Microfluidic systems integrated with two-dimensional surface plasmon resonance phase imaging systems for microarray immunoassay

Available online at www.sciencedirect.com

Biosensors and Bioelectronics 23 (2007) 466–472

Microfluidic systems integrated with two-dimensional surface plasmon resonance phase imaging systems for microarray immunoassay夽 Kuo-Hoong Lee a , Yuan-Deng Su a , Shean-Jen Chen a , Fan-Gang Tseng b , Gwo-Bin Lee a,∗ a

b

Department of Engineering Science, National Cheng Kung University, Taiwan Engineering and System Science Department, National Tsing Hua University, Taiwan

Received 24 February 2007; received in revised form 10 May 2007; accepted 22 May 2007 Available online 2 June 2007

Abstract This study reports a microfluidic chip integrated with an arrayed immunoassay for surface plasmon resonance (SPR) phase imaging of specific bio-samples. The SPR phase imaging system uses a surface-sensitive optical technique to detect two-dimensional (2D) spatial phase variation caused by rabbit immunoglobulin G (IgG) adsorbed on an anti-rabbit IgG film. The microfluidic chip was fabricated by using micro-electromechanical-systems (MEMS) technology on glass and polydimethylsiloxane (PDMS) substrates to facilitate well-controlled and reproducible sample delivery and detection. Since SPR detection is very sensitive to temperature variation, a micromachine-based temperature control module comprising micro-heaters and temperature sensors was used to maintain a uniform temperature distribution inside the arrayed detection area with a variation of less than 0.3 ◦ C. A self-assembled monolayer (SAM) technique was used to pattern the surface chemistry on a gold layer to immobilize anti-rabbit IgG on the modified substrates. The microfluidic chip is capable of transporting a precise amount of IgG solution by using micropumps/valves to the arrayed detection area such that highly sensitive, highly specific bio-sensing can be achieved. The developed microfluidic chips, which employed SPR phase imaging for immunoassay analysis, could successfully detect the interaction of anti-rabbit IgG and IgG. The interactions between immobilized anti-rabbit IgG and IgG with various concentrations have been measured. The detection limit is experimentally found to be 1 × 10−4 mg/ml (0.67 nM). The specificity of the arrayed immunoassay was also explored. Experimental data show that only the rabbit IgG can be detected and the porcine IgG cannot be adsorbed. The developed microfluidic system is promising for various applications including medical diagnostics, microarray detection and observing protein–protein interactions. © 2007 Elsevier B.V. All rights reserved. Keywords: Microfluidics; MEMS; SPR; Immunoassay; Microarray

1. Introduction

Abbreviations: ADC, analog-to-digital converter; Au, gold; Bio-MEMS, bio-micro-electro-mechanical-systems; CCD, charge-coupled device; DI water, deionized water; DNA, deoxyribonucleic acid; EDC, 3-(3-dimethylaminopropyl)-1-ethylcarbodiimide; ELISA, enzyme-linked immunosorbent assay; EMV, electromagnetic valve; IgG, immunoglobulin G; IR, infrared; LOC, labon-a-chip; MEMS, micro-electro-mechanical-systems; NHS, N-hydroxysulfosuccinimide sodium salt; PDMS, polydimethylsiloxane; Pt, platinum; PR, photoresist; PSI, phase-shift interferometry; PWM, pulse-width-modulation; SAM, self-assembled monolayer; SEM, scanning electron microscope; Si, silicon; SPR, surface plasmon resonance; Ti, titanium; UV, ultraviolet; 2D, two-dimensional 夽 Preliminary results have been presented in the 19th IEEE International Conference on Micro Electro Mechanical Systems (IEEE MEMS 2006), 22–26 January 2006. ∗ Corresponding author. Tel.: +886 6 2757575x63347; fax: +886 6 2761687. E-mail address: [email protected] (G.-B. Lee). 0956-5663/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2007.05.007

