Ion-Exchange Synthesis of Surrounded CoNi@Al₂O₃ Catalyst for Levulinic Acid Hydrogenation to γ-Valerolactone under Mild Conditions

Abstract

The use of eco-friendly biomass as a resource is an efficient way to address the problems of fossil fuel depletion and climate change. In biomass conversion, versatile γ-valerolactone (GVL) is generally obtained from levulinic acid (LA) hydrogenation via a multimetallic catalyst system. Despite conversion efficiency being enhanced in mild conditions due to metal interactions, maintaining high catalyst stability is still a challenge. In this study, we synthesized a surrounded Co_0.52Ni_0.48@Al₂O₃-IE catalyst that exhibited excellent alloying and synergistic interaction between the metal constituents. Under relatively mild reaction conditions, the GVL yield over the catalyst exceeded 99% in LA hydrogenation. The catalyst showed no deactivation in a test of five cycles, displaying superiority in stability, possibly due to reasons of the physical isolation of the shell and the alumina retention on the Co-Ni alloys surface caused by the reversibility of exchange equilibrium. The present work demonstrated that a surrounded structured catalyst fabricated by ion exchange (IE) with active metals physically enclosed can lead to high catalytic activity and superior stability.

1. Introduction

In light of the depletion of fossil fuels and escalating concerns about global warming, the scientific communities have redoubled their efforts to explore sustainable and renewable resources. Biomass, characterized by its renewability, environmental friendliness, and sustainability, has emerged as a promising solution [1,2]. However, an uneconomic aspect lies in the fact that a substantial portion of biomass undergoes natural decomposition by microorganisms, resulting in the release of carbon dioxide, water, and heat, and this natural breakdown process fails to fully harness the energy and organic chemicals inherent in biomass. Therefore, further exploration and optimization of strategies for comprehensive biomass utilization are indispensable. LA is recognized as one of the “Top Ten” biomass-derived platform molecules by the U.S. Department of Energy due to its tremendous commercial attributes, which can be obtained through acid-catalyzed hydrolysis of cellulose [3,4,5]. The selective hydrogenation of LA or its esters holds the potential for producing a variety of high-value derivatives, including GVL, 1,4-butanediol (PDO), 2-methyltetrahydrofuran (MTHF), valeric acid, etc. [6]. Among these biomass-derived compounds, GVL has been found in widespread applications, such as food components, fuel additives, green solvent, advanced renewable fuel, pharmaceutical intermediates, and so on [7,8].

In the realm of catalytic processes from LA to GVL, both noble metals (e.g., Ru, Rh, Pt, Pd, and Au) [9,10,11,12,13,14], and non-precious metals (e.g., Ni, Cu, an Co) are used in the hydrogenation of LA to GVL [5,7,15,16,17,18,19,20]. Noble metal catalysts generally operate under mild reaction conditions and demonstrate notable catalytic efficiency. Nevertheless, it is true that their high costs pose challenges for large-scale industrial applications [21]. Moreover, the extensively employed Ru/C catalyst is plagued by relatively weak metal–support interactions, resulting in the facile loss of active components and consequently diminishing catalyst stability [11,14,18]. Non-precious metals have received widespread attention due to their cheapness and easy availability. Nevertheless, non-noble metals also manifest typical drawbacks, such as suboptimal catalytic activity and stringent reaction conditions, limiting their developments [22]. Over the past several decades, there has been a large amount of studies dedicated to the augmentation of catalytic efficiency in non-noble metals through the synergistic effects of multimetallic systems. These endeavors aim to harness the combined impact of various metal species to modify electronic configurations and improve adsorption properties, thereby delving into the intricate interplay of factors that influence catalytic performance. This multidimensional approach seeks to unravel the intricate mechanisms underlying the cooperative behavior of multiple metals in the catalytic system, offering insights that contribute to the refinement and optimization of non-noble metal catalysts for diverse applications [23,24,25,26,27,28]. For example, Huang et al. reported the effectiveness of the bimetallic Fe-Re catalyst supported by TiO₂, which contained metallic Re nanoparticles coated by FeO_x species and small quantities of Fe-Re alloys, giving a GVL yield of 95% at 180 °C under 40 bar of H₂ in water [6]. Tang et al. successfully synthesized the bimetallic Ni₁-Zn₁@OMC catalyst, achieving a 93% GVL yield at 180 °C within a 90 min reaction time. The incorporation of Zn metals led to the formation of Ni-Zn alloys, playing a pivotal role in activating the carbonyl group in LA. Simultaneously, the active sites of Ni^δ–Zn^δ+ facilitated the weakening and cleavage of C–O bonds in 4-hydroxyvaleric acid (4-HA) [29]. Wang et al. designed a bimetallic Ni₃Fe NPs@C catalyst by a simple one-step carbothermal reduction method, which was capable of showing high activity for selective conversion of LA to GVL by both direct and transfer hydrogenation mechanisms. Staggeringly, the bimetallic Ni₃Fe NPs@C catalyst was 6 and 40 times more reactive than both monometallic Ni NPs@C and Fe NPs@C in direct hydrogenation, respectively [30]. Notwithstanding the considerable body of researches on bimetallic catalysts, normally, elevating temperatures and pressures are necessary for achieving the optimal reaction conditions of the hydrogenation of levulinic acid (LA) to gamma-valerolactone (GVL). As such, the quest for bimetallic catalysts capable of facilitating LA hydrogenation under mild conditions, coupled with outstanding cyclic stability, still represents a formidable challenge [2,21,31,32,33].

