XPS studies on surface reduction of tungsten oxide nanowire film by Ar + bombardment

Article abstract of DOI:10.1016/j.elspec.2012.01.004

WO₃ nanowire film was bombarded by Ar ion beam in the analysis chamber of an X-ray photoelectron spectroscopy (XPS) system to produce uniform tungsten cone arrays. The WO₃ nanowire film itself served as an etching mask during the Ar⁺ bombardment. The changes of surface chemical states and electronic structures during bombardment were monitored by in situ XPS. The morphological evolution with different Ar⁺ bombardment time was observed by ex situ scanning electron microscopy (SEM). At the start of Ar⁺ bombardment partial W⁶⁺ in WO₃ was reduced to W⁵⁺ immediately, subsequently to W⁴⁺ and then to W^x+ (intermediate chemical state between W⁴⁺ and W⁰), finally to W⁰. Multiple oxidation states of tungsten coexisted until finally only W⁰ left. SEM images showed that the nanowires were broken and then fused together to be divided into clusters with a certain orientation after long-time high-energy ion beam bombardment. The mechanism of the ion-induced reduction during bombardment and the reason of the orientated cone arrays formation were discussed respectively.

Full text of DOI:10.1016/j.elspec.2012.01.004

Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
Contents lists available at SciVerse ScienceDirect
Journal of Electron Spectroscopy and
Related Phenomena
journal homepage: www.elsevier.com/locate/elspec
XPS studies on surface reduction of tungsten oxide nanowire film by
Ar+ bombardment
F.Y. Xiea, L. Gonga, X. Liub, Y.T. Taob, W.H. Zhanga, S.H. Chenb, H. Mengc, J. Chena,b,∗
a Instrumental Analysis & Research Center, Sun Yat-sen University, Guangzhou 510275, PR China
b Guangdong Province Key Laboratory of Display Materials and Technologies, School of Physics and Engineering, Sun Yat-sen University, Guangzhou 510275, PR China
c The Key Laboratory of Low-carbon Chemistry & Energy Conservation of Guangdong Province, School of Physics and Engineering, Sun Yat-sen University, Guangzhou 510275, PR
China
a r t i c l e i n f o
a b s t r a c t
Article history:
WO3 nanowire film was bombarded by Ar ion beam in the analysis chamber of an X-ray photoelectron
spectroscopy (XPS) system to produce uniform tungsten cone arrays. The WO3 nanowire film itself served
as an etching mask during the Ar+ bombardment. The changes of surface chemical states and electronic
structures during bombardment were monitored by in situ XPS. The morphological evolution with dif-
ferent Ar+ bombardment time was observed by ex situ scanning electron microscopy (SEM). At the start
of Ar+ bombardment partial W6+ in WO3 was reduced to W5+ immediately, subsequently to W4+ and then
to Wx+ (intermediate chemical state between W4+ and W0), finally to W0. Multiple oxidation states of
tungsten coexisted until finally only W0 left. SEM images showed that the nanowires were broken and
then fused together to be divided into clusters with a certain orientation after long-time high-energy ion
beam bombardment. The mechanism of the ion-induced reduction during bombardment and the reason
of the orientated cone arrays formation were discussed respectively.
Received 10 September 2011
Received in revised form 28 January 2012
Accepted 28 January 2012
Available online 6 February 2012
Keywords:
X-ray photoelectron spectroscopy
Ar ion bombardment
Ion-induced reduction
Tungsten oxide
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
The implantation of bombardment particles into the surface is real-
ized at lower sputtering rate of the bombarded components. The
Tungsten trioxide (WO3) is an important semiconductor due to
its wide applications in electrochromic devices [1], gasochromic
devices [2], gas sensors [3], field emission devices [4] and pho-
tocatalysts [5]. Its excellent photochromic, electrochromic and
gasochromic properties are attributed to easy transformation
between stoichiometric WO3 and nonstoichiometric WO3−x. Non-
stoichiometric WO3−x such as W18O49 [6] and W20O58 [7] consist
of W6+ and W5+ as a result of inadequate oxidation in the process
of sample preparation. Different tungsten oxidation states can be
obtained by Ar+ bombardment on WO3 nanowires owing to ion-
induced reduction effect.
gross surface heating is caused by the interaction between the
surface material and the high energy incident ions. In the ion-
induced reduction process, the ion energy transferred to the lattice
atoms may exceed the binding energies in molecules by several
orders of magnitude, which therefore suffices to break up chemical
bonds and to create new ones [9,10]. These side effects of ion bom-
bardment can make depth profile analysis more complicated. The
preferential sputtering usually causes errors in the components’
depth profile analysis. In particular, the chemical states of some
metal oxides are difficult to be estimated in a depth profile because
of the ion-induced reduction. For many transition metal oxides such
as TiO2 [11], V2O5 [12], Ni2O3 [13], MoO3 [14], WO3 [15], CeO2
[16] and PdO [8] the reduction to the corresponding metal or lower
valence oxide is often observed under ion bombardment.
