Undecanal

Adsorption / desorption

Link to relevant study record(s)

Reference

Endpoint:

adsorption / desorption, other

Type of information:

(Q)SAR

Adequacy of study:

key study

Study period:

May 2018

Reliability:

1 (reliable without restriction)

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

Remarks:

Results from two relevant supporting models corroborate key model result.

Justification for type of information:

See attached QPRF and QMRF section "Attached justification".

Guideline:

other: ECHA Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.6: QSARs and grouping of chemicals

Version / remarks:

2008

Guideline:

other: ECHA Practical guide How to use and report (Q)SARs

Version / remarks:

2016

Principles of method if other than guideline:

Key model:
The organic carbon normalized adsorption coefficient (Koc) was calculated by “OPERA-model for organic carbon-sorption coefficient”, V1.5 (EPA, 2018). The model is freely accessible via the US EPA Chemistry Dashboard.

Supporing models used are implemented in Chemical Properties Estimation Software System (ChemProp) version 6.6 (UFZ, 2017). ChemProp includes different methods for calculation of the organic carbon normalized adsorption coefficient (Koc). It includes applicability domain checks specific for the respective models. As such, not all implemented models for Koc calculation could be used for the submission substance.
Supporting model (1):
Model by Huuskonen (Sw or Kow; 2003a). The log Koc is estimated based on water solubility, molecular weight, and structural determinants (number of aromatic 5 or 6 rings; number of oxygen and nitrogen atoms; presence of carboxylic acid residue).

Supporting model (2):
Model by Poole and Poole (LSER; 1999). Koc is calculated from linear solvation energy relationship (LSER) parameters.

Supporting model (3):
Model by Huuskonen (e-states; 2003b). The log Koc is estimated based on first order molecular connectivity index (MCI) and the sums Si of particular electrotopological states.

References:

- EPA, Environmental Protection Agency, 2018 (date of access)
Chemistry Dashboard, for calculation of Koc by “OPERA-model for organic carbon-sorption coefficient”, V1.5
Online: https://comptox.epa.gov/dashboard/
U.S. Environmental Protection Agency, Washington, DC, USA

- Huuskonen, J. (2003a)
Prediction of soil sorption coefficient of a diverse set of organic chemicals from molecular structure
Journal of Chemical Information and Computer Sciences, 43, 1457-1462

- Huuskonen, J. (2003b)
Prediction of soil sorption coefficient of organic pesticides from the atom-type electrotopological state indices
Environmental Toxicology and Chemistry, 22, 816-820

- Poole, S.K.; Poole, C.F. (1999)
Chromatographic models for the sorption of neutral organic compounds by soil from water and air
Journal of Chromatography A, 845, 381-400

- UFZ, Department of Ecological Chemistry (2017)
ChemProp 6.6
http://www.ufz.de/ecoc

Type of method:

other: QSAR

Media:

other: Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient,Koc

Specific details on test material used for the study:

• CAS number: 112-44-7
• EC number: 203-972-6
• Chemical name / common name: n-undecanal / 1-undecanal
• Structural formula: C11-H22-O
• Canonical SMILES: CCCCCCCCCCC=O
• Partition coefficient n-octanol water, log Kow: 4.8 at 20°C (EU method A.8; WoE)
• Water solubility: 12 mg/L at 20°C (OECD 105, flask method)

Test temperature:

Not applicable, QSAR

Details on sampling:

Not applicable, QSAR

Details on matrix:

Not applicable, QSAR

Details on test conditions:

Not applicable, QSAR

Computational methods:

For key model calculation by “OPERA-model for organic carbon-sorption coefficient”, V1.5, accessed via US Environmental Protection Agency Chemistry Dashboard, see attached QMRF and QPRF!
For supporting estimations using models included in Chemical Properties Estimation Software System (ChemProp) version 6.6 (UFZ, 2017), see below section "Any other information on materials and
methods including tables"!

