Selective Production H2 Via Oxidative Steam Reforming

ARTICLE IN PRESS

Chemical Engineering Science

(

)

– www.elsevier.com/locate/ces

Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts Sanjay Patel a,b , K.K. Pant a,∗ a Department of Chemical Engineering, Indian Institute of Technology, Hauz Khas, New Delhi 110016, India b Department of Chemical Engineering, Institute of Technology, Nirma University of Science and Technology, Ahmedabad 382481, India

Received 15 June 2006; received in revised form 19 January 2007; accepted 29 January 2007

Abstract The oxidative steam reforming of methanol (OSRM) was carried out to produce the hydrogen selectively for polymer electrolyte membrane (PEM) fuel cell applications over Cu–Zn–Ce–Al oxide and Cu–Zn–Al oxide catalysts of varying compositions prepared by co-precipitation method. Catalyst performance was evaluated in a packed bed reactor over a wide range of operating conditions, and reaction parameters were optimized in order to maximize the hydrogen production with minimum carbon monoxide formation. The incorporation of ceria in Cu–Zn–Al oxide catalysts enhanced the activity greatly compared to without it. The Cu/Zn/Ce/Al:30/20/10/40 exhibited 100% methanol conversion and 244 mmol s−1 kg−1 cat hydrogen rate at 553 K with carbon monoxide as low as 995 ppm, which reduces the load on preferential oxidation of CO to CO2 significantly before feeding the hydrogen rich stream to the PEM fuel cell as a feed. Ceria had improved the dispersion and specific surface area of copper in multi-component Cu–Zn–Ce–Al oxide catalysts which were confirmed by the physicochemical properties, X-ray diffraction (XRD), temperature programmed reduction (TPR) and CO chemisorption studies. The chemisorption studies were performed at 193 K in order to hinder the spillover of carbon monoxide to ceria. The time-on-stream stability test had shown Cu–Zn–Ce–Al oxide catalysts as more stable compared to Cu–Zn–Al oxide catalysts. The amount of carbon deposited onto the catalysts was determined using TG/DTA thermogravimetric analyzer and the type of carbon species were identified using C1s X-ray photoelectron spectroscopy (XPS) spectra. 䉷 2007 Elsevier Ltd. All rights reserved. Keywords: Hydrogen; OSRM; Ceria; PEM fuel cell

1. Introduction Hydrogen is expected to play a major role in the future as a carbon free energy carrier. Its use in the vehicles via polymer electrolyte membrane (PEM) fuel cells can offer the nontoxic tail-pipe emissions and high over all efficiency compared to conventional internal combustion engines (Edwards et al., 1998; Bowers et al., 2006). The on-board storage of hydrogen with high energy density is facing some technical problems (Breen and Ross, 1999; Raimondi et al., 2002). One promising solution to this can be an on-board production of hydrogen using liquid hydrocarbons like methanol (Patel and Pant, 2006c; Yan et al., 2006), ethanol (Sahoo et al., 2007), ∗ Corresponding author. Tel.: +91 11 26596172; fax: +91 11 26581120.

E-mail address: [email protected] (K.K. Pant).

dimethyl ether (Semelsberger et al., 2006), etc. Methanol offers several advantages for the hydrogen production compared to other liquid organics (Velu et al., 2001; Patel and Pant, 2006a). There are different routs available for the hydrogen production from methanol as follows. One is partial oxidation of methanol (POM), CH3 OH + 0.5O2 ←→ 2H2 + CO2 ,

H 0 = −192 kJ mol−1 . (1)

This is a highly exothermic reaction which leads to the problem of heat removal and reactor temperature control, and also produces significant amount of CO (Wang et al., 2003). Another route is the steam reforming of methanol (SRM), CH3 OH + H2 O ←→ 3H2 + CO2 ,

H 0 = 49.5 kJ mol−1 . (2)

0009-2509/$ - see front matter 䉷 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2007.01.066 Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS 2

S. Patel, K.K. Pant / Chemical Engineering Science

SRM can produce H2 /CO2 in the molar ratio of 3/1. It produces relatively small amount of carbon monoxide, usually less than 1%, as a by-product over copper-based catalysts (Choi and Stenger, 2002; Lindstrom et al., 2003; Patel and Pant, 2006b). But it is highly endothermic and has a slow rate of reaction. The limitations of partial oxidation and steam reforming of methanol for the hydrogen production for PEM fuel cell applications could be overcome by combining them as an oxidative steam reforming of methanol (OSRM) that can be performed auto-thermally with idealized reaction stoichiometry. CH3 OH + (1 − 2a)H2 O + aO2 ←→ (3 − 2a)H2 + CO2 , H 0 = (49.5 − 241.8 ∗ 2a) kJ mol−1 ,

