In Situ Infrared Spectroscopy Study on Imidization Reaction and

These decreases in the d indicate that in this temperature region the d reductions caused by the evaporation of residual solvent still override the d ...
3 downloads 0 Views 226KB Size
Download PDF
Langmuir 2005, 21, 6081-6085

6081

In Situ Infrared Spectroscopy Study on Imidization Reaction and Imidization-induced Refractive Index and Thickness Variations in Microscale Thin Films of a Poly(amic ester) Tae Joo Shin* National Synchrotron Light Source, Brookhaven National Laboratory, Upton, New York 11973

Moonhor Ree* Department of Chemistry, Center for Integrated Molecular Systems, Polymer Research Institute, and Division of Molecular and Life Sciences (BK21 Program), Pohang University of Science & Technology, Pohang 790-784, Republic of Korea Received February 22, 2005. In Final Form: April 14, 2005 Poly(amic ester) (PAE) is a soluble precursor of polyimide that has attracted interest from both the microelectronic and the flat-panel display industries because of its several important advantages, including excellent solubility, high hydrolytic stability, and solvent-free film formation, over the polyimide precursor, poly(amic acid), for which monomer-polymer equilibration always occurs in solution due to its carboxylic acid groups. In this study, poly(3,4′-oxydiphenylene pyromellitamic diethyl ester) (PMDA-3,4′-ODA PAE) was chosen as a PAE precursor, and its thermal imidization behavior in microscale thin films was investigated quantitatively for the first time using time-resolved infrared (IR) spectroscopy. In addition, the variations of the film refractive index and thickness with temperature and time were determined in detail from the time-resolved IR spectra and are fully interpreted in this paper by considering the imidization kinetics of the precursor.

Introduction Aromatic polyimides (PIs) have excellent mechanical, electrical, chemical, and thermal properties and are therefore widely used as electrical insulators, passivation layers, liquid-crystal alignment layers, gas separation membranes, and matrix resins for reinforced plastics.1-5 Most aromatic PIs are insoluble and have relatively high glass transition temperatures (Tg) and are thus usually * To whom correspondence should be addressed. (T.J.S.) Tel: +1-631-344-5778. Fax: +1-631-344-3238. E-mail: [email protected]. (M.R.) Tel: +82-54-279-2120. Fax: +82-54-279-3399. E-mail: [email protected]. (1) (a) Ghosh, M. K.; Mittal, K. L. Polyimides: Fundamentals and Applications; Marcel Dekker: New York, 1996. (b) Czornyj, G.; Chen, K. J.; Pradasilva, G.; Arnold, A.; Souleotis, H.; Kim, S.; Ree, M.; Volksen, W.; Dawson, D.; DiPietro, R. Proc. Electron. Comput. Technol. (IEEE) 1992, 42, 682. (c) Yu, J.; Ree, M.; Shin, T. J.; Wang, X.; Cai, W.; Zhou, D.; Lee, K.-W. Polymer 2000, 41, 169. (d) Pyo, S. M.; Kim, S. I.; Shin, T. J.; Park, Y. H.; Ree, M. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 937. (e) Robertson, W. M.; Arjavalingam, G.; Hougham, G.; Kopcsay, G. V.; Edelstein, D.; Ree, M.; Chapple-Sokol, J. P. Electron. Lett. 1992, 28, 62. (f) Ree, M.; Kim, K.; Woo, S. H.; Chang, H. J. Appl. Phys. 1997, 81, 698. (2) (a) Lee, S. W.; Kim, S. I.; Lee, B.; Kim, H. C.; Chang, T.; Ree, M. Langmuir 2003, 19, 10381. (b) Chae, B.; Lee, S. W.; Lee, B.; Choi, W.; Kim, S. B.; Jung, Y. M.; Jung, J. C.; Lee, K. H.; Ree, M. Langmuir 2003, 19, 9459. (c) Lee, S. W.; Kim, S. I.; Lee, B.; Choi, W.; Chae, B.; Kim, S. B.; Ree, M. Macromolecules 2003, 36, 6527. (d) Chae, B.; Kim, S. B.; Lee, S. W.; Kim, S. I.; Choi, W.; Lee, B.; Ree, M.; Lee, K. H.; Jung, J. C. Macromolecules 2002, 35, 10119. (e) Lee, S. W.; Chang, T.; Ree, M. Macromol. Rapid Commun. 2001, 22, 941. (3) (a) Han, H.; Chung, H.; Gryte, C. C.; Shin, T. J.; Ree, M. Polymer 1999, 40, 2681. (b) Kim, S. I.; Shin, T. J.; Ree, M.; Hwang, G. T.; Kim, B. H.; Han, H.; Seo, J. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 2013. (4) (a) Pyo, S. M.; Kim, S. I.; Shin, T. J.; Ree, M.; Park, K. H.; Kang, J. S. Macromolecules 1998, 31, 4777. (5) (a) Kim, Y.; Kang, E.; Kwon, Y. S.; Cho, W. J.; Chang, C.; Ree, M.; Chang, T.; Ha, C. S. Synth. Met. 1997, 85, 1399. (b) Ree, M.; Goh, W. H.; Kim, Y. Polym. Bull. 1995, 35, 215. (c) Ree, M.; Shin, T. J.; Kim, S. I.; Woo, S. H.; Yoon, D. Y. Polymer 1998, 39, 2521.

