Photochemical and Electronic Properties of Conjugated Bis(azo) Compounds: An Experimental and Computational Study

Article abstract of DOI:10.1002/chem.200305590

We have investigated the photophysical, photochemical and electrochemical properties of two bis(azo) derivatives, (E,E)-m-1 and (E,E)-p-1. The two compounds, which can be viewed as being composed of a pair of azobenzene units sharing one of their phenyl r

Full text of DOI:10.1002/chem.200305590

FULL PAPER
Photochemical and Electronic Properties of Conjugated Bis(azo)
Compounds: An Experimental and Computational Study
Federico Cisnetti,[a] Roberto Ballardini,[b] Alberto Credi,*[a] Maria Teresa Gandolfi,[a]
Stefano Masiero,[c] Fabrizia Negri,*[a] Silvia Pieraccini,[c] and Gian Piero Spada[c]
Dedicated to Professor Vincenzo Balzani on the occasion of his birthday
Abstract: We have investigated the
photophysical,photochemical and elec-
trochemical properties of two bis(azo)
derivatives,( E,E)-m-1 and (E,E)-p-1.
The two compounds,which can be
viewed as being composed of a pair of
azobenzene units sharing one of their
phenyl rings,differ only for the relative
position of the two azo groups on the
central phenyl ring–meta and para for
m-1 and p-1,respectively. The UV-visi-
ble absorption spectra and photoiso-
merisation properties are noticeably
different for the two structural isomers;
(E,E)-m-1 behaves similarly to (E)-
azobenzene,while ( E,E)-p-1 exhibits a
substantial red shift in the absorption
bands and a decreased photoreactivity.
The three geometric isomers of m-1,
namely the E,E, E,Z and Z,Z isomers,
cannot be resolved in a mixture by ab-
sorption spectroscopy,while the pres-
ence of three distinct species can be re-
vealed by analysis of the absorption
changes observed upon photoisomeri-
sation of (E,E)-p-1. Quantum chemical
ZINDO/1 calculations of vertical exci-
tation energies nicely reproduce the
observed absorption changes and sup-
port the idea that,while the absorption
spectra of the geometrical isomers of
m-1 are approximately given by the
sum of the spectra of the constituting
azobenzene units in their relevant iso-
meric form,this is not the case for p-1.
From a detailed study on the E!Z
photoisomerisation reaction it was ob-
served that the photoreactivity of an
azo unit in m-1 is influenced by the iso-
meric state of the other one. Such ob-
servations indicate a different degree
of electronic coupling and communica-
tion between the two azo units in m-1
and p-1,as confirmed by electrochemi-
cal experiments and quantum chemical
calculations. The decreased photoiso-
merisation efficiency of (E,E)-p-1 com-
pared to (E,E)-m-1 is rationalised by
modelling the geometry relaxation of
the lowest p p* state. These results are
expected to be important for the
design of novel oligomers and poly-
mers,based on the azobenzene unit,
with predetermined photoreactivity.
Keywords: ab initio calculations ¥
absorption spectroscopy ¥ azoben-
zenes ¥ electrochemistry ¥ photo-
chemistry
Introduction
have been widely exploited to modify chemical systems by
means of light stimulation. For instance,such a reaction has
been used to modulate the properties of polymers,[5,6] liquid
crystals,[7] biomolecules,[8] and dendritic[9] and supramolec-
ular[10,11] systems,as well as to construct molecular-level
memories,[12] devices and machines.[13,14]
The physicochemical and structural changes associated with
the E Z photoisomerisation of azobenzene derivatives[1
4]
[a] F. Cisnetti,Dr. A. Credi,Prof. M. T. Gandolfi,Prof. F. Negri
Dipartimento di Chimica ™G. Ciamician∫
Universit‡ di Bologna,via Selmi 2
40126 Bologna (Italy)
Molecular systems containing two or more azobenzene
units are of interest for at least two reasons: 1) because of
their multiphotochromic nature,such compounds can exist
in a wealth of different states (up to 2n,where n is the
number of photochromic units),a behaviour that may be
useful for molecular-level information processing and stor-
age; 2) the cooperation of the different photoisomerisable
units can produce an overall amplification of the geometri-
cal changes related to the E Z transformation,leading to
new light-induced functions.
Fax : (+39)051-2099456
E-mail: alberto.credi@unibo.it
fabrizia.negri@unibo.it
[b] Dr. R. Ballardini
Istituto ISOF-CNR,via Gobetti 101
40129 Bologna (Italy)
[c] Dr. S. Masiero,Dr. S. Pieraccini,Prof. G. P. Spada
Dipartimento di Chimica Organica ™A. Mangini∫
Universit‡ di Bologna,via San Donato 15
40127 Bologna (Italy)
Several compounds containing two azobenzene groups
have been studied[10,15] in the past,whereby,in most cases,
2011
Chem. Eur. J. 2004, 10,2011 2021
DOI: 10.1002/chem.200305590
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

FULL PAPER
the two azobenzenes were found to be basically noninteract-
ing,giving rise to a behaviour determined by the simple su-
perimposition of the properties of the two isolated units.
Here we report on the photochemical properties in solution
of two compounds, m-1 and p-1,composed of two conjugat-
ed azobenzene units.[16] These compounds can be viewed as
two azobenzene units sharing one of their phenyl rings,and
differ only for the relative position of the two azo groups on
the central phenyl ring. Interestingly, p-1 may be considered
as the azo analogue of bis(p-phenylene vinylene),which is
the first member of the oligo(p-phenylene vinylene) (OPV)
family. Enantiopure derivatives with camphanic ester side
groups have been prepared with the aim of obtaining molec-
ular chiral switches for doping nematic liquid crystals. Upon
E Z photoisomerisation it could be expected the induction
of skewed chiral conformations that should modify the pitch
of the induced cholesteric phases.[17] Ester substituents in
the para-position have little or no effect[18] on the photo-
chemical properties of the azobenzene group,and plain azo-
benzene has been taken as a reference compound for the
E!Z photoisomerisation.
Results and Discussion
Absorption spectra: The absorption spectra of the E,E iso-
mers of m-1 and p-1 in acetonitrile are shown in Figure 1,
together with twice the spectrum of (E)-azobenzene. The
Abstract in Italian: Sono state studiate le propriet‡ fotofisi-
che, fotochimiche ed elettrochimiche di due composti
bis(azo), (E,E)-m-1 and (E,E)-p-1. Tali composti, che posso-
no essere considerati come due unit‡ azobenzene che condi-
vidono uno dei loro anelli benzenici, differiscono solo per la
posizione relativa dei due gruppi azo sull×anello benzenico
centrale (ripettivamente meta e para per m-1 e p-1). Gli spet-
tri di assorbimento UV-visibile e le caratteristiche di fotoiso-
merizzazione dei due isomeri strutturali sono notevolmente
diverse: (E,E)-m-1 si comporta in maniera analoga all×(E)-
azobenzene, mentre (E,E)-p-1 mostra un considerevole spo-
stamento verso il rosso delle bande di assorbimento ed una
diminuzione della fotoreattivit‡. I tre isomeri geometrici di
m-1, E,E, E,Z e Z,Z, non sono distinguibili in miscela me-
diante spettroscopia di assorbimento; le variazioni negli spet-
tri di assorbimento osservate durante la fotoisomerizzazione
di (E,E)-p-1, invece, rivelano la presenza di tre specie distin-
te. I valori dell×energia di eccitazione verticale ottenuti me-
diante calcoli quantomeccanici ZINDO/1 sono in ottimo ac-
cordo con gli spettri di assorbimento. Tali calcoli suggerisco-
no inoltre che gli spettri di assorbimento degli isomeri geome-
trici di m-1 sono sostanzialmente dati dalla somma degli
spettri delle unit‡ azobenzene che li costituiscono, ciascuna
nella corrispondente forma isomerica, mentre la situazione di
p-1 õ completamente diversa. Studi dettagliati sulla reazione
di fotoisomerizzazione E!Z indicano che la fotoreattivit‡ di
un×unit‡ azo in m-1 dipende dalla forma isomerica dell×altra
unit‡. Tali osservazioni mostrano che l×accoppiamento e la
comunicazione elettronica tra le due unit‡ azo in m-1 e p-1
sono diversi, come confermato da esperimenti elettrochimici
e da calcoli quantomeccanici. La diminuzione dell×efficienza
di fotoisomerizzazione di (E,E)-p-1 rispetto a (E,E)-m-1 õ
interpretata modellando il rilassamento geometrico dello
stato p p* pifl basso. Questi risultati sono rilevanti per la
progettazione di nuovi oligomeri e polimeri contenenti unit‡
azobenzene aventi una fotoreattivit‡ predeterminata.
