Dehydrogenation of Diethylene Glycol to Para-Dioxanone over Cu/SiO₂ Catalyst: Effect of Structural and Surface Properties

Abstract

Para-dioxanone, a typical monomer for high-performance biomedical polymers, is generally obtained from the catalytic dehydrogenation of diethylene glycol. In this work, Cu/SiO₂ catalysts were prepared by two different methods and applied in diethylene glycol dehydrogenation. The effects of catalyst properties on the performance of para-dioxanone production were systematically explored by combined techniques. The results showed that the high specific surface area and regular pore structure of the catalyst promoted and stabilized the high dispersion of the copper species with an abundant Cu-Si interface, thereby providing numerous reactive sites. The appropriate Cu⁰/Cu⁺ ratio in the active components produced efficient synergy, which was responsible for the excellent dehydrogenation performance. In addition, the low surface acid density (0.532 μmol/m²) of the catalyst greatly reduced the occurrence of diethylene glycol dehydration, thus improving para-dioxanone selectivity over 90% with promising stability.

1. Introduction

As a kind of degradable aliphatic polyester, poly-para-dioxanone (PPDO) is extensively used in producing biomedical materials like surgical sutures, bone lamella, etc. The PPDO-based product reveals excellent flexibility and biocompatibility due to the presence of abundant ether and ester bonds in the molecular chains. Generally, PPDO is synthesized by ring-opening polymerization of the monomer para-dioxanone (PDO) [1]. The persistent supply of high-quality PDO monomers guarantees excellent performance and high-end application of the PPDO materials. Until now, nevertheless, PDO is still not an industrialized product that can be commonly available, which hinders the wide application of advanced biomedical materials. The easy and low-cost PDO production came to be a pivotal breakpoint. In the scattered reports, the direct synthesize of PDO from diethylene glycol (DEG) catalytic dehydrogenation was the most efficient route with high atom economy (Figure 1). It offers a sustainable path for PDO production, providing an alternative to traditional, energy-intensive processes with complicated steps. Towards the dehydrogenation method, developing a stable and effective catalyst for the reaction is decisive and highly desired but still challenging, which attracts much attention from researchers. In recent years, related works have been raised, and some meaningful results have been present. Guest et al. [2] selected alumina-supported copper chromite as a catalyst for gaseous dehydrogenation of DEG to PDO; the highest yield of the product reached 83.6%. Mayhew et al. [3] carried out the dehydrogenation reaction at 250 °C over 5% reduced copper on a zeolite catalyst in the presence of hydrogen, giving a PDO yield of 90.2%. Hattori et al. [4] applied spherical silica supported copper in the DEG-to-PDO reaction, restricting the number of acid sites with a strength of H₀ over 1.5 on silica support to less than 1.5 × 10⁻⁷ eq/m². The regulated catalyst surface greatly reduced the formation of byproducts, thus elevating the selectivity of PDO. Sun et al. [5] modified the Cu/SiO₂ catalyst with various amounts of Al. The conditions of catalyst modification and reaction temperature, pressure, and gas velocity of DEG dehydrogenation were further optimized, achieving the best DEG conversion and PDO selectivity at 97.2% and 96.4%, respectively. By organizing previous studies, it can be found that copper-based catalysts with inert support revealed promising dehydrogenation performance of DEG. The synergy between copper species and silica support in Cu/SiO₂ catalysts presents an intriguing avenue for catalytic exploration. But so far, the relevant reports have stopped with the screening of different catalyst components and the evaluation of various dehydrogenation conditions. The comprehensive understanding of the intrinsic properties of catalysts that determine catalytic activity and selectivity is quite limited, which greatly impedes the development of sustainable and economically viable processes for the synthesis of valuable chemicals like PDO. Herein, by employing Cu/SiO₂ catalysts from different preparation methods in DEG dehydrogenation, the influence of catalyst structure and surface properties on the dehydrogenation performance of DEG to PDO was studied, with a correlation between the reaction results and catalyst characterizations. Some key factors that govern various catalytic behaviors were found and fully elucidated. By addressing these objectives, we expect the present research could deepen some insights into the structure-activity relationship in terms of PDO production and bring new opinions on the construction of a Cu-based alcohol dehydrogenation catalyst.

