[No public or meaningful name is available]

Adsorption / desorption

Link to relevant study record(s)

Reference

Reference 1

Endpoint:

adsorption / desorption, other

Remarks:

multitracer experiment

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Qualifier:

no guideline followed

Principles of method if other than guideline:

Multitracer experiment. The adsorption of elements on model samples of marine particulates, including deep-sea and near-shore sediments, was studied.

GLP compliance:

no

Type of method:

other: multitracer experiment

Media:

sediment

Radiolabelling:

yes

Remarks:

radioactive multitracers are used

Test temperature:

25°C

Analytical monitoring:

yes

Details on sampling:

The suspension was shaken in an 8-shape mode with a shaker at 25° C. After centrifugation, a portion of the supernatant solution was pipetted.

Details on matrix:

Matrix 1
COLLECTION AND STORAGE
- Geographic location: deep-sea sediment Penrhyn Basin (12°26.44'S, 157°57.20'W, depth 5351 m)
- Storage conditions: in artificial seawater (storage 2 mg/cm³ artificial seawater), pH of suspension adjusted to 7.5

Matrix 2
COLLECTION AND STORAGE
- Geographic location: near-shore sediment Suruga Bay (34°44.52.'N, 169°27.05'E, depth ca. 600 m)
- Storage conditions: in artificial seawater (storage 2 mg/cm³ artificial seawater), pH of suspension adjusted to 7.5

Details on test conditions:

TEST CONDITIONS
10 cm3 of artificial sea water and 0.01 cm3 of a multitracer solution were put into a polyethylene bottle. After the pH was adjusted to 7.5 with a 1 M Na2CO3 solution, the adsorbent suspension was added. Resulting concentrations of adsorbent were 0.1, 0.05 and 0.01 mg/cm3. pH of suspension was readjusted to 7.5 when necessary.

Artifical seawater containing 23.94 g NaCl and 0.196 g NaHCO3 / 1000 cm3 water was used.

TEST SYSTEM
- Type, size and further details on reaction vessel: polyethylene bottle.
- Amount of soil/sediment/sludge and water per treatment: 10 cm3 artificial seawater + 0.01 cm3 of a multitracer solution + 0.1, 0.05 or 0.01 mg adsorbent/cm3.
- Suspension was shaken in a 8-shape mode with a shaker at 25°C.
- To reach equilibrium, 1-2 days of shaking was needed (not further specified for Fe).

Computational methods:

Kads = (Aads/m)/(Asoln/V) = ((Ai - Af)/Af)(V/m) (in cm3/g = L/kg))
Aads = radioactivity in the adsorbent after adsorption equilibrium
Asoln = radioactivity in the solution after adsorption equilibrium
V = volume of the solution (in cm3)
m = amount of adsorbent (in g)
Ai = radioactivity in the solution before adsorption equilibrium
Af = radioactivity in the solution after adsorption equilibrium

Phase system:

sediment-water

Type:

log Kp

Value:

5.76 L/kg

Temp.:

25 °C

pH:

7.5

Matrix:

deep-sea sediment

The log Kp value for near-shore sediment was determined for several elements in the publication, but for Fe.

Conclusions:

In this multitracer experiment, the log Kp for Fe for deep-sea sediment in artificial seawater was determined to be 5.76 L/kg.

Reference 2

Endpoint:

adsorption / desorption, other

Remarks:

Article on the Unit World Model citing Kd values

Type of information:

other: Article on the Unit World Model citing Kd values

Adequacy of study:

weight of evidence

Reliability:

4 (not assignable)

Rationale for reliability incl. deficiencies:

secondary literature

Qualifier:

no guideline followed

Principles of method if other than guideline:

The purpose of this paper is to explore and discuss the challenges in applying the Unit World Model (UWM) to metals with a view to developing an UWM critical loading model, EQCMetal, that can be applied to both metal ions and organic substances. In this context, Kp values for suspended mater/water, sediment/water and soil/water are cited for Fe.

