Study of the correlations of the MLCT Vis absorption maxima of 4-pentacyanoferrate-40′-arylsubstituted bispyridinium complexes with the Hammett substituent parameters and the solvent polarity parameters E NT and AN

Article abstract of DOI:10.1002/poc.1514

In this work six new 4-aryl substituted bispyridinium salts have been synthesized using 2, 4-dinitrophenyl-bispyridinium chloride as the starting material, which was arenamine exchanged via the Zincke reaction. These aromatic heterocyclic salts were used

Full text of DOI:10.1002/poc.1514

Research Article
Received: 17 September 2008,
Revised: 24 November 2008,
Accepted: 24 November 2008,
Published online in Wiley InterScience: 15 January 2009
(www.interscience.wiley.com) DOI 10.1002/poc.1514
Study of the correlations of the MLCT Vis
absorption maxima of 4-pentacyanoferrate-
40-arylsubstituted bispyridinium complexes
with the Hammett substituent parameters and
the solvent polarity parameters ENT and AN
Raffaello Papadakisa and Athanase Tsolomitisa
*
In this work six new 4-aryl substituted bispyridinium salts have been synthesized using 2,4-dinitrophenyl-
bispyridinium chloride as the starting material, which was arenamine exchanged via the Zincke reaction. These
aromatic heterocyclic salts were used as ligands, acting as Lewis bases to form a series of 4-pentacyanoferrate-4(-aryl
substituted bispyridinium complex salts in high yields. The complex salts were characterized using several spectro-
scopic techniques including 1H as well as 13C NMR spectroscopy, UV–Vis and FTIR spectrophotometry. These materials
present a number of interesting electronic and optical properties. Herein we mainly focused on the electronic
absorption of these materials. The visible metal-to-ligand-charge-transfer (MLCT) band of these materials assigned as
II
*
dp(Fe ) ! p (L) was proved to be affected by substituent changes (of the benzene ring of the complexes) but mainly
by solvent polarity changes. Band shifts of even 4000 cmS1 were induced by small solvent polarity changes (e.g., water
to methanol). The Vis absorption spectra were investigated in four protic solvents (water, 2,2,2-trifluoroethanol,
ethylene glycol, and methanol). The substituent effects were then quantified using the Hammett equation which
correlates the MLCT absorption wavenumbers with the Hammett constant (sx). Furthermore, two of the most
successful solvent polarity parameters (the acceptor number AN and the normalized solvent polarity scale ETN) were
used to quantify the solvent polarity effects on the MLCT absorption wavenumbers. The correlations obtained in all
cases proved to be satisfactory. The dominant interaction responsible for the solvent polarity effects proved to be the
hydrogen-bond formation between the cyano groups of the complex salts and the protic solvents. Copyright ß 2009
John Wiley & Sons, Ltd.
Keywords: pentacyanoferrate complex salts; substituent effects; solvatochromism; Hammett equation; LFERs
For substituent effects on the electronic absorption maxima,
the simple type of the Hammett equation was used [Eqn (1)]. In
INTRODUCTION
~
Complexes of several transition metals, such as ruthenium,
osmium, and iron with aromatic heterocyclic ligands such as
N-monosubstituted bispyridines are interesting materials which
in recent years have been examined for their unique optical and
electronic properties. Coe et al.[1] have reported the synthesis and
Non-Linear-Optical properties (NLO) of such products, Toma
and Takasugi[2] as well as Blandamer et al.[3] have reported that
complexes like these are markedly solvatochromic, Pinheiro et
al.[4] have also studied some ruthenium (II) complexes for use as
highly sensitive sensors for water in aprotic solvents using simple
UV–Vis spectrophotometry. Dyes with an analogous molecular
structure have been also used as molecular switches,[5] as
solvatochromic probes,[6] and in many other new technologies. In
this work six new 4,40-bispyridinium-1-aryl-10-pentacyanoferrate
(II) complex salts 4a-f (Scheme 1) were synthesized, character-
ized, and their UV–Vis spectra were recorded in four protic
solvents, [water, 2,2,2-trifluoroethanol (TFE), ethylene glycol (EG)
and methanol], and the electron donor–acceptor effect of the aryl
substituents as well as the influence of the solvent polarity on the
Vis spectra were examined.
this equation nx is the wavenumber of the absorption maximum
of a compound with substituent X on its benzene ring, sx is the
Hammett constant of the substituent X, ~no is the absorption
maximum wavenumber of the compound for X ¼ H and D~n is the
[7,8]
~
~
difference between nx and no.
~
~
~
~
~
Dn ¼ nx ꢀ no ¼ r:sx or nx ¼ no þ r:sx
(1)
The solvatochromism of these products was also examined. For
the solvent polarity effects on the electronic absorption maxima
of these compounds we have examined the correlation of the
absorption wavenumbers with the normalized Reichardt solvent
polarity scale ETN[9] and Gutmann’s acceptor number (AN),[10] two
*
Correspondence to: A. Tsolomitis, School of Chemical Engineering, National
Technical University of Athens 15780, Greece.
E-mail: tsolom@chemeng.ntua.gr
a
R. Papadakis, A. Tsolomitis
School of Chemical Engineering, National Technical University of Athens
15780, Greece
J. Phys. Org. Chem. 2009, 22 515–521
Copyright ß 2009 John Wiley & Sons, Ltd.

R. PAPADAKIS AND A. TSOLOMITIS
Table 1. Wavenumbers of the MLCT Vis absorption maxima
of 4a-f, measured in four solvents of different polarity
~nMLCT /(103 cmꢀ1
)
H2O
TFE
EG
MeOH
4a (p-OMe)
4b (p-Me)
4c (m-Me)
4d (p-Cl)
4e (p-Br)
4f (p-CN)
17.699
17.668
17.759
17.483
17.452
17.065
17.482
17.483
17.094
17.007
16.920
16.447
14.970
14.970
14.981
14.641
14.620
14.205
13.755
13.699
13.680
13.500
13.495
13.245
Scheme 1. Synthetic route for the preparation of complex salts 4a-f.
