Catalytic Reaction of Carbon Dioxide with Methane on Supported Noble Metal Catalysts

Abstract

The conversion of CO₂ and CH₄, the main components of the greenhouse gases, into synthesis gas are in the focus of academic and industrial research. In this review, the activity and stability of different supported noble metal catalysts were compared in the CO₂ + CH₄ reaction on. It was found that the efficiency of the catalysts depends not only on the metal and on the support but on the particle size, the metal support interface, the carbon deposition and the reactivity of carbon also influences the activity and stability of the catalysts. The possibility of the activation and dissociation of CO₂ and CH₄ on clean and on supported noble metals were discussed separately. CO₂ could dissociate on metal surfaces, this reaction could proceed via the formation of carbonate on the support, or on the metal–support interface but in the reaction the hydrogen assisted dissociation of CO₂ was also suggested. The decrease in the activity of the catalysts was generally attributed to carbon deposition, which can be formed from CH₄ while others suggest that the source of the surface carbon is CO₂. Carbon can occur in different forms on the surface, which can be transformed into each other depending on the temperature and the time elapsed since their formation. Basically, two reaction mechanisms was proposed, according to the mono-functional mechanism the activation of both CO₂ and CH₄ occurs on the metal sites, but in the bi-functional mechanism the CO₂ is activated on the support or on the metal–support interface and the CH₄ on the metal.

1. Introduction

Nowadays it is no longer necessary to emphasize that the increase of the CO₂—one of the main greenhouse gases—concentrations in the atmosphere and what kind of dangerous changes it can cause. To date, fossil fuels as the main sources of energy, the relatively new technologies (cement production, iron, steel industry etc.), the revolution in transportation, and the rapid growth in global population have resulted in the dramatic increase of CO₂ emission. Despite different international agreements CO₂ emissions have continued to increase. The CO₂ concentration in the atmosphere in the last few 100,000 years always remained below 300 ppm, but now the CO₂ concentration is over 415 ppm (in May 2020 the monthly mean value was 417 ppm) [1]. According to lots of meteorologists and climate experts, higher CO₂ concentration in the atmosphere results in climate change, the rise in temperature. The global annual mean temperature increased in the last 100 years by 1.4 degrees [2]. The climate change seems to have become generally accepted—although there are always those who disagree with it—and this has resulted in a worldwide increase in interest in reducing greenhouse gases, particularly CO₂, emissions.

If we want to prevent the increase of CO₂ concentration in the atmosphere we have to (i) hinder the enhancement of the carbon dioxide emission; (ii) use CO₂ capture and storage (CCS) technologies; or (iii) use CO₂ capture and utilization (CCU) technologies.

There is an ongoing debate as to whether CO₂ is a waste, or a wealth as feedstock. If CCS technology is used, CO₂ is considered waste, but if CO₂ is converted into a more valuable compound in a chemical reaction, it is not a waste but a feedstock.

The total amount of products that can be produced using technologies available today for CO₂ conversion is only a few hundred Mt, while CO₂ emissions from large stationary sources alone exceed 10 Gt. This means that traditional products and technologies do not really affect the CO₂ concentration in the atmosphere. One of the highest carbon emission sectors is transportation, so we have to look for the solution in these fields.

The reaction of carbon dioxide with methane, and dry reforming in particular utilizes two abundantly available greenhouse gases, to produce the industrially important syngas (CO + H₂ mixture). From the CO + H₂ mixture both fuel and methanol or plenty of other products can be produced using available technologies. This process could constitute a viable first step towards efficient CO₂ capture and utilization. The main problem is that dry reforming is a highly endothermic and equilibrium reaction (see later). However, CO hydrogenation is an exothermic process so these technologies could help in the storage of electricity because the greatest problem with renewable energy (wind and solar) production is the fluctuation, and the energy consumption is also uneven.

Opponents of the CO₂ + CH₄ reaction argue that carrying out this reaction requires a significant amount of energy. It is true that endothermic reaction consume energy. Synthesis gas is now produced worldwide by the steam reforming of CH₄, which is also an endothermic process and the dry reforming requires only 20% more energy, but converts greenhouse gases into a chemical feedstock. The disadvantage of steam reforming is that the H₂/CO ratio is too high to be used for methanol synthesis, although in the case of the CO₂ + CH₄ reaction this value is too low. The correct H₂/CO ratio could be obtained by reforming of methane with CO₂ + H₂O mixture.

Extensive research work has been conducted on different catalytic systems for the conversion of CO₂ + CH₄, various catalysts have been developed but there have been only a few attempts to industrialize this reaction, so it is not surprising that the reaction of carbon dioxide with methane is in the focus of academic and industrial research.

The present review attempts to summarize the results obtained on supported noble metal catalysts. The main aim is to show the different results obtained on the similar catalysts, the different theories regarding the interaction of the reactants and the catalysts and the reaction mechanism.

2. The Thermodynamics of the CO₂ + CH₄ Reaction

Carbon dioxide reforming of methane is a highly endothermic reaction (1). Thermodynamic calculation indicates that this process is not spontaneous at 1 atm pressure below 906 K. When the reaction takes place between the products and between the products and the reactants, a number of secondary reactions can occur (2)–(7).

$$\begin{array}{}\mathrm{(1)}& {\mathrm{CO}}_2+{\mathrm{CH}}_4\rightleftharpoons 2\mathrm{CO}+2{\mathrm{H}}_2\hfill & \Delta H_{298}^0=+247\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{dry}\mathrm{reforming}\hfill \mathrm{(2)}& {\mathrm{CH}}_4+{\mathrm{H}}_2\mathrm{O}\rightleftharpoons \mathrm{CO}+3{\mathrm{H}}_2\hfill & \Delta H_{298 }^0=+206\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{steam}\mathrm{reforming}\hfill \mathrm{(3)}& {\mathrm{CH}}_4\rightleftharpoons \mathrm{C}+2{\mathrm{H}}_2\hfill & \Delta H_{298}^0=+75\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{methane}\mathrm{decomposition}\hfill \mathrm{(4)}& {\mathrm{CO}}_2+{\mathrm{H}}_2\rightleftharpoons \mathrm{CO}+{\mathrm{H}}_2\mathrm{O}\hfill & \Delta H_{298}^0=+41.2\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{reverse}\mathrm{water–gas}\mathrm{shift}\hfill \mathrm{(5)}& {\mathrm{CO}}_2+4{\mathrm{H}}_2\rightleftharpoons {\mathrm{CH}}_4+2{\mathrm{H}}_2\mathrm{O}\hfill & \Delta H_{298}^0=-164.9\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{methanation}\hfill \mathrm{(6)}& 2\mathrm{CO}\rightleftharpoons \mathrm{C}+{\mathrm{CO}}_2\hfill & \Delta H_{298}^0=-171\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{Boudouard}\mathrm{reaction}\hfill \mathrm{(7)}& \mathrm{C}+{\mathrm{H}}_2\mathrm{O}\rightleftharpoons \mathrm{CO}+{\mathrm{H}}_2\hfill & \Delta H_{298}^0=+131\mathrm{kJ}/\mathrm{mol}\hfill & \mathrm{carbon}\mathrm{gasification}\hfill \end{array}$$

The calculated equilibrium conversions of methane and carbon dioxide at different temperatures are collected in

Table 1. The calculated equilibrium conversions at different temperatures for the CO₂ + CH₄ reaction at 1 atmosphere pressure [3].

Temperature K	CH4 Conversion %	CO2 Conversion %
673	6.9	9.9
773	25.6	32.9
873	57.7	66.8
973	84.0	89.6
1073	95.0	97.3
1173	98.3	99.2
1273	99.3	99.7

[3]. Similar results were obtained by Bradford and Vannice [4] for a feed stream at 1 atm total pressure (CH₄/CO₂/He = 1/1/1.8). The CO₂ conversion was always higher than that of methane, due to the reverse water gas shift reaction. The methanation (5) and the reverse reaction of steam reforming (2) may also have contributed to the lower methane conversion.

Different equilibrium conversions were calculated by Istadi and Amin [5] and by Nikoo and Amin [6]. The equilibrium methane conversion at 873 K was about 42% [5] while in the other work a value of 82% was reported [6]. The high CH₄ conversion in the latter case was explained by stating that the CH₄ decomposition (3) is the predominant reaction.

Figure 1 shows the equilibrium product distribution for the CO₂ + CH₄ reaction using CO₂/CH₄ feed ratio of 1/1 at a pressure of 1 atm [7]. This calculation cannot take into consideration the possibility of carbon deposition. According to this calculation water is formed up to 1173 K in the reaction and the CO/H₂ ratio is greater than one. These results indicate that the reverse water gas shift reaction also takes place.

The reactions responsible for carbon formation are the methane decomposition (3) and the Boudouard reaction (6). Carbon formation cannot be ruled out in the hydrogenation of CO and CO₂. It was found [6] that of these processes only the methane decomposition is favored at high temperatures; while the other reactions are favored when the reaction takes place at lower than 800 K temperature.

When the carbon formation, i.e., the methane decomposition (3) and the Boudouard reactions (6), were taken into account in the calculation entirely different equilibrium composition was found (Figure 2) [8]. In this case the amount of H₂ is more than that of CO, the H₂/CO ratio is far greater than unity because the increased carbon formation lowers the amount of CO formed and hence increases the H₂/CO ratio.

According to Wang et al. [9] the reverse water–gas shift (4) and the Boudouard reaction (6) will not occur at above 1093 K, and the Boudouard reaction and methane decomposition (3) will be mostly responsible for the formation of carbon at temperatures ranging from 830 K to 973 K.

3. Comparing the Activity of Different Catalysts in the CO₂ + CH₄ Reaction

The possibility of a reaction between methane and carbon dioxide was already mentioned by Lang [10] in the late 19th century, and in 1928, Fischer and Tropsch [11] showed that different metals have different activities for the dry reforming of methane and that the group VIII metals show appreciable activity for this reaction. However, intensive research did not begin until the second half of the 20th century. Since then, a variety of different supported metals were used as catalysts. Most of the group VIII metals, except Os, are more or less active in the CH₄ + CO₂ reaction. The results obtained are summarized in different reviews [4,12,13,14,15]. The biggest problem was in most cases the rapid deactivation of the catalysts due to the significant carbon deposit. This may also be the reason of different activity orders published by different research groups. Catalysts based on noble metals are less sensitive to coking than Ni containing samples.

Theoretically the H₂/CO ratio is unit, but due to the reverse water gas shift reaction and to the other secondary reactions this ratio is often less.

