Prolonged Glass Transition due to Topological Constraints in

Communication Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

pubs.acs.org/JACS

Prolonged Glass Transition due to Topological Constraints in Polyrotaxanes Kazuaki Kato,*,†,‡ Akihiro Ohara,† Hideaki Yokoyama,† and Kohzo Ito*,† †

Downloaded via KEAN UNIV on August 1, 2019 at 22:08:40 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Department of Advanced Materials Science, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8561, Japan ‡ Research and Services Division of Materials Data and Integrated System, National Institute for Materials Science, 1-2-1 Sengen, Tsukuba, Ibaraki 305-0047, Japan S Supporting Information *

prolong the glass transition dynamics accompanied by negligible cooperativity. Mechanical interactions in the form of topological constraints among mechanically interlocked components were synthetically tuned by changing the coverage of a glassforming polyrotaxane, as shown in Figure 1a. We synthesized three polyrotaxanes with different coverages composed of two common components: polyethylene glycol (PEG, Mn = 20 000) and methoxyethylated α-cyclodextrin (ME-CD). Methoxyethylation was selected, because it yields glass-formers with relatively low glass transition temperatures that resist thermal decomposition during measurements at considerably higher temperature than the transition. The coverage was controlled by additives that promote the nucleation of crystals of the host α-cyclodextrin (CD). It was recently demonstrated that the molecular weight of the inclusion complex between PEG and CD increased in the presence of oligoethylene glycols (MW < 700), likely because the additives act as temporal guests to form channel-type CD crystal nuclei that can take PEG directly in exchange for the temporal guests, resulting in the increase of coverage.14 Triethylene glycol (TEG) and oligoethylene glycol with MW ∼ 200 (OEG) were used as guest additives, and the obtained inclusion complexes were end-capped to isolate the polyrotaxanes. Figure 1b,d shows the 1 H nuclear magnetic resonance (NMR) spectra and size exclusion chromatography (SEC) traces, respectively, of two polyrotaxanes containing additives and without additives. Although all the 1H NMR spectra contain peaks attributed to PEG and CD, the PEG/CD ratios differed. The single peak of each SEC trace corresponds to a molecular weight substantially higher than that of PEG, suggesting successful isolation of polyrotaxanes. From the molar ratio between CD and the PEG repeating unit obtained from the NMR spectra, the coverages were calculated to be 38, 34, and 28% for the polyrotaxanes containing OEG, TEG, and no promoter, respectively. The increased coverage obtained with the additives was supported by the increased apparent molecular weight observed from the SEC traces. Subsequent methoxyethylation of the ring components yielded different coverage glass-forming polyrotaxanes, named GF38, GF34, and GF28 according to the coverage ratios. Incidentally, the degrees of

ABSTRACT: Topological constraints in polyrotaxanes significantly affected their glass transition dynamics. The effects of the constraints were systematically studied using a series of different coverage glass-forming polyrotaxanes consisting of a common polymer and threaded ring molecule of varying ratios. Although their ratios were similar and hence exhibited similar Tg values by differential thermal analysis, mechanical relaxation was considerably prolonged with increasing coverage. The relaxation became a two-step process: a faster step at a common temperature near the Tg and another which was prolonged by the coverage increase. Relaxation dynamics analysis revealed that segment motions, which are cooperative translations of different components, freeze at considerably higher temperatures than the Tg with increasing coverage. This suggests that although the rings are released from conventional interactions at the Tg, their cooperative translational motions are significantly constrained by the threading polymers with increasing coverage.