In this study, a novel bimetallic catalyst for the hydrogenation of LA to GVL was developed by ion exchange directly. Notably, the ion exchanging reactions are more inclined to proceed in the direction of a lower dissolution equilibrium constant (A(OH)_x + B^y+ $\leftrightharpoons$ B(OH)_y + A^x+_, Ksp_B(OH)x < Ksp_A(OH)y). CoNi-MMH (mixed-metal hydroxide) precursors were synthesized according to methodologies outlined in the prior literature, exhibiting a well-defined nanosheet structure (Scheme 1) [34,35,36]. Consequently, the bimetallic hydroxide served as the primary precursor, and then, the Co²⁺ and Ni²⁺ were partially exchanged by Al³⁺ due to the solubility constants. Lastly, alumina-encapsulated Co-Ni alloys could be obtained after H₂ reduction (Scheme 1). The catalyst exhibited a high level of alloying effect and the synergistic interaction between metals, resulting in improved LA hydrogenation activity. The alumina encapsulation and the residual traces of alumina after ion exchange (IE) on the alloy surface increased the stability of the catalytic system. Our findings served to underscore the high efficacy and stability of the bimetallic catalysts synthesized through the ion-exchange method in the context of LA hydrogenation for GVL production.

2. Results and Discussion

Previous studies by Guo and coworkers [37] demonstrated that the spontaneous replacement of Ni²⁺ by Al³⁺ can be facilitated by utilizing the discrepancies in solubility constants. In this work, Co²⁺ and Ni²⁺ were directly exchanged by Al³⁺ because they have similar exchange rates due to the comparable solubility constants (Ksp_Al(OH)3 = 1.3 × 10⁻³³ < Ksp_Ni(OH)2 = 2 × 10⁻¹⁵ ≈ Ksp_Co(OH)2 = 1.6 × 10⁻¹⁵) of Ni(OH)₂ and Co(OH)₂, resulting in the formation of Co_xNi_y(OH)₂@Al(OH)₃ which exhibited diffraction peaks of hydrotalcite-like structure (Figure 1a) [35,38]. Following a controlled calcination process, the XRD patterns of the sample revealed distinctive peaks at 2θ of 31.4°, 36.9°, 45.0°, 59.5°, and 65.3°. These discernible peaks can be confidently attributed to Co₃O₄ (PDF#43-1003). Additionally, other notable peaks detected at 2θ of 37.3°, 43.4°, and 62.9° can be ascribed to NiO (PDF#44-1159) [3,39,40]. Compared to the catalyst prepared by the impregnation method [37], the diffraction peaks of Co₃O₄ and NiO shifted to higher angles, plausibly due to the entry of Al³⁺ into the Co₃O₄ and NiO lattice (Figure 1b). After H₂ reduction, the oxides were transformed into the corresponding metal components. Due to the similarity in structural features and closeness in peak position, it is difficult to distinguish the Co, Ni, and alloy phases by XRD (Figure 1c,d). Within a narrow range, the Co-Ni alloys peaks were located between the metal peaks of Co and Ni, and negatively shifted with increasing Co/Ni ratio, indicating enhanced formation of Co-Ni alloys. The shift is primarily attributed to the fact that Ni and Co atoms are alike in crystal structure, atomic radius, and valence state [41,42]. Furthermore, the peak intensity increased with the rise i Ni content, signifying improved crystallinity and enlarged particle size.

