Utilization of CO₂ in micropacked bed reactors for enhanced synthesis of cyclic carbonates

Abstract

Conversion of carbon dioxide (CO₂) to cyclic carbonates is important for carbon capture and utilization and enables the production of valuable, non-toxic chemicals from a climate-relevant greenhouse gas. This paper reports a catalytic strategy for synthesizing cyclic carbonates using commercial solid mesoporous silica supports with controlled particle sizes. Tripropylammonium groups were grafted onto the silica surface to activate it for CO₂ conversion. Batch reactor experiments confirmed that the functionalized silica catalyzes the reaction under milder temperature and pressure conditions than previously reported. Importantly, this strategy effectively suppresses polymer by-product formation, enabling selective cyclic carbonate production while offering potential applications for controlled polymer synthesis through reverse optimization. A subsequent gas-phase cycloaddition reaction was conducted in a micropacked bed reactor. The reactor performance was assessed using the chemical regime method to separate intrinsic kinetics from mass transfer effects. Compared to conventional fixed-bed reactors, the reaction times were significantly reduced, from several days in batch systems to ∼1 minute in centimetric reactors and <1 second in microreactors. These findings support the potential for commercial-scale implementation.

---

Introduction

Beyond mere capture and storage, the utilization of carbon dioxide (CO₂) offers the opportunity to implement carbon-neutral processes that can yield significant economic benefits for industrial CO₂ emitters, especially in countries with carbon taxes.^1–3 Indeed, CO₂ can be used as a carbon building block and reincorporated into various commodity chemicals through catalytic pathways, thus replacing carbon sources that usually originate from fossil feedstock.^4–10 The catalytic coupling reaction between CO₂ and epoxides is one of the pathways that is currently leveraged by commercial CO₂ capture and conversion technologies.^11,12 This reaction pathway enables the production of cyclic carbonates, which have applications as sustainable alternatives in polycarbonate monomers, green solvents (particularly polar aprotic solvents), and electrolytes in lithium-ion batteries and serve as organic synthesis intermediates via base-catalyzed ring-opening reactions.^13–19 Although several homogeneous catalysts have shown impressive turnover frequencies for the formation of cyclic carbonates from epoxides and CO₂,^20–24 the operating conditions that are typically required (high temperature, high pressure, addition of co-catalysts and solvent) make the integration of these processes difficult on an industrial scale. Thus, the development of heterogeneous catalysts that can operate at low temperature and low pressure without co-catalysts is highly attractive. To this end, several classes of catalysts have been developed, including metal oxides,²⁵ metal–organic framework-supported ionic liquids,^26,27 transition metal complexes, and organic salts.^28,29

Inspired by Werner et al.,³⁰ we reported that ordered silica mesoporous materials containing grafted ammonium moieties were efficient catalysts for the formation of cyclic carbonates from CO₂ at ambient pressure.³¹ The high catalytic activity value of these heterogeneous materials, far beyond what Henry's law predicts, was attributed to the oversolubility of CO₂ within the nanometric pores of the material.^32,33 Indeed, a 7.5-fold rate enhancement of the catalytic activity was observed in grafted MCM-41, as compared to homogeneous catalysis in bulk solution under similar conditions.³⁴ Monte Carlo simulations revealed that this improvement was due to a lower density of the absorbed solvent in the nanopore, a phenomenon that is optimal with a pore size of approximately 4 to 6 nm.³⁴

Although ordered mesoporous silicas such as MCM-41 and SBA-15 exhibit high surface area, often close to 1000 m² g⁻¹, they are expensive to produce compared to irregular silica. If the hypothesis is correct that the enhanced catalytic activity observed for the formation of cyclic carbonate using epoxides and CO₂ is caused by the nanoporosity of the catalytic support, commercial silica materials with pores in the optimal range of 4 to 6 nm should also exhibit enhanced catalytic activity. For validation, we investigated the catalytic activity of commercial irregular silica materials SiliaFlash Silica Gel, both in batch and flow processes, as reported herein.

The industrial relevance of the present results lies in process intensification rather than direct geometric scale-up of a single microreactor. Micro-packed beds are attractive because their small characteristic dimensions enhance heat and mass transfer and may allow access to reaction conditions that are closer to intrinsic kinetic behaviour than in larger reactors.³⁵ In practice, translation to larger throughput would more realistically proceed through numbering-up of multiple micro-packed-bed channels operated in parallel, rather than by increasing the diameter of a single reactor and sacrificing the transport advantages of miniaturization. This perspective is also consistent with earlier analyses from our research group emphasizing that micro-packed-bed performance results from confinement-enhanced transport and hydrodynamics, but that practical implementation must still address pressure-drop management, flow distribution, and operability.^35,36 In addition, we conducted a performance assessment of the micropacked bed reactor, which encompassed parameters such as catalyst dilution, reactor temperature, operating pressure, residence time, and the molar ratio of carbon dioxide to propylene oxide. The limitations of fluid–solid mass transfer were studied by extensively comparing the performance of the microreactor to that of conventional centimetric fixed beds. This comparative analysis serves as a preliminary step towards extrapolating the findings to commercial-scale units.

Overall, this study demonstrates that confinement-enhanced CO₂/epoxide coupling with supported ammonium sites is achievable on inexpensive commercial silica gels and quantifies, via micro- versus centimetric-scale packed-bed benchmarking, the reactor-transport contribution to the observed flow performance.

Synthesis methodology

Tripropylamine (≥98%, Sigma-Aldrich), 3-iodopropyltrimethoxysilane (≥95%, Sigma-Aldrich), and silica gels (SiliCycle R10114B (5–20 μm), R10150B (60–120 μm), and R10170B (200–350 μm)) were used as received without further purification. For grafting of the ammonium salts, toluene was pre-treated and distilled under a N₂ atmosphere over sodium/benzophenone before use. The liquid reagents were bubbled under a N₂ atmosphere, and the glassware was placed in an oven at 200 °C to remove all traces of moisture. The reactions were carried out under a N₂ atmosphere using conventional Schlenk techniques. For the catalytic tests, different oxides (Sigma-Aldrich) and gaseous carbon dioxide (CO₂) (Praxair, bone dry) were used as received. Detailed characterization methods used to evaluate the raw and synthesized materials are provided in the SI, section S1.

Synthesis of 3-trimethoxysilyltetrapropylammonium iodide (TPA)

Tripropylamine (2.3 mL, 12.1 mmol) was added to 3-iodopropyltrimethoxysilane (9.0 mL, 46.0 mmol), and the mixture was stirred for 5 min (Fig. 1a). Anhydrous toluene (50 mL) was added to the mixture, and the solution was stirred vigorously under reflux for 72 h. The desired product accumulated in a yellow-brown phase at the bottom of the flask. The resulting oily phase was placed in a rotatory evaporator to remove the remaining solvent. To remove traces of toluene from the oily liquid product, the solution was washed three times with 100 mL of a 1 : 1 ether/hexane mixture (Fig. S1 and S2 in the SI).

Fig. 1 Synthetic routes for (a) TPA, (b) grafting of TPA catalyst onto silica, and (c) TMDS passivation of TPA-grafted silica material.

