Synthesis of Network Biobased Aliphatic Polyesters Exhibiting Better Tensile Properties than the Linear Polymers by ADMET Polymerization in the Presence of Glycerol Tris(undec-10-enoate)

Development of biobased aliphatic polyesters with better mechanical (tensile) properties in film has attracted considerable attention. This report presents the synthesis of soluble network biobased aliphatic polyesters by acyclic diene metathesis (ADMET) polymerization of bis(undec-10-enyl)isosorbide diester [M1, dianhydro-D-glucityl bis(undec-10-enoate)] in the presence of a tri-arm crosslinker [CL, glycerol tris(undec-10-enoate)] using a ruthenium–carbene catalyst, and subsequent olefin hydrogenation using RhCl(PPh3)3. The resultant polymers, after hydrogenation (expressed as HCP1) and prepared in the presence of 1.0 mol% CL, showed better tensile properties than the linear polymer (HP1) with similar molecular weight [tensile strength (elongation at break): 20.8 MPa (282%) in HP1 vs. 35.4 MPa (572%) in HCP1]. It turned out that the polymer films prepared by the addition of CL during the polymerization (expressed as a 2-step approach) showed better tensile properties. The resultant polymer film also shows better tensile properties than the conventional polyolefins such as linear high density polyethylene, polypropylene, and low density polyethylene.

In this paper, we focus on the synthesis of the soluble network polymers by ADMET polymerization.These polymers were prepared by conducting the polymerization in the presence of a crosslinker (CL) possessing three terminal olefins [31,41], as reported previously by us [31], and the resultant polymer should exhibit better tensile strength due to their network framework.We thus herein report the synthesis of network biobased aliphatic polyesters, which show better tensile properties (tensile strengths and elongation at breaks) than the linear one (HP1).These polymers were also depolymerized to afford the corresponding diesters and diols (and triols) through transesterification with ethanol in the presence of CpTiCl 3 [12,14].
All 1 H and 13 C NMR measurements were performed at 25 • C on a Bruker AV500 spectrometer (500.13MHz and 125.77MHz, respectively) using CDCl 3 as solvent.Chemical shifts were reported as ppm with reference to SiMe 4 at 0.00 ppm.Gel permeation chromatography (GPC) was used for the analysis of molecular weights and molecular weight distributions for the resultant polymer.The GPC measurements were carried out at 40 • C on a SCL-10A (Shimadzu Co., Ltd., Kyoto, Japan) connected columns (ShimPAC GPC-806, 804 and 802, 30 cm × 8.0 mm diameter, spherical porous gel made of styrene/divinylbenzene copolymer, ranging from <10 2 to 2 × 10 7 MW, Kyoto, Japan), using a Shimadzu RID-10A detector in THF (>99.8%,Kanto Chemical Co., Inc., Tokyo, Japan) served as the eluent with a flow rate 1.0 mL/min.

