Owing to its environmental and sustainable economic advantages, ethanol has received much attention as a promising alternative to fossil fuels. In particular, ethanol is the second-most used fuel oxygenate for preventing groundwater pollution caused by methyl tert-butyl ether [1-3]. Although the ethanol-from-corn industry has traditionally played an important role in ethanol production, it has many issues, such as the dilemma of whether to use corn crops for food or fuel [2]. Therefore, finding new methods to produce ethanol is of great importance. As the hydrogenation of methyl acetate (MA) to ethanol is highly active and selective, it is a promising alternative method for ethanol production [4]. Traditionally, MA is mainly produced from coal and natural gas [5]. Increasing interest in shale gas and considerable coal reserves makes the hydrogenation of methyl acetate a promising ethanol production process.

Copper-based chromium oxide-supported catalysts are traditional catalysts for ester hydrogenation. However, the potential harm caused by Cr to health and the environment limits its application [6, 7]. Therefore, the exploration of chromium-free catalysts has attracted much attention. Much research has focused on SiO₂-supported Cu catalysts, mainly due to their safety and low cost, which meet the needs of industrial catalysts [8]. However, problems such as short lifetime and poor stability will severely restrict the industrial application of Cu/SiO₂ catalysts. There has been much research into improving the stability and performance of Cu/SiO₂ catalysts, mainly focused on support modifications, different catalyst synthesis methods, and doping additives. Qin et al. [9] synthesized Cu/MgO-SiO₂ containing different MgO/SiO₂ mass ratios and found that catalysts with a MgO/SiO₂ mass ratio of 9 showed the best catalytic performance for MA hydrogenation. Yin et al. [10] studied Cu/SiO₂, Cu/TiO₂, and Cu/SiO₂-TiO₂ catalysts in the hydrogenation of dimethyl oxalate, and found that Cu/SiO₂-TiO₂ catalyst displayed high catalytic activity and long stability, which could be attributed to highly dispersed copper species and intimate interaction between copper species and the support. Dong et al. [11] reported a comparative study of different silica-supported copper catalyst preparation methods and found that Cu/SiO₂ catalysts synthesized using an ammonia-evaporation method exhibited the best activity in MA hydrogenation. As for doping additives, introducing noble metal additives, such as Au [12] and Pt [13], to synthesize bimetallic catalysts has been widely used to promote catalytic performance. However, the rising costs of noble metals limit further industrial use. Some noble-metal-free Cu/SiO₂ catalysts have been reported. He et al. [14] introduced B into Cu/SiO₂ catalysts, which showed greatly enhanced catalytic performance with a Cu/B atomic ratio of 6.6. Furthermore, Yin et al. [15] added an appropriate amount of Ni to increase the activity of Cu/SiO₂ catalysts, while Zheng et al. [16] synthesized lanthanum oxide-modified Cu/SiO₂ catalysts that showed better catalytic performance than Cu/SiO₂. Therefore, modifying Cu/SiO₂ with noble-metal-free additives is an efficient and low-cost approach to improving catalyst life and performance.

In the hydrogenation of biomass and steam reforming of methanol to produce H₂, the addition of indium has been shown to significantly enhance the catalytic performance of copper-based catalysts and stabilize copper surface activity [17, 18]. Onyestyák et al. [19] also studied zeolite-supported Cu and Cu₂ In nanoparticles for the hydrogenation of carboxylic acids to alcohols, finding that the conversion and selectivity of the catalysts was increased by In₂O₃ doping. When doped with In₂O₃, Cu-based catalysts exhibit excellent performance in some hydrogenation reactions. However, copper-based catalysts modified with In₂O₃ have not been reported in MA hydrogenation. Furthermore, changes to the structure and chemical state of copper species after In₂O₃ doping must be investigated.

Herein, xIn-Cu/SiO₂ catalysts doped with different indium oxide loadings were synthesized using the ammonia-evaporation (AE) method. The catalytic activity and stability were evaluated in the vapor-phase hydrogenation of MA. Furthermore, a series of catalyst characterization techniques were used to investigate the promotional influence of In₂O₃ on the Cu/SiO₂ catalyst.

xIn-Cu/SiO₂ catalysts with a Cu loading of 30 wt% were synthesized using the AE method, where x represents the In₂O₃ mass loading.

In(NO₃)₃·xH₂O was purchased from Aladdin, China. Other reagents used were purchased from Tianjin Yuanli Chemical Institute, China. All reagents were of analytical grade and used directly without purification.

