X-ray photoelectron spectroscopy of alkali tellurite glasses

X-ray photoelectron spectroscopy of alkali tellurite glasses

JOURNA ELSEVIER I, O F Journal of Non-Crystalline Solids 211 (1997) 64-71 X-ray photoelectron spectroscopy of alkali tellurite glasses Y. Himei a,...

603KB Sizes 0 Downloads 9 Views

JOURNA

ELSEVIER

I, O F

Journal of Non-Crystalline Solids 211 (1997) 64-71

X-ray photoelectron spectroscopy of alkali tellurite glasses Y. Himei a, y . Miura a,*, T. Nanba a, A. Osaka b a Department of Environmental Chemistry and Materials, Faculty of Environmental Science and Technology, Okayama University, 2-1-1 Tsushima Naka, Okayama-shi 700, Japan b Department of Bioengineering Science, Faculty of Engineering, Okayama University, 3-1-1 Tsushima Naka, Okayama-shi 700, Japan

Received 27 May 1996; revised 19 September 1996

Abstract X-ray photoelectron spectra of R20-TeO 2 (R: Li, Na, K, Rb and Cs) glasses were measured, using a fresh surface fractured in an ultra high vacuum ( = 7 × 10 -8 Pa) and irradiated with a monochromatic A1 K a X-ray ( h v = 1486.6 eV). The O ls photoelectron spectra showed only a single Gaussian-Lorentzian peak and the peak shifted toward smaller binding energy with increase in Lewis basicity of oxide ions in the glasses. Two peaks attributed to BO and NBO were not observed. In the near valence band spectra for the lithium tellurite glasses, the spectral profile gradually became similar to that of a Li2TeO 3 crystal with increase in LizO content up to 30 mol% Li20. This variation of the profile is correlated to the change in the coordination structure of the tellurium atoms (TeO4 trigonal bipyramids ~ TeO 3 trigonal pyramids) with the addition of the alkali oxides. PACS: 42.70.Ce; 73.20.At; 79.60.-1

1. Introduction Tellurite glass is one of the candidates for new optical materials [1] because of its superior properties, such as wide band infrared transmittance, large refractive index and large third-order non-linear optical susceptibility [1,2]. These interesting properties of tellurite glasses are possibly due to its anomalous network and electronic structures. The network structure of the tellurite glass has been investigated by various spectroscopic methods such as infrared [3-5],

* Corresponding author. Tel.: +81-86 251 8100; fax: + 81-86 253 5755; e-mail: [email protected]

Raman [5-9], nuclear magnetic resonance [10] and X-ray absorption spectroscopy [11-13] as well as diffraction methods such as X-ray [13,14] and neutron diffraction [15,16]. From these studies, it is concluded that TeO 4 trigonal bipyramids (tbp) change to TeO 3 trigonal pyramids (tp) via TeO3+ , polyhedra in the glasses with addition of network modifying oxides; TeO 4 tbp has a lone pair of electrons at one of the equatorial sites of the Te sp3d hybrid orbitals and TeO 3 tp has also a lone pair of electrons at one of the Te sp 3 hybrid orbitals. Then, the electronic structure o f tellurite glass, discussed by Kim et al. [2], is based on the bond-orbital theory in which an influence of cationic empty d-orbitals on the non-linear optical properties was taken into account.

0022-3093/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved. PII S0022-3093(96)00628-X

Y. Himei et al. / Journal of Non-Crystalline Solids 211 (1997) 64-71

The X-ray photoelectron spectroscopy (XPS) is an effective technique for the analysis of the electronic structure of atoms. Since the study performed by Briickner et al. [17], XPS has been widely used for the bonding state analysis of anions and the structural analysis in various glass systems (oxide [17], halide [18,19] and chalcogenide [20]). In this study, X-ray photoelectron spectra for R 2 0 - T e O 2 (R: Li, Na, K, Rb and Cs) glasses were measured. Electronic structures of the binary tellurite glasses and the bonding states of T e - O bonds are discussed.

