Anal. Chem. 2000, 72, 4242-4247
A Miniaturized DNA Amplifier: Its Application in Traditional Chinese Medicine Thomas M. H. Lee, I-Ming Hsing,* Alex I. K. Lao, and Maria C. Carles†
Department of Chemical Engineering, The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong
A Si-based miniaturized device for the polymerase chain reaction (PCR) has been developed. The device has Pt temperature sensors and heaters integrated on top of the reaction chamber for real-time accurate temperature sensing and control. Reaction temperature of the device is digitally controlled to achieve a temperature accuracy of (0.025 °C. The effects of PCR protocol optimization on the amplification performance of the surface-passivated chip reactor have been investigated in detail and quantitatively compared with that of the conventional thermal cycler. In this study, four traditional Chinese medicine genes including Fritillaria cirrhosa, Cartharmus tinctorius, Adenophora lobophilla, and Stephania tetrandra are used as model template. With appropriate chamber surface treatment (chlorotrimethylsilane/polyadenylic acid or SiO2 coatings), the device demonstrates amplification as efficient as that in the conventional thermal cycler at optimized MgCl2 concentration. The amplified DNA has concentration higher than 27 ng/µL, which is sufficient for subsequent on-chip analyses and detection. Experimental results reveal the importance of inclusion of BSA for an efficient amplification in the SiO2passivated device and the excellent reusability of the device with a simple cleaning step. The formal introduction of the polymerase chain reaction1-4 (PCR) in 1986 by Mullis to replicate defined DNA fragments in vitro has revolutionized the routine practice in molecular biology. The use of a thermal cycler reduces the time required for amplifying DNA from several days in the conventional cloning method to a few hours in PCR and thereby has received wide applications in medical diagnostics, genetic analyses, forensic science, and basic research (cloning, restriction analysis, sequencing, fingerprinting, and others). Most medical applications involve time-consuming DNA analysis procedures such as sample preparation, DNA amplification, product separation, and detection. However, the utilization of bulky benchtop instrumentation in these procedures limits clinical diagnostics or on-site DNA † Department of Biology (1) Mullis, K.; Faloona, F.; Scharf, S.; Saiki, R.; Horn, G.; Erlich, H. Cold Spring Harbor Symp. Quant. Biol. 1986, LI, 263-273. (2) Mullis, K. B.; Ferre, F.; Gibbs, R. A. The Polymerase Chain Reaction; Birkhauser: Boston, 1994. (3) Innis, M. A.; Gelfand, D. H.; Sninsky, J. J.; White, T. J. PCR Protocols: A Guide to Methods and Applications; Academic Press: San Diego, 1990. (4) Eeles, R. A.; Stamps, A. C. Polymerase Chain Reaction (PCR): The Technique and Its Applications; R. G. Landes Co.: Austin, TX, 1993.
4242 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
analyses. A portable Si/glass-based DNA analyzer that features short analysis time, small sample volume, and automated operation can be fabricated by integrated circuit (IC) and microelectromechanical systems (MEMS) technologies, offering the possibility of finishing routine and tedious medical analysis work in a few minutes. The polymerase chain reaction (PCR) device is an indispensable module for a Si-based DNA analyzer when the starting amount of DNA is too small for direct detection or analysis. The first miniaturized DNA amplifier was developed by Northrup et al.5 in 1993. It has polysilicon heaters patterned on the bottom Si3N4 membrane and is sealed from the top by a glass slide. Since then, many studies have been carried out to enhance the design of the microreaction chamber.6-13 Efforts have been made to improve temperature control, enhance temperature ramping rates, and reduce power consumption. Meanwhile, some research groups14-19 have been working on the integration of the amplification module with sample preparation as well as product separation together with detection units. However, comparison of amplification efficiency between Si/glass-based and the conventional polypropylene tube-based PCR has not received much attention. (5) Northrup, M. A.; Ching, M. T.; White, R. M.; Watson, R. T. Proc. 1993 IEEE Int. Conf. Solid-State Sens. Actuators 1993, 924-926. (6) Wilding, P.; Shoffner, M. A.; Kricka, L. J. Clin. Chem. 1994, 40, 18151818. (7) Northrup, M. A.; Gonzalez, C.; Hadley, D.; Hills, R. F.; Landre, P.; Lehew, S.; Saiki, R.; Sninsky, J. J.; Watson, R. Proc. 1995 IEEE Int. Conf. SolidState Sens. Actuators 1995, 