The reforming reaction of CH₄ with CO₂ has been studied over different supported transition metals. It was found that at 1073 K the turnover frequencies (TOF) on Al₂O₃ supported catalysts (Ru, Rh, Pt, Ir, Ni) were within the same order of magnitude. The value obtained for the best Ru (4.0 s⁻¹) was only slightly more than twice as for the weakest Ni catalyst (1.8 s⁻¹) [16]. This order deviate from those found by Rostrup-Nielsen and Bak Hansen [17] who obtained Ru > Rh ≥ Ni > Ir > Pt > Pd order (

Table 2. Comparison of turnover frequencies of supported noble metal catalysts for CO₂ reforming of methane.

	Al2O3 CO2/CH4 = 1/1.5 1073 K [16]	Al2O3 CO2/CH4 = 1/1 973 K [17]	Al2O3 CO2/CH4 = 1/1 823 K [18]	Alumina Stabilized Magnesia CO2/CH4 = 4/1 823 K [19]	Al2O3 CO2/CH4/He = 1/1/8 723 K [20]
Ru	4.0	2.0	1.36	2.9	0.2
Rh	3.4	4.0	0.32	1.9	0.8
Pd		< 0.1	0.64	0.18
Ir	2.0	2.0	0.18	0.44	0.5
Pt	3.8	< 0.1	0.36	0.36	0.2
Ni	1.8			1.9	0.6

).

No significant differences were found at 1050 K in the activities of the alumina supported metal catalysts [21,22] only in the cases of Ru and Pd were significantly lower conversions observed; in the other cases they approached the equilibrium values. The Ni and Pd catalysts were very active at first (especially Ni), then rapidly showed a decline in conversions, and a considerable deposition (up to 40%) of carbon occurred within a few hours. The Ir/Al₂O₃ catalyst maintained its very high performance for over 200 h, with no evidence for deactivation, and only less than 0.3% carbon was observed on the used catalyst.

The order of catalytic activity at 723 K on turnover frequency basis was found to be Rh > Ni > Ir> Ru~Pt > Co over the alumina supported catalysts [20] (Table 2). A different sequence was found for SiO₂ supported metals: Ni > Ru > Rh~Ir > Co~Pt [20]. This indicates that the support has a great influence on the specific activity of the catalysts.

Over different MgO supported metal catalysts the activities depended upon the metals in the sequence Ni, Ru, Pt >> Rh > Pd [23]. The amounts of CO were greater than those of H₂ possibly due to the reverse water gas shift reaction. Different results were obtained at 1073 K on MgO supported noble metals [24]. The methane conversion was the highest on Rh and the activity decreased following the order Rh > Ru > Ir > Pt > Pd.

CO₂ reforming of CH₄ has been also studied on catalysts containing less than a monolayer of Rh, Ru, and Ir deposited on α-Al₂O₃, MgO, CeO₂, La₂O₃, and TiO₂ [25]. It was found that the activity, the turnover frequency of CH₄ consumption on Rh/α-Al₂O₃ was four times higher than on Ru and Ir. The high stability of the noble metal catalyst comparing with the Ni was attributed to the high oxidative properties of compounds formed on the catalyst surface, which inhibit the carbon growth reactions [25].

The turnover frequency for CO formation in the dry reforming of CH₄ on SiO₂ and TiO₂ supported transition metals depends on both the support and the %-d character of the metal [26]; however, different trends were observed for metals on SiO₂ and TiO₂. The activities of the TiO₂ supported metals were higher than those of SiO₂ based catalysts in all cases. Pt/SiO₂ has an outstanding activity among the silica supported samples and it decreased in the following order at 723 K: Pt >> Pd > Ni > Rh > Fe. Among the TiO₂ supported catalysts the Rh activity was the highest. In this case the turnover frequency for CO formation decreased in the Rh > Ir > Ru > Pt > Pd > Ni > Fe~Co~Cu order [26].

The methane conversion at 923 K decreased on pillared clay impregnated by different metals along the 3% Rh > 10% Ni > 3% Pd > 3% Ni > 3% Ce sequence [27]. The H₂/CO ratio was 1.1 for Rh and 0.85 for Pd which is metallic at the temperature of the reaction but it was even lower for Ni and Ce.

Comparing the activities of the ZnLaAlO₄ supported Ru, Ni, and Pt catalysts reveals that the best performance was observed for 3% Ru/ZnLaAlO₄ while the 3% Pt shows the lowest activity. These results were attributed to the synergistic effect of the interaction between the support and the active metals [28].

Only the activities of noble metals have been compared in several publications [18,29,30,31,32,33] and great differences were found in their catalytic behavior as concerns the CH₄ + CO₂ reaction. The reaction occurred at the highest rate on Rh, followed by Pt, Pd, Ru, and Ir. Little or no deactivation of the catalyst occurred during the reaction at 823 K.

The conversion of CO₂ exceeded that of CH₄ for every catalyst sample indicating that the dry reforming reaction was followed by several secondary processes, including the methanation of CO₂ and CO, reverse water gas shift reaction and the Boudouard reaction. The turnover frequency for CH₄ or CO₂ consumption at 823 K on the alumina supported 1 wt% of noble metals decreased in the sequence Ru > Pd > Rh > Pt > Ir [18] (Table 2). This is practically the same as the order for CO methanation [34], and the sequence for CO₂ methanation differed only in that Pd was the least active sample [35]. In contrast at 923 K the order Rh > Ru > Ir >> Pd, Pt was determined for catalysts with 0.5% metal supported on Al₂O₃ [17] (Table 2) but others found that the activity decreased in the order Rh > Pd > Pt >> Ru [29]. Rh, Ru, and Ir catalysts showed stable long term activity but Pd and Pt containing catalysts deactivated rapidly. On alumina-stabilized magnesia and Mg-Al mixed oxide supported Pt metals at 973 K a different sequence was observed [30,32,33] (Ru > Rh > Ir > Pt > Pd).

The order of activities for ZrO₂-supported catalysts was Rh > Pt > Ir, Pd > Ru [36]. This sequence is different from that reported for MgO-supported materials [19,24]. The observed differences suggest that the basicity of the support may have some significant effect.

Stable CH₄ and CO₂ conversion was observed for 50 h on Ru substituted LaAlO₃ (LaAl_0.98Ru_0.02O_3−δ) catalyst. This indicates that it is a highly stable and coke resistant catalyst. It has been suggested that the thermal stability of the sample is due to a strong interaction between Ru and Al. In contrast, Pt and Pd incorporated into LaAlO₃ lost their activities due to sintering and carbon deposition [37].

The activities of Rh, and Pt supported α-Al₂O₃ doped with La₂O₃ and BaO were compared in the dry reforming of methane [38]. It was found that Rh is more active than Pt independently of the support composition. La incorporation induces a rate enhancement in the conversion of methane in all cases, but Ba incorporation in Pt/Al₂O₃ leads to loss of activity because there is a preferential interaction between Pt and Ba, in the case of Rh this had no effect.

Ru- and Rh-catalysts supported on Mg-Al mixed oxides revealed activity in the combined partial oxidation and dry reforming of methane to synthesis gas at 1123 K which was markedly higher than that of Pt. Ru and Rh approached their maximal activity which was limited by the equilibrium [39].

Ru and Ir supported on Eu₂O₃ were active in the dry reforming reaction, but Ru showed an onset of activity around 673 K, whereas the Ir showed initial activity at a much higher temperature, around 823 K [40].

The activity of BaZr_(1−x)Me_xO₃ (Me: Ru, Rh, Pt) perovskite type oxides [41] decreased with the active species as Rh > Ru > Pt. The better activity of Rh- and Ru-containing samples was linked to their highl- reducible nature.

The brief summary above clearly shows that noble metals are effective in the CO₂ + CH₄ reaction. Despite the high costs of noble metals, their superior catalytic properties at low temperature still make them indispensable to understand the interaction between the metal and the support, to check the surface compounds formed during the reaction and to show the elementary steps of the reaction, and as additives appear to be unavoidable for the development of an effective and stable catalyst.

4. Dry Reforming of Methane on Supported Noble Metal Catalysts

4.1. Supported Rhodium

The above summary clearly shows that Rh is one of the best and most stable catalysts of the reaction.

The specific activity of Rh catalysts was found to strongly depend on the support [42,43,44,45]. It decreased in the order yttria-stabilized zirconia (YSZ) > Al₂O₃ > MgO > TiO₂ > SiO₂ [43,44]. This order correlates directly with the acidic character of the support [43]. A strong dependence was observed on the specific activity of the methane reforming reaction and the rate of deactivation of the catalysts on rhodium particle size in the range 1–6 nm. Both activity and rate of deactivation decrease with increasing metal particle size. The degree of these dependencies was largely affected also by the nature of the support. Relatively high deactivation rates were observed over Rh/TiO₂ and on Rh/MgO, but on Rh/YSZ and on Rh/SiO₂ no deactivation was observed. Three kinds of factors were suggested which contribute to the catalyst deactivation carbon deposition, metal sintering, and the poisoning of surface Rh sites by species originating from the carrier [43]. The TOF for 0.5% Rh/YSZ at 923 K was about five times higher than for the less active SiO₂ or TiO₂ supported catalysts in the steady state [44].

Others found no significant difference in the catalytic activity at 773 K on 1% Rh supported on A1₂O₃, TiO₂, SiO₂, MgO (

Table 3. Turnover frequencies on different supported Rh catalysts in the CO₂ + CH₄ reaction.

Catalyst	Temperature K	${\mathbf{N}}_{\mathbf{C}{\mathbf{H}}_4} {\mathbf{s}}^{-1}$	${\mathbf{N}}_{\mathbf{C}{\mathbf{O}}_2} {\mathbf{s}}^{-1}$	${\mathbf{N}}_{\mathbf{C}\mathbf{O}} {\mathbf{s}}^{-1}$	${\mathbf{N}}_{{\mathbf{H}}_2} {\mathbf{s}}^{-1}$	References
Rh/Al2O3	723			0.52		[47]
Rh/TiO2	723			0.29		[47]
Rh/Nb2O5	723			0.48		[47]
Rh/NaY	723			0.08		[47]
3.8% Rh/SiO2	723	0.08		0.18		[48]
3.3%Rh/VOx/SiO2	723	1.3		3.4		[48]
1% Rh/Al2O3	773			0.39	0.15	[46]
1% Rh/MgO	773			0.85	0.5	[46]
1% Rh/TiO2	773			0.37	0.22	[46]
1% Rh/SiO2	773			0.32	0.2	[46]
1% Rh/α-Al2O3	873	1.5				[17]
1% Rh/ZrO2	873	0.9				[17]
1% Rh/TiO2	873	1.3				[17]
1% Rh/SiO2	873	1.2				[17]
Rh/γ-Al2O3	875		4.6			[49]
Rh/SiO2	875		1.2			[49]
0.5% Rh/Al2O3	893	25.5				[50]
0.5% Rh/SiO2	893	6.1				[50]
0.5% Rh/Al2O3	1013	6.35				[45]
Rh/α-Al2O3	1023	5.02	5.08			[25]
Rh/MgO	1023	1.04	1.06			[25]
Rh/CeO2	1023	2.82	3.04			[25]
Rh/TiO2	1023	0.74	0.85			[25]
Rh/La2O3	1023	4.01	4.06			[25]

) and no decay in the activity was experienced and small amount if any carbon deposit was observed [46].

