A Crysiallographic Study of the Second-order Phase Transition in Bis(p-toluene sulphonate) Diacetylene Polymer Crystals B Y RICHARD L. WILLIAMS, DAVID BLOOR AND DAVID N. BATCHELDER Department of Physics, Queen Mary College, Mile End Road, London E l 4NS AND MICHAEL B. HURSTHOUSE Department of Chemistry, Queen Mary College, Mile End Road, London El 4NS AND WILLIAM B. DANIELS Department of Physics, University of Delaware, Newark, Delaware 1971 1, U.S.A. Received 10th January, 1980 The second-order phase transition which occurs in bis(p-toluene sulphonate) diacetylene polymer crystals has been studied by X-ray diffraction. Measurements of the temperature dependence of the integrated intensity of Bragg reflections from several crystals provide evidence that the phase transition is very sensitive to the presence of defects.For one crystal several Friedel pair reflections have integrated intensities which are very asymmetric in the vicinity of the transition temperature. This suggests that the crystal was not centrosymmetric in this region and provides a possible explanation for the previous observation of pyroelectricity in pTS crystals. The second-order phase transition which occurs in crystals of the polymer of the bis (p-toluene sulphonate) of 2,4-hexadiyne-l,6 diol (pTS) has been studied by a variety of experimental techniques. The crystal structures of both the high- ' and the low- t e r n p e r a t ~ r e ~ ? ~ phases have been determined and the gradual onset of order in the low-temperature phase has been observed in X-ray s t u d i e ~ .~ , ~ Order parameter analysis has been applied to the splittings of the optical absorption and Raman lines observed in the low-temperature phase and to the intensities of " hard " vibrational modes observed by far-infrared spectrocopy.6 The order parameter fitting suggested that at low temperatures the transition had a two-dimensional character while closer to the transition temperature it appeared to be three-dimensional. There were indica- tions of a broad transition region which might be expected for a predominantly two- dimensional phase transition in which fluctuations dominate the behaviour of the system. Fig. 1 shows a projection of the P2Jc unit cell of pTS onto the ac plane at 300 K in the high temperature phase.6 The polymer chains have the chemical structure with R = CH20S02C6H4CH3 and are oriented parallel [.>C-C=C-C CR],,50 DIACETYLENE POLYMER CRYSTALS FIG.1.-Projection of the P2Jc unit cell of pTS crystals at 300 K onto the uc crystallographic plane. The polymer backbone is parallel to the b axis which is perpendicular to the uc plane. The atoms, which are identified in fig. 2, are represented by thermal ellipsoids. The solid lines outline the unit cell in the high-temperature phase. In the low-temperature phase with P21/n structure, the unit cell is doubled in the Q direction (dotted lines) and there are two inequivalent polymer chains, A and B, in the unit cell. to the b axis. The doubling of the unit cell in the a direction, which occurs in the low- temperature P2Jn structure, and the two types of polymer chain (those with type A and those with type B side-groups), are also indicated. Fig.2 shows the anisotropic thermal motion of the non-hydrogen atoms in PTS.~ The central figure is a plot of a backbone segment and one sidegroup at 300 K, and the plots labelled A and B are similar parts of the two distinct types of polymer chain at 120 K. Despite the fact FIG. 2.