Figure 1. Absorption spectra of (E,E)-m-1 (dashed line) and (E,E)-p-1
(full line) in acetonitrile at room temperature. The absorption spectrum
of (E)-azobenzene,multiplied by 2,is also shown for comparison
(dashed-dotted line).
spectroscopic data are gathered in Table 1. Similarly to that
of azobenzene,the spectra of ( E,E)-m-1 and (E,E)-p-1 ex-
hibit a two-band pattern,with well-separated p p* and n
p* absorption features in the 270 520 nm range,and an in-
tensity which is roughly that expected for the presence of
two azobenzene chromophores.[19] The absorption maxima
of (E,E)-m-1 are only slightly shifted relative to azobenzene,
and very similar to those of an azobenzene derivative bear-
ing alkyl ester groups in the 4,4’ positions.[18] In contrast,the
absorption bands of (E,E)-p-1 (in particular the p p* one)
are noticeably displaced to lower energy.
2012
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10,2011 2021

Bis(azo) Compounds
2011 2021
Table 1. Absorption,photochemical and electrochemical data.
Absorption[a]
E!Z Photoreaction[a]
p p* irradiation
Electrochemistry[b,c]
n p* irradiation
p p*
n p*
lmax [nm] (e [mÀ1 cmÀ1])
lmax [nm] (e [mÀ1 cmÀ1])
F
%Z PSS[d,e]
F
%Z PSS[d]
E [V vs SCE][f]
(E)-azobenzene
(E,E)-m-1
(E,E)-p-1
316 (29000)
326 (52000)
361 (53000)
442 (770)
0.14
0.12
0.017
95
79
32
0.35
0.30
0.03
15
18
~1
À1.38
436 (2500)[g]
~450 (~4000)[g]
À0.86
À0.84; À0.85
[a] Air-equilibrated acetonitrile,298 K. [b] Argon-purged acetonitrile with 0.05 m TEAP. [c] First reduction process of the azobenzene unit(s). [d] Fraction
of (Z)-azo units at the photostationary state (PSS). [e] Irradiation wavelength of maximum conversion (345 nm for (E)-azobenzene and (E,E)-m-1,
384 nm for (E,E)-p-1). See the text for details. [f] Irreversible processes; peak potential estimated from DPV peaks. [g] Partially overlapped with the p
p* band.
Compounds (E,E)-m-1 and (E,E)-p-1 are not luminescent
either at room temperature or in a frozen glass at 77 K,in
line with the behaviour of azobenzene.[1,2]
Since the composition of the photostationary state ob-
tained with 365 nm light,expressed in azo units,is E:Z=
35:65,the photoreaction must span all three geometric iso-
mers,namely, E,E, E,Z and Z,Z. The presence of a set of
clean isosbestic points,which are maintained throughout the
irradiation,suggests that the spectrum of ( E,Z)-m-1 is
almost the sum of half the spectrum of (E,E)-m-1 and half
that of (Z,Z)-m-1,as expected for two weakly interacting
chromophores. Alternative explanations for this behaviour
could be based on the sequential reaction scheme E,E!
E,Z!Z,Z,assuming that the quantum yield for the second
photoisomerisation step is much higher than that of the first
step,or considering that the absorption of one photon
causes the isomerisation of both azo units. These models,
however,can be discarded since they are in contrast with
the behaviour observed for (E,E)-m-1,which shows a de-
crease of the apparent quantum yield of photoisomerisation
with irradiation time,and are not in agreement with the re-
sults of the quantum chemical modelling (see below). It can
therefore be concluded that,from UV-visible absorption
data,only the total amount of E and Z isomeric forms of
the azo units of m-1 can be determined,irrespective of their
belonging to the same or different molecules. As will be
shown in the next section,the computed vertical excitation
energies and intensities for the three isomers of m-1 agree
with the experimental observations. The inset of Figure 2
shows the spectrum of the final photoproduct,( Z,Z)-m-1,
obtained by subtraction of the absorption spectrum of the
reactant,together with the absorption spectrum of ( Z)-azo-
benzene. These observations,together with the absorption
behaviour and the similarities with azobenzene,indicate
that m-1 is composed of two weakly interacting azo units. A
quantitative discussion of the electronic coupling in these
bis(azo) derivatives,fully supporting the above conclusions,
will be given in the next section.
E!Z photoisomerisation: The E!Z photoisomerisation
quantum yields and the composition of the photostationary
states at selected irradiation wavelengths are reported in
Table 1. The photoisomerisation of azobenzene was also
studied under our experimental conditions as a control ex-
periment. The quantum yield values determined are in
agreement with those found in earlier accurate studies.[3]
Compound (E,E)-m-1: Figure 2 shows the absorption
changes observed upon irradiation of (E,E)-m-1 in its p p*
band at 365 nm. Such changes are consistent with the E!Z
isomerisation of the azobenzene units of m-1. Similarly to
that of (E)-azobenzene,the photoisomerisation of ( E,E)-m-
1 is more efficient if irradiation is performed in the n p* ab-
sorption band. It should be noted that since (E,E)-m-1 con-
tains two azobenzene units,the E!Z photoisomerisation
For photochromic systems in which the two forms absorb
in the same spectral region,such as those described here,
the absorbance changes at short irradiation times must be
employed for the determination of the photoreaction quan-
tum yield. At higher photoconversions,the absorption of
light by the photoproduct becomes substantial and back-
photoisomerisation cannot be neglected. However,if the ir-
radiation is performed at a wavelength at which the product
exhibits a very low molar absorption coefficient relative to
the reactant,the quantum yield is expected to show small
changes also for higher degrees of photoconversion. An ™in-
stantaneous∫ apparent quantum yield for the disappearance
of the E species can be defined [Eq (1)]:
Figure 2. Absorption spectral changes on 365 nm irradiation of (E,E)-m-1
(1.1î10 5 m in acetonitrile,room temperature). Irradiation times are 0,
0.33,0.66,1,1.5,2,3,5,7.5 and 30 minutes. Inset: absorption spectrum of
(Z,Z)-m-1 (full line) and twice the spectrum of (Z)-azobenzene (dashed
line).
quantum yield per azobenzene unit is actually higher than
that of (E)-azobenzene (Table 1). Such an increased photo-
reactivity of the (E)-azo group in (E,E)-m-1 can be rational-
ised on the basis of quantum chemical calculations (vide
infra).
