0CATALY5Í5 FOR 5U5TAiNABL£ £N£RGY
Research Article • DOI: 10.2478/cse-2013-0001 • CSE • 2013 • 60-70
Production of hydrogen from steam reforming of glycerol using nickel catalysts supported on Al2O3, CeO2 and ZrO2
Abstract
Nickel catalysts supported on Al2O3, CeO2 and ZrO2 were prepared by wet Impregnation method and evaluated in steam reforming of glycerol. The catalysts were characterized by chemical composition, textural analysis, crystalline structure and reducibility. The structural characterization of the catalysts revealed a good dispersion of Ni particles using the Al2O3 support, needing higher reduction temperature. The reactions were performed at 500°C with 10 vol.% glycerol solution in a continuous flow reactor. All catalysts showed conversions close to 100%. The selectivity to gas products and formation of liquid by-products were found to be dependent on the type of support. The H2 selectivity showed the following trend: ZrO2 > Al2O3 « CeO2. The catalyst supported on CeO2 showed low activity for water-gas shift reaction, with the highest CO selectivity. All catalysts presented a low formation of CH4. In the liquid phase some by-products were identified (hydroxyacetone, acetic acid, lactic acid, acetaldehyde, acrolein and ethanol) and secondary reaction routes were proposed. Coke formation was higher on Ni/Al2O3 catalyst, but no deactivation was observed during 8 h of reaction.
Robinson L. Manfro, Nielson F.P. Ribeiro, Mariana M.V.M. Souza*
Escola de Química,
Federal University of Rio de Janeiro - UFRJ, Centro de Tecnologia, Bloco E, sala 206, CEP 21941 -909, Rio de Janeiro/RJ, Brazil
Keywords
Hydrogen • Glycerol • Steam reforming • Nickel • Catalysts
Received 27 November 2012 Accepted 04 January 2013
© Versita Sp. z o.o.
1. Introduction
The search for alternative energy sources is increasing in the world motivated by predictions that point to a progressive decrease in the production of fossil fuels. Besides the shortage, another problem associated with the use of fossil fuels is the continuing increase in emissions of pollutants, especially those related to global warming. These greenhouse gases affect human health and also cause imbalances in fauna and flora, such as acid rain. Therefore, there is a great necessity for alternative fuels that do not affect the environment [1].
Motivated by concerns about air pollution, energy security and climate change, the notion of "hydrogen economy" is moved beyond the area of the scientists and engineers, it is a matter of policy issues and treated by business leaders. The interest in hydrogen, the simplest element and most abundant in the universe, is increasing due to technological advances in fuel cells, a potential successor to batteries in portable electronics, power stations and the internal combustion engine [2]. For hydrogen to become a truly sustainable energy source, it should be promoted its production from renewable resources; more than 95 % of hydrogen produced today comes from nonrenewable resources, based on fossil fuels [3,4].
Glycerol is a chemical used for more than two centuries in a variety of applications, such as production of nitroglycerin and esters. Glycerol is also widely used in food, pharmaceutical, cosmetics, toiletries and cleaning industries. The demand and supply of glycerol in the world market was in equilibrium by the end of the 1990s. With the production of biofuels, especially biodiesel, this equilibrium has been completely changed. Biodiesel is produced by the transesterification of vegetable oils and animal fats, and glycerol is a byproduct of this reaction. One ton of biodiesel yields about 110 kg of crude glycerol (glycerin) or 100 kg of pure glycerol [3]. In Brazil, according to National Agency of Petroleum, Natural Gas and Biofuels (ANP), the production of biodiesel (B100) in 2010 was approximately 2.4 million m3, generating 240,000 m3 of glycerin, creating a surplus of glycerin in the Brazilian market. Thus, glycerol is a product that has a large potential market, however, with increasing biodiesel production, their supply has risen substantially and the price of glycerin has fallen to about U$ 0.11 kg -1 [5].
Dumesic et al. [6,7] have produced hydrogen from oxygenated biomass-derived compounds, including glycerol, through the aqueous phase reforming (APR). Although the catalyst has provided stability for a long period, the high pressure and low
* E-mail: mmattos@eq.ufrj.br
reaction rates have prevented its use as a commercially viable process. On the other hand, the steam reforming process can be performed at atmospheric pressure. The steam reforming is the most efficient energy technology available, and with greater profitability [8]. It is highly endothermic and should ideally be performed at high temperatures, low pressures, and high steam/ glycerol ratios to obtain high conversions [9].
The steam reforming of glycerol generates H2 and CO2 as the main gas products and CH4 and CO at low concentrations, but also other alkanes and liquid by-products can be formed [6]. The series of reactions most widely accepted to represent the steam reforming of glycerol is Eq. (1)-(5) [8,10]:
Decomposition of glycerol:
C3H8O3 ^ 4H2 + 3CO
Shift reaction:
CO + H2O o H2 + CO2
Methanation reactions:
CO + 3H2 o CH4 + H2O
CO2 + 4H2 o CH4 + 2H2O
(1) (2)
The overall reaction of steam reforming of glycerol can be written as:
C3H8O3 + 3H2O ^ 3CO2 + 7H2 AH298K= 346.4 kJ/mol (5)
Supported Pt, Ru, Pd and Ni are the most used catalysts in the reforming of oxygenated organics, in both APR and steam reforming, showing good catalytic activity and selectivity for hydrogen production. Unwanted products can be formed due to the occurrence of parallel reactions by breaking of the C-O bonds of the oxygenated organic, forming alcohols or organic acids. Thus, a good catalyst to produce hydrogen by reforming reactions has to be more active for cleavage of the C-C, O-H and C-H bonds rather than the C-O bonds and promote the removal of adsorbed carbon monoxide by the water gas shift reaction [11].
