|
|
|
|
Mark J. Milgrew1,2
and David R. S. Cumming1
1. University of Glasgow, Department of Electronics and Electrical Engineering, Oakfield Avenue, Glasgow, G12 8LT, Scotland.
2. Tel: +44 (0) 141 330 5226, Fax: +44 (0) 141 330 6010, E-mail: mark@elec.gla.ac.uk, Web: http://www.elec.gla.ac.uk/~mark.
|
|
There is considerable interest in developing solid state based sensors for integration with medical diagnostic and analytical devices. One important area is the development of multi-sensor array nanotechnology that can perform both real-time and non-invasive measurements on an analyte.
The analysis of cultured cell populations is the main area of application for this technology. The ability to conduct detailed studies on ion channels (proteins) of cell membranes is crucial to the development of new and improved drugs (Figure 1). It is also desirable to record temporally and spatially the electrical activation sequence of cells to gain a deeper insight into disease pathology. Furthermore there is a demand for high throughput cell screening and a need to be able to pinpoint damaged cell tissue with high precision.
Figure 1. The sodium-potassium pump: a specific case of cell ion transport.
|
|
The aim of this work is to develop arrays of highly functional sensors for the analysis of cell tissue in-vitro in a two dimensional environment. The technology employed must be sufficiently generic so that it is scalable and allows more than one kind of sensor to be implemented in an array. The devices will be used to monitor the behaviour of proteins embedded in cell membranes by measuring the extracellular hydrogen, potassium, sodium, and calcium cation concentrations.
The first objective is to develop a sensor array of electrochemical cells using an unmodified standard CMOS process. All components of the design should be incorporated on-chip to form a complete integrated sensor system. The second objective is to functionalise the cells in the array with minimal post-processing using standard photolithographic techniques. The third objective is to demonstrate the functionality of the devices by capturing data from a variety of cell cultures.
|
|
An ion selective FET (ISFET) can be fabricated in an unmodified standard CMOS process (Figure 2). The transducer of the sensor is formed by connecting a floating electrode to the gate of a MOSFET. The passivation layer is sensitive to H+ and gives a linear response. The silicon nitride (Si3N4) membrane behaves like a variable voltage source in series with the floating gate electrode.
Figure 2. Cross-section through a MOSFET-based sensor.
A biocompatible photocurable membrane can be formed by creating a basic polymer composition consisting of the reagents: urethane Ebecryl 270, crosslinker hexanediol diacrylate, and photoinitiator Irgacure 651. A cation selective membrane is then formed by adding a plasticizer, an ionophore, and a lipophilic salt to the basic composition. The final membrane is spun on to the silicon substrate and cured by ultraviolet light. Ionisable sites on the membrane surface react with the analyte to create a pX-dependent potential difference across the membrane.
A complete integrated sensor system is developed by eliminating the need for a conventional reference electrode. An electrochemical cell can be created using an electrically identically ISFET and reference FET (REFET) pair biased by a common pseudo reference electrode (PRE). These components are integrated into a sensor interface circuit that performs a differential potentiometric measurement to establish the change in pX of an analyte.
|
|
An nxn array is designed by considering a single column of n electrochemical half-cells (Figure 3). The source-drain voltage and the drain current are fixed by the source-and-drain follower circuit and the current source for the ISFET in the enabled row.
Figure 3. Schematic of a column of n electrochemical half-cells.
For each column of n electrochemical cells, the circuit configuration is employed for both ISFETs and REFETs. The source voltage of the devices are connected to an instrumentation amplifier to allow each cell in the column to make a differential potentiometric measurement. The readout circuitry for the nxn array consists of an n-row decoder, an n-column sequencer, and an analog-to-digital converter.
|
|
A 2x2 sensor array is developed using Cadence v4.4.5 and the Austria Micro Systems (AMS) HIT-Kit v3.40. The devices are fabricated through the Europractice ASIC Service using the AMS 0.35µm CSI standard CMOS process (Figure 4). This is a 3-metal, 2-poly layer process with an operating voltage of 3.3V.
Figure 4. Micrograph and photograph of the nanosensor array integrated circuit.
|
|
All of the signal processing and readout circuitry has been characterised and is in good agreement with simulation results. A plot of the ISFET source voltage against the PRE voltage illustrates that the linear operating range of a sensor is 1.25V (Figure 5). For a theoretical sensitivity of 57mV/pH, an ISFET is in principle capable of detecting a range of 1-14pH.
Figure 5. Source-PRE voltage characteristic for an ISFET.
|
|
A complete multi-sensor array integrated circuit has been designed and fabricated in an unmodified standard CMOS process. Although the devices are sensitive to pH on return from the foundry, the design is generic and can be functionalised for various ions. Further work will involve developing optimal photocurable cation selective membranes and demonstrating the functionality of the devices with different cell cultures.
|
|
Copyright © Mark J. Milgrew 2003. All rights reserved.
|
|
|