In previous work [17], we obtained completely metallic W cone
arrays from WO3 nanowire film by long-time Ar+ bombardment
with WO3 nanowire film itself serving as the etching mask. The
field-emission measurements indicated that the tungsten cone
arrays had a better performance of field emission than that of WO3
nanowire film. In the process of Ar+ bombarding on WO3 nanowire
film the change of W chemical states and the morphological change
from nanowires to cone arrays should be concerned.
In XPS study Ar+ bombardment is used to peel away surface
atoms so as to achieve the depth profile of sample composition.
The exchanged energy from the high energy ion to the surface
atoms results in the removal of surface atoms from the sam-
ple [8,9]. In addition to the removal of the surface layer, ion
bombardment may also cause other effects such as preferential
sputtering, implantation of bombardment particles, gross surface
heating and ion-induced reactions [8]. The preferential sputtering
is caused by different sputtering rates of individual constituents.
In this work the whole ion-induced reduction process of WO3
nanowires was monitored by in situ XPS measurement to study
the evolution of tungsten chemical states. W5+ oxidation state
∗
Corresponding author at: Instrumental Analysis & Research Center, Sun Yat-sen
University, Guangzhou 510275, PR China.
E-mail address: puscj@mail.sysu.edu.cn (J. Chen).
0368-2048/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.elspec.2012.01.004

F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
113
appeared immediately when the bombardment began, an interme-
diate chemical state between W4+ and W0, Wx+ started to appear
from 140 s of bombardment, and from 160 s on, all oxide states of
W including W6+, W5+, W4+, Wx+, W0 coexisted until there was only
W0 in the end. The morphological change of the samples with dif-
ferent Ar+ bombardment time was observed by ex situ SEM. SEM
images revealed that the nanowires were broken and then fused
together to be divided into wire cluster with a certain direction
after long-time high-energy ion beam bombardment due to the
long-time gross heating effect.
film before Ar+ bombardment FWHM of W4f7/2 was 1.0 eV, while
FWHM of W4f7/2 in W6+ oxidation state was 1.8 eV at 300 s of bom-
bardment time. These parameters are summarized in Table 1.
3. Results and discussion
3.1. XPS analysis
3.1.1. W4f core level spectra
W4f core level XPS spectra of WO3 nanowires with different
reduction degrees were obtained by Ar+ bombardment on the
same WO3 nanowire film (Fig. 1a–k). During Ar+ bombardment
the peak BE (binding energy) of W4f7/2 and the spectral line-shape
of W4f core level varied drastically. The W4f7/2 peak BE shifted
from 35.6 eV of the initial WO3 nanowire film to 31.0 eV of the
final etched sample. Intermediate oxidation states were observed
in the reduction process of WO3 to metallic W. The spectral shape
of W4f core level changed from narrow doublet peaks to broad
bands which included several oxidation states of W, in the end the
W4f spectra changed into narrow doublet peaks corresponding to
metallic W.
In order to understand the evolution process of W oxidation
states change, the spectra with different Ar+ bombardment time
were fitted on the basis of the parameters in Table 1. Fig. 1a
shows a W4f XPS spectrum of the stoichiometric WO3 nanowire
film before Ar+ bombardment where only W6+ spin–orbit dou-
blet existed. W5+ oxidation state emerged immediately after the
bombardment (Fig. 1b). A couple of peaks assigned to W4+ oxida-
tion state appeared with the bombardment time increasing to 70 s
(Fig. 1c). After 140 s of bombardment a new W chemical state with
W4f7/2 peak BE of 31.9 eV began to appear (Fig. 1d). The spectrum
could not be well fitted without the addition of a fourth component
(Fig. 1d, 140 s). From 160 s on, metallic W0 was presented (Fig. 1e).
