Hydrogen production from supercritical water reforming of glycerol over Ni/Al2O3–SiO2 catalyst

Depósito de Investigación de la Universidad de Sevilla https://idus.us.es/ This is an Accepted Manuscript of an article published by Elsevier in Energy, Vol. 84, on May 2015, available at: https://doi.org/10.1016/j.energy.2015.03.046 © 2015 Elsevier. En idUS Licencia Creative Commons CC BY-NC-ND 1 Hydrogen production from supercritical water reforming of glycerol over Ni/Al2O3-SiO2 catalyst F. J. Gutiérrez Ortiz, F. J. Campanario, P. G. Aguilera, P. Ollero Departamento de Ingeniería Química y Ambiental, Universidad de Sevilla Camino de los Descubrimientos, s/n. 41092 Sevilla, Spain Phones: + 34 95 448 72 68 / 60 e-mail: “Francisco Javier Gutiérrez Ortiz” [email protected] Abstract Hydrogen production from the supercritical water reforming of glycerol was studied in a tubular fixed-bed reactor by using a Ni-based catalyst supported on Al2O3 and SiO2. Tests were carried out at a pressure of 240 bar, temperatures of 500-800 ºC, glycerol feed concentrations of 5-30 wt.%, and weight hourly space velocity from 1.25 to 22.5 gGlyꞏh-1ꞏgCat-1 (residence time from 1.6 to 4.8 s through the bed). The dry gas is mainly composed of H2, CO2, CO, CH4. The results showed that the glycerol conversion was almost complete, except at the highest glycerol feed concentration and lowest temperature. Hydrogen yields were very close to those values predicted by equilibrium at a short residence time. Nickel on catalyst was completely reduced, and structured carbon nanotubes were encountered at glycerol concentrations higher than 20 wt.%. This study illustrates that the reforming of glycerol using supercritical water over a Ni catalyst makes it possible to reduce the reforming temperature needed when no catalyst is used (from 800 ºC to 600 ºC), achieving a high-yield hydrogen production, very close to equilibrium, and requiring less energy. Keywords Reforming, supercritical water, glycerol, hydrogen, catalyst, nickel 2 1. Introduction The rising surplus of biodiesel-derived crude glycerol from the transesterification process requires a further processing, but conventional options for crude glycerol consist of refining it to a higher purity that is costly. Therefore, glycerol valorization is essential. The transformation of glycerol by reforming into an energy carrier (syngas or hydrogen) is one of the most attractive routes of valorization. There exist different reforming processes that continue under study, such as steam reforming [1], autothermal reforming [2] and aqueous phase reforming [3]. However, the use of supercritical water makes it possible to perform another type of reforming process. Supercritical water (SCW) has relevant thermophysical properties [4, 5], and it is extremely reactive, so it is possible to perform the process without adding a catalyst, which has been experimentally verified. This is possible at high temperatures (about 800 ºC, and above), and long residence times, which favors the carbon -to-gas efficiency and the hydrogen yield [6]. Earlier studies about biomass gasification by SCW without catalyst [7-10] and with catalyst [11-16] have been published. Only some of the above-cited papers point out that a very high temperature (> 700 ºC and even more) is required to achieve acceptable conversions. In addition, most of the studies performed with SCW have used very small biomass feed concentrations, generally lower than 10 wt.%. Previously, using Aspen Plus and assuming equilibrium conversion, we have verified that a high hydrogen yield can be obtained at temperatures higher than 800 ºC and high glycerol concentrations [17]. Additionally, we have also developed a heat-integrated process focused on using a high temperature and a glycerol feed concentration that assures the energy selfsufficiency [18, 19]. This condition requires a relatively high glycerol feed concentration, which decreases from about 27 wt.% at 800 ºC, to 21 wt.% at 700ºC, 18 wt.% at 600 and 15 wt.