0 a 0.5,

(3)

where a is an oxygen to methanol (O/M) molar ratio. For a = 0.125 OSRM can be operated auto-thermally at 573 K. The limited work in the field of OSRM has been reported over the different catalysts like CuO/ZnO/Al2 O3 (Murcia-Mascaros et al., 2001; Turco et al., 2004a,b), Cu/Zn/Zr/Al oxide (Velu et al., 2001), Ce/CuO (Shan et al., 2004), Pd/ZnO (Liu et al., 2003), Pd–Zn/Cu–Zn–Al (Chen et al., 2007), etc. Turco et al. (2004b) observed the 100% methanol conversion at a temperature greater than 623 K. They observed the methane and other oxygenates in addition to H2 , CO2 and CO over CuO/ZnO/Al2 O3 . The copper–ceria based catalysts have been used widely for the preferential oxidation of carbon monoxide (PROX) (Wang et al., 2002; Papavasiliou et al., 2006). Also ceria is known to improve the stability of catalysts (Mattos et al., 2002). Therefore, in the present study we have devised the incorporation of ceria in Cu–Zn–Al oxide catalysts in order to improve the catalyst performance in OSRM process in terms of methanol conversion, hydrogen selectivity, suppression of CO formation and deactivation compared to Cu–Zn–Al oxide catalysts. The detailed characterization and activity study of Cu–Zn–Ce–Al as well as Cu–Zn–Al oxide catalysts have been carried out over the wide range of operating conditions of OSRM process. 2. Experimental All the Cu–Zn–Ce–Al oxide and Cu–Zn–Al oxide catalysts, abbreviated as CZCeAi and CZAi , respectively, were prepared by co-precipitation method as follows. The 1.25 M solutions of Cu(NO3 )2 3H2 O, Zn(NO3 )2 6H2 O, Ce(NO3 )3 6H2 O and Al(NO3 )3 9H2 O were prepared separately and then mixed in a volume proportion corresponding to the final composition of the catalysts to be obtained. The resulting solution was stirred and heated to 343 K in a round-bottomed flask. The aqueous 0.5 M Na2 CO3 solution was added drop-wise to the nitrate solution under vigorous stirring until pH 7 was attained at a temperature of 343 K. The precipitate was allowed to age for 2 h at temperature 343 K with stirring. The excess solution was removed by filtration. The precipitate was washed three times by double distilled water at a temperature of 343 K followed by several times by double distilled water at room temperature in order to remove the sodium salts. The drying was carried out at 383 K for 12 h. This material was the catalyst precursor,

(

)

–

which was crushed to a fine powder and subsequently the catalyst was produced by calcination in the presence of air at 673 K for 4 h. The pellets of 3 mm size from fine catalyst powder were made in an automatic palletizing press Techno Search AP-15 and subsequently crushed and sieved to a size of 20/25 mesh. The same procedure was followed for the preparation of Cu–Zn–Al oxide catalysts. The final elemental composition was determined using atomic absorption spectroscopy (AAS) Varian AA420FS. Surface area and pore volume of catalysts were measured using ASAP 2010 Micromeritics micro-pore surface area analyzer. The copper surface area, dispersion and particle size were measured by CO chemisorption at 193 K using Micromeritics Pulse Chemisorption ChemiSorb 2720 for Cu–Zn–Ce–Al oxide catalysts (Holmgren and Anderson, 1998; Holmgren et al., 1999; Cheekatamarla et al., 2005). It has been reported that the spillover of CO from metal onto the ceria support gives a higher dispersion of metal than the actual metal dispersion. Chemisorption of CO at 193 K alleviates this phenomenon. A 0.5 g catalyst was reduced in the stream of 9.9% H2 /Ar with the flow of 30 cc min−1 at 533 K for 3 h with a final temperature achieved by 10 K min−1 ramp rate. Then the sample was flushed with N2 for 30 min and subsequently cooled to room temperature at a ramp rate of 10 K min−1 . Then 193 K temperature of sample was achieved by immersing the sample holder in the dewar of iso-propanol gel that was prepared by quenching it with liquid nitrogen. At this temperature, 0.2 ml pulses of CO were injected periodically in the stream of nitrogen in order to find out the total CO uptake. Temperature programmed reduction (TPR) of calcined catalysts was carried out using Pulse Chemisorption ChemiSorb2720, Micromeritics. Different crystalline phases were identified by means of Philips X’PERT PRO PW 3040/60 (PANalytical) ˚ powder diffractometer using monochromatic Cu-K 1.5418 A radiation at a current of 30 mA with diffraction angle ranging from 10–60◦ . The amount of carbon deposited on catalysts was determined by oxidizing it in the air by means of thermogravimetric analyzer (TGA, Seiko TG/DTA 32 SSC 5100). Surface analysis of post reaction catalysts was done by X-ray photoelectron spectroscopy (XPS) using Perkin Elmer-1257. Catalyst performance was evaluated in a fixed bed reactor assembled with electrically heated furnace and PID controllers. All the catalysts were reduced in situ in the stream of 10% H2 /N2 mixture before OSRM reaction. The reduction of Cu–Zn–Ce–Al oxide catalysts was carried out at a ramp of 10 K min−1 with final temperature of 500 K and dwelled for 1 h; on the other hand Cu–Zn–Al oxide catalysts dwelled at 530 K for 1 h. A mixture of methanol and water vapor generated in an evaporator was passed through a preheating zone along with oxygen and nitrogen prior to entering the reactor. The stainless steel reactor was used which consisted of a cylindrical cup of 19 mm diameter and 30 mm length, perforated on the bottom with orifices of 0.5 mm diameter to enable the flow of gases. The catalyst was retained by a plug of glass wool. An equal amount of inert quartz particles of 20/25 mesh were mixed with the catalyst to dilute the heat transfer effects. A thermowell passed through the catalyst bed so that temperature could be measured at different positions along the bed using

