Environ. Sci. Technol. 2007, 41, 7857-7863
MEMS Needle-type Sensor Array for in Situ Measurements of Dissolved Oxygen and Redox Potential JIN-HWAN LEE,† YOUNGWOO SEO,‡ TAE-SUN LIM,† PAUL L. BISHOP,‡ AND I A N P A P A U T S K Y * ,† Department of Electrical and Computer Engineering and Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio 45221
Microelectrodes are among the most accurate and reliable monitoring devices for measuring the dynamics of biofilm processes. This paper describes a novel needletype microelectrode array (MEA) for simultaneous in situ measurements of dissolved oxygen (DO) and oxidation reduction potential (ORP) fabricated using microelectromechanical systems (MEMS) technologies. The MEA exhibits fast response times for both DO and ORP measurements and shows a substantial increase in DO sensitivity. To demonstrate the versatility of the new sensor, it was applied to the measurement of DO and ORP microprofiles in a multispecies biofilm. This work demonstrates that the MEA is able to monitor local concentration changes with a high spatial resolution and provide the versatility of the microelectrode technique needed for biofilm studies as well as the capability for repetitive measurements. In addition, the use of MEMS technologies and batch fabrication approaches enables integration, high consistency, high yields, and mass production. With further development, it may be possible to add additional sensors to the MEA (e.g., pH, phosphate) and integrate them with a reference electrode.
Introduction Microelectrodes are electrochemical sensors with micrometer or smaller dimensions and are a powerful tool for microscale measurements (1, 2). In recent years, microelectrodes have been widely applied to the analysis of environmental samples, such as probing chemical composition or events in small soil pores or monitoring the concentration changes within biofilms or bacterial flocs (3-10). Unlike electrodes of conventional size, microelectrodes can be used in measurements with high spatial resolution. Other advantages include a smaller double-layer capacitance, faster response, and enhanced signal-to-noise ratio. Microelectrodes have been used to analyze redox potential, dissolved oxygen (DO), and pH in environmental samples on the microscale in water, biofilms, activated sludge flocs, and microbial granules (410). Recently, arrays of microelectrodes attracted even more interest as they can provide simultaneous measurements at multiple positions or of multiple analytes to enhance measurement reliability (11-13). * Corresponding author phone: (513) 556-2347; fax: (513) 5567326; e-mail:
[email protected]. † Department of Electrical and Computer Engineering. ‡ Department of Civil and Environmental Engineering. 10.1021/es070969o CCC: $37.00 Published on Web 10/19/2007
2007 American Chemical Society
Microelectrodes can be fabricated in a number of ways (1, 14, 15). Most commonly, microelectrodes are fabricated by pulling a glass pipet or capillary and inserting a metal wire (Pt or Au), or by filling with a low-melting-point alloy. Alternatively, a metal wire can be inserted into a glass pipet first and the metal/glass assembly pulled under heat to simultaneously decrease the wire diameter and tightly seal the metal within the glass capillary. Fine Au or Pt wires (∼10 µm diameter) can also be sealed into glass by inserting them into a 1 mm i.d. glass capillary and melting it around the wire. With these techniques, microelectrodes with 1-10 µm tips can be fabricated. Although these microelectrode fabrication methods are well-established, a number of inherent disadvantages still exist, such as low success rate, poor reproducibility, fragility, and difficulty in making a multisensor device (1, 5, 16-18). Microelectromechanical systems (MEMS) technologies permit the integration of different mechanical elements, sensors, and electronics in a small device and therefore can be suitable for the development of integrated multi-analyte sensors for biological applications (19, 20). In particular, MEMS silicon probes for electrical signal recording from neurons have been used for some time (21-23). Because MEMS devices are manufactured using batch fabrication techniques similar to those used for integrated circuits, unprecedented levels of reliability and sophistication can be implemented at a relatively low cost. In this work, MEMS technologies are used to successfully demonstrate a novel needle-type microelectrode array sensor for the simultaneous measurements of DO and oxidationreduction potential (ORP). The micrometer tip size and the needle nature of the sensor permit analysis to be performed in situ by penetrating samples. Gold was selected as the electrode material, as it could be used to directly measure both DO (17) and ORP (11, 12) and has a great potential for monitoring a large number of other analytes (2, 24). Multispecies biofilms may be comprised of numerous combinations of microbial processes, such as aerobic oxidation, nitrification, denitrification, sulfate reduction, and methanogenesis (3, 6). Thus, the evaluation of DO and ORP profiles in biofilms may lead to better understanding of the microbial ecology of biofilms.