% at 500 ºC, operating at 240 bar. 3 This work is focused to carry out the supercritical water reforming (SCWR) process aided by a Ni-based catalyst in order to inspect the minimum feasible reforming temperature (and, hence, the energy required for the process) and residence time (and the reactor size) required, by operating at 240 bar and glycerol feed concentrations from 5 to 30 wt.%. Another aim is to compare the process performance with and without using a catalyst. Finally, results are also compared to other few studies of SCW reforming over Ni catalyst. 2. Materials and methods The experiments were carried out with pure glycerol (Panreac PRS-Codex 100% ± 1% purity) and deionized water (to avoid material corrosion) obtained by a high-performance water treatment process that combines an activated carbon and an ion-exchange resin. A commercial Ni-based catalyst supported on Al2O3 and SiO2 was used (Sigma-Aldrich 208779), with 66.4 wt.% Ni, measured by inductively coupled plasma atomic emission spectroscopy (Horiba Jobin Yvon Ultima 2 High resolution ICP-OES spectrometer). The catalyst had the following main characteristics: - BET surface area: 166.2 m2/g; - Crystal structure: amorphous (50%), crystalline (balance, identifying nickel (15%) and NiO (5%)); - Specific pore volume: 0.27 mL/g; average pore diameter: 6.02 nm; - Shape: dust, which is pressed to make a compact tablet, and then milled and sieved between 1.00 and 2.83 mm. The BET surface area as well as the micropores and mesopores areas and volumes were obtained by means of physisorption with nitrogen as adsorptive (Micromeritics ASAP 2420 System). Structural information on an atomic scale of both crystalline and non-crystalline 4 (amorphous) materials was acquired by X-ray Diffraction (XRD) (Bruker D8 Advance A25 equipment). Fig. 1 illustrates a basic diagram of the experimental plant. A glycerol-water mixture is pumped (HPLC) to a desired pressure and then heated to the operating temperature by an electrical tubular furnace, in which the supercritical water reforming reactor is (only 2.0 m length are inside the furnace; 7.92 mm ID) made of Inconel 625. The product gas from the reformer is cooled down to 40 ºC, thus obtaining a condensate consisting mainly of water. Then, the mixture of liquid and gas is expanded through a back-pressure valve. Next, the condensate is separated from the gas in a knock-out drum. The mass-flow rate of the gas stream is measured, and this stream is sampled and analyzed in a gas chromatograph (GC Agilent 7890A) continuously. More details can be found elsewhere [6], in which an initial catalytic effect of the reactor material on the gas product yields was verified. This catalytic activity of the reactor material was decreasing to zero after a few tens of operating hours. The current catalytic inactivity has been confirmed in subsequent tests, so there was no other catalytic activity than that of the added heterogeneous catalyst in the tests carried out in this study. The tests lasted 3-6 hours. All of the experiments were conducted at 240 bar, and the furnace temperature was changed from 500 to 800 ºC to measure the effect of the temperature on the reforming. The weight hourly space velocity (WHSV), defined as mass flow-rate of glycerol divided by mass of catalyst, was varied from 1.25 to 22.5 gGlyꞏh-1ꞏgCat-1 by changing the feed flow-rate along with the glycerol concentration, which was varied from 5 to 30 wt.