The perplexing problem of the extremely large activation energy for nucleat
ion U(ac)similar to 10(6) eV, which, as was shown by Landauer, appears in t
he Merz thermal-fluctuational theory of ferroelectric switching, is examine
d. Such enormous values of U-ac are a result of the combined contribution o
f the energy of the depolarization field and the surface energy of the nucl
eus. While the depolarizing field term can be reduced by a redistribution o
f free charges, the surface energy term alone is too large (greater than 10
(4) eV) for thermal-fluctuation-aided nucleation to be possible. This concl
usion contradicts the numerous observations of a rapid increase in the numb
er of repolarization nuclei with temperature, which suggest that the actual
activation energy is on the order of an electron volt. This article is dev
oted to the problem of how the surface energy of a nucleus can be compensat
ed. It is assumed that repolarization nuclei form by a fluctuon mechanism,
and it is shown how this assumption can help in resolving the conflict betw
een theory and experiment. It is suggested that upon application of an exte
rnal electric field, electrons tunnel into the region where the spontaneous
polarization vector fluctuates. This fluctuation region maintains its stab
ility owing to the field of the electrons localized within the region. This
leads to the appearance of a multielectron fluctuon, which is a bound stat
e of electrons and a fluctuation and can be regarded as a repolarization nu
cleus. The electronic structure of these nuclei is calculated with density-
functional theory. The calculations show that fluctuon effects can fully co
mpensate the surface energy term. The assumption that repolarization nuclei
have a fluctuon nature is shown to account for the principal laws governin
g the switching process, including low activation energy values. The result
s of possible experiments that could verify the proposed mechanism are pred
icted. <(C)> 2000 American Institute of Physics. [S0021-8979(00)06920-6].