Green Synthesis of Novel Polyesterurethane Materials from Epoxides and Carbon Dioxide by New Set of One-Dimensional Coordination Polymer Catalyst.

Two novel polyesterurethane materials, PEU1 and PEU2, were synthesized via nontoxic and isocyanate-free route by simple conversion of two epoxides 1,2-epoxy-3-phenoxy propane (2) and styrene epoxide (3) utilizing CO2. Epoxides 2 and 3 were converted to the respective cyclic carbonates 4 and 5 by a new set of cobalt-based catalyst 1a in the presence of 10 bar of CO2 and 80 °C temperature without using cocatalyst tetrabutylammonium bromide (TBAB). The mechanistic pathway of the catalysis reaction for the cycloaddition of epoxides with CO2 to generate the cyclic carbonates was investigated by several spectroscopic techniques and utilizing analogous zinc-based 1D coordination polymer 1b, which does not act as an efficient catalyst in the absence of TBAB. Cyclic carbonates 4 and 5 were converted to the respective polyesterurethanes PEU1 and PEU2 sequentially by first synthesizing the ring-opened diols 6 and 7 reacting with ethylenediamine and subsequently annealing the respective diols 6 and 7 at 120 °C in the presence of terepthalyl chloride and triethylamine. The polyesterurethanes PEU1 and PEU2 were characterized by multinuclear NMR and FTIR. PEU1 was also characterized by MALDI-TOF mass spectrometry. The thermal studies of PEU1 and PEU2 showed the stability up to 200-270 °C. The number-average and weight-average molecular weights were determined for PEU1 and PEU2 by GPC analysis. The weight-average molecular weight for PEU1 was found to be 5948 with a polydispersity of 1.1, and PEU2 showed the weight-average molecular weight as 4224 with a polydispersity of 1.06.

Introduction

Polyurethanes[1−5] are known to be very promising materials widely used in home furnishing, coatings, etc. Apart from wide spread application of polyurethanes, the modified polyurethanes known as polyhydroxyurethanes[6−10] and polyesterurethanes[11−13] are becoming increasingly demanding materials because of their application for building smart biodegradable materials including materials for tissue engineering in biomedical science.[12,14−17] Most of the cases, the synthesis of polyesterurethane involves the toxic disocyanates,[18] and an isocyanate-free synthetic route for the polyesterurethanes is rare.[11] Here, we report a green method for the synthesis of novel polyesterurethane materials, PEU1 and PEU2, from cyclic carbonates (4 and 5), ethylenediamine, and terepthalyl chloride. Although there is a report of green synthesis of the polyesterurethane material,[11] but the involvement of cyclic carbonates incorporating the phenyl substituent is not known. Synthesis of polyesterurethane materials involving the phenyl-containing cyclic carbonates could bring additional properties including the thermal stability of the resulting polyesterurethanes. The global warming and climate change have caused alarming impacts on earth’s environment, and the major cause of the global warming is happened to be the uncontrolled increase of level of CO2 in earth’s atmosphere, which has crossed 405 ppm in the present decade. Several methods have been developed to reduce the level of CO2 in the atmosphere,[19−22] among which the catalytic conversion of CO2 and epoxides to produce cyclic carbonates has been one of the major routes to convert CO2 into chemical feedstock.[23,24,33,34,25−32] In order to synthesize the cyclic carbonates, the majority of catalysts include a Lewis acidic metal center for the binding of the epoxide and a subsequent nitrogen base, which activates the CO2 for insertion into the metal–oxygen bond.[35,36] Tetrabutylammonium bromide (TBAB) has been widely used as the cocatalysts for the conversion of CO2 and epoxides to cyclic carbonates.[37,38] TBAB plays a dual role in the catalytic cycle[39] where it provides the bromide anions for the ring opening of the epoxide and produces tetrabutyl amine for converting CO2 into carbamate, which subsequently inserts into the metal–oxygen bond. Herein, we report two novel one-dimensional coordination polymers 1a and 1b as catalysts for the conversion of CO2 and respective epoxides 1,2-epoxy-2-phenoxy propane (2) and styrene epoxide (3) to corresponding cyclic carbonates 4 and 5 in the absence of TBAB cocatalyst where the cobalt-based framework 1a acts as an efficient catalyst, whereas the zinc-based framework 1b shows poor activity in the absence of TBAB (Scheme ). The catalysts 1a and 1b were synthesized from the precyclophane ligand L by the reaction with respective metal salts (Scheme ). The involvement of the cobalt-based catalyst 1a in cyclic carbonate synthesis was investigated by EPR, UV, and Raman spectroscopy where the intermediate carbonate species bound to the catalyst 1a was characterized by UV and Raman spectroscopy. Cyclic carbonates 4 and 5 as obtained by the catalysis reaction in Scheme with catalyst 1a were converted to the polyesterurethane materials (PEU1 and PEU2) sequentially by first ring opening of 4 and 5 with ethylenediamine to convert into the respective diols 6 and 7, which were subsequently annealed with terepthalyl chloride in the presence of triethylamine at 120 °C as shown in Scheme . The polyesterurethanes PEU1 and PEU2 were characterized by FTIR, NMR, and MALDI-TOF MS analyses. The thermal properties of the PEU1 and PEU2 were determined by TGA and DSC, and the surface analysis was performed with SEM. The molecular weight of the polyesterurethanes has been determined by GPC analysis. Hence, novel polyesterurethane materials PEU1 and PEU2 were synthesized from ring opening of five-membered cyclic carbonates 4 and 5 having phenyl-based substituents as compared to the previously reported procedure with unsubstituted ethylene carbonate.[11] Moreover, these cyclic carbonates were synthesized in moderate yields by heterogeneous catalysis by a novel cobalt-based 1D coordination polymer without using cocatalyst TBAB or any ionic liquids, which were previously reported to be present in the majority of cases for ambient catalytic conversion of CO2 and epoxides to cyclic carbonates.

