Solid State Communications,
Vol. 14, pp. 267—269, 1974.
Printed in Great Britain
OSCILLATORY PHOTOCONDUCTIVITY OF GALLIUM ARSENIDE A. Barbaric and E. Fortin University of Ottawa, Department of Physics, Ottawa, Canada (Received 4 October 1973 by M.F. Collins)
The photoconductivity of GaAs in the energy range 1.5—1.8 eV at 4.2 K exhibits two types of quantum oscillations: Landau oscillations for epitaxial material in high magnetic fields or LO phonon oscillations for high resistivity material. The purity and surface treatment of the samples seem to determine the type of oscillations observed.
ONE of the difficulties in observing interband transitions between Landau levels in GaAs arises because of the heavy conduction band electron effective mass 0.067 rn.1 The analysis is further complicated by exciton participation. Vrehen2 presented some magneto-absorption studies and later Reine et al.3 reported a stress-modulated magneto-reflectivity spectrum; both showed oscillatory structures associated with interband transitions between Landau levels in GaAs.
PHOTOCONDUCTIVITY B I
60 kG 42 K
The photoconductivity of epitaxial GaAs in an intense magnetic field reported here also exhibits oscillations associated with interband Landau transitions. Oscillatory photoconductivity has been reported by Nahory in 1968 ‘ in epitaxial GaAs; however, these oscillations were associated to phonon transitions. Such an oscillatory behavioris also observed here but in high resistivity material. The purity and surface
FIG. Photoconductivity spectrum of epitaxial GaAs at 60 I. kG. made of two parts H
treatment of the samples are taken to be decisive parameters in determining the type of oscillations observed, Two materials were studied, an epitaxial and a high resistivity sample designated GaAshad (1) and GaAs (2) respectively. The epitaxial material a carrier concentration of 1.4 X 1012 cm3 with a mobility = 202.000 cm2[V-sec at 77 K. A slow chemical etch (-‘ 0.1 pm/mm), a solution made of 20 parts H 2S04, one part 30% H202 and one part H20 was used room temperature. The high resistivity material at was a chromium doped single crystal with resistivity p = 2 X iO~~2-cm.The etch for this material was a solution
2SO4, one part 30% H202 and two parts H2O used at 70°C; the surface was lapped and mechanically polished with 5~iAl2 03 powder prior to etching. After the etch both samples were chemically reduced with a hydrogen gas and 5sealed by The conan amorphous aluminum oxide (Al2 03) film. tacts consisted of an Sn—In alloy for the epitaxial samplc and simply In for the high resistivity sample. Both were mounted in a Voigt configuration within the experimental chamber of the magnet. A description of 6 the apparatus can be found in previous work. The spectral distribution of the photoconductivity in the epitaxial material for a magnetic field of 60 kG
OSCILLATORY PHOTOCONDUCTWITY OF GALLiuM ARSENIDE
Vol. 14, No.3
GaAs (I) l.50 T~4.2K
GaAs (2) T~4.2K
./ / 1.55
> .! L5
CD ~E .56 w w
// / S ,//// • /~/___5 •
•8 5r’_. . • .~: i •
•v._:_.._.. — • — 5— — 5—
w Z W
z o 10
lqd.rc.pl I 520 SV Slope 00418 •V
~ •—5 -.-.‘.~.:—:--~_.— — — — — — • —.
20 50 40 50 MAGNETIC FIELD
60 C kG)
FIG. 2. Photon energy values at minima vs the magnetic field.
I C I 2 3 4 ORDER NUMBER
FIG. 4. Photon energy value at minima vs the phonon transition order.
_____________________________________________ Ga As C 2) • 4.2 K B
exciton effect giving 3.4 meV as the binding energy of the exciton in GaAs, the same value that was obtained 8 The scatteringof the experimental points by Sturge. arises from the changing shape of the successive spectra as the magnetic field is increased. Such changes remain unexplained at present.