(i) EtOH, reflux, 4–48 h (ii) H2O, r.t., Ar atmosphere, 6 h
electron-donor ability of the FeII center and the electron-acceptor
ability of the ligand (L).[1] These bands are affected mainly by
solvent polarity changes, but also by substituent changes as
shown in Table 1. Their position in the UV–Vis spectrum lies in the
region of lmax ¼ 550–750 nm and their molecular extinction
coefficients (of the maximum absorptions) vary from 3.103 to
4.103 (Mꢀ1 cmꢀ1). In this work we focus on the substituent
and solvent polarity effects on the MLCT bands of these complex
salts.
of the most successful quantitative measures of solvent polarity.
In both cases regression analysis resulted in six linear relations
(one for each compound) which were of the type
~n ¼ ~nDo þ aDETN
(2)
(3)
~
~
n ¼ nGo þ aGAN
Bathochromic band shifts of the MLCT bands are induced
when electron-withdrawing substituents are introduced in the
benzene ring. These bathochromic band shifts are probably
associated with the stabilization of the excited state by
delocalization of the charge transferred from the metal to the
aromatic system of the ligand taking place during the excitation
of molecules (Scheme 2). Electron-withdrawing substituents
stabilize more the MLCT excited state than electron-donating
substituents, by attracting effectively the charge transferred from
the metal to the ligand, so the MLCT bands are shifted to longer
wavelengths (bathochromism) in the case of electro-
n-withdrawing substituents. The steady red shift of the MLCT
bands is in agreement with the increasing acceptor strength
order p-OMe < p-Me < m-Me < p-Cl < p-Br < p-CN of substitu-
ents on the benzene ring of the examined compounds.
The Hammett equation is satisfactory even in its simplest type
in quantifying substituent effects in spectral data (obtained from
UV–Vis, NMR, or IR spectroscopy) and that is why it has been
widely used in the past for this purpose for many cases in
organic[7,13–16] and coordination chemistry.[17] For the use of this
equation we consider that the benzene ring which is attached
to the 4,40-bispyridine moiety, plays little role in the electronic
MLCT excitation process or in other words the electron density
changes in the benzene ring are minimal. Under this condition
the substituents attached to the ring may influence the position
of the MLCT band indirectly by their electron-donating or
withdrawing effect and thus the transition energy of the MLCT
band will be in proportion to the Hammett substituent constant
sx.[7]
~
In Eqns (2) and (3) a is the slope and no is the intercept resulting
from the analysis mentioned before (the subscript D corresponds
to the correlation with the ETN solvent polarity scale and the
subscript G to the correlation with the AN parameter). The
intercept in both cases corresponds to the regression value of
the absorption wavenumber of a certain compound in the
solvent where the ETN scale [for Eqn (2)] or the AN parameter [for
Eqn (3)] is zero. The slope a in both equations has a physical
meaning by indicating the sensitivity of the absorption
wavenumbers to the solvent polarity. Afterwards the experimen-
tal absorption wavenumbers were compared to those calculated
from the above two equations.
RESULTS AND DISCUSSION
The 4,40-bispyridinium-1-aryl-10-pentacyanoferrate (II) complex
salts 4a-f were prepared in two steps. 2,4-dinitrophenyl-
4,40-bispyridinium chloride 1 was used as starting material for
the synthesis of the N-aryl substituted 4,40-bispyridinium chloride
salts 3a-f (Scheme 1). 1 was arenamine-exchanged via the Zincke
reaction[11] giving products 3a-f in high yields. Then these
products were reacted with Na3[FeII(CN)5NH3] to afford
the complex salts 4a-f.[12]
These new compounds have interesting optical and electronic
properties. In this work we are interested in the UV–Vis spectra of
these materials.
The 4,40-bispyridinium-1-aryl-10-pentacyanoferrate(II) complex
salts exhibit two absorption bands in their UV–Vis spectra. The
first one lies in the region of lmax ¼ 310–330 nm and presents
molecular extinction coefficients (emax) which vary from 5.103 to
6.103 (Mꢀ1 cmꢀ1) corresponding to p-p intraligand excitations.
*
In all cases these UV bands are sharp and generally their position
in the UV–Vis spectrum does not seem to be considerably
affected neither by substituent changes (of their benzene ring)
nor by solvent polarity changes. The second Vis absorption band
is an intense broad metal-to-ligand-charge-transfer (MLCT) band
II
*
assigned as dp(Fe ) ! p (L), the energy of which depends on the
Scheme 2. Limiting resonance structures of the complex anion 4f
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 515–521

STUDY OF THE CORRELATIONS OF THE MLCT VIS ABSORPTION
Figure 2. Vis absorption spectra of 4a, measured in H2O, TFE, EG, and
MeOH at room temperature
Figure 1. Correlation between Hammett’s substituent constants, sx, and
~
the wavenumbers of 4a-f, nx, measured in four protic solvents at room
temperature
polarity scales.[21] These empirical scales can be used to quantify
solvent effects on the UV–Vis–NIR spectra.[9] In the past years
several successful attempts for the correlation of the MLCT bands
of complexes of several transition metals such as Fe, Ru, W, etc.
which contain organic ligands such as 2,20-bispyridine or N-alkyl
substituted 4,40-bispyridines, with the Dimroth–Reichardt ET(30)
polarity scale and the Kosower’s Z scale[22,23] have been made as
well as with Gutmann’s acceptor and donor number.[2,20]
In order to quantify the impact of the solvent polarity on the
MLCT absorption maxima wavenumbers, we have chosen to use
the normalized, dimensionless ETN solvent polarity scale and
Gutmann’s acceptor number, AN, two of the most satisfactory
solvent polarity scales. The polarity information for the solvents
used in this work is listed in Table 2.
