Barium-Promoted Yttria-Zirconia-Supported Ni Catalyst for Hydrogen Production via the Dry Reforming of Methane: Role of Barium in the Phase Stabilization of Cubic ZrO2.

Developing cost-effective nonprecious active metal-based catalysts for syngas (H2/CO) production via the dry reforming of methane (DRM) for industrial applications has remained a challenge. Herein, we utilized a facile and scalable mechanochemical method to develop Ba-promoted (1-5 wt %) zirconia and yttria-zirconia-supported Ni-based DRM catalysts. BET surface area and porosity measurements, infrared, ultraviolet-visible, and Raman spectroscopy, transmission electron microscopy, and temperature-programmed cyclic (reduction-oxidation-reduction) experiments were performed to characterize and elucidate the catalytic performance of the synthesized materials. Among different catalysts tested, the inferior catalytic performance of 5Ni/Zr was attributed to the unstable monoclinic ZrO2 support and weakly interacting NiO species whereas the 5Ni/YZr system performed better because of the stable cubic ZrO2 phase and stronger metal-support interaction. It is established that the addition of Ba to the catalysts improves the oxygen-endowing capacity and stabilization of the cubic ZrO2 and BaZrO3 phases. Among the Ba-promoted catalysts, owing to the optimal active metal particle size and excess ionic CO3 2- species, the 5Ni4Ba/YZr catalyst demonstrated a high, stable H2 yield (i.e., 79% with a 0.94 H2/CO ratio) for up to 7 h of time on stream. The 5Ni4Ba/YZr catalyst had the highest H2 formation rate, 1.14 mol g-1 h-1 and lowest apparent activation energy, 20.07 kJ/mol, among all zirconia-supported Ni catalyst systems.

Introduction

The Paris Agreement set a goal for this century of keeping global warming below 2 °C and preferably at 1.5 °C. Aside from reducing anthropogenic greenhouse gas emissions (e.g., CO2 and CH4), turning these gases into value-added chemical feedstock is a more enticing way to accomplish this aim. In this context, the dry reforming of methane (DRM) is a potential and viable option because it yields hydrogen from the conversion of two major greenhouse gases (i.e., CH4 and CO2). Catalysts based on noble metals have been reported to be effective for DRM.[1−5] The total methane dissociation energy among the transition metals was found to follow the order Ni < Pd = Pt, so the experimental order of methane conversion was observed to be Ni > Pd = Pt.[6] Among Ni and Co, Gallego et al. found that the electronic configuration of Ni in Ni–CH4 is s0.54d9.42 (with respect to the d8s2 electronic configuration of metallic Ni), indicating smaller steric repulsion between a closed shell of Ni and CH4.[7] However, the electronic configuration of Co remains the same (either in Co–CH4 or in metallic Co), causing a large repulsion between a closed shell of Co and CH4. Importantly, the interaction energy of CH4 with Ni is 18 kcal/mol, and with Co it is 0.7 kcal/mol. As a result, from a catalytic activity standpoint, Ni-based catalysts are more appealing for industrial applications than Co catalysts. However, high-temperature Ni sintering, which induces pronounced coke deposition and, eventually, catalyst deactivation, is a major challenge.

To stabilize the Ni, it has been dispersed on several metal oxides including Al2O3,[8] SiO2,[9] zeolites,[10] ZrO2,[11] TiO2,[12] and MgO.[13] Furthermore, the addition of a promoter over supported Ni catalysts had brought about major physiochemical changes over the catalyst surface in favor of DRM. In brief, Mg incorporation added alkalinity to the catalyst system,[2,14−17] Sr boosted Lewis basicity,[18] Yb brought about a high edge of reducibilty,[19] Sc induced basicity and a metal–support interaction,[20] W stabilized the NiO phase and modified the redox behavior,[8,21,22] Ce or Y advanced lattice ion mobility together with reducibility,[23−39] and B or La induced carbon gasification (through B–OH species and La2O2CO3 formation, respectively).[40−46] Likewise, the addition of Sm, Gd, or Mn–Al (equal proportions) optimized the Ni size and enhanced the metal–support interaction.[47−49] Among the various supports, ZrO2 has the advantage of being able to withstand harsh thermal conditions while also supplying mobile oxygen species that can facilitate the oxidation of CH4-derived coke deposits.[50,51] In the case of supported Ni/ZrO2 catalysts, after reductive treatment, metallic Ni sites (i.e., Ni0) and oxygen vacancies are formed. The reaction scheme over the Ni-supported catalyst system is shown in Figure . Generally, C–H cleavage occurs at Ni0 sites, whereas CO2 dissociation occurs preferably at oxygen vacancies. Because Ni has a strong interaction with CH4,[7] CH4 is decomposed over Ni0 into CH(4– and x(1/2)H2 (where x = 1, 2, 3, 4). CO2, on the other hand, is adsorbed over basic surface sites and dissociates into CO and atomic oxygen/adsorbed oxygen at the Ni-support interface/boundary as well as on oxygen vacancies. Subsequently, the adsorbed oxygen oxidizes the formed CH(4– species into CO and (4–x)(1/2)H2.[52] At the same time, the carbon deposit is oxidized by lattice oxygen from the support, leaving an oxygen vacancy behind. Following that, the oxygen vacancy is replenished by CO2. This emphasizes the significance of adsorbed oxygen, which is directly involved in the oxidation of CH(4– species.

