&EPA
United States
Environmental Protection
Agency
Robert S. Kerr Environmental
Research Laboratory
Ada, OK 74820
Research and Development
EPA/60Q/M-91/QQ9 Mar. 1991
ENVIRONMENTAL
RESEARCH BRIEF
Solubility, Sorption and Transport of Hydrophobic
Organic Chemicals in Complex Mixtures
P.S.C. Rao1, LS. Lee1 and A.L Wood8
Introduction
Environmental contamination problems commonly involve
wastes consisting of complex mixtures of chemicals. The
behavior of these mixtures has not been well understood
because the primary chemodvnamlc properties (e.g.,
solubility, sorptlon, transport) of organic chemicals have
usually been characterized in aqueous solutions which are
simple in composition relative to many waste mixtures fpynd
at or near disposal/spill sites. Typically, laboratory studies
have focused on ertemodynamtcs of single solutes in water or
In dilute electrolyte solutions.
The research summarized in this report focuses on the
effects which organic cosolvents have on the sorption and
mobility of organic contaminants. This work was initiated in
an effort to improve our understanding of the environmental
consequences associated with complex mixtures and to
enhance our ability to deal with these consequences in a
technically responsible manner,
Specific objectives of the project were to:
(1) measyre solubility and sorption for selected organic
chemicals in complex solvent mixtures consisting of
mixtures of organic cosolvents and water; and
(2) utilize isocratte- and gradient-elution techniques to
characterize the Impacts of organic cosoivents on the
transport of hydrophobia organic chemicals In soils and
aquifer media,
The results of this work have application to the definition,
prediction and remediation of sol and groundwater conta-
mination problems. Since an increase in the concentration of
organic cosoivents is reflected in decreased sorption, organic
contaminants at waste disposal/spill sites are liKely to be
present at higher concentrations In pore water. Decreased
sorption, In turn, will lead to organic contaminants being
transported further than predicted from aqueous-based
'Soil Sd. Dept.. Univ. of Florida, Gainesville, FL 32611; *R.S. Kerr
Env. Res. Lab., U.S. EPA, Ada, OK 7482Q
transport data. While this manifestation of cosolvency can
exacerbate environmental problems, judicious application of
the principles of cosolvancy can assist in alleviating existing
problems. For example, removal of contaminated soils from
a disposal/spill site and extraction with solvent mixtures (so-
called "solvent washing*) is one remediation technique that is
receiving considerable attention. The results presented here
should be of direct use in selection of the appropriate solvent
mixtures to extract the otganic contaminants of interest,
Background
We define here complex mixtures as those systems having
multiple solutes and multiple solvents. The solute mixtures of
interest might consist of various combinations of nonpolar,
hydraphobic organic chemicals (HOCs); hydrophoblc,
tonizable organic chemicals (HIOCs); and Ionic organic
chemicals (IOCS). The solvent may be a mixture of water
and one or mors organic cosolvents. Cosolvents that are
soluble in water in all proportions will be referred to as
eomptetely-misetble organic solvents (CMQSs), Other
solvents which have only a finite solubility in water will be
referred to as partially-miscible organic solvents (PMGSs),
Two types of solvent mixtures are of interest: (I) solvents
consisting of water and eosofvents in a single, homogeneous
tfqyid phase; and (II) solvents containing water and
cosolvents that form at least two distinct liquid phases. In
this report, the first type will be referred to as mixed solv&nts,
while the second type wilt be designated as muKiphasic
solvents,
In recent years, several researchers have recognized the
need to study the ehemodynamlcs of complex mixtures; and
coordinated efforts were Initiated to develop theoretical
approaches and bases. This work has funded
primarily by the U.S. Environmental Protection Agency (EPA),
the U.S. Department of Energy {DOE), and the U.S. Air
Force (USAF), Research funded by DOE, conducted
primarily at Battelle PNL, has focused on competitive sorption
by soils and clays from HIOC mixtures found In energy
wastes (Felice et al., 1985; Zachara et a!., 1987), With
funding from EPA, researchers at the University of Florida
hava also studied the sorptlon from mixtures of HOCs and
Printed on Recycled Paper
-------
HIOCs (selected to represent the so-called Appendix VIII
compounds) that are commonly found at hazardous-waste
land-treatment facilities (Rao et al,, 1986), EPA has also
funded much of the work completed to date on the effects of
cosolvents on chemodynarnics of organic contaminants
(Yalkowsky, 1985, 1987; Rao et al., 1985; Rao and Lee,
1987; Nkedi-Kizza et at, 1985,1987,1989; Woodbum et
Guissepi-EIIle, 1988; Walters et al., 1989). Sorptlon of jet
fuel constituents from liquid and vapor phases has been
examined with funding from the U.S. Air Force (Mclntyre and
deFur, 1985; Raoet al., 1988),
The log-linear cosolvency model and the UNIFAC model are
among the theoretical approaches that have been used to
examine cosolvent effects on solubility and sorption. The log-
linear cosolvency model (Yalkowsky and Roseman, 1981) is
based on the central assumption that the logarithm of the
solute solubility in a mixed solvent is given by the weighted-
average of the logarithms of solubilities in the component
solvents in the mixture; the weighting coefficient is taken to
be the volume fraction of each solvent component. Thus,
tog Sm = £ f, log S,
(1)
where S is solubility (mg/L), f Is volume fraction of the
solvent, and the subscript m denotes mixed solvent while i
denotes a specific solvent component. Note that averaging
the logarithms of solubilities is equivalent to averaging the
free energies of solution in different solvents in the mixture.
The UNIFAC model (Fredenslund et al., 1975) Is a group-
contribution scheme for calculation of the activity coefficients.
This model Is based on the UNIQUAC model {Abrams and
Prausnitz, 1975) and the so!ution-of-group concept (Wilson
and Deal, 1982), In this model, a mixture of different
chemicals is treated as a mixture of the functional groups
constituting the components in solution.
In many cases, the UNIFAC model may be preferred over the
log-linear model because: (i) it has a more sound theoretical
basis, (ii) activity coefficients In mixtures can be calculated
given only pure component data, and (Hi) all possible
interactions among the components in the mixture are
explicitly considered. A limitation of the UNIFAC model,
however, is that although the group interaction parameters
required to estimate the solute activity coefficients are
continuously reviewed and updated, their values are not
available for a number of systems of interest hare.
