i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
Available online at www.sciencedirect.com
ScienceDirect
journal homepage: www.elsevier.com/locate/he
Extraordinary room-temperature hydrogen
sensing capabilities of porous bulk PteTiO2
nanocomposite ceramics
Wan Ping Chen a,*, Yao Xiong a, Ye Sheng Li b, Ping Cui c, Shi Shang Guo a,
Wei Chen c, Zi Long Tang b, Zijie Yan d, Zhenyu Zhang c
a
School of Physics and Technology, Wuhan University, Wuhan, Hubei 430072, China
School of Materials Science and Engineering, State Key Laboratory of New Ceramics and Fine Processing, Tsinghua
University, Beijing 100084, China
c
International Center for Quantum Design of Functional Materials (ICQD), Hefei Laboratory for Physical Sciences at
Microscale (HFNL), and Synergetic Innovation Center of National Quantum Information and Quantum Physics,
University of Science and Technology of China, Hefei, Anhui 230026, China
d
Department of Chemical and Biomolecular Engineering, Clarkson University, Potsdam, NY 13699, United States
b
article info
abstract
Article history:
Metal oxide semiconductors (MOSs) are among the most promising materials for sensing
Received 31 October 2015
hydrogen in distributed use, yet prevalent hydrogen sensors based on porous thick-film
Received in revised form
MOSs have the standing and serious drawback of demanding high working tempera-
22 December 2015
tures. Here we report on the surprising experimental observation of porous bulk PteTiO2
Accepted 23 December 2015
nanocomposite ceramics that exhibit ultrahigh hydrogen sensitivities of the order of 104
Available online 14 January 2016
for 1000 ppm H2 in N2 at room temperature, with fast response and short recovery times of
10 s and 20 s, respectively. We also demonstrate that hydrogen diffuses in the bulk ce-
Keywords:
ramics with an unusually high diffusion coefficient, and a novel hydrogen sensing
Hydrogen
mechanism rooted in hydrogen chemisorption on TiO2 is established. The central findings
Sensing
help to enrich our understanding of hydrogen interaction with MOSs, and are also expected
Metal oxide
to be instrumental in developing related technologies with hydrogen as the energy carrier.
Semiconductors
Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights
reserved.
Catalysis
Introduction
Hydrogen is expected to have immense potentials in future
fuel cells and hydrogen-powered automobiles, which will
unavoidably face mounting safety concerns [1]. Among the
several different classes of hydrogen sensors already available
on the market, the one based on metal oxide semiconductors
(MOSs) is especially attractive for sensing hydrogen in
distributed use, as it is relatively compact, durable, and simple
in operation [2]. Unfortunately, prevalent commercial MOS
hydrogen sensors such as those based on porous thick-film
SnO2 all have to work at temperatures (300e700 C) much
higher than room temperature [3], which naturally requires
high power consumption, and also results in undesirable high
cross-sensitivity
and
decreased
lifetime.
Various
* Corresponding author.
E-mail address: [email protected] (W.P. Chen).
http://dx.doi.org/10.1016/j.ijhydene.2015.12.151
0360-3199/Copyright © 2015, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.
3308
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
nanostructured MOSs with large surface-to-volume ratios
have been successfully synthesized in the past decades,
which have been found especially promising for photocatalysis and low-temperature gas sensing [4e7]. While lowdimensional nanostructured MOSs can be directly applied
for photocatalysis [8,9], for gas sensing nanostructured MOSs
have to be connected to electric circuits. For this reason onedimensional (1D) MOS nanomaterials are relatively suitable
and have been fabricated as MOS hydrogen sensors with
impressive low-temperature hydrogen sensing capabilities
[10,11]. Nevertheless, such materials still encounter the
standing shortcomings of difficulties in forming lowresistance contacts, low mechanical and thermal robustness
[12], and strong adsorption of other unwanted gases in surrounding atmosphere [13,14], hindering their potential practical gas-sensing applications. So for gas-sensing, MOSs in
bulk forms, such as in porous thick films, are highly preferred
in numerous circumstances. It is strange that bulk MOSs have
been rarely investigated for room-temperature hydrogen
sensing applications up to date, and few systematic investigations have been conducted on room-temperature reactions between hydrogen and MOSs, either. A clear
mechanistic understanding for room-temperature hydrogen
sensing of MOSs is thus still lacking, in many cases with
mutually contradictory assumptions [2,6,15].
