J Mol Model (2015) 21:179 DOI 10.1007/s00894-015-2733-y
ORIGINAL PAPER
Theoretical investigation of the mechanism for the cycloaddition of CO2 to epoxides catalyzed by a magnesium(II) porphyrin complex Qin Wang 1 & Cai-Hong Guo 1 & Jianfeng Jia 1 & Hai-Shun Wu 1
Received: 8 April 2015 / Accepted: 8 June 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract The cycloaddition of CO2 to epoxides, catalyzed by Mg(TPP)/TBAI (TPP = tetraphenylporphyrin; TBAI = tetrabutylammonium iodide), was investigated using DFT methods. Epoxides with various substituents were studied to explore steric and electronic effects on the reaction mechanism. Computational results show that the cycloaddition proceeds according to a much easier mechanism in the presence of Mg(TPP) and TBAI than the mechanism that takes place when Mg(TPP) is used as the catalyst. A preference for the epoxide ringopening to occur at the methine (C α ) or methylene (Cβ) carbon was noted. The ring-closing step leading to the formation of a five-membered carbonate is predicted to determine the reaction rate. For alkylsubstituted epoxides, the β pathway is favorable since steric factors are dominant; for epoxides with a strongly electron-donating group and styrene oxide, the reaction is mainly controlled by electronic factors and proceeds along the α pathway. When the epoxide has a strongly electronwithdrawing group (CF3), both steric and electronic effects play important roles. The calculated reactivity of epoxides with CO2 catalyzed by Mg(TPP)/TBAI is in good agreement with that observed experimentally. Electronic supplementary material The online version of this article (doi:10.1007/s00894-015-2733-y) contains supplementary material, which is available to authorized users. * Cai-Hong Guo [email protected] * Hai-Shun Wu [email protected] 1
School of Chemistry and Materials Science, Shanxi Normal University, Linfen 041004, People’s Republic of China
Keywords Porphyrin . Carbon dioxide . Epoxide . Cycloaddition . DFT
Introduction The utilization of carbon dioxide (CO2) has attracted much attention in recent years due to the growing concern about global warming and the need to find new renewable carbon sources [1]. One reaction that utilizes CO2 as a carbon source is the cycloaddition of CO2 to epoxides [2–7], leading to useful cyclic carbonate products (Scheme 1). Cyclic carbonates are used as solvents and electrolytes in lithium batteries, and as starting materials for polymers, pharmaceuticals, and fine chemicals [8–13]. Various catalysts for this cycloaddition have been developed, such as quaternary ammonium salts [14, 15], alkali metal halides [16–18], ionic liquids [19–22], functional polymers [23, 24], and transition metal complexes [25–28]. However, most of these catalytic systems suffer from disadvantages such as low production rates, low catalyst reactivity, and harsh reaction conditions. Recently, Jing reported the cycloaddition of CO2 to epoxide using a biomimetic metalloporphyrin catalyst, which generated an excellent yield of cyclic carbonate in mild conditions [29]. Although there have been several theoretical studies of the mechanisms involved in the coupling of CO2 to epoxides using catalysts such as heterobimetallic Ru–Mn complexes [30], metal salen complexes [31, 32], ionic liquids [33, 34], N-heterocyclic carbenes [35], cobalt-substituted 12tungstenphosphate [36], and phenolic compounds [37], the reaction mechanisms involved in the metalloporphyrincatalyzed cycloaddition of CO2 to epoxide are still ambiguous. More recently, there has been an article reporting a mechanistic study of the cycloaddition of CO2 to epoxides catalyzed by bifunctional porphyrin catalysts [38], in which only
179
Page 2 of 9
Scheme 1 Catalyzed cycloaddition of CO2 to an epoxide, generating a cyclic carbonate
one pathway was calculated. We calculated all possible pathways for this reaction; moreover, we explored the influence of substituents on the epoxides on the reaction pathway. However, there was a big difference between the catalysts used in the experiments reported in [29] and [38]. Tetraalkylammonium bromide groups were attached to the phenyl ring of the porphyrin in the bifunctional porphyrin catalysts reported by Ema and coworkers [38], while tetrabutylammonium iodide (TBAI) and Mg(TPP) (TPP = tetraphenylporphyrin) were applied separately by Jing and coworkers [29]. Herein, we present a theoretical investigation of the cycloaddition of CO2 to epoxides catalyzed by the binary system Mg(TPP)/TBAI (see Scheme 2a) using the density functional theory (DFT) method. We also systematically investigated the reactivities of a variety of epoxides (Scheme 2b, epoxides A– E) with CO2.
Computational details All calculations were carried out with the M06 [39–41] DFT functional implemented in the Gaussian09 program software package [42]. Geometries were optimized with the BSI basis set system. The effective core potentials of Hay and Wadt with a double-ξ valence basis set (LANL2DZ [43, 44]) were used to describe the I atom in BSI. The standard 6-31G* [45] basis
Scheme 2 a Schematic diagram of the Mg(TPP) complex; b epoxides A–E
J Mol Model (2015) 21:179
set was used for all of the other atoms. Frequency calculations were performed at the same level of theory to determine whether the stationary points were true minima (no imaginary frequency) or transition states (one imaginary frequency). Zero-point vibrational corrections and thermal corrections to the Gibbs free energy were calculated at 298 K and a pressure of 1.4 MPa (the experimental conditions). All transition-state structures were further analyzed by carrying out intrinsic reaction coordinate (IRC) [46, 47] computations to verify that the structures did indeed connect two relevant minima. To obtain the refined energies, we performed single-point energy calculations for all species with a large basis set (BSII) at the M06 level. In BSII, the quadruple zeta valence def2QZVP [48] basis set was used for the I atom and the 6-311+ G(2d,P) basis set was used for the other atoms. To take solvent effects into account, we carried out single-point energy computations with the integral equation formalism polarizable continuum model (IEFPCM). The radii and non-electrostatic terms were taken from Truhlar and coworkers’ universal solvation model (SMD [49]). The IEFPCM parameters for Et2O, which is expected to have a dielectric constant similar to that of epoxide, was used. NBO analysis was carried out on some structures to aid interpretation [50]. We used unsubstituted porphyrin as a model for the catalyst Mg(TPP) (TPP = tetraphenylporphyrin). The mechanism of the reaction catalyzed by M(TPP) can be clarified sufficiently using this model, as confirmed by a previous study of C–H activation using DFT theory [51–53]. In order to strongly validate the reliability of the chosen level of theory, we performed specific studies of geometry optimization and energy computations (see the BElectronic supplementary material,^ ESM).
