Tools for exposure scenarios in REACH evaluations 2'

1
Tools for exposure scenarios in REACH
evaluations 2. Parameter evaluation and
validation for FATEMOD model.
Seija Sinkkonen and Jaakko Paasivirta,
Department of Chemistry, University of
Jyväskylä, Finland
University of Jyväskylä
2
History of parametrization studies for
environmental fate modelling
Application of structural chemistry to
environmentally relevant compounds in Jyväskylä
started in 1970s as cooperation of biology and
chemistry researchers. Environmental fate
modeling was realized to be central for
estimation and management of risk from toxic and
ecotoxic pollutants. Scientific base for
practical models was developed first in 1970s
and 1980s especially in Canada, USA, Germany and
Switzerland. It was adopted to organic chemistry
education and studies in Jyväskylä university at
1980s. Practical modeling was developed first
alone, but then by international cooperation
since 1989, when ESF project Chemical Exposure
Prediction (1989-1995) started in Miland, Italy,
J.Paasivirta as partner. S.Sinkkonen joined to
this project in second workshop in France 1990.
Later, several other environmental chemists from
Jyväskylä contributed. Furthermore, to build
models for Baltic Sea Area projects were active
within Nordic Council in 1991-1993 and as EU
project in 1996-1999. Our role in these programs
was parametrization and testing of exposure
models. We learnt that temperature correction of
several compound properties was essential to
achieve realistic fate predictions for different
climates. Therefore, we worked for fast automatic
procedure, to be included in model code. The
pioneer paper was J.Paasivirta, S.Sinkkonen,
P.Mikkelson, T.Rantio and F.Wania (1999)
Estimation of vapor pressures, solubilities, and
Henrys law constant or selected persistent
organic pollutants as functions of temperature.
Chemosphere 39, 811-832.
3
FATEMOD database parametrization of the values
for properties of the environments and chemicals
Properties of the
environments. Instead using unit world box 1 x 1
x 1 Km as suggested by D.Mackay Multimedia
Environmental Models L-242, Lewis, Chelsea, MI,
USA) suitable for general risk estimation of
chemicals, we adopted natural catchment areas as
model environments to achieve more flexibility
for different cases of risk evaluations.
Properties of the chemical
compounds. Molecular properties Name, Group,
Subgroup, CAS register number, Molar mass (WM),
Melting point (Tm K), Entropy of Fusion (?Sf),
Liquid state molar volume (Vb), pKa (for acids or
bases)
Temperature-dependent properties Log(pr) Apr
Bpr. Vapor pressure in liquid state (Pl Pa),
Solubility in water (S mol m-3), Henrys law
function (H Pa m3 mol-1), Hydrophobity LogKow
(where Kow is the octanol-water partition
coefficient) and. Degradation half-life
times HL(i) (i 1 air, 2 water, 3
soil/plants and 4 sediment reference time HLT
(usually 20 or 25 C)
4
/ Plants
5
FATEMOD window for editing property values of the
environment box
Southwest Finland (SWF) catchment area of the
Finnish Rivers flowing to the Bothnian Sea.
Major compartments for mass balance Air,
surface Water, Soil (including surface plants),
and Sediment. Minor compartments for
concentration data Suspended sediment and Fish
(aquatic biota).
6
FATEMOD editing window for substance parameters
7
Herbicide DNOC evaluation of parameters for
FATEMOD
CAS 534-52-1, WM 198.122, Mp 86.5 C ?Tm 359.65
K Enthalpy of fusion ? Hf 20515 J mol-1 (DSC by
C.Plato (1972) Anal. Chem. 44, 1531-1534). Entropy
of fusion ? Sf ?Hf / Tm 57.04 J K-1 mol-1.
Liquid state molar volume Vb 137.4 cm3 mol-1
from increments of P.Ruelle et al. (1991) Pharm.
Res. 840-850. pKa 4.31
Solubility parameter DB S Fdi / Vb according to
P.Ruelle (2000) Chemosphere 40, 457-512. S Fdi is
the dispersion component of molar attraction
constant calculated from increments of C.W.van
Krevelen (1990) in Properties of Polymers,
Elsevier, Amsterdam, pp. 212-213. Value calcd.
for DNOC 18.20.
Parameters needed for estimation of water
solubility and hydrophobity of the chemicals are
association terms P.Ruelle (2000) Chemosphere
40, 457-512. vAcc and vDon are the numbers of
active sites. KAccW(i) and KDonW(i) are
stability constants for proton acceptor and donor
groups of the compound in the water. Similar
terms for the compound in n-octanol are KAccO(i)
and KDonO(i). The greatest value of these
association terms, MAXW or MAXO are also needed
in evaluation. Additionally, sum of the hydroxyl
groups is NOH, and parameter boh has value of 1,
2 or 2.9 for primary, secondary of tertiary OH
group, respectively.
Example association terms for DNOC are (KAccO
values are zeros) vAcc vDon KAccW(i)
KDonW(i) MAXW KDonO(i) MAXO
2 1 100,100 5000
5000 5000 5000
8
Determination of the compound property as
function of temperature
(SUBCOOLED) LIQUID STATE VAPOR PRESSURE
VPLEST for evaluation the coefficients Apl and
Bpl for Log Pl Apl -
Bpl / T
Method is from Clark F. Grain in Handbook of
Chemical Estimation Methods, W.J.Lyman, W.F.Reehl
and D.H.Rosenblatt (Eds), ACS, Washington, DC
(1990) in Chapter 14. Liquid state vapor
pressures are computed in one Celsius intervals
at environmental range (e.g. -2 to 30C) by
Grains equation 14-25 using one known Vp and
temperature as reference. Then, the coefficients
are determined by linear regression.
The reference Vp can be for either solid or
liquid state (Ps or Pl). They can be converted to
each other by equation
Log Ps Log Pl ?Sf x (1-Tm/T) / (R x Ln10)
0bs. R x Ln10 19.1444
Conversions between temterature coefficients for
Vps are Aps Apl ?Sf / (RxLn10) and Bps
Bpl ?Sf x Tm / (RLn10)
VPLEST result for liquid state Vps of DNOC is
Compound Mp C ?Sf Pl(25) Apl Bpl
Aps Bps DNOC 86.5 57.04
0.243 11.31 3496 14.29 4567
9
Validation of Pl estimates by two independent
methods
Lei YD, Wania F and Shiu WY (1999)
J.Chem.Eng.Data 44, 577-582.
10
Solubility in water S mol m-3
WATSOLU.bas for evaluation the coefficients for
Log S As - Bs / T
WATSOLU is based on mobile order thermodynamics
estimation for log S at 25 C (P.Ruelle et al.
(1997) Int. J. Pharm. 157, 219-232). We have
divided equations to temperature dependent (Bs/T)
and non-dependent (As) parts
As 5.154 ?Sf / (RxLn10) - 0.036xVb-0.217xLnVb
SNOHx(2boh) / Ln10 SvAcc(i)xLog(1KaccW(i)/1
8.1) SvDon(i)xLog(1KDonW(i)/18.1)
Bs ?Sf x Tm / (RxLn10) (DB- 20.5)2 x Vb /
(RxLn10) x Log (1MAXW / 18.1)
Example Output from WATSOLU for DNOC As
4.617, Bs 1071.7
VOLATILITY Henrys law fuction
Simple conversions for Log H Ah Bh /
T At the narrow temperature range of environments
values of Ah and Bh are in fair agreement with
the relation H Pl / S. Therefore, FATEMOD model
automatically calculates them by conversions Ah
Apl As, and Bh Bpl - Bs .

