Thermoplastics are indispensable to modern society, yet the environmental concerns surrounding plastic production and waste accumulation have inspired efforts to rethink how plastics are synthesized and treated at end of life. [1][2][3][4] While they are inexpensive to manufacture and have exceptional properties, petroleum-based polymers are responsible for more than 2 billion metric tons of CO2 equivalent emissions and 300 million metric tons of waste generated annually. 5- 7 The environmental costs associated with plastics have motivated a reimagining of the plastics economy based on circular materials, with an emphasis on utilizing bio-derived polyesters that can be recycled or degraded back to CO2 and H2O at end-of-life . [8][9][10][11] Poly(lactic acid) (PLA) currently comprises more than 70% of global bio-based polyester production and is valued for its ability to be degraded under industrial composting conditions. However, while PLA has high modulus, strength, and optical clarity, it suffers from brittleness -fracturing at elongations < 10%. [12][13][14][15][16] Numerous strategies to improve the toughness of PLA have been reported, but these often suffer an accompanying loss in thermal or mechanical performance. [12][13][14][15][16] For bio-derived polyesters to replace petroleum-based plastics, advances are needed to achieve optimal thermomechanical properties without sacrificing other performance metrics.
Bicyclic lactones have emerged as a promising class of monomer to produce highperformance polyesters. [17][18][19][20][21][22][23][24] Compared to traditional monocyclic lactones, bicyclic lactones produce polyesters with enhanced thermal properties, such as higher glass transition temperatures (Tg). [17][18][19][20][21][22][23][24] While ring opening polymerization (ROP) of bicyclic lactones has been investigated for decades, [21][22][23][24][25][26][27][28] little emphasis has been placed on bicyclic lactones derived from renewable feedstocks. Recent reports detailing the polymerization of lactones derived from camphor, (±)norcamphor, and 3-carene are rare examples which investigate polyesters from bio-derived bicyclic lactones. 22,25,26 In this work, we investigate the synthesis and polymerization of the bicyclic lactone 2-oxo-3,8-dioxabicyclo[3.2.1]octane (ODO) from bio-derived 5-hydroxymethyl furoic acid (HMFA). 27 Polymerization of ODO yields poly(oxo-3,8-dioxabicyclo[3.2.1]octane) (PODO), a ductile polyester with a moderately high Tg. To identify the impact of the tetrahydrofuran moiety, a monocyclic analog of ODO, ε-caprolactone (CL), is used as a comparison point for monomer polymerizability and material properties of the corresponding polymers PODO and poly(ε-caprolactone) (PCL). By leveraging the ductility and relatively high Tg of PODO, we demonstrate that statistical copolymers of PODO and PLA can retain the strength, stiffness, and Tg of PLA, while substantially increasing the material toughness (Figure 1. Comparing feedstock, ductility, Tg, stiffness, and strength of a) PLA, b) PCL, c) PODO, and d) PLA-stat-PODO.), illustrating how ODO can significantly improve the performance of a vital biobased plastic.
ODO has previously been synthesized from HMFA via hydrogenation and acid-catalyzed cyclization in a 20% overall yield. 28 We sought to improve this synthesis by examining these reactions in greater detail. The hydrogenation of HMFA catalyzed by Rh/C in H2O proceeds in high yield with 10 bar H2 at ambient temperature to form the desired cis-5-hydroxymethyl-2tetrahydrofuroic acid (HMTA), but 5-10% of the hydrogenolysis product 5-methyl-2tetrahydrofuroic acid (MTA) is also formed under these conditions (Figure S1). Previous studies of the hydrogenation of benzyl alcohols showed that Rh/C resulted in more hydrogenolysis than Rh supported on other substrates. 29 Switching to Rh/Al2O3 for HMFA hydrogenation reduced the reaction time, allowed for lower catalyst loading, and provided a small selectivity benefit, resulting in 98% HMTA yield on a 1 g scale (Table S1). Applying the optimized reaction conditions on a 10 g scale, however, led to some loss in selectivity, resulting in a 91% yield of HMTA (Figure 2). The Rh/Al2O3 catalyst was readily recycled for multiple hydrogenation batches without noticeable loss in reactivity.
The cyclization of HMTA to ODO was investigated using a variety of heterogeneous and homogeneous acid catalysts across several solvents (Tables S2-S4). The highest yields were obtained using p-toluenesulfonic acid (p-TSA), which is commonly used for the cyclization of hydroxy acids. [30][31][32] Using optimized conditions of 1 mol% p-TSA in refluxing toluene with a Dean-Stark apparatus to remove water, ODO was obtained in 18% isolated yield on a 10 g scale after purification by using a silica plug and performing recrystallization (Figure S3-Figure S9). NMR analysis of the material remaining after ODO isolation suggests that oligomers (o-PODO) are the major side products of the reaction. Inspired by industrial lactide synthesis, 33 deoligomerization of o-PODO was investigated as a method to improve overall ODO yield. The crude mixture of oligomers from HMTA cyclization were directly deoligomerized in a short path cold finger distillation set up using a tin catalyst and glycerol ethoxylate (GEO) additive. Approximately 45% of the o-PODO side products can be recovered as ODO in 2 h at 200 °C (Table S5), which enables a combined overall yield of 61% for the cyclization of HMTA to ODO. Starting from HMFA, we have demonstrated a 56% isolated yield of ODO on a 10 g scale. We anticipate further improvements upon increasing the scale and optimizing the reaction vessel for the cyclization. Notably, this route uses no stoichiometric reagents other than H2 (Figure 2). 33,34 Figure 2. Synthesis of ODO from HMFA via hydrogenation, cyclization, and deoligomerization.