2. Experimental procedure The glass compositions investigated were x L i 2 0 • ( 1 0 0 - x)TeO 2 (x = 15-30 mol%), xNa20 • (100 - x ) T e O : (x = 10-35 mol%), x K 2 0 • ( 1 0 0 x)TeO 2 ( x = 10-25 mol%), xRb20 • ( 1 0 0 x)TeO 2 (x = 10-20 tool%) and x C s 2 0 . ( 1 0 0 x)TeO 2 (x = 5-12.5 mol%). Appropriate mixtures of reagent-grade a-TeO 2 and Li2CO 3, NazCO 3, K2CO3, Rb2CO 3 or Cs2CO 3 were melted in a platinum crucible at 750°C for 20 min in air and the melts were poured into stainless steel molds to form rod-shaped samples approximately 3 × 3 × 30 mm in size. The quenched glasses were annealed at T g - 20°C (Tg: glass transition temperature) for 30 min. The pure TeO 2 glass (x = 0 mol%) was prepared by rapid quenching by immersing the bottom of the platinum crucible containing the a-TeO 2 melt ( = 0.2 g) into ice water. A Li2TeO 3 crystal was obtained by gradually cooling the corresponding melt in the furnace and was identified by X-ray diffraction (JCPDS: 26-1192). XPS measurements were carried out with an SProbe ESCA SSX100S (Fisons Instruments, UK) using a monochromatic AI K a radiation (h v = 1486.6 eV). The fresh surfaces of rod-shaped glass samples were analyzed just after being fractured in an ultra high vacuum ( = 7 × 10 -8 Pa). The crystal samples were analyzed as powders. Neutralization of the surface charge was performed by setting an electrically grounded Ni mesh screen 1 mm above the sample surface and flooding with low energy ( -~ 7 eV) electrons [21 ]. Bryson [22] reported that the placement of a metal screen 1 to 2 mm above an insulating sample increased the effectiveness of an

65

electron flood gun for charge control in XPS measurements and led to better resolution and reproducibility. This charge control method is effective for the neutralization of differential charging which distorts photoelectron peak shapes [21,22]. The photoelectron spectrometer was calibrated by using Au 4f7/2 binding energy (83.96 eV) for the etched surface of an Au-metal reference sample and then the measured C ls binding energy for adventitious hydrocarbon accumulated on the Au-metal surface in a vacuum was 284.6 eV. Since the C ls peak was hardly detected for the fresh glass surfaces just after being fractured in a vacuum, the correction of binding energy was made by referencing the measured C ls binding energy for adventitious carbon accumulated in a vacuum after about 3 h as 284.6 eV. Peak fitting was performed by using a mixed GaussianLorentzian function for obtaining peak position and full width at half maximum (FWHM). In case of performing curve fitting for Au 4f7/2 peak of an Au-metal reference sample, the best fit was achieved by adopting the Ganssian/Lorentzian ratio of 85/15%. Therefore, this ratio was adopted in the curve fitting for photoelectron peaks of the glasses and crystals.

3. Results Fig. l(a) shows the O ls photoelectron spectra for the sample surfaces exposed in air (not fractured) in the system L i 2 0 - T e O 2. A shoulder is observed on the larger binding energy side of the O 1s main peak and the intensity of the shoulder increases with increase in the Li20 content. Heo et al. [5] measured a similar shoulder on the larger binding energy side of the O ls photoelectron spectrum of 3 0 N a 2 0 . 70TeO 2 sample and they attributed this shoulder (--- 538 eV [5]) to the oxygen atom in TeO 3 tp and attributed the main peak ( - - 5 3 6 eV [5]) to that in TeO 4 tbp. Fig. l(b) shows the O ls photoelectron spectra for the sample surfaces fractured in a vacuum in the system L i 2 0 - T e O 2. In this case, the shoulder on the larger binding energy side disappears and only a single and symmetric Gaussian-Lorentzian peak with a small FWHM (--- 1.6 eV) is observed. The assignment of these O 1s peaks is discussed in Section 4.1. As shown in Fig. l(b), the O ls peak

66

E Himei et al. / Journal of Non-Crystalline Solids 211 (1997) 64-71 xLi20" (100 - x)TeO2 ractured in a vacuum

(a) xLi20- (100 - x)TeO2 exposed in air

Te3ds~

r-

k . 30 ..

