Recently, the CO2 reforming of CH4 (CRM) has been widely investigated as one of the most promising options (from industrial, environmental, and research standpoints) for the use of greenhouse gases CO2 and CH4, as it generates valuable synthesized gas [1-3]. Ni-based catalysts are more interesting than noble metal catalysts because Ni metal has a higher availability and lower cost [4]. However, previous studies have shown that Ni-based catalysts are more prone to carbon deposition and deactivation [4, 5].
Numerous studies have been devoted to the understanding of the CRM elementary steps over Ni-based catalyst as well as their promotion to enhance the anti-carbon property [6-10]. So far, a general agreement in the literature is that the title reaction involves CH4 decomposition to CHx species on the Ni metal surfaces (reaction (1)), CO2 decomposition to absorbed oxygen-containing species (reactions (2) and (3)) and their following oxidation reactions (reactions (4) and (5)). When the above process is out of balance, the excessive carbon containing species on the Ni surface can undergo further dehydrogenation, polymerization, and rearrangement into highly stable carbon species, which not only show low reactivity toward the gasification but also lead to block Ni metal sites and destruction of the catalyst structure. In the early studies, it was believed that CH4 deprotonation was the rate-determining step at the reaction temperature (550-750 ℃) for the CRM [10]. However, it has been reported that CH4 dissociation is considered reversible since H2 and absorbed H species on the Ni metal surface can reach rapid equilibrium during the CRM [11]. Thus, additional experimental evidence and theoretical modeling results have proposed that the surface reaction between carbon and containing oxygen species is the rate-determining step at the higher reaction temperature (650-750 ℃) [7, 12, 13]. Besides, several studies have proposed that the rate of dissociation of CO2 on Ni crystallites is not significant compared with that of CH4 [7, 14]. Therefore, supplying more oxygen containing species to participate in the gasification process of carbon deposition has been considered as the predominant solution to the problem of having highly stable carbon species.
The ready deactivation situation leads to the need for developing Ni catalysts with improved stability to prevent or reduce the formation of carbon through research into the following aspects: (ⅰ) the nature of the support [15-18], (ⅱ) the catalyst preparation method [19-21], and (ⅲ) the addition of basic promoters [22, 23]. In relation to the latter variable, the role of basic metal oxide promoters is generally considered to enhance the basicity of the Ni-based catalysts, which could promote CO2 adsorption [24]. Interestingly, among the published literature, there exist some contradictory results about the effect of the basic metal oxide promoters with small amounts on the inhibition of carbon deposition in the CRM. Castro Luna and Iriarte [25] reported that the incorporation of potassium with 0.5 wt.% could hinder the accumulation of carbon on the Ni-Al catalyst surface by decreasing CH4, probably owing to the neutralization of a fraction of the most active sites for the reforming reaction. Hou et al. [26] observed that 0.32 wt% of Ca hindered the accumulation of coke over the Ni/Al catalyst by promoting the dispersion of Ni. Besides, it has been reported that catalysts with a low concentration of MgO promoter (0.4 wt%) had a higher activity and stability, although such small quantities of MgO had no influence on the basic strength of the catalyst [22].
To study the effects of a promoter on the catalytic performance, the spatial distribution of the promoter is equally important to represent its effect. Introduction of the promoter was achieved through common impregnation of the as-synthesized solid catalysts. This may not, however, ensure a homogeneous distribution of the promoter on the Ni-based catalyst surface. Until now, it has been reported that hydrotalcite-type anionic clays (also called layered double hydroxides), having all the cations homogeneously distributed inside the brucite type sheets of the layered structure, could give rise to calcination and to highly homogeneous dispersion of the elements into the oxide matrix [27]. These mixed oxides prepared from hydrotalcite-type anionic clays have an interesting property, a "memory effect", which allows the reconstitution of the original hydrotalcite structure under mild conditions when the product of the thermal treatment is brought into contact with aqueous solutions containing various anions [28]. Takehira et al. [29] reported that they prepared particles of Mg-Al mixed oxide by calcining Mg-Al hydrotalcite and dipping it in an aqueous solution of Ni2+ nitrate. They found that the reconstitution of Mg(Ni)-Al hydrotalcite through the "memory effect" took place on the surface layer of Mg-Al mixed oxide, resulting in a uniform distribution of Ni on the catalyst surface after calcination and reduction. Following this technique, to obtain a highly homogeneous distribution of the MgO promoter, we selected a Ni-Al mixed oxide which was prepared using calcinations of Ni-Al hydrotalcite. We then introduced the MgO promoter through the "memory effect". For comparison, a non-promoted Ni/Al catalyst was also prepared. This paper analyses the effects of a small quantity of MgO promoter on the catalytic performance and catalyst structure of a Ni/Al2O3 catalyst in the CRM reaction. We used a series of surface reaction experiments to examine the effect of a small quantity of MgO promoter on eliminating carbon deposition in the CRM.
