Preferential solvation studies using the solvatochromic dicyanobis(1,10-phenanthroline)iron(II) complex

Article abstract of DOI:10.1039/b104093f

Solvent effects on the electronic spectra of dicyanobis(1,10-phenanthroline)iron(II) have been investigated in thirteen pure solvents and twenty binary solvent mixtures. In pure solvents, the shifts of the v_max values are found to depend on more than one of the known solvent parameters (AN, α, β, DN and π*). AN, α and π* are found to be the most important solvent parameters, exerting a considerable effect on the solvatochromic shifts of the title complex. The preferential solvation of dicyanobis(1,10-phenanthroline)iron(II) in binary solvent mixtures has been investigated by monitoring the metal-ligand charge transfer band of the indicator complex. Organic solvents are preferred near the indicator complex in aqueous binary solvent mixtures (negative deviation), except in regions rich in MeCN, Me₂CO and 1,4-dioxane, where water molecules are preferred over the organic component (dual behavior). However, the indicator complex is preferentially solvated by the component which has the higher acceptor number in non-aqueous binary solvent mixtures. Negative deviation was observed in binary mixtures ofand CHCl₃ Me₂CO with alcohols and positive deviation in mixtures of CHCl₃ with THF, DMSO and Me₂CO. Different criteria were considered to evaluate the extent of preferential solvation in different solvent mixtures, viz., the local molar fraction (X_A^L), the excess function (ΔX), the iso-solvation point (X_B^iso) and the preferential solvation constant (K_A/B). The preferential solvation data have been linearly correlated with the different solvent parameters.

Full text of DOI:10.1039/b104093f

View Article Online / Journal Homepage / Table of Contents for this issue
Preferential solvation studies using the solvatochromic
dicyanobis(1,10-phenanthroline)iron(II) complex¤
A. Taha,*” A. A. T. Ramadan, M. A. El-Behairy, A. I. Ismail and M. M. Mahmoud
Department of Chemistry, Faculty of Education, Ain Shams University, Roxy, Cairo, Egypt
Received (in Montpellier, France) 9th May 2001, Accepted 21st June 2001
First published as an Advance Article on the web 4th September 2001
Solvent e†ects on the electronic spectra of dicyanobis(1,10-phenanthroline)iron(II) have been investigated in thirteen
pure solvents and twenty binary solvent mixtures. In pure solvents, the shifts of the l values are found to depend
max
on more than one of the known solvent parameters (AN, a, b, DN and n*). AN, a and n* are found to be the most
important solvent parameters, exerting a considerable e†ect on the solvatochromic shifts of the title complex. The
preferential solvation of dicyanobis(1,10-phenanthroline)iron(II) in binary solvent mixtures has been investigated by
monitoring the metalÈligand charge transfer band of the indicator complex. Organic solvents are preferred near the
indicator complex in aqueous binary solvent mixtures (negative deviation), except in regions rich in MeCN,
Me CO and 1,4-dioxane, where water molecules are preferred over the organic component (dual behavior).
2
However, the indicator complex is preferentially solvated by the component which has the higher acceptor number
in non-aqueous binary solvent mixtures. Negative deviation was observed in binary mixtures of CHCl and
3
Me CO with alcohols and positive deviation in mixtures of CHCl with THF, DMSO and Me CO. Di†erent
2
3
2
criteria were considered to evaluate the extent of preferential solvation in di†erent solvent mixtures, viz., the local
molar fraction (XL), the excess function (*X), the iso-solvation point (Xiso) and the preferential solvation constant
(K ). The preferential solvation data have been linearly correlated with the di†erent solvent parameters.
A
B
A@B
a color indicator for providing an insight into the microscopic
characteristics of the cybotactic zone of solutes in pure and
mixed solvents.
Introduction
The study of soluteÈsolvent and solventÈsolvent interactions,
and how they a†ect the intimate structure of the solute, has
attracted much attention as they play a major role in all phe-
nomena occurring in the liquid phase. In many cases, it has
been found that solute properties depend upon more than one
solvent parameter.1,2 Solvent mixtures have become an
important subject of research because of their frequent use
and the wide Ðeld of application they o†er.3h10 The most
important feature of these mixed solvents is the gradual varia-
tion of properties they show when their composition is grad-
ually modiÐed. Therefore, there is currently considerable
interest in the study of physicochemical phenomena in mixed
solvent systems and their interpretation in terms of prefer-
ential solvation of solutes by one of the component solvents in
the mixture.1,2,11 In connection with this, we have previously
studied the preferential solvation of solvatochromic mesoionic
2,3-diaryl-2H-tetrazolium-5-thiolate derivatives.12
The goal of this study was to monitor the e†ect of soluteÈ
solvent and solventÈsolvent interactions on the preferential
solvation characteristics. For this purpose, the MLCT bands
of Fe(phen) (CN) in pure and binary solvent mixtures have
2
2
been used as an indicator solute.
