Novel high-pressure/high-temperature reactor cell for in situ and operando x-ray absorption spectroscopy studies of heterogeneous catalysts at synchrotron facilities

Heterogeneous catalysts play a crucial role in various environmental and industrial applications, including the manufacturing of chemicals, pharmaceuticals, petroleum products, pollution control, and in the energy sector.¹ High-pressure gas phase reactions are encountered in various fields of chemistry. For instance, in the field of heterogeneous catalysis, different environmental and industrial reactions, such as CO₂ methanation, dimethyl ether synthesis, and methanol synthesis reactions, are thermodynamically favorable under high-pressure conditions.² Furthermore, catalytic processes at high pressure often lead to enhanced selectivity, minimizing the formation of undesirable products.^3,4 In addition, the increase in pressure can influence reaction pathways and promote the formation of targeted products by suppressing secondary reactions.² In the case of CO₂ hydrogenation reaction, this improvement in selectivity is crucial, enabling the targeted production of chemicals, fuels, or other value-added materials from CO₂.⁵ Moreover, in addition to enhancing kinetics and selectivity, high-pressure catalytic processes can also improve the stability and durability of catalysts.⁶ The operation of metal catalysts under high pressure/temperature gaseous environments often leads to continuous modifications of their structural properties (crystalline structure, size, morphology, and composition), which dictate their catalytic performance. The fundamental understanding of the relationships between the structures and catalytic properties of catalysts at the molecular and atomic scale is crucial for identifying the nature of their active sites, understanding their activation and deactivation mechanisms, and ultimately, for designing more active catalysts.⁷ A detailed understanding of reaction mechanisms in heterogeneous catalysis requires a precise description of the geometry and local environment of active sites, which are critical for the catalytic performance (conversion, selectivity, and stability). X-ray absorption spectroscopy (XAS), including x-ray absorption near-edge structure (XANES) and extended x-ray absorption fine structure (EXAFS), is currently one of the most well-established characterization techniques for studying heterogeneous catalysts.^8–15 Indeed, XAS provides information on the detailed local electronic (oxidation state) and geometric (nature, number, and distance of surrounding atoms) structure of the absorbing atom of the catalyst.

In the late 1970s, at the beginning of the synchrotron era, Lytle et al. in collaboration with the Exxon research group made important contributions to the development of in situ XAS cell technology.^16,17 The design of the cell enabled the regulation of gaseous or vaporous atmospheric conditions during XAS measurements. The cell material, composed of boron nitride, was chosen to minimize the x-ray absorption by the cell and ensure a sufficiently low level of absorption while housing the sample. This preliminary design allows for XAS experiments to be conducted on powdered and pelletized samples under conditions of up to 700 °C and 100 bar in transmission and fluorescence mode. Despite the significant range of applicable reaction conditions, the design has a drawback in that temperature regulation occurs on the furnace rather than in proximity to the catalyst sample. Furthermore, the cell large volume and associated external and internal mass transport limitations render it unsuitable for operando studies.¹⁸ To avoid the issue of pore diffusion during sample measurements, a widely employed solution is the implementation of a highly versatile plug-flow cell design. This design uses a quartz glass capillary reactor that is heated by a hot air gas blower, thereby minimizing the impact of diffusion. The initial proposal of this idea was presented by Clausen et al.¹⁹ and Sankar et al.²⁰ Subsequent advancements were made by Grunwaldt et al.,^18,21 which facilitated the integration of XAS with scattering techniques, such as x-ray diffraction (XRD), as well as optical spectroscopies, including UV–vis and Raman spectroscopy.^21–26 Despite the advancements achieved in in situ/operando studies, this design still poses some limitations, specifically the difficulty of conducting reactions at high pressures (>5 bar) and achieving uniform temperature distribution throughout the catalyst bed.^27–29 