With a fourth component it was possible to well fit the spectra in
Fig. 1d and e. The fourth component had a BE between W4+ (33.2 eV)
and W0 (31.1 eV) which was assigned as Wx+ in this work. Similar
situation was observed when Ar+ bombarded to V2O5 pellet where
2. Experiment
Stoichiometric WO3 nanowire film was prepared on silicon sub-
strate by thermal evaporation method (details in Ref. [18]). The
WO3 film was loaded into Thermo Fisher Scientific ESCALAB 250
XPS system, where Ar+ bombardment was performed and surface
chemical state of the sample was characterized. XPS spectra were
recorded using a monochromated Al K␣ (1486.6 eV) X-ray source
with the take-off angle of 90◦ after different Ar+ bombardment time.
The energy resolution of the instrument at 20 eV of pass energy is
0.6 eV, as estimated from the full width at half maximum (FWHM)
of Ag 3d5/2 peak. The Ar+ bombardment was operated with EX05
argon ion gun whose angle arrangement given the sample surface
normal is 45◦ at a chamber pressure of 1 × 10−6 Pa. The ion beam
voltage was 3 kV and the emission current was 2 ␮A. The etching
area was 2 mm × 2 mm, which was much larger than X-ray spot
(320 ␮m in diameter). The spectra in this paper were calibrated
using W4f7/2 peak of WO3 at 35.6 eV.
In order to observe the morphological change during Ar+ bom-
barding on WO3 nanowire film, the surface morphology of the
samples with different Ar+ bombardment time were observed by
field emission environmental SEM (FEI Quanta 400F) whose energy
of the primary electron beam was 20 kV with the Everhart-Thornley
detector. The samples for SEM observation were taken out from the
UHV chamber after WO3 nanowire film was bombarded for a period
of time.
The composition and chemical states changes induced by Ar+
bombardment were analyzed by XPS. The W4f core levels were
recorded with a step of 0.05 eV and a pass energy of 20 eV. The
Shirley-type background due to inelastic scattering electrons was
taken into account. The XPS spectra of W4f core level were fitted
into peak doublets with parameters of spin–orbit separation ꢀEP
V
2+/V1+ presented [12,20]. Subsequently, from 160 s to 4800 s, var-
ious tungsten oxidation states including W6+, W5+, W4+, Wx+ and
W0 coexisted (Fig. 1f–i). From 4800 s on, only metallic W0 doublet
peaks with W4f7/2 peak BE at about 31.0 eV was observed (Fig. 1j
and k).
By analyzing the W4f core level XPS spectra from 140 s to 4800 s
(Fig. 1d–j), the W4f7/2 peak of Wx+ was related to the shape of the
curve in the range between 30.0 eV and 32.5 eV. Compared with
pure metallic W state in Fig. 1j, the tail height of the curve in this
range was much higher in Fig. 1e–i. When fitting this range with the
same asymmetric parameters of the pure metallic W, there should
be a Wx+ peak centered at 32.0 eV. Increasing the bombardment
time to 4800 s the tail heights of the curve in the range between
30.0 eV and 32.5 eV dropped to zero where only W0 existed. It
was concluded that the Wx+ state should be an intermediate state
between W4+ and W0, which might originate from the W3+, W2+
or W1+ oxidation state. Wx+ may also originate from the tungsten
ions with partial unbroken oxygen ions and W W bonds. Or there
would also be ‘beta’ shift induced from species adjacent to the ion
attached to the photoionized W atoms to form this intermediate
chemical state. It is difficult to assign the intermediate chemical
state to a specific formal oxidation state.
(4f5/2–4f7/2) = 2.1 0.1 eV and intensity ratio IW4f /IW4f
= 0.75
5/2
7/2
using Thermo Avantage v4.30 software. The ratio of Lorentzian to
Gaussian varied in the range of 30 5% and the spectra of pure
metallic W were asymmetric ascribed to the coupling of the core
hole with the Fermi sea whose asymmetric parameters were used
in fitting metallic W state. The asymmetric peak shape was deter-
mined using the equation [19]
T = TM · CT + (1 − TM)exp(−Dx · ET)
(1)
where Dx is the separation from the peak center in channels, ET is
the exponential tail, TM was the tail mixing ratio, CT is the constant
tail height. For metallic W state the values of ET, TM and CT were
0.0373, 77.7% and 0, respectively.
