Methanol Steam Reforming on Ru/m-ZrO₂: Sodium Promotion of the CO₂-Forming Pathway

Abstract

Sodium (Na) promotion of Ru/m-ZrO₂ was investigated to elucidate how an alkali modification tunes selectivity in methanol steam reforming (MSR). H₂-TPR/XANES/EXAFS show that Na increases surface basicity and strengthens Ru–O interactions, shifting RuO_x reduction and H₂ spillover to a higher temperature. DRIFTS reveals Na-induced red shifts of the formate ν(CH) band and changes in OCO vibrational splitting, consistent with weakening of the formate C–H bond and an altered binding geometry. CO₂-TPD confirms a monotonic shift toward stronger basic sites with increasing Na concentrations. Under MSR conditions, Na selectively increases CO₂ concentration at the expense of CO. At ~80% conversion and 325 °C, CO₂ selectivity increases from 12.0% (unpromoted) to 16.2, 21.0, and 26.5% for 0.5, 1.0, and 1.8% Na, respectively; at ~300 °C and ~66–69% conversion, CO₂ selectivity increases from 8.6% to 23.7% at 1.8% Na. Transient MSR experiments further show earlier and larger H₂ evolution upon Na addition, corroborating the promotion of the dehydrogenation/decarbonylation route to CO₂ + H₂. We propose that Na increases basicity and modifies the Ru–support interface to favor formate dehydrogenation/decarboxylation, thereby increasing the H₂ yield and lowering CO formation. Ru’s higher-energy, less occupied d-band stabilizes CO and oxygenated intermediates more strongly in the reforming environment, making the CO-forming pathway more resistant to suppression than on Pt.

1. Introduction

Methanol steam reforming (MSR) is an attractive strategy for distributed hydrogen production because it operates at moderate temperatures and can be integrated with proton exchange membrane fuel cells (PEMFCs), which require hydrogen streams containing extremely low CO concentrations. Achieving this requires catalysts that direct methanol-derived intermediates (methoxy, formate, and carbonate) toward the CO₂-forming dehydrogenation/decarboxylation pathway instead of the CO-forming dehydration/decarbonylation route. Several recent reviews highlight the importance of surface basicity, metal–support interactions, and oxygen mobility in determining MSR selectivity [1,2,3]. Methanol’s stability, liquid-phase handling, and suitability as a hydrogen carrier further motivate its use in compact fuel processors [4].

On Pt-based catalysts, alkali promotion has proven highly effective for tuning methanol steam reforming (MSR) selectivity. On Pt/YSZ, Na doping increases CO₂ selectivity to >90%, with DRIFTS studies showing that Na weakens the formate C–H bond and redirects decomposition toward the H₂-forming dehydrogenation/decarboxylation pathway [5]. Similar behavior has been reported on Pt/CeO₂, where Li and Na promoters facilitate the scission of methoxy and formate C–H bonds, accelerating the dehydrogenation steps central to MSR and low-temperature water–gas shift chemistry [6]. More broadly, alkali modification of Pt–oxide interfaces consistently strengthens surface basicity, perturbs formate binding geometries, and biases reaction pathways toward CO₂ formation during MSR [7].

Prior work on Pt–oxide interfaces provides a useful mechanistic baseline for understanding how alkali promoters operate in MSR. On Pt/ThO₂, MSR was shown to share mechanistic features with H₂O-assisted formic acid decomposition and the water–gas shift reaction, with formate decomposition direction determined by the acid–base character of the support [8]. On Pt/CeO₂, in situ DRIFTS studies resolved the methoxy → formate transition and revealed carbonate formation during methanol conversion, providing spectroscopic markers for pathway selection [9]. Several reviews reaffirm that modifying support basicity and the interfacial oxygen content can steer formate decomposition toward CO₂ [10,11].

Alkali promotion has also been shown to improve CO₂ selectivity during MSR on metals that are not intrinsically highly active. In Na-promoted Ag/m-ZrO₂ catalysts [12], substantial increases in CO₂ selectivity were reported despite Ag exhibiting significantly lower reforming activity than Pt. These results indicate that alkali promotion can improve CO₂ selectivity during MSR even when the intrinsic reforming activity of the metal is low, supporting the broader relevance of alkali-driven selectivity control established for Pt-based catalysts.

In contrast, Ru-based catalysts—despite their well-recognized activity in reforming and hydrogen-producing reactions—show a greater propensity for CO formation under MSR conditions than Pt- or Ag-based systems [10,11,12], which complicates the suppression of the CO-producing route. Reviews of MSR catalyst design emphasize that achieving high CO₂ selectivity on Ru requires careful tuning of basicity and the interfacial oxygen content to weaken the formate C–H bond, although Ru is inherently less electronically tunable than Pt [10,11]. Insights into Ru oxide behavior is provided by mechanistic studies on reducible oxide systems that share key features with Ru/m-ZrO₂. For example, in situ DRIFTS and XPS analyses of Ni–CeO₂ reveal a methoxy → formate → carbonate progression that is strongly influenced by interfacial oxygen transfer, with CO₂ formation favored when the support actively participates in oxygen exchange [13,14,15].

The effects of the support on Ru have also been clarified through studies of CO₂ hydrogenation over supported Ru, which show that the support identity and Ru–O coordination strongly reshape adsorbate stabilization and reaction routes, highlighting why Na may influence Ru less dramatically than Pt [16]. Additional analyses of MSR catalysts, reactor designs, and integrated fuel–processor systems further contextualize the interplay of the catalyst composition, support identity, and process configuration [17,18,19,20,21,22].

These collective findings frame the central question of the present work: to what extent can Na, through its ability to increase surface basicity and modify Ru–O interfacial environments, redirect methanol-derived intermediates toward the H₂ + CO₂ pathway on Ru/m-ZrO₂? To address this, we synthesize a series of Na-promoted Ru/m-ZrO₂ catalysts and characterize them using H₂-TPR, XANES/EXAFS, CO₂-TPD, in situ DRIFTS, transient MSR experiments, and steady-state MSR testing. By correlating Na-induced changes in basicity, Ru–O coordination, and oxygenate vibrational signatures with CO₂ vs. CO selectivity and H₂ production behavior, we directly compare the magnitude and mechanistic basis of Na promotion on Ru with its well-documented effects on Pt [5,6,7].

2. Results and Discussion

Table 1. BET surface area, porosity, and average pore diameter for the prepared catalysts. Uncertainty ranges are a BET surface area of ±0.5–1.0 m²/g; BJH pore volume of ±0.01 cm³/g; and BJH pore diameter of ±1–2 Å.

Sample ID	Expected As (BET) (m2/g)	As (BET) (m2/g)	Vp (BJH Des) (cm3/g)	Dp (BJH Des) (Å)
m-ZrO2	-	95.4	0.29	95.0
1.0% Ru/m-ZrO2	94.1	94.8	0.27	93.8
0.5% Na-1.0% Ru/m-ZrO2	93.1	90.5	0.26	92.5
1% Na-1.0% Ru/m-ZrO2	92.0	83.0	0.25	90.3
1.8% Na-1.0% Ru/m-ZrO2	90.4	73.4	0.24	94.6
2.5% Na-1.0% Ru/m-ZrO2	89.0	67.1	0.22	95.2
5% Na-1.0% Ru/m-ZrO2	84.4	48.2	0.18	109.4

shows the BET surface area and pore volume diameter for the prepared catalysts. Estimates of expected BET surface areas were made assuming that Na is present as Na₂CO₃ in the as-prepared catalysts, and assuming that the promoter contributes to the mass but not the surface area (i.e., A_BET,expected = A_BET,ZrO2 × m_ZrO2/(m_ZrO2 + m_RuO2 + m_Na2CO3). The surface area and pore volume progressively decreased with increasing Na loading; a comparison of measured values with the expected values suggests some pore blocking by Na above 0.5% Na, and pore blocking is exacerbated with increasing Na contents. The average pore diameter ranged from 90.3 to 109.4 Å, and the increasing trend above 1% Na loading may be attributed to the preferential blocking of narrower pores.

