Stimuli-Responsive Mixed Grafted Polymer Films with Gradually

Aug 10, 2005 - Infrared spectroscopic ellipsometry (IRSE) and visible monochromatic ellipsometry (VISE) approaches were applied to investigate the che...
0 downloads 0 Views 271KB Size
Download PDF
Langmuir 2005, 21, 8711-8716

8711

Stimuli-Responsive Mixed Grafted Polymer Films with Gradually Changing Properties: Direct Determination of Chemical Composition Leonid Ionov, Alexander Sidorenko, Klaus-Jochen Eichhorn, and Manfred Stamm Leibniz Institute for Polymer Research Dresden, Dresden, Germany

Sergiy Minko* Chemistry Department, Clarkson University, Potsdam, New York 13699

Karsten Hinrichs ISAS - Institute for Analytical Sciences, Department Berlin, Albert-Einstein-Strasse 9, 12489 Berlin, Germany Received March 7, 2005. In Final Form: June 21, 2005 Infrared spectroscopic ellipsometry (IRSE) and visible monochromatic ellipsometry (VISE) approaches were applied to investigate the chemical structure and thickness of ultrathin polymer films. Mixed polystyrene-poly(2-vinylpyridine) and polystyrene-poly(tert-butyl acrylate) polymer grafted films (mixed brushes) with gradually changing composition (1D gradient mixed brush) along the sample were prepared on a temperature gradient stage via two subsequent “grafting to” reactions. The films were characterized by high-precision mapping VISE at a single wavelength (632.8 nm) and IRSE. The set of 1D IRSE spectra of the polymer brush films obtained by mapping the 1D gradient brush were used to estimate the thickness and the local composition of the film and to construct the 1D map of the film in terms of the chemical composition of the brush. The results were compared with the data obtained using monochromatic ellipsometry where the brush composition was estimated from the results of two subsequent measurements followed each grafting step. The measurements of the brush thickness and composition with both methods were found to be in gratifying agreement. The results demonstrate the high potential of IRSE methods for the one-step characterization (by thickness and chemical composition) of ultrathin polymer films of complex composition.

Introduction Complex thin polymer films are important for various applications such as biomaterials, sensors, microfluidic devices, and so forth. Thin films responsive to external stimuli (pH, solvent, light, electric field)1-5 are broadly used for the regulation of wettability,6,7 adhesion,8,9 and the design of sensors.10 Tuning a large variety of structural parameters is required for the effective design and study of such thin films.11 A combinatorial approach recently * Corresponding author. E-mail: [email protected]. (1) Ito, Y.; Park, Y. S.; Imanishi, Y. J. Am. Chem. Soc. 1997, 119, 2739-2740. (2) Russell, T. P. Science 2002, 297, 964-967. (3) Lahann, J.; Mitragotri, S.; Tran, T.-N.; Kaido, H.; Sundaram, J.; Choi, I. S.; Hoffer, S.; Somorjai, G. A.; Langer, R. Science 2003, 299, 371-374. (4) Ito, Y.; Nishi, S.; Park, Y. S.; Imanishi, Y. Macromolecules 1997, 30, 5856-5859. (5) Minko, S.; Usov, D.; Goreshnik, E.; Stamm, M. Macromol. Rapid Commun. 2001, 22, 206-211. (6) Wang, R.; Hashimoto, K.; Fujishima, A.; Chikuni, M.; Kojima, E.; Kitamura, A.; Shimohigoshi, M.; Watanabe, T. Nature 1997, 388, 431432. (7) Ichimura, K.; Oh, S.; Nakagawa, M. Science 2000, 288, 1624. (8) Lemieux, M.; Usov, D.; Minko, S.; Stamm, M.; Shulha, H.; Tsukruk, V. V. Macromolecules 2003, 36, 7244-7255. (9) Minko, S.; Muller, M.; Motornov, M.; Nitschke, M.; Grundke, K.; Stamm, M. J. Am. Chem. Soc. 2003, 125, 3896-3900. (10) Ionov, L.; Minko, S.; Stamm, M.; Gohy, J. F.; Jerome, R.; Scholl, A. J. Am. Chem. Soc. 2003, 125, 8302-8306. Tokareva, I.; Minko, S.; Fendler, J. H.; Hutter, E. J. Am. Chem. Soc. 2004, 126, 15950-15951. (11) Luzinov, I.; Minko, S.; Tsukruk, V. V. Prog. Polym. Sci. 2004, 29, 635-698.