~

L.. 25

25

¢/)

~_3o

e-

_

_..

~

~

t,._ 15

:::~..~, X = 15 mol% i

540

20

i

x ~ 0 mol°,~

i

535 530 525 Binding energy / e V

520

595

i

i

|

585

575

565

Binding e n e r g y / e V

Fig. 2. Te 3d photoelectron spectra for the glasses in the system Li20-TeO 2.

(b) xLi20. (100 - x)TeO2 fractured in a vacuum i

30

~

2s

®

20

-

15

!

540

!

535 530 525 Binding e n e r g y / e V

520

Fig. 1. (a) O Is photoelectron spectra for the glass surfaces exposed in air in the system Li20-TeO 2. Solid lines and dotted lines represent the experimental spectra and resolved GaussianLorentzian components, respectively. (b) O ls photoelectronspectra for the glass surfaces fractured in a vacuum in the system Li,O-TeO 2.

shifts toward smaller binding energy with increase in the lithium oxide content in the L i z O - T e O 2 system. Fig. 2 shows the Te 3d photoelectron spectra for the samples in the same system. Doublet peaks attributed to Te 3d3/2 and 3d5/2 are observed and both peaks shift toward smaller binding energy with increase in the L i 2 0 content. Similar phenomena (shape and shift of the peaks) were observed for other alkali tellurite samples. The observed binding energies and full widths at half maximum (FWHM) for the core levels of the O ls and Te 3d5/2 for these samples are

summarized in Table 1. Experimental uncertainties of the binding energies are less than _0.1 eV. The F W H M of the O ls peak was about 1.6 eV and independent of glass compositions (Table 1). Fig. 3 shows the O ls photoelectron spectra for the binary tellurite samples containing 20 tool% alkali oxides. The O ls peak shifts toward smaller binding energy with increasing ionic radius of the alkali ions in the order of Li < Na < Rb. Fig. 4(a), (b) show the O ls and Te 3d5/2 binding energies as a function of the sample composition, respectively. The O ls and the Te 3d5/2 binding energy decrease

20R20.80TeOz fractured in a vacuum

"~

~0

i

540

i

535 530 525 Binding energy / eV

520

Fig. 3. O ls photoelectron spectra for 20R20.80TeO 2 (R: Li,

Na, K and Rb) glasses•

}I. Himei et al. / Journal of Non-Cr3.,stalline Solids 211 (1997) 64-71

530.5

Table 1 Binding energies of XPS O Is and Te 3d5/2 peaks for alkali tellurite glasses and crystals. Experimental uncertainties of the binding energies are less than +0,1 eV Composition

pure TeO 2 glass

Te 3d5/2

530.32(1.59) "

576.11 (1.50) ~

x L i 2 0 • ( 1 0 0 - x)TeO 2 x = 15 mol% 530.13 (1.56) 20 529.98 (1.58) 25 529.95 (1.47) 30 529.82 (1.46)

575.96 575.82 575.79 575.68

(1.54) (1.53) (1.48) (1.45)

x N a 2 0 • ( 1 0 0 - x)TeO 2 x = 10 tool% 530.07 15 529.96 20 529.76 25 529.76 30 529.53 35 529.34

575.92 575.83 575.61 575.63 575.45 575.26

(1.60) (1.49) (I.62) (1.50) (1.57) (1.55)

T.

530,C

•D~

~

O O []

529.5

t• "1o t..Q 529.0 /k [] O • 528.5

TeO=glass

A /k

Li20 NazO K20 RbzO CszO '

0

[]

O

'

'

10 20 30 a20 content x / mol%

576.5 i

>

i-

40

(b)

576.0'

&rq~

2 0 0 []

575.5

• TeOa glass O Li20 O

e.,Q

Z~

[]

/"

575.0 i /k Na20

(1.61) (1.63) (1.58) (1.56)

575.73 575.57 575.45 575.38

(1.54) (1.57) (1.46) (1.52)

x R b 2 0 • ( 1 0 0 - x)TeO 2 x = 10 mol% 530.03 (1.59) 15 529.75 (1.60) 20 529.38 (1.59)

