J O U R N A L O F C H E M I S T R Y Materials X-ray and optical studies of the tilted phases of materials exhibiting antiferroelectric, ferrielectric and ferroelectric mesophases Joanne T. Mills,a Helen F. Gleeson,*a John W. Goodby,b Michael Hird,b Alexander Seedb and Peter Styringb aThe Department of Physics and Astronomy, The University of Manchester, Manchester, UK M13 9PL bThe School of Chemistry, The University of Hull, Hull, UK HU6 7RJ Received 20th July 1998, Accepted 2nd September 1998 X-Ray and optical techniques have been employed to probe the physical properties of six materials that exhibit various frustrated chiral smectic phases.The temperature dependent evolution of the layer spacing, measured by small angle X-ray diVraction, is discussed with respect to the molecular structures of the materials.The layer spacings are used to deduce the steric tilt angle of the systems across the ferro-, ferri- and antiferro-electric phase ranges. Measurements of the steric and saturated optical tilt of the phases are compared. In general, for most materials and at most temperatures, the steric tilt is lower than the X-ray tilt as would be expected.However, in three of the materials the X-ray tilt is higher than the optical tilt over part of the high temperature tilted phase regime, providing evidence of conformation driven inversion phenomena. Further, the ratio of the steric to optical tilt angle is strongly temperature dependent in all but two of the materials. Introduction The occurrence of antiferroelectric and ferrielectric phases in liquid crystals is well documented,1 though details of the structures of these phases are still the subject of some debate.The currently accepted structures of the ferroelectric, ferrielectric and antiferroelectric phases have been described in detail elsewhere.1 Briefly, all the phases include layers in which the director is tilted with respect to the layer normal at a temperature dependent angle.In the ferroelectric phase, all the layers tilt in the same direction, the antiferroelectric structure has alternating layers tilted in opposite directions, while the ferrielectric phases exhibit some proportion of layers tilted in either direction. The existence of antiferroelectricty and ferrielectricity in liquid crystals provides a challenge to both experimentalists and theoreticians who continue to carry out elegant work both deducing the structures and providing theories for their stabilisation.Many of the studies of antiferroand ferri-electric systems are on the well known material MHPOBC2 and a variety of experimental techniques have Fig. 1 The structures of the materials studied. been employed in the task of deducing the structures of the phases exhibited.It is clear that the most useful information is provided when several complementary techniques are used AS661, AS656, AS620 and AS657 all have the same high to probe the physical properties of the complex phases and temperature phase sequence with the SmC* phase occurring subphases that can occur in antiferroelectric liquid crystals.directly below the SmA phase. In AS618 there is a SmC*a This paper reports the phase behaviour of six new antiferroelec- phase intervening between the SmA phase and the SmC* tric materials and describes the behaviour of their tilt as a phase. AS666 lacks a SmC* phase exhibiting a direct phase function of temperature. Two diVerent techniques are transition from the SmA to a SmC*c phase.AS620 also employed to deduce the tilt of the systems. Both the optical exhibits a monotropic phase transition to a SmI*A phase at tilt and the tilt deduced from X-ray layer spacings are reported 42 °C which is not shown in the Table as it is well below the and compared. The results are interpreted in terms of the temperature range of the measurements reported. conformational structures of the molecules.The phase behaviour of materials that exhibit ferro-, ferriand antiferro-electricity is complex and there is still much debate about the subphases that appear in them. The diYculty Materials of identifying various of the mesophases that occur is reviewed by Itoh et al.4 who describe how misidentification occurs as a The molecular structures and phase transitions of the materials studied are presented in Fig. 1 and Table 1, respectively.The result of supercooling, phase sequences and surface interactions. For example, the phase sequences reported for AS573 materials were synthesised at Hull University and their synthesis