Short-circuiting azobenzene photoisomerization with electron-donating substituents and reactivating the photochemistry with chemical modification

Article abstract of DOI:10.1002/ejoc.201100216

Azobenzene (AB undergoes (E → (Z isomerization upon exposure to light. Whereas the light-induced structural change can be exploited for numerous applications, a second, light-orthogonal switch would facilitate the development of new uses for AB derivative

Full text of DOI:10.1002/ejoc.201100216

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201100216
Short-Circuiting Azobenzene Photoisomerization with Electron-Donating
Substituents and Reactivating the Photochemistry with Chemical Modification
H. M. Dhammika Bandara,[a] Shannon Cawley,[a] José A. Gascón,*[a] and
Shawn C. Burdette*[a]
Keywords: Photochemistry / Isomerization / Azobenzene / Density functional calculations / Fluorescence
Azobenzene (AB) undergoes (E) Ǟ (Z) isomerization upon
exposure to light. Whereas the light-induced structural
change can be exploited for numerous applications, a sec-
mental observations and computational studies suggest that
the inclusion of multiple electron-donating groups can short-
circuit the concerted inversion isomerization mechanism of
ond, light-orthogonal switch would facilitate the develop- AB by providing new conical intersections between excited
ment of new uses for AB derivatives. Electron-donating
groups on the AB ring system change not only the absorption
wavelengths, but also the isomerization properties. Experi-
states. This phenomenon has been exploited in a unique AB
derivative where the conversion of a phenol into an ester re-
stores the isomerization activity.
Introduction
Azobenzene (AB) photoisomerization has been explored
extensively since (Z)-AB was isolated over 80 years ago.[1]
Whereas (E)-AB is more stable, (Z)-AB forms upon irradia-
tion with near-UV light. The (Z) Ǟ (E) isomerization oc-
curs with visible light or heating.[2] Illumination of (E)-AB
affords an S2(ππ*) excited state, which rapidly decays to an
S1(nπ*) state where isomerization of the molecule occurs
accordding to a concerted inversion mechanism.[3,4]
Recently, we described the unexpected photochemistry of
three ortho-substituted aminoazobenzene derivatives.[5]
AzoAMP-1 (1) (Figure 1) does not photoisomerize owing
to intramolecular hydrogen bonds. Replacement of the anil-
ino protons with methyl groups restores the isomerization
behavior. The hydrogen bonds in AzoAMP-1 create an en-
ergy barrier in the S1 state that prevents aryl ring distortion,
a prerequisite for isomerization through the concerted in-
version pathway.[5]
Figure 1. Structure of AzoAMP-1. Intramolecular hydrogen bond-
ing prevents photoisomerization. The name AzoAMP refers to the
azobenzene (Azo) and (aminomethyl)pyridine (AMP) components.
The AB photoisomerization depends on the ring substit-
uents.[11] Aminoazobenzene derivatives typically exhibit an
S2ǟS0 transition at longer wavelengths, which causes over-
lap with the S1ǟS0 transition.[12] Given the isomerization
behavior observed with AzoAMP derivatives without hy-
drogen bonds, we reasoned that introduction of additional
electron-donating groups into the AzoAMP ring system
could provide the desired secondary isomerization control.
Light is usually the sole input for controlling the AB
isomerization.[6–10] An additional light-orthogonal switch
could lead to new applications for AB. The hydrogen bonds
in AzoAMP-1 were envisioned as a secondary means to
control the isomerization, but no means to disrupt the
hydrogen bonds could be found.
Results and Discussion
AzoAMP-4 (4) was synthesized by an azo coupling reac-
tion between the diazonium salt 3 and the aminophenol 2
(Scheme 1). The hydroxy group in 2 directs the site of the
electrophilic attack and introduces another electron-donat-
ing group into the AB ring system. The methyl group of 2
also blocks a site susceptible to electrophilic attack and
adds another electron-rich substituent.
Only small changes in the absorbance of AzoAMP-4, in-
dicating minimal (E) Ǟ (Z) isomerization, were measured
after prolonged irradiation (Figure 2). No isomerization oc-
[a] Department of Chemistry, University of Connecticut,
55 North Eagleville Road Unit 3060, Storrs, CT 06269-3060,
USA
Fax: +1-860-486-2981
E-mail: jose.gascon@uconn.edu
shawn.burdette@uconn.edu
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201100216.
2916
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2916–2919

