Planar 555

Biosensors and Bioelectronics 22 (2007) 1902–1907

A disposable on-chip phosphate sensor with planar cobalt microelectrodes on polymer substrate Zhiwei Zou a,∗ , Jungyoup Han a , Am Jang b , Paul L. Bishop b , Chong H. Ahn a a

Microsystems and BioMEMS Laboratory, Department of Electrical and Computer Engineering, University of Cincinnati, Cincinnati, OH 45221-0030, USA b Department of Civil and Environment Engineering, University of Cincinnati, Cincinnati, OH 45221-0030, USA Received 23 April 2006; received in revised form 2 August 2006; accepted 9 August 2006 Available online 18 September 2006

Abstract Disposable microsensors on polymer substrates consisting of fully integrated on-chip planar cobalt (Co) microelectrodes, Ag/AgCl reference electrodes, and microfluidic channels have been designed, fabricated, and characterized for phosphate concentration measurement in aqueous solution. The planar Co microelectrode shows phosphate-selective potential response over the range from 10−5 to 10−2 M in acidic medium (pH 5.0) for both inorganic (KH2 PO4 ) and organic (adenosine 5 -triphosphate (ATP) and adenosine 5 -diphosphates (ADP)) phosphate compounds. This microfabricated sensor also demonstrates significant reproducibility with a small repeated sensing deviation (i.e. relative standard deviation (R.S.D.) < 1%) on a single chip and a small chip-to-chip deviation (i.e. R.S.D. < 2.5%). Specifically, while keeping the high selectivity, sensitivity, and stability of a conventional bulk Co-wire electrode, the proposed phosphate sensor yields advantages such as ease of use, cost effectiveness, reduced analyte consumption, and ease of integrating into disposable polymer lab-on-a-chip devices. The capability to sense both inorganic and organic phosphate compounds makes this sensor applicable in diverse areas such as environmental monitoring, soil extract analysis, and clinical diagnostics. © 2006 Elsevier B.V. All rights reserved. Keywords: Phosphate sensor; Cobalt electrode; Polymer biosensor; Lab-on-a-chip

1. Introduction Aqueous phosphate ion has been the subject of continued research for over three decades (Engblom, 1998a) because of its ubiquitous significance. Determination of its concentrations in aqueous samples is important in applied analytical chemistry and clinical, horticultural, or environmental sample analysis (Engblom, 1998a; Antonisse and Reinhoudt, 1999; Moorcroft et al., 2001). For example, phosphate is the major source of eutrophication of rivers and lakes. Therefore, sensitive, cheap, and portable phosphate sensors are in high demand for monitoring the effective eutrophication process. Clinical diagnostics is another field where phosphate measurement is in demand. Hyperparathyroidism, Vitamin D deficiency, and Fanconi syndrome can be diagnosed based on specific phosphate concentrations in body fluids. Because phosphate is an essential nutrient for all plants, monitoring its concentration in soil extract is

∗

Corresponding author. Tel.: +1 513 556 0852; fax: +1 513 556 7326. E-mail address: [email protected] (Z. Zou).

0956-5663/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2006.08.004

another highly desired application for sensing phosphate level in fertilizer to serve the agricultural science (Engblom, 1998a,b). The diversity of applications and examples represents a significant need for sensitive and affordable phosphate sensors. The ion selective electrode (ISE) is a normally used method for phosphate detection and has demonstrated a lot of promise. Considerable efforts have been directed to develop ISE for monitoring phosphate concentrations sensitively and selectively. For instance, liquid-membrane electrode with different membrane materials (Carey and Riggan, 1994; Liu et al., 1997; Nishizawa et al., 2003; Ganjali et al., 2003a,b, 2006) have been explored and utilized to provide phosphate selective sensing and exhibited good sensitivity and selectivity. Although this type of phosphate sensor experiences limitations such as relatively complicated membrane structure and complicated preparation steps, it still holds great potential in some applications. Enzymebased amperimetric or potentiometric biosensing is another very commonly used method for phosphate ion detection. Pyruvate oxidase (POD) is one of the most widely used enzymes for phosphate-selective biosensors. Phosphate biosensors using POD have been realized with different sensing mechanisms