H₂-TPR (Figure 1), TPR-MS (Figure S1), TPR-XANES (Figure 2 and Figure 3;

Table 2. Temperatures from LC-XANES fits of the points where composition reaches 50% Ru oxide/50% Ru⁰ during H₂ reduction.

% Na Loading	T (°C) 50% Ru Oxide/50% Ru0
0 Na	119
0.5% Na	131
1.0% Na	145
1.8% Na	159
2.5% Na	166
5.0% Na	215

), TPR-EXAFS (Figure S2), and DRIFTS (Figure 4) together characterize the reduction behavior of Ru oxide, the onset of H₂ dissociation and spillover, the removal and formation of surface oxygenated species, and the influence of Na as a promoter. For the unpromoted 1% Ru/m-ZrO₂ catalyst, H₂ uptake begins at approximately 100 °C, appearing as a small shoulder before the main reduction peak with a range of 180–250 °C and a maximum of 225 °C. At higher temperatures, broad H₂ consumption features appear with maxima at 456 °C and 744 °C. The corresponding H₂-MS trace (Figure S1a) mirrors these features, confirming that they arise from true H₂ uptake rather than thermal conductivity artifacts. TPR-XANES measurements (Figure 2a) show that the Ru oxide is reduced readily: linear combination fitting (Figure 3a) yields a 50% RuO_x/50% Ru⁰ temperature of 119 °C (Table 2), and more than 80% of the Ru is reduced by ~150 °C. TPR-EXAFS (Figure S2a) confirms this finding, as the Ru–O scattering peak (1–2 Å, phase-uncorrected) diminishes strongly between 100–150 °C, while the Ru–Ru peak (2–3 Å) grows, indicating the formation of metallic Ru.

After RuO_x is reduced, Ru⁰ enables H₂ dissociation and spillover onto m-ZrO₂, producing additional H₂ consumption above ~250 °C. This region is accompanied by CO evolution (Figure S1a), consistent with hydrogen-induced carbonate decomposition on the support. This assignment is validated by DRIFTS (Figure 4, left panel), which shows strong carbonate bands in the 1400–1600 cm⁻¹ range that are removed during H₂ treatment. Dashed horizontal lines delineate two dominant surface carbonate band clusters present on the calcined catalysts after air exposure; these are designated Cluster A and Cluster B. Using the assignments of Daturi et al. [23], the 1650–1450 cm⁻¹ region is dominated by hydrogen carbonate/bicarbonate species formed by the interaction of CO₂ with surface hydroxyl groups, comprising mainly ν_as(OCO) (~1610–1630 cm⁻¹) and ν_s(OCO) (~1410–1440 cm⁻¹) modes. The 1450–1300 cm⁻¹ region is dominated by the ν_s(OCO) modes of bidentate carbonates (~1320–1360 cm⁻¹) with overlapping contributions from more strongly bound carbonate geometries. Introducing the alkaline promoter Na increases surface basicity. This in turn increases the population of surface carbonates with increasing Na content (vertical arrows).

The temperature at which carbonate bands disappear matches the spillover-related high-temperature H₂ consumption features in the TPR profiles. DRIFTS also reveals the formation of OH groups, as seen by the growth of the broad OH-stretching envelope near 3650–3700 cm⁻¹ (Figure 4, right). The emergence of these OH bands demonstrates that H₂ spillover not only removes carbonates but also hydroxylates the ZrO₂ surface.

Dashed horizontal lines delineate the two dominant hydroxyl stretching bands on m-ZrO₂. Using the assignments of Daturi et al. [23], the 3720–3690 cm⁻¹ region (designated Cluster A) corresponds to Type I isolated (terminal) Zr–OH groups, while the 3690–3620 cm⁻¹ region (designated Cluster B) corresponds to Type II bridging Zr–OH–Zr species. As indicated by the vertical arrows, an increasing Na content modifies both the intensity and position of these OH features, indicating that Na affects spillover-induced hydroxylation.

As Na is introduced and its loading increases, all reduction-related features shift systematically toward higher temperatures. For the 0.5% Na catalyst, the primary H₂-TPR peak shifts to a range of 210–295 °C with a maximum at 253 °C, with higher temperature events having a maxima at 375 °C and 750 °C; the XANES midpoint rises to 131 °C, and EXAFS shows that Ru–O scattering persists at higher temperatures than in the unpromoted sample. At 1.0% Na, the main reduction feature appears over a range of 215–320 °C with a maximum at 275 °C, followed by a broader event (320–460 °C) having a maximum at 364 °C, and the XANES midpoint increases to 145 °C. For 1.8% Na, primary reduction in H₂-TPR occurs at ~215–445 °C with maxima at 277 °C, 331 °C, and 360 °C, with a XANES midpoint of 159 °C. Increasing the Na content to 2.5% shifts the main event to the 220–450 °C range and maxima at 291 °C, 347 °C, and 385 °C, with spillover-related uptake extending toward ~500 °C; the XANES midpoint increases to 166 °C, and Ru–O coordination is retained across a wide temperature window. With 5.0% Na, inhibition is strongest: the primary reduction appears at 235–470 °C with maxima at 307 °C and 418 °C, with high-temperature features extending above 800 °C; the XANES midpoint rises to 215 °C, with Ru–O scattering persisting to the highest temperatures. The sharper rise in the XANES midpoint temperature between 2.5 wt.% and 5.0 wt.% Na (166 → 215 °C) suggests that high Na loadings produce a more extensively oxygenated Ru–support interface, further inhibiting RuO_x reduction.

Across these techniques, Na consistently inhibits catalyst activation by delaying RuO_x → Ru⁰ reduction, suppressing H₂ dissociation and spillover, and stabilizing Ru–O coordination at higher temperatures. Stronger CO and CO₂ MS signals with Na reflect larger populations of surface carbonates, consistent with the more intense carbonate bands observed in DRIFTS (Figure 4). The simultaneous removal of carbonates and formation of OH groups upon H₂ exposure demonstrates that Na influences both surface basicity and hydrogen spillover pathways.

EXAFS fits for the Na-promoted catalysts are reported in Figure 5 and Figure S3 and

Table 3. EXAFS fits of Ru K-edge data for catalysts following reduction in flowing H₂ at 400 °C and cooling. Ranges: Δk = 2.75–10 Å⁻¹; ΔR = 1.5–3.0 Å. S₀² was fixed at 0.90. Bold text represents fixed parameters.

