Cu–Ni _x O _y /SiO ₂ Catalysts for Hydrogenation of Dimethyl Oxalate to Ethylene Glycol: Regulation of Coprecipitation Process and Structure–Activity Relationship Study

Abstract

The silica-supported Cu–Ni _x O _y hybrid nanostructures (Cu–Ni _x O _y /SiO₂) were fabricated via a coprecipitation method with systematic optimization of precipitation temperatures and the reagent addition sequence. At room temperature, the first addition of aqueous ammonia to the silica colloids with subsequent addition of Cu precursors and further treatments obtained Cu/SiO₂–FA-RT with better performance for the hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG). The introduction of Ni _x O _y by the coprecipitation of Cu and Ni precursors further enhances catalytic performance due to the formation of Cu–Ni _x O _y hybrid nanostructures. During the hydrogenation of DMO to EG in a fixed-bed reactor, the Cu–Ni _x O _y /SiO₂ catalysts achieve 98.9% DMO conversion and 89.2% EG selectivity, outperforming the individual Cu/SiO₂–FA-RT. The stability test of Cu–Ni _x O _y /SiO₂ demonstrates nearly no activity/selectivity decay, with only 0.07 wt % coke deposition. The performance enhancement should originate from the formation of the close-contact Cu–Ni _x O _y interfaces, which modulate the electronic/geometric properties of the active surfaces, thus improving the hydrogenation performance and preventing coke deposition.

1. Introduction

In recent decades, the catalytic utilization of coal resources for synthesizing high-value chemicals has received more attention. Among the various catalytic applications of coal, the syngas-based route to manufacture ethylene glycol (EG) is one of the most important processes, , offering an economical alternative to the conventional petroleum-derived method. EG is extensively utilized in polyester fibers, antifreeze agents, lubricants, plasticizers, surfactants, and polyester films, − and the catalytic hydrogenation of dimethyl oxalate (DMO) to EG over transition-metal catalysts serves as a pivotal step in coal-based EG production. With the global demand for EG continuously rising, the development of more efficient catalysts for EG production is highly desirable. ,,

Two main categories of catalysts have been developed for DMO hydrogenation: noble-metal-based catalysts (including Ag, Au, and Pd) and non-noble Cu-based systems. Silver-based catalysts feature moderate hydrogenation activity and negligible C–C bond cleavage tendency, delivering exceptional selectivity toward methyl glycolate (MG) in the selective hydrogenation of DMO. Meanwhile, rationally designed bimetallic noble-metal systems (such as Pd–Au alloys) and noble-metal-promoted Cu catalysts , have also been demonstrated to be effective for the deep hydrogenation of DMO to EG. Nevertheless, the high cost and natural scarcity of noble metals severely restrict their large-scale deployment in industrial coal-to-EG production, rendering cost-effective non-noble Cu-based catalysts as the mainstream research direction.

Cu-based catalysts are widely employed in DMO hydrogenation due to their high efficiency in selectively hydrogenating CO bonds without cleaving C–C bonds coupled with their relatively low raw material costs. Their exceptional catalytic performance originates from the synergistic effect between Cu⁰ and Cu⁺ species, where Cu⁰ species facilitate H₂ dissociation and dominate the hydrogenation process, while Cu⁺ species promote DMO activation and stabilize reaction intermediates. , It is generally accepted that the Cu precipitation process for supported Cu catalysts has a significant effect on their hydrogenation performance. For example, Ding et al. utilized ammonium carbonate deposition-precipitation to prepare Cu/SiO₂ for selective hydrogenation of DMO to EG, while Zhao et al. fabricated ordered mesoporous silica-supported Cu catalysts via an ammonia evaporation method under weak alkaline conditions. The underlying principle is to obtain optimized Cu dispersion and metal–support interaction, thus achieving good performance for hydrogenation of DMO to EG. −

Besides the synthetic parameters during the Cu precipitation process, the metal components of Cu-based catalysts also have a significant effect on their hydrogenation performance. In the literature, a second metal or metal oxide such as W, Ag, , Zr, , Ce, La, In, Ni, , Pt, Pd, and Fe has been used to promote the performance of Cu-based catalysts for hydrogenation reactions. For instance, lanthanum oxide-modified Cu/SiO₂ (La–Cu/SiO₂) catalysts greatly improve the catalytic performance and stability, while bimetallic Cu–Ag/SiO₂ enhances the catalytic efficiency for the hydrogenation of DMO to EG. These modifications could tune surface electronic/geometric properties and/or the Cu⁰/Cu⁺ ratio, thereby improving their hydrogenation performance.

