D~ICE AND LFAQlABILITY OF RmIOOAL IJ3CP AND Em

FIOf SOILS AND SAPRCLITE

Richard E. Green

Frank L. Peterson

Donna S. Buxton

November 1985

SPECIAL REPCm' 7.1: 85

PREPARED FOR

Hawaii State Legislatureand

Office Qf Environmental. Quality ControlState of Hawaii

WATER RESCXJRCES RESFARCB CEmERUniversity of Hawaii at Manoa

HonolulU, Hawaii 96822

Dr. G.L.Dr. P.C.Dr. R.S.Mr. H.K.Dr. T.W.Dr. R.E.Dr. J.W.Dr. R.C.Dr. F.D.

Mr. J.F.Mr. E.T.Dr. F.L.

SUBSURFACE WATER QUALITY: PESTICIDES CONTAMINATION

PERSCHm.

Dr. L. stephen Lau, Director, Water Resources Research center; andProfessor, Department of Civil Engineering(Project Principal Investigator)

Dugan, Professor and Chainnan, Civil EngineeringEkern, Hydrologist, WRRC; Professor, Agron~ and SoilsFuj ioka, Virologist, WRRCGee, Environmental Engineering Specialist, WRRCGiambelluca, Assistant Hydrologist, WRRCGreen, Professor, Agro~ and SoilsHylin, Professor, Agricultural BiochemistryJones, Associate Professor of SoilsMiller, Associate Researcher, WRRC; Associate Professor,

Public HealthMink, Hydrologist-Geologist; Research Affiliate, WRRCMurabayashi, Land-Use Specialist, WRRCPeterson, Professor and Chainnan, Geology and Geophysics

Federal

U. S. Depa.rtment of the NavyU.S. Geological SUrvey

State of Hawaii

Depa.rtment of Health

City and County of Honolulu

Board of Water SUpply

A-iii

PREFACE

'Ibis rep:>rt is Pirt of the -SUbsurface water Quality: Pesticides Con­

taminationR project authorized in Act 285, Section 38F, by the 'lWelfth Legis­

lature, State of Hawaii, and SUPIX>rted by the Office of &1Vironmental QualityControl with the cooperation of several data collecting agencies. Other proj­

ect activities currently in progress focus on the follOlling topics: geologic

factors, mineralogic parameters, chro:oo1ogy of deep water percolation through

pinea.wle fields, leaching properties of fumigants fran soils, temp:>ral and

spatial distribJtions of contaminants in basal groundwaters, well am cquifer

rehabilitation, and methods of contaminant removal. Forthcaning reIX>rts will

present the results of these activities.

A-v

Pesticides applied to pineaWle fields in central O' ahu to control nema­tode I:X>pulations have been detected in groun<¥ater drawn fran the underlying

aquifer. '!hese nematicides have persisted in surface BlUS and deep saprolite

despite their discontinued use. ibe leachability of these pesticide residues

fran soil and their subsequent movement will deteImine whether or JX>t they

constitute a continued threat to groun<¥ater quality. In this study', leach­

ability is assessed t¥ sorption-desorption measurements and is characterized

t¥ distriwtion coefficients and the kinetics of release of sorbed residues.

Two methods for characterizing sorption-desorption processes have been

developed that exploit the tendency of these fumigant pesticides to v8J:X>rize.

Both methods analyze the val:X>r Iilase of a soil-pesticide system. '!be indirect

sorption method results in a determination of distribution coefficients and

the turge system results in a characterization of the kinetics of desorption.

Experiments using the indirect sorption method to determine distriwtion

coefficients for a surface soil obtained fran central 0' ahu have been ini­

tiated. Preliminary work using the J.X1Cge system is also lIlderway. '!be lerl

sorption-desorption methods will be used to deteImine the leachability of OOCP

and EIB residues in soil and saprolite obtained fran central 0' ahu.

A-vii

PRE:f?Ac:E • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • A-iii

AB~. • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• A-v

• • • • • • • • • • • • • • • • • •

· . . . . . . . . . . . . . . . . .INIR.CDJCl'ION. • • • • • • • • • • •

Sorption-Desorption Processes.

Methodology. • • • • • • • • • · . . . . . . • • • • • • • • •

A-IA-lA-2

ColllDIl Study • • • • • • • • • • • • • • • • • • • • • • • • • • •

selection of Soil samples for Sorption-Desorption Studies • • • •

Measuring DistriJ:ution Coefficients-Background••••••••

Direct SOrption Methcx1 • • • • • • • • • • • • • • • • • • • • • •

Indirect Sorption Methcx1 • • • • • • • • • • • • • • • • • • • • •

A-3

A-3A-4A-4

A-7

A-9A-IIA-I2

. .

. . . . .

. . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . .

Purge System • • •

CBJ~IVES•••••

APPROl\Oi AND ME'!B(J) DE.VELOPMENT • • • • • • • • • • • • • •

A-I2A-I2

A-I3

A-I3

A-I4A-ISA-ISA-I8

• •

. .

. . .

. . .

• •

. . .. . . . . .

. . . . .

. . . . . . . .