In recent studies, there has been a focus on developing innovative designs to simultaneously characterize the catalyst and evaluate its catalytic performance. One notable approach by Bare et al. involved the development of a plug-flow reactor using beryllium tubes.³⁰ The enhanced design has demonstrated favorable potential for conducting XAS experiments at an elevated temperature (600 °C) and pressure (14 bar). In addition, this design has effectively mitigated mass transfer issues and facilitated precise replication of catalytic properties (conversion and selectivity) achieved under controlled laboratory conditions. The modifications made to this setup by Fingland et al. using borosilicate glass or polyimide tubes as reactor materials, both significantly less expensive and easier to work with, have further enabled the acquisition of high-quality data and accurate rate measurements.³¹ However, the polyimide and borosilicate glass reactors possess a limited pressure rating, capable of accommodating only a few atmospheres. Moreover, polyimide tubing is constrained to temperatures below 400 °C and may deform even at lower temperatures. Indeed, at this temperature range, the authors reported the capillary deformation, resulting in its inclination with respect to the incident beam. As a result, the sample will no longer be exactly at the same position with respect to the beam, thus causing a signal modulation with temperature (the beam path in the sample is not constant) and so a decrease in the signal-to-noise ratio (even a complete loss of the signal if the sample is no more on the beam). This effect will directly impact the quality of XAS measurements. On the other hand, thin-walled borosilicate glass tubes are fragile and prone to cracking when compression fittings are used to create gas-tight seals, despite the use of PTFE ferrules. Three years later, the same research group proceeded to design a reactor by employing vitreous carbon as the material, aiming to increase its capability to withstand elevated temperatures and pressures for catalytic applications.³² Thanks to its various properties, such as inertness, robustness, and high x-ray transmission, the use of this reactor revealed an enhanced capability to withstand high temperature (up to 600 °C) and high pressure (up to 25 bar). Rochet et al. reported the development of a plug-flow cell utilizing a “sandwich-like” reactor design at the SAMBA beamline at SOLEIL synchrotron to facilitate pressures up to 50 bar at the same temperature (600 °C) during the XAS experiment.³³ This XAS cell can be used with either powder samples or pellets under different gas atmospheres. However, this system can only be used for transmission mode XAS, which means that information about elements at lower concentration cannot be examined. Another issue is that the horizontal flow direction can result in the accumulation of unwanted liquid or waxy substances in the catalytic bed, potentially leading to blockages in the reactor. Very recently, Pandit et al. have developed a high-pressure reactor for the measurement of the structure and catalytic performance of copper-based catalysts during the hydrogenation of CO₂ to methanol.³⁴ Compared to the previous study conducted by Rochet et al.,³³ this new reactor shares a “sandwich design” similar to the one described in their study, but it offers certain advantages. It is capable of operating not only in transmission mode but also in fluorescence XAS mode and at high pressure of up to 50 bar. However, the operating temperature is limited to 450 °C. Furthermore, this reactor is applicable to other techniques, such as small-angle x-ray scattering (SAXS) and x-ray diffraction (XRD). Despite the inherent toxicity of beryllium and the associated handling challenges, the authors opted to use it as the reactor material due to its robustness and adequate x-ray transmission capabilities. These advancements have played a crucial role in the progression of catalytic reaction studies and the understanding of material behavior within these reactions.^34–38 They have facilitated direct observation of the chemical and structural transformations taking place in materials during catalysis, particularly for high-pressure (up to 50 bar) reactions covering a wide temperature range (up to 600 °C). According to the various studies and developments described previously, achieving temperatures exceeding 800 °C and preventing temperature gradients on the catalyst bed poses a challenge. In order to address these challenges, our research group has recently made significant progress by developing a reaction cell using a glassy carbon reactor that can operate at elevated temperatures up to 1000 °C.³⁹ This innovative design ensures uniform heating of the sample region, thereby overcoming the mentioned difficulties. Various in situ/operando XAS studies of heterogeneous catalysts in a gaseous environment have further confirmed the significant progress achieved through the development of this setup.^40–45 However, this reaction cell is limited to catalytic reactions occurring under atmospheric pressure conditions.

