ARTICLE pubs.acs.org/Langmuir
Nonfunctionalized Polydimethyl Siloxane Superhydrophobic Surfaces Based on Hydrophobic-Hydrophilic Interactions Georgios Polizos,† Enis Tuncer,*,†,‡ Xiaofeng Qiu,†,|| Tolga Aytug,† Michelle K. Kidder,† Jamie M. Messman,† and Isidor Sauers† †
Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States Dielectrics & Electrophysics Lab, GE Global Research Center, Niskayuna, New York 12309, United States IBM TJ Watson Research Center, P.O. Box 218, Rte 134, Yorktown Heights, New York 10598, United States
)
‡
ABSTRACT: Superhydrophobic surfaces based on polydimethyl siloxane (PDMS) were fabricated using a 50:50 PDMS-poly(ethylene glycol) (PEG) blend. PDMS was mixed with PEG, and incomplete phase separation yielded a hierarchic structure. The phase-separated mixture was annealed at a temperature close to the crystallization temperature of the PEG. The PEG crystals were formed isothermally at the PDMS/PEG interface, leading to an engineered surface with PDMS spherulites. The resulting roughness of the surface was studied by scanning electron microscopy (SEM) and atomic force microscopy (AFM). The PDMS spherulites, a few micrometers in diameter observed from SEM images, were found to have an undulated (rippled) surface with nanometer-sized features. The combination of micrometer- and nanometer-sized surface features created a fractal surface and increased the water contact angle (WCA) of PDMS more than 60°, resulting in a superhydrophobic PDMS surface with WCA of >160°. The active surface layer for the superhydrophobicity was approximately 100 μm thick, illustrating that the material had bulk superhydrophobicity compared to conventional fluorocarbon or fluorinated coated rough surfaces. Theoretical analysis of the fractal surface indicates that the constructed surface has a fractal dimension of 2.5, which corresponds to the Apollonian sphere packing.
S
urface properties of materials with respect to water wettability are determined by both the water-material interaction and the surface microstructure of the material.1-8 Early investigations by Wetzel9 and Cassie and Baxter10 presented the importance of surface texturing on hydrophobicity. Later in the late 1990s, publications by a Japanese group11-13 initiated discussion and reporting of superhydrophobic surfaces and materials for numerous applications. Since then, many different material systems and surface structures have been reported.2-8 It is obvious that once the smooth surface of a hydrophobic material is converted to a large, structured surface area superhydrophobicity is obtained. Here, we present a method of creating a rough surface structure and a superhydrophobic material using hydrophobic-hydrophilic interactions in a polymer blend composed of polydimethyl siloxane (PDMS) and poly(ethylene glycol) (PEG). The material does not require any fluorocarbon groups or fluorination; therefore, it is environmentally friendly. PDMS is one of the engineering materials14 known to be a hydrophobic elastomer with a water contact angle (WCA) around 100° ( 3°. Careful processing of PDMS and altering its surface topology would yield a superhydrophobic material with contact angle of >150°.15 Different methods other than the one presented here have been reported to improve the WCA of PDMS in the literature.16-18 We have structured the PDMS surface using hydrophobic-hydrophilic interactions. A commercial-grade silica-filled PDMS with the trade name Sylgard 184 was supplied from Dow Corning Inc., USA. The un-cross-linked PDMS and r 2011 American Chemical Society
manufacturer-recommended amount of curative were added in a solution of deionized water and PEG with mol wt = 3400 (supplied from Alfa Aesar, USA) in a weight ratio WPDMS : Wcurative : WPEG : WH2 O ¼ 1 : 0:1 : 0:3 : 0:3: The immiscible blend was mixed at 1200 rpm for 10 min in a planetary mixer, which provides high shear for better homogenization and degassing capability during mixing. The resulting colloidal, phase-separated (opaque) solution was annealed at 310 ( 2 K for 15 h. The annealing activated three different processes: (i) evaporate the water, (ii) isothermally crystallize the PEG phase, and (iii) cross-link the PDMS phase. The annealing temperature (Ta) was selected according to the differential scanning calorimetry (DSC) thermogram of the pure PEG, presented in Figure 1. Taking into account the crystallization temperature (Tc = 307 K) and the two melting peaks Tm at 325 and 333 K (which correspond to different crystal sizes19), the selected annealing temperature Ta was close to Tc but far from the melting temperatures. After the removal of water, the PEG crystallization kinetics were fast and the crystallization process was complete in 20-30 min at 310 K.19 The crystallization time was significantly shorter than the time needed for the crosslinking of the PDMS phase. The cross-linking took place in Received: October 25, 2010 Revised: January 2, 2011 Published: February 04, 2011 2953
dx.doi.org/10.1021/la1042712 | Langmuir 2011, 27, 2953–2957
Langmuir
Figure 1. DSC thermograms of the PEG (mol wt = 3400) presenting the crystallinity melting peaks and the exothermic cold crystallization. The temperature scan rate was 10 K/min.
