Room-temperature hydrogen gas sensor

Article abstract of DOI:10.1063/1.2077865

The hydrogen gas sensing ability of n -type PtSiporous Si Schottky junctions is investigated at room temperature. These junctions exhibit a breakdown-type current-voltage curve, whose breakdown voltage depends on the gas content inside the pores. Hydrogen replaces gases with an inherent dipole moment and increases the breakdown voltage. The response time is very fast, around 6 s, and the recovery time is about 60 s. Concentrations as low as 10 ppm have been detected in the presence of methanol and acetone. Detection has been tested at different temperatures ranging from 273 to 300 K.

Full text of DOI:10.1063/1.2077865

Room-temperature hydrogen gas sensor
Farshid Raissi and Rashid Farivar
Citation: Applied Physics Letters 87, 164101 (2005); doi: 10.1063/1.2077865
View online: http://dx.doi.org/10.1063/1.2077865
View Table of Contents: http://scitation.aip.org/content/aip/journal/apl/87/16?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
PtSi dominated Schottky barrier heights of Ni(Pt)Si contacts due to Pt segregation
Appl. Phys. Lett. 102, 123507 (2013); 10.1063/1.4799277
Schottky barrier height extraction from forward current-voltage characteristics of non-ideal diodes with high
series resistance
Appl. Phys. Lett. 102, 042110 (2013); 10.1063/1.4789989
Barrier tuning in thin Pt Si Si contacts
Appl. Phys. Lett. 89, 093508 (2006); 10.1063/1.2344847
Quantum efficiencies of PtSi p - Si composites in the 3 – 5 m wavelength range
Appl. Phys. Lett. 87, 233510 (2005); 10.1063/1.2140881
Ion-bombardment effects on PtSi /n -Si Schottky contacts studied by ballistic electron emission microscopy
J. Vac. Sci. Technol. B 18, 1942 (2000); 10.1116/1.1305269
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 71.185.28.15
On: Sun, 27 Apr 2014 13:38:26

APPLIED PHYSICS LETTERS 87, 164101 ͑2005͒
Room-temperature hydrogen gas sensor
a͒
Farshid Raissi and Rashid Farivar
K. N. Toosi University of Technology, P. O. Box 16315-1355, Tehran 16314, Iran
͑
Received 2 June 2005; accepted 23 August 2005; published online 11 October 2005͒
The hydrogen gas sensing ability of n-type PtSi/porous Si Schottky junctions is investigated at
room temperature. These junctions exhibit a breakdown-type current-voltage curve, whose
breakdown voltage depends on the gas content inside the pores. Hydrogen replaces gases with an
inherent dipole moment and increases the breakdown voltage. The response time is very fast, around
6
s, and the recovery time is about 60 s. Concentrations as low as 10 ppm have been detected in the
presence of methanol and acetone. Detection has been tested at different temperatures ranging from
73 to 300 K. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.2077865͔
2
Hydrogen sensing has been the subject of much research
taining HF, deionized water, and ethanol ͑25, 25, and 50%͒
1
–3
in recent years. Industrial applications—which deal with
the use, transportation, and storage of hydrogen as well as
the automotive industry, investing heavily in hydrogen fuel
cells as a pollution-free alternative to today’s internal com-
bustion engines—need rugged, stable, durable, and compact
sensors for standard leak detection and also area monitoring.
Several different types of detectors, working based on varied
to produce the porous samples. Application of a current den-
2
sity of 50 mA/cm for 20 min results in 70% porous
samples. Scanning electron microscope ͑SEM͒ micrograph
of such a sample is provided in Fig. 1. The pores size is
about 2–5 ␮m and their depth is about 3 ␮m. The porous
2
surface to which gas is exposed was 0.5 cm and the pore
size, depth, and distribution were uniform ͑and reproducible͒
across the samples. Pt was electrodeposited into the pores
4
–10
physical and chemical properties, have been suggested
ranging from metal-oxide gas sensors to ChemFET, and tran-
using
a solution containing H ͑PtCl ͒.6H O 4g/liter,
2 6 2
11
sient response devices. Of prime importance is the fact that
hydrogen is to be detected in air and in the presence of other
͑NH4͒2HPO4 20 g/liter, Na2HPO4 100 g/liter, at 90 °C.
Electrodeposition was performed by applying 20 mA of cur-
rent for 45 min which deposited a thick Pt layer into the
pores filling them completely. The samples were subse-
quently annealed at 400 °C for 20 min to create a thin PtSi
layer. Under such annealing conditions, the thickness of the
3
combustible hydrocarbons and alcohols. Selective hydrogen
detection based on different types of molecular membranes
1
2,13
or filters have been suggested.
We have recently shown that n-type PtSi/porous Si
1
5
PtSi layer had been estimated to be between 20 to 100 Å.
The excess unreacted Pt which fills the pores was then re-
Schottky junctions are able to sense certain gases, such as
1
4
CO, CO , and O , at room temperature. The reverse bias
2
2
1
5
moved using a standard solution. This leaves empty pores
behind. The ambient gas, therefore, fills the pores and
reaches the PtSi layer which has uniformly covered the pore
current-voltage ͑I-V͒ curve of these junctions exhibits a
breakdown-type behavior at small reverse bias voltages of
about 10 to 15 V, for the typical porosity levels of about 50%
to 70%. Breakdown-type behavior has been attributed to the
large fringing fields developed at the sharp edges of the po-
rous surface which can be orders of magnitude larger than
that of a smooth area Schottky junction under the same
1
5,16
conditions.
It seems that gases with inherent dipole mo-
ment change the strength of the electric fringing fields and
decrease the breakdown voltage. Gases without a dipole mo-
ment affect the I-V curve by replacing the gases inside the
pores which have dipole moment. These types of gases
1
4
increase the breakdown voltage.
In this letter, we propose these junctions as novel hydro-
gen sensors, which can detect hydrogen at room temperature
and below. When hydrogen is introduced, the ambient gas
inside the pores is replaced and the breakdown voltage
increases. The low-temperature operation of these detectors
as well as their fast response and recovery time could be
of great practical interest. These detectors offer a certain de-
gree of selectivity as they are not sensitive to
ethanol, methanol, or acetone.
n-type Si wafers which were ͑100͒ oriented, with a re-
sistivity in 13–33 ⍀ cm range were used. Anodic etching
was performed under ultraviolet radiation in a solution con-
FIG. 1. SEM micrograph of a porous sample. The pore sizes vary between
2
to 5 ␮m and the their depth is about 3 ␮m ͑which is not shown in the
a͒Electronic mail: raissi@kntu.ac.ir
figure͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 71.185.28.15
003-6951/2005/87͑16͒/164101/3/$22.50 87, 164101-1 © 2005 American Institute of Physics
On: Sun, 27 Apr 2014 13:38:26
0

