Catalytic hydrogen combustion over supported metal catalysts; comparison and kinetic study

Catalytic


Introduction
H 2 , as a carbon-free energy vector, aids the transition to net-zero emission and paves the way for a potential green energy cycle.As a result, H 2 has gained tremendous interest in the industrial sectors as an efficient solution to reduce carbon footprint.Moreover, H 2 exhibits the greatest energy density (33.3 kWh/kg) among all other energy molecules [1].
H 2 could be used directly as a fuel or converted into another energy vector.In either of these applications, two main concerns need to be addressed: (1) the efficiency of H 2 utilization and (2) safety aspects with respect to the H 2 concentration in the exhaust gas of the system.The H 2 content in the exhaust gas must retain below the H 2 flammability limit.This value is around 4 vol% of H 2 in air at normal pressure and temperature conditions (NTP, 293 K and 1 atm) [2].Therefore, removing unreacted H 2 from the exhaust fumes, known as H 2 slip, is crucial due to the risk of explosion.The remaining H 2 in the exhaust gas could be mitigated by combusting the H 2 .However, conventional H 2 combustion in an open flame mode would not only be a waste of resources, but also has drawbacks, such as the drastic risk of flashbacks and NO x emission.Hence, catalytic H 2 combustion (CHC) is a promising solution to overcome these obstacles [3][4][5][6].Throughout this process, H 2 get oxidized to water through a highly exothermic reaction (Eq.( 1)).This heat can be subsequently used to increase the overall efficiency of the H 2 -involved process.
Moreover, since no carbon is involved in this process, the cycle is naturally closed and the water vapor could be released directly to the atmosphere.Currently, the CHC reaction has been implemented in H 2 cookers [6][7][8][9], Passive Auto-Recombiners (PARs) [10][11][12], methanation reactors and fuel cell anode exhaust [13].Recently this reaction has been used in H 2 sensors [14] and thermoelectric generators [15].Furthermore, due to the increasing deployment of green H 2 production, there is prospective application of the CHC reaction in the combined H 2 , heat and power (CHHP) systems [16][17][18] to produce heat and ensure H 2 concentration in the exhaust gas to be below the flammability limit of H 2 in the air.
H2O(g) = − 241.82 kJ / mol Eq. 1 Supported Pt and Pd catalysts are almost exclusively used catalysts in the CHC reaction [19][20][21][22][23][24][25].However, their utilization raises several concerns, including catalyst deactivation due to metal particle agglomeration, water poisoning, and, more importantly, the high cost and scarcity of precious metal resources.To address these issues, researchers have adopted approaches such as substituting precious metals with other transition metals (TMs) [25][26][27].Kozhukhova et al. [25] have investigated the effect of Co addition on the performance of Pt-Al 2 O 3 catalyst under the CHC reaction.Although long-term catalytic activity is achieved, there is an urge to replace precious metals with other active catalysts.Recently, Erturk and Elmaci [27] analyzed different support materials and highlighted the importance of replacing precious metal catalysts with Non-precious TM catalysts to develop a sustainable future.
The primary motivation for using the precious metals in the CHC reaction is the relatively low temperature reaction start-up.However, this low-temperature start-up is not necessary for many applications, such as anode exhaust of the solid oxide fuel cell.Despite the high abundance and low cost of non-precious TMs, there is limited literature on the application of non-precious TM catalysts in the CHC field, making them attractive candidates for further research.
The kinetic parameters are essential for reactor design and process optimization.Currently, this information is only available and limited to precious metals catalysts [28][29][30][31][32][33].Kim et al. [34,35] have determined the rate law for Pd-Cu/Al 2 O 3 and Cu/Al 2 O 3 catalysts.However, they have used the available kinetic parameters in the literature for this aim.As a result, the kinetic parameters need to be determined to extend the consumption of non-precious TM catalysts toward the CHC field.
The kinetic study is only valid if the kinetics determine and control the reaction, where the mass and heat transfer limitations are negligible.In this regard, Joshi et al. [36] have determined the different mass and kinetic control regions for a monolithic Pt-based catalytic reactor for the CHC reaction.However, to the best of our knowledge, this practical information is not available for any other supported TM catalysts under the CHC reaction.
Hence, the focus of the current study is to investigate the CHC reactivity of different supported TM catalysts systematically.First, improvement in the metal-support interaction is achieved using an ammonia solution-based pre-treatment on the γAl 2 O 3 support material.Then, through a highly reproducible and simple Multi-step incipient wet impregnation (MIWI) method, M-γAl 2 O 3 (M = Pt, Ru, Co, Ni and Mo) are synthesized.The catalytic performance of the synthesized M-γAl 2 O 3 and their long-term stability under the CHC reaction are investigated.The catalyst surface chemistry is studied using X-ray Photoemission Spectroscopy (XPS).The nanoparticle (NP) size distribution is investigated through scanning transmission electron microscopy (STEM).The kinetically controlled region of the CHC reaction on each catalyst is determined.The empirical activation energies and pre-exponential factor values of the catalyst toward the CHC reaction are then calculated using the Arrhenius kinetic model in the kinetically controlled region.Based on the performance data, we provide a series of catalytic activity and kinetics parameters that will facilitate reactor design and CHC process optimization in the future.

