C. Seko et K. Takatsuka, NONERGODICITY AND 2 SUBPHASES IN THE COEXISTENCE REGION IN ISOMERIZATION DYNAMICS OF AR-7-LIKE MOLECULES, The Journal of chemical physics, 104(21), 1996, pp. 8613-8626
It is well-known that a single cluster like Ar-7 undergoes ''melting''
from solidlike to liquidlike states as the energy is increased, the t
ransition of which is not as sudden as the ordinary phase transition t
hough and has a somewhat broad energy range in which solid and liquid
coexist. We study a very anomalous dynamics of the coexistence region
in the structural isomerization. It is explicitly shown that the time-
series of the structural changes both in the purely solidlike and liqu
idlike phases are stationary, while the coexistence region is found to
generate a strongly nonstationary dynamics. The calculated distributi
on of the residing times for the cluster to stay in one of the possibl
e structures exhibits a nonexponential form having a large hole around
the zero lifetime in the coexistence region. Motivated by these stran
ge behaviors, we have calculated the phase-space volumes that are assi
gned to the individual potential basins, and verified directly that wh
ile the pure liquid region is of ergodic nature, the dynamics in the c
oexistence region is indeed strongly nonergodic. The steep rises of th
e Lindemann index and the maximum Liapunov exponent in the coexistence
region, which were reported before by other authors, are found to be
ascribed to the statistical nature rather than the dynamical propertie
s as opposed to the picture suggested by the physical;meaning of the i
ndices. It also turns out that the energy range for the coexistence re
gion should be taken wider than considered before and thus extends bey
ond the ''melting point'' that is defined usually on the basis of the
Lindemann index. Therefore it is appropriate to divide the coexistence
region into two subphases. A ''temperature'' in a microcanonical ense
mble is defined so as to characterize the distribution of phase-space
volume on a given energy plane. Based on this distribution, we describ
e a statistical reason why the onset energy of the melting is much hig
her than those of the transition states. (C) 1996 American Institute o
f Physics.