Materials Science Forum Vols. 527-529

Abstract: New lines are observed in the photoluminescence of 6H and 4H SiC epitaxial layers grown in cold wall CVD reactors and doped with phosphorus. These lines are associated with neutral phosphorus donor four particle bound exciton complexes with the phosphorus substituting on both the carbon and silicon sublattices. Assignments are made for the (h) hexagonal and (k) cubic sites of the phosphorus donor substituting on the two SiC sublattices.

Abstract: We report the results of a photoluminescence (PL) study of n-type phosphorus-doped SiC epilayers. The PL spectra consist of a set of sharp lines that are interpreted as excitons bound to the P donor with zero-phonon lines which have photon energies very close to the nitrogen-bound excitons and followed by phonon assisted replicas.

Abstract: EPR spectra originating from phosphorus shallow donors occupying silicon sites in 3C-, 4H-, and 6H-SiC are identified by using CVD grown films in which the interference from the signals from the nitrogen shallow donors is practically absent. Phosphorus donors occupying both silicon and carbon sites are observed in high-energy phosphorus ion implanted semi-insulating 6H-SiC which was also free from the interference from the signals from the nitrogen shallow donors.

Abstract: 4H-SiC implanted with high dose of phosphorus has been shown to exhibit lower sheet resistance than 4H-SiC implanted with high dose of nitrogen. In this paper, we have implanted various doses (1x1014cm-2, 2x1014cm-2, 1x1015cm-2 and 4x1015cm-2) of phosphorus into 4H-SiC in order to extract the ionization energy of phosphorus in 4H-SiC as a function of the doping concentration. Variable temperature Hall effect measurements were performed in the temperature range from 60-600K. Least square fits using the charge neutrality equation with two donor levels were used to extract the ionization energies and donor concentrations from the measured data. The ionization energies for both, the hexagonal (53meV, 49meV and 26meV) and the cubic (109meV, 101meV and 74meV) site decreased as the donor concentration (5x1018cm-3, 9.8x1018cm-3 and 3.4x1019cm-3) increased.

Abstract: The procedure of fitting the spectra associated with donor-acceptor pair luminescence arising from nitrogen-aluminum and phosphorus-aluminum pairs in 4H SiC is described in detail. We show that the fitting can be used not only for accurate evaluation of the ionization energies of the different donors and acceptors involved, but also for unambiguous determination of their lattice sites.

Abstract: Ab initio supercell calculations have been carried out to investigate the doping of phosphorus in chemical vapor deposited (CVD) SiC layers. CVD conditions have been simulated by using the appropriate chemical potentials for hydrogen and phosphorus (P). We found that the most abundant defect is P at Si-site followed by P at C-site. The calculated concentrations of the P-donors and free carriers in CVD grown SiC agree with the experimental results.

Abstract: We have employed low-temperature photoluminescence to estimate the total residual N concentration in semi-insulating (SI) SiC substrates where all N shallow donors are compensated in the dark. The ratio of the nitrogen-bound exciton line (Qo) to the free excitonic emission (I77) as a function of excitation power density (Pexc) was tracked for several SI 4H-SiC samples with varying residual N concentration (~ 7x1014 – 5.2x1016 cm-3). Most notably, a linear relationship was found between Qo/I77 and [N] for [N] < 1x1016 cm-3 while a sub-linear behavior was observed for samples with higher N levels. This technique should be particularly valuable to map [N] where the levels are close to or below the present SIMS detection limit of ~ 5-7 x 1014 cm-3. Results obtained for a limited number of low n-type and SI 6H-SiC substrates are also presented.

Abstract: A multiple data point version of the industry standard, two data point raster-changing procedure is employed to measure low levels (< 1 x 1017 atoms/cm3) of nitrogen (N) in silicon carbide (SiC) by SIMS (Secondary Ion Mass Spectrometry). A current-changing procedure is also employed. Together, these are used evaluate the assumptions of the standard method, to separate and measure the components of background signal, and to improve upon the precision and accuracy of the standard method. The risk of poor precision in the two-point method is demonstrated, as is the improvement provided by the multiple-point method. Results show that, in addition to the wellknown N memory background, adsorption background can contribute significantly to the N signal. In general, establishing the presence of adsorption gas in this way can be used to warn of the presence of ionization background, which is not measurable per se.

Abstract: Kinetic mechanisms for the deactivation of nitrogen are investigated by ab initio theory. We find that the interaction of nitrogen with self-interstitials can lead to a deactivation of nitrogen, yet it cannot explain the experimentally observed nitrogen deactivation at high temperatures in silicon co-implanted samples. Our analysis suggests the aggregation of vacancies at high temperatures and the subsequent formation of passive nitrogen-vacancy complexes as a likely explanation.