GLP compliance:

not specified

Type of method:

other: Article on the Unit World Model citing Kd values

Media:

other: suspended matter, sediment and soil

Phase system:

sediment-water

Type:

log Kp

Value:

4.9

Remarks on result:

other: average value from 11 data

Phase system:

sediment-water

Type:

log Kp

Value:

4.31 - 5.767

Remarks on result:

other: range for 11 data

Phase system:

suspended matter-water

Type:

log Kp

Value:

6.63

Remarks on result:

other: average value from 5 data

Phase system:

suspended matter-water

Type:

log Kp

Value:

6.106 - 7.111

Remarks on result:

other: range for 5 data

Phase system:

soil-water

Type:

log Kp

Value:

2.74

Remarks on result:

other: average value from 4 data

Phase system:

soil-water

Type:

log Kp

Value:

2.394 - 3.079

Remarks on result:

other: range for 4 data

The literature data from which the above log Kp values were averaged are cited in the Supporting Information to the article:

Log Kp for sediment/water:

van Hattum et al. (1989) - 2 locations: 4.310, 4.972

Davies et al (1996) - 1 location: 5.181

Diamond et al. (1990) - 1 location: 5.767

Besser et al. (2001) - 7 locations: 4.401, 4.507, 4.706, 4.805, 4.994, 5.022, 5.213

Log Kp for suspended matter/water:

Diamond et al. (1990) - 1 location: 6.106

Cuthbert et al. (1993) - 1 location: 7.050

Gobeil et al. (2005) - 1 location: 6.496

Rondeau et al. (2005) - 2 locations: 6.374, 7.111

Log Kp for soil/water:

Thibault et al. (1990) - 4 locations: 2.394, 2.519, 2.954, 3.079

------------------------

Besser JM, Brumbaugh WG, May TW, Church SE, Kimball BA. 2001. Bioavailability of metals in  stream food webs and hazards to brook trout (Salvelinus fontinalis) in the upper Animas River Watershed, Colorado. Arch Environ Contam Toxicol 40:48–59.

Cuthbert ID, Kalff J. 1993. Empirical models for estimating the concentrations and exports of metals in rural rivers and streams. Water Air Soil Pollut 71:205–230.

Davis A, Sellstone C, Clough S, Barrick R, Yare B. 1996. Bioaccumulation of arsenic, chromium, and lead by fish: Constraints imposed by sediment geochemistry. Applied Geochemistry 11: 409–423.

Diamond ML, Mackay D, Cornett RJ, Chant LA. 1990. A model of the exchange of inorganic chemicals between water and sediments.Environ Sci Technol 24:713–722.

Gobeil C, Rondeau B, Beaudin L. 2005.Contribution of municipal effluents to metal fluxes in the St. Lawrence River.Environ Sci Technol 39:456–464.

Rondeau B, Cossa D, Gagnon P, Pham TT, Surette C. 2005.Hydrological and biogeochemical dynamics of the minor and trace elements in the St. Lawrence River. Applied Geochemistry 20: 1391–1408.

Thibault DH, Sheppard MI, Smith PA. 1990. A critical compilation and review of default soil solid/liquid partition coefficients, Kd, for use in environmental assessments. Report 10125. Atomic Energy of Canada, Pinawa, MB.

Timmermans KR, van Hattum B, Kraak MHS, Davids C. 1989. Trace metals in a littoral food web: Concentrations in organisms, sediment, and water. Sci Total Environ 87/88:477–494.

Conclusions:

The purpose of this paper is to explore and discuss the challenges in applying the Unit World Model (UWM) to metals with a view to developing an UWM critical loading model, EQCMetal, that can be applied to both metal ions and organic substances. In this context, log Kp values for suspended mater/water, sediment/water and soil/water are cited for Fe: 6.63, 4.90 and 2.74, respectively. These values are averaged from several literature data.

Reference 3

Endpoint:

adsorption / desorption, other

Remarks:

Field study

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Reason / purpose for cross-reference:

reference to same study

Qualifier:

no guideline followed

Principles of method if other than guideline:

Sediment, pore water, suspended matter, and water samples were taken in the Forsmark area, Baltic Sea, and analysed for Fe.

GLP compliance:

not specified

Type of method:

other: field study

Media:

other: sediment and suspended matter

Radiolabelling:

no

Test temperature:

between 2.3 and 12.3°C during sampling

Analytical monitoring:

yes

Details on sampling:

- Location and timing: all samples collected between 12-20 April 2005 in the bay between the north side of the islands of Stor and Lill-Tixlan, between the Forsmark nuclear power station and the town of Öregrund, NW Baltic Proper.
- Integrated water samples (10 L) collected from 0-4 m depth using a metal-free pump.
- Particulate organic matter in the water column was sampled through filtration of the water samples.
- Sediment: Kajak cores were taken at 7-8 m depth and sliced into two sections (0-3 cm and 3-6 cm).
- Pore water was extracted by centrifugation (20 min at 4500 rpm) from sediment samples.

