Simultaneous Tailoring of Surface Topography and Chemical

Langmuir 2004, 20, 8131-8136

8131

Simultaneous Tailoring of Surface Topography and Chemical Structure for Controlled Wettability Nui Takeshita,† Linda A. Paradis,† Didem O ¨ ner,‡ Thomas J. McCarthy,‡ and ,† Wei Chen* Chemistry Department, Mount Holyoke College, South Hadley, Massachusetts 01075, and Polymer Science & Engineering Department, University of Massachusetts, Amherst, Massachusetts 01003 Received March 8, 2004. In Final Form: June 29, 2004 Wettability was controlled in a rational manner by individually and simultaneously manipulating surface topography and surface chemical structure. The first stage of this research involved the adsorption of charged submicrometer polystyrene latex particles to oppositely charged poly(ethylene terephthalate) (PET) film samples to form surfaces with different topographies/roughness; adsorption time, solution pH, solution ionic strength, latex particle size, and substrate charge density are external variables that were controlled. The introduction of discrete functional groups to smooth and rough surfaces through organic transformations was carried out in the second stage. Amine groups (-NH2) and alcohol groups (-OH) were introduced onto smooth PET surfaces by amidation with poly(allylamine) and adsorption with poly(vinyl alcohol) (PVOH), respectively. On latex particle adsorbed surfaces, a thin layer of gold was evaporated first to prevent particle redistribution before chemical transformation. Reactions with functionalized thiols and adsorption with PVOH on patterned gold surfaces successfully enhanced surface hydrophobicity and hydrophilicity. Particle size and biomodal particle size distribution affect both hydrophobicity and hydrophilicity. A very hydrophobic surface exhibiting water contact angles of 150°/126° (θA/θR) prepared by adsorption of 1-octadecanethiol and a hydrophilic surface with water contact angles of 18°/8° (θA/θR) prepared by adsorption of PVOH were prepared on gold-coated surfaces containing both 0.35 and 0.1 µm latex particles. The combination of surface topography and surface-chemical functionality permits wettability control over a wide range.

Introduction The interfaces between solid objects and solutions in contact with them impact a wide range of real-world situations, and wettability is the central issue. A fundamental understanding of the basis for wettability will impact technologies that are concerned with solidsolution interactions, such as sensors, biomedical implants, anticorrosion and antifouling coatings, and membrane separations. The wettability of solids by solutions has been addressed, in scientific terms, for the past two centuries. Young described the equilibrium contact angle of a liquid droplet on a solid surface in 1805.1 If you place a drop of liquid on a surface, it forms a lens with a shape that is a section of a sphere. This shape is defined by the contact angle, which is determined by the interfacial free energy values of the air-liquid interface, the air-solid interface, and the liquid-solid interface. The contact angle assesses the wettability of the outermost few angstroms of the surface and is affected by both surface topography and surface chemical structure. Various theoretical approaches have taken Young’s equation beyond wetting on ideal (flat, smooth, and homogeneous) surfaces and have explored the effect of heterogeneities, both geometric and chemical, on wetting. For smooth, chemically heterogeneous surfaces, Cassie defined an “average” contact angle in 1948,2 which is the weighted mean of the angles that the drop would take on * To whom correspondence should be addressed. Email: [email protected]. † Mount Holyoke College. ‡ University of Massachusetts. (1) Young, T. Philos. Trans. R. Soc. London 1805, 95, 65. (2) Cassie, A. B. D. Discuss. Faraday Soc. 1948, 3, 11.

pure substrates. Wenzel recognized the contribution of surface roughness to wettability in 1936.3 He introduced eq 1

cos θrough ) r cos θ

(1)

where θrough is the “average” contact angle on a rough, chemically homogeneous surface of a given material, θ is the thermodynamic contact angle on a smooth surface of that material, and r is the roughness factor, r ) actual surface area/geometric surface area. Cassie and Baxter looked at wetting on porous substrates and derived eq 24

cos θporous ) f1 cos θ - f2

(2)

where θporous is the apparent contact angle, f1 is the fraction of fluid area in contact with the material, and f2 is the fraction of the fluid area in contact with air in the porous material. Johnson and Dettre5 predicted that a surface should transition from a nonporous (Wenzel regime) to a porous surface (Cassie regime) as roughness increases and that hysteresis (the difference between advancing and receding contact angles) should increase until this critical roughness is reached and then decrease. These simplified approaches, however, look only at the “average roughness” and fail to address the effects of different length scales of roughness, uniformity of roughness, and “density” of roughness. Superhydrophobic surfaces have been fabricated using various approaches.6-12 In one of the reports, the effects (3) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988. (4) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546. (5) Johnson, R. E., Jr.; Dettre, R. H. Adv. Chem. Ser. 1963, No. 43, 112. (6) O ¨ ner, D.; McCarthy, T. J. Langmuir 2000, 16, 7777.