Direct hydrogenation of CO₂ to methanol, using pure H₂ and CO₂ as reactants, represents a highly promising approach for CO₂ utilization, particularly when green hydrogen becomes accessible.^46–50 Furthermore, methanol has combined significant interest among the potential CO₂-derived products due to its applicability not only as an alternative fuel (e.g., in the production of dimethyl ether) but also as a convenient chemical feedstock (e.g., for the synthesis of formaldehyde and acetic acid).^51–55 In a study conducted by Bavykina et al.,⁵⁶ the authors prepared a novel and highly selective catalyst based on a combination of cobalt metal (Co) and oxidized indium (In). Separately, Co and oxidized In would produce methane and CO, respectively. However, the combination of these two elements showed a synergetic effect, leading to remarkable catalytic performances in terms of activity, selectivity, and stability during CO₂ hydrogenation reaction at 300 °C and 50 bar. More precisely, the InCo catalyst displayed significant selectivity toward methanol (69%) compared to CO (23%) and methane (8%), and it exhibited remarkable stability, maintaining around 19% CO₂ conversion rate over a duration of 100 h at 300 °C and 50 bar⁵⁶ In order to corroborate the catalytic results of the InCo catalysts obtained during this study, we were motivated to develop a high-pressure/high-temperature cell to monitor the catalyst structures under gaseous environment, high pressure, and temperature, using the XAS technique.

In the present work, we report a novel high-pressure/high-temperature reaction cell designed to work in both transmission and fluorescence detection modes (for the characterization of samples containing low element concentration). The objective is to integrate the high-temperature catalytic cell³⁹ into a high-pressure autoclave, which was also developed by our research team in 2005 for reproducing the experimental conditions associated with geological fluids using x-ray absorption and Raman spectroscopies.⁵⁷ For the catalytic application, we were inspired by the results reported by Bavykina et al.,⁵⁶ as an initial case study, to showcase the application of the novel high-pressure/temperature reactor setup in standard operando characterization within the field of heterogeneous catalysis. In this context, we conducted operando XAS measurements at the Co K-edge for an InCo catalyst during the CO₂ hydrogenation to methanol reaction at 300 °C and 50 bar. The high-pressure/high-temperature reaction cell is completed with an EcoCat-P portable mass spectrometer system provided by ESS. The mass spectrometer was positioned at the reactor outlet to enable the measurement of the composition of the gas products. This is equipped with two identical capillary inlets allowing the analysis of gas composition at both reactor outlets. The system has the capability to monitor up to 64 species in real time and offers detection levels down to ppb. This in situ/operando setup is available for users at the CRG-FAME and FAME-UHD (BM30 and BM16) beamlines at the European Synchrotron Radiation Facility (ESRF).

The XAS reaction cell was designed to perform in situ and operando gas-phase experiments for heterogeneous catalysis applications. In our case, the samples, in the form of powder, necessitate gas flow, thermal treatment, and pressure regulation to reproduce the actual conditions implemented in a conventional laboratory microreactor. The optimization of our setup offers a safer handling of the device in comparison to systems using gas blower heaters. The heating part of our cell is isolated from the ambient so the body of the cell remains cold, which allows a safe handling and the possibility to approach detectors near the cell, which is important to improve the signal detection by fluorescence in diluted samples. Gas delivery is ensured by a fully automated and transportable gas distribution system. The gas distribution system includes a set of high-pressure mass flow controllers to precisely deliver the gas flow to the reaction cell. Notably, this setup is designed to be user-friendly, modular, and portable. Figure 1(a) shows the three-dimensional configuration of the novel high-pressure/temperature reactor, which draws inspiration from the previous designs outlined in our last high-temperature reaction cell development.³⁹ The reactor body is constructed entirely from X13 stainless steel material, chosen for its non-oxidizing properties and high strength, enabling it to withstand the specified pressures and temperatures. The glassy carbon tube is positioned within the heating elements. Concerning the sample heating, the previous design (high-temperature reactor cell) consists of placing the sample inside the reactor, which, in turn, is placed in a 30 mm molybdenum tube that allows heating of the sample through the surrounding resistive molybdenum metallic wires.³⁹ This configuration uses a pair of thermocouples: (i) one housed in a hole drilled in the molybdenum tube to control the temperature of the oven and (ii) another located inside the reactor in close proximity to the sample to monitor and control the sample temperature. Continuous external flow of ultra-pure He serves the dual object of protecting the carbon tube and molybdenum resistance against oxidation, particularly at high temperatures. The design of the current reactor heating system is based on the same design. However, this heating system includes just one thermocouple, inserted in a hole drilled inside the molybdenum tube for precise temperature control of the furnace (see more details in Sec. Ⅱ B). Furthermore, the window materials of the new high-pressure/high-temperature reactor cell have changed compared to the old design. While the old design used Kapton windows (25 µm thick and 19 mm in diameter) to ensure high transmittance to x rays, the current design utilizes beryllium windows (Be SR200) with low absorption characteristics as well, but most importantly, excellent mechanical properties to ensure resistance of the cell during high-pressure operation (see more details in Sec. Ⅱ C). Figure 1(b) shows a photograph of the fully assembled reaction cell at the SAMBA beamline in SOLEIL synchrotron. The image showcases one of the two transmission windows and the fluorescence window.

