Nonfouling Hydrophilic Poly(ethylene glycol) Engraftment Strategy for

Oct 30, 2012 - Canadian Food Inspection Agency, Ottawa Fallowfield Laboratories, .... Developing an ultra non-fouling SU-8 and PDMS hybrid microfluidi...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Langmuir

Nonfouling Hydrophilic Poly(ethylene glycol) Engraftment Strategy for PDMS/SU‑8 Heterogeneous Microfluidic Devices Po Ying Yeh,† Zhiyi Zhang,*,‡ Min Lin,§ and Xudong Cao*,† †

Department of Chemical and Biological Engineering, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada Institute of Microstructural Science, National Research Council Canada, Ottawa, Ontario K1A 0R6, Canada § Canadian Food Inspection Agency, Ottawa Fallowfield Laboratories, Ottawa, Ontario K2H 8P9, Canada ‡

S Supporting Information *

ABSTRACT: We report a novel nonfouling passivation method using poly(ethylene glycol) (PEG) engraftment on the surfaces of poly(dimethylsiloxane) (PDMS) microfluidic devices sealed with SU-8. To achieve bonding between the PDMS and SU-8 surfaces, the PDMS surface was first functionalized with amines by treatment with 3-aminopropyltrimethoxysilane (APTMS) for subsequent reaction with epoxide functional groups on SU-8 surfaces. To modify the heterogeneous surfaces of the resulting PDMS/SU-8 microfluidic device further, the remaining SU-8 surfaces were amino functionalized using ethylene diamine (EDA), followed by treating both amino-functionalized PDMS and SU-8 surfaces with mPEG-NHS (N-hydroxysuccinimide) through an amine-NHS reaction for facile PEG immobilizations, thus simultaneously modifying both PDMS and SU-8 surfaces in one reaction. Detailed surface analyses such as the water contact angle, X-ray photoelectron spectroscopy (XPS), and atomic force microscopy (AFM) were conducted to confirm the chemical reactions and characterize the resulting surface properties. To test the efficacy of this surface-modification strategy, we conducted nonspecific protein and particle binding tests using microfluidic devices with and without modifications. The PEG-modified PDMS/SU-8 device surfaces showed a 64.5% reduction in nonspecific bovine serum albumin (BSA) adsorption in comparison to that of the unmodified surfaces and 92.0 and 95.8% reductions in microbead adhesion under both stagnant and flowing conditions, respectively.

I. INTRODUCTION Polymers used in micrototal analysis systems (μTASs) have grown increasingly important as research in the μTAS field rapidly progresses. This is because polymers offer many advantages over other materials, such as rapid fabrication at relative low cost and a wide variety of polymeric material choices for different applications.1 In addition, these polymerbased μTASs have provided important platforms for many biomedical applications, including biomedical analysis,2 proteome profiling,3 chromatography,4 and cell sorting.2,5 Among those polymers, poly(dimethylsiloxane) (PDMS) and epoxybased SU-8 are of particular interest. PDMS is a popular material because of its elasticity, moldability, chemical inertness, optical transparency in the visible spectral region, and outstanding gas permeability to CO2 and O2.6 In comparison, SU-8, a material originally used mainly in master fabrication processes for the mass production of PDMS microfabricated devices,7 is now increasingly being used as a material for fabricating microfluidic devices directly in applications such as microfluidic flow cytometry8 and chemical assays.9 Other applications, such as micromanipulation10 and magnetic bead separation,11 have also used SU-8-based μTAS devices. Sealing PDMS-based μTASs by either glass or another piece of PDMS is relatively simple. This is generally done by initially treating the two surfaces with O2 plasma to generate silanol © 2012 American Chemical Society

groups on the two separate surfaces and then bringing them together to form strong chemical bonding via silanol condensation.12 Although this works well for microfluidic devices with relatively simple structures, it is not suitable for complex microfluidic devices that need precise alignment because irreversible bonding between the two surfaces is immediately established once they are brought together.13,14 In contrast, the sealing of SU-8-based microchips has proven to be much more difficult. This is because SU-8 is a thermosetting polymer; once cured, it is already extensively cross-linked and therefore will not allow molecular diffusion for molecular entanglements with other layers to form stable bonds.15,16 To solve this problem, recently we introduced a novel method of sealing SU-8 microchannels with PDMS by introducing 3aminopropyltrimethoxysilane (APTMS) onto a PDMS surface and subsequently establishing covalent bonds between the SU8 and PDMS surfaces through an amine−epoxide reaction by moderate heating.17 This bonding process of SU-8 with PDMS offers several advantages. For example, the bonding of the two surfaces (i.e., PDMS and SU-8) becomes irreversible only after heating, which is especially beneficial for microfluidic devices Received: August 7, 2012 Revised: October 30, 2012 Published: October 30, 2012 16227

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

of the PEG-grafted PDMS/SU-8 hybrid microfluidic device, bovine serum albumin (BSA) and hydrophobic polystyrene microbeads are used as models for proteins and cells, respectively, to test the nonfouling properties of the modified surfaces.

that need precise alignment in the fabrication processes. Moreover, the unreacted amino functional groups on the PDMS surfaces would allow additional functionalizations, such as poly(ethylene glycol) (PEG) engraftment7 and DNA immobilization.18 In addition, the PDMS/SU-8 hybrid microfluidic device offers many advantages over PDMS/PDMS microdevices in many applications. For example, Ou et al. have reported that protein separation in a PDMS/SU-8/quartz hybrid chip demonstrates an improved detection limit and heat dissipation in comparison to the control PDMS/PDMS microdevice.19 Moreover, the SU-8 layer, which is an excellent optical waveguide material, can be integrated with optofluidic devices. For example, using standard photolithography technique, Watts et al. have shown that SU-8 can be easily and precisely fabricated as both the microchannel wall and a built-in lens to focus the excitation beam for microfluidic online optical detection.8 Therefore, there is a real need for such a hybrid device, and passiving the PDMS/SU-8 device surfaces after bonding remains a significant challenge because one has to deal with two distinct surfaces and therefore surface chemistries and the passivation of the microfluidics device surface is necessary to inhibit the nonspecific adsorption of reagents, improve surface wettabilities, and stabilize surface electro-osmotic mobilities.20,21 To passivate PDMS surfaces alone, covalently tethering hydrophilic polymer chains to PDMS surfaces proves to be a popular approach.21−25 For example, glycidyl methacrylate-graf ted-poly(vinyl pyrrolidone) (PVP-g-GMA) and glycidyl methacrylate-graf ted-poly(vinyl alcohol) (PVA-gGMA) have been used to modify O2-plasma-treated and silanized PDMS surfaces, respectively.20 In addition, PDMS electrophoresis microchips have also been successfully modified by PEG to prevent protein adsorption and enhance the separation performance of the microchip.22 Other approaches are also reported to modify the surfaces of PDMS channels. For example, an undecylenic acid-doped PDMS channel has been modified with n-dodecyl-β-D-maltoside to provide stable electro-osmotic flow (EOF) and reduce protein adsorption on channel walls, thus increasing the electrophoretic separation efficiency of proteins.26 In another study, PDMS channels modified by 2-(4-chlorosulfonylphenyl) ethyltrimethoxysilane are reported to enhance the electro-osmotic pumping and stability over a wide range of pH conditions ranging from 3 to 8, thus extending the microfluidic device application to lower pH ranges.27 Similarly, modifications of the SU-8 surface alone are also reported.21,28 For example, PEG-OSiCl3 has been utilized to attach covalently to a preoxidized SU-8 surface to improve the nonfouling properties of the surface while enhancing its biofunctionality.21 SU-8 microdevices are also reported to be modified by PEG-stabilized phosphatidylcholine lipid aggregates (PEG-lipid) to reduce biofouling on the SU-8 microdevice efficiently.28 Although passivation strategies of individual PDMS-based and SU-8 based microfluidic devices by polymer engraftment are well known,29 a facile approach to modifying PDMS/SU-8 hybrid microfluidic devices has been elusive. In this study, for the first time, we report a modification strategy to graft PEG covalently onto PDMS/SU-8 hybrid microfluidic device surfaces. Our approach is first to aminate PDMS surfaces to achieve PDMS/SU-8 bonding, followed by SU-8 surface amination, and the resulting aminated PDMS and SU-8 surfaces are both modified with amine-reactive mPEG-NHS in one facile chemical reaction. To evaluate the nonfouling properties