Example conversion result for DNOC Ah 6.693
Bh 2424.3
11
(No Transcript)
12
Validation of S estimate by two independent
methods
pKa 4.31
WATSOLU
HPLC
pH of the eluent 5.60
Tam D, Varhanikova D, Shiu WY and Mackay D
(1994) J.Chem.Eng.Data 39, 82-86.
13
Hydrophobity (lipophility) as Log Kow is also
temperature-dependent!
TDLKOW.bas for octanol/water partition LogKow
Aow Bow / T Is based on thermodynamic
estimation of LogKow at 25 C of P.Ruelle (2000)
Chemosphere 40, 457-512. We have divided Ruelles
equations in two parts to obtain the temperature
coefficients Aow and Bow

Aow ?B ?F ?Acc
?Don ?B (0.5 x Vb x (1/124.2-1/18.1)
0.5 x Ln(18.1/124.2) / Ln10 ?F (vB x
(rw/18.1 ro/124.2) SNOH x (boh rw ro) /
Ln10 ?Acc SvAcc x Log(1 KaccO(i) /
124.2)/(1 KaccW(i) / 18.1) ?Don
SvDon x Log(1 KdonO(i) / 124.2) / (1
KdonW(i) / 18.1) Bow (Vb/(RxLn10)x(DB-20.5)
2/(1MAXW/18.1)(DB-16.38)2/(1MAXO/124.2)
Where 18.1 is the molar volume of pure water,
124.2 the reduced molar volume of water-
saturated n-octanol, rw structuration factor for
water (2.0) and ro structuration factor for
wate-saturated n-octanol. Observe that
association coefficients for water are the same
as those in WATSOLU.bas (see above). The
temperature coefficient Bow is practically
zero for compounds (often POPs) having only one
kind of substituents, but with several polar
and different substituents in structure Bow can
be significant.