Despite the progress achieved with Cu-based catalysts for DMO hydrogenation, several critical challenges remain. The most prominent issues include insufficient catalytic activity, poor stability, carbon deposition susceptibility, and a short lifespan during industrial applications. − From an industrial economy perspective, conventional Cu-based catalysts require elevated reaction temperatures and high H₂/DMO molar ratios to achieve target conversion, which inevitably increases the energy consumption of hydrogen circulation compression and raises overall operating costs. , Regarding catalyst deactivation, sintering of Cu nanoparticles and carbonaceous deposition are recognized as the two dominant deactivation pathways during long-term operation. ,, Carbon deposition not only covers the active Cu sites and blocks catalyst pore channels but also gradually raises the pressure drop of the fixed-bed reactor, forcing frequent catalyst regeneration or replacement and further increasing production costs. In addition, current catalytic processes generally suffer from unsatisfactory EG selectivity, generating byproducts such as ethanol, 1,2-butanediol, and other high-boiling alcohol impurities. − These impurities exhibit physicochemical properties highly similar to those of EG, which greatly complicates downstream distillation and purification and in turn incurs extra equipment investment and operating costs for the coal-to-EG route.

Among various promoters, Ni is one of the most investigated modifiers for Cu-based DMO hydrogenation catalysts. , Existing studies on Ni-modified Cu catalysts are mainly divided into two categories: high-content NiO modifiers that improve catalytic activity mainly by enhancing Cu dispersion and constructing Cu–Ni metallic alloys that shift product selectivity toward MG. However, most reported Ni-modified Cu-based catalysts rely on lab-scale synthetic routes with presynthesized mesoporous or phyllosilicate supports, which involve complicated preparation procedures and are difficult for industrial scale-up. Meanwhile, the interfacial valence regulation effect of low-content amorphous Ni _x O _y on Cu active sites has rarely been systematically clarified, and the anticoking performance under industrial-like operating conditions remains to be further investigated.

Herein, we synergistically optimized the preparation process of Cu/SiO₂ catalysts and used a second metal to promote their catalytic efficiency for the gas-phase hydrogenation of DMO to EG. Different from most reported lab-scale synthesis routes, we developed a facile one-step coprecipitation method using industrial fumed silica as the raw material, which is suitable for large-scale extrusion molding. By controlling precipitation temperatures and the reagent addition sequence, Cu/SiO₂ with good catalytic performance was achieved. The introduction of Ni oxides into Cu further enhances its catalytic performance through the formation of close-contact Cu–Ni _x O _y interfaces. At an extremely low Ni/Cu molar ratio of 1/10, amorphous Ni _x O _y precisely regulates the surface Cu⁰/Cu⁺ ratio rather than merely improving Cu dispersion. The resulting Cu–Ni _x O _y /SiO₂ demonstrated 98.9% DMO conversion and an EG selectivity of 89.2% in a fixed-bed reactor, with no apparent performance degradation observed during the stability test. Only 0.07 wt % coke deposition is detected after long-term operation, demonstrating excellent anticoking ability critical for industrial application. The catalytic performance enhancement of Cu–Ni _x O _y /SiO₂ originates from the optimal synthetic parameters and the synergistic effect between Cu and Ni _x O _y .

2. Experimental Section

2.1. Chemicals and Materials Characterizations

The chemicals used and the techniques for material characterization are shown in the Supporting Information (SI).

2.2. Catalyst Preparation

2.2.1. Optimization of Synthetic Parameters

A portion of 57.0 g of Cu­(NO₃)₂·3H₂O and 60.0 g of fumed silica powder were dispersed in 600 g of deionized water under vigorous magnetic stirring at a constant speed of 500 rpm at room temperature for 30 min, followed by the addition of 30.3 g of aqueous ammonia solution in one portion within approximately 10 s under continuous stirring. After aging for 30 min, filtration, and washing, the precipitates were extruded into cylindrical shapes. The precursor was then heated from room temperature to 80 °C at a ramp rate of 2 °C/min and dried at 80 °C for 12 h, followed by calcination at 450 °C for 3.0 h with an identical heating ramp from room temperature, yielding the CuO/SiO₂ catalyst. The catalyst was further crushed into pieces with diameters of 5–7 mm and heights of 3–5 mm. The obtained materials and the subsequently reduced ones (reduction in the reactor) were named as CuO/SiO₂–FC-RT and Cu/SiO₂–FC-RT, respectively, where FC and RT represent the first addition of copper and room temperatures during the precipitation process, respectively. For CuO/SiO₂–FA-RT, the synthetic process is same as those of CuO/SiO₂–FC-RT except that aqueous ammonia was added first, where FA represents the first addition of ammonia.

For the synthesis of CuO/SiO₂–FC-70 and CuO/SiO₂–FA-70, their synthetic procedures are the same as those of CuO/SiO₂–FC-RT and CuO/SiO₂–FA-RT, respectively, except that the precipitation process was performed at 70 °C. The Cu loadings determined by an inductively coupled plasma optical emission spectrometer (ICP-OES) are fixed at 20.0 wt % for Cu/SiO₂–FC-RT, Cu/SiO₂–FA-RT, Cu/SiO₂–FC-70, and Cu/SiO₂–FA-70.

