J. CHEM. SOC. FARADAY TRANS., 1994, 90(11), 1537-1540 Glass Transition of Liquid-crystalline 4-Alkoxyphenyl and 4-Cyanophenyl 4=(2,4=Dialkoxybenzoyloxy)Benzoates Shunsuke Takenake and Hiroshi Yamasu Department of Materials Science and Engineering, Faculty of Engineering, Yamaguchi University, Ube Yamaguchi 755,Japan Some 4-alkoxyphenyl and 4-cyanophenyl 4-(2,4-dialkoxy)benzoates show a nematic phase and a glassy phase at low temperatures. The glass-transition phenomena were examined as a function of the sweep rate in both heating and cooling processes. The derivatives intrinsically experience two kinds of transition process in the nematic phase. Two transitions were observed when the sweep rate was <1 K min-', and these overlapped when the sweep rate was >1 K min-'.The transition temperatures show an interesting dependence on the chain length of three alkoxy groups; i.e. lengthening the lateral alkoxy group at position 2 lowers the glass- transition temperature (T,) while lengthening both alkoxy groups at the terminal positions increases Tg. The thermodynamic parameters are discussed in terms of the McMillan theory of the glass transition. Pure compounds sometimes exhibit a glassy phase when re- crystallization does not occur as they are cooled from the molten state. Some liquid-crystalline materials are also known to exhibit glassy phases corresponding to pre-states such as cholesteric'*2 and nematic phase^.^-^ A nematic glassy phase is also formed by a eutectic mixture.' The for- mation of a glassy phase is usually very undesirable in liquid- crystal cells that are driven electrically.In pure compounds, however, the observation and thermodynamic consideration of the glassy phase is not easy since in many cases re-crystallization precedes the transition. In such cases, the glassy phase is usually observed only during rapid cooling of the sample. In this paper, we describe the thermal properties of the liquid-crystalline compounds shown in Table 1. The derivatives have a long alkoxy group in the lateral position so that these form only a nematic phase mono-tropically. However, the lateral alkoxy group prevents re- crystallization during the cooling process and makes observation of the glassy phase easy. The results will be dis- cussed in terms of the McMillan theory of the glass tran- si tion.' Experimental The materials were synthesized according to the method described previously.' The purity and structures were con- firmed by HPLC, NMR and elemental analysis.The thermal properties were determined by a Seiko-denshi SSC-5200 work station differential scanning calorimeter Table 1 Liquid-crystalline compounds studied R2 R' RZ R3 17O OC4H9 OCBH17 'aH, 7O 0C6H 13 OCBH17 7O OC4H9 'sH17 'aH 17' 0C6H 13 CN 17O oclOH21 OCEHl7 C12H250 OC4H9 OClzH25 C12H250 oc1 OH21 OCl 2H25 'aH 17O H OCaH, 7 (DSC) and a Nikon POH polarizing microscope fitted with a Mettler FP-5 heat controller. The DSC thermograms were calibrated with indium (99.9%, mp 439.8 K, AH = 28.59 mJ mg-I).Results and Discussion The transition temperatures and latent heats for compounds 1-7 are summarized in Table 2. The melting points in Table 2 were obtained during the heating process of the virgin sample, as were the N-I transition temperatures. The liquid-crystalline phase showed a typical schlieren texture and it was identified as a nematic phase. 8 has a fun-damental skeleton of the present series, and the structure may keep a rod-like nature. 8 is known to exhibit a mesophoric sequence of the crystal-smectic C-smectic A-nematic-isotropic type. Therefore, the N-I transition temperature is high. As is evident from a comparison of 1-8, a lateral substituent at position 2 reduces the N-I transition temperature markedly.In addition, the presence of a lateral substituent tends to prevent the formation of smectic layers. The effect of a lateral substituent on the mesomorphic properties has been already discussed.' The DSC thermograms for 1and 4 are shown in Fig. 1 and 2. Fig. 1 shows the DSC thermograms for 1. In the cooling processes the thermograms showed an exotherm at 342 K (0.5 K min-') due to the I-N transition. The same transition occurred at ca. 342 K for cooling rates of 2, 5 and 10 K min-', and ca. 341 K for cooling rates of 20, 25 and 30 K min-' (the temperature was obtained from the onset of the exothermic peak). As shown in Fig. l(a), the exothermic Table 2 Transition temperatures (T/K) and latent heats (kJ mol-') for 1-7 compound K N I AHm,, * 344 (. 