Aging Effects on Molecular Orientation, Structure, and Morphology in

Hélène Bourque, Thierry Buffeteau, Daniel Blaudez, Bernard Desbat, and Michel Pézolet. The Journal of Physical Chemistry B 2002 106 (8), 1968-1...
0 downloads 0 Views 1MB Size
J. Phys. Chem. B 2000, 104, 1183-1190

1183

Aging Effects on Molecular Orientation, Structure, and Morphology in Langmuir-Blodgett Films of 2-Dodecyl-, 2-Pentadecyl-, and 2-Octadecyl-7,7,8,8-tetracyanoquinodimethane Studied by Infrared and Ultraviolet-Visible Spectroscopy and Atomic Force Microscopy Shin-ichi Morita,† Keiji Iriyama,‡ and Yukihiro Ozaki*,† Department of Chemistry, School of Science, Kwansei-Gakuin UniVersity, Uegahara, Nishinomiya 662-8501, Japan, and The Jikei UniVersity School of Medicine, Nishi-shinbashi, Minato-ku 105-0003, Japan ReceiVed: August 24, 1999; In Final Form: NoVember 3, 1999

Time-dependent changes in molecular orientation, structure, and morphology in one-layer Langmuir-Blodgett (LB) films of 2-dodecyl-, 2-pentadecyl-, and 2-octadecyl-7,7,8,8-tetracyanoquinodimethane (dodecyl-TCNQ, pentadecyl-TCNQ, and octadecyl-TCNQ) have been investigated by means of infrared and ultraviolet and visible (UV-vis) spectroscopy and atomic force microscopy (AFM). The intensities of bands at 2920 and 2850 cm-1 due to CH2 antisymmetric and symmetric stretching modes of the alkyl chain and those of bands at 2222, 1547, and 1529 cm-1 assigned to CtN, CdC, and CdC stretching modes of the TCNQ chromophore increase significantly with time in the infrared reflection-absorption (RA) spectrum of a one-layer LB film of pentadecyl-TCNQ after the film deposition on a gold-evaporated glass slide. These observations indicate that the alkyl chain becomes more tilted with respect to the surface normal during the time course and the TCNQ plane becomes more perpendicular with respect to the substrate surface. The time-dependent changes are also observed for one-layer LB films of dodecyl-TCNQ, but the changes are much smaller than those for the films of pentadecyl-TCNQ. One-layer LB films of octadecyl-TCNQ do not show appreciable infrared spectral changes with time. A UV-vis spectrum in the 350-450 nm region of the one-layer LB film of pentadecyl-TCNQ on a CaF2 plate consists of two bands due to the stacked and aggregation forms of the TCNQ chromophore. As a function of time, the intensity of the band due to the aggregation form increases while that of the band due to the stacked form decreases, indicating that some of the stacked TCNQ chromophores change into the aggregation with the change in the molecular orientation. The increase in baseline is also noted in the UV-vis spectrum. The baseline change may be due to the partition of the large domains of the one-layer film into small domains with the size of ∼1 µm. Time-dependent measurements of the AFM image of a one-layer LB film of pentadecyl-TCNQ on mica have revealed that the thickness of the film changes from 4.6 to 8.2 nm and that the flat and uniformed structure of the film changes into the condensed and accumulated structure during the time course from 15 to 180 min after the film deposition. Aged onelayer LB films of dodecyl- and octadecyl-TCNQ are similar to the aged and fresh one-layer LB films of pentadecyl-TCNQ, respectively, in terms of both the film morphology and thickness. On the basis of the results of infrared and UV-vis spectroscopy and AFM, we propose a model for the time-dependent morphological and structural changes in the one-layer LB films of the alkyl-TCNQ. The differences in the aging effects among the one-layer LB films of the three kinds of alkyl-TCNQ may be concerned with the differences in the strength of the hydrophobic interaction between the interdigitated alkyl chains and in the degree of three-dimensional microcrystal growth of the one-layer LB films.