Synthesis of TPA-catalytic silica materials

The silica gel materials R10114B (R114), R10150B (R150), R10170B (R170), MCM-41, and SBA-15 were preliminarily degassed at 150 °C in a vacuum oven for at least 24 h. For grafting, 1.0 g of silica was dispersed in 100 mL of dry toluene, and 3-trimethoxysilyltetrapropylammonium iodide (300 mg, 0.69 mmol) was added (Fig. 1b). This corresponds to a nominal precursor feed of 0.69 mmol g⁻¹ silica, which also represents the theoretical maximum TPA loading assuming quantitative grafting. The resulting mixture was refluxed for 24 h under a N₂ atmosphere at 110 °C. The solid material was then filtered, and a Soxhlet extraction of the solid in dichloromethane was conducted for 24 h to remove the remaining ungrafted ammonium salt. Finally, the resulting yellowish gels were dried at 60 °C in a vacuum oven for 24 h before physisorption analyses or catalysis tests were performed. The products are referred to as MCM-41-TPA, SBA-15-TPA, and RX-TPA. The actual grafted loading was not independently quantified in this study; therefore, only the nominal precursor/silica ratio is reported.

Passivation of the silica materials with 1,1,3,3 tetramethyldisilazane (TMDS)

To determine the importance of silanol groups on catalysis, the raw materials (R114, R150, R170, MCM-41, SBA-15) and the grafted counterparts (R114-TPA, R150-TPA, R170-TPA, MCM-41, SBA-15) were passivated using TMDS (Fig. 1c). To do this, the materials were degassed at 150 °C (MCM-41, SBA-15, and RX) and 80 °C (RX-TPA). Then, 1.0 g of silica was suspended in 50 mL of n-hexane, and TMDS (2.0 mL, 1.50 g, 11.2 mmol) was added dropwise over 15 min under vigorous stirring. This corresponds to a nominal TMDS/silica feed ratio of 11.2 mmol g⁻¹. The reaction mixture was stirred for 24 h, and the solid was transferred to a cellulose Soxhlet thimble and washed with CH₂Cl₂ for 12 h. These materials were then degassed under vacuum for 4 h at 60 °C before use.

Tables S1–S5 and Fig. S3–S7 (in the SI, section S3) compare the contractions in pore diameter after grafting to the non-functionalized material, confirming the success of the procedure. The quantitative differences observed between the two functionalized catalysts, R170-TPA and R150-TPA, can be attributed to the higher amount of quaternary ammonium species grafted onto the surface of R170. Grafting success was also confirmed by thermogravimetric analysis during air burn-off of the grafted organic materials (SI, Fig. S9–S13, section S4). Fig. S14 shows the FTIR spectra after the grafting and passivation steps, thus confirming the attachment of different groups of TPA and TMDS, like C–N symmetrical elongation, Si–C, and Si–H elongations. The details on all spectra can be found in the SI, section S5.

Catalytic conversion of CO₂ and epoxide to cyclic carbonates in batch experiments

The batch catalytic experiments were performed in 10 mL vials with 100 mg of silica material and the respective amount of styrene oxide. For reactions with higher catalyst loading (over 8 mol%), 5 mL of anhydrous diethyl ether was added to the mixture to ensure the diffusion of styrene oxide within the pores, and the solution was stirred for 30 min at 400 rpm to ensure that the epoxide was mixed well with the silica. Subsequently, the ether was evaporated under vacuum, and the reaction medium was placed under one atmosphere of CO₂. The mixture was stirred at 300 rpm for 24 h, then 5 mL of acetone was introduced into the vials at the same stirring intensity. Again, a 30-min stirring period was required to ensure homogenization and solubilization of the reagents and products within the pores of the materials. The contents of the vials were transferred into conical tubes for centrifugation at 10 000 rpm for 10 min. The acetone supernatant was then pipetted into 20 mL scintillation vials. The acetone solvent was evaporated to isolate the reaction product. The isolated yields were obtained, and the purity was determined using ¹H NMR (Fig. S15–S22, section S6). All catalytic reaction experiments were performed in triplicate, and the mean of the three measurements was calculated for each condition. The error bars in the figures represent the standard deviation.

Catalytic conversion of CO₂ and epoxide to cyclic carbonates in flow systems

In the flow system, the cycloaddition reaction was hosted in a micropacked bed reactor (Fig. S23) that was subject to a gas-phase throughflow. This microreactor consisted of a stainless-steel tube (15 cm length and 0.9 mm internal diameter) that was previously filled with grafted silica gels. A comparison was performed between the microreactor system and a conventional laboratory packed bed of the same length (15 cm) and 10 mm diameter (i.e., 11.1 times that of the microreactor) while keeping the same volumetric flow rates between the two scales. The purpose of this comparison was to anticipate the gain facilitated by a microreactor in comparison to the centimetric laboratory catalytic tests usually performed as a preamble before extrapolation to industrial macro-scale reactors.

Before beginning the catalytic experiments, the catalysts were prepared and activated by removing any moisture present by heating at 100 °C overnight. Liquid propylene oxide (PO) was vaporized inside a pressurized stainless-steel vessel with the liquid supply and vapor outlet placed on its top surface. The vessel was then placed in a water bath that was thermostatted at different temperatures to control the evaporation rate of the PO. To avoid condensation of PO vapors during their transit before entering the reactor, the line was wrapped with a silicone rubber heating tape to maintain a constant and sufficiently high temperature (60 °C). The gaseous stream, consisting of mixtures of high-purity CO₂ and vaporized PO with a mole ratio in the range of 1 to 5 and total volumetric flow rates sweeping the range of 13.4 to 14.6 mL h⁻¹, was fed into the packed bed microreactor placed inside a furnace at a temperature between 25 °C and 120 °C, depending on the experiment.

At the end of the test period, the amount of epoxide added to the container was measured and considered as the input PO. Some of the materials in liquid form were deposited in the product while the remaining amount left the system in the form of unreacted gaseous epoxide. The quantity of gaseous epoxide that left the system was determined using mass balance calculations that involved the input PO composition, the amount used to produce the product (based on stoichiometry), and the amount of epoxide measured by ¹H NMR, leading to the amount of conversion during the experiment. After the reaction, the content of the reactor was cooled to ambient temperature. Deuterated chloroform was added as a solvent for ¹H NMR spectroscopy analysis, the mixture was stirred for another 15 min, and centrifugation was performed to separate the catalyst from the solution. Finally, the reaction products were identified and quantified using ¹H NMR spectroscopy. The conversion of propylene oxide was estimated by analyzing the ¹H NMR data in combination with mass-balance calculations. This method involves first calculating the integral of the NMR peak for the cyclic carbonate and PO, ensuring that these integrals are normalized based on the number of contributing nuclei, if necessary. The conversion of the cyclic carbonate was then determined using eqn (1). This formula quantitatively reflects the proportion of the cyclic carbonate in the mixture relative to the total amount of both reactant and product and offers a direct insight into the efficiency of the catalytic reaction under study. The turnover frequency (TOF) was calculated as the moles of product formed per mole of catalyst per unit time, which aligns with the principle of quantifying the catalyst's activity over time (eqn (2)).

Results

A catalytic system for the synthesis of cyclic carbonates from CO₂ typically requires two components: a nucleophile to open the ring-strained epoxide and a Lewis acid or Bronsted acid to increase the electrophilicity of the epoxide. As shown in Fig. 2, the cyclization of CO₂ to epoxides typically occurs in two different ways, depending on the catalyst and reaction conditions.³⁷ A nucleophile can either attack the epoxide (A.1) to open the ring or attack the carbon (A.1′) of the CO₂ molecule. In both cases, highly reactive negatively charged reaction intermediates (Int. 1) and (Int. 1′) are produced. These intermediates can be stabilized with positively charged compounds, such as onium salts, to better control the reaction. Then, depending on the catalytic pathway used, the negatively charged intermediate can attack the electrophilic carbon of the coupling partner. The final step of the reaction, however, is the same for both synthetic pathways because it involves intramolecular attack to close the ring, produce cyclic carbonate, and facilitate the removal of the nucleophile as a leaving group.