10.1021/la049404l CCC: $27.50 © 2004 American Chemical Society Published on Web 08/10/2004

8132

Langmuir, Vol. 20, No. 19, 2004

of topography on hydrophobicity have recently been reported on surfaces prepared by photolithography with asperity length scales ranging from 2 to 128 µm; the maximum length scale of roughness that imparts “ultrahydrophobicity” is ∼32 µm according to the study.6 The effect of topography on hydrophilicity and the effect of roughness length scales less than 2 µm on wettability have not been reported. The availability of monodisperse particles with diameters ranging from 50 nm to 10 µm has allowed us to examine the effects of surface topography at this length scale on wettability. Wettability, hydrophilicity, hydrophobicity, lyophilicity, and lyophobicity are currently qualitative terms that cannot be specifically related to molecular structural features. A molecular basis for why surfaces are wet to varying extents remains unaddressed. It is well understood that incorporating perfluoroalkyl groups to surfaces makes them less wettable because the surface tension of substituent groups decreases in the order of CH2 (36 dyn/ cm) > CH3 (30 dyn/cm) > CF2 (23 dyn/cm) > CF3 (15 dyn/ cm).13 It is certainly well-known that introducing polar functionality to nonpolar surfaces makes them more wettable. The combination of concentration (surface density) and identity of chemical functionality required to impart a particular level of wettability to a given surface, however, is unknown. We measure dynamic wettability by assessing both advancing and receding contact angles. The importance of the three-phase contact line structure (length and continuity) and stability (how close in energy are static metastable states) has been stressed.8 Our objectives are to rationally control wettability of surfaces by individually and simultaneously manipulating surface topography and surface chemical structure and to help provide a fundamental understanding of the basis of wettability. Our strategy involves two stages of research: (1) the adsorption of charged polystyrene latex particles to oppositely charged poly(ethylene terephthalate) (PET) surfaces to form surfaces with different topographies/roughness; (2) the introduction of discrete functional groups to smooth and rough surfaces through organic transformations. The two research stages combine to form a method for preparing robust stable surfaces of variable wettability. The combination of surface topography and surface-chemical functionality allows us to control wettability over a wide range. Experimental Section General. Polybead carboxylate microspheres (0.1, 0.35, and 0.79 µm) in 2.5% solids-latex aqueous dispersions and poly(vinyl alcohol) (PVOH) (Mw ) 108 000, 99.7% hydrolyzed) were purchased from Polysciences, Inc. Poly(allylamine hydrochloride) (PAH) (Mw ) 50000-65000), poly(sodium styrene sulfonate) (PSS) (Mw ) 70 000), 1-octadecanethiol, 11-mercapto-1-undecanol, ethanol (anhydrous), and heptafluorobutyryl chloride (HFBC) were obtained from Aldrich. Gold shot (99.95%) was purchased from Electronic Space Products International. Sodium chloride, 1 M hydrochloric acid, sodium hydroxide, methanol (HPLC grade), and hexane (HPLC grade) were purchased from (7) Chen, W.; Fadeev, A. Y.; Hsieh, M. C.; O ¨ ner, D.; Youngblood, J.; McCarthy, T. J. Langmuir 1999, 15, 3395. (8) Youngblood, J. P.; McCarthy, T. J. Thin Solid Films 2001, 382, 95. (9) Erbil, H. Y.; Demirel, A. L.; Avci, Y.; Mert, O. Science 2003, 299, 1377. (10) Lau, K. K. S.; Bico, J.; Teo, K. B. K.; Chhowalla, M.; Amaratunga, G. A. J.; Milne, W. I.; McKinley, G. H.; Gleason, K. K. Nano Lett. 2003, 3, 1701. (11) Tadanaga, K.; Morinaga, J.; Matsuda, A.; Minami, T. Chem. Mater. 2000, 12, 590. (12) Feng, L.; Song, Y.; Zhai, J.; Liu, B.; Xu, J.; Jiang, L.; Zhu, D. Angew. Chem., Int. Ed. 2003, 42, 800. (13) Wang, J.; Ober, C. K. Macromolecules 1997, 30, 7560.