Catalysts based on noble metals have a lower sensitivity to carbon deposition and higher activity. Nevertheless, considering the high cost and limited availability of the noble metals, it is more economical to develop catalysts based on non-noble metals such as nickel, with good performance and high resistance to carbon deposition [12].
Alumina is one of the most widely used supports in catalysis due to its high surface area, which allows a great dispersion of the active phase. In addition, the alumina has relevant characteristics such as high porosity, good mechanical strength and high thermal stability, forming a diffusion barrier that prevents the active phase migration to form clusters of larger particles with lower activity [13].
There is evidence that the support has a significant effect on the overall catalytic behavior and the use of reducible oxides, like ZrO2 and CeO2, can result in additional benefits when compared to irreducible oxides, such as Al2O3. The reducibility and oxygen transfer capacity of ZrO2 and CeO2 have shown to be fundamental in keeping the active phase surface free of carbon deposits [14,15]. Zhuang et al. [15] have reported that CeOx increases the dissociation of water and accelerates the reaction of steam with adsorbed species on the nickel surface near the boundary area between the metal and support, thereby decreasing the carbon deposition and promoting the stability of the catalyst during the reforming of hydrocarbons.
In this contribution, we have investigated the effect of different supports (Al2O3, ZrO2 and CeO2) on the activity and hydrogen selectivity of Ni catalysts in steam reforming of glycerol. For this purpose, structural and morphological properties of the catalysts were studied and secondary reaction routes were proposed based on identification of the by-products formed in liquid phase.
2. Experimental
2.1. Catalyst preparation
The Ni/Al2O3 (NiAl), Ni/ZrO2 (NiZr) and Ni/CeO2 (NiCe) catalysts were prepared by the wet impregnation method, with 20 wt.% of NiO. For preparation of NiAl and NiZr, y-Al2O3 and ZrO2 were calcined at 500°C for 16 h under air flow (60 mL.min-1) to remove water and possible undesirable organic materials, because the bare supports were supplied in the form of pellets (Saint Gobain). For NiCe, CeO2 was prepared from cerium nitrate (Vetec) by calcination at 500°C for 3 h under air flow (60 mL.min-1).
Appropriate amount of nickel nitrate (Vetec) was dissolved in deionized water and mixed to the support in a rotary evaporator. The mixture was heated to 80°C under slight vacuum for evaporation of water, followed by drying at 110°C overnight. Then the catalysts were calcined at 500°C for 3 h using a heating rate of 10°Cmin"1 under air flow (60 mLmin-1).
2.2. Catalyst characterization
The chemical composition of the catalysts was determined by X-ray fluorescence (XRF) using a Rigaku (RIX 3100) spectrometer.
X-ray powder diffraction (XRD) patterns were recorded in a Rigaku Miniflex II X-ray diffractometer equipped with a graphite monochromator using CuKa radiation (30 kV and 15 mA). The measurements were carried out with steps of 0.05° using a counting time of 1 second per step and over the 20 range of 10° and 90o.
The textural properties of the catalysts were determined by N2 adsorption-desorption at -196oC in a Micromeritcs ASAP 2000. The specific area was obtained using the BET method and pore volume and diameter were obtained by BJH method. Prior to the analysis the samples were outgassed for 24 h at 200°C.
The reducibility of the catalysts was analyzed by temperature programmed reduction (TPR), carried out in a microflow reactor operating at atmospheric pressure. The samples were firstly dehydrated at 150°C under flowing Ar before the reduction. A mixture of 1.59% H/Ar flowed at 30 mLmin-1 through the sample, raising the temperature at a heating rate of 10°Cmin-1 up to 1000°C. The outflowing gases were detected by thermal conductivity detector (TCD).
The formation of coke on the catalysts was analyzed by thermogravimetric analysis (TGA) in equipment TA SDT Q600. The used catalysts were heated to 1000°C with a rate of 10°Cmin-1 in air flow (50 mLmin-1).
The morphology of the coke formed on the catalysts was analyzed by scanning electron microscopy (SEM) using Hitachi TM-1000 equipment. The acceleration voltage was 15 kV, using backscattering electron.
2.3. Catalytic tests
The reactions of steam reforming of glycerol were carried out at atmospheric pressure in a fixed bed quartz reactor using 150 mg of catalyst diluted in 75 mg of inert material (SiC) with similar particle size. Catalysts were reduced in situ under 20 % H./N2 flow (75 mLmin-1) with a heating rate of 10°Cmin-1 up to 900, 650 and 700°C, for NiAl, NiZr and NiCe catalysts, respectively.
