CO₂ reforming of methane over Mg-promoted Ni/SiO₂ catalysts: the influence of Mg precursors and impregnation sequences

Abstract

Mg-promoted Ni/SiO₂ catalysts with different Mg precursors were prepared by different impregnation sequences and used in CO₂ reforming of methane. The catalysts were characterized by XRD, TPR-H₂, TPD-CO₂, TG-DTA and SEM techniques. The use of Mg(CH₃COO)₂ as Mg precursor favored the performance of the catalyst. The impregnation of Mg(CH₃COO)₂ prior to Ni led to stronger interaction of nickel species with the support and the formation of stable Ni₂SiO₄ and Mg₂SiO₄ species, which inhibited the sintering of metallic Ni and then made the catalyst show better activity and stability. The catalyst thus prepared also exhibited enhanced capacity of CO₂ adsorption, which accelerated the CO₂ activation and the elimination of deposited carbon. No significant carbon deposition was observed on the surface of the catalysts, keeping the catalyst stable.

---

1. Introduction

Because of the increasing interest in reduction of CO₂ emissions and effective utilization of natural gas, CO₂ reforming of methane to produce synthesis gas has been paid much more attention from both academic and industrial standpoints.^{1–18,20–30} This process provides a good approach to convert the two greenhouse gases, CO₂ and methane, into synthesis gas with a theoretical CO/H₂ ratio of unity, which could be preferentially used for the production of oxygenated hydrocarbons. However, the industrialization of CO₂ reforming of methane remains a great challenge due to the lack of an efficient and stable catalyst, since it suffers from severe catalyst deactivation by coke deposition and metal sintering.

It is well known that noble metal-based catalysts exhibited high activity and good coke resistance.^25–33 Nevertheless, by consideration of the availability and economy, the development of non-noble metal-based catalysts is a promising route for this process. Many researchers reported that Fe, Co and Ni-based catalysts possessed high activity, but deactivated with time on stream due to coke formation and/or metal sintering.^5–7,13,14 Therefore, much more efforts were mainly focused on the enhancement of catalytic stability.

The catalytic performance of supported metallic catalysts for CO₂ reforming of methane is mainly affected by the type of support, the nature of metal and corresponding precursor, as well as the strategy of catalyst preparation and the conditions of pretreatment.^34–42 He et al.³⁴ investigated the catalytic performance of Ni/SiO₂ and Ni–Al₂O₃/SiO₂ catalysts prepared by citrate and nitrate precursors for the combination reactions of CO₂ reforming and partial oxidation of methane. They found that the catalytic activity significantly depended on the precursor of nickel and the interaction between NiO and support. Nickel citrate might be an excellent precursor for the preparation of Ni catalyst with high dispersion. The strong interaction between NiO and SiO₂ resisted the catalyst sintering at high temperature. They also reported that Ni/SiO₂ catalysts prepared with citrate precursor had excellent performances for their high nickel dispersion derived from the nickel silicate like species.³⁵ Özkara-Aydınoğlu et al.³⁹ studied the influence of cerium amount and the impregnation strategy on the activity and stability of Ce-promoted Pt/ZrO₂ catalysts for CO₂ reforming of methane. It was reported that the influence of the Ce promoter on catalytic activity depended on the cerium amount and impregnation strategy. The co-impregnation method resulted in extensive and strong Pt–Ce surface interaction and enhanced the Pt dispersion. Wang and Lu⁴⁰ studied the effect of nickel precursors on the catalytic performance of Ni/Al₂O₃ catalysts for CO₂ reforming of methane, and found that nickel nitrate was superior to nickel chloride and nickel acetylacetonate in terms of activity and stability. They also reported that Ni/Al₂O₃ catalysts prepared with different nickel precursors resulted in different carbon species formed on the catalyst surface, influencing the catalytic stability.

In our previous work, we investigated the effect of La, Mg, Co and Zn promoters on the catalytic performance of Ni/SiO₂ with an emphasis on the activity and stability. We found that Ni–La/SiO₂ catalysts exhibited excellent performance as compared to other catalysts. Ni–Mg/SiO₂ and Ni–Co/SiO₂ catalysts showed high initial activity, but poor stability. Meanwhile, Mg was reported to improve the catalytic performance due to its strong basicity, which enhanced the capacity of CO₂ adsorption, altered the acid–base property of catalysts and improved the dispersion of active phase.^11,16–18 Therefore, based on our previous investigation,¹⁸ the influence of Mg precursors and impregnation sequences on the performance of Mg-promoted Ni/SiO₂ catalysts for CO₂ reforming of methane was studied, aiming to develop a catalyst with enhanced activity and stability.

2. Experimental

2.1 Catalyst preparation

By modification of the catalyst preparation methods reported in previous literature,^18,19 the catalysts used in the present work were prepared by an incipient wetness impregnation method. For all catalysts, the loadings of Ni and Mg were 10 wt% and 5 wt%, respectively.

(1) The Ni–Mg/SiO₂ catalysts with different Mg precursors were prepared by the following procedure: the SiO₂ support (20–40 mesh, 320 m² g⁻¹) was impregnated with aqueous solution containing appropriate amounts of Ni(NO₃)₂ and Mg(CH₃COO)₂, or Mg(NO₃)₂, or MgCl₂, or MgSO₄ for 24 h, and dried at 110 °C for 4 h, then calcined in a muffle furnace at 800 °C for 5 h. These catalysts were denoted as Ni–Mg/SiO₂–C, Ni–Mg/SiO₂–N, Ni–Mg/SiO₂–Cl and Ni–Mg/SiO₂–S, respectively.

(2) The Ni/Mg/SiO₂ catalysts with different Mg precursors were prepared by the following procedure: the SiO₂ support (20–40 mesh, 320 m² g⁻¹) was firstly impregnated with aqueous solution containing an appropriate amount of Mg(CH₃COO)₂, or Mg(NO₃)₂, or MgCl₂, or MgSO₄ for 24 h, and dried at 110 °C for 4 h, then calcined in a muffle furnace at 800 °C for 5 h. The obtained samples were subsequently impregnated with aqueous solution containing an appropriate amount of Ni(NO₃)₂ for 24 h, and dried at 110 °C for 4 h, finally calcined in a muffle furnace at 800 °C for 5 h. These catalysts were denoted as Ni/Mg/SiO₂–C, Ni/Mg/SiO₂–N, Ni/Mg/SiO₂–Cl and Ni/Mg/SiO₂–S, respectively.