Takeshita et al. Fisher. All materials were used as received. Water was purified using a Millipore Milli-Q system that involves reverse osmosis followed by ion-exchange and filtration steps (18.2 MΩ cm). Solution pH for layer-by-layer adsorption studies was adjusted with either HCl or NaOH solution using a Fisher 825MP pH meter. Latex particle agglomeration during the adsorption process was prevented by using a Branson sonifier cell disruptor (model 185). Gold (∼120 Å) was evaporated onto smooth as well as patterned substrates at a pressure less than 10-5 Torr and deposition rate of ∼0.1 nm/s using a Thermionic VE-90 vacuum evaporator. Scanning electron micrographs were recorded on goldcoated samples with an ISI-DS 130 dual stage scanning electron microscope. Contact angle measurements were performed with a Rame´-Hart telescopic goniometer and a Gilmont syringe with a 24-gauge flat-tipped needle. The probe fluid used was water, purified as described above. Dynamic advancing (θA) and receding angles (θR) were recorded while the probe fluid was added to and withdrawn from the drop, respectively. The values reported are averages of three to five measurements made on different areas of the sample surface. X-ray photoelectron spectra (XPS) were recorded with a Perkin-Elmer-Physical Electronics 5100 with Mg KR excitation (15 kV, 400 W). Spectra were obtained at two takeoff angles, 15° and 75° (between the plane of the surface and the entrance lens of the detector optics). Each reported XPS datum is an average of at least two experimental data points. Substrate Preparation.14 PET film samples (Mylar, 5 mil) were rinsed with distilled water and methanol, extracted in refluxing hexane for 2 h, and then dried at reduced pressure. PET-NH3+ was prepared by immersing clean PET film in PAH solution (167 mg of PAH in 120 mL of water, pH ) 11.5) for 1 h at room temperature. The film was removed from the solution, rinsed with water three times, and introduced to water adjusted to pH ) 2.2 for 30 min. After being rinsed with three aliquots of water, the PET-NH3+ film samples were dried at reduced pressure. Four additional substrates bearing positive charges were prepared using a layer-by-layer adsorption technique: PETNH3+-[PSS(salt)-PAH(salt)]5 (PET-NH3+ with ten layers adsorbed, both polyelectrolyte solutions have 1 M NaCl), PET-NH3+[PSS-PAH(salt)]5 (only the PAH solution has 1 M NaCl), PETNH3+-[PSS-PAH]5 (no salt is added), and PET-NH3+-[PSS(salt)PAH]5 (only the PSS solution has 1 M NaCl). Adsorptions were carried out for 20 min at room temperature in open beakers containing unstirred polyelectrolyte solutions that were prepared fresh daily. After each layer deposition, film samples were rinsed with water three times. After deposition of 10 layers, films were dried at reduced pressure before the adsorption of latex particles. Adsorption of Carboxylated Polystyrene Beads. Dispersions for adsorption were prepared by adding three drops of latex solution and a desired amount of NaCl to 100 mL of water. The pH of the solutions was adjusted to the desired value. Film samples were immersed in the latex solution for the desired amount of time in an ice bath. Aggregation of the colloids was prevented by sonication of the solution every 10 min for 5 s over the first 2 h of adsorption. When adsorption of the latex particles was complete, the film samples were rinsed with water for 3 min (3 times) prior to drying at reduced pressure. PVOH Adsorption.15,16 Stock solutions of 0.1 M PVOH (based on repeat units) were prepared by heating (90-100 °C) the polymer-water suspension for ∼1 h and allowing the resulting solution to cool to room temperature. Solutions of lower concentration were prepared shortly thereafter by diluting the stock solutions. Film samples were submerged in 0.01 M PVOH solutions at room temperature for 24 h, removed, rinsed with copious amounts of water, and dried at reduced pressure. Reaction with HFBC. In a nitrogen-purged Schlenk flask, film samples were suspended and allowed to react with HFBC (0.5 mL) in the vapor phase for 24 h. After the reaction, the films were dried at reduced pressure. Adsorptions of 1-Octadecanethiol and 11-Mercapto-1undecanol. Gold-coated substrates were immersed in a freshly prepared solution of 0.01 M 1-octadecanethiol or 11-mercapto(14) Chen, W.; McCarthy, T. J. Macromolecules 1997, 30, 78. (15) Coupe, B.; Chen, W. Macromolecules 2001, 34, 1533. (16) Kozlov, M.; Quarmyne, M.; Chen, W.; McCarthy, T. J. Macromolecules 2003, 36, 6054.