The Ni-Al hydrotalcite precursor (molar ratio of Ni:Al was 2:1) was prepared using the co-precipitation method that has been reported previously [30]. The Ni-Al hydrotalcite precursor was divided into two parts, both of which were calcined at 700 ℃ for 1 h, to form Ni-Al mixed oxides. Reconstitution was achieved using the procedure reported previously by Takehira et al. [29]. The mixed oxide powder was pressed at 22 MPa/cm2, crushed, and sieved to 0.36-0.60 mm, before being used in the reconstitution procedure. The particles were immersed in a Mg(NO3)2 aqueous solution (1 mol/L) at ambient temperature for 45 min, washed with de-ionized water, dried in air at 80 ℃, and calcined at 800 ℃ for 5 h to form the MgO-promoted Ni-Al catalysts (referred to herein as Ni-Mg/Al). The Ni-Al metal oxide, selected as a reference sample, was not treated using the reconstitution procedure, but was first calcined 700 ℃ for 1 h and then was calcined at 800 ℃ for 5 h (referred to herein as Ni-Al).
X-ray diffraction (XRD) patterns of the samples were recorded on a Rigaku (D/Max-3B) X-ray powder diffractometer with monochromated Cu Ka radiation at 40 kV/100 mA in the 2θ range of 10°-80°, with a scan speed of 4 °/min. The amount of Mg2+ loading on the catalyst was determined by inductively coupled plasma mass spectrometry (ICP-MS). The reduction behavior of the fresh catalysts was checked using the H2 temperature programmed reduction (H2-TPR) in a Micromeritics Autochem Ⅱ 2920 instrument. The tests were performed from 100to 900℃ at a heating rate of 10℃/min under a flow of 10%H2/90%Ar gas (50 mL/min). Analysis of the carbon that had been deposited after the 10 h reaction was carried out by temperature programmed oxidation (TPO) for the used catalysts in a Micromeritics Autochem Ⅱ 2920 apparatus. The temperature was increased at 10℃/min from 50 to 900 ℃ under an oxidizing gas flow of 3%O2/97%He gas (50 mL/min). The spent catalyst was identified by Raman spectroscopy (LabRam-HR800) using a 514 nm He-Cd laser at a laser power of 30 mW.
The CO2 temperature programmed desorption (CO2-TPD) experiment was carried out in a Micromeritics Autochem Ⅱ 2920 apparatus and measured using a quadrupole mass spectrometer (OmniStar GSD 301). The samples were reduced in situ in a flow of 50 mL/min 10%H2/90%Ar gas for 0.5 h; then purged with helium gas for 0.5 h to remove any adsorbed H2. Later, it was cooled down and exposed to a flow of 50 mL/min 10%CO2/90%He gas at 50 ℃ for 0.5 h. After being purged with a helium gas for 30 min, the sample was heated again from 50 to 900 ℃ at a heating rate of 10 ℃/min in a helium gas flow of 50 mL/min.
The CO2 activation was investigated using pulse-injected surface reactions in a Micromeritics Autochem Ⅱ 2920 apparatus. The samples were reduced in situ at 800 ℃ for 0.5 h in a flow of 50 mL/min 10%H2/90%Ar gas and exposed to helium gas to remove any adsorbed H2 for 0.5 h, before being cooled to 750 ℃. 10%CO2/90%He gas was then injected into the reduced samples with a 6-plot gas sampling value under a stream of purified helium gas (50 mL/min). The effluents were detected by a quadrupole mass spectrometer (OmniStar GSD 301).