The EDX maps of Co_0.52Ni_0.48@Al₂O₃-IE showed that the cobalt and nickel signals highly overlap at localized positions, while those of aluminum are highly dispersed, suggesting a surrounded structure of the alumina-coated Co-Ni alloy [39]. The signal of Mg was extremely weak which meant that there was almost no retention of Mg (Figure 2a). Further evidence of the generation of the surrounded structure, as well as the partial residual of Al₂O₃ on the metal surface, could be found on the basis of TEM (Figure 2b). Upon reduction, the lattice spacing converged between that characteristic of metallic cobalt (111) and metallic nickel (111), offering further substantiation for the emergence of an alloy phase. Conversely, the catalyst synthesized through impregnation exhibited a particle size of 14.3 nm, and a notable increase compared to the Co_0.52Ni_0.48@Al₂O₃-IE catalyst (10.2 nm) (Figure 2b). This observation served as confirmation that the encapsulated structure fabricated through ion exchange adeptly restrained the growth of metal particles during the processes of calcination and high-temperature reduction. This encapsulation mechanism, arising from ion exchange, played a pivotal role in controlling the particle size distribution and preserving the nanostructure integrity of the Co_0.52Ni_0.48@Al₂O₃-IE catalyst, contributing to its enhanced stability and catalytic performance, eventually.

Moreover, we investigated the reduction behavior of the catalysts by H₂-TPR. Ni@Al₂O₃-IE had two hydrogen-consuming peaks at 485 °C and 588 °C attributable to β-NiO which had strong interaction with alumina [37,43]. The TPR curves of Co@Al₂O₃-IE showed peaks at 327 °C and 422 °C corresponding to the reduction of Co₃O₄ and CoO, respectively [16,44]. The peak at 678 °C was due to CoO_x-Al₂O₃ or CoAl_xO_y. Owing to the similar crystal structures of Co₃O₄ and γ-Al₂O₃, it was likely to have ion migration from cobalt oxide into the Al₂O₃ support [45,46] (Figure 3a). For bimetallic catalysts with different Co/Ni ratios, the reduction peaks gradually shifted to lower temperatures with increasing Ni content, suggesting interaction between cobalt–nickel oxides. Meanwhile the rise in peak area with an escalating Co/Ni ratio was attributed to the higher molar consumption of H₂ by cobalt compared to nickel, partly due to the presence of trivalent Co. In contrast to the catalysts synthesized through impregnation, the Co_0.52Ni_0.48@Al₂O₃-IE catalyst exhibited a comparatively elevated reduction temperature. This observation implied the more robust interactions with the support, a factor known to enhance catalyst stability, as illustrated in Figure 3b. The higher reduction temperature signified enhanced metal–support interactions, leading to a more secure anchoring of the active metal species on the Al₂O₃ support. This strengthened interaction was instrumental in fortifying the catalyst against sintering and agglomeration during the reduction process, thereby contributing to its sustained catalytic activity over prolonged reaction cycles [31,47].

The BET surface area, total pore volume, and pore size of the catalysts are listed in

Table 1. N₂ adsorption–desorption isotherms, pore size distributions, and pore volume.