In contrast a significant effect of the support was found on the reaction rate on Rh supported on the same oxides, where the order was Rh/Al₂O₃ > Rh/TiO₂ > Rh/SiO₂. The catalytic activity of Rh/SiO₂ increased significantly by physically mixing the catalyst with metal oxides such as A1₂O₃, TiO₂, and MgO, indicating a synergistic effect. The role of the metal oxides used in mixture according to the authors may be ascribed to promoting in dissociation CO₂ dissociation on the surface of Rh, since the CH₄ + CO₂ reaction is first order in the pressure of CO₂, and −0.6 for CH₄, suggesting that CO₂ dissociation is the rate-determining step [50,51].

The reforming of methane with carbon dioxide was investigated on Rh supported on reducible (CeO₂, Nb₂O₅, Ta₂O₅, TiO₂, and ZrO₂) and on irreducible (γ-Al₂O₃, La₂O₃, MgO, SiO₂, and Y₂O₃) oxides [45]. The reducible oxides supported Rh provided much lower conversion of CH₄ or CO₂ than the irreducible one. One possible explanation is that the Rh particles were partially covered by partially reduced oxide species (SMSI effect). Among the irreducible oxides supported catalysts, γ-Al₂O₃, La₂O₃, and MgO showed stable activities, and it increased in the sequence: La₂O₃ < MgO ≈ γ-Al₂O₃. Possible explanations for stability, based on the strong interactions between Rh₂O₃ and support were proposed. Among the reducible oxides, Nb₂O₅ supported Rh has a low activity, Rh/ZrO₂ and Rh/CeO₂ exhibited a very long activation period, and deactivation occurred over Ta₂O₅ and TiO₂. MgO and γ-Al₂O₃ supported Rh are the most promising catalysts; they provided a stable high activity with a CO yield of 83–85% and a H₂ yield of 76–79% at the high space velocity of 60 000 mL/h/g [45].

The turnover frequency for methane conversion in the CO₂ + CH₄ reaction on different oxide supported Rh catalysts decreased in the following order: Rh/TiO₂ > Rh/La₂O₃ = Rh/CeO₂ > Rh/ZrO₂ = Rh/MgO = Rh/SiO₂ = Rh/MCM-41 > Rh/Al₂O₃. The outstanding activity of Rh/TiO₂ was attributed to the fact that Rh remained in metallic state during the reaction, while on Rh/Al₂O₃ according to XANES characterization cationic Rh was generated [42].

The highest yield in hydrogen in the dry reforming of CH₄ was obtained on differently doped alumina supported Rh in the following order: Rh/NiO-Al₂O₃ > Rh/MgO-Al₂O₃~Rh/CeO₂-Al₂O₃ > Rh/ZrO₂-Al₂O₃ > Rh/La₂O₃-Al₂O₃. The high activity of Rh/NiO-Al₂O₃ was explained by the presence of NiAl₂O₄ spinel preventing the deactivation of Rh migration in alumina and the high dispersion of Rh favored by the presence of Ni particles at the support [52].

Rh/NaY effectively catalyzes the conversion in the CO₂ + CH₄ reaction to CO and H₂ with a H₂/CO ratio near unity even with low metal loadings [53,54]. Remarkable is the stability of these catalysts under operating conditions. The reason for this, according to the authors is the total absence of carbonaceous deposits. The Rh particles are quite uniform, having a size of~4 nm [53]. These particles are available in the early stages of the catalysis and the zeolite framework protects the encaged particles against further growth. Therefore, catalyst deactivation is not observed under these conditions which are significant for Rh on amorphous supports [53,54].

Unlike other results [54,55] no carbon formation was observed on Rh/Al₂O₃ in the temperature range of 873–1073 K [56]. Although the activity of the catalyst decreased slightly over time, this was attributed to the sintering because the catalysts could not be regenerated. This catalyst has been successfully tested in large scale experiments in solar panels. The reforming reaction at 923 K over Rh/Al₂O₃ results in the accumulation of three kinds of carbon species. The ratio of different carbon species depended on the reaction temperature and on the time of exposure. These results support the earlier finding that the transformation of the active form into a less active form occurs with increasing reaction time and the temperature [57]. The amount of carbon deposit decreased at temperatures higher than 923 K probably because of the prevalence of the carbon gasification (8) and the reverse Boudouard reaction (6). It was found that the origin of carbon is mainly the CO₂ molecule, with a small contribution from the CH₄ molecule [44,55].

Bitter et al. [49] also found that the activity of supported Rh catalysts for CO₂ + CH₄ reaction is mainly determined by the availability of Rh irrespective of the support. Contrary to this, Portugal et al. [57] found that the turnover frequency for CO formation over supported Rh at 723 K is strongly affected by the support; it decreased in the following order: Rh/Al₂O₃ > Rh/Nb₂O₅ > Rh/TiO₂ > Rh/NaY (Table 3). The higher activity of the oxide-supported catalysts was explained by the higher degree of contribution of the reverse water–gas shift reaction. Similarly, a dominant role of the carrier was observed, the turnover frequency for CH₄ or CO₂ consumption at 1023 K decreased in the Rh/α-Al₂O₃ > Rh/La₂O₃ > RhCeO₂ > Rh/MgO > Rh/TiO₂ order [25] (Table 3).

Alumina supported Rh catalysts with a wide range of dispersions (25.1–69.0%) and Rh particle size (1.4–4.0 nm) were prepared by varying the metal content (0.1–1.6 wt%) and the thermal treatment temperature (923–1123 K) [58]. It was found that the turnover rates increased monotonically with increasing Rh dispersion. A similar result was obtained on different NaY zeolite supported Rh, too [47]. This observation suggests that coordinative unsaturated Rh surface atoms, which are more common in small clusters, are more active than those in low-index surface planes on larger Rh crystallites. Edge and corner atoms appear to bind CH_x and H more strongly and to decrease the energy required to form relevant transition states for the initial C–H bond activation. They assume that supports can influence metal dispersion and, only in this manner, affect forward turnover rates [58]. It should be noted that a reduction of Rh particle size result in an enhanced rate of deactivation in the case of supported Rh [43].

Contrary to the above findings, it was found that over a wide range of metal content of the catalysts (0.5–5%) and supports (α- and γ-Al₂O₃, ZrO₂, TiO₂, SiO₂), the turnover frequency for CH₄ consumption at 1023 K does not depend significantly on dispersion, indicating that CO₂ reforming at high temperatures is structure insensitive. No direct influence of the microstructure of the support on the reaction rates was detected. The effect of the support seems limited to the stabilization of the metal surface area, which is responsible for the activity of the catalyst [17].

Rh/Al₂O₃ promoted with V₂O₅ or TiO₂ were more active in the CO₂ + CH₄ reaction at 773 K than the unpromoted catalyst [59]. The increase of the conversion was attributed to the oxygen vacancies which formed on the promotors during the pretreatment of the catalysts. Vanadium oxide promotion enhances the activity of Rh/SiO₂ too and decreases the deactivation by carbon deposition [48]. This effect was attributed to the formation of VO_x overlayer on the Rh which thereby reduces the size of accessible ensembles of Rh atoms required for coke formation. The new sites at the Rh-VO_x interface were considered to be active for the dissociation of CO₂.

The activity and coke resistance of Rh/Al₂O₃ were enhanced by the addition of CeO₂. It was suggested that this effect is associated with the presence of two redox couples, which favors the activation of both CH₄ and CO₂ [60]. A Rh–CeO₂ interaction was induced by high temperature reduction, which resulted in the formation of oxygen vacancies in ceria. The electrons could transfer from Rh to CeO₂ on the perimeter. The Ce⁴⁺/Ce³⁺ and Rh⁰/Rh^δ+ redox couples were found to coexist in Rh–CeO₂/Al₂O₃ catalyst. CH₄ decomposition on Rh released electrons, whereas CO₂ fill up the oxygen vacancies by accepting electrons. The Ce³⁺ species promote CO₂ dissociation into CO and surface oxygen, which reacts with CH_x to form CO [60].

Slow deactivation of 2% Rh/CeO₂ was observed at 1023 K during 70 h of the reaction, the methane conversion decreased from the initial value of 48.6 to 44.1%., which was attributed to catalyst sintering [61,62], the role in the deactivation of the coking was neglected.

Rh/ZSM-5 zeolite was found to be an active catalyst for dry reforming of methane, the reaction at 773 K was approaching the equilibrium product distribution, but adding 4–5% ethane to the reacting gas mixture lead to carbon deposition and to the deactivation of the catalyst [63].

Surprisingly, only a few studies reported on the effect of sulfur-containing compounds on the CO₂ + CH₄ reaction, although natural sources like natural gases and bio gases as possible candidates for being starting materials in this reaction, often contain more or less sulfur-containing compounds. Prichard et al. [64] did not observe any deactivation for 0.5% Rh/Al₂O₃ catalyst at 1068 K in the presence of 543 ppm dimethyl sulfide. In their opinion, this is caused by the high temperature, because at this temperature the interaction between the sulfur and the catalyst was limited.

The effect of H₂S (22 ppm) on the reaction of CO₂ with CH₄ at 773 K over TiO₂, SiO₂, Al₂O₃, ZrO₂, and CeO₂-supported Rh catalysts was examined [65] (Figure 3).

It was found that H₂S suppresses the formation of CO and H₂ in all cases, but on Rh/TiO₂ and on Rh/ZrO₂ the CO/H₂ ratio increased in time, indicating that H₂S poisons the secondary processes of the reaction to different extents. Therefore, the effect of H₂S on the CO₂ hydrogenation and the steam reforming of methane were investigated. It was found that the rate of methane formation in the CO₂ + H₂ reaction increased on TiO₂ and on CeO₂ supported metals (Ru, Rh, Pd), but on all other supported catalysts or when the H₂S content was higher than 22 ppm the reaction was poisoned [66]. The promotion effect of H₂S was explained by the formation of new centers at the metal–support interface.

Rh supported on La stabilized γ-Al₂O₃ has an outstanding catalytic activity for dry reforming of methane at 1073 K, but it was severely poisoned by even small amounts (below 1 ppmv) of sulfur in the feed [67]. The reverse water–gas shift reaction, occurring simultaneously with dry reforming, was only slightly affected by sulfur poisoning, approaching the equilibrium even at high S-levels. Sulfur inhibition was reversible and the Rh-catalyst slowly recovered its initial activity after the removal of sulfur from the feed.

4.2. Supported Ruthenium

Supported Ru catalyst is the best in some comparisons, while other studies show that it is one of the less active catalysts.

The support affected significantly the catalytic activity of supported Ru in the CO₂ reforming of methane at low temperature. With 0.5% Ru/La₂O₃ at 873 K CH₄ conversion reached 28.7% in stoichiometric CH₄ and CO₂ (1:1) mixture in a continuous flow reactor. On 0.5% Ru/Al₂O₃, it was only 5.7%. Y₂O₃ and ZrO₂ exhibited higher catalytic activities in this reaction (25–29%), but SiO₂ or TiO₂ gave lower conversions of CH₄ (less than 12%) [68].

On a turnover frequency basis, the activity decreased in the order of Ru/TiO₂ > Ru/Al₂O₃ >> Ru/C under conditions where heat and mass transfer limitations do not exist [69]. The catalytic activity under 0.1 MPa at 1023 K depends on the basicity of the supports (Ru/MgO > Ru/Al₂O₃ > Ru/TiO₂ > Ru/SiO₂) indicating that the CO₂ adsorption ability, the formation of carbonates on the supports determine the efficiency of the catalysts. Similar conclusion was reached on supported Rh catalysts [50]. Under 2 MPa pressure the activity correlated with the BET surface area, the activity order was Ru/SiO₂ > Ru/Al₂O₃ > Ru/MgO > Ru/TiO₂. On Ru/TiO₂ carbon deposition was not observed, the lowest activity of this catalyst was attributed to the SMSI effect [70].

Comparing the activity of Ru/Al₂O₃, Ru/SiO₂, and Ru/graphite [71] at 823 K it was found that at the beginning of the reaction the alumina supported catalyst has the highest activity, but it decreased continuously. The graphite-supported sample exhibits a rather stable activity, in contrast with the oxide-supported catalysts. The fastest deactivation rate is observed for Ru/SiO₂, its activity decays dramatically during the first 5 h. For the explanation of this observation, it was supposed that the activation of both reactants takes place on the metal surface and the accumulation of carbon species formed from CH₄ decomposition on the metal finally hinders CO₂ dissociation and results in a rapid deactivation of this catalyst [72]. The outstanding activity of Ru/TiO₂ compared to Ru/Al₂O₃ or Ru/C was attributed to the presence of interfacial Ru-Tiⁿ⁺-O_x species which markedly enhanced the activation of the reactants.

Several research groups investigated the activity of Ru/Al₂O₃ in the CO₂ + CH₄ reaction and obtained quite different results [17,18,25,31,62,69]. Direct comparison of the catalytic activities published in different papers is difficult due to the different reaction conditions used. Bradford and Vannice [69] extrapolated the results of each author to the same conditions using kinetic equations and it was found that there are two magnitude differences between the calculated turnover frequencies. These values correlate with the space velocity, the lower the space velocity the lower the TOF value. This observation was explained by the fact that the space velocity, the residence time influence reverse and secondary reactions and these result in differences in TOF values. The other possibility of the great differences in the turnover number is the effect of chlorine. Most of the catalysts were prepared using RuCl₃ and Cl could be detected with XPS on the surface. The Ru/Al₂O₃ catalyst with the lowest TOF value was prepared with Ru₃(CO)₁₂ [25]. Finally, it was concluded that it seems possible that Cl could promote the dry reforming of methane over supported Ru by facilitating CH₄ activation [69]. The particle size also influenced the activity of Ru. Silica supported Ru with lower Ru loading performs a better activity in comparison to the sample with higher Ru content [73].

Ru/NaY activity was higher than that of Ru/SiO₂, but unlike Ru/SiO₂ did not deactivate during the reaction, since the charge compensation protons caused dealumination accompanied by the local destruction of the zeolite structure [74].

Three percent Ru/Al₂O₃ exhibited good stability during the 70 h exposure to undiluted equimolar CH₄/CO₂ gas stream at 1023 K but the deactivation of the catalyst was observed during temperature programmed activity tests in the examined range of operating conditions, which was caused by metallic cluster sintering. A general trend was observed, namely that increasing CH₄/CO₂ ratio in the feed stream positively influenced carbon deposition, which revealed CH₄ as the primary carbon source [62].

The catalytic activities of 5 wt% Ru supported on alumina, ceria, and ceria/alumina with different proportions of ceria were studied for the methane reforming reaction with CO₂ in the temperature range of 673–1073 K. Ru/Al₂O₃ was a stable catalyst presenting no coke deposition. The addition of ceria on alumina can enhance the activity of ruthenium-based catalyst by the presence of Ru–Ce interaction [75].

2% Ru supported on Mg and/or Al oxides including γ-Al₂O₃, MgAl₂O4, Mg₃(Al)O (prepared from Mg-Al layered double hydroxide), and MgO showed great differences in the catalytic performance for the CH₄ + CO₂ reaction [76]. The activities of Ru/MgO and Ru/Mg₃(Al)O were higher than those of Ru/MgAl₂O₄ and Ru/γ-Al₂O₃, which might be related to the strong base intensity of the support and to the more accessible surface Ru⁰ atoms, respectively. In a catalytic test at 1023 K for 30 h, Ru/Mg₃(Al)O showed excellent stability, while a significant deactivation was observed on Ru/γ-Al₂O₃, Ru/MgAl₂O₄, and Ru/MgO. It was demonstrated that the sintering of Ru particles was mainly responsible for the catalyst deactivation.

In contrast with it, it was found that the activity of 4% Ru/MgAl₂O₄ in the 723–1023 K temperature range was stable in the CO₂ + CH₄ reaction, although the carbon dioxide and methane ratio was 3:1 [77]. The Ru particle size was found to remain stable during repeated methane reforming experiments, thus preserving the nanostructure.

A 0.15% Ru/MgAl₂O₄ catalyst prepared by physical vapor deposition method exhibits 1 or 2 orders of magnitude higher activity than Ru catalysts prepared by incipient wetness impregnation [78]. This sample, instead of deactivation or coke deposition, demonstratec an activity increase of 1.5 times after reaction at 1123 K for 600 h. Ru initially was dispersed as isolated atoms on the support, then condensed into clusters (1.1 nm), and finally retained as faceted nanoparticles (2.7 nm) during the reaction, becoming more and more active by decreasing the barrier of breaking the C−H bond of methane.

Ru/La₂O₃ catalyst remained stable for more than 80 h in the 823–903 K temperature range when the CO₂/CH₄ ratio was equal to unity. However, a significant deactivation was observed at 823 K when the reaction was carried out in CO₂ rich stream. It was demonstrated by XPS that the partial re-oxidation of metallic Ru could be one of the factors that produces the catalyst deactivation [79]. In the case of Ru supported Al₂O₃ promoted with different oxides negative temperature dependence was observed, which was explained by the loss of structure of the solids and complete oxidation of ruthenium species at the reaction temperature [80].

Substitution of just 0.5 atom% of Ru in La₂O₃ (La_1.99Ru_0.01O₃) enhanced the conversion by 20 times when compared to the activity exhibited by La₂O₃ [81]. The oxygen storage capacity of the Ru doped sample was considerably higher than that of clean La₂O₃, which resulted in higher conversions of CH₄ and CO₂. When the Ru concentration was higher (1 atom% La_1.98Ru_0.02O₃), the activity decreased considerably, indicating the role of the ionic nature of substituted Ru at low dopant concentrations. DRIFTS studies demonstrated the role of a specific type of carbonates in promoting the activity of the catalyst [81].

It was shown that a very small amount of Ru (0.13 wt%) supported on ZrO₂-SiO₂ is very active and stable for the dry reforming of methane and the coke formation was suppressed at 1073 K while Ru/SiO₂ and ZrO₂-SiO₂ showed poor activity. When Ru was deposited at the SiO₂, Ru particle size was relatively large at 6.3 nm, but the Ru particles significantly decreased to 1.4 nm when Ru was deposited at the ZrO₂-SiO₂ support [82].

4.3. Supported Platinum

The reaction of CH₄ + CO₂ on Pt wire up to 1373 K produced mainly CO even at a CH₄/CO₂ ratio of 4.3, but if coke is also present on the Pt the dominant reaction becomes the pyrolysis of CH₄ to form C₂H₂ and C₆H₆ [83].

The activities of different supported Pt catalysts were initially similar at 875 K [84,85]. However, the Pt/γ-Al₂O₃ deactivated almost completely in the first 10 h of the reaction. Pt/TiO₂ lost 80% of its activity within 50 h, whereas the Pt/ZrO₂ showed only a 5% deactivation within 50 h on stream [84]. For all catalysts, the decrease in activity is caused by carbon formation, which blocks the active metal sites for reaction since the carbon formation rate decreased in the order Pt/γ-Al₂O₃ >> Pt/TiO₂ > Pt/ZrO₂ [85]. Evidences were presented that deactivation is caused by carbon formed on the metal and it is associated with overgrowth of the catalytic active perimeter between the support and the metal. Bradford and Vannice [86] also found that Pt/ZrO₂ and Pt/TiO₂ exhibited high stability even after 80 to 100 h on-stream although the Pt/SiO₂ and Pt/Cr₂O₃ catalysts deactivated significantly. The differences in the stability of the catalysts were explained in the case of Pt/SiO₂ and Pt/Cr₂O₃ by the apparent absence of any remarkable metal–support interaction which resulted in significant carbon deposition and so rapid deactivation. Pt on TiO₂ suppresses the carbon deposition, due to an ensemble effect of TiO_x species migrating onto the Pt surface. The low carbon deposition on Pt/ZrO₂ was explained by an interaction of Pt with Zrⁿ⁺ centers, the participation of lattice oxygen species at the metal–support interface, or by the partial coverage of Pt by ZrO_xspecies, as occurs with titania [86]. The same activity order was found by van Keulen et al. [36] and they interpreted the results in similar way, but the lower carbon deposit on Pt/ZrO₂ was explained that the carbon formed on the catalyst is gasified by reaction with CO₂.

The TOF value of CO₂ consumption at 925 K for Pt/ZrO₂, Pt/TiO₂, and Pt/γ-Al₂O₃ was two orders of magnitude higher compared to Pt-black or Pt/SiO₂ [87]. From these observations it was concluded that in supported Pt catalysts not all the surface Pt atoms contribute equally to the activity of the catalyst; Pt atoms on the support–metal perimeter determine the activity. This can be explained by the fact that the CO₂ activation occurs via carbonate species on the support that must be in the surroundings of the Pt particles to react with the methane which is activated on the Pt.

The catalytic efficiency of Pt/ZrO₂ in the CO₂ + CH₄ reaction was studied by different research groups [36,87,88,89,90,91,92,93,94]. It was observed that on Pt/ZrO₂ both CH₄ and CO₂ dissociate independently of each other in the CO₂ + CH₄ reaction [94]. The dissociation of carbon dioxide produces oxygen, while the decomposition products of methane react with the oxygen. When there is a large quantity of oxygen on the surface, the CH₄ can be oxidized to CO₂. CO₂ dissociation and its reverse reaction is also effected by the amount of surface oxygen

CO_2ads ⇌ CO_ads + O_ads.