575.82 (1.50) 575.60 (1.53) 575.24 (1.49)

x C s 2 0 • ( 1 0 0 - x)TeO 2 x = 5 mol% 530.12 (1.57) 10 529.86 (1.51) 12.5 529.76 (1.58)

575.95 (I .49) 575.68 (1.44) 575.60 (1.48)

c~-TeO2 crystal Li2TeO 3 crystal

576.12 (1.70) 575.47 (1.63)

a

G)

r-

O Is

x K ~ O . ( 1 0 0 - x)TeO~_ x = 10 mol% 529.90 15 529.74 20 529.55 25 529.53

(a)

>

Binding energy (eV)

(1.64) (1.56) (1.66) (1.50) (1.54) (1.47)

67

530.51 (1.70) 529.66 (1.53)

"o 03 ® I-

[] K20 O Rb20 •

574.5

0

~

CszO

10 20 30 a20 content x / mol%

40

Fig. 4. (a) O Is binding energy as a function of the alkali oxide content. Experimental uncertainty of the binding energy is less than +0.1 eV. (b) Te 3d5/2 binding energy as a function of the alkali oxide content. Experimental uncertainty of the binding energy is less than + 0.1 eV.

ions and the cations. Fig. 5 shows the X-ray photoelectron spectra near the valence band for the samples in the system Li~O-TeO 2. The spectra for

Full width at half maximum in the parentheses.

xLi20.(100- x)TeO2 i t--

monotonically with increase in the alkali oxide content. The binding energy of core level electrons in the alkali cations as well as the O ls and the Te 3d5/2 decreased with increase in the alkali oxide content. The chemical shifts of the O ls binding energy with increase in N a 2 0 content were larger than those of the Te 3d5/2 and the Na ls binding energy in N a 2 0 - T e O 2 system. This difference of the chemical shift is not certainly due to the shift of the standard C ls peak, but due to a change of electron density in the valence shell on the oxide

~

~

L

i

2

T

e

O

3

¢.{-

25

15 5 Binding energy / eV

Fig. 5. X-ray photoelectron spectra near the valence band for the glasses and the crystals in the system L i 2 0 - T e O >

68

E

Himei et al. / Journal of Non-Crystalline Solids 211 (1997) 64-71

a-TeO 2 and Li2TeO 3 crystal samples are also shown in Fig. 5. It is found in Fig. 5 that the profile of the valence band spectrum gradually became similar to that of LiaTeO 3 crystal with increasing L i 2 0 content.

Bo A

t--

2

if/ e"

4. Discussion

2 0 N a 2 0 - 80SIO2

m

4.1. Assignment of the shoulder in the 0 ls peak

20N~O"

As shown in Fig. 1, the shoulder observed in the O 1s spectra for the glass surfaces exposed in air was not detected for the surfaces fractured in a vacuum. We assumed that the shoulder was due to some contamination on the glass surface. Fig. 6 indicates the C ls photoelectron spectrum of the surface exposed in air and that of the surface fractured in a vacuum for 3 0 L i 2 0 . 7 0 T e O 2 glass. Two components of C 1s are observed for the surface exposed in air, while only a single component is observed for the fresh fractured surface. Since the amplitude of the C Is peak around 284.6 eV increased gradually with the passage of measuring time, this peak is attributed to a hydrocarbon accumulated on the glass surface in a vacuum. On the other hand, as shown in Fig. 6, the C ls peak observed at about 289 eV for the surface exposed in air did not increase in the case of the measurement for the fresh surface. It is known that the C 1s binding energy in a carbonate is larger

30Li20-70TeO2

540

80TeO2

!

i

!

