Rh promoted Ni over yttria–zirconia supported catalyst for hydrogen‐rich syngas production through dry reforming of methane

Rh‐promoted YZr‐supported Ni catalyst (5Ni/YZr) is investigated for DRM and characterized with X‐ray diffraction, Raman, infrared spectroscopy, cyclic reduction–oxidation–reduction temperature programmed experiment, thermogravimetry, and transmission electron microscope. Over 5Ni/YZr, some active sites become inactive under the CO2 stream and limit H2 yield to ∼71%. Upon 4 wt% Rh addition over 5Ni/YZr; more than one type of stable active sites (Rh and Ni) generates, moderate basic sites are enhanced, wide ranges of CO2‐interacted surface species (especially bidentate CO2‐adsorbed species) are grown and graphitic carbon proportion over spent catalyst declines. This resulted in 87.35% H2 yield and 86.73% CO yield up to 420 min. 5Ni4Rh/YZr catalyst maintains ∼80% H2 yield at the end of 27 h of DRM reaction.

for the catalytic conversion of CH 4 and CO 2 into hydrogen-rich syngas.
The reaction can be confined into two steps, majorly decomposition of CH 4 over catalytic active sites (into CH x and x 2 H 2 ) followed by the oxidation of CH x by CO 2 (Equations 1 and 2). The delay in CH x oxidation invites pivots in the path of CH x polymerization and the formation of carbon deposits from amorphous to inert graphitic type over the catalytic surface. The catalytic activity is unaffected if carbon diffusion occurs far away from the catalytic active sites. Otherwise, the shading of active sites by inactive carbon deactivates the catalyst seriously. 1 There is a possibility of interaction between H 2 (one of the DRM products) and CO 2 (one of the DRM reactants) over the same catalytic surface, which ultimately decreases the H 2 yield and increases the CO yield. The reverse water gas shift reaction (Equation 3) competes parallelly with the DRM reaction (Equation 4).  x CH CH + (4 − )/2H , This highly endothermic reaction makes preserving the catalytic active site difficult during high-temperature DRM reactions. The supported catalyst for DRM has garnered considerable attention as a potential solution. Ni catalyst supported by alumina, silica, and zirconia has been utilized frequently for DRM. The lattice oxygen endowing capacity of zirconia may make it more suitable for DRM type reaction as mobile lattice oxygen oxidizes surface carbon instantly and leaves the catalytic active sites exposed for continuous reaction. However, the phase transition of ZrO 2 against high temperature renders inferior catalytic activity towards DRM. Ni/ZrO 2 catalyst showed 35% H 2 yield during 460 min at 700°C. 2 Recently, mixing other metal oxides like tungsten, lanthana, and yttria along with zirconia was found to stabilize the zirconia phases against high temperatures, ensuring constant high catalytic performance. [2][3][4] The use of second metal oxide along with zirconia had an additional benefit in favor of DRM also. Tungsten-zirconia-supported Ni catalyst was additionally benefited by the additional CH 4 decomposition sites and redox behavior of W during carbon removal (WO 3 → WC → WO 3 ). It ensured a constant 43% H 2 yield for 420 min at 700°C. 3 In the same way, lanthana-zirconia supported Ni catalyst had enhanced basicity which led to enhanced CO 2 interaction and coordination of oxycarbonate species to La +3 . Such widened interaction of CO 2 with catalytic surface pushed H 2 yield about 80% over lanthana-zirconia supported Ni catalyst. 2 Enrichment of the oxygen layer over the catalyst surface after the addition of yttria into zirconia, as well as the involvement of this oxygen in the oxidation of the carbon deposit oxidation is sparking features of yttria-zirconiasupported Ni catalyst (Ni/YZr) in favor of DRM. 4 So, yttria-zirconia-supported Ni catalyst needs greater industrial catalyst development attention from the DRM community. Ni/YZr catalyst showed about 63.7% H 2 yield for 420-min time on steam at 700°C. Recently, barium-, ceria-, and holmium-like promoters over Ni/ YZr catalyst were investigated for DRM and showed constant ∼79%, ∼80%, and ∼84% H 2 yield during the 420min time on stream at 800°C, respectively. [5][6][7] The selection of effective promoters over Ni/YZr catalyst still needs a committed investigation.
