Faraday Discuss. Chem. SOC., 1985, 79, 191-199 Dielectric, Nuclear Magnetic Resonance and Electron Spin Resonance Studies of Relaxation Processes in a Liquid-crystalline Polyester BY FRANCOISE LAUPRETRE AND CLAUDINE NOEL* Laboratoire de Physico-Chimie Structurale et Macromoleculaire, E.S.P.C.I., 10 rue Vauquelin, 75231 Paris Cedex 05, France AND W. N. JENKINS AND GRAHAM WILLIAMS Edward Davies Chemical Laboratories, Aberystwyth, Dyfed SY23 1 NE Received 10th December, 1984 The molecular dynamics of the polyester (-OC-[ Ph]3-CO-O-[CH2-CH2-O-]4)n, which has a liquid-crystalline smectic C (S,) phase, have been investigated by the e.s.r. spin-probe, dielectric relaxation and carbon-13 solid-state n.m.r. techniques. The low- temperature, high-frequency y relaxation preserves the characteristics of the local in-chain motions of polyethers.The apparent activation energy corresponds to a very simple flexible structure. The most likely mechanism of this y relaxation might be the local tg+t - tg-t transition in the central 'ether' units. Two /3 relaxation processes were found: the PL relaxation is believed to be caused by diffusional segmental motion of the flexible spacers located in isolated amorphous isotropic regions of the material and the flu process is associated with the 'ether' sequences located in the glassy Sc domains. The increasing rate of rotation of the phenyl rings about their CI-C4 axis is related to the crystal * Sc transition. Note that this motional process of the phenyl rings does not alter the mean orientation of the mesogenic groups.Considerable research effort has been expended in trying to gain a better understanding of the properties of thermotropic liquid-crystalline polymers. So far, however, little is known about the relaxation behaviour exhibited by these materials.'-" Since the detection and evaluation of polymer motions may aid in correlating polymer structure with mechanical properties, it seemed of interest to investigate the dynamics of the polyester TO1 1: ( -0C-Ph-Ph-Ph-CO-O-[CH2-CH2-0-]4),, which exhibits the following thermotropism: 12-14 115-125°C 245-255°C crystal - smectic C - isotropic liquid. This polymer is partially crystalline and, in addition to the melting and clearing endotherms, the d.s.c. curves show two transitions at low temperature: a small increase in heat capacity between -40 and 0 "C ( TgL) and a larger increase between 20 and 60 "C (TgU).l4 Two distinct relaxation processes are observed in the tem- perature dependence of the dynamic mechanical properties.l4 Recent e.s.r. spin- probe studies2 and high-power proton decoupled 13C solid-state n.m.r. measure- ments* have suggested that TgL, the lower of the glass transitions, is due to the flexible 'ether' sequences located in the amorphous isotropic regions of the material, while Tgu is associated with the flexible spacers located in the glassy smectic domains. 191192 DYNAMICS OF A LIQUID-CRYSTALLINE POLYESTER contact Fig. 1. Cross-polarization pulse sequence used in the n.m.r. experiments. It is the purpose of the present work to obtain new information on the molecular motions which occur in polyester TO11 using dielectric relaxation and to correlate these results with e.s.r.and n.m.r. results already reported for this polymerlS2 or reported in this paper. Indeed, the e.s.r. spin-probe technique provides information about any molecular motion which may occur in the polymer, dielectric relaxation measures the backbone motions of the ether and ester electric dipole groups and C solid-state n.m.r. is a powerful tool for studying the intramolecular motional processes which affect either the aromatic or the aliphatic part of the polymer. 13 EXPERIMENTAL The polyester TO1 1 was prepared at the Centre de Recherches des Carrikres de Rh6ne- Poulenc, Saint-Fons, France by standard methods, as described elsewhere.I2 Details of the properties of the polyester are given in ref.