synthesized via soluble precursor forms that are processed in various ways before their conversion to PIs.2-8 Poly(amic acid) (PAA) is a representative of one type of soluble PI precursors, which are soluble in aprotic solvents such as N-methyl-2-pyrrolidone (NMP) and N,N-dimethylacetamine.6,7 PAA is composed of two carboxylic acids and two amide groups per repeat unit and favorably forms complexes with the molecules of these solvents via hydrogen bonding. As a result, films of PAA always contain some residual solvent even after drying. During the thermal imidization of PAA, the residual solvent molecules are evaporated along with the water byproduct generated by imidization. This evaporation results in a significant reduction in the thickness of the film and makes the quantitative determination of the imidization behavior difficult.6,7 Poly(amic ester) (PAE) is representative of another type of soluble PI precursor.8 PAE has better solubility than PAA, as well as higher hydrolytic stability due to the absence of the carboxylic acid groups that cause monomerpolymer equilibration in solution.8 Moreover, films of PAE (6) (a) Kim, S. I.; Pyo, S. M.; Ree, M. Macromolecules 1997, 30, 7890. (b) Ree, M.; Yoon, D. Y.; Volksen, W. J. Polym. Sci., Part B: Polym. Phys. 1991, 29, 1203. (c) Kim, K.; Ree, M. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 1755. (d) Kim, S. I.; Pyo, S. M.; Kim, K.; Ree, M. Polymer 1998, 39, 6489. (7) (a) Shin, T. J.; Lee, B.; Youn, H. S.; Lee, K.-B.; Ree, M. Langmuir 2001, 17, 7842. (b) Shin, T. J.; Ree, M. Macromol. Chem. Phys. 2002, 203, 781. (8) (a) Labadie, J.; Lee, H.; Boese, D.; Yoon, D. Y.; Volksen, W.; Brock, P.; Cheng, Y. Y.; Ree, M.; Chen, K. R. Proc. Electron. Comput. Technol. (IEEE) 1993, 43, 327. (b) Ree, M.; Goh, W. H.; Park, J. W.; Lee, M. H.; Rhee, S. B. Polym. Bull. 1995, 35, 129. (c) Kim, S. I.; Ree, M.; Shin, T. J.; Lee, C.; Woo, T.-H.; Rhee, S. B. Polymer 2000, 41, 5173. (d) Chang, H.; Kim, K.; Ree, M.; Lee, K.-W. Macromol. Chem. Phys. 1999, 200, 422. (e) Kim, Y.; Ree, M.; Chang, T.; Ha, C. S.; Nunes, T. L.; Lin, J. S. J. Polym. Sci., Part B: Polym. Phys. Ed. 1995, 33, 2075.