Results and discussion The possible mechanisms for the cycloaddition of epoxides to CO2, based on those described by Lu [54], Lau [55], and Jing [29], are illustrated in Scheme 3. In route I, the coordination of epoxide to Mg(TPP) activates the former for nucleophilic ring opening by I−. Subsequently, the insertion of CO2 into a metal–alkoxy complex results in a metal carbonate species, and an SN2-type ring-closing step then leads to the cyclic carbonate product and regenerates the catalyst. In route II, the opening of the epoxide ring, which is coordinated to Mg(TPP), takes place via nucleophilic attack by iodoformate instead of I−. A ring-closing step from species 7 then proceeds to give the cyclic carbonate. We first performed comprehensive computations for the two abovementioned possible mechanisms for the cycloaddition reaction of CO2 to propylene oxide (PO) reported by Jing [29]. To take the effect of entropy into account, we used the free energies (ΔG) of activation and reaction. In order to compare the potential energies with the reported energies,
J Mol Model (2015) 21:179
Page 3 of 9 179
Scheme 3 Proposed mechanisms for the coupling of epoxide and CO2 catalyzed by Mg(TPP)/TBAI
Some relative potential energies given in parentheses are referred in Fig. 5. For route I, we considered two branched pathways, Iα and Iβ. The main difference between these two routes is the initial attack of I− on C1 (route Iα) or C2 (route Iβ) of PO, resulting in the ring-opening of PO via scission of the C1–O or C2–O bond, respectively. Fig. 1 shows the optimized structures for the intermediates and transition states involved in route Iα. Initially, PO coordinates to the Lewis-acidic Mg center to form complex 1A, which lies 2.6 kcal mol−1 lower than the initial reactants. The Lewis-acidic Mg center can activate the C–O bond of the epoxide to facilitate ring-opening. The nucleophilic I− then attacks the activated C1–O epoxide bond to form the intermediate 2Aα via transition state TS(1A-2A)α, which has a barrier of 16.7 kcal/mol. In TS(1A-2A)α, the lengths of the C1–O and C1–I bonds are 1.859 and 2.923 Å, respectively. These geometric parameters along with the normal mode corresponding to the imaginary frequency 359.7i cm−1 of TS(1A-2A)α clearly indicate that the nucleophilic attack of I− leads to the scission of the C1–O bond. Subsequently, the insertion of CO2 into the Mg–O bond leads to the formation of the linear carbonate 3Aα. This step proceeds via transition state TS(2A-3A)α. Its imaginary frequency is 172.9i cm−1, which is associated with the formation of the C–O and Mg–O1 bonds and the scission of the Mg–O bond. There is only a small barrier of 4.5 kcal mol−1 for this step. The linear carbonate 3Aα undergoes an intramolecular cyclic SN2-type reaction to convert into the product-like intermediate 4A. The authentic transition state TS(3A-4A)α was located. In TS(3A-4A)α, the C1–I bond breaks due to O2 attacking from the back, and the forming C1–O2 bond has a length of 2.045 Å while the breaking C1–I bond has a length of 2.798 Å. The imaginary frequency of TS(3A-4A)α is 350.9i cm−1. This ring-closing step has a barrier of 18.6 kcal mol−1, and was
found to be the rate-determining step in this pathway. In addition, the product-coordinated intermediate 4A was calculated to be 5.0 kcal mol−1 more stable than the reactants. Finally, the dissociation of cyclic carbonate from 4A and the release of the Mg(TPP) complex allows further PO turnover. The overall reaction was calculated to be exergonic by 4.8 kcal mol−1. Since route Iβ is analogous to route Iα, we only present the optimized structures (Fig. 2) and the free-energy profile along route Iβ (Fig. 3) here. The energy barriers to the three steps were found to be 16.2, 9.0, and 16.0 kcal mol−1, respectively. The barriers of the ring-opening and ring-closing steps in route Iβ are 0.5 and 2.6 kcal mol−1 lower, respectively, than those in route Iα, partly because C2 exerts only a small steric effect without the methyl group in route Iβ. Similar to Cα attack, the ring-closing step in C β attack is the ratedetermining step, with a relative energy of 17.4 kcal mol−1. Now we consider route II. As shown in Scheme 3, the initial step is also the coordination of PO to the Mg(TPP) complex, leading to intermediate 1A (Figs. 4 and 5), from which the ring-opening of PO proceeds through an attack by iodoformate. We optimized the geometry of iodoformate (see Fig. S2 in the ESM). The length of I–C(CO2) is 3.43 Å, which is slightly shorter than the sum of the van der Waals radii of the relevant elements (3.75 Å for C–I) [56]. The Wiberg bond index of I–C(CO2) is 0.0503. Moreover, we did not find any stable structures for the attack by iodoformate on the intermediate 1A to form the ring-opening intermediate. Therefore, it is impossible for iodoformate to open the epoxide, so CO2 may directly attack intermediate 1A. This step associated with the scission of the C1–O bond produces the coordinated cyclic carbonate 5Aα, which lies 1.7 kcal mol−1 below the reactants. It proceeds via the transition state TS(1A-5A)α, which has a barrier of 40.4 kcal mol−1. In TS(1A-5A)α, the lengths of the forming C1–O2 and C–O bonds are 2.162 and 2.098 Å. The
179
Page 4 of 9
J Mol Model (2015) 21:179
Fig. 1 Optimized structures for the intermediates and transition states involved in route Iα. Bond distances are in angstroms
imaginary frequency is 300.9i cm−1, and the vibrational mode corresponds to large amplitudes of C1, O2, C and O in the desired direction. Here, we also considered the scission of the C2–O bond of PO through the direct attack of CO2 along route IIβ, which is analogous to route IIα. The calculated barriers show that route IIα is favored by 8.3 kcal mol−1 compared with route IIβ. The relative potential energy barrier of the Mg(TPP)-catalyzed reaction is 32.1 kcal mol−1, which is much less than that of the uncatalyzed cycloaddition, ∼55 kcal mol−1 [33, 57]. To rigorously account for the solvent effect, we performed geometry optimization in solution for the most critical points. The key structures along with selected parameters are presented in Fig. S3 in the ESM. It was found that the optimized structures in solution are very similar to the structures obtained in the gas phase. Also, the free-energy profile is shown in Fig. S4 in the ESM. The activation free energies for 3A → 4A via the α and β pathways were calculated to be 18.6 and 16.0 kcal mol−1 when optimized in the gas phase, respectively. The activation free energies for 3A → 4A via the α and β pathways were slightly reduced to 17.6 and 14.6 kcal mol−1, respectively, when optimized in solution. The calculated ΔΔG‡ value of 3.0 kcal mol−1 when optimized in solution is closer to the value (2.6 kcal mol−1) obtained for optimization in the gas phase. There were
no major changes to the molecular structures and relative free energies in this critical case. In order to understand the influence of the substrate epoxide on the reaction, we examined the reactions of various epoxides A–E that have different substituents. According to the computational results depicted in Table 1, the ring-closing process was found to be the rate-determining step regardless of the epoxide considered, whereas the preferred reaction pathway was dependent on the substrate. Route Iβ is preferred when epoxides A–C are used as substrates, but route Iα is favored when epoxides D and E are used. In contrast to the reaction of CO2 with A, the barriers of the transition states for B are significantly increased by the increased steric bulk of the tert-butyl substituent (R = tBu). In particular, TS(3B-4B)α (Fig. 6, relating to the ring-closing step) is highly destabilized by the steric effect of the tBu group. This indicates that more sterically hindered substrates are more difficult to transform, consistent with the experimental findings reported by Jing and coworkers [29]. Moreover, when using epoxides A and B as substrates, there is a clear increase in the free-energy difference for the ring-closing step TS(3A−1 4A)α/β → TS(3B-4B)α/β (1.6 → 4.2 kcal mol ), implying that route Iβ is the only path and that route Iα is not competitive in the coupling reaction of B with CO2.
J Mol Model (2015) 21:179
Page 5 of 9 179
Fig. 2 Optimized structures for the intermediates and transition states involved in route Iβ. Bond distances are in angstroms
To gain insight into electronic effects on cycloaddition, we chose two epoxides (C and D) with different electronic substituents to react with CO2. We found that the electronic effect plays an important role in transition state TS(3C-4C)α. If we were to assume that only steric effects play a role in TS(3C4C)α, the relative free-energy value of TS(3C-4C)α should lie between 19.0 kcal mol−1 of TS(3A-4A)α and 24.2 kcal mol−1 of TS(3B-4B)α. However, the relative energy of TS(3C-4C)α is 26.2 kcal mol−1, which indicates that the electronic effect is also important in route Iα. For epoxide 1C, which has a strongly electron-withdrawing group (CF3), route Iβ is preferred. The rate-determining step remains the ring-closing process, with an energy value of 17.6 kcal mol−1, which is almost equal to that for epoxide A −1 in route Iβ (17.4 kcal mol−1). The reaction free energy for the cycloaddition of epoxide C with Fig. 3 Free-energy profiles for the Mg(TPP)/TBAI-catalyzed cycloaddition of PO to CO2 via route I. Free energies are given in kcal mol−1
CO2 catalyzed by Mg(TPP) is −2.4 kcal mol−1, which shows that the reaction is less exergonic than that of epoxide A (−4.8 kcal mol−1). Thus, the cycloaddition of epoxide C to CO2 is thermodynamically less favorable than that of epoxide A to CO2, which requires experimental validation. Interestingly, for epoxide D, which has a strongly electrondonating group (MeO), the relative free-energy values are much lower than those for epoxide A, and the Iα pathway is preferred. In the transition state TS(3D-4D)α (Fig. 7), relating to the ring-closing step, the charge resulting from the release of iodide is stabilized by the resonance effect of the methoxyl group [58]. In route Iα for epoxide D, the ring-closing step has a relative energy of 10.9 kcal mol−1, which is 6.5 kcal mol−1 less than that for epoxide A in route Iβ. Meanwhile, the reaction free energy of the cycloaddtion of epoxide D with
179
J Mol Model (2015) 21:179
Page 6 of 9
Fig. 4 Optimized structures for the intermediates and transition states involved in route II. Bond distances are in angstroms
CO2 catalyzed by Mg(TPP) was predicted to be −7.3 kcal mol−1, which is lower than that for epoxide A (−4.8 kcal mol−1). This indicates that the cycloaddition of epoxide D to CO2 should be both kinetically and thermodynamically favored over the cycloaddition of epoxide A to CO2. The α pathway is also preferred for styrene oxide E. In the transition state TS(3E-4E)α, relating to the ring-closing step, the charge resulting from the release of iodide is stabilized by the resonance effect of the phenyl group [58]. In route Iα for epoxide E, the ring-closing step has an relative energy of 17.5 kcal mol−1, which is almost equal to the value of 17.4 kcal mol−1 for the Iβ pathway for epoxide A. The reaction free energy for the cycloaddition of epoxide E with CO2 catalyzed by Mg(TPP) is −2.6 kcal mol−1, while that for epoxide A is −4.8 kcal mol−1. This indicates that the cycloaddition of epoxide E to CO2 is thermodynamically disfavored compared to the cycloaddition
Fig. 5 Free-energy profiles for the Mg(TPP)/TBAI-catalyzed cycloaddition of PO to CO2 via route II. Free energies are given in kcal mol−1
of epoxide A to CO2, in accordance with the experimental results reported by Bai and coworkers [29]: they observed the cycloaddition of epoxide E to CO2 to be a little sluggish compared with the cycloaddition of epoxide A to CO2.
