Microstructures by Solvent Drop Evaporation on Polymer Surfaces

Langmuir 2006, 22, 11395-11399

11395

Microstructures by Solvent Drop Evaporation on Polymer Surfaces: Dependence on Molar Mass Guangfen Li,† Hans-Ju¨rgen Butt,† and Karlheinz Graf*,†,‡ Max Planck Institute for Polymer Research, Ackermannweg 10, 55128 Mainz, and Centre for Micro- and Nanochemistry and -engineering (Cµ), UniVersity of Siegen, 57068 Siegen, Germany ReceiVed July 4, 2006. In Final Form: August 25, 2006 When a solvent drop evaporates from a polymer surface, it leaves behind a characteristic structure, typically a crater. We deposited toluene drops with a microsyringe onto planar polystyrene (PS) surfaces and analyzed the surface topography after drying. For low molar mass PS (Mw ) 20.9-24.3 kDa) dotlike protrusions with a ridge at the periphery formed on the polymer surface. With increasing molar mass the central region decreased in height. At Mw ) 29.6-643 kDa a craterlike structure with a depression in the center and a ridge was observed. At even higher molar mass, irregular structures without rotational symmetry occurred. We explain the observed dependence on the molar mass with a different degree of entanglement, leading to different dissolution rates and different diffusion constants.

The evaporation of drops of pure liquids from planar, inert solid surfaces has been studied in detail and is well understood.1-12 More complex is the situation for the evaporation of liquid mixtures,13,14 for solutions with nonvolatile solutes,14-21 and for colloidal dispersions.22-28 By far less understood is the evaporation of liquids from soluble surfaces. In particular, the evaporation of solvent drops from polymer surfaces is a complex process because the polymer is dissolved in the solvent and at the same

time solvent molecules diffuse into the polymer.29,30 In addition, the viscosity of the mixture and the diffusion constant of liquid molecules depend on the local concentration. Interest in a better understanding of the evaporation of solvent drops from polymers recently increased because they are used to microstructure polymer surfaces for a variety of applications.14,20,29,31,32 Here we show that the topographic structure left by a drying solvent drop in a polymer surface strongly depends on the molar mass of the polymer. As an example we studied toluene drops drying on polystyrene surfaces.

* To whom correspondence should be addressed. E-mail: [email protected]. Phone: +49-6131-379 115. Fax: +49-6131 379 310. † Max Planck Institute for Polymer Research. ‡ University of Siegen.

