Efficient Cu/SiO₂ catalysts for methyl glycolate synthesis via dimethyl oxalate hydrogenation under atmospheric pressure

Abstract

An innovative gel-hydrothermal (Gel-HT) method was developed to synthesize Cu/SiO₂ catalysts with excellent catalytic activity for the selective hydrogenation of dimethyl oxalate (DMO) to methyl glycolate (MG). A series of characterization techniques, including N₂ physisorption–desorption, N₂O titration, Fourier-transform infrared spectroscopy, H₂ temperature-programmed reduction, X-ray diffraction, transmission electron microscopy, H₂ temperature-programmed desorption and X-ray photoelectron spectroscopy were employed to elucidate the source of catalyst activity. The catalyst 35% Gel-HT achieved 92.5% DMO conversion and 84.5% MG selectivity at 0.1 MPa and 170 °C, and the catalyst maintained stable performance for over 5000 h. The results of the characterization and activity tests indicated that the activation capacity of hydrogen is a critical factor controlling the atmospheric pressure activity of the catalysts.

---

1. Introduction

The methyl glycolate (MG) molecule contains both ester and hydroxyl groups, which can undergo various reactions such as carbonylation, oxidation, and hydrolysis, thus MG serves as an important organic raw material for the production of high value-added commodities, such as pharmaceutical intermediates, perfumes, and degradable plastics.¹ Numerous methods have been explored to produce MG, however, harsh conditions and low yields hinder industrial applicability.²

Selective hydrogenation of dimethyl oxalate (DMO) is a promising catalytic reaction because the selective availability of methyl glycolate, ethylene glycol, and ethanol by controlling the degree of hydrogenation of the feedstock confers great production flexibility to the DMO hydrogenation process. A great deal of effort has been expended by researchers on the study of catalytic systems for the hydrogenation of DMO.³ Results show that the weak ability of Ag to activate hydrogen favored only partial hydrogenation of DMO rather than complete hydrogenation to EG, and the silver-based catalysts readily achieve satisfactory MG yields, but the expensive price and high-temperature sintering susceptibility of metallic silver limited the large-scale application of the silver-based catalysts to a certain extent.^3e,4

Copper-based catalysts have a significantly higher hydrogenation capacity than silver, and they selectively break C=O bonds rather than C–O, thus exhibiting excellent activity in DMO hydrogenation reactions.^4a,c,5 Copper-based catalysts catalyze DMO hydrogenation easily with high EG yields, but MG is the major product only when the DMO conversion is at a low level, and any measures to promote DMO conversion, such as increasing the temperature and decreasing the LHSV, will synchronously promote the hydrolysis of MG.⁶ It remains a great challenge to obtain satisfactory MG selectivity at high DMO conversion using copper-based catalysts. To cope with this dilemma, various strategies have been proposed: changing the type of catalyst carrier to investigate the effects of metal-carrier forces and carrier surface properties on catalyst activity; doping the catalyst with a second metal component to modulate the dispersion and valence distribution of copper species, which in turn affects the activation process of ester groups and hydrogen, and regulates the activity of the catalyst.⁷ Although researchers have done extensive work on the improvement of catalysts, the activity of Cu-based catalysts for selective hydrogenation of DMO to MG has not fundamentally improved, and there is still a big gap between the key indexes, such as target product yield and catalyst life, and the requirements of industrial production.

The catalytic conversion of DMO is typically conducted at pressures between 2 and 3 MPa.⁸ The high hydrogen pressure not only implies potential safety risks and high plant operating costs but is also likely to profoundly affect the catalyst activity, especially considering that the actual hydrogen-ester ratio employed is much higher than the theoretical value for the DMO hydrogenation reaction.⁹ Few researchers have focused on the study of DMO low-pressure hydrogenation reaction characteristics and the preparation of DMO atmospheric pressure hydrogenation catalysts. Although Zheng et al. achieved success in the hydrogenation reaction of DMO at atmospheric pressure to make EG using C₆₀-modified Cu/SiO₂, the complex preparation process of this catalyst makes it face greater difficulties in large-scale preparation.¹⁰

In this paper, Cu/SiO₂ catalysts with a theoretical copper content of 15–45% were prepared for the atmospheric pressure hydrogenation of dimethyl oxalate by the gel-hydrothermal method, and comparative catalysts with suitable copper contents were also prepared by the ammonia evaporation and precipitation methods. The physicochemical properties of the catalysts were systematically characterized by applying N₂-physical adsorption–desorption, FT-IR, XRD, H₂-TPD and XPS, and the main factor affecting the low-pressure activity of the catalysts was investigated in the context of the differences in the activity of these catalysts. In addition, a feasible strategy to improve the selectivity of copper-based catalysts for MG and extend their lifetime was proposed based on the effects of reaction pressure and hydrogen supply on the reaction product distribution and catalyst lifetime of DMO hydrogenation to MG.

2. Experimental

2.1 Catalyst preparation

Cu/SiO₂ catalysts were prepared via gel-hydrothermal method using copper acetate and ethyl orthosilicate as the copper and silicon sources, the theoretical copper content of the catalyst was set at 15%, 25%, 35% and 45%. A certain amount of copper acetate was dissolved in 250 ml of deionized water and 34.7 g of TEOS was added with strong stirring, a blue viscous gel formed after continuous stirring in a water bath 50 °C for 4 h. The gel was then transferred to a tetrafluoro-lined hydrothermal kettle and allowed to stand at 110 °C for 96 h. The obtained samples were separated by centrifugation and washed with deionized water and ethanol thoroughly. The washed sample was dried at 110 °C for 12 h, and calcined at 450 °C for 4 h. The calcined sample was labeled as xGel-HT, where x represents the theoretical copper loading of the catalyst.

For comparison, Cu/SiO₂ with 25% copper loading were prepared by ammonia evaporation and precipitation methods. For the ammonia evaporation method, a certain amount of Cu(NO₃)₂·3H₂O was dissolved in deionized water, after that 25 wt% ammonia was added and stirred for 30 min with the initial PH of 11–12. The silica sol was added and stirred for another 4 h, then the temperature was raised to 80 °C to remove ammonia. When the pH of the suspension dropped to 6–7, the ammonia evaporation process was terminated. The resulting mixture was washed thoroughly with deionized water, dried at 110 °C for 12 h, and calcined in air at 450 °C for 4 h, the sample was labeled as AEP.