342).19.9 -311 * 337 * 27.6 -324 (* 312) -34.2 -348(* 315). 14.4 * 347(* 322). 27.5 -342 * 342 31.5 * 318 * 322 * 34.7 -357 * 461 AH,, 0.8 0.7 0.3 0.1 0.3 0.6 0.4 Parentheses indicate a monotropic transition. K, N, and I indicate crystal, nematic and isotropic phases, respectively. " This compound undergoes smectic C-smectic A and smectic A-nematic transitions at 417 and 436 K also, respectively." J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 glass20852800< 2025 I U I /A/ 10U "V198 227 256 285 315 343 TIK 0 -750 z $-1500 n -2250 N -3000 I I 1 in the thermograms indicate the sweep rate (K min-'). peaks tend to become broader with increasing cooling rates.The small change in the transition temperature is due to instrumental error. The DSC thermograms in Fig. l(a) show a remarkable deviation of the base line around 240 K due to a glass tran- sition of the nematic phase. Therefore < values were obtained from the peak of the differential curve, according to the McMillan method.' The values decreased gradually with increasing cooling rate (Fig. 3). During the heating process, a glass-N-crystal transition can be observed in Fig. l(b). The glass transition accompa- nied a small endotherm arising from the enthalpy relaxation when the heating rate was slow, e.g. 2-10 K min-'. The endotherm is more pronounced for terminal groups with higher carbon numbers, e.g. 6 and 7.Interestingly, two kinds of glass transition were observed at 230 and 247 K when the heating rate was 1 K min-', while only one kind was observed when the heating rate was > 1 K min-'. The deviation from the baseline was larger for the former than for the latter. This shows that this compound intrinsically has at least two kinds of glassy phase, and the transitions merge when the sweep rate is >2 K min-'. Recrystallization occurred whatever heating rate was used ; recrystallization occurred at 263 K for a heating rate of 2 K min-' and at 298 K for a heating rate of 30 K min-'. The endothermic peak around 340 K arises from melting of the crystalline phase formed during the heating process. Inter- estingly, a similar deviation of the baseline to the glass tran- sition is observed during the recrystallization process.A 1850 1325 800 275 I I I I I I 213 237 261 285 309 333 TIK -1 00 -575 32 -1050 n -1 525 glass I-2000 213 237 261 285 309 333 T/Vlib Fig. 2 DSC thermograms for 4: (a) cooling, (b) heating. The numbers in the thermograms indicate the sweep rate (K min-'). similar phenomenon was observed for related compounds with a long substituent in the lateral position of liquid- crystalline molecules.' ' The DSC thermograms for 4 exhibited a similar feature (Fig. 2). During the cooling process the I-N transition occurred at 318-317 K for a cooling rate of 1-25 K min-'. The onset of the transition peak was almost independent of the cooling rate, while the peak became broader with increas- ing cooling rate.During the heating process, the monotropic nematic phase recrystallized during heating at any rate, but the transition temperature increased with the heating rate. A distinct glass transition for 1 was not observed even for heating and cooling rates of 1 K min-'. Fig. 3 shows plots of the % against sweep rate. For 1 T value varies from 240 K (sweep rate 1 K min-') to 246 K (30 K min-') in the heating process and from 240 K (sweep rate 1 K min-') to 235 K (30 K min-') in the cooling process. The characteristics of the heating and cooling processes are almost symmetric with respect to the 240 K line. We propose that is proportional to the sweep rate within the tem- perature range studied.The temperature dependence of 3 was obtained from the slope of the plots, giving 0.27 and 0.23 for the heating and cooling processes, respectively. Similar results were obtained for 2-5. The features for 6 and 7 are somewhat different. In the heating process, the temperature dependence of the < value on the sweep rate is 0.07, while for the cooling process it is 0.5. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 I ^^^I .== 1220p *** sweep rate/K min-' sweep rate/K min-' sweep rate/K min-' 2250 46r Y h' 242 0 10 20 30 0 10 20 30 40 0 10 20 30 40 sweep rate/K min-' sweep rate/K min-' sweep rate/K min-' 260(g)l .-240-2401 230 0 10 20 30 40 sweep rate/K min -' Fig. 3 Tgus.sweep rate for compounds: (a)1, (b)2, (c) 3, (d) 4, (e) 5,(I)6, (9)7. Upper and lower plots correspond to the heating and cooling processes, respectively. Extrapolation of the plots for the heating and cooling pro- cesses in Fig. 3 give the 'average' transition temperatures for heating and cooling rates of 0 K min-(Table 3). This table shows some interesting trends in connection with the molecu- lar structure. The ratios of 5 and the sweep rate (T, in K min-') in the heating and cooling processes are symmetrical for 1-5, but those for 6 and 7 are asymmetrical, i.e. the Tg values for the heating process are almost independent of the rate, while those for the cooling process change markedly with T,. Generally, the ratio of <to Tmpis ca. 0.7 due to the similarity in the entropy change between the melting and the glass-transition processes.' As can be seen from Table 3, the T$Tmpvalues for 1-5 fall in this category, while those for 6 and 7are significantly greater than 0.7.Kirov et al. have demonstrated that the values for sub- stituted two-ring compounds show a good correlation with their molecular weights.6 We attempted to correlate q with the carbon numbers of both lateral and terminal alkyl chains Table 3 Thermodynamic parameters for the glass transition compound Tmp heating cooling Tgb </TmP 1 344 0.27 0.23 240 0.70 2 331 0.23 0.23 234 0.71 3 324 0.27 0.23 225 0.69 4 348 0.2 0.2 244 0.70 5 347 0.23 0.23 231 0.67 6 342 0.07 0.5 268 0.78 7 318 0.07 0.5 250 0.79 is the sweep rate.'Values extrapolated from Fig. 3. (Fig. 4). This figure clearly shows the role of the alkoxy group in the glass transition, viz. the lateral alkoxy group reduces the transition temperature, probably due to an increase in the molecular breadth. On the other hand, the terminal alkoxy groups increase the transition temperature. Therefore, these results indicate that the glass transition is dependent on the molecular shape rather than the molecular weight. As mentioned above, both Tmpand % for the cyano com- pound, 4, are higher than those for the other derivatives. However, there is no fundamental difference in the ratio of q to Tmpcompared with the ratios for the other derivatives. The present results are discussed in terms of the McMillan model for the glass (g) transition of the nematic phase (N);' g eN (1)l-x x dx/dt = (1 -x)nk (k, T/h)exp(AS/R)exp(-AH/RT) (2) where k is the transmission coefficient for activation, k, is Boltzmann's constant, h is Plank's constant, R is the gas con- stant, AS is the entropy of activation, AH is the enthalpy of activation and n is the order of the reaction.The solution of eqn. (2) gives : (3) Eqn. (3) indicates that a plot of log[Ti/(dT,/dt] US. 1/< is linear, the slope giving the enthalpy of activation for the glass 220 4 6 8 10 12 carbon number 220 4 6 8 10 12 carbon number Fig. 4 Tgus. the carbon numbers of the terminal (a)and lateral (b) alkyl chains ern rnc rn rn 3.0 400 410 420 4 1 420 440 460 lo5 KIT, lo5 KIT, w .5.0-4.0-rn rn 4.0-*rn *m 't 4 3.0.-3.0 .I.,.I. 400 410 420 390 400 410 420 4d~ 1O5 KIT, 1O5 KIT, Fig.5 McMillan plots for: (a)1, (b) 3, (c) 4 and (d) 7. The plots for the left and right sides in each figure are for the heating and cooling processes, respectively. transition. Such plots are shown in Fig. 5. Interestingly, the plots for the cooling and heating processes are symmetrical, except for compounds 7 and 6. In order to clarify the effect of the conditions of the glass transition such as the cooling rate and an anneal, the values were examined under the various conditions (Table 4). The results in Table 4 show that the Tg values obtained for the heating process are completely inde- pendent of the cooling conditions and the anneal of the glassy state.Therefore, the symmetric and asymmetric fea- tures in Fig. 5 are due to the intrinsic nature of