SPECIFICHEATS OF LIQUIDCRYSTALS
895
Specific Heats of Nematic, Smectic, and Cholesteric Liquid Crystals by Differential Scanning Calorimetry1
by Edward M. Barrall, 11, Roger S. Porter, and Julian F. Johnson Chevron Research Company, Richmond, California (Received August 19, 1966)
Differential scanning calorimetry has been applied to the study of three pure compounds exhibiting liquid crystal or mesophase behavior. The compounds were p-azoxyanisole, anisaldazine, and cholesteryl myristate. The first two compounds exhibit a single mesophase of the nematic type, and the last compound forms both a smectic and a cholesterictype mesophase. The three basic types of mesophases are thus evaluated in this study of specific heats. Measurements were made in the temperature ranges for the crystalline solid, the mesophase, and the corresponding isotropic liquid. Specific heats were also measured on supercooled phases and at temperatures close to the first-order mesophase transitions. The measurements on p-azoxyanisole compare well with previous literature values. Specific heat measurements on cholesteryl esters have not been presented heretofore. The results on the three basic mesophase types are intercompared.
A group of organic compounds designated as “liquid materials have the unusual property of exhibiting one or more phases or mesophases which are liquid in mobility yet solid in structure. This is accomplished by having crystal-like order in one or more directions and liquid-like disorder in the other spatial orientations. In excess of 1500 compounds are known to exhibit liquid crystal properties.”6 The various liquid crystal phases which exist between true solid and isotropic liquid, viz., mesophases, are classified as smectic, nematic, and cholesteric. Briefly, the smectic mesophase is characterized by having the molecules stratified, Le., arranged in layers with long axes approximately normal to the plane of the layers. The molecules can move in two directions (in plane) and rotate in plane. The nematic structure is somewhat less restricted. The molecules are arranged parallel to one another. Movement can occur in three directions, and rotation in one. The cholesteric mesophase is exhibited principally by the esters of cholesterol and is due to the helical stacking of the cholesterol plates. The plates in each layer of the mesophase are displaced from the next layer by a small amount due to the ester tail and the 4-hydrogen which leads to a helical stacking order. The cholesteric mesophase
is exhibited by cholestane or cholestene derivatives only when rings A and B are in the trans configuration. Surprisingly few thermodynamic data are available on these basic types of mesophases. Until recently,’ only one heat of transition had been reported for a cholesteryl ester, by Hulett for cholesteryl benzoate,3E and less than a dozen transition heats were known*-l2 for all other liquid crystal systems. Specific heat data determined by a direct method were nonexistent until (1) Part X of a series on order and flow of liquid crystals. (2) 0. Lehmann, Z . Physik. Chem., 5 , 427 (1890). (3) (a) G.H. Brown and W. G. Shaw, Chem. Reu., 5 7 , 1049 (1957); (b) V. A. Usol’tseua and I. G. Chistyakov, Russ. Chem. Rev., 3 2 , 495 (1963). (4) R. vonBrauns, “Flussige Kristalle und Lebewesen,” Schweizerbartische Verlagsgesellschaft, Stuttgart, 1931. (5) D. Vorlander, Ber., 41, 2033 (1908). (6) W.Kast, Angew. Chem., 67, 592 (1955). (7) E.M. Barrall, R. S Porter, and J. F. Johnson, “Heats of Transition of Some Cholesteryl Esters,’’ to be published. (8) E. M. Barrall, R. S. Porter, and J. F. Johnson, J. Phys. Chem., 68, 2810 (1964).
(9) H. Martin and F. H. Mtiller, Kolloid-Z., 187, 107 (1963). (10) R. Schenck, “Kristalline Flnssigkeiten und Fltksige Kristalle,” Leipzig, 1905,pp 84-89. (11) K. Kreutzer, Ann. Physik, (5)3 3 , 192 (1938). (12) K. Kreutzer and W. Kast, Naturwissenschaften, 2 5 , 233 (1937).