The pulse experiments were carried out in a Micromeritics Autochem Ⅱ 2920 apparatus. Typically, samples were reduced in situ in a gas flow of 50 mL/min 10%H2/90%Ar at 800 ℃ for 0.5 h, then purged with helium gas to remove the adsorbed H2 for 0.5 h, before being cooled to 750 ℃ in helium gas. Then, 10%CH4/90%He and 10%CO2/90%He were pulsed in turn under a stream of He carrier gas (50 mL/min, 5 mL pulse, 5 min interval). The effluents were detected by a quadrupole mass spectrometer (OmniStar GSD 301).
The CO2 reforming of CH4 was conducted using a fixed-bed, down flow, tubular quartz reactor (i.d. 8 mm and a length of 560 mm) in a reaction mixture consisting of a stoichiometric (1:1) CH4(99.99%) and CO2(99.99%) under atmospheric pressure at 750 ℃ and at a gas hourly space velocity of 36000 mL/(g·h). Before testing, the samples were reduced in situ at 800 ℃ for 0.5 h, using a flow of 10%H2/90%Ar gas (50 mL/min). The product gas was cooled in ice water and then analyzed by an online gas chromatograph (HaiXin-950 China) equipped with the TDX-01 packed columns and a thermal conductivity detector. The conversions of CH4 and CO2 were defined as below:
CH4 conversion = (FCH4, in -FCH4, out)/FCH4, in
CO2 conversion = (FCO2, in -FCO2, out)/FCO2, in
Typical XRD patterns of the selected catalysts during preparation are shown in Fig. 1. The Ni/Al hydrotalcite precursor (molar ratio Ni:Al = 2:1) was successfully synthesized using the co-precipitation method (Fig. 1(1)), and converted to Ni/Al mixed oxide during the first calcination at 700 ℃ for 1 h. This mixed oxide was composed of a NiO phase and an amorphous phase of Al2O3 (Fig. 1(2)). After the first calcination, the mixed oxide was divided into two parts and one part was immersed in Mg(NO3)2 aqueous solution for 45 min, and the reconstitution of the Ni-Al hydrotalcite structure was clearly observed (Fig. 1(3)). The other part was immersed in de-ionized water for 45 min. Then, both parts were subjected to a final calcination at 800 ℃ for 5 h. XRD measurements of both samples after the final calcination detected similar diffraction lines, which could be attributed to NiO and NiAl2O4 for the Ni/Al catalyst (Fig. 1(4)) and the Ni-Mg/Al catalyst (Fig. 1(5)), respectively. This was in agreement with other reported results for this calcination temperature [31, 32]. For the Ni-Mg/Al catalyst, no relevant peaks in the MgO phase were detected by XRD. Because the diffraction peaks of MgO and NiO are so similar to each other, they almost overlap and are difficult to be distinguished. ICP-MS was employed to determine the amount of MgO loaded on the Ni-Mg/Al catalyst. It was found that a Ni-Mg/Al catalyst containing approximately 0.42 wt.% of Mg was obtained. Elemental mapping of the Ni-Mg/Al catalyst showed that all elements were homogeneously distributed over the sample (Fig. 2). It was well established that the addition of the MgO promoter by the "memory effect" had a highly homogeneous distribution on the alumina support. The nitrogen adsorption and desorption isotherms of both the catalysts obtained from N2 physical adsorption are displayed in Fig. 3. The textural properties of both the catalysts suggested that the surface area of the Ni-Mg/Al catalyst displayed a small degradation in comparison with that of the Ni/Al catalyst (102 m2/g for Ni/Al catalyst and 79 m2/g for Ni-Mg/Al catalyst), in agreement with a previous report. Fig. 4 presents the H2-TPR profiles of the Ni/Al catalyst and the Ni-Mg/Al catalyst. Both samples showed the same H2 consumption profile, with two peaks at 427 and 705 ℃, which indicated a similar interaction between nickel species and the alumina support. X-ray diffraction patterns of each catalyst after reduction are shown in Fig. 5. For both catalysts, the phase structure was very similar; typical peaks of the Ni metal phase (JCPDS 04-0850) and weak diffraction lines of the NiAl2O4 spinel phase (JCPDS 10-0339) could be observed. Noticeably, no MgO diffraction peaks were observed for the modified catalyst owing to the highly homogeneous distribution of the MgO promoter on the surface of the Ni-Mg/Al catalyst. This result was consistent with the analysis of the elemental mapping of the Ni-Mg/Al catalyst. The average Ni crystallite sizes were determined for the reduced samples by application of the Scherrer equation to the Ni (200) peak and similar average crystallite sizes were obtained for both catalysts (29.6 nm for Ni/Al catalyst and 29.8 nm for Ni-Mg/Al catalyst).
