Controlled Organocatalytic Ring-Opening Polymerization of ε-Thionocaprolactone

Abstract

For the first time, the controlled ring-opening polymerization (ROP) of ε-thionocaprolactone (tnCL) is conducted. The organocatalytic ROP of tnCL occurs without carbonyl scrambling, leading to homopoly(ε-thionocaprolactone) (PtnCL). The ROP by base catalysts alone is proposed to proceed via a nucleophilic mechanism, while the addition of an H-bond donating thiourea (TU) is shown to provide excellent reaction control. The increased reaction control provided by the TU occurs in the virtual absence of binding between tnCL and TU, and a mechanistic account for this observation is discussed. The monomer ring strain is measured and found to be similar to δ-valerolactone (VL). Copolymers with VL are synthesized, and the resulting analysis of the copolymer materials properties provides the only known physical characterizations of poly(thio(no)ester-co-ester)s.

Graphical abstract

INTRODUCTION

Over the past decade, developments in organocatalytic ring-opening polymerization (ROP) have demonstrated the remarkable ability of these catalysts to generate well-defined and highly functionalized polyesters and other polymers.^1–3 The H-bond-mediated organocatalysts, which are a paragon of highly controlled polymerization techniques, stand out in their ability to generate precisely tailored materials.^4–6 These catalysts are believed to operate by H-bond-mediated activation of monomer by thiourea and of growing polymer chain by base.^7,8 Given the remarkable control of these polymerization systems and the mild nature of their reactivity, the paucity of polymer backbone linkages which have been explored is surprising. The mild reactivity of organocatalysts for ROP perfectly position them for the generation of new polymer backbones and, hence, new materials and applications.

Our group recently disclosed the ROP of an S-substituted lactone, ε-thiocaprolactone (tCL).⁹ The ROP of this monomer was postulated to proceed through a classic transesterification mechanism mediated by H-bond activation of thioester by thiourea and thiol end group by base.⁹ The other S-substituted caprolactone, ε-thionocaprolactone (tnCL, Scheme 1), has been the subject of only two published reports.^10,11 Under cationic polymerization conditions, the ROP of tnCL proceeds with quantitative inversion of substitution at the thionoester to generate the same poly(thiocaprolactone) previously reported by Overberger and our group.^9,13,14 The anionic ROP of tnCL from alkyllithium reagents retains the S-carbonyl substitution, but reaction control suffers, and this method does not allow for M_n control, copolymerization, or end group selection.¹⁰ Partial inversion of substitution occurs with weak nucleophiles, resulting in a mixed polymer backbone. Polymerization conditions which result in inversion of S/O substitution are postulated to operate via an S_N2 propagation (Scheme 1).¹⁰ If the chain end/monomer activation mechanism of H-bond-mediated ROP of tCL is correct, we reasoned that organocatalytic ROP of tnCL should allow for the retention of the S/O substitution and controlled-generation of homopoly(thionocaprolactone).

Scheme 1.

Endo’s Anionic and Cationic ROPs of tnCL Have Been Shown To Proceed with S/O Scrambling

RESULTS AND DISCUSSION

Organic Base Catalyzed ROP

The organocatalytic room temperature ROP of tnCL initiated from alcohol or thiol initiators proceeds with retention of the S/O substitution. The ε-thionocaprolactone (tnCL) is generated from ε-caprolactone (CL) via a one-step reaction with P₄S₁₀ (see Experimental Section); the reaction is workable on at least a 2 g scale (75% yield).^15,16 The application of DBU (5 mol %; Table 1) to a C₆D₆ solution of tnCL (2 M) and octadecylthiol (1 mol %) results in 90% conversion to polymer in 25 h. ¹³C NMR analysis of isolated tnCL suggests quantitative retention of the thiono-substitution, forming poly(ε-thionocaprolactone) (PtnCL) (see Supporting Information, Figure S8). The room temperature application of DBU for the ROP of tnCL from benzyl alcohol results in a linear evolution of M_n versus time and an initially narrow M_w/M_n that broadens throughout the ROP (Figure 1). The guanidine bases, MTBD and TBD, were also applied for the ROP of tnCL. MTBD exhibited a similar activity in the ROP of tnCL to that of DBU whereas TBD effected a faster but less controlled ROP, resulting in erosion of M_w/M_n (Table 1).

