Environ. Sci. Technol. 2007, 41, 5447-5452
An Innovative Microelectrode Fabricated Using Photolithography for Measuring Dissolved Oxygen Distributions in Aerobic Granules SHAO-YANG LIU,† GANG LIU,‡ YANG-CHAO TIAN,‡ YOU-PENG CHEN,† H A N - Q I N G Y U , * ,† A N D F A N G F A N G † School of Chemistry, University of Science and Technology of China, Hefei, 230026, China and National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei, 230026, China
In this work an innovative microelectrode was successfully fabricated using photolithography for determination of dissolved oxygen distributions in aerobic granules, which were sampled from a nitrifying sequencing batch reactor. A negative photoresist, SU-8, was used as a substrate for the microelectrode and a 70 µm wide needle was photopatterned on it. The microelectrode could be renewed several times. Cyclic voltammetry analysis and dissolved oxygen measurement demonstrated that the microelectrode was stable and reliable. Dissolved oxygen distribution in a nitrifying granule was successfully monitored with the microelectrode. The profiles indicated that the main active part of the nitrifying granule was the upper 150 µm layer. Using the procedures developed in this work, microelectrodes of the desired shape could be constructed with precise size control at micrometers-scale.
Introduction Microelectrodes (MEs) have been regarded as powerful tools for microscale measurements, and they are widely used for analysis of environmental samples over the last decades (13). They have several advantageous properties, including small double-layer capacitance, fast response, high signalto-noise ratio, and low iR drop. Microelectrode arrays (MEAs) have attracted even more interest. They provide the possibility of simultaneous measurements of various parameters and/ or at different positions (4), and they could provide redundant information to enhance measurement reliability (1). Because of their specific properties along with their sturdy nature, metal and metal-metal oxide MEs have been intensively used to analyze environmental samples at microscales and to measure dissolved oxygen (DO), redox potential, and pH in water, wastewater, biofilms, activated sludge flocs, and microbial granules (5-11). There are two widely accepted methods to fabricate metal and metal-metal oxide MEs for environmental studies (8). One is to shield a tapered metal wire with a glass micropipette, and the other is to fill a glass micropipette with a low melting point alloy. Although many efforts have been made to improve these methods, some inherent disadvantages still exist, such as * Corresponding author fax: +86 551 3601592; e-mail: hqyu@ ustc.edu.cn. † School of Chemistry. ‡ National Synchrotron Radiation Laboratory. 10.1021/es070532g CCC: $37.00 Published on Web 06/22/2007
2007 American Chemical Society
complicated fabrication procedures, low success rate, poor reproducibility, frangibility, nonrenewability, and difficulty in a making multisensor device (5, 8, 12, 13). Thus, innovative fabrication methods should be developed to overcome these shortcomings. Photolithographic techniques, which are used for manufacturing small electrodes of any required planar patterns with excellent reproducibility, have attracted increasing interest and have been widely employed to fabricate ME(A)s in electrophoresis microchips (14), flow-injection systems (15), other microelectromechanical systems (16, 17), or electrochemical multianalyte detecting devices (18). Refined needle-type and other electrodes with a special shape are essentially desired for environmental purposes, e.g., exploring the inner parts of biofilms or microbial aggregates in natural and artificial systems. Needle-type MEAs (19, 20) and microchannels (21) based on silicon or glass have be developed by etching methods, but the fabrication processes are complex. Therefore, a microfabrication procedure with an easy shape-control is desirable to extend the application of lithographic techniques in environmental fields. In the present work, a novel needle-type gold microelectrode was fabricated using a lithography method for environmental sample analysis. An innovative procedure based on the utilization of a negative photoresist, SU-8, was proposed, and ME(A)s of any desired entire shape could be constructed at micrometer-scale readily and reproductively. The needle of the microelectrode was deliberately designed, and could be cut and renewed several times, resulting in a significantly prolonged lifetime. Gold was selected as the electrode material as it could be directly used to measure an important environmental parameter, DO (13), and has a great functionalizing potential to monitor a large number of environmental factors (22, 23). Recently aerobic granules have received increasing interest because they have a dense and strong microbial structure, and good settling ability, compared with activated sludge flocs (24). The evaluation of the DO distribution in aerobic granules will produce a better understanding their functional and structural characteristics in bioreactors. To elucidate the DO distribution in granules, it is important to carry out measurements inside granule samples by using oxygen microelectrodes. Therefore, the main objective of this study was to fabricate new and reliable DO microelectrodes that are small and sturdy enough for in-situ microscale DO measurements of aerobic granules. This paper will present the procedures to develop the microelectrodes, and to evaluate their performance for DO measurement.