Materials and Methods

Introduction

(1) Morse, H. W. Proc. Am. Acad. Arts Sci. 1910, 45, 363. (2) Langmuir, I. Phys. ReV. 1918, 12, 368. (3) Picknett, R. G.; Bexon, R. J. Colloid Interface Sci. 1977, 61, 336. (4) O’Brien, R. N.; Saville, P. Langmuir 1987, 3, 41. (5) Rowan, S. M.; Newton, M. I.; McHale, G. J. Phys. Chem. 1995, 99, 13268. (6) Bourges-Monnier, C.; Shanahan, M. E. R. Langmuir 1995, 11, 2820. (7) McGaughey, A. J. H.; Ward, C. A. J. Appl. Phys. 2002, 91, 6406. (8) Erbil, H. Y.; McHale, G.; Newton, M. I. Langmuir 2002, 18, 2636. (9) Takata, Y.; Hidaka, S.; Yamashita, A.; Yamamoto, H. Int. J. Heat Fluid Flow 2004, 25, 320. (10) Hu, H.; Larson, R. G. Langmuir 2005, 21, 3972. (11) Poulard, C.; Guna, G.; Cazabat, A. M.; Boudaoud, A.; Ben Amar, M. Langmuir 2005, 21, 8226. (12) Bonaccurso, E.; Butt, H.-J. J. Phys. Chem. B 2005, 109, 253. (13) Sefiane, K.; Tadrist, L.; Douglas, M. Int. J. Heat Mass Transfer 2003, 46, 4527. (14) Tekin, E.; de Gans, B. J.; Schubert, U. S. J. Mater. Chem. 2004, 14, 2627. (15) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Nature 1997. (16) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S. R.; Witten, T. A. Phys. ReV. E 2000, 62, 756. (17) Blossey, R.; Bosio, A. Langmuir 2002, 18, 2952. (18) Pauchard, L.; Allain, C. Phys. ReV. E 2003, 68, 052801. (19) Gorand, Y.; Pauchard, L.; Calligari, G.; Hulin, J. P.; Allain, C. Langmuir 2004, 20, 5138. (20) de Gans, B. J.; Schubert, U. S. Langmuir 2004, 20, 7789. (21) Kajiya, T.; Nishitani, E.; Yamaue, T.; Doi, M. Phys. ReV. E 2006, 73, 011601. (22) Adachi, E.; Dimitrov, A. S.; Nagayama, K. Langmuir 1995, 11, 1057. (23) Parisse, F.; Allain, C. J. Phys. II 1996, 6, 1111. (24) Conway, J.; Korns, H.; Fisch, M. R. Langmuir 1997, 13, 426. (25) Deegan, R. D. Phys. ReV. E 2000, 61, 475. (26) Uno, K.; Hayashi, K.; Hayashi, T.; Ito, K.; Kitano, H. Colloid Polym. Sci. 1998, 276, 810. (27) Morozumi, Y.; Ishizuka, H.; Fukai, J. Solute deposit during eVaporation of binary liquid micro-droplet on a substrate, The International Symposium on Micro-Mechanical Engineering, Japanese Society of Mechanical Engineers: Tokyo, Japan, 2003; p 24. (28) Ma, X.; Xia, Y.; Chen, E. Q.; Mi, Y.; Wang, X.; Shi, A. C. Langmuir 2004, 20, 9520.

Polystyrene (PS) was synthesized by anionic polymerization. The molecular masses and polydispersities were Mw ) 20.9, 24.3, 29.6, 62.5, 97.5, 210, 339, 643, 1102, and 1445 kDa and Mw/Mn ) 1.03, 1.05, 1.03, 1.06, 1.04, 1.07, 1.07, 1.10, 1.09, and 1.16, respectively. To produce flat, smooth surfaces, PS powder was compressionmoulded at 160 °C under vacuum into 25 mm wide transparent disks (except for PS of Mw ) 20.9 kDa, which formed an opaque sample). The PS samples were cleaned with methanol (HPLC grade) in an ultrasonic bath for 15 min prior to the deposition of toluene (HPLC grade) drops. Toluene drops were placed onto the polymer surfaces by a syringe (Figure 1). The flow is controlled by a homemade syringe pump, which consisted of a glass syringe (Hamilton, Germany), a Teflon tube, and a fused silica capillary (Upchurch Scientific, Oak Harbor, WA) with outer and inner diameters of 360 and 20 µm, respectively. The polymer substrate was moved up and down with a computercontrolled and stepper-motor-driven translator with a velocity of 11 mm/s toward the pendant drop at the capillary. Immediately after contact between the drop and the polymer substrate it was retracted again with the same velocity to its initial position, which was at least 1 mm below the capillary. The evaporating drop on the substrate was recorded with a camera system including a 5× objective (Mitutoyo, Japan), a zoom tube (Navitar, Inc., New York), and a digital camera (Basleer Vision Technology, Germany). For better contrast the substrate was illuminated from the back with a cold light source (KL 2500 LCD, Schott, Mainz, Germany) through a diffuser. The approach, retraction, and recording were controlled (29) Kawase, T.; Siringhaus, H.; Friend, R. H.; Shimoda, T. AdV. Mater. 2001, 13, 1601. (30) Bonaccurso, E.; Butt, H.-J.; Hankeln, B.; Niesenhaus, B.; Graf, K. J. Appl. Phys. 2005, 87, 124101. (31) Bonaccurso, E.; Butt, H.-J.; Graf, K. Eur. Polym. J. 2004, 40, 975. (32) Karabasheva, S.; Baluschev, S. Appl. Phys. Lett. 2006, 89, 031110.

