Semiconductivity and band gap of Tin-Doped indium tellurate

Semiconductivity and band gap of Tin-Doped indium tellurate

Mat. Res. Bull., Vol. 27, pp. 693-698, 1992. Printed in the USA. 0025-5408/92 $5.00 + .00 Copyright (c) 1992 Pergamon Press Ltd. SEMICONDUCTIVITY AND...

336KB Sizes 0 Downloads 3 Views

Mat. Res. Bull., Vol. 27, pp. 693-698, 1992. Printed in the USA. 0025-5408/92 $5.00 + .00 Copyright (c) 1992 Pergamon Press Ltd.

SEMICONDUCTIVITY AND BAND GAP OF TIN-DOPED INDIUM TELLURATE

B. SHEMIRANI and F. P. KOFFYBERG Department of Physics Brock University St. Catharines, Ontario Canada L2S 3A1 (Received March 17, 1992; Communicated by A.W. Sleight)

ABSTRACT: Tin doped In2TeO6 is prepared at 700° C as a n-type semiconductor. Its carrier mobility is less than 10-3 cm2/Vs. From photo-electrochemical experiments the optical transition across the band gap is determined to be direct forbidden at 1.56 eV. A further indirect allowed transition occurs at 2.93 eV. The electron affinity of In2TeO6 is 4.5 eV. MATERIALS INDEX:

tin, indium, tellurates

Introduction Indium TeUurate, In2TeO6, has previously been prepared as a conducting material by heating at 1200° C under 65 kbar pressure followed by quenching (1); it then has a nearly temperature independent resistivity of --- 3 x 10-2 f2 cm. This material is oxygen deficient, resulting in degenerate n-type conductivity similar to the semiconductivity of In203-x or SnO2-x. The doped material In2Teo.98Reo.0206 was also conducting but with an unexpectedly higher resistivity. The crystal structure of In2TeO6 is isotypic with that of Na2SiF6 (2, 3); both In 3+ and Te6+ are octahedrally coordinated by oxygen. We report here on the semiconductivity of In2TeO6 prepared at ambient pressure and at lower temperature. Under these conditions only properly doped material is conducting, but the conductivity mechanism may be different from that of the high temperature, high pressure material. We also report the results of some photo-electrochemical experiments with doped In2TeO6 as a photoanode, from which we have determined the electron affinity and the optical band gap. 693

694

B. SHEMIRANI et al.

Vol. 27, No. 6

Experimental Samples were prepared from 5N-pure In203, TeO2 and the dopant oxides SnO2, ZrO2 and ZnO. Polycrystalline discs (1 mm thick) were made by mixing, grinding, pressing at 1.4 GPa and then reacting the mixture in air for three days at 650° C followed by four days at 700 ° C and rapid cooling to room temperature. Reaction at 900° C was much faster, but led to noticeable loss of volatile TeO2. All stoichiometric and doped samples had an X-ray diffraction pattern equal to that reported for the material prepared at high pressure. (1) The presence of unreacted oxides, of the indium tellurite In2Te309 (4, 5), or of vitreous material was specifically looked for but was not found. The actual density of the four samples was (5.4 - 5.7) g cm -3, 80% of the theoretical density of 6.89 g cm-3; the average grain size was (20-40) p m. Electrical properties were measured with standard techniques; contacts were copper metal or indium solder. The photo-electrochemical currents were measured in a cell with the sample as anode, Pt as cathode and a saturated calomel electrode (SCE) as reference. Electrolyte was 0.1 M Na2HPO4, PH = 9.1, flushed with H2 gas. Currentvoltage curves were measured potentiostatically, with "white" Xenon arc illumination (0.8 W / c m 2 intensity). The quantum efficiency (q), defined as the ratio of photon-produced holes (measured via the photocurren0 to the incident photon flux was measured using chopped (f=37 Hz) monochromatic light. The photon flux was measured with a calibrated radiometer, and the small photocurrents with a lock-in amplifier. Results and Discussion Electrical Properties The light-yellow coloured u n d o p e d In2TeO6 samples were insulators (p > 106 ~2 cm). Samples in which the In/Te ratio was deliberately changed to 1.95:1 or 2:0.97 were also insulating; apparently the difference in chemical bonding and ionic radii between In 3+ and Te6+ makes it impossible that they can substitute for each other, to an appreciable extent, in the structure and act as donors or acceptors. n-type, pale green, conducting samples were made by doping with Sn (radius Sn4+: 0.69~) or Zr (radius Zr4+:0.72~) for indium (radius In3+:0.80 ~). Doping with Zn (radius Zn2+:0.74A) for In did not give p-type samples, but produced insulators. Qualitatively these results are consistent with In2TeO6 being a wide band gap material similar to In203; when preparation at low temperatures does not give donors related to an oxygen deficiency, then doping with M 4+ for In 3+ produces donor-states by charge compensation. The resistivity of four Inl.99 Sn0.01 TeO6 samples was p =(45-55) f2 cm, and their thermo-electric power Q = - (85-90) laV/K; a Hall effect could not be measured. Comparison of 2-probe and 4-probe resistivity measurements showed that the soldered indium- Inl.99 Sno.01 TeO6 contact resistance was (10-20) ft per cm 2 of contact area; such contacts were non-rectifying. The small Q value indicates that the Fermi level is close, within a few kT, to the conduction band edge. The absence of a measurable Hail effect, given the sensitivity limit of our equipment, sets an upper limit to the Hall coefficient of RH ~ 0.1 cm3/ C. The carrier mobility must therefore

