 55.3 49.71 49.19 48.7 49.79 49.27 47.9 49.93 49.41

46.1 46.77 45.40 46.4 46.84 45.47 48.4 46.92 45.55 48.5 47.49 46.12

[N2] (vol.%, wet)

0.6 0.56 0.56 0.6 0.56 0.56 0.6 0.56 0.56 0.6 0.56 0.55

0.6 0.52 0.51 0.6 0.52 0.51 0.6 0.52 0.51 0.5 0.52 0.50

[Ar] (vol.%, wet)

623 645 266 429 628 251 338 611 238 86 609 237

461 114 57 330 114 56 290 112 55 115 111 55

[C2H2] (ppmv, wet)

334 406 398 382 408 397 379 409 395 428 408 394

236 479 474 243 482 476 247 485 476 257 480 471

[NH3] (ppmv, wet)

61 2 6 60 3 6 55 4 7 27 4 7

30 0.6 1 30 0.6 1.3 30 0.8 1.6 19 0.8 1.6

[HCN] (ppmv, wet)

Table 6.2b Overview of the PFBG gasification results with wood pellets as fuel; comparison of model and experimental results

n.d. 12 3 n.d. 8 3 n.d. 6 3 n.d. 6 3

n.d. 27 7 n.d. 22 5 n.d. 18 3 n.d. 18 3

[HNCO] (ppmv, wet)

0 24 24 0 23 21 0 21 19 0 21 19

0 16 21 0 16 20 0 15 19 0 15 19

[NO] (ppmv, wet)

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0

[NO2] (ppmv, wet)

0 0 0 0 0 0 0 0 0 0 0 0

0 0 0.1 0 0 0 0 0 0 0 0 0

[N2O] (ppmv, wet)

98.7 92.1 80.2

97.9 97.9 90.9

-

Carbon Conv. (%)

191

Experiment/ Simulation Probe position

Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1

Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1

Experiment P1.1 Model P1.1 Experiment P2.1 Model P2.1 Experiment P3.1 Model P3.1 Experiment P8.1 Model P8.1

Set point

020319

020409

020416

n.d.: not detectable

13.7 14.04 13.6 13.93 13.6 13.83 10.9 13.78

11.4 8.97 11.5 8.91 11.4 8.84 9.5 8.81

14.2 19.66 13.8 19.64 14.6 19.62 14.1 19.45

[CO] (vol.%, wet)

6.5 6.19 6.5 6.26 6.6 6.32 7.5 6.29

9.5 5.15 10.0 5.20 9.9 5.24 10.1 5.22

8.5 8.82 8.4 8.80 8.8 8.78 9.2 8.71

[H2] (vol.%, wet)

n.m.: not measured

0.9 0.48 0.9 0.49 0.9 0.50 0.9 0.50

1.5 1.99 1.6 2.00 1.6 2.00 1.5 2.00

1.5 1.89 1.4 1.90 1.4 1.91 1.4 1.89

[CH4] (vol.%, wet)

0.1 0.03 0.08 0.023 0.07 0.020 0.05 0.020

0.2 0.10 0.2 0.09 0.2 0.09 0.09 0.09

0.1 0.06 0.06 0.05 0.05 0.05 0.02 0.05

[C2H4] (vol.%, wet)

0.008 0.001 0.004 0.001 0.005 0.001 0.02 0.001

0.05 0.02 0.04 0.02 0.04 0.02 0.05 0.02

0.01 0.01 0.004 0.01 0.004 0.01 0.021 0.01

[C2H6] (vol.%, wet)

12.7 10.33 12.8 10.40 12.7 10.46 14.3 10.43

13.4 11.90 14.0 11.95 14.4 11.99 14.2 11.95

11.0 9.50 11.1 9.48 11.3 9.47 11.7 9.39

[CO2] (vol.%, wet)

7.9 9.11 8.9 9.02 7.6 8.92 7.1 8.89

16.9 17.86 13.1 17.78 11.9 17.71 14.5 17.64

6.7 3.47 7.1 3.46 6.1 3.46 7.2 3.43

[H2O] (vol.%, wet)

58.0 59.04 57.1 59.11 58.3 59.19 58.5 59.33

46.9 53.18 49.4 53.24 50.5 53.30 49.3 53.46

57.3 55.72 57.4 55.77 57.0 55.83 55.5 56.20

[N2] (vol.%, wet)

n.m. 0.69 n.m. 0.69 n.m. 0.69 0.7 0.68

n.m. 0.61 n.m. 0.61 n.m. 0.61 0.58 0.61

0.6 0.64 0.5 0.64 0.6 0.64 0.7 0.64

[Ar] (vol.%, wet)

60 56 49 47 29 40 0 40

49 62 32 60 16 57 0 57

49 26 36 24 7 22 0 22

[C2H2] (ppmv, wet)

1228 611 1125 588 1163 569 1016 567

1900 1940 1838 1920 1774 1900 1690 1890

1470 2250 1496 2230 1225 2220 1296 2220

[NH3] (ppmv, wet)

32 8 26 11 24 13 9 13

20 9 22 11 21 13 9 13

57 7 55 9 41 10 23 10

[HCN] (ppmv, wet)

Table 6.3 Overview of the PFBG gasification results with brown coal as fuel; comparison of model and experimental results

n.d. 5 n.d. 5 n.d. 5 n.d. 5

n.d. 10 n.d. 9 n.d. 9 n.d. 9

n.d. 15 n.d. 15 n.d. 15 n.d. 15

[HNCO] (ppmv, wet)

0 191 0 162 0 140 0 139

0 102 0 76 0 57 0 57

0 76 0 54 0 38 0 0

[NO] (ppmv, wet)

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0

[NO2] (ppmv, wet)

0 0.1 0 0 0 0 0 0

0 0.3 0 0.1 0 0 0 0

0 0.7 0 0 0 0 0 0

[N2O] (ppmv, wet)

91.3 81.5

90.4 77.4

94.9 97.1

Carbon Conv. (%)

191

6.2.2 IVD DWSA experimental and simulation results Figure 6.7 a and b show the simulated profiles and the measured filter outlet concentrations of the main gas components for the IVD DWSA test rig using Hambach brown coal as fuel (experiment 991115). More details and the results of other brown coal gasification experiments and accompanying simulations are presented in table 6.4. Due to the existing reactor configuration it was not possible to perform measurements in the bed zone nor in the freeboard, so only experimental data from a position downstream of the hight temperature ceramic candle filter unit are available. The agreement for the main gaseous species is reasonable, although the H2 concentration is underpredicted. This is probably the result of inaccuracies introduced in the simulations because the yield of hydrogen in the initial fast pyrolysis step is based on experimental values of [Rüdiger, 1997]. Different compositions of the fuel, in particular the ash composition, lead to a different product yield spectrum. The differences in the calculated and measured CO and CO2 concentrations are probably a result of the simplified tar cracking kinetics that is assumed in the model. A simple one-step reaction is assumed here, which leads to CO, CH4 and carbon as products. The kinetic rate data used for this reaction have been determined for a beech wood derived tar [Van den Aarsen, 1985]. Because tar from brown coal is different in chemical nature from biomass-derived tar (e.g. a lower oxygen content is expected), this is also a source of inaccuracy. The agreement between the simulations and the experimental results is quite good for NH3 as can be observed in figure 6.8a. HCN was experimentally not detected. The simulations for this case predict approximately 10 ppm, which is close to the detection limit of the FTIR analyser used. For brown coal in general comparatively low HCN concentrations are found, which is attributed to the relatively high Ca content of the fuel. This very probably contributes to enhanced conversion into NH3, as has been pointed out in the literature review (paragraph 2.3.2).

192

Bed

Outlet Filter

Freeboard

Volume fraction (-)

0.14 0.12 CO2 model CO model

0.1 0.08 0.06

H2O model

H2 O2 H2O CO CH4 CO2 H2 exp. H2O exp. CH4 exp. CO exp. CO2 exp.

0.16 0.14 0.12 0.1

H2Omodel

0.08 0.06

CO2 model

H2 model

0.04

O2 model

0.04

COmodel H2 model CH4 model

0.02 0.001

CH4 model O2 model

0 0

10

20

30

Volume fraction (-)

0.16

0.02 0

0.01

0.1

1

Residence time (s)

40

50

Residence time (s)

a

b

Figure 6.7 Brown coal gasification - main species behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with exp.991115 (a); Tbed = Tfreeboard =1075 K, p= 0.51 MPa, λ=0.51, other exp. conditions see table 4.7(chapter 4).

1.E-03

0

10

20

Bed

Freeboard

30

40 NH3 model

50 Outlet Filter NH3 exp.

Volume fraction (-)

1.E-04 NO model HCN model HNCO model

H3N

N2O model

1.E-07

HNO model

1.E-08

H2NO model HCNO model NH2 model HOCN model

1.E-09 1.E-10

NO2 model N2H2 model

N2H3 model HONO model

1.E-12

0.01 NO

N2O H3N

NH

CHN

NH2

CN

NH

NH2

H2CN

N2H2

CNO

H2CN

HNO

CNO

HNCO

HOCN

HNO

HNCO

HOCN

HCNO

HCNO

H2NO

H2NO

HONO

HONO

NO3

0.1

NO3

C2N2

C2N2

N2H4

1.E-02

N2H4

N2H3

NO model

HCN model

1.E-03 1.E-04

NO2 model HNCO model H2NOmodel N2O model HNO model

1.E-05 1.E-06

NH2 model N2H3 model HOCN model HONO model HCNO model N2H4 model C2N2model CNO model

1.E-07 1.E-08

N2H2model NO3model

NH model

1.E-09 1.E-10

CN model H2CN model

1.E-11 1.E-12

N2H3

NH3 exp.

1

NH3 model

NO2

CN

N2H2

1.E-06

0.001

N2O

CHN

1.E-05

1.E-11

NO NO2

Volume fraction (-)

1.E-02

Residence time (s)

Residence time (s)

a

b

Figure 6.8 Brown coal gasification - fuel bound nitrogen compound behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with exp. 991115; Tbed = Tfreeboard =1075 K, p= 0.51 MPa, λ=0.51, other exp. conditions see table 4.7(chapter 4). The simulation- and experimental results for wood gasification (experiment 991206) are shown in figure 6.9 and 6.10, for the main- and nitrogen species respectively. Details of simulation and experimental results for this and other experiments can be found in table 6.4. As a consequence of feeder limitations, smaller fuel particle sizes were applied in comparison to the experiments using the Delft gasifier. As a result, the oxygen is consumed in mainly the hydrogen and carbon combustion reactions during a shorter time as compared to the Delft wood-fuelled gasification experiments, at temperatures even lower than in the Delft PFBG experiments. Discrepancies exist between the simulations and the measured values for CO and CO2 concentrations. The same explanation as given before, i.e. the application of a simple tar cracking kinetics, is probably the background of this difference. No HCN was measured in the wood-fuelled DWSA experiments. The predicted value is a few ppmv for all simulated cases sub-ppmv to 9 ppmv-, which is near the detection limit so this is possible. The NH3 concentration is well predicted and within the experimental error range of the FTIR instrument.

193

NO is predicted in the freeboard at a concentration of a few ppm till 12 ppmv. This is a concentration above the detection limit of the FTIR analyser. It is possible that heterogeneously catalysed reduction takes place to some extent, resulting in lower NO concentrations than predicted. HNCO concentrations predicted by the simulations are in the sub-ppm range, which cannot be detected by the FTIR instrument that has been used. Freeboard

Bed

Outlet Filter

0.18

Volume fraction (-)

0.16 0.14

H2O model CO model

0.12

H2 O2 H2O CO CH4 CO2 H2 exp. H2O exp. CO exp. CH4 exp. CO2 exp.

0.2 0.18 0.16

H2Omodel

0.14 0.12

0.1

0.1

CO2 model

0.08

0.08 O2 model

0.06

0.06

COmodel

H2 model

0.04

0.04

CO2 model H2 model

0.02

CH4 model

0.02

0.001

CH4 model O2 model

0 0

5

10

15

20

25

30

Volume fraction (-)

0.2

0

0.01

0.1

1

Residence time (s)

35

40

Residence time (s)

a

b

Figure 6.9 Wood gasification - main species behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with experimental set point 991206 (Labee wood gasification); Tbed = Tfreeboard =1097 K, p= 0.51 MPa, λ=0.48, other exp. conditions see table 4.7(chapter 4).

5 Bed

10

15

20

25

Freeboard

35

40

Outlet Filter NH3 model

1.E-04

NH3 exp.

NO

NO2 N2O

H3N

NO model

1.E-05

Volume fraction (-)

30

HCN model

1.E-06

HNCO model

1.E-07

CHN O NH NH2

HNO model N2O model

1.E-09

1.E-11

C2N2 model

HNCOmodel

CN

CNO

CNO HNO

HNCO

HOCN

HOCN

HCNO

HCNO

C2N2

H2NO

HONO

1.E-05

NOmodel NO2 model

1.E-06

H2NOmodel HCNOmodel HNOmodel N2Omodel NH2 model

1.E-07

HOCNmodel HONOmodel N2H3 model NO3 model CNOmodel

1.E-08 1.E-09

N2H4 model C2N2 model

1.E-10

NO3 C2N2

N2H2 model

N2H4

N2H4

CNmodel

N2H3

N2H3

NH3 exp.

1

1.E-04

NH

H2CN

HCN exp.

1.E-12

N2O H3N

CHN

N2H2

HONO

NH2 model HCNO model HOCN model NO2 model

NH3 model HCNmodel

H2CN

HNCO

0.1

1.E-03

NO2

NH2

NO3

1.E-10

0.01 NO

N2H2

HNO

1.E-08

0.001

Volume fraction (-)

0 1.E-03

NHmodel

H2CNmodel

1.E-11 1.E-12

Residence time (s)

Residence time (s)

a

b

Figure 6.10 Wood gasification – fuel bound nitrogen compound behaviour predicted by the model for the entire reactor (a) and the lower bed zone (b) and comparison with exp. 991206 (Labee wood gasification); Tbed=Tfreeboard=1097 K, p=0.51 MPa, λ=0.48, other exp. conditions see table 4.7 (chapter 4).

194

195

Experiment Model Experiment Model Experiment Model Experiment Model Experiment Model Experiment Model

991103

991208

991206

991202

991118

991115

Experiment/ simulation

Set point

WP

WP

WP

BC

BC

BC

Fuel type

[CO] (vol.%, wet) 14.0 23.54 8.8 10.09 5.7 7.71 7.8 12.13 5.3 12.25 7.2 9.90

[H2] (vol.%, wet) 10.4 10.39 7.0 4.23 4.5 3.50 5.2 6.55 3.8 4.67 5.3 4.47 [CH4] (vol.%, wet) 2.0 2.31 1.4 0.85 0.6 0.04 2.8 1.6 1.9 0.69 1.9 1.10

[C2H4] (vol.%, wet) 0.1 0.03 0.1 0.05 0.1 0 0.5 0.03 0.3 0.04 0.7 0.09

[C2H6] (vol.%, wet) n.m. 0.02 n.m. 0.01 n.m. 0 n.m. 0.01 n.m. 0.008 n.m. 0.008

[CO2] (vol.%, wet) 11.2 5.81 12.1 10.34 12.8 11.22 9.8 5.28 13.6 7.96 13.3 10.08

[H2O] (vol.%, wet) 5.2 1.51 5.3 5.92 6.4 6.59 8.2 7.06 11.3 12.68 10.6 11.41

[N2] (vol.% , wet) 56.4 55.64 64.6 67.81 69.2 70.33 65.4 67.02 63.2 61.18 60.4 62.39

[Ar] (vol.%, wet) 0.5 0.46 0.5 0.56 0.6 0.60 0.3 0.30 0.5 0.51 0.5 0.53

[C2H2] (ppmv , wet) 0 3 0 16 3 0 126 4 78 17 144 85

[NH3] (ppmv, wet) 2290 2870 1339 1440 770 104 176 151 149 142 176 153

[HCN] (ppmv, wet) 0 2 0 9 0 0 0 0.2 0 3 0 1

[HNCO] (ppmv, wet) 0 14 0 8 0 0 0 0.6 0 0.9 0 1

[NO] (ppmv, wet) 0 8 0 22 0 20 0 3 0 12 0 11

[NO2] (ppmv, wet) 0 0 0 0 0 0 0 0 0 0 0 0

[N2O] (ppmv, wet) 0 0.1 0 0.1 0 0 0 0 0 0.01 0 0.03

Table 6.4 Overview of the DWSA gasification results with wood pellets(WP) and brown coal (BC) as fuel; comparison of model and experimental results

Carbon Conv. (%) 94 99.2 98 93.1 99 98.9 100 91.3 100 90.1 100 94

6.3 Discussion of the results The agreement between model predictions and experimental results is in general reasonably good, especially regarding the light hydrocarbon species (CH4, C2H4) and NH3. Thermodynamic equilibrium modelling by means of Gibbs energy minimisation for conditions prevailing in the experiments leads to hydrocarbon and NH3 species concentrations, which are much lower than experimentally observed. This is confirmed by [Padban, 2000] who performed thermodynamic equilibrium calculations for a wide range of pressurised fluidised bed gasification process conditions. The differences between simulations and experiments as observed in the case of H2 concentrations for the biomass fuels used in this study can be attributed to the differences in the actual flash pyrolysis behaviour of the wood and miscanthus fuels that were applied in our research and the H2 pyrolysis yields that were obtained from the literature which were determined for beech wood. For brown coal the difference in hydrogen yield during pyrolysis can be attributed to differences in composition of the brown coal used and the one for which pyrolysis yields of hydrogen were derived. Differences in simulated and observed CO and CO2 concentrations also occur, and mainly for biomass, especially wood. The tar cracking kinetics description in our model has been assumed to be relatively simple: in a one step reaction CO, CH4 and carbon are formed. Reality is much more complex. Tar is a complex mixture of organic, aromatic, hetero-atom structures. It is likely that besides CO as oxygen containing gas product from tar cracking, CO2 will be formed as well. This can be partly the result of decarboxylation reactions. The problem is that data about tar cracking both in terms of explicit reaction mechanisms with product yields as function of e.g. temperature and kinetic rate data for these reactions are lacking. Both reactivity and product yield of tar cracking should be studied in detail for tars released under conditions characteristic for the fluidised bed gasification process. The carbon conversions predicted by the model for most cases do not deviate much from the measured values despite of the fact that the reaction kinetic data of the main heterogeneous gasification reactions for the biomass and brown coal fuel under consideration have been taken from the literature. The pyrolysis product yields of HCN and NH3 were obtained from model simulations with the biomass version of the FG-DVC model. No tar-N nor char-N formation is assumed and the rest of the nitrogen is instantaneously converted into HCN. This is based on the assumption that char bound nitrogen is to a major extent converted into HCN. The N2 release during the TG-FTIR measurements could not be determined directly, which is inherent to the analysis method, and could not be determined following from an element balance as char amounts were too low to determine the N contents. This, however, is expected to result in only a small contribution to the observed deviations between the simulation results and the experiments. Further differences between model and experimental results for the minor nitrogen species can be attributed to the neglection of sulphur and chlorine species, which compete for radicals with the other and thus also the nitrogen containing radicals. For wood this does not seem to play an important role, as S and Cl contents are very low. For miscanthus and brown coal the S and Cl contents are higher, leading to potential gas species concentrations (H2S+COS) and HCl of above 100 ppmv. This should be further studied. Char nitrogen reactions have not been specified. This simplification is justified by the fact that relatively reactive fuels have been used in this research work, for which the gas phase is much more important than the solid phase for the nitrogen chemistry. Differences in nitrogen species measured and simulated can be partly attributed to this simplification in the model. Heterogeneous catalytic effects of ash elements on the different reactions that play a major role in the conversion of the fuel particles and gas species have not been taken into account in the present model.

196

The predicted NO and HNCO concentrations, being principally in the range where they could be detected by FTIR analysis, have not been observed, which is probably due to these catalytic effects. The largest differences in absolute concentration terms are found for miscanthus for both NO and HNCO. Due to the larger ash content and lower carbon conversions of miscanthus as compared to wood it is possible that catalytic effects are more prominent for miscanthus. In several simulation cases, the HCN concentration in the product gas in and downstream of the freeboard is underpredicted. This is probably due to the assumption that HCN is heterogeneously hydrolysed at char surfaces with water, forming NH3 and CO. This reaction is considered to be catalysed on a char surface and fast with respect to external mass transfer of HCN, see [Shimizu et al., 1993]. The reaction is possibly slower at the char surfaces involved, so that the assumption of complete mass transfer limitation is inaccurate. This would lead to higher HCN concentrations at the bed exit and with the homogeneous reactions forming HCN in the freeboard this can lead to more accurately predicted HCN concentrations. 6.4 Conclusions and recommendations A comparison has been made between simulations and experiments for both biomass (wood and miscanthus) and brown coal pressurised fluidised bed gasification. This comparison has been performed for experimental results obtained from a pilot scale (Delft 1.5 MWth PFBG) and lab scale (Stuttgart 50 kWth DWSA) facility. The agreement between nitrogen species prediction and measurements is quite good for the fuels and more in particular for the main bound nitrogen component found in the product gas: ammonia. For HCN the concentrations are often underpredicted, possibly due to the heterogeneous hydrolysis reaction in the model, taking place at the char surface, which can have slower kinetics than assumed in literature. The model predicts the formation of HNCO and NO to super-ppmv concentration levels. These species have not been found experimentally in the product gas of the applied fuels. Probably catalytic hydrolysis can convert HNCO into NH3 and CO; this reaction has not been taken into account in the model. For NO, catalytic reduction by mineral ash constituents probably plays a role at the temperatures prevailing in the gasifiers. It is also possible that neglection of sulphur and chlorine chemistry causes deviations between model and experimental results for the minor species; this is not expected for wood, but more for miscanthus and brown coal. Simplified tar- and char-nitrogen reactions can be significant in gaseous nitrogen species formation prediction. The present model assumes that the nitrogen which is not available as gas species will be released initially in the form of HCN. The agreement between model and experimental results is reasonably good for the main gaseous constituents of the product gas. The differences between the model and experimental results can be attributed to the assumed devolatilisation yields regarding H2 in the biomass pyrolysis step. Due to experimental limitations using TG-FTIR, no kinetic data for the FG-DVC pyrolysis sub-model could be determined for H2 and N2. The uncertainty in accuracy of the tar cracking kinetics, in terms of possible reactions, product yields (the expected product spectrum) and rate also plays a major role in the deviations between the model and experimental data, especially for CO and CO2. The deviations are the highest for wood, and less for miscanthus and - more in particular - brown coal. This is in-line with the hypothesis that inaccuracies in tar cracking kinetics plays a major role, because wood shows the highest initial tar yields in the fast devolatilisation step.

197

The carbon conversions are generally quite well predicted, so that the rates of the heterogeneous combustion and kinetics of the gasification reactions are expected be accurate enough. Both the reactivity and product yield of tar cracking should be studied in detail for tars released under conditions characteristic for the fluidised bed gasification process; this probably leads to better predictions of the model for in particular the CO and CO2 concentrations. It is recommended that the concentrations of the molecular species NH3, HCN, HNCO are measured in the bed zone, where the main releases from the fuel and formation as well as destruction reactions are mostly concentrated. Furthermore, it is suggested that the pyrolysis of biomass is further studied with emphasis on a comparison between nitrogen species release under slow and high heating rate conditions. For this purpose a study with a (pressurised) heated grid reactor and a TGA coupled to FTIR analysis is well suited. As the fate of HCN in the model is so much dependent on the heterogeneous reaction of HCN and water on char surfaces, special attention to this reaction is recommended. A kinetic study using a TGFTIR combination can shed more light on this reaction.

198

Chapter 7 Conclusions and recommendations for further research 7.1

Conclusions

PFB experiments An experimental programme involving pressurised fluidised bed (PFB) gasification of miscanthus, wood and brown coal using a bubbling bed reactor has been carried out at the 50 kWth (max) DWSA unit at IVD (University of Stuttgart) and at the 1.5 MWth (max) PFBG test rig at Delft University of Technology. In contrast to significant radial profiles found in CFB gasifiers, no significant radial concentration profile of the main and minor gaseous products was observed in the PFB freeboard measurements performed at the PFBG unit. In this respect the assumption of a plug flow regime in the bubbling bed gasifier model appears to be justified. The concentrations of the main gasification product gas components are comparable to the limited open literature data from other PFB units operated at similar air stoichiometry values. During the PFBG tests, for acetylene (C2H2) axial gas concentration gradients could be clearly observed, which is related to reactions involving tar and soot precursor formation and destruction. For other species measured, such concentration gradients were not observed experimentally. Under the PFB gasification conditions studied, the main fuel bound nitrogen (FBN) component produced was NH3, whereas HCN was formed to a minor extent of only a few percent of the fuel bound nitrogen content. HNCO was never detected, even by means of a high resolution FTIR spectrophotometer under the pressurised gasification test conditions studied. Conversion of FBN to NH3 and HCN was found to be comparable with other bottom-fed FB gasifiers. On the other hand much lower values have been found for a top-fed pressurised FB. An increased Ca inventory in the gasifier by application of additive supply or from the fuel’s inorganic constituents increases the NH3/HCN ratio significantly. Fuel characterisation experiments Flash pyrolysis experiments with miscanthus were conducted at TU Eindhoven using a heated grid reactor (HGR) equipped with in-situ infrared absorption spectroscopy analysis using a tuneable laser. This research was focused on analysis of the pyrolysis yield of CO, CO2 and NH3 at heating rates between 280 and 320K/s and a final temperature in the range between 1050 and 1400K. The trends of CO and CO2 formation were quite well predicted qualitatively by the FG-DVC biomass pyrolysis model. However, the use of the model to predict the pyrolysis product yield at high heating rates, based on a kinetics determined by using low heating rate TG-FTIR analysis experiments, was found to be precarious. This resulted in a reasonable yield prediction for CO and a considerable underprediction for CO2. The background of this observation is probably the competition between the formation of primary products like primary tar fragments and carboxylic acids on one side and light gases like CO, CO2 an H2O on the other side. This competition is heating rate dependent making yield predictions by means of extrapolation from low to high heating rate pyrolysis precarious. The primary pyrolysis products, containing precursor groups for formation of CO and CO2, are quickly removed from the reaction zone and quenched immediately. Therefore, no time is available for further decomposition of these products into CO and CO2, respectively, resulting in low yields.

199

Carboxylic acids are oxygen rich and contain the precursor carboxyl group for CO2 formation. Primary tars contain ether links, precursors for CO formation during cracking. According to this hypothesis the yields of primary tar and acetic acids must be significant. This is confirmed by the observations that at high heating rates the tar yield increases for biomass pyrolysis, which is also observed in literature data. It is observed that due to the formation of transition metal carbonyls on the HGR wall CO disappears. This phenomenon did not influence the yield results, since just enough time was available to measure the maximum CO yield before the CO disappeared. An attempt to detect NH3 in-situ during flash pyrolysis experiments in the HGR has failed. No absorption peak could be found for the NH3 reference gas in the reactor. The reason could be condensation together with water on relatively cold walls. Also, it can be the case that the frequency range of the laser was shifted, so that in fact another laser would be necessary as the applied laser can only be tuned in a very narrow range. There were, however, no time and money resources left to further study this item. Gasifier modelling The effect of process conditions and gas addition in the bed and freeboard zone of the gasifier has been studied theoretically using a plug flow model study with detailed nitrogen chemistry for woody biomass, assuming fast pyrolysis. This was done in view of the possibility to convert NH3 to N2. The product yields of this fast pyrolysis sub-process were obtained from fluidised bed pyrolysis experiments reported in literature. A parameter analysis has been carried out, including an N-mass balance over the reactor sections (bed and freeboard), obtaining a clear N-partitioning between the fixed nitrogen compounds. The results can be summarised as follows: • • • • •

•

• •

NH3, which represents the major fuel-NOx precursor in biomass gasification, is a very stable compound, which is hardly converted in PFB gasifiers. The conversion of NH3 into N2 can only be slightly favoured by increasing the temperatures up to 1200 K. However, above 1150 K sintering of bed particles might occur when using alkali containing biomass fuels, posing a limit to higher reactor temperatures. The NH3 conversion is only slightly dependent on reactor pressure. A minimum in NH3 conversion was obtained around 2.5 bar. At higher pressures (10 bar) the NH3 conversion slightly increased (just 6 ppm less when varying the pressure from 5.1 to 10.2 bar). Gas residence time in the reactor practically did not affect the fuel-N conversion. Destruction of NH3 is only taking place in the presence of O, H and OH radicals, which react very fast in the initial part of the bed. More relevant NH3 conversions can be reached when adding NO or NO2 into the bed. The NH3 concentration decreased from 154 to 121 ppm when adding NO or NO2 in 1:1 ratio to NH3 on molar basis. On the other hand, HCN was formed (30 ppm) and unreacted NO was also predicted among the undesired emissions (114 ppm). Addition of O2 (as secondary air or as primary air) favours the NH3 conversion significantly. However, the main nitrogen species was NO rather than N2, or in the most favourable case (λ → 1, thus stoichiometric combustion) FBN is converted for 50% into NO and 50% into N2. The addition of secondary air lowers the already low energy content of the produced LCV-gas. The addition of CH4 in the bed part of the gasifier reduces the NH3 conversion, probably due to the competition for radicals between CH4, its intermediates (mainly CH3 radicals) and NH3. H2O2 and H2O (steam) addition into the bed did not affect NH3 conversion.

For N species, the agreement between the model prediction and measurements is quite good for all fuels; it is even better for the main fuel bound nitrogen component found in the product gas: NH3. For HCN often the concentrations are underpredicted, probably due to the heterogeneous hydrolysis

200

reaction, taking place at the char surface, which can have slower kinetics than assumed in literature. The model predicts the formation of HNCO and NO to super-ppmv concentration levels. These species have not been found experimentally in the product gas of the applied fuels. Probably catalytic hydrolysis converts HNCO into NH3 and CO. This reaction has not been taken into account in the model. Mineral ash constituents could play a role in catalytic NO reduction at the temperatures prevailing in the gasifiers. It is also possible that neglection of sulphur and chlorine chemistry causes deviations between model and experimental results for the minor species. Also, simplification of tar- and char-nitrogen reactions can be significant to some extent. The model assumes that the nitrogen, which is not available as gas species will be released initially in the form of HCN, which is a simplification. For the main gaseous products, the agreement between model and experimental results is reasonably good. The differences between the model and simulation results can be attributed to unknown pyrolysis yields and kinetic data regarding H2 in the FG-DVC biomass pyrolysis sub-model. Also, the uncertainty in accuracy of the tar cracking kinetics, in terms of possible reactions, product yields and rate plays a major role in the deviations between the model and experimental data. The deviations are the highest for wood, and less for miscanthus and brown coal. This is in-line with the idea that inaccuracies in tar cracking kinetics plays a major role here, as wood shows the highest initial tar yields in the fast pyrolysis step. The carbon conversions are generally quite well predicted, so that the rates of the heterogeneous combustion and kinetics of the gasification reactions are expected be accurate enough. 7.2

Recommendations

PFB experiments Further study of the reasons for the influence of the feeding location on the fuel bound nitrogen speciation behaviour is needed. This can be accomplished by performing fast pyrolysis experiments in different oxidizing media (variable O2 and H2O contents) and also by studying the nitrogen partitioning behaviour for particles with different moisture content. The observation of a clear gradient in the axial acetylene concentration profile and the influence of steam addition should be studied in more detail in relation to the fate of tars and soot, as it is related to proper functioning of high temperature dry gas filtration (prevention of blockage by fine carbonaceous material). It is recommended that the concentrations of the molecular species NH3, HCN, HNCO are measured in the bed zone, where the main releases from the fuel and formation as well as destruction reactions are taking place. Experimental fuel characterisation More experimental research must be performed on the pyrolysis of different biomass species at high heating rates to validate the FG-DVC pyrolysis sub-model, preferably starting at lower temperatures and at elevated pressures. This should include a selected range of major (CO, CO2, H2O) and minor nitrogen species (NH3, HCN, HNCO, NO). The feasibility to detect selected carboxylic acids during pyrolysis, by means of tuneable diode laser infrared absorption spectroscopy should be studied as well. Detection of large carboxylic acid concentrations released during flash pyrolysis experiments in the HGR used in this thesis could prove the hypothesis explaining the low CO2 yield.

201

Combining FTIR analysis or multi-laser IR absorption spectrometry with an HGR and thermogravimetric equipment suitable for high heating rates could provide the necessary experimental research equipment for this research. The hypothesis that CO disappears due to formation of transition metal carbonyls at the HGR reactor wall during flash pyrolysis should be investigated. A coating on the stainless steel reactor shell inside the grid reactor can eliminate formation of such metal carbonyls. The in-situ detection of NH3, HCN or HNCO in biomass pyrolysis experiments using the HGR should be pursued to determine the product end yields. Therefore, an investigation of the wavelengths of the absorption peaks to be used is needed. For this purpose, the minimum NH3 concentration in the reactor necessary to get absorption peaks well above the detection level should be determined. Furthermore, it is suggested that the nitrogen species release during the pyrolysis of biomass and nitrogen containing model compounds is further studied under slow and high heating rate conditions. For this purpose a study with a (pressurised) heated grid reactor and a TGA coupled to FTIR analysis is well suited. As the fate of HCN in the model is so much dependent on the heterogeneous reaction of HCN and water on char surfaces, special attention to this reaction is recommended. A kinetic study using a TGFTIR combination can shed more light on this reaction. It is recommended to study more in depth the influence of moisture content of the fuel on the pyrolysis of fuel bound nitrogen and its relation to the feeding location in an FB reactor. Modelling To improve the FG-DVC biomass pyrolysis sub-model, a mechanism should be developed accounting for the competition between primary pyrolysis products like primary tars and carboxylic acids on one side and light gases on the other side. The new version of the FG-DVC biomass model, to be released in the near future, will probably include such a mechanism. For the gasifier modelling it is recommended to study the influence of radical quenching on solid surfaces in the bed section of the fluidised bed. In the future, a model for gasification should be developed that combines knowledge of detailed reaction kinetics concerning N species with computational fluid dynamics. These CFD models are good for improving design and scale-up of fluidised beds. However, they are comparatively complicated and therefore require a large amount of computational power. In view of the quite good results obtained with a simple plug flow model, according to the opinion of the author, however, more can be attained for NOx precursor emission prediction by further developing correct reaction kinetics schemes, including more species interactions, than by a more detailed fluid dynamics description.

