Sensors andActuators B 81 -2001) 49±54
Micromachining of non-fouling coatings for bio-MEMS applications Yael Haneina,*, Y. Vickie Panb,1, Buddy D. Ratnerb,c, Denice D. Dentona, Karl F. BoÈhringera aDepartment of Electrical Engineering, University of Washington, Seattle, WA 98195, USA bDepartment of Bioengineering, University of Washington, Seattle, WA 98195, USA cDepartment of Chemical Engineering, University of Washington, Seattle, WA 98195, USA Received2 April 2001; receivedin revisedform 12 August 2001; accepted18 August 2001
Abstract
Standard photolithography is used to pattern a poly -ethylene glycol) -PEG)-like polymer onto silicon substrates. The coating has excellent non-fouling properties and good adhesion to various substrate materials, such as silicon, oxide, nitride, gold, and platinum. This methodallows precise control of the shape, size andalignment of the polymer, thus providinga reliable tool to pattern protein sheets as well as cell cultures. This method also enables the incorporation of patterned cell cultures with various prede®ned elements such as electrodes, channels, and sensors. To demonstrate the properties of our technique, we apply it to build cell cultures and to protect metallic electrodes from protein and cell adhesion. We show that the thin coatings provide excellent protection without compromising the conductivity of the electrodes. # 2001 Elsevier Science B.V. All rights reserved.
Keywords: Bio-MEMS; Bio-fouling; Proteins; Cell cultures
1. Introduction example, reducing its life span or increasing its power consumption [6]. The rapidgrowth of micro electro mechanical systems A convenient approach to control protein adsorption is by -MEMS) research in the last few decades has given rise to a surface modi®cation [7]. This approach is particularly con- wide variety of novel applications. Among the most notable venient for bio-MEMS development since it allows the use are small, portable systems for medical and biological of commonly usedsubstrates andmaterials -as long as they applications -bio-MEMS). These devices are currently are non-toxic), andrequires only a ®nal coating procedure. developed with the vision of building miniaturized, fast, Many common non-fouling coating techniques are based and portable analytical instrumentation for drug delivery, on poly -ethylene glycol) -PEG) [8,9]. PEG is a hydrated electrochemical monitoring [1] andDNA analysis [2]. These [10], non-toxic coating andFDA-approvedfor use in bio- devices would speed the analysis time for researchers in technology andconsumer applications. In recent years, the lab, for medical staff at hospitals or might be used as various methods have been adopted to coat surfaces with carry-on devices by chronically ill patients. PEG, including silicon surfaces [11]. In our study, we use The interaction of a device with a biological environment glow discharge plasma polymerization of tetraglyme leads to various challenges that have to be taken into account -CH3ÀOÀ-CH2ÀCH2ÀO)4ÀCH3) to obtain a PEG-like in order to allow its proper operation. A major issue is bio- material. Plasma polymerization is a gas-phase, thin-®lm fouling, the strong tendency of proteins and organisms to deposition process. This process is capable of producing thin physically adsorb to synthetic surfaces [3]. The adsorbed coatings -of the order of 10 nm) with superb surface con- protein layer tends to create undesired perturbations to the formality [10] which is an extremely important factor in operation of devices such as pH [4] and glucose [5] sensors. determining the success of the PEG coating [12]. The excel- In addition, the protein layer can mediate various biological lent surface coverage of the plasma polymerization process responses such as cell attachment andactivation, which may also contributes to its excellent non-fouling properties. interfere with the optimal operation of the device by, for In previous studies, it was demonstrated that plasma polymerizedtetraglyme -pp4G) coatedsurfaces show good resistance to protein adsorption [10]. The focus of our * Corresponding author. E-mail address: [email protected] -Y. Hanein). current study is to examine the compatibility between the 1 Present address: Cellomics, Pittsburgh PA, USA. pp4G coatings andphotolithography andmicro-machining