(12)-( 14). The dielectric measurements were made in the temperature range -80-36 "C using a three-terminal parallel-plate dielectric cell together with a General Radio 16 15-A capacitance- measuring assembly. The disc sample was prepared by compression-moulding the powdered material at room temperature and ca. 150 bar pressure, giving a sample of 22 mm diameter and 1.55 mm thickness. The temperature of the sample in the dielectric cell was controlled to *O.l "C by use of liquid circulated from a Lauda Ultrakryostat. The e.s.r. samples were prepared by addition of nitroxide radical to the polyester, taking care that the concentration of spin probe did not exceed 0.01 wt%. Samples were sealed under dynamic vacuum in e.s.r.tubes after repeated freeze-pump-thaw cycles. The e.s.r. measurements were performed on a Varian E-4 X-band spectrometer. The temperature in the active region of the cavity was controlled by a Varian E-257 variable-temperature control unit to a long-term stability of *O-1 "C. The e.s.r. spectra were analysed according to the method previously described.2 Carbon-1 3 cross-polarization, proton dipolar decoupling and magic-angle-spinning n.m.r. experiments were performed at 75.47 MHz on a Bruker CXP 300 spectrometer, employing quadrature detection and a single r.f. coil, which was double-tuned for both 13C and 'H. The cross-polarization pulse sequence used in the experiments is shown in fig. 1. Matched spin-lock cross-polarization transfers employed 13C and ' H magnetic-field strengths of 64 kHz.The rises in I3C polarization were obtained from a plot of the relative magnitude of the carbon magnetization as a function of the time of carbon-proton contact for very short contact times (10-200 ps). In all the spectra, spin-temperature inversion techniques were employed to minimize base-line noise and r011.15 Flip-backI6 was also used systematically to shorten theF. LAUPRETRE, C. NOEL, W. N. JENKINS AND G. WILLIAMS 193 delay time between two successive pulse sequences. Spinning experiments at the magic angle using boron nitride and [2H8] PMMA rotors were performed with spinning speeds of 3.5 kHz. THEORETICAL BACKGROUND In the domain of fast motions, information can be derived from the rise in carbon polarization in spin-lock experiment^.^'-^* When some protons are strongly coupled to carbons, as for example the protons of a methyne or a methylene group, the cross-polarization dynamics can no longer be described by a single cross- relaxation time.The initial step of the contact is coherent energy transfer between the strongly coupled carbons and protons. This oscillatory transfer is damped by the coupling of the carbon and its directly attached protons to the remote protons. At the same time, this coupling induces (i) the establishment of a quasi-equilibrium state within the tightly coupled group and (ii) a cross-relaxation energy transfer from the tightly coupled carbon and protons to the remote protons viewed as a thermal bath. At long times the rise in polarization is well described by an exponential dependence on contact duration.Motional information is provided by the short-time behaviour (first tenths of ps), which is governed by the coherent energy transfer. In the case of a powder sample, the short-time olarization rise can be approximated by a quadratic function of contact duration.P9p22 The rise time finf to half of the total polarization depends on the strength of the dipolar interaction (M,,)' of the carbon under interest and its bound protons. Values of tl12 as short as 20 ps for a CH2 group or 28 ps for a CH group are indicative of rigid-lattice behaviour. Longer tl12 values are evidence for motional reduction of (M,,)'. RESULTS AND DISCUSSION ')' RELAXATION From the representative e.s.r. spectra in fig.2 it can be seen that the motion of the probe varies from the rigid limit to the fast region over the temperature range from -160 to +150 "C. At low temperatures, the shape of the slow-motion spectra remains unaltered: the separation between the extrema is experimentally indisting- uishable from the rigid-limit value. However, at a temperature that depends on the probe size but is in the range from -70 to -4O"C, the separation of the outer hyperfine extrema decreases slightly with increasing temperature, indicating the onset of slow motion of the spin probe caused by local main-chain motion. If correlation times for probe tumbling in this region are plotted again reciprocal temperature in Arrhenius fashion,2 then an activation energy of 12 * 1 kJ mol-' is derived.Note that although the frequency and energetics of probe tumbling are intimately affected by the dynamics of the host polymer, it must not be assumed that the frequency and energetics of motion of the probe are strictly equal to those of the host polymer. A dielectric dispersion and absorption region is also seen above -80 "C in the frequency range 1 02- l O5 Hz (fig. 3 and 4). The loss curve clearly narrows and moves to higher frequencies as the temperature is increased. If the frequency of the maximum of dielectric loss for given temperatures is plotted against the reciprocal temperature, then an activation energy of 16.5 kJ mol-' is found, which is similar in magnitude to the activation energy determined by e.s.r. The present dielectric data indicate that, in the frequency and temperature ranges under investigation, local motions do occur: dipoles have limited freedom.Note also that the loss occursI94 Fig. 2. DYNAMICS OF A LIQUID-CRYSTALLINE POLYESTER /--- /---/&----. 