Catalysts synthesis and preparation
γAl 2 O 3 is chosen as the support material due to its low cost and high surface area.The γAl 2 O 3 support is provided from Jiangxi Co. LTD and has a surface area of 250 m 2 /gr (Table 1).The M-γAl 2 O 3 catalysts are synthesized through highly reproducible and simple MIWI method.A metal salt is used as the metal precursor and DI water as the solvent.The list of the metal salts, the corresponding metal loading, and other relevant synthesis parameters are summarized in Table S1.
In order to improve the metal-support interaction, γAl 2 O 3 support is pre-treated (activated).In this procedure, the γAl 2 O 3 support is immersed in an ammonia solution (NH 4 OH).After washing thoroughly with DI water, the γAl 2 O 3 is filtered and dried at 120 • C [37,38].This activated support is later impregnated with a metal salt solution.A metal solution with a desired concentration (5 mg/ml solution ) is added step by step to the activated γAl 2 O 3 .In each step, 0.4 ml of the metal solution, equivalent to the pore volume of support, is added to 1 gr of the support.This synthesis step is crucial to ensure the homogenous distribution of subsequently formed NPs.After drying in an oven at 100 • C, the subsequent hydroxide powders reduced at the relevant temperature (Table S1) under forming gas (FG, 5 vol% H 2 diluted in N 2 ) atmosphere.The reduction temperatures are chosen based on the H 2 temperature programmed reduction (H 2 -TPR) tests of each catalyst and their reduction properties.The TPR tests are provided in the supplementary data (SI) section SI1 (Fig. S1).An extra step is performed to remove the Cl ion in the case of Ru and Pt catalysts to prevent the poisoning effect of the Cl ion originating from the metal salt.In this step, the Cl ions of the reduced Ru and Pt catalysts are removed using an ammonia solution.Later, this step is followed by washing with DI water, filtration and drying at 100 • C overnight.Cl ion is believed to be poisonous for the catalyst active sites.Later, the granules are formed from the synthesized catalyst powders using a 250-500 mesh sieve.

Characterization
Catalysts phase characterization is carried out using X-ray diffraction (XRD) on a Bruker D8 advanced with Cu K α (λ = 1.54 Å) radiation and 40 kV, 40 mA.Transmission electron microscopy (TEM) is performed in order to analyze catalyst morphology, particle size, and size distribution.TEM images are acquired using a Thermo Fischer Scientific Tecnai-Osiris at 200 kV in scanning (STEM) mode.This machine is equipped with X-FEG and Super-X energy dispersive x-ray spectroscopy (EDXS) detectors.STEM images are acquired in high-angle annular dark-field (HAADF) conditions.STEM-EDXS elemental maps are obtained using a beam current of 900 pA in STEM mode.The electron microscopy images are analyzed by TIA, Velox, Esprit, and ImageJ software.The oxidation state before and after the CHC reaction is studied in detail using XPS (Kratos Axis Supra with Monochromated Al Ka X-ray source).The CasaXPS software is used to fit the XPS curve with the help of references and databases such as [39][40][41][42][43].The TPR tests are performed using Thermogravimetry analysis coupled with a Mass Spectrometer (TGA-MS by Netzsch).Surface area determination, as well as pore size and distribution, are obtained using N 2 adsorption/desorption experiments Table 1 The average particle size and the BET surface area (S BET ) of theγAl 2 O 3 and fresh catalysts.(Belsorp II max).H 2 adsorption measurement is also done using the same instrument to study the H 2 adsorption capacity.All samples are degassed at 180 • C for 3 h before the adsorption/desorption measurements.The metal catalyst concentration is confirmed using an inductively coupled plasma-mass spectrometer (ICP-MS, nexION35010, Perkin Elmer).

