347J. MATER. CHEM., 1991, 1(3), 347-356 Vitrification in Low-molecular-weight Mesogenic Compounds Wolfgang Wedler,” Dietrich Dernus/ Horst Zaschke,a Kristina Mohr,” Wolfgang Schaferb and Wolfgang Weissf Iog “Martin-Luther-Universitat HaIle- Wittenberg, Sektion Chemie, 0-4020 HaIle (Saale), Weinbergweg 16, Germany Werk fur Fernsehelektronik, 0-1160 Berlin, Ostendstr. 1-5, Germany “Spezia Ich em ie Leipzig, 0- 7030 Leipzig, Elstera ue 9, Germany Systematic investigations have allowed us to formulate a structural model representing the relationship between vitrification and mesogenity in low-molecular-weight liquid-crystalline compounds. The tendency to transform the highly viscous mesophases into glass phases is caused mainly by the bulkiness of the molecules.The freezing of such mesophases, oriented by external influences at high temperatures, can be used for information storage. Keywords: Liquid crystal; Vitrification; Differential scanning calorimetry; Viscosity According to the well known Tammann rule,’ all substances length :breath ratios. The common feature of all these mol- in nature can be transformed into the glassy state below their ecules is their unconventional structure. l l melting temperatures. However in most cases, the glassy state In the following, we present a survey of some important cannot be reached, because a strong tendency for crystalliz- unconventionally constructed mesogenic compounds with ation at temperatures between the melting point and the glass- strong glass-forming tendencies.transition temperature, Tg, exists. The glassy state in nematic, smectic and cholesteric liquid- One-string Compounds crystalline phases can be useful for the construction of infor- (1) Lateral branches: mation-storing materials. Information can be transferred to ref. 12 the mesophase at temperatures above Tg and can be stored in the glassy state for a long Though glass-transition phenomena seemed to be a privilege of polymeric liquid crystals, they have also been observed in some low-molecular- ref. 13weight mesophases.6-8 In most of these mesophases the glass-transition temperature was much lower than room temperature. ‘c0,-x 0..Kuhrmann’ and Baentsch’ synthesized and described liquid-crystalline low-molecular-weight substances, which ref.3showed glass-transformation behaviour. Based on their exper- imental results and on assumptions of the free-volume model of glass-forming liquids, we have developed a generalized structural model. This model allows us to find a compromise between the mesogenity of the substances on the one hand, and the glass-forming tendency on the other. The aim is to create enantiotropic, mostly calamitic, liquid-crystalline sub- (2) Terminal branches: stances with glass transitions above room temperature. ref. 14 The calorimetric and viscosimetric measurements reported here enable the systematic investigation of the relationship ‘swallow-tailed’ compounds between the glass-forming tendency and molecular structure. (b) Ro~C;c~~~~2C~co~~c~=c,c~*~,C02R’ ref.14 RO,C Materials ‘bi-swallow-tailed’ compounds Molecular Structure (c) Laterally and terminally branched, non-symmetric com- The substances of ref. 9 and 10 show a significant deviation pounds: from the rod-like shape of mesogenic molecules. They are branched in the centre of the molecular unit by condensed ring systems and are ‘bulky’ in shape, e.g. Twin Structure (1) ‘Siamese twin’ molecules: lS ref. 10 cr 142.2 N 207.0 IS (see Tables 1-9, later) “;px (x=O, 1, 2) In the recent years other types of mesogenic compounds have been synthesized, which have molecules with low 348 J. MATER. CHEM., 1991, VOL. 1 (2) Twin molecules with flexible bridge groups:' (C) Synthesis of compound /C02-CH2-@02 'CO2-R'Q /CO2-iy The elongated shape of the molecules, which guarantees the existence of the niesophases, can be disturbed by several structural fragments.In almost all cases the mesophase stab- ility is weakened, whereas the tendency of vitrification is strengthened. Synthesis