10.1021/la061929l CCC: $33.50 © 2006 American Chemical Society Published on Web 11/01/2006

11396 Langmuir, Vol. 22, No. 26, 2006

Li et al.

Figure 1. Schematic of the experimental setup for defined droplet deposition onto a polymer surface by a syringe. with a self-written macro in the program DTAcquireCL (Data Translation, Germany). After drop evaporation the topography of the polymer surface was imaged with a three-dimensional optical measurement system based on confocal white-light microscopy of the disk-scanning type (µsurf, NanoFocus AG, Oberhausen, Germany). Using a 20× or 50× objective, the resolution was 1.5 and 0.6 µm laterally and 5 and 2 nm in height, respectively.

Results and Discussion The surface structures left by evaporating sessile drops on a polymer surface showed a steep ridge at the periphery after drying (Figure 2). Typically, this ridge was 3 µm high. On low molar mass PS (Mw e 24.3 kDa) an evaporated toluene drop left a protrusion or dotlike structure in the center, which filled the inner part of the ridge and rose to roughly 3-4 µm height. Between the ridge and the dot a depression was observed. Similar dotlike deposits were observed after the evaporation of PS solutions from solvent mixtures, which contained one low- and one highboiling-point solvent.20 In our experiments only one solvent was present. With increasing molar mass, the center of the dot decreased in height and a central depression or crater was formed. Such craterlike structures have been described before.29,30 For Mw ) 97.5 kDa the bottom of the crater was even below the original surface and reached a depth of up to 3 µm. When the molar mass was further increased, the depth of the craters decreased until at 1445 kDa they disappeared completely. In addition, a new characteristic feature was visible: the craters became asymmetric. With increasing molar mass the degree of symmetry and the length scale, on which the asymmetric feature appeared, decreased. At Mw ) 339 kDa the outer ridge typically showed a height modulation; it was significantly higher on one side than on the opposite one. The length scale of the asymmetry is on the order of 400 µm. At Mw ) 1445 kDa the crater is filled with an irregular pattern of small depressions and protrusions with a width of typically 50 µm. Before dealing with the question of how the structures form, we describe basic features of the evaporation process itself. When a drop of toluene was deposited onto PS, the liquid spread. After 0.3-0.5 s it reached its maximal lateral extension of ∼180 µm. After that the toluene evaporated, keeping the contact radius constant while the contact angle decreased (Figure 3). For comparison, we placed toluene drops onto a nonsoluble, hydrophobic silicon surface silanated with trichlorooctadecylsilane. In this case the toluene drops evaporated with a constant contact angle of θ ) 32° and diminishing contact area. The evaporation rate, which is the decrease in solvent volume per

Figure 2. Typical surface plots (left) and cross-sections (right) of structures left by toluene drops on polystyrene with different molar masses 1-2 h after evaporation. The drops were deposited on the polymer substrate with a syringe, and the resulting microstructures were imaged with a confocal white-light microscope. The surface plots have automatic height rescaling so that the heights are in relative units. Only results for molar masses showing qualitative distinct features are shown.

Figure 3. Contact radius, contact angle, drop volume V, and V2/3 versus time for a toluene drop evaporating from a polystyrene surface with Mw ) 20.9 kDa (red diamonds), 210 kDa (green squares), and 1445 kDa (blue triangles). For comparison the results obtained with a silanized silicon wafer are shown (black).

unit time, |dV/dt|, on the silicon wafer was higher than on the polymer surface. When V2/3 was plotted, with V being the volume

SolVent Drop EVaporation on Polymer Surfaces

Langmuir, Vol. 22, No. 26, 2006 11397

Figure 4. Vapor pressure of toluene drops evaporating from polystyrene surfaces of different molar masses versus time. The relative vapor pressure, that is, the vapor pressure calculated with eq 2 divided by the vapor pressure of pure toluene (P0 ) 3.79 kPa), is plotted.

