speciation and bioavailability of trace elements in contaminated soils sébastien sauvé université...

Speciation and Speciation and Bioavailability of Trace Bioavailability of Trace

Elements in Contaminated Elements in Contaminated SoilsSoils

Sébastien Sauvé

Université de Montréal (Montréal, QC, Canada)

email:sebastien.sauve@umontreal.cahttp://mapageweb.umontreal.ca/sauves/

© Sauvé 2003

ObjectivesObjectives Determine the free metal speciation of divalent

metals in soil solutions

Identify the physico-chemical characteristics of the soil which control metal solubility and speciation

• Quantify the contributions of pH, total metal and organic matter

Propose simple semi-mechanistic regression models to estimate metal solubility and free Me2+ speciation in contaminated soils

Link within the context of a large data acquisition projet

© Sauvé 2003

SoilsSoils Multiple dataset of field-collected soils

Metals originating from smelting/battery recycling operations, long-term phosphate fertilizers, aerial deposition, sewage sludge application, diffuse and point source industrial contamination

• Montréal (QC), Ithaca (NY), Québec, France, Denmark & Colorado

Field « equilibrium », in most cases, contamination has occured at least ten years before sampling

Uncontaminated « controls »

© Sauvé 2003

Soil PropertiesSoil Properties Soil pH in 0.01 M CaCl2 or KNO3 extract (from

3.5 to 8.9)

Soil organic matter of 8.0 to 108 g C kg-1

Dissolved organic carbon 1.1 to 140 mg C L-1

Metal levels from background to high industrial range

• Soil totals of 0.1 to 56 mg Cd kg-1

• Dissolved Cd of 0.03 to 3500 µg Cd L-1

• Free Cd2+ of 10-10 to 10-5 M

© Sauvé 2003

Analytical MethodologyAnalytical Methodology «Totals» by HNO3 reflux digestion

Soil solutions obtained using 1:2 soil:0.01 M KNO3 or CaCl2 extractions filtered to <0.22µm (or <0.45µm)

• Total dissolved metal by GFAAS (Zeeman) or ICP-AES

• Electrochemically labile Cd, Pb and Zn by differential pulse anodic stripping voltammetry (DPASV)

• Free Cd2+ Pb2+, Zn2+ speciation by partitioning ASV-labile metal into inorganic ion-pairs

• Free Cu2+ by ion-selective electrode potentiometry

© Sauvé 2003

Electrochemically

active metals are

reduced into the

Hg drop electrode

Each metal has a

specific reduction

potential, peak

position identifies

metal, peak height

is proportional to

its concentration

Differential Pulse Anodic Stripping Differential Pulse Anodic Stripping VoltammetryVoltammetry

© Sauvé 2003

Calibration by comparison of known standards with samples

y = 0.0119x - 0.0002

R2 = 0.9985

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0 10 20 30 40 50 60 70

DPASV Peak Height (nA)

ASV-

Labi

le Cd

(µM)

Differential Pulse Anodic Stripping Differential Pulse Anodic Stripping VoltammetryVoltammetry

© Sauvé 2003

Free CdFree Cd2+2+ Speciation Speciation

Assuming that ASV is not sensitive to metals strongly complexed with dissolved organic matter

ASV-labile Cd is composed mainly from inorganic species

)CdCdClCdNO)Cd(COCdCO

CdHCOCd(OH)Cd(OH)(CdOH ASV)by Cd Labile(2++

3-2

2303

-3

+3

02

+

Cd) LabileASV()CdBoundOM(CdDissolved

© Sauvé 2003

CuCu2+2+ by potentiometry by potentiometry

Ion selective electrode very sensitive for Cu2+

Not prone to interferences (except very high levels of chloride or mercury)

y = -0.0299x + 10.107R2 = 0.9948

2.0

4.0

6.0

8.0

10.0

12.0

14.0

-1000100200300

Electrode potential (mV)

Free

Cu

(pCu

2+)

© Sauvé 2003

Fractionation/SpeciationFractionation/Speciation

Soil total

Bound to DOM

Free metal

Cl complexes

SO4 complexes

1 mg Cd kg1 mg Cd kg-1-1

pH ~ 5pH ~ 5

MineralMineral Solubility Solubility EquilibriaEquilibria

3 4 5 6 7 8 9pH

0

2

4

6

8

10

12

p (a

ctiv

ity)