2. Results and Discussion

2.1. Dehydrogenation Performance of the Catalysts

Figure 2 displays the catalytic performance in the DEG dehydrogenation to PDO reaction of two copper catalysts obtained with different preparation methods. It could be found in Figure 2a that the catalysts both performed fine reactivity at 260 °C and liquid hourly space velocity (LHSV) of 1 h⁻¹. DEG conversion of CuSi-ae reached 90% and CuSi-dp approached 99% at the early stage of the reaction, demonstrating copper-based catalysts revealed superior catalytic activity in DEG gas phase dehydrogenation. In addition, the activity of CuSi-ae was well-maintained with no obvious decline in reaction time over 80 h, while the activity of CuSi-dp gradually decreased to 90%, indicating the inferior catalytic stability of CuSi-dp, but no remarkable disparity existed.

Correspondingly, Figure 2b shows the reaction product selectivity of two catalysts. For CuSi-dp, despite the catalyst’s high reactivity with DEG, the selectivity of the target product PDO was only ca. 30–50%. A part of DEG proceeded with intra-molecular dehydration to dioxane with a selectivity of 40–50%, which was the main side reaction of DEG dehydrogenation. In contrast, a greatly distinct reaction behavior was observed on CuSi-ae. The selectivity of DEG to PDO over CuSi-ae was nearly 90%, while dioxane was barely detected during the whole test range. In addition, we also detected dioxene, methyldioxolane, etc. as by-products for both catalysts, but they were not shown here due to the negligible amount compared with PDO and dioxane. The result indicated CuSi-ae could effectively convert DEG to PDO with a favorable yield. Despite the fact that two catalysts were roughly the same in chemical composition, the distinct structural and surface properties of various preparation methods may contribute to the significant difference in catalytic performance. Therefore, a series of characterizations and analyses of the catalyst’s microscopic properties were carried out.

2.2. Structural Properties of Catalysts

N₂-sorption measurements on the catalysts were conducted to explore their texture properties; the obtained isotherms and pore size distribution are plotted in Figure 3. The N₂ adsorption-desorption curve of CuSi-dp showed a hysteresis loop of nearly H3 type, suggesting abundant crevice mesoporous crevice in the catalyst, which mainly originated from the stacking of oxide particles. This kind of pore characteristic was inherited from pristine SiO₂. The deposition of copper oxides reduced pore volume but kept physisorption behaviors. The pore size distribution of CuSi-dp in Figure 3b showed a broad curve with an average pore size of 10.8 nm, which is consistent with the feature of stacking pore structure. The minor increase in pore size from silica was probably due to the aggregation of isolated Cu ions in the precipitant during calcination, which expanded the stacking framework between oxide particles to some extent. On the other hand, CuSi-ae showed a slightly distorted H2 type hysteresis loop in the N₂ adsorption-desorption curve, indicating the presence of a tube-like pore structure with an irregular arrangement [6]. It also revealed a highly uniform pore size distribution in the range of 3–4 nm that was quite distinguished from CuSi-dp. This fine pore structure was believed to be beneficial for the effective diffusion of reactants, further elevating the catalytic efficiency. In addition, the specific surface area of CuSi-ae, 574 m²/g (

Table 1. Textural properties of the catalysts.

Catalyst	BET Specific Surface Area/m2·g−1	Pore Volume/cm3·g−1	Average Pore Size/nm	CuO Crystal Size a/nm	Cu Crystal Size a/nm	Cu Surface Area/ m2·g−1	Cu Particle Size b/nm	Cu Dispersion/%
SiO2	402	0.93	7.5	--	--	--	--	--
CuSi-dp	248	0.80	10.8	9.7	--	--	--	--
CuSi-ae	574	0.70	3.7	--	--	--	--	--
CuSi-dp-r	213	0.54	11.2	--	18.6	14.5	13.4	7.1
CuSi-ae-r	361	0.20	2.9	--	4.3	46.1	4.1	22.3

^a Calculated by the Scherrer equation from XRD data. ^b Obtained from N₂O chemisorption.