The aqueous solution of glycerol (10 vol.%) was injected to the reactor by an HPLC pump (Eldex 1SAM) with feed flow of 0.061 mLmin-1 and vaporized in He flow in steel tube heated at 225°C. The composition of the feed gas was glycerol/water/ He= 8/72/20 vol.%. All experiments were conducted at 500°C with gas hourly space velocity (GHSV) of 50,000 h-1. A blank test showed a glycerol conversion of about 20% at these conditions, without formation of gaseous products.
The gas products from the reactor passed through a heat exchanger at 10°C resulting in the separation of gaseous and liquid phases. The gas phase was analyzed online by gas chromatograph Shimadzu GC-2014, equipped with two columns (RT-QPLOT and Carboxen 1010) and thermal conductivity (TCD) and flame ionization (FID) detectors. Liquid phase was analyzed by a Shimadzu HPLC with Bio-Rad Aminex HPX-87H column, using 0.005 M H2SO4 as eluent at 0.6 mLmin-1, and both UV and refractive index detectors.
The catalyst performance is presented in terms of glycerol conversion and H2, CO2, CO, and CH4 selectivity. Performance parameters were calculated based on the following equations:
XGlvOH>%
Glycerol in - Glycerol out
H2 Selectivity,%=
Glycerol in
H2 moles produced C atoms produced in gas phase RR
x — x 100
where RR is the H2/CO2 reforming ratio; it is 7/3 in the case of the glycerol steam reforming.
Selectivity of i,%=
C atoms in species i
-x 100
C atoms produced in gas phase
where species i = CO, CO2, and CH4.
3. Results and discussion 3.1. Catalyst characterization
The chemical composition of the synthesized catalysts is presented on Table 1. As expected, the NiO loading of the catalysts was very close to the nominal value (20 wt.%) and the small differences could be related to experimental errors during synthesis procedure. Table 1 also summarizes the textural characteristics of the catalysts. The BET areas of the supports are 260, 85 and 70 m2g-1 for Al2O3, CeO2 and ZrO2, respectively. After impregnation with NiO all catalysts showed a reduction in the BET area in relation to the bare support. This area reduction can be due to partial obstruction of the pores of the support by NiO particles. Analyzing the average pore diameter of the catalysts, it was observed that all solids are mesoporous (20 - 500 A).
The XRD patterns of catalysts after calcination are shown in Figure 1(a). The catalysts presented two phases, one related to the support and another to nickel species. NiAl sample presented low crystallinity, with broad peaks at 20 = 37.5°, 45.7° and 66.8°, which are characteristic of y-alumina. NiCe presented fluorite CeO2 crystalline phase (JCPS 34-0394) with peaks at 20 = 28.6°, 33.1°, 47.5° and 56.3°, which represent the (111),
Table 1 Chemical composition, textural characteristics (BET area, pore volume and average pore diameter), and crystallite size of NiO and Ni of the nickel catalysts.
Catalyst NiO (wt.%) Sbet (m2g-1) Vpore (cm3g-1) Dp (A) dNiO (nm)a d№ (nm)b
NiAl 22.1 123 0.44 112.7 - 6.3±1.5
NiCe 23.7 62 0.15 84.2 25.0±2.3 44.8±2.5
NiZr 20.2 37 0.18 190.5 22.7±4.6 25.7±1.5
a Calculated by the Scherrer equation using (200) plane of NIO. 0 Calculated by the Scherrer equation using (111) plane of NI.
(200), (220) and (311) planes, respectively. NiZr showed peaks related to monoclinic phase of ZrO2 (JCPDS 37-1484) with peaks at 20 = 28.1°, 31.4°, 34.1°, 35.3°, 44.2° and 50.1° representing the (111), (111), (200), (002), (220) and (022) planes, respectively.
The cubic phase of NiO (JCPDS 47-1049) is observed at 20 = 37.2°, 43.3° and 62.9°, representing the (111), (200) and (220) planes, respectively. The diffraction peaks of the nickel phase are sharp and well defined in NiCe and NiZr samples. The presence of broad NiO peaks in NiAl sample can be related to a good nickel dispersion, resulting in particles too small to be detected by XRD, and/or the formation of NiAl2O4 spinel phase during calcination, also with very small particle size. Spinel phase not detected by XRD was also observed in other Ni/Al2O3 related works [16,17].
Figure 1(b) shows the XRD patterns of the NiAl, NiCe and NiZr catalysts after reduction at 900°C, 700°C and 650°C, respectively. The final reduction temperature corresponds to the end of H2 consumption in TPR profiles, considering the peaks associated with nickel reduction. The XRD patterns showed that NiO was completely reduced to metallic Ni (JCPDS 4-850), with diffraction peaks at 20 = 44.5°, 51.8° and 76.3°, correspondent to (111), (200) and (220) planes, respectively. The phases relative to the supports did not present any significant modification.
Table 1 shows the average crystallite size of NiO and Ni, obtained by the Scherrer equation, using the (200) plane of NiO and (111) plane of Ni. Due to the absence of well-defined diffraction peaks corresponding to NiO phase for NiAl catalyst, it was not possible to calculate the NiO crystallite size for this sample. The average crystallite sizes of NiO for NiCe and NiZr samples are very similar. After the reduction process NiZr sample showed a slight increase in the crystallite size of Ni phase. On the other hand, NiCe sample exhibited a large increase in the size of the Ni crystallite phase. This increase can be related to a weak interaction between NiO and CeO2 phases [18]. The Ni crystallites in NiAl sample presented the smallest average size, approximately 6 nm, which can be associated with the large dispersion of the NiO combined in the spinel phase and stronger interaction between nickel and alumina, reducing the mobility of the ni ckel particles during the reduction.