764-767. (8) Shoffner, M. A.; Cheng, J.; Hvichia, G. E.; Kricka, L. J.; Wilding, P. Nucleic Acids Res. 1996, 24, 375-379. (9) Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3407-3412. (10) Poser, S.; Schulz, T.; Dillner, U.; Baier, V.; Kohler, J. M.; Schimkat, D.; Mayer, G.; Siebert, A. Sens. Actuators 1997, A 62, 672-675. (11) Taylor, T. B.; Winn-Deen, E. S.; Picozza, E.; Woudenberg, T. M.; Albin, M. Nucleic Acids Res. 1997, 25, 3164-3168. (12) Daniel, J. H.; Iqbal, S.; Milington, R. B.; Moore, D. F.; Lowe, C. R.; Leslie, D. L.; Lee, M. A.; Pearce, M. J. Sens. Actuators 1998, A 71, 81-88. (13) Kopp, M. U.; de Mello, A. J.; Manz, A. Science 1998, 280, 1046-1048. (14) Woolley, A. T.; Hadley, D.; Laudre, P.; de Mello A. J.; Mathies R. A.; Northrup M. A. Anal. Chem. 1996, 68, 4081-4086. (15) Waters, L. C.; Jacobson, S. C.; Kroutchinina, Natalia.; Khandurina, J.; Foote, R. S.; Ramsey, J. M. Anal. Chem. 1998, 70, 158-162. (16) Wilding, P.; Kricka, L. J.; Cheng, J.; Hvichia, G.; Shoffner, M. A.; Fortina, P. Anal. Biochem. 1998, 257, 95-100. (17) Cheng, J.; Waters, L. C.; Fortina, P.; Hvichia, G.; Jacobson, S. C.; Ramsey, J. M.; Kricka, L. J.; Wilding, P. Anal. Biochem. 1998, 257, 101-106. (18) Northrup, M. A. Anal. Chem. 1998, 70, 918-922. (19) 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. Science 1998, 282, 484-487. 10.1021/ac000384b CCC: $19.00
© 2000 American Chemical Society Published on Web 08/03/2000
Figure 1. Illustrations of (a) the top and (b) the bottom views of the chip reactor. Schematic representations of the metal pattern and the cross-sectional structure are shown in the two figures, respectively. Note that TS stands for temperature sensor and H for heater.
In this work, we report a systematic study on the design, fabrication, and implementation of a Si-based miniaturized reactor with integrated heaters and temperature sensors for an efficient DNA amplification. Thermal characteristics, surface passivation, and reaction protocol optimization issues of the device are discussed in detail, in particular the effects of PCR protocol optimization on the amplification performance of the surfacepassivated chip reactor. Motivated by the ever increasing interests in the Asia Pacific communities on standardizing traditional Chinese medicine (TCM) using state-of-the-art molecular biology techniques, the amplification of four genes from TCM preparations used to treat ailments of the upper respiratory tract and in stroke therapy will be demonstrated in the chip reactor, the amplification efficiency of which will be compared with that of the conventional thermal cycler. EXPERIMENTAL SECTION The miniaturized PCR device has a reaction chamber of ∼8 µL (dimension of 5.8 mm × 3.7 mm × 0.35 mm, the length and width are reported at half depth) and Pt temperature sensors and heaters integrated on top of the reaction chamber for real-time and accurate temperature monitoring and control. Figure 1a shows the metal pattern on top of the chip reactor, and Figure 1b shows the bottom and cross-sectional structures. The average temperature reading of the two sensors on the left (TS1 and TS2) is used to control the left heater (H1) while TS3/TS4 and H2 form the
other pair. To have an efficient integrated heating and temperaturesensing systems20 on the device, the resistance and geometry of the heater and sensor should comply with certain design rules. These design rules and the fabrication procedures of the device are delineated in the Appendix. After clean room fabrication, the Si-based device was sealed from the bottom with Corning 7740 glass utilizing an anodic bonding process.21 Surface Passivation. Prior to chip-based PCR experiments, the inner surface of the reaction chamber has to be passivated to avoid any possible nonspecific adsorption of the reagents, enzyme, and DNA used in the PCR on the chamber wall, which inhibits the reaction and results in poor amplification. Methods of surface passivation reported to date include silanization (silane or silanebased compounds), bovine serum albumin (BSA) incubation, and polymer solution coating.8 Two surface preparations including coating of polymer and deposition of a thin thermal oxide, similar to the approach proposed by Shoffner and co-workers,8 will be examined in this study. In the polymer coating, chlorotrimethylsilane (CTMS, MerckSchuchardt) was pipetted to fill the reaction chamber and was dried in a nitrogen hood for 1 h after a 15-min incubation at room temperature. The chamber was then rinsed 3 times with auto(20) Lao, I. K.; Lee, M. H.; Hsing, I. M.; Ip, N. Y. Sens. Actuators A 2000, 84, 11-17. (21) Lee, M. H.; Lee, H. Y.; Liaw, Y. N.; Lao, I. K.; Hsing, I. M. Sens. Actuators A. In press.