0 --/ \ \ /- - /’ ---- -/ 4- // : /’ ‘ / / ‘.,I ‘ 1 \ I E.s.r spectra at different temperatures (in “C) of polyester Toll doped with: in the frequency and temperature ranges where loss is also observed for poly(oxy- methylene),23-’’ poly( ethylene oxide)26-28 and poly(tetramethy1ene oxide).” According to Wetton and Williams? the y process found for these materials must arise from a special kind of main-chain motion which is different from the cooperative micro-brownian motions of the chain responsible for the p relaxation. A local twisting motion was suggested to explain the y processes: the dipoles undergo a damped torsional oscillation within potential-energy minima prescribed by theF. LAUPRETRE, c.NOEL, w. N. JENKINS AND G. WILLIAMS 195 4 .O- E' - - - - 80 2.5- ! I I I I 4 5 l o g ( f l W 3 Fig. 3. Plot of the dielectric constant E' as a function of frequency for given f 36 16 0 -20 -40 -60 -80 temperatures 0 3 4 5 log(flHz) ("0 Fig. 4. Plot of the dielectric loss factor E" as a function of frequency for given temperatures196 DYNAMICS OF A LIQUID-CRYSTALLINE POLYESTER 0 10 20 30 40 50 60 70 tcdw Fig. 5. Plot of the variation of carbon magnetization in a spin-lock experiment as a function of contact duration (0, a, 0, b, and 0, c p s ) for polyester Toll. conformation of the chain.3* Dielectric and dynamic mechanical measurements carried out on multiblock copolymers composed of poly(ethy1ene oxide) and bis- phenol-A polycarbonate have clearly indicated that PEO segments as small as a dimer or a sequence of five or six bonds can accommodate the intrachain motions involved in the y relaxation of the PEO homo polymer^.^^ A small-scale motion, i.e. a two-site model similar to Monnerie's three-bond explains the low-temperature-high-frequency relaxation of these copolymer^.^' The y-process mechanism might be mainly due to local tg+tc*tg-t transition in the PEO blocks. To a certain extent polyester TO1 1 can also be considered as an alternating (AB).-type multiblock copolymer composed of soft (PEO) and hard (p-terphenyl) segment blocks, and our dielectric data suggest that the y relaxation is occurring via a local-model process, the mechanism of which might be the flip-flop transition between tg+tt* tg-t conformation in the PEO segments.Support for this local-mode model has been obtained using I3C solid-state n.m.r. techniques. From proton-decoupled 13C n.m.r. lineshape analysis it appears that the spectra of the aromatic carbons of the polymer are those of the rigid limit below 120 OC.l As regards the carboxy groups, measurement of the principal values crll and c33 of the C=O chemical-shift tensor from the side-band intensities of the carboxy 13C line in a sampie spinning at the magic angle34 leads to a u33-u11 difference of 124 ppm at room temperature. This is close to the value of 128 ppm reported for the same carbon in a single crystal of benzoic acid,35 which indicates that, at 25 "C, there exist no fast motions of the carboxy groups able to reduce the carboxy chemical-shift anisotropy.The increase of the magnetization in a spin-lock cross-polarization experiment as a function of the contact duration is shown in fig. 5 for the various aliphatic carbons of polyester TO1 1. The tl/Z value for the methylene carbon a, which is next to the carboxy group, is 21 ps, which is the expected valueF. LAUPRETRE, c. NOEL, w. N. JENKINS AND G. WILLIAMS 197 for the rigid lattice. The carbons a appear as frozen on the timescale of the experiments, lo5 Hz. By contrast, the tl12 of the methylene carbons 6 and c are longer, indicating substantial but incomplete motional averaging of the dipolar interaction between the carbon of interest and its directly bonded protons. In the case of carbons b the carbon-proton second moment averaged by the motion is 60% of the rigid-lattice value.Such a reduction is too low to