Co-Polymers based on Poly(1,4-butylene 2,5-furandicarboxylate) and Poly(propylene oxide) with Tuneable Thermal Properties: Synthesis and Characterization

Poly(ether ester)s (PEEs) represent a promising class of segmented co-polymers, nevertheless the synthesis of PEEs based on renewable 2,5-furandicarboxylic acid (FDCA) is still scarce. In this context, a series of poly(1,4-butylene 2,5-furandicarboxylate)-co-poly(poly(propylene oxide) 2,5-furandicarboxylate) co-polyesters with different composition of stiff poly(1,4-butylene 2,5-furandicarboxylate) (PBF) and soft poly(poly(propylene oxide) 2,5-furandicarboxylate) (PPOF) moieties were synthesized, via a two-step bulk polytransesterification reaction. The molar ratio of PBF/PPOF incorporated was varied (10 to 50 mol%) in order to prepare several novel materials with tuned properties. The materials were characterised in detail through several techniques, namely ATR FTIR, 1H and 13C NMR, TGA, DSC, DMTA and XRD. Their hydrolytic and enzymatic degradation evaluation was also assessed. These new co-polymers showed either a semi-crystalline nature when higher PBF/PPOF ratios were used, and for approximately equal amounts of PBF and PPOF an amorphous co-polyester was obtained instead.

Hereafter, the co-polyesters will be referred to as PBF-co-PPOF 90/10, 80/20 and 50/50, according to the BD/PPO molar ratio used as feed. Table 1 presents the molar amounts of each monomer used as well as the weight average molecular weights (M w ) and dispersity (Đ = M w /M n ) of the polymers.

Characterization Methods
Size-exclusion chromatography (SEC) analyses of co-polyesters were performed on a Viscotek (Viscotek TDAmax) (Malvern, Gondomar, Portugal) equipped with a differential viscometer (DV) and right-angle laser-light scattering (RALLS, Viscotek) and refractive index (RI) detectors. The column set consisted of a PLgel 5 µm guard column followed by two columns, namely Viscotek T5000 and T4000 column, respectively. A dual piston pump was set with a flow rate of 1 mL min −1 . The eluent (DMF with 0.03% LiBr) was previously filtered through a 0.2 µm filter. The system was also equipped with an on-line degasser. The analyses were performed at 60 • C using an Elder CH-150 heater. Before injection, the samples were filtered through a PTFE membrane with 0.2 µm pore. The system was calibrated with narrow poly(methyl methacrylate) standards and the molecular weight and dispersity of the polymers were determined by conventional calibration.
Attenuated total reflectance Fourier transform infrared (ATR FTIR) spectra were obtained using a PARAGON 1000 FTIR spectrometer equipped with a single-horizontal Golden Gate ATR cell (Perkin-Elmer, MA, United States). The spectra were recorded after 128 scans, at a resolution of 4 cm −1 , within the range of 500 to 4000 cm −1 .
1 H, 13 C nuclear magnetic resonance (NMR) spectra were recorded in CDCl 3 using a Bruker AMX 300 spectrometer (Bruker, Madrid, Spain), operating at 300 or 75 MHz, respectively. All chemical shifts (δ) are expressed as parts per million (ppm), downfield from tetramethylsilane (used as the internal standard). Further, the calculation of the real incorporation of BD/PPO ratio (PBF/PPOF real ) the integration areas of OCH 2 proton resonance of F-BD diad (δ ≈ 4.40 ppm) and of F-PPO diad (δ ≈ 3. Thermogravimetric analyses (TGA) were carried out with a Setaram SETSYS analyzer (Setaram, Caluire, France) equipped with an alumina plate. Thermograms were recorded under a nitrogen flow of 20 mL min −1 and heated at a constant rate of 10 • C min −1 from room temperature up to 800 • C. Thermal decomposition temperatures were taken at 5% weight loss step and at maximum decomposition temperatures from the heated samples (T d,5% and T d,max , respectively).