Synthesis of crosslinker glycerol triundec-10-enoate (CL).
In the dry box, glycerol 300 mg, 3.25 mmol) and triethylamine (2.0 g, 19.8 mmol, 2.0 eq.) were added to THF (30 mL) and then cooled to 0 • C. 10-undecenoyl chloride (2.0 g, 9.9 mmol) was then added dropwise to the cooled glycerol solution.The reaction was allowed to react at room temperature overnight until completion by confirmation via TLC.The THF solvent was then removed with a rotary evaporator and then diluted with chloroform and washed with 2M HCl, 5% NaHCO 3 , deionized water, and brine.The washed product was then dried over MgSO 4 , filtered through a filter paper, evaporated using a rotary evaporator, and further dried under vacuum.The crude product was then purified using column chromatography (9:1, n-hexane: ethyl acetate), collecting a colorless oil of CL (1.54 g, 80% yield).The resultant product was further purified inside the dry box by dissolving the crosslinker in hexane and passing through a short column of alumina and celite. 1 H NMR (CDCl 3 ): δ 1.28 (br s, 24H, -CH 2 -), 1.37 (t, J = 6.9 Hz, 6H, -CH ADMET Polymerization.ADMET polymerizations were conducted by the analogous procedure reported previously [14,32].In the dry box, the monomer (M1), crosslinker (CL), solvent, and ruthenium catalyst (HG2) were charged with the prescribed amounts in Table 1 into a sealed Schlenk-type tube (25 mL volume), and the reaction mixture was stirred for a specific time at 50 • C in an oil bath.The formed ethylene gas was removed continuously by freezing the reaction medium using liquid nitrogen, and the tube was shortly connected to the vacuum line with a certain time interval (each 15 min at the first 1 h, each 30 min in the following 2 h, and then each 1 h).The solvent exchange technique was conducted by solvent replacement [35,42] with a fresh solvent two or three times in the first two hours under N 2 atmosphere.Two drops of EVE were used to quench the ADMET polymerization and stirring for 1 h.The reaction mixture was diluted using 4 mL chloroform and precipitated in 100 mL cold methanol.The resultant polymers were collected via filtration and dried in vacuo and were characterized using 1 H (500.13 MHz) and 13 C{ 1 H} (125.77MHz) NMR spectra in CDCl 3 at 25 • C, GPC (SEC), and DSC.
CP1. 1   Olefin Hydrogenation.The resultant polymers (200 mg), toluene (3.0 mL), and RhCl(PPh 3 ) 3 (3 mg) were added into the autoclave (20 mL scale) [36].The stainless steel autoclave was then pressurized with hydrogen (1.0 MPa) and was placed into a heating Al block preheated at 50 • C. The solution was magnetically stirred for 24 h.The reaction mixture was then poured into a mixed solution of MeOH.The polymer precipitates were collected on a filter paper, and the collected white precipitates were then dried in vacuo for several hours.No significant differences in the M n , Ð (M w /M n ) were observed in the polymer samples before/after the hydrogenation.The resultant polymers were identified by NMR spectra (shown in Supplementary Materials), and the molecular weights and the distributions were analyzed by GPC.