Vol. 27, No. 6

TELLURATES

695

be p = (RH/P) < 2x10-3 c m 2 / V s . Similar results were obtained on samples with doping levels of up to 3.5% Sn. Such a low mobility indicated that in our samples the carriers are localized instead of itinerant, and that conduction takes place by electron h o p p i n g between occupied and unoccupied sites. Since Te atoms can exist in more than one stable valence state, we speculate that localization occurs as Te 5+ ions, formed w h e n Sn 4+ substitute for In 3+, but we have no evidence. The transition from hopping conduction to band conduction takes place in oxides w h e n the order of magnitude value of the mobility is 1,1 > 0.1 cm2/Vs (6). The high temperature material In2TeO6_x of ref. (1) has a low resistivity; assuming it has x < 0.1, then its p > 0.5 cm2/Vs. Its conductivity is essentially i n d e p e n d e n t of temperature for T < 300 K. Therefore it is likely a wide band semiconductor, like In203-x or In2-ySnyO3. The absence of band conduction in our material, prepared at low temperature with an apparently identical crystal structure, is as yet unexplained.

'

'

'

'

t

. . . .

I

'

'

'

'

150

,,<

d

E~ t._

100

?

o

o

i

/

4

J

E^

(4.50) -

O (8 D6)

n

~oo

50

o

V~.

.o

o

-

O,

,

E r

Vf b

0

d

5

.....

Eg

o

6

o"

(llss)

+1

VB

o o° .... 0 -0.5

m

CB

oP o

l

I

0.0

. . . .

I

. . . .

0.5

Potential V (SCE) Fig. 1 The photo current-voltage characteristic of Inl.95Sn0.05TeO6 in 0.1 M Na2HPO4 solution, PH = 9.1

.0

7 (~v)

SCE

Fig. 2 The positions of the conduction and valence band edges on the vacuum-zero and SCE scales; Vfb is positioned neglecting the Helmholtz potential VH.

