Chemistry Reference
In-Depth Information
Walker et al. (2007) developed 6 QSARs to predict the toxicity of 17 cations to
sunflower seeds (F.1.
Helianthus annuus
Sunspot) from the metal's physical proper-
based on the cation concentration producing a 50% inhibition of radicle growth one
day after emergence. The QSAR developed with density of the elements (ρ), enthalpy
of formation of metal sulfides (ΔH
s
), and the stability constants of metal ions with
occurrence levels, the QSAR developed with metal concentrations in soil (log M
soil
),
the median elemental composition of soils (mg X/kg soil), and the calculated mean
of the elemental content in land plants (Land Plants) produced the highest adjusted
r
2
(Table 5.20).
As discussed in Section 5.2.5, Van Kolck et al.
(2008) developed 4 QSARs to
predict the 96-hour LC
50
values of 5 cations to the mussel
Mytilis edulis
and 4
QSARs to predict the 96-hour LC
50
values of 6 cations to the mussel
Perna viridis
(
Table 5.17
). Six of these QSARs included 3 of the less numerous physical proper-
ties used to predict cation toxicity, viz., covalent index
X
r
m
( )
, absolute value of
the logarithm of the first hydrolysis constant (|log
K
OH
|), and ionic index (Z
2
/r).
The QSARs developed with the covalent index
X
r
m
(
)
produced the highest
r
2
value
2
(Table 5.20).
Kinraide (2009) developed softness and toxicity scales by compiling data from
8 previously published softness scales and 10 previously published toxicity scales
charge (Table 5.4). For this QSAR (T
Con obs
=
a
σ
Con comp
+
b
σ
Con comp
Z + cZ), r
2
= 0.923,
a
= 2.16,
b
= −0.521,
c
= 0.0778, observed toxicity = T
Con obs
, computed softness = σCon
comp, and ionic charge = Z.
Lepădatu et al.
(2009) applied molecular fingerprint descriptors representing the
electronegativities of the highest occupied molecular orbital (HOMO) and lowest
unoccupied molecular orbital (LUMO) quantum molecular states to 12 cations from
Walker et al. (2007) with
s
,
p
, and
d
N
valence shells that inhibited the growth of
sunflower. Coefficients of determination for descriptors representing HOMO and
LUMO for all 12 cations were consistently lower than coefficients of determination
for descriptors representing HOMO and LUMO for just the cations with
s
,
p
valence
shells (Ag
1+
, Al
3+
, Ba
2+
, Cd
2+
, Li
1+
, Pb
2+
, Zn
2+
). A similar, but less consistent, com-
parison also existed for cations with
d
N
valence shells (Co
2+
, Cu
2+
, Fe
3+
, Mn
2+
, Ni
2+
).
Based on those comparisons, descriptors representing HOMO and LUMO were used
to develop separate QSARs for cations with
s
,
p
valence shells and cations with
d
N
valence shells (Table 5.4). For the 7 cations with
s
,
p
valence shells, the highest
r
2
value was obtained for the QSAR that used the U-EO (electronegativity of UMO state
for EM [oxygen]) descriptor (Table 5.20). For the 5 cations with d
N
valence shells,
the highest r
2
values were obtained for the QSARs that used the HELH (electronega-
tivity of HOMO state for ELH [hydrogen]) and U-ELH (electronegativity of UMO
[unoccupied molecular orbital] state for ELH [hydrogen]) descriptors (Table 5.20).
Roy et al.
(2009) used the uncommon physicochemical property, electrophilic-
ity (ω), and a training set of 10 cations and experimental toxicity data for the soil
nematode,
Caenorhabditis elegans
, to develop 2 QSARs to predict the toxicity
of As
3+
and As
5+
ions (Table 5.4). The
r
2
value decreased when atomic number