2.2.2. Preparation of Cu–Ni _x O _y /SiO₂ Catalysts

The Cu–Ni _x O _y /SiO₂ catalysts were synthesized by similar procedures to those used for CuO/SiO₂–FA-RT except that the Ni precursors were coadded with the Cu ones at a molar ratio of Ni/Cu of 1/10, and the amount of ammonia solution was increased by an additional 10% molar equivalent. The additional ammonia compensates for the ammonia consumed by Ni²⁺ ammine complexation, ensuring an identical pH environment and copper-ammine equilibrium as those of the monometallic system to maintain a single preparation variable. The stirring speed, ammonia addition mode, and thermal treatment ramp rates are consistent with the above monometallic catalyst preparation protocol. The Cu and Ni loadings determined by ICP-OES are 20.0 and 1.85 wt %, respectively.

2.3. Catalyst Performance Evaluation

The catalytic activity tests were conducted in a fixed-bed tubular stainless steel reactor with a 40 mm inner diameter. The reactor was loaded with 40 mL of calcined catalysts in the center with glass wool placed in the upper and lower parts of the tube. The calcined catalysts were first reduced with an H₂/N₂ gas mixture (8/92 v/v) at a gas flow rate of 1800 mL/min at 230 °C for 5.0 h (heating rate of 1.0 °C/min). Then, the reactor was cooled to 185 °C and subsequently fed with a methanol solution containing 15.0 wt % DMO to initiate the reaction. The typical hydrogenation experiments were carried out at a temperature of 185 °C, an H₂ pressure of 2.6 MPa, an H₂/DMO molar ratio of 60/1, and a DMO liquid hourly space velocity (LHSV) of 0.8 h^–1 (based on 15 wt % DMO methanol solution). Sampling was performed after 5.0 h of stable operation. Product analysis was carried out using a GC9560 gas chromatograph equipped with a flame ionization detector (FID). The detailed analysis method can be found in the SI. The calculations for conversion (X), selectivity to EG (S_EG), and selectivity to methyl glycolate (S_MG) are provided in the SI. The carbon balance was above 98% in all catalytic tests.

3. Results and Discussion

3.1. The Effect of Synthetic Parameters of Cu/SiO₂ on DMO Hydrogenation

In this work, the influences of reagent addition sequence and coprecipitation temperatures on the catalytic performance of Cu/SiO₂ for DMO hydrogenation were first studied. As shown in entries 1 and 2 in Table , when aqueous ammonia was added first (Cu/SiO₂–FA-RT), the catalysts showed a higher conversion of 78.1% and EG selectivity of 78.4% compared with a conversion of 77.2% and EG selectivity of 77.8% for the one with the first addition of copper. Moreover, increasing the precipitation temperatures to 70 °C decreases the activity of Cu/SiO₂ for DMO hydrogenation. As illustrated in entries 3 and 4, Cu/SiO₂ prepared with a 70 °C precipitation temperature shows worse performance than those synthesized at room temperature, where Cu/SiO₂–FA-70 demonstrates the lowest activity with only 26.0% EG selectivity.

1. DMO Hydrogenation over Cu/SiO₂ Prepared with Different Reagent Addition Sequence and Precipitation Temperatures.

Entry	Catalysts	XDMO/%	SMG/%	SEG/%
1	Cu/SiO2–FC-RT	77.2	17.4	77.8
2	Cu/SiO2–FA-RT	78.1	16.6	78.4
3	Cu/SiO2–FC-70	76.5	18.7	76.6
4	Cu/SiO2–FA-70	65.0	73.0	26.0

^aReaction conditions: H₂ pressure, 2.6 MPa; reaction temperature, 185 °C; molar ratio of H₂/DMO, 60/1; LHSV of 0.8 h^–1, based on 15 wt % DMO methanol solution.

^bOther byproducts include ethanol, 1,2-butanediol, and extreme trace amounts of methyl formate and methyl acetate.

Figure S1 presents the images of transmission electron microscopy (TEM) for various samples. As shown in Figure S1, silica and Cu are uniformly dispersed, and differentiating Cu from silica is difficult due to the low contrast between Cu and SiO₂ even when the Cu loading is 20 wt %. Figure S2 presents the N₂ adsorption–desorption isotherms and the corresponding pore size distributions of Cu/SiO₂ catalysts prepared with different synthetic parameters, while their BET specific surface areas and average pore sizes are summarized in Table S1. As shown in Figure S2a, all materials present typical mesoporous structures with hysteresis loops, and 10–40 nm mesopores are clearly present in these materials (Figure S2b). Among these four materials, Cu/SiO₂–FC-RT demonstrates the highest BET specific surface area of 316.6 m²/g (Table S1), while Cu/SiO₂–FA-70 shows the lowest specific surface area of 241.0 m²/g. It is apparent that the ones with a precipitation temperature of 70 °C illustrate lower BET specific surface areas. For example, Cu/SiO₂–FC-70 shows lower specific surface areas than Cu/SiO₂–FC-RT, while Cu/SiO₂–FA-70 demonstrates lower specific surface areas than Cu/SiO₂–FA-RT, possibly due to the formation of larger particle sizes at higher precipitation temperatures. Here, it is mentioned that the first addition of ammonia and coprecipitation at 70 °C could cause the loss of ammonia at the relatively high temperature of 70 °C, which is possibly the reason why Cu/SiO₂–FA-70 showed the lowest catalytic activity (entry 4 in Table ).