Figure 2. Attenuated total reflection infrared (ATR-IR) spectra of PDMS, PEG, and constructed superhydrophobic PDMS surfaces.
several hours at Ta. The annealing temperature allowed us first to form the PEG crystals and subsequently force the PDMS polymer network to order in the presence of the PEG crystals. After the full curing cycle of the PDMS, an incomplete phase separation of the polymer blend was obtained—a layer rich in cross-linked PDMS was constructed on top of a PEG-rich layer. The sample was postcured according to the specifications of the manufacturer. The PDMS layer was later peeled from the PEG, rinsed with deionized water, and dried in a vacuum oven before the structural characterization and WCA measurements. The sample thickness was approximately 400 μm. The WCA measurements were performed with 5 μL droplets under a microscope using several droplets. The hydrophobic-hydrophilic interactions between the two phases cause the PDMS chains to form either micelles20,21 or a PDMS shell around the PEG spheres. However, the latter statement was not supported by the attenuated total reflection infrared (ATR-IR) spectroscopy analysis performed on the PDMS surface, shown in Figure 2. The PEG modes are visible around 1400 cm-1 and 2850 cm-1, while as shown in the figure, the SH PDMS and neat PDMS have similar ATR-IR characteristics. No modes related to PEG were observed that could indicate a possibility of the PEG phase being encapsulated in the PDMS spherulites, and all free PEG should have been washed away during the rinsing procedure.
ARTICLE
The spherulites on the surface were imaged with a scanning electron microscope (SEM), cf., Figure 3. The size of the PDMS particles in Figure 3 ranged from submicrometer to approximately 10 μm. The surface of the PDMS particles indicate a rippled micrometer-scale structure, similar to the corals, that increased the roughness of the surface as shown in Figure 3. The distance between self-assembled undulated stripes is between 1 and 2 μm. Using an atomic force microscope (AFM) further assisted us in extracting nanometer-scale features, which were like spikes in the range of several tens to a hundred nanometers in height with similar dimensions at the base as shown in Figure 4. The bright spots in the AFM images are the bumps that are the rippled structures in Figure 3, indicating that the AFM was performed on one of the spherulites. The distance between the bright stripes in the AFM picture correponds to the approximate distance observed between the strips in the SEM images in Figure 3. The SEM images could not show 120-nm-high nanospikes (nanostructures) due to the sample preparation procedure used for the SEM—the sample was sputtered with carbon to prevent charging of the sample surface. The procedure heats the sample, which could cause melting/softening of the PDMS. In addition, the carbon alters the nano features on the surface that were visible with AFM. The integration of micrometer- and nanometer-sized surface features created a hierarchy in the surface roughness, starting with the PDMS spherulites, their surface texture, and spikes. The fractal nature of the surface should lead to a superhydrophobic surface22 with a WCA more than 160° ((2°) because of the intrinsic hydrophobicity of the PDMS. Compared to the WCA (100°) of the smooth PDMS surface, the WCA of the constructed PDMS surface was larger than 160°, shown in Figure 5, which was the direct result of the surface roughness, a 60° improvement. The created structure was soft and did not lose properties when bent. Finally, to emphasize the benefit of the constructed PDMS, an optical microscope image of the cross section of the peeled layer is shown in Figure 6. The depth of the structured PDMS phase was approximately 100 μm. It should be noted that typical superhydrophobic surfaces are either based on surface chemical treatments, that are a few nanometers thick (i.e., monolayers of silanes, fluorocarbons, etc.), or fabricated by a coating process and suffer from poor adhesion performance. In contrast to the conventional methods, the structured surface presented herein has a persistent layer, which would have a long lifetime and durability—the superhydrophobic PDMS layer has an appreciable thickness, which indicates bulk superhydrophobicity. Performance of the constructed PDMS was tested at elevated temperatures to determine the amount of WCA degradation that would be of importance in outdoor applications at different climates. WCA measurements at 70 °C showed that the shape of the water droplet was not changed, and a slight increase in the WCA was observed (170°). The measurement were performed over a short period using a water-filled Petri dish to prevent water evaporation. Changes in superhydrophobic behavior were previously observed in porous materials.23 The increase in the WCA can be attributed to changes in the PDMS structure due to its thermal expansion. A representative model for rough surfaces, assuming a fractal surface, was presented as follows:1,11,13,24 cos θf ¼ ðL=lÞD-2 cos θ 2954
dx.doi.org/10.1021/la1042712 |Langmuir 2011, 27, 2953–2957
Langmuir
ARTICLE
Figure 3. Scanning electron microscopy images of the constructed PDMS surface with polydispersed spherulites are presented in different magnifications. The scale bars on the bottem left corners are (a) 50 μm, (b) 10 μm, (c) 5 μm, and (d) 2 μm, respectively.
