Registration Dossier

Environmental fate & pathways

Adsorption / desorption

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

A weight of evidence approach has been taken to determine the log Kp values for soil, suspended matter and sediment. The log Kp values that will be used in the assessment are: log Kp soil 2.07, log Kp suspended matter 3.65, and log Kp sediment 3.35.

In order to run EUSES, partition coefficients are required for soil, suspended matter, sediment, and raw, settled, activated and effluent sewage sludge. All of these are available except for the partition coefficients for sewage sludge. For organic substances EUSES can calculate a default partition coefficient based on the log Kow, but this is not applicable to inorganic substances. Instead, the individual Kp values available for soil, suspended matter and sediment have been used to back-calculate a theoretical log Kow based on the equations given in the TGD (ECB, 2003). The median theoretical log Kow for soil is 5.31, for suspended matter it is 6.98 and for sediment is also 6.98. The median of these three values (log Kow 6.98) has been entered into EUSES to calculate the partition coefficients for sewage sludge.

The calculated log Kp values are 4.13 for raw sewage sludge and settled activated sludge and 4.22 for activated sewage sludge and effluent sewage sludge. The compartment-specific log Kp values listed above have been used for soil, sediment and suspended matter.

Key value for chemical safety assessment

Other adsorption coefficients

Type:
log Kp (solids-water in soil)
Value in L/kg:
2.07

Other adsorption coefficients

Type:
log Kp (solids-water in sediment)
Value in L/kg:
3.35

Other adsorption coefficients

Type:
log Kp (solids-water in suspended matter)
Value in L/kg:
3.65

Other adsorption coefficients

Type:
log Kp (solids-water in raw sewage sludge)
Value in L/kg:
4.13

Other adsorption coefficients

Type:
log Kp (solids-water in activated sewage sludge)
Value in L/kg:
4.22

Additional information

Transport of metals between aqueous phase and soil/sediment/suspended matter should be described on the basis of measured soil/water, sediment/water and suspended matter/water equilibrium partition coefficients (Kp), instead of using common mathematical relationships based on, for example, octanol-water partition coefficients, as is usually done for organic chemicals (TGD, 1996).

Kpsoil

Four reliable studies were identified in the EU RAR for diantimony trioxide which report soil Kp.

Vangheluwe et al. (2001) report the results of a transformation experiment in soil using Sb2O3and SbCl3. The transformation experiment was performed for a duration of 24 weeks, with soils incubated at 20ºC, with a final humidity of 22g/100g dry substance (both soils). The incubation was initiated with 2 drying/rewetting cycles. Calculation of the Kpsoil, using a control soil in Vangheluwe et al. (2001), results in a partition coefficient in sandy soils of 81 l/kg (log Kpsoil= 1.91) and 150 l/kg (log Kpsoil= 2.18) in loamy soils. In the control the concentration was 0.57 mg Sb/kg in sandy soil and 0.45 mg Sb/kg in loamy soil. The corresponding pore water concentrations were 7mg Sb/l and 3mg Sb/l, respectively. The lowest applied Sb dose in this experiment (10 mg Sb/kg; Sb2O3), which is higher than normally expected in the environment, resulted in the Kpsoilvalues of 26 l/kg (log Kpsoil= 1.41) and 21 l/kg (log Kpsoil= 1.32) for sandy and loamy soil, respectively. The distribution coefficient Kd between soil and pore water was calculated to be 30 l/kg (log Kpsoil = 1.48) for 50 mg Sb/kg applied as SbCl3in sandy soil and 75 l/kg (log Kpsoil= 1.88) in a loamy soil.

Pauwels (1985) studied the distribution of Sb in soil and the uptake by plants for soils from Belgium and Algeria. The soils were treated with 32 or 1,000 mg Sb/kg (Sb2O3). The Kp in sandy soil (pH = 5.7) was 60 (log Kpsoil= 1.78) and 128 (log Kpsoil= 2.34) for the low and high concentrations, respectively, while in a soil with heavy clay (pH = 7.8) the Kp was 94 (log Kpsoil= 1.97) and 161 (log Kpsoil= 2.21) for the low and high concentrations, respectively.

King (1988) performed batch experiments with 13 soil series samples (pH: 4.2-6.5, clay 2-63%) taken from the south east of England). The equilibrium concentration of antimony was <1 to 100 mg/l, which ranges from background levels to much above background levels. As regards the importance of pH, no pH dependence of Kd values of Sb was found in the study of King (1988), but Kd values were correlated with oxide content and with clay content. Since the pH effect on Sb sorption is quite small, this pH effect is probably hidden by the effect of oxide content, when comparing different soils.