), was significantly higher than that of CuSi-dp (248 m²/g). Although both samples originated from the same support materials, the basic and hydrothermal environment in the synthesis procedure of CuSi-ae led to the silica dissolution and rearrangement with copper species, forming a brand-new spatial structure away from pristine silica. For the catalysts that were practically employed in the reaction, the reduced samples were also tested, and Figure S1 presents the results. It was found that H₂ reduction decreased the surface area and pore volume of both catalysts, while the adsorption-desorption curves remained unchanged. Despite the fact that the physical structure parameters of catalysts varied during reduction, the greatly expanded surface area of the CuSi-ae catalyst still provided much more anchorage for copper species, exposing abundant dehydrogenation active sites after catalyst reduction. Moreover, the smaller pore structure of CuSi-ae-r restricted the active component at high reaction temperatures, which enhanced the structural stability of the catalyst. The difference in continuous dehydrogenation reaction results from Figure 2 further verified this deduction.

Furthermore, the morphology features of the catalyst were examined by TEM; the images are present in Figure 4. As could be seen in the image of CuSi-ae, it showed an abundant tubular structure with random orientation, which were consistent with the speculation from N₂ adsorption-desorption data. Based on the previous literature, the typical tubular morphology of CuSi-ae was ascribed to the scrolling of lamellar copper phyllosilicate [7,8,9]. In addition, no aggregated CuO crystallite was observed in the image, indicating the uniform distribution of copper ions. These phenomena collectively implied that CuSi-ae was composed of copper silicate species formed during preparation. The TEM image of CuSi-dp displayed the evident stacking structure of oxide particles, presenting the interval pore between the bulk oxides. CuO crystals with a higher contrast in field could be clearly distinguished from those depositing on silica surfaces and demonstrated a typical morphology of the supporting catalyst. SEM and EDS mapping results of both samples in Figures S2 and S3 also demonstrated the homogeneous distribution of Cu and Si species, indicating a close combination between them. The microscopy results further proved the discrepancy between the two catalysts in texture, especially the morphology of support and Cu species. CuSi-ae revealed highly-uniformed dispersion of Cu species, generating expanded interfaces with support. The developed tubular pore structure brings opportunities for improving transfer efficiency and restraining the occurrence of DEG dehydration.