Differential scanning calorimetry (DSC) thermograms were obtained with a Hitachi DSC7000X calorimeter (Hitachi, Paris, France) equipped with a liquid nitrogen cooling system, using aluminum DSC pans. Scans were carried out under nitrogen with a heating rate of 5 • C min −1 in the temperature range from −90 to 200 • C. Two heating/cooling cycles were repeated. Glass transition temperature (T g ) was determined using the midpoint approach (second heating trace); and cold crystallization (T cc ) and melting (T m ) temperatures were determined as the maximum of the exothermic crystallization peak and the minimum of the melting endothermic peak during the second heating scan, respectively.
Dynamic mechanical thermal analyses (DMTA) were performed using a material pocket accessory with a Tritec 2000 DMA (Triton, WA, United States), operating in the single cantilever mode. Tests were performed at 1 and 10 Hz and the temperature was varied from -100 to 200 • C, at 2 • C min −1 . T g was determined as the maximum peak of tan δ.
X-ray diffraction (XRD) measurements were performed using a Philips X'pert MPD diffractometer operating with CuKα radiation (λ = 1.5405980 Å) (Malvern, Gondomar, Portugal) at 40 kV and 50 mA. Samples were scanned in the 2θ range of 5 to 50 • , with a step size of 0.04 • , and time per step of 50 s.
In vitro hydrolytic and enzymatic degradation tests were carried out using press-molded square-shape samples (ca. 69-113 mg) of the prepared polyesters and placed into closed containers with phosphate buffer saline solution (PBS) (10 mL) or with a PBS solution (10 mL) containing Porcine pancreas lipase (concentration of 0.1 mg mL −1 ), for each test, respectively. The specimens were taken out of the related solution at regular intervals (each 7 days), rinsed thoroughly with distilled water, dried at room temperature for 4 h and, weighed. To prevent saturation, both solutions were renewed every 7 days. Each measurement was repeated five times. The weight loss percentage was calculated using the expression: Weight loss (%) = [(W 0 − W d )/W 0 ] × 100, where, W o and W d stand for the specimens' weights prior and after incubation, respectively.

PBF-co-PPOF Co-Polyesters Synthesis and Structural Characterization
In this study the newly prepared poly(ester-ether) co-polymers are based on poly(1,4-butylene 2,5-furandicarboxylate) as rigid unit and on a soft segment derived from poly(propylene oxide) (Scheme 1). Interestingly, the poly(ether) selected has a methyl side group which plays an important role on the structure-properties features, as discussed ahead.
Typical ATR FTIR spectra of all PBF-co-PPOF co-polyesters and related PBF homopolyester ( Figure 1) displayed two weak bands near 3150 and 3115 cm −1 attributed to the ν C-H bond of the furanic ring. In addition, near 2968, 2930, 2893 and 2868 cm −1 there are four weak bands attributed to the anti-symmetrical and symmetrical stretching modes (ν C-H asym and ν C-H sym, respectively) of the C-H bond of methylene and methyl groups related to the BD and PPO moieties, respectively. Additionally, both PBF and co-polyesters spectra exhibited a very intense band near 1725 cm −1 , arising PBF-co-PPOF were synthesized via a two-step conventional melt polytransesterification approach (Scheme 1) [3], in the presence of Ti(OBu)4 catalyst and at relatively moderate temperatures, not exceeding 210 °C, to avoid undesirable side reactions involving the furan moiety (e.g., decarboxylation reactions which are commonly associated to color problem issues) [1]. The resulting polymers were isolated as powders (PBF, PBF-co-PPOF-90/10 and 80/20) or a viscous liquid (PBF-co-PPOF-50/50), in relatively good yields ranging from 65 to 71% (Table 1) in similarity to other FDCAbased polyesters [22]. Furthermore, these co-polyesters showed weight-average molecular weight (Mw) values between 36,700-48,500, and dispersity (Đ) around 2. Typical ATR FTIR spectra of all PBF-co-PPOF co-polyesters and related PBF homopolyester ( Figure 1) displayed two weak bands near 3150 and 3115 cm −1 attributed to the ν C-H bond of the furanic ring. In addition, near 2968, 2930, 2893 and 2868 cm −1 there are four weak bands attributed to the anti-symmetrical and symmetrical stretching modes (ν C-H asym and ν C-H