(3) The Mg/Ni/SiO₂ catalysts with different Mg precursors were prepared by the following procedure: the SiO₂ support (20–40 mesh, 320 m² g⁻¹) was firstly impregnated with aqueous solution containing an appropriate amount of Ni(NO₃)₂ for 24 h, and dried at 110 °C for 4 h, then calcined in a muffle furnace at 800 °C for 5 h. The obtained samples were subsequently impregnated with aqueous solution containing an appropriate amount of Mg(CH₃COO)₂, or Mg(NO₃)₂, or MgCl₂, or MgSO₄ for 24 h, and dried at 110 °C for 4 h, finally calcined in a muffle furnace at 800 °C for 5 h. These catalysts were denoted as Mg/Ni/SiO₂–C, Mg/Ni/SiO₂–N, Mg/Ni/SiO₂–Cl and Mg/Ni/SiO₂–S, respectively.

2.2 Catalytic activity test

The procedure of the activity test was reported in previous literature.¹⁸ The catalytic activity test was carried out in a fixed-bed flow under ambient atmospheric pressure. 0.25 g of catalyst was loaded in the middle of a micro-quartz-tube reactor. The catalyst was heated from room temperature to reaction temperature (800 °C) in argon flow. The catalyst was reduced in a H₂/Ar atmosphere for 1 h (H₂/Ar = 1, 60 ml min⁻¹) prior to reaction. Then the reactants CH₄ (99.99%) and CO₂ (99.5%) without any diluents were co-fed with a total flow rate of 60 ml min⁻¹ (CH₄/CO₂ = 1, F = 60 ml min⁻¹). The flow rate of reactant gases was controlled by mass flow controllers (D07-11A/ZM made by Beijing Sevenstar Huachuang Electronic). The effluent after the removal of water was analyzed by an on-line gas chromatograph (GC) with a Plot-C2000 capillary column for separating the products and a TCD to determine the concentration of each component.

In this work, the catalytic activity was calculated as mentioned in the literature.²⁴ All the activity results discussed in this paper were obtained by at least three parallel tests and the deviation of activity was about ±0.2%.

2.3 Catalyst characterization

The actual amounts of Ni and Mg on the fresh catalysts were determined by atomic absorption spectroscopy (AAS). The fresh and used catalysts were comparatively characterized by an XRD technique. The X-ray diffraction patterns of the catalysts were recorded with a DX-1000 CSC diffractometer using Cu Kα monochromatic X-ray radiation (40 kV and 25 mA).

The reduction behaviors of fresh catalysts were characterized by temperature-programmed reduction (TPR) with H₂ according to our previous report.¹⁸ In short, 0.25 g of catalyst was pretreated in N₂ flow at 150 °C for 30 min before reduction, and then cooled down to 80 °C. The TPR test was performed under a N₂/H₂ mixture (H₂/N₂ = 5/95 vol%) from 100 °C to about 830 °C with a heating rate of 8 °C min⁻¹. The amount of consumed H₂ was determined on-line by TCD. The conditions were selected to avoid the effect of mass transfer on the reduction behavior according to the literature.^18,21,22

Temperature-programmed desorption (TPD) of CO₂ was carried out on a conventional apparatus. 0.25 g of catalyst was pretreated in N₂ flow at 300 °C for 60 min before the adsorption of CO₂, and then cooled to below 50 °C. The samples were saturated by pure CO₂ stream for 30 min, then the gas was changed to pure helium to purge the gas line and remove weakly adsorbed CO₂. The TPD test was performed from room temperature to 800 °C with a heating rate of 8 °C min⁻¹. The amount of CO₂ desorbed was monitored on-line by TCD.

The morphologies of typical fresh and used catalysts were observed by scanning electron microscopy (SEM, FEI Inspect F). The samples were covered with a thin film of gold (ion sputtering) to improve the conductivity.

TG-DTA analysis was employed to determine the amount of carbon deposited on the surface of typical used catalysts. The TG-DTA measurements were carried out in air flow (70 ml min⁻¹) with a NETZSCH 409 PC/PG analyzer using 15–30 mg and a heating rate of 10 °C min⁻¹.

3. Results

3.1 Atomic absorption spectroscopy analysis

The actual amounts of Ni and Mg determined by AAS analysis are listed in Tables 1 and 2. It was found that the actual amounts are close to the controlled amounts with the exception of a relatively low Mg content for samples prepared with MgSO₄ precursor. This illustrated that Ni and Mg were successfully impregnated on the SiO₂ support during the preparation process.

Table 1 The catalytic activity of selected catalysts at 30 h^a

Catalyst	Actual amount^b (%)		Conversion^c (%)		Yield^c (%)		CO/H2^c
	Ni	Mg	CH4	CO2	CO	H2
a Conditions: 800 °C, F = 60 ml min^−1, CH4/CO2 = 1/1, m = 0.25 g. b According to AAS results. c Activity obtained at 30 h on stream.
Ni/Mg/SiO2–C	8.6	4.8	83.1	84.0	79.3	53.0	1.50
Ni/Mg/SiO2–N	8.4	4.5	78.5	80.6	74.6	49.5	1.51
Ni/Mg/SiO2–Cl	8.6	4.4	45.6	53.6	41.5	21.6	1.92
Ni/Mg/SiO2–S	7.9	2.0	44.6	52.4	40.0	20.7	1.93
Ni–Mg/SiO2–C	8.4	4.3	84.8	84.9	81.1	54.6	1.49
Mg/Ni/SiO2–C	8.3	4.5	50.0	52.8	43.1	27.9	1.54