Figure a presents the X-ray diffraction (XRD) patterns of Cu/SiO₂ synthesized with different synthetic parameters. The broad peak observed at ∼22° 2θ in Figure is assigned to SiO₂. The sharp diffraction peaks at 43.3°, 50.4°, and 74.1° could be ascribed to the (111), (200), and (220) crystal planes of metallic Cu, respectively, confirming the complete reduction of oxidized Cu to metallic Cu by H₂ reduction at 230 °C. According to the Scherrer formula, , the sizes of Cu nanoparticles (NPs) calculated by the full-width at half-maximum (fwhm) values in Figure are 30.5 nm for Cu/SiO₂–FA-RT, 31.7 nm for Cu/SiO₂–FC-RT, 33.8 nm for Cu/SiO₂–FC-70, and 34.8 nm for Cu/SiO₂–FA-70. The relatively larger particle sizes are consistent with their sharp XRD diffractions and high Cu loading. Moreover, coprecipitations at room temperature (Cu/SiO₂–FC-RT and Cu/SiO₂–FA-RT) achieved relatively smaller Cu NPs than those at 70 °C (Cu/SiO₂–FC-70 and Cu/SiO₂–FA-70). Among these four materials, Cu/SiO₂–FA-RT had the smallest Cu particle sizes and achieved the highest conversions and EG selectivity for DMO hydrogenation.

1.

a) XRD patterns of Cu/SiO₂ prepared with different synthetic parameters; b) XRD patterns of Cu–Ni _x O _y /SiO₂.

It is generally accepted that the reagent addition sequence directly modulates the pH evolution trajectory of the slurry throughout the precipitation process, thereby governing the particle size and dispersion of active NPs, metal–support interaction, and the textural properties of the resulting catalysts. For the FC-RT protocol (copper precursor added first, followed by ammonia addition), the initial mixture of copper nitrate and fumed silica colloid is weakly acidic (pH ≈ 3–4) due to the hydrolysis of Cu²⁺. Upon the introduction of aqueous ammonia, the bulk pH rises gradually from acidic to neutral and then to a weakly alkaline regime. The slow diffusion of ammonia creates transient spatial pH gradients within the slurry, causing locally high supersaturation of Cu²⁺ and triggering homogeneous nucleation in the bulk solution. The initially formed Cu-based nuclei then act as substrates for subsequent heterogeneous growth and aggregation, eventually yielding CuO aggregates with relatively large particle size and broad size distribution. Meanwhile, the silica surface carries only weak negative charges under acidic-to-neutral conditions, resulting in insufficient electrostatic attraction to Cu²⁺ and consequently weaker metal–support interaction. In contrast, for the FA-RT protocol (ammonia added first, followed by copper precursor addition), the silica colloid is initially dispersed in an alkaline ammonia solution (pH ≈ 10–11). In this alkaline environment, the surface hydroxyl groups on the silica surface are fully deprotonated to form negatively charged Si–O^– sites. When copper nitrate solution is introduced, Cu²⁺ preferentially forms stable tetraamminecopper­(II) complex ions ([Cu­(NH₃)₄]²⁺) in the presence of excess ammonia, rather than precipitating directly as hydroxide. These positively charged complex ions are uniformly adsorbed onto the negatively charged silica surface via strong electrostatic interaction. During the subsequent aging process, the copper–ammonia complexes release Cu²⁺ slowly, and the liberated Cu species preferentially undergo heterogeneous nucleation on the negatively charged silica surface, resulting in highly dispersed Cu species with smaller particle size and enhanced metal–support interaction.

This pH-mediated formation mechanism is in good agreement with the experimental characterization results. As calculated from XRD patterns, the average Cu particle size of Cu/SiO₂–FA-RT (30.5 nm) is smaller than that of Cu/SiO₂–FC-RT (31.7 nm), verifying that the FA-RT protocol induces more uniform precipitation and better Cu dispersion, which is also consistent with the corresponding BET specific surface area data. Since the FA-RT protocol delivered the best catalytic performance, it was adopted for the preparation of silica-supported bimetallic catalysts in subsequent studies.

