Presentation on theme: "Do we have a problem with freshwater Kd values? B. Howard and E. Tipping CEH, UK Analysis for discussion only – do not quote."— Presentation transcript:
Do we have a problem with freshwater Kd values? B. Howard and E. Tipping CEH, UK Analysis for discussion only – do not quote
ERICA ERICA uses Kd values to predict unknown water or sediment concentrations Water conc is used with CR to predict wholebody conc and internal dose Sediment conc is used for estimation of external dose Some ERICA values are based on sea water –-does this introduce larger error than for the other values used?
ERICA values from Ciffroy are AM from the reported GM
Comparison with TRS 364 (Onishi 81) Oxidising conditions
KDKD mol bound (g colloid) -1 mol L -1 in solution = But K D depends on: pH competing solutes competing ligands loading of the colloid ionic strength Modelling tries to explain variability in K D
H+H+ M z+ H+H+ - - - - N N Model VI Specific & non- specific proton & metal binding WHAM Key assumption – binding to organic matter dominates for metal ions
Humic substances Partial decomposition products of plants etc Principally composed of C, H and O, + N & S Possess weak acid groups - COOH, phenolic-OH Fulvic acidMWt ~ 1000dominant in waters Humic acidMWt ~ 10 000dominant in soils Heterogeneous, recalcitrant, yellow-to-brown The most abundant macromolecules on the planet!
Database for WHAM / Model VI ~ 20 data sets for protons ~ 100 data sets for metals Average proton binding for FA and HA Average binding for 23+ metals (Mg…Cu…Eu...Cm) Laboratory studies with isolated HA and FA Esp actinides
Model VI and cation binding : summary Proton and metal binding as function of [H+], [M z+ ] Proton-metal competition (pH dependence) Metal-metal competition (esp at high [M+]) Ionic (eg Na, Cl, )strength dependence of H and M binding due to interference with binding
Ion-binding models and their combinations “simple” solution chemistry Oxide model AlOx SiOx MnOx FeOx Clay cation exchanger Humic Ion- Binding Models V & VI WHAM SCAMP Na, Cl, OH, CO 3, SO 4
Wham 6 set up assumed ph8 for fw values also Freshwaters are for 3 different [DOC] - 1, 3 and 10 mg/L A range of pH's is generated by titrating an initially acid solution with Ca, to take us from pH ~ 4 to pH ~ 8.5 Seawater is assumed to be at pH 8, and with 2 mg/l DOC
WHAM IV Calculations assume that –DOC can be represented by average isolated fulvic acid, – OM in particulate matter (SPM) can be represented by average isolated humic acid Only organic matter in the SPM has any binding properties (oxides, clay etc ignored) Calculations take into account –competition between the element of interest and major ions (H+, Mg, Ca, Al, Fe etc), –complexation by inorganic ligands and natural organic matter (dissolved and particulate)
Kd estimates The Kd's are calculated for suspended particulate matter containing 10% organic matter results give some idea of –how Kd can vary with pH and [DOC], –comparisons between FW and SW
Health warning Elements which form hydrolysis reactions in solution at low pH may not be represented well as the model assumes organic complexation (eg Pu) The element concentrations are set to low levels and will be sensitive to the model's assumptions about small numbers of strong binding sites The model default database has differences in the binding strengths of fulvic and humic acid towards most metals, –these difference may not be real. (e.g. UO 2 and PuO 2 ) Some elements affected by redox, models assumes specifi oxidation state –Cr, Mn, Fe, Tc, Np
No Erica value (just WHAM) sw value similar to fw predictions at relevant pH Onishi Fe – 5000 Cr low Zn - 500
Erica - Ciffroy Am – ERICA high over most pH range Sw – lower Onishi 100x lower than ERICA Th – ERICA much higher Sw – lower, similar to fw model TRS – much lower Onishi (c.20000)
Changes with pH increase in Wham rises – Cr, Zn, Eu, Cm, Pb (Fe III, Am) rise and fall – Mn, Co, Sr, UO 2, Ni, Cd decrease – U IV (Th, Pu IV, PuO 2 ) Not possible to attribute differences systematically to only one causal factor – this would be misleading
Effect of DOC conc on Kd in FW in Wham IV High values are all metal ions with have the strongest binding to OM So more DOC = more metal in solution less DOC = less metal in soluton
SW vs FW – Erica vs model FW much higher than Wham SW –Am, Co, Mn, Sr, Th, PuIV, PuO 2 (Ciffroy) –Ni, Cd (sw values) –UIV (Onishi) Similar – Pb (ERICA is sw) FW much lower than Wham SW –UO 2, Eu, Cm (Onishi)
FW vs SW– model Wham FW higher than Wham SW –Cd, Mn, Sr, PuIV, PuO 2 UIV (Co, Eu, Ni,) Similar – Am, Cr, Cm, Fe III, Pb, Th, Zn FW lower than Wham SW –UO 2
Erica vs FW model Erica always higher than Wham –Co, Mn, Th, PuIV, PuO 2 (Ciffroy) Erica higher than Wham at low pH –Am, Sr, Ni, Cd, Pb (sw values) Erica lower than Wham –UO 2, Eu, Cm (Onishi) – except at pH 4 Similar at low pH, higher at high pH –UIV (Onishi)
Conclusions ERICA AM values often high Model rarely predicts SW > FW, often FW higher pH has large effect for many elements DOC important for Cr, Fe III, Pb, Am, Cm, Eu
Does it matter Too High Kd values –Will give low water conc – low whole body conc – therefore NOT conservative but more sensitive to error –Will give high sediment conc – higher external exposure - as >90% of most metals in sediment – less sensitive to error Can we “do” something in ERICA to assist user?