Waste Management 29 (2009) 934–944
Contents lists available at ScienceDirect
Waste Management journal homepage: www.elsevier.com/locate/wasman
Life cycle assessment for optimising the level of separated collection in integrated MSW management systems L. Rigamonti *, M. Grosso, M. Giugliano DIIAR – Environmental Section, Politecnico di Milano, P.zza Leonardo da Vinci, 32 - 20133 Milano, Italy
a r t i c l e
i n f o
Article history: Accepted 7 June 2008 Available online 5 August 2008
a b s t r a c t This life cycle assessment study analyses material and energy recovery within integrated municipal solid waste (MSW) management systems, and, in particular, the recovery of the source-separated materials (packaging and organic waste) and the energy recovery from the residual waste. The recovery of materials and energy are analysed together, with the final aim to evaluate possible optimum levels of source-separated collection that lead to the most favourable energetic and environmental results; this method allows identification of an optimum configuration of the MSW management system. The results show that the optimum level of source-separated collection is about 60%, when all the materials are recovered with high efficiency; it decreases to about 50%, when the 60% level is reached as a result of a very high recovery efficiency for organic fractions at the expense of the packaging materials, or when this implies an appreciable reduction of the quality of collected materials. The optimum MSW management system is thus characterized by source-separated collection levels as included in the above indicated range, with subsequent recycling of the separated materials and energy recovery of the residual waste in a large-scale incinerator operating in combined heat and power mode. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Life cycle assessment (LCA), originally developed for assessing environmental impacts of products, processes and activities with the so-called ‘‘cradle to grave” approach, has evolved in the last few years toward extended applications related to a broader range of human activities involving environmental interactions, such as waste management, treatment and disposal operations. LCA is becoming a tool commonly utilised for decisionmaking related to alternative waste management strategies (Finnveden, 1999; Rebitzer et al., 2004), but only a few studies have analysed municipal solid waste (MSW) management from a systems perspective (AEA, 2001; Eriksson et al., 2005; Heilmann and Winkler, 2005; Profu, 2004; Thorneloe et al., 2005). The main conclusion of all these studies is that reduced landfilling in favour of increased recycling of energy and materials leads to lower environmental impact and lower consumption of energy resources. On the basis of this result, we have analysed, from an energetic and environmental point of view, material and energy recovery within integrated MSW management systems, with the final aim to evaluate possible optimum levels of source-separated collection that lead to the most favourable energetic and environmental results. Trying thus to identify an optimum configuration of the MSW management system, this
* Corresponding author. Tel.: +39 02 23996415; fax: +39 02 23996499. E-mail address:
[email protected] (L. Rigamonti). 0956-053X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2008.06.005
LCA study analyses together the recovery of source-separated materials (i.e., the recycling of iron, aluminium, glass, paper, wood and plastic, and the composting of food waste and green fraction) and the high efficiency energy recovery from the residual waste (i.e., the incineration with production of electricity and heat). 2. Methodology In order to quantify the real energetic and environmental balance of the recycling of materials source-separated from MSW and of the energy recovery from the residual waste, the technique of LCA is used. This means taking into account that any recycling activity influences the environment by consuming resources and releasing emissions and waste streams, and by replacing conventional products from primary production, i.e., the production from virgin raw materials. Moreover, the energy recovered from the residual waste displaces the same quantity of energy produced in conventional power plants and boilers fuelled with fossil fuels. Standards ISO 14040 (2006) and ISO 14044 (2006) define the four basic steps of the assessment procedure, well described and commented in Rebitzer et al. (2004) and in Pennington et al. (2004): a. Goal and scope definition, which includes the preliminary assumptions about the aim of the study, the functional unit and the boundaries of the system.
935
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
b. Life cycle inventory (LCI), which focuses on the quantification of mass and energy fluxes. c. Life cycle impact assessment (LCIA), where the environmental impact of the activity is assessed with the use of impact indicators. d. Life cycle interpretation, which aims at evaluating possible changes or modifications of the system that can reduce its environmental impact. The Simapro 7 software, developed by PRè Consultants (2006a, b, c), is used for the evaluation of the energetic and environmental impacts of the various processing steps. Two characterisation methods have been chosen: the cumulative energy demand (CED) (Jungbluth and Frischknecht, 2004) and the CML 2 (CML, 2001). The first one is used to calculate the total energy demand of the activity under study. In fact, the CED method investigates the energy use throughout the life cycle of the analysed system, including direct as well as indirect consumptions of energy due to, e.g., the production of additives or construction materials. The CML 2 method, slightly modified in this study, is applied to evaluate the environmental impacts. In particular, the following environmental impact categories have been selected: – Global warming potential (GWP), which accounts for the emission of greenhouse gases; – Human toxicity potential (HTP), which addresses a wide range of toxic substances, including, in this study, the secondary particulate matter; – Acidification potential (AP), which accounts for the emissions of NOx, SOx and ammonia; – Photochemical ozone creation potential (POCP), which accounts for the substances that cause the photochemical ozone production in the troposphere. Finally, the effects of the variation of the most important input parameters on the results are evaluated and discussed, and in particular (Rigamonti, 2007) the role of very high recovery of organic fractions, the effects of the possible decrease of the quality of the recovered material when very high levels of source-separated collection are pursued, and the effects of assuming different types of conventional power plants for the evaluation of saved primary energy.
