Chemical Recycling of Poly(bisphenol A carbonate) by Glycolysis under 1,8-Diazabicyclo[5.4.0]undec-7-ene Catalysis

Abstract

The glycolysis reaction of poly(bisphenol A carbonate) (PC) has been explored under 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) catalysis as a potential route to valorize PC wastes by chemical recycling. The amidine base is an active catalyst of PC glycolysis and, under suitable conditions, promotes effectively and selectively the depolymerization of the polymeric material with 1,2-propanediol or glycerol to give the monomer bisphenol A (BPA) and the relevant cyclic carbonate. The depolymerization process has been investigated under solventless conditions, using diol/triol as the reagent and reaction medium, and also in an auxiliary solvent such as tetrahydrofuran (THF) that is able to dissolve the polymer. The influence of a few experimental parameters (temperature, catalyst load, and reaction time) on the selectivity to cyclic carbonate has been studied. High selectivity to cyclic carbonate has been attained by carrying out the depolymerization reaction in THF and using mild temperature conditions and a stoichiometric amount of polyol. The catalyst can be recovered from the reaction mixture as a BPA/DBU adduct and effectively recycled in a successive run.

1. Introduction

Chemical recycling of waste plastics is a useful strategy to reduce the environmental and social impact of this typology of wastes. This approach, which regards the waste as a potential resource rather than a mere refuse, implies the chemical conversion of waste polymers into valuable chemicals and provides not only a fascinating alternative to more conventional technological solutions (landfill, mechanical recycling, and energy recovery) but also a smart response to the current worldwide need of recycling carbon and saving energy.^1,2

Poly(bisphenol A carbonate) (PC; 1) is one of the most widely used thermoplastics whose market is in rapid expansion. The rise in the utilization of PC calls for the development of after-use treatments. Chemical recycling may be a suitable way of valorizing waste PC.^3,4 In recent years, besides pyrolytic approaches,^3,4 great attention has been paid to those protocols implying the cleavage of carbonate bond by hydrolysis,³⁻⁶ aminolysis,⁷ and alcoholysis.^3,4,8−10 These methods not only provide a potential route to the regeneration of the starting monomer bisphenol A (BPA; 2), which can be reused to produce new virgin PC, but can also allow the synthesis of chemicals with the added value.^3,4 In this paper, we have focused on the glycolysis reaction (eq 1; Scheme 1). This reaction is a potential route to recover BPA, and it also provides a synthetic entry into cyclic carbonates, a class of compounds widely used as solvents or intermediates in chemical synthesis. In principle, also the cyclic carbonate can be reused, together with recovered BPA, to regenerate virgin PC. In fact, cyclic carbonates, for a long time prepared from toxic phosgene or CO as sources of the carbonyl group, nowadays can be more safely obtained by the cycloaddition of CO₂ to epoxides:¹¹ this reaction is currently exploited at an industrial scale for the synthesis of dimethyl carbonate, used as a co-monomer (together with BPA) in the manufacture of PC.¹²

Scheme 1. PC Glycolysis (Eq 1).

Han et al. studied the glycolysis of PC in ethylene glycol (EG) in the absence of any catalyst.¹³ At 453 K, depolymerization proceeded slowly with low BPA yield (30%, after 6 h). At 493 K, the conversion of the polymer into the monomer was practically quantitative in 85 min, but, under the working conditions, the coproduced cyclic carbonate decomposed fast with the evolution of CO₂.

Inorganic bases, such as NaOH¹⁴ or also Na₂CO₃,¹⁵ turned out to be active catalysts of PC glycolysis and, at 453 K, under solvolytic conditions, promoted the depolymerization of PC in EG^14,15 or 1,2-propanediol (1,2-PD; 3a)¹⁵ to give a mixture of 2 and BPA mono- and bis-(hydroxyalkyl)-ethers. In these processes, the cyclic carbonate 4 formed as transient species as the used catalysts, under the working conditions, promoted fast not only the decomposition of 4 by decarboxylation¹⁴ but also the reaction of 4 with 2 to give BPA mono- (5) and bis-hydroxyalkylation (6) products (Scheme 2).^14,15

Scheme 2. Hydroxyalkylation of BPA (2) with Propylene Carbonate (4a): Formation of Mono-Hydroxypropyl-BPA (MHP-BPA) and Bis-Hydroxypropyl-BPA (BHP-BPA) Derivatives.

