Unveiling the organocatalytic pathway to glycerol carbonate from glycerol and CO₂: a comprehensive study using phase behavior, operando high-pressure FTIR, and DFT insights

Abstract

The direct synthesis of glycerol carbonate from glycerol and CO₂, using acetonitrile as a dehydration agent and DBU as a catalyst, was investigated to identify the mechanism involved during this reaction. DFT modeling revealed that the most efficient pathway involves forming a C–O bond between CO₂ and glycerol's secondary alcohol. The catalyst reduces energy barriers across steps, particularly for the rate-limiting intramolecular ring closure, making the carbonyl substitution route more favorable than hydroxyl dehydration. However, acetonitrile's hydrolysis proved less effective as a dehydrating agent due to its higher energy barrier compared to direct carbonylation. While reaction parameters like CO₂ pressure and acetonitrile volume had minimal impact on yield, they influenced glycerol conversion. Higher pressure and larger acetonitrile volumes reduced glycerol conversion without changing the yield of glycerol carbonate. Elevated temperatures and prolonged reaction times promoted the formation of side products, primarily monoacetin. The reaction's complex phase behavior, driven by pressure and temperature, revealed that glycerol and acetonitrile become fully miscible only under specific conditions (155 °C, 45 bar). In situ FTIR and HPLC analyses identified kinetic profiles for glycerol carbonate, acetins, and acetamide, with ammonia and urea detected in the gas phase. Complementary DFT calculations confirmed that monoacetin formation predominantly arises from the reaction of glycerol with acetamide, a pathway characterized by the lowest energy barriers, aligning with experimental findings.

---

Introduction

Glycerol is an interesting molecule that is obtained as a by-product of biodiesel. For every 10 kg of biodiesel produced, approximately 1 kg of crude glycerol is generated, which represents about 10% w/w of the total biodiesel yield. Beyond its direct application, glycerol can be upgraded into higher-value derivatives through various conversion processes.^1,2

Among all the products that can be obtained from glycerol, glycerol carbonate has attracted considerable interest due to its bio-degradability, low toxicity and attractive physical and chemical properties such as high viscosity, good water solubility, low inflammability, and high boiling point and flash point. Glycerol carbonate is versatile in its applications, serving directly as a solvent, intermediate, or component in batteries, coatings, paints, and detergents. Additionally, it serves indirectly as a monomer precursor in the synthesis of polymers such as polyurethane and polycarbonate.^3–5 Originally, organic carbonates were obtained by reacting an alcohol with phosgene. Over time, this process was abandoned owing to the severe toxicity of the starting material, which requires stringent safety measures, and the corrosive nature of the HCl by-product, which do not align with green chemistry principles.^6,7 Consequently, recent research has focused on finding more environmentally friendly and sustainable synthetic routes. Using glycerol as the initial feedstock, the synthesis of glycerol carbonate can proceed through two main methods: indirectly either by glycerolysis of urea^8,9 or by transesterification with cyclic or acyclic carbonates,^10–12 and directly through CO or CO₂-based carbonylation reactions.^13,14 The direct conversion of glycerol with CO₂ presents a promising and sustainable route since it simultaneously addresses the surplus of glycerol and reduces CO₂ emissions. Indeed, as the main contributor to climate change and global warming, CO₂ has become a central focus of efforts to address environmental challenges.^15,16 In this context, carbon capture, utilization, and storage (CCUS) has emerged as a key strategy to decarbonize the dominant fossil fuel-based energy landscape. By converting CO₂ – an abundant, non-toxic, and inexpensive C₁ resource, CCUS technologies enable the production of high-value chemicals, providing a pathway to reduce emissions, replace petroleum-based feedstocks, and support the rapid industrial deployment of sustainable solutions.^17–22 As a consequence, its use in the synthesis of glycerol carbonate will help to capture CO₂, and as glycerol is an abundant, inexpensive, non-toxic resources regarded as waste. The direct synthesis of glycerol carbonate is a simple reaction with a high atom economy (>87%) and generates only water as the by-product.¹³ However, two major challenges limit the direct synthesis of glycerol carbonate from CO₂ and glycerol. First, the high stability of CO₂ (ΔG²⁵₀ = −396 kJ mol⁻¹) necessitates a substantial energy input. Second, the reaction's poor thermodynamic equilibrium (1.506 × 10⁻³ at 180 °C, 5 MPa) leads to low yields.^23,24 To address these limitations, developing an efficient catalyst coupled with the in situ removal of the generated water is essential for enhancing overall performance. One can mention, that the direct synthesis of GC with glycerol and carbon dioxide was performed using a non-metal catalyst 1,8-diazabicyclo[5.4.0]undéc-7-ène (DBU) associated with CH₂Br₂ in the presence of an ionic liquid bmimPF6 as a solvent.²⁵ Although, this method used mild reaction conditions (70 °C, 10 bar), and a high yield of GC was obtained (86%), the toxicity and cost of CH₂Br₂ hinders the industrial application of this method. Besides, DBU what used in a stoichiometric amount (2 equiv.) leading to a lot of salts. To date, the optimum strategy is to use a dehydrating agent. Recent reviews highlights, in detail, the various homogeneous, heterogeneous, and metal-free catalysts, as well as the different types of dehydrating agents, used for the direct carboxylation of glycerol with CO₂.^13,26n-Bu₂Sn(OMe)₂ (ref. 27 and 28) and Zn(OTf)₂/Phen^29,30 have been utilized as homogeneous catalysts. For heterogeneous catalysis, a wide variety of materials have been tested, including metal oxides,^31,32 mixed metal oxides,³³ supported metal oxides, solid oxide solutions,³⁴ hydrotalcites,³⁵ zeolites,^36,37 photo-thermal catalysts,^38,39 and biochar.⁴⁰ In contrast, metal-free catalysis has received limited attention, DBU being among the few examples reported.²⁵ Two main types of dehydrating agents have been investigated: inorganic absorbents, such as molecular sieves, Zeolite 13,²⁸ CaC₂,²⁹ CaO,³⁰ and MgCO₃,⁴¹ and organic compounds, with acetonitrile⁴² and 2-cyanopyridine⁴³ being the most common. CeO₂ combined with 2-cyanopyridine as the dehydrating agent proved to be the most efficient system. Kulal et al.⁴⁴ reported that a 10.5% Zn/CeO₂ catalyst achieved optimal performance, with a glycerol conversion of 90.4% and a glycerol carbonate yield of 89.5%. These results were obtained under conditions of 10 mL DMF, 150 °C, 5 hours, and 4 MPa CO₂ pressure. Although, the use of 2-cyanopyridine seems to be a very interesting solution, its cost and toxicity lead to the needs to find another dehydrating agent. That is the reason why acetonitrile has been employed to overcome thermodynamic limitations, providing a cost-effective and less toxic alternative to 2-cyanopyridine. It has been extensively tested with various heterogeneous catalysts, such as La₂O₃CO₃–ZnO,⁴² Al–Zn (2 : 1),³⁵ 2.3% Cu/La₂O₃,⁴⁵ and Zn : Al : La (4 : 1 : 1).³⁵ However, its application demonstrates only moderate yield and selectivity that are primarily due to the formation of side products. Herein, we want to identify the different reaction pathways that occur during the direct synthesis of GC starting from glycerol and CO₂, in the presence of acetonitrile, using a non-metal catalyst, DBU. A correlation between the effects of key reaction parameters that could influence the conversion of glycerol and the selectivity of glycerol carbonate (i.e., reaction time, temperature, CO₂ pressure, amounts of catalyst, and dehydration agent) and the product distribution between the liquid and the gas phase was investigated using in situ FTIR spectroscopy. These studies coupled with DFT calculations help to identify the mechanism and the preferred reaction pathway for the direct carboxylation of glycerol with carbon dioxide combined with the hydration of the chosen dehydration agent within a unified framework. We emphasize that to the best of our knowledge, there is no detailed investigation on the various reaction pathways that leads to different side products observed during this reaction (acetamide, acetic acid, monoacetins and diacetins, …).