2013
Chem. Eur. J. 2004, 10,2011 2021
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

A. Credi,F. Negri,et al.
FULL PAPER
AEl ðt2ÞÀAEl ðt1Þ
1
zene; the E!Z quantum yield is only approximately 0.02
(Table 1). In irradiation experiments at 313,334,365 and
404 nm (Hg emission lines) very limited E!Z conversion
could be achieved,the best result being about 13% on
365 nm irradiation. In the attempt to isolate the absorption
spectrum of the photoproduct,an appropriate fraction of
the initial absorption spectrum was subtracted to the spec-
trum of the photostationary state. From this calculation it
can be noted that the photoproduct has an absorbance mini-
mum at about 384 nm. Therefore,irradiation at this opti-
mised wavelength was performed and,as expected,a higher
conversion (ca. 30%) was attained. Figure 4 shows the ab-
Fappðt2Þ ¼
V
ð1Þ
eEl ÀeZl
Nðt2Àt1ÞFE
in which AEl (t1) and AlE(t2) are the solution absorbance at
irradiation time t1 and t2,respectively,measured at a wave-
length l at which the absorbance of the Z form is negligible,
eEl and eZl are the molar absorption coefficient of the E and
Z forms,respectively,at the wavelength l, V is the volume
of irradiated solution, N is the moles of incident photons
per unit time in the volume V,and FE is the fraction of irra-
diation light absorbed by the E form. Figure 3 shows the
Figure 3. Changes of the instantaneous apparent quantum yield for E!Z
photoisomerisation of azobenzene (empty circles) and m-1 (full circles),
normalised to the inititial quantum yield value,as a function of the pho-
toconversion,expressed as the fraction of photogenerated ( Z)-azo units.
To ensure the same spectral conditions,the measurements have been car-
ried out on solutions with similar absorbance.
Figure 4. Absorption spectral changes on 384 nm irradiation of (E,E)-p-1
(1.1î10 5 m in acetonitrile,room temperature). Irradiation times are 0,2,
4,8.5,16,30 and 75 minutes. Inset: absorption spectrum of the primary
photoproduct.
sorption changes upon irradiation of (E,E)-p-1 at 384 nm;
the spectrum of the photoproduct,obtained as described
above,is represented in the inset.
changes of Fapp(t),normalised to the initial quantum yield,
as a function of the photoconversion,expressed as the frac-
tion of photogenerated (Z)-azo units. It can be clearly seen
that the E!Z photoisomerisation quantum yield of m-1 de-
creases much more rapidly than that of azobenzene,espe-
cially at low conversion. For example,at a E!Z conversion
of 10%,the quantum yield for ( E,E)-m-1 is approximately
18% lower than its initial value. Since we have checked that
the Z!E photoisomerisation in (Z,Z)-m-1 is not faster than
that of azobenzene,such an effect cannot be ascribed to the
back-photoisomerisation. Therefore,one can conclude that
the photoreactivity of one azo unit in m-1 depends on the
The presence of isosbestic points may suggest behaviour
similar to that of m-1,but the limited E!Z conversion does
not ensure the formation of species in which both azo units
have been isomerised into the Z form. Thus,it is possible
that the primary photoproduct observed is mainly the E,Z
isomer. This is in agreement with the spectrum shown in the
inset of Figure 4,which is very different from that of ( Z)-
azobenzene and (Z,Z)-m-1 (Figure 2,inset). In particular,
the intense absorption band with maximum at approximate-
ly 340 nm may be assigned to the p p* transition of the (E)-
azo unit in the E,Z photoproduct; such a band is more simi-
lar to that found for (E,E)-m-1 rather than that for (E,E)-p-
1,possibly because of partial loss of electronic conjugation
with the more distorted (Z)-azo unit (see the next section).
This would also imply the absorption spectrum of (E,Z)-p-1
is not the average of the spectra of (E,E)-p-1 and (Z,Z)-p-1.
Such a conjecture has been confirmed by consecutive irradi-
ation experiments with light of different wavelengths and
will be supported in the next section on the basis of quan-
tum chemical calculations of excitation energies.
[20]
isomeric form,that is, E or Z,of the neighbouring unit.
More specifically,the presence of a ( Z)-azo unit in (E,Z)-m-
1 reduces the quantum yield value for the E!Z photoiso-
merisation of the (E)-azo unit. Note that the presence of a
(Z)-azo unit in (E,Z)-m-1 introduces an asymmetry in the
torsional potential-energy curve that might be one of the
causes of the reduced photoreactivity,since isomerisation in
one of the two possible directions will be strongly hindered
by steric repulsions.
Compound (E,E)-p-1: The photoreactivity of (E,E)-p-1 is
much decreased relative to the meta-isomer and azoben-
From the absorption spectra shown in Figure 4 it can be
determined that selective irradiation of the primary photo-
2014
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10,2011 2021

Bis(azo) Compounds
2011 2021
product can be best carried out at 260 nm,whereby more
than 50% of the incident light is absorbed by the photo-
product. Therefore,a solution of ( E,E)-p-1 was first irradiat-
ed at 384 nm,until the photostationary state had been
reached,and subsequently irradiated at 260 nm. Upon
260 nm irradiation,the isosbestic points observed during ir-
radiation at 384 nm are lost; new isosbestic points,slightly
shifted from the previous location,appear on prolonged ir-
radiation. The time evolution of the absorbance at 241 nm
(which represents an isosbestic point for the 384 nm irradia-
tion) during 260 nm irradiation is reported in Figure 5. A
similar behaviour was observed for the other isosbestic
points. Such a result can be rationalised with the following
model,which involves the reactant (A),the primary photo-
product (B),and a third species (C) as the secondary photo-
product [Eq. (2)].
Figure 5. Time-dependent change of the absorbance at 241 nm for a solu-
tion of (E,E)-p-1 (pre-irradiated at 384 nm up to a photostationary state)
upon irradiation at 260 nm. The wavelength 241 nm was selected as the
observation wavelength,since it is an isosbestic point for the initial
384 nm irradiation. The line shows the simulation of the absorbance
change on the basis of the kinetic reaction model described in the text.
F1
F2
A
G
B
H G
C
H
ð2Þ
FÀ1
FÀ2
The following zero-order kinetic equations [Eq. (3) (5)]
can be derived from the definition of the photoreaction
quantum yield.
cating the occurrence of a thermal Z!E isomerisation proc-
ess,similar to that of azobenzene. The influence of such a
slow process on the photochemical studies presented in this
work can be safely neglected.
D½AŠ
Dt
eB260½BŠ N
eA260½AŠ N
¼FÀ1ð1À10ÀA
Þ
ÀF1ð1À10ÀA
Þ
Þ
260
260
ð3Þ
A260
V
A260 V
Quantum chemical modelling: To clarify the trend observed
in the absorption maxima and to assist the assignment of the
photoproducts,vertical excitation energies were obtained
with ZINDO/1 + CIS calculations carried out at the HF/6
31G computed equilibrium geometries. The computed exci-
tation energies of the lowest p p* transitions are compared,
in Table 2,with the correspondingly observed absorption
maxima. It is seen that the agreement between observed
and computed absorption maxima is very satisfactory.