Gases flow through the catalytic bed from the top to bottom through an innovative plug-flow reactor made of the glassy carbon tube [Fig. 1(c)] previously used in a high-temperature dedicated reaction cell developed in our research group.³⁹ The glassy carbon tubes were selected due to their distinct characteristics. They exhibit non-porosity and exceptional resistance to elevated temperatures, withstanding up to 3000 °C in an inert gas or vacuum environment and 600 °C in the presence of an oxidizing atmosphere.⁵⁸ The tube was machining in the window region, reducing its original wall thickness from 1 mm to 250 µm [Fig. 1(c)], thereby enabling effective transmission of the majority of the x-ray beam to the sample, with ∼83% of the beam transmitted to the sample at the Co K-edge. This value demonstrates excellent suitability in comparison to a standard quartz capillary, which enables 92% and 84% transmission for wall thicknesses of 10 and 20 µm, respectively, at the same energy level.

The reactor is connected to the gas inlet and outlet tubes by bellows in the internal part of the cell, and the sealing is assured by using Viton joints. The design comprises a dedicated metallic rod used as a support for holding the quartz wool of the catalytic bed. The reaction cell can be fully dismounted to recover the glassy carbon tube. The glassy carbon reactor is filled by placing it on a support designed for this purpose (Fig. S1 in the supplementary material). The reactor is filled in the upper part with the quartz wool and the powder sample, and the powder is gently pressed. Then, the carbon tube containing the catalyst is placed inside the reaction cell and placed together in the sample stage of the beamline. At the end of the test, the cell is demounted, the sample is recovered, and the carbon tube is cleaned with ethanol. Thereafter, it is ready to be refilled with a new sample for a new experiment. The dismounting, cleaning, and remounting process takes around 20 min.

The design of the current reactor heating system is based on the same design used for the development of our high-temperature reactor in 2018.³⁹ The heating element design was derived from the work of Tamura et al.⁵⁹ It involves positioning the reactor inside a 30 mm molybdenum tube [8 in Fig. 1(a)], which is heated to the desired experimental temperature using a resistive molybdenum metallic wire. In contrast to the previous system,³⁹ this heating system includes just one K-type thermocouple, inserted in a hole drilled inside the molybdenum tube to measure the temperature of our heating system. To prevent heat loss, all these components are enclosed in insulating ceramic elements. The cell design was based on a previously validated and certified catalytic reactor and autoclave development at the Néel Institute in Grenoble, and enables operation over a wide temperature range, from room temperature up to 1000 °C.^39,57 The cell is pressurized with ultra-pure He flow, which also prevents oxidation of the carbon tube and molybdenum resistance. The temperature control was achieved using a PID controller. After tuning the control system parameters, the reactor temperature could be smoothly ramped up from room temperature (RT) to the desired set point, employing a desired heating rate ranging from 2 to 20 °C/min, with minimal deviation during the ramping process and a stable control of the temperature at the set point with a deviation of ±0.66 °C.

