Proof for the concerted inversion mechanism in the trans→ cis Isomerization of azobenzene using hydrogen bonding to induce isomer locking

Article abstract of DOI:10.1021/jo100866m

(Figure Presented) Azobenzene undergoes reversible cis-trans photoisomerization upon irradiation. Substituents often change the isomerization behavior of azobenzene, but not always in a predictive manner. The synthesis and properties of three azobenzene derivatives, AzoAMP-1, -2, and -3, are reported. AzoAMP-1 (2,2′-bis[N-(2-pyridyl)methyl]diaminoazobenzene), which possesses two aminomethylpyridine groups ortho to the azo group, exhibits minimal trans→cis photoisomerization and extremely rapid cis→trans thermal recovery. AzoAMP-1 adopts a planar conformation in the solid state and is much more emissive (φ_fl = 0.003) than azobenzene when frozen in a matrix of 1:1 diethylether/ethanol at 77 K. Two strong intramolecular hydrogen bonds between anilino protons and pyridyl and azo nitrogen atoms are responsible for these unusual properties. Computational data predict AzoAMP-1 should not isomerize following S₂←S₀ excitation because of the presence of an energy barrier in the S₁ state. When potential energy curves are recalculated with methyl groups in place of anilino protons, the barrier to isomerization disappears. The dimethylated analogue AzoAMP-2 was independently synthesized, and the photoisomerization predicted by calculations was confirmed experimentally. AzoAMP-2, when irradiated at 460 nm, photoisomerizes with a quantum yield of 0.19 and has a much slower rate of thermal isomerization back to the trans form compared to that of AzoAMP-1. Its emission intensity at 77 K is comparable to that of azobenzene. Confirmation that the AzoAMP-1 and -2 retain excited state photochemistry analogous to azobenzene was provided by ultrafast transient absorption spectroscopy of both compounds in the visible spectral region. The isomerization of azobenzene occurs via a concerted inversion mechanism where both aryl rings must adopt a collinear arrangement prior to inversion. The hydrogen bonding in AzoAMP-1 prevents both aryl rings from adopting this conformation. To further probe the mechanism of isomerization, AzoAMP-3, which has only one anilinomethylpyridine substituent for hydrogen bonding, was prepared and characterized. AzoAMP-3 does not isomerize and exhibits emission (φ_fl = 0.0008) at 77 K. The hydrogen bonding motif in AzoAMP-1 and AzoAMP-3 provides the first example where inhibiting the concerted inversion pathway in an azobenzene prevents isomerization. These molecules provide important supporting evidence for the spectroscopic and computational studies aimed at elucidating the isomerization mechanism in azobenzene.

Full text of DOI:10.1021/jo100866m

pubs.acs.org/joc
Proof for the Concerted Inversion Mechanism in the transfcis Isomerization
of Azobenzene Using Hydrogen Bonding To Induce Isomer Locking
H. M. Dhammika Bandara,† Tracey R. Friss,† Miriam M. Enriquez,† William Isley,†
Christopher Incarvito,‡ Harry A. Frank,† Jose Gascon,† and Shawn C. Burdette*,†
†Department of Chemistry, University of Connecticut, 55 North Eagleville Road U-3060, Storrs, Connecticut
06269, and ‡Department of Chemistry, Yale University, 225 Prospect Street, P.O. Box 208107, New Haven,
Connecticut 06520-8107
shawn.burdette@uconn.edu
Received May 4, 2010
Azobenzene undergoes reversible cisTtrans photoisomerization upon irradiation. Substituents often
change the isomerization behavior of azobenzene, but not always in a predictive manner. The synthesis
and properties of three azobenzene derivatives, AzoAMP-1, -2, and -3, are reported. AzoAMP-1 (2,20-
bis[N-(2-pyridyl)methyl]diaminoazobenzene), which possesses two aminomethylpyridine groups ortho
to the azo group, exhibits minimal transfcis photoisomerization and extremely rapid cisftrans
thermal recovery. AzoAMP-1 adopts a planar conformation in the solid state and is much more
emissive (Φfl = 0.003) than azobenzene when frozen in a matrix of 1:1 diethylether/ethanol at 77 K.
Two strong intramolecular hydrogen bonds between anilino protons and pyridyl and azo nitrogen
atoms are responsible for these unusual properties. Computational data predict AzoAMP-1 should not
isomerize following S2rS0 excitation because of the presence of an energy barrier in the S1 state. When
potential energy curves are recalculated with methyl groups in place of anilino protons, the barrier to
isomerization disappears. The dimethylated analogue AzoAMP-2 was independently synthesized, and
the photoisomerization predicted by calculations was confirmed experimentally. AzoAMP-2, when
irradiated at 460 nm, photoisomerizes with a quantum yield of 0.19 and has a much slower rate of
thermal isomerization back to the trans form compared to that of AzoAMP-1. Its emission intensity at
77 K is comparable to that of azobenzene. Confirmation that the AzoAMP-1 and -2 retain excited state
photochemistry analogous to azobenzene was provided by ultrafast transient absorption spectroscopy
of both compounds in the visible spectral region. The isomerization of azobenzene occurs via a
concerted inversion mechanism where both aryl rings must adopt a collinear arrangement prior to
inversion. The hydrogen bonding in AzoAMP-1 prevents both aryl rings from adopting this
conformation. To further probe the mechanism of isomerization, AzoAMP-3, which has only one
anilinomethylpyridine substituent for hydrogen bonding, was prepared and characterized. AzoAMP-3
does not isomerize and exhibits emission (Φfl = 0.0008) at 77 K. The hydrogen bonding motif in
AzoAMP-1 and AzoAMP-3 provides the first example where inhibiting the concerted inversion
pathway in an azobenzene prevents isomerization. These molecules provide important supporting
evidence for the spectroscopic and computational studies aimed at elucidating the isomerization
mechanism in azobenzene.
DOI: 10.1021/jo100866m
r
Published on Web 06/14/2010
J. Org. Chem. 2010, 75, 4817–4827 4817
2010 American Chemical Society

Bandara et al.
JOCArticle
Introduction
from the S2 state resulted from an out-of-plane rotation after
scission of the NdN π-bond (Figure 1B).25 The excitation
wavelength dependence of the isomerization quantum yield
was attributed to the different isomerization mechanisms.
The dual isomerization mechanism was discarded when
femtosecond time-resolved spectroscopic studies revealed
that S1rS2 occurs with a quantum yield of ∼1, an efficiency
that precludes isomerization from the S2 state.26 Azoben-
zenes with rotation about the N-N bond restricted by steric
constraints have quantum yields of isomerization indepen-
dent of the excitation wavelength,27,28 which suggests photo-
isomerization from the S1 state dominates regardless of the
initial excited state.
Recent computational studies suggest that excitation of
azobenzene to the S1 state is followed by isomerization
through rotation.29-31 Excitation to the S2 state is followed
by rapid relaxation to the S1 state. The isomerization and
return to the S0 state occurs through a concerted-inversion
mechanism that involves simultaneous distortions of both
N-N-C bonds (Figure 1B).29 The lifetime of the S1 state
generated by the relaxation of the S2 state is 500 fs,26 but the
S1 state prepared by direct S1rS0 excitation has a longer
lifetime of 2.6 ps.32 The S1rS2 crossover creates a vibration-
ally excited S1 state that relaxes rotationally to the trans
isomer resulting in a lower quantum yield of isomerization.