The synthesis of one-string compounds type l(a), l(b), 2(a) and 2(b) as well as of the twin-structured mesogens has been discussed in detail in the references given. To elucidate the synthesis routes of the cases l(c), 3 and of some special compounds of type l(b) (see Table 2 later) we will give examples. (A) Synthesis of compound C3Hf19C029c02a Br CH,O 4-Propyloxy-3-bromobenzoic acid was esterified with 4-hydroxy-3-methoxybenzaldehyde(vanillin) according to the method of Einhorn.The obtained 4-(4-propyloxy-3-bromo) benzoyloxy-3-methoxybenzaldehydewas oxidized by Cr03 and concentrated acetic acid. This yielded 4-(4-propyloxy-3- bromo)benzoyloxy-3-methoxybenzoic acid, which was esterified again with 2-naphthole according to the Einhorn method. Repeated purification in mixtures of ethanol and toluene gave a compound with the following phase sequence (temperatures in "C): cr 151 (N 87) IS; TfN=31 "C (symbols and abbreviations see at the beginning of Tables 1-9). (B) Synthesis of compound The first step of the synthesis was the (4:2) cycloaddition (Diels-Alder reaction) of 2,3-dimethylbuta- lY3-diene with p-benzoquinone. The Diels-Alder adduct was isolated in high yield.The same good result was obtained by using other dienes, such as cyclopentadiene, cyclohexa- 1,3-diene or buta- 1,3-diene. In the following reaction the quinone adduct was transformed by isomerization with hydrochloric acid in etha- nol solution to 5,8-dihydro-6,7-dimethyl-1,4-dihydroxynaph-thalene. This hydroquinone derivative was acylated after the method of Einhorn with 4-ethyloxybenzoyl chloride at room temperature. For purification the prepared ester was recrys- tallized from ethanol to yield light crystals with the following phase sequences: cr 215 N 241.6 IS; T:N=52 "C (see also Table 3 later). The substance was prepared in two reaction steps. The esterification of 1,4-dihydroxy-2-naphthoicacid with 4-nitrob- enzyl bromide was performed in dried acetone.The formed hydrogen bromide can be eliminated by triethylamine without saponification of the halogenide. The 4-nitrobenzyl 1,4-dihydroxy-2-naphthoate was recrystallized from dimethylfor- mamide; melting range 218-220 "C. This phenolic intermedi- ate was acylated with 4-n-propyloxybenzoyl chloride in pyridine following the known procedure described by Einhorn. The crude product was recrystallized from pentan-1-01. The melting behaviour is given in Table 2: cr 204 (N 100) IS; Ty =49 "C. Experimental Calorimetric Measurements Calorimetric measurements were carried out on a Perkin-Elmer DSC-2 device which was cooled constantly by a COz-ethanol mixture. Heating and cooling experiments were made in amounts of 5-8 mg, placed in aluminium capsules.We performed the calorimetric observation in three steps: (1) heating the crystalline material into the isotropic liquid phase (heating rate 20 K min-'; observation of melting and other phase-transition processes); (2) quenching the sample (cooling rate, 40 K min-';observation of phase-transition processes and of the glass-transition step); (3) heating the liquid-crystal- line glassy material again into the isotropic liquid phase (heating rate 20 K min-';observation of the transformation interval, of the crystallization and of subsequent phase tran- sitions). Sometimes the supercooled liquid crystallized during quenching in the calorimetric device. In these cases the samples were heated and quenched outside the calorimeter using a heating plate and a cooling medium (e.g.dry ice). Step 3 of the procedure was then continued. Viscosimetric Measurements We made measurements on a Rheotest-2 viscosimeter to gain information about the dynamic viscosity of pure and mixed glass-forming liquid-crystalline phases. The substance was placed between a fixed metallic plate and a rotating metallic cone, The plate could be thermostatted. The cone had a diameter of 36 mm, was driven by a motor and had rotation velocities of 81 and 243 rpm. This enabled shearing tensions between 7540 and 45 200 Pa to be