Tin

Adsorption / desorption

Link to relevant study record(s)

Reference

Endpoint:

adsorption / desorption: screening

Type of information:

experimental study

Adequacy of study:

key study

Study period:

not reported.

Reliability:

1 (reliable without restriction)

Rationale for reliability incl. deficiencies:

comparable to guideline study

Qualifier:

equivalent or similar to guideline

Guideline:

OECD Guideline 106 (Adsorption - Desorption Using a Batch Equilibrium Method)

Deviations:

not applicable

GLP compliance:

not specified

Remarks:

No further information on GLP compliance rationale is available.

Type of method:

batch equilibrium method

Media:

soil

Radiolabelling:

no

Analytical monitoring:

yes

Details on sampling:

- Concentrations:
5mg Sn/kg

- Sampling interval:
Kd values were assesed by inductively coupled plasma-mass spectrometry every 10 minutes through the analysis of duplicate soils.

- Sample storage before analysis:
not reported.

Details on matrix:

A total of 4813 soils were sent to CSIRO by Eurometaux for this study. From these, a subset of samples were repeated by re-scanning to determine IR variability. A total of 500 soils were selected for IR analysis to be used in the development of calibration and validation models. The FTIR spectral variability in the analysis of soils can be found in the attachements. The soils were air dried and sieved to < 2 mm prior to shipment to CSIRO. Soils were oven dried at 40 °C for 12 h and cooled in a desiccator prior to MIR analysis or experimental Log-Kd value determinations. The soil pH data in 0.01 M calcium chloride (CaCl2) were provided to CSIRO by Eurometaux (EU).

Details on test conditions:

Approximately 2.0 ± 0.05 g of <2 mm sieved soils was weighed into 50 mL centrifuge tubes (Cellstar, Greiner Bio-one) and mixed end over end for 48 h with 20 mL of 0.01 M CaCl2 (1:10 m/v).
After this initial equilibration period, samples were spiked with 25 μL of 400 mg Sn(IV)/L solution as tin(IV) chloride pentahydrate (98 %, Sigma-Aldrich) (5 mg Sn(IV)/kg) and mixed end over end for a further 72 h. In this study, a single metal spike concentration of 5 mg Sn/kg was selected to be low enough to be in the linear region of the sorption curve. The Sn(IV) stock solution (400 mg Sn/L) was prepared in concentrated hydrochloride acid (HCl) and diluted prior to spiking to soil suspensions with ultrapure deionised water (Millipore) (final acid concentration in spike = 1 % HCl).

Duration:

72 h

Initial conc. measured:

5 mg/kg soil d.w.

Type:

other: Log Kd

Value:

2.1 - 4.3

Remarks on result:

other: Value in L/kg. Result from the validation of the 200 sample test set.

Adsorption and desorption constants:

Log Kd values are reported in Table 1 (below).

Recovery of test material:

not reported.

Concentration of test substance at end of adsorption equilibration period:

not reported.

Concentration of test substance at end of desorption equilibration period:

not applicable

Table 1. Results

	Log-Kd PLS DRIFT model	Log-Kd univariate model	Log-Kd PLS (DRIFT+pH) model	Log-Kd (DRIFT+pH) validation of the 200 sample Test set.
Min	1.8	1.8	1.8	2.1
Max	4.3	4.3	4.3	4.3
R2	0.32	0.12	0.32	0.31
RMSECV	0.47	0.53	0.47	0.43
SD	0.6	0.6	0.6	0.4
RPD	1.2	1.1	1.2	1

Results of the modelling of Log-Kd for Sn(IV) are shown for PLS modelling using only the DRIFT spectra, univariate regression from pH, and composite PLS model (DRIFT+pH). Statistics are described by coefficient of determination (R2), root mean square of error for cross validation (RMSECV), standard deviation (SD) and ratio of prediction to deviation (RPD). Statistics are also shown for the 200 sample validation soil set and the full set of unknown samples from the composite PLS model (Min, Max, SD and RMSECV are in L/kg units).

Validity criteria fulfilled:

not specified

Conclusions:

The experimentally derived Log-Kd of Sn(IV) in 200 soils was between 2.1 and 4.3 L/kg

Executive summary:

The aim of this project was to use mid-infrared spectroscopic analysis of soils to develop a calibration model to predict the solid-solution partitioning (Kd values) values for tin(IV) (Sn(IV)) in soils of the GEMAS soil sampling program.

Experimental Log-Kd values and a multivariate partial least squares (PLS) model were developed using 500 selected soils from the 4813 soils of the GEMAS sampling program. A composite PLS model was developed using a combination of both spectral and soil pH data and was found to predict with “poor” accuracy experimental Log-Kd values for Sn(IV) in calibration soils (n = 325, R2 = 0.32).

Validation of the model was carried out with the 500 samples being split into two sets; the first 300 for calibration and the remaining 200 for validation. Samples exceeding the maximum experimental determined Kd value for Sn(IV) of 25 500 L/kg were not included in the calibration model or as validation samples. The composite PLS model (spectra plus soil pH) developed was found to predict with “poor” accuracy the Log-Kd values for Sn(IV) in the validation soils (n = 176, R2 = 0.31). Compared to other metals, the variance in the Log-Kd values accounted for the model was lower for Sn(IV), this is likely due the instability of SnCl4 in water and the precipitation of Sn(IV) as an insoluble oxide (e.g. SnO2) in soil solutions.