The preparation process of the precipitation method refers to the reported method: Na₂SiO₃·9H₂O was used as the precipitant and the silicon source.¹¹ Cu(NO₃)₂·3H₂O and the appropriate amount of 65 wt% nitric acid were dissolved in 100 ml of ionized water to form solution A; 28.4 g of Na₂SiO₃·9H₂O was dissolved in 150 ml of deionized water to form solution B. Solution A and B were added to 400 ml of deionized water at the same time, stirred strongly for 1 h and then left to stand for 12 h, then washed thoroughly with deionized water, dried overnight at 110 °C and calcinated at 450 °C for 4 h, the as-prepared catalyst was named Des.

2.2 Catalyst characterization

The structural properties of the prepared samples were determined by N₂-physical adsorption–desorption at −196 °C using a Tristar II3020 analyzer. The samples were degassed under vacuum at 150 °C for 2 h before testing and the specific surface area of the samples was calculated by the Brunauer–Emmett–Teller method. The pore volume and pore size distribution of the samples were calculated from the desorption branch of the isotherms.

The actual copper content of all catalysts was measured by inductively coupled plasma optical emission spectroscopy (ICP-OES). The samples were dissolved with HF solution, then neutralized with HBO₃ and finally diluted with water.

X-ray diffraction (XRD) measurement was conducted on Shimadzu XRD-6000 diffractometer with Cu Kα radiation (λ = 1.5418 Å) in the 2θ range from 5° to 90° with a scanning rate of 4° min⁻¹.

Transmission electronic microscopy (TEM) images were obtained on JEM 2100F equipped with a field emission gun. The catalyst powder was dispersed in ethanol by ultrasonication and the suspension was dropped on a carbon foil supported by a copper mesh.

Fourier transform infrared spectra (FT-IR) were carried out on a Nicolet 6700 spectrometer. Sample powders were mixed well with KBr, pressed into tablets, and loaded into the sample holder. The data were recorded at a resolution of 4 cm⁻¹ in the region of 400–4000 cm⁻¹.

The reduction properties of the calcinated catalyst were investigated using a hydrogen-temperature-programmed reduction (H₂-TPR) technique. Approximately 50 mg of a 40–60 mesh sample was subjected to an Ar flow at 120 °C for a period of 2 hours, after which it was cooled to ambient temperature. Subsequently, the sample was heated from 50 °C to 900 °C at a rate of 10 °C min⁻¹ in a 5% H₂–95% N₂ flow. The quantity of hydrogen consumed during the reduction process was determined by TCD detection. The H₂-TPD test was conducted on the aforementioned equipment with an Ar purge for a period of 30 minutes before the commencement of the test. A quantity of 0.2 g of catalyst was subjected to reduction with pure hydrogen at a temperature of 300 °C for 2 h. Ar was switched after the temperature dropped below 50 °C. The desorption process commenced at 50 °C and concluded at 800 °C. The rate of temperature increase was 10 °C min⁻¹.

The N₂O titration experiment was employed to determine the copper dispersion of the catalyst. The experimental method and analytical approach were conducted following the established protocols documented in the literature.¹² A quantity of 50 mg of the sample was reduced with 5% H₂/N₂ at 300 °C for a period of 2 hours. The amount of hydrogen consumed was recorded as X, and then the reduced Cu⁰ was oxidized to Cu⁺ with 15% N₂O/He at 50 °C. Following the completion of the oxidation process, the catalyst was thoroughly purged with Ar, and the reduction was carried out again using 5% H₂/N₂. The amount of hydrogen consumed in the second reduction was recorded as Y, and the dispersion of metallic copper was calculated by the following equation:

D = 2Y/X × 100%

The surface area of metallic copper was calculated by following formula:

where, X is N is the number of atoms per m² of active Cu surface (1.46 × 10¹⁹), N_A is Avogadro constant, M_Cu = relative atomic mass (63.456 g mol⁻¹), and W is mass percentage of Cu in wt%.¹³

Turnover frequencies (TOF) were calculated according to the following equation:

where W is the DMO concentration in the DMO/methanol solution (mol L⁻¹), V is the flow rate of the DMO/methanol solution (L h⁻¹), C_DMO is the DMO conversion, N_Cu is the total amount of Cu (mol) in the catalyst and D is the Cu dispersion obtained from N₂O titration. For the TOF measurements, the DMO conversion was controlled below 30% by increasing the LHSV, while the other reaction conditions were kept the same (T = 170 °C, P = 0.1 MPa and H₂/DMO = 50).

X-ray photoelectron spectroscopy (XPS) and X-ray induced Auger spectra (XAES) analyses were conducted on a Thermo Scientific ESCALAB 250Xi apparatus equipped with an Al Kα X-ray source (hv = 1486.6 eV). The catalysts were pre-reduced at 300 °C for 2 h under a H₂ flow. The binding energies were calibrated with a C_1s binding energy of 284.6 eV.

2.3 Catalysts evaluation

The selective hydrogenation of dimethyl oxalate to methyl glycolate was conducted in a continuous-flow stainless steel tube fixed-bed reactor. 0.5 g of 40–60 mesh catalyst was placed in the middle of the tube, the top and bottom were filled with inert quartz sand. Before evaluation, the catalyst was reduced with pure hydrogen at 300 °C for 2 h, then cooled down to the desired reaction temperature, the pressure in the reactor was raised to 0.1 MPa with hydrogen, and a methanol solution of dimethyl oxalate with a mass content of 8% and hydrogen were pumped in simultaneously according to the set liquid-time air velocity and hydrogen-ester ratio. The liquid-phase feedstock and hydrogen were mixed and preheated in the upper part of the reaction tube and then passed through the catalyst bed for the reaction, after which the reaction mixture was condensed and collected in the storage tank. The post-reaction products were analyzed using gas chromatography equipped with a flame ion detector.

3. Results

3.1 Physicochemical properties of the catalysts

Actual copper loadings determined by ICP-OES are summarized in Table 1. The copper content of the gel hydrothermal series catalysts is lower than the pre-determined value due to the copper leaching during gel formation. Furthermore, the higher the theoretical amount, the higher the proportion of lost copper to the theoretical value.

