Point Defects in Ionic Crystals. I. Methods of Calculating the Electronic Structure

The authors present and describe a new embedded-cluster approach for calculating both the electronic and the spatial structure of point defects in ionic crystals with self-consistent incorporation of the polarisation effects into the eigenvalue spectrum and the total energy of a crystal. For the first time both the electronic structure and the ionic displacements within a large cluster containing a defect are obtained in the framework of the same calculating scheme. It permits one to describe correctly back-coupling between redistribution of the electron density in a defective region and polarisation of the rest of the crystal outside the cluster, which can play a vital role for the study of small polaron effects. Making use of a 45-atom cluster with 175 electrons and the semi-empirical INDO method, the authors have simulated the impurity-induced trapping of a hole in the (Li)0 centre in an MgO crystal. The calculations do not use a priori assumptions and do confirm that one-centre localisation of a hole in the ground state is energetically favourable. They show that polarisation energies can differ considerably for the ground and charge-transfer excited states of the defect. It is also demonstrated that the generally accepted model of the O- hole polaron is oversimplified: a considerable charge redistribution between Li+ and another O- along the (100) direction takes place.

The energy, the elastic dipole tensor and the entropy of point defects in ionic crystals are usually calculated by the Mott-Littleton approach, which treats a single defect in an infinite crystal. The authors suggest that there may be advantages, particularly for the dipole tensor and the entropy, in performing the calculations in periodically repeating geometry. They examine how this 'supercell' method can be used for the energy and the dipole tensor, paying attention to the problem that for charged defects the repeating unit carries a net charge, so that the Coulomb energy is divergent. They test the supercell method on a number of both charged and uncharged defects, and show that the results for the energy and dipole tensor are in close agreement with those obtained by the conventional approach.

Various mechanisms for the formation of point defects in ionic crystals are considered. Their probable importance as a function of the energy of the ionizing radiation and the temperature at which the irradiation takes place is discussed. Complex centers formed from aggregates of simple point defects are reviewed and their occurrence is considered in terms of the expected spatial distribution of the point defects. The effect of high-energy X-rays and protons on the ionic conductivity is dis- cussed in terms of a new interpretation of the conductivity curve as a function of temperature, and it is suggested that radiation-enhanced diffusion may be important in explaining the experimental data.

The subsurface electric potential in oriented single crystals of silver chloride has been directly mapped out by a technique based on the equilibrium distribution of a very small concentration of a charged radiotracer. The potential was found to vary with depth according to the Gouy-Chapman solution of Poisson's equation, with screening lengths in the range 10\char21{}20 nm. The surfaces were all negative, relative to the deep interior, by 0.10\char21{}0.30 V, depending on temperature and surface orientation. The present results on the subsurface potential difference can be employed to resolve the Frenkel-pair formation enthalpy and entropy, as determined from transport measurements, into the formation enthalpies and entropies of the individual point defects comprising the pair. For the silver chloride (110) surface, for example, ${G}_{v}$=0.74 eV and ${G}_{i}$=0.37 eV at 150 \ifmmode^\circ\else\textdegree\fi{}C; the best values obtained to date for ${H}_{v}$, ${H}_{i}$, ${S}_{v}$, and ${S}_{i}$, are 0.77 eV, 0.70 eV, 0.79k, and 9.0k, respectively.

The electrostatic potential which arises from a lattice array of point ions is computed in terms of a Taylor's series expansion for small distances from a lattice site. This expansion gives the change in electrostatic energy when an ion moves in the background of a perfect point ion lattice potential. The Taylor's series coefficients for terms up to fourth order in the ion displacement are evaluated for the NaCl and CaF2 lattice structures.

Any deviation from a perfect periodic lattice or structure in a crystal is an imperfection. There are three main types of imperfections, point, line and plane imperfections1, of which we shall deal only with the point defects in this chapter. The point imperfections encountered in ionic solids are: (i) chemical impurities (ii) vacant lattice sites and (iii) interstitial ions. Real crystals are always imperfect in some respect. In thermal equilibrium in an otherwise perfect crystal a certain number of lattice vacancies or interstitials are always present, because the entropy is increased by the presence of disorder in the structure and at a finite temperature, the equilibrium condition of a crystal happens to be a state of minimum free energy.

Investigation of the electronic structure of point defects in ionic crystals by the cluster scattered wave method with the self-consistent calculation of the lattice distortion and long-range polarization

The Stark effect of localized electronic centres is reviewed by using a few selected examples of impurity centres or simple defects. Important information on the symmetry of centres and on the position of energy levels is obtained through the electric-field-induced absorption changes. Typical results are summarized for the Stark effect on the zerophonon lines and on the broad absorption band of point defects in ionic crystals and of impurities in semiconductors.

