<%@ LANGUAGE="VBSCRIPT" %>
Use of humic acids to enhance the removal of aromatic hydrocarbons from contaminated aquifers.
DOWNLOAD PAPER
|
Proceedings of the 5th annual Symposium on Groundwater and Soil Remediation, Oct 2-6, 1995, Toronto Ontario. Proceedings on CD/ROM.
Use of humic acids to enhance the removal of aromatic hydrocarbons from contaminated aquifers.
Part II: Pilot Scale
Suzanne Lesage, Hao Xu, Kent S. Novakowski, Susan Brown and Louise Durham
Groundwater Remediation Project, National Water Research Institute, Environment Canada, Burlington, Ontario, L7R 4A6
|
ABSTRACT
Contamination of groundwater by gasoline and
diesel fuels is a widespread environmental problem in Canada.
Polynuclear aromatic hydrocarbons (PAHs) in the gasoline and diesel
fuels are of particular concern because many are carcinogenic
and degrade poorly. Aromatic hydrocarbons can represent up to
60 % of diesel fuels (Block, et al., 1991). Groundwater and soils
contaminated with petroleum products must therefore be remediated.
A technology commonly used to clean up groundwater
contaminated by petroleum hydrocarbons involves pumping the water
to the surface and then treating it. The low solubility of petroleum
hydrocarbons in water and unfavourable hydrogeological conditions
limit the efficiency of this pump-and-treat approach. Surfactants
have been used to enhance the solubilization process, however,
they can interfere with bioremediation taking place at the site,
or prove toxic to aquatic organisms in areas where the groundwater
discharges. Humic acids, found in soil and groundwater, have
a surface activity similar to that of surfactants. This study
was aimed at determining whether humic acids could be used cost-effectively
to enhance dissolution and transport of aromatic hydrocarbons
and provide an environmentally suitable alternative to artificial
surfactants for cleaning up soil and groundwater contaminated
by petroleum fuels
In order to test new technologies at near field
scale, a physical aquifer model was constructed and installed
in the Aquatic Ecosystem Restoration Evaluation Facility (AQUEREF)
at the National Water Research Institute in Burlington, Ontario.
The large-scale model aquifer (a rectangular stainless steel box,
2.4 m x 6m x 2m deep) was designed with a separate head tank to
provide uniform one-dimensional flow. Multi-level monitoring wells,
consisting of 72 bundles of stainless steel tubes, terminating
at 5 different depths provide a total of 360 monitoring points.
Petroleum was introduced as a saturated sand
slurry and the flow system was reestablished with water only for
five weeks, after which the experimental treatment was initiated.
The introduction of humic acid caused a rapid ten-fold increase
in the aqueous concentration of the PAHs found in diesel, which
was sustained for several weeks. The next experiment will explore
whether this increase in solubility translates into an increase
in the bioavailability of the PAHs in petroleum contaminated
soils.
INTRODUCTION
Contamination of groundwater by gasoline and
diesel fuels is a widespread environmental problem in Canada.
Polynuclear aromatic hydrocarbons (PAHs) in the gasoline and diesel
fuels are of particular concern because many are carcinogenic
and degrade poorly. Aromatic hydrocarbons can represent up to
60 % of diesel fuels (Block, et al., 1991). Groundwater and soils
contaminated with petroleum products must therefore be remediated.
A technology commonly used to clean up groundwater
contaminated by petroleum hydrocarbons involves pumping the water
to the surface and then treating it. The low solubility of petroleum
hydrocarbons in water and unfavourable hydrogeological conditions
limit the efficiency of this pump-and-treat approach. Surfactants
have been used to enhance the solubilization process, however,
they can interfere with bioremediation taking place at the site,
or prove toxic to aquatic organisms in areas where the groundwater
discharges. In this project, the enhancement of the aqueous solubility
of the aromatic hydrocarbons by humic acids was studied in laboratory
columns (Xu et al., 1994) and in a large scale model aquifer (this
article). Humic acids, found in soil and groundwater, have a
surface activity similar to that of surfactants. This study was
aimed at determining whether humic acids could be used cost-effectively
to enhance dissolution and transport of aromatic hydrocarbons
and provide an environmentally suitable alternative to artificial
surfactants for cleaning up soil and groundwater contaminated
by petroleum products.
