Pivotal role of organic adsorbates for the creation of catalytic sites during dry reforming of methane

Abstract

Inadvertent factors can sometimes be crucial for synthesis of catalysts. The use of polyalcohols is common in the synthesis of heterogeneous catalysts. Interactions between alcohols and heterogeneous catalysts have been shown to induce surface reconstructions that greatly impact catalytic performance. Thus, traces of these alcohol functionalities on the as-synthesized catalysts, combined with heat treatment, could be critical in the generation of catalytic sites. Here, we show that during the synthesis of a Ni–Mo/MgO catalyst using a polyol process, residual ethylene glycol (EG) on the surface plays a significant role in the generation of catalytic sites for dry reforming of methane (DRM). The as-synthesized catalyst presents dispersed cationic Ni. Under DRM reaction conditions, the presence of EG, and H₂ generated in situ, promote the generation of co-localized Ni–Mo nanoparticles (NPs). Greater amount of EG in the as-synthesized catalyst prevented sintering, leading to better catalyst stability and higher rates. If the residual EG remaining post-synthesis is removed through calcination, before conducting DRM, NiO NPs are formed and the material is completely inactive for catalyzing the reaction. When using a different support, denoted MgO*, EG also proved indispensable to generate active sites, although Ni–Mo co-localization was not evident, and a combination of DRM-related species was needed to activate the catalyst, not just H₂. This work systematically uncovers how the interactions between organic adsorbates, the supported metals and the catalyst support dictate the creation of catalytic active sites.

---

Broader context

This work underscores that even seemingly minor catalyst synthesis factors, such as residual organic compounds, can fundamentally alter a catalyst's active structure and performance. In energy applications like the dry reforming of methane (DRM), the formation of catalytic active sites is key to efficiently convert greenhouse gases into valuable fuels or chemical feedstocks, and this study reveals that trace amounts of organics remaining from synthesis, interacting with gas species present during DRM, can promote the generation of catalytic active sites. In environmental catalysis, where modifying reaction pathways to mitigate emissions is crucial, the unexpected influence of residual adsorbates suggests that careful control over catalyst synthesis can tailor surface chemistry to enable more sustainable and efficient chemical processes. Overall, these findings challenge traditional synthesis-characterization paradigms by demonstrating that small, often overlooked factors may be harnessed to optimize performance, opening new avenues to design catalysts that contribute significantly to energy efficiency and environmental sustainability.

Introduction

The polyol process is a widely used method for synthesizing nanoparticles (NPs) or clusters in polyalcohol media.^1,2 The high boiling point of polyalcohols enables synthesis of materials in a wide temperature range (∼80 to 300 °C). In this process, polyols can occupy the surface of as-synthesized NPs to prevent further aggregation. This method enables the synthesis of NPs of noble metals (e.g., Au, Ag, Pt, Pd) and transition metals (e.g., Ni, Fe, Cu, Co) with tunable properties, such as size, shape, and composition, which can directly influence catalytic performance.³ In our previous work, we discovered that the interaction of an alcohol with supported-metal catalysts can induce migration of the support over the supported NPs, even at temperatures below 200 °C.⁴ This catalyst reconstruction significantly impacted catalytic performance. It has also been shown that bulk metal oxides can spread onto surfaces of oxide supports upon thermal treatment with alcohols.⁵ These reports call attention to the alcohol-induced formation/reconstruction of active sites in heterogeneous catalysts. Common usage of alcohol media during catalyst synthesis, followed by a heat treatment, led us to investigate the influence of these alcohol residues on shaping the catalytic sites.

As a probe reaction, we study dry reforming of methane (DRM). DRM is a reaction of paramount importance since it converts two greenhouse gases, methane and carbon dioxide, into syngas, a mixture of hydrogen and carbon monoxide. Syngas is a platform mixture for the production of a variety of chemicals for mass consumption. Despite holding the promise of converting greenhouse gases into useful chemicals, DRM is currently not industrially adopted because catalysts suffer from deactivation through coking and sintering due to the elevated operating temperatures (>600 °C), demanded by the high endothermicity of the reaction (ΔH_298 K = 247 kJ mol⁻¹).⁶ Ni-based catalysts have been largely studied for DRM due to their promising performance and the low cost of Ni. In many cases, alcohols are involved in some step of the catalyst synthesis procedures.^7–15

Here, we focus on unveiling the impact of alcohol-induced formation of active sites on a Ni–Mo/MgO catalyst for DRM. This particular catalyst composition has shown outstanding performance for DRM and inhibition of coke.^10,16 Previous studies have attributed DRM activity in bimetallic Ni–Mo catalysts primarily to metallic Ni species, while Mo is proposed to stabilize Ni⁰ and promote carbon removal.¹⁷ We use a polyol process to load Ni and Mo species onto an MgO support in ethylene glycol (EG) media with polyvinylpyrrolidone (PVP) as the stabilizing agent. The amount of organics (EG and PVP) remaining on the catalyst post-synthesis was carefully tuned, and the bonding and electronic structure of the metal sites post-synthesis and post-DRM were characterized. The residual organics on the catalysts had a dramatic impact on catalytic conversion and catalyst stability. In particular, our study reveals that EG greatly influenced the formation of active sites on a Ni–Mo/MgO catalyst as evidenced via temperature-programmed characterization, spectroscopy and scattering techniques. EG and H₂, generated in situ during DRM, resulted in co-localized Ni–Mo sites active for DRM. The greater the amount of residual EG post-synthesis, the more active and stable the catalyst during DRM due to stabilization of Ni NPs of smaller size. To evaluate the broader applicability of our findings, the EG-assisted synthesis was repeated using another MgO support (denoted MgO*). EG was proven indispensable to create catalytic active sites, although evidence of Ni–Mo co-localization was not found, and a combination of DRM-related species was needed to activate the catalyst, not just H₂.

Calcination of the catalyst prior to reaction, burning away organics remaining post-synthesis, rendered catalysts inactive for DRM. The alcohol-induced formation of active sites in the Ni–Mo/MgO (or Ni–Mo/MgO*) catalyst turns an inactive catalyst into an active one. This work proves the pivotal role of residual organic species in creating active catalysts for important catalytic conversions such as DRM, and unveils yet another layer of precision catalyst design.

