Thermal donors in silicon: an investigation of their structure with electron nuclear double resonance
| Autor: | N. Meilwes, Johann Martin Spaeth, W Gotz, G Pensl |
|---|---|
| Rok vydání: | 1994 |
| Předmět: |
Electron nuclear double resonance
Silicon Doping Analytical chemistry chemistry.chemical_element Condensed Matter Physics Oxygen Spectral line Electronic Optical and Magnetic Materials law.invention Crystallography chemistry Impurity law Quadrupole Materials Chemistry Electrical and Electronic Engineering Electron paramagnetic resonance |
| Zdroj: | Semiconductor Science and Technology. 9:1623-1632 |
| ISSN: | 1361-6641 0268-1242 |
| DOI: | 10.1088/0268-1242/9/9/009 |
| Popis: | Singly ionized thermal double donors (TDD+), which have the so-called NL8 EPR spectrum, have been investigated with electron nuclear double resonance (ENDOR) in silicon doped with various acceptors (Al,B,In,Ga). No differences were detected in the ENDOR spectra and no signals due to any of the acceptors. TDDs appear to be heat treatment centres which do not involve any impurities except oxygen. The magnetic isotope 17O was diffused into B-doped oxygen-poor float-zone silicon and TDD+'s were formed with various durations of heat treatment at 470 degrees C up to 200 h. The total intensity of the 17O ENDOR lines increases overproportionally compared to the increase of the 29Si ENDOR lines with longer heat treatment, indicating that more oxygen is added upon TDD+ growth from earlier to later species. The 17O ENDOR lines of the latter have an almost vanishing superhyperfine interaction. Their quadrupole interactions are almost the same as those found in the earliest species detected with ENDOR. From its theoretical interpretation it is suggested that oxygen is always in an interstitial position between two Si atoms. The C2v symmetry appearing in the EPR and 29Si ENDOR spectra does not necessarily reflect the true defect symmetry, which may be lower since the addition of 17O does not noticeably influence the distribution of the unpaired spin density. Therefore the suggestion drawn from earlier measurements on the two early species observable with ENDOR, that four oxygen atoms may be in the defect core, should be revised. It seems that the model proposed by Deak et al. (1991-1992) is compatible with our ENDOR results. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|