W5p3/2 peak was in the higher BE (binding energy) side of
W4f5/2. W5p3/2 invariably accompanied with W4f peak and their
relative peak position and intensity were nearly unchangeable. The
relationships between the BE and peak intensity of W5p3/2 and
The evolution of the concentrations of the different tungsten
chemical states as function of Ar+ bombardment time is shown in
Fig. 2, expressed in percentage of the total W4f7/2 signal. The most
drastic changes could be found in the first 160 s. The W6+ decreased
drastically, W5+, W4+, Wx+ and W0 appeared gradually. The content
of W6+ decreased to zero after WO3 nanowire film was bombarded
W4f7/2 were as follows: BEW5p = BEW4f + 5.8, IW5p /IW4f
=
3/2
3/2
7/2
7/2
0.08 as shown in Fig. 1a and Table 1. Due to the mixing effect of the
ion beam the full width at half maximum (FWHM) of W6+ peak were
larger than those of WO3 nanowires film at the initial bombard-
ment state (before 70 s of bombardment time). For WO3 nanowires

114
F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
Fig. 1. Representative fitted W4f core level spectra under 3 kV × 2 ␮A Ar ion gun bombarding on WO3 nanowire film after different Ar+ bombardment times: (a) 0 s; (b) 10 s;
(c) 70 s; (d) 140 s; (e) 160 s; (f) 300 s; (g) 500 s; (h) 1500 s; (i) 3000 s; (j) 4800 s; (k) 8400 s.
when W5+ and W4+ appeared as shown in Figs. 1 and 2. The O/W
ratio decreased gradually along with Ar+ bombardment process,
and there was still O element remained in the sample at the final
stage. During the Ar+ bombardment process the oxygen content
reduced due to the preferential sputtering of oxygen in the metal
oxide compound.
for 4800 s, the contents of the intermediate W oxidation states
(W5+, W4+, Wx+) increased and then decreased to zero, whereas
W0 increased all the time. The O/W ratio at the different oxida-
tion stages is shown in Table 2. The O/W ratio was calculated using
the Scofield sensitivity factor method where the Scofield sensitiv-
ity factors of W4f and O1s are 8.22 and 2.93, respectively. During
the initial 70 s, the O/W ratio was close to the stoichiometric ratio
of WO3. However the chemical states of W changed significantly
At present in literature the reduction of metal oxides is mostly
explained by the preferential sputtering of oxygen in the metal

F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
115
Table 1
Overview of the W4f main and W5p3/2 peak fit parameters for different tungsten oxidation states.
Tungsten oxidation
states
W4f7/2 BE (eV)
%L-G
FWHM (eV)
ꢀBE (eV) (W4f7/2–W4f5/2
)
IW4f /IW4f
W5p3/2 BE (eV)
IW5p /IW4f
7/2
3/2
5/2
7/2
W6+
W5+
W4+
Wx+
aW0
35.6 0.1
34.3 0.2
33.2 0.2
32.0 0.2
31.0 0.1
b1.8
c1.5
1.5
1.2
0.8
41.30 0.1
40.10 0.1
39.00 0.1
37.80 0.1
36.75 0.1
30
5
2.1 0.1
0.75
0.08
a
The asymmetric peak-shape parameters of the metallic W: exponential tail (ET) is 0.0373, tail mixing ratio (TM) is 77.7%.
Before 70 s of bombardment FWHM of W6+ is 1.0 eV because there is no serious mixing effect induced by ion beam bombardment.
Before 70 s of bombardment FWHM of W5+ is 1.0 eV without serious mixing effect induced by ion beam bombardment.
b
c
Table 2
The O/W ratio during the whole Ar+ bombardment process.
Bombardment time (s)
O/W ratio
0
10
70
140
160
300
500
1500
3000
4800
8400
73:27
73:27
73:27
71:29
68:32
65:35
64:36
57:43
53:47
24:76
16:84
oxide compounds. However, our results showed that part of WO3
was reduced to W5+ and W4+ in the initial stage of Ar+ bombard-
ment while O/W ratio remained unchanged. The first stage of ion
bombardment led to partial break of the W O bonds. The oxygen
split off in this way did not necessarily leave the surface; instead, it
might remain at interstitial lattice sites of the surface layer. There
was no contradiction that the observed reduction of the O1s inten-
sity is slower than the change of the tungsten valence states as
shown in Fig. 2 and Table 2. This also happened for MoO3 [9]. Based
on above analysis one conclusion was reached: the ion-induced
reduction is not necessarily caused by the preferential sputtering
of oxygen.
where I (A) is the emission current of Ar ion beam, U (V) is the
ion beam voltage, t (s) is the duration of Ar+ bombardment. The
required energy for WO3 to lower valence WO2 and metallic W
(ETrans) is calculated by the following equation:
ꢁ × V
ETrans
=
× ꢀrH◦
(3)
M
where ꢁ (g/cm3) is the density of WO3, V (cm3) is the etched region
volume by Ar+ bombardment, M (g/mol) is the molecular weight
of WO3, ꢀrH◦ is the standard molar heat of reaction from WO3 to
lower valence WOx and metallic W at 298.15 K in kJ/mol.