202

Bibliography Aarna, I. and Suuberg, E.M. (1999) The Role of Carbon Monoxide in the NO-Carbon Reaction, Energy & Fuels, 13, pp. 1145-1153. Aarna, I. and Suuberg, E.M. (1997) A review of the kinetics of the nitric oxide-carbon reaction, Fuel, 76(6), pp. 475-491. Abatzoglou, N., Gagnon, M., Valsecchi, B., Toupin, R., Cabana, H., Sarbaek, L. and Chornet, E. (2003) The BiosyngasEstrie Pilot Project for the Gasification of Sorted Municipal Solid Waste. In: Bridgwater, A.V., Pyrolysis and Gasification of Biomass and Waste, proceedings of the expert meeting on September 30-October 1, Strasbourg, France, pp. 325-336. Abeliotis, K., Gyftopoulou, M.-E., Aleman-Mendez, Y.S. and Kyritsis, S. Gasification of Energy Crops in an Air-Blown Fluidized Bed, Proceedings of the 1st World Conference on Biomass for Energy and Industry, pp.1662-1664, June 5-9 Sevilla, Spain. Abul-Milh, M. and Steenari, B.-M. (2001) The Effect of Calcination on the Reactions of Ammonia over Different Carbonates and Limestones in Fluidized Bed Combustion Conditions, Energy & Fuels, 15, pp. 874-880. Acke, F. and Lindqvist, O. (1997) Reactions during the calcinations of a limestone under different atmospheres at fluidised bed combustion conditions: a fixed bed reactor study. In: Proceedings of the 14th International Conference on Fluidized Bed Combustion, ASME, 11-14 may 1997, Vancouver, Canada, pp. 405-413. Adlhoch, W., Keller, J., Herbert, P.K. (1992a) Das Rheinbraun-HTW-Kohlevergasungsverfahren. VGB-Kraftswerktechnik, 72(Heft 4), pp. 345-348. Adlhoch, W., Meyer, B., Schumacher, H.-J. (1992b) Stand der HTW-Vergasungstechnologie, Chem.Ing.Tech., 64(5), pp. 476. Aho, M.J. (1998) Effects of fuel properties, temperature, and pressure on fuel reactivity, formation and destruction of nitrogen oxides, and release of alkalis. Technical review 1993-1998 of the Liekki 2 research programme, Finland, pp. 383426. Aho, M.J., Hämäläinen, J.P. and Tummavuori, J.L. (1993a) Importance of Solid Fuel Properties to Nitrogen Oxide Formation Through HCN and NH3 in Small Particle Combustion. Combust. Flame, 95, pp. 22-30. Aho, M.J., Hämäläinen, J.P. and Tummavuori, J.L. (1993b) Conversion of peat and coal nitrogen through HCN and NH3 to nitrogen oxides at 800 °C. Fuel, 72, pp. 837-841. Aho, M.J. and Rantanen, J.T. (1989) Emissions of nitrogen oxides in pulverized peat combustion between 730 and 900 °C, Fuel, 68(5), pp. 586-590. Albrecht, W. (1992) NOx-Emissionen aus Kohlenstaubflammen, VGB Kraftwerkstechnik, 72(7), pp. 614-621. Alzueta, M.U., Tena, A. and Bilbao, R. (2002) Pyridine conversion in a flow reactor and its interaction with nitric oxide, Combust.Sci.Techn., 174(10), pp. 151-169. Åmand, L.E. and Leckner, B. (1991) Oxidation of Volatile Nitrogen Compounds during Combustion in Circulating Fluidized Bed Boilers, Energy & Fuels, 5, pp. 809 - 815. Amundson, N.R. and Arri, L.E. (1978) Char Gasification in a Countercurrent Reactor, AIChE J., 24(1), pp. 87-101. Anderl, H., Mory, A. and Zotter, T., (1999) BioCoComb – Gasification of Biomass and Co-combustion of the Gas in a Pulverized-Coal-Boiler, Proceedings of the 15th International Conference on Fluidized Bed Combustion, Paper No. FBC990029, May 16-19, Savannah, Georgia, USA. André, R.N., Pinto, F., Franco, C., Rosa, L., Barros, M.J., Gulyurtlu, I. and Cabrita, I. (2000) Study of gasification technology to convert biomass and plastic waste into an economical valuable gas. In: Kyritsis, S., Beenackers, A.A.C.M., Helm, P., Grassi, A. and Chiaramonti, D. (eds.) Proceedings of the 1st World Conference on Biomass for Energy and Industry, June 5-9 Sevilla, Spain, pp. 1709-1712. Andries J., Ünal Ö., de Jong W., and Hoppesteyn P.D.J. (2001) Operating experience with a ceramic channel-flow filter downstream of a coal-biomass fuelled pressurized bubbling fluidized bed gasifier. In: D.W. Geiling (eds.), Proceedings of the 16th international conference on fluidized bed combustion (CD-ROM), FBC01 (Reno), May 13-16, American Society of Mechanical Engineers, New York, pp. 1-11, ISBN: 0-7918-3533-5. Andries J, Ünal, Ö., De Jong, W. and Hoppesteijn, P. D. J. (2000) Hot gas clean up downstream of a coal/biomass fuelled pressurised fluidised bed gasifier using ceramic channel flow filter. The Journal of the Filtration Society, 1 (1), p. 10-11. Antal, M.J. and Varhegyi, G. (1995) Cellulose Pyrolysis Kinetics: The Current State of Knowledge, Ind.Eng.Chem.Res., 34, pp. 703-717.

203

Antal, M.J. (1985) Biomass Pyrolysis: a review of the literature, Part 2, Lignocellulose pyrolysis. In: Boer K.W., Duffie J.A. (eds.), Advances in Solar Energy, American Solar Energy Society. Anttikoski, T. (2002) Circulating Fluidized Bed Gasifier offers possibility for biomass and waste utilization and for substitution of natural gas by syngas from coal gasification. In: Proceedings of the IChemE conference “Gasification V: the clean choice for carbon mangament”, 8-10 April, Noordwijk, The Netherlands. Arri, L.E. and Amundson, N.R. (1978) An Analytical Study of Single Particle Char Gasification, AIChE J., 24(1), pp. 72-87. Arrieta, F.R.P. and Sanchez, C.G. (1999) BIG-GT and CEST Technologies for Sugar Cane Mill. Thermodynamic and Economic Assessments, Proceedings of the second Olle Lindström Symposium, pp 184-189, June 9-11, Stockholm, Sweden. Arthur, J.R. (1951) Reactions between Carbon and Oxygen, Trans.Farad.Soc.,47, pp. 164-178. Arvelakis, S., Gehermann, H., Beckmann, M. and Koukios, E.G. (2002) Effect of leaching on the ash behavior of olive residue during fluidized bed gasification, Biomass & Bioenergy, 22, pp. 55-69. Asikainen, A.H., Kuusisto, M.P., Hiltunen, M.A. and Ruuskanen, J. (2002) Occurrence and Destruction of PAHs, PCBs, CIPhs, CIBzs, and PCDD/Fs in Ash from Gasification of Straw, Environ. Sci. Technol., 36, pp. 2193-2197. Augustin, T. (1986) Wertstoffgewinnung durch Pyrolyse von thermisch konditioniertem Belebtschlamm in der Wirbelschicht, Ph.D. thesis, University of Hamburg, Germany. Avni, E., Davoudzadeh, F. and Coughlin, R.W. (1985) Flash Pyrolysis of Lignin. In: Overend, R.P., Milne, T.A. and Mudge, L.K. (eds.) Fundamentals of Thermochemical Biomass Conversion, Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 329-343. ISBN 0-85334-306-3 Axworthy, A.E., Dayan, V.H. and Martin, G.B. (1978) Reactions of fuel-nitrogen compounds under conditions of inert pyrolysis, Fuel, 57, pp. 29-35. Azuhata S., Akimoto H. and Hishinuma Y. (1982) Effect of H2O2 on Homogeneous Gas Phase NO Reduction Reaction with NH3. AIChE Journal, 28, (1), pp 7-11. Bacon, D.W., Downie, J., Hsu, J.C. and Peters, J. (1985) Modelling of Fluidized Bed Gasifiers. In: Overend, R.P., Milne, T.A., Mudge, L.K. (eds.) Fundamentals of Thermochemical Biomass Conversion, Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 717-732. ISBN 0-85334-306-3 Badr, O. and Probert, S.D. (1992) Nitrous Oxide in the Earth’s Atmosphere, Appl.Energy, 41, pp. 177-200. Badr, O. and Probert, S.D. (1993a) Environmental Impacts of Atmospheric Nitrous Oxide, Appl.Energy, 44, pp. 197-231. Badr, O. and Probert, S.D. (1993b) Oxides of Nitrogen in the Earth’s Atmosphere: Trends, Sources, Sinks and Environmental Impacts, Appl.Energy, 46, pp. 1-67. Banna, S.M. and Branch, M.C. (1981) Mixing and Reaction of NH3 with NO in Combustion Products, Combustion&Flame, 42, pp. 173-181. Barrio, M., Gøbel, B., Risnes, H., Henriksen, U., Hustad, J.E. and Sørensen, L.H. (2001) In: Bridgwater, A.V. (ed.) Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 32-46. Barrio, M. and Hustad, J.E. (2001) In: Bridgwater, A.V. (ed.) Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 47-60. Barducci, G., Ulivieri, P., Pike, D.C., McDonald, N., Repetto, F. and Cristo, F. (1999) The Greve in Chianti Project, Renewable Energy, 16, pp. 1041-1044. Bartle, K.D., Perry, D.L. and Wallace, S. (1987) The Functionality of Nitrogen in Coal and Derived Liquids: An XPS Study, Fuel Proc. Tech., 15, pp. 351-361. Bassilakis, R., Carangelo, R.M. and Wójtowicz M.A. (2001) TG-FTIR analysis of biomass pyrolysis, Fuel, 80, pp. 17651786. Bassilakis, R., Zhao, Y., Solomon, P.R. and Serio, M.A. (1993) Sulfur and Nitrogen Evolution in the Argonne Coals: Experiment and Modeling, Energy & Fuels, 7, pp. 710 – 720. Bauen, A. and Kaltschmitt, M. (2001) Reduction of energy related CO2 emissions - The potential Contribution of Biomass. In: Proceedings of the 1st World Conference on Biomass for Energy and Industry, pp.1354-1357, June 5-9 Sevilla, Spain. Baumann, H. and Möller, P. (1991) Pyrolysis of hard coals under fluidised bed combustor conditions – Distribution of nitrogen compounds on volatiles and residual char, Erdöl und Kohle – Erdgas – Petrochemie vereinigt mit BrennstoffChemie, 44 (1), pp. 29 - 33. Baumann, H., Schuler, J.,and Schiller, R. (1989) Drahtnetzversuche zum Verbleib des Kohlestickstoffs bei Pyrolyse und Verbrennung, Erdöl,Erdgas und Kohle, 105(5), pp. 227-232.

204

Baxter, L.L., Mitchel, R.E., Fletcher, T.H. and Hurt, R.H. (1996) Nitrogen release during Coal Combustion, Energy&Fuels, 10, pp. 188-196. Bend, S.L. (1992) The origin, formation and petrographic composition of coal, Fuel, 71(8), pp. 851-870. Berg, M., Vriesman, P., Heginuz, E., Sjöström, K. and Espenäs, B.-G. (2001) Fuel-bound Nitrogen Conversion: Results from Gasification of Biomass in two Different Small Scale Fluidized Beds. In: Bridgwater, A.V. (ed.) Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 322-332. Berg, M., Espenäs, B.-G. and Strömberg, B. (1999a) Minskning av emissionen av kväveföreningar från biobränslebaserade IGCC-processer Etapp 4, TPS report TPS-99/11, TPS Termiska Processer AB, Studsvik, Sweden. Berg, M., Espenäs, B.-G. and Strömberg, B. (1999b) Minskning av emissionen av kväveföreningar från biobränslebaserade IGCC-processer Etapp 5, TPS report TPS-99/44, TPS Termiska Processer AB, Studsvik, Sweden. Berger, A. and Rotzoll, G. (1995) Kinetics of NO Reduction by CO on quartz glass surfaces, Fuel, 74(3), pp. 452-455. Bettagli, N., Desideri, U. and Fiaschi, D. (1995) A Biomass Combustion-Gasification Model: Validation and Sensitivity Analysis, Trans.ASME J.Energy Res.Technol., 117, pp.329-336. Beuken, R. (2001) Grid reactor experiments on pyrolysis & gasification of biomass, MSc. Thesis, report nr. R-1528-A, Eindhoven University of Technology, Department of Applied Physics, Fluid Dynamics Laboratory, section Gas Dynamics, The Netherlands. Bi, J., Luo, C., Aoki, K., Uemiya, S. and Kojima, T. (1997) A numerical simulation of a jetting fluidised bed coal gasifier, Fuel, 76 (4), pp. 285-301. Biba, V., Klose, E., Malecha, J. and Macák, J. (1976a) Mathematisches Model zur Kohlevergasung unter Druck – Teil I, Energietechnik, 26 (1), pp. 28-32. (in German) Biba, V., Klose, E., Malecha, J. and Macák, J. (1976b) Mathematisches Model zur Kohlevergasung unter Druck – Teil II, Energietechnik, 26 (1), pp. 71-75. (in German) Biba, V., Macák, J., Klose, E. and Malecha, J. (1978) Mathematical model for the gasification of coal under pressure, Ind. Eng. Process Des. Dev., 17, pp. 92-98. Bilodeau, J.-F., Thérien, N., Proulx, P., Czernik, S. and Chornet, E. (1993) A Mathematical Model of Fluidized Bed Biomass Gasification, Can. J. Chem. Eng., 71, pp. 549-557. Bingyan, X., Zengfan, L, Chungzhi, W., Haitao, H. and Xiguang, Z. (1994) Circulating fluidised bed gasifier for biomass. In: Nan, L., Best, G. and Coelho de Carvalho Neto, C. (eds.) Integrated energy systems in China - The cold Northeastern region experience, report of the FAO (UN), Rome, paper no. 9407. Björkman, E., Nilsson, T. and Strömberg, B. (1997) The formation of nitrogen compounds at gasification/pyrolysis condition, TPS-report 97/35, Termiska Processer AB, Studsvik, Sweden. Björkman, E. and Larsson, C. (1996) The formation of nitrogen compounds at gasification conditions, with and without internal catalysts, TPS-report 96/48, Termiska Processer AB, Studsvik, Sweden. Björkman, E. and Sjöström, K. (1991) Decomposition of Ammonia over Dolomite and Related Compounds, Energy&Fuels, 5, pp. 753-760. Blackadder, W.H., Gustavsoon, P., Svensson, O. and Waldheim, L. (1995) BIG_GT project, Summary report 1993-1994, confidential internal report number TPS-95/27. Blair, D.W. Wendt, J.O.L. and Bartok, W. (1976) Evolution of nitrogen and other species during controlled pyrolysis of coal. In: Proceedings of the 16th Symposium (International) on Combustion/The combustion institute, pp. 475-489. Bliek, A., Lont, J.C. and van Swaaij, W.P.M. (1986) Gasification of coal-derived chars in synthesis gas mixtures under intraparticle mass transfer-controlled conditions, Chem.Eng.Sci., 41(7), pp. 1895-1909. Bos, D. (1998) Het hydrodynamisch gedrag van een wervelbed, , report EV 1978 Delft University of Technology, section Thermal Power Engineering (in Dutch). Bose, A.C., Dannecker, K.M. and Wendt, J.O.L. (1988) Coal Composition Effects on Mechanisms Governing the Destruction of NO and Other Nitrogenous Species during Fuel-Rich Combustion, Energy&Fuels, 2, pp. 301-308. Bosma, D. (1997) Calibratie van een FTIR spectrofotometer voor off-line gas analyse, Delft University of Technology, report number EV-1968 (in Dutch). Bowman, C.T. (1992) Control of combustion-generated nitrogen oxide emissions: technology driven by regulation, Proceedings of the 24th Symposium (International) on Combustion/The combustion institute, pp. 859-878. BP Amoco (2001) Statistical review of world energy – Primary Energy http://www.bp.com/worldenergy/primary/index.htm

205

Brage, C., Yu, Q., Chen, G. and Sjöström, K. (2000) Tar evolution profiles obtained from gasification of biomass and coal, Biomass & Bioenergy, 18 (1), pp. 87-91. Brage C., Yu Q., Chen, G. & Sjöström K. (1997) Use of amino phase adsorbent for biomass tar sampling and separation. Fuel, 76 (2), 137-42. Braun, R.L. and Burnham, A.K. (1987) Analysis of chemical reaction kinetics using a distribution of activation energies and simpler models. Energy & Fuels, 1, pp. 153-161. Brem, G. (1990) Mathematical Modelling of Coal Conversion Processes. PhD thesis. Twente University, Enschede, The Netherlands. Bridgwater, A.V. and Cottam, M.-L. (1992) Opportunities for Biomass Pyrolysis Liquids Production and Upgrading, Energy & Fuels, 6 (2), pp. 113-120. Bruinsma B. (2001) Pulverised Coal Gasification Under Simulated Blast Furnace Conditions. Section Thermal Engineering, Department of Mechanical Engineering, Twente University of Technology, Master Thesis. Bruinsma, O.S.L., Tromp, P.J.J., de Sauvage Nolting, H.J.J. and Moulijn, J.A. (1988a) Gas phase pyrolysis of coal-related aromatic compounds in a coiled tube flow reactor – 2. Heterocyclic compounds, their benzo and dibenzo derivatives, Fuel, 67, pp. 334-340. Bruinsma, O.S.L., Geertsma, R.S., Oudhuis, A.B.J., Kapteijn, F. And Moulijn, J.A. (1988b) Measurement of C, H, N-release from Coals during Pyrolysis, Fuel, 67, pp. 1190-1196. Buckley, A.N. (1994) Nitrogen functionality in coals and coal-tar pitch determined by X-ray photoelectron spectroscopy, Fuel Proc. Tech., 38, pp. 165-179. Buekens, A.G. and Schoeters, J.G. (1984) Mathematical modelling in gasification (chapter 11). In: Bridgwater, A.V. (ed.), Thermochemical Processing of Biomass. Butterworths, London, UK, pp. 177-199. Buffinga, G.J. (2002) Development and demonstration of a gasification process and integrated novel gas cleaning technology for the conversion of chicken manure into energy. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. And Grassi, A. (eds.) Proceedings of the 12th European Biomass Conference “Biomass for Energy, Industry and Climate Protection”, 17-21 June 20002, Amsterdam, The Netherlands, pp. 449-451. Burchill, P. and Welch, L.S. (1989) Variation of nitrogen content and functionality with rank for some UK Bituminous Coals, Fuel, 68, pp. 100-104. Burchill, P. (1987) In: International conference on Coal Science’, ed. J.A. Moulijn, K.A. Nater and H.A.G. Chermin, Elsevier, Amsterdam, pp. 5-8. Caballero, M., Corella, J., Aznar, M.-P. and Gil, J. (2000) Biomass Gasification with Air in Fluidized Bed. Hot Gas Cleanup with Selected Commercial and Full-Size Nickel-Based Catalysts, Ind. Eng. Chem. Res., 39, pp.1143-1154. Cai, H.-Y., Güell, A.J., Chatzakis, I.N., Lim, J.-Y., Dugwell, D.R. and Kandiyoti, R., (1996) Combustion reactivity and morphological change in coal chars: Effect of pyrolysis temperature, heating rate and pressure, Fuel, 75 (1), pp. 15-24. Cai, H.-Y., Güell, A.J., Dugwell, D.R. and Kandiyoti, R. (1993) Heteroatom distribution in pyrolysis products as a function of heating rate and pressure, Fuel, 72, pp. 321 – 327. Campbell, P.E., McMullan, J.T. and Williams, B.C. (2000) Concept for a competitive coal fired integrated gasification combined cycle power plant, Fuel, 79 (9), pp. 1031-1040. Caram, H.S. and Amundson, N.R. (1979) Fluidized Bed Gasification Reactor Modelling. 1. Model Description and Numerical Results for a Single Bed, Ind.Eng.Chem.Process Des.Dev., 18(1), pp. 80-96. Caram, H.S. and Amundson, N.R. (1979) Fluidized Bed Gasification Reactor Modelling. 2. Effect of the Residence Time Distribution and Mixing of the Particles. Staged Beds Modeling, Ind.Eng.Chem.Process Des.Dev., 18(1), pp. 96-102. Chakrabourty, R.K. and Howard, J.R. (1981) Combustion of char in shallow fluidized bed combustors: influence of some design and operation parameters, J.Inst. of Energy pp. 48-54. Chan, L.K., Sarofim, A.F. and Beér, J.M. (1983) Kinetics of the NO-Carbon Reaction at Fluidized Bed Combustor Conditions, Combustion&Flame, 52, pp. 37-45. Chang, L., Xie, Z., Xie, K.-C., Pratt, K.C., Hayashi, J.-I., Chiba, T. and Li, C.-Z. (2003) Formation of NOx precursors during pyrolysis of coal and biomass. Part VI. Effects of gas atmosphere on the formation of NH3 and HCN, Fuel, 82(10), pp. 11591166. Chen, C., Horio, M. and Kojima, T. (2000a) Numerical simulation of entrained flow coal gasifiers. Part I: modelling of coal gasification in an entrained flow gasifier, Chem.Eng Sci.,55, pp. 3861-3874. Chen, C., Horio, M. and Kojima, T. (2000b), Numerical simulation of entrained flow coal gasifiers. Part II: effects of operating conditions on gasifier performance, Chem.Eng Sci.,55, pp. 3875-3883.

206

Chen, C., Miyoshi, T., Kamiya, H., Horio, M. and Kojima, T. (1999) On the Scaling-up of a Two-stage Air Blown Entrained Flow Coal Gasifier, Can.J.Chem.Eng.,77, pp. 745-750. Chen, G., Andries, J. and Spliethoff, H. (2003) Biomass conversion into fuel gas using circulating fluidised bed technology: the concept improvement and modelling discussion, Renewable Energy, 28, pp. 985-994. Chen, G. (1998) The reactivity of char from rapid heating processes under pressure: the role of the time-temperatureenvironment history of its formation, Ph.D. thesis, Royal Institute of Technology Stockholm, Sweden. Chen, Y., Charpenay, S., Jensen, A., Wojtowicz, M.A. and Serio, M.A. (1998) Modeling of biomass pyrolysis kinetics, Proceedings of the 27th Symposium (International) on Combustion/The combustion institute, pp. 1327-1334. Chen, J.C. and Niksa, S. (1992 a) Suppressed Nitrogen Evolution from Coal-Derived Soot and Low-Volality Coal Chars. In: Proceedings of the 24th Symposium (International) on Combustion, The combustion institute, pp. 1269-1276. Chen, J.C. and Niksa, S. (1992 b) Coal Devolatilisation during Rapid Transient Heating, 1. Primary Devolatilisation, Energy & Fuels, 6, pp. 254 - 264. Chen, S.L., Heap, M.P., Pershing, D.W. and Martin, G.B. (1982) Influence of coal composition on the fate of volatile and char nitrogen during combustion. In: Proceedings of the 19th Symposium (International) on Combustion, The combustion institute, pp. 1271-1280. Chung, S.F. and Wen, C.Y. (1968) Longitudinal Dispersion of Liquid Flowing Through Fixed and Fluidized Beds, AIChE J., 14, pp.857-866. Coda Zabetta, E.G. and Kilpinen, P.T. (2001) Gas-Phase conversion of NH3 to N2 in gasification-Part II: Testing the Kinetic Model, IFRF Combustion Journal (on-line available at http://www.ifrf.net), Article number 200104, March 2001. Coda Zabetta, E.G., Kilpinen, P., Hupa, M., Ståhl, K., Leppälahti, J., Cannon, M. and Nieminen, J. (2000a) Kinetic Modeling Study on the Potential of Staged Combustion in Gas Turbines for the Reduction of Nitrogen Oxide Emissions from Biomass IGCC Plants, Energy & Fuels, 14, pp. 751-761. Coda Zabetta, E.G., Kilpinen, P., Hupa, M., Ståhl, K., Leppälahti, J., Cannon, M. and Nieminen, J. (2000b) Kinetic Modeling Study on the Potential of Staged Combustion in Gas Turbines for the Reduction of Nitrogen Oxide Emissions from Biomass IGCC Plants (correction), Energy & Fuels, 14, p. 1335. Corella, J., Aznar, M.-P., Gil, J. and Caballero, M.A. (1999) Biomass Gasification in Fluidized Bed: Where to locate the Dolomite to improve Gasification? Energy & Fuels, 13, 1122-1127. Corella, J., Herguido, J. and Alday, F.J. (1988) Pyrolysis and Steam Gasification of Biomass in Fluidized Bed. Influence of the Type and Location of the Biomass Feeding Point on the Product Distribution. In: Bridwater, A.V. and Kuester, J.L. (Eds.) Research in Thermochemical Biomass Conversion, Elsevier Applied Science: London, pp. 384-397. Corté, P., Hérault, V., Castillo, S. and Traverse, J.P. (1987) High-temperature gasification of carbonaceous materials by flash pyrolysis: thermal aspects, Fuel, 66, pp. 1107-1114. Costello, R. and Chum, H.L. (1998) Proceedings of the BioEnergy ’98: Expanding BioEnergy Partnerships conference, pp. 11-17, October 4-8 Madison, Wisconsin, USA. Cousins, W.J. (1978) A theoretical study of wood gasification processes, N.Z. J. Sci., 21, pp. 175-183. Craig, J.D. (1996) Development of a small scale biomass-fuelled integrated-gasifier gas turbine power plant: phase 1. Report published in partial fulfilment of grant DE-FG65-91WA08318 of the Western Regional Biomass Energy Program, Cratech Inc., Tahoma, Texas, United States of America. Dagaut, P., Luche, J. and Cathonnet, M. (2000) Experimental and Kinetic Modelling of the Reduction of NO by Propene at 1 Atm, Combustion & Flame, 121, pp. 651-661. Dagaut, P. and Cathonnet, M. (1998) A Comparative Study of the Kinetics of Benzene Formation from Unsaturated C2 to C4 Hydrocarbons, Combust.Flame, 113, pp. 620-623. De Bari, I, Barisano, D., Cardinale, M., Matera, D., Nanna, F. and Viggiano, D. (2000) Air Gasification of Biomass in a Downdraft Fixed Bed: A Comparative Study of the Inorganic and Organic Products Distribution, Energy&Fuels, 14, pp. 889898. de Jong , W., Pirone, A., Wójtowicz, M.A. (2003) Pyrolysis of Miscanthus Giganteus and wood pellets: TG-FTIR analysis and reaction kinetics, Fuel, 82(9), pp. 1139-1147. de Jong, W., van der Wel, M., Ünal, Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2002) Measurements and modelling of biomass and brown coal gasification in a pressurised fluidised bed gasifier with hot gas filtration using ceramic channel filters. In: Proceedings of the 12th European Conference and Technology Exhibition on Biomass for Energy, Industry and Climate Protection, 17-21 June, Amsterdam, The Netherlands.

207

de Jong, W., Andries, J. and Hein, K.R.G. (1999) Coal/Biomass co-gasification in a pressurised fluidised bed bed reactor, Renewable Energy 16, 1110-1113. De Lange, H.J. and Barbucci, P. (1998) The Thermie Energy Farm Project, Biomass and Bioenergy, 15 (3), pp. 219-224. Delgado, J., Aznar, M. and Corella, J. (1997) Biomass Gasification with Steam in Fluidized Bed: Effectiveness of CaO, MgO, and CaO-MgO for Hot Raw Gas Cleaning, Ind.Eng.Chem.Res., 36, pp. 1535-1543. Demirbas, A. (2000) Mechanisms of liquefaction and pyrolysis reactions of biomass, Energy conversion & Management, 41, pp. 633-646. Denn, M.M., Yu, W.-C. and Wei, J. (1979) Parameter Sensitivity and Kinetics-Free Modeling of Moving Bed Coal Gasifiers, Ind. Eng. Chem. Fundam., 18(3), pp. 286-288. De Soete, G.G. (1990) Heterogeneous N2O and NO Formation from Bound Nitrogen Atoms during Coal Char Combustion. In: Proceedings of the 23rd Symposium (International) on Combustion, The Combustion Institute, pp.1257-1264. De Sousa, L.C.R. (2001) Gasification of wood, urban waste wood (Altholz) and other wastes in a fluidised bed reactor Ph.D. thesis, Swiss Federal Institute of Technology, Zürich, Switzerland. De Sousa, L.C. and Stucki, S. (1997) Gasification of Urban Waste Wood in a Fluidized Bed Reactor, Proceedings of The Third Biomass Conference of the America’s: making a business from biomass in Energy, Environment, Chemicals, Fibers and Materials, Pergamon Press, August 24-29, Montreal, Canada, pp. 447-452. De Souza-Santos, M.L. Application of Comprehensive Simulation of Fluidized-Bed Reactors to the Pressurized Gasification of Biomass, J. of the Brazilian Soc. of Mech. Sci., 16 (4) , pp. 376-383. De Souza-Santos, M.L. (1989) Comprehensive modelling and simulation of fluidised bed boilers and gasifiers, Fuel, 68, pp. 1507-1521. Desrosiers, R. (1979) Thermodynamics of gas-char reactions. In: A Survey of Biomass Gasification – vol. 2, Principles and Technology, T.B. Reed (ed.), SERI report SERI/TR-33-239, USA, pp. 133-175. Di Felice, R., Coppola, G., Rapagna, S. and Jand, N. (1999) Modelling of Biomass Devolatilization in a Fluidized Bed Reactor, Can.J.Chem.Eng., 77(4), pp. 325-332. US Department of Energy (DOE) (1996) DOE Biomass power program, strategic plan 1996-2015, report number DOE/GO10096-345 by the National Renewable Energy Laboratory, Golden, Colorado, USA. Double, J.M., Smith, E.L. and Bridgwater, A.V. (1989) Computer modelling of fluidised bed gasification. In: Ferrero, G.L., Maniatis, K., Buekens, A. and Bridgwater, A.V. (eds.), Pyrolysis and gasification, Elsevier Applied Science, London, UK, pp. 651-655. Double, J.M. and Bridgwater, A.V. (1985) Sensitivity of theoretical gasifier performance to system parameters. In: Palz, W., Coombs, J. and Hall, D.O. (eds.), Energy from Biomass, 3rd E.C. Conference, Elsevier Applied Science, London, UK, pp. 915-919. Drummond, A.-R. and Drummond, I.W. (1996) Pyrolysis of Sugar Cane Bagasse in a Wire-Mesh Reactor, Ind. Eng.Chem.Res., 35, pp. 1263-1268. Duo, W. (1990) Kinetic Studies of the Reactions Involved in Selective Non-Catalytic Reduction of Nitric Oxide, Ph.D. thesis, Technical University of Denmark, Lyngby, Denmark.. Duurzame energie in uitvoering, (1999) report by the Dutch department of economic affairs, July 1999, The Netherlands. EC (1997) Energy for the Future: renewable sources of energy. White Paper for a Community Strategy and Action Plan, COM(97)599 final, European Commission, Brussels. http://www.ecn.nl/phyllis (2003) El Asri, R., Konnov, A.A. and De Ruyck, (2001) Reactor Network Modelling of a Biomass Dedicated Swirling Combustor and a Fluidized Bed Gasifier, Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September, Seefeld, Austria, pp. 599-613. El Asri, R., and De Ruyck, J. (1999) Fluidized Bed Flash Pyrolysis of Biomass. In Zanzi, R. (Ed.) Proceedings of the 2nd Olle Lindström Symposium, Stockholm, Sweden. Elliot, M.A. (ed.) (1981) Chemistry of Coal Utilization (second suppl.volume), John Wiley & Sons, New York. Energierapport (1999) report by the Dutch department of economic affairs, November 1999, The Netherlands. http://www.energyproducts.com (2004) Ergüdenler, A. and Ghaly, A.E. (1997) Mathematical Modeling of a Fluidized Bed Straw Gasifier: Part I-Model Development, Energy Sources, 19, pp. 1065-1084.

208

Esperanza, E., Aleman, Y., Arauzo, J., Gea, G. (1999) Fluidized bed gasification of sugar cane bagasse. Influence on gas composition, Proceedings of the second Olle Lindström Symposium, pp.128-133 , June 9-11, Stockholm, Sweden. Evans et al., (1985). Operating results from a pressurized fluidized bed biomass gasification process. In: Proceedings in Energy from biomass and wastes X (Ed. D.L. Klass), Institute of Gas Technology, Chicago, USA, pp. 573-594. Faaij, A.P.C. (1997) Energy from biomass and waste, Ph.D. thesis, University of Utrecht, The Netherlands. Fenimore, C.P. (1972) Formation of Nitric Oxide from Fuel Nitrogen in Ethylene Flames, Combustion & Flame, 19, pp. 289296. Fernando, R. (2002) Experience of indirect cofiring of biomass and coal. Report nr. CCC/64 IEA Clean Coal Centre, London. ISBN-92-9029-370-9. Franco, C., Gulyurtlu, I.,Azevedo, P. and Cabrita, I. (1998) Fluidised bed gasification of eucalyptus grown under different fertilisation conditions) , Proceedings of the 10th European Conference and Technology Exhibition “Biomass for Energy and Industry”, 8-11 June, Würzburg, Germany, pp. 1769 - 1772. Freihaut, J.D., Zabielski, M.F. and Seery, D.J. (1982) A Parametric Investigation of Tar Release in Coal Devolatilisation, Proceedings of the 19th Symposium (International) on Combustion/The combustion institute, pp. 1159-1167. Friebel, J. and Köpsel, R.F.W. (1999) The fate of nitrogen during pyrolysis of German low rank coals – a parameter study, Fuel, 78, pp. 923-932. Friedman, H.L. (1964) Kinetics of Thermal Degradation of Char-Forming Plastics from Thermogravimetry. Application to a Phenolic Plastic. J. Polym. Sci. part C, 6, pp. 183-195. Furusawa, T., Koyama, M. and Tsujimura, M. (1985) Nitric oxide reduction by carbon monoxide over calcined limestone enhanced by simultaneous sulphur retention, Fuel, 64(3), pp. 413-415. Furusawa, T., Tsunoda, M. and Kunii, D. (1982) Nitric Oxide Reduction by Hydrogen and Carbon Monoxide over Char Surface, ACS Symp.Ser., 196, pp. 347-. Gangwal, S.K., Gupta, R.P., Portzer, J.W., Turk, B.S., Krishnan, G.N., Hung, S.L. and Ayala, R.E. (1996) Catalytic Ammonia Decomposition for Coal-Derived Fuel Gases, DOE contract number DE-AC21-92MC29011, report number DOE/MC/29011-97/C0731. Garcia, L., Salvador, M.L., Arauzo, J. and Bilbao, R. (2001) CO2 as a gasifying agent for gas production from pine sawdust at low temperatures using a Ni/Al coprecipitated catalyst, Fuel Processing Technology, 69, pp. 157-174. Garcia-Ibañez, P., Cabanillas, A. and Sánchez, J.M. (2003) The First Gasification Tests of Leached Orujillo on a Circulating Fluidised-Bed Gasifier. In: Bridgwater, A.V., Pyrolysis and Gasification of Biomass and Waste, proceedings of the expert meeting on September 30-October 1, Strasbourg, France, pp. 477-485. Garcia-Ibañez, P., Cabanillas, A. and García-Ybarra, P.L. (2001) A Pilot Scale Circulating Fluidized Bed Plant for Orujillo gasification, Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 209-220. Gardiner, B. (1990) Energy Demands, Aladdin Books, 70 Old Compton Street, London, United Kingdom. Gaur S., and Reed, T.B. (1998) Thermal data for natural and synthetic fuels, New York: Marcel Dekker Inc., USA. Gavalas, G.R. (1982) Coal Science and Technology 4 - Coal Pyrolysis, Elsevier Science Ltd., Amsterdam. Geldart, D. (1973) Types of Gas Fluidization, Powder Technology, 7, pp. 285-292. Ghani, U.M. and Wendt, J.O.L. (1990) Early evolution of coal nitrogen in opposed flow combustion configurations. In: Proceedings of the 23rd Symposium (International) on Combustion, The Combustion Institute, pp.1281-1288. Gil, J., Caballero, M.A., Martín, J.A., Aznar, M.-P. and Corella, J. (1999a) Biomass Gasification with Air in a Fluidized Bed: Effect of the In-Bed Use of Dolomite under Different Operation Conditions, Ind. Eng. Chem. Res., 38, pp.4226-4235. Gil, J., Corella, J., Aznar, M.-P. and Caballero, M.A. (1999b) Biomass Gasification in atmospheric and bubbling fluidized bed: Effect of the type of gasifying agent on the product distribution, Biomass & Bioenergy, 17, pp. 389-403. Gil, J., Aznar, M.-P., Caballero, M.A., Francés, E. and Corella, J. (1997) Gasification in Fluidized Bed at Pilot Scale with Steam-Oxygen Mixtures. Product Distribution for Very Different Operating Conditions, Energy & Fuels, 11, 1109-1118. Glarborg, P., Jensen, A.D. and Johnsson, J.E. (2001) Some aspects of fuel nitrogen conversion in solid fuel fired systems. In: Proceedings of the sixth International Conference on Technologies and Combustion for a Clean Environment, Porto, Portugal, 9-12 July 2001, invited lecture no. 3. Glarborg, P., Kubel, D., Kristensen, P.G., Hansen, J. and Dam-Johansen, K. (1995a) Interactions of CO, NOx, and H2O under Post-Flame Conditions, Combust. Sci. and Tech., 110-111, pp. 461-485.