II. EXPERIMENTAL SECTION 1. Materials. Rhodamine-NHS (P146406) was purchased from Thermo Scientific Inc. (Rockford, IL). mPEG-NHS (PEG1-0003, MW 1000) was purchased from Nanocs Inc. (New York, NY). Polybead microspheres (19814-15, 2 μm in diameter) and Fluoresbrite YG microspheres (18338-5, 2 μm in diameter) were purchased from Polysciences Inc. (Warrington, PA). 2-Propanol (IPA, 5873-15) was purchased from JT Baker (Phillipsburg, NJ). Anhydrous ethanol was obtained from Commercial Alcohols Inc. (Boucherville, QC). A Sylgard 184 PDMS kit was purchased from Dow Corning (Midland, MI). An SU-8 2025 negative photoresist was purchased from MicroChem (Newton, MA). Other chemicals, such as bovine serum albumin−fluorescein isothiocyanate conjugate (BSA-FITC, A9771), ethylenediamine (EDA, E1521), ethanolamine (ETAM, 411000), (3aminopropyl) trimethoxysilane (APTMS, 281778), N,N-dimethylformamide (DMF, 227056), toluene (155004), and phosphate-buffered saline tablets (PBS, P4417-50TAB) were purchased from SigmaAldrich (Oakville, ON). 2. PDMS and SU-8 Preparation, Amination, and PEG Immobilization. A. PDMS Fabrication and PDMS and SU-8 Surface Preparation. PDMS microchannels were fabricated using a standard soft lithography process. Briefly, a 19-μm-thick layer of SU-8 2025 was patterned by photolithography on top of a 4 in. silicon wafer as a master for PDMS microchannels. Sylgard 184 and curing agent were mixed in a 10:1 ratio (w/w) using a glass rod for 5 min and slowly poured on top of the SU-8 master. The PDMS was left overnight at room temperature for degassing and subsequently cured at 90 °C for 2 h. For a PDMS surface without patterns, similar procedures were followed except for pouring the PDMS precursor on top of a 4 in. bare silicon wafer. Similarly, to prepare a SU-8 surface without patterns, the whole SU-8 surface was exposed to UV without using a mask. The resulting SU-8 surfaces were used to seal PDMS microchannels. B. Surface Amination. a. PDMS Surface Amination. To aminate a PDMS surface, we first activated it with air plasma using a plasma cleaner (PDC-001, Harrick Plasma Inc., Ithaca, NY) at 250 mTorr for 30 s. Immediately after the air plasma treatment, the PDMS surface was immersed in APTMS solution (1 wt % in toluene) for 30 min to allow APTMS binding on the PDMS surface through the silanol− silanol reaction. The aminated surface was then washed in 100% ethanol for 10 min. Finally, the APTMS-modified PDMS surface was held at 130 °C for either 30 or 60 min. b. SU-8 Surface Amination. To aminate a SU-8 surface, we treated it with EDA-based solutions. This enabled the amino groups on one end of the bifunctional EDA to react with epoxy groups on the cured SU-8 surface while allowing the other amine moiety to serve as the amino functionality after modification on the SU-8 surface.30,31 To minimize the reduction of the epoxy groups on the SU-8 surfaces, thereby allowing further surface modifications, the hard-bake step was deliberately omitted in the SU-8 fabrication processes.32 Specifically, a SU-8 surface was incubated with EDA solution (IPA/ethanol 1:1 v/v) for 1 h, washed with IPA/ethanol (1:1 v/v) 10 times, and subsequently dried at 40, 55, or 70 °C for 1 h. Two EDA concentrations (20 and 40 wt %) were used in the incubation step. In some cases, an additional 10 wt % ETAM was added to the EDA solution as a catalyst to aid the amination on the SU-8 surfaces. c. Quantification of Surface Amination. Fluorescent rhodamineNHS that readily reacts with primary amines was used to evaluate the amination level of modified surfaces. The fluorescence intensity of the rhodamine-NHS-labeled surface is established to be proportional to the quantity of surface amine groups reacting with the rhodamineNHS molecules.33,34 Briefly, rhodamine-NHS bulk solution (10 mg/ mL) was prepared in anhydrous DMF. The aminated surfaces were 16228