The utilization of nonconventional heating techniques such as microwave irradiation did not improve the selectivity to cyclic carbonate.^16,17 The success of reaction 1 depends (a) on the choice of a suitable catalyst that, although being effective in catalyzing carbonate formation (eq 1), should exhibit a poor tendency to promote both carbonate decomposition and BPA hydroxyalkylation (Scheme 2) and also requires (b) the careful tuning of right working conditions. To date, the isolation of cyclic carbonate from reaction 1 has been described only in rare cases. Oku, using NaOH or KOH as catalysts, first reported the synthesis of BPA and glycerol carbonate (4b) from the reaction of PC and a stoichiometric amount of glycerol (3b), at 373 K, in dioxane as the auxiliary solvent.¹⁸ Alkali bases are relatively cheap catalysts, but they cannot be recycled, are wasted at the end of the process, may cause equipment corrosion, and generate environmental problems. Recently, a more complex recyclable catalytic system based on an ionic liquid (Bu₄NCl) and ZnO nanoparticles was found to be effective in promoting, at 373 K in tetrahydrofuran (THF), the depolymerization of PC with diols (1,2-PD; 1,3-butanediol) or glycerol to give the relevant cyclic carbonates besides 2. However, the process required relatively long conversion times (7–14 h) as well as a strong excess (vs PC) of the used polyol.¹⁹ To date, the search for effective reusable catalytic systems able to promote reaction 1 selectively under conditions more appealing from a practical point of view is still a challenging task. In this work, we have focused on an organocatalyst such as 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU). Elsewhere, we have shown that DBU can promote the carbonylation of nucleophiles NuH with carbonic acid diesters (Scheme 3, eq 2):^10,20−22 DBU can work both as a base, activating NuH through the formation of the more nucleophilic anion Nu^– (Scheme 3; base catalysis),²² and as a nucleophile capable of activating the carbonate group through the formation of a N-carbonyl-substituted ketene aminal (Scheme 3; nucleophilic catalysis),²⁰⁻²² which can effectively transfer the CO₂R group to nucleophiles NuH such as N-heteroaromatics (pyrrole) and even alcohols.²⁰ This reactivity has been recently exploited in a very recent study, wherein we have shown that DBU is an effective selective catalyst of PC alcoholysis.¹⁰ Herein, we report on the activity of DBU as the catalyst of reaction 1 and describe a few novel simple efficient protocols for the glycolysis and glycerolysis of PC using the amidine superbase as the recyclable catalyst.

Scheme 3. DBU-Promoted Carbonylation of Nucleophiles NuH with Carbonic Acid Diesters (RO)₂CO.

2. Experimental Section

2.1. General Methods and Materials

In this study, pure PC pellets (3 mm length × 2 mm diameter) were used as a model of waste PC. The used polycarbonate may contain or not a chain terminator. Polycarbonate used in the experiments with 1,2-PD was from Aldrich (M_w ≈ 64 000 by gel permeation chromatography). Scheme S1 shows a sketch of the polymeric chain and highlights the nature of the additive, 4-cumylphenol (4-CP), used as a chain terminator.²³ The spectroscopic characterization [NMR and Fourier-transform infrared (FTIR)] of this polymeric material has been reported elsewhere.⁶ The molar ratio of 4-CP/BPA in the used polymer was equal to 0.035, as determined by NMR.⁶ Accordingly, the moles of BPA units (n_BPA°) and the moles of CO₃ groups (n_CO3) incorporated in w(g) of PC feed have been calculated by means of eqs S1 and S2, respectively.