The XRD patterns of calcined catalysts were depicted in Figure 5a, which gave their crystalline structure information. CuSi-dp revealed a broad diffraction band at 2θ = 20°–30°, which was attributed to the amorphous silica support. In addition, the pattern emerged from two intense diffraction peaks at 2θ = 35.5° and 38.7°, which belonged to crystallite CuO, and the crystal size was determined as 9.7 nm by the Scherrer equation, indicating the Cu species in calcined CuSi-dp is present as CuO. For CuSi-ae, the absence of a broad band from amorphous silica implied the transformation of support. Instead of that, the diffraction pattern showed two weak peaks at 2θ = 31° and 35.4°, corresponding to the signal of copper phyllosilicate (Cu₂Si₂O₅(OH)₃) [10,11]. No evident CuO peak appeared in the meantime, which is in line with the catalyst component information given by TEM. Furthermore, the formation of copper phyllosilicate could also be verified by the FT-IR results in Figure S4. The spectrum of CuSi-ae emerged peak at 675 cm⁻¹ in the magnified vision, which was ascribed to δ_OH band in copper phyllosilicate. Due to the asymmetric ν_SiO band of silica at 1116 cm⁻¹ and the ν_SiO band of copper phyllosilicate at 1035 cm⁻¹ overlapping with each other, the δ_OH band was sensitive to identifying the presence of phyllosilicate [12,13,14,15]. These results collectively illustrated that almost all the Cu precursors reacted with silica, giving copper phyllosilicate in the preparation process. On account of the requirement of catalyst reduction before applying to dehydrogenation, we also collected in situ XRD patterns of the catalysts after being reduced in hydrogen flow at 250 °C. Figure 5b shows the results. It could be found that both catalysts displayed only diffraction peaks of metallic Cu after reduction, suggesting the Cu species in calcined catalysts were all reduced. The Cu crystallite particle size of CuSi-dp after reduction were calculated at 18.6 nm, but the data were only 4.3 nm for CuSi-ae. The results obtained from N₂O chemisorption also showed a similar trend. The obvious distinction in Cu particle size demonstrated that the preparation method greatly influenced the catalyst structure. CuSi-ae with phyllosilicate as a Cu precursor effectively restricted Cu particle size in reduction and prominently promoted the dispersion of Cu species. Correspondingly, the Cu surface area estimated by N₂O titration also gave a distinct value of 14.5 m²/g for CuSi-dp and 46.1 m²/g for CuSi-ae, respectively. The increased active site exposure by high dispersion (22.3% vs. 7.1%) was believed to elevate the dehydrogenation efficiency [16,17].

In order to investigate the redox capacity, an H₂-TPR experiment was conducted on the two catalysts, and Figure 6 gives the reduction curves. Apparently, CuSi-dp emerged as the main peak at 162.8 °C, corresponding to the reduction of lattice oxygen in bulk CuO crystallite. The shoulder peak at 174.1 °C could be attributed to the reduction of partial CuO that closely contacts and interacts with support. Compared to bulk CuO, its reduction tended to postpone to a higher temperature [18]. CuSi-ae showed a single intense reduction peak at 186.0 °C, which originated from the reduction of copper phyllosilicate present in the catalyst. Nevertheless, the reduction peak clearly shifted to a higher temperature in comparison with CuSi-dp. In the procedure of CuSi-ae synthesis, the main phase copper phyllosilicate dominantly formed with solid chemical bonding between Cu and Si; the components strong interaction hindered the reduction of copper species [19]. In addition, all the reduction processes of two catalysts terminated before 250 °C in H₂-TPR, indicating the pretreatment of catalysts at 250 °C before reaction guaranteed total reduction of Cu²⁺ species in the catalysts.