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

Figure 1. Illustration of the general approach to fabricating PEG graft PDMS/SU-8 heterogeneous microchannels. immersed in PBS (pH 7.4), and the prepared rhodamine-NHS bulk solution was subsequently added to PBS (pH 7.4) in a quantity of 1 μL bulk solution per cm2 of surface area for labeling. For example, if the surface area of the sample was 10 cm2, then 10 μL of rhodamineNHS bulk solution was added. The surfaces were incubated in the labeling solution for 30 min, washed in fresh PBS (pH 7.4) 10 times, and stored in PBS (pH 7.4) until further analysis. It should be noted that the NHS functional groups are very easy to hydrolyze;33 therefore, the rhodamine-NHS labeling solution should be prepared immediately before the experiment. Fluorescent images were captured using an inverted fluorescence microscope (Olympus X70) equipped with a high-resolution camera (QIC-F-CLR-12, QImaging). To analyze the data, we first converted the collected color images to grayscale images, and the average intensities of the grayscale images were analyzed by GIMP 2.6 (freeware available at http://www.gimp.org/). The average intensities of the grayscale images represent the relative quantity of rhodamineNHS conjugated on the surface of interest and therefore an indirect indication of the relative amount of amines on the surface.33 To compare the amination levels of surfaces created by various treatments, the background intensity (i.e., the fluorescence intensity of the pristine surface before amination) was first subtracted from the measured average intensity on the surface. All measured intensities were then normalized by that of a standard PDMS surface. The standard PDMS surface was a PDMS surface aminated by APTMS treatment, stored in PBS (pH 7.4) for 6 days, and then labeled with rhodamine-NHS. C. PEG Surface Immobilization and Characterization. a. PEG Surface Immobilization. To immobilize PEG on PDMS and SU-8 surfaces, both aminated PDMS and SU-8 surfaces obtained using respective APTMS and EDA treatments were incubated with mPEGNHS at predetermined bulk concentrations (i.e., 1, 3, 6, 10, 12, 20 mg/ mL (w/v) in PBS (pH 7.4)) at room temperature for 1 h, washed in PBS (pH 7.4) to remove unreacted PEG, and then stored in PBS (pH 7.4) until future use. Similar to rhodamine-NHS, mPEG-NHS solutions should be prepared immediately before the experiment. b. Surface Characterization. To characterize the surface property changes following each treatment, the water contact angles of experimental surfaces were measured using a goniometer (FTA200, First Ten Angstroms, Inc., Portsmouth, VA). For every measurement, the volume of DI water droplets was controlled to ∼10 μL. The droplets were kept pendent on the needle tip until dropping onto the target surfaces for measurement. To minimize water evaporation, measurements were taken within 10 s once the droplet settled on the sample surface. For each sample surface, four measurements (for a total of eight readings, two readings per measurement) were taken at random and different locations on the sample surface, and the average values were reported. An X-ray photoelectron spectroscope (XPS) from Physical Electronics (PHI 5500, Physical Electronics, Chanhassen, MN) was used to characterize the PDMS and SU-8 surfaces w/o mPEG engraftment. The instrument used a monochromatic Al Kα X-ray

source and a hemispherical electrostatic analyzer. It was equipped with a duoplasmatron ion gun (Ar gas, operation ranges 1−4 kV and 10− 200 nA) for sputter depth profiling. The takeoff angle for all measurements was fixed at 45°. In all spectra, the C−C/C−H peak was fixed at a binding energy of 285 eV as the reference. To analyze a specific peak in detail, the peak of interest was scanned at high resolution and was deconvoluted using XPSPEAK 4.1 software (freeware available at http://www.uksaf.org/software.html). In analyzing the data, a Shirley-type background and the Gaussian/Lorentzian (80:20) profile were applied to fit all peaks of interest as detailed in a previous study elsewhere.35 A Veeco Di EnviroScope multimode atomic force microscope (AFM) (Santa Barbara, CA) equipped with a Di NanoScope controller was used to profile the surface morphology in contact mode. The raw data were processed with NanoScope software by first-order flattening. 3. Microfluidic Device Fabrication and Device Performance. A. Microfluidic Device Fabrication. To fabricate microfluidic devices, the APTMS-modified PDMS microchannel was brought into contact with the SU-8 surface. The assembly was uniformly heated at 100 °C for 30 min. To obtain a fully functional microchannel, the device was connected to stainless steel tubes by leak-proof pressure sealing at 130 °C for 30 min. For subsequent surface modifications, the obtained microchannel (both PDMS and SU-8 surfaces) was incubated with an EDA-based solution for 1 h and was subsequently held at 70 °C for 1 h to react the epoxy groups on the SU-8 surface with amino functionalities. Finally, the microchannel was incubated by mPEGNHS solutions of various bulk concentrations for 1 h at room temperature for PEG engraftment unto microchannel surfaces. B. Device Performance. To test the efficacy of the PEG-modified surfaces, BSA was used as a model protein in a nonspecific protein adsorption study. Specifically, PEG-modified microchannels were incubated with a 2 mg/mL (w/v) BSA-FITC solution in PBS (pH 7.4) for 1 h at room temperature followed by extensive washing in fresh PBS (pH 7.4). The fluorescence intensities on the surfaces of the microchannels were analyzed to quantify the nonspecific BSA adsorption on microchannel surfaces as discussed in section 2.B.c. In addition, hydrophobic polybead and Fluoresbrite YG microspheres were used as model particles for studying particle adhesion to the modified surfaces, under both stagnant and dynamic flow conditions. Briefly, 2 μm polybead microbeads at 5.68 × 109 particles/mL and 2 μm Fluoresbrite YG microbeads at 5.95 × 109 particles/mL were diluted to working solutions of 1 × 109 and 1 × 108 particles/mL by PBS (pH 7.4), respectively. For stagnant conditions, a polybead microbead working solution was injected into a microchannel to allow incubation for 2 min, and the microchannel was subsequently washed with PBS (pH 7.4) several times to remove unbound microbeads. For dynamic flow conditions, a Fluoresbrite YG microbead working solution was allowed to flow through the microchannel at a flow rate of 50 μL/h for 1 h and subsequently washed with PBS (pH 7.4) to remove unbound microbeads. The 16229

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

Figure 2. Effects of different parameters on the amination of PDMS and SU-8 surfaces as determined by relative fluorescence intensities. (a) Relative fluorescence intensities of PDMS surfaces aminated by APTMS and stored in PBS (pH 7.4) for 1, 2, and 6 days and (b) relative fluorescence intensities of SU-8 surfaces aminated under different conditions (listed below the graph). Control samples were pristine PDMS and SU-8 surfaces incubated with rhodamine-NHS. Error bars represent the standard deviation of six to eight measured fluorescence intensities on each sample surface. number of microbeads adsorbed on the microchannel surface was counted manually under a microscope.

contrast, SU-8 surfaces were amino functionalized via EDA. The EDA molecule is the smallest bifunctional molecule with primary amino groups on both ends. The bi-amino functionality on one end of the EDA molecule conjugated with the epoxy group on the SU-8 surface and on the other amine moiety of the EDA provided amino functionality of the SU-8 surface. The short length between the two amino functionalities within one EDA molecule likely reduced the possibility for both amines to interact with the surface epoxy groups at the same time. The SU-8 surface was reported to be amino functionalized as well by oxidation with sodium periodate, EDA treatment, and then reduction with sodium cyanoborohydride to induce more uniform and effective amination compared to that of the nucleophilic ring-opening reaction.38 Fluorescent rhodamine-NHS was used as a reporter molecule toward surface amino functionalities in this study. Although this is not a surface functional group quantification method due to nonspecific adsorption, fluorescence labeling has been widely reported as a very useful tool in the relative comparison of densities of surface functional groups of interest36,39,40 and is particularly helpful in this study in investigating parameters that affect the amination levels on PDMS or SU-8 surfaces. The relative fluorescence intensities of amino-modified PDMS and SU-8 surfaces are shown in Figure 2. The control samples were pristine PDMS and SU-8 surfaces incubated with rhodamineNHS, respectively. The weak fluorescence intensities exhibited by the controls, PDMS (0.07 ± 0.01) and SU-8 (0.04 ± 0.01), indicate a small degree of nonspecific adsorption of rhodamineNHS on the control surfaces. In contrast, for PDMS surfaces, a much stronger fluorescence intensity was observed on APTMS-