Azobenzene Photoisomerization
Scheme 1. Synthesis of AzoAMP derivatives.
curred in aqueous solutions of different pH (1 Ͻ pH Ͻ 14) tion, the lack of photoisomerization cannot be attributed
or in organic solvents with a broad range of polarities when to structural features. AzoAMP-4 could undergo tauto-
using excitation wavelengths between 300 and 600 nm. The merization by migration of the phenolic hydrogen atom;[13]
small amount of (Z)-AzoAMP-4 formed converts back into however, AzoAMP-5, which is obtained by methylating
the (E) isomer within several seconds. The isomerization AzoAMP-4 (Scheme 1), behaves identically to its phenolic
behavior is similar to that of AzoAMP-1;[5] however, since congener. The photochemistry of AzoAMP-4 and
AzoAMP-4 lacks hydrogen bonds, an alternative explana- AzoAMP-5 appears to be unique since 4-hydroxyazoben-
tion was required.
zenes and 4-methoxyazobenzenes are photoactive.[13,14]
Treatment of AzoAMP-4 with acetic anhydride provides
the acetate, AzoAMP-6 (Scheme 1). In contrast to
AzoAMP-4 and AzoAMP-5, the absorption spectrum of
AzoAMP-6 changes significantly when a 10 μm solution is
irradiated at 483 nm, which indicates (E) Ǟ (Z) isomeriza-
tion (Figure 3). AzoAMP-6 reaches a photostationary state
after ca. 15 min of irradiation. Similar to AzoAMP-4 and
AzoAMP-5, however, the emission of AzoAMP-6 at 77 K
is weak. The experiments suggest that all three new AB de-
rivatives have similar structures at room temperature and in
frozen matrices, so the differences in photoisomerization
stem from changes in the electronic properties of the oxy-
gen-based substituent. The oxygen lone pairs of AzoAMP-
4 and AzoAMP-5 can participate in resonance interactions
with the azo group. When esterified, however, the lone pairs
engage in resonance with the carbonyl group instead of the
AB ring system. The resonance interactions provide the
most reasonable explanation for the differences in the isom-
erization behavior.
Figure 2. Changes in the absorption spectrum of 50 μm AzoAMP-
4 in EtOH/Et2O (1:1). No further decrease in absorption occurs
after 30 min of irradiation at 500 nm, and the AzoAMP-4 spectrum
reverts to the initial state after ca. 10 s in the dark.
The absorption and emission spectra of AzoAMP-4 were
recorded at 77 K in a transparent glass of EtOH/Et2O (1:1).
AzoAMP-4 has no detectable fluorescence at room tem-
perature but is weakly fluorescent (Φ = 0.0001) at 77 K; termined by monitoring the changes in the H NMR spec-
however, its emission intensity is significantly weaker than trum. Growth of new peaks near both the NCH3 and
that of AzoAMP-1 (Φ = 0.003).[5] The vibronic structure NCH2Py resonances after irradiating a 3.1 mm solution of
of the S2ǟS0 transition of AzoAMP-4 remains unresolved, AzoAMP-6 at λmax for 30 min, corresponds to the forma-
whereas that of AzoAMP-1 becomes partially resolved at tion of (Z)-AzoAMP-6. The (E)/(Z) isomer ratio was calcu-
The isomerization quantum yield of AzoAMP-6 was de-
1
1
77 K. The spectroscopic properties of AzoAMP-1 at low lated by integrating the peak areas in the H NMR spec-
temperatures are attributed to structural rigidity caused by trum. Approximately 20% of AzoAMP-6 exists in the (Z)
intramolecular hydrogen bonds. Since these measurements form in the photostationary state compared to AB, which
indicate that AzoAMP-4 does not adopt a rigid conforma- reaches 95% conversion.[2] Steric interactions in AzoAMP-
Eur. J. Org. Chem. 2011, 2916–2919
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2917