209

Glarborg, P., Dam-Johansen, K. and Miller, J.A. (1995b) The Reaction of Ammonia with Nitrogen Dioxide in a Flow Reactor: Implications for the NH2+NO2 Reaction, Int. J. Chem. Kin.,27, pp. 1207-1220. Glarborg, P., Kristensen, P.G., Jensen, S.H. and Dam-Johansen, K. (1994) A Flow Reactor Study of HNCO Oxidation Chemistry, Combustion&Flame, 98, pp. 241-258. Glarborg, P., Dam-Johansen, K. And Kristensen, P. (1993) Final Report, Gas Research Institute Contract No. 5091-2602126, Nordic Gas Technology Centre Contract No. 89-03-11. Glarborg, P. and Hadvig, S. (1991) Report, Nordic Gas Technology Centre, pp. 83. Glarborg, P., Miller, J.A. and Kee, R.J. (1986) Kinetic Modeling and Sensitivity Analysis of Nitrogen Oxide Formation in Well-Stirred Reactors, Combustion&Flame, 65, pp. 177-202. Gockel, B. (1994) Ein mathematisches Modell zur Berechnung der Kohlenstaubverbrennung und –vergasung unter Berücksichtigung höherer Drücke, PhD thesis, Universität Bochum, Germany. (in German) Goldschmidt, B., Padban, N., Cannon, M., Kelsall, G., Neergaard, M., Ståhl, K. and Odenbrand, I. (2001) Ammonia formation and NOx emissions with various biomass and waste fuels at the Värnamo 18 MWth IGCC plant, Proceedings of the Progress in Thermochemical Biomass Conversion Conference, pp.524-535, September 17-22 Seefeld, Austria. Gómez, E.O., Cortez, L.A.B., Lora, E.S., Sanchez, C.G. and Bauen, A. (1999) Preliminary tests with a sugarcane bagasse fuelled fluidised-bed air gasifier, Energy Conv. & Management, 40, pp. 205-214. Govind, R. and Shah, J. (1984) Modelling and simulation of an entrained flow coal gasifier, AIChE Journal, 30 (1), pp. 7992. Greil, C. and Vierrath, H. (2000) Fuel Gas from Biomass. In: Kyritsis, S., Beenackers, A.A.C.M., Helm, P., Grassi, A. and Chiaramonti, D. (eds.) Proceedings of the 1st World Conference on Biomass for Energy and Industry, June 5-9 Sevilla, Spain, pp.2128-2130. Greil, C., Schelhaas, K.P. and Sturm, P. (2000) Abatement and removal of Noxious Components at Gasification of Biomass and rejects in Circulating Fluidised Bed Reactors. In: Proceedings of the IChemE conference “Gasification 4: The Future”, 11-13 April, Noordwijk, The Netherlands. Greil, C. (1998) Fuel gas from biomass – utilisation concepts, Proceedings of the 10th European Conference and Technology Exhibition “Biomass for Energy and Industry”, 8-11 June, Würzburg, Germany, pp. 1652-1653. Griffiths, P.R. and De Haseth, J.A. (1986) Fourier transform infrared spectrometry, John Wiley & Sons, New York, USA. Groeneveld, M.J. (1980) The co-current moving bed gasifier, PhD thesis, University of Twente, department of chemical engineering, The Netherlands. Gross, R., Leach, M. and Bauen, A. (2003) Progress in renewable energy, Environ.Int., 29, pp. 105-122. Gudenau, H.W., Schauermann, G. and Hahn, W. (1993) Vergasung von Weidelgras in der zirkulierenden Wirbelschicht, Brennstoff-Wärme-Kraft, 45(11), pp. 491-494. Gudenau, H.W. and Hahn, W. (1993) Schwachgaserzeugung – Vergasung von Miscanthus sinensis giganteus in der zirkulierenden Wirbelschicht, Erdöl und Kohle – Erdgas-Petrochemie vereinigt mit Brennstoff-Chemie, 46 (7/8), pp. 281285. Gumz, W. (1952) Vergasung fester Brennstoffe – Stoffbilanz und Gleichgewicht – Eine Darstellung praktischer Berechnungsverfahren, Springer Verlag, Berlin (in German). Guntermann, K. and Franke, F.H. (1985) Mathematisches Modell für einen Festbettreaktor zur Kohlevergasung, Erdöl und Kohle-Erdgas-Petrochemie vereinigt mit Brennstoffchemie, 38, pp. 456-461. Guo, B., Li, D., Cheng, C., Lü, Z.-A. and Shen, Y. (2001) Simulation of biomass gasification with a hybrid neural network, Bioresource Technology, 76, pp. 77-83. Guo, J. (2004) Pyrolysis of Wood Powder and Gasification of Wood-derived Char, Ph.D. thesis, Eindhoven University of Technology, The Netherlands. ISBN 90-386-1935-9. Hajaligol, M R, Howard, J.B. and Peters, W.A. (1993) An Experimental and Modeling Study of Pressure Effects on Tar Release by Rapid Pyrolysis of Cellulose Sheets in a Screen Heater, Combustion & Flame, 95, pp.47-60 Hajaligol, M.R., Howard, J.B., Longwell, J.P. and Peters, W.A. (1982) Product Compositions and Kinetics for Rapid Pyrolysis of Cellulose, Ind. Eng. Chem. Process Des. Dev., 21, pp. 457-65. Haldipur, G.B., Schmidt, D.K. and Smith, K.J. (1989) A 50-month gasifier mechanistic study and downstream unit process development program for the pressurized ash-agglomeration fluidized-bed gasification system. Final Report, Vol. 1, DOE/me/21063-2740. KRW Energy Systems, Madison (PA), USA.

210

Hall, D.O., Rosillo-Calle, F., Williams, R.H. and Woods, J. (1993) Biomass for energy: supply prospects, In: Johansson, T.B., Kelly, H., Reddy, A.K.N, Williams, R.H and Burnham, L. (eds.) Renewable Energy Sources for Fuels and Electricity, Island Press, Washington D.C., USA. Hallgren, A.L., Engvall, K and Skrifars, B-J. (1999) Ash-induced operational difficulties in fluidised bed firing of biofuels and waste. In: Overend, R.P. and Chornet, E. (eds.) Proceedings of the 4th Biomass conference of the Americas, 29 August – 2 September 1999, Oakland, California, USA, pp. 1365-1370. Hämäläinen, J.P. (1996) Conversion of fuel nitrogen through HCN and NH3 to nitrogen oxides at elevated pressure, Fuel, 75, pp. 1377-1386. Hämäläinen, J.P. and Aho, M.J. (1995) Effect of fuel composition on the conversion of volatile solid fuel-N to N2O and NO, Fuel, 74, pp. 1922 – 1924. Hämäläinen, J.P., Aho, M.J and Tummavuori, J.L. (1994) Formation of nitrogen oxides from fuel-N through HCN and NH3: a model-compound study, Fuel, 73 (12), pp. 1894 – 1898. Hamel, S. (2001) Mathematische Modellierung und experimentelle Untersuchung der Vergasung verschiedener fester Brennstoffe in atmosphärischen und druckaufgeladenen stationären Wirbelschichten, PhD thesis, Universität Siegen, Forschr.-Ber. VDI Reihe 6 Nr. 469. Düsseldorf, Germany. ISBN 3-18-346906-5. (in German) Hannes, J. (1996) Mathematical Modelling of Circulating Fluidized Bed Combustion, PhD thesis, Delft University of Technology, faculty of chemical engineering. ISBN 3-88817-002-8. Hansen, L.K., Rathmann, O., Olsen, A. and Poulsen, K. (1997) Steam Gasification of Wheat Straw, Barley Straw, Willow and Giganteus, Report nr. Risø-R-944(EN), Risø National Laboratory, Roskilde, Denmark. Hansen, P.F.B., Dam-Johansen, K., Johnsson, J.E. and Hulgaard, T. (1992) Catalytic Reduction of NO and N2O on Limestone during Sulfur Capture under Fluidized-Bed Combustion Conditions, Chem.Eng.Sci.,47(9-11), pp.2419-2424. Hansson, K.M., Samuelsson, J., Tullin, C. and Åmand, L.E. (2004) Formation of HNCO, HCN, and NH3 from the pyrolysis of bark and nitrogen-containing model compounds, Combustion&Flame, 137, pp. 265-277. Hansson, K.M., Åmand, L.E., Habermann, A. and Winter, F. (2003) Pyrolysis of poly-L-Leucine under combustion-like conditions, Fuel, 82, pp. 653-660. Hartiniati, Samoeradjo, A. and Youvial, M. Performance of a pilot scale fluidised bed gasifier fuelled by rice husks, in Pyrolysis and Gasification, Proceedings of an international conference, pp. 257-263, 23-25 May 1989, Luxemburg, Luxemburg. Hayhurst, A.N. and Lawrence, A.D. (1992) Emissions of Nitrous Oxide from Combustion Sources, Prog.Energy Combust.Sci., 18, pp. 529-552. Hofbauer, H. and Rauch, R. (2001) Stoichiometric Water Consumption of Steam Gasification by the FICFB-Gasification Process, Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 199-208. Haussmann, G.J. and Kruger, C.H. (1990) Evolution and Reaction of Coal Fuel Nitrogen during Rapid Oxidative Pyrolysis and Combustion, Proceedings of the 23rd Symposium (International) on Combustion/The combustion institute, pp. 12651271. Hayhurst , A.N and Lawrence, A.D. (1992) Emissions of Nitrous Oxide from Combustion Sources, Prog. Energy Combust. Sci., 18, pp. 529-552. Henderson, C. (2003) Clean coal technologies, IEA report number CCC/74, IEA Clean Coal Centre, United Kingdom, ISBN 92-9029-389-6. Hill, W.H. (1945) Recovery of ammonia, cyanogen, pyridine and other nitrogenous compounds from industrial gases. In Lowry, H.H. (ed.) Chemistry of Coal Utilisation, vol. II, Wiley, New York, USA. Hiltunen, M., Kilpinen, P., Hupa, M. and Lee, Y.Y. (1991) N2O Emissions from CFB Boilers: Experimental Results and Chemical Interpretation. In: Anthony, E.J. (ed.) ASME 11th International Conference on Fluidized Bed Combustion, pp. 687694. Hirama, T. et al. (1983) A Two-Stage Fluidized-Bed Coal Combustor for Effective Reduction of NOx Emission. Fluidization, Engineering Foundation, New York, p. 467. Henrich, E., Bürkle, S., Meza-Renken, Z.I. and Rumpel, S. (1999) Combustion and gasification kinetics of pyrolysis chars from waste and biomass, J.of Anal. & Appl. Pyrolysis, 49, pp. 221-241. Hobbs, M.L., Radulovic, P.T. and Smoot, L.D. (1992) Modelling Fixed-Bed Coal Gasifiers, AIChE Journal, 38, pp. 681-702.

211

Hofbauer, H., Rauch, R., Bosch, K., Koch, R. and Aichernig, C. (2003) Biomass CHP Plant Güssing – A Success Story. In: Bridgwater, A.V., Pyrolysis and Gasification of Biomass and Waste, proceedings of the expert meeting on September 30October 1, Strasbourg, France, pp. 527-536. Hoppesteyn, P.D.J. (1999) Application of Low Calorific Value Gaseous Fuels in Gas Turbine Combustors, Ph.D. thesis, Delft University of Technology, The Netherlands. ISBN 90-407-1982-9. Houser, T.J., McCarville, M.E. and Biftu, T. (1980) Kinetics of the Thermal Decomposition of Pyridine in a Flow System, Int.J.Chem.Kinet., 12, pp. 555-568. Hovmand, S. and Davidson, J.F. (1968) Chemical conversion in a slugging fluidised bed, Trans.Instn.Chem.Engrs., 46, pp.190-203. Howard, J.R. ed. (1983) Fluidized Beds Combustion and Applications, Applied Science Publishers London and New York. Howard, J.B. (1981) Fundamentals of Coal Pyrolysis and Hydropyrolysis, chapter 12. In: “The Chemistry of Coal Utilization”, Elliott, M.A. (ed.), John Wiley & Sons, New York, USA. IEA (International Energy Agency) report World Economic Outlook, 1998. International Energy Outlook (2000), report DOE/EIA-0484, published by the Energy Information Administration, Office of Integrated Analysis and Forecasting, US Department of Energy, Washington, USA. IIASA (International Institute for Applied Systems Analysis) / WEC (World Energy Council), Global Energy Prospective, 1998. Ising, M., Hölder, D., Backhaus, C. and Althaus, W. (1998) Holzvergasung in der zirkulierenden Wirbelschicht, Brennstoff, Wärme Kraft, 56 (1-2), pp. 59-62. Iso (1995) Natural gas — Calculation of calorific values, density, relative density and Wobbe index from composition, www.iso.org. Jahns P. Verbleib des Schwefels und Stickstoffs bei der Kohlenpyrolyse und dessen Bedeutung fur thermische Kohleprozesse. (1990) PhD Thesis. Energie- und Verfahrenstechnik, Universität-Gesamthochschule Essen, Germany. Janse, A.M.C., de Jonge, H.G., Prins, W. And van Swaaij, W.P.M. (1998) Combustion kinetics of Char Obtained by Flash Pyrolysis of Pyne Wood, Ind.Eng.Chem.Res., 37, pp. 3909-3918. Jensen, L.S. (1999) NOx from cement production – Reduction by primary measures, PhD thesis, Technical University of Denmark, department of Chemical Engineering, Denmark. Jensen, A. (1996) Nitrogen Chemistry in Fluidized Bed Combustion of Coal, PhD thesis, Technical University of Denmark, department of Chemical Engineering, Denmark. Jensen, A., Johnsson, J.E., Andries, J., Laughlin, K., Read, G., Mayer, M., Baumann, H. and Bonn, B. (1995) Formation and Reduction of NOx in pressurized fluidized bed combustion of coal, Fuel, 74(11), pp. 1555-1569. Jiang, H. and Morey, R.V. (1992) A Numerical Model of a Fluidised Bed Biomass Gasifier, Biomass&Bioenergy, 3, pp. 431447. Jiang, H. (1991) A numerical simulation of a fluidised bed gasifier, PhD thesis, University of Minnesota, United States of America. Johnsson, J.E. and Jensen, A. (1994) Formation and Reduction of NOx in PFB Combustion of Coals and Chars. CHEC report No. 9409, Department of Chemical Engineering, Technical University of Denmark, Lyngby, Denmark. Johnsson, J.E. (1994) Formation and reduction of nitrogen oxides in fluidised-bed Combustion, Fuel, 73(9), pp. 1398-1415. Johnsson, J.E. (1991) Nitrous Oxide Formation and Destruction in Fluidized Bed Combustion – A literature Review of Kinetics. In: Proceedings of the 23rd IEA-AFBC Meeting, Florence, Italy. Johnsson, J.E. and Dam-Johansen, K. (1991) Formation and reduction of NOx in a Fluidized Bed Combustor. In: Anthony, E.J. (ed.) Proceedings of the 11th International Conference on Fluidized Bed Combustion, Montreal, Canada, pp. 1389-1396. Jiang, H. (1991) A numerical simulation of a fluidised bed gasifier, PhD thesis, University of Minnesota, United States of America. Jones, J.M., Harding, A.W., Brown, S.D. and Thomas, K.M. (1995) Detection of reactive intermediate nitrogen and sulfur species in the combustion of carbons that are models for coal chars, Carbon, 33, pp. 833-843. Just, T. and Kelm, S. (1986) Mechanismen der NOx-Entstehung und Minderung bei technischer Verbrennung, Industriefeuerung, 38, pp. 96 -. Kambara, S., Takarada, T. Yamamoto, Y. and Kato, K. (1993) Relation between Functional Forms of Coal Nitrogen and Formation of NOx Precursors during Rapid Pyrolysis, Energy & Fuels, 7, pp. 1013 – 1020.

212

Kee, R.J., Rupley, F.M. and Miller, J.A. (1990) The Chemkin Thermodynamic Database. Report SAND 87-8215, Sandia National Laboratories, Livermore, CA, USA. Kelemen, S.R., Gorbaty, M.L., Kwiatek, P.J., Fletcher, T.H., Watt, M., Solum, M.S. and Pugmire, R.J. (1998) Nitrogen Transformations in Coal during Pyrolysis, Energy&Fuels, 12, pp. 159-173. Kelemen, S.R., Gorbaty, M.L., and Kwiatek, P.J. (1994) Quantification of Nitrogen Forms in Argonne Premium Coals, Energy&Fuels, 8, pp. 896-906. Kersten, S.R.A. (2002) Biomass Gasification in Circulating Fluidized Beds, PhD thesis, Twente University, department of chemical engineering, The Netherlands. ISBN 90-365-1832-6. Kidena, K. Hirose, Y., Aibara, T., Murata, S. and Nomura, M. (2000) Analysis of Nitrogen-Containing Species during Pyrolysis of Coal at Two Different Heating Rates, Energy&Fuels, 14, pp. 184-189. Kilpinen, P.T., Kallio, S., Konttinen, J. and Barišić, V. (2002) Char-nitrogen oxidation under fluidised bed combustion conditions: single particle studies, Fuel, 81, pp. 2349-2362. Kilpinen, P.T., Leppälahti, J.K., Coda Zabetta, E.G. and Hupa, M.M. (1999) Gas-Phase Conversion of NH3 to N2 in gasification. Part 1: A Kinetic Modelling Study on the Potential of the Method, IFRF on-line Combustion Journal (on-line available at http://www.ifrf.net), art.nr. 199901, September 1999. Kilpinen, P. and Hupa, M. (1998) Nitrogen chemistry in combustion and gasification – mechanism and modelling, Liekki 2 Technical Review 1993-1998; Åbo Akademi University, Turku, Finland; Vol.1, pp. 327-359, ISBN 952-12-0271-8. Kilpinen, P., Hupa, M., and Aho, M. (1997) Selective Non-Catalytic NOx Reduction at Elevated Pressures: Studies on the Risks for Increased N2O Emissions. In: Becker, K.H. and Wiesen, P. (eds.) Proceedings of the 7th International Workshop on Nitrous Oxide Emissions, Bergische Universität Gesamthochschule Wuppertal, Cologne, Germany, April 21-23, pp. 9-18. Kilpinen, P. (1992) Kinetic modeling of gas-phase Nitrogen reactions in advanced combustion processes, Ph.D. thesis, Report 92-7, Åbo Akademi University, department of chemical engineering, combustion chemistry research group, Finland. Kilpinen, P. and Hupa, M. (1991) Homogeneous N2O Chemistry at Fluidized Bed Combustion Conditions: A kinetic Modeling Study, Combustion&Flame, 85, pp. 94-104. Kilpinen, P., Hupa, M., Leppälahti, J. (1991) Nitrogen chemistry at gasification – a thermodynamic analysis. Report 91-14, Combustion Chemistry Research Group, Åbo Academi University, Finland. Kilzer, F.J. and Broido, A. (1965) Speculation on the Nature of Cellulose Pyrolysis, Pyrodynamics, 2, pp. 151-163. Kim, Y.J., Lee, J.M. and Kim, S.D. (2000) Modeling of coal gasification in an internally circulating fluidised bed reactor with draught tube, Fuel, 79, pp. 69-77. Kim, S., Park, J.K. and Chun, H.-D. (1995) Pyrolysis Kinetics of Scrap Tire Rubbers. I: Using DTG and TGA, J.Environmental.Eng., July, pp. 507 - 514. Kimball-Linne, M.A. and Hanson, R.K. (1986) Combustion-Driven Flow Reactor Studies of Thermal DeNOx Reaction Kinetics, Combustion&Flame, 64, pp. 337-351. Kinoshita, C.M., Wang, Y. and Takahashi, P.K. (1991) Chemical Equilibrium Computations for Gasification of Biomass to Produce Methanol, Energy Sources, 13, pp. 361-368. Klass DL. Biomass for renewable energy, fuels, and chemicals. 1998; Academic Press Limited, London, UK. Klein, J. (1973) Untersuchungen zur Abspaltung von Wasserdampf, CO, CO2, N2 und H2 bei der nicht isothermen Steinkohlenpyrolyse unter interter und oxydierender Atmosphäre, Ph.D. thesis, University of Aachen, Germany. Knight, R.A. (2000), Experience with raw gas analysis from pressurized gasification of biomass, Biomass & Bioenergy, 18, pp. 67-77. Kojima, T., Assavadakorn, P. And Furusawa, T. (1993) Measurement and evaluation of gasification kinetics of sawdust char with steam in an experimental fluidized bed, Fuel Processing Technology, 36, pp.201-207. Kosky, P.G. and Floess, J.K. (1980) Global model of coal gasifiers, Ind. Eng. Chem. Proc. Des. Dev., 19, pp. 586-592. Kovacik, G., Oguztöreli, M., Chambers, A. and Özüm, B. (1990) Equilibrium calculations in coal gasification, Int. J. Hydrogen Energy, 15(2), pp. 125-131. Kramlich, J.C., Cole, J.A., McCarthy, J.M., Lanier, W.S. and McSorley, J.A. (1989) Mechanisms of Nitrous Oxide Formation in Coal Flames, Combustion&Flame, 77, pp. 375-384. Krishnan, G.N., Wood, B.L., Tong, G.T., and McCarty, J.G. (1988) Study of Ammonia Removal in Coal Gasification Process, Report no.DOE/MC/23087-2667, US department of Energy.

213

Kunii, D. and Levenspiel, O. (1991) Fluidization Engineering (second ed.) Butterworth-Heinemann, Boston, USA. ISBN-0409-90233-0. Kunii, D. and Levenspiel, O. (1990) Fluidized Bed Reactor Models. 1. For Bubbling Beds of Fine, Intermediate and Large Particles, Ind. Engng. Chem. Res., 29, pp. 1226-1234. Kurkela E. (1996) Formation and removal of biomass –derived contaminants in fluidized-bed gasification processes. VTT publications report 287, VTT Technical Research Centre of Finland, Finland. Kurkela, E., Laatikainen-Luntama, J., Ståhlberg, P. and Moilanen, A. (1996) Pressurised fluidised-bed gasification experiments with biomass,, peat and coal at VTT in 1991-94, part 3 Gasification of Danish wheat straw and coal, VTT Publications 291, VTT Technical Research Centre of Finland, Finland. Kurkela, E., Laatikainen-Luntama, J. and Ståhlberg, P. (1995) Pressurized fluidized-bed gasification experiments with wood, peat and coal at VTT in 1991-94, part 2 Experiences from peat and coal gasification and hot gas filtration, VTT Publications 249, VTT Technical Research Centre of Finland, Finland. Kurkela, E., Ståhlberg, P., Laatikainen, J. and Simell, P. (1993a) Development of simplified IGCC-processes for biofuels: supporting gasification research at VTT, Bioresource Technol., 46, pp.37-47. Kurkela, E., Ståhlberg, P. and Laatikainen, J. (1993b) Pressurized fluidized-bed gasification experiments with wood, peat and coal at VTT in 1991-92, part 1 Test facilities and gasification experiments with sawdust, VTT Publications 161, VTT Technical Research Centre of Finland, Finland. Kurkela, E. and Ståhlberg, P., (1992) Air gasification of peat, wood and brown coal in a pressurized fluidized-bed reactor.II.Formation of Nitrogen compounds, Fuel Processing Technology, 31, pp. 23-32. Kwant, K.W. & Leenders, C. (1999) Development of green energy market in the Netherlands and the perspectives of biomass. In: Proceedings of the fourth biomass conference of the Americas, R.P. Overend & E. Chornet Ed., pp. 1629, Elsevier Science Ltd., Oxford, United Kingdom. Laser Photonics, Molspec computer program, 1992. Lau, F.S. (1998) The Hawaiian Project, Biomass & Bioenergy,15 (3), pp.233-238. Lau, F.S. and Carty, R.H. (1994) Development of the IGT RENUGAS® process. In: Proceedings of the 29th Intersociety energy conversion engineering conference, Monterey, California, August 7-12, USA. Laurendeau, N.M. (1978) Heterogeneous Kinetics of Coal Char Gasification and Combustion, Prog. Energy Combust. Sci., 4, pp. 221-270. Leckner, B. and Åmand, L.-E. (1992) N2O Emissions from Combustion in Circulating Fluidized Bed. In: Proceedings of the 5th International Workshop on Nitrous Oxide Emissions, pp. 325-339. Ledesman, E.B., Li, C.-Z., Nelson, P.F. and Mackie, J.C. (1998) Release of HCN, NH3 and HNCO from the Thermal GasPhase Cracking of Coal Pyrolysis Tars, Energy&Fuels, 12, pp. 536-541. Lee, H.-T., Wu, K.-T., Juch, C.I., Tsai, M.Y., Huang, C.C., Kylä-Sipilä, M. and Hiltunen, M. (2003) Gasification of Paper Reject in a CFB Pilot Plant. In: Bridgwater, A.V., Pyrolysis and Gasification of Biomass and Waste, proceedings of the expert meeting on September 30-October 1, Strasbourg, France, pp. 585-592. Leighton, P.A. (1961) Photochemistry of Air Pollution (chapter 10), Academic Press Inc., New York, USA. Leppälahti, J. (1998) Behaviour of fuel-bound nitrogen in gasification and in high-temperature NH3 removal processes, Ph.D. thesis, Åbo Akademi University, VTT Publications 369, VTT Technical Research Centre of Finland, Finland. Leppälahti, J., Kilpinen, P. and Hupa, M. (1998) Selective Catalytic Oxidation of NH3 in Gasification Gas. 3. Experiments at Elevated Pressure, Energy&Fuels,12(4), pp. 758-766. Leppälahti, J. and Koljonen, T. (1995) Nitrogen evolution from coal, peat and wood during gasification: Literature review, Fuel Process.Technol., 43, 1 – 45. Leppälahti, J. (1995) Formation of NH3 and HCN in slow-heating-rate inert pyrolysis of peat, coal and bark, Fuel, 74, pp. 1363 – 1368. Leppälahti, J. (1993) Formation and behaviour of fuel nitrogen compounds in gasification. Bioresource Technology, 46 (1-2), pp. 65-70. Leppälahti, J., Simell, P. and Kurkela, E. (1991) Catalytic conversion of nitrogen compounds in gasification gas, Fuel Processing Technology, 29, pp. 43-56. Leppälahti, J. and Kurkela, E. (1991) Behaviour of nitrogen compounds and tars in fluidised air gasification of peat, Fuel, 70, pp. 491-497.

214

Lewandowski, I., Clifton-Brown, J.C., Scurlock, J.M.O. and Huisman, W. (2000) Miscanthus: European experience with a novel energy crop, Biomass&Bioenergy, 19, pp. 209-227. Lewellen, P.C., Peters, W.A. and Howard, J.B. (1976) Cellulose Pyrolysis Kinetics and Char Formation Mechanism, , Proceedings of the 16th Symposium (International) on Combustion/The combustion institute, pp. 1471-1480. Lewis, R.S., Datar, R.P., Tanner, R.S., Cateni, B.G., Bowser, T.J., Bellmer, D.D. and Huhnke, R.L. (2002) Making the connection: conversion of biomass-generated producer gas to ethanol, Proceedings Bioenergy 2002 Conference, September 22-26 Boise, Idaho, USA. Li, C.-Z. and Tan, L.L. (2000a) Tan, L.L. and Li, C.-Z. (2000) Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part III. Further discussion on the formation of HCN and NH3 during pyrolysis, Fuel, 79, pp. 1899-1906. Li, C.-Z. and Tan, L.L. (2000b) Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part I. Effects of reactor configuration on the determined yields of HCN and NH3 during pyrolysis. Fuel, 79, pp. 1883-1889. Li, C.-Z., Buckley, A.N. and Nelson, P.F. (1998) Effects of temperature and molecular mass on the nitrogen functionality of tars produced under high heating rate conditions, Fuel, 77 (3), pp. 157 164. Li, C.-Z., Nelson, P.F., Ledesma, E.B. and Mackie, J.C. (1996) An experimental study of the release of nitrogen from coals pyrolyzed in fluidised-bed reactions. In: Proceedings of the 26th Symposium (International) on Combustion/The Combustion Institute, pp. 3205-3211. Li, X.T., Grace, J.R., Lim, C.J., Watkinson, A.P., Chen, H.P. and Kim, J.R. (2004) Biomass gasification in a circulating fluidised bed, Biomass&Bioenergy, 26, pp. 171-193. Liu, H. and Gibbs, B.M. (2003) Modelling NH3 and HCN emissions from biomass circulating fluidised bed gasifiers, Fuel, 81, to be published. Liu, P., Guo, X., Wu, C., Chen, Y. and Arai, N. (2000) Gasification Characteristics of Biomass Wastes in Fluidized Bed Gasifier, J. of Propulsion & Power, 16 (4), pp. 606-608. Loeffler, G. and Hofbauer, H. (2002) Does CO Burn in a Fluidized Bed ? - A Detailed Chemical Kinetic Modeling Study, Combustion&Flame, 129, pp. 439-452. Lucas, D. and Brown, N.J. (1982) Characterization of the Selective Reduction of NO by NH3, Combustion&Flame, 47, pp. 219-234. Lucas, J.P., Lim, C.J. and Watkinson, A.P. (1998) A nonisothermal model of a spouted bed gasifier, Fuel, 77 (7), pp. 683694. Lyon, R.K. and Benn, D. (1978) Kinetics of the NO-NH3-O2 Reaction. In: Proceedings of the 17th Symposium (International) on Combustion/The Combustion Institute, pp. 601-610. Lyon, R.K. (1975) Method for the Reduction of the Concentration of NO in combustion Effluents Using Ammonia, US Patent 3900554. Lumbreras, M., Riaño, A., Ibarra, M., Millera, A., Alzueta, M.U. and Bilbao, R. (2001) Nitrogen Distribution in Volatiles and Char during Biomass Pyrolysis. In: Proceedings of the sixth International Conference on Technologies and Combustion for a Clean Environment, Porto, Portugal, 9-12 July 2001, paper 18.3 pp. 607-612. Mackie, J.C., Colket, M.B., Nelson, P.F. and Esler, M. (1991) Shock Tube Pyrolysis of Pyrrole and Kinetic Modelling, Int. J.Chem.Kinet., 23, pp. 733-760. Man, C.K., Pendlebury, K.J. and Gibbins, J.R. (1993) Laboratory measurement of N release under combustion conditions and comparison with plant NOx formation, Fuel Proc.Technol., 36, pp. 117-122. Maniatis, K., Guiu, G. and Riesgo, J. (2003) The European Commission perspective in Biomass and Waste Thermochemical Conversion. In: Bridgwater, A.V., Pyrolysis and Gasification of Biomass and Waste, proceedings of the expert meeting on September 30-October 1, Strasbourg, France, pp. 1-18. Maniatis, K. (1986) Fluidized Bed Gasification of Biomass, PhD Thesis, Aston University, United Kingdom. Mann, R.M., Harris, G.E., Menzies, W.R., Simonson, A.V. and Williams, W.A. (1985) Environmental, health and safety data base for the KRW coal gasification process development unit. Report GRI-85/0123, Gas Research Institute, Chicage (IL), USA. Mansaray, K.G., Al-Taweel, A.M., Ghaly, A.E., Hamdullahpur, F. and Ugursal, V.I. (2000a) Mathematical Modeling of a Fluidized Bed Rice Husk Gasifier: Part I-Model Development, Energy Sources, 22, pp. 83-98. Mansaray, K.G., Ghaly, A.E., Al-Taweel, A.M., Ugursal, V.I. and Hamdullahpur, F. (2000b) Mathematical Modeling of a Fluidized Bed Rice Husk Gasifier: Part III-Model Verification, Energy Sources, 22, pp. 281-296. Mansaray, K.G., Ghaly, A.E., Al-Taweel, A.M., Hamdullahpur, F. and Ugursal, V.I. (1999) Air gasification of rice husk in a dual distributor type fluidized bed gasifier, Biomass & Bioenergy, 17 (4), pp. 315-332.