Experimental Section Chemicals. Potassium ferricyanide, potassium ferrocyanide, potassium chloride, sodium hydroxide, ethanol, and isopropanol, all purchased from Shanghai Chemical Reagent Company, were of analytical grade and used as received. Deionized water was used throughout the experiments. Microfabrication of ME. To facilitate mass production, each glass plate substrate was patterned to contain six MEs. The fabrication procedures were as follows (Figure 1): Bottom SU-8 layer fabrication: The glass plate (65 mm × 65 mm × 2 mm) was washed with detergent, rinsed with water and acetone, and dried at 110 °C for 30 min. The glass plate was then spin-coated with a thin layer of OmniCoat (MicroChem, Newton, MA) at 2000 rpm for 30 s as an adhesive and baked at 200 °C for 2 min on a hot plate. Later, a 50 µm thick SU-8 (SU-8 2050, MicroChem) was applied, and a twoVOL. 41, NO. 15, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5447
FIGURE 1. Schematic diagram of the fabrication procedures for the innovative microelectrodes (For clarity, only one microelectrode is shown). step soft-bake (65 °C for 10 min and 95 °C for 50 min) was performed on a level hot plate. The photoresist was then patterned using a photomask of the ME shape (i-line, 150 mJ/cm2). An oven was used for the post exposure bake (PEB) (65 °C for 5 min; rise to 95 °C with a ramp of 2 °C/min; 95 °C for 15 min) (Step 1 in Figure 1). After baking SU-8, the substrate was allowed to cool down to ambient temperatures on the hot plate or in the oven to release stress. Gold track formation: A thin layer of titanium and 500 nm thick gold were sputtering deposited on the SU-8 layer, followed by a spin-coated 600 nm thick positive photoresist (Ruihong, Suzhou Ruihong Co., China; soft-bake in oven: 65 °C for 5 min; rise to 95 °C with a ramp of 2 °C/min; 95 °C for 5 min). The substrate was then aligned and exposed using a photomask with the pattern of metal tracks (UV exposure: g-line, 58 mJ/cm2; development: 0.4% NaOH, 1 min). After the development, ion-beam etching (IBE) was conducted until the metal uncovered by the photoresist was completely removed (Step 2 in Figure 1). Subsequently, the remaining positive photoresist was removed using ethanol and the substrate was cleaned in oxygen plasma for 1 min (30 W, ME-3A MEIRE, Institute of Microelectronics, China). Lift-off procedure was not employed here because of the poor adherence between the metal and SU-8. Upper SU-8 Layer Fabrication: A second SU-8 layer (50 µm thickness) was applied (soft-bake in oven: 65 °C for 5 min; rise to 95 °C with a ramp of 2 °C/min; 95 °C for 15 min) and patterned using a photomask with the upper layer ME shape and the contact pads (i-line, 120 mJ/cm2). Relatively long time PEB was performed in the oven to achieve a desired cross-link density (65 °C for 5 min; rise to 95 °C with a ramp of 2 °C/min; 95 °C for 75 min) (Step 3 in Figure 1). Then, the substrate was left in air for 5 days to release stress. Thereafter, it was developed in SU-8 Developer (MicroChem) for 2 h. MEs would detach from the glass plate automatically. They were rinsed in isopropanol immediately after development (Step 4 in Figure 1). Packaging: In the final step, coated copper wires (the insulated coating at the very ends were sanded off) were 5448
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 15, 2007
