Ultra-durable Ni-Ir/MgAl2O4 catalysts for dry reforming of methane enabled by dynamic balance between carbon deposition and elimination

Abstract

Various methods have been attempted to suppress the carbon deposits in DRM.By using supports rich in oxygen vacancies such as CeO2 [22][23][24] or alkaline oxide supports like MgO [25][26] , CO2 activation can be strengthened to accelerate the elimination of carbon deposits.This enhances the redox property of the catalyst and inhibits the development of carbon deposition [27,28] .The accumulation and sintering of metallic Ni at high temperatures (700-1000 °C) can result in the formation of larger Ni particles, which promotes the carbon deposition [29] .The strong metal-support interaction (SMSI) between metallic Ni and the hydroxyapatite [Ca10(PO4)6(OH)2] (HAP) support [30] was found to enhance the dispersion of Ni particles and alleviate sintering and carbon deposits concurrently.In addition, it was reported that decorating the Ni surface with Co prevented the continuous formation of carbon nanotubes on the Ni surface [31] .The redox recycling on Ni-Fe/MgxAlyOz catalysts, where metallic Fe was oxidized by CO2 to form FeOx and the adsorbed C atom obtained from the dissociation of CH4 reduced FeOx back to Fe, can effectively eliminate carbon deposits [32] .Ni-Mo nanoparticles were stabilized on the edge of a single MgO crystal to inhibit the particle sintering at high temperatures and effectively prevent the generation of carbon deposits over a long reaction period (850 h) [33] .
The core issue for achieving practically-durable DRM catalysts with minimal carbon deposition is to establish an efficient equilibrium between CH4 dissociation and CO2 activation on the catalyst surface.In this work, we show that a highly efficient and stable DRM system can be built on bimetallic Ni-Ir/MgAl2O4 catalysts, brought forth by a synergy between the CH4 dissociation on Ni sites and MgAl2O4-enhanced CO2 adsorption and activation on Ir sites that can scavenge surface carbon species generated from the CH4 dissociation step.Through in-situ spectroscopic characterization of the generation and elimination of carbon species and theoretical calculations on the CO2 activation, we propose an equilibrium mechanism of carbon generation and elimination for designing the effective and durable catalysts for the DRM process.

Structural characterization of supported Ni-Ir catalysts
MgAl2O4-supported Ni, Ni3Ir1, and Ir catalysts were synthesized using a conventional co-impregnation method, and the surface areas of these catalysts were 126.0, 133.9, and 136.1 m 2 /g, respectively, as determined by nitrogen physical adsorption (Fig. S1).X-ray diffraction (XRD) results (Fig. 1a) showed that the diffraction peaks of metallic Ni for Ni3Ir1/MgAl2O4 had a much lower intensity compared to the pure Ni catalyst.Similarly, compared with the pure Ir catalyst, the diffraction peaks of Ir had a lower intensity for Ni3Ir1/MgAl2O4.Such weakened diffraction peaks for the Ni-Ir catalyst implies that the interaction of Ni and Ir increased metal dispersion compared to the single-metal catalyst system.The Ni-Ir interaction is further evidenced by a shift of the signal peaks of metallic Ir in the presence of Ni (e.g., The structural information for the Ni3Ir1/MgAl2O4 catalyst was further confirmed by transmission electron microscopy (TEM) and energy dispersive spectroscopy (EDS) elemental mapping (Figs.1d-g).The metal nanoparticles were uniformly dispersed on the MgAl2O4 spinel support surface with a typical bimodal particle size distribution, for which the corresponding average particle diameters were 1.2 and 4.3 nm and the smaller particles were predominant.In contrast, the average diameters of metal particles for Ir/MgAl2O4 and Ni/MgAl2O4 were 1.1 and 5.7 nm, respectively (Fig. S4).It is clearly indicated that the formation of the Ni-Ir alloy improved the dispersion of Ni, consistent with the XRD results shown above (Fig. 1a).In addition, the EDS analysis (Fig. 1e) showed that the small particles in the Ni3Ir1/MgAl2O4 catalyst possessed a higher

Catalytic performances of Ni-Ir catalysts in DRM
