Bimetallic Co-Rh Systems as a Prospective Base for Design of CH 4 Reforming Catalysts to Produce Syngas with a Controllable Composition

: Dry and bireforming (CO 2 -H 2 O) of methane are the most environmentally friendly routes involving two main greenhouse gases to produce syngas—an important building block for large-scale production of various commodity chemicals. The main drawback preventing their industrial application is the coke formation. Developing catalysts that do not favour or are resistant to coke formation is the only way to improve the catalyst stability. Designing an economically viable catalyst may be achieved by exploiting the synergic effects of combining noble (expensive but coke-resistant) and non-noble (cheap but prone to carbonisation) metals to form highly effective catalysts. This work deals with development of highly active and stable bimetallic Co-containing catalysts modiﬁed with small amount of Rh, 0.1–0.5 mass %. The catalysts were characterised by BET, XRD, TEM, SEM, XPS, and TPR-H 2 methods and tested in dry, bi-, and for comparison in steam reforming of methane. It was revealed that the bimetallic Co-Rh systems is much more effective than monometallic ones due to Co-Rh interaction accompanied with increasing dispersion and reducibility of Co. The extents of CH 4 and CO 2 conversion over the 5%Co-Rh/Al 2 O 3 are varied within 85–99%. Syngas with variable H 2 /CO = 0.9–3.9 was formed. No loss of activity was observed for 100 h of long-term stability test.


Introduction
Syngas, a mixture of H 2 and CO in any ratios has been considered as a potential and promising sustainable energy from an economic and environmental point of view [1,2]. Syngas is an intermediate in the multi-tonnage production of fuels and chemicals and can be produced by three methane reforming processes, including steam (SRM, Equation (1)) or dry reforming (DRM, Equation (2)) and partial oxidation, collectively named in literature as oxyforming processes [3], or also by any combination of these reactions-bireforming (BRM), tri-reforming, authothermal reforming of methane. CH 4 + H 2 O ↔ CO + 3H 2 ∆H = 206 kJ/mol (1) CH 4 + CO 2 ↔ 2CO + 2H 2 ∆H = 247 kJ/mol (2) Among these methods for syngas production, the DRM [4,5] supplies a clean and cheap oxygen source derived from CO 2 [6]. Biogas mainly composed of carbon dioxide and methane is a cheap renewable source for DRM [7]. Also, some reservoirs of natural gas can contain comparable concentrations of CO 2 and methane, which makes the reaction more desirable for commercialization [8]. The main benefit of DRM is its potential to diminish The main drawback for all the methane reforming processes is the catalyst deactivation due to the surface coke formation as a result of methane decomposition (Equation (5)) and Boudouard's reaction (Equation (6)) [17] and thermal sintering of the active sites under high operation temperature [5].
CH 4 ↔ C + 2H 2 ∆H = 74.9 kJ/mol (5) 2CO ↔ CO 2 + C ∆H = −172.5 kJ/mol (6) In terms of catalysts reported to be efficient for the DRM and SRM reactions, one of the most commonly reported involves nickel either directly or supported. Nickel-based catalysts are widely studied due to their low price and comparable activity with noble metal catalysts. However, the key issue of the Ni-based catalyst is the carbon deposition.
Recently supported Co catalysts began to be actively investigated in the reforming of methane due to their low cost and availability on a larger scale [4,5,[18][19][20][21]. It has been found that Co−based catalysts can be used as appropriate active metals with sufficient activity for the DRM reaction [5,22]. These catalysts have higher resistance to carbon deposition than Ni-based and can control the rate of carbon formation by oxidizing the surface carbon [19,20]. However, in addition to coke formation, the Co−based catalysts are prone to deactivation due to the oxidation of metallic Co, as well as the formation of inactive spinel structures, which lead to a reduction in the number of active sites [4,5,18,19,22].
From most of the previous research, transition metals of the VIII group except for the osmium, especially rhodium and ruthenium, are excellent reforming catalysts demonstrating both high catalytic activity and resistance to carbon formation and oxidation [21,[23][24][25][26][27]. Unfortunately, there are strong constraints, such as the low reserves and high cost of these metals, preventing the industrial application of these metals as catalysts. Nevertheless, noble metals are good candidates to be used in a small amount to promote the catalyst's resistance to carbonization.
It has been reported that introducing a second metal component to catalyst system forming bimetallic alloy sites could also improve the anti-coking property of monometallic catalysts [28]. The addition of another metal to form bimetallic Co-based catalysts could be the crucial technique to develop the effective stable DRM catalyst. Thus, it has been reported that the addition of an appropriate amount of Zn produced a noticeable improvement in the catalytic performance in DRM [5]. The enhanced stability of the Ru-promoted Co/α-Al 2 O 3 catalyst in DRM was observed by authors [29], which was caused due to facile removal of the deposited coke and suppression of metal oxidation. The authors concluded that the Ru promoter provided active sites during DRM, and structural stabilization of Co species [29]. In our previous works, the same effect of the second noble metal like Pt [30] and Pd [31] on the performance of Co-based catalysts in DRM and bireforming of methane were observed too. By authors [20] it has been found that the 10%Co/Al 2 O 3 with additives of 0.1-0.5% Rh is more active in the CO 2 + CH 4 reaction than the monometallic Co/Al 2 O 3 . Besides DRM, the Co-Rh bimetallic catalysts may be applied in other processes, for example, they are also active in ethanol conversion [26,32].
The design of cost-effective, efficient, and robust (coking-and sintering-resistant) DRM catalysts is a grand challenge in the topic. Bimetallic catalysts, providing synergistic effects via metal-to-metal interactions seems to be an effective strategy for achieving these goals [9,11,28,33,34].
Since, among the vast number of synthesized and widely studied catalysts, only noble metals have a noticeable resistance to coke formation, it is advisable to add them to the catalyst composition. Their high cost can be offset by their long service life, in addition, precious metals can be removed from the spent catalyst for reuse. The appropriate pair of noble-nonnoble metals with possible synergetic effect should be chosen. Cobalt is a prospective candidate to be used as a nonnoble metal because studies of the CH 4 reforming on Co-containing catalysts are far less in number than those on Ni or noble metal catalysts.
On the basis of literature [13,20,23,26,[35][36][37][38] and own studies [30,31,39], the bimetallic Co-Rh system was selected as a promising base for DRM catalysts. In contrast to the publication of the authors [20], in this work, the Co-Rh catalysts were studied with a lower amount of cobalt, 10 vs. 5 mass %, respectively while the amount of rhodium additive was the same-0.1-0.5 mass %. The synthesized catalysts were studied not only in the dry methane reforming reaction, but also in bireforming and steam methane reforming to elucidate their performance in presence of water with an aim to enhance hydrogen generation. In addition, the stability of the catalysts was investigated for 60-200 h. The effects of the reforming temperature, the amounts of the second metal-Rh, type of process, and steam amount on the conversion of initial reagents and yield and composition of syngas over the 5%Co-Rh/Al 2 O 3 catalysts were investigated. To compare and elucidate the effect of Rh on Co, the data on monometallic Co/Al 2 O 3 and Rh/Al 2 O 3 studied by us earlier are also included.