Polycarbonate used in the glycerolysis experiments was from SABIC (M_w ≈ 40 000 by GPC). Figures S1–S3 show, respectively, the ¹H and ¹³C NMR and attenuated total reflection (ATR)–FTIR spectra of the polymer. In this case, the moles of BPA (n_BPA°) and CO₃ groups (n_CO3) present in w(g) of feed PC have been calculated according to eq S3.

Yields of 2 and 4 were calculated through eqs 3 and 4, respectively,

3

4

where n_BPA and n_carbonate were, respectively, the moles of 2 and 4 determined by GC or isolated.

1,2-PD (3a) and glycerol (3b) were Fluka and Aldrich products, respectively. THF and diethyl ether were previously dried according to conventional methods (P₂O₅; Na/benzophenone)²⁴ and stored under N₂. DBU (Aldrich) was used as received and manipulated under an inert gas atmosphere to prevent any contamination from atmospheric CO₂ or moisture.

GC analyses were performed with a HP 5890 Series II or a THERMO Scientific TRACE 1310 gas-chromatograph (capillary column: Heliflex AT-5, 30 m × 0.25 mm, 0.25 μm film thickness). GC–(mass spectrometry) MS analyses were carried out with Shimadzu GC-17A linked to a Shimadzu GC-MS QP5050 selective mass detector (capillary column: Supelco MDN-5S, 30 m × 0.25 mm, 0.25 μm film thickness). IR spectra were recorded on a Shimadzu FTIR Prestige 21 spectrophotometer or a PerkinElmer Frontier MIR/FIR spectrophotometer equipped with a Pike GladiATR (diamond crystal) accessory. NMR spectra were recorded with a Varian Inova 400 spectrometer or with an Agilent 500 instrument. Chemical shifts are in δ (ppm) versus tetramethylsilane.

2.2. PC Glycolysis with 1,2-PD under Solvolytic Conditions: General Procedure

The glycolysis reaction was carried out in a ∼40 mL Schlenk tube equipped with a Sovirel screw cap and Torion stopcock. The polycarbonate was suspended in the diol under an inert gas (N₂) stream, and the catalyst was added. The reactor, once charged with the reactants, was sealed and dipped into an electrically heated silicon oil bath, and the suspension was stirred at the working temperature until complete depolymerization of PC (disappearance of the PC pellets). The reaction mixture was cooled to room temperature, diluted with THF, and analyzed by FTIR, GC–MS, and GC using n-dodecane as the internal standard.

In a few experiments, heating was prolonged until the full consumption of propylene carbonate (4a) initially formed (see, later, Table 4). BPA, as well as MHP-BPA and BHP-BPA species formed by the reaction of 2 with 4a (Scheme 2), was separated by chromatography as reported hereafter. The reaction mixture, cooled to room temperature, was treated with distilled H₂O and then extracted with diethyl ether. The organic phase was dried over MgSO₄ and fractionated on a silica gel column using, as eluent, petroleum ether/diethyl ether 1:1 (v/v) until elution of 4-CP and BPA and then petroleum ether/diethyl ether 1:2 (v/v) until elution of the mixture of MHP-BPA derivatives (5a′ and 5a″). Afterward, the composition of the mobile phase was progressively enriched in diethyl ether [petroleum ether/diethyl ether: 1:3 (v/v) → 1:5 (v/v) → 0:1 (v/v)] until elution of the mixture of BHP-BPA isomers (6a′–6a‴). The NMR characterization of the mixture of MHP-BPA derivatives (5a) as well as the mixture of BHP-BPA isomers (6a) has been reported in the Supporting Information (Figures S4–S7).