Three MSW integrated management systems are analysed (Fig. 1). They differ from each other in the quantities of waste sent to material recovery and to energy recovery, based on three different scenarios of source-separated collection (Table 1): – Scenario 35%, characterized by a source-separated collection of about 35%: this is the current target for Italy (year 2007), despite the fact that the actual average of source-separated collection of recyclables and compostable materials in 2005 was equal only to 24.3% of the total Italian MSW production (APAT-ONR, 2006); – Scenario 50%, characterized by a source-separated collection of about 50%: this level has been reached in recent years in some provinces in the North of Italy (APAT-ONR, 2006); – Scenario 60%, characterized by a source-separated collection of about 60%: we have considered this as a reasonable target level that can be reached in the medium term at the provincial scale (at least in the North and Centre of Italy). The composition of gross MSW was calculated based on several analyses and represents the Italian average (Rigamonti, 2007). The fractions collected separately are delivered to material recovery processes, whereas the residual waste is destined to energy recovery. Material recovery includes the recycling of packaging materi-
Table 1 Scenarios analysed: quantity collected for each fraction, expressed in kg per tonne of gross MSW produced Fractions
• 35% • 50%
Scenario 35% kg tMSW
Paper Wood Plastic Glass and inert material Metals without Al Aluminium Food waste Green waste Other Total collected Total residual waste
Steel Aluminium Glass Paper Wood Plastic
Source-separated collection level of: MSW
3. Integrated MSW management systems analysed
1
103 14 29 41 8 1 69 52 31 348 652
Scenario 50% kg tMSW 191 16 44 41 13 1 115 52 31 503 497
RECYCLING (substitution of primary production)
• 60% Green and food waste
Residual waste
COMPOSTING (substitution of peat and mineral fertilizers)
INCINERATOR: LARGE PLANT ONLY ELECTRICITY (substitution of a power plant fed with a mix of fossil fuels / coal / natural gas) LARGE PLANT CHP (substitution of boiler fed with natural gas + power plant fed with a mix of fossil fuels / coal / natural gas) SMALL PLANT CHP (substitution of boiler fed with natural gas + power plant fed with a mix of fossil fuels / coal / natural gas) Fig. 1. Integrated MSW management systems analysed.
1
Scenario 60% kg tMSW 191 21 66 48 13 3 160 69 31 601 399
1
936
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
als (iron, aluminium, glass, paper, wood and plastic) and the composting of food waste and green fraction. 4. LCA of material and energy recovery For the packaging materials, we have assumed that 1 kg of secondary material (produced from recycled materials) displaces 1 kg of the corresponding primary material (produced using virgin raw materials). In this sense, we do not consider the possible degradation of the material during the recycling process, which might lead to the consequence that the quality of the secondary material is worse than that of the primary material. Moreover, we do not consider that the recycled material can also compete on the market with materials of other types: for example recycled plastics can replace timber or concrete in structural items, as discussed by Ekvall (2000) and by Ekvall and Finnveden (2001). The compost obtained from food waste and green fraction is used as a substitute for peat and mineral fertilizers (AEA, 2001; Centemero and Caimi, 2002; Eriksson et al., 2005; Finnveden et al., 2005; Sonesson et al., 2000). Moreover, the application of compost as organic fertilizer promotes over time a build up of carbon in the soil that could prove to be a powerful ‘‘sink” (Barth and Favoino, 2005). Linzner and Mostbauer (2005) tried to give an estimation of this potential carbon sequestration. They concluded that the sequestered amount of carbon is 213 or 133 kg CO2 eq. per t of compost, in the hypothesis that the residual carbon after 50 years is 40% or 25%, respectively, of the original amount. However, due to the high uncertainty associated to these values, we have not included the carbon sequestration contribution in our analysis. The residual waste is sent to energy recovery in a waste-to-energy (WTE) plant. For the energetic and environmental balances, we have assumed that the electricity produced from the incinera-
tor displaces the same quantity of electricity produced by the thermoelectric Italian mix, composed by coal at 20%, fuel oil at 20%, natural gas at 20% and natural gas in a combined cycle at 40%. When combined heat and power (CHP) operation is considered, the heat produced displaces the same quantity of heat generated by household boilers fed with natural gas (thermal efficiency = 87%). 4.1. LCA of material recovery 4.1.1. Inventory: material flows, energy consumptions and emissions For the organic waste, data about emissions, energy consumptions and material flows have been gathered for the composting activity and the production of peat and mineral fertilizers. We have assumed that one can obtain 30 kg of compost starting from 100 kg of food and green waste (CITEC, 2004), with an electrical consumption of 50 kWh per tonne to be treated (Scaglia et al., 2004). Gaseous emissions are treated with biofilters. We have assumed that 34% of the produced compost is used in garden centres in substitution of peat, 62% in agriculture in substitution of mineral fertilizers with the same content of nutrients (N, P and K) and 4% in environmental restorations without substituting anything (Centemero, 2006). Data about the production of peat and mineral fertilizers were found in the Ecoinvent database (Swiss Centre for Life Cycle Inventories, 2004). For packaging materials, data about emissions, energy consumptions and material flows have been gathered for both the production from waste materials and from primary raw materials. While the latter are easily available from the literature and from international databases such as Ecoinvent and BUWAL250 (PRè Consultants, 2006d), the former were acquired mainly from direct contacts with the operators of the most important recycling plants
Table 2 Energy consumptions for materials recycling, expressed per tonne of recycled material (R-material) produced (for wood expressed per m3 of particle board produced and for plastic expressed per tonne of ‘‘total plastic”) Source of the datab Plant Literature
Steel recycling Electrical energy (Pre-treatment) Electrical energy (Melting) Total energy
Energy consumptions per tonne of R-steel produced 71 kW h 600 kW h 671 kW h (=6357 MJ)a
Aluminium recycling Electrical energy (Pre-treatment) Electrical energy (Melting) Thermal energy (Pre-treatment) Thermal energy (Melting) Total energy
Energy consumptions per tonne of R-Al produced 69 kW h 10 kW h 845 MJ (from natural gas) 4040 MJ (from natural gas) 5633 MJa
Literature Plant Plant Plant
Glass recycling Electrical energy (Pre-treatment) Thermal energy (Melting) Total energy
Energy consumptions per tonne of R-glass produced 18.4 kW h 5460 MJ (from fuel oil) 5634 MJa
Plant Literature
Paper recycling Electrical energy Thermal energy Total energy
Energy consumptions per tonne of R-pulp produced 7 kW h 15 MJ (from diesel) 81 MJa
Literature Literature
Wood recycling Electrical energy (Pre-treatment) Electrical energy (Production of particleboard) Thermal energy (Production of particleboard) Total energy
Energy consumptions per m3 of particleboard produced 36 kW h 95 kW h 239 MJ from fossil fuel + 2147 MJ from wood 3627 MJ a
Literature Literature Literature
Plastic recycling Electrical energy (Pre-treatment) Electrical energy (Recovery) Thermal energy (Pre-treatment) Thermal energy (Recovery) Total energy
Energy consumptions per tonne of ‘‘total plastic” 136 kW h 278 kW h 451 MJ (from diesel) 1840 MJ (from natural gas) 6212 MJa
Literature Literature Literature Literature
a b
An average electrical efficiency equal to 38% is used for the conversion of kWh in MJ (IPPC, 2006). ‘‘Plant”: data from direct contacts with the operators of the most important recycling plants located in the North of Italy; ‘‘literature”: data from literature.