Volume 71, Number 4
March 1967
896
Arnold's study of homologous dialkoxyazobenzoates.la3l4 Recently, a differential scanning calorimetric (dsc) method has been developed for the rapid, direct measurement of' specific heat.15 Since the method is continuously recording, it is possible to determine the specific heat of a phase over a very narrow, defined temperature range within the limits of a dynamic method. Such a method is ideal for the initial study of liquid crystals which are noted for pretransitional effects in narrow temperature intervals on each side of first-order mesophase t r a n ~ i t i o n s . ~ J This new technique is further substantiated by a direct correlation of data with those obtained by classic calorimetry. I n this study, the specific heats measured on p-azoxyanisole (PAA) by dsc are compared with Arnold's measurements obtained by classical calorimetry in the same temperature range.14 In addition, the specific heats of anisaldazine (AAD) and cholesteryl myristate (CM) are also developed over a wide temperature range using dsc. These measurements constitute the first determination of the specific heat of a cliolesteric mesophase. The thermodynamic reversibility and identity of mesophase formation from cooled melt and from heated solid is also explored. Classical thermodynamics, which disregards statistical considerations, treats crystals and liquids as completely independent entities. This gives no basis for either expecting or predicting pretransitional effects. However, calculation of pretransitional anomalies by the theories of cooperative fluctuations predict some temperature effects on both sides of first-order transitions. The parameter controlling the magnitude of the observed effects is the interfacial energy.I6 This parameter is difficult to estimate exactly. However, it is reasonable to estimate that the interfacial energy will be small for liquid crystal transitions due to the small size of the heat of transformation. This should and does lead to large pretransitional effects as reported in the following results. Transition temperatures and heats for the three compounds studied are compiled in Table I. Experimental Section The origin and purification of the liquid crystal materials has been discussed in detail in previous works.7,8,17 Briefly, PAA and AAD were obtained from Organic and the CM was obtained in a highly purified form from Applied Science Laboratories, State College, Pa. All -~ materials were recrystallized from ethanol three times. Carbon-hydrogen and spectrophotometric indicated a purity of better than 99.9%0,8*17 the acid The Journal of Physical Chemistry
E. M. BARRALL, 11, R. S. PORTER, AND J. F. JOHNSON
Table I: Transition Temperatures and Heats for Mesophase Systems Transition temp,
O c a
Transition heat, oal/g
p-Azoxyanisole Solid-nematic Nematic-iso tropic
117.6 133.9
28.1 == ! 0.9 0.68 & 0.02
Anisaldasine Solid-nematic Nem &-isotropic
168.9 180.5
26.5 & 0.5 0.59 =$ 0.02
73.6 79.7 85.5
18.7 0.52 0.41
Cholesteryl myristate Solid-smectic Smectic-cholesteric Cholesteric-isotropic
' By differential thermal analysis. chloride and cholesterol reactant content being used as an index of purity. Thp scanning calorimeter (Perkin-Elmer Model DSC 1-B) was calibrated using a 0.03-g sapphire standard plate supplied for this purpose by the PerkinElmer Co. The detailed technique of calibration has been given in full by Wunderlich.lB A useful description of the experimental details is available from the Perkin-Elmer CO.l9 The sapphire specific heat data used for calibration were obtained from the work of Ginnings and Furukawa.20 All calculations were carried out using a computer program developed at Chevron Research Coo Measurements of programmed base line displacement were made with a precision vernier rule at 2" intervals. Specific heats were calculated a t 0.5" intervals for the smectic mesophase and the isotropic liquid range. A lO"/min heating rate was used to cover the full temperature range of heat capacity measurements on all compounds. Detailed studies over narrow temperature ranges near transitions were carried out at a 5"lmin heating rate. Intervals of 30-40" were taken between reference base line determinations. All samples were contained within sealed planchettes which were weighed before (13) H.Arnold, 2. Chem., 6, 211 (1964). (14) H. Arnold, Z . P h w i k . Chem., 226, 146 (1964). (15) A. P. Gray and N. Brenner, Polymer Preprints, 6, 956 (1965). (16) R. R. Ubbelohde, "Melting and Crystal Structure," Clarendon Press, Oxford, 1965, pp 68,243. (17) E. M.Barrall, R. S. Porter, and J. F. Johnson, J . Phys. Chem., 7 0 , 385 (1966). (18) B. Wunderlich, ibid., 69, 2078 (1965). (19) Thermal Analysis News Letter No. 3, Perkin-Elmer CO., Norwalk, Conn., 1965, p 1. (20) D. C. Ginnings and G. T. Furukawa, J . A m , Chem. sot., 75, 522 (1953).