Herein, we present some striking findings on roomtemperature hydrogen sensing capabilities of bulk MOSs. We
show that a porous bulk PteTiO2 nanocomposite ceramic
material, readily fabricated from TiO2 and Pt nanoparticles
through pressing and sintering, exhibits extraordinary roomtemperature hydrogen sensitivities of the order of 104 for
1000 ppm H2 in N2, with fast response and short recovery
times of 10 s and 20 s, respectively. We further reveal a fast
diffusion of hydrogen at room temperature in the porous bulk
ceramics, and propose a hydrogen sensing mechanism rooted
in chemisorption of hydrogen on TiO2. These central findings
help to enrich our understanding of hydrogen interaction with
MOSs, and are also expected to be instrumental in developing
related technologies with hydrogen as the energy carrier.
Experimental methods
Material preparation
Commercial TiO2 nanoparticles, P25 (ca. 80% anatase and 20%
rutile), purchased from the Degussa Co., were used as the
starting material. The TiO2 powder was dispersed in deionized
water and magnetically stirred. Pt powder (particle sizes of
~10 nm) was added to the suspended system according to a
weight ratio of 5:95 for Pt:TiO2. After stirring for 2 h, the system was dried at 110 C in oven. The dried powder was then
pressed into cylindrical pellets (~1.7 mm in thickness and
~12 mm in diameter) by applying about 10 MPa pressure
through a hydraulic press. Pellets of pure P25 were also made
for comparing purpose.
These pellets were then heat treated in air for 2 h at a series
of temperatures, from 250 to 950 C with a step of 100 C. Due
to the large specific surface area of the nanoparticles, a
sintering-induced decrease in diameter was detected when
the pellets were treated at temperatures as low as 450 C.
Through DC magnetron sputtering, a pair of rectangular Pt
electrodes were coated on the major surfaces of those pellets
intended for hydrogen-sensing measurement.
Hydrogen sensing measurement
During the sensing measurements, a pellet was placed in a
sealed chamber (of volume ~350 ml) with a gas inlet and
outlet. In the response process, N2 was used as the carrier gas,
and H2 was mixed with N2 in appropriate ratios to realize
desirable hydrogen concentrations. In the recovery process,
air gas was pumped into the chamber. The total gas flow rate
was kept at about 300 ml/min. A 15 V DC voltage was applied
between the two electrodes and the electric current was
measured to calculate the real-time resistance response of the
sensor to hydrogen. All the data acquisition was carried out
automatically through a computer.
Results and discussion
The transition from anatase to rutile was observed in the
pellets heat-treated at and above 650 C, while such transition
was not observed in those pellets heat-treated at and below
550 C [16]. We have thus chosen 550 C as the sintering
temperature for those pellets used in further hydrogensensing investigation, at which there was a 3.5% decrease in
diameter after sintering. Fig. 1a shows a schematic of a sensor
and a scanning electron microscopy (SEM) image of the
PteTiO2 ceramics sintered at 550 C. TiO2 grains are ~30 nm in
diameter and numerous nanosized pores are present. According to energy dispersive spectroscopy (EDS) analysis, the
Pt clusters mostly exist as aggregates of a few micrometers in
size and ~10 mm apart from one another, with a typical
aggregate partially revealed in the bottom-right corner of
Fig. 1a. Obviously, porous bulk PteTiO2 nanocomposite ceramics have thus been prepared through convenient pressing
and sintering.
The sintered pellets of pure P25 show no detectable
response to hydrogen at room temperature. In contrast,
PteTiO2 nanocomposite ceramics exhibit surprising and
extraordinarily strong room-temperature hydrogen sensing
behaviors as shown in Fig. 1b. Upon exposure to 1000 ppm H2
in N2, the resistance of the ceramics quickly decreases by a
factor of over 6000 at room temperature, indicating a sensitivity of over 6000 in terms of Ra/Rg (where Ra and Rg are the
electrical resistances of the sensor in clean air and in the
measuring gas, respectively). Previously, such high hydrogen
sensitivities at room temperature could only be achieved
using much less robust MOS nanostructures, such as networks of Pd-decorated SnO2 nanowires [10]. Considering that
prevalent commercial SnO2 thick-film hydrogen sensors have
to work at elevated temperatures, the ultrahigh roomtemperature sensitivities of the PteTiO2 ceramics demonstrated here represent a breakthrough discovery in bulk MOSbased hydrogen sensing.