Catalyst Characterisation
The BET surface areas of the fresh (before reaction, no any preliminary treatment) 5%Co-Rh/Al 2 O 3 catalyst samples are varied within 144.5-156.5 m 2 /g depending on the Rh content (Table 1). After exploitation in methane reforming for 25-200 h, the specific surface areas of the spent catalyst were slightly decreased up to 126.8-147.6 m 2 /g. By X-ray analysis, no phases except γ-alumina were observed for all the Co-Rhbased catalysts including the fresh and spent samples. There are no reflexes related to metal phases. In Figure 1, almost the same XRD patterns are presented for the 5%Co-  (Figure 1a) and γ-alumina (Figure 1b)  By X-ray analysis, no phases except γ-alumina were observed for all the Co-Rh-based catalysts including the fresh and spent samples. There are no reflexes related to metal phases. In Figure 1, almost the same XRD patterns are presented for the 5%Co-Rh(98:2)/Al2O3 ( Figure 1a) and γ-alumina (Figure 1b)   The TEM study of a 'fresh' (preliminary reduced by H2, 300 °C and then passivated in Ar flow) sample of the 5%Со-Rh(9:1)/Al2O3 catalyst showed the highly dispersed and uniformly distributed spherical particles with the average size of 1.5-2 nm, microdiffraction analysis of which revealed the existence of metallic states both of Rh 0 and Co 0 (Figure 2a). The TEM study of a 'fresh' (preliminary reduced by H 2 , 300 • C and then passivated in Ar flow) sample of the 5%Co-Rh(9:1)/Al 2 O 3 catalyst showed the highly dispersed and uniformly distributed spherical particles with the average size of 1.5-2 nm, microdiffraction analysis of which revealed the existence of metallic states both of Rh 0 and Co 0 (Figure 2a). Rarely, the particles with a bigger size of 2.5 nm were observed. The last were presumably attributed to the oxidized state of metals; because microdiffraction was presented by small amounts of weak rings, their identification was difficult. The presence of metal oxides is not a surprise because the reduced and passivated sample has been undergone to air. It is notable that the reduced state of Co 0 in the bimetallic catalyst was kept in an air environment. The Co 0 state in a monometallic Со/Al2O3 catalysts with Co content of 2.5 and 5 mass % was not observed after passivation [31,39].
In a spent sample (after reaction and subsequent passivation by the Ar flow), the particles retained their spherical shape, but some of them increased in size varied within 1.5-3 nm (Figure 2b). Microdiffraction analysis showed the presence of Co 0 and two oxidized states-RhO2 (ASTM, 21-1315) and Rh2O3 (ASTM, . The presence of metallic state of Co in the catalyst used indicates that the Co was not oxidized in the reaction medium, which is a mixture of both oxidizing and reducing agents-CH4, CO2, H2O, CO, and H2. Besides, the metallic state of Co was kept even after air exposure when the catalyst was removed from the reactor. That is a result of Rh-Co interaction in the bimetallic catalyst. No carbon species were observed over the bimetallic Co-Rh catalysts exploited in dry/bireforming of methane for 100-200 h. Whereas for monometallic catalysts used in DRM and studied for comparison, the monolayer carbon (ASTM, 26-1083) was unexpectedly observed for monometallic 2.5% Rh/Al2O3, while graphite formation was observed on the spent monometallic 2.5% Co/Al2O3. It is well known that the noble metal catalysts are more resistant to coking [27], while the Co and Ni-based ones are less stable [20]. On a basis of the TEM data, it may be concluded that the addition of Rh to Co leads to formation of sub-5 nm particles, which are more stable and resistant to carbonization than Co and even Rh alone.
By means of SEM of the 5%Со-Rh/Al2O3 catalyst, the loose amorphous structures covered the entire catalyst surface of both fresh samples, Co:Rh(95:5) and Co:Rh(9:1), were revealed (Figure 3a,c). Rarely, the particles with a bigger size of 2.5 nm were observed. The last were presumably attributed to the oxidized state of metals; because microdiffraction was presented by small amounts of weak rings, their identification was difficult. The presence of metal oxides is not a surprise because the reduced and passivated sample has been undergone to air. It is notable that the reduced state of Co 0 in the bimetallic catalyst was kept in an air environment. The Co 0 state in a monometallic Co/Al 2 O 3 catalysts with Co content of 2.5 and 5 mass % was not observed after passivation [31,39].
In a spent sample (after reaction and subsequent passivation by the Ar flow), the particles retained their spherical shape, but some of them increased in size varied within 1.5-3 nm (Figure 2b). Microdiffraction analysis showed the presence of Co 0 and two oxidized states-RhO 2 (ASTM, 21-1315) and Rh 2 O 3 (ASTM, . The presence of metallic state of Co in the catalyst used indicates that the Co was not oxidized in the reaction medium, which is a mixture of both oxidizing and reducing agents-CH 4 , CO 2 , H 2 O, CO, and H 2 . Besides, the metallic state of Co was kept even after air exposure when the catalyst was removed from the reactor. That is a result of Rh-Co interaction in the bimetallic catalyst. No carbon species were observed over the bimetallic Co-Rh catalysts exploited in dry/bireforming of methane for 100-200 h. Whereas for monometallic catalysts used in DRM and studied for comparison, the monolayer carbon (ASTM, 26-1083) was unexpectedly observed for monometallic 2.5% Rh/Al 2 O 3 , while graphite formation was observed on the spent monometallic 2.5% Co/Al 2 O 3 . It is well known that the noble metal catalysts are more resistant to coking [27], while the Co and Ni-based ones are less stable [20]. On a basis of the TEM data, it may be concluded that the addition of Rh to Co leads to formation of sub-5 nm particles, which are more stable and resistant to carbonization than Co and even Rh alone.
By means of SEM of the 5%Co-Rh/Al 2 O 3 catalyst, the loose amorphous structures covered the entire catalyst surface of both fresh samples, Co:Rh(95:5) and Co:Rh(9:1), were revealed (Figure 3a,c).  After reforming processes over the catalysts the amorphousness became less pronounced especially for the lower content of Rh-0.25 mass% (Figure 3b,d). This sample was continuously exploited for more than 60 h in the DRM and BRM at 700 °C. It is difficult to determine the contribution of each process in changing the catalyst morphology. Nevertheless, the catalyst did not lose its activity for all periods of operation.
According to the XPS data ( Table 2) provided for the bimetallic 5%Со-Rh(1:1)/Al2O3 as well as for monometallic 2.5%Со/Al2O3 catalysts for comparison, both the metals are in oxidised states in the initial calcined samples (fresh). The XPS spectra of the calcined bimetallic Co-Rh catalyst are shown in Figure 4a,b.  After reforming processes over the catalysts the amorphousness became less pronounced especially for the lower content of Rh-0.25 mass % (Figure 3b,d). This sample was continuously exploited for more than 60 h in the DRM and BRM at 700 • C. It is difficult to determine the contribution of each process in changing the catalyst morphology. Nevertheless, the catalyst did not lose its activity for all periods of operation.