Scheme 2

Catalytic Cycloaddition of CO2 and Epoxides

Scheme 1

Synthesis of One-Dimensional Coordination Polymers 1a and 1b

Scheme 4

Synthesis of Polyesterurethanes PEU1 and PEU2 from the respective cyclic carbonates 4 and 5.

Results and Discussion

Synthesis and Structure of 1D Coordination Polymers

Many metal–organic frameworks and coordination polymers are known to perform interesting catalytic activities.[40−43] A few 1D coordination polymers synthesized from benzimidazole- and triazole-based precyclophane ligands are reported to perform catalytic organic transformations.[44,45] By considering the catalytic potentials of the precyclophane type of ligands, we synthesized two new 1D coordination polymers based on the ligand 2,6-bis(1H-benzotriazol-1-ylmethyl)pyridine (L). The 1D coordination polymers 1a and 1b were synthesized by the reaction of the precyclophane ligand 2,6-bis(1H-benzotriazol-1-ylmethyl)pyridine (L) with the respective metal salts as per Scheme . 1a was isolated as a blue crystalline solid, whereas 1b was isolated as a white crystalline solid from the respective acetonitrile solutions. The single crystals of 1a were grown by layering an acetonitrile solution of cobalt chloride onto an acetonitrile solution of ligand L, whereas the single crystals of 1b were grown by slow ether diffusion to a methanolic solution of 1b. 1a and 1b were characterized by FTIR and single-crystal X-ray diffraction methods. The single-crystal data and refinement parameters for 1a and 1b are provided in Table . Cobalt complex 1a crystallizes in triclinic P1 space group, and its asymmetric unit consists of one ligand L, one cobalt, and two chloride atoms. Ligand L adopts a tweezers conformation with the benzotriazole units remaining as two arms of the tweezers having an angle of 14.37° between the two benzotriazole planes. Two benzotriazole units remain in opposite orientation with the two coordinating nitrogen atoms of the benzotriazole units remaining in the opposite direction forming a one-dimensional zigzag chain along the crystallographic “a” axis (Figure ). The angle created by the benzotriazole donors from two different ligands at the cobalt center is found to be 11.64(7)°. Two chloride anions and two benzotriazole units from two different ligands coordinate to a cobalt center creating a pseudo-tetrahedral geometry with the angles ranging from 111.64(7)° [N(7)-Co(1)-N(1)] to 105.35(5)° [N(7)-Co(1)-Cl(2)] (Table S1). The cobalt–nitrogen bond distances appear as 2.0289(17) and 2.0299(17) Å (Table S1), which are slightly less than the cobalt–nitrogen bond distance of the reported cobalt-benzotriazole pentacoordinated complexes.[46] The benzene rings of the benzotriazole units from the same ligand remain in parallel orientation in the structure of 1a with the distance between two centroids of benzene rings measured as 3.79 Å, which is close to be considered as a π-stacking interaction.[47] The compound 1b crystallizes in the monoclinic C2/c space group, and its asymmetric unit consists of half of the ligand L, one Zn center, and one nitrate anion. In this coordination mode, L adopts a tweezers conformation with two benzotriazole units remaining as two arms of the tweezers with the angle between the two benzotriazole planes measuring 13.37°. The orientation of the two benzotriazole unit remains opposite to each other, which form a zigzag one-dimensional coordination polymer along the crystallographic c axis (Figure S1). The angle created at the zinc center by the two associated coordinating benzotriazole unit measures 115.79(10)°. With the two nitrate anions and two benzotriazole units coordinating to a zinc center, the coordination environment around each zinc center remains pseudotetrahedral with the angles ranging from 115.79(10)° [N(2)-Zn(1)-N(2)] to 133.42(12)° [O(1)-Zn(1)-O(1)] (Table S2). The zinc–nitrogen distance ranges 2.0323(17) Å, which is slightly less than the reported zinc–nitrogen bond distance in case of reported zinc benzotriazole complexes (2.1–2.2 Å) where the zinc atom remains in pentacoordinated environment.[46] The benzotriazole benzene rings from the same ligand L form π-stacking interaction in the structure with the distance between the two benzene rings measuring 3.87 Å. The TGA of the coordination polymer 1a shows thermal stability up to 300 °C, whereas the zinc-based coordination polymer 1b shows thermal stability up to 250°C (Figures S2 and S3).