3.2. DMO Hydrogenation over Cu–Ni _x O _y /SiO₂ Catalysts

A series of Cu–M _x O _y /SiO₂ (M = Ni, Cr, Zn, and Fe) catalysts were prepared with the first addition of ammonia and coprecipitation of silica, Cu, and M precursors at room temperature, where M/Cu molar ratios were maintained at 1/10. Table summarizes the catalytic performance of Cu–M _x O _y /SiO₂ for DMO hydrogenation. Among these catalysts, Cu–Ni _x O _y /SiO₂ shows much higher DMO conversions and EG selectivity than monometallic Cu/SiO₂–FA-RT and other silica-supported Cu–Cr _x O _y , Cu–Zn _x O _y , and Cu–Fe _x O _y bimetallic catalysts. In contrast, the introduction of Cr _x O _y , Zn _x O _y , and Fe _x O _y even deteriorates the hydrogenation performance to varying degrees. The remarkable performance divergence among different transition-metal promoters confirms that the promotion effect is not a universal feature of transition-metal oxides but arises from the specific interfacial interaction between Cu and amorphous Ni _x O _y . Unlike high-content NiO modifiers that enhance performance mainly by improving Cu dispersion, the low-content Ni _x O _y in this work is proposed to play a dominant role in regulating the electronic state of the surface Cu species.

2. DMO Hydrogenation over Silica-Supported Cu-Based Bimetallic Catalysts.

Catalysts	XDMO/%	SMG/%	SEG/%
Cu–Ni x O y /SiO2	98.9	5.5	89.2
Cu–Cr x O y /SiO2	73.8	56.2	40.7
Cu–Zn x O y /SiO2	64.6	70.7	25.8
Cu–Fe x O y /SiO2	60.1	74.6	22.3

^aReaction conditions: H₂ pressure, 2.6 MPa; reaction temperature, 185 °C; molar ratio of H₂/DMO, 60/1; LHSV of 0.8h^–1, based on 15 wt % DMO methanol solution.

^bOther byproducts include ethanol, 1,2-butanediol, and extreme trace amounts of methyl formate and methyl acetate.

Figure presents the effects of operational parameters on DMO hydrogenation over Cu–Ni _x O _y /SiO₂ catalysts. As shown in Figure a, the conversions of DMO exhibit a steady increase with rising temperatures, approaching near-complete conversion at 180 °C. The selectivity toward EG approaches the highest value of 89.2% at 185 °C. With continuously increasing operational temperatures, a slight decline in EG selectivity is observed, accompanied by a slight increase in MG selectivity, suggesting increased side reactions at elevated temperatures. Figure b shows the influence of LHSV on the conversions of DMO and the selectivity toward EG/MG. Within the range of LHSV from 0.6 to 0.8 h^–1, DMO conversions and EG selectivity remain relatively stable. However, when the LHSV is above 0.85 h^–1, apparent decline in DMO conversion and EG selectivity is observed, indicating the shortened residence time leading to incomplete conversion and byproduct formation. All the above tests were conducted at a fixed H₂/DMO molar ratio of 60/1. Therefore, the operational temperature of 185 °C and LHSV of 0.8 h^–1 are optimal conditions for DMO hydrogenation over Cu–Ni _x O _y /SiO₂.

2.

Effects of a) reaction temperatures and b) LHSV on DMO hydrogenation over Cu–Ni _x O _y /SiO₂. Reaction conditions: H₂ pressure, 2.6 MPa; reaction temperature for b) 185 °C; molar ratio of H₂/DMO, 60/1; LHSV for a) 0.8 h^–1 (based on 15 wt % DMO methanol solution).

Notably, the above favorable catalytic performance is achieved at an H₂/DMO molar ratio of only 60/1, which is markedly lower than the ratios of 80/1 to 100/1 commonly adopted in most reported lab-scale Cu-based DMO hydrogenation systems. ,, From an industrial perspective, a lower hydrogen-to-ester ratio can effectively reduce the energy consumption of hydrogen circulation and recompression, endowing the catalyst with prominent economic advantages for large-scale deployment. However, it should be clarified that direct head-to-head performance comparison between this work and most lab-scale literature results is not sufficiently rigorous due to fundamental discrepancies in catalyst fabrication routes and evaluation scales. The majority of reported Cu-based catalysts were synthesized using elaborately designed nanostructured supports (e.g., ordered mesoporous silica, copper phyllosilicate) via complicated laboratory-specific protocols, and their catalytic tests were typically performed in microreactors with catalyst loadings of only 0.5–5 mL, , where internal/external diffusion and heat/mass transfer limitations are largely eliminated. In contrast, the catalyst in this work was prepared from industrial-grade fumed silica via a scalable one-step coprecipitation method fully compatible with industrial extrusion molding, and the performance evaluation was conducted in a bench-scale fixed-bed reactor with an inner diameter of 40 mm and 40 mL of catalyst loading. This bench-scale system better mimics industrial reactors in terms of flow hydrodynamics, bed heat management, and mass transfer behavior, rendering the obtained performance data more practically relevant for engineering application.