Another vitally important characteristic of a sensor is its
response and recovery times. As shown in Fig. 1b, the ceramics exhibit a 90% response time of ~10 s and a
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
3309
Fig. 1 e (a) A SEM image taken from a fractured surface of a PteTiO2 ceramic pellet. A schematic of a sensor is shown in the
top-left inset, where two Pt electrodes are coated on the same surface of the pellet to measure its resistance change upon
hydrogen exposure. (b) Real-time electrical resistance response to 1000 ppm H2 in N2 for five successive cycles at room
temperature. (c) Electrical resistance response to hydrogen at different hydrogen concentrations at room temperature.
(d) Sensitivity dependence on hydrogen concentration.
comparatively short 90% recovery time of ~20 s. In contrast,
the 1D/quasi-1D nanostructure-based sensors that exhibit
fast response will often encounter the common drawback of
too slow room-temperature recovery [17,18]. Furthermore, the
overall response and recovery behaviors shown in Fig. 1b have
little variations in five successive sensing cycles, indicating
high repeatability in hydrogen sensing.
The ceramics also show excellent sensing capability over a
large range of hydrogen concentrations. As shown in Fig. 1c
and d, the sensitivity decreases with decreasing hydrogen
concentration, as expected; yet even at 30 ppm H2 in N2, the
sensitivity is still ~200. This is still an extraordinarily high
sensitivity at such a low hydrogen concentration.
Aside from ultrahigh sensitivities, the PteTiO2 composite
ceramics also possess high hydrogen sensing selectivity,
shown by the lack of any visible responses upon in-gassing of
CO and NO2 at room temperature. This high selectivity is tied
to the drastically different catalytic reactions of the Pt aggregates with H2, CO, and NO2 [11].
So far we have demonstrated that those PteTiO2 composite
ceramic materials are ideal for hydrogen sensing in many
important aspects, including the readily accessible and stable
bulk form, ultrahigh room-temperature sensitivities, fast
response and short recovery times, high repeatability, and
high gas selectivity. Next we turn our attention to the likely
underlying room-temperature hydrogen sensing mechanisms, which for bulk ceramic materials have so far been
largely overlooked. We first designed a special experiment to
study the diffusion of hydrogen in the bulk PteTiO2 ceramics,
in which a pair of Pt electrodes were coated on the two
opposite major surfaces of a sufficiently thick (~1.7 mm) sintered pellet, with its side surface sealed by epoxy as illustrated
in Fig. 2a. This configuration was adopted to avoid possible
formation of a short circuit on the surface layer between the
electrodes, to ensure that the electrical resistance change
must reflect the evolution of the bulk property during
hydrogen sensing. A representative hydrogen absorption and
sensing curve is shown in Fig. 2b. Compared with the results
shown in Fig. 1b, this configuration again yields ultrahigh
sensitivity, but its response to hydrogen clearly indicates
multiple electronic/ionic processes, reflected by the two
stages with distinctly different slopes. As the decrease in
resistance is primarily related to some hydrogen-induced reactions, the ultrahigh sensitivity demonstrates that a lowresistance path has been established along the thickness direction of the pellet when exposed to hydrogen, indicating
that hydrogen is accessible to the very interior of the pellet at
room temperature.
The slope of the first response stage is much lower than
that of the second stage, and the much faster second stage
resembles well the monotonous behavior shown in the first
configuration (see Fig. 1). So the first stage should correspond
to the hydrogen diffusion process over the relatively long
distance along the thickness direction of the sample in the
3310
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
Fig. 2 e (a) Schematic of a PteTiO2 nanocomposite ceramic
pellet with two electrodes coated on its two opposite major
surfaces, and the side surface of the pellet was sealed by
epoxy. (b) Real-time electrical resistance response to
1000 ppm H2 in N2 for three successive cycles (bottom) at
room temperature, with a typical response upon hydrogen
exposure magnified (top).
second configuration. The time when the measured resistance
begins to decrease upon hydrogen exposure is taken as t ¼ 0.
