CC BY
5Khyzhun O. Yu., Solonin Yu. M.
Institute for Problems of Materials Science, National Academy of Sciences of Ukraine, 3 Krzhyzhanovsky str., UA-03142 Kiev, Ukraine
X 3
X-Ray photoelectron (XPS), emission (XES) and absorption (XAS) spectroscopy methods were used to study the electronic structure of nanoparticles of hexagonal hydrogen tungsten bronze, HxWO3, which is a prospective sensor material. The content of hydrogen atoms, x, in the HxWO3 specimen investigated in the present work was found to be 0.24. For comparison, the electronic structure of the monoclinic and hexagonal modifications of WO3 was also studied. Both the XPS valence-band and the O Ka XES spectra for the HxWO3 and WO3 compounds were derived and compared on a common energy scale. It was established that, half-widths of the O Ka bands and the XPS valence-band spectra increase in the sequence monoclinic WO3 — hexagonal WO3 — H024WO3. The formation of an additional near-Fermi sub-band, which is absent in the both modifications of WO, was observed on the XPS valence-band spectrum of H024WO3. Binding energies of both W 4f and O 1s core-level electrons remain constant (within experimental error) for all the compounds studied. The above fact indicates that charge states of the tungsten and oxygen atoms in the hexagonal H0 24WO3 compound are close to those in the tungsten trioxides. The energy positions of the centers of gravity of the O Ka band do not change in the sequence monoclinic WO3 — hexagonal WO3 — HxWO3. A high-energy shift of the inflection point of the W Lm XAS spectrum when going from metallic tungsten to HxWO3 has been evaluated.
1. INTRODUCTION
Tungsten trioxide, WO3, is of great interest both in technology and in theory in particular because of its ability to incorporate alkali metals, ammonium ions and hydrogen to form tungsten bronzes, AxWO3. Physical and chemical properties of the bronzes can vary dramatically with concentration x, of an incorporating element, A. The hexagonal tungsten bronzes, HxWO3, are substoichiometric compounds and have been characterized in the range 0 < x < 0.46 [1-3]. The bronzes possess metallic conductivity and represent a new and important electrode material. The hexagonal HxWO3 compounds are prospective sensor materials which possess good electrochromic properties [3,4]. Some properties of the substoichiometric HxWO3 compounds can be understood by considering their electronic structure.
To our knowledge the electronic structure of the hexagonal HxWO3 compounds has not been studied either by theoretical band-structure calculations or by experimental methods. The electronic structure of two cubic hydrogen tungsten bronzes, HWO3 and H2WO3, has been investigated recently using a full-potential linear muffin-tin orbital (FP-LMTO) method [5]. Hjelm et al. [5] also used the FP-LMTO method to calculate band structures of the both cubic and hexagonal modifications of tungsten trioxide. It should be noted that the electronic structure of WO3 in the oversimplified cubic perovskite-type structure has been studied theoretically using a non-self-consistent and non-relativistic version of the Korringa-Kohn-Rostoker (KKR) method [6] and a self-consistent and non-relativistic version of the atomic-orbital-based method [7]. In the latter work, the full monoclinic distorted structure of tungsten trioxide was
also investigated. As shown by Bullett [7], the cubic • • monoclinic distortion of WO3 increases both the band gap and the W d-orbital occupation.
The electronic structure of the monoclinic modification of WO3 was studied experimentally in a series of work [8-12]. In the mentioned works, shapes of the valence band and/or O 2p-like emission band of monoclinic trioxide were determined. Nevertheless, the electronic structure of the hexagonal WO3 and H^WO3 phases has not been studied experimentally (see, e. g. [5,13]). Therefore, the purpose of the present work was to carry out a complex experimental study (using the X-ray photo-electron spectroscopy (XPS), X-ray emission spectroscopy (XES) and X-ray absorption spectroscopy (XAS) methods) of the electronic structure of hexagonal hydrogen tungsten bronze, H^WO3. For comparison, XPS and XES measurements for the two different tungsten trioxide phases, with the monoclinic and hexagonal structures, were carried out.
2. EXPERIMENTAL
The technique of the present experimental investigation of the electronic structure of the monoclinic and hexagonal WO3 phases and hexagonal hydrogen tungsten bronze, HxWO3, was analogous to those described elsewhere [11,12]. Therefore, in this paper only the main details of the experiment are reported.