According to the XPS data ( Table 2) provided for the bimetallic 5%Co-Rh(1:1)/Al 2 O 3 as well as for monometallic 2.5%Co/Al 2 O 3 catalysts for comparison, both the metals are in oxidised states in the initial calcined samples (fresh). The XPS spectra of the calcined bimetallic Co-Rh catalyst are shown in Figure 4a,b.  The binding energy (BE) of the Rh3d5/2 line is 308.9 eV. A comparison of the XPS data obtained (Table 2) with the reference data [40] shows that Rh exists in an oxidation state of 3 + , which can correspond to rhodium oxide or CoRh2O4. As for Co, the binding energy of the Co2p3/2 line is 781.3 eV for both monometallic and bimetallic Co-Rh catalysts. This value is higher than for Co oxides (778.3-780.2 eV) and close to Co 2+ in such compounds as CoRh2O4 (781.2 eV), СоАl2О4 (781.0 eV) or Co hydroxide (781.0 eV) according to the reference data [40]. By TEM data, no spinel and Co hydroxide were observed. A higher dispersed state of Co in both mono-and bimetallic catalysts and strong metal-support interaction can shift the BE to higher values by up to 1 eV. For these reasons it is possible to suppose that the BE at 781.3 eV can be assigned to Co n+ either in CoO or Co3O4, which have the standard BE Co 2p3/2-780.4 and 780.2 eV respectively [40]. TPR analysis described below indicates the existence of at least Co 3+ state.
To elucidate the effect of Rh on Co reducibility, the bimetallic Co-Rh/Al2O3 catalysts have been studied by H2-TPR. In Figure 5, the TPR profiles are presented for five catalyst compositions including monometallic ones. The binding energy (BE) of the Rh3d 5/2 line is 308.9 eV. A comparison of the XPS data obtained (Table 2) with the reference data [40] shows that Rh exists in an oxidation state of 3 + , which can correspond to rhodium oxide or CoRh 2 O 4 . As for Co, the binding energy of the Co2p 3/2 line is 781.3 eV for both monometallic and bimetallic Co-Rh catalysts. This value is higher than for Co oxides (778.3-780.2 eV) and close to Co 2+ in such compounds as CoRh 2 O 4 (781.2 eV), CoAl 2 O 4 (781.0 eV) or Co hydroxide (781.0 eV) according to the reference data [40]. By TEM data, no spinel and Co hydroxide were observed. A higher dispersed state of Co in both mono-and bimetallic catalysts and strong metal-support interaction can shift the BE to higher values by up to 1 eV. For these reasons it is possible to suppose that the BE at 781.3 eV can be assigned to Co n+ either in CoO or Co 3 O 4 , which have the standard BE Co 2p 3/2 -780.4 and 780.2 eV respectively [40]. TPR analysis described below indicates the existence of at least Co 3+ state.
To elucidate the effect of Rh on Co reducibility, the bimetallic Co-Rh/Al 2 O 3 catalysts have been studied by H 2 -TPR. In Figure 5, the TPR profiles are presented for five catalyst compositions including monometallic ones.
There is one peak at 160 • C for monometallic Rh/Al 2 O 3 catalyst corresponding to the reduction of Rh 2 O 3 to Rh 0 . The oxidized states of Co in monometallic 2.5% Co/Al 2 O 3 are reduced at higher temperatures and presented by two main peaks: at 386 • and broad peak at 621 with a shoulder at 538 • C.
The addition of a noble metal, Rh, has a significant effect on the reducibility of cobalt supported on the alumina. The temperature of the reduction of Co oxides is significantly shifted to lower values. The shift is strengthened with an increase in Rh amount. Three peaks were observed for the bimetallic catalysts with various content of Rh. The first one at 160 • C is the same for all catalysts and attributed to the reduction of Rh 3+ , which presence was shown by XPS. The second and third peaks are assigned to the staged reduction of Co oxides via Co 3 O 4 →CoO→Co 0 [20,28,39,41]. The positions of peaks are shifted to lower temperature from 386 • C to 258 → 227 → 210 • C, while the third peak assigned to the 2nd step of reduction, CoO→Co 0 , is shifted from 621 • C to 420 → 395 → 290 • C with an increase in Rh content: 0.25 → 0.5 → 2.5 mass% respectively. Thus, adding Rh to Co/Al 2 O 3 catalyst promotes Co reduction and shifts its reduction temperature by 128-176 • for Co 2 O 3 and by 201-331 • for CoO depending on Rh amount. The same effect of Rh on Co reducibility was reported in the literature [14,20]. There is one peak at 160 °C for monometallic Rh/Al2O3 catalyst corresponding to the reduction of Rh2O3 to Rh 0 . The oxidized states of Co in monometallic 2.5% Co/Al2O3 are reduced at higher temperatures and presented by two main peaks: at 386° and broad peak at 621 with a shoulder at 538 °C.
The addition of a noble metal, Rh, has a significant effect on the reducibility of cobalt supported on the alumina. The temperature of the reduction of Co oxides is significantly shifted to lower values. The shift is strengthened with an increase in Rh amount. Three peaks were observed for the bimetallic catalysts with various content of Rh. The first one at 160 °C is the same for all catalysts and attributed to the reduction of Rh 3+ , which presence was shown by XPS. The second and third peaks are assigned to the staged reduction of Co oxides via Co3O4→CoO→Co 0 [20,28,39,41]. The positions of peaks are shifted to lower temperature from 386 °C to 258 → 227 → 210 °C, while the third peak assigned to the 2nd step of reduction, CoO→Co 0 , is shifted from 621 °C to 420 → 395 → 290 °C with an increase in Rh content: 0.25 → 0.5 → 2.5 mass% respectively. Thus, adding Rh to Co/Al2O3 catalyst promotes Co reduction and shifts its reduction temperature by 128-176° for Co2O3 and by 201-331° for CoO depending on Rh amount. The same effect of Rh on Co reducibility was reported in the literature [14,20].
The reduced state of Co 0 promoted by Rh is kept even in an air environment as was confirmed by TEM studies. This effect of Rh on Co is very important, because the methane reforming is activated by metallic Co and oxidation of metallic sites can cause the decrease in activity [6,20] and should be avoided. The data obtained correspond to the reported The reduced state of Co 0 promoted by Rh is kept even in an air environment as was confirmed by TEM studies. This effect of Rh on Co is very important, because the methane reforming is activated by metallic Co and oxidation of metallic sites can cause the decrease in activity [6,20] and should be avoided. The data obtained correspond to the reported promotion effect of noble metals on Co. Thus, for the Co-Rh catalysts, Co 0 was stated to be the active sites of the catalyst, and the presence of the optimum amount of rhodium led to highly dispersed active species [20]. Authors [32] concluded that the modification with Rh has at least two different roles in the Co-Rh catalytic system. First, the reduction of Co in H 2 is much more efficient in the presence of Rh due to the hydrogen spillover phenomena and Rh may also help to keep the cobalt in the metallic state. In addition, adding Rh to Co leads to higher dispersion state of metals and its stabilisation.

