Addendum: Au/ZSM-5 catalyses the selective oxidation of CH₄ to CH₃OH and CH₃COOH using O₂

Addendum to: Nature Catalysis https://doi.org/10.1038/s41929-021-00725-8, published online 6 January 2022.

In our article “Au/ZSM-5 catalyses the selective oxidation of CH₄ to CH₃OH and CH₃COOH using O₂” we demonstrated that Au nanoparticles supported on H-ZSM-5 were active for the selective oxidation of methane and that the surface of the Au nanoparticles was the active site for the reaction¹. At that time we utilized HAADF-STEM imaging, which effectively distinguishes heavy atoms such as Au from the support. However, due to limited dynamic range of the HAADF imaging, we did not notice there could be Fe on the Au particles. Recently, we detected the presence of Fe on the catalysts, prompting us to conduct X-ray EDS mapping on the used catalysts. In this addendum we explore the effect of Fe on the catalysis and show that although incomplete iron oxide layers do form on the Au nanoparticles their presence does not alter the original conclusions and findings of our study.

Presence of iron oxide overlayers on Au/ZSM-5 during reaction

In Fig. 3 of the original paper, we utilized high-angle annular dark field (HAADF) imaging to demonstrate the absence of obvious particle agglomeration in the Au/ZSM-5 catalyst following its use in the methane oxidation reaction. However, it was only later that we realized the Au particles in the used catalysts were partially coated—a detail that was not apparent from the HAADF images due to their limited dynamic range. This coating was, however, visible in bright field (BF) images and detectable through spectroscopy techniques in electron microscopy, such as X-ray energy dispersive spectroscopy (X-EDS) and electron energy loss spectroscopy (EELS).

Representative HAADF images, BF images, and EELS elemental maps overlaid with corresponding annular dark field (ADF) images of the fresh and used Au/ZSM-5 catalysts are shown in Fig. 1. The Au particles in the fresh catalyst (Fig. 1a–c, Supplementary Figs. 1 and 2) are similar to those depicted in Fig. 3 of the original paper. Remarkably, after just 5 min of use in the methane oxidation reaction, a visible coating layer on the Au particles became evident in the BF images (Fig. 1e). In most instances, this coating layer appear to only partially cover the Au particles (Supplementary Fig. 3). X-EDS (Supplementary Fig. 4) and EELS (Fig. 1f, Supplementary Figs. 5 and 6) analysis revealed that this coating layer consistently contains Fe oxides, with occasional detection of other elements such as Cr, Ni, and Ti. Notably, these oxides preferentially coat the Au particles and are not detected on the bare support (Supplementary Fig. 7). Extending the reaction time to 120 min did not significantly alter the coating layers (Fig. 1g–i). Most particles remained partially coated (Supplementary Fig. 8), and the elemental analysis yielded similar results (Fig. 1i and Supplementary Figs. 9 and 10). We consider that the Fe and the other minor elements (Cr, Ni and Ti) are leached from the body of the stainless-steel batch reactor and stirrer under the reaction conditions we employ, i.e., water as solvent with a reaction temperature of 240 °C.

Fig. 1: STEM-HAADF and STEM-BF images with corresponding EELS mapping of Fe and O (overlaid with ADF image) of Au/ZSM-5 catalysts.

a–c, fresh Au/ZSM-5. d–f, after 5 min of reaction. g–i, after 120 min of reaction. j–l, Au/ZSM-5 with 300 ppm Fe(NO₃)₃ after 5 min of reaction. Yellow arrows highlight the amorphous overlayer on the crystalline Au particles in the bright field images.

Effect of Fe addition on methane oxidation catalysis

We compared the catalytic performance of methane oxidation over the Au/ZSM-5 catalyst with and without an additional Fe³⁺ source (Fe(NO₃)₃ or FeCl₃) (Fig. 2, Supplementary Table 1). In a 5 min reaction without any added Fe³⁺, the catalyst produced a liquid-phase oxygenate yield of 7.84 µmol/g_cat with 90 % selectivity. Extending the reaction to 120 min increased the yield to 20.3 µmol/g_at but reduced selectivity to 65 % (Supplementary Table 1, entries 1 and 2). The addition of Fe³⁺ sources, either as Fe(NO₃)₃ or FeCl₃, and at varying amounts (0.03 wt.% to 0.10 wt. %), did not significantly affect the liquid-phase oxygenates yields after 5 min of methane oxidation (Fig. 2a and Supplementary Table 1, entries 3, 5, 7, and 9). Although the addition of Fe³⁺ slightly increased gas-phase CO₂ yield, it resulted in a marginal decrease in liquid-phase oxygenate selectivity. Similarly, changing the Fe³⁺ source and amount has no obvious influence on the catalytic performance after 120 min of methane oxidation (Supplementary Fig. 2b and Table 1, entries 4, 6, 8, and 10). The liquid-phase oxygenates yield averages around 18.5 μmol/g catalyst for the reactions with Fe³⁺ addition, which is less than 10% lower than the value for the 120-min reaction without Fe³⁺ addition. The liquid-phase oxygenates selectivity and gas-phase CO₂ yield remains unchanged between the reactions with and without Fe³⁺ addition.

