Dry reforming of methane over Ni/CeO₂ catalysts prepared by three different methods

2.1 Catalyst preparation

Three Ni-based catalysts were prepared using sol-gel, microemulsion and autocombustion methods; they were labeled Ni-Ce SG, Ni-Ce ME and Ni-Ce AT, respectively. All the reagents were of analytical grade and purity.

The preparation of supported Ni-Ce ME (15%wt of Ni) consists essentially in the preparation of an aqueous solution containing Ni(NO₃)₂6H₂O (99.8%, Merck) and cerium nitrate Ce(NO₃)₃6H₂O (99.5%, Merck) in 50 ml of distilled water. A water-oil microemulsion was prepared by mixing, while stirring, 50 ml (15.5%) of an aqueous solution containing the solution of Ni and Ce, with 187 ml (58%) of cyclohexane (99.8%, SDS), with 39 ml (12%) of butanol (99.5%, Riedel-de Haen) which acts as a co-surfactant, and 36.48 g (14.5%) of cetyltriammonium bromide (CTAB) (99%, Biochem) used as a surfactant. The same microemulsion containing the precipitating agent was prepared using ammonium hydroxide (30%, Panreac). These two microemulsions were stirred separately for 1 h at room temperature, then the second microemulsion was added to the first one dropwise with vigorous stirring (300 rpm) at room temperature for 20 h. The precipitating agent reacts rapidly with the metal precursors through the collision and coalescence between two aggregates containing the different reactants. The reaction (nucleation) will take place essentially at the same time in the whole microemulsion media which will favor the formation of a large number of nuclei inside the water cores of the inverse micelles. These nuclei will grow very quickly through material exchange between micelles (particle growth) to give the final particle size.

At this stage, inorganic particles stabilized by the surfactant molecules, and water-containing micelles, are suspended in the oil phase [34]. The resulting suspension was filtered. The remaining solid was washed with ethanol and acetone. The resulting solid was first dried at room temperature for 12 h then at 383 K for 24 h, and finally it was calcined in air for 8 h at 973 K using a ramp of 5 K/min.

The Ni-Ce sol-gel catalyst (15%wt of Ni) was prepared using the same starting nickel and cerium precursors. A mixture of adequate amounts of Ni(NO₃)₂6H₂O and Ce(NO₃)₃6H₂O was dissolved in 40 ml of distilled water at 298 K. Subsequently, 20 ml of saturated solution of stearic acid C₁₈H₃₆O₂ (98%, Gpr Recctapur) was added to this mixture. The resulting solution was stirred at 353 K for 5 h. The gel prepared was dried at room temperature for 6 h then at 383 K for 24 h and finally calcined during 8 h in air at 973 K using a ramp of 5 K/min.

The Ni-Ce catalyst prepared by autocombustion (AT) (15%wt of Ni) was primarily prepared using the same procedure as the sol gel method; the only difference is in the addition of 10 g of CTAB and 30 ml of the saturated solution of stearic acid to the mixture in order to increase the dispersion of nickel over the support. Afterwards, the mixture was stirred, firstly, at 353 K for 5 h, and then when the temperature reached 553 K it gave rise to the autocombustion of the gel (exothermic reaction); the resulting solid was calcined for 8 h at 973 K.

2.2 Characterization methods

The elemental chemical analysis of catalysts was carried out by atomic absorption.

The BET surface areas of the catalysts were measured with N₂ adsorption at N₂ liquid temperature by using a micromeritics ASAP 2020 instrument. Prior to each measurement, the samples were degassed at 403 K in a vacuum for 1 h.

TPR was conducted with a BROOKS 5878 instrument in the Materials Research Laboratory. About 200 mg of the samples was loaded in a quartz reactor and heated from room temperature to 1023 K at a heating rate of 8 K/min in stream of (5%) H₂ in Ar with total flow of 100 ml/min; the hydrogen consumption is determined as a function of the temperature.

Power XRD patterns were recorded on an Xpert Pro X-ray diffractometer with Cu/Kα radiation (λ=0.01544 nm) operating at 45 kV and 40 mA, ranging from 4 to 90°. The crystallite size was determined from the Scherrer-Warren equation [35].

The characterization of the as-obtained catalysts by TEM was performed with a JEOL 2100F field-emission gun electron microscope operated at 200 kV. The TEM specimens were prepared by dispersing a small sample amount in ethanol and placing one drop of the dispersion on a carbon-film-coated copper grid (3.0 mm, 200 mesh) allowing solvent evaporation.