Figure 2 shows the TPR profiles of the catalysts. The NiAl catalyst presented a reduction peak at higher temperatures compared to the other catalysts, suggesting the existence of a strong interaction between nickel and alumina support. The single reduction peak at 730°C of NiAl catalyst is characteristic of incorporation of NiO into Al2O3 with formation of a spinel phase (NiAl2O4), which adversely affects the reduction of nickel (in general at temperatures above 600°C) [19,20].
Figure 1. XRD patterns of the nickel catalysts after calcination (a) and reduction (b).
Figure 2. TPR profiles of the nickel catalysts.
TPR profile of NiCe catalyst presents four reduction peaks, at 300°C, 420°C, 550°C and 925°C. Since Ni2+ is reduced to Ni0 without going through intermediate oxides, the hydrogen consumption peaks appearing in different temperature regions are assigned to the reduction of different species [21]. The peaks at temperatures below 300oC can be attributed to reduction of free NiO species segregated on the surface of the support [22]. Peaks observed at 420°C and 550°C may be related to the reduction of NiO strongly bounded to the CeO2 support [23] and partial reduction of the surface ceria forming nonstoichiometric cerium oxides (CeOx) [22]. The peak observed at higher temperature (925°C) can be attributed to the partial reduction of bulk CeO2 to Ce2O3. According to Shyu et al. [24] the reduction of bulk CeO2 to Ce3+ occurs at temperatures above 800 °C.
NiZr catalyst presented two reduction peaks: 445°C and 580°C. Generally, low temperature peaks are attributed to the reduction of large NiO particles, while the higher temperature peaks are attributed to the reduction of NiO in intimate contact with the oxide support [25]. Based on this analysis, the reduction peak at low temperature (445°C) corresponds to the reduction of relatively free NiO particles. This temperature is higher than that of NiCe due to the stronger interaction between NiO and ZrO2. The reduction peak at high temperature (580°C) can be attributed to NiO with strong interaction with the support, also observed by Xu and Wang [26], together with a partial reduction of zirconia.
Table 2 shows the experimental consumption of H2 calculated by integrating TPR profiles, theoretical consumption considering the reduction of Ni2+ to Ni° and the reduction degree (ratio of experimental to theoretical consumption). The catalysts supported on ceria and zirconia presented reduction degrees above 100 %, which is associated with a partial reduction of support, forming suboxide species. NiAl catalyst presented a reduction degree of 88 %, due to the difficulty to reduce nickel alumínate. According to Youn et al. [27] the interaction between the metal and support increases with decreasing size of metallic species, resulting in a poor reducibility of the supported metal. This is consistent with our XRD results, which showed a good dispersion of Ni on the alumina support.
3.2. Catalytic tests
Catalytic activity of nickel supported on Al203, Ce02 and Zr02in steam reforming of glycerol in terms of H2 selectivity is displayed in Figure 3. H2 selectivity of NiAl catalyst showed a constant
increase in the first 4 h of reaction reaching 94%. From the fourth hour of reaction there was a reduction in H2 selectivity, which remained relatively constant until the end of the catalytic test, with an average of 68%. The same behavior was observed with the NiCe catalyst, which presented an initial H2 selectivity of 55% that increased to 74% in the third hour of reaction and then remained almost constant until the end of the catalytic test, with an average of 67%. Zhang et al. [22] obtained an H2 selectivity of 87% in steam reforming of glycerol at 500oC using Ni/CeO2 prepared by deposition-precipitation method. H2 selectivity of NiZr catalyst presented an opposite behavior to the other catalysts in the first hours of reaction: an initial high H2 selectivity (87%), which was reduced to 76% in the third hour of reaction and then was kept constant up to the end of reaction, with an average of 78%.
Although the NiAl catalyst has reached the highest H2 selectivity (94%), NiZr catalyst had the best performance for most of the reaction time, with an average H2 selectivity 10% higher at the final part of the test. Comparing the H2 selectivities (Figure 3) and Ni crystallite sizes (Table 1) it seems that Ni dispersion is not clearly related to the catalytic activity for H2 formation, as observed by Iriondo et al. [28] for La2O3 modified Ni/Al2O3 catalysts.
Figure 4 displays CO, CH4 and CO2 selectivities of the catalytic tests performed with the NiAl, NiCe and NiZr catalysts. Among these three compounds it can be clearly observed that CO2 is the
2 3 4 5 6 7 Time (h)
Figure 3. H2 selectivity of NiAl, NiCe and NiZr catalysts. Reaction conditions: 500°C, feed gas composition of glycerol/water/ He= 8/72/20 vol.% and GHSVof 50,000 h-1.
Table 2. Temperature-programmed reduction results.
