PAPER www.rsc.org/loc Lab on a Chip Fabrication of complex multilevel microchannels in PDMS by using threedimensional photoresist masters Kwang-Seok Yun* a and Euisik Yoon b Received 22nd August 2007, Accepted 21st November 2007 First published as an Advance Article on the web 3rd December 2007 DOI: 10.1039/b712932g This paper demonstrates a new method of implementing complex microchannels in PDMS, which is simply constructed using three-dimensional photoresist structures as a master mold for the PDMS replica process. The process utilizes UV-insensitive LOR resist as a sacrificial layer to levitate the structural photoresist. In addition, the thickness of photoresist structures can be controlled by multi-step UV exposure. By using these techniques, various three-dimensional photoresist structures were successfully implemented, including the recessed cantilevers, suspended bridges, and the complex plates with micro-pits or micro-villi. We demonstrate that the three-dimensional photoresist structures are applicable to implementing complex multiple microchannels in PDMS by using the PDMS replica method. 1. Introduction In the last few years, polydimethysiloxane (PDMS) has been largely employed for implementation of microfluidic devices. 1 6 The advantages of PDMS are many: 7 optical transparency in a broad range of spectrum, bio and chemical compatibility with safe use, 8,9 low cost, easy and superior bonding property, 10 low water permeability 11,12 and simple processing using the micro molding technique. 7,13 The typical PDMS molding is a process of casting PDMS prepolymer into structured PDMS in solid state, which is composed of three steps: (1) pouring of PDMS prepolymer onto the structured master, (2) thermal curing, and (3) peeling-off of cured PDMS from the master. The structured PDMS can be utilized in a further process through bonding with other plates, such as silicon, glass, plain PDMS, and other structured PDMS. 7,13,14 For casting of PDMS prepolymer, a variety of masters have been employed, including structured silicon or glass plate, electroplated metal on substrate, and most frequently, a patterned photoresist on substrate because of its easy formation on substrate by simple photolithography. Although a simple single-level microchannel in PDMS can be easily formed by those masters, there are growing demands for more complex multilayer structures, such as multilevel microchannels or buried microchannels which are indispensable to implement integrated microfluidic components, 15 micro analytical systems, 16 and integrated microfluidic networks. 17 To date, the buried PDMS channel or multilevel PDMS channel structures have been formed by stacking and bonding of patterned PDMS plates. 5,6,14,15,17 19 These methods require an alignment between PDMS plates, which is not compatible with conventional alignment tools and is easily susceptible to a Department of Electronic Engineering, Sogang University, 1 Shinsoodong, Mapo-gu, Seoul, Korea. E-mail: ksyun@sogang.ac.kr b Department of Electrical and Computer Engineering, University of Minnesota, 5-125 EE/CSci Bldg, 200 Union Street S.E., Minneapolis, MN 55455, USA large misalignment. In addition, the via structures, which connect microchannels in different levels, are formed by physical clamping 5,14,18 or spin coating of PDMS prepolymer. 20 This requires a labor-intensive PDMS process, resulting in poor reproducibility. The other method to obtain multilevel microchannels is to utilize the three-dimensional suspended photoresist structures as masters for PDMS replica processes. It has been reported that a negative photoresist, such as SU-8, can easily be made into three-dimensional structures, such as suspended cantilevers or buried channels using photolithographic technology. 21 25 However, such a negative photoresist is hard to remove, and this difficulty restricts its application to sacrificial master structures for PDMS structures formation. Although there are several methods of obtaining three-dimensional masters through three-dimensional MEMS techniques, most of them have the same problem that the negative photoresist has. To address these issues, our research group has previously developed and reported a technique of forming suspended three-dimensional positive photoresist structures on a glass wafer using backside UV exposure. 