COVER SHEET
Tesfamichael, T. and Motta, N. and Bostrom, T. and Bell, J. M. (2006) Development of Porous Metal
Oxide Thin Films by Co-evaporation. Applied Surface Science 253(11):pp. 4853-4859.
Accessed from http://eprints.qut.edu.au
Copyright 2006 Elsevier
Development of Porous Metal Oxide Thin Films by Co-evaporation
a
*T. Tesfamichael, aNunzio Motta, bThor Bostrom, and cJ. M. Bell
a
Faculty of Built Environment and Engineering, School of Engineering Systems,
b
Faculty of Science, Analytical Electron Microscopy Facility
c
Centre for Built Environment and Engineering Research
Queensland University of Technology, 2 George Street, Brisbane, QLD 4000,
Australia
Fax: +61 7 38641516
*Corresponding author: Dr Tuquabo Tesfamichael, Phone: +61 7 38641988,
[email protected]
Abstract
This paper focuses on the development of mixed metal oxide thin films and physical
characterization of the films. The films were produced by co-evaporation of titanium
oxide and tungsten oxide powders. This allowed the development of titanium oxidetungsten oxide films as analyzed using XPS. Examination in the SEM and AFM
showed that the films were nano-porous with the pore size and pore orientation
varying as a function of the deposition angle. UV-Vis spectra of the films show an
increase of transmittance with increasing deposition angle which is attributed to the
structure and porosity of the films. Raman analysis indicated that the as-deposited
films have broad and weak Raman characteristics, attributed to the nanocrystal
nature of the films and the presence of defects, and the peak broadening deceases
after annealing the film, as expected.
PACS: 81.05.Rm; 61.43.Gt; 78.66.Qn; 68.55.Jk; 07.07.Df; 61.72.–y
1
Keywords: metal oxide gas sensors; oblique deposition; co-evaporated TiO2-WO3
thin film; surface characterization; porosity
1. Introduction
To protect people and property from gas-related accidents, gas sensors with higher
detecting capacity and gas selectivity are needed. A common sensing mechanism
involves measuring an electrical conductance change caused by gas adsorption.
This requires a sensitive layer that converts a change in the chemical state of some
species into an electrical signal. Metal oxide materials such as TiO2, SnO2, ZnO and
WO3 have been used for gas-sensing devices because of their simplicity and low
cost [1-9]. However, they have poor selectivity and are affected by humidity and thus
commonly operate at temperatures between 200-400 oC. The gas sensing properties
of these oxides are determined by their intrinsic properties, but can be modified by
addition of impurities and/or modifying the microstructure including particle size,
distribution and orientation, surface morphology
and porosity [6; 8; 10-13]. The
materials have band gap energies between 2.6 to 3.3eV at room temperature. They
behave like n-type semiconducting oxides due to non-stoichiometry and they show
good chemo-sensitivity toward reducing gas pollutants. Among these oxides SnO2 is
one of the most widely used sensors but recent research shows the other metal
oxides are able to detect pollutants and toxic gas emissions such as ammonia,
carbon monoxide, hydrogen sulfide, nitrogen dioxide. The response behavior of
various metal oxide films to a specific gas suggests that oxygen vacancy and charge
transport mechanisms are dependent on film microstructure [14]. Thin films are
usually compact and the active layer is limited to the surface whereas thick films are
commonly porous and hence the whole layer can interact with the gas species. If a
controlled porosity can be achieved in thin films, then the gas sensing properties can
2
be enhanced. Gas detecting sensitivity also depends on the reactivity of film surface
as sensors are strongly influenced by the presence of oxidizing or reducing gases on
the surface. The reactivity can be enhanced by impurities, defects and active species
on the surface of the films, increasing the adsorption of gas species. It has been
shown that inclusion of different doping metals in the oxide films increased the
sensitivity to specific gases [3; 10-12; 15-17].
