An investigation on the influence of catalyst composition, calcination and reduction temperatures on Ni/MgO catalyst for dry reforming of methane

Abstract

In the present paper, the following issues regarding the dry reforming of methane, over Ni/MgO catalyst prepared by microemulsion synthesis were investigated: (i) the effect of nickel content (20, 40 and 80 wt%) in the Ni/MgO catalyst (ii) the effect of calcination temperatures (450 °C, 600 °C and 800 °C) on the solid solution formation between NiO and MgO (iii) the effect of reduction temperatures (550 °C and 800 °C) on the catalytic activity and stability of the catalyst at a very high space velocity of the reactants (CH₄/CO₂ = 1, GHSV = 1.68 × 10⁵ ml h⁻¹ g⁻¹). Under optimum conditions, the Ni/MgO catalyst having higher Ni content (80%), calcined at higher temperature (800 °C) and reduced at lower temperature (550 °C) exhibited better catalytic activity and stability. This was attributed to the presence of higher Ni^o active sites, formation of a strong NiO–MgO solid solution and also the catalyst was less prone to sintering at a lower reduction temperature. Furthermore, BET analysis of the Ni/MgO catalysts indicates a decrease in surface area from 153.22 to 54.01 m² g⁻¹, as the Ni content was increased from 20% to 80%, respectively. Furthermore, fresh and spent catalysts were characterized by BET, XRD, TPR-H₂, CO₂-TPD, FESEM and TEM.

---

1. Introduction

Dry reforming of methane (eqn (1)) offers valuable environmental benefits such as: biogas utilization,^1,2 removal of greenhouse gases (methane and carbon dioxide) and conversion of natural gas with a high carbon dioxide content to a valuable synthesis gas (syn-gas).^3,4 The lower syn-gas ratio (H₂/CO = 1) produced by dry reforming of methane is considered suitable for the synthesis of oxygenated chemicals⁵ and hydrocarbons from Fischer–Tropsch synthesis.⁶

CH₄ + CO₂ → 2H₂ + 2CO, ΔH^°_298K = 247.1 kJ mol⁻¹ (1)

In dry reforming of methane, non-noble metals (Ni, Co and Fe) are preferred over noble metals (Rh, Ru, Pd and Pt) for dry reforming of methane due to their low cost but are more susceptible to coke formation compared to noble metals.^7–9 The literature on the dry reforming of methane suggests that most common supports applied were MgO,^10–12 Al₂O₃,^12,13 SiO₂,^14,15 CeO₂ (ref. 12 and 16) and ZrO₂.^12,17 Among the employed catalyst supports, MgO has strong Lewis basicity and its application in dry reforming reaction will be beneficial as its basic characteristics will enhance CO₂ chemisorption, as CO₂ is acidic in nature. The chemisorbed species will react with produced carbon to form CO, resulting in the reduction of carbon deposition.^18,19 Thus, MgO appears to be a suitable support, which can reduce or inhibit carbon deposition without adding extra cost²⁰ and also has high thermal stability and low cost.¹⁸

The investigation on the preparation of Ni/MgO catalyst reduced from NiO–MgO solid solution has drawn much attention over the past decade due to its high stability in dry reforming of methane. However, the exhibition of higher stability of NiO–MgO solid solution catalyst was suggested to be dependent on their composition, preparation conditions and also on the morphological properties of the MgO. Previous study suggested that both NiO and MgO can form theoretically proposed ideal solution in any molar ratio, however, there are factors influencing the strength and formation of NiO–MgO solid solution such as; calcination temperature and Ni weight percentage. It is suggested that the preparation history of the NiO–MgO solid solution would have a strong influence on the interaction between Ni metal and MgO support.²¹

Water-in-oil (W/O) microemulsion is considered as a system in which nanosize water droplets (dispersed phase) are present in continuous phase (oil) and stabilized by surfactant molecules. Transparent nature and thermodynamic stability are salient features of microemulsion system, a microenvironment is created by this unique type of surfactant covered water droplets, which plays its role to inhibit the agglomeration of synthesized nanoparticles.^22,23 The advantages of microemulsion synthesis approach over precipitation method, sol–gel process, hydrothermal method are its superior control over the morphology of the nanoparticles prepared^24,25 and also the synthesis of nanoparticles performed at room temperature is a very attractive feature of this process.²² Furthermore, the application of precipitation method for the synthesis of metal oxides suffer from its complexity and also requires longer aging time, sol–gel process uses metal alkoxides as raw materials, which are expensive and also demands long gelation time, while, hydrothermal method requires high temperature and pressure.²⁴ Therefore, in this work, the preparation of Ni/MgO catalyst was performed by microemulsion synthesis approach at room temperature and studied activity of these prepared catalysts in dry reforming of methane reaction. Furthermore, the influence of different weight percentages of Ni (20, 40 and 80 wt%), calcination temperatures (450 °C, 600 °C and 800 °C), reduction temperatures (550 °C and 800 °C) was investigated in order to better understand the formation of NiO–MgO solid solution and how the activity and stability of the Ni/MgO will be affected by these parameters for the catalysts prepared by microemulsion synthesis.

2. Materials and methods

2.1. Materials

Nickel(II) nitrate hexahydrate (Ni(NO₃)₂·6H₂O) and magnesium nitrate hexahydrate (Mg(NO₃)₂·6H₂O) were obtained from Acros Organics. Ammonium hydroxide (28 wt%) and ethanol was purchased from R&M solutions. Non-anionic surfactant Igepal CO-520 (polyoxyethylene(5) nonylphenylether) was acquired from Sigma-Aldrich. Gases used in our study, such as: H₂ (99.999%), CH₄ (99.995%) CO₂ (99.995%) and N₂ (99.99%) were acquired from Linde Malaysia.

2.2. Catalyst preparation

In this study, microemulsion system based on water/Igepal CO-520/cyclohexane was chosen and their ratio (water/surfactant/oil = 6/9/85) was selected from the ternary phase diagram. In the first step, 9 ml of Igepal was added into cyclohexane (85 ml) and left under stirring for 15 min. To prepare Ni/MgO catalyst having different weight percentage of Ni (20 wt%, 40 wt% and 80 wt%), desirable amount of Ni(NO₃)₂ was dissolved in 6 ml of water to prepare 2 M Ni metal solution and then added dropwise to the mixture of surfactant and oil. The solution was put on stirrer for 15 min and then sonicated for 15 min. The pH of the microemulsion was adjusted at 13 with the addition of ammonium hydroxide. Thereafter, the Mg(NO₃)₂ solution was prepared separately, in which desirable amount of magnesium nitrate was dissolved in a suitable amount of water to dissolve all salt and then ammonium hydroxide was added to convert it to magnesium hydroxide. The prepared solution was then added to the microemulsion system containing Ni metal solution and then sonicated for 25 min. The temperature was continuously monitored and controlled at room temperature by using ice cold water. After sonication, the sample was put on stirrer for 2 h at 650 rpm and then ethanol (40 ml) was added as a destabilizer. Later on, centrifugation was done at 4000 rpm for 30 min and then washing was done with ethanol. After washing, another cycle of centrifugation was repeated. The sample was dried for overnight at 100 °C and calcined at 450 °C for 2 h. Ni/MgO catalysts prepared with different weight percentages of Ni (20 wt%, 40 wt% and 80 wt%) are denoted as CS1, CS2 and CS3, respectively. To investigate the influence of calcination temperatures on the supported catalyst (CS3-450 °C), same preparation procedure was repeated and the catalyst was calcined at 600 °C and 800 °C for 2 h, denoted as CS4 and CS5, respectively.