III. RESULTS AND DISCUSSION In this article, we study covalent PEG engraftment on the surfaces of PDMS/SU-8 hybrid microchannels. As shown in Figure 1, our general approach is first to aminate PDMS microchannels with APTMS treatment through silanol condensation. This was followed by sealing with the SU-8 surface through an amine-epoxide reaction accelerated by the application of heat. Subsequently, the SU-8 surfaces of the microchannels were amino functionalized with EDA treatment. A common amine reactive reagent, mPEG-NHS, was used to graft PEG molecules onto surfaces of the PDMS/SU-8 hybrid microchannel through the amine-NHS reaction. It should be noted that the PDMS microchannel could also be amino functionalized by N2 plasma treatment as reported previously.16 Although the N2 plasma treatment includes fewer steps to amino functionalize PDMS, the amino functionalties introduced by this method are less stable because a faster hydrophobic recovery was observed when the sample was stored in PBS (fluorescence data not shown). Moreover, it seems that there were fewer amino functionalities on PDMS surfaces after EDA treatment because a very small PEG coating density was noted by the contact angle measurement (data not shown). 1. Material Surface Modification. A. Surface Amination. PDMS surfaces can be aminated by either physical methods, such as plasma activation,36 or chemical methods, such as silane coupling.17,37 In this study, APTMS coupling was used to introduce amino functionalities onto PDMS surfaces. In 16230

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

treated surfaces compared to the control as shown in Figure 2a. The fluorescence intensities seemed to have leveled off after 2 days of storage in PBS, suggesting stable, significant amination on the APTMS-treated PDMS surfaces and a slow hydrophobic recovery, a surface phenomenon due to the migration of small hydrophobic PDMS molecular chains from the bulk of PDMS to the surface of the material, causing surface rearrangement41 or polar functionality reorientation on the PDMS surface.35 The observation of retarded hydrophobic recovery in this study is consistent with the previous study demonstrating a slow hydrophobic recovery for the APTMS-treated PDMS surfaces after 1 h and up to 7 days of modification.35 In the case of SU-8, there was no noticeable amination when the SU-8 surface was treated in a 7 wt % EDA solution as shown in Figure 2b. However, a significant number of amino functionalities were introduced onto the SU-8 surface when the EDA solution concentration was increased to 20 wt % (Cond 2, 0.14 ± 0.01). As reported previously,42 the addition of a hydroxyl-containing amine such as ETAM, as a catalyst, was able to speed up the amine-epoxide reaction process, as indicated by a much stronger fluorescence signal (Cond 3, 0.54 ± 0.24). However, the fluorescence intensity was found to be unevenly distributed throughout the SU-8 surface when the ETAM catalyst was used. Alternatively, the amine-epoxide reaction could also be accelerated by heating, as evidenced by surfaces that were heat treated following the 20 wt % EDA solution treatment. The fluorescence intensities were 0.43 ± 0.05 (Cond 4), 0.54 ± 0.09 (Cond 5), and 0.71 ± 0.16 (Cond 6) for the surfaces that were heat treated at 40, 55, and 70 °C for 1 h, respectively. The heat-treated SU-8 surfaces showed relatively uniform fluorescence intensity distributions on the modified surfaces compared to those on the SU-8 surfaces catalyzed by ETAM, as evidenced by the smaller standard deviations of the surface fluorescence intensity. The fluorescence intensity was uniform and further increased to 1.1 ± 0.1 (Cond 7) for the SU-8 surface that was heat treated at 70 °C for 1 h following the 40 wt % EDA solution treatment. Because the heating and concentrated EDA are better than ETAM addition for uniformly functionalizing SU-8 surfaces, those treatments were mainly used in our work to engraft PEG further. B. PEG Engraftment. Amine reactive PEG-NHS was used to graft PEG onto aminated PDMS and SU-8 surfaces in this study. The hydrophilicity of the surfaces was characterized by water contact angle (θ) measurements. The measured water contact angles were further used to estimate the PEG surface coverage, assuming that the water contact angle of a PEG monolayer is ∼36° as reported previously in the literature.43 The water contact angles of PEG-modified PDMS and SU-8 surfaces are shown in Figures 3 and 4, respectively. For the convenience of discussion, abbreviations of PDMS(a)-APTMS(b)-PEG(c) and SU-8-EDA(d)-PEG(c) are used to represent the modified PDMS and SU-8 surfaces using different conditions. The designations a, b, c, and d indicate the air plasma treatment duration in seconds, the concentration of APTMS in weight percentage (wt %), the bulk concentration of PEG-NHS in mg/mL, and the concentration of EDA in wt %, respectively. The changes in water contact angles on PEG-grafted PDMS surfaces due to the plasma processing time (seconds) and the PEG bulk concentration are shown in Figures 3a,b, respectively. Surfaces in Figure 3a were all treated with 3 wt % APTMS and 6 mg/mL PEG. As shown in Figure 3a, the water contact angles

Figure 3. Water contact angles measured on PEG-grafted PDMS surfaces as a function of (a) air plasma processing time (using a constant 6 mg/mL for the bulk PEG concentration) and (b) bulk PEG concentration in milligrams per milliliter (using 15 s of air plasma activation). For both cases, the PDMS sample surfaces were activated by air plasma, followed by APTMS (3 wt % in toluene) silanization and PEG engraftment. Error bars represent the standard deviation of eight measured data points on each surface.

Figure 4. Water contact angles measured on a PEG-grafted SU-8 surface as a function of storage time. The SU-8 surfaces were aminated by EDA (20 and 40 wt %) with or without the addition of ethanolamine (ETAM) and then passivated by PEG engraftment (using a constant 6 mg/mL bulk PEG concentration). Error bars represent the standard deviation of eight measured data points on each surface.

are 72 ± 2, 61 ± 9, and 81 ± 3° on PDMS(15)-APTMS(3)PEG(6), PDMS(30)-APTMS(3)-PEG(6), and PDMS(60)APTMS(3)-PEG(6), respectively. This suggests that although the increase in the plasma treatment duration allowed more surface amination at the beginning, excessive plasma exposure likely caused surface deterioration because of radical bombardment. Similar observations have been reported elsewhere.35 As a result, 30 s of air plasma exposure was determined to be the 16231