937
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
located in the North of Italy. In addition, the reference documents on best available techniques (BAT), issued by the IPPC Bureau of the European Union, were utilised as a source of information (IPPC, 2001a, b, c). A brief description of the recycling of each packaging material is reported in the following paragraphs, including the basic data about mass balances and energy consumptions. 4.1.1.1. Steel. The reprocessing of scrap ferrous metal is a wellestablished industry. Households are a significant but relatively minor source of ferrous scrap, mainly in the form of tin cans. During recycling, these components are first shredded and then a magnetic separator is used to remove impurities (typically paper, plastics and non-ferrous metals) and to obtain separate ferrous metals, cleaned at 90–95% and ready to be sent to a steel smelter. The selection efficiency is equal to 80%, whereas the electric arc furnace efficiency, where the actual recycling takes place, is equal to 84%. Energy consumptions are shown in Table 2; altogether, the recycling requires 671 kW h per tonne of recycled steel. The production of steel, both from virgin raw materials and from scrap, releases air emissions that are taken into account in the environmental assessment (IPPC, 2001a; ENEA, 2002). 4.1.1.2. Aluminium. Most of the aluminium in the MSW stream derives from beverage cans. Magnetic and eddy current separation techniques can be employed to effectively remove ferrous metal from aluminium. The recovered aluminium undergoes a pre-treatment of pyrolysis and then it is melted in a rotary kiln fed with natural gas. The recycled aluminium is produced in the form of ingots, which are then sent to dedicated foundry for remelting. The selection efficiency is equal to 95%, whereas the melting efficiency is equal to 93%. The energy consumptions of the recycling activities are showed in Table 2. The production of aluminium, both from virgin raw materials and from scrap, releases gaseous emissions that are taken into account in the environmental assessment. In particular, we have used data from Ecoinvent database (Swiss Centre for Life Cycle Inventories, 2004) and the reference document on BAT in the non-ferrous metals industry (IPPC, 2001c) for emissions associated with primary production. We have used data from a state-of-theart Italian plant, Ecoinvent database and ENEA (2002) for emissions associated with secondary production. 4.1.1.3. Glass. The source-separated glass comes mainly in the form of food and beverage containers. This fraction includes both coloured and clear glass bottles. Glass recycling involves different activities such as manual selection, shredding, screening, magnetic and non-magnetic separation to remove impurities and inert materials (ceramics and gravels) and to obtain a proper size distribu-
tion. The glass cullet is then delivered to a glass manufacturing plant, where it is used in the production of new glass containers, together with ordinary virgin raw materials (silica, calcium carbonate, sodium hydroxide, additives). The presence of cullet, which is characterized by a lower melting temperature than virgin raw materials, allows the glass furnace to be operated at a lower temperature, thus leading to a significant savings of primary energy (up to 20% when 80% of cullet is utilised in the kiln feeding). The selection efficiency is equal to 94%, while the melting efficiency is equal to 100%. The energy consumptions of the recycling activities are shown in Table 2. We have assumed that the furnace would be fed with 83% of glass cullet and 17% of virgin raw materials, according to the current practice of the reference plants. The production of glass, both from virgin raw materials and from cullet, releases air emissions that are taken into account in the environmental assessment (IPPC, 2001b; Ecoinvent database; Glass Technology Services Ltd., 2004). 4.1.1.4. Paper. In this study, the production of pulp using recycled paper is compared to the production of thermo-mechanical pulp from wood. Virgin and recycled pulps are subsequently processed in essentially comparable ways, and so this stage was not considered in the LCA. Moreover, we have not included the phase of de-inking. To produce the recycled pulp, the source-separated paper undergoes a selection process, aimed to remove the impurities (like pieces of plastic), and then is sent to the pulper, where other residues are produced (ashes, sand and worn-out fibres). The overall activity has an average efficiency of 85.5%; the energy consumptions just for the selection are shown in Table 2 (Arena et al., 2004; AmbienteItalia – Comieco, 2003). It is important to underscore that paper fibres degrade in the recycling process, so they cannot be reused indefinitely. 4.1.1.5. Wood. Wood separated from MSW is mostly used for the production of particleboard. With this aim, wood is first shredded, then it undergoes a magnetic separation and finally it is reduced into chips. The efficiency of this pre-treatment is equal to 85.5%; the energy consumptions of the recycling activities are shown in Table 2 (Fruhwald and Hasch, 1999). In this study, the production of particleboard from wood source-separated from MSW is compared with the production of plywood from virgin material. 4.1.1.6. Plastic. Plastic materials include a wide range of different polymers. In this study we have considered the mechanical recycling of polyethylene terephthalate (PET), high density polyethylene (HDPE) and a mix composed by 57% low density
37.75% R-PET (75%) 12.25% residues 50% PET
recovery
10% HDPE
recovery
20% MIX
recovery
9% R-HDPE (90%) 100 plastic
1% residues
selection
20% residues
12% R- MIX (60%) 8% residues
Total residues = 41.25% Fig. 2. Mass balance of plastic recycling.
938
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
polyethylene (LDPE), 35% linear low density polyethylene (LLDPE) and 8% polypropylene (PP). The process of plastic recycling consists of separating the mixed plastic materials in the three fractions, PET, HDPE and the mix, using mainly near infra-red (NIR) detectors and manual sorting. Then, each stream of plastic undergoes a series of treatments, including pre-washing, manual separation, separation by X-ray and metal detectors, grinding, filtration, washing, flotation, drying and fine screening. At the end of the process, flakes or granules of the recycled polymers (R-PET, R-HDPE and R-mix) are obtained. The mass balance of plastic recycling is shown in Fig. 2. In this study, we have considered a product called ‘‘total plastic” made of R-PET (64.3%), R-HDPE (15.3%) and R-MIX (20.4%). The energy consumptions for the production of 1 tonne of ‘‘total plastic” are reported in Table 2 (Arena et al., 2003). 4.1.2. Results: recycling efficiencies Table 3 summarizes selection and recovery efficiencies of the materials analysed; the combination of these two values gives the overall recycling efficiency.
