J. Phys. Chem. 1994,98, 2608-2612
2608
A Study of Phase Change in Langmuir-Blodgett Films of Mesomorphic Side-Chain Polymers Louis Bosio laboratoire 'Physique des Liquides et Elecfrochimie", CNRS, ESPCI 10 rue Vauquelin, 75005 Paris. France Patrick Keller and Lay-Theng Lee laboratoire LPon Brillouin (CNRS-CEA), CEA Saclay. 91191 Gifsur Yvette, France Jean-Philippe Bourgoin and Miehel Vandevyver' Seruice de Chimie Moleculaire. CEA Saclay. 91191 Gif sur Yvette, France Receiued: July 6, 1993; In Final Form: October 27, 1993*
Langmuir-Blodgett films of mesomorphic side-chain poly(methacry1ates) are investigated from the point of view of their phase change versus temperature via X-ray and neutron reflectivity together with atomic force microscopy measurements. Upon heating, "thick" films undergo an irreversible transition toward the wellknown smectic A I lamellar phase. In contrast, a bilayer deposited onto a solid substrate appears to be very stable upon annealing. Such a behavior is believed to induce an epitaxial mechanism when cooling "thick" films from the isotropic melt phase.
Introduction The Langmuir-Blodgett (LB) method has been extensively worked out in the last two decades.) It has proven to be a very useful tool for deposition of ultrathin organic films onto a solid substrate. Films deposited in this manner generally exhibit a lamellar structure obviously imposed by the dipping and raising process. Hence, such a structure appear as a periodic molecular array (Le., a molecular order imposed by the LB method itself) along the normal N to the substrate. In the recent past, liquid crystalline compounds have been shown to yield exceptionally well controlled LB films.z The origin of such a behavior is likely to be found in the ability of these compounds to self-organize hence increasing the initial order induced by the LB method? At present, LB films of liquid crystalline compounds have been extensively studied by several authors- and ref 1. In contrast, little has been reportedon thepackingandordervariationsduring their thermally induced phase changes?-12 In a previous work,'] we investigated the phase change in LB filmsofa mesomorphicsidechainpolymer. This polymer, labeled l a in Figure 1, was chosen because of its relatively low phase transition temperature from glassy to smectic AI. More than 50 superimposed monolayers could be obtained when transferred at 25 mN/m despite a poor transfer efficiency (equal or lower than 0.5). The most striking feature observed for "as prepared" films was the lack of lamellar structure usually expected for LB films. In contrast, after heating the film at 55 OC, a lamellar phase stable at room temperature (of periodicity ca. 28 A) appeared andwasidentifiedtobethesmecticAlmesophase. Inthepresent work, we have improved the film quality by choosing more appropriate deposition conditions. This enabled the building up of'as prepared" filmsexhibitinga well-defined lamellar structure not observed before. Phase changes in these films have been investigated by X-ray and neutron reflectivity measurements using polymer laandselectivelydeuteratedpolymerslband le(Figure I). In addition, specific features of film texture were elucidated by atomic force microscopy (AFM). Materials and Method SamplePreparation. Polymers la, lb,and lcweresynthesized using the method given in the literature."Js For the LB *Abstract published in Advonee ACS Abstrocrs. February 1, 1994.
0022-3654/94/2098-2608$04.50/0
I lCllii
I I
0
c=o
I
- c - Clh," I
cn,
la
Ltt
I
Ir
Figure 1. Polymers la, lb. and le have been used in the prescnt work. The average number n of monomers in a polymeric molecule is estimated to lie between 500 and 1000.
