MICHIGAN IIIIIIIIIIIIIIIIIIIIII

miss illWill!(HWWW!!!{Hill}!HIHIHUIHHIHII

93 01561 0987

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled
W“ ofi Salami». m Sou
(pmpoglnon cm Seam i ' (new
Md Newton/tic. W

presented by

W OMWKWQ

has been accepted towards fulfillmentl . .-
of the requirements for

mQSER degree in SOl|SCARWC€

Major professor

Date 3136/96

0—7639 MS U i: an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

 

PLACE ll RETURN BOX to romovo this checkout from your ncord.
TO AVOID FINES rotum on or baton duo duo.

 

DATE DUE DATE DUE DATE DUE

mama '1111312
{3‘ “i a '

 

 

 

i—Ti:

l___l:
II I

MSU ioAn Affirmative Action/EM Opporhmky Initiation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

EFFECT OF SOLUTION AND SOIL
COMPOSITION ON SORPTI ON OF IONIZABLE
AND NONIONIC COMPOUNDS

By
Wendy Ona Werkheiser

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Crop and Soil Sciences
1996

ABSTRACT

EFFECT OF SOLUTION AND SOIL COMPOSITION ON SORPTION
OF IONIZABLE AND NONIONIC COMPOUNDS

By

Wendy Ona Werkheiser

Solution and soil composition affect sorption of ionizable (primisulfuron) and
nonionic (TCE and toluene) compounds. Sorption of primisulfuron increased with
decreasing pH. At pH 4.5, primisulfuron sorption increased with increasing 0C. In soils
with less than 1% OC, sorption at pH 4.5 increased with increasing Fe oxide content.
Surfactant caused increased sorption in a soil with 0.7% DC, but decreased sorption in a
soil with 1.7% OC. About 50% of sorbed primisulfuron did not desorb during a single 24—
h batch desorption. Sorption estimates based on log-log relations between water solubility
and K0c overestimated K“: by a factor of 4 to 40, even when K0c was corrected for
ionization of primisulfuron. Sorption of TCE and toluene by a soil with 2.1% DC
decreased log-linearly with increasing volume fraction methanol, although deviations from
log-linear behavior were noticed. Without accurate prediction of solubility enhancement, it

is impossible to predict sorption in cosolvent solutions.

ACKNOWLEDGMENTS

The research reported in Chapter 1 of this thesis was supported by the USDA National
Research Initiative Competitive Grants Program, Grant 92-37102-7405 and by the
Michigan Agricultural Experiment Station. Gratitude is expressed to Drs. Yan Pan and
Paul Laconto of the EPA Hazardous Substance Research Center at MSU for their

assisstance and cooperation in the headspace gas chromatography analyses in Chapter 2.

iii

TABLE OF CONTENTS

List of Tables ................................................................................. v
List of Figures ............................................................................... vi
Chapter 1. Effect of Soil Properties and Surfactant on Primisulfuron Sorption ...... 1
Abstract .................................................................................... 1
Introduction ............................................................................... 2
Materials and Methods ................................................................... 5
Materials ............................................................................... 5
Batch Sorption Isotherms ........................................................... 6
Batch Desorption Isotherms ......................................................... 8

Data Analysis ......................................................................... 8
Results and Discussion .................................................................. 9
Effect of QC on Sorption at pH 4.5 ................................................ 9
Effect of Iron Oxides on Sorption ................................................ 13
Effect of pH on Sorption .......................................................... 14
Effect of Surfactant on Sorption .................................................. l4
Desorption ........................................................................... 17
Conclusions ............................................................................. 19
References .............................................................................. 20

Chapter 2. Methanol Cosolvent Effects on TCE and Toluene

Solubility and Sorption ............................................... 23

Abstract .................................................................................. 23
Introduction ............................................................................. 24
Materials and Methods ................................................................. 27
Materials ............................................................................. 27
Solubility Enhancement Experiments ............................................ 27
Batch Sorption Isotherms ......................................................... 28
Gas Chromatography .............................................................. 29
Data Analysis ....................................................................... 29
Results ................................................................................... 29
Solubility Enhancement Experiments ............................................ 29
Batch Sorption Isotherms ......................................................... 32
Conclusions ............................................................................. 34
References .............................................................................. 35
Chapter 3. Conclusions ................................................................... 37

iv

LIST OF TABLES

Chapter 1

Table 1. Primisulfuron sorption coefficients (KI) and K0‘ values
at pH 4.51:0.1, and relevant soil properties (n=3) ........................... 10

Table 2. Primisulfuron sorption coefficients (K,; L kg“) at various
pH, with and without surfactant (X-77) ...................................... 15

Table 3. Primisulfuron sorption (K,) and desorption (KMP)
coefficients at natural soil pH. ................................................. 18

Chapter 2
Table 1. Solute Properties ................................................................ 27

Table 2. TCE and toluene sorption coefficients (Km) by Capac soil
at low volume fractions (f c) of methanol ..................................... 32

Table 3. Toluene sorption coefficients (K) by Kalamazoo soil
at low volume fractions (f J of methanol ..................................... 34

LIST OF FIGURES

Chapter 1

Figure 1. Structure of neutral and anionic forms of primisulfuron,
showing delocalization of electron density in the anionic species .......... 3

Figure 2. Sorption coefficient as a function of soil organic C
concentration for the rve mineral soils ...................................... 11

Figure 3. Effect of pH (top axis), fraction of neutral primisulfuron
(bottom axis), and X-77 surfactant on primisulfuron sorption
coefficients for Haplaquept (1.7% DC) and Hapludalf

(0.7% DC) soil samples. ...................................................... 16
Chapter 2
Figure 1. Log solubility of TCE as a function of volume fraction methanol ........ 31

Figure 2. Log solubility of toluene as a function of volume fraction methanol ..... 31

Figure 3. Log TCE sorption coefficient Km for Capac soil
as a function of methanol fc. .................................................. 33

Figure 4. Log toluene sorption coefficient K,n for Capac soil
as a function of methanol fc. .................................................. 33

vi

Chapter 1.

Effect of Soil Properties and Surfactant on Primisulfuron Sorption

ABSTRACT

Soil properties affect the environmental fate of herbicides. Little is known about
how soil organic matter and pH affect sorption and fate of the sulfonylurea primisulfuron
(2-[3-(4,6-bis(difiuoromethoxy)-pyn'midin-2—yl]-benzoic acid methyl ester).
Primisulfuron sorption and desorption by six soils with different organic carbon (0C),
iron oxide, and clay contents were investigated using batch reaction at controlled pH.
Sorption was measured at a soil:solution ratio of 1:2 with 0.01 M CaC12 as a background
electrolyte. Sorption and desorption isotherms were linear. When pH was adjusted to
values between 4 and 6.5, sorption coefficients decreased with increasing pH. At pH 4.5,
sorption coefficients increased with increasing OC and ranged from 0.6 for a soil with 0.2
% CC to 15.5 for a soil with 48 % OC. Surfactant (Triton X-77, 0.25% v v") caused
primisulfuron sorption to increase in a low-0C soil but caused decreased sorption in a
soil with 1.7 % OC. Sorption coefficients were correlated with dithionite-extractable iron
in soils with less than 1 % DC, but not in soils with greater OC. Desorption coefficients
at native soil pH were up to five times greater than were sorption coefficients; about 50 %
of sorbed primisulfuron did not desorb during a single 24-h batch desorption reaction.

Primisulfuron is most likely to leach in soils with high pH or low organic matter contents.

INTRODUCTION

Primisulfuron, trade name Beacon (CGA-136872), is a relatively new sulfonylurea
herbicide used to control grass and broad-leaved weeds in corn. In order to predict
primisulfuron fate in the environment, it is essential to understand the effect of soil
properties on primisulfuron sorption. Sorption may affect leaching, chemical and
microbial degradation, and carryover to subsequent crops. Primisulfuron is a
moderately weak acid with a pKa of 5.1 (US. Environmental Protection Agency, 1990)
and a low, pH-dependent water solubility (Fig. 1).