Figure b shows the XRD patterns of Cu–Ni _x O _y /SiO₂. Only diffractions from silica and Cu are obvious in Figure b while Ni-related diffractions are not observed, suggesting the formation of amorphous Ni _x O _y (discussed later). The average particle size of Cu NPs in Cu–Ni _x O _y /SiO₂ is calculated to be 27.3 nm using the Scherrer formula, while that of Cu/SiO₂–FA-RT is 30.5 nm, indicating that the introduction of nickel effectively inhibits the growth of copper particles. This mild geometric promotion effect moderately improves Cu dispersion, but the reduction in particle size is only ∼10%. Such a limited improvement in dispersion cannot fully account for the remarkable enhancement in catalytic performance, implying that electronic regulation at the Cu–Ni _x O _y interface should play a dominant role in boosting the hydrogenation activity. We further carried out X-ray photoelectron spectroscopy (XPS) measurements for Ni species of Cu–Ni _x O _y /SiO₂. As shown in Figure S3, the binding energies at 856.2 and 874.0 eV are assigned to 2p_3/2 and 2p_1/2 of Ni²⁺ species, , and the distinct satellite peaks at 861.8 and 879.5 eV are also observed. Moreover, no metallic Ni information was found in Figure S3, indicating that the Ni species is present as amorphous Ni _x O _y in Cu–Ni _x O _y /SiO₂. Indeed, literature has reported that H₂ reduction at >300 °C is generally required for the reduction of Ni _x O _y to metallic Ni, and H₂ reduction at 230 °C in this work is not enough for the reduction of Ni _x O _y .

Figure S4a shows TEM images of Cu–Ni _x O _y /SiO₂, where the contrast between the metals and silica is not obvious. Since TEM images cannot differentiate the Cu/Ni from silica, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) with energy-dispersive spectrometry (EDS) phase mapping was further employed to determine the dispersion of Ni _x O _y and Cu in Cu–Ni _x O _y /SiO₂. As shown in Figures and S4b, the metal particles are clearly visible in the HAADF-STEM model. Moreover, it is apparent in Figure that the highest concentration of Cu aligns with that of Ni, confirming the close contact of Cu and Ni species. With the combination of the studies of XRD (Figure b), XPS (Figure S3), and HAADF-STEM with EDS phase mapping (Figure ), it is concluded that the active particles of Cu–Ni _x O _y /SiO₂ are present in the form of Cu–Ni _x O _y heteroaggregate nanostructures. The literature has reported that such close contact of metal–metal oxide species may provide active interfaces for hydrogenation, and the metal–metal oxide interaction could modulate the electronic/geometric properties of active surfaces to facilitate hydrogenation reactions. −

3.

a) and d) HAADF-STEM images of Cu–Ni _x O _y /SiO₂ for EDS phase mappings; b) and e) the Cu phase mappings; and c) and f) the Ni phase mappings.

Figure presents the H₂-temperature-programmed reduction (H₂-TPR) profiles of Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂ catalysts. For Cu/SiO₂–FA-RT, three reduction peaks are resolved at 215, 250, and 288 °C. The low-temperature shoulder peak at 215 °C is assigned to the reduction of smaller CuO NPs. The dominant peak at 250 °C corresponds to the reduction of the majority of bulk CuO NPs, while the high-temperature shoulder peak at 288 °C arises from the reduction of CuO species strongly interacting with SiO₂ supports. , This multipeak feature reflects the heterogeneous reducibility of Cu species in the monometallic catalysts, originating from diverse coordination environments of Cu atoms. In contrast, Cu–Ni _x O _y /SiO₂ exhibits a peak centered at 258 °C with a shoulder around 215 °C. Literature has reported that the reduction of Ni oxides requires a relatively high temperature of ∼490 °C for silica-supported monometallic Ni catalysts. However, no additional peaks were found for Cu–Ni _x O _y /SiO₂, suggesting that Ni and Cu were reduced together at ∼258 °C. The Cu reduction generates an H₂ spillover effect to promote the Ni reduction, which has been found in other Ni–Cu/SiO₂ systems. It is mentioned here that the prepared catalysts for hydrogenation reactions were reduced by H₂/N₂ gas at 230 °C, and XPS studies of Cu–Ni _x O _y /SiO₂ reduced at 230 °C (Figure S3) show no presence of metallic Ni. Therefore, the Ni species of Cu–Ni _x O _y /SiO₂ reduced at 230 °C is present in the form of Ni²⁺, and Cu species are present in the form of Cu⁺ and Cu⁰ (as evidenced by XPS studies, discussed later), since the complete reduction of Cu species requires a temperature of 250 °C. Such a reduction process guarantees the formation of Cu–Ni _x O _y interaction and the presence of Cu⁺ species, which is crucial for enhanced DMO hydrogenation performance. −

4.

