Mercury in contaminated sediments and pore waters enriched insulphate (Tagus Estuary, Portugal)

J. Canario*, C. Vale, M. Caetano, M.J. Madureira

IPIMAR, Institute for Fisheries and Sea Research, Av. Brasılia, 1449-006, Lisbon, Portugal

Received 3 December 2002; accepted 6 June 2003

overlying water.

‘‘Capsule’’: Excess sulfate in sore water favours the binding of mercury into iron oxides, limiting escape into the

Abstract

Three sediment cores, collected nearby the effluent of a chlor-alkali industry, were sliced in 0.5-cm layers and centrifuged for porewater extraction. Mercury, Fe and Mn were determined in the solids as total concentration, hydroxylamine extractable fraction and

HCl extractable fraction. Sulphur was determined in the HCl extraction. Total and reactive mercury, chlorinity, S2�, SO42�, total

Fe, and total Mn were measured in pore waters. The solids contained 3.0–60 nmol g�1 of total Hg and pore waters 70–5800 pM oftotal Hg and 1.8–76 pM of reactive mercury. Pore waters presented 2.3–94 times more sulphate than the overlying estuarine watersdue to the input from the industry. In layers where hydroxylamine extractable Fe exhibited a broad maximum (precipitation of Fe-

oxides) sulphate was reduced to S2�. The competition between the high content of SO42� and Fe(III) as electron acceptors, in che-

mical reactions occurring in the upper sediments, may explain the co-existence of S2� and Fe-oxides in the same layers. Mercurywas detected in the hydroxylamine extracts (20–29 nmolg�1) in the layers where Fe-oxides were formed, and reactive dissolved Hg

showed minimum concentrations. The excess of sulphate in pore waters favoured the abundant Fe-oxides in the upper solid sedi-ments, which appear to work as a barrier limiting the escape of mercury to the water column.# 2003 Elsevier Ltd. All rights reserved.

Keywords:Mercury; Fe-oxides; Sulphate; Sediment; Tagus Estuary

1. Introduction

A large number of studies have reported the mercurydistribution in sediments, water and organisms of con-taminated areas (eg. Coquery et al., 1995; Bjerregaard etal., 1999). Mercury in sediments is mainly associatedwith organic matter (Mantoura et al., 1978) and withsulphur compounds (Drobner, 1990; Morse and Luther,1999). In several cases the vertical profiles of totalmercury have been related to the historical evolution ofmercury contamination in the area (Gobeil and Cossa,1993; Pereira et al., 1998). Although the availability ofmercury in contaminated sediments is of great concern,due to the toxicity of its organic forms (Covelli etal., 1999), only a few studies have approached thediagenesis and mobility of mercury in coastal con-

taminated sediments (Bothner et al., 1980; Gobeil andCossa, 1993; Gagnon et al., 1997). By determining thedistribution and speciation of mercury in both dissolvedand solid phases these works have showed that mercuryappears to be bound mainly to organic matter. Only aminor fraction is recycled with the Mn and/or Fe-oxidesnear the redox boundary, adsorbed to or co-precipitatedwith acid volatile sulphides and some incorporated inpyrite.The North Channel of the Tagus Estuary, which is 14

km long, 750 m wide and a maximum depth of 2 m,receives the discharge of a chlor-alkali industry. Highlevels of mercury are recorded in surface sediments,suspended particles, water and organisms (Figueres etal., 1985; Ferreira et al., 1997; Canario, 2000). Sedi-ments of this channel incorporate anthropogenic mate-rials associated with suspended particulate matter thatsettles (Vale, 1986; Ferreira, 1995) due to the lowhydrodynamics (Nunes, 1993). This paper reports thevertical distribution of mercury in sediments and pore

0269-7491/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/S0269-7491(03)00234-3

Environmental Pollution 126 (2003) 425–433

www.elsevier.com/locate/envpol

* Corresponding author. Tel.: +351-213027191; fax: +351

213015948.

E-mail address: jcanario@ipimar.pt (J. Canario).

waters of this Channel and examines the association ofmercury with Fe/Mn oxihydroxides in a sedimentaryenvironment characterized by excess of sulphate fromthe industrial effluent.

2. Materials and methods

2.1. Sampling and measurements in loco

Three cores of 10-cm were collected in mudflats in theNorth Channel (Fig. 1): cores A and B close to theindustrial effluent and core C in the opposite side of thechannel. The cores were sampled by hand and rapidlysliced in loco in layers of 0.5-cm thickness. Sampleswere stored in leak proof tubes that were completelyfilled up and sealed in order to avoid sulphide oxidation.Oxygen levels were measured in overlying water and inthe first 1-cm layer using a Diamond Electro-Tech Inc.needle electrode following the method described inBrotas et al. (1990). Redox potential and pH were mea-sured in each layer of duplicated cores. The pH wasdetermined with a glass combine electrode (Mettler)calibrated (20 �C) with 4.0, 7.0 and 10.0 buffers. Redoxpotential was measured using a platinum combine redoxelectrode (Ingold) calibrated with a solution of 220 mV

(vs. Ag/AgCl, 25 �C) and converted to hydrogen normalpotential.