of the toluene drop, versus the evaporation time t, a straight line was obtained (Figure 3). This linear decrease of V2/3 with time is well-known.1,8,33,34 It was derived by assuming that the diffusion of molecules in the gas phase is rate limiting, that a steady state is reached at each stage of evaporation, that the temperature and the contact angle are constant, and that the vapor pressure of the liquid far from the drop is zero:3

4 3 1/3 P0M f V2/3 ) V02/3 - π Dt t 3 π RT Fβ1/3

()

(1)

Here, V0 is the initial volume of the drop, Dt is the diffusion constant of toluene in air, P0 is the vapor pressure of toluene, M is the molar mass of toluene, R is the gas constant, T is the temperature, and F is the liquid density. β and f are functions of the contact angle θ: β ) (1 - cos θ)2(2 + cos θ) and f ) 0.00008957 + 0.6333θ + 0.116θ2 - 0.08878θ3 + 0.01033θ.4 With θ ) 32° we get β ) 0.0658 and f ) 0.376. From the slope of V2/3 versus t and by inserting P0 ) 3.79 kPa, M ) 92.13 g/mol, and F ) 866 kg/m3 at 25 °C, we find a diffusion coefficient Dt ) 8.1 × 10-6 m2/s. This agrees well with the value of 8.0 × 10-6 m2/s measured previously in tube and sessile drop evaporation experiments.35 The reduction in the evaporation rate on PS as compared to that on silicon can be understood by describing the evaporation of sessile drops with a more general equation, which is valid irrespective of whether the contact angle is constant or not:3

P0′M dV ) -2πDtRc f dt FRT

(2)

Rc is the radius of curvature of the drop, which depends on the contact angle and contact radius. We attribute the reduced rate of evaporation on PS to a decrease in the vapor pressure. The vapor pressure P0′ of the solution depends on the amount of dissolved PS. As more and more PS is dissolved in the toluene, P0′ decreases.36 It goes to zero as the volume fraction of toluene in the toluene/PS mixture tends to zero. To estimate this reduction in the vapor pressure, we applied eq 2. From the measured contact radii a and drop heights h0 the volumes, V ) πh0(3a2 + h02)/6, and the radii of curvature, Rc ) (h02 + a2)/(2h0), at different stages of evaporation were obtained (Figure 4). The vapor pressure decreases from the value of pure toluene to almost zero after 5 (33) Sresnewsky, B. Beibl. Ann. Phys. Chem. 1883, 7, 888. (34) Yu, H. Z.; Soolaman, D. M.; Rowe, A. W.; Banks, J. T. ChemPhysChem 2004, 5, 1035. (35) Erbil, H. Y.; Avci, Y. Langmuir 2002, 18, 5113. (36) Hao, W.; Elbro, H. S.; Alessi, P. Polymer Solution Data Collection. Part 1: Vapor-Liquid Equilibria; Dechema: Frankfurt, Germany, 1992.

Figure 5. Radial flow velocity in an evaporating drop calculated with eq 4 at the beginning and 1 and 2 s after evaporation has started. The parameters used were estimated from experiments of toluene drops evaporating from PS. We used h0 ) 60, 40, and 30 µm, θ ) 33°, 23°, and 17°, λ ) 0.39, 0.43, and 0.45, J0 ) 4.4, 4.0, and 3.8 × 10-3 kg m-2 s-1, and dh0/dt ) 26, 15, and 11 µm/s for t ) 0, 1, and 2 s, respectively. To estimate J0 from the experimental results, we integrated the flux over the whole drop surface: F(dV/dt) ≈ 4πλJ0a2. A value of dV/dt ≈ 10-12 m3/s was taken from the experiments. At the top left a schematic of an evaporating drop illustrates the parameters.