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS S. Patel, K.K. Pant / Chemical Engineering Science

thermocouple. Product effluent was passed through the condenser and separator in order to separate the unconverted methanol and water from the product gases. Operating temperature, contact time (W/F) and oxygen to methanol (O/M) molar ratio varied from 473 to 573 K, 3 to 15 kgcat s mol−1 and 0.1 to 0.5, respectively, with steam to methanol (S/M) molar ratio = 1.5 and pressure P = 1 atm being kept constant. Reaction products were analyzed by Nucon-5700 Gas chromatograph, equipped with thermal conductivity detector having carbosphere column for gaseous product concentration measurement and flame ionization detector with silica–alumina fused capillary column for unconverted liquid reactants. The very low concentration of CO was determined using molecular sieve-5A followed by methanizer installed with flame ionization detector. 3. Results and discussion 3.1. Catalyst characterization The specific surface area and pore volume of various catalyst samples shown in Table 1 were measured according to the Brunauer–Emmett–Teller (BET) method by nitrogen adsorption at 77 K after degassing the samples at least for 12 h at 523 K. There is no significant difference in the BET surface area was observed among the catalysts. The surface area, dispersion and particle size of copper in the catalyst samples shown in Table 1 were calculated assuming atomic adsorption of CO on a copper site and the site density of copper to be 0.069 nm2 atm−1 . All the copper particles were assumed to be spherical for calculating the mean particle size of copper. Stoichiometrically one atom of CO is adsorbed on each copper atom site. The CO chemisorption was carried out at 193 K in order to hinder the spillover of CO to the ceria that was confirmed by comparing the chemisorption results of catalysts which were obtained at room temperature 302 K. Holmgren and Anderson (1998) have confirmed the CO spillover to the ceria support by performing CO chemisorption at 298 K followed by temperature programme desorption (TPD) over

(

)

–

3

Pt/CeO2 catalyst. However, they observed that the spillover of CO to ceria support could be hindered by performing CO chemisorption at 195 K. The catalysts containing Ce exhibited higher copper surface area and better dispersion, and hence the smaller copper particle size compared to those containing only Cu–Zn–Al oxides. However, optimum loading of Ce was required as shown in Table 1. The alumina and ceria enhanced the dispersion of copper oxide species by stabilization of isolated Cu2+ ions in their matrix and moderately by formation of spinal like CuAl2 O4 . This is in agreement with the work of Fernandez-Garcia et al. (1997) who confirmed that the ceria increases the copper dispersion by reducing the particle size of copper in the CuO/CeO2 /Al2 O3 catalysts compared to only CuO/Al2 O3 catalysts. XRD spectra of catalyst precursors displayed in Fig. 1(a) exhibit mainly hydrotalcite (HT)-like layered double hydroxide (LDH; (CuZn)6 Al2 (OH)16 CO3 · 4H2 O; JCPDS file No. 38-487) with other phases such as aurichalcite (AC; (CuZn)5 (CO3 )2 (OH)6 ; JSPDS file No. 7-743) and bayerite (AH; Al(OH)3 ; JCPDS file No. 20-11). At pH of precipitates higher than 7 during catalyst preparation, aurichalcite was the dominated phase of catalyst precursors, and on the other hand at pH lower than 7, the hydroxynitrates were dominated. At pH 7 HT-like LDH phase was dominant which exhibited the higher catalytic activity. The calcination of these precursors resulted in the catalyst comprising copper, zinc, cerium and aluminum oxide. When catalysts were reduced in the stream of hydrogen at the beginning of activity test, CuO converted into copper where as zinc, cerium and aluminum remained in oxide form as shown in Fig. 1(b). A small peak of spinal CuAl2 O4 was also observed in the XRD spectra. As per the Scherrer formula the crystal size is inversely proportional to the peak width at constant diffraction angle and monochromatic wavelength. The peak width of copper for a CZCeA2 catalyst is broader than that for the CZA2 catalyst (Fig. 1(b)), which resulted in a smaller copper crystallite size. The smaller the crystallite sizes, the better the copper dispersion and the more the copper surface area for CZCeA2 catalyst compared to CZA2 catalyst. The reducibility of copper species in catalysts was investigated by TPR experiments, and profiles are displayed in

Table 1 Physiochemical properties of catalysts

Preparative composition (wt%) Final composition (wt%) BET surface area (m2 g−1 ) Pore volume (cm3 g−1 ) Cu dispersion (%) Cu surface area (m2 g−1 ) ˚ Cu particle size (A) Methanol conversiona (%) H2 ratea (mmol s−1 kg−1 cat ) CO formationa (ppm) a Results

CZA1 Cu/Zn/Al

CZA2 Cu/Zn/Al

CZCeA1 Cu/Zn/Ce/Al

CZCeA2 Cu/Zn/Ce/Al

CZCeA3 Cu/Zn/Ce/Al

30/30/40 29/26/45 92 0.26 9.4 18.3 108 60 132 9400

30/20/50 29/19/52 106 0.32 12.8 25.1 80 77 180 3400

30/25/5/40 30/24/6/40 96 0.28 10.2 20.2 101 69 160 1400

30/20/10/40 29/18/10/43 108 0.34 19.6 38.6 52 100 244 995

30/10/20/40 29/10/19/42 101 0.29 14.8 29.3 69 90 217 1240

at T = 553 K, S/M = 1.5 M, O/M = 0.15 M and W/F = 11 kgcat mol−1 s.

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS 4

S. Patel, K.K. Pant / Chemical Engineering Science

(

)

–

LDH AH AC (003)

(012)

Intensity (a.u.)

(006)

(015)

CZCeA2

CZA2

10

15

20

25

30

35 2θ

40

45

50

55

60

Intensity (a.u.)