Firstly, certain amounts of In(NO₃)₃·xH₂O and Cu(NO₃)₂·3H₂O were dissolved in distilled water in a beaker. Aqueous ammonia (25 wt%) was added dropwise until the mixture reached an initial pH of around 11, and then an appropriate amount of silica sol with 25% mass content was added. All operations above were carried out at room temperature. The beaker was immersed in a water bath at 303 K and stirred for 4 h. The water bath temperature was then increased to 363 K to evaporate the ammonia. During ammonia evaporation, the pH of the mixture gradually decreased. Evaporation was terminated when the mixture pH decreased to 6-7. The resulting suspension was dried for 12 h at 393 K and the obtained powder was calcinated for 4 h at 723 K.

For comparison, 1In₂O₃/SiO₂ samples, containing an In₂O₃ mass loading of 1%, and Cu/SiO₂ catalyst with a Cu mass loading of 30% were also prepared using the AE method.

N₂ physisorption was performed using a Quantachrome autosorbiQ₂ device. After degassing the sample for 3 h under vacuum at 573 K, N₂ adsorption-desorption isotherms were measured at 77 K. Specific surface areas (A_BET) were calculated using the Brunauer-Emmett-Teller (BET) method. The total pore volume (V_pore) and pore size (D_pore) derived from the desorption branches of the N₂ isotherm distribution were estimated using the Barrett-Joyner-Halenda (BJH) method.

Copper dispersion is a vital index related to catalytic performance. N₂O chemisorption was performed on a Quantachrome Instrument AMI-90 apparatus to determine the copper dispersion according to a literature procedure [20].

The X-ray powder diffraction (XRD) profiles were collected on a D/MAX-2500 diffractometer with a Cu K_α radiation source under measuring conditions of 40 kV and 40 mA. The catalysts were reduced under pure hydrogen (250 mL/min) at 623 K for 4 h, and then the gas was switched to pure nitrogen under the same flow rate. After cooling to room temperature, the catalysts were collected and stored under a nitrogen atmosphere before conducting XRD measurements. The scanning angle was from 20° to 90° with a scanning rate of 5°/min. The Cu crystallite size was obtained using the Scherrer equation.

Fourier transform infrared spectroscopy (FT-IR) patterns of xIn-Cu/SiO₂ catalysts were obtained using a Nicolet iZ10 spectrometer to detect characteristics vibration bonds. Catalysts were mixed with KBr using an agate mortar and compressed into pellets for FT-IR measurement in a wavenumber range of 2200-400 cm^-1.

Inductively coupled plasma optical emission spectroscopy (ICP-OES), performed using an Agilent 7700ce spectrometer, was used to determine the copper and indium loadings in all catalysts. All samples were dissolved in a mixture of H₃BO₃, HF, and HNO₃, and diluted with distilled water before testing.

H₂-temperature programmed reduction (H₂-TPR) was conducted on a Quantachrome Instrument AMI-90. Prior to analysis, xIn-Cu/SiO₂ catalysts were pretreated under Ar (30 mL/min) at 373 K for 30 min. The catalysts were then reduced under 10 vol% H₂/Ar (30 mL/min) from 373 to 1073 K with a heating rate of 10 K/min.

Transmission electron microscopy (TEM) images were obtained using a JEM-2100F JEOL microscope with a voltage of 200 kV.

X-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (XAES) were obtained using a Thermo ESCALAB 250XI with monochromatic Al K_α (hv = 1486.6 eV) and the binding energies were calibrated to the C 1s signal (284.8 eV). Prior to XPS measurements, the samples were reduced with pure hydrogen (250 mL/min) at 623 K for 4 h and then stored under a nitrogen atmosphere.

Catalytic activities were evaluated using a tubular reactor with a 10-mm internal diameter. A certain amount of catalyst (40-60 mesh) was loaded into a fixed place in the reactor with a thermocouple to control the temperature. Before each test, the catalyst was reduced under pure hydrogen (250 mL/min) at 623 K for 4 h. After cooling to 523 K, MA was added into the preheating chamber using a constant flow pump as a mixture with pure hydrogen (hydrogen/MA molar ratio of 15). The mixture entered the reactor with pressure and temperature controlled at 3 MPa and 523 K, respectively. The liquid hourly space velocity (LHSV) was from 1 to 6 h^-1, calculated based on the weights of catalysts and MA injection per hour. The products were collected and analyzed by gas chromatography (GC; SP-6890) equipped with a flame ionization detector (FID) and a PEG-20M capillary column.