s. Such a low vapor pressure reflects the fact that at the end of the evaporation process little toluene is dissolved in a lot of PS and the volume ratio of PS is almost 1. In fact, some toluene will remain in the PS and not evaporate even after days. One should, however, consider that the distribution of PS in the drop is not homogeneous and thus the vapor pressure varies locally. The evaporation is high at the rim (see below), and thus, the value of P0′ is dominated by the concentration of PS at the periphery. To explain the formation of surface structures, which are left by evaporating solvent drops, we can only attempt a qualitative interpretation, since the whole process is rather complex. The craterlike structures formed at intermediate molar masses can be understood on the basis of a model developed to interpret the formation of ring-shaped coffee stains.14-16,24,29,30,37,38 When the drop dries with a pinned contact line, a radial flow occurs to compensate the difference in volume change and evaporation rate across the drop (Figure 5). In the edge region of the drop, the evaporation rate is high because the available angle of evaporation is increased (for θ < 90°), but the change in volume is restricted by the pinned contact line. In the center, on the other hand, the evaporation rate is lower but the volume changes drastically. The radial flow supplies the solution from the center to the edge region to compensate the difference. As a result there is an enhanced deposition of dissolved material in the edge region. We estimate the velocity of the flow V in our system from the equation of continuity assuming laminar flow:16

V)-

1 Frh

∂h + J(r′,t)(1 + ( ) ) ∫0rr′(F ∂h ∂t ∂r′

2 1/2

) dr′

(3)

Here, h is the height of a drop at radius r and J is the evaporation flux from the surface of the drop (kg m-2 s-1). The height was (37) Hu, H.; Larson, R. G. J. Phys. Chem. 2002, 106, 1334. (38) Rieger, B.; van den Doel, L. R.; van Vliet, L. J. Phys. ReV. E 2003, 68, 036312.

11398 Langmuir, Vol. 22, No. 26, 2006

calculated with the assumption of a hemispherical drop: h(r,t) ) ((h02 + a2)2/(2h0)2 - r2)1/2 - (a2 - h02)/2h0. The evaporation flux was calculated according to ref 16, J ≈ J0[1 - (r/a)2]-λ with λ ) (90° - θ)/(180° - θ), which is a good approximation for the enhanced evaporation at the rim. For small contact angles and under constant contact radius conditions, we can integrate eq 3:

V)-

[

1 ∂h0 r2 2 1 (h0 - a2) {(h02 + a2)2 2 ∂t 2 2rhh0 12h02

4r2h02}3/2 +

]

(h02 + a2)3 12h02

+

2λa2J0 [1 - {1 - (r/a)2}λ] (4) Frh

The flow velocities were calculated in Figure 5 with parameters for a toluene drop with an initial volume of 2 nL and an initial contact angle of 33° at room temperature. The flow velocities are sufficiently high to transport a significant amount of material to the rim of the drop. This leads to the formation of the craterlike structures observed. A second factor, which might contribute to the formation of the peripheral ridge, is the direct mechanical tension of the liquid surface. The liquid surface tension at the three-phase contact line can mechanically deform the polymer surface, which is softened by toluene (see, e.g., ref 39). This we found in experiments in which we placed a water droplet on a PS substrate, softened by toluene vapor. However, usually the rim is rather narrow in contrast to the profiles presented here, and its rim height is below 1 µm. We additionally performed simulations on the structuring process, which are based on the evaporation-driven process, accounting for a flow of polymer toward the rim40 (copublished in ref 41). A good agreement was found between the simulated height profile and the experimental one. In contrast, simulations accounting for a mechanical effect suggest that the resulting rim heights are far below 1 µm. A further hint is our finding that a solvent droplet close to a PS surface, without any contact, leads to a concave surface topology as well. Since a meniscus is missing and thus an acting capillary force can be ruled out, a flow of softened polymer must occur, similar to that in a droplet. A better understanding of which process is predominant for the surface structuring can only be obtained on the basis of further experiments. To understand evaporation structures at low and high molecular weights, it is instructive to first analyze the time and length scales of the processes involved. As soon as the drop has spread on the surface, polymer starts to dissolve into the toluene. Polymer dissolution is slowed by entanglement. The entanglement molar mass of pure PS is Me ) 18.1-19.1 kDa.42,43 With increasing molar mass the rate of dissolution will decrease. This explains why for high molar mass only shallow structural changes are observed. After dissolution the polymer diffuses in the liquid drop. Within ∆t ) 1 s, a characteristic time scale for the evaporation, a typical diffusion length is (6D∆t)1/2, where D is the diffusion coefficient of PS in toluene. It can be estimated from the phenomenological (39) White, L. R. J. Colloid Interface Sci. 2003, 258, 82. (40) Stupperich-Sequeira, C.; Graf, K.; Wiechert, W. Math. Comput. Modell. Dyn. Syst. 2006, 12, 263. (41) Stupperich-Sequeira, C.; Graf, K.; Wiechert, W. Modelling of the Production of Microwells. In ARGESIM Reports; Argesim-Verlag: Technical University Vienna; Vol. 24. (42) Fetters, L. J.; Lohse, D. J.; Milner, S. T.; Graessley, W. W. Macromolecules 1999, 32, 6847. (43) Brandrup, J.; Immergut, E. H.; Grulke, E. A. Polymer Handbook, 4th ed.; John Wiley & Sons: New York, 1999.