	SBET (m2/g)	Vtotal (cm3/g)	Pore Size (nm)
Co0.52Ni0.48@Al2O3-IE	246.9	0.39	6.3
Co0.27Ni0.73@Al2O3-IE	199.8	0.40	8.1
Co0.71Ni0.29@Al2O3-IE	176.1	0.41	9.3
Co0.50Ni0.50/Al2O3-IM	132.5	0.67	21.8
γ-Al2O3	141.0	0.86	24.6

. The Co_0.52Ni_0.48@Al₂O₃-IE catalyst exhibited a large specific surface area of 246.9 m²/g and a relatively homogeneous mesoporous structure, whereas the specific surface area decreased with the variation in the Co/Ni ratio, and the slight deviations may be due to the slight difference in Ksp between Co and Ni, which caused inhomogeneity in the carrier layer after exchange and may have resulted in the accumulation of pore structures. Relatively, the Co_0.50Ni_0.50/Al₂O₃-IM catalyst prepared by the impregnation method, of which the specific surface area was only 132.5 m²/g, was significantly lower than that of the catalyst prepared by the ion-exchange method. Figure 4 displays the isothermal adsorption–desorption and pore size distribution curves of the catalysts. The Co_0.52Ni_0.48@Al₂O₃-IE catalyst exhibited a characteristic type IV isotherm profile with an H2-type hysteresis loop. Moreover, the pore size distributions of the ion exchange (IE) series were noticeably more uniform than the catalyst of the impregnation method (IM). In comparison to other catalysts, Co_0.52Ni_0.48@Al₂O₃-IE possessed substantial specific surface area and uniform pore structure, favoring the promotion of reactant diffusion and molecular transfer, and contributing to the enhancement in catalyst activity.

The investigation of catalyst chemical compositions and electronic states was performed through X-ray photoelectron spectroscopy (XPS) characterization. This analytical technique enabled a detailed examination of the surface elemental composition and oxidation states of the catalyst, providing valuable insights into its chemical structure and electronic configuration. After the reduction treatment of the calcined metal oxide catalysts, it was clearly observed that the catalysts were reduced to the metallic states. The presence of some oxide signals in the peaks may be attributed to oxidation during the testing process. Notably, the Ni 2p region (Figure 5a) showed two predominant peaks at 852.93 eV and 855.74 eV which could be ascribed to Ni⁰ and Ni²⁺, respectively [38]. Figure 5b shows that the Co 2p spectra had three main peaks including Co⁰ (~777.87 eV), Co³⁺ (~780.13 eV), and Co²⁺ (~781.86 eV) [39,40]. The O 1s fine spectrum revealed three distinct signal peaks, each associated with specific oxygen species (Figure 5c). These included lattice oxygen (O_I), surface defect-related oxygen (e.g., oxygen vacancies), and surface hydroxyl groups (O_II) encompassing O₂²⁻ and O⁻, as well as adsorbed CO₂ (O_III). Notably, the higher O_II/O_I suggested that the bimetallic catalysts possessed a higher density of oxygen vacancies. This increased abundance of oxygen vacancies enhanced the adsorption capacity of carbonyl groups present in the reactant molecules [48]. In contrast to the monometallic catalysts, the Co peaks of the bimetallic catalyst exhibited a discernible shift toward lower binding energy, whereas there was a noticeable increase in the binding energy of the Ni 2p peaks of Ni⁰ and Ni²⁺. These shifts served as a compelling indicator of intermetallic electron interactions between Ni and Co, indicating the generation of CoNi alloys [49]. As a consequence of this charge transfer mechanism, there was the generation of electron-enriched Co⁰ centers on the catalyst surface [50,51].

The catalytic efficiency of catalysts was assessed in the context of LA hydrogenation to GVL. Among the bimetallic catalysts with different Co/Ni ratios, Co_0.52Ni_0.48@Al₂O₃-IE showed the highest GVL yield, and thus we optimized the reaction conditions using this catalyst. Interestingly, the type of alcohol solvent had a significant effect on GVL selectivity, with isopropanol (IPA) leading to higher selectivity and yield compared to the primary alcohol. This finding was consistent with those of previous reports [5,30,52] (Figure 6a).