Sample Description	N Ru-O (Long)	R Ru-O (Long) (Å)	N Ru-Ru Metal	R Ru-Ru Metal (Å)	e0 (eV)	σ2 (Å2)	r-Factor	Est. # of Atoms	Est. Diam. (nm)
Ru0 foil	-	-	6 + 6 (fixed)	2.630 (0.0062) 2.697 (0.0056)	−2.3 −10.3 (0.52)	0.0015	0.0076	-	-
0% Na-1% Ru/m-ZrO2	0.6 (0.22)	1.947 (0.0535)	4.6 (0.76)	2.660 (0.0101)	12.1 −4.92 (0.950)	0.00276 (0.00167)	0.0067	8	0.69
0.5% Na-1% Ru/m-ZrO2	-	-	6.8 (0.65)	2.669 (0.0077)	−5.12 (0.661)	0.00194 (0.00111)	0.0118	18	0.90
1% Na-1% Ru/m-ZrO2	1.5 (0.39)	2.165 (0.0734)	5.4 (1.3)	2.654 (0.0113)	15.2 (6.31) −3.67 (1.16)	0.00175 (0.00218)	0.0038	10	0.76
1.8% Na-1% Ru/m-ZrO2	1.0 (0.51)	2.113 (0.0473)	5.3 (1.0)	2.661 (0.0122)	3.26 −3.74 (1.30)	0.00124 (0.00184)	0.0146	10	0.75
2.5% Na-1% Ru/m-ZrO2	1.4 (0.52)	2.093 (0.0931)	7.3 (1.2)	2.682 (0.0086)	6.35 (8.66) −2.88 (0.886)	0.00140 (0.00146)	0.0034	21	0.96
5% Na-1% Ru/m-ZrO2	1.6 (0.99)	2.044 (0.150)	6.0 (1.0)	2.679 (0.0207)	8.26 (13.4) −1.64 (2.16)	0.00125	0.0183	13	0.82

and Table S1. Figure 5 and Table 3 show that catalysts containing ≥1.0 wt.% Na exhibit a small but reproducible long Ru–O coordination at approximately 2.1–2.2 Å, with typical coordination numbers N ≈ 1.0–1.6. Although modest in magnitude, this contribution consistently improves the quality of the fits (specifically, as compared with Figure S3 and Table S1, where Ru–O coordination was excluded from the fitting model), indicating that it represents a real structural feature rather than a fitting artifact. In contrast, the unpromoted catalyst shows only a contribution of shorter Ru–O interactions with a lower coordination number, and the 0.5 wt.% Na sample does not require a Ru–O path, suggesting that Na loadings of 1 wt.% and above are needed to stabilize this longer Ru–O interaction at the support interface. Because the Ru particle sizes obtained from the EXAFS analysis remain small (0.7–1.0 nm across the series), a significant fraction of Ru atoms reside at the metal–support interface, where changes in oxygen coordination are detectable. The appearance of the long Ru–O feature only at ≥1 wt.% Na is therefore consistent with the Na-induced modification of the interfacial oxygen environment and correlates with the delayed RuO_X → Ru⁰ reduction observed in the TPR-XANES and TPR-EXAFS data. For clarity, Table S1 reports only the first-shell Ru-Ru coordination as a foil-referenced comparison, whereas Table 3 contains the full Ru-O and Ru-Ru model used for structural interpretation.

Figure 6 and Figure 7 present HR-TEM images of the unpromoted 1% Ru/m-ZrO₂ catalyst and the 1.8% Na-promoted counterpart, respectively. Both catalysts exhibit a high dispersion of Ru nanoparticles. For the unpromoted catalyst, the EDX-based particle size analysis shows that approximately 70% of the Ru nanoparticles are ≤1 nm, ~27% fall in the 1–2 nm range, and only 2.7% are between 2 and 3 nm, with the remainder present as larger agglomerates. This high dispersion maximizes the fraction of accessible Ru surface atoms and is therefore expected to enhance catalytic activity on a per-mass basis.

Upon Na addition, the Ru nanoparticles undergo a modest increase in size. For the 1.8% Na-promoted catalyst, ~5% of Ru nanoparticles are ≤1 nm, ~70% lie between 1 and 2 nm, ~17% are between 2 and 3 nm, and ~5% are in the 3–4 nm range, with the remainder forming agglomerates larger than 4 nm. Despite this slight growth, Ru remains highly dispersed. The Na promoter itself is also well distributed, with ~1% of Na species ≤1 nm, ~81% between 1 and 2 nm, ~15% between 2 and 3 nm, and only a small fraction present as larger agglomerates.

Spatial correlation analysis indicates that approximately 40–60% of Na species are co-located with Ru nanoparticles, suggesting a direct interaction between Na and Ru. Such proximity is consistent with the Na-induced modification of the electronic and adsorption properties of Ru, which can influence reaction selectivity. Overall, the TEM/EDX observations are in good agreement with the EXAFS results, confirming highly dispersed Ru nanoparticles and only a minor increase in particle size upon Na promotion.

CO₂-TPD profiles (Figure 8) were used to evaluate changes in surface basicity as Na was added to the Ru/m-ZrO₂ catalysts. Because CO₂ is an acidic probe molecule, stronger basic sites retain CO₂ more strongly, requiring higher temperatures for desorption. The unpromoted 1.0% Ru/m-ZrO₂ catalyst exhibits a broad distribution of basic sites, with nearly 48% of the CO₂ desorbing below 250 °C, 21% between 250 and 400 °C, and 31% above 400 °C (

Table 4. Results of fitting CO₂ TPD profiles with Gaussian peaks. Gaussian peak area percentages having maxima within various temperature ranges are shown. (BDL, below the detection limit).

Catalyst	% T < 250 °C	% 250 °C < T < 400 °C	% T > 400 °C
1.0% Ru/m-ZrO2	47.9	20.8	31.3
0.5% Na-1.0% Ru/m-ZrO2	31.0	30.9	38.1
1.0% Na-1.0% Ru/m-ZrO2	25.3	26.0	48.7
1.8% Na-1.0% Ru/m-ZrO2	4.1	12.2	83.6
2.5% Na-1.0% Ru/m-ZrO2	BDL	4.2	95.8
5.0% Na-1.0% Ru/m-ZrO2	BDL	BDL	100

). Upon adding Na, these proportions shift systematically toward higher temperatures. For example, at 0.5% Na, the low-temperature contribution drops to 31%, while high-temperature desorption increases to 38%, indicating that Na already strengthens a substantial fraction of basic sites. With 1.0% Na, the trend intensifies: only 25% of CO₂ desorbs below 250 °C, and nearly 49% desorbs above 400 °C. The shift becomes especially pronounced at 1.8% Na, where the low-temperature peak is almost completely eliminated (4%), and more than 83% of CO₂ desorbs above 400 °C. At even higher promoter levels (2.5% and 5.0% Na), the low- and mid-temperature features fall below the detection limits, and the desorption occurs exclusively at temperatures greater than 400 °C, with 96–100% of CO₂ released at the highest temperature range. This disappearance of low- and medium-strength basic sites indicates that at higher Na loadings, nearly all surface oxygen sites are converted into stronger carbonate-binding environments, consistent with the increased basicity inferred from DRIFTS and EXAFS. These changes are directly reflected in the shapes of the CO₂-TPD profiles in Figure 8, where the emergence and growth of the high-temperature peak dominate the spectra as Na loading increases. Together, the disappearance of low-temperature desorption and the strong growth of the high-temperature component demonstrate a clear, monotonic increase in catalyst basicity with Na addition. This is consistent with Na introducing or strengthening basic oxygen sites on the surface, resulting in progressively stronger CO₂ binding as the promoter concentration increases.

Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) during temperature-stepped methanol steam reforming provides insights into the evolution of surface intermediates and how the Na modification alters catalyst behavior. At 50 °C, the spectra for all catalysts (Figure 9, Figure 10, Figure 11 and Figure 12) are dominated by methoxy species formed by the dissociative adsorption of methanol on reduced sites of m-ZrO₂. The methoxy ν(OC) vibrations appear in the 1000–1200 cm⁻¹ region, and the ν(CH) modes are observed between 2800 and 3000 cm⁻¹. Even at this initial temperature, however, significant formate species are already present on the surface, as indicated by ν_asym(OCO) bands in the 1500–1650 cm⁻¹ region and ν_sym(OCO) bands in the 1300–1400 cm⁻¹ region. As the temperature increases, the methoxy features decrease and the formate-related bands intensify, indicating the progressive conversion of methoxy to formate. The formate ν(CH) vibration appears in the ~2850–2900 cm⁻¹ region and is clearly resolved in Figure 13.

Table 5. Formate ν(CH) band and ν(OCO) band positions observed in the range of 150–175 °C.

Catalyst	Main Formate ν(CH) Band	Main νasym(OCO) and νsym(OCO) Bands	Δ of Main Formate νasym(OCO) Minus Main νsym(OCO) Formate
1.0% Ru/m-ZrO2	2870	1579, 1376	203
w/0.5% Na	2864	1585, 1360	225
w/1.0% Na	2850	1616, 1348	268
w/1.8% Na	2790	1605, 1360	245
w/2.5% Na	2789	1605, 1359	246
w/5.0% Na	2792	1606, 1359	247

summarizes the band positions of the key formate vibrational modes for each catalyst. The unpromoted 1.0% Ru/m-ZrO₂ sample displays a main formate ν(CH) band at 2870 cm⁻¹, whereas the addition of Na causes this band to shift to lower wavenumbers. At 0.5% Na, the ν(CH) band moves to 2864 cm⁻¹, and at 1.0% Na, it shifts further to 2850 cm⁻¹. At higher promoter loadings (1.8–5.0% Na), the ν(CH) band appears near 2790–2792 cm⁻¹. The consistent red shift of this feature reflects the progressive destabilization of the formate C–H bond. This behavior is consistent with increased surface basicity, wherein Na strengthens the interaction between surface oxygen sites and the OCO moiety, leading to increased perturbation of the C–H bond and a corresponding decrease in vibrational frequency.

Further evidence of changing surface basicity is found in the splitting of the ν_asym(OCO) and ν_sym(OCO) formate bands. For the unpromoted catalyst, the split is 203 cm⁻¹. With Na addition, this increases, reaching 225 cm⁻¹ at 0.5% Na and a maximum of 268 cm⁻¹ at 1.0% Na, consistent with stronger perturbation of the OCO group due to increased basicity. At higher Na loadings (1.8–5.0%), however, splitting decreases slightly to Δ = 245–247 cm⁻¹. This subtle decrease correlates with the increasing intensity of carbonate-related bands at elevated temperatures in Figure 9, Figure 10, Figure 11 and Figure 12 and in the supplementary spectra (Figures S4 and S5). Highly basic surfaces promote the formation of polydentate carbonate species, which exhibit more delocalized O–C–O bonding and correspondingly smaller ν_asym–ν_sym separations compared to bidentate formates. Accordingly, the reduced band splitting at higher Na contents suggests a relative increase in multidentate carbonate-like species at the expense of bidentate formate on increasingly basic Na-modified surfaces.

As temperature increases above roughly 100 °C, the formate intermediates begin to decompose into carbonate species, which are precursors to CO₂ evolution. Carbonate features grow in the 1400–1700 cm⁻¹ region, and their earlier appearance and increasing prominence with Na loading indicate that Na accelerates formate decomposition. Figure 9, Figure 10, Figure 11 and Figure 12 show that formate bands decrease more rapidly with temperature as the Na concentration increases, consistent with the weakened formate C–H bond inferred from the ν(CH) shifts. By 400 °C, He purging (panels “p” in each figure) reveals residual surface carbonates characteristic of m-ZrO₂, although their persistence and intensity vary with the Na content. Overall, the DRIFTS results show that Na promotion increases surface basicity, modifies the vibrational signatures of formate and carbonate intermediates, and accelerates formate decomposition during methanol steam reforming.

Figure 14 shows the temperature-programmed evolution of H₂ during methanol steam reforming from pre-adsorbed CH₃OH and H₂O. For the unpromoted 1.0% Ru/m-ZrO₂ catalyst, H₂ evolution begins at a moderate temperature and reaches its maximum at ~300–325 °C. With Na addition, however, the onset of H₂ release shifts to lower temperatures, and the total H₂ signal increases. As the Na loading increases, this shift becomes more pronounced, with the most heavily promoted catalysts exhibiting earlier and higher H₂ evolution peaks.

This behavior reflects Na-induced changes in the decomposition pathways of surface formate species. On the unpromoted catalyst, formate decomposition proceeds through a mixture of decarbonylation/dehydration pathways (–H + –OOCH → CO + H₂O), which do not release H₂, and decarboxylation/dehydrogenation pathways (–H + –OOCH → CO₂ + H₂), which do release H₂. Adding Na increases the basicity of the catalyst surface, strengthening the interaction of the formate OCO fragment with nearby oxygen sites. As supported by the DRIFTS data (Figure 9, Figure 10, Figure 11, Figure 12 and Figure 13, Figures S4 and S5, and Table 5), increased basicity weakens the formate C–H bond and biases formate decomposition toward the dehydrogenation/decarboxylation pathway.

As a result, the Na-promoted catalysts produce more H₂ and do so at lower temperatures, consistent with the earlier loss of formate bands in DRIFTS spectra and the increased ν(CH) red-shifts associated with C–H bond weakening. At higher Na loadings, the effect is stronger, with the H₂ evolution peak shifting steadily to lower temperatures and increasing in intensity, matching the trend highlighted by the dashed red guides in Figure 14. This earlier onset of H₂ evolution aligns with the DRIFTS-observed weakening of the formate C–H bond (Figure 13), confirming that Na enhances the dehydrogenation/decarboxylation route. This shift in formate reaction selectivity is important for methanol utilization as a hydrogen-carrying liquid. By redirecting formate decomposition toward pathways that release H₂ and suppress pathways that yield CO, Na-promoted catalysts can increase the hydrogen yield from methanol and could be beneficial for applications such as feeding H₂ to proton exchange membrane (PEM) fuel cells.

Catalytic methanol steam reforming (MSR) performance was evaluated at 1 atm using a 1:1 CH₃OH:H₂O feed under identical space velocities (GHSV = 38,100 h⁻¹ for the Ru catalysts), with selectivities reported on a %C basis (

Table 6. Selectivities during MSR at the same space velocity at various temperatures without co-fed H₂ and using a 1:1 CH₃OH:H₂O ratio. Process conditions: P = 1 atm; GHSV = 38,100 h⁻¹ for the Ru catalysts; 2.9% CH₃OH/10.8% N₂/2.9% H₂O/83.4% He. Notes: All selectivities are reported on a % C basis. For the gray highlighted row, the catalyst was reduced at a lower temperature of 325 °C.