An extensive amount of data has shown that in binary mixed
solvents, HOC solubility increases and sorption decreases in
a log-linear manner as the volume fraction of the organic
cosolvent increases (Rao et al., 1985; Nkedi-Kizza et al.,
1985; 1987; Woodburn et al., 1988; Fu and Luthy, 1986a,
19865; Rubino and Yalkowsky, 1985; 1987a, 1987b, 1987c).
These experimental findings are consistent with the
predictions of both the UNIFAC model and the log-linear
cosolvency model. The successful extension of the log-linear
model to describe solubility and sorption in binary, ternary
and quinary mixed solvents has also been demonstrated
(Rao and Lee, 1987; Rubino et al.. 1984; Rubino and
Yalkowsky, 1985; Yalkowsky and Rubino, 1984),
In contrast to the large amount of data on solubility and
sorption of HOCs in mixed solvents with completeiy-mtscibie
organic cosolvents (e.g., alcohols), similar data for mixed
solvents Involving partially-misclble organic cosolvents were
essentially nonexistent prior to this study. Only limited data
for solubility and sorption of HIOCs in mixed solvents are
available (Fu and Luthy, 1986a; Zachara et al., 1988),
Cosolvent Impacts on kinetics of phase partitioning have
been examined, though mostly in a qualitative manner
(Freeman and Cheung, 1981; Nkedi-Kizza et al,, 1989;
Walters et al, 1989).
Laboratory studies were conducted during this project to
collect solubility data, and these were used to evaluate
two theoretical approaches (the UNIFAC model and the log-
linear model). Based on the solubility data collected,
modifications to the log-linear model ware proposed to
improve its predictive capabilities. Batch equilibration and
column displacement techniques were used to characterize
equilibrium and nonequllibrium sorption of HOCs and HIOCs
from mixed solvents. These data were used to compare
cosolvent effects on solubility and sorption and to assess
cosolvent interactions with the sorbent which impact
equilibrium and kinetics of sorption. Partitioning of HOCs into
water from complex mixtures of liquids was examined In
order to provide data essential lor estimating solubilization
and release of aromatic constituents from fuels such as
gasoline, diesei, and jet fuel, and from such wastes as coal
tar and creosote. Retention of several HOCs from mixed
solvents by synthetic sorbents (reversed-phase
chromatography supports) and a surface soil was measured
at several temperatures to characterize the energetics of
HOC sorption and to better understand the sorptive
mechanisms.
in the following sections, a brief summary of our findings in
each of these areas is presented. Details can be found In
several publications listed in Appendix A and Appendix B.
Cosolveney
The effects on solubility and sorption (hence, on transport) of
organic chemicals upon addition of one or more organic
cosoivents to an aqueous solution are defined here as
cosolvency, The alterations in solubility might result from the
following interactions; solute-solute, solute-cosolvent,
cosolvent-cosolvent, and water-cosolvent. For nonpolar
solutes and at low concentrations of polar/ionizable solutes,
solute-solute interactions are likely to be negligible and are
not considered further. The work presented here focuses on
the other interactions listed above.
Sorption of organic solutes, especially HOCs, is inversely
related to solubility. Thus, an Increase in solubility resulting
from the addition of a cosolvent leads to a proportional
decrease in sorption. In addition to the interactions listed for
solubility, sorption is influenced by solute-sorbent, and
sorbent-cosolvent interactions. Both solute and cosolvent
interactions with the sorbent were investigated in this project,
A convenient measure of cosoivency is the cosotvency
power, defined here on the basis of the inherent ability of a
cosolvent to produce an alteration in the solubility or sorption
upon addition of a cosolvent. Behavior In a pure (neat)
aqueous solution and a pure (neat) organic solvent will serve
as the basis for quantifying the cosolvency power.
-------
The eosolvency power (os c) of a cosolvent for a solute
(subscript s) may be defined as:
(2)
where S is the solute solubility (mg/L) in neat cosolvent
(subscript c) or pure water {subscript w). Since HOC solubility
in organic solvents is larger than that in water, 0gc > 0.
Larger values of as c indicate a greater sofubilizinig power of
the solvent for a specific soiute,
1 may be expected that with decreasing polarity of the
cosolvent, o e wlil increase for hydrophobia solutes. Thus,
0 values stiould be inversely related to various indices of
sowent polarity (e.g., dielectric constant, Rohrschneider
polarity Index, ET(30), and others). Rubins and Yalkowsky
(1987a, 1987b) have investigated such relationships for
soiubilization of pharmaceutical drugs. The inverse relation-
ship between os and ET(30) for solubiiization of anthracene
In several solveras is shown In Figure 1, for a given solvent,
os e values should increase with increasing hydrophobteity of
the solute, Morris et al. (1988) have shown that a values are
indeed positively correlated with log K^ values. Two
examples of such a relationship are shown in Figures 2A
and 2B, It is evident from the data in Figures 1 and 2 that
the cosoivency power (os c) values can be estimated given
specific properties of the solute and the solvent. Data
presently available for several solute-solvent systems can be
effectively utilized to estimate os e for other systems.
The measured HOC solubility profiles in binary solvent
mixtures generally deviate from the expected tog-linear plot,
primarily because of water-cosolvent interactions. The extent
of such deviations may then be taken as an index for the
magnitude of such interactions. The observed cosotvertcy
power in a binary mixed solvent is defined here for solubility
as,
and for sorption as,
°S,m « !°9fK8
(3)
(4)
where cs m is the exoerimentalfy-measured vaiue of
cQsolvency power, S is the measured HOC solubility {mg/L),
and K is the measured sorption coefficient (mL/g), with the
subscripts m and w denoting a binary mixed solvent and
water, respectively.
For operational convenience, in calculating o values in eqs
(3) and (4) we use solubility or sorption values measured in
50% (v/v) solvent mixture for CMOS-water systems and
saturated solutions for PMOS-water systems. This choice
also precludes problems associated with mutual miscibiiity of
organic coslvents (e.g., methanol, acetone) with liquid solutes
(e.g., benzene, toluene). When eosolverrf-water interactions
are significant, the predicted (eq 2) and measured {eqs 3 or
4) eosolvency powers are not equivalent. We may therefore
5.6
5.4
5.2
0)
O 5
CL
O 4.8
C
0)
| 4'6
O
a 4.4
4.2
4
ctloxane
trlchtoro»then»i
chlorobenzeno
m©thylena chloride
MTBE ^ nitrobenzene
m 2-butanone
30
dlmBthylsuItoxIdo
Isopropanol
35
40
45
ET(30)
50
55
60
Figure 1. Inverse relationship between cosolveney power (o$ c) and solvent polarily index E-^30).
