This document describes the design, construction, and operation of a funnel and gate permeable reactive barrier used to remediate petroleum hydrocarbons in groundwater. A white spirit spill contaminated groundwater near a factory, which flowed into a nearby river. A funnel and gate system was constructed, with an impermeable barrier funnel directing groundwater through an air sparging unit and blended peat gate for treatment before entering the river. Monitoring over 10 months showed the system effectively reduced contaminant concentrations below regulatory limits and prevented further pollution of the river.

1. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND
GATE IN-SITU PERMEABLE REACTIVE BARRIER FOR
REMEDIATION OF PETROLEUM HYDROCARBONS IN
GROUNDWATER
TERRY McGOVERN1, TURLOUGH F. GUERIN2∗, STUART HORNER1
and BRENT DAVEY3
1 SRS Australia Pty. Ltd., Environmental Consultants, Werribee, Victoria, Australia; 2 Shell
Engineering, NSW State Office, P.O. Box 26, Rosehilly, New South Wales 2142, Australia; 3 Prpic
Davey Consulting, Murrumbeena, Victoria, Australia
(∗ author for correspondence, e-mail: turlough.guerin@shell.com.au, phone: 61-417-124453)
(Received 27 November 2000; accepted 4 April 2001)
Abstract. A white spirit spill at a factory site located in a residential area of south eastern Australia
led to contamination of shallow groundwater that fed into a nearby river. The contaminated ground-
water contained toluene, ethyl benzene, and xylene and n-alkanes in the C6–C36 fraction range. A
funnel and gate permeable reactive barrier was designed and built, based on preliminary pilot scale
tests using peat as the medium for the gate and the work conducted is presented as a case study.
The technical effectiveness of the funnel and gate, over the 10 month operating period in which data
was collected, indicates that peat represents an effective material for use in the gate component of
funnel and gate remedial systems. The application of the funnel and gate technology represented a
substantial saving to the client and was effective in preventing ongoing pollution of the nearby river.
The construction of the funnel and gate system also incurred the minimum disturbance to the public
access areas between the facility and the river.
Keywords: case study, dissolved phase, efficiency, groundwater, passive remediation, peat, per-
meable barriers, petroleum hydrocarbons, remediation, sparging
1. Introduction
Treatment walls, or permeable reactive barriers, first reported by McMurty and
Elton (1985), involve construction of permanent, semi-permanent, or replaceable
units across the flow path of a dissolved phase contaminant plume (Starr and
Cherry, 1994; Vidic and Pohland, 1996). As the contaminated groundwater moves
passively through the treatment wall, contaminants are removed by physical, chem-
ical and/or biological processes, including precipitation, sorption, oxidation/reduc-
tion, fixation, or degradation. These barriers may contain agents that are placed
either in the path of contaminant plumes to prevent further migration or immedi-
ately downgradient of the contaminant source to prevent plume formation (Vidic
and Pohland, 1996).
Water, Air, and Soil Pollution 136: 11–31, 2002.
© 2002 Kluwer Academic Publishers. Printed in the Netherlands.

2. 12 T. McGOVERN ET AL.
Several methods have been developed for the installation of permeable treat-
ment walls (Eykholt and Sivavec, 1995; Steimle, 1995). The majority of experience
with installation of these walls is with relatively shallow emplacements (<10 m)
using standard geotechnical design and construction approaches, although a few
technologies for deeper installations have been identified. In the simplest case, a
trench of the appropriate width can be excavated to intercept the contaminated
strata and backfilled with reactive material. This method would normally be limited
to shallow depths in stable geologic materials.
The major advantages of permeable treatment wall technology over other ground-
water remediation approaches are the reduced operation and maintenance (O&M)
costs and the enhanced technical efficacy, particularly compared with pump-and-
treat approaches (Thomas and Ward, 1995; Clark et al., 1997). Other than ground-
water monitoring, the major factor affecting O&M costs is the need for periodic
removal of precipitates from reactive media or periodic replacement-rejuvenation
of the affected sections of the permeable wall. The key issues associated with the
design of a treatment wall include the selection of the reactive media (chemistry,
particle size distribution, proportion and composition of admixtures), residence
time in the reaction zone, and the reaction zone size for optimum life span, as
well as the effect of the reaction zone medium on groundwater quality and the
ultimate fate of a treatment wall, which may impact upon its disposal. Selection
of the reactive media is based on the type (i.e., organic vs. inorganic) and concen-
tration of groundwater contaminants to be treated, groundwater flow velocity and
water quality parameters, and the available reaction mechanisms for the removal
of contaminants (i.e., sorption, precipitation, and biodegradation).
There are disadvantages in using permeable treatment wall technology. A criti-
cism that the life expectancy of the reactive media in a treatment wall may degrade
with time has been addressed by developing a construction approach whereby the
reactive media is placed in the subsurface in removable cassettes (Day et al., 1999).
Another criticism is the high capital costs in setting up such systems and the highly
specificity of barrier wall materials towards particular contaminants.
Typically, treatment wall system design is based on the results of treatability or
effectiveness studies that can incorporate both batch reaction tests and laboratory-
or field-scale column experiments (Rael et al., 1995). Batch tests are conducted
to obtain initial measures of media reactivity (i.e. degradation half-life, sorption
kinetics and capacity) that form the basis of the treatment wall design. Column tests
are typically conducted by packing a vertical column with the reactive medium
and passing the contaminated groundwater through the column until steady-state
performance of the reactor is obtained. Flow velocities are adjusted to simulate
groundwater velocity and reactor residence time. In addition, the impact of the
treatment wall on groundwater quality can be assessed from these studies. The life
span of sorption and precipitation barriers is limited by the ultimate capacity of
the medium to facilitate appropriate removal reactions. Once the ultimate capacity
of the medium is exhausted, contaminant breakthrough will occur, and this has

3. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 13
previously been observed with peat used to remove dissolved and free phase and
petroleum hydrocarbons from water (Stehmeier, 1989). In addition, contaminant
release or resolubilisation may occur.
The funnel and gate system is an application of a permeable reactive barrier
for in situ treatment of dissolved phase contamination (Starr and Cherry, 1994).
Such systems consist of low hydraulic conductivity cut-off walls (e.g. 1 × 10−6
cm
s−1
) with one or more gaps that contain permeable reaction zones. Cutoff walls
(the funnel) modify flow patterns so that groundwater primarily flows through high
conductivity gaps (the gates). The type of cutoff walls most likely to be used in cur-
rent practice are slurry walls, sheet piles, or soil admixtures applied by soil mixing
or jet grouting. At the current time there are relatively few published reports on the
full-scale design and operation of funnel and gate systems and related systems for
dissolved phase petroleum hydrocarbons (Bowles et al., 2000; Schad and Schulze,
2000) with even fewer reporting the use of peat (Kao and Wang, 2000). The current
article presents the findings of a case study on the use of a funnel and gate system
in Australia.
2. Background
A white spirit petroleum hydrocarbon spill occurred at a factory facility in South
Eastern Australia in December 1997. This volatile petroleum hydrocarbon was
used routinely as a solvent at the facility (Table I). A leaking underground storage
tank (UST) caused a quantity of the white spirit (∼2000–3000 L) to leak through
the scoria fill material in which the UST was located, into the soil of the upper
embankment at the rear of the facility. A quantity of the white spirit emerged at
the base of the slope, discharging into a spoon drain, and then via a culvert, flowed
into the nearby river. A proportion of the leaked white spirit was absorbed by the
soil in the embankment and an amount entered the groundwater underlying the site
(Figures 1 and 2).
3. Study Objective
The aim of the remedial works conducted in this study was to intercept the dis-
solved phase plume of petroleum hydrocarbons present in the shallow aquifer
moving toward the river, using a funnel and gate permeable barrier, and reduce con-
centrations to less than regulatory limits. A funnel and gate system was considered
to be the most cost-effective way for treating petroleum hydrocarbon contaminated
groundwater and preventing further contamination of the river. In the proposed re-
mediation strategy, contaminated groundwater flow would be directed towards the
gate which would then be subject to two treatment processes, applied in sequence,
within the funnel and gate system. The processes applied were (1) air sparging, as

4. 14 T. McGOVERN ET AL.
Figure1.Planofstudyareashowingmainfacilityarea,locationofspillarea(darkshading),funnelandgatesystemandnearbyriver.

5. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 15
Figure2.Transectthroughstudyarea(refertoB–BonFigure1).

6. 16 T. McGOVERN ET AL.
TABLE I
Composition of White Spirita
Component Name % (v/v)
C6–C9 Benzene 0.07%
Toluene 0.04%
Ethyl benzene 1.10%
Total xylene 20%
Total C6–C9 ∼23%
C10–C14 n-Alkanes 77%
C15–C28 n-Alkanes <1%
C29–C36 n-Alkanes <1%
a CAS number is 8052-41-3. White spirit is
a dry cleaning solvent, also known as mineral
spirits or Stoddard solvent and is a colorless
liquid with a kerosene-like odor.
previously described (Pankow et al., 1993), and (2) absorption and biodegradation
within a mixture of blended peats as previously highlighted (Cohen et al., 1991;
Couillard, 1994).
The development of the remediation strategy for the site and treatment of the
contaminated groundwater after the white spirit spill is outlined in Table II.
4. Site Description and Methodology
4.1. SITE LAYOUT AND GEOLOGY
The area between the rear of the facility and the nearby river consists of a relatively
steep (∼45◦
) embankment, which is bisected by a cycle path of approximately
2 m width. A spoon drain runs parallel to the western side of the access track,
terminating approximately 20 m north of the northern boundary of the site in a
culvert, which runs under the cycle path and discharges directly to the nearby river.
The facility occupies relatively level ground, exhibiting a lesser incline towards the
river. On site structures consist of a large 2 storey building, the lower floor of which
is used for hide tanning activities. A tank farm is located in the north eastern corner
of the premises, consisting of four underground storage tanks, which are devoted
to either fresh or used white spirit (Figure 1).
The geology of the site consists of basement siltstones and sandstones of the
upper Silurian Dargile Formation, the upper profile of which has weathered to a
tan-orange mottled sandy clay. These sediments are visible in outcrops at several
locations along the banks of the nearby river. Variably weathered lavas of the Qua-
ternary Newer Basalts formation overlie the Silurian sediments in the eastern half

7. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 17
TABLE II
Development of the remediation strategy after the White Spirit spill
Key activity Date Tasks
Control and containment December 1997 Boom and sorbents placed in river to capture
of spill spilled solvent leaving the site
Removal of leaking tanks
Initial characterization December 1997– Sampling of soil and groundwater
of soil and groundwater February 1998 Developed model of contamination
contamination and its at the site
extent
Selection of remediation March 1998 Selection of funnel and gate option
strategy
April 1998 Feasibility study to determine optimum
configuration of peat
Identified options for remediating
petroleum hydrocarbon contaminated soil
Implementation of March–May 1998 Installation of funnel and gate
remediation strategy
Remediation of contaminated soil excavated
from the site
Monitoring July 1998– Initial 10 month groundwater
May 1999 sampling and analysis program was
conducted to assess effectives of funnel
and gate system
Review of effectiveness March 1998– Soil remediation was validated by sampling
of remediation strategy August 1998 and analysis of the recirculating air during
bioventing, and sampling and analysis of the
soil in the biovented piles for BTEX and
n-alkanes
September 2000– Three monthly groundwater sampling and
Onwards analysis program is planned
of the site, as identified in monitoring wells (MW) 9 and 10. Varying thicknesses of
fill comprise the upper and lower embankments which is of a highly heterogeneous
nature.

8. 18 T. McGOVERN ET AL.
4.2. SITE HYDROGEOLOGY
A highly localised, shallow perched water table is present in the fill material which
comprises the embankment between the factory and the nearby river. The regional
watertable is expected to be located at approximately the same depth as the level
of the river. Residual clays from the weathered Silurian siltstones tend to be imper-
meable, and as such perched water tends to be located at the fill-natural sandy clay
interface.
4.3. AQUIFER TESTING
Aquifer recovery tests were carried out on monitoring wells MW1, MW2, MW3,
MW4, MW5, MW6, MW7, and MW8 on 20 November 1998. The procedure
involved the removal of a known volume of groundwater from the well and mon-
itoring the subsequent recovery with time. The hydraulic properties of the fill
material were investigated (by slug test) to determine the hydraulic conductivity
of this material. The Hvorslev method (Fetter, 1988) was used to calculate the
hydraulic conductivity in the immediate proximity to each well tested.
4.4. CONSTRUCTION OF FUNNEL AND GATE SYSTEM
A volume of ∼1600 m3
of petroleum hydrocarbon contaminated soil was excavated
as part of establishing the funnel and gate system. This was placed in 3 biovented
piles at the facility (refer to biobanks in Figure 1). Details of the remediation of this
soil are not given in this report. The funnel and gate system was comprised of an
impervious barrier membrane (i.e. the funnel), directing groundwater into the treat-
ment area (i.e. the gate). The gate consisted of a sparging unit upgradient of a blend
of peat materials. The funnel component, designed to intercept the contaminant
plume before it entered the nearby river, consisted of a 0.75 mm thick High-
Density Polyethylene (HDPE) impervious barrier membrane positioned vertically
in the cut-off trench to capture and redirect incident groundwater over the length
of the spill area, parallel to the nearby river. The gate was composed of sequential
treatment systems comprising a sparging unit emplaced in basaltic scoria, followed
by blended peat materials. The funnel was designed to intercept groundwater flow
from areas directly downgradient of the spill site, as well as adjacent areas in which
lateral migration of the plume may be occurring. The dimensions of the funnel
trench are 27 m long × 5 m deep × 0.6 m wide, excavated to a level of 0.5 m
below the fill-natural siltstone interface.
4.4.1. Gate Construction
A stepwise procedure was undertaken so as to ensure the final integrity of the
gate and prevent the collapse of the riverbank. This was achieved in the following
manner. Chain mesh fencing (2 m × 15 m long section) was laid flat along the
river-embankment interface and fastened to the ground using 1.5 m long steel star

9. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 19
Figure 3. Transect through the funnel and gate system (as built) (refer to Section A–A, Figure 2).
pickets to aid long term stability. Sheet piling was inserted into the embankment
parallel to the river bank, at a distance of approximately 4.0 m from the river,
dividing the areas to be excavated for both components of the gate system. The area
designated for the peat (between the river bank and the sheet piling) was excavated
to a depth of 0.5 m below the siltstone-fill interface, and infilled with peat to a depth
of 0.5 m below surface. The remaining volume was filled with natural soil overlying
a layer of A15 Type Bidim Geofabric to prevent fine soil particles mixing with and
clogging the peat and scoria filtration media. The area was then sown with grass
seed in order to promote bank stability. Excavated soil was removed for bioventing.
Subsequent to the emplacement of the peat, soil adjacent to the opposite side of the
sheet piling was excavated to a depth of 0.5 m below the fill-natural siltstone strata
interface, and a horizontal sparging tube and scoria was emplaced. Thirty cubic
meters of peat (on a wet volume basis of estimated at 40% moisture) was used in
the gate.
4.4.2. Sparging Component of Gate
The section of the gate that initially encounters influent groundwater consists of a
volume (∼5 m3
) of porous basaltic scoria. The scoria was emplaced over a sub-
merged, perforated air sparging pipe, creating fine air bubbles, which permeate the
water column within the gate, delivering oxygen to enhance biodegradation. Air
was supplied to the system by a 12-cfm compressor, supplying air at 0.9 m3
hr−1
.
Soil gas and air concentrations of petroleum hydrocarbons were monitored at the
site using a Foxboro TVA 1000A Flame Ionization Detector (FID) and a Photo-
Ionization Detector (PID) unit. A transect diagram through the gate is given in
Figure 3.