Experimental Methods Microfabrication of DO and ORP MEA. The new needletype MEA is fabricated from 175 µm thick 45 mm × 50 mm borosilicate glass wafers (Erie Scientific, Pittsburgh, PA). To facilitate mass production, a batch fabrication approach is used. Twelve MEAs can be fabricated from a single glass wafer. The process, illustrated in Figure 1, has five major steps: dicing, etching, metallization, packaging, and sensor tip formation. Dicing. Glass wafers were cleaned with sulfuric peroxide solution (H2SO4 and H2O2 in a 7:3 (v/v) ratio) and cut with a dicing saw to yield an array of glass probes (Figure 1a). A 10 mil thick, 45 µm diamond grit resinoid blade (K&S MicroSwiss, Fort Washington, PA) was used to form 900 µm centerto-center spacing between each glass probe, 1 cm in length. Longer 1.5 cm cuts were made every four probes to define the edges of individual MEAs. Following metallization, an additional cross-cut indicated by a dashed line will be made to separate individual MEAs. However, keeping MEAs together at this time permits batch processing, increasing VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7857
FIGURE 1. Batch microelectrode array fabrication sequence: (a) dice glass wafer, (b) form glass probes by dicing, (c) taper probes in HF-based etchant, (d) use meniscus etching to sharpen probes, (e) deposit Au conductive layer, (f) pattern PCB, (g) fabricate PCB carriers, (h) use silver epoxy to establish electrical connections, and (i) coat microelectrodes with parylene insulating layer and fabricate recessed tips and exposed Au tips by beveling and etching. yield and reducing fabrication costs. The cut wafer was then annealed at 550 °C for 10 min in a programmable box furnace (Lindberg/Blue M, Thermo Scientific, Norwood, MA) to relieve stress from the dicing process. Etching. The glass probes were sharpened into needletype microelectrodes using an HF-based meniscus etching process we used previously in the fabrication of optic probes (25) and redox potential sensors (11, 12). The process uses an organic layer (paraffin oil) to modify the contact angle at the glass-etchant interface and consists of three steps. Initially, the probes were etched statically in the microelectrode etchant solution for 10 min with agitation to smooth the diced surface and reduce the probe dimensions to ∼80 µm in width. The microelectrode etchant solution was prepared by mixing HF, HNO3, and H2O in a 10:7:33 (v/v/v) ratio. Next, the glass probes were gradually pulled out using a computer-controlled (via LabView, version 7) motorized linear translation stage (Newport Corp., Irvine, CA) to taper them down to ∼10 µm in width at the tip (11). In the final etch step, a 1 mm length of the tapered probe was immersed into the same etchant for further sharpening using meniscus etching, yielding ∼200 nm tips (11). This final etch step is self-terminating, which permits consistent and reliable batch fabrication of microelectrode sensor tips. 7858
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 22, 2007
Metallization. The tapered glass probes with sharpened tips were metalized on both sides by thermal evaporation (Figure 1e). A 200 nm thick layer of gold was deposited as a conductive layer, with a 20 nm thick layer of titanium to enhance adhesion between glass and the gold layer. The gold layer serves a dual function as a sensing layer and to establish an electrical connection. Packaging. Following the batch fabrication steps, the metalized glass wafer was cross-cut as indicated by the dashed line in Figure 1a to separate individual MEAs. For easier handling and to establish electrical connections with individual sensors, MEAs were packaged with printed circuit board (PCB) carriers (Figure 1f). Microelectrodes were fixed to carriers with UV-cured epoxy (3301, Loctite, Rocky Hill, CT). Conductive silver epoxy (Ablebond 8700E, Emerson & Cuming, Billerica, MA) was used to establish the electrical connections between microelectrodes and carriers (Figure 1h). To insulate individual microelectrodes, a 1.5 µm thick layer of parylene C was coated over the entire substrate (PDS 