(1988).
for £^30) were taken from Reichardt
-------
K
Ftg«8 2, Direct relationship eosolvsncy {0S c) and octanol-wrator
acetone-water mixtures.
nwthanol-waief, and (8)
define for solubility,
*.e
and for sorptiort,
>apo,
9.C
(6)
where the empirical constants a and $ account for water-
eosolvent and sortwnt-cosolvent interactions, respectively,
that to deviations from the expected iog-Iinear behavior,
Note that p = l Implies the absence of water-cosoivsnt
interactions, and o <* 1 suggests that the sorber>t-cosol¥ent
interactions are negligible. Work done during this project
aimed at examining the deviations from the tog-linear
cosolvency model, and to provide a prtanomenoiogfcai
for tf» empirical constants a and p.
and
The solubilities of several HOCs were measured in binary
mixtures (water-CMOS, water-PMOS) and in ternary
mixtures (wale?-CMOS-PMOS). Severn! aromatic
hydrocarbons (al! solids) in studies in order
to minimize specie solute-solvent and scluts-solyta
interactions. The easoiverts permitted measurement of
cosoivency over a of solvent properties. The
HOC solubiitzation profiles compared with the
of the log-linear and the UNIFAC
[refs. 2r 12, A; ref, 7, 3 }.
OUT Indicate that PMOSs can significantly
increase the solubility of HOCs, provided that the PMOS
concentration is about 1% (v/v) or Iarg«r, In an or
predominantly aqueoys solvent, PMOSs such as o-crasol
and aniline which have strong polar functional groups (e.g.,
-OH, -NO4) would exhibit considerable cosolvency, Nonpolar
PMOSs, such as trfchloroethylene (and other haloalkenes)
and toluene (and other aromatic hydrocarbons), are not
to show eosotveney {< 20% increase).
Only In the presence of a CMOS can the concentration of
PMOSs be sufficientfy high to have a measurable
impact on HOC solubility.
Preliminary that the greater eosolvsney of
polar-PMOSs to that of rwnpoiar-PMQS can be
attributed, in part, to water-PMOS interactions arising
from the moieties such as -OH and -NOr The
presence of strong polar functional groups has a dua! effect:
(i) solubility of the PMOS is higher,
relatively high cosoteenl concentrations can be
(fi} the likelihood of watsr-PMOS Interactions is
Modifications to the tog-linear to
account for interactions [mf. I, B J,
In 3A 38, the solubility prairies for
anthracene in CMOS-water iacstone-water) and Irs PMOS-
waier {bulanone-water} mixtures are with
by the iog-i-near While
agreement between the and solubility
is encouraging, further work to understand the
-------
Anthracene Solubility
Acetone/Water Mixtures
* Observed
- - - Loglinear
Modified Log-!inear
I Aqueous-rich
4 '- Phase
Butanone/Water Mixtures B
>
2 Phases
,,••'' Organic
-rich
Phase
C , - , i> Observed
Loglinaaf
Modified Log-linear
0.2 0,4 0.6 0,8 1
Volume Fraction Cosolvent
figure 3. Predicted and measured solubility of anthracene in binary solvent mixtures: (A) acetone-water, and (B) butanone-water. Note fiat
butanona is misdble with water only up to fe • 0.25, while acetone Is misdbte with water in all proporttons.
specific nature of the cosofvent-water Interactions is
recommended.
Results similar to those for HOCs have been observed for
ionizable organic chemicals (Yalkowsky, 1985,1986;
Yalkowsky and Roseman, 1981; Fy and Luthy, 1i86a), As
solute polarity increased relative to the solvent, Yalkowsky
and Roseman (1981} observed that eosolvency curves
became increasingly more parabolic in shape until an inverse
relationship was actually Deserved (i.e., a decrease In
solubility with cosolvent additions). Sych behavior was
explained on the basis of the solute-solvent interactions.
HOC Partitioning from Complex Solvent Mixtures
An understanding of solubility (or partitioning) of HOCs from
complex liquids is essential for predicting contaminant
release from mixtures such as gasoline, coal tar, and
creosote. The solubility of a given component In a mixture
may be altered by other components that may act as
cosolutes or cosolvents. We investigated partitioning of
various aromatic compounds into water from several
gasolines and from known mixtures of aromatic and aliphatic
solvent mixtures [mf, 9, Appendix A J, The purposes of these
studies wsre to assess: (!) the variability In gasoline-water
partitioning of aromatic hydrocarbon constituents arising from
variability in gasoline composition (source variations); (ii) the
application of Raoulfs law for the partitioning of aromatic
hydrocarbons from complex solvent mixtures; and (iii) the
cosolvent affects from oxygenated additives (e.g., methanoi,
ethanol, MTBE).
Aromatic hydrocarbon concentrations in water extracts of 31
gasoline samples varied over an order of magnitude,
reflecting the diversity in gasoline composition. However, the
fuel-water partition coefficients (K^) varied by less than 30%
-------
among these samples. HOC partitioning between water and
known mixtures of aromatic and aliphatic solvents was
measured and used to estimate the upper and lower bounds
of Kto values for more complex solvent mixtures such as
gasoline and diesal fuel.
Assuming that gasoline is an ideal mixture of solvents, the
following relationship can be derived:
log Kte = A - log S *
A = log [ 103 {6/MWJ J
(7a)
(7b)
where § and MW0 are the average liquid density (g/mL) and
molecular weight (g/mole), respectively, of the gasoline; and
S* Is the aqueous solubility (ug/mL) of a specific gasoline
component. The observed inverse, log-log linear
dependence of K^ values on aqueous solubility (Figure 4)
could be well predicted by eq (7), These results suggest that
given the gasoline composition, reliable estimates of likely
concentrations in groundwater can be made.
CosoSvency and! Equilibrium Sorption
Satptlon of HOCs from Mlxod Solvents
Sorption of several HOCs by two soils was measured from
mixed solvents containing CMOSs and/or PMOSs, The utility
of the log-linear cosolvency model for predicting oosoivency
was evaluated. The cosolvent effects on HOC solubility and
sorption were compared In order to examine cosolvent-
sorbeot interactions (I.e., estimation of 0 and P) [fofs, 1,3,
Appendix A; mi 2, Appendix B ].
As anticipated from the solubility studies, nonpolar- PMOSs
(e.g., toluene, trichloroethytene) had small or no effect on
HOC sorption. Sufficiently high concentrations of these
PMOSs needed to measurably decrease HOC sorption could
only be achieved when CMOS concentrations were high (ca.
30% by volyme or larger). In contrast, the cosolvency of
polar-pyOSs was sufficiently high to cause significant
reduction in sorption in predominantly aqueous solutions;
however, the measured cosolvency for sorption was different
(usually higher) from that predicted from solubility data. For
example, sorption of anthracene by Eustis fine sand {organic
carbon content of 0.39%) with increasing amounts of o-creso!
was less than would be predicted based on measured
solubilities (i.e. a > 1) (Flgur® SA). However, fluoranthene
sorption measured in butanone-water mixtures (Figure SB)
was greater than that predicted from solubility measurements
(i.e., a<1).
A series of batch and column studies was performed to
further Investigate the interactions of polar- PMOS (e.g., o-
cresol, aniline, butanone, MTBE) with selected soils with a
broad range of organic carbon contents [fdf, 2, Appendix B J.
It was hypothesized that uptake of polar-PMGS by the
sorbent organic matter decreased sorbent hydrophobiclty
O
o
c
•2
S
•c
5,
I
7B
1,2,3'Trfmethylbenzene
n-Propy/betizene
Efftyflbenzene
i
!
Xy/snes
Toluene
Benzene
O Measured Values
~ §5% Confidence Interval
— Ideal Line (eq, 7)
-3.5
-3
-2.5
-2
-1,5
-1
-0.5
0
Log Sw
Figure 4. Relationship between fuel-water partition coefficients (K^) and aqueous solubility ($*) for major gasoline constituents. The line shown is
calculated assuming Raoulfs law to be applicable (eq 7).
-------
Anthracene Sorption
Cosolvency of o-Cresol
Measured
Reference
0.04 0,08 0,12
Volume Fraction o-eresol
Flyorantnene Sorption
Cosoivency of Butanone
0,16
0.05 0,1 0.15 0,2
Volume Fraction Butanone
0,25
Rgure 5. Cosolvency of (A) o-cresol on the solubility and sorpfcn of "C-anttracene in 50:50 dimeUTybutoxideWater mixtures, and (B) butanone
on tie solubility and sorptton of 14C-antirae8ne.
resulting in reduced HOC sorptton by the soil. The
magnitude of such an effect would depend on the following
factors: (1) the polarity and functionality of the PMOS, which
determine the manner and the amount of sorption/yptake by
soil; (li) the hydrophobicfty of the solute,,as indicated by is
KM»> anc* W. ^9 hydrophobicfty of the sorbent, as indicated
by tts organic carbon content,
Energetics of HOC Sorptlon from Mixed Solvents
Reversed-phase liquid chromatography (RPLC) techniques
have been used to Investigate retention of hydrophobia
solutes In soils and sediments (e.g., Veith et al, 1979; Swann
at al., 1981; McCall et al., 1980; Szabo et a!., 199Qa, 1990b,
1990c). We conducted a series of RPLC studies to evaluate
the syitabllity of RPLC supports as experimental surrogates
for soils and sediments with high OG content. The retention
data collected at several temperatures during Isocratic,
Isothermal elution with binary mixtures of methanotAwater and
acetonitrile/water mobile phases ware used to assess
selected chromatographic models and to investigate the
mechanisms of hydrophobfc solute retention in the presence
of mixed solvents [ret. 7, Appendix A; rets. 3,4,5,6,
Appendix B ],
Changes In chromatographlc retention factors (k) were
correlated with the following Indices of solute hydrophobiclty
and molecular topology: oetarol-water partition coefficient
-------
(K }, hydrophobia surface area |HSA), and first-order
molecular connectivity (1x); and with solvent Indices, sych as,
solvent surface tension (y) and dielectric constant index (I").
Also, the solvophoble model (Horvath et al, 1976) and the
entropy-enthalpy compensation model (Malander et al,,
1978)} were successfully applied to describe' the dependence
of solute retention on solvent composition and temperature,
Correlations obtained between solute retention factors and
various solute Indices, along with the thermodynamlc analysis
of the retention data using the entropy-enthalpy
compensation model showed that alkylbenzenes behaved
differently from the polyeycllc aromatic hydrocarbons and the
monofialobenzenes. Such distinctive thermodynamlc
behavior ts Indicative of the differences in mechanisms of
retention of these compounds,
Sorptlon of four polycyclic aromatic hydrocarbons by Webster
sltty loam (OC = 3%) from a methanol-water mixture (30:70
vrv) at three temperatures was measured using batch
equilibration methods {/»/, 7, Appendix A ]. Results from a
thermodynarnic analysis of the sorptioo data were similar to
those obtained from the RPLC sypports. This suggests that
mechanisms for sorption/retention of polycyclfc aromatic
hydrocarbons from methanol-water are similar for RPLC
supports and the Webster soil; however, comparable sorption
data for alkylbenzenes on soils have not been measured,
Sorptlon of HOCs from Multlphaslc Solvents
For HOC sorption in muftiphasic solvents distribution between
the three phases (soil, water, PMOS) can be described by
the soil-water and PMOS-water partition coefficients if HOC
sorptlon on soil Is assumed to occur only through the
aqueous phase (i.e., no direct solute transfer between the
PMOS and soil) and i the dissolved PMOS has a negligible
cosolvency effect, Sorption of the herbicide diuron (a
substituted urea) by Webster sirty loam and a Eustis sand
was measured from aqueous electrolyte solutions (0.01 N
CaCy saturated with a PMOS, and from several biphasle
solvents. The PMOSs used were n-octanoi, toluene, p-
xytene, and TCE, which are considered nonpolar PMOSs [ref.
3, Appendix A ]. The sorptlon data are shown in Figure §,
These results suggest that even If present In a separate liquid
phase, nonpolar PMOSs did not Influence HOC sorption for
these soils. As previously discussed, nonpolar PMOSs had
little effect on HOC sorption due to their limited aqyeoys
solubilities,
CosoSveney and Transport
IsocratSc Button
Mlsctole displacement experiments were conducted to
measure the transport of three HOCs (naphthalene,
phenanthrene, anthracene), a substituted urea herbicide
(diuron), and an lonizable compound (pentachlorophenot)
with methanol-water mixtures as the mobile phase {refs.