The effects of humic substances on the solubility
and mobility of organic contaminants have been the subject of
numerous studies. The apparent aqueous solubility of chlordane,
DDT, PCBs and chlorodioxins has been observed to increase in the
presence of humic substances (Chiou et al., 1987; Johnson-Logan
et al., 1992; Webster et al., 1986). This is because the binding
of a particular PAH compound by humic substances depends not only
on the hydrophobicity of the PAH solute but also on the size of
the solute molecule and its ability to fit into hydrophobic cavities
in humic substances. An increase in pH and ionic strength was
found to decrease the binding of PAHs by humic substances (Schlautman
and Morgan, 1993). Hydrophobic organic contaminants (PAHs, benzene,
carbon tetrachloride and DDT) can also be adsorbed by dissolved
humic substances (Carter and Suffet, 1982; Gauthler et al., 1987;
Herbert et al., 1993; McCarthy and Jimenez; 1985; Rutherford et
al., 1992). The corresponding Koc values was also
found to be strongly correlated with the degree of aromaticity
in the humic substances (Gauthler et al., 1987). The mobility
of PAHs, PCBs, and chlordane in groundwater systems was observed
to have increased in the presence of humic substances (Backhus
and Gschwend, 1990; Dunnivant et al., 1992; Johnson et al., 1993;
Johnson-Logan et al., 1992). Backhus and Gschwend (1990), concluded
that the presence of humic substances (<40 mg C/L) will double
the mobile load of hydrophobic pollutants such as benzo[a]pyrene
or perylene, but will have little effect on the mobility of less
hydrophobic pollutants. However, all these studies were conducted
with pure standards and no information was available on the effects
of humic substances on the solubility and mobility of the aromatic
compounds released from a petroleum NAPL phase in groundwater.
In the first phase of the study, it was found that humic acids
concentrations as low as 0.1 g/L could enhance the solubilization
of aromatics from diesel fuel. While pH and ionic strength did
affect the solubility, the concentration of humic acid was the
most important factor governing solubility.
Preliminary studies were conducted using silica
sand. For the system to be used in hte field, sorption effects
on soil need to be considered. The interactions between humic
acids and soil have been studied (Xu, 1991). The partitioning
of humic acids in soil/water phases is dependent on the pHzpc
( pH at zero point of charge) of the soil. Above the pHzpc
a large proportion of humic acids will reside in the aqueous phase.
Since most of the geological materials have pHzpc below
7-8, humic acids should reside primarily in the aqueous phase.
By altering the soil pH, the binding of the humic substance to
the mineral surface can be reduced and the contaminants mobilized.
Field observations of the mobility of naturally occurring substances
in groundwater are scarce. One experiment by McCarthy et al. (1993)
describes the injection and withdrawal of naturally occurring
brown water, high in organic content. A fraction of the compounds
was extremely mobile whereas some tailing and long desorption
times were observed for the larger fractions. This study did not
involve organic contaminants.
One of the reasons that few field-scale experiments
have been reported is that environmental authorities are increasingly
reluctant to allow controlled spills followed by a treatment that
is yet to be proven effective. In order to test new technologies
at near field scale, a model aquifer with a very dense monitoring
network was constructed, providing controlled conditions only
possible in a semi-artificial system. The physical aquifer model
was designed and installed at the National Water Research Institute
of Environment Canada in Burlington in the newly created Aquatic
Ecosystem Restoration Evaluation Facility.
METHODS
Tank Design
Some model aquifers are in fact large scale
columns where vertical transport of contaminants are measured
(FIAT, NHRI, Environment Canada, Saskatoon). Others are rectangular
containers used to isolate contaminated soil and re-create aquifers
(VEGAS, University of Stuttgart, Germany and US EPA R.S. Kerr
Laboratory in ADA Oklahoma) or are cells created in-situ in the
field by sheet piling walls (CFB Borden by the Waterloo Centre
for Groundwater Research). For this study, it was decided that
a rectangular tank to measure horizontal transport be constructed.
Figure 1 Model Aquifer
The tank (Figure 1) was constructed entirely of 1/4" industrial grade stainless
steel (rectangular 2m x 6m x 2m deep) with an external support structure made
of steel beams. While most physical aquifer models rely on a combination of
injection/withdrawal wells to induce water flow, in this case it was decided
to provide a head tank separated from the aquifer material by a porous plate.