experiment, the starting material la is the same as that in ref 13 in which, following refs 14 and 15, we stated the polymorphism to be g-38 oCSAI-106 'C-N-113 OC-I (where g stands for glassy state, SA, for smectic AI phase, N for nematic phase, and I for isotropicliquid). Sincethesevaluesdependon themolecular weight of the polymer, differential scanning calorimetry (DSC) was performed on the material used in the present study. We obtain g43.4°CSA~-99.9 OC-N-109 "C-I for la. Thesevalues are considered to be the appropriate values in the present work as well as that of ref 13. For l b and le we find respectively g 4 4 OCSAI-95.4 OC-N-107.5 OC-I and g47.7 'CSAI-IOS OCN-113 OC-I. Float glass (70 X 80 X 2 mml) is used as substrate for LB deposition. Theglass is first cleaned in boiling ethanol for several hours and subsequently treated by dry oxidation using UV and ozone atmosphere. Just before film deposition, it is rendered hydrophobic by ultrasonic treatment (IO min) in chloroform containing 5% ofdimethyldichlorosilaneand rinsed in chloroform containing a few percent of ethanol (IO min of ultrasonic 0 1994 American Chemical Society
- The Journal of Physical Chemistry, Vol. 98, No. 10, 1994
Phase Change in LB Films of Polymers treatment). The polymer is then deposited immediately onto the glass using the LB method without any additional underlayer. All attempts to deposit the multilayersonto a hydrophilic substrate failed systematically. LB films are built using a large area trough16 at room temperature under a continuous nitrogen flow. The surface pressure is measured by a Wilhelmy plate made of a single strip of filter paper. The subphase is Millipore grade water obtained by reverse osmosis, subsequently deionized (resistivity 18 M k m ) and finally transferred through an activated charcoal filter. Before compression, the surface pressureis always kept below 0.1 mN/m and the compression speed is the same as that in ref 13 (0.5 A2 per monomer X minute). A careful observationof the spreading condition reveals that the film homogeneity is considerably improved for a spreading solution of low molarity (ca. 10-4 M or lower) deposited on several places of the gas-water interface. The average molecular area per monomer is found to be significantly higher than that given in ref 13 where high film viscosity was obtained. In the present work, films are transferred at an unusually low surface pressure of 5 mN/m. The advantage of such a procedure is a reduction in film viscosity and therefore an increase in the transfer efficiency. The average area occupied by a monomer at this surface pressure is found to be 31-35 A2; however, the transfer efficiency remains lower than unity (ca. 0.8) and decreases slightly with increasing number of superimposed monolayers. In order to reduce this latter effect thick films are built by sequencesof 10-12 superimposedmonolayers. X-ray Technique. It has been demonstrated that X-ray reflectivity at grazing incidence can be used to obtain structural information on LB films even for a single layer.17 LB films are considered as a stratified medium in which the density of the matter varies as a periodic function of the Z coordinate measured in the direction of the normal N to the substrate. Variations of the density profile along N produce large modulations of the reflected beam versus the incident angle 8. The reflectivity can be calculated from the standard expression of the classical optics for a stratified medium18 by substituting the following expression for the complex optical index: n = l-b-ip
(1)
6 = (A2/2r)r,y‘+ Af’)p/M
(2)
0 = pA/4r
(3)
where Ais the X-ray wavelength, rethe classical radius of electron, (f Af’) the real part of the atomic scattering factor, p the density, M the atomic mass, and p the linear absorption coefficient.19 The most convenient way is to consider the film as a superposition of homogeneous lamellae.17J9~20 For the values of 6 and /3 ascribed to each lamella, one can probe density profiles by comparing the calculated reflectivity to the one measured. The reflectivity measurements are carried out using an experimental device21 with a conventional Cu Kal X-ray source (A =
+
1.54 A).
Neutron Reflectivity. Classical optical phenomena such as reflection, refraction, and interference are also observed with neutrons. The neutron refractive index of a medium, neglecting absorption, is defined as22 n = 1 - NbA2/2?T N b is the average coherent scattering length density of the medium and A the neutron wavelength. In this study, the reflectivity R is measured as a function of k,, defined as (2r/A) sin 8, where 8 is the grazing incidentangle. As in the case for X-ray reflectivity described above, the optical matrix method is used to analyze the reflectivity data. A F M Technique. The AFM technique, although developed r e ~ e n t l y , ~ ~now . * ~ iaswidely used technique. Oneof its interesting
2609
si----
:I
1
.
.
.
.
.
0
.
.
.