%. Likewise, the residence time through the bed, which was computed at the outlet reactor conditions by Eq. (1), changed between 1.6 and 4.8 s: feed RRproductR rm PTV  ),(  [1] 5 where VR is the volume of the reactoris the void fraction in the catalyst bed (0.70), product is the density of the product gas at the reformer temperature and pressure, and feed m  is the mass flow rate of the feed. The catalyst load was always 20 g, occupying about 400 mm inside the reactor, at 500 mm from the reactor outlet, and it was capped by pieces of glass fiber. The catalyst was not reduced before testing. Regarding the experimental procedure, each experiment started by flowing pure deionized water through the reactor, and increasing the set-point of the back-pressure regulator during 45 min up to the reference value of 240 bar. Then, the furnace was powered to reach the desired temperature (500-800 ºC) during the next 45 - 75 min. Therefore, the start-up took about 1.5 - 2.0 hours. Afterwards, the glycerol was added to the water to effectively start the test. At the end of the test, the furnace was switched off, the pressure was reduced, and only deionized water was entered. Thus, before and after each experiment, the water-insoluble catalyst was inserted in or picked up the reactor, and the reactor and the rest of the tubes were cleaned by pumping pure water to flush any remaining solids or other substances present in the reactor. After washing, carbonaceous residue was recovered by vacuum filtration, and quantified by drying and weighing the residue, i.e., the solid carbon content was determined by a gravimetric method. During the operation, once the system was stable in temperature, pressure and gas flow-rate, a fraction of the product gas was derived to the GC, and some liquid samples were taken from the bottom of the gas-liquid knock-out drum, and cooled-saved for analysis. The liquid samples were analyzed by a gas chromatograph equipped with a mass spectrometry detector Ultra Shimadzu QP 2010. The samples were prepared by diluting them with 1-Butanol to avoid damaging the column due to the very high water content in samples. The analysis method was performed at temperatures from 45 to 225 °C using heating ramps during 28.5 min. Mass range (m/z) is between 10 and 150. 6 The microstructure of the catalyst (before and after being tested) and carbon residues was characterized by SEM-EDS (PHILIPS XL-30 SEM microscope), TEM (PHILIPS CM-200 microscope), XRD and Raman spectroscopy (Confocal Raman Bruker Senterra Microscope). X-ray Photoelectron Spectroscopy (XPS) was used to discriminate on the different oxidation states (ESCA Phoibos HSA3500 150 MCD-9). Carbon balance was calculated from the carbon content in the feed, and the total carbon found in the gaseous and liquid products. Then, the result was compared with the carbon residue recovered, and the difference was the experimental error. The overall reforming reaction is shown in the Eq. (2), as follows: C3H8O3 + 3H2O  3CO2 + 7H2  H0 = + 122.83 kJ/mol [2] where H0 is the enthalpy of reaction at standard conditions (25ºC, 1 atm). The performance variables were computed as follows: Hydrogen yield: feedGly H N N YieldH , 2 2  [3] Glycerol conversion: 100 , ,,    feedGly liquidGlyfeedGly Gly N NN [4] Carbon-to-gas efficiency: 100 3, 42     feedGly CHCOCO GC N NNN [5] Carbon-to-liquid efficiency: 100 , ,  feedC liquidC LC N N [6] Carbon-to-solid efficiency: 100 , ,  feedC solidC SC N N [7] where N denotes the number of moles of the compound referred in the subscript (see Nomenclature), C-L gives the total organic carbon in the liquid product, and C-S takes into account the solid carbon recovered. Carbon balance involves the following relationship: 7 100  LCLCGC [8] Chemical equilibrium calculations were performed by Aspen Plus™ to compare them with the experimental results. The thermodynamic method used was the predictive Soave- Redlich-Kwong (PSRK), as the most suitable to describe the supercritical state [17]. The reforming reactor was simulated as an R-Gibbs reactor, in which the product gas composition and the heat of the overall reaction are calculated under conditions that minimize the Gibbs free energy. The simulation did not predict coke formation for any of the experimental conditions tested in this paper. For all of the operating conditions simulated, the glycerol conversion was always 100%, at equilibrium condition. 3. Results and discussion The conversion of glycerol was always greater than 99%, so it can be considered that the conversion was complete, except for the tests at temperatures of 500 ºC where much lower conversions were obtained (below 45%) and tests with concentrations higher than 20 wt.