Sensors and Actuators B 70 Ž2000. 83–86 www.elsevier.nlrlocatersensorb
Metal silicides as a novel electrode material in electrochemical sensors T.G.I. Ling ) , L. Montelius Lund UniÕersity, DiÕision of Solid State Physics and Nanometer Structure Consortium, Department of Physics, P.O. Box 118, S-221 00 Lund, Sweden Received 23 November 1999; accepted 13 April 2000
Abstract We have studied the suitability of titanium silicide as an electrode material in electrochemical cells. The high chemical stability, low resistivity and possibility to immobilise biomolecules in combination with the possibility to manufacture structured electrodes with methods from the microelectronics industry make silicides interesting. As an example results from measurements with the enzyme horseradish peroxidase are given and compared with similar measurements with graphite and carbon paste electrodes. q 2000 Elsevier Science S.A. All rights reserved. Keywords: Electrodes; Electrochemistry; Metal silicides; Graphite; Carbon paste
1. Introduction In the effort to miniaturise sensors a considerable interest has been focused on the methods developed in the microelectronics industry. This trend has primarily been driven by the need to reduce the cost of an analysis by reducing the amount of chemicals used. The miniaturisation may eventually lead to new ways to study molecules in sensor applications. It is now possible to structure semiconductor surfaces down to the biomolecular size range of a few nanometers w1,2x. Electrochemical measurements of biomolecules immobilised onto such structures will use an electric field that spans the volume of a nanoliter. Short distances between electrodes also facilitate the study of electron transfer mechanisms. One of the key problems has been to make reliable microelectrodes. Gold electrodes can easily be manufactured but are also known to give problems. In order to make gold stick to a surface of silicon dioxide a separate adhesion layer is required. In semiconductor processing chromium or titanium is commonly used for this purpose. As a consequence, a galvanic cell is formed as soon as the two metals come in contact with an electrolyte. The leakage of the less noble metal can be seen as additional peaks
) Corresponding author. Tel.: q46-46-222-4495; fax: q46-46-2223637. E-mail address:
[email protected] ŽT.G.I. Ling..
in cyclic voltammograms w3x. Gold also gives problems with unspecific adsorption of various compounds which in the macroscopic case can be overcome by polishing of the electrode surface before every experiment. Polishing of a microelectrode is, however, not feasible. Metal silicides have in recent years been used as connectors in integrated circuits, mainly because of their thermal stability that makes them compatible with processing technology together with their low bulk resistivity and good ohmic contacts, i.e., with low resistivity and capacitance values w4,5x. They have also an excellent stability towards different chemicals which should make them useful for electrochemical applications. Such usage seems however to have been overlooked so far. In this paper we have used titanium silicide with a resistivity of ; 100 mV cm, which should be compared with graphite Ž1.4 m V cm. another common choice of electrode material for bioelectrochemical sensors. Another advantage of silicon and silicides is that it is possible to derivatise the surface by using functional silanes. This chemistry has been previously applied in modification of silica gels used in liquid chromatography columns. This paper aims at showing the potential possibilities of using silicides by making a comparison with well known electrode materials. Three different materials were evaluated as enzyme electrodes by immobilising horseradish peroxidase on titanium silicide by adsorption on graphite and by inclusion in carbon paste. The performance was analysed by the currentrvoltage characteristics.
0925-4005r00r$ - see front matter q 2000 Elsevier Science S.A. All rights reserved. PII: S 0 9 2 5 - 4 0 0 5 Ž 0 0 . 0 0 5 6 3 - 3
84
T.G.I. Ling, L. Monteliusr Sensors and Actuators B 70 (2000) 83–86
2. Experimental Water processed in a Millipore MilliQ Plus with a resistivity of 18 M V P cm was used in all experiments. Horseradish peroxidase ŽHRP. was of type VI-A, and Tris buffer was made of Trizmaw base, both were purchased from Sigma Chemical. All other chemicals were of analytical grade. 2.1. Carbon paste electrodes The HRP modified Carbon Paste was obtained after vacuum drying a mixture of 100 mg of graphite powder ŽFluka Chemie, Switzerland. and 2 mg HRP dissolved in 10 mM sodium phosphate buffer at pH s 7.0. Water was removed by keeping the mixture under vacuum Ž; 10 mbar. for 4 h at 48C. The paste was then packed into a plastic syringe and electrical contact was established by insertion of a silver wire into the paste. This electrode preparation was performed at the Department of Analytical Chemistry, Lund University. 2.2. Graphite electrodes Rods of solid spectroscopic graphite ŽRingsdorff Werke, Germany, type RW001. having a diameter of 3.05 mm were cut, and then polished on white fine emery paper. After careful washing with water, the surfaces were again polished, this time with ordinary white paper to prepare a flat-mirror like surface. Then a small amount of HRP dissolved in 10 mM Tris–HCl buffer, pH s 7.1, was placed on top of the mirror surface allowing adsorption of HRP. After 10 min the electrodes were rinsed with Millipore water and then stored in Tris buffer until use. 2.3. Titanium silicide electrodes Titanium silicide was formed by sputtering an approx. 70 nm thick titanium layer onto a 6-in. silicon wafer and subsequent heating to 5958C in an atmosphere of N2 and H 2 . The surface nitride was removed by wet etching ŽH 2 O 2 :NH 4 OH:H 2 O 1:1:5. at 258C. The resulting surface resistivity was determined with a four-point probe and found to be 6.2 Vrcm. The resistivity and processing conditions are in agreement with the base centre orthorhombic structure ŽC49. of TiSi 2 w6x. The wafer was cleaved into smaller dices and carefully washed with an organic cleaning cycle consisting of trichloroethane, acetone, 2-propanol and de-ionised water. After cleaning, the dices were immediately placed into a modified vacuum drying oven in a nitrogen atmosphere. Then the silicide surface was chemically functionalised by a reaction with 3-aminopropyl-triethoxysilane, APTS, at 2008C and 10 mbar for 2 h w7x. AFM pictures showed a flat surface and the thickness of the layer was estimated to be in close agreement with the thickness of a monolayer. HRP was
immobilised using a microscale adaptation of the metaperiodate coupling method w8x. The cleaned APTS terminated silicide surface was then exposed to an excess of HRP. Here, the amino group in the APTS reacts with the sugar group of HRP through a Schiff–Base reaction. After this binding event the dices were carefully rinsed in order to wash away non-specific bound HRP. Then wires were bonded onto the top surface and carefully protected with silicone rubber. The size of the electrodes was 3 = 3 mm2 . 2.4. Measurement set-up The measurements were in all cases performed in a liquid cell and most often in a two-terminal scheme having a gold-electrode as a counter electrode. The current to voltage characteristics were obtained by utilising a computer controlled electrometer ŽKeithley Electrometer 617. by a Lab-View implemented virtual instrument. For voltage measurements a multimeter ŽHewlett-Packard 34401A. were employed. Both instruments were controlled through the IEEE-interface by the Lab-View program. For all measurements the enzymatic reaction of HRP with peroxide was monitored by measuring the current vs. voltage between the working and counter electrodes as a function of peroxide concentration. The full reaction mix consisted of hydrogen peroxide together with 0.8 mM 4-amino-antipyrine and 14 mM phenol dissolved in 10 mM Tris–HCl buffer pH 7.1. The peroxide concentration was varied between 0 and 500 mM.
3. Results and discussion The enzyme horseradish peroxidase was immobilised on electrodes of titanium silicide and the results were compared with similar experiments performed on electrodes of graphite and carbon paste. As far as possible the same conditions were used in these experiments. The difference was the way the enzyme was presented on the surface and the chemistry used for this purpose. On the
Fig. 1. Currentrvoltage characteristics of a graphite electrode with adsorbed HRP. Varying concentrations in mM of hydrogen peroxide in a mixture of 4-amino antipyrine and phenol.