Experimental setup and CHC catalytic activity measurement
A plug flow reactor is used to study the CHC catalytic performance.The reactor temperature is probed using two K-type thermocouples: one inside the reactor before the catalytic bed and the other outside the reactor in the center of the catalytic bed.Prior to the introduction of the reactive gas mixture, all catalysts are activated at 400 • C in the FG atmosphere for 1 h.Then, the reactive gas mixture is introduced into the reactor with a total flow rate of 20 ml/min and a space-time of 45 s.The reactive gas mixture contained 4 vol% of H 2 , with the stoichiometric ratio between H 2 and O 2 and the rest of N 2 as the carrier gas.It should be noted that the H 2 concentration is chosen as the maximum amount before the lower limit of H 2 flammability in air.To analyze the exhaust gas and calculate the H 2 conversion rate (r) (Eq.( 2) and Eq. ( 3)), it is directed to a quadrupole mass spectrometer (QMS, OmniStar 320, Pfeiffer Vacuum).Multi-ion detection (MID) mode in the QMS, coupled with an analogously inputted thermocouple, enabled fast gas detection with sufficient statistics (36 data points per minute).The performance is determined between room temperature (RT) -500 • C by continuous temperature ramping with a heating rate of 1 • C/min.More information on the setup can be found in our previous publication [44].Moreover, a 45 h stability test is performed under the same CHC reaction condition at 240 • C to compare the stability of the catalysts.

Hydrogen conversion
In order to cover the low conversion range at low temperatures, a new low-temperature setup is designed and used.Interested readers are referred to section SI2.
The kinetically controlled region of the CHC reaction is experimentally determined as explained in section 3.3.1 for each catalyst, and validated by the Weisz-Prater Criteria.