Fig. 2 Catalytic mechanism of CO₂ cyclization to epoxides.

Control experiments conducted in the presence of only tetrabutylammonium iodide in bulk solution show that very low yields are obtained and that several hydrolysis side products are produced when no Lewis acid catalyst is used.³⁸ Although silica materials do not possess inherent Lewis acids in their structure, the presence of silanol groups on the surface can help stabilize the system by hydrogen bonding to the oxygen atoms of the epoxide and/or activated carbon dioxide.^39,40 Although catalysts typically accelerate chemical reactions by reducing the energy of activation of the substrates, it is also possible to accelerate the same reaction by increasing the concentration of the reagents. By exploiting oversolubility, which can increase the local concentration of CO₂ by a factor of 7 within the pore structure of mesoporous silica, it is possible to increase a reaction rate without requiring higher pressures.⁴¹ To confirm the generality of this process in commercial materials rather than in well-ordered silica, we compared the catalytic activity of SBA-15 and MCM-41 to that of R114, R150, and R170 silica in the conversion of epoxides to cyclic carbonates in the presence of CO₂. These latter materials, produced by Silicycle, are irregular silica gels with pores that have diameters in the range of 6 to 7 nm and surface areas in the range of 400–600 m² g⁻¹. The grafting of 3-iodopropyltrimethoxysilane on these commercial silicas occurred readily and in high yield, offering a coverage of approximately 15% in mass, which is like what was observed with SBA-15 and MCM-41.⁴² A comparison of the physical properties of silica materials is presented in Tables S1–S5 in the SI. As shown by these data and the spectroscopic characterization that was carried out on all silica materials (see SI), there is no significant difference between the regular (SBA-15 and MCM-41) and irregular (R114, R150, and R170) silicas before and after grafting.

Catalytic conversion tests

Cycloaddition of CO₂ and styrene oxide in a batch system

The first parameter to be optimized was the catalyst loading (Fig. 3b). The variation in the catalyst loading in the reaction was achieved by modifying the amount of styrene oxide. Thus, each reaction used 100 mg of R150-TPA, and the amount of epoxide was varied between 21 mg and 2100 mg to study a catalyst ratio of 11.5% to 0.2%, respectively. The best results, with yields greater than 75%, were obtained using a catalyst loading of approximately 1% after 24 h of reaction time. This catalyst molar ratio corresponded to approximately 425 mg of styrene oxide per 100 mg of material at 15% grafting. Once the optimal catalyst ratio was found, the reaction time was varied. The isolated yields obtained after a 4 h reaction were about 20% lower than those after a 24 h reaction. Although reactions with yields greater than 60% allowed isolation of the cyclic carbonate without purification because no side reaction was observed, an initial washing with n-hexane and a second washing with a mixture of acetonitrile and n-hexane (1 : 1) were necessary when the yields were lower. Control reactions with the ammonium salt were performed to ensure that the catalytic support was necessary to obtain good yields. The results demonstrate that the ammonium salt under similar conditions gives a low yield of 4 ± 4% of the cyclic carbonate under optimal conditions. These findings corroborate studies with the mesoporous silicas MCM-41 and SBA-15, in which acceleration was possible due to the presence of the silica material.³⁴

Fig. 3 (a) Cyclization reaction of CO₂ to various epoxides for batch conditions (10 ml volume of vessel) (conditions: 4 or 24 h, 50 °C, and 1 bar CO₂), (b) optimization of catalyst time and molar ratio using a fixed quantity (100 mg) of R150-TPA. Styrene oxide was added to give various mol% of the catalyst. Purification is more challenging at a lower yield of product. 1 mol% is optimal for batch reactions. Yields were calculated as (moles of isolated cyclic carbonate/moles of limiting reactant) × 100%.

The yields obtained with catalyst loadings higher than 10% are low for two reasons. First, it was difficult to isolate the styrene carbonate that tends to remain within the pores. Second, the supernatant transfer steps during centrifugation resulted in significant losses. In comparison, the test with a catalyst loading of 0.2 mol% used 2500 mg of the epoxide, which made the purification and yield more satisfactory. However, because this carbonate is soluble in the epoxide, excess styrene oxide can act as a solubilizing agent while also acting as a reagent. Thus, the catalytic system developed using these materials is optimal when using a large quantity of the reagents. As a result, we decided to reproduce the CO₂ cyclization reactions using styrene oxide with commercial materials at a catalyst molar ratio of 1%, a temperature of 50 °C, a pressure of 1 bar, and a reaction time of 24 h.

In the next step, we studied catalytic conversion using pristine and passivated materials, as presented in Fig. 4a and b, respectively. These tests were carried out under the optimized conditions of 50 °C and 24 h, but with a silica mass corresponding to that used in the 1 mol% tests. These tests were necessary because the surface of silica materials contains silanol groups, which, through H bridges, can promote the opening of the epoxide and aid in the conversion. However, low conversions were expected because no nucleophiles were present in the reaction medium to catalyze the transformation. This assumption was proven correct based on the results in Fig. 4a. Although materials R114 and R150 yielded only 9 ± 4% and 7 ± 2%, respectively, material R170 yielded conversions of 30 ± 15% even without the anion. Although small-sized silica particles have difficulty achieving conversion, the significant conversion observed with material R170 could be explained by a greater abundance of silanol groups on its surface. Passivated materials were also tested in the catalytic transformation (Fig. 4c). Similar results were obtained for the raw materials, although a slight increase in yields was observed with increasing particle size. Likely, the smaller pore diameters of the R150-TMDS and R170-TMDS materials, 3.9 and 4.2 nm, respectively, instead of 5.8 nm for R140-TMDS, could accentuate the oversolubility effect.³⁴ Thus, the surrounding CO₂ crowding is higher, and this has the effect of increasing the rate of the homogeneous uncatalyzed transformation. This hypothesis may explain the average yield of R114-TMDS of 9 ± 2% compared to the yields of 29 ± 12% and 31 ± 8% obtained for R150-TMDS and R170-TMDS, respectively. The TMDS group allows the removal of certain silanol groups by substitution with a silane group SiH(CH₃)₂, on the surface, but complete passivation of the material is never truly possible, which suggests that certain silanol groups, or even the hydrogen on the TMDS group, may contribute to the observed conversion. It should be noted that the epoxide possesses a high ring strain energy caused by the 60° angles between these atoms, which gives it high reactivity. This “potential” chemical energy, combined with the addition of heat, pressure, and a few hydrogen bonds, can promote epoxide opening even in the absence of an added nucleophile. Fig. 4d presents the catalytic activation of epoxides using TPA-grafted materials under optimal conditions (1 mol%, 50 °C, 24 h). The results of the catalytic cyclization of CO₂ to styrene oxide confirm that the ammonium salt is essential to achieve yields greater than 70%. The R114-TPA catalyst yielded 75 ± 6%, R150-TPA yielded 67 ± 8%, and R170-TPA yielded 74 ± 5%. The iodide counterion of the salt promotes activation of the epoxide to open the ring. Compared to control reactions with only the catalyst, the catalytic support, which contains silanol groups and pores of an appropriate size for the supersolubility effect, has a favorable effect on ring opening. Although significant differences were sometimes observed among the yields of the triplicates, we nevertheless have a certain reproducibility of the yields. The commercial silicas, surface-grafted with the ammonium catalyst and passivated with TMDS, were also tested. The R114-TPA-TMDS catalyst yielded 82 ± 6%, R150-TPA-TMDS yielded 74 ± 5%, and R170-TPA-TMDS yielded 76 ± 6%. The TMDS-passivated materials exhibited a yield increase of 6%, 7%, and 2% for the R114-TPA-TMDS, R150-TPA-TMDS, and R170-TPA-TMDS materials, respectively, compared to their catalyst-only counterparts. It is possible that the passivating agent slightly increases conversion. This hypothesis stems from the observation that the starting epoxide is soluble in n-hexane, whereas styrene carbonate is not. Thus, styrene oxide will have a certain affinity for a hydrophobic surface, whereas styrene carbonate, which has a lesser affinity for hydrophobic substances, could be more easily dislodged from the active sites. The excess styrene oxide and its solubilizing effect on styrene carbonate should also help dislodge the formed product from the pores and aid in catalysis.

