Steam Reforming of Glycerol for Hydrogen Production over Catalyst
The performance of catalyst for glycerol reforming has been investigated in fixed-bed reactor using careful tailoring of the operational conditions. In this paper, a commercial Engelhard catalyst has been sized and compared to gas product distribution versus catalyst size, water-to-carbon ratio, and stability of the catalyst system. catalysts of three sizes (, , and mm) are evaluated using glycerol: water mixture at to produce 2 L H2 g−1 cat h−1. The results indicate that mm size pellet is showing minimum coking and maintaining same level of conversion even after several hours of reforming activity. Whereas studies on and mm pellets indicate that carbon formation is affecting the reforming activity. Under accelerated aging studies, with 1 : 9 molar ratio of glycerol to water, 3 mg carbon g−1 cat h−1 was generated in 20 cycles, whereas 1 : 18 feed produced only 1.5 mg carbon g−1 cat h−1 during the same cycles of operation. The catalysts were characterized before and after evaluation by X-ray diffraction (XRD), BET surface area, scanning electron microscopy (SEM) with energy dispersive X-ray spectroscopy (EDAX), CHNS analysis, transmission electron microscopy (TEM), and X-ray photo electron spectroscopy (XPS).
The search for alternative energy sources is becoming an important aspect in the present scenario due to diminishing petroleum reserves and increased environmental pollution. Hydrogen production from biomass has great interest because of the potential application in fuel cells. Significant amount of glycerol is produced as a by-product in biodiesel production by transesterification of vegetable oils, which are available at low cost in large quantity from renewable raw materials. With increased production of biodiesel, an excess amount of glycerol (C3H8O3) is expected in the world market . At present, glycerol is used in many applications including personal care, food, oral care, tobacco, polymer, and pharmaceutical applications. Besides converting glycerol into value-added chemicals [2–4], hydrogen production through reforming is alternative route [5–12]. Aqueous phase reforming of oxygenated hydrocarbons is extensively studied by the Luo et al. and Shabaker et al. [9, 13], and glycerol steam reforming is studied by Czenik et al. , Pompeo et al. , and Adhikari et al. [6, 15] over nickel-based catalysts and noble metal catalysts on different supports [16–19]. Chiodo et al. reported carbon formation of 2–6 mg carbon g−1 cat h−1 by steam reforming of glycerol by Ni over MgO, CeO2, Al2O3, and Ru/Al2O3 catalysts for 20 h . Adhikari et al. have reported 20–40 mg carbon/100 mg catalyst over Ni supported CeO2, MgO, and TiO2 for 2 h duration . The thermodynamic analysis reported by Adhikari et al. suggests the best conditions for hydrogen production at temperatures higher than 900 K under atmospheric pressure with 1 : 9 molar ratio of glycerol to water [6, 22]. Performance of metal-supported catalysts have been evaluated in terms of activity and hydrogen productivity, but no clear evidence has been reported to establish the most suitable catalyst system for glycerol steam reforming.
The steam reforming reaction of glycerol proceeds according to the following equations.
First, the decomposition of glycerol into syn-gas: followed by the water-gas-shift reaction: (2) The stoichiometric glycerol steam reforming process is represented as:
Steam reforming of glycerol involves complex reactions that results in several intermediates affecting the selectivity of the hydrogen. Furthermore, hydrogen production and carbon deposition strongly depend on different operating conditions, such as glycerol-to-water molar ratio, temperature, contact time, and pressure. The deactivation of catalysts can be minimized by operating at lower glycerol-to-water molar ratio (GWMRs). So far, all the published papers do not report the use of Ni/SiO2 catalyst for glycerol steam reforming except Pompeo et al.  with only 5 wt% Ni on SiO2 support. The results do not clear the evidence to establish the most suitable catalyst system for glycerol steam reforming:
In the present study, carbon formation during glycerol reforming process was investigated. In our experiments, coke formation was found to be the main reason for the deactivation of Ni/SiO2 catalysts. Small size of the catalyst and higher GWMRs promote carbon deposition. The use of small catalyst particles reduces the internal mass-transfer limitation even though they are limited by relatively high pressure drops, mal distribution of flows, and hot spots that alter selectivity of the process. The selections of the right size of the catalyst overcome these drawbacks. We report results of experimental studies on the rates of H2, CO2, CO, and CH4 products from glycerol over silica supported Ni catalyst. Under reaction conditions investigated in this study, Ni/SiO2 catalyst 3 × 5 mm size was found to be the best performing catalyst in terms of hydrogen production from glycerol.
