Langmuir 2004, 20, 5569-5574
5569
Tailoring the Surface Properties of Poly(dimethylsiloxane) Microfluidic Devices Shuwen Hu,† Xueqin Ren,† Mark Bachman,‡,§ Christopher E. Sims,| G. P. Li,*,†,§,‡ and Nancy L. Allbritton*,†,| Center for Biomedical Engineering, Integrated Nanosystems Research Facility, Department of Electrical and Computer Engineering, and Department of Physiology and Biophysics, University of California, Irvine, California 92697 Received January 3, 2004. In Final Form: April 18, 2004 Poly(dimethylsiloxane) (PDMS) is an attractive material for microelectrophoretic applications because of its ease of fabrication, low cost, and optical transparency. However, its use remains limited compared to that of glass. A major reason is the difficulty of tailoring the surface properties of PDMS. We demonstrate UV grafting of co-mixed monomers to customize the surface properties of PDMS microfluidic channels in a simple one-step process. By co-mixing a neutral monomer with a charged monomer in different ratios, properties between those of the neutral monomer and those of the charged monomer could be selected. Mixtures of four different neutral monomers and two different charged monomers were grafted onto PDMS surfaces. Functional microchannels were fabricated from PDMS halves grafted with each of the different mixtures. By varying the concentration of the charged monomer, microchannels with electrophoretic mobilities between +4 × 10-4 cm2/(V s) and -2 × 10-4 cm2/(V s) were attainable. In addition, both the contact angle of the coated surfaces and the electrophoretic mobility of the coated microchannels were stable over time and upon exposure to air. By carefully selecting mixtures of monomers with the appropriate properties, it may be possible to tailor the surface of PDMS for a large number of different applications.
Introduction In the decade since the first reports of electrophoretic separations on microfluidic devices, tremendous research effort has gone into characterizing surface and electroosmotic properties of the silica-based materials most commonly used for their manufacture.1-4 As a result of this work, the performance of these glass devices is highly predictable and commercial glass products are now available.5 Interest in the use of polymeric materials such as poly(dimethylsiloxane) (PDMS) and poly(methyl methacrylate) has increased over the past few years because of the potential for easily fabricated, low-cost, disposable microfluidic devices.6-10 PDMS is particularly attractive * Corresponding authors. N.L.A.: phone 949-824-6493, e-mail
[email protected], fax 949-824-8540. G.P.L.: phone 949-824-4194, e-mail
[email protected], fax 949-824-3732. † Center for Biomedical Engineering, University of California. ‡ Integrated Nanosystems Research Facility, University of California. § Department of Electrical and Computer Engineering, University of California. | Department of Physiology and Biophysics, University of California. (1) Dolnik, V.; Liu, S.; Jovanovich, S. Electrophoresis 2000, 21, 4154. (2) Bousse, L.; Cohen, C.; Nikiforov, T.; Chow, A.; Kopf-Still, A. R.; Dubrow, R.; Parce, J. W. Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 155-181. (3) Regnier, F. E.; He, B.; Lin, S.; Busse, J. TIBTECH 1999, 17, 101-106. (4) Verpoorte, E. Electrophoresis 2002, 23, 677-712. (5) Bruin, G. J. M. Electrophoresis 2000, 21, 3931-3951. (6) Becker, H.; Gartner, C. Electrophoresis 2000, 21, 12-26. (7) McDonald, J. C.; Duffy, D. C.; Anderson, J. R.; Chiu, D. T.; Wu, H.; Schueller, O. J. A.; Whitesides, G. M. Electrophoresis 2000, 21, 27-40. (8) Whitesides, G. M.; Ostuni, E.; Takayama, S.; Jiang, X.; Ingber, D. E. Annu. Rev. Biomed. Eng. 2001, 3, 335-373. (9) Sia, S. K.; Whitesides, G. M. Electrophoresis 2003, 24, 35633576. (10) Makamba, H.; Kim, J. H.; Lim, K.; Park, N.; Hahn, J. H. Electrophoresis 2003, 24, 3607-3619.