Bio-micro-electro-mechanical-systems (Bio-MEMS) and microfluidics technologies based on the principles of biology, chemistry and miniaturization techniques have proven to be a promising approach for biomedical and bioengineering applications (Whitesides et al., 2001; Nguyen et al., 2002). Compact biosensors developed by using Bio-MEMS and microfluidics are capable of selectively detecting chemical and biomedical entities with a relatively high sensitivity (Unger et al., 2000; Delamarche et al., 1998). Furthermore, a so-called lab-on-a-chip (LOC) system, which integrates sample pretreatment, transport, mixing, reaction, separation and detection on a single chip, may be realized using this approach (Delamarche et al., 1997). These microfluidic bio-sensing devices may provide faster and less expensive approaches when compared to existing methods to handle and detect aqueous bio-samples. Furthermore, the func-

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

tionality and reliability of a microfluidic bio-sensing system may be significantly enhanced if integrated with functional microfluidic components. Two of the crucial components of a microfluidic system are micropumps and microvalves, which have been extensively used for automation of detection process for miniaturized biosensing systems (Wang and Lee, 2005; Liao et al., 2004). Many micropumps and microvalves have been reported, which used different activation principles and face different challenges depending on their specific applications. Among them, micropumps and microvalves formed by soft lithographic methods have attracted considerable interest (Unger et al., 2000). Soft lithography is a technique for fabricating micro-structures from elastometric materials allowing modification of surface properties (Whitesides et al., 2001). Microfluidic structures with an inverse image of the template can be fabricated by casting or spin-coating a liquid elastomer onto a master with patterned micro-structures. For example, use of elastic polymer membranes coupled with an external compressed air source to control the movement of the microfluids was reported (Fu et al., 2002). Several types of pneumatic micropumps have been demonstrated in the literature, including three-membrane (Thorsen et al., 2002), serpentine-shape (Wang and Lee, 2006) and spider-web layouts (Wang and Lee, 2005). These pneumatic micropumps can be also used as microvalves if the membranes are deflected completely to shut off the flow inside the microchannels. Immunosensors based on the principle of solid-phase immunoassays and antigen–antibody coupling are one of the most popular techniques for bio-sensing. The specific binding of the complementary species can be successfully detected with this approach. Furthermore, immunoassays facilitated on a microfluidic platform have recently been under extensive investigation. Briefly, ligand molecules are usually immobilized on a microchannel surface and the binding activity of target molecules is detected by flowing the bio-samples through the substrate (Kanda et al., 2004). A microfluidic immunoassay is attractive because it can be performed automatically with reduced sample and reagent consumption, less power consumption and a lower unit cost. In addition, the high surface-to-volume ratio of the microfluidic device can enhance mass transport, resulting in shorter assay times and increased detection sensitivity. For example, Delamarche et al. used multiple channels to pattern proteins onto a chip surface for a direct chemical reaction (Delamarche et al., 1997, 1998). An automatic bio-sensing system which can perform enzyme-linked immunosorbent assay (ELISA) on a microfluidic system was successfully reported by the current research group (Wang and Lee, 2005). Detection of hepatitis C virus and syphilis was demonstrated using the bio-sampling chips. Surface plasmon resonance (SPR) sensing has been used for a variety of chemical and biochemical sensing applications. The SPR device is capable of performing real-time, high-throughput, biomolecular interaction analysis. It has major advantages in that prior labeling of the analytes is not required and it can rapidly monitor dynamic interactions in real-time (Homola et al., 1999). Furthermore, SPR intensity imaging systems even have

