phytoremediation of polychlorinated biphenyl-contaminated...
TRANSCRIPT
www.elsevier.com/locate/envint
Environment International 30 (2004) 799–804
Phytoremediation of polychlorinated biphenyl-contaminated soils:
the rhizosphere effect
Tesema Chekola,*, Lester R. Voughb, Rufus L. Chaneyc
aBattelle Memorial Institute, 1204 Technology Drive, Aberdeen, MD 21001, USAbDepartment of Natural Resource Sciences, University of Maryland, College Park, MD 20742, USA
cUSDA-ARS, Animal Manure and By-Products Laboratory, Beltsville, MD 20705, USA
Received 20 October 2003; accepted 28 January 2004
Abstract
The objective in the first phase of this study was to screen alfalfa, flatpea, sericea lespedeza, deertongue, reed canarygrass, switchgrass,
and tall fescue for phytoremediation of polychlorinated biphenyl (PCB)-contaminated soil. During the second phase, the focus was
rhizosphere characterization to optimize PCB phytoremediation. Aroclor 1248 (PCB) was added to soil at 100 mg kg� 1 of soil. In the first
phase, all of the plant species treatments showed significantly greater PCB biodegradation compared to the unplanted controls and the two
most effective species were selected for further study. During the rhizosphere characterization study, soil irradiation did not affect PCB
biodegradation, but planting significantly increased PCB biodegradation; 38% or less of the initial PCB was recovered from planted pots,
compared to more than 82% from the unplanted control soils. Presence of plants significantly increased the biological activity (microbial
counts and enzyme activity) of both irradiated and unirradiated soils. Greater bacterial counts and soil enzyme activity were closely related to
higher levels of PCB biodegradation. The data showed that Aroclor 1248 biodegradation in soil seem to be positively influenced by the
presence of plants and plant–bacteria interactions. Our results suggested that phytoremediation could be an environmentally friendly
alternative for PCB-contaminated soils.
D 2004 Elsevier Ltd. All rights reserved.
Keywords: PCB; Soil contamination; Phytoremediation
1. Introduction
Contamination of soils and waters with polychlorinated
biphenyls (PCBs) has often resulted from the manufacture,
handling, use, and disposal of these chemicals. Moreover,
their extreme persistence in the environment and ability to
bioconcentrate in the food chain makes them great environ-
mental and human health risks that need remedial action
(Cousins et al., 1998; Hickey, 1999). Due to the fact that
engineering-based remedial technologies are expensive and
disruptive, there is a growing interest in developing new
remediation technologies that are environment friendly and
less expensive (Cunningham et al., 1996). Phytoremedia-
tion, use of plants for remediation, is one such highly
appealing technology (Schnoor et al., 1995; Wenzel et al.,
1999).
0160-4120/$ - see front matter D 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.envint.2004.01.008
* Corresponding author. Tel.: +1-410-306-8641; fax: +1-410-306-
8422.
E-mail address: [email protected] (T. Chekol).
Past plant–PCB interaction experiments were done main-
ly for the purpose of looking into the food chain transfer
(Webber et al., 1990; Gan and Berthouex, 1994). The
possibility of PCB uptake and translocation by corn (Zea
mays L.), cabbage (Brassica oleracea var. capitata L.) and
carrot (Daucus carota L.) from contaminated sewage sludge
was explored and carrots had the highest PCB concentration
in the plant tissues followed by cabbage and corn (Webber et
al., 1990). However, these concentrations, with the exception
of peel layer of carrots, were very small and not related to the
soil PCB contamination levels. Other reports have also
confirmed that the risk of PCB translocation into the corn
grain or stover to be negligible (Gan and Berthouex, 1994).
Use of ectomycorrhizal fungi has shown some promise in
PCB remediation, particularly the lower chlorinated con-
geners (Donnelly and Flecher, 1995). In this mutually
beneficial system of plant–fungus interaction, PCB degrad-
ing ectomycorrhizal fungi were better adapted and this was
one crucial step in overcoming the competitive disadvan-
tages encountered in other introduced degraders.
T. Chekol et al. / Environment International 30 (2004) 799–804800
Actual PCB phytoremediation studies are very limited
and the reported cases refer mainly to plant tissue culture or
soil microcosm studies (Mackova et al., 1997; Epuri and
Sorensen, 1997). For example, hairy root cultures of Sola-
num nigrum L. were reported to cause 40% mineralization
of commercial PCB mixtures after 30 days of incubation
(Mackova et al., 1997). Enhanced mineralization of Aroclor
1260 was also reported in planted soil microcosm compared
to that in unplanted control (Epuri and Sorensen, 1997).
Available literature is very sketchy and phytoremediation
of PCB-contaminated soils is a largely untouched area of
research. This study was conducted to determine if some
forage and conservation crops could be used for phytor-
emediation of PCB-contaminated soils. The specific objec-
tives were to screen several legume and grass species for
their effectiveness in enhancing soil Aroclor 1248 biodeg-
radation and to determine rhizosphere characteristics, in-
cluding enzymes and microbial population and composition,
associated with contaminant biodegradation.
2. Materials and methods
Three legumes and four grass species were screened in the
first phase. The legumes were alfalfa (Medicago sativa L.),
flatpea (Wagner pea), (Lathyrus sylvestris L.) and sericea
lespedeza (Lespedeza cuneata Dum.-Cours.). Grass species
were deertongue (Panicum clandestinum L.), reed canary-
grass (Phalaris arundinacea L.), switchgrass (Panicum vir-
gatum L.), and tall fescue (Festuca arundinacea Schreb.).
The two sets of experiments were conducted at the
Department of Natural Resource Sciences growth chambers
facility. The growth chambers were Model BDR8 (Con-
viron, Winnipeg, Canada). The first phase of the trials
involved crop species screening, while in the second phase,
plant species that performed best in the first phase were
further evaluated for their rhizosphere characteristics.
Growth chamber temperatures were maintained at 25/16
jC, and the light regime was 16/8 h day/night cycle, with
photosynthetic photon flux rate of 400–500 Amol m� 2 s� 1
from metal halide bulbs. The relative humidity in the
chambers was set at 65F 5%.
The soil for these studies was obtained from a forested
area with no cropping and pesticide application history at
the Central Maryland Research and Education Center-
Clarksville Facility. The soil is classified as Hatboro
(course-loamy, mixed, nonacid, mesic, Typic Fluvaquent).
Soil testing results showed pH (1:1 water) 6.1; OM content
2.6% and CEC 5.72 cmol/kg.
Soil for the irradiation experiment in the second phase of
the study was irradiated at the University of Maryland
Radiation Engineering Facilities. The irradiation was done
at 25 krad/h for 100 h at 25 jC with a total of 2.5 mrad g-
irradiation from a Co60 source. The irradiated soil was used
to assess the impact of plants on contaminant biodegrada-
tion in the absence of microorganisms.
Both the analytical standard and pure chemical Aroclor
1248 were obtained from ChemService, West Chester, PA.
The targeted contaminant concentration of 100 mg/kg of soil
was used for this experiment. Right after spiking, the soil was
left under exhaust hood for 3 days with periodic mixing to
allow the solvent (hexane) to evaporate, subsequently cov-
ered and allowed to age at room temperature for 4 weeks
before transfer to pots and seeding. There were three repli-
cates of plant species treatment and unplanted control pots.
The unplanted control was used to assess contaminant
biodegradation in the absence of plant rhizosphere.
Constant soil moisture (f 28%) was maintained through
periodic refilling of reservoir of the Moisture Replacement
System (MRS) developed by Sardanelli and Kenworthy
(1997). This system consisted of plant pots with wick
inserts, a holding rack to support the pots, and an enclosed
bottom water reservoir with a disposable plastic lining.
During the second phase, the MRS was connected to a
permanent water source and refilling of the reservoirs was
regulated with a programmable One-Station DC controller
(The Drip Store, Valley Center, CA). Deionized water
flowing into the reservoirs in the irradiated chambers was
filtered using a microbial filter, 0.2 AM Super cap capsules
(VWR Scientific, Bridgeport, NJ).
The experimental design for the crop species screening
study was split-plot in a Randomized Complete Block
Design (RCBD). The chemical treatments in each box were
used as main plots, whereas the crop species were the sub-
plot treatments. Split–split-plot in a Randomized Complete
Block Design (RCBD) was employed for the second phase
study. Soil irradiation was the main plot level treatment and
chemicals and crop species were the sub and sub–sub-plot
treatments, respectively. Data was analyzed using Proc
MIXED and multiple mean comparisons were done with
Tukey–Kramer test (SAS, 2001).
At the end of each experiment, composite soil samples
were placed in QorpakR glass containers and stored in the
freezer (� 50 jC) until extraction. For bacterial and fungal
enumeration, subsamples were placed in WhirlpakR bags
and stored at 4 jC, until the microbial assays were per-
formed. In addition, plant biomass was separated into shoot
and root mass. Any plant part above the soil surface was
considered as the shoot weight and conversely, the biomass
below the soil surface was included in the root weight.
Soil–root mass was taken out of the pots, roots were
separated from the soil by shaking, and the soil was sieved
to remove any root fragments. Both shoot and root weights
are reported on a dry weight basis.
Soil PCB extractions and analysis were done using meth-
ods described by Lopez-Avila et al. (1995). Extraction was
done using a Microwave Extraction Unit (CEM, Mathews,
NC). Ten grams oven-dry equivalent weight (ODE) of soil
were quantitatively transferred into the Teflon-lined extrac-
tion vessels and 40 ml of 1:9 acetone/hexane was added to
each vessel. The extraction was performed at 110 jC for 10
min at 100% power (1000 W). The vessels were allowed to
Table 2
Aroclor 1248 levels (mg/kg) in irradiated and unirradiated soils after 4
months of plant growth
Treatment Soil PCB
(mg/kg)
Irradiated soil
Unplanted control 86c*
Reed canarygrass 31a
Switchgrass 39b
Unirradiated soil
Unplanted control 82c
Reed canarygrass 23a
Switchgrass 28a
*Means followed by the same letter are not significantly different as
determined by Tukey–Kramer multiple comparison procedure ( P < 0.05).
T. Chekol et al. / Environment International 30 (2004) 799–804 801
cool to room temperature before opening and 10 ml of the
extract was transferred into 40 ml hexane. Samples were
filtered into amber vials through 0.2 Am filter, the first 2–3ml
of filtrate was discarded and the remainder was retained for
analysis.
Gas chromatographic analysis of Aroclor 1248 was per-
formed using a Hewlett-Packard 5890 Series II gas chro-
matograph (Hewlett-Packard, Palo Alto, CA) equipped with
an electron capture detector (ECD). A J&WMegaboreR DB
608 GC column (30 m� 0.53 mm id) was used (J&W
Scientific). The column was temperature programmed at 80
jC as an initial (0.1 min), followed with two ramps at 15 jC/min to 180 jC and then at 4 jC/min to 300 jC. Injector anddetector temperatures were maintained at 270 and 330 jC,respectively. The column was operated with a helium carrier
gas at a flow rate of 1.3 ml/min. The detector makeup and
anode purge gases consisted of 5% methane and 95% argon.
The run time was set at 37 min.
Total Aroclor 1248 levels were quantified using area
counts of 10 major peaks of a 0.8 mg/kg Aroclor 1248
standard. Quality assurance–quality control checks were
conducted with each analysis. These included use of stand-
ards with known concentrations, material blank measure-
ments, repeat analysis, and fresh spike recoveries.
Microbial enumeration was accomplished using a mod-
ified Harris and Sommers (1968) plate dilution frequency
technique.
Rhizosphere dehydrogenase activity was assayed using a
modified Dick et al. (1996) method. Twenty grams (ODE)
of rhizosphere soil was thoroughly mixed with 0.2 g CaCO3
and three replicate samples of 6 g soil were placed in three
test tubes. In each tube, 1 ml of a 3% 2,3,5-triphenyltetra-
zolium chloride (TTC) and 2.5 ml distilled water were
added. Samples were then incubated while shaking at
medium speed for 24 h at 37 jC. After 24 h, 10 ml methanol
was added to each tube and samples were vortexed. The soil
suspension was filtered through a glass funnel plugged with
absorbent cotton. The filtrate was diluted with methanol to a
100 ml volume and intensity of the reddish color was
Table 1
Residual Aroclor 1248 levels (mg/kg) in soil after 4 months of plant growth
in PCB-amended soil
Treatment PCB
(mg/kg)
Legumes
Alfalfa 23a*
Flatpea 28ab
Sericea lespedeza 29ab
Grasses
Deertongue 28ab
Reed canarygrass 27a
Switchgrass 31ab
Tall fescue 33b
Control (no crop) 82c
*Means followed by the same letter are not significantly different as
determined by Tukey–Kramer multiple comparison procedure ( P < 0.05).
measured at 485 nm using a spectrophotometer. Dehydro-
genase activities in the samples were calculated by using
calibration graphs prepared from 500, 1000, 1500 and 2000
mg triphenyl formazan (TPF)/100 ml standards. Results are
presented as mg TPF/kg soil.
3. Results and discussion
3.1. Soil Aroclor 1248 levels
At the end of the 4-month species screening experiment,
all the planted treatments had significantly lower levels of
PCB compared to the unplanted control pots (Table 1).
Although no complete congener specific analysis was con-
ducted for this study, analysis on the major component of
Aroclor 1248, tetrachlorbiphenyl, showed that this congener
accounted for about 38% of the total the PCB in soil. This
shows that the composition of the final Aroclor 1248 was
very close to the initial (40% tetrachlorbiphnyl). It should be
noted, however, quantitation of individual congeners is
more accurate than estimating Aroclors and there is a need
for congener-specific measurement in future PCB phytor-
emediation studies.
This study is in complete agreement with the findings
of Epuri and Sorensen (1997) regarding the use of tall
fescue in PCB phytoremediation. However, due to the
Table 3
LS mean root and shoot weights (g) in control and PCB-amended soil
Species Root weight (g) Shoot weight (g)
Control 100 mg/kg Control 100 mg/kg
Alfalfa 25a* 8b 24a 5b
Flatpea 53a 33b 45a 44a
Sericea lespedeza 37a 19b 54a 50a
Deertongue 140a 128a 37a 28a
Reed canarygrass 145a 140a 64a 53a
Switchgrass 90a 89a 65a 66a
Tall fescue 139a 127a 68a 63a
*Means within variable and rows followed by the same letter are not
significantly different as determined by Tukey–Kramer test ( P< 0.05).
Table 6
Bacterial and fungal counts (MPN/g) in PCB-amended soils
Treatment Bacterial counts (� 107) Fungal counts (� 104)
Initial End Initial End
Irradiated soil
Unplanted control 0f* 3.9d 0e 0.9d
Reed canarygrass 0f 4.6c 0e 2.5c
Switchgrass 0f 4.5c 0e 2.9b
Unirradiated soil
Unplanted control 2.3e 4.9c 1.5d 3.1b
Reed canarygrass 2.1e 6.2a 1.2d 4.4a
Switchgrass 2.5e 5.6b 1.6d 4.5a
*Means within variable (bacteria and fungi) followed by the same letter
are not significantly different as determined by Tukey–Kramer multiple
comparison procedure ( P < 0.05).
Table 4
Root/shoot ratios of crop species in the control and PCB-amended soil
Species Root/shoot ratio
Control 100 mg/kg
Alfalfa 1.04 1.68
Flatpea 1.18 0.75
Sericea lespedeza 0.68 0.37
Deertongue 3.73 4.57
Reed canarygrass 2.27 2.64
Switchgrass 1.38 1.34
Tall fescue 2.04 2.03
T. Chekol et al. / Environment International 30 (2004) 799–804802
more comprehensive nature of this study, tall fescue,
although better than the unplanted controls, did not per-
form as well compared to other species. As a result, this
species was not considered for further detailed rhizosphere
evaluation. Based on the results from the screening study,
reed canarygrass and switchgrass were further evaluated
for their phytoremediation abilities under irradiated and
unirradiated soil conditions.
Results from the rhizosphere effects study showed that soil
irradiation did not significantly affect PCB biodegradations
(Table 2). However, planting significantly improved the
biodegradation of PCB in both irradiated and unirradiated
soils. More than 82% of the initial PCB was recovered from
the irradiated and unirradiated unplanted control soils com-
pared to less than 39% in the planted treatments. After 4
months of plant growth in the irradiated soil, planting with
reed canarygrass and switchgrass resulted in a 70% and 61%,
respectively, biodegradation of the initial PCB levels. In the
unirradiated soils, only 23% and 28%, respectively, of the
initial PCB levels were recovered from reed canarygrass and
switchgrass planted pots compared to 82% in the unplanted
control pots.
These results clearly showed that planted treatments had
significantly lower levels of PCB compared to unplanted
controls and reed canarygrass and switchgrass were the
most effective species for phytoremediation of PCB-con-
taminated soils. Moreover, statistically significant differ-
ences observed in soil Aroclor 1248 levels between
Table 5
Root and shoot weights (g) of reed canarygrass and switchgrass grown in
control and PCB-amended soils
Treatment Root weight (g) Shoot weight (g)
Control PCB Control PCB
Reed canarygrass
(unirradiated soil)
157a* 136a 61a 50a
Reed canarygrass
(irradiated soil)
156a 133b 57a 48b
Switchgrass
(unirradiated soil)
87a 86a 67a 62a
Switchgrass
(irradiated soil)
86a 66b 60a 57b
*Means within variable and rows followed by the same letter are not
significantly different as determined by Tukey–Kramer test ( P< 0.05).
planted and unplanted soils and lack thereof between the
planted irradiated and unirradiated soils suggested that
planting was the most important factor for a successful
phytoremediation of PCB in soil.
3.2. Plant species biomass
Root weights of all legume species and root and shoot
weights of alfalfa were significantly reduced because of
soil PCB contamination (Table 3). PCB spiking of the soil
did not significantly affect the biomasses of all grass
species and shoot weights of flatpea and sericea lespede-
za. This effect could be a result of greater sensitivity to
PCB of the root system and/or beneficial microorganisms
in the legume rhizospheres compared to those in the grass
species.
Root/shoot ratios of the plant species under investigation
changed as a result of PCB amendment of the soil (Table 4).
There was a slight decrease in the root/shoot ratios of most
of the plant species used in this experiment. The increase in
the root/shoot ratios of alfalfa was attributable to a more
pronounced negative effect of PCB on the shoot growth of
this crop.
Soil irradiation significantly reduced the root weights of
reed canarygrass and switchgrass in PCB spiked soil, but
did not affect shoot weights (Table 5). The decrease in the
root weights observed in the irradiated soil could be due to a
Table 7
Correlations (R2) of microbial counts and plant biomass to Aroclor 1248
biodegradation
Variables Bacterial
counts
Fungal
counts
Shoot
weight
Root
weight
Bacterial counts
Fungal counts 0.68*
Shoot weight 0.63* 0.34
Root weight 0.72* 0.45 0.87**
Soil PCB levels � 0.67* � 0.40 � 0.94*** � 0.95***
*Significance at P < 0.05.
**Significance at P < 0.01.
***Significance at P < 0.0001.
Table 8
Soil dehydrogenase activity and Aroclor 1248 levels
Treatment Soil dehydrogenase
activity (mg/kg/24 h)
Soil PCB
(mg/kg)
Irradiated soil
Unamended no crop 33f* 0
Unamended reed canary 386c 0
Unamended switchgrass 256d 0
Amended no crop 23f 86c
Amended reed canary 368c 21a
Amended switchgrass 299cd 39b
Unirradiated soil
Unamended no crop 218d 0
Unamended reed canary 604a 0
Unamended switchgrass 471b 0
Amended no crop 130e 82c
Amended reed canary 601a 23a
Amended switchgrass 450b 28a
*Means within column followed by the same letter are not significantly
different as determined by Tukey–Kramer multiple comparison procedure
( P< 0.05).
T. Chekol et al. / Environment International 30 (2004) 799–804 803
combined negative effect of soil irradiation and PCB on the
rhizosphere of reed canarygrass and switchgrass. Soil spik-
ing with PCB alone did not significantly affect the root and
shoot weights of these grasses.
3.3. Microbial assays
Microbial data at the end of experiment (Table 6)
showed that bacterial and fungal counts in the irradiated
soil were slightly increased compared to the initial counts.
The increase in both counts could be a result of prolif-
eration in the previously nonculturable organisms stimu-
lated by the favorable environment in the growth
chambers and/or some level of contamination during the
growing period.
Planting significantly increased the microbial numbers in
both the irradiated and unirradiated soil.
Plant shoots and root weights were significantly
correlated to Aroclor 1248 biodegradations in the soil
(Table 7). Bacterial counts were also significantly corre-
lated to the PCB biodegradation, plant biomass and fungal
counts.
These observations indicated that plant biomass was the
most important factor for Aroclor 1248 biodegradations in
the soil, hence selection of plant species and cultivars for
PCB phytoremediation. In addition, the positive correlation
observed between plant biomass and the number of bac-
teria in soil could be indicative of the proliferation of
bacteria stimulated by greater supply of nutrients from a
vigorous plant. These correlations could be an indication
for the biologically mediated biodegradations of PCB in
soil.
