Journal of Energy Chemistry 23(2014)639–644

Kinetic behaviour of commercial catalysts for methane reformingin ethanol steam reforming process

Jorge Vicentea, Javier Erenaa∗, Martin Olazara, Pedro L. Benitob, Javier Bilbaoa, Ana G. Gayuboa

a. Departamento de Ingenierıa Quımica, Universidad del Paıs Vasco UPV/EHU, Apartado 644, 48080 Bilbao, Spain;b. Departamento de Ingenierıa Minera y Metalurgica y Ciencia de los Materiales, Universidad del Paıs Vasco UPV/EHU,

Nieves Cano 12, 01006 Vitoria-Gasteiz, Spain

[ Manuscript received February 17, 2014; revised May 4, 2014 ]

AbstractEthanol steam reforming has been studied in a fluidized bed (in order to ensure bed isothermicity) on commercial catalysts for methanereforming. The results allow analyzing the effect of temperature (in 300−700 ◦C range), and both metal and support nature on the reactionindices (ethanol conversion, yields and selectivities to H2 and byproducts (CO2, CO, CH4 and C2H4O)). Special attention has been paid tocatalysts’ stability by comparing the evolution of the reaction indices with time on stream at 500 ◦C (minimum CO formation) and 700 ◦C(minimum deactivation by coke deposition). Although they provide a slightly lower H2 yield, the results evidence a good behaviour of Nibased catalysts, indicating that they are an interesting alternative of more expensive Rh based ones.

Key wordsethanol steam reforming; hydrogen; commercial catalyst; selectivity; deactivation

1. Introduction

The technological development of hydrolysis-fermentation of lignocellulosic biomass explains that secondgeneration bioethanol is considered to be an alternative sourceto oil for the production of automotive fuels and petrochem-icals [1−3] and to fulfill the increasing demand of H2 foruse as fuel and raw materials in the hydroprocessing units inrefineries.

Viikari et al. [4] review the aspects concerning the de-crease in lignocellulosic bioethanol production costs, withtheir estimation being in 0.41−0.57 euro/L range in the EUand in 0.24−0.34 euro/L range in the USA, which is encour-aging for the viability of its valorization at industrial scale ina near future. Moreover, bioethanol is a more attractive rawmaterials for use in fuel cell hydrogen production, which pro-vides higher electrical and thermal efficiency in solid oxidefuel cell (SOFC) than other renewable fuels, such as biogasand glycerol [5].

The steam reforming of bioethanol (SRE) is an attractiveroute to obtain H2, as it avoids the energy-demanding waterseparation processes needed for upgrading bioethanol as a fuel

or in its valorisation following other routes. The reaction takesplace with the following stoichiometry:

CH3CH2OH+ 3H2O→ 6H2 + 2CO2 (1)

Several undesired reactions also take place, which di-minish H2 yield, give way to the formation of byproducts inthe reaction medium (CO, CH4, acetaldehyde, ethylene) andcontribute to catalyst deactivation by coke deposition fromthese byproducts. These reactions include: ethanol dehydro-genation to acetaldehyde, ethanol decomposition to CO+CH4

and H2, ethanol dehydration to ethylene (which is consid-ered as one of the main responsible for coke formation byintermediate oligomers), acetaldehyde decomposition to COand CH4, which in turn contribute to H2 formation by water-gas shift reaction (WGS) and reforming reaction, respec-tively, Boudouard reaction, CO and CO2 methanation reac-tions, methane decomposition to carbon and H2 [6−8].

Catalyst deactivation by coke is one of the main prob-lems in SRE reaction and its attenuation requires operationto be carried out at a sufficiently high temperature for themain reactions reforming ethanol, intermediate compoundsand byproducts, such as CH4. Moreover, given that a cata-

∗ Corresponding author. Tel: +34-94-6015363; Fax: +34-94-6013500; E-mail: javier.erena@ehu.esThis work was supported by financial support of the Ministry of Science and Technology of the Spanish Government (Projects CTQ2009-13428 and

CTQ2012-35263), the University of the Basque Country (UFI 11/39) and the Basque Government (Project IT748-13).

Copyright©2014, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. All rights reserved.doi: 10.1016/S2095-4956(14)60195-9

640 Jorge Vicente et al./ Journal of Energy Chemistry Vol. 23 No. 5 2014

lysts with a good balance between activity, selectivity to H2

and stability is required, numerous metallic functions havebeen tested, such as transition metals (Ni, Co) and noble met-als (Pt, Pd, Rh), as well as bimetallic alloys, which have beensupported on different materials (Al2O3, SiO2, MgO, ZnO,TiO2, CeO2) [9−16].