attached to the contact pads with silver conductive paint (SPI no. 5002, Washington, DC), and the pads were insulated by silicon rubber (NANDA no. 703, Nanjing Daxue Ltd., China). Electrochemical Analysis. All electrochemical measurements were performed using an electrochemical detector (CHI 800, CH Instruments Inc., Austin, TX) inside a Faraday cage at 25 °C. Three-electrode configuration was employed. A commercial Ag/AgCl electrode (CHI111, CH Instruments) was used as reference and a platinum wire counterelectrode (CHI115, CH Instruments) was used, unless the integrated gold counterelectrode was investigated. The measurement system was connected to a ground line separated from the building grounding. Chronoamperometry was employed to measure DO concentration. The cathode was polarized at -0.75 V (vs Ag/ AgCl) and limiting current was recorded. The DO level was controlled by saturating the solution with standard gas containing 0, 5, 10, 15, or 21% oxygen. The gases were introduced to the reaction cell at least 20 min, to reach an equilibration between the gases and solvent. To avoid bubble agitation, the gas supply to the reaction cell was stopped in each measurement, after which gas supply was resumed immediately. DO Measurement for Nitrifying Granules. A columntype sequencing batch reactor (SBR) with a working volume of 4 L (a height of 120 cm and an internal diameter of 7 cm) was used for cultivating nitrifying granules (Figure 2a). The reactor was operated sequentially in 4 h cycle, including 3 min of influent filling, 221 min of aeration, 2 min of settling, 4 min of effluent discharging, and 10 min of idle. Effluent was drawn from the middle port of the reactor with a volumetric exchange ratio of 50%. Air was introduced at the reactor bottom. Synthetic wastewater used in this work was composed as follows (in mg/L): NH4Cl 800; K2HPO4 160; CaCl2 18.8; MgCl2‚6H2O 7.4; FeSO4‚7H2O 4.2. In addition, microelement solution was added with a dose of 1.0 mL per 10 L synthetic wastewater, which contained (in mg/L): EDTA 50.0; ZnSO4‚7H2O 22; CaCl2‚2H2O 8; MnCl2‚4H2O 5; FeSO4‚
FIGURE 3. Photograph of the innovative microelectrode and three magnified microscopic images of different sections of the needle.
FIGURE 2. (a) Microphotograph of the nitrifying granules investigated in this work. (b) Schematic diagram of the test chamber used for measurement of the DO microprofile of the nitrifying granules. 7H2O 5; (NH4)6Mo7O24‚4H2O 1.1; CuSO4‚5H2O 1.57; and CoCl2‚ 6H2O 1.6. The pH was kept in a range of 7.0-8.5 by a dose of NaHCO3. Seed sludge was obtained from an aeration tank in the Wangxiaoying Municipal Wastewater Treatment Plant, Hefei, China. Seed sludge was brown in color and had a fluffy, irregular, and loose-structure floc morphology. After 3 months of operation, the seed sludge in the reactor was nearly totally granulized. After approximately 150 days of operation, nitrifying granules with an average diameter of 1.2 mm were formed. The granules were regular, round spheres in shape and had an apparently compact structure. The Monod kinetic coefficients of the nitrifying granules were found to be vm of 18.0 mg g-VSS-1 h-1 and Km of 36.7 mg L-1, respectively. The Fluorescence in situ hybridization (FISH) technique was used to explore the microbial communities in the nitrifying granules, and the analytical procedures can be found in the Supporting Information. Granules were sampled at the middle of an operation cycle (after 110 min aeration) for DO profile measurements. They were immediately transferred to a test chamber (Figure 2b) and a granule was carefully settled on the nylon net. Then, it was cultivated over 1 h prior to test, to achieve pseudo-steady state. The bulk solution drawn from the SBR contained NH+ 4 -N of 234 mg/L, NO2 -N of 423 mg/L, and NO-N of 27.0 mg/L. 3 The MEs fabricated were employed to measure DO distribution in the granule. Two-point calibrations to the ME were performed in the bulk solution before and after the measurements by saturating the solution with standard gas containing 0% or 21% oxygen, respectively. The changes of response were less than 5%. A micromanipulator was used to adjust the fine positions of the ME tip at a spatial resolution
of better than 5 µm, and a microscope was used for precisely locating the granular surface. The movement of the ME tip was perpendicular to the granular surface and the tip could be inserted into the granule readily. No buckling of the needle and no significant changes of the granule were observed in the measurements.
