Semiconducting gas sensor incorporating a sparking decomposer

199

Sensors and Actuators B, 2 (1990) 199-203

Semicondncting Gas !SensorIncorporating

a Sparking Decomposer

J. MIZSEI Technical University

of Budapest,

Golbnann

Gy. ter 3, H-1521 B&pest

(Hungary)

(Received August 14, 1989; in revised form January 24, 1990; acceptedFebruary 15, 1990)

Abstract In this article a new method is demonstrated for improving gas sensor characteristics. An electronic sparker is included in a semiconducting SnO, thin-film sensor system for cracking the gas molecules to be detected. The distance between the electrodes is 0.2 mm, which conforms to a 1- 1.8 kV breakdown voltage at 0.1-0.5 mA coming from a current-limited power supply. A good correlation exists between the characteristics of this composite sensor system and the structure of detectable gas molecules. For small gas concentrations (below 1% for CH.,, 0.01% for normal C,H,, and C4H& the sensor resisitivity increases due to the spark discharge, as well as in clean air. At moderate and high concentrations of hydrocarbons, the spark discharge improves the sensitivity, depending on the structure and the hydrogen content of the molecule. Chlorine and other halogen-containing gases and solvent vapours, such as CCL, CHC&, CH&l, and CF2Clz, also yield a higher resistivity after sparking, even at high concentrations (5% or more). Introduction The theory, fabrication and application of semiconducting gas sensors have been well developed in the last twenty years. However, their limited selectively and sensitivity are still problematical. Several methods have been used for improving the previously mentioned characteristics (e.g., using catalytically active materials [ 11,chemical and physical filtering [ 21 and conversion [ 31 of the gases to be measured).

Experimental

A new method is demonstrated here for improving the sensor characteristics. In order to :rack gas molecules, an electronic sparker is in:luded in the complete sensor system, as shown in Fig. 1. The distance between the electrodes is 19254005/90/%3.50

sparker

~1 kV

PI

gas inlet

I spark gap (0.2 ml

T

rtal

box

Fig. 1. Schematic view of the experimental set-up.

0.2 mm, which conforms to a I- 1.8 kV breakdown voltage at 0.1-0.5 mA coming from a current-limited power supply. (The current limit and the output capacitor determine the number of sparks per second, thus the current limit is a useful parameter for characterization of the spark intensity.) The glass tube housing the sensor is closed (no continuous flow of gas). The gas concentration can be calculated as the volume ratio of the gas introduced into the glass tube by an injection syringe, and the glass tube itself respectively. The sensor element is a Pd-doped SnOz thin film. The technology and theory of the operation of this semiconducting sensor layer have been discussed earlier [4]. The sensor layer was sputtered onto a ceramic plate with a preprinted heating resistor layer and a gold contact layer. The complete sensor chip is mounted into a TO header with metal mesh in the cup. This construction is more stable and, therefore, better than the earlier one, but the gas-sensing properties are the same. The sensor temperature is about 120-M “C. At that temperature the element is basically sensitive for hydrogen, carbon monoxide and chlorine. For hydrogen and other reducing gases, which create donors (positive ions) at the surface, the resistivity of the sensor layer decreases; for chlorine and 0 Elsevier Sequoia/Printed

in The Netherlands

200

other oxidizing gases which behave as acceptors (negative ions), the resistivity increases.

Measurements The sensor response is expressed in logarithmic form as S = ln(%lR,) (1) where R,, is the steady-state resistivity at the starting point of the given experiment, while R, is the steady-state resistivity in the air/gas mixture, but it can also be considered as a time-dependent function (R, = R,(t)) for characterization of the velocity of time-dependent processes. For this case, the derivative of S can be calculated easily from the 4(f) plot as the relative change of the resistivity divided by the time interval:

as/at = a/at(h(R&J) = -( l/R,)aR,/iYt x - ARJAtR,

(2) Thus S and as/at (relative slope of the R,(r) plot) are positive for reducing, and negative for oxidizing gases and/or processes. (It was also shown earlier [4] that S is nearly equal to the chemical potential change in kT/q units in the gas-sensitive semiconductor film.) Figure 2 shows the time diagram of the measurement, and typical response plots of the sensor (the previously defined quantities are also included). The sensor itself is calibrated with Hz and Cl2 without the sparker, and by sparking in clean air (without gas). In this case the resistivity also increases, similar to the effect of Cl,. The relative humidity was about 50% in the clean air of the laboratory at 22-24 “C.