Based on the soil classification system developed by USDA (1967) the soil types sand (n = 2), sandy loam (n = 2), sandy loam-loam (n = 1), loam (n = 1), sandy clay loam-clay loam (n = 1), clay (n = 1) and silt loam (n = 1) were identified. In addition one humus-rich soil sample with no data on sand, silt or clay is also available. Two additional humus-rich soil samples were excluded since it was not possible to calculate partition coefficients from the data available. In order not to have a disproportionately large influence of the different soil types in this study (the majority consisting of only one or two values), when compared to data from other studies, it was decided to combine the different loam soils, resulting in one sand soil (n = 2; Lakeland-A, Norfolk-Ap), one loam soil group (n = 7; Cecil-Ap, Davidson-Ap, Mecklenburg-Ap, Wahee-A, Iredell-Ap, White Store-A, Portsmouth-Ap), one clay soil group (n = 1; Hayesville-A), and one humus-rich soil (n=1; Wasada-Ap). The corresponding (median; for the sand and loam soils) log values are 1.44, 2.35, 1.98, and 2.08, respectively.

Nakamura et al.(2006) studied the mobility of antimony in Japanese agricultural soils by radiotracer experiments using124Sb tracer. The Kd values for Sb (0.1 mg Sb/kg added as SbCl3, 7 days equilibration) measured in 110 Japanese soils ranged from 1 to 2065 l/kg, with a geometric mean of 62 l/kg (Log Kpsoil= 1.79) excluding one extremely high value of 2065 l/kg. Experimental measurements of Kd showed a decrease with both increasing pH and increasing phosphate concentration. The latter suggested that one aspect of the antimony sorption in Japanese soils was influenced by specific adsorption of anions such as phosphate. However, other aspects could not be explained by this specific adsorption mechanism, because only 20-40 % of soil-sorbed antimony could be extracted by phosphate solution.

Four sources for log Kpsoilvalues for antimony were identified in the EU RAR:

i) The study by Vangheluwe et al. (2001), which contain the log Kpsoilfor sandy and loamy soils. The corresponding partition coefficients for the two control soils were 1.91 and 2.18; for the 10 mg Sb/kg exposure group it was 1.41 and 1.32. In another experiment using 50 mg Sb/kg the partition coefficients 1.48 and 1.88 were obtained for sandy and loamy soil, respectively.

ii) The study by Pauwels (1985), which resulted in the log Kpsoilvalues of 1.78, 2.34, 1.97, and 2.21, with the first two values coming from a sandy soil and the last two values from a soil with heavy clay.

iii) The study by King (1988), which resulted in the values 1.44, 2.35, 1.98, and 2.08 for the sand, loam, clay and humus-rich soils, respectively.

iv) The study by Nakamura et al. (2006), which resulted in the log Kpsoil1.79.

However, in order to try to give equal weight to results from different studies performed in different types of soils at different concentrations only one value for each soil type from each study will be used. Since the “loading dependence” for antimony in soil does not seem to be important (see above) a median value of the partition coefficients will be taken for the different concentrations used for each soil type. This means that the data points (log Kpsoilvalues) resulting from the study by Vangheluwe et al. (2001), will be used in this CSR, and will be 1.48 for sandy soil (median value of 1.91, 1.32, and 1.48), and 1.88 for loamy soil (median value of 2.18, 1.41, and 1.88). The resulting data points from the study by Pauwels (1985) would be 2.06 for sandy soils (median value of 1.78 and 2.34) and 2.09 for the soil with heavy clay (median value of 1.97 and 2.21). The values from King (1988) are 1.44, 2.35, 1.98, and 2.08, and the value from Nakamaru et al.(2006) is 1.79. Thus, the resulting values (log Kpsoil) are 1.48, 1.88, 2.06, 2.09, 1.44, 2.35, 1.98, 2.08 and 1.79. Based on these nine data points, it was decided to use the median value in the EU RAR (log Kpsoil= 1.98).

A further seven studies have since been identified that report additional log Kp soil values for antimony. Four of these studies (Ettler et al. 2007, Fuentes et al. 2003, Lintschinger et al. 1998 and Clemente, Dickinson and Lepp 2008) investigate how easily antimony in soils is desorbed. All four studies used soils collected from sites which were potentially exposed to antimony either due to mining and smelting operations or industrial use of antimony. Following the approach taken in the EU RAR, log Kp values are calculated for each soil type reported in each study. This results in log Kp values of 2.85 (forest soil) and 3 (tilled soil) (Ettler et al., 2007), 3.2 (Fuentes et al., 2003), 3.95 (Lintschinger et al., 1998) and 2.1 (Clemente et al., 2008).