To rationalise the spectroscopical and photochemical be-
haviour of the bis(azo) derivatives upon excitation in the p
p* region,and to quantify the degree of electronic coupling
in the m-1 and p-1 compounds,we can analyse the electron-
D½BŠ
Dt
eA260½AŠ N
eC260½CŠ N
¼F1ð1À10ÀA
Þ
þFÀ2ð1À10ÀA
260
260
A260
V
A260 V
ð4Þ
ð5Þ
eB260½BŠ N
ÀðFÀ1 þ F2Þð1À10ÀA
Þ
260
A260
V
D½CŠ
Dt
eB260½BŠ N
eC260½CŠ N
¼F2ð1À10ÀA
Þ
ÀFÀ2ð1À10ÀA
Þ
260
260
A260
V
A260 V
in which the symbols used have the same meaning as those
in Equation (1). From a numerical solution of these equa-
tions,the photokinetic behaviour of the system can be simu-
lated (Figure 5). The above rate equations could be used in
principle to determine the unknown parameters from the fit
of absorbance versus time pro-
files. In the present case,how-
Table 2. Comparison between computed and observed p p* excitation energies.
ever,the absorbance changes
are very small and there are
many unknown parameters in
the above equations. Thus,the
simulation of Figure 5 should
be regarded only as a qualita-
tive proof that a third species,
with a characteristic absorption
spectrum,appears in the reac-
tion mixture upon irradiation.
On the basis of the above re-
sults,we infer that the B and C
species are in fact (E,Z)-p-1
and (Z,Z)-p-1.
observed[a]
lmax [nm]
computed[b]
DE
Wf composition
DE [nm]
(f[c]
Gap
(e [mÀ1 cmÀ1])
)
[eV]
[eV][d]
(E)-azobenzene
(E,E)-m-1
316 (29000)
326 (52000)
317 (1.04)
332 (1.92)
316 (0.27)
366 (1.89)
298 (0.00)
301 (0.27)
327 (1.18)
307 (0.09)
299 (0.17)
344 (1.17)
300 (0.11)
308 (0.31)
301 (0.27)
322 (0.60)
297 (0.00)
3.92
3.74
3.92
3.39
4.16
0.70 (H!L)
0.18
0.77
0.46 (HÀ1!L); 0.52 (H!L+1)
0.62 (H!L); 0.31 (HÀ1!L+1)
0.67 (H!L); 0.21(HÀ1!L+1)
0.46 (HÀ1!L); 0.52 (H!L+1)
(E,E)-p-1
361 (53000)
282 (6300)
(Z)-azobenzene
(E,Z)-m-1
(E,Z)-p-1
(Z,Z)-m-1
(Z,Z)-p-1
340 (28000)
~290[e] (10000)
After irradiation, m-1 and p-
1 showed a slow (timescale of
days) recovery of the original
absorption bands of their E,E
form if stored in the dark,indi-
[a] Air-equilibrated acetonitrile,298 K. [b] Vertical excitation energies DE from ZINDO/1 + CIS calculations.
[c] Oscillator strength. [d] Energy gap between the lowest two p p* states that would be degenerate in decou-
pled bis(azo) derivatives. See the discussion in the text. [e] Shoulder.
2015
Chem. Eur. J. 2004, 10,2011 2021
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

A. Credi,F. Negri,et al.
FULL PAPER
ic nature (i.e.,the wavefunction compositions) of the lowest
p p* excited states. These are collected in the last column
of Table 2 for the E,E isomers of the bis(azo) derivatives
and for (E)-azobenzene.
Note that,for simplicity,in the following discussion we
will focus only on the p p* excited states with dominant
pNN p* character and disregard those dominated by pphenyl
p* character,which are also involved in the excited-state re-
laxation of azobenzene.[21] Inspection of Table 2 shows that
while the lowest p p* state of (E)-azobenzene is dominated
by the HOMO LUMO (H!L) excitation,owing to the
presence of two azobenzene units,at least four molecular or-
bitals,namely the H À1,H,L and L +1 must be taken into
account to describe the lowest p p* excited states of the
bis(azo) derivatives. The HF/3 21G frontier molecular orbi-
tals are given in Figure 6 for (E)-azobenzene and in Fig-
ures 7 and 8 for (E,E)-m-1 and (E,E)-p-1,respectively. Note
Figure 7. HF/3-21G frontier orbitals of (E,E)-m-1 computed at the equili-
brium structure of the ground state. From bottom to top: HÀ1,H,L,L
+
1 molecular orbitals.
Figure 6. (E)-azobenzene: a) Ground state (HF/3-21G) and lowest p p*
excited state (CIS/3-21G,underlined) equilibrium structures (bond
lengths in ä). b) Frontier orbitals computed at the ground state geome-
try.
Figure 8. HF/3-21G frontier orbitals of (E,E)-p-1 computed at the equili-
brium structure of the ground state. From bottom to top: HÀ1,H,L,L
+
1 molecular orbitals.
that ZINDO/1 and HF/3 21G methods predict exactly the
same order and shape for the pNN molecular orbitals. The
frontier orbitals of both bis(azo) derivatives are delocalised,
as are the excited states built from them. Similar delocalisa-
tion does not imply,however,similar conjugation efficiency,
since the latter is determined by the magnitude of the elec-
tron electron coupling and is manifested by the energy split-
ting of the molecular orbitals. The increased degree of cou-
pling,or conjugation,in ( E,E)-p-1 versus (E,E)-m-1 can be
clearly seen from Scheme 1 in which the energies of the HF/
3 21G frontier orbitals are reported for (E)-azobenzene and
the two bis(azo) derivatives. The H,H À1 and L,L +1 mo-
lecular orbitals,(and similarly the lowest two p p* states)
would be degenerate for two fully decoupled azobenzene
units,while they are energetically separated in the conjugat-
ed compounds studied here and the splitting is larger for the
para compound.
The four-configuration nature of the lowest two p p* of
the bis(azo) derivatives implies that a discussion of their
photophysical and photochemical properties cannot be
based simply on the inspection of the HOMO and LUMO
orbitals. For instance,in contrast with ( E)-azobenzene and
(E,E)-p-1,the lowest p p* excited state of (E,E)-m-1 is do-
minated by the (HÀ1!L) and (H!L+1) excitations. The
larger energy separation computed for the lowest two p p*
excited states of (E,E)-p-1 (0.77 eV,see Table 2) with re-
spect to (E,E)-m-1 (0.18 eV) is also a consequence of the
more efficient conjugation in the former. We can now inter-
2016
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10,2011 2021

Bis(azo) Compounds
2011 2021
Scheme 1. Energies of the HF/3-21G frontier orbitals of (E)-azobenzene,
(E,E)-m-1 and (E,E)-p-1 bis(azo) derivatives.
pret the different spectroscopic behaviour of (E,E)-p-1 with
respect to (E,E)-m-1 in terms of the stronger coupling,in
the former,between the lowest two p p* states that would
be degenerate for a decoupled bis(azo) derivative. Because
of such coupling,the lowest of the two p p* excited states
is pushed at much lower energies relative to the correspond-
ing state in (E,E)-m-1. Conversely,the second lowest p p*
state is found at much higher energies in (E,E)-p-1
(4.16 eV) with respect to (E,E)-m-1 (3.92 eV,see Table 2).