2.3. Characterization techniques

In order to determine the crystallite size of catalysts powder X-ray diffraction (XRD) was done using Rigaku miniflex Cu-Kα radiation (45 kV, 40 mA) and diffractograms were collected in the 2θ range of 5–80°. Specific surface area was assessed from N₂ physisorption data on a liquid Micromeritics ASAP 2020 at −196 °C. To determine the catalyst reduction and desorption behaviour of samples (0.3 g), temperature programmed reduction-H₂ (TPR-H₂) and temperature programmed desorption (CO₂-TPD) experiments were carried out with Micromeritics-2720 (Chemisorb TPx) instrument. Surface morphology of catalysts were studied by using FEI Tecnai™ instrument controlled to obtain transmission electron micrographs. Furthermore, field emission scanning electron microscopy (FESEM) and energy dispersive X-ray (EDX) elemental mapping analysis was carried out in FEG Quanta 450, EDX-OXFORD instrument.

2.4. Experimental setup for catalytic activity test

Activity tests were carried out in a fixed-bed reactor made of stainless steel having 6.03 cm outer diameter, 0.87 cm wall thickens and 120 cm length. To avoid the interaction of feed gas with stainless steel, a quartz tube with 3.56 cm internal diameter, 4 cm outer diameter and 120 cm length was placed inside the reactor obtained from Technical Glass Products (Painesville, USA). For each run, 0.5 g of catalyst was immobilized between two quartz wool plugs to serve as catalyst bed. Reactant gases (methane and carbon dioxide) were fed into the reactor at a total flow rate of 1.4 l min⁻¹ (CH₄ : CO₂ = 1 : 1) and gas hourly space velocity (GHSV) of 1.68 × 10⁵ ml g⁻¹ h⁻¹ or 168 l h⁻¹ g⁻¹. Before activity tests, the reduction of catalysts was done with 30% H₂/N₂ at 550 °C for 2 h. Thereafter, the activity tests were conducted at different temperatures ranging from 550 °C to 850 °C. The stability of the catalysts was investigated at three different reaction temperatures (750 °C, 800 °C and 850 °C). To investigate the influence of reduction temperatures, the catalysts were reduced at two different temperatures (550 and 800 °C) at similar reduction conditions. Mass flow controllers (0–2 l min⁻¹) were used to control the flow of feed gas purchased from Dwyer, USA. Rosemount Analytical X-STREAM on-line analyzer was used to measure the mole percentage of unreacted methane and H₂ produced, while Guardian NG gas monitor supplied by Edinburgh Sensors (UK) was used to measure the composition of carbon dioxide and carbon monoxide at the outlet. The conversion and yields of both reactants and products are calculated as follows:

3. Results and discussion

3.1. Characterization of fresh catalysts

X-ray diffraction (XRD) patterns of NiO, MgO and NiO/MgO catalyst gave insight into the crystallite size as shown in Fig. 1. Pure NiO exhibits sharp peaks at 2θ values of 37.28°, 43.44°, 63.01°, 75.55° and 79.40° corresponding to the respective crystallite phase of (111), (200), (220), (311), and (222), which matches well with standard card of cubic NiO with JCPDS no. 01-073-1519. Furthermore, pure MgO and NiO/MgO catalyst matched well with the standard card of cubic JCPDS no. 01-079-0612 and JCPDS no. 00-024-0712, respectively. The presence of diffraction lines at 2θ values of 62.37°, 74.80° and 78.54° are attributed to the formation of NiO–MgO solid solution and in the present study for all the catalysts (CS1 to CS5), these diffraction line are present indicating the formation of NiO–MgO solid solution. Another study also reported the formation of NiO–MgO solid solution at similar 2θ values (62.32°, 74.72° and 78.66°).²⁶ However, the close similarity of NiO and MgO in their oxides structure both having face-centered-cubic structures and also having similar lattice parameters (4.1946 and 4.2112 Å) makes it quite difficult to distinguish between the diffraction peaks of NiO, MgO and NiO/MgO.²⁷ Furthermore, Fig. 1 clearly indicates that with the increase of Ni content (20% to 80%) diffraction peaks become narrow indicating the increase of crystallite size (Table 1). The broad diffraction peaks for CS1 indicates the presence of small crystallite size, which matches well with the BET results for CS1 exhibiting larger surface area (Table 1).

Fig. 1 XRD patterns of calcined catalysts, where (), (), and () presents NiO–MgO solid solution peaks, NiO crystallite peaks and MgO crystallite peaks, respectively.

Table 1 BET surface area, total pore volume and XRD crystallite size of calcined catalysts

Catalyst	BET SA m^2 g^−1	Pore volume cm^3 g^−1	BET pore size (nm)	Average crystallite size (nm)	Average Ni^o crystallite size (nm)^d
					Spent catalyst reduced at 550 °C	Spent catalyst reduced at 800 °C
a Based on all NiO peaks in JCPDS no. 01-073-1519.b Based on all MgO peaks in JCPDS no. 01-079-0612.c Based on all NiO–MgO peaks in JCPDS no. 00-024-0712.d Based on Ni^o peaks located at 44.48° and 51.83°.
NiO	23.87	0.1215	20.36	27.58^a	—	—
MgO	34.69	0.0836	9.643	25.69^b	—	—
CS1	153.22	0.3932	10.26	18.81^c	50.82	—
CS2	125.82	0.4401	13.92	26.69	56.71	—
CS3	54.01	0.2121	15.70	27.09	30.98	32.63
CS4	29.11	0.1634	22.46	28.56	44.07	47.63
CS5	15.19	0.0794	20.91	33.90	27.22	32.54