as a substrate for these devices for a variety of reasons. Once a master has been produced, multiple devices can be cast in PDMS quickly and easily.7 This ease of fabrication translates into a low cost and makes such devices available to a large number of investigators. PDMS is durable and can readily seal either reversibly or irreversibly to a variety of substrates including PDMS itself, other polymers, and glass.11 Such attributes make PDMS ideally suited for fabricating topologically complex systems. Another important characteristic of PDMS is its ability to generate strong electroosmotic flow (EOF) in buffer-filled channels which can be utilized in electrokinetic pumping and electrophoretic separations.11,12 Additionally, its optical transparency at wavelengths >280 nm makes PDMS amenable to the use of laser-induced fluorescence detection, permitting sensitive fluorescence detection.13-15 Despite its versatility, a number of characteristics have limited the use of PDMS in the fabrication of microfluidic devices. These limitations are pertinent to the biological analyses for which these devices are predicted to be of most utility: clinical and pharmaceutical assays.4 Foremost among PDMS’s disadvantages is its extreme hydrophobicity.7,11 This property makes wettability difficult, creating problems filling micrometer-sized channels with suitable buffers. Additionally, many biomolecules are hydrophobic or contain hydrophobic regions which results in significant adsorption to the walls of the device.11,12,16 (11) Duffy, D. C.; McDonald, J. C.; Schueller, J. A.; Whitesides, G. M. Anal. Chem. 1998, 70, 4974-4984. (12) Ocvirk, G.; Munroe, M.; Tang, T.; Oleschuk, R.; Westra, K.; Harrison, D. J. Electrophoresis 2000, 21, 107-115. (13) Auroux, P. A.; Iossifidis, D.; Reyes, D. R.; Manz, A. Anal. Chem. 2002, 74, 2637-2652. (14) Ro, K. W.; Lim, K.; Kim, H.; Hahn, J. H. Electrophoresis 2002, 23, 1129-1137. (15) Leach, A. M.; Wheeler, A. R.; Zare, R. N. Anal. Chem. 2003, 75, 967-972. (16) Hu, S.; Ren, X.; Bachman, M.; Sims, C. E.; Li, G. P.; Allbritton, N. L. Anal. Chem. 2002, 74, 4117-4123.
10.1021/la049974l CCC: $27.50 © 2004 American Chemical Society Published on Web 05/25/2004
5570
Langmuir, Vol. 20, No. 13, 2004
This adsorption leads to sample loss and diminished resolution in analytical separations. As mentioned above, strong EOF can occur in PDMS devices, but the surface must be oxidized or otherwise modified to support the EOF. In addition, EOF in oxidized devices is unstable, making reproducible electrophoretic separation or electrokinetic pumping a challenge.12 One approach to resolving the shortcomings of PDMS for bioanalytical applications is to chemically modify the surface of the device. A large body of literature exists on tailoring surface properties of polymers using both chemical and physical modifications.17,18 Recent descriptions of modifications of PDMS have been stimulated by its growing popularity in the manufacture of microfluidic devices for bioassays.7,8,10 Exposure of the PDMS surface to plasma oxidation renders the surface hydrophilic by ionization of silanol groups.7,19-22 This treatment improves