Catalysts physicochemical properties
The XRD is utilized as a preliminary technique to characterize the fresh (as-synthesized) catalyst phases, as shown in Fig. 1a.The cubic structure of γAl 2 O 3 is characterized by its broad peaks (Fig. 1a), which is not altered by the support activation process (Fig. S3).
Upon introducing 1 wt% Pt on the γAl 2 O 3 support, a small characteristic peak at 39.5 • could be observed, typical for Pt NPs.For the 1Ru-γAl 2 O 3 catalyst, such evidence is completely obscured by the γAl 2 O 3 phase (Fig. 1a), indicative of small particle sizes and relatively low loading.As the non-precious metals are expected to pose less catalytic activity toward the CHC reaction compared to Pt-group metals, a higher loading of 10 wt% is used.Unsurprisingly, more distinguishable peaks could be observed for Ni and Co (Fig. 1a).Interestingly, for the 10Mo-γAl 2 O 3 catalyst, no clear evidence of a metallic Mo phase could be discerned (Fig. 1a).Instead, a MoO 2 phase had formed with a peak around 26.0 • .
The catalyst surface chemistry, as well as the effectiveness of the Cl ion removal step, is investigated using the XPS method.γAl 2 O 3 is a hydrophilic oxide in which the presence of hydroxyl groups on the surface is inevitable.These hydroxyl groups are stable even at temperatures as high as 900 • C [45].Consequently, even the reduced surface of the γAl 2 O 3 poses a large amount of OH groups (Fig. S5).Fig. 1b provides the XPS graphs of Fresh catalysts.Pt 4f orbital, which is commonly used in the XPS analysis of Pt, overlaps with Al 2p.Hence, to achieve a reliable fitting, Pt 4d orbital is utilized [46,47].Concerning the doublet peaks observed at higher binding energy (BE) of 318.6 and 335.5 eV in the Pt 4d spectra, the associated peak to Pt 4d 5 / 2 (318.6 eV) has a higher BE than what is expected for Pt 2+ and lower than Pt 4+ [46].Thus, this peak could be associated with PtCl x , although due to the noise level in the Pt 4d region, one cannot exclude the presence of PtO and PtO 2 .Furthermore, the 1Pt-γAl 2 O 3 sample still contains traces of Cl ions (interested readers are referred to the SI4 and Fig. S6).The peak at a lower BE of 307.8 eV is associated with some C 1s energy loss features [48].
For the fresh 1Ru-γAl 2 O 3 sample, the parallel presence of both Ru oxide(s) and metallic Ru 0 is evident, as demonstrated in Fig. 1b.
The fresh 10Co-γAl 2 O 3 catalyst is likely to contain both Co(OH) 2 (at 780.6 eV) and CoO (at 780.2), as represented in Fig. 1b [42].This is in agreement with the XRD results in Fig. 1a, where only the broad peaks of CoO are observable in the fresh sample.The spectra obtained from the photoelectrons that exited the Ni 2p can be very complicated since these photoelectrons undergo multiple processes [43,49].In the current study, the curve-fitting parameters from Biesinger et al. [42] is used to fit the XPS spectra of Ni 2p of fresh and used 10Ni-γAl 2 O 3 catalysts.In the curve fitting, the FWHM of the photoelectron peaks are small, and the ones from the satellites are broad, as often.The XPS graph of fresh 10Ni-γAl 2 O 3 catalyst possesses a strong low BE shoulder of the metallic Ni at 852.6 eV.The high portion of NiO compared to the metallic Ni in the fresh sample arises from the fact that the XPS is surface sensitive while XRD is bulk sensitive.Therefore, the NiO is more on the surface, while the core is metallic Ni.
Regarding the 10Mo-γAl 2 O 3 sample, based on available references, the peak at BE of 229.3 eV and 232.6 eV correspond to Mo 4+ (MoO 2 ) and Mo 6+ (MoO 3 ) species, respectively.However, the lower BE's peak in the fresh sample is 231.8 eV (Fig. 1b).This means there is a contribution of Mo 5+ in the fresh sample [50][51][52][53].The presence of partially reduced Mo is believed to be crucial for the catalytic activity.Nevertheless, Mo 6+ is the dominant species in the fresh sample, accounting for 67.1% of the sample.
Aberration-corrected-STEM imaging is utilized to observe the supported catalyst particles' morphology, size, and size distribution.Using the HAADF mode, the catalyst particles are visible due to the Z-contrast compared to the γAl 2 O 3 support (Fig. 2).The bright contrasts in the STEM images are attributed to the catalyst phases.This is further supported by the EDXS analysis provided in Fig. 2. In all the samples, the nano-sized catalyst phase is uniformly distributed on the support surface, which grants a high surface-to-volume ratio.This efficient consumption of the catalyst is of great importance, especially in the case of precious metals, and it can only be achieved by precisely controlling the synthesis parameters.The average particle sizes of the catalysts are noted on the size distribution histograms in Fig. 2 and are summarized in Table 1.The metallic sub-nanosized Pt and Ru NPs are well-distributed on the support with a spherical morphology.However, in the case of 10Co-γAl 2 O 3 , it seems the sub-nano-sized Co oxides are agglomerated to form clusters with an average size of 5.9 ± 2.4 nm.The cluster formation can be a consequence of the magnetic properties of the Co oxides.Both Co 3 O 4 and CoO are antiferromagnetic materials at RT.However, nanosized Co 3 O 4 and CoO have shown ferromagnetic properties [54], which can explain the reason behind the particles' agglomeration.
Among all samples, Ni NPs have the largest particle size, while Mo sample has revealed the smallest particle size of MoO 2 with an average value of 8.7 ± 1.7 and 0.5 ± 0.1 nm, respectively.
The particle size obtained from STEM images, is used later to calculate the Turnover Frequency (TOF) of each catalyst under the CHC reaction condition.
In order to study the effect of introducing the catalyst phase on the surface area, pore volume, and pore size distribution of the γAl 2 O 3 support, the N 2 adsorption/desorption measurements are performed.As shown in Fig. 3a, the N 2 adsorption/desorption isotherms are classified as type IV physical adsorption.This indicates that the interaction between the sample surface and adsorbate is relatively strong, and pores feature as mesopores (2-50 nm).The surface areas are calculated using the BET (Brunauer-Emmett-Teller) method and are provided in Table 1.
The pore size and distribution are derived using the BJH (Barrett, Joyner, and Halenda) method (Fig. 3b).The average pore size (diameter) of γAl 2 O 3 support is found to be 7.4 nm with a maximum frequency size of 5 nm.
The introduction of Pt NP's has increased the BET surface area (S BET ).These NPs with an average size of 0.92 nm have created more micropores in the range of <2 nm.These micropores have the most pronounced effect on the S BET .Compared to other catalysts, Ru NPs have a relatively minor effect on pore size and distribution.This is likely due to the presence of sub-nano-sized Ru NPs that do not contribute to the pore structure.
On the other hand, introducing Co NPs and Ni NPs with average particle sizes of 6.36 and 8.79 nm, respectively, resulted in a decrease in the surface area.This is because they blocked the support pores in the size ranges of 5-11 nm and 6-12 nm, respectively.The size of the Co NPs and Ni NPs is approximately the same as the pores of the γAl 2 O 3 support, which led to the blockage of some of the support pores.However, the presence of Co NPs and Ni NPs has created more micropores with a size of <4 nm, suggesting the porous structure of the synthesized Co and Ni NPs.Additionally, Mo NPs have introduced more pores with a size of <1 nm, which could be attributed to the sub-nano-sized Mo NPs.
Anyhow, one should be cautious when interpreting N 2 adsorption/ desorption measurements since the synthesis procedure can affect the support properties.For instance, it has been shown that every 100 • C increase in the synthesis temperature reduces the surface area of the γAl 2 O 3 support by 20-40 m 2 /gr [55].