Results and discussion

Investigation of reaction conditions

In this study, metal-free catalyst was used to perform the direct conversion of glycerol with CO₂ to glycerol carbonate. As reported in the literature TBD (1,5,7-triazabicyclo[4.4.0]dec-5-ene) and DBU can be of interest in this reaction. The direct carboxylation of glycerol with carbon dioxide carried out in the presence of acetonitrile as a dehydrating agent using 5 mol% of DBU or TBD as catalysts under optimized conditions give similar conversion, yield and selectivity as shown by Table 1. Based on these results we decided to pursue the study with DBU as it is commonly used in organic synthesis with notable advantages over TBD including a lower cost, ease of handling (liquid), and less hygroscopic.

Table 1 Effect of the catalyst type on the reaction of glycerol to afford glycerol carbonate. Reaction conditions: 3.68 g glycerol (40 mmol), 5 mol% catalyst, 16 mL acetonitrile, 155 °C, 45 bar, 16 h

Catalyst	Glycerol conv. (%)	GC yield (%)	GC selectivity (%)
DBU	57	22	38
TBD	59	20	35

---

In order to understand the mechanism of the direct synthesis of glycerol carbonate from glycerol and carbon dioxide using a Lewis base catalyst and acetonitrile as a reactive dehydration agent, theoretical investigations into the interaction between CO₂ and glycerol were performed. Although, prior studies have illustrated various aspects of these reaction mechanisms, they have typically focused on isolated scenarios. Thus, Ma J. et al.⁴⁶ studied the non-catalyzed reaction between ethylene glycol and CO₂, elucidating the intrinsic reactivity of vicinal diols. In contrast, another study examined the catalyzed reaction of CO₂ with propylene glycol, employing 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) where it was demonstrated that propylene carbonate-activated mechanism was identified as the more probable pathway, rather than the CO₂-activated mechanism.⁴⁷

The reactivity of acetonitrile with water has also been investigated using TBD as a catalyst. Despite these valuable contributions, no theoretical study to date has examined both reactions within a unified framework using consistent computational parameters. Moreover, these studies were performed using TBD. Here we have performed a comprehensive modelling of the catalysis by DBU involved in both the carboxylation of glycerol with CO₂ and the hydration of acetonitrile to gain deeper insights into the overall reaction system and presenting a robust energy profile for the entire process.

DFT calculations were performed for both the DBU catalyzed and uncatalyzed reactions to enable a comparative analysis. We begin by examining the direct carboxylation of glycerol with carbon dioxide. Various reaction pathways were examined that are described in details for the uncatalyzed reaction in ESI† 3. Glycerol possesses two primary and one secondary hydroxyl group, each capable of participating in the reaction. The preferred pathway involves the secondary alcohol with the carbonyl substitution mechanism was identified as the more favorable pathway. Regarding the catalyzed pathway, the secondary hydroxyl group (C₂) was initially proposed as the preferred site for glycerol carbonate formation. However, the small difference in activation barriers between the secondary (+33.6 kcal mol⁻¹) and primary hydroxyl groups (+33.8 kcal mol⁻¹) suggests that both reaction pathways are energetically comparable and may proceed concurrently (ESI† 4). The first step consists of deprotonation by DBU of the hydroxyl group which activates the alcohol function and induces angular deformation of CO₂ leading to the formation of INT1 with an activation energy of +10.8 kcal mol⁻¹ (Fig. 1a). Following this, we have identified two pathways. The first one, the hydroxyl dehydration pathway (red), which involves the proton transfer from DBUH⁺ to the oxygen of the hydroxyl group, followed by cyclization to produce glycerol carbonate, along with water elimination and DBU regeneration (G(TS2) − G(INT1) = +49.4 kcal mol⁻¹). The second one, the carbonyl substitution pathway (green) includes two additional steps. First, the cyclization occurs to form a 5-membered ring accompanied by proton transfers from DBUH⁺ and the cyclizing hydroxyl group to reach the oxygen atom of the carboxylate leading to the INT2. This cyclisation step is the rate-determining step with an energy barrier of +33.6 kcal mol⁻¹ (G(TS2) − G(INT1)). The final step corresponds to the proton transfer through the DBU, along with elimination of water and regeneration of DBU (G(TS3) − G(INT2) = +15.1 kcal mol⁻¹). Therefore, the carbonyl substitution pathway is energetically more advantageous.

Fig. 1 Gibbs free energy profile (kcal mol⁻¹) of: a) the direct carboxylation of glycerol with carbon dioxide, b) the hydration of acetonitrile.

On the other hand, the hydrolysis of acetonitrile was investigated. Both the catalyzed and uncatalyzed reactions proceed through two elementary steps. The energy profile of the uncatalyzed reaction pathway is presented in ESI† 5. In the catalyzed pathway (Fig. 1b), initially, the oxygen atom of the water molecule is positioned near the carbon forming a C–O bond, while DBU plays a role in assisting the proton transfer from the water molecule to the nitrogen of the nitrile group. This step is the rate-determining step, with an activation energy of +44.2 kcal mol⁻¹. Subsequently, DBU facilitates proton transfer from the hydroxyl group to the nitrogen, allowing the oxygen to form a double bond with the carbon and thereby form an amide (G(TS2) − G(INT1) = +13.2 kcal mol⁻¹).