Oorts et al. (2008) spiked soil from an agricultural field with antimony trichloride at concentrations from 0-1000 mg Sb/kg. The concentration of antimony in the soil solution was monitored from 2-35 days after spiking. A linear relationship between soil solution Sb concentration and dose was observed up to 500 mg Sb/kg, above which the soil solution Sb concentration remained constant. The log Kp was calculated from the linear portion of the graph as 1.6.

Vangheluwe et al. (2003) spiked a sandy soil and a loam soil with antimony trichloride at concentrations from 0-50 mg Sb/kg. The concentration of antimony in the soil solution was monitored after 24 weeks at 22% water content, or after 30 weeks at field capacity. Concentrations in the soil solution ranged from 0.007-3.82 (sandy soil) and 0.003-0.86 (loam soil). The log Kp was calculated as 1.55 for the sandy soil and 1.75 for the loam.

Tighe, Lockwood and Wilson (2005) spiked two floodplain (loam) soils with potassium antimonate at a range of concentrations (0-11.3 mg Sb/l). After 1 day’s equilibration the concentration of Sb in the aqueous phase was analysed and Freundlich isotherms constructed. The log Kp values ranged from 1.9-3.1 for both soils. As both soils are classed as loams the median log Kp of 2.5 is used in this assessment.

Based on the nine data points from the EU RAR and the additional nine data points identified here, we have decided to use the median value in the assessment (log Kpsoil= 2.07). This value is very similar to the value used in the EU RAR (log Kpsoil= 1.98).

New data has recently become available. Janik et al (2010) determined the Log Kp of antimony in nearly 500 soils collected from across Europe. Antimony was applied as antimony (V) chloride (5 mg Sb/kg) and allowed to equilibrate for 72 hours. After this time the solid solution partitioning coefficient for each soil was determined. The measured Log Kp values ranged from 0 -3.74, with a median value of 1.8. The range of Log Kp values reported in this study spans the range reported by other researchers and reported here. As the results from this study cover a wide number of soils, of different types the median value is not used to influence the Log Kp used in the risk assessment, but is used to confirm that the Log Kp of 2.07 used for soil is an appropriate value.

Kpsuspended matter

Partition coefficients for the distribution of metals between water and suspended matter are used to calculate the dissolved concentrations from total concentrations in surface water. Seven reliable studies were identified in the EU RAR for diantimony trioxide which report suspended matter Kp.

Habib and Minski (1982) reported mean concentrations of antimony in soluble and suspended particulate fractions of 0.27 μg Sb/l and 8.4 mg Sb/kg, respectively, for three stations on the River Thames. The sampling covered a 10-month period, with a sampling frequency of every second week. Each sampling was performed and completed within five hours. The resulting log Kp value is 4.49 l/kg.

Martin and Whitfield (1983) report a log Kp suspended matter of 3.4 representing a world river average.

Van der Sloot et al.(1989) determined metal concentrations of particulate matter in the Solo River (Indonesia). Sampling of the particulate matter involved continuous flow centrifugation of water. Particulate matter was obtained in two size/density fractions. The metal content of the particulate matter was determined using neutron activation analysis (NAA). The metal concentrations of the water samples (0.45 µm filtered) were determined with NAA after a preconcentration procedure. For the calculation of the Kpvalues the average metal content of the two size fractions was used. The average log Kp value is 3.3 l/kg.

Veselý et al. (2001) derived partition coefficients for 41 elements, among them antimony, from water samples collected from 54 Czech rivers at 119 localities over the whole state territory in the summers of 1997 and 1998 under stable hydrological conditions. The analysed river waters had a mean pH of 7.74 (range 6.9 to 8.8), ionic strength 7.8 mmol/l, specific conductance 538mS/cm at 25 ºC, alkalinity of 1.9 mmol, and moderate mean contents of suspended particulate matter (SPM) of 9.9 mg/l (range 1.0-124 mg/l). The 10th, 50thand 90thpercentile of the log Kpsuspended matterwere 3.54, 4.14 and 4.64, respectively. The values of Kp were calculated by dividing the total concentration of the element in SPM by its concentration in filtered water. There was no statistically significant dependence between log Kp values for antimony and loading (element loading is the sum of element concentration in the “dissolved” fraction and in the SPM).