467

a capability for detection of arrayed analytes (Jordan and Corn, 1997). Conventionally, SPR imaging systems apply a collimating monochromatic light beam oriented such that it incidents on a gold film through a prism or a grating-coupling. The angle of incidence is close to the SPR resonance angle and the reflectivity pattern is then captured by a charge-coupled device (CCD) detector. The use of SPR intensity imaging in the investigation of label-free sensing of reversible protein adsorption and interactions was also demonstrated (Yu et al., 2003; Guedon et al., 2000). However, the resolution of this kind of SPR intensity imaging system is relatively low such that it is difficult to detect low molecular weight samples and low concentration samples. Alternatively, another SPR phase imaging system with prism coupling to produce interference between SPR phase shifts with respect to a reference light beam can significantly enhance the system sensitivity (Nikitin et al., 2000). Recently, a common-path phase-shift interferometry (PSI) SPR phase imaging system was successfully used to detect the binding affinity of unlabeled biological molecules onto DNA arrays by the current research group (Su et al., 2005). Briefly, the selectivity of a SPR immunoassay is provided by a biological recognition biosample (either antibody or antigen) immobilized on the surface of the metal film since the SPR signal is proportional to the mass of the binding biomolecule (Fontana et al., 1990). Previous studies have reported various approaches to enhance the performance of SPR imaging microarray biosensors (Kanda et al., 2004). However, they all used a single or multi-channel layout based on SPR imaging without using a microfluidic control device to automate the whole process. Besides, SPR detection is very sensitive to temperature variation. A large-scale temperature control system is inevitable while performing the SPR sensing, which may increase the size and complexity of the peripheral apparatus. In this study, we develop an automatic, chip-based microfluidic device that has a multi-channel configuration to detect micro-array immunoassay samples based on an SPR phase imaging detection system (Fig. 1(a)). More detailed information about SPR phase imaging setup can be found in our previous work (Su et al., 2005). When compared to our previous work using a microfluidic system to automate transportation of the sample for SPR sensing (Huang et al., 2006), the current study develops a new microfluidic system in which arrayed microchannels, peristaltic micropumps, microvalves, flow sensors, heaters and temperature sensors are all integrated. The new microfluidic chip with a multilayer PDMS and arrayed microchannels has the capability to transport multiple bio-samples and the potential of achieving high-throughput analyses. The microfluidic chips are designed to facilitate well-controlled and reproducible sample delivery on the arrayed detection area. In addition, platinum resistors for the micro-heaters and the temperature sensor are located inside the microchannel and can provide precise temperature control to reduce thermal noise during SPR sensing. High-throughput screening or sensing of bio-molecules can be achieved by using the microfluidic chips based on the SPR phase imaging method. The 2D SPR phase imaging detection system provides a useful detection scheme for microarray immunoassay with a high specificity and a high sensitivity.

468

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

Fig. 1. (a) Schematic illustration of a microfluidic system integrated with a 2D SPR phase imaging system for the detection of a microarray immunoassay. (b) Schematic illustration of a microfluidic chip comprising of a 3-to-1 converging microchannels, microvalves, micropumps, flow sensors, heaters and temperature sensors.

2. Experimental 2.1. Design and fabrication Fig. 1(b) illustrates schematically the proposed microfluidic chip. The microfluidic chip comprises microchannels, micropumps/valves, flow sensors and a micromachine-based temperature control system fabricated by using MEMS technology. The microfluidic chip is made of three layers of PDMS and a glass substrate. The first PDMS layer contains structures for micropumps/valves. The second PDMS layer is for arrayed microchannels. The third PDMS layer has an arrayed detection area. The micromachine-based heaters, temperature sensors and flow sensors are fabricated on the glass substrate. Elastic polymer PDMS is used to duplicate microchannels and thin membranes for micropumps and microvalves from SU-8 templates. The bio-samples are transported using a “spider-web” peristaltic micro-pneumatic pump (Wang and Lee, 2005) from the inlet to the detection area. The detailed fabrication process of the microfluidic chip can be found in literature (Liao et al., 2004). Briefly, the micromachine-based temperature sensors and heaters were fabricated on soda-lime glass substrates (G-Tech Optoelectronics