Table 1.

ROP of tnCL with Base Catalysts^a

entry	base	convb (%)	time (h)	Mnc (g/mol)	Mw/Mnc
1d	TBD	>99	0.33	17 700	1.41
2	DBU	89	25	17 200	1.30
3	MTBD	90	23	14 100	1.24
4	BEMP	0	25
5e	DBU	98	20	12 800	1.37
6f	DBU	98	19	19 700	1.20

^aReaction conditions: 2 M (0.77 mmol, 1 equiv) tnCL, 1 mol % octadecylthiol, 5 mol % base, and C₆D₆.

^bConversion to polymer obtained by ¹H NMR.

^cDetermined by GPC (CH₂Cl₂) vs polystyrene standards.

^d1 mol % TBD.

^eInitiation was from benzyl alcohol (1 mol %).

^finitiation was from 1-pyrenebutanol (1 mol %).

Figure 1.

Evolution of M_n vs conversion for the (upper) DBU catalyzed ROP of tnCL from benzyl alcohol; (lower) 1/BEMP (5 mol % each) catalyzed ROP of tnCL (2 M, 100 mg, 1 equiv) from benzyl alcohol (1 mol %).

PtnCL was previously only available through the application of alkyllithiums at elevated temperatures which resulted in the uncontrolled ROP of tnCL.¹⁰ Endo’s ROP of tnCL initiated from DBU at elevated temperatures was more controlled than the alkyllithium ROP but resulted in scrambling of the S/O substitution (Scheme 1).^{10 13}C NMR analysis of the polymer resulting from the repetition of our DBU-catalyzed ROP experiment at high temperature (1 equiv of tnCL, 5 mol % DBU, toluene, 100 °C) in the presence of an alcohol initiator reveals that S/O scrambling does occur (see Supporting Information, Figure S9), which suggests that the thiono/thio switching observed by Endo¹⁰ is simply due to heating the reaction solution.

The application of the phosphazene base, BEMP (Table 1), to a C₆D₆ solution of tnCL and benzyl alcohol does not result in ROP. The observation that the considerably more basic (vs DBU or MTBD) but non-nucleophilic BEMP (BEMP-H⁺; pK_a^MeCN = 27.6)¹⁷ does not effect ROP suggests a nucleophilic mode of action for DBU (DBU-H⁺; pK_a^MeCN = 24.3)¹⁸ and MTBD (MTBD-H⁺; pK_a^MeCN = 25.4)¹⁸ in the ROP of tnCL (Scheme 2). DBU and MTBD have previously been suggested to operate as nucleophiles in ROP.¹⁹ Conducting the DBU-catalyzed ROP of tnCL from alcoholic initiators (either benzyl alcohol or 1-pyrenebutanol) results in minimally altered M_w/M_n compared to when initiated from octadecylthiol (Table 1). This suggests no reduction in the “living” character of the ROP due to slower initiation from alcohols vs thiols.

Scheme 2.

Proposed Mechanism for the DBU Catalyzed ROP of tnCL

Organic Base and Thiourea Catalyzed ROP

The presence of thiourea 1 (Table 2) has a distinct impact upon the base cocatalyzed ROP of tnCL. The 1/DBU (5 mol % each) cocatalyzed ROP of tnCL from octadecylthiol in C₆D₆ lowers the reaction time and M_w/M_n versus the ROP with DBU alone. A similar effect is observed for MTBD (Table 2). The most striking results are observed with BEMP, which exhibits no activity in the absence of 1, but the 1/BEMP (5 mol % each) catalyzed ROP of tnCL (2 M, 1 equiv) from benzyl alcohol (1 mol %) achieves full conversion in 5 h. The reaction is highly controlled, exhibiting the characteristics of a “living” polymerization: linear evolution of M_n vs conversion, narrow M_w/M_n (=1.10) (Figure 1), first-order evolution of [tnCL] vs time (see Figure S6), and a M_n that is predictable from [M]₀/[I]₀ (Table 2), although polymers begin to become insoluble in benzene at elevated molecular weight (M_n ≥ 22 000). The ROP reaction proceeds in methylene chloride and chloroform but experiences reduced reaction control. Despite the narrow M_w/M_n, the GPC traces taken throughout the polymerization show the gradual growth of a high molecular weight tail, resulting in slight erosion of M_w/M_n (see Figure S2 and Figure 1). ¹³C NMR analysis of the isolated polymer confirms the quantitative retention of thiono-ester moieties, and the MALDI-TOF mass spectrum is consistent with linear benzyl alcohol-terminated PtnCL (see Figure S1). When initiated from 1-pyrenebutanol, the refractive index and UV GPC traces of PtnCL overlap, including the high molecular weight tail (see Figure S2), suggesting end group fidelity. When allowed to stir past full conversion, the high weight tail on the GPC trace grows in prominence and eventually merges with the “main” polymer peak (see Figure S2). The high molecular weight tail in the GPC trace arises from an unknown postpolymerization reaction.

