Faraday Discuss. Chem. Soc., 1985, 79, 33-39 Structure-Property Correlations in some Nematic Main-chain Polyesters BY ALEXANDRE BLUMSTEIN," MICHELLE M. GAUTHIER, OOMAN THOMAS AND RITA B. BLUMSTEIN Department of Chemistry, Polymer Science Program, University of Lowell, Lowell, Massachusetts 01854, U.S.A. Received 3 rd December, 1984 The homologous series of thermotropic nematic main-chain polyesters 0 has been investigated. In the case of even values of n, a smectic phase appears at n = 18. An odd-even oscillatih of nematic order parameters associated with the mesogen is observed near the clearing tempkrature, T,. At the same time, in the vicinity of T,, the relative order of mesogen and spacer is approximately the same for odd and even values of n. In unoriented specimens of polymer with spacer length n = 10, a layered structure develops progressively with increasing chain length, the sharpness and intensity of the SAX diffraction peak increasing with molecular weight.However, the degree of order remains that of a cybotactic nematic, based on measurements of the complex viscosity and miscibility behaviour. No SAX diff rac- tion peak is observed with spacer length n = 7. Connectivity between building blocks is the distinctive feature of polymeric liquid crystals (p,l.c.), as opposed to their low-molecular-mass (1.m.l.c.) counterparts. In p.l.c., mesogens can be located either in the side chains or in the main chain: a priori considerations would le+d one to predict a preponderance of smectic phases in the former and of nematic phases in the latter case.However, even a cursory glance at the review literat~rel-~ clearly shows that details of molecular structure profoundly influence phase behaviour in both cases. In the present paper we wish to illustrate this point by focusing on the homologous series of main-chain nematic polyesters of general formula: 0 1 where n = 3- 18. length n on mesophase order. Results are presented on the influence of the molecular mass and flexible spacer EXPERIMENTAL The synthesis and characterization of samples with n d 14 have been r e p ~ r t e d . ~ - ~ Inter- facial polycondensation was used to prepare molecular weights > 5000-6000 and solution 3334 STRUCTURE-PROPERTY CORRELATIONS polymerization was used for smaller values of Mn. Synthesis of diacid chlorides with n > 14 will be described elsewhere.The mesogen nematic order parameter S was measured by proton magnetic resonance (p.m.r.) as described in ref. ( 6 ) and (7) (2S,/kHz = 24.08s; see fig. 1 later). Rheological measurements were made on a rheometric viscoelastic tester. Data were obtained in the parallel-plate torsion geometry and oscillatory shear mode over the frequency range 2-100 rad s-’. A constant strain of 2% was maintained in order to keep the sample in the linear viscoelastic range. X-ray data on quenched samples were obtained by a technique described in ref. (5) and (8). RESULTS AND DISCUSSION CONNECTIVITY OF MESOGENIC GROUPS AND ODD-EVEN OSCILLATIONS AT THE CLEARING POINT L.m.l.c., in which the mesogen parts are not interconnected and the flexible portims of the molecule are free at one end, usually display a certain number of common characteristics.(i) The odd-even effect at the liquid-crystal/isotropic (LC/I) transition [ Tc( n), ASc( n) etc. as a function of the number of ‘flexible bonds’ in terminal chains] dampens out rapidly and is usually very weak for n > 4, (ii) the nematic phases often coexist with smectic phases at n > 7 - 8 and are eventually replaced by the latter at high values of n and (iii) no change of mesophase is observed in the odd-even alternation: both odd and even compounds display similar mesophases over the same range of n. This situation changes when two mesogens, or potential mesogens, are connected together via a sequence of flexible bonds. The odd-even effect is reinforced and the transition parameters [such as Tc( n), ASc( n) and Sc( n)] oscillate in a sustained fashion.Often, alternation between two different levels of order takes place. An example of such behaviour for a Schiff base phenyl alkyl ester of cinnamic acid was given by Coates and Gray’ for n = 0-3: However, systematic investigation of this phenomenon occurred only recently with the work on ‘Siamese twin’ model and the odd-even effect in p01ymers.