Chain Conformation, Molecular Dynamics, and Thermal Properties of

1 hour ago - Chain Conformation, Molecular Dynamics, and Thermal Properties of Poly(n-methylene 2,5-furanoates) as a Function of Methylene Unit ...
0 downloads 0 Views 3MB Size
Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Chain Conformation, Molecular Dynamics, and Thermal Properties of Poly(n‑methylene 2,5-furanoates) as a Function of Methylene Unit Sequence Length George Papamokos,*,† Theodoros Dimitriadis,† Dimitrios N. Bikiaris,‡ George Z. Papageorgiou,§ and George Floudas*,† †

Department of Physics, University of Ioannina, 451 10 Ioannina, Greece Laboratory of Polymer Chemistry and Technology, Department of Chemistry, Aristotle University of Thessaloniki, GR-541 24 Thessaloniki, Macedonia, Greece § Chemistry Department, University of Ioannina, P.O. Box 1186, 45110 Ioannina, Greece

Downloaded via NOTTINGHAM TRENT UNIV on August 22, 2019 at 20:54:44 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



S Supporting Information *

ABSTRACT: Poly(n-methylene 2,5-furanoates) is a family of biobased polymers with outstanding gas barrier and mechanical properties and with the potential to frame the future in certain applications (e.g., food packaging, fibers, and engineering thermoplastics). Herein, we used combined efforts by density functional theory calculations and experiments to explore in detail the conformational properties, the thermodynamics, and the molecular dynamics in the poly(n-methylene 2,5-furanoate) series as a function of n in the range from 2 poly(ethylene furanoate) (PEF) to 12 poly(dodecylene furanoate). The computational study employed the conformers suggested earlier [Macromolecules 2018, 51, 3515−3526] but used additional functionals and investigated, in addition to the monomer and trimer, the PEF nonamer with respect to conformations pertinent to the amorphous state. Depending on the conformer, variable dipole moments were obtained in the range from 2.1 to 6.1, 3.0 to 8.2, and 1.8 to 7.1 debye, respectively, for the monomer, the trimer, and the nonamer. Strikingly, both the trimer and more importantly, the nonamer exhibited very compact helical structures stabilized by π−π interactions of the furan rings. We suggest that the helical motifs within the amorphous state contribute to the barrier improvement for carbon dioxide in PEF as compared to PET. The distinct structural motifs of poly(n-methylene 2,5-furanoate)s exerted an influence on the sub-Tg and the segmental dynamics (average relaxation times and distribution of relaxation times, fragility, and dielectric strength). The segmental process shows Vogel−Fulcher−Tammann temperature dependence with distinctly different behaviors in the amorphous and crystalline states with Tg dependencies following an approximate linear dependence with n−1 as Tcrg = 249 ± 5 + (231 ± 18/n) and Tam g = 240 ± 5 + (232 ± 17/n). The large Tg reduction is compared with another homologous series, namely, poly(n-alkyl methacrylates), where the internal plasticization takes place at the side group. Internal plasticization is more efficient in the latter because of the mobile free end. Apart from Tg reduction, they show (i) subglass dynamics with activation energies that decrease with increasing alkyl length [from 57.8 kJ/mol in PEF (n = 2) to 47 kJ/mol in PNF (n = 9)], revealing the unlocking of local dipolar motions by the flexible spaces, (ii) a narrow distribution within the segmental process, αam, (corresponding Kohlrausch−Williams−Watts stretching exponent of 0.48, i.e., among the narrower for amorphous polymers), (iii) segments with locally nearly antiparallel dipolar orientation correlations, and (iv) a constant fragility in the amorphous state independent of alkyl chain length. We suggest that pertinent to these dynamic features is the local packing of chains composed of compact helical segments.

1. INTRODUCTION Growing environmental concerns associated with petrochemical extraction and the depletion of fossil resources have led to an increased interest in polymers from renewable resources.1−4 © XXXX American Chemical Society

Received: June 25, 2019 Revised: August 7, 2019

A

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

PEF/PPF and poly(butylene furanoate)/poly(propylene furanoate) (PBF/PPF) blends.25 On the theoretical/computational side, a thorough vibrational spectroscopy and molecular modeling study of PEF26 explored its conformational properties in the amorphous and crystalline regions. It was suggested that in the amorphous domains, where intermolecular interactions are weaker, PEF chains prefer a helical conformation. On the other hand, the zigzag conformation was the preferred crystalline motif with α and β PEF polymorphs. Surprisingly, the latter was energetically unfavorable, but it was stabilized by C−H···O bonds among different chain segments. In view of the growing interest in polyfuranoates and rather than reporting incremental work on each member of the homologous series, we decided to explore the poly(nmethylene 2,5-furanoate) series with n in the range from 2 (PEF) to 12 poly(dodecylene 2,5-furanoate) (PDoF) with respect to the thermodynamic and dynamic properties with DSC and DS, respectively. In addition, expanding on the results from the earlier computational study,26 we explore the conformations of PEF oligomers from the monomer to the trimer and to the nonamer. By employing computationally demanding density functional theory (DFT) calculations, we verify that helical motifs are the preferred structures in the amorphous phase. We further demonstrate a peculiarity of the helical segments; they are very compact (pitch of 0.35 nm) stabilized by π−π stacking of the furan rings in PEF. These distinct structural motifs in the amorphous phase are fundamental for the understanding of the mechanical and gas barrier properties of PEF and, to some extent, dictate the dynamics in the amorphous domains (mean relaxation times, fragility, dielectric strength, and distribution of relaxation times). The work is organized as follows: first, we present the results of the computational study with respect to the conformational properties of FDCA and all monomers of the series as well as the PEF trimer and nonamer provide the respective dipole moments. This is followed by the thermodynamic properties and by the subglass and segmental dynamics within the amorphous and RAF regions as a function of alkyl chain length. The results are compared with another homologous series, where the alkyl chain appears as a side group [poly(n-alkyl methacrylates)].

The design of new renewable polymers that can replace existing petrochemical polymers requires low-cost, abundant starting materials that can be produced on an industrial scale, and at the same time, a deep understanding of polymer structure−property relationships. Poly(n-methylene 2,5-furanoates) provide an exciting contemporary example of this route. Poly(n-methylene 2,5-furanoates) are 100% biobased and recyclable polymers derived from plants that offer a range of applications. In particular, the first member of the homologous series, the poly(ethylene 2,5-furandicarboxylate) [PEF or poly(ethylene furanoate)] in sort, attracted recent attention as it can be applied in the form of bottles and films to applications, including the packaging of liquids, food, and nonfood products. PEF is a polyester based on 2,5furandircaboxylic acid (FDCA) and ethylene glycol (EG). FDCA is an important monomer that can be obtained through oxidation of 5-hydroxymethylfurfural which in turn can be derived from sugars such as glucose or fructose by dehydration.5−7 It shows a potential to replace the plastic industry giant: poly(ethylene terephthalate) (PET), a durable material derived from conventional resources. PEF’s barrier and thermal properties are superior to PET. It shows improved barrier properties for gases such as carbon dioxide (19 times better) and oxygen (6 times better), leading to a longer shelf life of packaged products. It offers higher mechanical strength, meaning that thinner PEF packaging can be produced and less resources are required. It has a higher glass temperature, Tg, a lower melting temperature, Tm, and a higher modulus than PET. A series of experimental studies8−20 investigated the chain conformation,8 morphology/structure,9,10 the thermal behavior including the crystallization, the mechanical characteristics, and the gas barrier properties18−20 of PEF. The segmental and subglass dynamics of PEF were first addressed by Dimitriadis et al. by dielectric spectroscopy (DS) combined with differential scanning calorimetry (DSC).21 The segmental process was discussed in detail in both the quenched amorphous and semicrystalline states. It was shown that PEF can be described by a three-phase model composed from a mobile amorphous fraction, a crystalline fraction, and a RAF located at their interface. Combined DSC and DS results suggested that, in samples with degrees of crystallinity around 40%, approximately a third of segments are located within the rigid (or restricted) amorphous fraction (RAF) with possible implications for the mechanical and gas barrier properties.21 On the other hand, a two-phase model was adequate in describing the β-process within the glassy state.21 Later, Bourdet et al.22 investigated the dielectric properties of PEFs, and the authors found that PEFs with an asymmetrical position of the carbonyl group on the furan ring display higher relaxation times compared to their symmetrical counterparts.22 In addition, DS investigations on poly(butylene 2,5-furandicarboxylate) (PBF) and poly(propylene 2,5-furandicarboxylate) (PPF), respectively, by Soccio et al.23 and Genovese et al.24 focused on the subglass processes. Additional studies explored the dynamics in polymer blends comprised of furanbased polyesters differing by one or two methylene groups in their repeat units.25 Blends of PEF and PPF displayed a single composition-dependent glass temperature. DS revealed dynamically homogeneous mixtures when the backbones differ by a single methylene unit and dynamically heterogeneous mixtures when they differ by two methylene units. The most interesting finding was the single segmental process in the