Synthesis of Network Polymers (CP1 and HCP1) by ADMET Polymerization and Subsequent Hydrogenation
ADMET polymerization of bis(undec-10-enyl)isosorbide diester [M1, bis(undec-10enoate) with isosorbide] in the presence of glycerol tris(undec-10-enoate) (CL) was chosen because the resultant polymer film prepared by M1 (expressed as HP1, Scheme 1) exhibited good tensile strength and elongation at break [36].As shown in Scheme 2, two approaches involving adding CL from the beginning (1-step approach) or after 1 or 3 h (2-step approach) have been chosen for the synthesis since the approach might affect the network density or average polymer chain length between CLs (crosslinking point).The polymerizations of M1 by the ruthenium-carbene catalyst (HG2, 2.0 mol%) were conducted in solvent (toluene, chloroform, or tetrachloroethane, initial M1 conc.0.94 mmol/mL) in the presence of CL (0.5-5.0 mol%), which were prepared by glycerol with 3.0 equiv of 1-undecenoyl chloride (see Materials and Methods).The selected results conducted under the various conditions are summarized in Table 1.For the obtainment of high molecular weight polymers under these conditions (300 mg M1 scale), the solvent in the reaction mixture was removed in vacuo to replace the fresh one (called solvent replacement) every 30 min in certain experimental runs [42].The method is effective for the purpose (removal of ethylene remained) of condensation polymerization [42], although the method would not be appropriate in terms of a green sustainable process, and alternative methods (conducted in IL [14] or a molybdenum catalyst [36]) should be studied.
It was revealed that the ADMET polymerizations of M1 in the presence of 0.5-2.5 mol% CL (added at the beginning, called a 1-step approach) gave the highest molecular weight polymers (expressed as CP1s), and the M n values became higher upon presence of CL (0.5 or 1.0 mol%, runs 1-5) compared to that in the absence of CL (run 1).The M n value decreased upon further CL addition (2.5 mol%, runs 9 and 10), and the resultant polymers became swelled gels with stop stirring when the polymerizations were conducted in the presence of 5.0 mol% of CL (run 14).The results are highly reproducible, as demonstrated in runs 3-5, although the dispersity (PDI, M w /M n ) values are somewhat large, probably due to the difficulty of stirring the reaction mixture owing to high viscosity (run 14) [14].Indeed, the polymerizations in the presence of 5.0 mol% CL eventually afforded polymer gels irrespective of the kind of solvent employed (runs 13-17).It seems that the polymerization conducted in toluene and tetrachloroethane afforded polymers possessing rather low molecular weights compared to those conducted in chloroform, whereas the solvent replacement improved the M n value in CP1 [M n = 19,300 (run 6) vs. 26,000 (run 7)].
In contrast, the M n values in the resultant polymers (CP1s) were low when CL (1.0 mol%) was added after 1 h of the ADMET polymerization (expressed as a 2-step approach) irrespective of solvents (M n = 17,600-26,800, runs 18-20); the polymerization in chloroform gave CP1 with the highest molecular weight (run 18).Although the M w /M n values in the resultant polymers in the presence of 2.5 mol% CL conducted in chloroform were rather large [M w /M n = 4.13-5.45(runs [23][24][25]], the M w /M n values in CP1 became rather low (unimodal) when CL was added after 3 h of the ADMET polymerization (runs 29-32).As observed above, an increase in the number of solvent replacements led to an increase in the M n values because the ethylene that remained in the polymerization solution was removed in vacuo with the removal of solvent, which led to condensation polymerization [42].Polymer samples with different molecular weights for the analysis of tensile properties (shown below, Figures 1 and 2) were thus prepared by adopting this method.According to the reported procedure [14], olefinic double bonds in the resultan polymers (CP1s) were hydrogenated by RhCl(PPh3)3 catalysts in toluene (Scheme 2, H2 1. MPa, 50 °C, 24 h).The results are summarized in Table 2.It was revealed that, as reporte previously [36], no significant changes in the Mn (as well as Mw/Mn) values were see before/after hydrogenation, and resonances ascribed to the internal olefins disappeare in the polymer samples (HCP1s) after hydrogenation.Their uniform composition (completion of hydrogenation) were also confirmed by DSC thermograms observed a sole melting temperatures [32].As shown in Figure S20 (and Table S4), no significan differences in the melting temperature were observed between linear and networ polymers, whereas the Tm values in HCP1 are higher than CP1, as reported previously b P1 and HP1 [32].The resultant polymers are soluble in chloroform, THF, and toluene regardless of their network structure.According to the reported procedure [14], olefinic double bonds in the resultan polymers (CP1s) were hydrogenated by RhCl(PPh3)3 catalysts in toluene (Scheme 2, H2 1.0 MPa, 50 °C, 24 h).The results are summarized in Table 2.It was revealed that, as reported previously [36], no significant changes in the Mn (as well as Mw/Mn) values were seen before/after hydrogenation, and resonances ascribed to the internal olefins disappeared in the polymer samples (HCP1s) after hydrogenation.Their uniform composition (completion of hydrogenation) were also confirmed by DSC thermograms observed a sole melting temperatures [32].As shown in Figure S20 (and Table S4), no significan differences in the melting temperature were observed between linear and network polymers, whereas the Tm values in HCP1 are higher than CP1, as reported previously by P1 and HP1 [32].The resultant polymers are soluble in chloroform, THF, and toluene regardless of their network structure.According to the reported procedure [14], olefinic double bonds in the resultant polymers (CP1s) were hydrogenated by RhCl(PPh 3 ) 3 catalysts in toluene (Scheme 2, H 2 1.0 MPa, 50 • C, 24 h).The results are summarized in Table 2.It was revealed that, as reported previously [36], no significant changes in the M n (as well as M w /M n ) values were seen before/after hydrogenation, and resonances ascribed to the internal olefins disappeared in the polymer samples (HCP1s) after hydrogenation.Their uniform compositions (completion of hydrogenation) were also confirmed by DSC thermograms observed as sole melting temperatures [32].As shown in Figure S20 (and Table S4), no significant differences in Polymers 2024, 16, 468 8 of 12 the melting temperature were observed between linear and network polymers, whereas the T m values in HCP1 are higher than CP1, as reported previously by P1 and HP1 [32].The resultant polymers are soluble in chloroform, THF, and toluene, regardless of their network structure.1. 3 GPC data in THF vs. polystyrene standards.4 Isolated yields by precipitation as the methanol insoluble fraction.