Fig. 4 Results of catalytic tests for batch conditions (10 ml volume of vessel) with (a) pristine silica (purple), (b) RX-TPA-TMDS (blue), (c) RX-TMDS (green), and (d) RX-TPA (red) materials (conditions: 24 h, 50 °C, 1% mol, 1 bar of CO₂). All catalysts were used at an identical 1 mol% molar loading (ratio of catalyst sites to epoxide). Error bars represent standard deviation.

We tested the synthetic materials MCM-41 and SBA-15 and the grafted analogues under the same conditions to confirm that the irregular silicas are as efficient as the regular ones. The results, shown in Fig. 5a, demonstrate yields lower than those obtained with commercial materials. The yields obtained were 66 ± 24% for MCM-41-TPA and 65 ± 7% for MCM-41-TPA-TMDS, and 61 ± 6% for SBA-15-TPA silica and 70 ± 7% for SBA-15-TPA-TMDS silica. The lower yields with the new catalytic conditions could be explained by the high specific surface area of these MCM-41 and SBA-15 silicas, with specific surface areas ranging from 600 to 1000 m² and from 400 to 800 m², respectively.⁴³ This physical characteristic can cause strong adsorption of liquids and require a larger quantity of reagents to disperse and wet these materials. Without the addition of a solvent to produce better homogenization and dispersion of the epoxy on the material, lower yields (than those previously obtained) could be obtained. Fig. 5b shows the yields obtained for the larger-scale reactions (1 mol% of catalyst and 5 g of styrene oxide). Catalyst R114-TPA achieved a yield of over 78% during its first cycle and a better yield of 84% during the second cycle. The efficiencies obtained with the R114-TPA-TMDS catalyst were equally good for all three operating cycles. The efficiency was 80% for the first cycle, 78% for the second cycle, and 83% for the third cycle.

Fig. 5 (a) Results of catalytic tests for batch conditions (10 ml volume of vessel) with MCM-41 and SBA-15 material (conditions: 24 h, 50 °C, 1 bar CO₂, 1% mol) and (b) results of large-scale catalytic tests with different hybrid materials. In panel (b), each cycle represents a complete batch reaction at identical conditions, with the catalyst recovered and reused without pretreatment. Error bars indicate ±1 SD from triplicate runs within each cycle.

Conversion evolution as a function of reaction time

Monitoring of the reaction between styrene oxide and CO₂ using catalyst R114-TPA over a period of 24 h was carried out using ¹H NMR (Fig. S15), which was acquired after 0, 1, 2, 3, 4, 6, 8, 20, and 24 h. It is interesting that the conversion was 50% after only 3 h and reached a maximum value after 8 h, thus eliminating the need for longer reaction times. These results reveal biphasic kinetics: fast ring-opening at 16.7% h⁻¹ (0–8 h, 73% conversion) transitions to slow desorption at 0.125% h⁻¹ (8–24 h plateau, 75% final), a 133-fold rate decrease. TMDS passivation accelerates only the desorption phase (+7% yield), proving that product removal, not ring-opening, is rate-limiting at high conversion. Flow reactor performance confirming this mechanistic picture (microreactor 82% > centimetric 70% despite lower temperature and shorter residence time) suggests that rapid product extraction circumvents the desorption bottleneck. Thus, product desorption is the critical rate-limiting step at industrially relevant conversions.

With confirmation of the catalytic efficiency of commercial silica materials for the conversion of styrene oxide to styrene carbonate, we decided to study the efficiency and reactivity using other epoxides and 1 mol% of catalysts R114-TPA and R114-TPA-TMDS under the optimal conditions tested above (Fig. 6). The study of the two catalysts was carried out to determine whether the hydrophilic/hydrophobic surface affinities might cause differences depending on the reagents tested. Terminal epoxides 1 to 7 reacted well, and the two catalytic systems developed from commercial silica R114 proved effective in completing the cyclic carbonate formation reaction. The yields obtained with the terminal epoxides were mostly greater than 70%, except for 4-(fluoromethyl)-1,3-dioxolan-2-one (1) and 4-(bromomethyl)-1,3-dioxolan-2-one (2), which yielded around 50% for both hybrid materials (section S6). One reason for these low yields is epoxide volatility, and losses occur under the combined influence of heat and the gas sweeping flow. However, the reaction products are mainly the corresponding carbonates according to their ¹H NMR spectra, which agree with those listed in the literature.^44,45

Fig. 6 Results of catalytic reactions with different epoxides using R114-TPA and R114-TPA-TMDS.

Despite the lower yields of compounds 1 and 2, it is interesting to note that electron-withdrawing groups such as halides do not appear to influence catalytic activity. The yields of compounds 3 to 5, which are known for their electron-donating properties, are also satisfactory, with yields between 68% and 94%. Functional groups such as esters and chlorobenzene in compounds 6 and 7 also do not appear to have an impact on catalytic efficiency.

However, the solid catalyst was markedly less effective for internal or sterically hindered epoxides, since the corresponding carbonates 8, 9, and 10 were not detected by ¹H NMR. This behavior likely does not reflect a general intrinsic inactivity of the catalyst, because terminal epoxides bearing a range of electron-withdrawing and electron-donating substituents still afforded moderate to high yields under the same conditions. Instead, the poor reactivity of the internal/hindered substrates is more plausibly attributed to a combination of steric effects and restricted accessibility of the active sites in the porous heterogeneous catalyst. In such cases, bulkier epoxides may undergo less favorable nucleophilic ring opening and may also experience slower diffusion or less effective confinement within the pore network. Accordingly, the present results suggest a substrate-dependent limitation arising from both steric and transport factors, although dedicated mechanistic studies would be required to quantitatively distinguish between these contributions.

Cycloaddition of CO₂ and propylene oxide under flow conditions

Under flow conditions, we selected R170-TPA over the other two materials due to its higher ligand capacity coupled with its larger particle size, which should favor diffusion and limit the backpressure. This strategic focus allowed us to delve deeper into the relationship between microreactor design and catalytic efficiency and optimize the application of R170-TPA in catalyzing environmentally significant reactions.

As shown in Fig. 7, the ¹H NMR data for a cycloaddition reaction performed at room temperature and atmospheric pressure in the packed bed microreactor confirm the observation of two different types of products, namely, the cyclic propylene carbonate (PC) monomer and poly(propylene carbonate) (poly(PC)). Poly(PC) generation is a side reaction that often occurs in heterogeneous systems, even if it was not observed with styrene oxide in our batch experiments. The poly(PC) can be formed either by ring-opening polymerization of propylene carbonate or by the copolymerization of CO₂ with PO. This may lead to adsorption or deposition of the polymer onto the catalyst surfaces, thereby blocking active sites. This fouling effect can reduce the catalyst activity over time unless it is mitigated by polymer removal or catalytic reaction design.⁴⁶ Temperature, time, and catalyst concentration^47,48 are factors that can influence the rate of epoxide ring-opening, thus promoting or slowing the growth of the polymer chain towards poly(PC) production.

