No abstract

In previous work we have identified a near-gap photoluminescence in Ga0.52In0.48P which exhibits a strong dependence of emission energy on excitation intensity (‘‘moving emission’’) and correlated its presence and strength to conditions of growth. In this work we extend our investigations to the rise and decay lifetimes associated with the moving and nonmoving components of the emission. The two processes proceed simultaneously at the same energy. For the moving emission, the time constants scale approximately linearly with excitation intensity. Decaying luminescence can, in most cases, be well fitted with one or two exponentials with time constants as long as milliseconds. The rising luminescence is typically slower and in some cases has a nonmonotonic first time derivative. These results are discussed in terms of existing models of the microstructure of ordered Ga0.52In0.48P.

Four different metastable paramagnetic centers have been identified in the low-temperature, light-induced electron-spin resonance (ESR) spectrum of glassy As&S3. Two of the centers anneal at significantly lower temperatures than the other two, allowing the line shapes to be partially separated with isochronal annealing experiments. The two centers which anneal at lower temperatures, labeled type-I centers, constitute approximately 15% of the induced spins after long-time irradiation at high intensities ( 100 mW/crn~). These centers consist of a hole on a nonbonding 3p orbital of a sulfur atom (Si) and an electron on an s-p hybridized orbital of an arsenic atom (As&). Similar kinetic behavior suggests that the origin of these two centers is a single event, which may be the breaking of an arsenic-sulfur bond. The type-II centers, which are thermally more stable, represent -85%%uo of the induced spins and are concluded to be due to an electron in a nonbonding 4p wave function on a twofold-coordinated arsenic atom (As») and a hole on a nonbonding 3p orbital of a sulfur atom (Si&).The origin of these centers is suggested to be the breaking of As -As and S -S bonds. The densities of the different spins vary rapidly with the stoichiometry. Interpretation of the kinetic behavior of the type-I and type-II ESR signals suggests the existence of a third intermediate metastable state in addition to the ground state and the excited paramagnetic state. High-intensity () 100 mWcm ) light with energy above the band gap (A,~514.5 nm) creates new structural defects in the glass at densities which exceed 10' cm '. At high temperatures (T)250 K) the shift of the opticalabsorption edge to lower energies, which is known as the photodarkening effect, exhibits different kinetics from the electron-spin resonance. This fact suggests that there exists no one-to-one correlation between these two effects. There is, however, a close parallel at all temperatures between absorption well below the gap (midgap absorption) and the type-I ESR centers.

Calculations of the point-ion and oxygen-dipole electrostatic potentials and Born-Mayer repulsion energies at various sites throughout the rutile unit cell offer an explanation for the strong anisotropy observed in the diffusion of lithium in the crystal. On the basis of these three terms, the difference in the energy of a lithium ion in the §0 § and |0J sites is 0.11 eV, in reasonable agreement with the experimentally determined activation energy for diffusion of Li + along the c axis. The energy barrier in the [110] direction was calculated to be 3.5 eV, thus predicting an anisotropy of 10 30 in lithium diffusion at 300°C. These calculations also predict a possible competition between the two interstitial sites for equilibrium, depending on the size and charge of the interstitial ion. The complete solution to the problem awaits a detailed treatment of the polarization term and lattice distortion due to interstitial ions.

Rutile crystals which were doped by heating in contact with Li, Na, K, Ti, or H 2 were examined with x-band electron paramagnetic resonance at about 2°K. The same spectrum was seen as was previously reported for hydrogen-reduced samples. Experimental results and lattice-potential calculations are used to show that the defect is Ti 3+ located at the J0| interstitial site. The Ti 3+ result from interstitial Ti 4+ trapping conduction electrons below 8°K. Li-doping experiments in particular show that Ti 4+ must be present in §0 § sites in fully oxidized crystals. It is postulated that these Ti 4+ compensate for the numerous trivalent substitutional impurities. The absence of the characteristic spectrum in a sample doped with W 5+ is in agreement with this model.

Experimental results are presented demonstrating that the frequently observed "vacuum reduction" of rutile is, in fact, due to residual vacuum system impurities, and that no measurable reduction occurs in a clean vacuum at temperatures as high as lOOO'C for times of '" 100 h.

No abstract

No abstract