Hydrogen Production from Gadolinium-Promoted Yttrium-Zirconium-Supported Ni Catalysts through Dry Methane Reforming

Hydrogen production from dry reforming of methane (DRM) not only concerns with green energy but also involves the consumption of two greenhouse gases CH4 and CO2. The lattice oxygen endowing capacity, thermostability, and efficient anchoring of Ni has brought the attention of the DRM community over the yttria-zirconia-supported Ni system (Ni/Y + Zr). Herein, Gd-promoted Ni/Y + Zr is characterized and investigated for hydrogen production through DRM. The H2-TPR → CO2-TPD → H2-TPR cyclic experiment indicates that most of the catalytic active site (Ni) remains present during the DRM reaction over all catalyst systems. Upon Y addition, the tetragonal zirconia-yttrium oxide phase stabilizes the support. Gadolinium promotional addition up to 4 wt % modifies the surface by formation of the cubic zirconium gadolinium oxide phase, limits the size of NiO, and makes reducible NiO moderately interacted species available over the catalyst surface and resists coke deposition. The 5Ni4Gd/Y + Zr catalyst shows about ∼80% yield of hydrogen constantly up to 24 h at 800 °C.


INTRODUCTION
The green routes acclaimed for hydrogen production concern with electrolysis of water, steam reforming of methane (H 2 / CO ∼3), partial oxidation of methane (H 2 /CO ∼2), and dry reforming of methane (DRM) (H 2 /CO ∼1). 1 Although the H 2 /CO ratio of DRM is low, it has drawn huge attention because by this route, two global warming gases (CH 4 and CO 2 ) are precisely converted. Therefore, it bears the hope of reduction of the concentration of global warming gases. Additionally, this route generates energy prime hydrogen-rich syngas.
DRM reaction is a highly endothermic reaction, and it is catalyzed by different active metals Pt, Pd, Ru, Co, and Ni. The methane dissociation energy is found in the order Ni < Pd = Pt. 2 Noble metals show better activity, but they are extremely expensive; a good replacement would be nickel, which has shown the best activity among transition metals. The methane interaction energy over Ni is 18 kcal/mol, which is 25 times higher than that of Co (methane interaction energy over Co is 0.7 kcal/mol). 3 However, Ni is highly unstable at high temperatures where it is sintered into large sizes, stimulates large coke deposition, and finally ends with deactivation. Silica, alumina, aluminosilicate, and zirconia supports have a high metal−support interaction with Ni, and so Ni particles are held over these supports even against very high temperatures. Silica was used as a neutral support, alumino-silicate support offered acidity at a high Si/Al ratio, 4 and alumina support was known to provide acidity to the catalyst system. 5 Promotional addition of ZrO 2 over an Al 2 O 3 -supported Ni catalyst was found to increase Ni dispersion and decrease coke deposition. 6,7 Zhang et al. carried out CH 4 -temperature-programmed surface reaction over Ni/ZrO 2 and detected CO. 8 This observation indicates the involvement of lattice oxygen of ZrO 2 in carbon deposit oxidation. Therefore, ZrO 2 can endow lattice oxygen, withstand high-temperature conditions, and provide an efficient Ni anchoring effect by Zr +4 .