4,5,6,8,10, Appendix A ], The objectives of these studies
were: (I) to utilize the eolumn-measyred retardation factors at
several cosolverrt contents to estimate equilibrium sorption
constants (K, mL/g) lor aqueous systems, and (li) to
determine sorption rate coefficients (kg, hr1} to the
Impacts of cosolvents on sorptlon kinetics.
Solute retardation factors were determined from
breakthrough curves measured for HOC displacement with
metrtanol-water mixtures. The column- and batch-measured
K values were generally in agreement at all cosoivent
contents, Indicating that cosolvency power can be estimated
using either technique, Extrapolation of the mixed solvent
data to fe « 0 (I.e., aqueous solutions) yielded K values that
were generally equal to or larger (by a factor of ^ 2) than the
batch-measured aqueoys K values (Figure 7). It was
suggested that the extrapolated values were, In fact, more
— w/phase
w/phase
w/phase
w/phase j
4 6
Solution Cone. (ug/mL)
Figure 6, Equilibrium isotherms for sorptlon and diyron herbicide by Webster ami Eustis soils from biphasie solvents and from
aqueous solutions saturated with several PMOS.
-------
1.5
1
as
o
-1
W
O
0,8
0,6'
0.4
0,2
°
; -'- Mtttott Dt»p!«s*ff«»nt
"°-s;. KM,
-0,2 r
=0.4
• 1.51
IT
0,5 r
0 '
| -0 S ;
-1
0 0.1 0.2 0.3 0.4 0.5 0.6 0 0.2 0,4 0,8 0.8
Dluron
, PentacMorop/mio/
2 «-,
1r
-2 t-
[ y » -4.01 I * }
**• OJ3
0 0.1 0,2 0.3 0.4 0.5 0 0.2 0.4 0.8 0.8 1
Volume Fraction
Figure 7, Com'p»8on of batch- and column-measured K far at phenanttvene, Aufoo, and as a
of ¥Oltffne
maasursmenf of sorption for strongly
hydrophobte solutes subject to several artifacts {ret. 5,
A],
The validity of the log-linear model for describing the Impact
0} cosolvents on the mobility of organic solutes Is confirmed
by the batch and the column data. For transport, this model
can be written In terms of the commonly used retardation
factor (R) as follows:
ln(Rm-1).1n(Rw-1)-aowfc (8)
where the m and w in
solvents water, respectively; other
are as previously.
The column to
during HOC with soi¥ents [refs,
6,8,10, Appendix A j. The bieontinuum sorption model
to the curves to
sorplion constants fkg). In this model, sorption Is
as a two-step initial
(Instantaneous) sorption is foitowacJ by a slower, diffusion-
towards equilibrium; k2 Is the rate
constant lor the et at,, 1990), A fog-
linear in h_ was noted with increasing volume
fraction cosolvent p ) {Figure 8), Implying sorption
kinatles In mixed solvents. This relationship was expected
on the existence of a log-log, inverse relationship
between K and Kj and log-linear, inverse relationship
K and fc. The for the
measured log k- vs. fc were In agreement with
Independent estimates on simple empirical eqyations.
AJso, the measured kz values in aqueous solutions compared
well with the estimates on extrapolation of the mixed-
solvents data, lending further support to the log-linear model
ft proposed that sorption kinetics are controlled by
sorbate diffusion within the sorbent organic matter; the
of a cosolvent to after the conformation of
the polymeric matrix of the organic matter, hence, the rate of
solute diffusion [rets. 4,5.6, A }.
wtih HOC transport in mixed
used time-invariant (I.e.,
elutton). From a point of view, environmental
involving
a constantly phase. Thus, 8 is
importan? to in* baftavtor of contaminants
of cosofvent composlfion. Therefor©, ws
the utility of §rad»nt-eiutlon
chramatograpriy Jar HOC
under such conditens [ret, 11, Appendix A },
HOCs were through laboratory soil
columns ysing isocratic- and gradisnt-slutlon
chrornatography techniques. The mobile phase consisted of
binary mixtures of water and methane! or acetonltrlle. First,
Isoeratie elutlons were used to confirm the log-linear
-------
W
1.8
1.4
1.3
1.2
1.1
1
0.9
0.*
0.4
0.2
Naphthalene
logy- 1.67 x* 0.87
r*-0.9S
0.1
0.2
0.3
-0.2
-0.4
Phenanthrerts
tog y-2.10 x- 1.20;
! r*« 0,97
0.4
O.S
0.8
0.7
0
-0.1
-0,2
-0.3
-0,4
-0,5
-O.g
0.2
0,8
Diuron
logy 1.24 x-0.89
t'» 0.88
0.1
O,2
8.3
0.4
Volume Fraction Methanol
Figure 8. Relationship between sorption rats coefficient (kg.hr"1} and volume fractal eosolvent for sorpfon of several HOCs by Eustts soil from
malhanol-water mixtures.
functionality between solvent composition and the retention
(actor, k (equivalent to R-1). Second, retention factors were
measured for elutlons with several linear solvent gradients.
Third, the following equation, adapted from gradtent-eluflon
theory, was used to predtet observed R values:
(9)
e
where R Is the retardation factor (dimensionless) measured
for a specified gradient elution, v is average pore-water
velocity (cm/mln), og c is the eosolveney power, b Is the slope
of the linear solvent gradient (mirr1){i.e., the constant rate at
which the fraction cosolvent content Is changed), L is the
column length (cm), p Is the soil bulk density fg/cm3), e is the
volumetric liquid content (cm'/cm*), and K, is the sorption
coefficient (mUg) value at the initial fraction cosolvent when
the gradient elution is initiated. The sorption and cosolveney
parameters for this model were measured independently
from batch or isocratfc elution data.
In general, the agreement was good between measured and
predicted capacity factors (Flgor® 9). These results suggest
that gradient-elation techniques and theory developed from
reversed-phase chromatography are also ysefyl for
investigating eosolvent effects on HOG sorption and transport
in soils and aquifer media.
