Registration Dossier

Environmental fate & pathways

Adsorption / desorption

Currently viewing:

Administrative data

Link to relevant study record(s)

Referenceopen allclose all

Endpoint:
adsorption / desorption, other
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
not applicable
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
EpiSuite (v4.11)

2. MODEL (incl. version number)
(PC)KOCWIN v2.00 - Kow Method

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
[Explain how the model fulfils the OECD principles for (Q)SAR model validation. Consider attaching the QMRF or providing a link]

4.1 - Defined endpoint:
Soil adsorption coefficient KOC given as a logarithmic value. (Remark: Definition of KOC according to Lyman et al, 1990: “the ratio of the amount of chemical adsorbed per unit weight of organic carbon (oc) in the soil or sediment to the concentration of the chemical in solution at equilibrium” Koc = (µg adsorbed/g organic carbon) / (µg/mL solution) [L/kg or mL/g])

4.2 - Unambiguous algorithm:
* Non-polar chemicals (i.e. compounds where no correction factor is needed): log Koc = 0.8679 Log Kow - 0.0004
* Polar chemicals (i.e. compounds where a correction factor is needed): log Koc = 0.55313 Log Kow + 0.9251 + ΣPfN

The equations were developed in two separate regression calculations since this approach is statistically more accurate than the approach taken in the MCI-method
For more information, see KOCWIN help.

The employed descriptors (with units) are the Log Kow: logarithm of the n-octanol/water partition coefficient; Pf: correction factor for chemical class of functional group f; N: number of times chemical class or functional group f occurs. The list of the correction factors of the training set is available in KOCWIN Help, Appendix D.

The training set comprises of non-polar set (68 chemicals), see KOCWIN help appendix E and a polar set (447 chemicals), see KOCWIN help appendix F, taken from several literature sources. One compound of the original non-polar training set (hexabromobiphenyl) was not considered since there was no recommended experimental log KOW.

4.3 - Defined domain of applicability:

Currently there is no universally accepted definition of model domain. Log Koc estimates are less accurate for compounds outside the MW range of the training set compounds and/or that have more instances of a given fragment than the maximum for all training set compounds. It is also possible that a compound may have a functional group(s) or other structural features not represented in the training set, and for which no fragment coefficient or correction factor was developed. The limits are:
* Molecular weight: 32.04-665.02 g/Mol
* Fragments and Functional groups according to “Training Sets” and correction factors for best results

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

The Training set is constituted of 68 non-polar chemicals and 447 polar chemicals (since there was no recommended experimental log KOW for one non-polar chemical of the training set (hexabromobiphenyl)).

Nonpolar compounds: n = 68; r2= 0.877, std. dev. = 0.478, average dev. = 0.371
Polar compounds: n = 447, r²=0.855, std. dev. = 0.396, average dev. = 0.307
For more details, see KOCWIN Help, appendices E and F (non-polar and polar chemicals).

4.5 - Mechanistic interpretation:
The tendency of a compound to adsorb itself on organic carbon is linked with its lipophilicity. The n-octanol/water partition coefficient is one descriptor for lipophilicity

5. APPLICABILITY DOMAIN
- Descriptor domain: Yes, the substance is in the range of the model
- Similarity with analogues in the training set: no very similar substance found
- Other considerations: to perform the study, a covalent liaison has been drawn between the 2 components of the substance, which is not the case.
Qualifier:
no guideline followed
Version / remarks:
QSAR determination of the Log Koc/Koc values
Principles of method if other than guideline:
- Principle of test: QSAR determination of the Log Koc/Koc values
GLP compliance:
no
Type of method:
other: QSAR estimation: KOCWIN v2.00: Koc estimate from log Kow
Specific details on test material used for the study:
c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1
Key result
Type:
Koc
Value:
ca. 3 873 L/kg
Remarks on result:
other: QSAR determination
Key result
Type:
log Koc
Value:
ca. 3.588

KOCWIN v2.00 results:

Koc Estimate from Log Kow:

Log Kow (user entered) : 3.90

Non corrected Log Koc (0.55313 log Kow + 0.9251) = 3.0823

Fragment corrections:

2 Nitrogen to non-fused aromatic ring: -0.0432

2 Nitrogen to carbon (aliphatic) (-N-c): -0.0436

1 Azo (-N=N-) : 0.43410

1 Miscellaneous S(=0) group: 0.1614

Corrected Log Koc: 3.5880

Estimated Kow: 3873 L/kg.