While ODO has been previously polymerized utilizing titanium (IV) tert-butoxide in the melt, the kinetics and thermodynamics of ODO polymerization have not been reported and characterization of the corresponding polymer has been limited to 1 H NMR and IR spectroscopy. 28 Organocatalytic ring opening polymerization (OROP) was used to investigate the thermodynamics and kinetics of lactone polymerization. [35][36][37][38][39][40][41][42][43] Several catalytic systems for OROP of ODO were investigated. The combination of 1,3-bis(3,5-bis(trifluoromethyl)phenyl)urea (Schreiner's urea catalyst, SU) and 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) with a benzyl alcohol (BnOH) initiator was identified as the optimal catalyst system (Table S6). Optimization of the reaction conditions revealed THF to be a convenient solvent due to low catalyst solubility in other solvents. In addition, low catalyst loading (1 mol%) was found to be sufficient for ODO polymerization at higher concentration or longer reaction times (Table S7). Under the optimized conditions, the OROP of ODO exhibits a linear increase in Mn with conversion and narrow dispersity, thereby indicating the SU/DBU catalyst system facilitates controlled living polymerization (Figure 3). PODO structure was confirmed using NMR spectroscopy ( 1 H, 13 C, HSQC, HMBC, COSY) and MALDI (Figure S10-Figure S15). To investigate the impact of the tetrahydrofuran moiety on polymerizability, the thermodynamics and kinetics of polymerization for ODO were compared to CL. The thermodynamics of polymerization were studied by measuring the equilibrium monomer concentrations of each lactone at different temperatures by NMR spectroscopy (Table S8-Table S9). The equilibrium monomer conversions were then fit using a Van't Hoff analysis to obtain the enthalpy and entropy of polymerization in THF solution, ΔHp and ΔSp, respectively (Figure 4a). 36 These data reveal that the polymerization of the bicyclic ODO (ΔHp,ODO = -25.2 kJ/mol, ΔSp,ODO = -60 J/mol•K, ΔGp,ODO(293K) = -8 kJ/mol) is slightly less thermodynamically favorable than that of CL polymerization (ΔHp,CL = -29.3 kJ/mol, ΔSp,CL = -60 J/mol•K, ΔGp,CL(293K) = -10 kJ/mol) in THF at 293K. The lower ΔHp of ODO suggests that incorporation of the tetrahydrofuran ring into the monomer backbone decreases the amount of ring strain released during polymerization. The decreased ΔHp suggests that ring-closing depolymerization of PODO would be more feasible due to reduced thermal requirements, making PODO amenable to chemical recycling.
To study whether PODO can be chemically recycled, PODO depolymerization was investigated using a similar method as was optimized for deoligomerization, with an ultra-short path distillation apparatus and a tin catalyst (Table S14). When depolymerizing PLA, GEO has been shown to promote transesterification and increase the rate of depolymerization. 44 With SnO as the catalyst and the addition of GEO, PODO was depolymerized to ODO at 180 °C with a polymer-to-monomer yield of 66% (Figure S16, Table S14). While ring-closing thermodynamics would indicate that depolymerization of PODO should be more facile than PCL, decomposition reactions at 180-200 °C limit polymer-to-monomer yield.
While CL has a larger thermodynamic driving force for polymerization, ODO polymerizes more rapidly at room temperature. Where a polymerization of ODO (1 mol% SU, 1 mol% DBU,
Conversion (%) 0 20 40 60 80 100 M n,GPC (kDa) 0 20 10 15 5 Ð 1.0 1.2 1.4 1.6 1.8 2.0 r 2 = 0.99
1 mol% BnOH, 1 M) concentration reaches 90% conversion in approximately 6 hours, CL takes nearly three times as long to achieve comparable conversion under the same conditions (Figure 4b). The kinetic regime of each polymerization was fit to a first-order kinetics model, from which the initial apparent rate constant (kinit) was extracted (Figure 4b). Interestingly, kinit for ODO was observed to be approximately five-fold higher than that of CL. To obtain high molecular weight PODO, melt polymerization of ODO was investigated (Table S10). Melt polymerization of ODO with catalyst Sn(Oct)2 and initiator BnOH readily produced PODO up to approximately 60 kDa with molecular weight controlled by initiator loading (Table 1). However, when targeting molecular weights exceeding 100 kDa, the theoretical molecular weight controlled by initiator loading (Mn,theor) fell below the observed molecular weight measured by GPC (Mn,GPC). To obtain molecular weights exceeding 100 kDa, ODO was polymerized using only small catalyst loadings without the addition of an initiator (Table 1, Table S11). While the absence of an initiator limits control over molecular weight, high molecular weight PODO (Mn,GPC > 200 kDa) was successfully obtained. The thermodynamics of melt polymerization were studied by measuring the equilibrium monomer concentration of samples polymerized at different temperatures. The resulting enthalpy of polymerization was comparable to that obtained in solution (Figure S19, Table S12). Thermal and mechanical properties of PODO were examined relative to PCL to understand how the tetrahydrofuran ring impacts physical performance metrics. Thermogravimetric analysis (TGA) was used to determine the thermal stability of the different materials. PCL has an onset thermal decomposition temperature (Td,5%) greater than 300 °C, whereas PODO has a substantially lower Td,5% value of 262 °C (Figure 5a). The lower decomposition temperature is hypothesized to result from either depolymerization or decomposition of PODO due to thermodynamically easier ring-closing depolymerization or thermal instability imbued by the tetrahydrofuran ring, respectively. The melting temperature (Tm) and Tg were determined by differential scanning calorimetry (DSC). PODO exhibits a Tg of 30 °C, approximately 90 °C higher than that of PCL (Tg = -60 °C) (Figure 5b); this agrees well with previous work that found a tetrahydrofurancontaining analog of nylon 6 had a Tg of 70 °C greater than that of native nylon 6. 