wettability and supports a strong EOF for electrokinetic pumping provided the surface remains in contact with neutral or basic solutions. However, oxidized PDMS reverts to its hydrophobic character within a few hours after exposure to air.7,8,16,23 Dynamic coatings with polyelectrolyte multilayers of poly(styrene sulfonate) and poly(allylamine hydrochloride) or polybrene and dextran sulfate have been used to selectively coat microchannels for control and stabilization of EOF.24-26 Thormann and colleagues have recently described a three-layer modification of biotinylated IgG, neutravidin, and biotinylated dextran that considerably reduced adsorption of molecules to the channel walls while maintaining modest EOF.27,28 Regnier’s group published a combination of oxidation and radical polymerization to modify PDMS surfaces with C18 silanes.29 This combination gave the investigators efficient and reproducible separations of a variety of biomolecules. Chemical vapor deposition has been utilized by Langer and colleagues to place reactive groups on PDMS microdevices for subsequent attachment of biomolecules.30 Our laboratory recently described a UV grafting method to covalently link a variety of polymers to the surface of PDMS.16,31 This method yields a hydrophilic surface with reduced adsorption of peptides compared with oxidized PDMS, and excellent EOF stability even after exposure to air. A variety of other coating methods have been described and are reviewed in ref 10. Despite these advances, the choices available for modification of PDMS surfaces for microfluidic applications remain limited.10 (17) Chan, C. M. Polymer Surface Modification and Characterization; Hanser/Gardner Publications: Cincinnati, 1994; Chapters 1, 2, 5. (18) Garbassi, F.; Morra, M.; Occhiello, E. Polymer Surfaces, 2nd ed.; John Wiley and Sons: New York, 1998; Chapters 2, 6, 7. (19) Wang, B.; Abdulai-Kanji, Z.; Dodwell, E.; Horton, J. H.; Oleschuk, R. D. Electrophoresis 2003, 24, 1442-1450. (20) Spehar, A. M.; Koster, S.; Linder, V.; Kulmala, S.; de Rooij, N. F.; Verpoorte, E.; Sigrist, H.; Thormann, W. Electrophoresis 2003, 24, 3674-3678. (21) Sinton, D.; Escobedo-Canseco, C.; Ren, L.; Li, D. J. Colloid Interface Sci. 2002, 254, 184-189. (22) Efimenko, K.; Wallace, W. E.; Genzer, J. J. Colloid Interface Sci. 2002, 254, 306-315. (23) Ren, X.; Bachman, M.; Sims, C.; Li, G. P.; Allbritton, N. L. J. Chromatogr., B 2001, 762, 117-125. (24) Liu, Y.; Fanguy, J. C.; Bledsoe, J. M.; Henry, C. S. Anal. Chem. 2000, 72, 5939-5944. (25) Barker, S. L. R.; Ross, D.; Tarlov, M. J.; Gaitan, M.; Locascio, L. E. Anal. Chem. 2000, 72, 5925-5929. (26) Barker, S. L. R.; Tarlov, M. J.; Canavan, H.; Hickman, J. J.; Locascio, L. E. Anal. Chem. 2000, 72, 4899-4903. (27) Linder, V.; Verpoorte, E.; Thormann, W.; de Rooij, N. F.; Sigrist, H. Anal. Chem. 2001, 73, 4181-4189. (28) Linder, V.; Verpoorte, E.; de Rooij, N. F.; Sigrist, H.; Thormann, W. Electrophoresis 2002, 23, 740-749. (29) Slentz, B. E.; Penner, N. A.; Regnier, F. E. J. Chromatogr., A 2002, 948, 225-233. (30) Lahann, J.; Balcells, M.; Lu, H.; Rodon, T.; Jensen, K. F.; Langer, R. Anal. Chem. 2003, 75, 2117-2122.
Hu et al.