Catalyst Description	T (°C)	% CH3OH Conversion	Relative Apparent TOF Estimate (275 °C Case)	% CH4 Selectivity	% CO Selectivity	% CO2 Selectivity
1% Ru/m-ZrO2	325	80.1	-	0.8	87.3	12.0
0.5% Na-1% Ru/m-ZrO2	325	80.6	-	0.8	83.0	16.2
1.0% Na-1% Ru/m-ZrO2	325	80.0	-	0.8	78.2	21.0
1.8% Na-1% Ru/m-ZrO2	325	78.1	-	1.4	72.1	26.5
2.5% Na-1% Ru/m-ZrO2	325	71.2	-	1.4	72.4	26.2
5% Na-1% Ru/m-ZrO2	325	47.4	-	0.9	78.6	20.5
1% Ru/m-ZrO2	300	65.5	-	0.3	91.1	8.6
0.5% Na-1% Ru/m-ZrO2	300	69.4	-	0.3	89.1	10.6
1.0% Na-1% Ru/m-ZrO2	300	69.4	-	0.3	86.4	13.3
1.8% Na-1% Ru/m-ZrO2	300	55.5	-	0.5	81.0	18.5
2.5% Na-1% Ru/m-ZrO2	300	51.8	-	0.8	80.7	18.5
5% Na-1% Ru/m-ZrO2	300	26.7	-	0.7	78.9	20.4
1% Ru/m-ZrO2	275	42.1	1.00	0.1	91.9	8.0
0.5% Na-1% Ru/m-ZrO2	275	35.1	1.08	<0.1	87.6	12.4
1.0% Na-1% Ru/m-ZrO2	275	35.8	0.93	<0.1	84.6	15.4
1.8% Na-1% Ru/m-ZrO2	275	29.1	0.74	1.0	80.2	18.8
2.5% Na-1% Ru/m-ZrO2	275	26.3	0.75	0.4	83.6	16.0
5% Na-1% Ru/m-ZrO2	275	11.0	0.30	0.3	79.2	20.6

and

Table 7. Selectivities during MSR at similar conversion levels (GHSV was varied to match conversion) at two temperatures using a 1:1 CH₃OH:H₂O ratio. Process conditions: P = 1 atm; 2.9% CH₃OH/10.8% N₂/2.9% H₂O/83.4% He. Note: All selectivities are reported on a % C basis.

Catalyst Description	T (°C)	% CH3OH Conversion	% CH4 Selectivity	% CO Selectivity	% CO2 Selectivity
1% Ru/m-ZrO2	325	80.1	0.8	87.3	12.0
0.5% Na-1% Ru/m-ZrO2	325	80.6	0.8	83.0	16.2
1.0% Na-1% Ru/m-ZrO2	325	80.0	0.8	78.2	21.0
1.8% Na-1% Ru/m-ZrO2	325	78.1	1.4	72.1	26.5
1% Ru/m-ZrO2	300	65.5	0.3	91.1	8.6
0.5% Na-1% Ru/m-ZrO2	300	69.4	0.3	89.1	10.6
1.0% Na-1% Ru/m-ZrO2	300	69.4	0.3	86.4	13.3
1.8% Na-1% Ru/m-ZrO2	300	67.6	0.8	75.4	23.7

). At 325 °C, the unpromoted 1% Ru/m-ZrO₂ catalyst achieves 80.1% methanol conversion, producing primarily CO (87.3% selectivity) and a smaller fraction of CO₂ (12.0% selectivity). Adding Na modifies the product distribution. At low Na loadings (0.5–1.0% Na), methanol conversion remains similar (80.6% and 80.0%, respectively), but CO₂ selectivity increases to 16.2–21.0%, with a corresponding decrease in CO selectivity to 78–83%. At higher Na contents (1.8–2.5% Na), methanol conversion decreases moderately (≈71–78%), but CO₂ selectivity increases further to 26–26.5%, indicating a shift toward the dehydrogenation/decarboxylation route. The highest Na loading (5% Na) yields the lowest conversion (47.4%), but CO₂ selectivity remains elevated (20.5%) relative to the unpromoted catalyst.

At 300 °C, similar trends are observed. The unpromoted catalyst converts 65.5% of methanol with 91.1% CO selectivity. Low-Na catalysts show slightly higher conversion (≈69%) and increased CO₂ selectivity (≈10–13%), while CO selectivity decreases proportionally. At intermediate Na loadings (1.8–2.5%), methanol conversion drops to ~52–55%, and CO₂ selectivity increases to ~18–20%. The 5% Na catalyst again shows the lowest conversion (26.7%), but maintains the characteristic Na-induced increase in CO₂ selectivity (20.4%).

At the lowest temperature examined, 275 °C, the behavior is consistent: Na-free Ru/m-ZrO₂ achieves 42.1% conversion with 91.9% CO selectivity. Low Na loadings maintain similar conversion (~35–43%) but shift selectivity toward CO₂ (up to 12.4%). At 1.8–2.5% Na, conversion decreases to ~26–28%, and CO₂ selectivity increases to ~18–19%. The 5% Na catalyst again shows the lowest conversion (11.0%) but elevated CO₂ formation (20.6%).

For this analysis, apparent turnover frequencies (TOFs) were estimated using low-conversion methanol steam reforming data obtained at 275 °C (Table 6), where all catalysts were evaluated under identical gas hourly space velocity and feed composition. Surface Ru dispersion, which is proportional to accessible Ru atoms, was estimated from EXAFS-derived Ru particle sizes by assuming spherical nanoparticles and constant Ru site accessibility across the catalyst series, which serves as a simplifying assumption. Apparent TOFs are reported relative to the unpromoted Ru/m-ZrO₂ catalyst to emphasize comparative trends rather than absolute kinetic values. Surface Ru dispersions were estimated from the particle size analysis rather than from chemisorption measurements, as hydrogen uptake on Ru/ZrO₂ systems is dominated by spillover onto the support. Consequently, the apparent TOFs should be regarded as approximate, since possible Na-induced changes in Ru site accessibility, interfacial structure, and surface coverage are not explicitly accounted for, and are therefore used only for qualitative comparison.

The estimated apparent TOFs show no enhancement of intrinsic Ru activity with Na addition and instead decrease gradually with increasing Na loading. Importantly, the magnitude of this decrease is modest at low-to-moderate Na loadings and is small relative to the much larger changes observed in CO₂ selectivity. This decoupling indicates that Na promotion primarily alters reaction pathways and surface coverage rather than increasing the intrinsic turnover capability of Ru sites. Partial Na decoration or masking of Ru and Ru–support interfacial sites may reduce the number of accessible Ru sites across the catalyst series such that the intrinsic TOFs per truly accessible site could be more similar than the apparent values suggest. Accordingly, the apparent TOF trends are interpreted as reflecting effective catalytic behavior under Na-modified surface conditions rather than definitive intrinsic site kinetics.

At a high Na loading (5 wt.% Na), the catalyst surface is already highly basic and strongly populated by stabilized oxygenated species at low temperature, as evidenced by CO₂-TPD and DRIFTS. Under these conditions, the reaction pathway is biased toward CO₂ formation already at low temperature, resulting in high CO₂ selectivity with little further temperature dependence. The accompanying suppression of methanol conversion and lower apparent TOF may arise from a combination of effects, including Na decoration of Ru or Ru–support interfacial sites, strong stabilization of oxygenated surface species, and restricted accessibility or dynamic regeneration of active ensembles, all of which reduce effective catalytic turnover under these highly basic conditions.