Results and discussion

Significant organic residues on the surface of as-synthesized Ni–Mo/MgO catalysts

The Ni–Mo/MgO catalysts were synthesized via a polyol process as previously reported.^10,18 Nickel chloride and ammonium molybdate hydrates were dissolved in EG, mixed with MgO support and PVP. The mixture was then heated to 80 °C for 1 h, washed with a solution of ethanol and EG, and dried at room temperature overnight. The organic residue content was controlled by varying the number of washing cycles from 3 to 9. Scanning electron microscopy (SEM) images revealed that the Ni–Mo/MgO samples consisted of aggregates of microplates with a uniform size of approximately 2–5 µm (Fig. S1). Notably, the SEM images of Ni–Mo/MgO washed 9 times with EG + ethanol displayed greater charging than the sample washed 3 times, suggesting a higher concentration of organic residues with additional washing steps. Thermogravimetric analysis (TGA) was conducted to quantify the organic residues by measuring weight loss relative to the pristine MgO support (Fig. 1A). The weight-based content of organics for the four catalyst samples was 8 wt%, 12 wt%, 14 wt%, 38 wt%, corresponding to 3, 5, 7, and 9 washes with EG and ethanol. When 9 washes were performed, the catalyst was soaked in EG and ethanol for extra ∼3 min. Additionally, the TGA curves for the four samples exhibited three distinct weight loss stages: 50–153 °C, 153–260 °C, and 260–400 °C. As references, TGA measurements were performed for pristine MgO, PVP, and EG-impregnated MgO (“EG on MgO”, 23 wt% EG). Neither the MgO support nor PVP displayed a significant weight loss in the 50–153 °C region, which was present in the Ni–Mo/MgO samples and in the “EG on MgO” sample, suggesting that the diol solvent is the main residual organic component on the surface after synthesis. This phenomenon can be attributed to electrostatic interactions between Mg²⁺ ions and organic solvents, which facilitates the adsorption of EG during synthesis and washing.¹⁹ Temperature-programmed oxidation (TPO) combined with online mass spectrometry (MS) was employed to further characterize the organic species remaining on the surface. As shown in Fig. 1B, the TPO profiles for all four samples exhibited similar characteristics, with an onset temperature at 67 °C and a peak at 328 °C. Comparison with the TPO profile of “EG on MgO” and pure PVP supports that EG is the main component left on the surface after synthesis. However, a shoulder peak at around 500 °C in the TPO profile of the Ni–Mo/MgO samples suggested that a trace amount of PVP might be present on the surface. To characterize the organic residues, solid-state ¹³C Nuclear Magnetic Resonance (NMR) spectroscopy and Raman spectroscopy were conducted on the catalyst samples. NMR of the as-synthesized Ni–Mo/MgO 38% (Fig. S1C) showed the characteristic peaks at 17 ppm (–CH₂–) and 184 ppm (C=O), confirming the presence of PVP on the surface. A peak at 62 ppm, corresponding to –CH₂– units of EG, confirmed the presence of EG on the surface as well. As depicted in Fig. S2, characteristic Raman shifts for EG and PVP were observed around 920 cm⁻¹ and 1620 cm⁻¹ in all Ni–Mo/MgO samples, respectively. In summary, the polyol synthesis process left a significant amount of EG and PVP on as-synthesized samples. By varying the number of EG + ethanol washes at the end of the synthesis, the amount of residual surface organic species was tuned, with multiple washes specifically enriching the EG content. In the following section, we show that the amount of organics in the surface dramatically impacted catalytic performance.

Fig. 1 TGA and TPO suggest EG is the predominant organic on the catalyst samples. (A) TGA and (B) TPO of as-synthesized Ni–Mo/MgO samples. The m/z =44 signal for PVP and EG on MgO were multiplied by 5 for comparison.

EG and H₂ drive the generation of catalytic active sites

DRM was performed at 650 °C using a gas mixture of 8.3% CH₄, 8.3% CO₂, balanced with Ar at a flow rate of 60 cm³ min⁻¹, without any catalyst pretreatment. To capture the catalyst deactivation, the reaction was maintained below equilibrium conversion. Conversion of reactants and catalyst stability (Fig. 2) varied substantially with the organic content of catalysts. Notably, the catalyst with the highest concentration of organics (Ni–Mo/MgO 38%) exhibited the highest conversion and the most stable performance. The site time yield (STY) normalized to Ni loading (Table S1) further highlighted the strong influence of organic residues on catalytic performance (Fig. S3). The Ni–Mo/MgO 38% sample demonstrated a ∼5-fold higher STY compared to Ni–Mo/MgO 8%. Normalized CH₄ conversion data (Fig. 2C) showed a clear stability trend: Ni–Mo/MgO 38% > Ni–Mo/MgO 12% ∼ Ni–Mo/MgO 14% > Ni–Mo/MgO 8%, suggesting that the organic residues also contributed to enhanced catalyst stability. The trend in stability was independent of the maximum conversion level at early (<1 h) TOS (Fig. S4).

Fig. 2 Organic content greatly affects conversion and stability during DRM at 650 °C. Conversion of (A) CH₄, (B) CO₂ and (C) normalized CH₄ conversion over 5 mg Ni–Mo/MgO samples at 650 °C with 60 cm³ min⁻¹ 8.3% CH₄, 8.3% CO₂ balanced by Ar. The catalyst bed temperature was ramped up from room temperature to the reaction temperature under reactant flow without pretreatment. The dashed lines are used to guide the eyes.

To further confirm the role of organic residues, the Ni–Mo/MgO catalysts were pretreated in air at 550 °C for 1 h to remove EG and PVP. As shown in Fig. S5, this pre-calcination step rendered the catalysts inactive, confirming that attempting to remove the residual organic species via calcination was detrimental for the generation of catalytic sites.

To further understand the deactivation behavior of the samples, we performed temperature-programmed oxidation (TPO) of the spent samples. Organics burned off the surface are a combination of remaining PVP- and EG-derived species, and coke produced during DRM, thus convoluting the analysis of TPO results. However, the sample with the highest initial concentration of organics and the most stable performance, Ni–Mo/MgO 38%, showed negligible concentration of surface carbonaceous species according to TPO results (Fig. S6). It appears that the concentration of organics in the as-synthesized catalyst even alters the catalytic pathways towards coke deposition, and coking was suppressed for the Ni–Mo/MgO 38% sample.

After showing the dramatic effect of residual EG on the catalytic performance, we studied whether the gas phase species involved in DRM participated in the generation of catalytic sites, along with the residual organic species. To determine the gas components driving the formation of active sites along with EG, the as-synthesized Ni–Mo/MgO 38% was pretreated in individual gas species (10% CH₄, 10% CO₂, 20% CO, or 20% H₂) before conducting DRM, to unveil the reactants/products that drive the generation of the active sites in conjunction with EG. As illustrated in Fig. 3A and B, only the reduction with 20% H₂ led to the activation of Ni–Mo/MgO 38%, achieving CH₄ and CO₂ conversion comparable to the results presented in Fig. 2. Thus, adsorbed EG and H₂ were responsible for the generation of the catalytic active sites.