Controlled Wettability

Langmuir, Vol. 20, No. 19, 2004 8133

1-undecanol in anhydrous ethanol. The adsorptions were carried out for 24 h at room temperature. The substrates were rinsed for 5 min in ethanol (3 times) before they were dried at reduced pressure.

Results and Discussion Topographic Control. Adsorption of Latex Particles onto PET Surfaces. The adsorption of charged latex particles onto oppositely charged surfaces is controlled by the attraction between the oppositely charged particles and the substrate and the repulsion among adjacent adsorbing particles. The irreversibility of adsorption, however, cannot be fully described by thermodynamic considerations, and kinetic control must also be taken into account. The random sequential adsorption model (a kinetic model) predicts a jamming limit of 54.7% surface coverage for monodisperse spheres.17 The adsorption rate and ultimate surface coverage should depend on particle size, solution ionic strength, solution pH, and the charge density of both the surface and the particles. Only the first two variables have been studied experimentally for particle adsorption.18,19 Since the substrate and latex particles have to be oppositely charged, adsorptions of negatively charged latex particles to positively charged substrates were carried out in this study. We have prepared surfaces with different topographies (particle size and distance between adjacent particles) and characterized them by scanning electron microscopy (SEM). The charge density of substrates, adsorption time, particle size, solution pH, solution ionic strength, and sequential adsorption of different size latex particles were the variables studied. PET-NH3+ was prepared by reacting PET with poly(allylamine). This is an autoinhibiting reaction that renders a layer of reactive amine groups that are covalently attached to the substrate by amide bonds.14 The surface charge density was varied through polyelectrolyte multilayer deposition. Alternating layers of poly(sodium styrenesulfonate) (PSS) and poly(allylamine hydrochloride) (PAH) were layer-by-layer assembled on PET-NH3+, and the resulting surface amine group density (and the thickness of individual layers) was controlled by the ionic strength of the solutions (added salt screens like-charge repulsions resulting in more dense and thicker layers).14 Five positively charged substrates were prepared: PETNH3+, PET-NH3+-[PSS-PAH(salt)]5, PET-NH3+-[PSS(salt)-PAH(salt)]5, PET-NH3+-[PSS-PAH]5, and PETNH3+-[PSS(salt)-PAH]5. Because the addition of salt results in thicker layers with more charges and a thick PSS layer should neutralize a large portion of positive charges in the outermost PAH layer, we expect that surface positive charge excess decreases in the order of PET-NH3+[PSS-PAH(salt)]5 > PET-NH3+-[PSS(salt)-PAH(salt)]5 > PET-NH3+-[PSS-PAH]5, > PET-NH3+-[PSS(salt)-PAH]5. PET-NH3+ should fall somewhere in this series. This simplified model is based on the assumption that only the outermost layer and its underlying layer control the substrate charge density. Substrate charge density affects both the rate of latex adsorption and the ultimate surface coverage. Figure 1 illustrates surface coverage (calculated based on SEM images) as a function of adsorption time for 0.1 µm latex particles at pH 8.8 (the substrate and the latex particles (17) Oberholzer, M. R.; Stankovich, J. M.; Carnie, S. L.; Chan, D. Y. C.; Lenhoff, A. M. J. Colloid Interface Sci. 1997, 194, 138. (18) Hayes, R. A.; Bo¨hmer, M. R.; Fokkink, L. G. J. Langmuir 1999, 15, 2865. (19) Krozer, A.; Nordin, S.-A.; Kasemo, B. J. Colloid Interface Sci. 1995, 176, 479.

Figure 1. Adsorption kinetics of 0.1 µm carboxylated polystyrene particles at pH 8.8 to various substrates: PET-NH3+ (O), PET-NH3+-[PSS-PAH(salt)]5 (0), PET-NH3+-[PSS(salt)PAH(salt)]5 (4), PET-NH3+-[PSS-PAH]5 (+), and PET-NH3+[PSS(salt)-PAH]5 (1).