-Comparison of a pTS side-group at 300 K (centre diagram) with that of the two structures (top and bottom diagrams labelled A and B) in which the side-groups are found in the low-tempera- ture phase at 120 K. The anisotropic thermal motion is depicted by 50% probability ellipsoids for the non-hydrogen atoms. The atomic labels are: C, carbon; 0, oxygen; S, sulphur. Atoms C(l)’, C(1), C(2), and C(2)’ are on the backbone with the vector b defining the polymer chain direction.WILLIAMS, BLOOR, BATCHELDER, HURSTHOUSE AND DANIELS 51 that the atoms in the side-groups labelled A and B occupy positions which are slightly displaced from the positions they would have occupied at room temperature, the ther- mal ellipsoids of side-groups A and B could be completely enveloped within the thermal ellipsoids of the room-temperature molecule.A model for the phase transition has been suggested in which ordering in two dimensions predominates.6 The side-groups along a single side of a polymer chain are considered to form an ordered stack which has either the A or B structure. In the low temperature phase these stacks are ordered in two dimensions (ac plane) while the high temperature phase would consist of a random array of A and B stacks.The third dimension would only be important for a stack in the process of changing from A to B or vice versa. This model would appear to satisfy in a qualitative way the requirements of both the order parameter analysis and the X-ray structural data. The present X-ray study was undertaken in order to provide further evidence against which this model could be tested. In addition it was hoped to find a possible explanation for the observation of the pyroelectric effect in pTS crystal^.^ The struc- tures of both the low- and high-temperature phases determined by X-ray analysis are centrosymmetric, while the pyroelectric effect can only be observed in crystals which lack a centre of symmetry.EXPERIMENTAL High-purity TS monomer crystals were grown by evaporation from acetone solution.' Small as-grown crystals of typical dimensions 0.2 x 0.8 x 0.2 mm were selected for study on a four-circle X-ray diffractometer. Larger crystals typically 2 x 6 x 4 mm in size were cleaved parallel to the (100) face to provide a clean surface for study by a rotating back- reflection X-ray camera. Both types of crystal were then thermally polymerized to 100% conversion at 333 K.' INTENSITY MEASUREMENTS O N THE BACK-REFLECTION CAMERA The large crystals were glued along one edge with low-temperature varnish to the alu- minium cold finger of a continuous flow cryostat. The temperature was controlled to f0.02 K and measured by an NBS calibrated platinum resistance thermometer located only 1 mm from the crystal. With the crystal fixed in the cryostat a rotating back-reflection cameralo*" was used to scan the (D,O, 3) Bragg reflection of the low-temperature phase. This difference reflection does not appear in the high-temperature phase. The normal to the (33,0, 3) planes lies at about 12" to the (100) face and the Bragg angle was approximately 68.6" with Cu Koc radiation.The pinhole X-ray collimator had an angular divergence of &20' and was 19 cm from the crystal. A scintillation counter with a 1.5 cm square NaI crystal was placed 20 cm from the crystal to measure the intensity of diffracted X-rays. Rotating the camera over a 2" range produced a profile of the diffracted Kal and Kaz peaks on a chart recorder." The integrated intensities at each temperature were then determined graphically.INTENSITY MEASUREMENTS O N THE FOUR-CIRCLE DIFFRACTOMETER The smaller crystals were studied on a Nonius CAD-4 X-ray diffractometer with com- puter control. Intensities were collected with Ni-filtered Cu Koc radiation and an 0 / 2 8 scan mode. In each 96-step scan the outer 16 steps on each side constituted left (B,) and right (B,) backgrounds, and the central 64 steps the peak count (C). The integrated in- tensity (lo) of a reflection was calculated from the equation lo = IC - 2(B1 + &)]. The temperature of the crystals was held constant during the measurements to f0.5 K in a stream of cold nitrogen gas. The accuracy of temperature measurement by a thermocouple in the52 DIACETYLENE POLYMER CRYSTALS 1.0 0.8 gas flow was estimated to be 5 2 K.At each new temperature the lattice parameters were calculated by least-squares analysis of setting angles for 25 reflections which were auto- matically centred. I l l I I l l I I I 1 I I I I -OP - 0 - 0 - a 0 - a 0 - RESULTS INTENSITY ME AS U RE ME N TS ON THE B A C K-RE F LE CTI 0 N CAMERA Fig. 3 shows the integrated intensity of the (=,O, 3) Bragg reflection of the low- temperature phase, relative to the value at 80 K, as a function of temperature for two different pTS crystals. The intensities have been ratioed against those of the (12,0,6) 0.0 a 0 . I I I I I I l l I I l l I l l 1 80 100 120 110 160 180 200 220 tempera t ure/K FIG. 3.