Electron Affinity and Band Gap The electron affinity (EA) and the band gap Eg were determined from photo-electrochemical experiments. In this technique the semiconductor functions as a p h o t o a n o d e in electrochemical cell. H o l e - e l e c t r o n pairs p r o d u c e d u p o n i l l u m i n a t i o n are separated in the s e m i c o n d u c t o r d e p l e t i o n layer w h i c h is

696

B. SHEMIRANI et al.

Vol. 27, No. 6

established by biasing the semiconductor, against the Pt cathode, beyond its flatband potential VFB. The photo generated holes are driven by the electric field in the depletion layer to the semiconductor-electrolyte interface, where they are used in an oxidation reaction, thus producing a photocurrent ip through the cell.(7) On illuminating the front of the sample with a light pulse, the cell current increased in less than one millisecond from a small value (dark curren0 to a relatively large and steady value; the change is the photocurrent ip, which was found to be stable for hours under constant illumination, ip was proportional to the light intensity; it is shown as a function of the sample potential V in Fig. 1. V is actually measured at the indium back contact of the sample, but corrections for the iR drops over the sample and indium contact are negligible. The nature of the chemical reaction at the indium tellurate-electrolyte interface is unknown; as long as the reaction rate is fast compared to the arrival rate of the photogenerated holes at the sample surface the quantum efficiency q will be determined by the solid state properties of the semiconductor. The potential is which ip is first observed in the onset potential Von; for In2TeO6 in 0.1 M Na2HPO4 (pH = 9.1) we find (Fig. 1) that Von = -0.30 Volt (SCE), with a variation of +0.05 V between samples. Since photocurrents can only occur when the energy bands in the surface depletion layer are bent, Von is equal to the flatband potential Vfb. For a ntype material Vfb (which in the experiment is measured on the SCE scale) is related to the electron affinity EA (the position of the conduction band edge with respect to the vacuum zero energy level) by (8) EA = eVFB -eVH- AEF+ 4.74 (eV)

[1]

Here 4.74 eV is the difference between the zero of the SCE scale and the vacuum zero level and AEF is the position of the Fermi level below the conduction band edge (estimated as --- 0.05 eV from the thermoelectrical power). VH is the Helmholtz potential due to preferential H + / O H - ion adsorption on the semiconductor surface; it is estimated as as -0.15 V for In2TeO6 in contact with PH = 9.1 solution (8). We find EA = 4.5 +0.1 eV for indium tellurate comparable to values for In20 3 (4.45 eV, ref. (9) and SnO2 (4.5 eV. ref. (7). The energy levels are shown in Fig. 2. The quantum efficiency spectrum q (hv) for Inl.95 Sno.05TeO6, measured at +0.70 V (SCE) anode potential, is shown in Fig. 3.; it clearly shows a threshold at --- 1.6 eV. q (hv) is proportional to the optical absorption coefficient ~ ( h v ) if both the depletion layer width and the hole diffusion coefficient are less than the optical absorption length ~-1 (10). For our heavily doped samples both conditions should be satisfied. Then the optical transition energy E0 can be determined (11) from q(hv) by fitting to" q(hv) = K(hv - E0)n / h v [2] The value of n is best determined by forming plots of (qhv) 1/n versus n which are linear for the correct choice of n (12). For h v < 2.9 eV analysis of the q (hv) data gives EO = 1.56 + 0.06 eV; n is found to be 1.5, indicating that the transition across the band gap is direct but forbidden (11).

TELLURATES

Vol. 27, No. 6

''''

I ' ' ' ' I

....

I ....

697

| ....