is reported elsewhere.3 Fig. 1 shows that the materials (the opposite enantiomer of AS661)5–7 are reported to be diVerent according to the techniques employed to study the studied are structurally similar, including identical terminal chains and chiral centres.The diVerences between the materials temperature dependent material properties, though it seems likely that at least two ferrielectric phases and possibly two involve substitution on the ring systems. Table 1 shows that J.Mater. Chem., 1998, 8(11), 2385–2390 2385Table 1 Phase transition temperaturesa Transition temperature/ °C Material K-SmC*A SmC*A–SmC*c SmC*c–SmC* SmC*–SmC*a SmC*a–SmA SmC*–SmA SmA–I AS618 72.9 99.9 103.5 117.0 122.2 129.3 AS661 53.3 78.3 82.0 — — 90.7 105.7 AS666 39.6 108.4 — — — 118.6 126.7 (c–A) AS656 52.8 94.0 95.2 – – 99.5 110.8 AS620b 67.7 97.8 99.0 — — 109.4 116.6 AS657 46.3 79.7 83.3 — — 84.3 93.7 aNo distinction is made between the SmC*b and SmC* phases.The SmC*c phase encompasses possible subdivisions into other ferrielectric phases. bAS620 also exhibits SmI*A and SmI* phase transitions at 42.2 and 33.3 °C (not shown). antiferroelectric phases exist in the system. The complex phase thin Mylar windows to allow the passage of incident and scattered X-rays with minimal attenuation.The thickness of behaviour of AS661 and AS573 is the subject of a further publication.8 In this paper, no distinction is made between the the liquid crystal sample was approximately 1 mm. The layer spacing in the smectic mesophases was deduced from the SmC* phase and the SmC*b phase, nor is any subdivision made of the SmC*c phase into other ferrielectric subphases position of the first order Bragg scattering peak with an accuracy of 0.5%, which for a layer spacing of 35 A° equates since the techniques employed for the work reported here do not allow these subphases to be distinguished.to an uncertainty of ±0.2 A° . The steric tilt angle d was deduced from the layer spacing measurements using the equation cos d=d/l where d is the smectic layer spacing and l is Experimental the molecular length.The way in which the molecular length was determined is discussed later. The transition temperatures of the materials were determined by optical microscopy and diVerential scanning calorimetry to within ±0.2 °C. In the optical experiments the sample was Results and discussion held on a Linkam THMS 600 hot stage and the temperature was maintained with a relative accuracy of ±0.1 °C using a Layer spacing measurements Linkam TMS 91 control unit.In the X-ray measurements a The layer spacings of the materials are shown across their specially modified Linkam hot stage DSC system was used to mesophase ranges as a function of reduced temperature in control the sample temperature9 again with a relative accuracy Fig. 2. The phase transition from the orthogonal to tilted of ±0.1 °C. smectic phases is clear on the figures, occurring at the point The liquid crystal samples were held in commercially produced10 devices of nominal thickness 5 mm for the optical tilt angle measurements. The inner surfaces of the devices had been treated with rubbed polyimide to promote antiparallel alignment and the devices included transparent indium tin oxide electrodes to allow the application of electric fields.Alternating fields of up to 40 V mm-1 were produced across the device from a signal generator connected to a wide band amplifier constructed in-house. The optical tilt angles of the materials were determined with an accuracy of ±0.5° across the mesophase range by observation of the extinction angles of the samples when viewed between crossed polarisers.These experiments were performed using white light on an Olympus polarising microscope. The optical tilt angle of the materials are field dependent due to both the chiral and antiferro- or ferri-electric nature of the materials. Thus, suYciently large fields (±10 V mm-1) were applied during the measurements to the samples to ensure that the tilt angle value was saturated.Observation of the samples via polarising microscopy during the optical tilt angle measurements ensured that the high fields necessary to ensure saturation did not have the adverse eVect of distorting the alignment of the sample, a well known phenomenon that would result in erroneous tilt angle measurements.X-Ray diVraction was employed to probe the temperature dependent layer spacing of the materials. The small angle Xray scattering (SAXS) experiments were carried out on station 