26 However, this method can be adopted only on transparent substrate, and it requires additional metal deposition/patterning for the backside UV mask. Another problem is a long photoresist development time. The fabrication technique for similar three-dimensional photoresist structures on an opaque substrate like silicon has also been reported on. 27 We adopted LOR resist (Microchem. Co.), which is used for the lift-off process as a sacrificial photoresist. We have fully utilized the properties of LOR resist: insensitivity to UV exposure and selective development in some specific developers. This paper reports on the details of a fabrication technique for the formation of various three-dimensional positive photoresist structures. This technique can be generally applied without limitation in substrate types. Also as an application of this technique, the formation of three-dimensional PDMS microchannel structures will be demonstrated. This journal is ß The Royal Society of Chemistry 2008 Lab Chip, 2008, 8, 245 250 245
Table 1 Fabrication process recipe for the measurement of development rate of LOR LOR coating (LOR 10B) 1. Spin coating of LOR resist at 700 rpm for 10 s. 2. Baking on hot plate at 150 uc for 1 min. 3. Spin coating of LOR resist at 700 rpm for 10 s. 4. Baking on hot plate at 150 uc for 1 min. Thin PR coating and photolithography (AZ 6612 K) 1. Spin coating of AZ 6612 K on LOR layer at 4000 rpm for 30 s. 2. Soft bake on hot plate at 85 uc for 1 min. 3. Exposure: 45 mj cm 22. 4. Development in (AZ 340 : DI water = 1 : 5) for about 30 s. Development of LOR 2. Three-dimensional photoresist structures 2. 1. Experiment The three-dimensional photoresist structures were fabricated by using LOR as a sacrificial layer and multi-step exposure of thick photoresist. To measure the development rate of LOR experimentally, samples were prepared where the LOR is undercut in the developer. The details of procedures are described in Table 1. We used LOR 10B (MicroChem Co.) as an undercut resist and AZ 6612 K (Clariant) as a masking resist. The target thicknesses of resist are 30 mm for LOR and 1. Development in (AZ 400 K : DI water = 1 : 4). 2. Rinse with DI water and dry. Fig. 1 Test of LOR development rate. (a) Cross-sectional SEM picture of LOR after 5 min development in AZ 400K developer with a thin AZ 6612K photoresist as a mask. (b) Diagram showing lateral development depth of LOR vs. development time. Linear development rate of 5.7 mm min 21 was measured. Fig. 2 Fabrication process for three-dimensional positive photoresist structures. (a) Spin coating of LOR on silicon substrate. (b) Photolithography of thin photoresist using AZ 340 developer. (c) Development of unmasked LOR in AZ 400K developer. (d) UV exposure and dissolution of thin photoresist in AZ 340. (e) Spin coating of thick photoresist. (f) Shallow exposure. (g) Deep exposure. (h) Development of exposed thick photoresist and whole LOR in AZ 400K. 246 Lab Chip, 2008, 8, 245 250 This journal is ß The Royal Society of Chemistry 2008
1.7 mm for AZ 6612 K, respectively. We tested two different solutions as developer; one is AZ 400 K (Clariant), which is a potassium-based developer, and the other is AZ 340 (Clariant), which is a sodium-based developer. The development rates were measured by examining the undercut with a scanning electron microscope (SEM) for various development times (Fig. 1). Fig. 1 (a) shows the cross-sectional SEM picture of LOR that is masked with a thin AZ 6612 K photoresist after 5 min development in AZ 400 K developer. The development depth was determined at the mean value of the undercut morphology for various development times. As shown in Fig. 1 (b), the development rate of LOR in AZ 400 K developer is about 5.7 mm min 21. However, a similar experiment using an AZ 340 developer shows that LOR is hardly developed in this developer, showing a development rate of less than 1.5 mm min 21. The three-dimensional photoresist structures were fabricated by using the property of LOR and positive photoresist, Table 2 Proposed fabrication process recipe for three-dimensional positive photoresist structures. LOR coating (LOR 10B) Fig. 2 (a) i.e., in the case of the positive photoresist, only the UVexposed portion was dissolved in the developer AZ 400 K or AZ 340, but the LOR was developed in AZ 400 K regardless of the UV exposure. The proposed fabrication processes and recipe for three-dimensional photoresist structures are shown in Fig. 2 and Table 2. First, the 30 mm-thick sacrificial LOR layer was obtained by double spin coatings of LOR 10B (MicroChem Co.) resist on silicon substrate (Fig. 2 (a)). Then, a thin positive photoresist, AZ 6612 K, was spin coated to the thickness of about 1.7 mm. After that, some areas were UVexposed