sym, respectively) of the C-H bond of methylene and methyl groups related to the BD and PPO moieties, respectively. Additionally, both PBF and co-polyesters spectra exhibited a very intense band near 1725 cm −1 , arising from the C-O stretching vibration, typical of ester groups. Two bands at 1506 and 1573 cm −1 , arising from the C-C bond of the furan ring, and C-O-C stretching vibrations appeared at around 1271 cm −1 and the typical vibration modes of 2,5-disubstituted furans were observed at 966, 822, and 769 cm −1 in the case of PBF and PBF-co-PPOF materials. The presence of the abovementioned bands confirmed the success of the polymerization reactions.   The chemical structure characterization of all PBF-co-PPOF co-polyesters and PBF homopolyester was also studied by 1 H (( Figure 2 and Table 2 and Error! Reference source not found.), 13 C NMR (Error! Reference source not found. and Figure S1) and 2D analysis ( Figure S2). The main 1 H NMR resonances and respective assignments of all polymers studied are summarized in in Table 2 and Figure 2 displays the 1 H NMR spectrum of PBF-co-PPOF-90/10. The chemical structure characterization of all PBF-co-PPOF co-polyesters and PBF homopolyester was also studied by 1 H (Figure 2 and Table 2), 13 C NMR (Table 4 and Figure S1) and 2D analysis ( Figure S2). The main 1 H NMR resonances and respective assignments of all polymers studied are summarized in Table 2 and Figure 2 displays the 1 H NMR spectrum of PBF-co-PPOF-90/10. Furthermore, the 1 H NMR spectra data was used to access the real molar percentage of PBF and PPOF moieties in the co-polyesters backbone, due to the important impact this ratio has on the ensuing co-polyesters properties. The PBF/PPOF real incorporation was determined using the integration areas of C(O)OCH2 proton resonances (in the neighboring of furan ring) of the F-BD (δ at 4.40 ppm) and F-PPO (δ at 3.93 ppm) diads, respectively, and the main results are presented in Table  3.
From Table 3, it is possible to observe that despite the BD and PPO feed ratio, there was a tendency for a higher incorporation of PPOF into co-polyesters chains, most probably associated with BD lost during the polytransesterification step due to the high BD volatility. However, this trend is almost negligible in the case of the PBF-co-PPOF-90/10 and 80/20 co-polyesters. The number average sequence length of BF unit (Ln,BF) was also assessed assuming that PBF-co-PPOF co-polyesters are random co-polyesters. It was found that Ln,BF increased with the BF content increasing, according with the theoretically expected values [11,22,29,30]. Importantly, the copolyester with the highest amount of PBF (PBF-co-PPOF-90/10) had a Ln,BF of 6.5. This is an important  The 1 H NMR spectra of all polymers ( Figure 2 and Table 2) displayed the typical resonances attributed to the F-BD diad at approximately δ 7.21, 4.40 and 1.91 attributed to the H3 and H4 protons of the furan ring, and to the C(O)OCH 2 CH 2 and C(O)OCH 2 CH 2 protons of the BD moiety, respectively. In the co-polymers spectra, the corresponding resonances associated to the PPO-F diads were also detected: δ 5.26, 3.67 and 1.34 ppm, arising from the C(O)OCH, C(O)OCH 2 and C(O)OCHCH 3 protons, in the neighboring of the furan ring, respectively. Moreover, the protons related to the PPO-PPO units were also identified at δ 3.54, 3.40, and 1.15 ppm, related to OCH 2 (ether linkage), OCH and OCHCH 3 , respectively. Furthermore, the 1 H NMR spectra data was used to access the real molar percentage of PBF and PPOF moieties in the co-polyesters backbone, due to the important impact this ratio has on the ensuing co-polyesters properties. The PBF/PPOF real incorporation was determined using the integration areas of C(O)OCH 2 proton resonances (in the neighboring of furan ring) of the F-BD (δ at 4.40 ppm) and F-PPO (δ at 3.93 ppm) diads, respectively, and the main results are presented in Table 3. From Table 3, it is possible to observe that despite the BD and PPO feed ratio, there was a tendency for a higher incorporation of PPOF into co-polyesters chains, most probably associated with BD lost during the polytransesterification step due to the high BD volatility. However, this trend is almost negligible in the case of the PBF-co-PPOF-90/10 and 80/20 co-polyesters.