Primisulfuron sorption depends on the affinity of soil for both the neutral molecule,
which predominates when pH<pKa, and the anionic species, which is dominant when
pH>pK.. Neutral primisulfuron may be sorbed by soil organic matter, whereas the
anionic species may be sorbed by protonated hydroxyl groups on iron and aluminum
oxides. As pH increases, the proportion of neutral primisulfuron decreases, which causes
sorption by soil organic matter to decrease. The proportion of anionic primisulfuron
increases, yet the concentration of positively charged sorption sites decreases, which may
cause sorption of anionic primisulfuron to decrease as pH increases. Thus, net sorption
of primisulfuron should decrease as pH becomes less acidic.

No comprehensive, systematic studies of the effect of soil properties on
primisulfuron sorption appear to have been published, but more is known about sorption
of chlorsulfuron, an older sulfonylurea herbicide with pKa near 3.7 (Zahnow, 1982;
Shea, 1986). When soil pH was increased by long-term liming, chlorsulfuron sorption
decreased with increasing pH (Nicholls and Evans, 1985). A similar decrease in
sulfonylurea sorption with increasing pH has been reported when pH was adjusted to
values between 4 and 8 (Mersie and Foy, 1985; Shea, 1986; Wehtje et al., 1987;
Borggaard and Streibi g, 1989).

Sorption of nonionic herbicides generally increases with increasing soil organic

carbon (0C) content, but the relationship between OC and sorption of weakly acidic

0 fit N/O\CHF pK 51H4<N:§O\CHF2
‘—a—-> O/NY N
H30 WE WNZO/ CHF2 H30 0 00. EC‘Y - ‘10 I“ O/CHFz
Neutral Anionic
Sw = 0.37 ppm (0.79 pM) at pH 4.5 SW =70 ppm (149 pM) at pH 7

SW =600 ppm (1281 pM) at pH 8

Figure 1. Structure of neutral and anionic forms of primisulfuron, showing

delocalization of electron density in the anionic species.

herbicides such as sulfonylureas is more complex. Humic acid sorbs chlorsulfuron
strongly at low pH (K. = 35 at pH 4.2), but sorption decreases exponentially with
increasing pH, paralleling the decrease in both neutral herbicide and in undissociated
functional groups on humic acid (Borggaard and Streibi g, 1988). Some authors have
reported a positive correlation between chlorsulfuron sorption and soil 0C in soils of
varying pH (Mersie and Foy, 1985; 1986; Walker et al., 1989). In these studies,
however, soil OC generally increased as soil pH decreased (Mersie and Foy, 1985; 1986;
Walker et al., 1989), and there was no trend when soils with > 20 % organic matter were
excluded (Walker et al., 1989). Even for soils at the same pH, chlorsulfuron sorption was
greater in a soil with 0.8 % organic matter than in a soil with 1.3 % organic matter
(Nicholls and Evans, 1985). The absence of a relationship between chlorsulfuron
sorption and soil organic matter indicates that soil components other than soil organic
matter likely play an important role in chlorsulfuron sorption.

To determine the effect of soil 0C and of Fe and Al oxides, Borggaard and Streibi g
(1989) measured chlorsulfuron sorption by ten soil samples at pH 4.2 and 7 after a series
of sequential extractions to remove Fe and Al incrementally. Multiple regression
analysis showed that sorption coefficients at each pH depended on soil Fe and AI
concentrations, as well as on soil OC (Borggaard and Streibi g, 1989). On a mass basis,
amorphous Fe(OH)3 sorbs ten times as much chlorsulfuron as does goethite, but only
one-fourth as much as does humic acid (Borgaard and Streibi g, 1988).

Because sulfonylurea sorption depends in part on interactions between the neutral
herbicide and soil organic matter, particularly at pH < 5, sorption may be affected by
surfactants (wetting agents) that are used in the field to increase herbicide foliar
absorption and water solubility. If surfactants cause substantial increases or decreases in
herbicide sorption, then sorption coefficients measured without surfactant would lead to
incorrect predictions of herbicide mobility and availability in the field. Sun and Boyd

(1993) have reported that surfactants cause decreased sorption of compounds with low

water solubility and high sorption coefficients, but cause increased sorption of
compounds with higher water solubilities and lower sorption coefficients. In soils with
little organic carbon, sorption of the surfactant itself may substantially increase soil
organic C and thus cause an increase in herbicide sorption (Sun and Boyd, 1993).

The objectives of this study were to determine the effect of pH, organic matter,
variable-charge minerals, and surfactant (Triton X-77) on primisulfuron sorption by
several soils. Careful selection of soils and choice of reaction pH permits unambiguous
interpretation of the effect of soil pH and QC on primisulfuron sorption. Desorption was

measured at native soil pH to determine the amount of sorption hysteresis.

MATERIALS AND METHODS
Materials

l4C-phenyl-labeled primisulfuron with a specific activity of 52.4 pCi mg" and a
radiochemical purity of 94.3 % was obtained from Ciba-Geigy (Greensborough, NC).
Non-radioactive analytical grade (99% pure) primisulfuron also was obtained from Ciba-
Geigy. Triton X-77, a non-ionic surfactant typically applied with primisulfuron in the
field (0.25% v v") was obtained from Valent Corp. All other chemicals were analytical
grade.

Samples of Kalamazoo sandy clay loam A and B horizons (fine-loamy, mixed,
mesic Typic Hapludalfs), Misteguay clay A horizon [fine, mixed (calcareous), mesic
Aerie Haplaquepts], and Houghton muck (euic, mesic Typic Medisaprists) were collected
from Michigan State University research farms. Samples of two Oxisol B horizon
samples [fine (siliceous), isothermic Typic Haplorthoxes, and very-fine (oxidic),
isothermic Typic Haplorthoxes] were collected in Lavras, Brazil and provided by Dr. J.
M. Lima (Univ. Fed., Lavras). These soils were selected to provide a range of OC and

iron oxide concentrations (Table 1). The soils were air-dried and passed through a 2-mm

sieve. Organic carbon (OC) was determined by the Walkley-Black method (Nelson and
Sommers, 1982), dithionite-extractable Fe (FeDCB) was determined using the method of
Jackson et al. (1986), and clay content was determined by the hydrometer method (Gee
and Bauder, 1986).

Batch Sorption Isotherms

The effect of soil OC and extractable Fe on primisulfuron sorption was
determined by batch sorption experiments at pH 4.5. This pH was chosen in part because
four of the six soil samples had supernatant solution pH values near 4.5 when neither acid
nor base was added, and in part because neutral primisulfuron composes 80% of total
solution-phase primisulfuron at this pH. The effect of pH on primisulfuron sorption by
four of the soils was determined by adjusting suspension pH to a range of values between
4 and 6.5. The amounts of acid or base required to bring suspension pH to the desired
value were determined in preliminary experiments. To measure primisulfuron sorption,
5.0-g subsamples of each soil ( 1.0 g for the muck soil) were weighed into screw-top, 25—
mL glass Corex centrifuge tubes in triplicate for each herbicide concentration and pH.
Solutions were prepared by adding either HCl or NaOH to 0.01 M CaClz solutions so
that each soil suspension would be at the desired pH after the solutions and soils were
shaken for 24 h; CaCl2 was used as a background electrolyte to minimize
dissolution/dispersion of organic matter. Ten mL of appropriate solution was added to
soil subsamples in the centrifuge tubes to give a 1:2 soil:solution ratio (1:10 for muck
soil). Tubes then were sealed with tefion-lined caps and shaken for 1 h to bring each soil
suspension to the desired pH before primisulfuron was added. For each soil and pH, four
of the tubes were centrifuged, the supernatant solution pH was measured, and soil was
resuspended using a vortex mixer. Finally, aliquots of 1000 mg L" analytical grade
primisulfuron and 5.0 ”L of 100 mg L“ l4C-primisulfuron (both in methanol) were added
to all tubes to give total radioactivity of 2.6 yCi L'1 and eight initial herbicide

concentrations ranging from 0.05 to 1.0 mg/L (for pH < 5.5; 0.11 to 2.13 yM) or 0.05 to
5 mg L" (for pH 5.5 to 6.5; 0.11 to 10.67 pM). The lowest initial primisulfuron
concentration was equivalent to 100 pg kg" soil (0.21 pmol kg"), a typical field
application rate. For each concentration and pH, a tube with solution only (that is, soil-
free "blank') was prepared. For each soil and pH, triplicate samples were also prepared
with no primisulfuron so that the concentration of dissolved and colloidal organic C
(DOC) could be measured without releasing l“C-CO, into the laboratory air. Methanol-
containing samples explode in the hi gh-temperature C analyzer, so no methanol was
added to these samples. The tubes were placed on a reciprocating shaker and shaken for
24 h at 23 :1; 2 °C. Loss of primisulfuron from soil-free blank solutions during shaking
and centrifugation was negligible. A preliminary kinetic study showed that steady state
was reached in 24; within experimental error, the concentration of primisulfuron sorbed
after longer reaction times (up to 5 d) was the same as that sorbed after 24 h. The
supernatant solutions were separated from the sorbents by centrifuging for 10 min at
8000 x g.
A l-mL aliquot from each sample supernatant and “blank” solution was added to