Regarding H₂ pressure, a pivotal reducing agent in the hydrogenation of LA to GVL, its significance is demonstrated in Figure 6b. Notably, the GVL yield reached nearly 94% when the H₂ pressure was maintained at 2 MPa, underscoring the facilitating role of moderate H₂ pressure in the reaction. Furthermore, elevating the H₂ pressure from 1 MPa to 3 MPa led to LA conversion and GVL yield exceeding 99.5%. Nevertheless, raising the H₂ pressure beyond 2 MPa failed to yield further enhancement in GVL yield, signifying that a pressure of 2 MPa was adequate for the hydrogenation of LA to GVL. Figure 6c illustrates the outcomes of temperature adjustments spanning from 110 °C to 150 °C. Notably, no significant alterations were observed beyond this temperature range (130 °C to 150 °C). The reaction time played a pivotal role in the hydrogenation of LA to GVL. We meticulously optimized the reaction time, and the results are depicted in Figure 6d. The optimal conditions were reached with a 4 h reaction time, where a remarkable 97.6% GVL yield and complete conversion of LA were achieved. Extending the reaction time beyond 4 h did not result in a significant improvement in GVL yield, indicating that the GVL product is stable in the current system. Overall, we determined that the optimal reaction conditions are as follows: 4 h of reaction time, 130 °C, and 2 MPa H₂. These conditions were notably milder than those of typical ones reported in the literature, and we could reach a remarkable GVL yield exceeding 99% (

Table 2. Hydrogenation of LA to GVL over various catalysts under optimal reaction conditions.

Entry	Catalyst	Solvent	T/°C	H2/MPa	t/h	Y/%	Cycles	Ref.
1	Co0.52Ni0.48@Al2O3	IPA	130	2	4	>99	5	this work
2	Ni–MoOx/C	—	250	5	24	>99	—	[53]
3	Ni–Sn(1.4)/AlOH	H2O	120	4	2	>99	—	[54]
4	25%Fe-25%Ni/MMT	IPA	200	3.5	1	>99	6	[22]
5	Ni1-Zn1@OMC	H2O	180	2	1.5	>93	2	[29]
6	Ni(0)@boehmite	H2O	200	3	6	>99	2	[55]
7	Fe-Re/TiO2	H2O	180	4	4	95	—	[6]
8	RuHZSM-5	1, 4-dioxane	220	3	10	93.1	3	[56]

).

When comparing bimetallic catalysts to their monometallic counterparts (Figure 7), it was evident that the bimetallic catalyst exhibits significantly higher activity. The enhanced activity could be attributed to the synergistic interaction between Ni and Co, and the prominent specific surface area facilitating the reactant diffusion and molecular transfer [50]. Ultimately, we speculate that the reaction mechanism is as follows: In comparison to their monometallic counterparts, the bimetallic CoNi nanoparticles (NPs) demonstrated a heightened abundance of surface oxygen vacancies. These vacancies readily coordinated with oxygen atoms present on the carbonyl groups of LA, resulting in the elongation and weakening of the C=O double bond. Apart from the generation of active hydrogen through the dissociation of molecular hydrogen, the bimetallic CoNi NPs exhibited an additional capability: they facilitated the dehydrogenation of isopropanol (IPA), leading to the production of active hydrogen. This dual activation of hydrogen proved to be effective in attacking the less robust C=O double bond, thereby expediting a swift hydrogenation reaction to yield the 4-HA intermediate [48,49]. Both the partial acid sites provided by acidic support Al₂O₃ and a relatively high temperature facilitated the dehydration cyclization of 4-HA into GVL [3,57,58].

We assessed the recyclability of the catalyst (Figure 8). The capacity of a catalyst to consistently sustain its catalytic activity and selectivity across successive reaction cycles holds paramount significance for its pragmatic utilization and economic feasibility in industrial applications. A catalyst characterized by superior cyclic stability not only ensures prolonged effectiveness over an extended operational duration but also plays a crucial role in diminishing the overall cost associated with the catalytic process. This dual advantage enhances the environmental sustainability and economic viability of the process. When comparing the recyclability performance between the ion-exchange (IE) and impregnation (IM) catalysts, a discernible difference emerged. The former demonstrated consistent and stable activity over five consecutive test runs, whereas the latter experienced deactivation after only two runs. This observation underscores the pivotal role played by the encapsulation of alloy particles and the potential influence of residual Al₂O₃ on the surface of the cobalt–nickel alloy, both possibly contributing to the stabilization of Co_0.52Ni_0.48@Al₂O₃-IE.