be interpreted in terms of three rotational states on the valence cone or in terms of jumps between two equilibrium conformations. It corresponds to oscillations on the valence cone of ca. 20" about one equilibrium conformation. In the case of carbons c, the reduction is larger and indicates either oscillations of larger amplitude or jumps between two equilibrium conformations. From these results we conclude that the y process is associated with small-scale motion of the flexible central -CH2-CH2-0 units. A two-site model similar to Monnerie's three-bond might reasonably explain the low-temperature-high-frequency y relaxation of polyester TO1 1, in agreement with the models proposed for the y relaxation of PE03' and bisphenol-A polycarbonate-polyoxyethylene system^.^' Note, however, that the activation ener- gies determined from e.s.r.(ca. 12 kJ mol-') and dielectric (16.5 kJ mol-') data are much smaller than the values of ca. 35-50 kJ mol-' reported for the y relaxation of these polymers .2733 1,36 Several explanations could be offered for this low apparent activation energy, possibly in combination, since it is known that when an activation energy is determined for an overall composite process, it may not be representative of the average of the activation energies of the various processes involved: (i) At low temperatures the loss peaks are very broad (fig.4), which suggests that there are a wide variety of local environments for the dipoles and that the environment places great constraints on the motion.A given 'ether' group can move only in cooperation with that environment, this being overall a slow process. As the temperature is increased, the whole system cooperatively moves faster and the overall loss peak is narrowed. As a consequence, at high enough temperatures a reference group moves in an average environment. (ii) The y relaxation may consist of two main components associated with a small amount of an isolated disordered structure and a 'glassy' smectic C phase. (iii) Above TgL (ca. -20 "C), overlapping loss processes PL and y may be present.Assuming the same frequency against temperature dependence for the PL process of polyester TO11 and the /3 process of PEO, the PL and y processes could merge in the temperature range 16-30°C. Note that marked narrowing of the loss curve occurs in this region (fig. 4). p RELAXATION AND THE CRYSTAL * SMECTIC c TRANSITION Above a critical temperature, which depends on the nature of the spin probe but which is in the range 0-40 "C, complex e.s.r. spectra are observed (fig. 2), which can be resolved into a mobile component and a solid-state component. The narrow component shows insignificant dependence of its outer peak resonance positions on temperature in contrast to the inward shift of the outer peaks of the broad line spectrum with increasing temperature.As a result, at 45-85 "C, depending on the size of the spin probe, the broad-line and narrow-line spectra coalesce and at higher temperatures the spectra show the motionally narrowed three-line pattern. However, the e.s.r. spectra are asymmetric. The low-field and the high-field lines are broader than the centre line. These features are characteristic of a spin probe undergoing anisotropic rotational reorientation. In this region, the correlation times T~ have the temperature dependence expected from WLF theory. At the crystal-smectic C198 DYNAMICS OF A LIQUID-CRYSTALLINE POLYESTER transition ( 1 15-125 "C), sharpening of the low-field and high-field lines occurs and the probes rotate isotropically.2 The superimposed spectral feature is believed to arise from two different environ- ments for the probes, i e .both rigid and rubbery regions are present. Such an assumption is consistent with a system exhibiting a double glass transition with the TgL transition occurring ca. 40-50 "C below the Tgu transition. Evidence in support of this assumption comes from d.s.c. and dynamic mechanical investigations." Two glass transitions, TgL and Tgu, separated by ca. 50 "C are detected. However, d.s.c. data for a series of samples submitted to different cooling cycles have shown that only the samples which have been rapidly quenched from the isotropic state ( T > 250 "C) exhibit a marked TgL. This suggests that the PL process is due to diffusional segmental motion of the flexible 'ether' sequences located in isolated amorphous isotropic regions of the material.This assignment seems