The quantum yield of isomerization also decreases with
increasing vibrational energy in the S1rS0 inversion path-
way. Irradiation of the trans isomer with 436 nm radiation
results in a quantum yield of isomerization of 0.27, which
decreases to 0.21 with 405 nm radiation.24 Both theoretical
and experimental evidence suggest that isomerization always
occurs from the S1 state; however, additional isomerization
channels are opened when the S1 state is accessed by the
relaxation of the S2 state. Theoretical studies have suggested
that such additional isomerization channel must involve a
concerted inversion pathway, where both the C-NdN and
NdN-C angles change simultaneously.29,33
Azobenzene undergoes a reversible transfcis conforma-
tional change upon photoexcitation,1 which has been utilized
as a light-triggered switch in a variety of polymers,2 surface-
modified materials,3 protein probes,4 molecular machines,5 and
metal ion chelators.6-9 The change in geometry upon isomer-
ization orients the molecules to perform a task,10-12 modulates
interactions that change the structure of the bulk material,13-15
changes the spectroscopic properties,16-19 or moves a substi-
tuent that blocks or unblocks activity.4,20-22 Despite the pre-
valence of azobenzene derivatives in a broad spectrum of
chemical applications, predicting the photochemical properties
of azobenzene derivatives remains difficult, and the mechanism
of isomerization continues to be a topic of interest.
The ground-state (S0) absorption spectrum of trans-azo-
benzene has two well-defined bands in the UV-vis region. The
symmetry-forbidden S1(nπ*)rS0 transition appears as a weak
band at ∼450 nm, while the symmetry-allowed S2(ππ*)rS0
transition absorbs at ∼320 nm (Figure 1A). The cis isomer is
thermodynamically unstable, and hence cis to trans thermal
isomerization occurs in the dark; however, thermal isomer-
ization proceeds more slowly than photoisomerization.23
Excitation to both S1 and S2 states leads to the transfcis
transformation, but the S1 state relaxes to the ground state
with a higher quantum yield of isomerization.1,24 Original
investigators of azobenzene photochemistry suggested that
isomerization from the S1 state occurred by in-plane inversion
centered at one of the azo-N atoms, whereas isomerization
(1) Bortolus, P.; Monti, S. J. Phys. Chem. 1979, 83, 648–652.
(2) Puntoriero, F.; Ceroni, P.; Balzani, V.; Bergamini, G.; Vogtle, F. J.
Am. Chem. Soc. 2007, 129, 10714–10719.
(3) Ferri, V.; Elbing, M.; Pace, G.; Dickey, M. D.; Zharnikov, M.;
Samori, P.; Mayor, M.; Rampi, M. A. Angew. Chem., Int. Ed. 2008, 47,
3407–3409.
(4) Banghart, M. R.; Fortin, A. M. D. L.; Yao, J. Z.; Kramer, R. H.;
Trauner, D. Angew. Chem., Int. Ed. 2009, 48, 9097–9101.
(5) Muraoka, T.; Kinbara, K.; Aida, T. Nature 2006, 440, 512–515.
(6) Evangelio, E.; Saiz-Poseu, J.; Maspoch, D.; Wurst, K.; Busque, F.;
Ruiz-Molina, D. Eur. J. Inorg. Chem. 2008, 2278–2285.
(7) Shinkai, S.; Nakaji, T.; Nishida, Y.; Ogawa, T.; Manabe, O. J. Am.
Chem. Soc. 1980, 102, 5860–5865.
Azobenzene exhibits very weak fluorescence emission due
to deactivation of the excited state by isomerization. Azo-
benzene does not obey Kasha’s rule, but at room tempera-
ture the S0rS1 quantum yield of fluorescence is 2.53 ꢀ
(8) Luboch, E.; Wagner-Wysiecka, E.; Poleska-Muchlado, Z.; Kravtsov,
V. C. Tetrahedron 2005, 61, 10738–10747.
(9) Luboch, E.; Wagner-Wysiecka, E.; Rzymowski, T. Tetrahedron 2009,
65, 10671–10679.
10-5 26
An increase in emission intensity can be observed
.
from the S0rS2 state in a frozen matrix when relaxation to
the S1 state is impaired.34 Engineering bonding interactions
can provide conformationally locked azobenzene derivatives
that show strong fluorescence at room temperature.35,36
(10) Cacciapaglia, R.; Di Stefano, S.; Mandolini, L. J. Am. Chem. Soc.
2003, 125, 2224–2227.
(11) Stoll, R. S.; Hecht, S. Org. Lett. 2009, 11, 4790–4793.
(12) Stoll, R. S.; Peters, M. V.; Kuhn, A.; Heiles, S.; Goddard, R.; Buhl,
M.; Thiele, C. M.; Hecht, S. J. Am. Chem. Soc. 2009, 131, 357–367.
(13) Yamada, M.; Kondo, M.; Miyasato, R.; Naka, Y.; Mamiya, J.;
Kinoshita, M.; Shishido, A.; Yu, Y. L.; Barrett, C. J.; Ikeda, T. J. Mater.
Chem. 2009, 19, 60–62.
(14) Zhang, Q.; Bazuin, C. G. Macromolecules 2009, 42, 4775–4786.
(15) Kumar, S. K.; Pennakalathil, J.; Kim, T. H.; Kim, K.; Park, J. K.;
Hong, J. D. Langmuir 2009, 25, 1767–1771.
(16) Khan, A.; Hecht, S. Chem.;Eur. J. 2006, 12, 4764–4774.
(17) Khan, A.; Hecht, S. J. Polym. Sci. A 2006, 44, 1619–1627.
(18) Khan, A.; Kaiser, C.; Hecht, S. Angew. Chem., Int. Ed. 2006, 45,
1878–1881.
(25) Monti, S.; Orlandi, G.; Palmieri, P. Chem. Phys. 1982, 71, 87–99.
(26) Fujino, T.; Arzhantsev, S. Y.; Tahara, T. J. Phys. Chem. A 2001, 105,
8123–8129.
(27) Rau, H.; Lueddecke, E. J. Am. Chem. Soc. 1982, 104, 1616–1620.
(28) Lednev, I. K.; Ye, T.-Q.; Abbott, L. C.; Hester, R. E.; Moore, J. N. J.
Phys. Chem. A 1998, 102, 9161–9166.
(29) Crecca, C. R.; Roitberg, A. E. J. Phys. Chem. A 2006, 110, 8188–
8203.
(19) Mathews, M.; Tamaoki, N. J. Am. Chem. Soc. 2008, 130, 11409–
11416.
(30) Ishikawa, T.; Noro, T.; Shoda, T. J. Chem. Phys. 2001, 115, 7503–
7512.
(20) Banghart, M.; Borges, K.; Isacoff, E.; Trauner, D.; Kramer, R. H.
Nat. Neurosci. 2004, 7, 1381–1386.
(31) Tiago, M. L.; Ismail-Beigi, S.; Louie, S. G. J. Chem. Phys. 2005, 122,
094311.
(21) Gorostiza, P.; Volgraf, M.; Numano, R.; Szobota, S.; Trauner, D.;
Isacoff, E. Y. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10865–10870.
(22) Volgraf, M.; Gorostiza, P.; Numano, R.; Kramer, R. H.; Isacoff,
E. Y.; Trauner, D. Nat. Chem. Biol. 2006, 2, 47–52.
(23) Blevins, A. A.; Blanchard, G. J. J. Phys. Chem. B 2004, 108, 4962–
4968.
(32) Lednev, I. K.; Ye, T. Q.; Matousek, P.; Towrie, M.; Foggi, P.;
Neuwahl, F. V. R.; Umapathy, S.; Hester, R. E.; Moore, J. N. Chem. Phys.
Lett. 1998, 290, 68–74.
(33) Diau, E. W. G. J. Phys. Chem. A 2004, 108, 950–956.
(34) Nepras, M.; Lunak, S., Jr.; Hrdina, R.; Fabian, J. Chem. Phys. Lett.
1989, 159, 366–370.
(24) Zimmerman, G.; Chow, L.-Y.; Paik, U.-J. J. Am. Chem. Soc. 1958,
80, 3528–3531.
(35) Yoshino, J.; Furuta, A.; Kambe, T.; Itoi, H.; Kano, N.; Kawashima,
T.; Ito, Y.; Asashima, M. Chem.;Eur. J. 2010, 16, 5026–5035.