%, in which the conversion ranges from 90% to 99%. Table 1 shows some significant results. Most of the tests were replicated two or three times, and those with an average error in the carbon balance higher than 10% were discarded. Almost all of the product gas (> 98.2 vol.%) always consisted of H2, CO2, CO and CH4. Apart from unconverted glycerol, some other products were detected as traces (acetaldehyde, hydroxyacetone, propionaldehyde, acetone, methanol, ethanol, and acrolein) in the analysis of a few samples of liquid corresponding to tests with glycerol concentrations equal or greater than 20 wt.%. Because the glycerol was entered the reactor in an aqueous solution, which is heated from ambient temperature to the reaction temperature, glycerol undergoes a progressive thermal decomposition through reaction intermediates before arriving at the catalyst section, at the end of the reactor. Based on the results of our study performed without adding a catalyst [6], two reaction pathways of dehydration of glycerol may be probable to occur. Thus, in the first, 8 glycerol (C3) decomposes to acetaldehyde (C2) and formaldehyde (C1), and in the other one, glycerol is converted to acrolein (C2), in the section of hot-compressed water. In the supercritical region, at high temperature, the acetaldehyde decomposes into methane and carbon monoxide, while the formaldehyde decomposes into carbon monoxide and hydrogen. Additionally, acrolein may decompose to ethylene and carbon monoxide; ethylene is cracked, and solid carbon may be hydrogenated or gasified into hydrogen, methane and carbon monoxide [6]. Next, two reactions can take place: Water-gas shift (WGS): CO + H2O  CO2 + H2  H0 = - 41.17 kJ/mol [9] Methane reforming: CH4 + H2O  CO + 3H2  H0 = + 206.18 kJ/mol [10] Consequently, as depicted in Fig. 2, in any of the two dehydration pathways, the thermal decomposition of one molecule of glycerol leads to one molecule of hydrogen, one molecule of methane and two molecules of carbon monoxide, if the glycerol conversion is complete and no solid carbon is formed. Therefore, if the reaction (9) were much faster than the reaction (10), the number of moles of CH4, CO2 and H2 per mol of glycerol fed would be one, two and three, respectively. Equally, if the methane reforming were much faster than the WGS reaction, the result would be three moles of CO and four moles of H2 per mol of glycerol. Indeed, both reactions (9) and (10) will simultaneously occur in a more or less extent depending on the operating conditions. However, as the catalyst shows a high activity, gas product yields become close to equilibrium. 3.1. Effect of the temperature Fig. 3 depicts the gas product yields at different temperatures (500-800 ºC), at 10 and 20 wt.% glycerol feed concentration, 240 bar and 1.0 L/h feed flow-rate, i.e., at two WHSV of 5 and 10 gGlyꞏh-1ꞏgCat-1 through the bed (both corresponding to a residence time of about 2.4 s). The solid lines represent equilibrium values. 15 Subscripts: feed: feed entering the reactor liquid: liquid product from the knock-out drum product: product gas leaving the reactor solid: solid residue generated R: reactor Gly: glycerol Cat: catalyst C: carbon CO: carbon monoxide CO2: carbon dioxide CH4: methane H2: hydrogen Superscripts: C: carbon S: solid phase L: liquid phase G: gas phase Acknowledgments The Science and Technology Ministry of Spain supported this research under the research project ENE2009-13755. References [1] Nichele V, Signoretto M, Menegazzo F, Gallo A, Dal Santo V, Cruciani G, Cerrato G. Glycerol steam reforming for hydrogen production: Design of Ni supported catalysts. Appl Catal B-Environ 2012;111–112:225–232 16 [2] Wang H, Wang X, Li M, Li S, Wang S, Ma X. Thermodynamic analysis of hydrogen production from glycerol autothermal reforming. Int J Hydrogen Energy 2009;34(14): 5683-5690 [3] Guo Y, Liu X, Azmat MU, Xu W, Ren J, Wang Y, Lu G. Hydrogen production by aqueous-phase