2. EXPERIMENTAL SECTION 2.1. Synthesis. 2.1.1. Synthesis of 2,5-Dimethylfuran-dicarboxylate. A mass of 15.6 g of 2,5-furandicarboxylic acid, 200 mL of methanol anhydrite, and 2 mL of concentrated sulfuric acid was transferred into a random flask (500 mL), and the mixture was refluxed for 5 h. Excess methanol was distilled, and the solution was filtered through a disposable Teflon membrane filter. During filtration, dimethylester was precipitated as white powder and after cooling, 100 mL of distilled water was added. The dispersion was partially neutralized by adding Na2CO3 5% w/v during stirring while pH was measured continuously. The white powder was filtered, and the solid was washed several times with distilled water and dried. The isolated white methylester was recrystallized with a mixture of 50:50 v/v methanol/water. After cooling, 2,5-dimethylfuran-dicarboxylate (DMFD) was precipitated in the form of white needles. The reaction yield was calculated at 83%. 2.1.2. Polyester Synthesis Using DMFD and Diols with 2−6 Number of Methylene Groups.27−30 The polyesters with diols having methylene groups 2−6 were prepared by the two-stage melt polycondensation method (esterification and polycondensation) in a glass batch reactor. For the preparation of PEF, PPF, and PBF, the proper amounts of DMFD/diol = 1:2.2 (diol: EG, 1,3-propanediol, B

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

temperature T, and pressure P, ε* = ε*(ω, T, P).35−37 In the analysis of the DS spectra, we have used the empirical equation of Havriliak and Negami (HN)38

1,4-butanediol, 1,5-pentanediol, and 1,6-hexanediol) were charged into the reaction tube of the polyesterification apparatus. TBT (400 ppm) was added as the catalyst, and the apparatus with the reagents was evacuated several times and filled with argon in order to remove the whole oxygen amount. The reaction mixture was heated at 433 K under argon flow (5 mL/min) for 2 h, subsequently at 453 K for additional 2 h, and finally at 453 K for 2 h. This first step (transesterification) is considered to be complete after the collection of almost all theoretical amounts of CH3OH, which was removed from the reaction mixture by distillation and collected in a graduate cylinder. In the second step of polycondensation, vacuum (5.0 Pa) was applied slowly over a period of time of about 30 min to remove the excess of diols and to avoid excessive foaming and, furthermore, to minimize oligomer sublimation, which is a potential problem during the melt polycondensation. The temperature was increased to 493 K while stirring speed was increased to 720 rpm. The reaction continued at this temperature for 1 h and after that time, the temperature was increased to 508 K for 2 h and at 523 K for additional 2 h. After the polycondensation reaction was completed, the polyester was easily removed, milled, and washed with methanol. 2.1.3. Polyester Synthesis Using DMFD and Diols with High Number of Methylene Groups (7−12).31−34 The polyesters with diols having methylene groups 7−12 were prepared by applying a variation of the two-stage melt polycondensation method (esterification and polycondensation) in a glass batch reactor. DMFD and the appropriate diols at a molar ratio of diester/diol = 1:2 (diol = 1,7heptanediol, 1,8-octanediol, 1,9-nonanediol, 1,10-decanediol, and 1,12-dodecanediol) were charged into the reaction tube of the polyesterification apparatus with 400 ppm TBT. The reaction mixture was heated at 423 K under argon atmosphere for 2 h, then at 433 K for additional 2 h, and finally at 443 K for 1 h. This first step (transesterification) is considered to complete after the collection of almost all theoretical amounts of CH3OH, which was removed from the reaction mixture by distillation and collected in a graduate cylinder. After this stage, the corresponding bishydroxyalkylene-2,5furan carboxylate monomers have been formed. In the second stage, these monomers reacted with DMFD in a molar ratio of 1:1.05 at 423 K under the argon atmosphere for 2 h, at 433 K for additional 2 h, and finally at 443 K for 1 h. During this stage, methanol was also removed as the byproduct. After that time, in the third step of polycondensation, vacuum (5.0 Pa) was applied slowly over a period of time of about 30 min. The temperature was increased to 483 K, and the polymerization was continued for 1 h at this temperature, at 493 K for 1 h, and 503 K for 0.5 h using a stirring speed of 720 rpm. After the polycondensation reaction was completed, the polyesters were easily removed, milled, and washed with methanol. 2.1.4. Characterization. PEF: IV = 0.45 dL/g and Mn = 11 200 g/ mol; PPF: IV = 0.57 dL/g and Mn = 13 900 g/mol; PBF: IV = 0.45 dL/g. Poly(pentylene furanoate) (PPeF): IV = 0.53 dL/g and Mn = 12 600 g/mol; poly(hexamethylene furanoate) (PHF): IV = 0.47 g/ dL and Mn = 13 100 g/mol; poly(heptylene furanoate) (PHepF): IV = 0.38 dL/g and Mn = 12 400 g/mol; poly(octylene furanoate) (POF): 0.43 dL/g and Mn = 34 500 g/mol; poly(nonylene furanoate) (PNF): IV = 0.58 dL/g and Mn = 32 800 g/mol; poly(decylene-2,5furanoate) (PDeF): IV = 0.47 dL/g and Mn = 36 700 g/mol; PDoF: IV = 0.49 dL/g and Mn = 39 400 g/mol. 2.2. Differential Scanning Calorimetry. Thermal analysis studies were carried out using TA Instruments temperaturemodulated DSC (TA Q2000). The instrument was calibrated with indium for the heat flow and temperature, while the heat capacity was evaluated using the sapphire standard. Nitrogen gas flow of 50 mL/ min was purged into the DSC cell. The sample mass was around 5 mg. The Al sample and reference pans are of identical mass with an error of ±0.01 mg. 2.3. Dielectric Spectroscopy. DS measurements were made as a function of temperature in the range from 163 to 303 K using a Novocontrol Alpha frequency analyzer (frequency range from 10−2 to 107 Hz). The complex dielectric permittivity ε* = ε′ − iε″, where ε′ is the real and ε″ is the imaginary part, is a function of frequency ω,

ϵ*HN(ω , T ) = ϵ∞(T ) +

σ (T ) Δϵ(T ) + 0 [1 + (iω · τHN(T ))m ]n iϵf ω

(1)

where τHN(T,P) is the characteristic relaxation time, Δε(T,P) = ε0(T,P) − ε∞(T,P) is the relaxation strength of the process under investigation, m and n (with limits 0 < m, mn ≤ 1) describe the symmetrical and unsymmetrical broadening of the distribution of relaxation times, respectively, σ0 is the dc-conductivity, and εf is the permittivity of the free space. In the fitting procedure, we employed the dielectric loss data at every temperatureas a consistency check, the dielectric permittivity data at some temperatures were also employedand extracted the pertinent parameters (Δε, τHN, m, n). From τHN, the relaxation time at maximum loss, τmax, is obtained analytically as follows i πm yz 1/ mij πmn yz zz· sin jj z τmax = τHN· sin−1/ mjjjj z j 2(1 + n) zz k 2(1 + n) { k {

(2)

In the temperature range where two relaxation processes contribute to ε*, as in the present case with the α- and β-processes, there are two ways of representing the data: the first one, followed here, is based on a summation of two HN functions and assumes statistical independence in the frequency domain. The second one proposed by Williams and Watts is a molecular theory for the dipole moment time-correlation function Cμ(t).39 2.4. Computational Methods. Initially the methylene furanoate (MF) was chosen as a model for the study of all members of the homologous series. The purpose of this choice was to explore the twodimensional dihedral space by altering the two scan coordinates defined by the two dihedral angles d1 and d2 (Oring−Cring−CO) and abbreviated as SC1 and SC2, as shown in Figure 1. Relaxed

Figure 1. Two scan coordinates SC1 and SC2 defined by two dihedral angles d1 and d2 (Oring−Cring−CO) of the MF monomer. optimization took place for 1296 initial geometries obtained by the variation of each coordinate by an incremental step of 10° (36 × 36 initial points). The DFT-ωB97X-D40,41 level of theory and the 631+(d,p) basis set were applied. The same procedure was repeated at the DFT-M06-2X42 level of theory and the 6-31+(d,p) basis set. The resulting scan grid after optimization is shown in Figure S1, Supporting Information section. Four conformers possessing the lowest energy minima were located and abbreviated as C1, C2, C3, and C4. These conformers were adopted as starting geometries for the homologous series of the molecules with n = 0−8. For each molecule of the series, four full unconstrained optimizations took place (with the exception of FDCA because there are only three different conformations for it) at the DFT-ωB97X-D level of theory and the aug-cc pVTZ basis set. Tight optimization criteria were employed, and subsequent frequency calculations confirmed that optimized geometries are true minima. Additionally, five selected geometries of PEF oligomers (EG4FDCA3) were built. These geometries were replicas of the structures retrieved from the work of Araujo et al.26 The five possible conformations selected for treatment in that work were anti FDCA gauche E G , anti−syn FDCA gauche EG , anti FDC A trans EG , synFDCAgaucheEG, and synFDCAtransEG. The antiFDCA and synFDCA C

