A micromachined electrochemical sensor for free chlorine monitoring in drinking water

Transcription

1 A micromachined electrochemical sensor for free chlorine monitoring in drinking water A. Mehta*, H. Shekhar**, S.H. Hyun***, S. Hong**** and H.J. Cho* *Mechanical, Materials and Aerospace Engineering, University of Central Florida, Orlando, FL , USA ( **Electrical Engineering, University of Central Florida, Orlando, FL , USA ***Civil and Environmental Engineering, University of Central Florida, Orlando, FL , USA ****Civil and Environmental Engineering Department, Korea University, 1, 5-ka, Anam-dong, Sungbuk-ku, Seoul, , Korea Abstract In this work, we designed, fabricated and tested a disposable, flow-through amperometric sensor for free chlorine determination in water. The sensor is based on the principle of an electrochemical cell. The substrate, as well as the top microfluidic layer, is made up of a polymer material. The advantages include; (a) disposability from low cost; (b) stable operation range from three-electrode design; (c) fluidic interconnections that provide on line testing capabilities; and (d) transparent substrate which provides for future integration of on-chip optics. The sensor showed a good response and linearity in the chlorine concentration ranging from 0.3 to 1.6 ppm, which applies to common chlorination process for drinking water purification. Keywords Chlorination; drinking water; electrochemical sensor; microsensor Introduction Required potable water quality has changed dramatically from the late 1970s due to enhanced regulation of pathogens and disinfection by-products: for example, Surface Water Treatment Rule and Disinfection By-Product Rule (Taylor and Hong, 2000). Consequently, new treatment technology and reliable analytical techniques are required to meet this challenge. In the United States, waterborne disease has been controlled primarily by chlorination, which is an essential part of typical drinking water treatment processes to prevent bacterial proliferation in the water distribution systems. The use of chlorine for disinfection of pathogens, however, requires caution since it also reacts with organics and forms disinfection by-products (DBPs). Regulated carcinogenic DBPs formed during chlorination include trihalomethanes and haloacetic acids. Excess chlorination also results in the corrosion of water distribution system and adds odour and taste to drinking water. To ensure public health and safety, it is extremely important to accurately and effectively monitor chlorine residuals during the treatment and transport of drinking water. HACH DR/4000 Spectrophotometer and DPD (N, N-diethyl-p-phenylenediamine) ferrous titrimetric methods are the standard methods used for measuring free chlorine (APHA, 1995; Hach, 1995). These methods are not portable and require expensive instrumentation. Several alternative methods have been suggested. One approach is based on optical detection using differential absorption spectroscopy (Belz et al., 1997). This method requires UV or laser light sources and complicated instrumentation. Electrochemical sensing techniques, and more specifically amperometry, have been employed for their simple sensor structure and good sensitivity (Matsumoto et al., 2002). Although different types of disposable electrochemical microsensors have been studied (Yang et al., 1996, 1997), most of doi: /wst Water Science & Technology Vol 53 No 4 5 pp Q IWA Publishing 2006

2 A. Mehta et al. them are paper-based, dip-type sensors. They have limitations in lifetime and on-line measurement. A miniaturized amperometric flow-through sensor for residual chlorine measurement has been studied (Okumura et al., 2001), but the sensor cell consists of protruded electrodes on silicon and a glass cover, which requires relatively complicated fabrication process. The microsensor presented here has been developed based on a transparent polymer substrate (Trichur et al., 2002) and is inexpensive and offers advantages such as future integration of on-chip optics. For improved linearity over a wide range of chlorine concentrations, three electrode configurations have been selected (Zhu et al., 2002). Microfluidic packaging allows easy approach to external connections with the sensor. Methods Sensor design The sensor is comprised of an electrode layer which is used to conduct amperometric measurements corresponding to the changing analyte concentration and a fluidic layer on top which allows for sample transport over the sensor electrodes. The two layers bonded together provide a packaged device for flow-through measurement of free chlorine concentration. In addition to the electrochemical sensor, the electrode layer also contains a temperature sensor for sample temperature monitoring. Electrode layer. A spherical design was used to design a 3-electrode amperometric sensor. The working (WE) and counter (CE) electrodes were fabricated using Au, while Ag/AgCl was also deposited on top of the Au for the reference electrode (RE). The principle of amperometry is based on the measurement of the current between the working and counter electrode, which is induced by a redox reaction at the working electrode. The conditions are chosen in such a way that the current is directly proportional to the concentration of a redox active species in the analyte solution. In the case of free chlorine measurement, the current corresponds to the movement of OCl - ions. Figure 1a, shows the structure and dimensions of the amperometric sensor. In addition to the sensing electrodes, a meander type Au temperature sensor was designed. It is a resistance temperature sensor (RTS) which is based on the change of electrical resistivity with temperature in metals. The active area shown in Figure 1b is the region that is in direct contact with the substance (gas or liquid) of which the temperature is to be determined. The area of the region determines the correct operating range. For our application, with a known gold film thickness and temperature range, an area of a b c d e Active region Area 1 mm 2 (a) (b) 404 Figure 1 (a) Free chlorine sensor dimensions (mm): a ¼ 400, b ¼ 100, c ¼ 350, d ¼ 400 and e ¼ 250. a, c and e correspond to the dimensions of RE, WE and CE respectively. (b) Resistive meander type temperature sensor

3 approximately 1 mm 2 is ideal. This area represents the area of the Au in the active region. The inclusion of a temperature sensor adds to the functionality of design by allowing compensation for temperature variation during measurements. Fluidic layer. The fluidic layer provides standard 1/16 00 and 1/32 00 tubing connections for sample flow over the bottom electrode layer. Figure 2 shows the sample injection mechanism into the electrode layer using the top fluidic layer. Sensor fabrication Au was deposited on a 3 00 COC (cyclic olefin copolymer) wafer using electron beam evaporation. Specific sites corresponding to the reference electrode were opened using photolithography for Ag deposition via electroplating. To complete the Ag/AgCl fabrication process, chlorination was carried out through electroplating with a 0.1 M KCl solution as the analyte. The Au was selectively patterned by a photolithographic step which defined the electrochemical sensor, temperature sensor and contact pads. Finally, the remaining Au was etched away using the standard potassium iodide-iodine gold etch solution. Figure 3 illustrates the fabrication steps. The substrate was then diced to produce individual devices. Figure 4 shows one of the assembled and packaged devices with the fluidic layer on top. A. Mehta et al. Sample preparation Free residual chlorine sample solutions were prepared by diluting standard chlorine solutions (Hach Company, USA). The 63.5 ppm standard solution was diluted using the deionized (DI) water to obtain samples in the ppm range. This was done to simulate the free chlorine concentration found in drinking tap water. Free chlorine is the sum of hypochlorous acid (HOCl) and hypochlorite ions (OCl 2 ). The ratio HOCl to OCl 2 is ph and temperature dependent. As the ph increases, the fraction of hypochlorous acid decreases and the fraction of hypochlorite ions increases (Rosemount Analytical, 2002). Since the sensor responds to OCl 2 ions alone, the operating ph was kept at 9 by adding appropriate amount of 0.1N NaOH. Free chlorine concentrations for each sample were measured using HACH DR/4000 spectrophotometer (Hach Company, U.S.A). To ensure the accuracy of the Hach measurements, DPD (N, N-diethyl-p-phenylenediamine) ferrous titrimetric method was also performed. Figure 2 (a) Fluidic and electrode layers (b) Cross sectional schematic showing a flow-through sample test 405

4 A. Mehta et al. Figure 3 Fabrication steps: (a) Au deposition, (b) photolithography and Ag electroplating, (c) chlorination and photoresist stripping, (d) Au etching Figure 4 Packaged sensor with top fluidic layer and electrical contacts 406 Test setup For sample injection, the injection pump (NE 1000, New Era Pump Systems) and valves were preset. The flow rate was fixed under 5 ml/min. Cyclic polarization was performed between 21V and 1V with respect to the Ag/AgCl electrode to determine the reduction potential for free chlorine. The measured potential was obtained with respect to the Ag/AgCl reference potential. Then, the potentiostatic measurements were performed using the reduction potential for free chlorine. These experiments were conducted on samples in the ppm free chlorine range. The sensor was first run with a calibration cycle, then the samples. The signal from the sensor was obtained with a potentiostat (Multistat II, Sycopel Scientific) from which the data was transferred to a PC and displayed using the multistat II software package. Figure 5 shows the experimental setup.

5 Calibration Sample Pump Valve Fluidic path Waste reservoir Disposable sensor Potentiostat Figure 5 Schematic view of the setup for sensor testing PC Electrical path A. Mehta et al. Results and discussions Amperometric sensor The current flow observed during the potentiostatic measurements was in direct proportion with the free chlorine concentration (OCl 2 ions) exclusively. The working potential of 100 mv was determined using cyclic polarization. For the initial sample measurements, a stability period of 100 s was needed before the sensor signal attained a constant value. Figure 6 shows the potentiostatic curves plotted for chlorine concentration variation in the range ppm. Current outputs from all the experiments were tabulated and plotted against the sample concentration as shown in Figure 7. Using the calibration curve, the sensor was also tested to measure free chlorine in a tap water sample. It was measured to be in the range of 0.7 ppm. The sensor shows linear operation in the test range of ppm chlorine solutions. The correlation factor, R 2, is found to be This linearity in operation can be used to correlate the sensor output to the chemically Figure 6 Time variation of sensor output measured in a solution with 1.6 ppm chlorine Figure 7 Sensor output as a function of free chlorine concentration where the correlation factor is R 2 ¼