2. Materials and Methods
2.1. Catalysts and Chemicals
Commercial Engelhard Ni-5256 E 3/64′′ catalyst containing highly dispersed Ni (57%) on a silica support with surface area 260 m2/g was obtained. The catalyst samples were ground, and the particles with diameter as denoted in Table 1 were used in the fixed bed reactor at 2 g catalyst scale of operation. Glycerol laboratory grade from Qualigens Fine Chemicals India, Ltd. was used.
The X-ray diffraction (XRD) patterns of the Ni/SiO2 fresh and used catalysts were recorded with Siemens D-5000 X-ray diffractometer using Ni-filtered Cu Kα radiation ( nm) and range between 2–80°. The BET surface areas of fresh and used samples were measured by N2 physical adsorption at −196°C in an Autosorb-I (Quantachrome) instrument. The scanning electron microscopy (SEM) analysis of the catalysts was performed on Hitachi S-520 SEM unit. Elemental analysis was carried out using Link ISIS-300 Oxford energy-dispersive analysis of X-ray spectroscopy detector (EDAX). TEM studies were conducted on TECHNAI 20B2 S-Twin unit operated at 120 kV with a filament current of 28 mA. X-ray photoelectron spectroscopies (XPSs) were recorded on a KRATOS AXIC 165 photoelectron spectroscopy using the Mg Kα radiation. The carbon contents were measured using VARIO EL, CHNS elementary analysis that was carried out on spent catalysts.
2.3. Catalyst Evaluation
Glycerol steam reforming reactions were carried out using a fixed bed, tubular down flow quartz type reactor of 18 mm dia with a thermocouple of 2 mm width. The reactor was provided with a preheater, a syringe pump, a cold condenser, and gas flow meter. The catalyst (2 g) was loaded which is supported by quartz wool at bottom of the catalyst in middle of the reactor. The reactor is placed in a tubular furnace with inner dia of 25 mm. The feed mixture (1 : 9 to 1 : 18 mole ratio of glycerol to water) was fed into the vaporizer using a syringe (B. Braun) pump. The feed entering the pre-heater was maintained at 500°C before reaching the catalyst bed. A Nippon (NC-2538) temperature controller was used for maintaining the temperature of preheating zone and catalyst bed of the reactor. The conversion of glycerol was calculated from the volume of the condensate, and the gas composition was determined by gas chromatography (SHIMADZU GC-2014) with TCD (thermal conductivity detector) using Carboxen 1000 column and Helium as carrier gas. 6 mL of feed was used to get 2 L H2 g−1 cat h−1. Prior to the reaction, catalysts were reduced using 10% H2/N2 in the temperature range of 450–500°C/4 h with a heating rate at 10°C/minute. The evaluation studies were carried out for 20 h to 100 h depending on the activity level.
3. Results and Discussion
3.1. Catalyst Characterization
X-ray diffraction (XRD) patterns of fresh and used catalysts are shown in Figure 1. The diffraction pattern of fresh catalyst (NS) shows the characteristic peaks of Ni at (111), whereas in used catalysts, additional characteristic Ni peaks at (200) and 76.38° (220) are observed along with the peak at (JCPDS no. 04-0850) . The sharp characteristic peaks of Ni in the XRD of used samples assign the formation of crystals with defined planes during the reaction. No interactions were observed between nickel and silica. The average crystallite size calculated from Scherrer equation for fresh catalyst is 5 nm, and in used samples, the size varied from 6 to 20 nm depending on the reaction conditions (Table 1). Peaks around (JCPDS no. 75-1621) attributed to graphitic carbon corresponding to the 002 plane of carbon in the used catalyst. Initially mm size catalyst was used, where the XRD graph clearly indicates high intensity of carbon than Ni peak due to the carbon deposition on the surface. Further mm size catalysts show a decrease in the peak intensity at , simultaneously increasing intensity of Ni peaks. This indicates minimum carbon deposition on the surface of large pellets. The XRD patterns show that these catalysts have minimum coking rate and maintain the activity due to the surface active sites (Ni crystallites).
3.1.2. BET Surface Area
The surface area values of Ni/SiO2 fresh and used catalysts are shown in Table 1. The surface area of the fresh NS catalyst is high (260 m2/g) when compared to used catalysts due to coke formation on the surface .
Morphology of fresh and used catalysts that are investigated by scanning electron microscopy is shown in Figure 2. The fresh catalyst shows homogeneous texture, whereas heterogeneous morphology is observed in used catalysts due to the presence of carbon content. Thus, the SEM images could depict the change in the morphology of fresh and used samples. EDAX analysis of fresh and used catalysts is provided in Table 1. In used catalyst where the carbon content is more, the Ni content on the surface was minimum indicating that the Ni is covered by carbon deposition.