G

lass transition dynamics varies with molecular structures and influences the properties of glass.1,2 The rapid slowdown of the dynamics when approaching the glass transition temperature, Tg, is attributed to cooperative motions,3−6 though differences in cooperativity remain difficult to be interpreted from the molecular structures. The Arrhenius behavior of so-called strong glass formers, such as SiO2 and GeO2, has been considered due to the directional atomic interactions that suppress cooperative motions.7−9 On the other hand, in glass-forming polymers, the stiffness of the main chains dominates the cooperativity along with intermolecular interactions.10−12 This suggests that the one-dimensional connectivity in polymers significantly influences cooperativity. Recently, we developed a novel class of glass-forming supramolecular polymers consisting of mechanically interlocked ring molecules and a threading polymer.13 These socalled polyrotaxanes showed almost Arrhenius-type glass transition dynamics, unlike conventional polymers, suggesting potential effects of the intramolecular mechanical connectivity on the glass transition behaviors. Herein, we demonstrate that experimentally enhanced mechanical interactions significantly © XXXX American Chemical Society

Received: June 6, 2019

A

DOI: 10.1021/jacs.9b06063 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

Figure 1. (a) Reaction schemes and summary for the different coverage polyrotaxanes (PRs) and their glass-forming derivatives (GFs). (b) The corresponding 1H NMR spectra (400 MHz, DMSO-d6, 343 K) of PRs and (c) those (400 MHz, D2O, 298 K) of GFs, and (d) SEC traces (eluent, DMSO; detector, RI) of PRs and (e) those of GFs. DMSO = dimethyl sulfoxide.

methoxyethylation were similar ranging from 46 to 50% (Figure 1c,e). The glass-forming polyrotaxanes were melt-press molded to obtain glassy films and their amorphous structures were confirmed by analysis of the corresponding X-ray diffraction profiles (Figure S1). All glasses exhibited a common amorphous halo at approximately q = 4.7 nm−1 corresponding to the correlation distance between their ring components.15 It should be noted that the glasses mostly consisted of the ring component (ME-CD), and the weight fractions of ME-CD, ϕring, increased slightly with increasing coverage: 0.82 < ϕring < 0.85. The thermal properties of the films were first analyzed by DSC and the profiles are shown in Figure 2, compared with the profile of a pure ME-CD glass (ϕring = 1). Independent of the coverages, the glass transition temperatures determined by DSC, Tg(d), of the polyrotaxanes were constant at approximately 280 K, which is approximately 10 K lower than that of ME-CD. The similar Tg(d) values can be explained by regarding the polyrotaxanes as miscible blends of the two components: PEG and ME-CD. From the Fox equation,16 the small difference in ϕring (0.82−0.85) by the coverage is predicted to result in Tg(d) changes of 4 K at the most. Incidentally, the exothermic peaks at 325−330 K correspond to the melting of PEG crystallized during the first cooling process at approximately 250 K. The threading PEG tends to crystallize in lower coverage polyrotaxanes because it is covered with ring components.17 The significantly decreased intensity of the exothermic peaks with increasing coverage suggests no significant phase separations occurred where the naked PEG was homogeneously divided by ME-CD. On the other hand, the viscoelastic relaxation at the glass transition region is strongly affected by the coverage. Figure 3 shows the temperature dependence of the dynamic Young’s

Figure 2. DSC profiles (second heating, 10 K/min) of the glassforming polyrotaxanes (GFs) with different coverages and of a pure ring component glass (ME-CD).

modulus (E′), loss modulus (E″), and loss tangent (tan δ) of the polyrotaxanes and ME-CD. E′ values on the order of 1 GPa, which is common for polymer glasses, sharply decreased near room temperature for GF28 and ME-CD, corresponding to the glass transition. The peak temperature of E″, which is regarded as the Tg determined by viscoelastic measurements, Tg(v), of GF28 was slightly lower than that of ME-CD. This is consistent with the above discussion based on the DSC results, although each Tg(v) is approximately 10 K higher than the corresponding Tg(d). However, with increasing coverage, the Tg(v) values shifted to considerably higher temperatures. The Tg(v) values of GF34 and GF38 were approximately 30 and 50 B

DOI: 10.1021/jacs.9b06063 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

dependences of the viscoelasticity, master curves at the glass transition regimes were generated (Figure S3). Here, 282 K was set as the reference temperature because it is close to the common Tg(d) in the GFs. The time−temperature superposition principle was applicable to GF28 and GF34 throughout the entire glass transition regime, whereas the data of GF38 at T < 360 K were not superimposable. The shift factors, aT, depending on the temperature normalized by Tg(d) are plotted in Figure 4. At the temperatures considerably