Table 3 Recycling efficiency (found from the combination of selection and recovery efficiency) for the materials analysed Material
Selection efficiency (% in weight) (A)
Recovery efficiency (% in weight) (B)
Recycling efficiency (% in weight) (A B)
Steel
80
67.2
Aluminium Glass Paper Wood
95 94 85.5 85.5
84 (melting furnace) 93 (melting kiln) 100 100 100
Plastic Food and green wastes
80 80
73.5 37.5 (composting)
88.35 94 85.5 85.5 (44.5 after drying) 58.75 30
Table 4 Energy savings when recycling instead of producing starting from virgin raw materials (values expressed: in MJeq per tonne produced, except for that of wood which is expressed in MJeq per m3 of particleboard produced, as percentage and in MJeq per tonne source-separated) Material
Steel Aluminium Glass Paper (pulp) Wood Plastic Compost (from food and green wastes) a
Saved energy MJeq per tonneproduced
%
MJeq per tonnesource-separated
27,176 187,834 6,424 42,044 29,438a 72,573 1,080
81.2 93.5 36.1 99.4 76.9 91.4 41.0
18,275 165,824 7,231 35,929 23,391 43,170 324
Expressed in MJeq per m3produced .
The recycling of glass is the most efficient because, starting from 100% of source-separated glass, it yields 94% of recycled glass. It is followed by the recycling of aluminium, which has a yield of about 88%. Composting is the least efficient material recovery process; the yield of this process is only 30%. 4.1.3. Results: cumulative energy demand The energy balance of material recovery is based on a comparison between the consumptions for recycling and those required for the production from virgin raw materials. In particular, using the CED characterization method, this consists in subtracting, for each material, the energy consumption associated with the production from virgin raw materials from that required by the recycling processes. Both direct and indirect energy consumptions are considered. The results of this operation are shown in Table 4. The main considerations that can be drawn are: For all the materials analysed, energy consumption for the virgin production is higher than for recycling; this means that recycling always allows energy savings. The highest savings is related to the aluminium recycling; this activity allows a savings of 187,834 MJeq per tonne produced (corresponding to 165,951 MJeq per tonne collected). The second process that allows a large energy saving is plastic recycling, with 72,573 MJeq saved per tonne produced (equal to 42,637 MJeq per tonne collected). If we express the energy savings in relative terms, the recycling of paper allows the highest savings, equal to 99%; this means that the production of pulp from recycled paper requires only 1% of the energy necessary for pulp production from wood. This is due to the fact that the energy consumptions for the growth and maintenance of the forest and for the production of fibres from wood are absent in pulp production from recycled paper. Aluminium and plastic recycling allows significant savings too, 94–91%. Finally, glass recycling and composting are the activities that allow for the lowest energy savings when compared to the corresponding primary production.
4.1.4. Results: environmental impact indicators Environmental assessment is performed with the same approach utilised for energetic balance: for each material, the emissions released during the production from virgin raw materials are subtracted from the emissions derived from the recycling processes. The assessment follows an LCA approach, including both direct and indirect emissions. Results are reported in Table 5, and lead to the following considerations: All the packaging materials show negative values for all the impact indicators; this means that the collection and recycling of 1 tonne of each of these materials with its substitution for virgin production is environmentally advantageous. The collection and recycling of aluminium is the process that allows the highest environmental advantages, for all the ana-
Table 5 Environmental impact indicators for material recovery (expressed per tonne of source-separated material) Per 1 source-separated tonne Global warming (kg CO2 eq.) Acidification (kg SO2 eq.) Human toxicity (kg 1,4-DCB eq.) Photochemical ozone creation (kg C2H4 eq.)
Steel 405 0.06 247 0.587
Aluminium 9855 52 47001 2.9
Glass 722 2.9 141 0.185
Paper
Wood
557 3.3 126 0.237
Note: a negative value indicates an advantage for the environment whereas a positive value indicates a disadvantage.
166 1.2 93 0.317
Plastic 1120 7.1 248 1.2
Food and green wastes 26.8 +0.07 +5.6 +0.025
939
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
lysed impact indicators. The benefit in human toxicity is even two orders of magnitude higher than that of plastic, iron and paper, the other packaging materials whose recycling results more convenient. This is mainly due to the avoided emissions from electrolysis, a basic process in the primary aluminium production, which is obviously not required when producing recycled aluminium. The composting of food waste and green fraction appears neutral from an environmental point of view. 4.2. LCA of the energy recovery from residual waste 4.2.1. Inventory and main hypotheses The residual waste is sent to energy recovery in a waste-to-energy (WTE) plant, without any further pre-treatment. Strategies based on the production of refuse derived fuel (RDF) and its subsequent combustion in a dedicated plant have not been considered here because a previous study (Consonni et al., 2005a,b) demonstrated that this is less efficient than the direct combustion of residual waste from an energetic, environmental and economic point of view. The lower heating value (LHV) of the residual waste was calculated based on the different levels of source separation hypothesised for the various scenarios. This equals 10,249 kJ per kg, 10,090 kJ per kg and 10,393 kJ per kg for scenarios 35%, 50% and 60%, respectively. We have considered three different WTE plants (Federambiente, 2005; Consonni et al., 2006): – a large plant, designed for a MSW management system of about 1,200,000 inhabitants, producing only electricity (yearly average net electrical efficiency = 28.8%) (LP); – a large plant, designed for a MSW management system of about 1,200,000 inhabitants, operating in CHP mode (LP CHP). We have assumed that the amount of steam sent to district heating equals 30% of the total flow entering the steam turbine (yearly average net electrical efficiency = 24.6%; yearly average net thermal efficiency = 19.2%);