Details on matrix:

no details reported

Details on test conditions:

field test

Computational methods:

Partitioning coefficients were calculated by dividing Fe concentration in solid phase by Fe concentration in water (L/kg).
Kpsuspended matter-water was calculated as well as Kpsediment-pore water for upper layer (0-3 cm) and lower layer (3-6 cm).

Phase system:

suspended matter-water

Type:

log Kp

Value:

6.08 L/kg

Remarks on result:

other: concentrations in filtered water and particulate organic matter used

Phase system:

sediment-water

Remarks:

(pore water)

Type:

log Kp

Value:

5.32 L/kg

Remarks on result:

other: concentrations in sediment and pore water of upper sediment layer (0-3 cm) used

Details on results (Batch equilibrium method):

For the lower sediment layer (3-6 cm), a log Kpsediment-pore water of 4.0 was obtained.

Conclusions:

In this study, water, suspended matter, sediment, and pore water samples were taken in the Forsmark area, Baltic Sea, and analysed for Fe. The log Kpsuspended matter-water was calculated to be 6.08. The log Kpsediment-pore water for the upper sediment layer (0-3 cm) was determined to be 5.32.

Reference 4

Endpoint:

adsorption / desorption, other

Remarks:

Field study

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Reason / purpose for cross-reference:

reference to same study

Qualifier:

no guideline followed

Principles of method if other than guideline:

Sediment and pore water samples were collected from two rivers which receive wastewater from urban Hanoi (Vietnam) and analysed for Fe.

GLP compliance:

not specified

Type of method:

other: field study

Media:

sediment

Radiolabelling:

no

Test temperature:

not reported

Analytical monitoring:

yes

Details on sampling:

- Sediment samples collected from To Lich and Kim Nguu rivers, both of which receive wastewater from urban Hanoi, Vietnam. Samples collected from three sites in both rivers with a stainless steel Kajak core sediment sampler equipped with a polymethylmethacrylat inner liner with an inner diameter of 46 mm. Three replicates were collected from each site. Core samples were subdivided into sections 0-10cm, 10-20cm, 20-30cm. Sediment samples were dried at 45°C until constant weight, passed through a 2 mm stainless steel sieve and pulverised in an agate mortar.
- Pore water was extracted from sediment samples: sediment from 0-10cm depth was transferred to polypropylene büchner funnel with 25 µm mesh nylon filter and a minimum of 15 mL pore water was extracted under suction of 10 kPa. Pore water was filtered through a 0.45 µm nylon filter (Millipore) and acidified with 0.1 mL 70% HNO3 (Baker Instra-Analysed).

Details on matrix:

- % organic carbon: 1.2-5.3% in To Lich river samples, 1.8-10.6% in Kim Nguu rivers
- pH of pore water was 7.4-8.1
- redox potential of pore water was -257 to -185 mV

Details on test conditions:

field study

Computational methods:

Partitioning coefficients were calculated by dividing Fe concentration in sediment by Fe concentration in pore water (L/kg).

Phase system:

sediment-water

Remarks:

(pore water)

Type:

log Kp

Value:

3.97 - 5.66 L/kg

% Org. carbon:

1.2 - 10.6

Remarks on result:

other: range for all paired samples

Conclusions:

In this study, samples of sediment and pore water were taken along two rivers receiving wastewater from Hanoi, Vietnam, and analysed for Fe. Log Kpsediment-pore water values of 3.97 to 5.66 were reported.

Reference 5

Endpoint:

adsorption / desorption, other

Remarks:

Field study

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Qualifier:

no guideline followed

Principles of method if other than guideline:

Sediment and surface water samples were collected from Onion Creek and one tailing pond (Van Stone mine (VSM) region, Washington, USA) and analysed for Fe.

GLP compliance:

not specified

Type of method:

other: field study

Media:

sediment

Radiolabelling:

no

Test temperature:

The average temperature of Onion Creek water was 9.5°C.

Analytical monitoring:

yes

Details on sampling:

During the 1992 annual low-discharge season (August-October), 23 water and 18 sediment samples were collected from VSM over a distance of 5.3 km:
- Water samples: Water samples were collected in acid washed 1000-ml Nalgene ® bottles. Sample bottles were immersed upside-down below the water surface (0-20 cm) on the upstream side, inverted and allowed to fill.
- Sediment samples: Stream sediments were scooped from the middle of Onion Creek so that samples collected were representative of the entire drainage area and all the locations of the studied area. Samples were collected immediately upstream of confluences. Sampling of sediment from the tailing pond was restricted to its periphery.

For background concentrations, uncontaminated water and sediment samples were collected from locations where the stream flows upstream and is away from the mining area and the tailings ponds.