10.1021/la050470c CCC: $30.25 © 2005 American Chemical Society Published on Web 05/17/2005

6082

Langmuir, Vol. 21, No. 13, 2005

Figure 1. PMDA-3,4′-ODA PAE precursor and its PI.

can be produced that are free of residual solvent, which means that there is no reduction in the thickness of the film that is associated with the removal of residual solvent in the PI film formation process. For these reasons, PAE has recently gained increased attention from industry.8 Despite this interest in PAE, its imidization behavior and resulting properties, which are essential to understanding PI formation and the structure and properties of the polymer, have rarely been investigated in detail. In the present study, we chose poly(3,4′-oxydiphenylene pyromellitamic diethyl ester) (PMDA-3,4′-ODA PAE) as our model PAE precursor and then investigated its thermal imidization behavior in microscale thin films using timeresolved infrared (IR) spectroscopy. From the timeresolved IR spectra obtained during the thermal imidization of the PAE precursor, we were able to determine the variations of the refractive index and film thickness with temperature and time, which we then interpreted in detail by considering the imidization kinetics of the precursor in the film.

Shin and Ree

Figure 2. Interference IR spectral patterns of a microscale film of PMDA-3,4′-ODA PAE precursor measured over the range of 4000-6200 cm-1 during thermal imidization with a heating rate of 2.0 °C/min in the range of 30-400 °C. at a rate of 2.0 °C/min. The IR beam irradiation area in the samples was around 10 mm in diameter. Temperature-dependent background spectra were measured and then used to correct all of the measured spectra.

Results and Discussion Figure 2 shows typical transmission IR spectra of a film of the PMDA-3,4′-ODA PAE precursor obtained for the range of 4000-6200 cm-1 during heating to 400 °C at a rate of 2.0 °C/min. For this spectral region corresponding to the 2.5-1.6 µm wavelength range, the PAE precursor and its PI do not exhibit any characteristic IR absorption peaks, as is observed for some PIs.9 Thus, the oscillations within the spectral region are interference fringes attributed to the phase difference δ and reflectivity R of the precursor film. The interference fringe pattern at a given temperature can be interpreted by considering the spectral transmittance T of the precursor film, which is a function of its phase difference δ and reflectivity R. The spectral transmittance T can be expressed as follows:10

Experimental Section PMDA-3,4′-ODA PAE was synthesized in dry NMP by the polycondensation of equivalent pyromellitic diethyl ester diacyl chloride and 3,4′-oxydiphenylene diamine (Figure 1), according to a previously reported method.8 The PAE precursor was determined to have an intrinsic viscosity of 1.102 dL/g in NMP at 25.0 °C and a weight average molecular weight of 25 000 with gel permeation chromatography. A solution of the PAE precursor was prepared in dry NMP with a solid content of 13.5 wt %. The precursor solution was spin-coated onto silicon wafers and then dried on a hotplate at 80 °C for 2 days in ambient air, resulting in thin PAE precursor films 3.8-4.0 µm in thickness. The dried films were determined to contain residual solvent of around 7 wt % by using proton nuclear magnetic resonance spectroscopy. Each thin precursor film was carefully detached from the silicon substrate and mounted into a specially designed IR sample cell, in which the thin film was held tightly between two aluminum plates with a 10 mm diameter hole in the center. This IR cell was then mounted into a heating chamber with a nitrogen gas blowing hole to prevent any undesirable oxidation of the precursor film at high temperatures; the temperature was controlled by a Eurotherm temperature controller with a K type thermocouple. The heating chamber was installed onto the sample stage of an IR spectrometer (Mattson Research Series) equipped with an MCT detector cooled with liquid nitrogen. The spectrometer was calibrated using a film of polystyrene standard. The spectra were recorded in transmission mode using 42 scans at a spectral resolution of 4.0 cm-1. The IR spectra were collected as a function of temperature and time while the sample was heated to 400 °C