535

530

525

520

Binding energy / eV

Fig. 7. O ls photoelectron spectra for 20Na20-80TeO 2 and 20Na20 • 8 0 S i O 2 [21] glass. Solid lines and dotted lines represent the experimental spectra and resolved Gaussian-Lorentzian components, respectively.

than that in a hydrocarbon. The former is at about 289.2 eV (Na2CO 3 [23]) and the latter is at about 284.7 eV ((CH2) . [23]). Therefore, the peak around 289 eV in the C ls photoelectron spectrum for the glass exposed in air can be attributed to the carbon atom in a carbonate which contaminates the glass surface. The O ls binding energy for the oxygen atom in a carbonate (CaCO3:530.9 eV [23]) and a hydroxide (AI(OH)3:533.2 eV [23]) is larger than that in ordinary oxide [23]. Consequently, the shoulder measured on the larger binding energy side of the O 1s peak was assigned to the oxygen atoms in carbonates or hydroxides which cover the sample surface and not to the oxygen in TeO 3 tp as claimed by Heo et al. [5].

4.2. Profile of 0 ls photoelectron spectra ""

exposed in a i r

295

i 290

i 285

i 280

i

275

Binding energy / e V

Fig. 6. C ls photoelectron spectra for the surface exposed in air and that for the surface fractured in a vacuumfor 30Li20 • 70TeO2 glass.

As shown in Fig. l(b) and Fig. 3, only a single and symmetric Gaussian-Lorentzian peak was observed in the O ls photoelectron spectra for the alkali tellurite glasses. In the case of alkali silicate glasses [17,21] two components attributed to bridging oxygen (BO) and non-bridging oxygen (NBO) were observed. Fig. 7 shows the O ls photoelectron spectra for 2 0 N a 2 0 • 80TeO 2 and 2 0 N a 2 0 • 80SIO 2 glass [21]. The O ls binding energy for the alkali tellurite glasses is considerably smaller ( = 2 eV)

E Himei et al./Journal of Non-Crystalline Solids 211 (1997) 64-71

than that for BO atoms of the silicate glass. In general, the O Is binding energy for alkali tellurite glasses is small and close to that of the NBO for alkali silicate glasses, which implies that the electronic density of the valence shell on the oxide ions in the tellurite glasses is larger. We assume that some of the electrons of the lone pair on the Te atom in TeO n polyhedron may be donated to the ligand oxide ions through the T e - O o- bonds, thus the electronic density of the valence shell on the oxide ions in the alkali tellurite glass becomes larger. From the Raman spectroscopic study for alkali tellurite glasses, Sekiya et al. [7] reported that TeO 3+ 1 polyhedra and TeO 3 tp units with including NBO atoms were formed in the glasses with addition of the alkali oxides. Tatsumisago et al. [9] reported that the Tg decreased with addition of Li20 in the system L i 2 0 - T e O 2 due to the cleavage of the networks formed by TeO 4 tbp units and the increase of NBO atoms in the glasses. As mentioned above, though both BO ( T e - O - T e ) and NBO ( T e - O - R ) atoms should certainly exist in the alkali tellurite glasses, only a single component was observed in the O 1s photoelectron spectra and NBO atoms could not be resolved from BO atom. The difference of the O 1s binding energy between BO and NBO in a-TeO: and Li2TeO 3 crystal, respectively, is smaller ( = 0.9 eV (Table 1)) than that in 2 0 N a 2 0 . 8 0 S i O 2 glass ( = 2.1 eV [21]), but the increase in the FWHM and the asymmetry of the O I s peak should be observed with the addition of alkali oxides. However the FWHM of the O 1s peak for the alkali tellurite glass was about 1.6 eV and independent of the glass compositions (Table 1). We suggest that the electronic density of the valence shell on NBO is almost equal to that on BO atom. Here, a theoretical explanation of equalization of the electronic density in the valence shell between BO and NBO is to be proposed. It is based on one of the ideas of the chemical bonding that pTr-dTr bonds are formed between O 2p and empty Te 5d orbitals. Grutsch et al. [24] reported variation of Sn 3d5/2 binding energy by forming 7r back-bonding between tin atoms and ligands in the tin compounds. Blackburn et al. [25] suggested that o- and ~- bonds were formed between phosphorus and oxygen atoms with the cr bond involving P ~ O electron donation and the ~r bond involving O --* P donation in P - O bonds.