Fig. 2: Effect of Fe³⁺ precursor type and amount on methane oxidation performance of Au/ZSM-5.

a,b, Yield of product measured after 5 min (a) and 120 min (b) with varying amounts of either Fe(NO₃)₃ or FeCl₃ as the Fe³⁺ source. Reaction conditions: catalyst (0.1 g), Fe³⁺ (0.03 or 0.10 wt.% of catalyst), H₂O or Fe³⁺ solution (15 mL), CH₄ (20.7 bar), O₂ (3.5 bar), temperature 240°C. The wt.% Fe reported represents the weight percentage of iron per gram of catalyst.

Analysis of the Fe concentration on the fresh and used catalysts (Supplementary Table 2) reveals that the Fe content increases on the used catalysts, even without additional Fe³⁺ sources. For the reaction without added Fe³⁺, the Fe concentration increases from 0.05 w.% to 0.12 and 0.14 wt.% on the used catalysts after 5 and 120 min of methane oxidation respectively in the stainless-steel autoclave reactor, suggesting deposition of Fe due to reactor corrosion under the hydrothermal conditions. For catalysts with added 0.03 wt.% and 0.1 wt.% Fe³⁺, the Fe concentration increases to approximately 0.20 wt.% and 0.30 wt.% after 5 min and 120 min, respectively. This trend is consistent for both Fe(NO₃)₃ and FeCl₃ as sources of the added iron. Based on our analysis, it appears that Fe³⁺ in the reaction solution acts primarily as a spectator, exerting minimal influence on the catalytic performance of the Au/ZSM-5 catalyst.

We suspect the oxide coating layers are from the corrosion of the stainless-steel reactor. We repeated the experiment using a fresh batch of Au/ZSM-5 catalyst and conducted the reaction for 5 min or 120 min, this time adding 300 ppm of Fe³⁺ to the reaction mixture. Similar oxide coating layer was again observed (Fig. 1j–l) and Supplementary Figs. 11–13). In this case, we even noted additional coating and the formation of what are likely Fe oxide nanoparticles on the zeolite support, away from the Au particles.

A detailed theoretical analysis of the effect of Fe impurities was not performed at this stage. However, we considered that the partial covering of the gold nanoparticles could provide a small percentage of potentially more active gold/iron-oxide interface sites. Using a model that was previously developed to study Au supported on Fe₂O₃ for CO oxidation², methane activation at the interface sites of Au₁₀O₆/Fe₂O₃ was calculated, showing a prohibitively high reaction barrier of 95 kJ mol⁻¹ — similarly, methane activation over the pure (0001) facet of α-Fe₂O₃ features hydrogen transfer barrier >100 kJ mol^{−1 3}. While these models do not capture the complexity of the current system, this qualitative result is consistent with the experiments that show how the presence of Fe impurities has no major impact on catalytic activity (see Supplementary Fig. 14).

Conclusions

During the oxidation of methane using a Au/ZSM-5 catalyst in a stainless steel autoclave under the reaction we used¹, a partial overlayer of iron oxide forms. In this addendum we have presented evidence to show that the presence of this partial overlayer does not alter the original conclusions and findings of our study.

Methods

Note on safe operation of experiments

For any reaction involving catalytic oxidation, care must be taken to work under conditions outside of the explosive mixture composition of the reagents. In the case of methane with oxygen, Cooper and Wiezevich have shown that as long as the experiments are conducted at ≤14% O₂, even at elevated temperature and pressure the experiment is outside of the explosive regime⁴. This is the case for all our experiments.

Catalyst preparation and testing

Au/ZSM-5 was prepared as described previously¹.

The methane oxidation reaction was conducted in a stainless-steel Parr autoclave reactor (total volume, 25 mL). Typically, the catalyst (0.1 g), deionized water (15 mL), then the reactor was sealed and then followed by purging with pure N₂ for 0.5 h to remove dissolved gases in water. After purging the reactor three times with methane to remove the residual nitrogen, the reactor was pressured with methane (20.7 bar) and oxygen (3.5 bar) at room temperature. Next, the mixture system was stirred at 1000 rpm for 10 min at room temperature to ensure that the reaction system reaches a completely stable state. Subsequently, the reactor was heated to the required reaction temperature (240 °C). After the reaction was maintained for 5 min or 120 min, the reactor was rapidly cooled in ice water bath to below 10 °C to avoid the loss of liquid products. The gas-phase products were collected with a gas sampling bag and the liquid-phase products were filtered and collected for liquid NMR analysis. The data was obtained by averaging the results of three repeat experiments.

For the control reactions with the addition of Fe³⁺ ions, Fe(NO₃)₃ (or FeCl₃) solution (15 mL) containing 0.03 or 0.10 wt. % of Fe³⁺ (weight of the catalyst) was used instead of pure water. The reaction procedures are consistent with those outlined above.

Catalyst characterization

The Fe content of fresh and used Au/ZSM-5 samples were analyzed by using an Agilent ICP-OES 730 inductively coupled plasma optical emission spectrometer.

The Au/ZSM-5 catalysts were characterized using a combination of transmission electron microscopy (TEM), scanning transmission electron microscopy (STEM), and associated spectroscopic techniques, including energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS). For TEM and STEM analysis, the samples were dry-mounted onto lacey carbon film TEM copper grids. TEM bright field (BF) imaging and EDS experiments were conducted on a JEOL 2800 microscope operating at 200 kV, equipped with a JEOL EDS detection system featuring 100 mm² Silicon Drift Detectors (SDDs) with a collection angle of approximately 0.95 steradians. STEM HAADF/BF imaging and EELS analysis were performed on an aberration-corrected JEOL ARM-200CF microscope, also operating at 200 kV, which is equipped with a cold field-emission gun and a Gatan GIF Quantum spectrometer.