The plate was also constructed of 1/4" stainless steel, but perforated
with 1" holes so as not to impede water flow. The aquifer material (sand)
was retained by a polyester geotextile, a material compatible with most organic
solvents. The purpose of using a head tank instead of injection wells was to
provide an even flow field across the entire width and depth of the tank, a
condition more closely resembling natural groundwater flow systems. A walkway
was attached to the perimeter of the tank so as to provide for a working area
and prevent uneven compaction by walking on top of the sand.
Instrumentation
In order to provide an experimental control,
the aquifer model was divided in half longitudinally using a series
of stainless steel plates which were sealed and bolted together.
The bottom of the plates were secured with bentonite clay before
the tank was filled with sand. The monitoring well network consists
of 72 bundles of 1/8" stainless steel tubes. Each bundle
consists of 5 sampling tubes terminating at 30 cm depth intervals.
Each tube is completed with a stainless steel frit. The bundles
were spaced at 30 cm and 25 cm centres, parallel and perpendicular
to the flow direction, respectiviely. In addition, two withdrawal
wells were installed, one on each side of the tank (5 cm i.d.,
10 cm o.d. with their own packing and screened over the entire
length). A medium- to coarse-grained mixed sand was obtained from
a local sand and gravel supplier (particle size range 75 :m
to 2.4 mm; hydraulic conductivity of approximately 0.04 m/s).
The model was filled with dry sand, using a moving conveyor belt.
In order to prevent layering, the sand was mixed manually around
the monitoring tubes, using edge cutters. The sand was flushed
for 24 hours with helium to prevent any air entrapment. The aquifer
material was saturated from the head-tank end over a period of
one week using tap water. The tap water was continually purged
with helium to remove dissolved air and residual chlorinated compounds.
The sand was allowed to settle for a few weeks prior to conducting
the tracer experiments.
Tracer Experiment
The details of the tracer test have been reported
(Lesage et al., 1995) and are therefore only summarized here.
A conservative tracer experiment was conducted by introducing,
simultaneously, a mixture of Lissamine (2 mg/L) and sodium
bromide (100mg/L) into the head tank (total volume 3120 L). Over
the course of the following two weeks, over 5000 samples were
collected and analyzed for Lissamine (by fluorescence) and for
bromide (by conductivity).
Figure 2
Estimates of the longitudinal dispersivity
and porosity of the sand were obtained by interpreting breakthrough
concentrations of bromide from selected sampling points. This
was accomplished using a one-dimensional transport model (Novakowski,
1992) which accounts for an exponentially decaying source in the
head tank and assumes the aquifer to be of semi-infinite extent.
Modelled fits to the concentration data were obtained by adjusting
the independent parameters, longitudinal dispersion and porosity,
and determining the average linear velocity from the pumping rate
and specified porosity.
Emplacement of the Petroleum Source
The results of the conservative tracer experiment
were used to determine the optimal placement of the petroleum
source. Because the main goal of this study is to investigate
methods for the cleanup of petroleum present at residual saturation
in the groundwater zone, it was necessary to devise a method of
emplacement that would provide a well defined source in terms
of both mass and geometry. Thus, it was determined to emplace
a rectangular source centered at depth of 1.2 m, approximately
0.5 m down-gradient from the head tank. Five additional monitoring
well bundles were added along the center-line. The residual capacity
of the sand was determined using a column experiment and it was
found that approximately 500 mL of diesel fuel could be retained
by 20 kg of sand.
To emplace the source, the water table was
lowered to a depth of 1.60 m. In order to prevent disruption of
the sand in the area of the emplaced source, a wood frame (20 cm
X 40 cm X 2 m) was lowered into the tank while the sand inside
the box was removed by vacuum. The excavated sand was temporarily
placed on a clean tarp. A 25kg portion was weighed and 500 mL
of diesel (ESSO Petroleum, local supplier) was added and mixed.
The mixture was placed back into the bottom of the hole and covered
with a layer of clean sand while the wooden form was pulled upwards
using an overhead pulley. The water table was gradually reestablished
by refilling the head tank while pumping at the withdrawal wells.
This insured that flow proceeded in the desired direction.