Z. Zou et al. / Biosensors and Bioelectronics 22 (2007) 1902–1907

(Nakamura et al., 1997; Mak et al., 2003; Rahman et al., 2006). Biosensors based on phosphate binding protein (Salins et al., 2004) and ion-selective-channels (Aoki et al., 2003) have also been reported for phosphate sensing. However, the comparatively high cost and instability of enzyme materials limit the use of enzyme-based phosphate sensors. Both low cost and high stability are necessary in disposable biochips for point-of-care testing (POCT) and mass environmental data collection. In addition to the use of biological components for phosphate detection, other non-biological approaches are also under investigation. Xiao et al. (1995) introduced cobalt (Co) metal as a phosphate-sensitive electrode material. They showed that the metallic Co-wire has a selective electromotive force (EMF) response to dihydrogen phosphate (H2 PO4 − ) in acidic medium. Meruva and Meyerhoff (1996) reported that the Co-wire also responded to hydrogen phosphate (HPO4 2− ) and phosphate (PO4 3− ) ions in different pH solutions. Several other groups (Chen et al., 1997; Engblom, 1998b; Parra et al., 2005) showed that this Co-wire based phosphate sensor had excellent sensitivity and low detection limit in a broad detection range. The Co-wire sensor is particularly attractive because of its ease to make, long lifetime, and produces low noise or interference from other common anions (De Marco and Phan, 2003). Most reports on Co-based phosphate sensors have used a bulk Co-wire as the working electrode and used another isolated cell as the reference electrode. More recently, miniaturized on-chip electrochemical sensors with planar microelectrodes draw great attention for their numerous benefits (Bakker, 2004). Work has been done in Ahn’s research group to develop various micro electrochemical biosensors with fully integrated on-chip working electrodes, reference electrodes, and microfluidic channels for monitoring pH, pO2 , glucose, lactate (Ahn et al., 2004), insulin (Gao, 2005), and heavy metal ions (Zhu et al., 2005). These on-chip microsensors have also been integrated as parts of the micro total analysis system (microTAS) and lab-on-achip device, which provides a platform to conduct chemical and biological analysis in a miniaturized format and is a rapidly growing field for biochemical analysis and clinical diagnostics (Manz et al., 1990; Ahn et al., 2004; Janasek et al., 2006). Polymer substrates such as cyclic olefin copolymer (COC) have been extensively utilized for lab-on-a-chips instead of traditional silicon and glass substrates due to their unique properties of biocompatibility, high optical transparency, and very low cost (Ahn et al., 2004). The main goal of this work was to develop a miniaturized phosphate sensor with on-chip planar Co microelectrode and integrated microfluidic channels (Fig. 1) using standard BioMEMS fabrication technology. The proposed sensor has been realized very cheaply and is suited for large-scale mass production and disposable usage without cross contamination. Further benefits of the proposed sensor include low volume of analyte consumption and waste generation, rapid sensing time, and elimination of the extensive polishing step used for bulk Co-wire, while maintaining comparable stability and sensitivity to traditional Co-wire electrodes. Eventually, this sensor can be used for large-scale field deployment for environment applications and disposable POCT in clinical diagnostics. Moreover,

1903

Fig. 1. Schematic view and working principle of the on-chip phosphate sensor with planar Co electrodes on polymer substrates.