Gas sensitivity and selectivity can also be enhanced by appropriate choice of
material and proper control of the film microstructure. Each material has its own
response and co-deposition of various materials can add selectivity to specific gas
species and also often improves sensor quality and performance [8; 18]. Various
techniques have been used to develop films for gas sensor applications. This
includes sol-gel, chemical vapor deposition (CVD) and physical vapor deposition
(PVD) [19-24]. Each of the film deposition techniques has its own advantages and
limitations. TiO2-WO3 thin film sensors have been produced by sputtering and sol-gel
techniques [18; 25]. In this work, we expand the use of metal oxide sensors by coevaporation of titanium oxide and tungsten oxide producing porous metal oxide thin
films. The experiment investigated the development of titanium oxide-tungsten oxide
films by adjusting deposition angle to optimize the microstructure of the thin film to
achieve high performance gas sensors. Deposition onto a substrate held at an angle
to the incoming atom or ion flux is known to create columnar structures which often
exhibit higher porosity and rougher surfaces than films deposited with the ion or atom
flux normal to the substrate [26]. The combination of co-deposition to provide doping
and off-normal deposition to change the porosity and surface roughness should
provide real opportunities to improve sensitivity and selectivity.
3
Physical characterization of various films to determine film structure, composition,
crystallinity and optical properties have been performed. Atomic Force Microscopy
(AFM) and Scanning Electron Microscope (SEM) were used to study the surface
structure and morphology of the films. The chemical composition was investigated
using X-Ray Photoelectron Spectroscopy (XPS). The optical properties and
crystalline nature of the films have been characterized using UV-Vis spectroscopy
and Raman spectroscopy, respectively.
2. Experimental
2.1. Thin Film Preparation
The present work focuses on a development of nanoporous metal oxide thin film by
evaporation method. A conducting SnO2:F glass substrate (Libbey-Owens Ford TEC
8 – 8 Ω/square nominal resistance) was obtained from Libbey-Owens Ford (USA).
The substrate was cut in to 5.5x5.5 cm squares and cleaned using methylated spirit.
The substrate was attached to a variable-angle substrate holder and mounted 10 cm
above a heat resistive boat (see Figure 1). High purity commercial titanium oxide
(99.9%) and tungsten oxide (99.9%) pellets were obtained from Pure Tech Inc.
(USA). The pellets of equal proportion (by volume) were crushed to a fine powder,
mixed and put onto the boat in a chamber (Figure 1). The chamber was evacuated
by mechanical and diffusion pumps to a base pressure of about 1.5 x10-4 Torr. Coevaporation of the mixture was carried out at a partial pressure of about 10-2 Torr by
applying an electric current through the boat. Film thickness was controlled using
quartz thickness monitor and films of about 100 to 200 nm thick were obtained by
controlling the deposition rate and time. The samples analyzed in this paper have film
thickness of about 200 nm. Various samples were evaporated at different angles of
incidence between normal angle and 80 degree. One sample was annealed in air in
4
a temperature-substrate-specimen controlled oven at 450
o
C for 24 hours to
investigate the change in crystalline property of the film. The effect of annealing on
crystallinity and gas selectivity of the films will be studied systematically at a later
stage.
2.2. Film Characterization
The surface microstructure of the films deposited at various angles of incidence was
examined using an FEI Quanta 200 scanning electron microscope operated at 25kV.
The samples for SEM were coated with a very thin layer of gold in a sputter coater to
provide electrical conduction and thereby reduce any charging effects. AFM images
of the film surface were obtained using an NT-MDT Solver P47 scanning probe
microscope (NT-MDT Co., Moscow, Russia) with "Golden" Si cantilevers operated in
semi-contact (vibrating tip) mode. Optical properties of the film were measured using
Varian Cary 50 UV-Visible conventional spectrophotometer in the wavelength range
300 to 1100 nm. As a reference, zero and 100% baseline signals were displayed
before each measurement of transmittance. The composition of the film was
determined using an Kratos AXIS ULTRA multi-technique X-ray Photoelectron
Spectrometer (XPS) incorporating a 165 mm hemispherical electron energy analyser.
XPS measurements were made at the Brisbane Surface Analysis Facility at the
University of Queensland. Survey (wide) scans were taken at an analyser pass
energy of 160 eV. Survey scans were carried out over 1200-0 eV binding energy
range with 1.0 eV steps and a dwell time of 100 ms and atomic concentration of each
element in the film were determined from peak areas of the profile. Narrow multiplex
high resolution scans were run with 0.1 eV steps and 250 ms dwell time. Raman
measurements were carried out using Renishaw System-1000 Raman spectrometer
to determine the chemical structure and physical state of the film. A HeNe laser
5
excitation source of wavelength 633 nm with a 10 mW power at the sample was
used. A Raman shift between the wavenumbers 200 cm-1 to 1000 cm-1 has been
measured.
3. Results And Discussions
3.1 Scanning Electron Microscopy
Figure 2 shows SEM images of titanium oxide-tungsten oxide films deposited at
various angles of incidence. Figure 2a shows a film deposited at near normal angle
of incidence. A dense film with a range of submicron particle sizes from
approximately 100 nm to 500 nm can be observed.
Figures 2b and c show images of titanium oxide-tungsten oxide films deposited at
70° and 80° angle of incidence. From the images, it can be observed that at higher
angles of incidence nanoporous films have been obtained. The microstructures of
these films are considerably different from the film deposited at near normal angle of
incidence (Figure 2a). The particle size is similar, with some smaller crystallites
evident in the films deposited at 80° angle of incidence, but there is considerable
porosity evident at the surface of the film, with pores of about 20 nm to 100 nm in
size distributed across the film. From the results it can be concluded that the porosity
of the co-evaporated films was found to increase with increasing angle of deposition
and those films which possess nanopores are especially desirable for gas-sensor
design [10; 27; 28].
3.2 Atomic Force Microscopy
AFM images of four samples deposited between 30 to 80 degrees of incidence are
shown in Figure 3a-d. The porosity of the films was found to increase with increasing
deposition angle which is consistent with the SEM results discussed above. During
deposition of the films, the particles cluster together forming bigger grains with grain
6
boundaries clearly shown between particles and also between grains. From the
images, it is apparent that angular deposition at a higher angle can significantly
change the porosity and morphology of the film. The effect of surface morphology on
gas sensitivity has been reported elsewhere [13].
Using the Nova and Image Analysis software from NT-MDT, the average grain size
was calculated from the AFM scans and is shown in Table 1. Generally the grain size
remained similar as a function deposition angle with pronounced clustering of the
individual particles at lower angles. At the highest angle of deposition (80 degree) the
clustering of the particles was lower and as a result the average grain size was
reduced (Table 1). In order to improve the gas-sensing characteristic of a film,
optimizations of the grain size and porosity are important factors. Overall the
analyses have shown that a change in deposition angle leads to the change in film
microstructure.
3.3 X-Ray Photo-electron Spectroscopy
Figure 4a shows XPS spectra obtained from survey scans on the surface of the coevaporated film between binding energies 0 to 1200 eV. As expected elements of C,
O, Ti and W were detected. From the spectra, impurities of K and P were also
detected. From high resolution spectra (Figure 4b), the core level of Ti 2p1/2 and Ti
2p3/2 are found at binding energy of 465.6 eV and 459 eV, respectively with a peak
separation of 6.6 eV. Figure 4c shows the core level of W 4f5/2 (38.2 eV) and W 4f7/2
(36 eV) obtained from high resolution measurements and the two peaks are
separated by 2.2 eV. The Ti 2p3/2 peak obtained from the co-evaporated film (459
eV) is in close agreement with the literature value (459.2 eV) which corresponds to
the fully oxidized Ti+4 in titanium dioxide [29]. This shows that the Ti in the metal
oxide film can be assigned to the TiO2. The binding energy of W 4f7/2 obtained in this
7
work (36 eV) was found to be in reasonable agreement with the corresponding
characteristic binding energy of W 4f (35.8 eV) in tungsten trioxide film reported
elsewhere [5; 14; 24]. The peak position of O 1s was at 531 eV (see table 2) but this
value is shifted, for example by 1.1 eV when compared with the core level of O 1s
(529.9 eV) of the anatase TiO2 film reported elsewhere [30]. This can be due to the
nanocrystal nature of the mixed oxide thin film which may cause a shift of the core
levels.
From the peak area profile and the FWHM, the atomic concentrations of the
elements before and after etching the film using argon gas have been determined. As
shown in Table 2 and Figure 5, the amount of C decreased significantly after 5
minutes of etching which indicated that the C was not introduced during the
deposition process. However, the concentrations of K and P remained significant
after etching for up to 40 minutes. The source of these impurities was not known. The
slight shift of the core level peaks of Ti 2p and W 4f observed above may be
influenced by the K and P impurities in the nano-porous metal oxide film. The
concentrations of Ti and W across the depth of the film remained fairly constant with
about 8 at% and 19 at%, respectively. This is equivalent to a volume fraction of about
0.3 of TiO2 compared to WO3 in the film.
