Cyclic and Linear Polyhydroxylbutyrates from Ring-Opening

Dec 19, 2018 - ... on the positive resolution mode with a laser power set to 400 mW. ...... Effective Construction of Hyperbranched Multicyclic Polyme...
3 downloads 0 Views 2MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Cyclic and Linear Polyhydroxylbutyrates from Ring-Opening Polymerization of β‑Butyrolactone with Amido-Oxazolinate Zinc Catalysts Muneer Shaik, Jhaiquashia Peterson, and Guodong Du* Department of Chemistry, University of North Dakota, 151 Cornell Street Stop 9024, Grand Forks, North Dakota 58202, United States

Macromolecules Downloaded from pubs.acs.org by UNIV OF WINNIPEG on 12/19/18. For personal use only.

S Supporting Information *

ABSTRACT: A series of amido-oxazolinate zinc complexes have been employed for the ring-opening polymerization of rac-β-butyrolactone (BBL). Experimental results show that these complexes efficiently catalyze the reactions, yielding cyclic poly(3-hydroxybutyrate)s (PHBs) with high molecular weights (Mn up to 197 kg/mol) and relatively low dispersity. In comparison, in the presence of alcohol cocatalysts, the zinccatalyzed ROP reactions lead to the formation of linear PHBs end-capped by the alcohol initiator and hydroxylbutyrate. A possible mechanism for cyclic PHBs is proposed, in which the zinc catalysts function as a loose Lewis pair at elevated temperature, followed by a fast propagation through a zwitterionic intermediate. Use of diols such as 1,4-cyclohexanediol and 1,4-benzenedimethanol results in the formation of telechelic polyester diols. The thermal analysis by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) shows higher thermal resistance of cyclic PHBs than linear ones.



INTRODUCTION Growing environmental concerns of persistent plastic wastes have prompted the development of biodegradable and biocompatible polymers, which can be decomposed by living organisms without harming them and the environment.1−3 Among these polymers, poly(hydroxyalkanoates), such as poly(3-hydroxybutyrate) (PHB), a biodegradable aliphatic polyester, have attracted significant attention in applications ranging from green packaging to drug delivery systems.4−6 To further their applications, it is essential to control the polymer architectures, which can directly affect the physical properties. Ring-opening polymerization (ROP) of cyclic esters and ringopening copolymerization (ROCOP) of epoxides and cyclic anhydrides are promising techniques for the synthesis of polyesters in a well-controlled fashion, and a wide range of catalysts have been developed.7 Distinct types of metal catalysts such as chromium,8 zinc,9−11 aluminum,12 indium,13−15 tin,16,17 yttrium,18,19 and lanthanides20,21 bearing different supporting ligands have been employed for the ROP of β-butyrolactone (BBL) to synthesize PHBs with good control over molecular weights and stereoselectivity. While most of the resultant polyesters possess linear structure, various architectures such as star-shaped, grafted, cross-linked, and hyperbranched structures are possible with an appropriate choice of catalysts and initiators or chain transfer agents.22−25 Specially, macrocyclic polyesters have received considerable interest because of their appealing properties different from their linear counterparts, including glass © XXXX American Chemical Society

transition temperature (Tg), morphologies, melt viscosities, thermal stability, and biodistribution profile.26−28 However, it has been challenging to access cyclic polymers with good control, and most studies have focused on the ROP of lactide29−31 and ε-caprolactone.32−35 Waymouth and coworkers reported the cyclic polyesters using N-heterocyclic carbene (NHC) catalysts via zwitterionic polymerization of lactide and lactones.36−39 The zwitterionic interaction between the intramolecular end groups is the crucial step to furnish the cyclic structure. Interestingly, polymerization with similar catalyst systems in the presence of alcohol led to the formation of linear or branched polyesters. Inspired by the versatility of conventional β-diketiminate− zinc complexes in catalytic ROP and ROCOP,40,41 we have developed a chiral analogue derived from the amidooxazolinate framework (Figure 1)42 and demonstrated that their zinc complexes are effective catalysts for the ROP of lactide43 and ROCOP of epoxide and CO244 or cyclic anhydrides.45 Herein, we describe the amido-oxazolinate zinc complexes as active catalysts for the controlled ROP of BBL into cyclic PHBs. As far as we are aware, this represents a rare example of cyclic PHB synthesis from BBL with zinc catalysis. In the presence of mono- and bifunctional alcohol initiators such as ethyl alcohol (EtOH), benzyl alcohol (BnOH), and Received: October 2, 2018 Revised: November 29, 2018

A

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

min−1 and a flow rate of 100 mL min−1 of N2 (furnace purge gas). The differential scanning calorimetry (DSC) thermograms were collected on a PerkinElmer Pyris differential scanning calorimeter using a 10.0 °C/min heating rate from −30 to 160 °C with 20 mL/ min nitrogen flow. Electrospray ionization mass spectrometry (ESIMS) analysis was performed on a high-resolution time-of-flight G1969A instrument (Agilent). The analyte-containing solution was introduced into the instrument by direct infusion using a syringe drive (5 mL min−1). The electrospray ionization (e.g., capillary) and collision-induced dissociation (e.g., fragmentor) potentials were set to 3500 and 150 V, respectively. Acetic acid was used as an electrolyte at 25 mmol L−1. Mass Hunter Qualitative Analysis software was used for data processing. The MALDI HR-MS experiment was performed on a MALDI SYNAPT G2-Si Q-ToF instrument equipped with a solid state, high repetition laser (355 nm). The matrix solution was prepared by dissolving 10 mg of 2,5-DHB (2,5-dihydroxybenzoic acid) in 10 mL of acetonitrile, and the resulting solution was mixed with 20 mL of 0.1% TFA in acetonitrile. The polymer sample and matrix mixture was deposited on a sample plate to form a thin film after solvent evaporation. The mass spectra were obtained on the positive resolution mode with a laser power set to 400 mW. General Procedure for ROP of BBL. The following procedure is representative: an oven-dried 10 mL Schlenk flask equipped with a stir bar was charged with BBL (323 mg, 3.75 mmol, 200 equiv) and zinc catalyst (1 equiv) in toluene (4.0 mL) in a glovebox under nitrogen. The flask was capped and taken out and then heated in an oil bath preset at 100 °C. The reaction was monitored by 1H NMR spectroscopy until the complete conversion of BBL. After removal of the volatile components, the residue was dissolved in DCM (1−3 mL), followed by addition of hexane (4−5 mL). The precipitation of the polymeric products was facilitated by cooling briefly the flask in liquid nitrogen. The supernatant was decanted, and the residues were washed and dried under reduced pressure. The purified polymers were then characterized by various techniques including NMR, ESIMS, and GPC. Characterization data are compiled in the Supporting Information.