W4+ started to appear after 70 s of bombardment (shown in
Figs. 1c and 2). The energy of Ar ion beam (EAr+ ) and the required
energy (ETrans) from WO3 to lower valence WO2 were calculated
using Eqs. (2) and (3), respectively. The energy of Ar ion beam (3 kV,
2 ␮A) was 4.2 × 10−1 J calculated by Eq. (2) where I was 2 × 10−6 A,
U was 3 × 103 V, and t was 70 s.
The first ion-induced reduction study in XPS was presented by
Kim et al. [8]. They proposed a model based on bulk thermodynamic
free energy considerations to explain the reduction of metal oxides
to the corres◦ponding metal or lower valence oxides. All oxides
whose −ꢀGf value at room temperature were below a certain
value could be reduced by exposure to Ar ions whereas those oxides
As for ETrans we should obtain the standard molar heat of reac-
tion transformed from WO3 to lower valence WO2 (ꢀrH◦) which
was determined by the standard molar heat of formation (ꢀfH◦) at
298.15 K of the reactants and products in the following reaction
◦
with −ꢀGf above another definite value were stable to this expo-
sure. Using Kim’s proposed model we calculated the energy of Ar
ion beam and the required energy for the transformation from WO3
to lower valence WO2 and metallic W, respectively. The energy of
Ar ion beam (EAr+ ) sputtering for a period of time (t) was calculated
by
WO3 → WO2 + (1/2)O2
where the ꢀfH◦ values of WO2, O2 and WO3 are −589.7 kJ/mol,
0 kJ/mol and −842.9 kJ/mol [21]. The value of the standard molar
heat of reaction ꢀrH◦ is 253.2 kJ/mol calculated by the above ꢀfH◦
values. The number of moles of WO3 etched by Ar+ bombardment
was given by (ꢁ × V)/M, where ꢁ of WO3 is 7.16 g/cm3 and M is
231.85 g/mol of WO3. The etched region volume V is 4 × 10−8 cm3
estimated by the product of etched area (2 mm × 2 mm) and the
etched depth is 10 nm for 70 s (the etched depth is obtained from
the calibrated etching rate 0.154 nm/s for SiO2 under the same
Ar+ bombardment condition). Putting the above values into Eq.
(3), ETrans from WO3 to WO2 in the etched region for 70 s was
3.1 × 10−4 J. Compared with the energy of Ar ion beam (EAr+ ) and
the required transformation energy (ETrans) from WO3 completely
to WO2 in the etched region, the value of EAr+ was about 1000 times
as that of ETrans from WO3 to WO2 shown in Table 3. This result indi-
cated that the energy required from WO3 to WO2 was only a very
small fraction of the Ar+ beam energy.
EAr+ = I × U × t
(2)
After 4800 s of bombardment WO3 was transformed into metal-
lic W completely as shown in Figs. 1j and 2, the values of
EAr
+
and ETrans from WO3 to metallic W (see Table 3) were
2.9 × 101 J and 7.7 × 10−2 J, respectively where ꢀrH◦ of the reaction
WO3 → W + (3/2)O2 was 842.9 kJ/mol and the etched depth 740 nm
for 4800 s was estimated from the calibrated etching rate of SiO2 at
Fig. 2. Evolution of the different tungsten oxidation states as function of Ar+ bom-
bardment time. The inset shows the first 500 s in more detail.

116
F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
Table 3
The energy of Ar ion beam (EAr+ ) and the required transformation energy from WO3 to lower valence W states (ETrans) for different bombardment time.
Bombardment
time
Energy of Ar ion beam/EAr
+
Involved reaction
Required
transformation
energy/ETrans
70 s
4800 s
4.2 × 10−1
J
WO3 → WO2 + (1/2)O2
WO3 → W + (3/2)O2
3.1 × 10−4
7.7 × 10−2
J
J
2.9 × 101
J
0.154 nm/s. The value of EAr+ was about 350 times as that of ETrans
from WO3 to metallic W.
essential reason was the varied valence states of W element which
led to the coexistence of multiple oxidation states in XPS W4f core
level spectra besides the preferential sputtering of oxygen was an
important factor. Various oxidation states had been observed in
many transition metal oxides such as MoO3, ReO3, TiO2, V2O5 and
ZrO2 after Ar+ bombardment. A small amount of Ti3+ had been
observed even in the case of 10 V bombardment for TiO2 sample
[11]. Ion-induced reduction of oxides should be associated with
the inherent properties of elements in oxide compounds, such as
transition metal’s electron occupied states of d or f orbits, the varied
valences of the composed metal constituents in oxides, the value
of M O (metal oxygen) bond energy in oxide compounds, besides
the preferential sputtering of oxygen. Detailed studies about ion
induced reduction of oxides were reported in our other work [22].
