SOVIET PHYSICS JOURNAL

79

ELECTRICAL CONDUCTIVITY OF BARIUM TITANATE A. Yu. Kudzin Izvestiya VUZ. Fizika, Vol. 10, No. 7, pp. 124-126, 1967 Much attention is given to the investigation of the conductivity of titanium-containing materials. In particular, a great deal of research has been devoted to investigation of the electrical conductivity of barium titanate. So far, however, the nature of its conductivity has remained unexplained. Serious objections have been raised [1] to the common view that it is electronic in nature. The results from conductivity measurements by various authors vary widely. To a large extent this arises from the fact that titanium-containing compounds show-strong changes in their electrical properties under the influence of an applied field (aging) [2], and their conductivity depends on the method of preparation [3] and on the method of application and material of the electrodes [4]. It was shown earlier [5] that the electrical conductivity of barium titanate single crystals depends largely on the heat treatment. Crystals subjected to prolonged annealing at high temperature in air (5-7 hr, 1O00 ~ C) have a conductivity which considerably exceeds that of uuannealed crystals. Moreover, as a result of heating, the activation energy of the charge carriers in the crystals decreases, strong deviation from Ohm's law is found, and the current through the specimen increases with time until it reaches a stationary value. In the present work some new results are given from an investigation into the effect of annealing on the electrical conductivity of /7 5

'

\

!

N,

!

ation (Fig. 3). These characteristics are somewhat different for crystals of different thicknesses. The value of the finat conductivity is considerably higher in thin than in thick crystals. It was found that after the action of a strong field at temperatures above the Curie point

Z'min

25 20-/5

\\ v \"\

10 5

I00

150

~00 "~

Fig. 2. Dependence of time for establishment of stationary conductivity value (r) on temperature for crystals of various thicknesses: Field E = 800 V/era; 1) d = 0.09 mm; 2) d = 0.13 ram; 3) d = 0.27 ram; 4) d = 0.4 mm. the erystais had considerable remanent polarization, integration of curves for the depolarization current as a function of time gives the r value of the charge as several tens o~ microeoulombs per square centimeter of crystal surface. Some of the phenomena observed, such as, for example, deviation from Ohm' s can be explained by processes which take place

1

0

03

Of

0.5

ffmm

Law, 18"S~ o ohm "1 cm "1

Fig. i. Dependence of nonlinearity coefficient (I = A V n) on crystal thickne~,s (d): 1) 50 ~ C; 2) 130 ~ C; 3) 180" C, barium titanate single crystals. The measurements were made on single crystals out of a single batch grown from a solution in molten potassium fluoride. They had various thicknesses. They were first annealed in air for 5 hr at 1000 ~ C. Most of the results were obtained on specimens with silver electrodes deposited by cathode sputtering. Use of electrodes of other materials (Pt, In, A1, and Te) gave qualitatively similar results. The volt-ampere characteristics of these crystals have the form I = AV n, where I is the current passing through the specimen, V is the applied potential, and A and n are constants for a given specimen. The vaiues of n, which characterize the deviation from linear relationship, vary for crystals of various thicknesses (Fig. 1). The deviation from Oban's Law increases as the crystal becomes thinner. When a constant voltage is applied to the crystal the current increases with time. The current increases quickly at first and then tends towards a saturation value. The time for establishment of the stationary current depends largely on the thickness of the crystal, Measurements at various temperatures showed that this stationary value fails as the temperature increases (Fig. 2), The relationship of electrical conductivity to the applied field at high values has an interesting form. At first the conductivity increases rapidly, but with a field greater than 4 kV/cm it reaches satur-

lO'Z~ 108~ 0

f

d

kV/cm

Fig. 8. Dependence of conductivity on field for crystals of various thicknesses at 150" C: 1) 0.2 mm; 2) 0.27 mm. at the point of contact with the metallic electrode, The vok-ampere relationship found in the experiment can be attributed to injection of charge carriers with the electrodes. However, in this case a fall in current with time should be observed at constant potential (with prolonged