Conversion and selectivity of the catalysts were calculated using the following equations:

(1)

(2)

The activities of all catalysts were examined in a tubular reactor under the same reaction conditions. As listed in Table 1, the conversion and selectivity of the catalysts increased and then decreased with increasing indium oxide loading. With the addition of 0.5-2 wt% indium oxide, the catalysts exhibited better catalytic performance than Cu/SiO₂. With an indium oxide mass content of 1%, 1In-Cu/SiO₂ catalyst exhibited the best MA conversion and ethanol selectivity. Notably, 1In/SiO₂ showed no activity in MA hydrogenation. These results indicated that doping the Cu/SiO₂ catalyst with an appropriate content of indium oxide significantly improved its catalytic activity.

To investigate the influence of liquid hour space velocity (LHSV) on the catalytic activities of Cu/SiO₂ and 1In-Cu/SiO₂, changes in their conversion and selectivity with LHSV were evaluated, as shown in Fig. 1. LHSV was found to have a great influence on catalyst performance. When the LHSV was 1-3 h^-1, both catalysts displayed relatively high conversion and selectivity, and the difference in selectivity between the two catalysts remained almost the same. However, the difference in conversion between 1In-Cu/SiO₂ and Cu/SiO₂ became wider with increasing LHSV. When the LHSV reached 6 h^-1, the MA conversion of Cu/SiO₂ was only 32.3%, while that of 1In-Cu/SiO₂ was 57.6%. Compared with Cu/SiO₂, 1In-Cu/SiO₂ exhibited higher catalytic activity under higher LHSV conditions, which implied that 1In-Cu/SiO₂ catalyst possessed good tolerance to LHSV.

The textural properties of calcined xIn-Cu/SiO₂ catalysts are summarized in Table 2. The BET surface area, pore volume, and average pore diameter of the catalysts did not change obviously when the indium oxide content was changed. Therefore, we deduced that the surface area and pore structure of the catalysts were not the main factors affecting catalytic activity.

The copper dispersion was determined using an N₂O chemisorption method (Table 2). With the addition of indium oxide, the copper dispersion was significantly improved, with 1In-Cu/SiO₂ presenting the highest copper dispersion (26.0%). The results indicated that doping with an appropriate amount of indium oxide significantly improved the copper dispersion. Combined with activity results, the improvement in copper dispersion might be related to high catalytic performance. Furthermore, it was suggested that the degree of copper dispersion was important to the performance of the Cu-based catalysts [21]. Therefore, the higher copper species dispersion of 1In-Cu/SiO₂ contributed to its superior catalytic performance compared with other catalysts.

The calcined and reduced catalysts were characterized by XRD. Fig. 2(A) shows XRD patterns of all catalysts after calcination. The broad peak at 2θ of 22° was attributed to amorphous silica [22, 23], while two diffraction peaks at 2θ of 35.6° and 38.7° were assigned to CuO species (JCPDS 05-0661). The diffraction peaks for CuO first became weak and then strong with increasing indium oxide loading. When the mass loading of indium oxide was 1%, the diffraction peaks for CuO were the weakest, attributed to the smaller size and better CuO dispersion.

As showed in Fig. 2(B), the diffraction peaks at 2θ of 43.3°, 50.4°, and 74.1° were assigned to the (111), (200), and (220) lattice planes of Cu⁰, while the characteristic peak positioned at 2θ of 36.4° attributed to the Cu₂O (111) plane was observed in all catalysts. The intensity of the characteristic peaks of Cu⁰ became weaker with the addition of indium oxide, and was weakest when the indium oxide loading was 1 wt%. The peak intensity might be related to active phase dispersion in addition to the crystallite size [14]. The Cu⁰ crystallite size (Table 1) initially decreased and then increased with increasing indium oxide loading. The 1In-Cu/SiO₂ catalyst exhibited diffraction peaks with the weakest intensity and the smallest crystallite size (16.3 nm). The weak characteristic peaks and small crystallite size implied that 1In-Cu/SiO₂ had the better active species dispersion among all catalysts, which was consistent with the copper dispersion obtained by N₂O chemisorption characterization. Furthermore, all reduced catalysts had a weak and broad diffraction peak assigned to Cu₂O, indicating that both Cu⁰ and Cu₂O were present in all reduced samples.