UV-mediated grafting has been utilized to polymerize various monomers onto the surface of a PDMS microfluidic device.16 However, there were a number of limitations to this method. The electroosmotic mobility (µeo) could only be varied by altering the graft density or by utilizing a different monomer. The graft density could not be kept constant while the surface charge or µeo was varied. This suggests that for a single monomer µeo of the device can only be varied by altering the degree of surface coverage. A method in which the surface coverage and charge density could be varied independently would be of great utility for microfluidic devices. In addition, Hu et al. demonstrated the grafting of only neutral and negatively charged surface coatings.16,31 In this paper, we demonstrate the ability to tailor the surface properties of PDMS microfluidic devices by grafting mixtures of monomers onto the surface. UV grafting was used to polymerize a mixture of neutral and negative monomers or a mixture of neutral and positive monomers onto the surface of PDMS. The properties of these covalently attached coatings were investigated by determining the graft density on the PDMS surface, the contact angle, and the chemical stability of the surface. To determine whether the coatings might also be useful to fine tune the surface properties of channels, µeo and the stability of µeo were measured for microfluidic devices grafted with a wide range of monomer ratios. Experimental Section Reagents. Sylgard 184 was purchased from Dow Corning (Midland, MI) and silicon nitride coated silicon wafers were obtained from Wafernet, Inc. (San Jose, CA). Acrylic acid (AA), acrylamide (AM), N,N-dimethyl acrylamide (DMA), 2-hydroxy ethyl acrylate (HEA), poly(ethylene glycol) monomethoxyl acrylate (PEG), poly(ethylene glycol) diacrylate, 2-methacryloxyethyltrimethylammonium chloride (MATC), NaIO4, and benzyl alcohol were all obtained from Aldrich and used without further purification. All other reagents and materials were purchased from Fisher Scientific (Pittsburgh, PA). Microchip Fabrication. Microfluidic channel patterns and the corresponding master were designed and fabricated as described previously.23 Both micromolded and flat PDMS films were obtained as described.16 The microdevice was fabricated by sealing the micromolded PDMS half of the microfluidic device against a flat PDMS substrate. Grafted microdevices were fabricated by first grafting the two PDMS halves with a polymer as described in the next section and then mating together. Ultraviolet Polymer Grafting. Micromolded or flat PDMS films were immersed in an aqueous solution containing NaIO4 (0.5 mM), benzyl alcohol (0.5 wt %), and monomers at the concentrations and ratios indicated in the text.16 The immersed films were placed in a custom-built irradiator (200-W mercury lamp) for 3 h unless stated otherwise. The distance between the sample and the lamp was 5 cm. Uniform UV exposure was ensured by rotating the films (10 rpm) under the UV source. The samples were then washed in distilled water at 80 °C under constant stirring for 24 h to remove adsorbed monomer and polymer. The PDMS films were then vacuum-dried (0.01 atm, 23 °C). The graft density was defined as the difference in the film weight before and after grafting divided by the total surface area of the film. For measurement of graft density, the PDMS sample was approximately 4.5 cm × 2.5 cm × 0.2 cm. Measurement of Contact Angle. Contact angles were measured on flat PDMS films with or without grafted polymer. A droplet of deionized water was placed on the surface of a film at room temperature, and after 30 s the contact angle was measured using a contact angle goniometer (NRL-100, RameHart). The average of five measurements was utilized for each droplet. Measurement of µeo. The current monitoring method was used to measure µeo in the microfabricated channels.23,32,33 Measurements were performed as described previously.23 The composition of the higher ionic strength buffer was 20 mM
Properties of PDMS Microfluidic Devices
Langmuir, Vol. 20, No. 13, 2004 5571
Figure 1. Reaction scheme for graft polymerization of co-mixed monomers onto a PDMS surface. The initial free radical (left-most structure) was generated at the surface of the PDMS by irradiation with UV light. The PDMS surface reacts with one or both of the reactive species [CH2C(R3)COR1 or CH2C(R4)COR2]. R1 and R2 represent the monomer side groups that were employed. R2 and R4 represent the presence or absence of a methyl group on the acrylate backbone. The right-most structure represents one of many possible products. phosphate (pH 7.0), while the lower ionic strength buffer was composed of 18 mM phosphate (pH 7.0). A minimum of three measurements were performed on each device. The temperature was 23 °C with humidity at 20-30%. To measure the stability of µeo after exposure to air, the channels were flushed with water, dried under a vacuum, and then exposed to air at room temperature. At the times indicated, the channels were filled with aqueous buffer and µeo was measured. After the measurement, the channels were flushed with water, dried under a vacuum (0.1 atm, 23 °C), and again exposed to air at room temperature until the next measurement of µeo.
