ISFET.pdf

(536 KB) Pobierz

Sensors and Actuators B 83 (2002) 90±97

ISFET glucose sensor system with fast recovery

characteristics by employing electrolysis

Keun-Yong Park

, Sang-Bok Choi

, Minho Lee

Byung-Ki Sohn

, Sie-Young Choi

a,*

School of Electrical Engineering and Computer Science, Kyungpook National University, 1370 Sankyuk-dong, Puk-Gu, Taegu, 702-701, South Korea

Department of Sensor Engineering, Kyungpook National University, Taegu, 702-701, South Korea

Abstract

The long recovery-time involved with an ion-sensitive ®eld-effect transistor (ISFET)-based glucose sensor, in relation to long-term

monitoring is one of the main problems preventing widespread commercialization. Accordingly, this paper proposes a new ISFET glucose

sensor system with rapid recovery characteristics by an electrolysis method. In addition, since practical application requires a detailed

understanding of the in¯uence of electrolysis, the current article also investigates the recovery characteristics of an ISFET glucose sensor

including the long-term stability, and the recovery time when using electrolysis. As a result, it was found that the recovery time of an ISFET

glucose sensor was reduced to less than two minutes when using electrolysis in contrast to the 10 or 20 min required using a conventional

method. Furthermore, no variation in the long-term stability at biasing the reduction potential was observed for 15 days.

2002 Elsevier

Keywords:

Enzyme ®eld-effect transistor; Glucose sensor; ISFET; Electrolysis

1. Introduction

Ion-sensitive ®eld-effect transistors (ISFETs) have been

developed on the basis of a metal oxide ®eld effect transistor

(MOSFET). Since, Bergveld introduced the ISFET for

neurophysiological measurement in 1970, a lot of funda-

mental researches have been performed on these devices and

various applications proposed [1,2]. Among them, ISFET

glucose sensors are very attractive for the fabrication of

instruments and measurement systems that need to be com-

pact, simple, and reliable to allow their application to the

continuous ¯ow analysis of chemical parameters in complex

systems like monitoring the environment or process control.

ISFET glucose sensors with such advantages can also be

widely used in areas such as medical diagnostics, monitoring

clinical or environmental samples, fermentation and bio-

process control, testing pharmaceutical or food products,

and early warning systems for chemical and biological

warfare [3,4]. In particular, as the demand for monitoring

continues to increase, the real-time measurement of glucose

is an important area for industrial development. Although

ISFET glucose sensors have an especially appropriate con-

®guration for application to in situ and in vivo monitoring

Corresponding author. Tel.:

82-53-950-5518;

fax:

82-53-950-5508.

E-mail address:

sychoi@ee.knu.ac.kr (S.-Y. Choi).

[1,2], the recovery-time delay has prevented the widespread

commercialization of ISFET glucose sensors. Accordingly,

this paper presents a novel method for improving the recov-

ery time through applying a reduction potential to a platinum

(Pt) working-electrode, and shows the recovery characteris-

tics of the ISFET glucose sensor system including the long-

term stability and the recovery time when using electrolysis.

1.1. Theory

ISFETs are electronic devices that have been developed

to measure pH and other ions. The ion sensitivity of these

devices depends on the nature and physicochemical proper-

ties of the membranes in contact with the electrolyte ana-

lysis. Using these basic mechanisms, a conventional ISFET

glucose sensor determines the quantity of glucose by mea-

suring the pH variation due to the dissociation of gluconic

acid generated by the enzyme reaction and the electrolysis of

hydrogen peroxide (H

). The sensing mechanism of an

ISFET glucose sensor is shown in Eqs. (1)±(3) respectively

[5,6].

b-d-glucose

Gluconic acid

Gluconate

0:7 V vs:Ag=AgCl

GOD

(1)

(2)

(3)

0925-4005/02/$ ± see front matter

PII: S 0 9 2 5 - 4 0 0 5 ( 0 1 ) 0 1 0 4 9 - 8

K.-Y. Park et al. / Sensors and Actuators B 83 (2002) 90±97

The hydrogen ions generated by Eqs. (2) and (3) exist in the

enzyme-immobilized membrane of the ISFET glucose sen-

sor. These hydrogen ions delay the recovery time due to the

continuation of their chemical reaction to equilibrate the

concentrations inside and outside the enzyme-immobilized

membrane after a response. This delay in the recovery

time creates a problem for applications that require repeat

measurements. Accordingly, this paper proposes a novel

electrolysis method where a

À0.7

V reduction potential is

applied to the platinum working-electrode to improve the

recovery time of an ISFET glucose sensor [3,4]. After the

response of the sensor, hydroxyl (4OH

) ions are generated

in the enzyme-immobilized membrane by applying a reduc-

tion potential to a combination of O

, 2H

O, and 4e

. The

generated hydroxyl ions then reduce the recovery time of

the ISFET glucose sensor because they combine quickly

with any hydrogen ions. Finally, an ISFET glucose sensor

obtains an output value similar to the initial state, namely,

the concentration of hydrogen ions present in the enzyme-

immobilized membrane is diminished by the hydroxide ions

generated by the reduction potential. This reaction is shown

in Eqs. (4) and (5), respectively.

4OH

À0:7

(4)

(5)

To prove the effectiveness of the proposed method, changes

in the recovery time were monitored when repeating the

experiment with different glucose concentrations.

2. Experiment

2.1. Materials

The glucose oxidase (GOD, EC 1.1.3.4) from Aspergillus

niger and bovine serum albumin (BSA, 100 mg) were obtained

from Sigma. The

g-aminopropyltriethoxysilane

(g-APTES,

100 ml) was obtained from Aldrich. A 25 wt.% aqueous

solution of glutaraldehyde (GA) was purchased from Sigma.