To isolate the effect of Na on selectivity independent of activity, Table 7 compares catalysts at similar conversion levels by adjusting the space velocity. At 325 °C, all catalysts were evaluated at ~78–80% conversion. Under these matched-conversion conditions, Na increases CO₂ selectivity from 12.0% (unpromoted) to 16.2% (0.5% Na), 21.0% (1.0% Na), and 26.5% (1.8% Na), with corresponding decreases in CO selectivity. A similar trend is observed at 300 °C, where at ~66–69% conversion, CO₂ selectivity rises from 8.6% (unpromoted) to 23.7% at 1.8% Na. These matched-conversion data confirm that the higher CO₂ selectivity obtained with Na loading is not simply due to lower conversion at higher Na levels, but reflects a real shift in the reaction pathway.

Catalyst stability was assessed at ~317–325 °C for extended time-on-stream (Figure 15 and Figure 16). The unpromoted 1% Ru/m-ZrO₂ catalyst (Figure 15) maintains stable methanol conversion (~88%) and steady CO, CO₂, and CH₄ selectivities over 360 min. The 1.8% Na-promoted catalyst (Figure 16) also shows stable CH₄ and CO₂ selectivities, though methanol conversion exhibits a gradual decline over the first ~100 min before stabilizing. Even under these mildly deactivating conditions, the Na-promoted catalyst maintains higher CO₂ selectivity than the unpromoted catalyst at comparable times.

Overall, the reaction testing data show that Na shifts MSR selectivity away from CO and toward CO₂ across a wide temperature range. This effect persists even when conversions are matched, demonstrating that Na modifies the decomposition pathway of surface intermediates rather than simply lowering activity. Taken together with the DRIFTS evidence of formate C–H destabilization, the catalytic data show that Na favors H₂-producing dehydrogenation/decarboxylation pathways (→CO₂ + H₂) while suppressing CO-forming decarbonylation processes (→CO + H₂O).

The combined TPR, EXAFS, CO₂-TPD, DRIFTS, and catalytic testing results allow construction of a consistent mechanistic picture for how Na modifies methanol steam reforming (MSR) over Ru/m-ZrO₂. The overall pathway involves sequential methoxy → formate → carbonate intermediates on the m-ZrO₂ support, with Ru providing sites for H₂ formation and spillover. Na affects multiple steps in this sequence.

During CH₃OH activation (Step 1 in Figure 17), methanol dissociates at reduced ZrO₂ sites to form surface methoxy species, which rapidly convert into formate even at low temperature, as observed using DRIFTS. Without Na, the formate species can decompose through two competing reactions: dehydration/decarbonylation (producing CO + H₂O) or dehydrogenation/decarboxylation (producing CO₂ + H₂). The DRIFTS ν(CH) formate frequencies and ν_asym–ν_sym(OCO) splitting demonstrate that Na increases the basicity of the surface, which strengthens binding of the OCO fragment and weakens the formate C–H bond. This perturbation stabilizes formate in a geometry that favors dehydrogenation/decarboxylation (blue pathway in Figure 17) and disfavors the CO-producing route (red pathway). The emergence of the long Ru-O coordination at ≥1 wt.% Na further supports the interfacial oxygen stabilization illustrated in Figure 17, reinforcing Na’s role in modifying the Ru–support environment.

In Step 2, the weakened C–H bond of formate at Na-modified sites facilitates dehydrogenation to HCOO* and H*, which readily recombine on Ru to form H₂. This is consistent with transient MSR data (Figure 15), which show that Na-promoted catalysts release H₂ at lower temperatures and in greater amounts, reflecting the stabilization of the H₂-forming pathway. CO₂-TPD also shows that Na increases the strength and population of basic surface sites, supporting this mechanistic shift.

In Step 3, Na also influences how the oxygenated intermediates bind to the support. The EXAFS analysis reveals a distinctive long Ru–O interaction (2.1–2.2 Å) that appears only in Na-containing catalysts. This suggests that Na modifies the Ru–support interface, introducing or stabilizing oxygen-containing species that interact with Ru. These modified interfacial sites may further favor formate orientations and O-binding motifs associated with the dehydrogenation/decarboxylation pathway. This structural change also correlates with the observed increase in CO₂ selectivity across temperatures and reaction conditions (Table 6 and Table 7).

As formate decomposes to CO₂, carbonate species form on the ZrO₂ support (Step 4). DRIFTS data show that these carbonates are more strongly bound when Na is present, shifting carbonate ν(OCO) bands and increasing their thermal stability. During H₂-TPR, hydrogen spillover decomposes these carbonates, and Na influences both the temperature and manner in which this occurs. Defect-associated hydroxyl groups observed in DRIFTS spectra after H₂ exposure (Figure 4) confirm that hydrogen spillover and carbonate removal are linked, and that Na modifies these processes as well.

Overall, the mechanistic picture is that Na increases surface basicity, stabilizes oxygen-rich environments at the Ru–support interface, weakens the formate C–H bond, and shifts the reaction pathway toward H₂-forming dehydrogenation/decarboxylation rather than the CO-forming route. This selective promotion of the blue pathway in Figure 17 leads to greater H₂ production from methanol and higher CO₂ selectivity, aligning with the steady-state, transient, DRIFTS, and structural characterization results.

The more modest effect of Na on Ru/m-ZrO₂ compared to Na-doped Pt/YSZ arises from intrinsic differences in the electronic structures of Ru and Pt that govern their reactivity. Pt has a more filled and lower-energy d-band, which results in weaker M–O and M–CO interactions and makes Pt far more sensitive to electronic perturbation by alkali promoters. As shown in the Na–Pt/YSZ study, Na substantially weakens the formate C–H bond on Pt, enabling >90% CO₂ selectivity at 2.5 wt.% Na. Ru, in contrast, has a higher-energy, less filled d-band that binds oxygenated intermediates and CO much more strongly. This makes C–O bond scission intrinsically easier on Ru and limits how much Na can shift the reaction pathway. Consistent with this, the Na-Ru/m-ZrO₂ catalysts exhibit smaller ν(CH) red shifts, only modest interfacial Ru–O perturbation in EXAFS spectra, and maximum CO₂ selectivities of ~20–27% at matched conversion. Thus, while Na pushes both metals toward decarboxylation, Ru’s electronic structure inherently restricts the extent to which Na can suppress the CO-forming pathway.

3. Materials and Methods

3.1. Catalyst Synthesis

The 1.0 wt.% Ru/m-ZrO₂ catalyst was synthesized via incipient wetness impregnation using monoclinic zirconia as the support (Thermo Scientific, Waltham, MA, USA). Commercial ZrO₂ pellets (1/8″) were crushed and sieved to obtain particles in the 63–106 µm size range prior to use. An aqueous solution of ruthenium nitrosyl nitrate (Alfa Aesar, Haverhill, MA, USA) was added dropwise to the support until incipient wetness was achieved. The impregnated material was dried and subsequently calcined in static air at 350 °C for 4 h in a muffle furnace (Thermolyne, Thermo Scientific).

The resulting Ru/m-ZrO₂ material served as a parent catalyst and was subdivided to prepare a series of Na-promoted samples. Sodium was introduced at nominal loadings of 0.50, 1.0, 1.8, 2.5, and 5.0 wt.% using sodium nitrate (Alfa Aesar) as the precursor. Sodium addition was carried out using a second incipient wetness impregnation step from aqueous NaNO₃ solutions, followed by drying and calcination in air at 350 °C for 4 h under identical conditions.