T)

(1 - R) I ) Io 1 + R2 - 2Rcosδ

(1)

where Io and I are the intensities of the incident and transmitted IR beams, respectively; R ) [(1 - nf)/(1 + nf)]2, where nf is the refractive index of the film; δ ) (4π/ λ)nfd ) 4πνnfd, where λ is the wavelength, ν is the wavenumber, and d is the film thickness. In eq 1, the minimum T value arises when cosδ ) -1. From the minimum T value, nf can be determined using eq 1. The film thickness can be further determined from this value for nf, using either the relationship cosδ ) -1 () cos{4πνnfd}) or the periodic distance (i.e., peak-to-peak distance: frequency) of the interference fringe pattern; in general, thicker films produce higher frequencies in the interference fringe pattern. Therefore, the variations with temperature of the interference fringe patterns’ amplitude and frequency indicate changes with temperature in the refractive index nf and film thickness d, respectively. Taking these considerations into account, the interference fringe patterns in Figure 2 were analyzed in detail (9) (a) Saito, M.; Gojo, T.; Kato, Y.; Miyagi, M. Infrared Phys. Technol. 1995, 36, 1125-1129. (b) Zhang, Z. M.; Lefever-Button, G.; Powell, F. R. Int. J. Thermophys. 1998, 19, 905. (10) Hecht, E. Optics, 2nd ed.; Addison-Wesley: New York, NY, 1987.

Microscale Thin Films of a Poly(amic ester)

Langmuir, Vol. 21, No. 13, 2005 6083

Figure 3. Representative interference IR patterns of a PMDA3,4′-ODA PAE film selected from the data of Figure 2: The symbols represent the measured data, and the solid lines were obtained by fitting the data with eq 1.

Figure 5. IR spectra of a microscale film of PMDA-3,4′-ODA PAE precursor measured over the range of 1200-1830 cm-1 during thermal imidization with a heating rate of 2.0 °C/min over the range of 30-400 °C.

Figure 4. Variations with temperature of the refractive index nf, thickness d, degree of imidization Ximidization, and its first derivative dXimidization/dT for a microscale film of PMDA-3,4′ODA PAE precursor, obtained from analyses of the interference IR spectral patterns in Figure 2 and the IR absorption spectra in Figure 5, which were measured during thermal imidization with a heating rate of 2.0 °C/min over the range of 30-400 °C.

using eq 1. As can be seen in Figure 3, the interference fringe patterns can be satisfactorily fitted with eq 1, demonstrating that this formula is a powerful tool for determining both the refractive index nf and the thickness d of a thin film from its spectral interference fringe pattern. The results for nf and d are plotted in Figure 4a,b, respectively. During the heating run, nf gradually decreases from 1.4804 at 30 °C to 1.4684 at 236 °C, then suddenly drops to its minimum value, 1.4586, at 256 °C, and thereafter increases with temperature, reaching its maximum, 1.5675, at 336 °C (Figure 4a). After reaching this maximum, nf again decreases very slowly with further increases in temperature (Figure 4a). The film thickness d varies slightly in the 3.897-3.961 µm range with increases in temperature up to 216 °C, and then rapidly drops with further increases in temperature, ultimately reaching 2.852 µm at 400 °C (Figure 4b). The variations of nf and d with temperature are due to the thermal imidization and associated structure formation that occurs

