ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Distribution and mobility of metals in contaminated sites.
Chemometric investigation of pollutant profiles
Ornella Abollino
a
, Maurizio Aceto
b
, Mery Malandrino
a
, Edoardo Mentasti
a,
*,
Corrado Sarzanini
a
, Renzo Barberis
c
a
Department of Analytical Chemistry, University of Torino, Via P. Giuria 5, 10125 Torino, Italy
b
Department of Science and Advanced Technologies, University of East Piedmont, Corso Borsalino 54, 15100 Alessandria, Italy
c
Environmental Protection Agency of the Regional Government of Piedmont (ARPA Piemonte), Via della Rocca 49, 10123 Torino, Italy
Received 10 July 2001; accepted 9 November 2001
‘‘Capsule’’: Chemometrics allowed identification of groups of samples with similar characteristics.
Abstract
The distribution and mobility of heavy metals in the soils of two contaminated sites in Piedmont (Italy) was investigated, evalu-
ating the horizontal and vertical profiles of 15 metals, namely Al, Cd, Cu, Cr, Fe, La, Mn, Ni, Pb, Sc, Ti, V, Y, Zn and Zr. The

present paper describes the characterisation of heavy
metal pollution in the soils of two sites formerly used
for industrial waste disposal. The horizontal and ver-
tical distribution of contaminants was investigated and
the concentrations were compared with the acceptable
limits imposed by Italian and Dutch legislation (Minis-
try of Housing, 1994; Ministerial Decree, 1999b) for soil
reclamation. A chemometric treatment of the data was
performed.
The toxicity of metals depends not only on their total
concentration, but also on their mobility and reactivity
with other components of the ecosystem. The most com-
mon way to study element mobility in soils is by treat-
ment with extractants of different chemical properties
0269-7491/01/$ - see front matter # 2001 Published by Elsevier Science Ltd. All rights reserved.
PII: S0269-7491(01)00333-5
Environmental Pollution & (&&&&) &–&
www.elsevier.com/locate/envpol
* Corresponding author. Tel.: +39-011-6707625; fax: +39-011-
6707615.
E-mail address: (E. Mentasti).
1
2
3
4
5
6
7
8
9

40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69

100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
(Nowak, 1995; Szulczewski et al., 1997; Rauret, 1998).
In this work the release of metals into water, dilute
acetic acid and EDTA was investigated, and the Tes-
sier’s partitioning scheme (Tessier et al., 1979) was
applied to selected samples.
The results obtained can be of use for the local
authorities to decide about the necessity of reclamation
of the two sites and the level of priority of the interven-
tion, with respect to the situation of other polluted areas.
Moreover, the data can be of interest to the European
Environment Agency for its activities of soil monitoring.
2. Description of the experimental procedures

province of Novara. The contamination occurred
because of the repeated floods of a small stream, which
today has a new course, caused by the insufficient size of
the stream bed with respect to the flow in rainy periods.
The stream collected wastewaters of local industries,
some of which operating in the electroplating field, and
its floods caused an accumulation of contaminants,
mainly of inorganic nature, in the soil. The extension of
the polluted area is estimated between 20,000 and
100,000 m
2
. The core of the contaminated zone is about
3000 m
2
wide: it is a flat, uncultivated area, covered by a
layer of black sludge about 1.50 m deep carried by
the floods, where a scant vegetation grows. The rest of
the area is covered by trees and spontaneous plants. The
land in the zone is made of alluvial deposits. The top
layer of the soil, down to a depth of from 0.6 to 2 m, is
composed of sand with silt and clay, with a low gravel
content. This layer gives a discrete impermeability to the
soil. Below there is an alluvial layer with sand and gravel,
down to groundwater which flows at 4–5 m depth.
Table 1 reports a brief description of the location of
the single sampling points, which were chosen in a ran-
dom fashion in order to cover the whole areas. A total
of 33 samples was collected at site A, both at different
points of the presumably most contaminated zone and
in the surroundings. Some were sampled from the sur-

Cadmium and lead, when present below the ICP–AES
detection limits, were determined with a Perkin Elmer
5100 (Perkin Elmer, Norwalk, Connecticut, USA) elec-
trothermal atomic absorption spectrometer (ETAAS)
equipped with Zeeman-effect background correction.
Sample dissolutions for the determination of total
concentrations were performed in tetrafluormethoxyl
(TFM) bombs, with a Milestone MLS-1200 Mega
(Milestone, Sorisole, Italy) microwave laboratory unit.
Analytical grade reagents were used throughout. Stan-
dard metal solutions were prepared from concentrated
2 O. Abollino et al. / Environmental Pollution & (&&&&) &–&
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78

109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Merck Titrisol stock solutions (Merck, Darmstadt,
Germany).
2.3. Procedures
All experiments were performed in triplicate and
blanks were run simultaneously. The relative standard
deviations of the results were typically below 10%.
Higher deviations were observed, for some data, for
extractions in water and acetate, owing to the low con-
centrations measured; in some cases also the total
concentrations in the polluted area at A site showed a
variability higher than 10%, especially for copper,
because of the heterogeneity of the samples.
The evaluation of pH and EDTA-extractable frac-
tions was performed according to the official methods of
soil analysis envisaged by the Italian legislation (Minis-
terial Decree, 1992). After completion of the experi-
mental work, a new revision of official methods was
issued (Ministerial Decree, 1999a), which in any case
only slightly differs from the previous one. The leaching
test with acetic acid was performed according to the
Italian official methods for sludge analysis (Water