Results In the air/gas and air/vapour systems the relative resistivity change depends on the concentration and chemical structure of the gas to be measured, and the intensity of the sparking. The characteristics for H,, Cl, and for some hydrocarbons are summarized in Fig. 3. For small concentrations (below 1% for CH.+ and 0.01% for normal C6Hr4 and C4HIO), the resistivity increases due to the spark discharge. At moderate and high concentrations of hydrocarbons the spark discharge improves the sensitivity, depending on the structure and the hydrogen content of the molecule. As can be seen from Fig. 4, chlorine and other halogen-containing gases and solvent vapours,

I

7

(t) -. -

Ro-

1. ‘\ I

2d

tim I I

point for

dS/dt talc.

lyt1 RJ, I

%z~~e time point

for

dS/dt talc.

Fig. 2. Time diagram of the measurement, and different responses of the sensor. u: Cl, or other oxidizing gas, CH,, inert gas, cool. uu: Cl, or other oxidizing gas, CH, at low concentration, clean air, inert gases. udz CH, at high concentration. d: reducing gases, H,, organic solvent vapour. du: reducing gases, H,, organic solvent vapour at low concentration, chlorine-containing gases and vapours. dd: hydrogen-containing gases and vapours at higher concentration.

such as CC&, CHC& , CH,Cl, and CFzClz also yield higher resistivity after sparking (the relative slope of the R,(t) curve is positive, thus S is negative), even at high concentrations (5% or more). The sensor -A IAt& response is also plotted for benzene an“a chlorobenzene in Fig. 4. Discussion As can be seen from the results, for low gas concentrations the spark discharge yields acceptor-type oxidizing products (probably the ozone from the air). These chemical species dominate over the low concentration of hydrogen or other reducing substances produced by the sparker. The donor-type reducing radicals are dominant only at higher concentrations of halogen-free gases.

201 6.

6.

s

s

.

H 5.

5.

4.

H

4.

H

CZHB C2HB

II

3.

H

31

C2%

2. CZHS

1

I

H

Ii

C2%

1. H

CZHS

H

C2%

C2%

0.. H

C29 Cl -1 .

1

Cl

Cl

(4 6

C4Yo

S

H

H

csHT4

5

H

H

4

CH

H

3

C496>

H

%Yr

100

2

H I

0

-1

-1

Cl

Cl Cl

-2

(Cl

Cl Cl

1,,,,,,,,1 10 PPI

Fig. 3. Steady-state C,H,; (c) C,H,,;

100 PP=

I,,II,,I 0.1 x

1 x

11,

I I, 10 cont. x

Cl

-2

,,,,,,,,,,1,,,,,,111llll 10 100 (d) PP= PP= ,

0.1 x

characteristics of the semiconducting gas sensor for different hydrocarkms, and for (d) C6H,,. *: after sparking.

1 x

H,

and cl,. (a),

10 x

CH,;

cont.

(b)

202

N/At*, thus for the reducing radicals

From the calibration for H, and CIZ the net rate of development of the reducing or oxidizing radicals can be estimated. A time interval At* can be expressed as

ahyat

= 8.66

x 1016

as/at

(4)

and for the oxidizing radicals

aivlat

ARIR, -- ARIR, AR,/R,At = as/at where AR/& is the relative resistivity change from the steady-state measurements: AR/& = -0.63 for 1OOppm Hz and AR/R,,= 0.54 for 100 ppm Cl,. The number of radicals developed by the spark discharge (within the time interval At*) corresponds to the number of chlorine or hydrogen molecules: N = 5.5 x lOI for 20 cm3 volume and 1OOppm concentration at atinospheric pressure. The net developing rate aN/at is At* = -

= -1017 as/at

(5)

Based on the previous method, these net rates of development are summarized in Table 1, as well as the measured sensitivities (S) calculated from the steady-state resistivities. This calculation contains an implicit assumption: the adsorbed donors and acceptors add or subtract the same number of mobile electrons to or from the semiconductor space-charge layer, independently of their real chemical structure or nature.