For the quantitative assessment of the substituent effects on
the MLCT band absorption wavenumbers of our products we
have used the simple type of the Hammett equation [Eqn (1)]
using the sp substituent constants for the para substituents
(p-OMe, ꢀ0.268; p-Me, ꢀ0.170; p-Cl, þ0.227; p-Br, þ0.232),[18] the
sm substituent constant for m-Me (ꢀ0.069)[18] and finally the sꢀa
constant for p-CN (þ1.00).[18,19] The sꢀa value for p-CN (þ1.00) was
derived by Zollinger, Wittwer,[19] concerning the acidity constants
of substituted anilinium ions. Four linear relationships were
gained from the Hammett correlation, one for each solvent.
These correlations are shown in Fig. 1. The same slopes were
obtained for both donors (p-OMe, p-Me, m-Me) and acceptors
(p-Cl, p-Br, p-CN). In all four solvents used, the correlation
between the measured wavenumbers of the MLCT bands and the
Hammett substituent constant s, was quite good (in all cases
R > 0.970). There was no reason in using a more sophisticated
type of the Hammett equation.
The ETN solvent polarity scale is the normalized Reichardt
solvent polarity scale which associates with the Dimro-
th–Reichardt solvent polarity parameter [ET(30)] as shown in
Eqn (4).[9]
In all cases the slope r of the correlation equations was
negative which means that an increase in the acceptor strength
of a substituent of a compound induces a decrease of the MLCT
maximum absorption wavenumber of the UV–Vis spectrum of
this compound. The sensitivity of the MLCT absorption
wavenumbers to the substituent changes, proved to be greater
in the case of ethylene glycol as solvent (in that case r had the
highest absolute value).
ET ð30ÞðsolventÞ ꢀ ET ð30Þ
ET ð30ÞðsolventÞ ꢀ 30:7
32:4
ðSiMe4Þ
ðSiMe4Þ
ETN
¼
¼
(4)
ET ð30Þ
ꢀ ET ð30Þ
ðH2OÞ
The regression analysis resulted in six linear relations of the
type
~n¼~nDo þ aDETN
(2)
The influence of the solvent polarity on the absorption spectra
of 4a-f was also investigated. As it is shown in Fig. 2 when the
solvent polarity is reduced (going from water to methanol) a
bathochromic band shift is induced (negative solvatochro-
mism).[9] The complexes of the transition metals (such as Fe,
Ru, Os, etc.) with heterocyclic ligands are known as solvato-
chromic and have been examined intensively in the past.[2,3,20]
4a-f present a very intense solvatochromism. For example, a
relatively small change in the solvent polarity, e.g., from water
In this equation [Eqn (2)] ~n is the wavenumber of a MLCT band
(absorption maximum) of a compound in a certain solvent, aD
~
and nDo are the slope and the intercept respectively (obtained
from the regression analysis) and ETN is the value of the
normalized Reichardt polarity scale of the solvent. The regression
~
value nDo corresponds to the wavenumber of the MLCT
absorption maximum of a substance when the ETNscale is set
to zero [i.e., for tetramethylsilane (SiMe4) as shown in Eqn (4)]. The
~
regression values of aD and nDo as well as the correlation
coefficients for each case are shown in Table 3. The correlation of
(with
ET(30) ¼ 63.1 kcal molꢀ1
)
to
methanol
(with
ET(30) ¼ 55.4 kcal molꢀ1), induces a bathochromic band shift of
Dl ffi 200 nm (Dn ¼ ꢀ3944 cmꢀ1) as shown in Fig. 2.
~
This was observed also for some similar complexes of
ruthenium by Curtis et al.[20] In general, it is known that
substances the dipole moment (m) of which decreases upon
excitation in solution show negative solvatochromism (mg > me,
where the subscript g corresponds to the ground state and the
subscript e to the excited state). Generally, such substances
usually present intramolecular charge-transfer bands as it
happens in the case of the substances studied here.[9]
Table 2. Polarity parameters of the four solvents used
ET (30)[9]
(kcal molꢀ1
[9]
Solvents
m (D)[24]
)
ETN
AN
H2O
TFE
EG
MeOH
1.85
2.46
2.28
1.70
63.1
59.8
56.3
55.4
1.000
0.898
0.790
0.762
54.8[4]
53.3[25]
43.5[26]
41.3[4]
The lack of theoretical expressions for calculating solvent
effects has driven to attempts to introduce empirical solvent
J. Phys. Org. Chem. 2009, 22 515–521
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

R. PAPADAKIS AND A. TSOLOMITIS
Table 3. Results of the regression analysis for the linear
correlation between the wavenumbers of the MLCT absorp-
tion maxima of 4a-f and the ETNsolvent polarity scale
Table 4. Results of the regression analysis for the linear
correlation between the wavenumbers of the MLCT absorp-
tion maxima of 4a-f and the AN solvent polarity parameter
a
a
Rb
Compound
aG
Rb
~
nDo
~
nGo
Compound
aD
4a
4b
4c
4d
4e
4f
1.865
1.918
1.609
1.336
1.440
1.417
16.300
16.162
16.544
16.604
16.443
16.027
0.9445
0.9513
0.9554
0.9496
0.9536
0.9630
4a
4b
4c
4d
4e
4f
2.753
2.934
2.721
2.283
2.434
2.486
0.273
0.268
0.273
0.277
0.274
0.265
0.9909
0.9876
0.9910
0.9931
0.9930
0.9950
a~n/(103 cmꢀ1).
b Correlation coefficient.
a~n/(103 cmꢀ1).
b Correlation coefficient.
the wavenumbers of the MLCT bands with the ETN solvent polarity
scale is successful. In all cases of the six products in all four
solvents used, the correlation coefficient was greater than 0.900
as shown in Table 3. It should be mentioned that aD has a physical
meaning. The slope (aD) indicates the sensitivity of the MLCT
absorption maximum of a product to the polarity of the solvent in
which it is dissolved. Increased values of aD correspond to a more
intense solvatochromism or in other words when aD increases,
the MLCT band wavenumbers become more sensitive to the
changes of solvent polarity.