Results and Discussion Microelectrode Design and Fabrication. The ME contained a small needle 70 µm of width, 100 µm thickness, and 5 mm length, on which two working microelectrodes and a counterelectrode were integrated (Figure 3). The needle with this dimension and hardness is suitable to be inserted into biofilms or microbial aggregates. The gold tracks of the working electrodes (10 µm in width) were embedded in crosslinked SU-8, while that of the counterelectrode (15 µm in width) was not covered by the upper layer of SU-8 at the needle. The gold traces were separated from each other by 10 µm. Before the ME was used, the needle tip was severed by a sharp knife to reveal the cross-sections of the gold tracks of the working electrodes. The newly exposed Au surfaces were used as the active surfaces of working electrodes. Thus, the working electrodes could be renewed through another cutting of the needle tip. The nominal shape of the active surface on the working electrode was a rectangle (0.5 × 10 µm), but its actual shape depended on the fabrication and cutting processes. The working electrodes could be modified to improve their performance or to fabricate other microsensors in a further work. The adhesion between the two SU-8 layers was sufficiently strong, and thus their detachment was not observed in the fabrication or in the measurement and renewal. Cyclic Voltammetry Tests for the Working Electrodes. Linear sweep cyclic voltammetry analysis of the working electrodes was carried out in a solution with 0.0033 M ferro/ ferricyanide and 0.1 M KCl at a scan rate of 0.1 V/s (Figure 4). The cyclic voltammogram showed a sigmoidal transition to the diffusion limiting current, which was in good agreement with the diffusion features of the microelectrode. Fine reversible oxidation and reduction waves were observed, indicating the reliable state of the working electrodes. VOL. 41, NO. 15, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
5449
FIGURE 4. Cyclic voltammogram of a working electrode of the ME in 0.0033 M ferro/ferricyanide and 0.1 M KCl solution at a scan rate of 0.1 V/s (Three circles were recorded). FIGURE 6. Lifetimes of individual working electrodes for DO measurement. The electrodes were stored in air.
FIGURE 5. DO calibration curves of three individual working electrodes recorded in 0.1 M KCl solution. Performance of the Microelectrode as a DO Sensor. Dissolved oxygen is a crucial indicator in a variety of environmental processes. Chronoamperometry was employed to measure DO concentration in 0.1 M KCl solution. Three calibration curves of individual working electrodes of MEs are illustrated in Figure 5. High correlation coefficients were found for each case. The response of DO ranged from 36 to 169 pA per mg O2/L. This was likely to be attributed to the different electrode surface areas formed in the fabrication and cutting processes. The responses of the conventional pulled-pipet microelectrodes with tip diameters of several micrometerss are usually several pA per mg O2/L (8, 13). The MEs developed in the present work have a larger electrode area than the conventional ones, resulting in a greater response. The time for the 90% response was typically less than 15 s, which was much shorter than those of usually sized commercial oxygen electrodes. The fabricated working electrodes had a good response to DO, and they could be used for several days without renewal (Figure 6). The response waved and tended to drop as time elapsed. A decrease in the response was typically attributed to the decrease in the active electrode area. This might often be caused by the adsorption of some substances on the gold surface. Similar response changes have been reported by Linsenmeier and Yancey (13). Although the response was not a constant, the MEs could be used to monitor DO concentration precisely and reliably after calibration. At the end of the lifetime, the response almost disappeared (