Table 2.

Thiourea plus Base Cocatalyzed ROP of tnCL^a

entry	initiator	base	convb (%)	time (h)	Mnc (g/mol)	Mw/Mnc
1	octadecylthiol	DBU	93	17	22 500	1.14
2	octadecylthiol	MTBD	92	19	14 800	1.16
3	octadecylthiol	BEMP	84	5	14 600	1.25
4	1-pyrenebutanol	BEMP	89	6	15 200	1.11
5	benzyl alcohol	BEMP	91	5	12 000	1.14
6d	benzyl alcohol	BEMP	89	4.5	8 200	1.11
7e	benzyl alcohol	BEMP	90	7	20 900	1.10

^aReaction conditions: 2 M (0.77 mmol, 1 equiv) tnCL, 1 mol % octadecylthiol, C₆D₆ and given amount of catalyst.

^bConversion to polymer obtained by ¹H NMR.

^cDetermined by GPC (CH₂Cl₂) vs polystyrene standards.

^d2 mol % benzyl alcohol, [M]₀/[I]₀ = 50.

^e0.5 mol % benzyl alcohol, [M]₀/[I]₀ = 200.

Role of Thiourea in ROP

The presence of the thiocarbonyl in tnCL was expected to perturb the ability of H-bond donors to activate the monomer for ROP, yet the addition of thiourea 1 to the ROP solution clearly affects the course of the reaction. An NMR titration study^20–22 was conducted in C₆D₆ to determine the binding constant between 1 and tnCL, K_eq = 1.6 ± 0.2 (in eq 1). The comparable binding between CL and TU was measured to be K_eq = 42 ± 5,⁷ and a similarly dramatic perturbation from this latter strong binding value was previously measured for tCL, K_eq = 2.7 ± 0.5.⁹ The remarkable ability of 1 to activate tnCL and tCL toward ROP despite the weak binding exhibited by 1 toward these monomers suggests an incongruity in the approximation of “magnitude of binding” as “extent of activation”.

(eq 1)

The clear effects of 1 upon the ROP of tnCL in the absence of strong binding suggests that 1 plays a mechanistic role that cannot be fully understood by the magnitude of a binding constant between 1 and monomer. Because 1 does not measurably bind to ethyl acetate (a surrogate for open polymer),⁸ an approximation of the kinetic bias of 1 for polymerization vs depolymerization (or transesterification) is at most the magnitude of the 1/monomer binding constant, K_eq = 1.6, or ΔΔG^‡ = 0.27 kcal/mol (see Figure S7). This would only be true if 1 activates s-cis and s-trans esters equally. Indeed, DFT (B3LYP/6-31G**) calculations suggest that 1 is equally effective at the activation of (i.e., increasing the electophilicity of)¹⁰ CL, tnCL, tCL, methyl thionoacetate, and methyl acetate. For these compounds, both the electrostatic charges at carbon (C═X) and polarity of the C═X bond increase by ~5–10% upon the binding of 1 (see Figure S15). The effects rendered by 1 upon ROP must then be due to interactions in the transition state that are not adequately reflected in the magnitude of the binding of 1 to monomer (vs polymer) in the reactants/products. The increased reaction control provided by 1 during the ROP of tnCL could arise from the suppression of transesterification events due to prominent secondary interactions (e.g., 1 to base cocatalyst).^5,22 These results suggest that despite minimal binding to tnCL (or tCL), the H-bond mediated ROP of tnCL is operative by dual activation of monomer by 1 and of chain end by base (Scheme 3).

Scheme 3.