2.3 Catalytic test

The catalytic activity of the prepared samples was performed in a fixed-bed tubular reactor (with an inner diameter of 9.5 mm) that was heated in an electric furnace equipped with a programmable temperature controller. Fresh 100 mg of catalysts, with a particle size between 150 and 250 μm, were diluted with silicon carbide (SiC) to obtain a 50 mm bed height and packed in the middle of the reactor. The temperature was monitored by a K-type thermocouple placed in the center of the catalyst bed. Before starting the reaction, the catalyst was reduced in situ at 873 K (maximum operation temperature) for 2 h with a mixture of 25 vol. % H₂ in helium at a flow of 100 ml/min. After reduction, helium gas was used during 30 min to sweep the H₂ from the reactor. The feed steam gas mixture consisted of CH₄, CO₂ and He to balance 1:1:8. The total flow rate was 100 ml/min. The reactant gases were measured-controlled by mass flow meters (Brook). The catalytic activity was measured at 873 K. Gas analyses of both reactants and products were carried out by on-line gas chromatography (Varian 3400) equipped with the thermal conductivity detector (TCD). Porapaq Q and Chromosorb 102 columns were used to separate the sample gas (CH₄, H₂, CO and CO₂). Blank experiments were done to verify the absence of catalytic activity under the conditions used in this study, either with the reactor empty or filled with SiC. To discard the presence of diffusion problems, the experiments performed were replicated with other particle sizes, and both the flow rate and the catalyst amount were significantly changed while keeping the mass/flow rate ratio constant. The results obtained suggest the absence of both internal and external mass transfer effects. The carbon balance was close to 100% in all cases. The conversions (X), the yields (Y) and the H₂/CO ratio are calculated as follows:

XCH4(%)=(moles of CH4 converted)∗100moles of CH4 fed

XCO2(%)=(moles of CO2 converted)∗100moles of CO2 fed

YH2(%)=(moles de H2 produced)∗1002 moles of CH4 fed

YCO(%)=(molesdeCOproduced)∗100moles of CH4 fed+moles of CO2 fed

H2CORatio=(molesdeH2 produced)∗100moles CO produced

To investigate the stability of the Ni-Ce catalysts, the deactivation after 12 h of reaction was calculated as follows:

D(%)=(CH40-CH412)/CH40

3.1 Catalyst characterization

In all cases, the chemical composition of samples is in good agreement with the nominal value (Table 1). The BET surface areas of the catalysts are shown in Table 1. The Ni-Ce ME catalyst presents a higher specific surface area (49 m²/g) than that of Ni-Ce SG catalyst (45 m²/g). This result is in good agreement with those reported by Shan et al. [36], who found that the BET surface area of catalysts prepared by microemulsion is much higher than those prepared by the sol-gel method, with a large pore volume and pronounced mesoporosity. However, the Ni-Ce AT catalyst has a low specific surface area (34 m²/g) compared to the two others, which can be related to the high particles size of CeO₂ support compared to that of catalysts prepared by microemulsion and sol-gel (Table 1).

Figure 1 shows the XRD patterns of fresh Ni-Ce catalysts. It can be seen that the as-obtained materials have a good crystallization. As shown from the patterns of the samples, the peaks at scattering angles of 28.54, 33.07, 47.47, 56.33, 59.07, 69.40 and 76.68° correspond to the CeO₂ cubic system (00-034-0394, JCPDS) (Fm-3m, a=5.41 Å) for the three catalysts. Also, peaks attributed to NiO, cubic phase (01-078-0423, JCPDS) (Fm-3m, a=4.1 Å) are observed at 37.23, 43.26, 62.24 and 75.37° in case of Ni-Ce Me; however, NiO monoclinic (C2/m) (03-065-6920, JCPDS) are observed for Ni-Ce AT and Ni-Ce SG at 2θ 37.79, 43.6, 63.25 and 77.22°.

Figure 1:

XRD patterns of studied catalysts prepared by autocombustion, sol-gel and microemulsion methods.

The comparison of particle sizes calculated by applying the Scherrer equation to the main NiO patterns show that Ni-Ce SG and Ni-Ce AT have a smaller particle size of nickel oxide phase (11 nm) than that of the catalyst obtained by microemulsion (34 nm) (Table 1), which also has a higher specific surface area.

The TPR profiles of H₂ consumption for Ni-Ce catalysts are shown in Figure 2. The TPR profiles of the three samples showed a main peak located at 633, 622 and 628 K for Ni-Ce SG, Ni-Ce AT and Ni-Ce ME, respectively. This peak can be assigned to the reduction of bulk NiO phase to Ni⁰ [37, 38]. Morever, for the catalyst prepared by the sol-gel method, small peaks are observed at 517 K and 787 K. The first one can be attributed to the reduction of non-stoichiometric species Ni⁺³ [39, 40], the presence of free NiO particles [41] or hydrogen spillover effect [42], while the second one, much broader, can be assigned to the reduction of NiO interacting with support [43] or CeO₂ reduction [44–46].

Figure 2:

TPR profiles of Ni-Ce catalysts prepared by sol-gel, autocombustion and microemulsion.

On the other hand, the TPR profile of the Ni-Ce catalyst obtained by the microemulsion method shows a small hydrogen consumption around 900 K which can be due to the reduction of bulk CeO₂ [11, 47].