Table 4. BPA-Hydroxyalkylation Times by the Glycolysis of Polycarbonate in 1,2-PD (1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6) in the Presence of DBU.

entry	3a (mL)	PC (g)	DBU (μL)	DBU loada (mol %)	DBU concentration gDBU/mL1,2-PD	T (K)	tdepolymb (min)	tc(h)
1	2	0.209	12	9.9	0.006	436	10	4.5
2	2	0.199	12	10.4	0.006	453	5	1.5
3	10	1.009	30	5.1	0.003	453	9	5
4	2	0.204	1.2	1.0	0.0006	453	21	10

^a(mol_DBU/mol_CO3) × 100.

^bAt that time, the depolymerization of PC was quantitative (disappearance of the PC pellets).

^cOverall reaction time (i.e., until the complete conversion of 4a).

2.3. PC Glycolysis with 1,2-PD in THF: General Procedure

The glycolysis reaction was carried out in a suitable glass reactor analogous to that described above. The reactor, once charged with the reactants and the catalyst, was sealed, and the reaction mixture, containing also n-dodecane (internal standard) if used, was stirred at the working temperature. The progress of the reaction was monitored by measuring the formation of 4a over time by GC. The complete conversion of the polymer was further confirmed spectroscopically by observing the disappearance of the FTIR absorption of the polycarbonate at 1778 cm^–1. The products can be separated by column chromatography. Herein, we report the details of one of these experiments. To a THF (5 mL) solution of PC (0.512 g), 1,2-PD (150 μL) and DBU (30 μL; 10 mol % vs n_CO3) were added. The reaction mixture was reacted at 373 K for 2.5 h and, then, analyzed by GC (4a yield: ≥99%). Both 4a and 2 were separated with high yield (91 and 94%, respectively) on a silica gel column using, as eluents, CHCl₃ and, after elution of 4a, diethyl ether.

2.3.1. Catalyst Recovery and Recycling

The catalyst can be recovered through a procedure analogous to that described in ref (10). At the end of the catalytic run, the solvent (THF) was evaporated in vacuum. The residue was washed with diethyl ether. The ethereal phase, after addition of n-dodecane, was analyzed by GC. The material insoluble in diethyl ether was dried in vacuum and characterized by ¹H NMR (Figure S8) as a BPA/DBU adduct.¹⁰ In accordance with the outcomes of our previous studies on DBU-promoted PC alcoholysis,¹⁰ the BPA/DBU ratio was found to be equal to 2.5:1 mol/mol as established by means of the ¹H integral spectrum. The recovered adduct, once dissolved in THF and after addition of fresh PC and the diol, was reusable in a new run.

2.4. PC Glycerolysis under Solvolytic Conditions: General Procedure

The experimental apparatus was similar to that described for solventless glycolysis experiments (see 2.2). To the suspension of PC in the triol, the catalyst was added under a N₂ stream. The glass reactor was sealed, and the reaction mixture was stirred at the working temperature until complete PC depolymerization (disappearance of the PC pellets). The reaction mixture was cooled to ambient temperature, analyzed qualitatively by FTIR and GC–MS, treated with H₂O, and extracted several times with diethyl ether. The resulting ethereal solution was dried over MgSO₄ and, after addition of n-dodecane (internal standard), analyzed by GC for the quantitation of 2. BPA was also isolated by chromatography on a silica gel column using, as eluent, a petroleum ether/ethyl acetate gradient [from 6:1 to 2:1(v/v)].

2.5. PC Glycerolysis in THF: General Procedure

Into a suitable glass reactor (see 2.3), containing the THF solution of the polymer, glycerol, the catalyst, and n-hexadecane (internal standard), if used, were introduced under a N₂ stream. The glass reactor was sealed, and the reaction mixture was stirred at the working temperature. The progress of the reaction was monitored by measuring the formation of 4b over time by GC. The quantitative conversion of the polymer was further confirmed spectroscopically by observing the disappearance of the FTIR absorption of the polycarbonate at 1778 cm^–1. The products can be separated by column chromatography. Herein, we report the details of one of these experiments. To the THF (5 mL) solution of PC (0.517 g), glycerol (150 μL) and DBU (15 μL; 5 mol % vs n_CO3) were added. The reaction mixture was reacted at 333 K for 3 h and cooled to room temperature. After addition of CH₃COOH (6 μL), the solution was evaporated in vacuum, and the residue was fractionated on a silica gel column, first with petroleum ether/ethyl acetate 9:1 (v/v), afterward with a mobile phase progressively enriched in ethyl acetate until elution of BPA (94% yield), and, finally, with only ethyl acetate until elution of 4b (90% yield).