A bimetallic Ni-Co particle dispersion over the ZrO 2 -MgO (9:1) support was found effective in H 2 yield (85%) in the presence of 7 mol % oxygen along with a 3:1 CO 2 and CH 4 ratio. 9 However, in DRM, CO 2 is a promising oxidizing agent for oxidation of CH 4 and use of oxygen is completely restricted. Furthermore, "Ni impregnated over ZrO 2 " was employed by different groups for DRM but due to instability of ZrO 2 phases against high temperature, H 2 yield never exceeded more than 50%. 10−12 Silica-promoted ZrO 2 -supported Ni catalysts have brought attention due to 1.5% H 2 yield up to 15 h at a very low temperature (400°C) compared to the conventional 700−800°C reaction temperature. 13 ZrO 2 -rich ceria (28 wt %)-supported Ni catalysts had better stability under a thermally reductive environment than individual metal oxide-supported Ni catalysts. 14 They also induced Ni dispersion and gave 35% H 2 yield up to a 24 h time on stream (TOS) at a 700°C reaction temperature 14 and about ∼40% H 2 yield at 750°C for 50 h. 15 The presence of ceria along with ZrO 2 had a stabilizing effect on tetragonal ZrO 2 , enhancing lattice oxygen mobility and increasing reducibility in favor of DRM and ceria-zirconia-supported Ni catalysts, conveying more than 57% H 2 yield. 16−18 Upon 0.02 mol Ca incorporation with 0.04 mol Ni and 0.1 mol ZrO 2 (via refluxmediated coprecipitation), both the surface parameter and the metal−surface interaction were enhanced and about 65% H 2 yield was obtained. 10 Lanthana addition in the ZrO 2 support brought stability of the tetragonal phase of ZrO 2 and formation of La 2 O 2 CO 3 species for efficient removal of carbon deposit. A 10 wt % lanthana−90 wt % ZrO 2 -supported Ni catalyst showed 74% H 2 yield up to 400 min. 19 Further promotional addition of basic Ca over a lanthana-zirconia-supported system brought a more efficient CO 2 interaction, and the Ni−O-Ca interphase helped Ni to restore Ni particles to the original state 20 for efficient CH 4 decomposition. It brought interest in lowtemperature DRM reaction as it showed ∼6.1% H 2 yield at a 450°C reaction temperature after 30 min of reaction. Ceria promotional addition over lanthana-zirconia-supported nickel reduced the band gap and added additional lattice oxygen mobility, prominent CH 4 decomposition sites, resistance of ZrO 2 phase transition, and prominent interaction of CO 2 . 11 It gave 75% H 2 yield at 460 min TOS. A chromium-promoted lanthana-zirconia-supported system stabilized the tetragonal ZrO 2 phase, stabilized the lanthana-zirconia phase, and showed excellent oxygen replenishment capacity as reduced NiO was re-oxidized by CO 2 up to the optimum level. It showed about 80% H 2 yield. 21 The promotional addition of Ce over the tungsten-zirconia support also had a stable tetragonal zirconia phase. 22 It had additional basic sites, a ceria tungsten oxide phase for instant release of oxygen, and an excellent reoxidizing capacity of reducible NiO. It resulted in 78% H 2 yield at 420 min TOS. Y addition with zirconia was found to stabilize 12 tetragonal ZrO 2 as well as facilitate additional O 2− species in the lattice. A 15 wt % yttria−85 wt % zirconiasupported Ni catalyst had a wide range of basic sites, and it showed 78% H 2 yield. A nonmetal oxide−metal oxidesupported Ni catalyst, as well as nonmetal-promoted DRM catalyst systems, was also tried; however, we are limiting our literature to metal-promoted and binary metal oxide-supported DRM catalysts. 23,24 Gd was used as a spacer over iron oxide-based catalysts, which caused the increase in specific area. Its presence inhibited the reduction of Fe +3 to Fe +2 and the formation of iron carbide. 25 Gadolinium ferrites had p-type conductivity, which favors the chemisorption of CO molecules. 39−41 Gadolinium-doped ceria depressed the bulk but increased the dynamic oxygen exchange capacity (OEC). 26,27 There is a strong 3d−4f electron exchange, and a spin−orbit interaction is noted with Gd and Ni. 28,29 A 0.2 wt % Gd-doped MCM-41supported Ni catalyst had a GdNi 5 phase (2−10 nm crystallite), which prevented the agglomeration of Ni and provided oxygen atoms for carbon deposit oxidation. 29 The addition of just 0.1 wt % Gd over the MCM-41 (or Al 2 O 3 )supported Ni catalyst caused retention of the size of active metal Ni before and after the reaction ensuring coke resistance and high catalytic performance. 