4818 J. Org. Chem. Vol. 75, No. 14, 2010

Bandara et al.
JOCArticle
FIGURE 1. (A) Changes in the absorption spectrum of azobenzene upon irradiation with 320 nm light. The S1(nπ*)rS0 transition appears as
a weak band at ∼440 nm (inset) and the S2(ππ*)rS0 transition absorbs at ∼320 nm. (B) Historic erroneously presumed azobenzene
isomerization pathways upon S1 (bottom) and S2 (top) excitation and the currently accepted concerted inversion isomerization mechanism
(center) following S2 excitation that is supported by time-resolved spectroscopy and theoretical studies.
Several studies have attempted tousesteric congestionaround
the azo group to inhibit isomerization and enhance fluores-
cence emission.37,38 Hydrogen bonding has been reported to
enhance the emission of 2-hydroxyazobenzene and related
derivatives, but not the aniline analogues.39 Metal binding
induces emission enhancement of 2,20dihydroxyazobenzene;40,41
however, these compounds tautomerize readily unlike other
azobenzene derivatives and therefore exhibit different
photochemistry.42
Mg2þ chelator resembling an elongated EDTA motif.43 The
Mg2þ ligand incorporated iminodiacetate ligands on two of
the carbon atoms adjacent to the azo group. While the apo
chelator exhibited the expected reversible transfcis inter-
conversion, no photoisomerization was observed in the
presence of Mg2þ; in addition the parent 2,20-diaminoazo-
benzene (5) isomerized efficiently. Since the presumed Mg2þ
complex was coordinatively saturated, we hypothesized that
eliminating one ligand from each anilino nitrogen atom
would provide a less rigid metal complex that could isomerize.
The 2,20-diaminoazobenzene ligand framework was pre-
pared by the oxidation of o-phenylenediamine.44 Initially
AzoAMP-1 (2,20-bis[N,N0-(2-pyridyl)methyl]diaminoazoben-
zene) was synthesized by a nucleophilic substitution reaction
between 5 and 2-picholylchloride, but the desired product
only was obtained in low yields. The name AzoAMP-1 is
derived from the azobenzene (azo) and 2-aminomethylpyr-
idine (AMP) components of the molecule. Alternatively,
reductive amination of 2-pyridinecarboxaldehyde and 5
provided AzoAMP-1 as a red crystalline solid in 30.1% yield
(Scheme 1A). Only the more thermodynamically stable trans
isomer was observed by 1H NMR spectroscopy.
The λmax of the S2rS0 transition of AzoAMP-1 appears at
490 nm compared to 320 in unsubstituted azobenzene. The
large bathochromic shift of the transition compared to the
parent chromophore indicates a significant delocalization of
the anilino lone pair into the π-system, which is typical of
diaminoazobenzenes.23 To interrogate the photochemistry,
a solution of AzoAMP-1 was irradiated with a 150 W Xe
lamp, and the absorbance was recorded at 10 min intervals
for 4 h. Although no evidence of significant AzoAMP-1
isomerization was observed under these conditions, the
isomerization of 2,2-diaminoazobenzene was reproduced in
analogous experiments.43 To further interrogate AzoAMP-1
isomerization, several organic solvents with a broad range of
polarities were screened, as well as aqueous solutions with
different pH (1 < pH < 14). In addition to the media, the
excitation wavelength was varied between 250 and 600 nm,
but only minimal changes in absorbance (<10% reduction)
On the basis of the light-driven changes in metal ion
binding observed with crown ether-azobenzene dyads, we
envisioned a ligand with light-tunable affinity for metal ions
based on 2,20-diaminoazobenzene. Installing a pyridylmethyl
group on each aniline nitrogen atom would provide a binding
pocket that could accommodate different metal complexes in
each isomeric form. Preliminary modeling studies suggested
that in the trans conformation, the azobenzene would provide
a prearranged structure with a higher affinity for Zn2þ
;
however, canting of the aryl rings in the cis conformer would
prevent simultaneous binding of the guest by both halves of
the ligand. With the objective of developing new applications
for photoisomerizable metal ion chelators, we initiated a
synthetic and spectroscopic study of these ligands. In the
course of our investigation, we discovered some unusual and
unexpected photochemistry of ortho disubstituted derivatives
that has provided insight into underlying aspects of azoben-
zene isomerization as well as the possibility to uncover new
applications for a well-known molecule.
Results and Discussion
Design and Photochemistry of AzoAMP-1. The use of 2,20-
disubstituted azobenzene scaffold was inspired in part by a
(36) Yoshino, J.; Kano, N.; Kawashima, T. Chem. Commun. 2007, 559–561.
(37) Bunce, N. J.; Ferguson, G.; Forber, C. L.; Stachnyk, G. J. J. Org.
Chem. 1987, 52, 394–398.
(38) Han, M.; Ishikawa, D.; Muto, E.; Hara, M. J. Lumin. 2009, 129,
1163–1168.
(39) Nurmukhametov, R. N.; Shigorin, D. N.; Kozlov, Y. I.; Puchkov,
V. A. Opt. Spectrosc. 1961, 11, 327–330.
(40) Diehl, H.; Olsen, R.; Spielholtz, G. I.; Jensen, R. Anal. Chem. 1963,
35, 1144–1154.
(41) Freeman, D. C.; White, C. E. J. Am. Chem. Soc. 1956, 78, 2678–2682.
(42) Dedkov, Y. M.; Kotov, A. V. Russ. Chem. Bull. 1973, 22, 1334–1336.
(43) Momotake, A.; Arai, T. Tetrahedron Lett. 2003, 44, 7277–7280.
(44) Crank, G.; Makin, M. I. H. Tetrahedron Lett. 1979, 20, 2169–2170.
J. Org. Chem. Vol. 75, No. 14, 2010 4819

Bandara et al.
JOCArticle
SCHEME 1. Synthesis of AzoAMP-1 (A), AzoAMP-2 (B), and
AzoAMP-3 (C)
FIGURE 2. Relative emission intensity of azobenzene, AzoAMP-
1, AzoAMP-2, and AzoAMP-3 in a frozen 1:1 ether/ethanol matrix.
Excitation provided at the λmax for each compound.
were observed even after prolonged irradiation. In addition
to minimal formation of the cis isomer, the thermal cisftrans
isomerization appears to occur within seconds, unlike 2,20-
diaminoazobenzene that takes 30 min to recover completely.
To interrogate the electronic structure of AzoAMP-1
further, spectroscopic studies were conducted on AzoAMP-1
at 77 K in a transparent glass of 1:1 EtOH/Et2O. At room
temperature azobenzene exists in a large number of confor-
mations due to rotation around the two N-C bonds, which
results in broad absorption bands. Unlike the parent com-
pound, the disubstituted azobenzene, bis-4,40-diethylami-
noazobenzene (BDAAB), exhibits a partially resolved vibro-
nic structure at room temperature that becomes clearer at
77 K.34 Delocalization of the anilino lone pairs into the
π-system of BDAAB causes the molecule to adopt a rigid
planar conformation by preventing rotation around the
N-C bonds. The vibronic structure of the AzoAMP-1
S2rS0 transition also becomes partially resolved when
trapped in a frozen matrix, which suggests that AzoAMP-1
also exists in a rigid planar conformation.
Although AzoAMP-1 exhibits no detectable fluorescence
emission at room temperature, it becomes emissive when
frozen in an EtOH/Et2O glass. The fluorescence exhibits a
bimodal emission profile with an emission maximum at
566 nm and a slightly weaker peak at 602 nm (Figure 2).