Since CO molecules showed specific adsorption on copper species, the chemical state of copper in the reduced catalysts was rationally characterized by the CO-probed in situ FT-IR technique. The self-supporting catalyst disk with a diameter of ca. 10 mm was reduced in the IR cell at 250 °C for 1 h after being purged by N₂ at 150 °C. Subsequently, CO was introduced to the catalyst surface to conduct adsorption for 30 min, followed by evacuation and CO desorption at elevated temperatures. The CO-adsorbed IR spectra were recorded in this process at different temperatures, and the results are depicted in Figure 7. As reported in previous reports, the IR signal of CO adsorbed on the Cu⁺ site (CO-Cu⁺) usually gives a higher wavenumber than that of CO on the Cu⁰ site (CO-Cu⁰), due to the stronger electron-donating effect from CO to the Cu⁺ site [20,21,22]. In addition, the adsorption stability of CO-Cu⁺ is much higher than that of CO-Cu⁰; the latter configuration will desorb rapidly under heating or flow purging. Accordingly, for CuSi-dp, the IR signal of CO adsorption is located at 2120 cm⁻¹, and during temperature increases, the adsorbed CO on the catalyst surface almost desorbed completely at 150 °C, suggesting the adsorption peak was mainly contributed by CO-Cu⁰. While CO adsorbed on CuSi-ae, it gave the IR signal at 2126 cm⁻¹. The obvious blue shift in signal wavenumber implies that, apart from CO-Cu⁰, the content of CO-Cu⁺ prominently increased on the CO-adsorbed CuSi-ae surface. On the other hand, the IR signal showed stronger stability; it gradually disappeared until temperatures exceeded 200 °C. The phenomenon also reflected the increased ratio of CO-Cu⁺ species on the surface of CuSi-ae compared to CuSi-dp. In addition, the 2045 cm⁻¹ species was assigned to dicarbonyl geminal complexes formed on copper atoms with a low coordination number, while the 2003 cm⁻¹ species could be a monocarbonyl one [23]. However, both species were present at only some specific adsorption conditions (low pressure or bigger metallic copper particles) and were not stable; they would rapidly diminish upon more CO feeding or temperature elevation, as we could observe in the spectra. In order to further clarify the relative proportions of Cu⁰ and Cu⁺ in catalysts with different structures after reduction, the Auger spectra of Cu LMM were collected; Figure 8 shows the results. The spectra were first deconvoluted to two peaks locating at 916.4 eV and 918.5 eV, which were ascribed to the signals of Cu⁺ and Cu⁰ species, respectively [20,24,25,26]. The integration of peak area showed that the ratio of Cu⁺/(Cu⁺+Cu⁰) in reduced CuSi-ae, 65.97%, was evidently higher than 40.42% of Cu⁺/(Cu⁺+Cu⁰) in reduced CuSi-dp. CO-adsorbed IR and Auger results collectively manifested that the reduced copper species in CuSi-ae with silicate as the precursor were mainly in the form of Cu⁰ and Cu⁺, while in CuSi-dp with oxide as the precursor, Cu⁰ was present as the majority. Combining the evaluation of catalytic performance, the promising activity of CuSi-ae indicated that a higher proportion of Cu⁺ sites could show more effective synergistic interaction with other Cu⁰ sites, which was conducive to the adsorption of DEG reactants and the stability of hemiacetal intermediates, thus promoting DEG dehydrogenation efficiency [13,27,28].

2.3. Surface Acid Properties of Catalysts

Apart from DEG conversion, the selectivity of product PDO was also a crucial indicator of catalytic dehydrogenation performance. In terms of two catalysts with different structure properties, a notable distinction in efficiency in directly converting DEG to PDO was exhibited, as illustrated clearly in Figure 2. A considerable part of DEG dehydrated on the CuSi-dp surface, forming dioxane, which obviously deteriorated PDO yield. According to previous studies, the alcohol dehydration over the oxide catalyst surface was generally catalyzed by acid sites [29,30,31,32]. Therefore, the NH₃-TPD method was employed to investigate the acid properties of two catalyst surfaces; the profiles are shown in Figure 9. It could be distinguished from the desorption curves that two catalysts both arose clear NH₃ signals at 100–300 °C and 300–400 °C, which were on account of NH₃ molecules adsorbed on weak acid sites and medium-strong acid sites, respectively. The similar peak positions suggested the types of acid sites were roughly the same on the two catalysts. However, the great difference was revealed in the intensity of desorption peaks, i.e., the surface acid amounts of the two catalysts. Both the weak and medium-strong acid amounts of CuSi-dp were much higher than those of CuSi-ae. For quantitative comparison, the desorption peak area was further integrated, and the total acid amount of CuSi-ae surface was calculated as 0.192 mmol/g, while CuSi-dp reached 0.521 mmol/g. The much lower surface acid amount combined with a higher specific surface area meant the density of the surface acid site was significantly reduced (0.532 μmol/m² vs. 2.45 μmol/m²), which largely decreased the adsorption amount of DEG on the acid site. The protonation of the DEG hydroxyl group at the acid site and the subsequent dehydration reaction were accordingly inhibited. In addition, the highly dispersed copper species provided more dehydrogenative sites, and a suitable tubular pore structure promoted efficient diffusion of products. These aspects collectively improved the PDO selectivity.