215

Martens, F.J.A. (1984) Freeboard Phenomena in a Fluidized Bed Combustor. Ph.D. thesis, Delft University of Technology, The Netherlands. McDonald, K.M., Hyde, W.D. and Hecker, W.C. (1992) Low temperature char oxidation kinetics: effect of preparation method, Fuel, 71, pp. 116. Mengis, W. (1983) Mathematische Modellierung des stationären und dynamischen Verhaltens des Lurgi-Druckvergasers unter Berücksichtigung der Kohleschwelung, Ph.D. thesis, Universität Dortmund, Germany. (in German) Meyers, R.A. (ed.) (1982) Coal Structure, chapter 5 Functional Groups and Heteroatoms in Coal, Academic Press, New York, pp. 162-171. Miccio, F., Moersch, O., Spliethoff, H. and Hein, K.R.G. (1999a) Generation and conversion of carbonaceous fine particles during bubbling fluidized bed gasification of a biomass fuel, Fuel, 78, pp. 1473-1481. Miccio, F., Moersch, O., Spliethoff, H. and Hein, K.R.G. (1999b) Gasification of Two Biomass Fuels in Bubbling Fluidized Bed, Proceedings of the 15th International Conference on Fluidized Bed Combustion, paper FBC99-0014, May 16-19 Savannah, Georgia, USA. Miller, R.S., Bellan J. (1997) A generalised biomass pyrolysis model based on superimposed cellulose, hemi-cellulose and lignin kinetics. Combust. Sci. and Technol. 126, pp. 97-137. Miller, J.A. and Bowman, C.T. (1989) Mechanism and Modeling of Nitrogen Chemistry in Combustion, Prog. Energy Combust. Sci.,15, pp. 287 - 338. Milne, T.A., Abatzoglou N. and Evans, R.J. (1998) Biomass Gasifier Tars: Their Nature, Formation and Conversion. IEA Biomass Utilization Task XIII, "Thermal Gasification of Biomass", activity report NREL/TP-570-25357. Mitra-Kirtley, S., Mullins, O.C., Van Elp, J. and Cramer, S.P. (1993) Nitrogen chemical structure in petroleum asphaltene and coal by X-ray absorption spectroscopy, Fuel, 72(1), pp. 133-135. Minchener, A.J. and Kelsall, G.J. (1990) The control of NOx emissions from PFBC systems, Journal of the Institute of Energy, pp. 85-91. Moersch, O., Spliethoff, H. and Hein, K.R.G. (2000) Tar quantification with a new online analyzing method, Biomass & Bioenergy, 18 (1), pp. 79-86. Moilanen, A. and Kurkela, E. (1998) Gasification reactivity and ash behaviour of biomass fuels, Liekki 2 Combustion and Gasification Research Programme, Technical Review 1993-1998, Ed. Hupa, M and Matinlinna, J., Åbo University Press, pp. 529-563. Mojtahedi, W., Ylitalo, M., Maunula, T. and Abbasian, J. (1995) Catalytic decomposition of ammonia in fuel gas produced in pilot-scale pressurized fluidised-bed gasifier, Fuel Processing Technology, 45, pp. 221-236. Mojtahedi, W. and Abbasian, J. (1995) Catalytic decomposition of ammonia in a fuel gas at high temperature and pressure, Fuel, 74 (11), pp. 1698-1703. Molina, A., Eddings, E.G., Pershing, D.W. and Sarofim, A.F. (2000) Char nitrogen conversion: implications to emissions from coal-fired utility boilers, Progr.in Energy&Combust.Sci., 26, pp. 507-531. Monchik, L. and Mason, E.A. (1961) Transport Properties of Polar Gases Journal of Chemical Physics, 35(5), pp. 1676-1695. Monson, C.R., Germane, G.J., Blackham, A.U. and Smoot, L.D. (1995) Char Oxidation at Elevated Pressures, Combustion & Flame, 100, pp.669-683. Moors, J.H.J. (1998) Pulverised char combustion and gasification at high temperatures and pressures, Ph.D. thesis, Eindhoven University of Technology, The Netherlands. ISBN 90-386-0797-0 Mori, H., Asami, K. and Ohtsuka, Y. (1996) Role of Iron Catalyst in Fate of Fuel Nitrogen during Coal Pyrolysis, Energy&Fuels, 10, pp. 1022-1027. Moritomi, H., Suzuki, Y., Kido, N. and Ogiso, Y. (1991) NOx Formation Mechanism of Circulating Fluidized Bed Combustion. In: Anthony E.J. (ed.) Proceedings of the 11th International Conference on Fluidized Bed Combustion, Montreal, Canada, pp. 1005-1011. Mörsch, O., Spliethoff H. and Hein K.RG. (2000) Tar quantification with a new online analyzing method. Biomass & Bioenergy 18(1), pp. 79-86. Mörsch, O. (2000) Entwicklung einer online Methode zur Bestimmung des Teergehalts im Gas aus der Vergasung von Biomasse, Ph.D. thesis, Stuttgart University, Fortschritt-Berichte VDI Reihe 8, Nr. 853, VDI Verlag GmbH, Düsseldorf, Germany. Mullins, O.C., Mitra-Kirtley, S., van Elp, J. and Cramer, S.P. (1993) Molecular Structure of Nitrogen in Coal from XANES spectroscopy, Appl.Spectroscopy, 47(8), pp. 1268-1275.

216

Murphy, M.L. (2001) Repowering options: Retrofit of coal-fired power boilers using Fluidized Bed Biomass Gasification. In: Donald W. Geiling (ed.) Proceedings of the 16th International Conference on Fluidized Bed Combustion, paper FBC01-0031, May 13-16 Reno, Nevada, USA. Nagel, H. (2002) Untersuchungen zum Emissionsverhalten und Wirkungsgradpotential von Druckwirbelschichtfeuerungen der ersten und zweiten Generation, Forschritt-Bericht Nr. 484, VDI, Reihe 6, Düsseldorf, Germany, ISBN 3-18-348406-4. Nagel H., Spliethoff H. and Hein K.R.G. (1998) Untersuchungen zum Einfluss des Hybridkonzeptes auf den Betrieb einer Druckwirbelschicht, Proceedings of the VGB conference: “Forschung der Kraftwerkstechnik 1998”, C5, pp.1-18.Essen. Nakata, T., Sato, M., Ninomiya, T. and Hasegawa, T. (1996) A study on low NOx combustion characteristics in LBG-fueled 1500 °C-class gas turbine, Journal of engineering and power, 118, pp. 534-540. Nandi, S.P. and Orischak, M. (1985) Gasification of chars obtained from maple and jack pine wood, In Overend, R.P., Milne, T.A. & Mudge, L.K (eds.) conference proceedings Fundamentals of Thermochemical Biomass Conversion, pp. 567-587. Narvaéz, I., Orío, A., Aznar, M.P. and Corella, J. (1996) Biomass Gasification with Air in an Atmospheric Bubbling Fluidized Bed. Effect of Six Operational Variables on the Quality of the Produced Raw Gas, Ind. Eng.Chem. Res., 35, pp. 2110-2120. Naumann, D. (2000) Infrared Spectroscopy in Microbiology. In: Meyers, R.A. (ed.) Encyclopedia of Analytical Chemistry, John Wiley & Sons Ltd., Chichester, UK, pp. 102-131. Neeft, J. et al. (2002) www.tarweb.net Neeft, J.P.A. (2000) Teren uit pyrolyse en vergassing van biomassa en reststromen – definities, vorming, eigenschappen, en bemonstering en analyse, ECN report nr. ECN-C—99-102, Petten, The Netherlands (in Dutch). Nelson, P.F., Li, C.-Z. and Ledesma, E. (1996) Formation of HNCO from the Rapid Pyrolysis of Coals, Energy & Fuels, 10, pp. 264 – 265. Nelson, P.F., Buckley, A.N. and Kelly, M.D. (1992) Functional Forms of Nitrogen in Coals and the release of Coal Nitrogen as NOx Precursors (HCN and NH3), 24th Symp (Int) Combust, The Combustion Institute, Pittsburgh, PA, pp. 1259-1267. Nelson, P.F., Kelly, M.D. and Wornat, M.J. (1991) Conversion of fuel nitrogen in coal volatiles to NOx precursors under rapid heating conditions, Fuel, 70, pp.403 – 407. Neogi, D., Chang, C.C., Walawender, W.P. and Fan, L.T. (1986) Study of Coal Gasification in an Experimental Fluidized Bed Reactor, AIChE Journal, 32 (1), pp. 17-28. Nichols, K.M., Hedman, P. and Smoot, L.D. (1987) Release and reaction of fuel-nitrogen in a high-pressure entrained-coal gasifier, Fuel, 66, pp. 1257-1263. Nieminen, J. and Kivelä, M. (1998) Biomass CFB gasifier connected to a 350 MWth steam boiler fired with coal and natural gas – Thermie demonstration project in Lahti in Finland, Biomass & Bioenergy,15 (3), pp.251-257. Niksa, S. (1996) Coal Combustion Modelling, IEAPER/31, IEA Coal Research, London. Niksa, S. and Cho, S. (1996) Conversion of Fuel-Nitrogen in the Primary Zones of Pulverized Coal Flames, Energy&Fuels, 10, pp. 463-473. Nordin, A., Kallner, P. and Johansson, E. (1997) Gas quality from biomass gasification; an extensive parametric equilibrium study. In: Developments in Thermochemical Biomass Conversion, Bridgwater A.V. (ed.), Blackie Academic & Professional, pp. 838-850. Norman, J., Pourkashanian, M. and Williams, A. (1997) Modelling the formation and emission of environmentally unfriendly coal species in some gasification processes, Fuel, 76, pp. 1201-1216. Nunn, T.R., Howard, J.B., Longwell, J.P. and Peters, W.A. (1985a) Product Composition and Kinetics in the Rapid Pyrolysis of Sweet Gum Hardwood’, Ind. Eng. Chem. Process. Des. Dev, 24(3), pp. 836-844. Nunn, T.R., Howard, J.B., Longwell, J.P. and Peters W.A. (1985b) Product Compositions and Kinetics in the Rapid Pyrolysis of Milled Wood Lignin. Ind. Eng. Chem. Process Des. Dev., 24, pp. 844-852. Ohtsuka, Y. and Wu, Z. (1999) Nitrogen release during fixed-bed gasification of several coals with CO2: factors controlling formation of N2, Fuel, 78, pp. 521-527. Ohtsuka, Y., Zhiheng, W. and Furimsky, E. (1997) Effect of alkali and alkaline earth metals on nitrogen release during temperature programmed pyrolysis of coal, Fuel, 76, pp. 1361-1367. Ohtsuka, Y. and Furimsky, E. (1995) Effect of Ultrafine Iron and Mineral Matter on Conversion of Nitrogen and Carbon during Pyrolysis and Gasification of Coal, Energy&Fuels, 9, pp. 141-147. Ohtsuka , Y., Mori, H., Watanabe, T., Asami, K. (1994) Nitrogen removal during atmospheric-pressure pyrolysis of brown coal with iron, Fuel, 73(7), pp.1093-1097.

217

Ohtsuka, Y., Mori, H., Nonaka, K, Watanabe, T. and Asami, K. (1993) Selective Conversion of Coal Nitrogen to N2 with Iron, Energy&Fuels, 7, pp. 1095-1096. Okumura, Y., Sugiyama, Y. and Okazaki, K. (2002) Evolution prediction of coal-nitrogen in high pressure pyrolysis processes, Fuel, 81, pp. 2317-2324. Olanders, B. and Strömberg, D. (1995) Reduction of Nitric Oxide over Magnesium Oxide and Dolomite at Fluidized Bed Conditions, Energy & Fuels, 9, pp. 680-684. OPET Report nr. 4 (2001) Review of Finnish biomass gasification technologies, VTT Energy, Espoo, Finland. Oude Lohuis, J.A., Tromp, P.J.J. and Moulijn, J.A. (1992) Parametric study of N2O formation in coal combustion, Fuel,71, pp. 9-14. Overend, R.P., Milne, T.A., and Mudge, L.K. (eds.), Fundamentals of Thermochemical Biomass Conversion, Elsevier, New York, 1985. Padban, N. (2000) PFB Air Gasification of Biomass, Investigation of Product Formation and Problematic Issues Related to Ammonia, Tar and Alkali, Ph.D. thesis, Lund University, department of Chemical Engineering II, Sweden. Padban, N., Wang, W., Ye, Z., Bjerle, I and Odenbrand I. (2000) Tar formation in pressurized fluidised bed air gasification of woody biomass, Energy & Fuels, 14 (3), pp. 603-611. Padban, N. and Odenbrand, I. (1999) Polynuclear aromatic hydrocarbons in fly ash from pressurized fluidised bed gasification of fuel blends. A discussion of the contribution of textile to PAHs, Energy & Fuels, 13 (5), pp. 1067-1073. Paisley, M.A., Farris, G., Slack, W. and Irving, J., (1997) Commercial development of the Battelle/Ferco biomass gasification process - initial operation of the McNeil gasifier, Proceedings of The Third Biomass Conference of the America’s: making a business from biomass in Energy, Environment, Chemicals, Fibers and Materials, Pergamon Press, August 24-29, Montreal, Canada. Paisley, M.A. and Farris, G. (1995) Development and Commercialization of a Biomass Gasification Power Generation System, Proceedings of The Second Biomass Conference of the America’s: Energy, Environment, Agriculture and Industry, Pergamon Press, August 21 – 24, Portland, Oregon, USA. Pan, Y.G., Velo, E., Roca, X., Manyà, Puigjaner, L. (2000) Fluidized-bed co-gasification of residual biomass/poor coal blends for fuel gas production, Fuel, 79, pp. 1317-1326. Pan, Y.G., Velo, E. and Puigjaner, L. (1999) Removal of tar by secondary air in fluidised bed gasification of residual biomass and coal, Fuel, 78, pp. 1703-1709. Patel, J. (2004) Demonstration of New Gasification Technology. In: Proceedings of the 2nd World Conference and Technology Exhibition on Biomass for Energy, Industry and Climate Protection, May 10-14, Rome, Italy (to be published). Paterson, N., Zhuo, Y., Dugwell, D.R. and Kandiyoti, R. (2002) Investigation of Ammonia Formation during Gasification in an Air-Blown Spouted Bed: Reactor Design and Initial Tests, Energy&Fuels, 16, pp. 127-135. Paterson, N. (1997) Fuel behaviour studies in the air-blown gasification cycle, Fuel, 76, pp. 1319-1325. Peck, R.E., Glarborg, P. and Johnsson, J.E. (1991) Kinetic Modeling of Fuel-Nitrogen Conversion in One-Dimensional, Pulverized-Coal Flames, Combust.Sci.Tech., 76, pp. 81-109. Pels, J.R. (1995) Nitrous Oxide in Coal Combustion, PhD Thesis, Delft University of Technology, The Netherlands. ISBN 90-5166-466-X. Pels, J.R., Kapteijn, F., Moulijn, J.A., Zhu, Q. and Thomas, K.M. (1995) Evolution of Nitrogen Functionalities in Carbonaceous Materials During Pyrolysis, Carbon, 33, pp. 1641- 1653. Perry, D.L. and Grint, A. (1983) Application of XPS to coal characterization, Fuel, 62(9), pp. 1024-1033. Pfab, F. (2001) Vergasung biogener Feststoffe in einem Wirbelkammerreaktor, Ph.D. thesis, Berlin Technical University, Germany. ISBN 3-89825-373-2. Phong-Anant, D., Wibberley, L.J. and Wall, T.F. (1985) Nitrogen Oxide Formation from Australian Coals, Combustion&Flame, 62, pp. 21-30. Phyllis database (2003), http://www.ecn.nl/phyllis Pinto, F., Franco, C., Andre, R.N., Miranda, M., Gulyurtlu, I and Cabrita, I. (2002) Co-gasification study of biomass mixed with plastic wastes, Fuel, 81, pp. 291-297. Pitcher, K., Paterson, B., Weekes, A., Neergaard, M. and Ståhl, K. (2002) Progress achieved in BIGCC projects and prospects for the future. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. And Grassi, A. (eds.) Proceedings of the 12th European Biomass Conference “Biomass for Energy, Industry and Climate Protection”, 17-21 June 20002, Amsterdam , The Netherlands, pp. 71-76.

218

Pitcher, K., Hilton, B and Lundberg, H. (1998), The ARBRE project: Progress achieved, Biomass & Bioenergy,15 (3), pp.213-218. Ploeg, J.E.G. (2000) Gasification Performance of the Demkolec IGCC. In: Proceedings of the Gasification 4 The Future IchemE conference,Session 3 – Operations, Noordwijk April 11-13, 2000, The Netherlands. Pohl, J.H. and Sarofim, A.F. (1976) Devolatilization and Oxidation of Coal Nitrogen. In: Proceedings of the 16th Symposium (International) on Combustion/The combustion institute, pp. 491-501. Prater, C.D. (1958) The temperature produced by heat of reaction in the interior of porous particles, Chem.Eng.Sci., 8, pp. 284-286. Purdy, M.J., Felder, R.M., Ferrell, J.K. (1981) Coal gasification in a pilot scale fluidized bed reactor. 1. Gasification of a devolatilized bituminous coal, Ind. Eng. Chem. Process Des. Dev., 20, 675-682. Purvis, C.R. and Craig, J.D. (1998) A small scale biomass fuelled gas turbine power plant, , Proceedings of the BioEnergy ’98: Expanding BioEnergy Partnerships conference, pp.578-584, October 4-8 Madison, Wisconsin, USA. Raman, P., Walawender, W.P., Fan, L.T. and Chang, C.C. (1981) Mathematical model for the fluid-bed gasification of biomass materials. Application to feedlot manure, Ind. Eng. Chem. Process Des. Dev., 20, pp. 686-692. Rapagna, S., Jand, N., Kiennemann, A. and Foscolo, P.U. (2000) Steam-gasification of biomass in a fluidised-bed of olivine particles, Biomass & Bioenergy, 19, pp. 187-197. Rapagna, S., Jand, N. and Foscolo, P.U. (1998) Catalytic gasification of biomass to produce hydrogen rich gas, Int. J. Hydrogen Energy, 23(7), pp.551-557. Raskin, N., Palonen, J. and Nieminen, J. (2001) Power boiler fuel augmentation with a biomass fired atmospheric circulating fluid-bed gasifier, Biomass & Bioenergy, 20, pp. 471-481. Reed, T.B. (1981) Biomass Gasification Principles and Technology, T.B. Reed (ed.), Noyes Data Co., New Jersey, USA. Reed, T.B. and Gaur, S. (1999) A survey of biomass gasification 2000 – Gasifier Projects and Manufacturers Around the World, The Biomass Energy Foundation, T.B. Reed (ed.), Golden, Colorado, USA. Reed, T.B. and Bryant, B. (1978) Densified Biomass a new form of solid fuel with a forward for the 21st century. The Biomass Energy Foundation (BEF), Golden, Colorado, USA. Redlich, P.J., Jackson, W.R., Larkins, F.P and Rash, D. (1989) Studies related to the structure and reactivity of coals, Fuel, 68, pp. 222-230. Reid, R.C., Prausnitz, J.M. and Sherwood, T.K., (1977) The properties of gases and liquids, 3rd. Ed., New York, McGrawHill. Rensfelt, E.K.W. (1997) Atmospheric CFB gasification – The Greve plant and beyond, Proceedings of the International Conference of Gasification and Pyrolysis of Biomass, pp.139-159, April 9 –11, Stuttgart, Germany. Rensfelt, E., Blomkvist, G., Ekström, C., Engström, S., Espenäs, B.G. and Liinanki, L. (1978) Basic gasification studies for development of biomass medium-BTU gasification processes, Proceedings of the symposium “Energy from biomass and wastes”, Washington D.C. (USA), 14-18 August, pp. 465-494. Rhinehart, R.R., Felder, R.M. and Ferrell, J.K. (1987) Dynamic Modeling of a Pilot-Scale Fluidized-Bed Coal Gasification Reactor, Ind. Eng. Chem. Res., 26, pp. 738-745. Röhricht, C. and Beier, T. (1998) The suitability for cultivation, yield potential and ecological performance of energy crops grown on various soils in the German state of Saxony, Proceedings of the 10th European Conference and Technology Exhibition “Biomass for Energy and Industry”, 8-11 June, Würzburg, Germany, pp. 784 - 786. Roll, H. (1994) Vergasung von grob gemahlenem Schilfgras (Miscanthus Sinensis Giganteus) im Flugstrom, Ph.D. thesis, University of Karlsruhe, faculty of chemical engineering, Germany. Rosén, C., Björnbom, E., Yu, Q and Sjöström, K. (1997) Fundamentals of pressurized gasification of biomass. In: Developments in Thermochemical Biomass Conversion (Eds. Bridgwater, A.V. and Boocock, D.G.B.), Blackie Academic & Professional, London, pp. 817-827. Rota, R., Morbidelli, M. and Carrà, S. (1998) Combustion Kinetics of Light Hydrocarbons in the Presence of Nitrogen Oxide, Ind.Eng.Chem.Res., 37, pp. 4241-4252. Rota, R., Bonini, F., Servida, A., Morbidelli, M. and Carrá, S. (1997) Modeling of the Reburning Process, Combust.Sci.Technol., 123, pp. 83-105. Rowe, P.N. (1984) The effect of pressure on minimum fluidisation velocity, Chem.Engng.Sci., 39, pp. 173-174. Rüdiger, H. (1997) Pyrolyse von festen biogenen und fossilen Brennstoffen zur Erzeugung eines Zusatzbrennstoffs für Feuerungsanlagen. PhD thesis, Stuttgart University.

219

Rüdiger, H., Kicherer, A., Greul, U., Spliethoff, H. and Hein, K.R.G. (1996) ‘Two concepts for a combined combustion of biomass with hard coal’, Proceedings of the 21st Internat. Technical Conf. on Coal Utilization & Fuels Syst., March 18-21, Clearwater, USA., pp. 557-567. Rüdiger, H., Kecherer, A., Greul, U., Spliethoff, H. and Hein, K.R.G., (1996) Investigations in Combined Combustion of Biomass and Coal in Power Plant Technology, Energy & Fuels, 10 (3), pp. 789-796 Rüdiger, H., Greul, U., Spliethoff, H. and Hein, K.R.G. (1995) Co-pyrolysis of Coal/Biomass and Coal/Sewage Sludge mixtures in an Entrained Flow Reactor’, Final Report. In: Bemtgem, J.M., K.R.G. Hein, and A.J. Minchener (Eds.), COGasification of Coal/Biomass and Coal/Waste Mixtures, EC-Research Programme, APAS-Contract COAL-CT92-0001, Institute for Process Engineering and Power Plant Technology, University of Stuttgart, Germany Rumpel, S. (2000) Die autotherme Wirbelschichtpyrolyse zur Erzeugung heizwertreicher Stützbrennstoffe, Ph.D. thesis, University of Karlsruhe, faculty of Chemical Engineering, Germany. Saastamoinen, J. and Hämäläinen, J. (1999) Release of Nitrogen from wood. In: Proceedings of the IEA-workshop “NOx/N2O Formation and Destruction in Fluidized Bed Combustors”, May 16th 1999, Savannah, USA. Sadaka, S.S., Ghaly, A.E. and Sabbah, M.A. (2002) Two phase biomass air-steam gasification model for fluidized bed reactors: Part I – model development, Biomass & Bioenergy, 22, pp. 439-462. Sadaka, S.S., Ghaly, A.E. and Sabbah, M.A. (2002) Two phase biomass air-steam gasification model for fluidized bed reactors: Part II – model sensitivity, Biomass & Bioenergy, 22, pp. 463-477. Sadaka, S.S., Ghaly, A.E. and Sabbah, M.A. (2002) Two phase biomass air-steam gasification model for fluidized bed reactors: Part III – model validation, Biomass & Bioenergy, 22, pp. 479-487. Salo, K., Horvath, A., Mojtahedi, W. and Patel, J. (1998b) Pressurized Gasification of Biomass, Paper 98-GT-349 presented at the ASME Turbo Expo ’98, June 2-5, 1998, Stockholm, Sweden. Salo, K. and Patel, J.G. (1997) Integrated Gasification Combined Cycle Based on Pressurized Fluidized Bed Gasification. In: Developments in Thermochemical Biomass Conversion (Eds. Bridgwater, A.V. and Boocock, D.G.B.), Blackie Academic & Professional, London, pp.994-1005. Samson, R. and Duxbury, P. (2000) Assessment of Pelletized Biofuels, "Resource efficient Agriculture Production - Canada" report, Quebec, Canada. Scheibner, B. and Wolters, C. (2002) Schumacher hot gas filter long-term operating experience in the NUON power Buggenum IGCC power plant. In: 5th International Symposium on Gas Cleaning at High Temperature, Morgantown, September 17-20, 2002, USA. Schellberg,W. (2000) Project Status of the Puertollano IGCC Plant. In: Proceedings of the Gasification 4 The Future IchemE conference, Session 3 – Operations, Noordwijk, April 11-13, 2000, The Netherlands. Schlich, E. (1977) Mathematisches Modell für die Vergasung und Verbrennung von Steinkohle unter Druck, Ph.D. thesis, RWTH Aachen, Germany. Schoderböck, P. and Lahaye, J. (1996) The influence of impurities contained in quartz sand on the catalytic reduction of nitric oxide by carbon monoxide, Appl. Surf. Sci., 93, pp. 109-118. Schot, M.S. and Laro, D.A.H. (2000) Hydrodynamisch Gedrag van een Wervelbed, BSc final assignment, Delft University of Technology, section Thermal Power Engineering (in Dutch). Seidle, J.P. and Branch, M.C. (1983) Hydroxyl Concentration Measurements in the NH3-NO-O2 Reaction in Postflame Gases, Combustion&Flame,52, pp.47-57. Selinger, A., Steiner, C., Shin, K. (2003) TwinRec – Bridging the Gap of Car Recycling in Europe. Proceedings of the International Automobile Recycling Congress, March 12-14, Geneva, Switzerland. Serio, M.A., Charpenay, S., Bassilakis, R., and Solomon, P.R. (1994) Biomass & Bioenergy, 7 (1-6), pp. 107-124. Sett, A. and Bhattacharya, S.C. (1988) Mathematical Modelling of a Fluidised-Bed Charcoal Gasifier, Applied Energy, 30, pp. 161-186. Shafizadeh, F. (1984) in The Chemistry of Solid Wood (R.M. Rowell, ed.), Advances in Chemistry Series no. 207, American Chemical Society, Washington DC, U.S.A. Shand, R.N. and Bridgwater, A.V. (1984) Fuel gas from biomass: status and new modelling approaches. In: Bridgwater A.V. (ed.), Thermochemical Processing of Biomass. Butterworths, London, UK, pp. 229-254. Shesh, K.K. and Sunavala, P.D. (1990) Thermodynamics of pressurized air-steam gasification of biomass, Indian J. Technol., 28, pp. 133-138. Shimizu T, Ishizu K, Kobayashi S, Kimura S, Shimizu T, Inagaki M. (1993) Hydrolysis and Oxidation of HCN over Limestone under Fluidized Bed Combustion Conditions, Energy & Fuels, 7, pp. 645-647.

220

Shimizu, T., Tachimaya, Y, Fujita, D., Kumazawa, K., Wakayama, O., Ishizu, K, Kobayashi, S., Shikada, S. and Inagaki, M. (1992) Effect of SO2 Removal by Limestone on NOx and N2O emissions from a Circulating Fluidized Bed Combustor, Energy & Fuels, 6, pp. 753-757. Siemons, R.V. (2002) A development perspective for biomass-fuelled electricity generation technologies – Economic technology assessment in view of sustainability, Ph.D. thesis, University of Amsterdam, The Netherlands. ISBN 90-3651805-9. Simell P, Stahlberg P, Kurkela E, Albrecht J, Deutsch S, Sjöström K. (2000) Provisional protocol for the sampling and analysis of tar and particulates in the gas from large-scale biomass gasifiers. Version 1998. Biomass & Bioenergy, 18 (1), pp. 19-38. Simell, P.A., Hepola, J.O. and Krause, A.O.I. (1997) Effects of gasification gas components on tar and ammonia decomposition over hot gas cleanup catalysts, Fuel, 76(12), pp. 1117-1127. Simell, P., Kurkela, E., Ståhlberg, P. and Hepola, J. (1996) Catalytic hot gas cleaning of gasification gas, Catal. Today, 27, pp. 55-62. Simons, G.A. and Finnson, M.L. (1979) The structure of coal char: Part I. Pore branching, Combustion Science and Technology, 19, pp. 217-25. Sjöström, K., Chen, G., Yu, Q., Brage, C. and Rosen, C. (1999) Promoted reactivity of char in co-gasification of biomass and coal: Synergies in the thermochemical process, Fuel,78 (10), 1189-1194. Skoog, D.A., West, D.M. and Holler, F.J. (1992) Fundamentals of Analytical Chemistry, sixth edition, Saunders College Publishing, Fort Worth. Slabbekoorn, A. (2002) Flash Pyrolysis of Miscanthus – Validation of the FG-DVC biomass model with heated grid reactor experiments, report EV-2090, Delft University of Technology, section Thermal Power Engineering. Slaughter, D.M., Overmoe, B.J. and Pershing, D.W. (1988) Inert pyrolysis of stoker-coal fines, Fuel, 67(4), pp. 482-489. Sloss, L.L. (2002) Non-CO2 greenhouse gases-emissions and control from coal, IEA report number CCC/62, IEA Clean Coal Centre, United Kingdom, ISBN 92-9029-375-6. Smeenk, J., Brown, R.C. and Eckels, D. (1999), Determination of vapor phase alkali content during biomass gasification, Proceedings of the Fourth Biomass Conference of the Americas: Biomass a Growth Opportunity in Green Energy and ValueAdded Products, pp. 961-967, Oakland, California, USA. Smeenk, J. and Brown, R.C. (1998) Experience with atmospheric fluidised bed gasification of switchgrass, Proceedings of the BioEnergy ’98: Expanding BioEnergy Partnerships conference, pp.600-606, October 4-8 Madison, Wisconsin, USA. Smith, G.P., Golden, D.M., Frenklach, M., Moriarty, N.W., Eiteneer, B., Goldenberg, M., Bowman, C.T., Hanson, R.K., Song, S., Gardiner, W.C., Lissianski, V.V. and Qin, Z. (2000) http://www.me.berkeley.edu/gri_mech/ Smith, I.W. (1982) The Combustion Rates of Coal Chars: A Review, Proceedings of the 19th Symposium (International) on Combustion/The combustion institute, pp. 1045-1065. Smith, J.M., Van Ness, H.V. and Abbott, M.M. (2001) Introduction to chemical engineering thermodynamics, 6th edition, McGraw-Hill, Singapore. ISBN 0-07-240296-2. Smith, K.L., Smoot, L.D., Fletcher, T.H. and Pugmire, R.J. (1994) The Structure and Reaction Processes of Coal, Plenum Press, New York, USA. Solomon, P.R., Hamblen, D.G., Serio, M.A., Yu, Z.-Z. and Charpenay, S. (1993) A characterization method and model for predicting coal conversion behaviour, Fuel, 74 (2), pp. 469-488. Solomon, P.R., Serio, M.A. and Suuberg, E.M. (1992) Coal Pyrolysis: Experiments, Kinetic Rates and mechanisms, Prog.Energy Comb.Sci., 18, pp. 133 - 220. Solomon, P.R., Hamblen, D.G., Serio, M.A., Yu, Z.-Z. and Serio, M.A. (1990) Network models of coal thermal decomposition, Fuel, 69 (6), pp. 754-763. Solomon, P.R., Hamblen, D.G., Carangelo, R.M., Serio, M.A. and Deshpande, G.V. (1988) General model of coal devolatilization, Energy & Fuels, 2, pp. 405-422. Solomon, P.R., P.R., and Hamblen, D.G. (1985) in Chemistry of Coal Conversion (R.H. Schlosberg, ed.), Plenum Press, New York, chapter 5, pp. 121-251. Solomon, P.R., Hamblen, D.G., Carangelo, R.M. and Krause, J.L. (1982), Coal Thermal Decomposition in an Entrained Flow Reactor: Experiments and Theory, 19th Symp. (Int.) Combust., The Combustion Institute, Pittsburgh, PA, pp. 1139 – 1149. Solomon, P.R. and Colket, M.B. (1978) Evolution of fuel nitrogen in coal devolatilization, Fuel, 57, pp. 749-755. Solomons, T.W.G. (1984) Organic Chemistry, 3rd edition, John Wiley & Sons, New York.