and developed in a sodium-based developer, AZ 340, which dissolves only exposed thin PR layer without developing LOR, as shown in Fig. 2 (b). Then, the unmasked LOR area was developed in a potassium-based developer, AZ 400 K (Fig. 2 (c)). Next, the entire area of the wafer was exposed to UV light without a mask, and only the thin positive photoresist was selectively developed in a AZ 340 developer, in which the LOR resist was hardly dissolved, as in Fig. 2 (d). Now, the thick positive resist, AZ 9260 (Clariant), as a structure material, was spin-coated with a thickness of about 75 mm on the patterned LOR resist, as in Fig. 2 (e). A shallow 1. Spin coating of LOR resist at 700 rpm for 10 s. 2. Baking on hot plate at 150 uc for 1 min. 3. Spin coating of LOR resist at 700 rpm for 10 s. 4. Baking on hot plate at 150 uc for 1 min. Thin PR coating and photolithography (AZ 6612K) Fig. 2 (b) 1. Spin coating of AZ 6612K on LOR layer at 4000 rpm for 30 s. 2. Soft bake on hot plate at 85 uc for 1 min. 3. Exposure: 45 mj cm 22. 4. Development in (AZ 340 : DI water = 1 : 5) for about 30 s. Development of LOR Fig. 2 (c) 1. Development in (AZ 400K : DI water = 1 : 4) for about 4 min 30 s. 2. Rinse with DI water and dry using nitrogen blow. Removal of thin AZ 6612K layer Fig. 2 (d) 1. Exposure: 45 mj cm 22. 2. Development in (AZ 340 : DI water = 1 : 10) for 1minand30s. Spin coating of thick PR (AZ 9260) Fig. 2 (e) 1. 0 rpm 2(0.5 s) A 1500 rpm, spin for 1.2 s 2(0.5 s) A 0 rpm. 2. Stabilization on flat table for 60 min. 3. 1st soft bake on hot plate at 85 uc for 70 min. 4. Edge bead removal by using acetone spray on the edge of wafer. 5. 2nd soft bake on hot plate at 110 uc for 2 min. 6. Water absorption in humidified chamber (humidity of 70%) for 50 min. Exposure Fig. 2 (f)(g) 1. Shallow exposure: 600 mj cm 22. 2. Deep exposure: 3600 mj cm 22. Development of exposed PR and entire LOR 1. Development in (AZ 400K : DI water = 1 : 4) for about 30 to 40 min. 2. Rinse with DI water and dry using nitrogen blow. Fig. 2 (h) Fig. 3 SEM pictures of suspended three-dimensional photoresist structures. The levitation height is about 30 mm and recessed depth by shallow exposure is about 25 mm. Suspended bridges with various lengths and suspended plates with micro-pits and micro-villi are successfully formed. This journal is ß The Royal Society of Chemistry 2008 Lab Chip, 2008, 8, 245 250 247
UV exposure was performed to a desired depth (about 25 mm in this work) using a photomask, as in Fig. 2 (f), followed by a deep exposure all the way down to the substrate using another mask, 28 as in Fig. 2 (g). Finally, the three-dimensional photoresist structures were obtained after a single development in the AZ 400 K developer, which dissolved all of the LOR and the exposed thick photoresist (Fig. 2 (h)). 2. 2. Result The SEM photographs of various three-dimensional photoresist structures fabricated by applying the proposed process are shown in Fig. 3. The structures, which are unexposed photoresist structures, are suspended about 30 mm from the bottom by removal of the sacrificial LOR layer and recessed about 25 mm from the top by shallow UV exposure, as shown in Fig. 3 (a). A number of photoresist structures with various dimensions and formations were also fabricated to validate the fabrication process and demonstrate the possible shapes and resolutions. Fig. 3 (b) shows the suspended bridge structures with various three-dimensional shapes. The experiment shows that 500 mm-long bridge structures with a thickness of 20 mm can be successfully formed. Sometimes there was a stiction problem in long and wide bridge structures, where the center of the bridge was stuck on the substrate, as shown in Fig. 3 (c). But in most of the structures, we could successfully obtain the suspended three-dimensional PR structures. Fig. 3 (d) and (e) show the various shapes of photoresist structures, such as suspended plates with micro-pits or micro-villi. During the process and in the final structures, no issues related to stress, such as cracking, were observed. In addition, the surface of the recessed region by shallow exposure appears to be smooth according to SEM pictures. The suspended thickness can be adjusted by controlling the thickness of the LOR layer, and the recessed thicknesses can be adjusted by controlling the energy dose of shallow exposure. 3. Complex microchannels in PDMS 3. 