Chemical mineralogical characterisation of clay sediments around ferrara (italy) a tool for an environme

Chemical mineralogical characterisation of clay sediments around ferrara (italy) a tool for an environme

Chemical–mineralogical characterisation of clay sediments around Ferrara (Italy): a tool for an environmental analysis

a

Dipartimento di Scienze della Terra, Universita` di Ferrara, corso Ercole I D’Este 32, 44100 Ferrara, Italy

b

Dipartimento Geomineralogico, Universita` degli Studi di Bari, via E Orabona 4, 70125 Bari, Italy Received 27 October 2000; received in revised form 18 June 2001; accepted 19 July 2001

Abstract

The content of heavy metals in water and soil is a key parameter for evaluating the geochemical vulnerability of an ecosystem These elements display a limited solubility and are easily trapped and adsorbed by phyllosilicate minerals; they are thus preferentially partitioned in the fine fraction of sediments In this light, an analysis of recent river sediments gives information on possible water pollution, and more in general on the related ecosystem We therefore investigated the chemical – mineralogical features of clay sediments outcropping around the town of Ferrara, paying particular attention to their fine fraction (grain size < 2 Am)

X-ray fluorescence (XRF) analyses indicate that the abundance of transition trace elements, such as Cr and Ni, is positively correlated with MgO wt.%, and discriminates two well-delineated populations of samples, respectively characterised by high (Cr > 180 ppm; Ni>100 ppm) and low (Cr < 180 ppm; Ni < 100 ppm) contents of these elements The mineralogical composition of the fine fraction ( < 2 Am) was investigated through X-ray powder diffraction (XRPD) integrated with differential thermal (DTA) and thermogravimetric analyses (DTG), showing that: low-Cr samples are characterised by a higher proportion of clay minerals in which smectite + mixed layers are more abundant than chlorite (Sm + ML/Chl>1); on the other hand, the high-Cr samples have a coarser grain size, and a lower abundance of clay minerals in which chlorite (Mg-rich chlorite in this group of samples) predominates over smectite + mixed layers (Sm + ML/ Chl < 1) These two distinct groups of samples are ascribed to different sources: high-Cr lithologies are related to the sedimentary contribution of the Po river, whereas low-Cr sediments plausibly derive from small rivers of Apennine origin (e.g the Reno river) Within the high-Cr group, concentrations of Ni and Cr tend to be higher than those indicated by the current environmental Italian legislation However, in the study – case presented here, the detected high heavy-metal concentrations are not related to urban – industrial – agricultural activities, but instead appear to be typical of the original lithologies

An integration of similar scientific contributions would be useful to set up a geochemical – mineralogical database as a first step toward the preparation of more complete thematic maps These would provide information relative to the behaviour (e.g distribution and abundance) of chemical elements within the different geochemical spheres, and would be

0169-1317/02/$ - see front matter D 2002 Elsevier Science B.V All rights reserved.

PII: S 0 1 6 9 - 1 3 1 7 ( 0 1 ) 0 0 0 8 6 - 2

* Corresponding author.

E-mail address: gbianch@libero.it (G Bianchini).

www.elsevier.com/locate/clay Applied Clay Science 21 (2002) 165 – 176

useful for recognising and interpreting possible geochemical anomalies induced by pollution processes D 2002 Elsevier Science B.V All rights reserved

Keywords: Clay sediments; Environment; Geochemical monitoring

1 Introduction

The Po river alluvial plain in the province of

Ferrara (Northern Italy) constitutes the terminal part

of the most important drainage basin of the Italian

peninsula, and is a delicate ecosystem The Po river

flows through highly urbanised and industrialised

areas, and consequently its superficial waters,

associ-ated ground waters, and deposited sediments have to

be carefully monitored for pollution

To recognise and interpret possible anthropogenic geochemical anomalies within the soils and sediments outcropping in the area, it is necessary to assess the

‘‘intrinsic baseline’’ (for each monitored parameter) typical of the natural system and related to the chemical characters of the original lithologies (Marini

Fig 1 Simplified map of the area (province of Ferrara) reporting the sampling locations, the estimated ages of sedimentation, and the chemical – mineralogical affinity of the studied fine sediments The characters of the mentioned sample groups (i.e low Cr, high Cr) and the related origin are discussed in the text.