Figure 2 shows a 3D drawing of the window configuration. The high-pressure/temperature cell was designed with three windows: two positioned along the beam axis, one for entrance and the other for exit of the beam for measurements using transmission mode, and an additional one at 90° for working at fluorescence or inelastic scattering photons mode. One of the advantages of this cell is its ability to work in fluorescence mode, enabling the characterization of samples with very low concentrations of the elements of interest. Among the most versatile windows, suitable for the XAS technique, we use Beryllium as a material. Specifically, extruded Be SR200 was chosen for its favorable balance between excellent mechanical properties and low absorption characteristics. Window thickness is calculated as a function of working pressure. We use 0.8 mm Be windows for a working pressure of up to 300 bar, and then, we use 1 mm Be windows at a working pressure from 300 to 1000 bar. The high-pressure cell is cooled by the circulation of water to protect the windows from heating.

Because the sample is positioned at the center of the heating zone, the heating element was designed with three apertures with a diameter of 5 mm (22° wide aperture). These apertures allow the transmission of x-ray photons. This configuration additionally facilitates the vertical measurements of chemical inhomogeneity within the catalytic bed (∼4 mm). This is achievable as the FAME beamline provides a beam spot of 200 × 100 µm² (HxV, FWHM), which can be moved along the reactor vertical axis. This is of particular importance when the reactor is operated under conditions of high reactant conversion. In such scenarios, the gaseous composition through the catalytic bed diverges substantially from that at the reactor inlet.

The reaction setup includes a completely automated high-pressure gas distribution system (Fig. S2 in the supplementary material) from Serv’Instrumentation to deliver up to four different gases. Each gas line is regulated by a high-pressure mass flow controller (in a range of 20–120 bar). Despite the limitation of our gas distribution system to introduce a maximum pressure of 120 bar, our high-pressure cell complies with European regulations, meeting the requirements of the Pressure Equipment Directive DESP 2014/68/UE, and has been certified to operate at pressures of up to 1000 bar, depending on the gas environment (see more details in the S3 section of the supplementary material).

The pressure in the cell increases as the reaction mixture flows through the glassy carbon reactor; simultaneously, helium flows through the cell body at the working pressure. The introduction of an inert gas, such as helium, allows for balancing the pressure throughout the system, followed by the extraction of the gaseous products formed at the reactor outlet. A pressure transmitter model Rosemount 2051, enables the equilibrium of the internal pressure between the glassy carbon tube and the body of the cell. The mixture of the products of the reaction and the He gas occurs immediately prior to the cell outlet [9 in Fig. 1(a)]. Precise pressure regulation is achieved at the reaction cell exit using the Brooks model SLA5820 pressure regulator. Furthermore, this regulator ensures release of the reaction mixture at atmospheric pressure. Finally, the gas distribution setup is provided with a liquid trap in order to protect the pressure regulator and the mass spectrometer.

The desired pressure set point can be manually adjusted through a control software, developed by Serv’Instrumentation, according to the specific requirements of the reaction conditions.

The main drawback of our setup is the lack of the temperature measurement near the sample. To overcome this issue, a temperature profile test was conducted to assess the longitudinal temperature distribution within the 30 mm oven [8 in Fig. 1(a)]. The test was performed by flowing pure He through the cell and a second test by flowing the CO₂/H₂ reaction mixture, using different flow rates (10, 50, and 100 ml/min) in both cases. The test was performed maintaining a set point of 300 °C at atmospheric pressure. Figure 3 shows the temperature profile across the oven under varying He (A) and CO₂/H₂ reaction mixture (B) flow rates.