Introduction Langmuir-Blodgett (LB) films having a tetracyanoquinodimethane (TCNQ) chromophore have recently received keen interest because some of them have shown fairly high conductivity.1-3 To understand the interesting properties that the conducting LB films with a TCNQ chromophore show, it is very important to investigate the relationship between the function and structure of the films. As the first step of the studies of the relationship, we have been investigating the structural characterization of LB films of 2-alkyl-7,7,8,8-tetracyanoquino* To whom correspondence should be addressed. Fax: 81-798-51-0914. E-mail: [email protected]. † Kwansei-Gakuin University. ‡ The Jikei University School of Science of Medicine.

dimethane (alkyl-TCNQ; Figure 1).4-10 Molecular aggregation, orientation, and the structure, morphology, and thermal behavior have been explored for LB films of 2-dodecyl-7,7,8,8-tetracyanoquinodimethane (dodecyl-TCNQ; Figure 1), 2-pentadecyl7,7,8,8-tetracyanoquinodimethane (pentadecyl-TCNQ; Figure 1), and 2-octadecyl-7,7,8,8-tetracyanoquinodimethane (octadecylTCNQ; Figure 1) by use of ultraviolet-visible (UV-vis) and infrared spectroscopies and atomic force microscopy (AFM).4-10 We have also been studying the structure, morphology, and thermal behavior for mixed-stack charge-transfer (CT) LB films of octadecyl-TCNQ and 3,3′,5,5′-tetramethylbenzidine (TMB)11,12 and of octadecyl-TCNQ and 5,10-dimethyl-5,10-dihydrophenazine ((Me)2P).13,14 Thus far, the following conclusions could be reached from our studies on the structure of the LB films of alkyl-

10.1021/jp992988t CCC: $19.00 © 2000 American Chemical Society Published on Web 01/21/2000

1184 J. Phys. Chem. B, Vol. 104, No. 6, 2000 TCNQ:4-10 (i) The LB films of octadecyl-TCNQ consist of numerous platelet microcrystal domains, which have the layered assembly formed by bimolecular layers with a thickness of 3.7 nm.9,10 A periodic arrangement of octadecyl-TCNQ molecules with a period of 0.85 nm can be observed inside the domains.9,10 The domains in the first layer (4.3 nm) are thicker than those above the first layer (3.7 nm). In the case of the first layer, the direct interaction between the substrate and octadecyl-TCNQ molecules plays an important role in determining the thickness. (ii) The hydrocarbon chains are tilted considerably with respect to the surface normal in the three kinds of TCNQ LB films.4,5 The TCNQ planes and their long axes are also inclined with respect to the surface normal in the LB films.4,5 (iii) Thermally induced structural changes occur gradually in a one-layer LB film of octadecyl-TCNQ but take place abruptly in its multilayer films.9,10 The order-disorder transition in the multilayer LB film of octadecyl-TCNQ with the longer even-numbered hydrocarbon chain occurs at a temperature higher than that in the corresponding LB film of dodecyl-TCNQ with the shorter even-numbered hydrocarbon chain.6 The LB film of pentadecylTCNQ with the odd-numbered hydrocarbon chain shows a transition temperature similar to the film of dodecyl-TCNQ.6 The purpose of the present study is to explore further the structural characterization of one-layer LB films of alkyl-TCNQ. Aging effects on molecular orientation, structure, and morphology have been studied for one-layer LB films of dodecyl-TCNQ, pentadecyl-TCNQ, and octadecyl-TCNQ. We have employed infrared transmission and reflection-absorption (RA) spectroscopy, ultraviolet-visible (UV-vis) spectroscopy, and atomic force microscopy (AFM) to monitor the time-dependent phenomena. By combining three complementary techniques, we can investigate comprehensively the aging effects on one-layer LB films of the three kinds of alkyl-TCNQ. It has been found that the time-dependent changes depend on the length of alkyl chain substituted and the substrate used. It seems that the hydrophobic interaction among the alkyl chains, microcrystallization of the LB films, and the substrate-first layer interaction play key roles in the length of alkyl chain and substrate dependences. To our best knowledge, this is the first time that the time-dependent structural and morphological changes have ever been investigated for one-layer LB films. The present study may provide new insight into the control in molecular orientation in one-layer LB films and the interaction between the first layer and the substrate. Experimental Section Sample Preparation. Dodecyl-, pentadecyl-, and octadecylTCNQ were purchased from the Japanese Research Institute for Photosensitizing Dyes Co., Ltd. (Okayama, Japan), and used without further purification. The thin-layer chromatographic examinations revealed that they did not contain any other colored components. To prepare water for the fabrication of LB films, water was passed through activated charcoal and reverse osmosis filters and then distilled. Finally, it was purified by a Ultrapure Water System model CPW-101 (Advantec, Japan). The resistance of the finally prepared water was larger than 18.1 MΩ. The Z-type LB films of dodecyl-, pentadecyl-, and octadecylTCNQ were prepared by use of a Kyowa Kaimen Kagaku model HBM-AP Langmuir trough with a Whilhelmy balance. Chloroform solutions (∼1 × 10-3 M) of dodecyl-, pentadecyl-, and octadecyl-TCNQ were placed onto an aqueous subphase of water. After evaporation of the solvent, the monolayer was compressed at a constant rate of 30 cm2 min-1 up to the surface pressure of ∼1 mN m-1 and then the rate was kept at 20 cm2

Morita et al.