At larger p h o t o n e n e r g y q ( h v ) i s deconvoluted b y subtracting from the experimental q values the extrapolated q values from the band gap transition, 10-2 and fitting the differences again to eq. [2]. For h v > 3.0 eV an acceptable fit (.) C with E0 = 2.93 + 0.08 eV is obtained; this transition is i n d i r e c t allowed, °~ n=2. In a simple tight-binding a p p r o x i m a t i o n , for o xides without ~ 10_3 (if 0 transition metal ions, the b a n d gap 0 transitions are from a valence band m a d e u p mainly of oxygen-2p wave t) functions to a conduction b a n d made u p m a i n l y of m e t a l ion w a v e functions. For In203, C d I n 2 0 4 a n d 10-4 .... I,,,,I .... I,,, ,I .... In2TeO6 the metal ions all have a 4d 10 1.5 2.0 2.5 3.0 3.5 4.0 c o n f i g u r a t i o n , a n d the c o n d u c t i o n b a n d will h a v e m a i n l y metal-5s p h o t o n e n e r g y (v) character. Their band gap transitions are all relatively small and forbidden; Fig. 3 for In203 : 2.62 eV indirect forbidden The q u a n t u m efficiency (q) spectrum (13), for C d I n 2 0 4 : 2.23 eV direct of Inl.95Sn0.05TeO6; forbidden (14) and for In2TeO6 : 1.56 eV -o-o-o- experimental values; direct forbidden. Even though these ..... calculated due to the band gaps are in the visible or near direct absorption at Eg = 1.56 eV; infrared, the f o r b i d d e n nature of the calculated sum of direct and transitions results in weak ab-sorption indirect absorption with Eo = 2.93 eV. a n d m a k e s the m a t e r i a l s n e a r l y transparent semiconductors. Finally we note that the position D of the valence band edge on the vacuum-zero scale (Fig. 2) is small in In2TeO6: D = EA + Eg = 6.06 eV. For most oxides with mainly ionic binding and without transition metal ions D is found to be 7.0 - 7.6 eV, close to the value of the Millikan electronegativity for the oxygen ion, 7.54 eV (8, 15). The small value for In2TeO6 m a y be due to covalent bonding between the tellurium and oxygen, which m a y have raised the top of the valence b a n d e d g e above the - 7.5 eV position set by the mainly oxygen-2p character of the wave functions of the valence band. Conclusions I n d i u m tellurate p r e p a r e d at low temperature and ambient oxygen pressure is only semiconductive when d o p e d with Sn or Zr, as in Inl.99 Sn0.01TeO6. Its carrier mobility is small; the electrons m a y be localized and conduction m a y be by hopping.

698

B.

SHEMIRANI

et al.

Vol.

27,

No.

6

This is in contrast to In2TeO6-x prepared at high temperatures and pressure which is a wide band conductor with large conductivity. The difference between the two materials, with apparently identical crystal structure, is unexplained. The optical bandgap of indium tellurate is smaller, at 1.56 eV, than the gap in other indium oxides. The direct forbidden nature of the bandgap transition is in common with other indium oxides, and makes these materials nearly transparent semiconductors. By comparison with oxides with ionic binding the valence band of In2TeO6 lies high, at 6.1 eV; this points to covalent bonding between the tellurium and oxygen atoms.

References 1. R.D. Shannon, J. L. Gillson and R. J. Bouchard. J. Phys. Chem. Solids 3__G8877 (1977). 2. S. Natansohn. J. Inorg. Nucl. Chem. 30, 741 (1968). 3. A. Zalkin, J. D. Forrester and D. H. Templeton. Acta Crystallogr. 1_7,71408 (1964). 4. B. Frit and P. Hagenmuller. Bull. SOc. Chim. 1970, 2491. 5. M.J. Redman, W. P. Binnie and W. J. Mallio. J. Less-Common Metals 23, 313 (1971). 6. N. Tsuda, K. Nasu, A. Yanase and K. Siratori. Electronic Conducting in Oxides. Springer Verlag, Berlin (1991). 7. A.J. Nozik. Ann. Rev. Phys. Chem. 29, 189, (1978). 8 M.A. Butler and D. S. Ginley. J. Electrochem. SOc. 125, 228 (1978). 9. E.Y. Wang and Lan Hsu. J. Electrochem. Soc. 25, 1328 (1978). 10. M.A. Buffer. J. Appl. Phys. ~ 1914 (1977). 11. E.J. Johnson in Semiconductors and Semimetals, Vol. 3. Edited by R. K. Willardson and A. C. Beer, Academic Press, New York, (1967) pp. 153-258. 12. F.P. Koffyberg, K. Dwight and W. Wold. Solid State Commun. 30, 433 (1979). 13. R.L. Weiher and R. P. Ley. J. Appl. Phys. 377,299 (1966). 14. F.P. Koffyberg and F. A. Benko. Appl. Phys. Lett. 37, 320 (1980). 15. D.E. Scaife. Solar Energy ~ 41, (1980).