The shaded area in the figure (positioned at 12–17 mm along the oven) corresponds to the sample position. Remarkably, within the initial 5 mm of the oven, rapid heating of the gas occurs, in both cases [Figs. 3(a) and 3(b)], resulting in a homogeneous and stable temperature profile along the catalytic bed. This result is significant because it shows that the temperature remains constant and homogeneous regardless of the gas composition. The uniform temperature profile achieved within the reaction cell oven plays a crucial role in the study of catalysts under operando conditions. This feature stands as a distinct advantage compared to gas blower heaters,³⁰ in which temperature gradients may introduce imprecision in the measurement of the catalytic behavior. As previously discussed, gas blower heaters do not allow for homogeneous heating of the catalyst. In addition, the positioning of the gas blower with a capillary tube and the flow of the heating stream directly impact the sample temperature,³⁰ leading to a significant drawback when studying catalysts under operando conditions.

The operando XAS measurements were performed at the SAMBA (spectroscopy applied to material based on absorption) beamline at the French National Synchrotron Facility (SOLEIL), Saint Aubin, France. The details of the beamline and its optical devices have been described by Belin et al.⁶⁰ The beamline was equipped with a double-crystal Si 220 monochromator. The presence of two Pd-coated mirrors ensured the rejection of the harmonics to better than 0.1%, i.e., the intensity of the harmonics was less than 0.1% of the intensity of the fundamental x rays. The incident (I₀) and transmitted (I₁) intensities were measured by using ionization chambers. The measurements were carried out at the Co K-edge using fluorescence mode, and data acquisition employed a Canberra 35-element monolithic planar Ge pixel array detector.

The operando XAS characterization was made following the next protocol: a total amount of 100 mg InCo catalyst (sieved to a particle size range of 200–250 µm) was packed into the reactor to serve as the catalytic bed. The InCo catalyst was prepared following the same preparation method used by Bavykina et al.⁵⁶ Co K-edge XAS spectra were acquired under the helium atmosphere at room temperature to assess the initial state of the catalyst. Subsequently, the reactor was gradually heated up to 300 °C at a rate of 10 °C/min and then pressurized to 50 bars under He. XANES spectra were continuously recorded during the thermal treatment to monitor the change in the metal oxidation state. After thermal treatment under the He atmosphere, 50 ml/min of the reaction mixture comprising 20% CO₂, 79% H₂, and 1% Ar, diluted with 150 ml/min of He, was introduced. The reaction was stopped after 24 h, and then, EXAFS spectra were recorded to characterize the catalyst’s final state. The XANES and EXAFS data were processed using the ATHENA program from the HORAE software package.⁶¹ 

To extract information about the proportion of the Cobalt species in the InCo catalyst, we applied linear combination fitting (LCF) analysis to the Co K-edge XANES spectra. The LCF method involves modeling the XAS spectra by least-squares fitting using a linear combination of known species to adjust an unknown spectra.⁶² Indeed, this analysis method will determine the proportions of spectra (containing multiple species) for the selected reference compounds, which, when summed, yield the least-squares fit to the spectrum of an unknown sample.⁶³ The principle of this method is based on the additive nature of the absorption of each species in the sample. LCF analysis can be applied to both XANES and EXAFS spectra. The relative weights of the fractional components are allowed to vary from 0 to 1, and the resultant sum of all components is 1. In the present paper, LCF analysis was performed using Athena software and has been carried out in the normalized XAS spectra using the reference spectra from −30 to 80 eV.

The objective of our study was to gain insights into the chemical state of cobalt in a methanol producer catalyst, specifically the InCo catalyst, during the induction phase of the reaction. To achieve this, catalyst characterization was conducted under high-pressure reaction conditions at the SAMBA beamline in the synchrotron SOLEIL, under real conditions for the CO₂ hydrogenation reaction at 300 °C and 50 bar.