45,46 While PCL has a defined Tm of 54 °C, no such melt transition is observed for PODO (Figure 5b). As PODO is obtained from racemic ODO monomer, the lack of a melting transition is consistent with an atactic, amorphous material, 47,48 which was confirmed by powder X-ray diffraction (Figure S20). To examine mechanical properties, PODO and PCL were solution cast into films and subsequently cut into dogbones for tensile testing. PODO films were transparent and ductile (Figure S21-Figure S22). In comparing PODO and PCL via a tensile test at room temperature, we observe that PODO has a modulus and yield strength 1-2 orders of magnitude lower than semicrystalline PCL and behaves as a soft elastomer (Figure S21-Figure S22). The high Tg and ductility of PODO make it an attractive copolymer to improve the toughness of brittle thermoplastics. PLA is a promising bio-derived thermoplastic that is valued for its high Young's modulus (E = 4.2 GPa) and tensile strength (sy = 43.6 MPa) but suffers from a low elongation at break (εb = 7.4%) (Table 2, Figure 6a). 46 Strategies to improve the toughness of PLA would ideally do so without sacrificing other key properties such as high Tg, strength, or degradability, with significant effort being devoted to copolymerization, plasticization, and blending. 14,[49][50][51][52] Statistical copolymerization of L-lactide (LA) with other lactones can increase εb, but often with a significant decrease in Young's modulus (E < 1 GPa) or glass transition temperature (Tg < 40 °C) (Figure 6b, Table S15). [53][54][55][56][57][58][59][60][61] Amorphous PODO has a significantly higher Tg than many other lactones due to the rigidity of its tetrahydrofuranic backbone, yet it still exhibits elastomer-like properties. We hypothesized that the copolymerization of ODO with LA could improve ductility without drastically decreasing Tg. Preliminary copolymerizations with ODO and LA demonstrate that comonomer incorporation corresponds well to comonomer loading (Figure S23). NMR spectroscopy and X-ray diffraction suggest that copolymerization occurs statistically and crystalline regions are composed exclusively of PLA segments respectively (Figure S23-Figure S26).
Table 2. Thermal and mechanical properties of PLA-stat-PODO copolymers. [53][54][55][56][57][58][59][60][61] Melt copolymerization was used to produce high molecular weight copolymers (Mn > 20 kDa). Copolymerization of 0-19% ODO with LA resulted in semicrystalline copolymers with melting temperatures between 124-153 °C and glass transition temperatures between 54-60 °C (Table 2, Table S13). An inverse relationship between the incorporation of ODO and copolymer molecular weight was also observed. Integration of the melting peaks shows that greater incorporation of ODO decreases the enthalpy of fusion (ΔHf) of the resulting copolymer, suggesting a decrease in crystallinity (Figure S27). While Tm and ΔHf decay significantly, Tg decreases more modestly, following the Flory-Fox relationship (Figure S29).
PLA and PLA-stat-PODO copolymers were melt-pressed into films for mechanical and thermal analysis (Table 2, Figure 6a). Copolymer films were optically clear and did not exhibit discoloration (Figure S30). Each film was then cut into at least three dogbones for replicate measurements (Figure S31). PLA-stat-PODO copolymers had an average elongation at break (εb) greater than 12× that of native PLA (Figure 6a), with 10:1 PLA-stat-PODO able to elongate up to 144% (Figure S31). In addition, only a modest change in Young's modulus and yield strength was observed for any of the copolymers. Retention of high modulus and Tg while enhancing εb is exceedingly rare for statistical copolymers of PLA and has only been achieved in a few instances of PLA block copolymers containing polyethylene glycol or polypropylene glycol as the soft block (Figure 6b). 60 We hypothesize that the 5-9% incorporation balances polymerizability to obtain a high-molecular weight polymer with enough ductility from the ODO monomer to substantially improve the mechanical performance. The ability of ODO to increase the εb without significantly decreasing Tg, modulus, or strength makes it a highly attractive comonomer for producing more ductile PLA-based materials. In comparing PLA-stat-PODO to the mechanical performance of industrial polymers, we observe that PLA-stat-PODO exhibits competitive mechanical properties, comparable in elongation to polypropylene (PP) and polycarbonate (PC). 62 Degradable polymers tend to either have high modulus or high ductility, with PLA-stat-PODO blending ductility and strength to yield a unique class of degradable, renewably-sourced materials (Figure 7).