Fig. 7 1H NMR spectra of PC and poly(PC) produced using R170-TPA as catalysts in the micro-reactor at T: 25 °C, P: 1 bar, t: 30 min.

Dilution of the catalyst in an inert matrix was examined as a countermeasure to the loss in selectivity of the catalytic reaction for the targeted monomer.⁴⁹ The influence of catalyst dilution on the PC/poly(PC) ratio was demonstrated using reduced amounts of the catalyst (R170-TPA) and by adding R170 as a diluent to keep a constant mass in the reactor inventory (Fig. 8). The dilution ratios (mass_diluent/mass_catalyst) were varied at ratios of 1 : 1, 1 : 2, 1 : 3, and 1 : 4 for both the microreactor (Æ = 0.9 mm) and the centimetric laboratory reactor (Æ = 1 cm) while holding the total inventory mass constant at 75 and 515 mg, respectively. For the conditions described in Fig. 8, the molar flux of CO₂ was 40% more than that of PO. Therefore, the reaction was followed in terms of partial PO conversions to PC and to poly(PC). Unwanted conversion of PO to poly(PC) was eliminated in the microreactor if the diluent and catalyst were mixed at a ratio greater than or equal to 1 : 2, whereas higher diluent mass (ratios of 1 : 3 and 1 : 4) resulted in lower conversions to PC (51% and 33%, respectively) versus 80% at the 1 : 2 ratio. The use of catalyst dilution in the centimetric reactor could not completely prevent poly(PC) generation, as reflected in lower PC conversions. However, the two scales having been subject to an identical volume flow rate implies a longer gas residence time in the centimetric reactor (about 2 orders of magnitude) and a consequent loss in selectivity towards the key product (PC). Although RTD was not measured directly in the present work, the comparison between the two reactors can be interpreted considering previous micro-packed-bed hydrodynamic studies from our research group. Prior work has shown that deviations from plug flow in micro-packed beds are generally limited and that near-wall flow can be characterized by locally enhanced velocities arising from the confined geometry.⁵⁰ These observations are in line with the broader conclusion that micro-packed beds can approach more ideal contacting behaviour than larger packed systems under suitable operating conditions.³⁵ In this context, the much higher conversion observed in the microreactor at very short contact times, compared with the lower conversion observed in the centimetric reactor despite its longer residence time and higher temperature, indicates that reactor performance cannot be explained by residence time alone. Rather, the centimetric reactor likely suffers from broader hydrodynamic non-idealities and stronger external transport limitations, whereas the microreactor operates closer to an apparent kinetic regime.³⁵

Fig. 8 Synthesis of PC in the gas flow reactors with different catalyst dilution loadings: (a) micrometric reactor, (b) centimetric reactor (in each case, the evaporation rate of PO was 1.6 mL h⁻¹, the flow rate of CO₂ was 13 mL min⁻¹, T: 25 °C, P: 1 bar).

Conversely, a significant gain was obtained by favoring the use of microreactors that require lower catalyst quantities and shorter contact times, while the dilution acts as an anti-polymerization barrier. Hence, an effective strategy to suppress this undesired polymerization is to rapidly remove PC product from the reactor before it undergoes further conversion into poly(PC) through parasitic chain-growth processes. However, it should be kept in mind that under the reaction conditions of the test, the reaction products accumulate as a viscous liquid phase that is further trapped by capillary forces within the porous medium. Given the identical micro-mixing quality of the active/inert powders used in both reactors, it is likely that the longer contact times in the centimeter reactor favor mutual diffusion of the PC monomers and thus their polymerization, as demonstrated by analysis of liquid samples after 30 min of reaction.

The best trade-off in terms of selectivity and conversion was achieved with the 1 : 2 powder mixture (Fig. 8). The 1 : 2 dilution ratio is also noteworthy from a practical standpoint, since it reduced the amount of active catalytic material required while preserving high selectivity toward propylene carbonate and suppressing poly(propylene carbonate) formation. This result suggests that catalyst dilution with virgin silica may be a useful strategy not only for hydrodynamic and selectivity control, but also for improving the practical viability of the process.

Although the micro-packed bed exhibited substantially higher conversion and TOF than the centimetric packed bed, this intensification benefit must be interpreted together with the hydraulic penalty intrinsic to micro-structured packed beds. Micro-packed beds are typically associated with large axial pressure gradients because of their small hydraulic diameter and confined pore network, as shown in our previous hydrodynamic studies.^35,36,50,51 Their hydrodynamics are also strongly influenced by capillary effects and hysteresis, which further affect pressure-drop behavior under gas–liquid flow.^36,51 Therefore, the present comparison should be understood as a performance trade-off: the microreactor offers shorter characteristic diffusion lengths and improved fluid–solid contacting, while the centimetric reactor is less hydraulically demanding but more vulnerable to transport limitations. This interpretation is consistent with prior analyses from our research group showing that confinement and particle-scale hydrodynamics strongly govern reactor behavior in micro-packed bed flow.^36,51

Hence, the effect of temperature and reaction time on the cycloaddition reaction catalyzed by the dilute catalyst (1 : 2 ratio) was examined for both reactor scales in the temperature and time ranges of 25 to 120 °C and 30 to 240 min, respectively (Fig. 9). Not surprisingly, the PO partial conversion to PC in the microreactor increased from 80% to approximately 95% when the temperature was increased from 25 °C to 120 °C. However, a temperature above 120 °C led to the formation of side products, which resulted in a significant decrease in product selectivity.^52–55 When the reaction was run for 30 min in the centimetric reactor, the conversion to PC was increased from 43% to 68% by increasing the temperature from 25 °C to 120 °C. Although polycarbonate formation was gradually decreased at higher temperatures (up to 120 °C, at which no polymer was observed), some polymerization activity could not be suppressed for a longer reaction time (>30 min).^47,56 The significance of selecting an appropriate microreactor becomes unquestionable, particularly when operating at elevated temperatures and short reaction times, thus avoiding the formation and growth of polymeric chains and consequently favoring the selective and entropy-driven production of cyclic carbonate.⁴⁸

Fig. 9 Synthesis of PC in the flow reactors at different reactor temperatures using a 1 : 2 dilution ratio: (a) micrometric reactor, (b) and (c) centimetric reactor (in each case, the evaporation rate of PO was 1.6 mL h⁻¹, the flow rate of CO₂ was 13 mL min⁻¹, P: 1 bar).

The next operating variable studied was the effect of partial pressure of CO₂ on the conversion of PO to PC in both reactors (Fig. 10a and b). The conversion to PC in the microreactor first increased from 80% to 93% when the CO₂ pressure was increased by a factor of 4. Beyond this factor, the conversion was reduced, but there was no information on whether the inhibition was due to chemical or physical factors. However, due to the drawbacks of the centimetric reactor enumerated above, although it qualitatively exhibits the same CO₂ pressure dependence as the microreactor, it nevertheless achieves much lower PC conversions.

Fig. 10 (a) and (b) Effect of CO₂ pressure on conversion of PC at 25 °C and 120 °C, and residence time of 0.44 s and 54.2 s, in micro-reactors and macro-reactors, respectively. Reaction conditions: CO₂/PO (mol mol⁻¹) = 1 : 1. (c) and (d) Effect of the molar ratio of CO₂/PO on the conversion of PC. Reaction conditions: micro-system (T: 25 °C, P: 1 bar, and residence time: 0.44 s) and macro-system, respectively (T: 125 °C, P: 1 bar, and residence time: 54.2 s).