Fig. 3 EG and H₂ are responsible for the generation of the active sites. Conversion of (A) CH₄ and (B) CO₂ during DRM with 60 cm³ min⁻¹ 8.3% CH₄, 8.3% CO₂ balanced by Ar at 650 °C over 5 mg Ni–Mo/MgO 38% after the sample undergoes pretreatment in 20% H₂, 20% CO, 10% CO₂ or 10% CH₄ at 650 °C for 2 h. The dashed lines are used to guide the eyes.

Chemical and bonding properties of the catalytic sites promoted by EG and H₂

Given the significant impact of the organics content on catalyst performance, we proceeded to study whether the presence of organics impacted the crystal, electronic and/or bonding structure of the as-synthesized and spent catalysts. X-ray diffraction (XRD) patterns (Fig. S7) of the as-synthesized samples showed typical diffraction peaks at 43.06° and 62.46°, corresponding to the (200) and (220) crystal planes of the MgO rock-salt structure. No distinct diffraction peaks related to metallic or alloy phases were observed, likely indicating the formation of highly dispersed metal/oxide species. High-angle annular dark-field scanning transmission electron microscopy (HAADF–STEM) and energy-dispersive X-ray spectroscopy (EDS) images supported homogenous distribution of Ni and Mo on the surface of all samples (Fig. 4). The Mo signal was significantly lower in Ni–Mo/MgO 38% (Fig. S8) compared to the other samples. Considering the blurred SEM images (Fig. S1B), the weakened Mo signal may be attributed to the coverage of the surface by organic residues.

Fig. 4 Homogenous distribution of dispersed Ni and Mo in the Ni–Mo/MgO samples. HAADF–STEM images and EDS elemental mapping of as-synthesized Ni–Mo/MgO samples with 8% (A), 12% (B) and 14% (C) organic contents.

To study the impact of the organic content on the local environment of Ni in the Ni–Mo/MgO samples, X-ray absorption spectroscopy (XAS) was performed. As shown in Fig. 5A, all four samples exhibited a common pre-edge absorption peak at 8333.2 eV, characteristic of Ni²⁺.²⁰ Furthermore, the absorption edges align with those of the NiO reference, indicating the predominance of Ni²⁺ species in the synthesized samples. Extended X-ray absorption fine structure (EXAFS) spectra revealed the absence of second-shell scattering peaks associated with Ni–Ni contribution (Fig. 5B), typical of bulk NiO phase. This suggested that no Ni²⁺ aggregates were present in the as-synthesized catalysts. Additionally, the distinct oscillatory patterns observed in the k-space EXAFS spectra of Ni–Mo/MgO samples highlighted significant structural differences from bulk NiO (Fig. 5C). Despite being synthesized in a reductive solvent environment, the catalysts predominantly featured ionic Ni²⁺ rather than metallic Ni, differing from previous reports.¹⁰ Since only Ni–O can be clearly observed in the raw data, the fitting model only included the Ni–O bond for the as-synthesized catalysts. The fitting results (Table S2) showed a coordination number (CN) of 4.1–4.2 for the first shell Ni–O, lower than those in bulk NiO and Ni in the Ni–MgO solid solutions.^21,22 The relatively lower coordination number could be attributed to the bonding of Ni²⁺ ions on the surface of MgO. Interestingly, despite the different concentrations of organics in the as-synthesized catalyst samples, and their different catalytic activity and stability, the local environment of Ni²⁺ was essentially the same for all as-synthesized samples. In other words, all the as-synthesized samples can be regarded as precatalysts sharing the same cationic Ni²⁺ species, which then undergo divergent evolutions controlled by the residual organic species.

Fig. 5 As-synthesized catalysts present isolated cationic Ni species. Spent catalysts present metallic Ni species. Ni K edge ex situ XAS spectra of as-synthesized Ni–Mo/MgO samples. (A) XANES, (B) R-space and (C) k-space EXAFS spectra. Ni K edge ex situ XAS spectra of spent Ni–Mo/MgO samples after DRM at 650 °C for 16 h. (D) XANES, (E) R-space and (F) k-space EXAFS spectra.

For spent catalyst samples after DRM at 650 °C, XAS characterization confirmed the formation of metallic Ni (Fig. 5D), as indicated by the shift in the absorption edge to lower energy, approaching that of the Ni foil reference. However, a significant amount of cationic Ni species remains, as evidenced by the white line peak features. These oxidized species could be formed due to air exposure during ex situ characterization. EXAFS of the spent samples showed a peak at 2.1 Å, corresponding to Ni–Ni scattering within metallic clusters (Fig. 5E). The distinct oscillatory pattern observed in the k-space EXAFS spectra of spent Ni–Mo/MgO samples showed a structure that did not fully resemble NiO or Ni foil (Fig. 5F). The fitting results (Table S3) show coordination number (CN) of 2.8–4.9 for the first-shell Ni–Ni scattering at 2.462–2.471 Å (typical bonding distance in metallic phase). This much lower value, compared to metallic Ni foil (CN = 12), indicates the smaller particle size of formed Ni NPs.

XAS analysis indicated that the electronic and coordination states of Ni species in both the as-synthesized and spent catalysts were largely unaffected by the amount of organic adsorbates present. However, the organic content strongly influenced catalytic rates and stability (Fig. 2 and Fig. S3). It is also worth noting that estimating particle sizes based on EXAFS coordination numbers has known limitations.²³ Because XAS is a bulk-averaging characterization, it can easily overlook a significant fraction of sub-nanometer clusters. Thus, significant particle size differences could be present with varying residual EG contents, despite the identical local environments suggested by EXAFS analysis. To better understand the role of EG in tuning catalyst performance and its influence on Ni–Mo interactions, high-resolution STEM imaging combined with EDS mapping was performed on Ni–Mo/MgO catalysts with 8%, 14%, and 38% organic loadings. Samples were reduced in 20% H₂ at 650 °C for 2 h, conditions previously shown to activate them for DRM (Fig. 3). STEM images revealed that higher EG surface coverage led to smaller average Ni NP sizes (Fig. 6 and Fig. S9). Elemental mapping of individual high-contrast particles showed significant overlap between Ni and Mo signals (Fig. 6F and G and Fig. S9 (top)), indicating the formation of co-localized Ni–Mo sites. The MgO support did not appear to encapsulate the NPs (Fig. 6H and I and Fig. S10). Overall, varying EG loadings produced co-localized Ni–Mo particles with similar electronic structures (consistent with XAS results in Fig. 5), while higher EG content suppressed particle sintering under reducing conditions.