3. Materials and Methods

3.1. Materials

Magnesium carbonate basic (AR, 40–45% MgO basis) was procured from Rhawn (Shanghai, China). Nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O, AR, ≥98.0%) and cobalt nitrate hexahydrate (Co(NO₃)₂·6H₂O, AR, ≥98.5%) were acquired from Sinopharm (Shanghai, China). Levulinic acid (LA, AR, 99.0%) was sourced from Shanghai Macklin Biochemical Co., Ltd. (Shanghai, China). γ-Valerolactone (GVL) and γ-Alumina (Al₂O₃) were obtained from Aladdin (Shanghai, China). Aluminum nitrate (Al(NO₃)₃·9H₂O, AR, 99.5%) was procured from Nanjing Reagent (Nanjing, China). All of the aforementioned chemicals were utilized as received without further purification, ensuring their chemical integrity and suitability for the intended experimental procedures.

3.2. Catalyst Preparation

3.2.1. Synthesis of CoNi-MMH Precursors

Magnesium carbonate basic underwent calcination at 750 °C for 2 h in a static air atmosphere, resulting in the formation of magnesium oxide (MgO). In a standard synthesis protocol, the MgO obtained (1.21 g, 0.03 mol) was immersed in a solution containing cobalt and nickel ions with varying Co/Ni ratios, maintaining a total “Co + Ni” concentration equivalent to that of MgO, stirring at 500 revolutions per minute for 24 h at room temperature. Following this, the resulting dark green CoNi-MMH products were filtered and washed seven times with deionized water (the green color deepened as the proportion of Co increased). Ultimately, the products were dried at 60 °C overnight. This meticulously executed procedure ensured the complete exchange of cobalt and nickel ions with magnesium ions of the template, resulting in the desired CoNi-MMH for further study.

3.2.2. Synthesis of Co_xNi_y(OH)₂@Al(OH)₃

CoNi-MMH and aluminium nitrate were dissolved in 60 mL of deionized water solution at a molar ratio of 2/1, followed by rapid stirring for 30 min, otherwise a gel may form. Subsequently, this mixture was transferred into a 100 mL Teflon-lined stainless steel autoclave, then hermetically sealed, and maintained at a temperature of 120 °C for a duration of 12 h. Following the completion of the reaction, the resulting precipitate, denoted as Co_xNi_y(OH)₂@Al(OH)₃, was isolated through centrifugation, followed by thorough rinsing with deionized water. Lastly, the obtained product underwent a drying process at 80 °C overnight. This meticulous procedure ensured the successful formation of Co_xNi_y(OH)₂@Al(OH)₃, a precursor material for subsequent investigations.

3.2.3. Synthesis of Co_xNi_y@Al₂O₃-IE

Following calcination of Co_xNi_y(OH)₂@Al(OH)₃ under a static air atmosphere at 400 °C for 2 h, a gray powder was obtained. Subsequently, the powder underwent a reduction in a 5% H₂/Ar gas flow at 650 °C (with a heating rate of 5 °C min⁻¹) for a duration of 2 h, resulting in the formation of Co_xNi_y@Al₂O₃-IE. Various samples with different metal loadings can be prepared by carefully adjusting the Al³⁺/bimetallic hydroxide molar ratio. In Co_xNi_y@Al₂O₃-IE, the variables x and y represent the Ni and Co mass percentages determined by ICP-OES, respectively. The cumulative metal loading was meticulously controlled to be 25%. This systematic procedure ensured the reproducibility and accuracy of the metal loadings in Co_xNi_y@Al₂O₃-IE, facilitating precise analysis and interpretation of its catalytic properties.

3.2.4. Synthesis of Comparative Catalysts

The catalyst for comparison was prepared using a conventional impregnation method, noted as Co_0.50Ni_0.50/Al₂O₃-IM. The preparation process for surrounded single (Co or Ni) catalysts was like that of bimetallic catalysts described earlier: Through ion exchange, single-metal nanosheets were prepared. Initially, a concentrated 1 M solution of nickel nitrate or cobalt nitrate was prepared. Subsequently, 6.40 g of freshly prepared magnesium oxide was added to the above solution. The mixture was stirred at room temperature for 48 h. The excess nickel nitrate or cobalt nitrate and an extended reaction time facilitated complete ion exchange between metal ions and Mg(OH)₂ templates. Following filtration, the resulting product was washed multiple times with water and ethanol, filtered, dried, and then ground to obtain light green Ni(OH)₂ or dark green Co(OH)₂ nanosheets. The ensuing preparation process was similar to that of the previous bimetallic catalysts.

