CO₂ hydrogenation to CH₃OH promoted by strong Cu_xO–MgO interactions and non-thermal plasma under mild conditions

Abstract

Herein, we report the strong interaction between Cu_xO and MgO and its crucial role in promoting plasma-catalytic CO₂ hydrogenation to produce CH₃OH. Four MgO supports and corresponding Cu_xO/MgO catalysts were prepared and systematically characterized and tested in plasma-catalytic CO₂ hydrogenation. The results show that MgO supports with lower crystallinity and more defects favor stronger interactions with Cu_xO species, which result in better CO₂ conversion and CH₃OH selectivity in plasma-catalytic CO₂ hydrogenation, achieving 9.1% CO₂ conversion and 46.2% CH₃OH selectivity under ambient reaction conditions (30 °C and 0.1 MPa). In situ FTIR spectra and catalyst characterization results demonstrate that basic sites of MgO and O_ads. species activate CO₂ molecules into CO₃* species, which are further hydrogenated by H species to form the HCOO* and HCO* species on the MgO support. Cu_xO sites further promote the subsequent hydrogenation steps to move the reaction forward more easily to form the CH₃O* intermediate and the CH₃OH product, which could be the main reason why the strong Cu_xO–MgO interactions promote CH₃OH production in plasma-catalytic CO₂ hydrogenation.

---

Green foundation

1. This study achieves efficient low-temperature CO₂ conversion at 30 °C and ambient pressure via a plasma-driven Cu_xO/MgO catalytic system. The high-energy electrons from the plasma activate inert CO₂ molecules, while the strong Cu_xO–MgO interaction directs hydrogenation pathways, circumventing traditional high temperature and pressure requirements (200–300 °C and 2–5 MPa).

2. A cost-effective Cu-based catalyst exhibits 9.1% CO₂ conversion with 46.2% CH₃OH selectivity, transforming the greenhouse gas CO₂ into methanol—a key organic chemical feedstock.

3. Future work will design bifunctional and tandem catalysts to hydrogenate the byproduct CO into methanol in situ, alongside developing MgO-confined structures to enhance mass transfer efficiency, aiming for higher methanol selectivity.

1. Introduction

The increase in global population has resulted in an escalating demand for primary energy, more than 70% of which is provided by fossil fuels (coal, natural gas, and oil). This upsurge has led to an increasing atmospheric concentration of carbon dioxide (CO₂), which is considered to be the principal driver of adverse climate change and ocean acidification.¹

Extensive research studies have been directed toward the capture, separation and conversion of CO₂ to yield valuable chemicals (e.g., methanol, formic acid, ethanol and hydrocarbons),^2–8 among which CO₂ hydrogenation to methanol (CH₃OH) is an important strategy,^9–11 given that CH₃OH is a versatile molecule widely used in the chemical and energy industries. However, thermodynamic stability of the CO₂ molecule poses challenges in terms of activation under conventional conditions. Thus, the activation of CO₂ and the generation of methanol has traditionally been achieved at relatively high temperature and high pressure,^9,10,12–15 yet this approach engenders substantial economic burdens.^16,17 Additionally, the reactions to produce CO are favorable under high temperature conditions in terms of thermodynamics.¹⁸

Non-thermal plasma (NTP), an emerging technology, utilizes energetic electrons that are capable of activating CO₂ and H₂ to form active species, including radicals and excited (electronic and vibrational) species, which facilitate the CO₂ hydrogenation reaction at low temperature and atmospheric pressure.^18,19 Additionally, NTP offers high productivity and quick on/off capability, reducing operating costs and enhancing production efficiency. To date, NTP-assisted CO₂ hydrogenation has been investigated by many groups,^20–25 and excellent performances have been achieved at low temperature and atmospheric pressure using Cu-, Pt- and Fe-based catalysts.^18,26–29

MgO, a metal oxide with moderate basicity, is widely used as a catalyst, support, catalytic promoter, flame retardant and adsorbent.¹ With suitable basicity, MgO demonstrates moderate adsorption for CO₂, making it a popular choice for CO₂ capture.^30–33 Furthermore, MgO is capable of enhancing the adsorption capacity for CO₂ as a promoter, and thus it has been used to improve the performance of CO₂ hydrogenation in thermal catalysis.^13,34,35

In this study, Cu-based catalysts with MgO as the support were designed aiming to strengthen plasma-catalytic CO₂ hydrogenation to CH₃OH. Four types of MgO materials were synthesized and used as supports to prepare Cu_xO/MgO catalysts, which were tested in the plasma-catalytic CO₂ hydrogenation reaction. By combining the catalytic performances, the catalyst characterization results, and in situ FTIR spectra, it was found that the strong interactions between Cu_xO and MgO are favorable for plasma-catalytic CH₃OH generation. Experimental details are shown in the SI.

2. Experimental section

2.1 Catalyst preparation

Materials. Cu(NO₃)₂·3H₂O, MgCl₂·6H₂O, ammonia, Mg(CH₃COO)₂·4H₂O, C₆H₈O₇·H₂O, Mg(NO₃)₂·6H₂O, H₂C₂O₄, MgSO₄·7H₂O and KOH were purchased from Tianjin Kemiou Chemical Reagent Co., Ltd.

2.2 Synthesis of MgO supports

Hydrothermal method for MgO-1. 200 mL of 1 mol L⁻¹ MgCl₂ solution was added to a 500 mL beaker, and 27 mL of ammonia (25–28 wt%) was slowly added at a rate of 1 mL min⁻¹ to adjust the pH value to 10–11 under magnetic stirring. The mixture was then transferred to a Teflon-coated stainless-steel autoclave for a hydrothermal reaction at 150 °C for 6 h. After the reaction, the mixture was washed and filtered with deionized water until the filtrate pH reached neutrality. Subsequently, the gel was dried in a hot-air oven at 80 °C for 12 h, and then calcined in an air oven at 520 °C for 3 h.³⁶ The synthesis process of MgO-1 is shown in Fig. S1.

Sol–gel method for MgO-2. Initially, 0.2 mol of Mg(CH₃COO)₂·4H₂O and 0.2 mol of C₆H₈O₇·H₂O, serving as the precursor and gelling agent, respectively, were dissolved separately in 200 mL of ethanol and stirred for 2 h using a magnetic stirrer. The C₆H₈O₇·H₂O solution was added dropwise to the Mg(CH₃COO)₂·4H₂O solution under magnetic stirring, and the mixed solution was stirred for an additional 2 h to form a white gel. This white gel was washed and filtered with ethanol until the filtrate reached neutrality. Afterwards, the gel was dried in a hot-air oven at 100 °C for 12 h, followed by calcination in an air oven at 600 °C for 2 h.³⁷ The synthesis process of MgO-2 is shown in Fig. S2.