3. Results and Discussion

3.1. DBU-Promoted Glycolysis of PC with 1,2-PD

In this work, 1,2-PD (3a) was selected as the reference diol. The glycolysis reaction was preliminarily investigated in the absence of any catalyst, under solvolytic conditions, using diol as the reactant and reaction medium (Table 1). At 453 K (1,2-PD/PC ≈ 12 m/m), the depolymerization of 1 proceeded heterogeneously because the polymer was poorly soluble in 3a under the working conditions. The reaction was stopped after ∼2 h (entry 1, Table 1), when depolymerization was quantitative (i.e., disappearance of the polymer pellets) and a homogeneous system was obtained. Both 2 and 4a were obtained in high yield (95 and 94%, respectively). However, prolonging the reaction time caused a sensible diminution of both 2 and 4a yield (81 and 78%, respectively; entry 2, Table 1) because of side formation of MHP-BPA and BHP-BPA derivatives, as well as other species, identified by GC–MS, such as 1-[4-(2-phenylpropan-2-yl)phenoxy]propan-2-ol (the 2-hydroxypropylether of 4-CP; 270 m/z) and PhOH, 4-isopropenylphenol (4-IPP) and 4-isopropylphenol (4-IPPH), which are products of thermal decomposition of BPA.²⁵

Table 1. PC (∼0.200 g) Depolymerization in 1,2-PD (2 mL; 1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6), at 453 K, in the Absence of Any Catalyst.

entry	ta(h)	4ab (%)	BPAb (%)
1	1.83 (1.83)	94	95
2	28 (2)	78	81

^aReaction time. The value in parentheses is the time after which the depolymerization of PC was quantitative (disappearance of the PC pellets).

^bGC yield.

DBU is an active catalyst of PC glycolysis in 1,2-PD. In the presence of the amidine base, the depolymerization of 1 was markedly faster and, using a catalyst load as high as 10 mol % (mol_DBU/mol_CO3), was complete in 5 min (Table 2 and Figure 1). The reacting system remained heterogeneous throughout the reaction time, that is, until complete disappearance of the PC pellets. The FTIR analysis of the reaction mixture (diluted in THF) showed the disappearance of the band of PC at 1778 cm^–1 and the appearance of the new absorption at 1809 cm^–1 assigned to 4a (Figure 2): the FTIR spectrum does not show any other carbonyl absorption at lower wavenumbers excluding, thus, the presence in the reaction mixture of significant amounts of soluble oligomers or hydroxyalkyl-aryl-carbonate species. This suggests that the latter species, once formed, react fast to give the final products.

Table 2. Propylene Carbonate Yields and PC (∼0.200 g) Depolymerization Times in 1,2-PD (2 mL; 1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6), at 453 K, in the Presence of DBU.

entry	DBU (mL)	DBU molar loada (%)	DBU concentration (gDBU/mL1,2-PD)	tb(min)	4a GC yield (%)
1	0.012	10.4	0.006	5	84
2	0.006	5.1	0.003	8	85
3c	0.0012	1.0	0.0006	23	87
4				110	94

^a(mol_DBU/mol_CO3) × 100.

^bReaction time: at that time, the depolymerization of PC was quantitative (100%; disappearance of the PC pellets).

^cBPA was isolated with 82% yield: at the end of the catalytic run, the reaction mixture was partitioned between diethyl ether and distilled H₂O. The ethereal phase was dried over MgSO₄ and fractionated on a silica gel column with petroleum ether/diethyl ether 1:1 (v/v) until elution of 4-CP (87%) and, afterward, petroleum ether/diethyl ether 1:2 (v/v) until elution of BPA.