Fig. 6 The more EG in the as-synthesized catalyst, the smaller the Ni particle size. Co-localized Ni–Mo particles are formed due to EG + H₂ treatment. (A) Average Ni particle size for reduced (20% H₂, 650 °C, 2 h) Ni–Mo/MgO 8%, Ni–Mo/MgO 14% and Ni–Mo/MgO 38%. (B) STEM images (ADF = annular dark field), (C) EDS elemental mapping and (D) particle size distribution of Ni NPs. (E) High-resolution STEM image and (F)–(I) elemental mapping for reduced (20% H₂, 650 °C, 2 h) Ni–Mo/MgO 14%.

As presented before, pre-calcination of the catalyst samples led to inactive catalyst samples (Fig. S5). XAS characterization of these inactive samples showed that Ni species were present in the oxidized state, as indicated by the overlapping absorption edge and pre-edge feature peaks (Fig. S11A). However, the second-shell scattering peak at 2.6 Å increased significantly, suggesting the formation of NiO NPs (Fig. S11B). In the k-space EXAFS spectra, the oscillation patterns of the calcined samples closely resembled that of the NiO reference (Fig. S11C), unlike the as-synthesized samples. To evaluate the reactivation of the catalyst samples, the calcined Ni–Mo/MgO 38% sample was reduced in 4% H₂/Ar at 650 °C for 3 h and tested for DRM at 650 °C. No activity was observed. Taking a step further to reactivate the calcined deactivated samples, the calcined Ni–Mo/MgO 38% sample was impregnated with an EG solution containing PVP, followed by washing with a mixture of EG + ethanol for 5 times (to replenish the organics content in the original sample, Fig. S12). However, the samples remain inactive for DRM at 650 °C. This suggested that it is crucial to preserve the isolated Ni²⁺ ions in the as-synthesized state to form active species with the assistance of EG and H₂. Calcination altered the nickel dispersion and interaction with the support, therefore the inactivity of this sample was not exclusively related to the absence of EG on the surface but also to the generation of inert Ni species.

Broader impact: EG is also key when using other MgO support

To evaluate the broader applicability of our findings, the EG-assisted synthesis was repeated using a commercial MgO support from a different supplier (denoted MgO*). MgO and MgO* differ on their surface areas (21 m² g⁻¹ and 40.2 m² g⁻¹, respectively), basicity (MgO is more basic than MgO* (Fig. S13A)), and the mechanism for EG decomposition that they enable (Fig. S13B).

After optimizing the synthesis procedure (see Methods), we found that EG was essential for producing active DRM catalysts on MgO*. EG loading was introduced through EG + ethanol washing steps; in contrast, catalysts washed with ethanol alone showed no catalytic activity. To assess reproducibility, multiple batches were prepared. Despite batch-to-batch variability, the presence of EG on the surface clearly determined whether the catalyst was active or inactive (Fig. S13C–E). A representative Ni–Mo/MgO* sample (21% organics) was evaluated for DRM under different pretreatment conditions: no pretreatment, air calcination at 550 °C for 1 h, and treatment in individual DRM-related gases (10% CH₄, 10% CO₂, 20% CO, or 20% H₂) at 650 °C for 2 h (Fig. 7A and B). As observed with the original MgO support, MgO* became inactive after calcination (which removes EG) or pretreatment in Ar (not shown), CO, CO₂, or CH₄. Pretreatment in H₂ resulted in some activity, but performance remained lower than that of the untreated catalyst. These results suggest that although both MgO and MgO* supports can host the formation of active species for the reforming reaction in the presence of residual EG, the gas-phase components, either from the feed or generated in situ, also dictate the activity of the formed species. For MgO, activity is primarily driven by H₂ generated in situ, whereas for MgO*, activation appears to involve a combination of DRM-related species. TPO analysis of treated catalysts (Fig. S14) confirmed that activity was not correlated with the retention of organic species. Notably, CO-treated samples retained the highest level of surface organics but were completely inactive. Temperature-programmed desorption (TPD) characterization of EG impregnated on MgO* showed that most EG decomposes below 650 °C (Fig. S13B); therefore, after pretreatment at 650 °C, organic residues have a different nature compared to pristine EG. In the case of the CO-treated sample, at least some of the residue became graphitic carbon (2θ = 26°, Fig. 7C-ii) and it was more difficult to oxidize compared with the residue left by other treatments (see TPO in Fig. S14A). In summary, the organic residues on the surface are one of the key parameters, along with the gas-phase composition and surface properties of the support, that determine the catalytic activity.

Fig. 7 In addition to EG and H₂, other chemical species during DRM are responsible for the generation of the catalytic active sites in the MgO* support. Conversion of (A) CH₄ and (B) CO₂ during DRM with 60 cm³ min⁻¹ 8.3% CH₄, 8.3% CO₂ balanced by Ar at 650 °C over 5 mg Ni–Mo/MgO* 21% after the sample undergoes no pretreatment, pretreatment in air at 550 °C for 1 h, or in 20% H₂, 20% CO, 10% CO₂, 10% CH₄ at 650 °C for 2 h. Note the partition in the y-axis signaled by the dashed line. Pretreatment in Ar at 650 °C for 2 h did not show reactant conversion (not shown). (C) XRD patterns of Ni–Mo/MgO* 21% after conducting DRM for 2 h at 650 °C, and after pretreatment in 20% H₂, 20% CO, 10% CO₂, 10% CH₄ at 650 °C for 2 h (MgO* reference is included). (i) full spectra (ii) graphitic carbon region, (iii) metallic Ni region. Spectra were normalized by the MgO diffraction peak at 2θ = 43°. (D) Ex situ Ni 2p_3/2 XPS spectra of Ni–Mo/MgO* 21%: as-synthesized, after conducting DRM for 2 h at 650 °C, and after pretreatment in in 20% H₂, 20% CO, 10% CO₂, 10% CH₄ at 650 °C for 2 h.

To better understand the origin of the different catalytic performances following various pretreatments, STEM characterization was performed on catalyst samples after DRM at 650 °C (no pretreatment) and after H₂ pretreatment at 650 °C. The images showed that H₂ pretreatment preserved a fraction of Ni in a dispersed state, whereas DRM conditions led to the formation of a greater number of Ni NPs (Fig. S15). In contrast to MgO, co-localization of Ni and Mo was not observed on the MgO* support, which could potentially contribute to the relatively lower activity and quicker deactivation. To further probe the metal/alloy phases formed under these conditions, XRD characterization was conducted. XRD patterns were collected for the as-synthesized Ni–Mo/MgO* 21% catalyst, as well as after DRM at 650 °C for 2 h (without pretreatment) and after pretreatment for 2 h at 650 °C in 20% H₂, 20% CO, 10% CO₂, or 10% CH₄ (Fig. 7C). Diffraction patterns revealed that treatment with reductive components in DRM, H₂, and CO promoted the generation of metallic Ni more than treatment in CO₂ and CH₄ (Fig. 7C-iii).