(8)

The activity and stability of Pt/Al₂O₃ and Pt/ZrO₂ was compared in different papers [89,90,95,96,97]. The activity decrease for Pt/Al₂O₃ was rather slow and minor above 1070 K, while it was fast and almost complete at low temperature (< 875 K) [89,95]. Others found that Pt/Al₂O₃ deactivated significantly at 1073 K (deactivation rate was 4%/h) [90,97]. The cause of deactivation on Pt/Al₂O₃ was attributed to coke deposition on Pt. On the other hand, coke hardly deposits on Pt/ZrO₂ which resulted in the high stability of this sample. It was found that the reactivity of coke on Pt with hydrogen was higher on Pt/Al₂O₃ than on Pt/ZrO₂ but the reactivity of carbon with CO₂ was the opposite. These results and the observation that the methane decomposition is slower on Pt/ZrO₂ than on Pt/Al₂O₃, resulted in the small carbon deposition on Pt/ZrO₂ under reaction conditions [96]. The main difference between the two samples was found in the CO₂ dissociation. On Pt/Al₂O₃ this step was assisted by hydrogen, while on Pt/ZrO₂ the dissociation takes place over the support involving the oxygen vacancies.

The reduction of carbon deposition over zirconia-alumina catalyst was related to Pt-Zrⁿ⁺ interactions and to the high oxygen mobility on zirconia [90,97].

Ballarini et al. [93] also found that Pt/Al₂O₃ lost its activity at 1073 K due to the carbon deposition. Na or K addition to the catalyst enhances the stability since they provide basic sites, which promote the dissociation reaction of CO₂ into CO and O, and the O species can react with the surface carbon deposited on the Pt, thus cleaning the metallic phase of the catalysts. A similar effect can be achieved if at least 1% of CeO₂ is added to the Pt/Al₂O₃ catalyst [98] since CeO₂ promote the gasification of coke deposition, because the dispersity and thus the metal–support interfacial region will be higher.

Comparing the activity of Pt/ZrO₂ and Pt/SiO₂ it was found that the Pt/ZrO₂ was the best sample due to the ability of ZrO₂ to promote the CO₂ dissociation [99] which aids in the removal of carbon deposit from the metal. High reforming activity (CO₂ conversion was 96.4% at 1073 K) was observed for the Pt/ZrO₂/SiO₂ (4:1) sample among all the prepared ZrO₂/SiO₂ mixed oxides with different ratios (2:1 to 4:1) supported Pt [91]. The better catalytic properties of the Pt/ZrO₂/SiO₂ (4:1) were explained by higher Pt dispersion due to the absence of remaining free silica and the presence of higher amount amorphous ZrSiO₄.

The dry reforming of methane was compared on different alumina supported Pt catalysts [100]. It was observed that the morphological differences between the supports affect in general the catalyst performance. Pt is present as Pt⁰ on all catalysts, but the dispersion and stability was higher on Pt supported on a nanofibrous alumina, which presents a better behavior in terms of reactivity. It was found that CO₂ activation is promoted by the CH_X species, which probably initiates the reforming reaction.

Deactivation of Pt/Al₂O₃ was prevented by the addition of Pr-, Zr-, or Nb-doped CeO₂ to the Pt/Al₂O₃ catalyst [101]. The presence of dopants decreased the carbon accumulation on Pt surface. It was stated that during the reduction of the catalysts the Ce⁴⁺/Ce³⁺ ratio decreased on the surface—which was more significant for Pt/CePr/Al₂O₃—this resulted in the increase of oxygen storage/release capacity, which promotes the carbon removal from metallic surface.

High conversions and close to equal H₂/CO ratios were observed in the dry reforming of methane on 4% Ru/TiO₂ and 4% Pt/TiO₂ but on the Pt-Ru/TiO₂ bimetallic samples these values were lower than on monometallic catalysts [102]. Interestingly, the H₂/CO ratio for bimetallic catalysts decreased with increasing Pt content under the same conditions; for 3% Pt1% Ru/TiO₂ this ratio was below 0.2, while for Pt/TiO₂ it was found above 0.8.

4.4. Supported Palladium and Iridium

Nakagawa et al. [103] ranked the activity of the supported Ir catalysts as follows: TiO₂ ≥ ZrO₂ ≥ Y₂O₃ > La₂O₃ > MgO ≥ Al₂O₃ > SiO₂. Carbon deposition was observed only on Al₂O₃ and SiO₂ supported catalysts and this resulted in the low stability of the catalysts. The effect of the support was attributed to the activation of CO₂ on the metal oxides. Similar order was observed on 5% Ir supported on TiO₂, MgO, Al₂O₃, and SiO₂ when the turnover rate for H₂ formation was compared at 773 K [104]. In these cases, a significant amount of carbon formation on SiO₂ (C/Ir_surf = 1.95) and MgO (C/Ir_surf = 1.08) supported catalysts was observed after 1 h of reaction at 773 K, on Ir/Al₂O₃ and Ir/TiO₂ only traces of carbon were detected (C/Ir_surf ≤ 0.1).

On Ir/Al₂O₃ doped with both alkaline (Na and K) and alkaline earth metals (Ba, Ca and Mg) of different concentrations (1, 5 and 10 wt%) was studied the dry reforming of methane [105]. The catalytic performance was affected by acid–base functions of the support that modify the metal–support interaction. It was found that the higher the dopant concentration, the better the performance. The highest activity was obtained for 1% Ir/Al₂O₃-10% Mg catalyst.

The turnover rate for CH₄ consumption in the CH₄ + CO₂ reaction increase linearly with increasing CH₄ pressure at 873 K on 0.8 wt% Ir/ZrO₂, but it were not influenced by CO₂ concentration [106]. These results indicate that Ir surfaces remain essentially uncovered by CH₄-derived intermediates or by reaction products, because co-reactant or product concentrations would have otherwise influenced the forward rates. CH₄ consumption rates increased with increasing Ir dispersion, as it varied with metal content (0.2–1.6 wt%) or support (ZrO₂ or Al₂O₃). These data contradict with the results of Mark and Maier [17] since they observed that there is no effect of catalyst particle size on the methane conversion at 1073 K in the case of 1% Ir/Al₂O₃ catalyst.

Long-term experiments show the excellent stability of the Ir/Ce_0.9Gd_0.1O_2−x catalyst [107]. Carbon formation was totally inhibited or very limited under the most severe conditions of the study (1073 K, CH₄/CO₂ = 2).

Ir/Ce_0.9Pr_0.1O₂ catalysts prepared by different methods were also stable at 1023 K in the dry reforming reaction [108]. The highest activity of the sample synthesized by deposition–precipitation method was associated with the more enhanced basicity of the support, the more accessible Ir nanoparticles to transform CH₄ and the Ir-support boundary to remove the coke precursors.

The support significantly influenced the rate of the CO₂ + CH₄ reaction on the palladium catalyst. The specific activities of Pd/TiO₂ exceeded by almost one order of magnitude the TOF value for CO formation of the less active Pd/Al₂O₃ and Pd/SiO₂, and it was more than twenty times higher than on Pd/MgO [109]. It was suggested that the activity order of the Pd catalysts may be related to their efficiency in the dissociation of CO₂, where the same sequence of activity was found. It could mean that on supported palladium, the activation of carbon dioxide should play an important role in the CH₄ + CO₂ reaction. The high activity of Pd/TiO₂ may be associated with the extended electronic interaction between Pd and n-type TiO₂. The oxygen vacancy present on Pd titania interface could also promote the dissociation of carbon dioxide [109].

Pd/Al₂O₃ shows high initial activity but poor stability in the CO₂ + CH₄ reaction at 973 K due to Pd sintering and to a large amount of carbon deposition. Of the various promoters added to the catalyst (K, Ca, Y, Mn, and Cu), Y or Ca slightly increased the initial conversions, which can be attributed to the moderate reducibility of Pd species and to the increased Pd dispersion. In contrast, while the addition of K, Mn and Cu similarly reduced the reducibility of Pd species, the metal dispersions decreased owing to decoration effect or alloy formation [110]. Y-modified Pd/Al₂O₃ showed the best stability without activity loss in 20 h of reaction. This is probably because the addition of Y to the catalyst suppresses carbon formation and metal sintering and maintains Pd particle size below 10 nm. This particle size is crucial, below which carbon formation can be completely avoided under the reforming conditions.

A 0.86% Pd/CeO₂ catalyst showed activity in the dry reforming already at 623 K due to the presence of highly dispersed and very small Pd-nanoparticles (~2.9 nm), to metal–support interactions, and to the redox property of the support CeO₂ [111]. The high reducibility of CeO₂ (~32.8% Ce³⁺) influenced methane activation at as low as 623 K. Only a slight decrease in conversion was observed due to sintering, re-oxidation and hydroxylation of the active Pd-nanoparticles at 1073 K during 12 h of reaction. The promoting effect of Ce³⁺ was also observed when the activity of $1\% \mathrm{P}\mathrm{d}/\mathrm{M}\mathrm{g}{\mathrm{O}}_{1-\mathrm{x}}\mathrm{C}{\mathrm{e}}_{\mathrm{x}}^{3+}\mathrm{O}$ was found to be greater than that of $1\% \mathrm{P}\mathrm{d}/\mathrm{M}\mathrm{g}{\mathrm{O}}_{1-\mathrm{x}}\mathrm{C}{\mathrm{e}}_{\mathrm{x}}^{4+}\mathrm{O}$ in the reaction at 1173 K [112]. This effect was attributed to the smaller size of Ce⁴⁺.

5. Adsorption, Dissociation, and Activation of CO₂ on Noble Metals and on Supported Noble Metal Catalysts

It is known that CO₂ adsorbs on the surface of oxides used as catalyst support with the exception of some acidic oxides such as SiO₂. Spectroscopic analyses show that a variety of surface species can be formed, depending on the basicity of the surface, forming mixture of bidentate, monodentate, bridged carbonates, and bicarbonates [26]. In some cases, formate species were also detected after CO₂ adsorption. These observations indicate that hydrogen remained on the catalyst surface after reduction [26,113].

Any reaction of CO₂ requires its activation on the surface of the catalyst. There are different possibilities and therefore there are different theories for CO₂ activation on noble metals and on supported metal catalysts.

CO₂ could dissociate on metal surfaces to produce CO and adsorbed oxygen [114,115].

The reduction reaction could proceed via the formation of carbonate or formate species on the basic catalytic sites of the support [55].

The activation of CO₂ could occur on the interfacial sites of the catalysts, on the oxygen vacancies formed on the metal–oxide interface [49,88].

Another possibility for the activation of CO₂ is that the hydrogen formed in the decomposition of methane could promote the dissociation of CO₂, the formation of CO.

