Accelerating Dewetting on Deformable Substrates by Adding a Liquid

Feb 9, 2010 - Hermann-Herder-Strasse 3, 79104 Freiburg, Germany. Received November 22, 2009. Revised Manuscript Received January 18, 2010...
1 downloads 0 Views 4MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Accelerating Dewetting on Deformable Substrates by Adding a Liquid Underlayer Lin Xu,† G€unter Reiter,‡ Tongfei Shi,*,† and Lijia An*,† †

State Key Laboratory of Polymer Physics and Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun 130022, PR China, and ‡Physikalisches Institut, Universit€ at Freiburg, Hermann-Herder-Strasse 3, 79104 Freiburg, Germany Received November 22, 2009. Revised Manuscript Received January 18, 2010

We investigated the dependence of the dewetting velocity of a thin, low-viscosity polystyrene (PS) top film on a poly(methyl methacrylate) (PMMA) double layer consisting of a low-viscosity underlayer of thickness hL coated with a high-viscosity middle layer of thickness hM. The addition of the liquid underlayer generated complex nonmonotonic behavior of the dewetting velocity as a function of increasing hM. In particular, we observed an acceleration of dewetting for an intermediate range of hM. This phenomenon has been interpreted by a combination deformation of the middle elastic layer and a concurrent change in the contact angle. On one hand, deformation led to the formation of a trench that dissipated energy during its movement through the liquid underlayer and thus caused a slowing down of dewetting. However, with an increase in the thickness of the elastic middle layer, the size of the trench decreased and its influence on the dewetting velocity also decreased. On the other hand, the deformation of the elastic layer also led to an increase in the contact angle. This increase in the driving capillary forces caused an increase in the dewetting velocity.

Introduction The dewetting of thin liquid films is a common phenomenon that has generated important scientific interest in its fundamental physical processes because of its crucial impact on various technological processes, such as coating, painting, diffusion, multilayer adsorption, and so on. The dynamics of dewetting has been widely studied for several experimental conditions, and detailed theories have been developed for many cases. In particular, on solid substrates, our understanding of the experimentally observed dewetting dynamics is quite profound.1-5 As shown in previous papers on the dewetting of viscous fluids, the driving force for dewetting is proportional to the spreading parameter S = γs/g - γp/g - γp/s (γs/g, γp/g, and γp/s are the interfacial energy (tension) of the substrate/air, the polymer/air, and the polymer/substrate, respectively). In contrast to solid substrates, the dewetting dynamics on liquid or deformable *To whom correspondence should be addressed. E-mail: [email protected] Or [email protected]. Tel: þ86-431-85262137 or þ86-431-85262206. Fax: þ86431-85262969.

(1) Thiele, U. Eur. Phys. J. E 2003, 12, 409. (2) Seemann, R.; Herminghaus, S.; Neto, C.; Schlagowski, S.; Podzimek, D.; Konrad, R.; Mantz, H.; Jacobs, K. J. Phys: Condens. Matter. 2005, 17, S267. (3) Sharma, A.; Reiter, G. J. Colloid Interface Sci. 1996, 178, 383. (4) Reiter, G. Phys. Rev. Lett. 1992, 68, 75. (5) Gabriele, S.; Damman, P.; Sclavons, S.; Desprez, S.; Coppee, S.; Reiter, G.; Hamieh, M.; Akhrass, S. A.; Vilmin, T.; Rapha€el, E. J. Polym. Sci., Part B: Polym. Phys. 2006, 44, 3022. (6) Brochard-Wyart, F.; Martin, P.; Redon, C. Langmuir 1993, 9, 3682. (7) Lambooy, P.; Phelan, K. C.; Haugg, O.; Krausch, G. Phys. Rev. Lett. 1996, 76, 1110. (8) Qu, S.; Clarke, J.; Liu, Y.; Rafailovich, M. H.; Sokolov, J.; Phelan, K. C.; Krausch, G. Macromolecules 1997, 30, 3640. (9) Wang, C.; Krausch, G.; Geoghegan, M. Langmuir 2001, 17, 6269. (10) Segalman, R. A.; Green, P. F. Macromolecules 1999, 32, 801. (11) Faldi, A.; Composto, R. J.; Winey, K. I. Langmuir 1995, 11, 4855. (12) Pan, Q.; Winey, K. I.; Hu, H. H.; Composto, R. J. Langmuir 1997, 13, 1758. (13) Morariu, M. D.; Sch€affer, E.; Steiner, U. Eur. Phys. J. E 2003, 12, 375. (14) Lin, Z.; Kerle, T.; Russell, T. P.; Sch€affer, E.; Steiner, U. Macromolecules 2002, 35, 3971. (15) de Silva, J. P.; Geoghegan, M.; Higgins, A. M.; Krausch, G.; David, M.-O.; Reiter, G. Phys. Rev. Lett. 2007, 98, 267802.

7270 DOI: 10.1021/la904420d

substrates is more complex because of viscous dissipation within or elastic deformation of the substrate.6-23 On a liquid substrate of a low-viscosity fluid, the normal force component resulting from the balance of interfacial forces at the three-phase contact line (as expressed by S) shifts the contact line out of the plane of the initial interface. This, in turn, has a direct influence on the shape of the dewetting rim, which thus deviates from the circular shape. The more viscous the substrate, the less pronounced this deviation will be, eventually recovering the behavior on solid substrates. Typically, one observes that, all other parameters being the same, the higher the deformation of the substrate, the slower the dewetting process. This slowing down is mainly due to a dissipation of energy caused by either an increase in the higher interfacial friction (proportional to the size of the rim in the case of interfacial slippage) or viscoelastic effects causing dissipation within a rubbery substrate. In general, additional sources of energy dissipation cause a slowing down of dewetting. However, it is surprising to find that dewetting can also be accelerated even on deformed substrates. In this article, we investigate the dewetting behavior of polymer films on a complex layered substrate consisting of a double layer where a thin layer of high-molecular-weight polymer covers a thin layer of a lowmolecular-weight polymer. Both layers are supported by a solid silicon substrate. The chosen trilayer system is schematically shown in Figure 1. (16) Pototsky, A.; Bestehorn, M.; Merkt, D.; Thiele, U. Phys. Rev. E 2004, 70, 025201. (17) Pototsky, A.; Bestehorn, M.; Merkt, D.; Thiele, U. J. Chem. Phys. 2005, 122, 224711. (18) Bandyopadhyay, D.; Sharma, A. J. Chem. Phys. 2006, 125, 054711. (19) Bandyopadhyay, D.; Gulabani, R.; Sharma, A. Ind. Eng. Chem. Res. 2005, 44, 1259. (20) Fisher, L. S.; Golovin, A. A. J. Colloid Interface Sci. 2005, 291, 515. (21) Besancon, B. M.; Green, P. F. J. Chem. Phys. 2007, 126, 224903. (22) Bandyopadhyay, D.; Sharma, A. J. Phys. Chem. B 2008, 112, 11564. (23) Al Akhrass, S.; Reiter, G.; Hou, S. Y.; Yang, M. H.; Chang, Y.; Chang, F. C.; Wang, C. F.; Yang, A. C.-M. Phys. Rev. Lett. 2008, 100, 178301.

Published on Web 02/09/2010

Langmuir 2010, 26(10), 7270–7276

Xu et al.

Article

Figure 1. Schematic diagram of the trilayer.

To minimize or even avoid slipping of the dewetting polymer, the molecular weight of this polymer was chosen to be well below the entanglement molecular weight.24,25 At the temperature chosen for dewetting, the low-molecular-weight layer behaved like a viscous liquid whereas the high-molecular-weight middle layer behaved like a viscoelastic fluid in the plateau region. At room temperature, both the dewetting polymer and the substrate polymers were in a glassy state and the morphology of the polymer/polymer interface could be frozen in. The experiments presented here show that, depending on the respective thicknesses of substrate layers, the velocity of the dewetting film can be faster in the trilayer systems than in the bilayer system without the low-viscosity underlayer. A novel phenomenon is presented: the addition of a liquid underlayer can accelerate dewetting.

