Langmuir 1992,8, 887-893
887
Spectroscopic Determination of Diffusion in Langmuir-Blodgett Films. 1. Interlayer Diffusion of Cadmium Arachidate and Its Deuterated Analogue through a Polymer Interface M. Shimomurat Department of Chemistry, University of California a t Berkeley, Berkeley, California 94720
K. Song$ and J. F. Rabolt* IBM Research Division, Almaden Research Center, 650 Harry Road, San Jose, California 95120-6099 Received August 27, 1991 The functional specificity of polymer monolayers can be controlled through molecular architecture. A series of preformed polymers containing hydrophobic side chains of various chemical compositions has been investigated in order to assess their interfacial properties at diffusion barriers. Fourier transform infrared and neutron reflectivity studies of multilayered Langmuir-Blodgett films of cadmium arachidate and ita deuterated analogue indicate that rapid interdiffusion occurs at elevated temperatures (85 "C). Results also indicate that the extent of this interdiffusion can be controlled through the introduction of a polymer bilayer at the interface and can be varied depending on the chemical structure of the side chain. Introduction Much structural information about the order and orientation of molecules comprising Langmuir-Blodgett (LB) monolayers and multilayers has become available in the last decade. X-ray,1,2IR,3-8andRamang-12studies, in particular, have provided an insight into the way amphiphilic molecules organize in these two-dimensional films and what the effect of intermolecular environment is on the orientation of various parts of the molecule. Results indicate that LB films made up of fatty acid molecules are highly ordered with their n-alkyl tails organized into twodimensional (2D) crystals. The long-term stability of these organized assemblies on solid surfaces as a function of temperature has also been investigated by IR measurements8 and ellipsometry.13 At temperatures below 60 OC very little mobility is detectable but defect structures such as gauche bonds are detected a t the end of the n-alkyl tails. Exposing these t Permanent address: Research Institute for Polymers and Textiles, Tsukuba, Ibaraki 305, Japan. Permanent address: Hoechst-CelaneseResearch Lab, 86 Morris Ave, Summit, NJ 07901. (1) Garoff, S.;Deckman, H. W.; Dunsmuir, J. H.; Alvarez, M. S. J. Phys. (Paris) 1986, 47, 701. (2) Grundy, M. J.; Richardson, R. M.; Roses, S. J.; Penfold, J.; Ward, R. C. Thin Solid Films 1988,159,43. (3) Chollet, P.; Messier, J.; Rosilio, C. J. Chem. Phys. 1976, 64, 1042. (4) Ohnishi, T.;Ishitani,A.; Ishida, H.;Yamamoto, N.;Teubomura,H. J . Phys. Chem. 1978,82, 1989. (5) Takenaka,T.;Nogami, K.;Gotoh, H. J. ColloidInterface Sci. 1971,
40,409. (6) Allara, D. L.; Swalen, J. D. J. Phys. Chem. 1982,86, 2700. (7) Rabolt, J. F.; Burns, F. C.; Schlotter, N. E.; Swalen, J. D. J . Chem. Phys. 1983, 78, 946. (8)Naselli, C.; Rabolt, J. F.; Swalen, J. D. J. Chem. Phys. 1985, 82, 2136. (9) Knoll, W.; Philpott, M. R.; Golden, W. G. J. Chem. Phys. 1982,77, 219. (10) Masson, M.; Caillaud, M.; Zhi-nan, P.; Dupeyrat, R. Opt. Commun. 1985, 53, 33. (11) Barbaczy, E.; Dodge, F.; Rabolt, J. F. Appl. Spectrosc. 1987,41, 176. (12) Rabe, J. P.; Swalen, J. D.; Rabolt, J. F. J. Chem. Phys. 1987,86, 1601. (13) Rabe, J. P.; Novotny, V.; Swalen, J. D.; Rabolt, J. F. Thin Solid Films 1988,159, 369.