Results and Discussion Simultaneous Grafting of Neutral and Negative Monomers. Copolymerization of different monomers has been used extensively throughout polymer chemistry to achieve fine control of surface properties.34,35 For surface coatings on PDMS microfluidic devices, the degree of coverage, the hydrophobicity, the chemical stability, and µeo are important properties that must be controlled. To determine whether a copolymerized coating might yield superior surface properties compared to that grafted by a single monomer, a variety of monomers were co-mixed and then UV grafted onto the surface of PDMS (Figure 1). The monomers were selected predominantly by their reported ability to resist protein adsorption and their range of hydrophilicities.36-38 Initially, a neutral monomer (HEA, AM, DMA, or PEG) was mixed with varying concentrations of a negative monomer (AA) to create a surface with a variable negative charge. The total concentration of monomer (AA plus neutral monomer) was held constant while the concentration of AA was varied. To determine (31) Hu, S.; Ren, X.; Bachman, M.; Sims, C. E.; Li, G. P.; Allbritton, N. L. Electrophoresis 2003, 524, 3679-3688. (32) Liu, Y.; Fanguy, J. C.; Bledsoe, J. M.; Henry, C. S. Anal. Chem. 2000, 72, 5939-5944. (33) Huang, X.; Gordon, M. J.; Zare, R. N. Anal. Chem. 1988, 60, 1837-1838. (34) Garbassi, F.; Morra, M.; Occhiello, E. Polymer Surfaces, 2nd ed.; John Wiley and Sons: New York, 1998; Chapters 8, 12. (35) Stevens, M. P. Polymer Chemistry, 3rd ed.; Oxford University Press: New York, 1999. (36) Ikada, Y. Biomaterials 1994, 15, 725-36. (37) Belanger, M. C.; Marois, Y. J. Biomed. Mater. Res. 2001, 58, 467-477. (38) Jagur-Grodzinski, J. Heterogeneous Modification of Polymers, John Wiley and Sons: New York, 1997; Chapters 7, 8.
the quantity of monomer deposited, the graft density was measured (Figure 2A). As the concentration of AA was increased, the graft density for all the mixtures increased but only slightly. For example, quadrupling the concentration of AA yielded only a 7% increase in the graft density of the AM/AA mixture. This small increase is likely due to the fact that although the concentration of AA was increasing the total monomer concentration was held constant. The nearly constant graft density also suggests that the thickness of the coating might have remained nearly constant despite the increasing concentration of charged monomer. The relative ordering of the graft density (AM > HEA > DMA > PEG) for any concentration of AA was also consistent with previous experimental work utilizing the grafted, single monomers.16 This is not unexpected because the neutral monomer was in a molar excess at all data points (except the 50% AA point in the AM/AA comixture). Properties of Surfaces Grafted with Co-Mixed Neutral and Negative Monomers. The contact angle that a water droplet makes with a surface is frequently used as a measure of the hydrophobicity of that surface. To determine the surface hydrophobicity of PDMS grafted with the different co-mixed monomers, the contact angle of a water droplet on the surface was measured after grafting with a 10:1 ratio of neutral monomer/AA (Table 1). For reference, the measured contact angles of native and oxidized PDMS were 109 and 15°, respectively.16 For these conditions, all of the grafted surfaces were substantially more hydrophilic than native PDMS but less hydrophilic than oxidized PDMS. Because the grafted surfaces possess a mixture of charged and neutral monomers, they most likely have a lower net charge density than oxidized PDMS. All of the neutral monomer/AAgrafted surfaces possessed smaller contact angles than surfaces grafted with similar densities of only neutral monomer and larger contact angles than surfaces grafted with only AA. Thus, the hydrophilicity of the surfaces was intermediate to that of surfaces with only neutral monomer or only AA, suggesting that both types of monomers were grafted onto the surface.16 The co-mixtures of HEA/AA and AM/AA were the most hydrophilic presumably as a result of the polarity of the hydroxyl or amide group, respectively, in addition to that contributed by AA.
5572
Langmuir, Vol. 20, No. 13, 2004
Hu et al.