1
s(A-1)
Figure 2. X-ray reflectivity curve versus q = 4s sin Bjh for a “thick”film made of 3 1 superimposed bilayers of IC (deuterated backbone): (a) as prepared sample; (b) after annealingfor 1 h at 57 O C ; (c) after annealing for 1 hat 118OC. Curvecreferstotherightordinatescale(measurements carried out at room temperature). Due to the reduced scale, Kiessig fringesarenotvisibleinthelowqrange. In thebulkmatcrial, thetransition toward the isotropical melt occurs at 113 OC.
features is the possibility for real space investigation. The AFM25 operates in the socalled “height” mode (Le., tip tosample distance is adjusted by a feedback to keep the force constant). Imaging was done with a 15 X 15 pm scanner, with a scanning speed in the range 3-8 Hz. Both microfabricated silicon nitride (spring constants in the range 0.1-0.4 N/m) and silicon (spring constants in the range 0.05-0.6 N/m) cantilevers were used. Samples were imaged under water in order to reduce the applied force. For the most part of the experiment, a force load lower than 3 nN was used. Reclult.9 X-ray Reflectivity. Figure 2(a-c) gives the X-ray reflectivity curves versus q = 4 1 sin 8/A for a “thick” film made of 31 superimposed bilayers of l c (deuterated backbone). Before heating (Figure 2a), three main Bragg peaks are visible at q = 0.147,0.296, and 0.446 A-l together with Kiessig fringes of weak intensities in the vicinity of q = 0.06 A-l. The rapid damping of the Kiessig fringes is a signature of some inhomogeneity of the film thickness. However, the angular difference between two successivemaxima can be measured. The film thickness is found to be e = 1350 50 A. Bragg peaks are rather broad, showing a noticeable dispersion in the interplanar distances. However, the 001,002, and 003 reflections lead to relatively well-defined average reticular distances, Le., 42.5,21.2, and 14.1A (corrected for the refractive index in the Bragg formula) and the corresponding parameter is 42.4 0.1 A. Such a finding is obviously consistentwith the presence of an ordered phase, unknown before, built up from IC by the use of the LB method. In its most extended configuration, the length of a mesogenic side chain is close to 28 A. The value of 42.4 A obtained for the interplanar distance necessarily involves coupling of adjacent monolayers in a Y-type configuration. In addition, this result agrees with the observed behavior of the meniscus during the transfer which leads to the clear conclusion that the transfer is of Y type. We conclude that 42.4 A is the thickness of a bilayer, the total thickness of the film being 62 X (42.4/2) = 13 14 A, in close agreement with the above value e = 1350 f 50 A deduced from the Kiessig fringes. The thickness of the individual monolayers (21.2 A) is substantially lower than the length of a stretched side chain (28 A). This is understood in terms of a pronounced average tilt (ca. 40°) or a more or less “coiled” situation for the mesogenic side chains. Such a behavior is not surprising, keeping in mind that the film has been transferred at an unusually low surface pressure and that, consequently,the molecular array is not expected to bevery compact. In the following, films are heated at a given temperature
*
*
2610 The Journal of Physical Chemistry, Vol. 98, No. 10, 1994
Figure 3. X-ray reflectivity curve versus q = 4r sin @/A for a 'thin" film made of six superimposed bilayers of le: (a) as prepared sample; (b) after annealing for 1 h at 57 OC;(c) after annealing for 1 h at 118 OC (measurements carried out at room temperature). and subsequentlyinvestigated by X-ray in a frozen, indefinitely stable state at room temperature. After heating for 1 h at 57 "C (Figure 2b), the Bragg peaks strongly decrease but still remain clearly visible while inspection of the Kiessig fringes does not reveal any significant change in the film thickness. In contrast, a new Bragg peak can be detected at q = 0.224 A-1 which correspondsto the well-known interplanar distance of the smectic A, phase in the bulk material. After heating for 1 h at 97 "C(Figure 2c), the previous Bragg peaks completely disappear while two new ones (001, 002) are observed at q = 0.224 and 0.446 A-I. As expected these peaks correspond to the reticular distance observed in the bulk material, Le., 28.2 f 0.1 A for the smectic A1 phase. In addition, they are very strong (note the change in the ordinate scale) and very sharp hence showing a considerable increase in the molecular order along the normal to the substrate. At the same time, inspection of the Kiessig fringes located in the low q region reveals a moderate decrease in the film thickness (5-10%). This is understood in termsof theincreasein the film compactness. It is worth noticing that no further change in the reflectivity curve is recorded after the film has been heated for 1 h at 118 "C,Le., 5' beyond the N to I transition temperature. The behavior of "thin" films is basically not very different from that of "thick" ones. Figure 3a shows the reflectivity curve of a film made of 6 superimposedbilayers of IC. Before heating, the Kiessig fringesare not very strong and