Sol–gel method for MgO-3. Initially, 0.2 mol of Mg(NO₃)₂·6H₂O and the 0.2 mol of H₂C₂O₄, acting as the precursor and gelling agent, respectively, were dissolved separately in 200 mL of ethanol and stirred for 2 h using a magnetic stirrer. Next, the H₂C₂O₄ solution was slowly added to the Mg(NO₃)₂·6H₂O solution under magnetic stirring, and the mixed solution was stirred for an additional 2 h until a white gel formed. This white gel was washed and filtered with ethanol until the filtrate was neutral. Subsequently, the gel was dried in a hot-air oven at 100 °C for 12 h, followed by calcination in an air oven at 600 °C for 2 h.³⁸ The synthesis process of MgO-3 is shown in Fig. S3.

Sol–gel method for MgO-4. Initially, 0.2 mol of MgSO₄·7H₂O and 0.2 mol of KOH, serving as the precursor and gelling agent, respectively, were dissolved separately in 200 mL of deionized water and stirred for 2 h using a magnetic stirrer. Then, the KOH solution was slowly added to the MgSO₄·7H₂O solution under magnetic stirring, and the mixed solution was stirred for an additional 2 h to form a white gel. This white gel was washed and filtered with deionized water until the filtrate reached neutrality. Afterward, the gel was dried in a hot-air oven at 80 °C for 12 h, and subsequently calcined in an air oven at 520 °C for 3 h.³⁹ The synthesis process of MgO-4 is shown in Fig. S4.

2.3 Preparation of Cu_xO/MgO catalysts

The Cu_xO/MgO catalysts were prepared using an impregnation method with Cu(NO₃)₂·3H₂O as the copper source, as shown in Fig. S5. The amount of Cu_xO is expressed as the Cu loading, which was 10 wt% for all samples. The impregnated precursor was dried at 110 °C overnight and then calcined in air at 540 °C for 5 h. The prepared catalysts were sieved into target particles of 20–40 mesh for filling in the reaction area.

2.4 Catalytic tests

Fig. S6 illustrates the schematic configuration of the plasma-catalytic CO₂/H₂ conversion system. The reaction occurs within a coaxial dielectric barrier discharge (DBD) reactor designed to generate CO₂/H₂ plasma, featuring a dual-electrode configuration where a stainless-steel rod of 2 mm diameter serves as the high-voltage central electrode. A water-cooling system circulates deionized water through the inter-cylinder space, simultaneously functioning as the grounding electrode. The discharge gap with discharge length of 60 mm was filled with catalyst granules (0.42–0.85 mm in diameter). Gas feed streams were precisely regulated through calibrated mass flow controllers (MFC, CS200, SevenStar, China) with volumetric verification being conducted using a soap-film flowmeter maintaining flow rates of 18 mL min⁻¹ (CO₂) and 54 mL min⁻¹ (H₂) under ambient pressure.

The plasma-catalytic system employed a DBD reactor powered by an AC high-voltage generator with a frequency range of 7–12 kHz (CTP-2000K, Suman, China). Real-time monitoring of electrical parameters,including applied voltage, discharge current, and external capacitor voltage, was achieved through a two-channel digital oscilloscope (Tektronix, MDO 3024). Plasma power quantification employed the Q-U Lissajous methodology, with operational parameters maintained at an excitation frequency of 9.2 kHz and a net discharge power of 24 W.

A gas chromatograph (Agilent 8860) equipped with a thermal conductivity detector (TCD) connected to a Porapak Q + 5A molecular sieve column (2 m × 2.1 mm) and a flame ionization detector (FID) connected to a HP-PLOT/Q capillary column (30 m × 0.53 mm × 40 μm) was used to analyze the gaseous products. CH₄ and CH₃OH were analyzed using the FID, while CO₂, CO and H₂ were analyzed using the TCD. The reaction temperature in the discharge area was close to the circulating water temperature (30 °C).

To evaluate the reaction performance, the CO₂ conversion was calculated using eqn (1):

In the tail gas, CH₃OH, CO and CH₄ were detected using the gas chromatograph. The selectivity of CH₃OH, CO and CH₄ was calculated by using eqn (2), (3) and (4), respectively:

Based on the results of thermogravimetric analysis-mass spectrometry (TGA-MS), as shown in Fig. S10 and Table S11, the coke deposition on the catalyst was found to be negligible.

The carbon balance was calculated using eqn (5):

2.5 Catalyst characterization

The crystalline phase of the catalysts was analyzed using a SmartLab 9KW (Rigaku, Japan) powder X-ray diffractometer (PXRD) with Cu Kα radiation (λ = 0.15406 nm) under operational parameters of 240 kV and 50 mA. Diffraction patterns were acquired in the 2θ range of 10°–80° with a scanning rate of 10° min⁻¹ and step size of 0.02°. The specific surface area, N₂ adsorption–desorption isotherms, and Barrett–Joyner–Halenda (BJH) pore size distributions were determined using a low-temperature physisorption analyzer (ASAP 3020, Micromeritics, USA), following sample degassing at 350 °C for 5 h under a 10⁻³ Torr vacuum. The surface morphology and particle size of the catalysts were investigated via field emission scanning electron microscopy (FESEM, SU5000, Hitachi, Japan) at 5 kV accelerating voltage. Atomic-scale microstructure was resolved using a high-resolution transmission electron microscopy system (HRTEM, JEM-F200, JEOL, Japan) operating at 200 kV voltage, demonstrating 0.23 nm point resolution and 0.10 nm lattice resolution. Quasi in situ X-ray photoelectron spectroscopy (quasi in situ XPS) measurements were performed using a spectrometer (PHI 5000 Versaprobe II, ULVAC-PHI, Japan) with a micro-focused monochromatic Al Kα X-ray source and scanning X-ray beam-induced secondary electron image. Binding energy calibration utilized adventitious carbon (C 1s = 284.8 eV) as the internal reference. Immediately after the plasma-catalytic reaction, the entire reactor was transferred into a nitrogen-filled glovebox. Inside the glovebox, the spent catalyst was carefully retrieved and sealed in a sample vial. Prior to XPS scanning, the sealed vial was moved into a dedicated glovebox that was directly connected to the X-ray photoelectron spectrometer. The entire sample transfer process was conducted under an inert atmosphere without exposure to air, which ensured the stability of the Cu valence state and the accuracy of the measurements. Hydrogen temperature-programmed reduction (H₂-TPR) and CO₂ temperature-programmed desorption (CO₂-TPD) experiments were conducted with Quantachrome ChemBET Pulsar TPR apparatus (Anton Paar, Austria). For H₂-TPR, 0.15 g samples were pretreated in He (550 °C, 1 h) followed by 10% H₂/Ar temperature ramping (50–500 °C, 10 °C min⁻¹) with TCD monitoring. The CO₂-TPD involved saturation adsorption (120 mL min⁻¹, 99.99% CO₂, 80 min) prior to programmed desorption (50–760 °C, He flow, 10 °C min⁻¹), and the signal of the CO₂ desorption was recorded using a TCD detector. Raman spectra of the samples were collected using a micro-confocal Raman spectrometer (InVia Qontor, Renishaw, UK) with a 532 nm excitation laser (5 mW) and an integration time of 10 s. The dispersion, particle size and specific surface area of metallic copper (Cu) were further quantified by N₂O chemisorption, conducted on the same apparatus used for the H₂-TPR analyses. The detailed experimental procedures and calculation methods are provided in the SI.