Figure 1

Reaction scheme over a Ni-supported catalyst system. Ovac is atomic oxygen that is formed after the dissociation of CO2 at an oxygen vacancy. □ is an oxygen vacancy.

According to the literature, the conventionally synthesized Ni-impregnated ZrO2 catalyst exhibited good DRM activity initially. However, because of a lack of optimal metal–support interactions, it underwent a high degree of graphitization and continuous catalyst deactivation.[53,54] On the other hand, sol–gel-derived catalysts demonstrated a strong metal–support interaction feature but required a costly synthetic procedure and produced fewer exposed active metal sites. However, by applying high Ni loading (10 wt %) and high-volume expansive carrier gas argon (8 times the feed gas volume), good catalytic activity was noticed.[55]

The ultimate challenges needed for a DRM reaction are the inhibition of carbon deposition on the catalyst surface and active metal sintering. Numerous approaches have been used by researchers to improve stability and to circumvent coke formation over Ni-based catalysts. It has been proposed that basicity improves the catalyst’s ability to adsorb CO2, facilitating coke gasification via the reverse Boudouard reaction (i.e., 2CO ↔ C + CO2). The addition of alkali or alkaline earth metals as promoters can improve the basic properties of the catalysts. These promoters may also enhance other features such as active metal dispersion and the metal–support interaction. For instance, 0.6% Na addition to a ZrO2-supported Ni catalyst was found to increase the metal–support interaction by the formation of NiOH species and to inhibit the hydrogenation of carbon deposits.[55] Similarly, adding Ca to a Ni/ZrO2 catalyst improves its basicity and textural properties, which in turn helps to avoid carbon deposition.[53,54] In the 13CH4 isotope experiment, it was found that the ratio of “13CO derived from CH4” and “CO derived from CO2” is 5/10 over a lanthana-zirconia-supported Ni catalyst, however, when Ca was used as a promoter, this ratio increased to 8/10.[56] That means that without Ca the majority of the CO was derived from CO2, but upon Ca promotion, more interaction of CO2 with the carbon impurity has taken place and more CO is generated by CH4.

Among other basic promoters, barium (Ba) has previously been used to improve the thermal and catalytic properties of Ni-based materials. For instance, when Ni was deposited by chemical vapor deposition over the BaO-ZrO2 support, it demonstrated the self-decoking ability of a carbon deposit by −OH and −O species with a negligible sign of Ni sintering for up to 50 h TOS at 700 °C.[57] Similarly, when barium is added to alumina, barium hexa-aluminate is formed, which has excellent thermal stability.[58,59] You et al. demonstrated that the addition of Ba to γ-Al2O3 can significantly neutralize the acidity of alumina.[60] Gomes et al. found that by substituting La with Ba in LaNiO3 perovskite, resistance against deactivation had been improved.[61] BaO (4 wt %) addition over SiO2-supported Ni enhanced the CO2 methanation activity.[62] The BaO/Ni interface is known for the water-mediated oxidation of carbon deposits to CO.[63] Ersolmaz demonstrated that BaCO3 is effective at oxidizing carbon through the formation of complexes between BaCO3 and C, which can be decomposed to CO2 at higher temperatures.[64] A barium zirconate-supported Ni catalyst has been utilized for a dry reforming reaction by Seo et al.[57] When BaO is combined with the ZrO2–Y2O3 support, it forms the BaZr0.9Y0.1O3−δ mixed oxide, which has a higher basicity than ZrO2–Y2O3.[65] BaZr0.9Y0.1O3−δ also exhibits a high proton conductivity, which may accelerate H-abstraction from the methyl group on the surface.[66] Promotional loading of BaCO3 over the ZrO2–Y2O3-supported Ni catalyst was well utilized in solid-oxide fuel cells for the direct utilization of methane[67] and the electrolysis of H2O to H2 and CO2 to CO.[68] On the basis of these findings, we anticipate that BaO-promoted ZrO2–Y2O3-supported Ni materials will have well-dispersed catalytically active sites (Ni0), improved basic properties, high proton conductivity, and coke resistance in favor of DRM.