reforming of glycerol over Ni-B catalysts. Int J Hydrogen Energ 2012;37:227-234 [4] Savage PE. Organic chemical reactions in supercritical water. Chem Rev 1999;99:603- 622 [5] Guo Y, Wang SZ, Xu DH, Gong YM, Ma HH, Tang XY. Review of catalytic supercritical water gasification for hydrogen production from biomass. Renew Sust Energ Rev 2010;14:334–343 [6] Gutiérrez Ortiz FJ, Serrera A, Galera S, Ollero P. Experimental study of the supercritical water reforming of glycerol without the addition of a catalyst. Energy 2013;56:193-206 [7] Yu D, Aihara M, Antal MJ. Hydrogen production by steam reforming glucose in supercritical water. Energ Fuel 1993;7(5):574-577 [8] Buhler W, Dinjus E, Ederer HJ, Kruse A, Mas C. Ionic reactions and pyrolysis of glycerol as competing reaction pathways in near- and supercritical water. J Supercrit Fluid 2002;22(1):37–53 [9] Arita T, Nakahara K, Nagami K, Kajimoto O. Hydrogen generation from ethanol in supercritical water without catalyst. Tetrahedron Lett 2003;44(5):1083-1086. [10] Susanti RF, Nugroho A, Lee J, Kim Y, Kim J. Noncatalytic gasification of isooctane in supercritical water: A Strategy for high-yield hydrogen production. Int J Hydrogen Energ 2011;36:3895-3906 17 [11] Byrd AJ, Gupta RB, Pant KK. Hydrogen production from glycerol by reforming in supercritical water over Ru/Al2O3 catalyst. Fuel 2008;87:2956-2960 [12] May A, Salvadó J, Torras C, Montané C. Catalytic gasification of glycerol in supercritical water. Chem Eng J 2010;160:751–759 [13] Hammerschmidt A, Boukis N, Hauer E, Galla U, Dinjus E, Hitzmann B, Larsen T, Nygaard SD. Catalytic conversion of waste biomass by hydrothermal treatment. Fuel 2011;90:555–562 [14] de Vlieger DJM, Chakinala AG, Lefferts L, Kersten SRA, Seshan K, Brilman DWF. Hydrogen from ethylene glycol by supercritical water reforming using noble and base metal catalysts. Appl Catal B-Environ 2012;111– 112:536– 544 [15] van Bennekom JG, Kirillov VA, Amosov YI, Krieger T, Venderbosch RH, Assink D, Lemmens KPJ, Heeres HJ. Explorative catalyst screening studies on reforming of glycerol in supercritical water. J Supercrit Fluid 2012;70:171– 181 [16] Guo S, Guo L, Cao C, Yin J, Lu Y, Zhang X. Hydrogen production from glycerol by supercritical water gasification in a continuous flow tubular reactor. Int J Hydrogen Energ 2012;37:5559-5568 [17] Gutiérrez Ortiz FJ, Ollero P, Serrera A, Sanz A. Thermodynamic Study of the Supercritical Water Reforming of Glycerol. Int J Hydrogen Energ 2011;36:8994-9013 [18] Gutiérrez Ortiz FJ, Ollero P, Serrera A, Galera S. An energy and exergy analysis of the supercritical water reforming of glycerol for power production. Int J Hydrogen Energ 2012;37:209-226 [19] Gutiérrez Ortiz FJ, Ollero P, Serrera A, Galera S. Optimization of Power and Hydrogen Production from Glycerol by Supercritical Water Reforming. Chem Eng J 2013;218:309-318 18 [20] Suelves I, Lázaro MJ, Moliner BM, Corbella BM, Palacios JM. Hydrogen production by thermo catalytic decomposition of methane on Ni-based catalysts: influence of operating conditions on catalysts deactivation and carbon characteristics. Int J Hydrogen Energy 2005;30:1555-1567 [21] Jun J, Dhayal M, J-Shin H, Han YH, Getoff N. Surface chemistry and catalytic activity of Ni/Al2O3 irradiated with high-energy electron beam. Appl Surf Sci 2008;254:4557- 4564 [22] Buffoni IF, Pompeo F, Santori GF, Nichio NN. Nickel catalysts applied in steam reforming of glycerol for hydrogen production. Catal Commun 2009;10:1656–1660 [23] Sierra Gallego G, Mondragón F, Tatibouët J-M, Barrault J, Batiot-Dupeyrat C. Carbon dioxide reforming of methane over La2NiO4 as catalyst precursor-Characterization of carbon deposition. Catal Today 2008;133–135:200–209 [24] Savva PG, Polychronopoulou K, Ryzkov VA, Efstathiou AM. Low-temperature catalytic decomposition of ethylene into H2 and secondary carbon nanotubes over Ni/CNTs. Appl Catal B-Environ 2010;3-4:314–324 [25] Schubert M. Catalytic hydrothermal gasification of biomass–salt recovery and continuous gasification of glycerol solutions, PhD Thesis, ETH, Zürich, 2010 19 Tables and Figures Captions Table 1. Some relevant results (240 bar and 1.0 L/h). Figure 1. Scheme of the experimental plant. Figure 2. Potential glycerol dehydration mechanism. Figure 3. Effect of the temperature on the gas product yields at (a) 10 wt.