Table 1 Physicochemical properties of the as-prepared catalysts

Catalyst	Cu loading^a (%)	S BET (m^2 g^−1)	V p (cm^3 g^−1)	D p (nm)	Cu dispersion^c (%)	S Cu^0 (m^2 g^−1)	Relative content of copper silicate^d	Cu^0 content^e (%)
a Determined by ICP-OES. b Determined by N2 isotherm adsorption. c Determined by N2O titration method. d Calculated as I670/I800, where I is peak intensity. e Cu^0/(Cu^+ + Cu^0) obtained by deconvolution of Cu LMM XAES spectra.
Des	23.62	516.7	1.11	8.6	36.3	43.0	0.0348	43.3
AEP	24.71	480.2	0.85	7.1	14.5	22.9	0.1074	62.5
15% Gel-HT	12.20	390.1	1.02	9.4	25.2	19.7	0.0526	41.8
25% Gel-HT	19.56	491.8	0.85	6.0	20.9	26.3	0.1619	49.7
35% Gel-HT	24.78	541.7	0.71	5.1	18.2	20.2	0.4311	56.4
45% Gel-HT	30.86	592.5	0.55	3.6	10.4	20.7	0.8932	65.7
SiO2	—	732.3	2.06	11.2	—	—	—

---

The structural properties of all the catalysts were listed in Table 1, and the corresponding N₂ adsorption–desorption isotherms and pore size distribution curves were presented in Fig. 1 and 2, respectively. The specific surface area of the pure SiO₂ carrier was as high as 732.3 m² g⁻¹, which dropped to 390.1 m² g⁻¹ with 12.2% Cu loading and increased to 592.5 m² g⁻¹ as the Cu loading increased from 12.2% to 30.86%, and the relative copper silicate content of the catalysts increased steadily. As demonstrated by J. Rouquerol et al., the abundant cross-linked pores of copper silicate species in Cu/SiO₂ significantly increased their specific surface area.¹⁴ However, the trends in pore volume and pore diameter for Cu-containing catalysts with varying copper loadings were opposite to those of the specific surface area, the increased copper loading decreased the pore volume and pore diameter of Cu/SiO₂ catalysts. The increase in copper loading caused the growth of catalytic copper species particles which evidenced by the deteriorated Cu dispersion, and more carrier pores were blocked by copper species particles; the increased copper loading improved the relative copper silicate content of the catalyst and the large amount of in situ-generated copper silicate generated numerous fine cross-linking pore channels in the original pore channels of the carrier. Consequently, the pore volume and pore size of the copper-containing catalysts decreased monotonically as the copper content increased.

Fig. 1 N₂ adsorption–desorption isotherm of calcinated catalysts.

Fig. 2 Pore distribution curves of calcinated catalysts.

All Gel-HT samples exhibited type IV isothermal features accompanied by type H1 hysteresis loops, indicating that the material possesses a regular mesoporous structure. The pore size distribution curve of the SiO₂ carrier displays a predominant peak at approximately 11 nm. In regard to copper-containing catalysts prepared by the gel-hydrothermal method, the intensity of the peaks at the larger sizes remained stable while a new peak appeared around 3 nm. It is noteworthy that the intensity of this secondary peak was found to be significantly enhanced with the increase of copper loading. This can be attributed to structural reorganization results from the increase of the relative content of copper silicate, which leads to the formation of new cross-linked pore structures in the carriers.

3.2 FT-IR spectroscopy

FT-IR spectroscopy was employed to analyze the structural evolution of copper species. As illustrated in Fig. 3. The absorption peak at 1096 cm⁻¹ and its broad shoulder at 1210 cm⁻¹ corresponds to the asymmetric stretching vibration of the Si–O–Si bond in the SiO₂ carrier, while the Si–O–Si bond symmetric stretching vibration is observed at 800 cm⁻¹.¹⁵ Copper-containing catalysts showed an absorption peak at 1036 cm⁻¹ for Cu–O–Si bond stretching vibration, and the peak intensity of this absorption peak increased with copper content in the gel-hydrothermal series catalysts, which indicated that the increase of copper loading significantly promoted the formation of Cu–O–Si bonds, and more copper ions were embedded in the SiO₂ skeleton at high copper content, leading to partial replacement of Si–O bonds by Cu–O–Si bonds.

Fig. 3 The FT-IR spectra of calcinated catalysts.

The presence of copper silicate species in the copper-containing catalysts was confirmed by the absorption peak of δ_OH at 670 cm⁻¹. The relative abundance of copper phyllosilicates was semi-quantified using the intensity of the absorption peaks at 800 and 670 cm⁻¹ and the obtained results were listed in Table 1. With the actual copper loading increased from 12.20% to 30.86%, the relative copper silicate content of the Gel-HT series catalysts increased from 0.0526 to 0.8392, and this trend aligns with the enhancement of the intensity of the absorption peaks of the Cu–O–Si bonds with the increase of copper content. These observations underscored the critical role of copper loading in the context of copper phyllosilicate formation and silica–copper bonding.

The FT-IR spectra of AEP, Des, and Gel-HT series catalysts revealed that the form of copper species present was significantly modulated by the preparation method. The AEP catalyst exhibited a strong Si–O–Si absorption peak at 1096 cm⁻¹, which confirms that the carrier structure is intact. It is noteworthy that the weak intensity of the absorption peak at 1036 cm⁻¹, combined with its low copper silicate content, reflected the limited formation of Cu–O–Si bonds in this preparation method.

The absorption peak at 1036 cm⁻¹ of the Des catalyst was not significant and the high Cu dispersion measured by nitrous oxide titration suggested that the slow nucleation process of the precipitation method, although favorable for Cu dispersion, is difficult to drive the formation of Cu–O–Si bonds. When Cu/SiO₂ catalysts were prepared using the gel-hydrothermal method, the slow hydrolysis of TEOS in a weakly acidic solution produced a large number of silica hydroxyl groups. These groups achieved the pre-dispersion of copper species by anchoring free Cu²⁺ in the solution through electrostatic action. The rearrangement of the silica–oxygen network in the subsequent hydrothermal crystallization process led to the embedding of the copper ions in the SiO₂ skeleton structure to realize the bonding of Cu–Si. The 35% Gel-HT sample exhibited significantly higher copper silicate content than Des and AEP, emphasizing the superiority of the gel-hydrothermal approach in fostering copper–silica interactions.¹⁶