Addition of Humic Acid
Humic acid was obtained as the sodium salt
(Aldrich Chemicals, Milwaukee, WI) and prepared as a solution
using tap water at a concentration of 1g/L. The resulting pH was
8.5. Initially a concentrate was added to the head tank to the
same concentration such as to provide a well-defined concentration
gradient. The effluent was discharged to the sewer (through a
charcoal filter for the treatment side) until a constant concentration
of humic acid was obtained at the withdrawal wells. The effluent
was then collected and recirculated. A humic acid concentrate
was added through a metering pump to make up for losses on the
charcoal filter.
Analysis
Monitoring of transport in the model aquifer
required a very large (>1000) number of analyses. It was initially
assumed that TPH analysis with an Infra-Red (IR) detector would
be a suitable surrogate parameter and that selected samples would
be analyzed by HPLC for PAHs. Preliminary tests showed that because
TPH requires the extraction of the sample in a solvent such as
CFC-113, it would be more labour intensive than direct injection
into the HPLC. Also, the fluorescence detector is orders of magnitude
more sensitive than the IR detector commonly used for TPH. This
allows the analyses to be conducted without sample preconcentration.
Analysis of the full range of PAHs typically requires solvent
programming and approximately 45 minutes per sample. Because diesel
fuel contains mostly methyl naphthalenes and phenanthrenes, these
compounds can be grouped together using a much shorter column
and an isocratic system. Using this method, an analysis can be
performed in only 15 minutes.
RESULTS
Tracer Experiment
Lissamine is often used as a tracer in hydrogeology
because it is a highly fluorescent compound which can be measured
rapidly over a wide range of concentrations. It is not entirely
conservative and will adsorb to organic carbon, but its behaviour
was expected to be representative of the behaviour of humic acid.
By comparison of the arrival time for Lissamine and bromide tracers,
a retardation factor of 1.3 was measured for lissamine. Knowing
that the aquifer material has less than 0.1% or organic carbon,
this retardation is most likely due to ionic interactions.
Interpretation of the tracer experiment yielded
estimates of the transport properties of the aquifer material.
Modelling of the concentration profiles for bromide at four sampling
locations indicated a longitudinal dispersivity of 0.005-0.015
m, a porosity of 28-32% and an average linear-velocity of 0.020-0.022
m/hr. The porosity values determined from the modelling results
are consistent with independent determinations. Also the values
for the longitudinal dispersivity suggest that the influence of
macroscopic dispersion is negligible for this material.
As can be seen in a vertical contour (adjacent
to the center line) of the concentration of Lissamine at 310 hours
(Figure 2), the groundwater velocity was slower in the upper part
of the tank. This effect is probably attributable to the presence
of the capillary fringe and the free-surface at the head tank
end. The results were used to direct the placement of the contaminated
source. Although more difficult to achieve, it was determined
that the source should be placed at 1.2 m, so that a flow
field of even velocity would be experienced by the source.
Figure 3 PAH plume - Cross section
Diesel Plume
During the emplacement of the source, there
was concern that some residual diesel fuel might float as an undissolved
phase as the water table rose. To determine whether this occurred,
the wells in the vadose zone were monitored for BTEX (benzene,
toluene, ethyl benzene and xylenes) using headspace analysis and
for PAHs by HPLC. BTEX and PAHs were found only in level four
monitoring points, none above or below, clearly indicating that
NAPL movement did not occur (Figure 3). A similarly-shaped plume
of BTEX was also found moving ahead of the PAHs. This is further
evidence that the aromatic hydrocarbons were only transported
in the dissolved phase. Had the compounds been transported as
NAPL droplets, the volatile hydrocarbons and PAHs would not have
been separated spatially with time.
Figure 4 Degradation of BTEX in space and time.
 |
There was also evidence of biodegradation of
the volatile hydrocarbons. In spite of all the efforts to remove
oxygen by sparging the incoming water with nitrogen and the fact
that no nutrients were added, the concentration of volatiles decreased
in the aqueous samples. These losses could not be attributed to
sorption, because, as can be seen in Figure 4, the more water
soluble compounds disappeared the fastest, or volatilization,
because the samples were all from the saturated zone at a depth
of 1.2 m. |
Effect of Humic Acid
Water was pumped through for a period of seven
weeks prior to initiating the humic acid treatment. A total PAH
concentration of approximately 500 :g/L
was observed in the monitoring wells just beyond the source (Figure
5). An average three-fold increase in PAHs was observed when
humic acid was added. A ten-fold increase in solubilization was
observed for trimethyl naphthalene. This is consistent with what
had been observed in our column studies and is generally the case
with surfactants, where the least soluble compounds benefit the
most of the addition of a solubilization agent.