0925-4005/01/$ ± see front matter # 2001 Elsevier Science B.V. All rights reserved. PII: S 0925-4005-01)00925-X 50 Y. Hanein et al. / Sensors and Actuators B 81 +2001) 49±54 processes in order to allow pp4G patterning as a standard for our experiments because it has better corrosion resis- MEMS process. In this paper, we focus on devices with a tance in aqueous environments than oxide [14]. To achieve topography suitable for photolithography. Such conditions the pp4G patterning, photoresist -Clariant AZ 1512, 1.5 mm) are common for many Bio-MEMS devices which consist of was deposited, and then exposed and developed in order to two bonded elements with at least one suitable for photo- create openings in the resist for the later pp4G lift-off. After lithography. Patterning the pp4G directly on high aspect a plasma oxygen descum etch of 15 s in an RIE chamber ratio structures is limitedby the ability to achieve uniform -Trion Technology), the wafers were introduced to the photoresist coatings on these structures. However, such plasma reactor for the pp4G deposition. The substrates were coatings may be achievedby using very thick photo-resist brie¯y cleanedin an argon plasma, the tetraglyme vapor was or a special application procedure [16]. introduced to the chamber and the samples were coated. The We describe below the integration of pp4G coatings into details of the monomer preparation and the deposition microfabrication processes for bio-MEMS applications. We process will be discussed elsewhere [15]. Following the characterizedthe coatings using environmental scanning plasma deposition process the wafers were soaked in acetone electron microscopy -ESEM) which we usedto verify the an ultrasonic bath for approximately 20 s. Finally, the wafers pattern de®nition. Electron spectroscopy for chemical ana- were rinsedwith methanol anddriedwithnitrogen. The lysis -ESCA), a surface sensitive analytical technique, was microfabrication processes were done at the Washington also utilizedto analyze the coating surface chemistry after Technology Center -WTC) andat the Electrical Engineering the photoresist lift-off. We foundthat the patterning process Micro-fabrication Labs at the University of Washington. The did not affect the pp4G coating. pp4G deposition was done at the University of Washington Finally, we discuss experiments which demonstrate the EngineeredBiomaterials -UWEB) labs. capabilities of our technique: A protein adsorption experi- Integrating the lift-off andthe plasma deposition pro- ment demonstrates the spatial control of these interactions cesses was rather straightforward. However, special atten- on patternedpp4G surfaces. We use the patternedpp4G tion hadto be directedto the location of the wafer in the coatings to con®ne cell cultures to speci®c regions on the plasma chamber. We placedthe wafers in low monomer substrate. The advantages over contemporary methods that condensation regions to prevent tetraglyme vapor from consist of soft lithography [13] are discussed. The pp4G was dissolving the photoresist. After the plasma deposition, also testedas a protection layer for electrodesin electro- the wafers were rinsedin acetone -in an ultrasonic bath) chemical sensing. to remove resist leftovers.