Figure 1. Structure of 2-alkyl-7,7,8,8-tetracyanoquinodimethane.

Figure 2. Infrared RA (a) and transmission (b) spectra of one-layer LB films of pentadecyl-TCNQ prepared on a gold-evaporated glass slide and CaF2 plate, respectively. The spectra were measured at 6 and 5 min after the film deposition, respectively. An intense absorption band due to CO2 near 2350 cm-1 was cut artificially.

min-1 up to the surface pressure of 5 mN m-1 (293 K). The π-A isotherm revealed that the monolayers were solidcondensed films at this surface pressure. These monolayers were transferred by one upstroke of the vertical dipping method onto CaF2 plates (for infrared and UVvis transmission measurements), gold-evaporated glass slides (for infrared RA and UV-vis reflection measurements), and mica. In the case of mica, the films were deposited on a cleavage plane. These substrates were cleaned by ultrasonifications in acetone, chloroform, and acetone, then in distilled water, and then by a homemade UV ozone cleaner. The transfer ratio was found to be nearly unity throughout the experiments. Spectroscopy. Infrared spectra were measured with a Nicolet Magna 550 FT-IR spectrometer equipped with a MCT detector. The spectra were taken at a 4 cm-1 resolution, and typically, 512 or 1024 interferograms were coadded to yield the spectra of high signal-to-noise ratio. UV-vis spectra were obtained by a Shimadzu UV-3101PC UV-vis spectrophotometer. Microscopy. AFM images of one-layer LB films of the three kinds of alkyl-TCNQ on the mica were measured with a Shimadzu SPM-9500 by a contact mode with constant force. We used silicon nitride tips on a cantilever with spring constants of 0.16 and 0.68 N/m. Results Infrared Spectra. Parts a and b of Figure 2 show infrared RA and transmission spectra of one-layer LB films of pentadecyl-TCNQ prepared on a gold-evaporated glass slide and CaF2 plate, respectively. These spectra were measured at 6 and 5 min

LB Films of Three Decyl-TCNQs

J. Phys. Chem. B, Vol. 104, No. 6, 2000 1185

Figure 4. Possible model for time-dependent (a) orientational and (b) morphological changes in a one-layer LB film of pentadecyl-TCNQ.

Figure 3. Enlargement of time-dependent changes in the infrared RA spectrum of a one-layer LB film of pentadecyl-TCNQ on a goldevaporated glass slide: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 1580-1500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 6 to 131 min after the film deposition.

after the film deposition, respectively. Assignments for infrared bands of alkyl-TCNQ have been well established,4-7 and those for key bands in the infrared spectra of the LB films are as follows: 2918 cm-1, CH2 antisymmetric stretching; 2850 cm-1, CH2 symmetric stretching; 2222 cm-1, CtN stretching; 1547 and 1529 cm-1, CdC stretching; 1471 and 1462 cm-1, a doublet due to the CH2 scissoring. It is noted that the band due to the CH2 antisymmetric stretching mode splits into two.5 We found that the degree of the splitting depends on the length of the alkyl chain substituted. The cause of the splitting is now under investigation and will be reported separately. In Figure 3a-d are shown time-dependent changes in the 3000-2800, 2250-2190, 1580-1500, and 1500-1440 cm-1 regions of the infrared RA spectrum of the one-layer LB film of pentadecyl-TCNQ, respectively. The spectra were measured between 6 and 131 min after the film deposition. Of note is that time-dependent intensity changes are observed for almost all the bands arising from both the alkyl chain and TCNQ plane. As a function of time, the bands due to the CH2 antisymmetric and symmetric stretching modes increase. In the spectrum measured at 6 min after the deposition (the bottom trace in Figure 3a), the intensities of two CH2 stretching bands are weak and nearly the same, while in the spectrum measured at 131 min after the deposition (the top trace in Figure 3a), the intensity of the antisymmetric stretching band is much stronger than that of the symmetric stretching band. According to the surface selection rule in infrared RA spectroscopy,15-18 bands due to the modes with their transition moments perpendicular to the surface are enhanced in an RA spectrum. Therefore, the above observations suggest that the alkyl chain is nearly perpendicular to the substrate surface in the LB film just after the film deposition but that it becomes tilted gradually with time. It is noted that the intensity of the CH2 antisymmetric stretching band increases markedly with time while that of the symmetric stretching band increases a little. The directions of

