Hydrogen Sensing Behavior of Tellurium Thin Films Studied by A.C. Measurements
Hydrogen Sensing Behavior of Tellurium Thin Films Studied by A.C. Measurements
Hydrogen Sensing Behavior of Tellurium Thin Films Studied by A.C. Measurements
For the first time it is pointed out that tellurium films exhibit sensitivity to H 2 at room temperature
along with sensitivity to NO2. The hydrogen gas sensing performance of tellurium thin films was
investigated by method of impedance spectroscopy. The impedance spectra are strongly influenced by
gaseous environment but the effect of target gas is mainly due to variation of resistance of the film. It is
assumed that the sensitivity of tellurium films to H2 arises because of reducing effect of oxygen
priory absorbed on the surface of the film from carrier (dry air) gas. The high concentration of oxygen
in carrier gas promotes the formation of a catalytic gate, which can be removed by other gases
including hydrogen. Removing of the priory-adsorbed oxygen results in decreasing of both, hole
concentration and conductivity of the surface and intragrain regions of tellurium film. Due to
impedance change in different direction, reducing H2 may be distinguished from oxidizing NO2, hence
the effective, operating at room temperature H2 sensors can be manufactured using tellurium-based
films.
Key words: Gas sensing, Tellurium, Hydrogen, A.C. measurements
VEGA TESCAN TS 5130 MM scanning electron is assigned to a sum of Ohmic resistance due to
microscope (SEM). electric connection, but resistance R0 and capacity C
The film was encapsulated in a standard TO 8 are distributed to others contributors, the grain
sockets and then the contacts were thermally bounded boundary resistance and capacity being the main.
to socket pins, using the copper wires. The circle of Nyquist diagram shown in fig. 1b
The sockets with thin film sensing devices were is depressed owing to the dependence of both C and
put into a test cell (of 10ml volume) in which the R on frequency. From the left and right intercepts
gases were injected with a flow rate of 100ml / min , of semi circle with the Re(Z ) axis the values of
parallel to the film surface. R 0 and Rr R R0 can be estimated. Thus, R 0
Hydrogen gaseous media, with concentration 1%
by volume was obtained from cylinders (Linde, was found to be very small, only about 5Ohm . That
Germania). Nitrogen dioxide (NO2) vapors with is the arc practically passes through the origin and the
concentrations of 1,5 ppm were obtained by using right intercept gives the value of Rr 20 kOhm .
calibrated permeation tubes (Vici Metronics, USA),
For a parallel R C circuit the impedance is
which were incorporated subsequent into the
experimental set up. Dry synthetic air was used as given as:
the carrier and reference gas. 1 1
Impedance measurements were carried out in Z ( )
frequency range of 5Hz to 13MHz using a HP 4192
Y ( ) 1
i C
A impedance analyzer. R
R R (1 i C R )
3. RESULTS AND DISCUSSION
3.1 Impedance behavior under dry air 1 i C R 1 ( C R ) 2
(1)
Before checking the effect of different harmful
gases on a.c. conductivity the tellurium films were where Y ( ) is the admittance, i 1 is the
aged by 12 months in normal conditions and the imaginary number , 2f , f -the frequency.
measurements have been performed under synthetic
Thus, the real and imaginary parts of the
dry air. Figure 1b shows the typical complex
impedance are:
impedance diagram in Nyquist plot obtained in pure
synthetic dry air from a thin film device at room R
Re( Z ) (2)
( 22 C ) temperature. 1 2 R2 C2
and
C R
Im( Z ) (3)
1 2 R2 C2
From these system of equations the values of
R and C of the film can be evaluated as:
Im 2 ( Z ) Re 2 ( Z )
R (4)
Re( Z )
Figure 1 - a) Interdigital electrode structure used to and
measure the a.c. conductivity; b) Nyquist diagram of
Im( Z )
C
an aged at 22oC tellurium thin film in pure synthetic
dry air; c) Suggested equivalent circuit
Im ( Z ) Re 2 ( Z )
2
(5)
The diagram shows a slightly depressed semi From equations (4) and (5) the resistance Rm,
circular arc with a center displaced below the real
capacitance Cm and time constant m ( 2f m ) of
1
axis, owing to presence of distributed elements in
tellurium-based device [8]. A simplified equivalent the film can be estimated at characteristic frequency
circuit inserted in Fig. 1 (c) can interpret the Nyquist fm, that is the frequency at which the imaginary part -
plot. The frequency independent serial resistance R0 Im(Z) reaches its maximum value. Because of he-
terogeneity of the material-electrode system the where fm - is the characteristic frequency at which the
relaxation time (time constant) m , estimated from imaginary part Im(Z) reaches the maximum value,
the complex impedance represents a mean value for Rm and C m are the resistance and capacity of the
the complete thin film device: film at characteristic frequency fm. The characteristic
frequency (fm), impedance (Z) and estimated from
1
m m1 Rm C m (6) equation (6) the time constant ( m ) of the sample in
2f m
dry synthetic air, are listed in Table 1.