---

BET surface area and pore size distribution of the catalysts were measured by N₂-physisoorption. The isotherms of nitrogen adsorption and desorption for these catalysts were found to be type III isotherm with a large type H3 hysteresis loop according to the IUPAC classification. The formation of type H3 hysteresis loop at relatively high pressure indicates the formation of large pore volumes in the samples. N₂ adsorption/desorption isotherms for NiO, MgO, CS1 and CS2 are depicted in Fig. 2(a). N₂ isotherms for catalysts (CS3, CS4 and CS5) having higher Ni weight percentage (80%) and calcined at different temperatures are exhibited in Fig. 3(a). The surface area of pure NiO nanoparticles and MgO was 23.87 m² g⁻¹ and 34.69 m² g⁻¹, respectively. An increase in surface area was observed for the supported catalysts (CS1, CS2, CS3) calcined at 450 °C with the addition of MgO support as depicted in Fig. 1(a), 3(a) and Table 1. However, with the increase of Ni content (CS1 to CS3), there was seen a decrease in surface area from 153.22 to 54.01 m² g⁻¹. The decrease of surface area with the increase of Ni content can be attributed to the deposition of Ni onto smaller pore of support. A further decrease in the surface area from 54.01 to 15.19 m² g⁻¹ was observed with the increase of calcination temperature from 450 °C (CS3) to 800 °C (CS5), respectively as depicted in Table 1 and Fig. 3(a). Similar type of results was reported by Feng et al.,¹⁰ for impregnated NiO/MgO catalysts, when the calcination temperature was increased from 600 °C to 800 °C, surface area decreased from 38.0 m² g⁻¹ to 27.4 m² g⁻¹, respectively. Pore size distribution of catalysts was determined by Barret–Joyner–Halenda (BJH) method based on adsorption branch of N₂ isotherm as shown in Fig. 1(b) and 2(b). Furthermore, BJH average pore width sizes for NiO, MgO, CS1, CS2, CS3, CS4 and CS5 were calculated to be around 22.56 nm, 12.41 nm, 15.06 nm, 18.04 nm, 15.34 nm, 27.04 nm and 30.85 nm, respectively.

Fig. 2 (a) N₂ adsorption–desorption isotherms and (b) BJH pore width distribution of calcined catalysts.

Fig. 3 (a) N₂ adsorption–desorption isotherms and (b) BJH pore width distribution of calcined catalysts.

Surface reducibility of the catalysts was examined by TPR-H₂, which has been recognized as a technique to discriminate various species in solid solutions. TPR profiles of the pure NiO and Ni/MgO catalysts are exhibited in Fig. 4. The higher reducibility of pure NiO exhibited by higher H₂ uptake refers to the reduction of bulk NiO as shown in Fig. 4(a). TPR profiles of supported catalysts (CS1, CS2 and CS3) exhibited two peaks one at very low temperature (T_L) and second peak at quite higher temperature (T_H) except for CS1 for which the second peak also appear at medium temperature (T_M) around 365.1 °C. The presence of temperature peaks (<400 °C) can either be attributed to the reduction of Ni³⁺ surface species located at surface sites for Ni/MgO catalyst or to the reduction of NiO which was uninfluenced by the MgO support. The presence of higher reduction peaks temperature in the range of 500–700 °C for supported catalysts indicates the reduction of Ni²⁺ ions in the outermost and sub-surface layers of the MgO lattice, which is in accordance with the literature cited.^10,21,28

Fig. 4 TPR-H₂ profiles of (a) NiO and (b) Ni/MgO catalysts.

Previous studies suggested that calcination temperature has a strong influence on the incorporation of NiO into NiO–MgO solid solution.^10,21 This is the reason that for CS4 and CS5, application of higher calcination temperature (600 °C and 800 °C, respectively) shifts the reduction peak temperature from 698.8 °C (CS3-450 °C) to 875.5 °C and 884.1 °C, respectively (Fig. 4(b)). The shift of reduction peak to higher temperature (>800 °C) indicates the reduction of lattice Ni²⁺ ions in the MgO matrix, which indicates the formation of strong NiO–MgO solid solution. Furthermore, the increase of calcination temperature from 600 °C to 800 °C, shifted the medium reduction peak at 555.9 °C to 656.9 °C, which indicates the presence of strong metal–support interaction for CS5. The detail of the reduction peaks temperature along with their reducibility extent are mentioned in Table 2. Similar type of shift in reduction peaks to higher temperature was reported for Ni/MgO catalyst by Feng et al.,¹⁰ as the calcination temperature was increased from 600 °C to 800 °C. Fig. 4(a) and (b) also showed that the addition of MgO leads to the shift of reduction peaks to higher temperature and also the interaction between NiO and MgO probably hindered the reduction of NiO owing to the formation of NiO–MgO solid solution.

Table 2 Reduction peaks temperatures and reducibility of Ni/MgO catalysts

Catalyst	TL (°C)	TM (°C)	TH (°C)	Reducibility (%)
NiO	—	347.2		—
CS1	122.1	365.1		11.26
CS2	169.5	—	611.3	20.96
CS3	190.8	—	698.8	31.42
CS4	230.5	555.9	875.5	46.55
CS5	—	656.9	884.1	46.97

---

The application of basic support (MgO) will have a strong influence on the strength of basic sites in Ni/MgO catalyst. Temperature programmed desorption of CO₂ (CO₂-TPD) was performed in order to investigate the strength of basic sites in Ni/MgO catalyst. The temperature at which chemisorbed CO₂ on the basic sites is desorbed, indicates the strength of basic sites. Weak basic sites are said to be formed in Ni/MgO catalyst, when CO₂ desorption peaks occur at lower temperature, whereas strong basic sites exhibit an opposite trend.²⁹ CO₂-TPD profiles for the catalysts having higher MgO content (CS1 and CS2) exhibited higher peak temperature at 653.4 °C and 729.2 °C, respectively, which indicates the presence of strong basic active sites. The lower peak temperature for CS1 at 252.4 °C indicates the presence of weak basic sites. CO₂-TPD profiles for the catalyst having lower MgO content (CS3) exhibited the peak temperature at 367.1 °C, which indicates the presence of medium strength basic sites. The investigation on the influence of calcination temperature over the strength of basic sites indicated that CO₂ desorption peaks were shifted to the higher temperature as the calcination temperature increased from 450 °C to 600 °C and 800 °C for CS4 and CS5, respectively. For CS4, the peak temperature shifted from 367.3 °C to 834.6 °C as depicted in Fig. 5, whilst for CS5 the peak temperature shifted to further higher temperature (862.0 °C). CO₂-TPD profiles for CS4 and CS5 indicates the existence of very strong basic sites, which indicates the interaction of metal sites with basic sites.

Fig. 5 CO₂-TPD profiles for Ni/MgO catalysts.

TEM images of Ni/MgO catalyst having higher MgO content (CS1, Fig. 6(a) and (b)) and lower MgO content (CS3, Fig. 6(c) and (d)) exhibited that the addition of MgO in the microemulsion system leads to the better nanoparticle distribution, however, the influence of the addition of MgO to resist agglomeration was more pronounced in CS1 compared to CS3, as depicted by its higher surface area in BET results (Table 1). Furthermore, EDX analysis of NiO, CS1, CS2 and CS3 indicates the weight percentages of Ni and Mg before the reduction are shown in Fig. 7.