Conclusions The mechanisms for the Mg(TPP)/TBAI-catalyzed cycloaddition of CO2 to epoxides were studied using DFT calculations at the M06 level. Two possible mechanisms were Table 1 The relative Gibbs free-energy values for the transition states involved in route I
J Mol Model (2015) 21:179
Fig. 6 Optimized structures for the transition state TS(3–4) in the α and β pathways for epoxides B and C. Bond distances are in angstroms
considered and compared. The calculated results show that cycloaddition proceeds through an easier path that is synergistically catalyzed by Mg(TPP) and TBAI. This favored reaction mechanism includes four elementary steps: (i) coordination of the epoxide to Mg(TPP), (ii) nucleophilic ring-opening of the epoxide by I−, (iii) insertion of CO2 into a metal–alkoxy complex, and (iv) an SN2-type ring-closing step. The other possible mechanism for cycloaddition, involving iodoformate attack, is not feasible. In all of the reactions of interest, the rate-determining step is ring closure from metal carbonate species 3. However, the preferred reaction pathway (Iα or Iβ) is dependent on the substrate. Steric factors are predicted to dominate for alkyl-substituted epoxides (A and B). For substrates A and B, ring opening and ring closure tend to occur at the unsubstituted carbon atom (β pathway), and this preference becomes more obvious upon increasing the steric bulk of the substituent on the epoxide, as demonstrated by a clear increase in the energy difference for the ringclosing step TS(3A-4A)α/β → TS(3B-4B)α/β (1.6 → 4.2 kcal mol−1). The β pathway is also preferred by substrate C, an epoxide with a strongly electron-withdrawing group (CF3). For epoxide D (with a strongly electron-donating group) and styrene oxide E, the reaction is mainly controlled by electronic factors and the α pathway is preferred. The resonance effect may be
Page 7 of 9 179
Fig. 7 Optimized structures for the transition state TS(3–4) involved in the α and β pathways for epoxides D and E. Bond distances are in angstroms
responsible for stabilizing the charge resulting from the release of iodide. The calculated reactivities of these different epoxides with CO2, where the reactions are catalyzed by Mg(TPP)/TBAI, are in good agreement with the experimentally observed trend in reactivities. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (21373131 and 21203115), the Program for New Century Excellent Talents in University (NCET-121035), the Key Project of the Chinese Ministry of Education (212022), and the Research Fund for the Doctoral Program of Higher Education (20111404120004). Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest.
References 1.
Klaus S, Lehenmeier MW, Anderson CE, Rieger B (2011) Recent advances in CO2/epoxide copolymerization—new strategies and cooperative mechanisms. Coord Chem Rev 255(13–14):1460– 1479
179 2.
3.
4.
5.
6.
7.
8. 9. 10. 11.
12.
13.
14.
15.
16.
17.
18.
19. 20.
21.
22.
J Mol Model (2015) 21:179
Page 8 of 9 Kim HS, Kim JJ, Kim H, Jang HG (2003) Imidazolium zinc tetrahalide-catalyzed coupling reaction of CO2 and ethylene oxide or propylene oxide. J Catal 220(1):44–46 Xiao LF, Li FW, Peng HH, Xia CG (2006) Immobilized ionic liquid/zinc chloride: heterogeneous catalyst for synthesis of cyclic carbonates from carbon dioxide and epoxides. J Mol Catal A Chem 253(1–2):265–269 Xiao LF, Li FW, Xia CG (2005) An easily recoverable and efficient natural biopolymer-supported zinc chloride catalyst system for the chemical fixation of carbon dioxide to cyclic carbonate. Appl Catal A Gen 279(1–2):125–129 Sun J, Wang JQ, Cheng WG, Zhang JX, Li XH, Zhang SJ, She YB (2012) Chitosan functionalized ionic liquid as a recyclable biopolymer-supported catalyst for cycloaddition of CO2. Green Chem 14(3):654–660 Zhao YC, Yao CQ, Chen GW, Yuan Q (2013) Highly efficient synthesis of cyclic carbonate with CO2 catalyzed by ionic liquid in a microreactor. Green Chem 15(2):446–452 Foltran S, Alsarraf J, Robert F, Landais Y, Cloutet E, Cramail H, Tassaing T (2013) On the chemical fixation of supercritical carbon dioxide with epoxides catalyzed by ionic salts: an in situ FTIR and Raman study. Catal Sci Technol 3(4):1046–1055 Aresta M (2010) Carbon dioxide as chemical feedstock. Wiley, Weinheim Sakakura T, Choi J-C, Yasuda H (2007) Transformation of carbon dioxide. Chem Rev 107(6):2365–2387 Sakakura T, Kohno K (2009) The synthesis of organic carbonates from carbon dioxide. Chem Commun 11:1312–1330 Schäffner B, Schäffner F, Verevkin SP, Börner A (2010) Organic carbonates as solvents in synthesis and catalysis. Chem Rev 110(8): 4554–4581 Darensbourg DJ, Wilson SJ (2012) What’s new with CO2? Recent advances in its copolymerization with oxiranes. Green Chem 14(10):2665–2671 Lu X-B, Darensbourg DJ (2012) Cobalt catalysts for the coupling of CO2 and epoxides to provide polycarbonates and cyclic carbonates. Chem Soc Rev 41(4):1462–1484 Caló V, Nacci A, Monopoli A, Fanizzi A (2002) Cyclic carbonate formation from carbon dioxide and oxiranes in tetrabutylammonium halides as solvents and catalysts. Org Lett 4(15):2561–2563 Buckley BR, Patel AP, Wijayantha KGU (2011) Electrosynthesis of cyclic carbonates from epoxides and atmospheric pressure carbon dioxide. Chem Commun 47(43):11888–11890 Kihara N, Hara N, Endo T (1993) Catalytic activity of various salts in the reaction of 2,3-epoxypropyl phenyl ether and carbon dioxide under atmospheric pressure. J Org Chem 58(23):6198–6202 Guo L, Wang C, Luo X, Cui G, Li H (2010) Probing catalytic activity of halide salts by electrical conductivity in the coupling reaction of CO2 and propylene oxide. Chem Commun 46(32): 5960–5962 Aresta M, Dibenedetto A (2002) Carbon dioxide as building block for the synthesis of organic carbonates: behavior of homogeneous and heterogeneous catalysts in the oxidative carboxylation of olefins. J Mol Catal A Chem 182–183:399–409 Peng J, Deng Y (2001) Cycloaddition of carbon dioxide to propylene oxide catalyzed by ionic liquids. New J Chem 25(4):639–641 Yang H, Gu Y, Deng Y, Shi F (2002) Electrochemical activation of carbon dioxide in ionic liquid: synthesis of cyclic carbonates at mild reaction conditions. Chem Commun 3:274–275 Kawanami H, Sasaki A, Matsui K, Ikushima Y (2003) A rapid and effective synthesis of propylene carbonate using a supercritical CO2-ionic liquid system. Chem Commun 7:896–897 Sun J, S-i F, Arai M (2005) Development in the green synthesis of cyclic carbonate from carbon dioxide using ionic liquids. J Organomet Chem 690(15):3490–3497
23.