5.5 mL of toluene-based scintillation cocktail and counted with a Packard 1500 Tricarb
Liquid Scintillation Analyzer (Packard Inst. Co., Downer's Grove, IL). The pH of the
remaining supernatant solution was measured. The supernatant solution was decanted,
and the tubes were reweighed to determine the mass of solution entrained in the soil
pastes at the bottom of each tube. Samples were saved for desorption experiments. The
DOC concentration in the supernatant solution of samples with no 14C was determined
using a Dohrmann DC- 190 hi gh-temperature TOC analyzer (Rosemount Analytical,
Santa Clara, CA).

To assess the effect of surfactant on primisulfuron sorption by two agricultural soils

at three suspension pH values, primisulfuron sorption by the Hapludalf and Haplaquept A

horizon samples was measured as described above, except that the surfactant Triton X-77

was added to all 0.01 M CaC12 solutions at a rate of 0.25 % (v v").

Batch desorption Isotherms
To determine whether primisulfuron exhibits sorption hysteresis at natural soil
pH, primisulfuron desorption was measured in those samples to which no acid or base
had been added. Ten mL of herbicide-free 0.01 M CaC12 were added to the soil pastes
that remained in the tubes after the sorption experiments, and the suspensiOns were
shaken for 24 h and centrifuged as described above. A l-mL aliquot was removed from

each tube and counted by liquid scintillation counting.

Data analysis
The concentration of sorbed primisulfuron (x.; mg kg") was calculated with the
equation:

Xs = (Cblank - Ce)Ve MI [1]
where Cblank and Ce are the ”equilibrium” primisulfuron concentrations (mg L") in the
soil-free blank and the sample supernatant solutions, respectively, Ve is the volume (L) of
sample solution reacted with the sorbent, and M is the dry mass (kg) of soil used in the
experiment. Sorption isotherms were constructed by plotting x3 vs. Ce. Sorption
isotherms were linear at the primisulfuron concentrations used in these experiments, so
sorption data were fit by linear regression to give the sorption coefficient K. (L kg"),
which is the slope of the regression line.

For the desorption experiments, the concentration of primisulfuron not desorbed
in 24 h (xd) was calculated with the equation:
Xd = [(XsM + CeVem) - CdVd)] M'1 [2]
where Vent is the volume (that is, mass) of solution entrained after the sorption

experiment, Cd is the primisulfuron concentration in the desorption supernatant solution,

and Va is the volume of entrained solution plus the volume of 0.01 M CaClz solution
added during the desorption experiment. Desorption isotherms were constructed by
plotting qd vs. Cd. The linear desorption coefficients (Kdesorp) were determined from the

slope of the desorption isotherms.

RESULTS AND DISCUSSION
Effect of QC on Sorption at pH 4.5

At pH 4.5, 80% of solution-phase primisulfuron is neutral, and primisulfuron
sorption coefficients generally increased as soil OC increased (Table 1). There is a
correlation between Ks and OC, even when the muck soil is omitted and only the mineral
soils are included (P = 0.037; Fig. 2), which suggests that soil organic matter is important
for primisulfuron sorption. However, the non-zero intercepts for the Ks vs OC
correlation lines and the ten-fold increase in the OC-norrnalized sorption coefficient, Koo,
with decreasing OC (Table 1) suggest that primisulfuron sorption is more complex than
predicted by partition theory.

Green and Karickhof f (1990) have suggested that partition theory (Karickhof f et al.,
1979; Karickhoff, 1981; Chiou et al., 1979; 1983) should be used to predict Koo only if
an herbicide's water solubility is less than about 1 mM and a soil’s clay:OC ratio is less
than 40. When the clay:OC ratio is greater than about 40, sorption by minerals should
cause the measured Koo to be greater than Koc calculated from partition theory.
According to these criteria, partition theory should provide a reliable estimate of K00 in
the muck soil but not in the other five soils, where the clay:OC ranges from 39 to 94.

To determine whether partition theory gives reasonable estimates of Koc for any of
the soils at pH 4.5, Koo was calculated with an equation adapted from Chiou et al. (1983):

log K0c = -0.813 (log Sscl + log V) - 0.993 + log 1.74 + log (Co Ca") [3]

10

Table l. Primisulfuron sorption coefficients (Ks) and Koo values at pH 4.5101, and

relevant soil properties (n=3).

 

Soil K,’ K0: oc‘ FeDcB’ Clay Doct
— L kg" — % g kg '1 % mg L-1
Medisaprist I,

(muck) 15.471090 32 47.7 8.9 ND. 150:1:15
Haplauqept 1.56:1:0. 13 93 1.67 5.0 60 85:10
Oxidic Oxisol 0.80:0.04 101 0.79 84.8 74 308:1:08
Hapludalf A 0.551005 83 0.66 1 1.6 26 57:8
Hapludalf B 0671004 279 0.24 32.6 23 26.7106
Siliceous Oxisol 0671005 319 0.21 38.1 42 20:3

 

‘l

interval

* Koo 2 KS (0.01 x % OC)". Predicted K0c is 1325; see text for explanation

‘ SD < 0.03 % 0C for mineral soils; 0.3 % 0C for muck soil

’ SD 5 0.2 g Fe kg"

Not determined. Ash content of muck soil was 18 %

Sorption coefficient K, determined from slope of isotherm :1: 95% confidence

 

 

 

 

 

pH
6.1 5.1 . 4.1
+x_-77 ng x-77 +
Cl 177.06 I
A 0.7%oc A
C]
2..

A
A

 

 

 

0.0 0.2 0.4 0.6 0.8 1.0
Neutral Primi ITotal Primi

Figure 2. Sorption coefficient Ks as a function of soil organic C concentration for the
five mineral soils. Error bars represent 95% confidence limits of the slope of sorption

isotherms and generally are smaller than the size of data points.

12

where Sac] is the water solubility of primisulfuron as a hypothetical supercooled liquid, V
is primisulfuron's molar volume (0.289 L moi"), log 1.74 converts from log Km to log
Koc, and log (Co Cu") adjusts for the fraction of primisulfuron that is neutral at pH 4.5.
The solubility of the hypothetical supercooled liquid may be calculated as (Chiou et al.,
1982)

AH — T
loS=loS+ if“) 4
3 ed 8 ma "rg'r [ 1
where S is the measured solubility of solid primisulfuron, AH; is the enthalpy of fusion,
Tan is the melting temperature (476.1 K), and T is temperature. With the approximation

that AHf (cal) = 13.5 Tm (Chiou et al., 1982) and setting T = 298 K, Eq. [3] simplifies to
log K0c = -0.813 (log S + log V) - 0.00805 ('1‘m - T) - 0.849. [5]

At pH 4.5, Eq. [5] gives K0c equal to 1325, which is 4 to 40 times greater than the
measured Koc values (Table 1). The discrepancy between measured and calculated K0‘: is
greatest for the muck soil, the only soil for which partition theory was expected to be
valid. No other solubility-based equation in the literature gave a better estimate of Koo
than did Eq. [5]. Furthermore, the discrepancy between measured and calculated K0c was
even worse (that is, the calculated K0c is near 49,100) if corrections for the solid-liquid
phase transition, molar volume, and (C0 Cm") were omitted. The less-than-predicted Koo
values at pH 4.5 cannot be caused by primisulfuron sorption by minerals because
additional sorption sites on minerals should cause measured sorption to be greater, not
less, than predicted from partition theory.