Figure 4. Arrhenius plots of the shift factors aT for different coverage polyrotaxanes. A representative Tg(d) (=282 K) was used as a common reference temperature. Physically meaningless data points due to inaccurate time−temperature superposition are shown as open circles. Figure 3. Temperature dependence of (a) the complex dynamic Young’s moduli and (b) loss tangent of the glass-forming polyrotaxanes (GFs) with different coverages and a pure ME-CD glass. All measurements were performed at 1 Hz.

higher than Tg(d), all polyrotaxanes exhibited Arrhenius behavior. Their similar activation energies indicate a common elementary molecular motion in the melt, likely the segment motion of the threading chain accompanied by the threaded rings. For GF28, the Arrhenius dependence continues until close to the Tg(d) without a general increase in the apparent activation energy, reproducing the “strong” glass former behavior reported previously.13 However, increasing coverage resulted in a negative deviation from the Arrhenius dependence, which is opposite to the general deviation attributed to effects of cooperativity near the Tg. For GF34, the negative deviation starts at approximately 320 K (Tg(d)/T = 0.87) with decreasing apparent activation energy approaching the Tg(d). This indicates that the segment motion of the polyrotaxane was frozen and the dominant molecular motion shifted to other local motions. Similar freezing of the segment motion occurred at higher temperatures in GF38, as indicated by the invalid superposition principle at T < 360 K (Tg(d)/T = 0.78). In other words, higher coverage prolongs the release of the segment motion. The PEG and ME-CD components exhibit no β-relaxation near the Tg(d), though much faster motions are observed, including the rotation of ring side chains.13 Thus, the smaller relaxation observed separately from the main relaxation was attributable to precursor local motions of the main relaxation, resembling the so-called Johari−Goldstein relaxation.24,25 Regardless of the coverage, the precursor molecular motions of the segment motions should be released at the Tg(d). Once

K higher, respectively, than that of GF28, and significantly exceed that of ME-CD. Notably, the highest-coverage GF38 exhibited a clear two-step relaxation. E′ began decreasing at approximately 280 K and then mainly decreases from 345 K, and the corresponding shoulder and peak of E″ appear at approximately 295 and 350 K, respectively. GF34 also exhibited a shoulder of E′ near 300 K (see the isolated linear graph in Figure S2). The first small relaxations at the lower temperature are close to the Tg(v) of GF28, and the temperature is nearly independent of the coverage, similar to Tg(d). Thus, the first relaxation is likely associated with common molecular interactions. Although the strong interactions between different components or negative free volume effects in binary blends occasionally make the Tg values higher than their intrinsically higher-T g components, the T g discrepancy is 15 K at its highest.18−21 Analogous two-step prolonged relaxations were observed in several nanocomposite systems because the polymers were strongly bound to the surface of nanofillers.22,23 However, these conventional interactions cannot explain the significant prolongation of the main relaxation induced by increasing coverage. The origin of the significantly prolonged main relaxation of the polyrotaxanes can be discussed from a perspective of glass transition dynamics. From the frequency and temperature C

DOI: 10.1021/jacs.9b06063 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society

behaviors. Unlike conventional binary blends, the influence of the minor component polymer was enhanced by increasing the topological connections as opposed to intermolecular interactions. Therefore, the threading of the polymer and its related efficiency are critical for glass transition. Efficient threading enhanced the constraints and prolonged segment motion, even though the polymer fraction decreased. To date, the glass transition dynamics of polymers are usually focused on the one-dimensional connectivity and intermolecular interactions. The effects of mechanical interactions in polyrotaxanes on the glass transition suggest that systematic studies using various mechanically interlocked supramolecules will help elucidate universal glass transition mechanisms. Multiple analyses of the polyrotaxanes focusing on microscopic dynamics using NMR or neutron scattering and free volume effects are necessary for understanding their unique properties, including unsolved origins of the abnormally low cooperativity. Microscopic dynamics analyses using neutron scattering and free volume measurements are in progress.