Table 1

Crystal Data and Structure Refinement for 1a and 1ba

identification code	1a	1b
empirical formula	C19H15Cl2CoN7	C19H15 N9 O6 Zn
formula weight	471.21	530.79
temperature	296(2) K	296(2) K
wavelength	0.71073 Å	0.71073 Å
crystal system	triclinic	monoclinic
space group	P1	C2/c
unit cell dimensions	a = 8.9044(4) Å	a = 12.9885(6) Å
	b = 10.4660(5) Å	b = 18.7571(6) Å
	c = 11.4088(5) Å	c = 8.7867(3) Å
	a = 69.988(2)°	a = 90°
	b = 86.245(2)°	b = 93.556(2)°
	g = 84.542(2)°	g = 90°
volume	993.88(8) Å3	2136.55(14) Å3
Z	2	4
density (calculated)	1.575 mg/m3	1.650 mg/m3
absorption coefficient	1.153 mm–1	1.209 mm–1
F(000)	478	1080
crystal size	0.4 × 0.2 × 0.2 mm3	0.4 × 0.2 × 0.2 mm3
θ range for data collection	2.932 to 29.055°	1.910 to 29.128°
index ranges	–12 ≤ h ≤12, −11 ≤ k ≤ 14, –15 ≤ l ≤ 15	–17 ≤ h ≤ 17, –25 ≤ k ≤ 25, –11 ≤ l ≤ 11
reflections collected	16857	20440
independent reflections	5254 [R (int) = 0.0256]	2860 [R (int) = 0.0243]
completeness to θ = 25.242°	99.4%	99.9%
refinement method	full-matrix least-squares on F2	full-matrix least-squares on F2
data/restraints/parameters	5254/0/322	2860/0/191
goodness-of-fit on F2	1.002	1.280
final R indices [I > 2σ (I)]	R1 = 0.0403, wR2 = 0.1238	R1 = 0.0455, wR2 = 0.1437
R indices (all data)	R1 = 0.0499, wR2 = 0.1326	R1 = 0.0489, wR2 = 0.1468
extinction coefficient	n/a	0.0091(14)
largest diff. peak and hole	0.965 and −0.358 e Å–3	1.064 and −0.573 e Å–3

R1 = [Σ||F0| – |Fc||/Σ|F0|], wR2 = [Σw(F02 – Fc2)2/Σw(F02)2]1/2; n/a, not applicable.

Figure 1

Crystal packing diagram of the coordination polymer 1a forming a 1D zigzag chain along the crystallographic a axis.