X-ray photoelectron spectroscopy (XPS) was conducted to provide deeper insights into the oxidation state of the metals after different pretreatments (Fig. 7D). XPS revealed that Ni was predominantly in the Ni²⁺ state, with a minor contribution from Ni⁰, in the as-synthesized catalyst. The sample with “no pretreatment”, which showed the greatest conversion, presented mostly metallic Ni, with a small fraction remaining oxidized. Samples pretreated in CO₂ and CH₄ exhibited a higher ratio of cationic to metallic Ni. Regarding Mo, the as-synthesized sample exhibited a Mo 3d binding energy near ∼234 eV, consistent with Mo⁶⁺. After various treatments, signal intensity remained in this region, but the spectra were poorly defined (Fig. S16). This likely resulted from overlapping Mo 3d doublets from lower oxidation states and/or sample charging effects.

XRD and XPS provided congruent findings regarding the oxidation state of Ni after different treatments. However, nanoscale features or active site poisoning that might impact the catalytic performance may remain obscure. For instance, close inspection of XRD patterns revealed that some graphitic coke was generated after 20% CO pretreatment (Fig. 7C-ii), which could potentially lead to coking of active Ni sites. However, the most active sample (“No pretreatment” in Fig. 7A and B) also presented a subtle XRD peak corresponding to graphitic carbon (Fig. 7C-ii), which highlights the need for a judicious study to truly pinpoint the active site and its deactivation mechanisms. In fact, the active site could be a combination of Ni sites with varied oxidation states. Recent studies have shown that an adequate balance of Ni²⁺ and Ni⁰ species is required to complete the catalytic cycle for DRM, where CH₄ activation primarily occurs on Ni⁰, and cationic Ni assists CO₂ activation and oxidation of carbonaceous surface species.^16,24,25 All in all, the experimental evidence collected here allows to state that the formation of metallic Ni alone did not guarantee catalytic activity and the unequivocal identification of the active sites exposed at the surface will be matter of future work.

This work showed that EG remaining on the surface after synthesis of Ni–Mo/MgO, and H₂ generated in situ during DRM, were crucial to reduce isolated cations of Ni to produce NPs of co-localized Ni–Mo active for DRM (Fig. 8). The crucial role of EG to generate catalytic active sites was also proven on other magnesium oxide support (MgO*); however, not only H₂ but also other DRM-related species were needed to activate the catalyst.

Fig. 8 Schematic of conditions required for formation of co-localized Ni–Mo particles on MgO for DRM. MgO* generates active sites due to EG under DRM conditions, but Ni–Mo were not co-localized.

Conclusion

We challenged a common catalyst synthesis procedure and found that residual ethylene glycol (EG) on the catalyst surface plays a critical role in the performance of Ni–Mo/MgO catalysts for high-temperature dry reforming of methane (DRM). Although EG decomposes below the reaction temperature (650 °C), it interacts with the supported metals during its removal, facilitating the formation of active catalytic sites. Specifically, EG, together with H₂, promotes the reduction of isolated Ni²⁺ species in the as-synthesized catalyst, leading to the formation of Ni–Mo nanoparticles that are active for DRM. The amount of residual EG can be tuned during a washing step using EG and ethanol. Higher EG content suppressed sintering of Ni–Mo nanoparticles, resulting in improved stability and higher reaction rates. The catalyst with the highest residual organic content (Ni–Mo/MgO 38%) exhibited the greatest site time yield, best stability, and lowest carbon deposition during DRM. These findings have broader implications. Using an alternative support (MgO*), EG remained essential for generating active sites; however, Ni–Mo co-localization was not observed, and activation required a combination of DRM-related species rather than H₂ alone. Overall, this work highlights that interactions between organic adsorbates, supported metals, and the support, are key to enable catalytic properties, offering new insights for precision catalyst synthesis across a range of chemical transformations.

Methods

Catalyst synthesis

Nickel chloride hexahydrate (NiCl₂·6H₂O, 238 g mol⁻¹, Sigma Aldrich), ammonium molybdate tetrahydrate ((NH₄)₆Mo₇O₂₄·4H₂O, Sigma Aldrich), sodium hydroxide (NaOH, Sigma Aldrich), polyvinylpyrrolidone (PVP, M_w = 55 000, Sigma Aldrich), MgO (Dickinson Corporation, 21 m² g⁻¹), MgO* (Sigma Aldrich, 40.2 m² g⁻¹), hydrazine hydrate (N₂H₄·xH₂O, 50–60%, Sigma Aldrich) and ethylene glycol (HOCH₂CH₂OH, Fisher Chemical) were used as received without further purification.

To synthesize the MgO support, sourced from Dickinson Corporation, particles of a lansfordite (MgCO₃·5H₂O) precursor were precipitated from ∼1 M aqueous Mg(HCO₃)₂ solution in a batch reaction at 5 °C using air sparging from a gas-dispersing blade at 700 rpm. The lansfordite particles were rinsed with ethanol, dried and then calcined under flowing air at 1050 °C for 30 minutes to form the polycrystalline MgO assembly (PMA) particles.

To synthesize the Ni–Mo/MgO catalysts, the general procedure reported in ref. 10 was followed. The metal precursor solution (marked as solution A) was prepared by adding 1.4140 g NiCl₂·6H₂O into 93.49 g EG, followed by addition of 0.1288 g (NH₄)₆Mo₇O₂₄·4H₂O in 3.5 g of DI water. The mixture was stirred at 500 rpm overnight for complete dissolution. The basic solution (designated as solution B) was prepared by dissolving 0.677 g of NaOH in 36 g of EG, followed by dilution to 50 mL with additional EG. The mixture was stirred at 500 rpm overnight for complete dissolution. The reducing agent solution (designated as solution R) was prepared by mixing 2.558 g of hydrazine hydrate with 5 g of EG, then adjusting the total volume to 10 mL by adding EG in a volumetric flask. In a typical synthesis, 28.164 g of solution A were added into a preheated Teflon liner (80 °C) followed by addition of 0.38 g PVP and stirring for approximately 1 h for complete dissolution. Then, 0.9 g of MgO were added, and the solution was further stirred for approximately 15 min. Next, 5.432 g of solution R and 11.505 g of solution B were added. Then, the Teflon liner was introduced in an autoclave reactor and the reactor was sealed. The autoclave was introduced in an oil bath at 80 °C for 1 h. The contents of the autoclave were stirred at 500 rpm using a magnetic stirbar. Afterwards, the autoclave was cooled down in running water and the solid was separated by vacuum filtration. The sample cake was further washed with 45 mL of a EG + ethanol mixture (volumetric ratio of 3 to 7) for 3, 5, 7 or 9 times to tune the organic contents. For the sample washed 9 times, the mixture was soaked in the EG/ethanol mixture for ∼3 min in the last wash. After washing, the sample was dried in the fume hood at room temperature overnight.