Xie Y, Zhang Z, Jiang T, He J, Han B, Wu T, Ding K (2007) CO2 cycloaddition reactions catalyzed by an ionic liquid grafted onto a highly cross-linked polymer matrix. Angew Chem Int Ed 46(38): 7255–7258 24. Xiong Y, Wang H, Wang R, Yan Y, Zheng B, Wang Y (2010) A facile one-step synthesis to cross-linked polymeric nanoparticles as highly active and selective catalysts for cycloaddition of CO2 to epoxides. Chem Commun 46(19):3399–3401 25. Kruper WJ, Dellar DD (1995) Catalytic formation of cyclic carbonates from epoxides and CO2 with chromium metalloporphyrinates. J Org Chem 60(3):725–727 26. Li F, Xia C, Xu L, Sun W, Chen G (2003) A novel and effective Ni complex catalyst system for the coupling reactions of carbon dioxide and epoxides. Chem Commun 16:2042–2043 27. Clegg W, Harrington RW, North M, Pasquale R (2010) Cyclic carbonate synthesis catalysed by bimetallic aluminium–salen complexes. Chem Eur J 16(23):6828–6843 28. Paddock RL, Nguyen ST (2004) Chiral (salen)Co-III catalyst for the synthesis of cyclic carbonates. Chem Commun 14:1622–1623 29. Bai DS, Duan SH, Hai L, Jing HW (2012) Carbon dioxide fixation by cycloaddition with epoxides, catalyzed by biomimetic metalloporphyrins. ChemCatChem 4(11):1752–1758 30. Man ML, Lam KC, Sit WN, Ng SM, Zhou ZY, Lin ZY, Lau CP (2006) Synthesis of heterobimetallic Ru–Mn complexes and the coupling reactions of epoxides with carbon dioxide catalyzed by these complexes. Chem Eur J 12(4):1004–1015 31. Wang TT, Xie Y, Deng WQ (2014) Reaction mechanism of epoxide cycloaddition to CO2 catalyzed by salen-M (M = Co, Al, Zn). J Phys Chem A 118(39):9239–9243 32. Castro-Gomez F, Salassa G, Kleij AW, Bo C (2013) A DFT study on the mechanism of the cycloaddition reaction of CO2 to epoxides catalyzed by Zn(salphen) complexes. Chem Eur J 19(20):6289– 6298 33. Sun H, Zhang D (2007) Density functional theory study on the cycloaddition of carbon dioxide with propylene oxide catalyzed by alkylmethylimidazolium chlorine ionic liquids. J Phys Chem A 111(32):8036–8043 34. Foltran S, Mereau R, Tassaing T (2014) Theoretical study on the chemical fixation of carbon dioxide with propylene oxide catalyzed by ammonium and guanidinium salts. Catal Sci Technol 4(6):1585– 1597 35. Ajitha MJ, Suresh CH (2011) NHC catalyzed CO2 fixation with epoxides: probable mechanisms reveal ter molecular pathway. Tetrahedron Lett 52(41):5403–5406 36. Chen F, Li X, Wang B, Xu T, Chen S-L, Liu P, Hu C (2012) Mechanism of the cycloaddition of carbon dioxide and epoxides catalyzed by cobalt-substituted 12-tungstenphosphate. Chem Eur J 18(32):9870–9876 37. Whiteoak CJ, Nova A, Maseras F, Kleij AW (2012) Merging sustainability with organocatalysis in the formation of organic carbonates by using CO2 as a feedstock. Chemsuschem 5(10):2032–2038 38. Ema T, Miyazaki Y, Shimonishi J, Maeda C, Hasegawa J (2014) Bifunctional porphyrin catalysts for the synthesis of cyclic carbonates from epoxides and CO2: structural optimization and mechanistic study. J Am Chem Soc 136(43):15270–15279 39. Zhao Y, Schultz NE, Truhlar DG (2006) Design of density functionals by combining the method of constraint satisfaction with parametrization for thermochemistry, thermochemical kinetics, and noncovalent interactions. J Chem Theory Comput 2(2):364– 382 40. Zhao Y, Truhlar DG (2006) Density functional for spectroscopy: no long-range self-interaction error, good performance for Rydberg and charge-transfer states, and better performance on average than B3LYP for ground states. J Phys Chem A 110(49):13126–13130 41. Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics,
J Mol Model (2015) 21:179
42. 43.
44.
45.
46. 47. 48.
49.
50.
noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor Chem Accounts 120(1–3): 215–241 Frisch MJ, Trucks GW, Schlegel HB et al (2010) Gaussian 09, revision C.01. Gaussian, Inc., Wallingford Hay PJ, Wadt WR (1985) Abinitio effective core potentials for molecular carculations—potentials for the transition-metal atoms Sc to Hg. J Chem Phys 82(1):270–283 Wadt WR, Hay PJ (1985) Abinitio effective core potentials for molecular calculations—potentials for main group elements Na to Bi. J Chem Phys 82(1):284–298 Harihara PC, Pople JA (1973) Influence of polarization functions on molecular-orbital hydrogenation energies. Theor Chim Acta 28(3):213–222 Fukui K (1970) Formulation of the reaction coordinate. J Phys Chem 74(23):4161–4163 Fukui K (1981) The path of chemical reactions—the IRC approach. Acc Chem Res 14(12):363–368 Weigend F, Furche F, Ahlrichs R (2003) Gaussian basis sets of quadruple zeta valence quality for atoms H–Kr. J Chem Phys 119(24):12753–12762 Marenich AV, Cramer CJ, Truhlar DG (2009) Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions. J Phys Chem B 113(18):6378–6396 Reed AE, Curtiss LA, Weinhold F (1988) Intermolecular interactions from a natural bond orbital, donor–acceptor viewpoint. Chem Rev 88(6):899–926
Page 9 of 9 179 51.
52.
53.
54.
55.
56.
57.
58.