Experimental
Chemicals
The reagents used were Merck and Aldrich chemicals. The
water used in all experiments was distilled twice. All organic
solvents were of the highest grade available and were puriÐed
using standard methods.23,24
Studying the strong solvatochromicity of the metal-to-
Synthesis
ligand charge transfer (MLCT) band of FeII(LL)(CN) com-
2
plexes,13h18 where LL \ diimine, 1,10-phenanthroline (phen)
Dicyanobis(1,10-phenanthroline)iron(II) complex was synthe-
sized according to the reported method,13 by heating a
mixture of 6.0 g of phenanthroline monohydrate (0.03 mol)
and 3.90 g of ammonium iron(II) sulfate hexahydrate (0.01
mol) in 400 cm3 of water, followed by addition of 9.80 g of
KCN (0.15 mol) with continuous stirring for 2 h. The obtained
product was Ðltered, then dissolved in 30 cm3 of concentrated
sulfuric acid followed by the addition of 1 dm3 of bidistilled
water. The resulting dark violet crystals were dried in vacuo
for several hours at 35 ¡C. The composition of the prepared
indicator complex, Fe(phen) (CN) É 2H O was characterized
or 2,2@-bipyridine, enables one to measure the extent of prefer-
ential solvation. Although the solvation of Fe(phen) (CN) in
2
2
pure solvents has been extensively studied,19h21 only limited
attention has been paid to its behavior in binary solvent mix-
tures.22 This is surprising since this compound can be used as
¤ Electronic supplementary information (ESI) available: tables listing
various solvent parameters (Table S1), l
values (Table S2) and local
max
molar fractions (Table S3). See http://www.rsc.org/suppdata/nj/b1/
b104093f/
2
2
2
by elemental analysis, found: C 66.60, H 3.40, N 18.0; calc. for
” Present Address: Chemistry Department, UAE University Faculty
of Science, Al-Ain, P.O. Box 17551, United Arab Emirates. E-mail:
A.Taha=uaeu.ac.ae
C
H
N O Fe: C 66.69, H 3.44, N 17.95%. Furthermore, IR
26 20
6 2
and magnetic susceptibility measurements were used to
1306
New J. Chem., 2001, 25, 1306È1312
DOI: 10.1039/b104093f
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2001

View Article Online
solvent parameters, and a, b and c are coefficients of X , X
1
2
and X , which can be obtained by multiple linear regression
3
analysis.
Correlation of l
values with the p* and a solvatochromic
max
parameters of Kamlet et al.,26 yields l \ 15.10 ] 2.32a
max
] 1.44n*, r \ 0.99. The relative percentage inÑuences of a and
n* on the l
values were calculated as described in previous
max
work12 and found to be 61.7 and 38.3%, respectively. This
correlation demonstrates that the capability of the indicator
complex to form hydrogen bonds with proton-donor solvents
(as measured by the a term) plays an important role in deter-
Scheme
1
The structure of the dicyanobis(1,10-phenanthroline)-
iron(II) indicator complex.
mining the shift of the l
values in di†erent solvents. The
max
positive sign of the a coefficient indicates that the hydrogen
bonds formed with the indicator complex in protic solvents
may stabilize the ground state rather than the excited state,
resulting in an hypsochromic shift.
conÐrm the proposed structure of the complex. The IR spec-
trum of a KBr disc of the complex shows the following
stretching frequencies: 3430, 2071 and 1628 cm~1, corre-
Another good correlation was also found when the n* and
sponding to l , l
and l
, respectively.25 Magnetic sus-
OH CK N
CJN
AN parameters were used as independent variables, l
\
ceptibility measurements show that the complex is
diamagnetic, which suggests that the indicator complex is low
spin octahedral (Scheme 1).
max
15.00 ] 3.84AN ] 0.385n*, r \ 0.98. This result demonstrates
the importance of both the solvent Lewis acidity (measured by
AN) and n* parameters to explain the observed variation in
the shift of l
values of the indicator complex, with relative
max
Measurements
contributions of 91 and 9% for AN and n*, respectively. The
positive sign of the AN coefficient suggests that as the solvent
acceptor number increases, the ground state of the indicator
complex is preferentially stabilized. This might be attributed
to the removal of electron density from the cyanide ligand and
increasing p back bonding with the metal ion, leading to hyp-
sochromic shifts in the MLCT bands as a function of AN.
