Development of a simple

phát triển một phương pháp chiết đơn giản để xác định các dạng kim loại nặng trong đất

Development of a simple extraction procedure using

ligand competition for biogeochemically available

metals of estuarine suspended particulate matter

D.-J Whitworth, E.P Achterberg*, V Herzl, M Nimmo, M Gledhill, P.J Worsfold

Department of Environmental Sciences, University of Plymouth, Plymouth, PL4 8AA, UK Received 25 January 1999; received in revised form 8 March 1999; accepted 11 March 1999

Abstract

Sorption of trace metals by suspended particulate matter (SPM) in estuarine systems has important implications for the fate of dissolved metals in these waters This paper describes the development of a single extraction procedure for SPM-associated trace metals, using a ligand competition approach with EDTA as the added complexing ligand The use of EDTA allows the determination of available particulate trace metals using well de®ned constraints with respect to the competition for trace metals between EDTA and the particles Incubation experiments showed that equilibrium times between EDTA and particulate material of 72 h were required to reach equilibrium for most of the metals studied (Cu, Zn, Mn, Ni, Co, Al, Fe, Pb and Mg) Optimum conditions included a 0.05 M EDTA concentration and the use of an extractant: particulate matter ratio of 200 : 1 (v : w) Kinetic calculations on data from the incubation experiments were used to calculate the apparent stability constants (KMeS) for the metal-particulate matter interaction and indicated values ranging from 10ÿ2.1for KMgSto 10ÿ13.5 for KCuS

# 1999 Elsevier Science B.V All rights reserved

Keywords: Extraction; Metals; Suspended particulate matter; Kinetics; EDTA; Apparent stability constant; Estuary

1 Introduction

Trace metal behaviour in estuaries is strongly

in¯u-enced by suspended particulate material (SPM)

Par-ticle-water interactions of trace metals determine

whether they are ¯ushed from an estuary in the

dis-solved phase, or because of adsorption onto particles

are retained within the internal cycle of the estuary [1]

In many estuaries removal of dissolved metal

concen-trations has been observed in the turbidity maximum

zone (TMZ), which is an estuarine region with strongly enhanced SPM levels (e.g >600 mg lÿ1 in the Tamar estuary (U.K.), salinity 0.5±5 psu) [2,3] Desorption of SPM bound metals (e.g Mn and Zn [2] and Cd, Cu and Zn [4]) has been observed in the higher salinity regions of estuaries and has been explained by

an increase in major cation concentrations Suspended particulate matter may consist of biological, organic and mineral phases [5] and each of these phases has a different af®nity for trace metals

Studies involving the determination of trace metals

in SPM and sediments often determine total metal concentrations This approach does not provide

infor-*Corresponding author Tel.: 233-036; fax:

+44-1752-233-035; e-mail: eachterberg@plymouth.ac.uk

0003-2670/99/$ ± see front matter # 1999 Elsevier Science B.V All rights reserved.

PII: S 0 0 0 3 - 2 6 7 0 ( 9 9 ) 0 0 2 8 5 - 8

mation about the biogeochemical availability of the

particulate matter associated trace metals For soils

and sediments, workers have employed sequential

chemical extraction schemes in order to investigate

trace metal association with organic and mineral

phases in their particles Commonly employed

sequential extraction procedures for sediments include

the ®ve step scheme developed by Tessier et al [6] and

the three step BCR scheme [7,8] and variations on

these schemes [9,10] A large number of studies have

been published employing the multi-step sequential

extraction schemes on sediments [11±13] and soils

[14±16]