CdOH2

CdCO 3

CdSO4·2Cd(OH)2

Cd3(PO4)2

3 5 7 9

0

4

8

123 5 7 9

0

4

8

123 4 5 6 7 8 9

pH

0

2

4

6

8

10

12

Cu(OH) 2

CuO

Cu4(OH)6SO4

CuCO 3

Cu3(PO4)2·H2O

3 5 7 9

0

4

8

123 5 7 9

0

4

8

123 4 5 6 7 8 9

pH

0

2

4

6

8

10

12

Pb2(CO)2(OH)2

PbOPb(OH)2

PbSO4PbHPO4

Pb5(PO4)3OH

Pb5(PO4)3Cl

3 5 7 9

0

4

8

123 5 7 9

0

4

8

12

Sauvé S. 2002. «The Role of Chemical Speciation in Bioavailability » In: Naidu R., Gupta V.V.S.R., Kookana R.S., Rogers S., Adriano D. (Eds.),

Bioavailability, Toxicity and Risk Relationships in Ecosystems. Science Publishers Inc., Enfield, NH, pp 21-44.

© Sauvé 2003

Solid/liquid PartitioningSolid/liquid Partitioning

• Assumes a unique and constant ratio between solution and solid phases:

• Total metal is in mg/kg dry soil and dissolved metal is in mg/L, hence Kd´s are usually reported as L/kg

• Sensitive to determination method, solid:liquid ratio, extracting solution, time of extraction and filtration

MetalDissolvedMetalTotal

Kd

© Sauvé 2003

Dependence of KDependence of Kdd on pH on pH

For a compilation of literature Kd’s, 29 to 58 % of the variability depends on soil solution pH.

Soil Solution pH

2 4 6 8 10

Soil Solution pH

2 4 6 8 10

Soil Solution pH

2 4 6 8 10

Soil Solution pH

2 4 6 8 10

Soil Solution pH

2 4 6 8 10

Kd(L

kg

-1)

10-1100101102103104105106107

Cd Cu Ni

Pb

Zn

Sauvé S. Hendershot W., Allen H.E. 2000. «Solid-Solution Partitioning of Metals in Contaminated Soils: Dependence on pH, Total Metal and Organic Matter  ». Environ. Sci. Technol. 34:1125-1131 .

© Sauvé 2003

Dissolved Cd - KDissolved Cd - Kdd Partitioning Partitioning(Field-collected soils only)(Field-collected soils only)

0.101.00

10.00

Soil Total Cd (mg/kg)

10

100

1000

10000

100000

Kd (

kg/L

)

A

3 4 5 6 7 8 9Soil Solution pH

3 4 5 6 7 8 93 4 5 6 7 8 93 4 5 6 7 8 9

B

1 10Dissolved OM (mg C/L)

C

Janssen et al. 1996 Data Lee et al. 1996 Anderson and Christensen 1988

© Sauvé 2003

Dissolved CdDissolved Cd

Total Cd

pH

Field & spiked datasets are similar at pH<8

KOH effect on DOM at pH>8 Field

Spiked

TYPE

© Sauvé 2003

Predictive RegressionsPredictive Regressions

Field-collected dataset

Field & spiked soils (pH<7).).(,64,001.0,759.0

)(log)07.0(77.0

)04.0(54.0)28.0(23.3)41.0)((log

2

10

10

ESnpR

CdSoilTotal

pHCdDissolved

.).(),7(,70,001.0,861.0

)(log)06.0(07.1

)05.0(57.0)28.0(54.3)40.0)((log

2

10

10

ESpHnpR

CdSoilTotal

pHCdDissolved

© Sauvé 2003

Adsorption ModelAdsorption Model

Assuming competitive binding of H+ and Me2+ to a deprotonated surface (S):

HyMeSurSurHMe y

© Sauvé 2003

Adsorption ModelAdsorption Model With some assumptions, then:

Assuming that adsorption capacity is dependent on organic matter content:

But could be oxyde content, clays, sulfides, etc.

)log(

) log(2

Surfaced

MetalTotalcpHbapMe

) log(

) log(2

MatterOrganicd

MetalTotalcpHbapMe

© Sauvé 2003

Adsorption ModelAdsorption Model

simplified further to:

) log(2 MetalTotalcpHbapMe

Applied with succes to the soil solution speciationof Cd2+, Cu2+, Pb2+ and Zn2+.