From the view of thermodynamics, the DRM process is favorable at high temperatures due to its endothermic nature.At medium to low temperatures (below 700 °C), the impact of carbon deposits will be significant.The results for the Ni/MgAl2O4, Ni3Ir1/MgAl2O4, and Ir/MgAl2O4 catalysts are shown in Figs.2a and S6 (650 °C, GHSV = 40,000 mL•g −1 •h −1 , 1 bar).For the pure Ni catalyst, the initial conversion of CH4 was 59.4% and decreased to 47.3% after 100 hours of testing, in which the corresponding conversion of CO2 decreased from 71.5% to 61.6%.It is reflected that the Ni catalyst was not able to maintain a good stability during DRM, although its initial activity was high.For the Ir catalyst, the initial conversions of CH4 and CO2 were only 50.2% and 63.6%, respectively, but the catalyst activity did not decrease significantly during the 100-hour test.For the Ni3Ir1/MgAl2O4 catalysts, the conversions of CH4 and CO2 slightly decreased from 61.5% to 60.2% and 73.3% to 72.8% during the 100-hour test, respectively, indicating the Ni-Ir alloy catalyst possessed both of high activity and improved stability.As shown in Table S2, we summarize the typical catalyst systems for DRM, and the Ni-Ir/MgAl2O4 system developed in our study had superior conversion rates of CH4 and CO2 than the state-ofthe-art catalysts [9,[34][35][36][37][38][39][40][41] .
For the DRM test with different temperatures (600-800 °C, Fig. S7), the initial conversion of CH4 on the Ni, Ni3Ir1 and Ir catalysts increased with increasing temperature from 42.1%, 46.8%, and 20.9% at 600 °C to 89.8%, 90.4%, and 90.2% at 800 °C, respectively.Although the three catalysts showed significant differences in activity at the lower temperature, the gap decreased as the temperature increased until the activity was almost identical at 800 °C.The ratio of H2/CO in the syngas showed a similar phenomenon, increasing from 0.67, 0.80, and 0.60 at 600°C to 0.96, 0.97, and 0.94 at 800 °C.It is suggested that DRM is preferred over the reverse water-gas shift reaction (RWGS; CO2 + H2 → CO + H2O) at temperatures above 600 °C, which leads to an increase in the H2/CO ratio and gradually approaches unity as the temperature rises.
The different performances of the three catalysts in DRM may be directly related to their intrinsic abilities for the adsorption and activation of the CH4 and CO2 reactants.Kinetic studies (Figs.2e-f) showed that the measured reaction order of CO2 for Ni/MgAl2O4 was 0.12, while for Ni3Ir1/MgAl2O4 and Ir/MgAl2O4, the reaction order values were −0.13 and −0.57, respectively.These data indicate the adsorption strength of CO2 on the three catalysts decreased with an order of Ir/MgAl2O4, Ni3Ir1/MgAl2O4, and Ni/MgAl2O4.As shown below, DFT calculations unveil that the higher oxophilicity of Ir than Ni accounts for the stronger CO2 adsorption on Ir/MgAl2O4.
The measured reaction orders for CH4 over the Ni, Ni3Ir1, and Ir catalysts were −0.38, −0.50, and −1.14, respectively, indicating the dissociative adsorption of CH4 on Ir/MgAl2O4 was also the strongest among the three catalysts.This outstanding adsorption ability of Ir made it difficult for the reactive species to desorb from the catalyst surface, which limited the catalytic activity at medium to low temperatures.As the temperature increased, the desorption of the reactive species became easier and the catalytic activity of Ir/MgAl2O4 increased faster compared with the other two catalysts as evidenced by the changes of CH4 conversion with increasing temperature (Fig. S7).

Characterization of spent Ni-Ir catalysts
It is generally accepted that carbon deposition is one of the main causes of catalyst deactivation in DRM [16,18] .SEM images (Figs.2b-d) of the spent catalysts clearly showed the appearance of carbon deposits after the DRM reaction, which were interlaced and attached to the catalyst surface, blocking contact between the active metal site and the reactants and thus inhibiting the catalytic activity.Specifically, Ni/MgAl2O4 had more carbon deposits than Ni3Ir1/MgAl2O4, and the carbon chains on Ni/MgAl2O4 were longer and thicker than the latter, with average diameters of about 27.9 nm versus 21.5 nm.The carbon deposits also showed different morphologies for each catalyst, presumably due to the different amount of carbon species provided for the carbon chain growth for each catalyst.It is worth noting that the carbon deposition was nearly negligible on the Ir catalyst (Fig. 2d), consistent with the high stability of this catalyst for DRM.