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

optimal condition to use in this study. As shown in Figure 3b, the effects of the PEG concentration follows the order 20 mg/ mL (68 ± 3°) ≈ 12 mg/mL (64 ± 3°) ≈ 10 mg/mL (66 ± 4°) > 6 mg/mL (72 ± 2°) > 3 mg/mL (78 ± 2°) > 1 mg/mL (80 ± 2°) and the PEG surface coverages on those surfaces were estimated to be 53.7, 60.6, 57.0, 46.7, 34.6, and 30.9%, respectively. The relationship between the PEG surface coverage and the PEG bulk concentration appears to fit well (R2 = 0.93) with the Langmuir adsorption isotherm,28 suggesting that an increased amount of PEG bulk solution would likely result in a higher PEG surface coverage. The PEG surface coverage on PDMS surfaces reached a plateau when the concentration of PEG solution was higher than 10 mg/mL. Water contact angles of PEG-grafted SU-8 surfaces that were aminated by EDA are shown in Figure 4. As shown in Figure 4, when incubated with the same PEG bulk concentration at 6 mg/mL, the water contact angles of SU-8-EDA(20)-PEG(6) and SU-8-EDA(40)-PEG(6) were 67 ± 3 and 55 ± 2°, and the PEG surface coverages of those surfaces were estimated to be ∼38.6 and ∼57.3%, respectively. The grafted PEG layers on the SU-8 surfaces were found to be stable for at least 5 days. However, unlike in Figure 3b, increasing the PEG bulk concentration did not significantly decrease the water contact angles on the amine-functionalized PDMS surfaces. In addition, when incubated with the same PEG bulk concentration, the SU-8 surface functionalized with 40% EDA resulted in a greater PEG surface coverage than that with 20% EDA. Taken together, Figures 3 and 4 have provided valuable information, such as an optimum air plasma time and the effect of PEG and EDA concentrations on the preparation of a stable PEG layer to passivate microfluidic channels. C. Surface Characterization. To characterize PEG immobilization on the surfaces further, XPS and AFM were uilitized to acquire detailed chemical and structural information about the PEG-grafted surfaces. A pristine SU-8 surface and a PEGgrafted SU-8 surface (i.e., SU-8-EDA(40)-PEG(6)) were studied by XPS (Figure 5a). The N 1s peaks shown in these PEG-grafted SU-8 sample surfaces confirm the introduction of nitrogen in the forms of −NH2 (free amine) and C−NH− (amide bond).6 The introduced atomic percentage of nitrogen was 3.2 atom % for SU-8-EDA(40)-PEG(6). Additionally, in comparison with pristine SU-8, the PEG-grafted SU-8 surfaces showed a significant shift to a higher bonding energy (i.e., more C−O bonds and less C−C bond) in high-resolution C1s spectra, strongly suggesting the success of PEG coupling on the surfaces.33 By quantifying the relative intensity of C−O to C−C bonds in the spectra, the thickness of grafted PEG was estimated using a uniform overlay model.21 Other parameters, such as the distance between two PEG chains (L, nm), the PEG surface concentration (Γ, g/nm2), and the chain density (Σ, 1/ L 2 ) were also evaluated (data shown in Supporting Information). In the case of SU-8, the grafted PEG thicknesses on SU-8-EDA(40)-PEG(6) surfaces were estimated to be ∼2.13 nm. The fact that the thickness of grafted PEG layers was between RF (1.7 nm) and 2RF (3.4 nm), where RF is the Flory radius of a single PEG coil, strongly suggests that the PEG chain is partially stretched from a mushroom-like configuration to a brush-like configuration and that the SU-8 surface is partially covered with PEG chains.21 The PEG surface coverage on the SU-8-EDA(40)-PEG(6) sample was 2.57 × 10−21 g/ nm2. Similarly, XPS spectra of PEG-grafted PDMS surfaces also showed a significant shift of C 1s to higher bonding energy and

Figure 5. XPS spectra of high-resolution C 1s and N 1s peaks of (a) pristine and PEG-grafted SU-8 surfaces and (b) pristine and PEGgrafted PDMS surfaces. (c) Atomic percentages of C, N, O, and Si on pristine and PEG-grafted PDMS surfaces.

the emergence of the N 1s peak, as shown in Figure 5b. Approximately 2.4, 2.5, and 6.3 atom % nitrogen were introduced onto PDMS(30)-APTMS(1)-PEG(6), PDMS(30)APTMS(1)-PEG(12), and PDMS(30)-APTMS(1)-PEG(20) samples, respectively. In addition, a detailed analysis of the N 1s spectra revealed that the nitrogen was in the form of both −NH2 and C−NH− on PDMS(30)-APTMS(1)-PEG(6) and PDMS(30)-APTMS(1)-PEG(12) surfaces whereas the introduced nitrogen was likely only in the form of C−NH− on the PDMS(30)-APTM(1)-PEG(20) surface. This suggests that on the PDMS(30)-APTM(1)-PEG(20) surface likely all amino functionalities have been conjugated with PEG by forming amide bonds. In analyzing detailed C 1s spectra, it was noted that whereas C−C/C−H at 285 eV was dominant on the pristine PDMS surface (C−Si, 5.5%; C−C/C−H, 94.5%) the component of C−N/C−O (70.9−78.6%) emerged and became significant on the PEG-modified surfaces as shown in Figure 5c. The increase of the atom % of elements O and N and the decrease in the atom % of Si and C are strong evidence of successful PEG modifications on the PDMS surfaces. Furthermore, the increase in O and the decrease in Si suggest 16232

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

Figure 6. Three-dimensional topology images of (a) primitive PDMS, (b) PDMS(30)-APTMS(1)-PEG(6), (c) PDMS(30)-APTMS(1)-PEG(20), (d) primitive SU-8, (e) SU-8-EDA(20)-PEG(6), and (f) SU-8-EDA(20)-PEG(20) surfaces. The inset beside the 3D image is the corresponding 2 μm × 2 μm 2D height image. (g) Tabulated surface roughness of the surfaces in a−f as determined by AFM. Note that the hill-and-valley topography shown in part a is due to the elastic deformation of PDMS during AFM scanning.

Table 1. Effects of Drying Time, Heating Time, Temperature, Air Plasma Treatment Time, and APTMS Concentration on the Bonding Strength between PDMS and SU-8 Surfaces drying 70 mina, heating 30 min

drying 30 mina, heating 1 h

APTMS conc (wt %) bonding temp (°C)

air plasma time (s)

0.5

1

3

100

15 30 15 30

peeled offb

fracturec

fracturec

b

c

c

130 a

1

3

fracturec peeled off peeled offb

fracture fracturec

b

fracture fracturec

fracturec

c

Drying is at room temperature. SU-8 together with PDMS peeled off of the silicon substrate. 95−100% fracture in PDMS.