3.1.4. CHNS Analysis
CHNS analysis was carried out to understand the amount of carbon deposition during the course of reaction, and the data is shown in Table 1. Steam reforming of glycerol using different catalyst sizes and feed compositions are found to be effecting the carbon deposition as is evident from the analysis.
The results of TEM analysis are shown in Figures 3 and 4. TEM images of fresh (NS) catalyst show the fine dispersion of Ni particles on SiO2 support . The size of Ni in fresh and used catalysts is about 5–8 nm and 10–20 nm, respectively, which is consistent with the crystallite size calculated from XRD. TEM images used catalysts that depict carbon nanofibers formed during glycerol reforming reaction over NS1, NS3, NS4, and NS5, with a diameter ranging from 30 to 60 nm along with Ni particles embedded in them.
Furthermore, the overall morphological and structural information for used samples is also obtained by SAED analysis. Figures 5(a) and 5(b) show SAED analysis of accelerated ageing with different GWMRs. The results show small difference in their morphologies. The SAED patterns of both 1 : 9 and 1 : 18 GWMRs feed for corresponding used catalysts (NS4 and NS5) show spotty ring patterns (Figures 5(a) and 5(b)) which are indexed as cubic Ni [26, 27]. The same results are further confirmed by the XRD results (JCPDS no. 04-0850). Apart from the above, there is a slight difference between the two SAED patterns (Figures 5(a) and 5(b)) which should not be ignored. The (111), (200), (220) and (311) reflections of Ni planes in Figure 5(b) appear to be more intense than those in Figure 5(a), signifying the better crystallinity of the catalyst in accelerating ageing study with 1 : 18 (GWMRs). The reflection of graphite plane (002) is also observed in Figures 5(a) and 5(b) having different intensities as studied at different GWMRs conditions.
XPS spectra of Ni in fresh catalyst (Figure 6(a)) show a characteristic peak at 853.7 eV attributed to and a satellite peak near 859.7 eV, whereas shows peak at 870.6 eV with a satellite peak around 876.9 eV respectively . Besides, Ni 2p spectra show a decrease in binding energy values in the used catalyst. The Si 2p spectra (Figure 6(b)) show a characteristic peak with high intensity at 103.4 eV assigned to SiO2 (fresh), while peak at 102.6 eV in used catalyst corresponds to silicate . The O1s binding energy observed at 532.6 eV ascribed to prominent SiO2 linkages. The decrease of 532.6 eV peak in the intensity in the used sample spectra may be due to carbon deposition (Figure 6(c)). A peak with the highest intensity is observed in C1s spectra at 284.6 eV corresponding to the coke deposition on the surface (Figure 6(d)). The results are in well agreement with the XRD analysis. Table 2 shows the binding energy values of Ni 2p, Si 2p, O1s, and C1s in NS and NS1 catalysts.
3.2. Catalyst Performance Testing
3.2.1. Effect of Pellet Size
Glycerol steam reforming activity studies are conducted over Ni/SiO2 catalyst of different sizes at 600°C as shown in Figure 7. The initial reactant conversion on , , and (mm) pellets is 100%. However, with time on stream, stable activity is observed on 3 × 5 mm size pellet. The activity is seen decreasing with time, on the other two sizes, which are smaller in size, and maximum deactivation is observed on the smallest pellet ( mm). This could possibly be seen as due to the constrain in flow of the reactants due to the increased pressure on the surface that could be resolved with increasing size of the pellet. This may also be seen as due to the heavy coking on the greater surface exposed to reaction in the smaller size . The catalyst particle size is critical since it alters the surface-to-volume (S/V) ratio of the catalyst. If carbon blocking of active sites is responsible for catalyst deactivation on decreasing the S/V ratio, then an enhancement of the deactivation process should be expected. The increase in the S/V ratio increases the rate of carbon gasification enhancing the catalyst stability. Thus, further studies are made on mm pellet size for obtaining better operating conditions.
