Anal. Chem. 2007, 79, 7104-7109
Stability of Virtual Air Walls on Micropallet Arrays Yuli Wang,†,§ Mark Bachman,†,‡ Christopher E. Sims,§,# G. P. Li,†,‡ and Nancy L. Allbritton*,†,§,#
Integrated Nanosystems Research Facility, Department of Electrical and Computer Engineering, and Department of Physiology and Biophysics, University of California, Irvine, California 92697
Arrays of micropallets have been used to pattern adherent cells as well to sort mixtures of cells. These artificial surfaces are composed of micrometer-sized, SU-8 structures formed on a hydrophobic glass surface. Successful application of these arrays requires stable Cassie-Baxter wetting by aqueous biological solutions. This paper systematically studies the factors governing the stability of Cassie-Baxter wetting on the arrays, including the surface properties of the array components as well as the physical and chemical properties of the wetting solutions. To establish stable Cassie-Baxter wetting with water, the surface of the array must be coated with a perfluoroalkylsilane of sufficient hydrophobicity, and the roughness of array must be greater than 1.6. Additionally, long-term stability of the Cassie-Baxter wetting depends on the properties of the wetting solutions, including the surface tension (>40 mM/m), salt concentration (>10 mM), and protein concentration (40 mM/m). Most biological solutions possess a surface tension similar to that of water; however, cells are often incubated with low concentrations of surfactants to assist in the dispersion and uptake of hydrophobic reagents. For example, many of calcium indicators are loaded into (29) Nakae, H.; Inui, R.; Hirata, Y.; Saito, H. Acta Mater. 1998, 46, 2313-2318. (30) Yoshimitsu, Z.; Nakajima, A.; Waanabe, T.; Hashimoto, K. Langmuir 2002, 18, 5818-5822. (31) Patankar, N. A. Langmuir 2004, 20, 7097-7102. (32) Lide, D. R. CRC Handbook of Chemistry and Physics, 73rd ed.; CRC Press: Boca Raton, FL, 1992. (33) Sung, J.; Park, K.; Kim, D. J. Phys. Chem. B 2005, 109, 18507-18514. (34) Flaming, J. E.; Knox, R. C.; Sabatini, D. A.; Kibbey, T. C. Vadose Zone J. 2003, 2, 168-176. (35) Alexander, V. K.; Irina, R. N.; Irina, V. A.; Elena, V. B.; Valery, Y. A.; Alexander, A. Y.; Victor, A. K. Macromolecules 1995, 28, 2303-2314.
Analytical Chemistry, Vol. 79, No. 18, September 15, 2007
7107
Figure 4. Influence of salt concentration on the long-term stability of Cassie-Baxter wetting on a 55 µm × 50 µm × 50 µm (h × b × a) pallet array. (A) Side view of the pallet array when wetted with water. Psolution and Pdroplet are the vapor pressures of the wetting solution and water droplet, respectively. (B) Side view of the pallet array when wetted with water containing salt. P′solution is the vapor pressure of the wetting solution with dissolved salt. (C) Images of pallet arrays immersed in 0, 1, 10, or 100 mM NaCl for 24, 48, or 192 h. (D) Closeup view of the upper, left corner of the pallet immersed in 0 mM NaCl for 24 h. The “/” overlies a region of trapped air while the arrow points to a water droplet covering the glass surface between two pallets. The “#” marks a pallet. (E) Close-up view of the upper, left corner of the pallet immersed in 1 mM NaCl for 24 h.