221

Sowa, F. (1991) Reaktionskinetische Grundlagen und Monte-Carlo Simulation der Kohlevergasung in der Wirbelschicht, Ph.D. thesis, Universität Essen, Germany. (in German) Speight, J.G. (1983) The Chemistry and Technology of Coal, Marcel Dekker, ISBN 0-8247-1915-8. Spliethoff, H. (2000) Verbrennung fester Brennstoffe zur Strom- und Wärmeerzeugung, Verfahren und Stand der Technik – Wirkungsgrad, Betrieb, Emissionen und Reststoffe, Fortschritt-Berichte VDI, Reihe 6 Energietechnik Nr. 443, VDI Verlag GmbH, Düsseldorf, Germany. Srinivas, B. and Amundson, N.R. (1980) A Single-Particle Char Gasification Model, AIChEJ, 26, pp. 487-496. Ståhl, K. (2001) Värnamo Demonstration Plant – A demonstration plant for biofuel-fired combined heat and power generation based on pressurized gasification, report, Berlings Skogs, Trelleborg 2001. Ståhl, K. and Neergaard, M. (1998) IGCC Power Plant For Biomass Utilisation, Värnamo, Sweden, Biomass & Bioenergy,15 (3), pp.205-211. Ståhl, K., Neergaard, M. and Nilsson, P., (1997) Pressurized CFB gasification – The Värnamo Plant, Proceedings of the International Conference of Gasification and Pyrolysis of Biomass, pp. 160-171, April 9 –11, Stuttgart, Germany. Ståhlberg, P., Lappi, M., Kurkela, E., Simell, P., Oesch, P. and Nieminen, M. Pressurised fluidised-bed gasification experiments with biomass,, peat and coal at VTT in 1991-94, part 3 Gasification of Danish wheat straw and coal, VTT Research Notes 1903, VTT Technical Research Centre of Finland, Finland. Stanczyk, K. and Boudou, J.P. (1994) Elimination of nitrogen from coal in pyrolysis and hydropyrolysis: a study of coal and model chars, Fuel, 73(6), pp. 940-944. Sternberg, J.C., Gallaway, W.S. and Jones, D.T.L. (1962) In: Third International Gas Chromatography Symposium (Brenne, Callen, Weis, Hrsg.). Academic Press Inc., New York, p. 231. Stoholm, P., Nielsen, R.G., Nygaard, H., Tobiasen, L., Fock, M.W., Richardt, K. and Henriksen, U. Low Temperature CFB Gasifier, conceptual ideas, applications and first test results. In: Kyritsis, S., Beenackers, A.A.C.M., Helm, P., Grassi, A. and Chiaramonti, D. (eds.) Proceedings of the 1st World Conference on Biomass for Energy and Industry, June 5-9 Sevilla, Spain, pp.1893-1896. Storm, C., Rüdiger, H., Spliethoff, H. and Hein, K.R.G. (1999) ‘Co-pyrolysis of Coal/Biomass and Coal/Sewage Sludge Mixtures’, Journal of Engineering for Gas Turbines and Power, 121, pp. 55-63. Stryer, L. (1988) Biochemistry, third edition, W.H. Freeman and Company, New York, USA. Stubington, J.F. and Aiman, S. (1994) Pyrolysis Kinetics of Bagasse at High Heating Rate, Energy&Fuels, 8, pp. 194-203. Tabasaran, O., Besemer, G. and Thomanetz, E. (1977) Pyrolyse organischer Abfälle als Behandlungsmethode und als Möglichkeit zur Gewinnung von Energie und Rohstoffen, Müll und Abfall, 10, pp. 293-300. Takagi, H., Isoda, T., Kusakabe, K. and Morooka, S. (1999) Effects of Coal Structures on Denitrogenation during Flash Pyrolysis, Energy&Fuels, 13, pp. 934-940. Tan, L.L. and Li, C.-Z. (2000) Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part I. Effects of reactor configuration on the determined yields of HCN and NH3 during pyrolysis, Fuel, 79, pp. 1883-1889. Tan, L.L. and Li, C.-Z. (2000) Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part II. Effects of experimental conditions on the yields of NOx and SOx precursors from the pyrolysis of a Victorian brown coal, Fuel, 79, pp. 1891-1897. Teng, H., Serio, M.A., Wójtowicz, M.A., Bassilakis, R. and Solomon, P.R. (1995) Reprocessing of used tires into activated carbon and other products. Ind. Eng. Chem. Res., 34, pp. 3102-3111. Thambimuthu, K.V. (1993) Gas cleaning for advanced coal-based power generation. Report nr. IEACR/53, IEA Coal Research, London. Thomas, K.M. (1997) The release of nitrogen oxides during char combustion, Fuel, 76 (6), pp. 457-473. Tian, F.-J., Li, B.-Q., Chen, Y. and Li, C.-Z. (2002) Formation of NOx precursors during the pyrolysis of coal and biomass. Part V. Pyrolysis of a sewage sludge, Fuel, 81, pp. 2203-2208. Tsubouchi, N., Ohshima, Y., Xu, C. and Ohtsuka, Y. (2001) Enhancement of N2 Formation from the Nitrogen in Carbon and Coal by Calcium, Energy&Fuels, 15, pp. 158-162. Tsujimura, M, Furusawa, T. and Kunii, D. (1983) Catalytic Reduction of Nitric Oxide by Carbon Monoxide over Calcined Limestone, J.Chem.Eng.Jpn., 16(2), pp. 132-136. Turn, S., Kinoshita, C., Zhang, Z., Ishimura, D. and Zhou, J. (1998a) An experimental investigation of Hydrogen production from biomass gasification, Int. J. Hydrogen Energy, 23 (8), pp. 641-648.

222

Turn, S.Q., Kinoshita, C.M., Ishimura, D.M. and Zhou, J. (1998b) The fate of inorganic constituents of biomass in fluidized bed gasification, Fuel, 77(3), pp. 135-146. Tullin, C.J., Sarofim, A.F. and Beér, J.M. (1993) Formation of NO and N2O in coal combustion: the relative importance of volatile and char nitrogen, J. Inst. Energy, 66, pp. 207-215. Ünal, Ö. (2005) Pressurised fluidised bed gasification of biomass and browncoal: Ash and trace element behaviour, Ph.D. thesis to be published, Delft University of Technology, The Netherlands. United Nations Framework Convention on Climate Change (1997) Kyoto protocol by the 3rd Conference of the Parties. Van den Aarsen, F.G. (1985) Fluidised bed wood gasifier performance and modeling, PhD thesis, University of Twente, department of chemical engineering, The Netherlands. Van den Enden, P.J. and Silva Lora, E. (2004) Design approach for a biomass fed fluidised bed gasifier using the simulation software CSFB, Biomass&Bioenergy (to be published) Van den Enden, P.J. (2000) Report on gasifier design for an experimental 45 kWe Biomass gasifier and micro turbine installation, report EV 2048 Delft University of Technology, section Thermal Power Engineering, pp. 101-102. Van den Heuvel, E.J.M.T. (1995) Pretreatment technologies for energy crops, report EWAB-9525, Novem, Utrecht, The Netherlands. Van den Heuvel, E.J.M.T. and Stassen, H.E.M., (1994) Elektriciteit uit biomassa een vergelijk tussen verbranding en vergassing, EWAB report 9414 (in Dutch), NOVEM, Utrecht, The Netherlands. Van der Drift, A., Van der Meijden, C.M. and Strating-Ytsma, S.D. (2002) Ways to increase the carbon conversion of a CFBgasifier. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. And Grassi, A. (eds.) Proceedings of the 12th European Biomass Conference “Biomass for Energy, Industry and Climate Protection”, 17-21 June 2002, Amsterdam, The Netherlands, pp. 604-606. Van der Drift, A., Van Doorn and Vermeulen, J.W. (2001) Ten residual biomass fuela for circulating fluidized-bed gasification, Biomass & Bioenergy, 20, pp. 45-56. Van der Drift and Van Doorn, J. (2001) Effect of fuel size and process temperature on fuel gas quality from CFB gasification of biomass. In: Bridgwater, A.V. (ed.) Proceedings of conference Progress in Thermochemical Biomass Conversion, 17-22 September 2000, Seefeld, Austria, pp. 265-271. Van der Drift, A. and Olsen, A. (1999) Conversion of biomass, prediction and solution methods for ash agglomeration and related problems. Final Report EU-project JOR3-95-0079 (ECN-report ECN-C-99-090). Van Krevelen, D.W. (1993) Coal – Typology – physics – chemistry – constitution, 3rd edition, Elsevier, Amsterdam, ISBN 0 444 89586 8. Van Paasen, S.V.B., Bergman, P.C.A., Neeft, J.P.A. and Kiel, J.H.A. (2002) Primary measures for tar reduction. Reduce the problem at the source. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. And Grassi, A. (eds.) Proceedings of the 12th European Biomass Conference “Biomass for Energy, Industry and Climate Protection”, 17-21 June 20002, Amsterdam , The Netherlands, pp. 597-599. Van Smeerdijk, D.G. and Boon, J.J. (1987) Characterisation of subfossil Sphagnum leaves, rootlets of Ericaceae and their peat by pyrolysis-high-resolution gas chromatography-mass spectrometry, J.Anal.Appl.Pyrolysis, 11, pp. 377-402. Varey, J.E., Hindmarsh, C.J and Thomas, K.M. (1996) The detection of reactive intermediates in the combustion and pyrolysis of coals, chars and macerals, Fuel, 75, pp. 164 176. Venendaal, R. and Van Haren, P. (1999) Learning from experiences with Alternative Fuels in Electric Power Generating Plants, Centre for the Analysis and dissemination of demonstrated energy technologies (IEA-OECD) CADDET Analyses Series No. 26, ISBN-90-72647-44-0. Verloop, C.M. (1998) The chemistry in the freeboard of a coal-fired pressurized fluidised bed combustor, Ph.D. thesis, Delft University of Technology, The Netherlands. ISBN 90-407-1666-8. Vriesman, P., Heginuz, E. and Sjöström, K. (2000) Biomass gasification in a laboratory –scale AFBG: influence of the location of the feeding point on the fuel-N conversion, Fuel, 79, pp. 1371-1378. Wahlers, W. (1989) Beitrag zur Strukturaufklärung der Stickstoffverbindungen in Steinkohlen, Ph.D. thesis, University of Essen, Department of Chemistry, Germany. Wall, T.F. (1987) The Combustion of Coal as Pulverized Fuel Through Swirl Burners, chapter 3 in: “Combustion Treatise. Principles of Combustion Engineering for Boilers”, Lawn, C.J. (ed.), Academic Press. Wallace, S., Bartle, K.D. and Perry, D.L. (1989) Quantification of nitrogen functional groups in coal and coal derived products, Fuel, 68(11), pp. 1450-1455.

223

Waldheim, L., Fredriksson, C. and Svensson, O. (1999) GEF/UNDP Bagasse project, Summary report. Confidential internal TPS-report number TPS-99/37. Wallman, P.H., Carlsson, R.C.J. and Leckner, B. (1993) NOx reduction in Pressurized Fluidized-Bed Combustion. In: Proceedings of the 12th International Conference on Fluidized Bed Combustion, San Diego, USA, pp. 789-794. Wang, W. and Olofsson, G. (2002) Reduction of Ammonia and Tar in Pressurized Biomass Gasification. Proceedings of the 5th International Symposium on Gas Cleaning at High Temperature, 17-20 September 20002, Morgantown, USA. Watkinson, A.P., Lucas, J.P. and Lim, C.J. (1991) A prediction of performance of commercial coal gasifiers, Fuel, 70, pp. 519-527. Weeda, M. (1995) Kinetics of coal gasification under industrial conditions. PhD thesis, University of Amsterdam, Amsterdam, The Netherlands. Weeda, M., Abcouwer, H.H., Kapteijn, F. and Moulijn, J.A. (1993) Steam gasification kinetics and burn-off behaviour for a bituminous coal derived char in the presence of H2, Fuel Process.Techn., 36 (1-3), pp. 235-242. Weimer, A.W. and Clough, D.E. (1981) Modeling a low pressure steam-oxygen fluidized bed coal gasifying reactor, Chem. Eng. Sci.,36, pp. 549-567. Wen, C.Y. and Chaung, T.Z. (1979) Entrainment coal gasification modeling, Ind. Eng. Chem. Process Des. Dev., 18 (4), pp. 684-695. Westerterp, K.R., Van Swaaij, W.P.M. and Beenackers, A.A.C.M. (1984) Chemical reactor design and operation, John Wiley & Sons. Weterings, R.A.P.M., Bergsma, G.C., Koppejan, J. and Meeusen-van Onna, (1999) Beschikbaarheid van afval en biomassa voor energieopwekking in Nederland, GAVE-report 9911 (in Dutch), NOVEM, Utrecht, The Netherlands. Wiant, B., Bryan, B. and Miles, T. (jr.) Hawaiian biomass gasification facility: testing results and current status, Proceedings of the BioEnergy ’98: Expanding BioEnergy Partnerships conference, pp.414-423, October 4-8 Madison, Wisconsin, USA. Will E.M. (1999) Gasification of coal and wood. Eindhoven University of Technology, Department of Applied Physics, The Netherlands, Master Thesis no. R-1484-S. Willeboer, W. (1998), AMER demolition wood gasification project, Biomass & Bioenergy,15(3), pp. 245-249. Williams, A., Pourkashanian, M, Jones, J.M. and Skorupska, N. (2000) Combustion and Gasification of Coal, Taylor & Francis, New York, USA. Winter, F., Wartha, C. and Hofbauer, H. (1999) NO and N2O formation during the combustion of wood, straw, malt waste and peat, Bioresource Technol., 70, pp. 39-49. Wittler, W., Schütte, K., Rotzoll, G. and Schügerl, K. (1988) Heterogeneous reduction of nitric oxide by carbon monoxide on quartz surfaces, Fuel, 67(3), pp. 438-440. Wójtowicz, M.A., Bassilakis, R., Smith, W. and Serio, M.A. (2001) TG-FTIR characterization of biomass and coal samples – final report, Advanced Fuel Research, Inc., East Hartford CT USA. Wójtowicz, M.A., Pels, J.R. and Moulijn, J.A. (1995) The Fate of Nitrogen Functionalities in Coal during Pyrolysis and Combustion, Fuel, 74, pp. 507-516. Wójtowicz, M.A., Pels, J.R. and Moulijn, J.A. (1993) Combustion of coal as a source of N2O emission, Fuel Process.Technol., 34, pp. 1-71. Woudstra, Th. and Woudstra, N. (1995) Exergy analysis of hot-gas clean-up in IGCC systems, J. Institute of Energy, 68, pp. 157-166. Wu, Z. and Ohtsuka, Y. (1997) Key Factors for Formation of N2 from Low-Rank Coals during Fixed Bed Pyrolysis: Pyrolysis Conditions and Inherent Minerals, Energy&Fuels, 11, pp. 902-908. Wu, Z. and Ohtsuka, Y. (1996) Remarkable Formation of N2 from a Chinese Lignite during Coal Pyrolysis, Energy&Fuels, 10, pp. 1280-1281. Xie, Z., Feng, J., Zhao, W., Xie, K.-C., Pratt, K.C. and Li, C.-Z. (2001) Formation of NOx and SOx precursors during the pyrolysis of coal and biomass. Part IV. Pyrolysis of a set of Australian and Chinese coals, Fuel, 80, pp. 2131-2138. Yan, H. and Zhang, D.K. (2000) Modelling of fluidised-bed coal gasifiers: elimination of the combustion product distribution coefficient by considering homogeneous combustion, Chem.Eng.&Processing, 39, pp. 229-237. Yan, H., Heidenreich, C. and Zhang, D.K. (1999) Modelling of bubbling fluidised bed coal gasifiers, Fuel, 78, pp. 1027-1047. Yan, H., Heidenreich, C. and Zhang, D.K. (1998) Mathematical modelling of a bubbling fluidised-bed coal gasifier and the significance of ‘net flow’, Fuel, 77 (9/10), pp. 1067-1079.

224

Yin, X.L., Wu, C.Z., Zheng, S.P. and Chen, Y. (2002) Design and operation of a CFB gasification and power generation system for rice husk, Biomass & Bioenergy, 23, pp. 181-187. Yoon, H., Wei, J. and Denn, M.M. (1978) A Model for Moving-Bed Coal Gasification Reactors, AIChE J., 24(5), pp. 885903. Zand, A., Felder, R. and Ferrell, J. (1985) Analysis of nitrogeneous compounds in the effluent streams from a fluidised-bed coal gasification reactor, Fuel Processing Techn., 10, pp. 249-259. Zanzi, R. (2001) Pyrolysis of biomass. Rapid pyrolysis at high temperature. Slow pyrolysis for active carbon preparation, PhD thesis, Royal Institute of Technology, Stockholm, Sweden. Zanzi, R., Sjöström, K. and Bjornbom, E. (1998) Pyrolysis of Agricultural residues at high temperatures, Proceedings of the conference “BioEnergy ’98: Expanding Partnerships”, pp. 569-577. Zanzi, R., Sjöström, K. and Bjornbom, E. (1997) Rapid Pyrolysis of Straw at high Temperatures. In: A.V. Bridgwater & D.G.B. Boocock, (Eds.) Developments in Thermochemical Biomass Conversion, vol. 1, pp. 61-66, Blackie, London. Zanzi, R., Sjöström, K. and Bjornbom, E. (1996) Rapid high-temperature pyrolysis of biomass in a free-fall reactor, Fuel, 75, pp. 545-550. Zdenek, S., Ladislav, O. and Van Tuyen, N. (2002) Biomass and waste gasification in AFB reactor. Proceedings of the “New and Renewable Technologies for Sustainable Development” conference, June 24-26, Ponta Delgada, S. Miguel Island, Azores, Portugal. Zhang, W., Dai, X., Söderlind, U., Wu, C. and Luo, X. (2002) A preliminary test on an industrial biomass CFB gasifier. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. And Grassi, A. (eds.) Proceedings of the 12th European Biomass Conference “Biomass for Energy, Industry and Climate Protection”, 17-21 June 20002, Amsterdam , The Netherlands, pp. 745-748. Zhou, J., Masutani, M., Ishimura, D.M., Turn, S.Q. and Kinoshita, C.M (2000) Release of Fuel-Bound Nitrogen during Biomass Gasification, Ind. Eng. Chem. Res., 39, pp. 626-634. Zhou, J. (1998) Fuel-bound Nitrogen evolution during biomass gasification, PhD thesis, University of Hawaii, USA. Zhuo, Y., Paterson, N., Avid, B., Dugwell, D.R. and Kandiyoti, R. (2002) Investigation of Ammonia Formation during Gasification in an Air Blown Spouted Bed: The Effect of the Operating Conditions on Ammonia Formation and the Identification of Ways of Minimizing Its Formation, Energy&Fuels, 16, pp. 742-751. Zygourakis, K. (1988) The effects of pyrolysis conditions on the macropore structure of coal chars, Prepr.Pap.Am.Chem.Soc., Div.Fuel Chem., 33, pp. 951-.

225

APPENDIX 1 Details of analytical measurements 1.1

Spectra, spectral windows and calibration curves used for quantitative species analysis with FT-IR

Table A1.1 Spectral windows used for calibration and analysis of gaseous species Species

Regions [cm-1] start end

Windows [cm-1] start end

(corrected for interferring species)

Species

Regions [cm-1] start end

Windows [cm-1] start end

(corrected for interferring species)

CO

2096.9

2148.8

CO2

2392.4

3541.9

CH4

2846.6

2871.9

(CO2) C2H4

998.2

1016.7

(NH3) C2H2

3232.7

3286.1

(H2O) COS

2034.4

2075.9

(CO2, C2H4, CO) H2O

3970.0

3996.0

2096.9 2101.0 2108.9 2133.1 2141.2 2148.2 2392.4 3515.7 3532.3 3541.3

2097.7 2101.8 2109.5 2133.8 2143.4 2148.8 2394.9 3516.8 3534.0 3541.9

2846.6 2848.9 2851.7 2856.6 2857.7 2861.0 2866.6 2870.1 998.2 1001.2 1008.3 1014.7 3232.7 3246.2 3252.6 3262.5 3265.6 3269.1 3284.7 2034.4 2044.7 2047.8 2052.5 2070.8 2075.0 3970.0 3973.6 3979.1 3984.2 3988.8 3992.2 3994.0

2848.2 2849.9 2852.7 2857.3 2858.5 2862.4 2867.4 2871.9 999.9 1005.1 1010.1 1016.7 3234.4 3247.9 3253.9 3263.1 3266.8 3270.5 3286.1 2036.0 2045.3 2048.7 2053.4 2071.7 2075.9 3973.2 3975.5 3981.4 3986.9 3991.4 3993.0 3996.0

HCN

3345.0

3363.9

3345.0 3350.1 3352.9 3358.0 3363.2

3345.6 3351.0 3353.8 3358.8 3363.9

(NH3, C2H2) NO

1818.3

1937.6

(NH3, H2O, C2H4) NO2

1584.9

1600.2

1818.3 1831.5 1846.3 1911.7 1928.3 1933.9 1936.2 1584.9 1596.8

1820.8 1832.3 1846.8 1913.1 1930.8 1934.9 1937.6 1586.3 1600.2

(NH3) N2O

2184.2

2202.2

(CO, CO2) NH3

1112.6

1178.9

2184.2 2193.9 2197.4 2201.3 1112.6 1133.5 1150.2 1156.9 1165.2 1178.0

2185.1 2194.9 2199.3 2202.2 1113.5 1134.4 1150.7 1157.6 1165.8 1178.9

(C2H4) HCl

2702.5

2799.6

2702.5 2727.10 2751.4 2775.0 2798.2

2703.5 2728.3 2752.6 2776.5 2799.6

227

Figure A1.1a Calibration spectrum of CO.

Figure A1.1b Calibration curve of CO, calibration range: 0 – 20 vol%.

228

Figure A1.2a Calibration spectrum of CO2.

Figure A1.2b Calibration curve of CO2, calibration range: 2-20 vol%.

229

Figure A1.3a Calibration spectrum of CH4.

Figure A1.3b Calibration curve of CH4, calibration range: 0.5-5 vol%.

230

Figure A1.4a Calibration spectrum of C2H4.

Figure A1.4b Calibration curve of C2H4 , calibration range: 90-9000 ppmv.

231

1.6 1.5 1.4 1.3 1.2 1.1 1.0 Absorbance

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 3400

3300

3200

Wavenumbers (cm-1)

Figure A1.5a Calibration spectrum of C2H2.

Figure A1.5b Calibration curve of C2H2 , calibration range: 0-1.3 vol%.

232

3.0 2.8 2.6 2.4 2.2

Absorbance

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 2080

2060

2040

2020

Wavenumbers (cm-1)

Figure A1.6a Calibration spectrum of COS.

Figure A1.6b Calibration curve of COS, calibration range: 51-510 ppmv.

233

Figure A1.7a Calibration spectrum of H2O.

Figure A1.7b Calibration curve of H2O, calibration range: 3 - 21 vol%.

234

Figure A1.8a Calibration spectrum of HCN.

Figure A1.8b Calibration curve of HCN, calibration range: 50-500 ppm.

235

Figure A1.9a Calibration spectrum of NO.

Figure A1.9b Calibration curve of NO, calibration range: 20-200 ppm.

236

Figure A1.10a Calibration spectrum of NO2.

Figure A1.10b Calibration curve of NO2, calibration range: 20-200 ppm.

237

Figure A1.11a Calibration spectrum of N2O.

Figure A1.11b Calibration curve of N2O, calibration range: 20-200 ppm.

238

Figure A1.12a Calibration spectrum of NH3.

Figure A1.12b Calibration curve of NH3 , calibration range: 300-3000 ppm.

239

0.230 0.220 0.210 0.200 0.190 0.180 0.170 0.160 0.150 0.140 0.130

Absorbance

0.120 0.110 0.100 0.090 0.080 0.070 0.060 0.050 0.040 0.030 0.020 0.010 0.000 -0.010 -0.020 -0.030

4000

3800

3600

3400

3200

3000

2800

2600

2400 2200 Wavenumbers (cm-1)

2000

1800

1600

1400

Figure A1.13a Calibration spectrum of HCl.

Figure A1.13b Calibration curve of HCl , calibration range: 9 - 91 ppm.

240

1200

1000

800

1.2

Calibration curves used for quantitative species analysis with gas chromatography

In this section calibration curves of the off-line operated gas chromatograph are presented, applied for the measurements at the PFBG test rig. Oven temperature-ramp data: at time t0, the oven temperature is 50 °C and the oven is kept at that temperature for 9 minutes, then the temperature is increased with 10 °C/min till 180 °C and kept at that temperature for 5 minutes.

Figure A1.14

Calibration curve for CO; calibration range:5-30 vol%; Columns: 1) 0.5 m, 2.0 mm ID Ni column with 80-100 mesh Hayesep T support, 2) 0.5m, 2.0 mm ID Ni column with 80-100 mesh size Hayesep Q 3) 1.5 m, 2.0 mm ID Stainless steel column with Molsieve 13X support; retention time: 12.90 minutes, oven end-temperature: 180 °C.

Figure A1.15 Calibration curve for CO2; calibration range:2-30 vol%; Columns: 1) 0.5 m, 2.0 mm ID Ni column with 80-100 mesh Hayesep T support 2) 0.5m, 2.0 mm ID Ni column with 80-100 mesh size Hayesep Q 3) 1.5 m, 2.0 mm ID Stainless steel column with Molsieve 13X support; retention time: 3.44 minutes, oven end-temperature: 180 °C.

241

Figure A1.16

Figure A1.17

242

Calibration curve for H2; calibration range: 1-10 vol%; Columns: 1) 1.0 m, 2.0 mm ID Stainless Steel column with 80-100 mesh Hayesep Q support 2)1.0 m, 2.0 mm ID Stainless steel column with 80-100 mesh Molsieve 5A support; retention time:1.44 minutes, oven end-temperature: 180 °C.

Calibration curve for CH4,, calibration range: 0.36 - 7.6 vol%, Columns: 1) 2.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:4.00 minutes, oven end-temperature: 180 °C.

Figure A1.18 Calibration curve for C2H4,, calibration range: 0.07 - 1.5 vol%,%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:5.00 minutes, oven end-temperature: 180 °C.

Figure A1.19 Calibration curve for C2H6,, calibration range: 0.07 - 1.5 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:4.41 minutes, oven end-temperature: 180 °C.

243

Figure A1.20 Calibration curve for C2H2,, calibration range: 0.07 - 1.52 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:14.85 minutes, oven end-temperature: 180 °C.

Figure A1.21 Calibration curve for C3H6,, calibration range: 0.01 - 0.20 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:11.40 minutes, oven end-temperature: 180 °C.

244

Figure A1.22 Calibration curve for C3H8, calibration range: 0.01 - 0.23 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time:6.66 minutes, oven end-temperature: 180 °C.

Figure A1.23 Calibration curve for C3H4 (Propyne), calibration range: 0.008 - 0.18 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time: 20.13 minutes, oven end-temperature: 180 °C.

245

Figure A1.24 Calibration curve for C3H4 (Propadiene), calibration range: 0.009 - 0.20 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time: 14.00 minutes, oven end-temperature: 180 °C.

Figure A1.25 Calibration curve for C4H8 (1-Butene), calibration range: 0.01 - 0.20 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time: 17.16 minutes, oven end-temperature: 180 °C.

246

Figure A1.26 Calibration curve for C4H6 (1,3-Butadiene), calibration range: 0.01 - 0.20 vol%, Columns: 1) 12.5 m, 0.32 mm ID fused silica column with CP-SIL 5 CB stationary phase material 2) 50 m, 0.53 mm ID fused silica column with Al2O3/Na2SO4 stationary phase material 3) 3 m, 0.32 mm ID fused silica column with deactivated stationary phase; retention time: 19.93 minutes, oven end-temperature: 180 °C

Figure A1.27 Calibration curve for Ar, calibration range: 0.4-2 vol%, Columns: 1) 0.5 m, 2.0 mm ID Ni column with 80-100 mesh Hayesep T support, 2) 0.5m, 2.0 mm ID Ni column with 80-100 mesh size Hayesep Q 3) 1.5 m, 2.0 mm ID Stainless steel column with Molsieve 13X support; retention time:10.70 minutes, oven end-temperature: 180 °C.

247

Figure A1.28 Calibration curve for N2, calibration range: 70 - 100 vol%, Columns: 1) 0.5 m, 2.0 mm ID Ni column with 80-100 mesh Hayesep T support, 2) 0.5m, 2.0 mm ID Ni column with 80-100 mesh size Hayesep Q 3) 1.5 m, 2.0 mm ID Stainless steel column with Molsieve 13X support; retention time:11.10 minutes, oven end-temperature: 180 °C

248

APPENDIX 2 Relevant chemical & physical properties of the gasification product gas components 2.1

Gas phase viscosity

The viscosity of the gas phase is calculated according to the Chapman-Enskog model [Reid et al., 1977]. The pure component low pressure vapor viscosity, ηi, is retrieved from the following relation:

MiT η = 2.669 *10 − 26 2 (2,2)∗ i σkΩ

(A2.1)

where: Ω (2,2)∗ = f (Tk* =

k BT µ2 ,δ = ) εk 2εσ 3

(A2.2)

with Tk* the reduced temperature [-], εk the Lennard-Jones potential well depth [J], kB is Boltzmann’s constant [=1.3807.1023 J/K], σk the Lennard-Jones collision diameter [m] and Ωv the collision integral. Ω(2,2)* [-] is determined by a quadratic interpolation of the tables based on Stockmayer potentials given by [Monchick & Mason, 1961]. For the gas mixture the low-pressure vapor mixture viscosity is calculated by the Wilke approximation of the Chapman-Enskog equation: y iηi η mix = ∑ i ∑ y jΦ ij j

(A2.3)

For Φij the formulation by Brokaw is used: ⎡η Φ ij = ⎢ i ⎢ηj ⎣

1

⎤2 ⎥ S ij A ij ⎥ ⎦

(A2.4)

where

-

1

A ij = m ij M ij2

⎡ ⎢ ⎢ M ij − M ij0.45 ⎢1 + 1 ⎢ 0.45 2 1 + M m ⎢ ij ij ⎢ 2 1+M ij + 1 + M ij ⎣⎢

(

)

(

)

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦⎥

(A2.5)

with ⎡ 4 m ij = ⎢ -1 ⎢⎣ 1 + M ij 1 + M ij

(

)(

1

)

⎤4 ⎥ ⎥⎦

(A2.6)

and Mi (A2.7) Mj The term Sij is equal to unity for non-polar gases and for polar gases the following relation holds: M ij =

249

⎛δ δ ⎞ 1 + Ti*Tj* + ⎜ i j ⎟ 4 ⎠ ⎝

Sij = S ji =

2 ⎡ * ⎛δi 1+T + i ⎢ ⎜

⎝

⎣

1 2

⎞⎤ ⎡ 1+Tj* + ⎜ 4 ⎟⎠ ⎥ ⎢⎢ ⎦ ⎣ ⎝

⎛ δ j2

⎞⎤ 4 ⎟ ⎥⎥ ⎠⎦

1 2

(A2.8)

Here δ is the dipole moment parameter, related to the dipole moment µ by:

δ= 2.2

µ2 2εσ 3

(A2.9)

Diffusion coefficients of gas phase components

The binary diffusion coefficent at low pressures, Dvij is calculated using the Chapman-Enskog-WilkeLee model: ⎡ ⎧ ⎫⎤ 3 ⎢3.03 − ⎪⎨ 0.98 ⎪⎬⎥ (10−7 ) T 2 1 ⎢ 2 ⎥ ⎩⎪ M ij ⎭⎪⎦ ⎣

Dijv =

1

P M ij 2σ ij2 Ω(1,1)*

(A2.10)

where: ⎡ (M + M ) ⎤ i j ⎥ M ij = 2.⎢ ⎢ (M * M ) ⎥ i j ⎦ ⎣

−1

(A2.11)

The diffusion coefficient is expressed in the unit [m2/s] Mi, Mj in [g/mol], T in [K], P in [bar] in the above equations (A2.10) and (A2.11). The collision integral for diffusion is: Ω (1,1)* = f (T,

ε ij k

)

(A2.12)

The binary size and energy parameters are defined as: σ ij =

(σ i + σ j ) 2

(A2.13)

[

(A2.14)

ε ij = ε i * ε j

]12

Also here, polar parameter δ, defined by (A2.9),is used to determine whether to use the Stockmayer (in case of polar species) or Lennard-Jones potential parameters: ε/k (energy parameter) and σ (collision diameter). To calculate δ, the dipole moment, p, and either the Stockmayer parameters (when available in the applied database) or the dipole moment Tb and Vb are needed. In the last case the following equations are applied: 1

σ = 1.18Vb 3 and

ε k

= 1.15Tb

(A2.15)

(A2.16)

with Vb being the liquid molar volume in [cm3/mol] and Tb the normal boiling point at 1 atm in [K]. The diffusion coefficient of a gas into a gas mixture is calculated using Blanc’s law:

250

⎡ Dijv ⎤ Div = ∑ y j ⎢∑ ⎥ j≠ i ⎢⎣ j≠i y j ⎥⎦

(A2.17)

The binary diffusion coefficients Dvij at high pressure are determined from Chapman-Enskog-WilkeLee model, see above. 2.3

Gas Phase Thermal Conductivity

The pure component vapour thermal conductivity for low pressure gases is calculated according to the Stiel-Thodos method. Strictly spoken this method was suggested for nonpolar gases, but as the largest part of practical producer gas is nonpolar, this method can be used with a sufficient degree of accuracy. The equation is the following [Reid et al., 1977]:

⎡1.15 ( Cigp,i - R ) + 2.03 R ⎤ ⎦ λ i = ηi ⎣ (A2.18) Mi in which ηi is calculated according to the Chapman-Enskog-Brokaw method described in 2.1. The mixture vapour thermal conductivity is calculated with the Wassiljewa-Mason-Saxena mixing rule: y i λi λ mix = ∑ i ∑ y j A ij j

(A2.19)

with

A ij =

⎡ ⎧ ⎫ 12 M 14 ⎤ ⎢1 + ⎪⎨ηi ⎪⎬ ⎨⎧ j ⎬⎫ ⎥ ⎢ ⎪⎩η j ⎪⎭ ⎩ M i ⎭ ⎥ ⎣ ⎦ ⎡ ⎧⎪ ⎛ M ⎞ ⎫⎪⎤ ⎢8 ⎨1 + ⎜ i ⎟ ⎬⎥ ⎢⎣ ⎩⎪ ⎜⎝ M j ⎟⎠ ⎭⎪⎥⎦

1

1

2

(A2.20)

2

251

2.4

Thermodynamic data

In the calculations, thermodynamic data are used, obtained from the CHEMKIN III thermodynamic database [Kee et al., 1990] with revisions in the ’97 version of the program.