Practical Applications of Project Findings
Magnitude of Cosolvent Effects
A vartety of organic cosolvents may be expected at and near
waste disposal sites, especially If codlsposal of a number of
wastes had been practiced. However, protocols for site
Investigations usually do not call for monitoring of organic
eosolvents In groundwater. Ste-spectfic Information on
concentrations and types of cosolvents that may be present
can be surmised only ff waste composition Is known. Thus,
estimating the magnitude of eosolvent Impacts in enhancing
solubility and decreasing sorption/retardation is often difficult.
Although direct field-scale evidence for cosolveney Is lacking,
anecdotal evidence Irani site investigations suggests thai
organic cosolvents are present and may have contributed to
facilitated transport of organic contaminants at waste
disposal facilities.
This study was aimed at developing the necessary data base
from laboratory studies and theoretical approaches that can
be used in evaluating the likely magnitude of cosolvent
10
-------
2.4
1.8
Borden Aquifer
* Dlurw
A Anthr»c»n»
O f"yr««*
Eusf is Soli
TJ
©
«»•*
O
""5
9
A Anthracene
0,8 1.8
k
2,4
of and (k) fee of so!
column with mixtures.
effects, Our and model calculations suggest that
solubility enhancement for most organic contaminants Is
likely to be small (< 20% increase) as long as cosolvent
concentrations in pore water are < 2% by volume {or about
20,000 rog/L). Thus, of champ-dynamic
are to be only in proximity to a
sits (La., the "near-f»W* high
of are likely. With increasing
from the site, cosolvent should diminish in
proportion to decreasing cosofvent concentrations.
Partially-rritecto'e coso'vents (e.g., MTBE, o-creso!,
have greaser cosotvencv to
(e.g., rnsthanc!, ethanol acetone} and as
a they should be to or
retafdaiion to the a! much
concentrations. Usually equivalent in retardation
may be noted with a partially-miscibte organic solvent at
concentrations 2 to 4 times than that of a eomptetely-
mlscibte organic solvent Properties, such as high
solubily, of the polar-PMOS that them to
cosoivaney than nonpolar-PMOS also them to ba
and to ha» a grealw mobility in
aquifers. Tfim, of polar-PMOSs
be at from
as
alcohols, are generally not by and they can
readily teach with water, Htnce, cosolvent of CMOSs
and polar-PMOSs may be of concern in the lar-flekf regions
as well,
Our Raoult's law can be to
erf in
in wrth mtiti-componsnt
(NAPLs) as fuel.
Even though are mixtures, the
from with
is sufficiently to be for most
Equilibrium or
of a fuel constituent in groundwater, Cwt (mg/L), cart 130
as follows:
C «F C / K 1 l'10^
C,, Is the ieng/L} of the constituent in the
fuel and Iv, is the foei-wattr can
be using eq {?),
For as
(a CMOS) or nrwtrtyl-tertiafy butyl (a
11
-------
PMOS), solubility enhancement of aromatic hydrocarbons by
these eosolvents appears to be small. The concentrations of
the oxygenated additives in fuels on the market now Is
usually less than 10% by volume. Assuming that residual
NAPLs in the saturated zone do not occupy more than 20-
30% of the pore spaces, the likely concentrations of the
oxygenated additives in groundwater will be less than 1% by
volume. Our fuel-water partitioning data are consistent with
this expectation.
Altornatl¥6 Fo@te
Consideration is being given to wide-scale use of fuels that
have much greater quantities of oxygenates, especially
methanoi. For example, the often mentioned "alternative
fuel* M-85 is a 85:15 mixture of methanoi and gasoline. This
mixture will be completely misclble in water, and the aromatic
constituents are likely to be transported with little or no
retardation as a result of cosolvency from high methanoi
concentrations.
Consider a hypothetical scenario of p-xylene transport in
groundwafer as a result of a continuous teak of gasoline
containing 10% methanoi contrasted to a leak of M-85 fuel.
The volume fraction of methanoi in groundwater would be
about 1 % in the first case and about 80% in the latter case,
Assuming a = 1 and 0e = 2.5, eq (8) would predict an 8-fold
increase in the mobility of p-xylene as a result of the
presence of methanoi from the M-85 fuel. For the case of M-
10 fuel, the enhancement in p-xylene mobility woyld be too
small to be measurable and may be neglected for all pratcical
purposes,
In the foregoing analysis we have assumed that the methanoi
content would be constant alt along the flow path; this
assumption would be reasonable for a large fuel spill/leak.
For a smaller spill/teak, however, the methanoi content in
groundwater will decrease with increasing distance. For such
a case, the concepts developed for gradient-elation {eq 9)
could be used to provide an approximate estimate of the
likely reduction In retardation.
Accidental release of alternative fuels such as M-85 into
rivers or other surface water bodies with contaminated
sediments might also result in the release of highly-
hydrophobic organic contaminants (such as PCBs, PBBs,
dioxlns, etc.) Into the water column. The environmental
consequences of release of such contaminants and their
transport further downstream ought to be carefully evaluated.
Remediation of Contaminated Soils
The remediation of contaminated soils or sediments removed
from waste disposal/spill sites by extraction with organic
solvents Is being evaluated at present. The data collected
during this project may be used in developing criteria for the
selection of solvents or solvent mixtures to achieve optimal
extraction of the contaminants of interest. For example,
other factors being equal, it may be preferable to choose a
solvent with greater cosoivency power so that the
contaminants can be extracted using less solvent, tf
economic or other constraints prevent the use of pure
organic solvents and a mixture of water and cosolvents are to
be used, the log-linear cosolvency model might provide
reasonable estimates of the level of extraction achievable
with various combinations of cosolvents. If a CMOS-water
mixture is to be used as the extracting solution, the
concentration of a CMOS needed may be decreased by
adding a small amount of a polar-PMOS (a better solvent).
Our data suggest that for a given addition of PMOS, the
reduction in the amount of CMOS is proportional to the ratios
of the cosolvency powers of the two solvents. If a mixture of
water and a polar-PMOS are to be used as trie extractant,
the addition of a small amount of a CMOS can increase the
maximum concentration of PMOS and still achieve the
level of extraction efficiency.