2010 Parylene Labcoter, Specialty Coating Systems). Sensor Tip Formation. Two ORP and two DO sensors were fabricated in each MEA with different tip structures (Figure 1i). Microelectrode tips were first beveled (BV-10 Beveler, Sutter Instrument Co.) at 45° above horizontal (for better penetration) for 30 min on a rotating plate under visual control through a microscope to remove parylene and Ti/Au layers. The resulting structure with exposed Au, schematically shown in Figure 2a, formed the solid-state ORP sensors. Gold gives more reliable measurement of ORP than platinum for this application, as platinum may catalyze some additional reactions at its surface (26). The DO sensors are polarographic recessed cathode Au electrodes. Thus, for DO microelectrodes, a recess was created at the tip of each microelectrode (Figure 2b). To create the recess, the glass core and Ti exposed by beveling were etched using HF-based etchant for 5 min. The beveling and etching steps permit precise control of the recess opening size and depth. The exposed Au was etched in a 1:4:40 (m/m/v) mixture of I2, KI, and H2O for 3 min to relocate the Au sensing area inside the formed recess. Microelectrodes were cleaned ultrasonically in deionized (DI) water after each etching step. For both sensors, tip diameters were on the order of 1-2 µm. Fabrication of Conventional ORP Microelectrodes. The conventional ORP microelectrodes (MEs) are solid-state electrodes made from a platinum wire fused into a glass micropipette with a tip diameter of 10 µm (Figure 2c). ORP MEs were fabricated by following the procedures described by Bishop and Yu (6). Briefly, a 10 cm long lead-glass micropipette with 1.50 mm o.d. and 0.75 mm i.d. (PG1-15-4, World Precision Instruments) was melted under an open flame and pulled to shrink the diameter. A 99.99% pure Pt wire (26,716-3, Aldrich) was etched in a 2 M KCN solution to reduce its diameter to ∼5 µm. The Pt wire was then inserted into the pulled glass pipet, and the glass micropipette tip was then melted to coat the tip of the platinum wire. The electrode was then beveled to expose the platinum wire surface. The microelectrode was completed by connecting the wire to a BNC cable. Fabrication of Conventional DO Microelectrodes. The conventional DO microelectrodes are polarographic recessed cathode Au electrodes with a tip diameter of 2-5 µm (Figure 2d). DO MEs were fabricated by following the procedures described by Yu and Bishop (27). Briefly, a 15-cm long lead glass micropipette with 1.20 mm o.d. and 0.69 mm i.d. (B12069-15, Sutter Instrument Company) was pulled with a micropipette puller (Sutter Instrument Co.). The tips of pulled pipettes were broken with tweezers and then filled half way with melted bismuth alloy (Belmont Alloy 2451: 44.7% bismuth, 22.6% lead, 19.1% indium, 8.3% tin, 5.3% cadmium). Glass pipettes filled with bismuth alloy were beveled at 45°
FIGURE 2. Cross-sectional diagrams of sensor tips: (a) ORP MEA, (b) DO MEA, (c) ORP ME, and (d) DO ME. and a 15 to 25 µm recess was created using 2M KCN solution. Gold plating was applied on the bismuth alloy surface to complete the microelectrode (ME) fabrication. Characterization. ORP Characterization. Three ORP reference solutions of 450, 228, and 90 mV at 25 °C (Sensorex Corp., Garden Grove, CA) were used to investigate the performance of the ORP sensors. A commercial Ag/AgCl millielectrode (MI-401, Microelectrodes, Inc.) was used as the reference. If the sensor readings were outside the ( 10 mV range of the reference solution recommended by the American Society for Testing and Materials (ASTM) standard D1498 (28), the sensors were discarded. Electrodes were rinsed with distilled water after each measurement. DO Characterization. The DO sensor characterization was performed inside a Faraday cage to minimize electromagnetic interference. A commercial Ag/AgCl milli-electrode was used as reference. DO sensors were polarized for several hours prior to measurements (17, 27). A Chemical Microsensors II picoammeter (Diamond General Development Corp., product no. 1231) was used to supply the -750 mV potential (vs