Table A2.1 Thermodynamic data for applied species in kinetic mechanism. Component C (g) H2 O2 H2O CO CH4 CO2 C2H4 C2H6 N2 NO NO2 N2O NH3 Ar HCN H N O CH CH2 CH3 C2H2 C2H3 C2H5 CN NH NH2 NNH N2H2 OH HO2 H2O2 H2CN CHO CH2O CH3O CH2OH CH2CO HCCO NCO HNO HNCO HOCN HCNO C4H2 C2H C3H2 CH2 (S) H2NO HNNO HONO NO3 C2N2 NCN N2H4 N2H3 C3H3

252

∆f H0 (298.15 K) [J/kmol] 7.17.108 0 0 -2.41.108 -1.11.108 -7.49.107 -3.94.108 5.25.107 -8.39.107 0 9.03.107 3.31.107 8.21.107 -4.59.107 0 1.33.108 2.18.108 4.73.108 2.49.108 5.94.108 3.87.108 1.46.108 2.27.108 2.86.108 1.17.108 4.35.108 3.57.108 1.90.108 2.45.108 2.13.108 3.90.107 1.05.107 -1.36.108 2.47.108 4.35.107 -1.16.108 1.63.107 -1.72.107 -5.19.107 1.78.108 1.32.108 9.96.107 -1.18.108 -1.48.107 1.61.108 4.67.108 5.65.108 5.42.108 4.25.108 6.62.107 2.31.108 -7.67.107 7.11.107 3.09.108 4.50.108 9.54.107 1.54.108 5.18.108

∆f S0 (298.15 K) [J / (kmol.K)] 1.52.105 0 0 -4.44.104 8.93.104 -8.09.104 2.96.103 -5.35.104 -1.74.105 0 1.24104 -6.09.104 -7.41.104 -9.90.104 0 3.50.104 4.93.104 5.74.104 5.84.104 1.12.105 5.92.104 -7.58.103 5.88.104 2.41.104 -8.63.104 1.01.105 2.01.104 -3.17.104 -3.24.104 -1.04.105 1.58.104 -4.13.104 -1.03.105 -7.84.103 5.10.104 -2.02.104 -7.57.104 -5.78.104 -2.78.103 7.49.104 2.25.104 -4.30.104 -3.06.104 -2.13.104 -4.42.104 9.66.104 1.31.105 1.23.105 5.24.104 -9.59.104 -1.06.105 -1.17.105 -1.51.105 3.86.104 3.19.104 -2.14.105 -1.59.105 3.91.104

∆f G0 (298.15 K) [J/kmol] 6.71.108 0 0 -2.29.108 -1.37.108 -5.08.107 -3.94.108 6.84.107 -3.19.107 0 8.66.107 5.12.107 1.04.108 -1.64.107 0 1.23.108 2.03.108 4.55.108 2.32.108 5.61.108 3.69.108 1.48.108 2.09.108 2.79.108 1.43.108 4.05.108 3.51.108 2.00.108 2.55.108 2.44.108 3.43.107 2.28.107 -1.05.108 2.50.108 2.83.107 -1.10.108 3.89.107 6.43.104 -5.10.107 1.55.108 1.25.108 1.12.108 -1.09.108 -8.39.106 1.74.108 4.39.108 5.26.108 5.05.108 4.09.108 9.48.107 2.63.108 -4.19.107 1.16.108 2.98.108 4.41.108 1.59.108 2.01.108 5.06.108

Table A2.2 Polynomial data for applied species in kinetic scheme of Cp (T); Cp (T) = C1 + C2.T + C3.T2 + C4.T3 + C5.T4 (Cp in [J/kmol.K]; T < 1000 K). Component C (g) H2 O2 H2O CO CH4 CO2 C2H4 C2H6 N2 NO NO2 N2O NH3 AR HCN H N O CH CH2 CH3 C2H2 C2H3 C2H5 CN NH NH2 NNH N2H2 OH HO2 H2O2 H2CN CHO CH2O CH3O CH2OH CH2CO HCCO NCO** HNO HNCO* HOCN* HCNO C4H2 C2H C3H2 CH2 (S) H2NO** HNNO** HONO NO3 C2N2 NCN** N2H4 N2H3 C3H3 * **

C1

2.07739770E+04 2.74215924E+04 2.67133138E+04 2.81592204E+04 2.71249963E+04 6.47469045E+03 1.89210521E+04 -7.16266968E+03 1.21599863E+04 2.74261902E+04 2.80735748E+04 2.22041696E+04 2.11437388E+04 1.83276355E+04 2.07857500E+04 2.01022065E+04 2.07857500E+04 2.08112832E+04 2.44974863E+04 2.66074395E+04 3.12803671E+04 2.02074239E+04 1.67413585E+04 2.04471584E+04 2.23712953E+04 3.04569770E+04 2.77677499E+04 2.85387765E+04 2.91112244E+04 1.34525324E+04 3.02413207E+04 2.47763064E+04 2.81751091E+04 2.37095651E+04 2.40975768E+04 1.37413022E+04 1.75116119E+04 2.38007480E+04 3.24213218E+04 4.19702954E+04 2.79326641E+04 2.31503535E+04 3.20804522E+04 3.15064080E+04 2.64798649E+04 3.33003595E+04 2.27620924E+04 2.63290019E+04 3.30182886E+04 2.10400761E+04 1.86098811E+04 1.90431808E+04 1.01523947E+04 3.54643058E+04 2.57848892E+04 5.35657508E+02 2.63912760E+04 3.15839554E+04

C2

6.72275674E-01 6.85883273E+00 9.37426018E+00 2.88920428E+01 1.25707302E+01 1.45306361E+02 8.24950832E+01 2.32481297E+02 1.28827310E+02 1.17085332E+01 1.04183450E+01 6.51716406E+01 7.89209403E+01 8.40971906E+01 0.00000000E+00 7.50935603E+01 0.00000000E+00 -1.81253237E-01 -1.36202077E+01 1.72345046E+01 9.64308394E+00 9.24890963E+01 1.26297925E+02 6.12886629E+01 7.24934875E+01 -9.61572906E+00 1.04178894E+01 2.74333654E+01 1.70741301E+01 1.08610715E+02 1.53890210E+00 4.15440379E+01 5.46185157E+01 4.73518757E+01 5.15415596E+01 1.05021573E+02 6.00009358E+01 8.32699843E+01 8.06576479E+01 3.70275521E+01 4.48409987E+01 5.49545797E+01 5.31312122E+01 4.47972904E+01 8.10836809E+01 1.64706283E+02 6.69171946E+01 2.06408401E+02 -1.41267274E+00 7.14700138E+01 1.13007941E+02 1.17225170E+02 1.56208819E+02 9.91278154E+01 8.29906321E+01 2.28620718E+02 3.92094656E+01 7.27423262E+01

C3

-2.24294622E-03 -6.77034696E-03 -4.78583579E-03 -5.28348490E-02 -3.22740756E-02 -2.31420974E-01 -8.65444882E-02 -2.81744772E-01 4.80608694E-02 -3.29514167E-02 -2.74600543E-02 -6.70453845E-02 -8.14200692E-02 -1.21826511E-01 0.00000000E+00 -9.20997127E-02 0.00000000E+00 4.50679043E-04 2.01291780E-02 -4.26891997E-02 2.06991566E-03 -1.39698531E-01 -1.34385602E-01 1.75421088E-02 3.67478591E-02 1.79872314E-02 -2.90305840E-02 -5.49874545E-02 5.96169315E-03 -1.42649360E-01 -1.39361353E-02 -3.15194864E-02 -1.23468394E-03 8.90578180E-03 -8.00092073E-02 -1.56987952E-01 4.43856577E-02 -4.39446922E-03 -2.59348708E-03 1.88591770E-03 -6.77165148E-03 -7.73248441E-02 -7.49669419E-03 -5.41948439E-03 -1.06439901E-02 -8.20278611E-02 -7.68599666E-02 -3.81762475E-01 8.52522465E-03 -4.54877847E-02 -9.80981642E-02 -1.13728957E-01 -1.11770898E-01 -1.11579087E-01 -8.24836843E-02 -2.41069054E-01 1.10984855E-01 2.09770953E-02

C4

2.52815331E-06 -7.87816009E-07 1.09239700E-05 5.79388730E-05 4.64099570E-05 2.53561872E-04 5.70916874E-05 2.31565893E-04 -1.04579918E-04 4.69052482E-05 4.33824377E-05 5.12303387E-05 5.20794782E-05 1.20327460E-04 0.00000000E+00 6.63492863E-05 0.00000000E+00 -4.69555081E-07 -1.33265184E-05 4.76732816E-05 7.31727908E-06 1.34843712E-04 7.54854632E-05 -1.09885289E-05 7.76447784E-06 1.54164416E-06 3.50764686E-05 7.14277106E-05 4.09175637E-06 1.33495058E-04 1.98479136E-05 1.95734585E-05 -3.84603305E-05 -1.34908830E-05 9.06113117E-05 1.70445727E-04 -6.13398790E-05 -4.27233548E-05 -4.60423652E-05 -1.23225783E-05 -1.59041584E-05 7.84701971E-05 -1.57824038E-05 -1.18076654E-05 -5.12418956E-05 -5.51666942E-05 5.42529526E-05 3.54855904E-04 2.07238085E-05 1.89254171E-05 4.48387705E-05 6.23471066E-05 1.05974176E-05 7.64275149E-05 3.95670802E-05 1.45104406E-04 -1.59608287E-04 -1.27156659E-05

C5

-9.20103506E-10 3.43785663E-09 -7.29043885E-09 -2.08405246E-08 -2.05774851E-08 -1.01761270E-07 -1.76037011E-08 -8.09636474E-08 3.81315997E-08 -2.03272496E-08 -2.03389561E-08 -1.92904231E-08 -1.58123436E-08 -4.43028224E-08 0.00000000E+00 -1.92155113E-08 0.00000000E+00 1.74592318E-10 3.23484138E-09 -1.62588880E-08 -6.09640643E-09 -4.87629704E-08 -1.59031441E-08 -9.85064795E-09 -3.26566831E-08 -6.82994386E-09 -1.29505025E-08 -2.96990621E-08 -8.04090087E-09 -5.06643344E-08 -7.01015382E-09 -6.72545722E-09 2.05489089E-08 -1.95478172E-09 -3.80369664E-08 -6.99501764E-08 1.72572442E-08 1.86742587E-08 2.05008356E-08 1.87133359E-09 6.51574564E-09 -3.12047818E-08 6.36158604E-09 4.46309047E-09 2.68242182E-08 5.05294349E-08 -1.61262417E-08 -1.23230597E-07 -1.64728399E-08 -3.86454733E-09 -8.40458415E-09 -1.56051512E-08 1.12580619E-08 -2.31049492E-08 -7.45678472E-09 -3.67681792E-08 6.22538451E-08 -1.17278348E-08

valid for T<1400 K valid for T<1500 K

253

Table A2.3 Polynomial data for Cp (T); Cp (T) = C1 + C2.T + C3.T2 + C4.T3 + C5.T4 (Cp in [J/kmol.K]; T > 1000 K). Component C (g) H2 O2 H2O CO CH4 CO2 C2H4 C2H6 N2 NO NO2 N2O NH3 AR HCN H N O CH CH2 CH3 C2H2 C2H3 C2H5 CN NH NH2 NNH N2H2 OH HO2 H2O2 H2CN CHO CH2O CH3O CH2OH CH2CO HCCO NCO** HNO HNCO* HOCN* HCNO C4H2 C2H C3H2 CH2 (S) H2NO** HNNO** HONO NO3 C2N2 NCN** N2H4 N2H3 C3H3 * **

C1

1.23896905E+04 2.48715882E+04 3.07427728E+04 2.22170152E+04 2.51514060E+04 1.39969411E+04 3.70287577E+04 2.93363258E+04 4.01242963E+04 2.43329630E+04 2.69835202E+04 3.89346946E+04 3.92349905E+04 2.04690084E+04 2.07857500E+04 2.84885914E+04 2.07857500E+04 2.03722632E+04 2.11354411E+04 1.82600569E+04 3.02341787E+04 2.36462932E+04 3.68886368E+04 4.93326330E+04 5.97838079E+04 3.09301854E+04 2.29495383E+04 2.46212280E+04 3.67104780E+04 2.80290434E+04 2.39678820E+04 3.38574176E+04 3.80226824E+04 4.33150337E+04 2.95762183E+04 2.49063670E+04 3.13515541E+04 5.26088995E+04 6.09298842E+04 5.61886463E+04 5.04873063E+04 3.00573918E+04 5.44196377E+04 5.00696458E+04 5.56427211E+04 7.50898272E+04 3.31438428E+04 6.37788373E+04 2.95397767E+04 4.71699006E+04 1.86098811E+04 4.56196662E+04 5.92003685E+04 5.44420530E+04 5.53077213E+04 4.13829067E+04 3.69308402E+04 6.35230895E+04

valid for T>1400 K valid for T>1500 K

254

C2

1.38194109E+01 5.82054544E+00 5.10098684E+00 2.54109369E+01 1.19949450E+01 8.51154513E+01 2.61082988E+01 9.54912736E+01 1.15073479E+02 1.23714855E+01 1.05519966E+01 2.04733734E+01 2.38929120E+01 5.03777239E+01 0.00000000E+00 3.26268929E+01 0.00000000E+00 8.86425602E-01 -2.29064037E-01 1.94586297E+01 1.60720075E+01 5.10329572E+01 4.46980011E+01 3.34047373E+01 5.39105614E+01 1.26240224E+00 1.14350417E+01 2.43833393E+01 1.34225053E+01 5.02181059E+01 8.43048652E+00 1.77202343E+01 3.60519355E+01 2.46875762E+01 2.78160893E+01 5.55505072E+01 6.54459875E+01 3.00002393E+01 2.77399635E+01 1.66319257E+01 7.67229387E+00 2.67095640E+01 1.63439101E+01 1.60426913E+01 1.96912555E+01 5.02786673E+01 2.61328676E+01 2.28539238E+01 1.71838955E+01 1.91132122E+01 1.13007941E+02 3.50702495E+01 2.69901135E+01 3.31300494E+01 5.07840271E+00 7.97800236E+01 5.99816051E+01 4.35133547E+01

C3

-5.55994202E-03 -4.68413361E-04 -1.04663900E-03 -7.25860007E-03 -4.68163849E-03 -3.22189767E-02 -1.06290884E-02 -3.67357784E-02 -3.78904102E-02 -4.72648001E-03 -4.17036142E-03 -8.66564985E-03 -9.95633933E-03 -1.66699720E-02 0.00000000E+00 -1.33123400E-02 0.00000000E+00 -6.20690514E-04 -2.57976350E-05 -5.86840006E-03 -1.40263571E-03 -1.85437574E-02 -1.59037261E-02 -3.29806581E-03 -5.34448525E-03 1.65236818E-03 -3.70145486E-03 -7.53574895E-03 -1.35763731E-03 -1.91549250E-02 -1.89306384E-03 -4.41335100E-03 -1.22610051E-02 -2.37422236E-03 -1.10996404E-02 -2.18579122E-02 -2.20859735E-02 -2.66186222E-03 -2.51483048E-03 -1.68581329E-03 -8.18590559E-04 -1.04788199E-02 -1.29924598E-03 -1.20975443E-03 -1.97174456E-03 -1.62028081E-02 -1.05362393E-02 -3.63413231E-03 -1.59145347E-03 -1.47532681E-03 -9.80981642E-02 -1.37114663E-02 -1.19028633E-02 -1.35873654E-02 -1.15546039E-03 -2.94961349E-02 -2.07498655E-02 -4.20174375E-03

C4

1.07327603E-06 -7.67541090E-08 1.47602188E-07 9.98544437E-07 8.46879050E-07 5.64173894E-06 1.99044009E-06 6.52223578E-06 5.59133931E-06 8.39498030E-07 7.62361696E-07 1.64365563E-06 1.87117562E-06 2.60736697E-06 0.00000000E+00 2.62895339E-06 0.00000000E+00 1.56279906E-07 3.78389364E-08 7.48917390E-07 -8.39660658E-07 3.14709641E-06 2.73239409E-06 -1.19831229E-06 -1.95209704E-06 -3.15807960E-07 6.39601722E-07 1.34463640E-06 -7.11691276E-07 3.37792466E-06 1.80809669E-07 5.08192381E-07 1.95294842E-06 -1.35984567E-06 2.05410768E-06 3.93861112E-06 3.27951827E-06 -1.61193491E-06 -1.48086496E-06 -8.65628212E-07 -3.96103479E-07 1.88509874E-06 -8.93219882E-07 -8.69518889E-07 -1.06049171E-06 2.29047574E-06 2.43140313E-06 -5.36737868E-07 -9.18458522E-07 -9.17467873E-07 4.48387705E-05 2.47090686E-06 2.32555378E-06 2.52637987E-06 2.24116031E-08 5.09192592E-06 3.25967453E-06 -2.42758603E-06

C5

-7.65359085E-11 1.31594741E-11 -9.44866485E-12 -5.31418295E-11 -5.74597199E-11 -3.74428098E-10 -1.38768436E-10 -4.37901543E-10 -2.99161900E-10 -5.61493862E-11 -5.21757162E-11 -1.15711510E-10 -1.30978244E-10 -1.61157490E-10 0.00000000E+00 -2.02274448E-10 0.00000000E+00 -8.53033794E-12 -3.63172864E-12 -3.20519591E-11 1.50343745E-10 -2.03879856E-10 -1.79315256E-10 1.97767515E-10 3.22667756E-10 1.10432994E-11 -4.17177652E-11 -1.00120826E-10 1.34258560E-10 -2.25578932E-10 -4.26216377E-12 -2.36222898E-11 -1.19031975E-10 2.53025683E-10 -1.42494631E-10 -2.67098301E-10 -1.75649232E-10 2.91807320E-10 2.69813003E-10 1.63389630E-10 7.55806869E-11 -1.27727253E-10 1.55866436E-10 1.49226220E-10 2.00136592E-10 -1.15203606E-10 -2.25842994E-10 1.38340590E-10 1.68061020E-10 1.54626192E-10 -8.40458415E-09 -1.68044308E-10 -1.67367441E-10 -1.75520610E-10 1.38844146E-11 -3.35049246E-10 -1.91141517E-10 4.52771835E-10

APPENDIX 3 Detailed homogeneous reaction scheme “Kilpinen 97” Table A3.1 Homogeneous reaction rates (A in units mole-cm-s-K, Ea in cal/mole) [Zabetta et al., 2000] REACTIONS CONSIDERED A b Ea --------------------------------------------------------------------------2.20E+16 0.0 93470.0 1. NH3+M=NH2+H+M 2. NH3+H=NH2+H2 6.40E+05 2.4 10171.0 9.40E+06 1.9 6460.0 3. NH3+O=NH2+OH 2.00E+06 2.0 566.0 4. NH3+OH=NH2+H2O 3.00E+11 0.0 22000.0 5. NH3+HO2=NH2+H2O2 4.00E+13 0.0 3650.0 6. NH2+H=NH+H2 6.60E+14 -0.5 0.0 7. NH2+O=HNO+H 6.80E+12 0.0 0.0 8. NH2+O=NH+OH 9. NH2+OH=NH+H2O 4.00E+06 2.0 1000.0 5.00E+13 0.0 0.0 10. NH2+HO2=H2NO+OH 1.00E+13 0.0 0.0 11. NH2+HO2=NH3+O2 12. H2NO+O=NH2+O2 2.00E+14 0.0 0.0 8.50E+11 0.0 0.0 13. NH2+NH2=N2H2+H2 5.00E+13 0.0 10000.0 14. NH2+NH2=NH3+NH 15. NH2+NH2 (+M)=N2H4 (+M) 1.50E+13 0.0 0.0 Low pressure limit: 0.10E+19 0.0 0.0 Enhanced by 2.5 N2 Enhanced by 5.0 H2O Enhanced by 10.0 NH3 5.00E+13 0.0 0.0 16. NH2+NH=N2H2+H 17. NH2+N=N2+2H 7.00E+13 0.0 0.0 8.90E+12 -0.3 0.0 18. NH2+NO=NNH+OH 1.30E+16 -1.3 0.0 19. NH2+NO=N2+H2O Declared duplicate reaction… 8.90E+12 -0.3 0.0 20. NH2+NO=N2+H2O Declared duplicate reaction... 3.20E+18 -2.2 0.0 21. NH2+NO2=N2O+H2O 3.50E+12 0.0 0.0 22. NH2+NO2=H2NO+NO 3.00E+13 0.0 0.0 23. NH+H=N+H2 24. NH+O=NO+H 9.20E+13 0.0 0.0 25. NH+OH=HNO+H 2.00E+13 0.0 0.0 5.00E+11 0.5 2000.0 26. NH+OH=N+H2O 4.60E+05 2.0 6500.0 27. NH+O2=HNO+O 28. NH+O2=NO+OH 1.30E+06 1.5 100.0 3.00E+13 0.0 0.0 29. NH+N=N2+H 2.50E+13 0.0 0.0 30. NH+NH=N2+2H 31. NH+NO=N2O+H 2.90E+14 -0.4 0.0 Declared duplicate reaction... -2.20E+13 -0.2 0.0 32. NH+NO=N2O+H Declared duplicate reaction... 2.20E+13 -0.2 0.0 33. NH+NO=N2+OH 1.00E+13 0.0 0.0 34. NH+NO2=N2O+OH 35. N+OH=NO+H 3.80E+13 0.0 0.0 6.40E+09 1.0 6280.0 36. N+O2=NO+O 3.30E+12 0.3 0.0 37. N+NO=N2+O 38. NO+O+M=NO2+M 7.50E+19 -1.4 0.0 Enhanced by 1.7 N2 Enhanced by 1.5 O2 H2O Enhanced by 10.0 39. NO+OH+M=HONO+M 5.00E+23 -2.5 -68.0 Enhanced by 1.0 N2 H2O Enhanced by 5.0 2.10E+12 0.0 -480.0 40. NO+HO2=NO2+OH 8.40E+13 0.0 0.0 41. NO2+H=NO+OH 42. NO2+O=NO+O2 3.90E+12 0.0 -238.0 1.30E+13 0.0 0.0 43. NO2+O(+M)=NO3(+M) Low pressure limit: 0.1E+29 – 4.08 2470.0 Enhanced by 1.5 N2 O2 Enhanced by 1.5 Enhanced by 18.6 H2O 1.60E+12 0.0 26123.0 44. NO2+NO2=NO+NO+O2 9.60E+09 0.7 20900.0 45. NO2+NO2=NO3+NO 46. HNO+M=H+NO+M 1.50E+16 0.0 48680.0 Enhanced by 10.0 H2O O2 Enhanced by 2.0 Enhanced by 2.0 N2 Enhanced by 2.0 H2

4.40E+11 47. HNO+H=NO+H2 48. HNO+O=NO+OH 1.00E+13 3.60E+13 49. HNO+OH=NO+H2O 1.00E+13 50. HNO+O2=NO+HO2 2.00E+13 51. HNO+NH2=NO+NH3 2.00E+12 52. HNO+NO=N2O+OH 6.00E+11 53. HNO+NO2=HONO+NO 4.00E+12 54. HNO+HNO=N2O+H2O 55. HONO+H=NO2+H2 1.20E+13 1.20E+13 56. HONO+O=NO2+OH 4.00E+12 57. HONO+OH=NO2+H2O 58. HONO+NH=NH2+NO2 1.00E+13 5.00E+12 59. HONO+NH2=NH3+NO2 2.30E+12 60. 2 HONO=NO+NO2+H2O 61. H2NO+M=HNO+H+M 2.50E+16 3.00E+07 62. H2NO+H=HNO+H2 5.00E+13 63. H2NO+H=NH2+OH 3.00E+07 64. H2NO+O=HNO+OH 2.00E+07 65. H2NO+OH=HNO+H2O 2.00E+07 66. H2NO+NO=HNO+HNO 67. H2NO+NH2=HNO+NH3 3.00E+12 68. H2NO+NO2=HONO+HNO 6.00E+11 6.00E+13 69. NO3+H=NO2+OH 70. NO3+O=NO2+O2 1.00E+13 1.40E+13 71. NO3+OH=NO2+HO2 1.50E+12 72. NO3+HO2=NO2+O2+OH 73. NO3+NO2=NO+NO2+O2 5.00E+10 1.30E+13 74. N2H4+H=N2H3+H2 8.50E+13 75. N2H4+O=N2H2+H2O 76. N2H4+OH=N2H3+H2O 4.00E+13 3.90E+12 77. N2H4+NH2=N2H3+NH3 3.50E+16 78. N2H3+M=N2H2+H+M 1.60E+12 79. N2H3+H=NH2+NH2 80. N2H3+O=N2H2+OH 5.00E+12 1.00E+13 81. N2H3+O=NH2+HNO 1.00E+13 82. N2H3+OH=N2H2+H2O 83. N2H3+OH=NH3+HNO 1.00E+12 2.00E+13 84. N2H3+NH=N2H2+NH2 5.00E+16 85. N2H2+M=NNH+H+M Enhanced by 15.0 H2O Enhanced by 2.0 H2 Enhanced by 2.0 N2 Enhanced by 2.0 O2 5.00E+13 86. N2H2+H=NNH+H2 1.00E+13 87. N2H2+O=NH2+NO 88. N2H2+O=NNH+OH 2.00E+13 1.00E+13 89. N2H2+OH=NNH+H2O 1.00E+13 90. N2H2+NH=NNH+NH2 91. N2H2+NH2=NNH+NH3 1.00E+13 3.00E+12 92. N2H2+NO=N2O+NH2 1.00E+07 93. NNH=N2+H 94. NNH+H=N2+H2 1.00E+14 1.00E+14 95. NNH+O=N2O+H 96. NNH+O=NH+NO 5.00E+13 5.00E+13 97. NNH+OH=N2+H2O 2.00E+14 98. NNH+O2=N2+HO2 5.00E+13 99. NNH+O2=N2+H+O2 5.00E+13 100. NNH+NH=N2+NH2 101. NNH+NH2=N2+NH3 5.00E+13 5.00E+13 102. NNH+NO=N2+HNO 4.00E+14 103. N2O+M=N2+O+M Enhanced by 1.7 N2 Enhanced by 1.4 O2 Enhanced by 3.0 CO2 H2O Enhanced by 12.0 3.30E+10 104. N2O+H=N2+OH Declared duplicate reaction...

0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 2.0 0.0 2.0 2.0 2.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

650.0 0.0 0.0 25000.0 1000.0 26000.0 2000.0 5000.0 7350.0 6000.0 0.0 0.0 0.0 8400.0 50000.0 2000.0 0.0 2000.0 1000.0 13000.0 1000.0 2000.0 0.0 0.0 0.0 0.0 2940.0 2500.0 1200.0 0.0 1500.0 46000.0 0.0 5000.0 0.0 1000.0 15000.0 0.0 50000.0

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1000.0 1000.0 1000.0 1000.0 1000.0 1000.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 56100.0

0.0

4729.0

255

REACTIONS CONSIDERED (continued) A b Ea --------------------------------------------------------------------------4.40E+14 0.0 19254.0 105. N2O+H=N2+OH Declared duplicate reaction... 2.90E+13 0.0 23150.0 106. N2O+O=NO+NO 1.40E+12 0.0 10800.0 107. N2O+O=N2+O2 2.00E+12 0.0 40000.0 108. N2O+OH=N2+HO2 2.00E+14 -0.4 0.0 109. O+OH=H+O2 110. O+H2=OH+H 5.10E+04 2.7 6290.0 2.10E+08 1.5 3450.0 111. OH+H2=H2O+H 4.30E+03 2.7 -2486.0 112. OH+OH=H2O+O 113. H+OH+M=H2O+M 8.40E+21 -2.0 0.0 Enhanced by 2.6 N2 16.5 H2O Enhanced by 1.90E+13 0.0 -1788.0 114. O+O+M=O2+M N2 Enhanced by 1.5 1.00E+18 -1.0 0.0 115. H+H+M=H2+M Enhanced by 0.0 H2 H2O Enhanced by 0.0 9.20E+16 -0.6 0.0 116. H+H+H2=H2+H2 6.00E+19 -1.3 0.0 117. H+H+H2O=H2+H2O 118. H+O+M=OH+M 6.20E+16 -0.6 0.0 Enhanced by 1.5 N2 2.10E+18 -1.0 0.0 119. H+O2+M=HO2+M Enhanced by 0.0 N2 Enhanced by 1.5 O2 Enhanced by 10.0 H2O 120. H+O2+N2=HO2+N2 6.70E+19 -1.4 0.0 4.30E+13 0.0 1411.0 121. HO2+H=H2+O2 1.70E+14 0.0 875.0 122. HO2+H=OH+OH 123. HO2+H=O+H2O 3.00E+13 0.0 1721.0 3.30E+13 0.0 0.0 124. HO2+O=OH+O2 2.90E+13 0.0 -497.0 125. HO2+OH=H2O+O2 126. HO2+HO2=H2O2+O2 1.30E+11 0.0 -1630.0 Declared duplicate reaction... 4.20E+14 0.0 11980.0 127. HO2+HO2=H2O2+O2 Declared duplicate reaction... 1.30E+17 0.0 45500.0 128. H2O2+M=OH+OH+M Enhanced by 1.5 N2 Enhanced by 1.5 O2 H2O Enhanced by 10.0 1.00E+13 0.0 3576.0 129. H2O2+H=H2O+OH 1.70E+12 0.0 3755.0 130. H2O2+H=HO2+H2 131. H2O2+O=HO2+OH 6.60E+11 0.0 3974.0 7.80E+12 0.0 1330.0 132. H2O2+OH=H2O+HO2 Declared duplicate reaction... 5.80E+14 0.0 9560.0 133. H2O2+OH=H2O+HO2 Declared duplicate reaction... 6.20E+14 0.0 3000.0 134. CO+O+M=CO2+M Enhanced by 1.5 N2 Enhanced by 1.5 O2 Enhanced by 16.0 H2O 135. CO+OH=CO2+H 1.40E+05 1.9 -1347.0 6.00E+13 0.0 22950.0 136. CO+HO2=CO2+OH 2.50E+12 0.0 47700.0 137. CO+O2=CO2+O 138. CN+H2=HCN+H 3.60E+08 1.6 3000.0 139. HCN+O=NCO+H 1.40E+04 2.6 4980.0 140. HCN+O=CN+OH 2.70E+09 1.6 29200.0 141. HCN+O=NH+CO 3.50E+03 2.6 4980.0 8.00E+12 0.0 7450.0 142. CN+H2O=HCN+OH 143. HCN+OH=HOCN+H 5.90E+04 2.4 12500.0 144. HCN+OH=HNCO+H 2.00E-03 4.0 1000.0 7.80E-04 4.0 4000.0 145. HCN+OH=NH2+CO 1.50E+07 1.7 153.0 146. HCN+CN=C2N2+H 4.60E+12 0.0 8880.0 147. C2N2+O=CN+NCO 148. C2N2+OH=CN+HOCN 1.90E+11 0.0 2900.0 149. NCN+H=HCN+N 1.00E+14 0.0 0.0 150. NCN+O=CN+NO 1.00E+14 0.0 0.0 151. NCN+OH=HCN+NO 5.00E+13 0.0 0.0 1.00E+13 0.0 0.0 152. NCN+O2=NO+NCO 153. CN+O=CO+N 7.70E+13 0.0 0.0 154. CN+OH=NCO+H 6.00E+13 0.0 0.0 7.50E+12 0.0 -389.0 155. CN+O2=NCO+O 3.70E+06 2.2 26900.0 156. CN+CO2=NCO+CO 157. CN+NO=NCO+N 1.00E+14 0.0 42100.0

256

2.40E+13 158. CN+NO2=NCO+NO 159. CN+HNO=HCN+NO 1.80E+13 1.20E+13 160. CN+HONO=HCN+NO2 3.80E+03 161. CN+N2O=NCN+NO 162. HOCN+H=HNCO+H 2.00E+07 6.40E+05 163. HOCN+OH=NCO+H2O 164. HOCN+O=NCO+OH 1.50E+04 165. HNCO+M=CO+NH+M 1.10E+16 Enhanced by 1.5 N2 166. HNCO+H=NH2+CO 2.20E+07 167. HNCO+O=NCO+OH 2.20E+06 9.60E+07 168. HNCO+O=NH+CO2 169. HNCO+O=HNO+CO 1.50E+08 6.40E+05 170. HNCO+OH=NCO+H2O 171. HNCO+HO2=NCO+H2O2 3.00E+11 172. HNCO+NH2=NH3+NCO 5.00E+12 173. HNCO+NH=NH2+NCO 3.00E+13 174. HNCO+NO2=HNNO+CO2 2.50E+12 175. HNCO+CN=HCN+NCO 1.50E+13 176. NCO+M=N+CO+M 3.10E+16 Enhanced by 1.5 N2 177. NCO+H=CO+NH 5.00E+13 178. NCO+O=NO+CO 4.70E+13 7.60E+02 179. NCO+H2=HNCO+H 180. NCO+OH=HCO+NO 5.00E+12 2.00E+12 181. NCO+O2=NO+CO2 182. NCO+HCO=HNCO+CO 3.60E+13 183. NCO+CH2O=HNCO+HCO 6.00E+12 184. NCO+N=N2+CO 2.00E+13 6.20E+17 185. NCO+NO=N2O+CO 7.80E+17 186. NCO+NO=N2+CO2 187. NCO+NO2=CO+2 NO 1.30E+13 5.40E+12 188. NCO+NO2=CO2+N2O 189. NCO+HNO=HNCO+NO 1.80E+13 190. NCO+HONO=HNCO+NO2 3.60E+12 1.80E+13 191 2 NCO=2 CO+N2 192. NCO+CN=NCN+CO 1.80E+13 2.20E+15 193. HNNO+M=N2O+H+M 194. HNNO+M=N2+OH+M 1.00E+15 2.00E+13 195. HNNO+H=N2O+H2 196. HNNO+H=NNH+OH 1.00E+13 2.00E+13 197. HNNO+O=N2O+OH 1.00E+13 198. HNNO+O=NNH+O2 2.00E+13 199. HNNO+OH=N2O+H2O 1.00E+13 200. HNNO+OH=NNH+HO2 1.00E+12 201. HNNO+NO=N2O+HNO 3.20E+12 202. HNNO+NO=NNH+NO2 203. HNNO+NO2=NNH+NO3 1.00E+13 204. HNNO+NO2=N2O+HONO 1.00E+12 205. HCO+M=H+CO+M 1.90E+17 Enhanced by 1.5 N2 Enhanced by 1.5 O2 CO Enhanced by 1.9 Enhanced by 3.0 CO2 Enhanced by 5.0 H2O 1.20E+13 206. HCO+H=CO+H2 207. HCO+O=CO+OH 3.00E+13 3.00E+13 208. HCO+O=CO2+H 1.10E+14 209. HCO+OH=CO+H2O 210. HCO+O2=CO+HO2 7.60E+12 3.30E+16 211. CH2O+M=HCO+H+M Enhanced by 1.5 N2 Enhanced by 1.5 O2 H2O Enhanced by 10.0 2.20E+08 212. CH2O+H=HCO+H2 1.80E+13 213. CH2O+O=HCO+OH 3.40E+09 214. CH2O+OH=HCO+H2O 215. CH2O+HO2=HCO+H2O2 2.00E+12 2.10E+13 216. CH2O+O2=HCO+HO2

0.0 0.0 0.0 2.6 2.0 2.0 2.6 0.0

-370.0 0.0 0.0 3700.0 2000.0 2560.0 4000.0 86000.0

1.7 2.1 1.4 1.6 2.0 0.0 0.0 0.0 0.0 0.0 -0.5

3800.0 11430.0 8520.0 44012.0 2560.0 29000.0 6200.0 23700.0 26200.0 0.0 48000.0

0.0 0.0 3.0 0.0 0.0 0.0 0.0 0.0 -1.7 -1.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 -1.0

0.0 0.0 4000.0 15000.0 20000.0 0.0 0.0 0.0 763.0 763.0 0.0 0.0 0.0 0.0 0.0 0.0 21600.0 25600.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 270.0 0.0 0.0 17020.0