The positive sign of aD corresponds to the bathochromism
(~nMLCT reduction) induced when ETN is being reduced, so the sign
(negative or positive) of the aD parameter reflects the type of
solvatochromism. The positive sign of the slop in all cases studied
here is in agreement with the observed negative solvatochro-
mism. In order to check the success of the correlation, we
compared the experimentally obtained MLCT maxima absorption
wavenumbers to the calculated ones from the Eqn (2) for all six
products in different solvents. The results are shown in the plot of
Fig. 3.
parameter introduced by Gutmann et al. quantifies the ability of a
solvent to act as a Lewis acid by accepting electron pairs and thus
quantifies the electrophilic properties of a solvent. That is why it
can be used in deciding which would be the most appropriate
solvent for a given reaction taking into account the electrophilic
properties of solvents.[10] In this work, we are interested in the
impact of the electrophilic properties of the solvents used in our
experiments, on the MLCT absorption maxima wavenumbers of
the compounds studied. The regression analysis between
~
nMLCT(max) and the AN parameter resulted in six linear
relationships (one for each product) of the type
~n¼~nGo þ aGAN
(3)
where ~n is the wavenumber of a MLCT band (absorption
~
maximum) of a product in a certain solvent, aG and nGo are the
slope and the intercept, respectively, and AN is the acceptor
number of the solvent. The results of the regression analysis are
shown in Table 4.
In addition to the previously described correlation, we
attempted to investigate the correlation of the experimentally
obtained MLCT absorption maxima wavenumbers, with Gut-
mann’s acceptor number (AN) which is another quantitative
empirical parameter of solvent polarity. This solvent polarity
In all cases the correlation coefficient (R) was close to 0.9900 or
more. In all cases we obtained positive slopes (aG) which means
that an increase of the acceptor number induces an increase of
the MLCT absorption maximum wavenumber of a product. It
means that solvents which do not behave as good Lewis acids by
acting as worse electron-pair acceptors stabilize more the MLCT
excited state of the products causing a bathochromic shift of the
MLCT band. On the other hand, solvents which act as better Lewis
acids (such as water) can donate more easily their proton in order
to form a hydrogen bond with the nitrogen atom of a cyano
group bound to the iron (II) atom of a complex molecule
(Scheme 3) and destabilize strongly the MLCT excited state.
It is known that in the presence of strong Lewis acids the
nitrogen atom of a cyano group of complexes of the general type
LFeII(CN)5 where L is a heterocyclic ligand, can be even
protonated causing then the destabilization of the MLCT excited
state since the electron transfer in that case becomes more
difficult.[27] The slope here (aG) reflects the impact of the Lewis
acidity of protic solvents on the MLCT bands energies or, in other
words, the sensitivity of the MLCT absorption maxima wave-
numbers of the substances to the Lewis acidity. The successful
correlation for all six cases of products indicate that probably the
dominating interaction, responsible for the negative solvato-
chromism of all the products, is the H-bond donating ability of
Figure 3. Correlation of the experimentally obtained wavenumbers of
the MLCTabsorption maxima versus the calculated ones (from Eqn (2)) for
4a-f in the four solvents used
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 515–521

STUDY OF THE CORRELATIONS OF THE MLCT VIS ABSORPTION
number proved to be a perfect solvent polarity parameter at least
in the case of the protic solvents used in this work.
EXPERIMENTAL SECTION
Scheme 3. Hydrogen bond between the hydrogen atom of a solvent (S
stands for Solvent) and a nitrogen atom of a cyano group of 4a-f
Materials and procedures
Sodium nitroprusside Na3[FeII(CN)5NO] and chloro-2,4-
dinitrobenzene were purchased from ACROS Organics,
4,40-bispyridine was obtained from Fluka as well as the
substituted anilines (p-methoxy, p- and m-methyl, p-chloro,
p-bromo, p-cyano). The solvents used were degassed for several
minutes with argon before use. Water was purified with a
Barnstead EASYpure RF compact ultrapure water system and
then distilled. In all cases of complex formation reactions, the
reaction mixture was kept under argon atmosphere in the dark.
The final products were dried under vacuum over anhydrous
CaCl2 for several days until they were brought to constant weight.
Physical measurements
NMR spectra were obtained using a Varian Gemini 300
spectrometer (300 MHz H, 75 MHz 13C). Both H and 13C NMR
spectra were recorded in deuterium oxide (D2O) at 25 8C. In the
case of 13C NMR measurements, [D6]DMSO was used as an
internal reference (39.39 ppm)[28]. The 1H NMR spectra were
calibrated by using residual undeuterated solvent (H2O) as an
internal reference (4.79 ppm)[28]. IR spectra were recorded on a
Perkin Elmer Spectrum 1 FTIR spectrophotometer in the solid
state (without any preparation of the samples) using the
attenuated total reflectance technique (ATR) in the region
600–4000 cmꢀ1. UV–Visible spectra were recorded using a Varian
CARY 1E UV–Visible spectrophotometer at 25 8C. Melting points
were determined in open capillary tubes using a Gallenkamp
HFB-595 melting point apparatus and are uncorrected. Elemental
and thermogravimetric analyses (TGA) were performed by the
microanalytical laboratory of CNRS (France).
1
1
Figure 4. Plot of the experimentally obtained wavenumbers of the MLCT
absorption maxima versus the calculated ones (from Eqn (3)) for 4a-f in all
four solvents used
the solvents used, with the cyano groups of the products, since
Gutmann’s acceptor number indicates the ability of solvents to
act as Lewis acids either by giving a proton in order to protonate a
group or by forming hydrogen bonds by proton donation to an
electronegative atom such as nitrogen.