Among the various synthesis methods, mechanochemical synthesis has received a considerable amount of attention because of its simplicity and use of cheap precursors as well as the possibility of realizing phases with different properties. Herein, we have systematically developed Ba-promoted (1–5 wt %) yttria–zirconia-supported Ni-based catalysts (5NixBa/YZr; x = 1–5 wt %). These materials were tested for DRM and characterized by surface area porosity measurements, infrared, ultraviolet–visible and Raman spectroscopy, and transmission electron microscopy. We demonstrated that adding a Ba promoter to the Ni/YZr catalyst inhibits carbon formation. During the DRM reaction, catalyst surfaces are exposed to reducing gas (H2) as well as oxidizing gas (CO2). The reduction–oxidation–reduction cycles over catalyst surfaces are regulated during the entire DRM reaction. To establish the function–activity correlations, we performed cyclic H2TPR-CO2TPR-H2TPR experiments in this study. These findings will contribute to advancing the knowledge spectrum of surface science toward DRM.

Experiment

Materials

Nickel nitrate hexahydrate (98%, Alfa Aesar), zirconia (gifted by Kagaku Daiichi Kogyo Co. Ltd Osaka), yttria (obtained from China), and deionized water were used.

Catalyst Preparation

The catalysts were prepared by the mechanochemical mixing of Ni(NO3)2·6H2O (equivalent to 5 wt % Ni loading), Ba(NO3)2 (equivalent to 0, 1, 2, 3, 4, 5 wt % BaO loadings), and a mesoporous yttria-stabilized zirconia (8 wt % yttria, 92 wt % zirconia) support, followed by drying and calcination at 600 °C for 3 h. For convenience, the prepared catalysts are abbreviated as 5NixBa/YZr, where Ni loading is fixed at 5 wt % and the Ba loading “x” varies from 0 to 5 wt % (i.e., x = 0, 1, 2, 3, 4, 5).

Catalyst Characterization

The catalysts that were synthesized were characterized using the Brunauer–Emmett–Teller (BET) surface area, X-ray diffraction (XRD), Raman spectroscopy, Fourier transform infrared spectroscopy (FTIR), ultraviolet–visible spectroscopy (UV–vis), transmission electron microscopy (TEM), H2 temperature-programmed reduction (H2-TPR), CO2 temperature-programmed desorption (CO2-TPD) and thermogravimetric analysis (TGA). Detailed descriptions of the instruments and characterization procedures are provided in Supporting Information S1.

Catalyst Activity Test

The DRM experiments were carried out in a tubular stainless-steel reactor at a space velocity of 42 000 mL/h gcat by passing a 30:30:10 mL/min volume ratio of a CH4/CO2/N2 gas feed through 0.1 g of prereduced catalyst. All of the catalysts were prereduced under H2 flow for 1 h at a flow rate of 30 mL/min at 600 °C. The DRM reaction was performed at 1 atm and 800 °C. The effluent was examined with an online GC equipped with molecular sieves 5A, Porapak Q columns, and a TCD detector using Ar carrier gas. The H2 yield % was estimated with the following expression:

Results

Characterization Results

A N2 adsorption isotherm and the porosity distribution, BET surface area, pore volume, and pore diameter results of 5Ni/Zr and 5NixBa/YZr (x = 0–5) catalysts are depicted in Figure , Figure S1, and Table S1. All materials have typical type IV isotherms with an H1 hysteresis loop indicating the presence of cylindrical mesopores. The dV/d log W vs W plot (where V is volume and W is the pore width) shows a rapid view of micropore, mesopore, and macropore distributions over the catalyst surface. The obtained results show that our catalysts have a bimodal pore size distribution. The marked change appeared in the lower pore width range of 10–50 nm and the intermediate pore width range of 100–150 nm, where the intensity of the earlier one was higher than that of the later.[69] The average pore size over Ni/YZr catalyst is 17.88 nm (Table S1). Interestingly, when yttria is incorporated, the pore size of the respective catalyst is increased to 24.78 nm. It is worth noting that the yttria–zirconia-supported Ni catalyst system has ∼50% less surface area but an ∼40% larger average pore size than the zirconia-supported Ni catalyst. However, no substantial structural changes in terms of pore volume and pore width are observed upon the incorporation of Ba into the Ni/YZr catalyst. In a Ba-promoted catalyst system, the pore size was typically in the 24–27 nm range (Table S1).

Figure 2

N2 adsorption isotherm and porosity distribution profiles of (A) 5Ni/Zr, (B) 5Ni/YZr, (C) 5Ni1Ba/YZr, (D) 5Ni2Ba/YZr, (E) 5Ni3Ba/YZr, (F) 5Ni4Ba/YZr, and (G) 5Ni5Ba/YZr.