introduced in material science and, particularly, in the study of complex thin polymer films has attracted great interest.12,13 The polymer films with gradual changes in composition and, therefore, physical properties allow for the fast and effective study of properties/morphologychemical composition relationships. An analytical method for the rapid investigation of gradient films is a critical issue for the high efficiency of the combinatorial approach. In this study, we provide experimental evidence for the high potential of visible monochromatic (VISE) and infrared spectroscopic (IRSE) ellipsometry in studying the thickness and chemical structure/composition of thin polymer films. Polymer chains grafted to solid substratesspolymer brushessrepresent a remarkable class of stimuli-responsive films.14 The conformation of the tethered polymer chains is strongly influenced by the brush environment affecting the switching properties of the grafted layers.15,16 The grafting of two or more different kinds of polymer (12) Smith, A. P.; Douglas, J. F.; Meredith, J. C.; Amis, E. J.; Karim, A. J. Polym. Sci., Part B: Polym. Phys. 2001, 18, 2141-2158. (13) Smith, A. P.; Meredith, J. C.; Douglas, J. F.; Amis, E. J.; Karim, A. Phys. Rev. Lett. 2001, 87, 015503. (14) Minko, S.; Mu¨ller, M.; Luchnikov, V.; Motornov, M.; Usov, D.; Ionov, L.; Stamm, M. In Polymer Brushes; Ruehe, J., Ed.; Wiley-VCH: Weinheim, Germany, 2004; pp 403-425. (15) Ionov, L.; Zdyrko, B.; Sidorenko, A.; Minko, S.; Klep, V.; Luzinov, I. Macromol. Rapid Commun. 2004, 25, 360-365. (16) Zhulina, E. B.; Birshtein, T. M.; Priamitsyn, V. A.; Klushin, L. I. Macromolecules 1995, 28, 8612-8620.

10.1021/la050620j CCC: $30.25 © 2005 American Chemical Society Published on Web 08/10/2005