Figure 1. Chiral zinc amido-oxazolinate complexes.

1,4-cyclohexanediol (1,4-CHD), the same catalysts afford wellcontrolled linear PHBs with low dispersity.



EXPERIMENTAL SECTION

Materials and Methods. Reactions with air- and moisturesensitive compounds were performed under dry nitrogen using standard glovebox (VAC) and/or Schlenk line techniques. Deuterated solvents were purchased from Cambridge Isotope Laboratories. Analytical grade THF was purchased from Fisher Scientific and used as received. Other chemicals were purchased from SigmaAldrich. β-Butyrolactone was distilled over CaH2 following three freeze−pump−thaw cycles. CDCl3 and C6D6 were distilled over CaH2 and Na/benzophenone, respectively, and degassed prior to use. Toluene was distilled under nitrogen from Na/benzophenone. Syntheses of zinc complexes 1a−1d were conducted according to literature methods.46 NMR spectra (1D and 2D) were recorded on a Bruker AVANCE 500 NMR spectrometer and referenced to the residual peaks in CDCl3. The microstructures of PHBs were characterized by examination of the carbonyl region in the inverse gated 13C NMR spectra recorded at room temperature in CDCl3 with concentrations in the range 1−1.5 mg/mL. Pm values were determined by Pm = (Im/ Im + Ir), and Pr = (Ir/Im + Ir) where Ir and Im are the integrations of corresponding inverse gated 13C{1H} peaks according to the literature.47 Gel permeation chromatography (GPC) analysis was performed on a Varian Prostar instrument with autosampler model 400, using a PLgel 5 mm Mixed-D column, a Prostar 355 RI detector, and THF as eluent at a flow rate of 1 mL min−1 (20 °C). Polystyrene standards from Agilent Technologies were used for calibration. Galaxie software was used to operate the instrument, and Cirrus software was used for data processing. TGA was performed on an SDT Q 600 instrument with Advantage software. The samples in Al2O3 cups were heated from 30 to 600 °C with a ramp rate of 20 °C



RESULTS AND DISCUSSION ROP of β-Butyrolactone. ROP of rac-β-butyrolactone (BBL) using 1a as a catalyst was first examined, and the progress of the reaction was monitored by 1H NMR spectroscopy. As shown in Table 1, when the reaction was performed in toluene at 100 °C with 0.5 mol % catalyst 1a, a complete conversion of BBL was achieved within 90 min (entry 1), leading to PHB with high molecular weight (Mn = 27.0 kg/mol) and narrow dispersity (Đ = 1.08). Reactions in DCM (entry 2) and THF (entry 3) at 60 °C achieved only 15% and 20% conversion after 90 min, respectively, and 100% conversion was attained at longer times (24 h) with relatively lower molecular weights. When the reaction was performed under bulk conditions without solvent, a maximum conversion of 79% was reached in 30 min with Mn of 20.1 kg/mol and a

Table 1. Polymerization of rac-BBL with Amido-Oxazolinate Zinc Complexes entry

complex

solvent

Mnb (kg/mol)

time (min)

Đ

conv (%)

Pr/Pmc

yield (%)

1 2 3 4 5 6 7

1a 1a 1a 1a 1b 1c 1d

toluene DCM THF bulk toluene toluene toluene

27.0 7.1 17.3 20.1 26.7 17.5 33.1

85 1440 1440 30 90 90 90

1.08 1.13 1.49 2.56 1.27 1.20 1.38

100 100 100 79 100 100 100

46/54 57/43 46/54 51/49 48/52 45/55 47/53

40 57 59 27 58 42 56

Polymerizations are run with [M]/[Zn] = 200/1 in toluene and bulk at 100 °C and in THF and DCM at 60 °C (heating bath temperature). Determined by gel permeation chromatography calibrated with polystyrene standards in THF. cDetermined by inverse gated 13C NMR in the carbonyl region.

a

b

B

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 2. Ring-Opening Polymerization of β-Butyrolactone with 1a in Toluene entry

[M0]/[Zn]a

time (min)

Mn(calcd)b

Mn(GPC)c

Đc

conv (%)

yieldd

1 2 3 4 5 6 7 8

20:1 100:1 200:1 400:1 600:1 800:1 1000:1 2000:1

70 90 90 85 80 60 180 180

1.7 8.6 17.2 34.4 51.7 68.9 86.1 172.2

2.8 14.6 27.0 45.3 76.4 87.3 140.0 196.5

1.21 1.31 1.08 1.53 1.35 1.23 1.81 1.97

100 100 96 99 99 98 99 98

94 31 40 53 61 48 72 84

Polymerizations were run with [BBL]/[Zn] = 20/1 to 2000/1 in toluene (4 mL) at 100 °C. Molecular weights reported in kg/mol. bCalculated on the basis of conversion and catalyst loading. cDetermined by gel permeation chromatography calibrated with polystyrene standards in THF. d Isolated yields. a