3.2. Textural Characterization by N₂ Physisorption

Specific surface areas and porosity characteristics were determined by performing nitrogen adsorption–desorption measurements using a Micromeritics 3-Flex analyzer (Norcross, GA, USA). Prior to analysis, the samples were degassed under vacuum (<6.7 Pa) at 160 °C for 12 h to remove physisorbed species. Adsorption and desorption isotherms were subsequently recorded at liquid nitrogen temperature. Surface areas were calculated using the Brunauer–Emmett–Teller (BET) method, while pore volumes and pore size distributions were derived from the desorption branch employing the Barrett–Joyner–Halenda (BJH) model.

3.3. H₂ Temperature-Programmed Reduction and TPR-MS

Temperature-programmed reduction (TPR) experiments were carried out using an Altamira AMI-300R system (Altamira Instruments, Twin Lakes, WI, USA) equipped with a thermal conductivity detector (TCD). Approximately 150 mg of catalyst was loaded into the reactor, with the thermocouple positioned directly within the catalyst bed. A reducing gas mixture consisting of 10 vol.% H₂ in Ar (Airgas, San Antonio, TX, USA, UHP grade) was passed through the reactor at a flow rate of 30 cm³ min⁻¹. The temperature was increased from 30 to 1000 °C at a linear ramp rate of 10 °C min⁻¹.

For selected experiments, the reactor outlet was connected to a quadrupole mass spectrometer (Hiden Analytical, Warrington, UK) to monitor evolved gas species during reduction. All gases used in TPR and TPR-MS experiments were supplied by Airgas (San Antonio, TX, USA).

3.4. CO₂ Temperature-Programmed Desorption with Mass Spectrometry

CO₂-TPD measurements were performed using the same Altamira AMI-300R instrument coupled to a Hiden quadrupole mass spectrometer. Prior to CO₂ adsorption, the catalyst was reduced at 400 °C for 1 h under a flow of 10 cm³ min⁻¹ H₂ diluted with 20 cm³ min⁻¹ Ar. The sample was then purged in flowing Ar (30 cm³ min⁻¹) for 20 min and cooled to 50 °C.

CO₂ adsorption was carried out by exposing the sample to 25 cm³ min⁻¹ of 4 vol.% CO₂ in helium for 15 min. Following saturation, the system was purged briefly, and the temperature was ramped to 1000 °C at 10 °C min⁻¹ under a helium flow of 30 cm³ min⁻¹. CO₂ desorption was monitored by tracking the m/z = 44 signal during the temperature ramp.

3.5. High-Resolution Transmission Electron Microscopy and EDX

High-resolution transmission electron microscopy (HR-TEM) and scanning transmission electron microscopy (STEM) analyses were conducted using a JEOL JEM-2010F field-emission transmission electron microscope (JEOL, Peabody, MA, USA). Imaging was performed at accelerating voltages of 120 or 200 kV, enabling lattice-resolved imaging with a point-to-point resolution of approximately 1.9 Å.

Before microscopy, catalyst samples were activated at 350 °C for 1 h under flowing 33 vol.% H₂ in He (30 cm³ min⁻¹), followed by purging and cooling in He. The samples were then passivated in 1 vol.% O₂ in He for 10 h. Activated catalysts were finely ground, dispersed in ethanol via brief ultrasonication (~10 min), and deposited onto lacey carbon-coated copper grids.

Bright-field TEM, HR-TEM, and annular dark-field (ADF) STEM images were acquired using a TVIPS CMOS camera system (TVIPS GmbH, Gilching/Gauting, Germany). Elemental analysis and mapping were performed with an EDAX Genesis SiLi energy-dispersive X-ray detector (EDAX, Inc., Pleasonton, CA, USA). The data were processed using TVIPS EMMenu (v4.0) and EDAX Genesis (v5.2) software.

3.6. H₂-TPR-XANES and EXAFS Measurements

In situ H₂-TPR-XANES and EXAFS experiments were conducted at the MR-CAT beamline of the Advanced Photon Source (Argonne National Laboratory). Incident X-ray energies were selected using a Si(111) monochromator (Kohzu Precision Co., Ltd., Kawasaki City, Japan) in combination with a rhodium-coated mirror to suppress higher-order harmonics. The overall experimental configuration was adapted from a previously described setup reported by Jacoby [24].

Temperature-programmed reduction measurements were performed using a stainless-steel multi-sample holder designed to accommodate six samples simultaneously, with each channel having an inner diameter of 3 mm. Samples were prepared as self-supporting wafers using 21–23 mg of material pressed from a 60/40 wt.% mixture of the catalyst and SiO₂. This composition and mass were selected to optimize data quality for Ru K-edge measurements on ZrO₂-supported catalysts. The multi-sample holder was positioned inside a clamshell furnace mounted on a precision translation stage.

The samples were enclosed within a quartz reactor tube fitted with Kapton windows for X-ray transmission and equipped with gas inlet/outlet ports and a thermocouple for temperature monitoring. Alignment of the samples relative to the incident X-ray beam was achieved with an accuracy of approximately 20 µm. After alignment, the reactor was flushed with helium at a flow rate of 100 mL min⁻¹ for a minimum of 5 min prior to introducing the reducing gas mixture. The gas feed was then switched to 25 vol.% H₂ balanced with helium, maintaining the same total flow rate. The temperature was increased at a controlled rate of 1.0 °C min⁻¹ to a final temperature of 400 °C.

X-ray absorption spectra at the Ru K-edge were collected in transmission mode, with a metallic Ru foil placed downstream of the samples for simultaneous energy calibration. EXAFS data processing was carried out using the WinXAS software package (version 2.0, Thorsten Ressler, Berlin, Germany) [25]. Structural modeling and fitting were performed using the Atoms code (version 2.46b) [26] together with the FEFF/FEFFIT suite (version 2.54) [27,28]. The quantitative analysis focused on spectra acquired after the completion of the temperature ramp and subsequent cooling under flowing hydrogen. Fits were conducted over k-space and R-space ranges of 2.75–10 Å⁻¹ and 1.5–3.0 Å, respectively.

3.7. DRIFTS Analysis of Methanol Steam Reforming

Diffuse reflectance infrared measurements were carried out using a Nicolet iS-10 Fourier transform infrared spectrometer (Thermo Scientific, Waltham, MA, USA) equipped with a Praying Mantis high-temperature reaction cell (Harrick Scientific, Mt. Kisco, NY, USA). The experiments were designed to monitor surface species during temperature-programmed methanol steam reforming under well-defined flow conditions.

Initial reference spectra (512 scans) were collected for the calcined catalyst under a helium flow of 100 cm³ min⁻¹ at room temperature. The catalyst was then reduced at 400 °C for 1 h using a 1:1 mixture of H₂ and He at a total flow rate of 200 cm³ min⁻¹, after which a background spectrum consisting of 512 scans was recorded. Following reduction, the catalyst was purged with helium (100 cm³ min⁻¹) at 400 °C, cooled to 50 °C under flowing helium, and an additional background spectrum (512 scans) was obtained.