3.2. Impedance behavior under mixture of dry air more the addition of NO2 decreases both impedance
with H 2 and NO2 and Rm (at characteristic frequency, which also is gas
influenced) but addition of H2 increases these
Figure 2 reports the complex impedance spectra parameters.
of aged tellurium-based films upon exposure to
different test gases that is NO2 and H2. It is seen that 3.3. Real and imaginary parts of impedance for Te
addition of these gases to dry synthetic air does not films in H 2 and NO2 gaseous media.
change the general shape of curve, i.e. they influences Figure 3 reports the spectra of the real part of
all elements of the equivalent circuit. The values of impedance of tellurium films upon exposure to
characteristic frequency, impedance and time constant different test gases. It is seen that addition of 1,5 ppm
m of the film at this frequency, by indicated concen- of NO2 to dry synthetic air diminishes the real part of
trations of NO2 and H2 at room temperature, are impedance by ~ 10 kOhm in the frequency range 1,0-
summarized in Table 1. 103 kHz. On the contrary, the addition of 1 % of
Listed in this table values of Rm and Cm (the volume of H2 to dry synthetic air enhances the real
resistance and capacity at characteristic frequency) part of impedance by ~7,0 kOhm in the much shorter
have been obtained from Eq. (2) and (3) applied to frequency range: 1,0 100 kHz.
the data of Figure 2.
the peak of imaginary part of impedance shifting it to Figure 5 shows the sensor sensitivity as a function
higher frequencies but the addition of H2 vapors of the measurement frequency during the exposure to
results in a vice-versa behavior. Analysis of these 1,5 ppm NO2. The sensitivity (here and further) is
spectra help determining the influence of tested defined as absolute variation of measured value
harmful gases on all elements of the equivalent circuit (impedance or imaginary part of impedance) for a
of the sample showed above in the Table 1. selected frequency in mixture of carrier gas with NO2
divided by the measurement value in the carrier gas at
the same frequency, in percents per ppm.
The response curves for either impedance or
imaginary part are nearly independent on frequency
until approximately 300 kHz, then go down, but
sensitivity to NO2 is maintained until 10 MHz. The
sensitivity in d.c. and impedance measurements
amounts to approximately 30 % / ppm, but evaluating
the imaginary part as the sensor response results in an
increasing of sensitivity until ~50 % / ppm. The high
sensitivity, as well as the large frequency range of
response to NO2 supports the early-proposed mecha-
nism of nitrogen dioxide interaction with chalco-
genides [9], which involves strong chemisorption
due to interaction between odd electrons of NO2
molecules and lone pair electrons of tellurium based
Figure 4 - Spectra of imaginary part of impedance chalcogenides.
upon exposure to different test gases
As the response to different gases strongly 3.4.2. Hydrogen
depends of applied frequency, it becomes interesting Figure 6 shows the sensor sensitivity as a function
to analyze the frequency dependences of sensitivity to of measurement frequency using the hydrogen as a
different target gases. test gas. It is observed that sensitivity to hydrogen is
by four orders of magnitude smaller that to NO2, but
3.4. Sensor sensitivity versus applied frequency
also cover a large range of frequencies and can be
3.4.1. Nitrogen dioxide clearly detected. Unlike exposure to NO2 the
D. c. resistance of tellurium films is known to impedance response spectra to hydrogen go down
decrease reversibly in presence of NO2 due to inter- starting with approximately 150 kHz but at 1,0 MHz
action of adsorbed species with lone pair electrons, the sensitivity to H2 practically disappears.
which from the upper part of the valence band [9].
Apparently, by changing from d.c. to a.c. technique
the mechanism of interaction cannot be modified but
the sensitivity (or selectivity) can be increased.
60
Sensitivity , [ % / ppm ]
50
40 NO2
30
20
Im (Z)
IZI
10
IZVOD
PROUAVANJE PONAANJA VODONIKA NA OSETLJIVIM TANKIM
FILMOVIMA OD TELURA POMOU AC MERENJA
Po prvi put je pokazano da su Telurovi tanki filmovi osetljivost na H 2 na sobnoj temperaturi uz osetljivost na
NO2. Reagovanje vodonika na Telurove tanke filmove ispitan je metodom spektroskopske impedanse. Spektri
impedanse su pod jakim uticajem gasovitog okruenju, ali je efekat ciljanog gasa uglavnom zbog varijacija
otpora filma. Pretpostavlja se da osetljivost Telurovih tankih filmova za H 2 nastaje zato - da se smanji
dejstvo kiseonika apsorbovanog na povrini filma od provodnog gasa (suv vazduh). Visoka koncentracija
kiseonika u noseem gasu podstie formiranje katalitikog kapije, koja moe ukloniti druge gasove,
ukljuujui i vodonik.
Uklanjanje apriori apsorbovanog kiseonika smanjenjuje koncentracije i provodljivosti na povrinskim i
intrgranularnim regionima Telurovih tankih filmova. Zbog promene impedanse u drugom pravcu, redukcija
H2 moe se razlikovati od oksidacije NO2, stoga efikasno, rad na sobnoj temperaturi H2 senzora mogu biti
proizvedeni pomou telur osnovnih filmova.
Kljune rei: vodonik, osetljivi tanki filmovi, telur, AC merenja
Rad primljen: 18.02.2013.
Originalni nauni rad