Ag/AgCl) during polarization (17). A mineral-salt solution was prepared as a test solution by aeration with pure nitrogen gas (0% O2 saturation or 0 mg/L), a gas containing 10% O2 and 90% N2 (10% O2 or 4.1 mg/L), and air (21% O2 or 8.7 mg/L). The aeration was applied for at least 20 min to establish a stable concentration. A commercial DO milli-electrode (MI730, Microelectrodes, Inc.) was used to confirm the DO content in the test solution and during calibration. Development of a Multispecies Aerobic Biofilm. A multispecies biofilm was developed using activated sludge from a municipal wastewater treatment plant aeration tank (Mill Creek WWTP, Cincinnati, OH). Activated sludge was decanted, washed with DI water several times, and transferred to Petri dishes. Frosted glass slides (12-544-5CY, Fisher Scientific) were placed inside the Petri dishes for biofilm formation and growth. After 24 h, glass slides containing biofilm were suspended from the top of a closed reactor, schematically illustrated in Figure 3. A two-ring polycyclic aromatic hydrocarbon, naphthalene, was used as the sole carbon source. To obtain a stable influent concentration of naphthalene (around 18-20 mg/L), a 20 L glass jar was used as a feed solution tank, containing an excess amount of crystal naphthalene (4 g/L) mixed with mineral-salt solution (29). The mineral-salt solution was prepared by mixing 32 mg/L of NaNO3, 10 mg/L of NH4Cl, 40 mg/L of Na2HPO4, 10 mg/L of KH2PO4, 1.4 mg/L of CaCl2, 3.8 mg/L of MgSO4, 0.65 mg/L of FeCl3, 11.2 µg/L of MnSO4, 0.7 µg/L of CuSO4, 0.4 µg/L of NaMoO4, and 12 µg/L of ZnSO4. To remove naphthalene
FIGURE 3. Experimental setup for biofilm development control system: (1) feed tank; (2) peristaltic pump; (3) closed reactor; (4) DO probe; (5) DO controller; (6) micropump; (7) H2O2 tank. particles from the influent, a fabric filter was attached to the outlet line of the naphthalene tank and the flow was cycled through the closed biofilm reactor with a peristaltic pump (Cole-Parmer Instrument Co., Chicago, IL). Feed solution with naphthalene was prepared in advance; naphthalene was allowed to dissolve for 3 days before it was used. The feed solution was changed every 3 days. Hydrogen peroxide (0.5%) was added as a supplemental oxygen source to maintain 5.0 ( 0.5 mg/L DO in the reactor. Biofilm Microprofile Measurement. Figure 4 shows the experiment setup for the microprofile measurements inside the biofilm. Multispecies biofilm grown on slides were taken from the closed reactor and placed in the open-channel test chamber to obtain oxygen and redox potential profiles. The open-channel chamber (7.85 cm long, 2.9 cm wide, 1.2 cm high) was mounted under a stereomicroscope with a CCD camera (model JE-3662 HR, Javelin Electronics, Torrance, CA) situated on a Micro-g series high-performance vibration isolation table (63-527-01, TMC, Peabody, MA) inside a Faraday cage (TMC, Peabody, MA). Feed solution with the same composition as that in the closed biofilm reactor was continually aerated and recycled (2.35 mL/min) through the open-channel chamber with a peristaltic pump mounted outside of the Faraday cage. The biofilm sample was placed in the chamber for at least 30 min before performing measurements. Microelectrodes were mounted and positioned using a motor-driven 3-D micromanipulator (model 11N, Narisige, Japan). Oxygen profiles were measured at 1050 µm intervals into the biofilm; ORP profiles were measured at 50 µm intervals. To monitor the reproducibility of the electrodes and any possible damage to the biofilm by the VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7859
FIGURE 4. Experimental setup for the microprofile measurements: (1) video display; (2) data acquisition system; (3) three-dimensional (3-D) micromanipulator; (4) ME or MEA; (5) charge-coupled device (CCD) camera; (6) stereomicroscope; (7) open-channel chamber; (8) vibration isolation table; (9) Faraday cage; (10) peristaltic pump; (11) feed tank; and (12) oxygen cylinder. microelectrode penetration, measurements in the biofilm were conducted during both the microelectrode penetration and withdrawal.