LB films to even higher temperatures can cause interlayer diffusion and Hblation13 of molecules from the surface. This latter aspect of LB films has become increasingly more important as the quest to make optoelectronic and microelectronic devices gains momentum, since, in many cases, the LB film component of such devices will be exposed to high humidity and elevated temperatures which may cause an alteration of their desired properties. In particular, with the wide scale availability of alternating LB troughs, increasingly more device applications have become available since now layers having differing functionalities can be assembled on a single substrate. In fact, recent work by Hayden et al.14 demonstrated that alternating LB layers of a hemicyanine dye interleaved with tricosenoic acid to form Y-type films enhanced its second harmonic generation output relative to Z-type deposited dye layers without spacers. IR15studies showed that in the latter case the inherent disorder of Z-type layers decreased the output of SHG signal while the use of spacer layers improved significantly the orientation and order of the dye chromophores giving a larger SHG response. The expanding interest in alternating layered structures has been a primary motivation for our study of interlayer diffusion in LB films. Certainly the exposure of layered structures to elevated temperatures could provide the molecular mobility required for molecules in one layer to diffuse into an adjacent one as shown in Figure 1. The purpose of this work was to first establish that interlayer diffusion occurs and then to explore the use of various polymeric bilayers to provide an interfacial barrier to transport. The various polymers chosen differed by the choice of chemical architecture of the side chain and, as such, provided different amounts of resistance to interlayer diffusion. The precise role of side chain packing on interlayer diffusion is currently under study and will be the subject of an upcoming report.'6 ~
~~
(14) Hayden, L. M.; Kowel, S. T.; Srinivasan, M. P. Opt. Commun. 1987, 61, 351. (15) Stroeve, P.; Saperstein,D. D.; Rabolt, J. F. J. Chem. Phys. 1990, 92, 6958. (16) Agosti, E.; Rabolt, J. F. To be submitted for publication.
0 1992 American Chemical Society
Shimomura et al.
888 Langmuir, Vol. 8, No. 3, 1992
11111111 TTTTTTTT
-
llblllbl
TPTPTT~P
Figure 1. Schematic diagram of interlayer transport in alternating layer LB films.
Doub I e t
+
Singlet
Experimental Section LB multilayer8 of cadmium arachidate (CdA) and polymers were prepared on a commercially available film balance (Joyce Loebl) with a solid Teflon trough instead of the originallysupplied glass tank. The water in the subphase was purified by deionization and then passed through a Barnstead Nanopure filtration system. The surface pressure of the monolayers was monitored by a Wilhelmy pressure pickup system. The substrates were zinc selenide disks (25 mm diameter and 2 or 4 mm thickness) which were cleaned with a chloroform/methanol mixture ( l / l ) , treated by water spray, deionization bath, and 2-propanol water exchanger, and finally dried at 100 OC. The monolayer-forming fatty acids, arachidic acid, and deuterated arachidic acid (Cambridge Isotope Labs) were used as received. The fatty acids were spread from chloroform solutions having a concentration of 12.5 g/L. The aqueous subphase contained 2.5 X mol/L CdClz for arachidic acid or 1.0 X 10-3 mol/L CdCla for deuterated arachidic acid. These conditions guarantee full conversion7of acid into the cadmium salt. To prepare the mixture LB films of CdA-hag and CdA-d39 (3:7, 55, and 7:3 in molar ratios), protonated and deuterated arachidic acids were mixed in chloroform solution (1g/L) and spread on the subphase c o m n i n g 1.0 X mol/L CdC12. Surface pressure during the deposition process of CdA monolayers was 40 mN/m and the dipping speed about 2 mm mi+. A drying period of 15 min between subsequent dips was necessary to prevent retransfer of the previously deposited layers during the next dip downward. The polymers17used as a diffusion barrier contained an alkyl side chain (CISpolymer), a hemicyanine side chain (PECH), and a fluorocarbon side chain (CFCH polymer). The exact chemical structure will be discussed in a later section. The CISpolymer and PECH were spread from chloroform solutions having a concentration of 1g/L. The CFCH polymer was spread from a chloroform/methanol mixture solution (9/1 by volume) of 0.5 g/L. Surface pressures during the deposition process for CIS polymer, PECH, and CFCH polymer were 32.5,35, and 40 mN/ m, respectively, and the dipping speed was about 2 mm mi+. As before, the drying period between subsequent dips was 15 min. All infrared measurements were made with an evacuated IBM IR98 FT-IR. Spectra were recorded using a room-temperature deuterated triglycine sulfate (DTGS) detector. Since CdA molecules in LB monolayers are aligned nearly perpendicular to the surface,7J the induced dipole momenta of the CH2 and CD2 bending modes are almost parallel to the surface. Thus, zinc selenidedisks were used in the transmission arrangement (electric field vector is parallel to the surface) to detect these bending modes. To measure the band splitting of CH2 or CDZbending, which has maximum value of about 10 and 7.5 cm-l, respectively, spectra were recorded at 2 cm-1 resolution. To obtain spectra with high signal-to-noise ratio, from 2000 to 8000 scans had to be collected depending on the number of monolayers and the strength of the bands. Though the interferometer was evacuated, residual water vapor interfered with the CH2 bending region and these water vapor bands were removed by the spectral subtraction method. Heat treatment of the samples was done using a high-temperature attachment in the evacuated FT-IR. After a roomtemperature measurement, the sample was heated in vacuum to the desired temperature in about 10 min and held at that temperature for 70 min. After the spectrum at elevated tempera(17) The (2x8 was a gift from the s. C. Johnson Co. while PECH was synthesized by Jeff Lindsay of the Naval Weapons Center, China Lake. The CFCH polymer was one of a seriee synthesized by C. Erdelen and H. Ringsdorf of the University of Mainz.