The X-ray diffraction pattern of different catalyst samples and the NiO and BaZrO3 crystallite sizes are shown in Figure and Table S2. The zirconia-supported Ni catalyst (Ni/Zr) has a monoclinic zirconia phase (at 2θ = 23.93, 28.18, 31.48, 34.78, 38.58, 40.85, 44.70, 49.37, 50.16, 54.04, 55.34, 58.14, 59.98, 61.68, 62.92, 65.60, 69.13, 71.20, 75.39, 78.97, and 83.56°; JCPDS reference no. 00-007-0343) and a cubic NiO phase (at 2θ = 37.12, 43.24, 62.92, 75.39, and 78.97°; JCPDS reference no. 00-004-0835). The presence of the monoclinic ZrO2 phase in the 5Ni/Zr catalyst is also verified by Raman spectra (Figure A). The Raman bands related to the monoclinic ZrO2 phase appeared at 179, 335, 379, 476, and 610 cm–1.[70,71] Interestingly, over the yttria–zirconia-supported Ni catalyst were found more intense peaks of the cubic ZrO2 phase (2θ = 30.08, 34.97, 50.10, 59.64, 62.59, 74.07, 75.39, 81.67, and 84.31°; JCPDS reference no. 00-003-0640) than of monoclinic ZrO2. This indicates that yttria stabilizes the cubic phase of ZrO2. The crystallite size of cubic NiO is increased to 38.8 nm in 5Ni/YZr (against 18.4 nm in the 5Ni/Zr catalyst) (Table S2). The 1 wt % barium-promoted yttria–zirconia-supported Ni catalyst mainly contains the cubic ZrO2 phase. However, cubic BaZrO3 Bragg reflections at 2θ = 30.10, 43.26, and 62.67° (JCPDS reference no. 00-006-0399) were also evident in this catalyst (Figure B). The presence of Ba–O in the structure is also verified by the Raman spectra of the 5Ni1Ba/YZr catalyst. The Raman band at around 220–280 cm–1 over 5Ni1Ba/YZr is due to the overtones of TA, TA + TO, and TO of Ba–O vibrational modes[72] (Figure A). It can be said that Ba incorporation stabilizes the cubic phase of ZrO2 pronouncedly. On increasing the Ba loading up to 4 wt %, the minimum sizes of NiO (18.7 nm) and BaZrO3 (28.4 nm) crystallites were found. At 5 wt % Ba loading, selected planes (111, 200, 220, and 311) of cubic ZrO2 peaks are shifted to a higher Bragg’s angle, indicating a decrease in interplanar spacing (Figure C–F).

Figure 3

X-ray diffraction (XRD) profile of different catalyst samples (A) 5Ni/Zr and 5Ni/YZr. (B) Comparative XRD profiles of 5NixBa/YZr (x = 0, 1, 2, 3, 4, and 5 wt %). (C–F) Peak shifts of the 5Ni5Ba/YZr catalyst around the (111), (200), (220), and (311) planes, respectively, as compared to other barium-promoted catalysts.

Figure 4

(A) Raman spectra. (B) IR spectra. (C) UV–vis spectra. (D) Band gaps of different catalyst samples.

IR spectra, UV–vis spectra, and corresponding band-gap energy profiles of 5Ni/Zr and 5NixBa/YZr (x = 0, 1, 2, 3, 4, and 5 wt %) catalyst systems are shown in Figure B–D, respectively. IR peaks due to the bending and stretching vibration of O–H are present at 1630 and 3444 cm–1, respectively, in all catalysts.[22,73] The zirconia-supported Ni catalyst has vibrational peaks of Zr–O at 497 and 750 cm–1,[22] a broad peak of the bidentate format at 1355 cm–1,[74] and a unidentate carbonate peak at 1380 cm–1.[37,73] Interestingly, in the yttria–zirconia-supported Ni catalyst, the vibrational peaks of Zr–O and bidentate format peaks disappeared, indicating that the addition of yttria brought about major changes in the bonding pattern. At higher Ba loadings (4 to 5 wt %), the stretching vibrations of CO3–2 are observed at 851 and 1460 cm–1.[75] However, at 2 wt % Ba, the symmetric stretching vibrational peak of CO3–2 (C2 or C symmetry) at 1084 cm–1[76] and the stretching vibration of C=O at 1712 cm–1 are also noticed.[77]

The zirconia-supported Ni catalyst had O2– (2p, valence band) to M (4d, conduction band) charge-transition bands at 229 and 290 nm in the UV–vis spectra.[22] In comparing the UV spectra of Ni/Zr to those of ZrO2, it is found that the peak intensity at 229 nm remains the same but the peak intensity at about 290 nm is increased as well as broadened upon Ni anchoring over ZrO2. (Figure S2). This indicates that the peak at 229 nm is due to the charge-transfer band from O2– to Zr4+ and that the peak at 290 is due to the charge-transfer band from O2– to Ni2+ as well as that from O2– to Zr4+. On the other hand, for the yttria–zirconia-supported catalyst, the peak at 229 nm disappears, which indicates that yttria incorporation into the support changes the coordination environment of Zr4+ exclusively. For most of these catalysts, the d–d transition bands at 378 and 418 nm for the d–d transition from the 3A2g(F) energy state to the 3T1g(P) energy state of Ni2+ (in the octahedral environment) and at 718 nm for the d–d transition from the 3A2g(F) energy state to the 3T1g(F) energy state of Ni2+ (in an octahedral environment) are found.[73] Indeed, these findings confirm the octahedral environment of Ni2+ in the 5NixBa/YZr (x = 0–5 wt %) catalyst system. Interestingly, at a 4 wt % Ba loading, the charge transition from the O2– (2p, valence band) to the Zr4+/Ni2+ peak has the highest intensity but the d–d transition band for the 3A2g(F) energy state to the 3T1g(P) energy state of the Ni2+ octahedral environment disappeared. However, in this catalyst, the band gap was not affected by the addition of the Ba promoter (Figure D).