Figure 4. Atomic force microscope (a) two- and (b) three-dimensional images of the constructed PDMS surface. Because of the undulated patterns on the spherulite surface shown in Figure 3, needle-like spikes of 60 to 120 nm in height formed, resulting in a superhydrophobic PDMS surface. The scanning area is 5 μm.
where θf and θ are the contact angles for the fractal and smooth surfaces, respectively; the ratio L/l is the roughness correction factor,1 which is the largest (L) to smallest (l) length scale for a given surface; and D is the fractal dimension for a given surface.
The above equation does not take into account the surface coverage of the fractal structure and would be applicable to solid coverage of the roughness. However, as in the Cassie-Baxter formalism,10 one needs to include the fractional area for the 2955
dx.doi.org/10.1021/la1042712 |Langmuir 2011, 27, 2953–2957
Langmuir
ARTICLE
Figure 5. Water droplets on the constructed PDMS surface. Observe that the image on the right is for a bent PDMS surface.
Figure 6. Cross-sectional view of the constructed PDMS surface. The scale bars are 100 μm. The hierarchical layer is more than 200 μm thick.
fractal structure, which was reported by Feng and collaborators3 as cos θf ¼ fs ðL=lÞD-2 cos θþfs -1 where fs is the fractional coverage. This equation becomes very important in determining the fractal dimensions and the surface structure of materials just by using contact angle measurement. We will show this as follows. Consider that the smooth surface contact angle θ is 100° and the ratio of largest to smallest structural differences (L/l) is 12. This number was estimated using the size ratio of large to small spherulites from the SEM images. Any small deviation from this value would yield a slight deviation of the measurement point presented in Figure 7. Although the nano spikes observed with the AFM would be important, their high aspect ratio and surface coverage is not high enough to achieve the observed WCA as presented next. We have simulated a chart for constant contact angles for a given fractal dimension D and fractional coverage fs, as shown in Figure 7, where contact angles for the fractal surface are plotted at 5° intervals from 90° to 180°. To better understand the chart, let us assume that the fractal dimension D of the rough surface is 2.6; then to have a contact angle of 150, one needs fs = 0.6, or a fractional coverage fs of 0.6 would yield a contact angle value of 150°. This assumption is shown in Figure 7 with vertical and horizontal lines that intersect the contact angle 150° curve. When we applied the same procedure to our PDMS sample, the result was a revelation. We first took the right-hand SEM image in Figure 3 and estimated the fractional coverage fs of the top layer of spherulites, as shown in the inset of Figure 7, fs = 31/176. This SEM image was representative of the whole surface. Later, using that value and the measured contact (160°), we realized that the
Figure 7. Equi-contact angle values at 5° intervals between 90° and 180° as a function of fractal dimension D and fractional coverage fs for L/ l = 12. The contact angle of constructed PDMS is shown with an open symbol (O). The angles 150° and 160° are marked with arrows. The inset illustrates the calculation of fractional coverage, and the rectangular spots indicate the top layer of the rough surface with fs ≈ 0.17 (the SEM image is the same one shown in Figure 3). The ratio of the largest to smallest (L/l) features is taken to be 12 in the calculations. Observe that the Apollonian sphere packing yields a fractal dimension 2.47.