. . . . . . . . . .• • •

Purge System • • • • • • • •

Initial Developnent of Vap:>r sampling Techniques • • • •

Using Indirect Sorption Methcx1 • • • • • • •

Solving Mass Balance Relations Using Henry I s Law • • • • •

Results of Henry I S Law Experimental Work • • • • • • • • • •

Modification of AR?lication of Henry's Law •••••

Prel irninary Kd Experirnent. • • • •

Ac:x::x:>MPLISHMENTS • • • • • •

~ .....REFERIK:&S CITED.

. . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . .A-IS

A-20

Figures

1.

2.

SChematic Diagram Oltlining Basic Step; in DirectSOrption Methcx1 (Batch EkIUilibraticn Methcx1). • ••

SChematic 'Diagram Oltlining Basic SteIE in IndirectSOrption .Methcx1 • • • • • • • • • • • • • • • • • •

. . . . . . . .

. . . . . . . .A-6

A-a3 • Diagram of Purge System Used for Collecting Pesticide

ValX>I Purged fran Soil in SOlution. • • • • • • • • • • • • • • • •

4. Linear Plot Shatling Dependence of Kh en Temperatureand Varying Concentrations of mcP in SOlution. • • • • • • • • • •

A-I0

A-16

A-viii

5. Linear Plot Showing Dependence of Kh a1 VaryingConcentrations of lJ3CP in Solution••••••• • • • • • • • • • • A-16

6. Logarithmic Plot Showing Linearity of IJ3CP VaIX>rConcentrations at LQi Solution Concentrations amDeviation fran Linearity at High Solution Concentrations. • • • •• A-17

Tables

1. CClDparison of Kh Values Derived fran Measurements ofOOCP Vcqx>r in Soil SOlution am Tap-water samples 1¥Using Indirect Sorption Method. • • • • • • • • • • • . . . . . . . A-15

2. Data Used to Calculate DBCP Distribution Coefficient <Kd)by Indirect SOrption Method and Calculation Sa;Iuencefor SOil sample • • • • • • • • • • • • • • • • • • • • • . . . . . A-I9

For a number of years various :Pesticides have been awlied to surface

soils of central O· ahu in an attenpt to reduce nematode infestation of pine­

aWle plants. Pesticides have since been detected in grolmdwater drawn fran

the lmderlying aquifer. '!he COI'ClX>unds being detected are classified as fumi­

gant nematicides, dibranochloropropane (DBCP) and ethylene dibranide (Em),

both halogenated hydrocarbons. Although these cantx>W1ds have been successful

in controlling nematode tx>pulations, they have been banned fran use in pine­

awle alltivation because of their carcinogenicity and presence in ground­

water.

Besides detection in groundwater, JECP and Em residues have been

detected in soils and underlying saprolite, despite their discontinued use

on O· ahu some years ago (Dept. of Agriculture 1983). Another contaminant,

trichloropropane (TCP), which originated in the nematicide ro awlied over

30 years ago, has also been detected in grolDldwater samples taken fran several

locations on central O' ahu. Acceptable levels for JECP and Em, set by the

Envirorunental Protection Agency and the state of Hawaii, are in the p:lIts ~r

trillion (ppt) range. Both JECP and TCP have been detected at the p:lIts ~r

billion (wt» level in water and soil scUnples obtained fran the SChofield

Plateau in central O' ahu.

The Olerall objective is to examine the leachability of sorbed :Pesticide

residues and to assess the tx>ssibility that continued groundwater contamina­

tion can result fran the desorption of these residues. Leachability can be

studied through the analysis of soil-:Pesticide systems. ltt>re specifically,

leachability can be assessed by sorption distribJtion coefficients and by

characterizing the kinetics of sorption-desorption processes.

Sorption-Desorption Processes

The extent of contamination of a water resource is related to the leach­

ability of sorbed caDIX>unds (i.e., their ranova! fran the soil by :Percolating

water) • The leachability of a :Pesticide depends on the sorption-desorption

process, a };henomenon that involves two-way exchange of soluble COI'ClX>nents.

Sorption describes the process by which soil particles take up :Pesticide;

desorption is the process by which soil particles release sorbed :Pesticide to

soil solution.

A-2

Partitioning of a ~sticide between soil solution and soil particles at

equilibrium is expressed in terms of a distribution coefficient, Kd. 'Ibis

distribJtion coefficient is defined as the ratio of the concentration of the

sorbed phase to the concentration of the solution PJase for a bYstem at B:Iui­librium. At low solution concentrations (e.g., one-tenth the solubility in

water for IBCP and EIB), sorption Kd values are relatively constant, i.e., the

sorption isotherm is linear. The magnitude of Kd values reveals infonnation

concerning both sorption and desorption processes. For instance, a relativelylarge desorption Kd value indicates a pesticide may be "highly" sorbed or is

not readily or easily leached fran the soil site, whereas a relatively small

value suggests that the pesticide is quite easily leached.

Previously, a single Kd value determined fran sorption experiments was

awlied to both sorption and desorption processes. Kd values of sorption anddesorption have been shown to differ significantly (Green, Liu, and Tamrakar

1986) • 'Ibis implies the i.mJ.x>rtance of determining Kd values for the awro­

priate ~ocesses.

To understand the rates of release of a pesticide from soil particle

surfaces and subsequent movement, it is i.nqx>rtant to dlaracterize the kinetics

of desorption. Previous studies by DiToro and Horzanpa (1982) and Karickhoff

and Morris (1985) have shown that the initially desorbed IX>rtion of the total

amount of ~sticide in a soil system is easily ranoved while the remainder ismore resistant to desorption.

Methodology

Conventional methods used to obtain distribJtion coefficients (Kd) in:­

volve measuring a COJlllX)und in the solution Ihase. However, due to the vola­

tile nature of these nematicides, methodologies were developed that exploit

the tendency of these particular COJlllX)unds to vaIX>rize.

Two different methods were developed that quantify the volatile COJlllX)nent

of IBCP and EIB. The first method, a sinple "headspace analysis" technique,

is referred to as the indirect sorption method. It involves characterizingdesorption processes within a closed soil-water-~sticide system that reaches

equilibrium (given sufficient time) between the sorbed, solution, and vaIX>rphases of the pesticide. '!be second method, referred to as a "purge system"

and which is modeled after the awroach of Karickhoff and Morris (1985),

characterizes desorption in an open system under dynamic conditions. Both

A-3

methods require an analysis of the vaIX>r Iilase and allow the determination of

the rates and the extent of desorption. '!he indirect sorption method is cap­

able of determining sorption distrilxltion coefficients. 1he plrge ~stem

prOV'ides data which can yield kinetic coefficients for the desorption process.

Soil properties, such as organic carbon content, pH, and moisture con­

tent, will also be determined for soil samples. Using the techniques devel­

oped in this study, relationships between soil properties, sorption distrilxl­

tion coefficients, and desorption kinetics will be thoroughly examined. SUch

relationships will enable an assessment of the IX>tential for leaching of pes­

ticide residues and the IX>ssible contamination of groundwater lDlderlying the

study area.

The oojectives of this project are

1. To determine the rate and extent of IBCP and EIB desorption fran soils in

relation to soil properties

2. To determine the relationship between hydraulic properties of soils and

the displacement of residual pesticides t:¥ transient water JOOVement

3 • To assess the IX>tential of groundwater contamination t:¥ residual IBCP and

EIB in soils in central 0' ahu.

The developnent of nethods used to determine sorption coefficients (Kd)

and to characterize the kinetics of release are described. 1he batch BIuili­

bration method is a traditional method used to OOtain Kd values, and involves

analysis of pesticide in the solution P1ase. Two additional techniques have

been developed in this study to characterize sorption-desorption processes

fran measuranents of pesticide in the vaIX>r Iilase in a soil ~stem.

The first technique, the indirect sorption method, relates the vaIX>r

phase to the solution Ihase, the concentration of which is used to calculate

Kd. The second technique developed in this study, the p.trge ~stem, charac­

terizes the kinetics of desorption via measuranents of the amount of pesticide

in the vaIX>r Iilase p.trged fran a soil solution, analyses being conducted at

A-4

varying time intervals.

In this stUdy, the conventional batch 8Iuilibration method has been used

to deteDlline sorption coefficients for CCIllJ?arison with results obtained by the

more versatile indirect sorption method. '!he indirect sorption method has

been awlied to determine the relationship between the vaIX>I phase pesticide

in equilibrium with a pesticide in the solution phase in a solution-vaIX>I

system. Preliminary neasurenents have also been made using a soil-solution­vcqx>r system to obtain sorption coefficients for a surface soil sample by the

indirect sorption method.

Measuring DistribJtion Coefficients-Background

A major effort in this study will be directed tONard measuring distribJ­

tion coefficients, whether for a system at equilibrium or at various stages in

the equilibration process. Kd values can be calculated fran the change in

concentration of a pesticide in a soil solution before and after 8Iuilibra­

tion. By definition,

Kd = Slee (1)

where S is the amolDlt of pesticide sorbed onto soil particles (ng/g) and eeis the concentration of pesticide in the soil solution after equilibration

(nglmR,). Kd is expressed in units of mVg of oven-dry soil.

The amount of pesticide sorbed onto soil particles is given by

S - (Ci - ce)V- M

where Ci is the initial concentration of pesticide in the soil solution (ng/

roR,), ce is as defined above, V is the total volume (roR,) of solution (the

amount added containing pesticide plus that amount existing as soil IOOisture),

and M is the mass of soil expressed as oven-dry weight (9). The methods used

in this project involve either a direct measurenent of Ci and ee, as with the

batch EqUilibration method, or an analysis of the vaIX>r phase which can be

used to determine the concentration of pesticide in the solution phase, aswith the indirect sorption method.

Direct SOrption Method

The batch EqUilibration method is a conventional method that has been

used to determine Kd values in the study of sorption-desorption processes. It

A-S

is also referred to here as the direct sorption method since the pesticide re­

leased to a>lution is determined l:!i a direct analysis of the Equilibrium solu­

tion Ptase. A schematic diagram that outlines the ste~ in the procedure is

shown in Figure 1.

To measure the desorption Kd value, a soil sample is weighed directly

into a screw-cap teflon centrifuge tube into which tap water is also weighed.

The ratio of solution to soil is variable. The tube is agitated for 24 to

72 hr to insure full contact between tap water and sorbed pesticide resicile,

resulting in a well-mixed soil solution. The agitated solution Ptase is then

separated fram soil particles ~ centrifugation.

An aliquot of this supernatant solution Ptase is pipetted into a boiling

flask containing tap water and benzene; the supernatant is added to the flask

under a layer of benzene so that pesticide loss dIe to volatilization is mini­

mized. The benzene solution is left to sit (with periodic agitaticn) to en­courage the p:lrtitioning of the pesticide into the benzene. After one hour,

the ~stem is distilled. The refluxed benzene/pesticide azeotrope is collect­

ed and dried with sodium sulfate. An aliquot of this azeotrope is injected

into a gas chranatograph to determine the ooncentration of the pesticide in

the original soil solution Ba1Ti?le.

The same procedure is followed to determine sorption information except

that the tap water initially added to the soil sample contains a known pest!­

cide ex>ncentration. During the sorption process, the pesticide is taken up ~

the soil ~stem instead of being released to solution.

The direct sorption method has the advantage of being a traditional pro­

Cedure, one that has been tested numerous times, and is fairly simple to ex>n­

duct. However, the method has certain disadvantages. It can take up to five

days to complete depending on the equilibration time and each Ba1Ti?le provides

only one measured value. Pesticide losses can be substantial cile to the

number of times the supernatant solution is transferred to various containers

according to the steps in the procedure. ltDre i.nqx>rtantly, however, the

efficiency of this procedure depends principally on pesticide losses in the

distillation step. NlInerous distillations with both water and soil samples

conducted in this laboratory resulted in 87% (5%) recovery, a value c0m­

parable to others obtained fran the literature. Also, desorption rates cannot

be measured accurately because of the difficulty in interrupting the desorp­

tion process at specific stages of equilibration; only the extent of desorp-

A-6

Tap Water

Equilibration by Agitation (24-72 hr)

ICentrifugation

Aliquot of Supernatant into Boiling Flask(solution released under benzene layer)

BenzeneLayer

Dis til 1a t i on

IRefluxed Benzene/Pesticide Collected

and Dried with Na2S04

IAnalysis by Gas Chromatography

Figure 1. SChematic diagram outlining resic step:; in directsorption methcx1 (retch equilibration methoo)

(3)

A-7

tion at any given time can be measured.

These disadvantages encouraged the developnent of new methodologies-the

indirect sorption method and the p.1rge system-to inprove the quantification

of pesticide distribution in various Ptases within a soil system.

Indirect Sorption Method

The indirect sorption method was developed for the specific puqx>se of

inferring solution Ptase concentration fran measured pesticide in the vap>r

phase. A generalized schematic diagram is shown in Figure 2 that outlines the

basic steps to the procedure.

An amount of soil containing sorbed pesticide residues is weighed direct­

ly into a 16Q-m£ 8yIX>vial, a glass bottle with a very narrow neck that is

closed (stopper fashion) by a tight-fitting teflon "mininert" valve that can

be accessed by a b)1ringe needle. Tap water is then added by weight until the

total volume of solution is 60 rot, leaving a total headspace volume of 100 mR,.

The system is inmediately closed with a mininert valve.

The oontents of the HyJ;x>vial are agitated on a gyrotory shaker for 24 to

72 hr after which time a vap>r sample (maximum volume 0.5 mR,) is drawn. 'Ibis

vap>r sample is injected directly into a gas chranatograph and its ooncentra­

tion determined.

Two equations are applied in the indirect sorption method that result in

a determination of the amolD'lts of pesticide coexisting in the sorbed, the

solution, and the vap>r Ptases. The first equation makes use of sinple mass

balance relations. The second equation is empirically derived and states the

relationship between the vap>r concentration of a particular pesticide and its

concentration in the solution Ptase as determined for a solution-vcqx>r &ystem

only (i.e., no roil particles present). The application of these equations is

described more fully in the Acconplishments section of this rep>rt although a

brief description is given here.

Mass balance in a soil-pesticide system is defined as

Total amount of pesticide =sorbed Ptase + solution Ptase + vap>r Ptase.

The total arnotmt of pesticide is determined by codistillation of a p>rtion of

the soil sample with benzene. The amount in the vap>r };ilase is measured fran

another p>rtion of the same soil sample contained in a closed HyJ;ovial. The

A-8

Tap Water

Placed Directly into Hypovial

System Closed (vapor tight) by Mininert Valve

Equilibration by Agitation (24-72 hr)

Syringe

Hypovial

Vapor

Solution

Sample Withdrawn by Vapor Syringe

I

Vapor Sample Analyzed byInjection into Gas Chromatograph

Figure 2. SChematic diagram outlining basicsteps in indirect sorption method

A-9

concentration of the solution Ptase is calculated fran the previously deter­

mined eI1i>irical EqUation that relates the concentrations of the vaIX>r and

solution Ptases of the pnticular pesticide system. The only Wlknown term

remaining in the mass balance equation, the amolDlt present in the sorbed

phase, is then calculated. Kd is determined I:¥ dividing the sorbed Ptase con­

centration I:¥ the solution Ptase concentration.

As with the direct sorption method, the indirect sorption method has

advantages and disadvantages. ibe disadvantages include having to experi­

mentally derive the relationship between the vaIX>r and solution Ptases of a

particular pesticide (solution-vaIX>r) system. Another disadvantage lies in

the propagation of errors in the calculations reqUired to obtain a Kd value

fran a measurEment of the vaIX>r Ptase. Advantages of this method include

reduction of pesticide loss under vaIX>r-tight conditions, and minimal handling

of the soil sample once the procedure is underway. Also, a single sample is

able to provide many data IX>ints; the examination of nultiple measurements

taken at various time intervals enables the rate and the extent of sorption

and desorption to be determined.

Ad:litionally, va.IX>r samples result in a clean, distinct peak on gas chrer

matograms. The snall number of canponents in a vaIX>r sample decreases the

chance of column contamination occurring during gas chranatographic analyses.

It is rx>t unusual for samples obtained fran soil or solution extractions to

contain canponents derived fran the extracted medium that cause contamination

of a gas chromatogra};tlic column or erroneous peaks on a gas chranatogram.

Peaks other than those of concern make interpretation of gas chranatograms

more difficult than need be.

Purge System

A second method, the p.1rge system, was developed to measure the instan­

taneous release of sorbed pesticide (Ner time. 'Ihis approach has been used

by others to study the desorption of highly hydrophobic organic canpounds

(Karickhoff and Morris 1985; Garbarini and Lion 1985; and Oliver 1985).

Figure 3 is a schematic diagram of the !Urge system developed for our studies.

Filtered laboratory air forced through a fritted glass disc results in a

continuous supply of fine air tutbles that stream through a soil solution

contained within the glass p.1rge chamber. '!his ex>nstant supply of uncontami­

nated air heightens the effect of concentration gradients between the sorbed,

A-IO

Air Filter

Teflon Tubing

Glass Purge Chamberwith Fritted Disc

Tenax Cartridges in Series(vapor collection)

r

.......------.. } Aluminum Collar

Teflon Gasket

Soil solution

[tt)}!l! Va por

rCheckValve

Flow Meter

LaboratoryAir Source

1

NeedleValve

PressureGage

Figure 3. Diagram of :purge system used for collectingpesticide vatx>r :purged from soil in solution

A-II

solution, and vap>r Plases (i. e., the continuous flushing of pesticide-laden

vap>r out of the system causes pesticide in the solution Plase to diffuse/

disperse out which in turn encourages the diffusion of sorbed pesticide fran

soil sorption sites into the solution Iilase>. A series of Tenax cartridges

<consisting of the same components as in a gas chranatograpuc column­

diatanaceous earth and silica gel) capture the pesticide in the vaIX>r Ptase asit is forced out of the p,lrge system.

After several minutes to hours of p,lrging, the cartridges are removed

fran the system. Their contents are anptied into aliquots of benzene which

are then analyzed by gas chranatography. This analysis reveals the amount of

pesticide p.1rged fran the soil system in the vap>r Plase for a given time

interval. The soil system is assumed to be completely purged of leachable

sorbed residues when IX) pesticides are detected on the Tenax cartridges. With

this method as well as with the indirect sorption method, a anall number of

comp:>nents present in the analyzed vap>r sample results in a clean gas chro­

matogram.

By removing and replacing spent cartridges at different time intervals

throughout the p,lrge, rates of desorption as well as the extent can be deter­

mined. A plot showing amount of pesticide released per time interval versus

el~ed time will provide &ta required for analysis of desorption kinetics.

IIat1 readily and at what rates sorbed pesticide residues are released fran soil

sites is directly related to the leachability of a particular compound. We

are particularly interested in determining why pesticide residles in surface

soils resist leaching and volatilization CNer long periods of time (years).

Colunm study'

A fourth method, not yet developed, will serve to correlate sorption

information obtained fran the methods outlined above with pesticide IOOVement

through a soil layer as a flUlction of flow velocity and soil properties. This

study will be ex>nducted on recently applied pesticides as well as long-term

sorbed residues already present in field sanples. Attenpts will be made to

a~re undisturbed field-structured columns (surface soils) once laboratory

packed columns have been analyzed. This p>rtion of the project will provide

an i.mI;x>rtant link to future IOOdeling efforts.

A-12

Selection of SOil samples for SOrption-Desorption studies

Soil and saprolite samples will be taken fran pineapple fields where ED3

and IBCP have been applied in the plst. High priority will be given to areas

associated with the ongoing deep drilling project conducted by the OHM Water

Resources Research Center. ibe three canpolDlds of interest (Em, IECP, TCP)

have been detected in various locations and at various depths in soil cores

taken fran central 0' ahu. Soil samples containing the highest concentrations

of these pesticides will be analyzed in this study.It has been stated that the leachability of a compound can be character­

ized by desorption distritution coefficients and kinetic measurements of the

release of sorbed pesticides. All of the above methods accomplish these goals

by providing infonnation about the plrtitioning of pesticide into each of the

three };hases (sorbed, solution, and vapor) present in a soil system.

Several nonths were conmitted to refining vapor sampling and injection

techniques for use in gas chranatographic analyses. l)Jring this Ittase, sev­

eral difficulties were encolDltered that had to be worked through before the

studies could continue.

Initial Developnent of Vapor sampling Techniques

It was found that the vapor };hase sample of a soil solution had to con­

sist solely of vapor for oonsistent gas chranatographic results to be 00­

tained, i.e., no moisture could be present in the syringe used to inject the

vapor sample into the gas chranatograph (OC). 'Ibis necessitated construction

of an apparatus to prevent a>ndensation fran occurring on the mininert valves

used to close the 8yIx>vials. Ultraviolet light and a cooling fan were added

to the apparatus used for agitating the sanples to prevent the occurrence of

1x>th condensation and excessive heating of sanples in the 8yIx>vials dlring theequilibration period.

'!he technique of removing a vapor sample fran the 8yIx>via! also required

refinement. It was fOlD1d that the barrel of the glass syringe used for 00­

taining a vapor sample became coated with pesticide roolecules so that a p>r­

tion of the sample was sorbed within the syringe after its contents had been

A-13

expelled, thus reducing the measured vap:>r ooncentration. '!his problan was

overcome t¥ fully exposing the &yringe barrel to pesticide vaIXlr before sample

analysis. '!his was oone t¥ flushing the syringe with three volumes of valX>r

prior to injection of a fi.n:U. sample into the GC.

using Indirect Sorption Method

Initial experiments with the indirect sorption method involved the E:qui­

libration of ~eous IJ3CP solutions of varying concentrations with the result­

ing vap:>r Plase. '!he original intention was to aWly the concepts of Henry's

law to derive a oonstant equilibrium value calculated as the ratio of the

vap:>r Plase concentration to the rolution Plase concentration. '!he resulting

ratio, a theoretical constant as predicted t¥ Henry's lat/, could then beawlied to solving mass balance relations (eq. [3]).

However, it was fOWld that for IJ3CP the Henry's law ratio is not constant

over a range of solution concentrations, theret¥ invalidating its use in the

indirect sorption method. As a result, the aWlication of the Henry's law

ratio was modified into the aWlication of an empirically derived equation

that expresses the relationship between a canpound' s vaIXlr and solution Ihases

as measured after the equilibration of a solution-vap:>r system. But since

much of the initial work with the indirect sorption method was completed with

the assumption of a constant vaIXlr concentration to solution concentrationratio, this topic will be discussed briefly.

Solving Mass Balance Relations Using Beru:y' s Law

Henry's law states basically that the ratio of a solute's vaIXlr pressure

to its solubility in aqueous solution, lDlder temperature ex>ntrolled condi­

tions, is a>nstant. For this study, the Henry's law constant was represented

t¥ the ratio of the vaIXlr Ihase concentration to the solution Plase concentra­tion at equilibrium within a solution-vap:>r system. '!his ratio (Kh) was cal­culated as follows:

K- Cv·Vvh-es. Vs-Cv. Vv (4)

where Cv and es are respectively the ex>ncentrations of the vaIXlr and the solu­

tion J;ilases (ng/mR,), and VV and VB are respectively the volumes (mR,) of vap:>r

(heads:Pace) and solution. Here, Cs is the concentration of solution when it

A-14

is placed in the 8ylXwial before Equilibration with the vcqx>r Ptase.The PJIlX>se of defining the Kh ratio was to provide terms necessary for

solving the mass balance Equation that accolBlts for all three ];i1ases coexist­

ing in the system at ~ibrium (eq. [3]). The total amount of pesticide in

the system was determined by refluxing a soil sample with benzene and water

followed by co-distillation of benzene and the pesticide. Equation (4) was

rearranged to solve for the solution concentration given a predetermined Kh

value and the vap:>r ex>ncentration as measured by the indirect sorption method

and GC analysis. Application of the total, solution, and vaIX>r ];i1ase concen­

trations in equation (3) left only the sorbed phase as an unknown. This

"sorbed phase value" divided by the concentration of solution at equilibrium

yielded the value of the distrihltion coefficient, Kd (eq. [1]).

Results of Henry's Law Experimental Work

The application of Henry's liM to indirect sorption measurements was

evaluated using OOCP. First, the behavior of the vaIX>r Ptase was fOlUld to be

unaffected by the presence of organic solutes in the solution Ptase. second,

Kh was found to be dependent on the concentration of pesticide in solution and

on temperature. 'lhi.rd, it was fOlDld that a relationship between the concen­

trations of the vaIX>r Ptase and the solution ];i1ase of a solution-vap:>r system

at equilibrium could be expressed by a mathenatical e.quation derived by inter­

IX>lation of a logarithmic plot of the solution and vaIX>r concentration values.

These findings are discussed more fully belOtl.

Initial experiments using the indirect sorption method involved sepn-ate

equilibrations of OOCP-spiked soil solutions and OOCP-spiked tap water samples

to see if soluble organic CXJIlX>nents in soil solutions significantly affected

the concentration of IBCP in the vaIX>r Ptase at equilibrium. Table 1 coorpa.res

Kh values derived fran soil solution· and tap water samples for two different

concentrations of IBCP. No statistically significant difference occurred

between Kh values obtained fran IBCP-spiked tap water solutions or fran IJ3CP­

spiked soil solutions. This result shows that soluble organic material does

not significantly affect the behavior of sorbed OOCP as it is released to the

vap:>r Ptase. 'lhi.s conclusion enabled vap:>r studies to be conducted with tap

water solutions instead of soil solutions which saved considerable time in

experimental procedures.The vaIX>r pressure of a volatile OOfillX>und is a temperature dependent

A-IS

TABLE 1. CDMPARISCIl OF HmRY'S INl CONSTANT (Kh) VALUES DERIVEDFRCJtf~ OF DBCP VAroR IN roIL OOLurION ANDTAP-WATER SAMPLES BY USm:; INDIREC1' SORPl'ION ME'lH(])

IBCP MEASURED KhSCLurION

<nnN.lRATION SOil Tap(ng/mR,) SOlution Water

2,348 0.00575 0.00571

628 0.00963 0.00969

property (MacKay et ale 1982). ibis behavior was examined experimentally for

IBCP by ~librating and analyzing solutions in a laboratory that was not

temperature controlled. Figure 4 shows the results of duplicate samples

analyzed at different times (and temperatures) during a single day. Kh values

are plotted against the corresp:>nding solution ex>ncentrations. The tempera­

ture of the room was ex>olest (average of 25.9°C> in the zoorning when the first

set of samples was analyzed and warmest (average of 29.5°C) in the afterJ'X)()n

when the second set was analyzed. The differences in Kh values for a single

solution ex>ncentration are statistically significant. ibis conclusion

demonstrated the i.mIX>rtance of conducting va.{X>r studies under temperature­

controlled conditions.

Figure 4 also shows a dependence of Kh on solution concentration. The

value of Kh decreases significantly as solution concentration increases. To

remove the effects of fluctuating temperature, an experiment identical to that

which resulted in Figure 4 was conducted Wlder temperature controlled condi­

tions. The same trend in variation of Kh with solution concentration wasrevealed as shown in Figure 5.

Modification of Awlication of Henry's Law

Nonconstancy of the Kh ratio might awear to invalidate the use of an

equilibrium relationship between the vaIX>r and solution Iilases to determine

sorption by the indirect sorption method. Yet closer inspection of the data

shows that the E.qUilibrium relationship between a solution Iilase and its vaIX>rPlase can be applied with confidence.

A logarithmic plot of IBCP vaIX>r concentrations versus corresp:>nding

aqueous solution concentrations is sham in Figure 6. The graIil shCMs that

A-16

O.010.-------.-----....---....---~....--- .......----..---~

0.009

0.008

0.007

0.006

--0- 29.5·C

- .... - 25.9·C

... -... ----... --- ..... -... .....---~-... ..... ..........-- .....-..........-..... ..... -.. ........

...-... _- .... -.......... ----- .... -------e

700060001000 2000 3000 .000 5000

SOLUTION CONCENTRATION (ng/mt)

NJI'E: Data is for duplicate samples in a single day.

0.005 a..----__---I- --.A- '"""--__----Il-.-__---I.. ...£-__----..

o

Figure 4. Linear plot showing dependence of Kh on teJI1?eratureand varying concentrations of JECP in solution

10 I I , ,

•9_

8 • -

...... 7 •~

~I

'-' 60

x 5 -......,.s:::. 4~ -~

3-

2~

10

I I I

1000 2000 3000 .000 SOOO 6000SOLUTION CONCENTRATION (ng/ml)

Figure 5. Linear plot showing dependence of Khon varying concentrations of JECP insolution

A-17

t

l

00

I

LLJ .....-..C~

- Eu ..........- Ol~ c:CJ") .......,

UJ0..>

ULL.o ~ 1.0

LLJZ V>0<3::- :c~o...

~a::~OZo...LLJ«U>ZoU

10

z

0.110 100 1000 10000

CONCENTRATION OF DBCP IN SOLUTION PHASE, Cs (ng/m£)

Figure 6. Logarithmic plot showing linearity of DBCPvapor concentrations at low solution con­centrations and deviation fran linearityat high solution concentrations

vapor concentration deviates fran linearity at high solution concentrations

but is approximately log-linear at low solution concentrations. This trend is

consistent with theories that describe the nonideal behavior of the vapor

phase as the solubility limit of a COJnIX>W1d is approached. A linear inter­

polation of the data plotted in Figure 6 results in an equation relating Cs

and Cv as

Cv=aesb (5)

where Cv and Cs are respectively the concentrations of pesticide in the vapor

and solution I,ilases in (ng/mR,) i for DBCP a is 0.0258 and b is 0.653.

The fact that the soils of the Schofield Plateau contain relatiVely low

levels of total pesticide (tens of wb> places the concentration of pesticide

in the solution I,ilase at the dilute solution end and in the linear region of

the gral,Xl. Thus, for analyzing soils that contain residual pesticide concen­

trations within the range CNer which this equation aWlies, the aWlication of

the enpirical relationship Ceq. [5]) is justified.

A-18

Preliminary Kd Experiment

A single experiment was conducted using the indirect sorption method to

determine desorption Kd values for a surface soil obtained fran Ik>le Field

4201 located near Mililani Town in central 0' ahu. 'lWo samples fran a depth of

170 to 360 Dm were analyzed in dlplicate.Soil and tap water in a ratio of 2:1 were EqUilibrated for 72 hr in HyJ;x>­

vials after which time vaIX>r measuranents were taken. '!he soil samples as

obtained fran the field contained a total of about 59 wb of long-teDtl sorbed

pesticide residues. Table 2 shows the data that were used in calaIlating Kdfran a vaIX>r measuranent and an application of EqUations (5), (3), and (1)

(in that order). '!he resulting Kd values are less than expected canpared to

values previously determined in this laboratory by the direct sorption method

(unplblished data). we plan to conduct more solution-vapor equilibration

studies to confirm the reprocl1cibility of the a and b terms in equation (5).

The indirect sorption method will be run concurrently with the direct sorption

method to assess the applicability of the indirect sorption method.

Purge System

Duplicate setups of the plrge &ystem have recently been canpleted and

checked to be vapor tight. However, no soil solutions have yet been tested,

nor has the efficiency of the a};Paratus been evaluated. '!herefore, no defini­

tive results are available to date.

The plrge &ystem will be initially tested using aqueous solutions to

detennine the efficiency of the apparatus. Mass balance relations will be

applied in which the total anount of pesticide in the initial solution will be

canpa.red to the amolDlt in the final solution plus that amount of pesticide

captured on the Tenax cartridges after a period of plrging.

Developnent of the indirect sorption method and the plrge &ystem is can­plete and much of the preliminary work has been refined. Experiments are

being conducted using the indirect sorption method that estimates the pesti­

cide concentration in solution fran a vapor concentration measurement, allow­

ing determination of sorption Kd values.

A-19

TABLE 2. DATA USED 'lO CALaJIATE IBCP DISlRIBurION OOEFFICIENT <Kd) BYINDlRECr SORPl'I~ Mm'H(» AND CALOJIATION SEOuao: FUR OOIL SAMPLE

~ ~ S ~

sample Measured calculated calculated calculated----Cng/mR,) Cng/g) cmVg)

1

2

0.2077

0.2048

24.39

23.87

5.00

6.74

0.205

0.282

Calculation Sequence for Soil Sample 1:

Step 1 Ceq. [5]): Cv = 0.0258 es CO•• 51 )

or

(0 .2077J(1/0.8 5 I )Cs = ~.02S8, = 24.39

Step 2 (eq. [3]): Total = sorbed phase + solution J.:hase + vap:>r J.:hase

or

sorbed phase = [(total) (g soil) - (Cs) (mR, solution)

- Cv (mt vap:>r)] / (g soil)

= (59 • 22.78 - 24.39 • 49.58

- 0.2077 • 100)/22.78= 5.00

Step 3(eq. [1]): Kd = (2t~~) =0.205

The application of these techniques and the infonnation they are designed

to provide in terms of distribution coefficients and kinetics of desorption

will be used to ex>rrelate desorption processes with soil properties (particu­

larly organic carbon). Investigations of pesticide behavior in soil systems

aid in defining the leachability of a sorbed pesticide and the related poten­

tial for grounCMater contamination.

A-20

a intercept of logarithnic plot of OJ and Cs determined experimentally

b slope of logarithmic plot of OJ and Cs determined experimentally

ce equilibrilDll (final concentration (ng/mRJ of pesticide in mlution !base

Ci initial concentration (ng/mRJ of pesticide in solution Plase beforeequilibration with vaIX>r Ptase

Cs concentration (ng/mR,) of pesticide in solution Ptase

Cv concentration (ng/mR,) of pesticide in vaIX>r Ptase

Kd distriootion coefficient; defines p:lrtitioning of a pesticide betweensorbed and solution Ptases of a soil system

I<h ratio of OJ in equilibrium with Cs (based on Beru:y's law)

M mass (g) oven-dry soil

S concentration (ng/g) of pesticide in sorbed phase (sorbed onto soilparticles)

V volume (mR,) of noisture in soil system (preexisting soil lOOisture plusadditional solution ex>ntaining pesticide)

Vv volume (mR,) of vaIX>r Ptase (headspace)

Vs volume (mRJ of solution Ptase

Department of Agriculture. 1983. Preliminary reIX>rt on soil sampling for Emon Oahu. Pesticides Branch, Division of Plant IndIstry, state of Hawaii,HonolulU, Hawaii, 19 5eptenber.

DiToro, D.M., and Horzenpa, L.M. 1982. Reversible and resistant componentsof Pa3 adsorption-desorption: Isotherms. Environ. SCi. Technol. 16 (9):594-602.

Garbarini, D.R., and Lion, L.W. 1985. E.Valuation of sorptive partitioning ofnonionic IX>llutants in closed systems t¥ heacEpace analysis. Note sub­mitted to Environ. SCi. Technol. 19(11) :1122-1128.

Green, R. E.; Liu, C.C.K.; and Tarnrakar, N. 1986. "Modeling pesticide IOO\1e­

ment in the unsaturated zone of Hawaii soils under agricultural use."(To be pJblished in Evaluation of Pesticides in Groundwater, Am. Clem.Soc. SyIr(x>sium series).

Karickhoff, S.W., and Morris, K.R. 1985.pollutants in sediment susJ:ensions. II

Athens, GeOrgia (to be p1blished).

A-21

·Sorption dynamics of 1¥drophobicEnviron. Res. Lab., u.s. EPA,

MacKay, D.; Shiu, W.Y.; Bobra, A.; Billington, J.; Chau, E.; Yetm, A.; Ng, C.;and Szeto, F. 1982. Volatilization of organic pollutants fran water.Project SUrrmary EPA-600/S3-82-019, Environ. Res. Lab., u.S. EPA, Athens,Georgia, August.

Oliver, B.G. 1985. Desorption of chlorinated hydrocarbons fran spiked andanthropogenically contaminated sediments. Chemosphere 14 (8) :1087-1106 •