Figure 2. Illustration of changes in the orthorhombic unit cell environment when deuterated (D)and protonated (H)chains intermix. The IR spectrum of the bending and rocking modes would be split into a doublet for structures on the left but would appear as a singlet for isolated chains (right side). ture was obtained, the sample was cooled down to room temperature and a subsequent spectrum obtained.
Results and Discussion A. CH2 and CD2 Bending Vibrations. A number of over the last 25 years have been devoted to the investigation of the CH2 (1470cm-l) and CD2 (1090 cm-l) bending vibrations in selected n-alkanes and polyethylene. Along with their corresponding rocking vibrations, they have been shown to split into two components due to the nature of the intermolecular coupling of two chains in an orthorhombic unit cell. The splitting of these modes has been shown to be very sensitiveto the local envir0nment~~3~ and, as shown in Figure 2, goes from a doublet as the number of dissimilar (by virtue of their different masses) oscillators in the vicinity increases to a singlet when only an isolated chain remains. This can be qualitatively18J9 understood by considering that the origin of the splitting is due to the weak coupling of two identical oscillators on adjacent chains in the unit cell. If one of the surrounding chains is replaced by one whose vibrational frequencies are quite different, e.g., a CD2 oscillator, the intermolecular coupling to that oscillator is removed. In the extreme case depicted on the far right of Figure 2, if an oscillator is completely decoupled from its environment, then the CD2 bending and rocking modes will occur at a single frequency. One clear demonstration of the effect of the unit cell environment has been shown by Bank and KrimmZ2in a study of mixed crystals of C36H74 and C36D74. In their study the splitting of CH2 and CD2 bending modes was measured as a function of composition and a plot of their results is shown in Figure 3 (solid and open triangles). Both the experimental data and subsequent calculations= indicate that both the CH2 and CD2 splitting5 decrease as the amount of the second component is included in the lattice. It should be pointed out that an average splitting is plotted since the IR beam encompasses a large number of unit cells and hence contributions from all the structures in Figure 2 will be represented in the observed CH2 and CD2 bending bandshapes. (18) Krimm, S. J. Chem. Phys. 1954, 22, 567. (19) Snyder, R. G. J. Mol. Spectrosc. 1961, 7, 116. (20) Tasumi, M.; Shimanouchi, T. J. Chem. Phys. 1965,43,1245. (21) Zerbi, G.; Piseri, L. J. Chem. Phys. 1968, 49, 3840. (22) Bank,M. I.; Krimm, S. J.Polym. Sci., Polym. Phys. Ed. 1969, 7, 1785. (23) Tasumi, M.; Krimm, S. J. Polym. Sci., Polym. Phys. Ed. 1968,6, 995. (24) Casal, H.; Mantach, H.; Cameron, D.; Snyder, R. G. J. Chem. Phys. 1982, 77, 2825. ( 2 5 ) Ungar, G.; Keller, A. Colloid Polym. Sci. 1979, 257, 90.