Catalysts activity measurements
The catalytic activity of the different catalysts is measured under the CHC reaction condition, as explained in the experimental section.The calculated H 2 conversion sigmoidal curves (X H2 in Eq. ( 2)) are represented in Fig. 4. At low temperatures where, X H2 < 10%, the reaction rate is mainly determined by the reaction kinetics.In this region, the reaction rate has an exponential temperature dependence based on the Arrhenius kinetic model.For X H2 < 10%, the H 2 conversion is considered to be small enough that the mass and/or heat transfer do not limit the reaction rate.This is because the time for the CHC reaction is much larger than that for gas species and heat diffusion.Moreover, this approximation is necessary to validate the kinetic control region based on the Weisz-Prater criteria (Table S2).
With increasing temperature, since the CHC reaction is a thermally activated phenomenon that is not limited by thermodynamics, H 2 conversion increases exponentially based on the Arrhenius model.However, since the diffusion of the gases in the gas phase only slightly changes with increasing temperature, the reaction enters a mixed control region where the kinetic and mass/heat transport controls the overall reaction rate.
While the reaction rate has an exponential dependence on the temperature, the diffusion rate of gas species has a weaker temperature dependence.Thus, with a further increase in temperature, at some point, the X H2 is high enough that reactant depletion and product accumulation in the vicinity of the catalyst limit the reaction.Therefore, hereafter, the reaction is dominantly controlled by mass transport.Although all the supported metal catalysts follow the same trend, their catalytic activity is not the same.In the following, the origin of different CHC catalytic activities, as well as the reaction start-up properties, are compared between different M-γAl 2 O 3 catalysts.

Comparison of the CHC catalytic activity
As evidenced by the CHC catalytic activity measurement, the Pt catalyst has the highest CHC catalytic activity.Among all group 10 TMs, Pt has the lowest O 2 adsorption energy (− 60.79 kJ/mol on (111) facet for peroxo species) and highest O 2 dissociation energy (60.79 kJ/mol on (111) facet for peroxo species).This means that adsorbed O 2 on Pt does not tend to dissociate and oxidize the Pt [56,57].On the other hand, H 2 adsorption on the Pt surface is thermodynamically more favorable than  O 2 , given the fact that H 2 adsorption energy (− 88.77 kJ/mol) is higher than the O 2 adsorption energy [58].It is well-known that Pt has an activation toward H 2 molecules and not O 2 molecules.Since hydrogen has a lower molecular mass and a higher sticking coefficient than oxygen, before the start-up of the CHC reaction, the surface is mainly covered by hydrogen.Hence, the reaction rate is low.Increasing the temperature, a transient step occurs, and the surface adsorbed hydrogens start to oxidize [59].This is while early TMs from groups 4 to 8 have relatively high O 2 adsorption energies with a very small energy barrier for the O 2 dissociation [57].As a result, the M − O binding energy is higher in these metals compared to the precious metals.Moreover, since TMs have a large density of electrons in their fermi level, the H 2 adsorption on these metals is spontaneous and increases exponentially with the temperature [60,61].The difference in the CHC catalytic activity for these metals arises from the reactivity of the adsorbed hydrogen and oxygen species.A relatively strong M − O and M − H bindings (Table S5) will cause a low catalytic activity and delay the CHC start-up.Interested readers are referred to section SI6, which compares the values of O 2 and H 2 adsorption energy as well as M − H binding energy on different TMs.For these metals, the O 2 adsorption energy has a higher absolute value than that of the H 2 adsorption energy.This means that, thermodynamically, O 2 dissociation adsorption is more favorable than H 2 adsorption.As a result, upon introducing the reactive gas, these metals tend to oxidize and further show sluggish activation toward the CHC reaction.
For example, in the case of Co, the O 2 adsorption energy is reported to be − 218.06 kJ/mol, while the O 2 dissociation energy is only about 3.86 kJ/mol on the (0001) facet [57,62].This also means that Co and O have a strong bond, and despite the preformed reduction activation step, the reduced Co surface undergoes oxidation upon introduction of the reactive gas to the 10Co-γAl 2 O 3 sample, as proved by XPS (Fig. S4b).However, since the induced Co oxide in the surface (Co 3 O 4 ) is reducible, lattice oxygens participate in the CHC reaction, which results in a high catalytic activity of the Co-γAl 2 O 3 sample which is comparable to Ru-γAl 2 O 3 sample.This is specially important in designing a catalyst for the CHC reaction since although Ru is the cheapest precious metal, the Co price is almost 500 times less than Ru (The metal prices are compared based on www.dailymetalprice.comas of June 2024.).Furthermore, by considering the diverse temperature requirements across different applications, our study can assist in selecting catalysts that are effective within specific temperature ranges.
In terms of the CHC activity of different catalysts, the precious metals are the state-of-the-art catalysts for the CHC reaction.However, to the best of our knowledge, it is the first time that the catalytic activity of the different supported metal catalysts is investigated under the CHC reaction.It should be noted that the catalytic activity of the synthesized Pt-γAl 2 O 3 in the current study is higher than what has been reported for the γAl 2 O 3 -supported Pt catalysts, with a similar Pt loading in the previous studies [10,36,63].This can be attributed to sub-nano-sized Pt NPs well dispersed on the γAl 2 O 3 support, resulting from enhanced metal-support interaction through the support activation procedure.
Overall, based on the XPS results provided in section 2.2, the key contributing factor in the CHC catalytic performance seems to be the oxidation state of the metal catalyst.Apart from Pt, during the CHC reaction, the catalysts (be it metallic or a lower-order oxide) tend to undergo further oxidation despite the presence of both H 2 and O 2 in stochiometric amounts (Fig. S4).This leads to competitive oxidation between hydrogen species and metal catalysts during the CHC reaction, significantly hindering its activity.Consequently, it can be inferred that the poor catalytic activity of TMs (excluding Pt) stems from their tendency to undergo oxidation during the CHC reaction.However, gaining deeper insights into the relative timing of catalyst oxidation requires insitu experiments, which fall beyond the scope of this study.