H. M. D. Bandara, S. Cawley, J. A. Gascón, S. C. Burdette
SHORT COMMUNICATION
tal difference between the S0-S1 energy surfaces for
AzoAMP-4 and AzoAMP-6. AzoAMP-4 presents a conical
seam in the interval φ1 = 168°–170°, and φ2 = 170°–172°,
whereas AzoAMP-6 does not present any conical intersec-
tion in this region. This suggests that as AzoAMP-4 pro-
ceeds by concerted inversion, with a slight asymmetry in
the angles φ1 and φ2, isomerization becomes frustrated by
an early conical intersection (a “short-circuit”). After
reaching the S0 state, AzoAMP-4 still needs to proceed
5 kcal/mol uphill to the transition state in order to iso-
merize. Simultaneously, the system reaches the S0 state near
a classical turning point with the lowest kinetic energy.
Once in the S0 state, the molecule will return to the (E)
configuration rather than continuing.
Figure 3. Photoisomerization of AzoAMP-6. A 10 μm solution of
AzoAMP-6 in benzene was irradiated at 483 nm, and absorption
spectra were recorded at 1 min intervals. No change in absorbance
was observed after 14 min. After 40 min in the dark, the absorption
returns to its original state.
6 may contribute to the decreased the (E) Ǟ (Z) conversion.
The isomerization experiment was repeated in C6D6,
CDCl3, and CD3OD, and the quantum yields were deter-
mined to be 0.17, 0.18 and 0.20, respectively. Like AB,[2]
AzoAMP-6 photoisomerizes with higher quantum yields in
more polar solvents. Since the (E) Ǟ (Z) isomerization is
not complete, the (E) and (Z) isomer absorptions overlap,
and the (E) isomer absorbs more strongly than the (Z) iso-
mer, it is difficult to determine the quantum yield of the (Z)
Ǟ (E) isomerization. When kept in the dark, AzoAMP-6
returns to the (E) isomer thermally after ca. 40 min.
Computational Approach
A question remains as to what features in the potential
energy surface determine the differential behavior of isom-
erization between AzoAMP-4 and AzoAMP-6. To answer
these questions density functional theory (DFT) calcula-
tions were performed on the concerted inversion pathway.
Previous DFT studies showed that the concerted inversion
pathway was a contributing channel for photoisomeriza-
tion,[5,15] which is consistent with CASSCF studies.[16]
Furthermore, our DFT calculations of the potential energy
surface, along different isomerization paths, predicted that
AzoAMP-1 does not isomerize, in agreement with our
measurements of the transient absorption spectrum.[5]
As concerted inversion proceeds on the S1 energy surface
[after S2ǞS1 relaxation around the (E) position] the system
returns to the ground electronic state through conical inter-
sections between S1 and S0. If such intersection occurs near Figure 4. (A) Potential energy surface showing no differences along
a strict concerted inversion pathway. (B), (C) S1-S0 energy gap as a
the S0 transition state and enough kinetic energy is avail-
function of the angles C–N=N (φ1) and N=N–C (φ2). Isosurface
able, concerted inversion can proceed forward, and the (Z)
values are in kcal/mol. White areas represent conical intersections
isomer is formed. Inspection of the S0-S1 energy surfaces for
in AzoAMP-6 (B) and AzoAMP-4 (C).
AzoAMP-4 and AzoAMP-6 reveals no general qualitative
differences between the two molecules (Figure 4A), which
The photoisomerization efficiency of ABs can be in-
would suggest no difference in the isomerization behavior. fluenced by steric interactions[17,18] or intramolecular hy-
Upon closer inspection, however, small differences are ap- drogen bonds.[5,19] Substituens increase the rate of thermal
parent around the inflection point at φ1 = 170°, φ2 = 170°. isomerization, which results in lower quantum yields and
This observation led us to perform a finer-grid calculation shifts in the energies of both S2ǟS0 and S1ǟS0 transi-
around that point, allowing for slight asymmetries in the tions.[20] AzoAMP-4 and AzoAMP-5 both contain three
change of both angles. Figure 4B and C show a fundamen- strong electron-donating substituents and do not photoiso-
2918
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 2916–2919