Diffusion in LB Films
Langmuir, Vol. 8, No. 3, 1992 889 0.01
10 7
5
\ 0
1
9
0
al C
; n
A
U
0 A
I:
-
0.00 1490
1480
1470
1460
1450
Wavenumber / cm-'
I00 0.0
-I 0.5
0.005
1.0
Composition o f CDz
LuuLAJ
1.0
0.5
0.0
0 W
n
k
Composition o f CH2
Figure 3. Observedsplittingof CH2 (open points) and CD2 (solid points) bending modes in mixed crystals of C38H74 and C36D74 (triangles from ref 22) and mixed monolayers of CdA and CdAd39 (circles). In the text this is referred to as the calibration
n U
curve.
Doublet
n
Singlet
n
Figure 4. Schematicof mixed layer LB f i i showing the presence of some representative unit cell structures which could be found. In order to ascertain the applicability of this calibration curve obtained for 3D mixed crystals to 2D LB films of CdA and its deuterated analogue,stacked samples of mixed monolayers of the type schematically shown in Figure 4 were constructed with the intent that incorporating two dissimilar chains in the lattice would also result in a change in the CH2 bandsplittings previouslyobserved. The roomtemperature IR spectra obtained from the mixed monolayers are shown in parts a and b of Figure 5, which contain the CH2 and CD2 bending regions, respectively, as a function of layer composition. As can be seen in both spectra, with the introduction of isotopically different species, the splitting of the doublet is diminished until only a singlet remains. If the observed splitting is plotted on the same axes as the mixed crystals, one observes an excellent fit with the previous data (Figure 3 open and solid circles). In fact the CH2 splittings observed for the mixed monolayers overlaps that observed for the mixed crystals. There is a small systematic offset for the observed
0.000
-
1110
1100
1090
1080
1070
Wavenumber / cm-'
Figure 5. (a)IR spectra in CH2 bending region of stacked mixed layersas a function of composition. Number on right is percentage of CdA molecules. (b) IR spectra in the CD2 bending region of stacked mixed layers as a function of composition. Number on right is percentage of CdA-dag molecules in layer. splitting of the CD2 bending vibration which may be due to a slight change in the unit cell parameters in CdA-dag since the offset is still present in the 100% deuterated monolayer. The important point demonstrated by these data are that the observed splittings in mixed monolayers are very similar to that found for the orthorhombic nalkanes, indicating that the cadmium head group does not perturb the lattice in any significant way. Hence this calibration curve can be used to determine average composition of the unit cell from the observed splittings. B. Disorder in LB Monolayers Deposited on a Dielectric Surface. With the above calibration curve firmly established, it was then possibleto explore a number of different layer geometries as shown in Figure 6 to address the question of surface disordering and interdiffusion. Shown schematically in the top two examples of Figure 6 are three-layer systems which alternatively have a monolayer in contact with the hydrophilic surface (ZnSe) followed by a bilayer of isotopically different material. The spectra obtained for both structures in the CH2 bending region are shown in Figure 7. Although the signal to noise (S/N) ratio is not very large due to the small number of layers, it is clearly observed by bandwidth considerations that when CdA monolayer is on the surface, the CH2 bending vibration is a singlet, whereas if it is isolated from the surface by a CdA-d3g layer, it is in fact a doublet. Likewise (but not shown) the CD2 bending vibration is a singlet when CdA-da~monolayer is in contact with the surface and a doublet when isolated in bilayer form by a CdA monolayer. This then indicates that contact with the surface in some way changes the intermolecular arrangement of chains in the unit cell of a monolayer.
Shimomura et al.
890 Langmuir, Vol. 8, No. 3, 1992 0.1 C- dsg-CdA
E
I 1 I
I
1
I
I
Polymer
Figure 6. Schematic of multilayered structuresinvestigatedby IR transmission studies. The vertical line next to the sample designation represents the hydrophilic ZnSe substrate surface.
3000
I
2500 2000 1500 Wavenumbers, em-'
I
1000
Figure 8. IR transmissionspectra of multilayered f i b s of CdA (9H), CdA-dse (9D), and an alternating bilayer of each on ZnSe.