© Sauvé 2003

Free CdFree Cd2+2+

Total Cd

pH

Field & spiked datasets are similar

No apparent effects of KOH-induced DOM

FieldSpiked

TYPE

© Sauvé 2003

Predictive Regressions for Predictive Regressions for Free CdFree Cd2+2+

Spiked dataset

Field & spiked soils

.).(,102,001.0,736.0

)(log)08.0(97.0)05.0(69.0)32.0(39.4)66.0(2

102

ESnpR

CdSoilTotalpHpCd

.).(,35,001.0,822.0

)(log)11.0(76.0)07.0(66.0)27.0(96.3)51.0(2

102

ESnpR

CdSoilTotalpHpCd

© Sauvé 2003

Free CuFree Cu2+2+

Tight relationship to soil solution pH and total metal content

N=94

© Sauvé 2003

Free PbFree Pb2+2+

For Pb…

N=84

© Sauvé 2003

Free ZnFree Zn2+2+

Preliminary speciation data for a free zinc regression

N=30 (Tambasco et al., Sauvé unpublished and and Knight et al. 1999)

© Sauvé 2003

Predictive Regressions for Predictive Regressions for Free MetalFree Metal

Pb2+

Cu2+

Zn2+

Should be possible to derive similar regressions for other divalent cationic metals or anionic elements.

.).(,94,001.0,921.0

)(log)08.0(84.1)06.0(47.1)39.0(20.3)58.0(2

102

ESnpR

CuSoilTotalpHpCu

.).(,84,001.0,643.0

)(log)10.0(84.0)05.0(84.0)28.0(78.6)47.0(2

102

ESnpR

CdSoilTotalpHpPb

.).(,30,001.0,760.0

)(log)22.0(71.1)10.0(95.0)59.0(70.4)50.0(2

102

ESnpR

ZnSoilTotalpHpZn

© Sauvé 2003

Free Ion Activity ModelFree Ion Activity Model

7.58.08.59.09.5

Free Metal (pCu2+ )

0

20

40

60

80

100

% S

urv

ival

B0 1 2 3 4 5

Total Dissolved Cu (µM)

0

20

40

60

80

100

% S

urv

ival

A

Ma H, Kim S, Cha D, Allen H (1999) Effect of kinetics of complexation by humic acid on toxicity of copper to Ceriodaphnia dubia. Environ Toxicol Chem 18: 828-837.

InhibitionInhibition

1 10 100 1000

Total Pb (mg·kg -1)

0

20

40

60

80

100%

Inhi

bitio

n R2=0.127

From: Sauvé et al. 1998. Derivation of soil quality criteria using predicted chemical speciation of Pb2+ and Cu2+. Environ. Toxicol.

Chem. 17:1481-1489.

InhibitionInhibition

1 10 100 1000

Total Pb (mg·kg -1)

0

20

40

60

80

100%

Inhi

bitio

n R2=0.127

From: Sauvé et al. 1998. Derivation of soil quality criteria using predicted chemical speciation of Pb2+ and Cu2+. Environ. Toxicol.

Chem. 17:1481-1489.

© Sauvé 2003

InhibitionInhibition

6789101112

Predicted pPb2+

0

20

40

60

80

100

% In

hibi

ti on

BR2=0.409

1 10 100 1000 10000Total Pb (mg/kg)

0

20

40

60

80

100

% In

hibi

ti on

AR2=0.127

From: Sauvé et al. 1998. Derivation of soil quality criteria using predicted chemical speciation of Pb2+ and Cu2+. Environ. Toxicol.

Chem. 17:1481-1489.

© Sauvé 2003

InhibitionInhibition

6789101112

Predicted pPb2+

0

20

40

60

80

100

% In

hibi

ti on

BR2=0.409

1 10 100 1000 10000Total Pb (mg/kg)

0

20

40

60

80

100

% In

hibi

ti on

AR2=0.127

From: Sauvé et al. 1998. Derivation of soil quality criteria using predicted chemical speciation of Pb2+ and Cu2+. Environ. Toxicol.

Chem. 17:1481-1489.

© Sauvé 2003

ConclusionsConclusions Reasonable predictions of the speciation of

metals in soils can be realized from simple regressions with total metal burden, soil solution pH and other soil characteristics

Within a large sampling program, as much care should be devoted to « other » physicochemical parameters as to soil metal analyses per se.

email: sebastien.sauve@umontreal.cahttp://mapageweb.umontreal.ca/sauves/