To synthesize the Ni–Mo/MgO* catalysts, 28.164 g of solution A were added into a preheated Teflon liner (located inside an autoclave reactor) at 80 °C, followed by addition of 0.38 g PVP and stirring for approximately 1 h for complete dissolution. Then, 0.9 g of MgO were added, and the solution was further stirred for approximately 15 min. Next, 5.432 g of solution R and 11.505 g of solution B were added. The final solution was poured into another Teflon liner (located in an autoclave reactor preheated to 65 °C inside an oil bath). The Teflon liner was covered with a Teflon lid, but the autoclave reactor was not sealed. The catalyst synthesis reaction was allowed to proceed for 1 h. The contents of the autoclave were stirred at 500 rpm using a magnetic stirbar. Afterwards, the autoclave was cooled down in running water and the solid was separated by vacuum filtration. The sample cake is further washed with 45 mL of an EG + ethanol mixture (volumetric ratio of 3 to 7) for 9 times. After washing, the sample was dried in the fume hood at room temperature overnight.

Catalyst characterization

Powder X-ray diffraction (XRD) analysis of catalysts was conducted using a PANalytical X’Pert Pro MPD system equipped with an X’Celerator solid-state detector (Cu Kα source, λ = 0.1542 nm, 45 kV, 40 mA). Scanning was performed over a wide 2θ range, utilizing a step size of 0.007° and a rate of 0.03° s⁻¹. The metal contents of the catalysts were determined by inductively coupled plasma atomic emission spectroscopy (ICP–AES), conducted by Galbraith Laboratories, Inc.

Microscopy was conducted using an FEI Talos F200X TEM, equipped with four X-ray detectors and operated at 200 kV. Quantitative EDS analysis was performed using TFS Velox software version 3.10. For scanning transmission electron microscope (STEM) imaging, the collection angle was 0–9 mrad for bright-field (BF) imaging and 61–200 mrad for annular dark-field (ADF) imaging. The size of Ni–Mo particles was quantified using superimposed STEM–ADF imaging and Ni energy-dispersive X-ray spectroscopy (EDS) mapping.

Raman spectroscopy was performed using both a 244 nm and a 532 nm excitation laser. Measurements were performed ex situ at room temperature. For the measurements taken using a 244 nm laser excitation (2 mW at sample position), a multiwavelength Raman system²⁶ was used. Spectra were collected via a customized ellipsoidal mirror and directed by a fiber optics bundle to the spectrograph stage of a triple Raman spectrometer (Princeton Instruments Acton Trivista 555). To block the laser irradiation, an edge filter (Semrock) was used in front of the UV-vis fiber optic bundle (Princeton Instruments). A UV-enhanced liquid N₂-cooled CCD detector (Princeton Instruments) was employed for signal detection. The samples were analyzed on a stationary sample holder at different exposure times until a clear signal was observed. For the measurements taken using a 532 nm laser excitation, a Renishaw inVia Qontor confocal Raman microscope was used. The instrument was operated with a long-working-distance 50× objective, employing a 2400 l mm⁻¹ grating. Spectra were acquired using a laser power of 1% and an exposure time of 60 s over a spectral range of 300–2000 cm⁻¹. Background correction for the MgO samples was performed using the WiRE 5.6 software, applying the subtract baseline tool with an intelligent polynomial of order 3.

Temperature-programmed oxidation (TPO) was conducted using a commercial AMI-300 instrument (Altamira Instruments, Inc.). For the as-synthesized Ni–Mg/MgO samples, 15 mg of sample was loaded into the U-shape reactor without pelletizing or sieving, fixed in place between two plugs of quartz wool. The reactor was sequentially flushed with 75 cm³ min⁻¹ Ar for 0.5 h and 40 cm³ min⁻¹ 5% O₂/He for 1 h at 30 °C, followed by heating up to 900 °C at 10 °C min⁻¹ and holding for 30 min. Quantification of coke on spent catalysts was conducted using the same procedure, without unloading or separating the catalyst after 16 h DRM reaction at 650 °C with 60 cm³ 8.3% CH₄, 8.3% CO₂ balanced by Ar. The amount of deposits was quantified by pulse-calibration of the MS signal.

Temperature-programmed desorption (TPD) on the EG/MgO and EG/MgO* samples was conducted by mixing 25 mg of sample with 100 mg of quartz sand, and loading the mixture into a packed-bed reactor. The bed was heated to 50 °C and flushed under 30 cm³ min⁻¹ He for 30 min. After the flush, the reactor was heated to 880 °C at a rate of 10 °C min⁻¹ under 30 cm³ min⁻¹ He. The evolved gases were monitored using an online MS.

CO₂-TPD was conducted by mixing 25 mg of MgO (or MgO*) with 100 mg of quartz sand, and loading the mixture into a packed-bed reactor. The reactor bed was pretreated in 30 cm³ min⁻¹ air at 650 °C for 2 h using a heating rate of 10 °C min⁻¹. Following pretreatment, the bed was cooled to 50 °C and purged with 30 cm³ min⁻¹ Ar for 30 min. The reactor was then exposed to 30 cm³ min⁻¹ of 10% CO₂/Ar for 20 min, followed by an additional 30 min Ar purge. After the purge, the reactor was heated to 850 °C at a rate of 10 °C min⁻¹ under 30 cm³ min⁻¹ Ar. The evolved gases were monitored using an online MS. The reported CO₂ signals (m/z = 44) for both samples were normalized to the Ar signal (m/z = 40) of their respective run.

Liquid-state ¹³C nuclear magnetic resonance (¹³C NMR) spectra were recorded on a JEOL 400YH nuclear magnetic resonance spectrometer using a deuterated dimethyl sulfoxide (DMSO-d6) capillary as the solvent. Chemical shifts were referenced to the residual solvent signal. Solid State NMR (SSNMR) measurements were performed on a 400 MHz Varian Inova solid state NMR spectrometer with a 5 mm Chemagnetic probe. ¹³C SSNMR experiments were recorded at room temperature with magic angle spinning (MAS) between 6.5 and 7.0 KHz. ¹³C spectra was acquired with a tangent shape cross polarization (TAN-CP) pulse sequence and Spinal 1H decoupling during the acquisition.