---

Table 2 The catalytic activity of catalysts within 10 h^a

Catalyst	Actual amount^b (%)		Conversion^c (%)		Conversion^d (%)		CO/H2^d
	Ni	Mg	CH4	CO2	CH4	CO2
a Conditions: 800 °C, F = 60 ml min^−1, CH4/CO2 = 1/1, m = 0.25 g. b According to AAS results. c Activity obtained at 1 h on stream. d Activity obtained at 10 h on stream.
Ni–Mg/SiO2–N	8.4	4.2	86.2	87.6	74.6	78.6	1.50
Mg/Ni/SiO2–N	8.2	4.6	33.9	36.4	46.9	52.1	1.66
Ni–Mg/SiO2–Cl	8.1	4.4	35.1	42.8	36.7	44.7	2.14
Mg/Ni/SiO2–Cl	8.4	4.6	34.4	42.4	11.2	12.7	1.62
Ni–Mg/SiO2–S	7.8	2.2	31.5	34.7	46.4	53.1	1.77
Mg/Ni/SiO2–S	7.5	1.9	15.7	20.7	37.0	47.5	1.76

---

3.2 XRD characterization

The X-ray patterns of several typical fresh catalysts are plotted in Fig. 1. By comparison of Ni/Mg/SiO₂ catalysts with different Mg precursors, it was found that the XRD peaks attributed to NiO phase were detected on all the four catalysts. Especially, NiO was highly dispersed on Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂. Nevertheless, NiO species with particle sizes of 17 and 13 nm calculated from the Scherrer–Warren formula was observed on Ni/Mg/SiO₂ prepared with MgCl₂ and MgSO₄ precursors. Moreover, as Mg(CH₃COO)₂ and Mg(NO₃)₂ were used, new phases of Ni₂SiO₄ and Mg₂SiO₄ appeared on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts. This demonstrated that the interaction between Ni, Mg species and SiO₂ support was enhanced on the two catalysts, leading to the formation of Ni₂SiO₄ and Mg₂SiO₄ species (it is hard to discriminate between Ni₂SiO₄ and Mg₂SiO₄ as the diffractions of these phases overlap partly). Ni₂SiO₄ and Mg₂SiO₄ species were formed due to the interaction of nickel oxide and magnesium oxide with SiO₂ support as reported in other literature.^43–47 However, other new peaks corresponding to MgSiO₃ were detected on Ni/Mg/SiO₂ catalysts prepared with MgCl₂. For the series of Ni–Mg/SiO₂ and Mg/Ni/SiO₂ catalysts, NiO phase was found on all these catalysts. Ni₂SiO₄ and Mg₂SiO₄ species were only detected on Ni–Mg/SiO₂–C, while MgSiO₃ species was detected on Ni–Mg/SiO₂–Cl and Mg/Ni/SiO₂–Cl catalysts (see ESI†, Fig. S1).

Fig. 1 XRD patterns of the typical fresh catalysts. (a) Ni/Mg/SiO₂–C; (b) Ni/Mg/SiO₂–N; (c) Ni/Mg/SiO₂–Cl; (d) Ni/Mg/SiO₂–S; (e) Ni–Mg/SiO₂–C; (f) Mg/Ni/SiO₂–C; (○) NiO; (●) Ni₂SiO₄; (△) Mg₂SiO₄; (◆) MgSiO₃.

As Mg(CH₃COO)₂ was used, except for highly dispersed NiO species, Ni₂SiO₄ and Mg₂SiO₄ phases appeared on Mg-promoted Ni/SiO₂ catalysts prepared by impregnating Mg prior to Ni and co-impregnation. It can be inferred that the two impregnation sequences were more beneficial to enhance the interaction with the support, leading to the formation of Ni₂SiO₄ and Mg₂SiO₄ species. However, only NiO species with a particle size of about 15 nm was detected on catalysts prepared by impregnating Ni prior to Mg. As Mg(NO₃)₂ was used, except for highly dispersed NiO phase, Ni₂SiO₄ and Mg₂SiO₄ appeared on the catalysts prepared by impregnating Mg prior to Ni, while only NiO phase was found on the samples prepared by other two impregnation sequences. Nevertheless, there was no significant influence of impregnation sequences on the structure of catalysts as MgCl₂ and MgSO₄ precursors were employed.

In brief, the structure of the catalysts was influenced significantly by the addition of different Mg precursors. Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂ possessed high NiO dispersion and stronger surface interaction, leading to the formation of Ni₂SiO₄ and Mg₂SiO₄ species. Meanwhile, the impregnation sequence influenced the structure of the catalysts. Especially, as Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors were employed, the catalysts prepared by impregnating Mg prior to Ni exhibited higher NiO dispersion and enhanced interaction with the support, resulting in the formation of Ni₂SiO₄ and Mg₂SiO₄ species.

3.3 TPR results

The reduction behaviors of the catalysts were characterized by TPR experiments. The TPR-H₂ profiles of typical samples are presented in Fig. 2. By comparison of Ni/Mg/SiO₂ catalysts with different Mg precursors, it was found that Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors had only one reduction peak with a maximum temperature of higher than 800 °C, which was assigned to the reduction of nickel oxides bearing strong interaction with the support. Zhang et al.⁴⁷ reported that the reduction of Ni₂SiO₄ needs a higher temperature of about 630–700 °C. Actually, the Ni₂SiO₄ and Mg₂SiO₄ species were detected by XRD characterization mentioned above. In addition, the low intensity of the reduction peak indicated that the amount of reducible species on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts was low. It is well known that the compounds bearing strong interaction with the support provided higher metal dispersion and high resistance to sintering.^15,16 On the other hand, except for the not ended peak at about 800 °C, a very broad peak starting from 435 to 730 °C was found on the TPR profile of Ni/Mg/SiO₂ prepared with MgCl₂. Meanwhile, three well-defined peaks with maxima at 440, 508, and 595 °C were detected on Ni/Mg/SiO₂ catalysts prepared with MgSO₄. This illustrated that nickel oxides bearing different interactions with the support existed on Ni/Mg/SiO₂–Cl and Ni/Mg/SiO₂–S catalysts. Moreover, the higher area of these peaks on the two catalysts demonstrated that the amount of reducible species was more than that on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts. Similar results were obtained on the series of Ni–Mg/SiO₂ and Mg/Ni/SiO₂ catalysts. Especially, as MgCl₂ and MgSO₄ were used, the relatively lower reduction temperature and higher area of peaks were found on the profiles of TPR (see ESI†, Fig. S2).

