Propylene oxide as a dehydrating agent: potassium carbonate-catalyzed carboxylative cyclization of propylene glycol with CO₂ in a polyethylene glycol/CO₂ biphasic system

Abstract

The synthesis of propylene carbonate (PC) from 1,2-propylene glycol (PG) and CO₂ was smoothly performed in a PEG₈₀₀ (polyethylene glycol)/CO₂ biphasic system with K₂CO₃ as a catalyst and propylene oxide (PO) as a dehydrating agent. In the reaction of PG with CO₂, PO presumably removes the water produced, and simultaneously generates more PG, both of which shift the thermodynamic control process and thus accelerate the PC synthesis. The PC yield directly from PG and CO₂ reached 78% under relatively mild reaction conditions (4 MPa, 120 °C, 10 h). Notably, no additional by-product was detected in this process, resulting in economic benefits and the ease of workup procedure.

---

Introduction

Ethylene carbonate (EC), propylene carbonate (PC) and dimethyl carbonate (DMC) are important industrial organic carbonates,¹ which have been widely applied as polar aprotic solvents,^2–4 electrolytes for lithium ion batteries, raw materials for polycarbonates/polyurethanes,⁵ green carbonylation reagents⁶ and fuel additives.⁷ Generally, PC and EC have also been used to prepare DMC through the transesterification with methanol.⁸

Currently, there are two conventional accesses to produce DMC: (a) phosgenation of methanol;⁹ (b) oxidative carbonylation of methanol with carbon monoxide (Scheme 1a and b).¹⁰ Corrosive and toxic reagents are commonly required in those procedures. As the most straightforward and challenging method for direct DMC synthesis, the dehydrating condensation of methanol with CO₂ is highly attractive (Scheme 1c). Although DMC synthesis directly from CO₂ and methanol catalyzed by transition metals has been developed, the process has not been yet applied for industrial application because of low productivity originating from catalyst deactivation and DMC hydrolysis caused by co-produced H₂O. Therefore, considerable efforts have been devoted to overcoming such limitations.^11–13 On the other hand, one of the most efficient and practicable strategies is the two-step process consisting of cycloaddition of epoxides with CO₂, and subsequent transesterification of cyclic carbonate e.g. PC with methanol (Scheme 2a), which has widely been applied in industrial DMC production.¹⁴

Scheme 1 DMC synthetic methodologies.

Scheme 2 Two-step process for DMC production.

In this process, however, an equivalent of propylene glycol (PG) is inevitably released as co-product.¹⁵ Therefore, if PC can be synthesized directly from PG and CO₂, the process can be regarded as DMC synthesis indirectly from MeOH and CO₂ (Scheme 2b). In this context, the synthesis of cyclic carbonates from CO₂ with glycol could be promising but remains in its infancy. We have already developed efficient metal catalysis for PC synthesis from PG and CO₂, such as Bu₂SnO or Bu₂Sn(OMe)₂,¹⁶ magnesium and its oxide.⁹ Despite high selectivity, PG conversion and PC yield were unsatisfactory [e.g. 3.4% yield with Bu₂Sn(OMe)₂] even under harsh reaction conditions (180 °C, 15 MPa). Such low conversion and yield are likely ascribed to the thermodynamic limitations and/or deactivation of the catalysts by the co-product water as shown in Scheme 3. Therefore, removing water to shifting the equilibrium to PC is the key to accomplishing high conversion. Various dehydrating agents have been introduced for this purpose. Inorganic dehydrating agents such as molecular sieve, CaCl₂ are easily recyclable and do not form by-products. However, PC yields were still low because of the dehydration reversibility under the reaction conditions. In other words, it is difficult to physically absorb water at high temperature.

Scheme 3 PC synthesis from PG and CO₂ using PO as dehydrating agent.