it can be easily integrated into lab-on-a-chip devices, coupled with sample preparation and additional analyses. In addition, while most of previously reported Co-based phosphate sensors have only been tested for the inorganic phosphate salt (e.g. KH2 PO4 and NaH2 PO4 ), another aim of this work is to investigate the potential of the proposed sensor for organic phosphate compounds measurement. Adenosine 5 -triphosphate (ATP) and adenosine 5 -diphosphates (ADP) have been selected as analytes for organic phosphate measurement in this work. ATP and ADP are good indicators of cellular viability due to their critical roles as the energy source for many biochemical reactions. Cellular contractile phenomena are directly related to ATP and ADP concentration locally. For these reasons, a reliable technique to measure the intracellular free ATP and ADP concentration on isolated or cultured single cells addresses worthy actual physiological and pharmacological interests (Bernengo et al., 1996; Kueng et al., 2004). 2. Theoretical background Several theories have been introduced to explain the sensing mechanism of Co towards phosphate ions. Xiao et al. (1995) first proposed a host–guest mechanism in which the non-stoichiometric CoO layer provides specific cavities that can accommodate H2 PO4 − and where the specific equilibrium of H2 PO4 − within these cavities is responsible for the phosphateselective EMF response. A more broadly accepted explanation was given by Meruva and Meyerhoff (1996), in which the potentiometric response originates from a mixed potential resulting from the slow oxidation of Co and the simultaneous reduction of oxygen (Eqs. (1a), (1b) and (1)). In the presence of phosphate ions in the solution, Co3 (PO4 )2 is formed at the electrode surface (Eq. (2)). This coupled reaction shifts the equilibrium of the net electrochemical reaction (Eq. (1)), hence alters the steadystate mixed potential due to the combined anodic and cathodic components of these reactions.

1904

Z. Zou et al. / Biosensors and Bioelectronics 22 (2007) 1902–1907

Under acidic conditions,

3. Experimental

2Co + 2H2 O ⇔ 2CoO + 4H+ + 4e−

(1a)

O2 + 4H+ + 4e− ⇔ 2H2 O

(1b)

2Co + O2 ⇔ 2CoO

(1)

3CoO + 2H2 PO4 − + 2H+ ⇔ Co3 (PO4 )2 + 3H2 O

(2)

Briefly, when phosphate contacts the Co electrode, the Co2+ /Co0 redox couple will be influenced at the electrode surface due to the formation of Co3 (PO4 )2 . The electrode potential is determined by the Co2+ concentration at the electrode surface and is therefore dependent on the mass transport of phosphate ions to the electrode surface as shown in Fig. 1. The electrode potential response can be derived in terms of the Nernst equation with this assumption that the electrode potential is determined by bulk concentrations of phosphate ions, and thus a linear potential response to the logarithmic phosphate concentration can be expected (Chen et al., 1998).

3.1. Materials and apparatus KH2 PO4 (Fisher Scientific International Inc., NH, USA) was used as the reference inorganic phosphate source and was diluted to several different concentrations using buffer solution. The buffer solution was made by 25 mM potassium hydrogen phthalate (KHP, Sigma–Aldrich Corp., MO, USA) and 1 mM KCl (Fisher Scientific International Inc.) in de-ionized (DI) water at pH 5.0. Disodium adenosine 5 -triphosphate and disodium adenosine 5 -diphosphate were obtained from Sigma–Aldrich as organic phosphate sample. The buffer solution for ATP and ADP was prepared by 15 mM KHP and 1 mM KCl in DI water at pH 5.0. Co rods (99.95%) were purchased from Alfa Aesar (MA, USA) and used as metal source for the e-beam evaporator to fabricate the planar Co microelectrodes. The fabricated phosphate microsensor was electrically connected to the model 215 benchtop research-grade pH/mV meter (Denver Instrument Corp., CO, USA). The potential was measured at room temperature and the data was collected and analyzed by BalanceTalk SLTM Software (Labtronics Inc., Ontario, Canada). The sample solution was injected into the sensing chan-

Fig. 2. Fabrication processes of the on-chip phosphate sensor with polymer microfluidic channels.

