Max-Planck-Institüt für Medizinische Forschung, D-6900 Heidelberg, Germany. (Rep le 19 août 1981, révisé le 26 octobre, accepte le 28 octobre 1981). Résumé ...
J.
FTVRIER 1982,1
Physique 43 (1982) 421-430
421
Classification
Physics Abstracts
61.30- 76.70 61.30
2013
76.70 K
A deuterium NMR study of solute molecules dissolved in the discotic mesophase of p-n-hexahexyloxytriphenylene D.
Goldfarb, Z. Luz
The Weizmann Institute of Science, 76100 Rehovot, Israel
and H. Zimmermann Max-Planck-Institüt für Medizinische Forschung, D-6900 Heidelberg, Germany
(Rep le 19 août 1981, révisé le 26 octobre, accepte le 28 octobre 1981)
Résumé. 2014 On présente les spectres RMN du deutérium de douze composés deutérés, principalement aromatiques, dissous dans la mésophase discotique de p-n-hexahexyloxytriphénylène normal. Les splittings quadrupolaires des deutériums du soluté sont en général beaucoup plus petits que ceux observés dans des cristaux liquides thermotropiques et en même temps sont très fortement dépendants de la température, changeant souvent de signe lorsque l’on fait varier la température dans la région de la mésophase. Les résultats sont interprétés par un modèle dans lequel les molécules du soluté peuvent occuper deux sites ayant des paramètres d’ordre de signes opposés. On propose d’identifier les deux sites (I) à des molécules intercalées dans la structure en colonnes, et (II) à des molécules situées dans la région aliphatique des chaînes mésogènes latérales. Les résultats sont en accord avec un modèle dans lequel un équilibre dynamique rapide (à l’échelle de temps RMN) s’établit entre les deux sites, lequel est déplacé du site I au site II par un accroissement de température. Ces résultats indiquent que les paramètres magnétiques sont différents pour chaque site. Deuterium NMR spectra of twelve, mostly aromatic, deuterated compounds dissolved in the discotic mesophase of normal p-n-hexahexyloxytriphenylene are reported. The solute deuterium quadrupole splittings are in general much smaller than normally observed in thermotropic liquid crystals and at the same time they exhibit a conspicuous dramatic temperature dependence, often changing sign as the temperature is varied within the mesophase region. The results are interpreted in terms of a model in which the solute molecules can occupy two sites having order parameters of opposite signs. The two sites are tentatively identified with (I) molecules intercalated within the columnar structures, and (II) molecules within the aliphatic region of the mesogen side chains. The results are consistent with a model in which there is fast dynamic equilibrium (on the NMR timescale) between the two sites, which shifts from site I to site II with increasing temperature. The results also indicate that the magnetic parameters in the two sites are different. Abstract
-
1. Introduction. In a previous publication [1] we have reported deuterium NMR measurements on the discotic mesophase of p-n-hexahexyloxytriphenylene (THE6). This compound belongs to a novel class of liquid crystals composed of disc-like, rather than rodlike molecules, which exhibit new types of mesomorphic phases [2, 3]. Several different types of discotic mesophases have been identified, including nematic and various columnar smectic mesophases [4-9]. The mesophase of THE6 belongs to the Dho class [ 10,1] whose structure is characterized by columns of regularly stacked molecules, arranged in an hexagonal array (Fig. 1). The deuterium NMR measurements on this mesophase indicate [1] that the orientational -
core of the molecule is and only weakly temperature dependent, while the side chains seem to be quite disordered. These properties suggest that the discotics could serve as hosts for probe molecules with a variety of orientational ordering, depending on the nature of the guest compounds. We would, e.g., expect that aromatic probes will prefer to intercalate between the host molecules within the columns and thus achieve a high degree of orientation, while aliphatic compounds would prefer to dissolve within the side chain region with a considerably smaller orientational order. The purpose of the present work is to investigate
order of the
rigid (aromatic)
relatively high (0.90-0.95)
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphys:01982004302042100
422
these
expectations experimentally emphasis on aromatic probes. As
with particular in our previous
study we use deuterium NMR of deuterated probe compounds. Preliminary proton NMR measurement with normal probe molecules gave unresolved and unintelligible spectra. On the other hand, the deuterium spectra are usually well resolved and can be readily interpreted in terms of the molecular ordering. The results were quite surprising in that the orientational order parameters of all molecules studied, including large fused-ring aromatics, were very low compared to corresponding values in usual thermotropic liquid crystals, and at the same time they were also strongly temperature dependent, often changing signs within the mesophase temperature range. We interprete these results in terms of a model in which the solute molecules diffuse rapidly between various sites of different orientational order. Such models were employed previously for liquid crystalline solutions [11, 12] but it appears that in the present case of the Dho discotic mesophase the effect is particularly
pronounced. 2. Basic equations A schematic diagram of the columnar mesophase Fig. 1. describing the stacking of molecules into columns, and their hexagonal arrangement.
The deuteand definitions. rium NMR spectrum in liquid crystalline phases is dominated by the quadrupole interaction. The spectrum of a single deuteron consists of a doublet with an overall splitting given by [13] :
This equation applies to a uniaxial mesophase as e.g. the discotic phase under consideration, and a molecule with D2 or C2, symmetry. The term e2 qqlh is the deuterium principal component of the quadrupole interaction tensor assumed to be along the C-D bond direction is the asymmetry parameter, 11 ( V n~ - V,)/V,4, and Saa, Sbb and S~~ are the elements of the ordering matrix of the probe molecules. The axes q, ~, ~, are the principal direction of the quadrupole tensor (where ~ is taken to coincide with the C-D bond direction) and a, b, c are the molecular fixed coordinates such that c coincides with the molecular C2 (or higher C) axis, and ac, bc are symmetry planes. Finally, a, #5 y are Euler angles which transform the 11(Ç coordinate system to that of abc, and 00 is the angle between the magnetic field direction and the director. From our previous studies [1]J on the neat mesogen and on a benzene solution we know that under the experimental conditions used in the present work i.e. in which the samples are allowed to cool slowly inside the magnetic field from the isotropic liquid (see the Experimental Section) the mesophase consists of many domains whose directors lie in a plane perpendicular to the external field,
and are randomly distributed in this plane. Thus for all domains 00 in equation (1) is n/2 and the term ’(3 cos2 00 - 1) becomes - 2. Most of the solutes on which we report below are aromatic compounds. For the final presentation of , the results on these compounds we have chosen a new set of labelling, x, y, z such that z is the normal to the molecular plane and x is the direction of the shortest dimension within the plane. We shall be mainly concerned with the S~~ element of the ordering matrix. Thus S~~ > 0 means that the molecular plane of the solute prefers to align parallel to the mesogen mole0 corresponds to the situation in cules while Szz which the solute molecules prefer the perpendicular orientation. For molecules having a symmetry axis Cn with n > 3 we take the direction of this axis parallel to z so that only the element S~~ is required to describe the molecular orientation. For the calculation of the ordering matrix elements we used the quadrupole parameters given in table I. These parameters were determined in various solids and liquid crystalline solvents and they depend slightly on the system used for the measurements. However the variations in these parameters are quite small
-
=
-
423
The list of deuterated solutes and data THE6 solutions which were studied in the concerning the present work.
Table I.
-
under reduced pressure were used. Before the actual NMR measurement the samples were heated several times to above the clearing point and shaken vigorously in order to achieve a homogeneous solution. They were then placed into the preheated NMR probe head and the temperature slowly lowered to the desired value. Depending on the probe concentration, between 100 and 2000 FID signals were accumulated.