Methanol adsorption was achieved by bubbling helium (75 cm³ min⁻¹) through liquid methanol for approximately 15 min. Subsequently, gas-phase methanol and weakly adsorbed species were removed by purging with 100 cm³ min⁻¹ of helium, after which a spectrum comprising 512 scans was collected. Water vapor was introduced by passing helium at 30 cm³ min⁻¹ through a deionized water saturator maintained at 31 °C, corresponding to an H₂O concentration of 4.4 vol.% at the same flow rate. The introduction of water vapor led to the reaction of adsorbed methoxy species with H₂O to form surface formates, followed by their subsequent decarbonylation and/or decarboxylation.

The sodium content of the catalysts was varied between 0 and 5.0 wt.%, and spectra were collected as the temperature was increased stepwise from 50 to 400 °C in increments of 25 °C. All gases used in these experiments were supplied by Airgas (San Antonio, TX, USA).

3.8. Temperature-Programmed Surface Reaction of Methanol Steam Reforming

Temperature-programmed surface reaction experiments were conducted using an Altamira AMI-300R unit (Altamira Instruments, Twin Lakes, WI, USA) equipped with a quadrupole mass spectrometer (Hiden Analytical, Warrington, UK) for online gas analysis. Prior to the reaction, catalysts were activated by heating to 400 °C under a flow of 33 vol.% H₂ in argon (30 cm³ min⁻¹). After activation, the catalysts were purged in flowing argon for 20 min and subsequently cooled to 50 °C.

Surface saturation with methanol was achieved by injecting 100 µL of liquid methanol directly into the reactor, followed by purging with argon at 30 cm³ min⁻¹ for 15 min to remove gas-phase and weakly adsorbed species. The water delivery system was purged with argon prior to use. Water vapor was then introduced by bubbling argon at 30 cm³ min⁻¹ through the water saturator for 10 min, after which the catalysts were again purged in argon (30 cm³ min⁻¹) for 15 min.

Following these surface preparation steps, the catalysts were heated from 50 °C to 1000 °C at a linear ramp rate of 10 °C min⁻¹. Hydrogen evolution during the temperature ramp was monitored by mass spectrometry to evaluate the influence of sodium promoter loading on methanol dehydrogenation behavior.

3.9. Fixed-Bed Reactor Evaluation

Catalytic performance was evaluated in a continuous-flow stainless-steel tubular reactor operated under steady-state conditions. The reactor had an internal diameter of 0.42 in. and contained a fixed catalyst bed. For each experiment, 400 mg of catalyst particles (60–90 µm) were mechanically mixed with 800 mg of SiO₂ beads of a comparable particle size to improve bed packing and heat transfer.

Prior to the reaction, the catalyst bed was activated under a reducing atmosphere consisting of 50 vol.% H₂ in helium flowing at 150 cm³ min⁻¹. The temperature was increased to 400 °C at a heating rate of 3 °C min⁻¹ and maintained for 30 min. After activation, the feed was switched to a reactant mixture containing 2.9 vol.% methanol, 2.9 vol.% water vapor, 10.8 vol.% nitrogen, and helium comprising the remainder. Reactions were carried out at atmospheric pressure over a temperature range of 275–325 °C, corresponding to a gas hourly space velocity (GHSV) of 38,100 h⁻¹.

The reactor effluent was passed through a cold trap maintained at 0 °C to remove condensable products prior to analysis. The remaining gas-phase components were analyzed online using an SRI 8610 gas chromatograph (SRI Instruments, Torrance, CA, USA). The GC was equipped with a 3.658 m silica gel-packed column and a 1.829 m molecular sieve-packed column, along with both thermal conductivity (TCD) and flame ionization (FID) detectors. A built-in methanizer was employed to increase the sensitivity of the FID toward carbon monoxide and carbon dioxide.

Apparent turnover frequencies (TOFs) were estimated for comparative purposes using the methanol steam reforming data obtained at 275 °C, where all catalysts were evaluated in a fixed-bed plug-flow reactor with an identical gas hourly space velocity, feed composition, and catalyst mass. Under these low-to-moderate conversion conditions, methanol conversion was used as a proxy for the bed-averaged reaction rate, enabling a qualitative comparison of apparent TOFs across the catalyst series.

Surface Ru site densities were estimated from EXAFS-derived Ru particle sizes, assuming spherical nanoparticles and a uniform particle size. Site densities were not determined by chemisorption, as hydrogen uptake on Ru/ZrO₂ systems is dominated by spillover to the support. To minimize uncertainties associated with absolute site counting and plug-flow integration effects, TOFs were reported relative to the unpromoted Ru/m-ZrO₂ catalyst. The resulting values represent apparent, PFR-averaged TOFs and are intended only for a qualitative, comparative interpretation, as possible Na-induced changes in Ru site accessibility and surface coverage are not explicitly accounted for.

Surface Ru dispersion was estimated from EXAFS-derived Ru particle sizes assuming spherical nanoparticles and a uniform particle size. Ru dispersion ($D_i$) was approximated using the following geometric model:

$$D_i={\displaystyle \frac{6}{{\rho }_{\mathrm{Ru}}\hspace{0.17em}{d}_i}}$$

where $d_i$ is the average Ru particle diameter obtained from the EXAFS analysis and ${\rho }_{\mathrm{Ru}}$ is the density of metallic Ru (12.37 g cm⁻³). Because Ru loading was constant across the catalyst series, differences in dispersion reflect relative differences in the number of surface-accessible Ru atoms.

Hydrogen chemisorption was not used to estimate Ru dispersion or site density, as hydrogen uptake on Ru/ZrO₂ systems may be dominated by spillover to the support. Dispersion estimates derived from the particle size analysis were therefore used to provide a relative basis for comparing apparent catalytic behavior under identical plug-flow reactor conditions.

4. Conclusions

Na promotion of Ru/m-ZrO₂ modifies both the physicochemical properties of the catalyst and the reaction pathway for methanol steam reforming. Structural characterization shows that Na stabilizes Ru–O interactions and increases surface basicity, shifting RuO_x reduction and H₂ spillover to higher temperatures. Spectroscopic measurements confirm that Na increases the strength and population of basic sites, weakens the formate C–H bond, and stabilizes carbonate and OH surface species, consistent with a more strongly oxygen-binding environment.

Catalytic testing demonstrates that Na shifts methanol steam reforming selectivity away from CO and toward CO₂ across all temperatures. Even when conversions are matched, CO₂ selectivity increases substantially with Na addition. At approximately 80% conversion and 325 °C, CO₂ selectivity increases from 12.0% for the unpromoted catalyst to 16.2% at 0.5% Na, 21.0% at 1.0% Na, and 26.5% at 1.8% Na. At roughly 66–69% conversion and 300 °C, CO₂ selectivity increases from 8.6% for the unpromoted sample to 23.7% at 1.8% Na. Transient reaction experiments further show that Na-promoted catalysts release H₂ at lower temperatures and in greater amounts, consistent with enhanced dehydrogenation and decarboxylation of surface formate.

Collectively, the results point to a consistent mechanism in which Na increases surface basicity, modifies the Ru–support interface, destabilizes formate C–H bonds, and selectively favors the H₂-producing CO₂ pathway over CO formation. The extent of this selectivity shift is inherently limited by the electronic structure of Ru; under MSR conditions, its higher-energy, less occupied d-band strongly stabilizes CO-forming surface ensembles and oxygenated intermediates, making the CO-forming pathway more resistant to suppression even when Na is present. As a result, Na-modified Ru/m-ZrO₂ improves the H₂ yield and lowers CO selectivity, but the degree of tunability is intrinsically lower than what can be achieved with Na-promoted Pt catalysts.