From above calculation the energy of Ar ion beam exceeded by
several orders of magnitude the required transformation energy
from WO3 to lower valence states, which was in agreement with
Holm’s proposed opinion [9,10]. Holm et al. considered the energy
received by the lattice atoms from the bombarding ion might
exceed by several orders of magnitude the binding energies in
molecules so that chemical bonds could be broken. Theoretically
the reduction of all metal oxides could happen while there was no
ion induced reduction of Al2O3, MgO, ZnO, SnO2, Ga2O3, In2O5 and
SiO2 in experiments. The ion-induced reductions were observed
in some metal oxides, which indicated that most of the energy of
Ar ion beam was used to remove the surface atoms and the ion-
induced reduction was merely a side-effect of ion bombardment for
some compounds. Therefore the reaction heat involved during Ar+
bombardment was not the main reason for ion-induced reduction.
The ion-induced reduction of some metal oxides by ion bom-
bardment was very complicated which could not be explained
simply by the preferential sputtering of oxygen and the reac-
tion heat involved in ion bombardment independently. For
the ion-induced reduction of WO3 we believe that the more
3.1.2. XPS valence band spectra
The XPS valence band spectra were used to characterize the sur-
face electronic structure of the materials. The energy bands within
the valence band of WO3 have almost entire characteristics of tung-
sten 5d or oxygen 2p orbital [23]. The typical XPS valence band
spectra of WO3 nanowires film at different bombardment time are
shown in Fig. 3. Fig. 3a shows the XPS valence band spectrum of
Fig. 3. XPS valence band spectra of the different bombardment time on WO3 nanowire film: (a) 0 s; (b) 70 s; (c) 160 s, (d) 300 s; (e) 3000 s; (f) 8400 s.

F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
117
Fig. 4. The surface morphologies of the different bombardment time during Ar+ bombardment on WO3 nanowire film: (a) 0 s; (b) 300 s; (c) 500 s; (d) 1500 s; (e) 3000 s; (f)
4800 s; (g) 8400 s; (h) magnified image of g.
pure WO3 nanowire film. Compared with the literature results [24],
the spectrum in Fig. 3a was characteristic of a monoclinic phase
␥-WO3 consistent with the results from Raman analysis in our pre-
vious work [18]. Covalent bonding interactions transformed O2p
and W5d orbitals into bonding and antibonding orbitals. The peak
located at ∼4.1 eV was assigned to an O2p-derived band while the
peak centered at ∼6.2 eV corresponded to a hybridized ‘W5d–O2p’
band. The onset of the valence band spectrum at ∼2.5 eV could be
assigned to the valence band maximum (VBM).
A new peak near the Fermi level appeared in the XPS valence
band spectrum of the sample after the WO3 nanowire film was
bombarded for 70 s (Fig. 3b). As shown in Fig. 1c, there were W5+
and W4+ together with W6+ in this sample so that residual W5d
electrons appeared. Due to the empty conduction band of tungsten
oxide associated with the W5d electrons [25], the peak near the
Fermi level was attributed to the conduction band filled with the
residual W5d electrons which was associated with lower oxidation
states.
At the bombardment time of 160 s, the edge of the Fermi level
appeared at 0 eV where metallic W started to emerge (Fig. 3c) and
the peak of O2p decreased a lot compared with that of pure WO3
nanowires film. Once metallic W came into existence the peak near
Fermi level corresponding to W5d electrons became not obvious
because the density of metallic W increased. The intensity of the
hybridized W5d–O2p decreased while the density of the metallic
W increased with the increase of Ar+ bombardment time (Fig. 3d

118
F.Y. Xie et al. / Journal of Electron Spectroscopy and Related Phenomena 185 (2012) 112–118
and e). Fig. 3f corresponds to the metallic W which had a sharp edge
of the Fermi level. The O2p and hybridized W5d–O2p disappeared
in the valence band spectrum of metallic W. In conclusion, the XPS
valence band spectra further confirmed the results of W4f core level
XPS spectra.
the long-time gross surface heating of ion bombardment with the
orientated high-energy Ar ion beam emitted by EX05 ion gun.