2.2. Sediment analysis

In the laboratory, samples were centrifuged at 3000rpm for 20 min. (4 �C) and the supernatant filteredthrough 0.45 mm membranes in a N2 chamber. Thisprocedure was used in previous works and errors asso-ciated with iron and sulphur determinations were lessthan 6% (Madureira et al., 1997). The obtained porewaters were divided in two aliquots: one was stored in10 ml tubes for sulphide and chloride analysis and theremainder being acidified to pH<2 with HNO3 fordeterminations of sulphate and ‘‘dissolved’’ metals. Thesolid fraction was oven dried at 40 �C, desegregated,homogenised and stored in polyethylene bottles forfuture analysis. Duplicates were frozen for determina-tion of the water content and loss on ignition (LOI).The water content was estimated by weight loss at105 �C and LOI at 450 �C.

2.2.1. Total analysis of solid sedimentTotal determinations of Al, Si, Ca, Mg, Fe and Mn

were performed bymineralization of the sediment sampleswith a mixture of acids (HF, HNO3 and HCl) accordingto the method described by Rantala and Loring (1975).Metal concentrations were obtained by flame-AAS(Perkin-Elmer Aanalist 100) using direct aspiration intoN2O-acetylene flame (Al, Si, Ca and Mg) or air-acety-lene flame (Fe and Mn). For total Hg determinations,sediment was digested overnight at room temperaturewith 4 M HNO3 in borosilicate glass erlenmeyers andthen heated (60–70 �C) for 2 h in a sand bath (Pereira etal., 1998). Mercury was determined by cold vapourAAS (Perkin-Elmer FIMS-FIAS-100) using SnCl2.2H20as reduction agent. The precision expressed as relativestandard deviation was less then 4% for all metalinvestigated (P<0.05). International certified stan-dards were used to ensure the accuracy of our proce-dure and precision was determined by analysingreplicate samples.

2.2.2. Extraction with hydroxylamine chlorideEach sediment sample was stirred for 6 hours with a

NH2OH.HCl (0.04 M) solution in CH3COOH (25%)according to themethod described by Chester andHughes(1967). The extraction was performed in leak proof tubessealed with Parafilm to avoid possible losses of Hg0. Thesupernatant solution was removed by centrifugation at3000 rpm for 10 minutes and filtered through 0.45 mmmembranes. Iron and Mn in the overlying solution weredeterminated by flame-AAS and Hg by CV-AAS asdescribed above. Detection limits for Fe, Mn and Hgwere 0.80, 0.60 mmol g�1 and 0.02 nmol g�1 and preci-sion errors were 7, 6 and 3% (p<0.05), respectively.

Fig. 1. Map of the Tagus Estuary with the location of the North

Channel and sampling sites (A, B and C).

426 J. Canario et al. / Environmental Pollution 126 (2003) 425–433

2.2.3. Extraction with 1 M HClAcid volatile sulphides, mainly amorphous iron sul-

phides and poorly crystallised Fe-oxides were extractedwith 1M HCl (Luther et al., 1991; Henneke et al., 1991).Sulphide was trapped in a de-aerated NaOH solutionand analysed by Differential Pulse Polarography (DPP)using a Metrohm apparatus equipped with a 693 VAProcessor and a 694 VA Stand. Iron, Mn and Hg weredetermined in the extracted solution by flame-AAS (Feand Mn) and CV-AAS (Hg) as described above.Recovery of standard sulphide solutions was 97%(Madureira et al., 1997). Detection limits for sulphide,Fe, Mn and Hg were 0.10, 0.14, 0.36 mmol g�1 and 0.04nmol g�1 and precision errors were 5, 2, 3 and 6%(p<0.05), respectively.

2.2.4. Determination of pyrite and elemental sulphurPyrite was determined in 100 mg sediment sample

using the chromium reduced sulphur (CRS) methoddescribed by Canfield et al. (1986). This method hasproven to be specific for inorganic sulphur (AVS+-FeS2+S

0) with an accuracy that is not affected by thepresence of organic sulphur. Before analysis, S0 wasextracted from the sample by 16-h stirring with 20 cm3

of acetone followed by centrifugation (3000 rpm/10min) and filtration through 0.45-mm membranes. Theresidue was then placed in the reaction vessel with 10cm3 HCl 1M and purged with N2 for 20 min to releaseAVS. Finally the CRS method was used in the last resi-due to analyse the pyrite content. Elemental sulphurwas determined using the same CRS method in theacetone extracts. The measurements of the released H2Swere made by DPP.