3. Experimental

3.1. Catalyst Preparation

A total of 30.37 g of Cu(NO₃)₂·3H₂O (AR, Sinopharm, Co., Ltd., Shanghai, China) was dissolved with deionized water in the IKA LR1000 reactor. Approximately 12 g of amorphous silica (Qingdao Haiyang Chemical Co., Ltd., Qingdao, China) was then put into the above solution, and the suspension was stirred and heated to 60 °C. In another beaker, 14.32 g of (NH₄)₂CO₃ (AR, Sinopharm Co., Ltd.) was dissolved with deionized water, and the solution was injected into the reactor with a peristaltic pump to initiate the precipitation. The obtained mixture was further aged for 4 h. After that, the precipitate was collected by filtration, followed by washing and drying at 120 °C. The solid was finally calcined in a muffle furnace at 350 °C for 4 h to obtain the catalyst, which was labeled CuSi-dp. The hydrogen-reduced sample before use in the reaction was denoted as CuSi-dp-r.

A total of 33.42 g of Cu(NO₃)₂·3H₂O (AR, Sinopharm Co., Ltd.) was dissolved with deionized water in a beaker, and NH₃·H₂O (25 wt%, Sinopharm Co., Ltd.) was subsequently added until pH reached 12. Then, 12 g of amorphous silica (Qingdao Haiyang Chemical Co., Ltd.) was put into the above solution and kept stirring for 4 h. The suspension was transferred into an autoclave and placed in the oven at 180 °C for 24 h to reinforce the combination between silica and copper species in solution. After cooling down to room temperature, the mixture in the autoclave was filtrated to obtain a solid product. The cake was washed with deionized water and dried at 120 °C. The solid was finally calcined in a muffle furnace at 350 °C for 4 h to obtain the catalyst, which was labeled CuSi-ae. The reduced sample was denoted as CuSi-ae-r.

The CuO content in both catalysts was determined to be 40 wt% by XRF measurement.

3.2. Catalyst Characterizations

The XRD patterns of catalysts were collected by the D/Max-1400 diffraction spectrometer, Rigaku, Japan, with Cu Kα radiation (λ = 0.154 nm), the scanning range of 2θ was 10–80° with a speed of 2°/min, and the particle size of crystallite in the sample was calculated by the Scherrer equation. In situ XRD patterns were recorded on the same instrument, while the sample was placed and reduced in the flow of hydrogen at 250 °C for 1 h with a heating rate of 10 °C/min before measurement. The TEM images were taken by a JEM-2100 microscope (JOEL, Tokyo, Japan), and the acceleration voltage of the electron was 200 kV. Micromeritics Tristar 3000, Norcross, GA, USA was used to analyze the pore structure and specific surface area of the catalysts with the Barret–Joyner–Halenda (BJH) and Brunauer–Emmett–Teller (BET) methods. The sample was degassed at 300 °C for 4 h and tested at liquid nitrogen temperature. The chemical composition of the catalysts was determined by X-ray fluorescence spectroscopy (XRF) using a Rigaku, Tokyo, Japan ZSX PrimusⅡX-Ray spectrometer, which is equipped with a Rh target X-ray tube at a power of 3 kW. FT-IR spectra of the samples were recorded on a Nicolet Is10 FT-IR spectrometer (Thermofisher scientific, Waltham, MA, USA) in an ATR pattern from 400 to 4000 cm⁻¹. The background is collected and subtracted for every sample. The number of scans was 32 at a resolution of 4 cm⁻¹. NH₃-TPD was carried out on TPD/TPR PX200A from Tianjin Pengxiang Technology Co. Ltd., Tianjin, China. The sample was purged with hydrogen at 250 °C and cooled down to 100 °C, allowing the saturation adsorption of NH₃. After purging the physically adsorbed NH₃, the desorption started with a temperature increase. H₂-TPR was carried out on Micromeritics AutoChem Ⅱ 2920 Norcross, GA, USA, equipped with TCD. After being purged at 200 °C by Ar flow, the sample was heated from room temperature to 800 °C in 10% H₂/Ar. The Cu surface area, particle size, and dispersion were determined in the same instrument by N₂O titration on Cu particles at 50 °C after the sample was reduced to 250 °C, assuming 1.46 × 10¹⁹ copper atoms/m². The XPS test was conducted using the PHI, Tokyo, Japan 5000 VersaProbe system with monochromatic Al Kα radiation. The samples were reduced before testing, and the binding energy of all samples was corrected by adventitious C 1s (284.8 eV) to compensate for the surface charging of the samples. The in situ FT-IR experiments were performed on a Nicolet 6700 spectrometer (Thermofisher scientific, Waltham, MA, USA) equipped with a mercury-cadmium-telluride detector. The spectrum was collected in the range of 400–4000 cm⁻¹ with a resolution of 4 cm⁻¹, averaging 32 scans. The samples were pressed into self-supporting discs before collection and placed in the center of a closed, circulated IR cell for pretreating, degassing, and adsorption experiments.