These results were virtually identical to those
observed in the laboratory for the enhanced dissolution of the
residual phase (Xu et al., 1994). In small columns where the
amount of oil exceeded the holding capacity of the solid phase,
a two-stage removal of the hydrocarbons had been observed: one
where the oil was removed as droplets followed by a slower phase
where dissolution was the predominant mechanism for removal. In
the larger model, because the emplaced diesel source was designed
such as not to exceed the holding capacity of the sand, dissolution
only was observed. Unlike what is generally observed with surfactants
(Thangamani and Shreve, 1994), the increase in solubility occurred
below the critical micelle concentration for humic acids, (7.4 g/L
for Aldrich humic acid, Guetzloff and Rice, 1994).
The sand used in the tank had a very low carbon
content thus any interaction between the sand and the humic acid
were expected to be ionic. Initially the humic acid had a pH of
8.5, but this was buffered down to 7.5 by the contact with the
sand. At that pH sorption was not expected to be significant and
indeed most of the humic acid was recuperated at the withdrawal
wells. This allowed for recirculation of the treatment in a closed
loop system, resulting in significant cost savings. The effluent
was collected in a reservoir which was assayed for humic acid
concentration and replenished as necessary by the addition of
a concentrate through a metering pump.
 |
The evolution of the PAH plume with time is
depicted in Figure 6. The plume labeled 0 days was just before
the addition of humic acid, 51 days after the emplacement of the
source. As was shown above, a much higher concentration of PAHs
was achieved after the addition of PAHs. The depression on the
longitudinal direction of the contour is attributable to the lower
permeability in the source zone. From the tracer test, it was
determined that one pore volume would take ten days to travel
through the tank. The last contour on Figure 6 represents 9 pore
volumes. It is therefore evident that in spite of the fact that
humic acids do enhance dissolution significantly, rapid cleanup
will not occur using dissolution alone. The next step will therefore
be to study the potential of bioremediation under these conditions. |
CONCLUSION
A large scale aquifer model to study groundwater
remediation technologies was successfully designed and constructed.
The transport characteristics of the model aquifer were measured
using a tracer experiment. A residual diesel source was emplaced
and dissolved slowly such as would occur in a typical pump and
treat scenario. The addition of humic acids was found to increase
the concentration of PAHs dissolving from diesel as much as ten-fold.
Future plans involve the addition of oxygen and nutrients to promote
bioremediation and the use of a heavier fuel containing more recalcitrant
PAHs.
ACKNOWLEDGEMENT
The authors wish to thank GASReP, Environment
Canada and the National Energy Board for their financial support.
The advice of Dr. Carl Enfield of the U.S. EPA R.S. Kerr Laboratory,
Ada OK, on the design and packing of the model aquifer and on
installing the monitoring well network, is very much appreciated.
REFERENCES
Backhus, D.A. and P.M. Gschwend, 1990. Fluorescent
polycyclic aromatic hydrocarbons as probes for studying the impact
of colloids on pollutant transport in groundwater. Environ.
Sci. Technol., 24, 1214-1223.
Block, R.N., N. Allworth and M. Bishop, 1991.
Assessment of diesel contamination in soil. In (E.J. Calabrese
and P.T. Kostecki, ed.) Hydrocarbon Contaminated Soils, Volume
1, Lewis Publishers, Inc., pp. 135.
Carter, C.W. and I.H. Suffet, 1982. Binding
of DDT to dissolved humic materials. Environ. Sci. Technol.,
16, 735-740.
Chiou, C.T., Kile, D.E., Brinton, T.I., Malcolm,
R.L., Leenheer, J.A. and MacCarthy P., 1987. A comparison of water
solubility enhancements of organic solutes by aquatic humic materials
and commercial humic acids. Environ. Sci. Technol., 21,
1231-1234.