Corp., Taiwan) by using standard lithography and lift-off processes. The glass substrates were first cleaned and then followed by patterning a thin layer of positive photoresist (PR, AZ4620, Clariant, USA). An electron-beam evaporator was used to deposit a 0.1 ␮m platinum (Pt) layer on the glass surface that had previously been deposited with a thin layer of titanium (Ti, 0.02 ␮m) to enhance Pt adhesion. The width of the temperature sensor was measured to be 50 ␮m. Similarly, electrical leads were patterned using a 0.2 ␮m gold layer. The resistances of the sensor and heaters were measured to be 700 and 30 , respectively. After the patterning process, a cover glass plate (Marienfeld Corp., Germany) with a thickness of 100 ␮m was then bonded on top of the glass substrate utilizing ultravioletsensitive glue to prevent electric breakdown during operation. Additionally, multiple microchannels and micropumps/valves were fabricated by using a SU-8 molding and a PDMS casting processes (Liao et al., 2004). Initially, the silicon (Si) substrate as the master mold was cleaned by using piranha solution (H2 SO4 :H2 O2 = 3:1) with a temperature of 180 ◦ C for 10 min, rinsed by deionized (DI) water and blow dried with nitrogen gas. A SU-8 negative thick-film PR (MicroChem Corp., MA, USA), was spin-coated on the substrate and followed by a 30-min soft-baking process. A standard lithography process was then performed. Microstructures were patterned by using a mask aligner with an exposure of 575 mJ/cm2 . A post-exposure bake was then conducted at 65 ◦ C for 3 min and 95 ◦ C for 10 min, respectively. After the baking process, the substrate was immersed into a developer solution with an ultrasonic agitation for 5 min to develop the micro-structures. After the SU-8 molding process, PDMS (Sylgard 184A/B, Dow Corning Corp., USA) was used to replicate the inverse structures of the microchannels and micropumps/valves by spin-coating a PDMS layer onto the SU-8 mold. The elastomer and curing agents in the mixed ratio of 10:1 was poured onto the master mold and cured at a temperature of 75 ◦ C overnight. The PDMS layers were then released from the SU-8 mold. Another PDMS membrane with a thickness of 300 ␮m was formed by using a spinning speed of 500 rpm for 30 s. The surfaces were treated with an oxygen plasma for a subsequent bonding process. Via holes for sample reservoirs were then mechanically drilled such that a horizontal channel can transport samples from the bottom channel to the top detection area. Finally, the completed microfluidic chip was assembled by bonding two substrates together using an oxygen plasma treatment. An assembled microfluidic chip is shown in Fig. 2(a). Note that the microchannels are filled with red dye in this figure in order to observe the micropumps and microvalves. The dimensions of the microfluidic chip are measured to be 6 cm × 3.5 cm. Fig. 2(b)–(d) shows scanning electron microscope (SEM) photographs of the micropumps/valves SU-8 molds, the PDMS replicas of the arrayed detection area, the temperature sensor and the heaters, respectively. 2.2. Materials A model involving the interaction between anti-rabbit IgG and IgG was used to verify the capability of the developed

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

469

Fig. 2. (a) Photograph of the completed microfluidic chip with dimensions of 6 cm × 3.5 cm. The microchannels are filled with red dye. The SEM images of the (b) SU-8 molds of the micropumps/valves, (c) PDMS replica of the arrayed detection area and (d) the temperature sensor and heaters. The width of the temperature sensor is 50 ␮m. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of the article.)

bio-sensing system. An Au film in the SPR imaging system had a thickness of 47.5 nm and was coated on the SF-11 slide (Hsin Yu Tech. Co., Taiwan) by using a sputtering deposition process. Prior to depositing the Au film, the SF-11 slide was pre-coated with a 2-nm chromium layer to enhance the adhesion of the film. A self-assembled monolayer (SAM) using thiols on the Au thin film was first formed (Guedon et al., 2000). The thiol treatments are particularly useful for assays of specific binding of proteins to ligands on the Au surface. The formation of SAM on Au surfaces has been one of the most popular protocols to modify a surface chemically in order to immobilize complex biomolecules for bio-sensing. Briefly, the slide was first immersed in a 1-mM thiol solution (HS(CH2 )15 COOH) for 6 h. The Au-film slide with the carboxyl function group thiol SAM was then immersed into a fresh sulfoNHS (N-hydroxysulfosuccinimide sodium salt) (2 mM) and 3-(3-dimethylaminopropyl)-1-ethylcarbodiimide (EDC) solution (5 mM) to activate the carboxyl group for 12 h at room temperature. The activated Au-film slide was then rinsed with the reaction buffer solution (sodium phosphate) twice and blown dry with nitrogen. The Au-film slide was finally bonded with

the upper PDMS layer of the microfluidic chip. The slide was then aligned with the SPR prism. In this study, a reaction buffer (sodium phosphate, pH 7.4) was prepared from Na2 HPO4 and NaH2 PO4 in DI water. Anti-rabbit IgG (R2004, Sigma Chemical Company, Switzerland) was immobilized on the activated slide and then used to identify the rabbit IgG in the fluid. Rabbit serum (I5006, Sigma Chemical Company, Switzerland) with different concentrations of IgG were detected using the microfluidic chip integrated with the SPR phase imaging system. 2.3. SPR instrumentation The SPR phase imaging system is based on the Kretschmann configuration incorporated with a common-path PSI technique (Su et al., 2005). The incident light source is a He-Ne laser (Melles Griot, USA) with a wavelength of 632.8 nm and the interference imaging signals are acquired by using a CCD camera. The interference images associated with these phase shifts are acquired and displayed in the form of the digital images. The common-path PSI technique has the advantage of long-term stability, even when subjected to external disturbances. The SPR