4. Results and discussion. In this section we present the results of the deuterium quadrupole splittings of the various probe molecules investigated. We first consider the more symmetric probe molecules, i.e. benzene, triphenylene, acetonitrile and cyclohexane for which a single ordering matrix element is sufficient for the analysis of the quadrupole splitting. These examples will bring out the main points characterizing the behaviour of solute molecules in the mesophase. We then describe results for several other probe molecules with lower symmetry (D2 and C2,) for which two elements of the ordering matrix are required for the analysis. -
(*) Numbers in brackets correspond to the references for the quadrupole interaction constant. and did not significantly affect the resulting values of the S~~’s. In practice the term in q made a very little contribution to the final results and therefore for q gg 0.05 it was generally neglected. For the case of pyridine (vide infra) for which more extensive experimental values were obtained, all q-values were included in the computation of the S;;’s.
Experimental. The synthesis of p-n-hexahexyloxytriphenylene (THE6) is described in references [4] and [1]. In the neat form its phase transition temperatures are 68 OC and 99 ~C for solid to mesophase, and mesophase to isotropic respectively. Solutions of probe molecules in THE6 were prepared by adding weighted amounts of solutes into known quantities of mesogen. The concentrations used ranged between 2 and 7 wt. %, and resulted in lowering of the clearing temperature by several degrees. Some of the deuterated compounds (benzene, naphthalene, triphenylene) were obtained commercially while others were leftovers from previous studies in the Heidelberg laboratory. The list of compounds studied, the solute concentrations and the corresponding clearing temperatures are summarized in table I, in which the quadrupole interaction constants used for the calculation and the corresponding references are also indicated. 3.
-
MEASUREMENTS. - The NMR measurements were performed on a Bruker WH-270 spectrometer operating at 41.45 MHz for deuterium, and using the pulsed Fourier transform mode. The temperature was controlled with a BST 100/700 unit and its absolute value was calibrated using a Fluke 2190A digital thermometer. For convenience of presentation the results in each case are given relative to the corresponding clearing temperature, Tc. Five millimeter sample tubes containing approximately 0.2 g of solvent plus solute which were sealed
4.1 BENZENE-d6 : THE TWO SITE MODEL. Examof deuterium NMR a of wt. 5.0 ples spectra % solution of deuterated benzene (consisting of a 2 : 3 ratio C6D6 : C6H6) in normal THE6 at different temperatures are shown in figure 2. The same general behaviour of the splitting was observed in two other solutions that we have studied viz. containing 1.6 and 6.6 wt. % benzene-d6. The magnitude of the observed splitting,vQ 19 in the latter solution is plotted in the upper part of figure 3a versus the temperature difference, T~ - T, from the clearing point. Since ben-
NMR
of a 5.0 w~ % solution of deuterated benzene in THE6 at different temperatures. The deuterated benzene consists of a 2 : 3 ratio of
Fig. 2. - 20 NMR spectra
C6D6 : C6H6.
424
temperature. The two possibilities are plotted in the lower part of figure 3a. In principle one could determine the sign of vQ if the spectrum would exhibit
Fig. 3. The experimental quadrupolar splittingvQI and the derived order parameter Szz for (a) 6.6 wt. % benzene-d6 and (b) 3.1 wt % triphenylene-d12 in THE6 respectively. The two sets of data in the lower parts of the figure correspond in each case to the two possible signs of vQ. For triphenylene Szz was calculated from the deuteron data of position 2. -
features due to other interactions for which the absolute signs are known, e.g. the indirect coupling between the deuterons. However the resolution of the deuterium spectrum is much too low for these interactions to show up. We have tried to record the 1 H NMR spectrum of normal benzene dissolved in THE6 with the hope that it will provide the sign of the dipolar interaction, but were unable to obtain high resolution proton spectra of this solute. To explain the results of the benzene solutions we propose a model in which the solute molecules occupy several sites with different ordering matrix elements undergoing fast (on the NMR timescale) dynamic equilibrium. In principle there could be many « sites » representing different structures and different solutesolvent complexes. However in order to explain the main features of figure 3a it is sufficient to consider just two sites having S~~ values of opposite signs, and a population ratio that is temperature dependent The average order parameter is then given by :