Catalytic Cycloaddition of Epoxide and CO2

The catalytic cycloaddition reaction of the epoxides 2 and 3 to the corresponding cyclic carbonates 4 and 5 as shown in Scheme was tried at 10 bar pressure of CO2 and 80o C in a 100 mL reactor with one-dimensional coordination polymers 1a and 1b as catalysts. In the previously reported studies of conversion of CO2 to cyclic carbonates, it was observed that the reaction proceeds with the addition of the cocatalyst[37,39,48,49]tetrabutylammonium bromide along with the metal-based catalysts. In recent years, ionic liquid-based heterogeneous catalytic systems have been developed, which usually perform the catalytic reaction at higher temperature and pressure[50−52] but the reaction yield is found to be very high and with higher turnover numbers. However, the conversion of 2 and 3 to the respective cyclic carbonates 4 and 5 were achieved in 66 and 59% yield, respectively, with more than 1000 turnover number (Table ) by cobalt-based catalyst 1a at 10 atmospheric pressure of CO2 and 80 °C in the absence of cocatalyst TBAB. The cyclic carbonates 4 and 5 were obtained exclusively without the appearance of any side product or polymers, which was confirmed by NMR spectroscopy. Although, in this case, the reaction condition is not relatively milder as compared to the previously reported metal-based catalysts, but it is interesting from the viewpoint of the TBAB-free catalysis reaction. The zinc-based catalyst 1b on the other hand showed lower activity in the absence of the cocatalyst TBAB under the same reaction condition with a turnover number less than 300 and yields less than 10% in both cases (Table ). In order to see the effect of TBAB on the cycloaddition of CO2 and epoxide, we performed the blank experiment with styrene epoxide under the same reaction condition without any catalyst and found the reaction yield to be 9% and a turnover number of 260 (Table ). However, the involvement of catalyst 1a for the cycloaddition of CO2 and epoxide was found indispensable by carrying out the blank experiment without the catalyst and only with CO2 and styrene oxide, which shows no cyclic carbonate peak in FTIR spectra, whereas the presence of the catalyst 1a in the same reaction under the same reaction condition showed an intense cyclic carbonate peak at 1813 cm–1 (Figures S20 and S21). In order to establish the reason behind the cobalt-based one-dimensional coordination polymer 1a acting as an efficient catalyst for the synthesis of the cyclic carbonates 4 and 5, it is presumed that the chloride ions bound to the free coordination sites of the one-dimensional framework are first replaced by the epoxide substrates following the ring opening of the epoxide by the replaced chloride anions (Scheme ). The activation of the CO2 is supposed to take place via the free pyridine base present in the precyclophane ligand (L) in the complex 1a followed by the insertion of the CO2 into the metal–oxygen bond. However, in case of zinc-based catalyst 1b, the presence of the nitrate anion and a different coordination environment may possibly slow down the epoxide ring opening and initiation of the catalysis reaction,[53] which might be the reason behind the poor efficiency of 1b for the cycloaddition of epoxide and CO2. This is supported by the fact that the coordination polymer 1b catalyzes the cycloaddition reaction of the epoxides and CO2 with a high turnover number in the presence of the cocatalyst TBAB.

Table 2

Catalytic Cycloaddition of CO2 and Epoxides Carried Out at 10 Atmospheric Pressure of CO2 and 80 °C Temperature

Scheme 3

Proposed Catalytic Cycle for the Catalytic Reaction of Epoxide 2 with Catalyst 1a

The mechanistic investigation was further supported by the UV–vis titration study of the catalyst 1a with epoxide 2, which indicates the binding of the epoxide with 1a as shown in Figure S4. The 1:1 mixture of 1a and the epoxide 2 were kept in 10 bar of CO2 pressure at 80 o C for 2 h, and the change in the UV–vis spectra of the mixture shows the formation of the carbonate-bound proposed intermediate 1a′ (Figure ).

Figure 2

UV–vis spectra of the Co-based 1D coordination polymer 1a, 1a and epoxide 2 (both 8.575 × 10–6 M in DMSO), and 1:1 mixture of 1a and 2 after 2 h reaction with CO2 at 80 °C and 10 bar of CO2.

Resonance Raman spectroscopic analysis of the 1:1 mixture of 1a and 2 before and after the reaction for 2 h at 80 °C and 10 bar of CO2 also clearly reveals the formation of the proposed intermediate 1a′ (Figure ). After the reaction with CO2, the appearance of the new peaks in Raman spectra at 1708 and 1221 cm–1 can be attributed to the carbon–oxygen double bond and carbon–oxygen single bond stretching, respectively, whereas appearance of the new peak at 781 cm–1 can be attributed to the Co–O stretching of the intermediate species 1a′. The peaks in the FTIR spectra for the carbon–oxygen double bond and carbon–oxygen single bond for cyclic carbonate appear at 1790 and 1250 cm–1,respectively, whereas the peaks corresponding to polycarbonate appear at 1747 and 1244 cm–1, respectively.[54] Hence, the new signals in resonance Raman spectra are due to the intermediate carbonate species 1a′.