30 s), bonding temperature (100 and 130 °C), and heating duration (30 min and 1 h) on the PDMS/SU-8 bonding strength were investigated, as shown in Table 1. To evaluate the bonding strength, a simple test to scrape an entire piece of PDMS off the SU-8-coated silicon substrate was performed. When the bonding strength of PDMS/SU-8 is stronger than the fracture strength of PDMS (∼2.24 MPa) or the binding strength between the SU-8 layer and the silicon substrate, the test will result in either the fracture of PDMS or the peeling off of the entire PDMS/SU-8 from the silicon substrate. The test results in Table 1 suggest that the sealing of the aminated PDMS surface by the SU-8 surface is robust and reliable for the fabrication of PDMS/SU-8 microfluidic devices without leaks. B. Nonfouling Performance of Microchannels. The procedures used to fabricate microchannels had been mentioned earlier, and the drying time after APTMS coating and air plasma treatment were chosen to be 30 min and 30 s, respectively, for all microchannel fabrication. The control is the microchannel without PEG immobilization and was not processed by EDA after microchannel bonding. The relative fluorescence intensities (Figures 7) and adherent microbead number counts (Figures 8) were normalized against the respective values on control microchannels. Fluorescent BSA was used as a model protein to test protein nonspecific adsorption on the control and PEG-immobilized PDMS/SU-8 microchannels. The fluorescent images of the

an increasing PEG surface coverage with the increase in the PEG grafting solution concentrations. The surface morphology was investigated by AFM in contact mode. The surface 3D topology of pristine PDMS, pristine SU8, PEG-grafted PDMS, and PEG-grafted SU-8 surfaces are shown in Figure 6a−f. Located beside the 3D images are the corresponding 2 μm × 2 μm 2D height images. The calculated roughnesses of all surfaces are tabulated in Figure 6g. As shown in Figure 6, the pristine PDMS and SU-8 surfaces were relatively smooth (roughnesses of 1.17 and 0.23 nm, respectively) and both became rough immediately after PEG immobilization. In addition, it is interesting that both PDMS and SU-8 surfaces incubated in more concentrated PEG solution appeared to result in rougher surfaces. Furthermore, the immobilized PEG molecules organized more uniformly on SU-8 surfaces than on PDMS surfaces when incubated in PEG solutions of the same concentrations. A closer examination of the AFM 3D images showed an evenly PEG-covered surface without any polymer crystallization or visible defects on the modified surfaces. 2. Channel Surface Modification. A. Studies of Bonding Strength. The processed SU-8 without a hard baking step contained residual epoxy groups, making it possible to seal with aminated PDMS surfaces.17 The effects of the APTMS concentration (0.5, 1, and 3 wt %), drying time after APTMS coating (30 and 70 min), air plasma treatment duration (15 and 16233

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

Figure 7. Fluorescent images of PDMS/SU-8 microchannels and a test of protein adsorption using BSA. (a) Microchannels without modification, (b) microchannel modified by 20 wt % EDA and then 6 mg/mL bulk PEG, (c) microchannel modified by 20 wt % EDA and then 12 mg/mL bulk PEG, and (d) microchannel modified by 40 wt % EDA and then 6 mg/mL bulk PEG. Prior to bonding, the PDMS surfaces were treated with air plasma (30 s) and then APTMS (1 wt %). Error bars represent the standard deviation of six measured intensities from each microchannel.

control and PEG-modified microchannels (i.e., PDMS(30)APTMS(1)-SU-8-EDA(20)-PEG(6), PDMS(30)-APTMS(1)SU-8-EDA(20)-PEG(12), and PDMS(30)-APTMS(1)-SU-8EDA(40)-PEG(6)) are shown in Figure 7a−d, respectively. The fluorescence intensity of the control microchannel had more than twice the brightness as that of PEG-grafted microchannels. Because the fluorescence intensity positively correlates with the amount of BSA adsorbed on the surface,34 it is clear that PEG grafting significantly reduces the nonspecific BSA adsorption on microchannel surfaces (64.45, 60.37, and 63.91% reductions on PDMS(30)-APTMS(1)-SU-8-EDA(20)PEG(6), PDMS(30)-APTMS(1)-SU-8-EDA(20)-PEG(12), and PDMS(30)-APTMS(1)-SU-8-EDA(40)-PEG(6), respectively). However, it is also worthwhile to note that there were no significant differences among the three PEG-grafted microchannels in their nonfouling performances. This is likely due to insufficient PEG surface coverage (estimated from both contact angle and XPS measurements) to shield the microchannel surfaces from BSA adsorption efficiently. In addition, polybead and Fluoresbrite YG microbeads were used to test hydrophobic particle adhesion on microchannels with and without PEG engraftment under both stagnant and flowing conditions, the results of which are shown in Figure 8a,b, respectively. The number of polybead microbeads adhering to the walls was counted manually, with the numbers of adherent microbeads on either PDMS or the SU-8 surface individually counted using an optical microscope. In comparison to the control (i.e., surfaces without PEG modification), significant reductions in microbead adhesion were observed in PEG-immobilized microchannels (i.e., PDMS(30)-APTMS(1)SU-8-EDA(20)-PEG(6), 85.6 ± 6.9%; PDMS(30)-APTMS(1)SU-8-EDA(20)-PEG(12), 91.4 ± 2.2%; and PDMS(30)APTMS(1)-SU-8-EDA(40)-PEG(12), 92.0 ± 4.2%) as shown in Figure 8a. In addition, regardless of control microchannel or PEG immobilized microchannels, generally there were more

Figure 8. Changes in particle adhesion within microchannels of different PEG coverage. (a) Particle adhesion on PDMS/SU-8 microchannel surfaces. A microbead solution of 1 × 109 particles/ mL was incubated within microchannels for 2 min. (b) Particle adhesion on PDMS/SU-8 microchannel surfaces. A microbead solution of 1 × 108 particles/mL was allowed to flow in microchannels at 50 μL/h for 1 h. Error bars represent the standard deviation of four measured data points from each microchannel.

microbeads adhering to PDMS surfaces than to SU-8 surfaces, most likely because of the more hydrophobic nature of PDMS surfaces The behavior of microbead adhesion under continuous flow was also studied. PEG-grafted microchannels showed remarkable reductions (84.1−95.8%) in microbead adhesion as evidenced in Figure 8b. Interestingly, the PEG modification on the microchannel surfaces seemed to reduce the microbead adhesion to the same extent regardless of the surface modification parameters used. In summary, more than 90% reductions in particle adhesion can be achieved under stagnant or flowing conditions by our PEG engraftment approach. The significant suppression of particle adhesion on the microfluidic channels as a result of the PEG modification is particularly interesting for many applications. For example, this surface passivation technique can be used to modify the surfaces of microcytometers44 that have been used to detect low concentrations of food-borne pathogenic bacteria in food samples where pathogen adhesion to the microcytometer channel, in which the pathogen will not be detected by the cytometer, will result in false negative results and should be completely prevented. A linear PEG chain (1000 Da) was used to passivate the microchannel surfaces in this work. Studies are in process to improve the PEG passivation performance further to resist proteins, microbeads, and cell adhesion by engrafting longer PEG chains21,45 and branched dendrimers.33 Longer 16234

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

PEG chains have larger hydration shells46 whereas branched dendrimers possess higher extension and chain flexibility.33 It has been previously shown that the steric interactions of polymer chains are proportional to the compressed length of the molecular chains.47 Therefore, when the polymer chain is long enough strong steric interactions can effectively shield attractive interactions that contribute to nonspecific adsorption, including hydrophobic interactions, van der Waals forces, and electrostatic interactions.