30 Over silica-supported Ni, Gd/Ni = 0.45 was found to enhance the DRM activity due to enhanced CO 2 adsorption (formation of surface carbonate species), increased metal−support interaction, and Ni dispersion. 31,32 The promotional addition of 0.5% Gd 2 O 3 caused a stronger interaction with Ru (causes smaller Ru particle) over the Zr 0.5 Ce 0.5 O 2 support and decreased the apparent activation energy of methane conversion. 33 A 1 wt % Gd-promoted yttria-supported Ni catalyst had easy reducibility, high surface area, high basicity, and strong carbon resistance compared to the nonpromoted one. 34 Zhang et al. prepared a Gd-promoted alumina-silica-supported Ni catalyst by the one-pot Pechini method and found that 1.2 wt % Gd addition tended to weaken NiAl 2 O 4 formation, facilitate the reduction of Ni, and enhance the catalytic activity of CH 4 :CO 2 :N 2 = 35:35:20 gas feed. 35 ZrO 2 -Gd 2 O 3 (90:10 mol ratio) solid solutions were sinter reactive. 36 The Zr +4 cation was too small (r = 0.084 nm) to support a full eightfold oxygen coordination. Increasing the dopant size like Gd +3 led to an increase in the concentration of the vacancies. Among Gd +3 , Y +3 , Yb +3 , and Nd +3 , doping of Gd +3 caused an increase in conductivity. 37 Overall, Gd was successfully incorporated in various Nicontaining supports like silica, alumina, alumina−silica, yttria, and zirconia and tested for DRM. The oxygen endowing capacity of zirconia, stabilization of zirconia by yttria, and oxide layer enrichment over yttria-zirconia has drawn much interest to use yttria-zirconia as support for DRM. Herein, we have investigated a Gd-promoted yttrium-zirconia-supported Ni catalyst system for hydrogen production through DRM. Again, a simple catalyst preparation procedure is needed to make it handy for less skilled workers in industry. Here, the catalyst is prepared through simple steps like mechanical mixing followed by calcination. The phase distribution, Ni coordination environment and band gap between valence and conduction bands, CO 2 -bonding surface species, and presence of reducible NiO-interacting species are studied by X-ray diffraction, Raman spectroscopy, UV−vis spectroscopy, and infrared spectroscopy, respectively. The retention of Ni active sites during DRM reaction are studied by a H 2 -TPR → CO 2 -TPD → H 2 TPR cyclic experiment over a fresh catalyst. Morphology is depicted by transmission electron microscopy, and last weight loss (%) of the spent catalyst system is investigated by thermogravimetric analysis. By comparison of different activities (like H 2 -yield (Y Hd 2 ), CO yield (Y CO ), CH 4 conversion (C CHd 4 ), and Y Hd 2 /C CHd 4 )), the possible DRM reaction mechanism and competitive RWGS reaction are outlined. Finally, a scientific correlation of characterization results and catalytic activity over the Ni-Gd/Y + Zr catalyst system is determined.

Catalyst Preparation.
The wet impregnation method was used to prepare the zirconium-supported Ni catalyst, yttrium-zirconium-supported Ni catalyst, and gadoliniumpromoted yttrium-zirconium-supported nickel catalyst. Yttrium and zirconium oxides were mixed together mechanically. An aqueous solution of (5 wt %) nickel nitrate precursor and (1.0, 2.0, 3.0, 4.0, 5.0 wt %) gadolinium nitrate using purified water were then added to the support (8 wt % yttrium-zirconium). The prepared mixture was stirred at 80°C and dried at 120°C overnight in an oven. Subsequently, the dried product was calcined at 600°C for 3 h. The zirconium-supported Ni catalyst and gadolinium-promoted yttrium-zirconium-supported nickel catalyst were abbreviated as 5Ni/Zr and 5NixGd/Y + Zr (0, 1, 2, 3, 4, 5), respectively. The scanning electron microcopy image and energy-dispersive X-ray (EDX) elemental analysis of the 5Ni4Gd/Y + Zr and 5Ni5Gd/Y + Zr catalysts are shown in Figure S1. SEM images show the salty texture of the catalysts, and EDX analysis shows the presence of all claimed elements in synthesis. Upon increasing Gd loading, the atomic percentage of Gd is also found to increase. The surface area, pore volume, and pore diameter of the Gdpromoted catalysts were found lower than the unpromoted catalysts ( Figure S2). No substantial changes in surface parameter is observed upon increasing Gd loading. In the Gd-promoted catalyst system, the surface area, pore volume, and pore diameter were typically in the 26−27 m 2 /g, 14−16 cm 3 /g, and 21.36−23.49 nm-pore-diameter ranges, respectively.

Catalyst Activity Test.