Catalyst Test
Due to the highly endothermic character of reforming reactions (Equations (1) and (2)), the temperature has the most significant effect on the dry reforming (DRM), steam reforming (SRM), and the combined DRM-SRM or bireforming of methane (BRM). The effect of temperature on the performance of the bimetallic Co-Rh supported catalysts in methane reforming by CO 2 and/or H 2 O has been studied. Furthermore, the effect of the amount of Rh and feed composition on methane reforming has been elucidated.

Dry Reforming of Methane
Temperature has the same effect on the conversion of CH 4 and CO 2 and yield of hydrogen and carbon monoxide for all three Co-Rh compositions in DRM. The extents of the conversion of methane (X CH4 , %) and carbon dioxide (X CO2 , %), as well as the yield of hydrogen (YH 2 , µmol/g cat ×s) and carbon oxide (YCO, µmol/g cat ×s), are raised with an increase in temperature as it shown for example the 5%Co-Rh(98:2)/Al 2 O 3 (Figure 6) catalysts. The eligible activity of the catalysts (X CH4 > 25%) in the DRM is occurred at temperatures above 500 • C. Both reactants converted up to extents of 90-99% at 700-800 • C depending on the content of Rh.

Catalyst Test
Due to the highly endothermic character of reforming reactions (Equations (1) and (2)), the temperature has the most significant effect on the dry reforming (DRM), steam reforming (SRM), and the combined DRM-SRM or bireforming of methane (BRM). The effect of temperature on the performance of the bimetallic Co-Rh supported catalysts in methane reforming by CO2 and/or H2O has been studied. Furthermore, the effect of the amount of Rh and feed composition on methane reforming has been elucidated.

Dry Reforming of Methane
Temperature has the same effect on the conversion of CH4 and CO2 and yield of hydrogen and carbon monoxide for all three Co-Rh compositions in DRM. The extents of the conversion of methane (XCH4, %) and carbon dioxide (XCO2, %), as well as the yield of hydrogen (YH2, µmol/gcat×s) and carbon oxide (YCO, µmol/gcat×s), are raised with an increase in temperature as it shown for example the 5%Co-Rh(98:2)/Al2O3 (Figure 6) catalysts. The eligible activity of the catalysts (XCH4 > 25%) in the DRM is occurred at temperatures above 500 °C. Both reactants converted up to extents of 90-99% at 700-800 °C depending on the content of Rh. In Table 3, the data on all three reactions are listed for three compositions of Co:Rh studied. As it can be seen, the increase in Rh content from 0.1 to 0.5 mass% leads to no substantial growth of conversion of both reactants: XCH4 is raised from 87.6 to 94.4.0%, while XCO2 from 86.8 to 91.1% at 700 °C ( Table 3). As it was earlier shown, the monometallic Co catalyst is much less active than Rh modified ones [31,39]. Thus, for the 2.5%Co/Al2O3 catalyst XCH4 = 60.4, XCO2 = 44.3% and H2/CO = 0.7 at higher temperature-900 °C. That corresponds to the results of Erdohelyi et al. reported the promoting effect of Rh on 10%Co/Al2O3 in DRM [20]. In Table 3, the data on all three reactions are listed for three compositions of Co:Rh studied. As it can be seen, the increase in Rh content from 0.1 to 0.5 mass% leads to no substantial growth of conversion of both reactants: X CH4 is raised from 87.6 to 94.4.0%, while X CO2 from 86.8 to 91.1% at 700 • C ( Table 3). As it was earlier shown, the monometallic Co catalyst is much less active than Rh modified ones [31,39]. Thus, for the 2.5%Co/Al 2 O 3 catalyst X CH4 = 60.4, X CO2 = 44.3% and H 2 /CO = 0.7 at higher temperature-900 • C. That corresponds to the results of Erdohelyi et al. reported the promoting effect of Rh on 10%Co/Al 2 O 3 in DRM [20]. The ratio of H 2 /CO in the syngas formed on the catalysts with lower Rh content (0.1-0.5 mass %) is less in comparison with the Co-Rh(1:1) [31], H 2 /CO is 0.99-1.0 and 1.1 respectively. Higher than stoichiometric H 2 /CO ratio may be the result of the WGSR reaction (Equation (3)), which is strengthened with an increase in Rh content. In contrast, monometallic 2.5%Co/Al 2 O 3 is prone to RWGSR (Equation (4)), which is responsible for decreasing H 2 yield and accordingly for lower ratio of H 2 /CO, which is equal to 0.7 in DRM at 900 • C [39].