Bicyclic lactones can be polymerized to form rigid polyesters with valuable thermal and mechanical properties. Herein, we highlight the synthesis and polymerization of ODO, a bicyclic lactone derived from HMFA that introduces a tetrahydrofuran ring in the polymer backbone. We demonstrate how ODO can be synthesized at a multi-gram scale using H2 as the only stoichiometric reactant without the use of stoichiometric reagents and how it can be controllably polymerized via organocatalytic and melt methods. Subsequent comparison between polymerization of ODO and CL reveal that the tetrahydrofuran moiety reduces ring strain released during polymerization but increases the rate of polymerization. The introduction of the rigid tetrahydrofuran ring also increases the Tg of PODO by approximately 90 °C relative to PCL but reduces the thermal stability. Furthermore, the tetrahydrofuran ring introduces two stereocenters that make PODO synthesized from racemic monomer atactic. While PCL behaves like a tough thermoplastic, atactic PODO behaves like a soft elastomer (εb > 500%). Copolymerization of racemic ODO with LA, however, results in thermoplastics with improved ductility and comparable E, sy, and Tg to PLA, demonstrating how tetrahydrofuran-containing lactones could be valuable as toughening comonomers for brittle thermoplastics.
• Experimental procedures, instrumentation details, additional structural characterization
( 1 H NMR, 13 C NMR, COSY NMR, HMBC NMR, and MALDI), thermal analysis (DSC), mechanical data, reaction optimization data, and miscellaneous data (PDF)
1-8 2. Supporting figures 3. Supporting tables 9-30 31-38
5-hydroxymethyl furoic acid was purchased from Arctom Chemical and was used to synthesize Oxo-3,8-dioxabicyclo[3.2.1]octane (ODO) in two steps. 1-cyclohexyl-3-phenylthiourea (TU) and Triazabicyclodecene (TBD) were purchased from Arctom Chemical and were stored under N2 in a glovebox prior to usage. 1,3-bis(3,5-bis(trifluoromethyl)phenyl)urea (SU) was synthesized using a previously reported literature method and was stored under N2 in a glovebox prior to use. Tin octoate (Sn(Oct)2) was purchased from AK Scientific. 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) and tin oxide (SnO) were purchased from AmBeed Chemical and stored under N2 in a glovebox prior to use. Glycerol ethoxylate (GEO) was purchased from SigmaAldrich. Rhodium on carbon (Rh/C) was purchased from Oakwood Chemical. Rhodium on Alumina (Rh/Al2O3) was purchased from VWR International. Epsilon-caprolactone (CL) was purchased from Fisher Scientific and stored in an N2-filled glovebox prior to use. L-lactide (LA) was purchased from Purac, recrystallized from methylene chloride (DCM) and stored under N2 in a glovebox prior to use. Unless otherwise indicated, chemicals were used as received.
NMR spectra were obtained at 23 °C on either a 600, 500, or 400 MHz Varian Unity Inova spectrometer or a 400 MHz Bruker spectrometer. Deuterated solvents were used as received from Cambridge Isotopes. Differential scanning calorimetry (DSC) was performed on a TA Instruments 2500 Differential Scanning Calorimeter at the Stanford Soft Materials Facility. Samples of ~5 mg were sealed in standard aluminum "Tzero" pans with lids, purchased from TA Instruments. Heat flow to the sample in mW was recorded relative to a reference pan and lid prepared without sample. Measurements were performed under 50 mL/min dry N2 flow with a temperature ramp rate of 10 °C/min. Glass transition temperatures (Tg) were determined by taking the midpoint of the transition curve on the second heating cycle. Thermogravimetric analysis (TGA) was performed on a TA Instruments 5500 thermogravimetric analyzer at the Stanford Soft Materials Facility. Samples of ~10 mg were loaded onto 100 μL platinum crucibles and heated to desired temperature under 50 mL/min N2 flow at a ramp rate of 10 °C/min. Powder X-ray diffraction was performed under ambient conditions on a Bruker D8 Advance diffractometer equipped with a Cu anode (Kα1 = 1.54060 Å, Kα2 = 1.54443 Å, Kα2/Kα1 = 0.5). Matrix-Assisted Laser Desorption Ionization Time of Flight Spectrometry (MALDI) was performed at the Stanford Peptide and Nucleic Acid Facility (Stanford, CA). Polymer samples were dissolved in THF (~5 mg/mL). The MALDI matrix was prepared using sinapinic acid (~10 mg/mL) in 1:1 acetonitrile/water with 1 wt% TFA. Each MALDI sample was prepared by combining 2 μL of the MALDI matrix with 1.5 μL of the polymer solution and depositing them onto a MALDI plate. The solution was dried before analysis with an Applied Biosystems Voyager DE RP MALDI-TOF Spectrometer (Laser energy ~2.5 μJ). Gel permeation chromatography (GPC) was performed using two Agilent PolyPore columns in a THF mobile phase (~1 mg/mL) connected to a Wyatt Instruments DAWN multiangle light scattering (MALLS) detector and a Wyatt Instruments Optilab T-rEX differential refractometer. Prior to injection, samples were dissolved in THF and filtered through a 0.2 μm pore size PTFE syringe filter. Liquid chromatography-mass spectrometry (LC-MS) was performed on a Waters SQD2 LC/MS system (RRID:SCR_022217). LC-MS samples were dissolved in acetonitrile (~0.1 mg/mL). Tensile properties were investigated using a Linkam Mechanical Tester with an extension rate of 10 μm/s and gauge length of 15 mm using a 200 N load cell.