and Ottonello, 1999) The definition of the ‘‘blank’’

for each chemical parameter is critical for

discrim-ination between ‘‘natural’’ anomalies typical of the

deposited sediments and the chemical fingerprint

induced by subsequent human activity In particular,

this study is focused on the chemical – mineralogical

characterisation of sediments outcropping around the

town of Ferrara, paying particular attention to the fine

fraction (grain size < 2 Am) within which some key

chemical species (e.g heavy metals) tend to

concen-trate, frequently exceeding the permitted limits set by

the Italian national environmental regulations The

study approach outlined here is a suitable method

for evaluating the environmental conditions and

investigating contamination and pollution in an area

containing sensitive natural systems, including

impor-tant parks and water resources (e.g the Ostellato

natural oasis and the Po delta nature park)

2 Location of the sampling areas—preliminary

description of the investigated sediments

Fine sediment outcrops were selected for the

sam-pling areas (Fig 1) using a lithological map of Ferrara

(edited by the provincial administration of Ferrara)

Samples were collected using a manual drill at depths

of 70 – 100 cm, to exclude levels containing high

amount of vegetable matter and/or disturbed by

agri-cultural activity

From available geomorphological studies, it is possible to associate these sediments with a fluvial – lagoon – lacustrine environment (Bondesan, 1990; Ste-fani et al., 1999) These geomorphological investiga-tions further define the sedimentary history of the area, enabling the samples to be assigned to three different chronological intervals (Bondesan, personal communication):

1 sediments over 2000 years old (samples 1, 2,

3, 4, 6, 15, and 23);

2 sediments ranging between 2000 and 1000 years old (samples 5, 13, 14, 16, 17, 18, 19, and 24);

3 sediments less than 1000 years old (samples 7,

8, 9, 10, 11, 12, 20, 21, and 22)

Granulometric analysis was performed on a set of these samples (covering the three mentioned sedimen-tation periods and representative of the various sectors

of the area) in a two-stage process Firstly, the samples were wet sieved to separate the coarser particles (>63 Am), then decantation experiments were carried out (gravity settling in deionized water) to characterize the finer fraction The analysis indicated that the clay fraction ( < 2 Am) ranges between 44 and 74 wt.%, the silt fraction (2 – 63 Am) between 25 and 49 wt.%, and that the sand fraction (>63 Am) is subordinate ( < 9 wt.%) (Table 1, Fig 2)

Fig 2 Grain size distribution of the studied sediments Symbols:

n = low-Cr sediments; 5 = high-Cr sediments.

Table 1

Grain size distribution of the studied sediments (wt.%)

Samples Sand

(>63 Am)

Silt (2 – 63 Am)

Clay ( < 2 Am)

x = Average value.

G Bianchini et al / Applied Clay Science 21 (2002) 165–176 167

CO2content analyses were carried out (by simple

volumetric technique; Jackson, 1958) to constrain the

amount of carbonates present The maximum values

were up to 17.8 wt.% in sample 16 (see Table 2),

which indicates that some samples could be better

classified as marly – clay; in this sample, petrographic

microscope analysis indicates that the recorded high

amount of calcite is present as micrite, whereas

detrital calcite (coarse grains/rock fragments) was

not detected

3 Chemical – mineralogical data set

Chemical compositions of the major and trace

ele-ments (reported in Table 3) of the sampled sediele-ments

(tout-venant) were determined by X-ray fluorescence

(XRF) using a Philips PW 1400 spectrometer,

follow-ing the methodology of Franzini et al (1975) and Leoni

and Saitta (1975)

Major elements present the following

composi-tional ranges: SiO2= 47.0 – 54.6%, TiO2= 0.5 – 0.8%,

A l2O3= 1 5 1 – 2 1 0 % , F e2O3t o t = 5 1 – 8 2 % ,

MnO V 0.1%, MgO = 2.5 – 4.7%, CaO = 1.3 – 11.1%,

Na2O = 0.3 – 0.8%, K2O = 2.2 – 3.6%, P2O5V0.3%,

LOI = 9.0 – 17.9 Further examination of these data

(Fig 3) highlights:

(1) a lack of correlation between the SiO2% and

oxides of the other major elements, indicating that the

SiO2% content is mainly related to the abundance of

quartz;

(2) that K2O% and Rb ppm (usually characterised

by similar geochemical behaviour) are positively correlated with Al2O3% (correlation coefficient

r2>0.75), as usually observed in illite-rich fine sedi-ments;

(3) a negative correlation between CaO% and

Al2O3% (r2= 0.60), indicating that CaO is mainly hosted within carbonates, thus precluding a significant presence of CaO-bearing silicates;

(4) a coherent positive correlation between the CaO wt.% and CO2(r2= 0.95) content;