470

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

phase imaging system meets the requirements of the real-time kinetic studies involved in biomolecular interaction analysis. The SPR phase imaging system demonstrates a detection limit of 2 × 10−7 refractive index change, a long-term phase stability of 2.5 × 10−4 π (root-mean-square) over 4 h, and a spatial phase resolution of 10−3 π with a lateral resolution of 100 ␮m (Su et al., 2005). Since the property of the Au film is rather stable, it is then chosen as the material for the metallic thin film for SPR sensing. The Au film with a thickness of 47.5 nm is calculated to be the most sensitive for SPR measurement for a laser light of wavelength 632.8 nm. Adopting the Kretschmann configuration, an SF-11 right angle prism with a refractive index of 1.779 at 632.8 nm was used to increase the wave vector of the incident light and decrease the resonant angle to below 55◦ in the phosphate buffer solution (PBS, pH 7.4) with a refractive index of 1.334. A SF-11 slide is used for the gold film deposition and is in contact with the prism base using an index matching oil (Cargille Labs, NJ, USA) having the same refractive index n = 1.779 of both the prism and the slide. Finally, the assembled microfluidic chip is aligned with the SPR phase imaging system. 3. Results and discussion In this study, we used a peristaltic pneumatic micropump for driving the fluid flow (Wang and Lee, 2005). The micropumps controlled by three electromagnetic valves (EMV, SMC Inc., S070M-5BG-32, Japan) can transport a precise amount of bio-samples flowing through multiple microchannels by using a simple control circuit. The peristaltic sequence could be precisely modulated to control the input and output of compressed air. The pumping rate of the micro-pneumatic pumps was first explored in terms of operating frequency and applied air pressure. As the compressed air travels along the microchannel it causes the membranes to deflect. The time-phased deflection of successive membranes along the microchannel length generates a peristaltic effect that drives the fluid along the microfluidic channel (Wang and Lee, 2005). Experimental data show that the pumping rate significantly increases with the increasing frequency. Similarly, the pumping rate also increases with the driving air pressure. The pumping rate will ultimately reach a saturation value limited by the operation frequency of the EMVs. The maximum value of the fluid pumping rate is measured to be 40 ␮l/min at the driving frequency of 17 Hz and an air pressure of 25 psi. The minimum driving air pressure is experimentally found to be 10 psi to assure that PDMS membranes can be completely deflected. In this study, a pumping rate of 15 ␮l/min at a driving frequency of 15 Hz was used to pump the bio-samples through the detection area to assure stable sample transportation. The permittivity of the base film (i.e. Au film) is sensitive to the variation of the bulk temperature in the arrayed detection area. Hence the stability of the temperature is a crucial issue in SPR phase imaging system. Temperature variation can affect the sensitivity of the SPR system (Ozdemir and Turhan-Sayan, 2003). Therefore, a micro temperature control module was used to maintain a stable temperature field. Note that both the temperature sensors and the micro-heaters were located within the arrayed microchannels. This arrangement facilitates the real-

time measurement of the temperature field and improves the response of the temperature control system. An ATMEGA8535 microcontroller (ATMEL Corp., USA) was used to operate as an analog-to-digital (ADC) converter and a pulse-widthmodulation (PWM) module to measure the signals from the temperature sensor and adjust the heaters. The micromachinebased temperature control system can be used to maintain a temperature field at 30 ◦ C during the SPR sensing process. With this approach, it can dramatically reduce the thermal noise common to SPR sensing. Bio-samples can be heated up to 30 ◦ C within 20 s and then kept at a constant value. An infrared (IR) camera (ThermaCAM P25, FLIR systems Inc., USA) was used to capture thermography in the detection area (3 chambers of each channel). Fig. 3 shows the temperature distribution inside the arrayed detection area. It is found that the temperature is uniformly distributed with a variation less than 0.3 ◦ C when a buffer solution flows through the micromachine-based temperature control system. Therefore, the steady temperature at each channel enhances the stability of the SPR detection system during the testing process. Note that the SPR phase imaging system has the capability to detect a protein immunoassay with an arrayed configuration. The developed system has successfully detected the interaction between anti-rabbit IgG and rabbit IgG. After characterization of the micropumps and the micro temperature control system, the IgG interaction procedures were carried out as follows. First, the immobilization of the anti-rabbit IgG on the Au surface was performed. Then the microfluidic chip was coupled with the SPR phase imaging system. Micropumps were used to transport the buffer solution into the multiple microchannels for 40 min. The micromachine-based temperature control system was used to maintain a stable temperature of 30 ◦ C during the whole testing process. Then, a solution of 0.01 mg/ml anti-rabbit IgG was constantly pumped through the immobilized Au film of the arrayed detection area such that the anti-rabbit IgG can be conjugated on the detection area. The reaction procedure is described as follows: 0–40 min for buffer injection, 41–190 min for buffer and