Ni K-edge X-ray absorption spectroscopy (XAS) measurements were conducted at beamline BL 4-1 of the Stanford synchrotron radiation lightsource (SSRL), SLAC National Accelerator Laboratory. Prior to measurement, all samples were diluted with varying amounts of boron nitride and pressed into thin pellets. Spectra were acquired in transmission mode, with energy calibration performed using a nickel metal foil. The data analysis and reduction process were carried out utilizing the Demeter software suite.²⁷

Ex situ X-ray photoelectron spectroscopy (XPS) measurements were performed using a Thermo Scientific Nexsa G2 XPS instrument (Waltham, MA, USA). The system employs a monochromatic, micro-focused Al K α X-ray source (1486.6 eV) with an adjustable spot size ranging from 30 to 400 µm. The instrument is equipped with a hemispherical electron energy analyzer and a 128-channel detector array. The analysis chamber achieves a vacuum of at least 2 × 10⁻⁹ mbar. Data acquisition and analysis were carried out using the Thermo Scientific Avantage XPS software.

Catalyst testing

Dry reforming of methane (DRM) was conducted on a commercial µBenchCAT4000 unit from Altamira Instruments. Typically, 5 mg of sample were pelletized and sieved to particles size 180–250 µm and diluted with 180–250 µm quartz sand with a weight ratio of 1 : 10, followed by loading into a U-shape quartz reactor (inner diameter 4 mm) with a fixed-bed configuration. A K-type thermocouple, covered with a quartz sleeve, was located on the top of the catalyst bed for temperature control. For the DRM reaction at 650 °C, 60 cm³ min⁻¹ with composition of 8.3% CH₄, 8.3% CO₂ (diluted by Ar) was flowed through the catalyst bed at 30 °C for 70 min. Next, the reactor temperature was increased at 10 °C min⁻¹ to the desired reaction temperature. In the figures presented, TOS = 0 is when the temperature reached the desired reaction temperature. To control the conversion away from equilibrium at 800 °C, the reactant gas flow was increased to 96 cm³ min⁻¹ with the same composition. The reactor exhaust was monitored online with a mass spectrometer (MS, Pfeiffer Vacuum) and a gas chromatograph (Agilent) equipped with a thermal conductivity detector (TCD).

The conversion of CH₄ and CO₂ was calculated using eqn (1):

where X_reactant (%) is the conversion for CH₄ and CO₂, C_inlet is the concentration of reactant feeding into the reactor, C_outlet is the concentration of reactant in product stream, F_inlet is the flow rate of the feed and F_outlet is the flow rate of the outlet stream.

Author contributions

J. Zhang: conceptualization, data curation, formal analysis, investigation, methodology, validation, writing – original draft, writing – review & editing. N. B. Abdul Rahman: investigation, writing – review & editing. B. Zingg: investigation. Y. Li: formal analysis, methodology, writing – review & editing. Y. Wu: investigation, writing – review & editing. L. Zhang: investigation, writing – review & editing. Y. Lin: formal analysis, investigation, writing – review & editing. J. D. Arregui-Mena: formal analysis, investigation, writing – review & editing. L. Qiu: formal analysis, investigation, writing – review & editing. Z. Yang: formal analysis, investigation, methodology, writing – review & editing. H. Meyer: investigation, formal analysis. Z. Wu: funding acquisition, investigation, methodology, supervision, writing – review & editing. F. Polo-Garzon: conceptualization, funding acquisition, investigation, methodology, project administration, supervision, validation, writing – review & editing. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: SEM images, NMR spectra, Raman spectra, catalytic performance, TPO, XRD, STEM images, EDS mapping, XAS spectra, TGA, TPD, XPS spectra, ICP-AES, EXAFS fitting results (Tables S1–S4). See DOI: https://doi.org/10.1039/d6ey00088f.

Acknowledgements

We are grateful to Dickinson Corporation (https://dickinsoncorp.com/) for supplying the MgO samples free of charge. B. Z. was supported by an appointment to the National Nuclear Security Administration Impact Internship Program (NNSA-IMPACT), sponsored by the U.S. Department of Energy and administered by the Oak Ridge Institute for Science and Education. This research was sponsored by the U.S. Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division, Catalysis Science program. Some of the work including Raman spectra measurement was conducted as part of a user project at the Center for Nanophase Materials Science (CNMS), which is a U.S. DOE Office of Science User Facility located at Oak Ridge National Laboratory. This research used resources of the Stanford Synchrotron Radiation Lightsource (SSRL) of SLAC National Accelerator Laboratory supported by Basic Energy Sciences under contract No. DE-AC02-76SF00515. This research used FEI Talos F200X TEM of the Center for Functional Nanomaterials which is a U.S. DOE, Office of Science User Facilities located at Brookhaven National Laboratory under Contract No. DE-SC0012704.