At room temperature, azobenzene fluorescence intensity
from the S2 and S1 states is nearly identical.26,45 The emission
bands of azobenzene are separated significantly as is observed
for the two absorption peaks. Since the two absorption
bands of AzoAMP-1 are coincidental, the emission bands
FIGURE 3. (A) ORTEP diagram of AzoAMP-1 showing 50% ther-
mal ellipsoids and selected atom labels. (B) Side view of AzoAMP-1
showing the coplanarity of all non-hydrogen atoms.
might correspond to overlapping emission from the S2 and S1
states; however, additional experiments will be required to
confirm this hypothesis. The emission of AzoAMP-1 is weak
(Φ = 0.003) but considerably higher than the quantum yield
of emission of azobenzene under the same conditions (Φ =
0.0001). The increased quantum yield of AzoAMP-1 sug-
gests the compound adopts a more rigid conformation
than azobenzene at low temperature, which prevents deacti-
vation of the excited state by isomerization or vibra-
tion. Use of dative covalent bonding produces more emis-
sive azobenzene derivatives;35,36 however, weaker hydrogen
bonding interactions have not been examined in this
capacity.
X-ray quality crystals of AzoAMP-1 were obtained by
diffusing Et2O into CH3OH/CH3CN; however, AzoAMP-1
forms large single crystals under a variety of conditions.
AzoAMP-1 adopts the trans conformation in the solid state
(Figure 3). The geometry of the two anilino nitrogen atoms is
trigonal planar despite formal sp3 hybridization. Hydrogen
bonding of the anilino protons to the azo and the pyridinium
nitrogen atoms enforces the planar geometry; however, π-
stacking in the crystal lattice reinforces this planarity.
AzoAMP-1 exhibits a 5-6 chelate ring motif, where the
aniline hydrogen atom coordinates to the distal azo nitrogen
atom in AzoAMP-1 forming a six-membered ring instead of
the proximal one that would result in a five-membered ring.
The preference for adopting this structure may reflect the bite
angle requirements for interaction with a hydrogen atom. The
hydrogen bonds also persist in solution as indicated by the
(45) Fujino, T.; Arzhantsev, S. Y.; Tahara, T. Bull. Chem. Soc. Jpn. 2002,
75, 1031–1040.
4820 J. Org. Chem. Vol. 75, No. 14, 2010

Bandara et al.
JOCArticle
1
downfield signal at ∼9 ppm in the H NMR spectrum in
CDCl3. The anilino protons of 2,20-diaminoazobenzene
appear at ∼5.50 ppm, which indicates the significant inter-
action of the pyridyl hydrogen bond acceptors to strong
intramolecular hydrogen bonds in AzoAMP-1. Just as elec-
tronics and steric requirements contribute to the isomeriza-
tion properties of azobenzene derivatives, the hydrogen
bonds enhance the thermodynamic stability of trans-
AzoAMP-1 that provides insight into the fluorescence and
isomerization behavior.
To quantify the effect of the hydrogen bonding interaction
on the isomerization, we performed computational studies
on AzoAMP-1 to probe the energetics of various excited
states involved in the rotational, inversion, and concerted
inversion isomerization pathways. Ground and excited
states were obtained using density functional theory (DFT)
and time-dependent DFT with the hybrid functional B3LYP
and the split-valence double-ζ basis set 6-31 g**. A similar
approach was recently used at the same level of theory to
study the photoisomerization of azobenzene,29 which show
consistent results with respect to other wave-based ab initio
calculations.31,33,46-51
For the rotation pathway, one NdN-C angle was fixed at
120° while the other angle was scanned from 0 to 180° at 10°
intervals. For the inversion pathway one of the NdN-C
angles was fixed at its minimum energy value in the trans
conformation (∼115°), while the dihedral C-NdN-C angle
was varied from 100° to 250° at 10° intervals (Figure 4B). In
the concerted inversion pathway both NdN-C angles were
distorted simultaneously from 100° to 180°. The calculations
reveal the absence of a probable isomerization pathway in
AzoAMP-1 for the inversion and rotation pathways, which
is similar to the results obtained with azobenzene.
Potential energy curves calculated previously and repro-
duced for this study indicate that azobenzene isomerization
occurs through a concerted inversion pathway when the
molecule is excited into the S2 state. Clear conical intersec-
tions are observed between the S1rS2 and the S0rS1 states
allowing the excited molecule to relax into the S1 state with
excess potential energy and return to the ground state
(Figure 4A). The question arises as to whether the concerted
inversion pathway is also an isomerization pathway in
AzoAMP-1. The calculations on AzoAMP-1 reveal a possi-
ble conical intersection for the concerted inversion pathway
although a gap exists (Figure 4B). In contrast to azobenzene
however, the hydrogen bonding in AzoAMP-1 prevents the
two required NdN-C distortions that bring the two aryl
rings into the same plane prior to the inversion. Such a strong
restraint affects rotation in both the S0 and S1, making the
energy basin more pronounced. Thus, upon S1rS2 conver-
sion around 120° the system will not have enough kinetic
energy around the NdN bond to reach the potential near the
FIGURE 4. Calculated energy surfaces for the concerted inversion
isomerization pathway for azobenzene (A) and AzoAMP-1 (B).
S0rS1 intersection. Since the coplanarization prerequisite
cannot be achieved, the computation predicts that AzoAMP-1
should not isomerize, which is confirmed by the spectro-
scopic studies.
Design and Photochemistry of AzoAMP-2. Because the
intramolecular hydrogen bonds in AzoAMP-1 appear to be
responsible for the unusual photochemistry, we rationalized
that replacing the two anilino protons with methyl groups
would provide an active azobenzene. The analogous com-
puted potential energy curves for the rotation, inversion, and
concerted inversion isomerization pathways of AzoAMP-2
are similar to those of azobenzene; therefore, we predicted
that AzoAMP-2 would photoisomerize by the concerted
inversion when excited into the S2 state (Figure 5).
Attempts to prepare AzoAMP-2 using AzoAMP-1 as a
starting material were unsuccessful. AzoAMP-1 is highly
unreactive with formaldehyde using reductive amination
conditions. Strong alkylating reagents such as CH3I failed
to give the desired product by nucleophilic substitution even
when used in conjunction with a strong base such as lithium
diisopropylamine, so an alternative one-step preparation
based on reduction of nitrobenzene derivatives was explored.52
A nucleophilic aromatic substitution reaction between the
amine 7 and 2-bromonitrobenzene (6) gave the nitroaniline 8
with a satisfactory yield. The nitroaniline was reduced with
(46) Biswas, N.; Umapathy, S. J. Phys. Chem. A. 1997, 101, 5555–5566.
(47) Cattaneo, P.; Persico, M. Phys. Chem. Chem. Phys. 1999, 1, 4739–
4743.
(48) Cembran, A.; Bernardi, F.; Garavelli, M.; Gagliardi, L.; Orlandi, G.
J. Am. Chem. Soc. 2004, 126, 3234–3243.
(49) Fliegl, H.; Kohn, A.; Hattig, C.; Ahlrichs, R. J. Am. Chem. Soc. 2003,
125, 9821–9827.
(50) Gagliardi, L.; Orlandi, G.; Bernardi, F.; Cembran, A.; Garavelli, M.
Theor. Chem. Acc. 2004, 111, 363–372.
(51) Ootani, Y.; Satoh, K.; Nakayama, A.; Noro, T.; Taketsugu, T. J.
Chem. Phys. 2009, 131, 194306.
(52) Tanino, T.; Yoshikawa, S.; Ujike, T.; Nagahama, D.; Moriwaki, K.;
Takahashi, T.; Kotani, Y.; Nakano, H.; Shirota, Y. J. Mater. Chem. 2007,
17, 4953–4963.
J. Org. Chem. Vol. 75, No. 14, 2010 4821

Bandara et al.
JOCArticle
FIGURE 5. Calculated energy surfaces for the concerted inversion
isomerization pathway for AzoAMP-2.