Cu CuO ZnO Al2O3 CeO2 CuAl2O4

CZCeA2

CZA2

10

15

20

25

30

35 2θ

40

45

50

55

60

Fig. 1. (a) X-ray diffraction patterns of catalysts precursors. Layered double hydroxide (LDH) (CuZn)6 Al2 (OH)16 CO3 · 4H2 O; aurichalcite (AC): (CuZn)5 (CO3 )2 (OH)6 ; bayerite (AH): Al(OH)3 . (b) X-ray diffraction patterns of reduced catalysts.

Fig. 2. TPR profiles of calcined catalysts.

Fig. 2. All the catalysts exhibited composite peaks, with two or three unresolved signals. These might be due to reduction of different Cu(II) species such as, CuO, CuAl2 O4 and Cu2+ ions incorporated in octahedral sites of the alumina phase. The pure CuO sample had shown a single large peak at 613 K. For CZA2 catalyst two peaks were observed, the peak at 513 K is

Fig. 3. Catalytic activity in terms of methanol conversion for different catalysts as a function of temperature (W/F = 11 kgcat s mol−1 methanol , O/M = 0.15 M, S/M = 1.5 M, P = 1 atm).

attributed to the reduction of CuO into Cu whereas the peak at 459 K might be due to spinal CuAl2 O4 . But it is expected that the limited amount of surface CuAl2 O4 species contribute to the TPR signals; therefore, the peak at 459 K might be a composite peak of CuAl2 O4 and Cu2 O reduction. The formation of Cu2 O is also confirmed by Lindstrom et al. (2002) and Reitz et al. (2001) in which they found two step reduction of copper: Cu(II) → Cu(I) → Cu(0). For CZCeA2 catalyst three composite peaks were observed with shoulders at 443, 463 and 488 K. The lower valent Cu species cause structural defects of ceria lattice. However, cerium sublattice is not strongly perturbed by these defects, which instead cause the formation of a defective oxygen sublattice (Shan et al., 2004). In this way many oxygen vacancies are generated with the formation of solid solutions. Therefore, oxygen vacancies of CZCeA2 catalyst were reduced at a low temperature of 443 K during TPR to give first peak. The peak at temperature 463 K was attributed to spinal CuAl2 O4 and Cu2 O, and a peak at 488 K was due to CuO reduction to Cu. Breen and Ross (1999) and Agrell et al. (2001) reported that the highly dispersed CuO gave the TPR signals at lower temperature than the bulk CuO. Dow et al. (2000) and Liu and Flyzani-Stephanopoulos (1996) found that strong interaction occurs at the interface between highly dispersed copper oxide and ceria (interfacial metal oxide–support interaction (IMOSI)) due to their close contact with each other that leads to the low temperature reduction of copper oxide. This is also observed in the present study with the catalyst CZCeA2 for which CuO was reduced at lower temperature than that of CZA2 due to better dispersion of copper in CZCeA2 . 3.2. Effect of ceria over catalysts performance Figs. 3 and 4 clearly depict that the addition of ceria in CZAi catalysts improved the performance greatly in terms

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS S. Patel, K.K. Pant / Chemical Engineering Science

(

)

–

5

Fig. 4. Catalytic activity in terms of hydrogen production rate for different catalysts as a function of temperature (W/F = 11 kgcat s mol−1 methanol , O/M = 0.15 M, S/M = 1.5 M, P = 1 atm).

Fig. 5. Comparison of CO formation for different catalysts as a function of temperature (W/F = 11 kgcat s mol−1 methanol , O/M = 0.15 M, S/M = 1.5 M, P = 1 atm).

of methanol conversion as well as hydrogen production. The CZCeA2 exhibited 100% methanol and 244 mmol s−1 kg−1 cat hy−1 drogen rate at 553 K whereas 50% and 112 mmol s kg−1 cat given by CZA1 . The methanol conversion was increased as a function of temperature for all the catalysts. The enhanced activity of CZCeA2 could be due to higher copper surface area and better copper dispersion (Table 1 and XRD spectra of Fig. 1(b)). Cheekatamarla et al. (2005) observed the better copper dispersion in ceria promoted copper–alumina catalyst than catalyst without it. The addition of zinc oxide and ceria caused the synergistic effect to give the activity enhancement. Ceria has been widely used as a promoter in three-way catalysts for removal of carbon monoxide, nitrogen oxides and hydrocarbons under conditions with rich-lean air/fuel in automotive exhaust (Golunski et al., 1995). It has also been widely studied for PROX to clean the hydrogen rich feed stream for PEM fuel cell applications (Wang et al., 2002; Pillai and Deevi, 2006) and methane partial oxidation (Pantu and Gavalas, 2002). This is because ceria has a fluorite-like cubic structure in which each cerium site is surrounded by eight oxygen sites in fcc arrangement and each oxygen site has a tetrahedron of cerium sites. Under various redox conditions, the oxidation state of cerium may vary between 3+ and 4+. The better redox properties than those CeO2 alone can be obtained by the incorporation of metal ions into the CeO2 lattice forming Ce1−x Mx Oy solid solutions (Shan et al., 2004). Surface oxygen and oxygen vaccines are involved in the catalytic activity. This enhanced oxygen-mobility facilitates the occurrence of redox process like OSRM at lower temperature. Therefore, enhanced performance of Cu–Zn–Ce–Al oxide catalysts in terms of methanol conversion, high hydrogen production rate and low CO formation was observed in the present study. It can be seen that optimum loading of ceria was required among the Cu–Zn–Ce–Al oxide catalysts. The 15% and 5% Ce catalysts had given lower methanol conversion and

