The effect of the conformation on the quadratic nonlinear optical response of metal carbonyl based chromophores with one-dimensional charge transfer capabilities: A computational investigation

Article abstract of DOI:10.1039/b601315e

The synthesis and the quadratic nonlinear optical (NLO) properties of (4-(4′-dimethylaminophenyl)pyridine)chromium and tungsten pentacarbonyl are reported. Experimentally, the chromophores exhibit molecular hyperpolarizabilities (β) equal to 15.5 and 21.7 × 10^-30 cm⁵ esu^-1, respectively at 1.064 μm, which arise from a charge transfer capability along the molecular C₂ symmetry axis, in relation to the π-conjugated structure of the ligand. At a more fundamental level, these molecules have been investigated as possible synthetic targets to illustrate the ZINDO prediction that, in chromophores with one-dimensional charge transfer capabilities a specific set of substituents could be found to switch the β direction, on passing from a ground state geometry with co-planar aromatic rings to a conformation in which the rings are perpendicular. Due to the C₂ symmetry requirement, this situation implies that β is necessarily vanishing for a specific conformation. The possibilities to observe experimentally and to use this intriguing effect in a perspective of a molecular switch are critically evaluated. the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006.

Full text of DOI:10.1039/b601315e

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/njc | New Journal of Chemistry
The effect of the conformation on the quadratic nonlinear optical
response of metal carbonyl based chromophores with one-dimensional
charge transfer capabilities: a computational investigation
Jean-Franc¸
ois Lamere,a Isabelle Sasaki,a Pascal G. Lacroix*a and
Keitaro Nakatanib
Received (in Montpellier, France) 27th January 2006, Accepted 23rd March 2006
First published as an Advance Article on the web 10th April 2006
DOI: 10.1039/b601315e
The synthesis and the quadratic nonlinear optical (NLO) properties of (4-(40-
dimethylaminophenyl)pyridine)chromium and tungsten pentacarbonyl are reported.
Experimentally, the chromophores exhibit molecular hyperpolarizabilities (b) equal to 15.5 and
21.7 ꢀ 10ꢁ30 cm5 esuꢁ1, respectively at 1.064 mm, which arise from a charge transfer capability
along the molecular C2 symmetry axis, in relation to the p-conjugated structure of the ligand. At
a more fundamental level, these molecules have been investigated as possible synthetic targets to
illustrate the ZINDO prediction that, in chromophores with one-dimensional charge transfer
capabilities a specific set of substituents could be found to switch the b direction, on passing from
a ground state geometry with co-planar aromatic rings to a conformation in which the rings are
perpendicular. Due to the C2 symmetry requirement, this situation implies that b is necessarily
vanishing for a specific conformation. The possibilities to observe experimentally and to use this
intriguing effect in a perspective of a molecular switch are critically evaluated.
In this equation, m0 is the permanent dipole moment, a is the
Introduction
polarizability, and b the quadratic hyperpolarizability, respon-
sible for the NLO properties. Up to now, the most convincing
NLO switch seems to be the report by Coe et al. that b is 10 to
20 times reduced by oxidation of a ruthenium(II) complex.5a
Nevertheless, the ultimate molecular NLO switch implies that,
in the off-state, a strict cancellation of the b tensor must be
achieved on a single entity.
The concept of molecular switches is currently attracting a
growing interest from a variety of perspectives,1,2 and could
lead to a family of derivatives with intriguing applications in
biology, electronic, and photonic technologies. For instance,
molecules with nonlinear optical (NLO) properties have led to
a multidisciplinary research area of great importance for the
emerging optoelectronics and photonic technologies relevant
to all optical computing applications.3 Furthermore, the in-
herent tailorability of molecular compounds renders them
ideally suited for the creation of materials possessing switch-
able properties.
In the present contribution, we wish to report on a compu-
tational approach towards this intriguing situation, within the
framework of the proven ZINDO/SOS formalism (vide infra).
In previous studies, we have provided evidences for the
possibility of an absolute switch obtained by 901 b rotation
achieved upon protonation.9 However, this idea is based on
the assumption that b is a vector in order for the on-state (b
parallel to the electric field of the light), to be turned into an
off-state after rotation (b orthogonal with respect to the field).
In fact, the hyperpolarizability defined in eqn (1) is a third-
rank tensor, with 27 components. Except for molecules having
pure one-dimensional charge transfer character, in which b ꢂ
Nevertheless, the possibilities for achieving a switch in the
quadratic (pE2) NLO response has become considered only
recently.4 Most chromophores are built up from three basic
components: (i) an electron rich substituent connected, through
(ii) a p-electron bridge, to (iii) an electron acceptor counterpart,
leading to charge delocalization by resonance. Therefore,
switching the NLO response has been achieved either by
changing the donating5 or accepting6 strength of the substi-
tuents or by changing the conjugated capabilities7 of the bridge.
In all instances, the (on - off) switch invariably arises from a
reduction of the magnitude of the quadratic hyperpolarizability,
defined from the molecular polarization as follows:8
bzzz along the OZ charge transfer direction, the contribution of
off-diagonal tensor components (bijk) may have to be taken
into account, for instance in the case of the promising octu-
polar geometries (e.g. Td, D3h, D2d).10 By contrast, we focus
here on a situation in which single charge transfer axis and
symmetry requirement lead to the assumption that b ꢂ bzzz is
valid. The methodology employed for the selection of a
suitable chromophore will be presented first. Then, the possi-
bility of perfect b cancellation by mean of intramolecular
rotation will be discussed for (4-(40-aminophenyl)pyridine)-
chromium pentacarbonyl. To illustrate this computational
approach, a chromium and a tungsten analogue will be
mi(E) = m0i þ aijEj þ bijkEjEk þ . . .