Proposed Mechanism for the H-Bond-Mediated ROP of tnCL

Thionocaprolactone vs Other Monomers

The terminal conversion of the DBU catalyzed ROP of tnCL from alcohol initiators showed a strong temperature dependence; we sought to measure the thermodynamics of ROP to energetically place tnCL among other cyclic (thio)lactone monomers. The equilibrium monomer concentration, [tnCL]_eq, of a TBD catalyzed ROP of tnCL from benzyl alcohol in C₆D₆ was monitored as a function of temperature (see Experimental Section). The resulting Van’t Hoff plot allowed for the extraction of the thermodynamic parameters of ROP:²³ ΔH° = −5.79 ± 0.32 kcal/mol (298 K); ΔS° = −13.5 ± 1.0 cal/(mol K); [tnCL]_eq = 0.05 M at 300 K, and T_ceiling = 156 °C (see Figure S3). These values suggest that tnCL is most energetically similar to VL:²³ T_ceiling = 149 °C. For comparison, the ceiling temperatures (T_ceiling) of CL and tCL are T_ceiling(CL)²³ = 261 °C and T_ceiling(tCL)⁹ = 7000 °C. The low ceiling temperature of tnCL accounts for the low monomer conversions which are observed when the ROP of tnCL is attempted at elevated temperatures.¹⁰ Kinetically, tnCL is more reactive than VL. VL will not undergo ROP in the presence of MTBD or DBU alone (no 1), and the increased reactivity of tnCL (vs VL) is attributed to the increased electophilicity of thionoesters (vs esters). In contrast, the thioester, tCL, was observed to exhibit behavior that is both more and less reactive than VL.⁹

Copolymerization with δ-Valerolactone

The observation of similar ROP thermodynamics for tnCL and VL suggests that random copolymerizations of these two monomers are possible. When 1/BEMP (5 mol % each) is added to a mixture of VL (1 M, 0.5 equiv), tnCL (1 M, 0.5 equiv) and benzyl alcohol (1 mol %) in C₆D₆, both monomers are observed to undergo ROP at approximately equal rates in a first-order evolution of [monomer]s vs time plot (k_tnCL/k_VL = 1.07), suggesting random copolymer formation (see Figure S4). ¹³C NMR analysis of the copolymer confirms random monomer incorporation as evidenced by the equal intensities of the well-resolved tnCL–tnCL vs tnCL–VL resonances (72.15 vs 71.78 ppm) (see Figure S10). The monomer feed can be adjusted to higher or lower VL/tnCL ratios to give gradient copolymers. ¹³C NMR analyses also confirm the retention of C═S substitution in the copolymers (see Figure S10). Whereas PtnCL is an oil at room temperature for all molecular weights examined in our lab (<20 kg/mol), copolymers of tnCL and VL with greater than 70% VL are solid at room temperature. The materials properties (T_m, T_c, and T_deg) of P(tnCL-co-VL) with varying tnCL content were analyzed by differential scanning calorimetry (DSC) and thermal gravimetric analysis (TGA, under N₂). Polymers with increasing tnCL content show predictably reduced T_m, T_c, and T_deg (Table 3). The hydrolytic stability of copolymers was measured under basic, acidic, and neutral conditions by established methods²⁵ (Figure 2). Increased tnCL content in copolymers with VL is associated with reduced hydrolytic stability under basic conditions, increased stability toward hydrolysis under acidic conditions, and minimally altered stability in neutral water. These observations are consistent with general trends of thio(no)ester stability.¹² To our knowledge, these are the only known characterizations of poly(thionoester-co-ester)s.

Table 3.

Copolymers of tnCL and VL with Varying Monomer Feeds^a

entry	tnCL (% feed)	VL (% feed)	time (h)	convb (%)	Mnc	Mw/Mnc	Tmd (°C)	Tcd (°C)	Tdege (°C)
1	0	100	5	0:93	12 300	1.06	53	27	380
2	5	95	4	56:90	19 600	1.02	49	22	440
3	10	90	5	73:93	19 200	1.02	43	22	360
4	20	80	4	56:90	19 200	1.03	40	8	340
5	30	70	5	79:96	18 200	1.05	31	−8	320
6	50	50	5	95:92	29 800	1.25	18	n/a	310
7f	100	0	7	89:0	20 900	1.10	9	n/a	260

^aPolymerization conditions: 4 M ([VL] + [tnCL]) (2 mmol total), 2.5 mol % 1/BEMP (each), 0.5 mol % benzyl alcohol in C₆D₆.