Bireforming of Methane
For bireforming of methane, temperature has the same effect on the catalyst performance as for DRM, as is shown in, for example, the 5%Co-Rh(98:2)/Al 2 O 3 (Figure 7). The ratio of H2/CO in the syngas formed on the catalysts with lower Rh content (0.1-0.5 mass %) is less in comparison with the Co-Rh(1:1) [31], H2/CO is 0.99-1.0 and 1.1 respectively. Higher than stoichiometric H2/CO ratio may be the result of the WGSR reaction (Equation (3)), which is strengthened with an increase in Rh content. In contrast, monometallic 2.5%Co/Al2O3 is prone to RWGSR (Equation (4)), which is responsible for decreasing H2 yield and accordingly for lower ratio of H2/CO, which is equal to 0.7 in DRM at 900 °C [39].

Bireforming of Methane
For bireforming of methane, temperature has the same effect on the catalyst performance as for DRM, as is shown in, for example, the 5%Co-Rh(98:2)/Al2O3 (Figure 7). The subsequent twofold increase in the amount of Rh from 0.25 to 0.5 wt% does not lead to a noticeable increase in activity. The close values of methane conversion-99.2 and 98.8% and carbon dioxide conversion-97.3 and 96.5% for Co-Rh(9:1) and Co-Rh(95:5) respectively were reached at 800 °C and CH4/CO2/H2O = 1/1/1.
The data on the effect of steam amount on methane reforming studied for the two Co-Rh compositions, 95:5 and 9:1, are summarized in Table 4. As it can be seen, the increase in steam volume from 0 to 2 parts in a feed of CH4:CO2 = 1:1 leads to enhancing the H2/CO ratio in syngas formed from 1.0 to 2.7-3.1. The higher conversions of methane and carbon dioxide were observed at adding of 1. The subsequent twofold increase in the amount of Rh from 0.25 to 0.5 wt% does not lead to a noticeable increase in activity. The close values of methane conversion-99.2 and 98.8% and carbon dioxide conversion-97.3 and 96.5% for Co-Rh(9:1) and Co-Rh(95:5) respectively were reached at 800 • C and CH 4 /CO 2 /H 2 O = 1/1/1.
The data on the effect of steam amount on methane reforming studied for the two Co-Rh compositions, 95:5 and 9:1, are summarized in Table 4. As it can be seen, the increase in steam volume from 0 to 2 parts in a feed of CH 4 :CO 2 = 1:1 leads to enhancing the H 2 /CO ratio in syngas formed from 1.0 to 2.7-3.1. The higher conversions of methane and carbon dioxide were observed at adding of 1.5-2 vol. parts of steam depending on Rh amount. At a lower volume of steam, CH 4 :CO 2 :H 2 O = 1:1:0.5, the extents of conversions of carbon dioxide and methane are almost the same for the Co:Rh(95:5), at following increase in steam volume from 0.5 to 2 the difference between X CH4 and X CO2 become more notable and methane conversion prevails. This effect is more pronounced for composition, with higher Rh amount, Co:Rh(9:1) ( Table 4). The possible reason of that is the WGSR (Equation (3)) accelerating by Rh. It is obviously that the rate of WGSR is enhanced in excess of steam. Running the two competitive reactions-CO 2 and H 2 O decomposition over the same active sites may also be a reason for decline of CO 2 conversion. The data on hydrogen and carbon oxide yields in BRM are given in Table 3. In comparison with DRM, hydrogen yield grows up over all catalysts studied when steam is added to the CH 4 -CO 2 = 1:1 feed, while the CO content is decreased. These are two reasons of enhancing H 2 /CO ratio. For example, the Y H2 is increased from 3.2 to 4.2-4.3 µmol/g cat ×s for the catalysts with 0.25-0.5 mass % of Rh, while CO yield are decreased from 3.2 to 2.5-3.1 µmol/g cat ×s. The same behaviour was observed for the catalyst with the smaller amount of Rh-0.1 mass %: Y H2 is increased, while Y CO is decreased from 16.2 and 16.4 to 19.6 and 15.0 µmol/g cat ×s respectively (Table 3). It needs to note that the concentrated feed of CH 4 -CO 2 without dilution with inert gas was used for this catalyst testing, that is why a higher yield of products occurred. WGSR (Equation (3)) is the most likely reason for a lowering conversion of CO 2 in BRM in comparison with DRM. In fact, the values of CO 2 have been calculated on a base of Formula (7), which considers only the content of CO 2 in inlet and outlet flows but does not consider the formation of CO 2 by WGSR and its contribution to carbon dioxide amount in the outlet flow. That is why a real conversion of carbon dioxide may be higher than the calculated one for BRM.
Comparison of DRM and BRM over Co-Rh/Al 2 O 3 ( Table 3) demonstrates such benefits of the BRM reaction as a controllable syngas composition and deeper conversion of methane. Intensification of methane reforming is caused by reducing carbon formation because of increased gasification [2,8,10]. Nevertheless, the steam amount to be added is limited for Co-based catalysts. It is known that excess of water can poison the catalyst surface; therefore, the catalysts with a good performance in DRM may be inactive in BRM. By TEM, the formation of Co hydroxide due to the water effect in BRM was found for monometallic 2.5%Co/Al 2 O 3 . The cobalt hydroxide is not active in DRM and cause the drastic decreasing activity of the catalyst, X CH4 < 5% at 900 • C. The addition of Rh not only increases the activity of Co-based catalyst in DRM, but also makes it active in BRM. So, this is evidence of synergistic effect of Co and Rh interaction.