The high resistivity material exhibits quantum oscillations at B = 0 kG and B = 55 kG as shown in Fig. 3. No pertinent change seems to result from the application of a magnetic field. Figure 4 shows the plot of the photon energy for the minima9 at B = 8 • 5590
FIG. 3. Photoconductivity spectra of high resistivity GaAs forB = 0kG and B = 55 kG. is given in Fig. 1. A plot of the photon energy of the minima7 versus the magnetic field is shown in Fig. 2. Their convergence at B = 0 kG yields a value of the direct bandgap energy at 1.520 ± 0.002 eV. This value agrees with the value E 0 = 1.521 ±0.0015 eV reported by Sturge 8 using optical absorption. The 1M minimum does not converge as the other minima and at B = 0 kG its energy value is 1.5 166 eV. This was taken as an
0 kG versus their order number. An extrapolation to zeroeth order gives the direct bandgap energy value at 1.520 ±0.001 eV. The slope gives a periodicity of 41.8 meV for the oscillations. Band parameters calculated by Lawaetz1 give melmh = 0.108, where me is the conduction band effective mass and mh is the heavy hole effective mass. Using that ratio, the periodicity of 41.8 meV and the following relation ~: o < < h~, =
[(hv—Eo)/(l +melmh)1 —nhw~
where is the energy of the conduction band carrier, hv the photon energy, E0 the direct bandgap energy,
Vol. 14, No.3
OSCILLATORY PHOTOCONDUCTWITY OF GALLIUM ARSENIDE
h~,the LO (longitudinal optical) phonon energy, and n the number of possible phonon transitions, one obtains for the LO phonon energy a value of 37.7 meV, in good agreement with the value of 36.8 meV obtained by Mooradian’° Previous work on GaSb ‘~.12 pointed out that an increase in the impurity concentration at the surface brought about by oxidation diminished both the collision and recombmation times thus quenching the Landau oscillations and promoting the phonon osdillations. This could possibly explain the difference between Nahory’s results and the present results on epitaxial material: the impurity concentration at the surface of the sample is lower here, either because the material is a hundred times purer or because an A12O3 film is protecting the surface from oxidation. The proposition is reinforced by a second experiment on a high resistivity material which reproduces Nahory’s results An increase in impurity concentration reduces the collision and recombination times; the condition
~r> I (where ~ is the cyclotron frequency and r the collision time) necessary to observe Landau oscillations, is no longer satisfied. Furthermore the decrease in recombination time T,. prevents thermalization of the conduction band carriers permitting the observa. tion of optical phonon interaction via an energy dependent free carrier mobility or lifetime ~~ the condition r450 ~ r~ ) Top (where r~,is the acoustical phonon interaction time, r0~,the optical phonon interaction time) is now satisfied. Therefore indications are that the impurity concentration at the sample’s surface governed by both the surface oxidation and the material’s purity is a decisive parameter in the type of quantum oscifiations to be observed.
Acknowledgements ThanksHill, are due Dr. DL. Rode of Bell Laboratories, Murray Newto Jersey for providing the epitaxial GaAs. This research was conducted with the financial support of the National Research Council of Canada. —
J.AWAETZ P.,Phys. Rev. B4, 3460(1971).
VREHEN Q.H.F.,J. Phys. Chem. Solids 29, 129 (1967).
REINE M., AGGARWAL R.L. and LAX B.,Phys. Rev. B5, 3033 (1971). NAHORY R.E.,Phys. Rev. 178, 1.293 (1968).
ABOAF J.A., J. Electrochem. Soc.: Solid State Science 114,949 (1965).
BARBARIE A. and FORTIN E., Can J. Phys. 50, 1593 (1971).
UCHINOKURA K. and FAN H.Y.,Phys. Rev. Bi, 3395 (1969).
STURGE M.D.,Phys. Rev. 127, 768 (1962).
HABEGGER M.A. and FAN H.Y., Phys. Rev. Lett. 12,99(1964). MOORADIAN A. and URIGHT G.B., Solid State Commun 4,431(1966).
BARBARIE A. and FORTIN E., Proc. 11th mt. Conf on the Physics ofSemiconductors 1,300(1972).
FILION A. and FORTIN E.,Phys. Rev. B, Oct. (1973).
STOCKER H.J., STANNARD C.R., KAPLAN H. and LEVINSTEIN H.,Phys. Rev. Lett. 12, 163 (1964).
La photoconductivité du GaAs dans le domaine spectral 1.5—1.8 eV révèle deux types d’oscillations quantiques: des oscillations du type Landau observées dans le materiel epitaxial en champs magnétiques intenses ou des oscifiations du type phonons LO observCes dans le materiel de haute résistivitC. La puretC et le traitement de Ia surface des Cchantillons semble ëtre le facteur determinant pour l’observation de l’un ou de l’autre type d’oscillations.