~”~ The explanation, similar to the one forwarded by Coates and Gray,g is based on the alignment of two consecutive mesogen units via an extended flexible spacer. The two mesogen axes fall alternately in and out of alignment as n changes from even to odd. Several theoretical treatments of this phenomenon have recently appeared.l 4 MESOGEN AND SPACER ALIGNMENT IN THE ODD-EVEN SERIES OF 4,4‘-DIOXY-2,2‘- DIMETHYLAZOXYBENZENE ALKANEDIOYLS For the homologous series based on 2,2’-dimethylazoxybenzene, odd-even oscillations of AHIN and A S I N were investigated in the interval of n = 3-15 for high-molecular-weight samples with small biphase intervals.478 The experimental points fit the following equations: ( 1 4 - I N -4.7+0.16n kJ mru-* AS? = 7.55 + 0.87n J K-’ mru-’A. BLUMSTEIN, M. M. GAUTHIER, 0. THOMAS AND R. B. BLUMSTEIN 35 ( 2 4 (2b) AHodd- IN -0.94+0.19n kJmru-' AS:;" = 1.41 + 0.57n J K-' mru-' where mru is a mole of the repeating unit. Synthesis of the homologous series has recently beeen extended to n = 20. The range of the nematic phase of both odd- and even-numbered samples narrows progressively after passing through a maximum at n =: 6-8.For the odd-numbered series it seems to collapse for n = 17, while for the even-numbered samples the mesomorphic range is maintained for n = 18. A smectic mesophase appears at n = 18 and will be discussed e1~ewhere.l~ If one considers the intercepts and slopes in eqn (1) and (2) as the respective contributions of mesogen order and spacer conformational changes at the N/I transition, it is clear that the mesogen order is much larger for n = even than n = odd. An increase in the trans-conformer population at the I / N transition is apparent in both the odd and the even series, from the slopes of AH,,( n ) and ASIN( n). However, such measurements provide insufficient insight into the drastic differences in overall mesophase order that may be associated with even small differences in alkyl-chain flexibility. The odd-even oscillation of the mesogen orientational order was also investigated by p.m.r.and is shown in table 1. As high-molecular-weight samples do not align in the spectrometer field, polymers with = 8- 10 were prepared. Phase transitions are listed in table 1. Table 1. Mesogen order parameter S as a function of spacer length phase transitions biphasic supercooling n heating cooling range/"C at T,,/"C S" 5 6 7 8 9 10 1 1 12 ~~ ~~~ g28N 1541 g23K164N224I g12N136I 22K110N1661 g17K114N1321 g9KlO 1 N 1381 g20K96N 1221 g19K105N 1331 I1 50N20g I208N 122K13g 1135N8g I1 58N15g I1 27N73K9g I1 32N6g I1 18N15g I1 20N54K8g 10.5 23 6" 23 12 18 1 1 18 4 16 1 12 5 6 4 15 0.35 0.49 0.72 0.35 0.7 1 0.44 0.75 b " At Tred = 0.98.DP-= 15 ; sample alignment in spectrometer field was not investigated. " Narrow fraction (M,/ M, = 1.06). With the exception of n = 7, samples were unfractionated and thus display a fairly broad nematic-isotropic (N + I) biphase, as previously reported for samples of low degree of polymerization.6 The biphasic range reported here was measured as the width of the d.s.c. I/N peak on cooling. This appears to be a more accurate measure of the N + I biphase range than microscopic observation. The supercool- ing at the I/N transition is reported as the supercooling of the d.s.c. peak maximum at scanning rates of 20 K min-'. Note the odd-even