The effect of the molar ratio of CO₂ to PO on the PC conversion at constant pressure was evaluated next (Fig. 10c and d). The experiments were conducted over a range of PO evaporation rates (0.4–1.6 mL h⁻¹), whereas the CO₂ flow rate was fixed at 13 mL h⁻¹. In the microreactor, an increase in the molar ratio of CO₂ to PO from 1 : 1 to 5 : 1 induces a negligible change in the conversion of PC, a quite expected result considering that the equimolecular reaction involves PO as the minority reactant. The centimetric reactor shows atypical behavior in which a stoichiometric excess of CO₂ leads to a decrease in PC conversion. This inhibition by CO₂ was also found when the partial pressure of CO₂ was increased (see Fig. 9a and b and the discussion above).

The residence time (also referred to as contact time) should not be confused with the preset reaction time discussed in connection with Fig. 10a and b; the former is controlled by the total incident flux of reagents. To evaluate the effect of residence time on PC conversion, the volumetric flow rates of evaporated PO and CO₂ were adjusted in the ranges of 0.2–4.8 mL h⁻¹ and 2.1–40 mL min⁻¹, respectively, for the microreactor and 1.6–11.2 mL h⁻¹ and 13–90 mL min⁻¹, for the centimetric reactor, while holding constant the molar ratio of the reactants (Fig. 11). The PC conversion in both systems increases with increasing residence time, whereas the turnover frequency (TOF) values decrease (Fig. 11). In addition, the TOF values are much larger in the microreactor (51–715 h⁻¹) than in the centimetric reactor (2–9.1 h⁻¹). If the reaction is kinetically controlled, the residence time and conversion should have a direct relationship.^57,58 Thus, if the chemical regime were to prevail in the two reactors while it is subjected to an identical feed composition, then the one with the longer contact time and higher temperature should lead to higher conversions. However, the opposite is observed (Fig. 11). The microreactor converts 85% of the reactants in 2.6 s at 25 °C, compared to less than half (i.e., 40% conversion) during a three-fold greater run time (7.7 s) in the centimeter reactor at 125 °C. Therefore, it is not possible to rule out the negative interference of the external diffusional transport limitations of the reactants and products (as well as hydrodynamic imperfections related to feed distribution and velocity profile developments) on the decrease of conversion in the centimetric reactor, which evidences a rapid cycloaddition reaction for a given catalyst granulometry. If the catalytic reaction is intrinsically rapid, then operation of a miniaturized reactor comes closer to the desirable chemical regime due to the attenuation or complete riddance of external diffusion limitations, in contrast with a cm-scale packed bed reactor. Therefore, it can be concluded that the cycloaddition reaction is controlled by external mass transfer in the case of the centimetric reactor, but that (presumably) an apparent kinetic (at most dependent on intraparticle diffusion limitations) may be in effect in the microreactor.

Fig. 11 Effect of residence time on PC conversion in (a) 0.9 mm and (b) 1 cm reactor. CO₂/PO (mol mol⁻¹): 1; P: 1 bar, T: 25 °C and 125 °C for the micro- and macro-systems.

To contextualize our findings, Table 1 compares the operational performance of our centimeter- and micrometer-scale systems against established continuous flow reactors reported in the literature. This comparison demonstrates how reactor geometry and scale influence key metrics such as conversion efficiency, selectivity, and overall productivity across different operational regimes.

Table 1 Continuous-flow reactors, with different catalysts and geometries, reported for CO₂ cycloaddition reactions to produce cyclic carbonates

Catalyst (type)	Support	Epoxide	Reactor ID	T (°C)	P (bar)	PC yield (%)	Ref.
TPA (tethered ammonium iodide)	Commercial silica (R170)	Propylene oxide	0.9 mm	25	1	82	This work
TPA (tethered ammonium iodide)	Commercial silica (R114)	Propylene oxide	1 cm	125	1	70	This work
[bmim]Br	—	Propylene oxide	2 mm	140	20	77.8	Wu, 2021 (ref. 59)
tBu	Amorphous silica	Propylene oxide	1 cm	170	1	50	North, 2009 (ref. 60)
Mesoporous melamine-formaldehyde resin (MMFR)	—	1,2-Butylene oxide	1 cm	120	13	100	Bui, 2020 (ref. 61)
DBU-based ionic liquid	SBA-15	Propylene oxide	—	90	20	57.10	Sun, 2021 (ref. 62)
Cs–P–Si mixed oxide	—	Propylene oxide	1 cm	200	80	94	Yasuda, 2006 (ref. 63)
MOF-100 (sc)	—	Propylene oxide	—	100	4	57	James, 2018 (ref. 64)

---

To comprehensively quantify process intensification, space–time yields were calculated for all reactor configurations, revealing a ∼228-fold improvement in the microreactor compared with batch operation (Table 2).

Table 2 Space–time yield analysis and performance comparison of batch and continuous-flow reactor systems

Reactor type	Volume (mL)	Residence/reaction time	Yield (%)	STY (mol L^−1 h^−1)
Batch	10	24 h	75	36
Centimetric	11.78	54.2 s	70	45
Microreactor	0.0955	0.44 s	82	8200

---

Conclusion

The cycloaddition of epoxide and CO₂ was evaluated in both batch and micro-packed bed reactors. Compared to conventional fixed-bed reactors, the reaction times were significantly reduced, from several days in batch systems to ∼1 minute in centimetric reactors and <1 second in microreactors, due to the benefits of gas flow chemistry. In the microsystem, efficiency was examined by investigating the effect of operating conditions such as catalyst dilution, temperature, pressure, residence time, and CO₂/epoxide molar ratio.

The following conclusions can be drawn. The cyclic carbonate reaction can be intensified by diluting the catalytic silica powders (to 1 : 2 virgin silica/grafted silica) at room temperature in the microreactor to achieve 80% conversion. However, in the centimetric reactor, only 68% cyclic carbonate conversion was observed at 125 °C. By increasing the pressure from 1 to 3 bar, the conversion in the microreactor was increased from 80% to 92%, in contrast to the centimetric reactor, which displayed much lower performance; from 1 to 5 bar, the conversion increased from 68% to 83%. The residence time can be significantly reduced from several hours in a conventional batch reactor to nearly a minute in a continuous centimeter reactor and to only less than a second in the microreactor for comparable conversions. The external diffusional limitations in terms of CO₂ and epoxide mass transfer can be significantly improved by using packed-bed microreactors that provide a closer approach to the chemical regime of catalysts or even the fast ‘intrinsic’ reaction regime.

In this study, we have purposely not addressed issues related to chemical reaction engineering scale-up methodology, such as numbering up a multiple microreactor array. In addition, while the micropacked-bed benchmarking indicates that reactor-scale transport effects contribute to the observed flow performance, we did not complete a full quantitative decoupling of mass transfer from intrinsic kinetics (e.g., via Thiele modulus/effectiveness factor or a comprehensive dimensionless analysis). These experiments, coupled with an effectiveness-factor framework, will be the focus of follow-up work.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI).

Supplementary information is available. See DOI: https://doi.org/10.1039/d6re00100a.