Figure 2. NMR spectra of the cyclic PHB obtained with 1a (Table 2, entry 1). Top: 1H NMR spectrum. Bottom: IG-13C NMR spectrum. The inset shows the racemic and meso diads in the carbonyl region.

comparable, affording 100% conversion of monomer within 90 min. The steric hindrance increase at the R2 position seemed to lower the molecular weight. It was also noted that the isolated yields of PHBs were somewhat low due to the loss of materials in the purification by a dissolution−precipitation method.

broad dispersity (Đ = 2.56) (entry 4). Consequently, the subsequent polymerizations were conducted in toluene as solvent. Under these conditions, PHBs with high molecular weights and narrow Đ values were obtained when other zinc complexes 1b−1d were employed in the ROP of BBL (Table 1, entries 5−7). The activity of different zinc catalysts was C

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 3. Polymerization of BBL in the Presence of Alcohols as Initiatora

Polymerizations were run in toluene (4 mL) at 100 °C. The reactions were monitored by NMR until all BBL were consumed. Molecular weights reported in kg/mol. bCalculated on the basis of conversion and catalyst/initiator loading. cDetermined by the end group/main chain ratio in 1H NMR. dDetermined by gel permeation chromatography calibrated with polystyrene standards in THF. eIsolated yields. a

Figure 3. 1H (top) and 13C (bottom) NMR of PHB obtained with [BnOH]/[Zn] = 50 (Table 3, entry 3). The inset shows the alkenic protons of the trans-crotonate end group.

D

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 4. Polymerization of BBL with 1a in the Presence of Diolsa

a

Polymerizations were run in toluene (4 mL). The reactions were monitored by NMR until all BBL were consumed. Molecular weights reported in kg/mol. bCalculated on the basis of conversion and catalyst/initiator loading. cEstimated from the end group/main chain ratio in 1H NMR. d Determined by GPC calibrated with polystyrene standards in THF. eIsolated yields.

region around 169 ppm (Figure 2), which suggested that mostly atactic PHB was obtained, and the present catalysts were not stereoselective for ROP of rac-BBL. This was different from ROP of rac-lactides catalyzed by these same zinc complexes, in which highly isotactic polylactides were obtained.43 Similar contrasts have been noted for the classical diketiminate zinc complex catalyzed ROP of BBL and lactides.48,49 Synthesis of Linear PHBs with 1a/Alcohols. Next, ROP of BBL was performed with catalyst 1a in the presence of an alcohol as cocatalyst. Varied amounts of ethanol (EtOH) and benzyl alcohol (BnOH) were used, and the results are summarized in Table 3. In all cases 100% conversions were achieved within 3 h, and the resulting PHBs were largely atactic, as determined by IG-13C NMR. Good agreements between theoretical and experimental molecular weights and relatively narrow dispersities (Đ = 1.09−1.74) were noted. Up to 50 equiv of alcohols (versus catalyst) could be employed, and the molecular weights of the polymers could be controlled by changing the Zn/alcohol ratios, indicating that alcohols also acted as chain transfer agents in the reaction. Higher loading of alcohol resulted in the formation of oligomeric PHBs, though the polymerization became slower (entries 3 and 7). The microstructure of the resulting PHBs was investigated by NMR and ESI-MS techniques. Along with the four large 1H NMR peaks at 1.3 ppm (methyl), 2.5 and 2.6 ppm (methylene), and 5.2 ppm (methine) assignable to the main chain repeating butyrate units, small yet consistent peaks were notable that could be attributed to the end groups (Figure S7). A representative 1H NMR spectrum of an oligomeric PHB sample obtained with 50 equiv (versus Zn) BnOH initiator (entry 7 in Table 3) is shown in Figure 3. The peaks at 5.1 and 7.3 ppm could be assigned to the benzylic and aryl protons of the benzyloxyl group at one end, whereas the other end of the chain was characterized by a terminal hydroxybutyrate unit featuring a singlet at 4.1 ppm for methine protons and a broad peak at 3.2 ppm for hydroxyl protons (other signals overlapped with large main chain signals). The integration of the hydroxyl (3.2 ppm) versus benzylic (5.1 ppm) protons was 1:2, in agreement with them being the two ends of the chain. In the 1 H−1H COSY NMR spectrum (Figure S8), only one weak

We further examined the activity of 1a by varying the BBL/ Zn ratios from 20:1 to 2000:1. Regardless of the catalyst loading, the polymerization generally reached completion within 90 min (Table 2). For example, at [BBL]/[Zn] = 400 (entry 4), the polymerization yielded PHB with Mn = 45.3 kg/ mol and 100% conversion. Only at low catalyst loadings ([BBL]/[Zn] = 1000 and 2000) did the reaction require longer reaction time (∼3 h). Under these conditions, high molecular weight PHBs could be obtained: Mn = 140.0 kg/mol at 99% conversion for [BBL]/[Zn] = 1000 (entry 7), and Mn = 196.5 kg/mol at 98% conversion for [BBL]/[Zn] = 2000 (entry 8). The experimental molecular weights determined by GPC of the isolated PHBs were generally higher than the Mn calculated from the initial catalyst loading and conversion. At the same time, the molecular weights increased with the [BBL]/[Zn] ratio in a roughly linear manner (Figure S1). These observations suggested that the initiation step was slow compared to propagation in the polymerization, and not all catalysts were actively involved in the initiation, but the polymerization was reasonably controlled. The microstructure of the resultant PHBs was evaluated by NMR spectroscopic techniques. Two peaks at 2.45 and 2.56 ppm in the 1H NMR spectrum (Figure 2) were assigned to the two diastereotopic methylene protons, whereas the peaks at 1.24 and 5.22 ppm were assigned to methyl and methine protons, respectively. These data agreed with the expected PHB main chain structure; however, the most remarkable feature was the absence of any other signals typically associated with chain ends. A similar observation was noted in the 13C NMR, where only four signals corresponding to the main chain carbons were present (Figure 2). In other words, no indication of any end groups could be detected, regardless of the molecular weights of the polymers. Lack of end groups suggested the formation of cyclic PHBs in the reaction, which was supported by an ESI-MS experiment (Figure S22): the primary series of peaks of 86 Da difference between consecutive peaks corresponded to the PHB main chain and could be assigned to the n(C4H6O2) + (CH3CN) series, in agreement with the macrocyclic polymer structure. Additionally, inverse-gated 13C (IG-13C) NMR showed approximately equal intensities of meso and racemic signals in the carbonyl E