3.3 H₂-TPR

The H₂-TPR measurements were conducted to investigate the reducibility of the catalysts prepared by different methods and the results were shown in Fig. 4. For the Gel-HT series, a systematic trend emerged as the Cu loading increased from 15% to 45%. The intensity of the reduction peak progressively intensified, accompanied by noticeable shifts in peak shape. At 15% Cu content, a single broad peak centered around 220 °C was observed, characteristic of the reduction of highly dispersed CuO species in the SiO₂. As the Cu loading was elevated to 25%, the peak intensity increased slightly, and its width narrowed, suggesting a partial aggregation of Cu species. A pronounced transformation was observed at 35% and 45% Cu loadings, where the original reduction peak split into two overlapping peaks in the low and high temperature regions. The lower-temperature peak is attributed to the reduction of highly dispersed CuO particle, while the higher-temperature peak corresponds to the reduction of Cu²⁺ to Cu⁺ in the Cu–O–Si structure.¹⁷ This bifurcation suggests that excessive Cu loading induces phase segregation or structural rearrangement within the catalyst matrix. Notably, the total intensity of the peaks scaled linearly with Cu loading, confirming the quantitative relationship between metal loading and reducible copper species. Such behavior aligns with previous studies on Cu/SiO₂ systems, where optimal copper loading and phase balance were critical for catalytic performance.¹²

Fig. 4 The H₂-TPR profiles of calcinated catalysts.

AEP catalyst exhibited a dual-peak profile with pronounced peaks at 250 °C and an inconspicuous shoulder peak at a slightly lower temperature. This indicates the presence of two distinct copper species in the calcinated AEP catalysts, akin to the Gel-HT series catalysts. However, the AEP catalysts contain minor copper silicate, with bulk CuO emerging as the predominant copper species. Des catalyst exhibited a single broad peak centered at 230 °C, similar to that of 15% Gel-HT sample, suggesting minimal Cu aggregation despite its higher loading. Comparatively, the Gel-HT, AEP and Des methods yield contrasting Cu distributions under identical loading conditions. The Des catalyst's resistance to aggregation is advantageous for maintaining high surface activity, whereas the AEP catalyst's tendency to form bulk CuO may compromise reducibility; the type of copper species in the catalyst prepared by Gel-HT method can be controlled by varying the copper loading. These differences underscore the importance of methodology in controlling the type of copper species present as well as their reduction behaviors.^12b,18

3.4 XRD

Fig. 5 illustrated the XRD patterns of all calcinated and reduced catalysts. The broad diffraction peak at 2θ of 22° was attributed to the amorphous silica carriers. For Gel-HT catalysts, the XRD patterns showed characteristic peaks of CuO (JCPDS 48-1548) at 2θ = 35.5° and 38.7°, confirming the presence of crystalline CuO species. As the copper content increases from 15% to 45%, the CuO diffraction peaks intensified and narrowed, indicating enhanced crystallinity and larger crystallite sizes. This trend suggests that higher copper loading promotes agglomeration of CuO particles, likely due to reduced dispersion efficiency on the SiO₂ carrier at elevated metal loading. The actual copper content of catalysts AEP and Des was found to be almost equivalent to that of 35% Gel-HT but exhibited distinct structural characteristics. The AEP catalyst shown a weak CuO signal, which is very similar to the XRD pattern of 45% Gel-HT; while the XRD pattern of the Des catalyst was very similar to that of 15% Gel-HT and no observable copper-related diffraction peaks were detected. Therefore, in terms of maintaining the dispersion state of copper species, Des is the best, followed by the gel-hydrothermal method, and the ammonia evaporation method is the worst.

Fig. 5 XRD patterns of catalysts: (a) after calcination and (b) after reduction.

After reduction, the characteristic diffraction peaks of Cu⁰ (PDF# 04-0836) and Cu₂O (PDF# 05-0667) were clearly observed in the XRD spectra of Gel-HT catalysts with different copper loadings. The intensity of the Cu⁰ (111) diffraction peak at 2θ = 43.3° increased with the increase of Cu loading, indicating a propensity for Cu⁰ agglomeration at elevated loading. The expansion in particle dimensions at higher loadings likely stems from enhanced precursor aggregation during synthesis, favoring sintering upon reduction. Concurrently, the (111) peak of Cu₂O at 2θ ≈ 36.4° diminished in intensity at elevated copper loadings (35–45%), suggesting a higher reduction efficiency to metallic copper under these conditions. In contrast, lower copper loadings (15–25%) retain detectable Cu₂O phases, reflecting incomplete reduction due to limited precursor availability. These observations highlight the interplay between copper loading and particle dispersion, where optimal loadings balanced crystallite growth and spatial distribution.

When comparing AEP, Des, and Gel-HT catalysts at equivalent copper content, marked differences in structural features emerged. The AEP sample exhibited a sharp and intense Cu⁰ (111) peak, whereas the Des catalyst showed a slightly broader peak. In contrast, the 35% Gel-HT displayed the broadest Cu⁰ peak signifying smaller particle dimensions and superior dispersion. These findings underscore the critical role of synthesis strategy in tailoring Cu/SiO₂ catalyst properties, gel-hydrothermal method offering a robust route to optimize metal dispersion and metal particle size, while AEP and Des may require additional modifications to address their inherent limitations in particle size control and reduction efficiency.

3.5 TEM images

The morphology of the reduced catalysts and the size of the copper particles were observed by TEM characterization. Fig. 6 showed the TEM images of the reduced catalysts and their respective particle size distribution statistics. For the catalysts prepared by the gel-hydrothermal method, the increase in copper loading only led to an increase in the average particle size of copper species in the reduced catalysts from 2.03 nm to 3.94 nm without significantly affecting the structure of the SiO₂ carriers. The catalyst Des that prepared by the precipitation method comprises a substantial quantity of irregular particles that stacked in a disordered manner and embedded with copper species with an average particle size of approximately 3 nm. In the case of the catalyst AEP, the copper species are found to be homogeneously dispersed in the carrier, existing in the form of larger-sized but mutually independent particles. The combination of N₂O chemisorption and XRD results demonstrated that the gel-hydrothermal method is highly effective in promoting the homogeneous dispersion of copper species in Cu/SiO₂ catalysts as well as inhibiting aggregation at high copper loadings.

Fig. 6 TEM images of reduced catalysts: (a) Des, (b) AEP, (c) 15% Gel-HT, (d) 25% Gel-HT, (e) 35% Gel-HT, (f) 45% Gel-HT.