2.3. Pore water analysis

Pore water chlorinity was determined using an argento-metric method (Paustian, 1987). Concentrations of sul-phate were determined by turbidimetry using a HitachiU-2000 Spectrophotometer providing a light path of 5-cm(Rossum and Villarruz, 1961). Total dissolved inorganicsulphide, [HS�]t (� [H2S]+[HS

�]+[S2-]+[Sx2�]), was

measured by differential pulse cathodic stripping vol-tammetry (DPCSV) (Luther et al., 1985; Luther andTsamakis, 1989) using the Metrohm apparatus descri-bed above. Total dissolved Fe and Mn were determinedby AAS using direct aspiration into air-acetylene flame.Reactive dissolved mercury was measured directly fromthe filtered acidified solutions by cold vapour atomicfluorescence spectroscopy (CV-AFS) using a coldvapour generator PSA, model 10.003 associated to afluorescence detector PSA Merlin 10.023. This mercuryfraction includes forms of inorganic-complexes andweakly bound organic-complexes easily reducible bySnCl2 (NOAA, 1996). Total dissolved mercury was alsomeasured in the same equipment after a UV oxidation

step with a 1000 W UV lamp following the methoddescribed by Mucci et al. (1995). The detection limits forCl�, SO4

2�, [HS-]t, Fediss, Mndiss and Hgdiss were 21, 2.1mM, 0.10, 0.20, 0.11 mM and 0.01 nM and precisionerrors were 3.6, 10, 5, 2.3, 5.0 and 4.0% (p<0.05),respectively.

3. Results

3.1. Water content, LOI, chlorinity, pH and redoxpotential

Visual inspection of the sediment slices showed nomacro fauna or signs of bioturbation. Samples consistedof fine particles and sand was virtually absent. Watercontent in the three sediment cores was higher in thefirst 2-cm (70–85%) and decreased gradually to 40–45%with the depth. Vertical variation of LOI showed asimilar pattern: a gradual decrease with the depth from13–18 to 8.5% in cores B and C and from 5.5 to 3.5% incore A. Chlorinity increased with the depth in core A(41–386 mM) and core B (270–402 mM), reflecting thedensity stratification and the presence of freshwater(core A) from the industrial effluent. Otherwise core Cexhibited a slight increase of chlorinity in the uppersediment layers (186–334 mM). Dissolved oxygen inoverlying water were around 100% saturation (240mM) and concentrations decreased sharply in the firstmillimetre of the sediment (approximately 120 mM)being undetectable below 2-mm depth. Values of pHwere around 8.0 in core A, 7.0 in core B and in core Cincreased from 7.0 to 8.0 with the depth. The first0.5-cm sediment showed higher pH values in all cores.Redox potentials were positive in the topmost layers(50–100 mV) and decreased gradually with depth to�300 mV.

3.2. Major and minor elemental composition of solidsediments

The ranges of Al, Ca and Mg concentrations and ofSi/Al, Fe/Al and Mn/Al ratios are presented in Table 1.In core C, these parameters were relatively constantwith the depth, while in cores A and B Ca, Mg and Fevaried with the depth. The higher values of Ca, Mg, Si/Al and Fe/Al ratios were found in core A, the Fe/Alratio showing a sharp increase at 4-cm depth. The Mn/Al ratios presented an increase in the upper sedimentlayers of cores B and A.

3.3. Sulphate, sulphide, iron and manganese in pore waters

Sulphate concentration was higher in pore watersfrom upper layers of the three cores and decreased pro-nouncedly with the depth (Fig. 2). However, the levels

J. Canario et al. / Environmental Pollution 126 (2003) 425–433 427

and the depth at which sulphate decreased differed con-siderably in the cores. Sulphate concentration in core Cdecreased almost linearly from 23 to 2.4 mM within thefirst 3 cm, and sulphide increased downward to 8.8 mM.Pore waters of the first 2.5-cm of core B contained 112mM of sulphate and presented an abrupt decrease to 7.1mM around 3-cm depth. These levels exceeded largelythe sulphate concentration of the estuarine water in theNorth Channel (4.7 mM for salinity 29.8). Besides thehigh concentration of sulphide in the upper layers (100mM) values increased sharply to a maximum of 4800mM in deeper sediments. The situation in core A wasequally unusual for estuarine sediments: sulphatereached similar high levels (84 mM) in the 6 cm sedi-ment depth. An additional unexpected situation is theco-existence of sulphide (422 mM) and sulphate (76 mM)peaks. Dissolved Fe and Mn in pore waters of upperlayers of the cores A and B were lower than 12 and 25mM respectively, while core C exhibited a classic profileof sequential maximum values in the topmost layersexceeding 100 mM.