XRD patterns of the spent Ni/MgAl2O4 catalyst showed an obvious graphite-2H signal (2θ = 25.9°), which is typical of amorphous carbon species produced in DRM (Fig. 2g).The intensity of the graphite-2H signal for Ni3Ir1 was lower than that for Ni, while this signal was not detectable on the Ir catalyst.Raman spectra of these spent catalysts (Fig. 2h) showed peaks at 1334 and 1598 cm −1 for both Ni and Ni3Ir1, which were assigned to the D-band and G-band signals of carbon species.Although the peak intensity of Ni was higher, the peak intensity ratios of the D-band and G-band signals were similar between the two catalysts, indicating no significant differences in the types of carbon deposits formed.For the spent Ir catalyst, no corresponding peaks were observed, consistent with the XRD characterization.To quantify the amount of carbon deposits formed on the spent catalysts, TGA analysis was conducted (Fig. 2i).No significant weight loss was observed for the Ir catalyst during the TGA process, while the Ni and Ni3Ir1 catalysts showed weight losses of 16.7% and 10.2%, respectively.These data unambiguously reflect that the addition of Ir suppressed the formation of carbon deposits in DRM.

In-situ DRIFTS study of DRM
Adsorption of CO2 on the MgAl2O4 support can form carbonate species, which may promote the fixation and activation of CO2 during the DRM.In-situ diffuse reflectance infrared fourier transform spectroscopy (DRIFTS) analysis showed that MgAl2O4 adsorbed CO2 and converted it into carbonate species readily at 650 °C in a CO2 atmosphere (Fig. S9a).TGA was further performed to assess the CO2 adsorption capacity of MgAl2O4 (Fig. S9b), which was approximately 1.0 wt% at 650 °C.As the temperature decreased, the adsorption capacity of MgAl2O4 continued to increase, and reached about 2.6 wt% at 50 °C.Although the adsorption capacity was not significant at high temperatures, the adsorption and desorption of CO2 occurred simultaneously during the DRM process and could dynamically provide CO2 for the active metal sites, beneficial for the capture and utilization of gaseous CO2.
To further ascertain the specific impact of carbonates on MgAl2O4 for DRM, insitu DRIFTS was employed to analyze the activation process through the step-by-step introduction of CO2-CH4-CO2 atmosphere.Within the first minute of introducing CO2, all the three catalysts showed vibrational peaks at 1643 and 1542 cm −1 (attributed to monodentate carbonate species [42][43] ), indicating all the catalysts can adsorb CO2 to form carbonate (Fig. S10).Subsequently, the IR reactor cell was purged with Ar, and CH4 was introduced in the second stage.The in-situ DRIFTS spectra for the three catalysts during the CH4 stage are shown in Figs.3a-c; the carbonate signals slowly decreased over time and those for Ni/MgAl2O4 decreased faster than the other two catalysts.It is worth noting that, at the beginning of the CH4 stage for Ni/MgAl2O4, the gaseous CO2 signal initially increased and then slowly decreased; however, no gaseous CO2 was detectable on the other two catalysts.In addition, all the catalysts were observed generating CO (twin peak at 2150 cm -1 ), and the generation rate of CO was relatively slow on the Ni/MgAl2O4 catalyst, with a noticeable CO signal observed only after 1 minute of introducing CH4, while the other catalysts showed obvious CO formation upon the initiation of the stage.
In the third stage, we again used Ar to purge the residual CH4 and then reintroduced CO2.The in-situ DRIFTS spectra of the catalysts during this stage again showed that carbonate species were formed on the catalysts (Figs.3d-f).This occurred more slowly on the Ni/MgAl2O4 catalyst, requiring 2.5 minutes to reach carbonate species saturation while on the other two catalysts, the saturation was reached after just 1 minute.Further, the signal peak for gaseous CO only appeared for the Ni3Ir1/MgAl2O4 catalyst.