Experimental Section The system under investigation was a polymeric trilayer with a polystyrene (PS4.1K, Mw = 4.1 kg/mol, Mw/Mn = 1.05) film coated onto a high-molecular-weight poly(methyl methacrylate) (PMMA365K, Mw=365 kg/mol, Mw/Mn < 3) layer deposited on a low-Mw fluid PMMA layer consisting of poly(methyl methacrylate) (PMMA15K, Mw = 15 kg/mol, Mw/Mn < 3) of different thicknesses between 18 and 240 nm (as measured by ellipsometry) on top of a silicon substrate. PMMA15K films were spin coated from a chloroform solution onto a Si wafer with a native oxide layer. PMMA365K films were spin coated from a chloroform solution onto cleaved mica. The PMMA365K films were floated onto deionized water and deposited on the PMMA15K film. The trilayer films were prepared by spin coating the cyclohexane solution of PS4.1K onto the PMMA bilayer films. The thickness of the films was measured by ellipsometry. We measured three points on every polymer film to ensure that the thickness of these films was uniform. To initiate the dewetting process, the samples were heated in air to 160 °C (i.e., far above the glass-transition temperature of PS4.1K, Tg = 73 °C). The morphology of the surface was observed by optical microscopy (OM) in reflection mode and atomic force microscopy (AFM) in tapping mode using an SPA300HV with an SPI3800N controller (Seiko Instruments Inc, Japan). The spring constant of the cantilever is 2 N/m. The cantilever oscillated close to its resonance frequency between 65 and 75 kHz. The PS film was removed by immersing the samples in cyclohexane. After being dried, the remaining PMMA365K surface was imaged to obtain the morphology of the PS/PMMA interface directly.

Results and Discussion It can be indirectly seen from Figure 2a how VPS changed with increasing hM. As shown in Figure 2b, the radius of the holes increased in an approximately linear fashion with annealing time. This implies that the PS film does not slip significantly in the course of the dewetting process. In Figure 2e, we present the dewetting velocity (VPS) of the PS4.1K layer as a function of the film thickness (hM) of the PMMA365K middle layer. When the PMMA15K (24) de Gennes, P.-G. C. R. Acad. Sci. 1979, 228B, 219. (25) Brochard-Wyart, F.; de Gennes, P.-G.; Hervert, H.; Redon, C. Langmuir 1994, 10, 1566.

Langmuir 2010, 26(10), 7270–7276

underlayer was absent or its thickness (hL) was only 18 nm, an increase in hM did not measurably influence VPS. However, when the PMMA15K underlayer became thicker (hL g 52 nm), the behavior of VPS (with respect to dewetting on only the middle layer without any liquid underlayer) can be divided into three regions of increasing value of hM: (I) for small values of hM, adding a PMMA15K underlayer decreased VPS; (II) for intermediate values of hM, adding a PMMA15K underlayer accelerated dewetting (increased VPS); (III) for the largest values of hM, adding a PMMA15K underlayer did not affect VPS. Consequently, in comparison with the simple bilayer system (without the PMMA15K underlayer), an “overshoot” in dewetting velocity was found in region II of the trilayer system (with a PMMA15K underlayer). Figure 3 presents AFM profiles of the PS4.1K/PMMA365K interface taken at the location of the rim for holes of approximately equal diameter. When the PMMA15K layer was absent or very thin (18 nm), the PS/PMMA interface hardly deformed (Figure 3a,b) and stayed approximately planar. With the PMMA15K layer becoming thicker, in region I the PS/PMMA interface deformed during the course of dewetting and a trench was clearly observed underneath the rim (black line in Figure 3c-f). In region II, the volume of the trench is small (red line in Figure 3c-f), and in region III, the PS/PMMA interface hardly deformed at all. No obvious trench was observed (green line in Figure 3c-f). From Figures 2 and 3, it may be concluded that the presence of the PMMA15K layer not only decreased VPS but also could cause an increase in VPS. This was found even for a deformable substrate where deformation could cause an increase in energy dissipation and thus would be expected to slow down dewetting. In the following text, we will discuss these results. From Figure 3, it can be clearly seen that the contact line can be lifted in the direction normal to the substrate. When the PMMA15K underlayer was very thin (18 nm), the height of the lifted contact line was very small, probably also because the PMMA chains of the underlayer were adsorbed onto the solid silicon support.26-28 When the added liquid underlayer became thick, a trench was formed with the position of the height of the contact line increasing. Obviously, with hL increasing, the influence of the solid silicon support on the mobility of the PMMA chains decreased. The deformation of the PS/PMMA interface near the contact line became more pronounced during the dewetting of the upper PS film. Figure 4 shows that the profile of the PS/PMMA interface at the location of the rim varied with increasing hL. To see the influence of the added liquid underlayer clearly, we chose a thin hM. From Figure 4, with hL increasing, the height of the contact line increased and the depth of the trench became larger. In the course of the dewetting process, the PS/PMMA interface deformed in order to minimize the Gibbs free energy. After the formation of a hole, the vertical component of the upper PS surface tension, γLV sin θ1, raised the contact line29-41 (the (26) Forrest, J. A.; Dalnoki-veress, K.; Stevens, J. R.; Dutcher, J. R. Phys. Rev. Lett. 1996, 77, 2002. (27) Chen, X. C.; Anthamatten, M. Langmuir 2009, 25, 11555. (28) Xue, L. J.; Han, Y. C. Langmuir 2009, 25, 5235. (29) Shanahan, M.; Carre, A. Langmuir 1994, 10, 1647. (30) Carre, A.; Shanahan, M. Langmuir 2001, 17, 2982. (31) Pericet-Camara, R.; Best, A.; Butt, H.; Bonaccurso, E. Langmuir 2008, 24, 10565. (32) Lester, G. R. J. Colloid Interface Sci. 1961, 16, 315. (33) Rusanov, A. I. Colloid J. USSR 1975, 37, 614. (34) Extrand, C. W.; Kumagai, Y. J. Colloid Interface Sci. 1996, 184, 191. (35) White, L. R. J. Colloid Interface Sci. 2003, 258, 82. (36) Carre, A.; Gastel, J. C.; Shanahan, M. E. R. Nature 1996, 379, 432. (37) Fredrickson, G. H.; Ajdari, A.; Leibler, L.; Carton, J. P. Macromolecules 1992, 25, 2882. (38) Long, D.; Ajdari, A.; Leibler, L. Langmuir 1996, 12, 5221.