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 4. 1H (top) and 13C (bottom) NMR spectra of isolated PHB with [1,4-BDM]/[Zn] = 0.5 (Table 4, entry 8) at 80 °C.

of BBL were carried out at 100 °C in toluene (Table 4) in the presence of variable amounts of diols such as 1,4-benzenedimethanol (1,4-BDM), 1,6-hexanediol (1,6-HD), and 1,4cyclohexanediol (1,4-CHD). As expected, the reactions reached completion within 3 h, and linear PHBs were isolated in good to high yields. The molecular weights generally agreed with the theoretical values determined from the loadings of zinc and diols, demonstrating the ROP processes were wellcontrolled under these conditions. When large excess of diols (20 equiv of 1,4-BDM or 1,4-CHD) were employed, the reactions did not proceed to give PHBs, likely due to catalyst deactivation. A representative 1H NMR spectrum of the PHB synthesized from 1,4-BDM (Table 4, entry 8) is shown in Figure 4. Along with the major main chain peaks, similar small resonances peaks at 3.1 and 4.1 ppm for the hydroxyl butyrate group and 5.0 and 7.3 ppm for the BDM unit were also observed, which supported that the diol unit was incorporated between the two growing chains with the terminal hydroxybutyrate units. Although it was possible that diol could be present as a chain end with only one of the two hydroxyl groups reacting, the lack of the NMR signals for the unreacted corresponding benzylic protons (around 4.7 ppm) suggested that the diol was incorporated in the chain. α,ω-Dihydroxyl

cross-peak with methine proton at 4.1 ppm was observed for the hydroxyl protons (3.2 ppm), whereas two additional, stronger cross-peaks with methylene (2.3 ppm) and methyl protons (1.2 ppm) were noted for methine protons. In addition, the 3.2 ppm signal showed no correlation with 13C signals in the 1H−13C HETCOR NMR spectrum (Figure S9). In the 13C NMR, besides the main chain carbons, the rest of the peaks can be assigned to benzyl and hydroxybutyrate end groups (Figure 3). Together, these observations supported that the polymer end groups were hydroxybutyrate at one end and benzyloxyl at the other. The linear structure and the end groups were further established by the ESI-MS analysis, which showed a major series of peaks at 107 + 86n + 1 + 23 that could be assigned to the BnO + n(C4H6O2) + H + Na+ structure (Figure S23). ROP of BBL in the Presence of Diol Initiators. Given the previous observations that simple alcohols such as BnOH led to formation of benzyl and hydroxyl end-capped polyesters, we sought to use the ROP reaction to produce α,ω-dihydroxyl end-capped PHBs by employing diol initiators. Such polyester diols, or telechelic polymers in general, are of great interest and have found uses in various applications such as thermoplastic elastomers and polyurethane coatings.50 Thus, ROP reactions F

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

interactions (or a Lewis pair) as in zwitterionic ring-opening polymerization36,51 or by a tethered initiating group as in ring expansion polymerization.52−54 The present zinc catalysts can be viewed as a classical Lewis adduct. The Lewis acidity of the zinc center is significantly attenuated, as they do not react with bases such as pyridine at room temperature.46 However, at elevated temperature, the Lewis adduct may loosen up to become a “loose” pair that initiates the polymerization. 1H NMR studies (Figure 5) of compound 1a in C6D6 at variable

end-capped PHBs were also obtained with an aliphatic diol, 1,4-CHD, which indicated that secondary alcohols were effective initiators in these reactions as well. In this case, the butyrate end group was characterized by the hydroxyl (3.1 ppm) and methine (4.1 ppm) peaks, while the incorporation of CHD unit was evidenced by the α-OCH signal of the C6H10 ring at 4.75 ppm (Figure S14). Again, the lack of the NMR signals for the unreacted α-OHCH (part of C6H10OH) protons (around 4.0 ppm) suggested that the diol was incorporated in the chain, not at the chain end. These assignments were further supported by the 13C and 1H−1H COSY NMR spectra as well as the ESI-MS analysis of the isolated products. The ESI-MS spectrum of BDM-derived PHB showed a dominant series of peaks at 136 + 86n + 2 + 23 that could be assigned to the structure of C8H8O2 + n(C4H6O2) + 2H + Na+ (Figure S24). Similarly, the ESI-MS spectrum of CHD derived polymer showed a dominant series of peaks at 114 + 86n + 2 + 23 that could be assigned to the C6H10O2 + n(C4H6O2) + 2H + Na+ structure (Figure S25). It was noted that the integrations of diol unit and the end group signals deviated somewhat from the expected 1:2 ratio. Close inspection of the 1H NMR spectra revealed the presence of two minor peaks at 5.7 and 6.9 ppm when 1,4-BDM and 1,4-CHD were used (Figure 4). These signals could be assigned to the alkenic protons in the trans-crotonate group, presumably derived from the dehydration of alcohols, and they could amount to up to half of the end groups. These assignments are further supported by the presence of two alkenic carbon peaks at 123 and 145 ppm in the 13C NMR. The minor series of peaks in the ESI-MS spectrum (Figure S24) could be attributed to the product with crotonate end groups (18 Da less than the main series of peaks). Similar dehydration has been reported in the polymerization of BBL in the presence of BnOH, which was attributed to elimination reaction promoted by the metal complexes.49 In our case, the crotonate end groups were mostly noticeable when diols were used as initiators, and only minimal formation (∼2%) of crotonate signals was detected when BnOH was used as initiator (see Figure 3). Because our goal was to produce polyester diols, we attempted to explore the conditions to minimize the dehydration reaction. When the reaction was allowed to proceed at a lower temperature (80 °C), no crotonate groups were observed in the initial reaction mixture. However, after work-up the crotonate signals showed up in the NMR spectrum. Prolonged reaction times and repeated precipitation of the polymer products during purification also tended to increase the percentage of crotonate groups. Mechanistic Consideration of β-Butyrolactone Polymerization by Zn Complexes. The cyclic polymers are commonly obtained from two pathways. In the first case, a linear precursor bearing a good leaving group was initially produced, and backbiting (or ring closing) under high dilution conditions in the end furnished the cyclic structure. This is usually accompanied by the presence of both linear and cyclic polymers in the products. The present zinc catalysts feature a silylamido −N(SiMe3)2 that could serve as an initiating group; however, it tends to be sluggish due to its steric bulk. Furthermore, the chain end groups were not observed even at low conversion of BBL, and cyclic PHBs were obtained exclusively. These observations were inconsistent with such a pathway. In the second scenario, which could lead to exclusive formation of cyclic polymers, the two ends of the chain were never truly separated. They were held together by electrostatic