The sequential extraction procedures have a

number of disadvantages limiting their widespread

use for studies of SPM associated trace metals Firstly,

the procedures often use 1 g of dry material, an

amount which is often dif®cult to obtain for SPM

by ®ltration of natural waters Furthermore, parts of

sequential extraction schemes suffer from

re-adsorp-tion of the extracted metal onto the residual phases

remaining after each extraction step, and a limited

speci®city of the reagents for the targeted phases

of the soil or sediment [17,18] In addition, the

procedures are labour-intensive and because of

the numerous steps involved have enhanced

associated errors and an enhanced risk of sample

contamination

We, therefore, propose the use of a well de®ned

ligand competition procedure for the investigation of

non-lattice bound trace metals associated with SPM

Our approach uses EDTA as the added competing

ligand The use of EDTA for soil and sediment

extraction procedures has been reported by other

workers [16,19] However, little or no work appears

to have been carried out using EDTA extraction

for SPM-associated metals The analytical procedure

for SPM-associated trace metals reported in this

paper complements the ligand competition

techni-ques used in our laboratory for the determination

of trace metal complexation by dissolved organic

ligands in natural waters [20,21] The proposed

extraction scheme allows the application of a well

de®ned binding strength, and provides an indication

of the biogeochemical availability of particle

associated metals The approach is simple, requires

little sample handling, and has a small requirement

of SPM (minimum 15 mg) A total digestion using

HF complements the ligand competition extraction scheme, and allows the assessment of the contribution

of biogeochemically available trace metals to the total particulate metal concentration in SPM

2 Materials and methods 2.1 Reagents and labware All reagents and wash solutions were prepared in water puri®ed by reverse osmosis (RO, Milli-pore) followed by ion exchange (Milli-Q, MilliMilli-pore) Reagents were purchased from Merck and were of AnalaR grade unless otherwise stated Concentrated acids were puri®ed by sub-boiling quartz distillation and NH3was puri®ed through isopiestic distillation

To ensure chemical consistency of the EDTA extrac-tion soluextrac-tions, a single 2.5 l stock soluextrac-tion of 0.5 M EDTA was freshly prepared and used for all the extraction studies The pH of the EDTA stock solution was set at pH 7.6 using an appropriate volume of concentrated NH3 (ca 6.5 ml) Standard solutions utilised for dissolved metal analysis by inductively coupled plasma-mass spectrometry (ICP-MS) were prepared from Spectrosol standard solutions (1000 mg lÿ1 Cu, Ni, Co, Fe, Mn, Pb, Al, Mg and Zn), and acidi®ed to pH 2.2 using concentrated HNO3

(1 ml per 1 ml of solution) When not in use, reagents were stored in high density polyethylene (HDPE, Nalgene) containers at 48C in the dark

Prior to use, all the sample bottles and reagent containers were soaked in 2% (v/v) Decon 90 for

24 h, then washed in copious quantities of Milli-Q water and transferred to a 50% (v/v) HCl bath and left for one week They were subsequently rinsed with Milli-Q water and placed in a 20% (v/v) HNO3bath After a further week, the bottles were thoroughly washed with Milli-Q water and stored inside re-seal-able polythene bags

2.2 Sample treatment Suspended particulate material samples were col-lected during a longitudinal transect in the Scheldt Estuary (Belgium) during a survey with the research vessel Belgica (December 1996) Water samples were collected using 10 l Niskin sampling units deployed at

3 m depth After three rinses with Scheldt water, a

sample of 2.5 l was collected in an acid washed

HDPE container In the ship's laboratory, the water

samples were ®ltered using a polysulfone vacuum

®ltration unit (Nalgene) ®tted with acid washed

(1% HCl), pre-weighed membrane ®lters (47 mm

diameter, 0.45 mm porosity, cellulose nitrate,

What-man) Seawater salts were rinsed from the ®lters

containing SPM by washing with 50 ml of Milli-Q

water The ®lters were dried at 458C for 24 h and

frozen at ÿ178C for transport to the laboratory in

Plymouth In the laboratory, the ®lters were kept for a

further 24 h at 458C and subsequently re-weighed on a

precision balance (Sartorius) The weight of SPM on

the ®lters was calculated as the difference between the

weight of the ®lter containing SPM and the original

®lter

In order to obtain a large quantity of material for

optimising a particulate metal extraction protocol

using EDTA, approximately 300 g of freshly

depos-ited particulate material was collected from the

sedi-ment-water interface at Halton Quay on the Tamar

Estuary (UK) It was postulated that a surface

sedi-ment sample from this locality would closely re¯ect

characteristics of estuarine SPM [3] The sample was

collected from the sediment-water interface (<2 cm

depth) at a water depth of ca 30 cm and transferred

into a re-sealable polythene bag using a HDPE

scrap-per Air was expelled and the bag was resealed and

then stored at 48C for transport to the laboratory In the

laboratory, the particulate material was immediately

air dried at 458C for 48 h In order to achieve a

homogenous ®ne grained material, the dried sample

was crushed into a ®ne powder (<200 mm) using an

acid washed agate mortar and pestle, subsequently

placed in a polythene bag and left for 24 h on a

motorised end-to-end shaker (Baird and Tatlock,

UK) operating at 40 rpm

The organic carbon (OC) content in the Halton

Quay particulate material was determined in triplicate

by the loss on ignition method [22] For this purpose

acid was added to the material (10 ml of 1 M HCl to

1 g of particulate matter) to remove inorganic carbon

(mainly calcium carbonate), and subsequently the

material was dried at 1058C until constant weight

was achieved after cooling in a dessicator (weight

A) Subsequently, the material was ashed in a muf¯e

furnace at 6008C for 8 h, cooled in a dessicator and

weighed (weight B) The organic carbon content was calculated as:

2.3 Use of EDTA as extractant The interaction between an added chelating ligand and metals complexed by naturally occurring ligands

in the aquatic environment may be slow Experiments involving competition for dissolved Cu in seawater between an added metal chelating ligand (salicylal-doxime) and naturally occurring dissolved organic ligands have indicated that the establishment of an equilibrium may take more than 8 h [23] In order to investigate the rate of interaction between EDTA and particle bound trace metals we performed incubation experiments The aim of these studies was to establish the minimum time required for the EDTA-particulate metal extraction procedure

The incubation experiments were performed using

2 different ratios of extraction solution to particulate matter (200 : 1 and 2000 : 1 (v : w)), in order to investigate the in¯uence of particulate matter concen-tration on the extraction ef®ciency Furthermore, two EDTA concentrations (0.005 M and 0.05 M) were investigated Table 1 includes the experiments that were undertaken The pH of the EDTA solution was set at 7.6, which is close to the natural pH observed in large parts of estuarine systems The pH of the EDTA extraction solutions was determined after the experi-ments to ensure that dissolution of material from the particles did not modify the pH of the experiments No change in pH was observed at the end of the EDTA extraction experiments

Fig 1 gives a schematic representation of the par-ticle extraction procedures applied during all extrac-tion experiments Samples were agitated during the incubation period using an end-to-end shaker set at

40 rpm A centrifuge (Sanyo, Centaur 2, 3000 rpm for

10 min) was utilised to separate the extraction solution from the particulate material The supernatant was acidi®ed to pH ca 2 using concentrated HNO3(1 ml per 1 ml of solution) to avoid loss of metals onto the wall of the bottles, and then stored at 48C prior to metals analysis by ICP-MS All experiments were carried out in triplicate using separate fractions of

freshly deposited particulate material collected from

Halton Quay or SPM from the Scheldt

2.4 Other extraction protocols

Commonly employed single extraction protocols

for marine SPM and sediment particles were utilised

to allow comparison with the ef®ciency of the EDTA

protocol The protocols applied included the extrac-tion of metals from particulate material using 1 M HCl for 6 h [24±26], 25% acetic acid at pH 4.5 for 16 h [1],

1 M ammonium acetate at pH 7 for 6 h [19] and 0.05 M EDTA at pH 7.6 for 72 h The extraction procedures are summarised in Table 1 and Fig 1 and were applied to certi®ed reference materials (CRMs), BCR 320 (riverine sediment) and BCR

Table 1

Procedures used for particle extraction experiments using particulate material from Halton Quay with EDTA, HCl, ammonium acetate and acetic acid as extractants

Experiment Reagent Ratio extractant:

particle (v : w) Dilution for metalsanalysis by ICP-MS Incubationtime (h) Metals determined

3 0.05 M EDTA 200 : 1 10 18, 24, 36, 48, 60, 72 Cu, Co, Ni, Zn, Fe, Al, Pb, Mn, Mg

4 0.05 M EDTA 2000 : 1 10 18, 24, 36, 48, 60, 72 Cu, Co, Ni, Zn, Fe, Al, Pb, Mn, Mg

Fig 1 Schematic representation of the procedure adopted for particulate metal extraction experiments.

