N. Alov et al. / Surface Science 600 (2006) 1628–1631
1629
of comparison the oxide films were also grown by thermal
oxidation of vanadium by molecular oxygen.
2. Experimental
The experiments were carried out using a VG ESCA 3
Mk II and Gammadata Scienta ESCA 310 electron spec-
trometers equipped with a source of AlKa (1486.6 eV)
radiation, hemispherical electron analyzer and a VG AG
2 cold cathode ion gun. The pressure in the analyzer cham-
ber during spectra acquisition was 10À8 Pa. Polycrystalline
vanadium foil 0.127 mm thick (99.7%, Aldrich Chemical
Co.) was used as a starting material. Prior to oxidation
the vanadium surface was cleaned by sputtering with argon
ions (E = 6 keV, I = 40 lA). No annealing was applied
after sputtering. The oxide layers were grown in the prepa-
ration chamber of the spectrometer. In the present experi-
ments molecular oxygen ion bombardment was carried out
with ion energy of 6 keV and fluence ranging from 1015 to
1017 ions/cm2. The ion beam was directed normal to the
sample surface. The projected range of 3 keV oxygen atoms
produced by surface neutralization and dissociation of
6 keV molecular oxygen ions into vanadium metal and cal-
culated using TRIM code is ꢀ6 nm. During oxygen ion
bombardment the sample was kept at room temperature.
The purity of oxygen used (Linde AG) was 99.998%. The
samples with oxide layer were transferred into the analyzer
chamber of the spectrometer under ultrahigh vacuum. The
spectra of V 2p and O 1s electrons were measured at several
different detection angles. The electron analyzer was oper-
ated in the fixed transmission mode. The energy scale of the
spectrometer was calibrated to the Au 4f7/2 binding energy
fixed at 84.0 eV. No static charging of oxide layer was ob-
served; the spectra measured with flood gun on and off
coincided. The estimated error in the determination of
the binding energies was 0.1 eV. Fitting of the V 2p–O
1s spectra was essential for determination of oxidation
states of V present in the oxide layer. This was done after
subtraction of Shirley background using Gaussian–
Lorentzian sum function and XPSPEAK 4.1 software [9].
530
520
510
Binding energy, eV
Fig. 1. Fitted spectra of V 2p and O 1s core electrons taken from clean V
surface (a), after oxidation by molecular oxygen ions (E = 6 keV) with
fluence D = 9.4 · 1014 ions/cm2 (b) and D = 1.7 · 1017 ions/cm2 (c).
Detection angle is 45ꢁ.
[12–15]. The binding energies of the peaks located at
515.8 eV and 517.2 eV are close to those published for bulk
VO2 oxides (515.7–516.2 eV) [12–15] and bulk V2O5 oxides
(516.9–517.2 eV) [12–15], respectively.
The dependence of the population of individual oxida-
tion states of vanadium on the oxygen ion fluence is given
in Fig. 2. It follows from this figure that with increasing
3. Results and discussion
Fig. 1 shows the high-resolution spectra of V 2p–O 1s
electrons of clean vanadium surface and surface oxidized
with oxygen ion beam using low and high ion fluence,
respectively. The fits of the spectra show in addition to V
2p3/2 line located at 512.4 eV and belonging to metallic V
the presence of components centered at 513.6, 515.2,
515.8 and 517.2 eV. Comparison of these values with the
data found in the literature [10–15] allows attributing these
components to V2+, V3+, V4+ and V5+ oxidation states of
vanadium, respectively. The binding energy of the peak lo-
cated at 513.6 eV compares to value 513.5 and 513.7 eV
published for thin films of VO [10–12]. The binding energy
of the peak centered at 515.2 eV is in the range of the
values reported for bulk V2O3 oxides (515.2–515.7 eV)
Fig. 2. The population of different oxidation states of vanadium as a
function of oxygen ion fluence calculated from the fitted XP spectra.
Alov