the transition moments of the CH2 antisymmetric and symmetric stretching modes are perpendicular to each other. Therefore, the above observation suggests that the transition moment of the antisymmetric stretching mode is gradually oriented toward the direction of the substrate normal. The relative intensity of the doublet near 1470 cm-1 due to the CH2 scissoring mode changes as a function of time (Figure 3d). The two bands at 1471 and 1462 cm-1 are assigned to CH2 scissoring modes of noninterdigitated and interdigitated parts of the alkyl chain, respectively.19 The observed spectral changes in Figure 3d suggest that the proportion of the interdigitated parts increases with time probably because of the evaporation of water molecules. It can be seen from Figure 3b,c that the bands due to the CtN and CdC stretching modes increase with time and that the relative intensity of the two bands at 1547 and 1531 cm-1 also varies. The two bands arise from the CdC stretching modes whose transition moments are parallel and perpendicular to the molecular axis of the TCNQ chromophore, respectively.4 Thus, it may be concluded from the results in Figure 3b,c that the TCNQ plane becomes more perpendicular with respect to the substrate surface during the time course and the molecular axis of the TCNQ chromophore becomes more tilted with respect to the surface normal. All the conclusions reached from the infrared RA spectra in Figure 3 lead us to propose a model for the time-dependent changes in the LB film of pentadecyl-TCNQ shown in Figure 4a. Parts a-d of Figure 5 show infrared transmission spectra in the 3000-2800, 2250-2190, 1580-1500, and 1500-1440 cm-1 regions of the one-layer LB film of pentadecyl-TCNQ as a function of time, respectively. The spectra were measured between 5 and 120 min after the film deposition. Again, timedependent changes were observed for most of infrared bands. The band at 2918 cm-1 due to the CH2 antisymmetric stretching band decreases considerably, while that at 2848 cm-1 due to the CH2 symmetric stretching band increases slightly as time increases. These observations indicate that, as in the case of the gold-evaporated substrate, the alkyl chain becomes more tilted with respect to the surface normal during the time course. The bands assigned to the CtN and CdC stretching modes decrease with time (Figure 5b,c), indicating that the TCNQ plane becomes more tilted. Therefore, the time-dependent changes in the molecular orientation in the one-layer LB film of pentadecylTCNQ on the CaF2 plate may also be depicted by the model shown in Figure 4a.

1186 J. Phys. Chem. B, Vol. 104, No. 6, 2000

Figure 5. Time-dependent changes in the infrared transmission spectrum of a one-layer LB film of pentadecyl-TCNQ on a CaF2 plate: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 1580-1500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 5 to 120 min after the film deposition.

Figure 6. Time-dependent changes in the infrared RA spectrum of a one-layer LB film of dodecyl-TCNQ on a gold-evaporated glass slide: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 15801500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 20 to 180 min after the film deposition.

Parts a-d of Figure 6 show time-dependent changes in an infrared RA spectrum in the 3000-2800, 2250-2190, 15801500, and 1500-1440 cm-1 regions of a one-layer LB film of dodecyl-TCNQ on a gold-evaporated glass slide, respectively. The spectra were obtained between 20 and 180 min after the film deposition. In Figure 7a-d are exhibited time-dependent changes in an infrared transmission spectrum in the 3000-2800, 2250-2190, 1580-1500, and 1500-1440 cm-1 regions of a one-layer LB film of dodecyl-TCNQ on a CaF2 plate, respectively.

Morita et al.

Figure 7. Time-dependent changes in the infrared transmission spectrum of a one-layer LB film of dodecyl-TCNQ on a CaF2 plate: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 15801500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 5 to 210 min after the film deposition. Spectra (c) and (d) include some “ghost” bands due to noise.