The samples were acidified, refrigerated, and later processed in the laboratory within a week. Temperature and pH (Oakton Model 35624) conditions were measured in the field. Conductivity was measured in the laboratory (Radiometer Copenhagen CDM 83) in unacidified samples.

Matrix type:

other: sediment collected in the field

Details on test conditions:

Field study

Computational methods:

Partition coefficient (Kd) expresses the fractionation of a contaminant between the solid and liquid (Zhang et al., 1994) and is represented as:
Kd (in L/kg) = 10E6 Cp / Cw
where Cp is the concentration in the solid phase (mg/kg); and Cw is the concentration in stream water (ng/L).

Phase system:

sediment-water

Remarks:

(surface water)

Type:

log Kp

Value:

4.5 L/kg

Remarks on result:

other: values calculated based on average concentration near the mine and tailings ponds

Conclusions:

In this study, samples of sediment and surface water were taken from Onion Creek and one tailing pond (Van Stone mine (VSM) region, Washington, USA) and analysed for Fe. A log Kpsediment-surface water value of 4.5 was reported.

Reference 6

Endpoint:

adsorption / desorption, other

Remarks:

Field study

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Qualifier:

no guideline followed

Principles of method if other than guideline:

A suite of trace metals (including Fe) was analyzed in surface water and sediment samples from the Blesbokspruit, a Ramsar certified riparian wetland, to assess the impact of mining on the sediment quality and the fate of trace metals in the environment.

GLP compliance:

not specified

Type of method:

other: field study

Media:

sediment

Radiolabelling:

no

Test temperature:

Temperature ranged from 5.7 to 24.2 °C over the 20 sampling sites.

Analytical monitoring:

yes

Details on sampling:

Surface water and sediment samples were collected from 20 randomly chosen sites that were geographically dispersed throughout the marsh. Care was taken to include sites both upstream and downstream of a liming plant that treats and ultimately drains the effluent from tailings dams into the wetland, and the site where mine water from Grootvlei Gold Mine is discharged into the Blesbokspruit.

Water samples: At each site, two surface water samples were collected and filtered through a 0.45 µm nylon membrane filter using a hand-held vacuum pump. One of the samples was acidified with 3 M HNO3 to a pH of less than 2 to prevent metal precipitation. The water samples were stored in plastic bottles pre-rinsed with HNO3 and deionized water. The bottles were sealed and stored at 4 °C before analysis.

Sediment samples: At each of the 20 sample sites, surface sediments were sampled close to the sediment–water interface by completely inserting an inverted 50 mL polypropylene centrifuge vial. After digging the vial out, they were capped without leaving any head space. Care was taken to prevent further exposure of the sample to the atmosphere by storing the vials in an anaerobic jar where anoxic conditions were maintained using a BBL(R) gas pack. The jars were stored on ice for transportation to the laboratory for further analyses.

Matrix type:

other: sediment collected in the field

Details on matrix:

The ranges below cover the measurements performed over the 20 sampling sites:
- % gravel: 0 to 44
- % sand: 11 to 85
- % silt: 8 to 73
- % clay: 43
- % organic carbon: 0.5 to 9

Sediment samples were analyzed for their grain size following a protocol adopted from the non-affiliated soil analysis work committee (Anon., 1990) and from Moore and Reynolds (1997). Organic C content was determined using a CHN analyzer. Prior to the analysis, 0.5 g of wet sediment samples were first treated with 2 mL of 50% (vol/vol) HCl to dissolve any carbonate fraction. The carbonate free sediment was washed with 5 mL of 1 M HCOONH4 and dried at 40 °C. Subsequently, a known weight of dried sediment sample was combusted in the CHN analyzer to determine the organic C content.

Details on test conditions:

Field study

Computational methods:

To what extent the trace metals are partitioned between the solution and the solid phase is quantified in terms of partition coefficient (Kd) and is calculated according to: Kd (in L/kg) = Cs / Cw, where Cw and Cs denote the trace metal concentration in the water and solid phases, respectively.
Under field conditions, the Kd approach describes removal of trace metals from solution by multiple mechanisms such as, adsorption, precipitation and coprecipitation with major elements (Siegel, 2002). Such a lumped process approach provides a Kd value that is valid only for the system in which it is measured (Langmuir, 1997). When the contaminant of interest is already present in the sediments, the in situ Kd or apparent Kd values retrieved following the lumped approach are far more relevant, though (Langmuir, 1997).