Figure 3

Resonance Raman spectra with 785 nm laser of the catalyst (1a) and epoxide (2) in 1:1 ratio (10–2 M in DMSO) before (blue) and after (brown) the reaction with CO2 at 10 bar pressure and 80 °C temperature.

The EPR spectra of catalyst 1a and the epoxide 2 before and after the reaction with CO2 (Figure ) also indicate the formation of a new species with a g value for the mixture after the reaction with CO2 appearing at g⊥ = 2.27, whereas the g value for the same mixture before the reaction with CO2 appears at g⊥= 2.13. Hence, UV–vis, resonance Raman, and EPR spectra coherently suggest the formation of the intermediate carbonate species 1a′ as proposed in Scheme .

Figure 4

EPR spectra of 1a (brown) and 1a + 2 after (blue) the reaction with CO2.

Regeneration of the catalyst 1a was studied by recovering the catalyst from the reaction with styrene oxide and CO2 after 12 h. The recovered catalyst showed identical FTIR spectra (Figure S22), which indicate that the catalyst remains after the first catalytic cycle and the second catalytic cycle with the regenerated catalyst 1a showed the formation of the respective cyclic carbonates with a slightly less yield (Figure S23).

Synthesis of the Polyesterurethanes PEU1 and PEU2

The cyclic carbonates 4 and 5 as obtained from the catalytic reaction were first treated with 0.5 equiv of ethylenediamine at 100 °C to synthesize the ring-opened diols 6 and 7 (Scheme ). The use of the phenyl substituent-based five-membered cyclic carbonates 4 and 5 gives the ring-opened diols with higher selectivity as previously reported analogous compounds[9] and is confirmed from the HPLC data of the respective ring-opened diols 6 and 7 (Figures S25 and S26).The formation of the diols 6 and 7 was fully confirmed via 1H and 13C NMR as well as the ESI-MS studies (Figures S5–S10). The subsequent annealing of 6 and 7 with terepthalyl chloride in equivalent ratios at 120 °C in the presence of 2 equiv of triethylamine produces the polyesterurethanes PEU1 and PEU2, respectively. The PEU1 and PEU2 were isolated as thick viscous materials. Formation of the polyesterurethane was initially checked by the FTIR spectra, which show (Figure ) the characteristic peaks at 3390 (ν N–H), 2960 (ν C–H), 2880 (ν C–H), 1720 (ν C=O), 1263 (ν C–O ester), and 1103 (ν C–O urethane) cm–1 corresponding to the polyurethane materials.[11] In case of proton NMR spectra of PEU1, the peaks at 8.12 and 7.98 ppm correspond to the aromatic protons of the terepthalyl group, which indicate the incorporation of the terepthalyl moiety in the PEU1 as compared to the 1H NMR of the ring-opened diol 6 (Figure S11). The 13C NMR spectra of the respective polyesterurethane PEU1 also include the peaks corresponding to the terepthalyl moiety at 165 and 133 ppm. In case of PEU2, the proton NMR shows the peaks at 8.19 and 7.98 ppm corresponding to the terepthalyl moiety as compared to the 1H NMR of ring-opened diol 7. The 13C NMR of PEU2 shows the peaks at 165 and 139 ppm corresponding to the terepthalyl unit in the polymer (Figures S13 and S14). In order to study the thermal behavior of both the polyesterurethane materials, TGA and DSC studies were carried out. PEU1 shows thermal stability up to 270 °C (Figure ), whereas PEU2 shows thermal stability up to 250 °C (Figures S15 and S16) in the TGA data.

Figure 5

FTIR spectra of PEU1.

Figure 6

TGA and DSC curves for PEU1.

The differential scanning calorimetry (DSC) technique is an important method to study the thermal transformation of polymer materials, and it is necessary to study the melting point (Tm) of PEU1 and PEU2. As shown in Figure and Figure S16, with the temperature increasing, both PEU1 and PEU2 show an endothermic peak at the range of 200 to 400 °C. The melting range of the curve is long, the shape of the peak is not sharp, and it can be attributed to the microphase-separation structures of the polyesterurethane.[55] The higher value of the Tm (280 °C) in case of PEU2 can be due to the strong H-bonding interactions between the NH groups of the respective polyesterurethane molecules along with the π-bonding interaction between the aromatic rings of the polyesterurethanes. The MALDI mass of PEU1 (Figure ) shows a typical polymeric pattern with an average 240 mass difference, which is corresponding to the respective fragment-I coming out of the PEU1. Although the MALDI mass of PEU2 was not determined, but the analogous FTIR peaks (Figure S24) at 3408, 2940, 2879, 1707, 1267, and 1149 cm–1 as compared to PEU1 confirm the formation of the polyesterurethane material.