during the heating run. These results are discussed by considering the imidization kinetics of the PAE precursor film later in this paper. Figure 5 shows the IR absorption spectra of a film of the PMDA-3,4′-ODA PAE precursor in situ monitored over the range of 1200-1830 cm-1 during heating to 400 °C at a rate of 2.0 °C/min. In these spectra, some characteristic IR peaks can be assigned with the aid of previously reported results for PAA precursors, their PIs, and model compounds.7,11 In the spectrum of the PAE precursor obtained at 30 °C, the absorption peak centered at 1662 cm-1 is assigned to the characteristic vibration of amide I, namely, carbonyl (-CdO-) stretching in the amide linkage, and the double peaks at 1540 cm-1 are assigned to vibrations of amide II, namely, -CNH- bend-stretch vibrations. The single peak at 1363 cm-1 is assigned to characteristic imide -CNstretching in the imide ring. Taking these IR peak assignments into account, the in situ monitored IR spectra in Figure 5 show that during the heating run the characteristic amide peaks vary little with increasing temperature below 216 °C and then drastically decrease in intensity with further increases in temperature, before ultimately disappearing. These decreases in the intensity of the amide peaks above 216 °C are attributed to the consumption of the amide linkages by imide ring formation. Other the other hand, the imide peak appears for the first time around 216 °C and then becomes stronger with further increases in temperature, indicating the formation of imide rings above 216 °C. These results indicate that the imidization reaction takes place mainly above 216 °C during the heating run. Taking these results into account, the imide peak at 1363 cm-1 was chosen from the assigned IR peaks and used for estimating the degree of imidization (Ximidization) as a function of temperature; here, the Ximidization value at a chosen temperature was determined from the area of the imide peak measured at that temperature with respect to the imide peak area measured at 400 °C as described in the literature.7 The (11) (a) Ishida, H.; Wellinghoff, S. T.; Baer, E.; Koenig, J. L. Macromolecules 1980, 13, 826-834. (b) Kumar, D. J. Polym. Sci., Part A: Polym. Chem. Ed. 1981, 19, 795-805. (c) Thomson, B.; Park, Y.; Painter, P. C.; Snyder, R. W. Macromolecules 1989, 22, 4159-4166. (d) Synder, R. W.; Thompson, B.; Bartges, B.; Czerniawski, D.; Painter, P. C. Macromolecules 1989, 22, 4166-4172. (e) Ortelli, E. E.; Geiger, F.; Lippert, T.; Wei, J.; Wokaun, A. Macromolecules 2000, 33, 5090-5097. (f) Pryde, C. A. J. Polym. Sci., Part A: Polym. Chem. Ed. 1989, 27, 711-724.

6084

Langmuir, Vol. 21, No. 13, 2005

variation of the degree of imidization with temperature is shown in Figure 4c. As can be seen in Figure 4c, the PAE precursor film starts to imidize at 216 °C and has undergone 97% imidization by 356 °C, finally reaching 100% imidization above that temperature. This progression of thermal imidization is quite different from those of PAA precursors in films.6,7,11 Figure 4c also presents a plot of d(Ximidization)/ dT vs T; the maximum at 256 °C indicates that the imidization reaction takes place with a maximum rate at that temperature. At this temperature, the maximum imidization rate is much higher than those observed for films of PAA precursors.7 Considering these imidization kinetics, the variations of the refractive index nf and film thickness d with temperature in Figure 4a,b can be interpreted as follows. The temperature range of the heating run can be split up into several regions bounded by the drying temperature (Tdrying) and the onset imidization temperature (Timidization) of the precursor film. In the temperature region of room temperature to 100 °C, 20 °C higher than Tdrying, the film thickness increases slightly with increases in temperature. The precursor film was dried for 2 days at 80 °C; thus, there should be no further significant removal of residual solvent in this temperature region. Taking these data into account, both the slow decrease in nf and the slow increase in d with temperature in this temperature region are due mainly to the film’s thermal expansion. In the second temperature region (100-135 °C), with increases in temperature, the nf turns to very slightly increase but the d also turns to slowly decrease. No imidization occurs in this temperature range (Figure 4c). Thus, because of thermal expansion, the refractive index nf is expected to continuously decrease with temperature and the film thickness d is expected to continuously increase with temperature. However, the observations are opposite to these expectations. Consequently, the observed d variations with temperature indicate that in the temperature region the evaporation of the residual NMP solvent (about 7 wt %) in the film significantly takes place in comparison to that which occurred below 100 °C, leading to reductions in the film thickness, which can override the film thermal expansions. Here, it is noteworthy that residual NMP has a relatively lower refractive index (nNMP ) 1.4684) than that of the film itself. Taking the d variations and the relatively lower refractive index of NMP into account, the observed nf variations with temperature are attributed to the increases in the nf caused by the evaporation of residual NMP solvent, which override the contributions due to the film’s thermal expansion. In the third temperature region (135-186 °C), the d continues to slowly decrease with increases in temperature but the nf again turns to slightly decrease with temperature. These decreases in the d indicate that in this temperature region the d reductions caused by the evaporation of residual solvent still override the d increases due to the film’s thermal expansion. Taking the d reductions into account, the nf is expected to continuously increase with temperature. However, the observed nf variations are contradicted to the expectation. Therefore, the observed nf reductions with temperature suggest that the nf decreases due to the film’s thermal expansion in the competition with the shrinkage by the continuous evaporation of the residual solvent override the nf increases due to the evaporation of the residual solvent. In the fourth temperature region (186 °C to Timidization), the nf continues to slightly decrease with increases in temperature. The film thickness d is expected to continuously decrease with temperature because of the evapora-