This seems an entirely reasonable assumption since the tran-
sition involved is likely to be metal-to-ligand charge transfer
from the iron coordination center to the p* orbitals of the
phenanthroline ligands.17 Electron withdrawal due to inter-
action with Lewis acids thus leads to changes in the r
The electronic absorption spectra of 2.5 ] 10~5 M solutions
of the indicator complex in thirteen pure solvents and their
binary solvent mixtures were recorded on a Jasco V-550 spec-
trophotometer equipped with a thermostatted cell com-
partment and using a cell of path length 1 cm. The solvent
mixtures included both aqueous and non-aqueous mixtures.
The aqueous binary solvent mixtures were formed by mixing
water with acetonitrile (MeCN), 1,4-dioxane (Diox), acetone
(Me CO), N,N-dimethylformamide (DMF), dimethylsulfoxide
2
(DMSO), ethanol (EtOH), 2-propanol (2-PrOH) and meth-
anol (MeOH). The non-aqueous binary solvent mixtures were
obtained by mixing chloroform with EtOH, 2-PrOH, MeOH,
bonding molecular orbitals (including the e levels of the iron),
g
which in turn leads to a deformation of Fe(phen) (CN)
DMSO, Me CO and tetrahydrofuran (THF), and acetone
2
2
2
complex and the concomitant p orbitals (including the split t
with EtOH, 2-PrOH, MeOH and DMSO solvents and the
2g
levels of the coordination center), thus inÑuencing the energies
of the antibonding p* ligand orbitals.21 Therefore, as the
polarityÈpolarizability (n*) of the solvent increases, the
ground state is more stabilized than the excited state. This
produces an hypsochromic shift of the absorption band
(positive n* coefficient).
temperature was kept at 25 ¡C. The infrared spectrum in KBr
(400È4000 cm~1) was recorded using a Shimadzu FTIR 8101
spectrometer. The magnetic moment was measured using the
Gouy method and
a Johnson Matthey Alfa Products
MSB-MK I magnetic balance.
The quality Ðt obtained with the presented multi-
parametric correlations for the indicator complex is similar to
that reported for mesoionic compounds.12 According to these
results, it can be concluded that both the solvent AN and a
parameters, as well as the n* parameter, are the most impor-
Results and discussion
Pure solvents
tant factors necessary to explain the dependence of the l
max
When, the shift of the lowest energy metal-to-ligand charge
transfer band (MLCT) of the indicator complex, correspond-
shift on the solvent nature. The relative percentage contribu-
tions of these correlations suggest that the spectral changes in
ing to the t ] p* transition, was correlated vs. the acceptor
l
depend essentially on the a and AN parameters rather
2g
max
number (AN) or ReichardÏs
E
solvent parameter, two
than the n* parameter. Hence, the MLCT transition band
values of the indicator complex reÑect the soluteÈsolvent
interaction at the microscopic level, embodying both the spe-
ciÐc and non-speciÐc modes of interaction.