Figure 7

MALDI mass of the polyesterurethane PEU1.

The molecular weight determination by GPC analysis showed the weight-average molecular weight (Mw) in case of PEU1 and PEU2 as 5948 and 4224, respectively (Figures S17 and S18). The polydispersities for PEU1 and PEU2 were found to be 1.105 and 1.069, respectively. The molecular weight for the polyesterurethanes is in the lower range as compared to the literature,[11] which might be due to faster chain termination for the presence of the acid chloride. The surface of the PEU1 and PEU2 were analyzed via SEM analysis (Figure and Figure S19). Corresponding EDX spectra confirm the presence of polyesterurethane materials by the appearance of the constituting elements N, O, and C peaks (Figure and Figure S19).

Figure 8

SEM image of the PEU2 (left) and the corresponding EDX spectra (right)

Conclusions

Two novel polyesterurethane materials PEU1 and PEU2 have been synthesized and successfully characterized. The polyesterurethane materials were synthesized from cyclic carbonates 4 and 5 via ring opening with ethylenediamine and subsequent polymerization with terepthalyl chloride. Cyclic carbonates 4 and 5 were synthesized by a green method involving two novel 1D coordination polymers 1a and 1b, among which cobalt-based 1a catalyzes the reaction with a high turnover number in the absence of the cocatalyst TBAB, whereas zinc-based 1b could not perform the catalysis reaction with high efficiency in the absence of TBAB. The chloride anions in case of 1a is supposed to act as an efficient epoxide ring opening agent, which is the basic requirement for the initiation of the cycloaddition reaction of epoxide and CO2; however, the nitrate anion does not perform the epoxide ring opening very efficiently. The presence of a basic free pyridine moiety in 1a and 1b is supposed to activate the CO2 to respective carbamate, which subsequently insert into the metal–epoxide bond to form the carbonate intermediate. The mechanistic investigation by monitoring the reaction with UV–vis, resonance Raman spectroscopy, and EPR spectroscopy showed the involvement of the carbonate intermediate species 1a′ in case of cobalt-based catalyst 1a. In order to establish this mechanistic pathway unambiguously, the design of new catalysts incorporating the free nitrogen bases are under investigation.

Experimental Section

Materials and Methods

Styrene oxide, 1,2-epoxy-3-phenoxy propane, 2,6-bis(hydroxymethyl)pyridine, phosphorous tribromide, benzotriazole, and terepthalyl chloride were purchased from Sigma-Aldrich chemical company and were used without further purification. The solvents acetonitrile and methanol were obtained as analytical grade and were distilled following the standard procedure. Ligand L was synthesized according to the reported literature procedure.[56] Deuterated solvents for NMR experiments were purchased from Sigma-Aldrich. UV–vis experiment was performed on a UV-1800 Shimadzu model instrument. TGA data were collected on a Mettler Toledo STAR system. 1H NMR spectra were recorded on a Bruker 300 MHz NMR spectrometer. 1H NMR spectra were referenced internally to residual solvent peaks, and chemical shifts are expressed relative to tetramethylsilane, SiMe4 (δ = 0 ppm). Fourier transform infrared (FTIR) spectra were recorded with an IRAffinity-1S Shimadzu spectrophotometer. ESI-MS was recorded in a high-resolution mass spectrometer, Agilent QTOF 6520. MALDI mass was taken in a Bruker Autoflex Speed TOF instrument. ESR was recorded in a JEOL JES-FA200 instrument with solid samples. Resonance Raman spectra were recorded in a Horiba Jobin-Yvon LabRam HR. SEM analysis was performed with an FESEM Sigma 300 machine. GPC was performed on a Water instrument with UV/V visible detection-2489, refractive index detector-2414.

Crystallographic Methods

Single-crystal X-ray diffraction data were collected on a Bruker Smart Apex II CCD diffractometer equipped with a graphite monochromator and a Mo Kα fine-focus sealed tube (λ = 0.71073 Å). Data integration was done using SAINT. Intensities for absorption were corrected using SADABS. Structure solution and refinement were carried out using a Bruker SHELXTL.[57,58] The hydrogen atoms were refined isotropically, all the other atoms were refined anisotropically, and C–H hydrogens were fixed using the HFIX command in SHELXTL. Molecular graphics were prepared using X-SEED 68 and Mercury licensed version 3.9.69.