Given the high energy barrier associated with the rate-determining step (TS1), an alternative reaction pathway was considered, as illustrated in Fig. 1b (blue). The formation of acetamide follows a two-step mechanism, in which two water molecules act cooperatively to lower the energy barriers to +39.7 kcal mol⁻¹ (TS1) and +5.5 kcal mol⁻¹ (TS2). The reaction mechanism is detailed in ESI† 6.

Examination of the energy profiles reveals that for glycerol carbonate synthesis to be favored, the energy barrier for acetonitrile hydration must be lower than that for glycerol carbonate hydration (i.e., the reverse reaction to reform glycerol). This is not the case, as the hydration barrier height of glycerol carbonate (less than 30 kcal mol⁻¹) appears to be more favorable than that of acetonitrile (+44.2 kcal mol⁻¹).

Therefore, from this reaction mechanism overview, we clearly put in evidence that the efficiency of the direct carboxylation of glycerol with carbon dioxide should largely depend on the effectiveness of the Lewis base DBU to catalyze both the carboxylation reaction and the hydration of acetonitrile. Following this detailed mechanistic study, in order to find the optimum conditions and the underlying thermochemical properties that governs the selectivity and yield of the DBU catalyzed synthesis of glycerol carbonate using acetonitrile as a dehydrating agent, the effect of important reaction parameters such as the amount of DBU and acetonitrile, the CO₂ pressure, the reaction time, and the temperature is reported below. Depending on the parameter investigated, upon completion of the reaction and at room temperature, we found that the mixture could be either monophasic or biphasic with various colors (ESI† 8) that indicate the presence of side products at different level of concentration, mainly monoacetin and diacetin (ESI† 9).

Temperature and carbon dioxide effect. The influence of reaction temperature, evaluated within the range of 135–175 °C, is depicted in Fig. 2. The conversion, yield, and selectivity are significantly affected by the reaction temperature. The conversion of glycerol increased progressively as the temperature increased. From 135 °C to 155 °C, there is a noticeable increase in both yield and selectivity, reaching 22% and 38%, respectively, at 155 °C. With further increases in temperature, although there is an increase of the conversion up to 74% at 175 °C, the yield and selectivity decrease significantly, reaching only 13% and 18%. At the end of the reaction, the color of the reaction mixture changed from transparent at 135 °C, where a notable amount of unreacted glycerol led to a biphasic system, to brown at 175 °C, resulting in a monophasic system (ESI† 8). This change was supported by an increase in the intensity of the peaks corresponding to monoacetin and diacetin (ESI† 9). A suitable reaction temperature for the reaction should be about 155 °C.

Fig. 2 Effect of reaction temperature. Reaction conditions: 3.68 g glycerol (40 mmol), 5 mol% DBU, 16 mL acetonitrile, 45 bar, 16 h.

The conversion of glycerol and carbon dioxide into glycerol carbonate is an equilibrium-driven process. Increasing the partial pressure of CO₂ should shift the equilibrium towards the products side, thereby favoring a higher yield of glycerol carbonate. Accordingly, the pressure of the system was varied between 30 and 106 bar (Fig. 3). As the pressure rises, there is a slight increase of the conversion between 30 and 46 bar from 52 to 57% that remains constant up to 85 bar and a slight decrease to 49% at the highest pressure of 106 bar. A similar trend is observed for the yield of glycerol carbonate that displays an optimum value between 46 and 67 bar while the selectivity remained relatively constant above 46 bar. Therefore, a pressure range of 45 bar appears to give the highest yield and selectivity and will be selected for subsequent experiments. Surprisingly, higher pressures do not favor the conversion and yield of the reaction. Therefore, in order to try to understand this behavior and rationalize the CO₂ pressure effect, we have performed an in situ FTIR investigation of the phase behavior of the ternary mixture glycerol–CO₂–acetonitrile in the temperature range between 25 and 140 °C under different CO₂ pressure (0, 10, 40, 70, and 100 bar) at the same glycerol/acetonitrile ratio than that used in the optimized reaction conditions.

Fig. 3 Effect of reaction pressure. Reaction conditions: 3.68 g glycerol (40 mmol), 5 mol% DBU, 16 mL acetonitrile, 155 °C, 16 h.

In situ FTIR investigation of the phase behavior. At room temperature, as glycerol and acetonitrile are immiscible forming a biphasic system in the liquid phase, we can expect a complex phase behavior in the presence of CO₂ at different pressures and temperatures that can explain the results obtained by varying the temperature and the CO₂ pressure. In order to determine the concentration of each species in both the liquid and the gas phase, we have used an analytical method based on the monitoring of the evolution of both phases by infrared absorption spectroscopy. The details of the method and its validation are given in detail in ESI† 10 and ESI† 11. The variation in the concentration of glycerol, acetonitrile, and CO₂ in the liquid phase (in mol L⁻¹) with the temperature at constant pressure is reported in Fig. 4. One, two, or three phases are observed based on the pressure of the system and temperature.

Fig. 4 Concentration (in mol L⁻¹) of CO₂, acetonitrile, and glycerol in the liquid phase at fixed pressure and various temperatures.