% and (b) 20 wt.% (240 bar and 1.0 L/h; WHSV of 5 and 10 gGlyꞏh-1ꞏgCat-1). Figure 4. Effect of the glycerol feed concentration the gas product yields at (a) 600 ºC and (b) 800 ºC (240 bar and 1.0 L/h). Figure 5. XRD of catalyst after being tested. Figure 6. XPS of high-resolution for the zones Ni 2p (a) and C 1s (b), and Raman spectroscopy of carbonaceous residue of catalyst after being tested (c). Figure 7. Microstructure of carbon nanotubes obtained by high-resolution SEM/TEM and EDS spectrogram. 20 Temperature (ºC) Glycerol concentration (wt.%) Product gas composition (vol.%) (dry basis) Hydrogen yield (mol/mol) Gly  (%) GC  (%) LC  (%) SC  (%) H2 CO2 CH4 CO 500 5 55.3 33.4 11.3 0.0 1.14 44.3 30.7 55.8 3.0 10 44.5 41.1 14.4 0.0 0.75 43.5 31.2 56.6 3.8 20 39.6 29.8 21.2 9.4 0.43 32.7 21.9 67.5 4.7 550 5 47.3 34.1 18.7 0.0 2.43 99.6 90.3 0.4 0.3 10 33.0 35.9 31.1 0.0 1.39 99.4 94.1 0.5 0.4 20 23.5 38.2 38.3 0.0 0.87 98.3 94.4 1.8 0.7 30 14.6 39.0 46.4 0.0 0.41 90.1 80.0 9.0 1.5 600 5 53.3 32.4 14.3 0.0 3.27 100 95.7 0.0 0.1 10 41.6 35.4 23.0 0.0 2.02 99.9 94.6 <0.01 0.2 20 26.9 37.2 35.9 0.0 1.08 99.2 97.8 0.8 0.3 30 22.7 37.1 38.9 1.3 0.77 96.8 87.4 3.3 0.5 700 5 70.3 29.0 0.7 0.0 5.70 100 80.3 0.0 0.05 10 58.7 29.7 9.7 1.9 3.85 100 89.1 0.0 0.05 20 42.1 33.2 22.3 2.4 2.14 99.8 98.1 0.1 0.12 30 32.9 34.3 29.3 3.5 1.35 98.2 91.8 1.8 0.3 800 5 71.0 29.0 0.0 0.0 6.52 100 88.7 0.0 < 0.01 10 66.1 26.5 4.0 3.4 5.27 100 90.1 0.0 0.04 20 53.7 28.4 12.7 5.2 3.36 99.9 96.6 0.02 0.07 30 46.6 29.1 17.8 6.5 2.37 99.0 90.5 1.0 0.2 Note: Tests at 30 wt.% and 550 ºC were unsuccessful. Table 1 21 Figure 1 Figure 1 Electrical Furnace TT 107 TT 107 TT 107 High pressure pump Aqueous solution of Glycerol TT 101 TIC 101 SSR TT 102 TIC 102 SSR TT 104 TIC 104 SSR TT 105 TIC 105 SSR SP SP TT 103 TIC 103 SSR SP SP SP TT 106 TIC 106 SSR SP TT 101 TIC 101 SSR TT 102 TIC 102 SSR TT 104 TIC 104 SSR TT 105 TIC 105 SSR SP SP TT 103 TT 103 TIC 103 TIC 103 SSRSSR SP SP SP TT 106 TT 106 TIC 106 TIC 106 SSRSSR SP TT 109 TT 109 TT 109 Cooler TT 108 TT 108 TT 108 PT 110 PT 110 PT 110 PC 110 PC 110 V1 Gas-liquid separator PT 111 PT 111 PT 111 LSH 113 LSH 113 LSH 113 LSL 113 LSL 113 LSL 113 LC 113 LC 113 TT 112 TT 112 TT 112 Liquid V3 PC Gas chromatograph Coalescing filter FT 114 FT 114 FT 114 PC 111 PC 111 V2 SP = 3-10 atm V4 V5 FI 115 FI 115 FI 115 SP= 240 bar To Vent To analysis V0 22 HYDROXYACETONE -H 2 O H – C – C – C – H OH OH OH HHH GLYCEROL -H 2 O 3-HYDROXYPROPANAL H – C – C – C – H OOH HH H H 2 + CO C 2 H 4 + CO ACETALDEHYDE H – C – C – H + O H H ACROLEIN O H – C – C – C – H H H HOHO H – C – C – C – H H H HOH -H 2 O ACETALDEHYDE ACROLEIN H – C – H O FORMALDEHYDE CH 4 + CO Figure 2 23 (a) (b) Figure 3 20 wt.%; 10 g Glyꞏh-1ꞏ g Cat-1; 240 ba r 0 1 2 3 4 450 500 550 600 650 700 750 800 850 Temperature (ºC) Gas yield (mol/mol) H2_Catalyst H2-equilibrium CO2_Catalyst CO2-equilibrium CH4_Catalyst CH4-equilbrium CO_Catalyst CO-equilibrium 10 wt.%; 5 g Glyꞏh-1ꞏ g Cat-1 ; 240 ba r 0 1 2 3 4 5 6 450 500 550 600 650 700 750 800 850 Temperature (ºC) Gas yield (mol/mol) H2_Catalyst H2-equilibrium CO2_Catalyst CO2-equilibrium CH4_Catalyst CH4-equilbrium CO_Catalyst CO-equilibrium 24 (a) (b) Figure 4 800 ºC; 1.0 L/h; 240 ba r 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 5 10 15 20 25 30 35 Glycerol feed concentration (wt.%) Gas yield (mol/mol) H2_Catalyst H2-equilibrium CO2_Catalyst CO2-equilibrium CH4_Catalyst CH4-equilbrium CO_Catalyst CO-equilibrium 600 ºC; 1.0 L/h; 240 ba r 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 5 10 15 20 25 30 35 Glycerol feed concentration (wt.%) Gas yield (mol/mol) H2_Catalyst H2-equilibrium CO2_Catalyst CO2-equilibrium CH4_Catalyst CH4-equilbrium CO_Catalyst CO-equilibrium