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
4243
Figure 2. Schematic diagram of the chip PCR system setup.
claved double-deionized water and dried in a vacuum dryer for 1 h. After that, polyadenylic acid (Sigma Chemical Co., St. Louis, MO), 10 mg/mL in 10 mM Tris-HCl (pH 8.5), was injected into the chamber and incubated at room temperature for 1 h. Finally, the polymer solution was pipetted out and the chamber was again rinsed 3 times with autoclaved double-deionized water and dried. For the passivation by oxide, a 300-nm thermal SiO2 layer was deposited on the inner wall of the Si substrate after the opening of reaction chamber by time-stop KOH etching. PCR Protocol and Experiment. For each PCR experiment, 10 µL of master mix was prepared. The mixture contains 1 µL of 10× PCR buffer solution (200 mM Tris-HCl, pH 8.4, 500 mM KCl), 0.4 µL of 50 mM MgCl2, 0.2 µL of 10 mM dNTPs (GIBCO, BRL, Life Technologies, MD), 0.4 µL of 10 µM S-1 forward primer (5′GGATTCGTGCTTGGGCGAGAGTAG TA-3′) (Synthetic Genetics), 0.4 µL of 10 µM AS-1 reverse primer (3′-GGATCCTTAGTGCTGGTATGATCGCA-5′) (Synthetic Genetics), 1 µL of 5 µg/ µL BSA (Sigma Chemical Co., dissolved in autoclaved doubledeionized water), 0.4 µL of 1 ng/µL template DNA, 0.4 µL of 5 units/µL Taq DNA polymerase (GIBCO, BRL), and 5.8 µL of autoclaved double-deionized water. The template DNA consists of the 5S-rRNA spacer gene of one of the four TCM herbs cloned into pCR2.1-TOPO TA vector (3.9kb) (Invitrogen). The sizes of the gene amplified are for Fritillaria cirrhosa ∼600 bp, Cartharmus tinctorius ∼300 bp, Adenophora lobophilla ∼340 bp, and Stephania tetrandra ∼500 bp. The conventional and chip PCRs are subject to the same thermal cycling profile: initial denaturation at 95 °C for 4 min, 35 cycles of 95 °C for 30 s, 55 °C for 30 s, 72 °C for 1 min, and a final extension at 72 °C for 5 min. Instrumentation. The conventional PCR was performed in the GeneAmp PCR System 9600 (Perkin-Elmer). The chip-based PCR apparatus consists of four main units including analytical prober, data acquisition system, digital control unit, and power supply source. A schematic drawing of the system setup is illustrated in Figure 2. The analytical prober (Karl Suss) was used to align heater and temperature sensor contact pads to a custom4244
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
ized probe card for electrical connection between the chip and external electronic components. The temperature sensors were connected to a data acquisition (DAQ) card (PCI-MIO-16E-1, National Instruments) via a signal conditioning board (SC-2042RTD, National Instruments), which gave the temperature sensors an excitation current of 1 mA. The conversion from voltage reading to temperature was based on the sensor resistance-temperature correlation as described in the Results and Discussion section. A digital feedback proportional and integral (PI) with gain scheduling control algorithm20,22 was implemented in LabVIEW platform (National Instruments) to control the output of an external power source (HP6629A, Hewlett-Packard) via a GPIB interface (HP82350A). Product Analysis. For both GeneAmp and chip PCRs, 4 µL of each reaction product was mixed with 2 µL of 5× loading buffer (0.25% bromphenol blue, 30% glycerol) and 4 µL of water. Then, the samples were loaded in a 1% agarose/TBE containing 0.5 µg/ mL ethidium bromide. The electrophoresis buffer used was Tris/ borate/EDTA buffer (TBE) (44.5 mM Tris, 44.5 mM boric acid, and 1 mM EDTA). The gel was electrophoresed at 100 V for 1 h, and the product was visualized by UV transillumination. RESULTS AND DISCUSSION Thermal Characteristics. Resistance of the Pt temperature sensors has a linear relationship with temperature, given by
R ) Ro[1 + R(T - To)] ) RRoT + (Ro - RRoTo)
(1)
where R is the sensor resistance (Ω) at temperature T (°C), Ro is the reference sensor resistance (Ω) at reference temperature To (°C), and R is the temperature coefficient of resistance (TCR) of the sensor. By equating the slope of the sensor resistancetemperature plot (performed in the mechanical convection oven 650, Precision Scientific) with RRo, the TCR of the Pt temperature (22) Seborg, D. E.; Edgar, T. F.; Mellichamp, D. A. Process Dynamics and Control; Wiley: New York, 1989; Chapters 7 and 12.