T.G.I. Ling, L. Monteliusr Sensors and Actuators B 70 (2000) 83–86
85
Fig. 4. Current–concentration relation for hydrogen peroxide at an applied voltage of 400 mV between the carbon paste graphite working electrode and the gold counter electrode. Data taken from Fig. 3.
Fig. 2. Current–concentration relation for hydrogen peroxide at an applied voltage of 400 mV between the graphite working electrode and the gold counter electrode. Data taken from Fig 1.
graphite electrodes the enzyme was adsorbed in mild buffer, while the carbon paste also included vacuum drying. In case of the silicide electrode the peroxidase was immobilised to the amino groups on the surface by oxidation of the carbohydrate shell using the periodate method. By performing the silylation in gas phase only a monolayer of APTS is produced which minimises the surface resistivity between the electrode surface and the enzyme. The electrochemical measurements reported in this paper are all made without a reference electrode and a potentiostat which would be present in a standard set-up and compensate for any voltage drop including variations in surface impedance of the tested electrode. Omitting the reference electrode is not recommended for analytical purposes since processes on the electrode surface would affect the measurement. However, when the electrode itself is the subject of study, the voltage compensation performed by a potentiostat may diminish the differences between the electrodes and jeopardise the possibility to make a fair comparison. This means that care should be taken when comparing with graphs showing voltage vs. a reference electrode. In the measurements of current vs. voltage characteristics as a function of peroxide concentration, the different
Fig. 3. Currentrvoltage characteristics of a carbon paste electrode with included HRP. Varying concentrations of hydrogen peroxide ŽmM. in a mixture of 4-amino antipyrine and phenol.
electrodes displayed somewhat different behaviour. For the graphite electrode ŽFig. 1. the dependence on applied voltage indicates that only a limited potential window around 400 mV is suited for qualitative determinations. Outside this window either the high or low concentrations give a low resolution and consequently limits the working concentration range for these applied potential regions. By choosing a voltage of 400 mV, the relationship between concentration and the current could be established ŽFig. 2. displaying a quasi-linear behaviour. The carbon paste electrode ŽFig. 3. has a similar behaviour but a somewhat lower current sensitivity. At a voltage of 650 mV the different curves merge together and even cross each other. An offset in voltage can thus seriously affect a measurement. The potential window is wider than for graphite electrodes, but it was not possible to find a point matching the performance of the graphite electrode and the concentration dependence deviates considerably from linearity ŽFig. 4.. The titanium silicide electrode shows a different type of current to voltage characteristics ŽFig. 5.. At a voltage range around zero we have a very low current and outside this range the current increases almost exponentially. This difference is probably due to the surface layer on the electrode. The aminopropyl groups produce a thin layer of approx. 1 nm including counter ions that should be expected to give a significant resistance, meaning that there will be a resistive voltage drop over this layer. Also, we cannot rule out the possibility that we have a thin oxide layer on the surface. By choosing a voltage of 800 mV and by making a current vs. concentration graph in a similar
Fig. 5. Currentrvoltage characteristics of a titanium silicide electrode with immobilised HRP. Varying concentrations of hydrogen peroxide ŽmM. in a mixture of 4-amino antipyrine and phenol.
86
T.G.I. Ling, L. Monteliusr Sensors and Actuators B 70 (2000) 83–86
ULSI ŽUltra Large Scale Integration. applications in the microelectronics industry. Thus, the suitability for usage as material in enzyme electrodes together with the possibility to manufacture electrode patterns using established chip technology makes this material an extremely interesting candidate as electrode material for future bioelectronical applications, e.g., structured micro- and nano-biosensors w10x.