1. Experiment In this work, the three-dimensional photoresist structures are used as a master mold for PDMS (Sylgard 184, Dow Corning Co.) replica formation to fabricate complex microchannels in PDMS. After the fabrication of photoresist structures, the surface of a master plate is modified using tridecafluoro- (1,1,2,2-tetrahydrooctyl)-1-trichlorosilane (United Chemical Technologies) to prohibit permanent bonding between the Fig. 4 Application of three-dimensional photoresist structures. (a) Photoresist master structure for PDMS replica mold (left) and PDMS micromixer structure fabricated using this master structure (right). (b) Microfabricated photoresist structure with a solenoidal shape (left) and PDMS microchannel with green ink filled for visualization (right). 248 Lab Chip, 2008, 8, 245 250 This journal is ß The Royal Society of Chemistry 2008
substrate and PDMS. Next, the PDMS prepolymer is cast onto the substrate, followed by degassing in a vacuum chamber to remove air bubbles and fill the entire gap between the substrate and the suspended photoresist structure. Then the PDMS prepolymer is cured at 70 uc for 3 h on a hotplate. Finally, the hardened PDMS is gently peeled off from the substrate in an acetone bath, dissolving the master photoresist structures. After surface activation by using oxygen plasma treatment, the fabricated PDMS layer is bonded with the substrate or another PDMS layer after alignment if required. 3. 2. Result Using the three-dimensional photoresist structures as a mold for the PDMS replica process, we were able to simply fabricate various kinds of PDMS microfluidic structures, such as complex channels, filters, and mixers, as shown in Fig. 4. Our previous work has shown that the formation of a micromixer requires two PDMS structures aligned and bonded together, which has resulted in misalignment in many cases. 9 By using the proposed process, all the micro-mixer channels are accurately formed by using only one molding process, without any misalignment error. Fig. 4 (a) shows the PDMS micro-mixer structure before the bonding on the substrate (right figure) and the photoresist structure used as the master for PDMS molding process (left figure). Fig. 4 (b) shows the three-dimensional photoresist in a solenoidal structure and the corresponding PDMS microchannel where ink flows for visualization. The PDMS prepolymer was cast onto the photoresist structure (left figure) and detached from the substrate after thermal hardening. Because the detachment was performed in an acetone bath, the photoresist was removed from the silicon and PDMS substrates. After rinsing and surface activation, the PDMS was bonded with a glass plate, forming a three-dimensional solenoidal microchannel, as shown in the right figure. We fabricated several types of microchannel structures in PDMS by using the proposed process. More complex Fig. 5 Various types of photoresist structures and multilevel microchannels. The figures on left column are the schematic drawing describing the shape of microchannels and the SEM pictures on middle column show two photoresist structures used as masters to construct a pair of PDMS plates which are aligned and bonded together to form complex PDMS microchannels shown in pictures on right column. The microchannels with the shape of single solenoid (a), double solenoid (b), solenoid with core (c), and transformer (d) have been successfully demonstrated. This journal is ß The Royal Society of Chemistry 2008 Lab Chip, 2008, 8, 245 250 249
microfluidic structures can be fabricated if we use two different photoresist structures and corresponding two PDMS plates which are bonded together with a single PDMS alignment. Fig. 5 shows the pictures of various photoresist structures and PDMS microfluidic structures implemented by using each mold. The figures in the left column are the schematic drawing describing the shape of microchannels. The SEM pictures in the middle column show two photoresist structures used as masters to construct a pair of PDMS plates, which are aligned and bonded together to form complex PDMS microchannels shown in pictures in the right column of Fig. 5. The microchannels with the shape of a single solenoid (a), double solenoid (b), solenoid with core (c), and transformer (d) have been successfully demonstrated. Comparing with other PDMS multilayer fabrication techniques that can utilize the master structures over and over again, 5,14,18 the master photoresist structures in our proposed process are only for single use because they are dissolved in acetone during PDMS detachment. However, our fabrication approach provides accurate alignment between multiple microchannels in different levels by using three-dimensional photoresist masters formed by simple photolithography. 