Shin and Ree

tion of residual solvent. However, the d turns to increase slightly with temperature. No imidization still occurs in this temperature region (Figure 4c). Thus, the small increases in the d might be due to the film’s thermal expansion, which overrides the contribution of shrinkage due to any further evaporation of residual solvent. In fact, the observation of these increases in the d indicates that in this temperature range there is only negligible shrinkage of the film due to evaporation of residual solvent. These d variations further indicate that on the heating run the residual solvent is almost completely removed out from the film at temperatures below 186 °C, still far below Timidization. Here, it is worthy to additionally discuss the glass transition of the PAE precursor film. The glass transition temperature Tg of the PMDA-3,4′-ODA PAE precursor is not yet determined. However, poly(4,4′oxydiphenylene pyromellitamic diethyl ester) (PMDA-4,4′ODA PAE), a structural isomeric precursor of the PMDA3,4-ODA PAE, was determined to have a Tg of 189.8 °C for its powdery samples without any residual solvent and a Tg of 226.3 °C for its thin film samples without any residual solvent;12 here, the Tg is the onset temperature of glass transition. The PMDA-3,4′-ODA PAE film in our study has more link nature on the backbone, as compared to that of the PMDA-4,4′-ODA PAE film, and further contains 7 wt % residual NMP solvent; thus, its Tg is expected to be relatively lower than that of the PMDA4,4′-ODA PAE film. Taking these facts into account, the observed d increases in the range of 186 °C to Timidization are attributed to the relatively large thermal expansions of the film occurred above the Tg of the film. Considering these d increases associated with large thermal expansions above Tg, the PAE precursor thin film in our study is presumed to start glass transition at 186 °C on the heating run. In the fifth region (above Timidization), both nf and d vary significantly with temperature (Figure 4a,b). The film was found to shrink drastically with increases in temperature, which is due to the evaporation of the ethanol byproducts formed by imidization. Despite the initiation of imidization at 216 °C, nf continues the slow decrease observed below Timidization for increases in temperature up to 236 °C. At this temperature, the degree of imidization Ximidization is only about 4.8%. Thus, the slow decrease in nf might be due to the contribution of the film’s thermal expansion, which overrides that of the shrinkage associated with a degree of imidization of around 4.8%. Above 236 °C, nf starts to decrease rapidly and reaches a minimum at 256 °C, at which point the degree of imidization is about 33.6%. In this narrow temperature range, imidization reactions occur with very rapid rates (Figure 4c), indicating that the precursor polymer chains are highly mobile in the temperature range. Furthermore, such high polymer chain mobilities indicate that over the temperature range the film is above its Tg. These two factors lead a large increase in the film’s thermal expansion. Therefore, the rapid decrease in nf is due to the film’s thermal expansion, which significantly overrides the contribution of the imidizationinduced shrinkage. Taking into account the high polymer chain mobilities in the temperature range, the observed shrinkage might be mainly attributed to mass losses via the evaporation of the ethanol byproducts formed by imidization; in comparison, the contribution of densification to the film shrinkage might be small. In contrast, above 256 °C, nf increases steeply with temperature, which might be due to the significant densification and shrinkage of the film as well as the high chain rigidities of polymer (12) Kim, S. I.; Shin, T. J.; Ree, M. Polymer 1999, 40, 2263.