Acknowledgements Fig. 6. Current–concentration relation for hydrogen peroxide at an applied voltage of 800 mV between the titanium silicide working electrode and the gold counter electrode. Data taken from Fig. 5.
way as for the other electrodes ŽFig. 6., we can deduce that we have an increased sensitivity for this type of electrode. This is especially seen for the low concentrations where the slope of the curve is quite high. The overall shape of the current vs. voltage relationship in Fig. 5 is different from the previous by the monotonically increasing function indicating different electron transfer properties. Another consequence is that the selection of a voltage for amperometric determination becomes less critical. The enzyme is in this case attached to the surface in a different way with the help of a short spacer arm and will form a thin film of approximately one monolayer in thickness. It is thus probable that the surface is insulating as compared to the carbon based electrodes where the solution has access to a significant portion of the surface of the electrode. Titanium silicide warrants more extended studies and is now the subject of further electrical investigation in the authors lab.
4. Conclusion A new electrode material-titanium silicide-has been tested. An enzyme electrode consisting of horseradish peroxidase immobilised on titanium silicide gave a voltammetric response to hydrogen peroxide. The resistivity of the silicide used here lies between graphite and gold. The electrical properties in combination with the possibility to use well known covalent immobilisation methods makes the material promising for further studies. The electron transfer may be enhanced with alternate coupling chemistry. Titanium silicide may also be used as structured surfaces making it possible to fabricate patterns down to the sub 100 nm range w9x using technology developed for
This work is performed within the Nanometer Structure Consortium and financial support was given in parts from The Swedish Strategic Foundation and The Swedish Technical Science Research Council and the EU project DIAMONDS ŽBIO4-97-2199.. The authors would like to thank Ericsson Components for the titanium silicide wafer and the Division of Analytical Chemistry, Lund University for providing some of the electrode materials used. Technical assistance from Lena Timby, Mariusz Graczyk and Per ¨ Osterstrom ¨ is also gratefully acknowledged. References w1x A.N. Broers, Resolution limits for electron-beam lithography, IBM J. Res. Dev. 32 Ž1988. 502–513. w2x S.-B. Carlsson, T. Junno, L. Montelius, L. Samuelson, Mechanical tuning of tunnel gaps for the assembly of single-electron transistors, Appl. Phys. Lett. 75 Ž1999. 1461–1463. w3x S. Gernet, M. Koudelka, N.F. de Rooij, Fabrication and characterization of a planar electrochemical cell and its application as a glucose sensor, Sens. Actuators 18 Ž1989. 59–70. w4x U. Gottlieb, F. Nava, F. Affronte, O. Laborde, R. Madar, Electrical transport in metallic TM silicides, in: K. Maex, M. van Rossum ŽEds.., Properties of Metal Silicides, INSPEC, London, 1995, pp. 189–204. w5x S.P. Murarka, Self-aligned silicides or metals for very large scale integrated circuit applications, J. Vac. Sci. Technol. B 4 Ž1986. 1325–1331. w6x L.A. Clevenger, R.W. Mann, Formation of epitaxial TM silicides, in: K. Maex, M. van Rossum ŽEds.., Properties of Metal Silicides, INSPEC, London, 1995, pp. 61–70. w7x P. Wikstrom, ¨ C.-F. Mandenius, P.-O. Larsson, Gas phase silylation — a rapid method for preparation of high peformance liquid chromatography supports, J. Chromatogr. 455 Ž1988. 105–117. w8x M.B. Wilson, P.K. Nakane, in: W. Knapp, K. Holubar, G. Wick ŽEds.., Immunofluorescence and Related Staining Techniques, Elsevier, Amsterdam, 1978, pp. 215–224. w9x L. Montelius, B. Heidari, M. Graczyk, I. Maximov, E.-L. Sarwe, T.G.I. Ling, Nanoimprint- and UV-lithography: mix and match process for fabrication of interdigitated nanobiosensors, Microelectronic Engineering, in press. w10x L. Montelius, J.O. Tegenfeldt, T.G.I. Ling, Fabrication and characterization of a nanosensor for admittance spectroscopy of biomolecules, J. Vac. Sci. Technol. A 13 Ž1995. 1755–1760.