General Procedure: 5-hydroxymethyl furoic acid (HMFA), rhodium on alumina (5 wt%) (Rh/Al2O3), and water were combined to make a solution of 0.1 M in HMFA. The solution was transferred to a 250 mL Parr Instruments stirred reactor model 4576B. While stirring the solution, the reactor was filled and purged with N2 five times, followed by five times with H2. Reactor vessel was pressurized with H2 to 10 bar. Reactor vessel was left stirring for 4 hours, until complete conversion as determined by NMR. Upon completion, pressure from the reactor vessel was released and the vessel was flushed with N2. Solution was subsequently transferred to a centrifuge tube, centrifuged, and decanted. The solid catalyst was recovered and utilized in subsequent hydrogenation reactions. The solution was dried via rotary evaporation and under high vacuum. The resulting oil 5-hydroxymethyl-2-tetrahydrofuroic acid (HMTA) was used without further purification. Specific Example: Optimized conditions were applied on the 10 g scale. HMTA (10.0 g, 70 mmol), Rh/Al2O3 (1.0 g, 0.5 mmol, 0.7 mol% Rh), and water (50 mL) were combined in a 250 mL Parr reactor. While stirring the solution, the reactor was filled and purged with N2 five times, followed by five times with H2. Reactor vessel was pressurized with H2 to 10 bar. Reactor vessel was left stirring for 3 days, until complete conversion was determined by NMR. Upon completion, pressure from the reactor vessel was released and the vessel was flushed with N2. The reaction mixture was subsequently transferred to a centrifuge tube, centrifuged, and decanted to recover catalyst. The solution was dried via rotary evaporation and under high vacuum. The resulting oil 5hydroxymethyl-2-tetrahydrofuroic acid (HMTA) was used without further purification.
General Procedure: Resulting crude HMTA from hydrogenation, p-toluenesulfonic acid (p-TSA), and toluene were added to an oven-dried round bottom flask with a stir bar and an attached Dean-Stark apparatus with reflux condenser. The reaction mixture was stirred under reflux for 3 hours while conversion was monitored via NMR of aliquots taken at regular time intervals. Upon reaching equilibrium, the crude mixture was filtered using a silica plug to remove oligomeric species. Oligomeric species were recovered and separated by subsequently washing the silica plug with methanol. The filtered reaction mixture was concentrated via rotary evaporation, dissolved in diethyl ether, and partitioned in a separatory funnel between diethyl ether and saturated sodium bicarbonate (NaHCO3). The organic layer was saved over the course of three washes. The organic layer was subsequently dried with magnesium sulfate, filtered, concentrated via rotary evaporation, dissolved in an equal volume mixture of toluene and hexanes, and recrystallized in a freezer overnight to yield the white, low melting point solid oxo-3,8-dioxabicyclo[3. Specific Example: The resulting crude 5-hydroxymethyl-2-tetrahydrofuroic acid, p-TSA (255 mg, 1.5 mmol, 2 mol%), and toluene (200 mL) were added to an oven-dried 500 mL round bottom flask with a stir bar and an attached Dean-Stark apparatus. After refluxing overnight, the crude mixture was filtered using a silica plug to remove oligomeric species. Oligomeric species, o-PODO (7.5 g), were recovered and separated by subsequently washing the silica plug with methanol. The filtered reaction mixture was concentrated via rotary evaporation, dissolved in diethyl ether, and partitioned in a separatory funnel between diethyl ether and saturated sodium bicarbonate (NaHCO3). The organic layer was saved over the course of three washes. The organic layer was subsequently dried with magnesium sulfate, filtered, concentrated via rotary evaporation, dissolved in an equal volume mixture of toluene and hexanes, and recrystallized in a freezer overnight to yield the product lactone ODO (1.6 g, 20% yield relative to HMTA, 18% yield relative to HMFA).
General Procedure: o-PODO, tin octoate (Sn(Oct)2), and optionally glycerol ethoxylate (GEO) were added to a round bottom flask. The reaction mixture was then dispersed in methanol (MeOH), concentrated via rotary evaporation, and dried under high vacuum. The flask was then attached to a short-path distillation apparatus, evacuated under high vacuum, and heated to the desired temperature for the desired amount of time. The crude product was dissolved in toluene and filtered using a silica plug. The filtered reaction mixture was concentrated via rotary evaporation, dissolved in diethyl ether, and partitioned in a separatory funnel between diethyl ether and saturated sodium bicarbonate (NaHCO3). The organic layer was saved over the course of three washes. The organic layer was subsequently dried with magnesium sulfate, filtered, concentrated via rotary evaporation, dissolved in an equal volume mixture of toluene and hexanes, and recrystallized in a freezer overnight to yield the product lactone ODO. Specific Example: The resulting o-PODO (7.5 g), GEO (0.1 g, 1 wt%), and Sn(Oct)2 (0.31 g, 1 mol%) were added to a 25 mL round bottom flask. The reaction mixture was then dispersed in methanol (MeOH), concentrated via rotary evaporation, and dried under high vacuum. The flask was then attached to a short-path distillation apparatus, evacuated under high vacuum, and heated to 200 °C for 2 h. The crude product was dissolved in toluene and filtered using a silica plug. The filtered reaction mixture was concentrated via rotary evaporation, dissolved in diethyl ether, and partitioned in a separatory funnel between diethyl ether and saturated sodium bicarbonate (NaHCO3). The organic layer was saved over the course of three washes. The organic layer was subsequently dried with magnesium sulfate, filtered, concentrated via rotary evaporation, dissolved in an equal volume mixture of toluene and hexanes, and recrystallized in a freezer overnight to yield the product lactone ODO (3.4 g, 45% yield relative to o-PODO, 38% yield relative to HMFA).