(5) positive correlations between MgO% and tran-sition trace elements, such as Cr (r2= 0.58) and Ni (r2= 0.83); these elements are usually associated in sediments containing chlorite and serpentine, in turn derived from weathering of mother rocks rich in olivine, pyroxenes, and spinel;

(6) the existence of two well-delineated popula-tions of samples discriminated by Cr and Ni (Stu-dent’s t-test: t >10.84, P > 99.99% for Cr; t > 9.24,

P > 99.99% for Ni): these are characterised by high (Cr > 180 ppm; Ni > 100 ppm) and low (Cr < 180 ppm;

Ni < 100 ppm) concentrations of these elements, here-after named as high-Cr and low-Cr groups

To visualise the trace element distribution, the data presented here have been normalised to the composi-tion of fine sediments from the Po river (sampled and analysed by ourselves) In the normalised multi-ele-ment plot of Fig 4, most elemulti-ele-ments show only a limited scattering within the two sample groups On the other hand, the concentration of Ba and Sr (possibly hosted

in carbonates and feldspar) varies widely in both groups

The mineralogical composition of these lithotypes was carried out through X-ray powder diffraction (XRPD; Philips PW1010/80 diffractometer with graphite-filtered CuKa radiation) Particular attention was devoted to investigating the fine fraction of these sediments which, due to the high surface area and the particular nature of the related minerals (mainly clay minerals), tend to trap and concentrate possible pol-lution substances In this light, for a better character-isation of the constituent clay minerals, X-ray dif-fractometric analysis was carried out on the < 2-Am fraction (in which the different clay minerals are more clearly recognised) of each selected sample; in partic-ular, X-ray investigation was carried out on randomly oriented samples, and also on glycolated and

heat-Table 2

CO 2 and CaCO 3 content of the studied sediments

Samples CO 2 CaCO 3

10 5.08 11.55

11 4.73 10.76

14 6.54 14.87

High Cr 2 4.67 10.62

15 5.83 13.26

16 7.82 17.78

x = Average value.

Major (wt.%) and trace element (ppm) analyses of the studied sediments

Pb a

Zn a

x = Average value.

a Semiquantitative analyses.

treated oriented samples (Moore and Reynolds, 1997).

These data were further supplemented with differential

thermal (DTA) and thermogravimetric (DTG) analyses

(Fig 5), kindly performed by M.F Brigatti (University0

of Modena; see Brigatti et al., 1995, 1996 for

metho-dological details) Results indicate that the fine

frac-tion is made up of various proporfrac-tions of illite (Ill),

chlorite (Chl), kaolinite (K), smectite (Sm),

interstra-tified mineral phases (ML: chlorite – smectite and

subordinate kaolinite – smectite), and low percentages

of serpentine, quartz, and carbonate DTA – DTG measurements also indicate the presence of organic matter

A semiquantitative evaluation of the mineralogical composition (within the < 2-Am fraction; see Table 4) was obtained by applying the analytical methods of Schultz (1964) and Shaw et al (1971), modified by Laviano (1987)

Fig 3 Variation diagrams reporting XRF data carried out on the studied sediments: Ni (ppm) and Cr (ppm) vs MgO (wt.%); CaO (wt.%),

Fe 2 O 3 (wt.%), K 2 O (wt.%), and Rb (ppm) vs Al 2 O 3 Symbols: n = low-Cr sediments; 5 = high-Cr sediments.

Illite, is 2M polytype with Al3 + as the main

octahedral cation, K+ as the chief interlayer cation,

and the degree of paragonitization varying from 5% to

25% (Yoder and Eugster, 1955; Bradley and Grim,

1961; Dunoyer De Segonzac, 1970; Srodon and Eberl,

1984) The degree of crystallinity (E) of illite varies

from 150 to 200 A˚ (Weber et al., 1976; Wang and Zhou, 2000)

Smectite (montmorillonite type) is more dominant

in low-Cr samples, with more Ca2 + than Na+ exist-ing as interlayer cations and a medium degree of crystallinity (v/p = 0.6; Biscaye, 1965) In contrast,

Fig 4 Trace element distribution of the studied sediments Data are normalised to the composition of present Po river fine sediments (sampled and analysed by ourselves) around Ferrara (concentration expressed as ppm): Pb = 24, Zn = 91, Ni = 130, Co = 19, Cr = 221, V = 103, Th = 10,

Nb = 11, Zr = 88, Sr = 186, Ba = 306, Y = 15, and Ce = 52.

Fig 5 Representative TG, DTG, and DTA curves Similar analyses are available also for samples 10, 11, 14, and 19.