Sample Preparation for UV Experiment

A solution of 1a and 2 (both 8.575 × 10–6 M) was mixed in equal proportion, and UV–vis spectra were taken directly from one portion (2 mL) of this solution. Another portion (2 mL) was heated in a metal pressure reactor at 80 °C under 10 bar of CO2 for 2 h. After 2 h, the UV–vis spectra were recorded from the solution.

Sample Preparation for Resonance Raman Experiment

A solution of 1a and 2 (both 10–2 M) in DMSO was mixed in equimolar proportion, and one portion (2 mL) of this solution was directly submitted for resonance Raman experiment. Another portion of the solution (2 mL) was heated in a metal pressure reactor at 80 °C under 10 bar of CO2 for 2 h. After 2 h, the solution was submitted for resonance Raman experiment.

Sample Preparation for ESR Experiment

A 1:1 mixture of catalyst 1a (100 mg, 0.21 mmol) and epoxide 2 (31.5 mg, 0.21 mmol) was added to a metal pressure reactor and was heated to 80 °C for 2 h under 10 bar of CO2 pressure. After 2 h, the material was submitted for ESR analysis.

Synthesis of 1a

Ligand L (17.06 mg, 0.05 mmol) dissolved in 2 mL of acetonitrile was layered to cobalt(II) chloride hexahydrate (11.89 mg, 0.05 mmol) in 2 mL of acetonitrile. The deep blue crystal was obtained by the layering method. Yield: 10 mg (42.4%). FTIR: 3068, 2927, 1597, 1579, 1496, 1458, 1427, 1382, 1328, 1282, 1226, 1195, 1170, 1149, 995, 968, 788, 748 cm–1.

Synthesis of 1b

Ligand L (85.3 mg; 0.25 mmol) in acetonitrile (5 mL) was added to zinc nitrate hexahydrate (74.36 mg, 0.25 mmol) in acetonitrile (5 mL) and stirred at room temperature for 3 h. This yielded a white precipitate. The precipitate was isolated by centrifugation and recrystallized using methanol and diethyl ether to give 1b complex as colorless crystals. Yield: 75 mg (56.5%). FTIR: 3105, 2945, 1598, 1581, 1496, 1460, 1427, 1382, 1305, 1286, 1226, 1172, 1153, 1008, 970, 854, 806, 788, 750 cm–1.

Synthesis of Cyclic Carbonates

In a typical synthetic procedure, a 100 mL metal reactor was charged with 4 mL of epoxide (2/3), 5 mg of catalyst 1a, and magnetic stirrer bar. Then, the reactor was heated at 80 °C under 10 atmospheric pressure of CO2 for 20 h. After 20 h, the products were isolated and submitted for NMR. Yield and turnover numbers were calculated based on the NMR analysis. Cyclic carbonate obtained 4: 2.90 g (66% yield). 1H NMR(300 MHz, CDCl3): δ 7.34–7.27 (m, 2H, PhH), 7.04–7.02 (m, 1H, PhH), 6.93–6.90 (m, 2H, PhH), 5.07–5.02 (m, 1H, OCH), 4.65–4.52 (m, 2H, CH2), 4.27–4.22 (dd, J = 6 Hz, 1H, OCH2) ppm. 13C NMR (300 MHz, CDCl3): δ 157.65 (Ph–OC), 154.68 (C=O), 129.7 (Ph–C), 121.93 (Ph–C), 114.50 (Ph–C), 74.06 (CH2), 66.73 (OCH), 64.20 (OCH2) ppm. FTIR (KBr) ν: 2926, 2875, 1793, 1595, 1492 cm–1. HRMS (ESI+): [M + H]+ calcd for C10H10O4, 195.18; found, 195.07.

5: 2.47 g (59% yield). 1H NMR (300 MHz, CDCl3): δ 7.44 (m, 5H, PhH), 5.69 (m, 1H, OCH), 4.81 (m, 1H, OCH2), 4.33 (m, 1H, OCH2) ppm. 13C NMR (300 MHz, CDCl3): δ 154.83 (C=O), 135.59 (Ph), 129.52 (Ph), 129.01 (Ph), 125.9 (Ph), 77.88 (OCH), 71.03 (OCH2) ppm. FTIR (KBr): 2983, 2926, 1793, 1645, 1550, 1496, 1458, 1396 cm–1. HRMS (ESI+): [M + H]+ calcd for C9H8O3, 165.16; found, 165.05.