Two distinct reaction pathways are proposed to explain the observed difference during the in-situ DRIFTS experiments described above.The Ni/MgAl2O4 catalyst followed reaction path A (Fig. 3g).Namely, when CO2 was first introduced, the MgAl2O4 support adsorbed gaseous CO2 and formed monodentate carbonate species (CO2 (g) → CO3 was reintroduced in the third stage, only the accumulation of carbonate species on the support occurred without the formation of CO.
In contrast to Ni/MgAl2O4, the Ir/MgAl2O4 catalyst followed reaction path B (Fig. 3h).The difference was that the Ir active sites could directly use and activate the carbonate species on the support, eliminating the CHx* and forming gaseous CO (CO3 2-, ads + CHx* → CO (g)).The carbonate species did not have to decompose into CO2 (g) to be reused, therefore there was a large amount of CO formed without obvious CO2 formation.The direct utilization of carbonate on the Ir based catalysts increased the utilization rate of carbonate species, resulting in a slower rate of consumption compared to route A. Compared with Ni, Ir had an insufficient ability to activate CH4, therefore the carbonate species that adsorbed on the surface of MgAl2O4 in the first stage was sufficient to eliminate the CHx* generated during the second stage, which resulted in no CO formation in the third stage.We also increased the time for CH4 introduction in the second stage of the Ir/MgAl2O4 catalyst to ensure that enough CHx* were generated.
After CO2 was introduced again, gaseous CO signals appeared indicating that the Ir/MgAl2O4 catalyst could reduce the carbon species and CO2 to CO (Fig. S11).
The Ni3Ir1/MgAl2O4 catalyst also followed reaction path B. Due to the presence of Ni, the dissociation ability of CH4 was better than the Ir/MgAl2O4 catalyst, therefore CHx* species were not completely eliminated and grew into carbon chains in the second stage.In the third stage, gaseous CO formed in addition to the carbonate but was slower because the carbonate was not saturated during the initial stage of CO2 introduction.
With saturation of the carbonate species, the formation rate of gaseous CO increased accordingly, which further confirmed that carbonate was used as a bridge to adsorb and convert CO2 rather than directly utilize gaseous CO2.Both the Ni3Ir1/MgAl2O4 and Ni/MgAl2O4 catalysts generated carbon species, which grew into carbon chains in the second stage.However, in the third stage, only the Ni3Ir1/MgAl2O4 catalyst following path B eliminated the carbon species completely and formed gaseous CO.This suggests that path B was more beneficial to the activation of CO2 and the elimination of carbon species.
To verify the unique roles of MgAl2O4 in the DRM, an Ir/Al2O3 catalyst was further characterized by the in-situ DRIFTS experiment (Fig. S12).When CO2 was introduced in the first stage, no carbonate species formed and only obvious gaseous CO2 signal peaks appeared.In the second stage, no gaseous CO signal peak appeared during the introduction of CH4, because the carbon species were eliminated in a CH4 atmosphere without carbonates formed on the support surface.The third stage was the same as the second stage, with no obvious gaseous CO signal peaks.Although carbon species were produced by the activation of CH4 (TPSR-MS, in-situ DRIFTS) in the second stage, Ir weakly utilized gaseous CO2 to eliminate carbon species, and only a small amount of carbon species which had not continued to grow was further activated, forming a small amount of CO.Therefore, only the Ir-CO signal (2000 cm −1 ) was observed on the Ir catalyst, and the amount of gaseous CO did not reach the detection limit.This proves that the carbonate formed on the MgAl2O4 support played a key role in the effective removal of carbon by Ir during the DRM.
In order to further explore the adsorption and activation ability of CO2 on different catalysts, the catalysts were characterized with temperature programmed desorption of CO2 (CO2-TPD) showed a signal peak appeared only at 614 °C on MgAl2O4, and the peak position of Ni/MgAl2O4 was similar to that of MgAl2O4, but with a lower intensity (Fig. S13).For Ni3Ir1/MgAl2O4, in addition to a peak appeared at 621 °C (similar to the support), there were peaks at 778 and 873 °C, and the peaks at 621 °C and 809 °C were present for Ir/MgAl2O4.This was because Ir had better dispersion and CO2 could be effectively adsorbed and utilized at the Ir-MgAl2O4 interface, while the Ni-MgAl2O4 interface has no effect on CO2 adsorption.This further supports the findings from the kinetics studies and in-situ DRIFTS, which Ir sites impacted adsorption and activation of CO2 and restricted carbon deposits during the DRM.