FIGURE 7. Transient absorption spectra taken at different time delays
after excitation at 490 nm for AzoAMP-1 (A) and AzoAMP-2 (B).
absorption of trans and cis AzoAMP-2 overlap and the
extinction coefficient of the S2rS0 transition of the trans
isomer is large, it is not possible to selectively excite the
cisftrans isomerization with light because the initial con-
version is not complete. After reaching the photostationary
state, AzoAMP-2 returns to the trans isomer by thermal
isomerization but requires over 1 h to reach equilibrium.
No detectable fluorescence of AzoAMP-2 was observed at
room temperature, and the emission at 77 K is significantly
weaker than the intensity measured for AzoAMP-1 but
comparable to that of azobenzene. This observation suggests
that AzoAMP-2 adopts a less rigid conformation than
AzoAMP-1 at low temperatures. The lack of emission of
AzoAMP-2 also indicates that delocalization of the aniline
lone pair into the π-system does not contribute to the
fluorescence properties as in BDAAB. The composite results
confirm the hypothesis that hydrogen bonding in AzoAMP-1
induces a structural rigidity that provides an emissive azo-
benzene derivative at low temperatures.
Transient Absorption Studies of AzoAMP-1 and AzoAMP-2.
The inclusion of electron-donating aniline substituents in
AzoAMP-1 and AzoAMP-2 alters the energies of the fron-
tier molecular orbitals involved in isomerization. To confirm
that the AzoAMP compound retains excited state photo-
chemistry analogous to that of azobenzene, ultrafast tran-
sient absorption spectra of both compounds in the visible
spectral region were recorded after laser excitation at 490 nm
into their S2rS0 bands. While the spectral traces for both
molecules are weak (Figure 7), the ground state absorption
band below ∼525 nm clearly bleaches immediately and an
SnrS1 feature that maximizes in 600-700 fs appears. This is
similar to the transient absorption behavior of azobenzene,
which exhibits rapid decay to the S1 state following S2rS0
excitation. Like the steady-state S2rS0 absorption spec-
trum, the transient absorption spectrum of AzoAMP-1 is
sharper and more vibronically resolved than that of
AzoAMP-2, which suggests less conformational disorder in
both the S0 and S1 states for AzoAMP-1 than for AzoAMP-2.
The data imply that conformational disorder is induced in
the extended π-electron chain by the introduction of the
methylamino substituent in AzoAMP-2.
FIGURE 6. Spectroscopic changes upon irradiation of AzoAMP-2
at 451 nm. Isomerization efficiency can be measured by 1H NMR.
lithium aluminum hydride to give AzoAMP-2 in moderate
yields (Scheme 1B). After purification by column chroma-
tography, only the trans isomer of AzoAMP-2 was observed
by 1H NMR spectroscopy.
The S2rS0 transition of AzoAMP-2 shows a large batho-
chromic shift and overlaps with the S1rS0 transition analogous
to AzoAMP-1. Unlike AzoAMP-1, exposure of AzoAMP-2
at λmax 460 nm results in changes in the intensity, which
indicates transfcis photoisomerization (Figure 6). TLC
analysis confirms no photodegradation of AzoAMP-2, and
a photostationary state was reached after 15 min of irradia-
tion. The quantum yield of isomerization of AzoAMP-2 was
determined by monitoring the changes in the 1H NMR
spectrum after irradiating at 460 nm for 30 min. Growth of
new peaks was observed near both the NCH3 and NCH2Py
resonances, which correspond to cis-AzoAMP-2. These were
integrated to calculate the trans:cis isomer ratio. The quan-
tum yield of isomerization of AzoAMP-2 was determined to
be 0.19 using ferrioxalate actinometry to determine the
intensity of the radiation source, which is lower than that
of azobenzene under similar conditions. Infrared spectros-
copy, which can measure the trans:cis isomer ratio of azo-
benzene, cannot be used with AzoAMP-2 because of overlapping
absorption of the pyridyl groups. In the photostationary
state approximately 20% of AzoAMP-2 exists in the cis form
compared to azobenzene that exhibits 95% conversion.
Steric interactions between the anilino substituents presum-
ably contribute to the thermodynamic instability the cis
isomer of AzoAMP-2 decreasing the conversion. Since the
4822 J. Org. Chem. Vol. 75, No. 14, 2010

Bandara et al.
JOCArticle
AzoAMP-2. These represent the S1 lifetimes of the mole-
cules. A third and final EADS component (green lines in
Figure 8) is needed for a good fit to the data sets, and these
have time constants of 35 and 50 ps for AzoAMP-1 and
AzoAMP-2, respectively. These slower components have
been assigned by studies on other azobenzenes to vibronic
relaxation in the ground state.
AzoAMP-3 and Implications in the Mechanism of Azoben-
zene Isomerization. While the use of hydrogen bonding alone
seems unlikely to provide highly fluorescent azobenzenes at
room temperature that compete with the best examples,35,36
the observations suggest that ortho-substituted azobenzene
derivatives can be useful probes of the isomerization me-
chanisms. To further probe for the concerted inversion
isomerization mechanism, a monosubsituted version of
AzoAMP-1 was prepared.
Reductive amination of 2-pyridinecarboxaldehyde in the
presence of 1 equiv of 5 yielded AzoAMP-3 (Scheme 1C).
The signal of the single anilino hydrogen atom appears at δ
9.19 ppm in the 1H NMR spectrum, whereas the other two
anilino protons appear significantly upfield at δ 5.45 ppm.
Isomerization behavior of AzoAMP-3 is very similar to that
of AzoAMP-1. Very minimal transfcis conversion is ob-
served upon irradiation and thermal cisftrans isomerization
occurs rapidly. This behavior indicates that a hydrogen
bonding interaction on one side of the azobenzene is suffi-
cient to prevent isomerization. Both AzoAMP-1 and
AzoAMP-3 exhibit some minimal formation of the cis iso-
mer upon irradiation, but this may result from direct S1
excitation, which leads to isomerization via a different
mechanism.48 Since the S1 and S2 excitation bands overlap
in the AzoAMP compounds and the extinction coefficient
for S1 excitation is very small, it is difficult to probe this
process by conventional methods. The emission intensity of
AzoAMP-3 (Φ = 0.0008) at 77 K lies between those of
AzoAMP-1 and AzoAMP-2.
FIGURE 8. Global fitting results (EADS) of transient absorption
spectra of AzoAMP-1 (A) and AzoAMP-2 (B).
The composite results from AzoAMP-1, -2, and -3 provide
the first systematic demonstration that the concerted inver-
sion provides the best mechanistic explanation for azoben-
zene isomerization. Like azobenzene, AzoAMP-2 isomerizes
because both aryl groups can adopt the collinear arrange-
ment required (Scheme 2). The modest conversion to the cis
isomer of AzoAMP-2 compared to azobenzene may reflect
the steric restrictions of the ortho substituents, which might
make the collinear C-NdN-C arrangement more difficult
to achieve and the cis far more sterically congested. In
contrast, the hydrogen bonding interactions in AzoAMP-1
and AzoAMP-3 prevent isomerization.
In AzoAMP-1, hydrogen bonding stabilizes both C-
NdN bond angles at ∼120°, which would not preclude
isomerization by rotation after πfπ* (S2). The time-re-
solved absorption studies confirm that S2 excitation is fol-
lowed by rapid and efficient relaxation to the S1 state;
therefore the lack of isomerization in AzoAMP-1 is most
consistent with inhibition of the concerted inversion me-
chanism. By preventing the aryl groups from adopting the
prerequisite collinear arrangement, the hydrogen bonding
prevents isomerization. Further evidence for the concerted
inversion mechanism can be derived from the photoisome-
rization experiments with AzoAMP-3. Since 2,20-diaminoa-
zobenzene isomerizes, weak hydrogen bonding between an
azo group and unsubstituted aniline groups does not prevent
FIGURE 9. Kinetic traces probed at 589 nm for AzoAMP-1 (A)
and at 579 nm for AzoAMP-2 (B), with fits obtained from global
fitting.
The entire spectral and temporal transient absorption data
sets were fit by a global fitting analysis model that assumes a
sequential decay of the excited states formed after photo-
excitation into S2 (Figure 8). For both molecules, three
evolution associated decay spectra (EADS) kinetic compo-
nents were needed to fit the data (Figure 9). The first EADS
component (black lines in Figure 8) decays in 100 fs for
AzoAMP-1 and 55 fs for AzoAMP-2. These lifetimes are
shorter than the temporal resolution (∼125 fs) of the spectro-
meter system; nevertheless, the lineshapes in the region below
580 nm display a combination of ground state bleaching and
pronounced negative bands associated with Raman scatter-
ing from the solvent system. The broad positive band of the
first kinetic component sharpens and shifts to shorter wave-
length as the line shape evolves into the second EADS
component (red lines in Figure 8). This evolution may be
attributable to ultrafast vibronic relaxation in the S1 state.
The second EADS profiles are highly characteristic of SnrS1
transient spectral bandshapes of π-electron conjugated mo-
lecules and decay in 3.5 ps for AzoAMP-1 and 2 ps for
J. Org. Chem. Vol. 75, No. 14, 2010 4823

Bandara et al.
JOCArticle
SCHEME 2. Concerted Inversion Mechanism in Azobenzene Isomerization
distortions of the C-NdN bond angles. In AzoAMP-3,
however, the addition of a single methylpyridine group pre-
vents isomerization. Since only the C-NdN bond adjacent
to the AMP group is conformationally restricted, isomeriza-
tion could occur by inversion of the other aryl group. The
lack of isomerization provides experimental confirmation
that both aryl groups must move in a concerted process for
the inversion to proceed. These systematic steady-state ex-
periments provide strong supporting evidence for the con-
certed inversion isomerization mechanism, which could only
be probed by time-resolved optical spectroscopy and theo-
retical calculations with unmodified azobenzene.
and AzoAMP-3 more emissive than azobenzene at low
temperatures. Theoretical calculations demonstrate that
these hydrogen bonds prevent isomerization by creating
energetic barriers larger than the available kinetic energy
upon S1rS2 conversion. The energy barrier disappears upon
elimination of the hydrogen bonds in AzoAMP-2, and the
photoisomerization predicted by calculations was confirmed
experimentally.
To date, azobenzene derivatives have mostly been exp-
loited for applications that utilize the transfcis photoisome-
rization. With the exception of acting as a fluorescence
quencher, fewer applications utilize the other optical proper-
ties of azobenzenes. Conjugated polymers containing dipho-
sphene units, the phosphorus analogue of the azo group,
have been studied because of their unique spectroscopic
properties.53,54 Using hydrogen bonding to prevent isomer-
ization, new applications of azobenzene derivatives can be
explored, such as the development of conjugated polymers
with unique properties. Efforts to further delineate the
photochemistry of azobenzene derivatives and develop new
applications for conformationally restricted azobenzenes are
ongoing.
Conclusions
Incorporation of azobenzene into various biomolecules,
polymers, and ligands requires attaching at least one sub-
stituent to the phenyl rings. Substitution often changes the
electronic properties of the azobenzene and consequently the
transfcis isomerization behavior. Substituents may also be
responsible for higher rates of thermal cisftrans isomeriza-
tion, which is undesirable in most applications. Despite
numerous attempts, a predictive model that relates the
electronic nature, position, and number of substituents on
azobenzene with the isomerization has not been completely
elucidated. The results described above are the first steps in a
more extensive effort to study these structure-photochem-
istry relationships.
Experimental Section
General Synthetic Procedures. All reagents were purchased
from Acros at the highest commercial quality and used without
further purification. N-Methyl-2-(aminomethyl)pyridine (7)55
Substitution may also provide interesting photophysical
behavior, however. Strong intramolecular hydrogen bonds
in AzoAMP-1 and AzoAMP-3 introduce a barrier to photo-
isomerization. These hydrogen bonds force AzoAMP-1 to
adopt a rigid planar structure, which also makes AzoAMP-1
(53) Smith, R. C.; Protasiewicz, J. D. Eur. J. Inorg. Chem. 2004, 998–
1006.
(54) Smith, R. C.; Protasiewicz, J. D. J. Am. Chem. Soc. 2004, 126, 2268–
2269.
(55) Seitz, M.; Kaiser, A.; Tereshchenko, A.; Geiger, C.; Uematsu, Y.;
Reiser, O. Tetrahedron 2006, 62, 9973–9980.
4824 J. Org. Chem. Vol. 75, No. 14, 2010

Bandara et al.
JOCArticle
and 2,20-diaminoazobenzene (5)56 were synthesized according to
literature procedures. Dichloromethane (CH2Cl2) and tetrahy-
drofuran (THF) were sparged with argon and dried by passage
through a Seca solvent purification system. All chromatography
and TLC were performed on silica (230-400 mesh) from Sili-
cycle. TLCs were developed with mixtures of EtOAc/MeOH or
EtOAc/hexanes and were visualized with 254 and 365 nm light.
1H and 13C NMR spectra were recorded using a 400 MHz NMR
instrument, and chemical shifts are reported in ppm on the δ
scale relative to tetramethylsilane. IR spectra were recorded on a
FT-IR instrument, and the samples were prepared as KBr pellets
or thin films on KBr plates. High resolution mass spectra were
recorded using a electrospray mass spectrometer operating in
positive ion mode.
2,20-Bis[N,N0-(2-pyridyl)methyl]diaminoazobenzene (AzoAMP-
1, 1). The azobenzene 5 (320 mg, 1.5 mmol) and 2-pyridinecar-
boxaldehyde (0.29 mL, 3.0 mmol) were combined in CH2Cl2
(70 mL), and NaBH(OAc)3 (767 mg, 3.6 mmol) was added. The
mixture was stirred at 23 °C for 24 h. Water (40 mL) was added,
and the product was extracted into CH2Cl2 (3 ꢀ 40 mL). The
combined organic layers were dried over MgSO4, and the solvent
was removed. Flash chromatography on silica (24:1 EtOAc/
MeOH) followed by recrystallization (CH2Cl2/Et2O) yielded 1
as red flaky crystals (178 mg, 30.1%). TLC Rf = 0.45 (silica, 49:1
EtOAc:MeOH). Mp = 163-165 °C. 1H NMR (400 MHz,
CDCl3) δ 9.03 (s, 2 H), 8.61 (d, J = 4.7 Hz, 2 H), 7.87 (dd,
J = 7.8, 1.4 Hz, 2 H), 7.65 (dt, J = 1.6, 7.8 Hz, 2 H), 7.38 (d, J =
7.8 Hz, 2 H), 7.26-7.18 (m, 4H), 6.81 (t, J = 7.2 Hz, 2 H), 6.76
(d, J = 8.2 Hz, 2 H), 4.71 (s, 4 H). 13C NMR (100 MHz, CDCl3)
δ 158.3, 149.3, 143.0, 137.3, 137.1, 131.7, 128.0, 122.4, 121.6,
116.6, 112.4, 48.7. IR (neat, cm-1) 3207.9, 3066.7, 3020.8, 2857.9,
2840.7, 1606.5, 1594.5, 1497.6, 1447.7, 1439.2, 1415.9, 1309.9,
1285.9, 1205.4, 1153.2, 1081.1, 1043.9, 993.7, 843.4, 737.5, 671.3.
HRMS (þESI): Calcd for C24H22N6Hþ, 395.1984; Found,
395.1968.
Methyl-(2-nitro-phenyl)-pyridin-2-ylmethyl-amine (8). The
amine 7 (2.00 g, 16.4 mmol) and 2-bromonitrobenzene (3.30 g,
18.0 mmol, 6) were combined in dry 1,4-dioxane (80 mL) and
K2CO3 (8.48 g, 53.3 mmol) was added. The mixture was refluxed
at 100 °C for 72 h. Water (20 mL) was added and the product
was extracted into CH2Cl2 (3 ꢀ 40 mL). The combined organic
layers were dried over MgSO4 and the solvent was removed.