Figure depicts TEM images of fresh and spent catalysts as well as their particle size distributions. Mean NiO particle sizes of 3.25, 3.75, and 3.91 nm are observed for 5Ni/Zr, 5Ni/YZr, and 5Ni4Ba/YZr catalysts, respectively. After the reaction, the particle sizes (Ni species) are grown to 7.16, 7.60, and 7.64 nm, respectively. For spent catalysts, the formation of carbon nanotubes is easily visible.

Figure 5

TEM micrographs and Ni particle size distributions of different catalyst samples. (A and a) Fresh 5Ni/Zr, (B and b) fresh 5Ni/YZr, (C and c) fresh 5Ni4Ba/YZr, (D and d) spent 5Ni/Zr, (E and e) spent 5Ni/YZr, and (F and f) spent 5Ni4Ba/YZr.

H2-TPR, H2-TPR followed by CO2-TPD, and H2TPR-CO2TPD-H2TPR cyclic profiles of different catalyst samples are shown in Figure and Figure S3. The H2-TPR profile of the zirconia-supported Ni catalyst (Ni/ZrO2) shows a small reduction peak shoulder at about 235 °C for the reduction of free NiO, a sharp reduction peak at 335 °C for the reduction of NiO weakly interacting with the support, and a relatively smaller but broader peak at 490 °C for the reduction of NiO moderately interacting with the support (Figure A). When yttria was combined with the ZrO2 support, the reduction peak for weakly interacting NiO species almost vanished, whereas the reduction peaks for NiO moderately interacting with the support remained. This indicates that the catalyst had a smaller quantity of reducible, weakly interacting NiO species and that the addition of yttria resulted in a stronger metal–support interaction. Importantly, the reduction profile of Ba-promoted catalysts is identical to that of 5Ni/YZr. The CO2-TPD experiments were performed in conjunction with H2-TPR over reduced 5Ni/Zr, 5Ni/YZr, and 5Ni5Ba/YZr catalysts in order to estimate the basic sites in these materials (Figure B). During the H2-TPR, the reducible metal oxides are reduced to the respective metals and the surface hydroxyls are converted to water. As a result, the reducible NiO and surface hydroxyl ions should be eliminated during H2-TPR treatment for these catalysts. The reduced 5Ni/Zr catalyst had a significant concentration of weak basic sites (surface hydroxyls) at low temperatures, moderate-strength basic sites (surface oxide ions) at intermediate temperatures, and strong basic sites (thermally stable surface carbonates) at high temperatures. This indicates that the surface anion that is present on 5Ni/Zr may not be reducible but basic. The yttria–zirconia-supported Ni catalyst had a good quantity of moderately interacting reducible NiO species, and during H2-TPR treatment, they must be reduced to metallic Ni (by removing oxygen). Thus, the CO2 TPD profile of the reduced 5Ni/YZr catalyst showed the absence of moderate strength basic sites. It also indicates the greater oxygen-endowing capacity of the YZr-supported Ni catalyst than the ZrO2-supported Ni catalyst. The 5Ni4Ba/YZr catalyst had a good quantity of reducible, moderately interacting NiO species. However, a substantial number of intermediate-strength basic sites are present over the reduced 5Ni4Ba/YZr catalyst. These results imply that the basic nature of BaO contributes to retaining the high surface basicity of the 5Ni4Ba/YZr catalyst.

Figure 6

(A) H2-TPR profile of 5Ni/Zr, 5Ni/YZr, and 5Ni4Ba/YZr. (B) CO2TPD profile after H2-TPR of 5Ni/Zr, 5Ni/YZr, and 5Ni4Ba/YZr. (C) H2TPR-CO2TPD-H2TPR cycle of 5Ni/Zr, 5Ni/YZr, and 5Ni4Ba/YZr. (D) TGA profile of different catalyst samples.