3.4 Raman spectroscopy
Raman spectroscopy was used to characterize the chemical and crystalline nature of
the film. As shown in Figure 6, the as-deposited film has broad and weak Raman
characteristic peaks around 645 and 696 cm-1. Raman peak broadening is
associated with various structural and microstructural features, including nano-size
crystals, presence of grain boundary and defects [31]. The 645 cm-1 peak position
8
corresponds to anatase titanium oxide Raman active mode Eg and the 696 cm-1 peak
corresponding to the vibrational frequency Raman mode of monoclinic tungsten
trioxide. From literature the expected vibrational frequencies of the most intense
Raman mode of anatase titanium oxide and monoclinic tungsten oxide are 639 cm-1
and 700 cm-1 respectively. The results for these films show a small shift from the
anticipated results [18; 32; 33]. This can be due to the nano-structure of the film that
can have a wide distribution of Ti-O and W-O bond angles which may cause a shift of
Raman peaks. In Figure 6, Raman spectrum of the film after annealing at 450°C for
24 hours is also shown. The intensity of the absorption bands at 645 and 696 cm-1
increased and the peak width deceases. This indicates a more crystalline film after
heat treatment. In addition a weak Raman peak appeared at about 320 cm−1 which is
most likely from the tungsten trioxide hexagonal phase, which is normally seen at
310 cm-1 [18].
3.5 Optical Spectrophotometer
Transmittance of various films deposited at different angles of incidence measured in
the wavelength range 300 to 1100 nm is shown in Figure 7. From the figure, the
transmittance increases significantly with increasing deposition angle over the
measured wavelength range. The most likely cause for the increase is the reduced
overall electron density as the porosity increases. Luminous transmittance of the
films were determined by weighting the transmittance of the films (Figure 7) in the
wavelength range 380 to 760 nm to the intensity of the corresponding visible
spectrum for air mass 1.5 as shown in Table 3. The films deposited at higher angle
are fairly transparent having a luminous transmittance of about 60% which is closer
to the transmittance value of the conducting glass substrate (65%). The anatase
phase of titanium oxide has a band gap energy of 3.2 eV whereas tungsten trioxide
9
has a band gap energy of 2.6 eV. Their corresponding optical transition wavelengths
for these oxides are 385 nm and 475 nm, respectively. The co-evaporated
nanocrystalline metal oxide films containing defects thus have not shown a sharp
absorption edge (see Figure 7).
4. Conclusions
Co-evaporated titanium oxide-tungsten oxide films with submicron grain sizes have
been deposited at various angles of incidence. With angular deposition a nanoporous
film was produced with an increase of pore size and orientation at higher deposition
angles. Optimum porosity is one of the parameters in metal oxide films that best
control gas response and sensing properties. The transmittance of the films
increased with increasing angle of deposition most likely due to the reduced overall
electron density as the porosity increases. The XPS and Raman analyses show a
mixed oxide film containing anatase TiO2 and monoclinic WO3 phases. Due to the
mixed nature of the TiO2-WO3, it is anticipated that the material would have a broad
gas selectivity. The gas sensing properties and selectivity of the films will be further
investigated.
Acknowledgments
This work was financially supported by Queensland University of Technology (QUT).
The authors are indebted to Dr Barry Wood for assistance with the XPS analyses.
References
[1]
A. Dieguez, A. Romano-Rodriguez, J.R. Morante, U. Weimar, M. SchweizerBerberich, W. Gopel, Sensors and Actuators B: Chemical 31 (1996) 1-8.
[2]
X. He, J. Li, X. Gao, L. Wang, Sensors and Actuators B: Chemical
Proceedings of the Ninth International Meeting on Chemical Sensors 93
(2003) 463-467.
10
[3]
V. Khatko, E. Llobet, X. Vilanova, J. Brezmes, J. Hubalek, K. Malysz, X.
Correig, Sensors and Actuators B: Chemical 111-112 (2005) 45-51.