The incorporation of just 0.5 wt% Rh into Coimpregnated TiO 2 caused an improvement in oxygen mobility. 8 Electrons trapped in Rh (during photoirradiation) can reduce CO 2 . 9 Over alumina support, Rh was found to disperse 100%. 10 However, over Al 2 O 3 support, 0.1 wt% Rh needs a long activation period (∼50 h) 11 towards the DRM. Hou et al. observed that Rh/Al 2 O 3 exhibited 57% CH 4 conversion after 240 min at 800°C. 12 Mesoporous alumina had high initial activity, but suddenly the activity dropped due to heavy coke deposition. On increasing space velocity, coke deposit is also triggered. 12 In computation studies, it is found that CH 4 was dissociated more easily over Rh-Ni than Ni-Co and pure Ni. 13 The binding energy of CH x is also stronger over Rh-Ni than Cu-Ni and Pd-Ni. 14 RhNi dispersion over mesoporous alumina was found to have excellent coke resistance even against high space velocity. 12 The addition of Rh into Ni-Co/Al 2 O 3 catalyst showed 46% H 2 yield at 700°C, 15 whereas Al 2 O 3 supported Rh-Ni catalyst exhibited ∼70% CH 4 conversion for 4 h TOS at 800°C. 12 After alumina support, the Rh-Ni system is needed to investigate ZrO 2 and Y 2 O 3 -ZrO 2 support. Herein, we have investigated 1-5 wt% Rh promoted yttria-zirconia-supported Ni catalyst for DRM. The catalyst is characterized by X-ray diffraction (XRD), Raman, infrared spectroscopy, CO 2temperature programmed desorption, cyclic reduction-oxidation-reduction temperature programmed experiment, thermogravimetry and transmission electron microscope (TEM). The outcome of this research will advance the understanding of developing industrial suitable DRM catalysts in the near future.

| Catalyst preparation
Support is prepared by mechanical mixing of 8 wt% yttrium oxide and 92 wt% zirconium oxide. Furthermore, Ni(NO 3 ) 2 ·6H 2 O (equivalent to 5 wt% Ni), RhCl 3 (equivalent to 1-5 wt% Rh) and support are again mixed mechanically. The slurry is formed by adding distilled water dropwise under grinding. It is further dried at 110°C and calcined at 600°C for 3 h. The prepared catalysts were abbreviated as 5NixRh/YZr (where x = 0-5 wt%).

| Catalyst characterization
The detailed characterization procedure and instrument specifications are given in the Supporting Information S1.

| Catalyst activity test
Dry reforming of methane reaction is carried out over 0.1 g 5NixRh/YZr (x = 0-5 wt%) catalyst in a fixed bed stainless steel reactor (diameter = 9.1 mm and height = 300 mm) supplied by PID Eng. and Tech Micro activity Reference Company. A K-type thermocouple (placed axially in the catalyst bed) is used to monitor the temperature of the catalyst bed. Before the DRM reaction, catalyst reductive pretreatment is carried out under 30 mL/min H 2 flow for 1 h at 700°C. Furthermore, physiosorbed H 2 is flushed out under N 2 flow for 15 min. Now, a gas feed composed of CO 2 , CH 4 , and N 2 (30:30:10 mL/min; total 42,000 mL/h·g cat gas hourly space velocity) is allowed to pass over the catalyst at 800°C. DRM reaction is progressed, and the effluent is analyzed by an online GC (Molecular Sieve 5a and Porapak Q columns) equipped with a thermal conductivity detector under Ar carrier gas. H 2 yield % and CO yield % were estimated by the following expressions: 3 | RESULTS AND DISCUSSION
The Raman spectra of 5NixM/YZr (x = 1-5 wt%) catalyst catalysts are shown in Figures 3 and S3. Zirconia supported nickel catalyst system, as shown in Figure S3A, had plenty of peaks for monoclinic ZrO 2 phase patterns at 178, 334, 380, 476, and 610 cm −1 . 17,18 In the monoclinic ZrO 2 phase, the peak at 476 cm −1 is more intense than the peak at 630 cm −1 . 18 On incorporation of 8 wt% Yttria along with ZrO 2 support, 5Ni/YZr catalyst had intense peak at 258 cm −1 for cubic ZrO 2 19 and comparable peaks at 476 and 630 cm −1 for tetragonal ZrO 2 . The peak at 476 and 630 cm −1 are common in both monoclinic and tetragonal zirconia. However, in the tetragonal phase, the peak intensity at 476 cm −1 is comparable to that at 630 cm −1 (equal or slightly greater). 18 Upon promotional addition of 1 wt% Rh; cubic ZrO 2 peak retains, the tetragonal ZrO 2 peaks are intensified, and a new broad peak from 500 to 580 cm −1 for the Rh 2 O 3 phase 20 also appeared ( Figure S3B). It is interesting to note that on increasing Rh loading, both peaks due to cubic ZrO 2 as well as Rh 2 O 3 phase decreases. The spent 5NixRh/YZr catalyst showed the band at 1340 and 1560 cm −1 for the "defect carbon band" (I D ) and "graphite band" (I G ), respectively 6 ( Figure 3B). Interestingly, upon increasing the loading of Rh from 2 to 3 wt%; intensity of the defect carbon band (I D ) decreases whereas upon 3 to 4 wt% Rh loading, the intensity of both bands (I D and I G ) decrease. It pronounces the effective role of Rh in carbon deposit (defective as well as graphitic type) oxidation. 5Ni5Rh/ YZr catalyst shows the broad Raman band at 500 to 580 cm −1 for Rh 2 O 3 ( Figure S3B). This indicates that Rh 2 O 3 phase again assembles at 5 wt% Rh loading.