TPSR-MS analysis of the activation of methane and the elimination of carbon deposits
The activation and dissociation of CH4 is the main cause of carbon deposition Subsequently, TPSR-MS was performed on the catalysts after CH4 treatment to explore how the different catalytic systems eliminate carbon deposits in a CO2 atmosphere (Fig. S14b).Overall, the carbon removal temperature on Ni/MgAl2O4 was higher than that on Ni3Ir1/MgAl2O4 and Ir/MgAl2O4, which confirmed the Ni3Ir1/MgAl2O4 and Ir/MgAl2O4 catalysts had stronger carbon removal abilities.The Ni/MgAl2O4 catalyst had a relatively small amount of activated CH4 due to the blockage of active sites at lower temperatures, and the amount of carbon deposits and the resulting peak area were both relatively small.The above results show that Ni3Ir1/MgAl2O4 and Ir/MgAl2O4 which followed path B had advantages in the process of activating CO2 and removing carbon species.
Raman spectra of the catalysts treated with CH4 showed that all of them had Dband and G-band carbon peaks (Fig. S16) and the carbon content followed Ni-Ir > Ir > Ni, which was consistent with the TPSR-MS-CO2 results.In addition, the ratio of the D-band to the G-band indicated that the catalysts formed the same type of carbon species upon the dissociation of CH4.The different temperatures for CO2 removal reflect the difference in the activation mechanism of CO2 as described in the previous section.

Theoretical assessment of CO2 activation on supported metal catalyst
As shown above, the activation of CO2 over metal surface is critical for the efficient removal of carbon deposits during DRM.DFT calculations were further performed to understand the difference between Ir and Ni on CO2 activation at a molecular level.Herein, a Ni8 cluster, a Ir8 cluster, and two Ni6Ir2 clusters with distinct configurations (denoted as Ni6Ir2/MgAl2O4-a and Ni6Ir2/MgAl2O4-b) were constructed on a MgAl2O4 (100) surface for modeling the Ir/MgAl2O4, Ni/MgAl2O4 and Ni3Ir/MgAl2O4 catalysts, respectively (Fig. 4a).sidering that the elimination of carbon deposits during DRM mainly depends on the active O* species generated from CO2 dissociation [46][47][48][49][50] , we therefore focused on the adsorption and dissociation processes of CO2 at the metal-support interface.As shown in Fig. 4b, CO2 adsorption was much stronger on Ir8/MgAl2O4 than Ni8/MgAl2O4 (-2.14 eV vs -1.14 eV), consistent with the in-situ DRIFTS experiments.
For Ni6Ir2/MgAl2O4, the adsorption energy of CO2 was between those for the Ir8 and

Mechanism of carbon deposits-elimination balance in DRM and its application
With respect to the Ni3Ir1/MgAl2O4 catalyst, Ni mainly played the role of dissociating CH4, while MgAl2O4 not only acted as the support of the active metal sites, but also adsorbed CO2, forming carbonate species to enrich CO2.These carbonate species could be effectively activated by Ir to eliminate carbon species and inhibit carbon deposition.Since the main activation sites of CH4 and CO2 were different (Ni for CH4; Ir for CO2), the ratios of Ni and Ir atoms can greatly modulate the activation rates of CH4 and CO2.
During the DRM process, although the Ni-Ir/MgAl2O4 system had more carbon deposits, it maintained high stability.Interestingly, the formed carbon deposits seemed to have no effect on the catalytic activity.Therefore, carbon deposits formed on the catalysts along with the reaction were examined (Fig. 5a).We found that Ni/MgAl2O4 had a faster rate of carbon deposition at the beginning and steadily increased.Further, as the carbon deposits increased, carbon chains blocked the contact between the active metal site and CH4.The reduction in the amount of active sites for CH4 dissociation led to a decrease in the rate of carbon deposition and DRM activity.In contrast, the Ir/MgAl2O4 catalyst did not form carbon deposits during the 100-hour test, while carbon deposits increased within the first 20 hours for Ni3Ir1/MgAl2O4, which was then stabilized with no significant additional increase until 100 hours.