Azobenzene Photoisomerization
merize. When the 4-hydroxy group in AzoAMP-4 is con-
verted into an ester group, however, photoactivity is re-
stored. Since the lone pairs of electrons on the oxygen atom
are engaged in resonance with the ester carbonyl group in
AzoAMP-6, their contributions to the AB electronic struc-
ture is significantly mollified.
[1]
[2]
[3]
G. S. Hartley, Nature 1937, 140, 281.
P. Bortolus, S. Monti, J. Phys. Chem. 1979, 83, 648–652.
T. Fujino, S. Y. Arzhantsev, T. Tahara, J. Phys. Chem. A 2001,
105, 8123–8129.
T. Fujino, S. Y. Arzhantsev, T. Tahara, Bull. Chem. Soc. Jpn.
2002, 75, 1031–1040.
H. M. D. Bandara, T. R. Friss, M. M. Enriquez, W. Isley, C.
Incarvito, H. A. Frank, J. Gascon, S. C. Burdette, J. Org.
Chem. 2010, 75, 4817–4827.
[4]
[5]
Conclusions
[6]
[7]
F. Puntoriero, P. Ceroni, V. Balzani, G. Bergamini, F. Voegtle,
J. Am. Chem. Soc. 2007, 129, 10714–10719.
The current experiments suggest that removing lone pairs
from the π-system of AB containing electron-donating
groups attenuates the isomerization behavior. We showed
that this electronic effect leads to differences in the location
of conical intersections between S1 and S0. Whereas in
AzoAMP-6 conical intersection occurs closer to the transi-
tion state, in AzoAMP-6 it occurs much earlier in the inso-
merization path, leading to a frustrated isomerization. Fi-
nally, alternatives to esterification including metal-binding
are being explored in our laboratory as secondary switches
for the AB isomerization.
V. Ferri, M. Elbing, G. Pace, M. D. Dickey, M. Zharnikov, P.
Samori, M. Mayor, M. A. Rampi, Angew. Chem. 2008, 120,
3455; Angew. Chem. Int. Ed. 2008, 47, 3407–3409.
M. R. Banghart, A. Mourot, D. L. Fortin, J. Z. Yao, R. H.
Kramer, D. Trauner, Angew. Chem. 2009, 121, 9261; Angew.
Chem. Int. Ed. 2009, 48, 9097–9101.
T. Muraoka, K. Kinbara, T. Aida, Nature 2006, 440, 512–515.
E. Evangelio, J. Saiz-Poseu, D. Maspoch, K. Wurst, F. Busque,
D. Ruiz-Molina, Eur. J. Inorg. Chem. 2008, 2278–2285.
C. L. Forber, E. C. Kelusky, N. J. Bunce, M. C. Zerner, J. Am.
Chem. Soc. 1985, 107, 5884–5890.
[8]
[9]
[10]
[11]
[12]
A. A. Blevins, G. J. Blanchard, J. Phys. Chem. B 2004, 108,
4962–4968.
[13] E. Sawicki, J. Org. Chem. 1957, 22, 743–745.
[14] G. Gabor, Y. F. Frei, E. Fischer, J. Phys. Chem. 1968, 72, 3266–
3272.
Experimental Section
Computational Details: Ground and excited states were obtained by
using density functional theory (DFT) and time-dependent DFT
with the hybrid functional B3LYP and the split-valence double-zeta
basis set 6-31g**. We tested the theory to reproduce the known
conical intersections of azobenzene, the transition state, and the
(Z)-(E) energy difference, in agreement with wavefunction-based
methods such as CASSCF.[21] Our previous DFT results on azo-
benzene[5] reproduced the conical intersection between the ground
state S0 and the first excited state S1 near the midpoint of the rota-
tional pathway and the existence of an isomerization path from S2
along a concerted inversion pathway, both results in agreement
with CASSCF calculations.[16,21,22] The quantum chemistry pack-
age Gaussian 09[23] was used for all calculations. Electronic excited
states were computed at the ground-state geometry. All ground-
state geometries were obtained by relaxed scans (i.e. full geometry
optimization for all coordinates except for the constraints). For
each set of constrained angles we considered possible rotamers in-
volving the pyridine rings in 4 and 6. The lowest-energy rotamer
was considered for the computation of the potential energy surface
at each point.
[15] C. R. Crecca, A. E. Roitberg, J. Phys. Chem. A 2006, 110,
8188–8203.
[16] E. W.-G. Diau, J. Phys. Chem. A 2004, 108, 950–956.
[17] J. Yoshino, N. Kano, T. Kawashima, Chem. Commun. 2007,
559–561.
[18] C. A. Craig, R. J. Watts, Inorg. Chem. 1989, 28, 309–313.
[19] W. R. Brode, J. H. Gould, G. M. Wyman, J. Am. Chem. Soc.
1952, 74, 4641–4646.
[20] N. Siampiringue, G. Guyot, S. Monti, P. Bortolus, J. Pho-
tochem. 1987, 37, 185–188.
[21] L. Wang, W. Xu, C. Yi, X. Wang, J. Mol. Graphics Modell.
2009, 27, 792–796.
[22] A. Cembran, F. Bernardi, M. Garavelli, L. Gagliardi, G. Or-
landi, J. Am. Chem. Soc. 2004, 126, 3234–3243.
[23] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Starov-
erov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell,
J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M.
Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Ad-
amo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Mar-
tin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador,
J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B.
Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian 09, Re-
vision A.1, Gaussian, Inc., Wallingford, CT, 2009.
Received: February 18, 2011
Supporting Information (see footnote on the first page of this arti-
cle): Experimental procedures, additional spectroscopic data, 1H
and 13C NMR spectroscopic data for all new compounds.
Acknowledgments
J. A. G. acknowledges support from the University of Connecticut,
the Camille and Henry Dreyfus Foundation and the National Sci-
ence Fountation for a CAREER Award (CHE-0847340).
Published Online: April 18, 2011
Eur. J. Org. Chem. 2011, 2916–2919
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
2919