United States
                Environmental Protection
                Agency
 Risk Reduction
 Engineering Laboratory
 Cincinnati, OH 45268
                Research and Development
 EPA/600/SR-92/170 October 1992
EPA       Project  Summary
                Soil  Vapor  Extraction Column
                Experiments on  Gasoline
                Contaminated Soil
               Michael E. Miller, Tom A. Pederson, Carole A. Kaslick and George E. Hoag
                 Soil vapor extraction (SVE) is a tech-
               nique that is used to  remove volatile
               organic compounds from unsaturated
               soils. Air is pumped through and from
               the contaminated zone to remove vapor
               phase constituents. In  this work, labo-
               ratory soil column experiments were
               conducted using a sandy soil residu-
               ally saturated with gasoline to evaluate
               the performance of SVE under con-
               trolled conditions. Both vapor extrac-
               tion and aqueous leaching of the soil
               columns were conducted. The progress
               of the vapor extraction event was con-
               tinuously monitored by an in-line total
               hydrocarbon analyzer (THA). Perfor-
               mance of vapor extraction was evalu-
               ated  by a series of  soil  chemical
               analyses including total petroleum hy-
               drocarbons, headspace measurements,
               and extraction techniques with quanti-
               fication by GC/FID and GC/MS. Con-
               taminant levels in aqueous percolate
               were  compared before and after SVE.
               After 60 pore volumes of water flow
               through a column, the percolate from
               the contaminated soil still contained at
               least  100 mg/L of total hydrocarbons.
               Vapor extraction of contaminated soil
               reduced total hydrocarbons by 99.96%,
               and subsequent aqueous leaching re-
               sulted in percolate concentrations of
               3.7 mg/L initially and 0.6 mg/L after 60
               pore volumes of water flow.
                 This Project Summary was developed
               by EPA's Risk Reduction Engineering
               Laboratory, Cincinnati, OH, to announce
               key findings of the  research project
               that is fully documented in a separate
               report of the same title (see Project
               Report ordering information at back).
 Introduction
   SVE is an innovative technology used
 to remove volatile compounds from un-
 saturated soils. The basic principles of
 SVE technology are straightforward.  By
 inducing air flow from  and through the
 subsurface vadose zone, vapor phase
 contaminants are flushed from  the soil
 pores. Extraction wells are the usual route
 by which contaminants are recovered from
 the subsurface.
   The process consists of the following
 steps. Contaminant laden air is withdrawn
 through extraction wells under a vacuum
 created by an above-ground  blower or
 vacuum pump. The air flow is controlled
 by ball or butterfly valves and is monitored
 by vacuum gauges. The extracted vapor
 stream passes through an air/water sepa-
 rator to remove moisture and protect the
 blower. After the blower, the air stream
 passes through a heat exchanger to con-
 trol the relative humidity and improve the
 efficiency of the subsequent vapor treat-
 ment operation. Injection wells are optional
 as a means to enhance air flow. A cap or
 surface seal placed over the treatment
 area to control the vapor flow path is also
 an optional component of the system. The
 seal,  which could be as simple as plastic
 sheeting, serves to induce the air to flow
 in  a horizontal manner as opposed to a
vertical flow pafhway that may result from
air being drawn from the surface only near
the extraction well.
  Vapor extraction is most effective in re-
moving compounds that exhibit significant
volatility at ambient temperatures in soil
(i.e., vapor pressure greater than 0.5 mm
of  mercury at 20°C and  a dimensionless


          Gjj/0 Printed on Recycled Paper

-------
Henry's Law constant greater than 0.01).
This includes most gasoline constituents
and solvents. Additionally, the mobility of
the compound due to its affinity  for the
soil organic matter and small pore  spaces
affects  the  success of the treatment.
Compounds strongly  sorbed to the soil
matrix tend to be more difficult to extract.
  Although SVE has been used with great
success at numerous sites, the behavior
of the contaminants remaining in the soil
has  not been  completely assessed. The
efficacy of SVE as a remedial technology
for extraction  of  gasoline  contaminated
soils  was investigated  using laboratory
scale soil column experiments. One  di-
mensional air  flow rates were controlled
within a small soil sample with  distinct
boundaries. Other parameters controlled
Included moisture content, chemical con-
taminants, and soil characteristics.
  The  experimental program had three
main objectives:
  *  Determine the limitations on gasoline
     removal from soil using vapor extrac-
    tion under controlled, optimized labo-
     ratory conditions.
  •  Evaluate the aqueous mobility of the
    constituents in contaminated soil and
    those that remain in the soil following
     SVE.
  •  Assess the appropriateness of avail-
     able analytical techniques to measure
     son contaminant levels.

 Experimental Methods
   Stainless steel soil columns were 29.8
 cm (1 ft) in length with a 10.8 cm (4.25 in.)
 internal diameter. A rigid,  porous Teflon
 disk sat at the bottom of the column and
 supported the soil sample; a second  po-
 rous disk rested on top of the soil. Stain-
 less steel endplates were bolted to either
 end of the columns, and Won*  gaskets
 were used to create an air- and water-tight
 seal. The columns terminated at each end
 in needle valves, which could be closed to
 isolate the system. The internal volume of
 each column was 2.5 L
   The columns were weighed to the near-
 est gram following each manipulation. Four
 identical columns were packed with Con-
 necticut sand; 97% of those particles were
 between 0.1 mm and 5 mm in diameter,
 with an average diameter of 0.5 mm. The
 sand was predominantly quartz (35%) and
 feldspar (35%) and had a surface area of
 1.9 m^g. As  the soil was added to the
 columns in 2.5 cm lifts, the columns were
 vibrated to achieve maximum compaction.
 The process of soil addition and  shaking
 were repeated until the columns were full
(approximately 4.7 kg of soil). The in-place
bulk density was 1.9 g/cm3, and the po-
rosity was 30% as  determined  by  mass
difference (columns empty, packed with
soil, and saturated with water).
  The packed columns were  saturated
with water containing  100 mg  KCN/L  to
prevent microbial activity, drained to field
capacity moisture content, saturated with
gasoline, and allowed to drain to residual
saturation. At various stages of the ex-
periment, a column  was opened, the soil
was poured into a  stainless steel  bowl,
transferred to sample vials and  jars, and
analyzed by the battery of techniques listed
in Table 1. One column was sacrificed for
soil analysis after residual saturation with
gasoline.
  For headspace analysis, 1 g of soil was
spiked  with ~ 500  ppm  of""a~,a,~d-~
trifluorotoluene in a 10 mL Teflon-lined,
septum-sealed vial  and was heated in a
water bath at 90°C for 20 min. This treat-
ment was assumed  to displace all volatile
compounds in the  sample  into the
headspace.  A 200-u.L sample of the
headspace was injected directly into a GC/
FID to obtain a value for headspace vola-
tiles.
  Chromatographic  peaks eluting between
2-methylpentane and 1,2,4-trimethylben-
zene, defined as the gasoline  range or-
ganic (GRO) compounds,  were summed.
The  area of the internal standard was
subtracted from this total  area.  The sub-
sequent sample peak area was directly
correlated to a five  point calibration curve
utilizing the peak area summation of a 10-
compound   mixture  containing   2-
methylpentane; 2,2,4-trimethylpentane;
heptane; benzene; toluene; ethylbenzene;
m-xylene; p-xylene; o-xylene; and  1,2,4-
trimethylbenzene. Varying measured
              amounts of this mixture were added to
              clean soil, and 1 g portions were analyzed
              as above to create the calibration curve
              for soil GRO concentration.
                 The final two columns were vapor ex-
              tracted  simultaneously at a flow  rate of
              2.1 L/min (4.5 scfh). Air, supplied by com-
              pressed air cylinders, passed through an
              activated carbon column to remove hydro-
              carbons, and then through a flask of water
              via a diffuser stone. The  humidified air
              split into two equal  streams and flowed
              first through flow  meters,  then the  soil
              columns. The column exhaust air flowed
              to a sequential sampler. The vapors from
              one column, then the other, were sent
              every  minute to a  THA for continuous
              monitoring of total hydrocarbon levels.
                 After 6.5 days of vapor extraction,  one
              column  was sacrificed  for soil analysis;
              the other  received  60  pore volumes of
              rainwater flow as above. The soil from this
              final column was analyzed at the  conclu-
              sion of the experiment.

               Results and Discussion
                 Measurements are recorded  for a soil
              column residually saturated with gasoline,
              a column following aqueous leaching, and
              a vapor extracted column. Multiple values
              for any measurement are replicate sample
               analyses.
                 The  gasoline retained in each  column
              was measured directly by mass difference.
              The mass of each column at field capacity
               moisture content was subtracted from the
               mass of the same column residually satu-
               rated with gasoline.  Any additional water
              that was lost during the drainage of the
               excess gasoline was collected in a gradu-
               ated cylinder and was also subtracted to
               yield the final value. Initial gasoline  con-
               centrations ranged from 12,700 to 15,300
 Table 1. Soil Analytical Techniques-
 Measurement
                           USEPA SW-846
                           Method No. or Ret
                          Procedure
 • MwMfon of trad* names or commercial products does
  not constitute endorsement or recommendation for
  use.
 Total volatile solids
 Total petroleum
 hydrocarbons

 Gasoline range
 organics
 Headspace

 Semi-volatile organic
 compounds

 Volatile organic
 compounds	
160.4


9071/418.1


Enseco-RMAL, 1990



Method described below

8270


8260
Heat to 10&C
then 550°C

Freon extraction,
IP

Methanol extraction,
GC/FID or GC/MS,
integration of peaks

GC/FID

Methylene chloride
extraction, GC/MS

Methanol extraction,
GC/MS

-------
rag/kg. These values were greater than
the results of each of the analytical meth-
ods employed. The discrepancy can be
attributed to volatilization losses during the
transfer of soil samples from the columns
to vials and jars for storage before analy-
sis as well as to losses incurred during
the performance of the analytical methods
themselves.
  Aqueous leaching of a soil column re-
sidually saturated with gasoline  resulted
in the mobilization of some of the retained
GRO during  60 pore volumes  (about 43.2
L) of water flow. In the percolating water
during the leaching experiment, the peak
concentration of GRO was 3,000 mg/L at
the onset of  leaching. The concentrations
continually dropped, and after about 5 pore
volumes, they began to level off; however
by the end of the run, they were still as
high as 170 mg/L.
   Integration of the GRO  concentrations
during the aqueous leaching  experiment
yielded a total of 9.8  g of GRO removed.
The original  mass of gasoline in the  col-
umn was 79 g, so that about  12%  of the
contaminants were mobilized  in the per-
colating water. Although benzene dropped
below the detection limit of 1.25 mg/L by
60 pore volumes, the levels  of  toluene,
ethylbenzene, and xylene remained steady
at 70, 5, and 30 mg/L, respectively.
   Vapor  extraction was conducted con-
tinuously for 6.5 days, with only brief
breaks in the flow to make adjustments to
the system.  Vapor extraction of two con-
taminated columns brought the soil GRO
levels down as low as 0.5 mg/kg, a re-
moval of 99.96% of the original GRO.
  During the initial 7 min of vapor extrac-
tion, exhaust gas hydrocarbon concentra-
tions exceeded the  instrument's THA
maximum limit of detection of 100,000 ppm
(v/v). After this period, hydrocarbon levels
steadily declined with  brief  periods of in-
creased concentration spikes following
pauses in the flow. The vapor concentra-
tions exiting the two columns were similar
at first during the rapid decline and in the
final stages as the concentrations leveled
off. Hydrocarbon values, however, differed
by  as much as 50% during  the middle
stages of the  extraction process. These
differences  can be explained by different
hydrocarbon starting concentrations, since
the column  giving rise  to the higher levels
had retained 88 g of gasoline before SVE,
whereas the other began with only 66 g.
Final vapor hydrocarbon levels were be-
tween 20 and 40 ppm  (v/v).
  Integration of the vapor extraction  data
yielded a total hydrocarbon mass  removal
that was greater than the  initial gasoline
mass by about 70%. This  is understand-
able since the THA was operated under
circumstances that were  best suited for
generation  of  qualitative data only.  The
THA was calibrated only once and for one
concentration range at the start  of gas
flow, whereas several concentration ranges
were used  during the  course of  the ex-
periment.  Furthermore, the  calibration
mixture contained only butanes and pen-
tane; these were  not  completely repre-
sentative of the complex gasoline mixture.
The longer chain hydrocarbons and aro-
matic compounds also found in gasoline
exhibit greater THA response factors than
the calibration gas compounds. Concen-
trations fell from a maximum of 3.7 mg/L
at 2 pore volumes to a plateau of 0.6 mg/
L after 60 pore volumes.

Conclusions
  The gasoline contaminated sandy soil
was susceptible to  aqueous leaching  of
the hydrocarbon components.  Alter 60
pore volumes of water flow, however, the
majority of product  remained in the soil
and significant concentrations  of BTEX
were still found in the percolating water.
  SVE was effective in removing gasoline
from the soil, with a reduction in the GRO
of 3.5 orders of magnitude. After this pe-
riod, the vapor extracted soil still contained
low levels of water-mobile contaminants.
GRO was the most reproducible of the
analytical   techniques  tested,  and
headspace analysis by GC/FID recorded
the highest concentrations.
  Soil column experiments provided a well
controlled, effective  measurement of the
soil processes occurring  in the aqueous
leaching and vapor extraction of gasoline
contaminated soil.
  The full report was submitted in fulfill-
ment of Contract No. 68-03-3409 by COM
Federal Programs Corporation under the
sponsorship of the U.S.  Environmental
Protection Agency.
                                                                                    •U.S. Government Printing Office: 1992— 648-080/60108

-------
Michael £ Miller, Tom A. Pederson, and Carole A. Kaslick are with Camp Dresser
  andMcKee, Inc., Cambridge, MA 02141. George E Hoag is with University of
  Connecticut, Starrs, CT 06268.
 Chl-Yuan Fan is the EPA Project Officer (see below).
The complete report, entitled "Soil Vapor Extraction Column Experiments on
  Gasoline Contaminated Soil," (Order No. PB92-226430/AS; Cost: $19.00,
  subject to change)  will be available only from:
        National Technical Information Sen/be
        6285 Port Royal Road
        Springfield,VA22161
        Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
        Risk Reduction Engineering Laboratory
        U.S. Environmental Protection Agency
        Edison, NJ 08837
  United States
  Environmental Protection Agency
  Center for Environmental Research Information
  Cincinnati, OH 45268

  Official Business
  Penalty for Private Use
  $300

  EPA/600/SR-92/170
     BULK RATE
POSTAGE & FEES PAID
         EPA
   PERMIT No. G-35

-------