oscillation of the biphasic range and supercooling at TIN (with the exception of n = lo), for which we have no explanation at present.For n = 10 and 7 we have previously shown a 'plateau effect' with levelling off of order at 8- 10 repeating units per chain. With the exception of n = 6 (which was36 STRUCTURE-PROPERTY CORRELATIONS not investigated), the samples in table 1 are at, or slightly below, that value. The mesogen order parameters S shown in the last column of table 1 were measured at an arbitrary reduced temperature Tred = T / TIN = 0.98, where TIN is the d.s.c. peak minimum on cooling. This corresponds approximately to the end of the biphasic region, although some trailing of a minor isotropic component might still be present. Table 1 unequivocally shows the odd-even oscillation of mesogen order para- meters occurring in unison with the ASIN oscillation. Note that the order parameter increases rapidly with decreasing temperature in the N + I biphase.Thus, an arbitrary definition of T, probably results in the scatter of S values observed for n =odd. Siamese-twin model compounds formed by a sequence of mesogen-spacer- mesogen also display odd-even order oscillations. As pointed out previously,'2 even-numbered spacers in such systems might induce efficient alignment of guest molecules in mixed systems. On the basis of an X-ray diffraction investigation of quenched aligned samples we have proposed for the n = odd series a 'normal' nematic level of order' and for the n = even series a 'micellar cybotactic' nematic ~rganization,~ in which the polymer chains are extended and confined to layers skewed with respect to the nematic director by 45-41' (angle between director and normal to the plane of the layer). In l.m.l.c., such systems are precursors of smectic C mesophases.'' It is possible that in this homologous series lateral substitution at the 2 and 2' positions, which results in a distorted mesogenic part," leads to a 'frustration' of smectic packing.In the odd-numbered series, formation of cybotactic nematic domains can be induced by external elongational flow, provided n is high enough to allow micro- phase separation. Thus for n = 11 a fibre with a quenched cybotactic nematic structure was obtained _by extrusion from the nematic melt at Tred =: 0.95 (sample molecular weight was M, = 34 000). No odd-numbered polyester with n < I1 gave a quenched cybotactic SAX pattern on-extrusion and no odd-numbered polyester, including n = 11, gave a cybotactic SAX pattern when oriented by a magnetic field.As pointed out above, preliminary investigation of the higher homologues sug- gests that a smectic mesophase develops at n = 18. In contrast, for n = odd, the large orientational fluctuations of the mesogen axis seem to produce an early collapse of the nematic mesophase at n = 17. We expect that the smectic mesophase of the even-numbered homologues will collapse eventually because of dilution of the mesogen by the spacer. The X-ray scattering data obtained from samples oriented by a magnetic field of 10 T and quenched at Tred = 0.95 indicate that the methylene sequence of even- numbered spacers is in an extended state.l9 However, the uncertainty in the d values (of &0.5-0.8 A) is compatible with the existence of a non-negligible gauche com- ponent in the sequence. In the case of n = 10 (DDA-9), d.m.r. investigation of a labelled spacer2' suggests that the order is uniform along the eight internal methyl- enes in the spacer, which is extended, though not all trans, throughout the nematic range. The spacer reorientational mobility decreases by ca. 40% as the temperature decreases throughout the nematic phase, indicating considerable stiffening at lower temperatures (consistent with the above X-ray data). Conversely, it follows that the spacer is quite flexible near the N/I