2.1. Lift-off results 2. Fabrication To discuss the quality of our results we begin with a visual We usedsilicon wafers as substrates for the pp4G coat- examination of the pp4G coating after lift-off. In Fig. 1a and ings. The wafers were sputteredwith silicon nitrideor b, we show ESEM images of patternedpp4G coatings on oxidized for electrical passivation. Nitride is better suited thermally oxidized substrates. Here, arrays of Au squares
Fig. 1. ESEM images of pp4G features after lift-off. The scale bar is 200 mm for both images. -a) The pp4G structure -left) alignedto predefinedCrAu structure. Lift-off without ultrasonic power. -b) ExposedSiO 2 substrate -bright square) with pp4G background. Lift-off is done in an ultrasonic bath. Similar results were obtainedfor structures as small as 10 mm. While current results were produced with low-cost printed masks -features greater than 10 mm), minimum feature size andalignment precision for these protein-resistant coatings is limitedonly by the lithographic resolution. We recently achi evedsimilar results with 2 mm resolution. Y. Hanein et al. / Sensors and Actuators B 81 +2001) 49±54 51
Fig. 2. A fluorescence microscope image of a patternedprotein-resistant pp4G pad-1500 mm  1500 mm) with Au pads -200 mm  200 mm) on SiO2 substrate after incubation in a solution with fluorescently labeled proteins. The pp4G pad appears dark as the adsorbed proteins cover exposed SiO2 andAu areas while pp4G coatedareas resist protein adsorption. were ®rst patternedusing the lift-off technique. Then the the patterning process does not affect the chemistry of the pp4G was deposited and patterned using lift-off. Using a pp4G coating andthat there are no organic residues left on stylus stepper -Alpha-Step), the pp4G coating was measured the exposed or coated surfaces. The ESCA data are dis- to be approximately 20 nm. In Fig. 1a the pp4G padis shown cussedin detailelsewhere [15]. with a prede®ned chrome±gold -CrAu) feature. To perform To assess the effectiveness of the coating, we performed the pp4G lift-off, the sample in Fig. 1a was dipped in acetone protein adsorption tests. For this experiment we used sam- for several minutes and then rinsed in methanol and dried. ples similar to the samples in Fig. 1. In Fig. 2, we show Note the pp4G leftovers at the edges of the pattern. This is a ¯uorescence microscopy image of adsorbed ¯uorescently result of the conformal nature of the deposition process. It is labeled-Oregon Green, Molecular Probes) bovine serum important to note that conformality may be useful for albumin -BSA) on a patternedsurface. The sample consist of channel coating purposes andfor various three-dimensional an array of small, 200 mm  200 mm, Au pads -bright structures where shadowing is a major challenge. squares) andbig, 1500 mm  1500 mm pp4G pads -dark To improve the edge de®nition of the patterned pp4G we areas) on a SiO2 substrate. The adsorption experiment was usedan ultrasonic bath duringthe lift-off process -see carriedout at room temperature in citratedphosphate buffer Fig. 1b). As can be clearly seen in Fig. 1b, the edge de®nition solution at pH 7.4. After incubation -for 1.5 h) the surface is signi®cantly improvedcomparedto the sample in Fig. 1a. was thoroughly rinsedin a buffer solution. It is readilyseen, In addition, the adhesion of the coating is good enough to in Fig. 2, that the proteins non-speci®cally adsorbed only to keep the coating intact despite the aggressive ultrasonic exposedSiO 2 areas -bright areas) whereas the pp4G coated treatment. High resolution ESEM imaging andthe coating regions appear dark in the absence of adsorbed proteins. performances -discussed later in the text) demonstrate the viability of the coating despite the ultrasonic treatment. We can now summarize the advantages of our process: the 3. Applications coating has goodadhesion to the substrate andto the metal, it is conformal, it can be reliably patternedandcan be Micro-machinednon-fouling coatings can be useful for incorporatedonto complex structures using standardlitho- various applications, in particular when incorporatedinto graphy andlift-off processes. complex MEMS structures. Here, we choose to focus on two issues: patternedcell cultures andimprovedbiocompatibil- 2.2. Surface chemistry ity for electrochemical electrodes.