T
separate lines were obtained for dipolar aprotic solvents and
for alcohols. Points for water and carboxylic acids donÏt Ðt
with any one of these lines. This Ðnding is similar to that of
Al-Alousy and Burgess.19 For that reason, it was of interest to
re-examine the present data for the indicator complex solution
in thirteen pure solvents using multi-parametric correlations,
based on the linear solvation energy relationship (LSER) pro-
posed by Kamlet et al.26 Good multi-parametric correlations
were obtained using di†erent combinations of two or three
variables. These variables include: n* (polarityÈpolarizability),
a (hydrogen bond donor ability), b (hydrogen bond acceptor
ability), DN (donor number) and AN (acceptor number).2 The
general relationship can be expressed by eqn. (1):
Binary solvent mixtures
The shifts in the l
values of the indicator complex mea-
sured in pure solvents and their aqueous and non-aqueous
binary solvent mixtures (AB) at various molar fractions of the
max
component A are listed in the ESI (Table S2). The l
values
max
are shown as a function of the bulk molar fractions of the
component A in Fig. 1 and 2. The functional relationship of
l
vs. X is non-linear for all binary solvent mixtures in the
current study. This Ðnding is very similar to the partial vapor
max
A
l
\ l0 ] aX ] bX ] cX ] É É É
(1)
in a solvent for which the
max
max
1
2
3
pressure with X plots for binary solvent mixtures.28 It is well
A
where, l0 is the value of l
established that the non-linearity of the l
vs. X plots arises
max
max
max
A
properties X are zero for all i,27 X , X and X are di†erent
due to preferential solvation of the indicator complex, which is
i
1
2
3
New J. Chem., 2001, 25, 1306È1312
1307

View Article Online
Fig. 2 Dependence of the l
values for the Fe(phen) (CN) indica-
Fig. 1 Dependence of the l
values of the Fe(phen) (CN) indica-
max
2
2
max
2
2
tor complex on the bulk solution molar fraction of solvent (X ) in (a)
tor complex on the bulk solution molar fraction of water (X ) mixed
A
A
CHCl and (b) Me CO binary solvent mixtures.
with (a) EtOH, MeOH, 2-PrOH, DMF and DMSO, and (b) Diox,
3
2
Me CO and MeCN.
2
represent the situation where local and bulk molar fractions
are the same, is a further indication of the speciÐc interactions
of the indicator complex with the components of the solvent
mixture.4,7,30 Two distinct patterns of preferential solvation
for the indicator complex are observed in aqueous binary
solvent mixtures (Fig. 3), negative deviation and dual behav-
ior. Negative deviation is found in water mixed with alcohols
(EtOH, 2-PrOH and MeOH), DMF or DMSO, while dual
behavior (positive and negative deviations) is observed in
common, and because it modiÐes the neighborhood of the
solute.1,2,11,12
Di†erent criteria were used to assess the type and extent of
the preferential solvation of the indicator complex in binary
solvent mixtures, viz., the local molar fraction (XL), excess
A
function (*X), iso-solvation point (Xiso) and preferential solva-
B
tion constant (K ) parameters.
A@B
The local molar fraction of the component solvents can be
water mixed with MeCN, Me CO or Diox. In non-aqueous
2
calculated using the relation29
binary solvent mixtures, negative and positive deviations are
observed (Fig. 4 and 5). Negative deviation is found in
XL \ 1 [ XL \ (l [ l )/(l [ l )
AB
(2)
A
B
B
A
B
Me COÈalcohol and CHCl Èalcohol binary mixtures.
2
3
where l and l are the absorption frequencies in the pure
However, positive deviation is observed in chloroform mixed
A
B
solvents. The calculated values of the local molar fractions
with THF, DMSO or Me CO, and faint dual behavior is
2
(XL) are listed in the ESI (Table S3) and the functional
observed in Me COÈDMSO.
A
2
relationship of XL vs. the bulk molar fractions (X ) are shown
The type and extent of deviation from the straight line,
A
A
in Fig. 3È5. The deviation from ideality, straight lines which
excess function (*X; *X \ XL [ X ), may be taken as a
A
A
1308
New J. Chem., 2001, 25, 1306È1312

View Article Online
Fig. 4 XL vs. X for various binary solvent mixtures of CHCl with
A
A
3
alcohols, THF, Me CO and DMSO, using Fe(phen) (CN) as an indi-
2
2
2
cator solute.
Fig. 3 XL vs. X for various aqueous binary solvent mixtures, using
A
A
ascertained from the Xiso values. A negative deviation appears
Fe(phen) (CN) as an indicator solute. The straight line represents
B
2
2
ideal behavior.
when Xiso \ 0.5 and a positive deviation occurs when Xiso [
B
B
0.5, indicating a preference for component B over component
A for the former situation and the opposite trend for the
measure of preferential solvation. A positive value of *X indi-
cates a preference for component A over component B, while
a negative value signiÐes the reverse. A summation of *X
values at all fractions can be used to quantify the extent of this
preference (vide infra).
latter.12 Table 1 shows values of Xiso \ 0.5, for the indicator
B
complex in water, CHCl or Me CO mixed with alcohols, and
3
2
the mixture Me COÈDMSO, indicating preferential solvation
2
by component B.12,31 However, values of Xiso [ 0.5 are found
B
in chloroform mixed with DMF, THF or Me CO, and meth-
2
The iso-solvation point (Xiso) refers to the solvent composi-
anol with MeCN or THF, signifying preferential solvation by
tion in the bulk at which l
of the indicator complex in the
component A.