SPECIFIC HEATSOF LIQUID CRYSTALS
897
- _-
Agreement between Arnold's data and Figure 1 is excellent within the stated error limits of the two methods except in the low temperature limit where values obtained here are about 3% lower than given by Arnold. (See Table 11.) One earlier set of data also gave lower specific heats for this crystalline region. (See ref 10.) The data of Arnold reportedly have a high precision of =k0.3%. The isotropic liquid phase relationship dips down above 155" due to a small vaporization of sample, noted earlier. The dashed line in Figure 1 represents extrapolated heat capacity data.
086-
0.32-
078-
0 74-
0 70-
E ?
-
a u
-
w 4
-
5066-
,- 0 6 2 Table I1 : Comparison of Literature and Scanning Calorimetry Values for the Heat Capacity of pAaoxyanisole
S
V058-
I u
W
-
Temp,
p.
mo54-
OC
95 100 105 110 113 110 115 118.2 120 125.5 127 128 129 130 131 137 138 140 142 144 146 148 150
0 50-
L
0.42
-
0 38-
0.34
:/
-SOLID
+NEMATIC
+
ISOTROPIC LIQUID
-
1 I
I
I
I
I
I
I
Figure 1. Specific heat of p-azoxysnisole from 87 to 167".
and after each run. No weight losses were noted except in the case of PAA. Each data point shown represents an average of three points measured in each of three separate experiments. The average deviation was =1=357,. In some cases the error increased to h6Y0 near sharp transitions due to nonequilibrium conditions. Vertices of known transition peaks are not shown since these became indefinitely large, >10 cal/g deg.
Results p-Azoxyanisole ( P A A ) . A single, nematic liquid crystal phase is reported for PAA.8r9**4The transition heat from solid to nematic phase is large, 28.8 cal/g a t 117.6'. The transition from the nematic to isotropic liquid is small, 0.68 cal/g a t 133.9°.899 The transition heats and temperatures were determined by dta on pure PAA. Two determinations of specific heat on this compound have been made: by Arnold14 using classical calorimetric techniques, and by Martin and Muller using differential thermal analysis (dta) .g
-Heat capacity, cal/g deLit." This work
0.3631 0.3688 0.3762 0.3949 0.4359 0.4524 0.4596 0.4647 0.4682 0.4792 0.4853 0.4894 0.4934 0.4986 0.5058 0.4756 0.4734 0.4712 0.4701 0.4897 0.4696 0.4697 0.4698
0.3520 0.3595 0.3700 0.3895 0.4295 0.4549 0.4575 0.4600 0.4620 0.4740 0.4795 0.4842 0.4925 0.4970 0.5120 0.4850 0.4680 0.4650 0.4658 0.4660 0.4660 0.4662 0.4668
Phase
Solid
Nematic
Isotropic liquid
The transition from solid to the nematic mesophase is characterized by a gradual upward sweep of the specific heat prior to sharp transition; that is, the curve for the solid phase approaches infinity a t temperatures just above 112.5'. This type of curve has been described by Westrum and McCullough as either a type 21 or a type 31 transition.21 It was impossible to measure the descending portion of the specific heat curve just above the solid-nematic transition (21) E. F. Westrum and J. P. McCullough, "Physics and Chemistry of the Organic Solid State," D. Fox, M. M. Labes, and A. Weiss-
berger, Ed., Interscience Publishers Inc., New York, N. Y., 1963, 40 ff.