Fig. 6 TEM images of freshly calcined catalysts (a, b) CS1 and (c, d) CS3.

Fig. 7 EDX analysis of (a) NiO, (b) CS1, (c) CS2 and (d) CS3 catalysts.

3.2. Activity and stability test

The study of catalytic activity of catalysts (CS1, CS2, CS3, Ni) indicates that there was seen an increase in CH₄ and CO₂ conversions with the increase of temperature, which was attributed to the thermodynamic nature of the dry reforming reaction. Fig. 8(a) and (b) indicates that the catalyst having higher Ni content (CS3) exhibited higher catalytic activity among all the catalysts tested (CS1, CS2 and Ni). However, both CS1, CS2 exhibited comparable activity, while CS2 being slightly higher. Ni nanoparticles exhibited the lower methane (18.72%) and carbon dioxide (21.80%) conversion at final temperature 850 °C.

Fig. 8 Conversion–temperature relationship of different catalysts (reaction conditions: CH₄/CO₂ = 1/1 and GHSV = 1.68 × 10⁵ ml g⁻¹ h⁻¹).

The stability test of different catalysts (CS1, CS2 and CS3) at different reaction temperatures (700 °C, 800 °C and 850 °C) exhibited that CS3 has the higher catalytic activity and stability at 750 °C compared to CS1 and CS2 as shown in Fig. 9(a) and (b) and Table 3. The stability study of catalysts at higher reaction temperatures (800 °C and 850 °C), exhibited a severe decrease in the catalytic activity. However, with the rise of reaction temperature, a significant decrease in carbon deposition was observed as mentioned in Table 3. This can be attributed to the presence of more reactive carbon species produced by methane decomposition at these higher reaction temperatures, which are easily gasified by CO₂.³⁰ The catalytic activity of both CS1 and CS2 were comparable to each other with respect to reaction temperature (Fig. 8(a) and (b)), however, the stability of both catalysts exhibited a different trend as mentioned in Table 3. CS1 exhibited higher initial catalytic activity than CS2 at 750 °C, however, CS2 exhibited higher initial and final conversions at higher reaction temperatures (800 and 850 °C). Ni nanoparticles exhibited severe decrease in catalytic activity within 20 min of reaction period as shown in Fig. 9(a) and (b), which was attributed the agglomeration of Ni particles.

Fig. 9 Conversion–time relationship of different catalysts (reaction conditions 750 °C CH₄/CO₂ = 1/1, GHSV = 1.68 × 10⁵ ml g⁻¹ h⁻¹).

Table 3 Activity and stability of Ni/MgO catalysts at CH₄/CO₂ = 1/1 and GHSV = 1.68 × 10⁵ ml g⁻¹ h⁻¹

Sample ID	Tc (°C)	Tred (°C)	Treac (°C)	Initial % conv.		Final % conv. (after 140 min)		Carbon (gc gcat^−1)
				CH4	CO2	CH4	CO2
a Initial and final conversions of pure Ni after 20 min; Tc: calcination temperature; Tred: reduction temperature, Treac: reaction temperature.
CS1	450	550	750	46.13	51.40	34.30	37.60	2.648
			800	45.43	50.80	24.83	22.40	0.820
			850	46.44	55.20	27.98	27.00	0.122
CS2	450	550	750	35.87	41.80	34.90	37.20	3.700
			800	53.05	60.40	37.40	44.60	2.892
			850	63.72	70.20	42.16	48.60	0.806
CS3	450	550	750	49.76	63.80	44.21	56.40	6.684
			800	62.09	69.00	44.01	49.80	3.380
			850	73.84	79.40	43.89	53.20	2.680
		800	750	45.03	44.60	34.72	36.80	3.760
CS4	600	550	750	47.29	54.00	46.37	51.40	4.782
		800		62.48	64.80	38.82	46.00	3.884
CS5	800	550	750	59.22	65.60	52.09	59.80	3.580
		800		37.84	38.00	28.46	31.40	3.030
Ni^a	450	550	750	15.06	14.60	7.69	1.80	0.016

---

The study exhibited the fact that Ni content in Ni/MgO catalyst has an important role in catalytic activity and stability of catalysts. TPR-H₂ results suggested that with the increase of Ni content (20 wt% to 80 wt%) an increase in hydrogen consumption was observed, which indicates the reduction of more NiO to Ni^o. The higher Ni content in Ni/MgO catalyst will produce more Ni active sites upon reduction, which would be easily accessible by the reactants and thus exhibited higher catalytic activity as shown in Table 3. This can be justified on the basis of XRD analysis of spent catalysts (Fig. 12), which indicates the presence of more intense peak related to Ni^o for CS3. Thus, it can be concluded that the activity of the catalysts was related to the quantity of Ni^o present in the catalyst, which was obviously higher for the catalyst having higher Ni content. The surface area was decreased from 153.22 to 54.01 m² g⁻¹ with the increase of Ni content from 20% to 80%, however, the methane and carbon dioxide conversion were increased from 46.13% to 49.76% and from 51.40% to 63.80%, respectively at 750 °C. Similar type of results was reported by Therdthianwong et al.,³¹ for Ni/Al₂O₃ catalyst, when the Ni content was increase from 5% to 25%, the BET surface was decreased from 212.4 to 164.8 m² g⁻¹, respectively. However, both methane and carbon dioxide conversions were seen to increase with Ni content. For 5% Ni/Al₂O₃, CH₄ and CO₂ conversions were only 7.0% and 19.5% and increased to 63.4% and 66.9%, respectively for 25% Ni/Al₂O₃ at 700 °C.