Time-dependent spectral changes in the RA and transmission spectra of the one-layer LB films of dodecyl-TCNQ are similar to those in the RA and transmission spectra of the one-layer LB films of pentadecyl-TCNQ, respectively (Figures 3 and 5). However, the spectral changes are much smaller for dodecylTCNQ. Therefore, it seems that the time-dependent orientational changes in the one-layer LB films of dodecyl-TCNQ are similar to but smaller than those in the one-layer LB films of pentadecyl-TCNQ. Parts a-d of Figure 8 show time-dependent changes in an infrared RA spectrum in the 3000-2800, 2250-2190, 15801500, and 1500-1440 cm-1 regions of a one-layer LB film of octadecyl-TCNQ on a gold-evaporated glass slide, respectively. The corresponding infrared transmission spectra of a one-layer LB film of octadecyl-TCNQ on a CaF2 plate are shown as a function of time in Figure 9. Note that the spectral changes are very small for both LB films of octadecyl-TCNQ on the CaF2 plate and gold-evaporated glass slide. It is likely that the molecular orientation and structure in the one-layer LB films of octadecyl-TCNQ vary little after the film deposition. It is of note that the intensity ratio of the two CH2 stretching bands and the ratio of the two CdC stretching bands in the RA spectrum of the one-layer LB film of octadecyl-TCNQ are very close to those in the RA spectrum of the one-layer LB film of pentadecyl-TCNQ measured at 6 min after the deposition, respectively (Figures 3 and 8). Therefore, it seems that the molecular orientation in the LB film of pentadecyl-TCNQ just after the film deposition is very similar to that in the film of octadecyl-TCNQ. Probably, the LB film of octadecyl-TCNQ is more stable than that of pentadecyl-TCNQ because of the hydrophobic-hydrophobic interaction of the long alkyl chains. UV-Vis Spectra. Figure 10a,b compares time-dependent changes in UV-vis transmission and reflection spectra of onelayer LB films of pentadecyl-TCNQ on a CaF2 plate and a goldevaporated glass slide, respectively. The UV-vis spectrum in the 300-500 nm region of the film on the CaF2 plate consists

LB Films of Three Decyl-TCNQs

Figure 8. Time-dependent changes in the infrared RA spectrum of a one-layer LB film of octadecyl-TCNQ on a gold-evaporated glass slide: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 1580-1500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 5 to 90 min after the film deposition.

Figure 9. Time-dependent changes in the infrared transmission spectrum of a one-layer LB film of octadecyl-TCNQ on a CaF2 plate: (a) 3000-2800 cm-1 region; (b) 2250-2190 cm-1 region; (c) 15801500 cm-1 region; (d) 1500-1440 cm-1 region. The spectra were obtained from 5 to 60 min after the film deposition.

at least of two component bands near 370 and 420 nm; one (420 nm) is due to a π-π* transition of the aggregation form of the TCNQ chromophore while another (370 nm) arises from the corresponding transition of its stacked form.8,19 The intensity of the band due to the aggregation form increases, while that of the band due to the stacked form decreases, indicating that some of the stacked TCNQ chromophores change into the aggregation form with the change in the molecular orientation. It is also noted in Figure 10a that the baseline in the 450-600 nm region increases with time. The baseline shift in that region

J. Phys. Chem. B, Vol. 104, No. 6, 2000 1187

Figure 10. (a) Time-dependent changes in the UV-vis spectrum of a one-layer LB film of pentadecyl-TCNQ on a CaF2 plate. The spectra were obtained from 5 to 302 min after the film deposition. (b) Timedependent changes in the UV-vis reflection spectrum of a one-layer LB film of pentadecyl-TCNQ on a gold-evaporated glass slide. The spectra were obtained from 3 to 660 min after the film deposition.