cells in the presence of Pluronic F-127 detergent.36 Occasionally, small amounts of lipids or low surface tension liquids (e.g., N,N′dimethylformamide, dimethyl sulfoxide) are also added to biological solutions. In all of these cases, the virtual walls should remain stable since the surface tension does not decrease to less than 40 mM/m. The use of some ethanol-based fixatives or wash solutions, however, may dissipate the virtual walls if the surface tension is sufficiently low. Influence of Salt Concentration on the Stability of CassieBaxter Wetting. Salt is a component of virtually all biologic solutions with the concentration typically between 10 and 500 mM. Thus, it was important to understand how biologic salt solutions impacted the wetting properties of the pallet arrays. When water without salt was added to a perfluoroalkylsilane-coated pallet array with an R of 2.1 (a ) b ) 50 µm, h ) 55 µm), Cassie-Baxter wetting formed instantaneously. However, within a few minutes, tiny water droplets with diameters of 1-5 µm condensed on the hydrophobic glass surface (Figure 4A).3 This was likely due to saturation of the trapped air with water vapor and nucleation of water droplets on defects in the hydrophobic coating on the glass. Over time, the water droplets continued to grow in size on the glass surface. By 30 min, the water droplets coalesced to form larger water droplets and the Cassie-Baxter wetting began to deteriorate. After 24 h, the water droplets appeared to coalesce and form connections with the water overlying the pallets (Figure 4C). Thus, the side walls of many of the pallets were in contact with water. In addition, the air on the pallet area was no longer (36) Kao, J. P. Y. In Methods in Cell Biology; Nuccitelli, R., Ed.; Academic Press: New York, 1994; Vol. 40, p 155.
7108 Analytical Chemistry, Vol. 79, No. 18, September 15, 2007
contiguous but divided into discrete, segregated air bubbles (Figure 4D). After 48 h, many of the segregated air bubbles between the pallets were lost having been replaced by water. By 192 h, most of the array possessed Wenzel wetting although a few isolated air bubbles remained. Thus, Cassie-Baxter wetting by water was not stable over long periods of time even for arrays with R > 1.6. For two solutions separated by an air barrier, water will move spontaneously from the solution of high vapor pressure to the solution of lower vapor pressure. Addition of salt to the wetting solution will lower the vapor pressure of this solution (Raoult’s Law).37 However, the vapor pressure of the droplets beneath the trapped air should remain unchanged, i.e., that of pure water (Figure 4B). The lower vapor pressure in the wetting solution relative to the water droplets should act to limit the growth of the droplets and stabilize Cassie-Baxter wetting of the array. To test this hypothesis, pallet arrays were immersed in wetting solutions with varying concentrations of sodium chloride (Figure 4C). At a salt concentration of 1 mM, Cassie-Baxter wetting was present; however, large water droplets were present on the glass surface beneath the trapped air (Figure 4E). These droplets increased in size over time, and by 192 h much of the interpallet area was wetted by water. Addition of 10 mM salt led to stable Cassie-Baxter wetting for as long as 192 h. Small droplets were present on the glass between the pallets but they did not grow in size over time. Sodium chloride at 10 mM decreases the vapor pressure of water by only 99.98% yet significantly increases the stability of the virtual air walls. When NaCl was added at 100 or 1000 mM, water droplets forming on the glass beneath the trapped air were fewer in number and smaller in size than those forming at lower salt concentrations. These arrays possessed stable Cassie-Baxter wetting as long as 1 month after immersion in the salt solutions. Most biologic buffers possess substantial concentrations of salt. For example, phosphate-buffered saline (PBS) which is commonly used with cells contains sodium chloride (138 mM), potassium chloride (2.6 mM), and potassium phosphate (10 mM). Thus salt-containing, biologic buffers enhance the stability of the virtual walls. Effect of Protein on Wetting of the Pallet Arrays. Proteins are a frequent additive to biological solutions. For example, histone or bovine serum albumin (BSA) is often added as a carrier to deliver lipids to cells.38 Other proteins such as growth factors, cytokines, and antibodies are often used to modify cell behavior or label receptors on the surface of cells. Cells also secrete a protein-rich extracellular matrix as they attach to and grow on a surface. Since proteins can adhere to surfaces modifying the surface properties as well as altering the surface tension of liquids, it was important to assess virtual wall formation in the presence of protein-containing solutions. Three test proteins, BSA, lysozyme (chicken egg), and fibrinogen (bovine plasma) were used to assess the effects of proteins on virtual wall formation. The proteins were dissolved in PBS at different concentrations, and the solution was added to the pallet array. The arrays were then examined over time. For all three proteins at all concentrations (e60 mg/mL) tested, Cassie-Baxter wetting was stable in the first few hours. (37) Reid, R. C.; Prausnitz, J. M.; Poling, B. E. The Properties of Gases and Liquids, 4th ed.; McGraw-Hill: New York, 1987. (38) Weiner, O. P.; Neilsen, P. O.; Prestwich, G. D.; Kirschner, M. W.; Cantley, L. C.; Bourne, H. R. Nat. Cell Biol. 2002, 4, 509-513.