of C-H in format species and 2925 cm −1 for the combination of symmetric stretching of COO and bending vibration of C-H are frequently observed over 5NixRh/YZr catalyst system. 4 It is interesting to note that upon 4 wt% Rh incorporation over yttria-zirconia-supported Ni catalyst, two additional peaks at 1270 and 1730 cm −1 are observed ( Figure 3D). The unpromoted catalyst (5Ni/YZr) also has such additional peaks, which are attributed to bidentate CO 2 chemosorbed species or bidentate carbonate species. 21,22 The intensity of these peaks is higher over the 5Ni4Rh/YZr catalyst than 5Ni/YZr catalyst.
The CO 2 temperature programmed desorption profile of 5NixRh/YZr (x = 1-5 wt%) catalysts is shown in Figure 4A. The CO 2 -TPD profile of 5NixRh/YZr (x = 1-5 wt%) catalysts can be broadly classified into two regions, namely a low-temperature region below 175°C and an intermediate temperature region from 175°C to 450°C. CO 2 is deadsorbed from surface hydroxyl (weak basic sites) and isolated O 2− species (moderate basic sites) at low-temperature regions and intermediatetemperature regions, respectively. 4,23 1 wt% Rh promoted yttria-zirconia-supported catalyst has pronounced "weak basic sites" regions and diffuse "moderate basic sites" regions. On increasing Rh from 1 to 3 wt%, the total basic site is increasing up to the maximum extent over 5Ni3Rh/YZr catalyst. Upon 4 wt% Rh loading, the weak basic site is dropped, but moderate basic sites continued to grow over 5Ni4Rh/YZr. Furthermore, 5 wt% Rh promoted yttria-zirconia-supported Ni catalyst shows a drop of total basic sites. H 2 temperature programmed reduction profile of 1 wt% Rh-promoted yttria-zirconiasupported Ni catalyst and unpromoted catalyst are shown in Figure 4B-D. The H 2 temperature reduction profile of 5Ni/YZr catalyst has three merge peaks centered about 315°C, 360°C, and 450°C. It indicates that 5Ni/YZr had three types of NiO-interacted surface species, which are reduced at different temperatures as per the extent of interaction with support. The reduction peaks at about 315°C, 360°C, and 450°C are for weakly, moderately and strongly interacted NiO species, respectively. 5Ni1Rh/YZr catalyst is reduced at quite an early temperature (222-326°C) whereas 5Ni/YZr catalyst is reduced at high temperature range (286-486°C). It indicates that the addition of Rh brings easier reducibility of the catalyst. Upon 4 wt% Rh addition over yttria-zirconia-supported Ni catalyst, a new peak at early temperature range (149°C) is observed, which is attributed to the reduction of Rh 2 O 3 into metallic Rh 24 ( Figure 4E). The other two reduction peaks over 5Ni4Rh/YZr catalyst are observed at 280°C and 315°C. Rh favors H 2 dissociation followed by the spill-over of H species from Rh to NiO, which may promote easy reducibility of NiO. So, the reduction peak at about 280°C and 315°C may be due to the reduction of "interacted NiO species" which is reduced by "spill over hydrogen (from Rh)" and "H 2 stream". 5Ni5Rh/YZr catalyst had a similar reducibility profile as like as 5Ni4Rh/YZr catalyst ( Figure S4).