Figure 3. Typical thermal cycle of the chip reactor with denaturation at 95 °C for 45 s, annealing at 55 °C for 40 s, and extension at 72 °C for 1 min. Note that the time specified corresponds to that for set point. The actual hold time for the reaction chamber temperature is ∼30 s at 95 °C, ∼30 s at 55 °C, and ∼1 min at 72 °C.
sensor can be determined. In our case, the TCR was determined to be 1.59 × 10-3/°C. As PCR requires repetitive thermal cycling of reagent at three temperature set points, the application of gain scheduling along with the digital feedback PI control strategy offers a good temperature control22 for the chip reactor, as illustrated in Figure 3. The chip reactor achieves a temperature accuracy of (0.025 °C. With simple three-partition gain schedule input, corresponding to the three temperature set points, heating and cooling rates (by natural convection) of 5 and 3.5 °C/s have been achieved, respectively. A higher heating rate is possible with a more aggressive setting of controller parameters (larger proportional gain or smaller integral time), but this would inevitably result in a larger overshoot. In fact, with a modified gain scheduling algorithm (six-partition gain schedule input) and a more stringent tuning of controller parameters, a 3-fold increase in heating rate has been achieved with an overshoot of less than 1 °C. The four temperature sensors can provide information on the temperature uniformity across the chamber surface. At all three temperature set points, the difference in temperature readings of the four sensors is less than 0.1 °C (data not shown). This discrepancy is likely caused by the small but significant variations in the local resistance of the two heater lines. Also, we have conducted 3-D numerical simulation for the spatial temperature profile of the microreactor.20 Simulation results indicate that the temperature gradient within the reaction chamber is less than 0.1 °C, which does not have significant influence on the amplification efficiency. Moreover, the device has low power consumption. It requires only 1.5 W to heat the reaction chamber to 95 °C, the denaturing temperature of DNA. DNA Amplification. As Si inhibits the PCR, the inner chamber wall of the device has to be passivated prior to the thermal cycling experiments to minimize, if not eliminate, the nonspecific adsorption of the PCR reagents, enzyme, and DNA. To start with, the passivation was based on the work of Shoffner et al.8 (CTMS and polyadenylic acid, as mentioned in the Experimental Section). To achieve a high amplification efficiency, the required amount of MgCl2 in the PCR buffer solution has to be optimized in the chip device, as the concentration of which can be very different from that in the conventional PCR setting. Figure 4 shows the amplification results of a series of chip PCR experiments with varying MgCl2 concentration. Shown in lanes
Figure 4. Top gel electrophoresis series showing the amplification results of F. cirrhosa with varying MgCl2 and the bottom series showing DNA standards. The chip reactor used was passivated with CTMS and polyadenylic acid, and the amplification results are shown in lanes 2-5: (lanes 1 and 7) 1 kb molecular weight marker (GIBCO, BRL); (lane 2) 2 mM MgCl2; (lane 3) 3 mM MgCl2; (lane 4) 4 mM MgCl2; (lane 5) 5 mM MgCl2; (lane 6) conventional PCR with 2 mM MgCl2; (lanes 8-16) DNA standards of known amount, the amounts (ng) are marked in the figure.