C6 axis of symmetry the expression for vQ (taking fl n/2 in equation (1)) reduces to where
zene
has
a
=
so that the order parameter, Szz, is linearly related to vQ. The behaviour of vQ for the benzene probe is considerably different from that of the discotic bulk solvent, and is quite unexpected : while for the aromatic deuterons of the neat mesogen, large and essentially temperature independent quadrupole splittings have been observed [1], the magnitude of the benzene splitting is very small even compared to that found for benzene in normal thermotropic liquid crystals, and at the same time its temperature dependence is quite dramatic in particular note the vanishing of vQ at a particular temperature within the mesophase region. This temperature does not correspond to any mesomorphic or other discontinuous change in the neat mesogen. Also we have checked the deuterium splittings of a perdeuterated mesogen solvent containing 5 wt % benzene and found that they were essentially the same as in the neat mesogen and only very slightly affected by the presence of the solute. Since within the mesophase region we expect the quadrupole splitting to vary continuously, the vanishing of vQ for the benzene solute must correspond to a point at which its order parameter changes sign. Since the sign of vQ is not known it is not possible to tell in which direction it crosses the zero line, i.e. whether vQ increases or decreases with increasing -
the superscript I and II refer to the two sites, and we assumed that e2 qQ/h is the same in both of them. For the discotic columnar mesophase this is a very plausible model on chemical grounds : one site (site I) would correspond to benzene molecules intercalated within the discotic columns, with S’,, > 0, since the benzene molecules will most likely tend to lie parallel to the aromatic planes of the mesogen molecules. The second, less favourable, site (site II) is identified with the « aliphatic » region between the columns which is the space occupied by the side chains of the mesogen molecules. To be consistent with the experimental results we must assume that S’i is negative. This conclusion is more difficult to rationalize and it indicates that in the aliphatic region the benzene molecules prefer an orientation in which their plane is perpendicular to that of the triphenylene
rings. For a more quantitative analysis of the results we I express the population ratio P II~P of the two sites by an Arrhenius equation of the type
where OH is a positive enthalpy difference (H" - H’) between the two sites, and K a weighting factor related to the entropy difference between them. To simplify the analysis we assume that the temperature dependence of S~~ comes mainly from changes in P’ and PIT while S:z and Siz are temperature independent. The latter parameters reflect the ordering in the various sites of the mesogen molecules, which as indicated
425
above
found to be only very weakly temperature dependent. With this assumption the ratio Szz/Siz can be expressed in terms of a temperature To at which vQ vanishes [11] : was
substituting in equation (3) and using and the relation PI + P" 1 gives
equation (4)
=
where the second (approximate) equality applies since we consider the region where T - To. Since we assumed OH and S:z to be positive equation (6) predicts that vQ would change from positive to negative as T changes from below to above To. Thus we adopt the plot indicated by a full line in figure 3a rather than the one with a dashed line. We were unable to obtain a meaningful quantitative fit of the results of figure 3a to equation (6) using all three parameters, S’z, AH and AS ( = R In x) as free variables. Fixing S§z in the range 0.05 to 0.8 gave best fit values for AH and AS which varied in the narrow ranges 4.1 to 2.4 kcaL/mol-1 and 11.0 to 13.0 e.u. respectively. For Slz lower than 0.02 unreasonably high values of DH and AS were obtained, and the fit did not converge. The values obtained for AH and AS in the converging region seem quite reasonable : the result for AH is typical for solutesolvent interaction and in our model it reflects the specific interaction of the benzene with the aromatic core of the mesogen molecules in site I, while the high value of AS reflects the higher degree of disorder in site II. The actual magnitude of the parameters should however not be taken too seriously due to the simplified model used and the many assumptions made. 4. Z
TRIPHENYLENE-d12 : SITE DEPENDENT MAGNETIC Examples of deuterium spectra of a 3.1 wt. % solution triphenylene in THE6 and a plot of the corresponding splittings versus temperature are shown in figures 4 and 3b. The two peaks in the spectrum correspond to positions 1 and 2 as indicated in the figures. It may be seen that the centres PARAMETERS.