Fig. 3. The temperature control system can heated up bio-samples to 30 ◦ C within 20 s and kept them at a uniform temperature. The thermography shows a uniform temperature distribution in each detection chamber when buffer flows through the micromachine-based temperature control system.

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

471

Fig. 5. The association rate for 25 and 15 ␮l/min, respectively, for various rabbitIgG concentrations (0.01, 0.005, 0.001 mg/ml). Note that the association rate of the 25 ␮l/min is consistently higher than that of the 15 ␮l/min.

Fig. 4. (a) The phase difference caused by the injection of anti-rabbit IgG with a concentration of 0.01 mg/ml is measured to be approximately 0.19π. The operation procedure is as follows: 0–40 min for buffer injection, 41–190 min for buffer + anti-rabbit IgG, and 191–245 min for buffer washing, respectively. (b) The relation between phase difference and rabbit IgG interaction with various concentrations. The detection limit is found to be 1 × 10−4 mg/ml.

anti-rabbit IgG injection, and 191–245 min for buffer washing, respectively. The phase difference caused by the anti-rabbit IgG interaction is measured to be approximately 0.19π (Fig. 4(a)). After the anti-rabbit IgG was immobilized, various concentrations of the rabbit IgG were injected into the microchannels. The microfluidic chip could successfully detect the phase variation caused by antibodies adsorbed on the Au surface. Fig. 4(b) shows the relation between phase difference and the concentrations of rabbit IgG. Interaction between rabbit IgG with various concentrations and immobilized anti-rabbit IgG have been successfully measured using this method. Note that a pumping rate of 15 ␮l/min and a control temperature of 30 ◦ C were used in all cases. The detection limit was experimentally found to be 1 × 10−4 mg/ml. In addition, the association rate due to different pumping rates of the micropumps was also explored. Fig. 5 compares the association rate of IgG while different pumping rates are applied. It can be clearly seen that the association rate at 25 ␮l/min is higher than at 15 ␮l/min for a given constant

temperature and concentration. The pumping rate of 25 ␮l/min dramatically decreases the adsorption time and achieves a saturation state in a shorter period of time. The higher association rate observed should be attributed to more efficient interaction between the rabbit IgG and its antibodies (Fig. 5). Furthermore, the porcine IgG (I4381, Sigma Chemical Company, Switzerland) was used to test the nonspecific IgG adsorption (Fig. 6). Again, the anti-rabbit IgG with a concentration of 0.01 mg/ml was first conjugated on the detection area. The micropump then transported the porcine IgG (0.01 mg/ml), rabbit IgG (0.01 mg/ml), and a buffer solution through each microchannel to the detection area. After a 2 h interaction process followed by a washing process, it was clearly seen that only

Fig. 6. The specific targets of the IgGs (channels a, b, and c are for porcine IgG, rabbit IgG and a reference buffer, respectively) are tested in each microchannel. The low value of phase change on the porcine IgG region indicates that nonspecific adsorption can be ignored (image constructed from a 50 × 50 pixel region covering a 1.0 mm × 1.0 mm detection area).