Fig. 3 shows TEM and HRTEM images of reduced Cu/SiO₂ and 1In-Cu/SiO₂ catalysts. The light gray region and dark gray region were assigned to silica support particles and deposited-copper-containing materials, respectively. Compared with Cu/SiO₂ (Fig. 3(A)), 1In-Cu/SiO₂ (Fig. 3(B)) exhibited better copper species dispersion with smaller-scale dark areas, indicating that 1In-Cu/SiO₂ had a smaller crystallite size and better dispersion than Cu/SiO₂. Meanwhile, in the HRTEM images of both catalysts, (111), (200), and (220) lattice planes of Cu⁰ and (111) lattice planes of Cu₂O were found, which was consistent with the XRD results. Furthermore, no lattice plane assigned to indium or indium oxides was found, which might be attributed to the high dispersion of indium species.

To further investigate the structure properties, FT-IR spectra were recorded. The FI-IR spectra of SiO₂ and calcined xIn-Cu/SiO₂ are shown in Fig. 4. The band at 1637 cm^-1 was assigned to the bending mode of OH groups of absorbed water [24]. Absorption bands around 472, 1119, and 800 cm^-1, corresponding to different vibration modes of Si-O bonds, were assigned to amorphous silica [25]. The band at 800 cm^-1 corresponded to Si-O-Si symmetric stretching vibration [24, 25]. Furthermore, the band at 672 cm^-1 was assigned to the δ_OH band in copper phyllosilicate [25]. As shown in Fig. 4, calcined 1In-Cu/SiO₂ showed the strongest intensity of peaks corresponding to copper phyllosilicate. Notably, no obvious peaks at 601, 565, or 557 cm^-1, assigned to the presence of In-O phonon vibration, were observed [26]. This might be due to the low indium oxide content.

The relative content of copper phyllosilicate was evaluated by I₆₇₀/I₈₀₀, which represents the integrated intensity of the δ_OH band at 670 cm^-1 normalized to that of the ν_SiO band at 800 cm^-1 [25]. As shown in Table 2, calcined 1In-Cu/SiO₂ had the highest relative content (0.26) of copper phyllosilicate. This indicated that introducing indium species into Cu/SiO₂ could promote the formation of copper phyllosilicate. The presence of copper phyllosilicate and its plate-like structure could limit the diffusion of copper and restrict the probability for secondary nucleation during reduction under an H₂ atmosphere [27, 28]. Therefore, the presence of copper phyllosilicate in calcined catalysts might contribute to the smaller size of copper species and better copper dispersion after reduction, in good agreement with the XRD, TEM and N₂O chemisorption results. Furthermore, the small and well-dispersed copper species were considered the primary reason for the high activity and selectivity of the ester hydrogenation [29, 30], in accordance with the activity results.

To compare the reduction performance of the catalysts, the calcined xIn-Cu/SiO₂ samples were evaluated by H₂-TPR. As shown in Fig. 5, all samples exhibited two reduction peaks, which were attributed to the high copper species content (30 wt%) in all catalysts [25]. All samples displayed one peak, denoted as the α-peak, at around 528 K, thought to correspond with the reduction of well-dispersed copper species, and another small peak, denoted as the β-peak, at around 555 K, corresponding to the reduction of agglomerated bulk CuO crystallites [11, 20, 31]. More precisely, the α-peak was assigned to the collective contribution of the reduction of copper phyllosilicate and ion-exchanged Cu-O-Si species (a kind of cuprous species that strongly interacts with the silica support using the ammonia-evaporation method) to Cu⁺ and the reduction of well-dispersed CuO to Cu⁰ [32, 33]. Therefore, the strong α-peak and small β-peak observed for all xIn-Cu/SiO₂ samples compared with Cu/SiO₂ might partly result from the relatively high content of copper phyllosilicate and well-dispersed CuO, in accordance with the FT-IR and N₂O chemisorption results. Combined with the results shown in Table 2, it was reasonable to conclude that adding the appropriate content of indium species resulted in a relatively high content of copper phyllosilicate, well-dispersed copper species, and a small CuO crystallite size.