Microscale Thin Films of a Poly(amic ester)

molecules containing a large number of imide rings. Finally, above 336 °C, nf decreases slightly with temperature. This decrease in nf might be due to the thermal expansion of the film, which is >97% imidized in this temperature region. Finally, it is worthy to here compare the imidization behaviors of the PAE precursor films with those of the corresponding PAA precursor films. For the PAA films with a thickness of 5.5 µm prepared in the same manner as in our study, in the same heating run condition, the films were found to begin imidization reaction at 124 °C and undergo the majority of imidization by 210 °C; imidization then slowly proceeded until it was complete by 310 °C.7 The maximum rate of imidization was found to occur at 148.4 °C.7 These imidization behaviors were found to be associated strongly with residual NMP solvents complexed with the PAA precursor molecules; here, the PAA films contained about 32 wt % NMP.7 Overall, the PAA precursor films undergo an imidization reaction at much lower temperatures, as compared to the PAE precursor films. Comparing the results of our present study and those of the PAA films, the significant differences in imidization reaction are attributed to their chemically different characteristics in the nature of imidization as well as the residual solvent molecules strongly complexed with PAA precursor; in contrast to the PAA precursor, the PAE precursor molecules were found to undergo imidization reaction in free of residual solvent. Conclusions The microscale thin films of the PAE precursor and the resulting PI were found to exhibit characteristic vibrational peaks in the range of 600-3000 cm-1 but only interference fringe patterns without any vibrational absorption in the range of 4000-6200 cm-1. In addition, films of the PAE precursor were obtained free of residual solvent without any imidization. These IR vibrational characteristics and the formation of films without residual solvent enabled us to determine in detail the thermal imidization kinetics of microscale thin films of the PAE precursor for the first time by using in situ IR spectroscopy

Langmuir, Vol. 21, No. 13, 2005 6085

measurements during a heating run. Because the interference fringe patterns have no overlap with any vibrational absorption, we were able to use in situ IR spectroscopy to determine the variations of the PAE precursor film refractive index and thickness during thermal imidization. The in situ IR spectroscopy measurements were used to show that thin films of the PAE precursor start to undergo imidization at 216 °C during heating runs at a rate of 2.0 °C/min in nitrogen atmosphere and that the maximum imidization reaction rate arises at 256 °C. These imidization characteristics are quite different from those of PAA precursors in films, which unavoidably contain some amount of the aprotic solvents used in film preparation; imidization in PAA films is known to be activated and accelerated by the residual aprotic solvents in these films.6,7,11 The film refractive index was found to vary in a very complex manner during the thermal imidization of the PAE films, reflecting the changes in the film caused by both the imidization and the variations of thermal expansion and density with temperature. Through the whole thermal imidization process, the refractive index was found to vary by 7.4% with respect to that of the PAE film at room temperature. The film thickness was found to vary slightly with temperature due to thermal expansion and the evaporation of residual solvent in small amounts below the onset imidization temperature Timidization () 216 °C) but to undergo a significant reduction above Timidization due to the evaporation of the ethanol byproducts of the imidization reaction and the densification of the resulting PI molecules. During the heating run, the film thickness was found to vary by 28.2% with respect to that of the PAE film at room temperature. Acknowledgment. This study was supported by the Center for Integrated Molecular Systems (Korea Science & Engineering Foundation). LA050470C