Example1 TDLKOW output for DNOC
Aow 3.826 Bow - 0.439
Example 2 Musk xylene parameters from TDLKOW are
Aow 5.022 and Bow 361.6 in fair agreement of
HPLC and literature values /J.Paasivirta,
S.Sinkkonen, A-L.Rantalainen, D.Broman and
Y.Zebühr (2002) Environ Sci Pollut Res 9(5),
345-355/.
Musk xylene
14
WATSOLU output (Log S As Bs / T) Comp. Mp
C DSf Vb DB As Bs
LogS(25) S(25) mol L-1 Lindane 113.9 61.10
199.8 15.92 0.003 1454 - 4.874 1.37E -
05 CN66 206.0 56.29 219.6 21.40 - 0.991
1418 - 5.738 1.83E - 06 Parlar50 133.1
36.23 299.7 16.25 - 4.980 1052 - 8.507
3.11E - 09 TDLKOW output (Log Kow Aow
Bow / T) Comp. Log Kow(25) Vb
DB Aow Bow Lindane 5.505
199.8 15.92 6.231
216.7 CN66 7.829
219.6 21.40 6.890 - 279.8
Parlar50 8.609 299.7
16.25 9.556 282.5
15
HLT 20 OC reference values for DNOC are HL(1)
170 h, HL(2) 500 h,
HL(3) 720 h, HL(4) 1000 h
16
QSPR estimation of the
reference lifetimes. Example for
polychloronaphthalenes (PCNs). Based on maximal
and minimal HLT 25OC values in NCl classes of
PCDF mode of Mackay et al. and QSPR from
environmental data (J.Falandysz 1998). The most
abundant PCN congeners in Baltic Sea are included
here
Code Cl-subst. NCH-CH NßCls F
HL(1) h HL(2) h HL(3) h HL(4)
h CN42 1,3,5,7 0 2
13 522 1740 26100
87000 CN33 1,2,4,6 2
1 20 483 1610
24150 80500 CN28 1,2,3,5
2 2 26 444
1480 22200 74000 CN27
1,2,3,4 3 2 33
405 1350 20250
67500 CN35 1,2,4,8 2 3
33 405 1350 20250
67500 CN38 1,2,5,8 2
3 33 405 1350
20350 67500 CN46 1,4,5,8
2 4 39 366
1220 18300 61000 CN52
1,2,3,5,7 0 1 7
561 1870 28050
93500 CN58 1,2,4,5,7 0
2 13 522 1740
26100 87000 CN61 1,2,4,6,8
0 2 13 522
1740 26100 87000 CN50
1,2,3,4,6 1 1 13
522 1740 26100
87000 CN51 1,2,3,5,6 1
2 20 483 1610
24150 80500 CN57 1,2,4,5,6
1 2 20 483
1610 24150 80500 CN62
1,2,4,7,8 1 2 20
483 1610 24150
80500 CN53 1,2,3,5,8 1
2 20 483 1610
24150 80500 CN59 1,2,4,5,8
1 3 26 444
1480 22200 74000 CN66
1,2,3,4,6,7 0 0 0
600 2000 30000
100000 CN64 1,2,3,4,5,7 0
1 7 561 1870
28050 93500 CN69 1,2,3,5,7,8
0 1 7 561
1870 28050 93500 CN71
1,2,4,5,6,8 0 2 13
522 1740 26100
87000 CN63 1,2,3,4,5,6 1 1
13 522 1740
26100 87000 CN65 1,2,3,4,5,8
1 2 20 483
1610 24150 80500

F (NCH-CH Nß)6.5 HL(i) HL(i) max
(100 - F) / 100
17
The most persistent and bioaccumulative
chlorobornanes (Toxaphene congeners) are B8-1413
(Parlar 26) and B9-1679 (Parlar 50) /1,2,3/
FATEMOD PARAMETERS FOR TOX Parlar 26 evaluated by
us CAS 142534-71-2 /2/, WM 413.794, Vb 263.1
/4/, DB 16.46 /5/, ?Sf 53.47 /6/, Mp 149.6 C /7/,
Pl(25) 0.00281 Pa /8/, Apl 10.149, Bpl 3786.3
/VPLEST/, vAccW 8, KaccW 25, Aow 5.966, Bow 135.1
/TDLKOW/, As - 0.2260, Bs 1408. 4 /WATSOLU/, Ah
10.375, Bh 2377.9 /,Ah Apl As Bh Bpl
Bs/, HLT 25 HL(1) 600, HL(2) 2000, HL(3)
30000, HL(4) 100000.
References /in brackets/ to methods and
literature values used 1. Vetter and Kirchberg
(2001) Envir Sci Tech 35,5, 960-965. 2. Vetter
and Oehme (2000) The Handbook of Environmental
Chemistry Vol 3 Part K (J.Paasivirta ed)
Springer, pp 237-287. 3. Vetter and Scherer
(1998) Chemosphere 37,9/12, 2525-2543. 4. Ruelle
et al (1991) Pharm Res 8, 840-850. 5. Ruelle
(2000) Chemosphere 40, 457-512. 6. Chickos et al
(1991) J Org Chem 56, 927-938. 7. Lahtinen M et
al (2007) Unpublished DSC measurements. 8.
Bidleman et al (2003) J Chem Eng Data 48,
1122-1127.
Validation of the FATEMOD parameters of toxaphene
P26 (1) Our temperature dependence curve for Kow
(Log Kow Aow Bow / T yields for Log Kow (25)
5.513. Slow stirring method of Fisk et al
(1999) Chemosphere 39, 2549-2562 led to Log Kow
(25) 5.52
18
Validation of the FATEMOD parameters of toxaphene
P 26 (2)