Steam Reforming of Methane
Due to the revealed high activity of Co-Rh/Al 2 O 3 in BRM, the catalysts have been tested in the steam reforming of methane (SRM)-a method of industrial hydrogen production. The feed with ratio of CH 4 /H 2 O = 1/1 has been used that corresponds to the stoichiometry (Equation (2)). It has been observed that the Co-Rh systems performed high activity in SRM process too. In Figure 8a,b, the typical pattern of temperature effect on methane conversion is presented for catalysts with content of Rh-0.25 and 0.5%. X CH4 are 98.4 and 97.4% for Co-Rh (95:5) and Co-Rh(9:1) respectively at 800 • C. The yield of carbon monoxide in SRM is substantially less than in DRM and BRM, 1.3-1.5 and 2.5-3.2 µmol/g cat ×s respectively. This is due to a half lower intake of C from a feed consisting only of CH 4 , there is no CO 2 as the second supplier of C. The yield of hydrogen for SRM is slightly higher than for BRM: 4.7 and 4.3-4.5 µmol/g cat ×s respectively (Table 3). 98.4 and 97.4% for Co-Rh (95:5) and Co-Rh(9:1) respectively at 800 °C. The yield of carbon monoxide in SRM is substantially less than in DRM and BRM, 1.3-1.5 and 2.5-3.2 µmol/gcat×s respectively. This is due to a half lower intake of C from a feed consisting only of CH4, there is no CO2 as the second supplier of C. The yield of hydrogen for SRM is slightly higher than for BRM: 4.7 and 4.3-4.5 µmol/gcat×s respectively (Table 3). The ratio of H2/CO in SRM reaction is grown from 3.1 to 3.6 with an increase in Rh content from 0.25 to 0.5 mass%, while the yield of hydrogen is approximately the same-4.7 µmol/gcat×s. As for CO yield, it is moderately decreased from 1.5 to 1.3 µmol/gcat×s ( Table 3). The activity of both compositions is almost the same: XCH4 = 92.2-92.6%.
The higher H2/CO ratio in SRM in comparison with BRM and DRM occurs due to decrease in CO content in products formed. The SRM does not give a much higher yield The ratio of H 2 /CO in SRM reaction is grown from 3.1 to 3.6 with an increase in Rh content from 0.25 to 0.5 mass%, while the yield of hydrogen is approximately the same- 4.7 µmol/g cat ×s. As for CO yield, it is moderately decreased from 1.5 to 1.3 µmol/g cat ×s ( Table 3). The activity of both compositions is almost the same: X CH4 = 92.2-92.6%.
The higher H 2 /CO ratio in SRM in comparison with BRM and DRM occurs due to decrease in CO content in products formed. The SRM does not give a much higher yield of hydrogen than BRM.
In Figure 9, the comparison of methane reforming either by carbon dioxide or steam, or by their mixture is presented. As can be seen, all three processes may proceed with approximately the same activity over the same catalyst-5%Co-Rh(95:5)/Al 2 O 3 at 700 • C with production of syngas with H 2 /CO ratio varied in a wide region-1-3.
The ratio of H2/CO in SRM reaction is grown from 3.1 to 3.6 with an increase in Rh content from 0.25 to 0.5 mass%, while the yield of hydrogen is approximately the same-4.7 µmol/gcat×s. As for CO yield, it is moderately decreased from 1.5 to 1.3 µmol/gcat×s ( Table 3). The activity of both compositions is almost the same: XCH4 = 92.2-92.6%.
The higher H2/CO ratio in SRM in comparison with BRM and DRM occurs due to decrease in CO content in products formed. The SRM does not give a much higher yield of hydrogen than BRM.
In Figure 9, the comparison of methane reforming either by carbon dioxide or steam, or by their mixture is presented. As can be seen, all three processes may proceed with approximately the same activity over the same catalyst-5%Co-Rh(95:5)/Al2O3 at 700 °C with production of syngas with H2/CO ratio varied in a wide region-1-3.

Stability Test
To determine the stability, the catalysts with less amount of Rh-0.1-0.25 mass % performed the activity and yield of products comparable with the catalysts containing higher Rh content were selected for long-term testing in the dry reforming and in bireforming of methane under P = 1 atm, t = 670-700 °C, GHSV = 1000-1250 h −1 , and