Z. Zou et al. / Biosensors and Bioelectronics 22 (2007) 1902–1907

1905

Fig. 3. Photographs of the fabricated device and the microscope image of the onchip phosphate sensor composed of Co working electrodes (WE) and Ag/AgCl reference electrodes (RE).

nel through the inlet using the pump 33 dual syringes pump (Harvard Apparatus, MA, USA) and kept in the sensing chamber for test until washed out through the outlet. For each measurement, after obtained signals became stable for about 2 min, a washing step was performed using DI water, and then the next sample solution was applied. 3.2. Fabrication of phosphate sensor and microfluidic chip on polymer substrate Standard microfabrication processes were used and summarized in Fig. 2. Briefly, an Au layer of 100 nm and Co layer of 300 nm were deposited on the 3-inch blank COC wafer using the e-beam metal evaporator. Au and Co electrodes were patterned by photolithography technique and etched by Co (0.5% HNO3 ) and Au (TFA) etchant. The Ag/AgCl (∼1 ␮m) layer was deposited on the reference electrode using electroplating method on the Au seed layer (Gao, 2005). The analyte consumption and sensing time of the proposed sensor can be significantly reduced by using the integrated polymer microfluidic chip. The plastic injection molding and UV adhesive bonding technique have been developed in our group for high throughput polymer biochip fabrication. The fabrication detail has been reported previously (Choi et al., 2001) and summarized in Fig. 2. After drilling holes for fluidic interconnection at inlet and outlet, the microfluidic chip was bonded with the sensor chip using UV adhesive bonding technique at room temperature (Han et al., 2003) to achieve the final device. Photographs of the fabricated device have been shown in Fig. 3, which illustrate the microelectrode array, electrical connections, and microchannels. The entire chip size is 1.5 cm × 2 cm, and the inlet and outlet channels have width of 200 ␮m and depth of 100 ␮m. The reaction chamber has width of 2 mm, length of 10 mm, depth of 100 ␮m, and volume of 2 ␮l. The detail of the Co working electrode and the Ag/AgCl reference electrode has been clearly shown in Fig. 3 inset. Both electrodes have length of 1.5 mm, width of 200 ␮m, and a spacing of 200 ␮m.

Fig. 4. Potentiometric response of the phosphate sensor in different concentrations of KH2 PO4 at pH 5.0: (a) dynamic measurement and (b) calibration curve.

4. Results and discussion Bulk Co-wire based phosphate sensors have been characterized for inorganic phosphate (Meruva and Meyerhoff, 1996; Chen et al., 1997; Engblom, 1998b; Parra et al., 2005). These Co-wire electrodes show a very good response to inorganic phosphate (KH2 PO4 and NaH2 PO4 ) in a very wide dynamic range from 5 × 10−5 to 5 × 10−2 M with a detection limit less than 10−5 M. The Co-wire electrode has high selectivity for phosphate ions with respect to many other common anions (Chen et al., 1997). In this research, an on-chip Co microelectrode phosphate sensor was evaluated for its performance in comparison to traditional bulk Co-wire based phosphate sensors. Measurements were performed using this sensor as the concentration of KH2 PO4 was dynamically varied from 10−5 to 10−2 M. A stepwise response to different KH2 PO4 concentrations and the steady-state potential for each sample concentration has be observed and shown in Fig. 4a. It is also evident that due to the miniaturized sensing size and reaction volume, the on-chip sensor reaches an equilibrium response rapidly, in approximately 1 min for 10−5 M and in less than 30 s for higher concentrations above 10−5 M. Calibration curve derived from Fig. 4a is shown in Fig. 4b as well. The dynamic range is from

1906

Z. Zou et al. / Biosensors and Bioelectronics 22 (2007) 1902–1907

Fig. 5. Potentiometric response of the phosphate sensor in different concentrations of ATP and ADP at pH 5.0.

10−5 to 10−2 M by using the proposed sensor and is the same as most bulk Co-wire phosphate sensors (Xiao et al., 1995; Chen et al., 1998; De Marco and Phan, 2003). A higher base line potential can be observed in this sensing system compared to bulk Co-wire sensors but can likely be due to the different Cl− concentrations used for the Ag/AgCl reference electrode. The sensor response to the organic phosphate was performed using standard ATP and ADP samples. The KHP concentration in buffer solution was adjusted to 15 mM according to the optimized value for ATP and ADP (Xiao et al., 1995). As shown in Fig. 5, the sensor exhibits a potentiometric response to ATP and ADP in the range between 10−5 and 10−2 M. It is also noticed that ADP displays a larger potentiometric slope than ATP does. A similar phenomenon was observed and explained by Xiao et al. (1995). This is in agreement with the fact that the number of “additional” units of phosphate binding to the electrode is two for ATP and only one for ADP as illustrated in Fig. 5. The on-chip sensor presents a steady-state response for more than 30 min in 10−5 M KH2 PO4 solution (Fig. 6), which is sufficient for disposable sensor applications. In addition, this