It is worth noting herein that in DRM, CO 2 is an oxidant, and the active site is metallic Ni. What is the oxidizing strength of CO 2 ? Does it oxidize the carbon deposit as well as metallic Ni? If it oxidizes the metallic Ni (into NiO), then the catalytic active site would be lost, and the catalyst becomes inactive. To confirm this, we have carried out "H 2 TPR followed by CO 2 TPD and then again H 2 TPR" over 5NixRh/YZr (x = 0, 1, 4, 5) catalysts ( Figures 4C-E and S4). During initial H 2 TPR treatment, reducible interacted NiO-species is reduced into metallic Ni.
Furthermore, the CO 2 TPD of the reduced catalyst may provide the chance of oxidation of metallic Ni (into interacted NiO species) over the catalyst surface. After oxidation, it may be possible that these NiO species (which are oxidized by CO 2 ) may have a different extent of interaction with support than earlier. Whatsoever, such NiO species can be quantified by reducing it under H 2 TPR treatment again. After successive "H 2 -TPR-CO 2 -TPD-H 2 -TPR" cyclic treatment of 5Ni/YZr catalyst, a small reducible peak at an early temperature (150°C) is observed ( Figure 4C). This peak is abbreviated to reducible free NiO species. 25 However, no such peak is observed after "H 2 -TPR-CO 2 -TPD-H 2 -TPR" cyclic treatment of Rh-promoted yttria-zirconia-supported Ni (5NixRh/YZr; x = 1, 4, 5) catalyst ( Figures 4D,E and S4). These observations indicate that over yttria-zirconiasupported Ni catalyst, some portion of metallic Ni is oxidized into NiO under CO 2 stream whereas, over Rh promoted yttria-zirconia-supported Ni catalyst, catalytic active sites remain stabilized under CO 2 stream.
The TEM images of fresh and spent 5Ni4Rh/YZr catalyst are shown in Figure 5. No fibrous carbon deposit is seen in the image over the spent catalyst system. The Ni particle size in the fresh and spent catalyst sample is calculated by TEM image on a 200 nm scale. The Ni particle size over the catalyst has grown from 5.80 to 6.90 nm after the reaction.

| DISCUSSION
Zirconia-supported Ni catalyst was efficient toward DRM reaction as it showed an initial 50% H 2 yield which slowed down continuously as time on stream progressed. At the end of 420-min time on stream, H 2 yield declined to 43.5%. The low activity of the 5Ni/Zr catalyst is due to the unstable monoclinic ZrO 2 phase against high temperature during the DRM. Interestingly, by incorporating yttria into zirconia, the support yttria-zirconia is stabilized by forming cubic zirconia and tetragonal zirconia phases (verified in XRD and Raman). The 5Ni/ YZr catalyst had bidentate CO 2 chemosorbed species, monodentate carbonate species and format species even at room temperature (observed in IR). These DRM favorable surface characteristic over 5Ni/YZr ensures high and constant H 2 yield (∼71%) up to 420-min time on stream. Under the "H 2 -TPR-CO 2 -TPD-H 2 -TPR" cyclic experiment over 5Ni/YZr catalyst, it is found that some portion of catalytic active sites (metallic Ni) is oxidized under the CO 2 stream and becomes inactive toward DRM. Overall, the inactive portion of the catalyst limits the higher activity (H 2 yield) beyond the 70s.