DOI: 10.1021/la904420d

7271

Article

Xu et al.

Figure 2. (a) Comparison of OM images (size: 85 μm  75 μm) taken at similar annealing times (about 5 min; the time when the hole started to form was set to zero). The scale bar is 10 μm. (b-d) Dewetting dynamics for different thicknesses hL of the underlying liquid PMM15K layer. The symbols represent different thicknesses hM of the PMMA365K middle layer as indicated in the legend. (e) Plot of the dewetting velocity of the PS upper layer (hU = 66 nm) and the film thickness of the high-molecular-weight PMMA middle layer, hM. The symbols represent different thicknesses hL of the PMMA15K underlayer as indicated in the legend. The level of the dewetting velocity of the two-layer system (PS/PMMA365K) without any added liquid underlayer is shown by the horizontal dotted line in image e. For every dewetting event, the error in the dewetting velocity is smaller than 5%.

contact angle of PS on the substrate, θ=θ1 þ θ2 ; θ1 and θ2 are defined as contact angles with respect to a horizontal line at the position of the contact line). On such deformable PMMA365K/PMMA15K substrates, two factors can influence the dewetting velocity of the upper PS film. The presence of the trench can cause an increase in the dissipation of capillary energy, which leads to a decrease in the dewetting velocity. On the contrary, the lifting of the contact line may translate into an increase in the contact angle, which in turn accelerates dewetting because previous papers have reported a dewetting velocity of V ≈ θ3 for nonslipping films. For hL g 52 nm, in region I the increase in the contact angle θ can accelerate dewetting. However, during hole growth, obvious (39) Yu, Y.-S.; Yang, Z.; Zhao, Y.-P. J. Adhes. Sci. Technol. 2008, 22, 687. (40) Pericet-Camara, R.; Bonaccurso, E.; Graf, K. ChemPhysChem 2008, 9, 1738. (41) Pericet-Camara, R.; Auernhammer, G. K.; Koynov, K.; Lorenzoni, S.; Bonaccurso, E. Soft Matter 2009, 5, 3611.

7272 DOI: 10.1021/la904420d

trenches were observed at the PS/PMMA interface; especially for a thick underlayer with hL =181 nm, the depth of the trench is larger than hM=22 ( 3 nm (Figure 4). Energy is dissipated during the movement of these comparatively large trenches. The dewetting velocity decreased with the dissipated energy being proportional to the volume of the trench. The trench was so large that it became the main factor influencing the dewetting velocity. Thus, in region I the addition of the liquid underlayer decreased the dewetting velocity of PS films. Figure 5 shows the change in the rim shape and the profile of the PS/PMMA interface near the contact line for increasing hM. From Figure 5d, it is found that θ1 hardly varied with the thickness of the middle layer and the volume of the trench became small with increasing hM. Because of the diminution of the trench with increasing hM, the increase in the contact angle became the main factor influencing the dewetting velocity. Thus, in region II the addition of a liquid underlayer could increase VPS (with respect to a rather rigid substrate of a high-molecular-weight PMMA layer) although some dissipation Langmuir 2010, 26(10), 7270–7276

Xu et al.