Finally, the experimentally observed values of the wavenum-
bers of the MLCT absorption maxima of all products in all four
solvents were compared to the theoretical ones calculated from
the family of Eqns (3), and the plot of Fig. 4 was obtained.
Synthesis of Na3[FeII(CN)5NH3].3H2O
CONCLUSION
This compound was prepared according to the Kenney, Flynn and
Gallini method.[29] Sodium nitroprusside (15.00 g, 0.0526 mol)
was finely powdered and then added to concentrated NH4OH
(60 ml) containing anhydrous sodium acetate (5.00 g). The
reaction mixture was kept at 4 8C in the refrigerator overnight,
and the formed bright yellow solid was filtered off by suction. The
solid product was reprecipitated twice from an aqueous solution
by adding cold ethanol. Finally, the product was dried under
vacuum over anhydrous CaCl2 for about a week until its weight
was stabilized. The yield of the reaction was about 50% (8.60 g,
0.0263 mol).
The impact of the solvent polarity on the position of the MLCT
absorption maxima in the visible region, of 4a-f, proved to be
higher than the impact of the substituent in the benzene ring of
these compounds. Solvent polarity changes induced band shifts
of even D~n ꢁ 4000 cmꢀ1, however substituent changes induced
shifts up to D~n ꢁ 1000 cmꢀ1. The successful Hammett correlation
of the UV–Vis MLCT band maxima wavenumbers in all solvents
indicates that the Hammett equation is also applicable in the case
of the organometallics studied in this work and capable to
quantify substituent effects on UV–Vis spectra of such com-
pounds. The satisfactory results even when the simplest type of
the Hammett equation is used confirm the high applicability of
this linear-free-energy relationship (LFER). Concerning the solvent
polarity effects on the UV–Vis spectra of 4a-f, it was shown that
the MLCT bands are very sensitive in solvent polarity changes
even if that changes are small. The correlation with the acceptor
number (AN) was more successful (R ¼ 0.9912) than that
attempted one with the ETN solvent polarity scale, which was
also satisfactory (R ¼ 0.9538). This shows that probably the
dominating interaction between the solvent molecules and the
dye molecules is in all cases the H-bond formation. The acceptor
Synthesis of 2,4-dinitrophenyl-4,4(-bispyridinium chloride
[2,4-DNPhQR]Cl
According to Coe et al.[12] to
a solution of chloro-2,4-
dinitrobenzene (4.30 g, 21 mmol in 35 ml EtOH) 4,40-bispyridine
was added (3.35 g, 21 mmol) and the solution was refluxed for
24 h. The dark brown solution after cooling at room temperature
was added to diethyl ether (350 ml), with stirring. A golden-brown
precipitate was obtained which was filtered by suction and
washed several times with diethyl ether. The solid was stored
J. Phys. Org. Chem. 2009, 22 515–521
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc

R. PAPADAKIS AND A. TSOLOMITIS
under vacuum over anhydrous CaCl2 (higly hygroscopic solid
which turns to a sticky brown solid when exposed to air) to give
6.25 g (17.42 mmol) 82%: mp 158–159 8C.
J ¼ 4.5 Hz, 2 H; C5H4N), 8.61 (d, J ¼ 5.7 Hz, 2 H; C5H4N), 8.00 (d,
J ¼ 4.5 Hz, 2 H; C5H4N), 7.82 (m, 4 H; Ph); 13C NMR (75 MHz, D2O/
[D6]DMSO): d ¼ 156.03, 151.59, 146.10, 143.47, 142.13, 138.86,
132.10, 127.45, 127.02, 124.03; elemental analysis calcd (%) for
C16H12N2Cl2: C 63.37, H 3.96, N 9.24, found: C 63.63, H 3.79, N 9.43.
General procedure for the synthesis of N-substituted
4,4(-bispyridinium chloride (3a-f)
To a solution of [2,4-DNPhQR]Cl (1) (538 mg, 1.5 mmol in 8 ml
EtOH) an excess of the substituted aniline (2a-f) was added
(6.0 mmol). The solution turned immediately to deep red–violet
and it was refluxed under an argon atmosphere for a period
varying between 4 and 48 h depending on the substituent of
compounds 2a-f. After cooling the dark solution, water was
added (20 ml) and a yellow precipitate (2,4-dinitroaniline) was
formed immediately. The mixture was filtered and the filtrate was
concentrated to dryness. The solid was dissolved in EtOH and
reprecipitated by adding diethyl ether. Finally, the solid was
filtered off by suction, washed with ethyl acetate and diethyl
ether, and dried under vacuum for several hours. The products
were recrystalized from EtOH or EtOH/ iPrOH. Compounds 3a, 3d,
and 3e are known[30] as well as compound 3f[31] but their melting
points are not mentioned.
N-(4-bromophenyl)-4,4(-bispyridinium chloride
[p-BrPhQR]Cl (3e)
24 h reflux. Yellow powder 415 mg, 80%, mp: 284–286 8C; 1H NMR
(300 MHz, D2O): d ¼ 9.39 (d, J ¼ 6.3 Hz, 2 H; C5H4N), 9.00 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 8.77 (d, J ¼ 5.4 Hz, 2 H; C5H4N), 8.20 (d,
J ¼ 5.1 Hz, 2 H; C5H4N), 8.13 (d, J ¼ 8.1 Hz, 2 H; Ph), 7.94
(d, J ¼ 8.7 Hz, 2 H; Ph); 13C NMR (75 MHz, D2O/[D6]DMSO):
d ¼ 155.61, 151.55, 145.90, 143.20, 142.51, 138.86, 134.92, 127.21,
126.84, 123.90; elemental analysis calcd (%) for C16H12N2BrCl: C
55.25, H 3.45, N 8.06, found: C 55.02, H 3.21, N 7.89.