A10 0 cm below A9 B10 Just outside site core
A11 Relief, 30 m far from A9 B11 10 cm below B10
A12 10 cm below A11 B12 Vertical profile, 0–15 cm
A13 Hole in the relief B13 Vertical profile, 15–30 cm
A14 Border of the relief B14 Vertical profile, 30–40 cm
A15 Coloured material from the relief B15 Vertical profile, 40–50 cm
A16 Coloured material from the hole B16 Vertical profile, 50–65 cm
A17 Vertical profile, 0–30 cm B17 Vertical profile, 65–80 cm
A18 Vertical profile, 30–50 cm B18 Vertical profile, 80–100 cm
A19 Vertical profile, 50–60 cm B19 Vertical profile, 100–115 cm
A20 Vertical profile, 60–100 cm B20 Vertical profile, 115–130 cm
A21 Vertical profile, 100–135 cm B21 Vertical profile, 130–145 cm
A22 Vertical profile, 135–155 cm B22 Vertical profile, 145–160 cm
A23 Vertical profile, 155–160 cm B23 About 5 m North to the site core
A24 Vertical profile, 160–190 cm B24 About 5 m South to the site core
A25 Vertical profile, 190–218 cm B25 About 5 m West to the site core
A26 Vertical profile, 218–238 cm B26 About 5 m East to the site core
A27 Vertical profile, 238–260 cm B27 About 200 m S-E to the site core
A28 Vertical profile, 260–280 cm B28 City centre
A29 Vertical profile, 280–300 cm
A30 Vertical profile, 300–320 cm
A31 Vertical profile, 320–330 cm
A32 About 400 m far from the site
A33 Centre of the town
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 3
1
2
3
4
5

36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65

96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
of 5 min each (at a power of 250, 400, 500 W, respec-
tively) followed by a final 3 min step at 600 W. Then 0.7
g of boric acid were added and the bombs were further
heated for 10 min at 250 W. Finally the samples were
filtered and diluted to 100 ml (Bettinelli et al., 1989;
Aceto et al., 1994; Gulmini et al., 1994).
2.3.4. Available metal fraction
The extractant was a 0.02 mol dm

2.3.6. Sequential extractions
The sample (1.0 g) was sequentially extracted with
different reagents according to the following procedure
(Tessier et al., 1979): (1) 8 ml of 1 mol dm
À3
MgCl
2
, for
1 h, at room temperature; (2) 8.0 ml of 1 mol dm
À3
CH
3
COONa, added with CH
3
COOH (pH 5.0), for 5
hours, at room temperature; (3) 20 ml of 0.04 mol dm
À3
NH
2
OH. HCl in 25% CH
3
COOH, for 6 h at 96Æ 3

C;
(4) 5.0 ml of 30% H
2
O
2
and 3.0 ml of 0.02 mol dm
À3

(i.e. > 90%) for Cd, Cr, Cu, Ni, Pb, Zn (i.e. the heavy
metals of greatest interest from the environmental point
of view) and Al, whereas Fe, Mn, Ti, V, Zr were par-
tially lost (recoveries ranged between 67 and 82%). Zir-
conium was mostly lost in the first extraction step,
manganese in the fourth one, whereas losses of the other
elements took place in all the first four steps, probably
during filtration of the surnatant.
2.3.7. Chemometric data treatment
Two unsupervised methods (Hierarchical Cluster
Analysis, HCA, and Principal Component Analysis,
PCA) and a supervised one (Discriminant Analysis, DA)
were applied to the data. The treatment was performed
with XlStat, an add-in package of Microsoft Excel.
HCA was run applying Ward’s method of agglom-
eration and squared Euclidean distance as similarity
measure. All variables were standardised by transform-
ing data into Z-scores (i.e. (xÀx
m
)/, where x
m
stands
for the average). Dendrograms were obtained.
As to DA, two classes were defined a priori, con-
sidering samples from sites A and B respectively. Uni-
variate ANOVA was used to calculate F-ratios and find
out variables with higher discriminating power. Prob-
abilities of class membership were calculated for all
samples.
3. Results and discussion

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94

normal ranges and the most common values typically
present in soils (Halloway, 1990; Merian, 1991) and
with the maximum admissible levels in soils according
to Italian (Ministerial Decree, 1999a,b) and Dutch
(Ministry of Housing, 1994) legislations. These data are
collected in Table 5, which also reports, for comparison,
the mean content in the earth’s crust (Weast, 1974).
3.1.1. Site A
Abnormally high levels of Cd, Cu, Pb and Zn were
found at site A. In particular, the presence of copper is
related to the disposal of electric cables. The most pol-
luted zone is the relief, in which an overall increase of
these four elements and, to a lesser extent, of chromium,
manganese, nickel and zirconium can be observed. The
concentrations of these metals are smaller in the basin,
even if a contamination of cadmium, copper, lead and
zinc is present. Also the neighbouring zone under the
vegetation has relatively high levels of Cu, Pb and Zn.
The concentrations at the base of the relief usually fall
between the ones in the relief and under the vegetation.
The contents of Cr and Ni do not exceed the typical
ranges, but in many samples are above the common
values reported in Table 5 and, especially in the vertical
profile (whose behaviour will be discussed below), are
higher than in the surroundings: therefore an input of
these elements with the waste can be supposed. The
same hypothesis is valid for manganese, whose level in
the vertical profile, moreover, is higher than typical
values. Some samples on the relief (A9, A10, A12, A13)
are also rich in zirconium, which might have been con-