The dry reforming methane catalytic activity test started with a 1 g catalyst put in a stainless-steel vertical fixed tubular reactor (9.1 mm i.d. and 30 cm long) (PID Eng & Tech Micro Activity) using a ball of glass wool. The reaction was carried out under atmospheric pressure. A K-type stainless sheathed thermocouple was used to maintain the temperature of the reaction. The catalysts were activated with reductive treatment under the flow of hydrogen (20 mL/min) for 60 min at 600°C. The mixture of feed gas was CH 4 /CO 2 /N 2 at 30, 30, 10 mL/min (with a space velocity of 42,000 mL/h·g cat ) passed through the catalyst at 800°C. DRM reaction was progressed, and the effluent was analyzed by an online GC-2014 Shimadzu (Molecular Sieve 5A and Porapak Q columns) equipped with a thermal conductivity detector under Ar carrier gas. The H 2 yield %, CO yield %, CH 4 conversion, CO 2 conversion, and carbon formation rate were determined by the following expressions.

Catalytic Activity Results.
The catalytic activities of the promoted and nonpromoted catalysts at 800°C for 420 min are shown in Figure 1A. The zirconia-supported Ni catalyst (5Ni/Zr) shows the lowest H 2 yield (50%), which drops to 43% at the end of 420 min. When yttrium is added to the zirconia support, the catalytic activity jumps to 70% H 2 yield, which remains constant for up to 420 min. This means that 5Ni/Y + Zr got higher and stable catalytic activity than the 5Ni/Zr catalyst. 1 wt % addition of the Gd promoter to the yttrium-zirconia-supported Ni catalyst (5Ni1Gd/Y + Zr) results in further progress of catalytic activity to 71%. Incorporating more wt % of the Gd promoter up to 4 wt % into the yttrium-zirconia-supported Ni catalyst (5NixGd/Y + Zr; x = 2,3,4), the H 2 yield is increased progressively. 5Ni4Ga/ Y + Zr shows the highest H 2 yield (78%) up to 420 min. More than 4 wt % Gd incorporation into the yttrium-zirconiasupported Ni catalyst results in a drop in catalytic activity. 5Ni4Gd/Y + Zr is found best in mean of H 2 yield toward DRM, and so this catalyst system is further investigated at different reaction temperatures for DRM reaction. It is found that the H 2 yield increases monotonically as the reaction temperature is increased from 500 up to 800°C over the 5Ni4Gd/Y + Zr catalyst ( Figure 1B). The 5Ni4Gd/Y + Zr catalyst is further tested for a long time on stream, and 80% H 2 yield is obtained at 800°C constantly for 24 h of time on stream ( Figure 1C).
As per the DRM reaction, the H 2 yield and CO yield should be equal. However, it is observed that the CO yield always remains higher than the H 2 and CO 2 yields over each catalyst system ( Figure S3). It indicates that along with DRM reaction, another H 2 -consuming reaction also takes place over the catalyst system. A reverse water gas−shift reaction (RWGS) (CO 2 + H 2 → CO + H 2 O) is a thermodynamic feasible reaction, which happens in the same temperature range as that of DRM. 38 It is noticeable that during the 420 min time on stream, the CO yield is 30−40% higher than the H 2 yield over the 5Ni/Zr catalyst, 7−10% higher than the H 2 yield over the 5Ni/Y + Zr catalyst, and 4−5% higher than the H 2 yield over the 5NixGd/Y + Zr (x = 4,5) catalyst system. It indicates the successive suppression of the RWGS reaction over the gadolinium-promoted yttria-zirconia-supported Ni catalyst system.