Figure 5. Partial 1H NMR of catalyst 1a at variable temperatures in C6D6.

temperatures (from 25 to 100 °C) showed a gradual shift of signals with increasing temperature, and the largest shifts occurred for the oxazoline ring protons in the 3.4−4.4 ppm region (up to >0.2 ppm; see Figure S21 and Table S1 for details). This suggested that the imino nitrogen of the oxazoline underwent most significant changes in their environment with increasing temperature, presumably by weakening of the zinc−nitrogen bond, or a partial dissociation, leading to a loose Lewis pair. The notion that the chelating imino nitrogen weakens most instead of the monodentate silylamido nitrogen is not unreasonable, considering that Nimino is neutral while Namido is anionic. In the solid structure of 1a, the Zn−Nimino distance (1.968(9) Å) is significantly longer than the Zn−Namido distance (1.874(9) Å).46 The Zn and Nimino atoms are supposedly still associated, but the pair is loose enough to allow for coordination of the BBL monomer, followed by the typical insertion/propagation steps through a zwitterionic intermediate (Scheme 1). The scenario is analogous to the cyclic polyesters synthesis via ROP promoted by Zn(C6F5)2 in the presence of a Lewis base, despite the generation of Lewis adducts between Zn(C6F5)2 and Lewis bases.55−57 Additionally, a solution of 1a and BBL (molar ratio 1:1 or 1:15) in C6D6 showed no reaction at room temperature, while BBL was converted to polymers at elevated temperature (100 °C). In this process, only signals of 1a, BBL, and PHB were observed throughout the course of the reaction, showing no presence of any intermediate species. This suggested that the initiation step is slow relative to the propagation, which is in agreement with the GPC analysis that the resulting G

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

thus behaving as such. The glass transition temperature (Tg) determined by DSC for cyclic PHB is 7.6 °C, also slightly higher than the Tg values of linear PHBs (1.9−4.6 °C). These observations provide additional support that cyclic polymers have been obtained in the zinc-catalyzed ROP of BBL in the absence of alcohols.

Scheme 1. Proposed Mechanism for the Formation of Cyclic PHBs



CONCLUSION In summary, a series of Zn complexes have been demonstrated to be effective in polymerizing BBL. Cyclic PHBs with high molecular weight (up to 197 kg/mol) were obtained in the absence of an alcohol, whereas the addition of alcohol cocatalyst resulted in the formation of linear PHBs with welldefined end groups. On the basis of experimental observations and catalyst structures, a Lewis pair derived zwitterionic intermediate was proposed for the formation of cyclic polymers. The ROP reaction was employed to synthesize α,ω-dihydroxyl end-capped PHBs with diol initiators, which were incorporated into the polymer chain. However, the mechanistic details, especially those related to the cyclic polymer control and generation, warrant further investigation, which will be the focus of future efforts, as well as improving the stereocontrol of the ROP process and exploring the potential applications of cyclic architectures and polyester diols.

molecular weights are higher than the calculated values based on catalyst loadings and conversions (see Table 1). In the presence of an alcohol cocatalyst, the first step would likely be the substitution of the silylamide by an alkoxide, which can function as a better initiating group than the silylamide. The rest of the reaction proceeds through a typical coordination−insertion mechanism, leading to the formation of linear polyesters end-capped by the initiating alkoxide at one end and hydroxyl butyrate at the other. The absence of any detectable signals for benzyl ether protons in BnOH-initiated PHBs supported that only acyl−oxygen scission occurred in the presence of 1a, in accord with the general depiction of metal-catalyzed ROP reactions. Thermal Properties of Cyclic and Linear PHBs. The thermal properties of the PHBs have been studied by TGA and DSC methods, with special attention to the possible difference between cyclic and linear PHBs. In the literature, it has been reported that the cyclic topology markedly increases the thermal stability of polymers.58 Indeed, the linear PHBs in our case demonstrate the onset of thermal degradation at a lower temperature than the cyclic PHB. As summarized in Table 5



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.8b02096.