reached and a velocity gradient between 1620 and 4860 s-'. An external magnetic or electric field to orient the director of the sample could not be used.So we assume that, owing to mechanical forces, the molecules were oriented along the shearing direction. As Schneider and Kneppe' demonstrated, the effective shear viscosity under the condition of stream orientation is close to the shear viscosity q1 defined by Mieso~icz'~ (director parallel to the vector of shearing velocity). Provided that the Leslie- Ericksen coefficients a2 and a3are both negative, then accord- ing to ref. 16 one can suppose that under the influence of the shearing tension the director is oriented in the shearing plane. Moreover, the equilibrium angle of stream orientation, O,, between the director and the vector of shearing velocity has a small value (5-1 5" for MBBA [N-(4-methoxybenzylidene)- J. MATER. CHEM., 1991,VOL.1 20 K min-'-/Icr Tm IS -40 TNI -----\ K min-' glassy nematic- N \I II " iliIS W 50 100 150 ' ' 200-91°C Fig. 1 DSC plot of 2-naphthyl-4-(4-cyanobenzoyloxy)-3-methoxy benzoate: showing a glassy phase with nematic structure I I -SC -NV IS glassy SC Fig. 2. DSC plot of 4,4-dimethyl-(4-n-octyloxybenzoyloxy)benzoyl-oxybenzylidene rnal~nate,'~ showing a glassy S, phase 4-butylaniline]). From this point of view, we suppose that our measured viscosities are near to the appropriate ql Miesowicz viscosity, in the best case. On the other hand, this assumption cannot be true if the coefficients a2 and a3 have opposite signs or if instabilities above a critical shearing tension appear. l6 Results Calorimetric Measurements The calorimetric plots of glass-forming mesogenic materials will be explained for a nematic and a smectic compound.Fig. 1 shows a typical three-step DSC-plot of a compound4 that creates a nematic glassy liquid-crystalline phase. The crystalline substance was heated with a heating rate of 20 K min-'. The melting point could be detected at a temperature of 189.8 "C, the corresponding phase-transition enthalpy had a value of 33.8 kJ mol-' (see first run). The second run was a cooling process of the first isotropic liquid phase. Using a cooling rate of 40K min-', an isotropic-nematic phase transition could be registered at 175.8 "C, having an enthalpy of 1.24 kJ mol -'. With further cooling no crystallization occurred, and the highly viscous, nematic substance was frozen into the glassy state at ca.40 "C. The glass-transition interval had a width of ca. 10 K. After this cooling process a new heating period again showed the transformation from the Table 1 Calorimetric investigation of the homologous series no. n cr N IS Tp/ "C 138.8(-11 1.1)141.1(. 136.4) 112.6(*106.2)99.6 * 109.0 93.5(-87.4) 81.0 92.0 27 22 13 5 -5 -8 78.3 82.5 -15 84.0 84.4 -16 no. n A,H/kJmol-' ANiHlkJ mol-' ACplJmol-' K-' 32.8' 1.36 181 39.3b 2.52 206 35.1' 2.07 224 35.2' 2.72 200 51.9" 2.22 197 38.2" 2.66 268 38.6" 2.43 220 45.5" 2.66 274 " First heating; 'second heating. cr =crystalline phase; Sc =smectic C phase; S,=smectic A phase; Sx=smectic phase, type unclear; N= nematic phase; IS =isotropic phase; TF =lower limit of the glass- transition interval (onset of glass-transition temperature); AmH= transition enthalpy of the melting process; ANiH =transition enthalpy of the transition nematic phase/isotropic phase; $NH =transition enthalpy of the transition smectic/nematic; ACp=jump of heat capacity caused by glass transition. nematic glassy phase to the supercooled, highly viscous nematic phase. This transformation can be recognised by a step in the curve, which corresponds to a C, variation of 150 J K-' mol-'.Further heating yields a slow, steady crystallization, which is characterized by an exothermic peak with a maximum at ca. 125 "C. The now crystalline material begins to melt at 181 "C.In Fig. 2 the three-step DSC plot for a smectic compo~nd'~ is depicted. All conditions are the same as in Fig. 1. This 'swallow-tailed' compound has a phase transition from the crystalline to the