As a complement, irreversible hydration with organic compounds is more effective at high temperature. For example, the hydration of nitriles to the amides can conveniently and selectively consume water.^17–20 Therefore, acetonitrile has been applied as dehydrating agent to enhance the reaction between PG and CO₂ with catalysts CeO₂–ZrO₂,^21,22 alkali carbonate (i.e. Cs₂CO₃),²³ modified zinc oxide (i.e. ZnO/KI),²⁴ zinc acetate,²⁵ and organic base (i.e. 1,5,7-triazabicyclo[4.4.0]dec-5-ene, TBD)²⁶ respectively. The drawbacks associated with the use of chemical dehydrating agents are the high cost and the formation of hydrolytic products, which are hard to separate and recycle. The acetamide generated via hydrolysis of acetonitrile could further react with water to afford acetic acid which then reacts with PG to produce propylene glycol-2-acetate and thereby reducing the selectivity of PC.

Herein, we performed the K₂CO₃-catalyzed PC synthesis from PG and CO₂ in PEG/CO₂ biphasic system (Scheme 3). In this work, propylene oxide (PO) works as a dehydrating agent. The hydration of PO consumes water formed, and generates more PG, thus allows the equilibrium to shift to PC formation. The high efficiency may be explained by PEG coordinating with the potassium cation to enhance the basicity and nucleophilicity of carbonate anion.^27,28 In addition, good solubility of CO₂ and K₂CO₃ in PEG may also contribute to such promotive effect.^32,33 As a consequence, this PEG/CO₂ biphasic system by employing PO as a dehydrant could efficiently prepare PC with no by-product from the use of PO.

Results and discussion

In order to directly and efficiently prepare PC from PG and CO₂, various bases including Cs₂CO₃, K₂CO₃, Bu₂SnO, KF/Al₂O₃, CsOH and KOH were screened as the catalysts at 120 °C, 8 MPa. To our delight, when Cs₂CO₃ and K₂CO₃ used as catalysts with PO as dehydrating agent, PC was obtained in 19% and 12% yield, respectively, being consistent with the basicity (entry 1, 2, Table 1), while other catalysts gave only trace amounts of PC (entries 3–6). Control experiments were conducted to gain deeper insight into the role of PO in PC synthesis. Without PG or PO, almost no PC was detected by GC (entry 7, 8), suggesting that PG and PO are inevitable for PC formation. Either PG or PO alone could not react with CO₂ to give PC under otherwise identical reaction conditions in this study. To distinguish PC produced from PG or PO, ethylene epoxide (EO) was chosen as a dehydrating agent because EC and PC can be unequivocally identified by GC technique. As a result, 24% of PC and 4% of EC were detected, indicating that the PC came mainly from PG rather than PO (entry 9). Herein, each PC yield in this study was obtained from two control experiments: a normal one to get the total PC yield from both PG and PO, another without PG to calculate the PC yield directly from PO, then PC yield from PG was obtained by subtracting such amount of PC from PO.

Table 1 PC synthesis from PG and CO₂ using epoxide as dehydrating agent^a

Entry	Catal.	Dehydrating agent (2 eq.)	Selectivity^b (%)	Yield^b (%)
a Reaction conditions: PG (2.5 mmol, 0.1902 g), catalyst (5 mol%), dehydrating agent (200 mol%, 5 mmol), reaction pressure (8 MPa), 120 °C, 12 h. b Determined by GC using biphenyl as internal standard. c Without PG. d GC yield of EC.
1	Cs2CO3	PO	>99	19
2	K2CO3	PO	>99	12
3	Bu3SnO	PO	>99	<1
4	KF/Al2O3	PO	>99	3
5	CsOH	PO	>99	3
6	KOH	PO	>99	<1
7^c	Cs2CO3	PO	>99	<1
8	Cs2CO3	—	>99	<1
9	Cs2CO3	EO	>99	24 (4^d)