– a small plant, designed for a MSW management system of about 200,000 inhabitants, that produces electricity and heat (SP CHP). We have assumed that the amount of steam sent to district heating equals 60% of the total flow entering the steam turbine (yearly average net electrical efficiency = 11.5%; yearly average net thermal efficiency = 40.2%). The three WTE plants are assumed to be representative of the state-of-the-art for combustion, energy recovery and flue gas treatment. The latter consists of a dry system that starts with an electrostatic precipitator, followed by a scrubbing with sodium bicarbonate and activated carbon, a fabric filter and a selective catalytic reduction reactor fed with ammonia for the control of nitrogen oxides. Stack concentrations are assumed to be the same for all the three WTE plants (Table 6), and they comply with the indication of the BAT Reference Document for waste incineration (IPPC, 2006). As most recent incinerators have emissions that are often significantly lower than those imposed by law, values in Table 6 are based, rather than on current legislation, on direct measurements carried out on state-of-the-art WTE plants operating in Italy. Emission factors of fossil and non-fossil CO2 were calculated based on the actual carbon content of the residual waste, by combining the elementary composition of each fraction with the percentage of each fraction present in the residual waste. This actual carbon content, which has been split between fossil (contained in plastics) and biogenic (contained in the food waste, green fraction, paper and wood), is equal to 294 kg, 288 kg and 296 kg per tonne of residual waste of respectively scenario 35%, 50% and 60%. As stated previously, the electricity produced from the WTE plant displaces the same amount of electricity produced by the thermoelectric Italian mix and the heat produced displaces the same amount of heat generated by household boilers fed with natural gas. As the assumptions on the saved primary energy are always an important factor in LCA of waste management systems (AEA, 2001; Björklund and Finnveden, 2005; Eriksson et al., 2005; Finnveden et al., 2005; Moberg et al., 2005; Profu, 2004; Sonesson et al., 2000; Thorneloe et al., 2005), in the sensitivity
Table 6 Concentrations of the main pollutants at the stack of the WTE plant and emission factors expressed per tonne of residual waste for each scenario Pollutants
Concentrations (11% O2, dry gas)
Emission factors Scenario 35%
mg mn NH3 CO PM10 HCl HF N2O TOC NOX (as NO2) SOX (as SO2)
2 10 2 2 0.2 2 3 50 2
lg mn Cd Hg Pb PAH
3
3
0.015 0.425 0.5 0.0025
0.01
CO2 fossil CO2 biogenic
– –
1
lg t 90 2555 3006 15
1
ng t
60 kg t
Scenario 50% 1
gt
12.0 60 12.0 12.0 1.20 12.0 18 301 12.0
ng I-TEQ mn 3 Dioxin
1
gt
1
421 656
1
gt
11.9 59 11.9 11.9 1.19 11.9 18 296 11.9 1
lg t 89 2518 2963 15
1
ng t
59 kg t
Scenario 60%
1
492 563
12.1 61 12.1 12.1 1.21 12.1 18 303 12.1 1
lg t 91 2576 3031 15
1
ng t
61 kg t
1
501 585
940
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
Scenario 35%
Scenario 50%
Scenario 60%
232
230
175
kg CO2 eq. per t of residual waste
160 120 95
114 89
75
46 21
15 8
6 -1
-25
-61 -67
-66
-77
-116
-125
-129
-143 -194 -225
mix Italy LP coal LP NGCC LP mix Italy LP CHP coal LP CHP NGCC LP CHP mix Italy SP CHP coal SP CHP NGCC SP CHP
-298 -315
-325
-332 -352 -378 -413
-425
Fig. 3. Variation of the Global warming indicator as a function of the kind of primary energy displaced by the electricity produced from the WTE plants (LP = large plant; SP = small plant; CHP = plant operating in combined heat and power mode). Note: a negative value indicates an advantage for the environment whereas a positive value indicates a disadvantage.
CUMULATIVE ENERGY DEMAND (MJeq t MSW-1) Large plant producing Large plant operating in a Small plant operating in a only electricity CHP way CHP way 0
0
35%
50%
0
35%
60%
35%
-4000
-9913
-9913 -11304
-11304
-6000
-8000
-4961
-9913
-8000
-10000
-3765
-12000
-11304
-6000
-5934
-10000
60%
-5929
-4000
-6000
50%
-2000
-5929
-5929 -4000
-12000
60%
-2000
-2000
-8000
50%
-5428
-10000 -12000
-4496
-3135
-4117 -3429
-3742
-14000
-14000
-14000
-16000
-16000
-16000
Energy recovery
Energy recovery
Energy recovery
Material recovery
Material recovery
Material recovery
Total: Scenario 35%: -10890 Scenario 50%: -13678 Scenario 60%: -14440
Total: Scenario 35%: -11863 Scenario 50%: -14409 Scenario 60%: -15046
Total: Scenario 35%: -11357 Scenario 50%: -14031 Scenario 60%: -14733
Fig. 4. Cumulative energy demand indicator for the three MSW management systems analysed and for the three types of WTE plant considered (the electricity produced from the WTE plant displaces that produced by the thermoelectric Italian mix). Note: A negative value indicates an advantage for the environment whereas a positive value indicates a disadvantage.
941
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
Table 7 Environmental impact indicators for the three MSW management systems analysed and for the three types of WTE plant considered (the electricity produced from the WTE plant displaces that produced by the thermoelectric Italian mix) Large plant producing only electricity
Large plant operating in a CHP way
Small plant operating in a CHP way
M
M
M
E
Global warming (kg CO2 eq. tMSW Scenario 35% 138 Scenario 50% 209 Scenario 60% 257
1
Total
E
Total
E
Total
) 40 +7 +2
Acidification (kg SO2 eq. tMSW 1) Scenario 35% 0.7 Scenario 50% 1.1 Scenario 60% 1.4
178 202 255
1.6 1.2 1.0
Human toxicity (kg 1,4 DCB eq. tMSW 1) Scenario 35% 71 Scenario 50% 101 Scenario 60% 183
138 209 257
2.3 2.3 2.4
91 74 62
0.7 1.1 1.4
162 175 245
Photochemical ozone creation (kg C2H4 tMSW 1) Scenario 35% 0.08 0.09 Scenario 50% 0.12 0.07 Scenario 60% 0.15 0.06
93 33 31 1.5 1.1 0.9
71 101 183
0.17 0.19 0.21
231 242 288 2.2 2.2 2.3
95 77 65
0.08 0.12 0.15
138 209 257 0.7 1.1 1.4
166 178 248
0.10 0.08 0.07
71 101 183
0.18 0.20 0.22
0.08 0.12 0.15
44 +4 0.4 0.8 0.6 0.5 84 69 58 0.08 0.06 0.06
182 205 257 1.5 1.7 1.9 155 169 240 0.16 0.18 0.21
Note: M: contribute of material recovery; E: contribute of energy recovery. A negative value indicates an advantage for the environment whereas a positive value indicates a disadvantage.