Results and Discussion Micromachined DO and ORP MEA. Needle-type microelectrode arrays for DO and ORP measurements were successfully fabricated, as illustrated in Figure 5a. This new MEA sensor consists of four 1-cm long probes at 900 µm center-to-center spacing packaged on a PCB carrier. The size of the microelectrode tips are on the order of 1 µm. Twelve MEA sensors were fabricated from a single 45 mm × 50 mm borosilicate glass wafer using the described batch microfabrication process. This batch processing substantially simplified fabrication, permitting increased uniformity and consistency of microelectrodes, as well as savings in time and costs. Figure 5c compares MEA tips demonstrated in this work with the tips of commercially available millielectrodes and conventional MEs. The recess size (diameter and length) plays a key role in determining the sensitivity of recessed DO sensors (30). To fabricate a recess, one of the four pyramidal faces in a probe was ground at 45° to expose the glass core. The angled sensor tips are expected to permit easier penetration into biofilms. The demonstrated DO MEA sensors contain ca. 1-2 µm diameter recess (Figure 5b). To verify recess functionality, DO microelectrode sensors were tested in oxygenated saline before and after the glass- and gold-etching steps during recess fabrication. When the Au sensing area was exposed to saline before glass etching, a current of ∼0.1 µA was measured with respect to a Ag/AgCl reference. The current remained constant and independent of DO concentrations, exhibiting no selectivity toward DO. The same results were observed for microelectrodes following glass etching, but prior to Au etching. Only following the Au etching to relocate the sensing area inside the recess did the DO microelectrode exhibit sensitivity to different DO concentrations and currents in the picoampere (pA) range. Microelectrode Array Calibration. Dissolved Oxygen. The developed DO sensors were characterized using oxygenated mineral-salt solution in the 0-9 mg/L O2 range. DO microelectrode array sensors were polarized for at least several hours before calibration, depleting oxygen from the cathode and reducing disturbances in the measurements (31). Typical three-point calibration curves for the DO MEA in mineralsalt solutions with respect to a Ag/AgCl reference at 20 °C are shown in Figure 6a. The MEA data are compared with that from a conventional ME. High correlation coefficients 7860
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 22, 2007
FIGURE 5. (a) Photograph of the integrated DO and ORP microelectrode array sensor packaged on a PCB carrier. (b) A scanning electron micrograph of the recessed electrode tip with approximately 1.7 µm wide recess. Inset: a pyramidal-shape microelectrode tip prior to beveling. (c) Comparison of electrode tips of commercial DO and ORP milli-electrodes (COM), conventional DO and ORP microelectrodes (ME), and the integrated microelectrode array (MEA) used in this work. were found for each case. The MEA sensor performed linearly and exhibited a high sensitivity of ∼147 pA/mg/L. The time for 90% response was typically less than 20 s, which is much shorter than that of macroscale commercial oxygen electrodes. These characteristics are a marked improvement over the conventional MEs, which exhibited similarly fast response times but lower sensitivities of ∼6 pA/mg/L. While both MEAs and MEs have small surface areas, the substantial increase in sensitivity and higher generated currents are attributed to differences in the electrode structure. As Figure 2d illustrates, conventional MEs in this
FIGURE 6. (a) Calibration curves of the DO ME and MEA in the mineral-salt solution. (b) Standardization curves of the ORP MEA, ME, and commercial milli-electrode (COM) sensors. and other works (3-7, 17, 24, 27) use bismuth alloy (e.g., Belmont 2451, Belmont Metals, Brooklyn, NY) to establish an electrical connection between a gold cathode and copper wire. The electrical conductivity of such an alloy has been reported to be approximately 0.019 × 106 Ω/cm (32), which is less than 4.2% of the conductivity of gold (0.455 × 106 Ω/cm (33)) used as the conducting layer in the MEAs (Figure 2b). The use of materials with higher electrical conductivities in MEA fabrication yields lower current losses and thus higher sensitivity. Using more