69

Similarly the electrons on the NBO atoms in TeO, might be back donated to Te atoms and the electrons might be delocalized through pTr-dTr bonds. As a result, the electronic density of the valence shell between BO and NBO atoms is equalized to give only one component of the O ls peak (Fig. l(b), Figs. 3 and 7). In contrast, two components attributed to BO and NBO atoms were observed in O ls photoelectron spectra for N a 2 0 - S i O 2 glasses. This effect is because the S i - O bonds in silicate glasses have less 7r bonding character than the T e - O bonds in tellurite glasses and lone pair electrons are not donated to Si atoms and are localized on NBO atoms. As shown in Fig. 4(b), the Te 3d5/2 binding energy decreases as well as the O 1s binding energy in Fig. 4(a) with increasing alkali oxide content. The shift in the Te 3d5/2 binding energy can be explained by the increase in the extent of the pTr-dTr back donation to reduce the charge separation between tellurium and oxygen atoms in T e - O bonds. Recently, Uchino and Yoko [26] calculated the electronic states of TeO 2 glass (H6Te207 cluster composed of two TeO 4 tbp units) by ab initio molecular orbital method at the Hartree-Fock/3-21G+ d(Te) level. They suggested that the empty Te 5d orbital hardly participated in the bonding, but that three-center bonds were formed between one tellurium and two oxygen atoms at axial sites (Oaxia0 in a TeO 4 tbp unit and the electrons were delocalized in the Oaxial-Te-Oax~aI bonds. It is also probable that the electrons are delocalized in the N B O - T e - B O bonds by forming the three-center bonds and, as a result, a single component is observed in the O ls spectrum.

4.3. Chemical shifts of the core electron binding energies The O ls and the Te 3d5/2 peaks shifted toward smaller binding energy as the ionic radius of the alkali ions increases (Li ~ Na ~ K ~ Rb ~ Cs) as well as the alkali oxide content (Fig. 4(a), (b)). These chemical shifts are explainable by a change in Lewis basicity of oxide ions in the glasses. The basicity of oxide ions is assumed to indicate effective electronic density of the valence shell on oxide ions which interact with cations. As mentioned before, O 1s binding energy is influenced by electronic density of the valence shell on oxide ions. So, some

70

Y. Himei et al. / Journal of Non-C~stalline Solids 211 (1997) 64-71 530.5

(a)

> (9

~ , 530.0 II) ell)

o~ 529.5 .=_ "O ¢-

"Q 529.0 0~

8

r-] Z12~ • O /k [] O •

528.5 0.35

TeO2 glass Li20 Na20 K20 Rb20 Cs20

0.4

0.45

O

A

0.5

0.55

0.6

Optical basicity Ac.~/ 576.5

>

(b)

(9

576.0 e575.5 "O t'.,Q

~, 575.0 "O 03

• O A [] O



1--

574.5 0.35

TeO2 glass LizO Na20 K20 Rb20 Cs20

0.4

0.45

[]

L~

O

A

0.5

0.55

0.6

Optical basicity Ac~ / -

Fig. 8. (a) Relation between core electron O Is binding energy and optical basicity Ac,I. (b) Relation between core electron Te 3d5/2 binding energy and optical basicity A~,I, correlation between the O ls binding energy and the basicity of oxide ions is expected. Fig. 8(a), (b), respectively, show the O ls binding energy and Te 3d5/2 binding energy as a function of the optical basicity which represents the average Lewis basicity of oxide ions in a matrix. Duffy and Ingram [27] have given an empirical expression of the optical basicity, which is given by Eq. (1):

Zir, Acal = E -

i 2%

(I)

Here, Ac,t is the calculated optical basicity, Z i is the oxidation number of the cation i, r~ is its ionic ratio with respect to the total number of oxide ions and Ti is the basicity moderating parameter related to Pauling's electronegativity Xi [28] by Eq. (2): Yi = 1.36( X i - 0.26).