-
of these doublets do not coincide. This asymmetry is due to the chemical shift difference between the two positions, and on the basis of previous NMR measurements [16] in isotropic liquids the high field doublet was identified with position 2. We have taken the data of this nucleus for the calculation of
Fig. 4. As in figure 2 for a 3.1 wt % solution of triphenylene-d12 in THE6. The numbering of the peaks correspond to the deuteron positions as indicated in the structural formula in figure 3. -
S~~ in the lower part of figure 3b. The general behaviour of vQ and thus of S~~ is very similar to that of benzene although the overall changes in the splittings are much larger. A similar analysis as done above for the benzene (Eq. (6)) using the deuteron data of position 2 gave converging results of 2.13.7 kcal./mol- 1 for AH and 6.3-8.4 e.u. for AS in the range 0.2 S:z 0.8. For lower values of Siz no convergence was obtained. Although the results for the thermodynamic parameters seem reasonable, the resulting values for Sii are about - 0.6, which is just outside the allowed region for motional constants. We have so far disregarded the difference in the quadrupolar, splitting between the two positions 1 and 2. On first sight it might appear that this is due to a small difference in the factor eZ qQ/h in the two deuteron positions. However it is clear that this cannot be enough. Firstly because the difference in splitting is quite large compared to the overall splittings : when triphenylene is dissolved in normal thermotropic liquid crystals its deuterium NMR also exhibits a small difference in splittings for positions 1 and 2, but this difference amounts at most to a few tenth of a percent However most significant is the fact that the quadrupole splittings of the two deuterons do not vanish simultaneously, while from equa0 should be tion (3) the temperature at which vQ of the independent quadrupole parameter and should coincide for both deuterons. =
426
To account for this anomalous effect modify our two site model and allow the
slightly quadrupole we
interaction parameter A = - 3 e2 qQ (3 cos’ # - 1)/8 h (Eq. (2)), for each of the deuterons to be different in the two sites, so that equation (3) now reads
where i refers to the deuteron position in the triphenylene. It is clear that if the change in A on going from site I to II is different for the two deuterons, i.e. A lI/Ai 5~ Ai’/A2 then the vanishing of v’ will not occur at the same temperature as that for vQ. The number of parameters in the equations for vQ is now much too large to attempt a best fit analysis of the experimental results. An order of magnitude estimate indicates that a difference of a few percent in the change of A on going from site I to II can explain the observed results. Whether this change is due to redistribution of the electronic charge in the probe molecules or whether it is due to changes in the molecular geometry on going from one site to the other cannot be decided from the experimental results. It should be noted that similar effects i.e. site dependent geometry changes of solutes in liquid crystalline solvents were noticed previously in the NMR spectra of several molecules [11, 12]. 4.3 ALIPHATIC
PROBES:I
CYCLOHEXANE- d 12,
ACETO-
HEXYLALCOHOL-1 d2 . - The deuteNITRILE- d3 rium NMR of each of these compounds in the THE6 mesophase solution consists of a single doublet. The temperature dependence ofvQ ~I derived from the spectra is plotted in figure 5. For acetonitrile and hexylalcohol the results are similar to those obtained in normal thermotropic liquid crystals, i.e. a gradual AND
decrease in vQ with increasing temperature, although the splittings are relatively small. It seems that these probes dissolve predominantly in the aliphatic region (site II) where the magnitude of the ordering is low. The order parameter for acetonitrile which has a C3 axis of symmetry was calculated from the equation
have taken ~3 to be equal to the tetrahedral Since the sign of vQ is not known and there is no obvious way to guess it on chemical grounds we have used equation (8) to calculate the magnitude of Szz as shown on the right ordinate in figure 5. A similar calculation for hexylalcohol is not possible because of its low symmetry and high flexiwhere
we
angle [17].
bility. The quadrupole splitting of the cyclohexane probe is even smaller than for the other examples in this section and its temperature dependence is abnormal, i.e. the splitting increases with increasing temperature. As for the cases of benzene and triphenylene we interprete this behaviour in terms of the two site model in which the increase inI vQI is due to the equilibrium shift from site I to site II, which have opposite sign for Szz. Here too it is difficult to guess the preferred orientations in the two sites and accordingly we plot the magnitude of S~~ for this compound. Since only a single doublet is observed for the cyclohexane deuterons due to fast ring interconversion we use their average quadrupole constant for the calculation of SZZ[ 18] :
where in the last equality (In reality it is 2.60 [18].) 4.4 FUSED-RINGS
we
approximated ~3 by
0.