Guo Z, Guan XG, Huang JS, Tsui WM, Lin ZY, Che CM (2013) Bis(sulfonylimide)ruthenium(VI) porphyrins: X-ray crystal structure and mechanism of C-H bond amination by density functional theory calculations. Chem Eur J 19(34):11320–11331 Lu HJ, Subbarayan V, Tao JR, Zhang XP (2010) Cobalt(II)-catalyzed intermolecular benzylic C–H amination with 2,2,2trichloroethoxycarbonyl azide (TrocN(3)). Organometallics 29(2): 389–393 Lyaskovskyy V, Suarez AIO, Lu HJ, Jiang HL, Zhang XP, de Bruin B (2011) Mechanism of cobalt(II) porphyrin-catalyzed C–H amination with organic azides: radical nature and H-atom abstraction ability of the key cobalt(III)–nitrene intermediates. J Am Chem Soc 133(31):12264–12273 Lu X-B, Liang B, Zhang Y-J, Tian Y-Z, Wang Y-M, Bai C-X, Wang H, Zhang R (2004) Asymmetric catalysis with CO2: direct synthesis of optically active propylene carbonate from racemic epoxides. J Am Chem Soc 126(12):3732–3733 Sit WN, Ng SM, Kwong KY, Lau CP (2005) Coupling reactions of CO2 with neat epoxides catalyzed by PPN salts to yield cyclic carbonates. J Org Chem 70(21):8583–8586 Rowland RS, Taylor R (1996) Intermolecular nonbonded contact distances in organic crystal structures: comparison with distances expected from van der Waals radii. J Phys Chem 100(18):7384–7391 Guo CH, Wu HS, Zhang XM, Song JY, Zhang X (2009) A comprehensive theoretical study on the coupling reaction mechanism of propylene oxide with carbon dioxide catalyzed by copper(I) cyanomethyl. J Phys Chem A 113(24):6710–6723 Anslyn EV, Dougherty DA (2006) Modern physical organic chemistry. University Science Books, Herndon
ORIGINAL PAPER
Theoretical investigation of the mechanism for the cycloaddition of CO2 to epoxides catalyzed by a magnesium(II) porphyrin complex Qin Wang 1 & Cai-Hong Guo 1 & Jianfeng Jia 1 & Hai-Shun Wu 1
Received: 8 April 2015 / Accepted: 8 June 2015 # Springer-Verlag Berlin Heidelberg 2015
Abstract The cycloaddition of CO2 to epoxides, catalyzed by Mg(TPP)/TBAI (TPP = tetraphenylporphyrin; TBAI = tetrabutylammonium iodide), was investigated using DFT methods. Epoxides with various substituents were studied to explore steric and electronic effects on the reaction mechanism. Computational results show that the cycloaddition proceeds according to a much easier mechanism in the presence of Mg(TPP) and TBAI than the mechanism that takes place when Mg(TPP) is used as the catalyst. A preference for the epoxide ringopening to occur at the methine (C α ) or methylene (Cβ) carbon was noted. The ring-closing step leading to the formation of a five-membered carbonate is predicted to determine the reaction rate. For alkylsubstituted epoxides, the β pathway is favorable since steric factors are dominant; for epoxides with a strongly electron-donating group and styrene oxide, the reaction is mainly controlled by electronic factors and proceeds along the α pathway. When the epoxide has a strongly electronwithdrawing group (CF3), both steric and electronic effects play important roles. The calculated reactivity of epoxides with CO2 catalyzed by Mg(TPP)/TBAI is in good agreement with that observed experimentally. Electronic supplementary material The online version of this article (doi:10.1007/s00894-015-2733-y) contains supplementary material, which is available to authorized users. * Cai-Hong Guo [email protected] * Hai-Shun Wu [email protected] 1
School of Chemistry and Materials Science, Shanxi Normal University, Linfen 041004, People’s Republic of China
Keywords Porphyrin . Carbon dioxide . Epoxide . Cycloaddition . DFT
Introduction The utilization of carbon dioxide (CO2) has attracted much attention in recent years due to the growing concern about global warming and the need to find new renewable carbon sources [1]. One reaction that utilizes CO2 as a carbon source is the cycloaddition of CO2 to epoxides [2–7], leading to useful cyclic carbonate products (Scheme 1). Cyclic carbonates are used as solvents and electrolytes in lithium batteries, and as starting materials for polymers, pharmaceuticals, and fine chemicals [8–13]. Various catalysts for this cycloaddition have been developed, such as quaternary ammonium salts [14, 15], alkali metal halides [16–18], ionic liquids [19–22], functional polymers [23, 24], and transition metal complexes [25–28]. However, most of these catalytic systems suffer from disadvantages such as low production rates, low catalyst reactivity, and harsh reaction conditions. Recently, Jing reported the cycloaddition of CO2 to epoxide using a biomimetic metalloporphyrin catalyst, which generated an excellent yield of cyclic carbonate in mild conditions [29]. Although there have been several theoretical studies of the mechanisms involved in the coupling of CO2 to epoxides using catalysts such as heterobimetallic Ru–Mn complexes [30], metal salen complexes [31, 32], ionic liquids [33, 34], N-heterocyclic carbenes [35], cobalt-substituted 12tungstenphosphate [36], and phenolic compounds [37], the reaction mechanisms involved in the metalloporphyrincatalyzed cycloaddition of CO2 to epoxide are still ambiguous. More recently, there has been an article reporting a mechanistic study of the cycloaddition of CO2 to epoxides catalyzed by bifunctional porphyrin catalysts [38], in which only
179
Page 2 of 9
Scheme 1 Catalyzed cycloaddition of CO2 to an epoxide, generating a cyclic carbonate
one pathway was calculated. We calculated all possible pathways for this reaction; moreover, we explored the influence of substituents on the epoxides on the reaction pathway. However, there was a big difference between the catalysts used in the experiments reported in [29] and [38]. Tetraalkylammonium bromide groups were attached to the phenyl ring of the porphyrin in the bifunctional porphyrin catalysts reported by Ema and coworkers [38], while tetrabutylammonium iodide (TBAI) and Mg(TPP) (TPP = tetraphenylporphyrin) were applied separately by Jing and coworkers [29]. Herein, we present a theoretical investigation of the cycloaddition of CO2 to epoxides catalyzed by the binary system Mg(TPP)/TBAI (see Scheme 2a) using the density functional theory (DFT) method. We also systematically investigated the reactivities of a variety of epoxides (Scheme 2b, epoxides A– E) with CO2.
Computational details All calculations were carried out with the M06 [39–41] DFT functional implemented in the Gaussian09 program software package [42]. Geometries were optimized with the BSI basis set system. The effective core potentials of Hay and Wadt with a double-ξ valence basis set (LANL2DZ [43, 44]) were used to describe the I atom in BSI. The standard 6-31G* [45] basis
Scheme 2 a Schematic diagram of the Mg(TPP) complex; b epoxides A–E
J Mol Model (2015) 21:179
set was used for all of the other atoms. Frequency calculations were performed at the same level of theory to determine whether the stationary points were true minima (no imaginary frequency) or transition states (one imaginary frequency). Zero-point vibrational corrections and thermal corrections to the Gibbs free energy were calculated at 298 K and a pressure of 1.4 MPa (the experimental conditions). All transition-state structures were further analyzed by carrying out intrinsic reaction coordinate (IRC) [46, 47] computations to verify that the structures did indeed connect two relevant minima. To obtain the refined energies, we performed single-point energy calculations for all species with a large basis set (BSII) at the M06 level. In BSII, the quadruple zeta valence def2QZVP [48] basis set was used for the I atom and the 6-311+ G(2d,P) basis set was used for the other atoms. To take solvent effects into account, we carried out single-point energy computations with the integral equation formalism polarizable continuum model (IEFPCM). The radii and non-electrostatic terms were taken from Truhlar and coworkers’ universal solvation model (SMD [49]). The IEFPCM parameters for Et2O, which is expected to have a dielectric constant similar to that of epoxide, was used. NBO analysis was carried out on some structures to aid interpretation [50]. We used unsubstituted porphyrin as a model for the catalyst Mg(TPP) (TPP = tetraphenylporphyrin). The mechanism of the reaction catalyzed by M(TPP) can be clarified sufficiently using this model, as confirmed by a previous study of C–H activation using DFT theory [51–53]. In order to strongly validate the reliability of the chosen level of theory, we performed specific studies of geometry optimization and energy computations (see the BElectronic supplementary material,^ ESM).