Fig. 6. Long-term potentiometric response of the phosphate sensor in 10−5 M KH2 PO4 at pH 5.0.

Fig. 7. Reproducibility of the fabricated sensor: (a) potential responses to 10time repeated injections of 10−3 M ADP to the same phosphate sensor and (b) chip-to-chip deviation of four different phosphate sensors in measuring 10−3 M KH2 PO4 and 10−3 M ATP.

Z. Zou et al. / Biosensors and Bioelectronics 22 (2007) 1902–1907

sensor has high reproducibility which is another vital requirement for mass-produced microsensors. For example, injections of 10−3 M ADP into the same phosphate sensor for 10 times reveal good measurement reproducibility (i.e. 526 ± 4 mV or relative standard deviation (R.S.D.) of 0.6%) as shown in Fig. 7a. This result is comparable with reported data using bulk Cowire (3.0% R.S.D., Chen et al., 1997; 3.8% R.S.D., Chen et al., 1998; 2–4% R.S.D., De Marco and Phan, 2003). Reasonably low chip-to-chip deviation has been obtained by measuring KH2 PO4 and ATP at 10−3 M on four different sensors with variances of 2.5% R.S.D. for KH2 PO4 and 2.1% R.S.D. for ATP (Fig. 7b). The proposed on-chip sensor also exhibited high selectivity for H2 PO4 − (e.g. Ki,j (Cl− ) = 4.1 × 10−3 , Ki,j (NO3 − ) = 8 × 10−4 , Ki,j (SO4 2− ) = 8.2 × 10−4 , Ki,j (I− ) = 1.1 × 10−2 ) which is comparable to bulk Co-wire based phosphate sensors. 5. Conclusions The new on-chip phosphate sensor using planar Co microelectrodes has been developed and fully characterized in this work. The feasibility of this electrochemical sensor to monitor both inorganic and organic phosphate compounds has been fully demonstrated. By incorporating the mass-produced microfabrication technique and high throughput plastic micromachining, the proposed on-chip phosphate sensor with the integrated microfluidic chip can be batch fabricated with very low cost and high yield compared to the conventional bulk Co-wire based sensor, while maintaining the excellent performance. The miniaturized sensing system is especially suitable for large-scale field deployment for mass environmental data collections and disposable POCT in clinical diagnostics. Moreover, the proposed on-chip microsensor is fully integrated with polymer microfluidic system and can be easy developed as multi-analyte polymer lab-on-a-chips for a wide range of applications. Acknowledgements The authors gratefully thank Mr. Ron Flenniken in the Institute for Nanoscale Science and Technology at the University of Cincinnati, for his technical support, and also thank Mr. Andrew Browne for discussion. References Ahn, C.H., Choi, J.W., Beaucage, G., Nevin, J.H., Lee, J.B., Puntambekar, A., Lee, J.Y., 2004. Proc. IEEE 92 (1), 154–173.