During hydrogen diffusion along the thickness (~1.7 mm) direction of the pellet, hydrogenated layers grow from each of
the two major surfaces inwards. Since the presence of
hydrogen leads to a dramatic decrease in the resistivity of
TiO2, the measured resistance in the first stage of response in
Fig. 2 is dominated by that of the central un-reacted layer,
which is proportional to the thickness of (1.7 2x). At any
given time t, when the measured resistance is R, the corresponding
x
value
is
calculated
according
to
R=R0 ¼ ð1:7 2xÞ=1:7, where R0 is the resistance measured in
air. The diffusion coefficient D is then calculated using the
pffiffiffiffiffiffi
formula of diffusion length x ¼ 2 Dt, as shown in Fig. 3. The
transition from the first to the second stage corresponds to the
Fig. 3 e Plot of x (thickness of the hydrogenated layer from
one side of the pellet) vs t1/2.
effective merging of the two reaction fronts around the center
of the pellet. Based on this picture, we estimate the effective
room-temperature diffusion coefficient of hydrogen in the
ceramics to be 4.2 105 cm2/s, which is rather high for typical
bulk diffusion processes in oxide materials and should be
responsible for the observed fast response and short recovery
times of the bulk materials. As a matter of fact, for pure P25
pellets sintered at 550 C, the presence of numerous cylindrical mesopores has been revealed, and the BrunauereEmmetteTeller (BET) specific surface area of the
ceramics is calculated to be around 23 m2/g [16]. Such a BET
specific surface area is close to one half of that of as-received
P25. So the bulk ceramics prepared from TiO2 nanoparticles
through convenient pressing and sintering is highly porous
with cylindrical mesopores, which are ideal for gas-sensing
applications via molecular diffusion and explain the
observed high bulk diffusion coefficient of hydrogen as
detailed above.
Given that the sensitivities of those commercial hydrogen
sensors based-on thick-film MOSs have remained relatively
low up to date, it seems a little surprising that such ultrahigh
hydrogen sensitivities can be readily obtained for PteTiO2
nanocomposite ceramics in this study. Commercial MOS
thick-film hydrogen sensors are usually fabricated through
drop-coating, tape-casting or screen-printing, in which thickfilms are formed from pastes of active material, namely MOSs,
a low softening temperature glass frit, and organic carrier [2].
In a subsequent firing, organic carrier is burnt out while glass
frit forms a suitable matrix for the active material. As the glass
is inert to target gas, it will seriously decrease the overall
sensitivities of thick-films to the target gas due to a series
connection between glass frit and MOSs. In this study, glass is
not present and the distance between the two electrodes can
be connected continuously with TiO2 nanograins whose
resistance is sensitive to hydrogen. Compared with commercial thick-film MOSs hydrogen sensors, the nanoceramics in
this study not only have relatively high specific surface area as
discussed above, the negative influence of glass on the sensitivities have also been avoided. So ultrahigh hydrogen sensitivities at room temperature have been observed for them.
For hydrogen interaction and sensing of MOSs at elevated
temperatures, the most widely accepted mechanism involves
the chemisorption of oxygen on MOSs [2], which leads to the
formation of electron depleted layers beneath the surface of
MOSs. When hydrogen, and other reducing gases as well, reacts with such chemisorbed oxygen, the thickness of the
depleted layers will decrease, and the resistance of the MOSs
will decrease accordingly. To verify whether this prevailing
mechanism is valid for the room-temperature hydrogen absorption and sensing behaviors observed in this study, we
have conducted some special measurements as shown in
Fig. 4. The ceramics display relatively high and steady conductivity in a mixed atmosphere of 1000 ppm H2 in N2. When
the hydrogen flow was stopped, the surrounding atmosphere
was changed to N2 and the resistance was found to increase
slowly in N2. And when the atmosphere of 1000 ppm H2 in N2
was restored, the initial high conductivity corresponding to
the mixed atmosphere was promptly restored. As O2 was absent in this process, the drastic increase/decrease in the
resistance could hardly be explained in terms of oxygen
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
3311
nanoparticles with their electrons denoted [24]. Expressing a
molecular hydrogen as H2 ðgÞ, and a chemisorbed hydrogen ion
on TiO2 as H$i , the overall reaction in the response stage can be
described as:
Pt;TiO2
H2 ðgÞ!2H$i þ 2e0 ;
(1)
0
Fig. 4 e Real-time electrical resistance response to different
kinds of atmospheres at room temperature for a PteTiO2
nanocomposite ceramic pellet. The Pt electrodes were
coated on the same major surface of the pellet.
chemisorption. As a matter of fact, many studies have shown
that for oxygen chemisorption on MOSs, the leading oxygen
species is O which starts to be dominant at temperatures
above 150e220 C [19,20]. So for the room-temperature
hydrogen absorption and sensing behaviors observed in this
study, a mechanism rooted in hydrogen chemisorption on
TiO2 is highly favored: First, molecular hydrogen can easily
penetrate into the bulk of the porous ceramic materials.