The H 2 yield, CH 4 conversion, CO yield, and CO 2 conversion over the 5Ni/Zr and 5NixGd/Y + Zr (x = 0, 1, 2, 3, 4, 5) catalyst system at the end of 60 min are shown in Figure S3C. If it is proposed that CH 4 is dissociated into C + 2H 2 , then the stoichiometric Y Hd 2 /C CHd 4 ratio for CH 4 decomposition should be 2. Y Hd 2 /C CHd 4 < 2 may be indicative of the involvement of H 2 in the RWGS reaction potentially. It may also indicate the formation of CH y and expulsion of H 2 gas after decomposition of CH 4 . 39,40 In our case, the H 2 yield (Y Hd 2 ) is found always less than CH 4 conversion (C CHd 4 ) and the CO yield remains less than CO 2 conversion. Simply, the Y Hd 2 / C CHd 4 ratio remains between 0.95 and 0.99 over the 5Ni/Zr and 5NixGd/Y + Zr (x = 0, 1, 2, 3, 4, 5) catalyst system. Earlier, we have found that the RWGS reaction is suppressed potentially over the 5Ni4Gd/Y + Zr and 5Ni5Gd/Y + Zr catalyst. However, over these catalysts, also the Y Hd 2 /C CHd 4 ratio remains between 0.95 and 0.97. It indicates that over the 5NixGd/Y + Zr (x = 0, 1, 2, 3, 4, 5) catalyst system, CH 4 is decomposed into CH y and ( ) The size of cubic NiO crystallite is about ∼13 nm over 5Ni/Y + Zr. This means tetragonal yttria stabilize the tetragonal phase of zirconia by forming a mixed tetragonal zirconia yttrium oxide phase. Simply, doping of zirconia with a lower-valence ion Y +3 (radius of Y +3 = 1.06 Å) was found to stabilize the tetragonal phase of zirconia above 400°C. 41,42 The valence of yttrium is smaller than that of zirconium. Lower-valence Y +3 serves as a substitute for the Zr +4 ion; the Y +3 ion is incorporated into the zirconia crystallite structure and stabilizes the tetragonal phase of zirconia. Furthermore, tetragonal phases of yttria and tetragonal phases of zirconia turn into the formation of a mixed tetragonal zirconia-yttrium oxide phase. The cubic zirconium gadolinium oxide phases (at 2θ =24.34, 28.22, 37.19, 62.66°; JCPDS card reference number 01-080-0469) are also noticed over the 5Ni1Gd/Y + Zr catalyst, but these peaks are generally merged with the zirconia phase peaks (Figure 2A−C). It seems that due to the engagement of the ZrO 2 phase with Gd, the peak intensity of the ZrO 2 -related phases decreased upon Gd introduction. On further 3 wt % Gd addition, the peak intensity not only decreased but also shifted toward the lower brag angle relatively, which indicated the lattice expansion on 3 wt % Gd loading ( Figure 2D,E). On gadolinium addition at 3−5 wt %, the peak intensity decreased continuously ( Figure 2F,G). It indicates that Gd addition decreases the crystallinity and increases the amorphousity. Interestingly, on 4 wt % Gd promotional addition, the NiO crystallite size is a minimum of 7.5 nm (Table 1). In the spent catalyst system, all crystalline phases decrease and no crystalline carbon phases are observed ( Figure S4). Over the spent 5NixGd/Y + Zr (x = 0, 1, 2, 3, 5) catalyst system, the metallic Ni phase (at Bragg angle 44.50°; JCPDS card reference number 00-004-0850) is evident. Over the spent 5Ni1Gd/Y + Zr catalyst, a 15.6 nm metallic Ni phase is noticed whereas the Ni size is optimized to 7.6 nm over the spent 5NixGd/Y + Zr (x = 2−3) catalyst (Table 1). Interestingly, over the spent 5Ni4Gd/Y + Zr catalyst, no diffraction peaks for the metallic Ni phase is observed. It indicates the dispersion of metallic Ni over the spent catalyst system. The role of Gd in size optimization of Ni species was noticeable in the fresh as well as spent catalyst system.