SA phase at ca. 104.4"C with a transition enthalpy of 37.4 kJ mol-'. The smectic phase transforms at 124 "C into a nematic phase (phase-transition enthalpy: 0.59 kJ mol-') and at 161 "C from the nematic into the isotropic phase (phase-transition enthalpy: 1.16 kJ mol-'). During the cooling process the phase sequence is encountered in reverse and at a temperature of 66°C the SA phase transforms into an Sc phase. The compound does not crys- tallize, if cooling is continued. Consequently, the highly viscous Sc phase transforms into the glassy state at CQ.0°C. Sub-sequent heating induces a step in the curve, beginning at -5 "C. The step has a height, corresponding to ca. 225 J K-' mol -'. Soon after forming the highly viscous, supercooled Sc phase from the glassy Sc phase, there is an intensive exother- mic effect at 20 "C indicating crystallization. The crystalline material melts at CQ. 94 "C,and then the phase sequence SA, N, isotropic liquid can be observed again. Tables 1-9 give a representative review of the phase behav- iour and the calorimetric data of some important classes of liquid-crys talline glass-forming substances (all temperatures in "C). With the aid of Table 7 the relationship between the glass- transition temperature and the clearing temperature can be 3 50 J. MATER.CHEM., 1991, VOL. 1 Table 2Calorimetric investigation of some laterally aromatically branched compounds no. compound cr N IS -204.0 100.0) c3H709c02~02c0m3H7 Co2-CH2+o2 213 'BHI 70~ c 0 2 ~ 0 2 c ~ 0 c 8 H7 I * 90.0 (-61.0) 91.0 C02~CH2)2~C(CH313 214 -113.0 96.5)- compound no. Ty/"C A,H/kJ mol-' AsXNH/kJmol-' ANiH/kJmol-' ACJJ mol-' K-' 14 34.0 49 47.6 3 44.9 -7 53.4 demonstrated. The elongation of the alkyl chain in this homologous series of initially isotropic 1 -naphthyl esters causes the creation of an anisotropic, probably nematic, phase. The clearing temperature rises with chain length, whereas the glass-transition temperature decreases. In this interpretation, growing alkyl chains stabilize the liquid-crystalline state but effect a decrease of the glass-transition temperature.Another phenomenon was observed for compound 8/1 (Table 8). This compound with a glass-transition temperature of 58 "C forms an isotropic glassy phase immediately after quenching. Observations by polarization microscopy over ca. 1.85 1.35 250 - 1.60 284 4.56 1.60 I46 - 1.56 234 3 days indicate molecular mobility in the glassy phase. Even at room temperature, ca. 25 K below the glass-transition interval, after 21 h beginnings of an anisotropic phase are visible (see Fig. 3), which grow and after ca. 2 days form a final picture that is compatible with a nematic phase (see Fig. 4).X-Ray investigations gave no indication of a crystalline phase.The isotropic-nematic phase transition which should occur at temperatures below the glass-transition temperature, in this compound is strongly supercooled. Kresse et d.'* have already demonstrated by dielectric relaxation measurements that ca. 30 K below the calorimetric glass-transition temperature the motions of the whole mol- Fig.3 Observation of the isotropic glassy phase of compound 8/1, 21 h after freezing. Small domains of an anisotropic phase are visible. Fig. 4.Final state of sample from Fig. 3 after 3 days. No covering No covering plates are used; 80-fold magnification, crossed polarizers plates; 80-fold magnification, crossed polarizers J. MATER. CHEM., 1991, VOL. 1 351 Table 3 Calorimetric investigation of mesogenic compounds with condensed ring systems at the centre of the molecules (4 no.n X cr N IS Ty/"C 3/la 2 215.0 -241.6 528H3C CH3 3l2a 3 * 172.2 -212.5 42-fJH3C CH3 3/3a 2 160.6 -172.8 23 no. ACJJ K-' mol-' A,H/kJ mol-' 3/la 121 41.6 4.20 3/2a 112 39.7 (was not measured) 3/3a 170 39.1 2.65 ref. 3, 10 X-no. X cr N IS Tr/"C 03/lb 192.4 * 200.4 38 3/2b * 150.0 -167.6 17 H3CeCH=N-3/3b 102.2 -146.6 .14 3/4b * 130.0 87.5) 23(a no. AC&J K-' mol-' A,H/kJ mol-' ANiH/kJmol-' 3/lb 155 41.4 1.34 3/2b 207 31.2 0.46 3/3b 180 44.1 0.58 3/4b 188 37.4 