Dissolved and colloidal organic C (DOC) can enhance the affinity of large,
insoluble organic compounds for the solution phase and cause sorption to decrease with
increasing DOC. Dissolved OC concentrations measured in supernatant solutions after

the sorption experiment (Table 1) were greatest for the soils with large OC concentrations

13

and small Koo, a result that is consistent with the hypothesis that the decrease in Koc with
increasing soil OC was caused by DOC-primisulfuron interactions. However, even the
largest DOC concentration, 150 mg/L, is only about 7 times Sml for primisulfuron (21 mg
L") and is probably not large enough to affect primisulfuron sorption, based on data
reported by Chiou et al (1986; 1987). The DOC concentrations reported in Table 1 do
not include methanol, which was the solvent for the primisulfuron stock solutions.
Methanol contributed an additional 155 to 450 mg DOC L" and increased with increasing
primisulfuron concentration. The total DOC concentration still was only 8 to 25 times
S.d for primisulfuron.

Another possible explanation for the decrease in K0c with increasing soil 0C is
that the polarity of organic matter in the soils with low OC may “match” the polarity of
primisulfuron, which contains 59% O, N, S, and F by weight. In contrast, the muck
soil’s organic matter is highly humified and possibly less polar than is primisulfuron,
making the muck soil a poor sorbent for primisulfuron. Although further research is
needed to verify this last hypothesis, our results clearly show that partition theory cannot
reliably predict sorption of primisulfuron, even at an acidic pH where most of the
herbicide is nonionic.

Effect of Iron Oxides on Sorption

At pH 4.5, where 20% of solution-phase primisulfuron is anionic, there is a
correlation between Ks and dithionite-extractable Fe (Fem) in soils with less than 1%
OC (Table 1; P < 0.01). In contrast, the relation is not significant when soils with greater
than 1% 0C are included (P = 0.45). Minerals may play an important role in sorption of
polar organic compounds when soil DC is less than about 0.5% or when the clay:OC ratio
is high (Green and Karickhof f , 1990). The data in Table 1 show that primisulfuron
sorption depends on both OC and Fe”; Al oxides may also play a role, as shown
previously for chlorsulfuron (Borggaard and Streibig, 1988; 1989). Our results indicate

that even when pH is held constant and is < pKa, primisulfuron Ks values cannot be

14

predicted from any single soil property. Previous research has shown that Ks values for
chlorsulfuron sorption by humic acid are four or five times greater than for Fe oxides,
even when the anion is the predominant solution-phase sulfonylurea species (Borggaard
and Streibi g, 1988). Thus, at pH 4.5, organic C is probably a more important sorbent
for primisulfuron than are minerals, particularly because 80% of dissolved primisulfuron
is uncharged.

Effect of pH on Sorption

Primisulfuron sorption coefficients decreased with increasing equilibrium
supernatant solution pH values in all of the soils used in these experiments (Table 2).
Other researchers have reported similar pH effects for other sulfonylureas (Mersie and
Foy, 1985; Shea, 1986; Wehtje et al., 1987; Borggaard and Streibig, 1989). As pH
increases, sorption of neutral primisulfuron decreases because the proportion of the
neutral species decreases. In addition, sorption of the anionic form also decreases as pH
increases and surface charge on Fe and Al oxides becomes less positive.

Measured Ks values exhibit a non-linear dependence on the fraction of neutral
primisulfuron (Co Cm") with the steepest slope at low pH (Fig. 3). The nonlinear
dependence of Ks on (C0 Cm") may be caused by strong sorption of the anion on
positively charged oxides at low pH, or by a greater affinity of organic matter for
primisulfuron at low pH, where organic matter has a smaller net negative charge.

Effect of Surfactant on Sorption
When primisulfuron-soil suspensions at pH 4.5 contained the surfactant X-77, Ks for
primisulfuron sorption by the Hapludalf A horizon soil (0.7% OC) was twice that
measured in surfactant-free solutions. The addition of X-77 caused Ks for the Hapludalf
to increase by a constant amount, about 0.25 L kg'l , at all pH values between 4.5 and 5.6
(Table 2; Fig. 3). For the Haplaquept soil (1.7% OC), however, X-77 caused a slight
decrease in Ks at pH 4.1 and had a negligible effect at pH 6.5, the natural soil pH. When

15

Table 2. Primisulfuron sorption coefficients (Ks; L kg") at various pH, with and without

 

 

 

surfactant (X-77).
Surfactant (X-77)
pH Absent Present
Medisaprist (muck)
4.5 15.47 :1; 0.901 ND"
58’ 5.43 :i: 0.34 ND.
Haplaquept
4.1 2.71 :1; 0.17 2.14 :1: 0.06
4.5 1.56 :1: 0.13 N.D.
4.8 1.24 :i: 0.11 N.D.
4.9 1.17 :i: 0.08 0.92 a: 0.03
6.5’ 0.18 :i: 0.03 0.14 1 0.02
Emu—(121W).

45* 0.55 :i: 0.05 0.88 :t: 0.10
5.0 0.29 :1: 0.05 0.56 :1: 0.05
5.6 0.23 :1 0.03 0.46 i 0.07

MW

39‘ 1.62 :i: 0.11 ND.
4.3 0.90 x 004 ND.
4.4 0.67 :i: 0.04 NB.
5.0 0.39 :i: 0.03 N.D.

 

I slope of isotherm :1: 95% confidence interval

* Indicates supernatant pH when no pH adjustment

’ N.D. = not measured

16

 

2
2-0 - K S: 0.41 + 0.62 00 R = 0.81

-1
(L k9 ) -
1.0-

 

 

 

 

0.0 . . . , . . .
0 1 2

Soil Organic Carbon (%)

Figure 3. Effect of pH (top axis), fraction of neutral primisulfuron (bottom axis),
and X-77 surfactant on primisulfuron sorption coefficients for Haplaquept (1.7% DC) and
Hapludalf (0.7% OC) soil samples. Error bars represent 95% confidence limits of the

slope of sorption isotherms.

17

pH was less than the pKa of primisulfuron, the increase in K, for the soil with 0.7 % OC
and the decrease in Ks for the soil with 1.7 % OC can be understood readily in terms of a
model proposed by Sun and Boyd (1993). In this model, addition of surfactant creates
two new phases with which primisulfuron may associate: surfactant that is sorbed to the
soil, and surfactant that is in the water. Sorbed surfactant increases the affinity of an
herbicide for the soil, whereas “aqueous phase” surfactant increases an herbicide’s
affinity for the solution. The net effect of surfactant (that is, increased or decreased
herbicide sorption) depends upon whether sorption enhancement or solubility
enhancement is more important. For a single herbicide-surfactant combination in
different soils, the mass ratio of sorbed surfactant to natural soil organic matter should
determine whether a surfactant enhances or inhibits sorption of the solute, because the
solubility-enhancement effect of the surfactant should be the same for all soils.

In the soil with 0.7 % DC, the dominant effect of surfactant apparently is to increase
the solid-phase OC concentration and thus to enhance primisulfuron sorption. In
contrast, for the soil with 1.7% DC, surfactant sorption likely causes soil CC to increase
by a smaller proportion of total OC, and the dominant effect of surfactant appears to be
smaller K. values because of increased affinity of primisulfuron for the aqueous phase.