Experimental details including characterization data and selected spectra (PDF)

AUTHOR INFORMATION

Corresponding Author

*Tel +1-701-777-2241; fax +1-701-777-2331; e-mail guodong. [email protected]. ORCID

Guodong Du: 0000-0002-3833-233X

Table 5. Thermal Properties of Cyclic and Linear Poly(hydroxybutyrate)sa,b entry Table Table Table Table Table Table

1, 2, 2, 3, 3, 3,

entry entry entry entry entry entry

1 1 5 2 5 8

Notes

polymer

Tg

T−1%

T−5%

T−50%

T−99%

cyclic PHB EtO-PHB BnO-PHB HD-PHB CHD-PHB BDM-PHB

7.6 1.9 2.7 2.3 4.6 3.0

269 182 208 156 243 240

285 277 279 275 273 278

296 296 294 298 289 295

316 351 304 317 367 373

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation under Grant NSF EPSCoR Award IIA1355466. J.P. thanks the NSF for providing REU support through Grant CHE 1460825. We thank Mr. J. Hatton and Dr. A. Kubatova for help with ESI-MS analysis, Dr. A. Ugrinov for MALDI-MS analysis, and Dr. E. Kolodka for help with GPC.

a Temperatures in °C. bTg values were determined from the second heating cycle in DSC. T−1%, T−5%, T−50%, and T−99% refer to the temperatures at which 1%, 5%, 50%, and 99% weight losses were observed in TGA, respectively.



REFERENCES

(1) Kunduru, K. R.; Basu, A.; Domb, A. J. Biodegradable Polymers: Medical Applications. In Encyclopedia of Polymer Science and Technology; Wiley: New York, 2016. (2) Rathna, G. V. N.; Ghosh, S. Bacterial Polymers: Resources, Synthesis and Applications. In Biopolymers: Biomedical and Environmental Applications; Kalia, S., Avérous, L., Eds.; Wiley Online Library: 2011; pp 291−316. (3) Gross, R. A.; Kalra, B. Biodegradable Polymers for the Environment. Science 2002, 297, 803−807.

and Figures S27−S29, the thermal degradation temperature as characterized by T−5% for linear PHBs occurs around 273−279 °C, similar to the value reported for atactic linear PHBs; in comparison, the T−5% for cyclic PHB lies around 285 °C. We also note that the Tmax values for cyclic and linear PHBs are comparable, which actually makes sense: once the first breakage of a cyclic PHB occurs, it becomes a linear polymer, H

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (4) Brannigan, P. R.; Dove, P. A. Synthesis, properties and biomedical applications of hydrolytically degradable materials based on aliphatic polyesters and polycarbonates. Biomater. Sci. 2017, 5, 9− 21. (5) Yeo, J. C. C.; Muiruri, J. K.; Thitsartarn, W.; Li, Z.; He, C. Recent advances in the development of biodegradable PHB-based toughening materials: Approaches, advantages and applications. Mater. Sci. Eng., C 2018, 92, 1092−1116. (6) Philip, S.; Keshavarz, T.; Roy, I. Polyhydroxyalkanoates: Biodegradable Polymers with a range of Applications. J. Chem. Technol. Biotechnol. 2007, 82, 233−247. (7) Longo, J. M.; Sanford, M. J.; Coates, G. W. Ring-Opening Copolymerization of Epoxides and Cyclic Anhydrides with Discrete Metal Complexes: Structure−Property Relationships. Chem. Rev. 2016, 116, 15167−15197. (8) Zintl, M.; Molnar, F.; Urban, T.; Bernhart, V.; Preishuber-Pflügl, P.; Rieger, B. Variably Isotactic Poly(hydroxybutyrate) from Racemic β-Butyrolactone: Microstructure Control by Achiral Chromium(III) Salophen Complexes. Angew. Chem., Int. Ed. 2008, 47, 3458−3460. (9) Jaffredo, C. G.; Carpentier, J. F.; Guillaume, S. M. Poly(hydroxyalkanoate) Block or Random Copolymers of β−Butyrolactone and Benzyl β−Malolactone: A Matter of Catalytic Tuning. Macromolecules 2013, 46, 6765−6776. (10) Guillaume, C.; Carpentier, J. F.; Guillaume, S. M. Immortal Ring Opening Polymerization of β-Butyrolactone with Zinc Catalysts: Catalytic Approach to Poly(3-hydroxyalkanoate). Polymer 2009, 50, 5909−5917. (11) Kramer, J. W.; Coates, G. W. Fluorinated β-lactones and Poly(β-hydroxyalkanoate)s: Synthesis via Epoxide Carbonylation and Ring Opening Polymerization. Tetrahedron 2008, 64, 6973−6978. (12) Garcia-Valle, M. F.; Tabernero, V.; Cuenca, T.; Mosquera, M. E. G.; Cano, J.; Milione, S. Biodegradable PHB from rac-βButyrolactone: Highly Controlled ROP Mediated by a Pentacoordinated Aluminum Complex. Organometallics 2018, 37, 837−840. (13) Osten, K. M.; Mehrkhodavandi, P. Indium Catalysts for Ring Opening Polymerization: Exploring the Importance of Catalyst Aggregation. Acc. Chem. Res. 2017, 50, 2861−2869. (14) Quan, S. M.; Diaconescu, P. L. High activity of an indium alkoxide complex toward ring opening polymerization of cyclic esters. Chem. Commun. 2015, 51, 9643−9646. (15) Xu, C.; Yu, I.; Mehrkhodavandi, P. Highly controlled immortal polymerization of β-butyrolactone by a dinuclear indium catalyst. Chem. Commun. 2012, 48, 6806−6808. (16) Kemnitzer, J. E.; McCarthy, S. P.; Gross, R. A. Syndiospecific Ring Opening Polymerization of β-Butyrolactone to form Predominantly Syndiotactic Poly(β-hydroxybutyrate) using Tin(IV) Catalysts. Macromolecules 1993, 26, 6143−6150. (17) Kricheldorf, H. R.; Eggerstedt, S. Polylactones. 41. Polymerizations of β-D, L-Butyrolactone with Dialkyltinoxides as Initiator. Macromolecules 1997, 30, 5693−5697. (18) Ajellal, N.; Bouyahyi, M.; Amgoune, A.; Thomas, C. M.; Bondon, A.; Pillin, I.; Grohens, Y.; Carpentier, J. F. Syndiotacticm Enriched Poly(3-hydroxybutyrate)s via Stereoselective Ring Opening Polymerization of Racemic β-Butyrolactone with Discrete Yttrium Catalysts. Macromolecules 2009, 42, 987−993. (19) Ajellal, N.; Thomas, C. M.; Carpentier, J. F. Functional Syndiotactic Poly(β-hydroxyalkanoate)s via Stereoselective Ring Opening Copolymerization of rac-β-butyrolactone and rac-allyl-βbutyrolactone. J. Polym. Sci., Part A: Polym. Chem. 2009, 47, 3177− 3189. (20) Zhuo, Z.; Zhang, C.; Luo, Y.; Wang, Y.; Yao, Y.; Yuan, D.; Cui, D. Stereo-selectivity switchable ROP of rac-β-butyrolactone initiated by salan-ligated rare-earth metal amide complexes: the key role of the substituents on ligand frameworks. Chem. Commun. 2018, 54, 11998− 12001. (21) Ajellal, N.; Durieux, G.; Delevoye, L.; Tricot, G.; Dujardin, C.; Thomas, C. M.; Gauvin, R. M. Polymerization of Racemic βbutyrolactone using Supported Catalysts: A Simple Access to Isotactic Polymers. Chem. Commun. 2010, 46, 1032−1034.