Validity criteria fulfilled:
yes
Executive summary:

 (PC)KOCWIN v2.00 - Kow Method

1. Substance c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1
2. General information
2.1 Date of QPRF August 2017
2.2 QMRF author and contact details See “Data Source (Reference)”
3. Prediction
3.1 Endpoint (OECD Principle 1) Endpoint Adsorption to solid phase of soils etc
Dependent variable Organic carbon normalised adsorption coefficient (Koc)
3.2 Algorithm (OECD Principle 2) Model or submodel name KOCWIN
Model version v. 2.00
Reference to QMRF QMRF: Estimation of Soil Adsorption Coefficient using KOCWIN v2.00 (EPI Suite v4.11): Estimation using log Kow
Predicted value (model result) See “Results and discussion”
Input for prediction - Chemical structure via SMILES
- log Kow (for log Kow method; optional)
Descriptor values - Chemical structure
- Correction factors
3.3 Applicability domain (OECD principle 3) Domains:
1) Molecular weight Substance within range (620 g/mol)
(range of test data set: 32.04 to 665.02 g/mol; KOCWIN User’s Guide, Ch. 6.2.4 Domain)
2) log Kow (range of test data set: -2.11 to 8.12; KOCWIN User’s Guide, Appendices E & F) Substance within range (3.9)
3) Correction factors: Number of instances of the identified correction factor does not exceed the maximum number as listed in Appendix D (KOCWIN User’s Guide) Fulfilled.
3.4 The uncertainty of the prediction (OECD principle 4) Training Set:
Nonpolar compounds: n = 68; r2= 0.877, std. dev. = 0.478, average dev. = 0.371
Polar compounds: n = 447, r²=0.855, std. dev. = 0.396, average dev. = 0.307
3.5 The chemical mechanisms according to the model underpinning the predicted result (OECD principle 5) Adsorption is caused by temporary (reversible) or permanent bonding between the substance and a surface (e.g. due to van der Waals interactions, hydrogen bonding to hydroxyl groups, ionic interactions, covalent bonding, etc.). The organic carbon normalized adsorption coefficient (Koc) is the ratio of a substance concentration sorbed in the organic matter component of soil or sediment to that in the aqueous phase at equilibrium.                                                                                                                                                                         Estimation using log Kow: The model is based on the high correlation of log Kow and the adsorption potential.

 

References

- KOCWIN User’s Guide (EPI Suite v4.11).

- Meylan, W., P.H. Howard and R.S. Boethling. 1992. Molecular topology/fragment contribution method for predicting soil sorption coefficients. Environ. Sci. Technol. 26: 1560-1567.

Assessment of estimation domain (molecular weight, fragments, correction factors):

  Molecular weight (g/mol) log Kow
Minimum Maximum Average                       Minimum Maximum
Training set 32.04 665.02 224.4 -2.11 8.12
Validation set 73.14 504.12 277.8 -5.98 8.68
 
Assessment of molecular weight Molecular weight within range of training set.
Assessment of log Kow (only relevant for log Kow method) Log Kow within range of training and validation sets.
Appendix D. MCI & Log Kow Corrections Factors for 447 Compound trainig set
Correction Factor Descriptor Coefficient for Molecular Connectivity Index (MCI) Regression Methodology Occurrence No. of instances of each bond found for the current substance
  (new model) Remark Coefficient for log Kow Regression Methodology (number of compounds) (max per structure)
Nitrogen to non-fused aromatic ring -0.522510 (a) -0.021606 154 2 2
Nitrogen to carbon (aliphatic) (-N-c) -0.212730 c -0.021787 133 5 2
Azo (-N=N-) -0.647525   0.431031 3 1 1
Miscellaneous S(=0) group -1.298009 (a) 0.161446 11 1 1
(a) Counted up to twice per structure, regardless of number of occurrences.
(b) Either one or both carbons aromatic; if both carbons aromatic, cannot be cyclic.
(c) Any nitrogen attached to double bond is not counted; also, carbonyl and thiocarbonyl are not counted as carbons.
(d) This is the only fragment counted, even if other fragments occur.
(e) Not included in regression derivation; estimated from other carbonyl fragments.
(f) Counted only once per structure, regardless of number of occurrences.
Endpoint:
adsorption / desorption, other
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
not applicable
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
EpiSuite (v4.11)

2. MODEL (incl. version number)
(PC)KOCWIN v2.00 - MCI Method
Estimation using first-order Molecular Connectivity Index (MCI)

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
[Explain how the model fulfils the OECD principles for (Q)SAR model validation. Consider attaching the QMRF or providing a link]