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Table 1. Dipole Moments of Optimized Conformers at the DFT-ωB97XD Level of Theory and the aug-cc pVTZ Basis Set for the Molecules: FDCA, PEF, and Their Homologous Series with the Number of Methylene Groups n = 1, 3−8 between the Ester and the Hydroxyl Groupsa

a

Conformers of FDCA (n = 0) and of PEF (n = 2, C1, C2, C3, C4) are given in a pictorial representation. supported by the FAS Division of Science, Research Computing Group at Harvard University.

denote the same or the opposite orientation of both carbonyl bonds and the furanic oxygen, respectively, while transEG and gaucheEG correspond to the 180° or 60° dihedral angle of the EG monomer. Minimized structures from the work of Araujo et al. were requested from the authors and were kindly provided. These structures were the input for all calculations of this work regarding the trimer. Here, these conformers were minimized at the same level of theory and basis set employed by Araujo et al.26DFT-B3LYP/6-311+G(d,p)and additionally tested with three different functionals: ωB97X-D, M062X and CAM-B3LYP.43 For the M06-2X functional, the D3 version of Grimme’s dispersion44 was added, while for CAM-B3LYP, the D3 Grimme’s dispersion with Becke−Johnson damping45 terms were applied. The aug-cc pVDZ basis set was employed for all calculations of the trimer. The results of the minimized structures produced by the ωB97X-D functional and the aug-cc pVDZ basis set were minimized again with the same functional (ωB97X-D) and the 6-311+G(d,p) basis set (for consistency reasons) followed by frequency calculations which generated the IR and Raman spectrum and confirmed that these structures are true minima. The structures were compared with the results of the current work and were also investigated at the PM746 semi-empirical level of theory along with their nonamers. The latter was studied for the five initial conformations mentioned above as well. Optimization and frequency calculations produced their optimized structures and their IR spectrum. For all calculations, the Gaussian1647 software package was employed, and the results were analyzed with Gaussview 6.48 The computations in this paper were run on the Odyssey Cluster

3. RESULTS AND DISCUSSION 3.1. Computation. Dipole moments for the various conformers of the molecules FDCA, EF, and monomers are given in Table 1, while electronic energies and heat capacities at 298.15 K are given in Table S1, Supporting Information section. Results of electronic energies and calculated dipole moments for the trimer conformers of PEF are given in Table 2, while their minimized structures are shown in Figure 2. For the monomers, the most stable structures are those adopting the anticonformation. However, the energy difference between conformers of the same monomer does not exceed 5 kJ/mol (Table S1). The conformation that generates the highest dipole moment is C3 or syn because the molecule achieves the most favorable orientation of its dipoles for the maximization of the vector (Table 1). In Table 1, an odd−even effect can be observed for the dipole moment values. Conformers C1, C2, C3, and C4 with an odd (even) number of methylene units resulted to lower (higher) dipole moments. This is mainly caused by the hydrogens of the terminal hydroxyl groups that are oriented in such a way that create either a maximizing or a minimizing effect, and this is controlled by the number of methylene bridges (spacers). Each D

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

constrained optimization of the terminal hydroxyl groups adopted by Araujo et al.26). Additionally, treatment with functionals that take under consideration long-range correction or dispersion indicate an important difference: the lowest energy conformers (C1 and C2) adopt a more compact helical structure with a smaller helical pitch. The Cartesian coordinates of all minimized structures depicted in Figure 2 are given in the Supporting Information. Investigating oligomer chains containing more than three monomers is computationally very expensive at this level of theory and basis set. To deal with this problem, we tested a level of theory that is less expensive and can give comparable results. To this end, the five initial conformations of PEF were also minimized at the semi-empirical PM7 level of theory, and the minimized structures were very similar to those obtained with the three functionals: ωB97X-D, M06-2X, and CAMB3LYP. This result allowed for the reliable investigation of the nonamer of PEF at the same level of theory adopting the five initial conformations. Energetics and dipole moments are given in Table S2, Supporting Information section, while a pictorial representation of the minimized structures is given in Figure 3.

Table 2. Energetics and Dipole Moments of the PEF Trimer for the Five Initial Conformations Optimized at Various Levels of Theory conf

energy (kJ/mol)

C1 C2 C3 C4 C5

−5 998 399.16 −5 998 399.00 −5 998 387.27 −5 998 382.32 −5 998 371.12

C1 C2 C3 C4 C5

−5 995 630.91 −5 995 614.65 −5 995 560.21 −5 995 559.65 −5 995 545.25

C1 C2 C3 C4 C5

−5 995 481.65 −5 995 413.25 −5 995 402.92 −5 995 406.09 −5 995 391.70

C1 C2 C3 C4 C5

−5 995 219.05 −5 995 162.71 −5 995 153.31 −5 995 153.47 −5 995 138.46

dipole moment (debye)

energy difference (kJ/mol) with respect to C1 conformer

B3LYP 4.6 8.2 3.7 3 8.1 wb97XD 3.2 5.9 3.8 3.3 7.8 M06-2X/D3BJ 3.0 5.0 3.7 3.7 7.6 CAM-B3LYP/D3BJ 3.1 5.4 3.7 3.0 7.9

0.16 11.89 16.83 28.04

16.26 70.70 71.26 85.65

68.4 78.73 75.55 89.94

56.34 65.74 65.58 80.59

time a methylene group is added, the effect of the terminal hydroxyl groups is inverted. The five PEF oligomers EG4FDCA3 were minimized at the DFT-B3LYP level of theory and the 6-311+G(d,p) basis set. Our results at the same level of theory reproduced successfully the earlier results26 with one exception: the difference of the electronic energy between the lowest energy conformer and the second lowest energy conformer is 0.16 kJ/mol (Table 2) as opposed to 11 kJ/mol (the reason is attributed to the

Figure 3. Pictorial representation of the optimized structures for the five initial conformations of the PEF nonamer at various levels of theory. Notice the dense helical structures associated with the more preferred conformations C1 and C2.

Figure 2. Pictorial representation of the optimized structures for the five initial conformations of PEF trimer at various levels of theory (left). E

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Table 3. Estimated IR Spectra of Bands Sensitive to gauche−trans Isomerism of EG and FDCA EG CH2 wagging (ω)

B3LYPa, 6-311+G(d,p) ωB97X-D,b 6-311+G(d,p)

FDCA CH2 deformation (δ)

gauche

trans

gauche

trans

in vacuo

in vacuo

in vacuo

1351 1349

1316 1319

1444 1445

ring torsion (τ) FDCA

FDCA

CC stretching (ν) FDCA

antiFDCA

syn

anti

in vacuo

in vacuo

in vacuo

in vacuo

in vacuo

1470 1461

608 613

614 619

1545 1568

1559 1575/1582

syn

a

Scaling factor = 0.967 see ref 4 and present work. bScaling factor = 0.957 present work.

In an effort to identity the pertinent crystalline and amorphous conformations, we have explored the same IR bands but for the more compact helical structures found herein with the ωB97X-D, M06-2X, and CAM-B3LYP functionals. The estimated IR spectra of bands sensitive to gauche−trans isomerism of EG and to the syn−anti isomerism of FDCA for the five conformations of the trimer are given in Table 3. A pictorial representation of the calculated IR spectra is given in Table S3, Supporting Information section. By inspecting Tables 3 and S3, it is evident that the estimated vibrational spectra for both levels of theory and for all bands are in agreement with the experimental results and can be employed to identify crystalline and amorphous conformations.26 The IR bands corresponding to gauche EG and antiFDCA units clearly identify helices as parts of amorphous regions. In summary, the computational study employed the conformers suggested earlier26 but went further by employing additional functionals and by investigating, in addition to the monomer and trimer, the PEF nonamer with respect to the pertinent conformations within the amorphous state. Depending on the conformer, variable dipole moments were obtained in the range from 2.1 to 6.1, 3.0 to 8.2, and 1.8 to 7.1 debye, for the respective monomer, trimer, and nonamer. Strikingly, both the trimer and more importantly, the nonamer exhibited very compact helical structures (the two more preferred conformations have a helical pitchmeasured between two furanic rings belonging to two consecutive turns eachof ∼0.35 nm) stabilized by π−π interactions of the furan rings. By comparison to the IR spectra of PEF in the crystalline and amorphous samples, these structural motifs were identified within the amorphous domains. The distinct structural motifs of PEF and in poly(n-methylene 2,5-furanoates) in general are expected to influence the segmental dynamics (average relaxation times and distribution of relaxation times, fragility, and dielectric strength) in the amorphous and possibly crystalline domains. 3.2. Thermodynamics. The thermodynamics (crystallization, melting, and liquid-to-glass “transition”) of the poly(nmethylene 2,5-furanoates) series exhibit a strong function of methylene unit sequence length. Two thermal protocols were employed. In the first, samples were quenched from the melt to liquid nitrogen and subsequently heated with 10 K/min. The heating traces are depicted in Figure 4, and the glass temperatures, the apparent melting temperatures, and heats of fusion are shown in Table 4. In a second protocol, the samples followed a cooling−heating cycle with 10 K/min, and the results from the second heating run are summarized in Table S4, Supporting Information. The DSC traces from the first protocol revealed that only PPeF, PHeptF, and PNF are amorphous, whereas all homopolymers with even membered methylene units as well as the PPF (n = 3) undergo cold crystallization.