surface fractal dimension D should be around 2.5, which is a value very close to the one reported for Apollonian sphere packing by Borkovec et al.,25 D = 2.47. This is an important observation in that one can use straightforward contact angle measurements to determine the fractal dimensions of surfaces. However, one needs to know fractional surface coverage fs, structural information (L/l), and smooth surface contact angle value θ. In conclusion, a superhydrophobic material based on PDMS was fabricated by developing an imprint method that employed hydrophobic-hydrophilic reactions in a polymer blend. The PEG-PDMS mixture created a surface with a hierarchic, fractallike structure. The induced roughness resulted in an enhancement of the hydrophobicity of PDMS and in a superhydrophobic material with WCA larger than 160°. We have used a theoretical approach to determine the surface fractal dimension in our sample, which agreed well with the fractal dimensions of the Apollonian sphere packing. The presented method is straightforward and based on imperfect homogenization procedure in a two-phase mixture. The only drawback of the method is the peeloff stage of the superhydrophobic layer. However, improvements in the mixing and postcuring could avoid this step. Finally, the superhydrophobic surfaces in this study are nonfunctionalized with fluorocarbons or other reactive hydrophobic groups—an environmentally friendly material.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. 2956
dx.doi.org/10.1021/la1042712 |Langmuir 2011, 27, 2953–2957
Langmuir
ARTICLE
’ ACKNOWLEDGMENT Research sponsored by the U.S. Department of Energy, Office of Electricity Delivery and Energy Reliability, Advanced Cables and Conductors Program for Electric Power Systems under contract DE-AC05-00OR22725 with Oak Ridge National Laboratory, managed and operated by UT-Battelle, LLC. ’ REFERENCES (1) Hazlett, R. D. J. Colloid Interface Sci. 1990, 137, 527–533. (2) Nakajima, A.; Hashimoto, K.; Watanabe, T. Monatsh. Chem. Chemical Monthly 2001, 132, 31–41. (3) Feng, L.; Li, S.; Li, Y.; Li, H.; Zhang, L.; Zhai, J.; Song, Y.; Liu, B.; Jiang, L.; Zhu, D. Adv. Mater. 2002, 14, 1857–1860. (4) Roach, P.; Shirtcliffe, N. J.; Newton, M. I. Soft Matter 2008, 4, 224–240. (5) Nosonovsky, M.; Bhushan, B. Adv. Funct. Mater. 2008, 18, 843– 855. (6) Zhang, X.; Shi, F.; Niu, J.; Jiang, Y.; Wang, Z. J. Mater. Chem. 2008, 18, 621–633. (7) Jung, Y.; Bhushan, B. Nanotechnology 2006, 17, 4970–4980. (8) Shirtcliffe, N.; McHale, G.; Atherton, S.; Newton, M. Adv. Colloid Interface Sci. 2010, 161, 124–138. (9) Wetzel, R. N. Ind. Eng. Chem. 1936, 28, 988–994. (10) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546– 551. (11) Shibuichi, S.; Yamamoto, T.; Onda, T.; Tsujii, K. J. Colloid Interface Sci. 1998, 208, 287–294. (12) Onda, T.; Shibuichi, S.; Satoh, N.; Tsujii, K. Langmuir 1996, 12, 2125–2127. (13) Shibuichi, S.; Onda, T.; Satoh, N.; Tsujii, K. J. Phys. Chem. 1996, 100, 19512–19517. (14) Johnson, R. T.; Biefeld, R. M.; Sayre, J. A. Polym. Eng. Sci. 1984, 24, 435. (15) Wu, Y.; Chen, Z.; Zeng, X. Appl. Surf. Sci. 2008, 254, 6952– 6958. (16) Tserepi, A. D.; Vlachopoulou, M.-E.; Gogolides, E. Nanotechnology 2006, 17, 3977. (17) Tsougeni, K.; Tserepi, A.; Boulousis, G.; Constantoudis, V.; Gogolides, E. Plasma Proc. Polym. 2007, 4, 398–405. (18) Vlachopoulou, M. E.; Petrou, P. S.; Kakabakos, S. E.; Tserepi, A.; Gogolides, E. J. Vac. Sci. Technol., B 2008, 26, 2543–2548. (19) Bogdanov, B.; Vidts, A.; Schacht, E.; Berghmans, H. Macromolecules 1999, 32, 726. (20) Kickelbick, G.; Bauer, J.; Huesing, N.; Andersson, M.; Holmberg, K. Langmuir 2003, 19, 10073. (21) Zhang, L.; Lin, J.; Lin, S. Macromolecules 2007, 40, 5582. (22) Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K.; Watanabe, T. Langmuir 2000, 16, 5754. (23) Shirtcliffe, N.; McHale, G.; Newton, M.; Perry, C. C.; Roach, P. Chem. Commun. 2005, 24, 3135–3137. (24) Synytska, A.; Ionov, L.; Grundke, K.; Stamm, M. Langmuir 2009, 25, 3132–3136. (25) Borkovec, M.; Paris, W. D.; Peikert, P. Fractals 1994, 2, 521– 526.
2957
dx.doi.org/10.1021/la1042712 |Langmuir 2011, 27, 2953–2957