---

As an inexpensive, non-toxic, nearly non-volatile and environmentally benign reaction medium,^29–31 PEG is expandable with CO₂,^32,33 thus leading to changes in its physical properties, such as lowered melting points³⁴ and lowered viscosity.³⁵ When PEG was selected as solvent, the PC yield directly from PG distinctly increased to 55% after subtracting the PC amount formed from PO under the same conditions (entry 2, Table 2), indicating the introduction of PEG is beneficial to the carboxylation of PG and CO₂ to produce PC when using PO as dehydrating agent. Typically, EO as a dehydrant afforded 44% yield of PC accompanying with 0.70 mmol of EC (entry 3). Accordingly, an epoxide could act as an efficient dehydrating agent to promote the carboxylation of PG with CO₂ to produce PC in PEG/CO₂ biphasic system. Notably, no additional by-product was detected in this process. Subsequently, the influence of PEG with different molecular weight ranging from 400 to 6000 was explored. When the reaction was conducted in PEG₄₀₀, PO was quantitatively converted to PC but only 7% yield of PC came from PG (entry 3, Table 2). As PEG molecular weight increased to 800, PC from PG was increased to 66% (entry 4, 5). However, further increasing PEG molecular weight to 6000 leaded to a decrease in PC yield (entry 6), presumably being ascribed to the decreasing in coordination effect and CO₂ solubility, as well as the increase in viscosity as PEG molecular weight increasing. Therefore, PEG₈₀₀ was chosen as the appropriate solvent for further investigation.

Table 2 PC synthesis from PG in PEG/CO₂ with epoxide as dehydrant^a

Entry	PEG	PEG (g)	PC amount^b (mmol)	PC yield from PG^c (%)
a Reaction conditions: PG (2.5 mmol, 0.1902 g), Cs2CO3 (5 mol%, 0.0407 g), PO (5 mmol, 0.2904 g), reaction pressure (8 MPa), 120 °C, 12 h. b GC yield with biphenyl as internal standard. c Obtained by subtracting the PC yield from PO under the same conditions. d Without PG.
1^d	PEG1000	0.5	0.44	—
2	PEG1000	0.5	1.86	55
3	PEG400	0.2	0.10	7
4	PEG600	0.3	1.59	54
5	PEG800	0.4	1.27	66
6	PEG6000	0.5	1.08	44

---

Then, the influence of different basic catalysts was examined in PEG₈₀₀, as shown in Table 3. Without any catalyst, no PC directly from PG was detected (entry 1). Alkali carbonates e.g. Cs₂CO₃ and K₂CO₃ gave relatively higher yields (entry 2, 3); whereas Na₂CO₃ was less effective (entry 4). In particular, Cs₂CO₃ showed lower activity than K₂CO₃ due to favourable coordination between PEG₈₀₀ and K⁺. Others, such as CaCO₃, MgO, CaO did not work at all under the given conditions (entries 5–7). Therefore, K₂CO₃ was proved to be the most appropriate catalyst for PC synthesis. Furthermore, increasing PEG₈₀₀ amount gave no significant rise in PC yield (entries 8–10). Hence, 0.4 g of PEG₈₀₀ was chosen as reaction medium. In addition, changing the reaction temperature to 100 °C or 140 °C resulted in adverse effect (entry 9, 10), presumably being ascribed to intermolecular dehydrating condensation of PG and PEG.²⁸

Table 3 Influence of catalyst and PEG amount for PC synthesis from PG and CO₂^a

Entry	Catalyst	PEG800 (g)	Yield^b^,^c (%)
a Reaction conditions: PG (2.5 mmol, 0.1902 g), PO (2 equiv. relative to PG, 0.2904 g), catalyst (5 mol%), in PEG800 under 8 MPa pressure, 120 °C for 12 h. b Determined by GC using biphenyl as internal standard. c PC yield relative to PG. d 100 °C. e 140 °C.
1	—	0.4	<1
2	Cs2CO3	0.4	66
3	K2CO3	0.4	74
4	Na2CO3	0.4	15
5	CaCO3	0.4	<1
6	MgO	0.4	<1
7	CaO	0.4	<1
8	K2CO3	0.5	75
9^d	K2CO3	0.4	38
10^e	K2CO3	0.4	<1

---

With those results in hand, we then investigated the effect of catalyst loading, reaction pressure and time on the reaction outcome as listed in Table 4. Reducing K₂CO₃ loading from 5 to 3 mol% dramatically leaded to decrease in PC yield (entry 2 vs. 1). However, when increased the catalyst amount to 10 mol%, no beneficial effect was shown (entry 3).