4. Conclusions
In summary, WO3 nanowire film was reduced to metallic W by
Ar+ bombardment through a complicated process of chemical state
change of tungsten. During Ar+ bombardment the content of W6+
decreased to zero, the contents of the intermediate oxidation states
of W (W5+, W4+, Wx+) increased and then decreased, the content of
W0 increased all the time until the chemical state changed into
metallic W after long-time Ar+ bombardment. Ion-induced reduc-
tion of WO3 should be associated with the varied valence states of
W element besides the preferential sputtering of oxygen.
The different surface morphologies after the different etching
time during Ar+ bombardment on the WO3 nanowire film was
explained by the same mechanism proposed in our previous work.
The WO3 nanowire film itself served as an etching mask to produce
tungsten cone arrays due to the long-time gross surface heating of
ion bombardment with the orientated high-energy Ar ion beam.
3.2. SEM surface morphology characterization
The surface morphologies of the WO3 nanowire film (Fig. 4a)
and the samples obtained from Ar+ bombardment on the same
batch of WO3 nanowire film for different time were investigated
by SEM. Using thermal evaporation method high density WO3
nanowires were prepared on the silicon substrate. The average
length of WO3 nanowires was about 4 ␮m and the diameter varied
from 50 to 80 nm. The WO3 nanowires had a random distribution
on the surface of the silicon substrate. Though the W4f core level
XPS spectrum changed significantly from narrow doublet peaks to
very broad spectrum shown in Figs. 1 and 2, the morphology of
the sample after 300 s of bombardment was similar with that of
the as-prepared WO3 nanowire film which had a random distribu-
tion without obvious change (Fig. 4b). For the sample after 300 s
of bombardment W6+, W5+, W4+, Wx+, W0 coexisted as shown in
Fig. 2f. For the sample after 500 s of bombardment the morphology
has changed slightly with some of the wires broken or bent by con-
tinuous Ar+ bombardment (Fig. 4c). After 1500 s of bombardment
the wires were broken and started to be divided into wire clus-
ters with a certain orientation because the long-time high-energy
ion beam bombardment provided enough energy to fuse the bro-
ken wires together (Fig. 4d). The chemical states of W were W0,
Wx+ and W4+ together with a small amount of W5+ and W6+ are
shown in Figs. 1h and 2. The initial morphology of cone arrays
emerged after 3000 s of bombardment. The cones were consisted
of many broken wires fused together with apparent gaps inside
(Fig. 4e). After 3000 s of bombardment, there was mostly W0 with
small amounts of other residual W oxidation states (Figs. 1i and 2).
The cones became sharper after 4800 s, and the wires were fur-
ther fused together with gaps smaller than that of the sample after
3000 s (Fig. 4f). Only metallic W was observed after 4800 s. In the
end, metallic tungsten cone arrays with a uniform orientation were
obtained after 8400 s, the wires was completely fused together
without any slit on the cones (Fig. 4g and h) and the tungsten chem-
ical state did not change any more even if ion bombardment time
was increased.
Acknowledgments
This work was financially supported by the National Natural Sci-
ence Foundation of China (Grant Nos. 51072236, 21106190), the
Department of Education and Department of Science and Tech-
nology of Guangdong Province (Grant No. 1051027501000094),
the Doctoral Fund of Ministry of Education of China (Grant
No. 20100171120022), the Research Foundation of Instrumen-
tal Analysis & Research Center of Sun Yat-Sen University (Grant
Nos. IARC 2009005, IARC 2010002), the Analysis Foundation of
Guangzhou Scientific Instrument Network (Grant No. 201003), Fos-
han Research Project (Grant No. 2011BC100023).
References
[1] N.A. Galiote, R.L.T. Parreira, J.M. Rosolen, F. Hugurenin, Electrochem. Commun.
12 (2010) 733–736.
[2] H.J. Chen, N.S. Xu, S.Z. Deng, D.Y. Lu, Z.L. Li, J. Zhou, J. Chen, Nanotechnology 18
(2007) 205701.
[3] T. Tesfamichael, M. Arita, T. Bostrom, J. Bell, Thin Solid Films 518 (2010)
4791–4797.
[4] J. Zhou, Y. Ding, S.Z. Deng, L. Gong, N.S. Xu, Z.L. Wang, Adv. Mater. 17 (2005)
2107–2110.