[4]
T. Pagnier, M. Boulova, A. Galerie, A. Gaskov, G. Lucazeau, Sensors and
Actuators B: Chemical 71 (2000) 134-139.
[5]
B. Fruhberger, M. Grunze, D.J. Dwyer, Sensors and Actuators B: Chemical 31
(1996) 167-174.
[6]
I. Jimenez, J. Arbiol, G. Dezanneau, A. Cornet, J.R. Morante, Sensors and
Actuators B: Chemical Proceedings of the Ninth International Meeting on
Chemical Sensors 93 (2003) 475-485.
[7]
G. Sarala Devi, V. Bala Subrahmanyam, S.C. Gadkari, S.K. Gupta, Analytica
Chimica Acta: Molecular Electronics and Analytical Chemistry 568 (2006) 4146.
[8]
K. Zakrzewska, Vacuum Proceedings of the International Workshop on
Surface Physics: Metals on Solid Surfaces 74 (2004) 335-338.
[9]
N. Koshizaki, T. Oyama, Sensors and Actuators B: Chemical 66 (2000) 119121.
[10]
L.J. LeGore, R.J. Lad, S.C. Moulzolf, J.F. Vetelino, B.G. Frederick, E.A. Kenik,
Thin Solid Films 406 (2002) 79-86.
[11]
M.A. Ponce, C.M. Aldao, M.S. Castro, Journal of the European Ceramic
Society 23 (2003) 2105-2111.
[12]
D.H. Yoon, G.M. Choi, Sensors and Actuators B: Chemical 45 (1997) 251257.
[13]
C. Bittencourt, E. Llobet, M.A.P. Silva, R. Landers, L. Nieto, K.O. Vicaro, J.E.
Sueiras, J. Calderer, X. Correig, Sensors and Actuators B: Chemical 92
(2003) 67-72.
[14]
S.C. Moulzolf, S.-a. Ding, R.J. Lad, Sensors and Actuators B: Chemical 77
(2001) 375-382.
[15]
H. Kawasaki, T. Ueda, Y. Suda, T. Ohshima, Sensors and Actuators B:
Chemical 100 (2004) 266-269.
[16]
J. Shieh, H.M. Feng, M.H. Hon, H.Y. Juang, Sensors and Actuators B:
Chemical 86 (2002) 75-80.
[17]
A. Rothschild, F. Edelman, Y. Komem, F. Cosandey, Sensors and Actuators
B: Chemical 67 (2000) 282-289.
11
[18]
L.E. Depero, I. Natali Sora, C. Perego, L. Sangaletti, G. Sberveglieri, Sensors
and Actuators B: Chemical Materials for Sensors: Functional Nanoscaled
Structures of the 1995 E-MRS Spring Conference 31 (1996) 19-24.
[19]
C.G. Granqvist, Handbook of Inorganic Electrochromic Materials, Amsterdam,
Elsevier, 1995.
[20]
M. Grätzel, Prog. Photovolt. Res. Appl. 8 (2000) 171-185.
[21]
H. Lindström, A. Holmberg, E. Magnusson, L. Malmqvist, A. Hagfeldt, J.
Photochem. Phorobio. A: Chemistry 145 (2001) 107-112.
[22]
V.G. Bessergenev, I.V. Khmelinskii, R.J.F. Pereira, V.V. Krisuk, A.E.
Turgambaeva, I.K. Igumenov, Vacuum 64 (2002) 275-279.
[23]
H. Huang, O.K. Tan, Y.C. Lee, M.S. Tse, Journal of Crystal Growth
International Conference on Materials for Advanced Technologies (ICMAT
2005, Symposium M - Photonic Materials and Devices 288 (2006) 70-74.
[24]
C. Cantalini, H.T. Sun, M. Faccio, M. Pelino, S. Santucci, L. Lozzi, M.
Passacantando, Sensors and Actuators B: Chemical Materials for Sensors 31
(1996) 81-87.
[25]
J. Gottsche, A. Hinsch, V. Wittwer, Solar Energy Materials and Solar Cells 31
(1993) 415-428.
[26]
G.B. Smith, Appl. Opt. 29 (1990) 3685-3693.