The next step in characterizing the patternedpp4G coat- 3.1. Integrated micro-patterned cell cultures ings is to study their surface chemistry. The surface chem- istry of pp4G on SiO2 andon metal layers was examined In Section 2, we demonstrated our ability to pattern using ESCA. The ESCA characterization demonstrated that protein layers. An additional, related application is micro- 52 Y. Hanein et al. / Sensors and Actuators B 81 +2001) 49±54 patterning of cell cultures. It is widely known that proteins by the patternedpp4G. The cell cytoskeleton, f-actin, stained mediate cell adhesion. There are two general classes of cell with rhodamine phalloidin is false-colored red -appearing adhesion mechanisms: non-speci®c attachment -due to dis- bright). DNA inside the cell nucleus, stained with DAPI, is persion forces andthe hydrophobic effect) andspeci®c pro- false-coloredblue -appearing dark).Detailedinformation on tein-cell interactions. We use the selectivity of our patterned the cell experiments is discussed elsewhere [15]. surface to protein adsorption to enable selective cell adhesion. Micro-patternedcell culture devices are useful for the 3.2. Protection layer for electrochemical electrodes study of cellular activity and development. Such devices allow careful monitoring in a well de®ned environment by The hydrated nature of the pp4G coating and its thickness controlling the size of cell-adhesive islands [17] or the suggest its potential as an inert protection layer for electro- con®guration of neural networks [18]. Micro-patternedcell chemical electrodes. Good ionic conductivity is important cultures couldalso be useful when incorporatedinto more for various electro-chemical sensing applications such as complex structures such as micro-electrode arrays [19], micro-electrodes for neural recording [20]. We verify the micro-channels or micro-analyzers. Contemporary methods conductivity properties of pp4G by comparing electroplated to realize patternedcell surfaces consist of soft lithography electrodes with and without pp4G coatings. The pp4G coating [13]. While this technique offers the advantage of simplicity was integratedinto the following process: a CrAu seedlayer andthe ability to pattern unusual materials over large areas, for an electroplating step was sputteredon a nitridelayer. it is mainly limitedto single-layer structures dueto limited Photoresist was then patternedandthe wafers were electro- alignment capabilities. Our technique to incorporate cell platedwith platinum. The electroplating process was carried domains into MEMS devices using standard photolithogra- out in ultrasonic bath in chloroplatinic acidsolution -1 g of phy is demonstrated next. PtCl6H2 to 100 ml H2O) with Pt wire as a counter electrode. For the cell culture experiment we usedline shaped The dc current density was ®xed at approximately 5 nA/mm2 openings in pp4G on SiO2. The bovine aortic endothelial for 1 min. Next, the wafers were rinsedin acetone to remove cells -BAECs) we usedin the experimet were culturedin the photoresist. A secondlayer of photoresist was applied Iscoves Modi®ed Dulbeccos Medium supplemented with andpatternedandthe seedlayer was etchedin goldetchant. 10% fetal bovine serum, 1% 100 mM MEM sodium pyr- A thirdlithography process de®nedthe pp4G patterns. uvate solution, 1% 10 mM MEM non-essential amino acids Our devices consist of two sets of six square electrodes solution, 100 units/ml penicillin and100 mg/ml streptomy- with different sizes -250±2250 mm side lengths). One set is cin. BAECs were culturedon 75 cm 2 tissue culture ¯asks coated-Fig. 4a) while the secondis bare. To demonstrate the inside a humidi®ed incubator at 37 Cwith5%CO2 and9% effectiveness of the coating we show images of the electro- air. The cells were culturedon the patternedpp4G surfaces deswith andwithout pp4G -Fig. 4b andc, respectively) after for 4 days. After 4 days in culture, the cell cultures were 24 h in a solution with BAECs. By slightly shaking the dish, ®xed and stained. Fig. 3 shows lines of cell cultures de®ned the suspended cells from the coated areas are removed and