Excitation of Plasmons and Interband Transitions by Electrons. By H. Raether. (Berlin, Heidelberg, New York: Springer-Verlag, 1980.) [Pp. 192+subject index.] Price DM 59, – –/US $32.50 hardback. Physics of Superionic Conductors. Edited by M. B. Salamon. (Berlin, Heidelberg, New York: Springer-Verlag, 1979.) [Pp. xii+248+subject index.] Price DM 59, – –/US $32.50 hardback. Electronic and Vibrational Properties of Point Defects in Ionic Crystals. By Y. Farces and M. P. Fontana. (Amsterdam: North Holland, 1979.) [Pp. xix+247+author and subject index.] Price Dfl. 100.00/US $48.75 hardback. Defects and Diffusion in Solids. An Introduction. By S. Mrowec. (Amsterdam, Oxford, New York: Elsevier, 1980.) [Pp. 464+subject index.] Price Dfl. 170.00/US $83.00 hardback. Fluctuation Phenomena. Edited by E. W. Montroll and J. L. Lebowitz. (Amsterdam: North-Holland, 1979.) [Pp. 340+author and subject index.] Price Dfl. 120.00/US $58.50 hardback. Crystal Growth: A Tutorial Approach. Edited by W. Bardsley, D. T. J. Hurle and J. B. Mullin. (Amsterdam: North-Holland, 1979.) [Pp. vi+401+subject index.] Price Dfl. 110.00/US $53.75 hardback. Liquid Crystals. By S. Chandrasekhar. (Cambridge University Press, 1980.) [Pp. x+336+subject index.] Price £8.95 paperback. Light Scattering in Solids. Edited by J. L. Birman, H. Z. Cummins and K. K. Rebane. (New York: Plenum Press, 1979.) [Pp. xxi+523+author and subject index.] Price US $55.00 hardback (camera-ready typescript). Electrostatics 1979. Edited by J. Lowell. (London, Bristol: The Institute of Physics, 1979.) [Pp. 349+author index.] Price £32.00 hardback. Auger Electron Spectroscopy Reference Manual. By G. E. Mcguire. (New York: Plenum Press, 1979.) [Pp. ix+129.] Price US $39.50 paperback. Laser Handbook, Volume 3. Edited by M. L. Stitch. (Amsterdam: North-Holland, 1979.) [Pp. ix+846+author and subject index.] Price Dfl. 250.00/ US $122.00 hardback.

Methods described in a previous paper are applied to the calculation of the electronic structure of the $F$ center in KCl, KBr, KI, and NaCl. Structure of the ions is explicitly taken into account in a region close to the defect site, and an effective-mass method is used in the more distant regions. Dielectric polarization effects are included on the basis of the formulation due to Toyozawa, Haken, and Schottky, but results in best agreement with experiment are obtained by a procedure which deviates somewhat from that formulation, as might be expected also on purely theoretical grounds. It does not appear to be possible to obtain agreement with experiment even for absorption energies if polarization effects are neglected. The calculated energy differences of the $1s$, $2p$, $3p$, and $2s$ states in absorption agree very well with presently available experimental data. Agreement with experiment is fairly good for the $1s$, $2s$, and $2p$ energies in emission. The calculated optical transition matrix elements are very sensitive to the details of the model; those obtained from our best model and wave functions are compatible with experimental results. The results for the spin density at the nearest-neighbor ions are reasonably close to the experimental values.

A mechanism of interaction of dislocations with point defects is suggestedon the basis of a study of deformation-stimulated luminescence (DL)from irradiated KCl. DL is considered to be a result of destruction ofhole colour centres by dislocations. The centres destroyed by dislocationsare responsible for irradiation-induced strengthening of alkali halides.

The paper briefly reviews the development of the atomistic method of calculating the properties of defects in ionic crystals, as it derives from the early work of Mott and Littleton (1938). The physical principles and assumptions of the method are described, although the main emphasis is on the development of the technique. It is pointed out that appeal to the quasi-harmonic approximation has made it possible to extend the technique so that the characteristics of defects can now be successfully obtained as functions of temperature. The insights into atomic transport phenomena which these calculations provide are illustrated by three examples.

It is shown that simple band representations tables can be used to solve the following two problems: 1) to find the irreducible components of the space group representations induced by the irreducible representations of their site symmetry subgroups; 2) to find the irreducible components of the site symmetry group representations subduced by irreducible representations of a space group. The solution of both the problems requires also the induced representation table for the crystallographic point grousp which is given here. The space group D is considered as an example. The simple band representations table for this group is also given. The application of the band representations of space groups in the theory of second‐order phase transitions and in the theory of point defects in crystals is discussed.

Point defects may exist in a crystal on a set of crystallographically equivalent orientations or sites. The response of a crystal containing point defects to an externally applied stress takes the form of 'stress-induced ordering', or preferential alignment of the defects. Recent experiments have used optical and spin resonance techniques to observe this stress-induced ordering directly. This same process also gives rise to anelasticity, e.g. to an internal friction peak. In describing such behaviour, the concept of a point defect as an 'elastic dipole', introduced by Kröner and others, is very useful. The elastic dipole is here defined as a second-rank tensor, the 'λ-tensor', equal to the average strain per mole fraction of defects all aligned in a particular orientation. This definition is then used to develop a thermodynamic theory of stress-induced ordering, which includes the interaction among defects and the possibility of a 'reaction' or interconversion between two species of defects. Applications of the theory are made to defects of various point symmetries in cubic crystals, and a number of specific examples are discussed to illustrate each type of defect symmetry. The usefulness of the theory is that it enables one to calculate the values of the components of the λ-tensor from experimental information and the defect symmetry. Typical values of these λ-parameters range from 10-2 unity.

The method of classifying and enumerating close-packed clusters of vacancies or substitutional solutes in crystals by using representation matrices has been extended to include clusters with more than one type of point defect. Examples are provided for a few simple cases in order to clarify the notation adopted. Results are then presented for clusters of up to four point defects in all single-lattice and selected double-lattice structures. All possible cases are included, from pure clusters with just one type of solute to mixed clusters in which each solute is different. In each case the number of distinct configurations and the total number of variants is recorded. For example, in h.c.p. structures there are 1043 configurations of clusters with four points, each occupied by a different solute, and these have a total of 22 800 variants. In all there are 106 possible decorated topologies for the configurations and these are illustrated. Relationships between the results for different structures are discussed.