3.4. Acid volatile sulphides (AVS), pyrite and elementalsulphur

The AVS in the three cores varied irregularly withdepth. In cores A and C concentrations ranged between2.6–154 and 78–250 mmol g�1, respectively, while in core

B reached 737 mmol g�1. Depth profiles of pyrite incores A and B presented similar ranges (86–396 mmolg�1) with concentrations increasing with depth, andcore C exhibited a broad peak (940 mmol g�1) at 6-cmdepth. Vertical profiles of elemental sulphur in cores Aand B showed increases from 50 mmol g�1 at surface to275 mmol g�1 at 10 cm depth. A different pattern wasobserved in core C: low concentrations (13–32 mmolg�1) in the first 2 cm; increase to 500 mmol g�1 at 4-cmdepth; variable in deeper layers.

3.5. Mercury profiles in the solid sediments

The total mercury concentrations differed in the threeanalysed cores (Fig. 3). The higher levels were recordedin core A (23–60 nmol g�1) and in core B that presenteda broad maximum of 39 nmol g�1 between 2 and 4-cmdepth. Mercury concentration in core C was lower than15 nmol g�1 and showed no vertical trend in the first 10-cm depth. The vertical profiles of iron and mercuryextracted with the hydroxylamine solution are presentedin Fig. 4. Iron concentrations were higher in the upperlayers (3–4 cm) with values exceeding 700 mmol g�1

whereas in core C values were lower than 60 mmol g�1.A sharp peak of Hg-hydroxylamine was detected in thecores A and B, reaching 20 and 29 nmol g�1, respec-tively. The peaks were formed at 2.5-cm depth, corre-sponding to the lower depth where Fe-hydroxylamine

Table 1

Ranges of Al, Ca and Mg concentrations (mmol g�1) and Si/Al, Fe/Al and Mnx10�4/Al ratios in cores A, B and C

Cores

Al (mmol g�1) Ca (mmol g�1) Mg (mmol g�1) Si/Al Fe/Al Mn�10�4/Al

A

3.7–4.7 2.9–4.2 0.6–1.1 0.9–1.1 0.2–0.4 2.5–3.4

B

2.7–4.4 0.9–2.6 0.5–0.9 0.4–0.7 0.1–0.2 2.7–5.1

C

3.7–4.5 1.6–2.4 0.6–0.7 0.3–0.8 0.1–0.2 2.5–3.4

Fig. 2. Vertical profiles of sulphate (mM) and sulphide (mM) concentrations in sediment pore waters of cores A, B and C.

428 J. Canario et al. / Environmental Pollution 126 (2003) 425–433

was higher. As iron decreased with depth, Hg-hydro-xylamine was undetectable. In core C the peak wasabsent and levels were below detection limit (0.02 nmolg�1). In core C, Mn-hydroxylamine decreased almostlinearly in the first centimetre from 3 to 0.2 mmol g�1,and remained relatively constant thereafter. Verticalprofiles in the other cores varied irregularly with depth.

3.6. Mercury profiles in pore waters

The vertical profiles of reactive mercury in porewaters were different in the three cores (Fig. 5). Cores Aand B showed successive increase and decrease of mer-cury in the first 10-cm depth, resulting in broad intervalconcentrations, between 9.3 and 76 pM and from 1.8 to37 pM, respectively. Levels were more uniform in coreC (2.6–21 pM). Total dissolved mercury showed similarvalues in the three cores ranging between 0.07 and 5.8

nM. Reactive mercury concentrations in the cores B andC parallel to those observed for total mercury. A lessclear similarity was found for core A.

4. Discussion

The ability of iron oxihydroxides to scavenge traceelements is well documented (Jenne, 1968; Farrah andPickering, 1978). The scavenger of Hg has been recor-ded in a few aquatic environments presenting high con-tent of Mn and Fe-oxides in the upper sediments, likethe Saguenay Fjord (Gagnon et al., 1997), LaurentianTrough (Gobeil and Cossa, 1993) and Laurentian GreatLakes (Matty and Long, 1995). This sequesteringemerged as a relevant process in the biogeochemicalcycle of mercury in those environments limiting theirflux to water column. The results obtained in our study

Fig. 3. Vertical profiles of total mercury concentrations (nmol g�1) in sediment cores A, B and C.

Fig. 4. Vertical profiles of Fe (mmol g�1) and Hg (nmol g�1) concentrations simultaneously extracted with an hydroxylamine solution in sedimentcores A, B and C.