Some general observations are noticed over Rhpromoted yttria-zirconia-supported Ni catalyst. The Rhpromoted yttria-zirconia catalyst is more easily reducible than the unpromoted catalyst (verified by H 2 TPR). Its active site remains stabilized under CO 2 stream during DRM. The catalytic activity of the Rh-promoted catalyst system is improved as time on the stream is progressed. Upon 1 wt% Rh addition over yttria-zirconia Ni catalyst, tetragonal zirconia phases are pronounced, and bidentate CO 2 chemosorbed species are depleted (with respect to 5Ni/YZr). 5Ni1Rh/YZr catalyst shows relatively inferior initial catalytic activity. However, H 2 yield improves during the longer time on stream and at the end of 420 min, it is ∼72%. Up to the addition of 2 wt% Rh in yttria-zirconia-supported Ni catalyst; less crystallinity (high dispersion), enhanced "weak basic sites" (verified by CO 2 TPD) and less carbon deposit (verified by Raman and thermogravimetric analysis) are addressed by characterization results. Over 5Ni2Rh/YZr catalyst, initial activity jumps to 79% H 2 yield, which further improves to ∼81% during 420 min. 5Ni3Rh/YZr catalyst has an optimum amount of basic sites and relatively lower crystallinity with respect to 5Ni2Rh/YZr. The carbon deposit over spent 5Ni3Rh/YZr catalyst is marginally lower than 5Ni2Rh/YZr, but the amount of defective carbon type over spent 5Ni3Rh/YZr catalyst is noticeably lower than 5Ni2Rh/YZr. The optimum basicity, relatively lower crystallinity/carbon deposit over 5Ni3Rh/YZr, helps to attain a constant 84% H 2 yield during 420-min time on stream. The 5Ni4Rh/YZr catalyst had acquired moderate strength basic sites and a wide variety of CO 2 -interacted species (formate, bidentate, and monodentate CO 2 chemosorbed species). The intensity of bidentate CO 2 -adsorbed species is maximum at 5Ni4Rh/ YZr. 5Ni4Rh/YZr also showed enhancement of moderate strength basic sites. H 2 TPR profile of 5Ni4Rh/YZr indicates the presence of two types of reducible species. One is reducible Rh 2 O 3 species which is reduced under H 2 stream, and another is reducible NiO species which is reduced under H 2 stream as well as under spill-over hydrogen (from Rh). Furthermore, the amount of carbon deposit over spent 5Ni4Rh/YZr catalyst is not only found less but also the quantity of defective as well as graphitic type of carbon is substantially decreased (verified by Raman) than 5Ni3Rh/YZr. Altogether, the presence of more than one type of active sites (Ni, Rh), pronounced population of bidentate CO 2 -adsorbed species (along with another CO 2 -interacted species), and deficit of graphic carbon deposit (during DRM) enable 5Ni4Rh/ YZr to perform best towards DRM. 5Ni4Rh/YZr catalyst showed 85%-87% H 2 yield during 420-min time on steam, and activity does not drop below 80 up to 27 h time on stream. It is also noticeable that CO yield is found to be less than H 2 yield over 5Ni4Rh/YZr catalyst during entire time on stream study. It indicates efficient suppression of reverse water gas shift reaction over 5Ni4Rh/YZr catalyst system.
Certainly, upon 5 wt% Rh loading over yttria-zirconia-supported Ni catalyst, metallic Ni species is quite stable under the CO 2 stream like 5Ni4Rh/YZr catalyst, and the spent catalyst has the least amount of carbon deposit among rest catalyst. But the Rh 2 O 3 phase assembles over 5Ni5Rh/YZr, which indicates less dispersion of Rh. Again, 5Ni5Rh/YZr had the least amount of basic sites than the rest catalyst, which indicates the least interaction of CO 2 with the catalyst surface. Overall, the catalytic activity over the 5Ni5Rh/YZr catalyst drops relatively but remains ∼81%-83% during the tested time on stream. The CO yield also enhances over H 2 yield over 5Ni5Rh/YZr catalyst ( Figure S2).

| CONCLUSION
Yttria-zirconia-supported Ni catalyst has stable ZrO 2 phases and a wide presence of "surface CO 2 -interacted species" but a small portion of the catalytic active (metallic Ni) tends to oxidize (inactive) under CO 2 stream. Altogether, 5Ni/YZr ensures higher activity than 5Ni/Zr but not beyond the seventies (∼71% H 2 yield). Over Rh-promoted yttria-zirconia-supported Ni catalyst, catalyst is more easily reducible than an unpromoted catalyst. Over 5Ni/YZr, basicity is grown up to 3 wt% Rh loading; crystallinity is dropped up to 3 wt% Rh loading and carbon deposit is declined up to 5 wt% Rh loading. Optimum basicity, relatively lower crystallinity, and marked decreased carbon deposit over 5Ni3Rh/YZr resulted in a constant 84% H 2 yield during 420-min time on stream. 5NixRh/YZr (x = 2, 3, 4) catalyst ensures high catalytic performance (>80% H 2 yield); 4 wt% Rh loading is optimum. At 5Ni4Rh/YZr catalyst, additional surface features like the presence of more than one type of active sites (Ni, Rh), depletion of graphitic carbon proportion, enhancement of moderate strength basic sites, and the optimum population of bidentate CO 2 -adsorbed species are found favorable towards DRM. 5Ni4Rh/YZr shows 85%-87% H 2 yield and a little bit less CO yield during 420-min time on stream. In a long time of the study, it is found that about 80% H 2 yield is maintained up to 27 h time on stream. This indicates efficient suppression of RWGS reaction also over 5Ni4Rh/YZr. In the case of