Phase system:

sediment-water

Remarks:

(surface water)

Type:

log Kp

Value:

4.38 - 5.81 L/kg

Remarks on result:

other: range over the 20 sampling sites

Phase system:

sediment-water

Remarks:

(surface water)

Type:

log Kp

Value:

5.08 L/kg

Remarks on result:

other: average over the 20 sampling sites

The results above presented in the field "Partition coefficients" are the values measured in situ during the study. For a comparison purpose, the publication also reports log Kp for sediment/surface water measured in other riverine environments:

- World average: 6.08 (Li 2000)

- Acide drainage stream, Adak: 1.23 to 3.95 (Bhattacharya et al. 2006)

- Huanghe: 5.88 to 7.47 (Zhang et al. 1994)

- Mississippi: 6.95 to 7.62 (Trefry et al. 1986)

- St-Lawrence: 5.96 (Yeats & Bewers 1982)

-------------------------------

Bhattacharya, A., Routh, J., Jacks, G., Bhattacharya, P., Mörth, M., 2006. Environmental quality assessment of abandoned mine tailings in Adak, Västerbotten district (northern Sweden). Appl. Geochem. 21(10), 1760–1780.

Li, Y.-H., 2000. A Compendium of Geochemistry: From Solar Nebula to the Human Brain. Princeton University Press, Princeton.

Trefry, J.M., Nelson, T.A., Trocine, R.P., Metz, S., Vetter, T.W., 1986. Trace metal fluxes through the  Mississippi River delta system.In: Kullenberg, G. (Ed.), Rapports et Proces-Verbaux des Reunions Conseil International pour l’Exploration de la Mer, pp. 277–288.

Yeats, P.A., Bewers, J.M., 1982. Discharge of metals from St. Lawrence River. Can. J. Earth Sci. 19, 982–992.

Zhang, J., Huang, W.W., Wang, J.H., 1994. Trace-metal chemistry of the Huanghe (Yellow River), China – examination of the data from in situ measurements and laboratory approach. Chem. Geol. 114, 83–94.

Conclusions:

In this study, samples of sediment and surface water were taken from rom the Blesbokspruit, a Ramsar certified riparian wetland, and analysed for Fe. An averaged log Kpsediment-surface water value of 5.08 was reported.

Reference 7

Endpoint:

adsorption / desorption, other

Remarks:

Field study

Type of information:

experimental study

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

study well documented, meets generally accepted scientific principles, acceptable for assessment

Reason / purpose for cross-reference:

reference to same study

Qualifier:

no guideline followed

Principles of method if other than guideline:

Water and suspended matter samples were collected from 54 Czech rivers at 119 sampling locations and analysed for Fe.

GLP compliance:

not specified

Type of method:

other: field study

Media:

suspended matter

Radiolabelling:

no

Test temperature:

not reported

Analytical monitoring:

yes

Details on sampling:

Samples taken from 54 Czech rivers at 119 localities over the whole state territory in summers 1997 and 1998 under stable hydrological conditions.
- Water samples: filtered (0.4 µm) and unfiltered samples taken. Acidification by adding 1 mL of HNO3 1:1 purified by sub-boiling distillation on the day of collection, to attain a pH of about 1.5.
- Sediment samples: 200 mL water was generally filtered, and the SPM deposited on the filters was dried at 105°C.

Details on matrix:

The river waters had a mean pH of 7.74 (6.9-8.8), ionic strength (I) 7.8 mmol/L, specific conductance 538 µS/cm at 25°C, alkalinity 1.9 mmol/L, and moderate mean contents of SPM 9.9 mg/L (1.0-124 mg/L).

Details on test conditions:

Field study.

Computational methods:

Log Kp suspended matter was calculated dividing concentrations in SPM (suspended particulate matter) (mg/kg) by concentrations in water (mg/L). The Kp values for which log values were reported in the publication were different because they divided the concentrations in SPM (mg/kg) by concentrations in water in µg/L.

Phase system:

suspended matter-water

Type:

log Kp

Value:

5.18 L/kg

Remarks on result:

other: calculated from median reported Fe concentrations in SPM (4.66 % = 46600 mg/kg) and aqueous phase (0.31 mg/L)

Phase system:

suspended matter-water

Type:

log Kp

Value:

5.34 L/kg

Remarks on result:

other: reported median Kp suspended matter value, but corrected for using aqueous Fe concentrations in mg/L instead of µg/L

Conclusions:

In this study, samples of suspended matter and water were taken from Czech rivers and analysed for Fe. Log Kp suspended matter was reported to be 5.34 (median of all samplings), and was 5.18 when calculated using median concentrations in water and suspended matter.

Reference 8

Endpoint:

adsorption / desorption: screening

Type of information:

(Q)SAR

Adequacy of study:

weight of evidence

Reliability:

2 (reliable with restrictions)

Rationale for reliability incl. deficiencies:

results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification

Justification for type of information:

1. SOFTWARE

EPI Suite version 4.1 (U.S. Environmental Protection Agency).

2. MODEL (incl. version number)

KOCWIN version 2.00.

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL

CAS number: 30399-84-9.
SMILES: O=C(O)CCCCCCCCCCCCCCC(C)C.

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL

- Defined endpoint: Soil adsorption coefficient (Koc).

- Unambiguous algorithm: KOCWIN estimates Koc with two separate estimation methodologies: using first-order Molecular Connectivity Index (MCI) and using log Kow (octanol-water partition coefficient); both methodologies are described in section “Methodology” of the KOCWIN Help.

* Estimation Using Molecular Connectivity Index: Koc is estimated with a QSAR utilizing Molecular Connectivity Index (MCI). This QSAR estimation methodology is described completely in a journal article (Meylan et al, 1992) and in a report prepared for the US EPA (SRC, 1991). Two separate regressions were performed. The first regression related log Koc of non-polar compounds to the first-order MCI. Non-polar compounds are now designated as "compounds having no correction factors" which simply means the MCI descriptor alone can adequately predict the Koc. Measured log Koc values were fit to a simple linear equation of the form:
log Koc = a MCI + b; where a and b are the coefficients fit by least-square analysis. The 69 compounds used for this regression are listed in Appendix E of the KOCWIN Help.
The equation derived by the non-polar (no correction factor) regression is:
log Koc = 0.5213 MCI + 0.60 (n = 69, r2 = 0.967, std dev = 0.247, avg dev = 0.199)
The second regression included the 447 compounds having correction factors; these compounds are listed in Appendix F. The correction factors descriptors are listed in Appendix D. Correction factors are specific chemical classes or structural fragments. The regression coefficients were derived via multiple linear regression of the correction descriptors to the residual error of the prediction from the non-polar equation. Adding in the correction factor regression yields the final MCI equation:
log Koc = 0.5213 MCI + 0.60 + ΣPfN; where ΣPfN is the summation of the products of all applicable correction factor coefficients from Appendix D multiplied by the number of times (N) that factor is counted for the structure.

* Estimation Using Log Kow: A traditional method of estimating soil adsorption Koc involves correlations developed with log octanol-water partition coefficient (log Kow) (Doucette, 2000). Since an expanded experimental Koc database was available from the new MCI regression, it was decided to develop a log Kow estimation methodology that was potentially more accurate than existing log Kow QSARs for diverse structure datasets. Effectively, the new log Kow methodology simply replaces the MCI descriptor with log Kow and derives similar equations. The derivation uses the same training and validation data sets. The training set is divided into the same non-polar (no correction factors) and correction factor sets. The same correction factors are also used. Separate equations correlating log Koc with log Kow were derived for nonpolar and polar compounds because it was statistically more accurate to do so than to use the approach taken with the MCI-based method. The equation derived by the non-polar (no correction factor) regression is the following and is used for any compound having no correction factors:
log Koc = 0.8679 Log Kow - 0.0004 (n = 68, r2 = 0.877, std dev = 0.478, avg dev = 0.371)
For the multiple-linear regression using correction factors, log Kow was included as an individual descriptor. For compounds having correction factors, the equation is:
log Koc = 0.55313 Log Kow + 0.9251 + ΣPfN; where ΣPfN is the summation of the products of all applicable correction factor coefficients from Appendix D multiplied by the number of times (N) that factor is counted for the structure.

- Defined domain of applicability: The intended application domain of EPI Suite is organic chemicals. Inorganic and organometallic chemicals are generally outside the domain. The training and validation sets of KOCWIN include molecular weight ranges between 32.04 and 665.02 g/mol and between 73.14 and 504.12 g/mol, respectively. The correction factor descriptors covered by KOCWIN and the maximum number of instances of that factor in any of the 447 training set compounds can be found in Appendix D of the KOCWIN Help.