Results and discussion The possible mechanisms for the cycloaddition of epoxides to CO2, based on those described by Lu [54], Lau [55], and Jing [29], are illustrated in Scheme 3. In route I, the coordination of epoxide to Mg(TPP) activates the former for nucleophilic ring opening by I−. Subsequently, the insertion of CO2 into a metal–alkoxy complex results in a metal carbonate species, and an SN2-type ring-closing step then leads to the cyclic carbonate product and regenerates the catalyst. In route II, the opening of the epoxide ring, which is coordinated to Mg(TPP), takes place via nucleophilic attack by iodoformate instead of I−. A ring-closing step from species 7 then proceeds to give the cyclic carbonate. We first performed comprehensive computations for the two abovementioned possible mechanisms for the cycloaddition reaction of CO2 to propylene oxide (PO) reported by Jing [29]. To take the effect of entropy into account, we used the free energies (ΔG) of activation and reaction. In order to compare the potential energies with the reported energies,
J Mol Model (2015) 21:179
Page 3 of 9 179
Scheme 3 Proposed mechanisms for the coupling of epoxide and CO2 catalyzed by Mg(TPP)/TBAI
Some relative potential energies given in parentheses are referred in Fig. 5. For route I, we considered two branched pathways, Iα and Iβ. The main difference between these two routes is the initial attack of I− on C1 (route Iα) or C2 (route Iβ) of PO, resulting in the ring-opening of PO via scission of the C1–O or C2–O bond, respectively. Fig. 1 shows the optimized structures for the intermediates and transition states involved in route Iα. Initially, PO coordinates to the Lewis-acidic Mg center to form complex 1A, which lies 2.6 kcal mol−1 lower than the initial reactants. The Lewis-acidic Mg center can activate the C–O bond of the epoxide to facilitate ring-opening. The nucleophilic I− then attacks the activated C1–O epoxide bond to form the intermediate 2Aα via transition state TS(1A-2A)α, which has a barrier of 16.7 kcal/mol. In TS(1A-2A)α, the lengths of the C1–O and C1–I bonds are 1.859 and 2.923 Å, respectively. These geometric parameters along with the normal mode corresponding to the imaginary frequency 359.7i cm−1 of TS(1A-2A)α clearly indicate that the nucleophilic attack of I− leads to the scission of the C1–O bond. Subsequently, the insertion of CO2 into the Mg–O bond leads to the formation of the linear carbonate 3Aα. This step proceeds via transition state TS(2A-3A)α. Its imaginary frequency is 172.9i cm−1, which is associated with the formation of the C–O and Mg–O1 bonds and the scission of the Mg–O bond. There is only a small barrier of 4.5 kcal mol−1 for this step. The linear carbonate 3Aα undergoes an intramolecular cyclic SN2-type reaction to convert into the product-like intermediate 4A. The authentic transition state TS(3A-4A)α was located. In TS(3A-4A)α, the C1–I bond breaks due to O2 attacking from the back, and the forming C1–O2 bond has a length of 2.045 Å while the breaking C1–I bond has a length of 2.798 Å. The imaginary frequency of TS(3A-4A)α is 350.9i cm−1. This ring-closing step has a barrier of 18.6 kcal mol−1, and was
found to be the rate-determining step in this pathway. In addition, the product-coordinated intermediate 4A was calculated to be 5.0 kcal mol−1 more stable than the reactants. Finally, the dissociation of cyclic carbonate from 4A and the release of the Mg(TPP) complex allows further PO turnover. The overall reaction was calculated to be exergonic by 4.8 kcal mol−1. Since route Iβ is analogous to route Iα, we only present the optimized structures (Fig. 2) and the free-energy profile along route Iβ (Fig. 3) here. The energy barriers to the three steps were found to be 16.2, 9.0, and 16.0 kcal mol−1, respectively. The barriers of the ring-opening and ring-closing steps in route Iβ are 0.5 and 2.6 kcal mol−1 lower, respectively, than those in route Iα, partly because C2 exerts only a small steric effect without the methyl group in route Iβ. Similar to Cα attack, the ring-closing step in C β attack is the ratedetermining step, with a relative energy of 17.4 kcal mol−1. Now we consider route II. As shown in Scheme 3, the initial step is also the coordination of PO to the Mg(TPP) complex, leading to intermediate 1A (Figs. 4 and 5), from which the ring-opening of PO proceeds through an attack by iodoformate. We optimized the geometry of iodoformate (see Fig. S2 in the ESM). The length of I–C(CO2) is 3.43 Å, which is slightly shorter than the sum of the van der Waals radii of the relevant elements (3.75 Å for C–I) [56]. The Wiberg bond index of I–C(CO2) is 0.0503. Moreover, we did not find any stable structures for the attack by iodoformate on the intermediate 1A to form the ring-opening intermediate. Therefore, it is impossible for iodoformate to open the epoxide, so CO2 may directly attack intermediate 1A. This step associated with the scission of the C1–O bond produces the coordinated cyclic carbonate 5Aα, which lies 1.7 kcal mol−1 below the reactants. It proceeds via the transition state TS(1A-5A)α, which has a barrier of 40.4 kcal mol−1. In TS(1A-5A)α, the lengths of the forming C1–O2 and C–O bonds are 2.162 and 2.098 Å. The
179
Page 4 of 9
J Mol Model (2015) 21:179
Fig. 1 Optimized structures for the intermediates and transition states involved in route Iα. Bond distances are in angstroms
imaginary frequency is 300.9i cm−1, and the vibrational mode corresponds to large amplitudes of C1, O2, C and O in the desired direction. Here, we also considered the scission of the C2–O bond of PO through the direct attack of CO2 along route IIβ, which is analogous to route IIα. The calculated barriers show that route IIα is favored by 8.3 kcal mol−1 compared with route IIβ. The relative potential energy barrier of the Mg(TPP)-catalyzed reaction is 32.1 kcal mol−1, which is much less than that of the uncatalyzed cycloaddition, ∼55 kcal mol−1 [33, 57]. To rigorously account for the solvent effect, we performed geometry optimization in solution for the most critical points. The key structures along with selected parameters are presented in Fig. S3 in the ESM. It was found that the optimized structures in solution are very similar to the structures obtained in the gas phase. Also, the free-energy profile is shown in Fig. S4 in the ESM. The activation free energies for 3A → 4A via the α and β pathways were calculated to be 18.6 and 16.0 kcal mol−1 when optimized in the gas phase, respectively. The activation free energies for 3A → 4A via the α and β pathways were slightly reduced to 17.6 and 14.6 kcal mol−1, respectively, when optimized in solution. The calculated ΔΔG‡ value of 3.0 kcal mol−1 when optimized in solution is closer to the value (2.6 kcal mol−1) obtained for optimization in the gas phase. There were
no major changes to the molecular structures and relative free energies in this critical case. In order to understand the influence of the substrate epoxide on the reaction, we examined the reactions of various epoxides A–E that have different substituents. According to the computational results depicted in Table 1, the ring-closing process was found to be the rate-determining step regardless of the epoxide considered, whereas the preferred reaction pathway was dependent on the substrate. Route Iβ is preferred when epoxides A–C are used as substrates, but route Iα is favored when epoxides D and E are used. In contrast to the reaction of CO2 with A, the barriers of the transition states for B are significantly increased by the increased steric bulk of the tert-butyl substituent (R = tBu). In particular, TS(3B-4B)α (Fig. 6, relating to the ring-closing step) is highly destabilized by the steric effect of the tBu group. This indicates that more sterically hindered substrates are more difficult to transform, consistent with the experimental findings reported by Jing and coworkers [29]. Moreover, when using epoxides A and B as substrates, there is a clear increase in the free-energy difference for the ring-closing step TS(3A−1 4A)α/β → TS(3B-4B)α/β (1.6 → 4.2 kcal mol ), implying that route Iβ is the only path and that route Iα is not competitive in the coupling reaction of B with CO2.
J Mol Model (2015) 21:179
Page 5 of 9 179
Fig. 2 Optimized structures for the intermediates and transition states involved in route Iβ. Bond distances are in angstroms
To gain insight into electronic effects on cycloaddition, we chose two epoxides (C and D) with different electronic substituents to react with CO2. We found that the electronic effect plays an important role in transition state TS(3C-4C)α. If we were to assume that only steric effects play a role in TS(3C4C)α, the relative free-energy value of TS(3C-4C)α should lie between 19.0 kcal mol−1 of TS(3A-4A)α and 24.2 kcal mol−1 of TS(3B-4B)α. However, the relative energy of TS(3C-4C)α is 26.2 kcal mol−1, which indicates that the electronic effect is also important in route Iα. For epoxide 1C, which has a strongly electron-withdrawing group (CF3), route Iβ is preferred. The rate-determining step remains the ring-closing process, with an energy value of 17.6 kcal mol−1, which is almost equal to that for epoxide A −1 in route Iβ (17.4 kcal mol−1). The reaction free energy for the cycloaddition of epoxide C with Fig. 3 Free-energy profiles for the Mg(TPP)/TBAI-catalyzed cycloaddition of PO to CO2 via route I. Free energies are given in kcal mol−1
CO2 catalyzed by Mg(TPP) is −2.4 kcal mol−1, which shows that the reaction is less exergonic than that of epoxide A (−4.8 kcal mol−1). Thus, the cycloaddition of epoxide C to CO2 is thermodynamically less favorable than that of epoxide A to CO2, which requires experimental validation. Interestingly, for epoxide D, which has a strongly electrondonating group (MeO), the relative free-energy values are much lower than those for epoxide A, and the Iα pathway is preferred. In the transition state TS(3D-4D)α (Fig. 7), relating to the ring-closing step, the charge resulting from the release of iodide is stabilized by the resonance effect of the methoxyl group [58]. In route Iα for epoxide D, the ring-closing step has a relative energy of 10.9 kcal mol−1, which is 6.5 kcal mol−1 less than that for epoxide A in route Iβ. Meanwhile, the reaction free energy of the cycloaddtion of epoxide D with
179
J Mol Model (2015) 21:179
Page 6 of 9
Fig. 4 Optimized structures for the intermediates and transition states involved in route II. Bond distances are in angstroms
CO2 catalyzed by Mg(TPP) was predicted to be −7.3 kcal mol−1, which is lower than that for epoxide A (−4.8 kcal mol−1). This indicates that the cycloaddition of epoxide D to CO2 should be both kinetically and thermodynamically favored over the cycloaddition of epoxide A to CO2. The α pathway is also preferred for styrene oxide E. In the transition state TS(3E-4E)α, relating to the ring-closing step, the charge resulting from the release of iodide is stabilized by the resonance effect of the phenyl group [58]. In route Iα for epoxide E, the ring-closing step has an relative energy of 17.5 kcal mol−1, which is almost equal to the value of 17.4 kcal mol−1 for the Iβ pathway for epoxide A. The reaction free energy for the cycloaddition of epoxide E with CO2 catalyzed by Mg(TPP) is −2.6 kcal mol−1, while that for epoxide A is −4.8 kcal mol−1. This indicates that the cycloaddition of epoxide E to CO2 is thermodynamically disfavored compared to the cycloaddition
Fig. 5 Free-energy profiles for the Mg(TPP)/TBAI-catalyzed cycloaddition of PO to CO2 via route II. Free energies are given in kcal mol−1
of epoxide A to CO2, in accordance with the experimental results reported by Bai and coworkers [29]: they observed the cycloaddition of epoxide E to CO2 to be a little sluggish compared with the cycloaddition of epoxide A to CO2.