^bPercent conversion to polymer obtained by ¹H NMR.

^cDetermined by GPC (CH₂Cl₂) vs polystyrene standards.

^dDetermined by DSC (N₂); no T_g were observed >−70 °C, the limit of our DSC.

^eDetermined by TGA (N₂).

^fPolymerization conditions: tnCL (2M, 1 mmol), 5 mol % 1/BEMP (each), 1 mol % benzyl alcohol in C₆D₆. n/a = not observed above −70 °C, the limit of our DSC.

Figure 2.

Percent mass loss for PtnCL and copolymers with VL in acidic (0.25 M HCl), basic (0.25 M NaOH), and neutral (distilled water) conditions vs time. The error from multiple measurements is ±5%.

CONCLUSION

The organocatalytic ROP of tnCL exhibits the characteristics of a “living” polymerization, particularly in the presence of an H-bond donating thiourea, 1. The marked effect of 1 upon the course of the ROP is notable because it occurs in the absence of a strong binding between the H-bond donor and monomer. The increase in reaction rate and reaction control for the ROP of tnCL in the presence of 1 cannot be accounted for by the traditional model of selectivity in the differential ability of 1 to bind to monomer or polymer, and further studies are required to elucidate the source of selectivity, presumably due to interactions in the transition state. Copolymers of tnCL with VL are to our knowledge the first reported and characterized copolymers of S-lactone and lactone monomers. The incorporation of tnCL to construct random (statistical), gradient, or block copolymers with traditional esters offers a unique and convenient method for tuning materials properties for custom tailored applications; the multitude of possible copolymers provides a wealth of research opportunities.

EXPERIMENTAL SECTION

General Considerations

All chemicals were used as received unless stated otherwise. P₄S₁₀, hexamethyldisiloxane (HMDO), and 2-tert-butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorine (BEMP) were supplied by Acros Organics. Acetonitrile, potassium carbonate, magnesium sulfate, benzyl alcohol, benzoic acid, ethyl acetate, and hexane were purchased from Fisher Scientific. ε-Caprolactone (CL) and δ-valerolactone (VL) were supplied by Alfa Aesar and distilled from CaH₂ under high vacuum. Benzene-d₆ was supplied by Cambridge Isotope Laboratories and distilled from CaH₂ under a nitrogen atmosphere. Benzyl alcohol was distilled from CaH₂ under high vacuum. 1 [3,5-bis(trifluoromethyl)phenyl]-3-cyclohexylthiourea was synthesized and purified according to a literature procedure.⁷ 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU), 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene (MTBD), and 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) were purchased from TCI. All polymerization reactions were performed in an MBRAUN stainless steel glovebox equipped with a gas purification system under a nitrogen atmosphere using glass vials and stir bars which were baked overnight at 140 °C. NMR experiments were performed on a Bruker Avance 300 MHz (proton) spectrometer. The chemical shifts for proton (¹H) and carbon (¹³C) NMR were recorded in parts per million (ppm) relative to a residual solvent. Size exclusion chromatography (SEC) was performed at 30 °C in dichloromethane (DCM) using an Agilent Infinity GPC system equipped with three Agilent PLGel columns 7.5 mm × 300 mm (5 µm pore sizes: 10³, 10⁴, and 10⁵ Å). Molecular weight and M_w/M_n were determined versus polystyrene standards (500 g/mol–3150 kg/mol; Polymer Laboratories). Differential scanning calorimetry (DSC) curves were obtained on a DSC Q100 (TA Instruments) under N₂ calibrated with an indium standard. The heating and cooling curves of DSC were run under a nitrogen atmosphere at a heating rate of ±10 °C/min in a 40 µL aluminum crucible. Thermogravimetric analysis (TGA) was performed using a TGA Q500 from TA Instruments under a N₂ atmosphere at a heating rate of 20 °C/min from 25 to 600 °C. MALDI-TOF MS analysis was performed at the University of Akron: reflectron mode with trans-2-[3-(4-tert-butylphenyl)-2-methyl-2-propenylidene]-malononitrile (DCTB) matrix with sodium trifluoroacetate (NaTFA) salt.