For CO₂. Whatever the pressure, we observe that the concentration of CO₂ decreases as temperature rises. At the lowest pressure of 10 bar and a high temperature of 100–140 °C, there is almost no CO₂ present in the liquid phase, which explains the lower glycerol carbonate yield observed at a low pressure (Fig. 3). Then, increasing the pressure from 10 to 100 bar results in a systematic higher concentration of CO₂ in the liquid phase. At the highest temperature of 140 °C, a concentration of about 1 mol L⁻¹ is measured for a pressure of 40 bar and is increased up to about 2.5 mol L⁻¹ at 100 bar. We notice that at 100 bar, data are only available from 80 °C as below this temperature, acetonitrile and CO₂ are fully miscible and therefore, the concentration of CO₂ only depends on the initial volume of each component introduced in the cell.
Acetonitrile. Whatever the pressure, we observe that the concentration of acetonitrile decreases as temperature rises. We emphasize that for the measurements without CO₂, the pressure is equal to the vapor pressure of acetonitrile (around 10 bar at 140 °C). Then, increasing the pressure from 10 to 100 bar results in a systematic decrease in the concentration of acetonitrile which indicates a volume expansion of the liquid phase that is correlated with the higher CO₂ sorption. At the highest temperature of 140 °C, a concentration between 7 and 10 mol L⁻¹ is measured for pressures between 10 and 100 bar.
For glycerol. Whatever the pressure, for temperatures below 80 °C, glycerol is nearly immiscible in acetonitrile. Increasing the temperature beyond 80 °C leads to an increase in the miscibility of glycerol in acetonitrile that is particularly pronounced without CO₂ where the initial concentration of glycerol introduced in the mixture is reached. However, increasing the pressure of CO₂ lowers both the concentration and the solubility of glycerol in acetonitrile. Indeed, introducing 10 bar of CO₂ expands the liquid phase, leading to a decrease in glycerol concentration at the highest temperature. At 40 bar, the temperature required for glycerol to be completely miscible in acetonitrile shifts from 100 °C to 120 °C. At pressures of 70 bar and 100 bar, full miscibility of glycerol in acetonitrile was not achieved below 140 °C. In other words, high temperature of at least 140 °C at a pressure of 70 bar is required to reach the initial concentration of glycerol introduced in the mixture. Finally, at 140 °C and a pressure of 100 bar, glycerol is not fully soluble in acetonitrile and therefore some liquid glycerol should remain at the bottom of the cell that couldn't react with CO₂ in the homogeneous glycerol/acetonitrile/CO₂ liquid phase. Therefore, a lower quantity of glycerol carbonate is expected to be formed at pressures higher than 100 bar thus explaining the lower conversion of glycerol reported in Fig. 4 at the highest pressure of 106 bar.

By the same token, the gas phase has been investigated in the same temperature and pressure range and the main results are reported in ESI† 12.

To summarize the previous graphs, Fig. 5 illustrates the phase behavior of our mixture at various temperatures and pressures. Depending on these conditions, the mixture can have one, two, or three phases. Under our optimized reaction conditions, we expect that two phases are present, with glycerol and acetonitrile being completely miscible under the glycerol/acetonitrile ratio used in our study and we have put in evidence that a minimum concentration of CO₂ of about 1 mol L⁻¹ is present in the liquid phase to react with glycerol to produce glycerol carbonate.

Fig. 5 Schematic representation of the evolution of glycerol–acetonitrile–CO₂ mixture at different pressures and temperatures.

Catalyst amount. The direct carboxylation of glycerol with carbon dioxide was carried out using DBU as a catalyst in the presence of 16 ml of acetonitrile with 3.68 g of glycerol at the reaction temperature of 155 °C under a pressure of 45 bar of CO₂ during 16 h. To assess the impact of DBU concentration on the reaction's performance, the amount was varied from 2.5 mol% to 30 mol% with respect to glycerol. As shown in Fig. 6, adding 2.5 mol% to 5 mol% of DBU slightly increased the conversion of glycerol from 45% to 57% and the yield of glycerol carbonate from 17% to 22%, while the selectivity remained constant at around 38%. Increasing the DBU concentration from 10 mol% to 30 mol% resulted in a notable reduction in the yield and selectivity of glycerol carbonate, which dropped to 14% and 23%, respectively. However, the conversion of glycerol remained constant, reaching a maximum of 62%, regardless of the catalyst amount added. For further experiments, 5 mol% of DBU was utilized as the reference concentration.

Fig. 6 Effect of catalyst amount. Reaction conditions: 3.68 g glycerol (40 mmol), 16 mL acetonitrile, 155 °C, 45 bar, 16 h.

One can mention that DBU can be hydrolyzed in the presence of water.⁴⁸ When an experiment was performed using water (40 mmol) instead of glycerol, keeping all the parameters constant, the product of the hydrolysis of DBU was observed by GC-MS analysis. However, during the reaction the amount of water was low and was used by acetonitrile in large excess to produce acetamide. Moreover, it is mentioned that this hydrolytic instability mainly occurs in very basic solution with a pH above 11.6. In our investigation, in the presence of water and CO₂ in an organic solvent such acetonitrile, DBU will react with CO₂ and water to form the carbonate salt DBUH⁺ HCO₃⁻ at room temperature.⁴⁹ Then, upon the temperature increase, this equilibrated reaction will go forward to the molecular species DBU, CO₂ and H₂O. After a decrease of temperature, this reaction was found to be fully reversible showing no hydrolysis reaction of DBU.⁵⁰ In addition, the pH of the solution is expected to be slightly acidic in the presence of water and CO₂. Therefore, the hydrolysis of DBU under our experimental conditions is not a favored reaction.

Acetonitrile amount. The amount of acetonitrile was then studied (Fig. 7). The results of varying the amount of acetonitrile from 4 mL to 20 mL has a significant impact on the conversion of glycerol. It stays relatively constant at 65% from 4 mL to 8 mL. However, it gradually decreases to 52% as the volume reaches 20 mL probably due to dilution effect. The yield of glycerol carbonate increases gradually from 4 mL to 16 mL, rising from 15% to 22% and then decreases to 19% with 20 mL of acetonitrile. Hence, the reaction time was also studied for 8 ml to 20 mL of acetonitrile (Fig. S7.5†). It was shown that even after 36 hours of reaction the conversion of glycerol was lower for 20 mL (65%) of acetonitrile than for 16 mL (70%) and 8 mL (82%). Consequently, a volume 16 mL of acetonitrile was selected as the optimum value for further experimentation.

Fig. 7 Effect of acetonitrile volume. Reaction conditions: 3.68 g glycerol (40 mmol), 5 mol% DBU, 155 °C, 45 bar, 16 h.

Reaction time. The influence of reaction time was studied for a range of times from 5 h to 62 h. The evolution of glycerol conversion and glycerol carbonate yield with reaction time is shown in Fig. 8. An increase in reaction time enhances the conversion of glycerol, with the maximum yield and selectivity of glycerol carbonate reaching 22% and 38% after 16 h. Logically, a longer reaction time up to 62 h increases the conversion up to 83%. However, extending the reaction time beyond 16 h leads to a gradual decline in yield of glycerol carbonate showing that it is not stable upon time. Indeed, a prolonged reaction time promotes the formation of side reactions, as confirmed by HPLC analysis where monoacetin and diacetin are the main side products (ESI† 9). At the end of the reaction, and under ambient temperature and pressure conditions, the reaction mixture is biphasic from 5 h to 16 h and monophasic from 24 h to 62 h (ESI† 8). Moreover, as the reaction time increases, the reaction mixture changes color from light yellow to brown indicating qualitatively the presence of side products (ESI† 8). The reaction time was also studied at 145 °C and 175 °C, and a comparison at iso-conversion of glycerol (40%) (Fig. S7.4†) led to a similar yield to glycerol carbonate (around 17%) showing that a compromise between the temperature and the reaction time should be selected. Consequently, 16 h were selected for further experimentation.