277 (estuarine sediment) supplied by the Commission

of the European Communities, Community Bureau of

Reference Furthermore, a total digestion of the CRMs

was performed using HNO3 and HF, following a

method adapted from Rantala and Loring [27], to

verify the accuracy of the extraction and analytical

procedures The total digestion was also applied to the

SPM from the Scheldt Estuary For the total digestion,

150 mg of certi®ed reference material, or a

pre-weighed ®lter with a known amount of SPM (between

30 and 500 mg) was placed into a 30 ml PTFE

decom-position vessel, 10 ml of concentrated HNO3 was

added and the vessel was put on a hotplate (1208C),

for a re¯ux period of 24 h Subsequently, 2 ml of

concentrated HF was added and the re¯ux was

con-tinued for a further 48 h at 1208C The vessel was then

uncovered and 4 ml of HNO3added and the content of

the vessel was evaporated to dryness on the hotplate

(1208C) 10 ml of concentrated HNO3was added and

the content of the vessel was evaporated to dryness at

708C; this step was repeated twice Then 10 ml of

0.1 M HNO3was added to the vessel and the solution

was transferred into a 25 ml HDPE volumetric ¯ask

containing 0.93 g of H3BO3 The volumetric ¯asks

were made up to volume using 0.1 M HNO3and stored

in the refrigerator at 48C for subsequent metal analysis

by ICP-MS The total digestion of the CRMs was

performed in triplicate Procedural blanks were

pro-cessed and used for correction of particulate metal

data

2.5 Trace metal analysis by ICP-MS

The concentration of metals (Cu, Ni, Co, Fe, Mn,

Pb, Al, Mg and Zn) in the supernatant after

centrifu-gation was determined by ICP-MS using a VG

Elemental PQ2 Turbo instrument (Winsford,

Cheshire) The spectrometer was ®tted with an

Ebdon high solids `V' groove nebuliser (Ar gas ¯ow

set at 0.9 l minÿ1) connected to a Scott double pass

spray chamber (Ar coolant, 15 l minÿ1) and the

plasma gas ¯ow was ®xed at 1 l minÿ1 Samples

were introduced into the manifold at 1 ml minÿ1

Analytes were ionised at 1350 W and the detection

dwell time was 10.24  10ÿ3s Fe measurements

were determined from the 57Fe concentration, as

56Fe could not be directly determined due to

poly-atomic interferences (ArO) Sample solutions

contain-ing EDTA were diluted with Milli-Q to EDTA con-centrations of 0.005 M, in order to avoid interferences associated with EDTA For experiments involving EDTA, metal standards were matrix matched to the sample by preparing the standards in 0.005 M EDTA Calibration was undertaken immediately prior to sam-ple analyses In addition, the samsam-ples and standards were spiked with115In (100 mg lÿ1) in order to correct analytical drift during the operation of the spectro-meter

3 Results and discussion 3.1 EDTA extraction studies Experiments 1 and 2 (see Table 1) were designed to compare the effect of the EDTA concentration and the incubation time on the extraction ef®ciency for 0.005 M and 0.05 M EDTA at a ratio of extraction solution to particulate material of 200 : 1 (v : w; 30 ml EDTA solution with 150 mg of particulate material) Fig 2 shows the concentrations of Fe, Cu, Zn, Co, Ni, and Mg extracted using EDTA and normalised with respect to the particulate matter concentration, plotted against time for these experiments The maximum incubation periods used for the EDTA extractions were 48 and 72 h, respectively The results suggest that after 48 h the extracted concentrations of Fe, Mg and Cu were lower in the 0.005 M EDTA extraction solution compared with the 0.05 M solution The difference was small and within the experimental error for Zn, Co and Ni Furthermore, metals were extracted more rapidly and a plateau in metal concentration was reached earlier using the 0.05 M ETDA compared with the 0.005 M EDTA Therefore, the equilibrium between metals complexed by EDTA, and those asso-ciated with particulate matter, was reached more rapidly using the 0.05 M EDTA solution The use

of a lower EDTA concentration (0.005 M) will, there-fore, extract a lower amount of Fe, Mg and Cu from particulate material, but will also require a longer incubation time to attain equilibrium, compared to a higher EDTA concentration (0.05 M)

The in¯uence of the concentration of particulate material on the metal (Me) extraction ef®ciency was investigated by employing 2 different ratios of extrac-tion soluextrac-tion (0.05 M EDTA) to particulate material:

200 : 1 and 2000 : 1 (v : w; 30 ml of EDTA solution

with 150 mg and 15 mg of particulate material,

respectively) In addition to the metals measured

during the previous experiment, Mn, Pb and Al were

also determined The incubation period was 72 h in

order to allow more time for the attainment of

equili-brium (experiments 3 and 4, Table 1) The results of

this experiment are shown in Figs 2 and 3, and

indicate that increasing the extractant to particle ratio

from 200 : 1 to 2000 : 1 (v : w) had no effect on the particulate matter-normalised MeEDTA concentra-tions, with the differences between MeEDTA concen-trations within analytical errors (with the exception of