Raman and Infrared
Spectroscopy. The XRD results are also supported by Raman spectra. The Raman profile of the promoted and unpromoted catalysts is shown in Figure 3A. The peaks at 179, 334, 380, 476, and 610 cm −1 characterize the monoclinic zirconia where the 476 cm −1 peak is larger than that of 610 cm −1 . 43 The peak at 630 cm −1 characterizes tetragonal zirconia for the zirconia-supported nickel catalyst. The yttria-zirconia-supported Ni catalyst has a low-intensity monoclinic zirconia peak but an emerging tetragonal zirconia peak at 146, 260, and 625 cm −1 . 44 In the 1 wt % Gd (over 5Ni/ Y + Zr catalyst), the intensities of all major peaks decrease ( Figure 3A). These observations are in line with the XRD results (discussed above). Infrared spectra of fresh promoted and unpromoted catalysts are shown in Figure 3B 11 The free NiO in a cubic lattice was reported at 433 cm −1 wavenumbers, whereas in the ZrO 2supported Ni catalyst, a Ni−O− vibration peak at a lower wavenumber (420 cm −1 ) was observed. 45 It indicates the deformation of cubic NiO species and formation of "NiOinteracted support" species, which result in the weakening of the Ni−O bond and vibration of Ni−O at a lower wavenumber than in free Ni−O in a cubic lattice. However, in the yttriazirconia-supported Ni catalyst, all Ni−O and Zr−O vibration peaks disappear. It indicates that the presence of yttria brought a major change in bonding patterns. IR of the 5NixGd/Y + Zr (x = 0,1,2,3,4,5) catalyst system shows the presence of CO 2 as monodentate carbonate species at 1388 and 1516 cm −146 as well as physically adsorbed CO 2 at 2343−2352 cm −1 . 11 3.2.3. Ultraviolet−Visible Spectroscopy. The UV spectra and band gap of the 5NixGd/Y + Zr (x = 0, 1,2,3,4,5) catalyst system are shown in Figure 4. The zirconia-supported nickel fresh catalyst shows a peak at 230 nm for charge transfer of O −2 to Zr +4 , "258 and 286 nm" for charge transfer of O −2 to Ni +2 in the octahedral symmetry, 24,47 peaks at 371, 410, and 451 nm for the d−d transition from 3 A 2g to 3 T 1g (P) 48 of Ni +2 in the octahedral environment, and peak at 717 nm for the d− d transition from 3 A 2g to 3 T 1g (F) of Ni +2 in the octahedral environment. 12 Altogether, UV−vis spectra confirm the octahedral coordination of Ni in the zirconia-supported catalyst. It has a 3.11 eV band gap. The yttria-zirconiasupported catalyst had the same type of peak pattern and the least band gap (2.23 eV) with respect to the rest of the catalysts. It indicates that yttria incorporation does not change the coordination environment but the charge transfer of O −2 from the valence band to the conduction band becomes easier. Upon Gd loading, a peak at 230 nm (for charge transfer of O −2 to Zr +4 ) was diffused, which indicates that Gd loading inhibits such transition. The UV−vis spectra of 2 wt % Gd and 5 wt % Gd are noticed by the absence of the d−d transition peaks. Upon 4 wt % Gd loading, the lowest band gap (or similar to the nonpromoted catalyst; 5Ni/Y + Zr) is noticed (Table S1).

Temperature-Programmed Study.
In an XRD study of the fresh catalyst system, a cubic NiO phase is found, which is reducible (into metallic Ni) under H 2 stream. The amount of reducible NiO over different temperatures reflects the extent of interaction of such NiO species over the support. Now, to understand the type of reducible species over the fresh catalyst system, H 2 -TPR is discussed in more detail. Al-Fatesh et al. showed that H 2 -reduction peaks before 200°C are for reducible free NiO species over the zirconia-supported Ni catalyst. 49 In our case, there are no reduction peaks before 200°C in the 5NixGd/Y + Zr (x = 0−5) catalyst system ( Figure  5A and Figure S5). It indicates that upon introducing yttria or both yttria and gadolinium, all NiO species interact with the support. The H 2 -TPR peak pattern of the 5NixGd/Y + Zr (x = 0, 1, 4, 5) catalyst system is composed of peaks about 330°C for reducible "NiO species weakly interacting with the support," shoulder peaks about 350°C for reducible "NiO species moderately interacting with the support", and hightemperature peaks about 465°C for reducible "NiO species  strongly interacting with the support". 21 It is interesting to note that on introducing 1 wt % Gd, peak patterns of lower and intermediate temperatures increase at the expense of a hightemperature peak. This means that the amount of weak and moderately reducible NiO-interacted species is growing over the surface upon Gd introduction whereas hardly reducible NiO-interacted species are decreasing. On increasing loading of up to 5 wt % Gd, H 2 consumption is increasing in lower and intermediate temperatures continuously. It indicates that the amount of reducible NiO species keeps increasing on Gd increase of loading. Before the DRM reaction, the catalyst is pre-treated under H 2 and NiO is reduced into Ni, which is a catalytic active site. Furthermore, CH 4 , CO 2 , and N 2 gas feed are passed over the catalyst at a reaction temperature of 800°C. CO 2 is an oxidizing gas, which can oxidize the carbon deposit as well as metallic Ni. Oxidation of Ni to NiO may turn the catalyst inactive because of depletion of the active site. It is important to note how our catalyst system behaves in front of CO 2 stream. To understand this, the 5NixGd/Y + Zr (x = 0, 2, 4, 5) catalyst is sequentially treated with H 2 -TPR then CO 2 -TPD and then H 2 -TPR (known as a H 2 -TPR → CO 2 -TPD → H 2 TPR cyclic experiment, depicted in Figure 5B−D and Figure S6). It is found that fresh catalysts consumed H 2 prominently during reduction of the H 2 -TPR and catalysts. Thereafter, CO 2 -TPD of the reduced catalysts shows very little desorption of CO 2 by the reduced catalysts. Finally, treating the catalyst again with H 2 -TPR in sequence (H 2 -TPR → CO 2 -TPD → H 2 TPR), there is very little consumption of H 2 . It indicates that most of Ni is not oxidized and remains present under CO 2 stream.