transition. The Siamese-twin model com- pound C2H2,-9]-DDA-9, consisting of the sequence mesogen-spacer-mesogen, shows qualitatively the same behaviour although the temperature dependence of spacer mobility is much less pronounced.21 In the case of p.m.r.studies, simulation carried out by Martins et aZ? allowed decomposition of the spectrum of DDA-9 ( n = 10) into the respective contributions of spacer protons (26, fig. 1) and mesogen ortho protons (2SN, fig. 1). Thus theA. BLUMSTEIN, M. M. GAUTHIER, 0. THOMAS AND R. B. BLUMSTEIN 37 20 kHz Fig. 1. P.m.r. lineshapes of a representative even-numbered (n = 10; solid line) and odd- numbered ( n = 9; dashed line) polymer. Both spectra recorded at Tred = 0.92, with linewidth at 2/5 of peak maximum shown as 2S2,, and dipolar splitting due to mesogen ortho protons shown as 2SN.The spectrum is decomposed as outlined in ref. (21). temperature dependence of the ratio 82/5/8N can be used to follow disordering of the spacer relative to the mesogen.2' For DDA-9 this ratio varies from ca. 1.9 to ca. 1.6 as the temperature increases to the N/ I transition, suggesting that the spacer disorders faster than the mesogen. P.m.r. spectra of the rest of the homologous series were considered by analogy with that of DDA-9. In the vicinity of the N/I transition the ratio 82/5/8N was found to be ca. 1.6 for all polymers. This suggests that in the vicinity of T, mesogen and spacer display the same relative degree of order in both the odd- and even-numbered series. The odd-numbered chains may fluctuate more strongly from an extended conformation than the even-numbered ones. Thus, theoretical predictions of spontaneous chain stiffening in the vicinity of the I/ N transition 22 may have to incorporate consideration of details of chemical structure.INFLUENCE OF MOLECULAR WEIGHT It has been pointed out previously that the range of the nematic-isotropic (N + I) biphase narrows considerably with increasing molecular mass M. Other thermody- namic parameters, such as TIN, AHIN, ASIN and the order parameter S,, are increasing functions of M,6 varying rapidly for low M and levelling off at ca. 10 repeating units. Selective partitioning of the longest and best aligned chains into the anisotropic phase upon cooling through the I/N transition was inferred from the undulations observed in the S( T) plot of unfractionated samples.23 We have recently confirmed this partitioning by quenching DDA-9 samples in the N+ I biphase and physically separating the nematic and isotropic component^.^^ Another aspect of molecular-weight dependence of order in the DDA-9 ( n = 10) system is the progressive appearance of layered structures upon increasin the chain length, even in unaligned samples.An X-ray diffraction maximum at 16.5 1 develops38 STRUCTURE- PROPE RTY CO RRELATI 0 NS 1 0' lo-' 1 oo 1 0' lo2 frequency/s-' Fig. 2. Complex viscosity Iq*l for (0) DDA-9 and ( X) AZA-9 polyesters of Gn = 8000 as a function of frequency. DDA-9 at 180 "C ( Tred = 1.05) (- - -) and 140 "C ( Tred = 0.96) (-) ; MA-9 at 180 "c ( Tred = 1.05) (- - -) and I30 "c ( Tred = 0.94) (-). with increasing degree of polymerization, from a broad halo for m< 5 to a well defined (broad) peak for m> 10.(This corresponds to the spacing recorded for the quenched aligned samples.) The transition is progressive, the sharpness and intensity of the peak increasing with m.19 No such peak was found for MA-9 ( n = 7) on scanning the molecular weights up to 13 000. However, the mesophase of a DDA-9 with M, = 20 000 (m =r 45) is totally miscible with the mesophase of MA-9 and also with classical 1.m.n. nematics such as PAA. The nematic nature of DDA-9 is further illustrated by its bulk viscosity behaviour. Fig. 2 shows the evolution of complex viscosity Iq*l _as function of frequency for a sample of DDA-9 ( n = 