This paper deals with the kinetic behaviour of methanesteam reforming (SRM) commercial catalysts which are usedin SRE reaction, in 300−700 ◦C range and in an isothermalfluidized-bed reactor (in order to ensure bed isothermicity inthe experimental runs). The results are a basis for the sub-sequent studies involving the adaptation of catalysts preparedin the laboratory to industrial scale, bearing in mind that thefluidized bed has suitable features for large-scale operation.Ni and Rh based catalysts have been studied as alternatives,given that catalysts cost is a key factor for the viability of theindustrial implementation of SRE reaction.

2. Experimental

2.1. Catalysts

Five commercial reforming catalysts have been sup-plied by Sud Chemie (G90 and G117), BASF (SG9301 andSG9402) and Fuel Cell Materials (Rh/ZDC), with the follow-ing formulations: G90 (NiO, CaAl3O4, Al2O3), G117 (NiO,MgO), SG9301 (NiO, CaO, La2O3, Al2O3), SG9402 (NiO,CaO, Al2O3), Rh/ZDC (Rh, CeO2, ZrO2). More detailedinformation on the composition of these catalysts cannot beprovided due to a confidentiality agreement. In order to beused in the fluidized reactor, the catalysts provided by SudChemie and BASF (which are supplied as rings) were pre-viously ground and sieved to 150−250 μm range. Rh/ZDCcatalyst (supplied as powder) was pressed, ground and sievedto the same size, which is suitable for use in the fluidized re-actor. As stated in the confidentiality agreement, compositionand chemical properties of these catalysts were not measured.

The physical properties (Table 1) were determined by N2

adsorption-desorption in a Micromeritics ASAP 2010C andHg porosimetry in a Micromeritics Autopore 9220. The Mi-cromeritics ASAP 2010C was also used for the analysis of themetallic surface by H2 chemisorption. Rh/ZDC commercialcatalyst (with CeO2 and ZrO2 as support) had high values ofBET surface area and pore volume, and the values of bothproperties were very low for the rest of the catalysts.

Table 1. Metal content and physical properties ofcommercial catalysts used for SRE

Metal contenta SBET Vpore dporeCatalysts(wt%) (m2/g) (cm3/g) (A)

Rh/ZDC 2 76 0.28 127SG9301 10−15 21 0.07 162SG9402 11−13 26 0.06 135

G90 >1 19 0.04 122G117 5 8 0.02 149a Provided by the supplier

2.2. Reaction conditions and equipment for reaction andproduct analysis

The kinetic runs were carried out in automated reactionequipment (Microactivity Reference from PID Eng & Tech)provided with an isothermal fluidized-bed reactor (22 mm ofinternal diameter and total length of 460 mm) connected on-line to a gas chromatograph (Agilent Micro-GC 3000) forproduct analysis [17]. The hydrodynamic properties of thebed were improved by mixing the catalyst (particle size be-tween 150 and 250 μm) with an inert solid (CSi, particle sizebetween 60 and 90 μm) at a catalyst/inert ratio of 1 : 4. TheMicro-GC was provided with four modules for analyzing: (1)permanent gases; (2) oxygenates, light olefins (C2–C3) andwater; (3) C2–C6 hydrocarbons; (4) C6–C12 hydrocarbons andoxygenate compounds. The compounds were quantified andidentified using calibration standards of known concentration.The balance of atoms (C, H, O) was closed in all the runsabove 98%.

The kinetic runs were performed under the follow-ing operating conditions: 300−700 ◦C; pressure, 1.2 bar;ethanol partial pressure in the feed (diluted in He), 0.11 bar;steam/ethanol molar ratio, 6; space time, 0.14 gcat·h·g

−1ethanol;

time on stream, up to 50 h. These operating conditions wereselected on the basis of a previous study [18] for delimitingthe proper ranges of operating variables to obtain kinetic re-sults and attain a stable fluidization regime. As proven in thisprevious paper, a steam/ethanol ratio higher than 6 was notinteresting because the improvement in H2 was low and theenergy cost was excessive.