Figure 2. Measurement of the graft density and the stability of a water droplet contact angle for PDMS films grafted with co-mixed monomers. (A) Effect of the concentration of AA on graft density. The ratio of the weight of AA to the weight of AA plus neutral monomer was varied as shown on the x axis. The total weight of the monomers with respect to the total weight of the final solution was 10%. The neutral monomers were AM (open circles), HEA (open squares), DMA (closed circles), and PEG (closed squares). Each data point is the average measurement from four different films, and the error bars represent the standard deviation of the measurements. (B) Stability of the contact angle of a water droplet on PDMS films grafted with co-mixed monomers. The PDMS films were grafted with PEG/AA (20:1, w/w; open triangles) or PEG/MATC (10:1; solid triangles) or were treated with an oxygen plasma (open diamonds) as described previously.16 The total concentration of the co-mixed monomers was 10% (combined weight of the monomers to weight of the final solution). Immediately after grafting or oxidation (time 0, x axis), the contact angle of the films was measured. The films were then dried and exposed to air. At each of the times indicated on the x axis, the contact angle was measured. After each measurement, the devices were dried and stored in air until the next measurement of the contact angle. The data points are the average of six experiments. In most instances, the error bars were smaller than the symbols representing the data points on the graph. (C) Effect of the concentration of MATC on graft density. The ratio of the weight of MATC to the total weight of the monomers (neutral monomer plus MATC) was varied as shown on the x axis. The total weight of the monomers with respect to the total weight of the final solution was constant (10%). The neutral monomers were AM (open circles), HEA (open squares), DMA (closed circles), and PEG (closed squares). Each data point is the average measurement from four different films, and the error bars represent the standard deviation of the measurements. (D) Effect of salt on the graft density of PEG/MATC and AM/MATC. The conditions are identical to those of part C except that NaCl (500 mM) was added to the reaction mixture prior to exposure to UV light.
Although oxidized PDMS is extremely hydrophilic, the contact angle and surface properties of oxidized PDMS are unstable when exposed to air.12,23 To determine the stability of the co-mixed monomer-grafted surfaces, the contact angle was measured after exposure to air. The contact angle of PEG/AA-grafted surfaces (shown in Figure 2B) as well as the other co-mixed monomer-grafted surfaces (not shown) was constant for up to 3 months, the longest time measured. In contrast, the contact angle of oxidized PDMS increased from ∼15 to ∼100° over a 50-h time span. These results suggest that microfluidic devices grafted with these co-mixed monomers will have stable surface and separation properties after exposure to air.
Hu et al. reported that many of the devices grafted with single monomers became rigid, opaque, and rough at high graft densities and, therefore, were not useful for electrophoretic applications.16 Devices grafted with HEA/AA or PEG/AA were flexible and transparent at all tested graft densities and monomer concentrations. PDMS slabs grafted with these co-mixed monomers also sealed well and reversibly with other identically grafted slabs. AM/AA- and DMA/AA-grafted devices remained flexible, smooth, and transparent at all monomer ratios as long as the graft density was less than ∼60 µg/cm2 (Figure 3A). At these graft densities, the grafted PDMS slabs also sealed well and reversibly with like-grafted slabs. At graft densities greater than ∼60 µg/cm2, the AM/AA and
Properties of PDMS Microfluidic Devices
Langmuir, Vol. 20, No. 13, 2004 5573
Table 1. Contact Angles of Devices Grafted with Co-Mixed Monomersa mixture HEA/AA AM/AA DMA/AA PEG/AA HEA/MATC AM/MATC DMA/MATC PEG/MATC native PDMS oxidized PDMS
ratio (w/w)
graft densityb (µg/cm3)
contact angleb (deg)
10:1 10:1 10:1 10:1 10:1 10:1 10:1 10:1
58 ( 2 67 ( 2 43 ( 2 29 ( 1 52 ( 2 59 ( 2 36 ( 2 27 ( 1
52 ( 1 61 ( 1 75 ( 2 75 ( 2 55 ( 1 63 ( 1 77 ( 2 77 ( 2 109c 15c
a The total weight of monomer to that of the final solution was 10%. b Measurements are the average and standard deviation from measurements on five different films. c See refs 13 and 16.
Figure 3. Photographs of UV-grafted slabs of PDMS. (A) A slab of PDMS was grafted with AM/AA (10:1 ratio and 10% total monomer) for 1 h. The graft density was 20 µg/cm2. (B) PDMS was grafted with AM/AA (10:1 ratio and 10% total monomer) for 4 h. The graft density was 78 µg/cm2. (C) PDMS was grafted with AM/MATC (10:1 ratio and 10% total monomer) for 1 h. The graft density was 16 µg/cm2. (D) PDMS was grafted with AM/MATC (10:1 ratio and 10% total monomer) for 4 h. The graft density was 69 µg/cm2.