10) and AZA-9 ( n = 7) of M,, = 8000, in the nematic and isotropic phase.It is easy to see that lq*l is lower in the liquid-crystalline than in the isotropic phase for both samples. This indicates that both polyesters are nematic; similar results are obtained for high-molecular-weight samples of DDA-9 and MA-9 (A?,, = 20 000). The difference in shear sensitivity between the two samples in the nematic phase may be due to different molecular mass distributions. This difference seems to subside in the isotropic phase. The relatively high value of DDA-9 complex viscosity, compared with MA-9 of similar Mn, is compatible with cybotactic nematic ordering of the n = 10 polymer. The cybotactic domains probably increase in size and perfection with molecular weight as the sample viscosity increases and order disruption created by chain ends lessens.In summary, development of the broad SAX peak at 16.5 A does not signify the onset of a smectic mesophase: DDA-9 remains a thermotropic nematic p.1.c. of higher order, at least in the range of up to ca. 45 investigated to date. It would be interesting to see whether a continuous transition from 'cybotactic nematic' to 'smectic C' level of order can be detected upon further increase in DDA-9 molecular mass.A. BLUMSTEIN, M. M. GAUTHIER, 0. THOMAS AND R. B. BLUMSTEIN 39 This work was supported by NSF Grant DMR *8303989. The wide-line n.m.r. data were obtained at the Worcester Consortium NMR Facility. We thank Dr K. R. Wissbrun of the Celanese Research Co. for complex viscosity data on DDA-9. A. Blumstein, Macromolecules, 1977, 10, 872. H. Finkelmann, in Polymer Liquid Crystals, ed.A. Ciferri, W. R. Krigbaum and R. B. Meyer (Academic Press, New York, 1982). A. Blumstein, J. Asrar and R. B. Blumstein, in Liquid Crystals and Oriented Fluids, ed. A. C. Griffin and J. F. Johnson (Plenum Press, New York, 1984), vol. 4, p. 31 1. A. Blumstein and 0. Thomas, Macromolecules, 1982, 15, 1264. ’ A. Blumstein, S. Vilasagar, S. Ponrathnam, S. B. Clough, G. Maret and R. B. Blumstein, J. Polym. Sci., Polym. Phys. Ed., 1982, 20, 887. R. B. Blumstein, E. M. Stickles, M. M. Gauthier, A. Blumstein and F. Volino, Macromolecules, 1984, 17, 177. A. F. Martins, J. B. Ferreira, F. Volino, A. Blumstein and R. B. Blumstein, Macromolecules, 1983, 16, 279. A. Blumstein, 0. Thomas, J. Asrar, P. Makris, S. B. Clough and R. B. Blumstein, J. Polym. Sci., Polym. Lett. Ed., 1984, 22, 13. D. Coates and G. W. Gray, J. Phys. (Paris), 1975, 36, 365. l o A. C. Griffin and T. R. Britt, J. Am. Chem. SOC., 1981, 103, 4957. J. A. Bouglione, A. Roviello and A. Sirigu, Mol. Cryst. Liq. Cryst., 1984, 106, 169. l 2 R. B. Blumstein, M. D. Poliks, E. M. Stickles, A. Blumstein and F. Volino, Mol. Cryst. Liq. Cryst., in press. J. Asrar, H. Toriumi, J. Watanabe, W. R. Krigbaum, A. Ciferri and J. Preston, J. Polym. Sci., Polym. Phys. Ed., 1983, 21, 11 19. l4 See, among others: ( a ) D. Y. Yoon and S. Bruckner, ZBM Research Report 4330 (47278), 1984; (b) A. Abe, Macromolecules, 1984, 17, 2280. I’ R. S. Kumar, work in progress. R. B. Blumstein, 0. Thomas, M. M. Gauthier, J. Asrar and A. Blumstein, in Polymeric Liquid Crystals, ed. A. Blumstein (Plenum Press, New York, 1984). 13 16 l 7 A. de Vries, Mol. Cryst. Liq. Cryst., 1970, 10, 219. ’* J. Bergts and H. Pemn, Mol. Cryst. Liq. Cryst., 1984, 113, 1; 269. l9 A. Blumstein, 1st SPSJInt. Polym. Con$, Japan, 1984, in Polym. J. (Jpn), 1985, 17, 277. 2o E. T. Samulski, M. M. Gauthier, R. B. Blumstein and A. Blumstein, Macromolecules, 1984,17,479. 22 P. G. de Gennes, Mol. Cryst. Liq. Cryst., Lett., 1984, 102, 95. 23 F. Volino, J. M. Alloneau, A. M. Giroud-Godquin, R. B. Blumstein, E. M. Stickles and A. 24 Work in progress. F. Volino and R. B. Blumstein, Mol. Cryst. Liq. Cryst., in press. 21 Blumstein, MoZ. Cryst. Liq. Cryst., Lett., 1984, 102, 21.