Fig. 2 TPR-H₂ profiles of the typical fresh catalysts.

As Mg(CH₃COO)₂ was used, except for the peak located between 520–730 °C, another new obvious peak with maximum at 474 °C ascribed to “free state” nickel oxides bearing weak interaction with the support was detected on the catalyst prepared by impregnating Ni prior to Mg. The same peaks were reported in other literature.^18,23,24 Nevertheless, only peaks with lower intensity located at higher than 800 °C were detected on catalysts prepared by co-impregnation and impregnating Mg prior to Ni. It can be inferred that the amount of reducible species was low and the metallic oxides bearing enhanced interaction with the support existed on the two catalysts, resulting in high reduction temperature. On the other hand, when the Mg(NO₃)₂ precursor was used, the catalyst prepared by impregnating Mg prior to Ni exhibited higher reduction temperature with lower peak area, as compared with other two impregnation sequences. Nevertheless, the impregnation sequence did not exert a significant effect on the reduction behavior of catalysts prepared with MgCl₂ and MgSO₄ precursors (see ESI†, Fig. S2).

In summary, Mg precursors and impregnation sequences influenced the reduction behavior of the catalysts. Totally speaking, Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂ exhibited high reduction temperature and enhanced interaction, leading to lower reducibility. Meanwhile, as Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors were used, impregnating Mg prior to Ni increased the interaction of metallic oxide with SiO₂ support much more significantly, keeping the catalysts stable.

3.4 TPD-CO₂ results

Fig. 3 presents the CO₂-TPD profiles of several typical fresh catalysts. Generally speaking, the temperature of CO₂ desorption reflected the strength of basic sites, while the areas of desorption peaks reflected the amount of basic sites on the surface of catalysts. In analyzing the TPD-CO₂ profiles of Ni/Mg/SiO₂ with different Mg precursors, it was found that a CO₂ desorption peak around 110 °C was detected on all the Ni/Mg/SiO₂ catalysts, which was ascribed to the weak basic sites. A similar CO₂ desorption peak was also reported on Ni-based catalysts in the literature.^11,17,48,49 Moreover, the area of desorption peaks on the TPD profiles of Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–S was much more than that on Ni/Mg/SiO₂–N and Ni/Mg/SiO₂–Cl catalysts. This illustrated that Ni/Mg/SiO₂ prepared with Mg(CH₃COO)₂ and MgSO₄ precursors had much more basic sites and enhanced ability of CO₂ adsorption, which was much more beneficial to CO₂ activation and the removal of deposited carbon.^22,49 Especially, the higher intensity of a desorption peak around 110 °C on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–S demonstrated that the equilibrium of CO₂ adsorption–desorption could take place much faster and easier at low temperature. However, except for the desorption peaks at 210 and 565 °C, which was attributed to moderate basic sites, another not ended peak with much higher intensity at about 800 °C was found on the TPD profile of Ni/Mg/SiO₂–S. This illustrated that the amount of strong basic sites on Ni/Mg/SiO₂–S was much higher and the CO₂ desorption could take place at higher temperature. On the other hand, the high intensity of CO₂ adsorption peaks at about 210 and 430 °C ascribed to moderate basic sites and a peak at 560–740 °C attributed to strong basic sites were detected on Ni/Mg/SiO₂–C catalysts. Similar peaks were also found on the profiles of Ni/Mg/SiO₂–Cl catalysts. Nevertheless, only a CO₂ desorption peak at 430 °C with lower intensity and a not ended peak around 800 °C were detected on Ni/Mg/SiO₂–S catalysts.

Fig. 3 TPD-CO₂ profiles of the typical fresh catalysts.

When the Mg(CH₃COO)₂ precursor was used, by comparison with the TPD-CO₂ desorption profiles of catalysts prepared by different impregnation sequences, it can be seen that the desorption area on Ni/Mg/SiO₂–C and Ni–Mg/SiO₂–C was higher than that on Mg/Ni/SiO₂–C catalysts. This illustrated that the amount of basic sites on Mg-promoted Ni/SiO₂ catalysts prepared by impregnating Mg prior to Ni and co-impregnation was more than on that prepared by impregnating Ni prior to Mg. The same desorption peak at a low temperature of about 110 °C was also detected on all the three Mg-promoted Ni/SiO₂ catalysts prepared with Mg(CH₃COO)₂ precursor. Especially, the amount of weak basic sites on Ni/Mg/SiO₂–C was much higher than that on Ni–Mg/SiO₂–C and Mg/Ni/SiO₂–C catalysts. This illustrated that the establishment of CO₂ adsorption–desorption equilibrium was much easier on the former catalysts. Meanwhile, another CO₂ desorption peak at 400–800 °C with low intensity was detected on Mg/Ni/SiO₂. Several CO₂ desorption peaks at 210, 430, 565, 720, 800 °C, and CO₂ desorption peaks at 210, 430, 640, 720, 800 °C were found on the TPD-CO₂ profiles of Ni–Mg/SiO₂–C and Ni/Mg/SiO₂–C catalysts, respectively. This indicated that the basic sites with different strengths existed on the surface of the catalysts.

In short, the basicity of Mg-promoted Ni/SiO₂ catalysts was influenced by the addition of different Mg precursors and impregnation sequences. Three kinds of basic sites, named weak, moderate and strong basic sites existed on the surface of catalysts. The capacity of CO₂ adsorption on the catalyst prepared with Mg(CH₃COO)₂ and MgSO₄ was much stronger, especially the weak basic sites increased significantly, leading to the faster and easier establishment of CO₂ adsorption–desorption equilibrium. Meanwhile, as Mg(CH₃COO)₂ was used, the basicity of catalysts prepared by impregnating Mg prior to Ni and co-impregnation was increased much more remarkably, which were more beneficial to the CO₂ adsorption and the elimination of deposited carbon.