their intensity decreases rapidly as q increases. Again, we conclude that the homogeneity of the film thickness e is relatively poor. Nevertheless,inspection of the reflectivity curve in the low-q range allows a determination of e. We find e = 275 f 25 A. On the other hand, maxima located at q = 0.142 and 0.283 A-1 are proof of a periodic structure with a periodicity of 44.6 f 0.1 A. As above, we conclude that a Y-type structure is formed. The peaks are broad and relatively weak showing again some irregularity in the stackingof successive monolayers. This reflects the problems, already pointed out, occurring in the building up of the films and also some difficulties in obtaining reproducible samples, as seen in the value of periodicity of 44.6 A instead of 42.4 A for the 'thick" films previously studied. However, the film thickness deduced from the periodicity, i.e., 12 X (44.6/2) = 268 A, is in quite good agreement with that deduced from the Kiessig fringes. After heating the sample at 57 "C for 1 h (Figure 3b), the Bragg peaks almost disappear while the film thickness does not vary. At the same time, a new Bragg peak is visible at q = 0.23 A-I showing the appearance of the smectic A1 phase. After a subsequent heating at 97 "C (1 h) the Bragg peak shifts to q = 0.226 A-l while the film thickness decreases by 3-5%. As for "thick" films no significant change is recorded when the film is
Bosio et al.
-
Figwe 4. X-ray reflectivity curve versu q 4r sin @/A for a bilayer of l e deposited onto a hydrophobic float glass subetrate: (X) as prepared film; (+) after annealing for 1 h at 57 O C ; (squares) after annealing for 1 h at 118 OC (measurements carried out at room temperature); (solid line) best fit using the electron density profile given in inset. The filmsubstrate interface appearsas a narrow "dip", 5.9 A thick, reflecting the misfit between the methyl group grafted on the glass substrate by the dimethyldichlorosilanetreatment and thosebelongingto the mesomorphic side chains. heated for 1 h at 118 "C (Figure 3 4 . The lack of homogeneity in both the film thickness and the stacking of superimposed monolayers does not allow an accurate fit of the reflectivity curves given in Figure 3. Investigationof a single bilayer is most interesting and brings out an entirely new result which can be summarized as follows. Reflectivity curves for a single bilayer of IC recorded at room temperature and after successive heating for 1 h at 57, 97 and 118 "Cremain basically the same (Figure 4). From the present data we can already conclude that the LB method produces a particular molecular array for which the phase transition g I) SA1 does not occur. In this case, an accurate fit can be made. We first investigatethe glass coated with the hydrophobicmethyl monolayer from dimethyldichlorosilane treatment. The substrate is made of the bulk glass (6, = 7.74 X 10-6and 6, = 1.3 X coated with a surface layer of slightly modified glass, 30 A thick, as a result of the surface treatment with 61 = 7.31 X 10-6 and 81 = 1.31 X l t 7 . The film-substrate interface appears as a narrow "dip", 5.9 A thick, reflecting the misfit between themethyl groups grafted on the glass substrate by thedimethyldichlorosilane treatment and those belonging to the mesomorphic side chains. In order to reduce the number of flexibleparameters,we postulated that the electron density 6, is constant within the bilayer (insert in Figure 4). With 6, = 4.5 X 10-6, a bilayer thickness of 44.4 A and a roughness of 4.9 A for each interface, we obtain a reasonably good fit. A Y-type structure is again concluded. Neutron Reflectivity. Figure 5 shows neutron reflectivity data as a function of wave vector k, for 52 layers of polymer IC (deuterated backbone)deposited at a surface pressureof 5 mN/m on treated float glass. The reflectivity measurements are taken for the sample as prepared (X), and after annealing at 60 "Cfor 1 h (square). There appears to be no difference between the two curves at higher k, values; at lower k, values, however, the peaks are more pronounced for the sample which has been annealed. Since the peak positions remain more or less unchanged, this indicates that while the overall thickness of the film is unaffected after heat treatment, the interface is better defined. In these neutron reflectivity measurements, Bragg peaks are not observed due to the limited range of k, values. The Kiessig fringesobtained are rapidly dampened. There is great difficulty in fitting these data; it appears that there is a high degree of inhomogeneity not only in the overall thickness but also in the composition of the film. It is impossible to fit the data toa model of Y-type structure suggested by the X-ray reflectivity results taking into account neutron contrast between the deuterated backbone and protonated
Phase Change in LB Films of Polymers
The Journal of Physical Chemistry. Vol. 98, No. 10, 1994 2611
.'L 4
0.0t
Flgure 5. Neutron reflectivity data for 52 layer of polymer IC deposited on float glass: sample as prepared (X) and sample annealed at 60 OC for I h (squares) (measurements carried out at rmm temperature).