J. Canario et al. / Environmental Pollution 126 (2003) 425–433 429

are in line with those findings, Hg being incorporatedinto the abundant Fe-oxides layers in upper sedimentsnearby the discharge of the mercury enriched effluent.The driving factor for this incorporation appears to bethe extremely high levels of sulphate in sediment porewaters.The sulphate in pore waters from sediments nearby

the industrial effluent may derive from aluminium sul-phate used as coagulator in the plant sewage treatment.On the basis of sulphate and chlorinity data the excesssulphate (�SO4

2�) was calculated by subtracting themeasured sulphate concentration in pore waters fromthe sulphate value expected for pore water chlorinity:

DSO2�4 ¼ SO2�4� �

measured � Cl�½ �=19:64

The value 19.64 represents the molar ratio of Cl� toSO4

2� in seawater (Culkin, 1965). The vertical profiles ofexcess sulphate obtained for the three sediment coresare presented in Fig. 6. Clearly, the excess sulphate

(max. 94 mM) was observed in the first 3–6 centimetrelayers of the cores A and B, closer to the effluent dis-charge. In core C, collected in the other shoreline of theNorth Channel, only a slight enhancement was regis-tered in the first centimetre layer. Excess sulphate canchange dramatically the thermodynamic sequence ofelectron acceptors for organic matter oxidation (Froe-lich et al., 1979), increasing the reaction kinetics invol-ving the pair SO4

2�/S2�. Consequently, reactionsinvolving other electron acceptors with free energy clo-ser to that of sulphate, like Fe(III)/Fe(II), tend to beenergetically less favourable.The segregation of different terminal electron accept-

ing processes in separate zones during the degradationof organic matter was one of the fundaments in under-standing redox processes in nature (e.g., Berner, 1981;Stumm and Morgan, 1996). These conditions are notalways found as demonstrated, for example, by Postmaand Jakobsen (1996) showing that the presence of awide range of iron-oxide stabilities in natural sedimentshas a strong influence on whether Fe(III) or sulphatereduction is the most favourable. As a consequence aconsiderable overlap between zones of Fe(III) and sul-phate reduction may occur. In cores A and B, contain-ing high sulphate concentration, large amounts of Fe-oxides (700–800 mmol g�1) co-exist with sulphides (up to400 mM) in the same upper sediment layers. This meansthat sulphate is reduced to sulphide in sediment layerswhere Fe(III) has not been considerably consumed,which suggests that sulphate reduction is superimposedto Fe(III) reduction. These specific conditions favourthe presence of a thicker layer enriched in Fe-oxides.Below this layer, the abundant Fe-oxides and sulphateare consumed in the same sediment layers, as can beinferred from the abrupt decrease in their concentra-tions. This situation is hardly observed in freshwater(Postma and Jakobsen, 1996) and seawater environ-ments because sulphate in pore waters occurs in much

Fig. 5. Vertical profiles of total (nM) and reactive (pM) dissolved mercury in sediment pore waters of cores A, B and C.

Fig. 6. Vertical profiles of excess sulphate (mM) in sediment pore

waters of cores A, B, and C.

430 J. Canario et al. / Environmental Pollution 126 (2003) 425–433

lower concentrations. The overlap between zones ofsulphate and Fe(III) reduction in the Tagus study arearesults from the fact that SO4

2- concentrations were 100times higher than natural values, due the discharge ofan effluent enriched in sulphate. Because sulphate wasnot in excess in core C, sulphide was produced afterconsumption of Fe-oxides and the sediment layers enri-ched in Fe-oxides is much thinner. Low levels of dis-solved Fe in pore waters from cores A and B reinforcesthe hypothesis of low mobility of Fe(II). The possibilityof interactions between HS-, Fe-oxides and AVS in theupper sediment layers (Canfield, 1989; Lovley, 1991;Burdige, 1993) should not be excluded.Extraction of solid sediments with hydroxylamine-