8.2 of the Synchrotron Radiation Source (SRS), Daresbury, UK. The apparatus has been described in detail elsewhere11 and includes facilities to perform concurrent DSC and SAXS, allowing the smectic to isotropic phase transition temperatures of the materials to be determined in situ with an accuracy of ±0.5 °C.The X-ray camera was 1.0 m in length, was equipped with an area detector and X-rays of wavelength of 1.54 A° Fig. 2 The layer spacings of the six materials studied as a function of were incident on the sample. The measurements were made reduced temperature (the temperature below the tilted to orthogonal SmA phase transition).on unaligned samples which were held in DSC pans fitted with 2386 J. Mater. Chem., 1998, 8(11), 2385–2390Table 2 The molecular lengths of the six materials studied, deduced where the layer spacing reduces dramatically. The temperatures from the layer spacing in the SmA phase at which this marked change in layer spacing occurs is consistent with the SmA to SmC* (or SmC*a) phase transition Material Molecular length/A° temperatures given in Table 1.In all six of the materials studied, the layer spacing in the AS 618 38.6±0.1 AS661 39.3±0.1 SmA phase increases slightly as the temperature is reduced. AS666 37.4±0.1 This observation implies that contraction of the layers does AS656 38.4±0.1 not occur as the temperature decreases, a phenomenon that is AS620 37.9±0.1 common in other SmA systems.12–14 Several eVects could result AS657 37.5±0.1 in a slightly increasing layer spacing with reducing temperature, including changes in the conformation of the alkyl chain or a changing population of the conformers occurring as function making no assumptions about packing or thermal eVects of temperature, a suggestion that is supported by further within a mesophase, but can give erroneous results as molecuevidence discussed in later sections of this paper.lar conformations other than those that actually occur in the Comparing the temperature range of the SmC*A phase in mesophase may be modelled. Molecular modelling was underthe six materials studied materials to the smectic layer spacings taken using Cerius2 on a silicon graphics workstation for one in the antiferroelectric phase leads to the conclusion that a of the materials reported here, AS661.The modelling yielded wider SmC*A range correlates with a smaller layer spacing. a molecular length of 39.3 A° (measured from tip to tip of the This phenomenon has also been observed by Ikeda et al.15 molecule). This value is only 0.1 A° diVerent from that deterand supports the theory that the molecular pairing believed to mined from the layer spacing, implying that the layers are well occur between adjacent layers in the SmC*A phase1 plays an defined and that there is little or no interpenetration of layers important role in its stabilisation.A greater degree of pairing in the SmA phase of AS661.Further, it seems that the implies a stronger antiferroelectric attraction between adjacent molecules are in an almost completely extended configuration layers, the stronger attraction resulting in a shorter the layer in the SmA phase. Given the structural similarities of the spacing. Thus a more stable SmC*A phase would be expected materials studied and their molecular lengths (Table 2), it is to have a shorter layer spacing.likely that the remainder of the materials behave in a similar Below the orthogonal to tilted phase transition, the layer manner. spacings initially decrease rapidly, then change little with The molecular lengths given in Table 2 are approximately temperature as the tilt angles saturate.It can also be seen that the same for all of the systems studied. This is not surprising at low temperatures, within the SmC*A phases, the spacings as all the materials have a C12 alkyl chain on one end and a of all the materials increase as the temperature is reduced, an C6 alkyl chain on the other with the same number of atoms eVect that is very marked in AS666. The general trend of layer across the length of the core.The two materials containing Se spacing as a function of temperature is similar to that reported atoms, AS620 and AS657, have shorter lengths than all of the by Rao et al. for an antiferroelectric compound with quite others apart from AS666. The selenophene group has been diVerent terminal groups.16 They attributed the increase of shown previously17 to promote a bend in the molecule (18° in layer spacing at lower temperatures in the SmC*A phase to