Al (Fig 3(c)) This observation may be explained by the use of an excess concentration of EDTA during the experiments, of which a large fraction was not com-plexed to metal ions The different behaviour observed for Al may be attributed to the fact that an equilibrium

Fig 2 Concentration of particulate metal (Fe, Cu, Zn, Co, Ni, Mg) extracted using 0.05 M EDTA with an extraction solution to particulate matter ratio of 200 : 1 and 2000 : 1 (v : w), and 0.005 M EDTA with a ratio of 200 : 1, plotted against time Legend key presented on Fig 2(a), and number between brackets refers to experiment (see Table 1) Solid curve obtained from model calculations.

for Al between EDTA and sorption sites on the

particles is attained only very slowly (see below)

3.2 Modelling of EDTA kinetic incubation

experiments

Experiments 1±4 indicated that the competition

between EDTA and particle bound metals was not

instantaneous (see Figs 2 and 3) The data from

experiments 1 and 3 (0.05 M EDTA extraction

solu-tion to particle ratio of 200 : 1 (v : w)) were, therefore,

used to model the interaction between EDTA and

metals

The competition between the EDTA and the surface

sites of the particles (S) for the trace metals can be

described using Eq (2) The concentration of metal

associated with particles is denoted by [MeS] and the

EDTA complexed metal concentration by [MeEDTA]

The reaction can be characterised by two rate

con-stants: k1for the forward and k2for the reverse reaction

[28] The time dependent linear differential equation for reaction (2) is expressed by Eq (3), asuming that the EDTA concentration used is in excess of the metal concentrations, and, therefore, is constant We also use the assumption that the concentration of S is much greater than the concentration of MeS, and that an increase in concentration of S with time is negligible MeS ‡ EDTA $k1

d‰MeEDTAŠ

Using the assumption that at t ˆ 0 the amount of the metal complexed with EDTA is zero, the solution to

Eq (3) is

‰MeEDTAŠ ˆk k1

1‡ k2‰MeSŠtˆ0

ÿk k1

1‡ k2eÿ…k 1 ‡k 2 †t‰MeSŠtˆ0 (4)

Fig 3 Concentration of particulate metal (Mn, Pb and Al) extracted using 0.05 M EDTA with an extraction solution to particulate matter ratio

of 200 : 1 and 2000 : 1 (v : w) Legend key presented on Fig 3(a), and number between brackets refers to experiment (see Table 1) Solid curve obtained from model calculations.

With the use of curve ®tting software

(CurveFitEx-pert 1.3) and an exponential function of the form

y ˆ a(1ÿebx), the concentration of [MeS]t ˆ 0 was

estimated as being the concentration of [MeEDTA]

at equilibrium Subsequently, a computer programme

(written in Turbo-Pascal) was employed to calculate

the rate constants (k1and k2(hÿ1)) using Eq (4) and

the data obtained from experiments 1 and 3 The

curves obtained from the model calculations are

pre-sented in Figs 2 and 3 Table 2 shows the results of

calculations of the minimum time required for the

different elements to attain equilibrium with the

extraction solution A commonly used approach to

assess the state of equilibrium is the characteristic

reaction response time (tresp), which is de®ned as the

time required to achieve 63% of the equilibrium

concentration (or the time to reduce the imbalance

to eÿ1 (37%) of its initial imbalance) [29] We also

calculated the minimum times required to achieve

95% (t95%) and 100% (t100%) of the equilibrium

con-centration Furthermore, the estimated equilibrium

concentrations of [MeEDTA] for t100% determined

using this approach are presented in Table 2

Figs 2 and 3 show that the time required to obtain

95% (t95%) of the equilibrium concentration was 30 h

or less for all metals except Al (Table 2) For this

element we observed a slow desorption process that

can be attributed to (a) the release of Al from binding

sites due to the competitive action of EDTA, and (b)

the slow dissolution of Al from the lattices of clay

particles The slow release of Al is most likely

respon-sible for its t95%value of 149 h The calculated t100%

for Al was greater than 3000 h, and for Cu and Mn

greater than 400 h Extractions over this period of time are practically impossible, and would most likely result in analytical artefacts including re-adsorption

of metals on particles and walls of the sample con-tainer and perhaps bacterial alteration of the metal speciation An extraction time of 72 h for the experi-ments involving EDTA was, therefore, chosen as the optimum condition, because most of the metal extrac-tions reached a state of at least 95% of their equili-brium within this time period