164101-2
F. Raissi and R. Farivar
Appl. Phys. Lett. 87, 164101 ͑2005͒
curve. It saturates at about 500 ppm after which no change in
breakdown voltage is exhibited.
When doing measurements, the samples are placed up-
side down attached to the top cover of the gas enclosure. It is
possible to obtain larger sensitivity to gas with a funnel-type
top cover, in which the sample is placed at the tip. For ex-
ample, for the funnel shape top cover, the change in break-
down voltage due to 10 ppm of hydrogen is equal to 100
ppm of hydrogen in the regular gas enclosure. This is
because the effective hydrogen concentration is increased at
the sample surface as hydrogen replaces more air molecules
in this case.
FIG. 2. I-V curve of a sample in 15% humid air with no hydrogen and in the
presence of 500 ppm of hydrogen at room temperature. I-V curve moves to
larger voltages as hydrogen is introduced. It reaches its maximum at 500
ppm of hydrogen, after that it remains unchanged.
The response and recovery times were estimated to be
around 6 and 60 s, respectively, which are quite fast and are
due to the physical nature of the detection mechanism. The
longer recovery time may be due to hydrogen adsorption,
which take longer to be released. The same is observed for
1
4
walls. Connections were made to the PtSi layer and to the
substrate. For the tests which are carried out in air, we use
two closed gas chambers. The samples are placed inside one
cylindrical chamber which has an exact volume of 30
other gases.
The experiments were also carried at lower temperatures
down to 273 K and hydrogen gas sensitivity was exhibited in
a similar fashion. The change in the I-V curve was similar to
Fig. 2. Temperatures lower than 273 K were not tried be-
cause water vapor condenses on the sample and shields it
from hydrogen.
3
3
ϫ10 cm . Normal air is let inside the chamber to atmo-
spheric pressure. The air humidity is checked by a standard
humidity sensor, which for these experiments was 15% ͑we
have not introduced water vapor to change the humidity in
these measurements͒. This chamber is then isolated. A sec-
As was mentioned, it is very important to check hydro-
gen sensing in the presence of other reducing gases; there-
fore, experiments were carried out in the presence of differ-
ent concentrations of acetone, methanol, and ethanol up to
2000 ppm. The introduction of these gases did not have any
affect on the hydrogen gas sensing behavior. In other words,
these samples are able to sense hydrogen even in the pres-
ence of a large concentration of acetone, methanol, or etha-
nol. It seems that acetone, methanol, and ethanol cannot re-
place the gas content inside the pores perhaps because of
their large size and mass. As a result, they do not affect the
I-V curve and do not interfere with hydrogen gas sensing.
Durability of these samples was tested by biasing them
below their breakdown voltage and exposing them to hydro-
gen gas periodically. The samples have not shown any drift
in I-V curve for a period of 3 months. A current limit of 5
mA was set on the biasing circuit. Larger currents resulted in
a drift in I-V curve and caused some samples to be burnt.
We have demonstrated here that n-type PtSi/porous Si
Schottky junctions are capable of detecting hydrogen gas at
room temperature. Detection occurs in air and in presence of
hydrocarbons, such as acetone, methanol, and ethanol. The
detection mechanism is physical in nature, in which hydro-
gen replaces ambient gas inside the pores and this changes
the electric fringing fields. This results in a relatively fast
response and recovery times. The physical nature of detec-
tion mechanism also makes low-temperature operation pos-
sible. The detector is stable, robust, compact, easy to fabri-
cate, and is compatible with integrated circuit fabrication
technology.
3
ond smaller chamber with a volume of 150 cm is first con-
nected to a turbomolecular pump and its pressure is brought
−
5
down to 10 Pa. Then selected gases are let inside the cham-
ber to a predetermined pressure. This chamber is then dis-
connected from the turbopump and is connected to the larger
chamber in which the sample is placed. The relative concen-
tration is then calculated considering the volume of the two
chambers and their pressure before connection.
The reverse bias I-V curve of this device at 300 K in air
and in presence of 500 ppm hydrogen gas is provided in Fig.
2. As observed, I-V curve moves to larger voltages as hydro-
gen is introduced. In other words, the breakdown voltage
increases with hydrogen gas content. The change in the
breakdown voltage with hydrogen concentration is provided
in Fig. 3. We estimate an error of about 4 ppm in our mea-
surements. At 10 ppm of hydrogen concentration, the break-
down voltage increases by about 0.02 V. The change in
breakdown voltage follows more or less a piecewise linear
1
X. Bevenot, A. Trouillet, C. Veillas, H. Gagnaire, and M. Clementl, Sens.
Actuators B 67, 57 ͑2000͒.
W. Zhu, O. K. Tan, Q. Yan, and J. T. Ohl, Sens. Actuators B 65, 366
2
3͑
2000͒.
Y. D. Wang, C. L. Ma, X. H. Wu, X. A. Sun, and H. D. Li, Sens. Actuators
B 85, 270 ͑2002͒.
4
5
6
K. Hara and N. Nishida, Sens. Actuators B 20, 181 ͑1994͒.
V. N. Mishra and R. P. Agrawal, Sens. Actuators B 21, 209 ͑1994͒.
V. A. Chaudhary, I. S. Mulla, and K. Vijayamohanan, Sens. Actuators B
FIG. 3. The change in breakdown voltage as hydrogen concentration is
increased from zero. The relation can be estimated as a piecewise linear
curve which saturates at 500 ppm ͑not shown here͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 71.185.28.15
55, 366 ͑2000͒.
On: Sun, 27 Apr 2014 13:38:26