conductive materials in the construction of conventional MEs may improve their sensitivity. Redox Potential. Figure 6b shows the standardization curves for the three ORP electrodes against three ORP standard solutions. The slopes of the three curves are 0.92 for the MEA, 0.99 for the conventional ME, and 0.88 for the commercial milli-electrode. The slopes of the curves are very close to the theoretical value of 1.00. The comparison shows that the MEA has the same or more accurate readings than those of the other two electrodes. The response time of the ORP MEA was typically on the order of 1 s, which is substantially faster than that of the commercial millielectrode that exhibited response times in the range of minutes. Application of MEA to Microprofile Measurements in Biofilm. To demonstrate the versatility of the new MEA electrode technique in biofilm studies, DO and ORP micro-
profiles were obtained and compared using both the MEA and the conventional ME under the same conditions. Multispecies biofilm grown on slides for 1 month was taken from the closed reactor and placed in the open-channel test chamber to obtain oxygen and oxidation-redox potential profiles. The same biofilm was used with both MEA and ME sensors. Identical electrode positioning was achieved by using a small marker on the biofilm and a stereomicroscope with a CCD camera. In the case of the conventional MEs, DO and ORP electrodes were separately prepared and mounted on the 3-D micromanipulator before each measurement. The MEA was constructed with integrated DO and ORP sensors in one body and thus could obtain DO and ORP profiles simultaneously. Oxygen profiles were measured at 10 or 50 µm intervals, and ORP profiles were measured at 50 µm intervals. The entire measurement process was monitored using a color monitor connected to a stereomicroscope with a video camera; the image was lit from above with a highintensity lamp during video observation. The bottom substratum of the biofilm and biofilm thickness were defined as the point where the electrode hit the substratum and visually bent. The whole system had clean electrically grounded lines. DO Microprofile. Figure 7 shows the microprofile changes obtained using both the MEA and ME sensors. The DO in the bulk solution was around 8.5 mg/L and decreased through the biofilm’s mass transfer boundary layer to 5.9 mg/L with VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7861
FIGURE 7. (a) Microprofiles of DO (b, O) and ORP (9, 0) in biofilm. Filled shapes represent microprofiles with the MEA (one sensor of each type), and empty shapes represent microprofiles with the ME. The data are the average of six measurements. (b) DO microprofile recorded with the MEA during the biofilm insertion (b) and withdrawal (O). The data are the average of six measurements. the MEA or to 6.2 mg/L with the ME at the biofilm surface. The thickness of the DO mass transfer boundary layer was estimated to be around 200 µm, and DO decreased by about 2.6 mg/L in this region. Inside the biofilm, oxygen decreased continually and was totally depleted at 700 µm depth, according to both the MEAs and MEs. This result confirms that an oxic zone inside of the mixed species biofilm is several hundred micrometers thick. At the biofilm depth of 300500 µm, small concentration differences of about 1 mg/L DO concentration were observed between the two electrodes. It is not clear whether this was caused by the heterogeneity of the biofilm, due to slight differences of positioning of the electrodes (microelectrode spacing in the MEA is 900 µm center-to-center), or to signal differences between the MEA and ME. Nevertheless, there is a strong correlation between the MEA and ME measurements (r (82) ) 0.98, p < 0.01). To monitor reproducibility of the microprofile measurements and any possible damage to the biofilm by the MEA penetration, the DO measurements in the biofilm were performed during both penetration and withdrawal (i.e., inand-out technique). Figure 7b illustrates that the same microprofile is obtained using the MEA during both penetration and withdrawal. Correlation analysis on these data yields a coefficient of r (82) ) 0.97, p < 0.01. No structural damage to the biofilm was observed during these measurements. ORP Microprofile. The ORP in the bulk solution was ca. 180-190 mV (1000 µm from the surface of the biofilm) and decreased gradually to 160 mV at the biofilm surface (see 7862