Key result

Sample No.:

#1

Type:

Koc

Remarks:

QSAR

Value:

698 L/kg

Matrix:

Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient, Koc

Remarks on result:

other: Key model: US EPA OPERA-model for organic carbon-sorption coefficient V1.5

Key result

Sample No.:

#1

Type:

log Koc

Remarks:

QSAR

Value:

2.84 dimensionless

Matrix:

Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient, Koc

Remarks on result:

other: Key model: US EPA OPERA-model for organic carbon-sorption coefficient V1.5

Sample No.:

#2

Type:

log Koc

Remarks:

QSAR

Value:

2.83 dimensionless

Matrix:

Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient, Koc

Remarks on result:

other: Supporting model 1: Model by Huuskonen (Sw or Kow; 2003a) - log Koc estimated based on water solubility, MW, and structural determinants; model implemented in ChemProp v. 6.6 (UFZ, 2017)

Sample No.:

#3

Type:

log Koc

Remarks:

QSAR

Value:

2.89 dimensionless

Matrix:

Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient, Koc

Remarks on result:

other: Supporting model 2: Model by Poole and Poole (LSER; 1999) - log Koc calculated from linear solvation energy relationship (LSER) parameters; model implemented in ChemProp v. 6.6 (UFZ, 2017)

Sample No.:

#4

Type:

log Koc

Remarks:

QSAR

Value:

2.37 dimensionless

Matrix:

Soil / Sediment: partition coefficient to organic carabon in soil/sediment; organic carbon normalized adsorption coefficient, Koc

Remarks on result:

other: Supporting model 3: Model by Huuskonen (e-states; 2003b) - Model implemented in ChemProp v. 6.6 (UFZ, 2017)

Details on results (HPLC method):

Not applicable, QSAR

Adsorption and desorption constants:

Not applicable, QSAR

Recovery of test material:

Not applicable, QSAR

Concentration of test substance at end of adsorption equilibration period:

Not applicable, QSAR

Concentration of test substance at end of desorption equilibration period:

Not applicable, QSAR

Details on results (Batch equilibrium method):

Not applicable, QSAR

Statistics:

Not applicable, QSAR

Result from OPERA V1.5 Calculation (Key Model)

The organic carbon normalized adsorption coefficient (Koc) was calculated by “OPERA-model for organic carbon-sorption coefficient”, V1.5. The model is freely accessible via the US EPA Chemistry Dashboard ( https://comptox.epa.gov/dashboard/). The model fulfils the OECD principles for QSAR models with algorithm and used experimental data for model build and validation being freely available. A QSAR Model Reporting Format is published and available from the JRC QSAR Model Database (see IUCLID section "Attached justification").

The model is based on a distance weighted k-nearest neighbours (kNN) approach, with k=5. The method is a refinement of the classical k-NN classification algorithm where the contribution of each of the k neighbours is weighted according to their distance to the query point, giving greater weight to closer neighbours. The used distance is the Euclidean distance. kNN is an unambiguous algorithm that fulfils the transparency requirements of OECD principle 2 with an optimal compromise between model complexity and performance. The model descriptors were selected statistically but they can also be mechanistically interpreted. Details are given in the available QMRF attached to this RSS (see IUCLID section "Attached justification").

The model is applicable to heterogeneous organic chemicals and as such, the compound is in the general applicability domain. Also the global applicability domain was met, which is determined by means of the leverage approach that checks whether the query structure falls within the multidimensional chemical space of the whole training set. Further, a local applicability domain index is determined. The local approach investigates the vicinity of the query chemical. It provides a continuous index ranging from 0 to 1 and is relative to the similarity of the query chemical to its 5 nearest neighbours in the p dimensional space of the model. Based on a local index above average (0.53) the compound is overall within the applicability domain of the model. This is reflected by the confidence level index for n-undecanal of 0.65. This index is ranging from 0 to 1 relative to the accuracy of prediction of the 5 nearest neighbours to the query chemical. The higher this index, the more the prediction is likely to be reliable. In addition, predicted and calculated results for the 5 nearest neighbours used for the prediction were checked: these match closely, which increases confidence in the estimated value for the submission substance. As such, the estimated value for the Koc resulting from the “OPERA-model for organic carbon-sorption coefficient” is adequate for environmental exposure and risk assessment for the submission substance.

In line with this interpretation, the Koc prediction for the submission substance is also rated as ‘inside global applicability domain’ of the OPERA model on the EPA Chemistry Dashboard.

Estimation result:

log Koc (OPERA V1.5): 2.84 (Koc: 698 L/kg)

For further information, please see corresponding QMRF (QMRF_OPERA_KOC_2017-10-17.pdf) and QPRF (QPRF_n-Undecanal_OPERA-Koc.pdf) attached to this RSS (see IUCLID section "Attached justification").

Results from supporting models using ChemPropv. 6.6 (UFZ, 2017)

Results

The following results were obtained using the three different models implemented in ChemProp v. 6.6 as outlined in section "Any other information on materials and methods including tables:"

-      log Koc = 2.83 with Huuskonen (Sw) model (equation 2)

-      log Koc = 2.89 with Poole and Poole (LSER) model

-      log Koc = 2.37 with Huuskonen (e-states) model

 

 Discussion on supporting model results

The resulting Koc values from the Huuskonen (e-states and Sw) and Poole and Poole (LSER) models range between log Koc 2.37 to log Koc 2.89.