3.3. Catalytic Performance Evaluation

The catalytic performance was evaluated by a fix-bed reactor. Before use, the catalyst was pressed, grinded, and sieved to 10–20 mesh. A total of 12 mL (approximately 6.5 g) of catalyst particle was loaded in the reactor and reduced by 100 mL/min H₂ at 250 °C for 4 h. Then, the temperature of the fix-bed was adjusted to 260 °C, and the carrier gas was switched to a N₂-H₂ mixture with a ratio of 5:1 by volume. The presence of minor hydrogen aimed to initiate the dehydrogenation reaction in the starting stage and reduce carbon deposits during the reaction. The total gas flow was set at 300 mL/min at atmospheric pressure. After reaching a steady state, DEG was injected into the catalyst bed by pump with a flow rate of 0.2 mL/min. The outlet product was collected by condensation and analyzed by gas chromatography (GC9860 Agilent, San Clara, CA, USA) equipped with a flame-ionization detector. The conversion of DEG and selectivity of PDO were calculated by the peak area normalization method, respectively. The corresponding equations are presented as follows: f and A are the relative molar calibration factor and chromatography peak area of the corresponding reactant or product.

$$\mathrm{D}\mathrm{E}\mathrm{G} \mathrm{c}\mathrm{o}\mathrm{n}\mathrm{v}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}=100\%\times \left(1-\frac{A_{DEG,out}}{A_{DEG,in}}\right)$$

$$\mathrm{P}\mathrm{D}\mathrm{O} \mathrm{s}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{i}\mathrm{t}\mathrm{y}=100\%\times \left(\frac{f_{PDO}{A}_{PDO,out}}{f_{DEG}{A}_{DEG,in}-f_{DEG}{A}_{DEG,out}}\right)$$

4. Conclusions

In this work, the effects of the structure and surface properties of CuSi-based catalysts prepared by different methods on the catalytic performance of diethylene glycol dehydrogenation to p-dioxanone were systematically investigated. The results showed that, in comparison to CuSi-dp with copper oxide as the precursor of active species, CuSi-ae with copper phyllosilicate as the precursor showed excellent dehydrogenation performance, DEG conversion, and PDO selectivity that surpassed 90% with promising stability. The favorable catalyst revealed a high specific surface area and abundant tubular pore structure, which improved the high dispersion of Cu species and reactant transfer efficiency. The ultra-small Cu particles in the reduced catalyst exposed adequate sites for dehydrogenative reactions, and the developed spatial tunnel prevented their aggregation at high temperatures. In addition, the appropriate Cu⁰/Cu⁺ ratio of the catalyst and the synergistic effect between them played an essential role in the adsorption of reactants and the stability of hemiacetal intermediates, thus effectively enhancing the dehydrogenation performance. At the same time, the low surface acid density of the catalyst also greatly inhibited the occurrence of DEG dehydration and promoted PDO yield. These results are expected to inspire some insights and provide theoretical guidance on efficient PDO production by high-performance dehydrogenation catalysts.