2.6 Plasma diagnostics

In situ optical emission spectroscopy of the CO₂/H₂ plasma was performed using a Princeton Instruments SP-2758 intensified ICCD spectrometer equipped with a 300 g mm⁻¹ holographic grating covering the 200–1100 nm spectral range. The spectral resolution was optimized with fixed instrumental parameters: 20 μm entrance slit width and 2 s integration time per acquisition. Dynamic plasma parameters were captured through a digital oscilloscope (DPO 3012, Tektronix, USA) with specialized probes: a P6015A high-voltage passive probe for voltage waveform acquisition and a Pearson Electronics 6585 Rogowski coil current monitor. Power quantification employed the Lissajous graphical method with phase-resolved charge–voltage (Q–U) analysis.

2.7 In situ FTIR setup

In situ FTIR measurements were conducted using an FTIR spectrometer (Nicolet iS10, Thermo, USA) equipped with a rapidly recoverable detector containing deuterated triethylene glycol sulfate (DTGS) to observe surface species of catalyst and study the mechanism of CO₂ hydrogenation to CH₃OH. The catalyst, pressed into a tablet (8 mm), was placed into the custom-designed plasma reaction chamber (Scheme S1). Then, a flow of pure CO₂ (18 mL min⁻¹) was introduced for 30 minutes to allow for the saturation of adsorption sites on the catalyst surface, and the spectrum collected at this stage represents the initially adsorbed species. H₂ was then introduced to form a CO₂/H₂ mixture (25 vol% CO₂, 75 vol% H₂, 18 mL min⁻¹ CO₂ and 54 mL min⁻¹ H₂), and spectra were recorded for 10 minutes in the absence of plasma. After the signal stabilized, the plasma was ignited while maintaining the CO₂/H₂ flow, and spectra were then recorded sequentially over time to monitor the dynamic changes of surface species triggered by the plasma. In situ infrared spectra were recorded from 900 to 4000 cm⁻¹ with a resolution of 4 cm⁻¹, and the plasma discharge frequency and input power were set to 9.2 kHz and 24 W, respectively.

3. Results and discussion

3.1 Synthesized MgO supports

Four MgO supports were prepared using hydrothermal and sol–gel methods with different preparation procedures (Fig. S1–S4). The synthesized MgO samples were denoted MgO-1, MgO-2, MgO-3 and MgO-4, respectively. Accordingly, employing an impregnation method (Fig. S5), the prepared Cu_xO/MgO catalysts (ca. 10 wt% loading of Cu) were named Cu_xO/MgO-1, Cu_xO/MgO-2, Cu_xO/MgO-3 and Cu_xO/MgO-4, respectively.

Fig. 1A shows the XRD patterns of the synthesized MgO supports. Diffraction peaks of the (200) and (220) planes of MgO were observed, and the intensity gradually decreased following the sequence MgO-1 > MgO-2 > MgO-3 > MgO-4. That is, the MgO samples show decreasing crystallinity from MgO-1 to MgO-4. The N₂-physisorption isotherms are shown in Fig. 1B, and the corresponding texture information (Table S1) indicates an increasing trend for the specific surface area of the synthesized MgO samples, following the sequence MgO-1 < MgO-2 < MgO-3 < MgO-4. In other words, the MgO-1 sample possesses the highest crystallinity but the smallest specific surface area, while the MgO-4 sample has the largest specific surface area. Fig. 1C shows the SEM images of the MgO samples. It can be seen that MgO-1 and MgO-2 feature hexagonal and spherical particles, respectively, both with uniform morphology and high crystallinity. However, the images of MgO-3 and MgO-4 display irregularly shaped particles with relatively low crystallinity.

Fig. 1 Characterization of the synthesized MgO samples. (A) XRD patterns, (B) N₂ adsorption isotherms, and (C) SEM images.

3.2 Catalytic performance of Cu_xO/MgO catalysts

Fig. 2 shows the performance of the plasma-catalytic CO₂ hydrogenation reaction employing MgO and Cu_xO/MgO as catalysts. For the MgO supports, the CO₂ conversion is very low (Fig. 2A), and the main product is CO (Fig. 2C). However, the Cu_xO/MgO catalysts shows much improved performance with both enhanced CO₂ conversion (Fig. 2A) and CH₃OH selectivity (Fig. 2B), indicating that the Cu_xO active sites are crucial for the production of CH₃OH. Notably, for the Cu_xO/MgO catalysts employing different MgO supports, the CO₂ conversion and CH₃OH selectivity gradually increase following the sequence Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4. However, CO selectivity shows a contrary trend (Fig. 2C). That is, the MgO supports play an important role in plasma-catalytic CO₂ hydrogenation to CH₃OH, and the Cu_xO/MgO-4 catalyst shows superior performance, achieving 9.1% CO₂ conversion and 46.2% CH₃OH selectivity. From Table S3, the (space–time yield) STY of CH₃OH achieved in our work ranks at the middle level among the reported plasma-catalysis reactions. The long-term test (Fig. S7) exhibits excellent stability over a 24 hour period for all four Cu_xO/MgO catalysts, with no significant decline observed in either CO₂ conversion or CH₃OH selectivity.

Fig. 2 Plasma-catalytic CO₂ hydrogenation performances over different catalysts. (A) CO₂ conversion; (B) CH₃OH selectivity; (C) CO selectivity; and (D) comparison of reaction modes (1.5 g of catalyst, 9.2 kHz discharge frequency, 24 W input power, CO₂ : H₂ = 1 : 3, WHSV = 2400 mL g⁻¹ h⁻¹, 18 mL min⁻¹ CO₂, 54 mL min⁻¹ H₂, 30 °C circulating water).

Fig. 2D shows the performances of “plasma only”, “plasma + MgO-4” and “plasma + Cu_xO/MgO-4”, which confirms an obvious synergistic effect between plasma and the Cu_xO/MgO-4 catalysts in driving CO₂ hydrogenation to selectively produce CH₃OH.