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 22, 2007
Figure 7). In the biofilm, the redox potential profiles provided by both the MEA and the ME also exhibited a gradual decrease from the surface to the substratum. The MEA measured 119.6 mV near the substratum, which compares well with the 125.3 mV measured by the ME. The ORP profiles revealed that both the MEA and ME performed similarly. The ASTM standard D1498 (28) suggests that the measured redox potentials should be within (10 mV of the nominal redox potential for a good redox electrode. Both electrodes behaved within the error range. A correlation analysis on these data yields a coefficient of r (64) ) 0.96, p < 0.01, indicating a nearly perfect correlation. Among the 66 measurement points, all but five differed by less than 10 mV. As with the DO measurements, the difference between electrodes is possibly due to the biofilm heterogeneity or due to slight differences of positioning of the electrodes. Discussion. The new needle-type MEA sensors for simultaneous measurements of DO and ORP were developed using MEMS technologies and applied to in situ microprofiling of a multispecies biofilm. The MEA exhibited fast response times for both DO and ORP measurements and showed a substantial increase in DO sensitivity. To date, such measurements were possible only with conventional microelectrodes and had to be performed only one at a time. This work demonstrated that the MEMS MEA is able to monitor local concentration changes in small structures with a high spatial resolution and provide the versatility of the microelectrode technique needed for biofilm studies as well as the capability for repetitive measurements. In addition to having the versatility of conventional microelectrodes, the use of MEMS technologies and batch fabrication approaches enables integration, high consistency, high yields, and mass production. With further development, it may be possible to add additional sensors to the MEA (e.g., pH, phosphate) and integrate a reference electrode. Using separate reference and working microelectrodes works well for laboratory measurements and is relatively easy to fabricate (4, 5). However, the drawback is that their use requires good shielding and grounding to minimize electrical interference. Thus, integrating a reference electrode and a signal conditioning circuitry with the MEAs should overcome many of the shortcomings of today’s conventional microelectrodes. Fully integrated multi-analyte MEAs may be used to obtain direct information from measurements inside heterogeneous biological systems. In environmental monitoring, measurements are typically made on samples extracted from the site, yet such measurements may not be acceptable. Microscale in situ measurements are essential for proper monitoring of environmental conditions, to determine impacts of environmental stressors, and in laboratory reactors, to properly control the environment.
Acknowledgments This work was supported by grants from the National Science Foundation (CBET-0529217) and the National Institute of Environmental Health Sciences, under the Superfund Basic Research Program (SBRP) (P42ES04908-14/Project 5). The authors thank Dr. Peng Jin for assistance with beveling and characterization.
Literature Cited (1) Zoski, C. G. Ultramicroelectrodes: design, fabrication, and characterization. Electroanalysis 2002, 14, 1041-1051. (2) Wightman, R. M. Probing cellular chemistry in biological systems with microelectrodes. Science 2006, 311, 1570-1574. (3) Zhang, T. C.; Pang, H. Applications of microelectrode techniques to measure pH and oxidation-reduction potential in rhizosphere soil. Environ. Sci. Technol. 1999, 33, 1293-1299.