(1)
a Laboratoire de Chimie de Coordination du CNRS (UPR 8241), 205
route de Narbonne, 31077 Toulouse, France. E-mail: pascal@
lcc-toulouse.fr; Fax: þ33 5 61 55 30 03
b
´
PPSM, Ecole Normale Superieure de Cachan (UMR 8531),
61 Avenue du Pdt Wilson, 94235 Cachan, France
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 921–928 | 921

View Article Online
Scheme 1
Scheme 2
presented and their NLO properties will be reported (Scheme
1). In the last section, and in a perspective of material science,
the possibility to observe a b cancellation in a real molecular
device will be critically evaluated.
of the substituents. In other words, the b magnitude can
change, but not its sign. This situation is exemplified in Fig.
1 for 4-(40-nitrophenyl)aniline (2, Scheme 2), where the calcu-
lation indicates a b reduction from 52.8 to 11.6 ꢀ 10ꢁ30 cm5
esuꢁ1, the sign being not affected by the rotation.
Results and discussion
At this point, it must be emphasized that, if this limitation
could be over-passed, changing the b sign between y = 0 and
901 would necessarily imply that b is strictly vanishing for a
specific angular value. However, this situation being hardly
expectable within the standard ‘‘push–pull’’ model illustrated
by the archetypical structure 2, we have been screened sets of
R1/R2 substituents of different natures (organic vs. inorganic
and organometallic) and non-traditional electronic capabil-
ities. For instance, (CO)5Cr– or (H3N)5RuII– become intri-
guing fragments once combined to a pyridine ligand. While
electron-rich metal complexes of substituted pyridines have
long been reported to possess intense d - p* back-donation
transitions,11 it has also been observed that the weak acceptor
pyridine unit has the capability to be turned into a strong one
after metal complexation, thus providing the enhancement of a
potential intraligand p - p* charge transfer, when a donor
substituent is present in para position of the pyridine ring.12–16
These two processes might compete in the overall electronic
response of the complex, each of them becoming dominant for
different y ranges, with the outcome of the overall NLO
response being gradually tuned from positive to negative
values, as a function of the conformation. Finally, and after
various computational screenings, the chromium complexes 3
(Scheme 2) was found to be the one having the most intriguing
NLO behavior. The origin of its unexpected properties will
therefore be thoroughly discussed in the next section, in
relation to those of the parent molecular structures 4 and 5
(Scheme 2).
Selection of candidates
In order to ensure the one-dimensional topology required for
the design of chromophores in which b is a pure vector, the
investigations have been restricted to molecules built up from
the general skeleton 1, depicted in Scheme 2. In this picture, R1
and R2 are donor or acceptor substituents, and y is the angle
between the two aromatic rings. R1 and R2 are chosen (e.g.
–NO2 and –NH2 in compound 2) in order for the molecule to
possess a two-fold symmetry axis in its ground state. This
topology is made necessary to ensure that the symmetry axis is
maintained for any y rotation, which leads to the following
electronic features: (i) the b magnitude is y dependent, for a
given set of R1 (donor) and R2 (acceptor), but (ii) the charge
transfer direction (and hence b) stays strictly parallel to the
axis whatever the y value. It can be anticipated that the
situation with y = 01 will be that allowing the best charge
delocalization within the bridge (highest b value), while y =
901 will result in a poor orbital overlap (reduced b value).
Nevertheless, the overall R1 to R2 charge transfer process will
stay qualitatively the same, a reverse R2 to R1 charge transfer
being not possible, owing to the intrinsic electronic capabilities
ZINDO investigation of the angular dependence of b
The calculated hyperpolarizabilities at y = 0 and 901 are
gathered in Table 1, for molecules 2, 3 and 4 (the dimethyla-
mino-analogue of molecule 3). In the ground state (y = 01),
the b values are positive and equal to 5.8 and 12.8 ꢀ 10ꢁ30 cm5
esuꢁ1, for 3 and 4, respectively. This situation indicates that, in
3 and 4, the overall effect is that of a charge transfer directed
from the aniline to the pyridine, contrary to most (CO)5Cr–
and (CO)5W– based molecules investigated in the literature, in
which b is negative (d - p* charge transfer).13,15
Fig. 1 ZINDO calculated hyperpolarizabilties for compound 2, as a
function of the y rotation along the molecular C2 symmetry axis
(defined in Scheme 1).
The most striking result in Table 1 is the negative b value
calculated at 901 for compound 3, which suggests the
ꢃc
922 | New J. Chem., 2006, 30, 921–928 This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
Table 1 Quadratic hyperpolarizabilities (b in 10ꢁ30 cm5 esuꢁ1) at
1.064 mm, for 2, 3 and 4, as calculated by ZINDO for different
conformations
Table 2 ZINDO spectrum calculated at y = 01, with absorption
maximum (lmax), oscillator strengths (f), change in dipole moment
upon transitions (Dm in D), and composition of the dominant excited
states involved in the NLO response of 2
y = 01
y = 901
Statea
(%)
Composition of
CI expansiona
2
3
4
52.8
5.8
12.8
11.6
ꢁ1.6
1.3
Transition
l
max/nm
f
Dm/D
1 - 4
1 - 7
347
266
0.76
0.25
18.0
9.3
0.79
0.04
0.910 w40-41
ꢁ0.896 w40-42
þ 0.344 w37-41
a
possibility of b cancellation upon y rotation, as anticipated
from the C2 symmetry. The angular dependence of b is shown
more precisely in Fig. 2. As y increases, the b magnitude is
gradually reduced, as previously calculated in 2. Nevertheless,
and in striking contrast vs. 2, the reduction is such in the
chromium complex 3 that b is strictly zero at a y value equal to
631, then becomes negative.