3.4. Rhizosphere dehydrogenase activity
Results presented in Table 8 showed that PCB amend-
ment and irradiation caused a reduction in the dehydroge-
nase activity of soils. It should be noted that dehydrogenase
activity of the soil was significantly reduced due to
irradiation. However, with the exception of unplanted
control treatment in the unirradiated soil, the reductions
associated with soil PCB amendments were not statistically
significant.
In this study, the presence of plants significantly in-
creased the dehydrogenase activity of both irradiated and
unirradiated soils. Moreover, higher soil dehydrogenase
activities significantly correlated with higher levels of
PCB biodegradation in both the irradiated and unirradiated
soils. Based on these observations, it could be concluded
that planting increased the biological activity of soils
measured as dehydrogenase activity. The increased biolog-
ical activity appeared to be responsible for the significantly
higher levels of PCB biodegradation in the planted pots
compared to unplanted controls.
Results from these experiments showed that use of forage
crops for phytoremediation of PCB-contaminated soils was
an effective and environmentally friendly means of remedi-
ation. To the best knowledge of the authors, this study was
the first of its kind to show that plants in general, and forage
crops in particular, are very effective for phytoremediation
of PCB-contaminated soils.
References
Cousins IT, Mclachlan MS, Jones KC. Lack of an aging effect on the soil –
air partitioning of polychlorinated biphenyls. Environ Sci Technol
1998;32:2734–40.
Cunningham SD, Anderson TA, Schwab P, Hsu FC. Phytoremediation
of soils contaminated with organic pollutants. Adv Agron 1996;56:
55–114.
Dick RP, Breakwell DP, Turco RF. Soil enzyme activities and biodiver-
sity measurements as integrative microbiological indicators. In: Doran
JW, Jones AJ, editors. Methods of soil analysis: Part 2. Microbio-
logical and biochemical properties. Soil Science Society of America.
Madison, WI: Special Publication #49, 1996. p. 247–72.
Donnelly PK, Flecher JS. PCB metabolism by ectomycorrhizal fungi. Bull
Environ Toxicol 1995;54:507–13.
Epuri V, Sorensen DL. Benzo(a)pyrene and hexachlorobiphenyl con-
taminated soil: phytoremediation potential. In: Kruger EL, Ander-
son TA, Coats JR, editors. Phytoremediation of soil and water
contaminants. Washington, DC: American Chemical Society; 1997.
p. 200–22.
Gan R, Berthouex P. Disappearance and crop uptake of PCBs from sludge-
amended farming. Water Environ Res 1994;66:54–69.
Harris RF, Sommers LE. Plate-dilution frequency techniques for assay of
microbial ecology. Appl Microbiol 1968;16:333–4.
Hickey WJ. Transformation and fate of polychlorinated biphenyls in soil
and sediment. In: Adriano DC, et al., editors. Bioremediation of con-
taminated soils. Agron Monogr, vol. 37. Madison. WI: ASA, CSSA and
SSSA; 1999. p. 213–37.
Lopez-Avila V, Benedicto J, Charan C, Young R. Determination of PCBs in
soils/sediments by microwave-assisted extraction and GC/ECD or
ELISA. J Environ Sci Technol 1995;29:2709–12.
Mackova M, Macek T, Kucerova T, Burkhard P, Pazlarova J, Demnerova
K. Degradation of polychlorinated biphenyls by hair root culture of
Solanum nigrum. Biotechnol Lett 1997;19:787–90.
T. Chekol et al. / Environment International 30 (2004) 799–804804
Sardanelli S, Kenworthy WJ. Soil moisture control and direct seeding
for bioassay of Heterodera glycines on soybean. J Nematol 1997;29:
625–34 [Suppl.].
SAS. Systems Release, vol. 8. Cary, NC: SAS Institute; 2001.
Schnoor JL, Licht LA, McCutcheon SC, Wolfe NL, Carriera LH. Phytor-
emediation of organic and nutrient contaminants. Environ Sci Technol
1995;29:318–23.
Webber MD, Pietz RI, Granato TC, Svoboda ML. Plant uptake of PCBs
and other organic contaminants from sludge-treated coal refuse. J En-
viron Qual 1990;23:1019–26.
Wenzel WW, Adriano DC, Salt D, Smith R. Phytoremediation: a plant–
microbe-based remediation system. In: Adriano DC, et al., editors. Bio-
remediation of Contaminated Soils. Agron Monogr, vol. 37. Madison
WI: ASA, CSSA and SSSA; 1999. p. 457–508.