vvEPA
United States
Environmental Protection
Agency
Environmental
Research Laboratory
Athens, GA 30613
Research and Development EPA/600/M-89/016 May 1990
ENVIRONMENTAL
RESEARCH BRIEF
Assessing the Environmental Partitioning of Organic Acid Compounds
Chad T. Jafvert
Background
The ultimate disposition of organic compounds in the
environment is influenced tremendously by their tendency
to partition, or "adsorb" to sediment particles or other
distinct environmental components or phases. To date the
majority of sorption research for organic chemicals'has
focused on describing partitioning processes of neutral
compounds, such as the polycyclicaromatic hydrocarbons
polychlorinated biphenyls, and dioxins. Many environ-
mentally relevant organic compounds, however, contain acid
or base functional groups. Compounds containing acid
functional groups include the chlorinated phenoxyalkyl acid
herbicides, such as 2,4-D and 2,4,5-T; chloro- and nitro-
phenols, such as pentachlorophenol (POP); and sulfonated
compounds, such as azo and anthraquinone acid dyes and
linear alkylbenzenesulfonate (LAS) surfactants. Organic
bases of significance include the nitrogen heterocycles,
such as quinoline, and aromatic amines, such as benzidine.'
This Research Brief focuses on the partitioning of organic
acid compounds in octanol-water and in sediment-water
systems. The main emphasis is to characterize the
adsorption processes of these compounds "mechanis-
tically," to permit the development of quantitative structure-
activity relationships. In light of the differences in
adsorption reactions among neutral, cationic, and anionic
organic compounds, a brief theoretical description is given
as to why these compounds partition as they do in the
environment. For example, why do certain environmental
components, such as fish lipid, or the organic material in
soils and sediments preferentially adsorb neutral com-
pounds over charged species? An explanation is given
based on thermodynamic considerations, and various
assumptions and requirements necessary to quantify these
partitioning processes are examined.
Equilibrium Partitioning
If no kinetic limitations exist, the tendency to partition
expressed as an equilibrium adsorption coefficient (Krt) is
directly related to the standard free energy of adsorption
AG°ads as expressed by Equation 1. (See [1,2] for a more
thorough discussion).
AG°ads = - RT In Kd (1)
To characterize the adsorption of organic compounds to
environmental matrices, such as soils, sediments, and biota
this standard free energy" of adsorption must be factored
into adsorption components (i.e., soil organic carbon)
Adsorption to each of these components, in turn, results
from additive effects of various adsorption interactions or
energies." According to Stern [see 1,2], in the most basic
form these energies can be separated into an intrinsic
term (AGochem) and an electrostatic energy
AGoads = AG°chem + AGoe, (2)
For neutral organic compounds, the total energy of
adsorption contains no electrostatic component, and hence
is equal to the intrinsic chemical adsorption energy
(Ata°chem)- For moderately large (molecular weight > 100)
nonpolar compounds, this adsorption energy is dominated
by hydrophobic forces. Hence partitioning occurs "out" of
hydrophihc phases such as water to more hydrophobic
phases. Water solubility (corrected for crystal energy of
solids), octanol-water partitioning, connectivity indices
cnromatographic retention times, and estimated molecular
surface areas, all have been used to index chemical activity
(or free energy changes) of hydrophobic compounds for
estimating partitioning to media such as soil orqanic carbon
and fish lipid.
Because the intrinsic chemical adsorption energy (AG°rhB )
for organic compounds, as a rule, results from hydrophobic
interactions, this term for both cationic and anionic organic
-------
compounds is less than for neutral compounds of [similar
structure. For these charged species, however, the primary
change in the overall adsorption energy may resu|t from
electrostatic contributions to adsorption (AG°ei)r This
electrostatic contribution can be positive or negative and is
equal to the product of the compound's valence (z),
Faraday's constant (F), and the surface potential (ip). Most
natural soils and sediments contain mineral and organic
surfaces that possess negatively charged functional groups.