Fig. 3. Cell patterns achieved by culturing BAECs on line-shaped oxide domains with a pp4G background. The thinner line is 100 mm thick. Despite the multilayer formation of these cells on the substrate, they remain outside the pp4G boundary. Y. Hanein et al. / Sensors and Actuators B 81 +2001) 49±54 53
Fig. 4. -a) Electroplated platinum electrodes -dark squares) with pp4G coating on silicon nitride surface. The two smallest electrodes --b) bare and -c) coated) after 24 h in a medium solution with BAECs. The pp4G coating -in c) prevents cells from adhering to the electrode surface. the electrodes are exposed -seen in Fig. 4c). In comparison, adhesion while maintaining the electrode conductivity prop- without the protection of the pp4G coating -Fig. 4b) the erties. The effects of protein adsorption and cell adhesion on entire surface is coveredwith cells. Since no special effort electrode performance are known to occur during periods of was made to rinse the access cells form the coating, some hours [4], days [5] or even months. These effects may vary cells remain on the coatedarea andare still attachedto the for different applications and biological models. Further in cells that adhered to the edges of the pp4G coating. vivo experimental study with coated and bare electrodes The ac anddcresistivity of various bare andcoated couldclarify the signi®cance of the coatings to issues such as samples was measured. In Fig. 5, we show a typical ac tissue irritation, signal to noise ratio, signal stability, and -10 Hz) conductance versus size plot of a coated electrode signal distortion for various applications. set. The derived resistivity is 5 Â 109 Omm2. The resistivity for the bare set was foundto be 8 Â 109 Omm2. In other tests, including measurements in clean medium solutions and in 4. Conclusions solutions with proteins, similar results were obtained. The slight consistent difference in favor of the electrodes with In summary, using standard photolithography and a plasma coating can be attributedto some fouling of the bare electro- polymerization process we were able to realize patterns of des either during the processing or during the measurement. non-fouling coatings on silicon substrates. Key features of The results discussed above demonstrate the ability of this non-fouling technology include excellent coating quality, the pp4G to protect the electrodes from protein and cell goodbiocompatibility, andease of incorporation into MEMS fabrication. Because pp4G is patternedby a photoresist lift- off process, high alignment precision andgoodspatial reso- lution are attainable. This enabling technique, thus, offers precise spatial control of protein adsorption and cell attach- ment with goodalignment to underlying features. Such precision can be a great advantage in engineering the surface chemistry of bio-MEMS, such as biosensors andmicro¯uidic systems, to better interface devices with the biological envir- onment. Here, we demonstrated the usefulness of our tech- nique to build domains of proteins, as well as domains of cell cultures. We also demonstrated the effectiveness of the coating in protecting electro-chemical electrodes.
Acknowledgements
Fig. 5. Conductance data at 10 Hz plotted vs. electrode size for a set of pp4G coated electroplated electrodes in a medium solution with cells. The authors wish to thank Mark Holl, Mark Troll, Russell Similar results were obtainedfor bare electrodes. Wyeth andDennis Willows for very useful discussions. This 54 Y. Hanein et al. / Sensors and Actuators B 81 +2001) 49±54 work was supportedin part by the NSF±ERC program Grant [8] N.B. Graham, N.E. Nwachuku, D.J. Walsh, Interaction of poly-ethylene # EEC-9529161. ESCA Experiments done at NESAC/BIO oxide) with solvents. 1. Preparation and swelling of a crosslinked poly-ethylene oxide) hydrogel, Polymer 23 -1982) 1345±1349. were supportedby NIH NCRR Grant RR-01296. Work in [9] J.D. Andrade, V. Hlady, S.