The results revealed that the isolated PEF nonamer adopts a helical conformation which is very compact. The C1, C2, and C4 conformers produce helices with a pitch 0 value close to 0.35 nm for the first two conformations and 2 nm for the latter, with lengths of 2.1, 1.7, and 6.7 nm. Of central importance to the compact helices in the C1 and C2 conformers are π−π interactions between the furan rings having a spacing of ∼0.35 nm, for example, identical to the helical pitch. Note that the inner cylinder diameters in C1 and C2 are ∼0.4 and 0.3 nm, respectively. Although more studies are necessary to explore the variety of the helical conformations in the amorphous phase, it is evident that such compact structures will play an essential role in the barrier properties of PEF. Other structures and motifs cannot be excluded and must be investigated further. However, π−π interactions are dominant in the minimized structures reported in this work. We mention here the reduced carbon dioxide permeability of PEF as compared to PET.18 We suggest that the helical motifs within the amorphous state contribute to the reduced gas barrier properties. It is important at this point to clarify if these helical structurespreferred in vacuoare also the preferred structures in the amorphous segments or whether they correspond solely to crystalline domains. The earlier systematic study26 explored this possibility by studying the IR and Raman spectra of fully amorphous and semicrystalline samples against ab initio calculations (DFT-B3LYP level of theory) from oligomers. They suggested that certain regions of the spectrum can be useful to identify the trans and gauche conformation of EG and the anti, anti−syn, or syn conformations of the FDCA segment preferred conformations in the amorphous and crystalline regions. The experimental infrared spectra of PEF in the region from 1300 to 1500 cm−1 showed a strong band at 1340 cm−1 in the crystalline sample. For the amorphous sample, the intensity of the 1340 cm−1 band decreased, while a band at 1370 cm−1 increased. The authors attributed both bands to the wagging vibration of CH2 groups. Specifically, the band at 1340 cm−1 was attributed to the trans EG units (crystalline) while that at 1370 cm−1 was attributed to the gauche EG units (amorphous).26 Moreover, the intensity around 1474−1477 cm−1 (deformation of trans CH2 groups) decreased in the amorphous samples, while a band at 1455 cm−1 appeared due to the gauche conformers. The asymmetrical νCC stretching in semicrystalline and fully crystalline PEF produced a broad band at 1577 cm−1 with a shoulder at 1582 cm−1. On the other hand, the ring deformation mode in crystalline PEF gave a broad band with a maximum at 609 cm−1 and a shoulder at 618 cm−1. These vibrational modes were attributed to the syn−anti conformation of FDCA in the crystalline state. (For more information, see the work of Araujo et al.26 and their Supporting Information.) F

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

presumably, with the relaxation within the RAF in the semicrystalline samples. 3.3. Molecular Dynamics. The molecular dynamics of the poly(n-methylene 2,5-furanoates) series exhibit a strong function of alkylene chain length. As evidenced by the thermal properties, three of the homopolymers are amorphous (PPeF, PHeptF, and PNF), whereas all homopolymers with even membered methylene units as well as PPF (n = 3) undergo cold crystallization. The former samples display a single segmental process (αam) associated with the relaxation of the amorphous segments, whereas in the latter, the segmental process (αam) is replaced by the relaxation of amorphous segments now restricted by the crystalline domains (αRAF). Representative dielectric loss curves for PPF, PBF, PPeF, and PNF are shown in Figure 5. Additional dielectric loss curves and representative fits with a summation of two Havriliak− Negami functions are shown in Figure S2, Supporting Information section. Independent of crystallization, another weaker process exists in the glassy state.21−24 Subglass molecular dynamics, despite weak in dielectric strength, play an important role as they could associate with the gas barrier properties in glassy poly(nmethylene 2,5-furanoates). These are the only processes acting at ambient temperature for the lower members of the poly(nmethylene 2,5-furanoates) series (PEF, PPF, PBF). The βprocess has a broad distribution of relaxation times (typically m ≈ 0.2−0.4 and mn ≈ 0.23−0.5) and relaxation times with temperature dependence that conforms to Arrhenius dependence as

Figure 4. DSC traces for the different samples obtained on heating (rate 10 K/min) following quenching from the melt state. Glass temperatures are indicated with arrows.

Table 4. Glass Temperatures Obtained from DSC during Heating (with 10 K/min) Following Quenching from Higher Temperatures, Change in Specific Heat, Apparent Melting Temperatures, and Heats of Fusion sample

n

Tg (K)

Δcp (J/g K)

ΔH (J/g)

Tm (K)

PEF PPF PBF PPeF PHF PHeptF POF PNF PDeF PDoF

2 3 4 5 6 7 8 9 10 12

350.0 323.7 307.3 289.9 279.7 274.5 283.2 267.8 294.5 287.0

0.40 0.31 0.37 0.42 0.42 0.41

34.6 28.4 42.9

485.6 446.9 442.3

49.0

416.5

41.9

413.1

44.8 44.7

374.8 380.0

i E yz zz τ = τ0* expjjj k RT {

(3)

where τ*0 is the relaxation time in the limit of very high temperatures, and E is the activation energy. The respective relaxation times and activation energies are shown in Figures 6 and 7. DS measurements at these low temperatures were made following quenching from the melt state (first thermal protocol). Hence, they provide the dynamics within the amorphous state (PEF, PPF, PBF, PPeF, PHeptF, and PNF) with the exception of PHF, POF, PDeF, and PDoF that readily crystallize on cooling. The activation energy is the highest for PEF (E = 57.8 kJ/mol) and decreases with increasing alkyl chain length. This suggests the unlocking of local dipolar dynamics by the more flexible alkylene chain. In addition, polymers in their amorphous state tend to have somewhat higher activation energies as shown in Figure 7. A recent DS study suggested two subglass processes in PBF.23 They were assigned to the glycolic unit (faster process, β1) and to the C−CA link between the ester carbon and the furan ring (slower, β2) with respective activation energies of Eβ1 ≈ 50 kJ/mol and Eβ2 ≈ 89 kJ/mol. It is the faster one that is reported here. On the other hand, essentially a monomodal βprocess was found in a subsequent DS study in PPF with an activation energy of ∼50 kJ/mol in agreement with the present investigation (see Table 4, below).24 These activation energies for the subglass process in PPF are reminiscent to the β2 component in terephthalate-based polymers. Based on a systematic comparison of PPF (with the monomodal subglass process) with poly(trimethylene 1,4-cyclohexanedicarboxylate) (with multimodal subglass processes) where the furan ring is replaced by the cyclohexane ring, it was suggested24 that the appearance of subglass processes is controlled by the stiffness

0.39

In general, the data show a reduction in glass temperature (and of the apparent melting temperature) with increasing alkyl chain length. This effect is anticipated by the internal plasticization of the polymer backbone by the flexible methylene spacers. In addition, by following different thermal protocols of PEF, the reduction in the heat capacity step at Tg can be used to identify the RAF, for example, the fraction of amorphous PEF segments located in the vicinity of the crystalline domains that are practically frozen. The reduced step in heat capacity is consistent with a three-phase model composed of a fraction of amorphous (φam), crystalline (φCr), and restricted-amorphous segments with φam + φCr + φRAF = 1.21 The latter fraction was found to amount to 30% for PEF. When the same fraction was estimated from the dielectric strength, Δε, the RAF fraction was estimated at ∼35%. Overall, in PEF samples with a crystallinity of around 40%, a third of segments are located within the RAF.21 The helical structures found computationally (Figure 3) are associated with the amorphous segments in fully amorphous samples and, G

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 5. Dielectric loss curves at selected temperatures obtained on heating for PPF (n = 3), PBF (n = 4), PPeF (n = 5), and PNF (n = 9). The βprocess is shown in all cases followed by the segmental process (αam) in the fully amorphous samples (PPeF and PNF) as well as in the samples prior to crystallization (PPF and PBF) and the segmental process (αRAF) following crystallization (PPF and PBF).

of the ring; the stiffer the ring, the less intense and the higher the proximity of the two subglass processes. Hence, in polymers bearing the furan ring, essentially a single β-process could be resolved. At higher temperatures, both the αam and αRAF conform to the Vogel−Fulcher−Tammann (VFT) dependence as ij B yz zz τ = τ0 expjjj j T − T0 zz k {

(4)

Here, τ0 is the relaxation time at very high temperatures, B is the activation parameter, and T0 is the “ideal” glass temperature located below the conventional glass temperature. Values of these parameters, as well as TDS g (defined as the temperature where the segmental relaxation time is at 100 s) are listed in Table 5. The segmental relaxation times corresponding to the αcr process are slower as compared to the (extrapolated) times from the αam process, that is, there is discontinuous change at the cold crystallization temperature as was observed previously for PEF21 and PBF.23 The segmental relaxation times of the αam and αRAF processes in poly(n-methylene 2,5-furanoates) are depicted in Figure 8 with filled and open symbols, respectively. The data for the lower members of the series display a discontinuous change of relaxation times at the crystallization temperature with the relaxation times within the crystalline state being systematically longer revealing the restricted segmental dynamics by the crystal. One may define a characteristic temperature for the freezing of both segmental dynamics; in the case of αam, this refers to the liquid-to-glass temperature of the fully amorphous segments that is fully accessible, whereas for αRAF, it refers to the hypothetical (e.g., obtained by extrapolation from higher temperatures) freezing temperature of the dynamics of amorphous segments that are restricted by the crystal, for example, within the RAF. Typically, the difference in the corresponding glass temperatures is about 9 K.