Figure 1.

PC depolymerization time vs DBU molar load [(mol_DBU/mol_CO3) × 100]. Experimental conditions: PC, ∼0.200 g; 1,2-PD, 2 mL (1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6); T, 453 K.

Figure 2.

FTIR spectra. (a) Reaction mixture diluted in THF (2 mL) after disappearance of the PC pellets (100% depolymerization); the bands at 1612 and 1593 cm^–1 are due to the BPA product. (b) Propylene carbonate in 1,2-PD/THF (1:1 v/v). (c) PC in THF.

The use of lower catalyst loads caused the lengthening of the depolymerization time. However, even by reducing the DBU load to 1 mol %, the increase of depolymerization time was contained within more than acceptable limits (23 min, at 453 K). The selectivity to carbonate 4a, measured when the depolymerization of 1 was complete, was lower than in the absence of the catalyst (in Table 2, entries 1–3 compared with entry 4), despite the fact that in the presence of DBU the reaction times were markedly shorter. This suggests that DBU may promote, in some degree, the further conversion of 4a. However, the high selectivity observed (≥84%; entries 1–3, Table 2) indicates that 4a is quite stable under the working conditions and reacts with BPA only to a moderate extent to give MHP-BPA and BHP-BPA derivatives (Scheme 2).

As expected, the use of less severe reaction temperatures caused a significant slowdown of the depolymerization process (Table 3, Figure 3). The selectivity to 4a, measured when PC depolymerization was quantitative, tends to decrease, albeit slowly, on lowering temperature (Table 3). This effect was more evident at the lowest temperatures investigated at which the time required for the full depolymerization of 1 was markedly longer.

Table 3. Propylene Carbonate Yields and PC (∼0.200 g) Depolymerization Times in 1,2-PD (2 mL; 1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6) at Different Temperatures, in the Presence of ∼10 mol % DBU^a.

entry	T (K)	tb(min)	4a GC yield (%)
1	453	5	84
2	436	10	83
3	407	60	74
4	393	270	75

^a(mol_DBU/mol_CO3) × 100.

^bReaction time: at that time, the depolymerization of PC was quantitative (disappearance of the PC pellets).

Figure 3.

PC depolymerization time vs temperature. Experimental conditions: PC, ∼0.200 g; 1,2-PD, 2 mL (1,2-PD/PC ≈ 12 m/m; mol_1,2-PD/mol_CO3 ≈ 3.6); DBU molar load [(mol_DBU/mol_CO3) × 100], ∼10 mol %.

Prolonging heating after the complete depolymerization of 1 caused the quantitative conversion of 4a within times that increased with decreasing either temperature or the catalyst load (Table 4). The disappearance of 4a was accompanied by the significant formation of MHP-BPA and BHP-BPA derivatives (Scheme 2), and minor amounts of other species, a few of which have been identified by GC–MS [PhOH (94 m/z); 1,1′-oxydipropan-2-ol and 2-(2-hydroxypropoxy)-1-propanol (134 m/z); 4-IPPH (136 m/z); 4-IPP (134 m/z); 1-phenoxypropan-2-ol (152 m/z); 1-[4-(2-phenylpropan-2-yl)phenoxy]propan-2-ol (270 m/z); and 1-(4-(prop-1-en-2-yl)phenoxy)propan-2-ol (192 m/z)]. As an example, at 453 K, in the presence of 5 mol % of DBU, the quantitative depolymerization of 1 (1.009 g) in 1,2-PD (10 mL) required a time as long as 9 min (entry 3, Table 4). Prolonging heating for additional 5 h caused the complete conversion of 4a and afforded MHP-BPA (5a′ and 5a″) and BHP-BPA (6a′–6a‴) products, which were isolated (see 2.2 and Figures S4–S7), together with unconverted BPA, by column chromatography (isolated yields: BPA, 38%; MHP-BPA, 40%; BHP-BPA, 13%).