Therefore, the overall sign of AG°e| is ultimately determined
by the charge of the adsorbing ion. \
For adsorption of cationic organic compounds (z = ^ 1), the
overall sign of AG°8| is negative, and electrostatic
interactions can promote adsorption beyond what is
predicted for a similar neutral compound .based purely on
hydrophobic interactions. Examples of this in the literature
include the adsorption of benzidine (4,4-diaminobiphenyl,
pKa = 4.3) [3] and quinoline (pKa = 4.94) [4] onto
sediments or aquifer material. The adsorption of these
compounds to natural surfaces is highly pH-dependent
because of their relatively low pKa values. At pH, values
where the cationic form of these compounds predominates,
adsorption can be two orders of magnitude greater than the
hydrophobic adsorption of the corresponding (neutral
species. This type of reaction is generally referred to as
ligand (or cation) exchange, and modeling this process
relies on understanding the adsorption behavior of other
exchangeable ions, such as Ca + 2 and Mg + 2. Also
adsorption can occur to various subcomponents of soils and
sediments, such as the organic matrix and various; mineral
components. To effectively describe these processes,
model algorithms must be somewhat complex. Modeling the
adsorption of organic cations may best be accomplished by
incorporating adsorption algorithms into "electrostatic" or
ligand exchange adsorption models, similar to those used to
estimate metals adsorption.
For the adsorption of anionic organic compounds (z = -1),
the overall sign on AG°ei is positive, and electrostatic
interactions act to attenuate adsorption. Hence, adsorption
to natural sediments by the anionic form of an organic acid
is generally less than adsorption of the corresponding
neutral species. Because no specific electrostatic
interactions occur to promote adsorption, the partitioning
process must result from the existence of a significant
chemical adsorption energy term. This situation results in
adsorption occurring predominantly through hydrophobic
interactions, attenuated by electrostatic forces. Indeed, in
limited cases [5,6], the adsorption of anionic organic
compounds has been normalized to the organic carbon
content of the sediments and soils. In another study [7],
adsorption of LAS surfactants was shown to increase with
surfactant alkyl chain length through the surfactant series
investigated, indicating that hydrophobicity contributes to
the total energy of adsorption.
In quantitatively assessing these partitioning reactions,
separating the hydrophobic component of partitioning from
the electrostatic component is essentially impossible. As
'with neutral organic compounds, however, model; systems
can be used to assess and possibly quantitatively describe
sorption processes. Model systems include somie of the
same "systems" or parameters used to assess partitioning
of neutral compounds, such as solubility or octano.l-water
distribution. Indeed, many of the same factors that influence
the transfer of organic acids between octanol and water (or
affect the apparent water solubility) influence the transfer of
these compounds between the components of the more
heterogeneous system of sediment-water [8-10]. These
similar factors include the degree of dissociation and the
hydrophobicity and electrostatic influences of both sorbent
and sorbate.
As a result of these similarities, the distribution of several
organic acids between octanol and water phases has been
examined as a function of aqueous phase salt concentration
(for KCI, NaCI, LiCI, MgCI2, and CaCI2) and aqueous phase
pH [9]. Compounds that have moderately low (< 5) pKa
values were chosen. From these experiments, equilibrium
partition coefficients of the neutral and charged forms, as
well as several pKa values have been calculated. The
sediment-water distribution of selected organic acids also
has been investigated as a function of aqueous phase ionic
composition and pH [10.]. Experimental results and model
development for each system is briefly reviewed.
Octanol-Water Partitioning
The distribution of organic acids between water and octanol
in the presence of monovalent inorganic cations can be
described by Reactions 1 through 6 of Figure 1. The
reactions in Figure 1 are formulated in terms of KCI as the
monovalent salt, although NaCI and LiCI also have been
used. The mass action equations for these reactions, along
with the material balance and electroneutrality constraints,
comprise the chemical equilibrium model.
Reactions of HA
1.[HA]W = [H + ]w + [A-]w
2. [HA]W = [HA]0
Reactions in KCI, NaCI, and LiCI salt systems.
3. [K+]w
Expression
KOW
Ki
ipx
5- [K + ]w + [A-]w = [K + ]0 + [A-]0 N
6. [K + ]w + [A-]w = [K + A-]0 Kip
7. [K + ]w + [OH-]W = [K + ]0 + [OH-]0
8. [K*lw + [OH-]W = [K*(
Figure 1. Equilibrium reactions of monovalent organic
acids in octanol-water. Adapted from reference
[8]. HA represents a neutral organic molecule
and A" represents the corresponding mono-
valent organic anion. Terms subscripted with
"w" or "o" represent aqueous phase or octanol
phase concentrations, respectively.