Accordingly, we proposed a carbon deposits-elimination balance mechanism (Fig. 5b).With excessive Ni, the activation rate of CH4 would be faster than that for CO2, and the dissociated carbon species would not be eliminated in time, leading to the generation of carbon deposits.This type of carbon deposition occurred on Ni, the site activated by CH4, which would block some of the Ni active sites.With a decrease in the Ni active sites, the activation rate of CH4 decreased, resulting in a balance in the activation rate of CH4 and CO2.Similarly, excessive carbon deposits led to higher activation rates for CO2 than CH4, leading to the elimination of carbon deposits and exposing the Ni active sites previously covered.This caused the activation rate of CH4 to increase until it matched that for CO2 activation.
In the long-term test, this system gradually approached equilibrium until ν(CH4) ≈ ν(CO2) (ν(CH4): the activation rate of CH4; ν(CO2): the activation rate of CO2).This was also why numerous carbon deposits were observed in the Ni-Ir/MgAl2O4 catalytic system, but the effect on the catalytic activity was minimal.As ν(CH4) approached ν(CO2), the carbon deposits would no longer increase, but the existing carbon deposits would not be eliminated.In other words, the system achieved a balance between carbon deposition and elimination, and the overall reactivity was determined by the activation of CO2.Over the Ni/MgAl2O4 catalyst, the dissociation of CH4 and the activation of CO2 occurred at the same active site, which did not follow this mechanism; thus, carbon deposits increased throughout the test.For the Ni3Ir1/MgAl2O4 catalyst, the Ni sites were reduced sufficiently such that the reaction system reached ν(CH4) ≈ ν(CO2) equilibrium after 20 hours.increased.When the Ir/Ni ratio was 2, both of the initial and average conversions reached the highest level, and there was no further impact with an additional increase of the Ir content.In addition, the carbon deposits decreased from 14.2 wt% (Ni/Ir = 12/1) to 0 wt% (Ir/Ni = 2), and no carbon deposition occurred with further increasing the Ir content.This was because the increased Ir content brought more catalytic sites for CO2 activation, allowing the system to reach the equilibrium between carbon deposition and elimination faster.Therefore, while maintaining high activity, the amount of carbon deposits was regularly reduced during the test.As the number of CO2 activation sites increased, the activity of the rate-determining step in the reaction increased, and the system exhibited higher activity.This result was consistent with the balance theory, as when Ir/Ni = 2, ν(CO2) matched ν(CH4) during the initial stage of the test.Therefore, the carbon deposition was almost negligible, with almost no differences between the initial activity and the average activity of the reaction, indicating an extremely high stability of the catalytic system.
We anticipated that a match between ν(CO2) and ν(CH4) in the initial state of the reaction by tuning the Ni-Ir relative contents would achieve a balance between the generation and elimination of carbon deposits, leading to a DRM catalytic system with zero carbon deposits.Because Ir is a precious metal, it was necessary to reduce the Ir content as much as possible.As shown in Figs.5c-d, the Ir/Ni ratio of 2 was found as the best catalyst ratio.We then conducted a 600 hours long-time test under the conditions of 650 °C, GHSV = 40,000 mL•g −1 •h −1 , and 1 bar (Fig. 5e).The initial conversion rates of CH4 and CO2 were 62.67 % and 72.94 %, respectively, with no significant change in activity after 600 hours.Raman spectra of the spent catalyst showed no obvious signal peaks for amorphous carbon at 1334 cm −1 or 1598 cm −1 (Fig. 5f).In addition, no significant weight loss was observed from TGA (Fig. 5g).These results confirmed that no obvious carbon deposition occurred on the catalyst, consistent with the results of the catalytic system optimized based on our proposed theory.

from 40 .