DMA/AA-grafted PDMS became opaque and rigid (Figure 3B). Properties of Surfaces Grafted with Co-Mixed Neutral and Positive Monomers. To create a surface with a variable positive charge, the neutral monomers were co-mixed with varying concentrations of a monomer with a positive charge, MATC, and then UV-grafted onto PDMS. The total concentration of monomer (MATC plus neutral monomer) was held constant while the concentration of MATC was varied. To determine the quantity of monomer deposited, the graft density was measured (Figure 2C). In contrast to the monomer mixtures with AA, the graft density for the mixtures with MATC decreased as the concentration of MATC was increased. At the pH of the MATC reaction mixtures (pH 4.7), MATC possessed a positive charge. Consequently, one possible explanation is that the positive charge density increased
Figure 4. Measurement of µeo in microchannels grafted with PEG and AA or PEG and MATC. (A) Effect of the concentration of AA or MATC on µeo. The ratio of the weight of AA (circles) or MATC (squares) to the total weight of the monomers was varied as shown on the x axis. The total weight of the monomers with respect to the total weight of the final solution was 10%. (B) Stability of the EOF for the microchannel exposed to air. Microchannels were grafted with either PEG/AA (7:3, 10% total monomer) or PEG/MATC (7:3, 10% total monomer). The magnitude of µeo was measured after exposure of PEG/AAgrafted (open triangles, left y axis), PEG/MATC-grafted (solid triangles, right y axis), or oxidized (open diamonds, left y axis) microchannels to air. After fabrication of the channels, µeo was measured immediately, and the devices were filled with water and stored in water. After the second measurement of µeo at 2 h, the channels were filled with air and stored in air. After all subsequent measurements of µeo, the channels were again filled with air. For parts A and B, the data points are the average of three different devices and the error bars represent the standard deviation.
even though the graft density decreased. The increasing density of the positive charge might have then limited the graft density because of charge-charge repulsion. This hypothesis was supported by the fact that the absolute value of µeo increased with the concentration of MATC despite a decreasing graft density (see Figure 4A and next section). In contrast, the carboxyl group of AA (pKa 4.3) was protonated under the conditions used to graft AA (pH 2.7).39 The uncharged AA could be deposited at increasingly higher densities as the percentage of AA was increased in the reaction mixture. When AA was grafted under conditions which yielded a charged carboxylate (pH 7.5), the graft density decreased as the percentage of AA increased (data not shown). To confirm that the charge of MATC limited the graft density as the concentration of MATC increased, the grafting reaction was repeated for (39) The Merck Index; Merck & Co., Inc.: Whitehouse Station, NJ, 2001; p 24.
5574
Langmuir, Vol. 20, No. 13, 2004
the PEG/MATC and AM/MATC mixtures but with the addition of NaCl (500 mM). Substantially higher graft densities were obtained in the presence of the salt (Figure 2D). These additional ions shielded the charged group of MATC, suggesting that charge-charge repulsion might limit the density of MATC grafted onto the PDMS surfaces. At all tested graft densities and ratios of PEG/MATC and HEA/MATC, the grafted PDMS remained transparent, flexible, and sealed reversibly with like surfaces. At graft densities of less than 60 µg/cm2, AM/MATC and DMA/ MATC also remained transparent, flexible, and sealed reversibly with like surfaces (Figure 3C,D). From the measured contact angles of a water droplet on the neutral monomer/MATC-grafted surfaces, these coatings were considerably more hydrophilic (Table 1) than native PDMS but less hydrophilic than oxidized PDMS. The neutral monomer/MATC-grafted surfaces also possessed smaller contact angles than surfaces grafted with similar densities of neutral monomer alone, suggesting that both types of monomers were successfully grafted onto the surface. In addition, the contact angle of a water droplet on the neutral monomer/MATC-grafted surfaces remained constant over time even after exposure to air suggesting that