The adsorption and activation of CO₂ on clean noble metals has also been widely investigated [116,117]. It is generally accepted that CO₂ adsorbs weakly and molecularly on carefully cleaned noble metals at 100–300 K. In the case of Rh, there was a controversy concerning the nature of the interaction and particularly the dissociation of CO₂ on clean Rh surfaces. In the early studies, it was found that the adsorption of CO₂ is weak and associative on Rh metal [118]. The Somorjai group [119,120] reported that CO₂ dissociate on a Rh foil and on Rh single crystals, but Weinberg [121] and Goodman [122], on the basis of thermodynamic and kinetic results, indicated that the CO₂ sticking probability under the experimental conditions used would be by orders of magnitude smaller than needed to observe dissociation. It was demonstrated by field electron microscopy that the crystal structure of the Rh tip influenced the adsorption and dissociation of CO₂ [115], it was the fastest on {012} facets [114]. On Pt powder already at 293 K adsorbed CO was detected by DRIFTS indicating that CO₂ dissociation occurs [123].

Completely different results were obtained, when potassium or sodium adatoms, an electron donating promoter was deposited on Rh [124], Pt [125], Ru [126], and Pd [127]. After the low temperature (~100 K) CO₂ adsorption ${\mathrm{CO}}_2^{-}$ formed and it dissociates even at 150–200 K or can be transformed into carbonate-like species and CO, at a higher temperature.

The CO band appeared on the infrared spectra during CO₂ adsorption, or the metal is oxidized indicating the dissociation of CO₂ on supported noble metals on Rh [128], Pt [86], Ir [104], Ru [72,79], and Pd [109]. The preparation, in some cases the reduction temperature of the catalysts, the dispersity or the particle size of metal, the nature of the support, the adsorption temperature, and the CO₂ pressure all have an influence on this process.

The high temperature dissociation of CO₂ occurred on reduced alumina supported noble metals. The extent of CO₂ dissociation (the amount of CO formed related to the number of surface metal atoms) at 773 K was the largest for Ru and Rh, and the smallest for Pt and Ir [18]. On oxidized surfaces, dissociation of CO₂ was not observed. Not everyone has detected CO bands after or during CO₂ adsorption, that is, the dissociation of CO₂ on supported noble metals [129]. Adsorption of CO₂ on Ru/TiO₂ formation of formate occurs and carbonate species associated with the support but CO not up to 723 K whereas in the presence of H₂ part of these species is converted to Ru-bonded carbonyls via the reverse water–gas shift reaction at the metal–support interface [113].

$$\mathrm{C}{\mathrm{O}}_{2\left({\mathrm{TiO}}_2\right)}+{\mathrm{H}}_{\left(\mathrm{R}\mathrm{u}\right)}\to {(\mathrm{C}\mathrm{O}\mathrm{O}\mathrm{H})}_{\mathrm{R}\mathrm{u}-\mathrm{T}\mathrm{i}{\mathrm{O}}_2}\to \mathrm{C}{\mathrm{O}}_{\mathrm{R}\mathrm{u}}+{\mathrm{OH}}_{\mathrm{T}\mathrm{i}{\mathrm{O}}_2}$$

(9)

On Ru/SiO₂ linearly (2039 cm⁻¹) and bridge (1941 cm⁻¹) bonded CO, on Ru/Al₂O₃ only Ru-CO (2029 cm⁻¹) were detected by IR spectroscopy after CO₂ adsorption at 823 K [72]. On the IR spectra registered after CO₂ adsorption at 723 K on Pt/TiO₂, Pt/ZrO₂, Pt/Cr₂O₃, and Pt/SiO₂ bands characteristic for adsorbed CO (2060–2070 cm⁻¹ and 1860–1894 cm⁻¹) were detectable indicating the CO₂ dissociation [86], but the nondissociative adsorption of CO₂ was also observed.

(10)

The dissociation of CO₂ on Rh/TiO₂ depends on the reduction temperature, the higher the reduction temperature the greater the CO formation [130]. These observations were explained by the speculation that oxygen vacancies produced by the reduction of the catalysts promote the dissociation of CO₂. A similar explanation was given for the dissociation of CO₂ in the case of vanadia promoted Rh/SiO₂ [48], vanadia or titania promoted Rh/Al₂O₃ [59] and Pt/ZrO₂ [96]. The active participation of the support in CO₂ dissociation was suggested in the case of CeO₂ supported noble metals [131]. It was found that by increasing the reduction temperature, the reduction of the CeO₂ increased, and the amount of CO desorbed after CO₂ adsorption also increased. The CO₂ could be activated on Ce³⁺ sites with the formation of CO and Ce⁴⁺ [73,131].

M-CeO_2−x + x CO₂ → M-CeO₂ + x CO

(11)

CO₂ activation was related to the carbonate type species formed on the support [49,87]. On Pt/Al₂O₃ bicarbonates while on Pt/ZrO₂ bidentate carbonates were detected and the formation of latter species leads to easier activation of CO₂ than that of bicarbonates [95]. On Pt/SiO₂ did not form carbonate and on this sample CO was not detected after adsorption of CO₂ at 775 K [87].

At room temperature, CO₂ adsorbs on 1% Pt/10%ZrO₂/Al₂O₃, the peaks at 1646, 1444, and 1228 cm⁻¹ can be assigned to bidentate carbonate. While temperature increases, these species are activated and migrate to metal particles, being adsorbed as CO absorbing at about 2052 cm⁻¹ (Figure 4). Thus, CO₂ activation on zirconia catalysts involves support and interfacial sites [90]. This conclusion support the finding that CO₂ dissociate on PtZrO₂ at 775 K but adsorbed oxygen or Pt-O was not detectable by EXAFS. It was concluded that the oxygen is consumed by the support at the metal–support interface [49].

The presence of oxygen vacancies in the bulk will be an additional driving force for the reduction of CO₂.

XPS measurements show that the CO₂ oxidized Rh at 398 K on Rh/Al₂O₃ during the dissociative adsorption [132]

After the catalytic reaction at 903 K on Ru/La₂O₃ Ru⁰ and Ru³⁺ species were observed by XPS [79]. The oxidation of Ru supported on La₂O₃ by CO₂ was also found by Matsui et al. [68]. Pulsed reaction with ¹³CO₂ indicates that Ru could be oxidized to give ¹³CO; they proposed explaining this process by the following reaction step:

Ru/La₂O₃ + ¹³CO₂(ad) → ¹³CO(gas) + Ru-O_x/La₂O₃.

(12)

The direct participation of La₂O₃ was suggested in the CO formation from CO₂ in the case of La₂O₃ supported noble metals or La₂O₃ doped catalysts. The carbon dioxide rapidly reacts with La₂O₃ to generate oxycarbonate, which reacts slowly with carbon formed in the methane decomposition (Equation (3)) to generate CO [38,79,133,134].

CO₂ + La₂O₃ ⇌ La₂O₂CO₃

(13)

La₂O₂CO₃ + C–S → La₂O₃ + 2 CO + S (slow step)

(14)

In addition to the CO₂ dissociation induced by an oxygen vacancy on Pt/ZrO₂, O’Connor et al. [96] did not rule out the CO₂ dissociation through carbonate formation.

CO₂ + ☐_Zr → CO* + O_Zr

(15)

CO₂ǀ_Zr + OH_Zr → HCO₃ǀ_Zr

(16)

HCO₃ǀ_Zr + ☐_Zr + * → CO* + OH_Zr + O_Zr

(17)

☐_Zr means an oxygen vacancy at the zirconia surface and O_Zr is an surface oxygen species. It is most likely that these species are located at the Pt/zirconia interface.

Non dissociative adsorption of CO₂ during reaction occurs on the support, in the form of carbonates, as proved by FTIR of CO₂. Souza et al. [90] used in situ DRIFT experiments to prove that very weakly adsorbed carbonate species are formed, once the lattice oxygen of zirconia has been replenished by CO₂ which could decompose to CO.

$$2{\mathrm{CO}}_2+2\mathrm{z}\to \mathrm{C}{\mathrm{O}}_3^{2-}-\mathrm{z}+\mathrm{CO}-\mathrm{z}$$

(18)

$${\mathrm{CO}}_3^{2-}-\mathrm{z}+2\mathrm{z}\rightleftharpoons \mathrm{CO}-\mathrm{z}+2\mathrm{O}-\mathrm{z}$$

(19)

CO − z → CO + z (where z is a support site)

(20)

The role of CO₂ is frequently assumed to be only a source of oxygen to the support.

6. Interaction of Methane with the Noble Metals and Supported Noble Metal Catalysts

The earlier results of the adsorption and dissociation of CH₄ on metal films were summarized by Frennet [135]. The theoretical calculations also show that the noble metals are active in methane dissociation [136]. It was found that the dissociation enthalpy of CH₄ is an important factor determining the catalytic activity of the metal. According to the trend in the calculated enthalpies, the most efficient catalysts for methane dissociation were Rh, Ru, and Ir are more active than Pd and Pt. These results are in agreement with experimental observations. Methane activation studies on single crystal surfaces have shown that the activation of C-H bonds depends sensitively not only on the metals but also on surface structure; specifically, surface steps and kinks are much more active than atoms on close-packed surfaces [137].

The dissociation, the activation of methane and its interaction with the catalyst have been studied separately on various supported noble metal catalysts. Treating Pt/SiO₂ [138] or different noble metals [39] with methane the dissociation of it was observed above 473 K, hydrogen and surprisingly ethane were formed. However, the activity of the catalysts quickly dropped to a low value, and after 15–25 min only traces of H₂ and C₂H₆ were detected. This suggests an accumulation of carbon deposit on the catalysts during this process poisoning the reaction. The rate of the ethane evolution passed through a maximum. The amount of hydrogen greatly exceeded that of ethane on all catalyst samples. The most active sample was Rh, but the largest amount of ethane was measured on Pt/SiO₂ [139]. The formation of ethane has been explained by the dissociation of methane on the surface and the resulting methyl group could dimerize to ethane and could decompose further to produce carbonaceous deposit.

CH₄(g) ⇌ CH₃(a) + H(a)

(21)

2 CH₃(a) ⇌ C₂H₆(g)

(22)

The support influenced the hydrogen and ethane formation in the decomposition of methane. The most effective catalyst in the hydrogen formation was Pd/TiO₂ [140], but among the Rh catalysts, the alumina supported sample was the best followed by Rh/TiO₂, Rh/SiO₂, and Rh/MgO [46]. The effect of the supports on the CH₄ dissociation was explained by the different crystal size of the catalysts and/or with the ease of the carbon migration from the metal to the support [46,140]. The highest C₂H₆ formation was observed in all cases on silica supported samples.

It is generally assumed that the CH₄ dissociation occurs on metallic sites but on Rh/Al₂O₃ and Rh/ZrO₂ it was found that the CH₄ activation on Rh⁰ would be better conducted when RhO_x sites are also on the surface, probably via oxygen assisted dissociation [141].