As shown in Table S5, the selectivity of CH₄ is 2.6% in the plasma only experiment, and with the introduction of the Cu_xO/MgO catalyst, the CH₄ selectivity was significantly suppressed to 0.5%. The simultaneous increase in CO₂ conversion and decrease in CH₄ selectivity after introducing the Cu_xO/MgO catalysts strongly suggests that the catalysts effectively suppress the deep hydrogenation pathway that leads to CH₄.

3.3 Catalyst characterization

Fig. 3A shows the XRD patterns of the Cu_xO/MgO catalysts. Clearly, the samples show nearly identical XRD patterns dominated by the MgO species. Diffraction peaks at 35.55° and 38.73°, associated with the (−111) and (111) planes of CuO, respectively (PDF#05-0661), are observed. However, the intensities are much lower than those of the MgO diffraction peaks, which means Cu species may be highly dispersed on the MgO supports. Fig. 3B shows the H₂-TPR profiles of the Cu_xO/MgO catalysts. Since MgO cannot be reduced at temperatures below 600 °C,^13,35,40,41 the observed H₂ consumption peaks are attributed to the reduction of Cu species. After peak fitting, the H₂-TPR profiles show three peaks denoted α, β and γ, respectively. Generally, the α peak is caused by the reduction of CuO species with strong interactions with the MgO support;⁴² the β peak corresponds to the reduction of highly dispersed CuO_x clusters with smaller sizes; and the γ peak originates from the reduction of Cu₂O to metallic Cu⁰ species.^41,43 The γ peak is observed in the Cu_xO/MgO-3 and Cu_xO/MgO-4 catalysts but is absent for the Cu_xO/MgO-1 and Cu_xO/MgO-2 catalysts, which means that the Cu_xO/MgO-3 and Cu_xO/MgO-4 samples may have an increased content of Cu₂O. Additionally, the reduction temperature of Cu_xO shows an increasing trend: Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4, which suggests that the interaction between Cu_xO and MgO is gradually enhanced with the decrease of MgO crystallinity. That is, MgO supports with low crystallinity favor strong oxide–support interaction due to the existence of more defects, and the Cu_xO/MgO-4 catalyst exhibits the strongest oxide–support interaction between Cu_xO and MgO. Furthermore, this trend correlates perfectly with the significantly improved catalytic performance for CH₃OH synthesis.

Fig. 3 Characterization of the Cu_xO/MgO samples. (A) XRD patterns of the Cu_xO/MgO samples; (B) H₂-TPR profiles of the Cu_xO/MgO samples; (C) the particle size and dispersion of the Cu_xO on different MgO supports as obtained via N₂O chemisorption.

The dispersion of Cu (proportion of surface Cu atoms in relation to the total number of Cu atoms) in the Cu_xO/MgO-X (X = 1, 2, 3, 4) catalyst was measured by the N₂O chemisorption experiment, as shown in Fig. S8, while the corresponding experimental data for specific surface area (S_Cu) and particle size (D_Cu) of Cu are shown in Table S6. The dispersion of Cu for the samples Cu_xO/MgO-1, Cu_xO/MgO-2, Cu_xO/MgO-3 and Cu_xO/MgO-4 was calculated to be 15.7, 18.2, 20.8 and 25.7%, respectively. That is, with the decrease of crystallinity of the MgO supports, the average particle size of Cu in the Cu_xO/MgO catalysts presents a decreasing trend, and thus the dispersion of Cu shows an increasing trend, as presented in Fig. 3C. For the sample Cu_xO/MgO-4, the average particle size of Cu was estimated to be around 3.85 nm and the dispersion of the Cu was ca. 25.7%. Overall, the H₂-TPR and N₂O chemisorption experiments indicate that the oxide–support interactions between Cu_xO and MgO gradually intensifies following the sequence Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4, corresponding with the increasing Cu_xO dispersion.

Fig. 4A and B show the Cu 2p_3/2 spectra of the fresh and spent Cu_xO/MgO samples, respectively (the fitted values and deconvolution details are shown in Tables S7 and S8). For the fresh Cu_xO/MgO-1 sample (Fig. 4A), the satellite companion peak at 942.6 eV indicates the predominance of divalent Cu species (CuO and Cu²⁺ in the exchange state). The Cu 2p_3/2 peak at 933.9 eV corresponds to the Cu⁺ species and the Cu 2p_3/2 peak at 935.4 eV is attributed to the Cu²⁺ species.^44,45 We can also note that, compared with the Cu_xO/MgO-1 sample, the binding energies of satellite, Cu²⁺ and Cu⁺ are reduced for the samples Cu_xO/MgO-3 and Cu_xO/MgO-4, indicating electron shift from MgO to Cu species, which may be caused by interactions between Cu_xO and MgO. This result is consistent with the conclusion of enhanced interactions from H₂-TPR characterization. For the spent Cu_xO/MgO catalysts, a notable increase in the relative content of Cu⁺ is observed, suggesting that CuO is partially reduced to Cu₂O during the H₂/CO₂ plasma reaction. Fig. 4C shows the relationship of Cu⁺ content and reaction performances. The relative content of Cu⁺ for the spent Cu_xO/MgO catalysts shows a more pronounced increasing trend (Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4), which corresponds to CO₂ conversion and CH₃OH selectivity that are gradually improved, demonstrating the active site role of Cu₂O.

Fig. 4 Quasi in situ XPS characterization of the Cu_xO/MgO catalysts. (A) Cu 2p_3/2 spectra of the fresh Cu_xO/MgO samples, (B) Cu 2p_3/2 spectra of the spent Cu_xO/MgO catalysts, (C) the relationship between the relative content of Cu⁺ and the performance of plasma-catalytic CO₂ hydrogenation on Cu_xO/MgO catalysts; (D) O 1s spectra of the fresh Cu_xO/MgO samples, (E) O 1s spectra of the spent Cu_xO/MgO catalysts; (F) the relationship between the relative content of O_ads. species and the performance of plasma-catalytic CO₂ hydrogenation on the Cu_xO/MgO catalysts. The standard charge was calibrated with C 1s binding energy of 284.8 eV.

Fig. 4D and E show the O 1s spectra of fresh and spent Cu_xO/MgO samples, respectively (the fitted values and deconvolution details are shown in Tables S9 and S10). The O 1s spectra of the fresh and spent Cu_xO/MgO samples can be deconvoluted into two peaks, i.e. 529.5 eV and 531.5 eV, corresponding to surface lattice oxygen (O_latt.) and adsorbed oxygen (O_ads.) species, respectively.³⁵ Interestingly, the relative content of O_ads. gradually increases following the sequence Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4, and the content of O_latt. shows an inverse trend, irrespective of fresh or spent Cu_xO/MgO samples. The highest content of O_ads. but the lowest content of O_latt. species occurs in the Cu_xO/MgO-4 sample and originates mainly from the strong Cu_xO–MgO interactions, which are caused principally by low crystallinity and the presence of more defects on the MgO-4 support. Fig. 4F shows the positive relationship of O_ads. species with CO₂ conversion and CH₃OH selectivity. That is, Cu₂O and O_ads. species that originated from strong Cu_xO–MgO interactions play an important role in plasma-catalytic CO₂ hydrogenation for CH₃OH production.