-I. Jeon, Poly-ethylene oxide) and protein the UW MEMS lab by Y. Hanein andK. BoÈhringer was resistance in hydrophilic polymers, performance with environmental supportedin part by DavidandLucile PackardFoundation acceptance, Advances in Chemistry, Vol. 248, American Chemica Grant 2000-01763, DARPA Bio:Info:Micro Grant MD Society, 1996. A972-01-1-002, Washington Technology Center awards [10] G.P. Lopez, B.D. Ratner, C.D. Tidwell, C.L. Haycox, R.J. Rapoza, WTC MEMS 99-8 and99-A, NSF CISE Postdoctoral T.A. Horbett, Glow discharge plasma deposition of tetraethylene glycol dimethyl ether for fouling-resistant biomaterial surfaces, J. Research Associateship EIA-0072744 to Y. Hanein, NSF Biomed. Mater. Res. 26 -1992) 415±439. Career AwardECS-9875367 to K. BoÈhringer andby Micro- [11] M. Zhang, T. Desai, M. Ferrari, Proteins andcells on PEG soft Research, Intel Corporation, andTanner Research Inc. immobilizedsilicon, Biomaterials 19 -1998) 953±960. [12] K.L. Prime, G.M. Whitesides, Self-assembled organic monolayers: model systems for studying adsorption of proteins at surfaces, Science 252 -1991) 1164. References [13] Y.N. Xia, J.A. Rogers, K.E. Paul, G.M. Whitesides, Unconventional methods for fabrication and patterning nanostructures, Chem. Rev. 99 [1] D. Papageorgiou, S.C. Bledsoe, M. Gulari, J.F. Hetke, D.J. Anderson, -1999) 1823±1848. K.D. Wise, A shutteredprobe with in-line flowmeters for chronic in [14] C. Christensen, P. Kersten, S. Henke, S. Bouwstra. Wafer through vivo drug delivery, in: Proceedings of the 14th IEEEE International hole interconnections with high vertical wiring densities, IEEE Trans. Conference on MEMS, Piscataway NJ, USA, 2001. Componensts, Packaging Manufac. Technol. 19 -1996) 516-522. [2] M.A. Northrup, B. Benett, D. Hadley, P. Landre, S. Lehew, J. [15] K. Dyrbye, T.R. Brown, G.F. Eriksen, Packaging of physical sensors Richards, A miniature DNA-based analytical instrument based on for aggressive media applications, J. Micromech. Microeng. 6 -1996) micromachinedsilicon reaction chamber, Anal. Chem. 70 -1998) 187±192. 918±922. [16] Y.V. Pan, Y. Hanein, D. Leach-Scampavia, K.F. BoÈhringer, D.D. [3] T.A. Horbett, Protein Adsorption on Biomaterials in Biomaterials: Denton, B.D. Ratner, Microscale Cell Patternig on a Non-fouling Interfacial Phenomena andApplications, American Chemical Surface with Micro-Device Applications. Plasmasr Polymers, Sub- Society, Washington, DC, 1982. mitted, 2001. [4] A.H. Auerbach, B.R. Soller, R.A. Peura, Sources of error in [17] C.S. Chen, M. Mrksich, S. Huang, G.M. Whitesides, D.E. Ingber, measuring pH with microsensors, in: Proceedings of the 20th Annual Geometric control of cell life anddeath, Science 276 -1997) 1425±1428. Northeast Bioengineering Conference, New York, NY, USA, 1994, [18] D.W. Branch, J.M. Corey, J.A. Weyhenmeyer, G.J. Brewer, B.C. p. 108. Wheeler, Microstamp patterns of biomolecules for high-resolution [5] Y. Yang, S.F. Zhang, M.A. Kingston, G. Jones, G. Wright, S.A. neuronal netwroks, Med. Biol. Eng. Comput. 36 -1998) 135±141. Spencer, Glucose sensor with improvedhaemocompatibility, Bio- [19] C.D. James, R. Davis, M. Meyer, A. Turner, S. Turner, G. Withers, L. sens. Bioelectron. 15 -2000) 221±227. Kam, G. Banker, H. Craighead, M. Issacson, J. Turner, W. Shain, [6] M. Stelzle, R. Wagner, W. Jagermann, R. Frohlich, Ultrathin organic Alignedmicrocontact printing of micrometer-scale poly- l-lysine films for biocompatible low-impedance pacemaker electrodes, in: structures for controlledgrowth of culturedneurons on planar micro- Proceedings of the 18th Annual International Conference of the IEEE electrode arrays, IEEE Trans. Biomedi. Eng. 47 -2000) 17±21. on Bridging Disciplines for Biomedicine, Vol. 1, 1997, p. 114. [20] G.T.A. Kovacs, Introduction to the Theory, Design, and Modeling of [7] B.D. Ratner, A.S. Hoffman, Thin-films, Grafts, andCoatings in Thin-film Micro-electrodes for Neural Interfaces in Enabling Biomaterials Science: An Introduction to Materials in Medicine, Technologies for CulturedNeural Networks, Academic Press, San Academic Press, San Diego, 1996. Diego, 1994.