corresponding H2 rate compared to 10% Ce catalyst as shown in Figs. 3 and 4. The carbon monoxide formation was raised drastically at a temperature greater than 513 K for all the Cu–Zn–Al oxide catalysts, whereas for the 10% and 15% ceria-based catalysts it was increased gradually as a function of temperature as shown in Fig. 5. The CO concentration of 10% and 15% Cu–Zn–Ce–Al oxide catalysts was in the range of 0.05–0.1%, which is quite low. Incorporation of ceria leads to the inhibition of CO formation envisaged due to the oxygen storage–release ability of ceria; this was expected because copper–ceria based catalysts have been used for the CO oxidation and PROX. Pillai and Deevi (2006) reported that catalytic activity of CuO/ZnO could be improved significantly by incorporating CeO2 for the carbon monoxide oxidation. The CuO/ZnO/CeO2 based catalysts have also been quite active for the PROX in hydrogen rich stream for the PEM fuel cell applications (Jung et al., 2004; Cheekatamarla et al., 2005). 3.3. Effect of contact time and oxygen to methanol molar ratio on catalyst performance The effect of contact time (W/F) on the methanol conversion and CO formation at a temperature of 553 K is shown in Fig. 6. The methanol conversion reached 100% at contact time more than 11 kgcat s mol−1 . The CO formation increased from 300 ppm at contact time 3 to 2400 ppm at contact time 15 kgcat s mol−1 . The contact time 11 kgcat s mol−1 is the optimum because the 100% methanol conversion with CO formation as low as 995 ppm was obtained with it. It was found that as contact time was increased the hydrogen selectivity marginally decreased. The concentration of carbon monoxide for all the catalysts was always below 1% at all the operating conditions, which is well below the equilibrium concentration of water gas shift reaction. At a temperature less

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS 6

S. Patel, K.K. Pant / Chemical Engineering Science

Fig. 6. Effect of contact time over methanol conversion and CO formation for CZCeA2 catalyst (T = 553 K, O/M = 0.15 M, S/M = 1.5 M, P = 1 atm).

than 473 K and contact time 3 kgcat s mol−1 there was no CO detected; therefore it can be postulated that the carbon monoxide is not a primary product. All these suggest that the CO formation takes place through reverse water gas shift reaction (rWGS: Eq. (4)), which uses the hydrogen and carbon dioxide produced by reforming reaction (Eq. (3)). H2 + CO2 ←→ CO + H2 O.

(4)

It can be seen in from Fig. 4 that at 573 K temperature H2 rate is lower than that at 553 K even if methanol conversion was 100% at both the temperatures. This was because of conversion of hydrogen into carbon monoxide via WGS reaction as it was accelerated at higher temperature. Various routes have been proposed for the formation of CO in the literature of methanol reforming. Santacesaria and Carra (1983) have suggested the decomposition/water gas shift reaction sequence: CH3 OH → CO + 2H2 ,

(5)

CO + H2 O ←→ H2 + CO2 .

(6)

Jiang et al. (1993) proposed the kinetic model for steam reforming of methanol in which they suggested CO formation via decomposition of methyl formate: 2CH3 OH → CH3 OCHO + 2H2 ,

(7)

CH3 OCHO + H2 O → HCOOH + CH3 OH,

(8)

HCOOH → CO2 + H2 .

(9)

Peppley et al. (1999) reported that the methanol decomposition is much slower than the steam reforming methanol reaction but it must be included in the overall reaction scheme comprising methanol steam reforming, methanol decomposition and water gas shift reactions. Horng et al. (2006) reported the CO formation via WGS (Eq. (6)) for the auto-thermal reforming of methanol. Breen et al. (1999) have done extensive study to understand the CO formation mechanism through DRIFT

(

)

–

Fig. 7. Effect of oxygen to methanol molar ratio over methanol conversion and hydrogen production for CZCeA2 catalyst (W/F = 11 kgcat s mol−1 methanol , T = 553 K, S/M = 1.5 M, P = 1 atm).

analysis as well as kinetic studies. They confirmed that the CO formation over CuO/ZnO/ZrO2 /Al2 O3 catalyst for steam reforming of methanol occurs via rWGS (Eq. (4)). Recently, many researchers have also proposed the CO formation via rWGS reaction (Eq. (4)) (Agrell et al., 2001; Purnama et al., 2004; (Reuse et al., 2004; Horny et al., 2007). The O/M molar ratio is expected to have a strong influence on the catalytic performance because the OSRM is a combination of POM and SRM reactions. The data were collected by keeping the steam to methanol molar ratio at 1.5 as shown in Fig. 7. The methanol conversion and hydrogen production rate increased with increasing O/M ratio and attained maximum corresponding values at O/M = 0.15. With further increase in O/M ratio the methanol conversion was 100% but the rate of H2 production declined drastically up to O/M = 0.25 due to the increase in CO formation rate by rWGS and also due to some extent of methanol partial oxidation (Eq. (1)). At O/M ratio greater than 0.25 both methanol conversion and hydrogen production decreased steeply because methanol partial oxidation (Eq. (1)) was the dominant reaction. The high exothermicity of POM resulted in inefficient heat removal and consequently decline of methanol conversion. The S/M molar ratio 1.5 was kept constant in the present study based on our previous study (Agarwal et al., 2005). The excess steam retarded the methanol decomposition by providing sufficient amount of steam for the methanol reforming reaction (Eq. (2)), suppressed the CO formation by inhibiting the rWGS reaction (Eq. (4)) to proceed toward the right side and also reduced the catalyst deactivation. 3.4. Reaction pathway The OSRM involves the following reactions in addition to the OSRM reaction Eq. (3). The POM (Eq. (11)) and SRM (Eq. (12)) are expected to occur simultaneously at the entry of reactor, while at the downstream side of reactor the major reaction may be the SRM. CH3 OH + 0.25O2 ←→ 2H2 + 0.5CO2 + 0.5CO,