In order to provide an analysis for the microscopic origin of
the NLO response in the present derivatives, it has to be
reminded that within the framework of the sum over state
(SOS) perturbation theory, b is ultimately related to all
electronic transitions of a molecule.17 Nevertheless, there is a
single transition in most cases with low energy (E), large
intensity (oscillator strength, f), and large charge transfer
capability (dipole moment change, Dm), which significantly
contributes to the NLO effect, according to the widely used
‘‘two-level’’ description of b, according to the following rela-
tion:18,19
Orbital 40 is the HOMO and 41 the LUMO.
shown in Fig. 3. In this respect, molecule 2 behaves as a
traditional NLO chromophore.
The situation encountered in the chromium complex 3 is
more difficult to rationalize. In this case, the calculated
spectrum reveals that many transitions exhibit a potential
NLO effect (p fDm/E3), according to eqn. (2). Furthermore,
and in striking contrast to 2, the calculation indicates two
types of transitions in complex 3 (Table 3): (i) transitions
exhibiting charge transfers from the aniline to the
(CO)5Cr–pyridine fragment (Dm 4 0), and transitions with
charge transfers in the reverse (chromium to aniline, Dm o 0)
direction (ii). Owing to the b sign, the first group (i) is
dominant, and contains most of the electronic transitions of
the molecule. Interestingly, all of them involve charge transfers
through the C–C ring connection. The effect of y rotation is
likely to perturbate these charge transfers, and finally to
decrease their overall contribution to the optical nonlinearity.
The second group (ii) is roughly restricted to the 1 - 5
(HOMO - LUMO based) transition, which involves an
intense charge transfer from the chromium atom to the
pyridine, with almost no contribution of the aniline fragment,
as shown in Fig. 4. In contrast to the previous transitions (i),
1 - 5 is very slightly affected by the y rotation and its
contribution may become dominant at large y value, thus
providing a qualitative illustration for the change occurring in
the overall charge transfer properties. Owing to the complexity
of the NLO response in metal carbonyl chromophores,13 the
present description may be somewhat over simplified. Never-
theless, it provides a simple and coherent rationale to qualita-
tively account for the intriguing behavior of complex 3.
Finally, the NLO behavior of chromium derivative 4 ap-
pears to be intermediate between those of 2 and 3. Although
the dimethylamino-fragment is only slightly more donor than
the amino-fragment, the calculations indicate a two times
enhancement of the b value (from 5.8 to 12.8 ꢀ 10ꢁ30 cm5
esuꢁ1) between 3 and 4. The relatively small b value for 3
3e2ꢀhf Dm
2mE3
E4
2
b ¼
ð2Þ
2
ðE2 ꢁ ð2ꢀhoÞ ÞðE2 ꢁ ðꢀhoÞ Þ
In this equation, ꢀho is the energy of the incident laser beam.
The benchmark reference compound 2 provides a perfect
illustration of this model (Table 2), with a calculated spectrum
strongly dominated by a single intense low-lying 1 - 4
transition having large charge transfer capability (Dm = 18.0
D). It is therefore readily assumed that this transition accounts
for the qualitative understanding of the NLO response. 83%
of the transition (0.9102 in Table 2) is described as the HOMO
(aniline-based) - LUMO (nitrophenyl-based) excitation,
which fully agrees with the chemical intuition. The charge
transfers associated to this dominant 1 - 4 transition is
Fig. 3 Difference in electronic population between ground and
excited states in the dominant transitions responsible for the NLO
effect in chromophore 2, calculated with y = 0. The white (black)
contribution is indicative of an increase (decrease) in electron density
in the charge transfer process.
Fig. 2 ZINDO calculated hyperpolarizabilties for 3, as a function of
the y rotation along the molecular C2 symmetry axis.
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 921–928 | 923

View Article Online
Table 3 Description of the low-energy (lmax) ZINDO calculated transitions of 4, with large oscillator strength (f), and large charge transfer (Dm)
character, in the conformation in which y = 01
Transition
lmax/nm
Dm/D
Weighta
Compositionb
Charge transfer character
1 - 5
1 - 30
305
195
ꢁ9.8
þ6.7
1
0.4
94% (60 - 61)
28% (58 - 64)
23% (58 - 67)
22% (60 - 63)
21% (60 - 65)
81% (57 - 65)
75% (57 - 63)
86% (57 - 61)
(OC)5Cr - pyridine
(OC)5Cr - (OC)5Cr
(OC)5Cr - (OC)5Cr
Aniline - (OC)5Cr
Aniline - (OC)5Cr
Aniline - (OC)5Cr
Aniline - (OC)5Cr
Aniline - pyridine
1 - 40
1 - 36
1 - 17
a
180
184
267
þ30.8
þ30.3
þ7.4
0.25
0.2
0.1
b
Calculated with eqn. (2). A reference value of 1 is assigned to transition 1 - 5. Orbital 60 is the HOMO and 61 the LUMO.
suggests that –NH2 and –Cr(CO)5 exhibit rather comparable
donating capabilities. The larger difference (–NMe2 vs.