0.29 ecules are slowed down. This takes place according to a Viscosimetric Measurements dramatic drop in the free volume, available for mostly rotational motions around the long and the short molecular In addition to the calorimetric investigations we made viscosi- axes.The temperature behaviour of this relaxation mechanism metric measurements on pure compounds l/S-l/S (Table 1) follows the Vogel-Fulcher-Tammann-equation [see eqn. (1) and on mixtures. We were able to measure dynamic viscosities later]. Contrary to this, the motion of individual molecular only from ca. 150 mPa s (near to the clearing point) to ca. fragments, showing an Arrhenius-like behaviour, is unabatedly 1000 mPa s. Then, as a consequence of the mechanical forces, active. So we interprete the occurrence of such supercooling the emerging crystallization prevented further observation. effects in the glassy state as a microscopically visible expression In Fig.5 the log q us. T-' dependence is depicted. Clearing of molecular relaxation processes, which are still active in the temperatures can be recognized by a characteristic decrease glassy state. in the dynamic viscosity when the substances are cooled from J. MATER. CHEM., 1991, VOL. 1 Table 4 Calorimetric investigation of the homologous series no. n cr 411 1 412 2 413 3 414 4 no. ACJJ K-' mol-' 228 279 193 190 ref. 14 SC SA N IS Ty/"C 100.8( -69.0) 126.0 * 162.0 -5 104.0(*58.5) 110.0 144.0 -16 65.0( * 53.5) 99.0 130.0 -23 70.0(* 5 1 .O) -92.0 -115.0 -28 A,H/kJ mol-' AsANH/kJ mol-' ANiH/kJ moi- ' 37.4 0.59 1.16 48.6 0.31 0.82 45.9 0.33 0.88 38.3 0.41 0.88 Table 5 Calorimetric investigation of the homologous series no.n cr1 -176.0 * 195.0 * 102.3 * 80.0 * 81.4 -141.2 80.0 83.0 ACJJ K-' mol-' 208 252 251 26 1 274 239 222 218 cr2 SC N IS T:N/ "C -297.0 36 m232.0 2 a201.0 -5 112.5 94.5 (-60.0) (-60.0) ( * 74.2) ( -80.0) * 165.0-150.5 * 120.4 * 120.4 * 109.2 -15 -17 -16 -16 -16 A,H/kJ mol-' A,,,H/kJ mol-' ANiH/kJ mol-' 39.7 - 0.90" ? - 0.45' 45.0 - 0.35' 36.3 - 0.40b 19.4119.4 (2 peaks) 1.23 0.32 38.3 1.80 0.27 39.6 1.87 0.39 66.6 2.0 I 0.39 a Own measurement, decomposition probably in progress; 'values given in ref. 14. the isotropic into the nematic state. Furthermore, below the clearing temperatures in this Arrhenius plot a non-linear dependence occurs.This is characteristic behaviour of glass- forming liquids. In ref. 18 and 19 it was shown, that the Fig. 5. Arrhenius plot for temperature dependence of dynamic-vis- cosity for higher homologues in the series of 2-tert-butylhydroquinonebis(4-n-alkyloxy benzoates). Values of n: +, 5; 0,6; x ,7; 0,8 temperature dependence of relaxation processes is governed by an activation part and a free-volume part. At high tempera- tures the activation part dominates and at low temperatures the free-volume part dominates. We assume that according to the 'bulky' molecular shape the influence of the free volume dominates. Other influences are most effective in the vicinity of the clearing point.Consequently, we used the empirical Vogel-Fulcher-Tammann (VFT)equation,20, to fit our data. As Fig. 6 shows, this VFT model satisfactorily reflects the temperature dependence of the dynamic viscosity. Table 10 gives a s.ummary of all fitted and measured data of the four pure substances 1/5-1/8. Also in this table the temperature for a dynamic viscosity 9 = 10'' mPa s is given. According to an often cited rule2' this temperature should coincide with the calorimetric glass transition. Within the limits of error, the values of Table 10 are in good agreement, which also confirms the validity of the VFT model for our glass-forming nematic compounds. J. MATER. CHEM., 1991, VOL. 1 Table 6 Calorimetric investigation of several 2-naphthyl esters4 R R" R"' cr N ISno.R (aH CH30 -161.2 135.2) * 134.4 -218.8CH3 H H H 134.6 -250.0 204.6H H -177.2 H CH30 * 155.2 (* 89.4) H CH30 * 143.4 H Br * 136.8 * 185.2 H Br 102.6 124.0 H H 174.4 -308.0 H CH30 * 189.8 ( * 176.0) Ty/ "C ACp/J K-' mol-' A,H/kJ mol-ANiH/kJmol-' no. 25 137 34.5 0.90 2 196 28.6 0.78 15 219 27.1 not measured 12 218 47.5 0.68 26 179 43.6 0.79 7 ? not measured -0 141 35.0 1.12 -5 249 39.4 0.69 17 126 32.8 not measured 40 150 33.8 1.24 Table 7 Calorimetric observation of a homologous series of 1-naphthyl Table 8 Calorimetric observation of the homologous series esters4 / ref. 15so anisotropic no. n cr phase IS T:N/oC (Siamese-twin mesogens) 4 116.8 10 * 66.6 (~12.4) -12 no.m n cr N IS TY/"C * 87.8 ( * 30.0) -15 8/l 1 * 189.2 58 no. AC,/J K-' mol-' A,H/kJ mol-' A,,H/kJ mol-' 812 2 ~211.4 (. 