the molecules. McMillan' and Tsuji et aL2 calculated the activation enth- alpies for the glass transition of glycerol and a steroid deriv- ative by using eqn. (3) and obtained reasonable values. In the present case, however, Fig. 5 shows a remarkable non-linear behaviour in both heating and cooling processes and does not follow eqn. (3). The following facts are noteworthy in connection with the abnormal features in Fig. 5. Table 4 Effect of cooling rates on the Tgvalues in the heating process for 1 heating rate T,"/cooling rate T,b/cooling rate 1" 23011 230110 247 2471 2 24112 241110 5 24215 243110 243/1od 10 244110 244110 20 245120 2451 10 30 246130 246110 The values were rounded to the nearest whole number." Samples prepared by cooling from 373 to 193 K at the cooling rates indicated. Samples prepared by cooling from 373 to 193 K at a cooling rate of 10 K min-'. 'Two kinds of transformation could be observed. The glassy state was annealed for 2 h at 233 K before it was heated. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 As can be seen from Fig. 1 and Table 4,l shows two kinds of glass transition when the heating rate is 1 K min-'. A similar trend is observed for 2 and 5. The glass transition temperatures for 2 and 5 are 229 and 245, and 230 and 237 K, respectively. These results indicate that the transition in the low-temperature region is almost independent of the chain length of the lateral alkoxy chain, and the 3 in the high-temperature region decreases with increasing chain length.As a result, the difference between values becomes small with increasing chain length of the lateral alkoxy chain. This apparent separation of the glass transition was not observed in 6 and 7, where the alkoxy chains at the terminal and lateral positions are significantly longer. Sorai et al." reported that N-(4-n-pentyloxybenzylidene)4'-n-butylaniline showed two kinds of glassy phases of a smectic G phase, and the transitions were attributed to a freezing of the intrinsic molecular modes of the layer structure and anisotropic translational self-diffusion parallel and perpen- dicular to the long molecular axes.This explanation cannot be applied to the split in the glassy transition in the nematic phase. We assume that the freezing of motional freedom for the terminal and lateral alkoxy groups is responsible for the glass transitions in the low- and high-temperature regions, respec- tively, and the 5 values observed for the high sweep rates are a thermodynamical average of the two transition tem-peratures. Conclusions In the 4-substituted phenyl 4-(2,4-disubstituted benzoyl- 0xy)benzoate systems the substituent at position 2 affects the observation of the glass transition. Two kinds of glass tran-sition were observed when the sweep rate was slow, and only one kind when the sweep rate was fast.Chain elongation at both terminal alkoxy groups increases the transformation temperature, while elongation of the lateral groups reduces it. The formation of two kinds of glass phase in the nematic phase is quite rare. This work was supported by Grant-in aid (no. 04640504) from the Ministry of Education, Science and Culture, Japan. References 1 K. Adachi, H. Suga and S. Seki, Bull. Chem. SOC.Jpn., 1969, 41, 1073; 1970,43,1961; 1971,44,78. 2 K. Tsuji, M. Sorai and S. Seki, Bull. Chem. SOC.Jpn., 1971, 44, 1452. 3 M. Sorai and S. Seki, Bull. Chem.SOC.Jpn., 1971,44,2887. 4 M. Sorai and S. Seki, Mol. Cryst. Liq. Cryst., 1973,23, 299. 5 N. Kirov, M. P. Fontana and F. Cavatorta, Mol. Cryst. Liq. Cryst., 1979,54,207. 6 N. Kirov, M. P. Fontana and N. Affanassieva, Mol. Cryst. Liq. Cryst., 1982,89, 193. 7 J. Cognard and C. Ganguillet, Nol. Cryst. Liq. Cryst. (Lett.), 1978,49, 33. 8 J. A. McMillan, J. Chem. Phys., 1965,42, 3497. 9 S. Takenaka, Y. Masuda, M. Iwano, H. Morita, S. Kusabayashi, H. Sugiura and T. Ikemoto, Mol. Cryst. Liq. Cryst., 1989, 168, 111. 10 D. Demus, H. Demus and H.Zaschke, Flussige Kristalle in Tabellen, VEB Deutscher Verlag fur Grundstoff Industrie, Leipzig, 1976. 11 S. Takenaka and H. Yamasu, Mol. Cryst. Liq. Cryst., in the press. 12 M. Sorai, K. Tani and H. Suga, Mol. Cryst. Liq. Cryst., 1983,97, 365. Paper 3/07032H;Received 26th November, 1993