Table 4 Influence of catalyst loading, pressure and reaction time^a

Entry	K2CO3 (mol%)	Pressure (MPa)	Time (h)	Yield^b^,^c (%)
a Reaction conditions: PG (2.5 mmol, 0.1902 g), PO (2 equiv. to PG, 0.2904 g), in 0.4 g PEG800 with CO2 and K2CO3, 120 °C. b Determined by GC using biphenyl as internal standard. c Yield of PC relative to PG.
1	3	8	12	55
2	5	8	12	74
3	10	8	12	68
4	5	2	12	49
5	5	4	12	79
6	5	10	12	66
7	5	4	10	78
8	5	4	6	70

---

Furthermore, the dependence of PC yield on reaction pressure was carefully examined. PC yield gradually increased when raising reaction pressure from 2 to 4 MPa (entry 4, 5), being ascribed to CO₂ amount increase. However, too high reaction pressure would dilute the reactants concentration,²⁸ leading to a decrease in PC yield (entry 6). The effect of reaction time on PC synthesis was also explored. 78% and 70% yield of PC were obtained when the reactions stopped at 10 h and 6 h, respectively (entry 7, 8). Consequently, the most suitable reaction conditions for the carboxylation of PG with CO₂ using PO as dehydrate agent were 5 mol% K₂CO₃ in 0.4 g PEG₈₀₀ under 4 MPa pressure at 120 °C for 10 h.

The proposed mechanism for the present PC synthesis from PG and CO₂ in PEG/CO₂ biphasic system catalyzed by K₂CO₃ with PO as dehydrating agent was depicted in Scheme 4. The solubility of CO₂ and K₂CO₃ in PEG is excellent, and PEG can coordinate with K⁺ to enhance the basicity and nucleophilicity of the carbonate anion.^27,28,32,33 The deprotonation of PG by K₂CO₃ and subsequent insertion of CO₂ gives the carbonate intermediate.

Scheme 4 Proposed mechanism for PC formation from PG and CO₂ by using PO as dehydrating agent.

Then PC is generated along with the formation of an equivalent water. The water further reacts with PO, regenerating more PG; and thus both shifts the equilibrium to PC production. After the completion of the carboxylation of PG with CO₂, PC could be separated or subsequently reacts with methanol to produce DMC through transesterification reaction. Another reaction in this system is the direct cyclization of PO and CO₂ to afford PC.

Conclusions

In conclusion, we have developed a simple, ecologically safe, cost-effective and industrially feasible process for PC synthesis from PG with CO₂ using K₂CO₃ as catalyst in PEG₈₀₀/CO₂ biphasic system. 78% yield of PC from PG can be attained under mild reaction conditions (120 °C, 4 MPa). In this process, PO is utilized as dehydrating agent to react with co-produced water along with regenerating PG, both are favorable to shift the equilibrium to PC formation. Compared with common organic dehydrating agents, no additional by-product was generated in this process, resulting in workup procedure and economic benefits.

Experimental

General information

Alcohols, epoxides and PEGs were purchased from Aladdin. PG was dried with anhydrous Na₂SO₄, then distilled under reduced pressure prior to use. PEGs were dried in vacuum oven at 80 °C for two days and all the catalysts were dried in vacuo at 150 °C for 4 h. CO₂ was commercially available with a purity of 99.99%. Other reagents including diethyl ether, ethyl acetate, petroleum ether (60–90 °C) were purchased from Tianjin Guangfu Fine Chemical Research Institute without further purification. GC analyses were performed on Shimadzu GC-2014 equipped with capillary column (RTX-17 30 m × 0.25 μm) and flame ionization detector. GC-MS analyses were performed on Shimadzu 2014-QP2010 SE. NMR spectra were recorded on Bruker 400 in CDCl₃. ¹H and ¹³C NMR chemical shifts (δ) were given in ppm relative to CDCl₃ (7.26 ppm and 77.0 ppm).