Figure 6. Relaxation times corresponding to the β-process for the poly(n-methylene 2,5-furanoates) series. Measurements were made on quenched samples from the melt by heating. “Amorphous” and “crystalline” states, within the indicated temperature range, are shown with filled and open symbols, respectively.

Figure 7. Activation energies corresponding to the β-process for the poly(n-methylene 2,5-furanoates) series. Polymers measured within their amorphous and crystalline states are shown with red and blue symbols, respectively.

H

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Table 5. Parameters of the VFT Equation, Glass Temperatures, Fragilities, and Arrhenius Activation Energies in the Glassy State of Poly(n-methylene 2,5-furanoates) Obtained from DS sample

n

PEF am cr

2

PPF am cr

3

PBF am cr

4

PPeF am PHF cr PHeptF am POF cr PNF amb PDeF cr PDoF cr

5 6 7 8 9 10 12

−log(τ0/s)

B (K)

a

2300 ± 40 2300a 2150 ± 40 2150a 2110 ± 30 2110a 1145 ± 25 1745 ± 25 1130 ± 10 830 ± 20 1100 ± 20 665 ± 20 655 ± 20

14 14a 14a 14a 14a 14a 10.8 14a 10.9 11a 11a 10a 10a

T0 (K) 289 297 265 272 248 262 247.3 235.8 232.4 253 225.5 244 243

± ± ± ± ± ± ± ± ± ± ± ± ±

1 1 1 1 1 1 0.7 0.5 0.4 1 0.5 1 1

Tg (τ = 100 s) 352 359.4 323.4 330.5 305.5 319.0 286.3 283.1 270.5 280.9 262.4 268.1 266.7

± ± ± ± ± ± ± ± ± ± ± ± ±

5 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.4 0.5 0.4 0.5 0.5

fragility at Tg

E (kJ mol−1)

89.8 91.8 88.3 90.3 85.5 89.3 93.6 95.8 91.4 130.7 92.6 133.3 135.8

57.8 ± 0.5 48 ± 1 47.4 ± 0.2 50 39.2 43 47 47.9 47.0 44

± ± ± ± ± ± ±

1 0.4 2 1 0.5 0.7 1

a Values held fixed; (am): amorphous; (cr): semicrystalline phases. bThe reduced dielectric strength reveal some weak crystallization at temperatures above 318 K for PNF.

Figure 8. Segmental relaxation times plotted vs inverse temperature in poly(n-methylene 2,5-furanoates). Polymers in their amorphous and crystalline states are shown with filled and open symbols, respectively. Solid and dashed lines are fits to the VFT equation and correspond, respectively, to the segmental process in the amorphous state prior to crystallization (αam) and to the same process following crystallization (αRAF).

Figure 9. Glass temperatures as a function of alkyl chain length obtained from VFT fits to the segmental processes in poly(nmethylene 2,5-furanoates). Different symbols correspond to the freezing of the segmental process (at τ = 100 s) in the amorphous state prior to crystallization (red squares) and to the same process following crystallization (blue circles). In the inset, the same data are plotted as a function of inverse methylene unit sequence length. Blue and red lines represent linear fits to the corresponding data.

This difference can better be discussed by plotting the glass temperatures corresponding to the freezing of αam and αRAF. Note here that for the αRAF process, we had to fix the values of the limiting segmental times (τ0 in eq 4). Figure 9 indicates a decreasing glass temperature with increasing methylene unit sequence length in steps known in literature as the “odd−even” effect, meaning that samples with an even number of methylene units display a higher glass temperature. Actually, the analysis presented here with respect to Figure 8 show that this effect is a mere demonstration of the restricted segmental dynamics in the semicrystalline samples as compared to the relaxation of the same segments in their amorphous state. In the semicrystalline samples, the segmental dynamics freeze at a higher temperature (they display a higher glass temperature) due to the restricted segmental dynamics by the crystalline lamellae. The Tg dependence follows an approximate linear dependence with n−1 as TRAF = 249 ± 5 + (231 ± 18/n) and g Tam = 240 ± 5 + (232 ± 17/n) for the “crystalline” and g “amorphous” samples. Notice the same slope and an intercept that differs by ∼9 K.

Tg reduction in going from the amorphous PEF (n = 2) to PNF (n = 9) is about 89 K. This should be compared with the Tg reduction in the poly(n-alkyl methacrylate) series, where the internal plasticization takes place at the side group.49−51 In the latter, in going from PEMA to PNMA, with two and nine carbon atoms in the side group following the ester group, respectively, the Tg reduction amounts to about 105 K.49 Evidently, internal plasticization works well both as a side group and in the main chain. The more efficient plasticization in the former series [e.g., in poly(n-alkyl methacrylates)] reflects the mobility of the free end in the side group. On the other hand, a common feature of both systems is that the Tgreduction is more efficient for n < 9. For n > 9, there is a plateau in the Tg(n) dependence. The latter may associate with a polyethylene-like glass temperature as in the poly(n-alkyl methacrylate) case.49−51 We mention here that a Tg of 237 K has been reported for amorphous polyethylene.52 Another important measure of the dynamics is the distribution of segmental dynamics and its temperature dependence. The latter is reflected in the low-frequency HN parameter, m, plotted in Figure 10. The figure depicts a narrow I

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 10. Low-frequency Havriliak−Negami shape parameter (m) for the segmental processes in poly(n-methylene 2,5-furanoates) plotted vs inverse temperature. Polymers in their amorphous and crystalline states are shown with filled and open symbols, respectively. Vertical lines indicate the crystallization temperatures.

Figure 11. Inverse temperature dependence of the effective relaxation strength (TΔε) corresponding to the segmental relaxation process in poly(n-methylene 2,5-furanoates). Not all data are shown (for clarity). Vertical lines indicate cold crystallization temperatures. Shaded areas in blue and orange indicate data corresponding to the crystalline and amorphous ranges.

(m ≈ 0.8−0.9) distribution of relaxation times within the amorphous state (filled symbols). The high-frequency HN parameter for the αam process amounts to mn ≈ 0.42, resulting to a Kohlrausch−Williams−Watts (KWW) exponent of βKWW ≈ 0.48 [βKWW ≈ (mn)1/1.23 under the constraint of m and n: n = 1 − 0.812 × (1 − m)0.387].53 In fact, this is among the narrower distributions associated with the segmental dynamics ever reported for amorphous polymers.54 The finding suggests minimal dynamic heterogeneity at the segmental level meaning that all dipoles relax within a narrow distribution of relaxation times. This dynamic property of poly(n-methylene 2,5furanoates) may also reflect on the more compact helical configurations associated with the amorphous state. This is contrasted by the broad distribution of relaxation times within the semicrystalline state. In the latter, in addition, the distribution is strongly temperature-dependent narrowing with increasing temperature. This suggests a continuously changing local environment of relaxing dipoles within the RAF in semicrystalline polymers. Information on the dipole−dipole orientation correlations of the poly(n-methylene 2,5-furanoates) series pertinent to the helical conformations within the amorphous state can be extracted by studying the dielectric strength of the segmental process (αam). The product of the dielectric strength with temperature is plotted in Figure 11 as a function of inverse temperature for some of the polymers investigated. We pay particular attention to the lower members of the series (PEF, PPF, and PBF) that crystallize on heating. The effective dielectric strength (TΔε) for amorphous segments is high and decreases to about half the original value during crystallization. Furthermore, TΔε decreases with increasing n in the crystalline samples reflecting a higher degree of crystallinity. The values of TΔε can be employed as a means of obtaining an estimate of the orientation of molecules within the amorphous and crystalline domains. According to the Kirkwood−Fröhlich theory,35,37 the static dielectric permittivity of polar molecules with short-range interactions can be expressed as Δϵ = ϵ′S − ϵ∞ =