Fig. 5A shows the Raman spectra of the Cu_xO/MgO samples. The characteristic band that is peaked at 292 cm⁻¹ corresponds to the A_g modes of CuO.⁴⁶ The Raman shift bands at 330–760 cm⁻¹ and 830–1380 cm⁻¹ are associated with the MgO characteristic peaks.^47–49 The former band is attributed to the longitudinal optical (LO) mode of MgO particles, while the latter is attributed to the MgO symmetric stretching mode of A1 symmetry.⁴⁹ Notably, the two Raman shift bands of MgO show the characteristic broadening and softening effects from Cu_xO/MgO-1 to Cu_xO/MgO-4. That is, the Cu_xO/MgO-4 sample shows the strongest MgO characteristic peak with the lowest Raman shift. The variations of the position and intensity of MgO characteristic peaks are mainly caused by the gradually enhanced interactions between Cu_xO and MgO supports, as indicated by the quasi in situ XPS and H₂-TPR results.

Fig. 5 Characterization of the Cu_xO/MgO catalysts. (A) Raman spectra of the Cu_xO/MgO catalysts and (B) CO₂-TPD profiles of the Cu_xO/MgO catalysts. TEM images of spent catalysts showing the lattice fringe of Cu₂O (200): (C) Cu_xO/MgO-1; (D) Cu_xO/MgO-2; (E) Cu_xO/MgO-3; (F) Cu_xO/MgO-4.

The CO₂ adsorption capacity of the catalysts has been proved to be a critical factor in enhancing the performance of CO₂ hydrogenation to CH₃OH.^13,35Fig. 5B shows the CO₂-TPD profiles of the Cu_xO/MgO samples. The α, β and γ peaks correspond to weak, moderate and strong basic sites in the catalysts, respectively.⁴⁰ It is evident that the quantity of moderate and strong basic sites in the catalysts continually increase following the sequence Cu_xO/MgO-1 < Cu_xO/MgO-2 < Cu_xO/MgO-3 < Cu_xO/MgO-4. That is, the CO₂ adsorption capacity of the Cu_xO/MgO catalysts is gradually enhanced with the decrease of MgO crystallinity. From the results of the quasi in situ XPS study, we inferred that the enhanced CO₂ adsorption capacity is attributed to the increased content of O_ads. in the catalysts with the increase of Cu_xO–MgO interactions. Fig. 4F indicates that there is a positive correlation between the O_ads. content in the catalysts and the performance of plasma-catalytic CO₂ hydrogenation to CH₃OH. In conjunction with the CO₂-TPD, it is inferred that the O_ads. enhances the CO₂ adsorption capacity of the catalysts, thereby increasing the CO₂ conversion and the CH₃OH selectivity.

Fig. S9 displays the TEM images of the fresh Cu_xO/MgO catalysts. The results reveal that the different preparation methods for MgO influences the exposed planes of the CuO crystals. The crystalline lattice distances of 0.232 nm and 0.252 nm correspond to the (111) and (−111) planes of CuO, respectively. Moreover, the size of the CuO particles is observed to be about 5–8 nm, and these results of particle size are in accordance with the results of N₂O chemisorption with respect to the particle size (Fig. 3C).

To further elucidate the nature of the active Cu species under reaction conditions, the spent catalysts were examined using TEM, with representative images shown in Fig. 5C–F. Notably, crystalline lattices corresponding to Cu₂O were clearly observed in all spent samples. This direct microscopy evidence for the formation of Cu₂O during the plasma-catalytic process strongly corroborates our quasi in situ XPS findings (Fig. 4B and C), which indicated a significant increase in the relative content of Cu₂O species after the reaction. The convergence of these results provides compelling support for the hypothesis that Cu₂O species are the primary active sites for the hydrogenation of CO₂ to CH₃OH.

3.4 Plasma diagnosis

The plasma discharge behaviors are diagnosed using a two-channel digital oscilloscope. Fig. 6A shows that there is no significant change in the waveform of the voltage for different catalysts. In Fig. 6B, the current waveform for different catalysts is gradually enhanced except for the plasma only, indicating that the catalyst with a larger specific surface area facilitates the surface discharge. As shown in Fig. 6C, the Lissajous figure changes from a parallelogram shape (plasma only) to an oval shape (plasma + Cu_xO/MgO), indicating a significant shift in the discharge behavior from the crucial filamentary discharges to surface discharges.⁵⁰ The measured discharge power differs from the input power displayed by the plasma source (24 W), because we measured the actual discharge power across the reactor rather than the input power of the plasma source. The small differences in actual discharge power observed among the different catalysts at the same input power (24 W)arise from changes in the equivalent capacitance of the reactor, which varies with the dielectric properties of the materials packed into the system.

Fig. 6 Diagnosis of the CO₂/H₂ plasmas. (A) Waveforms of discharge voltages; (B) waveforms of discharge currents; (C) Lissajous plots; (D) mean electron energy of CO₂/H₂ plasmas as a function of E/N; (E) the electron energy distribution function of the CO₂/H₂ plasmas; (F) in situ OES spectra (1.5 g catalyst, 9.2 kHz discharge frequency, 24 W input power, CO₂ : H₂ = 1 : 3, WHSV = 2400 mL g⁻¹ h⁻¹, 18 mL min⁻¹ CO₂, 54 mL min⁻¹ H₂, 30 °C circulating water).

The electron energy distribution function (EEDF) of the CO₂/H₂ plasmas under our experimental conditions was calculated using BOLSIG+.⁵¹Fig. 6D shows the mean electron energy as a function of the reduced electric field (E/N). The mean electron energy is 1.7 eV under the condition of plasma only. Clearly, packing of Cu_xO/MgO catalysts in the CO₂/H₂ plasma system dramatically enhances the mean electron energy, reaching ca. 4.5–4.7 eV, and a closer inspection reveals a slightly increasing trend from Cu_xO/MgO-1 to Cu_xO/MgO-4. This suggests that the catalyst with a stronger oxide–support interaction may slightly enhance the electric field in the discharge gap, leading to a higher mean electron energy. Furthermore, it can be found that the integration of plasma and Cu_xO/MgO catalysts strongly changes the electron energy distribution in CO₂/H₂ plasma and the four different Cu_xO/MgO catalysts show a similar energy distribution as shown in Fig. 6E. We also identified a clear trend of increasing mean electron energy from Cu_xO/MgO-1 to Cu_xO/MgO-4. The enhancement of mean electron energy produces more energetic electrons and chemically reactive species for plasma reactions and consequently contributes to the conversion of CO₂.