The carbon deposition on CS3 was comparatively high (6.684 g_c g_cat⁻¹) compared to CS1 (2.648 g_c g_cat⁻¹) and CS2 (3.700 g_c g_cat⁻¹). However, when one correlates the amount of carbon deposition and catalyst stability, it can be found that larger amount of carbon deposition does not always lead to the sever decrease in activity. A closer look at Fig. 9(a) and (b) and Table 3, indicates even though carbon deposition was high for CS3 but it exhibited better catalyst stability. Similar type of results was reported by Liu et al.,³² for Ni catalysts supported on mesoporous material MCM-41 having different (gel Si/Ni ratio). They reported that Ni-MCM-41(B) and Ni-MCM-41(C) show poor stability, even though they have lower carbon deposition compared to Ni-MCM-41(D), which exhibited highest catalyst stability. The higher carbon deposition on CS3 can be attributed to the larger particle size, which leads to the production of higher carbon deposition.¹⁸

The influence of calcination temperature over catalyst stability was studied at 750 °C, by calcining the catalyst CS3-450 °C at higher calcination temperatures, 600 °C (CS4) and 800 °C (CS5). The study showed that CS5 exhibited the higher stability and both initial (59.22% CH₄ and 65.60% CO₂) and final (52.09% CH₄ and 59.80% CO₂) conversions were high compared to CS3 and CS4. Even though, with the increase of calcination temperature (450 to 800 °C), there was seen a further decrease in BET surface area from 54.01 to 15.19 m² g⁻¹. The stability of the catalyst was uninfluenced with the decrease of BET surface area (larger particle size), which showed that the catalyst particle size did not play major role in the catalytic activity and stability. Furthermore, the carbon deposition was low for CS5 (3.580 g_c g_cat⁻¹) compared to CS3 (6.684 g_c g_cat⁻¹) and CS4 (4.782 g_c g_cat⁻¹), which can be attributed to the strong NiO–MgO solid solution. BET analysis indicates that CS5 has the largest particle size among all the catalyst test, still it exhibited the better stability and activity. Similar type of results was reported by Horváth et al.,³³ in which active metal Ni, NiCo or NiRh supported on CeZr-mixed oxide exhibited stable activity for the samples containing larger particles size and suggested that larger particle size was also responsible for long term stability.

It has been suggested that reduction temperature plays an important role in catalytic activity and stability of the catalyst. The investigation on the influence of reduction temperature was done in order to optimize the activation conditions. Therefore, the influence of reduction temperatures (550 and 800 °C) on the catalytic stability of catalysts (CS3, CS4 and CS5) was investigated and the results are shown in Fig. 9(c) and (d). The study shows that the catalysts reduced at higher reduction temperature (800 °C) exhibited severe catalyst deactivation. The results indicated that the catalyst (CS5) calcined at higher temperature (800 °C) and reduced at lower temperature (550 °C) exhibited the better stability and activity of the catalyst compared to CS3 and CS4 as mentioned in Fig. 9(c) and (d) and Table 3. The catalysts (CS3 and CS5) reduced at higher temperature (800 °C), even lead to the lower initial reactants conversions, while, CS4 exhibited higher initial catalytic activity at 800 °C compared to the same catalysts reduced at 550 °C (Table 3). Takanabe et al.,³⁴ reported the influence of reduction temperature (700 °C, 750 °C, 800 °C and 850 °C, 900 °C and 950 °C) on the catalytic behaviour of Co/TiO₂ catalyst applied for dry reforming of methane. The study showed that the catalyst Co/TiO₂ exhibited higher methane (65.5%) and carbon dioxide (71.7%) conversion at lower reduction temperature (700 °C). Both methane and carbon dioxide conversions were decreased with the increase of reduction temperature such as: that at 750 °C it exhibited 55.7% CH₄ and 64.3% CO₂ conversion, at 800 °C exhibited 39.0% CH₄ and 51.2% CO₂ conversion, at 850 °C exhibited 5.6% CH₄ and 9.2% CO₂ conversion. Furthermore, at higher reduction temperatures of 900 °C and 950 °C, methane and carbon dioxide conversion become negligible. They suggested that catalyst deactivation of Co/TiO₂ catalysts with the increase of reduction temperature was attributed to the metal sintering. The study of the influence of reduction temperature (400 °C, 500 °C, 600 °C and 700 °C and 900 °C) on cobalt catalyst supported on SiO₂ applied for Fischer–Tropsch catalysis exhibited that the catalysts reduced at higher reduction temperature were more prone to sintering and exhibited lower catalyst performance compared to the catalyst reduced at lower reduction temperature.³⁵ Similar results were reported by Liu et al.,³² for Ni-MCM-41 in which the increase of reduction temperature from 550 to 750 °C leads to the decrease in methane conversion from 80 to 73% and carbon dioxide from 83% to 74%. The study leads to the conclusion that the reduction of catalysts at higher reduction temperature makes the catalysts more prone to sintering, which in turn reduces the number of available active sites and thus leads to the lower catalytic activities and deactivation. Therefore, it seems that the proper choice of reduction temperature is critical and should be investigated properly for each of the preparation method.

Carbon formation during the dry reforming reaction (eqn (1)) is suggested to be produced by two major reactions: methane decomposition (eqn (2)) and Boudouard reaction (eqn (3)) being endothermic and exothermic, respectively.^28,36,37

CH₄ → C + 2H₂, ΔH^°_298K = 74.6 kJ mol⁻¹ (2)

2CO → CO₂ + C, ΔH^°_298K = −172.46 kJ mol⁻¹ (3)

Thermodynamic analysis of dry reforming reaction suggests that this reaction becomes spontaneous at higher reaction temperature (>640 °C).^38,39 Furthermore, previous studies suggested that methane decomposition is favourable at higher reaction temperatures (>700 °C) due to its endothermic nature,^28,40 while, Boudouard reaction is not thermodynamically favourable at higher reaction temperatures (>700 °C) and above 700 °C becomes non-spontaneous.^36,38,40 Therefore, the application of higher reaction temperature not only eliminates the influence of Boudouard reaction³⁹ but also will shift the equilibrium to the left side and will favour the occurrence of reverse of Boudouard reaction as shown in eqn (4).⁴¹

CO₂ + C → 2CO, ΔH^°_298K = 172.46 kJ mol⁻¹ (4)

Furthermore, the spontaneous reaction temperature for reverse Boudouard reaction is 719 °C.⁹ Therefore, dry reforming of methane at higher reaction temperature (750 °C) can be regarded as a combination of one carbon formation reaction (methane decomposition) and one carbon elimination reaction (reverse Boudouard reaction).³⁶ The occurrence of reverse Boudouard reaction will assist in the removal of carbon produced by reacting with chemisorbed CO₂.^36,42 However, the occurrence of carbon deposition at 750 °C (Table 3) during this study indicates the existence of following conditions such as: the rate of methane decomposition (eqn (2)) was high compared to carbon removal reaction (eqn (4)) and reverse Boudouard reaction seems to be the limiting step as suggested by various studies.^27,36

Higher CO₂ conversion was exhibited by Ni/MgO catalyst compared to methane during the activity and stability tests. Furthermore, the presence of water at outlet indicates the occurrence of reverse water gas shift reaction (RWGS) as shown in eqn (5).^20,42 The study also indicated that during the reaction time CO yield was always higher than H₂ yield for all the catalysts as shown in Fig. 10, which is attributed to the RWGS reaction. The utilization of produced H₂ in the RWGS reaction by CO₂ leads to the production of higher CO and in turn higher CO₂ conversions are observed.

CO₂ + H₂ → CO + H₂O, ΔH^°_298K = 41.0 kJ mol⁻¹ (5)

Fig. 10 H₂ and CO yield of different catalysts (reaction conditions 750 °C CH₄/CO₂ = 1/1, GHSV = 1.68 × 10⁵ ml g⁻¹ h⁻¹).