reasonable since the temperature and frequency ranges at which the mobile component appears in the e.s.r. spectra of polyester TO1 1 also correspond to a marked change in the nitroxide- doped PEO spectrum: the broad-line spectrum, typical of a solid-state nitroxide, assumes the three-line rotationally narrowed form, which has been assigned to the onset of the p relaxation.37i38 On the other hand, the flu process seems associated with the flexible spacers located in the 'glassy' smectic C domains. At this point it seems interesting to recall that earlier proton-decoupled 13C solid-state n.m.r. studies' have shown that the linewidth of the methylene carbons regularly decreases from 40°C up to the crystal-smectic C transition and then remains unchanged.On the other hand, for the aromatic carbons agreement with the rigid lattice is observed below the crystal-smectic C transition and with a rapid rotation about the CI-C4 axis in the smectic C phase. Note that this motional process of the phenyl rings does not alter the mean orientation of the mesogenic groups. ' P. Sergot, F. Laupretre, C. Louis and J. Virlet, Polymer, 1981, 22, 1150. P. Meurisse, C. Friedrich, M. Dvolaitzky, F. Laupretre, C. Noel and L. Monnerie, Macromolecules, 1984, 17, 72. K. Mueller, K. H. Wassmer, R. W. Lenz and G. Kothe, J. Polym. Sci., Polym. Lett. Ed., 1983, 21, 785. K. Mueller, B. Hisgen, H. Ringsdorf, R. W. Lenz and G. Kothe, in Recent Advances in Liquid Crystalline Polymers, ed.L. Chapoy (Elsevier, Barking, in press). K. H. Wassmer, E. Ohmes, G. Kothe, M. Portugal1 and H. Ringsdorf, Makromol. Chem., Rapid Commun., 1982, 3, 281. C. Boeffel, B. Hisgen, U. Pschorn, H. Ringsdorf and H. W. Spiess, Isr. J. Chem., 1983, 23, 388. H. Ringsdorf, H. W. Schmidt, G. Strobl and R. Zentel, Polym. Prepr. Am. Chem. SOC., Div. Polym. Chem., 1983, 388. * N. A. Nikonorova, T. I. Borisova, L. L. Burstein and V. P. Shibaev, in 12th Europhysics Conference on Macromolecular Physics: Molecular Mobility in Polymer Systems (Leipzig, 198 I), p. 223. H. Kresse, S. Kostromin and V. P. Shibaev, Makromol. Chem., Rapid Commun., 1982, 3, 509. l o H. Kresse and R. V. Talrose, Makromol. Chem., Rapid Comrnun., 1981, 2, 369. l 1 H. Kresse and V. P. Shibaev, 2. Phys.Chem. (Leipzig), 1983, 264, 161. l 2 P. Meurisse, C. Noel, L. Monnerie and B. Fayolle, Br. Polym. J., 1981, 13, 55. l 3 L. Bosio, B. Fayolle, C. Friedrich, F. Laupretre, P. Meurisse, C. Noel and J. Virlet, in Liquid Crystals and Ordered Fluids, ed. A. C. Griffin and J. F. Johnson (Plenum Press, New York, 1984), vol. 4, p. 401. l4 C. Noel, C. Friedrich, L. Bosio and C. Strazielle, Polymer, 1984, 25, 1281. l 5 E. D. Stejskal and J. Schaefer, J. Magn. Reson., 1975, 18, 560. l6 J. Tegenfeldt and U. Haeberlen, J. Mugn. Reson., 1979, 36, 453. l 7 T. T. P. Cheung and R. Yaris, J. Chem. Phys., 1980, 72, 3804. '' D. Demco, J. Tegenfeldt and J. S. Waugh, Phys. Rev. B, 1975, 11, 4133. 2o F. Laupretre, L. Monnerie and J. Virlet, Macromolecules, 1984, 17, 1397. L. Muller, A. Kumar, T. Baumann and R. R. Ernst, Phys. Rev. Lett., 1974,32, 1402. 19F. LAUPRETRE, c. NOEL, w. N. JENKINS AND G. WILLIAMS 199 21 F. Laupretre and J. Virlet, to be published. 22 G. C. Chingas, A. N. Garroway, R. 0. Bertrand and W. B. Moniz, J. Chem. Phys., 1981,74, 127. 23 B. E. Read and G. Williams, Polymer, 1961, 2, 239. 24 Y. Ishida, Kolloidn. Zh., 1961, 174, 162. 25 G. Williams, Polymer, 1963, 4, 27. 26 T. M. Connor, B. E. Read and G. Williams, J. Appl. Chem., 1964, 14, 74. 27 Y. Ishida, M. Matsuo and M. Takayanagi, J. Polym. Sci., Part B, 1965, 3, 321. 28 K. Hikichi and J. Furuichi, J. Polym. Sci., Part A, 1965, 3, 3003. 29 R. E. Wetton and G. Williams, Trans. Furuduy SOC., 1965, 61, 2132. 30 K. Yamafuji and Y. Ishida, Kofloidn. Zh., 1962, 183, 15. 31 T. Suzuki and T. Kotaka, Macromolecules, 1980, 13, 1495. 32 B. Valeur, J. P. Jarry, F. Geny and L. Monnerie, J. Polym. Sci., Polym. Phys. Ed., 1975, 13, 667. 33 B. Valeur, L. Monnerie and J. P. Jarry, J. Polym. Sci., Polym. Phys. Ed., 1975, 13, 675. 34 J. Herzfeld and A. E. Berger, J. Chem. Phys., 1980, 73, 6021. 35 J. Kemf, H. W. Spiess, U. Haeberlen and H. Zimmermann, Chem. Phys., 1974, 4, 269; Chem. Phys. Lett., 1972, 17, 39. K. Se, K. Adachi, Y. Ishida and T. Kotaka, Rep. Prog. Polyrn. Phys. Jpn, 1977, 20, 375. 37 M. C. Lang, C. Noel and A. P. Legrand, J. Polym. Sci., Polym. Phys. Ed., 1977, 15, 1329. P. Tormala, J. Mucromol. Sci., Chem, 1979, C17, 297. 36 38