Plant Physiology and Biochemistry 45 (2007) 379e387

www.elsevier.com/locate/plaphy

Research article

Removal of trace metals by Sorghum bicolor and Helianthus annuus

in a site polluted by industrial wastes: A field experience
L. Marchiol a,*, G. Fellet a, D. Perosa b, G. Zerbi a
a
Dipartimento di Scienze Agrarie e Ambientali, Università degli Studi di Udine, Via delle Scienze 208, I-33100 Udine, Italy
b
Caffaro srl, piazzale Marinotti 1, I-33050 Torviscosa, Italy
Available online 15 March 2007

Abstract

Using the perspective of full scale application of phytoremediation techniques, research is focusing on the optimization of agronomic prac-
tices. Two annual high biomass yield crops, Sorghum bicolor and Helianthus annuus, were grown in a polymetallic soil. The experimental site,
polluted by pyrite cinders, is located in an industrial site that has been listed in the clean-up national priority list since 2001. Specific aims of this
work were to observe the concentration of metals in plants during the crop cycle and to establish the amount of metal removed by the crops. The
field trial, arranged in a randomized block design, started in 2005. The concentrations of heavy metals in the soil were: As 309, Cd 4.29, Co 50.9,
Cu 1527 and Zn 980 mg kg1. The crops grown on the polluted soil received mineral fertilization (Fert) and organic amendment (Org), while
plants in control soil (Ctrl) did not receive anything. The plots were watered during the crop cycle during two drought periods, using a sprinkler
irrigation system. The phytoextraction potential of crops was estimated during the whole growth cycle and the plant biomass that was collected
in each sampling date was ICP-analyzed. Plantebiomass growth curves were obtained. The concentrations of the metals in the shoots and in the
total plant biomass were recorded. Finally, the metal removal was calculated for the harvestable parts of the crops. The amelioration of the nu-
tritive status of the substrate that resulted, was highly effective for the biomass yield. However, fertilization and soil amendment did not heighten
the concentration of metals in the harvestable tissue of the plants during the crop cycle. In some cases, organic matter appeared to bind the
elements making them less available for the plants. The evaluation of the potential of phytoremediation of our plants compared to other crops
in terms of metal removal was positive. Our results of metal removal are consistent with the results from other in situ experiments. The Zn re-
moval by S. bicolor and H. annuus reached about 2000 g ha1 and 1000 g ha1, respectively.
Ó 2007 Elsevier Masson SAS. All rights reserved.

Keywords: Soil pollution; Pyrite; Heavy metals; Phytoextraction; Fertilization; Field trial

1. Introduction polluted sites [9,16]. Phytoextraction of heavy metals by

Land is being continuously degraded and polluted by hu-
man activity such as industrial development, urbanization,
mining and intensive agriculture. Soil pollution is a condition
of global concern with regard to its implications on ecology
and human health due to the presence of organics and heavy
metals.
Phytoremediation has a strong potential as a natural, solar-
energy driven remediation approach for the treatment of
* Corresponding author. Fax: þ39 432 558603.
E-mail address: marchiol@uniud.it (L. Marchiol).
growing crops and harvesting above-ground material is a prom-
ising technology to remediate lightly or moderately polymetal-
lic contaminated soils [4].
Assisted phytoextraction [25], involves high-yielding crop
plants which can take up large amounts of metal, and is a re-
sponse to those management practices that increase the bio-
availability of elements by the application of chemical
agents and that maximize the efficiency of metal uptake by
the plants [21]. But this can lead to the expected results only
if the crop management is able to counteract the pressure
that the hostile soil exerts on the plants. In fact, the soils of
polluted sitesdand in particular those contaminated by indus-
trial slags and mining wastesdare often characterized by

0981-9428/$ - see front matter Ó 2007 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.plaphy.2007.03.018
380 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387
a poor content of organic matter, low levels of nutrients, pH
imbalance, and other physical anomalies [7,33]. Therefore,
the contaminated soil requires additional treatment to permit
the adaptation of metal-extracting plants [9].
Beyond the research that deals with metal uptake, transport
and tolerance of heavy metals in plants, another area of great
importance is the planning of the extensive field applications
used in the agronomic management of crops. In fact, how
crop management influences the process of phytoextraction
must be clarified. With this aim in mind, Clemente et al. [8]
observed the uptake of heavy metals by Brassica juncea and
studied the effects of organic amendments in a soil contami-
nated by pyrite wastes after the mine spill of Aznalcóllar
(Spain). The positive relationship between the efficiency of
metal phytoextraction and soil water content have been ob-
served by Angle et al. [2] who grew the hyperaccumulators
Alyssum murale, Berkheya coddii and Thlaspi caerulescens.
Wu et al. [32] examined the effects of the major plant nutri-
ents, singly and in combination, on the phytoextraction of cop-
per by Brassica juncea and they found a significant stimulation
not only in plant growth, but also in the uptake of the contami-
nant and therefore in the amount of pollutant that was cleaned
up from the soil.
In the literature, very few papers report any information
about in situ trials that looked at the possible enhancement
of metal uptake in response to modifications of agronomic
practices. On the other hand, this information is vital for
evaluating the clean-up duration, the number of annual
crops required to reach the target of soil remediation, and
the cost.
In order to evaluate the extensive application of phytoreme-
diation, we need to know the mass of elements that is expected
to be transferred from the contaminated soil to the plant bio-
mass, while taking into consideration (i) the initial level of
contamination and (ii) the target value to be reached after re-
mediation. Finally, the output of metals obtained by a single
crop cycle, expressed in grams per hectare per year, must be
multiplied by the number of clean-up cycles. Some examples
of such calculations have been reported by Robinson et al. [24]
and Keller et al. [15] for the remediation of contaminated soils
and by Ernst [9] for mine wastes.
The potential of metal-tolerant crops for phytoextraction
could be more properly estimated by considering the concept
of bio-available contaminant stripping (BCS), as discussed by
Hamon and McLaughlin [12]. From this perspective, the phy-
toextraction appears as a multistep process, which consists of
a series of depletions and replenishments of the metal bio-
available fraction.
This paper reports the first experimental data from an in situ
phytoremediation pilot plant, which was established in the
early spring of 2005. The study was conducted in Torviscosa
(Italy) in an experimental site, which is located in the chemical
plant Caffaro. Since 1938 the main activity of this industry
was the production of synthetic fibers; later, the chemical plant
was revamped in order to produce primary base- and
fine-chemicals. The decree 468/2001, by the Italian Ministry
of Environment, included this chemical plant and its
surroundings in the national priority list of polluted sites
with the name of ‘Laguna di Grado e Marano’.
The whole area of the plantdapproximately 110 hadis
polluted by heavy metals and organic compounds and the
soil of the experimental site is contaminated by pyrite cinders.
Pyrite ore contains primarily pyrite (FeS2), lesser amounts of
chalcopyrite (CuFeS2), sphalerite (ZnS), and magnetite
(Fe3O4), and several trace metals (mainly As, Cd, Cu and
Zn). The industry generated pyrite cinders as a by-product
of their sulfuric acid manufacturing operations which ceased
in the late 1970s. The wastes produced after roasting the pyrite
ore, contain a large amount of iron and various other trace
metals, such as arsenic, cadmium, copper and zinc, which
are commonly associated with the mineral.
Two annual high-biomass yield crops, Sorghum bicolor and
Helianthus annuus, were grown in a polymetallic substrate,
which was added in with chemical fertilizers and manure. Spe-
cific aims of this work were to: (i) observe the concentration of
the elements in the plant fractions during the crop cycle and
(ii) establish the amounts of trace metals that could be
extracted per hectare per annum.
2. Materials and methods
2.1. Experimental design
This study was conducted in Torviscosa, NE Italy (45
490 N, 13 160 E and 14 m above sea level). The climate is tem-
perate, with a mean annual temperature of 13.5  C and an
average annual rainfall of 1200 mm.
The experimental site consisted of 18 plots that were
15.75 m2 (4.5  3.5 m) each. The field trial was arranged in
a randomized block design, with two factors (species and
treatment) and three replications. The species considered in
our experiment during the summer 2005 were Sorghum bi-
color (L.) Moench and Helianthus annuus L.; these plants
were chosen after taking into account the results of prelimi-
nary pot experiments, which indicated the plants’ tolerance
to different levels of pyrite cinders.
The experiment that ran during the first season, dealt only
with the nutritional state of the polluted soil, without any ad-
ditional practices to enhance phytoextraction. The plants that
grew in the field trial were subjected to three treatments: the
Ctrl treatment, consisting of the pure polluted soil; the Fert
treatment, where the native soil was subjected to mineral fer-
tilization; and the Org treatment, where the native soil was
amended with cow manure. Fert-treated S. bicolor and H. an-
nuus received 150 kg N ha1 ((NH2)2CO in granules) which
was split into 100 kg N ha1 at sowing and 50 kg N ha1 after
7 weeks. Phosphorus was provided to both crops in the form of
calcium phosphate (CaHPO4) at a dose of 60 kg CaHPO ha1.
Potassium fertilization was achieved with granular KCl, at the
dose of 60 kg K2O ha1 and 290 kg K2O ha1 for S. bicolor
and H. annuus, respectively. Organic fertilization was
achieved by adding 90 t ha1 of mature cow manure to the na-
tive soil, which provided 316 mg Corg g1 1
ss ; 19.5 mg N gss ;
1 1 1
8.26 mg Ptot gss ; 171 mg Cutot gss ; 1178 mg Zntot gss .
 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387 381

Immediately before the seeding, the soil was ploughed to
a 25 cm depth and tilled. After the seedbed preparation and
fertilization, on May 5th, 2005, the seeding of S. bicolor and
H. annuus took place. Seeds of H. annuus cv. Carnia were
sowed at the density of 8 plants m2; seeds of S. bicolor cv.
Isadei were sowed broadcast at the density of 300 seeds
m2. In some plots of H. annuus, seedlings failed to emerge.
The plots were re-seeded on June 2nd, 2005.
The plots were watered during two drought periods of the
crop cycle, using a sprinkler irrigation system that supplied
the plots with about 7 mm h1. We did not provide a chemical
control for weeds because they were removed manually during
the crop cycle.
examine the effects of plant species and treatments. Stu-
denteNewmanneKeuls’s test (P < 0.05) was used to compare
treatment means. Statistical analysis was conducted using
CoStat 6.02 (CoHort Software, Monterey, CA). Crop growth
curves were obtained by interpolating the data with a logistic
function (TableCurve 5.01, Systan Software Inc.).
The accuracy of the analytical procedure was checked by
running up the standard solutions every 10 samples. For the
statistical analysis, in those samples where an element was
not detected, its concentration was assumed to be one-half
of the respective limits of detection [23]. Detection limits
were: 3 mg L1 for As; 1.5 mg L1 for Pb; 0.9 mg L1 for
Cu; 0.4 mg L1 for Co; and 0.2 mg L1 for Cd and Zn.

2.2. Plant harvest

Sampling started 40 days after sowing the plants. Plants of 3. Results and discussion
S. bicolor were collected from 0.4 m2 areas within the plots.
In the case of H. annuus, two sub-samples per plot were col- 3.1. Soil conditions
lected by sampling three adjacent plants in 0.75 m in two dif-
ferent areas of the plot, for a total of six plants. The plant The basic soil parameters are showed in Table 1 and the
harvesting was performed by using a shovel and paying atten- concentration of the heavy metals and As are provided in
tion not to break the roots. Subsequently, all the plant fractions Table 2. For reference, Table 2 also shows the threshold limits
(roots, leaves, shoots and heads), were carefully washed with as assigned by Italian legislation (D.Lgs. 152/06) on the basis
de-mineralized water. of land use. The concentration of As and Cu exceed the limits
Basic plant growth parameters, such as, number of green for industrial sites and the values for Cd, Co, Pb and Zn are
leaves and nodes, plant height, leaf area, and fresh and dry bio- higher than the limits assigned for green and residential areas.
mass of plant fractions were observed. Leaf area was esti- Even though our experimental site was located within an exist-
mated by measuring the green leaf area of all leaves with ing chemical plant, for the purposes of our research, we took
a leaf area meter (LI-3100, LI-COR). Dry biomass of the plant both threshold values into consideration.
fractions were measured after drying for 24 h at 105  C in an The experimental site of Torviscosa appears to be particu-
air-forced oven. larly suitable for studies on phytoextraction, because of the
presence of several metals in the soil layer that can be ac-
2.3. Heavy metals content in soil and plant fractions cessed by the roots of the plants, and because the metals are
not extremely concentrated.
In order to determine the concentration of heavy metals, Before we started the field trial, we performed several
samples from each plot were collected, air-dried, screened experiments under controlled conditions using the substrate
by means of a 2 mm sieve and finally oven dried (105  C collected at Torviscosa. The results of these experiments
for 24 h). The samples were acid-digested in a microwave demonstrated that our plants tolerated the concentration of
oven (CEM, MARSXpress), according to the EPA method the pollutants [18].
3051 [29]. After the mineralization, the samples were filtered
(0.45 mm PTFE) and diluted. Total content of As, Cd, Cr, Cu,
Table 1
Ni, Pb and Zn in topsoil and in the pyrite cinders were deter-
Basic parameters measured in the soil of the field trial of Torviscosa (Italy)
mined by an ICP-OES (Vista MPX, Varian Inc.). The analysis
Parameter Method
for As were done separately by using a continuous-flow vapor-
generation system, which provides improved detection limits Sand (2e0.05 mm; %) 69.4
Silt (0.02e0.05 mm; %) 6.55
for this element (VGA-77, Varian Inc.). The standard solutions
Silt (0.02e0.002 mm; %) 18.8
used for the ICP analytical calibration have their Certificate of Clay (<0.002 mm; %) 5.17
Analysis (CertipurÒ Reference Material). Acidity pH (H2O) 7.75 ISO 10390
The plant samples collected were acid treated following the Organic C (g kg1) 9.10
EPA method 3052 [30] and analyzed by an ICP-OES (Varian Cation exchange capacity (Cmolþ kg1) 5.5 Ac. NH4 pH7
Electrical conductivity (dS m1) 2.65 ISO 11265
Inc., Vista MPX).
P (mg kg1) 7.78 Olsen
Exchangeable K (mg kg1) (BaCl2 TEA) 132 ISO 11260
2.4. Data analysis Exchangeable Ca (mg kg1) (BaCl2 TEA) 3041 ISO 11260
Exchangeable Mg (mg kg1) (BaCl2 TEA) 198 ISO 11260
Data were subjected to a two-way analysis of variance Exchangeable Na (mg kg1) (BaCl2 TEA) 13.3 ISO 11260
Total Fe (g kg1) (69.5% nitric acid digestion) 90.2 USEPA 3051
(ANOVA) at each sampling time in order to statistically
382 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387

Table 2 the 6th sampling date, 112 days after sowing, the highest pro-
Total trace metal concentration measured in the soil of the field trial of Torvis- duction of above ground biomass was recorded: 1.54, 22.1 and
cosa (Italy) and concentration thresholds fixed by the Italian legislation for (A)
residential and (B) industrial soil utilization
16.9 tons of dry matter per hectare for Ctrl, Fert and Org,
respectively (Fig. 1).
Element Soil (A) D.Lgs 152/06 (B) D.Lgs 152/06
(mg kg1) (mg kg1) (mg kg1)
The data recorded during the whole growth cycle of H.
annuus suffered a high variability because of the delay in seed-
As 309 20 50
Cd 4.29 2 15
ing. Notwithstanding that the experimental data of H. annuus
Co 50.9 20 250 were disturbed by this variability, ANOVA showed a significant
Cu 1527 100 600 effect of treatment (P < 0.001). But H. annuus behaved in an
Pb 233 150 1000 opposite manner compared to S. bicolor. In fact, the highest
Zn 980 150 1500 recorded biomass was for the plants grown in the manure-
amended plots. At the 6th sampling date, 112 days after sow-
ing, the above ground biomass was 0.37, 3.38 and 6.30 tons of
3.2. Plant growth dry matter per hectare for Ctrl, Fert and Org, respectively
(Fig. 1).
The growth curves of the crops in response to the treatment
established in the trial, are shown in Fig. 1. As expected, the
native soil was poor in nutrients and therefore the biomass 3.3. Metal accumulation in plants
recorded during the growth of S. bicolor and H. annuus was
10e20 times higher in those plots, which received a nutrient Table 3 shows the concentrations of the elements measured
supply (both in mineral and organic form), compared with in the roots and shoots of the plants of S. bicolor and H. an-
the crops growing in the control soil. nuus, which were collected at the moment of the highest bio-
In the case of S. bicolor, ANOVA revealed a significant mass yield. The roots showed higher values of concentration
effect of the treatment: a part of the control soil and biomass than those observed in the shoots. As a consequence, the trans-
production was larger in Fert soil than in Org (P < 0.001). At location factor (shoot/root calculated in terms of concentra-
tion), which is typically >1 in the case of hyperaccumulator
species, was <1 for all the elements of our crops.

3.3.1. Arsenic
Under normal conditions As concentration in terrestrial
plants is usually less than 10 mg As kg1 of dry matter [28].
In our experimental conditions, the roots of S. bicolor accumu-
lated an amount of As that ranged between 32.3e67.5 mg kg1
(Table 3). However, ANOVA revealed an insignificant differ-
ence between the treatments. Also, we verified a high data
variability for H. annuus: the As concentration in roots varied
from 12.2e142 mg kg1. In regard to the concentration of As
in the shoots of S. bicolor and H. annuus, ANOVA revealed
a significant difference between the crops (P < 0.005), but
not between the treatments. On average, the As concentration
was about 7 and 58 times lower than in the roots, for S. bicolor
and H. annuus, respectively.

Fig. 1. Plant aboveground biomass production as a function of days after sow-
ing of S. bicolor and H. annuus grown in control (B), mineral fertilized ( )
and organic amended soil (:). Vertical bars for each point represent the stan-
dard error for three replicates.
3.3.2. Cadmium
On average, the Cd concentration recorded in the roots of S.
bicolor and H. annuus ranged between 0.73e2.31 mg kg1.
The organic amendment resulted in significantly lower Cd
concentrations than after other treatments (P < 0.05). In re-
gard to the values of Cd concentration measured in the shoots
of the plants, we observed values in the range of 0.01e1 mg
kg1, which have been reported by Chaney [5] as being nor-
mal values for Cd concentration in plants. H. annuus accumu-
lated more Cd than S. bicolor (P < 0.05), but the levels of
concentration were insignificant from the perspective of phy-
toextraction. Even though ANOVA did not reveal a significant

effect of the treatments (P ¼ 0.0547), the Cd concentrations
observed were higher in the shoots of unfertilized plants
 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387 383

(0.64 mg kg1), followed by those in Fert and Org plants (0.34

As (mg kg1) Cd (mg kg1) Co (mg kg1) Cu (mg kg1) Pb (mg kg1) Zn (mg kg1) As (mg kg1) Cd (mg kg1) Co (mg kg1) Cu (mg kg1) Pb (mg kg1) Zn (mg k1)

118  9.09
112  6.27
140  33.4
55.5  3.6
115  2.3
86.4  19
and 0.2 mg kg1, respectively) (Table 3).

n.s.
n.s.
*
3.3.3. Cobalt

2.73  0.38
4.10  0.48

2.52  0.82
3.53  0.88
0.53  0.13

Data are the means of three replicates  standard error. Significance levels of Species (Sp), treatments (Trea) and the interaction Sp  Trea are shown: *P < 0.05; **P < 0.01; ***P < 0.001.
6.08  2.9
Whereas similar values of Co concentration were found in
the roots of the crops (7.31 mg kg1 in S. bicolor and 6.07 mg

n.s.
n.s.
n.s.
kg1 in H. annuus), the treatments had a statistically signifi-
cant effect on the uptake of the element (P < 0.05). The plants
42.0  5.22
23.2  0.81
28.6  4.5
29.6  6.3
48.9  14
36.2  12
grown in the plots amended with organic matter had the lowest
concentration of Co in the roots, but this was more evident in
n.s.
n.s.
n.s.
the case of H. annuus than in S. bicolor (Table 3).
The root/shoot translocation of Co was not very high and
0.47  0.05
0.91  0.03
1.77  0.63
0.55  0.18
0.71  0.13
0.19  0.02

showed a significant effect only for the species but not for
the treatment (P < 0.05). The concentrations of Co in the
Concentration of As, Cd, Co, Cu, Pb and Zn in roots and shoots of Sorghum bicolor and Helianthus annuus recorded in plant samples collected 112 DAS

n.s.

shoots of S. bicolor and H. annuus had concentrations that

*

*
Metal concentration in shoots

were, respectively, 7 and 12 times lower than in the roots.

0.20  0.06
0.26  0.05
0.25  0.05
0.64  0.08
0.34  0.06
0.22  0.09

n.s.

3.3.4. Copper
*

Cu toxicity is associated with a soil Cu level of 150e400 mg

5.23  0.93
7.28  1.30

0.62  0.18
1.93  0.07
0.98  0.39

kg1 [1]. In our experiment the soil contained about 800 mg

13.1  4.3

kg1 of Cu and the concentration of Cu measured in the roots

***
n.s.
n.s.

of S. bicolor and H. annuus was, on average, about 589 mg

kg1. While ANOVA did not show any significant effect of
242  43.4
222  23.6
121  22.4
328  105
466  110

the species on the Cu concentration (Table 3), the treatments

265  70

did. The tissues of Org plants had the lowest concentration of

n.s.
n.s.

Cu (significant at P < 0.05), compared to both the control

*

and mineral-
42.9  4.63
54.0  7.74
15.7  4.65
60.1  4.1

fertilized plants. The Cu concentration in the shoots of the

60.1  26

22.7  2

crops was 15e18.5 times lower than in the roots (Table 2),
n.s.

n.s.

but none of the experimental factors were found to be signifi-

*

cant with ANOVA.

336  26.1
594  121

837  141
706  195
535  36
468  67

n.s.

n.s.

3.3.5. Lead
*

ANOVA did not show any difference between the species.

7.48  1.54
8.73  1.27
1.99  0.25

On the contrary, as a consequence of the organic amendment,

7.14  0.4
5.39  0.5
9.42  3

the concentration of Pb measured in the crop roots was more

n.s.

n.s.

than halved, compared to the other plants (P < 0.05). Accord-

*

ing to Wallace and Romney [31] the accumulation of Pb

Treatments Metal concentration in roots

1.75  0.31
1.71  0.11

2.31  0.68

0.73  0.12

occurs mostly in the root tissues of S. bicolor and H. annuus

1.35  0.4

1.57  0.1

(Table 3), but the concentrations we observed in the shoots

were, respectively, 15 and 18.5 times lower than in the roots.
n.s.

n.s.
*
48.6  7.17
142  88.6
32.3  9.8

12.2  2.2
67.5  18
47.8  13

3.3.6. Zinc
Zn concentration in the roots was significantly higher in
n.s.
n.s.
Sp  Trea n.s.

S. bicolor than H. annuus (P < 0.05). After averaging the

treatments, the Zn values were 352 and 195 mg kg1, respec-
tively (Table 3). In contrast, after averaging the treatments, the
Trea
Fert

Fert
Ctrl

Helianthus annuus Ctrl

Org

Org
Sp

Zn concentration in the shoots was 85.9 and 123 mg kg1 in

S. bicolor and H. annuus, respectively (Table 3). Such values
Sorghum bicolor

equate with a translocation factor of 0.24 and 0.63. However

the concentration of Zn did not exceed the value of 100e
ANOVA
Species
Table 3

500 mg kg1 in dry matter, which is cited in literature as being

the upper toxic level of Zn for various plants [13].
384 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387
3.4. Removal of trace metals
The amount of metal that plants can extract from the pol-
luted soil of Torviscosa was calculated by multiplying the
metal concentration measured in the harvestable fraction of
the crop and the biomass weight. Fig. 2 shows the values of
metal removed by harvesting the sole above ground biomass
and the whole plant.
A recent review of the literature confirmed that there is lack
of experience in evaluating the phytoextraction potential of
hyperaccumulators and high biomass crops under field condi-
tions. This is in part due to the fact that managing of in situ
experiments is often complicated and when it is permitted
by the appropriate governmental agency, strict governmental
regulations must be adhered to during the experiments. Results
from field trials that observed the performance of Thlaspi
caerulescens have been carried out by Schwartz et al. [26],
Hammer and Keller [11] and McGrath et al. [20].
Our data are comparable to those discussed by Madejón
et al. [17] and Keller et al., [15]. Both these papers provided
data about phytoextraction of heavy metals by crops that
were cultivated in field trials in areas affected by the pyrite
tailings spill in Aznalcóllar, Spain [17] and at an experimental
site in Dornach, Switzerland [15].
The 0e30 cm soil layer in the area contaminated by the
mine sludge of Aznalcóllar showed high levels of As, Cu,
Pb and Zn (2900, 1550, 7900 and 7100 mg kg1 respectively).
The plant uptake of heavy metals has been studied at two crop-
land sites, one with Helianthus annuus and one with Sorghum
bicolor [22].
The soil of the Dornach experiment was polluted by the
emission of a brass smelter. Total metal content (pH 7.3) at
the depth of 20 cm was: Cd 2.1 mg kg1, Cu 568 mg kg1
and Zn 360 mg kg1 [14]. In that site Brassica juncea, Zea
mays, Nicotiana tabacum, stands of Salix sp. and also the hy-
peraccumulator Thlaspi caerulescens have been cultivated
over two years.
3.4.1. Arsenic
The highest value of metal output was reached by S. bicolor
which removed from the soil 219 g ha1 of As; this value is
about 10 times higher than the metal output obtained by H. an-
nuus. The absence of nutrients in Ctrl soil resulted in a signifi-
cant difference among the soil treatments: in Ctrl soil the mean
offtake of As by S. bicolor was 7.48 g ha1, Fert and Org
plants allowed the removal of 158 and 219 g ha1 of As,
respectively. Our data show the same magnitude as those
provided by Madejón et al. [17].
A noticeable increase in the metal removal could be ob-
tained by harvesting not only the plant shoots but also the
roots. In this way it was possible to remove approximately
370 g ha1 of As from the contaminated soil, or 2.4 times
more than by harvesting the plant shoots (Fig. 2).
In our trial, by harvesting the aboveground biomass of H.
annuus we obtained an As removal of 0.18, 6.74 and 3.96 g
ha1 for Ctrl, Fert and Org plants, respectively (Fig. 2). In
the experiments performed in the area of Aznalcóllar, after
the mine spill, in growing plants of H. annuus an offtake of
approximately 3 g ha1 of As was observed.
3.4.2. Cadmium
On average, S. bicolor permitted a removal of Cd that was
higher than H. annuus. Harvesting sorghum at the moment of
highest biomass production, we removed 5.62 g ha1 of Cd
with Fert plants, while in the case of Org plants, the removal
was of 4.31 g ha1. The maximum Cd output obtained by
plants of H. annuus was about of 1.59, in both Fert and Org
plants (Fig. 2). The control plants, which were affected by se-
vere nutritional stress, produced a very low biomass and there-
fore their removal of Cd was negligible.
Keller et al. [15] reported the following Cd removal from
the Dornach experimental site: Brassica juncea 6.95 g ha1,
Nicotiana tabacum 41.7 g ha1, Zea mays 9 g ha1 and
Thlaspi caerulescens 179 g ha1. Apart from the hyperaccu-
mulator T. caerulescens, the values of maize and Indian mus-
tard are consistent with our data on sorghum.
Fig. 2 also shows that by harvesting the whole plants of
S. bicolor, the output of Cd is increased by about 2.5 and
1.7-fold, respectively.
3.4.3. Cobalt
No data from field trials on Co phytoremediation were
found in the literature. In our experiment, a single growth-cy-
cle of S. bicolor allowed a Co removal of 20.2 g ha1 and
30.2 g ha1 for Fert and Org plants, respectively. The metal
output calculated for H. annuus was one order of magnitude
lower than S. bicolor (Fig. 2). By re-calculating the output
of Co and considering the contribution of the roots, a 2.7
and 1.4-fold increase in the removal of the metal by Fert
and Org plants of S. bicolor, respectively, was found.
3.4.4. Copper
The maximum removal of Cu by plants of S. bicolor
reached respectively 820 and 644 g ha1 in Org and Fert
plants. Lower values have been observed for H. annuus; Org
plants removed 151 g ha1 of Cu and the Fert ones removed
116 g ha1 of Cu (Fig. 2). The hypothetical harvest of the en-
tire plants of Fert S. bicolor could allow a removal of Cu of
3215 g ha1, which is about five times higher than the result
obtained by harvesting the sole aboveground biomass; in the
case of Org plants of sorghum the removal of Cu was more
than doubled (Fig. 2). The same result was reached by the
plants of H. annuus.
The removal of copper calculated on the base of data col-
lected by Keller et al. [15] at the experiment of Dornach
showed that T. caerulescens had an efficiency in the removal
of Cu lower than the crops. In detail, the amounts of Cu ex-
tracted from the soil were: 146 g ha1 for B. juncea, 474 g
ha1 for N. tabacum, 163 g ha1 for Z. mays, and finally
50 g ha1 for T. caerulescens [15]. These low values of Cu
removal are not surprising, since this species is highly special-
ized for Zn. On the other hand, this feature is one of the
limitations of hyperaccumulators.
 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387 385

Fig. 2. Removal of trace metals by the shoots and the whole plant of S. bicolor and H. annuus grown in control, mineral fertilized and organic amended soil.
Vertical bars for each bar represent the standard error for three replicates.
386 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387
3.4.5. Lead
Combining the biomass yield with the concentration of Pb
in the shoots, it was observed that the highest offtake of the
element, 107 g ha1, was obtained by harvesting the Org
plants of S. bicolor, while Fert plant removed 91 g ha1 of
Pb (Fig. 2). Hypothesizing the harvest of the total biomass
of S. bicolor, the Pb removal rose to 378 g ha1 and 164 g
ha1 for Fert and Org plants, respectively (Fig. 2). The values
of Pb removal calculated for the plants of H. annuus were
significantly lower than S. bicolor. However our values
showed with those provided by Madejon et al. [17] that cur-
rently no realistic perspective of large scale application of phy-
toremediation in Pb polluted soils can be predicted.
3.4.6. Zinc
The total amount of Zn removed from the soil of Torviscosa
by S. bicolor calculated for the plants in correspondence of the
maximum biomass growth gave the values of 1223 and 1944 g
ha1 for Fert and Org plants, respectively. These values could,
on average, double when the entire plant biomass is harvested.
Due to the low biomass of Ctrl plants, the Zn removal was
148 g kg1 (Fig. 2).
The metal removal calculated for H. annuus was also vari-
able. The values were 410 and 804 g kg1, respectively for
Fert and Org plants.
Observing sunflower plants, Madejón et al. [17] calculated
a Zn removal of 2050 g ha1, when only the shoots of the sun-
flowers were harvested. The offtake increased to 2140 g ha1,
when the roots were also considered in the calculations. The out-
put of metal computed from the data collected on crops by Kel-
ler et al. [15] showed the following values: B. juncea 894 g ha1,
N. tabacum 1834 g ha1, Zea mays 1998 g ha1. In the same ex-
periment the hyperaccumulator T. caerulescens was more effi-
cient than expected, showing a metal removal of 5052 g ha1.
3.5. Use of sorghum and sunflower for phytoremediation
Phytoextraction is a long-term remediation effort, requiring
many cropping cycles to reduce metal concentrations to ac-
ceptable levels. The time required for remediation is depen-
dent on the type and extent of metal contamination and the
length of the growing season. However, the most important
feature of the plants is the efficiency of metal removal.
In our case the highest amounts of trace elements removed
by S. bicolor equated to only 0.03%, 0.056%, 0.024%,
0.225%, 0.018% and 0.082% of the total soil As, Cd, Co,
Cu, Pb and Zn content, respectively. The offtake of H. annuus
had the same order of magnitude. We extracted a negligible
amount of trace elements in comparison to the magnitude in
the soil; therefore, under our experimental conditions, S. bi-
color and H. annuus showed a very poor potential. On the
other hand, the experimental design did not consider any spe-
cific practices with which to enhance the bioavailability of the
trace metals. We have improved the nutritional status of sor-
ghum and sunflower and recorded a positive feedback by the
crops, in terms of element removal. In the future we will con-
sider other agronomic factors.
Biomass crops tolerant to trace metals are studied following
the approach of assisted phytoextraction [25], in which plants
have to be managed with practices able to enhance the element
bioavailability and the uptake. All of the factors, which must
be considered in successful agriculture, must also be consid-
ered during phytoremediation. From this point of view a lot
of work is needed.
Currently phytoextraction is far from being considered
a mature technology. It is difficult to predict when sufficient
knowledge will be acquired about specific molecular and
physiological aspects of phytoremediation for practical appli-
cation. Neither the biomass crops, nor the hyperaccumulator
plants have been studied deeply enough in field conditions.
Most of data published has been extrapolated from experi-
ments performed under conditions that are not adequate to
give results applicable for the future clean-up of contaminated
areas. This has been highlighted by Baker and Whiting [3]
and, more recently, again underlined by McGrath [20].
Among the few experiments dealing with the in situ perfor-
mances of hyperaccumulators, it was observed that hyperaccu-
mulator plants under certain conditions could not express their
potential. For example, testing the field performances of the
hyperaccumulators Thlaspi caerulescens and Arabidopsis hal-
leri, McGrath et al. [20] found that A. halleri did not perform
as expected and removed only 0.02% and 0.066% of the total
soil content of Cd and Zn. These values are not so different
from those we observed in our experiment, not growing hyper-
accumulator plants.
As regard to Thlaspi caerulescens, probably the most stud-
ied metal hyperaccumulator, it was demonstrated that the bio-
mass yields of T. caerulescens are highly variable, ranging
between 0.5 and 13.4 t ha1 [10,27,19,11,20], depending on
different ecological site characteristics. However, the funda-
mental traits of plants useful for phytoremediation, summa-
rized by Chaney et al. [6], should be expressed in a wide
range of environmental conditions to transform phytoremedia-
tion from a natural phenomenon to a sustainable technology
for soil clean-up.
4. Conclusions
We observed the phytoextraction potential of heavy metals
in polymetallic soil of S. bicolor and H. annuus. The plants
grew in a polluted soil, which had, within the profile 0e
30 cm, six trace metals exceeding the thresholds established
by the Italian legislation. A soil contaminated by several
metals in the surface layer could be intended as a good exper-
imental site for studies on phytoremediation.
In the early phases, the aim of our experiments was the
evaluation of the behavior of plants to the substrate which re-
ceived only a supply of nutrients by adding mineral fertilizers
and a manure amendment. Neither the chelate addition nor
other treatments have been established to modify bioavailabil-
ity of metals. In the native soil, which was very poor in nutri-
ents, we verified that the amelioration of the nutritive status of
soil was highly effective for the biomass yield. However,
fertilization and soil amendment did not increase the
 L. Marchiol et al. / Plant Physiology and Biochemistry 45 (2007) 379e387
concentration of metals in the harvestable tissue of the plants
during the crop cycle. In some cases, the organic matter
seemed to bind the elements, which made them less available
for plants.
The evaluation of the potential of phytoremediation of our
plants, compared to other crops in terms of metal removal, was
positive. In particular, in our experimental conditions we veri-
fied a metal removal which appeared to be consistent with the
results reported by Madejón et al. [17], and Keller et al. [15].
The trials at the site of Torviscosa are still running. Once
we complete our observations on the uptake and removal of
heavy metals by S. bicolor and H. annuus grown in plots
with native and mineral and organic fertilized soil, we will
focus on other experimental factors and other species.
Acknowledgements
This research was part of the project ‘‘Phytoextraction of
heavy metals in sites polluted by industrial activities: effi-
ciency of soil-plant system’’ (2003e2005), which was granted
by the Italian Ministry of University and Research.
References
[1] D.C. Adriano, Trace elements in terrestrial environments biogeochemis-
try, bioavailability, and risks of metals, Springer, 2001.
[2] J.S. Angle, A.J.M. Baker, S.N. Whiting, R.L. Chaney, Soil moisture ef-
fects on uptake of metals by Thlaspi, Alyssum, and Berkheya, Plant
Soil 256 (2003) 325e332.
[3] A.J.M. Baker, S.N. Whiting, In search of the Holy Grail - a further step in
understanding metal hyperaccumulation? New Phytol. 155 (2002) 1e3.
[4] L.A. Bouwman, J. Bloem, P.F.A.M. Römkens, J. Japenga, EDGA amend-
ment of slightly heavy metal loaded soil affects heavy metal solubility,
crop growth and microbivorous nematodes but not bacteria and herbivo-
rous nematodes, Soil Biol. Biochem. 37 (2005) 271e278.
[5] R.L. Chaney, Toxic element accumulation in soils and crops: protecting
soil fertility and agricultural food-chains, in: B. BarYosef, N.J. Barrow,
J. Goldshmid (Eds.), Inorganic Contaminants in the Vadose Zone,
Springer-Verlag, Berlin, 1989, pp. 140e158.
[6] R.L. Chaney, Y.M. Li, S.L. Brown, F.A. Homer, M. Malik, J.S. Angle,
A.J.M. Baker, R.D. Reeves, M. Chin, Improving metal hyperaccumulator
wild plants to develop commercial phytoextraction systems: approaches
and progress, in: N. Terry, G.S. Ba~nuelos (Eds.), Phytoremediation of
Contaminated Soil and Water. Proceedings of the Symposium on Phytor-
emediation, Fourth International Conference on the Biogeochemistry of
Trace Elements, CRC Press, Boca Raton, FL, 2000, pp. 129e158.
[7] K.K. Chiu, Z.H. Ye, M.H. Wong, Growth of Vetiveria zizanioides and
Phragmities australis on Pb/Zn and Cu mine tailings amended with ma-
nure compost and sewage sludge: a greenhouse study, Bioresour. Tech-
nol. 97 (2006) 158e170.
[8] R. Clemente, D.J. Walker, M. Pilar Bernal, Uptake of heavy metals and As
by Brassica juncea grown in a contaminated soil in Aznalcollar (Spain):
the effect of soil amendments, Environ. Pollut. 138 (2005) 46e58.
[9] W.H.O. Ernst, Phytoextraction of mine wastes e options and impossibil-
ities, Chem. Erde. 65 (2005) 29e42.
[10] H. Felix, Field trials for in situ decontamination of heavy metal polluted
soils using crops of metal-accumulating plants, Z. Pflanzenernähr. Bod-
enk. 160 (1997) 525e529.
[11] D. Hammer, C. Keller, Phytoextraction of Cd and Zn with Thlaspi caer-
ulescens in field trials, Soil Use Manage. 19 (2003) 144e149.
[12] R.E. Hamon, M.J. McLaughlin, Use of the hyperaccumulator Thlaspi
caerulescens for bioavailable contaminant stripping, in: W.W. Wenzel,
D.C. Adriano, B. Alloway, H.E. Doner, C. Keller (Eds.), Proceedings
387
of the Fifth International Conference on the Biogeochemistry of Trace
Elements Vienna (1999), pp. 908e909.
[13] A. Kabata-Pendias, Trace elements in soils and plants, third ed. CRC
Press, Boca Raton, FL, 2000.
[14] A. Kayser, K. Wenger, A. Keller, W. Attinger, H.R. Felix, S.K. Gupta,
R. Schulin, Enhancement of phytoextraction of Zn, Cd and Cu from cal-
careoussoil: the use of NTA and sulphur amendments, Environ. Sci.
Technol. 34 (2000) 1778e1783.
[15] C. Keller, D. Hammer, A. Kayser, W. Richner, M. Brodbeck, M. Sennhauser,
Root development and heavy metal phytoextraction efficiency: comparison
of different plant species in the field, Plant Soil 249 (2003) 67e81.
[16] M.M. Lasat, Phytoextraction of toxic metals: a review of biological
mechanisms, J. Environ. Qual. 31 (2002) 109e120.
[17] P. Madejón, J.M. Murillo, T. Maranón, F. Cabrera, M.A. Soriano, Trace
element and nutrient in sunflower plants two years after the Anzalcóllar
mine spill, Sci. Total Environ. 307 (2003) 239e257.
[18] L. Marchiol, G. Fellet, G. Zerbi, D. Perosa, Caratterizzazione di un suolo
contaminato da ceneri di pirite ai fini dell’applicazione di tecniche di
phytoremediation nel sito nazionale ‘‘Laguna di Grado e Marano’’. Atti
del seminario ‘‘Tecniche biologiche e chimico-fisiche innovative per il
monitoraggio e la bonifica dei siti contaminati e nuovi approcci all’ana-
lisi del rischio’’. Maggioli Editore, Santarcangelo di Romagna, 2004, pp.
349e357.
[19] S.P. McGrath, S.J. Dunham, R.L. Correl, Potential for phytoextraction of
zinc and cadmium from soils using hyperaccumulator plants, in:
N. Terry, G. Banuelos (Eds.), Phytoremediation of contaminated soil
and water, CRC Press, Boca Raton, FL, 2000, pp. 109e128.
[20] S.P. McGrath, E. Lombi, C.W. Gray, N. Caille, S.J. Dunham, F.J. Zhao,
Field evaluation of Cd and Zn phytoextraction potential by the hyperac-
cumulators Thlaspi caerulescens and Arabidopsis halleri, Environ. Pollut.
141 (2006) 115e125.
[21] E. Meers, A. Ruttens, M.J. Hopgood, D. Samson, F.M.G. Tack, Compar-
ison of EDTA and EDDS as potential soil amendments for enhanced phy-
toextraction of heavy metals, Chemosphere 58 (2005) 1011e1022.
[22] J.M. Murillo, T. Maranón, F. Cabrera, R. López, Accumulation of heavy
metals in sunflower and sorghum plants afffected by the Guadiamar spill,
Sci. Total Environ. 242 (1999) 281e292.
[23] M. Nadal, M. Schuhmacher, J.L. Domingo, Metal pollution of soils and
vegetation in an area with petrochemical industry, Sci. Total Environ. 321
(2004) 59e69.
[24] B.H. Robinson, R.R. Brooks, B.E. Clothier, Soil amendments affecting
nickel and cobalt uptake by Berkheya codii: potential use for phytomin-
ing and phytoremediation, Ann. Bot. 84 (1999) 689e694.
[25] D.E. Salt, R.D. Smith, I. Raskin, Phytoremediation, Annu. Rev. Plant
Physiol. Plant Mol. Biol. 49 (1998) 643e668.
[26] C. Schwartz, G.J. Echevarria, L. Morel, Phytoextraction of cadmium
with Thlaspi caerulescens, Plant Soil 249 (2003) 27e35.
[27] C. Schwartz, Phytoextraction des métaux pollués par la plante hyperac-
cumulatrice Thlaspi caerulescens, ENSAIA thesis, Nancy France, 1997.
[28] C. Tu, L.Q. Ma, B. Bondada, Arsenic accumulation in the hyperaccumu-
lator Chinese Brake and its utilization potential for phytoremediation,
J. Environ. Qual. 31 (2002) 1671e1675.
[29] USEPA, EPA Method 3051: Microwave assisted acid digestion of sedi-
ments, sludges, soils, and oils, in: Test Methods for Evaluating Solid
Waste, third ed. (1995) Washington, DC.
[30] USEPA, EPA Method 3052: Microwave assisted acid digestion of sili-
ceous and organically based matrices, in: Test Methods for Evaluating
Solid Waste, third ed. (1995) Washington, DC.
[31] A. Wallace, E.M. Romney, Roots of higher plants as a barrier to translo-
cation of some metals to shoots of plants, in: H. Drucker, R.E. Wildung
(Eds.), Proceedings of the Fifteenth Annual Hanford Life Science Sym-
posium (1977), pp. 370e379, Richland.
[32] L.H. Wu, H. Li, Y.M. Luo, P. Christie, Nutrients can enhance phytoreme-
diation of copper-polluted soil by Indian mustard, Environ. Geochem.
Health 26 (2004) 331e335.
[33] Z.H. Ye, W.S. Shu, Z.Q. Zhang, C.Y. Lan, M.H. Wong, Evaluation of ma-
jor constraints to revegetation of lead/zinc mine tailings using bioassay
techniques, Chemosphere 47 (2002) 1103e1111.