0.3 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 400.0 81000.0

1.8 0.0 1.2 0.0 0.0

3000.0 3080.0 -447.0 11665.0 38950.0

REACTIONS CONSIDERED (continued) A b Ea --------------------------------------------------------------------------6.00E+16 -1.0 0.0 217. CH3+H (+M)=CH4 (+M) Low pressure limit: 0.80E+27 –3.0 0.0 SRI centering: 0.45 797 979 Enhanced by 2.0 H2 CO Enhanced by 2.0 Enhanced by 3.0 CO2 H2O Enhanced by 5.0 1.30E+04 3.0 8047.0 218. CH4+H=CH3+H2 6.90E+08 1.6 8484.0 219. CH4+O=CH3+OH 220. CH4+OH=CH3+H2O 1.60E+07 1.8 2782.0 4.00E+13 0.0 56908.0 221. CH4+O2=CH3+HO2 1.80E+11 0.0 18678.0 222. CH4+HO2=CH3+H2O2 4.30E+12 0.0 10034.0 223. CH4+CH2=CH3+CH3 224. CH4+CH2*=CH3+CH3 4.30E+13 0.0 0.0 6.00E+13 0.0 0.0 225. CH4+CH=C2H4+H 1.00E+16 0.0 90607.0 226. CH3+M=CH2+H+M 227. CH2*+H2=CH3+H 7.20E+13 0.0 0.0 8.40E+13 0.0 0.0 228. CH3+O=CH2O+H 5.00E+13 0.0 0.0 229. CH3+OH=CH2*+H2O 230. CH3+O2=CH3O+O 1.10E+13 0.0 27818.0 3.30E+11 0.0 9001.0 231. CH3+O2=CH2O+OH 2.00E+13 0.0 0.0 232. CH3+HO2=CH3O+OH 5.50E+03 2.8 5862.0 233. CH3+CH2O=CH4+HCO 1.20E+14 0.0 0.0 234. CH3+HCO=CH4+CO 4.20E+13 0.0 0.0 235. CH3+CH2=C2H4+H 236. CH3+CH2*=C2H4+H 2.00E+13 0.0 0.0 3.00E+13 0.0 0.0 237. CH3+CH=C2H3+H 5.00E+13 0.0 0.0 238. CH3+C=C2H2+H 239. CH3O+M=CH2O+H+M 1.90E+26 -2.7 30600.0 2.00E+13 0.0 0.0 240. CH3O+H=CH2O+H2 1.50E+13 0.0 0.0 241. CH3O+O=CH2O+OH 242. CH3O+OH=CH2O+H2O 1.00E+13 0.0 0.0 4.00E+10 0.0 2126.0 243. CH3O+O2=CH2O+HO2 244. CH2OH+M=CH2O+H+M 1.10E+43 -8.0 42999.0 245. CH2OH+H=CH3+OH 1.00E+14 0.0 0.0 6.00E+12 0.0 0.0 246. CH2OH+H=CH2O+H2 5.00E+13 0.0 0.0 247. CH2OH+O=CH2O+OH 0.0 0.0 248. CH2OH+OH=CH2O+H2O 1.00E+13 249. CH2OH+O2=CH2O+HO2 2.20E+14 0.0 4709.0 8.30E+15 0.0 69545.0 250. CH2O+M=CO+H2+M 9.50E+13 0.0 -517.0 251. CH2O+CH=CH2CO+H 252. HCO+CH2=CH3+CO 2.00E+13 0.0 0.0 6.00E+12 0.0 -1788.0 253. CH2+H=CH+H2 7.00E+13 0.0 0.0 254. CH2+O=CO+H+H 5.00E+13 0.0 0.0 255. CH2+O=CO+H2 3.00E+13 0.0 0.0 256. CH2+OH=CH2O+H 1.10E+07 2.0 2981.0 257. CH2+OH=CH+H2O 5.00E+13 0.0 8941.0 258. CH2+O2=CH2O+O 8.00E+12 0.0 1490.0 259. CH2+O2=CO+H2O 1.70E+13 0.0 1490.0 260. CH2+O2=CO+OH+H 261. CH2+CO2=CO+CH2O 1.10E+11 0.0 994.0 1.20E+14 0.0 0.0 262. CH2+CH2=C2H2+2H 4.00E+13 0.0 0.0 263. CH2+CH=C2H2+H 264. CH2+C=C2H+H 5.00E+13 0.0 0.0 1.00E+13 0.0 0.0 265. CH2*+M=CH2+M H Enhanced by 20.0 Enhanced by 3.0 H2O 4.0 C2H2 Enhanced by 3.00E+13 0.0 0.0 266. CH2*+H=CH+H2 267. CH2*+O=CO+2H 3.00E+13 0.0 0.0 3.00E+13 0.0 0.0 268. CH2*+OH=CH2O+H 3.10E+13 0.0 0.0 269. CH2*+O2=CO+OH+H 6.60E+12 0.0 0.0 270. CH2*+CO2=CH2O+CO 271. CH+H=C+H2 1.50E+14 0.0 0.0 272. CH+O=CO+H 6.00E+13 0.0 0.0 273. CH+OH=HCO+H 3.00E+13 0.0 0.0 4.00E+07 2.0 2980.0 274. CH+OH=C+H2O 3.30E+13 0.0 0.0 275. CH+O2=HCO+O 5.70E+12 0.0 -755.0 276. CH+H2O=CH2O+H 277. CH+CO2=HCO+CO 3.40E+12 0.0 686.0 278. C+OH=CO+H 5.00E+13 0.0 0.0 2.00E+13 0.0 0.0 279. C+O2=CO+O

3.60E+13 0.0 0.0 280. 2 CH3 (+M)=C2H6 (+M) Low pressure limit: 0.32E+42 –7.03 2762.0 TROE centering: 0.38 1180.0 73.0 1.40E+09 1.5 7411.0 281. C2H6+H=C2H5 + H2 2.70E+06 2.4 5842.0 282. C2H6+O=C2H5+OH 7.20E+06 2.0 854.0 283. C2H6+OH=C2H5+H2O 284. C2H6+O2=C2H5+HO2 6.00E+13 0.0 51861.0 1.50E-07 6.0 6040.0 285. C2H6+CH3=C2H5+CH4 0.0 2066.0 286. C2H4+H (+M)=C2H5 (+M) 2.20E+13 Low pressure limit: 0.64E+28 –2.6 54.0 Enhanced by 2.0 H2 CO Enhanced by 2.0 Enhanced by 3.0 CO2 H2O Enhanced by 5.0 1.00E+14 0.0 0.0 287. C2H5+H=CH3+CH3 8.40E+11 0.0 3875.0 288. C2H5+O2=C2H4+HO2 289. C2H4+M=C2H2+H2+M 1.50E+15 0.0 55437.0 1.40E+16 0.0 81268.0 290. C2H4+M=C2H3+H+M 5.40E+14 0.0 15002.0 291. C2H4+H=C2H3+H2 292. C2H4+OH=C2H3+H2O 1.20E+14 0.0 6140.0 6.60E+00 3.7 9538.0 293. C2H4+CH3=C2H3+CH4 0.0 2404.0 294. C2H2+H (+M)=C2H3 (+M) 5.50E+12 Low pressure limit: 0.27E+28 0.0 2404.0 Enhanced by 2.0 H2 CO Enhanced by 2.0 Enhanced by 3.0 CO2 Enhanced by 5.0 H2O 3.00E+13 0.0 0.0 295. C2H3+H=C2H2+H2 296. C2H3+O=CH2CO+H 3.30E+13 0.0 0.0 3.00E+13 0.0 0.0 297. C2H3+OH=C2H2+H2O 5.40E+12 0.0 0.0 298. C2H3+O2=C2H2+HO2 299. C2H2+M=C2H+H+M 4.00E+16 0.0 106801.0 1.50E+13 0.0 3100.0 300. C2H+H2=C2H2+H 7.00E+03 2.8 497.0 301. C2H2+O=CH2+CO 302. C2H2+O=HCCO+H 1.50E+04 2.8 497.0 2.20E-04 4.5 -994.0 303. C2H2+OH=CH2CO+H 3.40E+07 2.0 13909.0 304. C2H2+OH=C2H+H2O 1.20E+13 0.0 6557.0 305. C2H2+CH2=C3H3+H 306. C2H2+CH2*=C3H3 + H 1.70E+14 0.0 0.0 8.40E+13 0.0 0.0 307. C2H2+CH=C3H2+H 4.00E+13 0.0 0.0 308. C2H2+C2H=C4H2+H 309. CH2CO(+M)=CH2+CO(+M) 3.00E+14 0.0 70936.0 Low pressure limit: 0.36E+16 0.0 59272.0 3.60E+12 0.0 2345.0 310. CH2CO+H=CH3+CO 2.50E+11 0.0 1351.0 311. CH2CO+O=CH2O+CO 1.50E+12 0.0 1351.0 312. CH2CO+O=CH2+CO2 0.0 0.0 313. CH2CO+OH=CH2O+HCO 1.00E+13 1.50E+14 0.0 0.0 314. HCCO+H=CH2*+CO 315. HCCO+O=CO+CO+H 9.60E+13 0.0 596.0 316. HCCO+OH=HCO+CO+H 1.00E+13 0.0 0.0 1.60E+12 0.0 854.0 317. HCCO+O2=2 CO+OH 3.00E+13 0.0 0.0 318. HCCO+CH2=C2H3+CO 5.00E+13 0.0 0.0 319. HCCO+CH=C2H2+CO 320. 2 HCCO=C2H2+2CO 1.00E+13 0.0 0.0 1.00E+13 0.0 0.0 321. C2H+O=CH+CO 2.00E+13 0.0 0.0 322. C2H+OH=HCCO+H 323. C2H+O2=HCCO+O 2.30E+13 0.0 0.0 2.40E+12 0.0 0.0 324. C2H+O2=HCO+CO 2.70E+13 0.0 1709.0 325. C4H2+O=C3H2+CO 326. C4H2+OH=C3H2+HCO 3.00E+13 0.0 0.0 5.00E+13 0.0 2981.0 327. C3H3+H=C3H2+H2 1.40E+14 0.0 0.0 328. C3H3+O=CH2O+C2H 2.00E+13 0.0 0.0 329. C3H3+OH=C3H2+H2O 330. C3H3+O2=CH2CO+HCO 3.00E+10 0.0 2861.0 5.00E+13 0.0 0.0 331. C3H2+OH=C2H2+HCO 5.00E+13 0.0 0.0 332. C3H2+O2=HCCO+HCO 3.00E+14 0.0 21857.0 333. H2CN+M=HCN+H+M 7.10E+13 0.0 0.0 334. CH3+N=H2CN+H 5.30E+11 0.0 14902.0 335. CH3+NO=HCN+H2O

257

REACTIONS CONSIDERED (continued) A b Ea --------------------------------------------------------------------------5.30E+11 0.0 14902.0 336. CH3+NO=H2CN+OH 5.00E+13 0.0 0.0 337. CH2+N=HCN+H 1.00E+13 0.0 73519.0 338. CH2+N2=HCN+NH 3.50E+12 0.0 -1093.0 339. CH2+NO=NCO+H2 340. HCNO+H=HCN+OH 1.00E+14 0.0 11915.0 1.00E+14 0.0 0.0 341. CH2*+NO=HCN+OH 342. CH+N=CN+H 1.30E+13 0.0 0.0 4.40E+12 0.0 21897.0 343. CH+N2=HCN+N 344. CH+NO=HCN+O 1.10E+14 0.0 0.0 9.60E+12 0.0 -994.0 345. CH+N2O=HCN+NO 6.30E+13 0.0 45999.0 346. C+N2=CN+N 347. C+NO=CN+O 1.90E+13 0.0 0.0

258

348. C+NO=CO+N 349. C2H3+N=HCN+CH2 350. HCCO+N=HCN+CO 351. HCCO+NO=HCNO+CO 352. C2H+NO=HCN+CO 353. C3H3+N=HCN+C2H2

2.90E+13 2.00E+13 5.00E+13 2.00E+13 2.10E+13 1.00E+14

0.0 0.0 0.0 0.0 0.0 0.0

0.0 0.0 0.0 0.0 0.0 0.0

Reference: Coda Zabetta, E.G., Kilpinen, P., Hupa, M., Ståhl,

K., Leppälahti, J., Cannon, M. and Nieminen, J. (2000) Kinetic Modeling Study on the Potential of Staged Combustion in Gas Turbines for the Reduction of Nitrogen Oxide Emissions from Biomass IGCC Plants, Energy & Fuels, 14, pp. 751-761.

APPENDIX 4 Results of TG-FTIR measurements and comparison of FG-DVC model results with experiments at different heating rates Water

800

80

600

60

400

40 Temperature

200

20

Weight loss

0 0

50

100

0 200

150

Rate (%/min)

100

1.2 1 0.8 0.6 0.4

Rate (%/min)

0.2

Yield (wt%)

Rate model (% ar/min) Yield model (wt%)

0 25

45

65

Time (min)

Time (min)

7.0

0.5

6.0

0.4

4.0

0.3

3.0

0.2

Rate (%/min) Rate model (% ar/min) Yield (wt%) Yield model (wt%)

0.1 0

2.0 1.0 0.0

25

45

65

Time (min)

85

105

1.4

0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

1.2 1 0.8 0.6 Rate model (%/min) Yield (wt%)

0.2

Yield model (wt%)

0 25

45

65

Time (min)

85

105

Ethylene

Carbon Monoxide

10

0.025

0.4

8

0.02

0.3

6

0.5

0.4

Rate (%/min)

Yield (wt%)

5.0

105

Methane Rate (%/min)

0.6

Yield (wt%)

Rate (%/min)

Carbon Dioxide

85

16 14 12 10 8 6 4 2 0

Yield (wt%)

1000

1.4

Char mass(%)

T (°C)

Temperature & Weight Loss

0.25

Rate model (%/min) Yield (wt%)

0.1

2

Yield model (wt%)

0

0 25

45

65

Time (min)

85

105

Rate (%/min)

4

Rate (%/min)

0.2

Yield (wt%) Yield model (wt%)

0.015

0.15

0.01

0.1

0.005

0.05

0

Yield (wt%)

0.2

Yield (wt%)

Rate (%/min)

Rate (%/min) Rate model (% ar/min)

0 25

45

65

85

105

Time (min)

Figure A4.1a Weight loss curve and main species. Comparison between TG-FTIR measurements and FGDVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate was 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

259

Formaldehyde

0.06

1

0.4

0.05

0.8

0.04

0.6

0.03

0.4

Rate (%/min)

0.02

Rate model (% ar/min) Yield (wt%)

0.01

0.2

Yield model (wt%)

0

Rate (%/min)

0.5

0.3 0.2 Rate (%/min)

45

65

85

Yield (wt%) Yield model (wt%)

0 25

Rate model (% ar/min)

0.1 0

105

25

45

65

6

1

4 Rate (%/min)

0.5

Rate model (% ar/min)

2

Yield (wt%)

Yield model (wt%)

0

0 45

65

85

3

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

2.5 2 1.5 Rate model (% ar/min) Yield model (wt%)

0 45

Time (min)

Acetic Acid

3.5 2.5

0.2

2

0.15

1.5

0.1

Rate (%/min)

1

Rate model (% ar/min) Yield (wt%)

0.05

0.5

Yield model (wt%)

0

0 25

45

65

85

105

Time (min)

85

105

Acetone

0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0

2.5 2 1.5 1

Rate (%/min) Rate model (% ar/min)

0.5

Yield (wt%) Yield model (wt%)

0 25

Time (min)

65

45

65

Time (min)

85

105

Phenol 0.1

Rate (%/min)

Rate model (% ar/min)

0.08

Yield (wt%) Yield model (wt%)

0.06 0.04 0.02 0 25

45

65

Time (min)

85

Yield (wt%)

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Rate (%/min)

105

Figure A4.1b Oxygenated hydrocarbons. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

260

Yield (wt%)

0.25

Rate (%/min)

3

Yield (wt%)

Rate (%/min)

0.3

0.5

Yield (wt%)

25

105

1

Rate (%/min)

Yield (wt%)

8

1.5

Rate (%/min)

10

Yield (wt%)

Rate (%/min)

Formic Acid

Acetaldehyde

25

105

Time (min)

Time (min)

2

85

4 3.5 3 2.5 2 1.5 1 0.5 0

Yield (wt%)

1.2

Yield (wt%)

Rate (%/min)

Methanol 0.07

Yield (wt%) Yield model (wt%)

0.005

0.08

0.004

0.004

0.06

0.003

0.04

0.002

0.02

0.001 0 45

65

85

Rate (%/min)

0.12 0.1

0.003

0.08

Rate (%/min) Rate model (% ar/min) Yield (wt%)

0.002

Yield model (wt%)

0.06 0.04

0.001

0.02 0 25

105

Time (min)

45

65

85

105

Time (min)

Ammonia

0.0018 0.0016 0.0014 0.0012 0.001 0.0008 0.0006 0.0004 0.0002 0

Isocyanic Acid (HNCO)

0

0 25

Rate (%/min)

Rate model (% ar/min)

0.005

Yield (wt%)

Rate (%/min)

Rate (%/min)

0.006

0.1

Yield (wt%)

Hydrogen Cyanide 0.007

0.0012

Rate (%/min)

0.001

Yield model (wt%) Rate model (% ar/min) Yield (wt%)

0.0008 0.0006 0.0004 0.0002 0

25

45

65

85

105

Time (min)

Figure A4.1c Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

261

Water

800

100 80

600

60

400

40

200

20

Temperature Weight Loss

0 0

20

40

60

80

12

Rate (%/min)

120

10 8 6 4

0 100

1

Yield model (wt%)

Rate (%/min)

Yield (wt%)

0 33

1.2 1 0.8 0.6 Rate model (%/min)

Yield (wt%) Yield model (wt%)

0.5 0

Time (min)

33

35

Rate (%/min)

Rate (%/min)

Rate model (%/min)

31

0.2

Yield (wt%) Yield model (wt%)

0 27

29

31

33

35

0.25

0.25

0.2

0.2

0.15

0.15

0.1

0.1 Rate (%/min)

0.05

0.05

Yield model (wt%) Rate model (%/min) Yield (wt%)

0 25

27

29

31

33

Yield (wt%)

Rate (%/min)

29

0.4

Rate (%/min)

Time (min)

Yield (wt%)

8 7 6 5 4 3 2 1 0

2

27

35

Ethylene

3

25

33

1.4

25

35

2.5

1

31

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Carbon Monoxide

1.5

29

Yield (wt%)

2

Rate model (%/min)

Yield (wt%)

Rate (%/min)

6

Rate (%/min)

31

27

Methane

3

29

Yield model (wt%)

Time (min)

4

Time (min)

Yield (wt%)

25

5

27

Rate model (%/min)

0

Carbon Dioxide

25

Rate (%/min)

2

Time (min)

4.5 4 3.5 3 2.5 2 1.5 1 0.5 0

16 14 12 10 8 6 4 2 0

Yield (wt%)

1000

Char mass(%)

T (°C)

Temperature & Weight Loss

0 35

Time (min)

Figure A4.2a Weight loss curve and main species. Comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

262

Methanol

1

0.4

0.6

0.3

0.4

0.2

Rate (%/min)

0.1

0.2

Rate model (%/min) Yield (wt%) Yield model (wt%)

0 25

27

29

31

0

33

Rate model (%/min)

0.5

Yield model (wt%)

4 Rate (%/min)

2

Rate model (%/min) Yield (wt%) Yield model (wt%)

0 31

33

27

29

1.5 1

0.5

Rate model (%/min) Yield model (wt%)

0

0 25

27

29

33

35

Time (min)

2

1.5

1

1 Rate (%/min) Rate model (%/min)

0.5

Yield (wt%) Yield model (wt%)

0

Rate (%/min)

1.5

1.5

1.5

1

1

Rate (%/min) Rate model (%/min) Yield (wt%) Yield model (wt%)

0.5

0.5

0

0 33

2

0 25

35

Yield (wt%)

2

Yield (wt%)

Rate (%/min)

31

Acetone

2

31

0.5

Rate (%/min) Yield (wt%)

2.5

29

2

1

35

2.5

27

0 35

1.5

Acetic Acid

25

33

Time (min)

Time (min)

0.5

31

Formic Acid

2

Rate (%/min)

6

0.5

Yield (wt%)

Yield (wt%)

8 7 6 5 4 3 2 1 0

1

Rate (%/min)

0

Yield (wt%)

Rate (%/min)

1.5

1

25

8

29

2

1.5

Acetaldehyde

27

2.5

2

Time (min)

25

3

2.5

35

10

3.5

3

Yield (wt%)

Rate (%/min)

0.8

Yield (wt%)

0.5

Formaldehyde

3.5

Rate (%/min)

0.6

27

29

Time (min)

31

33

35

Time (min)

Phenol

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

0.6 0.5 0.4 0.3

Rate (%/min) Rate model (%/min)

0.2

Yield (wt%) Yield model (wt%)

0.1 0 25

27

29

31

Time (min)

33

Yield (wt%)

Rate (%/min)

0.7

35

Figure A4.2b Oxygenated hydrocarbons. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

263

0.08

0.012

0.06 0.04 Rate (%/min) Rate model (%/min)

0.02

Yield (wt%) Yield model (wt%)

0 25

27

29

31

Time (min)

33

Rate (%/min)

0.014

Isocyanic Acid (HNCO)

0.04

0.01 0.008

Rate (%/min)

0.006

Yield (wt%)

0.03

Rate model (%/min)

0.02

Yield model (wt%)

0.004

0.01

0.002 0

35

0.05

Yield (wt%)

0.1

Yield (wt%)

Rate (%/min)

Hydrogen Cyanide

0.04 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0

0 25

27

29

31

Time (min)

33

35

0.03

0.006

0.025

0.005

0.02

0.004

0.015

0.003

Rate (%/min)

0.01

Yield model (wt%)

0.002

Rate model (%/min)

0.005

Yield (wt%)

0.001 0

Yield (wt%)

Rate (%/min)

Ammonia 0.007

0 25

27

29

31

33

35

Time (min)

Figure A4.2c Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of wood pellets. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

264

T (°C)

600

60

400

40

200

20

Temperature

100

Time (min)

150

Carbon Dioxide

1.2

Rate model (%/min) Yield model (wt%)

0

Time (min)

0.8

4 2

Yield model (wt%)

0 45

65

Time (min)

85

1.4

0.03

0.6

0.02

0.4

Rate (%/min) Rate model (%/min) Yield (wt%)

0.01

0.2

Yield model (wt%)

0

0 45

65

Time (min)

85

Ethylene 9 8 7 6 5 4 3 2 1 0

0.03

0.6 Rate (%/min)

0.025

0.5

Rate (%/min) Yield (wt%)

0.02

0.4

Yield model (wt%)

0.015

0.3

0.01

0.2

0.005

0.1

0

0 25

45

65

Time (min)

85

Figure A4.3a Weight loss curve and main species. Comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

265

Yield (wt%)

Rate (%/min) Rate model (%/min) Yield (wt%)

Time (min)

85

1

Yield (wt%)

0.4

65

Methane

25

0.6

25

0 45

0.04

85

0.8

0.2

Yield model (wt%)

1.2

Rate (%/min)

65

Carbon Monoxide

1

Rate (%/min)

45

Yield (wt%)

0.05

0 25

Rate model (%/min)

0.06

6

Yield (wt%)

5

Rate (%/min)

10

0.6 0.2

1 0.5

0.07

8

Rate (%/min)

10

12

0.8 0.4

1.5

Yield (wt%)

1

15

2

25

Yield (wt%)

Rate (%/min)

50

0 200

Rate (%/min)

0

2.5

0

Weight Loss

0

20

Yield (wt%)

80

Char mass(%)

800

Water

3

100

Rate (%/min)

Temperature & Weight Loss

1000

Yield (wt%) Yield model (wt%)

0.05 0 25

45

Time (min)

85

1.5

6

1

4

Rate (%/min) Rate model (%/min) Yield (wt%) Yield model (wt%)

0 45

65

2

1.5

0.2

0.25

Yield (wt%) Yield model (wt%)

0

1

Rate (%/min)

Rate (%/min)

Rate model (%/min)

Yield (wt%) Yield model (wt%)

1 0.5 0

45

65

Time (min)

85

Acetone

2 1.5 1

0.15 0.1

Rate (%/min) Rate model (%/min)

0.05

Yield (wt%) Yield model (wt%)

0

85

Time (min)

Rate model (%/min)

0.2

0 65

Rate (%/min)

0.1

4

2

Time (min)

85

0.3

0.3

Rate (%/min)

65

0.5

Yield (wt%)

0.2

0.2

2

5

3

0.4 0

45

0

Yield (wt%)

0.3

45

Yield model (wt%)

25

0.4

25

Yield (wt%)

0.4

85

Acetic Acid

0.1

Rate model (%/min)

2.5

Time (min)

0.5

0.6 Rate (%/min)

0.5

0 25

0.8

Yield (wt%)

8

Yield (wt%)

2

0.5

1

Formic Acid

12 10

1.4 1.2

25

Acetaldehyde

2.5

Rate (%/min)

65

Rate (%/min)

Rate (%/min)

Rate model (%/min)

Formaldehyde

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

Yield (wt%)

0.1

Rate (%/min)

Yield (wt%)

0.15

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Rate (%/min)

Methanol

0.2

0 25

45

65

Time (min)

85

Phenol 2.5 2

0.15

1.5

0.1 Rate (%/min)

0.05

Rate model (%/min) Yield (wt%) Yield model (wt%)

0

1 0.5

Yield (wt%)

Rate (%/min)

0.2

0 25

45

65

Time (min)

85

Figure A4.3b Oxygenated hydrocarbons. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TGFTIR measurements for yields (secondary axis) and rates (primary axis).

266

Hydrogen Cyanide Rate model (%/min)

Yield model (wt%)

0.008

0.15

0.006

0.1

0.004

0.05

0.002

0 25

45

65

Time (min)

0.25

Rate (%/min) Rate model (%/min)

0.2

Yield (wt%) Yield model (wt%)

0.15 0.1 0.05 0 25

85

Ammonia

0.009 0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0

0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0

45

65

Time (min)

85

0.14 0.12 0.08 0.06 0.04

Yield (wt%)

0.1 Rate (%/min) Rate model (%/min) Yield (wt%) Yield model (wt%)

0.02 0 25

45

65

Time (min)

85

Figure A4.3c Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TGFTIR measurements for yields (secondary axis) and rates (primary axis).

267

Yield (wt%)

0.2

Yield (wt%)

0.01

Rate (%/min)

Rate (%/min)

0.012

0

Rate (%/min)

Isocyanic Acid (HNCO) 0.25

Yield (wt%)

Rate (%/min)

0.014

Temperature & Weight Loss

600

60

Temperature Weight Loss

400

40

200

20

0 0

20

40

60

80

0 100

Rate (%/min)

80

T (°C)

800

Water

20

20

15

15

10

10

5

Rate (%/min)

Yield model (wt%)

0

0 25

27

29

0.5

4 Rate (%/min)

2

Rate model (%/min) Yield (wt%) Yield model (wt%)

25

27

29

31

33

Rate (%/min)

Rate (%/min)

8

1.2 1 0.8

0.3

0.6

0.2

Rate model (%/min) Yield (wt%)

0 25

27

Rate (%/min)

2

Rate model (%/min) Yield (wt%)

1

Yield model (wt%)

0

0 33

35

Rate (%/min)

3

Time (min)

33

35

0.25

0.2

0.2

0.15

0.15

0.1

0.1 Rate (%/min)

0.05

0.05

Rate model (%/min) Yield (wt%) Yield model (wt%)

0

0 25

27

29

31

Time (min)

33

35

Figure A4.4a Weight loss curve and main species. Comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 100 K/min. Lines without markers: FGDVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

268

Yield (wt%)

4

3

31

31

Ethylene

0.25

Yield (wt%)

Rate (%/min)

7 5

29

29

Time (min)

4

27

0.2

Yield model (wt%)

0

6

25

0.4

Rate (%/min)

0.1

35

5

1

1.4

0.4

0

Carbon Monoxide

2

35

Methane

Time (min)

6

33

Yield (wt%)

0.6

Yield (wt%)

10

31

Time (min)

Carbon Dioxide

6

5

Rate model (%/min) Yield (wt%)

Time (min)

10 9 8 7 6 5 4 3 2 1 0

25

Yield (wt%)

100

Char mass(%)

1000

1.4

1.2

1.2

1.2

1

0.8

0.8

0.6

0.6

0.4

0.4

Rate (%/min) Rate model (%/min)

0.2

0.2

Yield (wt%) Yield model (wt%)

0 27

29

31

33

1.2 1

0.8

0.8

0.6

0.6

0.4

Rate (%/min)

0.2

Yield (wt%)

0 27

29

6

6

4

4

Rate (%/min)

2

Rate model (%/min)

2

Yield (wt%)

2

2

1.5

1.5

Rate (%/min) Rate model (%/min) Yield (wt%) Yield model (wt%)

1

1 0.5

0.5

Yield model (wt%)

0

0 27

29

31

33

Time (min)

Acetic Acid

3.5

Rate (%/min) Rate model (%/min)

0.5

Yield (wt%) Yield model (wt%)

0 25

27

29

31

33

Time (min)

33

35

Acetone

2.5 2

1.5

1.5

1

1 Rate (%/min) Rate model (%/min)

0.5

0.5

Yield (wt%) Yield model (wt%)

0

35

0 25

27

Phenol

29

31

Time (min)

33

35

2.5 2 1.5 1 Rate (%/min)

0.5

Rate model (%/min) Yield (wt%)

Yield (wt%)

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

31

2

Time (min)

Rate (%/min)

29

Yield (wt%)

2 1.5

27

2.5

Yield (wt%)

2.5

0 25

4 3.5 3 2.5 2 1.5 1 0.5 0

3

1

0

35

Rate (%/min)

25

Rate (%/min)

Rate (%/min)

8

35

Yield (wt%)

8

33

Formic Acid

2.5

Yield (wt%)

Rate (%/min)

10

31

Time (min)

Acetaldehyde 10

0.2

Yield model (wt%)

25

Time (min)

12

0.4

Rate model (%/min)

0

35

14

1.4

1

0 25

Formaldehyde

Yield (wt%)

1

Rate (%/min)

1.4

Yield (wt%)

Rate (%/min)

Methanol 1.4

Yield model (wt%)

0 25

27

29

31

33

35

Time (min)

Figure A4.4b Oxygenated hydrocarbons. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

269

0.03 0.02

Rate (%/min) Rate model (%/min)

0.01

Yield (wt%) Yield model (wt%)

0 25

27

29

31

Time (min)

33

Rate (%/min)

0.04

0.05

0.12

0.04

0.1 0.08

0.03

0.06

0.02

0.04

Rate (%/min)

0.01

Rate model (%/min)

0.02

Yield (wt%) Yield model (wt%)

0

0 25

35

Yield (wt%)

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0

Yield (wt%)

Rate (%/min)

Isocyanic Acid (HNCO)

Hydrogen Cyanide

0.05

27

29

31

Time (min)

33

35

0.12 0.1 0.08 0.06 0.04

Rate (%/min) Rate model (%/min)

0.02

Yield (wt%) Yield model (wt%)

Yield (wt%)

Rate (%/min)

Ammonia 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0 25

27

29

31

Time (min)

33

35

Figure A4.4c Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of miscanthus giganteus pellets. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

270

Temperature & Weight Loss 0.5

80

600

60

400

40

200

Temperature Weight Loss

0 0

100

Time (min)

150

0 200

Water

0.4

10

0.3 Rate (%/min)

0.2 0.1

Yield model (wt%)

0

0 10

0.1

0.5

10

0.4

8 6

Rate (%/min)

4

Rate model (%/min) Yield (wt%) Yield model (wt%)

2

0 50

Time (min)

70

Carbon Monoxide

Rate (%/min) Rate model (%/min) Yield (wt%) Yield model (wt%)

10

30

50

Time (min)

90

70

90

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

0.08 0.06

Rate (%/min) Rate model (%/min)

0.04

Yield (wt%) Yield model (wt%)

0.02 30

50

Time (min)

70

90

Ethylene

0.035

0.35

0.03

0.3

0.025

0.25

0.02

0.2

0.015

0.15

Rate (%/min)

0.01

0.1

Rate model (%/min) Yield (wt%)

0.005

Yield (wt%)

18 16 14 12 10 8 6 4 2 0

Yield (wt%)

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

70

Methane

10

90

Rate (%/min)

30

50

0

0 10

Rate (%/min)

Rate (%/min)

12

0.1

30

0.05

Yield model (wt%)

0

0 10

30

50

Time (min)

Yield (wt%)

0.6

Yield (wt%)

0.12

0.2

5

Rate model (%/min) Yield (wt%)

14

0.3

15

Time (min)

Carbon Dioxide

0.7

Rate (%/min)

50

20

Rate (%/min)

800

100

T (°C)

0.6

Yield (wt%)

120

Char mass(%)

1000

70

90

Figure A4.5a Weight loss curve and main species. Comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of Hambach brown coal. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

271

0.2

0.004

0.15 0.1

Rate (%/min) Rate model (%/min) Yield (wt%)

0.05

Yield model (wt%)

0 10

30

50

Time (min)

70

90

Rate (%/min)

0.005

Ammonia

0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0.003 0.002

Rate (%/min) Rate (%/min)

0.001

Yield (wt%) Yield model (wt%)

0 10

30

50

Time (min)

70

90

Figure A4.5b Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of Hambach brown coal. Heating rate: 10 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

272

Yield (wt%)

0.25

Yield (wt%)

Rate (%/min)

Hydrogen Cyanide

0.008 0.007 0.006 0.005 0.004 0.003 0.002 0.001 0

Temperature & Weight Loss

400

40

200

20

Weight Loss Temperature

0

4 3

40

60

Time (min)

8 6 4

5

10

15

Time (min)

0

Yield model (wt%)

10

15

20

Time (min)

Ethylene

0.3

0.25

0.25

0.2

0.2

0.35

0.15

0.15

0.1

0.1

Rate (%/min)

Rate model (%/min)

0.05

0.05

Yield (wt%)

Yield model (wt%)

0 5

10

15

Time (min)

Yield (wt%)

Yield (wt%)

Time (min)

20

Yield (wt%)

0.3

0 15

Rate model (%/min)

12

2

Yield (wt%) Yield model (wt%)

Rate (%/min)

0.2

0.35

4

10

0.4

14

6

5

0.6

5

8 Rate (%/min)

Methane

0

10

Rate model (%/min)

1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

20

0.8

20

Carbon Monoxide

4 3.5 3 2.5 2 1.5 1 0.5 0

15

1

2

Yield model (wt%)

10

14 12 10 8 6 4 2 0

Time (min)

10

Yield (wt%)

Yield model (wt%)

5

12

Rate model (%/min)

Yield (wt%)

0

14

Rate (%/min)

Rate model (%/min)

1

80

Carbon Dioxide

8 7 6 5 4 3 2 1 0

Rate (%/min)

2

Rate (%/min)

20

Yield (wt%)

Rate (%/min)

5

0 0

Rate (%/min)

Rate (%/min)

60

Rate (%/min)

T (°C)

600

6

Yield (wt%)

80

Char mass(%)

800

Water

7

100

Yield (wt%)

1000

0 20

Figure A4.6a Weight loss curve and main species. Comparison between TG-FTIR measurements and FG-DVC modelling results for rates and yields obtained for the pyrolysis of Hambach brown coal. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

273

Rate model (%/min) Yield (wt%)

0.15

Yield model (wt%)

0.1 0.05 0 5

10

15

Time (min)

20

Rate (%/min)

0.2

Rate (%/min)

0.1

0.04 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0

0.08 0.06 0.04

Rate (%/min) Rate model (%/min)

0.02

Yield (wt%) Yield model (wt%)

Yield (wt%)

0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

Yield (wt%)

Rate (%/min)

Ammonia

Hydrogen Cyanide

0 5

10

15

20

Time (min)

Figure A4.6b Nitrogen species. Comparison between TG-FTIR and FG-DVC modelling results for rates and yields obtained for the pyrolysis of Hambach brown coal. Heating rate: 100 K/min. Lines without markers: FG-DVC model predictions for yields (secondary axis) and rates (primary axis). Lines with ■ markers: experimental TG-FTIR measurements for yields (secondary axis) and rates (primary axis).