4. Conclusion We developed a new technology to implement complex PDMS microchannels. PDMS microchannels were simply constructed by using three-dimensional photoresist structures as a master for the PDMS replica process. This process has utilized LOR resist as a sacrificial layer to levitate the structural photoresist (AZ 9260) and a multi-step exposure to control the thicknesses of photoresist structures. Various shapes of photoresist structures were successfully fabricated, including cantilevers, suspended bridges, and complex plates with micro-pits or micro-villi. Using the PDMS replica method, we demonstrated that the three-dimensional photoresist structures are applicable to implementing complex microchannels in PDMS. Microchannels imbedded in PDMS were fabricated more easily than before by using a single photoresist mold without any alignment. In addition, more complex multilevel microchannels were constructed by bonding two PDMS layers with a single PDMS alignment. Thus, the proposed technique will allow simple and rapid implementation of various microfluidic structures in PDMS, including microchannels, micro-mixers, and microfilters. Acknowledgements This research was undertaken with the help of a research grant from Sogang University in the year 2007 and Korea Research Foundation Grant funded by the Korean Government. References 1 J. C. McDonald and G. M. Whitesides, Acc. Chem. Res., 2002, 35, 491 499. 2 B. G. Chung, L. A. Flanagan, S. W. Rhee, P. H. Schwartz, A. P. Lee, E. S. Monuki and N. L. Jeon, Lab Chip, 2005, 5, 401 406. 3 S. W. Rhee, A. M. Taylor, C. H. Tu, D. H. Cribbs, C. W. Cotman and N. L. Jeon, Lab Chip, 2005, 5, 102 107. 4 J. Millet Larry, E. Stewart Matthew, V. Sweedler Jonathan, G. Nuzzo Ralph and U. Gillette Martha, Lab Chip, 2007, 8, 987 994. 5 K.-S. Yun and E. Yoon, in Proceedings of mtas 2003, pp. 861 864. 6 K.-S. Yun, D. Lee, H.-S. Kim and E. Yoon, Meas. Sci. Technol., 2006, 17, 3178 3183. 7 Y. Xia and G. M. Whitesides, Annu. Rev. Mater. Sci., 1998, 28, 153 184. 8 L. Tang, M. S. Sheu, T. Chu and Y. H. Huang, Biomaterials, 1999, 20, 1365 1370. 9 M. Rosdy, B. Grisoni and L. C. Clauss, Biomaterials, 1991, 12, 511 517. 10 H. Makamba, J. H. Kim, K. Lim, N. Park and J. H. Hahn, Electrophoresis, 2003, 24, 3607 3619. 11 T. C. Merkel, V. I. Bondar, K. Nagai, B. D. Freeman and I. Pinnau, J. Polym. Sci., Part B: Polym. Phys., 2000, 38, 415 434. 12 S. K. Sia and G. M. Whitesides, Electrophoresis, 2003, 24, 3563 3576. 13 D. C. Duffy, O. J. A. Schueller, S. T. Brittain and G. M. Whitesides, J. Micromech. Microeng., 1999, 9, 211 217. 14 B.-H. Jo, L. M. Van Lerberghe, K. M. Motsegood and D. J. Beebe, J. Microelectromech. Syst., 2000, 9, 76 81. 15 M. A. Unger, H.-P. Chou, T. Thorsen, A. Scherer and S. R. Quake, Science, 2000, 288, 113 116. 16 M. A. Burns, B. N. Johnson, S. N. Brahmasandra, K. Handique, J. R. Webster, M. Krishnan, T. S. Sammarco, P. M. Man, D. Jones, D. Heldsinger, C. H. Mastrangelo and D. T. Burke, Science, 1998, 282, 484 487. 17 T. Thorsen, S. J. Maerkl and S. R. Quake, Science, 2002, 298, 580 584. 18 J. R. Anderson, D. T. Chiu, R. J. Jackman, O. Cherniavskaya, J. C. McDonald, H. Wu, S. H. Whitesides and G. M. Whitesides, Anal. Chem., 2000, 72, 3158 3164. 19 H. Wu, T. W. Odom, D. T. Chiu and G. M. Whitesides, J. Am. Chem. Soc.,, 2003, 125(2), 554 559. 20 E. P. Kartalov, C. Walker, C. R. Taylor, W. F. Anderson and A. Scherer, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 12280 12284. 21 F. G. Tseng, Y. J. Chuang and W. K. Lin, in IEEE 15th International Conference on Micro Electro Mechanical Systems 2002, 2002, pp. 69 72. 22 F. E. H. Tay, J. A. V. Kan, F. Watt and W. O. Choong, J. Micromech. Microeng., 2001, 11, 27 33. 23 H. Sato, T. Kakinuma, J. S. Go and S. Shoji, in IEEE 16th International Conference on Micro Electro Mechanical Systems 2003, 2003, pp. 223 226. 24 Y.-K. Yoon, J.-H. Park, F. Cros and M. G. Allen, in IEEE 16th International Conference on Micro Electro Mechanical Systems 2003, 2003, pp. 227 230. 25 M. Han, W. Lee, S.-K. Lee and S. S. Lee, in IEEE 16th International Conference on Micro Electro Mechanical Systems 2003, 2003, pp. 554 557. 26 B.-G. Kim, J.-H. Kim and E. Yoon, in Proceedings of mtas 2003, 2003, pp. 627 630. 27 K.-S. Yun and E. Yoon, in IEEE International MEMS Conference 2004, 2004, pp. 757 760. 28 J.-B. Yoon, J.-D. Lee, C.-H. Han, E. Yoon and C.-K. Kim, Proc. SPIE, 1998, 3512, 358 366. 250 Lab Chip, 2008, 8, 245 250 This journal is ß The Royal Society of Chemistry 2008