In situ optical emission spectroscopy (OES) was used to identify key species in the CO₂/H₂ plasma, as presented in Fig. 6F. The CO₂/H₂ plasma exhibits strong signal intensity, featuring several spectral lines and bands without a catalyst or support, such as the H_α line (656.3 nm, 3d²D → 2p²P⁰),⁵² two O atomic spectral lines (777.5 nm, 3s⁵S⁰ → 3p⁵P; 844.7 nm, 3s³S⁰ → 3p³P),^53,54 a band for H₂ (580–650 nm, d³Π_u → a³Σ_g⁺) and a CO band (450 nm → 580 nm, B¹Σ → A¹Π).⁵⁵ These findings suggest that H atoms and CO molecules are abundant in the CO₂/H₂ plasma. After packing the CO₂/H₂ plasmas with Cu_xO/MgO catalysts, the spectral intensities were dramatically weakened, which may be caused by both the shielding effect of catalyst granules on light propagation and chemisorption of excited species on catalyst surface. More importantly, the enlarged OES spectra shown in Fig. S11 clearly demonstrate that the hydrogen-related species (H_α at 656.3 nm and H₂ band at 580–650 nm) and CO band (450–580 nm) underwent significant intensity decreases. This trend confirms that the Cu_xO/MgO-4 catalyst, which possesses the strongest Cu_xO–MgO interaction, is more effective at adsorbing and consuming these plasma-generated active species for the hydrogenation reaction. That is, the gradually decreased OES intensities (from Cu_xO/MgO-1 to Cu_xO/MgO-4) were caused by variation of chemisorption of excited species on the catalyst surface. The Cu_xO/MgO-4 catalyst with strong Cu_xO–MgO interactions and abundant defects is more conducive to adsorb excited species from CO₂/H₂ plasma. This should be one of the most important reasons why the Cu_xO/MgO-4 catalyst exhibits a superior performance compared with the other Cu_xO/MgO catalysts in plasma-catalytic CO₂ hydrogenation to CH₃OH.

3.5 Discussion of the reaction mechanism

In situ FTIR studies were conducted to investigate the active intermediates and reaction pathways in plasma-catalytic CO₂ hydrogenation to CH₃OH on Cu_xO/MgO catalysts. As shown in Fig. 7A and B, the behavior of the bare MgO support under plasma was first investigated as a control. With increasing CO₂ exposure times, the amount of carbonate species (CO₃* at 1456, 1540 and 1649 cm⁻¹) on the catalyst surface gradually increased.^56,57 Upon plasma ignition, a significant increase in the intensity of gas-phase CO was observed at 2115 and 2170 cm⁻¹,⁵⁸ concurrently with the emergence of an infrared absorption band at 1338 cm⁻¹, which was assigned to the formyl (HCO*) intermediate.⁵⁹ Additionally, a weak peak attributed to formate (HCOO*) can be observed at 2841 cm⁻¹.⁵⁶ This observation suggests that on the MgO support, the reaction proceeds via two parallel pathways: (i) a plasma-induced gas-phase dissociation of CO₂ to form CO, which subsequently reacts with H species from gas through a reverse water–gas shift (RWGS) pathway to generate HCO*,^18,21 and (ii) the reaction between CO₃* on the MgO surface and H species from gas, which follows the formate pathway to produce CH₃OH.^21,22 Although HCO* and HCOO* are plausible precursors, their subsequent hydrogenation to CH₃OH was not detected, presumably due to their extremely low yield on the bare support.

Fig. 7 In situ FTIR spectra during plasma-catalytic CO₂ hydrogenation on (A) MgO-4 and (B) Cu_xO/MgO-4 (9.2 kHz discharge frequency, 24 W input power, CO₂ : H₂ = 1 : 3, WHSV = 2400 mL g⁻¹ h⁻¹, 18 mL min⁻¹ CO₂, 54 mL min⁻¹ H₂, 30 °C circulating water).

In stark contrast, the plasma-catalytic process over the Cu_xO/MgO catalyst exhibits a distinctly different reaction network. Under identical discharge conditions, the signal for gas-phase CO was significantly suppressed. However, the intensity of the HCO* intermediate was more pronounced than on MgO, and a new peak emerged at 2077 cm⁻¹, corresponding to CO adsorbed on Cu₂O active sites (M–CO).⁶⁰ This suggests that on Cu_xO/MgO, CO generated from plasma dissociation preferentially adsorbs onto the catalyst surface and reacts with gas-phase H species via an Eley–Rideal (E–R) mechanism to form HCO*, which is then hydrogenated to CH₃OH. More importantly, several new, crucial intermediates were observed exclusively on the Cu_xO/MgO surface. As the plasma discharge proceeded, the intensity of surface CO₃* showed a slight increase, which could be attributed to the plasma-enhanced adsorption of CO₂.⁶¹ Critically, absorption bands corresponding to HCOO* at 1592 cm⁻¹ (ref. 58) and methoxy (CH₃O*) at 2820 and 1464 cm⁻¹ (ref. 62 and 63) appeared and intensified over time. Their evolutionary trend mirrored that of the surface CO₃*, providing direct evidence for the formate-mediated pathway: CO₂ first adsorbs on the surface to form CO₃*, which are then hydrogenated to HCOO*, and subsequently to CH₃O*, the direct precursor to CH₃OH.

By comparing the relative intensities of the key intermediates on Cu_xO/MgO, the signals for CH₃O* and HCOO* were substantially stronger than that for HCO*. This strongly indicates that the formate pathway is the dominant reaction route for CH₃OH synthesis on the Cu_xO/MgO catalyst. These collective observations highlight the multifunctional role of the Cu₂O active sites. On the one hand, they adsorb plasma-generated CO, facilitating a minor reaction channel via the Langmuir–Hinshelwood (L–H) mechanism. On the other hand, and more critically, Cu₂O active sites are adept at adsorbing H, enabling the hydrogenation of surface carbonates. This process can occur both via an E–R mechanism (adsorbed carbonate reacting with gas-phase H) and an L–H mechanism (adsorbed carbonate reacting with adsorbed H on nearby Cu₂O active sites), creating parallel and efficient routes for CH₃OH formation (no CH₄ production was observed during the entire process from the results of in situ FTIR).