Reaction 1 describes acid dissociation in the aqueous
phase. Reaction 2 describes transfer of the neutral species
(HA) to, the octanol phase. This reaction is the basis of lineal-
free energy relationships relating ,KOC (the sediment or soil
equilibrium partition coefficient, normalized for organic
carbon content) and Kow,(the octanol-water partition
coefficient) for sorption of nonpolar organic compounds
,2
-------
[11,12]. For any icnizable compound, this mechanism is
important when a significant fraction of the total compound
exists in the neutral form. Reactions 3 and 4 represent
transfer of the free and ion-paired inorganic species
respectively, to the octanol phase. For both KCI and NaCl'
an aqueous phase concentration of 0.2 M (with no organic
compounds present) will result in approximately 40 uM total
salt partitioning into the octanol phase. Reactions 5 and 6
describe the transfer of the free and ion-paired organic
species, respectively, and the inorganic counter-ion to the
octanol phase. The transfer of the organic ion, as a free
spec.es or as an ion-pair, is highly dependent upon the
counter-ion. Also, as the aqueous phase inorganic 'salt
concentration is increased, the transfer of the organic ion as
an ion-pair is favored.
To illustrate these effects, the overall distribution ratio (D) of
total compound in the octanol phase to total in the aqueous
phase, can be calculated:
([HA]
D =
(3)
" W
where M + represents the monovalent electrolyte K+ Na+
?r ;of°rJ3AX1periments Performed at sufficiently high pH
%"tj '' M IS an insignificant species in either phase
and the overall distribution results from Reactions 5 and 6 of
Figure 1. Using the calculated partition coefficients Kix and
Kipx for the inorganic species KCI, NaCl, or LiCI and the
measured coefficients K,, and Kip for the organic
compounds pentachlorophenol (PCP), 4-chloro-o-(4-
chlorophenyl)-benzeneacetic acid (DDA) and 2 45-
trichlorophenoxyacetic acid (2,4,5-T), we can obtain' the
overall distribution of these compounds (Figure 2) as a
function of aqueous phase salt concentration For
comparison, the concentration of the organic species [A-l
was fixed at 10 PM. and the pH value was set at 12 Fo the
salts NaCl and KCI, little difference in the calculated
distribution is seen, whereas partitioning into the octanol
obv.ously is favored for the LiCI system. Also, the
importance of the hydrophobicity of these compounds is
apparent. The magnitude of partitioning into the octanol for
the organic anions follows the sequence PCP > DDA >
2,4,5-T for all three salt systems. This is the same orderinq
as the octanol-water distribution coefficients of the
corresponding neutral species (Kow). Indeed, for ten organic
compounds investigated, including various carboxylic acids
and phenols the calculated partition coefficients for
formation of the ion-pair in the octanol phase (KiD) correlates
extremely well to Kow values [9]. p
Results of four experiments in which pH was varied at
constant ionic strength ([K + ]w = 0.1 M) are presented in
Figure 3 The curves were calculated using the partition
coefficients given in Figure 3 for the organic compounds
and calculated values for the coefficients Kix and KiDX, for
KCI. At low pH values (approximately pH 1 to 2), the nlutral
species is dominant in both phases, and the value of too D
is independent of.pH. At intermediate pH values
(approximately 4 to 6), the neutral species is dominant in
the octanol phase and the anionic species becomes
dominant in the aqueous phase, and the distribution ratio
becomes pH dependent. At high pH (above pH 9), the anion
predominates m both phases, and the distribution ratio is
again independent of pH, but dependent on ionic strength
as previously discussed. '
The significance of each component of this equilibrium
vLlT ua1-be m°re Clearly seen in a bo [concentration]
versus pH diagram as presented in Figure 4 for 4,6-dinitro-
o-cresol (DNOC). All lines in this figure are calculated from
the six equihbrium constants for DNOC and KCI, the total
mass of DNOC and KCI in the system, and the volumes of
nt?nr>P I6' The difference between the concentration of
DNOC in the octanol at low pH and the concentration in the
water at high pH is due to the difference in the volumes of
the two phases.
Sediment-Water Partitioning
As in octanol-water partitioning, the distribution of organic
acids between sediment and water is highly dependent
upon the degree of chemical dissociation. To assess the
effects of pH (and hence, dissociation) on sediment-water
distribution initial experiments have been performed using a
pH-stat. The pH-stat can be used to titrate and maintain a
sediment slurry to a given or set pH value. Over the course
of an experiment, a slurry can be sampled and the pH then
can be lowered to a new set-point sequentially, resulting in a
titration _ curve." Considerable changes in sediment
characteristics result from such a titration. This approach
however, allows factoring of the overall chemical distribution
into discrete coefficients for the neutral species (KH) and the
be defined
[HA]
sed
[A ]
K,. =
di
sed
IA ]„
with
K
(4)
(5)
(6)
[HA]..