Acknowledgements

The authors thank the Fonds de recherche du Québec-Nature et technologies (FRQNT), Ministère de l'Économie, de l'Innovation et de l'Énergie (Québec), and the National Sciences and Engineering Research Council of Canada (NSERC) for financial support. We also acknowledge the contribution of Centre en Chimie Verte et Catalyse (FRQNT) and the Centre for Innovation and Research on Carbon Utilization in Industrial Technologies (NSERC). FGF acknowledges CRSNG for a Canada Research Chair. We acknowledge Silicycle for the gift of R10114B (R114), R10150B (R150), R10170B (R170).

References

1. E. Hanson, C. Nwakile and V. O. Hammed, Carbon capture, utilization, and storage (CCUS) technologies: Evaluating the effectiveness of advanced CCUS solutions for reducing CO₂ emissions, Results Surf. Interfaces, 2025, 18, 100381.
2. M. H. Barecka, J. W. Ager and A. A. Lapkin, Economically viable CO₂ electroreduction embedded within ethylene oxide manufacturing, Energy Environ. Sci., 2021, 14, 1530–1543.
3. I. Mengesha and D. Roy, Carbon pricing drives critical transition to green growth, Nat. Commun., 2025, 16, 1321.
4. M. Aresta, A. Dibenedetto and A. Angelini, Catalysis for the Valorization of Exhaust Carbon: from CO₂ to Chemicals, Materials, and Fuels. Technological Use of CO₂, Chem. Rev., 2014, 114, 1709–1742.
5. Q.-W. Song, Z.-H. Zhou and L.-N. He, Efficient, selective and sustainable catalysis of carbon dioxide, Green Chem., 2017, 19, 3707–3728.
6. Q. Liu, L. Wu, R. Jackstell and M. Beller, Using carbon dioxide as a building block in organic synthesis, Nat. Commun., 2015, 6, 5933.
7. M. Bui, et al., Carbon capture and storage (CCS): the way forward, Energy Environ. Sci., 2018, 11, 1062–1176.
8. J. A. Martens, et al., The Chemical Route to a Carbon Dioxide Neutral World, ChemSusChem, 2017, 10, 1039–1055.
9. J. Artz, et al., Sustainable Conversion of Carbon Dioxide: An Integrated Review of Catalysis and Life Cycle Assessment, Chem. Rev., 2018, 118, 434–504.
10. A. W. Kleij, M. North and A. Urakawa, CO₂ Catalysis, ChemSusChem, 2017, 10, 1036–1038.
11. M. Rebei, et al., Fast carbon dioxide–epoxide cycloaddition catalyzed by metal and metal-free ionic liquids for designing non-isocyanate polyurethanes, Mater. Adv., 2024, 5, 4311–4323.
12. Z. Yu, H. Ishitani and S. Kobayashi, Continuous-flow synthesis of cyclic carbonates with polymer-supported imidazolium-based ionic liquid (Im-PSIL) catalysts, Green Chem., 2024, 26, 11548–11555.
13. B. Schäffner, F. Schäffner, S. P. Verevkin and A. Börner, Organic Carbonates as Solvents in Synthesis and Catalysis, Chem. Rev., 2010, 110, 4554–4581.
14. P. P. Pescarmona, Cyclic carbonates synthesised from CO₂: Applications, challenges and recent research trends, Curr. Opin. Green Sustainable Chem., 2021, 29, 100457.
15. A. Kaithal, M. Hölscher and W. Leitner, Catalytic Hydrogenation of Cyclic Carbonates using Manganese Complexes, Angew. Chem., Int. Ed., 2018, 57, 13449–13453.
16. W. Guo, J. E. Gómez, À. Cristòfol, J. Xie and A. W. Kleij, Catalytic Transformations of Functionalized Cyclic Organic Carbonates, Angew. Chem., Int. Ed., 2018, 57, 13735–13747.
17. J. Mindemark, L. Imholt, J. Montero and D. Brandell, Allyl ethers as combined plasticizing and crosslinkable side groups in polycarbonate-based polymer electrolytes for solid-state Li batteries, J. Polym. Sci., Part A: Polym. Chem., 2016, 54, 2128–2135.
18. R. R. Shaikh, S. Pornpraprom and V. D'Elia, Catalytic Strategies for the Cycloaddition of Pure, Diluted, and Waste CO₂ to Epoxides under Ambient Conditions, ACS Catal., 2018, 8, 419–450.
19. J. Liang, et al., An imidazolium-functionalized mesoporous cationic metal–organic framework for cooperative CO₂ fixation into cyclic carbonate, Chem. Commun., 2018, 54, 342–345.
20. Q. He, et al., Synthesis of cyclic carbonates from CO₂ and epoxides using ionic liquids and related catalysts including choline chloride–metal halide mixtures, Catal. Sci. Technol., 2014, 4, 1513–1528.
21. J. W. Comerford, I. D. V. Ingram, M. North and X. Wu, Sustainable metal-based catalysts for the synthesis of cyclic carbonates containing five-membered rings, Green Chem., 2015, 17, 1966–1987.
22. C. Martín, G. Fiorani and A. W. Kleij, Recent Advances in the Catalytic Preparation of Cyclic Organic Carbonates, ACS Catal., 2015, 5, 1353–1370.
23. L. Xiao, et al., Ionic Conjugated Polymers as Heterogeneous Catalysts for the Cycloaddition of Carbon Dioxide to Epoxides to Form Carbonates under Solvent- and Cocatalyst-Free Conditions, Chem, 2022, 87, 1–23.
24. R. Huang, et al., Deciphering key intermediates in the transformation of carbon dioxide into heterocyclic products, Nat. Catal., 2018, 2, 62–70.
25. R. Khatun, et al., Palladium nanoparticles embedded over mesoporous TiO₂ for chemical fixation of CO₂ under atmospheric pressure and solvent-free conditions, New J. Chem., 2017, 41, 12937–12946.
26. S. M. M. Nataj, S. Kaliaguine and F.-G. Fontaine, Post-synthetic modification of Zr-MOFs using guanidine for cyclic carbonate formation catalysis, Catal. Today, 2023, 422, 114216.
27. S. M. M. Nataj, S. Kaliaguine and F. Fontaine, Highly Efficient Catalysts for CO₂ Fixation using Guanidinium-Functionalized Zr-MOFs, ChemCatChem, 2023, 15, e202300079.
28. J.-Q. Wang, K. Dong, W.-G. Cheng, J. Sun and S.-J. Zhang, Insights into quaternary ammonium salts-catalyzed fixation carbon dioxide with epoxides, Catal. Sci. Technol., 2012, 2, 1480.
29. M. M. Dharman, J.-I. Yu, J.-Y. Ahn and D.-W. Park, Selective production of cyclic carbonate over polycarbonate using a double metal cyanide–quaternary ammonium salt catalyst system, Green Chem., 2009, 11, 1754.
30. C. Kohrt and T. Werner, Recyclable Bifunctional Polystyrene and Silica Gel-Supported Organocatalyst for the Coupling of CO₂ with Epoxides, ChemSusChem, 2015, 8, 2031–2034.
31. M. V. Zakharova, F. Kleitz and F. Fontaine, Back Cover: Carbon Dioxide Oversolubility in Nanoconfined Liquids for the Synthesis of Cyclic Carbonates (ChemCatChem 11/2017), ChemCatChem, 2017, 9, 2021–2021.
32. N. L. Ho, J. Perez-Pellitero, F. Porcheron and R. J.-M. Pellenq, Enhanced CO₂ Solubility in Hybrid Adsorbents: Optimization of Solid Support and Solvent Properties for CO₂ Capture, J. Phys. Chem. C, 2012, 116, 3600–3607.