are oppositely charged) on all five substrates: the surface coverage reached ∼68% (a theoretical maximum coverage of 78.5% is calculated for a surface covered with close packed arrays of spheres) on PET-NH3+, PET-NH3+-[PSSPAH(salt)]5 and PET-NH3+-[PSS(salt)-PAH(salt)]5 at adsorption times g8 h. It took g24 h for the surface coverage to reach 64% on PET-NH3+-[PSS-PAH]5 and the surface coverage reached only 37% on PET-NH3+-[PSS(salt)PAH]5 after 48 h. The jamming limit of 54.7% surface coverage for monodisperse spheres as predicted by the random sequential adsorption model17 apparently does not agree with our experimental observations. These data also indicate that the charge densities of the first three substrates are high and indistinguishable based on latex particle adsorption amount; PET-NH3+-[PSS-PAH]5 has intermediate charge density, and the charge density is the lowest on PET-NH3+-[PSS(salt)-PAH]5. The adsorption of 0.35 µm particles on various substrates follows a similar trend: ∼35% surface coverage was reached on PET-NH3+, PET-NH3+-[PSS-PAH(salt)]5, and PET-NH3+-[PSS(salt)PAH(salt)]5 at adsorption times g120 h, no adsorption occurred on PET-NH3+-[PSS(salt)-PAH]5, and intermediate coverage was observed on PET-NH3+-[PSS-PAH]5. The experimental observations agree with the theoretical prediction except that the three substrates with the highest charge densities cannot be distinguished using 0.1 and 0.35 µm particles. Due to the high charge density and ease of preparation, PET-NH3+ was the substrate used for further adsorption experiments unless specified otherwise. Particle size also affects both adsorption rate and surface coverage. The adsorption of 0.1 µm particles at pH 8.8 on PET-NH3+ reached maximum surface coverage of 68% after 8 h of adsorption time. With the same substrate and the same conditions, the adsorption of 0.35 and 0.79 µm particles resulted in surface coverage of 34% after 120 h and 32% after 42 days, respectively. With particles of the same charge density, larger particles experience greater repulsion among adjacent adsorbing particles. From both thermodynamic and kinetic considerations, adsorption of smaller particles is faster and results in higher surface coverage. That repulsion among large particles results in very low surface coverage puts restrictions on our study of surfaces with varying topographies in terms of length scale as well as “density”. One of the important challenges in this study is to determine which variables, in addition to surface charge density and adsorption time for a given

8134

Langmuir, Vol. 20, No. 19, 2004

Takeshita et al.

Figure 3. SEM micrographs of 0.35 µm carboxylated polystyrene latex particles adsorbed onto PET-NH3+ at pH 5.0 for 120 h followed by adsorption of 0.10 µm particles at pH 8.8 (left) and pH 5.0 (right) for 24 h. (The 0.35 µm particles indicate the length scale.)

Figure 2. SEM micrographs of adsorbed 0.35 µm carboxylated latex particles on PET-NH3+ (120 h at pH 2.2 (top left), pH 5 (top right), and pH 8.8 (bottom)). (The 0.35 µm particles indicate the length scale.)

particle size, can be tuned to control surface coverage. If we can reduce the repulsion among like-charged particles, we should be able to increase surface coverage. The repulsion among charged particles can be screened by adding salt to the latex dispersion and by changing the pH of the latex solution. The addition of salt only caused particle aggregation (not shown) and did not improve adsorption in the NaCl concentration range studied (0.001-0.05 M). Carboxylated latex particles (0.35 µm) were adsorbed on PET-NH3+ for 120 h at pH 2.2 (particles are neutral), pH 5 (particles are partially charged), and pH 8.8 (particles are fully charged). SEM micrographs of these surfaces are shown in Figure 2. The aggregation of particles on the surface at low pH values is due to the absence of likecharge repulsion among adjacent particles which stabilizes them as individual particles. The adsorbed amount increased from 34% to 48% when the pH was decreased from 8.8 to 5.0. At pH 5.0, when the latex particles are partially charged, there is sufficient electrostatic attraction between the substrate and the particles and also enough repulsion among adjacent particles to cause discrete particle adsorption; fewer like-charge repulsions among particles at pH 5 allows particles to adsorb closer to one another than at pH 8.8. Maximum surface coverage of 0.1 µm particles on PET-NH3+-[PSS(salt)-PAH]5 also increased from 34% at pH 8.8 to 66% at pH 5.25. Reducing particle charge density by changing the latex dispersion pH decreases electrostatic attraction between the substrate and the particles as well as repulsion among particles. At the highest particle charge density, repulsion dominates and results in low surface coverage; at the other end of the spectrum, aggregation occurs. After samples were prepared with various surface topographies with monodisperse spheres, sequential adsorption using particles of different size was carried out. The stability of the adsorbed latex particles is important in the sequential adsorption study as well as in further chemical manipulation. Stability studies of surfaces adsorbed with latex particles were carried out in water at different pH values (2.2, 5.0, and 8.8) for extended periods of time (e.g., 1 week) at room temperature. While the adsorbed 0.1 µm particles are stable under all conditions tested, the results are mixed with the adsorbed 0.35 µm particles. The surface coverage changed from 48% to 30%, 46%, and 33% after PET-NH3+ samples with 0.35 µm particles (pH 5.0, 120 h) were exposed to aqueous solution at pH values of 2.2, 5.0, and 8.8 for 1 week,