acetic acid solution evidences the effect of Fe-oxides inthe mercury diagenesis. Simultaneous extraction of highlevels of Hg and Fe in the upper sediment layers of coresA and B suggests that Fe-oxides act as a sink of mobilemercury forms. The association of mercury with Fe-oxides and hydroxides has been considered in otherstudies (Gobeil and Cossa, 1993; Matty and Long,1995), although the possibility of mercury being alsoassociated with hydrous manganese oxides highlyabundant in sediments of Gulf of St. Lawrence has notbeen excluded (Gobeil and Cossa, 1993). However, thislast hypothesis should not be taken into account in ourstudy area because the enhancement of Mn-hydro-xylamine is negligable. The association of mercury withiron in the cores A and B was restricted to the lowerpart of Fe-oxides layers, near the suboxic-anoxic inter-face defined by sharp decreases of Fe-oxides and sul-phate and concomitant increase of sulphide. The depthof this mercury-enriched layer suggests an intense recy-cling of mercury near that redox boundary. As thoseelectron acceptors are used in the degradation oforganic matter, mercury is released from solids to porewater, and diffused upward towards the Fe-oxide enri-ched layer. The hydroxylamine method could alsoextract from sediments some amorphous iron sulphides(Anchutz et al., 1998). However, undetectable levels ofHg associated with AVS in the entire sediment column,including the Hg-hydroxylamine enriched layers, indi-cate a minor role of AVS in mercury incorporation insediments from this study area. Mercury is thereforeefficiently scavenged in Fe-oxides through sorption and/or co-precipitation reactions, as documented for a widerange of other trace elements (eg. Jenne, 1968). Theretention processes are confirmed by the inverse corre-lation (r2=0.85) between reactive mercury in porewaters and mercury associated with Fe-oxides (Fig. 7).These results strongly indicate that Fe-oxides are anefficient barrier for diffusion of mercury towards thesediment surface. It is noticeable that the mass of Hgassociated with Fe-oxides was similar in the sites A andB (11.7 and 10.3 nmol dm�2), despite the differentamount of total mercury in the solids. This barrier effect

has been recorded at the sediment-water interface ofother sediment environments (Gobeil and Cossa, 1993;Matty and Long, 1995) limiting the flux of mercury tothe water column.In addition to the upward diffusion, it should be

expected that mercury mobilised at the suboxic/anoxicinterface is also diffused downward. However, processesrelated to its incorporation in solids are less clearlydefined. In fact, sediments are highly sulphidric, butconcentration of reactive mercury in pore waters wasrelatively high (1.8–76 pM) namely in comparison to thevalues predicted from the thermodynamic solubility ofHgS (Lindberg and Harriss, 1974; Stumm and Morgan,1996). These labile forms accounted only for a max-imum of 8% of the total dissolved mercury, concentra-tions reaching 5.8 nM. The elevated concentration ofmercury in pore waters may be due to the formation ofpolysulphide as well as organic complexes (Dyrssen,1985; Paquete and Helz, 1995). Hence, mercury ispoorly associated with AVS (HCl extraction), as indi-cated by the undetectable levels found in the sedimentcolumn. Due to the high pyrite concentration theincorporation of mercury in these sulphur mineralshould be considered. Indeed, Cooper and Morse(1998a and b) have demonstrated that Hg rapidly formspersistent sulphide minerals that are poorly soluble inHCl, preferentially being incorporated into pyrite. Inaddition, the degree of trace metal pyritization (DTMP)for mercury is one of the highest in all trace metals(Morse and Luther, 1999). Due to the high affinity ofmercury to organic matter, association with the organicfraction of the sediment should also be considered (eg.Baldi and Bargagli, 1984; Mucci and Edenborg, 1992).

5. Conclusions

These results evidence the incorporation of Hg inthe abundant Fe-oxides present in sediment that are

Fig. 7. Relationship between reactive dissolved mercury (pM) and

mercury simultaneously extracted with the hydroxylamine solution

(nmol g�1) in cores A and B; one value (o) not considered.

J. Canario et al. / Environmental Pollution 126 (2003) 425–433 431

extremely enriched in sulphate due to the discharge ofan industrial effluent. The sequestering of mercury inthe upper sediment layers limits its diffusion to thewater column. The use of aluminium sulphate for thesewage treatment appears to extend its action to thenearby sediment environment.

Acknowledgements

The authors which to thank the colleagues EduardaPereira, Monica Valega, Pedro Brito and Joana Rai-mundo for the help on the analytical and field work.

References

Anschutz, A., Zhong, S., Sundby, B., Mucci, A., Gobeil, C., 1998.

Burial efficiency of phosphorus and the geochemistry of iron in

continental margin sediments. Limnol. Oceanog. 43, 53–64.

Baldi, F., Bargagli, R., 1984. Mercury pollution in marine sediments

near a chlor-alkali plant: distribution and availability of the metal.

Sci. Total Environ. 39, 15–26.

Berner, R.A., 1981. Authigenic mineral formation resulting from

organic matter decomposition in modern sediments. Fortschr.

Miner. 59, 117–135.

Bjerregaard, S.R., Bruty, D., Chester, R., Padgham, R.C., 1999.

Retention of methyl-mercury and inorganic mercury in rainbow

trout Orcorhynchus mykiss (W): effect of dietary selenium. Aquat.

Toxicol. 45, 171–180.

Bothner, S.S., Jahnke, R.A., Peterson, M.L., Carpenter, R., 1980.

Rate of mercury loss from contaminated estuarine sediments. Geo-

chim. Cosmochim. Acta 44, 273–285.

Brotas, V., Amorim-Ferreira, A., Vale, C., Catarino, F., 1990. Oxygen

profiles in intertidal sediments of Ria Formosa (S. Portugal).

Hydrobiol. 207, 123–129.

Burdige, D.J., 1993. The biogeochemistry of manganese and iron

reduction in marine sediments. Earth Sci Rev. 35, 249–284.

Canario, J., 2000. Mercurio em Sedimentos Contaminados e Aguas

Intersticiais da Cala do Norte do Estuario do Tejo. Master thesis,

New University of Lisboa, June, 98 pp.

Canfield, D.E., Raiswell, R., Westrich, J.T., Reaves, C.M., Berner,

R.A., 1986. The use of chromium reduction in analysis of reduced

inorganic sulphur in sediments and shale. Chem. Geol. 54, 149–155.

Canfield, D.E., 1989. Reactive iron in marine sediments. Geochim.

Cosmochim. Acta 53, 619–632.

Chester, R., Huges, M.J., 1967. A chemical technique for the separa-

tion of ferromanganese minerals, carbonate minerals and adsorbed

trace metals from pelagic sediments. Chem. Geol. 2, 249–262.

Cooper, D.C., Morse, J.W., 1998a. Extractability of metal sulfide

minerals in acidic solutions: application to environmental studies of

trace metal contamination within anoxic sediments. Environ. Sci.

Tech. 32, 1076–1078.

Cooper, D.C., Morse, J.W., 1998b. Selective extraction chemistry of

toxic metal sulfides from sediments. Aquat. Geochem. 5, 87–97.

Coquery, M., Cossa, D., Martin, J.M., 1995. The distribution of dis-

solved and particulate mercury in three Siberian Estuaries and

adjacent artic coastal waters. Water Air Soil Pollut. 80, 653–664.

Covelli, S., Faganeli, J., Horvat, M., Brambati, A., 1999. Pore water

distribution and benthic flux measurements of mercury and methyl-

mercury in the Gulf of Trieste. Est. Coast Shelf Sci. 48, 415–428.

Culkin, F., 1965. The major constituents of sea water. In: Skirrow, J.P.

(Ed.), Chemical Oceanography. Academic Press, London, pp. 121–

161.

Drobner, E., Huber, H., Wachterhauser, G., Rose, D., Steller, K.,

1990. Pyrite formation linked with hydrogen sulphide evolution

under anaerobic conditions. Nature 546, 742–744.

Dyrssen, D., 1985. Metal complex formation in sulphidic seawater.

Mar. Chem. 15, 285–293.

Farrah, H., Pickering, W.F., 1978. The sorption of mercury species by

clay minerals. Water Air Soil Pollut. 9, 23–31.

Ferreira et al., 1997. Nova travessia rodoviaria sobre o Tejo em Lis-

boa. Monitorizacao da qualidade dos sedimentos de fundo. IPI-

MAR Technical Report, December, 62 pp.

Ferreira, J.G., 1995. An object-oriented ecological model for aquatic

ecosystems. Ecol. Mod. 79, 21–34.

Figueres, G., Martin, J.M., Meybeck, M., Seyler, P., 1985. A com-

parative study of mercury contamination in the Tagus Estuary

(Portugal) and major French Estuaries (Gironde, Loire, Rhone).

Est. Coast Shelf Sci. 20, 183–203.

Froelich, P., Klinkhammer, G., Bender, M., Luedtke, N., Heath, G.,

Cullen, D., Dauphin, P., Hammond, D., Hartman, B., Maynard, V.,

1979. Early oxidation of organic matter in pelagic sediments of the

east equatorial Atlantic: Suboxic diagenesis. Geochim. Cosmochim.

Acta 43, 1075–1091.

Gagnon, C., Pelletier, E., Mucci, A., 1997. Behaviour of anthro-

pogenic mercury in coastal marine sediments. Mar. Chem. 59, 159–

176.

Gobeil, C., Cossa, D., 1993. Mercury in sediments and sediment pore

water in the Laurentian Trough. Can. J. Fish Aquat. Sci. 50, 1794–

1800.

Henneke, E., Luther, G.W., De Lange, G.J., 1991. Determination of

inorganic sulphur speciation with polarographic techniques: Some

preliminary results from recent hypersaline anoxic environments.

Mar. Geol. 100, 115–123.

Jenne, E.A., 1968. Controls of Mn, Fe, Co, Ni, Cu and Zn con-

centrations in soils and water: The significant role of hydrous Mn

and Fe-oxides. In: Gould, R.F. (Ed.), Trace Inorganics in Water.