Figure 6 . A F M image of a three-bilayer-thick film. The dark areas correspond to the substrate. This image has been recorded on a sample where the silanization pracess was not fuliy efficient. The horizontal streaks correspond to mattcr removed by the tip. The large holes are not seen ona majorityofsamples(see,for example,Figure8 for acamparison). 1nfact.theyhavebeenobservedoniyonsamples which werenot perfectly
hydrophobic. Webelieve that they result fromlacallackofhydrophabicity due to a poor silanization.
side chain. The best possible fit gives an overall thickness e = 1190 A with a large air-film interface of 250 A. This value of film thickness is smaller than what one would expect if all side chains were completely stretched: 52 X 28.2 = 1466 A. It is, however, not far from the result obtained from X-ray reflectivity: 52 X (42.412) = 1102 A. The neutron reflectivity results also suggest the presence of air or water, both of which can render the filminhomogenwus as well assmear out any existing interface between deuterated and protonated zones. Similar results are obtained for the polymer where the side chain end instead of the backbone is deuterated. The difference in sensitivity to foreign light elements may explain why a certain degree of ordering in the deposited layers can still be detected by X-ray measurements but not neutron reflectivity measurements. It is also possible that the ordered regions are unevenly distributed. In this case, due to the larger irradiated surface area in neutron reflectivity measurements, the neutron reflectivity results are averaged over a larger area of inhomogeneity. Atomic Force Microscopy. We have investigated more than 15 samples with various numbers of layers, before and after annealing. Representative examples are shown in Figures 6 and 7. No difference is observed between samples la. lb, and IC. It should be noted that the AFM investigations are difficult because the scanning itself can very easily destroy the film.
Figure 7. AFM image of a five-bilayer-thickfilm. The topmost layer (mark A) corresponds to a Small area of the film where the outer layer is an odd one (i.e..the methacrylate is observed). On the topmost layer the rms roughness is 0.75 nm whereas on the unperturbed areas of the layerjustbelow (markB),thermsraughnessis0.3nm. Thebottommost visible areas belonging to inner monolayers appear as darker spots (mark C). The white areas in the upper part of the figure are dust particles. However, it is possible to get stable images for I h or more, provided the applied load is kept a t a low value. Typical loads are helow 3 nN for as prepared samples and below about IO nN for samples annealed for 1 h around 60 "C. This difference was the only one we were able to observe between annealed and unannealed samples. In particular, due to the high root mean square (rms) roughness value (Figure 7). it was not possible to confirm the phase change from 22 A/layer to 28 Allayer due to annealing. However, useful data has been collected. First of all, Figure 6, and especially Figure 7, show that the film surface is far from being homogeneous and confirm the inhomogeneity deduced from X-ray and neutron reflectivity experiments. Thesamplesareinhomogenwus with regard toboththenumber of actually deposited layers and the texture. Inhomogeneities in the number of layers on a given sample show up as steps which are systematically found when scanning large areas (typically >1 pm). Step height measurements, based on calibration using steps in cadmium hehenate yield monolayer thickness of 25 5 A for both annealed and unannealed samples. The accuracy of the measurements is sufficient to show that the lowest steps do correspond to a monolayer, not a bilayer. However, the accuracy is not good enough to distinguish between the annealed and unannealed period. As layers are deposited on hydrophobic substrates, we expect the samples to have an even number of layers. Thus, for perfect films, we should measure only steps corresponding to bilayers. However, this is not the case because both even and odd number oflayersareohservedon thesamesample. Forexample,in Figure 7 two steps of monomolecular thickness are observed along the line passing through points A, B, and C. This finding is not surprising since, in the best case, the transfer ratio is 0.8, and hence a full coverage of the surface by each layer is not expected. Concerning film texture, two different types of texture are found on each sample. An example is given in Figure 7 where the upper layer (mark A in the figure) is of one type, and the lower one (mark B) of the other. Both textures are granular hut grain size and roughness are very different. In one case (upper layer) the layer is made of grains 30-60 nm in diameter and the rms surface roughness is 0.75 nm. In the second case, the grain diametersare 10-30nmand thermssurfaceroughnessis0.3nm. Given the monolayer thickness, the average area per monomer and thenumberofmonomersperpolymericmolecule(5~1000), we estimate the grains of diameter 3 M 0 nm and the grains of diameter 10-30 nm to be made up of 2-20 and 1 4 polymeric molecules, respectively. Moreover, we have observed that the