2-5 are chip-based DNA amplification results with 2-5 mM MgCl2. It is possible to quantify the amount of DNA amplified by comparing the fluorescence intensity of the amplified DNA bands in lanes 2-5 with the intensity of fixed amount DNA bands in lanes 8-16. Their concentrations are determined to be higher than 27 ng/µL (lane 2), ∼9 ng/µL (lane 3), and lower than 2.5 ng/µL (lane 4 and lane 5). With appropriate surface passivation and optimized MgCl2 concentration, the amplification performance demonstrated by the chip PCR device is as good as that of the conventional thermal cycler, as indicated by the similar fluorescence intensity of lane 2 and lane 6. However, the surface pretreatment by CTMS and polyadenylic acid presents a large overhead to device fabrication and is not a clean-room-compatible process. To find an alternative solution, the use of SiO2, a dielectric material commonly used in the IC process, for surface passivation is studied. Figure 5 shows the amplification results of the four TCMs in the GeneAmp PCR system and chipbased device. For all TCMs studied, the fluorescence intensities of the amplified products in both systems are very similar, indicating SiO2 can work as well as the polymer coating to passivate the chamber surface, resulting in high amplification efficiency. In addition to MgCl2, BSA has a significant influence on the amplification efficiency of the SiO2-passivated chip reactor. Figure 6 shows a series of both conventional (even lanes) and chip (odd lanes) PCR experiments with varying amount of BSA in the PCR master mix. For the conventional PCR, the amplification efficiencies at all BSA concentrations (from no BSA to 1 µg/µL) are very similar, whereas for chip PCR, the efficiencies are comparable Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
4245
Figure 5. Gel electrophoresis picture showing the amplification results of the four TCMs in conventional and chip settings (passivated with SiO2), with 2 mM MgCl2 for all reactions: (lane 1) 1 kb molecular weight marker; (lanes 2 and 3) F. cirrhosa (∼600 bp); (lanes 4 and 5) C. tinctorius (∼300 bp); (lanes 6 and 7) A. lobophilla (∼340 bp); (lanes 8 and 9) S. tetrandra (∼500 bp) in conventional and chip setting, respectively.
Figure 6. Gel electrophoresis picture showing the amplification results of C. tinctorius with varying BSA concentrations in both conventional (even lanes) and SiO2-passivated chip (odd lanes) PCR experiments: (lane 1) 1 kb molecular weight marker; (lanes 2 and 3) 1 µg/µL BSA; (lanes 4 and 5) 0.5 µg/µL BSA; (anes 6 and 7) 0.25 µg/µL BSA; (lanes 8 and 9) no BSA;
for those with BSA concentration higher than 0.25 µg/µL. If no BSA is present in the master mix for chip PCR experiments, the amplification efficiency will be dramatically lowered. Note that the amplified gene was C. tinctorius (∼300 bp) instead of F. cirrhosa (∼600 bp, the one used in MgCl2 optimization experiments), and the lower amplification efficiency in the chip reactor compared with the conventional setting is most likely due to the nonoptimized MgCl2 condition for C. tinctorius. The reusability of the chip reactor is very good. Lane 3 of Figure 7 shows the amplification result of F. cirrhosa in a “fresh” SiO2-passivated reactor at optimized PCR protocol, the fluorescence intensity of which is similar to that of the conventional one (lane 2) and is comparable to the previous results (lanes 2 and 3 of Figure 6). After the first reaction, the chamber was rinsed 3 times with boiling, autoclaved double-deionized water. Then, a 4246 Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
Figure 7. Gel electrophoresis picture demonstrating the reusability of the SiO2-passivated reactor: (lane 1) 1 kb molecular weight marker; (lane 2) F. cirrhosa in the conventional PCR; (lane 3) first chip PCR with F. cirrhosa as template; (lane 4) C. tinctorius in the conventional PCR; (lane 3) second chip PCR with C. tinctorius as template. Note that a single chip was used for the two chip PCR experiments, and the reaction chamber was rinsed 3 times with autoclaved doubledeionized water between the two chip PCR experiments.