Catalysts Experimental H2 uptake (mmol H2 gcat-1) Theoretical H2 uptake (mmol H2 gcat"1) Reduction degree (%)
NiAl 2.59 2.95 88
NiCe 4.21a 3.17 135
NiZr 3.01 2.69 112
a Obtained by integration of the reduction profile up to 700 °C.
main reaction product. Table 3 shows the average conversion and selectivities after 3 h of reaction, which is approximately the time needed to reach the steady-state operation. NiAl catalyst showed CO2 selectivity always above 90% and the average CO and CH4 selectivity from 3 to 8 h of reaction was 1.4% and 1.7%, respectively. The reaction with NiCe catalyst showed CO2 selectivity between 75 and 80%, average CH4 selectivity of 3.1% and the highest CO selectivity compared to other catalysts, of about 20.9%. Finally, the reaction with NiZr catalyst showed a high CO2 selectivity, with an average of 91.9%. In this test the average selectivity to CO and CH4 was similar, 4.2% and 3.8%, respectively.
It is well known that Ni has moderate activity in the watergas shift reaction (Eq. 2) [29], resulting in low formation of CO, as observed in the tests with NiAl and NiZr catalysts. It can be noticed that the support has definitely a great influence on the catalytic performance. These tests showed that nickel supported on CeO2 has lower activity for water-gas shift reaction, as a consequence it presented the highest selectivity to CO and the
lowest selectivity to H2 (Table 3). Supported Ni catalysts have high activity in the methanation reactions of CO and CO2 (Eq. 3 - 4) [7], and the formation of methane was significant, mainly for NiCe and NiZr catalysts.
The liquid phase of the reactions was analyzed using HPLC. Figure 5(a) shows the composition of the liquid phase and the conversion of glycerol obtained in test with NiAl catalyst. The conversion of glycerol has an initial increase, reaching almost 100% on the fourth hour of reaction and maintaining constant until the end of the reaction. In the liquid phase it was identified acetic acid, lactic acid, acetaldehyde, acrolein and hydroxyacetone (acetol). The presence of acetic acid was found only in the first 2 h of reaction, with very low amounts (less than 0.05 gL-1). The concentration of most of by-products decreased during the reaction until the fifth hour, reaching concentrations as low as 0.1 gL-1. The only by-products formed during the whole reaction period were hydroxyacetone and acrolein.
The test performed with NiCe catalyst (Figure 5(b)) also showed high glycerol conversion, close to 100% after 3 h of
Figure 4 CO, CH4 and CO2 selectivity of NiAl, NiCe and NiZr catalysts. Same reaction conditions of Figure 3. Table 3. Average glycerol conversion and selectivities after 3 h of reaction.
Catalyst Conversion (%) H2 Selectivity (%) CO Selectivity (%) CO2 Selectivity (%) CH4 Selectivity (%)
NiAl 99.0 74.9 1.4 96.9 1.7
NiCe 99.7 68.7 20.9 76.0 3.1
NiZr 99.9 77.5 4.2 91.9 3.8
reaction. Besides the by-products already identified in the test with NiAl, it was also observed the presence of ethanol at low concentrations, around 0.1 gL-1, and traces of pyruvaldeyde. In this test the main by-product formed was acetaldehyde, followed by hydroxyacetone, lactic acid and acrolein. The concentration of these by-products is initially high and then presents successive reductions, reaching a minimum around the fourth hour of reaction. After this period there is an opposite behavior, with successive increases up to the end of the reaction. The byproduct concentrations in this test were the highest between all tests.
Figure 5(c) presents the results obtained in the test with the NiZr catalyst. The conversion of glycerol was almost constant and close to 100%. The compounds found in the liquid phase were acetic acid, lactic acid, acetaldehyde, acrolein and hydroxyacetone, all of them with very low concentration, below 0.2 gL-1. As in the test of the NiAl catalyst, acetic acid was observed only in the first two hours of reaction.
The formation of lactic acid occurs by the reaction of isomerization of glyceraldehyde (Eq. 7), which is formed from the dehydrogenation of glycerol (Eq. 6) [30-32].
HO OH —>■ O^OH + H2
glycerol glyceraldehyde
O^-^OH ^ OH
glyceraldehyde
lactic acid
Hydroxyacetone and acrolein are formed from glycerol by dehydration reactions. According to Chai et al. [33] and Corma et al. [34] the dehydration reaction of glycerol can be initiated by either the central -OH (Eq. 8) or terminal -OH (Eq. 9), which results in parallel formation of two enol intermediates. These enols undergo a rapid rearrangement to 3-hydroxypropionaldehyde (3-HPA) and hydroxyacetone, respectively. The 3-HPA can be easily dehydrated to form acrolein (Eq. 8).
H^^Y^OH -^^/^OH
glycerol 3-HPA
acrolein
- H2O OH -
hydroxyacetone
Figure 5. Conversion of glycerol and concentration of liquid by-products formed in the catalytic test using (a) NiAl, (b) NiCe and (c) NiZr catalysts. Same reaction conditions of Figure 3.