Flash chromatography on silica (EtOAc) yielded 7 as a yellow-
orange oil (721 mg, 18.0%). TLC Rf = 0.51 (silica, EtOAc). 1H
NMR (400 MHz, CDCl3) δ 8.57 (d, J = 4.8 Hz, 1 H), 7.77 (dd,
J = 8.1, 1.0 Hz, 1 H), 7.71 (dt, J = 7.7, 1.3 Hz, 1 H), 7.44 (dd,
J = 7.9, 0.6 Hz, 1 H), 7.39 (t, J = 7.3 Hz, 1 H), 7.22 (dd, J = 7.0,
5.1 Hz, 1 H), 7.14 (d, J = 8.4 Hz, 1 H), 6.93 (t, J = 7.4 Hz, 1 H),
4.56 (s, 2 H), 2.88 (s, 3 H). 13C NMR (100 MHz, CDCl3) δ 157.7,
149.2, 145.8, 141.3, 137.4, 133.4, 126.5, 122.6, 122.2, 120.4,
119.9, 60.6, 41.5. IR (neat, cm-1) 3065.1, 3008.2, 2926.1,
1604.2, 1589.7, 1566.8, 1433.0, 1342.1, 1290.1, 1231.9, 1214.8,
1193.1, 1114.1, 992.7, 939.1, 852.4, 741.6, 708.3. HRMS (þESI):
calcd for C13H13N3O2Hþ, 244.1086; found, 244.1119.
13C NMR (100 MHz, CDCl3) δ 160.1, 150.1, 149.3, 143.7, 136.7,
131.2, 122.0, 121.9, 119.8, 118.1, 117.7, 64.3, 40.9. IR (neat,
cm-1) 3058.9, 2946.4, 2869.9, 2802.9, 1588.7, 1568.1, 1484.5,
1431.4, 1358.0, 1304.0, 1184.3, 1159.8, 1100.5, 1046.2, 991.8,
936.6, 747.3. HRMS (þESI): calcd for C26H26N6Naþ, 445.2117;
found, 445.2103.
2-Amino-20-[N-(2-pyridyl)methylamino]azobenzene (AzoAMP-3,
3). The diamine 5 (180 mg, 0.85 mmol), 2-pyridinecarboxalde-
hyde (81 μL, 0.85 mmol), and NaBH(OAc)3 (360 mg, 1.7 mmol)
were combined in CH2Cl2 (20 mL). The reaction mixture was
stirred at 23 °C for 18 h. Water (20 mL) was added, and the
product was extracted into CH2Cl2 (3 ꢀ 40 mL). The combined
organic layers were dried over MgSO4, and the solvent was
removed. TLC showed a mixture of 2,20-diaminoazobenzene,
AzoAMP-1, and AzoAMP-3. Addition of 6:4 EtOAc/hexanes
caused AzoAMP-1 to precipitate. Flash chromatography on
silica (6:4 EtOAc/hexanes) yielded AzoAMP-3 as a viscous red
oil (11 mg, 4.0%). TLC Rf = 0.62 (silica, 19:1 EtOAc/MeOH).
1H NMR (400 MHz, CDCl3) δ 9.19 (s, 1 H), 8.64 (d, J = 4.3 Hz,
1 H), 7.82 (dt, J = 1.5, 8.2 Hz, 2 H), 7.68 (dt, J = 1.7, 7.7 Hz,
1 H), 7.36 (d, J = 7.7 Hz, 1 H), 7.29-7.17 (m, 3 H), 6.86-6.79
(m, 4 H), 5.45 (s, 2 H), 4.69 (s, 2 H). 13C NMR (100 MHz,
CDCl3) δ 157.7, 149.2, 143.6, 142.7, 138.2, 137.2, 137.0, 131.9,
131.2, 129.6, 123.4, 122.5, 121.8, 117.8, 117.2, 116.4, 112.3, 48.5.
IR (neat, cm-1) 3456.4, 3342.9, 3064.5, 2921.8, 2849.4, 1607.1,
1561.2, 1507.7, 1483.7, 1448.6, 1423.5, 1387.5, 1330.8, 1260.6,
1234.4, 1211.4, 1153.6, 1082.2, 1048.6, 908.2, 798.5, 758.2, 735.8,
579.4. HRMS (þESI): calcd for C18H17N5Naþ, 326.1382;
found, 326.1354.
Crystals of AzoAMP-1. A 15 mg of sample of AzoAMP-1 was
dissolved in the minimum amount of boiling acetonitrile. The
resulting solution was allowed to stand at 23 °C for 30 min and
placed in an Et2O diffusion chamber where single crystals
gradually formed from solution.
Collection and Reduction of X-ray Data. Structural analysis
was carried out in the X-ray Crystallographic Facility at Yale
University. Crystals were covered with oil and an orange rod
crystal of N6C24H22 with approximate dimensions of 0.50 mm ꢀ
0.25 mm ꢀ 0.20 mm was mounted on a glass fiber at room
temperature and transferred to a Rigaku RAXIS RAPID
imaging plate area detector with graphite monochromated Cu
˚
KR radiation (λ = 1.54187 A) controlled by a PC running the
Rigaku CrystalClear software package.57 The data were col-
lected at a temperature of -50 ( 1 °C to a maximum 2θ value
of 55°. The data were corrected for Lorentz and polarization
effects. The structure was solved by direct methods58 and
expanded using Fourier techniques.59 The space group was
determined by examining systematic absences and confirmed
by the successful solution and refinement of the structure. The
non-hydrogen atoms were refined anisotropically. Hydrogen
atoms were refined using the riding model. All calculations were
performed using the CrystalStructure60 crystallographic soft-
ware package except for refinement, which was performed using
SHELXL-97.58 Relevant crystallographic information is sum-
marized in Tables 1 and 2, and the 50% thermal ellipsoid plot is
shown in Figure 3A.
2,20-Bis[N,N0-methyl-N,N0-(2-pyridyl)methyl]aminoazobenzene
(AzoAMP-2, 2). The amine 7 (400 mg, 1.64 mmol) was dissolved
in THF (80 mL) and treated with LiAlH4 (125 mg, 3.28 mmol).
The mixture was stirred at 23 °C for 2 h. Water (20 mL) was
added, and the product was extracted into CH2Cl2 (3 ꢀ 40 mL).
The combined organic layers were dried over MgSO4, and the
solvent was removed. Flash chromatography on silica (48:2
EtOAc/MeOH) yielded 2 as a red oil (43.4 mg, 12.5%). TLC
General Spectroscopic Methods. All solutions were prepared
with spectrophotometric grade solvents. Graphs were manipu-
lated and equations calculated by using Kaleidagraph 4.0.
(57) CrystalClear and CrystalStructure; Rigaku/MSC: The Woodlands,
TX, 2005.
(58) Sheldrick, G. M. SHELX97, Programs for Crystal Structure Analysis;
1
Rf = 0.44 (silica, 48:2 EtOAc/MeOH). H NMR (400 MHz,
CDCl3) δ 8.56 (d, J = 4.4 Hz, 2 H), 7.62 (dt, J = 7.8, 1.6 Hz,
2 H), 7.53 (d, J = 7.7 Hz, 2 H), 7.24-7.15 (m, 4 H), 7.06-7.01
(m, 4 H), 6.59 (dt, J = 7.7, 1.0 Hz, 2 H), 4.76 (s, 4H), 3.06 (s, 6H).
€
€
€
Universitat Gottingen, Gottingen, Germany, 1997.
(59) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.; de
Gelder, R.; Israel, R.; Smits, J. M. M. DIRDIF99; University of Nijmegen:
The Netherlands, 1999.
(60) CrystalStructure 3.8: Crystal Structure Analysis Package; Rigaku
and Rigaku Americas: The Woodlands, TX, 2007.