Laboratory Protocols
Oyr results suggest that sorptlon data (for both equilibrium
and kinetics) collected using mixed solvents can be reliably
extrapolated to estimate HOC sorption from aqueous
solutions. This is an Important finding because mixed
solvents can be ysed to facilitate the experimental
measurement of sorption and transport behavior of strongly
hydrophobic organic solutes (log K > 5). The use of mixed
solvents reduces the time needed to conduct the sorption
experiments and may also reduce various artifacts, Including
losses via volatilization, degradation, and sorption to
glassware. Our column studies with methanoi as the
cosolvent suggest that long-term exposore to this cosolvent
had little effect on sorptlve and transport properties of a
surface soil. This result further fends support to our
recommendation that mixed solvents, in particular methanol-
water mixtures, be used for experimental facilitation.
References Cited
Abrams, D.S.; Prausnltz, J.M., AlChEJ., 1975, 21,116-128.
Brusseau, M.L; Jessup, R.E.; Rao, P.S.C., £nv. Scl,
Techno!.. 1990, 24, 727-735.
Felice, L.J.; Zachara, J.M.; Schmidt, R.L.; Resch, R.T., pp.
39-41, IN: Free. 2nd Intaml. Conf, on Groundwater Quality
Research, Tylsa, OK. 1985.
Fredensiund, A.; Jones, R.; Prausnltz, J.M, AlChEJ,, 1975,
21,1086-1099,
Freeman, D.H.; Cheung, L.S.. Science, 1981, 214, 790-792,
Fu, J.K.; Luthy, R.G. J. Environ. Eng,, 1986a, 112,
328-345.
Fu, J.K.; Luthy, R.G. ASCEJ. Environ. Eng,, 1988b, 112,
346-366.
Horvath, C; Melander, W,; Molnar, I. J, Chromatogr., 1976,
125,129-156.
Melander, W; Campbell, D.E.; Horvath, C,, J. Cftromafogr,,
1978,158,215-225,
McCall, P.J.; Swann, R.L., Laskowski, D.A.; Ungen, S.M.;
Vrona, S.A.; Dishburger, H. So/I. Environ, Contain. ToxicoL,
1980,24,190.
Mctntyre, W.G.; deFur, P.O., Chemosphere, 1985,14,103-
112,
12
-------
Morris, K.R; Abramowitz, R,; Final, Ft; Davis, P.; Yalkowsky,
S.H. Chemosphere, 1988,17,285-288.
Nkedl-KIzza, P.; Rao, P.S.C.; Homsby, A.Q. Env. Scl,
Techno!., 1985,19, 975-979.
Nkedl-KIzza, P.; Rao, P.S.C,; Hamsby, A.G. Env, Scl.
Techno/,, 1987, 21,1107-1111.
Nkedi-Kizza, P.; Bmsseau, M.L.; Rao, P.S.C.; Homsby, A.Q.
Env, Sci. Techno!., 1989, 23, 814-820,
Rao, P.S.G.; Lee, LS. pp. 457-471, IN: Proc. 24th Hartford
Ufa Scl. Symp., DOE Symp. Series 62, Battelle PNL,
Richfand, WA, 1987,
Rao, P.S.C.; Nkedi-Kizza, P.; Davidson, J.M. pp, 63-72, IN:
Land Treatment: A Hazardous Waste Management
Alternative, Water Research Symp, Series 13, Austin, TX,
1986.
Rao, P.S.C.; Hornsby, A.G.; Kllerease, D.P.; Nkedt-Kizza, P.
J, Env. Qua/., 1985,14, 378-383,
Rao, P.S.C.; Rhue, R.D.; Johnston, C.T.; Ogwada, R.A.
Project Completion Report, ESL-TR-S8-G2, ESL/AFESC,
USAF, Tyndall Air Force Base, FL 1988, 196 p.
Reichardt, C. Solvents and Solvent Effects In Organic
Chemistry, VCH Publishers, W&inhelm, Germany, 1988,
Rubino, J.T.; Yaikowsky, S.H, J. Paren. Sci, Techno!., 1985,
39, 106-111,
Rubino, J.T; Yalkowsky, S.H. Pharm, Res,, 1987a, 4, 220-
230,
Rybino, J.T.; Yalkowsky, S.H, Pharm. Res,, 1987b, 4, 231-
236.
Rubino, J.T,; Yalkowsky, S.H. J. Paren. Scl Tachnol,, 1987c,
41, 172-176.
Rubino, J.T,; Blanchard, J,; Yalkowsky, S.H, J, Paren. Sci.
Techno!,, 1984, 38,215-221.
Swarm, R.L; McCall, P.J.; Laskowski, D.A.; Dlshburger, H.,
pp, 43-48, IN: Toxicology and Hazard Assessment: Fourth
Conference, ASJM STP 737, Philadelphia, PA, 1981.
Szabo, Q.; Prosser, 8 L: Bufman, R.A. Chemosphere,
19903,21,495-506,
Szabo, G.; Prosser, S.L; Bulman, R.A, Chemosphere,
19906,21.729-740.
Szabo, G.; Prosser,, S.L.; Bulman, R.A. Ghamssphere,
1990c,21,777-787,
Veith, J.T.; Austin, N.M.; Morris, R.T., Water Res., 1979,13,
43-47.
Walters, R.; Guissepi-EIIie, A, Env. ScL TechnoL, 1988, 22,
819-825.
Walters, R.; Ostazeskl, S.A.; Gutesepi-Ellie, A, Env, Sci.
Techno!,, 1989, 23,480-484,
Wilson, G.M.; Deal, C.H., InA Big. Chem. Fundam, 1962,1,
20-23,
Woodburn, K.W.; Rao, P.S.C.; Fuky!, M.; Nkedi-Kizza, P, J.
Confam. HydroL, 1986,1, 227-241,
Yaikowsky, S.H, Solubility of Organic Solutes In Mixed
Aqueous Solvents, CRf 811852-01, U.S. Env, Prot. Agency,
Ada, OK, 1985,
Yalkowsky, S.H. Solubility of Organic Solutes in Mixed
Aqueous Solvents, CR# 812581-01, U.S. Env, Prot. Agency,
Ada, OK, 1987.
Yalkowsky, S.H.; Hoseman, T. pp, 91-134, IN; Techniques of
Solubillzation of Drugs, Yalkowsky, S.H., Ed,; Marcel Dekker,
Inc., New York, 1981.
Yalkowsky, S.H.; Rubino, J.T, J, Pharmacoi. Sci, 74,1984,
416-421.
Zachara, J.A.; Ainsworth, C.C.; Schmidt, R.L; Reseh.C.T. J.
Contam. Hydroi, 1988,1, 343-364.