General procedure for the K₂CO₃-promoted PC synthesis from PG and CO₂

A mixture of PG (2.5 mmol, 0.1902 g), potassium carbonate (17.3 mg, 0.125 mmol, 5 mol% relative to PG), epoxide (5 mmol) and PEG₈₀₀ (0.5 mmol, 0.4 g) were placed in a 25 mL autoclave equipped with inner glass tube. 2 MPa CO₂ was sequentially introduced into the autoclave and heated to 120 °C. The final pressure at the reaction temperature was regulated by introducing necessary amount of CO₂ and the mixture was stirred continuously for 12 h. After the reaction, the reactor was cooled to 0 °C in ice water. Then CO₂ was released slowly, the product was extracted with diethyl ether after depressurization. Then gas chromatograph analysis was performed with biphenyl as internal standard. The product was further identified using GC-MS by comparing retention times and fragmentation patterns with authentic samples and were also isolated by column chromatography on silica gel (200–300 mesh, petroleum ether (60–90 °C)/EtOAc = 3 : 1) and identified by NMR and MS measurement.

4-Methyl-1,3-dioxolan-2-one (PC)

Colorless oil; ¹H NMR (400 MHz, CDCl₃): δ = 4.93–4.77 (m, 1H, CH), 4.55 (t, J = 8.0 Hz, 1H, CH₂), 4.02 (t, J = 7.8 Hz, 1H, CH₂), 1.48 (d, J = 6.2 Hz, 3H, CH₃); ¹³C NMR (101 MHz, CDCl₃): δ = 155.0 (C), 73.5 (CH), 70.7 (CH₂), 19.5 (CH₃). EI-MS, m/z (%): 103.14 (26) [M]⁺, 57.03 (100).

1,3-Dioxolan-2-one (EC)

Colorless crystal; ¹H NMR (400 MHz, CDCl₃): δ = 4.50 (s, 4H, CH₂); ¹³C NMR (101 MHz, CDCl₃): δ = 155.5 (C), 64.6 (CH₂).

Acknowledgements

This work was financially supported by the National Natural Sciences Foundation of China, the “111” Project of the Ministry of Education of China (No. B06005), Specialized Research Fund for the Doctoral Program of Higher Education (20130031110013), MOE Innovation Team (IRT13022) of China.