46° to 40.92° for the Ir(111) plane and from 47.12° to 47.66° for the Ir (200) plane, Fig. S2), consistent with the smaller atomic radius of Ni than Ir.As shown in XPS profiles of Fig. 1b, the Ni 0 2p3/2 (855.2 eV) and Ni 2+ 2p3/2 (858.2 eV) signals were observed for the Ni/MgAl2O4 catalyst.The presence of Ni 2+ cations indicates the formation of NiAl2O4 species on the support surface during the synthesis process.Compared with Ni/MgAl2O4, the Ni 2+ content and binding energy position in the Ni3Ir1/MgAl2O4 catalyst changed minimally, while the Ni 0 2p3/2 signal increased to 858.6 eV (by 0.4 eV).In contrast, the binding energy of Ir 0 4f2/7 was 62.2 eV for the pure Ir catalyst (Fig. 1c), 0.4 eV higher than that in the Ni3Ir1 catalyst (61.8 eV).With respect to the monometallic Ni and Ir catalysts, these binding energy changes indicate the electrons were transferred from Ni to Ir, suggesting the formation of Ni-Ir alloy in the Ni-Ir/MgAl2O4 system.Ni and Ir K-edge X-ray adsorption fine structure (XAFS) spectroscopy was also used to identify the localized structure of the Ni-Ir alloy (Fig. S3, Table S1), which shows the existence of Ni-Ir coordination bonds (2.63-2.65 Å) in the Ni3Ir1/MgAl2O4 catalyst.
content of Ir (Ni/Ir ≈ 2/3), while Ni dominated in the large particles (Ni/Ir ≈ 4/1).The uniform distributions of the Ni and Ir elements within each kind of the Ni-Ir particles (Figs.1f-g, S5) agree well with the formation of alloys in the Ni-Ir/MgAl2O4 catalyst.Taken together, XRD, XPS, XAFS, TEM, and EDS were combined to verify that Ni-Ir alloys were formed in the bimetallic Ni-Ir/MgAl2O4 system, which significantly improved the dispersion of the Ni nanoparticles on the MgAl2O4 support and provided a basis for the efficient coupling between CH4 dissociation and CO2 activation on the Ni and Ir active sites, respectively, as demonstrated next for DRM.
TEM characterization was further applied to analyze the structure and morphology of the spent catalysts after long-time reaction (Fig.S8).The average diameter of the Ni nanoparticles on Ni/MgAl2O4 increased from 5.7 to 8.2 nm (5.7 nm for the fresh sample) after 100 hours of testing, indicative a severe sintering of the Ni particles at the condition of DRM.In contrast, the metal particle sizes of the spent Ni3Ir1 and Ir catalysts (1.2 and 4.1 nm for Ni3Ir1/MgAl2O4; 1.1 nm for Ir/MgAl2O4) did not change significantly compared with the fresh ones.It is surmised that metallic Ir has a strong interaction with the MgAl2O4 support, which results in the smaller metal particles and stronger anti-sintering ability for the Ir-containing catalysts.

Fig. 3
Fig. 3 In-situ DRIFTS with (a-c) introduction of CH4 in the second stage, (d-f) introduction of CO2 in the third stage of Ni/MgAl2O4, Ni3Ir1/MgAl2O4, Ir/MgAl2O4, the schematic diagram of the mechanism to activate CO2 (g) with the help of carbonate and (h) without the help of carbonate.
during DRM.Temperature programmed surface reaction-mass (TPSR-MS) was used to explore the activation ability of CH4 during the DRM process.First, CH4 was introduced during the temperature-programming process (Figs.S14a and S15).The Ni/MgAl2O4, Ni3Ir1/MgAl2O4, and Ir/MgAl2O4 catalysts exhibited CH4 activation abilities at 289, 328, and 347 °C, respectively.The carbon species generated by the dissociation of CH4 blocked the active sites, restricting the activation of CH4.This caused the CH4 activation ability to decrease at 490, 610, and 590 °C, respectively.This indicates that Ni was better than Ir for dissociation of CH4, and the dissociation rate increased with raising temperature.