In the present study, the LCF analysis method was applied to determine the proportions of different cobalt phases based on the near-edge spectra region spanning from 7690 to 7790 eV, acquired during the thermal treatment under the He atmosphere. We primarily relied on detailed structural analysis of the same InCo samples provided by Bavykina et al.⁵⁶ to select appropriate references (Co₃O₄, CoO, and Co metal) for LCF adjustments. Figure 4 shows the Co K-edge XANES spectra of the InCo catalyst at the initial state (20 °C) after heating under He to 300 °C and then after pressurization to 50 bar at 300 °C. The evolution of the XANES spectra during temperature and pressure ramping is shown in Fig. S4 in the supplementary material. In addition, Fig. 4(b) shows the corresponding evolution of the phase amounts of different cobalt species. LCF analysis of various cobalt references reveals the presence of CoO and Co₃O₄ species predominantly at the initial stage of treatment. As the temperature rises, there is a partial reduction in cobaltite, leading to the emergence of Co metal species. The phase amount of the Co metal remains nearly constant after reaching 150 °C. After thermal treatment, the main detected species by LCF include CoO (∼64%), Co metal (∼25%), and Co₃O₄ (∼11%).

Following the thermal treatment under He, the reaction mixture was introduced into the reactor cell at 300 °C and 50 bar, and the reaction was monitored during 24 h. Figure 5(a) shows the temporal evolution of CO₂ conversion using the high-pressure/high-temperature reaction cell. The conversion remains relatively constant at around 16% throughout the duration of the reaction. These results closely reproduce those obtained in a catalytic test conducted in a laboratory microreactor⁵⁶ (see the experimental details in Sec. S5 in the supplementary material), wherein a stable CO₂ conversion at around 13% is shown [Fig. 5(b)]. This alignment validates the performance of our high-pressure/high-temperature reaction cell for operando characterization.

The InCo catalyst was characterized at the final state after 24 h of reaction. The XAS spectra were collected at 300 °C and 50 bar under the reaction mixture. The structural characterization of InCo catalysts under the same conditions by Bavykina et al.⁵⁶ demonstrated that the catalyst surface state is modified after the reaction. Specifically, surface characterization by high-resolution XPS revealed that this modification was induced by the transformation of a part of the cobalt oxide initially present in metallic/carbides. The formation of carbides was further confirmed by elemental analyses, which showed an increase in carbon composition from 13% before the reaction to 24% after the reaction.⁵⁶ Based on these results, metallic Co, CoO, and CoInC references were selected for the LCF analyses applied to the XANES spectra after 24 h of CO₂ hydrogenation. LCF of the spectra at the final state indicates a complete disappearance of Co₃O₄ species. The evolved material comprises CoO (∼41%) and Co metal (∼46%) phases, alongside a residual presence of InCo carbide (∼13%) species [Fig. 5(c)]. These findings align with the results obtained through ADF-STEM imaging and EELS-based elemental mapping of an identical sample after CO₂ hydrogenation, performed by Bavykina et al.⁵⁶ Consequently, these results demonstrate the reduction in Co₃O₄ oxide to metallic cobalt and carbide during the reaction.

A further qualitative examination of the EXAFS oscillation and Fourier transforms (FTs) at the initial state, after thermal treatment under He at 300 °C and 50 bar and after CO₂ hydrogenation reaction was carried out to have a clearer idea of the nature of the Co species present in the catalyst at different stages of the catalytic cycle. The k²-weighted EXAFS spectra are shown in Fig. 6(a), and the corresponding pseudo-radial distribution functions of the EXAFS spectra (FT-EXAFS) of the catalyst comparing to those of different references (Co foil, CoO, and Co₃O₄) are showed in Fig. 6(b).