D2 NAPHTHALENE-d8, ANTHRACENE-d10 AND PYRENE-djo. In the remaining part of this paper we shall discuss probe molecules with symmetry C2~ or D2 and thus two ordering matrix elements are required to describe their ordering in the mesophase. To derive these constants from equation (1) we need AROMATIC COMPOUNDS WITH
SYMMETRY :I
-
Fig. 5. The deuterium quadrupole interaction constants I vQI for cyclohexane-d, 2, acetonitrile-d3 and hexylalcohol-I d2 and the correspondingSzzI for the first two compounds. Note the change in scale for the cyclohexane -
data.
data from at least two inequivalent deuterons. Since the signs of the vQ’s are not known there are four possible sets of solutions for the S;;’s, corresponding to the following four possible combinations for the two vQ’s : (a) + + ; (b) - - ; (c) + - ;
quadrupole
427
(d) - +. The results of the S;;’s for cases (a) (b) are identical in magnitude but opposite in sign, and similarly for cases (c) and (d). We shall therefore present only two sets of results for each compound - the other two possibilities are obtained by inverting the signs of S~~ and of S.,.,-Sy. The sets of results that we shall prefer to present will be those that seem more likely on the basis of the model described in the beginning of this section, i.e. we shall prefer the solutions for which Szz becomes algebraically smaller (more negative) with increasing temperature. This is the expected change in Szz if we assume that upon heating, more of the intercalated molecules shift to ’the aliphatic chain region. Examples of spectra for perdeuterated naphthalene, anthracene and pyrene are shown in figure 6. Note that for the latter two compounds only two doublets are observed due to the effective equivalence of positions 1 and 9 in anthracene and positions 2 and 3 in pyrene. This equivalence also serves to identify the NMR signals. In naphthalene the of naphthalene-d, peak assignment was made on the basis of the small Fig. 6. 2D NMR spectra of solutions and (64 DC), °C~ (72 pyrene-d 10 (62 °C’~ in anthracene-d 10 chemical shift [19] between the two doublets as well where the numbers in brackets correspond to the THE6, as by comparison with the spectrum of naphthalenewere recorded. The at which the 1 di. The temperature dependence of the various temperatures of the solutionsspectra are given in table I and the compositions I vQ ~’s of these probes are shown in the upper dia- labelling of the deuterium positions are indicated in the grams of figure 7 and the corresponding calculated structural formulae in figure 7. S~~’s (for two sets of signs as indicated in the figure and the magnetic parameters [20, 15] indicated in pounds have C2, symmetry and three inequivalent table I) are plotted underneath each case. deuterons, due to the deviation of the ring structure 4.5 AROMATIC PROBES WITH C2, SYMMETRY : PYRI- from perfect hexagon. Indeed as may be seen in DINE- d s AND NITROBENZENE- d:5. - These two com- figure 8 three distinct doublets appear in the deuteand and
-
Fig. 7. The deuterium quadrupolar splittings for the same solutions as in figure 6, and two out of the four sets of the ordering matrix elements for each compound. The other two sets are obtained by inverting the signs of all S;;. -
I v.
possible
428
9. The deuterium quadrupole splittingI vQI for solutions of pyridine-ds and nitrobenzene-ds in THE6, and several possible sets of Si’s for each compound The composition of the solution is given in table I.
Fig.