Monomer lactone, 1,3-bis(3,5-bis(trifluoromethyl)phenyl)urea (Schreiner's urea catalyst, SU) (1 mol%), 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) (1 mol%), benzyl alcohol (BnOH) (1 mol%), and tetrahydrofuran (THF) (1 M final concentration relative to lactone) were added under an N2 atmosphere to an oven-dried vial with a stir bar. Reaction was stirred for desired reaction time before a superstoichiometric amount of benzoic acid was added to quench the reaction. Reaction mixture was homogenized with the addition of methylene chloride with an aliquot taken for NMR analysis and subsequently precipitated in methanol. Precipitated polymer was washed with methanol three times before drying under high vacuum overnight.
In a glovebox, ODO (18 mg, 140 µmol) was added to an oven-dried vial with a stir bar. Stock solutions of Schreiner's urea (0.1 M, 14 µL, 1 mol%), DBU (0.1 M, 14 µL, 1 mol%), and benzyl alcohol (0.05 M, 28 µL, 1 mol%) along with THF (84 µL), ensuring the final concentration with respect to monomer was 1 M. The vessel was sealed and stirred at room temperature. At specified intervals (5, 15, 60, 120, 180, 300, 420, 1440 min), a small aliquot (10 µL) was extracted from the reaction solution and added to a separate vial containing THF and excess benzoic acid to quench the reaction. The quenched samples were subsequently dissolved in CDCl3 and analyzed by 1 H NMR to determine monomer conversion.
In a glovebox, ODO (18 mg, 140 µmol) was added to an oven-dried vial with a stir bar. Stock solutions of Schreiner's urea (0.1 M, 14 µL, 1 mol%), DBU (0.1 M, 14 µL, 1 mol%), and benzyl alcohol (0.05 M, 28 µL, 1 mol%) along with THF (84 µL), ensuring the final concentration with respect to monomer was 1 M. The vessel was sealed, heated, and stirred at desired temperature until equilibrium was established (1-24 hours). At the end of the reaction, excess benzoic acid (20 mg) was added to quench the reaction. The crude reaction was dissolved in CDCl3 and analyzed by 1 H NMR to determine monomer conversion.
Monomer lactone, tin octoate (Sn(Oct)2) in THF, and benzyl alcohol (BnOH) in THF were added to an oven-dried vial under a N2 atmosphere. The reaction mixture was heated until all components melted at which point the vial was vortexed to homogenize reaction mixture. The vial was heated at the desired temperature for the desired reaction time before quenching with cold methylene chloride or deuterated chloroform depending on the scale of the reaction. Reaction mixture was homogenized with the addition of methylene chloride with an aliquot taken for NMR analysis and subsequently precipitated in methanol. Precipitated polymer was washed with methanol three times before drying under high vacuum overnight.
Melt polymerization of ODO was conducted following the general procedure by adding oxo-3,8dioxabicyclo[3.2.1]octane (570 mg, 4.5 mmol) and a 0.05 M solution of Sn(Oct)2 in THF (15 µL, 0.00075 mmol, 0.02 mol%) to a 2 mL analytical vial under a N2 atmosphere. The vial was sealed and heated at 100 °C until the reaction mixture was completely melted at which time the vial was vortexed until well-mixed. The vial was left at 90 °C for 24 h before being cooled to ambient temperature. The resulting crude polymer was then dissolved in 20 mL methylene chloride in a 40-mL scintillation vial with an aliquot taken for NMR analysis and subsequently precipitated in methanol. Precipitated polymer was washed with methanol three times before drying under high vacuum overnight to yield a soft white polymer identified by NMR to be poly(oxo-3,8dioxabicyclo[3.2.1]octane) (PODO) (440 mg, yield = 78%).
Melt polymerization of ODO was conducted modifying the general procedure by adding oxo-3,8dioxabicyclo[3.2.1]octane (310 mg, 2.4 mmol), Sn(Oct)2 (5.5 mg, 0.014 mmol, 0.5 mol%), and BnOH (2.5 mg, 0.024 mmol, 1 mol%) to a 2 mL analytical vial under a N2 atmosphere. The vial was sealed and heated at 50 °C until the reaction mixture was completely melted at which time the vial was vortexed until well-mixed. The reaction mixture was then divided into twenty 2 mL analytical vials, sealed, and grouped into sets of four. Each set was heated to the desired temperature (60-100 °C) for 1 week before being quenched and homogenized with the addition of cold deuterated chloroform. Conversion was analyzed via NMR analysis.