N-(4-cyanophenyl)-4,4(-bispyridinium chloride
[p-CNPhQR]Cl (3f)
48 h reflux. Pale yellow powder 330 mg, 75%, mp: 185–186 8C; 1H
NMR (300 MHz, D2O): d ¼ 9.34 (d, J ¼ 6.3 Hz, 2 H; C5H4N), 8.84 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 8.68 (d, J ¼ 5.7 Hz, 2 H; C5H4N), 8.21 (d,
J ¼ 8.1 Hz, 2 H; Ph), 8.07 (m, 4 H; arom); 13C NMR (75 MHz, D2O/
[D6]DMSO): d ¼ 156.76, 151.61, 146.54, 146.16, 143.43, 136.25,
127.55, 126.74, 124.09, 119.24, 116.17; elemental analysis calcd
(%) for C17H12N3Cl: C 69.51, H 4.09, N 14.31, found: C 69.66, H 3.94,
N 14.51.
N-(4-methoxyphenyl)-4,4(-bispyridinium chloride,
[p-MeOPhQR]Cl (3a)
4 h reflux. Pale yellow powder 380 mg, 85%, mp: 192–194 8C; 1H
NMR (300 MHz, D2O): d ¼ 9.18 (d, J ¼ 6.9 Hz, 2 H; C5H4N), 8.81 (d,
J ¼ 5.6 Hz, 2 H; C5H4N), 8.56 (d, J ¼ 6.9 Hz, 2 H; C5H4N), 8.00 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 7.60 (d, J ¼ 9.0 Hz, 2 H; Ph), 7.29 (d,
J ¼ 6.9 Hz, 2 H; Ph), 3.94 (s, 3 H; Me); 13C NMR (75 MHz, D2O/
[D6]DMSO): d ¼ 162.81, 155.29, 151.59, 145.93, 143.70, 136.95,
127.34,126.86, 124.00, 117.14, 57.38 (Me); elemental analysis
calcd (%) for C17H15N2ClO: C 68.34, H 5.03, N 9.92, found: C 68.52,
H 4.80, N 9.73.
General Procedure for the synthesis of compounds 4a-f:
Na2[FeII(CN)5(R-PhQR)]
To an aqueous solution of the N-substituted-4,40-bispyridinium
chloride 3a-f (0.900 mmol in 10 ml of water), 1.000 mmol
(325 mg) of Na3[FeII(CN)5NH3].3H2O was added. The color of
the solution directly turned to deep purple–blue. The solution
was stirred in the dark under an argon atmosphere at room
temperature for 6 h. A sixfold volume of ethanol was then added
and the reaction mixture was kept at 4 8C overnight. The
precipitate was filtered off by suction, washed with ethanol, and
then dried for several hours in vacuo before storage. The yield in
all cases was above 60%. All products have melting points above
300 8C.
N-(4-methylphenyl)-4,4(-bispyridinium chloride,
[p-MePhQR]Cl (3b)
4 h reflux. Pale yellow powder 372 mg, 88%, mp: 164–166 8C; 1H
NMR (300 MHz, D2O): d ¼ 9.24 (d, J ¼ 6.0 Hz, 2 H; C5H4N), 8,83 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 8.61 (d, J ¼ 6.3 Hz, 2 H; C5H4N), 8.04 (d,
J ¼ 5.1 Hz, 2 H; C5H4N), 7.72 (d, J ¼ 8.1 Hz, 2 H; Ph), 7.60 (d,
J ¼ 8.1 Hz, 2 H; Ph), 2.50 (s, 3 H; Me); 13C NMR (75 MHz, D2O/
[D6]DMSO): d ¼ 155.62, 151.58, 146.02, 144.31, 143.67,
141.35,132.54, 127.38, 125.04, 124.02, 21.77 (Me); elemental
analysis calcd (%) for C17H15N2Cl: C 72.21, H 5.31, N 9.92, found: C
72.48, H 5.02, N 10.10.
Na2[FeII(CN)5(p-MeOPhQR)] (4a)
1
Deep blue solid, 352 mg (0.567 mmol) 63%; H NMR (300 MHz,
N-(3-methylphenyl)-4,4(-bispyridinium chloride,
[m-MePhQR]Cl (3c)
D2O): d ¼ 9.21 (m, 4 H; C5H4N), 8.53 (d, J ¼ 4.5 Hz, 2 H; C5H4N), 7.73
(m, 4 H; Ph), 7.30 (d, J ¼ 6 Hz, 2 H;) 3.96 (s, 3 H; Me); 13C NMR
(75 MHz, D2O/[D6]DMSO): d ¼ 170.71 (CN), 162.52, 154.66, 151.47,
4 h reflux. Pale yellow powder 380 mg, 90%, mp: 246–247 8C; 1H
NMR (300 MHz, D2O): d ¼ 9.18 (d, J ¼ 6.3 Hz, 2 H; C5H4N), 8.74 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 8.55 (d, J ¼ 6.0 Hz, 2 H; C5H4N), 7.95 (d,
J ¼ 4.8 Hz, 2 H; C5H4N), 7.58 (m, 4 H; Ph), 2.47 (s, 3 H; Me); 13C NMR
(75 MHz, D2O/[D6]DMSO): d ¼ 155.36, 151.54, 145.88, 143.48,
143.32, 142.85, 133.78, 131.81, 127.21, 125.61, 123.92, 122.24,
21.91 (Me); elemental analysis calcd (%) for C17H15N2Cl: C 72.21, H
5.31, N 9.92, found: C 72.47, H 5.14, N 10.11.