max
binary solvent mixture lies midway between the l
values in
A more quantitative estimate of the extent of preferential
solvation of the indicator complex in binary solvent mixtures
can be made by empolying the preferential solvation param-
max
the pure solvent components.12 The calculated Xiso values for
the indicator complex in di†erent binary mixtures are given in
B
Table 1. The type and extent of preferential solvation can be
eter K , using the thermodynamic model of Frankel et al.,1
A@B
Table 1 Preferential solvation parameters of the Fe(phen) (CN) complex at various molar fractions of component A in di†erent binary solvent
2
2
mixtures at 25 ¡C
&*X
`vea
Xiso
K
B
A@B
Binary solvent
mixture (AÈB)
Deviation
type
~ve
`ve
~ve
`ve
~ve
WaterÈMeOH
WaterÈEtOH
WaterÈ2-PrOH
WaterÈDMF
WaterÈDMSO
WaterÈDiox
1.25
1.52
1.30
1.09
1.51
0.32
0.19
0.01
[
[
[
[
[
db
d
0.30
0.25
0.28
0.31
0.27
0.11
0.15
0.06
0.69
0.40
0.25
0.33
0.37
0.46
0.55
0.67
0.53
0.60
0.97
1.04
2.40
0.90
0.89
0.75
0.69
0.87
1.30
1.24
1.46
1.34
3.47
WaterÈMe CO
2
WaterÈMeCN
d
MeOHÈTHF
]
]
[
[
[
]
]
]
[
[
[
[
MeOHÈMeCN
CHCl ÈMeOH
1.53
1.54
0.62
0.22
0.25
0.42
0.14
0.23
0.44
3
CHCl ÈEtOH
3
CHCl È2-PrOH
3
CHCl ÈDMSO
13.39
5.69
4.82
0.86
0.87
0.89
2.76
È
4.10
3
CHCl ÈTHF
3
CHCl ÈMe CO
3
2
Me COÈMeOH
1.57
1.90
1.89
0.99
0.22
0.17
0.18
0.29
0.39
0.33
0.30
0.31
2
Me COÈEtOH
2
Me COÈ2-PrOH
2
Me COÈDMSO
2
a `ve preferentially solvated by solvent A; ~ve preferentially solvated by solvent B. b d dual behavior (], [).
New J. Chem., 2001, 25, 1306È1312
1309

View Article Online
aqueous mixtures of MeCN, Me CO or Diox. (b) Values of
2
K
[ 1 are found in chloroform mixed with DMSO or
A@B
2
Me CO, and mixtures of methanol with THF or MeCN, as
well as at low molar fractions of water in aqueous mixtures of
MeCN, Me CO or Diox. This indicates preferential solvation
2
of the indicator complex by component B in the former case
and A in the latter.
All possible combinations have been checked for the calcu-
lated preferential solvation parameters K , &*X and Xiso
A@B
B
given in Table 1 using multiple regression analysis. The best
Ðts obtained yield the following:
K
\ 1.77 ] 0.73&*X
A@B
[ 2.03 Xiso, r \ 0.999 and K \ 0.66 ] 0.06&*X [ 0.82Xiso,
B
A@B
B
r \ 0.83; for non-aqueous and aqueous binary solvent mix-
tures, respectively. The correlation coefficient values can be
used to judge the agreement between the di†erent criteria
chosen in the present work. The agreement is very good for
chloroform and acetone systems and poor for aqueous
systems, indicating the simplicity of the former systems and
the complexity of the latter.
It was thought that correlations of the preferential solvation
data given in Table 1 vs. well-known solvent parameters using
multiple linear regression analysis might quantify the role
which solventÈsolvent interaction plays in the preferential sol-
vation process. Hence, such treatments were carried out, the
results of which are collected in Tables 2 and 3. According to
the relative percentage contributions of the a, n* and DN, AN
Fig. 5 XL vs. X for various binary solvent mixtures of Me CO with
A
A
2
alcohols and DMSO, using Fe(phen) (CN) as an indicator solute.