3.5 Catalytic activity and stability

The activity of the catalysts was monitored by means of the conversions of methane and CO₂, the yields of CO and H₂, as well as the molar ratio of CO/H₂. The activity results of several typical catalysts at 30 h are listed in Table 1. Under the following conditions: 800 °C, F = 60 ml min⁻¹, CH₄/CO₂ = 1/1, 0.25 g catalyst, by comparison with the activity of Ni/Mg/SiO₂ catalysts with different Mg precursors, it was found that Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors possessed higher conversion of CH₄ and CO₂, as well as higher yield of CO and H₂. However, Ni/Mg/SiO₂ catalysts prepared with both MgCl₂ and MgSO₄ exhibited a relatively low activity. This indicated that Mg(CH₃COO)₂ and Mg(NO₃)₂ might be better candidates as Mg precursors for the preparation of Ni/Mg/SiO₂ catalysts for use in CO₂ reforming of methane as compared to other two Mg precursors. On the other hand, the catalytic performance was also influenced by the impregnation sequence remarkably. As seen in Table 1, it was found that Ni/Mg/SiO₂–C and Ni–Mg/SiO₂–C showed higher activity than Mg/Ni/SiO₂–C catalysts. That is to say, as Mg(CH₃COO)₂ was used, impregnating Ni prior to Mg was not an effective method to promote the catalytic performance, while the other two impregnation methods improve the activity more significantly. The six typical catalysts were selected for further investigation with an emphasis on the influence of Mg precursors and impregnation sequences on the catalytic performance.

Under the same reaction conditions mentioned above, the activity results of other six catalysts within 10 h are shown in Table 2. It was found that all the catalysts exhibited a relatively low activity and poor stability with the exception of Ni–Mg/SiO₂–N and Mg/Ni/SiO₂–Cl catalysts, which had a relatively higher initial activity but decreased with time on stream.

In all activity tests, the molar ratio of CO/H₂ was more than unity due to the occurrence of reverse water-gas shift reaction. This is in accordance with other reported literature.^4,7,18 In addition, higher conversions of CH₄ and CO₂ with lower CO/H₂ molar ratio were obtained on these catalysts. Therefore, lower CO/H₂ molar ratio was obtained on Ni/Mg/SiO₂–C, Ni/Mg/SiO₂–N and Ni–Mg/SiO₂–C catalysts, which illustrated that the reverse water-gas shift reaction was inhibited much more significantly on these catalysts, leading to higher H₂ yield.

Fig. 4 presents the tendency of CH₄ conversion on selected catalysts with time on stream. It can be seen that the CH₄ conversion on Ni/Mg/SiO₂–C decreased slightly, which was similar to that on Ni/Mg/SiO₂–N and Ni–Mg/SiO₂–C catalysts. However, the other three catalysts showed low activity and poor stability. Similar results were found on the conversion of CO₂ (see ESI†, Fig. S3).

Fig. 4 The tendency of CH₄ conversion on selected catalysts within 30 h.

3.6 The characterization of used catalysts

The XRD patterns of selected catalysts after 30 h reaction are comparatively plotted in Fig. 5. Typically, NiO species were totally reduced to metallic Ni. However, the peaks ascribed to carbon were not detected, which was different from our previous report.¹⁸ It might be inferred that low amount of carbon deposition and/or high dispersion of carbon occurred on all the used catalysts. On the other hand, as compared with the XRD profiles of fresh catalysts, the peaks attributed to Ni₂SiO₄ and Mg₂SiO₄ species still existed on used Ni/Mg/SiO₂–C, Ni–Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts, even after long-time reaction of 30 h. This implied the excellent stability of Ni₂SiO₄ and Mg₂SiO₄ species, as reported in other literature.^46,47 Nevertheless, the intensity of peaks corresponding to MgSiO₃ species on Ni/Mg/SiO₂–Cl catalysts weakened after 30 h reaction, indicating that MgSiO₃ species exhibited a poor stability with time on stream. Świerczyński et al.⁴⁴ reported that MgSiO₃ species could be diminished in the presence of MgO and/or NiO to form Mg₂SiO₄ and/or NiMgSiO₄ at high temperature. Mg₂SiO₄ and NiMgSiO₄ species were not observed on the used Ni/Mg/SiO₂–Cl catalyst in the present work, which might be due to their lower amount and/or their higher dispersion on the surface of the catalyst. By calculating the particle size of metallic Ni, it was found that Ni/Mg/SiO₂–C, Ni–Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts had the particle size of about 11 nm, which was smaller than that on used catalysts of Mg/Ni/SiO₂–C (18 nm), Ni/Mg/SiO₂–Cl (21 nm) and Ni/Mg/SiO₂–S (15 nm).

Fig. 5 XRD patterns of the typical used catalysts after 30 h reaction. (⋄) Ni; (●) Ni₂SiO₄; (△) Mg₂SiO₄; (◆) MgSiO₃.

The carbon deposited on the catalyst surface after 30 h reaction was quantified by TG-DTA analysis in air flow. The TG-DTA profiles are plotted in Fig. 6. In analyzing the TG profiles, the initial step of weight loss occurred over the temperature range of 110–300 °C, which is attributed to the removal of moisture and easily oxidized carbonaceous species.^18,41 Meanwhile, the oxidation of metallic Ni and carbonaceous species could take place simultaneously thereafter up to 500 °C. The net result depended on the relative amount of nickel and deposited carbon. An obvious weight increase was found on Ni/Mg/SiO₂–Cl, Ni/Mg/SiO₂–S and Mg/Ni/SiO₂–C. Meanwhile, a slight weight decrease was found on Ni/Mg/SiO₂–C, Ni/Mg/SiO₂–N and Ni–Mg/SiO₂–C catalysts. According to TG profiles, it might be deduced that the amount of carbon deposited on the catalysts was low, which could be also supported by the analysis of DTA profiles (Fig. 6b). No remarkable exothermic peaks were observed for the used catalysts in all temperature profiles, which indicated that the amount of carbon deposited on the catalyst surface might be low. This is in accordance with the TG results mentioned above.

Fig. 6 TG-DTA profiles of the typical used catalysts. (a) TG profiles; (b) DTA profiles.

The selected catalysts after 30 h reaction were characterized by a SEM technique and the results are presented in Fig. 7. The corresponding fresh catalysts were also characterized comparatively (see ESI†, Fig. S4). No significant difference in morphology was observed between the used and fresh catalysts.