146.27, 143.34, 136.67, 127.56, 126.90, 124.03, 117.02, 57.31 (Me);
—
IR: n(C N) ¼ 2045 cmꢀ1; UV–Vis (H2O, lnm (loge)): 332 (3,70)
—
—
~
*
*
(p-p ), 565 (3,45) (d-p ); elemental analysis calcd (%) for
C22H15N7FeNa2O. 7H2O: C 42.53, H 4.70, N 15.78, found: C
42.34, H 4.49, N 16.02; TGA: loss of ca. 6.5 molecules of H2O upon
heating.
N-(4-chlorophenyl)-4,4(-bispyridinium chloride,
Na2[FeII(CN)5(p-MePhQR)] (4b)
[p-ClPhQR]Cl (3d)
1
Blue solid, 355 mg (0.586 mmol) 65%; H NMR (300 MHz, D2O):
24 h reflux. Yellow powder 355 mg, 78%, mp: 293–294 8C; 1H NMR
(300 MHz, D2O): d ¼ 9.24 (d, J ¼ 5.7 Hz, 2 H; C5H4N), 8.79 (d,
d ¼ 9.18 (d, J ¼ 6 Hz, 2 H; C5H4N), 8.57 (d, J ¼ 6 Hz, 2 H; C5H4N), 8.50
(d, J ¼ 4.5 Hz, 2 H; C5H4N), 7.89 (d, J ¼ Hz, 2 H; C5H4N), 7.70 (d,
www.interscience.wiley.com/journal/poc
Copyright ß 2009 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2009, 22 515–521

STUDY OF THE CORRELATIONS OF THE MLCT VIS ABSORPTION
J ¼ 9 Hz, 2 H; Ph), 7.60 (d, J ¼ 6.6 Hz, 2 H; Ph), 2.76 (s, 3 H; Me); 13
NMR (75 MHz, D2O/[D6]DMSO): d ¼ 170.14 (CN), 154.97, 151.51,
146.96, 143.83, 143.39, 141.19, 132.43, 127.30, 125.06, 123.99,
C
115.86; IR ~n(C N): 2043 cmꢀ1; UV–Vis (H2O) lnm (loge)): 323
—
—
—
*
*
(3,83) (p-p ), 586 (3,63) (d-p ); elemental analysis calcd (%) for
C22H12N8FeNa2.9H2O: C 40.51, H 4.64, N 17.18, found: C 40.33, H
4.84, N 17.01; TGA: loss of ca. 8.4 molecules of H2O upon heating.
21.82 (Me). IR ~n(C N): 2043 cmꢀ1; UV–Vis (H2O, lnm (loge)): 327
—
—
—
*
*
(3,90) (p-p ), 566 (3,77) (d-p ); elemental analysis calcd (%) for
C22H15N7FeNa2.7H2O: C 43.65, H 4.83, N 16.20, found: C 43.46, H
5.05, N 15.84; TGA: loss of ca. 6.8 molecules of H2O upon heating.
REFERENCES
Na2[FeII(CN)5(m-MePhQR)] (4c)
1
[1] B. J. Coe, J. A. Harris, L. J. Harrington, J. C. Jeffery, L. H. Rees, S.
Houbrechts, A. Persoons, Inorg. Chem. 1998, 37, 3391–3399.
[2] H. E. Toma, M. S. Takasugi, J. Solution Chem. 1983, 12, 547.
[3] M. J. Blandamer, J. Burgess, R. I. Haines, J. Chem. Soc., Dalton Trans.
1976, 5, 1293–1298.
[4] C. Pinheiro, J. Carlos Lima, A. Jorge Parola, Sens. Actuators B, 2006,
114, 978–983.
[5] J. P. Launay, C. Coudret, Ann. N.Y. Acad. Sci., 1998, 852, 116–132.
[6] W. Linert, V. Gutmann, Coord. Chem. Rev. 1992, 117, 159–183.
[7] J. Griffiths, Color and Constitution of Organic Molecules, Academic
Press, London, 1976. pp. 102–104.
Blue solid, 405 mg (0.651 mmol) 72%; H NMR (300 MHz, D2O):
d ¼ 9.27 (d, J ¼ 6.3 Hz, 2 H; C5H4N), 9.16 (d, J ¼ 6.9 Hz, 2 H; C5H4N),
8.56 (d, J ¼ 6.6 Hz, 2 H; C5H4N), 7.71 (d, J ¼ 6.6 Hz, 2 H; C5H4N), 7.63
(m, 4 H; Ph), 2.51 (s, 3 H; Me); 13C NMR (75 MHz, D2O/[D6]DMSO):
d ¼ 171.34 (CN), 167.93, 155.71, 151.54, 146.32, 143.74, 142.90,
133.70, 131.77, 127.62, 125.90, 124.13, 122.43, 21.913 (Me); IR
n(C N): 2035 cmꢀ1; UV–Vis (H2O, lnm (loge)): 315 (3,70) (p-p ),
563 (3,60) (d-p ); elemental analysis calcd (%) for C22H15N7Fe-
—
*
—
—
~
*
Na2.8H2O: C 42.46, H 5.02, N 15.75, found: C 42.27, H 4.81, N 15.66;
TGA: loss of ca. 7.8 molecules of H2O upon heating.
`
[8] H. H. Jaffe, O Milton, Theory and Applications of Ultraviolet Spec-
troscopy, (4th edn), John Wiley and Sons Inc., 1966. pp. 258–259.
[9] C. Reichardt, Chem. Rev. 1994, 94, 2319–2358.
[10] U. Mayer, V. Gutmann, W. Gerger, Monatsh. Chem. 1975, 106,
1235–1257.