4.1 - Defined endpoint:
Soil adsorption coefficient KOC given as a logarithmic value. (Remark: Definition of KOC according to Lyman et al, 1990: “the ratio of the amount of chemical adsorbed per unit weight of organic carbon (oc) in the soil or sediment to the concentration of the chemical in solution at equilibrium” Koc = (µg adsorbed/g organic carbon) / (µg/mL solution) [L/kg or mL/g])

4.2 - Unambiguous algorithm:
* Non-polar chemicals (i.e. compounds where no correction factor is needed):
log Koc = 0.5213 MCI + 0.60

* Polar chemicals (i.e. compounds where a correction factor is needed):
log Koc = 0.5213 MCI + 0.60 + ΣPfN
The equation was developed in a two step regression approach: 1. Derivation of equation without correction factors using a set of non polar chemicals 2. Derivation of final equation using a set of non-polar chemicals.
This method estimated Koc solely 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). For more information, see KOCWIN help. The method does not consider log Kow for calculations.

The employed descriptors (with units) are the Log Kow: logarithm of the n-octanol/water partition coefficient; Pf: correction factor for chemical class of functional group f; N: number of times chemical class or functional group f occurs. The list of the correction factors of the training set is available in KOCWIN Help, Appendix D.

The training set comprises of non-polar set (68 chemicals), see KOCWIN help appendix E and a polar set (447 chemicals), see KOCWIN help appendix F, taken from several literature sources.

4.3 - Defined domain of applicability:
Currently there is no universally accepted definition of model domain. Log Koc estimates are less accurate for compounds outside the MW range of the training set compounds and/or that have more instances of a given fragment than the maximum for all training set compounds. It is also possible that a compound may have a functional group(s) or other structural features not represented in the training set, and for which no fragment coefficient or correction factor was developed. The limits are:
* Molecular weight: 32.04-665.02 g/Mol
* Fragments and Functional groups according to “Training Sets” and correction factors for best results

4.4 - Appropriate measures of goodness-of-fit and robustness and predictivity:
The Training set is constituted of 69 non-polar chemicals and 447 polar chemicals (total of 516 chemicals).
n = 516, r² = 0.916, std. dev. = 0.330, average dev. = 0.263
For more details, see KOCWIN Help, Chapter 6.2.1 Statistical Accuracy, Appendices E and F (non-polar and polar chemicals).

4.5 - Mechanistic interpretation:
The tendency of a compound to adsorb itself on organic carbon is linked with its lipophilicity. The n-octanol/water partition coefficient is one descriptor for lipophilicity

5. APPLICABILITY DOMAIN
- Descriptor domain: Yes, the substance is in the range of the model
- Similarity with analogues in the training set: no very similar substance found
- Other considerations: to perform the study, a covalent liaison has been drawn between the 2 components of the substance, which is not the case.
Qualifier:
no guideline followed
Version / remarks:
KOCWIN (v2.00): estimation using first-order Molecular Connectivity Index (MCI)
Principles of method if other than guideline:
KOCWIN (v2.00): estimation using first-order Molecular Connectivity Index (MCI)
GLP compliance:
no
Type of method:
other: QSAR estimation: KOCWIN v2.00: Koc estimate from MCI
Media:
soil
Specific details on test material used for the study:
c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1
Key result
Type:
Koc
Value:
ca. 347 000 000 L/kg
Remarks on result:
other: QSAR determination
Key result
Type:
log Koc
Value:
ca. 8.54 L/kg

KOCWIN v2.00 results:

Koc Estimate from MCI

First Orer Molecular Connectivity Index: 21.785

Non-corrected Log Koc (0.5213 MCI + 0.60): 11.9564

 

Fragment Corrections:

2 Nitrogen to non-fused aromatic ring : -1.0450

2 Nitrogen to Carbon (aliphatic (-N-C-): -0.4255

1 Azo (-N=N-):-0.6475

1 Miscellaneous S(=O) group: -1.2980

 

Corrected Log Koc : 8.5404

 

Estimated Koc: 3.47E8 L/kg

Executive summary:

KOCWIN v2.00

1. Substance c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1
2. General information
2.1 Date of QPRF August 2017
2.2 QPRF author and contact details See “Data Source (Reference)”
3. Prediction
3.1 Endpoint (OECD Principle 1) Endpoint Adsorption to solid phase of soils etc.
Dependent variable Organic carbon normalised adsorption coefficient (Koc)
3.2 Algorithm (OECD Principle 2) Model or submodel name KOCWIN
Model version v. 2.00
Reference to QMRF QMRF: Estimation of Soil Adsorption Coefficient using KOCWIN v2.00 (EPI Suite v4.11): MCI methodology
Predicted value (model result) See “Results and discussion”
Input for prediction - Chemical structure via SMILES
Descriptor values - MCI (first order molecular connectivity index)
- Correction factors
3.3 Applicability domain (OECD principle 3) Domains: Domains:
1) Molecular weight                              (range of test data set: 32.04 to 665.02 g/mol; KOCWIN User’s Guide, Ch. 6.2.4 Domain) 1) Molecular weight                (range of test data set: 32.04 to 665.02 g/mol; KOCWIN User’s Guide, Ch. 6.2.4 Domain)
 
2) Correction factors: Number of instances of the identified correction factor does not exceed the maximum number as listed in Appendix D (KOCWIN User’s Guide) 2) Correction factors: Number of instances of the identified correction factor does not exceed the maximum number as listed in Appendix D (KOCWIN User’s Guide)
3.4 The uncertainty of the prediction The Training datasets are combined in to one dataset of 516 compounds (69 having no corrections plus 447 with corrections). Statistical accuracy for training dataset with the MCI methodology is:
(OECD principle 4)  
  n = 516, r² = 0.916, std. dev. = 0.330, average dev. = 0.263
3.5 The chemical mechanisms according to the model underpinning the predicted result Adsorption is caused by temporary (reversible) or permanent bonding between the substance and a surface (e.g. due to van der Waals interactions, hydrogen bonding to hydroxyl groups, ionic interactions, covalent bonding, etc.). The organic carbon normalized adsorption coefficient (Koc) is the ratio of a substance concentration sorbed in the organic matter component of soil or sediment to that in the aqueous phase at equilibrium.
(OECD principle 5) MCI methodology: The first-order molecular connectivity index is a measure to describe a variety of properties of chemicals. According to Sabljic (1984; cited in Meylan et al., 1992), the soil sorption potential is highly correlated with the first order MCI. Therefore, it has been used to derive the adsorption coefficient.

References

- US EPA (2012). On-Line KOCWIN User’s Guide.

- Meylan, W., P.H. Howard and R.S. Boethling. 1992. Molecular topology/fragment contribution method for predicting soil sorption coefficients. Environ. Sci. Technol. 26: 1560-1567.

Model:  KOCWIN v2.00
SMILES:  c1(NC(=NOS(=O)(=O)c5cc(N=Nc3ccc(Nc4ccccc4)cc3)ccc5)Nc2c(CC)cccc2)c(C)cc(C)cc1
Molecular weight (g/mol): 620
  Molecular weight (g/mol)
  Minimum Maximum Average
Training set 32,04 665,02  224,4
Validation set 73,14 504,12 277,8
Assessment of molecular weight Molecular weight within range of training set
   
Appendix DMCI & Log Kow Correction Factors for 447 Compound Training Set
Correction Factor Descriptor Coefficient for Molecular Connectivity Index (MCI) Regression Methodology Occurence No. of instances of each bondfound for the current substance
  (new model) Remark (number of compounds) (max per structure)
Nitrogen to non-fused aromatic ring -0.522510 (a) 154 2 2
Nitrogen to carbon (aliphatic) (-N-C-) -0.212730 c 133 5 2
Azo (-N=N-) -0.647525   3 1 1
Miscellaneous S(=O) group -1.298009 (a) 11 1 1

Description of key information

Koc was estimated with two separate estimation methodologies:

Method 1 : using log Kow (octanol-water partition coefficient)

Method 2: using first-order Molecular Connectivity Index (MCI)

Results of the two models did not show values within the same range.The MCI method does not take the log kow for calculations and the Koc estimated was very high. As the method 1 takes into account the experimental value of log kow of the substance, the Koc/LogKoc is more accurate.

Key value for chemical safety assessment

Koc at 20 °C:
3 873

Additional information

With a Koc between 2000 -5000, the substance can be considered to adsorb rapidly to soil or sludge, showing little mobility (according to P.J. McCall et al., 1981)1.

 

1McCall P.J., Laskowski D.A., Swann R.L., and Dishburger H.J., (1981), “Measurement of sorption coefficients of organic chemicals and their use, in environmental fate analysis”, in Test Protocols for Environmental Fate and Movement of Toxicants. Proceedings of AOAC Symposium, AOAC, washington DC.

 

McCall classification scheme of mobility

 Koc range  Class of mobility
 0 - 50  Very high
 50 - 150  High
 150 - 500  Medium
 500 - 2000  Low
 2000 - 5000  Slightly
 > 5000  Immobile

[LogKoc: 3.588]