3.6 H₂-TPD

The activation of hydrogen represents a crucial phase in the hydrogenation process.¹⁹ To gain insight into the adsorption and activation of hydrogen by the catalyst, the H₂-TPD technique can be employed. The H₂-TPD plots of several catalysts were presented in Fig. 7, where the differences in copper loading and preparation methods caused great discrepancies in the H₂ activation capacity of the catalysts. Only one hydrogen desorption signal appeared in the range of 600–750 °C for pure SiO₂ carriers, and multiple new hydrogen desorption peaks appeared within the 100–450 °C range for copper-containing catalysts, confirming that the copper species is the main active sites for hydrogen adsorption and activation. Based on the relationship between desorption temperature and adsorption energy, the low-temperature peak near 120 °C can be attributed to weakly chemisorbed hydrogen bound to the copper particles; the medium-temperature peak corresponds to chemically adsorbed hydrogen on the defective sites of the copper lattice; high-temperature peaks with desorption temperatures higher than 400 °C are dominated by carrier contributions.^8a,20 The onset temperature of the high-temperature hydrogen desorption peaks associated with the carriers decreases with increasing copper loading, reflecting the modulation of the carrier surface properties by copper species.²¹

Fig. 7 H₂-TPD profiles of catalysts.

For the Gel-HT series of catalysts, the intensity of the low-temperature hydrogen desorption peak at 120 °C exhibited a significant volcano-type trend as the Cu content increased. This trend can be deconvoluted into two competing structural factors governed by the copper loading. Initially, increasing the Cu content up to 35% enhances the population of active sites for hydrogen adsorption. More importantly, as evidenced by FT-IR and the relative copper silicate content (Table 1), higher loadings within this range promote the formation of copper phyllosilicate. Upon reduction, these precursors yield a structure comprising small, well-dispersed Cu⁰ nanoparticles intimately anchored to the silica support via persistent Cu–O–Si interactions. This specific architecture, characterized by a high density of under-coordinated Cu⁰ sites (e.g., corners, edges) and strengthened metal–support interfaces, is highly conducive to the weak chemisorption of hydrogen, manifesting as the prominent ∼120 °C desorption peak. However, beyond the optimal loading, excessive copper leads to particle agglomeration, as confirmed by the decreased Cu dispersion in N₂O titration and larger particle size in XRD/TEM, which diminishes the abundance of these critical low-coordination sites and consequently reduces the intensity of the low-temperature H₂ desorption peak.

The presence of this weakly adsorbed hydrogen species is pivotal for atmospheric pressure hydrogenation. Its lower binding energy renders it highly mobile and readily active, effectively increasing the surface concentration of active hydrogen atoms and lowering the overall activation barrier for the reaction under low hydrogen under low hydrogen partial pressure.²² For the AEP catalyst, despite its high surface Cu⁰ content, comprises larger Cu⁰ particles with a lower fraction of defect sites and lacks the robust Cu–O–Si interface, thus failing to generate significant weak adsorption sites. The Des catalyst, although having high dispersion, possesses a low surface Cu⁰ content and insufficient metal–support interaction to stabilize the active structure necessary for optimal H₂ activation. Therefore, the unique low-temperature hydrogen activation capability of the Gel-HT catalyst is unequivocally attributed to its optimally structure—small Cu⁰ nanoparticles stabilized by strong interaction with the silica support, which create the ideal landscape for facile hydrogen activation at ambient pressure.

3.7 XPS and XAES

The surface copper species valence states of all reduced catalysts were examined using X-ray photoelectron spectroscopy (XPS) and X-ray induced Auger spectra (XAES), and the results obtained were shown in Fig. 8 and 9. The electronic peak at 952.3 eV was attributed to the binding energy of the 2P_1/2 orbital of Cu. No satellite peaks associated with Cu²⁺, which typically located in the range of 940–945 eV, were observed, suggesting that the copper species on the surface of the reduced catalysts were existed in low valence state. Typically, the binding energy of the 2p_3/2 orbital of Cu⁰ was around 932.5 eV, whereas that of the 2p_3/2 orbital of Cu⁺ floats in the region of 932–933 eV. The high degree of overlap between those makes it difficult to determine the attribution of the electronic peak at 932.6 eV accurately.^3b,5

Fig. 8 Cu 2p XPS spectra of reduced catalysts.

Fig. 9 Cu LMM Auger spectra of reduced catalysts.

To accurately distinguish the contents of Cu⁺ and Cu⁰ on the surface of the reduced catalysts, X-ray induced Auger spectra (XAES) tests were conducted. The Cu LMM spectra of all the reduced catalysts exhibited asymmetric and broad peaks, which indicated the presence of both Cu⁺ and Cu⁰, the exact contents of Cu⁺ and Cu⁰ were determined by integrating the test results. The Cu⁰ contents of the catalysts on the surface were listed in Table 1. The surface Cu⁰ content of the Gel-HT series catalysts showed a significant positive correlation with copper loading. When the copper loading increased from 12.2% to 30.8%, the Cu⁰ content on the surface of reduced catalyst increased correspondingly from 41.8% to 65.7%. This trend indicated that the increase in copper loading strengthened the reduction tendency of Cu/SiO₂ catalysts.

Numerous studies have demonstrated that the copper species in calcinated Cu/SiO₂ catalysts primarily exist in the form of CuO and copper silicate.^12b,17 The former can be directly reduced to Cu⁰ easily, whereas the latter exist as Cu⁺ after reduction due to strong metal–support interactions. The FT-IR characterization above has confirmed that more copper silicate was produced in catalysts containing high levels of copper. Meanwhile, N₂O titration, XRD and TEM results have confirmed that excessive loading deteriorated the dispersion state of copper species in the catalysts, contributing to their agglomeration into large-sized particles. At the same time, the fixed amount of TEOS in the gel formation stage means that the amount of silica hydroxyl generated by hydrolysis is also fixed. Increasing the amount of copper acetate to enhance copper loading inevitably results in a proportion of free Cu²⁺ that cannot be adsorbed and anchored by the carrier precursor thus agglomerated into CuO particles when calcinated at elevated temperature.

When the copper loading is similar, the preparation method plays a significant role in regulating the valence distribution on the surface of the reduced catalyst. The AEP catalyst contains little copper silicate, and the copper species primarily exists in the form of easily reduced CuO. Consequently, the Cu⁰ content on the surface of the reduced AEP catalyst is as high as 61.5%. For Des catalysts prepared with sodium silicate as silicon source and precipitant, the rapid nucleation process during precipitation formed a large number of tiny Cu-containing precipitate. The decomposition and reforming of the precursor during the high-temperature process led to partial embedding of the Cu species in the carrier and restricted contact with hydrogen. Thus the Cu⁰ content on the surface of the reduced Des catalysts was as low as 43.3%. The gel-hydrothermal method promotes copper dispersion by anchoring Cu²⁺ to silicon hydroxyl groups, and inhibits particle migration during the reduction process by strengthening copper–silicon chemical bonding in hydrothermal environment, which ensures that the copper species were duly reduced to obtain the appropriate Cu⁰ content while avoiding agglomeration and maintaining a high active specific surface area.