Desorption

Desorption isotherms were linear over the entire concentration range (not shown).
Primisulfuron desorption coefficients (Kdesap) at natural soil pH were greater than the
corresponding sorption coefficients (Table 3). In most of the soils, about 50 % of sorbed
primisulfuron did not desorb during a single 24—h batch desorption reaction, but 90% did not
desorb in the soil with the highest pH and lowest K. (Table 3). Irreversible sorption of nonionic
organic compounds has been reported by numerous authors (for example, Pi gnatello, 1989;
Scribner et al., 1992; Brusseau and Rao, 1989) and attributed to a dif f usion-controlled process
rather than to chemical transformation or chemisorption. In the present study, we cannot state

with certainty whether the "irreversibly sorbed" primisulfuron is nonionic or anionic,

18

Table 3. Primisulfuron sorption (K.) and desorption (K dmp) coefficients at natural soil

 

 

pH
Supernatant
Soil 0C pH Ks Kdesorp
% L kg" L kg"
Medisaprist (muck) 47.7 5.8 5.43 :1; 0.341 9.00 1 0.94
Haplaquept 1.67 6.5 0.18 :1: 0.03 0.67 :1: 0.13
Oxidic Oxisol 0.70 4.6 0.80 :1: 0.04 1.00 :1: 0.10
Hapludalf A horizon 0.66 4.5 0.55 :1; 0.05 0.61 :1: 0.09
Hapludalf B horizon 0.24 3.9 1.62 :1: 0.11 2.09 :1: 0.29
Siliceous Oxisol 0.21 4.4 0.67 :1: 0.05 1.17 :1: 0.11

 

I Slope of isotherm :1: 95% confidence interval

19

yet the large fraction of irreversibly sorbed primisulfuron for the pH 6.5 soil may be
caused by strong retention of anionic primisulfuron. Additional research is needed to

determine the effect of soil properties on primisulfuron sorption hysteresis.

CONCLUSIONS

In summary, primisulfuron sorption coefficients increase with increasing OC and
dithionite-extractable Fe, and with decreasing pH. At constant pH, both OC and Fe
oxides are important sorbents in soils with less than 1% DC, but 0C is dominant in soils
with higher OC. Sorption estimates based on log-log relations between water solubility
and Koo overestimate Koo by a factor of 4 to 40, depending on the soil. When pH is
varied, a ten-fold change in [H+] in any one soil has a greater effect on Ks than does a
ten-fold change in OC at constant pH. Sorption of primisulfuron is also affected by the
surfactant, X-77. Addition of X-77 to soil suspensions caused primisulfuron sorption to
increase in a soil with less than 1% OC but to decrease in a soil with greater than 1% OC.
Our results show that Ks cannot be predicted from partition theory, and that accurate
estimation of herbicide sorption and mobility requires that sorption be measured in the
surfactant solutions normally used in weed-control applications, not in surfactant-free

aqueous solutions.

20

REFERENCES

Borggaard, OK, and J.C. Streibi g. 1988. Chlorsulfuron adsorption by humic acid, iron
oxides, and montmorillonite. Weed Sci. 36: 530-534.

Borggaard, OK, and 1.0 Streibi g. 1989. Chlorsulfuron adsorption by selected soil
samples. Acta Agric. Scand. 39:351-360.

Brusseau, M. L., and P. S. C. Rao. 1989. Sorption nonideality during organic
contaminant transport in porous media. CRC Crit. Rev. Env. Contr. 19: 33-99.

Chiou, C. T., R. L. Malcolm, T. I. Brinton, and D. E Kile. 1986. Water solubility
enhancement of some organic pollutants and pesticides by dissolved humic and
f ulvic acids. Environ. Sci. Technol. 20: 502-508.

Chiou, C.T., LJ. Peters, and V.H. Freed. 1979. A physical concept of soil-water
equilibria for nonionic organic compounds. Science 206:831-832.

Chiou, C.T., RE Porter, and D.W. Schmedding. 1983. Partitioning equilibria of nonionic
organic compounds between soil organic matter and water. Environ. Sci. Technol.
17:227-231.

Chiou, C.T., D.W. Schmedding, and M. Manes. 1982. Partitioning of organic
compounds in octanol-water systems. Environ. Sci. Technol. 16:4-10.

Chiou, C. T., D. E. Kile, T. I. Brinton, R. L. Malcolm, J. A. Leenheer, and P. MacCarthy.
1987. A comparison of water solubility enhancements of organic solutes by aquatic
humic materials and commercial humic acids. Environ. Sci. Technol. 21: 1231-
1234.

Gee, G.W. and J.W. Bauder. 1986. Particle size analysis. 111 A. Klute (ed.) Methods of
Soil Analysis Part 1. 2nd ed. Agronomy 91383-411.

Green, R. E., and S. W. Karickhoff. 1990. Sorption estimates for modeling. p. 79-101 Ln_
H.H. Cheng (ed.) Pesticides in the Soil Environment: Processes, Impacts, and
Modeling. Madison, WI, Soil Sci. Soc. of Am. 79-101.

21

Jackson, M.L., C.H. Lim, and L.W. Zelazny. 1986. oxides, hydroxides, and
aluminosilicates. _I_r_r_ A. Klute (ed.) Methods of Soil Analysis. Part 1. 2nd ed. Agron.
9: 101-150.

Karickhoff, SW. 1981. Semi-empirical estimation of sorption of hydrophobic pollutants
on natural sediments and soils. Chemosphere. 10: 833-846.

Karickhoff, S.W., D.S. Brown, and T.A. Scott. 1979. Sorption of hydrophobic pollutants
on natural sediments. Water Res. 13:241-248.

Mersie, W., and CL Foy. 1985. Phytotoxicity and adsorption of chlorsulfuron as
affected by soil properties. Weed Sci. 33:564—568.

Mersie, W., and CL. Foy. 1986. Adsorption, desorption, and mobility of chlorsulfuron
in soils. J. Agric. Food Chem. 34:89-92.

Nelson, D.W., and LE. Sommers. 1982. Total carbon, organic carbon, and organic
matter. kl. AL. Page (ed.) Methods of Soil Analysis. Part 2. 2nd ed. Agron. 9:539-
579.

Nicholls, P. and AA. Evans. 1985. Adsorption and movement in soils of chlorsulfuron
and other weak acids. p.333-339. In Proc. 1985 British Crop Protection Conf.-
Weeds.

Pi gnatello, J.J. 1989. Sorption dynamics of organic compounds in soils and sediments. p.
45-80. In B.L. Sawhney and K. Brown (eds.) Reactions and movement of organic
chemicals in soils. SSSA Spec. Publ. 22. SSSA, Madison, WI.

Scribner, S.L., T.R. Benzing, S. Sun, S.A. Boyd. 1992. Desorption and bioavailability of
aged simazine residues in soil from a continuous corn field. J. Environ. Qual.

21: 1 15-120.
Shea, PJ. 1986. Chlorsulfuron dissociation and adsorption on adsorbents and soil. Weed

Sci. 34:474-478.

22

Sun, S., and SA. Boyd. 1993. Sorption of nonionic organic compounds in soil-water
systems containing petroleum sulfonate-oil surfactants. Environ. Sci. Technol.
27: 1340-1346.

U.S. Environmental Protection Agency. 1990. 1990. EPA Off. Pesticides Toxic Subst.
Pesticide Fact Sheet. 540-FS-90-093. Cincinnati, OH.

Walker, A., BC. Cotterill, and SJ. Welch. 1989. Adsorption and degradation of
chlorsulfuron and metsulfuron-methyl in soils from different depths. Weed Res.
29:281-287.

Wehtje, G., R. Dickens, J.W. Wilcut, and BF. Hajek. 1987. Sorption and mobility of
sulfometuron and imazapyr in f ive Alabama soils. Weed Sci. 35:858-864.

Zahnow, EW. 1982. Analysis of the herbicide chlorsulfuron by liquid chromatography.
J. A gric. Food Chem. 30:854-857.

Chapter 2.
Methanol Cosolvent Effects on TCE and Toluene
Solubility and Sorption

ABSTRACT

Solution composition can affect the environmental fate of hydrophobic organic
compounds (HOCs). Other researchers have proposed log-linear relationships to describe
the effect of cosolvents on HOC solubility and sorption. The solubility and sorption of
trichloroethylene (T CE) and toluene were investigated in low volume fraction (f 9 methanol
solutions. Sorption was measured in two soils with low and medium organic C using
batch reaction. Solubility of TCE and toluene increased log-linearly with increasing
methanol fraction, although data for toluene might be better described with a nonlinear
relationship. Values obtained for the cosolvency power, a, of methanol for TCE and
toluene were 16:02 and 1.4102, respectively. Toluene and TCE sorption coefficients
for a soil with 2 % organic C soil could be fit with a log-linear dependence on methanol fc,
although the plots exhibited some nonlinearity. Deviations from log-linear behavior may be
due to increasing solubility enhancement effects with increasing volume fraction cosolvent.