Reaction start-up properties
The reaction start-up properties of the different catalysts are sum-marized in Table 2. Light-off temperature (T L ) is considered to be where the H 2 conversion reaches 10%.T 50% and T 100% are associated with the temperatures of X H2 = 50% and 100%, respectively.Moreover, the reaction start-up speed is evaluated by the T L -T 50% value.The lower the value, the sharpest and faster the start-up will happen when the CHC reaction proceeds over the catalyst.This value for the different catalysts is in the following order: 1Pt-γAl 2 O 3 < 1Ru- This sharp startup in 1Pt-γAl 2 O 3 arises from the catalyst's high activity with fast kinetics, resulting in a sudden transition from kinetic control to mass/heat transfer control regime.More insights for better understanding this sharp reaction start-up are explained in the next section.

CHC stability of the catalysts
Given the promising CHC catalytic activity of 1Ru-γAl 2 O 3 and 10Co-γAl 2 O 3 , these two catalysts are opted for a stability test.The long-term catalytic activity is assessed over 45 h at 240 • C where X H2 ≅ 100% (Fig. 5a).The reaction conditions remained consistent with that described in section 2.3.It is evident that both catalysts are stable under the CHC reaction condition.However, while the 1Ru-γAl 2 O 3 catalyst does not exhibit any deactivation, the 10Co-γAl 2 O 3 catalyst experiences a 1% degradation in the CHC catalytic activity after 45 h time on stream.The slight deactivation of the 10Co-γAl 2 O 3 catalyst could be attributed to NPs agglomeration, as evidenced by the STEM-EDXS of the used 10Co-γAl 2 O 3 catalyst (Fig. S7), as well as oxidation of the catalyst (Fig. S4b).To investigate the latter, the used 10Co-γAl 2 O 3 catalyst is regenerated under a reductive atmosphere (under FG) for 1 h at 400 • C. The regenerated catalyst is again exposed to the reactive gas under the same condition as in section 2.3, for an additional 5 h (Fig. 5b).This figure illustrates 3 regions: I) the first 5 h of the stability test, II) the last 5 h of the stability test, and III) the 5 h activity test following the regeneration step.As the results show, the CHC catalytic activity of used 10Co-γAl 2 O 3 catalyst is fully recovered after the regeneration process.This recyclability makes 10Co-γAl 2 O 3 catalyst a suitable non-precious metal candidate for the CHC reaction.
It is important to note that while NPs agglomeration also happens for 1Ru-γAl 2 O 3 catalyst, the CHC catalytic activity remains relatively stable, indicating that after an initial agglomeration, oxidized Ru NPs are stable under the CHC reaction.
In conclusion, although 1Pt-γAl 2 O3 is the best catalyst to initiate the CHC reaction at RT, both 1Ru-γAl 2 O 3 and 10Co-γAl 2 O 3 achieve high X H2 at slightly elevated temperatures.Specifically, in applications where the exhaust gas is already at high temperatures, such as solid oxide cells and methanation reactors, the use of expensive Pt catalyst is unnecessary.