The activity of both samples for the CRM reaction at 750 ℃ and 36000 mL/(g·h) (CH4/CO2 = 1) was interpreted by analyzing the CH4 and CO2 conversions, while taking into account the time on stream (displayed in Fig. 6). In both cases, CH4 conversions were lower than the corresponding CO2 conversions, probably owing to the influence of the reverse water-gas shift reaction. As shown in Fig. 6, the initial conversions of CH4 and CO2 on both catalysts were similar. However, over the 10 h reaction time for the Ni-Al catalyst, the conversions of CH4 and CO2 decreased quickly from 78.8% to 48.6%, and from 82.6% to 54.7%, respectively. In contrast, no deactivation was observed during the 10 h reaction period on the Ni-Mg/Al catalyst. By comparing the catalytic stability of the Ni/Al catalyst and the Ni-Mg/Al catalyst, the importance of the MgO promoter in improving the stability was evident. All of these results were reproducible within ±2%.
The nature and amount of carbon that was deposited onto the catalysts were identified by applying the TPO tests. The results are shown in Fig. 7. The amount of carbon deposited on both samples, as calculated from the total amount of CO2 produced, was 12.3 mg/g for the Ni/Al catalyst and 9.1 mg/g for the Ni-Mg/Al catalyst. The CO2 profile on the used Ni-Mg/Al catalyst exhibited only a small broad peak at 450-645 ℃, corresponding to the filamentous carbon [33, 34]. Although the combustion temperature of deposited carbon is also 450 ℃, the CO2 profile of the used Ni/Al catalyst showed three peaks. The first broader peak was at 450-672 ℃, the second of high intensity was at 674 ℃, and the third one of weak intensity at a higher temperature of 703 ℃. Oxidation of deposited carbon took place at temperatures > 650 ℃ for the used Ni/Al catalyst, which is commonly associated with oxidation of graphitic carbon [33, 35]. The structure of the carbon residuals over both used samples was further investigated by Raman spectroscopy (results are shown in Fig. 8). Two peaks could be distinguished, which could be ascribed to the G band (1580-1600 cm-1) and D band (1330-1350 cm-1). The former reflection was related to the in-plane carbon-carbon stretching vibrations of a graphite sp2 structure, whereas the latter was contributed to the structural imperfection of graphite or amorphous carbon [36]. It was reported that the area ratios of the D and G bands (ID:IG) could be used as an index to characterize the statistical graphitization degree of the carbon residuals over the used catalysts, namely the higher the ID:IG ratio, the lower the graphitization [37]. By fitting the curves, the ID:IG ratios increased from 1.05 on the Ni/Al catalyst to 1.98 on the Ni-Mg/Al catalyst. According to the TPO results, the filamentous carbon, which was a major carbon species over the Ni-Mg/Al catalyst, had a lower graphitization degree than that over the Ni/Al catalyst. These Raman results were consistent with the report by Wang et al. [38], where the ID:IG ratios for filamentous carbon and graphite carbon were calculated to be 2.13:1 and 0.93:1, respectively. It has been proven that filamentous carbon accumulation has little effect on deactivation, since the active metallic particles are still in contact with the gas phase and can further perform an active role in the catalytic reaction [34]. Note also that this result is in agreement with the catalytic performance of the Ni-Mg/Al catalyst (as shown in Fig. 6). Based on the current level of knowledge, graphite carbon is the most inactive species and is responsible for catalyst deactivation [35, 39]. Using DFT calculations regarding the carbon deposition on Ni-based reforming catalysts, Xu et al. [40] reported that large graphitic carbon islands are the most stable form of carbon on reforming catalysts, and that their formation requires a high carbon coverage. As a result, if the carbon deposits form very stable graphite carbon, the Ni catalyst surfaces could be significantly covered and more severe deactivation could be caused (since the active metal is no longer accessible to reactants) [41, 42]. Thus, the observed deactivation of the Ni/Al catalyst during the CRM reaction (as shown in Fig. 6) was predominantly caused by the formation of graphite carbon. When comparing the catalytic performance of the Ni/Al and Ni-Mg/Al catalysts and the results of TPO tests and Raman spectroscopy, it can be concluded that the addition of a small quantity of MgO promoter would enable a relatively stable catalytic performance and prevent inactive carbon deposition during the CRM reaction.