1600
1550
1500
1450
1400
Wavenumbers,cm-l
Figure 7. IR spectra of D/2H and H/2D multilayeredstructures in the CH2 bending region. First indications that this reorganization of the surface monolayer occurred were reported by GarofP using low energy electron diffraction (LEED). His results suggested that the new packing arrangement may, in fact,. be hexagonal. From the spectroscopic perspective a change from an orthorhombic unit cell to one which is either hexagonal or triclinic would be sufficient to explain the presence of a single CH2 or CD2 bending mode in the monolayer directly adjacent to the surface. However, comparing the bandwidth24of the CH2 bending vibration in the single layer (H/2D) spectrum of Figure 7 with that of a multilayer (9H) sample heated to 85 "C, one can conclude that they are identical, suggesting then the surface layer is, in fact, hexagonal in agreement with the LEED results. It should be emphasized that the spectroscopic results indicate that the second, third, etc., monolayer has the n-alkyl tails packed in an orthorhombic unit cell; i.e., the surface induced reorganization which occurs in the first monolayer is not propagated through adjacent monolayers. This nontransferability of disorder in fatty acid monolayers is significant and has led to the use of these monolayers as spacers in NLO applications.15 C. Multilayer vs Mixed Layer Structures. An 9D, investigation of several multilayered structures (QH, H/2D/2H/2D/2H) shown in Figure 6 was undertaken to assess any spectral changes in CHZand C& band split-
tinge which might occur at room temperature due to the stacking of deuterated and protonated layers next to one another. Shown in Figure 8 are the spectra (3200-900 cm-') obtained for pure 9H and 9D layers and an alternating bilayer structure consisting of nine monolayers. Certainly one point which is clearly obvious from Figure 8 is the relative weakness of the CH2 and CD2 bending vibrations compared to their corresponding stretching vibrations. It should be noted that this is, in part, responsible for the long collection times (4000 scans) needed to obtain good S/N spectra in the CH2 and CD2 bending region. Considering the stacked mixed layer spectrum at the bottom of Figure 8, it becomes clear that both the CH2 (1460-1470 cm-l) and CD2 (1085-1090 cm-l) bending modes are both split and present, providing a means to monitor both deuterated and protonated layers in stacked layer systems. Close inspection of the bands shows that there is no effect on the amount of the observed splitting in either case indicating that the effect of stacking is %oninvasivew. D. Effect of Elevated Temperature on Multilayered Structures. It was determined, through a series of room-temperature IR measurements on multilayered structures of pure CdA and CdA-dag which had been previously heated to elevated temperatures, that 85 OC was the optimum temperature for investigation of interdiffusion. As shown in Figure 9 after heating a multilayered (9D) sample above 85 OC and returning to room temperature, the splitting and general bandshape of the CD2 bending vibration have changed considerably. Heating to still higher temperatures and returning to ambient cause this spectral region to change irreversibly. Hence 85 "C was chosen as the maximum temperature below which the monolayer retained "memory" as shown in Figure loa. In contrast there is a dramatic effect on this spectral region when a multilayered structure consisting of alternating bilayers of CdA and CdA-dss is heated to 85 "C and returned to room temperature as shown in Figure lob. It is clear that after heating at 85 "Cfor 1h the bandahape of the CD2 bending vibration has changed considerably. In fact, if one does rudimentary curve fitting to remove
Diffusion in LB Films 0.006I
Langmuir, Vol. 8,No. 3, 1992 891 I
I
I
5D
0.01 1
Aat I After heating 97oc
R.T.
.1
a, C
Io
60'C
n
.1
0
n
R.T.
U
.1 85'C
.1 0.00 1150
R.T.
1100
1050
1000
Wavenumber / cm-' 0.0001
1110
I
I
I
J
1100
1090
1080
1070
Wavenumbers (cm-')
Figure 9. Effect of heating to various temperaturesand returning back to room temperature on the CD2 bending doublet. All IR spectra were recorded at ambient after raising the CdA-das multilayered structure to the temperature listed on right of figure.