–Cr(CO)5) encountered in complex 4 is such that the rotation
lowers but cannot remove the dominant aniline to pyridine
charge transfer effect (Table 1). Along this line, the overall
behavior of complex 4 is qualitatively similar to that of the
organic chromophore 2.
which are generally closely related derivatives, however with
a better stability in the case of tungsten,23 while chromium
carbonyls are potentially explosive compounds and are also
carcinogenic.24 Although ZINDO parameters are not avail-
able for tungsten, the computational predictions made on
Cr(CO)5-based derivatives might qualitatively be used to
describe the electronic features of tungsten analogues as well.13
The calculated b value is extremely modest in complex 3
(5.8 ꢀ 10ꢁ30 cm5 esuꢁ1), and falls close to the limit range of the
detection of the EFISH equipment (vide infra). We have
therefore tested the quality of the NLO computation on
complex 4 (the hyperpolarizability of which is expected to be
larger), and on its tungsten homologue 5. Additionally, the
synthesis of 4-(40-dimethylaminophenyl)pyridine is more sim-
ple, compared to that of the aminophenyl homologue, which
requires a four-step synthetic procedure and expensive re-
agents.25
In order to illustrate the present approach, and to test the
accuracy of the computational prediction, the model com-
pounds 4 and 5 (tungsten homologue of 4) have been synthe-
sised, and the NLO response evaluated.
Experimental investigation on the model compounds
(4-(40-dimethylaminophenyl)pyridine)chromium and -tungsten
pentacarbonyl
Various M(CO)5 pyridine (M = Cr,20 Mo21 and W22) deriva-
tives have been reported in the literature. We have selected
here a chromium complex, and the tungsten homologue,
The model chromophores were synthesized using conven-
tional photochemical methodology:15,23,26,27
hn
MðCOÞ6 ꢁ! MðCOÞ5ðTHFÞ þ CO
ð3Þ
THF
M(CO)5(THF) þ L - M(CO)5L þ THF
(4)
The electronic spectrum of complex 4 is shown in Fig. 5, and
compared to the ZINDO calculated data available in Table 4.
The experimental spectrum is dominated by an intense and
low lying transition having an absorption maximum (lmax
)
Fig. 4 Orbitals (HOMO and LUMO) involved in the 1 - 5 transi-
tion of complex 3, which indicates that the aniline fragment is not
involved in the charge transfer process.
Fig. 5 UV-visible optical absorption spectrum of complex 4, re-
corded in dioxane.
ꢃc
924 | New J. Chem., 2006, 30, 921–928 This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
Table 4 Comparison of the experimental and ZINDO calculated
spectra (lmax, f and e) of 4
10ꢁ30 cm5 esuꢁ1 at 1.064 mm, an unusually small value for a
molecule of such a large size.31 Indeed, the donating strengths
are probably increased in –NMe2 and reduced in –W(CO)5,
thus restoring an observable push–pull character to 5, and
hence an experimental hyperpolarizability of 21.7 ꢀ 10ꢁ30
UV-Vis in dioxane
max/nm
ZINDO data
max/nm
l
e/molꢁ1 L cmꢁ1
l
f
374
B350 (sh)
22 500
20 700
305
267
0.66
0.27
cm5 esuꢁ1
.
Critical evaluation for an experimental observation of the b
cancellation
located at 374 nm (e = 22 500 L molꢁ1 cmꢁ1). An additional,
but less intense transition is visible as a shoulder around 350
nm. The comparison between experimental and calculated
spectra (Table 4) indicates that the experimental UV-visible
spectra are slightly red shifted, vs. the computational data.
Nevertheless, the agreement appears satisfactory and ensure
the reliability of the ZINDO analysis for the description of the
NLO response of these metal carbonyl derivatives.
In the past two decades, theoretical computational quantum
chemistry procedures have profoundly changed the science of
chromophore design. In the context of nonlinear optics, the
fruitful SOS perturbation theory has successfully been applied
to describe the NLO response of many organic and metal
organic chromophores. To make progress today, synthetic
chemists and experimentalists must apply their chemical in-
tuition in concert with insights provided by computational
chemistry to effectively guide synthetic strategies.
Before reporting the experimental NLO properties, the issue
of the molecular stability has to be addressed. Although 3 and
4 exhibit modest thermal stabilities (Td = 135 and 159 1C,
respectively), they are stable at room temperature in the solid
state, but decompose after only a few minutes in aerobic
solution. However, a stability of several days was observed
under argon, which was therefore sufficient to envision the
measurement of the NLO response by the EFISH technique.
The quadratic hyperpolarizabilities measured under argon
atmosphere by the Electric Field Second Harmonic (EFISH)
technique18,28 are shown in Table 5 and compared to the
computational values available. First of all, it is observed that
all the b values are positive, which confirms that the overall
effect is that of a resulting charge transfer from the aniline to
the pyridine–M(CO)5 fragments. The data indicates that the
agreement is satisfactory between experimental (15.5 ꢀ 10ꢁ30
cm5 esuꢁ1) and calculated (12.8 ꢀ 10ꢁ30 cm5 esuꢁ1) b values in
the case of 4, which suggests that the ZINDO prediction is
valid for this class of molecules. Interestingly, a larger experi-
mental b value of 21.7 ꢀ 10ꢁ30 cm5 esuꢁ1 is obtained for the
tungsten derivative. Although no ZINDO data are available in
this case, one may tentatively relate this difference to the well
known trend for larger ligand field splitting parameter (D) on
going down a group (e.g. Cr, Mo, W).29 In the present d6
configuration, it leads to a reduced donating capability of the
tungsten, and finally to a slight enhancement of the overall
aniline - pyridine charge transfer effect, in complete agree-
ment with previously reported observations.13,30
Nevertheless, and in a perspective of material science, the
issue of how to verify experimentally that the predicted b
cancellation by molecular rotation in 3 is real must necessarily
be addressed, with the underlying issue of how to monitor the
conformational change to tune the NLO response in an
operating device. For instance, temperature-dependent b mea-
surements could be performed to test the influence of Boltz-
mann energy on a population of molecules. In a simplified
description, with the assumption that b is restricted to a
ground state (b//) and an excited state (b>) value, the relative
averaged b value (b/bJ) is expressed as:
ꢀ
0:002T
ꢁ
ꢁEg
b
b==
1 þ ? exp
b
b==
0:002T
ꢀ
ꢁ
¼
ð5Þ
ꢁEg
1 þ exp
expression in which Eg is the energy barrier (in kcal) for the y
rotation, the Boltzmann constant being equal to 0.002 kcal
molꢁ1
K
ꢁ1. The possibility for vanishing averaged b value at
300 K would therefore imply the following relation:
b>/bJ = ꢁexp(Eg/0.6)
(6)
This situation is depicted in Fig. 6 as a function of the energy
barrier. The energy barrier calculated in 3 is equal to 1.4 kcal,
a value which is far too small to lead to a real convincing
switch according to the Figure. Nevertheless, this simplified
picture suggests that temperature dependent nonlinear optics,
which up to now has not attracted much attention, may
deserve some interest in the future.