(5) Yao, B.; Luo, G.-a.; Feng, X.; Wang, W.; Chen, L.-x.; Wang, Y.-m. A microfluidic device based on gravity and electric force driving for flow cytometry and fluorescence activated cell sorting. Lab Chip 2004, 4, 603−607. (6) Séguin, C.; McLachlan, J. M.; Norton, P. R.; Lagugné-Labarthet, F. Surface modification of poly(dimethylsiloxane) for microfluidic assay applications. Appl. Surf. Sci. 2010, 256, 2524−2531. (7) Yeh, P.-Y.; Rossi, N.; Kizhakkedathu, J.; Chiao, M. A siliconebased microfluidic chip grafted with carboxyl functionalized hyperbranched polyglycerols for selective protein capture. Microfluid. Nanofluid. 2010, 9, 199−209. (8) Watts, B. R.; Kowpak, T.; Zhang, Z.; Xu, C.-Q.; Zhu, S. Formation and characterization of an ideal excitation beam geometry in an optofluidic device. Biomed. Opt. Express 2010, 1, 848−860. (9) Ruano-López, J. M.; Aguirregabiria, M.; Tijero, M.; Arroyo, M. T.; Elizalde, J.; Berganzo, J.; Aranburu, I.; Blanco, F. J.; Mayora, K. A new SU-8 process to integrate buried waveguides and sealed microchannels for a lab-on-a-chip. Sens. Actuators, B 2006, 114, 542−551. (10) Jokinen, V.; Suvanto, P.; Franssila, S. Oxygen and nitrogen plasma hydrophilization and hydrophobic recovery of polymers. Biomicrofluidics 2012, 6, 016501. (11) Bu, M.; Christensen, T. B.; Smistrup, K.; Wolff, A.; Hansen, M. F. Characterization of a microfluidic magnetic bead separator for highthroughput applications. Sens. Actuators, A 2008, 145−146, 430−436. (12) Ng, J. M. K.; Gitlin, I.; Stroock, A. D.; Whitesides, G. M. Components for integrated poly(dimethylsiloxane) microfluidic systems. Electrophoresis 2002, 23, 3461−3473. (13) Meier, R. C.; Badilita, V.; Brunne, J.; Wallrabe, U.; Korvink, J. G. Complex three-dimensional high aspect ratio microfluidic network manufactured in combined PerMX dry-resist and SU-8 technology. Biomicrofluidics 2011, 5 (), 034111. (14) Yang, C.; Wang, W.; Li, Z. Optimization of corona-triggered PDMS-PDMS bonding method. Proceedings of the 4th IEEE International Conference on Nano/Micro Engineered Molecular Systems, Jan 5− 8, 2009, Shenzhen, China; pp 319−322. (15) Tsao, C. W.; DeVoe, D. L. Bonding of thermoplastic polymer microfluidics. Microfluid. Nanofluid. 2009, 6, 1−16. (16) Zhang, Z.; Zhao, P.; Xiao, G.; Watts, B. R.; Xu, C. Sealing SU-8 microfluidic channels using PDM. Biomicrofluidics 2011, 5, 046503−8. (17) Zhang, Z.; Zhao, P.; Xiao, G. Z. The fabrication of polymer microfluidic devices using a solid-to-solid interfacial polyaddition. Polymer 2009, 50, 5358−5361. (18) Zammatteo, N.; Jeanmart, L.; Hamels, S.; Courtois, S.; Louette, P.; Hevesi, L.; Remacle, J. Comparison between different strategies of covalent attachment of DNA to glass surfaces to build DNA microarrays. Anal. Biochem. 2000, 280, 143−150. (19) Ou, J.; Glawdel, T.; Ren, C. L.; Pawliszyn, J. Fabrication of a hybrid PDMS/SU-8/quartz microfluidic chip for enhancing UV absorption whole-channel imaging detection sensitivity and application for isoelectric focusing of proteins. Lab Chip 2009, 9, 1926−1932. (20) Wu, D.; Zhao, B.; Dai, Z.; Qin, J.; Lin, B. Grafting epoxymodified hydrophilic polymers onto poly(dimethylsiloxane) microfluidic chip to resist nonspecific protein adsorption. Lab Chip 2006, 6, 942−947. (21) Tao, S. L.; Popat, K. C.; Norman, J. J.; Desai, T. A. Surface modification of SU-8 for enhanced biofunctionality and nonfouling properties. Langmuir 2008, 24, 2631−2636. (22) Zhang, Z.; Feng, X.; Luo, Q.; Liu, B.-F. Environmentally friendly surface modification of PDMS using PEG polymer brush. Electrophoresis 2009, 30, 3174−3180. (23) Wong, I.; Ho, C.-M. Surface molecular property modifications for poly (dimethyl-siloxane) (PDMS) based microfluidic devices. Microfluid. Nanofluid. 2009, 7, 291−306. (24) Guo, D.-J.; Han, H.-M.; Jing, W.; Xiao, S.-J.; Dai, Z.-D. Surfacehydrophilic and protein-resistant silicone elastomers prepared by hydrosilylation of vinyl poly(ethylene glycol) on hydrosilanespoly(dimethylsiloxane) surfaces. Colloids Surf., A 2007, 308, 129−135.

IV. CONCLUSIONS We have devised a facile chemical approach to the modification of both PDMS and SU-8 surfaces to reduce surface fouling in microchannels. To modify PDMS surfaces, we find that amination levels on PDMS surfaces are mainly affected by the concentration of APTMS and air plasma activation duration. In the case of the SU-8 surface, the amination level can be accomplished by changing the EDA concentration. In addition, the grafted PEG density is closely related to the amination level of surfaces and the bulk concentration of the PEG solution. The PEG-grafted PDMS/SU-8 microfluidic devices have shown up to a 64.5% reduction in nonspecific BSA adsorption and up to 92.0 and 95.8% reductions in microbead adhesion under stagnant and flowing conditions, respectively. Longer PEG chains or branched dendrimers that are further able to improve the performance of the antifouling of proteins, microbeads, or cells will be investigated in future studies.



ASSOCIATED CONTENT

S Supporting Information *

XPS characterization of PEG grafted SU-8 surfaces. This material is available free of charge via the Internet at http:// pubs.acs.org.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The study was supported, in part, by a strategic grant from the Natural Sciences and Engineering Research Council of Canada (NSERC) to X.C. and M.L. and an Ontario Centres of Excellence (OCE) grant to X.C. and M.L. We thank Ms. Simona Moisa and Mr. Kodra Oltion for XPS measurements and Ms. Raluca Movileanu for AFM measurements.