Example Synthesis of ε-Thionocaprolactone (TnCL)

In a dried 100 mL round-bottom flask with stir bar, P₄S₁₀ (1.95 g, 4.38 mmol), ε-caprolactone (1.94 mL, 17.52 mmol), HMDO (6.18 mL, 29.08 mmol), and acetonitrile (17.5 mL) were added. The solution was refluxed for 1 h while stirring, and the reaction was cooled to room temperature with stirring. The flask was then placed in an ice–water bath, and aqueous K₂CO₃ solution (1.26 mL of 5.3 M solution per mmol of P₄S₁₀) was added followed by distilled water (1 mL per mmol of P₄S₁₀) with stirring. Once cooled to room temperature, organics were extracted with dichloromethane and washed with brine. The organics were dried over MgSO₄, filtered, and removed of volatiles to obtain crude product. The crude product was purified in two stages by silica gel flash chromatography: 75:25 toluene:hexanes mobile phase and a second column with 95:5 dichloromethane:hexanes mobile phase. Kugelrohr vacuum distillation at 80 °C and 200 mTorr pressure yielded tnCL, a pure yellow oil which was transferred to the glovebox. Yield: 1.7 g, 75%. Characterization matched the literature.²⁴

Example Ring-Opening Polymerization

In a typical polymerization, tnCL (0.100 g, 0.768 mmol) was added to a 20 mL glass vial with a stir bar. In another 20 mL glass vial with stir bar, 1 (0.0142 g, 0.0384 mmol), BEMP (11.1 µL, 0.0384 mmol), and benzyl alcohol (1.6 µL, 15.4 µmol) were added. Solvent (for C₆D₆, 0.38 mL, 2 M in tnCL) was divided equally between the two vials. After 2 min of stirring, the tnCL solution was transferred with a pipet to the other vial consisting of initiator and catalysts. The solution was then transferred via pipet to a NMR tube and taken out of the glovebox. Reaction conversion was monitored by ¹H NMR, and the reaction was quenched with benzoic acid (2 mol equiv to base) after desired conversion had been achieved. The polymer was treated with hexanes to precipitate the polymer. After decanting the hexanes supernatant, the polymer was removed of volatiles under reduced pressure. Yield 85%; M_w/M_n = 1.11; M_n(GPC) = 8200; M_n(NMR) = 5200. For [M]₀/[I]₀ = 200, M_w/M_n = 1.10; M_n(GPC) = 20 900; M_n(NMR) = 15 800. ¹H and ¹³C NMR spectra (see Figures S12 and S13) show resonances consistent with assignment of the polymer as quantitatively thionoester repeat unit, with the characteristic thiocarbonyl carbon peak at 224 ppm in the ¹³C spectrum. ¹H NMR (300 MHz, CDCl₃) δ: 7.38 (5H, aromatic); 5.48 (2H, benzylic); 4.43 (78H, R-CH₂-R′ of PtnCL); 2.73 (80H, R-CH₂-R′ of tnCL); 1.80 (159 H, R-CH₂-R′ of tnCL); 1.46 (88H, R-CH₂-R′ of PtnCL). ¹³C NMR (75 MHz, CDCl₃) (see Figure S13) δ: 224 (R-C═S-R′); 128 (benzyl aromatics); 72 (1C, Ar-CH₂-OR); 62 (benzylic CH₂); 47 (R-CH₂-O); 28 (2C, R-C═SCH₂-CH₂-R′); 25 (R-C═S-CH₂-CH₂-CH₂-R′).