Methanol fc had no effect on toluene sorption by a soil with 0.7% organic C.

24
INTRODUCTION

Sorption is an important process in determining the fate and transport of
hydrophobic organic compounds (HOCs) in the subsurface environment. These chemicals
may occur in a mixture of solvents (water plus an organic cosolvent). In other cases,
cosolvents may used to facilitate groundwater remediation. Research results from aqueous
systems may not be appropriate in some situations to describe HOC sorption and transport
in solvent-containing environments. '

A number of researchers have studied the effect of cosolvents on HOC solubility
and sorption. Yalkowsky et al. (1972) proposed that the solubility of an HOC in aqueous
(SW) and mixed (Sm) solvents can be related by the equation:

log Sm (moles/L) = log S, (moles/L) +Ofc [1]
where fc is the volume fraction of organic cosolvent and a is the cosolvency power. The
value of o is unique for every solute/solvent combination and is the slope of a plot of log
Sm vs. fc.

Nkedi-Kizza et al. (1985) found that a could be related to a solute’s
hydrocarbonaceous surface area (HSA, A2) by the equation:

a = Ay‘ HSA/k'l‘ [2]
where A7“ is the interfacial free energy (J A") of the solvent at the hydrocarbonaceous
surface area of contact with the solute, k is the Boltzmann constant (1.38 x 10 23 J K"),
and T is the absolute temperature. Woodbum et al. (1986) found Ay‘ was independent of
solute and determined a value of 2.30x10'22 J A" for water/methanol mixtures. Monis et
al. (1988) found that a could be related to a solute’s octanol-water partition coefficient
(K0,) by the equation:

o=a logKW+b [3]
where a and b are regression parameters unique for every solvent (for methanol a = 0.68

and b = 1.07).

25

Most researchers who found a log-linear relationship between Sm and fc used fc
values in the range of 0.1 to 1.0 (Fu and Luthy, 1986a, Yalkowsky et al., 1972,
Yalkowsky et al., 1976). Kimble and Chin (1994) found that the solubilities of three
polynuclear aromatic hydrocarbons deviated from the log-linear relationship in
methanol/water mixtures at low fc. Pinal et a1. (1990) found slight deviations from the
log-linear relationship for trichloroethylene and toluene at f, > 0.3.

Using the fact that sorption of HOCs can be related to solubility, Rao (1985)
expanded Yalkowsky’s theory to relate sorption coefficients in aqueous solution (K) and
mixed solution (Kn):

log K... (ng) = log K. (We) - 0«If. [4]
where or represents the effect of solvent/sorbent interactions on log K“, and o is
independent of sorbent. If 01 is greater than 1.0, a cosolvent causes sorption to decrease
more than expected from the increase in HOC solubility alone. If 01 is equal to 1.0,
solvent-sorbent interactions do not affect the sorption coefficient. If 01 is less than 1.0, the
decrease in HOC sorption with increasing fc is less than predicted based on increased
solubility. Therefore, solvent-sorbent interactions cause greater sorption in cosolvent
mixtures than in water. Most researchers have found a values less than 1.0 in
methanol/water mixtures (Fu and Luthy, 1986, Walters and Guiseppi-Elle, 1988, Wood et
al., 1990, Brusseau et al., 1991, Kimble and Chin, 1994).

Values of 01 less below 1.0 may be caused by swelling of organic matter in
cosolvent solutions, which makes nonpolar sorption regions more accessible to a solute
(Wood et al., 1990). Freeman and Cheung (1981) have shown that the Hildebrand
solubility parameter (0) of organic matter is between 8 and 15, and likely is near 10, at least
for nonpolar sorption regions of organic matter. Swelling of cross-linked polymers
increases as the solubility parameter of the solvent approaches that of the cross-linked
solid. Thus, as the volume fraction methanol (0:15) increases relative to water (0:23) in

a cosolvent mixture, the solubility parameter of the solvent becomes closer to that of

26

organic matter, which facilitates swelling and may expose additional sorption domains
(Wood et al., 1990).

Increased swelling of organic matter may also explain why sorption and/or
desorption rate constants increase as the fraction of cosolvent increases (Walters and
Guiseppi-Elle, 1988, Nkedi-Kizza et al., 1989, Brusseau et al., 1991) because solute
diffusion through swollen organic matter is faster than through the more densely entangled
structure of unswollen organic matter (Freeman and Cheung, 1981). Intraorganic matter
diffusion is believed to be the rate limiting step in obtaining sorption equilibrium (Freeman
and Cheung, 1981; Brusseau and Rao, 1989). Consequently, sorption and desorption are
faster in the presence of cosolvents that promote swelling of nonpolar regions of organic
matter.

The addition of cosolvent can help researchers not only by decreasing time to reach
equilibration, but also can be useful in determining the water solubility (S) or sorption
from water (K) for a highly hydrophobic compound. Results from solubility or sorption
experiments using several different fractions of a cosolvent can be extrapolated to an fc =
0.0, to predict Sw or Kw. A Kbe value extrapolated from dioxin sorption isotherms using
methanol as a cosolvent agreed well with values predicted from S“, (Wood and Guiseppi-
Elle, 1988).

Due to the fact that cosolvents can increase solubility, decrease sorption, and
decrease nonequilibrium sorption constraints, addition of cosolvents to ground water
would increase the amount of HOCs present in the liquid phase relative to the soil at
contaminated sites (Augustijn et al., 1994). This may enhance in situ remediation
techniques such as bioremediation or pump and treat technologies, depending on the
cosolvent fraction and properties.

The purpose of this research was to determine the effect of methanol on the

solubility and sorption of trichloroethlyene (TCE) and toluene by soils with a range of

27

organic C contents and to determine if equations 1-4 hold true for low (fc <020)

concentrations of methanol.

MATERIALS AND METHODS
Materials

The two soils selected for this study were chosen to have a range of organic carbon
contents, because the effect of cosolvents on organic matter swelling should increase with
increasing organic matter content. The first soil used in this study was an A-horizon
sample of Capac loam (fine-loamy, mixed, mesic Aerie Ochraqualfs) that contained 2.1%
organic C. The second soil was an A-horizon sample of Kalamazoo sandy clay loam (fine-
loamy, mixed, mesic Typic Hapludalfs) that contained 0.66% organic C. Soil was air-
dried and passed through a 2-mm sieve. The two solutes used in the experiment were
trichloroethylene (T CE) and toluene. Selected properties of the solutes are shown in Table
l. The cosolvent chosen for the experiments was methanol. All organic chemicals were

HPLC grade.

Table 1. Solute Properties.

 

 

Densityf K0: SJ HSA

Solute (g/mL) (ng) log Km] (mol L") (A’)
toluene 0.866 300 2.73 5.59 x 10'3 127’
TCE 1.465 126 2.38 8.37 x 10'3 ND’

 

I Pavlostathis and Mathavan (1992).
*Woodburn et al. (1986).
’Not determined.

Solubility Enhancement Experiments
The solubility of TCE and toluene was measured in systems composed of 0 to 0.2
volume fraction methanol in aqueous 0.01 M CaCl2 with 0.05% NaN3 added as a biocide.