The number average sequence length of BF unit (L n,BF ) was also assessed assuming that PBF-co-PPOF co-polyesters are random co-polyesters. It was found that L n,BF increased with the BF content increasing, according with the theoretically expected values [11,22,29,30]. Importantly, the co-polyester with the highest amount of PBF (PBF-co-PPOF-90/10) had a L n,BF of 6.5. This is an important structural feature that is in accordance with a crystalline domain dominated by PBF segments and corresponding melting behavior, as discussed above.
In terms of 13 C NMR analysis (Table 4 and Figure S1), the observed resonances were in agreement with their expected chemical structure and corroborated the above 1 H NMR results, as well as the ATR FTIR data.

X-Ray Diffraction Analysis
The XRD patterns of PBF-co-PPOF-90/10 and 80/20 co-polymers ( Figure 3) indicated that they were semi-crystalline polymers, displaying both an amorphous halo around 2θ~21 • and diffraction peaks at 2θ ≈ 18, 23 and 25 • , quite similar to the PBF pattern although in the former case peaks were more intense than for the co-polymers probably due to a higher crystallinity. In fact, the PBF XRD pattern (Figure 3), exhibited intense diffraction peaks at 2θ~10, 18, 23 and 25 • [16, [31][32][33]. This clearly indicates that the ability of PBF-co-PPOF co-polyesters to crystallize is mainly associated to PBF segments (with L n,BF equal to 6.5 and 4.2, respectively). Moreover, these results are in perfect agreement with the below DSC and DMTA data, and also with other FDCA or TPA-based PPEs reported in the literature [21,22,34,35].
indicates that the ability of PBF-co-PPOF co-polyesters to crystallize is mainly associated to PBF segments (with Ln,BF equal to 6.5 and 4.2, respectively). Moreover, these results are in perfect agreement with the below DSC and DMTA data, and also with other FDCA or TPA-based PPEs reported in the literature [21,22,34,35].
As expected, in the case of the viscous liquid PBF-co-PPOF-50/50 co-polyester, only a halo centered at 2θ ≈ 19 ° was observed, in accordance with an essential amorphous nature.  As expected, in the case of the viscous liquid PBF-co-PPOF-50/50 co-polyester, only a halo centered at 2θ ≈ 19 • was observed, in accordance with an essential amorphous nature.
In summary, the highly desirable properties obtained included, the T g of the co-polyesters decreased with the increasing content of soft PPOF segments (tailored behavior), although these results were still higher than those of amorphous PPO homopolyester (T g between −67 to −72 • C [36,37]). These results were expected since the incorporation of more PPO flexible moieties into FDCA-based co-polyesters typically gives rise to a decrease on the co-polymers' thermal features [21,22]. Importantly, the co-polymers with higher BF content, PBF-co-PPOF-90/10 and 80/20, showed a typical segmented polymers behavior [22], with a T m very close to that of stiff PBF and a T g below room temperature closer to that of soft PPO, revealing that PBF units were the main responsible for the crystalline behavior of the ensuing materials, whereas the PPOF segments were associated with the amorphous domain. Hence, for PBF-co-PPOF-90/10 the range of working temperatures was quite enlarged: within −39 and 139 • C.
In general the TGA thermograms (Table 5 and Figure S6) of the co-polyesters (carried out under nitrogen atmosphere) exhibited one major characteristic event at the maximum decomposition temperatures (T d,max ) of 340−365 • C. Also, the newly prepared co-polymers showed to be thermally stable up to T d,5% ≈ 308 • C.