- Appropriate measures of goodness-of-fit and robustness and predictivity:

* Estimation Using Molecular Connectivity Index:
Training dataset includes a total of 69 compounds having no correction factors and 447 compounds having correction factors, with the following statistics:
number in dataset = 69 (no correction factors) / 447 (with correction factors)
correlation coef (r2) = 0.967 (no correction factors) / 0.900 (with correction factors)
standard deviation = 0.247 (no correction factors) / 0.340 (with correction factors)
average deviation = 0.199 (no correction factors) / 0.273 (with correction factors)
The residual estimation error of the training dataset is the following:
within <= 0.20 - 44.2%
within <= 0.40 - 76.9%
within <= 0.60 - 93.0%
within <= 0.80 - 98.6%
within <= 1.00 - 100%
Validation dataset includes a total of 158 compounds with the following statistics:
number in dataset = 158
correlation coef (r2) = 0.850
standard deviation = 0.583
average deviation = 0.459
The residual estimation error of the validation dataset is not available in KOCWIN Help

* Estimation Using Log Kow:
Training dataset includes a total of 68 compounds having no correction factors and 447 compounds having correction factors, with the following statistics:
number in dataset = 68 (no correction factors) / 447 (with correction factors)
correlation coef (r2) = 0.877 (no correction factors) / 0.855 (with correction factors)
standard deviation = 0.478 (no correction factors) / 0.396 (with correction factors)
average deviation = 0.371 (no correction factors) / 0.307 (with correction factors)
Validation dataset includes a total of 150 compounds with the following statistics:
number in dataset = 150
correlation coef (r2) = 0.778
standard deviation = 0.679
average deviation = 0.494
The residual estimation error of the training and validation datasets is not available in KOCWIN Help

- Mechanistic interpretation:

* Estimation Using Molecular Connectivity Index: The first-order molecular connectivity index (MCI) and a series of group contribution factors (correction factors) are used to predict Koc values. In the Molecular Correction Index information on the chemical structure, i.e. molecular size, branching, cyclisation, unsaturation and (to a certain extent) heteroatom content are encoded. The different influences of chemical classes or functional groups are considered by correction factors.

* Estimation Using Log Kow: Koc estimate is based upon log Kow (rather than MCI). The tendency of a compound to adsorb to organic carbon is linked with its lipophilicity. The n-octanol/water partition coefficient is one descriptor for lipophilicity.

5. APPLICABILITY DOMAIN

- Descriptor domain: The molecular weight of the target chemical (isostearic acid, C18H36O2) is 284.48 g/mol, hence within the estimation domain of the training (32.04 and 665.02 g/mol) and validation (73.14 and 504.12 g/mol) sets.

- Structural and mechanistic domains: The KOCWIN results mention one fragment for which a correction factor was developed (organic acid, -CO-OH). The occurrence of this fragment does not exceed the maximum for all training set compounds (see detailed data in “Any other information on results incl. tables”).

- Similarity with analogues in the training set: It was checked whether there are close analogues in the training / validations sets of KOCWIN. No fatty acids were found.

6. ADEQUACY OF THE RESULT

As above described, the QSAR model is valid with respect to the five OECD principles on scientific validity (a defined endpoint, an unambiguous algorithm, a defined domain of applicability, appropriate measures of goodness-of-fit / robustness / predictivity, a mechanistic interpretation). The target substance (isostearic acid) was demonstrated to have a molecular weight (MW) falling within the MW range of the training set compounds and the occurrence of the organic acid (-CO-OH) fragment does not exceed the maximum for all training set compounds. As part of a weight-of-evidence approach, the present prediction is thus adequate to fill in the Koc requirement of REACH Annex VIII.

Qualifier:

no guideline followed

Principles of method if other than guideline:

QSAR estimation using KOCWIN version 2.00 (EPI Suite version 4.1, U.S. Environmental Protection Agency).

GLP compliance:

no

Type of method:

other: QSAR calculation

Media:

soil

Specific details on test material used for the study:

SMILES notation: O=C(O)CCCCCCCCCCCCCCC(C)C

Type:

Koc

Value:

9 838

Remarks on result:

other: Koc estimate from MCI

Type:

log Koc

Value:

3.99

Remarks on result:

other: log Koc estimate from MCI

Type:

Koc

Value:

32 270

Remarks on result:

other: Koc estimate from Log Kow

Type:

log Koc

Value:

4.51

Remarks on result:

other: log Koc estimate from Log Kow

KOCWIN Predictions for Isostearic acid:

 

Koc Estimate from MCI
First Order Molecular Connectivity Index	9.626
Non-Corrected Log Koc (0.5213 MCI + 0.60)	5.6178
Fragment Correction: Organic Acid (-CO-OH)	-1.6249
Corrected Log Koc	3.9929
Estimated Koc	9838 L/kg

 

