The field-effect passivation of the interface of thermal oxides on silicon
is experimentally investigated by depositing corona charges on the oxide of
solar cells and of lifetime test structures. The open circuit voltage of s
olar cells with interdigitated rear contacts can be increased by +12 mV or
decreased by -34 mV, respectively, by depositing positive or negative coron
a charges on top of the front oxide. The resulting effective surface recomb
ination velocity, S-eff, is determined on carrier lifetime test structures
for different injection levels and charge densities using microwave-detecte
d photoconductance decay and a new expression for the Auger-limited bulk li
fetime. S-eff can be varied between 24 cm/s and 538 cm/s on a 1 Omega cm p-
type wafer with a thermal oxide of 105 nm thickness. The measurements are c
ompared with theoretical predictions of an analytical model for the calcula
tion of the surface recombination. Measured values for the capture cross se
ctions and interface trap densities are used for the calculation. The model
predicts an optimum passivation for strong positive compared to strong neg
ative charge densities. This is due to the asymmetry of the capture cross s
ections for electrons and holes. This prediction is in very good agreement
with the measured S-eff values. However, the predicted S-eff values of well
below 1 cm/s for 1 Omega cm p-type silicon cannot be achieved in the exper
iment. This discrepancy can be explained by an inhomogeneous charge distrib
ution resulting in potential fluctuations and additional loss currents. Wit
h a new extended analytical model for the calculation of S-eff the measured
S-eff values can be described quantitatively. (C) 1999 American Institute
of Physics. [S0021-8979(99)09113-6].