With regard to the Huuskonen models, judging from the respective published information on test and training sets as well as statistical measures for prediction accuracy (see section "Any other information on materials and methods including tables"), the model Huuskonen (Sw) provides more accurate predictions and is based on a much larger training set (N=403, ca. 3-times larger compared to e-states model) and validation set (N=165, ca. 4-times larger compared to e-states model). Importantly, the compound set contained a large fraction of pesticides with diverse structures. Thus, the value of log Koc of 2.83 is regarded to be most relevant with regard to the two Huuskonen models.

The value according to the Poole and Poole (1999) model based on LSER of log Koc 2.89 is very close to the relevant value of 2.83 according to the Huuskonen (Sw) model. While the training set was smaller compared to the Huuskonen (Sw) model, i.e. 131 compounds, measures for prediction accuracy were excellent, and the chemical mechanism underlying partitioning from water to wet soil is very well accounted for using the solvation parameter model. The latter is giving confidence that the model reliably represents sorption over a varied range of compound types. This renders the prediction of Koc for n-undecanal from the Poole and Poole model at least equally reliable compared to the Huuskonen (Sw) model.

Validity criteria fulfilled:

yes

Remarks:

with regard to OECD principles for (Q)SAR validation as well as ECHA guidance document R.6 (2008)

Conclusions:

The organic carbon normalized adsorption coefficient (Koc) was determined within a reliable QSAR study. The key value was determined using the valid (QMRF) key model "OPERA-model for organic carbon-sorption coefficient" (V1.5) available from the US EPA Chemistry Dashboard. Results were demonstrated to be reliable (QPRF) and were corroborated by 2 valid supporting models.
Result:
log Koc = 2.84 (rounded)
Koc = 698 L/kg

Executive summary:

The organic carbon normalized adsorption coefficient (Koc) was determined within a reliable QSAR study. The key value was determined using the valid (see QMRF) key model "OPERA-model for organic carbon-sorption coefficient" (V1.5) available from the US EPA Chemistry Dashboard. Results were demonstrated to be reliable (see QPRF) and were corroborated by 2 valid supporting models.   

Overall, the following log Koc values were estimated:

Key model:                                                                          2.84 (rounded, equivalent to Koc = 698 L/kg)

Supporting model 1 (Huuskonen (Sw)):                                2.83

Supporting model 2 (Poole and Poole (LSER)):                    2.89

Both supporting models and their average (log Koc 2.86) are very close to the value calculated according to the key model OPERA V1.5 (log Koc 2.84) and thus in strong support of the key value used for environmental exposure and risk assessment.

In addition, a third model (Huuskonen, e-states) was applied, and a log Koc of 2.37 resulted from this model. However, with regard to the Huuskonen models, judging from the respective published information on test and training sets as well as statistical measures for prediction accuracy (see section "Any other information on materials and methods including tables"), the model Huuskonen (Sw) provides more accurate predictions and is based on a much larger training set (N=403, ca. 3-times larger compared to e-states model) and validation set (N=165, ca. 4-times larger compared to e-states model). Importantly, the compound set contained a large fraction of pesticides with diverse structures. Thus, the value of log Koc of 2.83 is regarded to be most relevant with regard to the two Huuskonen models and was selected as supporting model 1.

The value according to the Poole and Poole (1999) model based on LSER of log Koc 2.89 is very close to the relevant value of 2.83 according to the Huuskonen (Sw) model (supporting model 1). While the training set was smaller compared to the Huuskonen (Sw) model, i.e. 131 compounds, measures for prediction accuracy were excellent, and the chemical mechanism underlying partitioning from water to wet soil is very well accounted for using the solvation parameter model. The latter is giving confidence that the model reliably represents sorption over a varied range of compound types. This renders the prediction of Koc for n-undecanal from the Poole and Poole model at least equally reliable compared to the Huuskonen (Sw) model and the Poole and Poole (LSER) model was selected as supporting model 2.

The resulting log Koc of 2.84 (Koc: 698 L/kg) characterizes the submission substance as significantly adsorbing to the organic carbon fraction of solids (e.g. for freshwater, ca. 10% of total (bulk concentration) will be adsorbed to suspended particle fraction of freshwater). McCall et al. (1981) classified compounds with regard to mobility in soil as medium mobile with Koc between 150 and 500 L/kg and low mobile with Koc between 500 and 2000 L/kg. The submission substance must therefore be regarded as low mobile in soil.