Fig. 8. As in figure 6 for pyridine-ds (67 ~C~, nitrobenzene-ds (52 OC), o-xylene-dio (74 OC) and p-xylene-d,, (61 °Q. The deuteron labelling are as in figures 9 and 10. -
-
4.6 PARA- AND ORTHO-XYLENE-d10. - The situation here is similar to that as for the molecules discussed in subsections 4.4 and 4.5 except that the CD3 groups are considered as a pseudo-aromatic deuteron with an effective quadrupole interaction constant
rium spectra of these compounds. The corresponding plots ofI vQI versus temperature are depicted in figure 9. The identification of the peaks was based
their relative intensities and chemical shifts. Since there are now three sets of data one might hope that certain choices of relative signs for vQ can be eliminated For nitrobenzene this is however not the case since as may be seen from the plots ofI vQ 1, positions 1 and 2 are nearly equivalent The situation is thus similar to that for the group of probes naphthalene, anthracene, pyrene. The results for SzZ plotted in figure 9 were calculated using geometrical parameters and e2 qQ/h values given in the literature [21, 22], and assuming that the C-D bonds lies along the direction of the corresponding CCC bisectors. For pyridine on the other hand we could perform a more complete analysis : firstly because both deuterium quadrupole [23] and geometrical [24] data for this compound are available, and secondly because as may be seen in figure 9 the vQ’s for the various deuterons in the mesophase solution seem to be changing independently. Indeed from the four possible pairs of sign combinations only one pair was consistent with all experimental data and the solution for ‘which Szz decreases algebraically is shown in the bottom part of the figure. on
The deuterium quadrupole splittingI vQI for solutions of o-xylene-dio and p-xylene-d,, in THE6, and possible sets of Sii’s for each compound. The composition of the solutions is givenrin table I.
Fig.
10.
-
429
In
reality there may be a large number of discrete perhaps a continuous range of sites. It is not possible to identify or characterize such sites from the NMR data, but the picture of two sites corresponding to intra- and inter-columnar regions seems attractive. Also the semi-quantitative estimate for 5. Summary and conclusions. The orientational the enthalpy and entropy difference between the two order of solute molecules dissolved in the discotic sites is consistent with the above interpretation. mesophase of THE6 as measured by 2D NMR has Perhaps some support for the two-site model can several unexpected features : the magnitude of be obtained by performing similar experiments in the solute order parameters is considerably smaller polymorphic discotics which exibit both a columnar than in normal thermotropic liquid crystals despite phase and a nematic phase in which there is no (or the fact that the ordering of the triphenylene moiety much less) molecular stacking. We would then of the mesophase molecules is very high. Moreover expect that in the columnar phase the probe molewhile the order parameter of the neat mesophase is cules behave as in the THE6 mesophase while in the essentially temperature independent that of the solute nematic phase the characteristic features of two molecules changes dramatically with temperature sites will not appear or will be much less pronounced and often changes sign within the mesophase region. At the moment we do not possess materials suitable These effects can be qualitatively explained in terms for such experiments. of a model involving two solvation sites with oppoThe analysis of the results indicates a further site signs of motional constants, and assuming fast important fact, namely that the probe molecules diffusion of the probe molecules between the sites. It might have different magnetic parameters in the two is natural to associate (although impossible to prove) sites. In fact for the particular probe of perdeuterated these two sites with : (1) probe molecules intercalated triphenylene the results can only be understood within the stacked columns of the mesogen, and (II) under the above assumption. Whether the difference molecules dissolved in the space between the columns in the magnetic parameter in the two sites reflect which is occupied by the aliphatic chains of the redistribution of electronic charges due to different mesogen. Planar aromatic molecules will tend to solvent solute interaction or perhaps geometrical align parallel to the mesogen molecules in site I changes on going from one site to the other cannot while in site II they will be highly disordered and be inferred from the experimental results. Geomeapparently prefer a perpendicular orientation. The trical changes in molecules on changing sites has strong temperature dependence of the probe’s orien- been invoked previously [11, 12] to explain anomatational order is then attributed to a temperature lous dipolar splittings in liquid crystals, but it seems driven shift of equilibrium between the two sites. that the effect on the quadrupole splitting is consiSpecifically we assume that at low temperatures site I derably more significant (intercalation) is preferred while at high temperawhere a is the angle between the methyl C-D bond and the local symmetry axis of this group. Examples of spectra, experimental values forI vQI and the derived possible values for the elements of the ordering matrix are given in figures 8 and 10.
or
-
entropy effects drive the molecules to site II the side chain region). It is important to emphasize (i.e. that the two-site model represents the simplest one that can be made to fit the experimental observation. tures
This research was supported Research Council of Israel and by the National by the U.S.-Israel Binational Science Foundation.