p
Volume 71, Number 4 March 1967
898
with sufficient precision to determine if the transformation is truly isothermal at the finish. Therefore, further separations of the transition by empirical type are difficult. However, since the purity of the PAA is known to exceed 99.9%, it is difficult to assign the 21 transition type, which is supposed to be characteristic of the melting of an impure Premelting molecular rotation in the solid phase seems more likely. The size and shape of nematic aggregates may also change near the transition so that there may be a minor order and entropy change in the pretransition range. This places the solid-nematic transition in the 31 category with the higher normal paraffins. Since the PAA molecules are approximately rod-shaped about a flat benzene plane, such a rotational adjustment on tht: basis of molecular geometry is to be expected. Several workers have observed thermal changes in the nematic phase prior to the nematic-isotropic liquid t r a n s i t i ~ n .' 4~ ?Jlartin and Miillera reported a pretransition between 128 and 132" (the 132" temperature being taken as the clarification temperature of the melt). This corresponds to the possible break and gently upw:ird sweeping portion of the specific heat curve found in this study (see Figure 1 between 128 and 134.4') The sharp break a t 128" is comparable to a simple first-order melting range transition, classified type I 21 The portion of the nematic specific heat between 128 and 134.4" appears to be reproducible and stable. This section of the curve was repeated on heating and cooling four times on different samples with identical results and no supercooling or superheating. The resolution of this mesophase region is not evident on dta curves of normal resolution.8 The only other evidence for existence other than the present work is given by Martin and Muller.9 Arnold14 observed a change in his values but discounted the data. Since this minor anomaly has been observed by a t least three independent workers, it is apparently not a calorimetric artifact. However, specific volume, viscosity, and surface tension measurements in this range do not show a distinguishable anomaly.2,zz,23 An attempt was made to determine if the specific heat of the nematic mesophase formed by cooling the isotropic liouid was identical with that formed by heating the crystalline solid. The melt was cooled to 107" and heated, as previously described, to 121'. In a second experiment, the solid phase was heated into the nematic region, cooled to 107", and heated to 121". The results are shown in Figure 2. Within the limits ctf experimental error, the nematic mesophase strucl,ure of p u formed from the solid and from the isotropic liquid are identical and readily The JOUT?UZ~ of Physical Chemistry
E. M. BARRALL, 11, R. S. PORTER, AND J. F. JOHNSON
-
0.466
0.462
0
t
V U
w I
0.454-
V
E 0.450 U w p. v,
0.446-
~
110
~
112
~
1
114
"
116
"
"
118
"
120
122
SAMPLE TEMPERATURE, "C Figure 2. p-Azoxyanisole nematic mesophase: 0,formed from the solid; 0 , from the liquid; 0 , Arnold's data."
reversible. Actual cooling curves were not employed because of instrumental difficulties. Anisaldazine (AAD). The specific heat-temperature relationships of AAD (see Figure 3) are simpler than those observed for PAA. Both the solid-nematic and the nematic-isotropic liquid show premelting specific heat changes. They belong to the type 21 or 31 transitions of Westrum and McCullough. 21 Premelting or pretransition chain rotation and adjustment are prominent features in both dtag and depolarized light intensityz4studies of AAD. Data in the range of the nematic-isotropic liquid transition from 187 to 200" identifies it empirically as a type 31 nonisothermal chain rotation. The shape of such specific heat curves has been compared to the Greek capital A and referred to as A-type transitions. The general findings, consistent with these results, indicate that A-type transitions arise when the differences are small between the structures undergoing the transition. l6 No evidence was found for two nematic mesophases in AAD. The steep premelting slope of the nematic specific heat-temperature relationship could mask the kind of step observed for PAA. Specific heat measurements have not been reported heretofore on AAD. The results in combination with data on PAA (22) W. A. Hoyer and A. W. Nolle, J . Chem. Phys., 24, 803 (1956). (23) R. S. Porter and J F. Johnson, J . A p p l . Phys., 34, 55 (1963). (24) E. M. Barrall and E. J. Gallegos, "Depolarized Light and Dif-
ferential Thermal Analyses of Some Polyolefin Transitions," J . Polymer sei., in press.
SPECIFIC HEATSOF LIQUID CRYSTALS
899
I041
1.00-
096-
092-
074-
088-
3
084-
5066v U -
0 80-
t
k.062-
d 8
0 76-
*5 E
. -
2
068-
I W
2-
"
w
-
EU 0 5 8 -
0 72-
v 4
070-
e -
Y
-
n. YI
054-
\
-
0501
064-
\
RAPIDLY COOLEO
-
a
n. n 060-
056-
0.42
0 52-
036
F ISOTROPIC . LIQUIDP
0441
\ MECTIC
CHOLESTERIC
I
SOLID
NEMATIC
SOLID
SLOWLY C O O L E D SOLID
034-
b
IO
20
eo SAMPLE TEMPERATURE, 'C
30
40
50
m
en
ISOTROPIC LIQUID
-
90
100
0 44-
Figure 4. Specific heat of cholesteryl myristate from -3 to 97".
