The study of a Lennard-Jones (LJ)-based model for viscosities led us to con
centrate on a recent high-quality pressure-density-temperature LJ equation
of state (EOS). The application of LJ-EOSs to real fluids typically require
s introducing a temperature dependence on the LJ intermolecular potential p
arameters. Though this goal has been the matter of previous works, we have
found the following problems: (a) relatively low quality of the unmodified
LJ-EOS (U-W-EOS) used as the basis for real-fluid computations: (b) uncontr
olled use of the U-LJ-EOS outside the range of conditions of the supporting
molecular simulation data; (c) poor reproduction of the pure-compound crit
ical temperature and/or critical pressure and (d) need of iterative calcula
tions either to compute the vapor pressure at a given temperature or to reg
ress the values of the LJ parameters from experimental data at saturation.
In the present work, we address all these problems, and provide useful equa
tions to compute properties at saturation through non-iterative procedures.
These procedures conserve the quality of the output numerical result, in c
omparison to the corresponding iterative algorithms. We examine a number of
options for the LJ temperature dependence. Some of them make possible to r
epresent accurately the vapor pressure, in a wide range of temperature, for
non-polar, polar and heavy compounds. We also scrutiny the potential (unwa
nted) appearance of crossing pressure versus volume isotherms, which can be
produced by the introduction of a temperature dependence, and propose a si
mple test to systematically assess the range of applicability of a chosen t
emperature dependence. We also illustrate here the effect of a good vapor-p
ressure reproduction on the prediction of viscosities. (C) 2001 Elsevier Sc
ience B,V. All rights reserved.