76) 49 813 3 * 194.0 80.4) 46(a 71 1 156 32.3 -814 4 ~171.8 115.2) 41 0.60 815 5 -142.8 (a 109.7) 33(a712 I06 30.4 713 138 53.2 1.06 816 6 -148.0 ( 119.0) 31 817 7 -119.8 (* 1 1 3.6) 24 A,,H =enthalpy of phase transition from anisotropic, probably 818 8 132.2 (* 1 1 5.0) 24 nematic, to isotropic phase. 819 9 * 107.2 (* 87.0) 10 no. ACp/J K-' mol-' A,H/kJ mol-" ANiH/kJmol-' -1o4 273 40.1 a310 65.0 286 58.5 1.75 210 63.8 2.70 297 50.4 2.40:lo3 352 62.5 2.59 0 21 1 72.0 2.8 1 1 E 284 57.0 2.97e 1 40 44.6 2.14 1o2 a Could not be measured.Mixtures 15 10 5 The prevention of crystallization in liquid-crystalline phases (T-~,)-1/10-3 K-1 of pure low-molecular-weight compounds is one of the most Fig. 6. Vogel-Fulcher-Tammann plot of the same compounds as serious problems with respect to applications. It is mostly plotted in Fig. 5. Values of n: +, 5; 0,6; x ,7; a,8 pure substances with high glass-transition temperatures that J. MATER. CHEM., 1991, VOL. 1 Table 9 Calorimetric observation of some twin mesogens with aromatic bridges: C8H 17O~ c 0 2 ~ 0 2 c ~ o c 817 H X ref. 13 no. X cr N IS T:N/oC 911 -CO,fCH& 0 120.8 142.2 12 vOlcMirOpC- 912 125.6 12* 147.6 0 99.8 -179.6 21 914 * 140.6 ( * 127.4) 17 -Cop -CH2 -t$C:p -0pC - CH30 no.AC,/J K-' mol-' 371 554 475 296 A,H/kJ mol-' 45.8 33.4135.8 (two peaks) 50.7 75.7 ANiH/kJmol-' 4.67 3.38 4.25 3.54 Table 10 Viscosimetric data of the pure compounds l/5-1/8 0.099 885 241 (-32) 265 (-8) -5 0.461 607 253 (-20) 270 (-3) -8 0.020 1112 225 (-48) 254 (-19) -15 0.179 777 234 (-39) 256 (-17) -16 crystallize easily. Consequently, we prepared mixtures in order to decrease the formation rate of crystallites and to improve the conditions for freezing liquid-crystalline structures. Already in ref. 18 and 19 we have reported measurements on two glass-forming liquid-crystalline mixtures, consisting of four components. The principles for the creation of such glass- forming mixtures could be: (1) selection of not more than four components; (2) components should have preferably enanti- otropic liquid-crystalline phases; (3) creation of eutectic mix- tures.Composition and melting temperatures can be estimated by the Schroeder-Van Laar equation.22 The prerequisite is that the components should have different molecular struc- tures. (4) The beginning of the glass-transition interval changes approximately to a linear equation between the appropriate temperatures of the pure components (Ty):23 N Ty(mixture)= C xiT$N (2)i= 1 (5) The pure components should have low crystallization rates, but simultaneously, high glass-transition temperatures. For that reason we used compounds with alkyl chains contain- ing two to four methylene groups.As an illustration, we have chosen two mixtures (see Tables 11 and 12). Fig. 7 gives the calorimetric curves of both mixtures. The four-component mixture seems to be really eutectic. It does not crystallize, whereas the heating curve of the three-component mixture indicates a slight crystallization: at a heating rate of 20 K min-', ca. 9% of the substance crystallizes (exothermic effect at ca. 135 "C). Computed and measured melting temperatures also deviate from each other. This is a hint for the non-eutectic character which surely has its origin in the similar chemical structure of the 2-naphthyles- ter compounds. Evidently, eqn. (2) satisfactorily predicts the measured glass-transition temperatures. Discussion All the above experimental results allow us to formulate a structural model for the causes of glass-transition behaviour under moderate conditions at high temperatures in low- molecular-weight liquid-crystalline phases.