-Relative integrated intensity of the ( n , O , 3) Bragg reflection of the low-temperature phase as a function of temperature for two different crystals (open and closed circles).The data have been scaled to remove intensity changes associated with the Debye-Waller factor. reflection, which appears in both the high-temperature and low-temperature phases, in order to remove the intensity changes associated with the Debye-Waller factor. The measurements were made with both increasing and decreasing temperature with no evidence of significant hysteresis. The estimated uncertainty in the intensity measurements increased continuously from 3% at 80 K to 20% at the highest tem- perature s. The intensity data in fig. 3 suggest that the two pTS crystals have transition tem- peratures which differ by nearly 10 K. For both crystals the transition temperature is not well defined as the intensity has no sharp cut-off.There is even significant intensity in the (%,0,3) reflection for one crystal as high as 210 K, 15 K above the transition temperature previously estimated from a photographic X-ray inve~tigation.~ For both crystals the Ka, and Ka2 Bragg peaks were clearly resolved with no significant differences in peak width at 80 K between the two crystals. Thus it is not possible to say from these measurements alone which was the " better " crystal.WILLIAMS, BLOOR, BATCHELDER, HURSTHOUSE AND DANIELS 53 INTENSITY MEASUREMENTS ON THE FOUR-CIRCLE DIFFRACTOMETER Fig. 4 illustrates the temperature dependences of the profiles of the (5,0,1) Bragg reflection of the low temperature phase which were recorded on the four-circle diffractometer for three different crystals.Photographic studies of the three crystals T K 250 230 210 2 00 190 170 150 Crystal 1 5 L 5 L Crystal 2 5 4- 5 j - - - - - - - - 54- _ - - - - - - - - __ 5 i -- I Crystal 3 51- - I LO]- -/ FIG. 4.-X-ray diffraction profiles of the (5,0,1) Bragg reflection for three different pTS crystals at several temperatures. Solid (dotted) lines represent the smoothed data taken with decreasing (increasing) temperature. The setting angles for crystals 1 and 3 were not centred between tempera- tures. Shifts in the diffraction maxima along the Bragg angle axis probably represent small rotations of the crystal in its mount. at 300 K had not shown any significant difference in their quality. Crystal 1 would appear to have been the best of the three as the onset of the (5,0,1) difference reflec- tion with decreasing temperature was sharp and the profile symmetric.The onset for crystal 2 is less sharp and the profiles showed some asymmetry. Crystal 3 showed a weak maximum in the (5,0,1) reflection at 235 K and the profiles are very asym- metric. There was some indication of hysteresis of the profiles for crystals 2 and 3 during temperature cycling while those of crystal 1 were reproducible. Similar results54 DIACETYLENE POLYMER CRYSTALS to fig. 4 were observed for another difference reflection, the (7,2,1), for the three crystals. With crystal 2 on the diffractometer a study of the intensities of Friedel pairs was made. Fig. 5 is a comparison of typical profiles for the (5,0,1) and (5,O,T) at 185 K.It can be seen that the latter has much greater intensity with a narrower line-width. In fig. 6 data are plotted for the Bijvoet ratio A which is defined by the relation l3 A = 2(1+ - I - ) / ( I + + l - ) where I + and I - are the integrated intensities of the greater and less intense Friedel related reflections, respectively. The open (closed) circles in the upper part of the 9.5 10.0 10.5 11.0 11.5 12.0 Bragg angle/deg FIG. 5.