The ultrasoft X-ray emission transition), reflecting the energy distribution of filled valence states of p-symmetry of oxygen, in the studied compounds were obtained using an RSM-500 spectrometer. The energy resolution AE . of the RSM-500 spec-
Fig. 1. On a common energy scale the XPS valence-band spectra (solid curves) and O Ka emission bands (dashed curves) of (1) HxWO3, (2) hexagonal WO3, and (3)
monoclinic WO3.
trometer in the range corresponding to the energy of the O Ka band was found to be better than 0.45 eV. The detector was a secondary electron multiplier VEU-6 with a Csl photocathode. A diffraction grating (600 lines/mm and a radius of curvature of R «=6 m) and a filter mirror (R « 4 m) were used. Both the grating and the mirror were covered with gold (thickness of about 300 E ). Operating conditions of the X-ray tube were the following: accelerating voltage, Ua=5.5 kV; anode current, I =2.5 mA. The O Ka bands were recorded in a
' a
spectrometer chamber having a base pressure less than 5x10-6 Pa. For every specimen studied, 7 to 11 independently derived spectra of the O Ka band were chosen for averaging the results obtained.
The W Ljjj absorption spectra, which reflect the energy distribution of the empty W d-like states, were obtained for hexagonal hydrogen tungsten bronze, H^WO3, and metallic tungsten using the spectrometer with scintillation recording of the X-ray radiation intensity. The spectra were recorded us-
ing the method of la variable field of absorptionl [14,15]. A quartz crystal with the (1340) reflecting plane and a radius of curvature of R=» 500 mm was used. The method for preparing the absorber and selection of the optimum thickness was analogous to that used earlier [16,17]. A BHV-23 X-ray tube with a tungsten anode operating at Ua=14 kV and Ia=40 mA was used as the source of primary excitation. The quantum yield of the X-ray photoeffect in the area of the oxygen K absorption edge in the HxWO3 and WO3 compounds could not be studied, probably as a result of the strong screening of the oxygen atoms by the tungsten atoms [13].
Measurements of the XPS valence-band and core-level spectra of the HxWO3 and WO3 samples were carried out in a ion-pumped chamber of an ES-2401 spectrometer. The chamber was evacuated to 2x 10-7 Pa. The Mg Ka12 (£=1253.6 eV) excitation was used in the capacity of the source of X-ray radiation. The impurity carbon 1s line (285.0 eV) was taken as reference.
The method of synthesis of the specimens and the X-ray diffraction analysis data for the investigated HxWO3 and WO3 samples were reported in Refs. [3,4,18]. The content of hydrogen atoms, x, in the HxWO3 specimen investigated in the present work was found to be 0.24 [19]. The lattice parameters of the samples were the following: a=0.7356 nm, b=0.7612 nm (H0 24WO3); a=0.7276 nm, b=0,7800 nm (hexagonal WO3); a=0.7297 nm, b=0.7539 nm, c=0.7688 nm, /? =90,91° (monoclinic WO3) [18,19].
3. RESULTS AND DISCUSSION
Fig. 1 shows on a common energy scale the XPS valence-band spectra and O Ka emission bands.
Energy positions of the maxima of the spectra are presented in Table 1.
The XPS valence-band spectra shown in Fig. 1 have been normalized to one and the same integral intensity of the XPS W 4f?/25/2 core-level spectra of the corresponding compound. The results of such normalization indicate that the peak intensity of the XPS valence-band spectra remains constant within the experimental error for all the compounds studied. The relative intensities of the O Ka spectra were not measured because there was not a standard X-ray line within the range of energies corresponding to the energy of the O Ka band [13,20]. This is why all the O Ka bands in Fig. 1 were reduced to one and the same intensity of their peaks "b".
Table 1.
Some characteristics (in eV) of the XPS valence-band spectra and O Ka emission bands of the compounds studied.
Compound XPS valence-band spectrum O Ka emission band
Energy position of the maximum a) Half-width Energy position of the maximum a) Half-width
HxWO3 -7.4 6.56 -4.8 5.20
Hexagonal WO3 -7.1 6.03 -4.7 4.67
Monoclinic WO3 -7.1 5.52 -4.5 4.22
Uncertainty ±0.2 ±0.08 ±0.1 ±0.08
a) On a common energy scale.
Khyzhun O.Yu., Solonin Yu.M. Electronic structure of hexagonal hydrogen tungsten bronze HxWO3 nanoparticles, a prospective sensor material.
As can be seen from Fig. 1, the energy positions of the centers of gravity of the O Ka bands do not change in the sequence monoclinic WO3 i hexagonal WO3 i i HjWO3. The main peak "b" of the O Ka band shifts by 2.4-2.6 eV (Table 1) toward higher energies with respect to the position of the main peak IAI of the XPS valence-band spectrum of the corresponding compound. Nevertheless, the comparison on a single energy scale indicates that, for all the compounds studied, the main peak IAI of the XPS valence-band spectrum coincides with low-energy feature Ial of the O Ka band (Fig. 1).
