Chemistry Reference
In-Depth Information
Deschamps, Costa Gomes and Padua
75
computed the relative solubility
of argon, methane, oxygen, nitrogen, and carbon dioxide in [C
4
mim][PF
6
] and
[C
4
mim][BF
4
] at 1 bar and temperatures of 303, 323, and 343 K using the free-
energy perturbation algorithm in DL_POLY. The simulations gave the correct
relative order of solubility in [C
4
mim][PF
6
], but the temperature dependence
of the solubility for the nonpolar gases was exactly opposite that observed
experimentally. It has been found experimentally that the partial molar enthal-
py of solution for many nonpolar gases such as nitrogen and oxygen is zero or
slightly positive. In contrast, the simulations predicted that they were slightly
negative, thus giving the wrong temperature dependence. For CO
2
, however,
the simulations predict correctly that solvation is exothermic. The radial
distributions for the CO
2
/[C
4
mim][PF
6
] system showed that CO
2
does not
localize near the C2 carbon of the cation, but rather prefers the C4 and C5
carbons. The authors also found that the CO
2
molecules lay flat against the
PF
6
anion. Both of these results are consistent with the findings of Cadena
and co-workers.
74
Computing the solvation of gases in liquids involves two processes: crea-
tion of a cavity within the solvent capable of hosting the solute (often termed a
''free volume'' process), and activation of the solute-solvent interactions.
Deschamps, Costa Gomes and Padua
75
computed the free volume in ionic
liquids by performing hard sphere insertions into the pure ionic liquid and
determining the probability of finding a cavity of a particular size. They found
the probability of cavity formation in an ionic liquid to be lower than what is
observed in either water or
n
-hexane at the same temperature, and thus the
work of cavity formation in an ionic liquid is greater than for conventional
solvents. They also saw little difference in the free volumes between the two
ionic liquids they studied. To evaluate the interactions between solute and sol-
vent, the authors computed the solubility of CO
2
and N
2
in both ionic liquids
in which electrostatic terms were used to model the quadrupole moment of the
gases. They then repeated the calculations with CO
2
and N
2
models having no
3
Figure 8
(a) Cross section of the three-dimensional probability distributions around
benzene in [C
1
mim][Cl]. The sixfold axis of the benzene molecule lies in the vertical
direction, and the twofold axis of the molecule lies in the horizontal direction,
perpendicular to the field of view. The difference in distributions of the cations and
anions is shown; regions in the lightest shades are more likely to contain anions, while
dark shades correspond to regions containing cations (see original publication for color
plots). The scale is in multiples of the average concentration of cations (or anions) in the
solution. (Taken from Ref. 72. Used with permission.) (b) Cross section of the three-
dimensional probability distributions around uncharged ''benzene'' in [C
1
mim][Cl].
The difference in distributions of the cations and anions is shown; regions in the lightest
shades are more likely to contain anions, while dark shades correspond to regions
containing cations (see original publication for color plots). Note that the differences are
much less than in the simulations of the ''real'' benzene (a). (Taken from Ref. 72. Used
with permission.)