All the samples resulting from the TPR analysis have been subjected to XRD analysis with the aim to show possible phase transition (Figure 3). The obtained patterns show well-defined peaks confirming the good crystallization of catalysts after TPR. Morever, there is complete disappearance of NiO following its reduction under the H₂ flow, giving rise to the formation of Ni⁰ cubic system (Fm-3m, a=3.51 Å) (01-070-0989, JCPDS) (Figure 3) with the same crystallite size (11 nm) (Table 1).

Figure 3:

XRD patterns after TPR of the studied Ni-Ce catalysts.

3.2 Catalytic activity

The as-prepared catalysts were tested in DRM at 873 K after reduction at 823 K for 2 h with hydrogen gas. Figure 4 shows the CH₄ and CO₂ conversion and CO and H₂ yield. The catalysts prepared by autocombustion and sol-gel method result in active catalytic reforming of methane; the rate conversions of CH₄ and CO₂ are about 53%, 53% and 22%, 28%, for samples obtained by autocombustion and sol-gel, respectively.

Figure 4:

Conversion of CH₄ and CO₂, yield of CO and H₂ (CH₄:CO₂:He)=(1:1:8), reduction temperature (Tr=873 K), calcinations temperature (Tc=973), and reaction time (tr=12 h) obtained from Ni-Ce catalysts prepared by autocombustion and sol-gel methods.

However, the catalyst Ni-Ce prepared by the microemulsion method does not show any catalytic activity in this catalytic test.

In fact, the good activity of the Ni-CeO₂ catalyst prepared by sol-gel and autocombustion can be assigned to the small particle size of NiO (11 nm) and/or NiO monoclinic structure (Figure 1), compared to the catalyst prepared using microemulsion method which has, relatively, a large particle size (34 nm) and a NiO cubic structure (Table 1).

Ni-Ce AT catalyst is more active than Ni-Ce SG, the conversion rates of CH₄ and CO₂ in the range of 53%, and yields of H₂ and CO of about 44% and 36%, respectively, with H₂/CO ratio near to 0.82. During the reaction, the conversion rate of methane is similar to that of CO₂; even the yield of CO is higher than that of H₂, suggesting that besides methane dry reforming, reverse water-gas shift reaction (RWGS) CO₂+H₂→CO+H₂O is also occurring, thermodynamically favored at low temperature.

In general, CO₂ reforming of methane is typically accompanied by the simultaneous occurrence of a RWGS reaction [48–50]; the yield of CO is higher than that of H₂ for both catalysts prepared by autocombustion and sol-gel (Table 2 and Figure 3) because a part of the hydrogen produced by DRM is consumed to produce CO from CO₂, confirmed by the H₂/CO ratio being always <1 (Table 2).

The RWGS is very important in the case of Ni-Ce SG confirmed by H₂/CO ratio (0.59) (Table 2), which may be due to the presence of nickel reduced at 787 K as given by TPR analysis (Figure 2). Wang et al. [51] conclude in their study of RWGS over co-precipitated Ni-CeO₂ that the RWGS reaction is favored with catalysts presenting three kinds of nickel as detected by temperature-programmed reduction (TPR) analysis, and they assigned the main active sites for RWGS reaction to oxygen vacancies and highly dispersed nickel.

After 12 h of the reaction, a deactivation of 42% was observed with catalyst prepared by the autocombustion method. This latter can result from the sintering of metallic nickel particles or/and nanotube carbon deposition [10, 11].

For comparative purposes, we have compared the results obtained by the present work with those presented in the literature [17] (Table 2), Roh et al. obtained a high catalytic activity with Ni-CeO₂ and Ni-Ce-ZrO₂ catalysts, with Ni loading close to 15%wt and prepared by the co-precipitation method. The high stability was obtained when they promoted the cerium with ZrO₂ since no deactivation was detected after 100 h of the reaction.

The tested catalysts were characterized by TEM after 12 h of the reaction in order to show their structures and the possible relationship with their deactivation cause. In fact, the deactivation of Ni-Ce SG and Ni-Ce AT can result from the sintering of metallic nickel particles confirmed by TEM micrographs observation (Figure 5); we have also observed an accumulation of carbon nanotubes on both catalysts (Ni-Ce AT and SG) but in a bigger extension on the Ni-Ce SG sample.

Figure 5:

TEM image of Ni-Ce catalysts prepared by sol-gel (A, B, C), autocombustion (D, E, F) and microemulsion (G, H, I) before (A, D, G) and after 12 h of reaction.

In their study, Trovarelli et al. [24] showed that all carbon species formed on Ni-CeO₂ prepared by wetness impregnation method were present in oxidized carbon forms, and they found that the oxidized carbon species are less active towards gasification; this carbon can cover the nickel sites acting as the deactivating factor, thus resulting in fast deactivation.