Pompeo et al. [35] proposed some reaction pathways involved in steam reforming of glycerol. In a possible scheme of reactions, the acetaldehyde is formed from pyruvaldehyde coming from the dehydrogenation of hydroxyacetone; the break of a C-C bond of the pyruvaldehyde leads to CO and acetaldehyde (Eq. 10). In this same work, it was also proposed that hydration of acetaldehyde leads to 1,1-ethanediol, which then suffers a dehydrogenation reaction resulting in acetic acid (Eq. 11).
hydroxyacetone pyruvaldehyde acetaldehyde OH
acetaldehyde 1,1-ethanediol acetic acid
King et al. [36] suggested some reaction routes for ethanol production from aqueous phase reforming of glycerol. In their
reaction scheme, dehydrogenation, decarbonylation, and dehydration occur sequentially forming acetaldehyde, which can be hydrogenated to ethanol, according to Eq. 12.
glycerol
■s- ho^ -
glyceraldehyde
> - h2O H
HO^ z-^O — OH
ethylene glycol acetaldehyde
/""-oh
In our experiments, the ethanol is probable formed from the hydrogenation of acetaldehyde, which is formed from the break of the C-C bond of the pyruvaldehyde, according to Eq 10. This is reinforced by the presence of pyruvaldehyde in the reaction with NiCe catalyst, which was the only catalyst that showed formation of ethanol.
The catalysts after reaction were subjected to thermogravimetric analysis (TGA) to investigate the formation of carbonaceous residues. The quantification of carbonaceous products was carried out from the weight loss shown in Figure 6(a). From the derivative of the TG curves (DTG) (Figure 6(b)) it can be clearly seen that the main weight loss for all catalysts is in the range of 490-630°C, which is associated with the oxidation of coke deposited on the catalyst. The weight loss in this temperature range for NiAl, NiCe and NiZr catalysts was 23%, 16.8% and 13%, respectively. Only NiAl sample presented a small weight loss below 300°C.
According to Sánchez-Sánchez et al. [37] the thermograms of the Ni catalysts used in ethanol steam reforming can be divided in three different temperature regions: region I at temperatures below 300°C represents the loss of water and volatile species; region II between 300-530°C is attributed to filamentous coke associated with Ni particles; and region III at temperatures above 530°C is related to oxidation of coke deposits with different
graphitization degrees. In our case, the coke was mainly oxidized in the third region of temperature, showing that filamentous coke was not predominant, in agreement with SEM images that will be shown later.
The rate of carbon formation was calculated considering the amount of coke formed in 8 h of reaction (Figure 7). The coke formation was the highest on the NiAl catalyst. According to Srisiriwat et al. [38] the coke formation is associated with the mobility of oxygen in the catalyst supports. Supports as Al2O3 have lower oxygen mobility than CeO2 and ZrO2 and therefore present a higher coke formation. It is well known that CeO2 and ZrO2 have good capacity to store and transfer oxygen and as result promote the gasification of coke deposited [39,40]. Adhikari et al. [41] associated the formation of coke with acidity of the support; increasing the acidity of the support the coke formation increases. The presence of acid sites on Al2O3 is much more significant compared to CeO2 and ZrO2 [42]. This is in agreement with our results, the high rate of coke formation on the NiAl catalyst can be related to the lower oxygen mobility and higher acidity of Al2O3 than CeO2 and ZrO2.
Although all catalysts have shown coke formation, no deactivation was observed during the reaction period. Thus, this carbon formed is not of encapsulating type, which causes the blockage of active Ni sites [10]. SEM analysis was performed to investigate the type of coke formed on the used samples. Figure 8 shows the SEM images of the fresh and used NiZr catalyst. It is possible to observe the presence of agglomerated particles of graphitic carbon on the used sample, but neither filamentous nor encapsulating coke was formed during reaction. The coke morphology was similar for all three catalysts.
Figure 6. Thermogravimetric analysis (TGA) of the NiAl, NiCe and NiZr catalysts after reaction. (a) weight loss in percentage (%) and (b) derivative weight loss (%/°C).
Figure 7. Rate of coke formation on Ni catalysts supported on Al2O3, CeO2 and ZrOz
Figure 8. SEM images of the NiZr catalyst (a) before and (b) after the reaction. Arrows point to agglomerated particles of graphitic carbon.
4. Conclusions
Nickel catalysts supported on Al2O3, ZrO2 and CeO2 were prepared and their morphological and structural properties were evaluated. NiAl catalyst showed larger BET surface area and smaller Ni crystallites sizes, due to a stronger interaction between nickel and alumina, forming a spinel type phase that reduces at high temperatures (about 730 oC). NiCe catalyst presented the biggest Ni crystallite size, which may be related to a weak interaction between the support and NiO phase, with a partial sintering during the catalyst reduction. The catalysts supported on ceria and zirconia presented reduction degrees above 100 %, which is associated with a partial reduction of the support, forming suboxide species.
The catalysts were evaluated in steam reforming of glycerol at 500oC. The average H2 selectivity showed the following tendency: NiZr > NiAl « NiCe. NiAl catalyst showed the highest
CO2 selectivity (above 95%) while the CO selectivity was higher for NiCe catalyst (about 20%). The conversion of glycerol was almost total in all reactions, with no apparent deactivation during 8 h of reaction. Several liquid by-products were identified: hydroxyacetone, lactic acid, acetic acid, acetaldehyde, acrolein for all catalysts and ethanol for NiCe catalyst. The concentration of these by-products is relatively low in tests with NiAl and NiZr catalysts; however with NiCe catalyst the concentrations are significant. Based on these liquid by-products, secondary reaction routes were proposed.