Melt polymerization of CL was conducted following the general procedure by adding εcaprolactone (10 g, 88 mmol) and Sn(Oct)2 (35 mg, 0.1 mol%) to a 20 mL scintillation vial with a stir bar under a N2 atmosphere. The vial was sealed and heated at 140 °C and left to stir for 3 h before being cooled to ambient temperature. The resulting crude polymer was then dissolved in Sn(Oct) 2 (0.1 mol%) 140 °C, 3 h 100 mL methylene chloride in a 250-mL round bottom flask with an aliquot taken for NMR analysis and subsequently precipitated in methanol. Precipitated polymer was washed with methanol three times before drying under high vacuum overnight to yield a soft white polymer identified by NMR to be poly(ε-caprolactone) (PCL).
Melt polymerization of LA/ODO was conducted following the general procedure by adding Llactide (790 mg, 5.5 mmol), oxo-3,8-dioxabicyclo[3.2.1]octane (70 mg, 0.5 mmol, 10 mol%), and a 0.05 M solution of Sn(Oct)2 in THF (25 µL, 0.0013 mmol, 0.02 mol%) to a 2 mL analytical vial under a N2 atmosphere. The vial was sealed and heated at 100 °C until the reaction mixture was completely melted at which time the vial was vortexed until well-mixed. The vial was left at 100 °C for 24 h before being cooled to ambient temperature. The resulting crude polymer was then dissolved in 20 mL methylene chloride in a 40-mL scintillation vial with an aliquot taken for NMR analysis and subsequently precipitated in methanol. Precipitated polymer was washed with methanol three times before drying under high vacuum overnight to yield a white polymer identified by NMR to be poly(L-lactide-stat-oxo-3,8-dioxabicyclo[3.2.1]octane) (PLA-stat-PODO, 10:1 PLA/PODO) (770 mg, yield = 90%).
General Procedure: PODO, tin octoate (Sn(Oct)2), and glycerol ethoxylate (GEO) were added to a round bottom flask. The reaction mixture was then dissolved in methylene chloride, concentrated via rotary evaporation, and dried under high vacuum. The flask was then attached to a kugelrohr distillation apparatus and evacuated under high vacuum. The collection flask was cooled with dry ice. Rotation was applied and the reaction flask was then heated to the desired temperature for the desired amount of time. ODO crystals were obtained from the collection flask and analyzed by NMR to determine purity. Specific Example: Optimized conditions were applied on the 200 mg scale. PODO (190 mg), tin octoate (3 mg, 2 wt%), and glycerol ethoxylate (4 mg, 2 wt%) were added to a 25 mL round bottom flask. The reaction mixture was then dissolved in methylene chloride, concentrated via rotary evaporation, and dried under high vacuum. The flask was then attached to a kugelrohr distillation apparatus and evacuated under high vacuum. Rotation was applied and the reaction flask was then heated to 200 °C for 1 h. The crystals in the receiving flask were collected and identified by NMR to be ODO (125 mg, 66% yield).
To prepare solvent-casted films, purified polymer was dissolved and sonicated in a minimal amount of methylene chloride (DCM) (50-100 mg polymer/mL DCM) to obtain a slightly viscous liquid. Solution was cast onto PTFE. The cast was then covered and allowed to dry slowly over 24 hours before drying in an over at 50 °C for 24 hours. Films were then dried in a vacuum oven overnight at 80 °C to remove residual solvent. The edges of the films were then cut away and the film was peeled off.
To prepare melt-pressed films, purified polymer was dried in a vacuum oven overnight at 80 °C to remove residual solvent and ensure polymer was free of moisture. The polymer was then placed between PTFE-lined brass plates and loaded onto a Carver press preheated to 180 °C and allowed to anneal for 1-2 minutes above its melting point. After melting, a pressure of 1000 psi was then applied for 0.5-1 minutes before being released. The sample was then removed from the press and allowed to cool at ambient temperature.
Polymer films were cut using a "Print-A-Punch" 3D-printed dogbone die and aged overnight at ambient temperature. 63 2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 f1 (ppm) 0.23 1.04 2.21 1.16 2.00 1.02 0.94 1.19 1.65 1.98 2.25 3.58 4.11 4.46 O O O Key: a-e = 1 H 1-6 = 13 C a|2 1 b|3 c|4 d|5 e|6 CDCl 3 CDCl 3 2 6 5 3 4 1 c a e Integrations for PLA-stat-PODO with 9% PODO (10:1 PLA-stat-PODO) reported above. PODO% incorporation = d, e 3 f 2 + 𝑑, 𝑒 3 Figure S29. Fitting the copolymer Tg to the Flory-Fox equation ) * !,#$ = + % * !,% + + & * !,& where 𝑤 , and 𝑇 -,, are the weight percent and homopolymeric glass transition temperatures of monomer 𝑖, respectively.