CO2-TPD experiments were conducted to investigate the CO2 adsorption ability on the catalysts. The results are presented in Fig. 9. The reduced Ni-Mg/Al catalyst exhibited a CO2 desorption curve similar to that of the reduced Ni/Al catalyst, namely, only one asymmetric peak with the maximum rate of CO2 desorption occurring at 98-124 ℃, which could be attributed to the weak basic sites corresponding to surface hydroxyl groups [43, 44]. No CO signals were detected in either sample. The CO2 desorption peaks of both catalysts had a long tail extending up to 400 ℃, indicating that MgCO3 was not present on the Ni-Mg/Al catalyst. It has been reported that the release of CO2 from MgCO3 decomposition is only possible at and above 540 ℃ [45]. Note that the intensity of CO2 desorption over the Ni-Mg/Al catalyst was lower than that over the Ni/Al catalyst owing to a decrease in surface area, while the peak desorption temperature did not change with the addition of MgO. This observation indicates that the addition of the MgO promoter did not increase the basic strength for the Ni-Mg/Al catalyst towing to its low concentration.
The dissociative adsorption of CO2 in the absence of H2 and CH4 was investigated using the pulse technique. The results are shown in Fig. 10. In Fig. 11, the amount of CO generated as a function of the CO2 pulse number on each catalyst is shown. When CO2 was pulsed onto the catalysts by the helium carrier gas, the only signal detected was CO. It has been reported that CO2 could dissociate on the metallic Ni metal surface to produce CO in the air, and adsorb oxygen on the surface. For both catalysts, the amount of CO formed gradually decreased as the number of pulses was increased, indicating that the active Ni metal sites for CO2→CO+Oads were gradually filled with adsorbed atomic oxygen, Oads. It was observed that four pulses of CO2 were not sufficient to cover completely the active sites with atomic oxygen. Moreover, it can be observed that the amount of CO produced from sequential injections of CO2 over the reduced Ni-Mg/Al catalyst was more than that over the reduced Ni/Al catalyst (as shown in Fig. 11).
Although carbon deposition can be derived from two reactions, CH4 decomposition and CO disproportionation, the former reaction was considered as the main course in the investigated temperature range of 700-900 ℃ [7]. Thus, to obtain additional information about the effect of the MgO promoter with a low concentration on the removal process of carbon-containing intermediates created from CH4 cracking over the surface of the catalyst during the CRM reaction, each catalyst was subjected to turn-by-turn pulse experiments. Before the results are presented, some points should be discussed. The residence time of the gases in the reactor was determined by the degree of interaction (adsorption/desorption) with the catalyst surface, by the rate of the surface reaction on the catalysts and by gas phase diffusion. First, Ross et al. [46] proposed that the gas phase diffusion is governed by Knudsen diffusion only when the pulse intensity is less than 1016 molecules. In this pulse study, first CH4 then CO2 were introduced in pulses into the reactor under a flow of purified He gas (50 mL/min), which was too large for the gas transport mechanism to be determined by Knudsen diffusion. Second, CO adsorption enthalpies were 135 kJ/mol at low coverages (< 0.02) and much lower at higher coverages on Ni(1 0 0), indicating that CO coverages should be well below 0.01 monolayer even at 600 ℃ [10]. As can be seen in Fig. 10, sharp CO peaks were observed over both catalysts. Thus, the influence of CO desorption on residence time could be ruled out.