Fig. 8 Effect of reaction time. Reaction conditions: 3.68 g glycerol (40 mmol), 5 mol% DBU, 16 mL acetonitrile, 155 °C, 45 bar.

Operando FTIR investigation of the reaction kinetic

By HPLC analysis, it was shown that monoacetin is the principal side product, along with minor quantities of diacetin and some other unidentified by-products, under optimized reaction conditions (16 h, 45 bar, 155 °C (ESI† 9)). However, acetamide or acetic acid were not observed. Therefore, operando FTIR spectroscopy was employed to further identify potential intermediates and side products. After confirming that glycerol and acetonitrile are fully miscible under our reaction conditions indicating a biphasic system (liquid phase where glycerol/acetonitrile are completely miscible and the gas phase) we have performed two independent reactions to monitor the evolution over time of the presence of various products in the liquid and gas phases.

Liquid phase. Fig. S12† shows the operando FTIR measurements of the liquid phase and illustrates the change in the IR spectra during the reaction (ESI† 13). The spectral range of main interest between 1650 and 1900 cm⁻¹ is reported in Fig. 9 and allows to monitor the appearance of contribution associated with the C=O stretching mode of various species, mainly, the peak at 1809 cm⁻¹ for glycerol carbonate, 1738 cm⁻¹ for acetins, and 1674 cm⁻¹ for acetamide. The intensities of the bands corresponding to glycerol carbonate, acetins, and acetamide rise over the reaction time. From the evolution as a function of time of the intensity of these bands as reported in Fig. 10, we observe that the products are produced at distinct reaction rates over time. Acetins are initially formed at a rapid rate before stabilizing after 7.5 h. Both glycerol carbonate and acetamide are produced at a steady rate, with the glycerol carbonate peak reaching a maximum value after about 20 h. Analysis of the reaction spectra reveals the absence of clear bands characteristic of acetic acid, suggesting that acetamide is not further hydrolyzed to acetic acid. Instead, it is either rapidly converted to acetins or transferred to the gas phase.

Fig. 9 Evolution as a function of time of the IR spectra of the direct carboxylation of glycerol in the liquid phase. Reaction conditions: 1.23 g glycerol (13.33 mmol), 5 mol% DBU, 2.67 mL acetonitrile, 145 °C, 45 bar.

Fig. 10 The evolution of absorbance for glycerol carbonate, acetins, acetamide, and glycerol with time in the liquid phase.

Gas phase. Fig. S13† shows the operando FTIR measurements of the gas phase and illustrates the change in the IR spectra during the reaction (ESI† 14). Upon examining the overall spectra, no characteristic bands of glycerol, glycerol carbonate, acetins, or acetamide were observed, indicating they remained in the liquid phase. Nevertheless, the gas phase contains a given amount of acetonitrile (ESI† 12) and a high concentration of CO₂, as indicated by the characteristic saturated bands of CO₂. Besides that, two distinct weak bands at 3332 cm⁻¹ and 1729 cm⁻¹ are detected and found to increase over time. The band at 3332 cm⁻¹ is characteristic of ammonia whereas the peak at 1729 cm⁻¹ could be assigned to urea. Plotting the kinetics profiles of both compounds (Fig. 11) indicates a rapid increase in the intensity of the ammonia, reaching its maximum after 10 hours, and then decreasing slowly thereafter, indicating that ammonia either moves into the liquid phase or undergoes reactions in both phases to produce nitrogen-containing compounds. Conversely, urea shows a continual increasing trend that is probably related to the continuous reaction between CO₂ and ammonia in the gas phase.

Fig. 11 The evolution of absorbance for ammonia and urea with time in the gas phase.

Investigation of side reactions

Following the hydrolysis of acetonitrile that results in the formation of acetamide, a number of side reactions can occur from the reaction between acetamide and the different components present in the reaction mixture that can produce a number of different species (see ESI† 15). Our FTIR operando experiments revealed clear characteristic peaks of acetamide and monoacetin in the liquid phase, along with ammonia and urea in the gas phase. Conversely, no distinct peaks for acetic acid were identified in both phases. Further DFT calculations were carried out to improve our understanding of the side reactions such as the formation of acetic acid from the hydration of acetamide and the formation of monoacetin from acetamide or acetic acid.

The energy profiles presented in Fig. 12 provide insights into the catalyzed reaction of both the hydration of acetamide (a) and the formation of monoacetin from acetamide or acetic acid (b). The uncatalyzed reaction of both reactions are presented in ESI† 7. The catalyzed hydration of acetamide, reported in Fig. 12a, consists of two steps. The first step, which is the rate-determining step, is the addition of the water molecule on the carbon atom of the carbonyl group of acetamide while DBU abstracts a proton of the water molecule. The second step consists of a proton transfer from DBUH⁺ to the NH₂ group, leading to the formation of acetic acid and the release of ammonia. The overall energy barrier of the hydration of acetamide is calculated to be +32.4 kcal mol⁻¹. The two-steps mechanism for the synthesis of monoacetin (Fig. 12b) from acetamide or acetic acid is identical, ending with the release of either ammonia or water, respectively. The first step is the deprotonation by DBU of the hydroxyl group of glycerol followed by the formation of a C–O bond to form the INT1. Whether monoacetin formation from acetamide or acetic acid, the 2 reactions present their highest transition states at similar energy values (G(TS1) = +28.1 kcal mol⁻¹ for acetamide and G(TS2) = +27.8 kcal mol⁻¹ for acetic acid), suggesting that these 2 reactions are equiprobable.

Fig. 12 Gibbs free energy (kcal mol⁻¹) profiles of: a) the hydration of acetamide to form acetic acid, b) the formation of monoacetin from acetamide or acetic acid.