3. Results and discussion

3.1. Kinetic behaviour of the catalysts at zero time on stream

The following reaction indices have been defined in orderto quantify the kinetic behaviour of the catalysts:

Ethanol conversion:

X =FE,0−F E

FE,0(2)

where, FE,0 and FE are the molar flow rates of ethanol at thereactor inlet and outlet, respectively.

Yield of H2:

YH2 =FH2

6FE,0(3)

where, FH2 is the molar flow rate of H2 at the reactor outlet.Selectivity to i product (H2, CO2, CO, CH4 and C2H4O):

Si =Fi

νi(FE,0−F E)(4)

where, Fi is the molar flow rate of i product at the reactor out-let and νi is the corresponding stoichiometric coefficient byassuming the independent formation from ethanol: νi = 6 forH2, 2 for CO2, CO and CH4, 1 for C2H4O (acetaldehyde,which was formed by ethanol dehydrogenation and whose

Journal of Energy Chemistry Vol. 23 No. 5 2014 641

concentration in the product stream was significant at lowtemperature).

Table 2 shows the results of ethanol conversion (Equa-tion 2) and H2 yield (Equation 3) in 300−700 ◦C range andfor the same values of the remaining operating conditions(steam/ethanol molar ratio, space time, ethanol partial pres-sure in the feed). As observed, the catalysts were very activeand the effect of temperature on ethanol conversion was note-worthy, with conversion being almost full at 500 ◦C (exceptSG9402 catalyst).

Below 500 ◦C, ethanol conversion decreased as follows:Rh/ZDC>>SG9301>SG9402>G90>G117, which confirmsthe well known ability of Rh for C–C bond breakage [19−21],and also the higher activity of Ni catalysts supported on Al2O3

than that supported on MgO. The yield of H2 at low tem-peratures (≤400 ◦C) for different catalysts followed the sametrend as that of ethanol conversion, although the differenceswere lower than those in conversion. The differences in H2

yield were even lower at high temperature (500 ◦C).

In order to explain the differences in H2 yield in Table 2(which were not directly related to the differences observedin ethanol conversion), product distribution should be consid-

ered, which, apart from temperature, depends on the type ofmetal and the metal properties in the catalyst. Figure 1 showsthe effect of reaction temperature on the selectivity (Equation4) to H2 (Figure 1a), CO2 (Figure 1b), CO (Figure 1c), CH4

(Figure 1d) and C2H4O (acetaldehyde formed by ethanol de-hydrogenation) (Figure 1e) for different catalysts.

Table 2. Effect of temperature on the initial values of ethanolconversion and H2 yield for the different catalysts

Reaction temperature ( ◦C)300 400 500 600 700

Ethanol conversionRh/ZDC 0.80 0.97 1.00 1.00 1.00SG9301 0.51 0.78 0.97 1.00 1.00SG9402 0.52 0.64 0.83 0.95 1.00G90 0.32 0.58 1.00 1.00 1.00G117 0.16 0.56 0.99 1.00 1.00Hydrogen yieldRh/ZDC 0.21 0.30 0.46 0.66 0.84SG9301 0.11 0.27 0.45 0.66 0.84SG9402 0.09 0.22 0.42 0.65 0.83G90 0.08 0.19 0.46 0.67 0.84G117 0.03 0.17 0.51 0.68 0.85

Reaction conditions: steam/ethanol molar ratio, 6; space time,0.14 gcat·h·g−1

ethanol; ethanol partial pressure in the feed, 0.11 bar

Figure 1. Effect of temperature on the selectivity to H2 (a), CO2 (b), CO (c), CH4 (d) and C2H4O (e) at zero time on stream for different catalysts. Reactionconditions: steam/ethanol molar ratio, 6; space time, 0.14 gcat·h·g

−1ethanol; ethanol partial pressure in the feed, 0.11 bar

Figure 1(a) shows that H2 selectivity, which was similarfor different catalysts, increased almost linearly with temper-ature, and that CO selectivity (Figure 1c) reached a minimumat 500 ◦C. At low or moderate temperatures, CO methanationreaction was favoured, whereas an increase in temperaturefavoured CH4 reforming and C2H4O reforming and decompo-

sition (to CO and CH4), which gives way to a significant de-crease in CH4 selectivity above 500 ◦C (Figure 1d), whereasC2H4O selectivity (Figure 1e) was almost negligible. Roy etal. [22] have proven the significance of C2H4O as reactionintermediate (formed by ethanol dehydrogenation and decom-posed by C–C breakage to CO and CH4), whose presence in