μ2 N0 1 Fg 3ϵ0 kBT V

V is the number density of dipoles expressed as (ρ/M)NA, where ρ is the mass density, M is the molar mass, μ is the dipole moment, and g is the Kirkwood−Fröhlich dipole orientation correlation function. The latter is defined as the ratio of the mean-squared dipole moment measured in a dense system divided by the same quantity obtained in a noninteracting N

g=1+

∑i =01 ∑i < j μi ·μj N0μ2

(6)

case, that is, in the gas phase, as g = μ /μgas . To estimate g in the amorphous phase of PEF, the following experimental values are employed: εS′ = 8.1, ε∞ = 2.8 (F = 3.27), ρam = 1429.9 kg/m3, TΔε ≈ 1650 K and the average gas-phase dipole moment of the monomers as μgas = 3.7 debye (Table 1). These parameters result to a Kirkwood−Fröhlich dipole orientation correlation function of g ≈ 0.26. The low value suggests locally a nearly antiparallel orientation of dipoles from different chain segments within the amorphous phase of PEF. This result is not surprising if we consider the dipole moment vector of the nonamer conformations bearing the lowest energy (Figure 3, C2 and Figure S3, Supporting Information section). In most of the helical conformation found in the amorphous state, the dipole moment vectors are either perpendicular to the helical axis or bear a large component of dipole moment in the perpendicular direction. In a hexagonal closed packed arrangement of helical segments, reduction in the dipole moment by the antiparallel orientation of individual helices is feasible. In the semicrystalline state [εS′ = 4.9, ε∞ = 2.8 (F = 2.99), ρcr = 1562 kg/m3, TΔε ≈ 850 K] and for the same value of the gas-phase dipole momentunder the assumption that the conformations in the RAF are the same as in the amorphous phase of PEFthe estimated g value (∼0.14) suggests again a nearly antiparallel chain orientation of segments within the RAF. Overall, the combined results from DFT calculations and DS reveal destructive interference of dipoles associated with the packing of helical segments within the amorphous PEF. A property closely associated with the temperature dependence of the segmental relaxation times is the steepness of τ(T) in the vicinity of the glass temperature, that is, the fragility 2

(5)

The theory considers an infinite continuum of dielectric permittivity, ε′S, and within this, a spherical region containing N0 elementary dipoles that are treated explicitly. In the above equation, F = εS′ (ε∞ + 2)2/[3(2εS′ + ε∞)] is the local field, N0/ J

2

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules parameter or the steepness index, m*, defined as, ϑ log τ/ ϑ(Tg/T)|T=Tg, which is equivalent to the slope in the “fragility” plot of log τ versus Tg/T.55 The steepness index can readily be calculated from values shown in Table 5 as m* =

for example, poly(n-alkyl methacrylates). Apart from Tg reduction with increasing alkyl chain length, an effect known as “internal plasticization” (a shared feature of both series), they show (i) subglass dynamics with activation energies that decrease with increasing alkyl length, for example, revealing the unlocking of local dipolar motions by the flexible spaces, (ii) a narrow distribution within the segmental process, αam, (iii) segments with locally nearly antiparallel dipolar orientation correlations, and (iv) constant fragility in the amorphous state independent of methylene unit sequence length. We speculate that pertinent to these dynamic features is the local packing of chains composed of compact helical segments.

BTg 2.303(Tg − T0)2

(7)

This index has been shown to be sensitive to polymer architecture/chain topology.49,50,56,57 For example, in poly(nalkyl methacrylates),49 a systematic dependence of fragility was found with increasing alkyl side group length; the longer the flexible side group, the lower the steepness index. This dependence was shown to correlate with interchain distances (e.g., packing of chains).49 In general, polymers with stiff backbones are more “fragile” (e.g., PET has m* ≈ 15656) and exhibit high Tg, whereas polymers with flexible backbones are “strong”. The low fragility in the amorphous poly(n-methylene 2,5-furanoates) is consistent with the narrow segmental relaxation (Figure 10) as anticipated by an earlier study on amorphous polymers.58 Along these lines, a theoretical study59−61 suggested the rigidity/flexibility of the backbone and of the side groups as the essential parameters that govern the fragility of polymers. Values of the steepness index are plotted in Figure 12 as a function of the number of methylene

4. CONCLUSIONS This study combined efforts by DFT calculations and experiment to explore in detail the conformational properties, the thermodynamics and the molecular dynamics in the poly(n-methylene 2,5-furanoate) series as a function of n in the range from 2 (PEF) to 12 (PDoF). Studying the structure− property relationships in the poly(n-methylene 2,5-furanoate) homologous series provided additional insights and the possibility to compare with homologous series, where the alkyl chain appears as a side group. The computational study was based on an earlier molecular modeling and vibrational study on PEF but went further by exploring the conformations of PEF oligomers, from the monomer to the trimer and to the monomer. By employing additional functionals in the DFT calculations that do not ignore dispersion forces, we verify that helix is the preferred structural motif in the amorphous phase. We further demonstrate very compact helices (pitch of 0.35 nm) in the nonamer, stabilized by π−π stacking of the furan rings in the case of PEF. These structural motifs in the amorphous phase are important for a fundamental understanding of the mechanical and gas barrier properties of PEF. We suggest that the structural features are responsible for the distinct dynamics in the amorphous domains (mean relaxation times, fragility, dielectric strength, and distribution of relaxation times). Two main dynamic processes were found in poly(nmethylene 2,5-furanoates). In the glassy statepertinent to the gas barrier properties in the lower members of the series a local β-process with an activation energy decreases with increasing methylene unit sequence length (from E = 57.8 kJ/ mol in PEF to E = 47 kJ/mol in PNF). At higher temperatures, all members exhibit the segmental process with VFT temperature dependence and with distinctly different behaviors in the amorphous and crystalline states. In the semicrystalline samples, the segmental dynamics freeze at a higher temperature due to the restricted segmental dynamics by the crystalline lamellae within the RAF. The Tg dependence follows an approximate linear dependence with n−1 as Tcrg = 249 ± 5 + (231 ± 18/n) and Tam g = 240 ± 5 + (232 ± 17/n) for the “crystalline” and “amorphous” samples. The large Tg reduction in going from the amorphous PEF (n = 2) to PNF (n = 9) is still smaller than in the poly(n-alkyl methacrylate) series, where the internal plasticization takes place at the side group. Internal plasticization works well both as a side group, and in the main chain, it is more efficient in poly(n-alkyl methacrylates) because of the mobile free end. With respect to the distribution of relaxation times, we find a minimal dynamic heterogeneity at the segmental level meaning that all dipoles relax within a narrow distribution of relaxation times (low-

Figure 12. Fragility index at the respective glass temperature as a function of alkyl chain length in poly(n-methylene 2,5-furanoates). Different symbols correspond to the amorphous state prior to crystallization (red squares) and to the same process following crystallization (blue circles). In both cases, Tg is defined by extrapolation to a segmental relaxation time of 100 s.

units in the backbone for both segmental processes in the amorphous (αam) and semicrystalline (αcr) states. The data reveal intermediate fragilities for the amorphous state of poly(n-methylene 2,5-furanoates) possessing 2−9 methylene groups (m* ≈ 90). On the other hand, there is higher steepness index for the segmental process in the semicrystalline state (POF, PDeF, and PDoF). The lower fragility of poly(n-methylene 2,5furanoates) as compared to PET and the nearly constant value, irrespective of n, result from the packing of helical segments within the amorphous phase. In this respect, an investigation of the segmental dynamics as a function of pressure would be very informative.36,62 Overall, poly(n-methylene 2,5-furanoates) exhibit distinct dynamical features not shared with other homologous series, K