Article

Figure 3. AFM profiles of the PS/PMMA365K interface at the location of the rim for various thicknesses hL of the liquid underlayer. All holes have similar diameters.

Figure 4. AFM profiles of the PS/PMMA365K interface at the location of the rim for hM = 22 ( 3 nm and variable hL. The height (H) level H = 0 represents the position of the initial PS/ PMMA365K interface. All holes have similar diameters.

was still expected to occur during the deformation of the substrate. From Figure 6, it is found that the thickness h*M for the transition from region I to region II initially increased with hL and became constant for hL larger than about 100 nm. For increasing hL (but constant hM) the formation of a large trench became easier (Figure 3). In region III, the PS/PMMA interface did not deform measurably during hole growth. No trenches were observed, and θ2 was approximately zero. Then, the two factors discussed above did not influence the dewetting velocity anymore. Thus, the presence of the liquid underlayer did not affect the dewetting process and VPS was the same as for the bilayer system (without the liquid underlayer). Figure 7 shows a schematic diagram that Langmuir 2010, 26(10), 7270–7276

summarizes our main results. A deep, steep trench resulting from a thin middle layer may increase the dissipation. A thicker middle layer did not allow for such a deep trench, and contact angle θ2 (about 1-4°) still remained significantly high compared to θ (about 7-8°). For much thicker middle layers, no trench was formed at all and θ2 was approximately zero. To understand this interesting result further, Figure 8 shows the change in the shape of the PS/PMMA interface during hole growth at a location near the hole. From Figure 8a,b for a thin (18.5 nm) middle layer of PMMA365K, obvious trenches could be observed. The shape of these trenches at the PS/PMMA interface did not change significantly during hole growth. We also found that the deformed PS/PMMA interface within the hole did not recover measurably after the contact line passed. The height level of the hole floor initially decreased with hole growth and reached a value often significantly lower than the level of the unperturbed PS/PMMA interface. Energy was dissipated by the movement of the trench and the corresponding deformation of the PS/PMMA interface. Consequently, VPS was smaller than for dewetting on a simple bilayer (i.e., a 18.5-nm-thick PMMA356K layer without any liquid underlayer). However, when hM became thick (as shown in Figure 8c,d) we found that the PS/PMMA interface did not deform much although the same thick liquid underlayer was present. Inside the hole, the height level of the PS/PMMA interface was not decreased during hole growth. The results show that only a small amount of the invested capillary energy was dissipated by moving such small trenches. DOI: 10.1021/la904420d

7273

Article

Xu et al.

Figure 5. (a) AFM profile images of the PS/PMMA interface at the location of the rim and (b) AFM profile images of the rim. The curves in plots a and b are merged into the curve in plot c. (d) Magnification of the circular region in plot c. All holes have similar diameters.

Figure 6. Film thickness h*M at the transition between the reduced and increased dewetting velocity as a function of the thickness hL of the PMMA15K underlayer.

Figure 7. Schematic diagram summarizing the typical interfacial profiles for the three regions of dewetting velocity discussed in the text.

However, such small deformations of the PS/PMMA interface could not explain the increase in the dewetting velocity in region II. One would rather expect that the dewetting velocity would approach its limiting value for the infinitely thick middle layer, which is found to be identical to dewetting on a PMMA365K layer without any liquid underlayer. Thus, an increase in VPS required a source for an additional contribution to the driving capillary forces. This source was found in the increase in the contact angle θ because θ2 increased but θ1 did not 7274 DOI: 10.1021/la904420d