Conclusions The mechanisms for the Mg(TPP)/TBAI-catalyzed cycloaddition of CO2 to epoxides were studied using DFT calculations at the M06 level. Two possible mechanisms were Table 1 The relative Gibbs free-energy values for the transition states involved in route I
J Mol Model (2015) 21:179
Fig. 6 Optimized structures for the transition state TS(3–4) in the α and β pathways for epoxides B and C. Bond distances are in angstroms
considered and compared. The calculated results show that cycloaddition proceeds through an easier path that is synergistically catalyzed by Mg(TPP) and TBAI. This favored reaction mechanism includes four elementary steps: (i) coordination of the epoxide to Mg(TPP), (ii) nucleophilic ring-opening of the epoxide by I−, (iii) insertion of CO2 into a metal–alkoxy complex, and (iv) an SN2-type ring-closing step. The other possible mechanism for cycloaddition, involving iodoformate attack, is not feasible. In all of the reactions of interest, the rate-determining step is ring closure from metal carbonate species 3. However, the preferred reaction pathway (Iα or Iβ) is dependent on the substrate. Steric factors are predicted to dominate for alkyl-substituted epoxides (A and B). For substrates A and B, ring opening and ring closure tend to occur at the unsubstituted carbon atom (β pathway), and this preference becomes more obvious upon increasing the steric bulk of the substituent on the epoxide, as demonstrated by a clear increase in the energy difference for the ringclosing step TS(3A-4A)α/β → TS(3B-4B)α/β (1.6 → 4.2 kcal mol−1). The β pathway is also preferred by substrate C, an epoxide with a strongly electron-withdrawing group (CF3). For epoxide D (with a strongly electron-donating group) and styrene oxide E, the reaction is mainly controlled by electronic factors and the α pathway is preferred. The resonance effect may be
Page 7 of 9 179
Fig. 7 Optimized structures for the transition state TS(3–4) involved in the α and β pathways for epoxides D and E. Bond distances are in angstroms
responsible for stabilizing the charge resulting from the release of iodide. The calculated reactivities of these different epoxides with CO2, where the reactions are catalyzed by Mg(TPP)/TBAI, are in good agreement with the experimentally observed trend in reactivities. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (21373131 and 21203115), the Program for New Century Excellent Talents in University (NCET-121035), the Key Project of the Chinese Ministry of Education (212022), and the Research Fund for the Doctoral Program of Higher Education (20111404120004). Compliance with ethical standards Conflict of interest The authors declare that they have no conflict of interest.
References 1.
Klaus S, Lehenmeier MW, Anderson CE, Rieger B (2011) Recent advances in CO2/epoxide copolymerization—new strategies and cooperative mechanisms. Coord Chem Rev 255(13–14):1460– 1479
179 2.
3.
4.
5.
6.
7.
8. 9. 10. 11.
12.
13.
14.
15.
16.
17.
18.
19. 20.
21.
22.
J Mol Model (2015) 21:179
Page 8 of 9 Kim HS, Kim JJ, Kim H, Jang HG (2003) Imidazolium zinc tetrahalide-catalyzed coupling reaction of CO2 and ethylene oxide or propylene oxide. J Catal 220(1):44–46 Xiao LF, Li FW, Peng HH, Xia CG (2006) Immobilized ionic liquid/zinc chloride: heterogeneous catalyst for synthesis of cyclic carbonates from carbon dioxide and epoxides. J Mol Catal A Chem 253(1–2):265–269 Xiao LF, Li FW, Xia CG (2005) An easily recoverable and efficient natural biopolymer-supported zinc chloride catalyst system for the chemical fixation of carbon dioxide to cyclic carbonate. Appl Catal A Gen 279(1–2):125–129 Sun J, Wang JQ, Cheng WG, Zhang JX, Li XH, Zhang SJ, She YB (2012) Chitosan functionalized ionic liquid as a recyclable biopolymer-supported catalyst for cycloaddition of CO2. Green Chem 14(3):654–660 Zhao YC, Yao CQ, Chen GW, Yuan Q (2013) Highly efficient synthesis of cyclic carbonate with CO2 catalyzed by ionic liquid in a microreactor. Green Chem 15(2):446–452 Foltran S, Alsarraf J, Robert F, Landais Y, Cloutet E, Cramail H, Tassaing T (2013) On the chemical fixation of supercritical carbon dioxide with epoxides catalyzed by ionic salts: an in situ FTIR and Raman study. Catal Sci Technol 3(4):1046–1055 Aresta M (2010) Carbon dioxide as chemical feedstock. Wiley, Weinheim Sakakura T, Choi J-C, Yasuda H (2007) Transformation of carbon dioxide. Chem Rev 107(6):2365–2387 Sakakura T, Kohno K (2009) The synthesis of organic carbonates from carbon dioxide. Chem Commun 11:1312–1330 Schäffner B, Schäffner F, Verevkin SP, Börner A (2010) Organic carbonates as solvents in synthesis and catalysis. Chem Rev 110(8): 4554–4581 Darensbourg DJ, Wilson SJ (2012) What’s new with CO2? Recent advances in its copolymerization with oxiranes. Green Chem 14(10):2665–2671 Lu X-B, Darensbourg DJ (2012) Cobalt catalysts for the coupling of CO2 and epoxides to provide polycarbonates and cyclic carbonates. Chem Soc Rev 41(4):1462–1484 Caló V, Nacci A, Monopoli A, Fanizzi A (2002) Cyclic carbonate formation from carbon dioxide and oxiranes in tetrabutylammonium halides as solvents and catalysts. Org Lett 4(15):2561–2563 Buckley BR, Patel AP, Wijayantha KGU (2011) Electrosynthesis of cyclic carbonates from epoxides and atmospheric pressure carbon dioxide. Chem Commun 47(43):11888–11890 Kihara N, Hara N, Endo T (1993) Catalytic activity of various salts in the reaction of 2,3-epoxypropyl phenyl ether and carbon dioxide under atmospheric pressure. J Org Chem 58(23):6198–6202 Guo L, Wang C, Luo X, Cui G, Li H (2010) Probing catalytic activity of halide salts by electrical conductivity in the coupling reaction of CO2 and propylene oxide. Chem Commun 46(32): 5960–5962 Aresta M, Dibenedetto A (2002) Carbon dioxide as building block for the synthesis of organic carbonates: behavior of homogeneous and heterogeneous catalysts in the oxidative carboxylation of olefins. J Mol Catal A Chem 182–183:399–409 Peng J, Deng Y (2001) Cycloaddition of carbon dioxide to propylene oxide catalyzed by ionic liquids. New J Chem 25(4):639–641 Yang H, Gu Y, Deng Y, Shi F (2002) Electrochemical activation of carbon dioxide in ionic liquid: synthesis of cyclic carbonates at mild reaction conditions. Chem Commun 3:274–275 Kawanami H, Sasaki A, Matsui K, Ikushima Y (2003) A rapid and effective synthesis of propylene carbonate using a supercritical CO2-ionic liquid system. Chem Commun 7:896–897 Sun J, S-i F, Arai M (2005) Development in the green synthesis of cyclic carbonate from carbon dioxide using ionic liquids. J Organomet Chem 690(15):3490–3497
23.