Ni8 clusters, no matter the C atom of the adsorbed CO2 on the Ni6Ir2 clusters was bound to the Ni site (i.e., Ni6Ir2/MgAl2O4-a; -1.33 eV) or the Ir site (i.e., Ni6Ir2/MgAl2O4-b; -1.77 eV).DFT calculations showed that the CO2 dissociation (CO2* → CO* + O*, Fig. 4b) at the Ir8/MgAl2O4 interface was exothermic by -0.41 eV with an activation barrier of 1.48 eV, which was considerably lower than that on Ni8/MgAl2O4 interface (2.18 eV).The difference of the activation barrier for CO2 dissociation indicates, compared with the Ni8 cluster, the generation of O* species on the Ir8 cluster was much more efficient, rendering a fast elimination of carbon deposits and thus a high catalyst stability during DRM.For the two Ni6Ir2/MgAl2O4 models, CO2 was dissociated on Ni6Ir2/MgAl2O4-a with the formed CO* species bound to a Ni site via a C-Ni coronation and the formed O* atom bound to a vicinal Ir site, while these two moieties formed on Ni6Ir2/MgAl2O4b were bound to the Ni and Ir sites inversely.CO2 dissociation on Ni6Ir2/MgAl2O4-a showed a much lower activation barrier than that for Ni6Ir2/MgAl2O4-b (1.69 vs. 2.21 eV, Fig. 4b), further reflecting different stabilities of the incipiently formed CO* and O* species at the transition state (TS) of CO2 dissociation on the Ir and Ni sites (TS structures shown in Fig. S17).In order to unveil the determining factors on the activity of CO2 dissociation, adsorption energies for O atom and CO on the supported M8 clusters were calculated independently to compare the stabilities of these two species on different metal sites.As shown in Fig.4c, the CO2 dissociation barrier presented a nearly linear correlation with the O adsorption energy, while no apparent correlation was observed between the CO2 dissociation barrier and the CO adsorption energy.These data clearly imply that the activity of CO2 dissociation is mainly determined by the stabilization of the O* species at the TS, which prefers the Ir site over the Ni site.The projected density of states (DOS) distributions of adsorbed O and M8 clusters were further analyzed (Fig.4d) to obtain a deeper understanding of the higher oxygen affinity of Ir.It is found that the antibonding states of O2p became more populated in the trend of Ir8/MgAl2O4, Ni6Ir2/MgAl2O4-a, Ni8/MgAl2O4 and Ni6Ir2/MgAl2O4-b, accounting for the decrease of the stability of the O* species and the increase of the CO2 dissociation barrier with this trend.Accordingly, the excellent stability of Ir/MgAl2O4 and Ni3Ir/MgAl2O4 catalysts during DRM is attributable to their stronger oxophilicity derived from metal Ir.

Fig. 4 (
Fig. 4 (a) Four M8/MgAl2O4 catalyst models applied in the theoretical calculations of DRM (with CO2 adsorbed on the metal clusters).(b) DFT-derived energy changes of CO2 dissociative adsorption on M8/MgAl2O4.(c) Correlations of the CO2 activation barrier with the adsorption energies of O-atom and CO on M8/MgAl2O4.(d) Local density of states projected onto the adsorbed oxygen 2p state and M8 cluster d state for the oxygen chemisorption on M8/MgAl2O4.

Fig. 5 (
Fig. 5 (a) Relationship between carbon deposits and time on stream for Ni/MgAl2O4, Ni3Ir1/MgAl2O4, and Ir/MgAl2O4, (b) schematic diagram of the balance theory in DRM, (c) weight percent of carbon deposits from spent catalysts with varying Ir/Ni ratios, (d) relationship between the Ir content and catalytic activity, (e) catalytic performance during a long-term test of Ni1Ir2/MgAl2O4 for DRM, (f) Raman spectra, and (g) TGA of the spent catalysts after the long-term test of DRM.
2-, ads), thus allowing the carbonate species to reach saturation.When CH4 was introduced in the second stage, it was dissociated on the active metal sites, forming surface CHx* (x = 0-3) species (CH4 (g) → H* + CHx*), which continued to accumulate and form carbon deposits.The carbonate species on MgAl2O4 cannot be used by the active Ni sites directly, while it can decompose into gaseous CO2 (CO3 2-ads → CO2 (g)) again, which then adsorbed on the active Ni sites and further reacted with CHx* to form CO* (CO2* + CHx* → 2CO (g) + x/2 H2).CO* was finally released from the catalyst surface to form gaseous CO.Therefore, when CH4 was introduced during the second stage, CO was not produced until sufficient carbonate was consumed to generate CO2.Only easily eliminated carbon species were activated and removed by the reaction with CO2 released by the support in the second stage, leaving the catalyst surface covered by the recalcitrant carbon chains produced by overgrowth.Since Ni could not effectively activate and utilize CO2 to eliminate the carbon chains, when CO2