These sizeable modulations achieved on passing from chro-
mium to tungsten and from amino- to a dimethylamino-
substituent may appear somewhat surprising. On the other
hand, it seems to indicate that the situation encountered in the
chromium derivative 3 is close to balance in terms of relative
donor/acceptor electronic capabilities between –NH2 and
–Cr(CO)5, in agreement with a calculated b value of 5.8 ꢀ
The ultimate and most elegant observation would require
the measurements to be performed on a single molecule, pined
in a specific conformation. Single-molecule nonlinear optics
has occasionally been reported.32 Nevertheless, although con-
siderable progresses have been achieved, such experiments are
restricted to chromophores associated with extremely large b
values, and this possibility remains a dream in the present case.
A more realistic (but indirect) approach could be envisioned
on the basis of single-molecule conductance measurements,
following the suggestion by Aviram and Ratner, that a
donor–acceptor chromophore should behave as a diode, when
placed between two electrodes.33 These last few years, the
Table 5 Experimental (EFISH) and calculated (ZINDO) hyperpo-
larizabilities (b) for compounds 4 and 5. The laser wavelength is equal
to 1.064 mm
Compound
1030bEFISH/cm5 esuꢁ1
1030bZINDO/cm5 esuꢁ1
4
5
15.5
21.7
12.8
N.a.
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 921–928 | 925

View Article Online
and to test the accuracy of the ZINDO prediction, two
(chromium and tungsten) derivatives built up from a related
ligand have been characterized. A set of experimental and
computational data suggests that an adjustment of the donat-
ing strength can be achieved between –NH2 and –Cr(CO)5, for
a conformation, in which the aromatic rings are twisted by a
specific angle, found to be equal to 631 in the semi-empirical
analysis.
In a perspective of material science, the chromophores
would require additional chemical or physical capabilities, to
allow the conformation to be tuned by a suitable stimulus (e.g.
proton transfer or pulse of electric field). Along this line,
complex 3 is certainly not the most promising candidate.
Nevertheless, the present investigation suggests that molecules
with low NLO responses or modest air- and thermal stabilities
might be revisited in some cases, as intriguing chromophores
with non-traditional NLO capabilities.
Fig. 6 Temperature dependence of the averaged b value (expressed as
b/bJ), assuming y rotation barriers of 1.4 (1), 2.8 (2) and 5.6 (3)
kcal molꢁ1
.
Experimental
study of electron transport within a single molecule has been
achieved experimentally using scanning tunneling microscopy
(STM).34 In the case of chromophores such as 3, one could
expect that the linear current–voltage characteristics measured
as a function of the y rotation could indicate the possibility for
a reverse overall charge transfer behavior.
Starting materials and equipment
Chromium and tungsten hexacarbonyl (Strem Chemicals)
were used as purchased. High-grade solvents (THF, dioxane)
were used for the syntheses and EFISH measurements. THF
was distilled on Na/benzophenone. Pentane and dichloro-
methane were distilled prior to use. 4-(40-Dimethylamino-
phenyl)pyridine was synthesized following the procedure
previously reported.38 All synthetic operations and measure-
ments involving chromium and tungsten carbonyl derivatives
were carried out with exclusion of oxygen, using standard
Schlenk techniques. The UV-visible spectra were recorded on
The possibility for the control of molecular conformation
remains a fundamental issue. By definition, a switch is a
device, in which two stable states can be obtained upon an
external perturbation. This is not the case in 3, where the ‘‘off’’
state actually corresponds to an excited conformation rather
than a stable state. One might therefore wonder how the y
rotation could be reversibly tuned for a practical use in an
operating molecular device. Certainly, using intramolecular
motion has become a fascinating issue of contemporary
research, for instance in relation to the concept of molecular
machines.35 To make such a device work, various strategies
could be considered to supply energy in some way (via
chemical, photochemical or electrochemical reactions).36 In
the present case, the motion being an oscillation instead of a
rotation, no external energy would have to be supplied, but a
simple stimulus could be envisioned (e.g. an electric field
pulse). Reversible controls of these motions would probably
be extremely difficult to achieve in an operative device. Never-
theless, Nature has provided many examples of molecular
motions occurring in hundreds of different biological molecu-
lar machines, each of them specialized for a particular func-
tion.37 This seems not accessible for our generation of
scientists, but there is no reason to think that molecular
engineering will never be able to reach this goal.
1
a Hewlett Packard 8452A spectrophotometer. H NMR spec-
tra were acquired at 250.13 MHz on a Bruker WM250
spectrometer. The atom labelling used for the assignment is
given in Scheme 1. Elemental analyses were carried out at the
Laboratoire de Chimie de Coordination in Toulouse, for C, H
and N. Thermal measurements were performed by TG/DTA
analysis on a Setaram-TGDTA92 thermoanalyser. The experi-
ments were conducted under nitrogen on 5 mg of sample (rate
of heating 10 1C minꢁ1). The decomposition temperature (Td)
was assigned as the intercept of the leading edge of the
decomposition endotherm with the base line of the DTA scan.