2
2
parameters, it can be concluded that the values of K
are
A@B
essentially a†ected by the hydrogen bond donor ability (a) and
acceptor number (AN) parameters of the co-solvent, except for
the acetone system, which is mainly a†ected by the polarityÈ
according to the following equation:
(XL/XL)
A
B
K
\
(3)
polarizability (n*) parameter. However, the &*X and Xiso
A@B
(X /X )
B
A
B
values are found to be mainly a†ected by the n* and donor
where XL and XL represent local molar fractions of com-
strength (DN) parameters of the co-solvent, except for the
chloroform system, which is found to be mainly dependent
upon the a and AN parameters.
A
B
ponents A and B in the solvation shell and X and X refer to
A
B
the same quantity in the bulk solvent. The parameter K
A@B
measures the tendency of the indicator complex to be solvated
with solvent A in comparison to solvent B. According to eqn.
(3), a plot of (XL/XL) vs. (X /X ) will give straight line with
The positive signs of the AN, DN and a coefficients for the
aqueous solvent system suggest that as the values of these
solvent parameters increase, the extent of solventÈsolvent and
soluteÈsolvent interactions increase. This can be ascribed to
solventÈsolvent interaction, which might lead to the formation
A
B
A
B
slope K , which represents the preferential solvation con-
A@B
A@B
stant. K
can be obtained from Fig. 3È5 and the calculated
values are given in Table 1. These K
preferential solvation are the same as those discussed above
but they do allow a better inter-system comparison to be
values concerning
of a new solvent species (S ) via donorÈacceptor or
A@B
AB
hydrogen-bonding interactions. This new solvent species
could have properties which are quite di†erent from those of
pure solvents A and B.30 Scheme 2 suggests possible routes
made. K \ 1 indicates a preference for component B over
A@B
component A, in contrast, K [ 1 signiÐes the opposite
for the formation of S ; a 1 : 1 molar fraction ratio for sol-
A@B
AB
trend.12
vents A and B was assumed for the sake of simplicity. The
The conclusion given above can be extracted, for the indica-
tor complex, from the data in Table 1. These data show that:
co-solvent which has a higher DN or lower AN, e.g. aprotic
solvents, should interact with water according to route (a),
which leads to a decrease in the extent of the water cluster. In
contrast, the solvent which has a higher acceptor number and
lower donor number, e.g. protic solvents such as alcohols,
(a) K \ 1 in water, chloroform or acetone mixed with alco-
A@B
hols, water mixed with DMSO or DMF, and the mixture
Me COÈDMSO, as well as at high molar fractions of water in
2
Table 2 Parametric solvent coefficients on the extent of the preferential solvation (PS) of dicyanobis(1,10-phenanthroline)iron(II) complex in
binary solvent mixtures, obtained from the multi-parametric equation: PS \ PS ] aa ] bn*
0
Relative contribution (%)
PS
System
PS
a
b
r
a
n*
0
*X
aqueous
[1.28
[6.84
[3.22
[0.12b
0.64c
1.15
0.01
1.35
[1.55
19.60
2.23
0.44
6.28
0.92
0.98
0.97
0.99
È
0.99
0.93
0.99
È
46.6
23.1
10.4
62.4
35.8
73.8
10.5
82.7
43.3
61.7
38.9
53.4
76.9
89.6
37.6
64.2
26.2
89.5
17.3
56.7
38.3
61.1
CHCl
[5.88
3
Me CO
2
0.26
Xiso
aqueousa
0.73
3.50
A
CHCl
[0.81
[0.29
3
Me CO
2
0.03
0.28
K
aqueousa
0.63b
1.42
7.00
[0.30
[9.17
[2.95
0.35
A@B
[0.67c
CHCl
6.05
[4.75
0.92
0.98
3
Me CO
2
[0.04
0.22
a Two distinct types. b Protic solvent. c Aprotic solvent.
1310
New J. Chem., 2001, 25, 1306È1312

View Article Online
Table 3 Parametric solvent coefficients on the extent of the preferential solvation (PS) of dicyanobis(1,10-phenanthroline)iron(II) complex in
binary solvent mixtures, obtained from the multi-parametric equation: PS \ PS ] aAN ] bDN
0
Relative contribution (%)
PS
System
PS
a
b
r
AN
DN
0
*X
aqueous
1.41
10.70
[5.54
[0.20
0.59
[1.02
[9.12b
[0.59c
6.44
1.64
2.99
0.92
0.99
0.99
0.93
0.99
0.91
0.98
È
35.4
65.6
25.2
20.9
55.3
34.9
63.2
57.4
72.2
80.9
64.6
34.4
74.8
79.1
44.7
65.1
36.8
42.6
27.8
19.1
CHCl
[12.60
[6.60
3
Me CO
2
1.70
5.04
Xiso
aqueous
0.18
0.68
0.84
1.36
4.73
A
CHCl
[1.04
[0.73
8.11
3
Me CO
2
K
aqueousa
A@B
1.12
[0.83
CHCl
[5.81
2.24
0.99
0.92
3
Me CO
2
0.54
[0.55
0.13
a Two distinct types. b Protic solvent. c Aprotic solvent.