respectively. This indicates that the adsorbed latex particles are not stable on the substrate at pH 2.2 and 8.8. At pH 2.2, there is no electrostatic attraction between the neutral particles and the substrate. That particle aggregation was not observed indicates that desorption is the dominant mechanism and readsorption does not occur extensively. At pH 8.8, the particles are fully charged and a drop in coverage is observed due to the strong repulsive forces between adjacent particles adsorbed that overcomes the attractive force between the adsorbed particles and the substrate. The particles appear to be stable at pH 5.0; no decrease in surface coverage was observed. The particles are partially charged and apparently a balance between attraction and repulsion is reached. This is in agreement with the fact that the adsorption of 0.35 µm particles at pH 5.0 gave the highest surface coverage. In sequential adsorption experiments, the adsorption of 0.35 µm particles at pH 5.0 for 120 h was followed by the adsorption of 0.1 µm particles for 24 h at both pH 5.0 or 8.8. The SEM micrographs of the resulting surfaces are shown in Figure 3. After the subsequent adsorption of 0.1 µm particles at pH 5.0, the surface coverage of 0.35 µm particles decreased from 48% to 29% and that of 0.1 µm particles was 16%; when the subsequent adsorption of 0.1 µm particles was carried out at pH 8.8, the coverage of 0.35 µm particles decreased slightly to 45% and that of 0.1 µm particles was only 2%. Apparently, the total surface coverage of the latex particles remains the same; however, the relative percentage of 0.35 and 0.1 µm particles adsorbed varies greatly depending on the solution pH of the 0.1 µm particles. When the 0.1 µm particles were subsequently adsorbed at pH 5.0, they replaced a large percentage of 0.35 µm particles. Because the adsorbed 0.35 µm particles are shown to be stable against spontaneous desorption at this pH, the 0.1 µm particles must have a stronger affinity for the substrate due to their smaller size. At pH 8.8, even though 0.35 µm particles are shown to desorb spontaneously after 1 week, the adsorbed 0.35 µm particles appear to be stable during the 24 h adsorption of 0.1 µm particles. It is likely that the desorption of 0.35 µm particles and the “replacement” by 0.1 µm particles occur much more slowly at pH 8.8 giving a surface topography after 24 h that is still very close to the initial structure. Even though the total surface coverage remains the same, the relative distribution of the two types of particles was controlled by the pH of the latex dispersion and adsorption time in the second step. By use of the sequential adsorption approach, interesting surface topographies with bimodal distributions of particles were prepared. This type of surface may exhibit interesting wetting characteristicsssmall particles situated between larger particles may be effective at preventing the intrusion of liquid between the larger particles. Chemical Structure Control. Introduction of Discrete Functional Groups. Surface chemistry, as well

Controlled Wettability

Langmuir, Vol. 20, No. 19, 2004 8135

Table 1. Water Contact Angles (θA/θR) of Smooth and Patterned PET Surfaces after Condensation of Gold and Subsequent Chemical Transformations substrate

Au-octadecyl

Au

Au-1-undecanol

Au-PVOH

PET PET-NH3+-0.1 µm (pH 8.8)a PET-NH3+-0.35 µm (pH 8.8)b PET-NH3+-0.35 µm (pH 5)c PET-NH3+-0.35 µm-0.1 µmd

110°/90° 135°/86° 146°/114° 145°/110° 150°/126°

78°/20° 90°/15° 93°/13° 98°/17° 109°/10°

53°/28° 55°/8° 68°/11° 56°/7° 40°/8°

43°/11° 35°/13° 17°/6° 18°/7° 18°/8°

a Amidated PET was treated with a 0.1 µm aqueous latex dispersion at pH 8.8 for 24 h; surface coverage is 68%. b Amidated PET was treated with a 0.35 µm aqueous latex dispersion at pH 8.8 for 120 h; surface coverage is 34%. c Amidated PET was treated with a 0.35 µm aqueous latex dispersion at pH 5.0 for 120 h; surface coverage is 48%. d Amidated PET was treated with a 0.35 µm aqueous latex dispersion at pH 5.0 for 120 h followed by treatment with a 0.1 µm aqueous latex dispersion at pH 5.0 for 24 h; surface coverage of 0.35 µm is 29% and that of 0.1 µm is 16%.