In this experiment, CH4 and CO2 were injected in turn for four pulses each under a He steady carrier gas over both reduced catalysts at 750 ℃. The results are shown in Fig. 12. In Figs. 13 and 14, the product distribution as a function of the number pulse on each catalyst is shown. For the Ni-Mg/Al catalyst, when the first CH4 pulse was introduced, H2 and CO were generated immediately. The reaction process was likely preceded by the decomposition of CH4 over Ni metal sites to give H2 and absorbed CHx species, followed by an oxidation reaction of hydroxyl groups on the alumina support with absorbed CHx species to give H2 and CO, as has been reported previously over various catalysts [7, 47-49]. Fig. 13(b) shows that the CO yield after CH4 was pulsed was significantly lower than the amount of CH4 converted. Thus, the absorbed carbon containing species must have been left on the Ni-Mg/Al catalyst surface. After an interval of 5 min, only CO was detected as soon as the first quantity of CO2 was pulsed. A similar phenomenon was reported by Sun et al. [50]. As shown in Fig. 13(b) and Fig. 14(b), as the number of pulses increased, the reactant conversion and the product yield over the Ni-Mg/Al catalyst remained the same. For the Ni-Al catalyst, the same procedure was performed. Certain differences, when compared with the Ni-Mg/Al catalyst, were: (1) although there was a similar amount of CH4 conversion after the first CH4 pulse, the CO yield and CO2 conversion were lower over the Ni/Al catalyst, indicating that more carbon containing species existed over the Ni/Al catalyst after the first CO2 pulse; (2) the CO feedback signal of the Ni/Al catalyst after CO2 pulses gradually became broader, indicating that the rate of the oxidation reaction on the Ni metal surface became slower (as seen in the region indicated by the arrows in Fig. 12(a)); (3) as observed in Fig. 13(a) and Fig. 14(a), the amount of reactant conversion and product yield decreased gradually as the pulse number increased owing to the carbon deposition blocking the Ni metal sites on the Ni/Al catalyst.
As illustrated from the XRD patterns (as shown in Fig. 1(3)), after dipping in the Mg2+ nitrate solution, the reconstruction of the Ni-Al mixed oxide occurred and no other diffraction lines were observed. This result indicates that the reconstruction was completely finished within 45 min under the conditions applied in this work, in good agreement with a previous report [29]. Among the previous publications, a decrease in the particle size of Ni metal has been reported to limit the formation of coke in the CRM. Different researchers have reported different critical particle sizes — approximately 7 to 10 nm for the CRM — below which carbon deposition can be avoided [51, 52]. According to the XRD of the reduced samples (Fig. 5), the particle sizes of Ni metal on the Ni/Al and Ni-Mg/Al catalyst were much larger than the "critical size". As might be expected, the Ni/Al catalyst presented much higher deactivation rates in CRM owing to the graphite carbon formation that was confirmed by TPO and Raman spectroscopy (Fig. 7 and Fig. 8). In agreement with previous papers, a similar deactivation of the catalysts has been reported owing to the larger Ni particle sizes (> "critical size") [53]. However, the Ni-Mg/Al catalyst with a low concentration of MgO promoter (0.42 wt.% Mg2+) showed a higher stability in the CRM reaction owing to the inhibition of graphitic carbon formation (as shown in Fig. 6 and Fig. 7). Numerous efforts have been made to minimize carbon formation, and the consensus in the literature is that carbon atoms can accumulate on Ni surfaces when the carbon deposition rate is higher than the gasification rates [54]. Thus, at least three possible reasons to account for such an unexpected observation may be considered.