H2-TPR reduces NiO to metallic Ni while also creating oxygen vacancies in the underlying metal oxide support because of H2 spillover from adjacent metallic Ni species. In general, such oxygen vacancies could be refilled when CO2-TPD is combined with H2-TPR. In this CO2-TPD process, the reduced Ni is reoxidized to NiO. It is important to know the type of NiO that is regenerated following oxygen replenishment by CO2. This can be achieved by performing another H2-TPR over the H2TPR-CO2TPD treated catalyst (Figure C). The cyclic experiment (H2TPR-CO2TPD-H2TPR) will provide evidence of the CO2 replenishment capacity and the stability of Ni species across different catalyst systems. In the case of the zirconia-supported Ni catalyst, abundant reducible peaks for free NiO species and reducible moderately interacting NiO species are observed. Prominent reducible peaks for free NiO species indicate pronounced Ni sintering during the reduction–oxidation–reduction cycle, which may be the major cause of the inferior performance of the catalytic activity of the zirconia-supported Ni catalyst. In contrast, there is no free NiO reducible peak in the 5Ni/YZr catalyst, although there are peaks attributable to moderately and strongly interacting NiO species, which indicates that the yttria has improved the sintering resistance and strong metal–support interaction properties. Likewise, the broad reduction peak associated with strongly interacting NiO species is mainly observed in the case of the Ba-promoted catalyst.

The TGA profiles of the spent catalysts are shown in Figure D. The zirconia- and yttria–zirconia-supported Ni spent catalysts showed a significant weight loss due to the oxidation of surface carbon deposits. Notably, increasing the Ba loading reduces weight loss. It implies that the incorporation of Ba in the catalytic system decreases the extent of coke formation on the catalyst surface during the DRM reaction.

Catalytic Activity Results

The catalytic activity of 5Ni/YZr and 5NixBa/YZr (x = 1–5 wt %) catalyst systems for DRM in terms of hydrogen yield is shown in Figure . The H2 yield for a zirconia-supported catalyst (5Ni/Zr) is the lowest and is unstable with respect to the time on stream (TOS). It is 50% initially, which decreases to 45% within 420 min. In contrast, the yttria–zirconia-supported Ni catalyst (5Ni/YZr) has a higher stable H2 yield for up to 420 min on TOS. It remains nearly stable at around 71% for 420 min. This suggests that incorporating yttria into the ZrO2 support is beneficial to the catalyst system. Interestingly, when 1–5 wt % Bapromoter is added to a yttria–zirconia-supported Ni catalyst (5NixBa/YZr; x = 1–5 wt %), prominent changes in the H2 yield are observed. The H2 yield remains more or less at about 72.5, 73, and 77% (for up to 420 min) over 1, 2, and 3 wt % BaO-promoted catalysts, respectively. The H2 yield is highest (i.e., 78%) for a catalyst with 4 wt % Ba loading (5Ni4Ba/YZr), and it remains constant for up to 420 min on TOS. The 5Ni4Ba/YZr catalyst also maintains the highest H2/CO ratio (i.e., 0.94) throughout the TOS (Figure B). The H2/CO and H2 yields increase as the reaction temperature increases from 500 to 800 °C, confirming the endothermic nature of the DRM reaction. Upon 5 wt % Ba, the H2 yield drops sharply to 70% (even less than for the 1 wt % Ba-promoted sample) and decreases further to 65% within 420 min on TOS. Excess BaO may cover the available catalytic active sites at a high Ba loading (5 wt %), resulting in inferior performance compared to that of its counterparts. It appears that 4 wt % Ba is the optimal promotor loading for the yttria–zirconia-supported Ni catalyst to obtain the maximal H2 yield and high H2/CO. The hydrogen-formation rate over the 5Ni4Ba/YZr catalyst was found to be 1.14 (molH/gCat/h). The effect of temperature on the H2-formation rate was also investigated in the temperature range of 500–800 °C. The apparent activation energy of 20.07 kJ/K/mol was estimated for H2 formation over the 5Ni4Ba/YZr catalyst.

Figure 7

Catalytic activity results. (A) H2 yield of different catalysts at 800 °C. (B) H2/CO ratio of different catalysts at 800 °C. (C) H2 yield and H2/CO ratio of 5Ni4Ba/YZr at different reaction temperatures. (D) Influence of the reaction temperature on the H2 formation rate of 5Ni4Ba/YZr.

Discussion

The catalytic activity of the zirconia-supported Ni catalyst in DRM is due to Ni2+ in octahedral coordination, a large surface area, the pore volume, and the bidentate format/monodentate carbonate species. However, the presence of a reducible free/weakly interacting NiO species, a small pore size, and unstable monoclinic ZrO2 phases limits the activity. The H2TPR-CO2TPD-H2TPR cyclic experiment displayed a prominent quantity of a reducible free NiO species/weakly interacted NiO species. Weak metal–support interaction leads to Ni sintering at high temperatures, which causes a prominent carbon deposit. The 5Ni/Zr catalyst exhibited only 45% H2 yield on TOS, and the TGA results showed a huge weight loss due to coke removal from this catalyst.