as surface topography, has an impact on wettability. Our study of the effect of chemical structure on wettability involved the attachment of functionalities to surfaces with different topographies prepared by latex particle adsorption. We tried three approaches: (1) Adsorption/reaction of poly(allylamine) (PAA) to the PET/latex particle surfaces. PAA reacts with PET and should adsorb under the right conditions (at low enough pH so it is cationic) to negatively charged latex particles. The amine functionality could be derivatized to control surface chemistry/structure. (2) Adsorption of poly(vinyl alcohol) to the PET/latex particle surfaces by hydrophobic interaction and crystallinity.15,16 The resulting alcohol functionality should be versatile in controlling surface chemistry/structure. (3) Evaporation of gold onto the PET/latex particle surface and subsequent adsorption of thiols or poly(vinyl alcohol). The first two approaches did not meet the criteria necessary for our goals, but the third was successful. We show selected data for each of these approaches, explaining why the first two failed and why the third was successful. Attempts to react amine groups on the surfaces of PET/ latex particle substrates resulted in low reaction yields. XPS analysis of PAA-treated PET indicated that an atomic composition of 7% nitrogen was present in the outermost ∼10 Å (15° takeoff angle) of the sample. A significant amount of free amines (potentially reactive nucleophiles), in addition to amides is present. A high reaction yield of this surface with heptafluorobutyryl chloride (HFBC) should render a hydrophobic surface with a high F:N ratio observed by XPS. The results were that fluorine content was only ∼10% (15° takeoff angle) and water contact angles remained low (θA/θR ) 85°/20°). The presence of nonnucleophilic -NH3+ is a likely cause of the low yield, but adding a catalyst/base (pyridine) or rinsing the surface with a base prior to HFBC exposure did not produce more hydrophobic or higher fluorine content surfaces. Adsorption of poly(vinyl alcohol) (PVOH) to introduce nucleophilic -OH groups was another approach. A PET film sample exposed to 0.01 M PVOH for 24 h at room temperature exhibited water contact angles of θA/θR ) 40°/14° (PET exhibits 78°/55°). XPS data indicated that the PET film is covered by a several angstrom thick PVOH monolayer.15,16 Crystallization of PVOH on PET is suspected to be the dominant driving force for the adsorption. Hydrogen bonding interaction between the substrate and PVOH provides an additional driving force.16 PVOH adsorption on a smooth PET surface thus renders a significant amount of exposed -OH groups. Reaction of PVOH adsorbed PET with HFBC results in a surface that has higher fluorine content, 17% obtained by XPS at 15° takeoff angle, and higher water contact angles, 98°/35°, than the amidated PET. PVOH was also adsorbed on surfaces prepared by latex particle adsorption. SEM micrographs of these surfaces after PVOH adsorption (not

shown) indicated latex particle aggregation likely due to desorption and readsorption of particles in PVOH solutions. Chemical transformations of PET-NH2 were not very successful likely because of the presence of nonnucleophilic -NH3+ groups and a low concentration of -NH2 groups. On the other hand, PVOH adsorption introduces a large amount of -OH groups but alters the underlying surface topographies through particle desorption/readsorption. To preserve the surface topography prepared by latex particle adsorption, a ∼120 Å thick gold layer was condensed on surfaces prior to subsequent chemical transformations. Two types of reactions were carried out on gold surfaces: adsorption of 1-octadecanethiol and 11-mercapto-1-undecanol, and adsorption of PVOH. Self-assembly of thiols on gold surfaces to form close packed monolayers is well documented in the literature.20 Adsorptions of 1-octadecanethiol and 11-mercapto-1-undecanol result in monolayers of close packed octadecyl and undecanol chains. Five substrates with different topographies, smooth PET, PET-NH3+-0.1 µm (pH 8.8), PET-NH3+-0.35 µm (pH 8.8), PET-NH3+-0.35 µm (pH 5), and PET-NH3+-0.35 µm (pH 5)-0.1 µm (pH 5), were prepared for chemical transformations. Water contact angle data are given in Table 1. The high water contact angles on gold are due to the extensive adsorption of contaminants from air to the metal surface to lower its surface energy. Water contact angle hysteresis (θA - θR) on gold-coated rough surfaces is high probably due to the combination of the inhomogeneous distribution of contaminants and surface roughness. Water contact angles on 1-undecanol surfaces are lower since these surfaces contain both hydrophobic alkyl and hydrophilic -OH groups. The two most interesting sets of data are given in the left-most and right-most columns in the table suggesting very hydrophobic (high θA and θR) and very hydrophilic (low θA and θR) surfaces, respectively. A monolayer of hydrophobic octadecyl chains on flat PET gives rise to water contact angles of 110°/ 90°. The same chemical functionality on four rough PET surfaces resulted in both higher advancing and receding contact angles. Both higher advancing and higher receding contact angles (a result of surface roughness) indicate surfaces described by Cassie.5 This type of surface has been termed “ultrahydrophobic” because a drop of water slides readily off of it.7 The contact angles of this surface are not as high as others reported and we hesitate to label it “ultrahydrophobic”, but the effect is apparent. PETNH3+-0.1 µm (pH 8.8) with 68% coverage is the least hydrophobic. PET-NH3+-0.35 µm (pH 8.8) with 34% coverage and PET-NH3+-0.35 µm (pH 5) with 48% coverage exhibit intermediate hydrophobicity. The surface adsorbed with bimodal particles is the most hydrophobic in terms (20) Bain, C. D.; Troughton, E. B.; Tao, Y. T.; Evall, J.; Whitesides, G. M.; Nuzzo, R. G. J. Am. Chem. Soc. 1989, 111, 321.