Syntheses
(4-(40-Dimethylaminophenyl)pyridine)chromium pentacarbo-
nyl. Under a nitrogen atmosphere Cr(CO)6 (220 mg, 10ꢁ3
mol) was photolyzed for 2 h in 150 mL of freshly distilled
THF, using a Pyrex-filtered high-pressure Hg lamp. The
resulting yellow Cr(CO)5(THF) solution was then transferred
by cannula to a stirring solution of 4-(40-dimethylaminophe-
nyl)pyridine (198 mg, 10ꢁ3 mol) in 100 mL of freshly distilled
THF. The resulting solution was stirred at room temperature
for 1.5 h, then evaporated to dryness in vacuo. The crude
product was then purified by chromatography on silica, elut-
ing with 1 : 1 pentane–CH2Cl2. Yield: 230 mg (59%) of light
yellow material. Anal. Found: C, 55.23; H, 3.38; N, 6.93. Calc.
for C18 CrH14N2O5: C, 55.39; H. 3.62; N, 7.19%. Td = 138 1C.
1H NMR (CDCl3) for 4-(40-dimethylaminophenyl)pyridine: d
Conclusion
In a computational (ZINDO) approach, we have found that
metal carbonyl-based chromophores are intriguing molecules,
in which a single intramolecular rotation has the potential to
lead to a strict b cancellation. This unusual feature likely
results from the perfect adjustment of the charge transfer
capabilities of the substituents, modulated by the molecular
rotation. To illustrate the synthetic feasibility of such systems,
ꢃc
926 | New J. Chem., 2006, 30, 921–928 This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006

View Article Online
3.026 (s, 6H, H(1)), 6.784 (d, J = 6.9 Hz, 2H, H(2)), 7.497
(d, J = 6.3 Hz, 2H, H(4)), 7.586 (d, J = 6.9 Hz, 2H, H(3)),
8.541 (d, J = 6.3 Hz, 2H, H(5)). For (4-(40-dimethylamino-
phenyl)pyridine)chromium pentacarbonyl: 3.035 (s, 6H, H(1)),
6.774 (d, J = 8.1 Hz, 2H, H(2)), 7.343 (d, J = 4.8 Hz, 2H,
H(4)), 7.534 (d, J = 8.3 Hz, 2H, H(3)), 8.423 (d, J = 5.0 Hz,
2H, H(5)).
Hamiltonian incorporated in the commercially available soft-
ware package ZINDO.46 Tungsten being not parametrized by
ZINDO, no calculation was performed on compound 5.
NLO Measurements
The electric field induced second harmonic (EFISH) technique
was employed for the b measurements. The principle of the
EFISH technique is reported elsewhere.28 The data were
recorded using a nanosecond Nd-YAG pulsed (10 Hz) laser
operating at l = 1.064 mm. The concentrations were around
2 ꢀ 10ꢁ3 mol Lꢁ1. The samples were dissolved under nitrogen
in dioxane, using standard Schlenk techniques, and transferred
by cannula into the EFISH cell under an inert atmosphere.
The centrosymmetry of the solution was broken by dipolar
orientation of the chromophores with a high voltage pulse (5
kV) synchronized with the laser pulse. The SHG signal was
selected through a suitable interference filter, detected by a
photomultiplier, and recorded on an ultrafast Tektronic TDS
620 B oscilloscope. As the NLO response was induced by
dipolar orientation of the chromophores, and in order to
extract the b values, the dipole moments had to be measured
independently by a classic method based on the Guggenheim
theory.47 Further details of the experimental methodology and
data analysis are reported elsewhere.48 Under these experi-
mental conditions, the NLO signal was weak, but sufficient to
allow the b determination of 4 and 5.
(4-(40-Dimethylaminophenyl)pyridine)tungsten pentacarbo-
nyl. The compound was obtained by the same procedure
starting from W(CO)6 (352 mg, 10ꢁ3 mol) instead of Cr(CO)6.
Yield: 318 mg (61%) of light yellow material. Anal. Found: C,
41.28; H, 2.40; N, 5.21. Calc. for C18H14N2O5W: C, 41.40; H.
1
2.70; N, 5.36%. Td = 146 1C. H NMR (CDCl3): d 3.052 (s,
6H, H(1)), 6.807 (d, J = 8.9 Hz, 2H, H(2)), 7.387 (d, J = 6.2
Hz, 2H, H(4)), 7.575 (d, J = 8.9 Hz, 2H, H(3)), 8.661 (d,
J = 6.1 Hz, 2H, H(5)).
Computational methods
The molecular geometry used for the calculations of 4-(40-
nitrophenyl)aniline (2) was that reported in the literature.39
For (4-(40-aminophenyl)pyridine)chromium pentacarbonyl
(3), the geometry was built up from those of (CO)5Cr–
pyridine,40 and 4-phenylpyridine,41 for which crystal struc-
tures have previously been reported. The bond lengths were
fixed at their crystallographic value and the effect of molecular
rotation was studied by steps of 2.51. The all-valence INDO
(intermediate neglect of differential overlap) method,42 was
employed for the calculation of the energy of the chromo-
phores in various conformations. The ‘‘gas-phase’’ geometry
of 3 was investigated first around the metal atom. The stable
conformation was found to be that having two carbonyls
eclipsed with respect to the pyridine ring, while the staggered
conformation (angle of 451) was found to be destabilized by
about 3.3 kcal molꢁ1. Although the ‘‘solid state’’ geometry
available reveals an apparent trend for the staggered confor-
mation between the pyridine and the CO units, reduced
torsion angles (angle o201) are also observed.43–45 The
eclipsed conformation was therefore preferred for the consis-
tency of the INDO method, and any further calculations were
performed with this geometry.