After 16 h, Cassie-Baxter wetting became unstable in fibrinogen solutions at high concentrations (>20 mg/mL). However, at 16 h, the virtual walls remained intact for all concentrations of BSA and lysozyme. After 72 h, the virtual walls on arrays submerged in e5 mg/mL BSA and lysozyme remained stable (Figure 3B). Most applications of these proteins in cell biology utilize the proteins at e5 mg/mL so that the virtual walls should be stable. Only very low concentrations (e0.1 mg/mL) of fibrinogen yielded stable walls at 72 h. Fibrinogen efficiently adsorbs to surfaces forming a protein polymer.39 For this reason it is frequently used to coat surfaces for cell attachment. However, this efficient adsorption to surfaces including SU-8 results in a hydrophilic coating on the surface. In these experiments, it is likely that the fibrinogen adsorbed to the edges of the pallet and progressively infiltrated down the sides of the pallets. The coating of the hydrophobic perfluorocarbon layer by a hydrophilic protein then leads to loss the trapped air. Other investigators have demonstrated that fibrinogen adsorbs to hydrophobic surfaces much more rapidly than BSA or lysozyme supporting this conclusion.40 Impact of Cell Culture Media on Cassie-Baxter Wetting. A major usage of the pallet arrays is expected to be the sorting of cells while they remain adherent to a surface. Consequently, cells will likely be cultured for several weeks on the pallet arrays. Culture conditions can be harsh on surfaces since the solutions used contain significant quantities of both lipids and proteins and are maintained at 37 °C. The stability of Cassie-Baxter wetting of a pallet array (55, 50, 50 µm for h, b, a, respectively) was measured under typical cell culture conditions. The array was immersed in Dulbecco’s minimal essential media supplemented with FBS (10%), L-glutamine (584 mg/L), penicillin (100 units/ mL), and streptomycin (100 g/mL). Remarkably Cassie-Baxter wetting was present on the array for a month at 37 °C in the humidified 5% CO2 atmosphere of the cell incubator. While most cells are grown in 10% fetal calf serum, some cell types require higher concentrations for proper growth. For this reason the array was also immersed in tissue culture media composed of 50% fetal calf serum and then examined for Cassie-Baxter wetting over time. The virtual walls on the array were stable for up to 72 h despite the very high concentration of serum.
CONCLUSIONS
(39) Fuss, C.; Palmaz, J. C.; Sprague, E. A. J. Vasc. Interv. Radiol. 2001, 12, 677-682. (40) Wertz, C. F.; Santore, M. M. Langmuir 1999, 15, 8884-8894.
Received for review May 3, 2007. Accepted July 17, 2007.
We have systematically studied the Cassie-Baxter wetting of arrays of micropallets with a particular focus on the effects of the wetting solutions used in bioanalytical assays. To ensure stable Cassie-Baxter wetting, the interpallet regions of the array must be sufficiently hydrophobic and the array must have a roughness of greater than 1.6. The properties of the wetting solution also played a role in the wetting properties of the array. Optimal wetting solutions possessed a salt concentration greater than 10 mM and a protein concentration less than 5 mg/mL. Most biologic buffers have high salt concentrations which enhances the stability of the trapped air. Solutions with modified surface tension such as ethanol-based fixatives or detergent-containing buffers are also compatible within limits. Remarkably very high concentrations (over 5 mg/mL) of some proteins are compatible with the arrays. Fetal calf serum which at low concentrations (10%) is used for cell growth and at high concentrations (50%) is employed as a stabilizer for cell freezing supported Cassie-Baxter wetting of the pallet arrays. Even when placed at 37 °C in conventional tissue culture medium, the virtual walls of air remained stable for up to a month. Thus, the arrays are excellent platforms for the culture of mammalian cells and will find utility in both the patterning and sorting of cells. Characterization of the wetting stability of the arrays will likely broaden their utility to a broader range of bioanalytical applications, for example, the creation of threedimensional constructs and microfluidic air-based valves. ACKNOWLEDGMENT This research was supported by Grants from the NIH (EB004597 and EB007612). SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
AC070911S
Analytical Chemistry, Vol. 79, No. 18, September 15, 2007
7109