8712

Langmuir, Vol. 21, No. 19, 2005

chains provides even more possibilities for tuning the surface properties by external stimuli.17-19 There are numerous reports of the successful synthesis of mixed polymer brushes via “grafting to”18 and “grafting from”19 approaches. Theory predicts that mixed polymer brushes undergo various phase transitions upon changes in solvent selectivity.20 The surface morphology can be switched from dimples (clusters of the unfavorite polymer embedded into the matrix of the favorite polymer) to lamellas (alternating stripes of two different polymers) and vice versa upon treatment with selective or nonselective solvents, respectively. Stimuli-provoked phase transitions result in a change in the composition of the topmost layer of the brush. Therefore, exposing the brush composed of hydrophobic and hydrophilic polymers to the solvent selective for one of the components (favorite polymer) leads to switching between the hydrophilic and hydrophobic states. The favorite polymer is swollen and occupies the top layer of the brush, and another polymer is collapsed and occupies the brush bottom. Various intermediate states can be obtained by the exposure of the mixed brush to solvents of different quality.19 Thus, the mixed brushes represent an example of a complex film where morphology and surface properties are very sensitive to the environment. The analysis of the composition of such films is a rather complicated task. Different methods including X-ray photoelectron spectroscopy (XPS), ATR-FTIR spectroscopy, VIS ellipsometry, X-ray scattering, and so forth were recently applied. However, the applicability of these methods is severely limited. The resolution of XPS is limited mostly by the depth sensitivity. The maximal probing depth is about 10 nm. ATR-FTIR involves multiple reflections of the probing IR beam; therefore, the method is unable to provide a very local composition. The combination of visible single-wavelength ellipsometry and X-ray scatterings allows for the determination of changes in layer thickness via step-by-step measurements following each step of the surface modification, but the specific determination of the composition of an “unknown” mixture is impossible. Spectroscopic ellipsometry methods, which give access to the identification of molecules and quantitative analysis including the determination of thickness, density, and molecular orientation, would be favorable to solve the problem. In this article, we suggest an approach that allows for the direct determination of the chemical composition and thickness of very thin polymer mixed layers by a combination of IRSE and VISE. From IR spectroscopic ellipsometry (IRSE) measurements, molecular bands21,22 as well as structural properties of thin films and layered systems can be reliably interpreted on the basis of an appropriate optical model. The high analytical potential of IRSE is based on (i) noncontact and noninvasive measurements, (ii) high sensitivity for the investigation of thin films,23,24 (iii) identification of chemical bonds by (17) Zhao, B.; Brittain, W. J.; Zhou, W.; Chemg, S. Z. D. J. Am. Chem. Soc. 2000, 122, 2407-2408. (18) Minko, S.; Patil, S.; Datsyuk, V.; Simon, F.; Eichhorn, K. J.; Motornov, M.; Usov, D.; Tokarev, I.; Stamm, M. Langmuir 2002, 18, 289-296. (19) Sidorenko, A.; Minko, S.; Schenk-Meuser, K.; Duschner, H.; Stamm, M. Langmuir 1999, 15, 8349-8355. (20) Minko, S.; Muller, M.; Usov, D.; Scholl, A.; Froeck, C.; Stamm, M. Phys. Rev. Lett. 2002, 88, 35502. (21) Ro¨seler, A.; Korte, E. H. In Infrared Spectroscopic Ellipsometry: Handbook of Vibrational Spectroscopy; Griffiths, P. R., Chalmers, J., Eds.; Wiley: Chichester, U.K., 2001; Vol. 2. (22) Tompkins, H. G.; Tiwald, T.; Bungay, C.; Hooper, A. E. J. Phys. Chem. B 2004, 108, 3777. (23) Angermann, H.; Henrion, W.; Rebien, M.; Ro¨seler, A. Surf. Sci. 1997, 388, 15.

Ionov et al.