3.3. Catalyst Characterization

The chemical composition was analyzed using ICP-OES (Avio500, PerkinElmer, Waltham, MA, USA). XRD patterns were obtained using a Philips X′Pert XRD (PANalytical, Almelo, the Netherlands) with Co Kα radiation (35 kV, 40 mA, λ = 1.7902 Å). Scanning electron microscopy (Hitachi S-4800 of JPN) was employed to characterize the morphology of the products. HRTEM images were captured on a JEM-2100F JEOL electron microscope (Tokyo, Japan) operating at 200 kV. Element mapping results were acquired with an X-maxN80T (Oxford Instruments, Oxford, UK) equipped with an energy dispersive spectrometer (EDX). The Micromeritics ASAP 3020 (Norcross, GA, USA) was employed for the analysis of the specific surface area in the samples. The BET method within the relative pressure (P/Po) range of 0.04–0.3 was used to determine the specific surface area, and the pore size distribution curve was calculated using the BJH method. XPS analysis was performed on an Escalab-250Xi system using a monochromatic Al Kα X-ray source, and all binding energies were referenced to the C 1s line at 284.8 eV. H₂-TPR experiments were conducted on a FINESORB-3010 (Finetec Instruments, Zhejiang, China) instrument with a thermal conductivity detector (TCD). Prior to TPR measurements, a standard pretreatment procedure involved subjecting the sample (40 mg in all runs) to a 2 h pretreatment at 300 °C under Ar flow to eliminate any absorbed gases. Following pretreatment, the sample was gradually cooled to room temperature and subsequently heated to 800 °C at a rate of 10 °C min⁻¹ in a 5% H₂/Ar gas flow (25 mL min⁻¹).

3.4. Catalytic Test

The performance testing of the catalyst for LA hydrogenation was conducted in a high-pressure liquid-phase reaction vessel. Initially, 10 mL of isopropanol solvent, 102.4 μL of LA, 36 μL of internal standard 1,4-dioxane, and 35 mg of freshly reduced catalyst were placed in a Teflon liner. Following this, the reaction vessel was then sealed, with subsequent purging of 0.1 MPa H₂ to displace air from the reaction vessel multiple times. Finally, after introducing 2 MPa H₂, the reaction proceeded at 130 °C for 4 h. The results were then analyzed for composition by GC (GC-9860) with a capillary column and a flame ionization detector (FID). The chromatographic column was FFAP (length: 30 m, inner diameter: 0.32, and film thickness: 0.50 μm). The catalyst’s performance was evaluated based on several reference values, including the following:

The conversion of LA was calculated as follows:

$$\mathrm{Conv}.(\%)=(\frac{m_{LA,0}-m_{LA,T}}{m_{LA,0}})\times 100$$

GVL selectivity was calculated as follows:

$$Sel.(\%)=(\frac{m_{GVL}}{m_{LA,0}-m_{LA,T}})\times 100$$

The yield of GVL was calculated as follows:

$$\mathrm{Yield}(\%)=\frac{\mathrm{Conv}.\times Sel.}{100}$$

4. Conclusions

In this work, we successfully generated an encapsulated catalyst with good alloying, which can achieve more than 99% GVL yield in LA hydrogenation under relatively mild reaction conditions thanks to the synergistic interaction between metals. Both the active Co-Ni alloy particles physically encapsulated by alumina, because of the surrounded structure resulting from ion exchange, and some alumina on the surface of the alloy particles, due to the reversibility of exchange equilibrium, may have contributed to the stabilization of the catalyst. Compared to the catalyst prepared by impregnation where the activity started to decrease after two cycles, the encapsulated Co_0.52Ni_0.48@Al₂O₃-IE counterpart showed no deactivation in a test of five cycles, displaying superiority in stability. The results show that the bimetallic catalyst synthesized by the ion-exchange method has high efficiency and stability in the production of GVL for the LA hydrogenation reaction. Moreover, the inverse loading synthesis by the ion-exchange method can be extended to many other valuable catalytic systems based on the differences in solubility product constants.