Kinetically control region determination
To exclude the effects of catalyst particle size and dispersion, the TOFs are calculated according to Eq. ( 4) for all the samples.where isperssion = NS NT .In this equation, N S and N T are the number of the surface and total atoms in the particle, respectively.Interested readers are referred to SI7 for more information on calculating the N S , N T , and dispersion.
Upon calculation of TOF, plots of TOF -1000 / T are sketched.It should be emphasized that E a is independent of TOF calculation since only X H2 , and T define the slope of the Arrhenius plot, and other terms will only affect the y-intercept.According to the Arrhenius kinetic model, the kinetic control region is essentially part of the ln (TOF) -1000 / T plots with a constant slope, equivalent to a second derivative equal to zero, where there is no curvature and basically a flat line in the slope of ln (TOF) -1000 / T plots (Fig. 6).In this region, kinetics predominantly control the reaction, while mass/heat transfer limitations are negligible.The deep valley in the plot of slope of ln (TOF) - 1000  / T is due to the curvature change in the ln (TOF) -1000 / T plots.The minimum in these curves is the inflexion point of ln (TOF) -1000 / T plot.Finally, the temperature region in which the CHC reaction is kinetically controlled for each catalyst is summarized in Table 2.As can be seen, this region for all catalysts is around 20-30 • C wide.
This experimentally determined kinetic control region is validated by Weisz-Prater Criteria.Interested readers are referred to section SI3 for information on how these criteria are calculated.

Activation energy and pre-exponential factor
After determining the kinetically controlled region for each catalyst, the amount of E a and pre-exponential factor or frequency factor, known as the A value, are calculated based on the Arrhenius kinetic model.It is important to note that the calculated E a in this study are emperical activation energies.Fig. 7a illustrates the linear kinetically controlled region of the ln (TOF) - 1000  / T and the corresponding E a values obtained from the slope of the Arrhenius-type plot for each catalyst.
The calculated E a for Pt is consistent with previously reported values ranging from 7 to 46 kJ/mol [36,64].Additionally, Rinnemo et al. [65] demonstrated that the apparent activation energy of the CHC reaction over Pt wire can vary from 46.2 to 35.3 kJ/mol depending on the H 2 to O 2 ratio.However, it is essential to emphasize that the support material significantly influences the E a [66], which likely explains the lower E a observed in our study (31.7

kJ/mol).
A comparison of the E a of the other supported TM catalysts is challenging due to the lack of available data, underscoring the significance of the current study.
A higher activation energy implies a stronger dependency of the reaction rate on temperature, explaining the sharp reaction start-up observed with the 1Pt-γAl 2 O 3 catalyst.Whereas 10Mo-γAl 2 O 3 exhibits the lowest activation energy, it has the slowest reaction start-up, and the catalytic activity (X H2 ) increases more slowly compared to the other catalysts.
While these E a values represent observed activation energies, they cannot intuitively be associated with the light-off or T 50% conversion temperatures.However, it can be concluded that higher E a values lead to faster light-off temperatures and lower T 50% conversion temperatures (Table 1).
According to the Arrhenius kinetic model, the activation energy alone does not determine the reaction rate; the pre-exponential factor or frequency factor (A) also plays a significant role.Although it might be counter-intuitive that the lower activation energy does not necessarily resolve into a higher reaction rate, this observation has been made before for a different reaction by Mutschler et al. [67].
The observed E a in the case of 10Mo-γAl 2 O 3 (10.4kJ/mol) resembles the value calculated by Sha et al. [68] for H 2 adsorption on the MoO 3 (about 10 kJ/mol), indicating a high H 2 adsorption capacity.They showed that the hydrogen atoms, upon adsorption, form hydrogen molybdenum bronze (H x MoO 3 ) on the surface.Furthermore, Ikeda et al. [32] suggested that low E a can be attributed to the higher surface coverage.Thus, it can be concluded that the observed E a for the CHC reaction over the 10Mo-γAl 2 O 3 sample is associated with the chemisorbed hydrogen molecules.This is further supported by H 2 adsorption/desorption experiments (Fig. S8) and the highest M − H binding energy for Mo metal (Table S5).Consequently, Mo acts as a weak catalyst for the CHC reaction due to the strong bonding of adsorbed hydrogens to the Mo oxide surface.
The A value represents the reaction rate in the absence of E a or at infinitely high temperatures.A higher A value implies a higher probability of the reactant collisions that lead to the product formation at lower temperatures.Thus, the A value influences the reaction start-up temperature of the CHC reaction.Additionally, a higher A value ensures that the kinetic control region occurs at relatively lower temperatures.Consequently, as summarized in Table 3, since A value is significantly higher in 1Pt-γAl 2 O 3 compared to 10Mo-γAl 2 O 3 (Fig. 7b), the T L of 1Pt-γAl 2 O 3 is much lower than that of 10Mo-γAl 2 O 3 (Table 2).Moreover, if two catalysts have similar E a values, the one with a higher A value will initiate the CHC reaction at lower temperatures.This is the case for 10Co-γAl 2 O 3 and 10Ni-γAl 2 O 3 .
The relation between catalytic activity and the kinetic parameters can be summarized as (1) Higher E a will result in a sharper reaction start-up, and (2) Higher A will result in lower start-up temperature.