As a whole, the above results show that DBU allows to control the selectivity to cyclic carbonate much better than alkali catalysts, such as Na₂CO₃ or NaOH, do.^14,15 In principle, this aspect of the catalytic activity of DBU can be exploited to improve the selectivity to 4a by suitably modifying the depolymerization conditions. Accordingly, much better selectivities to 4a have been achieved by carrying out the depolymerization process in an auxiliary solvent, such as THF (that dissolves 1 easily), applying milder temperature conditions and using the diol in stoichiometric amount versus PC (mol_1,2-PD/mol_CO3 ≈ 1).

Figure 4 summarizes the results obtained at different temperatures (≤373 K) when the polymer (∼0.200 g) was reacted with 3a in THF (2 mL) in the presence of 10 mol % DBU. The progress of the process was monitored by following the formation of 4a over time. The depolymerization reaction 1 proceeded smoothly even at ambient temperature (300 K). After 22 h, 4a formed in 96% yield. Markedly, shorter reaction times were observed at moderately higher temperatures. At 373 K, the depolymerization of 1 was quantitative in little more than 2.5 h with a carbonate yield as high as 99%. Both 4a and 2 were isolated with high yield (>90%) by column chromatography (see 2.3).

Figure 4.

PC (∼0.200 g) depolymerization with 1,2-PD (mol_1,2-PD/mol_CO3 ≈ 1) in THF (2 mL) in the presence of DBU (∼10 mol %) at different temperatures.

The catalyst can be recovered from the reaction mixture as a BPA/DBU adduct,¹⁰ as described in 2.3.1. Figure S8 shows the ¹H NMR spectrum of the adduct. The BPA/DBU molar ratio was found to be equal to 2.5:1.¹⁰ The adduct, once isolated, can be readily reused in a successive run. Even after 5 cycles, both the productivity and selectivity of the process were maintained very high (Figure 5) in accordance with the fact that the catalyst recovered from the last run exhibited spectroscopic features (Figure S9) analogous to those of the catalyst recovered after a single cycle.

Figure 5.

Catalyst recycling: PC, ∼0.500 g; 1,2-PD, 145 μL; DBU (used in cycle 1), 30 μL (∼10 mol %); THF, 5 mL; 373 K; 2.5 h. After each run, the catalyst was recovered as the BPA/DBU adduct (2.5:1 mol/mol) and reused in the successive cycle. GC yields were determined after the separation of the adduct. This explains why the BPA yield after the first cycle was sensibly lower (∼75%) than in the successive cycles.

3.2. DBU-Promoted Glycerolysis of PC

The study has been extended to an industrially relevant triol such as glycerol (3b). No reaction was observed when a suspension of PC (0.195 g) in 3b (2.491 g) was stirred for 5 h at 453 K in the absence of any catalyst: the polymer was quantitatively recovered by filtration. Therefore, in the absence of any catalyst, the behavior of the polymer in glycerol differs slightly from that displayed in EG¹³ or 1,2-PD (see 3.1): according to the mechanism proposed by Han,¹³ the very poor reactivity shown by 1 in neat glycerol may reflect, most probably, the greater difficulty of the triol molecules in penetrating into the solid particles of the polymer.

However, the addition of a catalytic amount of DBU (10 mol %) caused a drastic change of reactivity and, under solventless conditions (3b, 2.500 g), promoted the fast depolymerization of 1 (0.200 g). Because of the poor solubility of 1 in glycerol, also, in this case, the reaction mixture remained heterogeneous throughout the reaction time, until the disappearance of the pellets. The depolymerization was complete in only 26 min. Nevertheless, under the above conditions, 4b was obtained in a negligible amount. The GC–MS analysis showed the formation of BPA as the major product together with other species, among which we have identified minor amounts of PhOH, 4-IPP, and BPA hydroxyalkylation products. Use of a less severe temperature (423 K; PC, 0.202 g; glycerol, 2.563 g; DBU, 10 mol %) caused the prolongation of the depolymerization time (disappearance of the PC pellets after 168 min) but did not improve the selectivity to 4b. The analysis (FTIR, GC–MS) of the reaction solution, once the depolymerization was complete, showed, in addition to BPA, the formation of minor amounts of glycerol carbonate (no other IR absorptions were noted in the carbonyl region; see also 3.1), which disappeared on prolonging the heating at the working temperature (423 K) for further 30 min. The monomer 2 can be extracted from the reaction mixture with diethyl ether (GC yield: 90%) and was isolated (yield: 84%) by chromatography on a silica gel column (see Experimental Section). According to the above results, DBU behaves differently from KOH, which under comparable solventless conditions (10 mol %; 423 K, 1 h) did not show any catalytic activity.¹⁸