Tensile Properties in the Polymer Films (CP1s, HCP1s)
Small dumbbell-shaped test specimens in the resultant polymer samples were prepared by cutting the polymer sheet for measurement of their tensile properties.The polymer sheets were prepared using a hot press method according to the reported procedure [36].The stress/strain experiments were conducted using a universal testing machine at 23 • C (speed of 10 mm/min, humidity 50 ± 10%).The selected results are shown in Figures 1 and 2, and the data are summarized in Table 3.  1 and 2. 3 Method in Scheme 2. 4 GPC data in THF vs. polystyrene standards. 5Data cited from reference [16].
It should be noted that, as shown in Figure 1a, the tensile strength (stress) in the resultant network polymer films after hydrogenation (HCP1) became higher than that prepared in the absence of CL (HP1, linear polymer).The elongation at break (strain) was affected by the method prepared, whereas no significant effects toward the stress were observed; the polymer film prepared by the 2-step approach (addition of CL after 1 h polymerization) showed higher strain compared to that prepared by the 1-step approach.This might be probably due to the difference in the average length of each polymer chain between CL units, although we do not have clear evidence at this moment.Increased CL (from 1.0 mol to 2.5 mol%) led to a decrease in the strain, probably due to increased network density.As reported previously [36], both the tensile strength and the elongation break in the linear polymer (HP1) are affected by the M n value [HP1, M n = 30,700 vs. 40,900.1b].In contrast, in the network polymer (HCP1), an increase in the tensile strength (strain) was observed upon an increase in the M n value [HCP1, M n = 31,400 vs. 37,300.Figure 1b].
Figure 2 shows tensile properties in the resultant unsaturated polyester films (CP1, before hydrogenation) prepared by the ADMET polymerization.It was also noted that both the tensile strength (stress) and the elongation at break in the resultant network polymer films (CP1) became higher than that prepared in the absence of CL (P1, linear polymer).As reported previously in HP1 [36], the resultant polymer films (CP1) showed higher stress (elongation at break) than the saturated ones (HCP1), although the tensile strengths (strain) were somewhat low.In contrast to the results in the saturated polymer films (HCP1), it seems that the tensile properties were not affected by the method employed (1-step or 2-step), whereas the strain (elongation at break) in CP1 decreased upon increasing the CL (from 1.0 mol to 2.5 mol%).
Figure 3 summarizes plots of tensile (fracture) strengths and strains (elongation at breaks) of HP1 and CHP1 for comparison with the conventional polymers such as linear high density polyethylene (HDPE), polypropylene (PP) and low density polyethylene (LDPE) [43].It is clear that the resultant polymer film shows better tensile properties (tensile strength, elongation at break) than the conventional polyolefins as well as the other conventional polymers.

HP1
ref  3 summarizes plots of tensile (fracture) strengths and strains (elong breaks) of HP1 and CHP1 for comparison with the conventional polymers such a high density polyethylene (HDPE), polypropylene (PP) and low density polye (LDPE) [43].It is clear that the resultant polymer film shows better tensile pr (tensile strength, elongation at break) than the conventional polyolefins as wel other conventional polymers.Figure 4 shows temperature dependence in the storage module (E′), loss mod and loss factor, tan δ, for HP1 and HCP1s measured by dynamic mechanical a (DMA, 1.0 Hz).The resultant polymers possessed tan d values with relatively widths, suggesting that the resultant polymer possessed uniform composition, results may also suggest that the crosslinking distributions in HCP1 were unifo However, no significant differences in their temperature dependences toward E′, tan δ were observed between linear (HP1) and network polymers (HCP1).It seem that the resultant polymers possessed relatively low crosslinking density, and p units between the crosslinking points are relatively long (reflect the polymer p measured by DMA analysis).Figure 4 shows temperature dependence in the storage module (E ′ ), loss module (E ′′ ), and loss factor, tan δ, for HP1 and HCP1s measured by dynamic mechanical analyses (DMA, 1.0 Hz).The resultant polymers possessed tan d values with relatively narrow widths, suggesting that the resultant polymer possessed uniform composition, and the results may also suggest that the crosslinking distributions in HCP1 were uniform [44].However, no significant differences in their temperature dependences toward E ′ , E ′′ , and tan δ were observed between linear (HP1) and network polymers (HCP1).It seems likely that the resultant polymers possessed relatively low crosslinking density, and polymer units between the crosslinking points are relatively long (reflect the polymer property measured by DMA analysis).

Table 1 .
ADMET polymerization of M1 by HG2 in the presence of the crosslinker (CL)1.