(10)

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS S. Patel, K.K. Pant / Chemical Engineering Science

Fig. 8. Time on stream (TOS) stability test of catalysts (W/F = 11 kgcat s mol−1 methanol , T = 553 K, O/M = 0.15 M, S/M = 1.5 M, P = 1 atm).

CH3 OH + 0.5O2 ←→ 2H2 + CO2 ,

(11)

CH3 OH + H2 O ←→ 3H2 + CO2 ,

(12)

CH3 OH ←→ 2H2 + CO,

(13)

CO + 0.5O2 ←→ CO2 ,

(14)

H2 + 0.5O2 ←→ H2 O,

(15)

CH3 OH + 0.25O2 ←→ HCHO + 0.5H2 O + 0.5H2 ,

(16)

The experimental hydrogen selectivity (moles of hydrogen produced per mole of methanol converted) was observed slightly lower than that stoichiometrically given by Eq. (3). The selectivity of carbon dioxide decreased with temperature, whereas that of carbon monoxide was increased. These above all suggest the occurrence of rWGS reaction (Eq. (4)), and also the possibility of some hydrogen produced via Eqs. (10) and (13). For the OSRM process, benefit in reduction of CO concentration might be obtained due to conversion of CO to CO2 via oxidation reaction (Eq. (14)). The formation of formaldehyde via Eq. (16) was prevented by operating the OSRM at higher space velocities (Velu et al., 2001). The ceriabased Cu–Zn–Ce–Al oxide catalysts enhanced the activity that facilitated the complete methanol conversion at relatively lower temperature of 553 K so that there was no formation of methane and other oxygenates which were observed by Turco (et al. (2004b). 3.5. Stability of catalysts The Cu/Zn/Al2 O3 catalysts face the great problem of deactivation (Agarwal et al., 2005). The CZA2 catalyst deactivated greatly by a magnitude of 26% methanol conversion over a 72 h TOS stability test as shown in Fig. 8. On the other hand the activity of CZCeA2 catalyst declined only by a magnitude of 8%. For a CZCeA2 catalyst activity declined initially and then became nearly constant. This initial decline in the

(

)

–

7

catalytic activity might be due to some transient transformation to a stable form under the reaction condition that occurred in the catalyst. Catalyst CZCeA2 containing ceria was found to be quite durable because ceria promoter has high oxygen storage–release capacity, which inhibits the coke deposition significantly by coke gasification. The higher amount of lattice oxygen near the metal particles promotes the mechanism of carbon removal from the metallic surface (Fernandez-Garcia et al., 1997; Mattos et al., 2002, Patel and Pant, 2006a). They also reported that ceria improves the stability against sintering of metallic copper phase formed during the reaction. The amount of coke deposited on the catalysts obtained using TGA was 0.18 and 1.55 wt% for the CZCeA2 and CZA2 catalysts, respectively. The C1s XPS spectra of spent catalysts revealed that the graphite carbon –C–C– (284 eV, BE) and oxidized carbon CO3 (288 eV, BE) were present on the catalyst surface. 4. Conclusion In the present study to investigate the effect of Ce loading on the performance of Cu–Zn–Al oxide catalysts for OSRM, the optimum 10 wt% doping of cerium improved the catalyst performance greatly in terms of methanol conversion, hydrogen production rate and CO suppression. The 100% methanol conversion with CO formation as low as 995 ppm was obtained that reduce the load on preferential oxidation of CO to CO2 (PROX) before feeding the hydrogen rich stream as a feed for the PEM fuel cells. The optimum operating parameters could be suggested as reaction temperature 553 K, contact time (W/F) 11 kgcat s mol−1 , oxygen to methanol (O/M) molar ratio 0.15 and steam to methanol molar (S/M) ratio 1.5. The spillover of carbon monoxide to ceria can be hindered by performing the CO chemisorption over ceria-based catalysts at low temperature of 193 K. The CO chemisorption, TPR and XRD studies have revealed that the incorporation of Ce in Cu–Zn–Al oxide catalysts improved the copper dispersion by reducing the particle size of copper, and also facilitated catalyst reduction at lower temperature. The incorporation of ceria greatly improved the stability of Cu–Zn–Al oxide catalysts.