As the Equation (21) shows CH₄ adsorbs reversibly on the surface of noble metals arriving at the equilibrium. This conclusion is supported by the results of steady state isotopic tracing kinetic analysis (SSITKA) which detected methane on the surface of Pd/ZrO₂ and Pd/γ-Al₂O₃ under reaction conditions [142]. It was found that the rate constant did not depend on the dispersity of Pd, but others reported that the CH₄ activation is structure sensitive [72,143], and influenced by metal dispersion of different noble metals [17,58,144,145,146]. A direct correlation was found between the Rh dispersion and the turnover rate of the reaction [142,143,146].

The question is whether the surface carbon produced at high temperature from methane decomposition contains any hydrogen or not? By in situ IR spectroscopy on supported Rh could not detect CH_x intermediates [46], and when the carbon was reacted with D₂ only CD₄ was formed [147]. Similar was observed on Ru/Al₂O₃ but on Ru/La₂O₃ CH₂D₂, on Ru/Y₂O₃ CH₃D, CH₂D₂, and CHD₃ were also detected [148].

Several manuscripts have concluded from the kinetic parameters of the reaction, from the dependence of the reaction rate on the particle size that methane dissociates on the supported noble metals. The activation and dissociation of methane can take place not only directly on the metals but also on the adsorbed O species [136,141,149], which may be formed from the dissociation of CO₂ [18], furthermore the oxygen in the support oxide [8,134], or the surface OH groups [71] can also facilitate this process.

Pakhare et al. [8,134] found by means of CH₄ TPR that the lattice oxygen on the surface of Ru or Pt containing pyrochlors (La₂Zr₂O₇) help react with CH₄. The oxygen or OH group can promote the dissociation of CH₄ by reacting with the CH_x groups formed in the first step.

Bradford and Vannice [26] detected by IR spectroscopy H₂C = O (at 1690 cm⁻¹) species formation after CH₄ adsorption on TiO₂ supported noble metals. In addition, change in the intensity of OH groups was observed and on Rh and on Pt adsorbed CO were identified.

7. Surface Species Formed in the Interaction and in the Reaction of CO₂ with CH₄ on Supported Noble Metals

Examination of the interaction of reactants and products formed in the reaction on the surface helps to understand the whole process.

DRIFT spectra registered during co-adsorption of CO₂ + CH₄ gas mixture on Pt powder at 293 K show a marked increase in the intensity of the absorption band characteristic for adsorbed CO compared to CO₂ adsorption alone, implying that CH₄ promotes the dissociation of CO₂ [123], as it was observed on supported Rh [25,46,48], Pd [109] and Ir [25,104]. When the temperature was raised the intensity of the CO band increased, and its position shifted to higher wavenumbers. The spectral feature of CO formed in the co-adsorption differed from that observed during the adsorption of gaseous CO. For example, in the case of supported Rh the doublet due to the twin structure (Rh(CO)₂) at 2096 and 2029 cm⁻¹ [150] was missing and the other CO bands, characteristic for linear and bridge bonded CO, absorbed at lower frequencies. The promotion of the dissociation of carbon dioxide is attributed to the effect of hydrogen—as it was found in the interaction of CO₂ + H₂ mixture [128,151]—formed in the decomposition of methane.

After the adsorption of CO₂ + CH₄ gas mixture at room temperature on Al₂O₃ supported noble metals mainly hydrogen carbonate were found on the IR spectra bands (1650, 1443, 1348, and 1302 cm⁻¹) [59]. With increasing the adsorption temperature the intensities of these bands decreased and above 548 K a band characteristic for linearly bonded CO was detected (Figure 5) [59]. Above 673 K a band was observed at about 1589 cm⁻¹ characteristic for the O–C–O vibration (ν_ass) of formate groups adsorbed on alumina [150]. On V₂O₅ or TiO₂ doped alumina supported Rh the CO band appeared at lower temperature (Figure 5).

Similar spectra were recorded during the reaction of CO₂ + CH₄ at 723 K [147] or at 773 K [152]. It was observed that the intensity of formate band increased continuously during the reaction. It is surprising to observe adsorbed CO and formate species on the surface well above their desorption temperature [57,152]. It means that the formation rate of these form is higher than their desorption, decomposition or further reaction rate. This is supported by the observation that flushing the DRIFT cell with He after the catalytic reaction all bands disappeared immediately from the spectra [55].

The linearly bonded CO formed in the reaction of CO₂ + CH₄ similarly as was shown in the low temperature interaction of CO₂ + CH₄ absorbs at lower frequencies (~2030–2020 cm⁻¹) than the CO formed after gaseous CO adsorption. This effect was attributed to Rh carbonyl hydride formation [128,129,130].

On Rh/La₂O₃ CO bands (at 2036 and 1825 cm⁻¹) and absorptions characteristic for different La₂O₂CO₃ (at 720, 842, 864, 1060 and 1550 cm⁻¹) were observed during the exposure of the reaction mixture onto the catalyst at and above 773 K [34]. La₂O₂CO₃ formation was detected in all cases when La₂O₃ supported catalysts were used [153].

In the presentation of the activity of the catalysts, it was often mentioned that the significant decrease in activity in time is caused by carbon deposition. There are several theoretical possibilities for the formation of carbon: Decomposition of methane (3), the Boudouard reaction (6), and the hydrogenation of CO (the reverse reaction of 7).

On theoretical calculation, it was found that with increasing the particle size increased the carbon formation. At low temperature CO was identified as the precursor for carbon, but CH₄ was found to be the main source of carbon at high temperature [154]. Transient isotopic experiments indicate [43,55,152,155] that on Rh/Al₂O₃ during the reforming reaction accumulated carbon is mainly derived from the CO₂ molecule, since after the ¹³CH₄ + ¹²CO₂ reaction on Rh/Al₂O₃ in the temperature programmed oxidation (TPO) mainly ¹²CO₂ was formed [44,55].

CO₂ + 2 s → CO-s + O-s

(23)

CO-s + s ⇌ C-s + O-s

(24)

where (s) is a site on the Rh surface or on the metal support interface.

In the case of Pt/ZrO₂ the main fraction of carbonaceous deposit was formed from the dissociation of CH₄ the preferred species was CH₂* [156].

Graphitic carbon was observed on the Pt/TiO₂ and Ru/TiO₂ [102] after the reaction by TEM but on Pt/MgO-Al₂O₃ [80], on Ru supported modified SiO₂ [73], on Rh/α-Al₂O₃-La₂O₃ [38] filamentous carbon was found. According to experience the extent of carbon deposition on noble metals is usually lower than on Ni because carbon atoms can dissolve in Ni crystal and carbon fibers could form. This mechanism does not work on noble metals [157]. Raman spectra of the used Pt/CeO₂-Al₂O₃ and Ru/MgO-Al₂O₃ catalysts [80] especially after the reaction at 923 K, on Rh/La₂O₃ [153], on different Pd/Al₂O₃ [110], on Pt/ZrO₂/SiO₂ [91] show intense bands at 1580–1595 and 1339–1350 cm⁻¹ which could assignation to the so called G and D band, respectively. The G band is related to the graphitic carbon and the D band to the disordered structure [80]. The detection of graphitic carbon by Raman spectroscopy was successful even when with TGA no carbon formation was detected [153].

The formation of graphitic carbon was detected on different Pd/Al₂O₃ [110], on Ru/SiO₂ [73] by using XRD spectroscopy.

Examination of the reactivity of the surface deposited carbon by temperature programmed hydrogenation (TPH) or oxidation as well as TEM images showed that carbon can occur in different forms on the surface, which can be transformed into each other depending on the temperature and the time elapsed since their formation. The reactivity of the surface carbon was almost the same regardless of whether it was formed either from the decomposition of CH₄ or from the disproportionation of CO or from the reaction of CO₂ + CH₄. The gasification of surface carbon deposit is an important step enhancing the stability of the catalysts.

The reactivity of the surface carbon formed from CH₄ decomposition at 523 K, studied by TPH [46,63], exhibited practically the same features as those established for supported Rh catalysts following the dissociation of CO [57]. Three different carbon forms were distinguished: (i) The highly reactive form (α), which can be hydrogenated even below 350 K; (ii) a less reactive deposit (β), with T_Max = 500 K; and (iii) the relatively inactive form (γ), which reacts with hydrogen only above 650 K. Similar TPH spectra were registered on different RhCu/Al₂O₃ after decomposition of CH₄ at 623 K [158]. The reactivity of surface carbon, that is, the relative ratio of the three forms, is very sensitive to the temperature of its formation and also to the duration of its thermal treatment. Above 623 K, a significant aging was observed, with transformation of the more reactive form into less reactive ones. When decomposition of CH₄ occurs at 923 K a relatively inactive form of carbon was obtained (T_Max = 973 K), suggesting that aging of the carbon produced occurs [155].

C_α-s → C_β-s → C_γ-s

(25)

These results supported by the observation that in the case of Ru/Al₂O₃ and Ru/La₂O₃ the TPH peak shifted to higher temperature as a function of the CH₄ decomposition temperature [148]. On Ru/Al₂O₃ there is a carbon form which did not react with H₂ up to 1073 K.

TPH for Pt/Al₂O₃ after exposure to CH₄ at 1070 K showed a peak above 873 K, while on Pt/ZrO₂ the CH₄ formation was observed above 1073 K. These indicate that some of the surface carbon on Pt/Al₂O₃ is more reactive with H₂ than on Pt/ZrO₂ [95], but some of the coke could not remove by H₂ even at 1273 K.

The TPH profiles of the used Ru/SiO₂ [73], or that of different noble metals supported on spinell [30] catalysts clearly showed the existence of two types of carbonaceous species on the catalysts. A first peak is observed at temperatures between 620 and 680 K on different catalysts depending on the metal. This peak is related to amorphous carbon located at the interior of the active metal particles. A second peak is observed at temperatures above 873 K, identified as whisker-like filamentous carbon. The THP of 0.5% Rh/Al₂O₃ used in the CO₂ + CH₄ reaction at 923 K for 10 min showed three well resolved peaks. The most reactive hydrogenated at low temperature (373 K), the most abundant reacted at 423–593 K, and the least active carbon form above this temperature [43,155].

The TPO profile of used noble metal catalysts supported on MgAl₂O₄ depends on the metals. Pt and Ir showed one peak in the temperature range of 573–700 K which was related to the oxidation of amorphous carbon. In addition to this form for the other samples a second peak was also detected; in the case of Pd it was found at higher temperature above 873 K, due to the formation of whisker carbon. On Ru and Rh supported on spinel a TPO peak was observed at about 373 K which could be related to (C_α) carbide like structure [30]. In contrast on Ru/SiO₂ TPO peak characteristic for whisker was observed and the carbide like species was missing [73].