(22) Ebrahimi, T.; Hatzikiriakos, S. G.; Mehrkhodavandi. Synthesis and Rheological Characterization of Star Shaped and Linear Poly(hydroxybutyrate). Macromolecules 2015, 48, 6672−6681. (23) Du, Y.; Yan, H.; Huang, W.; Chai, F.; Niu, S. Unanticipated Strong Blue Photoluminescence from Fully Biobased Aliphatic Hyperbranched Polyesters. ACS Sustainable Chem. Eng. 2017, 5, 6139−6147. (24) Linhardt, A.; Kö nig, M.; Iturmendi, A.; Henke, H.; Brüggemann, O.; Teasdale, I. Degradable, Dendritic Polyols on a Branched Polyphosphazene Backbone. Ind. Eng. Chem. Res. 2018, 57, 3602−3609. (25) Ren, J. M.; McKenzie, T. G.; Fu, Q.; Wong, E. H. H.; Xu, J.; An, Z.; Shanmugam, S.; Davis, T. P.; Boyer, C.; Qiao, G. G. Star Polymers. Chem. Rev. 2016, 116, 6743−6836. (26) Zhu, Y.; Hosmane, N. S. Advanced Developments in Cyclic Polymers: Synthesis, Applications, and Perspectives. ChemistryOpen 2015, 4, 408−417. (27) Kricheldorf, H. R. Cyclic Polymers: Synthetic Strategies and Physical Properties. J. Polym. Sci., Part A: Polym. Chem. 2010, 48, 251−284. (28) Endo, K. Synthesis and Properties of Cyclic Polymers. Adv. Polym. Sci. 2008, 217, 121−183. (29) Jeong, W.; Shin, E. J.; Culkin, D. A.; Hedrick, J. L.; Waymouth, R. M. Zwitterionic Polymerization: A Kinetic Strategy for the Controlled Synthesis of Cyclic Polylactide. J. Am. Chem. Soc. 2009, 131, 4884−4891. (30) Kricheldorf, H. R.; Lomadze, N.; Schwarz, G. Cyclic Polylactides by Imidazole-Catalyzed Polymerization of L-Lactide. Macromolecules 2008, 41, 7812−7816. (31) Hoskins, J. N.; Grayson, S. M. Cyclic polyesters: synthetic approaches and potential applications. Polym. Chem. 2011, 2, 289− 299. (32) Brown, H. A.; Xiong, S. L.; Medvedev, G. A.; Chang, Y. A.; Abu-Omar, M. M.; Caruthers, J. M.; Waymouth, R. M. Zwitterionic Ring-Opening Polymerization: Models for Kinetics of Cyclic Poly(caprolactone) Synthesis. Macromolecules 2014, 47, 2955−2963. (33) Chang, Y. A.; Waymouth, R. M. Ion pairing effects in the zwitterionic ring opening polymerization of δ-valerolactone. Polym. Chem. 2015, 6, 5212−5218. (34) Shin, E. J.; Jeong, W.; Brown, H. A.; Koo, B. J.; Hedrick, J. L.; Waymouth, R. M. Crystallization of Cyclic Polymers: Synthesis and Crystallization Behavior of High Molecular Weight Cyclic Poly(εCaprolactone)s. Macromolecules 2011, 44, 2773−2779. (35) Phomphrai, K.; Pongchan-o, C.; Thumrongpatanaraks, W.; Sangtrirutnugul, P.; Kongsaeree, P.; Pohmakotr, M. Synthesis of High-Molecular-Weight Poly(ε-Caprolactone) Catalyzed by Highly Active Bis(amidinate)tin(II) Complexes. Dalton Trans 2011, 40, 2157−2159. (36) Culkin, D. A.; Jeong, W.; Csihony, S.; Gomez, E. D.; Balsara, N. P.; Hedrick, J. L.; Waymouth, R. M. Zwitterionic Polymerization of Lactide to Cyclic Poly(Lactide) by Using N-Heterocyclic Carbene Organocatalysts. Angew. Chem., Int. Ed. 2007, 46, 2627−2630. (37) Brown, H. A.; De Crisci, A. G.; Hedrick, J. L.; Waymouth, R. M. Amidine-Mediated Zwitterionic Polymerization of Lactide. ACS Macro Lett. 2012, 1, 1113−1115. (38) Brown, H. A.; Waymouth, R. M. Zwitterionic Ring-Opening Polymerization for the Synthesis of High Molecular Weight Cyclic Polymers. Acc. Chem. Res. 2013, 46, 2585−2596. (39) Shin, E. J.; Brown, H. A.; Gonzalez, S.; Jeong, W.; Hedrick, J. L.; Waymouth, R. M. Zwitterionic copolymerization: synthesis of cyclic gradient copolymers. Angew. Chem., Int. Ed. 2011, 50, 6388− 6391. (40) Ajellal, N.; Carpentier, J. F.; Guillaume, C.; Guillaume, S. M.; Helou, M.; Poirier, V.; Sarazin, Y.; Trifonov, A. Metal-catalyzed immortal ring-opening polymerization of lactones, lactides and cyclic carbonates. Dalton Trans 2010, 39, 8363−8376. (41) Jeske, R. C.; DiCiccio, A. M.; Coates, G. W. Alternating Copolymerization of Epoxides and Cyclic Anhydrides: An Improved I