Stability Test
To determine the stability, the catalysts with less amount of Rh-0.1-0.25 mass % performed the activity and yield of products comparable with the catalysts containing higher Rh content were selected for long-term testing in the dry reforming and in bireforming of methane under P = 1 atm, t = 670-700 • C, GHSV = 1000-1250 h −1 , and CH 4 /CO 2 = 1/1 for DRM and CH 4 /CO 2 /H 2 O g = 1/1/1 for BRM. The 5%Co-Rh(95:5)/Al 2 O 3 catalyst with V cat = 12 mL was tested in a quartz reactor in DRM and then without stopping and regeneration in BRM for 25 h each process. The 5%Co-Rh(98:2)/Al 2 O 3 with V cat = 100 mL was tested in the pilot reactor made from Inconel in BRM and DRM for 100 h each.
In Figure 10, the dependence of CH 4 and CO 2 conversion ( Figure 10a) and products yields (Figure 10b CH4/CO2 = 1/1 for DRM and CH4/CO2/H2Og = 1/1/1 for BRM. The 5%Со-Rh(95:5)/Al2O3 catalyst with Vcat = 12 mL was tested in a quartz reactor in DRM and then without stopping and regeneration in BRM for 25 h each process. The 5%Со-Rh(98:2)/Al2O3 with Vcat = 100 mL was tested in the pilot reactor made from Inconel in BRM and DRM for 100 h each. In Figure 10, the dependence of CH4 and CO2 conversion ( Figure 10a) and products yields (Figure 10b Syngas formed has a ratio of H2/CO = 0.9. There was no loss of the catalyst activity for 25 h in DRM. Moreover, the conversion of both reactants was slightly increased with TOS: XCO2 from 81 to 90 and XCH4 from 90 to 93%. It may occur because of the 'catalyst elaboration' accompanied by changing such catalyst surface characteristics as dispersion, composition, and state of the active centres due to effect of the reaction medium during Syngas formed has a ratio of H 2 /CO = 0.9. There was no loss of the catalyst activity for 25 h in DRM. Moreover, the conversion of both reactants was slightly increased with TOS: X CO2 from 81 to 90 and X CH4 from 90 to 93%. It may occur because of the 'catalyst elaboration' accompanied by changing such catalyst surface characteristics as dispersion, composition, and state of the active centres due to effect of the reaction medium during the first hours of the catalyst operation. Such phenomena, which is typical for Co-based FTS catalysts, was found for the 5%Co-Pt/Al 2 O 3 -ZrO 2 catalysts tested in DRM and BRM [42]. This is evidence of the significant influence of the DRM reaction medium (mix of initial, intermediate, and final products-CO 2 , CH 4 , CO, H 2 , H 2 O, etc.) on the catalyst surface, as a result of which the catalyst is completely structured. In particular, this can be accompanied by an increase in dispersion or the formation of new active sites.
After 25 h testing in DRM, steam was added to a CH 4 /CO 2 = 1/1 feed flow to carry out bireforming. The data on reagents conversion and products yield are presented in Figures 11a and 11b respectively. As it can be seen, after previous 25 h operation in DRM the conversion of methane and carbon dioxide were stabilised in BRM. Their values were negligible fluctuated within 92.0-93.9 and 88.5-89.9% for X CH4 and X CO2 respectively. In comparison with DRM, the yield of CO was slightly decreased, in opposite, the yield of hydrogen was substantially increased, H 2 /CO = 1.6. This is evidence of H 2 O decomposition and contributing H 2 from water. The 5%Co-Rh(95:5)/Al 2 O 3 catalyst demonstrated stability in BRM with producing H 2 -rich syngas for all entire period of 25 h.
Syngas formed has a ratio of H2/CO = 0.9. There was no loss of the catalyst activity for 25 h in DRM. Moreover, the conversion of both reactants was slightly increased with TOS: XCO2 from 81 to 90 and XCH4 from 90 to 93%. It may occur because of the 'catalyst elaboration' accompanied by changing such catalyst surface characteristics as dispersion, composition, and state of the active centres due to effect of the reaction medium during the first hours of the catalyst operation. Such phenomena, which is typical for Co-based FTS catalysts, was found for the 5%Co-Pt/Al2O3-ZrO2 catalysts tested in DRM and BRM [42]. This is evidence of the significant influence of the DRM reaction medium (mix of initial, intermediate, and final products-CO2, CH4, CO, H2, H2O, etc.) on the catalyst surface, as a result of which the catalyst is completely structured. In particular, this can be accompanied by an increase in dispersion or the formation of new active sites.
After 25 h testing in DRM, steam was added to a CH4/CO2 = 1/1 feed flow to carry out bireforming. The data on reagents conversion and products yield are presented in Figure  11a and Figure 11b respectively. As it can be seen, after previous 25 h operation in DRM the conversion of methane and carbon dioxide were stabilised in BRM. Their values were negligible fluctuated within 92.0-93.9 and 88.5-89.9% for XCH4 and XCO2 respectively. In comparison with DRM, the yield of CO was slightly decreased, in opposite, the yield of hydrogen was substantially increased, H2/CO = 1.6. This is evidence of H2O decomposition and contributing H2 from water. The 5%Со-Rh(95:5)/Al2O3 catalyst demonstrated stability in BRM with producing H2-rich syngas for all entire period of 25 h. On the whole, the 5%Co-Rh(95:5)/Al 2 O 3 catalyst worked continuously without regeneration for about 70 h in DRM and BRM. No visible carbon formation was observed, by TEM and SEM any carbonaceous species were not found.
These results demonstrate the possibility to use the same catalyst for the production of syngas with any desirable H 2 /CO ratio by changing feed composition-CH 4 :CO 2 :H 2 O and the easy transition from one reforming process to another one depending on requirements. The additional technological benefit may be in mitigating coke formation in case of its accumulation in DRM by providing periodically BRM/SRM. Such alternation of processes DRM ↔ BRM/SRM will promote the removal of coke by water as a mild oxidizing agent and allow it to work without stops.
To develop the effective stable catalyst with a low content of noble metal, the 5%Co-Rh(98:2)/Al 2 O 3 with Rh content of 0.1 mass % (V cat = 100 mL) was selected to be tested in DRM and in BRM. Average values of degrees of CH 4 and CO 2 conversion in DRM are almost the same and equal to 89.0 and 89.7% respectively. Syngas formed has a ratio of H 2 /CO~1 ( Figure 12). of processes DRM ↔ BRM/SRM will promote the removal of coke by water as a mild oxidizing agent and allow it to work without stops.
To develop the effective stable catalyst with a low content of noble metal, the 5%Co-Rh(98:2)/Al2O3 with Rh content of 0.1 mass % (Vcat = 100 mL) was selected to be tested in DRM and in BRM. Average values of degrees of CH4 and CO2 conversion in DRM are almost the same and equal to 89.0 and 89.7% respectively. Syngas formed has a ratio of Н2/СО~1 ( Figure 12). The substantial changes in activity and productivity were observed, when steam in the amount of 0.5 volume parts was added to a feed of CH4:CO2 = 1:1 ( Figure 13). If the extent of conversion of carbon dioxide slightly decreased from 89.7 to 87% on average then X(CH4) was increased by almost 10% from 89 to 98.7%. The yield of reaction products, carbon oxide and hydrogen, were also grown. The last increased in higher degree, due to enhancing methane conversion and water contribution. It should be noted that such improved catalyst performance was reached at lower temperature-670 °C, which is less by 30 °C in comparison with DRM at the same catalyst. As far as it can be concluded on the basis of stable extent of conversion and product yields, the 5%Co-Rh(98:2)/Al2O3 catalyst worked with the stable activity and selectivity for 100 h of its testing. The substantial changes in activity and productivity were observed, when steam in the amount of 0.5 volume parts was added to a feed of CH 4 :CO 2 = 1:1 ( Figure 13). If the extent of conversion of carbon dioxide slightly decreased from 89.7 to 87% on average then X(CH 4 ) was increased by almost 10% from 89 to 98.7%. The yield of reaction products, carbon oxide and hydrogen, were also grown. The last increased in higher degree, due to enhancing methane conversion and water contribution. It should be noted that such improved catalyst performance was reached at lower temperature-670 • C, which is less by 30 • C in comparison with DRM at the same catalyst. As far as it can be concluded on the basis of stable extent of conversion and product yields, the 5%Co-Rh(98:2)/Al 2 O 3 catalyst worked with the stable activity and selectivity for 100 h of its testing. The results obtained demonstrate the ability of bimetallic Co-Rh-based catalysts with low content of Rh-0.1-0.25 mass% to catalyse methane reforming by either CO2, H2O, or combined CO2-H2O. It allows to apply the 5%Со-Rh/Al2O3 catalyst with low content of Rh for producing syngas with a wide range of H2/CO ratios from the various dry or wet feedstock-natural gas or biogas.

Catalyst Preparation and Characterisation
The catalysts were prepared by wet co-impregnation of γ-alumina with aqueous solutions of Co(NO3)2·6H2O and RhCl3·nH2O, purity is 99% for both compounds. Support is γ-alumina (XRD-pattern in Figure 1b) being granules in the form of balls with a diameter of 3-5 mm and BET surface area is 140 m 2 /g (IC-02-200, Novosibirsk, Russia, The results obtained demonstrate the ability of bimetallic Co-Rh-based catalysts with low content of Rh-0.1-0.25 mass% to catalyse methane reforming by either CO 2 , H 2 O, or combined CO 2 -H 2 O. It allows to apply the 5%Co-Rh/Al 2 O 3 catalyst with low content of Rh for producing syngas with a wide range of H 2 /CO ratios from the various dry or wet feedstock-natural gas or biogas.

Catalyst Preparation and Characterisation
The catalysts were prepared by wet co-impregnation of γ-alumina with aqueous solutions of Co(NO 3 ) 2 ·6H 2 O and RhCl 3 ·nH 2 O, purity is 99% for both compounds. Support is γ-alumina (XRD-pattern in Figure 1b) being granules in the form of balls with a diameter of 3-5 mm and BET surface area is 140 m 2 /g (IC-02-200, Novosibirsk, Russia, purity-99%). After drying the catalysts were calcined at 400 • C for 3 h and then prior testing was reduced by hydrogen at 300-400 • C for 1-3 h. The total nominal content of Co and Rh metals was equal to 5% of the total catalyst mass. The mass ratio of Co:Rh was 98:2, 95:5, and 9:1, which corresponds to 0.1, 0.25, and 0.5 mass% of Rh and 4.9, 4.75, and 4.5 mass% of Co respectively. Element analysis provided by means of the scan electron microscope JSM 6610 LV JOEL, Japan showed a correlation between nominal and actual weight and deviation of no more than 2% on average for each element involved.
The physico-chemical studies of the catalysts were, as a rule, undertaken prior to the reaction (fresh sample) and after the reaction (spent sample) to understand the effects of the reaction feed and process conditions on the catalyst characteristics, such as the specific surface area, reducibility of Co, morphology, particle size, element distribution, etc.
Specific surface area measurements were performed with a Micromeritics Accusorb apparatus, USA. Samples were preliminarily degassed in an Ar flow at 220 • C for 3.5 h. The BET specific surface area was deduced from the N 2 adsorption/desorption isotherms at 77 K.
X-ray diffraction (XRD) measurements were performed with the fresh and used catalysts using the CuK α or CoK α radiation of a "Dron-4" upgraded powder diffractometer, USSR.
Transmission electron microscopy at a resolution of approximately 0.5 nm was used to characterize the sizes of the metal particles, their distribution and state in the fresh and spent samples of catalysts. Electron microscopy studies were provided with a JEM-100CX unit, Japan. Phase identification was performed with the help of the ASTM (American Society for Testing and Materials).
Scanning electron microscopy (SEM) images were taken using the JSM 6610 LV JOEL, Equipment, Japan using a secondary electron detector operating under high and low vacuum. The external surface of the entire catalyst granule as well as the inner surface of a catalyst granule divided in half have been scanned.
Investigation of catalysts by the XPS method was carried out using the VG ESCALAB electron spectrometer from VG Scientific at the Boreskov Institute of Catalysis, Novosibirsk, Russia. The emission of electrons from the sample was carried out using the soft X-ray radiation of AlK, so that the average mean free path of electrons was 2.0-3.0 nm depending on the line analysed. Thus, the surface thickness of the samples studied was 6-9 nm. Before spectroscopic measurements, the samples were evacuated in a preliminary preparation chamber to P = 10 −7 mb and then moved to an analyser chamber, where the base vacuum was not higher than 10 −9 mb. The spectra were recorded using a hemispherical electron energy analyser in the FAT (Fixed Analyzer Transmission) mode. The binding energy obtained were compared with the tabular empirical data given in [40].
To determine the effect of Rh on Co reducibility, the temperature programmed reduction (H 2 -TPR) was provided for the bimetallic Co-Rh/Al 2 O 3 and for mono-metallic catalysts for comparison. TPR measurements were performed on equipment CETARAM, France using a thermal conductivity detector (TCD) and a 5%H 2 /N 2 mixture at a flow rate of 20 ccm/min. The catalyst samples were heated from ambient temperature to 800 • C at 5 • C/min.

Catalyst Test
The catalysts were tested in a flow quartz reactor supplied with programmed heating, a controlled feed velocity, a programmable syringe pump (Braun Melsungen AG, Melsungen, Germany), and GCs. To provide the stability test, the pilot flow installation included the H 2 :CO ratio is tuned from 0.9 to 3.9 covering a wide range of syngas ratios relevant to various applications.
The strong improvement in the activity and stability of Co-Rh catalysts in DRM, BRM, and SRM occurs due to the synergetic effect of Co-Rh interaction accompanied with an increase in the metal dispersion, the formation of more reactive intermediate carbonaceous species, and enhanced Co reducibility. No coke was observed on the studied bimetallic Co-Rh alumina supported catalyst.
Therefore, the bimetallic Co-Rh systems are considered as a potential and promising base for the catalyst to convert methane with producing syngas with desirable composition. Several additional topics should be studied in future work: lower Rh loading, optimisation of method of preparation, the following modification of the catalyst composition, etc.  Data Availability Statement: New own data is contained within the article. Other data have been cited and listed in the bibliography.