Table S1. Optimization of HMFA hydrogenation Entry Time (h) Catalyst x (mol%) Solvent Yield (%) a 1 16 Rh/C 3 EtOH 90% 2 16 Rh/C 3 iPrOH 92% 3 16 Rh/C 3 H2O 95% 4 16 Rh/C 6 H2O 97% 5 4 Rh/Al2O3 6 H2O 98% 6 4 Rh/Al2O3 3 H2O 98% 7 b 4 Rh/Al2O3 3 H2O 91% Conditions: HMFA (1g, 7 mmol), solvent (0.1 M), catalyst, room temperature, 4-24h. a Isolated yield of HMTA. b 10g scale. Table S2. Heterogeneous catalyst screen for cyclization of HMTA cyclization without H2O removal. Entry Catalyst Solv. Temp. (°C) Conv. a (%) 1 b Amberlyst r15(H) Toluene 100 6 2 NH4-ZSM-5 (23:1 Si/Al) Toluene 100 3 3 NH4-Zeolite β (25:1 Si/Al) Toluene 100 3 4 H-Zeolite β (360:1 Si/Al) Toluene 100 8 5 H-Zeolite Y (30:1 Si/Al) Toluene 100 27 6 H-Zeolite Y (30:1 Si/Al) o-Xylene 140 27 7 H-ZSM-5 (23:1 Si/Al) c Toluene 100 19 8 H-Zeolite β (25:1 Si/Al) d Toluene 100 22 9 None Toluene 100 14 Conditions: HMTA (25 mg), catalyst (30 mg), solvent (0.1 M), 3h. a Conversion based on NMR quantification using N,N-dimethylacetamide as an internal quantitative standard. b adapted on conditions from ref 28. c Prepared by calcination of NH4-ZSM-5 (23:1 Si/Al) at 515 °C following the procedure from ref 64. d Prepared by calcination of NH4-Zeolite β (23:1 Si/Al) at 515 °C following the procedure from ref 64. Table S3. Homogeneous catalyst screen for cyclization of HMTA cyclization without H2O removal. Entry Catalyst Solv. Temp. (°C) Conv. a (%) 1 -Toluene 100 14 2 p-TSA Toluene 100 31 3 p-TSA o-Xylene 140 7 3 HCOOH Toluene 100 7 4 TFA Toluene 100 19 3 H2SO4 Toluene 100 3 3 AcOH Toluene 100 34 4 -AcOH 100 38 5 -TFA 100 4 6 -Ac2O 20 48 Conditions: HMTA (25 mg), catalyst (10-20 mol%), solvent (0.1 M), 110 °C, 3h. a Conversion based on NMR quantification using N,N-dimethylacetamide as an internal quantitative standard Table S4. Optimization of HMTA cyclization with H2O removal via Dean-Stark Apparatus Entry Catalyst x (mol%) Solv. Time (h) ODO (%) a Other (%) b 1 --Toluene 3 17 (20) 73 2 p-TSA 40 Toluene 3 32 (33) 74 3 p-TSA 40 Toluene 6 28 (18) 86 4 p-TSA 3 Toluene 3 47 (48) 65 5 H-Zeolite Y (30:1 Si/Al) -Toluene 3 39 (44) 30 6 H-Zeolite β (25:1 Si/Al) d -Toluene 3 46 (25) 90 7 p-TSA 3 o-Xylene 3 45 (52) 35 8 e p-TSA 1 Toluene 12 45 (49) 23 2 t-BuOK 78 -6.4 1.24 3 DBU 12 ---4 DMAP 0 ---5 TBD 92 7.9 6.6 1.18 6 TU 5 ---7 SU 2 ---8 TU/DBU c 67 11.8 5.1 1.16 9 SU/DBU c 94 10.1 9.6 1.19 10 SU/t-BuOK c 73 11.3 3.5 1.91 Table S7. Optimization of organocatalytic ring opening polymerization of ODO. Entry x (mol%) Solvent Conc. (M) Time (min) Conv. (%) a Mn,NMR (kDa) b Mn,GPC (kDa) Ð 1 100 THF 1 30 94 10.1 9.6 1.19 2 10 THF 1 30 95 10.5 11.3 1.19 3 5 THF 1 30 92 11.1 10 1.14 4 1 THF 1 30 42 12.3 9.6 1.03 5 1 CH2Cl2 1 30 2 ---6 1 Toluene 1 30 18 ---7 1 Benzene 1 30 1 ---8 1 THF 2 30 95 13.1 11.8 1.23 3 Sn(Oct)2 -100 100 1 75 -8.9 1.04 4 Sn(Oct)2 BnOH 100 100 1 89 8.7 8.3 1.22 5 MgCl2 BnOH 100 100 1 38 3.5 4.7 1.07 6 Al(i-PrO)3 BnOH 100 100 1 82 7.6 8 1.40 7 Sn(Oct)2 BnOH 100 80 1 89 7.4 10.4 1.11 8 Sn(Oct)2 BnOH 100 120 1 78 5.8 13.3 1.33 9 Sn(Oct)2 BnOH 100 80 2 92 8.8 12.7 1.14 10 Sn(Oct)2 BnOH 100 80 4 94 8.1 8.9 1.41 11 Sn(Oct)2 BnOH 400 80 2 51 24.3 21.1 1.07 12 Sn(Oct)2 BnOH 600 80 2 29 24.0 16.5 1.06 13 Sn(Oct)2 BnOH 600 80 24 72 62.5 48.9 1.37 Table S11. Melt ring opening polymerization of ODO with different catalyst loading. Entry M/Cat Conv. (%) Mn,GPC (kDa) Ð 1 500 67 116 1.10 2 4000 35 187 1.06 3 8000 26 208 1.10 Table S12. Melt ring opening polymerization of ODO at different temperatures in quadruplicate. Entry Temp. (°C) Conv. (%) Sample 1 2 3 4 1 60 98.1 98.0 98.1 98.1 2 70 97.5 97.6 97.6 97.9 3 80 96.5 96.8 96.7 96.9 4 90 96.4 96.1 96.4 96.5 5 100 95.5 95.5 94.8 94.8
Experimental methods