TGA profiles of promoted and unpromoted spent catalysts after 7 h of DRM reaction are shown in Figure 6A. The unpromoted catalysts show a 19% weight loss. 1, 2, 3, 4, and 5  wt % Gd-promoted catalysts show 14, 11, 14, 6, and 13% weight losses. This result shows that 4 wt % Gd-promoted catalyst has the lowest carbon deposition over the surface. Even after a 25 h reaction, the spent 5Ni4Gd/Y + Zr catalyst only showed 12.5% weight loss ( Figure 6B). The carbon formation rate over different catalysts is shown in Figure 6C. The 5Ni4Gd/Y + Zr catalyst shows a minimum carbon formation rate (0.001255 mg/min) than other catalysts. To understand the type of carbon species present over the spent 5Ni4Gd/Y + Zr catalyst (after 24 h of DRM reaction), O 2 temperature-programmed oxidation is carried out ( Figure S7). The catalyst showed a single peak in the region of 400−700°C (peak maxima at 615°C) for β-carbon species. 12,22 The Raman spectra of the spent 5Ni4Gd/Y + Zr catalyst showed a "defect carbon band" (I D ) at 1336 cm −1 , a "graphite band" (I G ) band at 1573 cm −1 , and a 2D band at 2673 cm −1 52,59,60 ( Figure S8). Among all the Gd-promoted catalysts, the 5Ni4Gd/Y + Zr catalyst had a minimum peak intensity for the I D band, I G band, and 2D band ( Figure 6D).

Transmission Electron
Microscopy. TEM images of fresh and spent samples of the 4 wt % gadolinium-promoted yttrium-zirconia-supported nickel catalyst are shown in Figure  7. The fresh catalyst shows a spherical shape with a particle size of 5.21 nm, while the spent catalyst shows a carbon nanotube with encapsulated carbon particle size 6.79 nm. After catalytic reaction, the particle size is increased.

DISCUSSION
The ZrO 2 support is known for the presence of dual sites (acid−base sites), which can withstand and hold the active metal species Ni against high-temperature DRM conditions. 51 From X-ray diffractograms and Raman spectra, mostly unstable monoclinic zirconia phases and cubic NiO crystallites of crystallite size 7.6 nm are confirmed over the Ni/ZrO 2 catalyst. The UV profile of the catalyst confirms the octahedral sites of Ni +2 over the ZrO 2 support and the 3.11 eV band gap between the valence and conduction bands. Infrared spectra confirm the presence of CO 2 -interacting surface species. The "H 2 -TPR → CO 2 -TPD → H 2 TPR" cyclic experiment over the fresh catalyst system shows that most of the Ni (catalytic active sites) remained present even under oxidizing CO 2 stream. However, the catalytic activity toward DRM was found inferior (43% H 2 yield at 420 min). Al-Fatesh et al. showed that zirconiasupported Ni catalysts had reducible free NiO species (less interacting NiO species) over the catalyst surface. 49 Overall, the inferior activity of Ni/Zr toward DRM can be claimed to show unstable monoclinic zirconia phases, less-interacting NiO species, and a huge carbon deposit (∼19% weight loss in TGA). The CO yield over the 5Ni/Zr catalyst remains 30− 40% higher than the H 2 yield during the 420 min time on stream. It indicates that after DRM, the major competitive reaction is RWGS over the 5Ni/Zr catalyst.