-X-ray diffraction profiles for the (5,0,1) (lower points) and S,O,T) (upper points) Friedel pair related Bragg reflections at 185 K. The diffracted intensity has been plotted as a function of Bragg angle. Note that the scale for the (5,0,1) points has been expanded by a factor of four relative to those of the (S,O,T) points.figure refer to the (5,0,1)-(5,0,T) pair with decreasing - - - (increasing) temperature- Similar data have been obtained for the (7,2,1)-(7,2,1) pair. In the lower part of fig. 6, data are plotted for the (20,2,4)-(20,2,4) pair of Bragg reflections with de- creasing temperature which appear in both the high- and low-temperature phases. DISCUSSION Considerable variation in the temperature dependence of the profiles and inte- grated intensities of reflections at both high and low Bragg angles has been observed for different pTS crystals. Both large crystals and small crystals exhibit inconsistencies in their X-ray diffraction properties which suggest that the second-order phase transi- tion in pTS crystals is strongly affected by the presence of crystalline defects. Further- more, these defects need only be present in relatively low concentrations as their effect on the X-ray diffraction properties of the crystals at 300 K, far from the transi- tion temperature, was not readily apparent.WILLIAMS, BLOOR, BATCHELDER, HURSTHOUSE AND DANIELS 55 The sensitivity of the phase transition in pTS crystals to the presence of defects which is observed by X-ray diffraction should also be apparent in other properties.Thus far the various spectroscopic techniques which have been applied to the problem have not been sufficiently sensitive to show up any consistent differences among the variety of crystals studied. It is clear, however, that great caution must be used when applying order parameter analysis to properties of crystals which have not been characterized by X-ray diffraction.The extraordinarily large asymmetry between the diffracted intensities of various Friedel pairs as illustrated in fig. 5 could be due to one of two reasons. The Ren- ninger effect,14 whereby the diffracted beam from one set of planes is the incident beam 0 0 0 0 a "1 A A A . A -201 ' I I I I I I I I ' I I I I I I 120 110 160 180 200 220 210 260 280 temperature/K FIG. 6.-Temperature dependence of the Bijvoet ratio, A = 2(1+ - l - ) / ( I + + I-), where I + and I - are the integrated intensities of the greater and less intense Friedel related reflections, respectively. The open (closed) circles in the upper part of the figure refer to the (3,0,1)-(5,0,1) pair with de- creasing (increasing) temperature.Above 220 K the reflections were so weak that the data points do not have too much significance. In the lower part of the figure the triangles refer to the (20,2,4) (%,2,4) pair of Bragg reflections which appear in both the high- and low-temperature phases. for another, can be ruled out as a possible explanation since the scattering geometry was not appropriate. Thus the intensity asymmetry must have been caused by anomalous scattering which occurs in non-centrosymmetric crystals when absorp- tion by one or more atoms in the unit cell causes a phase shift in the X-rays scattered by those atoms. For most Bragg reflections inpTS crystals this effect should be very small as the 1.54 A Cu Koc radiation is very far from the 5.02 K absorption edge of56 DIACETYLENE POLYMER CRYSTALS sulphur, the heaviest atom in the unit cell.For difference reflections like the (5,0,1), however, the scattering amplitude becomes nearly zero in the vicinity of the phase transition and the effects of anomalous dispersion are greatly magnified. From currently available data it is impossible to tell whether this lack of a centre of sym- metry is an intrinsic property of pTS crystals or associated with the presence of de- fects in crystal 2. The pyroelectric effect observed inpTS crystals must be closely associated with the loss of the centre of symmetry. A possible origin for the required temperature de- pendent dipole moment can be found in the A-B stack model of the phase transition described above. If a single polymer chain has the stack of side-groups on one side in the A configuration and the stack on the other side in B then this A-B chain: (a) no longer has a centre of symmetry