Duplicate 5-mL aliquots of these mixtures were added to screw-top, 25—mL glass Corex

centrifuge tubes. Liquid TCE or toluene was added dropwise to the tubes until phase
separation was obtained. All tubes were sealed with tefion-lined caps and shaken for 36 h
at 25:1:1°C, then centrifuged at 10,000 rpm for 10 min. A 10-}4L aliquot was removed from
the aqueous phase of each tube, diluted to 5 mL with deionized water, and analyzed by gas
chromatography. Although solubility enhancement experiments did not occur in zero
headspace, this method was considered reliable based on good agreement between values
measured for water solubility (volume fraction methanol = 0, no salt or biocide) and those
previously reported (Table 1).
Batch Sorption Isotherms

Sorption of TCE and toluene from water and from mixed methanol-water cosolvent
solutions was measured in batch sorption experiments. Sorption of TCE and toluene by
Capac soil was measured in solutions that contained 0. 0.04, 0.08, 0.12, and 0.20 volume
fraction methanol in 0.01 M CaCl2 with 0.05% NaN3 to inhibit microbial activity. Sorption
of toluene by Kalamazoo sandy clay loam was measured at 0, 0.04, 0.08, and 0.12 volume
fraction methanol. For each soil and cosolvent composition and each of eight initial TCE
and toluene concentrations, duplicate 5.0:1:0.05-g subsamples of soil (dry weight) were
weighed into 16-mL glass septum vials. The vials were then filled to zero headspace with
the appropriate aqueous or cosolvent solution. Each vial was sealed with a tefion-lined
septum and an aluminum crimp seal. The vials were then weighed to determine the
solution volume. Aliquots (1 to 8 14L) of 10,000 14g mL" TCE or toluene in methanol were
injected into each vial to give the appropriate initial solution concentration. The volume of
methanol added was negligible (<0.075% v v"). For each TCE, toluene, and cosolvent
fraction, eight soil-free blank solutions were prepared with zero headspace and treated
identically to the soil-containing samples. Vials were shaken at room temperature for 24 h,
after which the suspensions were centrifuged at 3,000 rpm for 10 min. A 5-mL aliquot of
the liquid phase was transfened with a syringe to a headspace vial and analyzed for TCE 0r

toluene by gas chromatography.

29

Gas Chromatography

Solute analyses were performed with a Perkin-Elmer gas chromatography
autosystem. Samples were heated to 90 °C for 25 min using an HS40 headspace sampler.
Injector temperature was set at 250 °C. A 30m, 0.53mm I.D., 3.0 pm film thickness,
liquid cyano-methyl-phenylsilicone column was used to provide resolution for TCE and
toluene. The carrier gas was helium and the temperature program was 60-100 °C at 10 °C
per min. Solutes were detected with a flame ionization detector set at 250 °C. Samples
were calibrated against a standard curve from obtained known quantities of solutes
dissolved in the same solvent (volume fraction methanol, salt and biocide) as the samples.

Data Analysis
The concentration of solute sorbed (x', mg kg") was calculated with the equation:
X. = (CuflVm - C.VJ M" [5]
where Cbhnk and CC are the concentrations (mg L") of solute in the “equilibrated” soil-free
blank and sample supernatant solutions, respectively, Vb,“ and V, are the volumes (L) of
soil-free blank and sample solutions, respectively, and M is the dry mass (kg) of soil used
in the experiment. Sorption isotherms were constructed by plotting x, vs. C,. The linear
sorption coefficient (K; L kg") was determined from the slope of the isotherm. Standard
errors for each K value were calculated using the regression data analysis tool in Microsoft
Excel.
RESULTS
Solubility Enhancement Experiments

The solubilities of TCE and toluene increased log-linearly with increasing methanol
fraction (Figs. 1 and 2), although data for toluene might be better described with a
nonlinear relationship. Using Eq. 1, values of o for TCE and toluene were 16:02 and
1.4:1:0.2, respectively. These values are significantly less then the values of 2.7 and 2.9
for TCE and toluene, respectively, predicted from Eq. 3. By contrast, Eq. 2 gives a = 7.1

for toluene, which is almost 2.5 times greater than calculated from Eq. 1.

30

Kimble and Chin (1994) suggested that the log-linear relationship between Sn and
fc could be divided into two regions, low (fc<0.2) and high (fc>0.2) cosolvent
concentrations. This would give each solute two a values. They proposed that o for low
methanol concentrations could be related to the HOC’s aqueous activity coefficient (y) by
the equation:

0 = 1.03 log (y) - 3.56 . [6]
The aqueous activity coefficient calculated from the inverse liquid mole fraction solubilities
gives 7 values of 6640 and 9440 for TCE and toluene, respectively. Using these values,
Eq. 6 predicts a values of 0.38 and 0.56 for TCE and toluene, respectively, significantly
less that what was found in our experiments. Clearly, published equations for calculating
a give values that differ by greater than a factor of 10; the values determined from Eq. 1 in
the present experiments are well within the broad range of predicted values.

One possible explanation for the wide range of 0 values is that methanol solubility-
enhancement effects are nonlinear over the entire concentration range. Banerjee and
Yalkowsky (1988) hypothesized that at low fc, each solute molecule interacts with only one
hydrated cosolvent molecule in low cosolvent/water mixtures. As fc increases, the
cosolvent molecules can associate with one another, significantly increasing solubility
enhancement effects. If this hypothesis is true, then solubility enhancement effects should
increase over the entire range of cosolvent concentrations. This would yield a nonlinear
dependence of log Sm on fc. There need not be a sharp increase in solubility enhancement
at a certain concentration as in a critical micelle concentration for surfactants or as proposed
for cosolvency by Kimble and Chin (1994).

A nonlinear increase in solubility enhancement effect can be seen in our data for
toluene as well as in data from other researchers previously reported as linear (Fu and
Luthy, 1986a, Morris et al., 1988, Pinal et al., 1990). These researchers may have
mistaken nonlinear trends as linear by focusing on medium ranges of cosolvent (fc=02 to

0.8) and ignoring data for low and high fc, where nonlinearity is most pronounced.

31

 

 
   

-‘| .65 1

I

y = 1.6:t0.2x - 1.98
R2 = 0.93

 

log solubility (moles/L)

 

 

 

-2.05 I i .L i P
0 0.04 0.08 0.1 2 0.16 0.2

volume fraction methanol

Figure 1. Log solubility of TCE as a function of volume fraction methanol.

 

4.85
-1.9 --
y = 1.41.0.2x - 2.23
-1-95 " R2 = 0.87

 

log solubility (moles/L)

 

 

 

db

'2.25 I i If JI—
0 0.04 0.08 0.12 0.16 0.2

volume fraction methanol

Figure 2. Log solubility of toluene as a function of volume fraction methanol.

32

Batch Sorption Isotherms

Trichloroethylene and toluene sorption coefficients for by Capac soil at methanol
volume fractions from 0 to 0. 20 are presented in Table 2. Sorption of TCE and toluene
decreased log-linearly with increasing methanol fraction (Fr gs. 3 and 4). The data for TCE
might be better described with a nonlinear relationship, although most published reports
have shown similar nonlinearity but regarded the data as log-linear (Fu and Luthy, 1986b,
Walters and Guiseppi-Elle, 1988, Wood et al., 1990). Using Eq.4 and or values calculated
from corresponding solubility enhancement experiments, 01 values for TCE and toluene
were 0910.3 and 0.7:1:0.2, respectively. Because 01 represents cosolvent/sorbent
interactions and is independent of solute, the value for 01 should be constant. There has
been a wide range of 01 values (0.35-0.89) published for water/methanol mixtures (Fu and
Luthy, 1986b, Pinal et al., 1990, Wood et al., 1990, Kimble and Chin, 1994). Because 01
is calculated from the slope of Km vs. fc (slope = -ao), an incorrect value for 0 would yield
an incorrect 01. Further, if 0 increases as fc increases, then log K"l vs. fc would be
nonlinear, and the calculated value of 01 would depend on the fc range included in the

study.

Table 2. TCE and toluene sorption coefficients (Km) by Capac soil at low

volume fractions (1",) of methanol.

 

 

fc Km (L kg") toluene K£(L kg") TCE
0.00 l.44:i:0.08f 1. 1710.07
0.04 1.29:1:0. 11 1. 121:0.06
0.12 1.17:0.06 0.93:1:0. 12
0.20 0.91:1:0. 15 0.58:0.03

 

I Slope of sorption isotherm :1: standard error.

log Km (L/kg)

33

 

 

 

 

'02 " y = -1.5x + 0.098
Z — s“~
_0.3 .. R _ 0.92
-0.4 i 1 1 ,
0 0.04 0.08 0.1 2 0.1 6 0.2

volume fraction methanol

 

Figure 3. Log TCE sorption coefficient K"l for Capac soil as a function of methanol fc.
(dotted lines represent 95% confidence levels)

log Km (g/mL)

0.24

 

0.2
0.16

.1.-
0.12 -

0.08 -
0.04 -

0 4
-0.04 ~
-0.08 -

I.