As shown in Table 5, the co-polymers had both T d,5% and T d,max results lower than those observed to PBF. These less favorable thermal results could be associated with the presence of the appending methyl group, as already reported for other polyesters also having side groups, such as poly(2,3-butylene 2,5-furandicarboxylate) compared to PBF [17]. Nevertheless, all PBF-co-PPOF co-polyesters have higher T d,max than PPO.

Hydrolytic and Enzymatic Degradation Tests
As mentioned above, in addition to specific thermal properties, it was also important to understand if the newly prepared materials degraded under hydrolytic and enzymatic conditions for a relatively short period of time. The evaluation of the hydrolytic and enzymatic degradation behavior was performed in terms of weight loss percentage versus time (Figure 4) for PBF-co-PPOF-80/20 co-polyester, with real incorporation of PPOF moieties around 24 mol%. However, the polymer's weight loss under hydrolytic conditions was almost negligible for a relatively short period of 12 weeks, and a little bit higher under enzymatic conditions (2.3%). This result was in the same line with those reported for PBF-co-PEGF co-polyesters incorporating similar amount of PBF moieties [23], and also with studies on PBT-co-PEGT co-polyesters in similar hydrolytic conditions as the ones used in this study (pH~7 and 37 • C) [38].
Materials 2019, 12 FOR PEER REVIEW 10 80/20 co-polyester, with real incorporation of PPOF moieties around 24 mol%. However, the polymer's weight loss under hydrolytic conditions was almost negligible for a relatively short period of 12 weeks, and a little bit higher under enzymatic conditions (2.3%). This result was in the same line with those reported for PBF-co-PEGF co-polyesters incorporating similar amount of PBF moieties [23], and also with studies on PBT-co-PEGT co-polyesters in similar hydrolytic conditions as the ones used in this study (pH ~ 7 and 37 °C) [38]. The incubation with a Porcine pancreas lipase in PBS solution slightly increased the weight loss of the co-polyester, but it was still very low. Through hydrolytic and enzymatic degradation only 1.5 and 2.3% weight loss were achieved, respectively. Nevertheless, the data obtained are good preliminary results, and showed some evidence of degradation of PBF-co-PPOF (also checked by FTIR) associated with the hydrolysis of the ester groups, but further studies using other enzymes such as for example cutinases, could enhance the results obtained. Some positive results were previously obtained with cutinase from Humicola insolens or Thermobifida cellulosilytica to hydrolyze The incubation with a Porcine pancreas lipase in PBS solution slightly increased the weight loss of the co-polyester, but it was still very low. Through hydrolytic and enzymatic degradation only 1.5 and 2.3% weight loss were achieved, respectively. Nevertheless, the data obtained are good preliminary results, and showed some evidence of degradation of PBF-co-PPOF (also checked by FTIR) associated with the hydrolysis of the ester groups, but further studies using other enzymes such as for example cutinases, could enhance the results obtained. Some positive results were previously obtained with cutinase from Humicola insolens or Thermobifida cellulosilytica to hydrolyze PET and PEF [12,13]. Further, incorporation of PPO in the co-polymers, is also, expected to enhance degradability.

Conclusions
In conclusion, a new class of FDCA-based poly(ester-ether) co-polymers has been accomplished, incorporating both stiff and soft moieties into their backbone. The ensuing co-polyesters have shown high thermal stability (T d,max between 340 to 365 • C) and T g at sub-ambient temperatures, namely from −42.3 to −32.6 • C. Moreover, the semi-crystalline character was only observed for co-polyesters with higher BF content, revealing that PBF units were mainly responsible for the crystalline behavior of the ensuing materials, whereas the PPOF were associated with the amorphous domain. Furthermore, PBF-co-PPOF-80/20 co-polyester had shown a week hydrolysable behavior, presenting a maximum percentage weight loss of 2.3 %, after 12 weeks.
Finally, due to their high thermal stability, as well as the presence of both stiff and soft moieties in the co-polymer chains, these materials could find interesting industrial applications, namely as thermoplastic polymers.