Na2[FeII(CN)5(p-ClPhQR)] (4d)
Blue powder, 365 mg (0.552 mmol) 61%; 1H NMR (300 MHz, D2O):
d ¼ 9.23 (m, 4 H; C5H4N), 8.55 (d, J ¼ 4.2 Hz, 2 H; C5H4N), 7.62 (m, 4
H; arom.), 7.64 (d, J ¼ 4.5 Hz, 2 H; arom.); 13C NMR (75 MHz, D2O/
¨
[11] D. Bongard, M. Moller, S. N. Rao, D. Corr, L. Walder, Helv. Chim. Acta
2005, 88, 3200–3209.
[12] B. J. Coe, J. L. Harries, M. Helliwell, L. A. Jones, I. Asselberghs, K. Clays,
B. S. Brunschwig, J. A. Harris, J. Garin, J. Orduna, J. Am. Chem. Soc.
2006, 128, 12192–12204.
[D6]DMSO): d ¼ 171.17 (CN), 167.69, 151.50, 146.51, 143.46,
—
—
~
138.67, 132.03, 127.76, 127.62, 127.15, 124,12; IR n(C N):
—
[13] C. Reichardt, Angew. Chem. 1979, 91, 119–131.
[14] C. Reichardt, Angew. Chem. Int. Ed. Engl. 1979, 18, 98–110.
2038 cmꢀ1; UV–Vis (H2O, lnm (loge)): 321 (3,74) (p-p ), 572
*
ˇ
[15] G. S. Uscumlic, DZ. Mijin, N. V. Valentic, V. Vajs, B. M. Susic, Chem. Phys.
ˇ
Lett. 2004, 397, 148–153.
´
´
ˇ ´
*
(3,65) (d-p ); elemental analysis calcd (%) for C21H12N7Fe-
Na2Cl .9H2O: C 38.11, H 4.57, N 14.82, found: C 38.38, H 4.37, N
14.59; TGA: loss of ca. 8.2 molecules of H2O upon heating.
.
´
ˇ ´
´
[16] S. Jovanovic, D. Mijin, M. Misic-Vukovic, ARKIVOC 2006, X, 116–128.
[17] R. Tong Coord. Chem. Rev. 1998, 175, 43–58.
[18] L. P. Hammett, J. Am. Chem. Soc. 1937, 59, 96–103.
[19] H. Zollinger, C. Wittwer, Helv. Chim. Acta 1956, 39, 347–356.
[20] J. C. Curtis, B. P. Sullivan, T. J. Meyer, Inorg. Chem. 1983, 22, 224–236.
[21] C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, (3rd
edn), Wiley-VCH, Weinheim, 2003.
[22] J. Burgess, Spectrochim. Acta, Part A 1970, A26, 1369.
[23] D. R. Lide, CRC Handbook of Chemistry and Physics, (85th edn), RC
Press, 2004. Chapter 15, pp. 17–1725.
[24] D. R. Lide, CRC Handbook of Chemistry and Physics, (85th edn), RC
Press, 2004. Chapter 15, pp. 17–25.
[25] R. Streck, A. J. Barnes, Spectrochim. Acta, Part A 1999, 55, 1050.
[26] The value of the acceptor number of ethylene glycol (ANEG 43,5) was
estimated on the basis of linear correlation[32] between AN and the
parameter ET(30).
Na2[FeII(CN)5(p-BrPhQR)] (4e)
Blue powder, 425 mg (0.617 mmol) 68.5%; 1H NMR (300 MHz,
D2O): d ¼ 9.14 (m, 4 H; C5H4N), 8.31 (s, 2 H; C5H4N), 7.62 (s, 2 H;
C5H4N), 7.55 (m, 4 H; Ph); 13C NMR (75 MHz, D2O/[D6]DMSO):
d ¼ 170.90 (CN), 167.21, 151.50, 146.38, 143.39, 142.53, 135.04,
127.71, 127.32, 126.88, 124.12; IR ~n(C N): 2040 cmꢀ1; UV–Vis
—
—
—
*
*
(H2O, lnm (loge)): 321 (3,76) (p-p ), 573 (3,67) (d-p ); elemental
analysis calcd (%) for C21H12N7FeNa2Br.8H2O: C 36.65, H 4.10, N
14.25, found: C 36.90, H 3.87, N 14.44; TGA: loss of ca. 7.4
molecules of H2O upon heating.
[27] H. E. Toma, J. M. Malin, Inorg. Chem. 1973, 12, 1039–1045.
[28] H. E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 1997, 62,
7512–7515.
[29] D. J. Kenney, T. P. Flynn, J. B. Gallini, J. Inorg. Nucl. Chem. 1961, 20,
75–81.
[30] I. Yamaguchi, H. Higashi, S. Shigasue, S. Shingai, M. Sato, Tetrahedron
Lett. 2007, 48, 7778–7781.
[31] J-H. Ryu, Y-H. Lee, K-D. Suh, J. Appl. Polym. Sci. 2008, 107, 102–108.
[32] W. R. Fawcett, In Theoretical and Computational Chemistry: Solute/
Solvent Interactions, (2nd edn) Vol. 2, Elsevier, Amsterdam, 1994. 183.
Na2[FeII(CN)5(p-CNPhQR)] (4f)
Deep blue powder, 350 mg (0.537 mmol) 60%; 1H NMR (300 MHz,
D2O): d ¼ 9.30 (d, J ¼ 6.9 Hz, 2 H; C5H4N), 9.18 (d, J ¼ 5.1 Hz, 2 H;
C5H4N), 8.55 (s, 2 H; C5H4N), 8.16 (d, J ¼ 8.1 Hz, 2 H; Ph), 8.01 (d,
J ¼ 8.7 Hz, 2 H; arom), 7.70 (2H, J ¼ 6.3 Hz 2 H; arom); 13C NMR
(75 MHz, D2O/[D6]DMSO): d ¼ 171.27 (CN-FeII), 167.94, 151.46,
150.94, 146.74, 146.10, 136.24, 127.97, 126.99, 124.12, 119.276,
J. Phys. Org. Chem. 2009, 22 515–521
Copyright ß 2009 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/poc