On the other hand, the yttria–zirconia-supported Ni catalyst had a small surface area but a relatively larger pore size. Furthermore, it mainly featured a cubic ZrO2 phase and reducible Ni2+ species that were in octahedral coordination and strongly interacted with the support. This strong metal–support interaction, together with cubic ZrO2 phase stabilization, resulted in a 71% H2 yield over the 5Ni/YZr catalyst. The H2-TPR followed by the CO2-TPD experiment showed that the 5Ni/YZr catalyst has a higher oxygen-endowing capacity than the Ni/Zr catalyst. Although the 5Ni/YZr catalyst experienced similar weight loss due to carbon removal as the 5Ni/Zr catalyst, it demonstrated stable catalytic performance (71% H2 yield) for up to 420 min, which indicates that the type of carbon deposit is amorphous and oxidizable and so does not block the catalytic active sites.

Upon addition of Ba promoter, the cubic BaZrO3 phase additionally stabilizes the cubic ZrO2 phase. When the Ba loading increases, the TGA result shows less weight loss in the spent catalysts. This indicates a greater oxygen-endowing capacity of the catalyst upon increasing the Ba loading to oxidize carbon deposits during the DRM. Among all Ba-promoted samples, 5Ni4Ba/YZr possesses the smallest NiO crystals (18.7 nm), excess CO32– ionic species, a high-intensity charge-transition band from O2– (2p, valence band) to Zr4+/Ni2+, and an optimal metal–support interaction. H2TPR followed by the CO2 TPD experiment showed that the reduced 5Ni4Ba/YZr sample has a relatively more basic site concentration than the 5Ni/YZr catalyst. The H2TPR-CO2TPD-H2TPR cyclic experiment shows the presence of only reducible NiO species strongly interacting with the support. It can be said that the metallic Ni species anchored on the cubic zirconia support facilitates CH4 decomposition and that the resultant hydrocarbon intermediates are then oxidized by oxygen-containing surface species (e.g., ionic CO32–) or lattice oxygen. It conveys the highest H2 yield of 78% constantly up to 420 min on TOS. The 4 wt % Ba is the optimum loading for the highest H2 yield. When the Ba loading is increased to 5 wt %, the lattice planes are compacted and the NiO crystallite size increases. The excess Ba covers the accessible catalytic active sites, which in turn lowers the catalytic activity and stability. The 5Ni5Ba/YZr catalyst shows an inferior H2 yield even below that of the 5Ni1Ba/YZr catalyst.

The catalytic activity of the above-discussed DRM catalysts and the other set of 54 DRM catalysts[78−132] in terms of the H2 yield, H2 formation rate, and CO formation is shown in Table S3. The calculation details for hydrogen and CO formation rates are described in Supporting Information S2. Among the different catalysts synthesized in this study, the 5Ni4Ba/YZr catalyst showed the highest hydrogen formation rate (1.14 mol g–1 h–1). On the basis of the results in Table S3, it seems that the Ba promoter is a better choice than Ga,[94] Mn,[95] Al,[49] Al–Mn,[49] Pr,[44] Sm,[37] and Nd[44] promoters in terms of achieving a high hydrogen formation rate. At a ≤5 wt % Ni loading, a >0.9 H2/CO ratio, and a ≤0.1 g catalyst weight, Sc-,[20] La-,[41] Gd-,[48] and Ce[109]-promoted ordered mesoporous silica-supported Ni catalyst systems demonstrated a higher rate of H2 formation than our catalyst system. Nevertheless, the additional cost of structure-directing agents and complex catalyst preparation procedures may limit the industrialization potential of these materials. Following silica, some zirconia-based Ni catalysts were found to be more competent than our catalyst system in terms of H2 production via DRM, such as the Cr-promoted lantana–zirconia-supported Ni catalyst[93] Ce-promoted lantana–zirconia[96] catalyst, and tungstate–zirconia[22]-supported Ni catalyst. They showed 1.18 molg–1h–1, 1.23 molg–1h–1, 1.14 molg–1h–1 H2 formation rate, respectively. Furthermore, we compared the apparent activation energy for H2 formation among closely related zirconia-supported Ni catalyst systems (Table ). Among 5 wt % Ni-loaded catalysts, the apparent activation energies of the 5Ni4Ba/YZr catalyst (this work) and phosphate–zirconia-supported catalyst (5Ni/8PZr)[86] were 20.07 and 6.48 kJ/mol, respectively. However, in terms of activity, the 5Ni/8PZr catalyst had a much lower rate of hydrogen formation than our 5Ni4Ba/YZr catalyst.[86]