Yttria addition along with zirconia support has brought major physiochemical changes in the catalytic property of the surface. XRD and Raman confirm the presence of a stable tetragonal zirconia phase through the formation of the tetragonal zirconia-yttrium oxide (mixed oxide) phase. The band gap between the valence band and the conduction band decreased sharply after yttrium incorporation (2.23 eV). 5Ni/Y + Zr had reducible "NiO surface-interacted species" in which strongly interacted NiO species are a majority. These favorable surface properties over the 5Ni/Y + Zr catalyst push the H 2   Possibly the rate of carbon formation may properly match the rate of carbon diffusion away from the catalytic active center, which affects the catalytic activity to a less extent. 52 The CO yield remains 7−10% higher than the H 2 yield. It indicates that the RWGS reaction is suppressed greatly over the 5Ni/YZr catalyst than the 5Ni/Zr catalyst. Definitely from here, the carbon deposition problem should be overcome to achieve higher activity up to the industrial mark. Upon 1 wt % Gd promotional addition, a cubic zirconium-gadolinium oxide phase is built. This means now that zirconia is stabilized by both tetragonal zirconium-yttrium oxide and cubic zirconium-gadolinium oxide phases. The crystallinity of the catalyst is decreased, along with the amount of weakly and moderately interacted NiO species growing on the surface at the expense of strongly interacted NiO species. On increasing Gd incorporation up to 4 wt %, the crystalline size of NiO is decreased to the lowest value of 7.5 nm, the crystallinity of catalyst is decreased continuously, the band gap between the conduction and valence bands is decreased to the lowest magnitude of 2.23 eV (similar to 5Ni/Y + Zr), and the weakly and moderately interacted reducible NiO species are mounted over the catalyst surface. It was reported that controlling the Ni particle ensemble size <9 nm restricts the thermal sintering of Ni particles, enhances the metal−support interaction, and alleviates the carbon deposition. 53 Here also, the 5Ni4Gd/Y + Zr catalyst has a minimum size of Ni species compared to another Gd-promoted catalyst, and so the minimum coke deposit is expected over it during the DRM reaction. Overall, the 5Ni4Gd/Y + Zr catalyst showed about 78% H 2 yield constantly up to 420 min with minimum coke deposit (6% weight loss in TGA). The CO yield remains just 4−5% higher than the H 2 yield over the 5Ni4Gd/Y + Zr catalyst. It indicates that RWGS is greatly retarded over 5Ni4Gd/Y + Zr. The particle size of 5Ni4Gd/Y + Zr is grown from 5.21 to 6.79 nm after the DRM reaction after 7 h. The 5Ni4Gd/Y + Zr catalyst is tested for longer TOS (24 h) where it again shows a constant 80% H 2 yield with 12.5 wt % mass loss. The deposited carbon above the catalyst surface is βcarbon species having a ratio of both defect carbon and graphite carbon. It is noticeable that upon 5 wt % Gd incorporation, the total number of reducible NiO species increases but the NiO crystallite size and band gap between the valence band and conduction band increase. It results into a relative drop of catalytic activity compared to the 5Ni4Gd/Y + Zr catalyst.

CONCLUSIONS
Over the 5NixGd/Y + Zr (x = 0−5) system, most of the Ni (catalytic active sites) remained present even under oxidizing CO 2 stream. The inferior catalytic activity (43% H 2 yield) of the zirconia-supported Ni catalyst is due to an unstable monoclinic zirconia phase and huge carbon deposition. The use of yttria along with zirconia support stabilizes zirconia by nurturing a stable tetragonal zirconia-yttrium oxide (mixed oxide) phase and shooting up the catalytic activity up to 70% H 2 yield even against the prominent carbon deposition. Gadolinium promotional addition up to 4 wt % modifies the surface by formation of a cubic zirconium-gadolinium oxide phase, limits the size of NiO up to the lowest value of 7.5 nm, and makes reducible NiO-moderately interacted species available over the catalyst surface and resists the coke deposition. The 5Ni4Gd/Y + Zr catalyst showed about ∼80% constantly up to 24 h at 800°C. The relative rise of the NiO crystallite size and the band gap of the 5Ni5Gd/Y + Zr catalyst (than 5Ni4Gd/Y + Zr) resulted in a relative drop of catalytic activity over 5Ni5Gd/Y + Zr (than 5Ni4Gd/Y + Zr).
■ ASSOCIATED CONTENT