Can COSMOTherm Predict a Salting in Effect?

We have used COSMO-RS,
a method combining quantum chemistry with
statistical thermodynamics, to compute Setschenow constants (KS) for a large array of organic solutes and
salts. These comprise both atmospherically relevant solute-salt combinations,
as well as systems for which experimental data are available. In agreement
with previous studies on single salts, the Setschenow constants predicted
by COSMO-RS (as implemented in the COSMOTherm program) are generally
too large compared to experiments. COSMOTherm overpredicts salting
out (positive KS), and/or underpredicts salting in (negative
KS). For ammonium and sodium salts, KS values are larger for oxalates and sulfates, and smaller
for chlorides and bromides. For chloride and bromide salts, KS values usually increase with decreasing size
of the cation, along the series Pr4N+ < Et4N+ < Me4N+ ≤ Na+ ≈ NH4+. Of the atmospherically
relevant systems studied, salting in is predicted only for oxalic
acid in sodium and ammonium oxalate, and sodium sulfate, solutions.
COSMOTherm was thus unable to replicate the experimentally observed
salting in of glyoxal in sulfate solutions, likely due to the overestimation
of salting out effects. By contrast, COSMOTherm does qualitatively
predict the experimentally observed salting in of multiple organic
solutes in solutions of alkylaminium salts.