Fig. 7 SEM images of the typical used catalysts after 30 h. (a) Ni/Mg/SiO₂–C, (b): Ni/Mg/SiO₂–N, (c) Ni/Mg/SiO₂–Cl, (d) Ni/Mg/SiO₂–S, (e) Ni–Mg/SiO₂–C, (f) Mg/Ni/SiO₂–C.

Meanwhile, as shown in Fig. 7, by comparison of Ni–Mg/SiO₂ catalysts prepared with Mg(NO₃)₂ precursor reported previously,¹⁸ the carbon whisker was not detected on all these Mg-promoted Ni/SiO₂ catalysts. The IR spectra proved that no absorption peaks ascribed to carbon whisker were detected (see ESI†, Fig. S5). It might be deduced that the carbon whisker was eliminated quickly as formed for its low amount, or transferred into inactive carbon with time on stream, leading to the partial deactivation. However, by consideration of the TG-DTA results, it is reasonable to infer that the carbon deposited on these catalysts might be low, even on Ni/Mg/SiO₂–C, Ni/Mg/SiO₂–N and Ni–Mg/SiO₂–C catalysts with higher catalytic activity.

4. Discussion

One of the major obstacles for the industrialization of CO₂ reforming of methane into synthesis gas is the deactivation of catalysts caused by carbon deposition and metal sintering during reaction.^5–7,13,14 Therefore, in recent decades, many researchers were devoted to develop nickel catalysts with improved activity and stability. They mainly focused on the preparation method of catalysts, the nature of the support and the addition of different promoters.^{1,8,36–42,50}

4.1 The influence of Mg precursors

In the present work, four Mg precursors were employed to investigate the effect of Mg precursors on the structure, reducibility, basicity, as well as the catalytic activity and stability of the catalysts. Totally speaking, Mg(CH₃COO)₂ might be the better candidate as Mg precursor for the preparation of Mg-promoted Ni/SiO₂ catalysts as compared to Mg(NO₃)₂, MgCl₂ and MgSO₄. For the series of Ni/Mg/SiO₂ catalysts, except for NiO species observed on all the four catalysts, Ni₂SiO₄ and Mg₂SiO₄ species were formed on Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors, while MgSiO₃ species was observed on Ni/Mg/SiO₂ catalysts prepared with MgCl₂. Moreover, the intensity of the peaks attributed to MgSiO₃ species weakened, while Ni₂SiO₄ and Mg₂SiO₄ species still existed after long-time reaction. Meanwhile, higher reduction temperature observed on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts illustrated that the interaction between metallic oxide and support was enhanced more significantly as Mg(CH₃COO)₂ and Mg(NO₃)₂ were employed, resulting in the decrease of their reducibility and the formation of Ni₂SiO₄ and Mg₂SiO₄ species on the two catalysts. Nevertheless, lower interaction of metallic oxide with support was found on Ni/Mg/SiO₂ prepared with MgCl₂ and MgSO₄ deduced from their low reduction temperature. Bouarab et al.¹⁷ reported that Mg₂SiO₄ phase promoted the formation of well dispersed cobalt particles and inhibited the sintering at high temperature, and consequently improved the catalytic stability under severe dry reforming conditions. By consideration of the higher activity obtained on Ni/Mg/SiO₂–C and Ni/Mg/SiO₂–N catalysts, it can be inferred that the formation of stable Ni₂SiO₄ and Mg₂SiO₄ species on the two catalysts might be responsible for their better catalytic performance. However, the lower ability of CO₂ adsorption on Ni/Mg/SiO₂–N might lead to the slight decrease of its activity. In contrast, the basicity of Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ precursor was enhanced much more significantly, which promoted the ability of CO₂ adsorption, resulting in a better CO₂ conversion and removal of deposited carbon. On the other hand, the series of Ni–Mg/SiO₂ and Mg/Ni/SiO₂ catalysts prepared with Mg(CH₃COO)₂ precursor exhibited similar results, that is, they possessed better catalytic performance as compared to those prepared with other Mg precursors.

In summary, the Mg precursors employed in this work influenced the catalytic performance significantly. In the preparation process, different precursors might influence the interaction between Mg precursors and the support. Moreover, the decomposition process of precursors and the producing gas atmosphere may affect the formation of active phases by altering the dispersion, structure, reducibility and basicity. For the series of Ni/Mg/SiO₂ catalysts, the stable Ni₂SiO₄ and Mg₂SiO₄ species formed only when Mg(CH₃COO)₂ and Mg(NO₃)₂ precursors were used. Ni₂SiO₄ and Mg₂SiO₄ species might be beneficial to the formation of highly dispersed active phase and inhibit the sintering of metallic Ni. Consequently, the catalytic performance, especially the stability was improved.

4.2 The influence of impregnation sequences

In this section, the influence of impregnation sequences on the structure, reducibility, basicity as well as the catalytic activity and stability of the catalysts was discussed. Totally speaking, impregnating Mg prior to Ni might be more beneficial to improve the catalytic performance as compared to the methods of co-impregnation and impregnating Ni before Mg. As the Mg(CH₃COO)₂ precursor was employed, except for highly dispersed NiO, Ni₂SiO₄ and Mg₂SiO₄ species were detected on Mg promoted Ni/SiO₂ catalysts prepared by impregnating Mg prior to Ni and co-impregnation. Meanwhile, the higher reduction temperature observed on Ni/Mg/SiO₂–C and Ni–Mg/SiO₂–C catalysts indicated that the interaction of metallic oxide with SiO₂ support was enhanced more remarkably, correspondingly, resulting in the decrease of their reducibility and the formation of stable Ni₂SiO₄ and Mg₂SiO₄ species, which might prevent the sintering of catalysts. Moreover, the catalysts prepared by impregnating Mg prior to Ni and co-impregnation exhibited much more enhanced basicity, which was beneficial to CO₂ adsorption and activation. Increasing the capacity of CO₂ absorption was proposed to reduce carbon deposition and keep the catalyst stable.^22,49 Additionally, the adsorption of CH₄ and CO might be suppressed partially for the competitive adsorption of CO₂ and could decrease the carbon deposition due to methane cracking and CO disproportionation reaction.⁴⁹ In contrast, the catalyst prepared by impregnating Ni before Mg had a weak interaction of nickel oxide with the support and the relatively lower basicity, which might be related to its low activity and poor stability. As the Mg(NO₃)₂ precursor was used, although higher initial activity was obtained on Ni/Mg/SiO₂–N and Ni–Mg/SiO₂–N catalysts, the latter exhibited poor stability. Mg/Ni/SiO₂–N catalysts showed only lower activity. In addition, the higher reduction temperature observed on Ni/Mg/SiO₂–N illustrated its enhanced interaction of nickel oxide with the support. Therefore, impregnating Mg prior to Ni improved the catalytic performance more effectively as the Mg(NO₃)₂ precursor was used. On the other hand, as MgCl₂ and MgSO₄ precursors were employed, the structure, reduction behavior and the catalytic performance of the catalysts were not influenced significantly by the impregnation sequence. In a word, impregnating Mg prior to Ni was a more effective method for the preparation of Mg-promoted Ni/SiO₂ catalysts used in CO₂ reforming of methane.

