You are on page 1of 10

Effect of an electric field on the luminescence of GaAs quantum wells

E.E.Mendez, G. Bastard, L.L. Chang, L. Esaki, H.Morckoc and R. Fischer

Student: Enrico Bandiello Professor: Nuria Garro

Universidad de Valencia Facultad de Fsica - Mster en Fsica Avanzada

Valencia, Diciembre 2011

Agenda

1. Goals 2. Experimental details 3. Experimental results 4. Analysis 5. Conclusions

Experimental details
Each sample is formed by 6 GaAs QW delimited by two layers of Ga1-xAsxAl (barriers); x here is ~0.320.40; the width of the QWs is Lz~20-35 ; the multilayer has been grown by MBE.

Semitransparent Au film on top of the multiheterostructure Schottky barrier

Quantum wells well inside the space-charge region (no free carriers in QW).

Luminescence was excited using three different lasers: Ar+ He-Ne Kr+ 5145 ~2.41 eV 6328 ~1.96 eV 6471 ~1.92 eV

Some photo-voltage was induced by the laser even at low intensity (Vext=0.73 V @0.08 W/cm2).

Goals

Low T (6K) photoluminescence measurements in GaAsGa1-xAlxAs quantum wells with an electric field E perpendicular to the well plane.

Study of the variation of the position and intensity of the excitonic peak with E. Qualitative/quantitative explication of the results.

Experimental results (1/2)


P1 No external field

Two peaks can be identified: P1 1.681 eV e--hh ground state exciton recombination in GaAs QW. P2 1.662 eV ground state e--C recombination (C is often present as an impurity in MBE grown films). The different origin is confirmed by the different behavior of P1, P2 when Vext is decreased. P1 decreases at much faster rate than P2 (quenching). Please note that the spectra for Vext < +0.73 V are amplified by a factor 2.5x/10x. P1 completely disappears at Vext= -0.2 V and P2 disappears at Vext= -0.5 V. Remember: E=-Vext/z ~ 1x104-5x104 V/cm here (estimated)
5 P2

Experimental results (2/2)


Excitonic peak (P1)

The position of P1 doesn't vary significantly with Vext while P2 redshifts when Vext decreases (increasing total field). For Vext< 0.25 V P2 shifts linearly with a slope of 0.03 eV/V. Results only vary slightly for different samples but are completely independent of the excitation energy (i.e. the laser).
Impurity peak (P2)

Analysis (qualitative)
What causes exciton PL quenching?

Not due to exciton dissociation by impact ionization with accelerated e- (as seen in bulk GaAs using fields ~1 V/cm, associated to an increasing in PL of the impurity peak). Exciton PL quenching in this experiment occurs with fields E ~ 5x104 V/cm and the impurity peak decreases with E. Not due to ionization of the exciton by the E field. This occurs (in bulk) with fields EI =R/ea (~ 5 kV/cm for GaAs) but in QW the field is perpendicular to the well no ionization. Probably due to the spatial separation (polarization) of the carriers induced by E: E=0 e- and hole WF are symmetrical with respect to the center of the QW. With E0 and perpendicular to the well the e- and hole WF are displaced in opposite directions the probability of recombination is thus decreased.

Analysis (quantitative)
Analysis is based on theoretical calculations. The spatial separation of the carrier by the field is estimated by the overlap integral:

The radiative recombination rate is proportional to Mcv2; for Lz=30 and E ~ 50 kV/cm a 3% decrease in PL intensity is expected. Moreover the calculations show that the field induces a negative energy shift on both e- and hh (-0.2 meV, -6.4 meV @50 kV/cm). This result of 3% in the decrease of excitonic PL is too small with respect to measured value and the strength of E, while the the trend is predicted correctly. The difference with measured values has thus to be explained... the actual

Conclusions
The calculations have been performed for a particle confined in a single finite well and are based on the empirical 85%-15% Dingle rule (c/v fraction of the energy gap difference) 0.4 eV for the depth of e- wells, 0.07 eV for hh. Some hypothesis for the quantitative discrepancy between the theory and the experimental values: a) Deviations from the Dingle rule: shallower heavy hole wells would lead to larger polarization of the WF by the field reduction in recombination rate. b) QW are quite narrow the field induces a leakage of the WF in the Ga1-xAsxAl the presence of nonradiative recombination centers further reduces the PL peak. c) QW have quite narrow barriers the WF of the e- (and hh) in adjacent wells could couple leading to a more polarized distribution of the carrier (with respect to the single well case considered in the calculations) d) Lowering in the energy of the quantum states of e- (due to E) redshift of the impurity peak. In particular, c) and d) are believed to be the major reasons for the quantitative discrepancy. They have to be taken in account in subsequent calculations and then tested again experimentally.
9

Thank you for your attention!

Created with Libreoffice on Debian GNU/Linux

10

You might also like