Table 8 Scenario ‘‘60% organics” in comparison with the other scenarios already examined: values indicate the percentage of collection on the total production of each fraction Fractions Paper Wood Plastic Glass and inert material Metals without Al Aluminium Food waste Green waste Other Total (% of collection)
Scenario 35% 40 30 20 70.5 40 14 30 60 100 34.8
Scenario 50% 74 35 30 70.5
Scenario 60% 74 45 45 83
61 19 50 60 100
74 35 30 70.5
61 45 70 80 100
50.3
Scenario ‘‘60% organics”
Clearly these two cases are representative of the ‘‘dirtiest” and of the ‘‘cleanest” ways of producing power from fossil fuels. In the LCA, in addition to the emissions at the stack of the WTE plants and to the avoided emissions due to the production of energy, we have considered emissions due to the production of steel and concrete used in the construction of the plants, of reagents used in the flue gas cleaning and of additives for inertisation of fly ashes, together with the avoided emissions related to the recycling of steel and aluminium separated from the bottom ashes.
61 19 80 90 100
60.1
– case ‘‘NGCC”: the electricity produced from the WTE plant displaces the same quantity of electricity produced by a combined cycle power plant fed with natural gas (net electrical efficiency = 55%).
59.7
analysis we have considered two further hypotheses for the substitution of electricity: – case ‘‘coal”: the electricity produced from the WTE plant displaces the same quantity of electricity produced by a conventional power plant fed with coal (net electrical efficiency = 36.63%);
4.2.2. Results: cumulative energy demand and environmental impact indicators The CED of the three different types of WTE plants for the three different scenarios is obviously negative (meaning a savings) due to the energy produced by the combustion. The most convenient option is, for all the three scenarios, the combustion of the residual waste in a large plant operating in combined heat and power mode.
Global warming Scenario 35%
Scenario 50%
Scenario 60%
Scenario 60% organics
M SW
-50
kg CO 2 eq. per t
0
-100 -150 -200
-177
-193
-202 -250 -255 -300
Fig. 5. Global warming indicator for the scenario ‘‘60% organics” in comparison with that of the other scenarios (the WTE plant is the large one and produces only electricity that displaces that produced by the thermoelectric Italian mix).
942
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
Fig. 3 reports the global warming indicator for the three scenarios, the three WTE plants and the three different hypotheses about the displaced electricity. We can conclude that, even considering the other environmental indicators not shown here, the incineration with energy recovery of the residual waste, in comparison with the production of the same amount of energy from fossil fuels, is environmentally convenient when the replaced electricity is produced from coal or from a mix of fossil fuels (20% oil, 20% coal, 20% natural gas and 40% natural gas used in a combined cycle), whereas it is not environmentally convenient when the displaced electricity is produced from natural gas in a combined cycle plant. 5. LCA of the integrated MSW management systems The combination of the LCA results obtained for material recovery with those for energy recovery allows the calculation of the LCA for the whole MSW management systems analysed. Fig. 4 and Table 7 show the calculated indicators for the three MSW management systems analysed and for the three types of WTE plants assumed, when the electricity produced displaces the production of the thermoelectric Italian mix. For the three MSW management systems, all the impact indicators have a negative value; this means that the MSW management systems analysed are energetically and environmentally advantageous in comparison with the conventional method of material and energy production. In particular, the MSW management system more convenient is the one characterized by a source-separated collection of 60%. 6. Sensitivity analyses applied to the whole MSW management system 6.1. The role of very high recovery of the organic fractions The value of 60% of source-separated collection can be reached by collecting the different fractions as reported in Table 1, but there are obviously several other possibilities. Table 8 reports the percentage of collection of each fraction that was assumed in the analysis for the scenarios 35%, 50% and 60% (columns 1, 2 and 3). In column 4, an alternative option of
Table 9 Hypothesis on the increase of residues produced in plastic recycling (scenario ‘‘60% plastic”) Scenario
% of collection of plastic (on the total of the produced plastic) (Table 1)
Residues from selection
Total recycling residues (from selection + recovery)
35% 50% 60% 60% plastic
20 30 45 45
20% 20% 20% 45%
41% 41% 41% 60%
source-separated collection is described (scenario ‘‘60% organics”), characterized by a very high yield of the food and green wastes. Fig. 5 shows the variation of the global warming indicator among the different scenarios; the trend is the same even for all the other indicators, not shown here, whatever type of WTE plant one considers. The results thus show that the new scenario ‘‘60% organics” is located between the scenario 50% and the scenario 35%. This means that a very high collection of the organic fractions for their composting is less advantageous than a recovery at average levels (scenarios 60% and 50%). A source-separated collection of 50% appears then to be more advantageous than a source-separated collection of 60%, when the latter is obtained due to a very high efficiency recovery for food and green fractions at the expense of the other materials. 6.2. Possible decrease of the quality of collected materials when very high levels of source-separated collection are pursued It is likely that the quality of the collected fractions decreases when very high levels of source-separated collection are pursued. This is due to the necessity of collecting a higher amount of material, which might include fractions that are more contaminated or associated with other components and so more difficult to be recycled. In this sensitivity analysis, the effects of the decrease of the quality of the collected material are examined for plastic that, among the six packaging materials considered, is the one characterized by the highest production of residues during the recycling process. Table 9 shows, for each scenario analysed, the percentage of collection of plastic and the hypothesis about the production of residues during its recycling (phases of selection and recovery). Moreover, a new scenario is introduced; this is the same as scenario 60%, but the residues produced during the selection of the collected plastic are increased from 20% to 45%. Consequently, in this case, the total residues from plastic recycling are 60% instead of 41%. The LCA of this new scenario shows the energetic and environmental benefits of the collection and of the following recycling decrease (Table 10). In particular, this worsening is between 2%, for human toxicity indicator, and 12%, for photochemical ozone creation indicator. Indeed, the scenario ‘‘60% plastic” is less advantageous than the scenario 50% in the indicators of cumulative energy demand, acidification and photochemical ozone creation. This means that reaching a source-separated collection of the 50% is more efficient than reaching a source-separated collection of the 60%, if the latter implies a decrease of the quality of the collected material. This is even more true if we consider that this result were obtained assuming the decrease in the quality of the collected plastic only. The worsening of the impact indicators is more evident if we assume the decrease in the quality of all the collected materials. For example, if we assume that the selection efficiency of the packaging materials is the one reported in Table 3 reduced by 10% (25% for plastic), the consequent worsening of the impact indicators is between 11% (for acidification) and 24% (for human toxicity); moreover, if the reduction of the selection efficiency is 20% (again