164101-3
F. Raissi and R. Farivar
Appl. Phys. Lett. 87, 164101 ͑2005͒
7
12
13
S. Basu and A. Dutta, Sens. Actuators B 22, 83 ͑1994͒.
V. R. Katti, A. K. Debnath, S. C. Gadkari, S. K. Gupta, and V. C. Sahni,
Sens. Actuators B 84, 219 ͑2002͒.
T. Weh, M. Fleischer, and H. Meixner, Sens. Actuators B 68, 146 ͑2000͒.
A. Katsuki and K. Fukui, Sens. Actuators B 52, 30 ͑1998͒.
F. Raissi, S. Mirzakuchaki, H. M. Jalili, and A Erfanian, IEEE Sens. J. ͑to
be published͒.
8
1
4
9
K. Domansky, A. Rose, W. H. Grover, and G. J. Exarhos, Mater. Sci. Eng.,
B 76, 116 ͑2000͒.
15
16
F. Raissi and M. Mohtashami Far, IEEE Sens. J. 2, 476 ͑2002͒.
F. Raissi, M. S. Abrishamian, and T. Emadi, IEEE Trans. Electron Devices
51, 339 ͑2004͒.
1
0
D. L. Hetherington, R. W. Grant, K. L. Hughes, W. T. Corbett, R. C.
Hughes, J. L. Rodriguez, and P. J. McWhorter, Proc. IEEE 1, 199 ͑1994͒.
F. H. Babaei and M. Orvatinia, IEEE Sens. J. 4, 802 ͑2004͒.
11
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP: 71.185.28.15
On: Sun, 27 Apr 2014 13:38:26