Li et al.

scaling law D ) D0 exp(-Rcu),44 with D0 ) (2.8 × 10-8)Mw-0.55 m2/s45 (Mw in grams per mole) for PS in toluene, u ) 0.65, and R ) 15.46 The diffusion coefficient in the dilute regime is 1.2 × 10-10 m2/s for Mw ) 20.9 kDa and 1.1 × 10-11 m2/s for Mw ) 1445 kDa. Within 1 s a 20.9 kDa PS chain typically diffuses 27 µm. This is of the same order of magnitude as the height of the drops (typically 60 µm at the beginning) but is significantly smaller than the radial extension of ∼200 µm. Thus, concentration gradients in the vertical direction are at least partially equilibrated. For a 1445 kDa PS molecule the diffusion length is 8 µm. In this case, only a little PS will reach the top of the drop by diffusion and strong concentration gradients can persist even in the vertical direction. At the same time toluene will diffuse into the polymer, which leads to a swelling of the polymer. The diffusion of toluene in PS depends critically on the amount of toluene present. With increasing volume fraction of toluene and thus increasing degree of swelling of PS, the toluene molecules become more and more mobile. At a polymer weight fraction of 0.7 the diffusion coefficient of toluene as determined by NMR is 10-10 m2/s.47 For higher weight fractions, diffusion is much slower. Extrapolating experimental NMR results from higher temperature to room temperature, a value of D ) 1.6 × 10-13 m2/s is estimated at a weight fraction of 0.89. Toluene molecules diffuse only a typical distance of ∼1 µm within 1 s. This corresponds to the length scale of the depth of the craters. It remains to explain the dotlike structures observed at low molar mass. PS with a low molar mass dissolves and diffuses faster than high molar mass PS. More polymer is dissolved in the toluene, and it quickly diffuses to the top of the drop. As a result the critical concentration, where polymer chains in solution overlap, is quickly reached. This critical concentration can be estimated from C* ) 3Mw/(4πNARg3), where the radius of gyration is given by Rg2 ) (1.38 × 10-4)Mw1.19 nm2, with Mw in grams per mole.46 For 20.9 kDa PS we have C* ≈ 0.1 g/mL corresponding to a volume fraction of roughly 10%. From this concentration on, the overlapping PS chains start to become immobile and form a network. Further dissolution and diffusion of PS is hindered. At the same time the toluene molecules can still diffuse freely in the PS network; their diffusion coefficient is only reduced by 20% with respect to the value of pure toluene.47 This will diminish or even stop the radial flow of polystyrene. The shape of the PS/toluene mixture is dominated by the surface tension, which according to the Laplace equation, tends to form a hemispherical structure. In summary, at high molar masses of polymer, only low amounts of polymer will be dissolved and polymer will diffuse on a small length scale. Thus, typical nonequilibrium features are observed. At low molar mass dissolution and diffusion of polymer chains is fast. Thus, the critical concentration is reached, leading to an immobile polymer gel. A closer look at the surface topologies reveals different classes of symmetry. For low and intermediate molar masses between 24.3 and 210 kDa, the resulting topology shows nearly axial symmetry. The topology changes from centrally bulged (24.3 kDa) over dimpled (29.6 kDa) to concave (210 kDa). This observation is described in the literature for an evaporating sessile droplet with dissolved polymer on an insoluble substrate, when (44) Phillies, G. D. J. Macromolecules 1986, 19, 2367. (45) Min, G.; Savin, D.; Gu, Z.; Patterson, G. D.; Kim, S. H.; Ramsay, D. J.; Fishman, D.; Ivanov, I.; Sheina, E.; Slaby, E.; Oliver, J. Int. J. Polym. Anal. Charact. 2003, 8, 187. (46) Liu, R.; Gao, X.; Adams, J.; Oppermann, W. Macromolecules 2005, 38, 8845. (47) Pickup, S.; Blum, F. D. Macromolecules 1989, 22, 3961.