The first possibility is that the MgO promoter may neutralize a fraction of the most active sites over the Ni metal surface in the CRM. The first successful strategy to inhibit coke deposition on Ni catalyst in MSR is the SPARG (sulfur-passivated reforming) process developed by Rostrup-Nielsen et al. [55], whereby the under-coordinated sites on the Ni surface are blocked by a small portion of H2S in the feedstock. It has been proven that Au, Pb, K, and so forth, also process the coke-resistant properties in the CRM by blockage of carbon nucleation centers on the Ni metal surface [25, 56, 57]. However, blocking a fraction of active sites over the Ni metal surface by addition of similar promoters could decrease the CH4 conversion in the CRM. Nevertheless, as shown in Fig. 13(a) and (b), the CH4 conversion after the first CH4 pulse was similar for the Ni/Al and Ni-Mg/Al catalysts, indicating that there was no obvious difference in the activation of CH4 between the two catalysts.
Considering the carbon gasification process, it has been recognized that promoting CO2 activation is the key to restrict carbon deposition kinetically. Daniel et al. believed that the difference between noble metals and Ni regarding the carbon-resistance ability was a result of the different abilities of the metals for CO2 dissociation which could influence the performance of eliminating the carbon produced by CH4 decomposition [58]. There are two reaction pathways for CO2 activation and conversion which are generally agreed upon in the literature: (1) dissociative adsorption of CO2 on Ni metal sites forms CO and absorbed O atoms; the absorbed O atoms then react with carbon containing species to form CO; (2) adsorption of CO2 on a basic support forms carbonate species which are reduced by H species from CH4 cracking to form CO and -OH groups; then, there is an oxidation reaction of the surface -OH groups on the support with absorbed carbon-containing species on the Ni metal surface [7].
Therefore, a second possibility is that the presence of MgO may enhance CO2 absorption owing to its strong Lewis basic sites. However, this hypothesis is not supported by the result of CO2-TPD (as shown in Fig. 9). As the third possibility, the "memory effect" may facilitate the generation of new Ni-MgO active sites. It has been reported that metal-support interfacial sites can promote the dissociation and reduction of CO2 during the title reaction [54]. Tomishige et al. [59] suggested that CO2 plays an important role in the inhibition of carbon formation on Ni-MgO solid solution through the activation of CO2 at the Ni-MgO interface sites[42]. Besides, a DFT study showed that addition of Mg facilitated the interaction of gas phase oxygen with the adsorbed carbon promoting the formation of a CO precursor species [60]. In this work, although only 0.42 wt% Mg2+ ions were loaded, by using the "memory effect", a uniform distribution of Mg2+ is present on the surface of the Ni-Mg/Al catalyst (as shown in Fig. 2), which must favorable for generating more Ni-MgO sites. García et al. [22] also confirmed that the formation of the Ni-MgO interface sites would be favored for the lowest MgO content. As can be seen in Fig. 10 and Fig. 11, indeed, a low concentration of MgO promoter promotes CO2 dissociation on Ni metal sites owing to the formation of Ni-MgO interface sites. Combining the results of the catalytic tests and turn-by-turn pulse experiments suggests that Ni-MgO interface sites could improve the CO2 dissociation, which could supply more absorbed oxygen species, so that more carbon-containing intermediates from CH4 cracking could be quickly converted to CO product, thus preventing accumulation.
By applying the "memory effect" of Ni-Al hydrotalcite structures, when a low concentration of MgO promoter (0.42 wt% Mg2+) was added to a Ni-Al catalyst, the catalyst showed higher stability than that of a non-promoted Ni-Al catalyst. Owing to the low concentration, the MgO promoter did not influence Ni particle sizes and increased the amount CO2 adsorption for the Ni-Mg/Al catalysts, when compared with the Ni-Al catalysts. It has been confirmed by pulse experiments that the promotion with a low concentration of Mg shows a higher ability for decomposing CO2 to form active surface oxygen owing to the formation of Ni-MgO interface sites, and therefore the carbon-resistance promotion by nature is suggested to contribute to an oxidative environment around Ni particles. This means that more carbon-containing intermediates from CH4 cracking can be quickly converted into CO product, which prevents its accumulation.
2018,
Vol. 39 Issue (1): 88-98 
PDF
2018,
Vol. 39 Issue (1): 88-98 PDF