where [HA]sed (mol/kg) and [HA]W (M) are the neutral
species concentrations in sediment and aqueous phase
respectively, and [A-]sed and [A-]w are the anionic species
sediment and aqueous phase concentrations, respectively
Because only total concentration in each phase can be
determ.ned analytically, mass balance equations for each
phase and for the total system are needed,
+ [A~] (7)
(8)
[HAJt = [HA] + (m/v)[HA] (Q)
(mrV} H-thf- sed!"lent ^ass (kg) to aqueous volume
ratio. Combination of Equations 4 through 9 results in an
phase35'0" f°r the fraCti°n °f t0ta' compound in tne aqueous
-------
log(D)
2.5
:
2.0
1.5
1.0
0.5
0.0 _
.-.4
PCP, LiCI
PCP, NaCI
PCP, KCI
DDA, LiCI
DDA, NaCI
DDA, KCI
2,4,5-T, LiCI
2,4,5-T, NaCI
2,4,5-T, KCI
-.5
0.04
I I L
I L
J I
0.08
0.12!
0.16
0.20
0.24
Total KJNa, or Li (M) in Aqueous Phase
Figure 2. The octano.-water distribution of 2,4,5-T, DDA and PCP using the ca.cu.ated partition coeff.c.ents
from experimental data, and setting [A lw at 20 y.M and the pH at 12.
-------
[HA]
f =
aq
W
[HA],
(K
(10)
(Kd{H+>
6.0-
4.0-
2.0-
0.0-
-2.0
• POP
°DDA
«2,4,5-T
Kow
5.07 -4.85
4.64 -3.66
3.31 -2.83
2.14 -4.46
log(K)
Ka K,
-2.13
-3.60
-3.76
-3.92
Kip
2.67
1.84
0.60
0.016
0.0
2.0
4.0
6.0
1 i •
8.0
• i ' •
10.0
' I ' '
12.0
14.0
PH
Figure 3. The octanol-water distribution of four compounds
as a function of aqueous phase pH. Symbols are
data points, and lines are calculated by minimizing
residuals on Kow and Ka. In experiments, the
aqueous phase contained 0.1 M KCI, and HCI or
KOH for pH adjustment. The total mass of
compound per tube in each experiment was
constant.
Results of initial studies on several organic acid compounds
using a silty sediment from the Ohio River (EPA sediment
11) suggests a definite pH dependence on Kdi [10]. In
accord with electrostatic considerations described
previously, some type of pH dependence on the adsorption
of the anion may seem unequivocal, resulting from changes
in the surface charge or degree of protonation of the organic
carbon matrix. For the compounds studied on EPA sediment
11 to date, Kdi can be expressed as
= a.(pH) + logK.
(11)
where, the average pH dependence or slope fa) is equal to
0.30, and K, is the "intrinsic" chemical adsorption term
(given in Table 1 for each compound). Only one sediment
system was examined, however, and clearly more evidence
is needed to confirm this general pH dependence. In total 5
parameters are used in Equations 10 and 11. One constant,
pKa, is intrinsic to the adsorbate; two constants, Kd and K,,
describe the interactions between adsorbate and adsorbent-
one constant, a,, is contained in Kdi (along with «,) and is
intrinsic to the adsorbent; and pH (or {H+}) is the
independent variable.
The experimental data for four compounds are shown in
Figure 5, along with model results using the constants given
in Table 1. For POP, DDA and 2-(2,4,5-trichlorophenoxy)-
-2.0-
-4.0-
ra
8 -6.0-
o
O
- -8.0-
-10.0-
(=) Total Measured in
Octariol Phase
/ (•) Total Measured in
0 6 9-,
Total in Water
[HA + ATw.
Water Phase
\..»j»5>^ N. Total in Octan
\IHA]W
0.0
Figure 4.
. . i ,
2.0 4.0 6.0 8.0
Aqueous Phase pH
10.0
pC - pH diagram of DNOC species partitioning
between octanol and water as a function of pH.
Symbols are data points and lines are model fit. The
volumes of each aqueous phase = 10 ml, volume of
octanol = 2.5 ml, total mass of DNOC in each tube
= 1.25x10'6 moles. The aqueous phase contains 01
M KCI, and HCI or KOH for pH adjustment.