Given that the energy barrier for acetonitrile hydration exceeds that of glycerol carbonation, the kinetics of glycerol carbonate formation are expected to be slow—an observation supported by experimental data showing only 4% glycerol carbonate formation after 5 h. However, under elevated temperature conditions (155 °C), the water generated during the reaction can hydrolyze acetonitrile, leading to the production of acetamide. Acetamide, once formed, may proceed through two potential pathways: it can react with unreacted glycerol to yield monoacetin, or it can be hydrolyzed again in the presence of another water molecule. During operando FTIR measurements, acetic acid was not detected in either phase. This can be attributed to the low energy barrier of the initial acetic acid and glycerol reaction (+21.8 kcal mol⁻¹), suggesting that once acetic acid is formed, it rapidly reacts with glycerol to generate a relatively stable intermediate (+5.1 kcal mol⁻¹). In the case of the reaction of acetamide with glycerol, the first reaction step requires a higher energy (+28.1 kcal mol⁻¹), resulting in a slow reaction rate. However, this reaction proceeds faster than the hydration of acetamide with water (+32.4 kcal mol⁻¹), which leads to acetic acid formation. Overall, both reaction pathways (step 1 + step 2) are nearly equally probable, as each involves a transition state (TS) with an energy of approximately +28 kcal mol⁻¹, though for acetic acid, the second step being rate-determining. A summary of the possible main side reaction pathways is outlined in Fig. 13.

Fig. 13 Possible reaction pathway for the main side products (monoacetin and acetic acid).

Experimental

Materials

The chemicals were obtained from commercial suppliers and utilized without any further purification or special storage conditions. Glycerol (GL) (Acros Organics, 99+%), acetonitrile (Carlo Erba, HPLC Plus Gradient), 1,8-diazabicyclo[5.4. 0]undec-7-ene (DBU) (Sigma Aldrich, 98.0%), 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) (Sigma-Aldrich, 98%). Carbon dioxide (purity 99.985%) was purchased from Air Liquide (France).

Methods

General procedure for single batch reactions. A typical procedure for the synthesis of glycerol carbonate from glycerol and CO₂ is as follow: under air, glycerol (40 mmol, 3.68 g), DBU (5 mol%, 2 mmol, 0.30 g), and acetonitrile (16 ml, 12.58 g) were filled into a glass reaction inlet equipped with a magnetic stirring bar and inserted into a stainless-steel batch reactor of 75 cm³ (Multiple Reactor System Series 5000 by Parr Instrument Company) (see ESI† 1). The reactor was sealed, and then the air content was quickly purged by flushing CO₂ three times. At room temperature, a desired amount of CO₂ was introduced in the autoclave at a given pressure and the reactor was weighted before and after the CO₂ addition to determine exactly of the number of moles of CO₂. Then the autoclave was heated to the desired reaction temperature (typically, 428.15 K, and the heating time before reaching the desired temperature is about 30 min). The mixture in the reactor was constantly stirred during the reaction by a magnetic stirrer and the stirring rate was around 250 rounds per minute. At the end of the reaction time, the reactor was cooled in a water bath to room temperature, then it was slowly depressurized. Ultrapure water was added to the liquid phase as a washing solvent and to homogenize the various phases (25 mL volumetric flask).

Quantitative analysis on the amount of glycerol and glycerol carbonate in a sample from the reaction output was performed using HPLC-RID (LC-20AD Shimadzu-Japan) equipped with autosampler assembly SIL-10A and a controller CBM 20A. The products were separated on Aminex HPX-87H Ion Exclusion Column (300 × 7.8 mm) using 745 μL of H₂SO₄ 98% in 2 L of ultrapure water as eluent, the flow rate was kept 0.8 mL min⁻¹ at 35 °C. The refractive index detector (model no. RID-10A) was used for the detection and calibration of compounds. Products are quantified via the external calibration of commercially available standards. For each compound, conversion (X) and yields (Y) or selectivity (S) are calculated and expressed in percentage via the following equations with n₀, the initial number of moles of glycerol, and n(name), the number of moles of the given compound after reaction:

General procedure for operando experiments by infrared spectroscopy. The kinetic of the reaction was monitored using high pressure operando infrared absorption spectroscopy by coupling a homemade high-pressure/high-temperature cell with an FTIR spectrometer (see ESI† 2) and a typical procedure is as follow: glycerol (13.33 mmol, 1.28 g, 0.982 mL) was first introduced in the cap with a magnetic stirrer and placed at the bottom of the cell. A solution containing 5 mol% of DBU (0.67 mmol, 0.10 g, 0.10 mL) in 2.67 mL of acetonitrile was prepared and subsequently transferred into the cell. The setup is sealed utilizing the CO₂ feed line to enable pressurization. The cell was pressurized at room temperature with 40 bar of CO₂. The reaction mixture was heated to 145 °C and the pressure was adjusted to the desired pressure when the equilibrium temperature was reached. Spectra were recorded every 10 minutes for a duration of about 20 hours.

General procedure for phase behavior investigation by in situ infrared spectroscopy. In order to determine quantitatively the concentrations of various species in the liquid and gas phases, the monitoring of the evolution of both phases at a constant pressure while varying the temperature was performed using the same high-pressure set-up coupled with an FTIR spectrometer as described in ESI† 2. To simulate the phase behavior of our reaction system, we kept the same glycerol/acetonitrile ratio (1.05/5.71 ml) than that used during the optimization of the reaction conditions. The spectra were initially collected at room temperature under CO₂ pressure (0, 10, 40, 70, and 100 bar). Subsequently, the temperature was raised by 20 °C till it reached 140 °C. The system was maintained at constant pressure and constant temperature for a duration of 15 to 30 minutes. During the stabilization of the operating conditions, consecutive spectra were recorded every 5 min. The equilibrium was considered to be achieved when no changes in the spectral bands were noticed. The equilibrium was systematically reached for a period of time lower than 30 min. It was at the same time possible to look inside the cell and to visualize the expansion of the liquid phase during the increase of the pressure. The mixture was constantly homogenized during the experiment using a magnetically driven stirrer disposed at the bottom of the cell.

Computational details. Preliminary calculations of equilibrium structures were performed using a semi-empirical model (AM1-D3H4) to determine the most stable conformations. These semi-empirical calculations were performed using the AMPAC software. The CHAIN algorithm was used for locating intermediates and transition states along the reaction path. The lowest energy structures obtained at the AM1-D3H4 level were further investigated using the density functional theory method (DFT) implemented in the Gaussian 16 package. DFT calculations of geometries, energies, and vibrational frequencies reported in this paper were carried out with the M06-2X functional using the 6-311G++(d,p) basis set. All frequencies of each structure have also been calculated to verify the presence of a single imaginary frequency for transition states and the absence of an imaginary frequency for ground states. The intrinsic reaction coordinate (IRC) method has been used to verify that the obtained transition states were effectively connected to the desired minima. For all catalysts, a wide range of possible configurations and interactions have been modelled and the more stable of them are reported in this work. To consider entropic effects, the energies mentioned in this study correspond to the Gibbs free energy (ΔG).