Surface characterization of the synthesized catalysts was also investigated by conducting XPS with X-ray induced Auger electron spectroscopy (XAES). For the In 3d XPS patterns (Fig. 6), the spectra of all samples contained In 3d peaks at approx. 445 eV (3d_5/2) and approx. 452.5 eV (3d_3/2), and the energy differences of all catalysts agreed well with the standard value of 7.5 eV, indicating that the indium species in all doped catalysts was In₂O₃ [17, 34, 35]. In the Cu 2p XPS patterns of calcined Cu/SiO₂ and 1In-Cu/SiO₂ (Fig. 7), the peak at around 935.6 eV in the two calcined catalysts was assigned to Cu²⁺. Furthermore, the 2p→3d satellite peak at 940-945 eV was also characteristic of Cu²⁺ species in Cu/SiO₂ and 1In-Cu/SiO₂, while no peaks ascribed to Cu⁰ or Cu⁺ were observed [20]. These results indicated that, in calcined catalysts, copper existed in the Cu²⁺ oxidation state. After reduction, only two peaks at around 932.7 and 925.6 eV were observed (Fig. 8), assigned as Cu 2p_3/2 and Cu 2p_1/2 peaks, respectively. However, the satellite peak and Cu 2p binding energy at 935.6 eV attributed to Cu²⁺ were not found, indicating the reduction of Cu²⁺ to Cu⁰ and/or Cu⁺ [14, 36].

The difference between the binding energies of Cu⁺ and Cu⁰ was negligible, and Cu LMM spectrum (Fig. 9) were recorded to determine the amount of Cu⁺ and Cu⁰ [37]. After deconvolution of the Cu LMM spectra, two obvious overlapping peaks were observed at around 915 and 918 eV, which were assigned to Cu⁺ and Cu⁰, respectively. The deconvolution results are shown in Table 3, the Cu⁺ and Cu⁰ contents were affected by the amount of In₂O₃. The Cu⁺/(Cu⁰+Cu⁺) ratio reached a maximum of 57.4% when the indium oxide content was 1 wt% and decreased when the indium oxide content of was 2 wt%. Cu⁺ existed after reduction mainly because the reductions of ion-exchanged Cu-O-Si species and copper phyllosilicate ceased at Cu⁺ at the reduction temperature used (623 K), with temperatures above 873 K required to further reduce Cu⁺ to Cu⁰ [38]. The XPS result showing that 1In-Cu/SiO₂ had the highest content of surface Cu⁺ was consistent with the FT-IR and H₂-TPR results.

As listed in Table 3, X_Cu⁺ increased with the addition of indium oxide, reached a maximum of 57.4% when the indium oxide content was 1 wt%, and then decreased with further indium addition. This result indicated that an appropriate indium oxide content could enrich the surface Cu⁺ species content and enhance the activity of the catalysts in MA hydrogenation. The Cu⁰ species were thought to adsorb hydrogen molecules, while the Cu⁺ species stabilized the methoxy and acyl species [39]. In ester hydrogenation, the synergetic effect of Cu⁰ and Cu⁺ could obtain good catalytic performance [14]. The high Cu dispersion and abundant surface Cu⁺ species might account for the excellent performance of 1In-Cu/SiO₂.

The stability test is a critical index for the practical application of catalysts in MA hydrogenation. We conducted long-term tests for Cu/SiO₂ and 1In-Cu/SiO₂ catalysts, and the results are shown in Fig. 10. For the first 50 h, Cu/SiO₂ and 1In-Cu/SiO₂ both displayed relatively stable catalytic activity. However, with increasing time on stream, the MA conversion and ethanol selectivity of Cu/SiO₂ decreased to around 64% and 73% after 100 h reaction, respectively. In contrast, 1In-Cu/SiO₂ maintained a high level of activity, with a conversion and selectivity of around 98%. Therefore, adding an appropriate amount of indium oxide could stabilize the activity of Cu/SiO₂ catalyst during the reaction.

After stability tests, XRD, XPS, ICP-OES, and N₂ physisorption analyses were performed to determine structure changes in the catalysts after long-term testing. As shown in Fig. 11 and Table 4, for Cu/SiO₂, the intensities of characteristic diffraction peaks assigned to Cu⁰ and Cu₂O became significantly stronger after the stability test, and the copper size of Cu/SiO₂ increased from 21.7 to 24.9 nm. However, for 1In-Cu/SiO₂, there was no obvious intensification of the characteristic peaks assigned to Cu⁰ and Cu₂O and the copper size only slightly increased by 0.6 nm. These results indicated that the active species of 1In-Cu/SiO₂ was stable, while that of Cu/SiO₂ sintered into larger particle after 100 h of testing. The sintering of active copper species has been reported to lead to Cu/SiO₂ deactivation [40].