33. M. Pera-Titus, S. Miachon and J. Dalmon, Increased gas solubility in nanoliquids: Improved performance in interfacial catalytic membrane contactors, AIChE J., 2009, 55, 434–441.
34. S. Boroun, et al., Insights into the Solubility of Carbon Dioxide in Grafted Mesoporous Silica for the Catalytic Synthesis of Cyclic Carbonates by Nanoconfinement, ACS Appl. Mater. Interfaces, 2021, 13, 27019–27028.
35. A. Faridkhou, J.-N. Tourvieille and F. Larachi, Reactions, hydrodynamics and mass transfer in micro-packed beds—Overview and new mass transfer data, Chem. Eng. Process.: Process Intesif., 2016, 110, 80–96.
36. A. Faridkhou and F. Larachi, Two-phase flow hydrodynamic study in micro-packed beds – Effect of bed geometry and particle size, Chem. Eng. Process.: Process Intesif., 2014, 78, 27–36.
37. P. Goodrich, et al., Sustainable Cyclic Carbonate Production, Utilizing Carbon Dioxide and Azolate Ionic Liquids, ACS Sustainable Chem. Eng., 2017, 5, 5635–5641.
38. M. V. Zakharova, F. Kleitz and F. Fontaine, Carbon Dioxide Oversolubility in Nanoconfined Liquids for the Synthesis of Cyclic Carbonates, ChemCatChem, 2017, 9, 1886–1890.
39. T. Q. Bui, et al., Non-hydrolytic sol-gel synthesis of amine-functionalized silica: Template- and catalyst-free preparation of mesoporous catalysts for CO₂ valorization, Microporous Mesoporous Mater., 2025, 381, 113371.
40. S. Hasegawa, et al., Concerted Hydrosilylation Catalysis by Silica-Immobilized Cyclic Carbonates and Surface Silanols, JACS Au, 2023, 3, 2692–2697.
41. S. Boroun, et al., Insights into the Solubility of Carbon Dioxide in Grafted Mesoporous Silica for the Catalytic Synthesis of Cyclic Carbonates by Nanoconfinement, ACS Appl. Mater. Interfaces, 2021, 13, 27019–27028.
42. M. V. Zakharova, F. Kleitz and F. Fontaine, Back Cover: Carbon Dioxide Oversolubility in Nanoconfined Liquids for the Synthesis of Cyclic Carbonates (ChemCatChem 11/2017), ChemCatChem, 2017, 9, 2021–2021.
43. B. Lindlar, A. J. Kogelbauer, P. Kooyman and R. Prins, Synthesis of large pore silica with a narrow pore size distribution, Microporous Mesoporous Mater., 2001, 44–45, 89–94.
44. J. Bensemhoun and S. Condon, Valorization of glycerol 1,2-carbonate as a precursor for the development of new synthons in organic chemistry, Green Chem., 2012, 14, 2595.
45. M. K. Nielsen, C. R. Ugaz, W. Li and A. G. Doyle, PyFluor: A Low-Cost, Stable, and Selective Deoxyfluorination Reagent, J. Am. Chem. Soc., 2015, 137, 9571–9574.
46. R. Abdul-Karim, A. Hameed and M. I. Malik, Ring-opening polymerization of propylene carbonate: Microstructural analysis of the polymer and selectivity of polymerization by 2D-NMR techniques, Eur. Polym. J., 2018, 105, 95–106.
47. R. Abdul-Karim, A. Hameed and M. I. Malik, Ring-opening polymerization of propylene carbonate: Microstructural analysis of the polymer and selectivity of polymerization by 2D-NMR techniques, Eur. Polym. J., 2018, 105, 95–106.
48. X.-B. Lu and D. J. Darensbourg, Cobalt catalysts for the coupling of CO₂ and epoxides to provide polycarbonates and cyclic carbonates, Chem. Soc. Rev., 2012, 41, 1462–1484.
49. A. M. Elmér and P. Jannasch, Synthesis and characterization of poly(ethylene oxide- co -ethylene carbonate) macromonomers and their use in the preparation of crosslinked polymer electrolytes, J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 2195–2205.
50. A. Faridkhou, M. Hamidipour and F. Larachi, Hydrodynamics of gas–liquid micro-fixed beds – Measurement approaches and technical challenges, Chem. Eng. J., 2013, 223, 425–435.
51. A. Faridkhou and F. Larachi, Hydrodynamics of Gas–Liquid Cocurrent Flows in Micropacked Beds—Wall Visualization Study, Ind. Eng. Chem. Res., 2012, 51, 16495–16504.
52. Y. Zhao, et al. Quaternary ammonium salt-functionalized chitosan: An easily recyclable catalyst for efficient synthesis of cyclic carbonates from epoxides and carbon dioxide, J. Mol. Catal. A: Chem., 2007, 271, 284–289.
53. T.-Y. Shi, et al. Efficient fixation of CO₂ into cyclic carbonates catalyzed by hydroxyl-functionalized poly(ionic liquids), RSC Adv., 2013, 3, 3726.
54. C. Jing-Xian, et al. Catalytic fixation of CO₂ to cyclic carbonates over biopolymer chitosan-grafted quarternary phosphonium ionic liquid as a recylable catalyst, Appl. Catal., A, 2014, 484, 26–32.
55. J. Sun, et al. Chitosan functionalized ionic liquid as a recyclable biopolymer-supported catalyst for cycloaddition of CO₂, Green Chem., 2012, 14, 654.
56. Z. Guo and Q. Lin, Coupling reaction of CO₂ and propylene oxide catalyzed by DMC with co-complexing agents incorporated via ball milling, J. Mol. Catal. A: Chem., 2014, 390, 63–68.
57. J. Yue, G. Chen, Q. Yuan, L. Luo and Y. Gonthier, Hydrodynamics and mass transfer characteristics in gas–liquid flow through a rectangular microchannel, Chem. Eng. Sci., 2007, 62, 2096–2108.
58. Y. Zhao, C. Yao, G. Chen and Q. Yuan, Highly efficient synthesis of cyclic carbonate with CO₂ catalyzed by ionic liquid in a microreactor, Green Chem., 2013, 15, 446–452.
59. Y. Wu, et al. Kinetic study of highly efficient CO₂ fixation into propylene carbonate using a continuous-flow reactor, Chem. Eng. Process., 2021, 159, 108235.
60. M. North, P. Villuendas and C. Young, A Gas-Phase Flow Reactor for Ethylene Carbonate Synthesis from Waste Carbon Dioxide, Chem. – Eur. J., 2009, 15, 11454–11457.
61. T. Q. Bui, L. J. Konwar, A. Samikannu, D. Nikjoo and J.-P. Mikkola, Mesoporous Melamine-Formaldehyde Resins as Efficient Heterogeneous Catalysts for Continuous Synthesis of Cyclic Carbonates from Epoxides and Gaseous CO₂, ACS Sustainable Chem. Eng., 2020, 8, 12852–12869.
62. J. Sun, Z. Li and J. Yin, Continuous flow synthesis of propylene carbonate using DBU-based ionic liquid in a packed bed reactor, J. CO₂ Util., 2021, 53, 101723.
63. H. Yasuda, L.-N. He, T. Takahashi and T. Sakakura, Non-halogen catalysts for propylene carbonate synthesis from CO₂ under supercritical conditions, Appl. Catal., A, 2006, 298, 177–180.
64. B. R. James, J. A. Boissonnault, A. G. Wong-Foy, A. J. Matzger and M. S. Sanford, Structure activity relationships in metal–organic framework catalysts for the continuous flow synthesis of propylene carbonate from CO₂ and propylene oxide, RSC Adv., 2018, 8, 2132–2137.

---