vibrational absorption bands,24-26 (iv) high probing depth up to the micrometer range for many organic materials, and (v) optical modeling with respect to molecular orientations and structure.25,26 The ellipsometric parameters measured with high sensitivity in single-reflection geometry contain information on the phase as well as on the amplitude of the reflected light in the mid-infrared spectral range. These quantities are derived within one and the same experiment, and thereby optical simulations and quantitative interpretation are improved for many applications. Here for the ellipsometric experiments we synthesized mixed-polymer brushes with gradually changing composition27,28 prepared via the “grafting to” approach in the melt.18 Experimental Section Highly polished single-crystal silicon wafers of {100} orientation (Semiconductor Processing Co.) with a native SiO2 layer were first cleaned in an ultrasonic bath for 30 min, placed in hot piranha solution (3:1 concentrated sulfuric acid/30% hydrogen peroxide) for 1 h, and then rinsed several times with high-purity water. Carboxyl-terminated polystyrene (PS-COOH, Mn ) 45 900 g/mol, Mw ) 48 400 g/mol), carboxyl terminated poly(2vinylpyridine) (P2VP-COOH, Mn ) 39 200 g/mol, Mw ) 41 500 g/mol), and carboxyl-terminated poly(tert-butyl acrylate) (PBACOOH, Mn ) 42 000 g/mol, Mw ) 47 000 g/mol) were purchased from Polymer Source, Inc. (synthesized by anionic polymerization). Polyglycidyl methacrylate (PGMA) with (Mn ) 84 000 g/mol, PDI ) 3.4) was synthesized by free radical polymerization of glycidyl methacrylate (Aldrich).29 The polymerization was carried out in methyl ethyl ketone (MEK, VWR) at 60 °C. AIBN (Aldrich) was used as the initiator. The PGMA sample was purified by multiple precipitations from MEK solution in diethyl ether. Gradient PS-mix-P2VP and PS-mix-PBA brushes were prepared according to the following procedure. First, silicon substrates were covered by 50-nm-thick gold layers (vapor deposition). Afterward, a thin layer of PGMA (1.5 ( 0.1 nm) was deposited by spin coating from 0.01% solution in MEK on the gold-covered Si wafer and held at 110 °C for 10 min. PGMA forms a homogeneous and macroscopically uniform film.15 Shorttime heating leads to the cross linking of PGMA and the formation of a network. This PGMA monolayer served as a macromolecular anchoring layer.29 Afterward, a film of PS-COOH was spin coated from 2% solution in toluene and annealed for 1 h on the specially designed stage with a 1D gradient of temperature so that the temperature of the stage changed gradually from 90 °C on the left-hand side of the stage to 130 °C on the right-hand side (Figure 1) as described elsewhere.15 The distance between these two edges was 50 mm. The ungrafted polymer was removed using Soxhlet extraction in toluene for 3 h. In the second step, a film (500 nm thick) of P2VP-COOH or PBA-COOH was spin coated onto the top of the gradient PS brush. The film was annealed at 150 °C for 8 h to produce a 1D gradient of the composition.28 Afterward, the ungrafted polymers, PBA-COOH and P2VP-COOH, were removed by Soxhlet extraction in THF for 8 h. Finally, we prepared the samples of the 1D gradient mixed brushes PS-mix-P2VP and PS-mix-PBA of size XY ) 50 mm × 10 mm. Before measurements, the samples were rinsed in chloroform, which is a nonselective solvent. In the dried sample, both polymers are present on the top of the brush that indicates the dominated lateral segregation.5 (24) Hinrichs, K.; Gensch, M.; Ro¨seler, A.; Esser, N. J. Phys.: Condens. Matter 2004, 16, S4335 (25) Hinrichs, K.; Tsankov, D.; Korte, E. H.; Ro¨seler, A.; Sahre, K.; Eichhorn, K.-J. Appl. Spectrosc. 2002, 56, 737. (26) Tsankov, D.; Hinrichs, K.; Korte, E. H.; Dietel, R.; Ro¨seler, A. Langmuir 2002, 18, 6559. (27) Ionov, L.; Houbenov, N.; Sidorenko, A.; Luzinov, I.; Minko, S.; Stamm, M. Langmuir 2004, 20, 9916-9919. (28) Ionov, L.; Sidorenko, A.; Stamm, M.; Minko, S.; Zdyrko, B.; Klep, V.; Luzinov, I. Macromolecules 2004, 37, 7421-7423. (29) Iyer, K. S.; Zdyrko, B.; Malz, H.; Pionteck, J.; Luzinov, I. Macromolecules 2003, 36, 6519-6526.