Table 1. Constants Used in Modeling the Adsorption of
Organic Acid Compounds to Sediment 11
i\a
PCP
DDA
DNOC
Silvex
0.05
0.08
0.10
0.10
4.85
3.66
4.46
3.07
" u
3.0
2.45
1.93
1.87
3.33
2.31
2.83
2.03
aSediment mass to water volume ratio (kg/L)
faReference [9].
propionic acid (silvex), organic carbon-normalized partition
coefficients for the neutral species (Koc) are, within a factor
of two, equal to the Kovv. The K0? for DNOC is significantly
greater than its Kow, likely resuming from polar interactions
with the sediment by the nitro-groups on DNOC. A similar
trend exists for the anionic species.
Summary
Partitioning in octanol-water and sediment-water systems by
organic compounds that contain a single, relatively strong (2
< pKa < 5) acidic functional group has been examined.
Clearly, a significant conclusion of this work is that, at pH
values found in most natural environments (i.e., 5 to 9)
partitioning can occur by both the neutral and ionic forms of
these and similar compounds Therefore, to determine a
priori chemical distribution in the environment, partitioning
must be factored into measurable or predictable quantities
for each adsorbing species, as has been attempted here.
The adsorption of neutral organic compounds is the
simplest case, wherein the adsorption energy is largely
determined by hydrophobic interactions. Octanol-water
distribution (Kow), in turn, has been used as a measure of
chemical hydrophobicity. Octanol-water distribution for
organic anions is more complex and is a function of both
chemical hydrophobicity and inorganic ion composition. For
-------
the partitioning to sediments, an "ionic" or electrostatic
effect also is observed for organic ions. This effect is (most
obvious as a result of varying the pH of sediment systems,
and may result from changes in surface properties |(i.e.,
charge). These and other effects that determine; the
adsorption of organic anions to natural materials must be
examined further. :
• POP
o DNOC
• DDA
A Silvex
Figure 5. The pH-dependent adsorption of four organic acid
compounds to EPA sediment 11 (organic carbon =
1.77%), and the predicted partitioning using j
Equations 10 and 11, and the values of Kd, Kj, and
Ka given in Table 1, with oi equal to -0.3. !
8. Westall, J.C., C. Leuenberger, and R.P. Schwarzenbach.
1985. Influence of pH and ionic strength on the aqueous-
nonaqueous distribution of chlorinated phenols. Environ. Sci.
Technol., 19:193-198.
9. Jafvert, C.T., J.C. Westall, E. Grieder, and R.P.
Schwarzenbach. 1989. Distribution of hydrophobic ionogenic
organic compounds between octanol and water: organic
acids. To be submitted for publication.
10. Jafvert, C.T. 1989. Sorption of organic acid compounds
to sediments: initial model development. Submitted for
publication.
11. Leo, A. Hansch, C. Elkins, D. 1971. Partition coefficients
and their uses. Chemical Reviews. 71:525-616.
12 Lyman, W.J. Loreti, C.P. Prediction of Soil and
Sediment Sorption for Organic Compounds. Final Draft
Report to Office of Water, U.S. EPA, Washington D.C.,
1986.
References \
1. Stumn, W. and J.J. Morgan. 1981. Aquatic Chemistry, An
Introduction Emphasizing Chemical Equilibria in Natural
Waters, 2nd Ed. John Wiley and Sons, New York.
2. Morel, F.M.M. 1983. Principles of Aquatic Chemistry
John Wiley and Sons, New York.
3. Zierath D.L. J.J. Hassett, W.L. Banwart, S.G. Wood, and
J.C. Means. 1980. Sorption of benzidine by sediments and
soils. Soil Science. 129:277-281. |
4. Zachara J.M., C.C. Ainsworth, L.J. Felice, -and C.T.
Resch. 1986. Quinoline sorption to subsurface materials:
Role of pH and retention of the organic cation. Environ. Sci.
Technol., 20:620-627.
5. Nkedi-Kizza, P., P.S.C. Rao, and J.W. Johnson.: 1983.
Adsorption of diuron and 2,4,5-T on soil particle size
separates. J. Environ. Qua/. 12:195-197. ,
6. Ogram, A. V., R.E. Jessups, L.T. On, and P.S.C. Rao.
1985. Effects of sorption on biological degradation rates of
(2,4-dichlorophenoxy) acetic acid in soils. Appl. Environ.
Microbiol. 49:582-587.
7. Hand, V.C., G.K. Williams. 1987. Structure-activity
relationships for sorption of linear alkylbenzenesulfonates.
Environ. Sci. Technol., 21:370-373. '
-------
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information j
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/M-89/016
------- |