microfluidic devices with these coatings would have stable properties upon exposure to air (Figure 2B). Measurement of µeo in Grafted Microchannels. To fabricate microchannels with grafted surfaces, the two halves of a device were first constructed. The bottom half consisted of a thin slab of flat PDMS while the top half with the imprinted microchannels was formed by casting PDMS against a silicon mold. Both halves of the device were grafted with identical mixtures of PEG and either AA or MATC. The device was then assembled by placing the halves together. For both monomer mixes, the two halves sealed reversibly and the channels were easily filled with aqueous solutions. When the concentration of AA or MATC to total monomer was greater than 20%, aqueous buffers applied to the mouth of the channels spontaneously filled the channels by capillary action. At lower concentrations of AA or MATC, the channels were still easily filled with aqueous solutions, but an additional force was required, that is, application of a vacuum to the outlet end of the channel. To determine whether µeo could be varied, the ratio of AA or MATC to PEG grafted onto the two PDMS halves was varied. As the concentration of AA was increased, µeo increased from 1 × 10-4 cm2/(V s) (0% AA) to ∼4 × 10-4 cm2/(V s) (50% AA; Figure 4A). In contrast, as the concentration of MATC was increased, µeo decreased from 1 × 10-4 cm2/(V s) (0% MATC) to -1.8 × 10-4 cm2/(V s) (50% MATC). Any µeo between -1.8 and 4 × 10-4 cm2/(V s) was achievable with this simple, onestep grafting procedure. Stability of µeo in Grafted Microchannels. To determine the stability of µeo, an oxidized PDMS channel, a
Hu et al.
channel grafted with PEG/AA, or a channel grafted with PEG/MATC was filled with buffer and µeo was measured. After 1 h, µeo was again measured in the buffer-filled channels. After this measurement, the channels were filled with air for storage and were refilled with buffer for successive µeo measurements. The µeo for the PEG/AAgrafted and PEG/MATC-grafted channels remained constant while µeo of the oxidized channels decreased over time (Figure 4B). Conclusions We have demonstrated a simple, one-step process to tailor the surface properties of PDMS microchannels. A wide range of charge densities, both positive and negative, were imparted to PDMS surfaces by co-mixing neutral monomers with either positive or negative monomers at varying ratios. Mixtures of four different neutral monomers and two different charged monomers were analyzed. The contact angle of a water droplet on these grafted surfaces and, consequently, their hydrophilicity was intermediate to that of oxidized and native PDMS and intermediate to that of surfaces grafted with only the neutral or only the charged monomer. Thus, the coatings exhibited properties of both the neutral and charged monomers. At low graft densities (e60 µg/cm2) all of the mixtures yielded coatings that were transparent, flexible, and readily sealed with like surfaces. Functional microchannels were fabricated from PDMS halves grafted with each of the different mixtures. By varying the concentration of the charged monomer grafted with PEG, a wide range of µeo, both positive and negative, was created in the microchannels. In addition, µeo was stable in these devices after exposure to air. By carefully selecting mixtures of monomers with the appropriate properties, it may be possible to tailor the surface of PDMS for many different applications, such as electrokinetic pumping and electrophoretic separations. By grafting differentially charged polymers onto the top and bottom pieces of a microfluidic device, interfacial reactions can be studied, chemical syntheses performed, and mixing devices constructed.9,10,25,40-43 Acknowledgment. This research was supported by the NIH (RR14892, GM57015) and DARPA (N66001-01C8014). The authors thank Ruisheng Chang and Wei Wei for technical assistance. LA049974L (40) Stroock, A. D.; Whitesides, G. M. Acc. Chem. Res. 2003, 36, 597604. (41) Ng., J. M.; Gitlin, I.; Stroock, A. D.; Whitesides, G. M. Electrophoresis 2002, 3461-3473. (42) Zhao, B.; Viernes, N. O. L.; Moore, J. S.; Beebe, D. J. J. Am. Chem. Soc. 2002, 124, 5284-5285. (43) Kenis, P. J. A.; Ismagilov, R. F.; Whitesides, G. M. Science 1999, 285, 83-85.