Conclusion:

Relevant organic carbon normalized adsorption coefficient (Koc):

Koc = 698 L/kg

 

Reference:

McCall, P.J.; Laskowski, D.A.; Swann, R.L.; Dishburger, H.J. (1981)

Measurement of sorption coefficients of organic chemicals and their use in environmental fate analysis

In: AOAC, Association of Official Analytical Chemists, Test Protocols for Environmental Fate & Movement of Toxicants, Proceedings of a Symposium Association of Official Analytical Chemists, 94th Annual Meeting, October 21, 22, 1980, Washington, DC, 89-109

Description of key information

The organic carbon normalized adsorption coefficient (Koc) was determined within a reliable QSAR study. The key value was determined using the valid (see QMRF) key model "OPERA-model for organic carbon-sorption coefficient" (V1.5) available from the US EPA Chemistry Dashboard. Results were demonstrated to be reliable (see QPRF) and were corroborated by 2 valid supporting models.  

Result:

log Koc = 2.84 (rounded)

Koc = 698 L/kg

Key value for chemical safety assessment

Koc at 20 °C:

698

Additional information

The organic carbon normalized adsorption coefficient (Koc) was determined within a reliable QSAR study.The key value was determined using the valid (see QMRF) key model "OPERA-model for organic carbon-sorption coefficient" (V1.5) available from the US EPA Chemistry Dashboard. Results were demonstrated to be reliable (see QPRF) and were corroborated by 2 valid supporting models.   

Overall, the following log Koc values were estimated:

Key model:                                                                          2.84 (rounded, equivalent to Koc = 698 L/kg)

Supporting model 1 (Huuskonen (Sw)):                                2.83

Supporting model 2 (Poole and Poole (LSER)):                    2.89

Both supporting models and their average (log Koc 2.86) are very close to the value calculated according to the key model OPERA V1.5 (log Koc 2.84) and thus in strong support of the key value used for environmental exposure and risk assessment.

In addition, a third model (Huuskonen, e-states) was applied, and a log Koc of 2.37 resulted from this model. However, with regard to the Huuskonen models, judging from the respective published information on test and training sets as well as statistical measures for prediction accuracy(see section "Any other information on materials and methods including tables"), the model Huuskonen (Sw) provides more accurate predictions and is based on a much larger training set (N=403, ca. 3-times larger compared to e-states model) and validation set (N=165, ca. 4-times larger compared to e-states model). Importantly, the compound set contained a large fraction of pesticides with diverse structures. Thus, the value of log Koc of 2.83 is regarded to be most relevant with regard to the two Huuskonen models and was selected as supporting model 1.

The value according to the Poole and Poole (1999) model based on LSER of log Koc 2.89 is very close to the relevant value of 2.83 according to the Huuskonen (Sw) model (supporting model 1). While the training set was smaller compared to the Huuskonen (Sw) model, i.e. 131 compounds, measures for prediction accuracy were excellent, and the chemical mechanism underlying partitioning from water to wet soil is very well accounted for using the solvation parameter model. The latter is giving confidence that the model reliably represents sorption over a varied range of compound types. This renders the prediction of Koc for n-undecanal from the Poole and Poole model at least equally reliable compared to the Huuskonen (Sw) model and the Poole and Poole (LSER) model was selected as supporting model 2.

The resulting log Koc of 2.84 (Koc: 698 L/kg) characterizes the submission substance as significantly adsorbing to the organic carbon fraction of solids (e.g. for freshwater, ca. 10% of total (bulk concentration) will be adsorbed to suspended particle fraction of freshwater). McCall et al. (1981) classified compounds with regard to mobility in soil as medium mobile with Koc between 150 and 500 L/kg and low mobile with Koc between 500 and 2000 L/kg. The submission substance must therefore be regarded as low mobile in soil.

Conclusion:

Relevant organic carbon normalized adsorption coefficient (Koc):

Koc = 698 L/kg

 

Reference:

McCall, P.J.; Laskowski, D.A.; Swann, R.L.; Dishburger, H.J. (1981)

Measurement of sorption coefficients of organic chemicals and their use in environmental fate analysis

In: AOAC, Association of Official Analytical Chemists, Test Protocols for Environmental Fate & Movement of Toxicants, Proceedings of a Symposium Association of Official Analytical Chemists, 94th Annual Meeting, October 21, 22, 1980, Washington, DC, 89-109

[LogKoc: 2.84]