an the core of AS620 relative to a structure containing biphenyl underlying SmI*A phase.The only compound for which that rings) so the shorter molecular lengths AS620 and AS657 are is possible here is AS620 as none of the other materials studied not unexpected. have been observed to exhibit underlying hexagonal phases.All of the compounds have an ester linkage between the Further, the increase in layer spacing in the SmC*A phase aromatic rings which will induce some bend into the core. The occurs well above the temperature associated with the mono- shortest molecular length is observed for AS666, possibly tropic phase transitions to the hexagonal smectic phases in because of an increased molecular bend caused by repulsion AS620. The observed increase in the layer spacing is not between the fluorine atom and the end ester group near the reflected in the optical tilt angles, as is shown in a later section, chiral centre.In AS661, the longest molecule, the fluorine and so cannot be attributed to changes in the director tilt atom is on the opposite side of the molecule to the ethyl group angle.The low temperature increase in layer spacing continues so this phenomenon does not occur. Rather, the inward the trend observed in the SmA phase and is likely to be due pointing F atom may repel the ester dipole, straightening the to increasingly restricted conformational structures and core to some extent and resulting in the longest molecular hindered rotation of the molecules as the temperature reduces. length.Tilt angle measurements The molecular length The steric tilt angle d can be deduced from the layer spacing Fig. 3(a) to (f ) show the optical and steric tilt angles of the materials AS618, AS661, AS666, AS656, AS620 and AS657, data using the equation cosd=d/l, where d is the layer spacing and l is the molecular length.In order to employ this technique respectively. It should be noted that the uncertainty in temperature associated with the two diVerent measurement techniques of deducing the steric tilt angle, it is clearly necessary to have a measure of the molecular length l for the materials studied. could translate into oVsets of the order of a degree between the data sets plotted on each graph.For all the materials the The simplest method of deducing a value for this parameter is to assume that the layer spacing in the SmA phase is optical tilt angle h was greater than the steric tilt angle d over the majority of the phase range, implying that the molecular identical to the molecular length, though the validity of such an assumption clearly depends on whether or not the layers cores are more tilted than the terminal alkyl chains.This observation is in keeping with the Wulf model.18 There is are intercalated. It also neglects the temperature dependence of the layer spacing reported in the previous section and evidence in three of the materials that d is larger than h over part of the high temperature tilted phase region, a phenomenon assumes that the molecules are in their most extended form in the SmA phase, an assumption which is rarely valid.Table 2 that is discussed in more detail later. The maximum values of the steric and optical tilt angles shows the molecular length for each of the materials studied, considered to be the layer spacing at the point at which the attained in the materials are compared in Table 3.The optical tilt angles are all relatively large (around 30°). The two orthogonal to tilted phase transition takes place. An alternative method of deducing the molecular length materials containing Se (AS620 and AS657) have significantly lower steric tilt angles than the other compounds. These relies on molecular modelling, which has the advantage of J.Mater. Chem., 1998, 8(11), 2385–2390 2387Fig. 3 The steric and optical tilt angles of (a) AS618, (b) AS661, (c) AS666, (d) AS656, (e) AS620 and (f ) AS657 as a function of reduced temperature. materials also had amongst the shortest molecular lengths, though the small steric tilt angles cannot be attributed to that factor as it is taken account of in the calculation of d.Although Table 3 A summary of the saturated values of the optical tilt angle h and the steric tilt angle d AS661 is a shorter molecule than AS620 and AS657, it has a larger steric tilt than either, and while d is still lower than that Saturated steric Saturated optical of the other three materials, the optical tilt of AS661 is also Material tilt angle d (°) tilt angle h (°) low.It is possible