the contribution of the singlet to this profile, the splitting observed after heating is then found to be 5.7 f 0.2 cm-', compared to 6.5 X 0.2 cm-l before heating. By use of the calibration curve established previously (Figure 31, this splitting indicates that approximately 30% of the CdA molecules can be found in the CdA-d39 monolayers. If, instead, the CH2 bending region is used for analysis, then the observed splitting of the CH2 bending mode is observed to change from 9.6 to 6.3 cm-1 after heating, indicating that 30% of the CdA-das molecules have diffused into the CdA monolayers. Thus there is an internal consistency whether using the CH2 for CD2 bending region. Similar studies on stacked layers using neutron reflectivity were undertaken since previous25investigations of deuterated and protonated paraffin crystals put in contact with each other had yielded valuable information on the intermixing process. Our26studies on stacked LB layers by neutron reflectivity indicated that approximately 25 % of the CdAd39 molecules were replaced by CdA molecules when the multilayered film was heated to 70 "C. It is interesting to note that in the latter measurements, the results are not predicated on any splittings but on shifts in scattering maxima and changes in this intensity but essentially give the same result. This therefore presents a strong argument for IR spectroscopy being a viable method to measure molecular transport in LB films. E. Polymer Interfaces as Barriers to Interdiffusion. The final question raised in this study was to eliminate interdiffusion by the introduction of a polymer interface at the boundary between CdA and CdA-dsg multilayers as depicted in Figure 11. It was first necessary to assess the effect of a polymer bilayer on the organization in the adjacent layer in order to determine whether the polymer side chain diffuses into or disrupts the intermolecular packing in the CdA layer. For this purpose a stacked layer structure consisting of
'
(26) Stroeve, P.; Rabolt, J. F.; Hilleke, R.; Felcher, G.; Chen, S. H. Mater. Res. SOC.Symp. R o c . 1990, 166, 103.
--u
R.T.
.1 60'C I
E
L J I d,i --
"
n. I
Shimomura et al.
892 Langmuir, Vol. 8, No. 3,1992
L
I
\
I
E
111tt Ttttt1 lllllllllll
4D/3H
m
0.000 n ooo1110 1110
1090 1090
1070 1070
Wavenumber (cm-’)
Figure 11. Schematic of LB bilayers separated by a polymer bilayer at the interface. Polymer side chains are represented by rectangles.
Figure 14. Change in CD2 bending vibration of a 4D/2P/3H (P = PECH polymer) layered structure after heating to 85 OC. For comparison, the spectrum (bottom) of a similar multilayered structure (4D/3H) which does not contain a diffusion barrier layer is included. dsa-CdA
hsa-CdA
dsa-CdA
Polvmer
hsa-CdA
@ N+C I - I
C16H33 I
CHz
+cH2-cH@o 0 alkyl side chain polymer C I 8 polymer
f C H - CHz- 0%
Heating hemicyanine side chain polymer
Heating
fluorocarbon side chain polymer CFCH polymer
PECH
Figure 12. Chemical structures of the C18, PECH, and CFCH polymers used in these interdiffusion studies. 0.050
I
I
I
I
CD, BandingVibration after Heating to 8S°C
Figure 15. Schematic of the interdiffusion which occurs with and without a polymeric bilayer at the interface.
‘E
DDDDPPDDD
O.Oo0 1140
1100 Frequency (cm-’ )
1080
Figure 13. Room-temperature IR transmiasion spectra of CD2 bending region for 9D and 4D/2P/3D layered fiis after heating to 85 OC. The P used in the latter represents, in this example, the PECH polymer.
(bottom spectrum). In the case of the PECH bilayer, this splitting was measured from Figure 14and translates, using the established calibration c w e , intoa migration of 10% of the CdA and CdA-dag molecules through the polymer interface. This is schematically shown in Figure 15 and represents a 67% reduction in the number of molecules that interdiffuse when the hemicyanine side chain polymer layer is used as a barrier compared to that when no polymer bilayer is present. This rather exciting result provides
-
the first experimental evidence that polymeric bilayers can provide effective barriers to interdiffusion in LB multilayers. In order to further explore the role of the chemical architecture of the polymeric side chain on its properties as a diffusion barrier, polymers containing both a hydrocarbon side chain (CUpolymer) anda partially fluorinated side chain (CFCH) were also investigated. In the latter case it was the thermodynamic incompatibility of fluorocarbons and hydrocarbons which led to the choice of this polymer as a potential barrier between hydrocarbon layers. After preliminary room temperature measurements before and after heating provided convincing evidence that the presence of the Cl8 or CFCH polymer bilayers was noninvasive as observed earlier for PECH, changes in the CD2 splitting due to migration of CdA molecules through the barrier was investigated in both cases. As shown in Table I, the CISpolymer was a poor barrier to interdiffusion which can be attributed to the compatibility of the CIS side chains and the diffusing CdA and CdA& molecules. On the other hand, the use of the CFCH polymer bilayer reduced interdiffusion by 84% (30% 5%) due in part to the incompatibility of the fatty acid monolayers with the semifluorinated side chains. This same fluorophobic behavior has been shown to lead to
-
Diffusion in LB Films
Langmuir, Vol. 8, No. 3, 1992 893
Table I. Comparison of the Average Composition of a Mixed Layer after Heating to 85 O C for the Three Polymers Used aa Interfacial Barriers. barrier layer
amt interdiffused, %
none alkyl side chain polymer hemicyanine side chain polymer fluorocarbon side chain polymer
30 25 10
5
* T o p number is for a layered film without a barrier. Numbers come from measured splitting5 of CD2 or CH2 bending nodes which were converted to average composition through a calibration curve (Figure 3).