The rate of coke formation showed the following order NiZr < NiCe < NiAl. By SEM analysis it was identified the presence of large graphitic coke particles distributed over the catalyst. This coke formation is not associated with catalyst deactivation. Based on its high H2 selectivity and low coke formation, Ni/ZrO2 catalyst was found to be the best catalyst compared to Ni/Al2O3 and Ni/CeO2 at the reaction conditions investigated.
Acknowledgments
The authors thank Núcleo de Catálise/COPPE/UFRJ for providing FAPERJ and CNPq for the financial support granted to carry out the FRX and BET analysis. Robinson L. Manfro acknowledges this work.
References
[1] Shahid E.M., Jamal Y., A review of biodiesel as vehicular fuel, Renewable Sustainable Energy Rev., 2008; 12, 2484-2494.
[2] Dunn S., Hydrogen futures: toward a sustainable energy system, Int. J Hydrogen Energy, 2002; 27, 235-264.
[3] Adhikari S., Fernando S.D., Haryanto A., Hydrogen production from glycerin by steam reforming over nickel catalysts, Renewable Energy, 2008; 33, 1097-1100.
[4] Kirtay E., Recent advances in production of hydrogen from biomass, Energy Conv Manag, 2011; 52, 1778-1789.
[5] Johnson D.T., Taconi K.A., The glycerin glut: options for the value-added Conversion of crude glycerol resulting from biodiesel production, Environ Progress, 2007; 26, 338-348.
[6] Cortright R.D., Davda R.R., Dumesic J.A., Hydrogen from catalytic reforming of biomass-derived hydrocarbons in liquid water, Nature, 2002; 418, 964-967.
[7] Davda R.R., Shabaker J.W., Huber G.W., Cortright R.D., Dumesic J.A., A review of catalytic issues and process conditions for renewable hydrogen and alkanes by aqueous-phase reforming of oxygenated hydrocarbons over supported metal catalysts, Appl. Catal. B Environ, 2005; 56, 171-186.
[8] Adhikari S., Fernando S.D., Haryanto A., Production of hydrogen by steam reforming of glycerin over alumina-supported metal catalysts, Catal. Today, 2007; 129, 355364.
[9] Sehested J., Four challenges for nickel steam-reforming catalysts, Catal. Today, 2006; 111, 103-110.
[10] Chiodo V., Freni S., Galvagno A., Mondello N., Frusteri F., Catalytic features of Rh and Ni supported catalysts in the steam reforming of glycerol to produce hydrogen, Appl. Catal. A Gen., 2010; 381, 1-7.
[11] Huber G.W., Shabaker J.W., Dumesic J.A., Raney Ni-Sn Catalyst for H2 production from biomass-derived hydrocarbons, Science, 2003; 300, 2075-2077.
[12] Jun K.W., Roh H.S., Chary K.V.R., Structure and Catalytic Properties of Ceria-based Nickel Catalysts for CO2 Reforming of methane, Catal. Surv. Asia, 2007; 11, 97-113.
[13] Suzuki T., Morikawa A., Suda A., Sobukawa H., Sugiura M., Kanazawa T., Suzuki J., Takada T., Alumina-ceria-zirconia composite oxide for three-way catalyst, R&D Review of Toyota CRDL, 2002; 37, 28-33.
[14] Souza M.M.V.M., Aranda D.A.G., Schmal M., Reforming of methane with carbon dioxide over Pt/ZrO2/Al2O3 catalysts, .J Catal., 2001; 204, 498-511.
[15] Zhuang Q., Qin Y., Chang L., Promoting effect of cerium oxide in supported nickel catalyst for hydrocarbon steam-reforming, Appl. Catal., 1991; 70, 1-8.
[16] Tang S., Ji L., Lin J., Zeng H.C., Tan K.L., Li K., CO2 reforming of methane to synthesis gas over sol-gel-made Ni/n-Al2O3 catalysts from organometallic precursors, J. Catal., 2000; 194, 424-30.
[17] Therdthianwong S., Siangchin C., Therdthianwong A., Improvement of coke resistance of Ni/Al2O3 catalyst in CH4/ CO2 reforming by ZrO2 addition, Fuel Proc. Tech., 2008; 89, 160-168.
[18] Roh H.-S., Potdar H.S., Jun K.-W., Carbon dioxide reforming of methane over co-precipitated Ni-CeO2, Ni-ZrO2 and Ni-Ce-ZrO2 catalysts, Catal. Today, 2004; 93-95, 39-44.
[19] Kim P., Kim Y., Kim H., Song I.K., Yi J., Synthesis and characterization of mesoporous alumina with nickel incorporated for use in the partial oxidation of methane into synthesis gas, Appl. Catal. A Gen., 2004; 272, 157-166.
[20] Zhu X., Huo P., Zhang Y., Cheng D., Liu C., Structure and reactivity of plasma treated Ni/Al2O3 catalyst for CO2 reforming of methane, Appl. Catal. B Environ., 2008; 81, 132-140.
[21] Molina R., Poncelet G., a-alumina-supported nickel catalysts prepared from nickel acetylacetonate: A TPR study, J. Catal., 1998; 173, 257-267.
[22] Zhang B., Tang X., Li Y., Xu Y., Shen W., Hydrogen production from steam reforming of ethanol and glycerol over ceria-supported metal catalysts, Int. J. Hydrogen Energy, 2007; 32, 2367-2373.