642 Jorge Vicente et al./ Journal of Energy Chemistry Vol. 23 No. 5 2014

the product stream was noticeable at low temperature.Moreover, catalyst composition has a major effect on the

distribution of by-products. The selectivity to CO2 (Figure 1b)was higher for Rh based catalyst (Rh/ZDC catalyst), whichevidences that this catalyst is highly active for the direct re-forming of ethanol. It should be noted that there was no ac-etaldehyde in the product stream when this catalyst was used(Figure 1e), which is attributed to the well known ability ofRh for C–C breakage [23]. The high reforming activity ofG90 catalyst also explained the low selectivity to acetalde-hyde (Figure 1e). The higher selectivity to CH4 over Rh/ZDCcatalyst (Figure 1d), as well as its lower selectivity to CO(Figure 1c) evidenced the higher activity of this catalyst formethanation reaction than Ni catalysts supported on Al2O3

(SG9301, SG9402 and G90), and of the latter compared withthat supported on MgO (G117). This higher CH4 selectivityof Rh/ZDC explained that, although ethanol conversion wassignificantly higher, the yield of H2 obtained with this cata-lyst was only slightly higher than that obtained with Ni basedcatalysts.

It should be noted that the results obtained in SRE reac-tion using these commercial catalysts (specifically designedfor methane reforming and well contrasted for this industrialscale process) were close to those previously obtained un-der the same operating conditions with Ni or Co catalyst ondifferent supports (SiO2 and α-Al2O3 alone or doped withLa2O3), which have been synthesized following the recom-mendations in the literature for a good performance in the re-forming of ethanol [18]. As expected, these catalysts had agood kinetic behaviour for methane reforming above 600 ◦C(Figure 1d) and also for C2H4O decomposition, given thattheir selectivity was insignificant above 600 ◦C. This goodbehaviour of the catalysts makes the purification of H2 streameasier by the subsequent steps aimed at the reforming of theremaining methane.

3.2. Catalysts’ stability

In order to compare the stability of the catalysts in de-tail, the results corresponding to two relevant temperatures forethanol reforming have be considered: i) 500 ◦C, correspond-ing to the minimum selectivity to CO in the product stream(Figure 1c), which is important to facilitate the purification ofH2 by the subsequent reaction steps, and ii) 700 ◦C (with thecatalysts that give way to a higher stability at 500 ◦C), whichis suitable to avoid coke deposition on the metallic sites bygasification of the coke precursors, and so minimizing cata-lyst deactivation.

Figure 2 shows the results of evolution of ethanol con-version and H2 yield with time on stream for the commercialcatalysts used at 500 ◦C. It is observed that Rh/ZDC catalystwas very stable and that ethanol conversion (which was full)and H2 yield (around 45%) remained constant with time onstream. The capacity of CeO2-ZrO2 support for oxygen stor-age will contribute to enhancing the stability, as it facilitatesthe gasification of coke precursors [24].

Figure 2. Evolution of ethanol conversion and H2 yield with time on streamfor different catalysts. Reaction conditions: 500 ◦C; steam/ethanol molar ra-tio, 6; space time, 0.14 gcat·h·g

−1ethanol; ethanol partial pressure in the feed,

0.11 bar

G90 catalyst (of Ni/Al) was partially deactivated once thefirst 12 h time on stream have elapsed and reached a pseudo-stable state, with ethanol conversion having a constant valueof 0.83. Moreover, H2 yield was similar to that obtained withRh catalyst and remained almost constant with time on stream,because deactivation has a lower impact on steam reformingthan on byproduct formation reactions. Therefore, H2 selec-tivity increased and those of the other by-products decreased,except the one of acetaldehyde, whose yield was very low.The catalysts SG9301 and SG9402 (the latter is not shown inFigure 2), whose supports were based on Al2O3 like G90 cat-alyst, have a very similar kinetic behaviour to the latter. G117catalyst (of Ni/MgO) underwent steady deactivation whichwas the most unstable of all the catalysts studied. Conse-quently, the values of average H2 production rate for 20 h timeon stream (calculated with Equation 5) using Rh/ZDC, G90,SG9301, SG9402 and G117 catalysts were 422, 409, 410, 405and 278 mmolH2 ·g

−1cat ·h

−1, respectively.

rp,H2 =1

W · t

t∫0

FH2 dt (5)