Stimuli-Responsive Mixed Grafted Polymer Films

Langmuir, Vol. 21, No. 19, 2005 8713

Figure 1. (a) Principal scheme of the temperature gradient stage. (b) Example of the temperature gradient created on the surface of the stage (obtained with built-in thermocouples). AFM (Beetle UHV SPM 300, Carl Zeiss Jena) measurements were made at different spots on the brush samples. The rootmean-square roughness (rms) was estimated to be 0.4 nm. The high-precision single-wavelength ellipsometric measurements were carried out at a 70° angle of incidence and a wavelength of 632.8 nm (He-Ne laser) with a SENTECH SE-402 microfocus ellipsometer. (The lateral resolution is defined by a beam spot of about 20 µm.) For IRSE in the mid-infrared spectral range, a photometric ellipsometer30,31 has been employed, which is externally attached to a Bruker IFS 55 FTIR spectrometer. At an incidence angle of 70°, the probed area is about 5 × 10 mm2. In general, in ellipsometry the elliptically polarized radiation reflected from the sample is analyzed.21-26 The reflected radiation is described by the complex ratio of the two mutually orthogonal polarized components of the reflected wave, rs and rp: F ) rp/rs ) tan Ψei∆. Here, tan Ψ ) |rp/rs| is the absolute amplitude ratio, and ∆ is the phase-shift difference between the p- and s-polarized components of reflected waves. The ellipsometric measurements are used to obtain tan Ψ and ∆ values for the beam reflected from the investigated thin film. The tan Ψ and ∆ values are functions of the optical thickness of the films. A complex film can be represented as an effective multilayered optical medium. Measured ellipsometric parameters tan Ψ and ∆ are used for simulations of the multilayered film in order to estimate the thickness or (and) refractive index of each layer. VISE Experiments. Measurements were performed after each step of the modification of the Si wafer to use the measurements of the previous step as a reference for the simulation of ellipsometric data. The thickness of the polymer layer was determined from ellipsometric parameters ∆ and tan Ψ obtained for each grafting step. The thickness of the PGMA layer was evaluated by applying the layer model of PGMA on Au using the value n ) 1.525 for the high-frequency refractive index of PGMA and n ) 0.166 - i3.15 for gold. The thickness of the grafted PS was evaluated with the three-layer model Au/PGMA/ PS with the effective value of the high-frequency refractive index n ) 1.59 for PS. The mapping of the sample surface (2D ellipsometric mapping) with a spatial resolution of 1 mm in the X and Y directions was used to construct the thickness profiles. Finally, the thickness of the mixed-polymer brushes was calculated using the three-layer model Au/PGMA/mixed brush, considering the polymer brushes to be an effective optical medium with n ) 1.59.18 Thus, for each measured point we obtained the thickness of PGMA, hPGMA (the first mapping). Then, upon grafting PS we obtained the thickness of the PGMA plus grafted PS film, hPGMA+PS (the second mapping), and finally, upon grafting (30) Ro¨seler, A. Infrared Spectroscopic Ellipsometry; Akademie Verlag: Berlin, 1990. (31) Ro¨seler, A. In Handbook of Ellipsometry; Tompkins, A. G., Irene, E. A., Eds.; Noyes Press: Park Ridge, NJ, in press; Vol. 11.

Figure 2. (a) Scheme of the model thin film used as an optical model for the simulation of IRES data. The spectra for each measured position were simulated with the two-layer model as indicated. Positions (in mm) for the center of the foci measured from the left edge of the sample are given at the top. The foci are elliptical in shape, and the width in the X direction is about 5 mm. (b) Model for the structure of the mixed-polymer brush in a nonselective solvent. PBA we have obtained the total thickness of the mixed brush (PGMA + PS + PBA), hmix (the third mapping). The composition of the brush was calculated as follows

hPGMA+PS - hPGMA FPS φPS ) hPGMA+PS - hPGMA hmix - hPGMA+PS + FPS FPBA

(1)

where φPS is the fraction of PS in the mixed brush and FPS and FPBA are the densities of PS and PBA, respectively. IRSE Experiments. Figure 2 presents a schematic of the model used for the simulation of IRSE data. The calculations are based on the Wolter algorithm for multilayers32 implementing the isotropic Airy model27 for the single layer. A laterally dependent double-layer model (Figure 2a) was used to simulate the laterally dependent structure of the mixed brush (Figure 2b). In this simple approach, we assumed that within the beam spot the optical response of the mixed brush could be modeled by a laterally homogeneous composition of the brush and that the film can be described with a double-layer model. Therefore, each layer consists of only one of the two polymers, and the corresponding thickness of each layer is determined by the fraction of the single component in the double layer. Because the film is brushlike as shown schematically in Figure 2b, only the total thickness of the film has physical meaning. The thicknesses of the single layers are employed for the determination of the fraction of the single compounds. It should be noted that only IR ellipsometric spectra of the complete brush film have been evaluated within this model and no spectra of the single brush during the stepwise production (as in the case of singlewavelength ellipsometry) were needed. Infrared spectra taken at several sample positions, as indicated in the Figure, were simulated using this model. The positions of the foci are given in terms of a distance to the left edge of the sample in millimeters. (32) Hinrichs, K.; Gensch, M.; Ro¨seler, A.; Korte, E. H.; Sahre, K.; Eichhorn, K.-J.; Esser, N.; Schade, U. Appl. Spectrosc. 2003, 57, 1250.

8714

Langmuir, Vol. 21, No. 19, 2005

Ionov et al.

Figure 3. FTIR transmission spectra of films prepared by casting on KBr (black lines) and tan Ψ (gray lines) spectra of thin films on gold at an incidence angle of 75°. As input parameters we used the optical constants (refractive index n and absorption index k) as determined from thick films of the single compounds PS, PVP, PBA, and the gold-coated Si wafer. The optical constants of the polymers have been calculated by a best-fit simulation as described in detail elsewhere.25 In this simulation, the dielectric function  is presented as a sum of harmonic oscillators and the square of the high-frequency refractive index, where the vibrational bands are represented as Lorentzian oscillators at the wavenumbers (ν˜ i0)

′ ) ∞ +

∑(ν˜ i

′′ )

∑(ν˜ i

Fi(ν˜ i02 2 i0

2

- ν˜ )

- ν˜ 2)2 + (Γiν˜ )2 FiΓiν˜

2 i0

- ν˜ 2)2 + (Γiν˜ )2

(2)

(3)

For an isotropic sample, the dielectric function in the j ) x, y, z directions are related by x ) y ) z. The complex refractive indices can be derived from the sum of the real and imaginary parts:

nˆ j ) nj + ikj ) x(j′ + ij′′) n∞j ) x∞j

(4)

The high-frequency refractive indices (n∞) were taken from measurements of single-wavelength ellipsometry on thick films. The spectral minima of the bands in the tan Ψ spectra were used for the simulation as starting values for the resonance frequencies (ν˜ i0) of vibrational bands. This procedure is not valid for the general case. The similarity of transmission spectra and tan Ψ spectra is given only for thin organic films on metallic substrates. This comparability of tan Ψ spectra of thin organic films on metallic substrates and standard transmission spectra of thicker films originates from the so-called “surface selection rule”. For thin organic films (below a few hundred nanometers) on metallic substrates, the band amplitudes in s-polarized spectra are usually below 1% of the amplitude in p-polarized spectra. Figure 3 shows the measured tan Ψ spectra for thin spin-coated films of PS-COOH (d ) 122 nm), P2VP-COOH (d ) 192 nm), and PBA (d ) 127 nm) and the reference IR transmission spectra of the thick films cast from chloroform on KBr substrates. The similarity of ellipsometric and transmission spectra is obvious, and the same vibrational bands are observed in both types of spectra for the presented spectral range.

Figure 4. Ellipsometric tan Ψ spectra at different sample positions (shown in mm) of the PS-mix-P2VP brush: measured spectra (thick lines) and the corresponding simulated spectra (thin lines). The top spectrum is the reference spectrum of the PGMA-coated substrate. The probed area was about 5 × 10 mm2. For better presentation, the spectra have been shifted with respect to the top reference spectrum. Changes in band intensities are marked by arrows (upward P2VP, downward PS). The arrows on the plot show the direction of increasing intensity of the bands.