4. Catalytic performance

Gas-phase dimethyl oxalate hydrogenation reactions were conducted in a fixed-bed reactor to investigate the catalytic performance of catalysts. With regard to the impact of reaction temperature and LHSV on catalyst activity, numerous studies have demonstrated that increasing temperature or decreasing LHSV can promote the conversion of DMO at the expense of MG selectivity. Copper-based catalysts are susceptible to deactivation at low temperatures and high LHSV conditions, hindering stable operation over a long period of time.^3b,12,17,23 In this study, the LHSV of DMO was fixed at 0.31 h⁻¹ and the reaction temperature was determined to be 170 °C. Additionally, the molar ratio of H₂/DMO was varied from 25 to 200 to assess the impact of hydrogen supply; the reaction pressure was altered from 0.1 MPa to 2 MPa to examine the effect of pressure on catalyst activity.

Table 2 presents the catalytic activity of each catalyst in the hydrogenation of DMO at atmospheric pressure. From the reactivity data of the Gel-HT series catalysts, the DMO conversion increased from 59.8% to 92.5% when the actual copper loading was increased from 12.20% to 24.78%, which reflected that at least 20% of actual copper loading is necessary in order to ensure a sufficient number of catalytically active sites and the corresponding catalytic activity. The catalytic activity of 45% Gel-HT was lower than that of 35% Gel-HT despite its highest Cu loading, which is consistent with literature reports that excessive metal loading tends to result in the growth of the active metal particles and a decrease in the active specific surface area, thus weakening the catalytic activity.^12,24

Table 2 Catalytic performances and TOF values of the catalysts^a

Catalyst	DMO conversion (%)	Selectivity to MG (%)	Selectivity to EG (%)	TOF^b (h^−1)
a Reaction conditions: T = 170 °C, P = 0.1 MPa, H2/DMO = 50, LHSV = 0.31 h^−1. b TOF was calculated by copper dispersion and the DMO conversion was controlled below 30% by increasing the LHSV.
Des	37.5	96.2	3.2	0.83
AEP	67.4	94.2	4.7	2.73
15% Gel-HT	59.8	92.4	7.1	3.05
25% Gel-HT	94.7	65.2	32.7	6.14
35% Gel-HT	92.5	84.5	14.5	5.81
45% Gel-HT	74.3	75.1	24.1	4.50

---

In view of the crucial influence of copper loading on catalyst activity, we prepared AEP and Des catalysts with copper loading very close to that of 35% Gel-HT. The catalytic activity of both in the atmospheric pressure DMO hydrogenation reaction was far inferior to that of the catalyst 35% Gel-HT, with DMO conversion of 67.4% for AEP and as low as 37.5% for Des. According to the preceding characterization results, there are no considerable discrepancies in the parameters of specific surface area, average pore size and Cu dispersion. The results of XAES and N₂O titration tests showed significant differences in the Cu⁰ content on the catalyst surface and Cu⁰ specific surface area. Cu⁰ is the key center of H₂ activation in the DMO hydrogenation reaction; therefore, the difference in the Cu⁰ content of the three catalysts leads to the difference in their H₂ activation capacity, especially at low temperatures below the reaction temperature, as evidenced by dramatic difference in H₂-TPD spectra for the desorption peak at 120 °C.^3d,25

Fig. 10 presents the activity of all catalysts for catalyzing DMO hydrogenation at 0.1–2 MPa. As showed in Fig. 10(a), increasing the reaction pressure promoted the conversion of the feedstock. All samples except the 15% Gel-HT sample with low copper loading achieved over 90% DMO conversion at 2 MPa. Notably, the growth rate of DMO conversion with increasing pressure was significantly accelerated for Des catalysts when the reaction pressure exceeded 1.0 MPa, reflecting the strong dependence of its catalytic activity on the reaction pressure. One possible reason for this pattern is that the elevated hydrogen partial pressure encouraged the adsorption and activation of hydrogen on the catalyst surface, compensating for the insufficient hydrogen activation capacity of the Des catalyst. Balancing the selectivity of the semi-hydrogenated product MG and the fully hydrogenated product EG over Cu/SiO₂ catalysts in the selective hydrogenation of DMO is tricky, as shown in Fig. 10(b) and (c), the elevated reaction pressures significantly reduced MG selectivity while improving EG selectivity. This can be attributed to the fact that the augmented system pressure impedes the detachment of MG from the catalyst surface, compelling it to persist in the reaction within a hydrogen-rich environment. In order to further understand the effect of reaction pressure on the DMO hydrogenation performance of Cu/SiO₂ catalysts, Fig. 10(d) demonstrates the equilibrium relationship between DMO conversion and MG selectivity based on the reaction pressure for all catalysts. An increase in DMO conversion for each catalyst leads to a decrease in MG selectivity. These findings suggest that the reaction pressure exerts a significant influence on the product distribution of the DMO hydrogenation reaction. When the DMO conversion was increased by boosting the pressure, any slight increase in pressure resulted in a substantial decrease in MG selectivity. This strongly suggests that the strategy of balancing the DMO conversion and MG selectivity of Cu/SiO₂ catalysts by adjusting the reaction pressure is inadvisable.

Fig. 10 Effect of reaction pressure on catalytic performance: (a) conversion of DMO, (b) selectivity of MG, (c) selectivity of EG, (d) equilibrium relationship between DMO conversion and MG selectivity. Reaction condition: T = 170 °C, H₂/DMO = 50, LHSV = 0.31 h⁻¹.

To ascertain the impact of hydrogen supply on catalyst performance, the magnitude of the hydrogen ester ratio was employed as an indicator of the strength of the hydrogen supply. Data on the reaction performance of the catalyst at varying hydrogen ester ratios were obtained, and the results were presented in Fig. 11. It was observed that the feedstock conversion and the selectivity of the fully hydrogenated product EG increased in conjunction with the rise in the hydrogen ester ratio, which was mainly attributed to the increase of hydrogen supply facilitating the adsorption and activation process of hydrogen at the Cu⁰ active sites, and increasing the concentration of active hydrogen species on the catalyst surface. The high concentration of active hydrogen species on the catalyst surface can effectively reduce the hydrogenation reaction energy barrier and provide sufficient guarantee for the hydrogenation of DMO.

