Ph. Poole et al., SPINODAL OF LIQUID WATER, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics, 48(5), 1993, pp. 3799-3817
An open question in the study of water concerns the shape of the liqui
d spinodal line in the phase diagram of water, a boundary which repres
ents the limit of mechanical stability of the liquid state. It has bee
n conjectured that the pressure of the liquid spinodal P(s)(T) does no
t decrease monotonically with decreasing temperature T, but passes thr
ough a minimum and is ''reentrant'' from negative to positive pressure
P in a region of T in which the liquid is deeply supercooled. The con
jectured minimum in P(s)(T) has not been directly observed due to the
difficulties encountered in experiments which attempt to study liquid
water under tension. Here we exploit the ability of molecular-dynamics
computer simulations to model the behavior of liquid water deep into
its metastable region. We thereby attempt to observe a minimum in P(s)
(T). We first argue that the ST2 potential of Stillinger and Rahman [J
. Chem. Phys. 60, 1545 (1974)] is the best of several commonly used wa
ter interaction potentials for this purpose. Then, we conduct simulati
ons of a system of ST2 particles over a wide range of stable and metas
table liquid-state points, and demonstrate that P(s)(T) for ST2 is not
reentrant. In a second set of simulations we test if the behavior we
find is limited to the ST2 potential by exploring the relevant thermod
ynamic region of the liquid as simulated by the TIP4P interaction pote
ntial of Jorgensen et al. [J. Chem. Phys. 79, 926 (1983)]. We find tha
t the TIP4P potential confirms the absence of a reentrant spinodal. We
also show how the structural and energetic properties of both the ST2
and TIP4P liquids are consistent with the absence of a reentrant spin
odal.