REFERENCES

(1) Becker, H.; Gartner, C. Polymer microfabrication technologies for microfluidic systems. Anal. Bioanal. Chem. 2008, 390, 89−111. (2) Burns, M. A.; Johnson, B. N.; Brahmasandra, S. N.; Handique, K.; Webster, J. R.; Krishnan, M.; Sammarco, T. S.; Man, P. M.; Jones, D.; Heldsinger, D.; Mastrangelo, C. H.; Burke, D. T. An integrated nanoliter dna analysis device. Science 1998, 282, 484−487. (3) Thongboonkerd, V.; Songtawee, N.; Sritippayawan, S. Urinary proteome profiling using microfluidic technology on a chip. J. Proteome Res. 2007, 6, 2011−2018. (4) De Malsche, W.; Eghbali, H.; Clicq, D.; Vangelooven, J.; Gardeniers, H.; Desmet, G. Pressure-driven reverse-phase liquid chromatography separations in ordered nonporous pillar array columns. Anal. Chem. 2007, 79, 5915−5926. 16235

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236

Langmuir

Article

(46) Kingshott, P.; Thissen, H.; Griesser, H. J. Effects of cloud-point grafting, chain length, and density of PEG layers on competitive adsorption of ocular proteins. Biomaterials 2002, 23, 2043−2056. (47) Yeh, P.-Y.; Kizhakkedathu, J. N.; Chiao, M. A novel method to attenuate protein adsorption using combinations of polyethylene glycol (PEG) grafts and piezoelectric actuation. J. Nanotechnol. Eng. Med. 2010, 1, 041010.

(25) Hu, S.; Ren, X.; Bachman, M.; Sims, C. E.; Li, G. P.; Allbritton, N. Surface modification of poly(dimethylsiloxane) microfluidic devices by ultraviolet polymer grafting. Anal. Chem. 2002, 74, 4117−4123. (26) Luo, Y.; Huang, B.; Wu, H.; Zare, R. N. Controlling electroosmotic flow in poly(dimethylsiloxane) separation channels by means of prepolymer additives. Anal. Chem. 2006, 78, 4588−4592. (27) Wang, B.; Horton, J. H.; Oleschuk, R. D. Sulfonatedpolydimethylsiloxane (PDMS) microdevices with enhanced electroosmotic pumping and stability. Can. J. Chem. 2006, 84, 720−729. (28) Sikanen, T.; Wiedmer, S. K.; Heikkilä, L.; Franssila, S.; Kostiainen, R.; Kotiaho, T. Dynamic coating of SU-8 microfluidic chips with phospholipid disks. Electrophoresis 2010, 31, 2566−2574. (29) Norde, W.; Buijs, J.; Lyklema, H. Adsorption of globular proteins. In Fundamentals of Interface and Colloid Science; Lyklema, J., Ed.; Academic Press: San Diego, CA, 2005; Vol. 5, pp 1−59. (30) Mateo, C.; Fernandez-Lorente, G.; Abian, O.; FernandezLafuente, R.; Guisan, J. M. Multifunctional epoxy supports: a new tool to improve the covalent immobilization of proteins. The promotion of physical adsorptions of proteins on the supports before their covalent linkage. Biomacromolecules 2000, 1, 739−745. (31) Yang, W. C.; Sun, X. H.; Pan, T.; Woolley, A. T. Affinity monolith preconcentrators for polymer microchip capillary electrophoresis. Electrophoresis 2008, 29, 3429−3435. (32) Wang, Y. L.; Pai, J. H.; Lai, H. H.; Sims, C. E.; Bachman, M.; Li, G. P.; Allbritton, N. L. Surface graft polymerization of SU-8 for bioMEMS applications. J. Micromech. Microeng. 2007, 17, 1371−1380. (33) Yeh, P.-Y. J.; Kainthan, R. K.; Zou, Y.; Chiao, M.; Kizhakkedathu, J. N. Self-assembled monothiol-terminated hyperbranched polyglycerols on a gold surface: a comparative study on the structure, morphology, and protein adsorption characteristics with linear poly(ethylene glycol)s. Langmuir 2008, 24, 4907−4916. (34) Scopelliti, P. E.; Borgonovo, A.; Indrieri, M.; Giorgetti, L.; Bongiorno, G.; Carbone, R.; Podestà, A.; Milani, P. The effect of surface nanometre-scale morphology on protein adsorption. PLoS ONE 2010, 5, e11862. (35) Wu, C.-C.; Yuan, C.-Y.; Ding, S.-J. Effect of polydimethylsiloxane surfaces silanized with different nitrogen-containing groups on the adhesion progress of epithelial cells. Surf. Coat. Technol. 2011, 205, 3182−3189. (36) Siow, K. S.; Britcher, L.; Kumar, S.; Griesser, H. J. Plasma methods for the generation of chemically reactive surfaces for biomolecule immobilization and cell colonization - a review. Plasma Process. Polym. 2006, 3, 392−418. (37) Ouellet, E.; Yang, C. W. T.; Lin, T.; Yang, L. L.; Lagally, E. T. Novel carboxyl-amine bonding methods for poly(dimethylsiloxane)based devices. Langmuir 2010, 26, 11609−11614. (38) Qvortrup, K.; Taveras, K. M.; Thastrup, O.; Nielsen, T. E. Chemical synthesis on SU-8. Chem. Commun. 2011, 47, 1309−1311. (39) Scrimgeour, J.; Kodalj, V. K.; Koyari, D. T.; Curtis, J. E. Photobleaching-activated micropatterning on self-assembled monolayers. J. Phys.: Condens. Matter 2010, 22, 194103. (40) Holländer, A. Labelling techniques for the chemical analysis of polymer surfaces. Surf. Interface Anal. 2004, 36, 1023−1026. (41) Lee, J. N.; Jiang, X.; Ryan, D.; Whitesides, G. M. Compatibility of mammalian cells on surfaces of poly(dimethylsiloxane). Langmuir 2004, 20, 11684−11691. (42) Shechter, L.; Wynstra, J.; Kurkjy, R. P. Glycidyl ether reactions with amines. Ind. Eng. Chem. 1956, 48, 94−97. (43) Cai, Y.; Yun, Y. H.; Zhang Newby, B.-m. Generation of contactprinting based poly(ethylene glycol) gradient surfaces with micrometer-sized steps. Colloids Surf., B 2010, 75, 115−122. (44) Mu, C.; Zhang, F.; Zhang, Z.; Lin, M.; Cao, X. Highly efficient dual-channel cytometric-detection of micron-sized particles in microfluidic device. Sens. Actuators, B 2011, 151, 402−409. (45) Zhang, Z.; Feng, X.; Xu, F.; Liu, X.; Liu, B.-F. Click chemistrybased surface modification of poly(dimethylsiloxane) for protein separation in a microfluidic chip. Electrophoresis 2010, 31, 3129−3136. 16236

dx.doi.org/10.1021/la303196m | Langmuir 2012, 28, 16227−16236