An analysis of the evolution of by-products yields (CO2,CO, CH4 and C2H4O) with time on stream is interestingfor gaining knowledge on how coke deposition affects thedifferent individual reactions involving ethanol reforming andalso for assessing the perspectives of the reactor outlet streampurification in order to intensify H2 production and minimizethe concentration of undesired compounds for this stream val-orisation. The results (Figure 3) confirmed the excellent be-haviour of Rh/ZDC and G90 catalysts. With the former, theyields remain constant with time on stream, i.e., a CO yieldof 5% and a very low yield of C2H4O (0.2%), but with a lessfavourable result concerning CH4 yield (46%). G90 catalystalso performed well, given that CH4 yield in this case waslower (between 30% and 40%), although C2H4O yield washigher (3%).

Journal of Energy Chemistry Vol. 23 No. 5 2014 643

Figure 3. Evolution of the yields of CO2 (a), CO (b), CH4 (c) and C2H4O (d) with time on stream for different catalysts. Reaction conditions: 500 ◦C;steam/ethanol molar ratio, 6; space time, 0.14 gcat·h ·g

−1ethanol; ethanol partial pressure in the feed, 0.11 bar

Based on the aforementioned results for the commercialcatalysts, Rh/ZDC and G90 catalysts were very active and sta-ble for ethanol steam reforming at 500 ◦C, with a low yield ofCO, 5% for the former and 8% for the latter. This result isvery interesting for the use of these catalysts at low tempera-ture and evidences their versatility, given that they have beendesigned for operating above 700 ◦C in methane reforming.

As mentioned in Section 3.1, temperature has a sig-nificant effect on product distribution. Its effect on catalystdeactivation should also be noted, given that the deactiva-tion by coke is attenuated by the gasification of intermediatecompounds that are precursors of carbon deposits, and thisgasification is enhanced as temperature is increased. Never-theless, an excessive increase in temperature may favour theundesired reactions and also causes irreversible deactivationof the catalyst by metal sintering and subsequent decrease inmetal dispersion with time on stream. The latter effect shouldnot be a problem for the catalysts studied, which have beendesigned for operating above 700 ◦C in methane reforming.

Table 3 gathers the results of product yields for the cat-alysts of highest interest (G90 and Rh/ZDC) at 700 ◦C un-der given operating conditions. At this temperature, neitherC2H4O nor other oxygenates or hydrocarbon products ap-peared in the product stream. Furthermore, ethanol conversionwas almost full and both catalysts were totally stable, with

the individual yields remaining constant in the runs of 50 hduration. Both G90 and Rh/ZDC commercial catalysts, gaveway to a similar H2 yield (84%) and a low CH4 yield (lowerfor Ni catalyst (G90) than for Rh catalyst), which is relevantin order to simplify the purification treatments of H2 stream.The average H2 production rate at 700 ◦C (calculated withEquation 5) for both catalysts was almost the same, around770 mmolH2 ·g

−1cat ·h

−1.

Table 3. Product yields in SRE on G90 and Rh/ZDC catalysts

YieldCatalysts

H2 CH4 CO CO2

Rh/ZDC 0.84 0.05 0.40 0.60G90 0.84 0.03 0.41 0.58

Reaction conditions: 700 ◦C; steam/ethanol molar ratio, 6; spacetime, 0.14 gcat·h·g

−1ethanol; ethanol partial pressure in the feed,

0.11 bar

4. Conclusions

The results prove that methane reforming commercial cat-alysts are versatile and perform well for ethanol steam reform-ing. The comparison between the yields of H2 and reactionby-products (CO2, CO, CH4, C2H4O) evidences the majorinfluence of temperature and, to a minor extent, of catalyst

644 Jorge Vicente et al./ Journal of Energy Chemistry Vol. 23 No. 5 2014

composition on product distribution. A good behaviour hasbeen proven for Rh/ZDC catalyst at 500 ◦C under the condi-tions of minimum CO formation, which is attributed to thehigh reforming activity of Rh and the ability of CeO2-ZrO2

support for oxygen storage. Furthermore, as an alternative(preferable due to its lower cost), the excellent behaviour ofG90, SG9301 and SG9402 catalysts should be noted, which isconsequence of a suitable synergy of Ni with Al2O3 used assupport. These catalysts are stable at 700 ◦C, with the reactionindices being constant throughout time on stream

AcknowledgementsThis work has been carried out with financial support of the

Ministry of Science and Technology of the Spanish Government(Projects CTQ2009-13428 and CTQ2012-35263), the Universityof the Basque Country (UFI 11/39) and the Basque Government(Project IT748-13). The authors are grateful to Sud Chemie, BASFand Fuel Cell Materials for supplying the catalysts.