As shown in Table 4, the copper loading of Cu/SiO₂ decreased from 28.2% to 27.8% (approx. 1.40% loss) and that of 1In-Cu/SiO₂ decreased from 27.9% to 27.8% (approx. 0.36% copper loss) after the 100-h stability test, meaning that the change in copper content in 1In-Cu/SiO₂ was insignificant, while that in Cu/SiO₂ was not negligible. As shown in Table 4, the Cu/Si mole ratio of Cu/SiO₂ was 0.21 before the reaction test, and had decreased to 0.13 after reacting for 100 h. This significant reduction was attributed to the sintering of copper species together with copper mass loss. In contrast to Cu/SiO₂, the Cu/Si molar ratio of 1In-Cu/SiO₂ slightly decreased from 0.23 to 0.22 after the stability test, which was an insignificant reduction attributed to the dispersion of copper species and the mass content of copper species remaining almost the same. The results above were consistent with XRD and ICP-OES results.

The X_Cu⁺ of Cu/SiO₂ declined by 4.8%, from 47.7% to 42.9%, while that of 1In-Cu/SiO₂ decreased by 1.6%, from 57.4% to 55.8%, after the 100-h stability test, as shown in Fig. 12 and Table 4. The negligible decrease of X_cu⁺ in 1In-Cu/SiO₂ indicated the steady valence of surface active species. Combined with the results of the stability test, conversion of copper chemical state could be a reason for the loss in activity of Cu/SiO₂. Zheng et al. [16] similarly observed that destroying the balance between surface Cu⁰ and Cu⁺ active sites led to the poor catalytic performance of Cu/SiO₂.

Moreover, after a 100-h stability test, the A_BET of Cu/SiO₂ had decreased by 134 m²/g (from 357 to 223 m²/g), while that of 1In-Cu/SiO₂ had dropped by only 8 m²/g (from 346 to 338 m²/g; Table 4). Therefore, the structure of 1In-Cu/SiO₂ was more stable than that of Cu/SiO₂. The distinct decrease in specific surface area of Cu/SiO₂ was mainly attributed to copper sintering and the aggregation of copper crystallite size.

The results above indicated that the deactivation of Cu/SiO₂ mainly resulted from the agglomeration of active copper particles, but also resulted from the valence change of surface copper species, which could lead to the loss of active sites and a decrease in the surface Cu⁺ content. However, doping an appropriate amount of indium into Cu/SiO₂ could significantly enhance copper dispersion and inhibit sintering and aggregation of active copper species to stabilize the chemical state of copper. The enhancement in catalyst stability could be explained as follows. At a high indium loading, indium dopant can alloy with copper particles to exist mainly as Cu₂ In and other types of CuIn alloy, such as Cu₄ In and Cu₉In₄ [18]. Furthermore, the low indium oxide content could also stabilize copper species, mainly due to the presence of indium oxide near the copper species [17]. As shown in the XRD, FT-IR, and XPS results (Table 2 and 3), indium doping significantly affected the physical and chemical states of copper species. Moreover, the thermal and structural stability of indium oxide was excellent below 773 K, making it much better than copper [41]. Therefore, In₂O₃ could act a kind of stable isolating agent to inhibit the thermal transmigration and accumulation of copper nanoparticles during the reaction. In conclusion, the excellent thermal and structural stability of indium oxide and its effect on copper species resulted in 1In-Cu/SiO₂ being stable during the stability test.

A series of xIn-Cu/SiO₂ catalysts were found to be efficient for methyl acetate hydrogenation. The best indium oxide content was approximately 1 wt%. As demonstrated by TEM and XRD measurements, copper dispersion was improved and copper particle size decreased when the appropriate amount of indium oxide was doped into the catalyst. Furthermore, FT-IR and N₂O chemisorption showed an increased content of copper phyllosilicate and better dispersion of copper species, which was beneficial to improving the performance of the Cu/SiO₂ catalyst. Moreover, XPS and XAES results showed synergetic catalysis between surface Cu⁰ and Cu⁺ active sites might contribute to the improved performance of the indium-modified Cu/SiO₂ catalyst. After a 100-h stability test, N₂ physisorption, XRD, XPS, and ICP-OES analyses were performed and showed that deactivation of the Cu/SiO₂ catalyst was mainly attributed to the agglomeration of active copper particle sizes and also attributed to destruction of the appropriate surface Cu⁰ and Cu⁺ distribution. Copper loss also contributed to the lower activity of Cu/SiO₂. However, 1In-Cu/SiO₂ remained almost the same before and after the stability test, implying that a suitable amount of indium oxide could stabilize the Cu/SiO₂ catalyst and prolong the catalyst lifespan. In₂O₃ might act as a kind of stable isolating agent to inhibit the thermal transmigration and accumulation of copper nanoparticles.