A22 5.74 68348 44.6 279 13208 19066 14.7 3990 408 19657 5.71 2027 42.8 10.7 45221 64.8
A23 5.88 71978 120 459 22351 59650 15.4 10403 542 35613 4.50 1923 68.4 10.1 54422 46.3
A24 5.88 56434 77.8 302 22627 45086 13.9 18420 282 30069 4.93 1765 57.2 11.0 45304 41.7
A25 5.88 55615 46.3 483 28492 37571 9.4 17050 289 37050 3.00 1051 37.3 7.99 51774 48.7
A26 5.87 64672 62.4 386 26134 48444 14.3 13816 305 33045 4.85 1869 45.6 10.9 45108 39.2
A27 5.87 58161 45.6 311 21575 43769 11.57 10188 255 26449 4.89 1614 43.8 9.45 35475 32.5
A28 5.98 52281 37.7 352 21279 34177 10.2 11675 271 21279 3.44 1219 40.2 8.07 40064 41.0
A29 5.88 54322 48.4 259 20695 39293 12.9 7218 236 22613 4.82 1445 38.4 9.13 31020 31.0
A30 5.75 66399 39.6 356 22712 48785 18.3 6781 341 32283 4.90 1883 48.1 10.3 42636 44.0
A31 5.87 64248 27.8 337 21162 45689 16.7 7721 294 27291 5.58 1814 40.5 11.6 38733 41.1
A32 7.75 61730 < 7.50 47.2 41.8 32877 10.1 705 < 30.0 119 11.2 4115 36.8 16.9 108 < 10.0
A33 6.49 59380 < 7.50 63.8 73.6 45457 12.1 850 < 30.0 69.4 13.5 7600 66.2 14.1 255 < 10.0
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77

108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
contents of Cr, Cu, Pb, Ni, Mn and Zn are inside the
typical ranges.
Al, Fe, La, Sc Ti, V and Y concentrations do not
show a definite trend as a function of sampling position.
The level of iron on the relief is higher than at the base
and in the basin, but lower than in the city center: it is
likely that the disposed wastes contained iron, causing a
local increase in its concentration, even if the levels
reached are not abnormal. An input of lanthanum with
the wastes may have occurred, since its concentrations
are slightly higher at the site than in the city centre and
the surrounding area.
The pH is lower in the relief than in the surrounding
area (except under vegetation).
The concentrations of Cd, Cu, Mn, Ni, Pb, Zn and Zr
are lower in surface samples than 10 cm below, with a
few exceptions. In most cases the levels of La, Sc, Ti, V
and Y show the opposite behaviour. There is not a reg-
ular trend for Al, Cr or Fe.
The concentrations of Cd, Cr, Cu, Mn, Ni, Pb, Zn

extents) anthropogenic sources, whereas the ones of
group four have mainly a geochemical origin.
Table 3
Total metal concentrations (mg/kg) and pH at site B
Sample pH Al Cd Cr Cu Fe La Mn Ni Pb Sc Ti V Y Zn Zr
B1 5.00 58931 4.75 3593 5753 29933 28.3 259 1418 687 9.07 4712 43.9 20.2 679 42.8
B2 4.06 58150 8.05 2880 4013 29246 33.9 251 741 864 8.61 5027 43.6 19.5 417 44.8
B3 5.17 50390 5.54 3016 5743 29022 27.4 265 1983 703 3.20 3808 40.9 19.8 1053 42.4
B4 4.51 48639 2.74 3160 5032 28635 26.8 230 1140 657 7.26 4105 40.9 17.1 598 40.5
B5 4.97 64379 0.98 312 3953 29966 32.5 281 969 533 8.02 4611 36.7 19.1 517 38.0
B6 5.80 64067 0.54 4523 7657 28501 41.7 259 1964 1162 9.86 6273 46.3 22.0 868 50.9
B7 6.08 64961 3.03 3905 4092 29487 24.4 240 1021 393 8.23 5172 50.0 12.1 723 54.2
B8 5.30 31514 0.17 2781 2485 19005 19.8 521 1471 675 5.46 3888 27.2 13.6 738 20.9
B9 5.66 59799 0.15 2101 1537 27674 23.8 328 856 244 8.53 4693 40.9 11.9 618 51.5
B10 5.22 61453 1.34 179 3151 34306 24.6 372 1142 621 7.32 4613 43.4 17.9 849 38.5
B11 4.96 55802 1.43 338 5778 30075 21.9 259 1272 751 5.79 5273 38.6 14.9 804 37.8
B12 4.41 63029 0.84 3123 3478 29412 31.8 265 697 156 9.50 4431 38.2 23.1 346 43.6
B13 3.92 67849 1.37 4683 3310 31394 31.7 258 648 159 9.79 5978 37.1 22.9 466 43.8
B14 3.81 88936 0.55 299 1019 39147 27.7 311 128 15.7 12.8 3947 46.3 16.7 191 60.3
B15 3.43 86985 2.96 140 787 36757 26.3 320 261 10.3 11.9 3741 55.0 16.9 393 65.5
B16 3.42 83775 2.87 199 1721 33967 23.2 348 412 10.3 11.7 3588 52.3 19.0 713 59.6
B17 3.99 81668 0.89 184 373 36370 31.6 455 1037 10.4 11.9 3689 47.2 20.5 957 45.0
B18 5.49 78555 0.22 128 80.0 37125 26.2 473 132 9.16 11.4 3691 51.7 18.4 147 47.0
B19 4.56 77734 0.61 860 1110 35784 28.8 391 406 57.6 11.3 3916 49.0 19.8 355 47.5
B20 5.63 72181 0.12 170 126 33379 25.3 404 98.0 11.7 10.4 3350 49.5 17.8 127 35.2
B21 6.05 73021 0.14 152 82.0 36502 19.2 437 179 8.33 12.1 3638 53.1 16.5 142 33.6
B22 5.74 71518 0.10 < 20.0 86.2 39333 17.4 624 158 5.59 11.5 3773 43.7 15.1 138 66.5
B23 5.75 50079 < 0.05 2958 2075 34372 24.5 269 863 664 < 2.00 5365 57.8 10.8 909 63.6
B24 4.16 45356 < 0.05 55.28 74.3 38252 29.2 337 33.0 21.0 3.18 3782 56.5 3.48 109 88.9
B25 4.17 41888 < 0.05 37.7 62.6 28110 17.8 394 30.1 293 3.88 3210 37.3 5.66 108 67.2

27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED

decrease with depth, whereas those of zirconium do not
show any trend. The pH value is lower than 4 at a depth
of between 15 and 80 cm: it is likely that this low value
is due to an input of acidic wastewater.
In general, the metals can be divided into two groups:
(1) Cd, Cr, Cu, Ni, Pb and Zn, whose concentrations
are heavily affected by anthropogenic inputs, and (2) Al,
Fe, Mn, Sc, Ti, V and Zr, which are mainly of geo-
chemical origin.
3.1.3. Legislation limits
The results were compared with the maximum
acceptable concentrations in soils reported in the
Table 4
Mean, median, ranges of total concentrations (mg/kg) at sites A and B
Site Mean
a
Median
a
Range
a
Mean
b
Median
b
Range
b
Mean
c
Median
c

Zn A 12255 6205 108–47681 13980 28916 677–47681 45713 45221 31020–66457
B 592 679 103–1053 715 723 417–1053 361 346 127–957
Zr A – 10.3 < 10.0–44.4 – 14.6 < 10.0–29.3 43.4 41.7 20.6–66.3
B 50.3 44.8 20.9–88.9 42.0 42.4 20.9–54.2 49.8 47.0 33.6–66.5
a
All samples excluding A15, A16 and the vertical profile (A1–A14 and A32–A33; B1–B11 and B23–B28).
b
All samples excluding A15, A16, vertical profile and the ones outside the most polluted area (A1–A14; B1–B11).
c
Vertical profile samples (A17–A31; B12–B22).
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79

110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Italian legislation
6
for the reclamation of contaminated
sites and with Dutch intervention values (Ministry of
Housing, 1994), formerly known as C values (Table 5).
Italian limits depend on land use, and are lower for
public and private green areas and residential sites (‘‘A’’
limits) and higher for industrial areas (‘‘B’’ limits). At
site A, the levels of Cd, Cu, Pb and Zn exceed ‘‘A’’ and
‘‘B’’ limits in most samples; the concentrations of Cr
and Ni are higher than ‘‘A’’ limits, but below ’’B’’ ones.
Copper, chromium and nickel contents at site B are
above both limits in many samples, whereas lead and
zinc, and in some cases cadmium, are between ‘‘A’’
and ‘‘B’’ values.
All samples exceeding ‘‘A’’ and some of the samples
exceeding ‘‘B’’ levels have concentrations above Dutch
intervention values, which are intermediate between the
two Italian sets of limits, and are to be considered,
according to the official terminology, ‘‘seriously pol-
luted’’.
3.1.4. Chemometric processing

apart but not very far from them. The specimens
from the relief (A9–A13, with the exception of A14)
are in other zones of the plot. They are distanced
from each other, owing to the heterogeneity of the
wastes. The combined plot shows that they are mainly
characterized by high concentrations of the polluting
elements. One of the pieces of material (A16) is com-
pletely isolated from the other samples, confirming its
different characteristics, and is strongly characterised
by its copper content;
the metals belonging to the first two above identified
groups, together with zirconium, are correlated, with
the exception of copper which stands alone. They
have opposite values of PC1 with respect to the other
Table 5
Typical concentration ranges and most common values present in soils, average abundance in the earth’s crust, acceptable concentrations in soils for
Italian legislation (A: limits for public and private green areas and residential use; B: limits for commercial and industrial use of soil), target and
intervention values for Dutch legislation (values in mg/kg unless otherwise stated)
Range Common values
a
Earth’s crust Limit (A) Limit (B) Target value Intervention value
pH 4–8.5
Al 81,300
Cd 0.01–2.0 0.2–1 0.15 2 10 0.8 12
Cr 5–1500 70–100 200 150 800 100 380
Cu 2–250 20–30 70 120 600 36 190
Fe 0.7–4.2% 50,000
La 18
Mn 20–10,000 1000 1000
Ni 2–750 50 80 120 500 35 210