is an indicator for the variation in the domain structure of a LB film.9,10 If the domain size of a sample to be measured in a UV-vis experiment has the same order as that of the wavelength of the incident light, the light is diffused by the domains and thus the baseline shifts up.9,10 Probably, the baseline change corresponds to the partition of the large domain in the LB film into small domains with the size of ∼1 µm (Figure 4b). The UV-vis reflection spectra in the 350-550 nm region of the one-layer LB film of pentadecyl-TCNQ on the goldevaporated glass slide are composed of several component bands arising from the stacked form, monomeric form, and aggregated forms of the TCNQ plane. With time the baseline changes significantly in the UV-vis reflection spectrum, but a timedependent change in the overall band shape is rather small. In Figure 11a,b are shown the time dependence of UV-vis transmission spectra of one-layer LB films of dodecyl-TCNQ and octadecyl-TCNQ on CaF2 plates, respectively. Note that the UV-vis spectrum of the LB film of dodecyl-TCNQ shows similar time-dependent variations to that of pentadecyl-TCNQ while that of octadecyl-TCNQ depicts very little change. The spectral variations in the 350-450 nm region and the baseline shift in the UV-vis spectrum of the film of dodecyl-TCNQ suggest that as in the case of the film of pentadecyl-TCNQ the stacked TCNQ chromophores change into the aggregation form and that the large domain decomposes into smaller ones with time. Atomic Force Microscopy. Figure 12 shows AFM images of (a) the fresh (15 min after the film deposition) and (b) the aged (120 min after the deposition) one-layer LB films of pentadecyl-TCNQ on mica. It can be clearly seen from comparison between Figure 12a,b that flat and uniformed domains in the LB film change into smaller and more condensed domains whose size is ∼1 µm in a few hours (Figure 4b). This observation is in good agreement with the conclusion reached from the time-dependent UV-vis spectral changes (Figure 10).

1188 J. Phys. Chem. B, Vol. 104, No. 6, 2000

Morita et al.

Figure 11. (a) Time-dependent changes in the UV-vis spectrum of a one-layer LB film of dodecyl-TCNQ on a CaF2 plate. (b) Time dependent changes in the UV-vis spectrum of a one-layer LB film of octadecyl-TCNQ on a CaF2 plate.

Figure 13a depicts an AFM image of a one-layer LB film of dodecyl-TCNQ on mica measured 2 days after the film deposition. It is noted that the morphology of the aged film of dodecyl-TCNQ is very similar to that of the aged one-layer LB film of pentadecyl-TCNQ (Figure 12b). The domains of the aged one-layer LB films of dodecyl- and pentadecyl-TCNQ are located separate from one another with a certain distance, indicating that the condensation of the films occurs after the film deposition. Figure 13b presents an AFM image of a one-layer LB film of octadecyl-TCNQ on mica measured 2 days after the film deposition. In contrast to the morphology of aged one-layer LB films of dodecyl- and pentadecyl-TCNQ, the AFM image of the aged one-layer LB film of octadecyl-TCNQ shows more flat and uniformed morphology. This morphology is close to that of the fresh one-layer LB film of pentadecyl-TCNQ, suggesting that the morphology varies little with aging. It is obvious that the one-layer LB film of octadecyl-TCNQ with the longer alkyl chain is more stable than those of dodecyland pentadecyl-TCNQ with the shorter alkyl chains, when the films are transferred onto the solid substrates from the airwater interface. One of the strong points of AFM is that one can determine the film thickness. Table 1 shows the time-dependent variation in the thickness of a one-layer LB film of pentadecyl-TCNQ. The thickness of the fresh film of pentadecyl-TCNQ is estimated to be 4.6 nm, while that of the aged LB film is ∼8.0 nm, which is almost double of the original thickness. Discussion Aging on the Molecular Orientation, Structure, and Morphology in LB Films of the Three Kinds of AlkylTCNQ; Dependence upon the Length of the Hydrocarbon Chain Substituted. As described in the Results, the one-layer LB films of pentadecyl-TCNQ show marked time-dependent changes in the molecular orientation, structure, and morphology and those of dodecyl-TCNQ show moderate corresponding time-

Figure 12. AFM images of (a) a fresh (15 min after the deposition) and (b) an aged (120 min) one-layer LB films of pentadecyl-TCNQ on mica.

dependent changes. In contrast, the LB films of octadecyl-TCNQ do not show significant time-dependent changes in the molecular orientation, structure, and morphology. Therefore, one can conclude that the time-dependent structural and morphological changes depend strongly upon the length of the hydrocarbon chain substituted. There may be two major reasons for the dependence upon the length of the alkyl chains. One is the balance between the substrate-chromophore interaction and hydrophobic-hydrophobic interaction of the alkyl chains. The latter interaction is strong enough in the one-layer LB films of octadecyl-TCNQ, so that the molecular interaction, structure, and morphology change little with time. In contrast, the hydrophobic-hydrophobic interaction is not so strong in the LB films of pentadecyland dodecyl-TCNQ, and thus the interaction between the

LB Films of Three Decyl-TCNQs

J. Phys. Chem. B, Vol. 104, No. 6, 2000 1189

Figure 13. (a) AFM image of an aged (2 days after the deposition) one-layer LB film of dodecyl-TCNQ on mica. (b) AFM image of an aged (2 days after the deposition) one-layer LB film of octadecyl-TCNQ on mica.