Table 1

Comparison of Apparent Activation Energies for H2 Formation across Various Catalyst Systems

catalyst system	Ni wt %	reaction temp (°C)	RH2	slope	apparent activation energy (kJ/mol)	ref
5Ni4Ba/YZr	5	500	0.21	–2.41	20.07	our work
	5	550	0.30
	5	600	0.48
	5	650	0.70
	5	700	0.90
	5	750	1.08
	5	800	1.13
Ni/Zr	5	500	0.11	–3.27	27.19	(43)
	5	550	0.29
	5	600	0.44
	5	650	0.56
	5	700	0.79
Ni-CeO2/ZrO2	5	500	0.13	–3.08	25.61	(43)
	5	550	0.26
	5	600	0.46
	5	650	0.57
	5	700	0.79
Ni-La2O3/ZrO2	5	500	0.22	–2.56	21.28	(43)
	5	550	0.35
	5	600	0.54
	5	650	0.69
	5	700	0.98
Ni-K2O/ZrO2	5	500	0.22	–2.62	21.78	(43)
	5	550	0.33
	5	600	0.56
	5	650	0.70
	5	700	0.98
Ni/ZrO2-P	10	600	0.48	–1.57	13.05	(133)
	10	650	0.57
	10	700	0.67
	10	750	0.80
	10	800	0.92
Ni/ZrO2-C	10	600	0.38	–2.16	17.96	(133)
	10	650	0.43
	10	700	0.55
	10	750	0.74
	10	800	0.91
5Ni/8PZr	5	600	0.05	–0.78	6.48	(86)
	5	650	0.71
	5	700	0.07
	5	750	0.10
	5	800	0.18
10Ni/8PZr	10	500	0.01	–4.28	35.58	(86)
	10	550	0.05
	10	600	0.12
	10	650	0.11
	10	700	0.24
	10	750	0.28
	10	800	0.37
15Ni/8PZr	15	500	0.07	–2.20	18.29	(86)
	15	550	0.13
	15	600	0.14
	15	650	0.25
	15	700	0.29
	15	750	0.32
	15	800	0.40
20Ni/8PZr	20	500	0.07	–2.23	18.54	(86)
	20	550	0.13
	20	600	0.14
	20	650	0.25
	20	700	0.29
	20	750	0.32
	20	800	0.42
Ni-CaO-ZrO2	13.76	600	5.85	–0.17	1.41	(91)
	13.76	650	5.87
	13.76	700	5.81
	13.76	750	5.48
	13.76	800	5.25
	13.76	850	5.28
	13.76	900	5.44
	13.76	950	5.34
	13.76	1000	5.23
	13.76	1050	5.23
	13.76	1100	5.22
	13.76	1150	5.13
	13.76	1200	5.13
Ni/Ce50-Zr50		550	0.01	–4.64	38.58	(95)
		600	0.01
		650	0.03
		700	0.05
		750	0.08
		800	0.11
		850	0.13
Ni-Mn/Ce50-Zr50		550	0.04	–1.67	13.88	(95)
		600	0.06
		650	0.08
		700	0.10
		750	0.11
		800	0.12
		850	0.12
Ni/MgO-ZrO2	10	850	0.47	–4.46	37.08	(88)
	10	900	0.71
	10	950	0.82
	10	1000	1.06
Ni-0.5K/MgO-ZrO2	10	850	0.61	–3.49	29.01	(88)
	10	900	0.80
	10	950	1.11
	10	1000	1.08
Ni-0.9K/MgO-ZrO2	10	850	0.32	–4.02	33.42	(88)
	10	900	0.52
	10	950	0.53
	10	1000	0.71
Ni-1.4K/MgO-ZrO2	10	850	0.34	–3.44	28.60	(88)
	10	900	0.52
	10	950	0.53
	10	1000	0.65
Ni-1.9K/MgO-ZrO2	10	850	0.21	–6.12	50.88	(88)
	10	900	0.28
	10	950	0.40
	10	1000	0.61

Conclusions

Yttria–zirconia-supported Ni-based catalysts and 1–5 wt % Ba-promoted yttria–zirconia-supported Ni-based catalysts are characterized and tested in the dry reforming of methane. The 5Ni/Zr catalyst shows low catalytic activity of a 45% H2 yield due to an unstable monoclinic ZrO2 support and the presence of free/weakly interacting reducible NiO species. The high activity (71%) of the 5Ni/YZr catalyst is correlated with larger exposed pores and a stronger metal–support interaction through the thermally stable cubic ZrO2 phase and the presence of moderately interacting reducible NiO species. Upon increasing the barium loading, the oxygen capacity increases and the carbon deposition decreases. The addition of 4 wt % barium brings about the BaZrO3 cubic phase, cubic ZrO2 phase, optimum NiO crystallite size (18.7 nm), excess ionic CO32– species, improved basicity, and high intensity of the charge-transfer band. Reduction–oxidation–reduction treatment showed only reducible, strongly interacting NiO species over the catalyst surface. A 79% H2 yield and a 0.94 H2/CO ratio are achieved for up to 420 min over the 5Ni4Ba/YZr catalyst. Among different zirconia-supported Ni catalysts, the 5Ni4Ba/YZr catalyst had the highest H2 formation rate (1.14 mol g–1 h–1) and the minimum apparent activation energy of hydrogen formation (20.07 kJ/mol). For 5 wt % Ba-promoted catalysts, the excess Ba covers the accessible Ni active sites and reduces the catalytic activity and stability.