20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49

80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109

as to the loadings, the polluting elements are less
strictly correlated than in the horizontal profile, even
if they are in the same area of the plot and load posi-
tively on PC1, with the exception of Zr and Cr. pH is
anticorrelated to this groups of variables. V and Fe
behave like the pollutants, whereas Al, Sc, La, Ti and
Y are in other areas of the plot. It can be supposed
that PC1 is connected to the elements of mainly
anthropogenic origin and PC2 to the ones of a
mainly geochemical source;
HCA confirms the different characteristics of the first
two layers;
when the data for site A are processed all together,
the variance explained by the first two PCs is 50 and
19% respectively (69% in all). The samples for ver-
tical and horizontal profiles form two groups in the
plot of PC1 vs. PC2, the former being characterised
by their content in polluting elements; exceptions to
this distribution are A10 and A13, which show a
stronger similarity with the vertical samples, and A16,
the piece of material, which is isolated from the other
specimen. Two clusters corresponding to horizontal
and vertical (plus A10, A13, A16) profile samples are
also present in the dendrogram.
Data processing for site B gave the following results:
as to the horizontal profile, the variance explained by
the first two PCs is 38 and 19% respectively (57% in
all). According to the plot of PC1 vs. PC2 (Fig. 2a)
samples B1–B11, collected in the core or just outside,
and B23–B27, from the surroundings, form two

metal concentrations).
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88

A differentiation between groups B14–B17 and B18–
B22 (with sample B19 as an outlier) is also present;
the group of the polluting elements is more scattered
than in the horizontal profile, but it still has loadings
on PC1 of the opposite sign with respect to many
elements of a mainly geochemical source. PC1 is
almost unaffected by pH, which on the other hand
heavily loads on PC2. pH is anticorrelated to Cd and
Zn and, at a lower level, to Ni, La, Zr, Al;
HCA confirms the observations made above: the first
two layers from a separate cluster, and groups B14-
B17 and B18-B22 are again present, with sample B19
being closer to the former;
when all data for site B are considered together, the
variance explained by the first three PCs is 38 and
18% respectively (56% in all). The sampling locations
can be divided into three main groups: (1) the hor-
izontal profile in the site core or just outside it (B1–
B11), together with the first two layers of the vertical
profile (B12–B13), characterized by a high content of
contaminants; (2) the deeper layers of the vertical
profile (B14–B22); (3) the samples collected in the
surroundings of the site (B24–B27), excluding B23,
and in the city centre (B28). The corresponding den-
drogram showed a similar clustering.
The data for sites A and B were also treated together.
The variance explained by the first two PCs is 40 and
17% respectively (57% in all). Three groups are present,
corresponding to (1) most A samples, (2) B samples
from the horizontal profile, (3) vertical B samples and

applied only to site A, because the pH of the water sus-
pensions of most site B samples was already lower than
5.0. Tessier’s protocol was applied to two samples for
each site.
PCA and HCA were performed on the percentages
extracted in water, acetic acid and EDTA.
3.2.1. Leaching with water
The leaching test with pure water was performed in
order to evaluate the fraction of metals weakly bound to
the matrix, e.g. present as inorganic soluble salts. The
results can also give a preliminary indication on the
possible release of pollutants by rains, although of
course the laboratory experimental conditions are dif-
ferent from the on-site situation. Moreover, it is likely
that most of the very labile metal fraction has already
been leached over the years. The percentages of metals
solubilised by water, their median and ranges are
reported in Table 6. As can be seen, the extracted
Fig. 2. Combined plot of scores and loadings obtained by (a) PCA
and (b) dendrogram for horizontal profile samples at site B (total
metal concentrations).
10 O. Abollino et al. / Environmental Pollution & (&&&&) &–&
1
2
3
4
5
6
7
8

39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

99
100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Table 6
Percentages extracted in water
a
Sample Al Cd Cr Cu Fe La Mn Ni Pb Sc Ti V Y Zn Zr
Aa 0.007 0.524 < 0.005 0.169 < 0.001 < 0.004 0.022 0.579 0.019 < 0.088 < 0.001 < 0.055 < 0.032 0.041 < 0.093
Ab 0.002 1.162 < 0.003 0.056 < 0.001 < 0.003 0.122 0.234 0.006 < 0.100 < 0.001 < 0.057 < 0.031 0.512 < 0.048
Ac 0.002 0.457 < 0.003 0.033 < 0.001 < 0.004 0.015 0.127 0.005 0.140 0.003 < 0.062 < 0.031 0.333 < 0.066
A26 0.002 0.543 < 0.003 0.038 0.002 < 0.004 0.024 0.364 0.007 0.103 0.004 < 0.065 0.031 0.388 < 0.077
A27 0.001 0.654 < 0.004 0.040 0.001 < 0.004 0.019 0.149 0.007 < 0.102 < 0.001 < 0.052 0.040 0.333 < 0.092
A28 0.001 0.774 < 0.003 0.024 < 0.001 < 0.005 0.015 0.435 0.006 < 0.145 0.001 < 0.062 0.046 0.322 < 0.073
A29 0.001 0.666 < 0.005 0.036 0.001 < 0.004 0.037 0.154 0.007 < 0.104 0.002 < 0.070 0.056 0.364 < 0.097

