1.0
partially because of the lack of experimental data.
Therefore, at high concentrations or when there
are interacting forces between the molecules, esti-
0.8
mates of activity using solubility may be inaccu-
rate.
In this paper, we will use the term "relative
0.6
χ=0
solubility" to express Cw/Cws and thus estimate
V1
activity. However, we should point out that there
Rubbery
0.4
also can be significant imprecision with this term.
0.5
While it is known that aqueous solubilities vary
Softening Range
with temperature, we have also found that for
0.2
1.0
some organic compounds, the reported solubility
Glassy
1.5
can vary 1 or 2 orders of magnitude at the same
temperature. Montgommery and Welkom (1990)
0
0.2
0.4
0.6
0.8
1.0
have compiled aqueous solubility data from a
Solvent Activity
number of sources. As an example of the variabil-
Figure 1. Flory-Huggins isotherms for various
ity we have observed, two of the values they re-
port for the solubility of 1,1-dichloroethylene at
softening of PVC at room temperature (from
25C are 273 mg/L and 5,000 mg/L.
Berens 1985).
Berens (1985) concluded that if aqueous solu-
tions of solvents are going to cause softening of
rigid PVC, the pure (or neat) solvent must be able
than 0.5 reportedly can completely dissolve PVC
to soften or dissolve PVC and must exist in solu-
(Berens 1985). According to Vonk (1985), PVC is
nearly inert to organic solvents if their χ value is
tion at a concentration that exceeds one-half its
solubility in water (activity > 0.5). However, one
greater than 3 (e.g., alcohols, aliphatic hydrocar-
criticism of Berens' work is that the various test
bons, and organic acids).
(activities) solutions he used in his experiments
According to Berens (1985), at ambient tem-
were prepared by dissolving the organic solvents
peratures, rigid PVC can be softened only by PVC
in polyethylene glycol and thus he never actually
solvents or strong swelling agents at activities
tested aqueous solutions of these organic solvents.
greater than 0.5. The less aggressive the solvent
(i.e., the higher the χ value), the higher the activ-
Since one would expect that the partitioning of
these organic solutes between the dissolved state
ity required to cause softening. Figure 1 shows
and the polymer would differ between glycol and
the isotherms Berens derived using the Flory-
Huggins equation for various values of χ and the
water, studies that use aqueous solutions of PVC
approximate softening range for PVC at 30C.
solvents should be conducted to confirm Berens'
predictions. While Vonk did use aqueous solu-
Berens predicted that for lower temperatures (such
tions of organics in his studies, we were unable to
as those commonly encountered in groundwa-
find a thorough discussion of his methods that
ter), slightly higher activities would be needed to
was available in English. Therefore, our labora-
cause softening.
tory conducted a series of studies (Parker et al.
The activity of an organic chemical in aqueous
1992, Parker and Ranney 1994, 1995) to confirm
solution can be approximated by dividing the con-
Berens' predictions for PVC solvents and swell-
centration of the compound in solution (Cw) by
ing agents in aqueous solutions. Following is a
its solubility in water (Cws) (Berens 1985, Vonk
summary of our findings to date.
1986, Holsen 1988). However, the true relation-
ship between activity (a), Cw, and Cws is:
Methylene chloride studies
Cw
a=γ
The first test solutions that were tested con-
Cws
tained methylene chloride. Vonk (1985) reported
where γ is the activity coefficient (Holsen 1988).
that the χ value of methylene chloride is 0.55,
According to Holsen (1988), γ is equal to unity
while Berens (1985) did not give a value for χ but
when there are no interacting forces between the
reported that it is less than 0.53. We (Ranney and
molecules; these forces become important at high
Parker 1995) have observed pieces of PVC well
but still unsaturated conditions. At lower con-
casing dissolve when placed in neat methylene
centrations, γ has been considered equal to unity
chloride. Thus, if Berens (1985) is correct in say-
2
Previous Page