(2)

Kawazoe [29] reported a correlation between the chemical shifts in Keel, 2 emission for Mg 2÷ and

A13+ in various oxide glasses and the optical basicity. On the basis of the correlation, he concluded that 6-coordinated Mg 2÷ and A13+ predominated over 4-coordinated ones in less basic matrix and it became possible to evaluate the coordination structure of Mg 2÷ and A13÷ in the oxide glasses [29]. A relatively linear correlation between the binding energy and the AcaI is found in Fig. 8(a), (b) for the alkali tellurite glasses. The increase in the optical basicity, Aca1, is assumed to indicate an increase in the effective electronic density of the valence shell on oxide ions which can interact with a cation. Therefore, the increase in the covalency of T e - O bonds due to the increase in Lewis basicity of oxide ions, i.e. the extent of electron donation from oxygen to tellurium atoms has been reflected in the structural change from TeO 4 tbp to more covalent TeO 3 tp with addition of alkali oxides in the glasses. The average T e - O interatomic distance of the TeO 3 tp unit in Li2TeO 3 crystal (0.188 nm) [30] is smaller than that of the TeO4 tbp unit in a-TeO 2 crystal (Te-Oequatorial: 0.1903 nm, Te-Oaxial: 0.2082 nm) [31]. In L i 2 0 - T e O 2 glasses, the T e - O interatomic distance decreases with increase in the Li20 content [ 12,13]. Therefore we suggest that the covalency of T e - O bonds increases with addition of the alkali oxide. However, the calculated optical basicity, Ac, ~, is defined on the assumption that all of oxide ions have an identical chemical bonding state in the matrices and is independent of the local structure around the tellurium atoms. The experimental optical basicity is obtained from frequency shifts of Is 0 --->3P1 transition in the ultraviolet spectra of the probe ions T1 +, Pb 2+ and Bi 3÷ [27] and it reflects the basicity of individual oxide ions in the matrices. But this basicity cannot be applied to the tellurite glasses because of their opaqueness to ultraviolet light. On the other hand, the O Is binding energy which is determined by the electronic density of the valence shell on oxide ions seems to be better and a more universal index of the basicity. 4.4. Valence band spectra

The valence band spectra for the pure TeO 2 (x = 0 mol%) sample was similar to that of c~-TeO2 crystal except for disagreement in the depth of a dip near 6 eV (Fig. 5). Since the pure TeO 2 glass has the same

E Himei et al./Journal of Non-Crystalline Solids 211 (1997) 64-71

basic structural unit, TeO 4 tbp, as a - T e O 2 crystal [6], the electronic states must be similar to each other. Suehara et al. [32] calculated the electronic structure of the TeO4+ 2 (deformed TeO 6 octahedra) cluster in c~-TeO 2 crystal by the self-consistentcharge discrete-variational Xce (SCC-DV-Xc~) method and the calculated valence band spectrum was in good agreement with the experimental one for c~-TeO 2 single crystal. According to their assignments, the band at about 22 eV is mainly composed of O 2s orbital and involves a small amount of Te 5s and Te 5p orbitals. The band at about 12 eV consists mainly of Te 5s orbital admixed with O 2s antibonding states, and the band at about 8 eV consists of the bonding orbitals between Te 5p and O 2p containing a small amount of O 2s antibonding state. The band at about 3 eV is mainly composed of O 2p orbital. Therefore, the variation of the valence band spectrum may be due to the hybridization in the bonding orbitals, i.e. the coordination change of tellurium atoms (TeO 4 tbp ~ TeO 3 tp) with addition of the alkali oxide. However Suehara et al. [32] took no account of the Te 5d orbital. Thus, necessity for more detailed understanding is the molecular orbital calculation involving the Te 5d orbitals, which is under investigation.

5. Conclusions The single symmetric Gaussian-Lorentzian peak observed for O l s photoelectron spectra of alkali tellurite glasses, is attributed to an equalization of the electronic density in the valence shell of bridging oxygen (BO) and non-bridging oxygen (NBO). The shift of O Is and Te 3d toward smaller binding energies is correlated with the increase in Lewis basicity of oxide ions in the samples. In the valence band spectra, the variation of the spectral profile indicated a change in the coordination structure of tellurium atoms (TeO 4 trigonal bipyramids ~ TeO 3 trigonal pyramids) with addition of the alkali oxides.

References [1] H. Nasu, T. Uchigaki, K. Kamiya, H. Kanbara and K. Kubodera, Jpn. J. Appl. Phys. 31 (1992) 3899. [2] S.H. Kim, T. Yoko and S. Sakka, J. Am. Ceram. Soc. 76 (1993) 2486.