[5] D. Tanaka, Y. Oaki, H. Imai, Chem. Commun. 46 (2010) 5286–5288.
[6] Y. Li, Y. Bando, D. Gollberg, Adv. Mater. 15 (2003) 1294–1296.
[7] J. Chen, D.Y. Lu, W.H. Zhang, F.Y. Xie, J. Zhou, L. Gong, X. Liu, S.Z. Deng, N.S. Xu,
J. Phys. D: Appl. Phys. 41 (2008) 115305.
[8] K.S. Kim, W.E. Baitinger, J.W. Amy, N. Winograd, J. Electron Spectrosc. Relat.
Phenom. 5 (1974) 351–367.
[9] R. Holm, S. Storp, Appl. Phys. 12 (1977) 101–112.
[10] R. Holm, S. Storp, Phys. Scripta 16 (1977) 442–447.
[11] S. Hashimoto, A. Tanaka, Surf. Interface Anal. 34 (2002) 262–265.
[12] G. Silversmit, D. Depla, H. Poelman, G.B. Marin, R.D. Gryse, Surf. Sci. 600 (2006)
3512–3517.
[13] K.S. Kim, N. Winograd, Surf. Sci. 43 (1974) 625–643.
[14] N.V. Alov, D.M. Kutsko, K.V. Bordo, J. Surf. Investig.: X-Ray 2 (2008) 184–188.
[15] H.Y. Wong, C.W. Ong, R.W.M. Kwok, K.W. Wong, S.P. Wong, W.Y. Cheung, Thin
Solid Films 376 (2000) 131–139.
[16] L.M. Qiu, F. Liu, L.Z. Zhao, Y. Ma, J.N. Yao, Appl. Surf. Sci. 252 (2006)
4931–4935.
[17] F.Y. Xie, L. Gong, X. Liu, J. Chen, W.G. Xie, W.H. Zhang, S.H. Chen, Appl. Surf. Sci.
256 (2009) 693–697.
[18] D.Y. Lu, J. Chen, H.J. Chen, L. Gong, S.Z. Deng, N.S. Xu, Y.L. Liu, Appl. Phys. Lett.
90 (2007) 041919.
[19] P.M.A. Sherwood, in: D. Briggs, M.P. Seah (Eds.), Appendix 3 in Practical Surface
Analysis, Wiley, 1983, pp. 460–461.
[20] M.C. Biesinger, L.W.M. Lau, A.R. Gerson, R.S.C. Smart, Appl. Surf. Sci. 252 (2010)
887–898.
[21] D.R. Lide, CRC Handbook of Chemistry and Physics, 87th ed., 2006, pp. 5–17.
[22] F.Y. Xie, L. Gong, W.H. Zhang, S.H. Chen, J. Chen, J. Electron Spectrosc. Relat.
Phenom., submitted for publication.
[23] R.D. Bringans, H. Hochst, H.R. Shanks, Phys. Rev. B 24 (1981) 3481–3489.
[24] C. Bittencourt, A. Felten, F. Mirabella, P. Ivanov, E. Llobet, M.A.P. Silva, L.A.O.
Nunes, J.J. Pireaux, J. Phys. Condens. Matter 17 (2005) 6813–6822.
[25] B.A.D. Angelis, M. Schiavello, J. Solid State Chem. 21 (1977) 67–72.
The morphological change during Ar+ bombardment was in
agreement with the schematic diagram of the ion sputtering pro-
cess proposed in our previous work [17]. In brief, the top ends of
the nanowires served as the etching mask when ion beam bom-
barded the sample. Some of ion beam bombarded the top ends of
the wires to break them down, and the others entered into the gap
between the wires to divide the wires into some clusters. Because
the high-energy Ar+ beam was orientated at an incident angle of
45◦, the nanowires started to be divided into clusters with a cer-
tain direction. The long-time high-energy ion beam bombardment
provided enough energy to fuse the broken wire cluster together
and finally orientated cone arrays was obtained. The main forma-
tion mechanism of the orientated cone arrays was the gross surface
heating produced by ion bombardment. The gross surface heat-
ing of long-time ion bombardment caused the nanowires to be
broken and fused together. The morphological change was much
slower than the change of tungsten chemical states because the
energy used to break the nanowires was provided by gross sur-
face heating of long-time ion beam bombardment whereas W
O
chemical bonds were broken immediately once ion beam hit the
WO3 nanowires film. The uniform orientation of cone arrays was
caused by the fixed incident angle of 45◦ of the high-energy Ar ion
beam. Therefore the formation of the directional cone arrays was