TABLE 1: Time-Dependent Change in the Thickness of a One-Layer LB Film of Pentadecyl-TCNQ on Mica after the Film Deposition time/min thickness/nm

15 4.6

60 7.7

80 8.3

120 7.9

180 8.2

substrate and the TCNQ chromophore conquers the hydrophobichydrophobic interaction, causing the changes. Another possible origin for the dependence upon the length of the hydrocarbon chain is that dodecyl-TCNQ crystallizes more easily in the LB films than the others.20 The microcrystalline part in the LB films does not show a significant timedependent change, while the noncrystalline part shows the changes. This may explain the observation that the LB films of pentadecyl-TCNQ with the longer alkyl chain undergo much

larger time-dependent structural and morphological variations than those of dodecyl-TCNQ with the shorter chain. Dependence of the Structural and Morphological Changes upon the Substrate. The substrate dependence of the timedependent changes in the stacking and morphology in the onelayer LB films of pentadecyl-TCNQ is clearly observed in the UV-vis spectra (Figure 10). Particularly notable is that the large domains are partitioned into the smaller domains with the size of ∼1 µm in the LB films on both the CaF2 plate and goldevaporated glass slide, but the states in the molecular stacking and aggregation of the TCNQ chromophores are largely different between the two kinds of substrates (Figure 10). Since the interaction between the substrate and the TCNQ chromophore plays a key role in the time-dependent changes, the observed substrate dependence seems to be reasonable. Model for Aging Effects on the Structure and Morphology in LB films of Pentadecyl-TCNQ. By analyzing the height distribution in AFM images of one-layer and multilayer LB films of pentadecyl-TCNQ on a hydrophobidized quartz, Nichogi et al.20 showed that the first layer (5.7 nm) is thicker than the upper layers (3.4 nm). They also reported that the value of 3.4 nm is in good agreement with the periodic layer distance obtained from the X-ray diffraction pattern of the multilayer LB film of pentadecyl-TCNQ. The value of 4.6 nm, the observed height of the fresh onelayer LB film of pentadecyl-TCNQ, is larger than the layer distance of upper layers (3.4 nm). As described above, it is very likely in the fresh one-layer LB film of pentadecyl-TCNQ that the molecules direct more perpendicular to the substrate surface, compared to those in the aged films (Figure 4a). It is noted that the thickness of the aged one-layer LB film of pentadecyl-TCNQ is stabilized at ∼8.0 nm. This variation in the height value means that the one-layer LB film of pentadecyl-TCNQ changes its morphology from the unstable one-layer structure to the stable two-layer structure. The dramatic doubling of thickness of the LB film of pentadecyl-TCNQ requires a great deal of molecular motion, and thus may induce significant band shifts in the infrared spectra. However, all the spectral variations observed during the film folding transition are intensity changes in bands as described above, and no band shift was detected. Of note is that the relative intensity of two bands at 2926 and 2918 cm-1 change largely with time. This observation suggests that the alkyl chain has more all-trans nature in the stable two-layer structure.21-23 It is likely that water molecules attaching to the alkyl chain evaporate with time, leading to the stable crystalline state where the alkyl chain tends to the all-trans conformation. The film folding transition does not bring about frequency shifts for the bands due to the TCNQ chromophore. Thus, the structure of the TCNQ chromophore does undergo a significant change upon the formation of two-layer structure. In other words, the interaction between the first and second layers in the double layer structure is not so strong enough to cause a band shift in the infrared spectra. On the basis of the results of infrared and UV-vis spectroscopy and AFM, we have constructed a new model for the aging effects on the structure and morphology of a one-layer LB film of pentadecyl-TCNQ. Figure 14 illustrates the models. Two possibilities are presented in Figure 14. The top model shows that the thickness of the bottom layer of the aged one-layer LB film is very close to that of the fresh film and that the top layer has a thinner thickness with tilted hydrocarbon chains. The bottom model depicts that both the bottom and top layers have

1190 J. Phys. Chem. B, Vol. 104, No. 6, 2000

Morita et al. hydrophobic interaction between the alkyl chains, microcrystallization of the LB films, and the substrate-first layer interaction are key factors for the dependence upon the length of the alkyl chain. The changes are also dependent upon the substrate employed; it seems that the substrate-TCNQ chromophore interaction is quite different between the CaF2 plate and gold-evaporated glass slide. The present study has provided new insight into the control of molecular orientation in onelayer LB films and stability of the LB films. References and Notes