Thus, the study leads to the conclusion that the better catalytic activity and stability of the Ni/MgO catalyst was exhibited by the formation of strong “NiO rich” solid solution and obviously will be present in the catalyst having higher Ni content. For the catalysts having higher MgO content (CS1 and CS2), NiO would diffuse from the outermost layer into a deeper layer to form a more stable type of “MgO rich” solid solution. However, for catalytic reactions, a surface “NiO rich” solid solution will be more beneficial as it will produce more Ni^o active sites compared to the “MgO rich” solid solution.⁴³ That is the reason, that catalysts having higher Ni content (CS3, CS4 and CS5) exhibited higher catalytic activity, which are having “NiO rich” solid solution. Furthermore, with the increase of Ni content an increase in the reducibility of NiO in calcined NiO/MgO was observed (Fig. 4). Hu and Ruckenstein¹⁸ suggested that the occurrence of such type of phenomenon would lead to the formation of large Ni particle size in Ni/MgO catalyst similar to that observed in this study by BET results (Table 1), which will be eventually more prone to the sintering and coking at high Ni loading. Hence, larger Ni particles in Ni/MgO catalyst would lead to severe deactivation of the catalytic activity during the reaction time.¹⁸ Furthermore, the regenerability of the CS5 catalysts reduced at 550 °C was studied in the presence of air for 1 h and then the activity of the regenerated catalyst was studied under the same condition as used in the stability studies (750 °C). However, the catalyst did not exhibit excellent regenerability and deactivated very rapidly (within 25 min) as shown in Fig. 11.

Fig. 11 Stability and regenerability study of CS5 catalyst.

However, according to the present data for activity (Table 3 and Fig. 8 and 9(a) and (b)) and surface area of different catalysts (Table 1), particle size of Ni was not having a major influence over the stability of the reduced Ni/MgO catalyst. Similar type of results was reported by Wang et al.,²¹ for Ni/MgO catalyst in which the larger Ni particle size (20.0 nm) was observed on reduced Ni30Mg-6 compared to the smaller particle sizes (16.7 and 10.0 nm) observed on reduced Ni8Mg-4 and Ni8Mg-5, respectively. However, Ni30Mg-6 exhibited very stable activity, but both Ni8Mg-4 and Ni8Mg-5 catalysts were deactivated rapidly during the reaction time. These results lead to an important conclusion that not only the Ni loading and Ni particle size in the reduced catalyst, but also physicochemical state of NiO in the oxidized sample plays a major role in affecting the stability of the catalyst. Previous study suggested that for Ni/MgO catalyst it is important to have NiO entities that belong to NiO–MgO solid solution, otherwise it would not be possible to have stable Ni sites.²¹ Therefore, the key to form stable Ni sites in Ni/MgO catalyst, it is necessary to have a complete reaction of the NiO component with the MgO support during the calcination step to form NiO–MgO solid solution. Therefore, in this study the catalyst calcined at higher temperature (800 °C) exhibited better stability due to the formation of strong NiO–MgO solid solution compared to the catalysts calcined at a lower calcination temperature (CS3-450 °C and CS4-600 °C) having similar Ni content. Previous study suggested that calcination temperature has significant influence on the diffusion of Ni²⁺ ions into the MgO lattice and strong metal–support interaction was observed.¹⁰

The comparison of different catalysts applied to dry reforming of methane with the present study indicates that Ni/MgO catalyst prepared by microemulsion synthesis exhibited better performance compared to the previous studies^{20,26,44–47} as mentioned in Table 4. Previous studies reported that the catalysts at low GHSVs exhibited higher reactants (CH₄ and CO₂) conversion and opposite trend was observed at high GHSVs. The lower reactants conversion at higher GHSV values, was attributed to the fact that the residence or contact time will be lower, moreover, larger amounts of reactants will be flowing into the reactor and reactants will have limited opportunity to adsorb on active sites.^48–51 However, the comparison of Ni/MgO catalyst (present work) with previous studies exhibited that even though the GHSV was quite higher (1.68 × 10⁵ ml h⁻¹ g⁻¹), it exhibited higher reactants conversion indicating its better performance even under severe reaction conditions. Moreover, the study suggests that the preparation conditions, Ni metal content, calcination temperatures and reduction temperatures has much strong influence over this unique system of NiO–MgO solid solution prepared by microemulsion synthesis. Therefore, we can conclude that the influence of MgO addition was prominent not only on the activity of Ni/MgO catalyst, but also provide resistance towards the agglomeration of Ni particles.

Table 4 Comparison of catalytic activity of previous studies with Ni/MgO catalysts^a

Catalyst	Reaction conditions		Final conversion (%)			Ref.
	T	GHSV	t	CH4	CO2
a T: temperature °C; GHSV: gas hour space velocity ml h^−1 g^−1; t: min; PT: plasma treated catalyst; C: conventional impregnated catalyst.
Ni0.10Mg0.90O	600	1.4 × 10^4	300	25.0	30.0	20
Ni-MCM-41	600	3.6 × 10^4	240	28.0	38.0	44
Rh-MCM-41-V				32.0	39.0
5%Ni/ZrO2	700	1.5 × 10^4	300	54.0	59.0	45
PT-Ni/MgO	700	9.6 × 10^4	240	49.0	54.0	26
C-Ni/MgO				20.0	30.0
NiO/Al2O3	800	9.37 × 10^4	300	52.5	—	46
3 mol% Ni/MgO	850	5.6 × 10^4	240	50.0	62.0	47
3 mol% Pt/MgO				40.0	50.0
Ni/MgO (CS5)	750	1.68 × 10^5	140	52.09	59.80	This study
Ni/MgO (CS4)				46.37	51.40
Ni/MgO (CS3)				44.21	56.40

---

3.3. Characterization of spent catalyst

The characterization of spent catalysts (CS1–CS5) by XRD analysis indicates the presence of graphite peak (002) around 26°, which was attributed to the crystalline phase of CNTs matching with JCPDS no. 01-075-1621 (Fig. 12). Furthermore, XRD analysis of all the spent catalysts exhibited Ni^o peaks at 44.48° and 51.83°, which matched with JCPDS no. 01-070-1849. XRD analysis of spent catalyst (CS3) indicates the presence of broad and strong Ni^o peaks compared to CS1 and CS2, which suggests that CS3 have more Ni^o active sites at these reaction conditions and this can be attributed to the higher Ni content. Therefore, the catalyst having higher Ni content (CS3) exhibited higher catalytic activity compared to the catalysts having lower Ni content (CS1 and CS2). XRD analysis of spent catalysts reduced at higher reduction temperature (800 °C) are also shown in Fig. 12 and the average crystallite sizes of spent catalysts (reduced at 550 °C and 800 °C) are mentioned in Table 1. An increase in the crystallite sizes was observed for the catalysts reduced at 800 °C compared to the catalysts reduced at lower reduction temperature (550 °C) indicating the occurrence of Ni metal sintering, which matches well with the previous studies that the higher reduction temperature makes the catalyst prone to sintering.^34,35

Fig. 12 XRD analysis of spent catalysts, where (), () and () presents NiO–MgO solid solution peaks, Ni crystallite peaks and graphite peaks, respectively.