G Bianchini et al / Applied Clay Science 21 (2002) 165–176 171

high-Cr samples exhibit a very low degree of

crystal-linity (v/p = 0.3)

Mg – Fe-bearing chlorite, with a high degree of

crystallinity, is always present; however, high-Cr

samples are comparatively richer in chlorite that

appear to be distinctively richer in Mg2 +

It is not possible to distinguish the X-ray double

reflections in kaolinite; the crystallinity index reflects

this (Brindley, 1961; Hincley, 1963), and as a result, a

low – medium degree of crystallinity is found in both

sample groups

Interstratified minerals are characterised by 60 –

70% of smectite layers within low-Cr samples, and by

30 – 40% of smectite layers within high-Cr samples

As mentioned, subordinate K/Sm mixed layers are

also present

Summarising, the low-Cr samples are characterised

by a comparatively fine grain size, together with a

high proportion of clay minerals in which

smecti-te + insmecti-terstratified minerals are more represensmecti-ted than

chlorite (Sm + ML/Chl>1) On the other hand, the

high-Cr samples have a coarser grain size, and a lower

abundance of clay minerals in which chlorite (Mg-rich

chlorite in this group of samples) predominates over

smectite + mixed layers (Sm + ML/Chl < 1)

Examination of the >63-Am fraction, using a

trans-mitted light microscope and XRPD analysis, revealed

that these coarse grains are mainly characterised by

quartz, minor amounts of carbonates and feldspars, and lithic fragments containing amphiboles – pyrox-enes, muscovite, biotite, chlorite, and serpentine (the last-named mineral is ubiquitous only in high-Cr sediments), with magnetite as a main Fe oxide in both groups of samples

4 Discussion 4.1 Possible origin and provenance of the studied sediments

To address the implications of the analyses under-taken, and to interpret the significance of the two men-tioned sample groups (high Cr, low Cr), the presented data were compared with the composition of sediments related to rivers of Apennine provenance such as the Reno and Panaro (Dondi et al., 1993; Dinelli and Lucchini, 1998), and also with recent sediments of the Po river (Dinelli et al., 1999 and authors’ data re-ported in Fig 4)

It can be seen that the low-Cr samples show chemical analogies with sediments of rivers sourced from the Bolognese Apennine (in which femic and ultrafemic rocks do not outcrop), whereas the high-Cr samples show a chemical affinity with the Po sedi-ments Within the latter samples, the high Ni, Cr (and V) concentration is related to the high abundance of chlorite ( F serpentine), presumably formed by weath-ering processes of ultrafemic – femic mineral paragen-eses that are widespread in the western sector of the

Po drainage basin where igneous and metamorphic rocks (and also ophiolite complexes) are present Coherently, Tomadin and Varani (1998) indicate the predominance of smectite (typical of the low-Cr samples) as the mineral fingerprint of fine sediments

of Apennine origin

In this light, the tendency toward finer grain size (and higher homogeneity of grain size) envisaged in low-Cr samples can be interpreted considering that they mainly represent reworked sediments, mainly derived from erosion – weathering of sedimentary rocks outcropping in the Bolognese Apennine

It can be observed (in the map of Fig 1) that the samples included in the low-Cr group preferentially outcrop in the southern/southwestern sector of the investigated area, and are more widespread within

Table 4

Mineralogical composition (wt.%) of clay fraction ( < 2 Am)

Samples Cm Sm +

ML Ill K Chl Qtz Fld Cal

Low Cr 3 92 31 28 16 17 5 3 tr

6 92 30 29 17 16 4 2 2

7 91 28 34 11 18 4 2 3

10 90 32 29 18 15 3 2 1

11 95 33 30 16 16 3 2 tr

14 90 30 26 18 16 4 2 4

x 92 31 29 16 16 4 2 2

High Cr 2 81 16 26 11 28 9 4 6

15 83 15 26 13 29 5 4 8

16 82 17 27 12 26 5 3 10

17 87 16 31 13 27 4 3 6

19 87 18 28 11 30 7 4 2

20 91 24 27 14 26 5 4 tr

x 85 18 28 12 28 6 4 5

Cm = clay minerals; Sm = smectite; ML = mixed layers; Ill = illite;

K = kaolinite; Chl = chlorite; Qtz = quartz; Fld = feldspars; Cal =

cal-cite; tr = trace; x = average value.