Both the XPS valence-band spectra and O Ka emission bands broaden somewhat in the sequence mono-clinic WO3 i hexagonal WO3 i HiWO3 (Table 1). Our experimental results are in agreement with the theoretical FP-LMTO band-structure calculations [5]. When studying the electronic structure of the tungsten triox-ides and hydrogen tungsten bronzes, Hjelm et al. [5] showed that the O 2p-like band broadens somewhat in the sequence cubic WO3 i hexagonal WO3=i HWO3. According to the theoretical calculations [5-7], for the mentioned compounds, the valence band is dominated by the O 2p-like states, while the W 5d-like states dominate the conduction band. As a result, in the sequence monoclinic WO3 i hexagonal WO3 i HiWOJ one would expect broadening both the O 2p-like bands and XPS valence-band spectra. This is actually observed in the present experiment as mentioned.
As one can see from Fig. 1 (solid curve 1), the formation of an additional sub-band at the Fermi energy, EF, is characteristic for the XPS valence-band spectrum of the hexagonal tungsten bronze HxWO3 nanoparticles. The subband is absent on the XPS valence-band spectra of the both modifications of WO3. According to the results of Refs. [5,19,21], the formation of the near-Fermi sub-band IBI on the XPS valence-band spectrum of H024WO3 can be explained due to filling by the hydrogen electrons the W t2g-like sub-band, which is empty in tungsten trioxide. The formation of the similar near-Fermi sub-band was observed earlier during XPS studies of the sodium tungsten bronzes NaxWO3 in the range 0.5 < x < 0.97 [21]. The relative intensity of the sub-band increased with increasing content x of the sodium atoms in the Na WO bronze [21].
x 3 L J
Fig. 3. XPS O 1s core-level spectra of (1) HWO, (2) hexagonal WOf and (3) monoclinic WO3.
Fig. 2. XPS W4f7/25/2 core-level spectra of(1) HWO, (2) hexagonal WO, (3) monoclinic WO, and (4) Wmet.X
Fig. 1 shows that in the energy region corresponding to the near-Fermi sub-band IBI of the XPS valence-band spectrum of H0.24WO3, the formation of additional fine-structure features was not detected for the O Ka emission band of the compound (Fig. 1, dashed curve 1). The above fact indicates that the O 2p-like states do not take part in the formation of the mentioned sub-band on the XPS valence-band spectrum of the hexagonal H024WO3 nanoparticles. It should be mentioned that our attempts to study the fluorescent W Lb5 band (Lm i O V transition), reflecting primarily the occupied W 5d-like states, in the hexagonal H0.24WO3 tungsten bronze were unsuccessful. After accumulation times of about 190 h, the intensity of the W L band for the hexagonal H0.24WO3 specimen, studied in the present work was too small to be able to discuss its fine-structure features. The method was analogous to those used previously to study the XES W 5d-like bands of tungsten carbides and germanides [16,17].
The XPS W 4f and O 1s core-level spectra for the compounds, studied in the present work, are presented in Figs. 2 and 3, respectively. Fig. 2 shows also the spectrum of the W 4f core-level electrons of pure metallic tungsten.
As can be seen from Fig. 2 and Table 2, in the investigated WO3 and H024WO3 specimens the XPS W 4f core-level binding energies are higher by about 4.6 eV compared with that of pure Wmet. The above fact indicates that the charge transfer occurs from W to O for all the compounds studied.
Table 2 indicates that the both XPS W 4f and O 1s core-level binding energies remain constant within the experimental error for the WO3 specimens and the hexagonal hydrogen tungsten bronze H024WO3. Therefore, the charge states of the both tungsten and oxygen atoms do not change in the sequence monoclinic WO3 i i hexagonal WO3 i H0.24WO3. The charge state of hydrogen atoms in the H0.24WO3 compound could not be investigated experimentally because the hydrogen XPS core-level spectra are absent [22,23].
Fig. 4 shows the W L absorption spectra of the hexagonal hydrogen tungsten bronze H0 24WO3 and pure metallic tungsten. The spectra are normalized so that intensities of their "white" lines are equal. The shape of the W LIII absorption spectra do not exhibit any significant changes when going from Wmet to H024WO3 (Fig. 4).
Table 2.
Binding energies (in eV) of the XPS core-level spectra for metallic tungsten and the compounds studied.
Sample W 4f7/2 5/2 spectrum O 1s spectrum
HxWO3 36.05 530.84
Hexagonal WO3 36.02 530.86
Monoclinic WO3 35.95 530.78
W (metal) 31.40 —
Uncertainty ±0.05 ±0.08
Nevertheless, the inflection point of the W L XAS spectrum of H0.24WO3 shifted by about 2.9 eV toward higher energies with respect to its position on the spectrum of pure metallic tungsten. This fact also indicates the displacement of the electron density in the direction W { O for the hydrogen tungsten bronze nanoparticles studied in the present work.