1612 The Journal of Physicol Chemistry. Val. 98, No. IO, 1994
Bosio et al. techniques (such as the silanization of the glass substrate) developed in the field of liquid crystals to align homeotropically nematicor smecticmesophases.2' However, in the present study, contrary to the above-mentioned traditional methods, the alignmentisohtainedeven when thesamplesarecooledquicklywithout special care. 3. The Y structure of a single bilayer (Le., from the glass toward external air-film interface: side chains-polymeric hackbonepolymeric hackbone-side chains) is retained even after heating thesampleat 118OC. Thisstructuremight bestabilized by preferential interactions between the side chains and the surfaces (substrate or air). This 'surface phase" is similar to the ones classicallv observed in liauid crvstals both at the substrate/ liquid crystal or the liquid crystallair interfaces.28 This stable single bilayer might play a role in the growth of the SA, moncdomain in multilayer samples, probably by an epitaxial phenomenon.
Figure 8. A F M image of a one bilayer thick film. Except far defects appearing as bright spots, only a bilayer (over ca. 70% of the surface, mark A) and a monolayer (over ca. 30% of the surface, mark B) can be seen. No more than two layers are built up during the single up/down
stroke.
large grain texture correspondsto odd layers and the small grain texture to even layers indicating that the large grain texture corresponds to the hydrophilicmethacrylate plane and the small grain texture to the hydrophobic aliphatic tail plane. The results obtained on bilayer samples are similar to those obtained on multilayers: there is a lackof homogeneity and there exist two types of textures (Figure 8). Again, no remarkable differences have been observed between annealed and unannealed samples. Moreover, for bilayer samples the maximum load that a film can withstand is not increased by annealing, contrary to multilayer samples. This is consistent with the absence of structural change upon heating as deduced from X-ray experimentson bilayers andconfirms thedifferent behaviors of bilayers and multilayers. The thickness of a monolayer is measured to be 24 X 5 A in bilayer samples, which is in agreement with the X-rays results. Conclusion Theresultsohtained with polymers la, lb, and lcare basically the same and the main features of the Dresent work can he summarized as follows: 1. In contrast to the results of the previous workl3 in which films were transferred at 25 mNJm, the use of an extremely low surface pressure allows the building up of well-defined Y-type structure never observed before. Such a difference is understood as follows: As surface pressure increases, local inhomogeneities or aggregates result, leading to a lack of reliability in the film stacking. At the same time, a dramatic increase in the film viscosityreducesthetransfereffciencyandhencethe homogeneity of the film thickness. Conversely, working at a very low surface pressure clearly improves both transfer efficiency and film homogeneity, which remains however far from being perfect. 2. The new phase observed when transferred at low surface pressure remains present after heating the sample for I h at ca. 60 OC. However, as expected, it turns into the classical SA, mesophase when the sample is heated at 97 'C. This new phase which might he similar to the S A , phase found in liquid crystals is thus metastable. The fact that the SA, mesophase is retained after cooling to room temperature is not very surprising. Moreover, the retention of long-range lamellar molecular order parallel to the substrate by the same SA, after heating up to 1 18 'C (i.e., So heyondthenematictoisotropictransitiontemperature) is reminiscent of the results obtained with well-established
Acknowledgment. The authors are indebted to Dr. M. F. Achard from C.R.P.P. (Bordeaux-France) for differential scanning calorimetry measurements performed on the polymers used in the present work. References and Notes (I) See for example: Proceedings of the 5th International Conference on Langmuir-Blcdgctt Films. Thin Solid Films 1992, 210121 1. (2) Vandevyver, M.: Richard, 1.: Barraud. A,; Vcbcr, M.; Jallabcrt, C.; Strzelecka,H. 3. Chim. Phys. 1988.85.385, (3) Albrecht, O.;Cumming, W.; Krcuda, W.; Laschwky.A.; Rlngdorf. H. Colloid Polym. Sci. 1986. 264, 659. (4) Karthaus, 0.;Ringsdorf, H.: Tsukruk. V. V.; Wendorff, J. H.