The composite PLS model (spectra plus soil pH) was applied to the prediction of Log- Kd values for Sn(IV) in the 4313 soils of the GEMAS soil sampling program.

Description of key information

Janik et al (2010) determined the Log Kd of Sn (IV) in nearly 500 soils collected from across Europe. Tin was applied as tin(IV) chloride pentahydrate and allowed to equilibrate for 72 hours. After this time the solid solution partitioning coefficient for each soil was determined. The measured Log Kd values ranged from 0 -4.4, with a median value of 3.9. Janik et al report measured Kd values from a large number of European soils, and therefore the median log Kd value of 3.9 for soil is selected for use in the assessment.

Only lower quality studies are available for suspended matter and sediment Kd. Therefore, mean values from the available studies are selected for use in the assessment. This gives log Kd values of 5.59 for suspended matter and 4.61 for sediment.

Additional information

Key value for chemical safety assessment

Other adsorption coefficients: Log Kd = 2.1 - 4.3 L/kg

 

There is virtually no available knowledge with regard to the effect of varying environmental physicochemistry on Sn-distribution coefficients. In fact, due to low solubility of tin in water – often giving dissolved Sn-levels below detection limits – few distribution coefficients for inorganic tin are available in (peer-reviewed) literature. More information is available for organotin compounds, but as these compounds behave differently than inorganic tin in the environment and have different properties, it is not relevant to use partition coefficients for organotin compounds.

In the most recent study, Janik et al (2010) report measured Log Kd values for the 484 soils from across the EU27 (and Norway). No data were available for Malta and Romania and only one data point was available for Luxembourg. Tin was applied to the soils as tin (IV) chloride pentahydrate and allowed to equilibrate. The measured Log Kd values ranged from 0 to 4.4, with a median of 3.9. The 10th percentile value of these soils was 2.77 and the 90th percentile 4.36.

 

Allison & Allison (2005) performed for various metals a literature survey to determine the range and statistical distribution of partition coefficient values that have been observed in field scenarios, including the collection of published partition coefficients or the estimation of partition coefficients from reported metal concentration data when feasible. The few tin partition coefficients that were collected for the different compartments are presented in the Table below. No differentiation was made between divalent and tetravalent tin. Allison & Allison (2005) did not report a total number of data for the soil/water Kd as this values was taken from a previous compilation.

It is assumed that the Kd value for sediment-water is based on a single measurement.

Median, mean (+ standard deviation), range, and number of samples (N) for Sn-partition coefficients (Kd in L/kg) from literature search (Allison & Allison, 2005).

	Soil/Water	Suspended Matter/Water	Sediment/Water
Median (log-value)	2.8	5.6	4.7
Range (Min-Max)	2.1 – 4.0	4.9 – 6.3	--
No. of data	--	3	--

 

An overview of available tin partition coefficients in literature has also been given by Crommentuijn et al (1997). Only one value was found for each compartment, and these values were proposed as a typical value for each compartment. A summary of the data is given in the Table below. It should be noted, however, that Crommentuijn et al. (1997) did not mention whether the reported values were representative for divalent or tetravalent tin. The partition coefficient (log Kd) reported by Crommentuijn et al. (2000) for suspended matter is 5.57, which is in line with the median value of 5.6 for Sn derived by Allison & Allison (2005). For sediment the Crommentuijn et al (2000) value (log Kd of 6.09) is approximately 1.3 log units higher than the value given by Allison & Allison (2005) (log Kd of 4.7). It should be noted, however, that the Crommentuijn et al (2000) value is taken from a secondary source, who estimated this value on the basis of partition coefficients between particulate matter and water (i.e., not based on measured values). Therefore, this value is considered as a low quality data point and should not be used for risk assessment purposes if other data are available. For the soil compartment the recommended log Kd from Crommentuijn et al. (2000) is around 0.5 log units higher than the value reported by Allison & Allison (2005).

 

Sn-distribution coefficients (Kd in L/kg) reported in Crommentuijn et al (2000)

Log Kd	Reference
Suspended matter/water
5.57	Bockting et al., 1992^(1)
Sediment/water
6.09	Stortelder et al., 1989
Soil/water
3.28	Bockting et al., 1992^(2)

⁽¹⁾: value derived by Bockting et al (1992) using measurements of metals in fresh surface water samples of 4 locations from 1983-1986 and data from Popp and Laquer (1980) and Li et al. (1984), in North American rivers

⁽²⁾: values based on literature study using batch experiments

  

The Allison & Allison (2005) value of 4.7 for the sediment compartment is supported by data presented by Sun et al. (1996). These authors published their findings on adsorption behaviour of Sn4+ (added as SnCl4) on estuarine sediment, and calculated a log Kd of 4.49 (based on a Freundlich isotherm) using measurements of Sn in water and spiked sediment (pH 8.04; Clay, 11.33 %; OM-content 1.43 %).

Nakamaru & Uchida (2008) presented distribution coefficients of Sn in Japanese agricultural soils, based on 142 measurements in four types of soil (Andosols, Fluvisols, Cambisols and Regosols). Log Kds ranged between 2.11 and 6.20, with a geometric mean of 4.09. This is around one order of magnitude above the typical values that are reported by Allison & Allison (2005) and attributed to Bockting et al. (1992).

The table below summarizes the different log Kd values that are considered to be reliable and relevant. Based on these data, a typical log Kd value for Sn (valence state not specified) is proposed as default value for local and regional modelling purposes.