unique crystal packing arrangement^^'^^^ in semifluorinated n-alkane oligomers of the form F(CF2)n(CH2)mH. Results for all three polymers are summarized in Table I and they are quite revealing. It is definitely clear that the chemical nature of the side chain affects its barrier properties even at the bilayer level (-50 A thick). What cannot be assessed at this point is the effect of side chain packing on interdiffusion since a close packed "lattice" of side chains would be a more effective barrier to transport than a partially disordered layer of side chains which contained many vacancies and conformational defects. To address this question, a study is in progress16using CFCH polymers which contain differing amounts of backbone spacer groups ( min Figure 121, since previous result^^^^^^ indicated that side chain organization improves with an increasing amount of spacer groups between side chains. Conclusions The nature of the intermolecular interactions in LB films of CdA and CdA-dss has been investigated through IR studies of CH2 and CD2 bending vibrations and their subsequent splitting due to weak coupling of adjacent chains in the lattice. Through a comparison of the splittings observed in 2D films with those previously reported for n-alkane single crystals, it was concluded that the (27) Rabolt, J. F.; Russell, T. P.; Twieg, R. J. Macromolecules 1984, 17,2786. (28) Russell, T. P.; Rabolt, J. F.; Twieg, R. J.; Siemens, R. L.; Farmer, B.L.Macromolecules 1986, 19, 1135. (29)Schneider, J.; Ringsdorf, H.; Rabolt, J. F. Macromolecules 1989,
205. (30)Schneider, J.; Erdelen, C.; Ringsdorf, H.; Rabolt, J. F. Macromolecules 1989,22, 3475.
22,
structure of the unit cell is identical. This implies that in the LB films of CdA, the divalent cadmium head group does not perturb the packing of the n-alkyl chains relative to the n-alkanes, which are known to pack in an orthorhombic unit cell. Using the observed splitting of CH2 and CD2 bending modes to probe the surface of a dielectric substrate, it was determined that the first LB monolayer in contact with the surface disorders or at least undergoes a change in crystal structure from an orthorhombic unit cell to one which is most probably hexagonal. Interestingly enough this surface rearrangement of the first monolayer is not propagated through the adjacent layers. In fact these adjacent layers appear to be unperturbed and maintain their orthorhombic arrangement of chains within the unit cell. The functional specificity of polymer monolayers can be controlled through molecular architecture. A series of preformed polymers containing hydrophobic side chains of various chemical compositions (including NLO chromophores and perfluoroalkyl groups) were investigated in order to assess their interfacial properties as diffusion barriers. FTIR and neutron reflectivity studies of multilayered Langmuir-Blodgett films of cadmium arachidate and its deuterated analogue indicated that rapid interdiffusion occurs at elevated temperatures (85 "C) and 2 5 3 0 % of the chains undergo interlayer transport. Results also indicate that the extent of this interdiffusion can be controlled through the introduction of a polymer bilayer at the interface and can be varied depending on the chemical structure of the side chain. Perfluoroalkyl side chain polymers are the most effective barriers, reducing the interlayer diffusion by 84% compared to multilayered structures without diffusion barriers. Acknowledgment. J.F.R. acknowledges many helpful discussions with S. Krimm (University of Michigan), R. Snyder (University of California, Berkeley), G. Zerbi (Politecnico Milano), and I. Levin (NIH). M.S. also thanks the Research Institute for Polymers and Textiles for the opportunity and financial support to do research in the
us.
Registry No. PECH,84828-51-3;CFCH, 138093-10-4;Cls polymer, 138128-57-1;CdA, 14923-81-0;ZnSe, 1315-09-9.