References

1. F. Fiévet, S. Ammar-Merah, R. Brayner, F. Chau, M. Giraud, F. Mammeri, J. Peron, J.-Y. Piquemal, L. Sicard and G. Viau, The polyol process: a unique method for easy access to metal nanoparticles with tailored sizes, shapes and compositions, Chem. Soc. Rev., 2018, 47(14), 5187–5233.
2. K. Eid; H. Wang and L. Wang, Nanoarchitectonic metals, in Supra-materials nanoarchitectonics, Elsevier, 2017, pp. 135–171.
3. B. Wiley, Y. Sun, B. Mayers and Y. Xia, Shape-controlled synthesis of metal nanostructures: the case of silver, Chem. – Eur. J., 2005, 11(2), 454–463.
4. F. Polo-Garzon, T. F. Blum, Z. Bao, K. Wang, V. Fung, Z. Huang, E. E. Bickel, D.-e Jiang, M. Chi and Z. Wu, In Situ Strong Metal–Support Interaction (SMSI) Affects Catalytic Alcohol Conversion, ACS Catal., 2021, 11(4), 1938–1945,  DOI:10.1021/acscatal.0c05324.
5. C.-B. Wang, Y. Cai and I. E. Wachs, Reaction-Induced Spreading of Metal Oxides onto Surfaces of Oxide Supports during Alcohol Oxidation:  Phenomenon, Nature, and Mechanisms, Langmuir, 1999, 15(4), 1223–1235,  DOI:10.1021/la9807210.
6. M. Bradford and M. Vannice, CO₂ reforming of CH₄, Catal. Rev., 1999, 41(1), 1–42.
7. P. Djinović, I. G. Osojnik Črnivec, B. Erjavec and A. Pintar, Influence of active metal loading and oxygen mobility on coke-free dry reforming of Ni–Co bimetallic catalysts, Appl. Catal., B, 2012, 125, 259–270,  DOI:10.1016/j.apcatb.2012.05.049.
8. P. Djinović, I. G. O. Črnivec and A. Pintar, Biogas to syngas conversion without carbonaceous deposits via the dry reforming reaction using transition metal catalysts, Catal. Today, 2015, 253, 155–162,  DOI:10.1016/j.cattod.2015.01.039.
9. I. G. Osojnik Črnivec, P. Djinović, B. Erjavec and A. Pintar, Effect of synthesis parameters on morphology and activity of bimetallic catalysts in CO2–CH4 reforming, Chem. Eng. J., 2012, 207–208, 299–307,  DOI:10.1016/j.cej.2012.06.107.
10. Y. Song, E. Ozdemir, S. Ramesh, A. Adishev, S. Subramanian, A. Harale, M. Albuali, B. A. Fadhel, A. Jamal and D. Moon, Dry reforming of methane by stable Ni–Mo nanocatalysts on single-crystalline MgO, Science, 2020, 367(6479), 777–781.
11. H. Zhou, T. Zhang, Z. Sui, Y.-A. Zhu, C. Han, K. Zhu and X. Zhou, A single source method to generate Ru-Ni-MgO catalysts for methane dry reforming and the kinetic effect of Ru on carbon deposition and gasification, Appl. Catal., B, 2018, 233, 143–159,  DOI:10.1016/j.apcatb.2018.03.103.
12. Y. Wang, L. Li, G. Li, Q. Zhao, X. s Wu, Y. Wang, Y. Sun and C. Hu, Synergy of Oxygen Vacancies and Ni0 Species to Promote the Stability of a Ni/ZrO2 Catalyst for Dry Reforming of Methane at Low Temperatures, ACS Catal., 2023, 13(10), 6486–6496,  DOI:10.1021/acscatal.2c06412.
13. A. Beheshti Askari, M. al Samarai, N. Hiraoka, H. Ishii, L. Tillmann, M. Muhler and S. DeBeer, In situ X-ray emission and high-resolution X-ray absorption spectroscopy applied to Ni-based bimetallic dry methane reforming catalysts, Nanoscale, 2020, 12(28), 15185–15192,  10.1039/D0NR01960G.
14. G. Gunduz-Meric, S. Kaytakoglu and L. Degirmenci, Ni,Co/SiO₂ and Ni/SiO₂,Co bimetallic microsphere catalysts indicating high activity and stability in the dry reforming of methane, React. Kinet., Mech. Catal., 2020, 129(1), 403–419,  DOI:10.1007/s11144-019-01708-4.
15. S. Shah, J. Hong, L. Cruz, S. Wasantwisut, S. R. Bare and K. L. Gilliard-AbdulAziz, Dynamic Tracking of NiFe Smart Catalysts using In Situ X-Ray Absorption Spectroscopy for the Dry Methane Reforming Reaction, ACS Catal., 2023, 13(6), 3990–4002,  DOI:10.1021/acscatal.2c05572.
16. X. Liu, J. Wen, Y. xie, Z. Li, M. Zhong, R. Nie, S. Peng, J. Cao, J. Chen and P. Ning, et al., Modulating proportion of Ni0 species stabilized by Ni²⁺ on Ni-MgO catalyst with superior stability for dry reforming of methane, Chem. Eng. J., 2024, 493, 152499,  DOI:10.1016/j.cej.2024.152499.
17. X. Zhang, J. Deng, M. Pupucevski, S. Impeng, B. Yang, G. Chen, S. Kuboon, Q. Zhong, K. Faungnawakij and L. Zheng, High-performance binary Mo–Ni catalysts for efficient carbon removal during carbon dioxide reforming of methane, ACS Catal., 2021, 11(19), 12087–12095.
18. R. Eluri and B. Paul, Synthesis of nickel nanoparticles by hydrazine reduction: mechanistic study and continuous flow synthesis, J. Nanopart. Res., 2012, 14, 1–14.
19. S. Ahmed, J. Pan, M. N. Ashiq, D. Li, P. Tang and Y. Feng, Ethylene glycol-assisted fabrication and superb adsorption capacity of hierarchical porous flower-like magnesium oxide microspheres for phosphate, Inorg. Chem. Front., 2019, 6(8), 1952–1961.
20. I. J. Pickering, G. N. George, J. T. Lewandowski and A. J. Jacobson, Nickel K-edge X-ray absorption fine structure of lithium nickel oxides, J. Am. Chem. Soc., 1993, 115(10), 4137–4144.
21. T. Yoshida, T. Tanaka, H. Yoshida, T. Funabiki and S. Yoshida, Study on the dispersion of nickel ions in the NiO− MgO system by X-ray absorption fine structure, J. Phys. Chem., 1996, 100(6), 2302–2309.
22. T. H. Ulucan, J. Wang, E. Onur, S. Chen, M. Behrens and C. Weidenthaler, Unveiling the structure–property relationship of MgO-supported Ni ammonia decomposition catalysts from bulk to atomic structure by in situ/operando studies, ACS Catal., 2024, 14(5), 2828–2841.
23. J. Finzel, K. M. Sanroman Gutierrez, A. S. Hoffman, J. Resasco, P. Christopher and S. R. Bare, Limits of Detection for EXAFS Characterization of Heterogeneous Single-Atom Catalysts, ACS Catal., 2023, 13(9), 6462–6473,  DOI:10.1021/acscatal.3c01116.
24. W. Zhang, O. Akdim, N. Zhang, X. Deng, H. Zhao, J. Yang, H. Song, Y. Hu, L. Chou and G. J. Hutchings, Coke-resistant Ni/Al₂O₃ catalysts for low-temperature dry reforming of methane via Ni0/NiAl2O4 interface control, Appl. Catal., B, 2026, 391, 126705,  DOI:10.1016/j.apcatb.2026.126705.
25. S. Zhang, M. Ying, J. Yu, W. Zhan, L. Wang, Y. Guo and Y. Guo, NixAl1O2-δ mesoporous catalysts for dry reforming of methane: The special role of NiAl2O4 spinel phase and its reaction mechanism, Appl. Catal., B, 2021, 291, 120074,  DOI:10.1016/j.apcatb.2021.120074.
26. Z. Wu, M. Li, J. Howe, H. M. Meyer and S. H. Overbury, Probing Defect Sites on CeO2 Nanocrystals with Well-Defined Surface Planes by Raman Spectroscopy and O2 Adsorption, Langmuir, 2010, 26(21), 16595–16606,  DOI:10.1021/la101723w.
27. B. Ravel and M. Newville, Athena, Artemis, Hephaestus: data analysis for X-ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat., 2005, 12(4), 537–541.

---

Footnote

† This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (https://energy.gov/downloads/doe-public-access-plan).

---