Conclusion
The current study contributes to evaluating and comparing the catalytic activity and kinetic parameters of different supported TMs in the CHC reaction.Well-dispersed NPs are synthesized through MIWI methods.XRD and XPS results showed that during the CHC reaction, the catalysts are oxidized, and only the Pt catalyst stayed unaffected.This competitive oxidation of the surface, along with the higher M − H binding energy of non-precious metals, is assumed to be a contributing factor to the diminished catalytic activity observed for the TM catalysts in the CHC reaction.While precious metals are state-of-the-art catalysts, the catalytic activity of 10Co-γAl 2 O 3 is found to be comparable to that of 1Ru-γAl 2 O 3 .This is a pivotal observation which signifies the importance of catalyst selection and design considerations for diverse elevated temperature applications.Both these catalysts showed stable CHC catalytic activity for 45 h at 240 • C.
Additionally, the adoption of a new high-throughput data acquisition strategy has significantly increased the number of kinetically relevant data points (36 points per minute).The kinetically controlled region for each catalyst, which is 20-30 • C wide, is determined.In this region, the kinetic parameters are calculated using the Arrhenius kinetic model and a relation between kinetic parameters and the catalytic activity is established.While the 1Pt-γAl 2 O 3 sample shows the highest E a (31.7 kJ/ mol), the 10Mo-γAl 2 O 3 catalyst has the lowest E a of 10.4 kJ/mol, which is the result of full coverage of Mo oxide surface by strongly adsorbed hydrogen species.This is further supported by the high H 2 uptake of the 10Mo-γAl 2 O 3 catalyst and the highest M − H binding energy for Mo.
In conclusion, we believe the results of this study and the adopted approach to study the kinetics of the CHC reaction will guide researchers in designing and modelling industrial-scale CHC reactors in various applications.With an understanding of the temperature demands across different applications, our research provides valuable direction in catalyst selection, catering to specific temperature ranges.Furthermore, it will enlighten the path to discovering potential new catalysts in this field in the future.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Fig. 3 .
Fig. 3. N 2 adsorption (solid shapes) and desorption (hollow shapes) of fresh catalysts, b) pore size and pore size distribution of fresh catalysts.

Fig. 4 .
Fig. 4. The catalytic activity of different catalysts under the CHC reaction.The total flow is 20 ml/min and 4 vol% of H 2 with H 2 :O 2 of 2:1.

Fig. 5 .Fig. 6 .
Fig. 5. a) CHC stability test of 1Ru-γAl 2 O 3 and 10Co-γAl 2 O 3 .The stability tests are done under 4 vol% H 2 with H 2 :O 2 equal to 2:1 at 240 • C for 45 h, b) Effect of regeneration on the CHC activity of 10Co-γAl 2 O 3 catalyst.I) the first 5 h of the stability test, II) the last 5 h of the stability test, and III) the 5 h activity test after the regeneration.

Fig. 7 .
Fig. 7. A) Arrhenius-type plot of TOF for the CHC reaction in the kinetic control region and calculated activation energies for each catalyst, b) pre-exponential factor, calculated from Arrhenius kinetic model for different catalysts.

Table 2 T
L , T 50% , T L -T 50% and T 100% of different catalysts.

Table 3
Kinetically controlled temperature region, E a and A values for each catalyst.