10. 20 T. McGOVERN ET AL.
4.4.3. Peat Composition and Implications for Contaminant Remediation
After the groundwater was sparged (as it passed through the scoria as described
above), the groundwater then passed through a peat mixture, immediately down-
gradient of the sparging system. Previous studies indicate the high sorptive ca-
pacities of humic materials (such as peat) for petroleum hydrocarbons (Kao and
Borden, 1997). Peats with low fibre contents and high lignin pyrolysis material
and ash content have been shown to be the most effective peats to remove free and
dissolved phase petroleum hydrocarbons from groundwater (Cohen et al., 1991).
The ability to both adsorb dissolved and free phase petroleum hydrocarbons from
groundwater, and provide a catalytic surface on which microbial activity can occur,
indicated the suitability of peat for use in groundwater treatment (Cohen et al.,
1991). Cohen et al., (1991) have indicated that a wide range of peats is effective
in removal of petroleum hydrocarbon contamination from water, with rates of re-
moval efficiency between 63 and 97%. Using this information, a locally available
peat, known commercially as Biogreen Humic Reed Sedge, with low fibre and high
humic content was selected for use onsite.
4.4.4. Peat Blending and Implications for Hydraulic Conductivity
An important consideration in the design and construction of funnel and gate sys-
tems is that the treatment media be of an equal or greater hydraulic conductiv-
ity than that of the surrounding substrate, to allow incident groundwater to be
channeled through the treatment aperture. This is particularly important as pre-
vious studies have shown that the variability of the saturated hydraulic conduct-
ivity of peat can be as high as 10 orders of magnitude, depending on the type
of peat. Accordingly, hydraulic conductivity tests were performed on a variety
of peat/sphagnum peat/cocoa fibre blends, in order that an optimal mixture, with
respect to sorptive capacity, air circulation and hydraulic conductivity, be achieved.
Tests were conducted using the ‘falling head test’ method (Schwartz and Domen-
ico, 1997). Equipment consisted of a column of packed, saturated peat. The fall in
hydraulic head over time (K) was measured from a manometer tube, and substi-
tuted in the following equation:
K = 2.3(aL/A(t1 − t0)) log10 h0/h1 ,
where
a = cross sectional area of the stand pipe,
L = length of sample,
A = cross sectional area of sample tube,
(t1 − t0) = elapsed time required for a head fall from h1 to h0,
h0 = starting height of water in the manometer tube,
h1 = end height of water in the manometer tube.

11. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 21
A blend of 70% peat (Biogreen Humic Reed Sedge), 20% sphagnum peat and 10%
cocoa fibre was selected for use. This mix contained the maximum percentage of
the peat with the highest sorptive capacity (Biogreen Humic Reed Sedge), without
reducing the hydraulic conductivity to below that encountered in the surrounding
fill materials.
5. Monitoring Well Installation, Sampling and Analysis
Effective funnel and gate design ensures sufficient residence time of contaminants
within the treatment zone as well as downgradient substrata permeability, which
prevents contaminated water from ‘banking up’ and circumventing the funnel and
gate. Testing of the funnel and gate was achieved via two groundwater monitoring
wells, one emplaced immediately upgradient of the gate, and another downgradient
of the gate area infilled with peat. By comparing influent and effluent concentra-
tions of n-alkanes and BTEX compounds, it has been possible to determine the
effectiveness of the funnel and gate system for removing dissolved phase petro-
leum hydrocarbon contamination. The monitoring system was set up such that any
break-throughs of petroleum hydrocarbons (through the gate) could be detected.
Furthermore the bores at the site were arranged so that any petroleum hydrocarbons
escaping the gate (by going around the funnel) would be detected.
A total of 14 monitoring wells were installed at the site. All monitoring wells
were fully screened from the watertable to the end of the boreholes. Since the estab-
lishment of the monitoring network in May 1998, the wells have been sampled on a
fortnightly basis for a period of 5 months, followed by a less intensive sampling re-
gime up to September 2000, as agreed to by the local regulatory authorities. Stand-
ing water levels were measured prior to well purging and sampling on all occasions.
All groundwater samples were analysed for n-alkanes (by GC-FID method, US
EPA 8021B), and BTEX using the purge and trap method (US EPA Method 5030).
All soil and groundwater samples were collected in Teflon (PTFE)-lined screw cap
glass jars and keep on ice in coolers (∼4 ◦
C) and sent for analysis within 24 hr of
collection. Recoveries for spiked n-alkanes varied from 75–100% in groundwater
samples. Recoveries for BTEX were 85–100% in groundwater samples. Laborat-
ory duplicates reported variation for n-alkane fractions and BTEX of 5–25% in
groundwater samples.