0 40-
0 36-
032""'"""""""'" 40 60 80 100
120
140
160
180
200
220
240
TEMPERATURE, 'C
Figure 3. Specific heat of anisaldazine from 47 to 237".
help to establish the general calorimetric characteristics of nematic mesophases. Cholesteryl Myristate ( C M ) . The specific heattemperature relationship of CM, Figure 4, is noteworthy for the broad transition near ambient temperature and the steep slope of the specific heat curve prior to the solid-smectic transition. The broad transition fits the type H transition of Westrum and McCullough.21 The example chosen as typical by those authors was 1,Zdichloroethane at -93". These broad transitions are easily moved on the temperature axis by rapidly cooling the sample from room temperature. The dashed line in Figure 4 shows the effect of rapidly cooling the solid phase to -60" and then heating. The solid line was obtained by slow cooling, refrigerating, and then heating. Depolarized light studies have indicated that a large amount of structural adjustment occurs on heating prior to the solid-smectic tran~ition.~bThe solid-smectic
transition belongs normally to type 31. The smectic specific heat is almost identical with that of the isotropic liquid. However, it was difficult to supercool the smectic phase into the solid temperature range by more than 20°, so that only a short smectic specific heat curve could be obtained. The slope of the smectic curve is comparable to that expected for a true solid. The specific heat values are approximately the same for the smectic and solid phases for MC as for the previously discussed nematic systems. The transition from smectic to cholesteric mesophase appears to be a type 31 transition as have been the other liquid crystal transitions discussed in this work. Figure 5 shows higher resolution data (5"/min heating rate) for the smectic to isotropic liquid interval for CM. The cholesteric mesophase has a higher apparent specific heat than either the smectic or isotropic liquid. The cholesteric mesophase is a thermally stable mesophase. However, the temperature interval of stability is extremely narrow, and its specific heat curve is continuously changing in slope. Since the smectic(25) E. M. Barrall, R. S. Porter, and J. F. Johnson, Mol. Cryst., in press.
Volume 71, Number 4 March 1967
E. M. BARRALL, 11, R. S. PORTER, AND J. F. JOHNSON
900
~ 0 0 0 -
No evidence of two cholesteric mesophases is apparent in the heat capacity data. Studies using optical microscopy, depolarized light intensity, and dta have suggested two transitions in the cholesteryl mesophase.25 It is possible that the specific heat differences between two possible modifications of the cholesteryl helix may be too small to detect by the present calorimetric methods.
5 om-
Discussion
OW-
092-
088-
004-
e.5
s Y012-
2 . . k 0680.
lZDL 0.56
-SMECTIC+CHOLESTERIC
Ji- ISOTROPIC
LIQUID
-
0.40
Figure 5. Specific heat of cholesteryl myristate in the liquid crystal and isotropic liquid ranges.
cholesteric and the cholesteric-isotropic transitions are completely reversible with little supercooling, a more detailed study of the cholesteric mesophase specific heat over a broader temperature range has been impossible. It is not surprising that the duration of the mesophase is so narrow, since only a few tenths of a calorie separate it from the two neighboring phases.'
The Journal of Physical Chemhtsy
The specific heat curves of nematic, smectic, and cholesteric types of liquid crystals indicate that the transitions from solid to nematic or smectic as well as interphase transitions are not simple first-order phase changes. Using the classification system set up by Westrum and McCullough,21it appears that most of the transitions are either type 21 or 31. The premelting phase transition of PAA reported from other s t u d i e ~ has ~ ~ 'been ~ noted as a sharp step in the present work. The significance of the type 31 or 21 transitions structurally is that the phase changes proceed through chain rotation comparable to that observed prior to the true melting of the higher n-paraffins. I n addition to the liquid crystal transitions, CM exhibits a movable type H transition near 0". This transition is indicative of an intercrystal adjustment of the kind observed for l12-dichloroethane and certain simple metal chelates. This is probably a consequence of the platelike structure of the crystal. Further work is in progress on the specific heattemperature relationships of more complex cholesteryl liquid crystals.