Most likely, the very low selectivity to 4b was due to the modest stability of glycerol carbonate under the used conditions. The glycerolysis of 1 has been, therefore, investigated in THF, at milder temperatures, using a stoichiometric amount of the triol.

Figure 6 shows that, at 373 K, using a catalyst load ≥5 mol %, the quantitative depolymerization of 1 (∼0.20 g) in THF (2 mL) required a very short time (<1 h): moreover, a sensible increase of selectivity (90 → 96%) can be observed on decreasing the DBU load. With a catalyst load as low as 1 mol %, the formation of the cyclic carbonate was even more selective (∼99%), but the depolymerization process was slower and complete within 3.5 h.

Figure 6.

Glycerolysis of 1 (PC ≈ 0.20 g; glycerol: ∼60 μL, mol_glycerol/mol_CO3 ≈ 1) catalyzed by DBU, in THF (2 mL) at 373 K, with different catalyst loads.

The glycerolysis reaction proceeds quantitatively and selectively, within times acceptable from the applicative point of view, even at temperatures significantly lower than 373 K (Figure 7). At ambient temperature (293 K), with a 10 mol % DBU load, a carbonate yield close to 95% was achieved within 22 h. A comparable yield (97%) was obtained at 333 K after only 2.5 h by using an even lower catalyst load (5 mol %).

Figure 7.

Glycerolysis of 1 (PC ≈ 0.20 g; glycerol: ∼58 μL, mol_glycerol/mol_CO3 ≈ 1) catalyzed by DBU, in THF (2 mL), at different temperatures.

Also, in this case, the products (BPA and glycerol carbonate) were separated with high yield (≥90%) by chromatography, as described in 2.5.

4. Conclusions

In conclusion, under suitable conditions, DBU behaves as an active organocatalyst of PC glycolysis and promotes effectively and selectively the conversion of 1 with 1,2-PD or glycerol to give the monomer (BPA) and the relevant cyclic carbonate 4a or 4b, respectively.

The depolymerization process can be carried out under solventless conditions (using diol/triol as the reagent and reaction medium) quantitatively with selectivity to carbonate (4a or 4b), which depends on the used polyol (3a or 3b), temperature, catalyst load, and reaction time.

Very high selectivities to cyclic carbonate have been obtained by carrying out the depolymerization reaction in a cosolvent, such as THF, wherein the polymer is soluble, and using mild temperature conditions and a stoichiometric amount of polyol.

The depolymerization process does not require any complex apparatus and is simple from the operational point of view. The use of DBU as the catalyst of the process allows to conjugate high productivity with effective catalyst recyclability. Both BPA and the cyclic carbonates 4a or 4b have been isolated, at a laboratory scale, in high yield. The above features, while marking positively the described protocols, also provide a suitable starting point for challenging scale up, and, as a whole, make these processes attractive routes to chemical recycling and valorization of the waste polymer.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.8b01123.

• PC (SABIC) characterization (NMR, ATR–FTIR); ¹H and ¹³C NMR of the mixture of MHP-BPA derivatives; ¹H and ¹³C NMR of the mixture of BHP-BPA derivatives; and ¹H NMR of the BPA/DBU adduct (PDF)

The authors declare no competing financial interest.