Xie Y, Zhang Z, Jiang T, He J, Han B, Wu T, Ding K (2007) CO2 cycloaddition reactions catalyzed by an ionic liquid grafted onto a highly cross-linked polymer matrix. Angew Chem Int Ed 46(38): 7255–7258 24. Xiong Y, Wang H, Wang R, Yan Y, Zheng B, Wang Y (2010) A facile one-step synthesis to cross-linked polymeric nanoparticles as highly active and selective catalysts for cycloaddition of CO2 to epoxides. Chem Commun 46(19):3399–3401 25. Kruper WJ, Dellar DD (1995) Catalytic formation of cyclic carbonates from epoxides and CO2 with chromium metalloporphyrinates. J Org Chem 60(3):725–727 26. Li F, Xia C, Xu L, Sun W, Chen G (2003) A novel and effective Ni complex catalyst system for the coupling reactions of carbon dioxide and epoxides. Chem Commun 16:2042–2043 27. Clegg W, Harrington RW, North M, Pasquale R (2010) Cyclic carbonate synthesis catalysed by bimetallic aluminium–salen complexes. Chem Eur J 16(23):6828–6843 28. Paddock RL, Nguyen ST (2004) Chiral (salen)Co-III catalyst for the synthesis of cyclic carbonates. Chem Commun 14:1622–1623 29. Bai DS, Duan SH, Hai L, Jing HW (2012) Carbon dioxide fixation by cycloaddition with epoxides, catalyzed by biomimetic metalloporphyrins. ChemCatChem 4(11):1752–1758 30. Man ML, Lam KC, Sit WN, Ng SM, Zhou ZY, Lin ZY, Lau CP (2006) Synthesis of heterobimetallic Ru–Mn complexes and the coupling reactions of epoxides with carbon dioxide catalyzed by these complexes. Chem Eur J 12(4):1004–1015 31. Wang TT, Xie Y, Deng WQ (2014) Reaction mechanism of epoxide cycloaddition to CO2 catalyzed by salen-M (M = Co, Al, Zn). J Phys Chem A 118(39):9239–9243 32. Castro-Gomez F, Salassa G, Kleij AW, Bo C (2013) A DFT study on the mechanism of the cycloaddition reaction of CO2 to epoxides catalyzed by Zn(salphen) complexes. Chem Eur J 19(20):6289– 6298 33. Sun H, Zhang D (2007) Density functional theory study on the cycloaddition of carbon dioxide with propylene oxide catalyzed by alkylmethylimidazolium chlorine ionic liquids. J Phys Chem A 111(32):8036–8043 34. Foltran S, Mereau R, Tassaing T (2014) Theoretical study on the chemical fixation of carbon dioxide with propylene oxide catalyzed by ammonium and guanidinium salts. Catal Sci Technol 4(6):1585– 1597 35. Ajitha MJ, Suresh CH (2011) NHC catalyzed CO2 fixation with epoxides: probable mechanisms reveal ter molecular pathway. Tetrahedron Lett 52(41):5403–5406 36. Chen F, Li X, Wang B, Xu T, Chen S-L, Liu P, Hu C (2012) Mechanism of the cycloaddition of carbon dioxide and epoxides catalyzed by cobalt-substituted 12-tungstenphosphate. Chem Eur J 18(32):9870–9876 37. Whiteoak CJ, Nova A, Maseras F, Kleij AW (2012) Merging sustainability with organocatalysis in the formation of organic carbonates by using CO2 as a feedstock. Chemsuschem 5(10):2032–2038 38. Ema T, Miyazaki Y, Shimonishi J, Maeda C, Hasegawa J (2014) Bifunctional porphyrin catalysts for the synthesis of cyclic carbonates from epoxides and CO2: structural optimization and mechanistic study. J Am Chem Soc 136(43):15270–15279 39. Zhao Y, Schultz NE, Truhlar DG (2006) Design of density functionals by combining the method of constraint satisfaction with parametrization for thermochemistry, thermochemical kinetics, and noncovalent interactions. J Chem Theory Comput 2(2):364– 382 40. Zhao Y, Truhlar DG (2006) Density functional for spectroscopy: no long-range self-interaction error, good performance for Rydberg and charge-transfer states, and better performance on average than B3LYP for ground states. J Phys Chem A 110(49):13126–13130 41. Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics,
J Mol Model (2015) 21:179
42. 43.
44.
45.
46. 47. 48.
49.
50.
noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor Chem Accounts 120(1–3): 215–241 Frisch MJ, Trucks GW, Schlegel HB et al (2010) Gaussian 09, revision C.01. Gaussian, Inc., Wallingford Hay PJ, Wadt WR (1985) Abinitio effective core potentials for molecular carculations—potentials for the transition-metal atoms Sc to Hg. J Chem Phys 82(1):270–283 Wadt WR, Hay PJ (1985) Abinitio effective core potentials for molecular calculations—potentials for main group elements Na to Bi. J Chem Phys 82(1):284–298 Harihara PC, Pople JA (1973) Influence of polarization functions on molecular-orbital hydrogenation energies. Theor Chim Acta 28(3):213–222 Fukui K (1970) Formulation of the reaction coordinate. J Phys Chem 74(23):4161–4163 Fukui K (1981) The path of chemical reactions—the IRC approach. Acc Chem Res 14(12):363–368 Weigend F, Furche F, Ahlrichs R (2003) Gaussian basis sets of quadruple zeta valence quality for atoms H–Kr. J Chem Phys 119(24):12753–12762 Marenich AV, Cramer CJ, Truhlar DG (2009) Universal solvation model based on solute electron density and on a continuum model of the solvent defined by the bulk dielectric constant and atomic surface tensions. J Phys Chem B 113(18):6378–6396 Reed AE, Curtiss LA, Weinhold F (1988) Intermolecular interactions from a natural bond orbital, donor–acceptor viewpoint. Chem Rev 88(6):899–926
Page 9 of 9 179 51.
52.
53.
54.
55.
56.
57.
58.
Guo Z, Guan XG, Huang JS, Tsui WM, Lin ZY, Che CM (2013) Bis(sulfonylimide)ruthenium(VI) porphyrins: X-ray crystal structure and mechanism of C-H bond amination by density functional theory calculations. Chem Eur J 19(34):11320–11331 Lu HJ, Subbarayan V, Tao JR, Zhang XP (2010) Cobalt(II)-catalyzed intermolecular benzylic C–H amination with 2,2,2trichloroethoxycarbonyl azide (TrocN(3)). Organometallics 29(2): 389–393 Lyaskovskyy V, Suarez AIO, Lu HJ, Jiang HL, Zhang XP, de Bruin B (2011) Mechanism of cobalt(II) porphyrin-catalyzed C–H amination with organic azides: radical nature and H-atom abstraction ability of the key cobalt(III)–nitrene intermediates. J Am Chem Soc 133(31):12264–12273 Lu X-B, Liang B, Zhang Y-J, Tian Y-Z, Wang Y-M, Bai C-X, Wang H, Zhang R (2004) Asymmetric catalysis with CO2: direct synthesis of optically active propylene carbonate from racemic epoxides. J Am Chem Soc 126(12):3732–3733 Sit WN, Ng SM, Kwong KY, Lau CP (2005) Coupling reactions of CO2 with neat epoxides catalyzed by PPN salts to yield cyclic carbonates. J Org Chem 70(21):8583–8586 Rowland RS, Taylor R (1996) Intermolecular nonbonded contact distances in organic crystal structures: comparison with distances expected from van der Waals radii. J Phys Chem 100(18):7384–7391 Guo CH, Wu HS, Zhang XM, Song JY, Zhang X (2009) A comprehensive theoretical study on the coupling reaction mechanism of propylene oxide with carbon dioxide catalyzed by copper(I) cyanomethyl. J Phys Chem A 113(24):6710–6723 Anslyn EV, Dougherty DA (2006) Modern physical organic chemistry. University Science Books, Herndon