Fig. 11 Effect of H₂/DMO on catalytic performance: (a) conversion of DMO, (b) selectivity of MG, (c) selectivity of EG, (d) equilibrium relationship between DMO conversion and MG selectivity. Reaction condition: P = 0.1 MPa, T = 170 °C, LHSV = 0.31 h⁻¹.

Although the copper loading or preparation method of each Cu/SiO₂ catalyst varied considerably, the equilibrium relationship curves of DMO conversion and MG selectivity based on the H₂/DMO in Fig. 11(d) showed nearly parallel trends, it reflects that the strategy of regulating the activity by controlling the hydrogen supply conditions for atmospheric pressure hydrogenation of DMO to MG over Cu/SiO₂ catalyst. Moreover, compared to controlling the pressure of the reaction system, the simple adjustment of the hydrogen flow rate is a more rational and convenient way of operation.

The lifespan of a catalyst is an essential metric for evaluating its performance. Fig. 12 and 13 illustrated the results of long-term operation of 35% Gel-HT catalysts at 0.1 MPa and 2 MPa, respectively. As clearly evidenced by the data, the catalyst exhibits extraordinary stability throughout the entire 5000-hour duration. The DMO conversion rate is maintained steadily at approximately 92.5%, and the selectivity towards MG remains consistently high at around 84.5%. This simultaneous stability in both conversion and MG selectivity is a remarkable achievement, as it demonstrates that the catalyst not only remains active but also precisely preserves its intrinsic selectivity over prolonged operation. The constant low level of ethylene glycol (EG) selectivity confirms that the catalyst successfully suppresses the over-hydrogenation of MG to EG. The catalysts were observed to undergo three distinct stages of initial, stable, and deactivation when operated at elevated pressure. DMO conversion decreased from >99% to ∼93% and finally less than 60%. The decaying trend of DMO conversion with reaction time indicated that the catalysts were gradually deactivated. The notable discrepancy in the outcomes of long-term testing at atmospheric and elevated pressures indicates that low pressures are conducive to sustaining consistent catalyst activity.

Fig. 12 Stability test of catalyst at 0.1 Mpa. Reaction condition: T = 170 °C, H₂/DMO = 50, LHSV = 0.31 h⁻¹.

Fig. 13 Stability test of catalyst at 2 Mpa. Reaction condition: T = 170 °C, H₂/DMO = 50, LHSV = 0.31 h⁻¹.

5. Conclusion

An innovative gel-hydrothermal method has been successfully applied to prepare Cu/SiO₂ catalysts for the hydrogenation of DMO to MG. Compared with other preparation approaches, the gel-hydrothermal method developed in this work enables precise control over the anchoring of copper species and the structural reconstruction of the silica framework, which effectively promotes the formation of highly dispersed copper phyllosilicates and enhances the Cu–O–Si interfacial interaction. As a result, the reduced catalyst simultaneously achieves high copper dispersion and an optimal surface Cu⁰ content. This distinctive structural configuration imparts exceptional low-temperature hydrogen activation capability to the catalyst, facilitating the generation of abundant weakly adsorbed hydrogen species under ambient pressure conditions. Such a feature significantly increases the surface concentration of active hydrogen species, thereby driving the efficient hydrogenation of DMO with high selectivity toward MG. The study not only enriches Cu/SiO₂ catalyst preparation methods but also provides an effective strategy for the selective hydrogenation of DMO to MG, which may inspire more researchers to carry out meaningful work and promote the industrial application of Cu-based DMO atmospheric pressure hydrogenation catalysts.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper.

Acknowledgements

We especially acknowledge the support of National Key Reach and Development Program of China (2023YFB4103300) and the Fundamental Research Program of Shanxi Province (202303021212377).