Koc Estimate from Log Kow
Log Kow (Kowwin estimate)	7.87
Non-Corrected Log Koc (0.55313 logKow + 0.9251)	5.2782
Fragment Correction: Organic Acid (-CO-OH)	-0.7694
Corrected Log Koc	4.5089
Estimated Koc	3.227e+004 L/kg

Applicability domain:

The KOCWIN results mention one fragment for which a correction factor was developed. The occurrence of this fragment does not exceed the maximum for all training set compounds:

Correction Factor Descriptor	MCI Methodology			Log Kow Methodology*
	Coef.	Nb of compounds	Max. occurrence per structure	Coef.
Organic Acid (-CO-OH)	-1.6249	21	1	-0.7694

* The correction factors and occurrences are determined exactly the same as for the MCI methodology.

Conclusions:

The KOCWIN-estimated log Koc of isostearic acid is 3.99 when estimated from MCI and 4.51 when estimated from log Kow.

Description of key information

Although iron oxide isostearate is not expected to give rise to high dissolved concentrations available for adsorption to particulate matter (see IUCLID section 4.8), existing data on the inorganic iron part and the organic isostearate part were nevertheless examined because considered relevant and informative of the environmental fate of the registered substance with respect to adsorption/desorption behaviour:

Inorganic iron part: A total of 7 studies was used in a weight of evidence approach to cover the endpoint. Data were available for soil, suspended matter, and sediment. The following final key values were retained for iron: a log Kp of 6.30 for suspended matter-water, a log Kp of 5.25 for sediment-water, and a log Kp of 2.74 for soil-water.

Organic isostearate part: A prediction derived from a valid (Q)SAR model (KOCWIN version 2.00, EPI Suite version 4.1, U.S. Environmental Protection Agency) and falling into its applicability domain, with adequate and reliable documentation / justification, was used for isostearic acid. The KOCWIN-estimated log Koc of isostearic acid is 3.99 when estimated from MCI and 4.51 when estimated from log Kow.

Additional information

Inorganic iron part:

In total, 7 studies were selected as useful for covering the adsorption/desorption endpoint using a weight of evidence approach. Adsorption/desorption is evaluated on an elemental basis and data were available for soil, sediment, and suspended matter. To determine a final key value, a single average (arithmetic mean) log Kp value was retained for each study. When some articles reported log Kp values from other references (Roychoudhury and Starke 2006, Harvey et al. 2007), individual average values (arithmetic mean) were retained for each study cited.

For suspended matter, three studies were identified as useful (Veselý et al. 2001, Harvey et al. 2007, Kumblad and Bradshaw 2008). The 10th, 50th and 90th percentile of the retained values was 5.67, 6.30 and 6.90, respectively. The median of 6.30 was taken as key log Kp for characterising distribution between suspended matter and water.  

For sediment, six studies were included in the weight of evidence approach (Chen et al. 1996, Routh and Ikramuddin 1996, Roychoudhury and Starke 2006, Harvey et al. 2007, Kumblad and Bradshaw 2008, Marcussen et al. 2008). The 10th, 50th and 90th percentile of the retained values was 4.54, 5.25 and 6.50, respectively. The median of 5.25 was taken as key log Kp for characterising distribution between sediment and water.

For soil, the study of Harvey et al. (2007) was used to characterize the adsorption potential. Because there is a limited amount of values available (i.e. four locations investigated in one study), the average (arithmetic mean) log Kp of 2.74 is selected as key value for characterising distribution between soil and water.

Overall, the obtained adsorption coefficients were similar as for many other metals. Adsorption to soil appears to be mild, however a stronger adsorption of iron to suspended matter and sediment seems to occur.

Organic isostearate part:

The soil adsorption coefficient (Koc) of isostearic acid was predicted using the QSAR model KOCWIN v2.00 (EPI Suite version 4.1, U.S. Environmental Protection Agency). This QSAR model is valid with respect to the five OECD principles on scientific validity (a defined endpoint, an unambiguous algorithm, a defined domain of applicability, appropriate measures of goodness-of-fit / robustness / predictivity, a mechanistic interpretation). Isostearic acid was demonstrated to have a molecular weight (MW) falling within the MW range of the training set compounds and the occurrence of the organic acid (-CO-OH) fragment does not exceed the maximum for all training set compounds. KOCWIN estimates Koc with two separate estimation methodologies: using first-order Molecular Connectivity Index (MCI) and using log Kow (octanol-water partition coefficient). The KOCWIN-estimated log Koc of isostearic acid is 3.99 when estimated from MCI and 4.51 when estimated from log Kow.