Example Copolymerization of ε-Thionocaprolactone with δ-Valerolactone

VL (100 mg, 1 mmol) and tnCL (130 mg, 1 mmol) were added to a 20 mL glass vial with a stir bar. In another 20 mL glass vial with stir bar, 1 (0.05 mmol), BEMP (0.05 mmol), and benzyl alcohol (0.01 mmol) were added. Solvent (for C₆D₆, 0.5 mL) was added in equal proportion to both the vials. The monomer solution was transferred via pipet to the vial containing initiator after few minutes of stirring. The solution was then transferred to a NMR tube which was immediately taken out of the glovebox. Reaction progress was monitored by ¹H NMR, and the reaction was quenched with benzoic acid (2 mol equiv to base). The polymer was treated with hexanes to precipitate the polymer, the supernatant was decanted, and polymer was subjected to reduced pressure to remove volatiles. The polymer samples were dissolved in methylene chloride and dialyzed against methanol over 48 h, changing the methanol solution after 24 h. Yield 90%; M_w/M_n = 1.25; M_n(GPC) = 29 800. ¹H NMR (300 MHz, C₆D₆) (see Figure S11) δ: 7.30 (5H, aromatic); 5.08 (2H, benzylic); 4.40 (164H, R-CH₂-R′ PVL); 4.04 (176H, R-CH₂-R′ PtnCL); 2.69 (164H, R-CH₂-R′ PVL); 2.32 (176H, R-CH₂-R′ PtnCL); 1.85–1.53 (m, 680H, R-CH₂-R′ PtnCL and PVL); 1.40 (176H, R-CH₂-R′ PtnCL). ¹³C NMR (75 MHz, CDCl₃) (see Figure S10) δ: 224 (1C, R-C═S-R′); 173 (1C, R-C═O-R′); 128 (6C, aromatic C’s); 72.15 and 71.78 (ε-CH₂, tnCL–tnCL and tnCL–VL); 64.21 and 63.90 (δ-CH₂, VL–VL, and tnCL–VL); 53.5 (α-CH₂, VL); 46.6 (α-CH₂, tnCL); 33.7 (γ-CH₂, VL); 28.32, 28.07, 27.80, 27.57 (γ-CH₂ and δ-CH₂, tnCL–tnCL and tnCL–VL); 25.2 (β-CH₂, tnCL); 21.4 (β-CH₂, VL).

Determination of Binding Constant (K_eq) between TnCL and 1

The binding constant (K_eq) between 1 and tnCL was determined in benzene-d₆ by the titration method and curve fitting as previously described.^20–22 The K_eq values were determined by fitting the binding curve to the quadratic form of the binding equation with K_eq and Δδ as variables: δ_obs = δ_H – (Δδ/2[H]₀){[H]₀ + [G]₀ + 1/K – (([H]₀ + [G]₀ + 1/K)² – 4[H]₀[G]₀)^1/2}; Δδ is the difference in the chemical shift of host and complex, δ_obs is the observed chemical shift of the TU in the presence of monomer, and δ_H is the chemical shift of free TU in the absence of monomer.

Determination of Thermodynamic Parameters

A sample of tnCL (100 mg, 0.77 mmol), TBD (5.3 mg, 0.038 mmol), and benzyl alcohol (0.80 µL, 0.0077 mmol) in C₆D₆ (2 M in monomer) was heated in a variable temperature NMR probe, and ¹H NMR spectra were acquired at temperatures from 290 to 330 K. Data points were taken in duplicate, during heating and cooling. The heating and cooling [M]_eq values were within error; heating values are shown in the Supporting Information (Figure S3). The thermodynamic values of tnCL ROP were determined from a Van’t Hoff plot of the data, and error was calculated from linear regression at 95% confidence interval (see Figure S3).

Polymer Hydrolysis

Polymer samples (approximately 10 mg each) were loaded into empty 20 mL scintillation vials. The polymers were then dissolved in dichloromethane to evenly distribute the polymer on the bottom of the vial, and the dichloromethane was subsequently removed under vacuum. Each vial was charged with 10 mL of aqueous 0.25 M HCl, aqueous 0.25 M NaOH solution, or distilled water. Each hydrolysis medium was tested in quadruplicate. All vials were shaken on a rotary shaker for the duration of the study. To take a data point, the solutions were removed via syringe, and the polymer samples were rinsed with minimal distilled water (~1 mL). After removing the distilled water via syringe, the vials were put in a vacuum oven overnight (60 °C, 30 in.Hg vacuum). The vials were cooled and weighed. Percent mass loss is given by mass₀ – mass_i/mass₀. The same steps were repeated over a three week period at different intervals.

Computational Methods

Computational experiments using Endo’s methods¹⁰ were performed in Spartan ’14 (Windows 7). Structures were geometry optimized at the DFT B3LYP/6-31G* level of theory in the gas phase. Energies and electrostatic charges in toluene solvent were calculated as single point energies (DFT B3LYP/6-31G**) from the DFT-optimized structures. Energies, electrostatic charges, computed structures, and coordinates of optimized structures are given in the Supporting Information (Figure S15).