Zachara, J.A.; Ainsworth, C.C.; Cowan, C.E.; Thomas, B.L
Env. Scl, TechnoL, 1987, 21, 397-402.
Appendix-A (Published Papers)
(1) Sorption and Transport of Organic Pollutants at Waste
Disposal Sites. 1989, P.S.C. Rao, L.S. Lee, P. Nkedi-
Kizza, and S.H, Yalkowshy. pp. 176-192, IN: Toxic
Organic Chemicals In Porous Mediat Z. Gerstl, Y. Chen,
U. Mingelgrin, and 8. Yaron feds.), Springer-Verlag,
Berlin, Germany.
(2) Cosolveney of Partlally-Misclble Organic Solvents on the
Solubility of Hydrophobic Organic Chemicals, 1990, R.
Final, P.S.C. Rao, L.S. Lee, and P.V. Gline. Env. Sci,
Techno!., 24:639-647,
(3) Cosolveney and Sorption of Hydrophobic Organic
Chemicals. 1990, P.S.C.Rao, L.S. Lee, and R, Pinai.
Env. Scl. Techno!,, 24:647-654.
(4) Influence of Solvent and Sortwnt Characteristics on
Distribution of Pentaohloro-phenol in Octanol-Water and
Soil-Water Systems, 1990, L.S. Lee, P.S.C, Rao, P.
Nkedi-Kizza, and J j. Delfino. Env. Scl, TechnoL,
24:654-661.
(5) Cosolvent Effects on Sorption and Mobility of Organic
Contaminants In Soils. 1990. A.L Wood, D.C,
Bouchard, M.L Brusseau, and P.S.C. Rao.
Chemosphere, 21:575-587.
(6) Nonequilibrium Sorption during Displacement of
Hydrophobic Organic Chemicals and 4SCa through Soil
Columns wtth Aqueous and Mixed Solvents. 1989. P.
Nkedi-Kizza, M.L, Brusseau, P.S.G. Rao and A.G.
Hornsby, Env. Scl TechnoL, 23; 814-820.
13
-------
(7) Comparison of Sorption Energetics for Hydrophobia
Organic Chemicals by Synthetic and Natural Sorbents
from Methanol/Water Solvent Mixtures. 1989. K.B.
Woodburn, LS. Lee, P.S.C. Rao, and J.J. Delfino. Env.
Sci, Fecftf)0/.,23:407-413.
(8) Nonequilibrium Sorption and Transport of Neutral and
Ionized Chlorophenols, 1991. L.S. Lee, P.S.C. Rao, and
M.L. Brusseau. Env. Sci. Techno!, (in press).
(9) Partitioning of Aromatic Constituents into Water Irom
Gasoline and Other Complex Solvent Mixtures, P.V,
Cline, J.J. Delfino, and P.S.C. Rao, 1991. Env. Sci.
Technol. (in press).
(1Q)The Influence o! Organic Cosolvents on the Sorption
Kinetics of Hydrophobic Organic Chemicals. M.L.
Brusseau, A.L.Wood, and P.S.C, Rao. Env, Sci.
TechnoL (in press).
(11) Gradient-elution Techniques for Assessing Sorption of
Hydrophobic Organic Chemicals in Solvent Mixtures.
A.L. Wood and P.S.C. Rao, 1990, Agronomy Abstracts,
p. 51.
{12) Solubility of Anthracene in Complex Solvent System.
1989. B. Gupta. Thesis, Department of Pharmaceutical
Sciences, University of Arizona, Tucson, AZ.
Appendlx-B (Manuscripts in Preparation)
(1) Solubility of Hydrophobic Compounds in Solvent
Mixtures. R. Final, P.S.C. Rao, and L.S. Lee. (for
Chemosphere).
(2} Effects of Cosolvent-Sorbent Interactions on Organic
Contaminant Sorption. P.S.C. Rao, A.P. Gamerdinger,
L.S. Lee, and R. Pinal. (for Chsmosphara).
(3) Retention of Hydrophobic Solutes on Reversed-phase
Liquid Chromatography Supports: Correlation with Solute
Topology and Hydraphobictty Indices. K.B. Woodburn,
J.J. Delfino, and P.S.C. Rao. (for Chsmosphera).
(4) Retention of Nonpolar Solutes by Reversed-phase Liquid
Chromatography Sorbents: A Study of Homologous
Series of Compounds Using the Solvophobic Theory.
K.B. Woodburn, P.S.C. Rao.» and J.J. Delfino. (for J.
Chromatography).
(5) Energetics of Hydrophobic Solute Retention on
Reversed-phase Ghromatography Supports: Effects ol
Solute, Solvent, and Sorbent Properties. K.B. Woodburn,
P.S.C. Rao., and J.J. Delfino. (for Env. Sci. Tachnol.).
(6) Energetics of Hydrophobic Solute Retention on
Reversed-phase Chromatography Supports: Evaluation
of Enthalpy-entropy Compensation Model. K.B.
Woodburn, J.J. Delfino, and P.S.C. Rao. (for Env. Sci,
TechnoL).
(7) Solubility of Anthracene in Complex Solvent System.
1990. B. Gupta, D. Mishra, and S.H. Yalkowsky. (for
Pharmaceutical Research ).
Disclaimer
The information in this document has funded wholly or
in part by the United States Environmental Protection Agency
under cooperative agreement CR-814512 with the University
of Florida. This document has been subjected to the
Agency's pear and administrative review and has been
approved for publication as an EPA document.
All research projects making conclusions or recom-
mendations based on environmentally measurements
and funded by the Environmental Protection Agency are
required to participate in the Agency Quality Assurance (QA)
program. This project conducted under an approved QA
project plan. Information on the plan and documentation of
the QA activities are available from the Principal Investigator
(P.S.C. Rao),
14
-------
United Slates
Environmental Proiectign
Agency
Center (or Environmental Research
Information
Cincinnati OH 4526i
& PAID
EPA
No, G-35
Official Business
Penalty for Private Use, S300
PlBasd mjkft all rvtewssar^ changes on the above tab*!.
flc-Mih IH copy. J'i'i fi3i"'n lu tne addfess >n ihe upper
)( VOM d» n"i wish 10 i«c««ve (hes« fepotfs CHECK HERE u,
iteti«;H (a topy (hi!, tuvi-f .ind teturo to the address in the
EPA/000/y-91/009
------- |