the rings are free from the interactions between neighboring rings or polymers, they begin rotating. Subsequently, the rings diffuse in the same manner as the ME-CD in the absence of threading polymers (Figure 5a). In reality, to realize their

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.9b06063. Synthetic procedures for the different coverage glassforming polyrotaxanes, their 1H NMR spectra, SEC traces, XRD profiles, isolated viscoelastic spectrum for GF34, and master curves (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*[email protected] *[email protected] ORCID

Figure 5. Schematic illustrations of the effects of topological constraints on the translational motions of two connecting CDs in (a) ME-CD, (b) polyrotaxanes with low coverage, and (c) polyrotaxanes with high coverage.

Kazuaki Kato: 0000-0002-9997-8599 Hideaki Yokoyama: 0000-0002-0446-7412 Notes

The authors declare no competing financial interest.

■

translational motions, the rings must move cooperatively with the threading polymer. More importantly, the neighboring rings topologically connected via a common threading chain must also move cooperatively to some degree (Figure 5b). Thus, the degree of the constraint depends on the contour length of the partial chains between the connected neighboring rings. When the contour length is decreased as a result of increased coverage, the translational motion of the ring is constrained more strongly by the connected rings (Figure 5c). Because of the enhanced topological constraints, the translational motions associated with the main relaxation are sufficiently prolonged to be observed separately from the rotational motion. It should be noted that although the weight fractions of the threading polymer are low and are largely unchanged with the coverage (0.15 < ϕPEG < 0.18), the average contour length of the partial chains differed substantially with the length decreased to 63% with increasing coverage from 28% to 38%. By threading through the ring cavities, the minor component polymers dominate the glass transition behaviors due to topological constraints, whereas their contribution to conventional interactions is limited. In this study, we demonstrated that the topological constraints in polyrotaxanes tuned by synthetically changing the coverage have a significant impact on the glass transition

ACKNOWLEDGMENTS This work was supported by a JSPS KAKENHI Grant (16H06050) and the ImPACT Program of the Council for Science, Technology and Innovation (Cabinet Office), Government of Japan.

■

REFERENCES

(1) Angell, C. A. Relaxation in Liquids, Polymers and Plastic Crystals - Strong/Fragile Patterns and Problems. J. Non-Cryst. Solids 1991, 131−133, 13−31. (2) 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. (3) Adam, G.; Gibbs, J. H. On the Temperature Dependence of Cooperative Relaxation Properties in Glass Forming Liquids. J. Chem. Phys. 1965, 43, 139−146. (4) Bennemann, C.; Donati, C.; Baschnagel, J.; Glotzer, S. C. Growing Range of Correlated Motion in a Polymer Melt on Cooling Towards the Glass Transition. Nature 1999, 399, 246−249. (5) Weeks, E. R.; Crocker, J. L.; Levitt, A. C.; Schofield, A.; Weitz, D. A. Three-Dimensional Direct Imaging of Structural Relaxation Near the Colloidal Glass Transition. Science 2000, 287, 627−629. (6) Russell, E. V.; Israeloff, N. E. Direct Observation of Molecular Cooperativity Near the Glass Transition. Nature 2000, 408, 695−698. D

DOI: 10.1021/jacs.9b06063 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Communication