The kinetic calculations allowed us to determine the interaction between the metals and the sites on the surface of the particles For this purpose we separated

Eq (2) into two reactions MeS $ Me ‡ S KMeSˆ‰SŠ‰Me‰MeSŠ0Š (5) where KMeSis the apparent stability constant, and EDTA ‡ Me $ MeEDTA

K0

Me 0 EDTA 0 ˆ ‰MeEDTAŠ

where K0

Me 0 EDTA 0is the conditional stability constant [EDTA0] is the concentration of EDTA not complexed

by Me, and can be taken as [EDTA], as [EDTA] [Me] The conditional stability constant for Eq (2) can be written as

and is the product of the apparent and conditional stability constants of the separate reactions

K2ˆ KMeSK0

Table 2

Treshold times and equilibrium concentrations calculated using the kinetic model Conditional and apparent stability constants (defined in Eqs (5)±(7)) were obtained from the literature (log K 0

Me 0 EDTA 0 ) [28] and model calculations (log K 2 and log K MeS ) Metal t resp

(h) t(h)95% t(h)100% eq conc.(mg g ÿ1 ) LogK 0 Me 0 EDTA 0 Log

K 2

Log

K MeS

The conditional stability constants for the

metal-EDTA reactions (K0Me0EDTA0) were obtained from the

literature [28] and corrected for the side reactions of

EDTA and Me at pH 7.6 and ionic strength of 0.1 M

(Table 2) The desorption of metals (Me) from surface

sites (S) is simply described by our model as a

competition between MeS and the sum of the metal

EDTA complexes In practice, only one, or at most

two, metal-EDTA complexes will be involved in the

equilibrium The predominant species can be assessed

by comparing the a-coef®cients that describe the

reactions between the metal ion and the EDTA species

[30] under the given conditions of pH, temperature

and ionic strength Only the most important MeEDTA

complexes have been used for our calculations and

are included in Table 2 For most metals considered,

the important complex is MeEDTA, although for Al it

is Al(OH)EDTA However, for Fe and Zn both

MeEDTA and Me(OH)EDTA are important In order

to simplify the calculation, the predominant complex

was used in this study At pH 7.6 FeEDTA and

ZnEDTA formed the predominant species (60 and

90% of the total metal-EDTA complexes,

respec-tively), and these were used for further calculations

However, it must be noted that this simpli®cation will

result in less certainty for calculations of KFeS and

KZnS

K2can be calculated from the forward and reverse

reaction rate constants (k1and k2) [28,31]

K2ˆk1

and KMeS can then be derived using Eq (8) The

results are presented in Table 2 and show that KMeS

generally follows the Irving±Williams order

Accord-ing to this rule, the stability of metal complexes

increases in the series [29] Mn2‡< Co2‡<

Ni2‡< Cu2‡> Zn2‡ The approach used with EDTA

as extractant releases metals into solution from sites

with a lower `binding strength' than that of the EDTA

ligand The amount of MeEDTA extracted, therefore,

relates to the strength of the MeEDTA complex In the

case of Fe and Al, not just the partition between solid

and dissolved phases is important, but also the

dis-solution of solid phases, as precipitation and

coagula-tion strongly in¯uence their solid speciacoagula-tion

Potentially useful information on metal partitioning

between solid and dissolved phases may be gained by

comparing apparent stability constants obtained using the method presented here and conditional stability constants measured for naturally occuring dissolved organic material (K0Me0L0; de®ned using Eq (10) MeL $ Me ‡ L K0

MeLˆ‰L‰Me0Š‰Me0L00ŠŠ (10) For example, the current study indicates that the apparent stability constant for Cu (KCuSˆ 10ÿ13.5) of the particulate matter from the sediment-water inter-face at Halton Quay, compares well with reported constants for dissolved organic ligands and Cu:

10ÿ10±10ÿ14 for lacustrine, seawater and estuarine conditions [21,32,33] This observation may imply that organic ligands on particles are important for the complexation of Cu; sequential extraction schemes

on sediments have reported similar ®ndings [18,34] The Halton Quay particulate material contained an important fraction of OC: 1.01  0.05% In natural waters competition may, therefore, occur for Cu between particle surface sites and dissolved Cu com-plexing natural ligands In the case of Zn, reported conditional stability constants for the dissolved ligands and Zn (between 10ÿ7.4and 10ÿ9.3) are some-what higher than for KZnS(10ÿ10.9) Conditional sta-bility constants for the interaction between dissolved organic ligands and Pb in seawater are of the order

10ÿ8±10ÿ9 [35] The lower value for KPbS (10ÿ12.1) would suggest that estuarine particulate matter may actively remove dissolved Pb from solution A particle reactive behaviour has been observed for dissolved Pb

in estuarine systems [36] Conditional stability con-stants for dissolved ligands and Fe in coastal and oceanic conditions are reported to be ca 10ÿ18±

10ÿ23 [31,37] The apparent stability constant (10ÿ5.1) for FeS is subject to error (see above), how-ever the large difference in stability constant between the dissolved organic ligands and particulate matter for Fe suggests that processes other than complexation (i.e precipitation, coagulation) determines the beha-viour of Fe in estuarine systems

Further work will need to be undertaken to relate the apparent stability constants to different types of SPM (e.g riverine, marine, estuarine) This may allow us to link changes in apparent stability constants to different physico-chemical characteristics of SPM Further-more, the fraction of MeS released by the ligand competition approach will give information about

the geochemical availability of the metals Inclusion in

water quality models of conditional stability constants

for the metal-dissolved ligand interactions and the

apparent stability constants for metal SPM

interac-tions may also result in important improvements in our

ability to model contaminant behaviour in natural

waters

3.3 Certified reference materials: total digests and

comparison of EDTA extraction with other

single extractants

Certi®ed reference materials BCR 320 (riverine

sediment) and BCR 277 (estuarine sediment) were

analysed for total particulate Pb, Zn, Fe, Cu, Co, Ni,

Al, Mg and Mn, after digestion using HF and HNO3

(see Section 2.4) The results of the analysis are

pre-sented in Table 3 The data show that the analysed and

certi®ed values were in close agreement

Further experiments were performed using the BCR

320 and BCR 277 sediments in order to compare the

concentration of exchangeable metals extracted from

these sediments using 0.05 M EDTA with other

com-monly used single extraction procedures (1 M HCl,

25% (v : v) acetic acid and 1 M ammonium acetate;

experiments 5±8, Table 1) Figs 4 and show the

results of these experiments with the extracted metal

concentration presented as a fraction of the total metal

concentration obtained after total digestion of the

sample

Fig 4 shows that for most metals (with the

excep-tion of Al and Zn), 1 M HCl extracted a higher fracexcep-tion

of particulate metal in BCR 320 compared with the other extractants (e.g Co ˆ 44%, Ni ˆ 50% and

Fe ˆ 36%) This observation indicates that metals were not easily removed from the riverine sediment using mild extraction procedures at pH values between 4.5 and 7.6 The low pH of the HCl extraction solution, may have resulted in the dissolution of carbonate phases in the sediment particles, and hence released matrix bound particulate metals The fraction of par-ticulate metals extracted from the BCR 320 sediment using the 0.05 M EDTA (pH 7.6) and the 25% v : v acetic acid (pH 4.5) extraction solutions were similar (Fig 4) For most metals, the lowest extraction yield was obtained using 1 M ammonium acetate

The fractions of metal extracted using 1 M HCl were generally similar for BCR 277 compared with BCR 320 (Fig 5) However, the other extractants showed higher yields for BCR 277 The particulate metals in the BCR 277 estuarine sediment, therefore, appeared to be present in a more available form compared with BCR 320 As was the case for BCR

320, the ammonium acetate extractions with BCR 277 resulted in the lowest yield Furthermore, the 0.05 M EDTA and 25% acetic acid extractions resulted again

in comparable yields

3.4 Application to SPM from the Scheldt Estuary The 0.05 M EDTA extraction protocol was used to investigate the extractable metal concentrations of SPM in surface water samples from the Scheldt Estu-ary An incubation time of 72 h was employed and the

Table 3

Analysis of total particulate metal concentration in certified reference materials BCR 277 and BCR 320

Certified concentration a Observed concentration Certified concentration Observed concentration a

a Indicative values (uncertified values provided by BCR); N/A ˆ analysis not undertaken; mean  SD (n ˆ 3).