 

-0.12

T--
---
~---
~..
u..

y = 09410.1 3x + 0.16
R2 = 0.97

 

 

 

I I

0.04

0.08

0.12

0.2

volume fraction methanol

Figure 4. Log toluene sorption coefficient Km for Capac soil as a function of methanol fc.
(dotted lines represent 95% confidence levels)

34

Toluene sorption coefficients for Kalamazoo soil at methanol volume fractions from
0 to 0.12 are presents in Table 3. Volume fraction of methanol had no significant effect on
sorption. This may be due to the low amount of organic matter in the soil and subsequent
low sorption of toluene by the soil even at fc = 0. Methanol may be sorbed by the organic
matter, increasing the affinity of toluene for the soil. This increase in sorption may

counteract the decrease in sorption due to enhanced water solubility.

Table 3. Toluene sorption coefficients (K...) by Kalamazoo

soil at low volume fractions (ft) of methanol.

 

 

f¢ Km (L kg") toluene
0.00 0.58:1:0. 131
0.04 06210.09
0.08 0.49:1:0.08
0. 12 0.59:0.02

 

I Slope of sorption isotherm :1: standard error.
CON C LUSIONS

In summary, TCE and toluene solubility increased log-lineariy with increasing
methanol f c, and sorption decreased log-linearly with increasing fc, although deviations
from log-linear behavior were noticed. Values obtained for a were within the range of
those calculated from previously published equations. The wide range of reported or values
may be due to an incorrect assumption of a linear relationship between log Sm and fc used in
the equations proposed by other authors. Therefore 0 may not be a constant value, as
solubility enhancement effects may increase as fc increases. This may also explain the
wide variability of or values found in our experiments as well as those reported by other
authors. Volume fraction of methanol had no significant effect on sorption of toluene by a

low organic C soil.

35

REFERENCES

Augustijn, D.C.M., R.E. Jessup, P.S.C. Rao, and A.L. Wood. 1994. Remediation of
contaminated soils by solvent flushing. J. Environ. Eng. 120:42-57.

Banerjee, S. and SH. Yalkowsky. 1988. Co-solvent induced solubilization of
hydrophobic compounds into water. Anal.Chem. 60:2153-2158.

Brusseau, M.L. and P.S.C. Rao. 1989. The influence of sorbate-organic matter
interactions on sorption nonequilibrium. Chemosphere 18: 1691-1706.

Brusseau, M.L., A.L. Wood, and P.S.C. Rao. 1991. Influence of organic cosolvents on
the sorption kinetics of hydrophobic organic chemicals. Environ. Sci. Technol.
25:903-910.

Freeman, DH, and LS. Cheung. 1981. A gel partition model for organic desorption
from a pond sediment. Science 214:790-792.

Fu, J.-K., and R.G. Luthy. 1986a. Aromatic compound solubility in solvent/water
mixtures.J. Environ. Eng. 112:328-345.

Fu, J .-K., and R.G. Luthy. 1986b. Effect of organic solvent on sorption of aromatic
solutes onto soils. J. Environ. Eng 112:346-366.

Karickhof f , SW. 1984. Organic pollutant sorption in aquatic systems. J. Hydraulic Eng.
1 10:707-735.

Kimble, K.D., and Y.-P. Chin. 1994. The sorption of polycyclic aromatic hydrocarbons
by soils in low-methanol/water mixtures. J. Contam. Hydro. 17:129-143.

Morris, K.R., R. Abramowitz, R. Pinal, P. Davis, and SH. Yalkowsky. 1988. Solubility
of aromatic pollutants in mixed solvents. Chemosphere 17:285-298.

Nkedi-Kizza, P., P.S.C. Rao, and A.G. Homsby. 1985. Influence of organic cosolvents
on sorption of hydrophobic organic chemicals by soils. Environ. Sci. Technol.

19:975-979.

36

Nkedi-Kizza, P., M.L. Brusseau, P.S.C. Rao, and A.G. Homsby. 1989. Nonequilibrium
sorption during displacement of hydrophobic organic chemicals and “Ca through soil
columns with aqueous and mixed solvents. Environ. Sci. Technol. 23:814—820.

Pavlostathis SC. and G. Mathavan. 1992. Desorption kinetics of selected volatile organic
compounds from field contaminated soils. Environ. Sci. Technol. 26:532-538.

Pinal, R., P.S.C. Rao, L.S. Lee, P.V. Cline, and SH. Yalkowsky. 1990. Cosolvency of
partially miscible organic solvents on the solubility of hydrophobic organic
chemicals. Environ. Sci. Technol. 24:639-647.

Rao, P.S.C., A.G. Homsby, D.P. Kilcrease, and P. Nkedi-Kizza. 1985. Sorption and
transport of hydrophobic organic chemicals in aqueous and mixed solvent systems:
model development and preliminary evaluation. J. Environ. Qual. 14:376-383.

Walters, R.W., and A. Guiseppi-Elie. 1988. Sorption of 2,3,7,8-tetrachlorodibenzo-p-
dioxin to soils from water/methanol mixtures. Environ. Sci. Technol. 22:819-825.

Wood, A.L., D.C. Bouchard, M.L. Brusseau, and P.S.C. Rao. 1990. Cosolvent effects
on sorption and mobility of organic contaminants in soils. Chemosphere 21:575-587.

Woodbum, K.B., P.S.C. Rao, M. Fukui, and P. Nkedi-Kizza. 1986. Solvophobic
approach for predicting sorption of hydrophobic organic chemicals on synthetic
sorbents and soils. J. Contam. Hydro. 1:227-241.

Yalkowsky, S.I-l., G.L. Flynn, and G.L. Amidon. 1972. Solubility of nonelectrolytes in
polar solvents. J. Pharm. Sci. 61:983-984.

Yalkowsky, S.H., S.C. Valvani, and G.L. Amidon. 1976. Solubility of nonelectroytes in

polar solvents IV: nonpolar drugs in mixed solvents. J. Pharm. Sci. 65: 1488- 1494.

Chapter 3.
Conclusions

Sorption is an important process in determining the fate and transport of organic
compounds in the subsurface environment. It was shown that solution and soil
composition affect sorption of ionizable (primisulfuron) and nonionic (TCE and toluene)
compounds. Sorption of primisulfuron increased with decreasing pH because of increased
fraction of the molecular form which can be sorbed by organic matter. At pH 4.5
primisulfuron sorption increased with increasing 0C. In soils with less than 1% OC,
primisulfuron sorption at pH 4.5 increased with increasing Fe oxide content. The presence
of surfactant increased sorption of primisulfuron in a soil with 0.7% DC, but decreased in a
soil with 1.7% DC. About 50% of sorbed primisulfuron did not desorb during a single 24-
h batch desorption reaction. Sorption estimates based on log-log relations between water
solubility and K“ overestimated K0c a factor of 4 to 40, depending on the soil, even when
K0c was corrected for the ionization of primisulfuron.

Sorption of TCE and toluene by a soil with 2.1% OC decreased log-linearly with
increasing volume fraction methanol, although deviations from log-linear behavior were
noticed. Cosolvency power of methanol was within the broad range of those predicted by
equations proposed by other authors. Without an accurate prediction of solubility
enhancement it is impossible to predict sorption increases with increased volume fraction
methanol using a proposed equation. More research is needed (especially on the molecular
level) in order to better predict the effect of soil and solution properties on sorption of
organic compounds. This first step is necessary in order to determine the fate and transport

of organic compounds in the subsurface environment.

37

 

 

 

.. i
.. 3.0.0131:

:1
tot

.b .21.

j
. . . . . . a 02.816101098660026$116.85“

. . . . . . . . . . . . . . .31Jfinnfind .1ulutt.

. . . . . , . . .. H .. . .. .19.... 1112.1

.49?!
(Hi... 1

a... r}. 4 r

5. . . .
surf}... 1.11.3...
1.. mmw wens-“Bent
0.3%... 9.7

n .019...
1
.Wwe..... is
an. 13 .

v.1 u
. . . .
Nature,
(rut.

.
1......

. Times... «1.: . 4...".
.imgfififiisdw . . 1r. umwvrfw...