472

K.-H. Lee et al. / Biosensors and Bioelectronics 23 (2007) 466–472

the rabbit IgG (in channel b) was successfully conjugated with its antibody. The low phase change value inside the porcine IgG region indicates that nonspecific adsorption is not significant. The phase difference before and after injection of buffer, porcine IgG and rabbit IgG was used as detection signal. Therefore, the microfluidic chip can successfully detect 2D spatial phase variation caused by antibodies absorbed on a sensing surface composed of highly specific protein films. 4. Conclusions In this study, we have successfully demonstrated an automatic microfluidic chip integrated with an arrayed immunoassay for SPR phase imaging of specific bio-samples. Using this microfluidic technology, a precise amount of bio-samples inside multiple microchannels can be automatically transported to the detection area by using micropumps/valves. Meanwhile, the micro-heaters and the temperature sensors were used to maintain a constant temperature field in the detection area so that thermal noise during SPR sensing can be reduced. The SPR phase imaging system has a high sensitivity and a highthroughput screening capabilities. This method can be used for biomolecular interaction analysis and no labeling of reagents is required. The spatial phase resolution achieved is 10−3 π in a long-term measurement over 4 h and the detection limit is found to be 1 × 10−4 mg/ml (0.67 nM). The experimental data show that the specific protein can be measured by using the microfluidic devices integrated with the SPR phase imaging system. Therefore, the developed microfluidic chip is promising for a variety of automatic immunoassay applications. Acknowledgements The authors would like to thank partial financial support from the National Science Council in Taiwan. Access to major

fabrication equipment at the Center for Micro/Nano Technology Research, National Cheng Kung University is also greatly appreciated. References Delamarche, E., Bernard, A., Schmid, H., Michel, B., Biebuyck, H., 1997. Science 276, 779–781. Delamarche, E., Bernard, A., Schmid, H., Bietsch, A., Michel, B., Biebuyck, H., 1998. J. Am. Chem. Soc. 120, 500–508. Fontana, E., Pantell, R.H., Strober, S., 1990. Appl. Opt. 29, 4694–4704. Fu, A.Y., Chou, H.P., Spence, C., Arnold, F.H., Quake, S.R., 2002. Anal. Chem. 74, 2451–2457. Guedon, P., Livache, T., Martin, F., Lesbre, F., Roget, A., 2000. Anal. Chem. 72, 6003–6009. Homola, J.S., Yee, S., Gauglitz, G., 1999. Sens. Actuators B 54, 3–15. Huang, S.C., Lee, G.B., Chien, F.C., Chen, S.J., Chen, W.J., Yang, M.C., 2006. J. Micromech. Microeng. 6, 1251–1257. Jordan, C.E., Corn, R.M., 1997. Anal. Chem. 69, 1449–1456. Kanda, V., Kariuki, J.K., Harrison, D.J., McDermott, M.T., 2004. Anal. Chem. 76, 7257–7262. Liao, C.S., Lee, G.B., Wu, J.J., Chang, C.C., Hsieh, T.M., Luo, C.H., 2004. Biosens. Bioelectron. 20, 1341–1348. Nguyen, N.T., Huang, X.Y., Chuan, T.K., 2002. J. Fluids Eng. 124, 384– 392. Nikitin, P.I., Grigorenko, A.N., Beloglazov, A.A., Valeiko, M.V., Savchuk, A.I., Savchuk, O.A., Steiner, G., Kuhne, C., Huebner, A., Salzer, R., 2000. Sens. Actuators A 85, 189–193. Ozdemir, S.K., Turhan-Sayan, G., 2003. J. Lightwave Technol. 21, 805– 814. Su, Y.D., Chen, S.-J., Yeh, T.L., 2005. Opt. Lett. 30, 1488–1490. Thorsen, T., Maerkl, S.J., Quake, S.R., 2002. Science 298, 580–584. Unger, M.A., Chou, H.P., Thorsen, T., Scherer, A., Quake, S.R., 2000. Science 288, 113–116. Wang, C.H., Lee, G.B., 2005. Biosens. Bioelectron. 21, 419–425. Wang, C.H., Lee, G.B., 2006. J. Micromech. Microeng. 16, 1–8. Whitesides, G.M., Ostuni, E., Takayama, S., Jiang, X.Y., Ingber, D.E., 2001. Annu. Rev. Biomed. Eng. 3, 335–373. Yu, X., Wang, D., Wang, D., Ou, Y.J.H., Yan, Z., Dong, Y., Liao, W., Zhao, X.S., 2003. Sens. Actuators B 91, 133–137.