should interact with water molecules according to route (b),
which strengthens the water clusters. This interpretation is
supported by the data in Tables 2 and 3, which show splitting
in the data for the aqueous systems into two groups, aprotic
and protic. This Ðnding could explain the complexity found
for the aqueous system data.
The preferential solvation data and discussion given above
clarify that, for mixed aqueous solvents containing alcohols,
DMF and DMSO, the indicator complex is preferentially sol-
vated by the organic component (B). Whereas, mixed aqueous
mixing.35 Thus, the chance of the water molecules solvating
the indicator complex will be enhanced as the molar fraction
of the organic component of these mixtures increases. While in
the water-rich region, free water molecules become less avail-
able for solvation owing to strong self-association and the for-
mation of S
species through hydrogen bonding, so the
AB
organic component is preferred over water. On the other
hand, in non-aqueous binary solvent mixtures, the indicator
complex usually shows preference for the component which
has the higher Lewis acidity (AN).
solvents containing Diox, Me CO and MeCN show dual
behavior; the indicator complex is preferentially solvated by
Thus, in the mixed aqueous systems, the results are under-
standable in terms of the micro-heterogeneity of the binary
mixture.36 The breaking of the hydrogen-bonded network of
water and formation of hydrogen bonds in aqueous aprotic
solvent mixtures have been reported by other workers.34,37,38
Similar preferential solvation characteristics were also
observed for mesoionic compounds and ReichardtÏs pyridine
betaine in this type of solvent mixture.12,39 A theoretical study
of preferential solvation in a number of two-component
systems has also shown similar behavior.32 Ultimately, the
preferential solvation of the indicator complex in aqueous
mixed solvents is determined by soluteÈsolvent and solventÈ
solvent interactions, while soluteÈsolvent interaction is more
predominant in non-aqueous solvent mixtures.
2
water Ðrst (at low molar fractions of water), then by the
organic component at high molar fractions of water. The pref-
erence for the organic component over water, despite water
having a higher acceptor number, can be understood in terms
of the strong self-association of water through solventÈsolvent
hydrogen bonding,32 in addition to the considerably hydro-
phobic nature of the indicator complex over most of its struc-
ture. Furthermore, the water clusters are strengthened
through the substitutional interaction of alcohols with these
clusters,12,33 which agrees with the proposed mechanism in
Scheme 2 route (b). Consequently, the opportunity for water
molecules to solvate the indicator complex will decrease and
the number of hydrogen bonds will increase.
For the case of dual behavior, as the percentage of the
organic component increases, the self-associated structure of
water gradually breaks down and at a high molar fraction of
the organic component, preferential solvation by water,
through soluteÈsolvent hydrogen bonding, is observed. This
References
1
L. S. Frankel, C. H. Langford and T. R. Stangle, J. Phys. Chem.,
1970, 74, 1376.
2
3
Y. Marcus, Chem. Soc. Rev., 1993, 409.
J. G. Dawber, J. Ward and R. A. Williams, J. Chem. Soc.,
Faraday T rans., 1988, 84, 713.
may be ascribed to the additional mixing of MeCN, Me CO
2
4
5
P. Chatteejee, A. K. Laha and S. Bagchi, J. Chem. Soc., Faraday
T rans., 1992, 88, 1675.
and Diox with water,34 since all these organic molecules func-
tion only as hydrogen bond acceptors, as described by the
proposed mechanism in Scheme 2 route (a). These solvents
cannot form part of the hydrogen-bonding network of water,
as suggested from their lower a values,2 which means that the
organic solvent molecules exist in the space between the water
clusters.12,34 Hence, the water clusters becomes weaker with
these organic solvents, which was observed as endothermic
N. Papafopoulus and A. Avranas, J. Solution Chem., 1991, 20,
293.