Advanc Chem Ser 73. Am Chem Soc, Washington DC, pp. 337–387.

Lindberg, S., Harriss, R.C., 1974. Mercury organic matter associ-

ations in estuarine sediments and interstitial water. Environ. Sci.

Tech. 8, 459–462.

Lovley, D.R., 1991. Dissimilatory Fe(III) and Mn(IV) reduction.

Microbiol. Rev. 55 (2), 259–287.

Luther, G.W., Giblin, A.E., Varsolona, R., 1985. Polarographic ana-

lysis of sulphur species in marine porewaters. Limnol. Oceanog. 30,

727–736.

Luther, G.W., Tsamakis, E.J., 1989. Concentrations and form of dis-

solved sulfide in the oxic water column of the ocean. Mar. Chem.

27, 165–177.

Luther, G.W., Ferdelman, T.G., Kostka, J.E., Tsamakis, E.J.,

Church, T.M., 1991. Temporal and spatial variability of reduced

sulphur species (FeS, S2O32-) and porewater parameters in salt marsh

sediment. Biogeochem. 14, 57–88.

Madureira, M.J., Vale, C., Simoes Goncalves, M.L., 1997. Effect of

plants on sulphur geochemistry in the Tagus salt-marshes sediments.

Mar. Chem. 58, 27–37.

Mantoura, R.F.C., Dickson, A., Riley, J.P., 1978. The complexation

of metals with humic materials in natural waters. Est. Coast Mar.

Sci. 6, 387–408.

Matty, J.M., Long, D.T., 1995. Early diagenesis of mercury in the

Laurentian Great Lakes. J. Great Lake Res. 21, 574–586.

Morse, J.W., Luther, G.W., 1999. Chemical influence on trace metal-

sulphide interactions in anoxic sediments. Geochim. Cosmochim.

Acta 63, 3373–3378.

Mucci, A., Edenborn, H.M., 1992. Influence of an organic-poor land-

slide deposit on the early diagenesis of iron and manganese in

coastal marine sediment. Geochim. Cosmochim. Acta 56, 3909–

3921.

Mucci, A., Lacotte, M., Montgomery, S., Plourde, Y., Pichet, P., Tra,

N., 1995. Mercury remobilization from flooded soils in a hydro-

432 J. Canario et al. / Environmental Pollution 126 (2003) 425–433

electric reservoir of Northern Quebec. La Grande-2: results of a soil

resuspension experiment. Can. J. Fish Aquat. Sci. 52, 2507–2517.

NOAA, 1996. Contaminants in aquatic habitats at hazardous waste

sites: Mercury. Technical Memorandum NOS ORCA 100, 64 pp.

Nunes, L.M., 1993. Distribuicao de metais pesados nos sedimentos da

Cala do Norte do Estuario do Tejo. ECOTEJO report A-8402-01-

93. New University of Lisbon, Lisbon.

Paquette, K., Helz, G., 1995. Solubility of cinnabar (red HgS) and

implications for mercury speciation in sulfidic waters. Water Air

Soil Pollut. 80, 1053–1056.

Paustian P., 1987. A novel method to calculate the Mohr chloride

titration. In: Advances in Water Treatment, Standard Methods Proc

14th Annu AWWA Water Quality Technology Conf, 16–20

November, 1986, Portland, Ore., USA, pp. 673.

Pereira, M.E., Duarte, A.C., Millward, G.E., Abreu, S., Vale, C.,

1998. An estimation of industrial mercury stored in sediments of a

confined area of the lagoon of Aveiro (Portugal). Water Sci. Tech.

36 (6–7), 125–130.

Postma, D., Jakobsen, R., 1996. Redox zonation: equilibrium con-

straints on the Fe(III)/SO42- reduction interface. Geochim. Cosmo-

chim. Acta 60, 3169–3175.

Rantala, R.T.T., Loring, D.H., 1975. Multi-element analysis of silicate

rocks and marine sediments by atomic absorption spectro-

photometry. Atom Absorp. Newsletter 14, 117–120.

Rossum, J.R., Villarruz, P., 1961. Suggested methods for turbidimetric

determination of sulphate in water. J. Am. Water Works Assoc. 53,

873.

Stumm, W., Morgan, J.J., 1996. Aquatic Chemistry. John Wiley &

Sons, Canada.

Vale, C., 1986. Transport of particulate metals at different fluvial and

tidal energies in the Tagus River Estuary. Rapp P.-V Reun Cons.

Int. Expl. Mer. 186 306–312.

J. Canario et al. / Environmental Pollution 126 (2003) 425–433 433