6 determined values of free chlorine in water. Since the operational range of the developed microsensor is close to the free chlorine concentrations allowed in drinking water treatment, this sensor has applications in the free chlorine monitoring in drinking water. A. Mehta et al. Temperature sensor In view of the general temperature range under which the free chlorine measurement experiments are conducted, a temperature range of 5 258C was considered for testing. The change in resistance with increase in temperature of the sample was plotted and is shown in Figure 8. A linear response was observed. Once the resistance is calibrated, the temperature of the sample could be directly read. Flow-through measurement Figure 9a shows the results of time variation of the sensor output. The flow-through output required approximately 100 s for sensor stabilization. The decrease in free chlorine concentration was observed at a flow-through rate of 5 ml/min in the initial stage, which would be contributed by the decomposition of free chlorine in contract with polymeric tubing. After this initial soaking period, the measured values became stable. The temperature was also measured simultaneously to account for any change in concentration with Figure 8 Variation of resistance of the temperature sensor with respect to temperature where the correlation factor is R 2 ¼ Figure 9 (a) Time variation of concentration during flow-through measurement of sample with 8 ppm free chlorine. (b) Variation of temperature inside the sensor cell during the flow-through measurement

7 % Current deviation Figure 10 Interference effects on chlorine sensor measurements with [Cl] ¼ 1.6 ppm respect to temperature. Figure 9b shows the operating temperature variation measured during the sensor operation. This demonstrates the stability of the sensor at this flow rate. A. Mehta et al. Selectivity Selectivity testing is particularly important to ensure the absence of false alarms in free chlorine testing. Potassium iodide (KI) can produce fairly reactive ionic species that may interfere with chlorine detection. To determine the selectivity of free chlorine sensor, the interfering species, KI was added in a fixed 1.6 ppm chlorine solution. Since the chlorination in tap water can vary in a wide range from 0.4 to 2.0 ppm, this concentration is applicable. The interference testing was done by successive addition of increasing amounts of KI solution from the ppm range; this higher chlorine concentration (1.6 ppm) was intentionally chosen over the other samples to ensure an adequate supply of free chlorine in the repeated testing and to compensate for any chlorine absorption or loss. Concentration of KI was varied in the range of ppm and the deviation in current response with the 1.6 ppm chlorine solution was recorded. As shown in Figure 10, there was no significant deviation in the signal response despite the presence of almost equivalent amounts of interfering ionic species. Conclusions We designed, fabricated and tested a disposable microsensor for free chlorine monitoring in water. Amperometric measurements yielded a linear sensor response in the tested range of ppm of free chlorine. This operation range is within the allowed chlorine levels in drinking water. The sensor was also tested with a tap water sample and its chlorine content determined. The sensor showed a stable output response and good selectivity in the presence of interfering ionic species. A temperature sensor was also incorporated in the design to compensate for possible temperature variation effects on the measurements. The developed microsensor is a low cost device that is applicable to a portable on-site instrument for monitoring free chlorine in the water treatment process. Acknowledgements This work was partially supported by a National Science Foundation CAREER Award (ECS ) and by Korea Research Foundation Grant (KRF-2004-D00165). The authors would like to thank Dr. C. H. Ahn at the University of Cincinnati for technical discussions and assistance. References APHA-AWWA-WEF (1995). Standard Methods For The Examination Of Water And Wastewater, 19th Ed., Washington DC. Belz, M., Boyle, W.J.O., Klein, K-F. and Grattan, K.T.V. (1997). Smart-sensor approach for a fiber-opticbased residual chlorine monitor. Sensors Actuators B, 38 39, Hach Company (1995). HACH DR/4000 Spectrometer procedures manual. 409

8 A. Mehta et al. Matsumoto, T., Ohashi, A. and Ito, N. (2002). Development of micro-planar Ag/AgCl quasi-reference electrode with long term stability for an amperometric glucose sensor. Anal. Chim. Acta., 462, Okumura, A., Hirabayashi, A., Sasaki, Y. and Miyake, R. (2001). Simple miniaturized amperometric flow cell for monitoring residual chlorine in tap water. Anal. Sci., 17, Rosemount Analytical (2002). Water and wastewater industry application data sheet, PN Taylor, J. and Hong, S. (2000). Potable water quality and membrane technology. J. Lab. Med., 31, Trichur, R., Kim, S., Lee, S.H., Abdelaziez, Y.A., Starkey, D.E., Halsall, H.B., Heineman, W.R. and Ahn, C.H. (2002). A new plastic CE chip with wide optical clarity using cyclic olefin copolymers (COC). Proceedings of the 6th International Conference on Micro Total Analysis Systems, Yang, Z., Suzuki, H., Sasaki, S. and Karube, I. (1996). Disposable sensor for biochemical oxygen demand. Appl. Microbiol. Biotechnol., 46, Yang, Z., Suzuki, H., Sasaki, S. and Karube, I. (1997). Design and validation of a low-cost paper-based oxygen electrode. Anal. Lett., 30, Zhu, J., Zhu, Z., Lai, Z., Wang, R., Guo, X., Wu, X., Zhang, G., Zhang, Z., Wang, Y. and Chen, Z. (2002). Planar amperometric glucose sensor based on glucose oxidase immobilized by chitosan film on prussian blue layer. Sensors, 2,