71

[3] N. Mochida, K. Takahashi, K. Nakata and S. Shibusawa, Yogyo Kyokai Shi 86 (1978) 317. [4] T. Yoko, K. Kamiya, K. Tanaka, H. Yamada and S. Sakka, Nippon SeramikkusuKyokai GakujutsuRonbunshi97 (1989) 289. [5] J. Heo, D. Lam, G.H. Sigel, Jr., E.A. Mendoza and D.A. Hensley, J. Am. Ceram. Soc. 75 (1992) 277. [6] T. Sekiya, N. Mochida, A. Ohtsuka and M. Tonokawa, Nippon SeramikkusuKyokaiGakujutsuRonbunshi97 (1989) 1435. [7] T. Sekiya, N. Mochida, A. Ohtsuka and M. Tonokawa, J. Non-Cryst. Solids 144 (1992) 128. [8] Y. Himei, A. Osaka, T. Nanba and Y. Miura, J. Non-Cryst. Solids 177 (1994) 164. [9] M. Tastumisago, T. Minami, Y. Kowada and H. Adachi, Phys. Chem. Glasses 35 (1994) 89. [10] T. Yoko, M. Fujita, F. Miyaji and S. Sakka, Chem. Expr. 5 (1990) 549. [11] A. Osaka, Q. Jianrong,T. Nanba,J. Takada, Y. Miura and T. Yao, J. Non-Cryst. Solids 142 (1992) 81. [12] H. Yamamoto, H. Nasu, J. Matsuoka and K. Kamiya, J. Non-Cryst. Solids 170 (1994) 87. [13] Y. Shimizugawa, T. Maeseto, S. Suehara, S. Inoue and A. Nukui, J. Mater. Res. 10 (1995) 405. [14] G.W. Brady, J. Chem. Phys. 27 (1957) 300. [15] S. Neov, V. Kozhukharov,I. Gerasimova,K. Krezhov and B. Sidzhimov, J. Phys. C: Solid State Phys. 12 (1979) 2475. [16] K. Suzuki,J. Non-Cryst. Solids 95-96 (1987) 15. [17] R. Brtickner, H.-U. Chun and H. Goretzki, Glastech. Ber. 51 (1978) 1. [18] R.M. Almeida, J. Lau and J.D. Mackenzie, J. Non-Cryst. Solids 69 (1984) 161. [19] K. Kadono, K. Kinugawa and H. Tanaka, Phys. Chem. Glasses 35 (1994) 59. [20] J.S. Sanghera, J. Heo and J.D. Mackenzie, J. Non-Cryst. Solids 101 (1988) 8. [21] S. Matsumoto, Y. Miura, T. Nanba and A. Osaka, Proc. 17th Int. Congress on Glass, Vol. 3 (1995) p. 72. [22] C.E. Bryson, III, Surf. Sci. 189-190 (1987) 50. [23] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder and G.E. Muilenberg, eds., Handbook of X-ray Photoelectron Spectroscopy (Perkin-Elmer,Eden Prairie, Minnesota, 1978). [24] P.A. Grutsch, M.V. Zeller and T.P, Fehlner, Inorg. Chem. 12 (1973) 1431. [25] J.R. Blackburn, R. Nordberg, F. Stevie, R.G. Albridge and M.M. Jones, Inorg. Chem. 9 (1970) 2374. [26] T. Uchino and T. Yoko, N. Glass 10 (1995) 41. [27] J.A. Dully and M.D. Ingram, J. Non-Cryst. Solids 21 (1976) 373. [28] L. Pauling, The Nature of the Chemical Bond, 3rd Ed. (Cornell UniversityPress, New York, 1960) ch. 3. [29] H. Kawazoe, J. Non-Cryst. Solids 42 (1980) 281. [30] V.F. Folger, Z. Anorg. Allg. Chem. 411 (1975) 103. [31] O. Lindqvist,Acta. Chem. Scand. 22 (1968) 977. [32] S. Suehara, K. Yamamoto, S. Hishita and A. Nukui, Phys. Rev. 50 (1994) 7981.