Van der Sloot et al.(1985) measured several oxy-anionic metals, among them antimony, in water and particulate matter samples from the Dutch Wadden Sea, the North Sea, the estuaries of the Rhine and Scheldt and Lake Yssel. Kp values derived from data for Lake Yssel (the only freshwater samples) are presented above. Sampling of particulate matter involved filtration through 0.45 µm filters. The metal content was determined after leaching for 18 h with 0.1 N HCl.

Li et al.(1984) performed batch experiments with particulate matter from the Hudson River (USA). Unfiltered water samples (freshwater, CL= 18 mg/l) were spiked with 13 radiotracers, including antimony. The spikes did not greatly change the natural concentration of various trace elements, as most radiotracers were carrier free and the amounts of spikes added were small. The samples were shaken for 20 days at 20°C. The aqueous radioactivity was determined in filtered (< 0.4 µm) subsamples taken at predetermined intervals using a germanium-lithium detector. Kp values for (unfiltered) water and a 4:1 mixture with filtered seawater are presented (equilibration period 20 days).

Hoede et al.(1987) determined metal concentrations of particulate matter in Indonesian river water. Sampling of the particulate matter involved continuous flow centrifugation of water. Particulate matter was obtained in two size/density fractions. The metal content of the particulate matter was determined using neutron activation analysis (NAA). The metal concentrations of the water samples (0.45 µm filtered) were determined with NAA after a preconcentration procedure. For the calculation of the Kp values the average metal content of the two size fractions was used.

Seven sources for log Kpsuspended mattervalues for antimony were identified in the EU RAR:

i) The study by Habib and Minski (1982) in which the log Kpsuspended matterfor the river Thames in the UK was 4.5.

ii) The study by Martin and Whitfield (1983) including a log Kpsuspended matterof 3.4, representing a world river average.

iii) The study by Van der Sloot et al. (1989) including a log Kpsuspended matterof 3. 3 for the river Solo (Indonesia).

iv) The study byVeselý et al.(2001) in which the 50thpercentile of log Kpsuspended matterfor Czech rivers was 4.14.

v) The study by van der Sloot et al. (1985) including two log Kpsuspended mattervalues of 3.52 and 3.64 for the Dutch lake Ysselmeer.

vi) The study by Li et al. (1984) in which the log Kpsuspended matterfor the Hudson river (North America) was 4.07.

vii) The study by Hoede et al.(1987) in which the log Kpsuspended matterfor the Indonesian rivers Solo, Wono Kromo, and Porong were 3.17, 3.00, and 3.72, respectively.

However, in order to try to give equal weight to results from different studies performed in different waters, it was decided to use one value for each water system. This means that the data points (log Kpsuspended mattervalues) from the River Thames will be 4.5, for the world wide average 3.4, for the River Solo 3.2 (median value of 3. 3 and 3.17), for the Czech rivers 4.14, for Lake Ysselmeer 3.58 (median value of 3.52 and 3.64), for the Hudson River 4.07, for the River Kromo 3.00, and for the River Porong 3.72. Thus, the resulting values (log Kpsuspended matter) are 4.5, 3.4, 3.2, 4.14, 3.58, 4.07, 3.00, and 3.72.

Based on these eight data points, it was decided to use the median value (log Kpsuspended matter= 3.65) in the EU RAR. No additional studies have been identified since the EU RAR so this value is used here also.

Kpsediment

Partition coefficients for the partitioning of metals between water and suspended matter are used to calculate the concentration in sediment from the dissolved concentration in water on a local scale. When calculating the concentration in sediments on a regional scale the partitioning of metals between water and sediment should be used. Three reliable studies were identified in the EU RAR for diantimony trioxide which report a sediment Kp.

Brannon and Patrick (1985) determined Sb concentrations in the interstitial water phase of 10 North American sediments. Sediments were kept under anoxic conditions. Ten freshwater sediment samples were mixed with cellulose (1% weight: weight basis) to enhance reduction of the sediment. Sb(III) was added (75 mg/kg, sediment dry weight) in the form of an antimony potassium tartrate solution. The containers were sealed and incubated for 45 days under a nitrogen atmosphere at 20˚C. After this period interstitial water from the sediments was collected by centrifugation of the sample under nitrogen atmosphere. All aqueous concentrations for unamended sediments were below the detection limit. Log Kpsediment(l/kg) calculated from the total Sb content of 10 Sb amended freshwater sediments (native Sb-content and added Sb) and the corresponding interstitial concentrations ranged from 2.82 to 3.64. The mean log Kpsedimentwas 3.32.