In summary, Mg-promoted Ni/SiO₂ catalysts prepared by impregnating Mg prior to Ni with the Mg(CH₃COO)₂ precursor possessed stronger interaction between nickel species with the support, resulting in the formation of Ni₂SiO₄ and Mg₂SiO₄ species, which might be responsible for their high activity and good stability. Moreover, the catalyst simultaneously exhibited enhanced capacity of CO₂ adsorption, which accelerated the removal of deposited carbon and kept the catalyst stable. Nevertheless, Ni/Mg/SiO₂ catalysts prepared with Mg(NO₃)₂ exhibited higher interaction of nickel oxide with SiO₂ support but relative lower ability of CO₂ adsorption, while that prepared with MgSO₄ had strong basicity but relative lower interaction between metallic oxide and support. Moreover, Ni/Mg/SiO₂ catalysts prepared with MgCl₂ only showed moderate interaction of nickel oxide with the support and a lower capacity of CO₂ adsorption, resulting in their low activity and poor stability.

5. Conclusion

Mg-promoted Ni/SiO₂ catalysts with different Mg precursors were prepared by different impregnation sequences and investigated in CO₂ reforming of methane. Mg(CH₃COO)₂ might be the better candidate as Mg precursor as compared to Mg(NO₃)₂, MgCl₂ and MgSO₄. Moreover, impregnating Mg prior to Ni is more beneficial to improve the activity and stability than the other two impregnation sequences, especially as the Mg(CH₃COO)₂ precursor was employed.

Ni/Mg/SiO₂ catalysts prepared with Mg(CH₃COO)₂ exhibited lower reducibility and higher ability of CO₂ adsorption. Meanwhile, when Mg(CH₃COO)₂ was used, the catalyst prepared by impregnating Mg prior to Ni possessed enhanced interaction of metallic oxide with the support and resulted in the formation of stable Ni₂SiO₄ and Mg₂SiO₄ species, which might prevent the catalyst sintering and be responsible for the better catalytic performance. Moreover, the catalyst prepared by this impregnation method simultaneously exhibited improved capacity of CO₂ adsorption, which accelerated the activation of CO₂ and the elimination of deposited carbon. Correspondingly, no whisker carbon was detected and the amount of inert carbon deposited on the surface of catalysts was low, which kept the catalyst stable.

Ni/Mg/SiO₂ catalysts prepared with Mg(NO₃)₂ exhibited higher interaction of nickel species with SiO₂ support but lower basicity, while that prepared with MgSO₄ had enhanced capacity of CO₂ adsorption but lower interaction of NiO with the support. Moreover, low activity and poor stability obtained on Ni/Mg/SiO₂ catalysts prepared with MgCl₂ resulted from lower ability of CO₂ adsorption and weaker interaction of nickel oxide with the support.

Acknowledgements

The authors are grateful for financial support from the NNSFC (No. 20976109, 21021001), the Special Research Foundation of Doctoral Education of China (No. 20090181110046), and the characterization of the catalysts from Analytical and Testing Center of Sichuan University.

Notes and references

1. Y.-X. Pan, P. Y. Kuai, Y. Liu, Q. F. Ge and C.-J. Liu, Energy Environ. Sci., 2010, 3, 1322.
2. F. Meshkani and M. Rezaei, Int. J. Hydrogen Energy, 2010, 35, 10295.
3. I. P. Silverwood, N. G. Hamilton, C. J. Laycock, J. Z. Staniforth, R. M. Ormerod, C. D. Frost, S. F. Parkerc and D. Lennon, Phys. Chem. Chem. Phys., 2010, 12, 3102.
4. M. M. Barroso-Quiroga and A. E. Castro-Luna, Int. J. Hydrogen Energy, 2010, 35, 6052.
5. N. N. Sun, X. Wen, F. Wang, W. Wei and Y. H. Sun, Energy Environ. Sci., 2010, 3, 366.
6. Ş. Özkara-Aydınoğlu and A. E. Aksoylu, Int. J. Hydrogen Energy, 2011, 36, 2950.
7. Ş. Özkara-Aydınoğlu and A. E. Aksoylu, Catal. Commun., 2010, 11, 1165.
8. X. S. Li, J.-S. Chang, M. Y. Tian and S.-E. Park, Appl. Organomet. Chem., 2001, 15, 109.
9. M. García-Diéguez, I. S. Pieta, M. C. Herrera, M. A. Larrubia and L. J. Alemany, Appl. Catal., A, 2010, 377, 191.
10. L. L. Xu, H. L. Song and L. J. Chou, Catal. Sci. Technol., 2011, 1, 1032.
11. V. García, J. J. Fernández, W. Ruíz, F. Mondragón and A. Moreno, Catal. Commun., 2009, 11, 240.
12. K. Nagaoka, K. Takanabe and K.-I. Aika, Chem. Commun., 2002, 1006.
13. S. Therdthianwong, A. Therdthianwong, C. Siangchin and S. Yongprapat, Int. J. Hydrogen Energy, 2008, 33, 991.
14. D. P. Liu, W. N. E. Cheo, Y. W. Y. Lim, A. Borgna, R. Lau and Y. H. Yang, Catal. Today, 2010, 154, 229.
15. D. P. Liu, X. Y. Quek, W. N. E. Cheo, R. Lau, A. Borgna and Y. H. Yang, J. Catal., 2009, 266, 380.
16. Y. H. Wang, H. M. Liu and B. Q. Xu, J. Mol. Catal. A: Chem., 2009, 299, 44.
17. R. Bouarab, O. Akdim, A. Auroux, O. Cherifi and C. Mirodatos, Appl. Catal., A, 2004, 264, 161.
18. J. Q. Zhu, X. X. Peng, L. Yao, J. Shen, D. M. Tong and C. W. Hu, Int. J. Hydrogen Energy, 2011, 36, 7094.
19. J. Q. Zhu, S. Qin, S. T. Ren, X. X. Peng, D. M. Tong and C. W. Hu, Catal. Today, 2009, 148, 310.
20. R. Bouarab, O. Cherifi and A. Auroux, Green Chem., 2003, 5, 209.
21. A. Venugopal, S. N. Kumar, J. Ashok, D. H. Prasad, V. D. Kumari, K. B. S. Prasad and M. Subrahmanyam, Int. J. Hydrogen Energy, 2007, 32, 1782.
22. Q. S. Jing, H. Lou, J. H. Fei, Z. Y. Hou and X. M. Zheng, Int. J. Hydrogen Energy, 2004, 29, 1245.
23. J. Gao, Z. Y. Hou, J. Z. Guo, Y. H. Zhu and X. M. Zheng, Catal. Today, 2008, 131, 278.
24. Y. X. Pan, C. J. Liu and P. Shi, J. Power Sources, 2008, 176, 46.
25. M. García-Diéguez, E. Finocchio, M. Á. Larrubia, L. J. Alemany and G. Busca, J. Catal., 2010, 274, 11.
26. B. Nematollahi, M. Rezaei and M. Khajenoori, Int. J. Hydrogen Energy, 2011, 36, 2969.
27. N. E. McGuire, N. P. Sullivan, O. Deutschmann, H. Y. Zhu and R. J. Kee, Appl. Catal., A, 2011, 394, 257.
28. M. García-Diéguez, I. S. Pieta, M. C. Herrera, M. A. Larrubia and L. J. Alemany, J. Catal., 2010, 270, 136.
29. H. Arbag, S. Yasyerli, N. Yasyerli and G. Dogu, Int. J. Hydrogen Energy, 2010, 35, 2296.
30. G. K. Reddy, S. Loridant, A. Takahashi, P. Delichère and B. M. Reddy, Appl. Catal., A, 2010, 389, 92.
31. Z.-J. Wang, Y. Zhao, L. Cui, H. Y. Du, P. Yao and C.-J. Liu, Green Chem., 2007, 9, 554.
32. S. Damyanova, B. Pawelec, K. Arishtirova, M. V. M. Huerta and J. L. G. Fierro, Appl. Catal., B, 2009, 89, 149.
33. O. Gamba, S. Moreno and R. Molina, Int. J. Hydrogen Energy, 2011, 36, 1540.
34. S. F. He, H. M. Wu, W. J. Yu, L. Y. Mo, H. Lou and X. M. Zheng, Int. J. Hydrogen Energy, 2009, 34, 839.
35. S. F. He, Q. S. Jing, W. J. Yu, L. Y. Mo, H. Lou and X. M. Zheng, Catal. Today, 2009, 148, 130.
36. S. Tang, L. Ji, J. Lin, H. C. Zeng, K. L. Tan and K. Li, J. Catal., 2000, 194, 424.
37. R. Pereñíguez, V. M. González-DelaCruz, J. P. Holgado and A. Caballero, Appl. Catal., B, 2010, 93, 346.
38. C. E. Daza, S. Moreno and R. Molina, Int. J. Hydrogen Energy, 2011, 36, 3886.
39. Ş. Özkara-Aydınoğlu, E. Özensoy and A. E. Aksoylu, Int. J. Hydrogen Energy, 2009, 34, 9711.
40. S. B. Wang and G. Q. Lu, Appl. Catal., A, 1998, 169, 271.
41. A. Nandini, K. K. Pant and S. C. Dhingra, Appl. Catal., A, 2005, 290, 166.
42. M. Rezaei, S. M. Alavi, S. Sahebdelfar, P. Bai, X. M. Liu and Z. F. Yan, Appl. Catal., B, 2008, 77, 346.
43. F. Pompeo, N. N. Nichio, M. G. González and M. Montes, Catal. Today, 2005, 107–108, 856.
44. D. Świerczyński, C. Courson, L. Bedel, A. Kiennemann and J. Guille, Chem. Mater., 2006, 18, 4025.
45. Q. Zafar, T. Mattisson and B. Gevert, Energy Fuels, 2006, 20, 34.
46. Q. Zafar, T. Mattisson and B. Gevert, Ind. Eng. Chem. Res., 2005, 44, 3485.
47. L. F. Zhang, W. Li, J. Liu, C. L. Guo, Y. P. Wang and J. L. Zhang, Fuel, 2009, 88, 511.
48. H.-S. Roh, K.-W. Jun, W.-S. Dong, J.-S. Chang, S.-E. Park and Y.-I. Joe, J. Mol. Catal. A: Chem., 2002, 181, 137.
49. X. C. Li, M. Wu, Z. H. Lai and F. He, Appl. Catal., A, 2005, 290, 81.
50. P. Chen, Z.-Y. Hou and X.-M. Zheng, Chin. J. Chem., 2005, 23, 847.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1: XRD patterns of several fresh catalysts; Fig. S2: TPR-H₂ profiles of several fresh catalysts; Fig. S3: the tendency of CO₂ conversion on selected catalysts within 30 h; Fig. S4: SEM images of the typical fresh catalysts; Fig. S5: IR spectra of the typical used catalysts after 30 h reaction. See DOI: 10.1039/c1cy00333j

---