1907

Antonisse, M.M.G., Reinhoudt, D.N., 1999. Electroanalysis 11 (14), 1035– 1048. Aoki, H., Hasegawa, K., Tohda, K., Umezawa, Y., 2003. Biosens. Bioelectron. 18 (2), 261–267. Bakker, E., 2004. Anal. Chem. 76 (12), 3285–3298. Bernengo, J.C., Brau, F., Steghens, J.P., 1996. Proceedings of 18th International Annual Conference of the IEEE Engineering in Medicine and Biology Society, Amsterdam, the Netherlands, October 31–November 3, pp. 1907– 1908. Carey, C.M., Riggan Jr., W.B., 1994. Anal. Chem. 66 (21), 3587–3591. Chen, Z., De Marco, R., Alexander, P.W., 1997. Anal. Commun. 34 (3), 93–95. Chen, Z., Grierson, P., Adams, M.A., 1998. Anal. Chim. Acta 363 (2), 191– 197. Choi, J.W., Kim, S., Trichur, R., Cho, H.J., Puntambekar, A., Cole, R.L., Simkins, J.R., Murugesan, S., Kim, K.S., Lee, J.B., Beaucage, G., Nevin, J.H., Ahn, C.H., 2001. Proceedings of Fifth International Conference on Micro Total Analysis Systems (micro-TAS 2001), Monterey, CA, USA, October 21–25, pp. 411–412. De Marco, R., Phan, C., 2003. Talanta 60 (6), 1215–1221. Engblom, S.O., 1998a. Biosens. Bioelectron. 13 (9), 981–994. Engblom, S.O., 1998b. Plant Soil 206 (2), 173–179. Ganjali, M.R., Mizani, F., Niasari, M.S., 2003a. Anal. Chim. Acta 481 (1), 85–90. Ganjali, M.R., Mizani, F., Emami, M., Niasari, M.S., Shamsipur, M., Yousefi, M., Javanbakhtd, M., 2003b. Electroanalysis 15 (2), 139–144. Ganjali, M.R., Norouzi, P., Ghomi, M., Niasari, M.S., 2006. Anal. Chim. Acta 567 (2), 196–201. Gao, C., 2005. Ph.D. Dissertation. University of Cincinnati. Han, J., Lee, S.H., Puntambekar, A., Murugesan, S., Choi, J.W., Beaucage, G., Ahn, C.H., 2003. Proceedings of Seventh International Conference on Micro Total Analysis Systems (micro-TAS 2003), Squaw Valley, CA, USA, October 5–9, pp. 1113–1116. Janasek, D., Franzke, J., Manz, A., 2006. Nature 442 (7101), 374–380. Kueng, A., Kranz, C., Mizaikoff, B., 2004. Biosens. Bioelectron. 19 (10), 1301–1307. Liu, D., Chen, W.C., Yang, R.H., Shen, G.L., Yu, R.Q., 1997. Anal. Chim. Acta 338 (3), 209–214. Mak, W.C., Chan, C., Barford, J., Renneberg, R., 2003. Bioelectronics 19 (3), 233–237. Manz, A., Graber, N., Widmer, H.M., 1990. Sens. Actuators B 1 (5), 244–248. Meruva, R.K., Meyerhoff, M.E., 1996. Anal. Chem. 68 (13), 2022–2026. Moorcroft, M.J., Davis, J., Compton, R.G., 2001. Talanta 54 (5), 785–803. Nakamura, H., Ikebukuro, K., McNiven, S., Karube, I., Yamamoto, H., Hayashi, K., Suzuki, M., Kubo, I., 1997. Biosens. Bioelectron. 12 (9–10), 959–966. Nishizawa, S., Yokobori, T., Kato, R., Yoshimoto, K., Kamaishi, T., Teramae, N., 2003. Analyst 128 (6), 663–669. Parra, A., Ramon, M., Alonso, J.N., Lemos, S.G., Vieira, E.C., Nogueira, A.R.A., 2005. J. Agric. Food Chem. 53 (20), 7644–7648. Rahman, M.A., Park, D.S., Chang, S.C., McNeil, C.J., Shim, Y.B., 2006. Biosens. Bioelectron. 21 (7), 1116–1124. Salins, L.L.E., Deo, S.K., Daunert, S., 2004. Sens. Actuators B 97 (1), 81–89. Xiao, D., Yuan, H.Y., Li, J., Yu, R.Q., 1995. Anal. Chem. 67 (2), 288–291. Zhu, X., Gao, C., Choi, J.W., Bishop, P.L., Ahn, C.H., 2005. Lab Chip 5 (2), 212–217.