Figure 14. Model for time-dependent morphological and structural changes in the one-layer LB film of pentadecyl-TCNQ.

the similar thickness with tilted hydrocarbon chains. At the moment, we cannot conclude which model is more likely. Conclusion This paper has reported the time-dependent changes of molecular orientation, structure, and morphology in the onelayer LB films of dodecyl-, pentadecyl-, and octadecyl-TCNQ studied by infrared and UV-vis spectroscopies and AFM. The following conclusions can be reached for the LB films of pentadecyl-TCNQ on the CaF2 plate and gold-evaporated glass slide. In the LB films, the alkyl chain and the molecular axis of the TCNQ chromophore become more tilted with respect to the surface normal and the TCNQ plane becomes more perpendicular with respect to the substrate surface during the time course (Figures 4a and 14). (ii) Some of the stacked TCNQ chromophores become the aggregated forms in the LB film on the CaF2 plate with time. (iii) The large domains in the LB films are divided into the smaller ones with the size of ∼1 µm as a function of time (Figures 4b and 14). (iv) The flat and uniformed structure of the one-layer LB film changes into the condensed and accumulated structure; the unstable one-layer structure becomes the stable two-layer structure (Figures 4a and 14). These variations in molecular orientation, structure, and morphology show marked dependence upon the length of the alkyl chain substituted; the LB films of dodecyl-TCNQ show much smaller changes while those of octadecyl-TCNQ do not show a significant change. Probably, the hydrophobic-

(1) Roberts, G. G. Langmuir-Blodgett Films; Plenum Press: New York and London, 1990. (2) Nakamura, T. In Handbook of Organic ConductiVe Molecules and Polymers; Nalwa, H. S., Ed.; John Wiley & Sons: Chichester, U.K., 1997; Vol. 1, p 727. (3) Williams, J. M.; Ferrano, J. R.; Thorn, R. J.; Carson, K. D.; Geiser, U.; Wang, H. H.; Kini, A. M.; Whangbo, M. Organic Superconductors; Prentice Hall: Englewood Cliffs, NJ, 1992. (4) Kubota, M.; Ozaki, Y.; Araki, T.; Ohki, S.; Iriyama, K. Langmuir 1991, 7, 774. (5) Terashita, S.; Nakatsu, K.; Ozaki, Y.; Mochida, T.; Araki, T.; Iriyama, K.; Langmuir 1992, 8, 3051. (6) Terashita, S.; Ozaki, Y.; Iriyama, K. J. Phys. Chem. 1993, 97, 10454. (7) Terashita, S.; Ozaki, Y.; Yageta, H.; Kudo, K.; Iriyama, K. Langmuir 1994, 10, 1807. (8) Wang, Y.; Ozaki, Y.; Iriyama, K. Langmuir 1995, 11, 705. (9) Wang, Y.; Nichogi, K.; Terashita, S.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. 1996, 100, 368. (10) Wang, Y.; Nichogi, K.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. 1996, 100, 374. (11) Wang, Y.; Nichogi, K.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. 1996, 100, 17232. (12) Wang, Y.; Nichogi, K.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. 1996, 100, 17238. (13) Wang, Y.; Nichogi, K.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. B 1997, 101, 6367. (14) Wang, Y.; Nichogi, K.; Iriyama, K.; Ozaki, Y. J. Phys. Chem. B 1997, 101, 6372. (15) Greenler, T. G. J. Chem. Phys. 1966, 44, 310. (16) Chollet, P. A.; Messier, J.; Rosilio, C. J. Chem. Phys. 1976, 64, 1042. (17) Chollet, P.-A. Thin Solid Films 1978, 52, 343. (18) Umemura, J.; Kamata, T.; Kawai, T.; Takenaka, T. J. Phys. Chem. 1990, 94, 62. (19) Morita, S.; Nichogi, K.; Ozaki, Y. To be submitted. (20) Nichogi, K.; Nambu, T.; Miyamoto, A.; Murakami, M. Jpn. J. Appl. Phys. 1995, 34, 4958. (21) Umemura, J.; Cameron, D. G.; Mantsch, H. H. Biochim. Biophys. Acta 1980, 602, 32. (22) Shapper, H.; Cameron, D. G.; Mantsch, H. H. Can. J. Chem. 1981, 59, 2543. (23) Tian, Y. J. Phys. Chem. 1991, 95, 9995.