NomenclaturesFE, FE,0 molar flow rate of ethanol at the reactor outlet and in

the feed, respectively, mmol·h−1

FH2 molar flow rate of H2 at the reactor outlet, mmol·h−1

Fi molar flow rate of i component at the reactor outlet,

mmol·h−1

rp,H2 average H2 production rate, mmolH2 ·g−1cat ·h

−1

Si selectivity of i component

W catalyst weight, g

X ethanol conversion

YH2 hydrogen yield

νi stoichiometric coefficient of i component

AbbreviationsG117 catalyst supplied by Sud Chemie, with NiO, MgO in

the formulation

G90 catalyst supplied by Sud Chemie, with NiO, CaAl3O4,

Al2O3 in the formulation

Rh/ZDC catalyst supplied by Fuel Cell Materials, with Rh,

CeO2, ZrO2 in the formulation

SG9301 catalyst supplied by BASF, with NiO, CaO, La2O3,

Al2O3 in the formulation

SG9402 catalyst supplied by BASF, with NiO, CaO, Al2O3 in

the formulation

SOFC solid oxide fuel cell

SRM steam reforming of CH4

SRE steam reforming of ethanol

WGS water-gas shift reaction

References

[1] Gayubo A G, Alonso A, Valle B, Aguayo A T, Bilbao J. ApplCatal B, 2010, 97: 299

[2] Gayubo A G, Alonso A, Valle B, Aguayo A T, Olazar M, BilbaoJ. Fuel, 2010, 89: 3365

[3] Gayubo A G, Alonso A, Valle B, Aguayo A T, Bilbao J. Ind EngChem Res, 2010, 49: 10836

[4] Viikari L, Vehmaanpera J, Koivula A. Biomass Bioenerg, 2012,46: 13

[5] Saebea D, Authayanun S, Patcharavorachot Y, Paengjuntuek W,Arpornwichanop A. Energy, 2013, 51: 305

[6] Haryanto A, Fernando S, Murali N, Adhikari S. Energy Fuels,2005, 19: 2098

[7] Vaidya P D, Rodrigues A E. Chem Eng J, 2006, 117: 39

[8] Ni M, Leung D Y C, Leung M K H. Int J Hydrog Energy, 2007,32: 3238

[9] Sun J, Qiu X P, Wu F, Zhu W T. Int J Hydrog Energy, 2005, 30:437

[10] Akande A J, Idem R O, Dalai A K. Appl Catal A, 2005, 287:159

[11] Akande A, Aboudheir A, Idem R, Dalai A. Int J Hydrog Energy,2006, 31: 1707

[12] Casanovas A, Roig M, de Leitenburg C, Trovarelli A, Llorca J.Int J Hydrog Energy, 2010, 35: 7690

[13] Le Valant A, Bion N, Can F, Duprez D, Epron F. Appl Catal B,2010, 97: 72

[14] Wu C F, Williams P T. Environ Sci Technol, 2010, 44: 5993

[15] Wu C F, Williams P T. Appl Catal B, 2011, 102: 251

[16] Espinal R, Taboada E, Molins E, Chimentao R J, Medina F,Llorca J. Appl Catal B, 2012, 127: 59

[17] Erena J, Vicente J, Aguayo A T, Gayubo A G, Bilbao J. Int JHydrog Energy, 2013, 38: 10019

[18] Vicente J. [PhD Dissertation]. Bilbao: University of the BasqueCountry UPV/EHU, 2012

[19] Aupretre F, Descorme C, Duprez D. Catal Commun, 2002, 3:263

[20] Breen J P, Burch R, Coleman H M. Appl Catal B, 2002, 39: 65

[21] Frusteri F, Freni S, Spadaro L, Chiodo V, Bonura G, Donato S,Cavallaro S. Catal Commun, 2004, 5: 611

[22] Roy B, Martınez U, Loganathan K, Datye A K, Leclerc C A. IntJ Hydrog Energy, 2012, 37: 8143

[23] Karim A M, Su Y, Sun J M, Yang C, Strohm J J, King D L,Wang Y. Appl Catal B, 2010, 96: 441

[24] Roh H S, Wang Y, King D L. Top Catal, 2008, 49: 32