6
7
8
H. Kovacs and A. Laaksonen, J. Am. Chem. Soc., 1991, 113, 5596.
J. G. Dawber, J. Chem. Soc., Faraday T rans., 1990, 86, 287.
H. Schneider, Solute-Solvent Interactions, ed. J. Cotes and C. D.
Rich, Marcel Decker, New York, 1976, vol. 2, ch. 11.
A. K. Covington and K. E. Newman, Pure Appl. Chem., 1979, 51,
2041.
9
10 F. Tanaka, Y. Kawasaki and S. Yamashita, J. Chem. Soc.,
Faraday T rans., 1988, 84, 1083.
11 C. Reichardt, Solvents and Solvent E†ects in Organic Chemistry,
VCH, Weinheim, 1988.
12 A. Taha and A. M. Kiwan, New J. Chem., 2001, 25, 502.
13 A. A. Schilt, J. Am. Chem. Soc., 1960, 82, 3000; A. A. Schilt, J.
Am. Chem. Soc., 1960, 82, 5779.
14 J. Burgess, Spectrochim. Acta, Part A, 1970, 1369.
15 V. Gutmann, G. Resh and W. Linert, Coord. Chem. Rev., 1982,
43, 133.
16 B. F. Shraydeh and S. Salih, Spectrosc. L ett., 1994, 27, 333.
17 H. E. Toma and M. S. Takasugi, J. Solution Chem., 1983, 12, 547.
18 R. W. Soukup and R. Shmid, J. Chem. Educ., 1985, 62, 459.
19 A. Al-Aousy and J. Burgess, Inorg. Chim. Acta, 1990, 169, 167.
Scheme 2 SolventÈsolvent interaction to form S solvent species in
AB
aqueous (a) protic and (b) aprotic solvent mixtures.
New J. Chem., 2001, 25, 1306È1312
1311

View Article Online
20 M. Podsiadla, J. Rzeszotarska and M. K. Kalinowski, Monatsh.
Chem., 1994, 125, 827 and references therein.
30 C. Rafols, M. Roses and E. Bosch, J. Chem. Soc., Perkin T rans. 2,
1997, 243; C. Rafols, M. Roses and E. Bosch, J. Chem. Soc.,
Perkin T rans. 2, 1997, 1341.
31 B. F. Shraydeh and N. Zatar, Monatsh. Chem., 1994, 125, 655.
32 Y. Marcus, J. Chem. Soc., Faraday T rans., 1989, 85, 381.
33 L. S. Frankel, T. R. Stangle and C. H. Langford, Chem. Commun.,
1965, 393.
21 W. Linert, R. F. Jaeson, G. Bauer and A. Taha, J. Coord. Chem.,
1997, 42, 211.
22 B. F. Shraydeh, M. Abu-Eid and N. Abu-Ghulwah, Monatsh.
Chem., 1995, 126, 631.
23 G. Porter and V. Hanten, J. Inorg. Nucl. Chem., 1979, 18, 2053.
24 Autorenkollektiv, Organikum, 16th edn., VEB Deutscher Verlag
der Wissenschaften, Berlin, 1986.
25 K. Nakamoto, Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 4th edn., Wiley-Interscience, New
York, Chichester and Brisbane, 1986.
26 J. A. Kamlet, J. L. M. Abboud, M. H. Abraham and R. W. Taft,
J. Org. Chem., 1983, 43, 2877.
27 J. L. M. Abboud and R. Notario, Pure Appl. Chem., 1999, 71, 645.
28 C. H. Langford and J. P. K. Tong, Pure Appl. Chem., 1977, 49, 93.
29 P. Chatterjee and S. Bagchi, J. Chem. Soc., Faraday T rans., 1991,
87, 587.
34 A. Wakisaka, H. A. Carime, Y. Yamamoto and Y. Akyiyama, J.
Chem. Soc., Faraday T rans., 1998, 94, 369.
35 K. W. Morcom and R. W. Smith, J. Chem. T hermodyn., 1969, 1,
503.
36 Y. Migron and Y. Marcus, J. Chem. Soc., Faraday T rans., 1991,
87, 1339.
37 S. K. Garg and C. P. Smth, J. Chem. Phys., 1965, 43, 2959.
38 G. Atkinson and Y. Mori, J. Chem. Phys., 1966, 45, 4716.
39 E. M. Kosower, H. Kanety, H. Dodiuk and J. Hermolin, J. Phys.
Chem., 1982, 86, 1270.
1312
New J. Chem., 2001, 25, 1306È1312