Based on the catalyst characterization results and the in situ FTIR spectra, the possible mechanism of plasma-catalytic CO₂ hydrogenation to CH₃OH promoted by Cu_xO–MgO strong interactions is shown in Scheme 1. Base sites of MgO and O_ads. species promote activation of the CO₂ molecule through chemisorption, leading to formation of CO₃* species. Benefiting from the strong hydrogenation capacity of H radicals (generated by plasma), hydrogenation of CO₃* species occurs easily to form the HCOO* and HCO* species on the MgO support through the formate and the RWGS pathways, respectively. Furthermore, the subsequent hydrogenation steps proceed more easily in the presence of the Cu₂O active sites, leading to formation of the CH₃O* intermediate. That is, the synergy of Cu₂O active sites and MgO support is crucial for CH₃OH generation, which may be explained from two aspects. First, electron transfer from MgO to Cu_xO (confirmed by quasi in situ XPS) favors the formation of active Cu₂O sites. Second, strong interaction between MgO and Cu_xO means a shorter Cu_xO–MgO distance and more interfacial sites, which facilitate stepwise hydrogenation on both the MgO support and Cu₂O active sites. This may be the main reason why the strong Cu_xO–MgO interactions promote plasma-catalytic CO₂ hydrogenation to produce CH₃OH.

Scheme 1 The possible mechanism of plasma-catalytic CO₂ hydrogenation to CH₃OH (A) MgO support and (B) promoted by the strong Cu_xO–MgO interaction.

4. Conclusion

Four MgO supports have been synthesized, and corresponding Cu_xO/MgO catalysts have been prepared to combine with plasma for the CO₂ hydrogenation reaction. Catalytic tests and catalyst characterization results show that the MgO support with lower crystallinity and more defects favors stronger interaction with Cu_xO species, which results in better CO₂ conversion and CH₃OH selectivity in plasma-catalytic CO₂ hydrogenation. The Cu_xO/MgO-4 catalyst exhibits the best performance, reaching 9.1% CO₂ conversion and 46.2% CH₃OH selectivity. Furthermore, catalyst characterization and in-situ FTIR spectra demonstrate that strong base sites of MgO and O_ads. species activate CO₂ molecule into CO₃* species, which is further hydrogenated by H species to form the HCOO* and HCO* species on the MgO support, and the presence of the Cu₂O active sites promote the subsequent hydrogenation steps that proceed more easily to form the CH₃O* intermediate and CH₃OH product. This could be the main reason why the strong Cu_xO–MgO interaction promotes CH₃OH production in plasma-catalytic CO₂ hydrogenation.

Author contributions

Qian Chen: conceptualization, validation, formal analysis, resources, data curation, writing – original draft, and writing – review & editing. Shengyan Meng: conceptualization, validation, formal analysis, and data curation. Xiaohan Zhai: validation, formal analysis, and data curation. Li Wang: resources and data curation. Zhaolun Cui: resources and data curation. Dongxing Li: validation, formal analysis, and data curation. Chuang Li: resources and data curation. Chong Peng: resources and data curation. Yanhui Yi: conceptualization, validation, formal analysis, resources, data curation, writing – original draft, writing – review & editing, supervision, and funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Data availability

The authors declare that the data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information: details of catalyst preparation, the methods of catalytic tests and characterization, plasma diagnostics, in situ FTIR setup, morphological information, standard curve equation, comparison of this work with state-of-the-art plasma-catalysis and thermal-catalysis reported in the literature, the results of stability tests, CH₄ selectivity for plasma only and different catalysts, N₂O chemisorption, deconvolution details of Cu 2p_3/2 and O 1s spectra, TEM for different fresh catalysts, carbon balance analysis and the enlarged in situ OES spectra. See DOI: https://doi.org/10.1039/d5gc02218e.

Acknowledgements

We acknowledge the financial support from the National Natural Science Foundation of China [22272015, 22472018, 21503032].