As a result,the absorption spectrum of ( E,E)-m-1 almost
overlaps with twice the spectrum of a single (E)-azobenzene
unit. The spectroscopic features of the E,Z and Z,Z isomers
can be interpreted with similar considerations.
Figure 9. Computed equilibrium structures of the ground state (HF/3-
21G) and lowest p p* excited state (CIS/3-21G,underlined) for a) ( E,E)-
m-1 and b) (E,E)-p-1. Bond lengths are in ä.
As will be shown below,the larger energy separation be-
tween the lowest two p p* excited states in (E,E)-p-1 is also
associated with its reduced photoisomerisation efficiency.
To understand the remarkably different photoisomerisa-
tion efficiency of the two (E,E)-bis(azo) derivatives,we car-
ried out ab initio calculations (CIS/3 21G) of geometry opti-
misation and obtained the relaxed structure of the lowest p
p* state. The computed equilibrium structure of the p p*
state and,for comparison,of the ground state of ( E)-azo-
À
benzene are shown in Figure 6. It is seen that the N N bond
length increases from 1.24 ä in the ground state to 1.33 ä in
the excited state. The lengthening of this bond is associated
with the efficient photoisomerisation reaction. In Figure 9
the computed equilibrium structures of the ground and
lowest p p* excited state of the two (E,E)-bis(azo) deriva-
tives are shown. The p p* state of (E,E)-p-1 relaxes to a de-
À
localised structure characterised by two equal N N bond
lengths of 1.30 ä. The N N elongation upon excitation
(1.24 ä in the ground state) is considerably reduced with re-
spect to (E)-azobenzene and this accounts for its reduced
photoisomerisation efficiency. Indeed,the increased barrier
to isomerisation will favour other deactivation channels in-
volving,for instance,phenyl-ring dynamics. The wavefunc-
tion nature of this excited state does not change along the
relaxation path,and the frontier orbitals of ( E,E)-p-1 com-
puted at the equilibrium geometry of the p p* state are vir-
tually identical to those depicted in Figure 8. In contrast,the
p p* state of (E,E)-m-1 changes nature along the relaxation
path,as shown by the wavefunction composition and molec-
ular orbital shapes (see Figure 10),and relaxes to a localised
Figure 10. HF/3-21G frontier orbitals of (E,E)-m-1 computed at the equi-
librium structure of the lowest p p* excited state. From bottom to top:
HÀ1,H,L,L +1 molecular orbitals.
À
structure characterised by two N N bond lengths remarka-
bly different. Upon relaxation the molecule loses the sv
symmetry plane. As a result,one N N remains virtually un-
2017
Chem. Eur. J. 2004, 10,2011 2021
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

A. Credi,F. Negri,et al.
FULL PAPER
changed and the other increases to 1.34 ä. The remarkable
À
elongation of a single N N bond,which becomes even
longer than that of excited azobenzene (1.33 ä) supports,on
the one hand,the increased photoisomerisation efficiency of
(E,E)-m-1 relative to (E)-azobenzene and,on the other,it
suggests a sequential E,E!E,Z!Z,Z photoisomerisation
mechanism. The different excited state relaxation of (E,E)-
m-1 with respect to (E,E)-p-1 can be ascribed to the pres-
ence of a strong electron phonon coupling mediated by the
asymmetric NN stretching vibrational mode as shown in
Scheme 2. This vibrational mode drives the localisation of
Figure 11. Cyclic voltammetric curves for the first reduction process of
the azobenzene units of (E,E)-m-1 (dashed line) and (E,E)-p-1 (full
line). Conditions: 4î10 4 m Ar-purged acetonitrile,room temperature,
0.05m TBAP; scan rate=200 mVs.
energy of the LUMO (see Scheme 1 and the discussion in
the previous section),and 2) electron-withdrawing ester sub-
stituents. It is interesting to note that the first peak of the
para derivative is split into two processes,while the meta
compound shows a single peak. Apparently,uptake of one
electron by (E,E)-p-1 affects the reduction potential of the
resulting anion,while this does not happen for ( E,E)-m-1. A
simple,qualitative explanation for the different electro-
chemical behaviour can be inferred by inspecting the
LUMO orbital splitting in Scheme 1. Indeed,the small
energy separation between the LUMO and LUMO+1 orbi-
tals computed for (E,E)-m-1 implies similar reduction po-
tentials for both reduction processes,whilst the considerably
larger splitting computed for (E,E)-p-1 supports the splitting
measured for the para compound. Furthermore,if we
assume that the anion formed by the first reduction process
behaves in a similar fashion to the lowest p p* state of the
neutral molecule,we can provide an additional justification
for the observed electrochemical behaviour. Only one peak
is observed for (E,E)-m-1,due to the localised nature of the
reduced state; this leaves the second azobenzene unit and
its reduction potential almost unchanged with respect to the
neutral molecule. In contrast,two peaks are observed for
(E,E)-p-1,due to the delocalised nature of the anion,which
is a new species characterised by its own reduction potential.
Thus,once more,the experimental and computational re-
sults reinforce each other and indicate a more efficient con-
jugation between the azo units in the para than in the meta
compound.
Scheme 2. Schematic representation of the delocalised [(E,E)-p-1] and lo-
calised [(E,E)-m-1] relaxed structure of the lowest p p* state of the
bis(azo) derivatives. The approximate shape of the vibrational mode cou-
pling the two p p* states is also shown.
the excitation in the lowest p p* state by coupling and
mixing the almost degenerate couple of pNN p* states. Such
mechanism is not effective in (E,E)-p-1 owing to the large
energy difference between the two pNN p* states. In summa-
ry,it is the balance between the electron electron interac-
tion (conjugation) and the electron vibration interaction
that drives the localised or delocalised character of the p p*
excited state,which,in turn,affects the photoisomerisation
efficiency of the (E,E)-bis(azo) derivatives.
Electrochemical behaviour: The electrochemical properties
of (E)-azobenzene have been the subject of a detailed inves-
tigation.[22] In our conditions,azobenzene shows two not
fully reversible reduction processes,and two irreversible oxi-
dation processes. Compounds (E,E)-m-1 and (E,E)-p-1 ex-
hibit poorly reversible reduction processes and irreversible
oxidation processes,suggesting a very complex electrochem-
ical behaviour.[23] The peaks appearing at fairly negative po-
tentials (below À1.6 V),not observed for azobenzene,are
most likely related to the reduction of the ester side
groups.[24] For the sake of simplicity,we will discuss only the
first reduction process (Table 1),which is assigned to the
uptake of one electron per each azobenzene unit.
Conclusion
The UV-visible absorption spectra,the E!Z photoisomeri-
sation reaction and the electrochemical properties of two
bis(azo) derivatives,( E,E)-m-1 and (E,E)-p-1,have been in-
vestigated. These two compounds are composed of a pair of
azobenzene units sharing one of their phenyl rings,and
differ only for the relative position of the two azo groups on
the central phenyl ring. In summary,the absorption spectra
Figure 11 shows the CV peaks corresponding to the first
reduction process of (E,E)-m-1 and (E,E)-p-1 in acetonitrile.
The first reduction for both the bis(azo) compounds occurs
at a less negative potential relative to azobenzene,owing to
the presence of 1) an extended p system,which lowers the
2018
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10,2011 2021

Bis(azo) Compounds
2011 2021
sion percentages (<10%; if necessary,extrapolation to t=0 was made).