FT-EXAFS spectra of the Co foil exhibit one major feature around 2.2 Å corresponding to the Co–Co metallic bond contribution. However, the Co₃O₄ reference displayed three major signals at 1.5, 2.5, and 3 Å, identified as Co–O, octahedral Co–Co, and tetrahedral Co–Co bond contributions, respectively.^63–65 At the initial state, the cobalt in the InCo as-prepared sample predominantly exhibited a Co₃O₄ structure. This is evident from the appearance of two distinct peaks at 1.5 Å (Co–O) and 2.5 Å (octahedral Co–Co), indicative of its crystalline arrangement (the black spectra shown in Fig. 6(b). This observation aligns with the results of the LCF analysis [Fig. 5(c)], which show that the InCo catalyst in its initial state is predominantly composed of 1.5 Å Co₃O₄ species (more than 50%). Then, after thermal treatment at 300 °C and 50 bar under He, the sample retains its oxide state as indicated by the Co–O peak still present at 1.5 Å. During CO₂ hydrogenation reaction at 300 °C and 50 bar, we observe a decrease in the intensity of the peak at 1.5 Å corresponding to the Co–O bond [the blue spectra shown in Fig. 6(b)]. In parallel, a new peak emerges around 2.2 Å, corresponding to the Co–Co bond observed in the metallic cobalt foil. This transformation aligns with the LCF analyses and confirms the partial reduction of Co₃O₄ oxide to metallic cobalt during the reaction. This result may indicate that the active cobalt phase in the bimetallic InCo catalyst in the CO₂ hydrogenation reaction is present as a mixture of cobalt oxide and metallic cobalt. After the reaction (the green spectra shown in Fig. 6(b), the InCo catalyst retains the same cobalt species (mixture of oxide and metal) present during the reaction. These findings, in conjunction with the catalytic results obtained by mass spectrometry, demonstrate that our new high-pressure/temperature reactor is capable of performing operando characterizations of heterogeneous catalysts under operational (high temperature/pressure) conditions.

In summary, our research has led to the development of a high-pressure/high-temperature reactor cell dedicated for catalyst characterization using synchrotron x-ray absorption spectroscopy. The conception of the cell design enables operation at temperatures of up to 1000 °C and pressures of up to 1000 bar, depending on the gas environment. Configured for both XAS transmission and fluorescence modes, it facilitates analysis of samples containing low element concentration. In addition, the design ensures precise sample heating control, maintaining a stable and uniform temperature profile throughout the catalytic bed. Inspired by the work of Bavykina et al.,⁵⁶ our study highlights the practical application of the newly developed high-pressure/high-temperature reaction cell. The operando XAS measurements at the Co K-edge offer valuable insights into the behavior of the InCo catalyst during CO₂ hydrogenation at 300 °C and 50 bar, enhancing our understanding of its reactivity under operational conditions. On the one hand, the achieved results demonstrate the efficiency of the cell in replicating catalytic performances observed in the laboratory and, on the other hand, the potential for obtaining high-quality XAS spectra in complex conditions, providing an understanding of catalyst behavior in real-conditions.

The supplementary material contains more technical details regarding sample preparation within the carbon reactor, laboratory catalytic tests, and, notably, the certification of pressure tests conducted using the high-pressure/high-temperature cell.

This work was supported by the French National Research Agency through the PEPR PIA4 project DIADEM “ESRF.” We acknowledge SOLEIL for provision of the synchrotron radiation facilities, and we thank Dr. Emiliano Fonda and Dr. Andrea Zitolo for their assistance in using beamline SAMBA. The authors thank Dr. Anastasiya Bavykina for her contributions to the laboratory microreactor catalytic testing of CO₂ hydrogenation conducted at KAUST.

The authors have no conflicts to disclose.

Abdallah Nassereddine: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Writing – original draft (equal). Alain Prat: Methodology (equal); Writing – review & editing (equal). Samy Ould-Chikh: Conceptualization (equal); Data curation (equal); Methodology (equal); Visualization (equal); Writing – review & editing (equal). Eric Lahera: Methodology (equal). Olivier Proux: Conceptualization (equal); Data curation (equal); Methodology (equal); Visualization (equal); Writing – review & editing (equal). William Delnet: Methodology (equal). Anael Costes: Funding acquisition (equal). Isabelle Maurin: Writing – review & editing (equal). Isabelle Kieffer: Methodology (equal). Sophie Min: Methodology (equal). Mauro Rovezzi: Methodology (equal). Denis Testemale: Methodology (equal). Jose Luis Cerrillo Olmo: Data curation (equal). Jorge Gascon: Supervision (equal); Visualization (equal). Jean-Louis Hazemann: Methodology (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – review & editing (equal). Antonio Aguilar Tapia: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.