H₂-TPR profiles of Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂.

Figure a and b present the XPS studies of Cu 2p of Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂, respectively. For Cu/SiO₂–FA-RT (Figure a), the binding energies at 932.6 and 952.5 eV can be assigned to Cu 2p_3/2 and Cu 2p_1/2 of Cu⁰/Cu⁺ species, respectively, while those at 934.8 and 954.6 eV are ascribed to Cu²⁺ species. Moreover, the shakeup satellite peaks at ∼945.0 eV confirm the presence of Cu²⁺ species. As for Cu–Ni _x O _y /SiO₂ catalysts in Figure b, the binding energies at 932.3 and 952.2 eV could be assigned to 2p_3/2 and 2p_1/2 of Cu⁺/Cu⁰ species, while those at 934.5 and 954.3 eV are ascribed to 2p_3/2 and 2p_1/2 of Cu²⁺ species. Figure c and d display the Cu LMM Auger electron spectra for Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂, respectively. In Figure c, the Cu LMM spectrum for Cu/SiO₂–FA-RT exhibits two characteristic peaks at 916.1 and 912.9 eV (kinetic energies), indicative of Cu⁰ and Cu⁺ species, , respectively, while the corresponding Cu LMM spectrum of Cu⁰ and Cu⁺ species for Cu–Ni _x O _y /SiO₂ appears at 917.1 and 914.2 eV (Figure d).

5.

Cu 2p XPS spectra showing a) Cu/SiO₂–FA-RT and b) Cu–Ni _x O _y /SiO₂. Cu LMM Auger electron spectra showing c) Cu/SiO₂–FA-RT and d) Cu–Ni _x O _y /SiO₂.

Table summarizes the atomic percentages of Cu⁰, Cu⁺, and Cu²⁺ by XPS studies for Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂ catalysts. The percentage of Cu²⁺ for Cu–Ni _x O _y /SiO₂ is 15.7%, significantly smaller than that of Cu/SiO₂–FA-RT (28.7%). Moreover, the atomic percentage of Cu⁺ species in Cu–Ni _x O _y /SiO₂ is 48.8%, much more than that of Cu/SiO₂–FA-RT (26.2%). It is well established that Cu⁰ species are responsible for H₂ dissociation to generate active hydrogen species, while Cu⁺ sites function to adsorb and activate the carbonyl groups of DMO and stabilize acyl reaction intermediates. ,, The optimized Cu⁰/Cu⁺ molar ratio of 0.73/1, induced by the close-contact Cu–Ni _x O _y interfacial interaction, balances the hydrogen activation and substrate adsorption steps, thus promoting the deep hydrogenation of DMO to EG and suppressing the accumulation of the intermediate MG. Since the BET-specific surface area of Cu–Ni _x O _y /SiO₂ is only 231.8 m²/g (Figure S2 and Table S1), the better performance of Cu–Ni _x O _y /SiO₂ is not from its textural properties. Instead, the higher total fraction of active Cu⁰ and Cu⁺ species, together with the optimized Cu⁰/Cu⁺ ratio, accounts for the enhanced hydrogenation performance, and such modulation originates from the strong metal–oxide interaction at the Cu–Ni _x O _y interfaces. This precise valence regulation effect induced by low-content amorphous Ni _x O _y is distinct from the traditional Ni modification that mainly relies on improving Cu dispersion, which deepens the understanding of the promotion mechanism of Ni on Cu-based DMO hydrogenation catalysts. Notably, such a metal–metal oxide interfacial synergistic strategy has been widely verified in various hydrogenation reaction systems, proving its universality in tuning catalytic performance via electronic and geometric modulation. −

3. Cu XPS and LMM Auger Electron Spectra Analysis for Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂ .

	Kinetic energy (eV)		Binding energy (eV)	Contents of Cu species (%)			Molar ratios
Catalysts	Cu0	Cu+	Cu2+	Cu0	Cu+	Cu2+	Cu0/Cu+
Cu/SiO2–FA-RT	916.1	912.9	935.5	45.1	26.2	28.7	1.72/1
Cu–Ni x O y /SiO2	917.1	914.2	935.4	35.5	48.8	15.7	0.73/1

To further probe the surface properties, studies of diffuse reflectance Fourier transform infrared spectroscopy (DRIFT-IR) with CO probes were performed on the reduced Cu/SiO₂–FA-RT and Cu–Ni _x O _y /SiO₂ catalysts. As shown in Figure , both catalysts exhibit similar spectra, confirming that the Ni species in Cu–Ni _x O _y /SiO₂ are in the form of Ni oxides. Specifically, the bands centered at ∼2120 and ∼2160 cm^–1 are related to CO adsorption on Cu species, − where the former could be assigned to CO adsorption on Cu⁰ or defect sites, and the latter is due to CO adsorption on Cu⁺ species or isolated Cu atoms. ,−

6.