and (b) has a net dipole moment.The latter occurs since the dipole moments of the strongly polar sulphonyl groups in the A and B stacks are no longer anti-parallel. Even with A-B polymer chains the crystal would still be centrosymmetric if there were a corresponding number of B-A chains with dipole moments in the opposite direction. The model suggests that defects must be present which cause an imbalance between the number of A-B and B-A chains. Such a defect could conceivably be a screw dislocation since it has directional sense. Monomer crystals studied by etching reveal emergent screw dislocations parallel to the polymer axis.Often these are sufficiently numerous to render the end-facets non-planar due to the high density of growth spirals. A few crystals, however, show specular end facets indicating a much lower dislocation density.16 The peak in the Bijvoet ratio A in fig. 6 occurs over roughly the same temperature range in which the pyroelectric constant changes sign.7 At lower temperatures the mean value of A for the (5,0,1) reflection is about 0.13 suggesting that the crystal was non-centrosymmetric even away from the transition region. This would be in agreement with the pyroelectric data but the X-ray measurements need to be repeated on crystals where absorption corrections will be easier to make in order to confirm this result. The same conclusion holds for the (20,2,4) reflection which has a mean value for A of 0.13 between 120 and 300 K.At very low temperatures the X-ray data show that pTS crystals consist of a regular array of A-A and B-B polymer chains, that is the chains have a centre of symmetry with stacks of either type A or type B side-groups on both sides of the chain. The loss of order which occurs with increasing temperature could be due to the random formation of A-B and B-A chains. If defects with a directional sense were present then the numbers of A-B and B-A chains might be unequal and the crystal would no longer appear centrosymmetric. Well above the transition the X-ray structural data could be interpreted as arising from a random array of A-A, A-B, B-A, and B-B chains. This qualitative model is of assistance in the understanding of many of the effects which have been observed in pyroelectric, spectroscopic and X-ray diffraction investigations of the second-order phase transition in pTS crystals.Considerable further X-ray diffraction studies will be required to determine the role of defects in the phase transition and the origin of the unusually large asymmetry in the intensities of Bragg reflections from Friedel pairs. This research was supported by grants from the S.R.C., the National Science Foundation and the University of Delaware. The authors are grateful to Mr. D. J. Ando for the preparation of the polymer specimens and to Mr. S. Mehta for assistance with data processing.WILLIAMS, BLOOR, BATCHELDER, HURSTHOUSE AND DANIELS 57 D. Kobelt and E. F. Paulus, Acta Cryst. B, 1974, 30, 232. V. Enklemann and G. Wegner, Makromol. Chem., 1977,178, 635. V. Enklemann, Acta Cryst. B, 1977, 33, 2842. B. Reimer, H. Bassler, and T. Debaerdemaker, Chem. Phys. Letters, 1976, 43, 85. R. Clarke, personal communication. D. Bloor, D. A. Fisher, D. N. Batchelder, R. Kennedy, A. C. Cottle and W. F. Lewis, Mol. Cryst. Liquid Cryst., 1979, 52, 83. H. Kiess and R. Clarke, Phys. Stat. Solidi a, 1978, 49, 133. G. C. Stevens, D. J. Ando, D. Bloor, and J. S. Ghotra, Polymer, 1976, 17, 623. G. C. Stevens and D. Bloor, J. Polymer Sci., Polymer Phys. Ed., 1975, 13, 241 1. lo D. N. Batchelder, J . Polymer Sci., Polymer. Phys. Ed., 1976, 14, 1235. D. N. Batchelder and R. 0. Simmons, J. Appl. Phys., 1965, 36, 2864. l2 0. G. Peterson, D. N. Batchelder and R. 0. Simmons, Phil. Mag., 1965, 12, 1193. l3 S. Parthasarathy, Acta Cryst., 1967, 22, 98. l4 M. Renninger, Z. Phys., 1937, 106, 141. j5 A. F. Peerdeman, Anomalous Scattering, ed. S . Ramaseshan and S. C. Abrahams (International Union of Crystallography, Munksgaard International Publishers Ltd, Copenhagen, 1973, p. 3. l6 D. Bloor, J. Mat. Sci., 1979, 14, 248.