[27]
Z.W. Chen, J.K.L. Lai, C.H. Shek, Chemical Physics Letters 422 (2006) 1-5.
[28]
G. Korotcenkov, V. Brinzari, A. Cerneavschi, M. Ivanov, V. Golovanov, A.
Cornet, J. Morante, A. Cabot, J. Arbiol, Thin Solid Films 460 (2004) 315-323.
[29]
X. Lai, Q. Guo, B.K. Min, D.W. Goodman, Surface Science 487 (2001) 1-8.
[30]
F.-M. Liu, T.-M. Wang, Applied Surface Science 195 (2002) 284-290.
[31]
P. Scardi, S. Veneri, M. Leoni, R. Polini, E. Traversa, Thin Solid Films 290-291
(1996) 136-142.
[32]
L. Zhao, Y. Yu, L. Song, X. Hu, A. Larbot, Applied Surface Science 239 (2005)
285-291.
[33]
L. Escobar-Alarcon, E. Haro-Poniatowski, M.A. Camacho-Lopez, M.
Fernandez-Guasti, J. Jimenez-Jarquin, A. Sanchez-Pineda, Appl. Surf. Sci.
137 (1999) 38.
12
Figures and Figure Captions
Figure 1. Schematic diagram of the evaporative deposition as a function of angles of
incidence, θ.
13
Figure 2
SEM images of titanium oxide-tungsten oxide film co-evaporated at (a) near normal
(b) 70 degree and (c) 80 degree angles of incidence.
14
a
b
15
c
d
Figure 3 AFM images of co-evaporated thin films deposited at (a) 30 degree, (b) 50
degree, (c) 70 degree and (d) 80 degree angles of incidence.
16
Figure 4 XPS spectra of co-evaporated titanium oxide-tungsten oxide film: (a)
survey (wide) scan between binding energy 0 to 1200 eV on the surface of the film
before etching, and (b), (c) narrow high resolution peaks for Ti 2p and W 4f,
respectively after etching the film.
17
70
Atomic Concentration (%)
O
60
50
40
30
C
W
20
K
Ti
10
P
0
0
10
20
30
40
Etching Time (min)
Figure 5 Atomic concentrations of C 1s, O 1s, Ti 2p, W 4f, K 2s and P 2p as a function
of etching time for titanium oxide-tungsten oxide film analyzed using high resolution
XPS.
Figure 6 Raman shift of co-evaporated film before and after annealing at 450oC for 24
hours measured using 633 nm HeNe laser source and a power of 10 mW at the
sample.
18
80
Substrate
70
Transmittance (%)
80
o
50
60
o
30
50
o
40
30
20
10
0
300
500
700
900
1100
Wavelength (nm)
Figure 7 Transmittance of titanium oxide-tungsten oxide films deposited at different
angles of incidence as a function of wavelength between 300 nm to 1100 nm. For
comparison the transmittance of the substrate (conducting glass) is shown.
19
Tables and Table captions
Table Captions
Angle of deposition
Average grain size
30 °
300 nm
50 °
346 nm
70 °
387 nm
80 °
250 nm
Table 1. Grain sizes of titanium oxide-tungsten oxide film deposited between 30°and
80° angles of incidence, determined from AFM scans.
Element
Peak
at %
Position
min)
(0
at %
(5
at % (10
at % (20
at % (30
at % (40
min)
min)
min)
min)
min)
Be (eV)
C 1s
285
40.7
11.4
9.2
6.8
5.1
9.4
O 1s
531
43.2
55.9
61.7
62.6
56.6
54.8
Ti 2p
459
3.3
7.3
7.7
7.7
9.1
8.0
W 4f
36
8.7
15.7
18.7
21.4
20.5
19.9
K 2s
378
1.2
6.9
1.1
0.2
4.8
5.6
P 2p
134
3.0
2.8
1.7
1.3
3.9
2.3
20
Table 2. Atomic concentration of co-evaporated titanium oxide-tungsten oxide film for
various etching times obtained using XPS analysis.
Deposition
Luminous
Angle
Transmittance
30
o
53%
50
o
55%
70
o
59%
80
o
60%
substrate
65%
Table 3. Luminous transmittance (%) of co-evaporated titanium oxide-tungsten oxide
films for various deposition angles.
21