After CO₂ + CH₄ reaction at 923 K four TPO peak were registered on 0.5% Rh/Al₂O₃, the T_Max were at 373, 453, 543, and 603 K [43] similarly to on 0.5% Rh/La₂O₃ [30]. Ghelamallah and Grangers found that on Rh/La₂O₃ the first TPO peak appeared below 523 K and it was assigned to carbon deposited on metal, the second (635–661 K) type of coke is on the metal support interface, and the third one (693–798 K) accumulated on the support, and the most inactive carbon (above 973 K) was identified as graphite [159].

TPO profiles of carbon deposited on Pt during the reaction at 1123 K showed a main peak at 893 K, which is typical for the oxidation of carbon with a nanotube structure. Only a small peak was observed between 613 and 623 K, attributed to oligomeric and polymeric carbon. This indicates that, regardless of the source of carbon (CH₄ or CO₂), the majority of the carbon species is present on the catalyst surface in the form of carbon nanotubes [156]. Similar TPO curves were found on different Pd/Al₂O₃ [110].

With increasing the reaction temperature on Ru substituted La₂Zr₂O₇ the amount of carbon deposit decreased although the TPO peaks shifted to higher temperature [8]. This means that the reactivity of deposited carbon decreases with increasing the reaction temperature.

On Rh/Al₂O₃ the carbon produced in CH₄ decomposition at 773 K react with oxygen above 400 K but up to 570 K small amount of CO₂ was detected. The main CO₂ peak was observed at 745 K [46].

The reactivity of surface carbon was investigated by temperature programmed reaction with CO₂ (the reverse Boudouard reaction (6)). On Rh/Al₂O₃ the carbon—produced at 773 K in the decomposition of CH₄—reacted with CO₂ above 550 K and the reaction proceeded at a higher rate above 700 K [46]. Similar results were obtained on Pt/Al₂O₃ [89,95] peak was observed below and above the reaction temperature and these were assigned to carbon on Pt and on Al₂O₃, respectively. For Pt/ZrO₂ also two peaks were registered but both react below reaction temperature and these were attributed to coke on Pt [89,95].

8. The Proposed Mechanism for the CO₂ + CH₄ Reaction on Supported Noble Metals Catalysts

In the previous chapters the adsorption, dissociation, and activation of CO₂ and CH₄ on supported noble metals were discussed, we presented the different experimental results and the processes proposed based on them. It means there is still some debate on the reaction mechanism of this reaction. Regarding the reaction mechanism for CO₂ + CH₄ reaction basically three different, but same in a few steps’ reaction mechanisms were proposed. The first involves the adsorption of CH₄, the second the activation or dissociation of CO₂ and the third one when both component dissociate in the first step. Further differences between the reaction mechanisms are that the dissociation of CH₄ or CO₂ takes place directly or in some reaction, and that this process takes place on the metal, on the support, or at the interface.

The mechanisms of the reaction could be divided into two groups: Mono- and bifunctional mechanism. According to the monofunctional mechanism the activation of both CO₂ and CH₄ occurs on the metal sites, but in the latter cases the CO₂ is activated on the support or on the metal–support interface and the CH₄ on the metal.

A dissociative methane and carbon dioxide adsorption is proposed by Rostrup-Nielsen and Bak-Hansen [19] on different alumina stabilized magnesia supported noble metals

CH₄ + * = CH₃* → CH₂* → CH* = C

(26)

CO₂ + * = O* + CO

(27)

CH_x* + O* = CO + x/2 H₂ + 2*

(28)

In the case of supported Rh and Pd catalyst, although the dissociation of CO₂ has been shown on these samples, it was suggested that during the reaction, the main route for CO formation is the hydrogen assisted CO₂ dissociation [46,109].

CH₄ = CH_3(a) + H_(a)

(29)

CO₂ +H_(a) = CO + OH_(a)

(30)

CH₄ + O_(a) = CH_3(a) +OH_(a)

(31)

CH_3(a) = CH_2(a) + H_(a)

(32)

CH_2(a) = CH_(a) + H_(a)

(33)

CH_x(a) = C + xH_(a)

(34)

CH_x + O_(a) = CO + xH_(a)

(35)

CH_x(a) + CO₂ = 2 CO + xH_(a)

(36)

2 H_(a) =H_2(g)

(37)

2 OH_(a) = H₂O + O_(a)

In the case of Rh/Al₂O₃ the reverse Boudouard reaction (6) and the

C + H₂O = CO + H₂

(38)

reaction was also confirmed [46].

Raskó and Solymosi showed that the adsorbed CH₃ radical and CO₂ react with each other at the metal–substrate interface on Rh/SiO₂. CH₃ groups adsorbed on the support migrate from the support to the metal [160].

Direct dissociation of CO₂ [55,71,149] and CO [55,149] and the further reaction of adsorbed oxygen, OH groups were not considered in former mechanism [55,58,149] and therefore it was completed by the following reactions:

CO_2(a) → CO_(a) + O_(a)

(39)

CO_(a) → C + O_(a)

(40)

O_(a) + H_(a) → OH_(a)

(41)

CH_(a) + OH_(a) ⇌ CO_(a) + H₂

(42)

C + OH_(a) ⇌ CO_(a) + H_(a).

(43)

Mark and Maier [17] proposed a mechanism for the reaction on Rh/Al₂O₃ where the dissociative CH₄ adsorption is rate limiting step, followed by the direct reaction of CO₂ with adsorbed carbon. This reaction mechanism was also confirmed on Ru [147,161] and Ir catalysts [147].

Bifunctional mechanisms in which methane is activated on the metal and carbon dioxide on the support have been proposed for Pt metals dispersed on supports, such as ZrO₂ [49,86,87,90,94], TiO₂ [50,86], or Al₂O₃ [50,71].

The following bifunctional mechanism (where p is a platinum site and z a support site) was proposed by Souza et al. [90] for CH₄ + CO₂ reforming:

$${\mathrm{CH}}_4+\mathrm{p}\stackrel{{\mathrm{k}}_{{\mathrm{CH}}_4}}{\iff }{\mathrm{CH}}_{\mathrm{x}}-\mathrm{p}+\left(\frac{4-\mathrm{x}}{2}\right){\mathrm{H}}_2$$

(44)

$${\mathrm{CO}}_2+\mathrm{z}\stackrel{{\mathrm{K}}_{{\mathrm{CO}}_2}}{\leftrightarrow }{\mathrm{CO}}_2-\mathrm{z}$$

(45)

$$2{\mathrm{CO}}_2+2\mathrm{z}\to {\mathrm{CO}}_3^{2-}-\mathrm{z}+\mathrm{CO}-\mathrm{z}$$

(46)

$${\mathrm{CO}}_3^{2-}-\mathrm{z}+2\mathrm{z}\stackrel{ }{\iff }\mathrm{CO}-\mathrm{z}+2\mathrm{O}-\mathrm{z}$$

(47)

CO − z → CO + z

(48)

$${\mathrm{H}}_2+2\mathrm{p}\stackrel{}{\leftrightarrow }2\mathrm{H}-\mathrm{p}$$

(49)

$${\mathrm{CO}}_2-\mathrm{z}+2\mathrm{H}-\mathrm{p}\stackrel{}{\leftrightarrow }\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}$$

(50)

$${\mathrm{CO}}_2-\mathrm{z}+2{\mathrm{CH}}_{\mathrm{x}}-\mathrm{p}\stackrel{\mathrm{k}}{\to }2\mathrm{CO}+\frac{\mathrm{x}}{2}{\mathrm{H}}_2+\mathrm{z}+\mathrm{p}$$

(51)

The occurrence of the bifunctional mechanism is supported by the observation that when TiO₂, Al₂O₃, or MgO was mixed with the Rh/SiO₂ catalyst, the conversion increased significantly, but when mixed with SiO₂ it did not change [46]. This was explained by the fact that neither carbonates nor formate are formed on SiO₂ which forms may be responsible for the increase of reaction rate.

Bitter et al. found [49] that the activity of Pt/ZrO₂ increased linearly with increasing the Pt-ZrO₂ perimeter, since the carbonate on the interface is reduced by the CH_x species on the metal to form CO and formate. The formate species then decompose to CO and OH and subsequently the OH groups can either be desorbed as water or react with CH₄ to form CO and H₂.

On La₂O₃ [34,38,68,79,86,134] supported catalysts it was suggested that the La₂O₂CO₃, formed in the adsorption of CO₂ (Equation (13)), react with the surface carbon (Equation (14)) formed in the decomposition of methane.

We have seen that in recent years, many ideas have been developed for the CO₂ + CH₄ reaction, but to date there is no clear explanation for what and why influences CO₂ dissociation, what kind of secondary reactions take place on the surface of noble metal catalysts.

9. Conclusions

This review summarizes the studies on the CO₂ + CH₄ reaction, the dry reforming of CH₄ on supported noble metal catalysts. Comparing the activity and stability of different supported metals, it was found that the best catalysts are among the noble metals, Rh and Ru. However, these metals are expensive and so to make the process uneconomical, but noble metals can be used to promote the cheaper Ni or Co catalyst which could help in removal of carbon deposit, thus increasing the stability of the catalyst.

Comparing the activity of noble metals and the effect of carriers, there are significant differences in the literature. The acidity, basicity and the reducibility of the supports basically influence the efficiency of the catalyst. The oxygen vacancies formed on the metal–support interface improve the activity as these sites promote CO₂ activation but the SMSI interaction reduces it. Alkaline or alkaline earth metal additive increases CO₂ dissociation and thus the efficiency of the catalysts

The possibility of the activation and dissociation of CO₂ and CH₄ on clean metals and on supported noble metals were discussed separately. There are different possibilities and therefore there are different theories for CO₂ dissociation or activation. CO₂ could dissociate on metal surfaces, this reaction could proceed via the formation of carbonate or formate species on the acidic or basic sites of the support, or on the metal–support interface. In the reaction the hydrogen assisted dissociation of CO₂ was also suggested. It is generally accepted that CH₄ decomposes primarily on the metal thus having appropriate dispersion of the metal is important. The activation and dissociation of methane can take place also on the surface O species. The main problem of the reaction is the carbon deposition and the reactivity of the surface coke. Surface carbon can be formed from methane while others suggest that the source of the surface carbon is CO₂. Carbon can occur in different forms on the surface, which can be transformed into each other depending on the temperature and the time elapsed since their formation. These processes are influenced by, among others, the metal, the metal particle size, the support and therefore significant differences have been found in the order of activity of the catalysts, the carbon deposition, and in the reactivity of surface carbon.

Basically, two reaction mechanisms were proposed, according to the mono-functional mechanism the activation of both CO₂ and CH₄ occurs on the metal sites because the support is completely inert. In the bi-functional mechanism the CO₂ is activated on the support or on the metal–support interface and the CH₄ on the metal and the reaction intermediates react on the interface.