Kawamoto and Morisawa (2003) studied the distribution and speciation of antimony in river water, sediment and biota in tributaries of the Yodo River, Japan. Samples were taken both upstream and downstream of municipal sewage treatment plants (STP), accepting wastewater from dye works, which are the main industries in the area. The antimony concentrations in water did not correlate well with the antimony concentration in sediments, as the concentration of antimony in water upstream of the STP in the Katsura River was lower than downstream (0.5 and 3.4 mg Sb/l, respectively) but the opposite was true for the concentration in sediments (2.4 and 1.2 mg Sb/kg dry weight). Concentrations measured in water and sediments downstream from another STP in the Umi River were 3.4 mg Sb/l and 1.0 mg Sb/kg dry weight, which are very similar to those measured downstream of the STP located on the Katsura River. Speciation of antimony revealed that the proportion of dissolved Sb(III) and Sb(V) changed depending on whether or not the sampling site was upstream or downstream of a STP, and how far downstream of the STP the sampling was performed. The proportion of Sb(III) and Sb(V) upstream and far downstream of STP was 20/80 and 30/70, respectively, while at point close downstream of an STP it was 1/99. This difference, which was suggested to be the result of the processes in the STP, thus resulted in a non-typical water for speciation of antimony.

Mori et al. (1999) measured the concentration of antimony in water and sediments in Corsican Rivers upstream and downstream of an abandoned realgar mine. The Bravona River spring is located in central Corsica and after 38 km reaches the Tyrrhenian Sea, on the eastern coast of Corsica. One of its tributaries, the Presa River, crosses an abandoned realgar mine. Ten sampling sites were chosen: six were located on the axial course of the Bravona River, three on the Presa River (1 upstream of the mine and two downstream) and one on a secondary tributary (the Alzillelo). The Presa River connects to the Bravona river between stations B2 and B3.

Three sources for log Kpsedimentvalues for antimony were identified in the EU RAR:

i) The study by Brannon and Patrick (1985) resulting in an averaged Kpsedimentof 3.32 from 10 North American sediments.

ii) The study by Kawamoto and Morisawa (2003) in which the Kpsedimentfor Japanese rivers (upstream and downstream of STP connected to Sb-emitting industries) was given: upstream STP – 3.68, 3.95; downstream STP – 2.51, 2.55, 2.47. Since the discharge of antimony, via the STP, changed the proportion of Sb(III) and Sb(V) upstream and downstream from the STP, only the values obtained upstream of the STPs will be used here.

iii) The study by Mori et al. (1999) in which a number of Kpsedimentvalues for Corsican rivers (from reference sites to strongly polluted) were given: reference sites – 4.20, 3.3, 3.51, 3.12; polluted – 2.99, 3.03, 3.30, 2.69; heavily polluted – 3.65, 2.89.

However, in order to try to give equal weight to results from different studies performed in different waters, only one value for each water system will be used. This means that the data points (log Kpsedimentvalues) are: North American sediments = 3.32, River Katsura = 3.68, River Kamo = 3.95, River Bravona = 3.18 (median value of 4.2, 3.3, 2.99, 3.05, 3.3, 2.69), River Presa = 3.51 (median value of 3.51, 3.64, 2.89), and River Alzillelo = 3.12. Thus, the resulting values (log Kpsediment) are 3.32, 3.68, 3.95, 3.18, 3.51, and 3.12. Based on these six data points, it was decided to use the median value (log Kpsediment= 3.4).

A futher three studies have since been identified that report additional log Kp sediment values for antimony (Ettler et al., 2007, Sola and Prat, 2006; Fernandez-Turiel et al., 1995). All three of these studies report paired monitoring data from water and sediment which allow us to calculate log Kp sediment values. The studies by Ettler et al. and Sola and Prat are located in areas potentially exposed to antimony due to mining and smelting operations. The area sampled by Fernandez-Turiel et al. is relatively uncontaminated. Following the approach used in the EU RAR only a single log Kp is used for each water system. The measurements reported in these studies result in log Kp values of 3.35 (Ettler et al., 2007), 3.5 (Sola and Prat, 2006) and 2.7 (Fernandez-Turiel et al., 1995).

Based on the six data points from the EU RAR and the additional three data points identified here, we have decided to use the median value in the assessment (log Kpsediment= 3.35). This value is very similar to the value used in the EU RAR (log Kpsediment= 3.4).