second PCR was carried out with a different template (C. tinctorius). The single fluorescence band in lane 5 (close to the 298-bp marker, corresponding to C. tinctorius) of Figure 7 indicates that the cleaning procedure is sufficient to remove all the DNA that remained in the chamber after the first PCR experiment. Otherwise, a fluorescence band corresponding to F. cirrhosa will also appear in lane 5. The SiO2-passivated device can be reused many times (more than 5 times) without degradation in amplification efficiency and cross-contamination problems while the polymer-coated device cannot. A DNA concentration of higher than 27 ng/µL can be obtained in the chip-based PCR. This amount is sufficient for subsequent on-chip capillary electrophoresis, hybridization, and many other analyses. This chip PCR component will later be integrated with other lab-on-a-chip modules, including DNA extraction, separation, and detection modules, and be used in the standardization and toxicity assessment of traditional Chinese medicine. CONCLUSIONS We have developed an efficient Si-based DNA amplifier with integrated Pt temperature sensors and heaters for real-time accurate temperature sensing and control. With digital PI together with a gain scheduling control scheme, a temperature accuracy of (0.025 °C and heating rate of 15 °C/s as well as a cooling rate of 3.5 °C/s have been achieved. We have demonstrated that the 8-µL reaction chamber, with appropriate surface treatment and optimized PCR protocol, was capable of amplifying TCM DNA with a performance comparable to that of the conventional thermal cycler. Furthermore, the device has low power consumption, which is crucial to the development of a fully integrated and battery-operated instrument for on-site DNA analysis. ACKNOWLEDGMENT The authors thank the Biotechnology Research Institute of the Hong Kong University of Science and Technology (HKUST) and
Industry Department of the Hong Kong Special Administrative Region for the financial support. The authors also thank Dr. Karl Tsim and Dr. Nikolaus Sucher of the Biology Department at HKUST for providing raw materials of F. cirrhosa clone and raw materials for cloning A. lobophilla, C. tinctorius, and S. tetrandra, respectively. APPENDIX Chip Design. For the heater design in a constant-voltage operation mode, the resistance should be small enough to provide the required power for heating the reaction chamber to the denaturation temperature (95 °C) of PCR, which is given by
Rh ) V 2/P
(A1)
where Rh is the heater resistance (Ω), V is the voltage applied to the heater (V), and P is the power dissipation (W). Our target power requirement is 1.5 W. Thus, in a two-heater system with a voltage setting of 15 V, the resistance of each heater was designed to be 300 Ω. Knowing the sheet resistance of Pt (∼2.5 Ω/0), the number of squares of the heater line was determined to be 120 based on eq A2,
number of squares ) Lh/wh ) Rh/Rs
(A2)
where Lh and wh are length and width of the heater line, respectively, and Rs is the sheet resistance of the heater material. Furthermore, the width of the heater line should be designed to prevent electromigration (For Pt, current density of V/JRhth
(A3)
where J is the electromigration limit and th is the thickness of the heater line. For th ) 0.1 µm, wh > 50 µm. We therefore chose wh ) 250 µm, and the length of each heater line should be set to 3 cm (250 µm × 120). In designing the dimension of the temperature sensor, the following equation should be considered:
Received for review April 3, 2000. Accepted June 20, 2000.
b
Rs )
Vmax/G(2 - 1) IR∆T
where Rs is the resistance of the temperature sensor, Vmax is the maximum input voltage of the DAQ card, G is the gain of the DAQ card (in our case G is set to 100), b is the resolution of the DAQ card (12 bits), I is the excitation current of signal conditioning board (1 mA), R is the TCR of Pt (1.6 × 10-3/°C), and ∆T is the desired minimum detectable temperature change. With a desired minimum detectable temperature change of 0.025 °C, the temperature sensor resistance should be set to 610 Ω. Following the same procedure as in the heater design, the number of squares of the sensor line equals to 244. The width of the sensor line should be narrow enough to fit into the indentation of the meandering heater line (as shown in Figure 1a). In this case, the width of the sensor was set to 20 µm and resulted in a sensor length of 5.5 mm. Chip Fabrication. A P-type, double-side polished, and 〈100〉 oriented Si wafer (4-in. diameter and 400 µm thick) was used as a substrate. The wafer was coated with 0.1-µm SiO2 and 1-µm lowstress SixNy on both sides of the wafer by a dry thermal oxidation and low-pressure chemical vapor deposition (LPCVD) process, respectively. Then, photolithography was used to define the reaction chamber geometry on the bottom side of the wafer and plasma etched to expose the underlying oxide. On the top side of the wafer, heaters and temperature sensors were patterned using a lift-off method. Ti (10 nm) and Pt (100 nm) were sputtered on top of the developed positive photoresist (PR), and the unexposed PR was removed by acetone treatment in an ultrasonic bath. To complete the top-side fabrication, the wafer was annealed at 600 °C for 15 min. The next step was to remove the exposed bottomside SiO2 by a buffered oxide etch (BOE), followed by time-stop KOH etching down to a chamber depth of 350 µm. Finally, the wafer was diced into individual chip reactors and sealed from the bottom by bonding with Corning 7740 glass using an improved anodic bonding process.21 The glass had two holes drilled on it for reagent injection into the reaction chamber. Immediately after the injection, the holes were sealed by adhesive tape (3M 5419 low static polyimide tape), and the PCR was performed as described in the Experimental Section.
(A4)
AC000384B
Analytical Chemistry, Vol. 72, No. 17, September 1, 2000
4247