Langmuir 1992.8, 2279. ( 5 ) Dura", R.S.;Thibadeaui.A. F.:Rin~~sdarf,H.;Schustcr.A.;Skoulim. A,; Gramain. P.: Ford, W. Polym. Prep,. 1991.32, 246. (6) Rcttig. W.;Naciri.J.;Shashidhar, R.: Duran. R.S. Moeromoleculca 1991. 24, 6539. (7) Albauy. P. A,; Vandcvyvcr, M.: Perez. 1.; Eeaffet, C.; Markovitsi, D.: Vebcr, M.; Jallabcrt, C.; Strzclecka. H. Langmuir 1991, 8. 2262. (8) Vandcvyver. M.: Albauy. P. A,; Mingotaud, C.; Perez. 1.;Barraud. A,: Karthaus, 0.:Ringsdorf, H. Longmuir 1993, 9, 1561. (9) Penner. T. L.: Sehildkraut. J. S.;Ringsdorf. H.; Schustcr, A. 3. Am. Chem. Soc. 1991.24, 1041.
(IO) Erdelen, C.; Lasehewsky. A,; Ringsdorf, H.;Schncider. 1.;Schustcr. A. Thin Solid Film 1989. 180. 153. (11) Entin. I.; Coffer, R.: Davidaf, D.: Hcnht, I., to bc published. (12) Tredgold has published B number of papers on temperature induced recrystallization of LB films. See for example: Joncs, R.: Trcdgold. R. H.: Hmrfar, A,; Allen, R. A,; Hcdgc, P. Thin Solid Films 1985, 134, 57. (13) Vandevyver, M.; Kella, P.; Roulliay, M.: Bourgain. J. P.; Barraud. A. 3. Phys. Appl. Phys. 1993, 26. 1. (14) Portugal. M.; Ringsdorf. H.; Zentcl. R. Mokromol. Chem. 1982. 183,2311. (15) Davidson. P.; Naira, L.; Cotton, 1. P.; Keller. P. t i p . C1y31. 1991, 101,111. (16) ATEMETA. 35 Brd Anatolc Francc. 93200 St. Denis, France. (17) Pomerantz, M.: Segmullcr. A. Thin Solid Films 1980. 68, 33 and
rcferencrs cited herein. (18) Principles of Oprics. Eleclromogncfic Theory of propagalion. Inrerlerenee and diffrmrion oflinhr; . . Max Born and Emil Wall: Pergamon Pr&: Oxford. U.K. (19) Ricutord, F.; Bennatar, J. 1.: Basio, L.: Robin. P.; Blot. C.; de Kouchkovsky. R. J. Phys. 1986.48,679. (20) Benattar, J. 1.;Bosio, L.;Ricutord, F.: Vsndcvyvcr, M. Ann. Phys. Colloq. 1986, 11, No. I, 61. (21) Basio, L.: Cortcs, R.; Folchcr, G.;Oumcrinc. M. Reo. Phys. Appl. 1985. 20. 431. (22) Werner. S. A,: Klein. A. G. In Neurron Scnllerinr: Skold, K.,Price, D. L.,Academic hess: New York, 1986 (23) Binnig, G.;Quate.C. F.;Gerbcr, Ch.Phy8. Reu.LCtl. 1986,56,930. (24) Radmachcr, M.: Tillmann. R. W.;Fritz, M.; Oaub. H. E.Scieme 1992.257. 1900.
(25) Digital lnstrumcnu Inc., 520E. Montecito St.. Santa-Barbara. California 93103. (26) Schwvartz. D. K.: Gamacs. J.; Wiswanathan. R.; Zaaadrinski, J. A. N. Science 1992.257.45? (27) Cognard, J. Mol. Crysl. t i q . Crysl. 1982, Suppl. I, 1. (28) Pershan. P. S. Slrclurcof Liquid Crysrnl Phares; World Scientific Singapore, 1988.