that the small steric tilts of AS620 and AS657 are because the electron dense selenium atom in these AS618 21.4 33.0 AS661 20.0 27.7 systems biases the X-ray tilt to lower values, which could AS666 21.1 36.1 happen if, on average, it remained on the inside of the cone. AS656 22.3 31.4 Such an eVect may be the result of hindered rotation about AS620 18.3 30.6 the molecular long axis.Alternatively, packing constraints in AS657 16.8 31.1 the mesophases, imposed due to the molecular bend that is 2388 J. Mater. Chem., 1998, 8(11), 2385–2390ratio d/h that as both d and h change rapidly directly below the SmA phase transition, the uncertainty in the ratio is greatest in this region. As mentioned previously, the uncertainty occurs primarily because of the diYculty in registering the absolute temperature measurements in the two diVerent Fig. 4 A schematic diagram of conformational changes in zig-zag experiments. Consequently, the ratios calculated within 2 °C molecules that result in inversion phenomena. of the phase transition are discarded and not shown in Fig. 5. In spite of this precaution, it is recognised that the data of known to occur these molecules, could equivalently cause the Fig. 5 are least reliable in the vicinity of the tilted to orthogonal axis of electron density to appear tilted at a lower angle than phase transition.would occur for unbent molecules. Rieker et al. report that the ratio of d/h for a ferroelectric The data for AS661 [Fig. 3(b)] clearly show a change from material (not containing antiferroelectric subphases) is almost d>h to h>d in the middle of the temperature regime identified temperature independent and that d/h~0.85, in common with as a SmC* phase.There is also some evidence of a similar many ferroelectric materials.22 The factor of 0.85 is of course eVect in the data for AS618, close to the tilted to orthogonal material dependent and is determined by the relative orienphase transition.Such an eVect is consistent with a confor- tations of axes of the electron density and polarisability in the mational change (inversion) occurring in the zig-zag shaped system. For the materials described here, it is clear that the molecules of the sort depicted in Fig. 4.19–21 Optical obser- ratio can fall within a number of diVerent values, as may be vations made of the pitch of this system show inversion expected from the diVerent degrees of molecular bend in the phenomena and are reported elsewhere.8 various systems, together with the inclusion of electron dense It is apparent from Fig. 3(a) to (f ) that the temperature selenium atoms in two of the compounds. The ratio d/h is dependence of h and d appears to be diVerent for some of the almost temperature independent for two of the materials, materials.In order to investigate the eVect further for all of AS666 and AS620, and takes values of 0.64 and 0.6, respectthe materials the data were replotted to show the variation of ively. These materials are therefore considered to behave as the ratio d/h with respect to reduced temperature, as is shown would be expected for ferroelectric systems.The ratio exhibits a strong temperature dependence for the three of the materials in Fig. 5. It is worth noting when examining the data for the Fig. 5 The ratio of the steric (d) to optical (h) tilt angle for the six materials studied. Note that the scales on the graphs corresponding to the value d/h is identical for all the materials apart from AS661, where it is significantly diVerent.The temperature scales are identical for all of the graphs, with zero reduced temperature at the orthogonal to tilted phase transition. J. Mater. Chem., 1998, 8(11), 2385–2390 2389studied, AS618, AS661 and AS656, passing through 1 for the The authors gratefully acknowledge the support of the EPSRC (through grant number GR/L/76648) and the DERA Displays latter two.Most of the temperature dependence occurs close Group, Malvern. Thanks are also due to Dr B. Komanschek to the tilted to orthogonal transition, and is most likely to be of Station 8.2 at Daresbury Laboratories, UK, for his help attributed primarily to the conformational change depicted in during the XRD work.Fig. 4. However, the ratio continues to be temperature dependent well below this transition, in common with the behaviour of the ratio for AS657. We attribute such behaviour to factors References other than the zig-zag conformational change, including 1 See for example: A. Fukuda, Y. Takanishi, T. Isozaki, K. Ishikawa increasingly restricted conformational structures and hindered and H.Takezoe, J. Mater Chem., 1994, 4, 997; A. D. L. Chandani, rotation of the molecules, as was discussed earlier. E. Gorecka, Y. Ouchi, H. Takezoe and A. Fukuda, Jpn. J. Appl. Phys., 1989, 28 L1265; A. Ikeda, Y. Takanishi, H. Takezoe and A. Fukuda, Jpn J. Appl Phys., 1993, 32 L97. 