The fraction of light transmitted at the irradiation wavelength was taken
into account in the calculation of the yields. Where possible,the composi-
tion of the photostationary state and the absorption spectrum of the pho-
toproduct were evaluated by a calibrated subtraction of a fraction of the
initial absorption spectrum from the absorption spectrum at the photosta-
tionary state. The Z!E photoisomerisation experiments on (Z,Z)-m-1
were carried out with samples obtained by exhaustive irradiation of
(E,E)-m-1 and contained at least 60% of the Z,Z isomer. The experi-
mental error on the quantum yield values was estimated to be 10% for
p p* irradiation,and 20% for n--p* irradiation.
and photoisomerisation properties of (E,E)-m-1 are similar
to those of (E)-azobenzene,while ( E,E)-p-1 exhibits a sub-
stantial red shift in the absorption bands and a much de-
creased photoreactivity. From a detailed study on the quan-
tum yield for the E!Z photoisomerisation reaction in
(E,E)-m-1,it can be observed that the photoreactivity of an
(E)-azo unit in an m-1 molecule is quenched by a neigh-
bouring (Z)-azo unit. The three geometric isomers of m-1,
namely E,E, E,Z and Z,Z,cannot be resolved by absorption
spectroscopy,whereas a careful analysis of the absorption
changes observed upon photoisomerisation of p-1 reveals
the presence of three distinct species. These observations
can be rationalised in terms of a stronger electronic commu-
nication of the two azo units in p-1 with respect to m-1.
Quantum chemical calculations of molecular orbitals at
the ground-state geometry show similar delocalisation for p-
1 and m-1,but the computed energies do indeed confirm a
larger electronic coupling for the para compound. Accord-
ingly,the computed vertical excitation energies closely
follow the observed behaviour. The balance between the
electron electron coupling (dominant in the para com-
pound) and the electron vibration coupling (dominant in
the meta compound) determines the delocalised (p-1) or lo-
calised (m-1) nature of the relaxed p p* excited state com-
puted for the E,E isomers and accounts for the reduced
photoisomerisation efficiency of p-1.
Electrochemical experiments were carried out by employing cyclic vol-
tammetry (CV) and differential pulse voltammetry (DPV) techniques in
¾
argon-purged acetonitrile (Romil HiDry ) at room temperature,with an
Autolab 30 multipurpose instrument interfaced to a PC. The working
electrode was a glassy carbon electrode (Amel; 0.07 cm2); its surface was
routinely polished with a 0.3 mm alumina/water slurry on a felt surface,
immediately prior to use. The counterelectrode was a Pt wire,separated
from the solution by a frit; an Ag wire was employed as a quasi-reference
electrode,and ferrocene was present as an internal standard. The concen-
tration of the compounds examined was 4î10À4 m; tetrabutylammonium
hexafluorophosphate (TBAP) 0.05m was added as supporting electrolyte.
Cyclic voltammograms were obtained at sweep rates varying from 20 to
500 mVsÀ1; differential pulse voltammograms were recorded at sweep
rates of 20 and 4 mVsÀ1 with a pulse height of 75 or 10 mV,respectively,
and a duration of 40 ms. The potential window examined was from À2 to
+2 V versus SCE. The experimental error on the potential values was es-
timated to be 10 mV.
Synthesis of (E,E)-p-1: 4’-Aminoacetanilide (5.41 g,36 mmol) was dis-
solved in water (13 mL) and 37% HCl (13 mL). The mixture was cooled
to 08C and diazotised by adding a cold solution of NaNO2 (2.61 g,
37.8 mmol) in water (7 mL) at such a rate as to keep the temperature of
the reaction mixture below 28C. A few crystals of urea were then added
and the mixture was slowly transferred into a cooled solution of phenol
(3.29 g,35 mmol) and NaOH (7.0 g,0.17 mol) in water (25 mL). The
slurry was allowed to warm to room temperature and carefully acidified
with conc. HCl to about pH 6. Filtration of the resulting mixture afforded
4-(4’-hydroxyphenylazo)acetanilide as an orange-brown solid in a 76%
Similar considerations on the reduced forms of p-1 and
m-1 explain the results of electrochemical experiments.
Therefore, m-1 can be considered closer to a two-compo-
nent species composed of weakly interacting azobenzene
units,whereas p-1 is better described as a large molecule[25]
in which the two chromophoric groups are electronically
coupled.
These results may be useful for the design of novel
oligomers and polymers,based on the azobenzene unit,with
predetermined photoreactivity.
yield. ESI-MS (MeOH): m/z: 256 [M+H]+
;
1H NMR (200 MHz,
[D6]DMSO): d=2.11 (s,3H),6.94 (m,2H),7.78 (m,6H),10.23 (s,1H),
10.25 ppm (s,1H).
4-(4’-Hydroxyphenylazo)acetanilide (6.97 g,27.3 mmol) was dissolved in
10% NaOH (350 mL) refluxed for 2 h. The mixture was then cooled to
room temperature and the pH was adjusted to 6 with conc. HCl. The re-
sulting precipitate was filtered,washed with water and dried in vacuo to
afford 4-(4’-hydroxyphenylazo)aniline as pale brown solid (86%). The
compound was sufficiently pure to be used in the subsequent step with-
Experimental Section
out any further purification. ESI-MS (MeOH): m/z: 214 [M+H]+
1H NMR (200 MHz,CDCl 3): d=4.0 (brs,2H),5.07 (s,1H),6.68 (m,
2H),6.91 (m,2H),7.79 ppm (m,4H).
;
General methods: Reagents and dry solvents were purchased from Fluka
unless otherwise noted. (E)-Azobenzene was purchased from Aldrich.
NMR spectra were recorded with a Varian Gemini 200 instrument. ESI-
MS spectra were obtained with a Micromass ZMD 4000 spectrometer.
4-(4’-Hydroxyphenylazo)aniline (4.95 g,23.2 mmol) was suspended in
water (15 mL) and conc. HCl (8.4 mL); the mixture was then cooled to
08C. A cold solution of NaNO2 (1.61 g,23.3 mmol) in water (4 mL) was
added at such a rate to keep the temperature below 38C. The addition
was stopped when iodine-starch paper gave a positive test. A few crystals
of urea added and the mixture was slowly transferred into a cooled (08C)
solution of phenol (2.18 g,23.2 mmol) and NaOH (4.65 g,116 mmol) in
water (18 mL). When the addition was complete,the resulting viscous
dark brown suspension was allowed to reach room temperature. The mix-
ture was made slightly acidic with conc. HCl and filtered. The precipitate
was washed with water and dried in vacuo. The filtrate was extracted
with diethyl ether and the organic layer was dried over Na2SO4,concen-
trated in vacuo and added to the precipitate. The solid was purified by
column chromatography over silica gel (petroleum ether/ethyl acetate 7:3
as the eluent) and the product with Rf =0.2 was collected. 1,4-bis(4’-hy-
droxyphenylazo)benzene was thus obtained as an orange solid in a 76%
yield. ESI-MS (MeOH): m/z: 317 [MÀH]À; 1H NMR (200 MHz,
[D6]DMSO): d=6.96 (m,4H),7.85 (m,4H),7.98 (s,4H),10.42 ppm (s,
2H); 13C NMR (200 MHz,[D 6]DMSO): d=115.96 (CH),123.13 (CH),
125.04 (CH),145.32 (C),152.62 (C),161.22 ppm (C).