References Agarwal, V., Patel, S., Pant, K.K., 2005. H2 production by steam reforming of methanol over Cu/ZnO/Al2 O3 catalysts: transient deactivation kinetics modeling. Applied Catalysis A: General 279, 155–164. Agrell, J., Hasselbo, K., Jansson, K., Jaras, S.G., Boutonnet, M., 2001. Production of hydrogen from methanol over Cu/ZnO catalysts promoted by ZrO2 and Al2 O3 . Applied Catalysis A: General 211, 239–250. Bowers, B.J., Zhao J.L., Ruffo, M., Khan, R., Dattatraya, D., Dushman, N., Beziat, J., Boudjemaa, F., 2006. Onboard fuel processor for PEM fuel cell vehicles. International Journal of Hydrogen Energy, in press. Available online at sciencedirect.com. Breen, J.P., Ross, J.R.H., 1999. Methanol reforming for fuel-cell applications: development of zirconia-containing Cu–Zn–Al catalysts. Catalysis Today 51, 521–533. Breen, J.P., Meunier, F.C., Ross, J.R.H., 1999. Mechnistic aspects of the steam reforming of methanol over a CuO/ZnO/ZrO2 /Al2 O3 catalyst. Chemical Communications, 2247–2248.

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066

ARTICLE IN PRESS 8

S. Patel, K.K. Pant / Chemical Engineering Science

Cheekatamarla, P.K., Epling, W.S., Lane, A.M., 2005. Selective lowtemperature removal of carbon monoxide from hydrogen-rich fuels over Cu–Ce–Al catalysts. Journal of Power Sources 147, 178–183. Chen, G., Li, S., Yuan, Q., 2007. Pd −Zn/Cu −Zn −Al catalysts prepared for methanol oxidation reforming in micro channel reactors. Catalysis Today 120, 63–70. Choi, Y., Stenger, H.G., 2002. Fuel cell grade hydrogen from methanol on a commercial Cu/ZnO/Al2 O3 catalyst. Applied Catalysis B: Environmental 38, 259–269. Dow, W., Wang, Y., Huang, T., 2000. TPR and XRD studies of yttriadoped ceria/g-alumina-supported copper oxide catalyst. Applied Catalysis A: General 190, 25–34. Edwards, N., Ellis, S.R., Frost, J.C., Golunski, S.E., Keulen, A.N.J., Lindewald, N.G., Reinkingh, J.G., 1998. On-board hydrogen generation for transportation applications: the HotSpot methanol processor. Journal of Power Sources 71, 123–128. Fernandez-Garcia, M., Rebollo, E.G., Ruiz, A.G., Conesa, J.C., Soria, J., 1997. Influence of ceria on the dispersion and reduction/oxidation behavior of alumina-supported copper catalysts. Journal of Catalysis 172, 146–159. Golunski, S.E., Hatcher, H.A., Rajaram, R.R., Truex, T.J., 1995. Origins of low-temperature three-way activity in Pt/CeO2 . Applied Catalysis B: Environmental 3, 367–376. Holmgren, A., Anderson, B., 1998. Oxygen storage dynamics in Pt/CeO2 /Al2 O3 catalysts. Journal of Catalysis 178, 14–25. Holmgren, A., Anderson, B., Duprez, D., 1999. Interactions of CO with Pt/ceria catalysts. Applied Catalysis B: Environmental 22, 215–230. Horng, R., Chou, H., Lee, C., Tsai, H., 2006. Characterisation of hydrogen produced by partial oxidation and auto-thermal reforming in a small methanol reformer. Journal of Power Sources 161, 1225–1233. Horny, C., Renken, A., Kiwi-Minsker, L., 2007. Compact string reactor for autothermal hydrogen production. Catalysis Today 120, 45–53. Jiang, C.J., Trimm, D.L., Wainwright, M.S., Cant, N.W., 1993. Kinetic mechanism for the reaction between methanol and water over a Cu–ZnO–Al2 O3 catalyst. Applied Catalysis A: General 97, 145–158. Jung, C.R., Han, J., Nam, S.W., Lim, T.H., Hong, S.A., Lee, H.I., 2004. Selective oxidation of CO over CuO–CeO2 catalyst: effect of calcination temperature. Catalysis Today 93–95, 183–190. Lindstrom, B., Pettersson, L.J., Menon, P.G., 2002. Activity and characterization of Cu/Zn, Cu/Cr and Cu/Zr on -alumina for methanol reforming for fuel cell vehicles. Applied Catalysis A: General 234, 111–125. Lindstrom, B., Agrell, J., Pettersson, L.J., 2003. Combined methanol reforming for hydrogen generation over monolithic catalysts. Chemical Engineering Journal 93, 91–101. Liu, S., Takahashi, K., Ayabe, M., 2003. Hydrogen production by oxidative steam reforming of methanol on Pd/ZnO catalyst: effects of Pd loading. Catalysis Today 87, 247–253. Liu, W., Flyzani-Stephanopoulos, M., 1996. Transition metal-promoted oxidation catalysis by fluorite oxides: a study of CO oxidation over Cu–CeO2 . The Chemical Engineering Journal and the Biochemical Engineering Journal 64, 283–294. Mattos, L.V., Oliveira, E.R., Resende, P.D., Noronha, F.B., Passos, F.B., 2002. Partial oxidation of methane on Pt/Ce–ZrO2 catalysts. Catalysis Today 77, 245–256. Murcia-Mascaros, S., Navarro, R.M., Gomez-Sainero, L., Costantino, U., Nocchetti, M., Fierro, J.L.G., 2001. Oxidative methanol reforming reactions on CuZnAl catalysts derived from hydrotalcite-like precursors. Journal of Catalysis 198, 338–347. Pantu, P., Gavalas, G.R., 2002. Methane partial oxidation on Pt/CeO2 and Pt/Al2 O3 catalysts. Applied Catalysis A: General 223, 253–260. Papavasiliou, J., Avgouropoulos, G., Ioannides, T., 2006. In situ combustion synthesis of structured Cu–Ce–O and Cu–Mn–O catalysts for the production and purification of hydrogen. Applied Catalysis B: Environmental 66, 168–174.