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

(3) Mülhaupt, R. Green Polymer Chemistry and Bio-Based Plastics: Dreams and Reality. Macromol. Chem. Phys. 2013, 214, 159−174. (4) Schneiderman, D. K.; Hillmyer, M. A. 50th Anniversary Perspective: There is a Great Future in Sustainable Polymers. Macromolecules 2017, 50, 3733−3749. (5) Motagamwala, A. H.; Won, W.; Sener, C.; Alonso, D. M.; Maravelias, C. T.; Dumesic, J. A. Toward Biomass-Derived Renewable Plastics: Production of 2,5-Furandicarboxylic Acid from Fructose. Sci. Adv. 2018, 4, No. Eaap9722. (6) Sousa, A. F.; Vilela, C.; Fonseca, A. C.; Matos, M.; Freire, C. S. R.; Gruter, G.-J. M.; Coelho, J. F. J.; Silvestre, A. J. D. Biobased Polyesters and Other Polymers from 2,5-Furandicarboxylic Acid: A Tribute to Furan Excellency. Polym. Chem. 2015, 6, 5961−5983. (7) Vilela, C.; Sousa, A. F.; Fonseca, A. C.; Serra, A. C.; Coelho, J. F. J.; Freire, C. S. R.; Silvestre, A. J. D. The quest for sustainable polyesters - insights into the future. Polym. Chem. 2014, 5, 3119− 3141. (8) Van Berkel, J. G.; Guigo, N.; Visser, H. A.; Sbirrazzuoli, N. Chain Structure and Molecular Weight Dependent Mechanics of Poly(Ethylene 2,5-Furandicarboxylate) Compared to Poly(Ethylene Terephthalate). Macromolecules 2018, 51, 8539−8549. (9) Gandini, A.; Lacerda, T. M.; Carvalho, A. J. F.; Trovatti, E. Progress of Polymers from Renewable Resources: Furans, Vegetable Oils, and Polysaccharides. Chem. Rev. 2016, 116, 1637−1669. (10) Mao, Y.; Kriegel, R. M.; Bucknall, D. G. The Crystal Structure of Poly(Ethyelene Furanoate). Polymer 2016, 102, 308−314. (11) Sousa, A. F.; Guigo, N.; Pożycka, M.; Delgado, M.; Soares, J.; Mendonça, P. V.; Coelho, J. F. J.; Sbirrazzouli, N.; Silvestre, A. J. D. Tailored design of renewable copolymers based on poly(1,4-butylene 2,5-furandicarboxylate) and poly(ethylene glycol) with refined thermal properties. Polym. Chem. 2018, 9, 722−731. (12) van Berkel, J. G.; Guigo, N.; Kolstad, J. J.; Sbirrazzuoli, N. Biaxial Orientation of Poly(ethylene 2,5-furandicarboxylate): An Explorative Study. Macromol. Mater. Eng. 2018, 303, 1700507. (13) Papageorgiou, G. Z.; Papageorgiou, D. G.; Terzopoulou, Z.; Bikiaris, D. N. Production of Bio-Based 2,5-Furan Dicarboxylate Polyesters: Recent Progress and Critical Aspects in Their Synthesis and Thermal Properties. Eur. Polym. J. 2016, 83, 202−229. (14) Stoclet, G.; du Sart, G. G.; Yeniad, B.; De Vos, S.; Lefebvre, J. M. Isothermal Crystallization and Structural Characterization of Poly(Ethylene-2,5-Furanoate). Polymer 2015, 72, 165−176. (15) Guigo, N.; van Berkel, J.; de Jong, E.; Sbirrazzuoli, N. Modelling the non-isothermal crystallization of polymers: Application to poly(ethylene 2,5-furandicarboxylate). Thermochim. Acta 2017, 650, 66−75. (16) Morales-Huerta, J. C.; de Ilarduya, A. M.; Muñoz-Guerra, S. Poly(alkylene 2,5-furandicarboxylate)s (PEF and PBF) by ring opening polymerization. Polymer 2016, 87, 148−158. (17) Burgess, S. K.; Leisen, J. E.; Kraftschik, B. E.; Mubarak, C. R.; Kriegel, R. M.; Koros, W. J. Chain Mobility, Thermal, and Mechanical Properties of Poly(Ethylene Furanoate) Compared to Poly(Ethylene Terephthalate). Macromolecules 2014, 47, 1383−1391. (18) Burgess, S. K.; Kriegel, R. M.; Koros, W. J. Carbon Dioxide Sorption and Transport in Amorphous Poly(Ethylene Furanoate). Macromolecules 2015, 48, 2184−2193. (19) Burgess, S. K.; Karvan, O.; Johnson, J. R.; Kriegel, R. M.; Koros, W. J. Oxygen Sorption and Transport in Amorphous Poly(Ethylene Furanoate). Polymer 2014, 55, 4748−4756. (20) Burgess, S. K.; Wenz, G. B.; Kriegel, R. M.; Koros, W. J. Penetrant Transport in Semicrystalline Poly(Ethylene Furanoate). Polymer 2016, 98, 305−310. (21) Dimitriadis, T.; Bikiaris, D. N.; Papageorgiou, G. Z.; Floudas, G. Molecular Dynamics of Poly(Ethylene-2,5-Furanoate) (PEF) as A Function of the Degree of Crystallinity By Dielectric Spectroscopy and Calorimetry. Macromol. Chem. Phys. 2016, 217, 2056−2062. (22) Bourdet, A.; Esposito, A.; Thiyagarajan, S.; Delbreilh, L.; Affouard, F.; Knoop, R. J. I.; Dargent, E. Molecular Mobility in Amorphous Biobased Poly(Ethylene 2,5-Furandicarboxylate) and

frequency HN parameter with values from 0.8 to 0.9). Furthermore, combined results from DFT calculations and DS revealed destructive interference of dipoles associated with the packing of helical segments within the amorphous domains. The lower fragility of poly(n-methylene 2,5furanoates) as compared to PET and the nearly constant value, irrespective of n, result from the dense packing of helical segments within the amorphous phase as opposed to the more loose packing in the poly(n-alkyl methacrylate) case. Poly(n-methylene 2,5-furanoates) provides an excellent example of a family of biobased polymers with outstanding gas barriers and mechanical properties and with the potential to frame the future in certain applications (e.g., packaging). We suggest that the helical motifs within the amorphous state contribute to their gas barrier properties. In this endeavor, exploring the conformational properties in parallel with the dynamics and thermodynamics is a step forward in understanding the polymer structure−property relationships that govern these properties.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.9b01320. DFT calculations of the Gaussian scam grid, energetics and heat capacities for the monomers, dipole moments for the PEF nonamer, calculated IR spectra for the PEF trimer, experimental DSC results obtained on heating− cooling with 10 K/min, and indicative fits to the α- and β-processes as a function of alkyl chain length (PDF) (ZIP)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (G.P.). *E-mail: gfl[email protected] (G.F.). ORCID

George Papamokos: 0000-0002-7671-2798 Dimitrios N. Bikiaris: 0000-0001-8458-4952 George Z. Papageorgiou: 0000-0003-2239-6985 George Floudas: 0000-0003-4629-3817 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The current work was supported by the Research unit on Dynamics and Thermodynamics of the UoI cofinanced by the European Union and the Greek state under NSRF 2007−2013 (Region of Epirus, call 18). We are grateful to Prof. Tim Kaxiras for providing access to the Odyssey Cluster, Research Computing Group at Harvard University. We also thank the authors of ref 26 for sharing their minimized structures and details of their methodology, upon request.



REFERENCES

(1) Hillmyer, M. A. the Promise of Plastics from Plants. Science 2017, 358, 868−870. (2) Hernández, N.; Williams, R. C.; Cochran, E. W. the Battle for the “Green” Polymer. Different Approaches for Biopolymer Synthesis: Bioadvantaged vs. Bioreplacement. Org. Biomol. Chem. 2014, 12, 2834−2849. L