The absorption and luminescence spectra were recorded with a Perkin
Elmer l40 spectrophotometer and an LS 50 spectrofluorimeter,respec-
¾
tively,in air-equilibrated acetonitrile (Merck Uvasol ) at room tempera-
À6
ture,with concentrations ranging from 1î10
to 1î10À5 m,contained in
1 cm quartz cells. Experimental errors: wavelengths, ꢀ1 nm; molar ab-
sorption coefficients,5%.
Photochemical reactions were performed on the same solutions,thor-
oughly stirred,by using either a Hanau Q400 medium pressure Hg lamp
(150 W) or the Xe lamp (150 W) of a Perkin Elmer 650 10S spectro-
fluorimeter. In the first case,selection of the desired irradiation wave-
length (313,334,365,405 or 436 nm) was accomplished by the use of ap-
propriate combinations of interference,broad-band and cut-off filters,
while in the second case the monochromator of the instrument (band-
pass=5 nm) was employed. The number of incident photons,determined
by ferrioxalate actinometry[26] in its micro version,[27] was between 10À8
and 10À6 Nhn minÀ1. The E!Z quantum yields for p p* (lirr =313 or
365 nm) and n p* (lirr =436 nm) irradiation were determined from the
disappearance of the p p* absorption band of the reactant at low conver-
2019
Chem. Eur. J. 2004, 10,2011 2021
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim

A. Credi,F. Negri,et al.
FULL PAPER
(À)-Camphanic acid chloride (300 mg,1.3 mmol) was quickly added to a
solution of 1,4-bis(4’-hydroxyphenylazo)benzene (110 mg,0.35 mmol)
and redistilled Et3N (0.15 mL) in dry CH3CN (5 mL) under inert atmos-
phere. The mixture was allowed to react for 2 h at room temperature and
reduced to dryness in vacuo. The residue was redissolved in CHCl3
(30 mL) and washed with 5% NaHCO3 (3î30 mL) and water (15 mL).
The resulting organic phase was dried over Na2SO4,the solvent was re-
moved by distillation and the crude was purified by chromatography on
silica gel (eluent: petroleum ether/ethyl acetate 6:4). Compound (E,E)-p-
1 was obtained as an orange solid in a 45% yield. 1H NMR (200 MHz,
CDCl3): d=1.13 (s6,H),1.17 (s6,H),1.18 (s,6H),1.70 1.87 (m,2H),
1.94 2.10 (m,2H),2.16 2.30 (m,2H),2.51 2.68 (m,2H),7.31 (m,4H),
7.99 8.09 ppm (m,8H); elemental analysis calcd (%) for C 38H38N4O8
(678.74): C 67.24,H 5.64,N 8.25; found: C 67.10,H 5.59,N 8.18.
Tahara, J. Phys. Chem. A 2001, 105,8123; c) T. Fujino,S. Y. Arz-
hantsev,T. Tahara, Bull. Chem. Soc. Jpn. 2002, 75,1031; d) L. Ga-
gliardi,G. Orlandi,F. Bernardi,A. Cembran,M. Garavelli,
Theor.
Chem. Acc.,in press.
[5] a) A. Natansohn,P. Rochon, Chem. Rev. 2002, 102,4139; b) G.
Sudesh Kumar,D. C. Neckers, Chem. Rev. 1989, 89,1915.
[6] For a recent interesting example,see: Y. Lu,M. Nakano,T. Ikeda,
Nature 2003, 425,145.
[7] T. Ikeda,A. Kanazawa in Molecular Switches (Ed.: B. L. Feringa),
Wiley-VCH,Weinheim, 2001,Chapter 12.
[8] a) O. Pieroni,A. Fissi,N. Angelini,F. Lenci, Acc. Chem. Res. 2001,
34,9; b) I. Willner,B. Willner in Molecular Switches (Ed.: B. L. Fer-
inga),Wiley-VCH,Weinheim,
2001,Chapter 6; c) F. Ciardelli,O.
Pieroni in Molecular Switches (Ed.: B. L. Feringa),Wiley-VCH,
Weinheim, 2001,Chapter 13.
Synthesis of (E,E)-m-1: The preparation is analogous to that described
for (E,E)-p-1. Spectral data for the corresponding derivatives are report-
ed:
[9] For representative examples,see: a) A. Archut,F. Vˆgtle,L. De
Cola,G. C. Azzellini,V. Balzani,P. S. Ramanujam,R. H. Berg,
Chem. Eur. J. 1998, 4,699; b) D. M. Junge,D. V. McGrath, J. Am.
Chem. Soc. 1999, 121,4912; c) R. M. Sebastian,J. C. Blais,A. M.
Caminade,J.-P. Majoral, Chem. Eur. J. 2002, 8,2172.
3-(4’-Hydroxyphenylazo)acetanilide: ESI-MS (MeOH): m/z: 256
[M+H]+.
3-(4’-Hydroxyphenylazo)aniline: ESI-MS (MeOH): m/z: 214 [M+H]+.
[10] a) V. Balzani,F. Scandola, Supramolecular Photochemistry,Hor-
1,3-Bis(4’-hydroxyphenylazo)benzene: ESI-MS (MeOH): m/z: 317
wood,Chichester,
1991,Chapter 7; b) S. Shinkai in
Molecular
[MÀH]À; 1H NMR (200 MHz,CDCl 3): d=6.97 (m,4H),7.63 (t,
J=
Switches (Ed.: B. L. Feringa),Wiley-VCH,Weinheim, 2001,Chap-
ter 9.
8.0 Hz,1H),7.9 8 (m,6H),8.36 ppm (t,
J=1.9 Hz,1H).
1
(E,E)-m-1: H NMR (200 MHz,CDCl ): d=1.15 (s,6H), 1.18 (s,6H), 1.19
3
[11] For early,representative examples,see: a) A. Ueno,H. Yoshimura,
R. Saka,T. Osa, J. Am. Chem. Soc. 1979, 101,2779; b) S. Shinkai,T.
(s,6H),1.71 1.87 (m,2H),1.94 2.10 (m,2H),2.17 2.32 (m,2H),2.52
2.68 (m,2H),7.33 (m,4H),7.69 (t,1H),8.00 8.09 (m,6H),8.44 ppm (t,
1H); elemental analysis calcd (%) for C38H38N4O8 (678.74): C 67.24,H
5.64,N 8.25; found: C 67.11,H 5.60,N 8.20.
Nakaji,Y. Nishida,T. Ogawa,O. Manabe,
J. Am. Chem. Soc. 1980,
102,5860.
[12] S. Kawata,Y. Kawata, Chem. Rev. 2000, 100,1777.
Computational methods: Ground-state equilibrium structures of the
three geometric isomers (E,E, E,Z and Z,Z) of meta- and para-bis(azo)
derivatives and of the two isomers (E and Z) of azobenzene were opti-
mised at HF/6 31G and HF/3 21G levels of theory. For simplicity,cam-
phanic ester side groups were replaced by acetate groups in the calcula-
tions. The HF/6 31G equilibrium geometries were employed to calculate
electronic excitation energies with the ZINDO/1 semi-empirical hamilto-
nian[28] combined with configuration interaction singles (CIS) calcula-
tions,which included the highest 20 occupied and lowest 20 unoccupied
(20î20) molecular orbitals (MOs). In order to obtain a qualitative pic-
ture of the p p* excited-state geometry relaxation,excited-state geome-
try optimisations were performed at CIS/3 21G level of theory. CIS cal-
culations with moderate basis sets are known to overestimate excitation
energies,but provide a reliable description of the geometry relaxation for
excited states dominated by single excitations. Given the large dimension
of the molecules investigated,the orbital space was restricted to 10î10
for (E,E) -m-1 and p-1 and to 6î6 for (E)-azobenzene. All the calcula-
tions were carried out with the Gaussian 98 suite of programs.[29] The
visual program Molekel[30] was employed to produce molecular orbitals
pictures.