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Route to Aliphatic Polyesters. J. Am. Chem. Soc. 2007, 129, 11330− 11331. (42) Binda, P. I.; Abbina, S.; Du, G. Modular synthesis of chiral βdiketiminato-type ligands containing 2-oxazoline moiety via Palladium-catalyzed amination. Synthesis 2011, 2609−2618. (43) Abbina, S.; Du, G. Zinc-catalyzed Highly Isoselective Ring Opening Polymerization of rac-Lactide. ACS Macro Lett. 2014, 3, 689−692. (44) Abbina, S.; Chidara, V. K.; Bian, S.; Ugrinov, A.; Du, G. Synthesis of Chiral C2-symmetric Bimetallic Zinc Complexes of Amido-Oxazolinates and Their Application in Copolymerization of CO2 and Cyclohexene Oxide. ChemistrySelect 2016, 1, 3175−3183. (45) Abbina, S.; Chidara, V. K.; Du, G. Ring Opening Copolymerization of Styrene Oxide and Cyclic Anhydrides Using Highly Effective Zinc Amido-Oxazolinate Catalysts. ChemCatChem 2017, 9, 1343−1348. (46) Abbina, S.; Du, G. Chiral Amido-Oxazolinate Zinc Complexes for Asymmetric Alternating Copolymerization of CO2 and Cyclohexene Oxide. Organometallics 2012, 31, 7394−7403. (47) Ebrahimi, T.; Aluthge, D. C.; Hatzikiriakos, S. G.; Mehrkhodavandi, P. Highly Active Chiral Zinc Catalysts for Immortal Polymerization of β-Butyrolactone Form Melt Processable SyndioRich Poly(hydroxybutyrate). Macromolecules 2016, 49, 8812−8824. (48) Chamberlain, B. M.; Cheng, M.; Moore, D. R.; Ovitt, T. M.; Lobkovsky, E. B.; Coates, G. W. Polymerization of Lactide with Zinc and Magnesium β-Diiminate Complexes: Stereocontrol and Mechanism. J. Am. Chem. Soc. 2001, 123, 3229−3238. (49) Rieth, L. R.; Moore, D. R.; Lobkovsky, E. B.; Coates, G. W. Single-Site β-Diiminate Zinc Catalysts for the Ring-Opening Polymerization of β-Butyrolactone and β-Valerolactone to Poly(3hydroxyalkanoates). J. Am. Chem. Soc. 2002, 124, 15239−15248. (50) Barouti, G.; Guillaume, S. M. Polyhydroxybutyrate (PHB)based triblock copolymers: synthesis of hydrophobic PHB/poly(benzyl β-malolactonate) and amphiphilic PHB/poly(malic acid) analogues by ring-opening polymerization. Polym. Chem. 2016, 7, 4603−4608. (51) Piedra-Arroni, E.; Ladaviere, C.; Amgoune, A.; Bourissou, D. Ring-Opening Polymerization with Zn(C6F5)2-Based Lewis Pairs: Original and Efficient Approach to Cyclic Polyesters. J. Am. Chem. Soc. 2013, 135, 13306−13309. (52) Xia, Y.; Boydston, A. J.; Yao, Y. F.; Kornfield, J. A.; Gorodetskaya, I. A.; Spiess, H. W.; Grubbs, R. H. Ring-Expansion Metathesis Polymerization: Catalyst-Dependent Polymerization Profiles. J. Am. Chem. Soc. 2009, 131, 2670−2677. (53) Weil, J.; Mathers, R. T.; Getzler, Y. D. Y. L. Lactide Cyclopolymerization by an Alumatrane-Inspired Catalyst. Macromolecules 2012, 45, 1118−1121. (54) Chang, Y. A.; Waymouth, R. M. Recent progress on the synthesis of cyclic polymers via ring-expansion strategies. J. Polym. Sci., Part A: Polym. Chem. 2017, 55, 2892−2902. (55) Li, X.-Q.; Wang, B.; Ji, H.-Y.; Li, Y.-S. Insights into the mechanism for ring-opening polymerization of lactide catalyzed by Zn(C6F5)2/organic superbase Lewis pairs. Catal. Sci. Technol. 2016, 6, 7763−7772. (56) Li, X.; Chen, C.; Wu, J. Lewis Pair Catalysts in the Polymerization of Lactide and Related Cyclic Esters. Molecules 2018, 23, 189−202. (57) Wang, B.; Pan, L.; Ma, Z.; Li, Y. Ring-Opening Polymerization with Lewis Pairs and Subsequent Nucleophilic Substitution: A Promising Strategy to Well-Defined Polyethylene-like Polyesters without Transesterification. Macromolecules 2018, 51, 836−845. (58) Hong, M.; Chen, E. Y.-X. Completely recyclable biopolymers with linear and cyclic topologies via ring-opening polymerization of γbutyrolactone. Nat. Chem. 2016, 8, 42−49.

J

DOI: 10.1021/acs.macromol.8b02096 Macromolecules XXXX, XXX, XXX−XXX