12. 22 T. McGOVERN ET AL.
TABLE III
Volumetric budget summary for the steady state calibration
simulation
Flux Description/origin Volume
(L day−1)
Into model Recharge 76
Bank seepage 3539
Out of model Constant heads (river-seepage) 3591
% Discrepancy 0.65
6. Results and Discussion
6.1. WATER TABLE LEVEL AND HYDRAULIC CONDUCTIVITY TESTS
The water table configuration remained relatively constant throughout the monit-
oring period. The groundwater flow lines were observed to flow toward the east,
approximately perpendicular to the bank of the river. Figure 4 illustrates the config-
uration of the water table (m AHD) on 7 October 1998. Variations in the water table
were observed with monitoring wells MW1, MW2, and MW3 illustrating similar
temporal changes. These wells are furthest from the nearby river, located at the
base of the slope. Closer to the river the water table variation is significantly less,
indicating that the water table close to the river is being controlled by the water
level in the river itself. Those wells located furthest from the river were expected
to reflect rainfall (recharge) water table response relationships. As expected, the
wells nearest the slope fluctuate the most in response to rainfall events, suggesting
recharge is occurring from the slope to the west, most likely as interface drainage.
The results of the subsurface conductivity tests are presented in Table IV. As
expected the material was found to have extremely variable hydraulic properties.
Measurement of the conductivity of the fill material was in some instances erro-
neous as the well screens extended through two lithologies, the clay overlying the
siltstone and the fill material itself. The large variability of the fill material suggests
that the majority of groundwater and contaminant migration through this material
will be occurring along preferential flow paths. This is likely to result in faster
migration times and a larger spread of the plume.

13. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 23
Figure 4. Measured and interpolated piezometric surface on 7 October 1998 (units in m).
6.2. ESTIMATES OF GROUNDWATER VOLUMES AND PETROLEUM
HYDROCARBON MASS ENTERING FUNNEL AND GATE SYSTEM
From modelling conducted, the concentrations of ethyl benzene and xylene in the
groundwater discharging to the river (ignoring mixing zones) would exceed the
relevant criteria, 250 and 620 µg L−1
, respectively, within 6–12 months and would
continue to exceed these criteria for a period in excess of 6 yr. The modelling
estimated that the width of the discharge face (capture zone) was estimated at
∼20 m.

14. 24 T. McGOVERN ET AL.
TABLE IV
Measured hydraulic conductivity of
subsurface
Monitoring well Values
(m day−1)
MW01 0.18
MW02 1.63 × 10−8
MW03 1.26 × 10−5
MW04 0.0001
MW05 0.0046
MW06 0.0001
MW07 0.025
MW08 0.68
The maximum concentrations of n-alkanes and the total of benzene, toluene,
ethyl benzene and xylenes (BTEX) present in groundwater from bore wells across
the site were 1000 and 25 mg L−1
, respectively. These values were reported between
December 1997 and December 1998.
The average concentration of n-alkanes in groundwater over the period from
December 1997 to August 1999, was 26 mg L−1
of n-alkanes, and the value for
BTEX was considerably lower at <0.05 mg L−1
. The predicted amount of petro-
leum hydrocarbons likely to be intercepted by the proposed funnel and gate system
are conservative estimates designed to encompass the worst case scenario. Based
on a median hydraulic conductivity of 0.4 m day−1
(estimated from the site data),
the median volume of groundwater entering the gate was estimated to be ∼6 m3
day−1
. The maximum volume of water entering the gate, based on a conservative
estimate, was expected to be ∼43 m3
day−1
. Assuming the average concentration
of petroleum hydrocarbons in groundwater was 26 mg L−1
, the maximum mass of
petroleum hydrocarbons entering the funnel and gate was calculated to be 1.1 kg
day−1
, and a median value of 160 g day−1
.
Assuming a median removal capacity of 20 kg t−1
peat (dry weight basis), as
determined from previous studies with a range of peats tested for absorbing mono
aromatic hydrocarbons (Cohen et al., 1991), it was expected that the petroleum
hydrocarbons in the groundwater entering into the funnel and gate (i.e. 1.1 kg
day−1
), would be treated by the system (as designed) for up to ∼210 days before
break through of petroleum hydrocarbons. This period would be longer at 1448
days, assuming 160 g day−1
of petroleum hydrocarbons in the influent, entering the
gate. These estimates, derived from Cohen et al. (1991), are conservative because
they do not take into account a highly active microbial population which would

15. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 25
be contributing to the apparent absorption (i.e. removal) process in the gate at
the current site. Previous studies have shown that such microbial populations are
readily established in peat when oil is added (Jack et al., 1994). These estimates
do, however, provide some indication of the size and design of the gate required
for removing the dissolved phase petroleum hydrocarbons. At 306 days, the overall
performance of the funnel and gate has been satisfactory. Further monitoring will
reveal the effectiveness of the design (see Implications section of this article).
6.3. PETROLEUM HYDROCARBON REMOVAL EFFICIENCIES
Initial contaminant concentrations determined from samples collected on 1 July
1999, i.e day ‘zero’, in the well upgradient of the funnel and gate (referred to as
MW 13) were equivalent to those downgradient (of the funnel and gate treatment)
(referred to as MW 14) (Table V). The expected decrease of n-alkanes and toluene,
ethyl benzene, and xylene concentrations across the funnel and gate system was not
apparent in this initial round of sampling and analysis. No benzene was detected
in either MW 13 or MW 14 over the 10 month monitoring period. Subsequent
monitoring from groundwater monitoring wells MW 13 and MW 14, conducted
over the following 10 month period that has been reported, indicated a substantial
and consistent reduction of petroleum hydrocarbon contaminants in the ground-
water due to passage through the funnel and gate system. Concentrations of both
n-alkanes and toluene, ethyl benzene, and xylene, have substantially decreased in
concentration between MW 13 (upgradient) and MW 14 (downgradient) of the
funnel and gate.
The removal efficiencies for the as-built funnel and gate system varied from
63 to 96% for the monoaromatic hydrocarbons, toluene, ethyl benzene, and xy-
lene. Removal efficiencies for C6–C9, C10–C14, and C15–C28 n-alkane fractions
were 69.7, 81.1 and 67.2%, respectively. The lowest removal efficiencies were for
the C29–C36 n-alkane fraction at 54%. The overall efficiency for the funnel and
gate system was 72% during the 10 month period over which data was collected
(Table VI). Removal efficiencies tended to decrease at 188 and 306 days with
respect to toluene, ethyl benzene, and xylene and the C6–C9 n-alkane fraction.
Further sampling and analysis in the future will ascertain whether this represents a
trend in decreasing performance of the funnel and gate. The apparent increase in
concentrations of both the n-alkanes and toluene, ethyl benzene and xylenes in the
upgradient well at day 245 correlate with a preceding heavy rain in days before the
samples were taken at that monitoring time.
Previous studies have shown that peats from various sources can remove 63–
97% dissolved phase benzene, toluene, and xylenes from contaminated ground-
water in laboratory studies and this data was used in the selection of peat for the
current study (Cohen et al., 1991), and the same range has been obtained in the
current study. Kao and Borden (1997), reporting efficiencies of BTEX removal
from laboratory groundwater undergoing remediation with a briquet-peat barrier

16. 26 T. McGOVERN ET AL.
TABLEV
BTEXgroundwatermonitoringdatafromupgradientanddowngradientofdissolvedphaseplume(µgL−1)
DaysTolueneEthylbenzeneXyleneC6–C9C10–C14C15–C28C29–C36
UGDGUGDGaUGbDGUGDGUGDGUGDGUGDG
0116280990250410110037010019010120150
81115310260011029005010020010046317
341118530740060220023009501900710412032
4144213201900130780560440100100117062
48310311703100803500400100320100215126
584021780320011066003406305080391
153151123020006058015050018012013816
188211202400207507804801001002216
2151711270270030350027041014001<1<1
245111109402800270300044051015012016840250
2741113401400301700530730150017511
306115100220080310041074010010061018
Mean36.21.313.4358272411024186114973081434.314554.5
U95%C.I.50.71.6167733642173439711768234161645.538865.3
aReferringtoupgradientbore(BoreMW13).
bReferringtodowngradient(BoreMW14).

17. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 27
TABLEVI
Efficiencyofthefunnelandgatesystem
StatisticTolueneEthylXyleneC6–C9
bC10–C14
bC15–C28
bC29–C36
bOverall
benzeneefficiency
(%)c
Upper95%C.I.96.965.483.277.667.781.360.576.1
Average96.368.362.569.277.679.553.572.4
Modalaverage95.464.175.069.781.167.261.873.5
aNotdetected.
bn-Alkanes.
cEfficiency(%)ofremoval=(contaminantingroundwaterindowngradientwell)/(contaminantgroundwaterinupgradientwell)×
100.

18. 28 T. McGOVERN ET AL.
system, stated values of 86, 71, 43 and 28% for toluene, ethylbenzene, m-xylene
and o-xylene, respectively. The removal efficiencies reported in the current field
study were therefore considerably higher than those reported in previous laboratory
studies. In another laboratory study, a peat column decreased petroleum hydro-
carbon concentrations in an oily water mixture by 100 fold at 30 000–70 000 mg
L−1
(Stehmeier, 1989). However, with a much lower dissolved phase concentration
of 5 mg L−1
, the peat column was ineffective at removing the petroleum hydro-
carbons (Stehmeier, 1989). The literature values therefore vary considerably for
petroleum hydrocarbon removal effectiveness. These previous studies reinforce the
importance of testing local peat types prior to deploying peat at full-scale.
The downgradient monitoring well reported concentrations of toluene, ethyl
benzene, and xylene and n-alkanes in the range C6–C36 below the surface water
guidelines for these pollutants demonstrating that the funnel and gate system is
also meeting the regulatory requirements, for the shallow groundwater entering the
river (Table VII; see row with SEPP values). All the guideline values that were
reviewed were met with respect to the specific contaminants of concern in the
groundwater from the site. The remediation will be considered completed when the
downgradient effluent is consistently reporting toluene, ethyl benzene, and xylene
and n-alkanes in the range C6–C36 below the surface water guidelines, over a period
of approximately 1 yr (with quarterly reporting).
7. Implications
This case study has described the effectiveness of an application of the funnel and
gate technology at a contaminated site in South Eastern Australia. The implications
from the success of the funnel and gate permeable barrier reported in the current
study are:
1. Technical effectiveness. The technical effectiveness of the funnel and gate, over
the 10 month operating period in which data was collected, indicates that peat
represents an effective material for use in the gate component of funnel and
gate remedial systems. Though efficiencies were not high for all contaminants,
further ongoing monitoring will allow more definite conclusions to be drawn
regarding the selective degradation capacities of the peat for each of the con-
taminants tested. The use of sparging apparatus at the base of the gate would
have contributed to the technical effectiveness of the funnel and gate system.
2. Cost effectiveness. The costs incurred for this project were ∼71 500 USD (equiv-
alent to ∼130 000 AUD) which included the construction of the funnel and gate
system and bioremediation of the contaminated soil. Alternatively, the costs of
treating the groundwater by a pump and treat method were estimated at ∼USD
83 500 or 160 000 AUD (capital costs) and ∼USD 0.2 m−3
(∼5500 USD or
∼10 000 AUD) for annual operational costs and a total cost of ∼95 000 USD

19. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 29
TABLE VII
Summary of investigation levels of contaminants in surface waters (µg L−1)
Beneficial use Benzene Toluene Ethyl Xylene C6–C9 C10–C14
benzene
Production of edible 120a, 140b 250 250 380b – –
fish and crustaceaa
Recreation 470c – – – 10000d 10000d
Aesthetic objectives – – – – 10000d 10000d
SEPP groundwater – 250 250 – – –
qualityg
Protection of aquatic 300 300 140 380 –e –e
ecosystems
Groundwater ingestion – 5 1000 700 10000 –f –f
residentialf
Freshwater aquatic 300 300 – – – –
ecosystemsh
a US Code of Federal Register (Anonymous, 1994a).
b Dutch guidelines (RIVM, 1994).
c US Code of Federal Register (Anonymous, 1994b).
d Historical Australian and New Zealand guidelines (ANZECC, 1992).
e A dash indicates that ‘Information needed to select threshold concentrations is incomplete’. This
document states that the NSW Clean Waters Act (1970) and Clean Waters Regulations (1972)
require licenced discharges to be visually free of oil and grease (∼10 000 µg L−1) (Anonymous,
1994c).
f ASTM look up tables (ASTM, 1995).
g SEPP or State Environmental Protection Policy is the legal instrument by which the en-
vironmental regulators in the State of Victoria (Australia) can enforce clean up targets (refer
http://www.epa.vic.gov.au/wp/wp_sepp.htm).
h These are the current Australian groundwater quality guidelines and are not enforceable targets
per se (NEPC, 1999).
(170 000 AUD). The application of the funnel and gate technology therefore
represented a substantial saving to the client and was effective in prevent-
ing ongoing pollution of the nearby river. The construction of the funnel and
gate system also incurred the minimum disturbance to the public access areas
between the facility and the river.
3. Further research. Ongoing technical effectiveness of the funnel and gate bar-
rier will be conducted over the longer term, by monitoring at 3 month intervals
to ensure continued effectiveness of the funnel and gate system. This is import-
ant as previous laboratory studies have shown that BTEX removal efficiencies

20. 30 T. McGOVERN ET AL.
decrease with time (Stehmeier, 1989; Cohen et al., 1991). The key areas re-
commended for further research are as follows:
• Determining the effect of peat on the quality of the groundwater could also
be checked to ensure that there are no deleterious effects of the peat on the
effluent groundwater.
• Further research could be conducted to determine the mechanisms by which
the petroleum hydrocarbons are removed from the groundwater and to
monitor any deleterious effects on removal efficiencies from any potential
increase in the biomass associated with any microbial growth in the peat
within the gate.
• The bulk density of peat has also previously been shown to influence the
effectiveness of petroleum hydrocarbon removal from water (Stehmeier,
1989) and further research could be conducted to optimise this effect for
full-scale studies.
• Disposal of peat after the project is completed. Testing of the peat from
the gate should be conducted at completion of remediation. This should
be designed to determine if the petroleum hydrocarbons are irreversibly or
reversibly bound to the peat, and to determine any risks associated with the
used peat. This will have implications for disposal options.
References
Anonymous: 1994a, ‘Water Quality Criteria: Human Health Partial Body Contact’, U.S. Govern-
ment.
Anonymous: 1994b, ‘Water Quality Criteria Human Health Fish Consumption’, U.S. Government.
Anonymous: 1994c, ‘Guidelines for Assessing Service Station Sites’, Sydney, NSW, Australia, NSW
EPA, 31 pp.
ANZECC: 1992, Australian Water Quality Guidelines for Fresh and Marine Waters, Summary
of Water Quality Guidelines for Recreational Waters, Australian New Zealand Environment
Conservation Council and National Health and Medical Research Council.
ASTM: 1995, ‘Guide for Risk Based Corrective Action Applied at Petroleum Release Sites’, West
Conshohocksen, PA, ASTM Committee E50 on Environmental Assessment, 51 pp.
Bowles, M. W., Bently, L. R, Hoyne, B. and Thomas, D. A.: 2000, ‘In situ ground water remediation
using the trench and gate system’, Ground Water 38, 172–181.
Clark, D. K., Darling, D. F., Hineline, T. L. and Hayden, P. H.: 1997, Field Trial of the Biowall Tech-
nology at a Former Manufactured Gas Plant Site. 29th Mid-Atlantic Industrial and Hazardous
Waste Conference, Presentation by Stearns and Wheler, LLC, Cazenovia, NY, U.S.A.
Cohen, A. D., Rollings, M. S., Zunic, W. M. and Durig, J. D.: 1991, ‘Effects of chemical and phys-
ical differences in peats on their ability to extract hydrocarbons from water’, Water Resources
Research 25, 1047–1060.
Couillard, D.: 1994, ‘The use of peat in wastewater treatment’ Water Research 28, 1261–1274.
Day, S. R., O’Hannesin and Marsden, L.: 1999, ‘Geotechnical techniques for the construction of
reactive barriers, J. Hazard. Mat. B, 285–297.
Eykholt, G. R. and Sivavec, T. M.: 1995, Contaminant Transport Issues for Reactive-Permeable
Barriers, New York, NY, ASCE, pp. 1608–1621.

21. DESIGN, CONSTRUCTION AND OPERATION OF A FUNNEL AND GATE 31
Fetter, C. W.: 1988, Applied Hydrology, Oshkosh, Wisconsin, Merill Publishing Company, 592 pp.
Jack, T. R., Francis, M. M. and Stehmeier, L. G.: 1994, ‘Disposal of slop oil and sludges by
biodegradation’, Res. Microbiol. 145, 49–52.
Kao, C. M. and Borden, R. C.: 1997, ‘Enhanced TEX biodegradation in nutrient briquet-peat barrier
system’, J. Environ. Engin. (New York) 123, 18–24.
Kao, C. M. and Wang, C. C.: 2000, ‘Control of BTEX migration by intrinsic bioremediation at a
gasoline spill site’, Water Res. 34, 3413–3423.
McMurty, D. C. and Elton, R. O.: 1985, ‘New approach to in situ treatment of contaminated
groundwaters’, Environ. Progr. 4, 168–170.
NEPC: 1999, National Environmental Protection Measure – Schedule B, (1) Guidelines on the Invest-
igations Levels for Soil and Groundwater, Adelaide, Australia, NEPC (National Environmental
Protection Council), 11 pp.
Pankow, J., Johnson, R. L. and Cherry, J. A.: 1993, ‘Air sparging in gate wells in cutoff walls and
trenches for control of plumes of volatile organic compounds (VOCs)’, Ground Water 31, 654–
663.
Rael, J., Shelton, S. and Dayaye, R.: 1995, ‘Permeable barriers to remove benzene: Candidate media
evaluation’, J. Environ. Engin. 121, 411–415.
RIVM: 1994, Environmental Quality Objectives in The Netherlands – A Review of Environmental
Quality Objectives and Their Policy Frame Work in The Netherlands, Amsterdam, RIVM (Risk
Assessment and Environmental Quality Division, Directorate for Chemicals, External Safety and
Radiation Protection, Ministry of Housing, Spatial Planning and Environment, The Netherlands).
Schad, H. and Schulze, B.: 2000, Funnel and Gate at a Former Manufactured Gas Plant Site in
Karlesruhe, Germany: Design and Construction. Remediation of Chlorinated and Recalcitrant
Compounds – Chemical Oxidation and Reactive Barriers, Monterey, California, Battelle Press,
pp. 315–322.
Schwartz, F. A. and Domenico, P. A.: 1997, Physical and Chemical Hydrogeology, New York, NY,
John Wiley & Sons, 640 pp.
Starr, R. C. and Cherry, J.: 1994, ‘In situ remediation of contaminated groundwater: The funnel and
gate system’, Ground Water 32, 465–476.
Stehmeier, L. G.: 1989, Development of an Oclanosorb (TM) (peat) Filter for Separating Hydrocar-
bons from Bilge Water, Calgary, CA, Nova Husky Research Corporation, 12 pp.
Steimle, R.: 1995, In Situ Remediation Technology Status Report: Treatment Walls, Washington, DC,
US EPA, 31 pp.
Thomas, J. M. and Ward, C. H.: 1995, Ground Water for Bioremediation, New York, NY,
Geotechnical Special Publication, ASCE, pp. 1456–1466.
Vidic, R. D. and Pohland, F. G.: 1996, Treatment Walls, Pittsburgh, PA, Ground Water Remediation
Technologies Analysis Center, 38 pp.