80

IZVESTIYA

action of the field in each case [6]). In our case the current was found to increase with time. Anomalous conductivity in barium titanate single crystals which have been annealed is caused by changes in the crystal. One possible change is the appearance of vacancies in the crystal lattice owing to loss of oxygen. Since oxygen enters the BaTiO s Effect of Heat Treatment on Electrical Conductivity of BaTiO s Single Crystals in a Field of 250 V / c m at 200* C Treatment Etching without heat treatment

o / o h m , em

2.5.10 - m

Heating in air (I000 ~ C, 5 hr) 2.6 10- 7 Heating in KF vapor (1200" C, 5 hr)

1.5- I0 -1~

Heating in KF vapor (1200 ~ C, 5 hr) and etching

1.7.10-10

VUZ.

FIZIKA

Anomalous conductivity is not only found in crystals which have been annealed in a mildly reducing atmosphere or in air but also in crystals annealed in oxygen at atmospheric pressure. At the same time ceramic barium titanate which has been annealed at high temperature (~1400" C) has comparatively low conductivity. That is, there is a considerable difference in behavior between the polycrystalline m a t e rial and single crystal. This fact cannot be explained by losses of oxygen on annealing. The barium titanate crystals were grown from a solution in molten potassium fluoride. They therefore contained a certain amount of postassium fluoride. It is possible that the latter m a y leave the crystals during annealing. The change in composition would then lead to a change in conductivity. To check this assumption crystals were annealed at about 1200" C in potassium fluoride vapor. Since it is difficult to obtain the required vapor pressure of potassium fluoride the crystals were heated in a closed platinum crucible on the bottom of which was a saturated solution of potassium fluoride in barium titanate. With such annealing the electrical conductivity of the crystals changes relatively little (see table). This indicates that the defects which occur in the surface layer during annealing arise from loss of potassium fluoride or one of its components, possibly fluorine.

REFERENCES

lattice as doubly charged ions and leaves as neutral atoms, two free electrons remain in the lattice besides the oxygen vacancy. These electrons can be captured by the oxygen vacancies to form f-centers. After thermal excitation the electrons of the f - c e n t e r s can give rise to conductivity in the specimen. The concentration of defects must be highest in the surface layer of t h e crystal and progressively decrease towards the center. T h e oxygen vacancies are capable of moving in the crystal under the influence of a field [7]. When an electric field is applied to the crystal the defects must move from the surface layer into the depth of the crystal, and as the field grows the number of defects leaving the surface layer increases. Here an increase is observed in the conductivity of the crystal. The saturation value of the conductivity arises from depiction of the defect source in the surface layer. After a certain t i m e a specific distribution of defects corresponding to the given field is established. To explain the nature of the defects it is of interest to determine the sign of their charge. For this purpose one side of the crystal was etched, after which the crystals exhibited unipolar conductivity. Deviation from Ohm' s Law and increase of current with time in this case was only observed when the field applied was of such polarity that positive charges could m o v e from the surface which had not been etched. Oxygen vacancies in barium titanate can have a positive charge.

1. D. D. Glower and R. C. Heckman, I. Chem P h y s , 41, 877, 1964. 2. V. Ya. Kunin and A. N. Tsikin, FTT, 4, 3435, 1962. 3. C. A. Cox and R. H. Tredgold, Brit. J. AppL Phys., 16, 427, 1965. 4. A. Brandwood, O. H, Hughes, I. D. Hurd, and R. H. Tredgold, Proc. Phys. Soc., 79, 1161,1962. 5. E. V. Sinyakov and A. Yu. Kudzin, Izv. AN SSSR, ser. f/z., 28, 731, 1964. 6. I. Hurd, A. Simpson, and R, Tredgold, Proc. Phys. Soc,, 79, 442, 1959. 7. A. Many and G. Rakavy, Phys. Rev., 126, 1980, 1962.

26 July 1966

Dnepropetrovsk State University