(1) The main ingredient is a 'bulky' structural element at a central or terminal position of the rod-like molecule, resulting in a decreased 1ength:breadth ratio. Because of this, J. MATER. CHEM., 1991, VOL. 1 Table 11 Composition and calorimetric data of a four-component mixture (all temperatures in "C) ~~ component 6i9 1 /4 TF [according eq. (2)]: TfN(measured): T, (eutectic, computed and measured: TNi: 13 "C 10 "C 88 "C 127.5-131.5 "C X TY cr N IS 0.100 36 -188.2 263.0 0.124 17 174.4 a308.0 0.086 0.690 40 5 * 189.8 * 99.6 (* 176.0)-109.0 AC,= 189 J K-' mol-' A,H =38.6 kJ mol-' see Fig.7(a) ANjH= 1.45 kJ mol-' Table 12 Composition and calorimetric data of a three-component mixture (all temperatures in "C) component X TYI "C cr N IS c 3 H 7 0 ~ c 0 2 p c 0 2 ~ 0.365 31 -151.0 (-87.0) 615 Br CH30 0.352 26 -155.2 (-89.4) 5i 1 0.283 36 * 176.0 -297.0 T:N [according eqn. (2)] 31 "C T:N (measured) 30 "C AC,= 196 J K-' mol-' see Fig. 7(b) T, (eutectic, computed) T, (measured) 123.4 "C 146.3 "C second heating: AmH=3.9 kJ mol- ' TNi(wide clearing interval) 143-169 "C ANiH=0.55kJ mol-' x =molar fraction of the component monotropic phases are observed in most cases. The situation can be optimized, if the bulky element is situated in the centre of the molecule (e.g.1,4-disubstituted naphthalene derivatives show more enantiotropic phases, whereas 2-naphthylesters show monotropic phases and 1 -naphthylesters mostly give isotropic glassy phases; see Tables 3, 6 and 7). (2) Investigations into several homologous series has proved that the glass-transition temperature decreases if the number of methylene groups in the chains increases. The alkyl chains as well as alicyclic ring systems in the molecules decrease the glass-transition temperature (see Tables 1, 4, 5, 7 and 8). The systematic decay of the glass-transition tempera- ture for the substances from Table 1 is demonstrated in Fig. 8. nematic nematic isotropicglass T~~~= 10 oc no crystallization I I ....t... --4w1 100 1500 91°C --40 K min-' appearing of the nematic phase- 20 K min-'-L nematic glass nematic-w isotropiccrystallization____.._ = 30 "' (ca.9% of the mixture) 1-50 100 150 91°C The influence of alicyclic ring systems may be illustrated by the following example:I3 R T:N/oC cr N is GC5HIl 9 -142.8( 130.8) ---@Hl1 4 * 139.0(a 94.4) -(3) Additional lateral substituents increase the glass-tran- sition temperature because the bulkiness also increases. The length :breadth ratio becomes lower. For this reason the clearing temperature decreases or the liquid-crystalline phase generally vanishes. This may be demonstrated by the following 2-naphthyle~ters:~ X TfN/"C cr N is H 27 176.6 -217.8 176.8(0 92.0)CH30 46 (4) Polar substituents in the molecules increase both the clearing temperature and the glass-transition temperature.In our interpretation, the polar parts increase the density at constant temperature^.^^ This causes a comparatively lower free volume and higher viscosity and, in terms of the free- Fig. 7. DSC plots for two mixtures. See Tables 11 and 12 for volume a higher glass-transition temperature. We compositions also give an example to illustrate this4 Fig.8. Melting and clearing temperatures as well as lower limit of glass-transition intervals as a function of the chain length in the homologous series of 2-tert-butylhydroquinone bis(4-n-alkyloxy benzoates) T;"/'C cr N is CH30 25 161.2 (435.2) NC 40 * 189.8 (-176.0) The compound with the shortest alkyl chain in a homologous series has the highest glass-transition temperature. As a matter of experience, long alkyl chains cause the appearance of smectic phases.It consequently follows that the glass-tran- sition in smectic phases should predominantly be observed at low temperatures (e.g. Tables 4 and 5). In general, the above results indicate that vitrification at high temperatures and