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules Poly(Ethylene 2,4-Furandicarboxylate). Macromolecules 2018, 51, 1937−1945. (23) Soccio, M.; Martínez-Tong, D. E.; Alegría, A.; Munari, A.; Lotti, N. Molecular Dynamics of Fully Biobased Poly(Butylene 2,5Furanoate) as Revealed By Broadband Dielectric Spectroscopy. Polymer 2017, 128, 24−30. (24) Genovese, L.; Soccio, M.; Lotti, N.; Munari, A.; Szymczyk, A.; Paszkiewicz, S.; Linares, A.; Nogales, A.; Ezquerra, T. A. Effect of Chemical Structure On the Subglass Relaxation Dynamics of Biobased Polyesters as Revealed By Dielectric Spectroscopy: 2,5Furandicarboxylic Acid Vs. Trans-1,4-Cyclohexanedicarboxylic Acid. Phys. Chem. Chem. Phys. 2018, 20, 15696−15706. (25) Poulopoulou, N.; Pipertzis, A.; Kasmi, N.; Bikiaris, D. N.; Papageorgiou, D. G.; Floudas, G.; Papageorgiou, G. Z. Green Polymeric Materials: On the Dynamic Homogeneity and Miscibility of Furan-Based Polyester Blends. Polymer 2019, 174, 187−199. (26) Araujo, C. F.; Nolasco, M. M.; Ribeiro-Claro, P. J. A.; Rudić, S.; Silvestre, A. J. D.; Vaz, P. D.; Sousa, A. F. Inside PEF: Chain Conformation and Dynamics in Crystalline and Amorphous Domains. Macromolecules 2018, 51, 3515−3526. (27) Papageorgiou, G. Z.; Tsanaktsis, V.; Bikiaris, D. N. Synthesis of Poly(Ethylene Furandicarboxylate) Polyester Using Monomers Derived from Renewable Resources: Thermal Behavior Comparison With PET and PEN. Phys. Chem. Chem. Phys. 2014, 16, 7946−7958. (28) Papageorgiou, G. Z.; Papageorgiou, D. G.; Tsanaktsis, V.; Bikiaris, D. N. Synthesis of the Bio-Based Polyester Poly(Propylene 2,5-Furan Dicarboxylate). Comparison of Thermal Behavior and Solid State Structure With Its Terephthalate and Naphthalate Homologues. Polymer 2015, 62, 28−38. (29) Papageorgiou, G. Z.; Tsanaktsis, V.; Papageorgiou, D. G.; Exarhopoulos, S.; Papageorgiou, M.; Bikiaris, D. N. Evaluation of Polyesters from Renewable Resources as Alternatives to the Current Fossil-Based Polymers. Phase Transitions of Poly(Butylene 2,5-FuranDicarboxylate). Polymer 2014, 55, 3846−3858. (30) Papageorgiou, G. Z.; Tsanaktsis, V.; Papageorgiou, D. G.; Chrissafis, K.; Exarhopoulos, S.; Bikiaris, D. N. Furan-Based Polyesters from Renewable Resources: Crystallization and Thermal Degradation Behavior of Poly(Hexamethylene 2,5-Furan-Dicarboxylate). Eur. Polym. J. 2015, 67, 383−396. (31) Tsanaktsis, V.; Papageorgiou, G. Z.; Bikiaris, D. N. A Facile Method to Synthesize High-Molecular-Weight Biobased Polyesters from 2,5-Furandicarboxylic Acid and Long-Chain Diols. J. Polym. Sci., Part A: Polym. Chem. 2015, 53, 2617−2632. (32) Papageorgiou, G. Z.; Guigo, N.; Tsanaktsis, V.; Papageorgiou, D. G.; Exarhopoulos, S.; Sbirrazzuoli, N.; Bikiaris, D. N. On the BioBased Furanic Polyesters: Synthesis and Thermal Behavior Study of Poly(Octylene Furanoate) Using Fast and Temperature Modulated Scanning Calorimetry. Eur. Polym. J. 2015, 68, 115−127. (33) Tsanaktsis, V.; Bikiaris, D. N.; Guigo, N.; Exarhopoulos, S.; Papageorgiou, D. G.; Sbirrazzuoli, N.; Papageorgiou, G. Z. Synthesis, Properties and Thermal Behavior of Poly(Decylene-2,5-Furanoate): A Biobased Polyester from 2,5-Furan Dicarboxylic Acid. RSC Adv. 2015, 5, 74592−74604. (34) Papageorgiou, D. G.; Guigo, N.; Tsanaktsis, V.; Exarhopoulos, S.; Bikiaris, D. N.; Sbirrazzuoli, N.; Papageorgiou, G. Z. Fast Crystallization and Melting Behavior of A Long-Spaced Aliphatic Furandicarboxylate Biobased Polyester, Poly(Dodecylene 2,5-Furanoate). Ind. Eng. Chem. Res. 2016, 55, 5315−5326. (35) Kremer, F.; Schönhals, A. The Scaling of the Dynamics of Glasses and Supercooled Liquids. In Broadband Dielectric Spectroscopy; Kremer, F., Schönhals, A., Eds.; Springer Berlin Heidelberg: Berlin, Heidelberg, 2003; pp 99−129. (36) Floudas, G.; Paluch, M.; Grzybowski, A.; Ngai, K. Molecular Dynamics of Glass-Forming Systems; Springer-Verlag Berlin Heidelberg: Berlin, 2011; Vol. XII, p 176. (37) Floudas, G. 2.32-Dielectric Spectroscopy. In Polymer Science: A Comprehensive Reference; Matyjaszewski, K., Möller, M., Eds.; Elsevier: Amsterdam, 2012; pp 825−845.

(38) Havriliak, S.; Negami, S. A Complex Plane Representation of Dielectric and Mechanical Relaxation Processes in Some Polymers. Polymer 1967, 8, 161−210. (39) Williams, G.; Watts, D. C. In NMR Basic Principles and Progress; Diehl, P., Flick, E., Kosfeld, E., Eds.; Springer: Berlin, 1971; Vol. 4, p 271. (40) Chai, J.-D.; Head-Gordon, M. Long-Range Corrected Hybrid Density Functionals With Damped Atom-Atom Dispersion Corrections. Phys. Chem. Chem. Phys. 2008, 10, 6615−6620. (41) Chai, J.-D.; Head-Gordon, M. Systematic Optimization of Long-Range Corrected Hybrid Density Functionals. J. Chem. Phys. 2008, 128, 084106. (42) Zhao, Y.; Truhlar, D. G. the M06 Suite of Density Functionals for Main Group Thermochemistry, Thermochemical Kinetics, Noncovalent Interactions, Excited States, and Transition Elements: Two New Functionals and Systematic Testing of Four M06-Class Functionals and 12 Other Functionals. Theor. Chem. Acc. 2007, 120, 215−241. (43) Yanai, T.; Tew, D. P.; Handy, N. C. A New Hybrid Exchange− Correlation Functional Using the Coulomb-Attenuating Method (CAM-B3LYP). Chem. Phys. Lett. 2004, 393, 51−57. (44) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and Accurate Ab Initio Parametrization of Density Functional Dispersion Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010, 132, 154104. (45) Grimme, S.; Ehrlich, S.; Goerigk, L. Effect of the Damping Function in Dispersion Corrected Density Functional Theory. J. Comput. Chem. 2011, 32, 1456−1465. (46) Stewart, J. J. P. Optimization of Parameters for Semiempirical Methods VI: More Modifications to the NDDO Approximations and Re-Optimization of Parameters. J. Mol. Model. 2013, 19, 1−32. (47) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; Li, X.; Caricato, M.; Marenich, A. V.; Bloino, J.; Janesko, B. G.; Gomperts, R.; Mennucci, B.; Hratchian, H. P.; Ortiz, J. V.; Izmaylov, A. F.; Sonnenberg, J. L.; Williams-Young, D.; Ding, F.; Lipparini, F.; Egidi, F.; Goings, J.; Peng, B.; Petrone, A.; Henderson, T.; Ranasinghe, D.; Zakrzewski, V. G.; Gao, J.; Rega, N.; Zheng, G.; Liang, W.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Throssell, K.; Montgomery, J. A., Jr.; Peralta, J. E.; Ogliaro, F.; Bearpark, M. J.; Heyd, J. J.; Brothers, E. N.; Kudin, K. N.; Staroverov, V. N.; Keith, T. A.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A. P.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Millam, J. M.; Klene, M.; Adamo, C.; Cammi, R.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Farkas, O.; Foresman, J. B.; Fox, D. J. Gaussian 16, Revision B.01; Gaussian, Inc.: Wallingford, CT, 2016. (48) Dennington, R.; Keith, T. A.; Millam, J. M. GaussView 6.0.16; Semichem Inc., 2016. (49) Floudas, G.; Š tepánek, P. Structure and Dynamics of Poly(NDecyl Methacrylate) Below and Above the Glass Transition. Macromolecules 1998, 31, 6951−6957. (50) Floudas, G.; Placke, P.; Stepanek, P.; Brown, W.; Fytas, G.; Ngai, K. L. Dynamics of the ″Strong″ Polymer of N-Lauryl Methacrylate Below and Above the Glass Transition. Macromolecules 1995, 28, 6799−6807. (51) Beiner, M.; Huth, H. Nanophase Separation and Hindered Glass Transition in Side-Chain Polymers. Nat. Mater. 2003, 2, 595− 599. (52) Gaur, U.; Wunderlich, B. The Glass Transition Temperature of Polyethylene. Macromolecules 1980, 13, 445−446. (53) Gómez, D.; Alegría, A. On the Empirical Functions Describing the Α-Relaxation of Glass-Forming Systems. J. Non-Cryst. Solids 2001, 287, 246−251. (54) Ngai, K. Relaxation and Diffusion in Complex Systems; SpringerVerlag: New York, 2011. (55) Ngai, K. L.; Roland, C. M. Chemical Structure and Intermolecular Cooperativity: Dielectric Relaxation Results. Macromolecules 1993, 26, 6824−6830. M

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules (56) Kunal, K.; Robertson, C. G.; Pawlus, S.; Hahn, S. F.; Sokolov, A. P. Role of Chemical Structure in Fragility of Polymers: A Qualitative Picture. Macromolecules 2008, 41, 7232−7238. (57) Pipertzis, A.; Hossain, M. D.; Monteiro, M. J.; Floudas, G. Segmental Dynamics in Multicyclic Polystyrenes. Macromolecules 2018, 51, 1488−1497. (58) Böhmer, R.; Ngai, K. L.; Angell, C. A.; Plazek, D. J. Nonexponential Relaxation in Strong and Fragile Glass Formers. J. Chem. Phys. 1993, 99, 4201−4209. (59) Dudowicz, J.; Freed, K. F.; Douglas, J. F. The Glass Transition Temperature of Polymer Melts. J. Phys. Chem. B 2005, 109, 21285− 21292. (60) Dudowicz, J.; Freed, K. F.; Douglas, J. F. Fragility of GlassForming Polymer Liquids. J. Phys. Chem. B 2005, 109, 21350−21356. (61) Dudowicz, J.; Freed, K. F.; Douglas, J. F. Entropy Theory of Polymer Glass Formation Revisited. I. General Formulation. J. Chem. Phys. 2006, 124, 064901. (62) Floudas, G.; Mpoukouvalas, K.; Papadopoulos, P. The Role of Temperature and Density on the Glass-Transition Dynamics of GlassFormers. J. Chem. Phys. 2006, 124, 074905.

N

DOI: 10.1021/acs.macromol.9b01320 Macromolecules XXXX, XXX, XXX−XXX