Fig. 1. Top view of chip layout (a) and cross-sectional view (A-B) of ISFET glucose sensor with enzyme-immobilized membrane (b).

Fig. 2. Block diagram of ISFET glucose sensor system.

K.-Y. Park et al. / Sensors and Actuators B 83 (2002) 90±97

The silicone rubber used for encapsulation was obtained

from Dow Corning Korea Ltd. (Korea). All reagents were

of pure analytical grade. Deionized water was used throughout

the experiments for the preparation of the samples, buffers, and

other solutions.

2.2. Fabrication of ISFET

A pH-ISFET was fabricated in a p-well on an n-Si wafer

using the CMOS process. A well structure was used to

achieve a better electrical isolation between the ISFET

Fig. 3. Design of analog circuit: (a) null balance circuit; (b) biasing circuit for oxidation and reduction potentials.

K.-Y. Park et al. / Sensors and Actuators B 83 (2002) 90±97

and the sample solution. To minimize any drift caused by the

hydration of the ®lm, the gate structure of the pH-ISFET

consisted of multilayer of Ta

/Si

/SiO

, with respec-

tive thickness of 50 nm. The size of ISFET chip was

1:0 mm

1:8 mm and the gate size was 20

400

mm.

The sensitivity of the fabricated ISFET was 58 mV/pH from

pH 4.0 to pH 10 [11]. The drift of the output voltage was less

than 0.3 mV/h. For the electrolysis, a platinum working-

electrode was created on the pH-ISFET base device. The

Pt/Ti layer was formed around the gate region encompassing

an area of 0.056 mm

on the surface of ISFET by the lift-off

process, which is a basic semiconductor process technique.

The pH-ISFET based device was ®xed onto a header and

then the metal contact area of the ISFET chip was wire-

bonded. Next, the ISFET chip was encapsulated by hand

using silicon rubber, leaving a small opening over the gate

region, then was dried for 48 h at room temperature.

2.3. Fabrication of enzyme-immobilized membrane

The enzymatic solution consisted of 100

of a 20 mM

phosphate buffer solution in which 5 mg GOD and 5 mg

BSA had been dissolved. The surface of the ISFET gate

region and platinum working-electrode was pretreated in

diluted hydro¯uoric (HF) acid. To increase the adhesion

between the enzyme-immobilized membrane and the Ta

gate, the gate region was coated with 0.8

ml g-APTES

and

then heated for 6 min at 80

8C.

Next, 0.8

of the enzymatic

solution was dropped around the gate region including the

platinum working-electrode area. Then 1.3

of 25 wt.%

GA was dropped for a chemical cross-linking reaction.

Finally, drying at room temperature for 4 h formed the

enzyme-immobilized membrane that responds to glucose

[12,13]. A cross-sectional view of the completed ISFET

glucose sensor is shown in Fig. 1.

2.4. Developed ISFET glucose sensor system

Fig. 2 shows the block diagram of the developed ISFET

glucose sensor system. The sensor output signal is converted

into voltage signal using null balance circuit, and ®ltered

using LPF (low pass ®lter), then digitized by ADC (analog-

to-digital converter). This digitized input signal is controlled

using the microprocessor. The threshold voltage of the

ISFET is stabilized by the inner feedback loop using the

DAC (digital-to-analog converter). The 80C196 micropro-

cessor is used for control of the operation including the

switching time for electrolysis. The 0.7 V at platinum (Pt)

working-electrode is used for enhancing the sensitivity for

oxidation and the

À0.7

V at Pt working-electrode is used for

obtaining the fast recovery-time. We use the null balance

circuit for biasing the ISFET sensor as shown in Fig. 3(a).

The ISFET was operated in a constant drain±source current

126

mA)

and drain±source voltage (V

3 V) mode

[7]. The recovery characteristics were measured by using a

double electrode structure of the Pt working-electrode, and

Ag/AgCl conventional reference electrode that provided a

proper gate voltage for the ISFET operation. The sensitivity

or response of the sensor was determined to base on the

value of the slope of the curve (DV

) as a function of the

hydrogen ion concentration relative to the glucose concen-

tration. Also, the LPF shown in Fig. 3(a) eliminates the

60 Hz interference. Fig. 3(b) shows the biasing circuit for

oxidation and reduction potentials. Control strategy of the

switching of Fig. 3(b) is shown in Fig. 4. P1 and P2 are ¯ag

is output

voltage through ADC. The ``steady state'' in the ¯owchart of

Fig. 4 is de®ned if the following condition is satis®ed,

where

and

are the variation ratio of output voltage of

ADC and variation ratio of time, respectively, and the value

(0.01±0.09) is determined by each calibration process. If

P1 controlled by microprocessor is ``high'', the Pt working-

electrode is biased to 0.7 V for oxidation. Also, if P2

controlled by microprocessor is ``high'', the Pt working-

electrode is biased to

À0.7

V for reduction. The initial

values of P1and P2 are low. When the value of

is steady

state, P1 is high, and the Pt working-electrode is biased to

0.7 V for oxidation, and then current output voltages of the

Fig. 4. Flowchart for biasing circuit control.

K.-Y. Park et al. / Sensors and Actuators B 83 (2002) 90±97

Fig. 5. Configuration of the developed ISFET glucose sensor system.

Fig. 6. Voltammograms of platinum working-electrode for determining oxidation and reduction potential: (a) oxidation potential; (b) reduction potential.

Plik z chomika:

tho.doduc

ISFET.pdf

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: