Full text of DOI:10.1002/etc.5620211212

Environmental Toxicology and Chemistry, Vol. 21, No. 12, pp. 2606–2616, 2002
2002 SETAC
Printed in the USA
0730-7268/02 $9.00 .00
᭧
ϩ
INVESTIGATION OF AN ONSITE WASTEWATER TREATMENT SYSTEM IN SANDY
SOIL: SITE CHARACTERIZATION AND FATE OF ANIONIC
AND NONIONIC SURFACTANTS
A
LLEN M. NIELSEN,† ALVARO J. D
E
C
ARVALHO,*‡ DREW C. M
and DAMANN L. ANDERSON
†Sasol North America, 12024 Vista Parke Drive, Austin, Texas 78726-4050, USA
CAVOY,§ LOU KRAVETZ,ͦͦ MANUEL L. CANO,ͦͦ
#
‡The Soap and Detergent Association, 1500 K Street NW, Suite 300, Washington, DC 20005
§The Procter & Gamble Company, 6110 Center Hill Avenue, Cincinnati, Ohio 45224, USA
ͦͦShell Chemical Company, 3333 Highway 6 South, Houston, Texas 77082, USA
#Hazen & Sawyer, P.C., 10150 Highland Manor Drive, Suite 140, Tampa, Florida 33610, USA
(
Received 31 October 2001; Accepted 28 May 2002)
Abstract—This study reports on the fate of linear alkylbenzene sulfonate (LAS), alcohol ethoxylate (AE), and alcohol ether sulfate
(AES) surfactants in a home septic system near Jacksonville (FL, USA) that has been used since 1976. The drainfield at this site
resides in fine sand (Ͻ6% silt and clay) with an unsaturated zone that ranges from 0 to 1.3 m. During the wettest times of the
year, it is likely that effluent from the septic system passes directly into the groundwater without exposure to an unsaturated zone
of soil. Groundwater was collected during two sampling events, representing seasonal high and low groundwater table levels, and
analyzed for the surfactants LAS, AES, and AE. During the wet season, the unsaturated zone was approximately 0.01 m beneath
the drainfield. During the dry season, the unsaturated zone was about 0.4 m below the drainfield. Alcohol ethoxylate was not
detected in any groundwater samples during either sampling. Alcohol ether sulfate was not found in the dry season sampling, but
traces of AES had migrated downgradient about 4.7 m horizontally and 1.8 m vertically in the wet season. Linear alkylbenzene
sulfonate was detected in some dry season samples and had moved downgradient some 11.7 m horizontally and 3.7 m vertically
in the wet season. These observations demonstrate that these surfactants were removed to a great extent; otherwise, they would
have traveled more than 260 m downgradient, which is the calculated distance that a conservative tracer like bromide would have
moved downgradient over the life of the system. The most likely removal mechanisms for these surfactants were biodegradation
and sorption. Therefore, this study indicates that LAS, AE, and AES are readily removed from groundwater in soils below septic
system drainfields even in situations with minimal unsaturated soil zones.
Keywords—Groundwater
Surfactants
Fate
Septic systems
Soil
INTRODUCTION
of the chemical structures for commercial grades of these three
surfactants are shown in Table 1. Since these surfactants are
used primarily in cleaning and personal care products, their
usual route of disposal is down the drain to publicly owned
treatment works (POTWs) or OWTS, resulting in their wide
dispersal. Although the excellent environmental safety profiles
of these surfactants (including their high removals in POTWs)
are well documented [12,13], little information exists con-
cerning their fate beneath the drainfields of septic systems.
One exception is a series of articles reporting on a compre-
hensive study of a home OWTS in Ontario, Canada, that pro-
vided information on the fate and transport of LAS. As part
of this work, LAS was shown to be effectively removed within
5 cm of soil below the drainfield [14–18]. It was also dem-
onstrated that the remaining LAS residues were removed in
the unsaturated zone soil so that LAS levels in the aquifer
On-site wastewater treatment systems (OWTS) serve ap-
proximately 25% of the U.S. population and about 37% of
new development. Over 25 million U.S. households use
OWTS, predominantly septic systems, for wastewater treat-
ment, making them one of the largest sources of effluent dis-
charge to groundwater [1].
As wastewater progresses through a conventional OWTS
and into the groundwater, the fate and transport of a given
wastewater constituent is determined largely by the charac-
teristics of the soil and groundwater in this subsurface envi-
ronment, including physical, chemical, and biological char-
acteristics [2]. Impacts of OWTS to ground and surface water
quality under certain conditions have been well documented
[3–10]. However, most OWTS investigations have focused on
constituents such as nitrogen, phosphorous, fecal bacteria, and
viruses. Less information is available in the literature regarding
the fate and transport of household cleaning and personal care
product ingredients discharged to OWTS.
were below the detection limit of 10 ␮g/L.
In order to better understand the fate of these surfactants
used in household cleaning and personal care products dis-
charged to OWTS, an extensive study of a septic system serv-
ing an individual home near Jacksonville (FL, USA) was con-
ducted. This home was selected because it reflected a situation
typical in Florida where subtropical weather conditions prevail
and septic systems exist in fine sandy soils close to the water
table. During the wettest times of the year, it is likely that
effluent from the septic system will pass directly into the
The most widely used surfactants in North America are
LAS, AE, and AES. Annual consumption in North America
during 2000 for LAS, AE, and AES was estimated at 400,000,
256,000, and 370,000 metric tons, respectively [11]. Details
* To whom correspondence may be addressed
(adecarvalho@sdahq.org).
2606

Fate of anionic and nonionic surfactants in sandy soil
Environ. Toxicol. Chem. 21, 2002
2607
Table 1. Surfactants monitored in the on-site wastewater treatment system effluent and groundwater
a Common values for commercial-grade materials.
b Approximately 20 to 50% of the mixture has
ϭ 0 and thus represents an alkyl sulfate component.
y
groundwater. The first phase of this study involved a detailed
characterization of the site, including the home, OWTS, site
geology, soils, and groundwater characteristics. This was fol-
lowed by an evaluation of the fate and transport of three major
surfactants found in the household wastewater discharged after
septic system treatment. Specific chemical measurements were
made for LAS, AE, and AES in septic tank effluents and
groundwater during the seasonal high and low groundwater
table levels.
fer. The top of the limestone of the Floridan aquifer is en-
countered at approximately 122 m below mean sea level.
Surface soils throughout the subdivision were derived from
sandy marine sediments. The site is underlain with Adamsville
fine sand [20]. The Adamsville series is classified as a some-
what poorly drained sandy soil with rapid permeability. The
water table is typically at 51 to 102 cm BGS for approximately
two to six months of the year and greater than 102 cm BGS
for the rest of the year. The typical Adamsville profile has a
fine sand texture throughout. Limitations for a conventional
septic system drainfield are classified as severe because of
wetness and poor filtration.
SITE DESCRIPTION
The residence utilized for this investigation was a three-
bedroom, two-bathroom home constructed in 1976. During this
study, the house had four residents consisting of two adults
and two children. The home had a dishwasher and washing
machine but no garbage disposal or water softener. The waste-
water from the home flowed into a septic tank with a capacity
of 3,400 L. The septic tank effluent (STE) flowed by gravity
through the infiltration system that covered 19.5 m2 and was
approximately 0.6 m below ground surface (BGS). This system
MATERIALS AND METHODS
Soils characterization
Subsurface characteristics of the site were determined from
the exploratory soil borings taken to a depth of 15 m BGS
during monitoring well and piezometer construction. Repre-
sentative soil samples were taken from the borings at various
depths with a split-spoon sampler. Particle size analyses were
conducted according to accepted methods [22].
consisted of three 0.6
ϫ 0.6-m gravel-filled trenches that con-
tained a 10-cm-diameter polyvinyl chloride perforated pipe.
The home was in a 200-unit subdivision that was located ap-
proximately 800 m east of the St. Johns River in St. Johns
County, south of Jacksonville. A plan view of the home’s
OWTS is shown in Figure 1.
The Jacksonville study site is located in the physiographic
area referred to as the Coastal Lowlands [19]. The topography
of the lowlands is controlled by a series of marine terraces.
Elevations at the site range from 3 to 5 m above mean sea
level. Surface drainage in the area is primarily through the St.
Johns River and its tributaries.
The climate is subtropical and is characterized by warm,
humid summers and mild, dry winters with occasional frost
between November and February. Average annual rainfall is
approximately 137 cm. Rainfall is seasonal, with the majority
falling during the months of June through September [20].
St. Johns County is underlain by two major geologic units
of differing lithologies [21]. The uppermost unit consists of
clastic sediments including poorly to moderately consolidated
sand, clay, and shell. This overlies a thick sequence of lime-
stone and dolomite, commonly referred to as the Floridan aqui-
Additional soil samples were collected in the immediate
vicinity of the drainfield to delineate various soil parameters
that affect contaminant transport. Soil core samples were col-
lected at three locations below and downgradient of the drain-
field using a stainless-steel soil recovery auger with polycar-
bonate sample liners. Soil core 1 (SC-1) was a composite
sample of unsaturated soil collected from locations along the
infiltrative surface of trench 1 (directly below the trench) ap-
proximately 71 to 91 cm BGS. Soil core 2 (SC-2) was a com-
posite sample from the same boreholes as SC-1, but the cores
were taken from the saturated zone below SC-1 from 122 to
142 cm BGS. Soil core 3 (SC-3) was a composite sample of
the soil approximately 4.6 m downgradient (southwest) of
trench 1 between TW-4 and TW-5 (see Fig. 1). The SC-3
sample was taken from the same relative elevation as SC-2
(122–142 cm BGS) and was also in the saturated zone at the
time of sampling. These samples were analyzed for bulk den-
sity, porosity, water content, and saturated hydraulic conduc-
tivity according to standard methods [22].

2608
Environ. Toxicol. Chem. 21, 2002
A.M. Nielsen et al.
Fig. 1. Diagram of septic tank site showing monitoring wells, sampling sites, and groundwater flow.
Groundwater hydrology
by side with a hollow-stem auger drill rig. One well was in-
stalled to a depth of 15 m and the other to a depth of 1.5 m
BGS. Differences in water elevation in these side-by-side wells
indicated vertical groundwater flow, and these data were used
to determine whether the vertical flow component was great
enough to cause the contaminant plume to be transported into
a deeper zone.
Water table elevation data were obtained to determine the
thickness of the unsaturated soil beneath the bottom of the
infiltration system. Monitoring well W-5 was installed between
the trenches of the OWTS infiltration system for this purpose.
A remote data-sensing water level indicator was installed ad-
jacent to W-5 to automatically record water table elevations
at 6-h intervals. A data-logging rain gauge was installed at the
site to measure and record precipitation events. This auto-
mation allowed for a more detailed evaluation of water table
This investigation included measurements of groundwater
elevation and flow direction. Subsequent aquifer testing was
used to identify groundwater monitoring points in downgra-
dient locations likely to be impacted by advective transport of
septic system effluent constituents.
Shallow groundwater monitoring wells were installed via
a hand auger. The wells consisted of a 5-cm-diameter polyvinyl
chloride pipe with a 60-cm slotted well screen attached at the
bottom. Monitoring wells were installed to a depth just below
the water table and were used to determine water table ele-
vations and the general direction of groundwater flow. Three
pairs of nested piezometer sets were also installed at the site
to determine if a vertical component of flow was present. Each
nested piezometer set consisted of two 5-cm-diameter poly-
vinyl chloride wells with 60-cm screen length, installed side

Fate of anionic and nonionic surfactants in sandy soil
Environ. Toxicol. Chem. 21, 2002
2609
fluctuations and unsaturated zone thickness below the OWTS
in response to rainfall.
water and is not typically present in septic tank effluent to any
great extent. Bromide was added by flushing 8 kg of sodium
bromide down the household toilet on May 9, 1995. Bromide
ion concentrations were then measured in samples from the
septic tank effluent and shallow groundwater wellpoint net-
work using a bromide ion-selective electrode in the field. This
information was used to calculate groundwater transport char-
acteristics at the site.
Slug testing was performed on several monitoring wells at
the site to obtain estimates of the saturated hydraulic conduc-
tivity in the shallow water table zone. The slug tests were
performed by rapidly removing a large slug of water from the
well casing and measuring the rate of rise of water back into
the well. Slug tests were performed in wells W-5, SW-7, and
SW-9 (see Fig. 1). Data from these tests were used to calculate
saturated hydraulic conductivity using the method of Bouwer
and Rich [23,24].
Septic tank effluent characterization
A septic tank effluent sampling basin was installed between
the septic tank and the infiltration system. Grab samples were
obtained by pumping STE from the basin through a dedicated
Teflon tubing line with a peristaltic pump. Twelve STE samples
were taken from August 1993 through May 1995. The STE
samples were preserved and analyzed according to standard
methods [26] for the following parameters: biochemical ox-
ygen demand (BOD), total organic carbon (TOC), MBAS,
chloride, ammonia nitrogen (NH3-N), nitrate nitrogen (NO3-
N), and total phosphorous (TP).
Nine STE samples taken from August 1993 through May
1995 were analyzed for LAS. Single grab samples taken in
November 1993 and May 1995 were analyzed for AE and
AES. Septic tank effluent quantity was estimated using water
meters installed at various locations on the home water system.
Interior water use was then used as an estimate of wastewater
flow to the septic tank.
Groundwater transport monitoring
In addition to the permanent monitoring wells described
here, a network of discrete sample points was installed down-
gradient of the drainfield using the direct push method [25].
Two types of stainless-steel miniature wellpoint probes were
used to obtain groundwater samples at the site. The first probe
unit utilized a push-pull wellpoint attached to Teflon
௢ tubing
(0.95-cm i.d.). The Teflon tubing was inserted through 1.27-
cm-i.d. stainless-steel tubing that was then pushed or driven
into the saturated zone to the desired sample depth. The
groundwater samples were obtained from this device by open-
ing the wellpoint with a slight upward pull and then pumping
samples to the surface through the Teflon tubing using a peri-
staltic pump. The entire probe system could then be pulled
from the ground and installed at a new sample location. This
system allowed collection of groundwater samples from 10 to
15 different locations in a day and was especially useful for
initial identification of the contaminant plume. The second
direct push probe was similar to the first, except that a min-
iature stainless-steel wellpoint attached to Teflon tubing was
left in the ground after pulling the probe system out.
Groundwater sampling procedures for surfactant analysis
A peristaltic pump with Teflon tubing was utilized to pump
groundwater samples from the wellpoints directly into amber
glass bottles. Pump tubing was replaced between each sam-
pling event. Each wellpoint had dedicated Teflon tubing at-
tached to the desired depths. Prior to filling the sample bottles,
several liters of groundwater were pumped from the wellpoints
to purge the system of stagnant water. Samples were preserved
with 8% (v/v) formalin (37% formaldehyde solution) and
shipped immediately to the analytical laboratories for specific
surfactant analysis. Samples were also prepared and assayed
for the same parameters as was done with the STE [26]. Two
sampling events for analysis of LAS, AE, and AES were per-
formed. The first was from November 3 to 5, 1993 (wet
season). Since November is part of the wet season, the un-
saturated soil thickness was very small (0.01 m) at the time
of sampling. Nineteen sampling locations with 52 wellpoints
were available for sampling. Fifty of these wellpoints were
sampled and analyzed for MBAS and LAS. Both AE and AES
analyses were done on 15 and 21 samples, respectively. The
selection of wellpoints to sample and analyze for the surfac-
tants was based on MBAS results from the January 1993 sam-
pling event. Results of MBAS were used because methylene
blue responds to LAS and AES in water samples. Furthermore,
it was assumed that AE would be in samples that showed LAS
and AES.
A system of 30 direct push, stainless-steel miniature well-
points was installed in January 1993 at 13 surface locations.
Sampling points TW-1 through TW-6 were installed at three
depths (1.8, 3.7, and 6.1 m BGS); TW-7 through TW-9 were
added with wellpoints at 1.8 and 6.1 m BGS (see Fig. 1). One
sample point of TW-10 through TW-12 at 1.8 m BGS and an
upgradient background well BW-1 had wellpoints at 1.5, 3.1
and 4.6 m BGS. Groundwater samples were taken from these
wellpoints in January 1993, preserved with 1% (v/v) formalin
Ϫ
(37% formaldehyde solution), and analyzed for chloride (Cl ),
methylene blue active substance (MBAS), and six other pa-
rameters according to standard methods [26].
Twenty-two additional miniature stainless-steel wellpoints
were installed in August 1993. These wellpoints included sam-
ple points TW-13 through TW-18 with wellpoints at 1.8, 3.7,
and 6.1 m BGS, wellpoints at 3.7 m BGS in TW-7 through
TW-9, and a wellpoint at 6.1 m BGS in BW-1. Thirty-three
more miniature stainless-steel wellpoints were added in Feb-
ruary 1995. These new sample points were TW-19 through
TW-25 with wellpoints at the typical 1.8, 3.7, and 6.1 m BGS
and TW-26 through TW-28 with wellpoints at 1.8, 3.7, 6.1,
and 9.1 m BGS. Figure 1 shows the sample locations at the
site after February 1995.
The second sampling event was from May 8 to 12, 1995
(dry season). Even though this sampling event was in the
middle of the dry season, the unsaturated soil thickness was
only 0.4 m. By this time period, the monitoring system had
been expanded to 29 sample locations with 85 permanent stain-
less-steel wellpoints at depths from 1.5 to 9.1 m BGS. Of these
wellpoints, 57 were analyzed for MBAS and LAS, 20 for AE,
and 21 for AES. The selection of these wellpoints was based
on the results from the November 1993 sampling event.
Bromide tracer testing
A groundwater tracer test was conducted to determine
groundwater velocities, verify groundwater flow direction,
quantify the effects of dilution, and obtain more accurate hy-
draulic conductivity values of the saturated soils. Bromide was
selected as the tracer because it is highly mobile in ground-

2610
Environ. Toxicol. Chem. 21, 2002
A.M. Nielsen et al.
LAS analysis
Results were reported as total AE, and the detection limit was
approximately 10
November 1993 were spiked in the field with 100
terated C13EO9 alcohol ethoxylate.
␮g/L. Selected groundwater samples from
Concentrations of commercial LAS (C10–C14) were deter-
mined in STE and groundwater samples using a procedure
similar to the method of Castles et al. [27]. The sample was
acidified and spiked with a C9 LAS homolog (prepared by
Sasol North America, Houston, TX, USA) as an internal stan-
dard. An aliquot of the sample was passed through a C2 solid-
phase extraction (SPE) cartridge (Varian, Harbor City, CA,
USA), and the cartridge was rinsed with a 70:30 (v/v) water:
methanol. Linear alkylbenzene sulfonate was eluted with meth-
anol and then passed through a strong anion exchange (SAX)
cartridge (Varian). The SAX cartridge was rinsed with 2%
acetic acid in methanol, which was followed by 5 ml of meth-
anol. The LAS was eluted with 1:1 (v/v) 2-N hydrochloric
acid (HCl) and methanol. The SAX cartridge eluate was evap-
orated to dryness, and a C15 LAS homolog (prepared by Sasol
North America) was added as an internal standard. The dried
eluate was reconstituted in a solution of 1:1 (v/v) water and
methanol and 0.25 mM sodium dodecyl sulfate. All solvents
were high-performance liquid chromatography grade, and the
water was Nanopure (Macalaster Bicknell, New Haven, CT,
USA [18 milliohms]).
␮g/L deu-
Septic tank effluent samples from August 1993 were pre-
pared and analyzed by a methodology similar to that described
in Evans et al. [28]. Sample preparation used a C8 SPE car-
tridge as described previously. Alcohol ethoxylate analysis and
quantitation were done by reverse-phase thermospray liquid
chromatography/mass spectroscopy (LC/MS) using deuterated
C13EO9 alcohol ethoxylate as an internal standard (injected
before MS analysis). Results were reported as total AE, and
the detection limit was approximately 10
samples were spiked with 10 mg/L Neodol
ethoxylate (Shell Chemical, Houston, TX, USA).
␮
g/L. Selected STE
25-9 alcohol
௡
Groundwater and septic tank effluent samples from the May
1995 sampling campaign were prepared and analyzed in a
manner similar to the methods of Evans et al. [30]. Samples
were prepared by passing them through a C8 SPE cartridge,
which was preconditioned with methanol, followed by iso-
propanol, and finally deionized water. Alcohol ethoxylate was
eluted with methanol followed by isopropanol (2:1, v/v). Sol-
vents were removed by evaporation, and the AE was resus-
pended in methanol. Alcohol ethoxylate analysis and quanti-
tation were done by electrospray LC/MS using deuterated
C13EO9 alcohol ethoxylate as an internal standard. Results were
reported as total AE, and the detection limit was approximately
Linear alkylbenzene sulfonate was analyzed by high-per-
formance liquid chromatography similar to the method of Cas-
tles et al. [27]. The method was isocratic, using a 45:55 (v/v)
tetrahydrofuran and 0.1-M NaClO4 (in water) eluent. Detection
was with ultraviolet fluorescence, which was optimized for
detection of LAS (225 nm for excitation and 290 nm for emis-
5
␮
g/L. In addition, groundwater samples from this sampling
event were spiked in the field (100 g/L) and in the laboratory
(10.3 g/L) with Neodol 25-9 alcohol ethoxylate.
␮
sion). The limit of quantitation was 10 ␮g/L.
␮
Septic tank effluent and groundwater samples were spiked
with a commercial-type LAS with an average alkyl chain
length of 11.4 carbons. Septic tank effluent samples were
spiked in the field with 1 to 30 mg/L and in the laboratory
with 3 and 30 mg/L LAS. Field spikes were done at 0.1 to
5.0 mg/L LAS in groundwater samples. Some spiking of
groundwater samples was done after the samples were taken
to the laboratory at 0.1 to 1.0 mg/L LAS. Spikes were done
in the field and the laboratory to determine the effects of
sample manipulations and storage on the efficiency of the an-
alytical method to detect LAS. Procedural blanks were in-
cluded in all analyses.
The results for all three sets of samples were reported as
total AE, which is a sum of all the ethoxymers (C12–15EO2–18).
AES analysis
The STE and groundwater samples for AES analysis were
first passed through a preconditioned C2 SPE column for sur-
factant isolation. Surfactants were eluted from the SPE car-
tridge with methanol:ethyl acetate (1:1, v/v) followed by meth-
anol. Separation of anionic and nonionic surfactants was per-
formed using a SAX column. After elution with 2.4 M HCl
in methanol, the anionic surfactants were dried, derivatized
with HBr (33% in glacial acetic acid) at 10ЊC for 4 h, cooled
to room temperature, resuspended in water, and extracted three
times with methylene chloride.
Alkyl bromide analysis was accomplished with a Hewlett-
Packard 5890 gas chromatograph (GC) equipped with a Hew-
lett-Packard 5971 mass selective detector (Palo Alto, CA,
USA). Chromatographic conditions were as follows: column—
HP-1 (dimethylpolysiloxane) or HP-5 (5% diphenyl, 95% di-
AE analysis
Alcohol ethoxylate isolation, concentration, and analyses
were done using three different procedures for groundwater
samples from the November 1993 sampling event, septic tank
effluent samples from the August 1993 sampling event, and
groundwater and septic tank effluent samples from the May
1995 sampling campaign. Different procedures were used be-
cause the AE analytical technology was improving during this
time period.
Sample preparation for the November 1993 sampling event
was similar to the methods of Evans et al. [28]. A C8 SPE
cartridge was used to isolate the AE. The AE samples were
passed through a preconditioned C8 column (methanol fol-
lowed by deionized water). Following rinses of 0.001 N HCl
(to remove excess formaldehyde), AE was eluted with meth-
anol and isopropyl alcohol (2:1, v/v). Solvents were removed
by evaporation, and AE was resuspended in methanol:water
(1:1, v/v). Alcohol ethoxylate analysis was done in a manner
similar to that described in Dorn et al. [29]. This method
separated and quantified the AE by high-performance liquid
chromatography with evaporative light scattering detection.
methylpolysiloxane); 30 m
ϫ 0.25 mm, 0.25-␮m film thickness
bonded-phase capillary column; carrier gas—helium; heat lin-
ear flow velocity of 40 cm/s; injection—splitless; injector tem-
perature—280
creased at 10
tection was done at 280
Њ
Њ
C; GC temperature—50
C/min to 280 C and held for 2 min. Mass de-
C.
ЊC for 1 min and in-
Њ
Њ
The total AES concentrations were calculated from the mas-
ses of alkyl bromides as measured by GC/MSD. The average
number of moles of EO for AES was assumed to be 1.7 as
determined by LC/MS during another study [31]. The method
did not distinguish between alkyl bromides from AES or alkyl
sulfates (AS). The quantitation limit was 10
water samples were spiked in the field with approximately 100
g/L of a deuterated C13 AS.
␮g/L. Ground-
␮

Fate of anionic and nonionic surfactants in sandy soil
Table 2. Summary of soil data from cores
Sample identification
Environ. Toxicol. Chem. 21, 2002
2611
top of the unsaturated zone was routinely within 0.3 m of the
infiltrative surface of the drainfield, and occasionally the mea-
sured groundwater levels were higher than the infiltrative sur-
face of the drainfield. For example, on October 17, 1994, the
water table was 0.15 m above the infiltrative surface.
Soil parameter
SC-1a
SC-2b
SC-3c
Bulk density (g/cm3)
Porosity (%)
Water content (% wt)
Saturated hydraulic con-
ductivity (cm/d)
1.43
46.10
23.60
1.54
41.90
26.00
1.53
42.30
30.10
The saturated hydraulic conductivity estimated from the slug
tests ranged from 212 cm/d at SW-9 to 282 cm/d at well W-5.
Bromide transport
324.00
410.00
514.00
Bromide began to exit the septic tank shortly after input;
a Soil core 1 (SC-1) was a composite sample of unsaturated soil di-
rectly below trench 1 (see the Materials and Methods section).
b Soil core 2 (SC-2) was a composite sample from boreholes of SC-
1 but taken from the saturated zone (see the Materials and Methods
section).
the first STE sample taken 30 min after input contained
Ͼ600
mg/L bromide. The peak bromide concentration (2,000 mg/L)
in the STE was measured approximately 54 h after input. Four
subsequent bromide samplings were conducted with the push-
pull sampler as well as the permanent sample points to de-
termine where the bromide plume was located. The bromide
infiltrated along the length of trench 1, but it moved more
quickly in the area farther from the septic tank. This would
likely be the case if a biomat had developed in the near-tank
area, as biomats have been shown to slow effluent infiltration
into unsaturated zones [32,33]. After 150 d, the bromide plume
had extended past TW-4, TW-5, and TW-6 and had moved
approximately 7 m downgradient of trench 1 (Fig. 3).
These bromide data were based on groundwater samples
taken 1.8 m BGS. Vertical profile samples showed that the
peak bromide concentrations were at 1.8 to 2.4 m BGS and
that the shallow sample data used to delineate the plume shown
in Figure 3 were the highest bromide concentrations.
c Soil core 3 (SC-3) was composite sample about 4.6 m downgradient
of trench 1 (see Materials and Methods section).
RESULTS AND DISCUSSION
Soil characteristics
Particle size analyses indicated that soil samples were of
fine sand texture. Approximately 85 to 97% by weight of the
soil samples obtained were composed of fine sand, while the
percentage of silts and clays (fines) ranged from 0.9 to 6.0%.
Percentages of medium sand were generally less than 5% with
the exception of soil samples obtained at approximately 8.2
to 9.8 m BGS that contained approximately 10% medium sand.
Physical characteristics of the soil from the site are presented
in Table 2.
Average groundwater seepage velocities were calculated
from the bromide groundwater data (Fig. 3). These velocities
were based on the distance of the highest bromide concentra-
tion downgradient from trench 1 and the time of travel. The
velocity calculations are very consistent and indicate a rela-
tively uniform average seepage velocity of approximately 0.04
m/d downgradient.
Groundwater hydrology
Water level measurements indicated that the groundwater
flow at the site was south-southwest toward the St. Johns River.
Figure 1 shows the location of piezometers, monitoring wells,
and relative groundwater elevation contours calculated from
data collected in February 1995. The horizontal groundwater
elevation gradient measured during the wet sampling period
ranged from 0.0029 m/m (November 5, 1993) to 0.0043 m/m
(February 20, 1995) and was 0.0016 m/m during the dry season
(June 29, 1995). A vertical downward gradient that ranged
from 0 to 0.008 m/m was observed at the southwest corner
piezometers, indicating a potential for horizontal and down-
ward groundwater flow at the site.
The bromide tracer data also provided the opportunity to
more accurately estimate the average, in situ hydraulic con-
ductivity of the shallow water table soils at this site. The travel
time of the bromide peak between two points downgradient
is related to the saturated hydraulic conductivity of the soil
through which it travels since bromide is an inert compound
that moves readily with groundwater flow [34]. The saturated
hydraulic conductivity was estimated to be 257 cm/d by ap-
plying the relationship between the seepage velocity and the
saturated hydraulic conductivity described in the Darcy equa-
tion. This value for the saturated hydraulic conductivity is
lower than those determined on soil cores in the laboratory
(Table 2); however, it agrees well with the saturated hydraulic
conductivity values determined with the slug test. The differ-
ences observed between the saturated hydraulic conductivity
values determined in the laboratory with core samples and the
field values are probably due to two factors: The act of ex-
tracting a core sample from the field disturbs the soil and
changes the soil’s physical characteristics, and the results from
the cores reflect a very small portion of the actual field situ-
ation. These results illustrate the importance of determining
transport parameters in the field prior to predicting or modeling
contaminant transport.
The water table would fall through the dry season and rise
during the rainy season (Fig. 2). For example, a 20-cm rainfall
event in 1995 caused a 0.6-m rise in the water table. The
unsaturated soil thickness ranged from 0 to 1.26 m BGS. The
Surfactant analysis
Results of surfactant analysis described earlier (LAS, AE,
and AES) were corrected on the basis of field spike recoveries.
Recoveries from field spikes of LAS into STE (1–30 mg/L)
Fig. 2. Relative groundwater elevations and 7-d cumulative precipi-
tation at study site. Note the rapid response of water table to precip-
itation.

2612
Environ. Toxicol. Chem. 21, 2002
A.M. Nielsen et al.
Fig. 3. Groundwater bromide concentrations (mg/L) following addition to the septic tank.
ranged from 80 to 103% with a mean of 87% and a standard
deviation of 8%. Two laboratory spikes of STE with LAS at
3 and 30 mg/L were recovered at 107 and 108%. Groundwater
samples (1993 and 1995) spiked in the field with LAS from
0.1 to 5.0 mg/L showed recoveries ranging from 40 to 170%
with a mean of 100% and a standard deviation of 33%. Lab-
oratory spikes of groundwater (1993 and 1995) with LAS from
0.1 to 1.0 mg/L had a range of recovery from 70 to 121%
with a mean of 94% and a standard deviation of 14%. The
greatest variability in the recoveries was noted with the 0.1-
mg/L spike.
of the soil beneath the drainfield [14–18]. The opportunities
for the STE in the Florida system to be exposed to unsaturated
soil are diminished. The septic system drainfield is placed in
sand with a high water table, which results in an unsaturated
soil treatment zone from 0 to 1.3 m. Many times, the STE
passes directly into the groundwater without exposure to an
unsaturated zone. In order to follow the fate of key STE com-
ponents, including major surfactants that are used in cleaning
products, it was necessary to first determine the concentrations
of these and other key components in the STE. Results from
these analyses are detailed in Table 3. Values for the parameters
measured were typically in the range of previously reported
STE measurements by Sherman et al. [35]. The concentrations
of BOD5, MBAS, and nitrogen were on the high end of the
range reported in the literature.
Average removal of key components after passing through
the infiltration surface was estimated by comparing concen-
trations in groundwater samples (taken just below the infiltra-
tion surface and 0.4 m unsaturated soil, locations TW-19, TW-
20, and TW-21) to the STE concentrations. The results from
May 1995 (dry season) are summarized in Table 4. The organic
load from STE as measured by TOC was about 91 mg/L with
the surfactants accounting for approximately 13% of this TOC.
The recoveries for the field spikes of 100 ␮g/L deuterated
C13EO9 alcohol ethoxylate into the groundwater samples taken
in November 1993 ranged from 62 to 100% with a mean of
95% and a standard deviation of 19%. Recoveries for field
spikes of Neodol 25-9 alcohol ethoxylate into groundwater
(1995) of 100
␮g/L ranged from 23 to 48% with a mean of
34% and a standard deviation of 10%. Neodol 25-9 alcohol
ethoxylate was recovered more efficiently from spikes of 10.3
␮
g/L into groundwater (1995) done in the laboratory with a
mean of 60% and a standard deviation of 4%. The recovery
of Neodol 25-9 alcohol ethoxylate from 10-mg/L spikes into
STE (1993) ranged from 98 to 172% with a mean of 135%
and a standard deviation of 37%. For AES analysis, field spikes
The surfactants showed a higher level of removal (Ͼ96%)
of 100
␮
g/L of a deuterated C13 AS into groundwater (1993
across the aerobic infiltration surface than the TOC (90%).
After exposure to the infiltration surface, the surfactants’ con-
and 1995) was recovered with a range of 58 to 130% with a
mean of 99% and a standard deviation of 26%.
tribution was reduced to
Ͻ4% of the STE organic load, which
would pass to the groundwater. Ammonium ion was converted
to nitrate as the septic tank effluent passed through the highly
active biomass at the infiltration surface. The large variability
in the ammonium-to-nitrate ratio in the closest downgradient
samples suggests that pockets of anaerobic conditions may
have been present in the immediate subsurface soil environ-
ment. Chloride and total phosphorus concentrations do not
appear to be significantly reduced by exposure to the biomass.
STE monitoring
The soil and groundwater conditions at this single home
septic system near Jacksonville represent a worst case for dem-
onstrating the treatment of the organic load (including surfac-
tants) coming from STE. In contrast to this system was the
septic system in Canada that always has an unsaturated soil
zone of at least 2 m and showed removal of LAS within 5 cm

Fate of anionic and nonionic surfactants in sandy soil
Table 3. Septic tank effluent quality; samples from August 1993 to May 1995
Range
Environ. Toxicol. Chem. 21, 2002
2613
Standard
deviation
Parameter
Mean
Na
Min
Max
Methylene blue active substances (MBAS)
(mg/L)
Linear alkylbenzene sulfonate (LAS) (mg/L)
Alcohol ether sulfate (AES) (mg/L)
Alcohol ethoxylate (AE) (mg/L)
Total organic carbon (TOC) (mg/L)
23.8
14.6
4.8
0.64
90.7
6.3
2.8
12
9
2
15
9.5
32
18.1
NAb
0.18
47.0
93
4.13
0.44
42
160
41
6.9
48
0.01
5.47
0.94
190
419
65
10.0
86
0.07
8
12
12
11
12
12
12
Biochemical oxygen demand (BOD5) (mg/L) 217
Ϫ
Chloride (Cl ) (mg/L)
54.6
8.4
62.0
0.03
8.0
1.1
10.0
0.02
Total phosphorous (TP) (mg/L)
Ammonia-nitrogen (NH3ϪN) (mg/L)
Nitrate-nitrogen (NOϪ3 ϪN) (mg/L)
a Number of samples analyzed.
b NA
ϭ not applicable.
Groundwater monitoring
Since the groundwater seepage velocity was 0.04 m/d and
the house had been in use since 1976, the water-soluble com-
ponents of the STE that were passed to the groundwater should
have traveled more than 260 m downgradient from the drain-
field. If retardation by sorption and removal due to biodeg-
radation were occurring, it would be expected that STE com-
ponents should be removed from the groundwater within 260
m. In order to test this prediction, it was necessary to determine
the concentrations of key STE components in the contaminant
plume downgradient of the drainfield.
Visual presentation of the LAS, MBAS, nitrogen species,
and total phosphorous groundwater data are shown as con-
centration contours in Figures 4 and 5. Figure 6 illustrates the
vertical and horizontal migration of LAS through the ground-
water by showing LAS concentrations along the centerline of
the plume. The location of the centerline is shown in Figure
Table 4. Concentrations of components (May 1995) just below
drainfield and average removals compared to septic tank effluent
(STE) concentrations (see Table 3)
Component concn. (mg/L)b
Range
Standard
Removalc
(%)
Componenta Mean deviation Nd
Min.
Max.
MBAS
LAS
AES
AE
0.76
0.55
0.85
0.82
3
3
3
2
3
3
3
3
3
0.06
0.1
NA
NA
7.5
23
3.3
0.3
0.01
1.7
1.5
NA
NA
12.0
46
12.0
61.0
9.4
96.8
Ͻ
Ͼ
Ͻ
Ͼ
Ͼ
Ͼ
96.2
99.8
99.1
90.1
41.0
0
NDe
ND
NAf
NA
TOC
9.0
32.0
8.8
31.8
3.2
2.6
12.1
4.8
30.4
5.4
Ϫ
Cl
TP
NHϩ4 Ϫ
NOϪ3 Ϫ
N
N
NA
NA
Ͻ
Ͼ
Ͻ
a For descriptions of abbreviations, see Table 3.
b Values of 1.8 m wellpoints from sampling locations. TW-19, TW-
20, and TW-21 that are the closest downgradient points (see Fig.
1).
c % Removal
(mean concn. in STE)
Ϫ (mean concn. in samples)
ϭ
ϫ 100.
Fig. 4. Contours of (A) linear alkylbenzene sulfonate (LAS) plume
(1.8 m below ground surface [BGS]) and the centerline of the plume
cross section and (B) contours of methylene blue active substance
(MBAS) plume (1.8 m BGS) from sampling in 1993 and 1995. Con-
centrations in mg/L.
mean concn. in STE
d Number of samples analyzed.
e ND
not detectable for AE (
of quantitation (10 g/L).
f NA
not applicable.
ϭ
Ͻ5␮g/L), for LAS and AES Ͻ limit
␮
ϭ

2614
Environ. Toxicol. Chem. 21, 2002
A.M. Nielsen et al.
Fig. 6. Vertical distribution of linear alkylbenzene sulfonate (LAS)
along the centerline of the plume (cross section) from sampling in
1993 (A) and 1995 (B).
4A. The MBAS and LAS concentration contours (Fig. 4) are
very similar. This is an expected result since LAS is a major
component of the soluble chemicals detected by MBAS. Total
organic carbon concentrations were reduced to the levels seen
in the background well (3–5 mg/L) in samples
gradient from the drainfield.
Ն22 m down-
The ammonia and nitrate contours (Fig. 5) indicate that the
concentration of dissolved oxygen in the groundwater varies
with time and space, which apparently resulted in anoxic con-
ditions in certain locations where denitrification may have oc-
curred. Consequently, variation in the ammonia-to-nitrate ratio
in groundwater samples was observed. These conditions may
also have reduced the rate of LAS removal. The total phos-
phorous concentration contours are very similar between the
high and low water table sampling events.
The septic tank drainfield system was extremely effective
at removing AE and AES. Even in the worst-case sampling
regime (November 1993), nondetectable levels (
AE were measured, and only 3 of 21 samples analyzed for
AES were greater than the quantitation level of 10 g/L. Sam-
ples at 1.7 and 4.7 m downgradient of the drainfield showed
18 and 13 g/L of AES, respectively. The other sample was
directly beneath the drainfield and had an AES concentration
of 13 g/L. During the dry season sampling event, no de-
tectable levels of AE or AES were found in any groundwater
samples ( 0.2 m downgradient of drainage field).
Ͻ5 ␮g/L) of
␮
Ϫ
Fig. 5. Contours of (A) NH3-N plume, (B) contours of NO 3-N
plume, and (C) contours of total phosphorous plume (1.8 m) from
1993 and 1995 sampling. Concentrations in mg/L. BGS
ground surface.
␮
ϭ below
␮
Ͻ
Linear alkylbenzene sulfonate was found to migrate in the
groundwater with detectable levels as far as 11.7 m horizon-

Fate of anionic and nonionic surfactants in sandy soil
Environ. Toxicol. Chem. 21, 2002
2615
5. Sikora LJ, Corey RB. 1976. Fate of nitrogen and phosphorus in
soils under septic tank waste disposal fields. Trans ASAE 19:866.
6. Viraraghavan T, Warnock RG. 1976. Groundwater pollution from
a septic tile field. Water Air Soil Pollut 5:281–287.
7. Cantor LW, Knox RC. 1985. Septic Tank Effects on Ground
Water Quality. Lewis, Boca Raton, FL, USA.
8. Harman J, Robertson WD, Cherry JA, Zanini L. 1996. Impacts
on a sand aquifer from an old septic system: Nitrate and phos-
phate. Ground Water 34:1105–1114.
9. Robertson WD, Cherry AJ, Sudicky EA. 1991. Groundwater con-
tamination from two small septic systems on sand aquifers.
Ground Water 29:82–92.
10. Walker WG, Bouma J, Keeny DR, Olcott PG. 1973. Nitrogen
transformations during subsurface disposal of septic tank effluent
in sands: II. Groundwater quality. J Environ Qual 2:521–525.
11. Colin A, Houston. 2000. Detergent Alkylates World Market,
1995–2010. Pound Ridge, NY, USA, pp IV-7.
12. Van de Plassche EJ, DeBruijn JHM, Stephenson RR, Marshall
SJ, Feijtel TCJ, Belanger SE. 1999. Predicted no-effect concen-
trations and risk characterization of four surfactants: Linear alkyl
benzene sulfonate, alcohol ethoxylates, alcohol ethoxylated sul-
fates, and soap. Environ Toxicol Chem 18:2653–2663.
13. McAvoy DC, Dyer SD, Fendinger NJ, Eckhoff WS, Lawrence
DL, Begley WM. 1998. Removal of alcohol ethoxylates, alkyl
ethoxylated sulfates, and linear alkylbenzene sulfonates in waste-
water treatment. Environ Toxicol Chem 17:1705–1711.
14. Robertson WD. 1994. Chemical fate and transport in a domestic
septic system: Site description and attenuation of dichloroben-
zene. Environ Toxicol Chem 13:183–191.
15. Wilhelm SR, Schiff SL, Robertson WD. 1994. Chemical fate and
transport in a domestic septic system: Unsaturated and saturated
zone geochemistry. Environ Toxicol Chem 13:193–203.
16. Shimp RJ, Lapsins EV, Ventullo RM. 1994. Chemical fate and
transport in a domestic septic system: Biodegradation of linear
alkylbenzene sulfonate (LAS) and nitrilotriacetic acid (NTA). En-
viron Toxicol Chem 13:205–212.
17. McAvoy DC, White CE, Moore BL, Rapaport RA. 1994. Chem-
ical fate and transport in a domestic septic system: Sorption and
transport of anionic and cationic surfactants. Environ Toxicol
Chem 13:213–221.
tally (Fig. 4) and 3.7 m vertically (Fig. 6) from the drainfield
during the wet season. As reflected in Figure 4, the LAS plume
was smaller during the dry season. The concentrations of LAS
in the plume dramatically decreased during the dry season,
when more unsaturated soil was available for rapid removal
of the surfactant. These data suggest that LAS is being re-
moved from the groundwater environment because it should
have migrated some 260 m downgradient of the drainfield by
1995 as predicted by the seepage velocity if no removal mech-
anisms were present. The most reasonable removal mecha-
nisms for LAS are biodegradation and adsorption. Research
by Doi et al. [36] demonstrated that the subsurface soils at
this site have the potential for adsorbing and mineralizing LAS.
Also, chloride monitoring data show that dilution of surfactants
by groundwater in the main portion of the plume was insig-
nificant. For example, during the January 1993 sampling, the
mean chloride concentration at the 1.7-m-BGS depth in down-
gradient sample locations TW-1 through TW-12 (see Fig. 1)
was 58 mg/L, which compares to a mean of about 55 mg/L
in the STE.
CONCLUSIONS
The soil and groundwater conditions at a single home septic
system near Jacksonville represent a worst case for demon-
strating the treatment of septic tank effluent. Despite these
adverse circumstances, the septic treatment system (aerobic
infiltration surface/unsaturated soil and saturated soil) effec-
tively removes STE components. Most of the treatment occurs
at the infiltration surface since about 90% of the STE organic
load (TOC),
Ͼ96% of LAS, and Ͼ99% of AE and AES were
removed when a 0.4-m unsaturated zone existed. The remain-
ing 10% of the STE total organic carbon was reduced to back-
ground concentrations after moving 22 m downgradient of the
drainfield. Under the worst conditions, when very little (0.01
m) or no unsaturated treatment zone is present, LAS and AES
surfactant residues were detected in the groundwater; AE was
not detected. Alcohol ether sulfate was not detected beyond
4.7 m horizontally and 1.8 m vertically downgradient from the
drainfield. Linear alkylbenzene sulfonate was detected only up
to 11.7 m horizontally and 3.7 m vertically from the drainfield
during the wet season. Since it was possible for the surfactants
to have migrated as much as 260 m if they had moved un-
inhibited with the wastewater plume, it is clear that removal
was occurring. The most likely removal mechanisms for these
surfactants are biodegradation and sorption.
18. Shutter SB, Sudicky EA, Robertson WD. 1994. Chemical fate
and transport in a domestic septic system: Application of a var-
iably saturated model for chemical movement. Environ Toxicol
Chem 13:223–231.
19. Fernald EA, Purdum ED. 1992. Atlas of Florida. University Press
of Florida, Gainesville, FL, USA.
20. U.S. Department of Agriculture. 1983. Soil survey of St. Johns
County, Florida. Soil Conservation Service, Washington, DC.
21. Parker GG, Forguson GE. 1955. Water resources of southeastern
Florida. Water Supply Paper 1255. U.S. Geological Survey, Res-
ton, VA.
22. Black CA, ed. 1965. Methods of Soil Analysis, Part 1—Physical
and Mineralogical Properties, Including Statistics of Measure-
ment and Sampling. Agronomy 9. American Society of Agron-
omy, Madison, WI.
23. Bouwer H, Rice RC. 1976. A slug test for determining hydraulic
conductivity of unconfined aquifers with completely or partially
penetrating wells. Water Resour Res 12:423–428.
Acknowledgement—The authors would like to thank H.O. Perkins
and C.E. Beall of Sasol North America, S.T. Dubey of Shell Chemical
Company, W.E. Begley and N.J. Fendinger of the Procter & Gamble
Company, and S. Jackson of Ayres & Associates.
24. Bouwer H. 1989. The Bouwer and Rice slug test—An update.
Ground Water 27:304–309.
25. U.S. Environmental Protection Agency. 1998. Field sampling and
analysis technologies matrix and reference guide. EPA 542-B-
98-002. Washington, DC.
REFERENCES
26. American Public Health Association, American Water Works As-
sociation, Water Environment Federation. 1999. Standard Meth-
ods for the Examination of Water and Wastewater, 20th ed.
American Public Health Association, Washington, DC.
27. Castles MA, Moore BL, Ward SR. 1989. Measurement of linear
alkylbenzene sulfonates in aqueous environmental matrices by
liquid chromatography with fluorescence detection. Anal Chem
61:2534–2540.
28. Evans KA, Dubey ST, Kravetz L, Dridic I, Gumulka J, Mueller
R, Stork JR. 1994. Quantitative determination of linear primary
alcohol ethoxylate surfactants in environmental samples by ther-
mospray LC/MS. Anal Chem 66:699–705.
1. U.S. Environmental Protection Agency. 1997. Response to Con-
gress on use of decentralized wastewater treatment systems. EPA
832-R-97-001b. Office of Water, Washington, DC.
2. Wilhelm SR, Schiff SL, Cherry JA. 1994. Biogeochemical evo-
lution of domestic wastewater in septic systems: 1. Conceptual
model. Ground Water 32:905–916.
3. Jones RA, Lee GF. 1979. Septic tank wastewater disposal systems
as phosphorus sources for surface waters. J Water Pollut Control
Fed 51:2764–2775.
4. Peavy HS, Groves KS. 1978. The influence of septic tank drain-
fields on groundwater quality in areas of high groundwater. Pro-
ceedings, 2nd National Home Sewage Treatment Symposium,
American Society of Agricultural Engineers, Chicago, IL, De-
cember 12–19, 1977, pp 218–225.
29. Dorn PB, Rodgers JH Jr, Dubey ST, Gillespie WB Jr, Lizotte RE
Jr. 1997. An assessment of the ecological effects of a C9–11

2616
Environ. Toxicol. Chem. 21, 2002
A.M. Nielsen et al.
linear alcohol ethoxylate surfactant in stream mesocosm experi-
ments. Ecotoxicology 6:275–292.
30. Evans KA, Dubey ST, Kravetz L, Evetts SW, Dzidic I, Dooyema
CC. 1997. Quantitation of alcohol ethoxylate surfactants in en-
vironmental samples by electrospray mass spectrometry. J Am
Oil Chem Soc 74:765–773.
31. Popenoe DD, Moris SJ III, Horn PS, Norwood KT. 1994. De-
termination of alkyl sulfates and alkyl ethoxysulfates in waste-
water treatment plant influents and effluents and in river water
using liquid chromatography/ion spray mass spectrometry. Anal
Chem 66:1620–1629.
32. Otis RJ. 1985. Soil clogging: Mechanisms and control. Proceed-
ings, 4th National Symposium on Individual and Small Com-
munity Sewage Systems, American Society of Agricultural En-
gineers, New Orleans, LA, December 10–11, 1984, pp 238–250.
33. Siegrist RL. 1987. Soil clogging during subsurface wastewater
infiltration as affected by effluent composition and loading rate.
J Environ Qual 16:181–187.
34. Davis SN, Thompson GM, Bentley HW, Stiles G. 1980. Ground-
water tracers—A short review. Ground Water 18:14–23.
35. Sherman KM, Anderson DL. 1991. An evaluation of volatile
organic compounds and conventional parameters from onsite sew-
age disposal systems in Florida. Proceedings, 6th National Sym-
posium on Individual and Small Community Sewage Systems,
American Society of Agricultural Engineers, Chicago, IL, De-
cember 16–17, pp 62–75.
36. Doi J, Marks KH, DeCarvalho AJ, McAvoy DC, Nielsen AM,
Kravetz L, Cano ML. 2002. Investigation of an onsite wastewater
treatment system in sandy soil: Sorption and biodegradation of
linear alkylbenzene sulfonate. Environ Toxicol Chem 21:2617–
2622.