4. CONCLUSION
The formation of an additional near-Fermi subband, which is absent in the both monoclinic and hexagonal modifications of tungsten trioxide, was observed on the XPS valence-band spectrum of the hexagonal H0.24WO3 bronze. Half-widths of the O Ka bands and the XPS valence-band spectra increase in the sequence monoclinic WO3 { hexagonal WO3 { H0 24WO3. Measurements of binding energies of the both XPS W 4f and O 1s core-level electrons indicate that the charge states of the tungsten and oxygen atoms in the hexagonal H024WO3 nanoparticles are close to those in the mono-clinic and hexagonal tungsten trioxides. The inflection point of the W L XAS spectrum of H024WO3 shifted by about 2.9 eV toward higher energies with respect to its position on the spectrum of pure metallic tungsten. The energy positions of the centers of gravity of the O Ka bands remain constant for all the compounds studied.
Hd 1 /\ 1 \ 1 1 1 / /N.
2_____S J ----
-20 0 20 40
Energy, eV
[2]
3]
4]
Fig. 4. W Lm absorption spectra of (1) Wmet and (2) H WO .
REFERENCES
[ 1 ] Dickens P.G., in: Goodenough J.B. and Whittinghem M.S. (Eds.), Solid State Chemistry of Energy Conversion and Storage, Advances in Chemistry Series 163, American Chemical Society, Washington DC, 1977, p. 165. Drobasheva T.I., Spitsyn V.I., in: Spishyn V.I. (Ed.) Oxide Bronzes, Nauka, Moscow, 1982, p. 40 (in Russian). susic M.V., Solonin YuM., J. Mater. Sci., 1008, Vol. 23, P. 267. Solonin Yu.M., DSc. Thesis, Institute for Problems of Materials Science, National Academy of Sciences of Ukraine, Kiev, 1990.
Hjelm A., Granqvist C.G., Wills J.M., Phys. Rev. B, 1996, Vol. 54, P. 2436.
Kopp L., Harmon B.N., Liu S.H., Solid State Commun., 1977, Vol. 22, P. 677.
Bullett D.W., J. Phys. C: Solid State Phys., 1983, Vol. 16, P. 2197. Colton R.J., Rabalais J.W., Inorg. Chem., 1976, Vol.15, P. 236. de Angelis B.A., Schiavello M., J. Solid State Chem., 1977, Vol. 21P. 67.
10] Bringans R.D., Höchst H., Shanks H.R., Phys. Rev. B, 1981, Vol. 24, P. 3481.
11] KhyzhunO.Yu., Rep. Natl. Acad. Sci., 1999, Vol. 7, P. 86.
12] Khyzhun O.Yu., Metallofiz. Noveishie Tekhnol., 2000, Vol. 22, No. 3, P. 55 (in Russian).
13] Meisel A., Leonhardt G., Szargan R., X-Ray Spectra and Chemical Binding, Springer-Verlag, Berlin/ Heidelberg, 1989.
14] Zaulychny Ya.V., Khyzhun O.Yu., Zhurakovsky E.A., Dobrovolsky V.D., Phys. Met., 1991, Vol. 10, P. 620.
15] Zhurakovsky E.A., Khyzhun O.Yu., Sinelnichenko A.K., Kolyagin V.A., Chuzhko R.K., Dobrovolsky V.D., Phys. Met., 1993, Vol. 12, P. 444.
16] Khyzhun O.Yu., Zhurakovsky E.A., Zaulychny Ya.V., Sov. Powder Metal. Met. Ceram., 1990, Vol. 29, P. 732.
17] Khyzhun O.Yu., Zaulychny Ya.V., Zhurakovsky E.A., J. Alloys Comp., 1996, Vol. 244, P. 107.
18] Solonin Yu.M., Privalov Yu.G., Dokl. AN UkrSSR, Ser. B., 1985, Vol. 1, P. 46 (in Russian).
19] Khyzhun O.Yu., Solonin Yu.M., Rep. Natl. Acad. Sci., 2000, No. 7, P. 82.
20] Blokhin M.A., Shveitser I.G., Handbook on X-Ray Spectra, Nauka, Moscow, 1982 (in Russian).
21] Hochst H., Bringans R.D., Shanks H.R., Steiner P., Solid State Commun., 1980, Vol. 37, P. 41.
22] Siegbahn K., Nordling C., Fahlman A., Nordberg R., Hamrin K., Hedman J., Johansson G., Bergmark T., Karlsson S.-E., Lindberg I., Lindberg B., ESCA: Atomic, molecular and solid state structure studied by means of electron spectroscopy, Uppsala, 1967.
[23] Nefedov V.I., Handbook On X-Ray Photoelectron Spectroscopy of Chemical Compounds, Khimia, Moscow, 1984 (in Russian).