The air stability of metal carbonyl derivatives being modest,
the EFISH experiments were conducted in an anaerobic
argon-purged cell, specially designed for the purpose of the
present investigation. The stability of the EFISH signal in-
dicates that the metal carbonyl-based chromophores do not
undergo any decomposition during the time of the experiment.
Acknowledgements
The authors thank Dr Remi Metivier for dipole moment
´ ´
measurements.
References
1 J. M. Lehn, Supramolecular Chemistry–Concepts and Perspectives,
VCH, Weinheim, 1995.
2 M. D. Ward, Chem. Soc. Rev., 1995, 24, 121.
3 (a) Molecular Nonlinear Optics: Materials, Physics and Devices, ed.
J. Zyss, Academic Press, Boston, MA, 1994; (b) Nonlinear Optics of
Organic Molecules and Polymers, ed. H. S. Nalwa and S. Miyata,
CRC Press, New York, 1997.
4 (a) B. J. Coe, Chem. Eur. J., 1999, 5, 2464; (b) I. Asselberghs, K.
Clays, A. Persoons, M. D. Ward and J. McCleverty, J. Mater.
Chem., 2004, 14, 2831.
5 See, for example: (a) B. J. Coe, S. Houbrechts, I. Asselberghs and
A. Persoons, Angew. Chem., Int. Ed., 1999, 38, 366; (b) F. Paul, K.
Costuas, I. Ledoux, S. Deveau, J. Zyss, J.-F. Halrt and C. Lapinte,
Organometallics, 2002, 21, 5229.
In a next step, INDO was employed in connection with the
sum over state (SOS) formalism17 for the calculation of the
electronic spectra and the molecular hyperpolarizabilities, as a
function of the y rotation between the two aromatic rings (vide
infra). The energy barrier was found to be equal to 1.4 kcal
molꢁ1, which confirms that this rotation is that to be con-
sidered as the basic molecular motion away from the planar
ground state conformation. The NLO response of (4-(40-
dimethylaminophenyl)pyridine)chromium pentacarbonyl (4)
was carried out similarly.
6 See for example: E. Hendrickx, K. Clays, A. Persoons, C. Dehu
and J. L. Bredas, J. Am. Chem. Soc., 1995, 117, 3547.
´
7 See for example: (a) S. L. Gilat, S. H. Kawai and J. M. Lehn,
Chem. Eur. J., 1995, 1, 275; (b) K. Nakatani and J. A. Delaire,
Chem. Mater., 1997, 9, 2682; (c) J. A. Delaire, I. Fanton-Maltey
and K. Nakatani, Mol. Cryst. Liq. Cryst., 2000, 345, 233; (d) M.
The laser wavelength used for the calculation was 1.064 mm.
Details of the computationally efficient INDO-SOS-based
method for describing the quadratic molecular optical non-
linearities have been reported elsewhere.17 In the present
approach, the mono-excited configuration interaction (CIS)
approximation was employed to describe the excited states.
The lowest 100 energy transitions were chosen to undergo CI
mixing. All calculations were performed using the INDO/1
Sliwa, S. Letard, I. Malfant, M. Nierlich, P. G. Lacroix, T. Asahi,
´
H. Masuhara, P. Yu and K. Nakatani, Chem. Mater., 2005, 17,
4727.
8 D. J. Williams, Angew. Chem., Int. Ed. Engl., 1984, 23, 690.
ꢃc
This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006
New J. Chem., 2006, 30, 921–928 | 927

View Article Online
9 J. P. Costes, J. F. Lamere, C. Lepetit, P. G. Lacroix, F. Dahan and
K. Nakatani, Inorg. Chem., 2005, 44, 1973.
26 L. T. Cheng, W. Tam and D. F. Eaton, Organometallics, 1990, 9,
2856.
10 (a) J. Zyss and I. Ledoux, Chem. Rev., 1994, 94, 77; (b) R. Kannan,
G. S. He, T.-C. Lin, P. N. Prasad, R. A. Vaia and L.-S. Tan, Chem.
Mater., 2004, 16, 185; (c) J. Brunel, O. Mongin, A. Jutand, I.
Ledoux, J. Zyss and M. Blanchard-Desce, Chem. Mater., 2003, 15,
4139, and references therein.
11 For example: (a) P. Ford, F. P. De Rudd, R. Gaunder and H.
Taube, J. Am. Chem. Soc., 1968, 90, 1187; (b) A. B. P. Lever, J.
Chem. Educ., 1974, 51, 612.
27 B. S. Creave, C. Long, G. Russel, J. Hamiliton and G. P. Mc
Quillan, Inorg. Chim. Acta, 1988, 146, 25.
28 (a) B. F. Levine and C. G. Betha, J. Chem. Phys., 1975, 63, 2666;
(b) B. F. Levine and C. G. Betha, J. Chem. Phys., 1976, 65, 1989.
29 D. F. Shriver, P. W. Atkins and C. H. Langford, Inorganic
Chemistry, Oxford University Press, Oxford, 1994, p. 247.
30 J. Mata, S. Uriel, E. Peris, R. Llusar, S. Houbrechts and A.
Persoons, J. Organomet. Chem., 1998, 562, 197.
12 (a) D. W. Bruce and A. Thornton, Mol. Cryst. Liq. Cryst., 1993,
231, 253; (b) D. W. Bruce, Adv. Inorg. Chem., 2001, 52, 151.