References

1. O. A. Ojelade and S. F. Zaman, Catal. Surv. Asia, 2019, 24, 11–37.
2. Y. Sun, Z. Lin, S. Peng, V. Sage and Z. Sun, J. Nanosci. Nanotechnol., 2019, 19, 3097–3109.
3. L. Fu, Z. Ren, W. Si, Q. Ma, W. Huang, K. Liao, Z. Huang, Y. Wang, J. Li and P. Xu, J. CO2 Util., 2022, 66, 102260.
4. S. G. Gizer, O. Polat, M. K. Ram and N. Sahiner, Int. J. Energy Res., 2022, 46, 16241–16263.
5. W. Wang, C. Zeng and N. Tsubakia, Green Carbon, 2023, 1, 133–145.
6. H. Ju, Y. Zhang, J. Zhang, Z. Yu, Y. Zhang, X. Zhang, X. Guo, J. Liu, Q. Mao, Q. Liu, Y. Zhao, T. Qi, X. Jiang, Z. Guo and S. Chen, Smart Mol., 2025, 3, e20240027.
7. D. Li, M. Ye, C. Ma, N. Li, Z. Gu and Z. Qiao, Smart Mol., 2024, 2, e20240009.
8. T. Xia, Y. Wu, T. Ji, W. Hu, K. Yu, X. He, B. Yin and Y. Liu, Smart Mol., 2025, e20240066.
9. S. Zhang, Z. Wu, X. Liu, K. Hua, Z. Shao, B. Wei, C. Huang, H. Wang and Y. Sun, Top. Catal., 2021, 64, 371–394.
10. J. Niu, H. Liu, Y. Jin, B. Fan, W. Qi and J. Ran, Int. J. Hydrogen Energy, 2022, 47, 9183–9200.
11. K. Jiang, Y. Xu, F. Cao, B. Li, X. Xu, W. Wang, Y. Tang, L. Wu and Li Tan, Green Carbon, 2025, 3, 139–147.
12. S. Dang, B. Qin, Y. Yang, H. Wang, J. Cai, Y. Han, S. Li, P. Gao and Y. Sun, Sci. Adv., 2020, 6, eaaz2060.
13. T. Guo, Q. Guo, S. Li, Y. Hu, S. Yun and Y. Qian, J. Catal., 2022, 407, 312–321.
14. C. Wang, Y. Fang, G. Liang, X. Lv, H. Duan, Y. Li, D. Chen and M. Long, J. CO2 Util., 2021, 49, 101542.
15. Q. Wu, C. Shen, N. Rui, K. Sun and C. Liu, J. CO2 Util., 2021, 53, 101720.
16. X. Jiang, X. Nie, X. Guo, C. Song and J. Chen, Chem. Rev., 2020, 120, 7984–8034.
17. I. Ganesh, Renewable Sustainable Energy Rev., 2014, 31, 221–257.
18. L. Wang, Y. Yi, H. Guo and X. Tu, ACS Catal., 2017, 8, 90–100.
19. L. Wang, X. Du, Y. Yi, H. Wang, M. Gul, Y. Zhu and X. Tu, Chem. Commun., 2020, 56, 14801–14804.
20. X. Gao, J. Liang, L. Wu, L. Wu and S. Kawi, Catalysts, 2022, 12, 66.
21. R. Michiels, Y. Engelmann and A. Bogaerts, J. Phys. Chem., 2020, 124, 25859–25872.
22. M. D. Farahani, Y. Zeng and Y. Zheng, Energy Sci. Eng., 2022, 10, 1572–1583.
23. G. Chen, R. Snyders and N. Britun, J. CO2 Util., 2021, 49, 101557.
24. S. Xu, H. Chen, C. Hardacre and X. Fan, J. Phys. D:Appl. Phys., 2021, 54, 233007.
25. M. Liu, Y. Yi, L. Wang, H. Guo and A. Bogaerts, Catalysts, 2019, 9, 275.
26. Y. Men, Y. Liu, Q. Wang, Z. Luo, S. Shao, Y. Li and Y. Pan, Chem. Eng. Sci., 2019, 200, 167–175.
27. S. Meng, L. Wu, M. Liu, Z. Cui, Q. Chen, S. Li, J. Yan, L. Wang, X. Wang, J. Qian, H. Guo, J. Niu, A. Bogaerts and Y. Yi, AICHE J., 2023, e18154.
28. Z. Cui, S. Meng, Y. Yi, A. Jafarzadeh, S. Li, E. C. Neyts, Y. Hao, L. Li, X. Zhang, X. Wang and A. Bogaerts, ACS Catal., 2022, 12, 1326–1337.
29. Q. Chen, S. Meng, R. Liu, X. Zhai, X. Wang, L. Wang, H. Guo and Y. Yi, Appl. Catal., B, 2024, 342, 123422.
30. A. H. Ruhaimi and M. A. A. Aziz, J. Solid State Chem., 2021, 300, 122242.
31. A. M. Alkadhem, M. A. A. Elgzoly and S. A. Onaizi, J. Environ. Chem. Eng., 2020, 8, 103968.
32. C. Tan, Y. Guo, J. Sun, W. Li, J. Zhang, C. Zhao and P. Lu, Fuel, 2020, 278, 118379.
33. A. H. Ruhaimi, M. A. A. Aziz and A. A. Jalil, J. CO2 Util., 2021, 43, 101357.
34. S. K. Sharma, T. S. Khan, R. K. Singha, B. Paul, M. K. Poddar, T. Sasaki, A. Bordoloi, C. Samanta, S. Gupta and R. Bal, Appl. Catal., A, 2021, 623, 118239.
35. G. Tian, Y. Wu, S. Wu, S. Huang and J. Gao, Chem. Phys. Lett., 2022, 786, 139173.
36. W. Zhang, N. Zhou, Y. Zhang, Z. Huang, H. Hu, J. Liang and Y. Qin, J. Appl. Polym. Sci., 2021, 138, e50677.
37. C. W. Wong, Y. S. Chan, J. Jeevanandam, K. Pal, M. Bechelany, M. A. Elkodous and G. S. El-Sayyad, J. Cluster Sci., 2019, 31, 367–389.
38. P. Kasikamphaiboon and U. Khunjan, Int. J. Chem. Eng., 2018, 8, 1706405.
39. Q. Ao, S. Li, H. Li, H. Guo and X. Liu, Appl. Chem. Ind., 2021, 50, 1855–1858.
40. C. Liu, X. Guo, Q. Guo, D. Mao, J. Yu and G. Lu, J. Mol. Catal. A:Chem., 2016, 425, 86–93.
41. S. Goodarznia and K. J. Smith, J. Mol. Catal. A:Chem., 2012, 353, 58–66.
42. S. Goodarznia and K. J. Smith, J. Mol. Catal. A:Chem., 2010, 320, 1–13.
43. F.-W. Chang, W.-Y. Kuo and K.-C. Lee, Appl. Catal., A, 2003, 246, 253–264.
44. C. Mebrahtu, R. Sun, C. H. Gierlich and R. Palkovits, Appl. Catal., B, 2021, 287, 119964.
45. F. Bin, X. Wei, B. Li and K. S. Hui, Appl. Catal., B, 2015, 162, 282–288.
46. J. Xu, W. Ji, Z. Shen, W. Li, S. Tang, X. Ye, D. Jia and X. Xin, J. Raman Spectrosc., 1999, 30, 413–415.
47. S. Demirci, B. K. Yildirim, M. M. Tünçay, N. Kaya and A. N. Güllüoğlu, J. Sol-Gel Sci. Technol., 2021, 99, 576–588.
48. R. A. El-Salamony, A. S. Morshedy and A. M. A. E. Naggar, Environ. Technol., 2020, 43, 1860–1869.
49. P. S. Prabhu, P. Kathirvel, D. Maruthamani and S. D. G. Ram, Optics, 2023, 283, 170869.
50. X. Zhang, Y. Liu, M. Zhang, T. Yu, B. Chen, Y. Xu, M. Crocker, X. Zhu, Y. Zhu, R. Wang, D. Xiao, M. Bi, D. Ma and C. Shi, Chem, 2020, 6, 3312–3328.
51. D. Mei, X. Zhu, Y. He, J. D. Yan and X. Tu, Plasma Sources Sci. Technol., 2015, 24, 015011.
52. G. Zambrano, H. Riascos, P. Prieto, E. Restrepo, A. Devia and C. Rincón, Surf. Coat. Technol., 2003, 172, 144–149.
53. A. V. Kholodkov, K. M. Golant and I. V. Nikolin, Microelectron. Eng., 2003, 69, 365–372.
54. T. Czerwiec, J. Gavillet, T. Belmonte, H. Michel and A. Ricard, Surf. Coat. Technol., 1998, 98, 1411–1415.
55. A. Granier, M. Vervloet, K. Aumaille and C. Vallée, Plasma Sources Sci. Technol., 2003, 12, 89–96.
56. W. Wang, Z. Qu, L. Song and Q. Fu, J. Energy Chem., 2020, 40, 22–30.
57. C. Schild, A. Wokaun and A. Baiker, J. Mol. Catal., 1990, 63, 243–254.
58. C. Vogt, E. Groeneveld, G. Kamsma, M. Nachtegaal, L. Lu, C. J. Kiely, P. H. Berben, F. Meirer and B. M. Weckhuysen, Nat. Catal., 2018, 1, 127–134.
59. Y.-F. Zhao, R. Rousseau, J. Li and D. Mei, J. Phys. Chem., 2012, 116, 15952–15961.
60. Y. Huang, J. Am. Chem. Soc., 1973, 95, 6636–6640.
61. S. Xu, S. Chansai, S. Xu, C. E. Stere, Y. Jiao, S. Yang, C. Hardacre and X. Fan, ACS Catal., 2020, 10, 12828–12840.
62. J. Wang, G. Li, Z. Li, C. Tang, Z. Feng, H. An, H. Liu, T. Liu and C. Li, Sci. Adv., 2017, 3, e1701290.
63. I. A. Fisher and A. T. Bell, J. Catal., 1997, 172, 222–237.

---