Results and Discussions A direct determination of the grafted amount of each component in the mixed brush at different distances from the left edge of the sample was performed by IR ellipsometry. The measured tan Ψ spectra (thick line) for different positions of the 1D gradient PS-mix-P2VP sample are shown in Figure 4. The reference spectrum of the PGMAcovered gold substrate is shown on the top of Figure 4. In the mixed-brush IRES spectra, we identified the characteristic bands of PS (1452, 1493, 1601 cm-1) and P2VP (1434, 1473, 1568, 1590 cm-1). Clearly, the decrease of the P2VP-related band amplitudes and the increase of the PS-related band amplitudes are observed with increasing distance from the left edge of the sample. The simulated IRES spectra are shown in Figure 4 with the thin line. The thickness of each single-component layer in the two-layer model was used as a fitting parameter. The thicknesses were adjusted until the best fit to distinct bands in the experimental spectrum was achieved. For PS and PVP, the bands at 1493 and 1473 cm-1, respectively, were used in the fitting procedure. In general, good agreement between measured and simulated spectra was obtained for most measurements. Small differences are observed in the band range of 1400-1500 cm-1 and at 1590 cm-1. In the 39 and 44 mm measurement-point positions, the simulations slightly overestimate the band amplitudes. We should take into account that the input parameters for the optical constants of the single-polymer compounds were derived from the measurements of the thick reference films, which might be structurally different from the grafted polymers in the mixed brushes. Very similar results were obtained for the PS-mix-PBA brush. Figure 5 shows the measured and calculated tan Ψ spectra for the PS-mix-PBA film. PBA bands are present at 1278, 1368, 1392, 1479, and 1728 cm-1. Around 1452 cm-1, the bands of PS and PBA are overlapping. In the

Stimuli-Responsive Mixed Grafted Polymer Films

Langmuir, Vol. 21, No. 19, 2005 8715

Figure 5. Ellipsometric tan Ψ spectra for different positions of the PS-mix-PBA brush: measured spectra (thick lines) and corresponding simulated spectra (thin lines). The top spectrum is the reference spectrum of the PGMA-coated substrate. The probed area was about 5 × 10 mm2. For better presentation, the spectra have been shifted with respect to the top reference spectrum. The arrows on the plot show the direction of increasing intensity of the bands.

Figure 6. Ellipsometric mapping of the thickness (a) and composition (b) of the gradient of the PS-mix-P2VP brush. (0, O, 4) Data from single-wavelength ellipsometry; (9, b, 2) data from IR ellipsometry; (9, 0) thickness of the PS grafted layer; (b, O) thickness of the P2VP grafted layer; (2, 4) fraction of PS, %.

simulations, the thicknesses of the PS and PBA layers were adjusted by the best fit to the amplitudes of 1490 cm-1 (PS) and 1278 cm-1(PBA) bands. Overall good agreement between measured and simulated spectra was found. For the carbonyl group band at 1728 cm-1, some deviation between simulation and experiment is observed. The results of VISE and IRSE mapping of the 1D gradient of the PS-mix-P2VP and PS-mix-PBA brushes are compared in Figures 6a and 7a, respectively. The entire thickness along the X axis has a nearly constant value of about 7-8 nm (dry film) corresponding to a grafting

Figure 7. Ellipsometric mapping of the thickness (a) and composition (b) of the gradient PS-mix-PBA brush. (0, O, 4) Data from single-wavelength ellipsometry; (9, b, 2) data from IR ellipsometry; (9, 0) thickness of the PS grafted layer; (b, O) thickness of the PBA grafted layer; (2, 4) fraction of PS, %.

density of approximately 0.1 chains/nm2. The distance between grafting points of 3.9 nm is smaller than the radius of gyration (Rg) of PS and P2VP polymer coils at Θ conditions (RgΘ ≈ 5 nm29). Consequently, the PS-mixP2VP and PS-mix-PBA polymer grafted films can be considered to be brushlike layers. The results of VISE were used for the indirect estimation of the mixed-brush composition (Figures 6b and 7b) as described in the Experimental Section (eq 1). For the presented results, the differences between the IR and VIS data may be explained by several reasons. An independent determination of the complex refractive index