SolVent Drop EVaporation on Polymer Surfaces

the polymer concentration is decreased.21 It was discussed as a bulging in the center of the droplet for an initially high polymer concentration19 and a coffee-stain-like effect for lower initial concentrations. In our case the polymer concentration is mediated through the molar mass. Thus, it confirms our interpretation and additionally suggests that the basic principles for insoluble substrates can be utilized for the description of the process on soluble substrates. Consequently, for increasing molar masses above 210 kDa, the polymer concentration should decrease. Since it is associated with an increasing asymmetry of the resulting surface topology, instabilities as an origin for these structures are likely. For a decreasing polymer concentration with increasing molar mass of the substrate, the viscosity inside the droplet decreases. This leads to an increasing Marangoni number und thus might favor Marangoni-type convection inside the droplet.10,48,49 Thus, undulations with distances on the order of 50 µm were found for the dissolution of PMMA layers into solvent droplets.50 The undulations for our 1445 kDa PS have the same size.

Conclusions We have presented a study on the microstructuring of polymer surfaces with a pure solvent drop dependent on the molar mass (48) Bormashenko, E.; Pogreb, R.; Musin, A.; Stanevsky, O.; Bormashenko, Y.; Whyman, G.; Gendelman, O.; Barkay, Z. J. Colloid Interface Sci. 2006, 297, 534. (49) Chang, S. T.; Velev, O. D. Langmuir 2006, 22, 1459. (50) Gonuguntla, M.; Sharma, A. Langmuir 2004, 20, 3456.

Langmuir, Vol. 22, No. 26, 2006 11399

of the polymer. For a molar mass of 210 kDa we found a rotationsymmetric concave surface topology as a confirmation of earlier results. Most likely, this topology occurs because of an evaporation-driven flow of solvent and polymer to the outer rim of the droplet rather than owing to a mechanical deformation. When the molar mass is decreased, the outer rim is increasingly located toward the center. For 20.9 kDa a bump appears in the middle. This behavior was discussed as generated by an increasing polymer concentration in the droplet and thus a faster gelation when the entanglement molar mass is approached from above. For molar masses higher than 210 kDa an increasing degree of asymmetry was found. Concentration gradients within the droplet cannot be compensated by diffusion. Thus, instabilities of the Marangoni type might be likely to occur, leading to polymer bulges on the order of several tens of micrometers. The dependence of the resulting surface topology on the molar mass of the underlying substrate suggests that our technique has some potential for the estimation of the molar mass of polymers. Acknowledgment. We appreciate help from A. Best and N. Ho¨hn for technical support, J. Thiel and T. Wagner for the synthesis of the polymers, and T. Haschke for last-minute simulations. Funding from the DFG (grant GR2003/2 within the research group FOR516) is greatly acknowledged. LA061929L