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39

70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99

amount released is lower than at site A, owing to the
lower total concentrations present. The solubility of
the elements of mainly geochemical origin is in most
cases below 1%, with some exceptions for La, Mn, Y.
The highest percentages for the pollutants are encoun-
tered between 30 and 80 cm: it is possible that some of the
metals released by the rain from top layers precipitated
again as water was flowing through deeper layers.
When PCA was applied to the data obtained for site
A, the variance explained by the first two PCs was 36
and 23% respectively (59% at all). The plot of PC1 vs.
PC2 (Fig. 3a) shows that the extract from the first layer
(Aa) differs from the other ones for its high content of
Pb, Cu, Al and Ni. A similarity among the samples
between 238 and 300 cm (A27–A29) can be observed.
Few correlations among the variables are present, such
as Pb–Al–Cu–Ni (anticorrelated to pH), and Cd–Mn.
Therefore the solubility is not only related to the
(anthropogenic or geochemical) origin of the elements
but also to other factors, such as their chemical proper-
ties. The following clustering is observed in the dendro-
gram reported in Fig. 3b: sample Aa, which stands
alone; sample Ab (from 60 to 155 cm); samples A27–
A31 (from 238 to 330 cm); samples Ac–A26 (from 155
to 238 cm).
As to site B, the variance explained by the first two
PCs was 67 and 15% respectively (82% at all). Layer
B16 (50–65 cm) is clearly differentiated from the other
ones, owing to the higher percentages of the above
mentioned nine metals in the extract. The samples

The results of the leaching test at pH 5.0 for site A are
reported in Table 7. The extracted fractions are slightly
higher than the ones found with pure water, because the
slightly lower pH favours the dissociation of the existing
complexes.
The most extensively extracted metals were Cd, Cu,
Zn and Ni. In general, endogenous metals are more
strongly bound to the soil matrix, whereas the ones
introduced by anthropogenic activities are in a more
soluble form and therefore more easily released into the
environment and potentially more toxic. Moreover,
Fig. 3. Combined plot of scores and loadings obtained by (a) PCA
and (b) dendrogram for vertical profile samples at site A (extracts in
water).
12 O. Abollino et al. / Environmental Pollution & (&&&&) &–&
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75

106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
bivalent elements are extracted more easily than tri- and
tetravalent ones. the latter probably form complex
anions, and/or stronger complexes with soil organic
matter, thus limiting their extractability; finally, their
ionic radii are smaller than their divalent counterparts
and are thus more likely to enter pores in the mineral
phases, penetrate between the layers in aluminosilicates
or become incorporated in the crystal lattice.
As to PCA, the variance explained by the first two
PCs was 45 and 22% respectively (67% in all). No cor-
relation among the elements identified as pollutants
exists. The values of the scores and the dendrogram
confirm the different behaviour of the first layer
(Aa) already observed in the extracts from water as well
as the similarity between samples Ac–A26 (from 155 to
238 cm).
3.2.3. Leaching with EDTA
The fraction of metals extracted by EDTA solutions

metals such as Cu, Pb and Zn, extracted by EDTA in
unpolluted soils are much lower (a few percentage
units), because they bind more to the soil matrix,
although the trend of the higher lability of bivalent
metals still exists. Cr, whose total concentration is also
influenced by the work of man, is only weakly released,
probably because of its inertness. Therefore the
observed trend is influenced by the chemical properties
of the single analytes and by their origin.
Table 7
Percentages extracted in acetic acid
a
Sample Al Cd Cr Cu Fe La Mn Ni Pb Sc Ti V Y Zn Zr
Aa 0.319 2.941 0.326 6.966 0.174 < 0.023 0.117 1.777 0.178 0.546 0.092 < 0.306 1.402 2.418 < 0.497
Ab 0.211 10.03 0.115 5.913 0.103 < 0.020 0.538 0.973 0.398 < 0.602 0.055 < 0.361 0.541 4.613 < 0.254
Ac 0.148 4.619 < 0.039 4.941 0.035 < 0.023 0.101 0.625 0.333 < 0.724 0.019 < 0.295 1.010 4.436 < 0.351
A26 0.145 5.125 0.057 5.265 0.033 < 0.021 0.134 1.344 0.354 < 0.619 0.015 < 0.351 0.985 4.784 0.408
A27 0.162 4.913 0.109 5.516 0.134 < 0.026 0.164 0.769 0.469 < 0.613 0.050 < 0.365 0.794 3.614 < 0.492
A28 0.116 6.732 0.108 4.652 0.062 < 0.029 0.155 1.808 0.461 5.087 < 0.004 < 0.398 1.004 4.692 < 0.390
A29 0.138 5.665 0.185 6.562 0.079 < 0.023 0.316 0.831 0.451 < 0.622 0.036 < 0.417 0.964 4.707 < 0.516
A30 0.187 6.562 0.096 6.147 0.082 < 0.016 0.460 1.191 0.390 < 0.612 0.029 < 0.333 1.018 2.603 < 0.364
A31 0.088 10.00 < 0.047 5.869 0.033 < 0.018 0.254 1.578 0.366 < 0.538 0.014 < 0.395 0.801 3.578 < 0.389
Median 0.148 5.665 0.108 5.869 0.079 < d.l. 0.164 1.191 0.390 <d.l. 0.029 < d.l. 0.985 4.436 < d.l.
Range 0.088–0.319 2.941–10.03 < 0.039–0.326 4.652–6.966 0.033–0.174 0.101–0.538 0.625–1.808 0.178–0.469 < 0.538–5.087 < 0.004–0.092 0.541–1.402 2.418–4.784 < 0.254–0.516
a
Aa=samples A17–A19; Ab, samples A20–A22; Ac, samples A23–A25; d.l., detection limit.
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 13
1
2
3