13 D. R. Kanis, P. G. Lacroix, M. A. Ratner and T. J. Marks, J. Am.
Chem. Soc., 1994, 116, 10089.
14 (a) D. Roberto, R. Ugo, S. Bruni, E. Cariati, F. Cariati, P.
Fantucci, I. Invernizzi, S. Quici, I. Ledoux and J. Zyss, Organo-
metallics, 2000, 19, 1775; (b) D. Roberto, R. Ugo, F. Tessore, E.
Lucenti, S. Quici, S. Vezza, P. Fantucci, I. Invernizzi, S. Bruni, I.
Ledoux-Rak and J. Zyss, Organometallics, 2002, 21, 161.
15 L. T. Cheng, W. Tam, G. R. Meredith and S. R. Marder, Mol.
Cryst. Liq. Cryst., 1990, 189, 137.
31 R. A. Huijts and G. L. J. Hesselink, Chem. Phys. Lett., 1989, 156,
209.
32 For a review on NLO response in single molecules, see for example:
Ph. Tamarat, F. Jelezko, Ch. Brunel, A. Maali, B. Lounis and M.
Orrit, Chem. Phys., 1999, 245, 121.
33 A. Aviram and M. A. Ratner, Chem. Phys. Lett., 1974, 29, 277.
34 C. Joachim, J. K. Gimzewski and A. Aviram, Nature, 2000, 408,
541.
35 Molecular Machines, special issue of Acc. Chem. Res., 2001, 34(6).
36 R. Ballardini, V. Balzani, A. Credi, M. T. Gandolfi and M.
Venturi, Acc. Chem. Res., 2001, 34, 445.
16 For a review of various NLO properties implying ruthenium(II)-
based chromophores, see: H. Le Bozec and T. Renouard, Eur. J.
Inorg. Chem., 2000, 229.
37 R. D. Vale and R. A. Milligan, Science, 2000, 288, 88.
38 A. N. Kost, A. K. Sheinkman and N. F. Kazarinova, J. Gen.
Chem. USSR, 1964, 34, 2059.
17 (a) J. F. Ward, Rev. Mod. Phys., 1965, 37, 1; (b) D. R. Kanis, M. A.
Ratner and T. J. Marks, Chem. Rev., 1994, 94, 195.
18 J. L. Oudar, J. Chem. Phys., 1977, 67, 446.
39 E. M. Graham, V. M. Miskowski, J. W. Perry, D. R. Coulter, A. E.
Stiegman, W. P. Schaefer and R. E. Marsh, J. Am. Chem. Soc.,
1989, 111, 8771.
19 J. L. Oudar and D. S. Chemla, J. Chem. Phys., 1977, 66, 2664.
20 For recent examples, see: (a) M. M. Fares, M. El-Khateeb and K.
J. Asali, J. Inorg. Organomet. Polym., 2003, 13, 143; (b) E. Peris, J.
A. Mata and V. Moliner, J. Chem. Soc., Dalton Trans., 1999, 3893.
21 For recent examples, see: (a) H. A. Cruse and N. E. Leadbeater,
Macromolecules, 1999, 32, 4450; (b) G. M. Li, J. H. Reibenspies
and R. A. Zingaro, Heteroat. Chem., 1998, 9, 57.
22 For recent examples, see: (a) T. W. Stringfield and R. E. Shepherd,
Inorg. Chim. Acta, 2003, 343, 156; (b) C. J. Aspley, J. R. Lindsay
Smith, R. N. Perutz and D. Pursche, J. Chem. Soc., Dalton Trans.,
2002, 170; (c) H. P. Wu, R. Aumann, R. Frohlich, B. Wibbeling
and O. Kataeva, Chem. Eur. J., 2001, 7, 23.
40 W. Ries, I. Bernal, M. Quast and T. A. Albright, Inorg. Chim.
Acta, 1984, 83, 5.
41 A. E. Goeta, S. E. Lawrence, M. M. Meehan, A. O’Dowd and T.
R. Spalding, Polyhedron, 2002, 21, 1689.
42 (a) M. Zerner, G. Loew, R. Kirchner and U. Mueller-Westerhoff,
J. Am. Chem. Soc., 1980, 102, 589; (b) W. P. Anderson, D.
Edwards and M. C. Zerner, Inorg. Chem., 1986, 25, 2728.
43 W. Y. Wong, S. H. Cheung, S. M. Lee and S. Y. Leung, J.
Organomet. Chem., 2000, 596, 36.
44 S. Sakanishi, D. A. Bardwell, S. Couchman, J. C. Jeffery, J. A. Mc
Cleverty and M. D. Ward, J. Organomet. Chem., 1997, 528, 35.
45 U. Klement, Z. Kristallogr., 1993, 208, 114.
23 J. M. Kelly and C. J. Long, J. Organomet. Chem., 1982, 235, 315.
24 R. J. Lewis. Sr., Sax’s Dangerous Properties of Industrial Materials,
Van Nostrand Reinhold, NY, 8th edn, 1992.
25 M. Lamothe, P. J. Pauwels, K. Belliard, P. Schambel and S.
Halazy, J. Med. Chem., 1997, 40, 3542.
46 ZINDO, release 9.60, Molecular Simulations Inc., Cambridge, UK,
1996.
47 E. A. Guggenheim, Trans. Faraday Soc., 1949, 45, 714.
48 I. Maltey, J. A. Delaire, K. Nakatani, P. Wang, X. Shi and S. Wu,
Adv. Mater. Opt. Electron., 1996, 6, 233.
ꢃc
928 | New J. Chem., 2006, 30, 921–928 This journal is the Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2006