Dunnivant, F.M., P.M. Jardine, D.L. Taylor
and J.F. McCarthy, 1992. Cotransport of cadmium and hexachlorobiphenyl
by dissolved organic carbon through columns containing aquifer
material. Environ. Sci. Technol., 26, 360-368.
Gauthler, T.D., W.R. Seltz and C.L. Grant,
1987. Effects of structural and compositional variations of dissolved
humic materials on pyrene Koc Values. Environ. Sci. Technol.,
21, 243-248.
Guetzloff, T.F. and J.A. Rice, 1994. Does
humic acid form a micelle? The Science of the Total Environment,
152, 31-35.
Herbert, B.E., P.M. Bertsch and J.M.Novak,
1993. Pyrene sorption by water-soluble organic carbon. Environ.
Sci. Technol., 27, 398-403.
Johnson, W.P., H. Liu and G.L. Amy, 1993. Natural
organic matter (NOM) interactions with polynuclear aromatic hydrocarons
(PAH); facilitated transport and enhanced desorption under saturated
groundwater conditions. Presented before the Division of Environmental
Chemistry, American Chemical Society, Denver, CO, March 28 - April
2.
Johnson-Logan, L.R., Broshears, R.E. and Klaine,
S.J., 1992. Partitioning behavior and the mobility of chlordane
in groundwater. Environ. Sci. Technol., 26, 22344-2239.
Lesage, S., K.S. Novakowski, H. Xu, G. Bickerton,
L. Durham and S. Brown, 1995. A large scale aquifer model to
study the removal of aromatic hydrocarbons from the saturated
zone. Solutions'95, International Association of Hydrogeologists
Congress, June 4-10, Edmonton Alberta, 6 pages (Electronic Proceedings,
hardcopy indexed alphabetically by author).
McCarthy, J.F, L. Liang, P.M. Jardine and T.M.
Williams. 1993. Mobility of Natural Organic Matter Injected into
a Sandy Aquifer. In Manipulation of Groundwater Colloids for Environmental
Restoration. J.F. McCarthy and F.J. Wobber. Lewis Publishers,
Boca Raton FLA. pages 35-39.
McCarthy, J.F. and B.D. Jimenez, 1985. Interactions
between polycyclic aromatic hydrocarbons and dissolved humic material:
binding and dissociation. Environ. Sci. Technol., 19,
1072-1076.
Novakowski, K. S., 1992. An evaluation of boundary
conditions for one-dimensional solute transport. 1. Mathematical
development., Water Resour. Res. 28(9), 2399-2410.
Rouse, J.D., D. A. Sabatini, J. M. Suflita
and J. H. Harwell, 1994. Influence of Surfactants onMicrobial
Degradation of Organic Compounds. Critical Reviews on Microbial
Degradation of Organic Compounds, 24, 325-370.
Rutherford, D.W., C.T. Chiou and D.E. Kile,
1992. Influence of soil organic matter composition on the partition
of organic compounds. Environ. Sci. Technol., 26,
336-340.
Schlautman, M.A. and J.J. Morgan, 1993. Effects
of aqueous chemistry on the binding of polycyclic aromatic hydrocarbons
by dissolved humic materials. Environ. Sci. Technol., 27,
961-969.
Shinozuka, N., C. Lee and S. Hayano. 1987.
Solubilizing Action of Humic Acid from Marine Sediments.The
Science of the Total Environment . 62, 311-314.
Thangamiani, S. and G.S. Shreve. 1994. Effect
of Anionic Biosurfactant on Hexadecane Partitioning in Multiphase
Systems. Environ. Sci. Technol. 28, 1993-2000.
Webster, G.R.B., D.H. Muldrew, J.J. Graham,
L.P. Sarna, D.C.G. Muir, 1986. Dissolved organic matter mediated
aquatic transport of chlorinated dioxins. Chemosphere,
15, 9-12.
Xu, H., S. Lesage and L. Durham, 1994. The
use of humic acids to enhance removal of aromatic hydrocarbons
from contaminated aquifers. Part I: Laboratory studies. Fourth
Annual Symposium on Groundwater and Soil Remediation, Calgary
Alberta. September 21-23. Pages 635-666.
Xu, H., 1991. Effects of Humic Substances and
pH on the Speciation and Adsorption of Cadmium, Mercury and Arsenic.
Ph. D. Diss., Link ping University, Sweden, ISBN 91-7870-783-8.22