34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63

94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Table 8
Percentages extracted in EDTA
a
Sample Al Cd Cr Cu Fe La Mn Ni Pb Sc Ti V Y Zn Zr
Aa 0.781 5.020 0.752 28.32 0.524 0.764 0.183 2.685 10.18 2.468 0.023 < 0.096 4.953 5.739 0.292
Ab 1.080 14.62 0.765 35.36 0.956 0.723 0.793 2.351 6.130 3.028 0.032 <0.113 4.639 8.047 0.241

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94

Aa and Ab, corresponding to the first layers down to
155 cm, are far from each other, and from the deeper
layers, in the plot of PC1 vs. PC2. The distribution of
variable loadings is different from the one observed with
leaching in water or acetate, due to the different extrac-
tion mechanism involved. Cu, Zn, Mn and Cd are cor-
related, and anticorrelated to Pb. Trivalent elements lay
in one quadrant of the plot (together with Ti and Ni)
whereas the two other elements at high valence state, Zr
and V, are scattered elsewhere. Sample pH does not
seem to have a strong effect on most elements, probably
because extractability mainly depends on the pH of the
acetate buffer. Most variables have positive loadings on
PC2. HCA confirms the different characteristics of the
first layers.
For the data set on site B, the variance explained by the
first two PCs is 38 and 26% respectively (64% at all).
Variable loadings in Fig. 4a are scattered in the plots and
no clear explanation of the existing correlations in terms
of chemical or environmental behaviour exists. pH has
a negative loading on PC1, unlike most of the other
elements.
As to the scores, two main (loose) groups can be
identified, corresponding to the deepest layers (B18–
B22, excluding B19), mainly characterized by higher
pH, and the intermediate ones (B14–B17). The top layer
(B12) stands on its own for the high percentages of
extracted Cu, Pb and Fe.
A different clustering is visible in the dendrogram
(Fig. 4b), characterized by the presence of two main

second fraction are higher, but generally below 5%, one
exception being represented by copper. The third frac-
tion mainly contains Cu, Zn and most Mn, but also
significant amounts of other elements such as Cd, Cr
and Fe. Low percentages (generally < 5%) are present
in the fourth fraction. The highest levels (> 90%) of
several elements are in the residual fraction, which on
the other hand contains only 20% or less of Cu and Mn.
In many cases the order of extractability in the frac-
tions is 1< 2 < 4< 3 < 5, but the fraction associated to
carbonates exceeds the one associated to organic matter
in several samples.
The metals in site B have a higher mobility than at site
A, since significative percentages of elements such as
Fig. 4. Combined plot of scores and loadings obtained by (a) PCA
and (b) dendrogram for vertical profile samples at site B (extracts in
EDTA).
O. Abollino et al. / Environmental Pollution & (&&&&) &–& 15
1
2
3
4
5
6
7
8
9
10
11
12

43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72

103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED
PROOF
Cu, Ni, Pb, Zn are present already in the first fraction.
Similar levels are contained in the fraction associated to
carbonates. The percentages in the third fraction are
generally higher than in the first two and decrease again
in the fourth one. The residual is very high for elements
of mainly geochemical origin, such as Al or Fe, and
lower for pollutants, e.g. Cu, Ni or Zn.
The extractability order is 1< 2< 3 < 4< 5or1< 2<
4< 3< 5 for most elements of mainly geochemical ori-
gin, whereas the pollutants are distributed in the five
fractions in a less regular way; their concentration in the
residue is sometimes lower than the one found in other
fractions, and the content in exchangeable metal is
higher than in fraction 2 or 4.
The element content in the first two fractions is rela-