References

1. T. Sakakura and K. Kohno, Chem. Commun., 2009, 1312–1330.
2. B. Schäffner, J. Holz, S. P. Verevkin and A. Börner, ChemSusChem, 2008, 1, 249–253.
3. J. Bayardon, J. Holz, B. Schäffner, V. Andrushko, S. Verevkin, A. Preetz and A. Börner, Angew. Chem., Int. Ed., 2007, 46, 5971–5974.
4. B. Schäffner, F. Schäffner, S. P. Verevkin and A. Börner, Chem. Rev., 2010, 110, 4554–4581.
5. S. Fukuoka, M. Kawamura, K. Komiya, M. Tojo, H. Hachiya, K. Hasegawa, M. Aminaka, H. Okamoto, I. Fukawa and S. Konno, Green Chem., 2003, 5, 497–507.
6. P. Tundo and M. Selva, Acc. Chem. Res., 2002, 35, 706–716.
7. M. A. Pacheco and C. L. Marshall, Energy Fuels, 1997, 11, 2–29.
8. Z.-Z. Yang, L.-N. He, X.-Y. Dou and S. Chanfreau, Tetrahedron Lett., 2010, 51, 2931–2934.
9. Y. Du, L.-N. He and D.-L. Kong, Catal. Commun., 2008, 9, 1754–1758.
10. D. Delledonne, F. Rivetti and U. Romano, Appl. Catal., A, 2001, 221, 241–251.
11. M. Honda, S. Kuno, S. Sonehara, K.-i. Fujimoto, K. Suzuki, Y. Nakagawa and K. Tomishige, ChemCatChem, 2011, 3, 365–370.
12. M. Honda, M. Tamura, Y. Nakagawa, S. Sonehara, K. Suzuki, K.-i. Fujimoto and K. Tomishige, ChemSusChem, 2013, 6, 1341–1344.
13. J.-S. Tian, C.-X. Miao, J.-Q. Wang, F. Cai, Y. Du, Y. Zhao and L.-N. He, Green Chem., 2007, 9, 566–571.
14. M. He, Y. Sun and B. Han, Angew. Chem., Int. Ed., 2013, 52, 9620–9633.
15. Z.-Z. Yang, Y.-N. Zhao, L.-N. He, J. Gao and Z.-S. Yin, Green Chem., 2012, 14, 519–527.
16. Y. Du, D.-L. Kong, H.-Y. Wang, F. Cai, J.-S. Tian, J.-Q. Wang and L.-N. He, J. Mol. Catal. A: Chem., 2005, 241, 233–237.
17. T. Sakakura, J.-C. Choi and H. Yasuda, Chem. Rev., 2007, 107, 2365–2387.
18. K. Yamaguchi, M. Matsushita and N. Mizuno, Angew. Chem., Int. Ed., 2004, 116, 1602–1606.
19. J. N. Moorthy and N. Singhal, J. Org. Chem., 2005, 70, 1926–1929.
20. E. Da Silva, W. Dayoub, G. Mignani, Y. Raoul and M. Lemaire, Catal. Commun., 2012, 29, 58–62.
21. K. Tomishige, H. Yasuda, Y. Yoshida and M. Nurunnabi, Catal. Lett., 2004, 95, 45–49.
22. K. Tomishige, H. Yasuda, Y. Yoshida, M. Nurunnabi, B. Li and K. Kunimori, Green Chem., 2004, 6, 206–214.
23. S. Huang, S. Liu, J. Li, N. Zhao, W. Wei and Y. Sun, Catal. Lett., 2006, 112, 187–191.
24. S. Huang, S. Liu, J. Li, N. Zhao, W. Wei and Y. Sun, Catal. Lett., 2007, 118, 290–294.
25. X. Zhao, N. Sun, S. Wang, F. Li and Y. Wang, Ind. Eng. Chem. Res., 2008, 47, 1365–1369.
26. S. Huang, J. Ma, J. Li, N. Zhao, W. Wei and Y. Sun, Catal. Commun., 2008, 9, 276–280.
27. A.-H. Liu, R. Ma, C. Song, Z.-Z. Yang, A. Yu, Y. Cai, L.-N. He, Y. N. Zhao, B. Yu and Q.-W. Song, Angew. Chem., Int. Ed., 2012, 51, 11306–11310.
28. D.-L. Kong, L.-N. He and J.-Q. Wang, Synlett, 2010, 1276–1280.
29. Z. Hou, N. Theyssen, A. Brinkmann and W. Leitner, Angew. Chem., Int. Ed., 2005, 44, 1346–1349.
30. J. Q. Wang, F. Cai, E. Wang and L. N. He, Green Chem., 2007, 9, 882–887.
31. Y. Du, Y. Wu, A.-H. Liu and L.-N. He, J. Org. Chem., 2008, 73, 4709–4712.
32. Z.-M. Xue, J.-L. Zhang, L. Peng, J.-S. Li, T.-C. Mu, B.-X. Han and G.-Y. Yang, Angew. Chem., Int. Ed., 2012, 51, 12325–12329.
33. R.-C. Luo, X.-T. Zhou, S.-Y. Chen, Y. Li, L. Zhou and H.-B. Ji, Green Chem., 2014, 16, 1496–1506.
34. D. J. Heldebrant and P. G. Jessop, J. Am. Chem. Soc., 2003, 125, 5600–5601.
35. S. Yeo and E. Kiran, J. Appl. Polym. Sci., 2000, 75, 306–315.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra04422k

---