References

1. (a) R. J. Shang, S. F. Mu, L. C. Wei and Y. L. Jiang, Nat. Gas Chem. Ind., 2015, 40, 86–90; (b) Y. Cheng, S. Liu, J. Jiao, M. Zhou, Y. Wang, X. Xing, Z. Chen, X. Sun, Q. Zhu and Q. Qian, J. Am. Chem. Soc., 2024, 146, 10084–10096; (c) K. S. Yu, G. J. Geng, X. Z. Yang, P. Liang and Q. L. Shi, Technol. Dev. Chem. Ind., 2019, 48, 27–29; (d) B. W. Wang, K. S. Tian and G. H. Xu, Nat. Gas Chem. Ind., 2005, 30, 64–69.
2. (a) H. Wang, T. Lan and G. Zhao, ACS Catal., 2024, 14, 728–740; (b) T. Hayashi, T. Inagaki, N. Itayama and H. Baba, Catal. Today, 2006, 117, 210–213; (c) T. Xie, S. Ai, Y. Huang, W. Yu and C. Huang, Sep. Purif. Technol., 2021, 268, 118718.
3. (a) J. Gong, H. Zhang and W. Shu, Mol. Catal., 2023, 550, 113592; (b) Y. Zhao, H. Zhang, Y. Xu, S. Wang and X. Ma, J. Energy Chem., 2020, 49, 248–256; (c) Z. Zhuang, Y. Li, F. Chen, X. Chen, Z. Li, S. Wang, X. Wang, H. Zhu, Y. Tan and Y. Ding, Chem. Commun., 2022, 58, 1958–1961; (d) F. Li, C. S. Lu and X. N. Li, Chin. Chem. Lett., 2014, 25, 1461–1465; (e) S. Cheng, T. Meng, D. Mao, X. Guo and J. Yu, ACS Omega, 2022, 7, 41224–41235.
4. (a) G. L. Dong, Z. Luo, Y. Q. Cao, S. N. Zheng, J. H. Zhou, W. li and X. G. Zhou, J. Catal., 2021, 401, 252–261; (b) X. Xu, X. X. Hu, Z. Luo, Y. Cao, Y. Zhu, W. Li, J. Zhou and X. Zhou, New J. Chem., 2023, 47, 6045–6049; (c) L. Wu, M. Guo, X. Xu, J. Ding and Q. Zhong, J. Environ. Chem. Eng., 2024, 12, 113037.
5. G. Giorgianni, C. Mebrahtu and S. Perathoner, Catal. Today, 2022, 391, 343–353.
6. (a) X. Chen, C. Wen and Y. Cui, Appl. Catal., B, 2015, 162, 483–493; (b) R. P. Ye, L. Lin, J. X. Yang, M. L. Sun, F. Li, B. Li and Y. G. Yao, J. Catal., 2017, 350, 122–132; (c) C. Wen, Y. Cui, A. Yin, K. Fan and W. L. Dai, ChemCatChem, 2013, 5, 138–141.
7. (a) Y. Liu, J. Ding, J. Yang, J. Bi, K. Liu and J. Chen, Catal. Commun., 2017, 98, 43–46; (b) D. Wang, C. C. Zhang, M. Zhu, F. Yu and B. Dai, ChemistrySelect, 2017, 2, 4823–4829; (c) S. Zhao, H. Yue, Y. Zhao, B. Wang, Y. Geng, J. Lv, S. Wang, J. Gong and X. Ma, J. Catal., 2013, 297, 142–150; (d) Y. N. Wang, X. Duan, J. Zheng, H. Lin and K. Asakura, Catal. Sci. Technol., 2012, 2, 1637–1639.
8. (a) X. Zheng, H. Lin, J. Zheng, X. Duan and Y. Yuan, ACS Catal., 2013, 3, 2738–2749; (b) H. R. Yue, Y. J. Zhao, S. Zhao, B. Wang and J. L. Gong, Nat. Commun., 2013, 4, 2339–2346; (c) Z. Li, Y. Li and X. Wang, Chem. Eng. J., 2023, 454, 140001; (d) G. Liu, G. Yang, X. Peng, J. Wu and N. Tsubaki, Chem. Soc. Rev., 2022, 51, 5606–5659.
9. (a) A. Yin, X. Guo, W. L. Dai, H. Li and K. Fan, Appl. Catal., A, 2008, 349, 91–99; (b) S. Zhao, H. R. Yue, Y. J. Zhao, B. Wang, Y. C. Geng, J. Lv, S. P. Wang, J. L. Gong and X. B. Ma, J. Catal., 2013, 297, 142–150; (c) J. Zheng, L. Ye, S. C. E. Tsang and Y. Youzhu, Appl. Catal., A, 2015, 505, 344–353.
10. J. Zheng, L. Huang, C. H. Cui, Z. C. Chen, X. F. Liu, X. Duan, X. Y. Cao, T. Z. Yang, H. Zhu and K. Shi, Science, 2022, 376, 288–292.
11. Y. Zhao, X. Kan, H. Yun, D. Wang and J. Shen, Catal. Commun., 2021, 154, 106310.
12. (a) Y. Wang, X. B. Ma, Y. Q. Zhang, Y. Xu and Y. J. Zhao, Appl. Catal., A, 2017, 539, 59–69; (b) Y. J. Zhao, S. Li, Y. Wang, B. Shan, J. Zhang, S. Wang and X. B. Ma, Chem. Eng. J., 2017, 313, 759–768.
13. (a) X. B. Ma, H. Chi, H. R. Yue, Y. J. Zhao, Y. Xu, J. Lv, S. Wang and J. L. Gong, AIChE J., 2013, 59, 2530–2539; (b) G. C. Chinchen, C. M. Hay, H. D. Vandervell and K. C. Waugh, J. Catal., 1987, 103, 79–86.
14. J. Rouquerol, D. Avnir, C. W. Fairbridge, D. H. Everett, J. M. Haynes, N. Pernicone, J. D. F. Ramsay, K. S. W. Sing and K. K. Unger, Pure Appl. Chem., 1994, 66, 1739–1758.
15. T. Toupance, M. Kermarec and C. Louis, J. Phys. Chem. B, 2000, 104, 965–972.
16. (a) L. F. Chen, P. J. Guo, M. H. Qiao, S. R. Yan, H. X. Li, W. Shen, H. L. Xu and K. N. Fan, J. Catal., 2008, 257, 172–180; (b) T. Toupance, M. Kermarec, J. F. Lambert and C. Louis, J. Phys. Chem. B, 2002, 106, 2277–2286.
17. J. Ding, T. Popa, J. K. Tang, K. A. M. Gasem, M. H. Fan and Q. Zhong, Appl. Catal., B, 2017, 209, 530–542.
18. (a) C. J. G. Van Der Grift, A. F. H. Wielers, A. Mulder and J. W. Geus, Thermochim. Acta, 1990, 171, 95–113; (b) J. G. C. Van Der Grift, A. F. H. Wielers, B. P. J. Jogh, J. Van Beunum, M. De Boer, M. Versluijs-Helder and J. W. Geus, J. Catal., 1991, 131, 178–189.
19. (a) C. Ye, C. Guo, C. Sun and Y. Zhang, RSC Adv., 2016, 6, 113796–113802; (b) H. Qin, C. Guo, C. Sun and J. Zhang, J. Mol. Catal. A: Chem., 2015, 409, 79–84; (c) Y. J. Zhao, S. Zhao, Y. Geng, Y. Shen, H. R. Yue, J. Lv, S. Wang and X. B. Ma, Catal. Today, 2016, 276, 28–35.
20. J. Schittkowski, K. Tlle, S. Anke, S. Stürmer and M. Muhler, J. Catal., 2017, 352, 120–129.
21. (a) S. J. Techner, A. R. Mazabrard, G. Pajonk, G. E. Gardes and C. Hoang-Van, J. Colloid Interface Sci., 1977, 58, 88–99; (b) W. C. Conner Jr and J. L. Falconer, Chem. Rev., 1995, 95, 759–788; (c) A. G. Boudjahem, M. Pietrowski, S. Monteverdi, M. Mercy and M. M. Bettahar, J. Mater. Sci., 2006, 41, 2025–2030.
22. (a) D. Duprez, J. Barbier, Z. F. Hamida and M. Bettahar, Appl. Catal., 1984, 12, 219–225; (b) S. Cȩckiewicz and B. Delmon, J. Catal., 1987, 108, 294–303.
23. S. Cheng, T. Meng, D. Mao, X. Guo, J. Yu and Z. Ma, Nanomaterials, 2022, 12, 407–415.
24. B. Zhang, S. G. Hui, S. H. Zhang, Y. Ji, W. li and D. Y. Fang, J. Nat. Gas Chem., 2012, 21, 563–570.
25. (a) L. He, H. Cheng, G. Liang, Y. Yu and F. Zhao, Appl. Catal., A, 2013, 452, 88–93; (b) X. B. Ma, Z. Yang, X. Liu, X. Tan and Q. Ge, RSC Adv., 2015, 5, 37581–37584.

---