mesogenity are in contradiction to each other in low-molecular-weight substances. Using the concept of the free-volume theory25 for forming glass phases, the above rules may be interpreted by the reduction of the accessible free volume by polar groups, stiff molecules with strong attractive forces, bulky substituents in lateral or ter- minal position.26 Molecules following this structural arrange- ment have a low mesogenity so that compromises in the molecular structure are necessary.On the other hand, in order to observe glass transitions it is necessary to avoid crystalliz- ation. This may be reached by a non-symmetric molecular shape, elongation of alkyl chains (optimum C5-C7) and creation of mixtures. Obviously, this demands further compro- mises. Therefore, obtaining low-molecular-weight liquid crys- tals with high glass-transition temperatures is a difficult, although not impossible, task. J. MATER. CHEM., 1991, VOL. 1 The authors are indebted to Professor K. H. Dehne and Dr. A.Roger, Giistrow, for the syntheses of some of the substances mentioned in this work, as well as for fruitful discussions. References 1 G. Tammann, Aggregatzustande, Leipzig, 1923. 2 (a)D. Demus, G. Pelzl, W. Wedler, Anisotropes Festes Optisches Medium, DD WP C09 K/282 705/8, 1985; (b)D. Demus, G. Pelzl, Polarisatoren, DD WP 242 625 A 1, 1985; (c) D. Demus, G. Pelzl, Thermo-elektrooptisches Speicherdisplay, DD WP 242 624 A 1, 1985; (d) D. Demus, G. Pelzl and W. Wedler, Proc. Eurodisplay '87, London, September 1987, p. 71. 3 D. Demus, W. Wedler, K. Mohr, W. Weissflog, W. Schafer and R. Schmidt, Nematische Flussigkristalle mit Glasphasen, DD WP C09 K/307 180/2, 1987. 4 D. Demus, W. Wedler, W. Schafer, G. Uhlig and W. Weissflog, Anisotrope Optische Medien, DD WP C09 K/316 239/7, 1988.5 D. Demus, W. Weissflog, G. Pelzl, W. Wedler and A. Humke, Niedermolekulare Glasbildende Gemische mit Smektischen Phasen, DD WP C09 K/318 571/5, 1988. 6 J. Grebovicz and B. Wunderlich, Mol. Cryst. Liq. Cryst., 1981, 76, 287. 7 H. Yosioka, M. Sorai and H. Suka, Mol. Cryst. Liq. Cryst., 1983, 95, 11. 8 K. S. Kunihisa and Y. Satomi, Mol. Cryst. Liq. Cryst., 1986, 141, 1. 9 Ch. Kuhrmann, Dissertation, Halle, 1926. 10 S. Baentsch, Dissertation, Halle, 1931. 11 D. Demus, Liq. Cryst., 1989, 5, 75. 12 W. Weissflog, R. Schlick and D. Demus, Z. Chem., 1981,21,452. 13 W. Weissflog, D. Demus, S. Diele, P. Nitschke and W. Wedler, Liq. Cryst., 1989, 5, I 11. 14 W. Weissflog, A. Wiegeleben, S. Diele and D.Demus, Cryst. Res. Technol., 1984, 19, 583. 15 H. Dehne, A. Roger, D. Demus, S. Diele, H. Kresse, G. Pelzl, W. Wedler and W. Weissflog, Liq. Cryst., 1989, 6, 47. 16 F. Schneider and H. Kneppe, Fliejverhalten von Stoflen und Stoflgemischen, ed. W-M. Kulicke, Hutling und Wepf, 1986, ch. 8, pp. 318-368. 17 M. Miesowicz, Mol. Cryst. Liq. Cryst., 1983, 97, 1. 18 H. Kresse, S. Ernst, W. Wedler, D. Demus and F. Kremer, Ber. Bunsenges. Phys. Chem., 1990, 94, 1478. 19 R. Stannarius, W. Gunther, M. Grigutsch, A. Scharkowski, W. Wedler and D. Demus, Liq. Cryst., in the press. 20 (a) H. Vogel, Phys. Z., 1921, 22, 645; (b) G. S. Fulcher, J. Am. Chem. SOC., 1925,8,789; (c)G. Tammann and G. Hesse, Z. Anorg. Allg. Chem., 1926, 156, 245. 21 A. Feltz, Amorphe und Glasartige Anorganische Festkorper, Akad-emie-Verlag, Berlin, 1983. 22 (a) I. Z. Schroeder, 2. Phys. Chem., 1893, 11, 449; (b) J. J. van Laar, 2. Phys. Chem., 1908, 63, 216. 23 H. Schad and H. L. Zeller, Phys. Rev. A, 1982, 26, 2940. 24 D. Demus, 2.Chem., 1986, 26, 6. 25 (a) A. K. Doolittle and D. B. Doolittle, J. Appl. Phys., 1957, 28, 901; (b)W. Brostow, Polymer, 1980, 21, 1410. 26 (a) D. Demus, W. Wedler, W. Schafer, H. Zaschke, G. Uhlig, H. Kresse, S. Diele, G. Pelzl and W. Weissflog, Proc. 8th Liq. Cryst. Conf. Socialist Countries, Krakow, August-September, 1989, abstracts A-D; (b) W. Schafer, G. Uhlig, H. Zaschke, D. Demus, S. Diele, H. Kresse, S. Ernst and W. Wedler, Mol. Cryst. Liq. Cryst., 1990, 191, 269. Paper 0/04716C; Received 19th October, 1990