Cd, Cr, Cu, Ni, Pb, Zn are present at site B. The levels
of some heavy metals at both sites were higher than the
acceptable limits reported in Italian and Dutch legisla-
tions for soil reclamation. The pollution seems to be
limited to the sites (or, in case B, to its immediate sur-
roundings) and does not involve the nearby soil. There
Table 9
Percentages extracted in the five fractions according to Tessier’s procedure
Sample Al Cd Cr Cu Fe La Mn Ni Pb Sc Ti V Y Zn Zr
A9
1st fraction 0.003 2.410 < 0.263 0.285 0.004 < 0.009 0.336 < 0.405 1.551 < 0.473 0.003 < 0.547 < 0.113 2.338 < 1.112
2nd fraction 0.254 4.305 < 0.537 35.74 0.103 < 0.017 1.469 3.196 8.265 0.993 0.009 < 1.093 1.787 16.70 < 2.224
3rd fraction 2.632 28.51 2.818 45.42 14.76 4.254 77.33 4.484 12.70 < 1.892 0.041 12.97 3.480 31.53 < 4.448
4th fraction 0.596 6.658 < 1.051 6.760 0.290 1.642 6.056 1.619 1.752 < 1.892 0.019 < 2.186 1.299 2.269 49.51
5th fraction 96.52 58.12 > 95.33 11.80 84.84 > 94.08 14.81 > 90.30 75.73 > 94.75 99.93 > 83.20 > 93.32 47.16 > 42.71
A20
1st fraction 0.012 1.407 0.110 0.245 < 0.001 < 0.018 1.140 < 0.157 0.483 < 1.250 < 0.006 < 0.687 < 0.261 3.292 < 0.430
2nd fraction 1.181 3.417 0.804 29.74 0.378 1.086 2.782 7.468 2.575 6.125 0.050 < 1.374 7.983 12.10 < 0.859
3rd fraction 14.53 11.29 18.45 42.94 39.93 19.13 71.91 < 0.629 1.560 12.50 0.193 < 2.747 17.88 24.17 < 1.718
4th fraction 4.773 1.585 0.275 6.873 2.191 1.871 4.762 < 0.629 0.330 < 5.000 0.040 3.879 < 1.045 2.164 < 1.718
5th fraction 79.50 82.30 80.36 20.20 > 57.50 > 77.90 19.41 > 91.117 95.05 > 75.13 > 99.71 > 91.31 > 72.83 58.27 < 95.28
B1
1st fraction 0.029 < 4.000 0.024 9.901 0.018 < 0.009 2.430 10.32 11.05 < 0.551 0.007 1.024 0.687 19.44 < 0.584
2nd fraction 0.175 14.74 0.631 18.03 0.237 12.88 0.824 3.652 8.604 < 1.103 0.018 < 1.138 5.302 7.704 < 1.169
3rd fraction 3.131 31.58 24.64 9.276 19.43 12.88 5.908 59.54 39.90 < 2.205 0.031 9.413 16.10 35.46 < 2.338
4th fraction 5.366 29.05 33.45 46.45 1.390 12.13 2.557 16.61 13.53 < 2.205 2.186 12.69 29.12 14.20 < 2.338
5th fraction 91.30 > 20.63 41.26 16.34 78.93 > 62.10 88.28 9.878 26.92 > 93.94 97.76 > 75.74 48.79 23.20 > 93.57
B17
1st fraction 0.053 13.28 < 0.272 6.220 0.178 < 0.008 10.00 43.11 5.366 < 0.421 0.006 < 0.530 1.112 43.36 < 0.556
2nd fraction 0.716 14.26 3.424 12.60 0.899 < 0.016 2.813 17.07 9.194 1.220 0.042 1.059 2.643 15.99 < 1.111

27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
ENPO 2364 Disk used No. pages 17, DTD=4.3.1
Version 7.5
ARTICLE IN PRESS
UNCORRECTED

Bettinelli, M., Baroni, U., Pastorelli, N., 1989. Microwave oven sam-
ple dissolution for the analysis of environmental and biological
materials. Anal. Chim. Acta 225, 159–174.
Bombach, G., Pierra, A., Klemm, W., 1994. Arsenic in contaminated
soil and river sediment. Fresenius J. Anal. Chem. 350, 49–53.
Ferguson, C., Kasamas, H., 1999. Risk Assessment for Contaminated
Sites in Europe, vol. 2. Policy Framework, LQM Press Ed, Notting-
ham.
Gulmini, M., Ostacoli, G., Zelano, V., Torazzo, A., 1994. Comparison
between microwave and conventional heating procedures in
Tessier’s extractions of calcium, copper, iron and manganese in a
Lagoon Sediment. Analyst 119, 2075–2080.
Halloway, B.J., 1990. Heavy Metals in Soils. Blackie, London.
Hani, H., 1990. The analysis of inorganic and organic pollutants in
soil with special regard to their bioavailability. Intern. J. Environ.
Anal. Chem. 39, 197–208.
Jeong, J.H., Song, H.J., Jeong, G.H., 1997. Depth profiles and the
behaviour of heavy metal atoms contained in the soil around a Il-
Kwang disused mine in Kyung Nam. Anal. Sci. Technol. 10, 105–
113.
Li, X., Coles, B.J., Ramsey, M.H., Thornton, I., 1995. Chemical par-
titioning of the new national Institute of Standards and Technology
Standard Reference Materials by sequential extraction using ICP
AE Spectrometry. Analyst 120, 1415–1419.
Lund, W., 1990. Speciation analysis—why and how? Fresenius J.
Anal. Chem. 337, 557–564.
Lund, U., Fobian, A., 1991. Pollution of two soils by arsenic, chro-
mium and copper, Denmark. Geoderma 49, 83–103.
Merian, E., 1991. Metals and their Compounds in the Environment.
VCH, Weinheim.

O. Abollino et al. / Environmental Pollution & (&&&&) &–& 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89