Journal of the American Chemical Society (7) Tsuneyuki, S.; Tsukada, M.; Aoki, H.; Matsui, Y. First-Principles Interatomic Potential of Silica Applied to Molecular Dynamics. Phys. Rev. Lett. 1988, 61, 869−872. (8) Coslovich, D.; Pastore, G. Understanding Fragility in Supercooled Lennard-Jones Mixtures. I. Locally Preferred Structures. J. Chem. Phys. 2007, 127, 124504. (9) Geske, J.; Drossel, B.; Vogel, M. Fragile-to-Strong Transition in Liquid Silica. AIP Adv. 2016, 6, 035131. (10) Wind, M.; Graf, R.; Heuer, A.; Spiess, H. W. Structural Relaxation of Polymers at the Glass Transition: Conformational Memory in Poly(n-alkylmethacrylates). Phys. Rev. Lett. 2003, 91, 155702. (11) Stukalin, E. B.; Douglas, J. F.; Freed, K. F. Application of the Entropy Theory of Glass Formation to Poly(α-olefins). J. Chem. Phys. 2009, 131, 114905. (12) Xu, W.-S.; Freed, K. F. Influence of Cohesive Energy and Chain Stiffness on Polymer Glass Formation. Macromolecules 2014, 47, 6990−6997. (13) Kato, K.; Mizusawa, T.; Yokoyama, H.; Ito, K. Polyrotaxane Glass: Peculiar Mechanics Attributable to the Isolated Dynamics of Different Components. J. Phys. Chem. Lett. 2015, 6, 4043−4048. (14) Takahashi, S. Inclusion Complexation between Polymer Chains and Cyclodextrins. Tokyo, The University of Tokyo, 2016, Ph.D. thesis; S. Takahashi. K. Ito, H. Yokoyama. Mechanism of Inclusion Complex Formation between Poly(ethylene glycol) Brush and alphaCyclodextrin, and Synthesis of Highly Filled Polyrotaxane. Polymer Preprints, Japan 2015, 64, 2H06. (15) Kato, K.; Nemoto, K.; Mayumi, K.; Yokoyama, H.; Ito, K. Ductile Glass of Polyrotaxane Toughened by Stretch-Induced Intramolecular Phase Separation. ACS Appl. Mater. Interfaces 2017, 9, 32436−32440. (16)

1 Tg

=

ϕPEG Tg,PEG

+

ϕring Tg ,ME − CD

.

(17) Travelet, C.; Schlatter, G.; Hébraud, P.; Brochon, C.; Lapp, A.; Anokhin, D. V.; Ivanov, D. A.; Gaillard, C.; Hadziioannou, G. Multiblock Copolymer Behaviour of a-CD/PEO-Based Polyrotaxanes: Towards Nano-Cylinder Self-Organization of a-CDs. Soft Matter 2008, 4, 1855−1860. (18) Kwei, T. K. The Effect of Hydrogen Bonding on the Glass Transition Temperatures of Polymer Mixtures. J. Polym. Sci., Polym. Lett. Ed. 1984, 22, 307−313. (19) Lin, A. A.; Kwei, T. K.; Reiser, A. On the Physical Meaning of the Kwei Equation for the Glass Transition Temperature of Polymer Blends. Macromolecules 1989, 22, 4112−4119. (20) Alegria, A.; Telleria, I.; Colmenero, J. Miscibility and Dielectric α-Relaxation of PECH/PVME Polymer Blends. J. Non-Cryst. Solids 1994, 172−174, 961−965. (21) Roland, C. M.; Casalini, R. Dynamics of Poly(cyclohexyl methacrylate): Neat and in Blends with Poly(α-methylstyrene). Macromolecules 2007, 40, 3631−3639. (22) Fragiadakis, D.; Pissis, P.; Bokobza, L. Glass Transition and Molecular Dynamics in Poly(dimethylsiloxane)/Silica Nanocomposites. Polymer 2005, 46, 6001−6008. (23) Sargsyan, A.; Tonoyan, A.; Davtyan, S.; Schick, C. The Amount of Immobilized Polymer in PMMA SiO2 Nanocomposites Determined from Calorimetric Data. Eur. Polym. J. 2007, 43, 3113−3127. (24) Johari, G. P.; Goldstein, M. Viscous Liquids and the Glass Transition. II. Secondary Relaxations in Glasses of Rigid Molecules. J. Chem. Phys. 1970, 53, 2372−2388. (25) Ngai, K. L.; Paluch, M. Classification of Secondary Relaxation in Glass-formers based on Dynamic Properties. J. Chem. Phys. 2004, 120, 857−873.

E

DOI: 10.1021/jacs.9b06063 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX