protocol,organic-freewaterismixedwitha soil sampleand heated
to 40 °C.
An
inert
gas is bubbled through the water,
and
the
concentration of chemicals
in
the gas is measured with a
gas
chromatograph. This methodis effective only
if
VOCs
in
soils
rapidly desorb from the soil surface into surrounding water.
Laboratorycontrolsamplesusing
sea
sandin
place
ofsoilsindicate
that. for nonadsorptive solids, purge-and-trap recoveryis accept–
able. Recent sttldies2-
4
.8.1W on the physical inaccessibilityof
contamination in soils suggest that this
may
not
be
the
case ,
particularlyinlong-contaminatedsoils. It
has
beenhypothesized
that
aging
involves diffusionintosoil
micrcpcres.
partitioninginto
soilorganicmatter, strong surfaceadsorption, or a combination
of these
processeS.2.14.l1i.l7
Previousstudies6.18indicate that 20–
90%
of contamination may
be
located
in
the interior of the soil
matrix
andthusinaccessible forpurge-and-trapmeasurement
The
purposeofthepresent
study
wastocomparethe effectiveness of
three conunonly used techniques for measuringVOCs
in
soils:
purge-and-trap,methanol immersion,andhotmethanolextraction.
We found
that
purge-and-trap consistentlyunderestimated the
concentrationofVOCsinagedsoils
by
factors
ranging
from2 to
- 100. This consistent underestimation of soil concentrations
undermines the EPA's attempt to remediate contaminated soils
to levelsconsistent withhealth-based cleanupstandards.
The soildesorption
process
is
knownto involvetwo distinct
phases: arapiddesorption from the
soil
surfaceoccurring
within
24 h and a much slower diffusion-limited desorption from the
interior of the soil matrix
occurring'
over a period of days to
years.5.19 This
biphasic
desorptionpattern ismost pronounced
in
aged soils.
where
a
significant
fraction of contamination
is
located
in
the interiorofthe soilmatrix. For example, despite
its
highvolatilityanddegradability,I,2-dibromoetbane
(EnB),
asoil
fumigant,
was
foundS
in
agriculturaltopsoil19
years after
its last
known application. The persistence of
EDB
was attributed
to
desorption
~times
of2-3decades at 25°C.2 For trichloroet–
hylene
acE),
a continuousdesorption studylOof long-eontarni–
nated soils revealed persistence of 18% of the
initial
TeE
concentrationafterdesorption
with
24000porevolumesofwater.
In
a subsequent study3 on simultaneous desorption of
TeE,
tetrachloroethylene, toluene, and xylene, a substantial portion
(48-94%)
of
the
sorbed
contaminant
mass
resisted
desorptionafter
7
days ofcontacttime.
TeE
soilconcentrations
at
the Picatinny
(I4) Hatzinger. P.
B.: AIelcand~.
M.ElIlIi"",.
Sri.
TtdlaoJ.
1995, 29.
537-~5_
(15)
PignateJIo. ]' J~
fernndino.
f.]~
Huang.
1.
Q.
bllin>a.
Xi.
Ted<1UIi.
1993,
27, 1563- 1571.
(16)Brosseau.M. 1.;Jessup.
R.
Eo:
Rao.P. S. C.
EIIWrn.
Sci.
Ttdl,",l. 1991,
25. 134-142.
Ilil
wu.
S. C.:Gschwend. P. M.£,nn"",-.
Sci.
Ttdr~.
1986, 20. 117-725.
(18)
Smith.].
A.;
Chiou.C.T.;Krammer,J.
A.:
KiJe,
D.
E.
ElIlIi"",_
Sci.
Ttduull.
1990,24, 676-683.
(9) MKkay. D. M.: CherTy.
J.
A.
E.lIiroll.
Sri.
TtdrllllL
1989, 23. 63(\-636.
(ZO)
Di Toro. D.M.; Honempa.
1.
M.
E"Wtm_
Sci.
TfdrIUJI
1982, 16, 594-
sea
Gm Ccates. J. T.:
EIzerman.
A.
W.
J.
C4oIld",i".
HytMJ.
1986, 1. 191- 210.
(22)
PignatelJo.].
J.:
HlWIg.
1.
Q.J.
E"lIiroll.
QIuI.
1991,20, 222- 228.
(23)
Scribner.
5.1.:
&nz:ing.
T.
R.:
Sun.
5.:
Boyd.
5.
A.j.
b!lIi"",.
Otud.
1992,
21.115-120_
(24) Test
~stor~ting
solid.W$e.
EPAl6OQ.SW~. 3rd~.
(2';) HPWin.. A.D. Cold
Ref(ions
Research llIId
Engineering
W>omory. Speeial
RepcnNo. 93-5. May1993.
Arsenalwere found to be 1- 3orders ofmagnitude greater than
predicted using soil- gas concentrations and equilibrium condi–
ticns." The present studyfoundthat thepurge-and-trapmethod.
ascomparedtohotsolventextraction. recovered only42and4.8%,
respectively. of TeE in
lung-contaminated
clays and silty
loam
soils.
Even
in
freshly
spikedsoils,desorption rates ofpollutants
can
be
1-3 orders ofmagnitude smallerthanequilibrium-predicted
rates.' Cleansoils
spiked
withhalogenatedaliphatic hydrocarbons
for 24-72
h
resisted desorptionafter16extractionsof 24-72
h
each.
6
These observationsbringintoquestionthe occurrenceof
desorption equilibriumnecessary for
validity
of the purge-and–
trapmeasurements
in
the freshlyspikedand aged soils.
Previousanalysisof
EDB
in
long-contaminated soils
has
shown
that purge-and-trap is less effective
than
extractionat 75°Cwith
organic solvents such as methanol, acetonitrile, and acetone,2.6J
recoveringless than 11% of the total
EDB
found by hot solvent
extraction. 8 Our
purpose
in
thepresentsttldy
was
toextendthese
results to a larger
class
of VOCs in aged field samples. Hot
methanol extraction proved to
be
moreeffective than the
EPA–
approved
purge-and-trap technique. For
three
long-contaminated
soilscontainingsuchVOCsas trichloroethylene,benzene, toluene,
chloroform, methylene chloride, and cis-l,l·dichloroethylene.
recovery
frompurge-and-trap
ranged
from 1.5up to 41% that of
hotmethanol extraction.
Slow desorption is recognized as a serious obstacle to soil
remediation technologies.
lW 2tl-Zl
Forsuchtechnologiesaspump–
and-treat, vaporexlraCtion, andbioremediation
to
be
effective,
soil
cont3mina:ntsmust
be
accessible. Tosimulate the effectofvapor
extractionon the efficiency of the pw-ge-and-trap methodology,
wesubjectedthe
Louisiana soil
toa
week
of
air
stripping. Purge–
and-trap recovered only
58%
of the methylene chloride, 4.1% of
the
cir
U-dichioroetbyJene, and5.6%ofthe
TCE
that hotmethanol
extraction
was
able to
recover.
These
results
indicate that the
purge-and--trap method
is
not
a
reliablemethod for
evaluating
vaporextraction as a remediation technology.
It
is clear from the
results
of this and previous studies that
the best overall choice for measurement of
soil
VOCs is hot
methanol extraction, since
this
methodyields
a
more accurate
analysis, regardless of the age ofcontaminated soil. The VOC
data from
three
different soil
types
clearly
demonstrate the
limitations oftheEPA-approvedpurge-and-trapmethod,which
can
bias
analytical
resuJts
byseveralordersofmagnitude, depending
on soil
type
and chemical properties. We
suggest
that the
EPA
review
theuseofpurge-and-trapasamethod formeasuringVOCs
in
soils.
ACKNOWLEDGMENT
Oak
Ridge National Laboratory is managed by
Lockheed
Martin
Energy
Research
Corp. forthe
U.S.
Departmentof
Energy
under Contract
DE-AC05-960R22464.
Received for review January 3, 1996.
Accepted
June
17.
1996. e
AC960009C
• Abstm:1
pu.bJish~
in
AdIlQ~u
ACSAbslouctJ.
Allgust 1, 1996.
Analytical Chemistry, Vol.
68.
No,
19,
October
1. 1996 3433
