Air sparging performance in a NAPL-contaminated, sandy-gravel aquifer

Abstract

A controlled field study was designed and conducted to assess the performance of air sparging for remediation of petroleum fuel and solvent contamination in a shallow groundwater aquifer. Sparging was performed in an isolation test cell (5 x 3 x 8 m deep) constructed from interlocking sheet pile. A soil vapour extraction (SVE) system was installed within the cell perimeter for collecting the sparge vapours. Gas samples were collected during sparging from the SVE offgas and analysed for specific target compounds, total hydrocarbons, oxygen, and carbon dioxide. Air sparging was not very effective at removing the contamination because the vapour pressures of the predominant contaminants were low (

Full text

Groundwater Quality: Remediation and Protection (Proceedings of Ihe GQ'98 Conference held at 1
1
Q
Tubingen, Germany, September 1998). IAHS Pub!, no. 250, 1998.
Air sparging performance in a NAPL-
contarainated, sandy-gravel aquifer
JOHN S. GIERKE
Department of
Geological
Engineering and
Sciences,
Michigan Technological University,
1400
Townsend
Drive, Houghton, Michigan 49931-1295, USA
CHRISTOPHER L. WOJICK
Department of
Civil
and Environmental
Engineering,
Michigan Technological University,
1400 Townsend Drive, Houghton, Michigan 49931-1295, USA
JENNIFER M. MURASKI-SMITH
Montgomery Watson, 2100
Corporate
Drive, Addison, Illinois 60101, USA
NEIL J. HUTZLER
College of Engineering, Michigan Technological University, 1400
Townsend
Drive,
Houghton,
Michigan 49931-1295, USA
Abstract A controlled field study was designed and conducted to assess the
performance of air sparging for remediation of petroleum fuel and solvent
contamination in a shallow groundwater aquifer. Sparging was performed in
an isolation test cell (5 x 3 x 8 m deep) constructed from interlocking sheet
pile.
A soil vapour extraction (SVE) system was installed within the cell
perimeter for collecting the sparge vapours. Gas samples were collected
during sparging from the SVE offgas and analysed for specific target
compounds, total hydrocarbons, oxygen, and carbon dioxide. Air sparging
was not very effective at removing the contamination because the vapour
pressures of the predominant contaminants were low (<5 mm Hg).
Biodégradation was a significant removal mechanism during sparging, as
indicated in the high concentrations of C02 in the soil gas. Air sparging was
more effective than water flushing at removing contaminant mass.
INTRODUCTION
Air sparging is a technique used to remove volatile pollutants from contaminated
aquifers (Bausmith et al., 1996; Nyer & Suthersan, 1993). The sparging process
consists of injecting air within or below contamination present in the saturated zone.
Volatile pollutants partition into the air channels created by sparging and are carried
by advection and diffusion/dispersion into the overlying unsaturated zone (Clayton et
al., 1996). The lateral distribution of air channels depends on the injection rate and
pressure, depth of the injection, capillary pressures and permeabilities of the aquifer
materials, location of heterogeneities (due to variations in soil properties and the
presence of non-aqueous phase liquid contaminants), and degree of anisotropy
(McCray & Falta, 1996; Lundegard & LaBrecque, 1995).
Typically, a soil vapour extraction (SVE) system, which directs gas flow in the
unsaturated zone via blower-induced vacuums applied to vents, is installed to capture
the vapours emanating from the sparge zone. SVE is a well-established technique

