High-resolution Fourier-transform photoluminescence spectroscopy combined with the resonant photoexcitation technique was used to study in detail the Zeeman eff'ect on excitons bound to the neutral donors in high purity epitaxial GaAs. The neutral donor g-factor depends on the magnetic-field intensity and orientation in agreement with the predictions of the five-level k p theory. The neutral-donor excited-state energies measured in a range of magnetic field from 0 to l2 T are in an excellent agreement with appropriately scaled calculations for the hydrogen atom. The correspondence between transitions having the same initial state but having as a final state either the nuetral-donor ground state (principal transitions) or a neutral donor excited state (two electron satellites) was established using resonant excitation and was verified by the angular dependences of the peak energies. Linear and circular polarizations of the 2p, 2po, and 2p+ two-electron satellites are consistent with the assignment of zero angular momentum to the ground state of the donor-bound exciton and we show that the transition energies for these components can be calculated with a 0.03-meV accuracy over the range from 0 to 12 T.

No abstract

The fine structure in the 0.61-eV photoluminescence band from the deep defect EL2 in semiinsulating GaAs has been studied under uniaxial stress and magnetic field. The results show no deviation from full Td symmetry and hence support the isolated-arsenic-antisite model of EL2. Measurements of both the shift due to the hydrostatic component of the stress as well as the g factor of the final state of the transition confirm the identification of this photoluminescence with EL2. PACS numbers: 71.55.Eq, 78.55.Cr The EL2 defect in semi-insulating GaAs continues to attract much attention from semiconductor researchers. Much of this interest stems from the control that EL2 exerts on the electrical characteristics of as-grown bulk GaAs, but more intriguing is the optically accessible inactive metastable state. l Central to understanding this phenomenon is a greater knowledge of the microscopic structure of the defect. A crucial step in this direction is to determine its symmetry. We present here a study of the recently reported 0.61-eV EL2 photoluminescence (PL) band 2 under external perturbations which provides important new information regarding this question.It is well established that the EL2 defect involves the arsenic antisite, Asoa-3 The remaining question is whether it is an isolated Asc a 4 or an arsenic-antisitearsenic-interstitial pair, Asoa-As/, with C^v symmetry. 5 Both of these models are supported by strong experimental evidence and no clear advantage exists for either on purely theoretical grounds. In order to put the results presented here in perspective, it is necessary to quickly review the strengths and weaknesses of these two conflicting models.At the heart of the support for the isolated-Asca model is the splitting under uniaxial stress of the 1.039-eV no-phonon (NP) line of the intracenter absorption transition. 4 These results have been interpreted as evidence of an A\-*T2 transition in Td symmetry, with a Jahn-Teller relaxation of the final state. 6,7 There was some initial opposition to the isolated-Asca model on the grounds that no mechanism has been proposed by which the isolated Aso a could produce the metastable state 8 but these objections have been overcome by recent calculations. 9,10 On the other hand, the strongest evidence that EL2 is the Asoa-As, pair with C^v symmetry is the optically detected electron-nuclear double-resonance (ODENDOR) results which indicate the presence of an As,-along the [111] direction. 5 This view is supported by electronic structure calculations which show that such a pair can also account for the properties of EL 2. l xThe link between the 1.039-eV NP absorption line and EL 2 is well established 12 and it is agreed that its piezospectroscopy must therefore reflect the defect symmetry.Questions are directed instead at the interpretation of these results, and hinge largely on questions on the nature of the final state of the transition and the related lattice relaxation. 13,14 The relationship between the ODENDOR results and EL2 is, however, somewhat less we...

No abstract

In a previous paper it was shown that thick-source α counting is a simple means for determining Th, U, Th-230 excess, and Pa-231 excess concentrations, and hence sedimentation rates, for deep-sea sediments. Here it is shown that radon escape can lead to inaccurate results if powdered samples are used and that this problem can be overcome by preparing samples as a borate glass before measurement. Glassed samples also permit a novel measurement of the Ra-226 content by measurement of the post-fusion buildup of Rn-222, a determination shown to be necessary for near-surface sediments that have a deficit of Ra-226. It is also shown that Po-210 is lost during the fusion and that this loss can be allowed for in the calculations.The above observations are found on comparisons of measured and calculated α count rates for several Th and U standards and on comparisons with α spectrometry results from four sets of deep-sea core samples.