3.2.2. Effect of Temperature
The effect of temperature on glycerol steam reforming and product gas distribution over Ni/SiO2 mm catalyst with 1 : 9 mole ratios of glycerol to water is shown in Figure 8 and Table 3. The increase in temperatures from 500 to 600°C, increases the glycerol conversion (80–100%). Around 500°C, the product gas distribution shows more of CH4 and CO2 conforming methanation and water gas shift reaction. At 600°C increase in CO formation and decrease in CO2 show that the reverse water gas shift reaction is predominant in (5) and (2) [31–34]. A nominal increment is observed in at 600°C temperature:
3.2.3. Effect of GWMRs (Glycerol-to-Water Mole Ratio)
Figure 9 and Table 3 explain the gas distribution at 600°C temperature with different GWMRs. Decrease in GWMRs (1 : 9 to 1 : 18 mole ratio) shows an increase in H2 and CO2, and simultaneously and a decrease of CO and CH4 that may be seen due to more favored methane steam reforming and reverse water gas shift reaction under the conditions of study  as follows:
3.2.4. Accelerated Ageing Studies
Catalyst life is evaluated by accelerated ageing studies, and the importance of this study is to reduce time of evaluation and manpower wastage and to evaluate catalyst life time in a short period. Accelerating ageing studies over the Ni/SiO2 catalyst ( mm) are conducted by changing the operational temperatures in a cycling manner between 600 and 700°C temperature, and the feed rate is doubled. The effect of GWMRs on catalytic conversion and coke deposition is studied under similar conditions. Figure 10(a) shows the glycerol conversion as a function of accelerated ageing for 20 experimental cycles (each cycle is 600°C/1 h and 700°C/1 h; feed rate is doubled) at 1 : 9 GWMRs. After 20 cycles of reaction operation, 20% fall is observed for glycerol conversion, and 3 mg carbon g−1 cat h−1 has deposited on the catalyst. In order to minimize the coke deposition, accelerating ageing studies are conducted with 1 : 18 GWMRs in similar conditions and the results are shown in Figure 10(b). Only 10% fall is observed in glycerol conversion after 20 cycles, and the coke formation is minimized to 1.5 mg carbon g−1 cat h−1 (Table 1) .
3.3. Coke Formation
Carbon deposition occurs mainly by the decomposition of CO, CH4, and thermal cracking of hydrocarbons depending on the conditions like temperature and partial pressures of reactants and products. Different morphologies of coke may form, namely, pyrolytic carbon, encapsulating carbon, whisker, and carbides. However, the nature and morphology depend on the active metal, metal particle size, and hydrocarbon source. Besides, at high temperatures polyaromatic or even graphitic compounds are formed. Carbon depositions on the surface of the catalyst will result in several undesirable reactions, and products affecting the purity of reformation and carbon also cause loss of effective surface area. As discussed earlier, the coke formation is not desirable in the steam reforming process. Compared to higher pellet size ( mm) of the catalyst lower pellet size catalysts (, mm) exhibit a greater tendency to form carbon at 600°C with 1 : 9 GWMRs. Furthermore, higher particle size ( mm) is investigated subjecting to accelerating aging studies using different GWMRs to minimize coke formation. Figure 11 clearly explains carbon formation rate in steam reforming of glycerol; 10–12 mg carbon g−1 cat h−1 is observed on Ni/SiO2 (; mm) catalysts. With increased pellet size mm, the carbon formation observed is only 2 mg carbon g−1 cat h−1. In accelerated aging studies with 1 : 9 GWMRs, carbon formation rate is 3 mg carbon g−1 cat h−1, and with further decreasing GWMRs to 1 : 18, the formation rate has decreased to 1.5 mg carbon g−1 cat h−1. These observations evidence that minimization of the coking formation rate is by increasing the catalyst pellet size and decreasing GWMRs. The coke formation is estimated by CHNS analysis that is provided in Table 1.
The present study illustrates the effects of catalyst pellet size in H2 production through glycerol steam reforming, where the small pellets promote the coke formation at high temperature (600°C). The higher pellet size improves the catalytic performance as well as minimizes coke formation. A minimum amount of 1.5 mg carbon g−1 cat h−1 coke was observed under accelerating ageing studies for (GWMRs 1 : 18). The detailed characterization of the fresh and used catalyst allows a better understanding of catalysts behavior under glycerol steam reforming conditions. The results illustrate that carbon formation is mainly responsible for the deactivation of the catalyst, and the sustainability can be achieved by optimizing favorable reaction conditions.
G. Sadanandam thank CSIR, New Delhi, for SRF Grant and also the authors thank MNRE, New Delhi, for funding this project.
BIODIESEL 2020, A Global Market Survey, Feedstock Trends and Forecasts, Emerging Markets Online, 2nd edition, 2008.
S. Czernik, R. French, C. Feik, and E. Chornet, “Hydrogen by catalytic steam reforming of liquid byproducts from biomass thermoconversion processes,” Industrial and Engineering Chemistry Research, vol. 41, no. 17, pp. 4209–4215, 2002.View at: Google Scholar
C. D. Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder, and G. E. Muilenberg, Eds., Handbook of X-Ray Photoelectron Spectroscopy, PerkineElmer Corporation, 1979.
V. Durga Kumari, K. Venkateswarlu, M. Subrahmanyam et al., “H2 production activities at Indian Institute of Chemical Technology,” in International Conference on Hydrogen Production, The Korean H2 and New Energy Society, Seoul, Korea, June 2012.View at: Google Scholar