TY - JOUR

T1 - Quantitative absorption corrections for electron diffraction

T2 - Correlation between theory and experiment

AU - Rossouw, C. J.

AU - Miller, P. R.

AU - Drennan, J.

AU - Allen, L. J.

PY - 1990/12/1

Y1 - 1990/12/1

N2 - The absorptive potential due to thermal diffuse scattering (TDS) of fast electrons is calculated using an Einstein model. The TDS potential is substantially different from the elastic potential, being extremely peaked on atomic positions with virtually no absorption in the channels between atomic planes. This absorptive potential is used to determine imaginary (absorptive) components of the eigenvalues for each branch of the dispersion surface, using a Bloch wave description for propagating the fast electron wavefunction through a crystal. Large-angle convergent-beam electron diffraction (LACBED) contrast is calculated and compared with experiment using a 300 keV beam for various zone axes of silicon, GaAs and spinel. Changes in mean free paths for TDS under strong dynamical diffraction conditions as a function of crystal orientation are calculated, showing strong channelling/blocking characteristics due to the extremely localized TDS potential. The usual model for taking dynamical absorption effects into account, i.e. scaling the absorptive potential to the elastic potential, has severe limitations and fails to quantitatively predict LACBED contrast. We demonstrate that the Einstein model provides quantitative agreement with experiment, without the use of fitted parameters. Theory is also correlated with convergent-beam diffraction patterns, where rockin-curve contrast is used to determine the polarity of a multiply twinned epilayer of CdTe. This model is also applied to the thickness fringe method for determining the composition of AlGaAs quantum wells. Thickness fringe calculations using exact matrix methods for determining complex eigenvalues are compared with perturbative methodes.

AB - The absorptive potential due to thermal diffuse scattering (TDS) of fast electrons is calculated using an Einstein model. The TDS potential is substantially different from the elastic potential, being extremely peaked on atomic positions with virtually no absorption in the channels between atomic planes. This absorptive potential is used to determine imaginary (absorptive) components of the eigenvalues for each branch of the dispersion surface, using a Bloch wave description for propagating the fast electron wavefunction through a crystal. Large-angle convergent-beam electron diffraction (LACBED) contrast is calculated and compared with experiment using a 300 keV beam for various zone axes of silicon, GaAs and spinel. Changes in mean free paths for TDS under strong dynamical diffraction conditions as a function of crystal orientation are calculated, showing strong channelling/blocking characteristics due to the extremely localized TDS potential. The usual model for taking dynamical absorption effects into account, i.e. scaling the absorptive potential to the elastic potential, has severe limitations and fails to quantitatively predict LACBED contrast. We demonstrate that the Einstein model provides quantitative agreement with experiment, without the use of fitted parameters. Theory is also correlated with convergent-beam diffraction patterns, where rockin-curve contrast is used to determine the polarity of a multiply twinned epilayer of CdTe. This model is also applied to the thickness fringe method for determining the composition of AlGaAs quantum wells. Thickness fringe calculations using exact matrix methods for determining complex eigenvalues are compared with perturbative methodes.

UR - http://www.scopus.com/inward/record.url?scp=0025685297&partnerID=8YFLogxK

U2 - 10.1016/0304-3991(90)90069-X

DO - 10.1016/0304-3991(90)90069-X

M3 - Article

AN - SCOPUS:0025685297

SN - 0304-3991

VL - 34

SP - 149

EP - 163

JO - Ultramicroscopy

JF - Ultramicroscopy

IS - 3

ER -