DRIFT-IR spectra with CO probes of a) Cu/SiO₂–FA-RT and b) Cu–Ni _x O _y /SiO₂. The spectra from bottom to top were first recorded in a CO atmosphere and then recorded after N₂ purging.

The catalytic stability of Cu–Ni _x O _y /SiO₂ for DMO hydrogenation is shown in Figure a, where the DMO conversions and EG selectivity maintain above ∼98 and 89%, respectively. Figure b shows the thermogravimetric/derivative thermogravimetric (TG/DTG) analysis of the catalysts after the stability test. No significant mass loss was found in Figure b, suggesting negligible coke deposition on Cu–Ni _x O _y /SiO₂ during the 48-h stability test. Carbon deposition is one of the dominant deactivation pathways for Cu-based DMO hydrogenation catalysts in industrial operation. The ultralow coke formation is tentatively attributed to the modulated adsorption state of oxygenated intermediates at the Cu–Ni _x O _y interfaces, which suppresses the deep decomposition and condensation of intermediates into carbonaceous deposits. This excellent anticoking ability is critical for prolonging catalyst service life in practical production. ICP-OES was further performed to evaluate the antileaching property of active metals: the Cu and Ni loadings of the spent catalysts are 19.8 and 1.82 wt %, respectively, nearly identical to those of the fresh catalyst (20.0 wt % Cu, 1.85 wt % Ni), confirming no significant leaching or loss of active components during the stability test. Moreover, no apparent changes were found in the XRD patterns (Figure S5) of Cu–Ni _x O _y /SiO₂ after the reaction, further verifying the structural stability of the catalyst.

7.

a) Stability evaluation of Cu–Ni _x O _y /SiO₂ catalysts for DMO hydrogenation; b) TG/DTG curves of Cu–Ni _x O _y /SiO₂ catalysts after the stability test. Reaction conditions: 2.6 MPa of H₂; 185 °C; H₂/DMO of 60/1; DMO LHSV of 0.8 h^–1, based on 15 wt % DMO methanol solution.

To further evaluate its industrial application potential, Cu–Ni _x O _y /SiO₂ catalyst was tested in a long-term side-line experiment on a skid-mounted facility. The side-line unit adopted a five-tube tubular reactor (ϕ44 × 2 × 6000 mm per tube), with 35 L of catalyst loaded on the tube side and circulating water for heat removal on the shell side. As shown in Figure S6, the catalyst ran continuously for 4000 h under industrial operating conditions. Throughout the test, DMO conversion remained nearly 100% and EG selectivity stabilized at approximately 98.5% without any performance attenuation. This performance is comparable to, and even exceeds, that of most commercially applied industrial DMO hydrogenation catalysts, providing solid evidence for the practical application value of the developed catalyst system. The superior performance on the side-line unit is ascribed to the reasonable height-to-diameter ratio and efficient heat transfer of the tubular reactor, as well as the adoption of neat DMO as feedstock. The integrated skid-mounted structure enables flexible on-site deployment for practical production.

4. Conclusions

In conclusion, this work developed silica-supported Cu–Ni _x O _y hybrid nanocatalysts for DMO hydrogenation to EG. At room temperature, the first addition of ammonia to silica colloids with subsequent coprecipitation of silica and metal precursors resulted in the formation of relatively small active particles, and the introduction of Ni to Cu further enhanced their catalytic performance. The developed one-step coprecipitation method using industrial fumed silica as raw material is facile and scalable for large-scale production, showing prominent engineering application value. The resultant Cu–Ni _x O _y /SiO₂ catalysts achieved 98.9% conversion and 89.2% EG selectivity with stable catalytic performance during the 48-h time-on-stream testing, surpassing the performance of their monometallic counterparts. Only 0.07 wt % coke deposition is detected after long-term operation, which brings distinct economic and stability advantages for industrial application. By characterization, the enhancement can be assigned to the suitable Cu⁰/Cu⁺ molar ratio and the high percentage of the total Cu⁰ and Cu⁺ species induced by Ni _x O _y –Cu interaction. We believe that the synthetic method developed can be extended to other supported bimetallic systems for more applications in heterogeneous reactions.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c05343.

• Further chemicals and material characterizations; calculations for DMO conversion (X), EG selectivity (S_EG), and methyl glycolate selectivity (S_MG); GC analysis methods; TEM images; N₂ adsorption–desorption isotherms and the corresponding pore size distributions; the textural properties of silica-supported Cu-based catalysts prepared with different synthetic parameters; XPS 2p spectrum of Ni species for Cu–Ni _x O _y /SiO₂; and XRD patterns of Cu–Ni _x O _y /SiO₂ after catalytic reactions (PDF)

The authors declare no competing financial interest.