(4) Bishop, P. L.; Zhang, T. C.; Fu, Y. C. Effects of biofilm structure, microbial distributions and mass-transport on biodegradation processes. Water Sci. Technol. 1995, 31, 143-152. (5) Pang, H.; Zhang, T. C. Fabrication of redox potential microelectrodes for studies in vegetated soils or biofilm systems. Environ. Sci. Technol. 1998, 32, 3646-3652. (6) Bishop, P. L.; Yu, T. A microelectrode study of redox potential change in biofilms. Water Sci. Technol. 1999, 39, 179-185. (7) Zhang, T.; Fu, Y.; Bishop, P.; Kupferle, M.; Fitzgerald, S.; Jiang, H.; Harmer, C. Transport and biodegradation of toxic organics in biofilms. J. Haz. Mater. 1995, 41, 267-285. (8) Bezbaruah, A. N.; Zhang, T. C. Fabrication of anodically electrodeposited iridium oxide film pH microelectrodes for microenvironmental studies. Anal. Chem. 2002, 74, 5726-5733. (9) Bezbaruah, A. N.; Zhang, T. C. pH, redox, and oxygen microprofiles in rhizosphere of bulrush (Scirpus validus) in a constructed wetland treating municipal wastewater. Biotechnol. Bioeng. 2004, 88, 60-70. (10) Jang, A.; Szabo, J.; Hosni, A. A.; Coughlin, M.; Bishop, P. L. Measurement of chlorine dioxide penetration in dairy process pipe biofilms during disinfection. Appl. Microbiol. Biotechnol. 2006, 72, 368-376. (11) Lee, J.-H.; Jang, A.; Myers, R.; Bhadri, P.; Timmons, W.; Beyette, F.; Bishop, P.; Papautsky, I. Fabrication of microelectrode arrays for in situ sensing of oxidation reduction potentials. Sens. Actuators, B 2006, 115, 220-226. (12) Jang, A.; Lee, J.-H.; Bhadri, P.; Kumar, S. A.; Timmons, W.; Beyette, F. R.; Papautsky, I.; Bishop, P. L. Miniaturized redox potential probe for in situ environmental monitoring. Environ. Sci. Technol. 2005, 39 (16), 6191-6197. (13) Errachid, A.; Ivorra, A.; Aguilo, J.; Villa, R.; Zine, N.; Bausells, J. New technology for multi-sensor silicon needles for biomedical applications. Sens. Actuators, B 2001, 78, 279-284. (14) Electroanalytical Chemistry; Bard, A. J., Ed.; Marcel Dekker: New York, NY, 1994. (15) Scanning Electrochemical Microscopy; Bard, A. J., Mirkin, M. V., Eds.; Marcel Dekker: New York, 2001. (16) Revsbech, N. P.; Ward, D. M. Oxygen microelectrode that is insensitive to medium chemical composition: use in an acid microbial mat dominated by Cyanidium caldarium. Appl. Environ. Microbiol. 1983, 45, 755-759. (17) Linsenmeier, R. A.; Yancey, C. M. Improved fabrication of doublebarreled recessed cathode O2 microelectrodes. J. Appl. Physiol. 1987, 63, 2554-2557. (18) Bezbaruah, A. N.; Zhang, T. C. An innovative electro-corrodion recess creation technique for improved microelectrode fabrication. Water Res. 2002, 36, 4428-4432.
(19) Kovacs, G. T. A. Micromachined Transducers Sourcebook; McGraw Hill: New York, 1998. (20) Madou, M. Fundamentals of Microfabrication, 2nd ed.; CRC Press: Boca Raton, FL, 2002. (21) Najafi, K.; Tochizuki, T.; Wise, K. D. A high-yield IC-compatible multichannel recording array. IEEE Trans. Electron Devices 1985, 32, 1206-1211. (22) Campbell, P. K.; Jones, K. E.; Huber, R. J.; Horch, K. W.; Norman, R. A. A silicon-based, three-dimensional neural interface: manufacturing processes for an intracortical electrode array. Trans. Biomed. Eng. 1991, 38, 758-68. (23) Rousche P. J.; Norman, R. A. Chronic recording capability of the Utah intracortical electrode array in cat sensory cortex. J. Neurosci. Methods 1998, 82, 1-15. (24) McRipley, M. A.; Linsenmeier, R. A. Fabrication of a mediated glucose oxidase recessed microelectrode for the amperometric determination of glucose. Electroanal. Chem. 1996, 414, 235246. (25) Srinivasan, P.; Beyette, F. R.; Papautsky, I. Micromachined arrays of cantilevered glass probes. Appl. Opt. 2004, 43 (4), 776-782. (26) ORP Theory; Van London-pHoenix Company: Houston, TX. http://www.vl-pc.com/orptheory.html. (27) Yu, T.; Bishop, P. L. Stratification and redox potential change in an aerobic and sulfate-reducing biofilm studied using microelectrodes. Water Environ. Res. 2001, 73, 368-373. (28) American Society for Testing and Materials (ASTM) D1498. Standard practice for oxidation-reduction potential of water; ASTM International: West Conshohocken, PA, 2000. (29) Jin, P. Ph.D. Dissertation, University of Cincinnati, OH, 2007. (30) McGuyer, J. C.; Vasile, M. J.; Schubert, R. W. Modeling a micromanufactured open recess-tip oxygen microelectrode. IEEE Biomed. Eng. Conf. 1997, 382-385. (31) Ober, E. S.; Sharp, R. E. A microsensor for direct measurement of O2 partial pressure within plant tissues. J. Exp. Bot. 1996, 47, 447-454. (32) ASM Handbook: Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, 10th ed.; ASM International: Materials Park, OH, 1992; Vol. 2. (33) Introduction to Solid State Physics, 7th ed.; Kittel, C., Ed.; Wiley & Sons: New York, 1996.
Received for review April 24, 2007. Revised manuscript received September 5, 2007. Accepted September 7, 2007. ES070969O
VOL. 41, NO. 22, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
7863