Table 10 Variation of the impact indicators due to the decrease of the quality of the plastic collected (scenario ‘‘60% plastic”) (the WTE plant is the large one and produces only electricity that displaces that produced by the thermoelectric Italian mix)
CED Global warming Acidification Human toxicity Photochemical ozone creation
Per t of MSW
Scenario 35%
Scenario 50%
MJ eq kg CO2 eq kg SO2 eq kg 1,4-DCB eq kg C2H4 eq
10890 177 2.31 162 0.168
13678 202 2.32 175 0.188
Scenario 60% (A) 14440 255 2.38 245 0.209
Scenario ‘‘60% plastic” (B) 13542 231 2.23 240 0.184
D% (B-A) 6.2 9.4 6.3 2.2 12
943
L. Rigamonti et al. / Waste Management 29 (2009) 934–944 Scenario 35%
Scenario 50%
Scenario 60%
-10500
Scenario 35%
Scenario 50%
Scenario 60%
-160 -10890
-11000
-11357
-180
-11500
-193 -195
-12500 -13000 -13451 -13516
-13500 -13678
-14000
-14031 -14409
-14500
-13835 -13838 -14148 -14193
kg CO2 eq. per tMSW
-11863
-12000
MJ eq. per tMSW
-177 -181
-200
-202 -205
-220 -230 -231 -234
-231
-240
-242 -255 -258 -264
-260
-14440 -14733
-15000
-15046
-280 -288
-15500 LP LP plastic LP organics
LP CHP LP CHP plastic LP CHP organics
SP CHP SP CHP plastic SP CHP organics
-300 LP LP plastic LP organics
LP CHP LP CHP plastic LP CHP organics
SP CHP SP CHP plastic SP CHP organics
Fig. 6. Variation of the CED and of the global warming indicator with the type of WTE plant for all the scenarios analysed (35%, 50%, 60%, ‘‘60% organics”, ‘‘60% plastic”) when the electricity produced by the WTE plant displaces that produced by the thermoelectric Italian mix.
25% for plastic), the worsening of the impact indicators is between 15% (for acidification) and 30% (for human toxicity). In this sense, a very high source-separated collection of the packaging materials might be not so effective.
a major role in defining the optimum balancing between material and energy recovery.
7. Conclusions
The authors wish to thank the operators of the recycling plants that have supplied most of the primary data utilised in the analysis and the packaging consortia (CiAl, CNA, CoReVe, Comieco, Corepla and Rilegno) for their useful advice.
Fig. 6 shows, for all the MSW management systems analysed (scenarios 35%, 50%, 60%, ‘‘60% organics”, and ‘‘60% plastic”), the variation of the CED and of the global warming indicator with the type of WTE plant considered when the electricity produced displaces the production of the thermoelectric Italian mix. Scenarios ‘‘60% organics” and ‘‘60% plastic” turn out to be worse than scenario 50% for the CED indicator; moreover, scenario ‘‘60% organics” is worse than scenario 50% also for the GWP indicator. We can thus conclude that the combination of the results for material recovery and those for energy recovery together with the indications of the sensitivity analysis allows the identification of the optimum level of source separation, which is: – about 60%, when all the materials are recovered with high efficiency (70% paper, 40–50% wood, plastic and aluminium, 80% glass, 60% iron, 70% food waste, 80% green fraction); – about 50%, when the level of 60% is reached due to a very high efficiency recovery for food waste and green fraction (food waste 80% and green fraction 90%) at the expense of the other materials, or when the level of 60% is reached due to a high efficiency recovery of all the materials but with a reduction of the quality of the collected materials. Under the hypotheses considered, the optimum MSW management system is thus characterized by a source separation level as above indicated, with subsequent recovery of the separated materials and energy recovery of the residual waste in a large-scale WTE plant operating in a combined heat and power mode. Moreover, when a decision has to be made on how much to increase the overall source separation level in integrated waste management systems, the efficiency of energy recovery from residual waste plays
Acknowledgements
References AEA, 2001. Waste management options and climate change. Final Report to the European Commission, DG Environment.
. AmbienteItalia – Comieco, 2003. Studio su consumi energetici della raccolta e della selezione di carta e cartone (Study about Energy Consumptions of Paper Collection and Selection). . APAT-ONR, 2006. Rapporto rifiuti. . Arena, U., Mastellone, M.L., Perugini, F., 2003. Life cycle assessment of a plastic packaging recycling system. International Journal of Life Cycle Assessment 8 (2), 92–98. Arena, U., Mastellone, M.L., Perugini, F., Clift, R., 2004. Environmental assessment of paper waste management options by means of LCA methodology. Industrial & Engineering Chemistry Research 43, 5702–5714. Barth, J., Favoino, E., 2005. Composting biowaste: state of the arts, trends, drivers in environmental policy. In: Proceedings of the 1st BOKU Waste Conference, pp. 131–141, ISBN 3-85076-721-3. Björklund, A., Finnveden, G., 2005. Recycling revisited – life cycle comparisons of global warming impact and total energy use of waste management strategies. Resources, Conservation and Recycling 44, 309–317. Centemero, M., 2006. Il compostaggio nella gestione dei rifiuti urbani (Composting within municipal waste management). In: Il Compost di Qualità, Annuario 2006/2007- V edizione, Ed. Il Verde editoriale. Centemero, M., Caimi, V., 2002. Impieghi del compost: settori di maggior rilevanza, modalità d’uso, scenari attuali di mercato (Compost utilisation). In: Proceedings: Sep pollution 2002: Il compostaggio in Italia, Maggioli Editore, pp. 35–67. CITEC, 2004. Guidelines for the design, production and running of high technology plants for the disposal of urban waste. CITEC Working Group, General Coordinator Aulo Magagni. GEVA Edizioni. CML, Bureau B&G, School of System Engineering, Policy Analysis and Management – Delft University of Technology, 2001. Life Cycle Assessment: an Operational Guide to the ISO Standards. Consonni, S., Giugliano, M., Grosso, M., 2005a. Alternative strategies for energy recovery from municipal solid waste. Part A: mass and energy balances. Waste Management 25, 123–135.