Acknowledgments.
-
References
[1] GOLDFARB, D., LUZ, Z. and ZIMMERMANN, H., J. Physique 42 (1981) 1303. [2] CHANDRASEKHAR, S., SADASHIVA, B. K. and SURESH, K. A., Pramana 9 (1977) 471. [3] BILLARD, J., DUBOIS, J. C., TINH, N. H. and ZANN, A., Nouv. J. Chimie 2 (1978) 535. [4] DESTRADE, C., MONDON, M. C. and MALTHETE, J., J. Physique Colloq. 40 (1979) C3-17. [5] TINH, N. H., DESTRADE, C. and GASPAROUX, H., Phys. Lett. 72A (1979) 251. [6] DESTRADE, C., MALTHETE, J., TINH, N. H. and GASPAROUX, H., Phys. Lett. 78A (1980) 82. [7] DESTRADE, C., TINH, N. H., MALTHETE, J. and JACQUES, J., Phys. Lett. 78A (1980) 189. [8] LEVELUT, A. M., J. Physique Lett. 40 (1979) L-81. [9] BILLARD, J., in Liquid Crystals of One and Two Dimen-
sional
Order, eds. W. Helfrich and G. Heppke
(Springer Verlag, Berlin) 1980, [10] DESTRADE, C., BERNAUD,
M.
p. 383.
C., GASPAROUX, H.,
LEVELUT, A. M. and TINH, N. H., in the Procee-
dings of the International Liquid Crystal Conference, Bangalore (1979), ed. S. Chandrasekhar, p. 29.
[11] DIEHL, P., SYKORA, S., NIEDERBERGER, W. and BURNELL, E. E., J. Mag. Reson. 14 (1974) 260; DIEHL, P., REINHOLD, M., TRACEY, A. S. and WULLSCHLEGER, E., Molec. Phys. 30 (1975) 1781. [12] BURNELL, E. E., COUNSIL, J. R. and ULRICH, S. E., Chem. Phys. Lett. 31 (1975) 395. [13] DOANE, J. W., in Magnetic Resonance of Phase Transitions, eds. Owens, F. J., Poole, C. P. and Farach, H. (Academic Press) 1979, p. 171.
430
[14] MANTSCH, H. H., SAITO, H. and SMITH, J. C. P., in Progress in Nuclear Magnetic Resonance, eds. Emsley, J. W., Feeney, J. and Sutcliffe, L. H. (Pergamon Press, New York) 1977, Vol. II, p. 211. [15] ELLIS, D. M. and BJORKSTAIN, J. L., J. Chem. Phys. 46 (1967) 4460. [16] BRÜGEL, W., Handbook of NMR Spectral Parameters (Heyden & Son Ltd.) 1979, Vol. 3, p. 702. [17] ANGLERT, G. and SAUPE, A., Mol. Cryst. Liq. Cryst. 8 (1969) 233.
[18] POUPKO, R. and Luz, Z., J. Chem. Phys. 75 (1981) 1675. [19] Reference [16], Vol. 2, p. 452. [20] DAGRACA, M., DILLON, C. and SMITH, J. A. S., J. Chem. Soc. Faraday Trans. II (1972) 2183. [21] TROTTER, J., Acta Crystallogr. 12 (1959) 884. [22] WEI, I. Y. and FUNG, B. M., J. Chem. Phys. 52 (1970) 4917.
[23] BARNES, R. G. and BLOOM, J. W., J. Chem. Phys. (1972) 3082. J. P. and LEHN, J. M., Mol. Phys. KINTZINGER, [24] (1971) 273.
57 22
