Optical studies of highly-doped GaAs:C
Infrared reflectivity and transmittance measurements (200=5000 cm^-1) were carried out on heavily-doped GaAs:C films grown by molecular beam epitaxy. With increasing carbon concentration, a broad reflectivity minimum develops in the 1000=3000 cm^-1 region and the one-phonon band near 270 cm^-1 ride...
Main Author: | |
---|---|
Other Authors: | |
Format: | Others |
Published: |
Virginia Tech
2014
|
Subjects: | |
Online Access: | http://hdl.handle.net/10919/28927 http://scholar.lib.vt.edu/theses/available/etd-09102001-091811/ |
Summary: | Infrared reflectivity and transmittance measurements (200=5000 cm^-1) were carried
out on heavily-doped GaAs:C films grown by molecular beam epitaxy. With increasing
carbon concentration, a broad reflectivity minimum develops in the
1000=3000 cm^-1 region and the one-phonon band near 270 cm^-1 rides on a progressively increasing
high-reflectivity background. An effective-plasmon/one-phonon dielectric
function with only two free parameters (plasma frequency Ï p and damping constant
γp) gives a good description of the main features of the reflectivity spectra. The
dependence of effective plasma frequency on hole concentration p is linear. At each doping, the effective-plasmon damping
constant is large and corresponds to an optical hole mobility that is about half the
Hall mobility at that p. Secondary-ion mass spectroscopy and localized-vibrational-mode
measurements indicate that the Hall-effect-derived hole concentration is close
to the carbon concentration and that the Hall factor is close to unity, so that the Hall
mobility provides a good estimate of the actual dc mobility. Also, analysis shows that,
for our highly-doped samples, the observed dichotomy between the dc and infrared
mobilities is not a statistical-averaging artifact of the approximations involved in the
model. The explanation of the small infrared mobility resides in the influence of
intervalence band absorption on the effective-plasmon fit, which operationally defines
that mobility via the effective-plasmon damping.
The optical properties obtained with the use of the effective-plasmon model for
GaAs:C yield a phenomenological, approximate, overall picture of the infrared spectra.
But the neglect of intervalenceband transitions, for this p-type semiconductor,
is shown (in this dissertation) to be a serious drawback of this simple model. In
order to obtain the optical properties of GaAs:C in a model-independent way, and to
attempt to resolve the apparent dc/infrared mobility dichotomy, we made use of a
recently-developed spectroscopic-analysis procedure. Using direct numerical-solution
techniques for the reflectance (R) and transmittance (T) equations of a multilayer
structure, we analyzed our infrared R and T results for highly-doped films having
hole concentrations from 2Ã 10^19 up to 1.4Ã 10^20 cm^-3. The optical properties were
determined for photon energies from 0.07 to 0.6 eV, in which region plasmon (intraband)
and intervalenceband contributions are in competition. Our results for the
optical absorption coefficient resolve two separate peaks located (at high doping) at
about 0.1 and 0.2 eV. (The effective-plasmon model necessarily missed the two-peak
character of the actual absorption spectrum.)
By carrying out theoretical calculations of the intervalenceband (IVB) absorption
processes for our dopings, we identify the peak near 0.2 eV with light-hole
to heavy-hole IVB transitions, and we attribute the lower-energy peak to the hole
plasmon. Our experimental absorption spectra are very well described by a model
combining the intervalenceband contribution to the dielectric function with a plasmon
contribution. The hole-plasmon parameters omega_plasmon and gamma_plasmon that we obtain
for highly-doped p-GaAs yield an infrared mobility which (unlike the too-small IVB-entangled
infrared mobility implied by the use of the usual effective-plasmon model)
is in substantial agreement with the dc mobility. Therefore, in actuality, there is no
dc/infrared mobility discrepancy. The discrepancy implied by the use of the usual,
standard-operating-procedure, effective-plasmon model is a consequence of the inadequacy
of that model for p-type semiconductors exhibiting intervalenceband infrared
absorption.
Raman-scattering measurements were carried out on the GaAs:C films. Only the
phononlike coupled plasmon-phonon mode is observed. The non-occurrence of the
plasmonlike mode is due to the large damping of the hole plasmon and the competition
with strong Raman scattering by intervalenceband transitions among the
heavy-hole, light-hole, and split-off bands. Analysis of the phononlike coupled mode,
within the framework of the wavevector-dependent Lindhard-Mermin dielectric function,
supports the hole properties that we determined by Hall and infrared studies.
Photoluminescence measurements showed that the split-off band also participates in
the photoluminescence of GaAs:C, giving rise to an above-bandgap emission band
corresponding to transitions from the conduction band to the split-off valence band. === Ph. D. |
---|