TABLE 1. Net developing rates of the positive and negative radicals, and steady-state sensitivities of the sensor for different gases. (Steady-state situation has not been found: /) aN-/al

s without

(lo’%)

(10’4/s)

spark discharge

l-4 9

-0.01 0.05 1.16

i

0.1 x 1

52.1

-0.17 -0.03

i

0.5 1 0.1

0.1 x 1 0.1 x 1

5.3 2.9 6.6

0.14 0.25 0.06

i

0.5 1 0.1

0.1 x 1 0.1 x 1

1.7 5.4

0.2 0.35 0.05

i

0.5 1 3.5 10

0.5 0.5 0.5 0.5

x x x x

1.8 1.8 1.8 1.8

0.65 8.6 30

0.1 0.5 1 2.3 5

0.5 0.5 0.5 0.5 0.5

x x x x x

1.8 1.8 1.8 1.8 1.8

2 8.1 18.9 19

0.01 0.1 0.5 1

0.5 0.5 0.5 0.5

x x x x

1.8 1.8 1.8 1.8

2.4 5.4 6.7

0.01 0.05 0.1 0.5 1

0.5 0.5 0.5 0.5 0.5

x x x x x

1.8 1.8 1.8 1.8 1.8

1.7 2.3 8.2 9-12

Cont. %

Spark (mA x kV)

:?a CH$, ’

5 5

0.1 x 1 0.1 x 1

Ccl, F, ccl,

5

C,H,Cl

C,H, (34

C,H,

C4H,o

W&4

only Air

S with spark discharge

aN+/ar

Gas or vapour

4.8

1.7

6.2

-0.01 -0.01 -0.01 -0.02

I 0.1 3.51 6.04

2.8

0.07 0.26 0.67 1.4 1.67

-0.07 0.46 2.18 3.23 3.33

2

0 0.01 0.04 0.09

-0.07 0.86 2.67 3.12

1.6

0 0 0.03 0.13 0.09

-0.05 0.36 1.04 2.86 3.25

0.05 x 0.7 1 spark/s

0.9 0.2

\

0.2 0.1 x 1.6 1 0.5 x 1.8 0.6 x 2

2.4 1.1-1.6 3.4 6.4

;

203

excluded. Thus, the direction of further investigations is the analysis of molecules and radicals produced by the sparker. Finally, for further development, and also for practical purposes, the sparker can be integrated or matched with other types of gas sensors.

.3/s

6,

5

4

3

2

Acknowledgements 1

0

)

0

Ccl4

This work was established during the UNIDO scholarship of the author at the Technical University of Munich. The author is indebted to Professor Dr R. Miiller for the opportunity to work, and for helpful discussions. The development of the.new sensor construction was supported tinancially by the National Committee for Technical Development.

CC12F2

-1 I

-2

-3

-4

References

-6

-6

-7

-6

-9x1o-3/s

1% 0.5x

6X

5%

5%

5%

5%

0.1X

1X 0. sx 0.1%

Fig. 4. The measured %/a~ values for different hydrocarbons and halogenides at 0.1 mA x 1 kV sparking power. All results are derived from the relative slope of the R,(r) curves at the time the spark begins.

Conclusions

The sensitivity and selectivity of a semiconducting gas sensor can be improved by including an electronic sparker in the sensor system. A good correlation exists between the characteristics of this composite sensor system (i.e., S and aS/&) and the structure of detectable gas molecules. However, the complete mechanism of operation is not well understood. Molecules are usually cracked and decomposed in a gas mixture due to the spark discharge, but the possibility of oxidation or other chemical reactions also cannot be

1 J. F. McAleer, P. T. Moseley, J. 0. W. Norris, D. E. Williams and B. C. Tofield, Tin dioxide gas sensors, Part 2. The role of surface additives, J. Chem. Sot., Faraaby Trans. I, 84 (1988) 441-457. 2 R. Miiller and E. Lange, Multidimensional sensor for gas analysis, Sensors and Actuurors, 9 (1986) 39-48. 3 W. Homik, A novel structure for detecting organic vapours and hydrocarbons based on a Pd-MOS sensor, Sensors Md Actuators, BI (1990) 35-39. 4 J. Mizsei and J. Harsanyi, Resistivity and work function measurements on Pd-doped SnO, sensor surface, Sensors and Actuators, 4 (1983) 397-402.

J. Mizsei was born in Jaszladany, Hungary, in 1952. He received his diploma in electrical engineering ( 1976), and Dr.Techn. degree (1979) from the Technical University of Budapest, candidate for science degree from the Hungarian Acad. of Sci. (1987). He has worked at the Department of Electron Devices of the Technical University of Budapest since 1977. He had a sabbatical year in 1981-82 at the Enterprise for Microelectronic, Budapest. His main subjects are semiconductor technology (investigation and education), electron devices, surface physics of semiconductors, and sensors.