Carbon accumulation on spent catalysts can be observed by TEM and FESEM images, indicating the presence of carbon species having different morphologies. The analysis of spent catalysts (CS1, CS2, CS3 and CS5) indicates that the majority of the accumulated carbon were in the form of carbon nanotubes (whisker-like carbon species). Moreover, TEM and FESEM images (Fig. 13 and 14) indicate the presence of different types of carbon nanotubes (CNTs) such as: CNTs with Ni particle at the tip, CNTs with closed end but without the presence of Ni particle on the tip, CNTs with different diameters and CNTs with hollow internal channel having open end structure and no Ni particles at the tip. Previous studies suggested that the formation of single wall and multiwall carbon nanotubes either follow tip-growth or base-growth mechanism.^52,53 Moreover, the intensity of metal–support interaction plays a decisive role in the formation of carbon nanotubes mechanism either by tip-growth or base-growth mode. Tip-growth mode will be more pronounced for catalysts having weak metal–support interaction and metal particles will be lifted up by the growing carbon nanotubes. On the other hands, for the catalysts having strong metal–support interaction, base-growth mode will be more pronounced in the formation of carbon nanotubes.⁵⁴ Therefore, CNTs with hollow internal channel having open end structure and no Ni particles at the tip are more prominent for CS5 (Fig. 14(c) and (d)), which can be attributed to its strong metal–support interaction as discussed in TPR-H₂ results. Thus, indicates that base-growth mechanism was more pronounced for CS5 instead of tip growth model. Furthermore, previous studies suggested that the cracking of hydrocarbon on Ni based catalysts will lead to the production of filamentous (whisker-like) and encapsulating (shell-like) carbon species. However, it was reported that in terms of degree of toxicity, the former is considered less toxic from the point of view of deactivation process,^55–57 while the latter has serious influence on the deactivation of the catalysts by decreasing the total number of active sites. In the present study, all of the spent catalysts indicate the presence of filamentous type of carbon rather than encapsulating carbon species and previous studies suggested that this type of carbon does not cause major deactivation.^44,57–60 This conclusion matches well with the present study that the catalysts indicating the presence of whisker-like carbon species does not exhibited deactivation and were quite stable during the reaction period. Furthermore, severe deactivation exhibited by pure Ni nanoparticles can be attributed to the existence of strong agglomeration of Ni particles as described in the FESEM images (Fig. 14(e) and (f)).

Fig. 13 TEM and FESEM images of spent catalysts (a, b) CS1 and (c, d) CS2 after the dry reforming reaction at 750 °C.

Fig. 14 TEM images of spent catalysts (a, b) CS3, (c, d) CS5 and FESEM images of Ni (e, f) after the dry reforming reaction at 750 °C.

4. Conclusion

The catalytic performance of Ni/MgO catalyst in dry reforming of methane and the factors influencing the catalytic activity were investigated in this study. The major conclusion draws from this study are:

(1) Ni/MgO catalyst (CS3) having higher Ni content (80%) exhibited better catalytic activity compared to the catalyst prepared by 20% (CS1) and 40% (CS2) Ni content, even though it has higher carbon deposition. This can be attributed to the higher number of Ni^o active sites crucial for achieving higher reactants conversion. Even though BET results indicated that CS3 has the lower surface area (larger particle size) compared to CS1 and CS2 but this did not influence the activity and stability of the catalyst.

(2) The increase of calcination temperature from 450 °C to 800 °C for the catalyst having higher Ni content (80%) not only enhanced metal–support interaction but also lower carbon deposition was observed compared to the catalyst having similar Ni content but calcined at lower calcination temperature (450 °C). The increase of catalytic activity for CS5 can be attributed to the increased number of Ni²⁺ ions having strong interaction with MgO support and also better resistance towards sintering.

(3) The reduction of catalyst at higher reduction temperature (800 °C) exhibited severe catalyst deactivation and this can be attributed to the more susceptibility of catalyst towards sintering.

To sum up all the results, the key to form stable Ni/MgO catalyst prepared by microemulsion synthesis is the application of higher calcination temperature and lower reduction temperature.

Acknowledgements

The authors gratefully acknowledge the financial support from GSP-MOHE (MO008-2015), University of Malaya, Malaysia. We also sincerely acknowledge the technical assistance from Cik Norhaya Binti Abdul Rahim and Cik Fazizah Binti Abdullah, Department of Chemical Engineering, University of Malaya for BET and TPR-H₂ analysis.

References

1. A. F. Lucredio, J. M. Assaf and E. M. Assaf, Fuel Process. Technol., 2012, 102, 124–131.
2. M. P. Kohn, M. J. Castaldi and R. J. Farrauto, Appl. Catal., B, 2014, 144, 353–361.
3. M. C. J. Bradford and M. A. Vannice, Catal. Rev., 1999, 41, 1–42.
4. J. H. Lunsford, Catal. Today, 2000, 63, 165–174.
5. T. Wurzel, S. Malcus and L. Mleczko, Chem. Eng. Sci., 2000, 55, 3955–3966.
6. S. T. Oyama, P. Hacarlioglu, Y. Gu and D. Lee, Int. J. Hydrogen Energy, 2012, 37, 10444–10450.
7. Y. J. O. Asencios and E. M. Assaf, Fuel Process. Technol., 2013, 106, 247–252.
8. D. Liu, X. Y. Quek, W. N. E. Cheo, R. Lau, A. Borgna and Y. Yang, J. Catal., 2009, 266, 380–390.
9. L. Xu, Y. Liu, Y. Li, Z. Lin, X. Ma, Y. Zhang, M. D. Argyle and M. Fan, Appl. Catal., A, 2014, 469, 387–397.
10. J. Feng, Y. Ding, Y. Guo, X. Li and W. Li, Fuel, 2012, 109, 110–115.
11. M. Yang, H. Guo, Y. Li and Q. Dang, J. Nat. Gas Chem., 2012, 21, 76–82.
12. H. Y. Wang and E. Ruckenstein, Appl. Catal., A, 2000, 204, 143–152.
13. Z. Hao, Q. Zhu, Z. Jiang, B. Hou and H. Li, Fuel Process. Technol., 2009, 90, 113–121.
14. A. Effendi, K. Hellgardt, Z. G. Zhang and T. Yoshida, Catal. Commun., 2003, 4, 203–207.
15. Q. Jing, H. Lou, J. Fei, Z. Hou and X. Zheng, Int. J. Hydrogen Energy, 2004, 29, 1245–1251.
16. M. M. Barroso-Quiroga and A. E. Castro-Luna, Int. J. Hydrogen Energy, 2010, 35, 6052–6056.
17. J.-S. Chang, D.-Y. Hong, X. Li and S.-E. Park, Catal. Today, 2006, 115, 186–190.
18. Y. H. Hu and E. Ruckenstein, Catal. Rev., 2002, 44, 423–453.
19. H. Özdemir, F. Öksüzömer and M. Ali Gürkaynak, Int. J. Hydrogen Energy, 2010, 35, 12147–12160.
20. R. Zanganeh, M. Rezaei and A. Zamaniyan, Int. J. Hydrogen Energy, 2013, 38, 3012–3018.
21. Y.-H. Wang, H.-M. Liu and B.-Q. Xu, J. Mol. Catal. A: Chem., 2009, 299, 44–52.
22. S. Eriksson, U. Nylén, S. Rojas and M. Boutonnet, Appl. Catal., A, 2004, 265, 207–219.
23. D.-H. Chen and S.-H. Wu, Chem. Mater., 2000, 12, 1354–1360.
24. J. Chandradass and D.-S. Bae, Mater. Manuf. Processes, 2008, 23, 494–498.
25. M. H. A. Shiraz, M. Rezaei and F. Meshkani, Int. J. Hydrogen Energy, 2016, 41, 6353–6361.
26. W. Hua, L. Jin, X. He, J. Liu and H. Hu, Catal. Commun., 2010, 11, 968–972.
27. Y. J. Asencios, J. D. Bellido and E. M. Assaf, Appl. Catal., A, 2011, 397, 138–144.
28. M. Yu, K. Zhu, Z. Liu, H. Xiao, W. Deng and X. Zhou, Appl. Catal., B, 2014, 148, 177–190.
29. A. G. Bhavani, W. Y. Kim, J. Y. Kim and J. S. Lee, Appl. Catal., A, 2012, 450, 63–72.
30. Y. Kathiraser, U. Oemar, E. T. Saw, Z. Li and S. Kawi, Chem. Eng. J., 2015, 278, 62–78.
31. S. Therdthianwong, C. Siangchin and A. Therdthianwong, Fuel Process. Technol., 2008, 89, 160–168.
32. D. Liu, R. Lau, A. Borgna and Y. Yang, Appl. Catal., A, 2009, 358, 110–118.
33. A. Horváth, G. Stefler, O. Geszti, A. Kienneman, A. Pietraszek and L. Guczi, Catal. Today, 2011, 169, 102–111.
34. K. Takanabe, K. Nagaoka, K. Nariai and K.-i. Aika, J. Catal., 2005, 230, 75–85.
35. G. L. Bezemer, P. Radstake, V. Koot, A. Van Dillen, J. Geus and K. De Jong, J. Catal., 2006, 237, 291–302.
36. J. Zhang, H. Wang and A. K. Dalai, Appl. Catal., A, 2008, 339, 121–129.
37. L. Guczi, G. Stefler, O. Geszti, I. Sajó, Z. Pászti, A. Tompos and Z. Schay, Appl. Catal., A, 2010, 375, 236–246.
38. S. Wang, G. Q. Lu and G. J. Millar, Energy Fuels, 1996, 10, 896–904.
39. J. Zhang, H. Wang and A. K. Dalai, J. Catal., 2007, 249, 300–310.
40. P. Frontera, A. Macario, A. Aloise, P. Antonucci, G. Giordano and J. Nagy, Catal. Today, 2013, 218, 18–29.
41. H. Özdemir, M. F. Öksüzömer and M. A. Gürkaynak, Int. J. Hydrogen Energy, 2010, 35, 12147–12160.
42. F. Meshkani and M. Rezaei, Catal. Commun., 2011, 12, 1046–1050.
43. Y. Li, G. Lu and J. Ma, RSC Adv., 2014, 4, 17420–17428.
44. H. Arbag, S. Yasyerli, N. Yasyerli and G. Dogu, Int. J. Hydrogen Energy, 2010, 35, 2296–2304.
45. M. Rezaei, S. Alavi, S. Sahebdelfar, P. Bai, X. Liu and Z.-F. Yan, Appl. Catal., B, 2008, 77, 346–354.
46. X. Chen, K. Honda and Z.-G. Zhang, Appl. Catal., A, 2005, 288, 86–97.
47. K. Tomishige, Y. Matsuo, Y. Yoshinaga, Y. Sekine, M. Asadullah and K. Fujimoto, Appl. Catal., A, 2002, 223, 225–238.
48. L. Xu, H. Zhao, H. Song and L. Chou, Int. J. Hydrogen Energy, 2012, 37, 7497–7511.
49. D. San José-Alonso, M. J. Illán-Gómez and M. C. Román-Martínez, Int. J. Hydrogen Energy, 2013, 38, 2230–2239.
50. N. Rahemi, M. Haghighi, A. A. Babaluo, M. F. Jafari and P. Estifaee, J. Ind. Eng. Chem., 2013, 19, 1566–1576.
51. J. Newnham, K. Mantri, M. H. Amin, J. Tardio and S. K. Bhargava, Int. J. Hydrogen Energy, 2012, 37, 1454–1464.
52. S. Fan, M. G. Chapline, N. R. Franklin, T. W. Tombler, A. M. Cassell and H. Dai, Science, 1999, 283, 512–514.
53. S. Amelinckx, X. Zhang, D. Bernaerts, X. Zhang, V. Ivanov and J. Nagy, Science, 1994, 265, 635–639.
54. S. K. Saraswat and K. K. Pant, Int. J. Hydrogen Energy, 2011, 36, 13352–13360.
55. F. Frusteri, L. Spadaro, F. Arena and A. Chuvilin, Carbon, 2002, 40, 1063–1070.
56. S. Tsang, J. Claridge and M. Green, Catal. Today, 1995, 23, 3–15.
57. D. Chen, R. Lødeng, K. Omdahl, A. Anundskås, O. Olsvik and A. Holmen, Stud. Surf. Sci. Catal., 2001, 139, 93–100.
58. K. Tomishige, Y.-g. Chen and K. Fujimoto, J. Catal., 1999, 181, 91–103.
59. J. A. Montoya, E. Romero-Pascual, C. Gimon, P. Del Angel and A. Monzon, Catal. Today, 2000, 63, 71–85.
60. K. Tomishige, O. Yamazaki, Y. Chen, K. Yokoyama, X. Li and K. Fujimoto, Catal. Today, 1998, 45, 35–39.

---