[13] V. Balzani,A. Credi,M. Venturi,
Molecular Devices and Ma-
chines–A Journey Into the Nano World,Wiley-VCH,Weinheim,
2003.
[14] For a recent example of a single-molecule optomechanical device
based on azobenzene,see: T. Hugel,N. B. Holland,A. Cattani,L.
Moroder,M. Seitz,H. E. Gaub, Science 2002, 296,1103.
[15] Representative examples: a) S. Shinkai,O. Manabe, Top. Curr.
Chem. 1984, 121,67; b) G. Ritter,G. H‰felinger,E. L¸ddecke,H.
Rau, J. Am. Chem. Soc. 1989, 111,4627; c) N. Tamaoki,K. Koseki,
T. Yamaoka, Angew. Chem. 1990, 102,66; Angew. Chem. Int. Ed.
Engl. 1990, 29,105; d) F. Vˆgtle,W. M. M¸ller,U. M¸ller,M.
Bauer,K. Rissanen, Angew. Chem. 1993, 105,1356; Angew. Chem.
Int. Ed. Engl. 1993, 32,1295; e) O. Rˆttger,H. Rau, Mol. Cryst. Liq.
Cryst. Sci. Technol. Sect. A 1994, 246,143; f) A. Bencini,M. A. Ber-
nardo,A. Bianchi,M. Ciampolini,V. Fusi,N. Nardi,A. J. Parola,F.
Pina,B. Valtancoli, J. Chem. Soc. Perkin Trans. 2 1998,413.
[16] Compounds related to those object of this study,with appropriate
substituent groups,owing to their photochemical stability,are widely
employed as dyes. Commercial names are,for example,Bismarck
Brown R and Sudan 403. To the best of our knowledge,no informa-
tion is available on photoreactive conjugated bis(azo) compounds.
[17] Unfortunately,the pitches of the cholesteric phases obtained by
doping the compounds (E,E)-p-1 and (E,E)-m-1 in several nematic
phases do not change to an appreciable extent during the photoiso-
merisation process. Furthermore,the circular dichroism spectrosco-
py in ordinary solutions does not give any evidence of formation of
skewed,helicoidal conformations of the solute molecules [O. Pando-
li,Laurea Thesis,University of Bologna (Italy), 2003]. For a recent
successful bis(azo) chiral photochemical switch for liquid crystal
doping,see: S. Pieraccini,S. Masiero,G. P. Spada,G. Gottarelli,
Chem. Commun. 2003,598.
Acknowledgments
This research was supported by MIUR (PRIN programs),FIRB
(Nomade project) and the University of Bologna (Funds for Selected Re-
search Topics). F.C. contributed to this work as research student from the
…cole Normale Supÿrieure,Paris,which he gratefully acknowledges. We
are indebted with Giorgio Orlandi and Sandra Monti for useful discus-
sions.
[18] a) C. Ruslim,K. Ichimura, J. Mater. Chem. 1999, 9,673; b) C.
Ruslim,K. Ichimura, J. Mater. Chem. 2000, 10,2704.
[19] For an early report on the absorption spectra of related compounds,
see: K. Ueno, J. Am. Chem. Soc. 1952, 74,4508.
[20] A similar effect has recently been observed in a symmetric biphoto-
chromic system consisting of two naphthopyran units. See: W. Zhao,
E. M. Carreira, J. Am. Chem. Soc. 2002, 124,1582.
[1] H. Rau in Photochromism: Molecules and Systems (Eds.: H. D¸rr,
H. Bouas-Laurent),Elsevier,Amsterdam, 2003,Chapter 4.
[2] J. Griffiths, Chem. Soc. Rev. 1972, 1,481.
[3] P. Bortolus,S. Monti, J. Phys. Chem. 1979, 83,648,and references
therein.
[21] T. Schultz,J. Quenneville,B. Levine,A. Toniolo,T. J. Martinez,S.
[4] The mechanism of E!Z photoisomerisation of azobenzene was ex-
tensively investigated in the past (ref. [1]) and has recently received
a renewed attention. See ref. [21] and: a) T. Fujino,T. Tahara, J.
Phys. Chem. A 2000, 104,4203; b) T. Fujino,S. Y. Arzhantsev,T.
Lochbrunner,M. Schmitt,J. P. Shaffer,M. Z. Zgierski,A. Stolow,
Am. Chem. Soc. 2003, 125,8098.
[22] J. L. Sadler,A. J. Bard, J. Am. Chem. Soc. 1968, 90,1979.
J.
2020
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10,2011 2021

Bis(azo) Compounds
2011 2021
[23] A. M. Shams-El Din,T. M. H. Saber,N. M. Abed,
J. Electroanal.
Tomasi,V. Barone,M. Cossi,R. Cammi,B. Mennucci,C. Pomelli,
Chem. 1969, 21,377.
[24] Organic Electrochemistry (Eds.: H. Lund,O. Hammerich),Dekker,
New York, 2001.
[25] a) V. Balzani,F. Scandola, Supramolecular Photochemistry,Hor-
wood,Chichester, 1991,Chapter 3; b) V. Balzani,A. Credi,M. Ven-
turi, Chem. Eur. J. 2002, 8,5524.
[26] C. G. Hatchard,C. A. Parker, Proc. R. Soc. London Ser. A 1956,
253,518.
C. Adamo,S. Clifford,J. Ochterski,G. A.Petersson,P. Y. Ayala,Q.
Cui,K. Morokuma,D. K. Malick,A. D. Rabuck,K. Raghavachari,
J. B. Foresman,J. Cioslowski,J. V. Ortiz,A. G. Baboul,B. B. Stefa-
nov,G. Liu,A. Liashenko,P. Piskorz,I. Komaromi,R. Gomperts,
R. L. Martin,D. J. Fox,T. Keith,M. A. Al-Laham,C. Y. Peng,A.
Nanayakkara,C. Gonzalez,M. Challacombe,P. M. W. Gill,B. John-
son,W. Chen,M. W. Wong,J. L. Andres,C. Gonzalez,M. Head-
Gordon,E. S. Replogle,J. A. Pople,Gaussian,Inc.,Pittsburgh PA,
1998.
[27] E. Fischer, EPA Newsl. 1984, 20,33.
[28] A. D. Bacon,M. C. Zerner, Theo. Chim. Acta 1979, 53,21.
[29] Gaussian 98 (revision A.7),M. J. Frisch,G. W. Trucks,H. B. Schle-
gel,G. E. Scuseria,M. A. Robb,J. R. Cheeseman,V. G. Zakrzewski,
J. A. Montgomery,Jr.,R. E. Stratmann,J. C. Burant,S. Dapprich,
J. M. Millam,A. D. Daniels,K. N. Kudin,M. C. Strain,O. Farkas,J.
[30] Molekel Version 4.3,http://www.cscs.ch/molekel/; S. Portmann,H. P.
L¸thi, Chimia 2000, 54,766.
Received: October 1,2003
Revised: December 8,2003 [F5590]
2021
Chem. Eur. J. 2004, 10,2011 2021
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA,Weinheim