8716

Langmuir, Vol. 21, No. 19, 2005

and thickness of a thin organic film in the lower nanometer range is an unresolved problem in thin film metrology. Because the optical data of the thin film in general are not known, optical data of reference samples, such as thicker films or a bulk material, have to be used for the evaluations. In the presented work in the evaluation of infrared ellipsometric spectra, the thicknesses were determined by adjusting the simulated vibrational bands to the heights of specific measured vibrational bands that were not overlapping with other bands. For the optical simulations, the vibrational parameters derived from a thick film of the same polymer were used as input data. For the evaluation of thicknesses from VIS ellipsometry, the corresponding high-frequency refractive indices of the single polymers in thick films were used. However, because of differences in the molecular structure the parameters of the vibrational bands and of the high-frequency refractive indices must not be exactly the same in a thin film and a thick film. As can be seen from a comparison of simulated and measured infrared ellipsometric spectra for the PS-mix-P2VP brush (Figure 4) and the PS-mixPBA brush (Figure 5), the overall agreement is better for the PS-mix-PBA brush. Therefore, the observed differences in the composition for the PS-mix-P2VP brush (Figure 6) could be caused by the fact that the optical characteristics of the polymers in segregated microphases are different from that of the reference (thick film). Obviously, such differences can lead to larger deviations for the results from simulations of the IR ellipsometric data as compared to those from VIS ellipsometry. (VIS are less sensitive to the molecular structure.) The differences between VIS and IR results in Figure 6b become larger because the presented composition is calculated from the ratios of the volume fractions of the single polymers, which in turn were derived from the ratios of the corresponding thicknesses (Figure 6a). Another reason for the differences between the results from VIS and IR ellipsometry is that different evaluation schemes were employed for the VIS and IR data. Because VIS ellipsometry has a low spectral contrast for the investigated polymers, the thicknesses could be determined in a stepwise measurement only. This procedure consists of three steps: the first measurement is of the PGMA-covered substrate, the second measurement is of the grafted PS, and the third measurement is of the mixed

Ionov et al.

brush. In contrast, the thicknesses from IR ellipsometry were determined on the basis of two measurements: the first measurement was obtained for the substrate, and the second measurement was obtained for the mixed brush. Finally, the IRSE probes a much larger area of about 50 mm2 as compared to 0.0004 mm2 for VISE. Presently, it is unclear how the gradient character within the focus influences the measured IRSE spectra. We plan to perform investigations using a synchrotron source (at the storage ring of BESSY II in Berlin) of IR radiation allowing for a higher lateral resolution32 to elucidate the influence of the lateral gradient. (A resolution of approximately 100 µm × 100 µm with an equivalent signal-to-noise ratio should be possible.31) Conclusions This study presents the high potential of the thin polymer film analysis using single-visible-wavelength ellipsometry and IR spectroscopic ellipsometry. Both methods show very good agreement for the PS-mix-PBA 1D gradient brush, whereas some discrepancies were observed for the PS-mix-P2VP film. It is obvious that infrared optical properties of polymer thin films, which are correlated to the chemical structure of the thin film, are very well suited to the structural analysis. VISE application for the structural analysis is limited by the similarity of the refractive indexes of different polymers. Present limitations of IRSE are the signal-to-noise ratio and lateral resolution, which potentially could be improved using a synchrotron ellipsometer. Acknowledgment. We thank Dr. A. Ro¨seler for his valuable support and Dr. M. Gensch for critical discussions. Financial support from the Senatsverwaltung fu¨r Wissenschaft, Forschung und Kultur des Landes Berlin, and from the Bundesministerium fu¨r Forschung und Wissenschaft as well as the European Union through the EFRE program (ProFIT grant, contract nr. 10125494) is gratefully acknowledged. The authors from IPF are grateful to DFG and BMBF for financial support. S.M. acknowledges support from the NYSTAR Center for Advanced Materials Processing (Clarkson University). LA050620J