120 John S. Gierke et al.
with some appropriate design guidance (Gierke & Powers, 1997). Air sparging is
commonly used, but design guidance is limited because of the difficulties in
monitoring treatment performance (Nyer & Suthersan, 1993; Lundegard &
LaBrecque, 1995). Most monitoring efforts have focused on delineating the treatment
zone in terms of the distribution of air flow around sparge wells. The sparge zone is
only one consideration, as treatment may also occur outside the region where air is
flowing and volatilization rates within the sparge zone may vary because of the
heterogeneous nature of sparge air flow (Hein et al, 1997; Clayton et al., 1996).
FIELD TEST DESCRIPTION
An evaluation of the treatment performance of an air sparging system of typical
configuration was conducted in a controlled field setting in Chemical Disposal Pit 2
(CDP2) of Operable Unit 1 (OUI) at Hill Air Force Base (HAFB), Utah, USA. This
evaluation was performed in parallel to performance assessments of seven other
innovative technologies (Brusseau et al., 1998; Montgomery Watson Consultants,
1996).
Like the other technology assessments, the air sparging field test was
conducted in an isolation cell (Cell 1), which was located inside the former chemical
disposal pit. The soil beneath the pit was coarse, ranging from a gravelly sand to a
sandy gravel, underlain by a thick (>70 m) clayey silt unit, which acted as a very
low-permeability confining layer for a deeper sand aquifer.
The pit was used by HAFB until 1974 for the disposal of waste fluids from
aircraft maintenance at the base (Montgomery Watson Consultants, 1996). Primarily,
waste fuels and solvents were dumped at the site. Most of the petroleum-based
compounds are only slightly soluble in water and are less dense than water, so the
non-aqueous phase liquid (NAPL) contamination resides in a smear zone above the
fluctuating water table and some floats on the water table as free product. Floating
NAPL was not observed in any of the wells in Cell 1. Many of the contaminants in
Cell 1 were compounds with vapour pressures near or below 1 mm Hg. Removal by
volatilization was, therefore, impractical for most of the contamination (Bausmith et
al., 1996).
The test cell was approximately 3 m wide, 5 m long and 8 m deep, and was
isolated from the surrounding soil by driving interlocking sheet pile (Fig. 1) to a
depth of more than 2 m into the top of the clayey unit, which was present at an
average depth of 7.3 m below the ground surface. After installation, all sheet pile
joints were grouted to hydraulically isolate the cell. A flexible membrane liner was
installed across the ground surface to prevent short circuiting of surface air into the
SVE system and to enhance capture of the sparge vapours. The groundwater table in
the cell was maintained at a depth of 4.6 m below the ground surface (2.7 m above
the clay unit).
The AS/SVE system utilized six SVE vents and two sparge wells (Fig. 1). The
SVE vents were located near the cell walls, positioned to surround the centrally
located sparge wells. During operation of the AS/SVE system, air was injected
below the water table through the sparge wells, while soil gas was drawn through the
SVE vents screened within the vadose zone and directed to a common manifold
before treatment by granular activated carbon. Measurements of contaminant

Air
sparging
performance in a
NAPL-contaminated,
sandy-gravel aquifer 121
removal were made from the SVE offgas sampled at this common manifold.
The performance assessment consisted of pre- and post-treatment characterization
activities and gas monitoring during treatment. The pre-/post-treatment
characterization was consistent with the other demonstrations and included: soil
coring with subsamples analysed for target compounds, groundwater sampling and
analysis of target compounds, and partitioning tracer tests. The pre-/post-
characterization is reported elsewhere (Wojick, 1998; Gierke et al, 1998). The focus
of this paper is on the performance assessment during sparging operation.
During the treatment phase, soil gas and SVE offgas were monitored for 13
target compounds, total volatile organic chemicals, oxygen, and carbon dioxide.
Contamination was present in both the vadose and saturated zones. Because the
sparge air must pass through the vadose zone before removal by the soil vapour
extraction system, contaminants present in the SVE offgas may originate from both
the vadose and saturated zones. To evaluate air sparging for remediating the
saturated zone, the SVE system was operated initially without sparging in an attempt
to remove as much of the contamination from the vadose zone as practical. After
which, the coupled AS/SVE system was operated.
Operation of the SVE system commenced on 28 August 1996 at an initial flow
rate of 6.8 m3 h"1, which was then incrementally increased to 34 m3 h"1 before the
start of combined AS/SVE operation. Sparging was started on 20 September 1996,
initially at a flow rate of 14 m3 h"1 and then incrementally increased to 37 m3 h"1,
which was sustained until the system was shutdown on 22 October. A total of
45 000 m3 of gas (at STP) was flushed through the test cell. The SVE flow rate
during sparging was approximately
25 %
higher than the sparging rate to prevent the
escape of sparging vapours from the test cell.
The SVE offgas was sampled at the combined-well manifold (Fig. 1) through a
stainless-steel tubing. The tubing was directly connected to a Hewlett-Packard Model
!.!••!
0 1 2 m
§ Air Sparge Well Location
3 SVE Vent Location
• Pre-Treatment Core Location
> Post-Treatment Core Locatior
iS-Soil Gas Monitoring Location
- Outline of Sheetpile Wall
"SVE Manifold
- Air Sparging Piping
*-Gas Flow Direction
Offgas - f- -
Sampling
SVE Offgas-<-Ç.";
Fig. 1 Plan view of the AS/SVE configuration and the soil core and soil gas
monitoring locations in Test Cell 1 of Operable Unit 1 at Hill AFB, Utah, USA.

122 John S. Gierke et al.
5890 gas chromatograph (GC) equipped with a flame ionization detector (FID). A
vacuum pump was used to draw the gas sample from the SVE manifold. This GC
was configured to measure either total VOC concentration or the concentrations of
the target compounds. The SVE offgas was also sampled at the combined SVE
manifold with a syringe, which was used to transfer the sample to a 0.5
1
Tedlar bag.
These samples were analysed for 02 and C02 on a Microsensor Technology
Instruments GC with separate column/detector modules for 02 and C02. The
methods for gas analyses of contaminants are given in Gierke et al. (1998) and
Wojick (1998). The methods for analysing 02 and C02 are reported in Muraski
(1997) and Wojick (1998).
RESULTS
The SVE offgas samples were analysed to determine concentrations of total volatile
organic chemicals (TVOCs) and 13 target compounds. The offgas concentrations of
the 13 constituent compounds were measured at least daily. Since the TVOC
concentrations were a combination of many compounds and the FID responds
differently to each of them, the TVOC concentration changes were affected not only
by removal rates but also by changes in the offgas composition. Therefore the trends
in the individual target constituents are more appropriate for comparing removal by
volatilization to removal by other mechanisms.
Sparging was started on 20 September and caused a gradual increase in the
concentrations of, in order of decreasing amount: undecane, 1,2-dichlorobenzene,
decane, and m-xylene. The changes in concentrations of these compounds followed
the trend in the undecane concentration depicted in Fig. 2. The gradual increase in
target concentrations correspond to incremental increases in the sparging rate over a
period of four days (20-24 September) and continued for another four days (24-28
September) after the maximum sparging rate was achieved. The lateral extent of the
0.010
Offgas
Concentration
(dimensionless)
0 005
)
0.001 -
"e©
MA ®0®
©
' I'll' " I" ""I
e
©
28-Aug 4-Sep 11-Sep 18-Sep 25-Sep 2-Oct 9-Oct 16-Oct 23-Oct
Date
Fig. 2 Dimensionless concentrations (measured concentrations/pure vapour
concentration) of undecane in SVE offgas collected during SVE (28 August-
19 September) and AS/SVE (20 September-22 October).

Air
sparging
performance in a
NAPL-contaminated,
sandy-gravel aquifer 123
sparging zone increases as a result of increasing the sparging rate and this usually
occurs in a few hours in high-permeability materials (Lundegard & LaBrecque,
1995).
Yet the offgas concentration response was much longer than would be
expected based on the redistribution of the sparging zone. The concentration trend is
due to a combination of alterations in the sparge zone, imperfect contact between the
sparge air and the contamination, volatilization of contaminants in the unsaturated
zone,
and, possibly, partitioning of contaminants in the sparge vapours into NAPL
contamination remaining in the unsaturated zone. The more volatile constituents, i.e.
compounds with a vapour pressure >5 mm Hg, were not observed in detectable
concentrations in the offgas during sparging nor were they found at very high
concentrations in the soil and groundwater samples (Wojick, 1998).
The test cell was designed and operated so that all of the volatilized contaminants
could be collected in the offgas. The mass of undecane removed by volatilization
during air sparging (20 September-22 October in Fig. 2) was 280 g. This was higher
than the sum of all the other target compounds. The target compounds were only a
small percentage (less than
5 %)
of the TVOC concentration. The mass of undecane
collected during AS/SVE was only about 10% of the calculated mass removed based
on the soil core analyses (Gierke et al., 1998). This discrepancy may suggest that
other removal mechanisms, such as biodégradation, were significant; however, the
applicability or representativeness of using point measurements of contaminant
concentrations in the soil for extrapolating to a total contaminant mass in the
treatment zone is uncertain.
Monitoring of C02 in the offgas showed that elevated soil-gas concentrations of
C02
existed at all times in the vadose zone of the test cell, especially during periods
of blower shutdown (Fig. 3). Concomitantly, oxygen concentrations in the soil gas
would drop during blower shutdowns (Muraski, 1997). The production rate of C02
was used to estimate an overall aerobic biodégradation rate in the test cell. The
average rate of C02 production was about 15 mol day"1 (Fig. 3), which translates, for
example, to a stoichiometrically equivalent undecane degradation rate of 210 g day"1
(1 mol of undecane, CnH24, yields 11 mol of C02). Of course, the C02 produced was
100000
10000
co2
Concentration
(ppmj
1000
100
+ SG1
DSG3
XSG5
HjSVE Offgas
ASG2
0SG4
£ SG6
MfaftÉaSÏfEg
LJ-H
28-Aug 4-Sep 11-Sep 18-Sep 25-Sep 2-Oct 9-0ct 16-Oct 23-Oct
Date
Fig. 3 Carbon dioxide concentrations in soil gas and SVE offgas during SVE
(28 August-19 September) and AS/SVE (20 September-22 October).

124 John S. Gierke et al.
from a combination of many contaminants, both above and below the water table,
and soil organic matter as well. Nevertheless, this estimate illustrates at least the
magnitude of the biodégradation potential in the test cell when aerobic conditions
exist.
Even though the AS/SVE system was not successful in completely remediating
the saturated zone in Cell 1 during the one-month test, it was far more effective than
mass removal rates by water flushing (Wojick, 1998; Gierke et al., 1998). Enhanced
pump-and-treat techniques, such as surfactant-, co-solvent-, and cyclodextrin-
flushing, were more effective than air sparging at this site (cf. Brusseau et al., 1998).
The poorer performance of air sparging at this site was due in a large part to the fact
that the contaminant mixture was predominantly made up of compounds with vapour
pressures below 5 mm Hg.
Acknowledgements This work was funded by the Strategic Environmental Research
and Development Program, jointly sponsored by the US Department of Defense, US
Environmental Protection Agency, and US Department of Energy. The authors
gratefully acknowledge Drs C. G. Enfield and A. L. Wood (US Environmental
Protection Agency), Dr J. S. Ginn (Hill Air Force Base, Utah) and Mr D. L. Perram
(Michigan Technological University) for their technical contributions in performing
this study.
REFERENCES
Bausmith, D. S., Campbell, D. J. & Vidic, R. D. (1996) In situ air stripping.
Wat.
Environ.
Technol.
8,
45-51.
Brusseau, M. L., Amiable, M. A., Gierke, J. S. & Sabatini, D. A. (eds) (1998) Field
Testing
of
Innovative Subsurface
Remediation Technologies. ACS Symposium Series, American Chemical Society, Washington DC, USA. In
review.
Clayton, W. S., Bass, D. H., Ram, N. M. & Nelson, C. H. (1996) In situ sparging: mass transfer mechanisms.
Remediation
6, 15-29.
Gierke, J. S. & Powers, S. E. (1997) Increasing implementation of in situ treatment technologies through field-scale
performance assessments.
Wat.
Environ.
Res. 69, 196-205.
Gierke, J. S., Wojick, C. L. & Hutzler, N. J. (1998) Field test of air sparging. In: Field Testing of Innovative
Subsurface Remediation Technologies
(ed. by M. L. Brusseau, M. A. Amiable, J. S. Gierke & D. A. Sabatini).
ACS Symposium Series, American Chemical Society, Washington, DC, USA. In review.
Hein, G. L., Gierke, J. S., Hutzler, N. J. & Falta, R. W. (1997) Three-dimensional experimental testing of a two-phase
flow-modeling approach for air sparging.
Ground
Wat.
Monit.
Remediation
17, 222-230.
Lundegard. P. D. & LaBrecque, D. (1995) Air sparging in a sandy aquifer (Florence, Oregon, USA): actual and
apparent radius of influence. J.
Contain.
Hydrol. 19, 1-27.
McCray, J. E. & Falta, R. W. (1996) Defining the air sparging radius of influence for groundwater remediation. J.
Contain.
Hydrol. 24, 25-52.
Montgomery Watson Consultants (1996) Hill Air Force Base, Utah, Draft Phase II Work Plan for Eight Treatability
Studies at Operable Unit 1. April 1996.
Muraski, J. M. (1997) In situ and laboratory measurements of biodégradation rate constants for a sandy-gravel soil
contaminated with petroleum and chlorinated hydrocarbons. MSc Thesis, Dept Civil & Environ. Engng, Michigan
Technological Univ., Houghton, Michigan, USA.
Nyer, E. K. & Suthersan, S. S. (1993) Air sparging: Savior of ground water remediation or just blowing bubbles in the
bath tub?
Ground
Wat.
Monit.
Remediation
13,
87-91.
Wojick, C. L. (1998) Comparison of several alternative and traditional methods for the performance assessment of an air
sparging/soil vapour extraction field test. PhD Thesis, Dept Civil & Environ. Engng, Michigan Technological
Univ., Houghton, Michigan, USA.