Synthesis of Ring-Opened Diol 6

Cyclic carbonate 4 (4 g, 0.02 mol, 2 equiv) was taken in an RB flux and heated to melt, followed by the dropwise addition of ethylene diamine (0.69 mL, 0.01 mol, 1 equiv). The resulting mixture was stirred and heated at 100 °C for 2 h to obtain a colorless viscous mass. Yield: 4.6 g (99%). 1H NMR(300 MHz, CDCl3): δ 7.34–7.24 (m, 4H, PhH), 6.98–6.87 (m, 6H, PhH), 5.71 (m, 2H, CH of carbonate), 5.0 (broad singlet, NH), 4.31–4.20 (m, 4H, CH of carbonate), 4.12–3.86 (m, 4H, CH of carbonate), 3.29 (m, 4H, CH2–CH2), 2.4 (broad singlet, OH) ppm. 13C NMR(300 MHz, CDCl3): δ 158, 157, 129, 121, 114, 68, 66, 61, 40 ppm. FTIR: 3444, 2931, 1674, 1597, 1541, 1384, 1269, 1068, 881, 823, 761, 686 cm–1. HRMS (ESI+): [M + H]+ calcd for C22H29N2O8, 449.192; found, 449.18.

Synthesis of Ring-Opened Diol 7

Cyclic carbonate 5 (2 g, 12.2 mmol, 2 equiv) was taken in an RB flux and heated to melt, followed by the dropwise addition of ethylenediamine (0.41 mL, 6.1 mmol, 1 equiv) resulting in a light brown color viscous mass. Yield: 2.24 g (93%). 1H NMR (300 MHz, CDCl3): δ 7.31–7.26 (m, 10H, PhH), 5.97–5.77 (m, 2H, CH of carbonate), 4.9 (broad singlet, NH), 4.26–3.74 (m, 4H, CH of carbonate), 3.25 (m, 4H, CH2–CH2), 2.75 (broad singlet, OH) ppm. 13C NMR (300 MHz, CDCl3): δ 157, 139, 128, 126, 69, 65, 40 ppm. FTIR: 3442, 2962, 1707, 1544, 1452, 1409, 1382, 1261, 1095, 1022, 866, 800, 700 cm–1. HRMS (ESI+): [M + H]+ calcd for C20H25N2O6, 389.171; found, 389.172.

Synthesis of PEU1

The diol 6 (3.46 g, 7.7 mmol), 1.56 g of terepthalyl chloride (7.7 mmol), and 3.89 mL of triethylamine (38.7 mmol, 5 equiv) were added, and the mixture was put in a glass tube and sealed under nitrogen. Then, this mixture was first heated to 120 °C for 4 h and cooled over for 20 h in a preprogrammed oven. The PEU1 was obtained as a pale yellow thick viscous material. 1H NMR (300 MHz, CDCl3): δ 8.13–7.84 (m, 4H, terepthalyl), 7.32–7.26 (m, 4H, PhH), 6.99–6.92 (m, 6H, PhH), 5.6 (m, 2H, CH of carbonate), 5 (broad singlet, NH), 4.64–3.94 (m, 8H, CH of carbonate), 3.25–3.0 (m, 4H, CH2–CH2) ppm. 13C NMR (300 MHz, CDCl3): δ 165, 158, 156, 133, 129, 121, 114, 73, 66, 53, 44.9 ppm. FTIR: 3390, 2960, 2880, 1720, 1598, 1535, 1426, 1458, 1384, 1103, 1047, 877, 806, 756, 692 cm–1.

Synthesis of PEU2

The diol 7 (2.24 g, 5.8 mmol), 1.56 g of terepthalyl chloride (1.18 g, 5.8 mmol), and 3.89 mL of triethylamine (2.47 mL) were added, and the mixture was put in a glass tube and sealed under nitrogen. Then, this mixture was first heated to 120 °C for 4 h and cooled over for 20 h in a preprogrammed oven resulting in the formation of a pale yellow thick viscous material. 1H NMR (300 MHz, CDCl3): δ 8.19–7.9 (m, 4H, terepthalyl), 7.5–7.26 (m, 10 H, PhH), 6.29–5.72 (m, 2H, CH of carbonate), 4.91–4.09 (m, 4H, CH of carbonate), 3.25–3.04 (m, 4H, CH2–CH2) ppm. 13C NMR (300 MHz, CDCl3): δ 165, 155, 139, 129, 128, 126, 44.9, 40 ppm. FTIR: 3425, 2954, 2879, 1707, 1544, 1458, 1384, 1271, 1151, 1045, 763, 702 cm–1.