THE NGAI COUPLING MODEL OF RELAXATION - GENERALIZATIONS, ALTERNATIVES, AND THEIR USE IN THE ANALYSIS OF NON-ARRHENIUS CONDUCTIVITY IN GLASSY, FAST-IONIC MATERIALS
Jr. Macdonald, THE NGAI COUPLING MODEL OF RELAXATION - GENERALIZATIONS, ALTERNATIVES, AND THEIR USE IN THE ANALYSIS OF NON-ARRHENIUS CONDUCTIVITY IN GLASSY, FAST-IONIC MATERIALS, Journal of applied physics, 84(2), 1998, pp. 812-827
The ionic conductivity of glassy, fast-ion-conducting materials can sh
ow non-Arrhenius behavior and approach saturation at sufficiently high
temperatures [J. Kincs and S. W. Martin, Phys. Rev. Lett. 76, 20 (199
6)]. The Ngai coupling model was soon applied to explain some of these
observations [K. L. Ngai and A. K. Rites, Phys. Rev. Lett. 76, 1296 (
1996)], but detailed examination and generalization of the coupling mo
del suggested the consideration of a related, yet different, approach,
the cutoff model. Although both the coupling and cutoff models involv
e a shortest nonzero response time, tau(c), and lead to single-relaxat
ion-time Debye response at limiting short times and high frequencies,
they involve different physical interpretations of their low- and high
-frequency response functions. These differences are discussed; the pr
edictions of both models in the frequency and time domains are compare
d; and the utility of both models is evaluated for explaining the non-
Arrhenius conductivity behavior associated with the dispersed frequenc
y response of zAgI+(1-z)[0.525Ag(2)S+0.475B(2)S(3):SiS2] glass for z=0
and 0.4. The cutoff approach, using simulation rather than direct dat
a fitting, yielded semiquantitative agreement with the data, but simil
ar analysis using the coupling model led to poor results. The coupling
model leads to an appreciable slope discontinuity at the tau(c) trans
ition point between its two separate response parts, while the cutoff
model shows no such discontinuity because it involves only a single re
sponse equation with a smooth transition at tau(c) to limiting single-
relaxation-time response. The greater simplicity, utility, and general
ity of the cutoff model suggest that it should be the favored choice f
or analyzing high-conductivity data exhibiting non-Arrhenius behavior.
(C) 1998 American Institute of Physics. [S0021-8979(98)06714-0]