944
L. Rigamonti et al. / Waste Management 29 (2009) 934–944
Consonni, S., Giugliano, M., Grosso, M., 2005b. Alternative strategies for energy recovery from municipal solid waste. Part B: Emission and cost estimates. Waste Management 25, 137–148. Consonni, S., Giugliano, M., Grosso, M., Rigamonti, L., 2006. Energy and environmental balances of energy recovery from municipal solid waste with and without RDF production. In: Proceedings of the Biomass and Waste to Energy Symposium, Venice, November 29 to December 1. Ekvall, T., 2000. A market-based approach to allocation at open-looping recycling. Resource, Conservation and Recycling 29, 91–109. Ekvall, T., Finnveden, G., 2001. Alllocation in ISO 14041 – a critical review. Journal of Cleaner Production 9, 197–208. ENEA, Ministero dell’Ambiente e della Tutela del Territorio, Associazione Industriale Bresciana, 2002. Valutazione delle emissioni di inquinanti organici persistenti da parte dell’industria metallurgica secondaria (Assessment of the Emissions of Persistent Organic Pollutants from Secondary Metallurgy Industry). Eriksson, O., Carlsson Reich, M., Frostell, B., Björklund, A., Assefa, G., Sundqvist, J.-O., Granath, J., Baky, A., Thyselius, L., 2005. Municipal solid waste management from a system perspective. Journal of Cleaner Production 13, 241–252. Federambiente, 2005. Bilancio ambientale, energetico ed economico del recupero di energia da rifiuti urbani mediante produzione di CDR e co-combustione in impianti non dedicati (Environmental, Energetic and Economic Assessment of Energy recovery from Municipal Waste Based on RDF co-incineration in Existing inDustrial Plants). A cura di Consonni S., Giugliano M., Grosso M., Rigamonti L. Finnveden, G., 1999. Methodological aspects of life cycle assessment of integrated solid waste management systems. Resources, Conservation and Recycling 26, 173–187. Finnveden, G., Johansson, J., Lind, P., Moberg, A., 2005. Life cycle assessment of energy from solid waste – Part 1: general methodology and results. Journal of Cleaner Production 13, 213–229. Fruhwald, A., Hasch, J., 1999. Life cycle assessment of particleboards and fibreboards. In: Proceedings Second European Wood-Based Panel Symposium, Hannover. Glass Technology Services Ltd., 2004. A study of the Balance between Furnace Operative Parameters and Recycled Glass in Glass Melting Furnaces . Heilmann, A., Winkler, J., 2005. Influence of the source separation efficiency of recyclable materials on the environmental performance of municipal waste management systems. In: Proceedings Sardinia 2005, Tenth International Waste Management and Landfill Symposium; S. Margherita di Pula, Cagliari, 3–7 October 2005. IPPC, 2001a. Best Available Techniques Reference Document on the Production of Iron and Steel. . IPPC, 2001b. Reference Document on Best Available Techniques in the Glass Manufacturing Industry. . IPPC, 2001c. Reference Document on Best Available Techniques in the Non-Ferrous Metals Industry. .
IPPC, 2006. Reference Document on Best Available Techniques for Waste Incineration. . ISO, 2006. ISO 14040: Environmental Management - Life Cycle Assessment Principles and Framework. ISO, 2006. ISO 14044: Environmental Management - Life Cycle Assessment Requirements and Guidelines. Jungbluth, N., Frischknecht, R., 2004. Implementation of Life Cycle Impact Assessment Methods. Ecoinvent Report No. 3; <www.ecoinvent.ch>. Linzner, R., Mostbauer, P., 2005. Composting and its impact on climate change with regard to process engineering and compost application – a case study in Vienna. In: Proceedings Sardinia 2005, Tenth International Waste Management and Landfill Symposium; S. Margherita di Pula, Cagliari, 3–7 October. Moberg, A., Finnveden, G., Johansson, J., Lind, P., 2005. Life cycle assessment of energy from solid waste – part 2: landfilling compared to other treatment methods. Journal of Cleaner Production 13, 231–240. Pennington, D.W., Potting, J., Finnveden, G., Lindeijer, E., Jolliet, O., Rydberg, T., Rebitzer, G., 2004. Life cycle assessment Part 2: Current impact assessment practice. Environment International 30, 721–739. PRè Consultants, 2006a. SimaPro 7: Introduction into LCA. . PRè Consultants, 2006b. SimaPro 7: Tutorial. . PRè Consultants, 2006c. SimaPro 7: Database Manual – Methods Library. . PRè Consultants, 2006d. SimaPro 7: Database Manual – The BUWAL 250 Library. . Profu, 2004. Evaluating waste incineration as treatment and energy recovery method from an environmental point of view. Final version 2004-05-13. <www.profu.se>. Rebitzer, G., Ekvall, T., Frischknecht, R., Hunkeler, D., Norris, G., Rydberg, T., Schmidt, W.P., Suh, S., Weidema, B.P., Pennington, D.W., 2004. Life cycle assessment Part 1: Framework, goal and scope definition, inventory analysis, and applications. Environmental International 30, 701–720. Rigamonti, L., 2007. Valutazione dei percorsi di recupero di materiali e di energia in sistemi integrati di gestione dei rifiuti urbani (Assessment of material and energy recovery pathways within integrated waste management systems). Tesi di dottorato XIX Ciclo, Politecnico di Milano. Scaglia, B., Tambone, F., Centemero, M., Favonio, E., 2004. Il compostaggio dei rifiuti urbani (Urban waste composting). Quaderno di Ingegneria Ambientale No. 40: I processi aerobici per il trattamento dei rifiuti urbani, a cura di F. Adani, pp. 26–46. Sonesson, U., Björklund, A., Carlsson, M., Daremo, M., 2000. Environmental and economic analysis of management systems for biodegradable waste. Resources, Conservation and Recycling 28, 29–53. Swiss Centre for Life Cycle Inventories, 2004. Ecoinvent: The Life Cycle Inventory Data, Version 1.1. Thorneloe, S.A., Weitz, K.A., Jambeck, J., 2005. Moving from solid waste disposal to materials management in the United States. In: Proceedings Sardinia 2005, Tenth International Waste Management and Landfill Symposium; S. Margherita di Pula, Cagliari, 3–7 October 2005.