Conclusion

The direct synthesis of glycerol carbonate from glycerol and carbon dioxide in the presence of acetonitrile as a reactive dehydration system and an organocatalyst (DBU) was studied. Full modeling of the catalysis of the condensation reaction of glycerol and CO₂ in combination with the hydrolysis of acetonitrile was intended to have a better understanding of the overall reaction system. DFT calculations for the catalyzed reaction were conducted and compared to the uncatalyzed reaction pathway. Forming a C–O bond between carbon dioxide and the secondary alcohol of glycerol is the preferred route for synthesizing glycerol carbonate. The catalyst reduces the energy barriers across different steps, with the intramolecular ring closure being the rate-limiting step. The pathway involving the carbonyl substitution is more favorable compared to the hydroxyl dehydration route. The hydrolysis of acetonitrile exhibited a higher energy barrier (+44.2 kcal mol⁻¹) than the direct carbonylation of glycerol with CO₂ (+33.6 kcal mol⁻¹), explaining its lower efficiency as a dehydrating agent.

The influence of various reaction parameters was investigated. Variations in CO₂ pressure and acetonitrile volume did not impact the yield of glycerol carbonate. However, both high pressure and a large volume of acetonitrile decreased glycerol conversion, which improved the selectivity of glycerol carbonate. The formation of side products is promoted by both high temperatures and extended reaction times, with monoacetin being the main side product.

Furthermore, complex phase behavior was observed for the reaction. Starting from a biphasic glycerol/acetonitrile system, it depends on the reaction conditions, the reaction mixture at the end of the reaction could be monophasic or biphasic. The phase behavior varies with pressure and temperature, resulting in one, two, or three distinct phases. The high solubility of CO₂ in acetonitrile results in a notable expansion of the acetonitrile (CO₂ expanded liquids (CXL)). It was noted that at a low pressure of 10 bar, nearly no CO₂ is present in the liquid phase. Regardless of CO₂ pressure, glycerol remains nearly immiscible with acetonitrile at temperatures below 80 °C. As CO₂ pressure rises, the miscibility of glycerol decreases at higher temperatures. Here, we confirm that under our reaction conditions of 155 °C and 45 bar, we observe two phases where glycerol and acetonitrile are fully miscible (in the liquid phase). These findings are crucial for conducting an accurate operando FTIR spectroscopy study of the reaction in transmission mode. The operando FTIR spectroscopy monitoring served as a supplementary analytical technique to HPLC to identify potential intermediates and side products. Besides detecting acetins through HPLC, we were able to outline the kinetic profiles of our primary product, glycerol carbonate, as well as acetins and acetamide in the liquid phase, while ammonia and urea were monitored in the gas phase. Interestingly, acetic acid was absent in both the liquid and gas phases.

Complementary DFT calculations were performed to better understand the evolution of side products. Results obtained are aligned with our experimental observations. The reaction of acetamide with glycerol (+28.1 kcal mol⁻¹) is found to be faster than acetamide hydration (+32.4 kcal mol⁻¹). During operando FTIR measurements, acetic acid was not detected in either phase, as it rapidly reacts with glycerol due to a low energy barrier (+21.8 kcal mol⁻¹). Overall, the formation of monoacetin either from acetamide or acetic acid are similarly probable, as both involve a transition state with an energy of approximately +28 kcal mol⁻¹.

In summary, given the limited water produced during the reaction and the small differences in energy barriers across various side reactions, the principal source of monoacetin is the reaction of glycerol with acetamide.

Data availability

The data will be available upon request.

Author contributions

Taha Amine Chibane makes all the experiments and the original draft of the paper. Raphaël Méreau, Thierry Tassaing and Karine De Oliveira Vigier conceived the idea of the manuscript and designed the experiments. They reviewed and edit the original draft. Raphaël Méreau supervise the modeling part, Thierry Tassaing the spectroscopy and Karine De Oliveira Vigier, the catalytic test and the analysis of the products.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge the Region Nouvelle Aquitaine and the University of Poitiers for their financial support for the funding of the PhD of T. A. Chibane. They are also grateful to the GDR 2035 SolvATE and to the INCREASE Federation. The authors acknowledge financial support from the European Union (ERDF) and Région Nouvelle Aquitaine. Computer time was provided by the Pôle Modélisation HPC facilities of the Institut des Sciences Moléculaires UMR 5255 CNRS – Université de Bordeaux, co-funded by the Nouvelle Aquitaine region. This work pertains to the French government program “Investissements d'Avenir” (EUR INTREE, reference ANR-18-EURE-0010).