274

APPENDIX 5 Calculation of axial Péclet numbers for PFBG and DWSA tests simulated In order to determine the fluidised bed reactor regime, an approach given by [Westerterp et al., 1984] is followed. The Péclet number ( Pe = u ⋅ L D ax ) related to the reactor is the empirical mean by which the bed hydrodynamic regime is evaluated: if Pe is higher than 100, then a plug flow regime can be assumed for the model. To estimate Pe, the Dax coefficient must be evaluated first. This can be done by means of calculating the Bodenstein number, Bo, defined as:

Bo =

u ⋅d D ax

(A5.1)

where d is a dimension significant in view the mechanism causing the dispersion, in this case the particle diameter of the bed dp. To calculate Bo, for fluidised beds, the following empirical relation can be applied [Chung & Wen, 1968]:

ε g ⋅ Bo ⋅ Re Re mf

= 0.20 + 0.011⋅ Re 0.48

(A5.2)

in which Re is the dimensionless Reynolds number, εg is the fraction of fluid phase in fluid-solid system (gas porosity). Reynolds number is defined as Re = u·dp/ν where ν is the kinematic viscosity (m2/s) of the gas and dp (m) is the particle diameter. Remf represents the Reynolds number at incipient fluidisation. The correlation (A5.2) is given to be valid for εg = 0.4-0.8, Re = 10-3-103, and particle densities up to 8000 kg/m3. This is the case for the experiments that have been performed. The experimental set points simulated for the Delft PFBG and IVD DWSA gasifiers are characterised by the following data related to the calculation of the Péclet numbers, as presented in table A5.1. The minimum fluidisation velocity necessary to calculate Remf has been determined according to an empirical relation for pressurised fluidised beds, given by [Rowe, 1984]: ⎧ ⎪ ⎡ 0.0003112 Ga ε 3 η ⎪ mf u mf = 42.9 1 − ε mf ) ⎨ ⎢1 + ( ⎢ 2 ρgas d p (1 − ε ) ⎪⎢ mf ⎪⎩ ⎣ with Ga being the Galilei number, which is defined as: Ga =

ρgas ( ρ s − ρ gas ) gd p

1 ⎫ ⎤ 2 ⎪ ⎪ ⎥ − 1⎬ ⎥ ⎪ ⎥⎦ ⎪⎭

(A5.3)

3

η2

The average particle diameter of the bed material is taken to be the Sauter mean diameter: 1 dp = N⎛ Y ⎞ ∑⎜ ⎟ i=1 ⎜ d p ⎟ ⎝ ⎠i with Yi: mass fraction of solid bed material.

(A5.4)

(A5.5)

275

Table A5.1 Calculation of the Péclet numbers for simulated PFB gasifiers based on [Chung&Wen, 1968] Test rig

Exp.

Fuel

u (m/s)

umf (m/s)

PFBG

981216 990107_1 990107_2 020429 020513 011030 011127 020111 020129 020205 020212 020220 020226 020319 020409 020416 991103 991115 991118 991202 991206 991208

miscanthus miscanthus miscanthus miscanthus miscanthus wood wood wood wood wood wood wood wood brown coal brown coal brown coal brown coal brown coal brown coal wood wood wood

0.76 0.75 0.72 0.86 0.89 0.99 0.86 0.80 0.84 0.82 0.75 0.77 0.70 0.71 0.64 0.66 0.27 0.28 0.31 0.33 0.36 0.53

0.15 0.14 0.15 0.15 0.15 0.14 0.15 0.14 0.14 0.14 0.14 0.15 0.15 0.15 0.14 0.14 0.08 0.08 0.08 0.09 0.09 0.09

DWSA

276

ν (m2/s)

2.7 x 10-5 2.1 x 10-5 3.3 x 10-5 4.2 x 10-5 4.3 x 10-5 4.9 x 10-5 4.7 x 10-5 5.1 x 10-5 3.8 x 10-5 3.8 x 10-5 3.7 x 10-5 4.6 x 10-5 4.8 x 10-5 3.1 x 10-5 3.2 x 10-5 4.8 x 10-5 2.7 x 10-5 2.8 x 10-5 2.8 x 10-5 2.8 x 10-5 3.0 x 10-5 1.0 x 10-4

Bo (-)

7.5 x 10-2 7.3 x 10-2 7.9 x 10-2 6.3 x 10-2 6.0 x 10-2 4.9 x 10-2 6.0 x 10-2 6.3 x 10-2 5.9 x 10-2 6.0 x 10-2 6.7 x 10-2 6.9 x 10-2 7.4 x 10-2 7.7 x 10-2 8.4 x 10-2 7.8 x 10-2 1.4 x 10-1 1.1 x 10-1 1.0 x 10-1 9.6 x 10-2 8.8 x 10-2 4.7 x 10-2

Dax (m2/s)

5.4 x 10-3 5.4 x 10-3 4.8 x 10-3 7.2 x 10-3 7.9 x 10-3 1.1 x 10-2 7.5 x 10-3 6.8 x 10-3 7.5 x 10-3 7.3 x 10-3 5.9 x 10-3 5.9 x 10-3 5.0 x 10-3 4.8 x 10-3 4.0 x 10-3 4.5 x 10-3 7.0 x 10-4 8.8 x 10-4 1.1 x 10-3 1.2 x 10-3 1.5 x 10-3 4.0 x 10-3

Pe (-) 238 196 209 175 171 143 150 167 168 165 179 172 187 229 187 203 380 318 281 270 247 205

APPENDIX 6 Overview of publications 1993 Westerterp, K.R., de Jong, W. and van Benthem, G.H.W. (1993) “Comments on discrimination of three approaches to evaluate heat fluxes for wall cooled fixed bed chemical reactors”, Chem. Engng. Sci., 48, 2669-2670 1994 Westerterp, K.R., de Jong, W. and van Benthem, G.H.W. (1994) “A more strict criterion for application of homogeneous models in modelling wall cooled packed bed reactors”, Chem. Engng. Sci,. 49, 3830-3831. 1996 Willers, E., Groll, M., Isselhorst, A and de Jong, W. (1996) “Advanced concept of a metal hydride solid sorption device for combined heating and air-conditioning”. In: Proceedings of the International AB-sorption Heat Pump Conference, 17-20 September 1996, Montreal, CA, pp. 499-505. 1997 Andries, J., de Jong, W. and Hein, K.R.G. (1997) “Co-gasification of biomass and coal in a pressurised fluidised bed gasifier”. In: G. van der Bijl and E.E. Biewinga (eds.) ‘Environmental impact of biomass for energy, proceedings of a conference’, 4-5 November 1996, Noordwijkerhout, CLM, Utrecht, The Netherlands, pp. 45-47. ISBN 90-5634-054-9. Andries, J. and de Jong, W. (1997) “Wervelbedvergassing onder druk van biomassa en kolen-biomassa mengels”. In: ‘Ruimte voor Duurzame Energie’, conferentieboek Nederlandse Duurzame Energie Conferentie 1997, 17-18 november 1997, Ede, The Netherlands, pp. 165-166. ISBN 90-803981-1-x. Andries, J., de Jong, W. and Hoppesteyn, P.D.J. (1997) “Integration of high temperature pollutant control systems”. Final Report JOU2-CT93-0431, January 1997, EU, Brussels, Belgium. Andries, J. and de Jong, W. “Coal-biomass system components development and design”, Second progress report JOR3CT95-0018, January 1997, EU, Brussels, Belgium. Andries, J. and de Jong, W. “Coal-biomass system components development and design”, Midterm report JOR3-CT95-0018, July 1997, EU, Brussels, Belgium. Andries, J. de Jong, W. and Hein, K.R.G. (1997) “Co-gasification of Biomass and Coal in a Pressurized Fluidized Bed Gasifier”, in ‘Wirbelschichtfeuerungen: Erfahrungen und Perspektiven’, VDI Berichte 1314, VDI Verlag, Düsseldorf, Germany, pp. 211-215. ISBN 3-18-0913-14-2. Andries, J., de Jong, W. and Hein, K.R.G. (1997) “Co-gasification of biomass and coal in a pressurised fluidised bed gasifier”. In: M. Kaltschmitt and A.V. Bridgwater (eds.) ‘Biomass, Gasification and Pyrolysis’; CPL Press, Newbury, paper presented at the conference ‘Gasification and Pyrolysis of biomass’, 9-11 April 1997, Stuttgart, Germany, pp. 1282-1291. ISBN 1-872691714. Andries, J. de Jong, W. and Hein, K.R.G. (1997) “Co-gasification of biomass and coal in a pressurized fluidized bed gasifier”. In: A. Ziegler, K.H. van Heek, J. Klein and W. Wanzl (eds.) Proceedings of the 9th International conference on Coal Science, 7-12 September 1997, Essen, Germany, DGMK Tagungsberichte 9703, vol. 2, Essen, Germany, pp.12651267. Andries, J., de Jong, W. and Hoppesteyn, P.D.J. (1997) EV-1964, 'Wervelbedvergassing van biomassa', Eindrapport Fase 1A NOVEM EWAB project 355196/150, Delft, The Netherlands. De Jong, W., Andries, J. and Hein, K.R.G. (1997) “Co-gasification of biomass and biomass-coal mixtures in a bubbling pressurised fluidised bed reactor using air and steam”. In: R.P. Overend and E. Chornet (eds.) ‘Making a business from biomass’, Pergamon, Oxford, 1997, Paper and poster presented at the ‘third Biomass Conference of the Americas-making a business of biomass’, 24-29 August 1997, Montreal, Canada, pp. 559-570. ISBN 0-08-0429-693. Hoppesteyn, P.D.J., de Jong, W., Andries, J. and Hein, K.R.G. (1997) “Combustion of biomass-derived low calorific value fuel gas”. In: Proceedings of combustion & emissions control III. Elsevier, London, United Kingdom, pp. 293-303. ISBN 090-2597558. 1998 Andries, J. and de Jong, W. (1998) “Coal-biomass system components development and design”. Second annual report JOR3-CT95-0018, EU, Brussels, Belgium.

277

Andries, J., de Jong, W. and Hein, K.R.G. (1998) “Gasification of biomass and coal-biomass mixtures in a pressurised fluidised bed gasifier”. In: R.W.F. Riemer, A.Y. Smith, K.V. Thambimuthu (eds.) ‘Greenhouse gas mitigation, Proceedings of technologies for activities implemented jointly’, Elsevier, 26-29 May 1997, Vancouver, Canada, pp 537 – 543. ISBN 0 -8 0433 251. Andries, J., de Jong, W. and Hein, K.R.G. (1998) "Gasification of biomass and coal in a pressurised fluidised bed gasifier". In: Tagungsbericht 9802, Proceedings of the "Energetische und stoffliche Nutzung von Abfallen und nachwachsenden Rohstoffen", DGMK, 20 - 22 april 1998, Velen, Germany, pp 319-326. Andries, J., de Jong, W. and Hein, K.R.G. (1998) “Co-gasification of biomass and coal in a pressurised fluidised bed gasifier”. In: G.Q. Lu, V. Rudulph, P.F. Greenfield (eds.) ‘Proceedings of the 2nd Asia Pacific Conference on sustainable energy and environmental technologies- challenges and opportunities’, 14-17 June 1998, Gold Coast, Australia, pp 383-389. ISBN 0-86776-754-5. Andries, J., Ünal, Ö., de Jong, W.,. Hoppesteyn, P.D.J and Hein, K.R.G., (1998) “High temperature dust filtration downstream of a coal/biomass fuelled gasifier using a ceramic channel flow filter”, Paper presented at the 23rd International Technical Conference on Coal Utilization and fuel systems, March 9-13, 1998, Clearwater, Florida, USA. Dehouche, Z., de Jong, W., Willers, E., Isselhorst, A. and Groll, M. (1998) “Modelling and simulation of heating / airconditioning systems using the thermal wave metal hydride cascade concept”, Appl. Thermal Engng. 18 (6), 457-480 De Jong, W., Andries, J. and Hein, K.R.G. (1998) “Co-gasification of coal and biomass in a pressurised fluidised bed reactor”. In: Proceedings of the VGB conference ‘Forschung für die Kraftwerkstechnik 1998’, VGB-TB 233, contribution H1, pp 1-6, Paper presented at the 10th International VGB Conference ‘Forschung für die Kraftwerkstechnik 1998’, 11-12 February 1998, Essen, Germany. De Jong, W., Andries, J. and Hein, K.R.G. (1998) “Coal-Biomass Gasification in a Pressurized Fluidized Bed Gasifier”, ASME New York, paper 98-GT-159, Paper presented at the International Gas Turbine & Aeroengine Congress & Exhibition, 2-5 June, Stockholm, Sweden. De Jong, W., Andries, J. and Hein, K.R.G. (1998) “Pressurised fluidised bed co-gasification of coal and biomass”, in ‘Biomass for Energy and Industry’ Proceedings of the 10th European Conference and Technology Exhibition, Eds. H. Kopetz, T. Weber, W. Palz, P. Chartier and G. L. Ferrero, C.A.R.M.EN., Rimpar, Paper and poster presented the 10th European Conference and Technology Exhibition ‘Biomass for Energy and Industry’, 8-11 June 1998, Würzburg, Germany, pp. 1781-1784. Hoppesteyn, P.D.J., de Jong, W., Andries, J. and Hein, K.R.G. (1998) “Coal Gasification and Combustion of LCV gas”, Bioresource Technology, 65, 105-115. 1999 Andries, J., de Jong, W., Hoppesteyn, P.D.J., Ünal, Ö and Hein, K.R.G. (1999) “Fluidized bed gasification of miscanthus and coal, high temperature gas cleaning using a ceramic channel-flow filter and combustion of the low calorific value fuel gas in a gas turbine combuster”. In: ‘Fortschrittliche Energiewandlung und -Anwendung’, Tagung München, 16-17 March 1999, VDI Berichte 1457, VDI Verlag, Düsseldorf, Germany, pp.399-408. Andries, J., de Jong, W. and Hoppesteyn, P.D.J. (1999) “Integration of High-Temperature Pollutant Control Systems”. In: K.R.G. Hein, A.J. Minchener, R. Pruschek, P.A. Roberts (eds.) ‘Integrated Hot Fuel Gas Cleaning for Advanced Gasification Combined Cyle Processes’, vol. IV, EU-contract JOU2-CT93-0431. ISBN 3-928123-29-7. De Jong, W., Andries, J. and Hein, K.R.G. (1999) “Coal/Biomass co-gasification in a pressurised fluidised bed bed reactor”, Renewable Energy 16, 1110-1113. De Jong, W., Andries, J., Hoppesteyn, P.D.J. and Ünal, Ö, (1999) “Conversion of biomass and biomass-coal mixtures: Gasification, hot gas cleaning and gasturbine”. In: Proceedings of the ‘2nd Olle Lindström Symposium’, June 9-11 1999, Stockholm, Sweden, pp. 113-116. De Jong, W., Andries, J., Hoppesteyn, P.D.J. and Ünal, Ö, (1999) “Conversion of biomass and biomass-coal mixtures: Gasification, hot gas cleaning and gasturbine”. In: Overend, R.P & Chornet, E. (eds.) Proceedings of the ‘fourth biomass conference of the America’s’, 29 August – 2 September 1999, Oakland, USA, pp.1009-1016. 2000 De Jong, W., Andries, J., Hoppesteyn, P.D.J., Ünal, Ö. and Hein, K.R.G. (2000) “Thermochemical gasification of biomass: The fate of main components, fuel nitrogen and trace components in a pressurized fluidized bed gasification system”. In: Proceedings of the IchemE conference ‘Gasification 4 the future’, poster presented at the ‘Gasification 4 the future’, 11-13 April 2000, Noordwijk, The Netherlands. De Jong, W., Andries, J., Hoppesteyn, P.D.J., Ünal, Ö. and Hein, K.R.G. (2000) “Miscanthus gasification in a pressurised fluidised bed gasifier, hot gas cleanup and product gas combustion in a gas turbine combustor”. In: Proceedings of the ‘1st

278

World Conference on Biomass for Energy and Industry’, poster presented at the ‘1st World Conference on Biomass for Energy and Industry’, 5-9 June 2000, Sevilla, Spain, pp. 1595-1598. Ünal, Ö., de Jong, W. and Spliethoff, H. (2000) "Vergleich von Braunkohle und Holz für verschiedene Druckwirbelschichtfeuerungen der 2. Generation". In: ‘Wirbelschichtfeuerungen und -vergasung: Erfahrungen und Perspektiven’, Cottbus, 28-29 March 2000, VDI Berichte 1535, VDI Verlag, Düsseldorf, Germany, pp. .199-208. Ünal, Ö., Paul, S., de Jong, W. and Spliethoff, H. (2000) "Dynamic modelling and control of a biomass based pressurised fluidised bed gasifier". In: Proceedings of the VGB Kraftswerktechnik, 10-12 October, Düsseldorf, Germany, pp 1-12. 2001 Andries, J., de Jong, W., Hoppesteyn, P.D.J. and Ünal, Ö. (2001) "The conversion of fuel nitrogen in a biomass-fuelled pressurized fluidized bed gasification system''. In: Proceedings of the International conference on power engineering, 8-12 October 2001, Xian, China, pp. 1278-1283. De Jong, W., Ünal, Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2001) “Pressurised Gasification Experiments and Modelling of Biomass and Fossil Fuels in Fluidised Bed Gasifiers with Hot Gas Ceramic Filters”. In: Proceedings of the Clean Air Conference VI, Porto, 9-12 July 2001, pp. 1327-1336. De Jong, W., Andries, J., Hoppesteyn, P.D.J., Ünal, Ö. and Hein, K.R.G. (2001) “Pressurised gasification of biomass and fossil fuels in fluidised bed gasifiers, hot gas cleanup using ceramic filters and pressurised product gas combustion”. In: A.V. Bridwater (ed.) Proceedings of the conference ’Progress in Thermochemical Biomass Conversion’, 17-22 September 2000, Innsbruck, Austria, pp.473-487. De Jong,W. , Ünal Ö., Hein, K.R.G. and Spliethoff, H. (2001) “Pressurised fluidised bed gasification experiments of biomass and fossil fuels”, Thermal Science, 5(2), pp. 69-81. 2002 Adouane, B., Hoppesteyn, P., de Jong, W., van der Wel, M.C., Hein, K.R.G. and Spliethoff, H. (2002) “Gas turbine combustor for biomass derived LCV gas, a first approach towards fuel-NOx modelling and experimental validation”, Appl.Thermal Engng., 22, pp. 959-970. De Jong, W., Ünal. Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2002) “Thermochemical conversion of brown coal and biomass in a pressurised fluidised bed gasifier with hot gas filtration using ceramic channel filters, measurements and gasifier modelling”. In: Proceedings of the ‘Energex 3’ conference, 19-24 May, Krakau, Poland. De Jong, W., van der Wel, M., Ünal, Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2002) “Measurements and modelling of biomass and brown coal gasification in a pressurised fluidised bed gasifier with hot gas filtration using ceramic channel filters”. In: Palz, W., Spitzer, J., Maniatis, K., Kwant, K., Helm, P. and Grassi, A. (eds.) Proceedings of the ‘12th European Conference and Technology Exhibition on Biomass for Energy, Industry and Climate Protection’, 17-21 June, Amsterdam, The Netherlands, pp. 485-488. De Jong, W., van der Wel, M., Ünal, Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2002) “Gasification of biomass and brown coal in an advanced pressurised process development unit consisting of a fluidised bed with a hot gas ceramic filter and pressurised combustor for the low calorific fuel gas”. In: Bertucco, A. (ed.) Proceedings of the HIGH PRESSURE IN VENICE 4th INTERNATIONAL SYMPOSIUM ON HIGH PRESSURE PROCESS TECHNOLOGY AND CHEMICAL ENGINEERING, 22-25 September, Venice, Italy, pp. 13-18. De Jong, W., Slabbekoorn, A., Guo, J. and Veefkind, A. (2002) “Heated grid flash pyrolysis of Miscanthus with in-situ infrared spectrometry species analysis and comparison with FG-DVC biomass model simulations”. In: Bridgwater, A.V. (ed.) Proceedings of the expert meeting on pyrolysis and gasification of biomass and waste, 30 September – 1 October, Strasbourg, France, pp. 111-123. Guo, J., de Jong, W. and Veefkind, A. (2002) ``Gasification of wood char with CO2”. In: Proceedings of the conference Renewable energy - renewables: world's best energy option, pp. 1-5. Van der Wel, M.C., Staiger, B., Ünal, Ö., de Jong, W. and Spliethoff, H. (2002) "Impact of tars on combustion of biomass derived low calorific value gas in gas turbines". In: Proceedings of the DGMK-Fachbereichstagung "Energetische Nutzung von Biomassen", 22-24 April, Velen, Germany, pp. 151-158.

279

2003 Adouane, B., Witteveen, G., de Jong, W. and Van Buijtenen, J.P. (2003) “Towards the Design of an Ultra Low NOx Combustor for Biomass Derived LCV Gas”, Revue des Energies Renouvables, 11èmes Journées Internationales de Thermique, pp. 111-117. Adouane, B., Van der Wel, M.C., de Jong, W. and Van Buijtenen, J.P. (2003) “Low NOx emission in a newly designed combustor for LCV gas: A modelling and experimental investigation”. In: Proceedings of Bioenergy2003 conference, 2 - 5 September, Jyväskylä, Finland. Adouane, B., Van der Wel, M.C., de Jong, W. and Van Buijtenen, J.P. (2003) “Experimental investigation on a newly designed combustor for LCV gas”. Paper GT2003-38930. In: ‘ASME TURBO EXPO 2003 conference’, paper GT200338930, 16 - 19 June, Atlanta, Georgia, USA. Andries, J., de Jong, W. and Spliethoff, H. (2003) “Sustainable Production of Hydrogen using Thermochemical Gasification of Biomass.” In: Proceedings of the International Conference on Power Engineering-03, 9-13 November, Kobe, Japan, vol. 3, pp. 369-372. De Jong , W., Pirone, A., Wójtowicz, M.A. (2003) “Pyrolysis of Miscanthus Giganteus and wood pellets: TG-FTIR analysis and reaction kinetics”, Fuel, 82(9), pp. 1139-1147. De Jong, W., Ünal. Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2003) “Biomass and fossil fuel conversion by pressurised fluidised bed gasification using hot gas ceramic filters as gas cleaning”, Biomass & Bioenergy, 25(1), pp. 59-83. De Jong, W., Ünal. Ö., Andries, J., Hein, K.R.G. and Spliethoff, H. (2003) “Thermochemical conversion of brown coal and biomass in a pressurised fluidised bed gasifier with hot gas filtration using ceramic channel filters, measurements and gasifier modelling”, Applied Energy, 74(3-4), pp. 425-437. Van der Wel, M.C., de Jong, W. and Spliethoff, H. (2003) “Tars and the Combustion of Biomass Derived Low Calorific Value Gas in a Gas Turbine Combustor”. In: proceedings of the VDI conference Verbrennung und Feuerungen - 21. Deutscher Flammentag, 9-10 September, Cottbus, Germany, VDI Berichte Nr. 1750, pp. 187-193. Van der Wel, M.C., de Jong, W. and Spliethoff, H. (2003) “Tar and soot in Low Calorific Value gas combustion”. In: Proceedings of the International Conference on Power Engineering-03, 9-13 November, Kobe, Japan, pp. 2-363-368. 2004 Adouane, B., Witteveen, G., de Jong, W., van Buijtenen, J.P. (2004) “An experimental investigation of a newly designed combustor for lcv gas with low NOx emissions from fuel bound nitrogen”. In: ‘ASME TURBO EXPO 2004’ conference, paper GT2004-54038, 14-17 June, Vienna, Austria. De Jong, W., Andries, J. and Spliethoff, H. (2004) “Sustainable Hydrogen production by gasification of biomass”. In: Proceedings of the DGMK-Fachbereichstagung "Energetische Nutzung von Biomassen", 19-21 April, Velen, Germany, pp. 39-46. De Jong, W., Glazer, M., Siedlecki, M., Ünal, Ö. and Spliethoff, H. (2004) "High temperature gas filtration results obtained for fluidised bed gasification and combustion". In: Proceedings of the 13th European Conference and Technology Exhibition on Biomass for Energy, Industry and Climate Protection’, 10-14 May, Rome, Italy, pp. 1234-1237. Heikkinen, J.M., Hordijk, J.C., de Jong, W. and Spliethoff, H. (2004) “Thermogravimetry as a tool to classify waste components to be used for energy generation”, J. Anal. Appl. Pyrolysis, 71, pp. 883-900.

280

Dankwoord Aan het eind gekomen van de promotie, een belangrijke fase in mijn wetenschappelijke loopbaan, wil ik iedereen van harte dank zeggen die op wat voor manier dan ook constructief heeft bijgedragen aan het totstandkomen van dit werk. Alleen zou ik het niet gered hebben. Dit werk is mede gefinancieerd door NOVEM in het kader van het EWAB programma (projecten 355298/2010 en 355196/1150) en door de EU in het kader van het Joule-Thermie programma (contracten JOF3-CT95-0018, JOR3-CT95-0027 en ENK-CT-2000-00313). Voor het hierin gebodene werk- en discussieplatform ben ik zeer erkentelijk. Mijn bijzondere dank gaat uit naar Prof. Dr. –Ing. Klaus R.G. Hein, mijn promotor, die mij de mogelijkheid en vrijheid gegeven heeft om bij de sectie Energy Technology van de TU Delft mijn onderzoek uit te voeren met oog voor de industriële relevantie ervan. Ondanks zijn uitzonderlijk drukke bestaan heeft hij toch de tijd gevonden om mij te begeleiden. Van cruciaal belang voor het onderzoek was ook het feit dat hij mij, in de lange, moeilijke tijd dat werk in ons Delftse lab door de verhuizingsperikelen onmogelijk werd, in staat stelde aan de universiteit Stuttgart (Institut für Verfahrenstechnik und Dampfkesselwesen), alwaar hij eveneens hoogleraar is, experimenten onder druk op kleinere schaal uit te voeren. Voorts wil ik hier mijn co-promotor Prof. Dr. Jacob A. Moulijn, hoogleraar van de sectie Reactor and Catalysis Engineering (faculteit TNW), bedanken voor de erg enthousiaste en stimulerende discussies die we onder het genot van een uitstekend vers gezet kopje espresso over de voortgang van met name het chemische modelleringswerk mochten voeren. Ook past het om hier Prof. Dr. –Ing. Hartmut Spliethoff, hoogleraar van de sectie Energy Technology, te bedanken voor de prettige samenwerking en de ruimte die hij mij gaf om mijn promotie af te ronden. Hij stelde mij in de gelegenheid om nieuwe projecten op te zetten en om hier mijn praktische, onderwijs- en modelleerwerk voort te kunnen zetten in de richting van docent experimenteel onderzoek thermische conversieprocessen. Mijn dagelijks begeleider, Drs. Jans Andries, wil ik op deze plaats bedanken voor de vele kritische discussies over het werk en de inzet die hij tijdens mijn promotie getoond heeft bij het verwerven van projecten en persoonlijke contacten in het “biomassa wereldje”, passend binnen mijn promotie. Voorts wil ik hem bedanken voor het corrigeren van het manuscript. Zeer waardevol waren ook de discussies over experimenteel werk, modellering en literatuuruitwisseling met mijn collega experimentele promovendi: Peter Hoppesteyn, Ömer Ünal, Marco van der Wel, Johanna Heikkinen, Guanyi Chen en Belcacem Adouane. Mijn dank gaat zeer wel naar hen uit. Dit geldt ook voor collega-promovendi aan het IVD (Universität Stuttgart, Duitsland), en wel in het bijzonder voor Holger Nagel. Mijn begeleidingscommissie bestond naast enige uit bovengenoemde personen uit de volgende leden: Ir. Ad van Dongen (UNA/Reliant/Nuon), Dr.Ir. Joep van Doorn (ECN/Grontmij), Ir. Adri de Dooij (voorm. Schelde), Ir. Erik de Kant (HoSt) en Dr.Ir. Bert Wagenaar (BTG). Hen allen wil ik dankzeggen voor de constructieve discussies, met name over de industriële relevantie van het onderzoek, en voor de moeite die zij genomen hebben om naar het soms verre westen te reizen. Zonder de ondersteuning van de technici was er van het experimentele onderzoek niets terecht gekomen, daarom wil ik de volgende personen (in alfabetische volgorde) bedanken: Daniel van Baarle, Sam Berkhout (je was er altijd bij de metingen tot diep in de nacht!), Johan Boender, Duco Bosma (ook jij ging door tot het einde van de metingen), Tjibbe van Dijk, Stefan ten Hagen, Willem Middelkoop (jij ontwierp en implementeerde met verve de procesapparatuur en meetsondes), Jasper Ruygrok, Aad Vincenten (jij kreeg het ingewikkelde vergassingsproces goed geregeld) en Rick Weers.

281

Eveneens wil ik voor het beheer van de financiële kant van projecten rondom mijn promotie Maarten de Groot bedanken: hij is in staat financiële deuren te openen die voor menigeen gesloten blijven. In de computer hard- en software ondersteunende sfeer wil ik Rob Staal en Teus van der Stelt bedanken. Voor de ondersteuning bij de finishing touch van het proefschrift ben ik Jaap Keuvelaar zeer erkentelijk. Voor de hulp op organisatie- en secretarieel gebied wil ik Joke Ammerlaan, Marie-Thérèse Van en Janneke Kempes met name noemen. In het kader van hun afstuderen binnen de sectie Energievoorziening hebben de volgende studenten een opdracht bij mij uitgevoerd: Bauke Leentjes, Armando Pirone en Abel Slabbekoorn. Als afstudeerder van de universiteit Stuttgart heeft Thomas Forn belangrijk modelleerwerk uitgevoerd. Grote dank gaat uit naar jullie! Voorts zijn er studenten die een eerste opdracht binnen de sectie hebben uitgevoerd met een onderwerp uit mijn keuken: Ruben Dijkstra, Heiko de Jong en Ragnhild Smeets. Jullie wil ik hiervoor bedanken. Twee studenten, Marc Schot en Dick Laroo, hebben een K-opdracht uitgevoerd met de kleinste opstelling die we in het lab hebben: “het pruttelpotje”. Van de TH Rijswijk heb ik een aantal afstudeerstudenten gehad die mij vooral in de praktische sfeer, maar soms ook met modelleren, hebben geholpen: Raymond Asmoredjo, Dirk Bos, Robin van Dijk, Patrick Dijks, Han Engel, Mark Gonesh, Karin Hoek, Patrick Nieuwenhuijzen en René Pakvis. Van de HLO zijn er twee zeer waardevolle afstudeerders op analytisch chemisch gebied geweest: Duco Bosma (al eerder genoemd, ja) en Louisa Ruijters. Als afstudeerder van de Hogeschool Rotterdam e.o. heeft Alexander Wemmenhove erg nuttig modelleerwerk voor mij gedaan. Stefanie Albertus heeft een stage doorlopen voor de Hogeschool Rotterdam e.o. in de moeilijke wederopbouwfase van onze analyseruimte na de herinrichting van het laboratorium. Eveneens is er een aantal Antilliaanse TH stagiairs geweest die ook hun bijdrage hebben geleverd aan mijn promotie: David Maria en Tjon-a-Tjen. Voor de mogelijkheid om met het heated grid metingen te verrichten, menige discussie en medebegeleiding van Abel Slabbekoorn, ben ik Bram Veefkind van de TU Eindhoven (Technische Natuurkunde, sectie gasdynamica) zeer erkentelijk. Mijn dank gaat eveneens uit naar ing. Drenth (Agromiscanthus) die de hoofdbrandstof van het onderzoek, miscanthus, heeft geleverd en bereid was veel materiaal te leveren en op te slaan. Dr. Menden (RWE-Rheinbraun) wil ik bedanken voor levering van de bruinkool op wervelbed specificatie. Ard-Jan Mosterd doe ik mijn dank toekomen voor het creatieve ontwerp van de voorkant. Last-but-not-least wil ik noemen mijn ouders, schoonouders en familie, maar bovenal mijn lieve vrouw Klarine die mij altijd heeft ondersteund en gestimuleerd om voor mezelf op te komen, door te gaan en mezelf te blijven.

282

Curriculum Vitae

The author Wiebren de Jong was born on September 18th 1968 in Rotsterhaule (The Netherlands). He attended high school at the Rijksscholengemeenschap Heerenveen, where he graduated in 1986 (VWO-gymnasium). Then he started his study Chemical Engineering at Twente University. As part of his study, he had an internship at Shell Chimie in Berre l’Etang (France), where he worked on the optimisation of the distillation process of iso-propanol. He completed his study in 1991 with a Master thesis on the experimental determination of the kinetics of hydrogenation of cyclohexene in a LaNi5 hydride slurry. After that he followed a post-graduate design course (“tweede fase opleiding”) “Process Technology” at Twente University in cooperation with Groningen State University in the period 1992-1994. He completed this post-graduate course with a Master thesis on experimental research towards alternative synthesis routes of aramide, a work done at Groningen State University in close cooperation with AKZO Delfzijl. After this study, he was appointed research assistant at the University of Stuttgart on a 1-year project (1994-1995) in the framework of the Human Capital and Mobility training programme of the European Union. In this period he worked in the field of application of metal hydrides in CFC-free absorption air-conditioning systems. From 1996-2001, he was appointed as a promovendus at Delft University in the section Thermal Power Engineering. From August 2001 he was appointed toegevoegd onderzoeker in this section, where he works on a variety of projects with focus on experimental research in the field of combustion and gasification systems.

283