change with increasing hM (Figure 5). Although contact angle θ2 was small (about 1-4°) in region II, together with contact angle θ1 of PS on PMMA (which was also small, about 7°), this would account for an increase in VPS of about 50% (V ≈ θ3, θ=θ1 þ θ2), thus explaining the so-called overshoot phenomenon. To verify further if the above explanation of the overshoot phenomenon is correct, dewetting was performed for a four-layer system (Figure 9a) where a gold (Au) layer of variable thickness was inserted between the PMMA15K and PMMA365K layers of the trilayer system. The thicknesses of the PS4.1K, PMMA365K, and PMMA15K layers were 60, 16, and 97 nm, respectively. Figure 9b shows the dewetting velocity of the PS upper layer as a function of the film thickness (average value) of the Au layer. For short sputtering times (i.e., small amounts of deposited gold), gold did not form a continuous layer on the PMMA15K layer. In Figure 9c, this is represented by the large roughness of the gold layer (10 nm). The PMMA365K layer deformed during the dewetting process (Figure 9d). Thus, the existence of a trench in the PMMA15K layer decreased VPS. With increasing sputtering time, the roughness became small (2 nm) and a continuous Au layer was formed. The substrate (PMMA365K/gold layer/ PMMA15K) did not deform during dewetting. A continuous gold layer can be considered to be a solid with a Young’s modulus of about 104 MPa at 160 °C, a value far higher than that of PMMA365K, about 1 MPa, at this temperature. Thus, the capillary forces were not able to deform this gold layer although it was only 30 nm thick. Consequently, the overshoot region in dewetting (Figure 9b) disappeared. From the above experimental results, we may conclude that coating a liquid underlayer with a thin elastic layer has two consequences. The formation of a trench is possible only if the elastic layer can be deformed by capillary forces. Large trenches lead to a decrease in VPS. When the interface can deform but only small trenches can build up, the presence of the liquid underlayer provokes an acceleration of dewetting. For a nondeformable interface (i.e., a thick elastic layer or a thin elastic Langmuir 2010, 26(10), 7270–7276

Xu et al.

Article

Figure 8. (a, c) AFM profiles across a dewetted hole of radius R for the PS/PMMA365K interface for a rather thin and a comparatively thick middle layer, respectively. (b, d) Enlargements of the left parts (holes grow in the negative lateral direction) of the curves in plots a and c, respectively.

layer with a high modulus), the presence of a liquid underlayer does not have an influence on VPS. Previous papers7-9 have reported that the presence of a trench can cause a decrease in dewetting. This slowing down is mainly due to the dissipation of energy (e.g., viscoelastic effects causing dissipation within a rubbery substrate). However, it is surprising to find that a liquid underlayer can also lead to accelerated dewetting, even when the substrate is deformable. The overshoot in dewetting velocity (Figure 2) represents a novel phenomenon. Thus, dewetting dynamics on complex, layered substrates, such as the elastic layer/liquid layer substrate used in our experiment, can also exhibit complex behavior. Several questions, however, remain open. For example, the exact influence of the PMMA15K underlayer on the morphology of the PS4.1K/PMMA365K interface and the temporal evolution of the contact angle are not yet known in detail.

Summary

Figure 9. (a) Schematic diagram of the four layers. (b) Plot of the dewetting velocity of the PS upper layer and the film thickness of the Au layer. The level of the dewetting velocity of the trilayer system (PS/PMMA365K/Au layer) without any added liquid underlayer is shown by the horizontal dotted line in plot b. A series of AFM images of the PS/PMMA365K interface are shown: (c) sputtering time = 10 s (hgold = 3 nm), no annealing and (d) sputtering time = 10 s (hgold = 3 nm), annealed 12 min at the location of the hole. Langmuir 2010, 26(10), 7270–7276

We carried out a series of dewetting experiments on polymer trilayer systems. A liquid PS film was deposited on a liquid, lowMw PMMA layer coated with a high-Mw PMMA elastic layer. With the thickness of the high-Mw PMMA middle layer increasing, it was found that the addition of a liquid underlayer not only could decrease the dewetting velocity of PS films but also could accelerate the dewetting process, even for deformable substrates. This overshoot phenomenon has been interpreted by a combination of two effects. First, the middle elastic layer could deform, which led to the formation of a trench. The trench caused a dissipation of energy during its movement through the liquid underlayer and thus a slowing down of dewetting. However, with an increase in the thickness of the elastic middle layer, the size of the trench decreased and its influence on the dewetting velocity also decreased. Second, the deformation of the elastic layer also led to the formation of contact angle θ2. Thus, this increase in the driving capillary forces caused an increase in the dewetting velocity. In a word, dewetting on complex substrates may exhibit surprising behavior. Complex substrates can be widely found in natural systems. Thus, a DOI: 10.1021/la904420d

7275

Article

comprehensive understanding of the dewetting process on such complex substrates is needed. It is hoped that our study may stimulate more systematic experiments that further contribute to our understanding of dewetting behavior on complex substrates.

7276 DOI: 10.1021/la904420d

Xu et al.

Acknowledgment. This work was supported by the National Natural Science Foundation of China (50973110) Programs and the Fund for Creative Research Groups (50921062) and was subsidized by the Special Funds for National Basic Research Program of China (2010CB631100).

Langmuir 2010, 26(10), 7270–7276