References

1. N. Yadav, G. Yadav and M. Ahmaruzzaman, Ind. Crops Prod., 2024, 210, 117999.
2. K. Mohan, S. Pai, B. S. Reghunath and D. Pinheiro, ChemistrySelect, 2023, 8, e202204501.
3. M. Checa, S. Nogales-Delgado, V. Montes and J. M. Encinar, Catalysts, 2020, 10, 1279.
4. S. Christy, A. Noschese, M. Lomeli-Rodriguez, N. Greeves and J. A. Lopez-Sanchez, Curr. Opin. Green Sustainable Chem., 2018, 14, 99–107.
5. J. R. Ochoa-Gomez, O. Gomez-Jimenez-Aberasturi, C. Ramirez-Lopez and M. Belsue, Org. Process Res. Dev., 2012, 16, 389–399.
6. D. Procopio and M. L. Di Gioia, Catalysts, 2022, 12, 23.
7. P. Anastas and N. Eghbali, Chem. Soc. Rev., 2010, 39, 301–312.
8. M. Aresta, A. Dibenedetto, F. Nocito and C. Ferragina, J. Catal., 2009, 268, 106–114.
9. Y. Q. Li, H. M. Liu, Z. Zheng, Z. Fu, D. H. He and Q. J. Zhang, Ind. Eng. Chem. Res., 2022, 61, 5698–5711.
10. W. K. Teng, G. C. Ngoh, R. Yusoff and M. K. Aroua, Energy Convers. Manage., 2014, 88, 484–497.
11. Y. Ji, Catalysts, 2019, 9, 581.
12. A. O. Esan, A. D. Adeyemi and S. Ganesan, J. Cleaner Prod., 2020, 257, 120561.
13. N. Rozulan, S. A. Halim, N. Razali and S. S. Lam, Biomass Convers. Biorefin., 2022, 12, 4665–4682.
14. A. Kojcnovic, B. Likozar and M. Grilc, J. CO2 Util., 2022, 66, 102250.
15. V. Masson-Delmotte, P. Zhai, S. Pirani, C. Connors, S. Péan, N. Berger, Y. Caud, L. Chen, M. Goldfarb and P. M. Scheel Monteiro, Ipcc, 2021: Summary for policymakers, in Climate change 2021: The physical science basis Posted: 2021, Cambridge University Press, 2021, SPM-9.
16. S. I. Seneviratne, M. G. Donat, A. J. Pitman, R. Knutti and R. L. Wilby, Nature, 2016, 529, 477–483.
17. W. L. Gao, S. Y. Liang, R. J. Wang, Q. Jiang, Y. Zhang, Q. W. Zheng, B. Q. Xie, C. Y. Toe, X. C. Zhu, J. Y. Wang, L. Huang, Y. S. Gao, Z. Wang, C. Jo, Q. Wang, L. D. Wang, Y. F. Liu, B. Louis, J. Scott, A. C. Roger, R. Amal, H. Heh and S. E. Park, Chem. Soc. Rev., 2020, 49, 8584–8686.
18. M. Aresta and A. Dibenedetto, Dalton Trans., 2007, 2975–2992,  10.1039/b700658f.
19. T. Sakakura, J. C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365–2387.
20. M. Mikkelsen, M. Jorgensen and F. C. Krebs, Energy Environ. Sci., 2010, 3, 43–81.
21. D. Y. C. Leung, G. Caramanna and M. M. Maroto-Valer, Renewable Sustainable Energy Rev., 2014, 39, 426–443.
22. W. Wang, S. P. Wang, X. B. Ma and J. L. Gong, Chem. Soc. Rev., 2011, 40, 3703–3727.
23. J. Artz, T. E. Müller, K. Thenert, J. Kleinekorte, R. Meys, A. Sternberg, A. Bardow and W. Leitner, Chem. Rev., 2018, 118, 434–504.
24. J. B. Li and T. Wang, J. Chem. Thermodyn., 2011, 43, 731–736.
25. Y. N. Lim, C. Lee and H. Y. Jang, Eur. J. Org. Chem., 2014, 2014, 1823–1826.
26. P. Inrirai, J. Keogh, A. Centeno-Pedrazo, N. Artioli and H. Manyar, J. CO2 Util., 2024, 80, 102693.
27. M. Aresta, A. Dibenedetto, F. Nocito and C. Pastore, J. Mol. Catal. A: Chem., 2006, 257, 149–153.
28. J. George, Y. Patel, S. M. Pillai and P. Munshi, J. Mol. Catal. A: Chem., 2009, 304, 1–7.
29. Q. Zhang, H. Y. Yuan, X. T. Lin, N. Fukaya, T. Fujitani, K. Sato and J. C. Choi, Green Chem., 2020, 22, 4231–4239.
30. K. Takeuchi, K. Matsumoto, N. Fukaya, K. Sato and J. C. Choi, Asian J. Org. Chem., 2022, 11, e202200212.
31. L. P. Ozorio and C. J. A. Mota, ChemPhysChem, 2017, 18, 3260–3265.
32. Y. H. Ke, H. H. Xu, X. Wang, H. Liu and H. Yuan, J. CO2 Util., 2024, 83, 102813.
33. P. H. Pandey and H. S. Pawar, ChemistrySelect, 2022, 7, e202104264.
34. J. X. Liu, Y. J. Li, H. M. Liu and D. H. He, Biomass Bioenergy, 2018, 118, 74–83.
35. H. G. Li, X. Jiao, L. Li, N. Zhao, F. K. Xiao, W. Wei, Y. H. Sun and B. S. Zhang, Catal. Sci. Technol., 2015, 5, 989–1005.
36. G. Özçakir and C. Açikgöz, Res. Chem. Intermed., 2024, 50, 5655–5678.
37. Z. X. Gao, M. Xiang, M. Y. He, W. Y. Zhou, J. Y. Chen, J. M. Lu, Z. Y. Wu and Y. Q. Su, Molecules, 2023, 28, 9200031.
38. Y. J. Li, H. M. Liu, L. Ma, J. X. Liu and D. H. He, Acta Phys.-Chim. Sin., 2024, 40, 2308005.
39. J. X. Liu, Y. J. Li, H. M. Liu and D. H. He, Appl. Catal., B, 2019, 244, 836–843.
40. C. Collett, O. Mašek, N. Razali and J. McGregor, Catalysts, 2020, 10, 1067.
41. C. C. Hu, C. W. Chang, M. Yoshida and K. H. Wang, J. Mater. Chem. A, 2021, 9, 7048–7058.
42. H. G. Li, D. Z. Gao, P. Gao, F. Wang, N. Zhao, F. K. Xiao, W. Wei and Y. H. Sun, Catal. Sci. Technol., 2013, 3, 2801–2809.
43. M. Honda, M. Tamura, K. Nakao, K. Suzuki, Y. Nakagawa and K. Tomishige, ACS Catal., 2014, 4, 1893–1896.
44. N. Kulal, R. Vetrivel, N. S. G. Krishna and G. V. Shanbhag, ACS Appl. Nano Mater., 2021, 4, 4388–4397.
45. J. Zhang and D. H. He, J. Colloid Interface Sci., 2014, 419, 31–38.
46. J. Ma, J. L. Liu, Z. F. Zhang and B. X. Han, Comput. Theor. Chem., 2012, 992, 103–109.
47. J. Ma, X. L. Zhang, N. Zhao, A. S. N. Al-Arifi, T. Aouak, Z. A. Al-Othman, F. K. Xiao, W. Wei and Y. H. Sun, J. Mol. Catal. A: Chem., 2010, 315, 76–81.
48. J. Ma, X. L. Zhang, N. Zhao, A. S. N. Al-Arifi, T. Aouak, Z. A. Al-Othman, F. K. Xiao, W. Wei and Y. H. Sun, Org. Process Res. Dev., 2019, 23, 1860–1871.
49. D. J. Heldebrant, P. G. Jessop, C. A. Thomas, C. A. Eckert and C. L. Liotta, J. Org. Chem., 2005, 70, 5335–5338.
50. K. N. Onwukamike, T. Tassaing, S. Grelier, E. Grau, H. Cramail and M. A. R. Meier, ACS Sustainable Chem. Eng., 2018, 6, 1496–1503.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5cy00193e

---