the younger sediments This fact could be explained

by considering the significant man-made

hydrogeo-logical modifications that occurred in the XIV – XVI

centuries throughout the Ferrarese territory

(Bonde-san, 1990) During this period, important

hydrogeo-logical developments were carried out to divert certain

apenninic torrent – rivers (e.g Reno) into the southern

branches of the Po river (flowing south of Ferrara at

that time) These modifications induced the

progres-sive decadence of these southern Po branches, due to

the deposition of considerable amount of sediments

carried by these torrents The alteration of the

hydro-geological system can explain the widespread

pres-ence of Reno-like compositions (low-Cr) within the

most recent terrains

4.2 Environmental analysis and implications

Trace amounts of heavy metals are ubiquitous

within rocks, soils, surface, and ground water Natural

background concentrations vary from place to place,

owing to different bedrock composition and

hetero-geneous distribution among the various geochemical

environments

Organisms tend to concentrate these elements, and

consequently their presence in waters, even at low

concentrations, is considered dangerous due to their

toxic effects (Hg in aquatic ecosystems is a notorious

example; Jackson, 1998 and references therein) It

follows that, as part of an environmental assessment,

the content of heavy metals in waters and soils is a

key parameter necessary for evaluating the

geochem-ical vulnerability of an ecosystem

Heavy elements usually exhibit a limited solubility,

and are easily trapped and adsorbed by phyllosilicate

minerals; as a consequence, they may be preferentially

partitioned within the fine fraction of sediments (Baldi

et al., 1997) This fact is related to the typical layer –

lattice alluminosilicatic structure of clay minerals,

cha-racterised by well-developed basal cleavage (001

pla-nar faces) with permanent negatively charged sites

Such structures induce electrostatic binding of cations

within the 001 faces (Jackson, 1998 and references

therein)

The fact that heavy elements are preferentially

retained by the fine fraction of river sediments can

be used as a tool to detect pollution of an ecosystem

However, to understand the possible presence of a

pollution process induced by human activity, it is necessary to define the baseline (or ‘‘blank’’) of each monitored key parameter (i.e the typical concentra-tions in an unpolluted environment) Within the con-sidered area, this information can be obtained through

a geochemical characterisation of the sediments that have not been affected by pollution and anthropogenic activities

In this study, the baseline approach enabled us to highlight the concentration of the monitored parame-ters in the natural system, as the studied terrains sedimentated at times when the environment was still unaffected by anthropic activity Our data also indi-cate the geochemical evolution of the Po basin in the last 3000 years

The concentrations of transition elements recorded

in the sediments sampled in the Ferrara surroundings were compared with the maximum concentrations admissible in terrains used for agricultural (Italian Legislative decree 27/01/1992, no 99) and residential (Italian Legislative decree 25/10/1999, no 471) pur-poses

Within the high-Cr group of samples, the concen-trations of Ni, Cr, V, and Co tend to be higher than those indicated by the current environmental legisla-tion However, in the study – case presented here, the high concentrations detected are not associated to urban – industrial – agricultural activities, but instead appear to be typical of the original lithologies (Fig 6) This has been further verified by taking into account an extensive data set (authors’ data: ca 100 samples) of bricks from precisely dated historic build-ings of Ferrara In fact the composition of these bricks show remarkable analogies with the clay composi-tions presented in this study

Moreover, the chemical and mineralogical data presented here provide an initial insight into the elemental partition between competing clay minerals Different mechanisms for trapping solution-electro-lytes by the various clay minerals have been docu-mented, and many authors also underlined existing relationships between clay mineralogy and concentra-tions of metallic elements (Brigatti et al., 1995, 1996; Pitsch et al., 1992; Helios-Rybicka et al., 1995) The implications become apparent when one con-siders that clay (sl) is often used as a barrier to isolate waste disposal sites from the surrounding environ-ment These studies could be applied to defining and

G Bianchini et al / Applied Clay Science 21 (2002) 165–176 173

characterising the most suitable clay materials to be

used as liners and capping layers in landfills, as well

as to treat leachates

Baldi et al (1997) investigated the interaction

between solutions containing metallic ions and natural

clay sediments, thus individuating the key role of

smectite in buffering the abundance of heavy metals

Pusch (1998) confirmed the important role of smectite for landfill-sealing purposes because of its high exchange capacity

Taking into account ion-exchange values reported

in the literature for the various clay minerals (CEC values expressed as meq/100 g at pH 7; Faure, 1998) and the modal proportions reported in Table 4

(refer-Fig 6 Boxplots reporting the compositional distribution of Cr and Ni in the studied samples, in present-day sediments from the Po river and rivers sourced within the Bolognese Apennine, and in the bricks of the historic buildings of Ferrara (plausibly made with similar fine sediments; authors’ data) The limit concentrations admissible in terrains (Italian Legislative Decree 25/10/1999, no 471) have been also reported for comparison.