(

)

–

Patel, S., Pant, K.K., 2006a. Activity and stability enhancement of copper–alumina catalysts using cerium and zinc promoters for the selective production of hydrogen via steam reforming of methanol. Journal of Power Sources 159, 139–143. Patel, S., Pant, K.K., 2006b. Influence of preparation method on performance of Cu(Zn)(Zr)–alumina catalysts for the hydrogen production via steam reforming of methanol. Journal of Porous Materials 13, 381–386. Patel, S., Pant, K.K., 2006c. Production of hydrogen with low carbon monoxide formation via catalytic steam reforming of methanol. ASME Journal of Fuel Cell Science and Technology 3, 369–374. Peppley, B.A., Amphlett, J.C., Kearns, L.M., Mann, R.F., 1999. Methanol–steam reforming on Cu/ZnO/Al2 O3 catalysts. Part 2. A comprehensive kinetic model. Applied Catalysis A: General 179, 31–49. Pillai, U.R., Deevi, S., 2006. Copper–zinc oxide and ceria promoted copper–zinc oxide as highly active catalysts for low temperature oxidation of carbon monoxide. Applied Catalysis B: Environmental 65, 110–117. Purnama, H., Ressler, T., Jentoft, R.E., Soerijanto, H., Schlogl, R., Schomacker, R., 2004. CO formation/selectivity for steam reforming of methanol with a commercial CuO/ZnO/Al2 O3 catalyst. Applied Catalysis A: General 259, 83–94. Raimondi, A., Geissler, K., Wambach, J., Wokaun, A., 2002. Hydrogen production by methanol reforming: post-reaction characterisation of a Cu/ZnO/Al2 O3 catalyst by XPS and TPD. Applied Surface Science 189, 59–71. Reitz, T.L., Lee, P.L., Czaplewski, K.F., Lang, J.C., Popp, K.E., Kung, H.H., 2001. Time-resolved XANES investigation of CuO/ZnO in the oxidative methanol reforming reaction. Journal of Catalysis 199, 193–201. Reuse, P., Renken, A., Hass-Santo, K., Gorke, O., Schubert, K., 2004. Hydrogen production for fuel cell application in an autothermal microchannel reactor. Chemical Engineering Journal 101, 133–141. Sahoo, D.R., Vajpai, S., Patel, S., Pant, K.K., 2007. Kinetic modeling of steam reforming of ethanol for the production of hydrogen over Co/Al2 O3 catalyst. Chemical Engineering Journal 125, 139–147. Santacesaria, E., Carra, S., 1983. Kinetics of catalytic steam reforming of methanol in a CSTR reactor. Applied Catalysis 5, 345–358. Semelsberger, T.A., Ott, K.C., Borup, R.L., Greene, H.L., 2006. Generating hydrogen-rich fuel-cell feeds from dimethyl ether (DME) using Cu/Zn supported on various solid–acid substrates. Applied Catalysis A: General 309, 210–223. Shan, W., Feng, Z., Li, Z., Zhang, J., Shen, W., Li, C., 2004. Oxidative steam reforming of methanol on Ce0.9Cu0.1Oy catalysts prepared by deposition–precipitation, coprecipitation, and complexation–combustion methods. Journal of Catalysis 228, 206–217. Turco, M., Bagnasco, G., Costantino, U., Marmottini, F., Montanari, T., Ramis, G., Busca, G., 2004a. Production of hydrogen from oxidative steam reforming of methanol I. Preparation and characterization of Cu/ZnO/Al2 O3 catalysts from a hydrotalcite-like LDH precursor. Journal of Catalysis 228, 43–55. Turco, M., Bagnasco, G., Costantino, U., Marmottini, F., Montanari, T., Ramis, G., Busca, G., 2004b. Production of hydrogen from oxidative steam reforming of methanol II. Catalytic activity and reaction mechanism on Cu/ZnO/Al2 O3 hydrotalcite-derived catalysts. Journal of Catalysis 228, 56–65. Velu, S., Suzuki, K., Kapoor, M.P., Ohashi, F., Osaki, T., 2001. Selective production of hydrogen for fuel cells via oxidative steam reforming of methanol over CuZnAl(Zr)-oxide catalysts. Applied Catalysis A: General 213, 47–63. Wang, J.B., Tsai, D., Huang, T., 2002. Synergistic catalysts of carbon monoxide oxidation over copper oxide supported on samaria-doped ceria. Journal of Catalysis 208, 370–380. Wang, Z., Xi, J., Wang, W., Lu, G., 2003. Selective production of hydrogen by partial oxidation of methanol over Cu/Cr catalysts. Journal of Molecular Catalysis A: Chemical 191, 123–134. Yan, X., Wang, S., Li, X., Hou, M., Yuan, Z., Li, D., Pan, L., Zhang, C., Liu, J., Ming, P., Yi, B., 2006. A 75-kW methanol reforming fuel cell system. Journal of Power Sources 244, 1–9.

Please cite this article as: Patel, S., Pant, K.K., Selective production of hydrogen via oxidative steam reforming of methanol using Cu–Zn–Ce–Al oxide catalysts. Chemical Engineering Science (2007), doi: 10.1016/j.ces.2007.01.066