8136

Langmuir, Vol. 20, No. 19, 2004

Figure 4. Pictorial representation of a surface adsorbed with latex particles. The darker lines describe possible three-phase (air-liquid-surface) contact lines for a drop of water in contact with these surfaces: (a) a surface adsorbed with monodisperse particles; (b) surfaces with small and large particles, crosssectional view; (c) a surface containing bimodal particles.

of high water contact angles and low hysteresis. Advancing and receding contact angles are dictated by the shape and continuity of the three-phase (air-liquid-surface) contact line for a drop of water on the surface, which is affected by the topology of surface roughness.8 The contact line on surfaces containing latex particles is in contact with the uppermost section of the spheres (Figure 4a). On these composite surfaces where the contact line is in contact with both the solid and air, energy barriers between metastable states are low.5,8 High water contact angles and low hysteresis are observed on such surfaces. That surfaces containing 0.1 µm particles are not as hydrophobic as surfaces with 0.35 µm spheres is likely because the contact line on surfaces containing smaller particles is less “elevated” from the underlying surfaces (Figure 4b) and it intrudes into the cavities between particles and wets the underlying surface more readily when the cavity size between adjacent particles is similar. The two

Takeshita et al.

octadecyl-functionalized surfaces containing 0.35 µm particles have similar contact angles. This implies that the contact line does not intrude into the cavities between particles at either surface coverage (34% and 48%). The length scale of roughness appears to be more important than overall roughness in determining hydrophobicity. Hydrophobicity, however, could not be improved further by increasing the length scale of roughness since particle density decreases dramatically as electrostatic repulsion between particles increases as particles become larger. At low enough particle density, contact lines can readily intrude between particles to wet underlying surfaces. The advancing and receding water contact angles are the highest on the surface containing both 0.1 and 0.35 µm particles. The contact line is more tortuous on surfaces with bimodal particles (Figure 4c), thus the energy barriers between metastable states should be lower. That the contact line can “hop” from one metastable state to the other readily gives rise to high advancing and receding contact angles. Roughness on low contact angle surfaces leads to increased hydrophilicity (eq 1). Adsorption of PVOH on gold-coated flat PET gives surfaces that exhibit water contact angles of 43°/11°(θA/ θR). The same -OH functionality on four rough PET surfaces results in both lower advancing and receding contact angles (Table 1). PETNH3+-0.1 µm (pH 8.8) is the least hydrophilic while the water contact angles on the other three rough surfaces were very difficult to measure since water droplets spread readily as soon as they contact the surfaces. On latex particle adsorbed surfaces, large particle size and biomodal particle distribution improve both hydrophobicity and hydrophilicity; the effect of particle density on wettability is not as pronounced in the length scales studied. Acknowledgment. We thank the National Science Foundation (RUI, DMR-0209282 and RSEC, CHE0113643) and the ACS Petroleum Research Fund for financial support. We acknowledge the use of the NSFMRSEC central facilities at the University of Massachusetts. L. A. Paradis would like to express her gratitude to Solutia, Inc. LA049404L