Subsequently, the Pt aggregates in the ceramics are effective
in decomposing the incoming molecular hydrogen [21]. In the
third step, such decomposed hydrogen atoms can efficiently
migrate into the surfaces or PteTiO2 interfaces of the TiO2
nanoparticles via a spill-over process [22,23]. Lastly and most
crucially, the hydrogen atoms become chemisorbed on TiO2
where e represents a released electron contributing to the
conductivity.
To further understand hydrogen chemisorption on the
TiO2 porous ceramics, we have also tried to produce such
hydrogen chemisorption in an electrochemical way [25], in
which a sintered P25 pellet was used as the cathode to electrolyze water in a NaOH solution. To our surprise, the chemisorbed hydrogen was found to be so reactive that such
hydrogen atoms effectively reduce Cu2þ in aqueous solutions
to Cu2O deposited on TiO2 at room temperature, as shown in
Fig. 5. With such a high reactivity, the increase/decrease in
resistance in Fig. 4 now can be satisfactorily understood: In N2,
chemisorbed hydrogen atoms will combine with one another
and leave from TiO2 in the form of molecules, resulting in the
relatively slow resistance increase; when the surrounding
atmosphere is in air, chemisorbed hydrogen will react with
oxygen more intensively [26], leading to the fast recovery.
Conclusion
We have achieved some novel and surprising findings on the
room-temperature hydrogen absorption and sensing behaviors of properly functionalized bulk porous PteTiO2 nanocomposite ceramics. Intriguing aspects included fast effective
hydrogen diffusivity, ultrahigh hydrogen sensitivity, fast
response and short recovery times, and high gas selectivity.
Collectively, these results point to a novel hydrogen sensing
mechanism rooted in hydrogen chemisorption on TiO2. These
findings help to enrich our microscopic understanding of
room-temperature hydrogen interactions with different
MOSs, and should also have important application potentials
in various technological developments with hydrogen as the
energy carrier in view of the stability, readiness for mass
production, and suitability in everyday-life circumstances
inherent with bulk MOSs.
Acknowledgments
We thank Mr. Peng Lu for his help on hydrogen sensing
measurement. This work was supported by the National
Natural Science Foundation of China (Nos. 61434002, J1210061,
and 11204286), the National High Technology Research and
Development Program of China (No. 2013AA031903), and the
National Key Basic Research Program of China (No.
2014CB921103).
Fig. 5 e X-ray diffraction pattern taken from the surface of a
P25 ceramic pellet, which had been electrochemically
hydrogenated for 20 h and then immersed in a CuSO4
solution for 10 min.
references
[1] Moy R. Liability and the hydrogen economy. Science
2003;301:47.
3312
i n t e r n a t i o n a l j o u r n a l o f h y d r o g e n e n e r g y 4 1 ( 2 0 1 6 ) 3 3 0 7 e3 3 1 2
[2] Korotcenkov G. Handbook of gas sensor materials:
Properties, advantages and shortcomings for applications.
New York: Springer Press; 2013.
[3] Boon-Brett L, Bousek J, Black G, Moretto P, Castello P,
Hubert T, et al. Identifying performance gaps in hydrogen
safety sensor technology for automotive and stationary
applications. Int J Hydrogen Energy 2010;35:373.
[4] Liu M, Piao LY, Wang WJ. Hierarchical TiO2 spheres: facile
fabrication and enhanced photocatalysis. Rare Met
2011;30:153.
[5] Liu M, Sunada K, Hashimoto K, Miyauchi M. Visiblelight sensitive Cu(II)eTiO2 with sustained antiviral activity for efficient indoor environmental remediation.
J Mater Chem A 2015;3:17312.
[6] Varghese OK, Gong DW, Paulose M, Ong KG, Dickey EC,
Grimes CA. Extreme changes in the electrical resistance
of titania nanotubes with hydrogen exposure. Adv Mater
2003;15:624.