[23] Wang Y., Zhu A., Zhang Y., Au C.T., Yang X., Shi C., Catalytic reduction of NO by CO over NiO/CeO2 catalyst in stoichiometric NO/CO and NO/CO/O2 reaction, Appl. Catal. B Environ., 2008; 81, 141-149.
[24] Shyu J.Z., Weber W.H., Gandhi H.S., Surface characterization of alumina-supported ceria, J. Phys. Chem., 1988; 92, 4964-4970.
[25] Dong W.-S., Roh H.-S., Jun K.-W., Park S.-E., Oh Y.-S., Methane reforming over Ni/Ce-ZrO2 catalysts: effect of nickel content, Appl. Catal. A Gen., 2002; 226, 63-72.
[26] Xu S., Wang X., Highly active and coking resistant Ni/CeO2-ZrO2 catalyst for partial oxidation of methane, Fuel, 2005; 84, 563-567.
[27] Youn M.H., Seo J.G., Kim P., Kim J.J., Lee H.-I., Song I.K., Hydrogen production by auto-thermal reforming of ethanol over Ni/y-Al2O3 catalysts: Effect of second metal addition, Journal of Power Sources, 2006; 162, 1270-1274.
[28] Iriondo A., Barrio V.L., Cambra J.F., Arias P.L., Guemez M.B., Navarro R.M., Sanchez-Sanchez M.C., Fierro J.L.G., Influence of La2O3 modified support and Ni and Pt active
phases on glycerol steam reforming to produce hydrogen, Catal. Comm., 2009; 10, 1275-1278.
[29] Grenoble D.C., Estadt M.M., Ollis D.F., The chemistry and [37] catalysis of the water gas shift reaction : 1. The kinetics over supported metal catalysts, J. Catalysis, 1981; 67, 90-102.
[30] Ramírez-López C.A., Ochoa-Gómez J.R., Fernández- [38] Santos M., Gómez-Jiménez-Aberasturi O., Alonso-Vicario
A., Torrecilla-Soria J., Synthesis of lactic acid by alkaline hydrothermal conversion of glycerol at high glycerol concentration, Ind. Eng. Chem. Res., 2010; 49, 6270-6278. [39]
[31] Marins EP, Davis RJ. Hydrogenolysis of glycerol over carbon-supported Ru and Pt catalysts. J Catal 2007;249: 328-337.
[32] Chheda J.N., Huber G.W., Dumesic J.A., Liquid-phase catalytic processing of biomass-derived oxygenated hydrocarbons to fuels and chemicals, Angew. Chem. Int. [40] 2007; 46, 7164-7183.
[33] Chai S.-H., Wang H.-P., Liang Y., Xu B.-Q., Sustainable production of acroleína: investigation of solid acid-base catalysts for gas-phase dehydration of glycerol, Green Chem., 2007; 9, 1130-1136. [41]
[34] Corma A., Huber G.W., Sauvanaud L., O'Connor P., Biomass to chemicals: Catalytic conversion of glycerol/water mixtures into acrolein, reaction network, J. Catal. 2008; 257, 163-171.
[35] Pompeo F., Santori G., Nichio N.N., Hydrogen and/or syngas [42] from steam reforming of glycerol. Study of platinum catalysts,
In.t J. Hydrogen Energy, 2010; 35, 8912-8920.
[36] King D.L., Zhang L., Xi G., Karim A.M., Heldebrant D.J., Wang X., Petersona T., Wang Y., Aqueous phase reforming
of glycerol for hydrogen production over Pt-Re supported on carbon, Appl. Catal. B Environ., 2010; 99, 206-213. Sánchez-Sánchez M.C., Navarro R.M., Fierro J.L.G., Ethanol steam reforming over Ni/La-Al2O3 catalysts: Influence of lanthanum loading, Catal. Today, 2007; 129, 336-345. Srisiriwat N., Therdthianwong S., Therdthianwong A., Oxidative steam reforming of ethanol over Ni/Al2O3 catalysts promoted by CeO2, ZrO2 and CeO2-ZrO2, Int. J. Hydrogen Energy, 2009; 34, 2224-2234.
Iriondo A., Barrio V.L., Cambra J.F., Arias P.L., Guemez M.B., Sanchez-Sanchez M.C., Navarro R.M., Fierro J.L.G., Glycerol steam reforming over Ni catalysts supported on ceria and ceria-promoted alumina, Int. J. Hydrogen Energy, 2010; 35, 11622-11633.
Iriondo A., Barrio V.L., Cambra J.F., Arias P.L., Guemez M.B., Navarro R.M., Sanchez-Sanchez M.C., Fierro J.L.G., Hydrogen production from glycerol over nickel catalysts supported on Al2O3 modified by Mg, Zr, Ce or La, Top. Catal., 2008; 49, 46-58.
Adhikari S., Fernando S.D., Filip To S.D., Brick R.M., Steele P.H., Haryanto A., Conversion of glycerol to hydrogen via a steam reforming process over nickel catalysts, Energy Fuels, 2008; 22, 1220-1226.
Menezes A.O., Rodrigues M.T., Zimmaro A., Borges L.E.P., Fraga M.A., Production of renewable hydrogen from aqueous-phase reforming of glycerol over Pt catalysts supported on different oxides, Renewable Energy, 2011; 36, 595-599.