2 A. D. L. Chandani, Y. Ouchi, H. Takezoe and A. Fukuda, Jpn. Conclusions J. Appl. Phys., 1989, 28, L1261. 3 M. Hird, P. Styring, A. Seed, H. F. Gleeson and J. T. Mills, This paper presents layer spacing and tilt angle measurements unpublished work. for six diVerent materials that exhibit ferroelectric and anti- 4 K. Itoh, M. Kabe, K. Miyachi, Y. Takanishi, K. Ishikawa, ferroelectric phases and subphases. The tilt angles were meas- H. Takezoe and A. Fukuda, J Mater. Chem., 1997, 7, 407.ured both optically and deduced from X-ray layer spacing 5 J. W. Goodby and I. Nishiyama, unpublished DSC data. measurements. Several features are common to all the systems 6 W. K. Robinson, PhD Thesis, Manchester University, studied. Firstly, the layer spacing measurements in the ortho- Manchester, UK, 1995. 7 Y. Panarin, W. Kalinovskaya, J. K. Vij and J. W. Goodby, Phys gonal phases imply that the molecules are considerably bent, Rev.E, 1997, 55, 4345. in common with other systems that exhibit antiferroelectricity. 8 L. Baylis, unpublished work. There is evidence that the molecules are almost completely 9 W. Bras, G. E. Derbyshire, A. J. Ryan, J. Cooke, A. Devine, extended in the SmA phase. The layer spacing changes rapidly B. E. Komanschek and S.M. Clark J. Appl. Crystallogr., 1995, with temperature at the transition to the tilted phases, as 28, 26. 10 Lucid EEV, 106, Waterhouse Lane, Chelmsord, Essex, UK CM1 would be expected. The layer spacing reaches a minimum 2QU. value, then rises again as the temperature is reduced into the 11 W. Bras, G. E. Derbyshire, A. J. Ryan, G. R. Mant, A. Felton, antiferroelectric phase.Previous work has attributed such R. Lewis, C. Hall and N. Greaves, NIM, 1993, 587, A326. behaviour to underlying SmI*A and SmI* phases, though this 12 J. StamatoV, P. E. Cladis, D. Guillion, M. C. Cross, T. Bilash and cannot be the case here, since only one of the materials studied P. Finn, Phys. Rev. Lett., 1980, 44, 1509. exhibits such a sequence and none of the materials shows any 13 Y.Ouchi, Y. Takanishi, H. Takezoe and A. Fukuda, Jpn. J. Appl. Phys., 1989, 28, 2547. reduction in the optical tilt angle of the system in the range 14 A. S. Morse and H. F. Gleeson, Liq. Cryst., 1997, 23, 531. in question. The result must therefore be interpreted as being 15 A. Ikeda, Y. Takanishi, H. Takezoe and A. Fukuda, Jpn. J. Appl. due to increasingly restricted conformational structures and Phys., 1993, 32, L97.hindered rotation of the molecules as the temperature reduces. 16 D. S. S. Rao, S. K. Prasad, S. Chandrasekhar, S. Mery and Comparison of the optical and steric tilt angles shows R. Shashidhar, Mol. Cryst. Liq. Cryst., 1997, 292, 301. 17 J. T. Mills, R. J. Miller, H. F. Gleeson, A. J. Seed, M. Hird and convincing evidence of inversion behaviour occurring in the P. Styring, Mol. Cryst. Liq. Cryst., 1997, 303 145. regions corresponding to the SmC* phase in three of the 18 A. Wulf, Phys. Rev. A, 1975, 11, 365. materials. Such behaviour has been confirmed by independent 19 J. S. Patel and J. W. Goodby, Philos. Magazine Lett., 1987, 55, measurements of the pitch for AS661. It should be noted that 283. the phase behaviour of these materials, in particular AS661, 20 J. S. Patel and J. W. Goodby, J. Phys. Chem., 1987, 91, 5838. is the subject of considerable debate. Certainly inversion 21 J.-H. Kim, S.-D. Lee, J. S. Patel and J. W. Goodby, Mol. Cryst. Liq. Cryst., 1994, 247, 293. behaviour, in addition to the other factors mentioned pre- 22 T. P. Rieker, N. A. Clark, G. S. Smith, D. S. Parmar, E. B. Sirota viously, further complicates the phase identification process. and C. R. Safinya, Phys. Rev. Lett., 1987, 59, 2658. In particular, inversion behaviour will be important in many 23 T. Sako, Y. Kimura, R. Hayakawa, N. Okabe and Y. Suzuki, Jpn. of the measurements currently employed to distinguish the J. Appl. Phys., 1996, 35, L114. SmC*b phase from the SmC* phase as it will aVect 24 H. F. Gleeson, unpublished work. electroclinic23 and dielectric relaxation phenomena.7 This phenomenon will be the subject of a future publication.24 Paper 8/05611K 2390 J. Mater. Chem., 1998, 8(11), 2385–2390