[7] Wang YM, Du GJ, Liu H, Liu D, Qin SB, Wang N, et al.
Nanostructured sheets of TieO nanobelts for gas
sensing and antibacterial applications. Adv Funct Mater
2008;18:1131.
[8] Liu M, Li HM, Zeng YS, Huang TC. Anatase TiO2 single
crystals with dominant {0 0 1} facets: Facile fabrication from
Ti powders and enhanced photocatalytical activity. Appl Surf
Sci 2013;274:117.
[9] Liu M, Zhong M, Li HM, Piao LY, Wang WJ. Facile synthesis of
hollow TiO2 single nanocrystals with improved
photocatalytic and photoelectrochemical activities. Chem
Plus Chem 2015;80:688.
[10] Lee JM, Park J, Kim S, Kim S, Lee EY, Kim SJ, et al. Ultrasensitive hydrogen gas sensors based on Pd-decorated tin
dioxide nanostructures: Room temperature operating
sensors. Int J Hydrogen Energy 2010;35:12568.
[11] Zou XM, Wang JL, Liu XQ, Wang CL, Jiang Y, Wang Y, et al.
Rational design of sub-parts per million specific gas sensors
array based on metal nanoparticles decorated nanowire
enhancement-mode transistors. Nano Lett 2013;13:3287.
[12] Varghese OK, Gong D, Paulose M, Grimes CA, Dickey EC.
Crystallization and high-temperature structural stability of
titanium oxide nanotube arrays. J Mater Res 2003;18:156.
s N,
[13] Pavelko RG, Vasiliev AA, Llobet E, Vilanova X, Barrabe
Medina F, et al. Comparative study of nanocrystalline SnO2
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
materials for gas sensor application: thermal stability and
catalytic activity. Sens Actuators B 2009;137:637.
Bahu M, Kumar K, Bahu T. CuOeZnO semiconductor gas
sensors for ammonia at room temperatures. J Electron
Devices 2012;14:1137.
Chen KS, Xie K, Feng XR, Wang SF, Hu R, Gu HS, et al. An
excellent room-temperature hydrogen sensor based on
titania nanotube-arrays. Int J Hydrogen Energy
2012;37:13602.
Xiong Y, Tang ZL, Wang Y, Hu YM, Gu HS, Li YZ, et al. Gas
sensing capabilities of TiO2 porous nanoceramics prepared
through premature sintering. J Adv Ceram 2015;4:152.
Shen YB, Yamazaki T, Liu Z, Meng D, Kikuta T, Nakatani N,
Saito M, Mori M. Microstructure and H2 gas sensing
properties of undoped and Pd-doped SnO2 nanowires. Sens
Actuators B 2009;135:524.
Tien LC, Sadik PW, Norton DP, Voss LF, Pearton SJ,
Wang HT, et al. Hydrogen sensing at room temperature
with Pt-coated ZnO thin films and nanorods. Appl Phys Lett
2005;87:222106.
Barsan N, Weimar U. Conduction model of metal oxide gas
sensors. J Electroceram 2001;7:143.
Batzill M. Surface science studies of gas sensing materials:
SnO2. Sensors 2006;6:1345.
Herrmann JM, Pichat P. Metal-support interactions: An in
situ electrical conductivity study of Pt/TiO2 catalysts. J Catal
1982;78:425.
Roland U, Salzer R, Braunschweig T, Roessner F, Winkler H.
Investigations on hydrogen spillover. Part 1.eelectrical
conductivity studies on titanium dioxide. J Chem Soc
Faraday Trans 1995;91:1091.
Lin Y, Ding F, Yakobson BI. Hydrogen storage by spillover on
graphene as a phase nucleation process. Phys Rev B
2008;78:041402(R).
Chen WP, Wang Y, Chan HLW. Hydrogen: A metastable
donor in TiO2 single crystals. Appl Phys Lett 2008;92:112907.
Chen WP, He KF, Wang Y, Chan HLW, Yan Z. Highly mobile
and reactive state of hydrogen in metal oxide
semiconductors at room temperature. Sci Rep 2013;3:3149.
Kong J, Chapline MG, Dai H. Functionalized carbon
nanotubes for molecular hydrogen sensors. Adv Mater.
2001;13:1384.