(56) Crank, G.; Makin, M. I. H. Aust. J. Chem. 1984, 37, 845–55.
J. Org. Chem. Vol. 75, No. 14, 2010 4825

Bandara et al.
JOCArticle
TABLE 1. Crystallographic Parameters for AzoAMP-1 (1)
in the dark, and absorbance was recorded at 1 min intervals to
monitor the cis to trans thermal isomerization.
formula
formula wt.
space group
N6C24H22
394.48
P21/c
10.1401(11)
13.6843(15)
7.3008(8)
95.328(2)
Quantum Yield of Photoisomerization of AzoAMP-2. Inten-
sity of Source. To determine the intensity of the radiation source,
a 1.0-mL aliquot of a 10 mM potassium ferrioxalate solution
was placed in an NMR tube and irradiated with 460 nm
wavelength radiation for 5 min. This results in the reduction
of Fe(III) oxalate to Fe(II) oxalate. The irradiated solution was
combined with 15 mg of ferrozine (3 equiv), resulting in the
formation of a reddish-purple solution containing [Fe(ferro-
zine)3]2þ, which has a molar absorption coefficient of 27,900
cm-1 M-1 at 563 nm. A 100-μL aliquot of the resulting solution
was diluted by a factor of 30, and its absorbance was measured
at 563 nm. The concentration of Fe(II) produced by the re-
duction of Fe(III) oxalate is given by eq 1 where ε563 = 27,900
˚
a, A
˚
b, A
˚
c, A
β, deg
3
˚
V, A
1008.68(19)
2
1.299
0.808
223
Z
F
calcd (g cm-3
)
absorp coeff (cm-1
temp, K
)
total no. data
no. unique data
obs. dataa
10199
2301
2301
cm-1 M-1
.
no. parameters
R, %b
141
0.0490
0.1294
0.20, -0.17
wR2, %c
max/min peaks, e/A
A563 ꢀ 30
˚
½FeðIIÞꢁ ¼
ð1Þ
ε563
aObservation criterion: I > 2σ(I). bR = Σ Fo| - |Fc /Σ|Fo|. cwR2 =
[Σ (w(Fo2 - Fc2)2)/Σw(Fo2)2]1/2
The intensity of the radiation source at 460 nm is given by eq 2
where Φ = 1.25.
a
˚
TABLE 2. Selected Interatomic Distances (A) and Angles (deg)
½FeðIIÞꢁ=irradiation time
intensity ðquanta s- 1L- 1Þ ¼
NA
ð2Þ
bond lengths
bond angles
Φ460
N(3)-N(3)
1.2793(15)
1.4127(16)
1.3573(18)
1.4434(18)
0.865(15)
2.219
N(3)-N(3)-C(12)
116.87(10)
126.89(11)
121.64(11)
120.1(11)
115.3(11)
110.70(10)
118.15(12)
N(3)-C(12)
N(2)-C(7)
N(3)-C(12)-C(7)
C(12)-C(7)-N(2)
C(7)-N(2)-H(2A)
H(2A)-N(2)-C(6)
N(2)-C(6)-C(5)
C(6)-C(5)-N(1)
Quantum Yield of Photoisomerization. A 1.3-mg sample of
AzoAMP-2 was dissolved in 1.0 mL of CDCl3 to achieve a final
concentration of 3.1 mM. The solution was transferred into an
N(2)-C(6)
N(2)-H(2A)
N(1)-H(2A)
N(3)-H(2A)
1
NMR tube, and the H NMR spectrum of the solution was
2.046
recorded. The NMR tube was irradiated at 460 nm for 30 min,
and the 1H NMR spectrum of the solution was recorded.
Growth of new peaks was observed near NCH3 resonance and
NCH2Py resonance of the 1H NMR spectrum. The [cis]:[trans]
isomer ratio was calculated using integrated peak areas of the
NCH3 resonance.
aNumber in parentheses are estimated standard deviations in the last
digit(s). Atom labels are provided in Figure 3A.
Absorption spectra were recorded on a UV-vis spectrophoto-
meter under the control of a PC running the manufacturer
supplied software package. Spectra were routinely acquired at
25 °C, in 1-cm path length quartz cuvettes with a total volume of
3.0 mL. Fluorescence spectra were recorded on a spectrophoto-
meter under the control of a PC running the manufacturer
supplied software package. Excitation was provided by a 150 W
Xe lamp operating at a current of 5 A. Spectra were routinely
acquired at 25 °C, in 1 cm quartz cuvette with a total volume of
3.0 mL using, unless otherwise stated, 10 nm slit widths, and a
photomultiplier tube power of 700 V. Photoisomerization ex-
periments were performed at 25 °C, in 1-cm path length quartz
cuvettes illuminated by a 150 W Xe lamp of a spectrophot-
ometer with emission wavelength set to λmax of the AzoAMP
species.
Irradiation of AzoAMP-1 and AzoAMP-3. Solutions of
AzoAMP (10 and 25 μM) were irradiated at λmax in hexanes,
benzene, Et2O, THF, EtOAc, DMF, DMSO, MeOH, and Et2O/
EtOH (1:1). Aqueous solutions of AzoAMP-1 (pH 1, 4, 7, 10,
and 14 and in concentrated HCl) were also irradiated. In DMF
the excitation wavelength was varied between 250 and 600 nm
(10 nm intervals between 450 and 550 nm, 20 nm intervals
outside this range). Several different excitation wavelengths
were tried in water, MeOH, and DMSO. No significant isomer-
ization was observed under any of these conditions.
The quantum yield of photoisomerization (Φ) of AzoAMP-2
is obtained by solving eq 3:
Change in ½AzoMAP-2ꢁ=irradiation time
Φ ¼
NA ð3Þ
intensity of source
Low-Temperature Absorption Spectroscopy. A 1.5-mL aliquot
of EtOH and a 1.5-mL aliquot of Et2O were placed in a cuvette,
and the background UV-vis spectrum was recorded. A 20-μL
aliquot of an AzoAMP solution (5.00 mM) was added to achieve
a final concentration of 50 μM, and the spectrum was recorded.
The cuvette containing the AzoAMP solution was immersed in
liquid N2 to obtain a transparent glass and its absorption
spectrum was recorded.
Low-Temperature Fluorescence Spectroscopy. A 1.0-mL ali-
quot of EtOH and a 1.0-mL aliquot of Et2O were placed in
a quartz cuvette. A 40-μL aliquot of an AzoAMP solution
(5.00 mM) was added to achieve a final concentration of
100 μM, and the fluorescence spectrum was recorded. The
cuvette containing the AzoAMP solution was immersed in
liquid N2 to obtain a transparent glass, and its fluorescence
spectrum was recorded.
Quantum Yield of Fluorescence. Quantum yields were calcu-
lated by measuring the integrated emission area of the corrected
spectra and comparing that value to the area measured for
Isomerization of AzoAMP-2. A 1.0-mL aliquot of EtOH and a
1.0-mL aliquot of Et2O were placed in a cuvette and the back-
ground UV-vis spectrum was recorded. A 4-μL aliquot of an
AzoAMP-2 solution (5.00 mM) was added to achieve a final
concentration of 10 μM, and the absorption spectrum was
recorded. The cuvette was irradiated at 451 nm, and spectra
were recorded at 1 min intervals. No additional changes in
absorbance were observed after 15 min. The solution was kept
fluorescein in 0.1 N NaOH when excited at 490 nm (Φfl
=
0.85). The quantum yield of AzoAMP-1 (100 μM solution in
1:1 EtOH/Et2O frozen in liquid N2) was then calculated using
eq 1, where F represents the area under the emission spectra for
the standard and sample, η is the refractive index of the solvent,
and A is the absorbance at the excitation wavelength selected for
4826 J. Org. Chem. Vol. 75, No. 14, 2010

Bandara et al.
JOCArticle
the standard and sample. Emission was integrated between 502
and 700 nm.
Note Added after ASAP Publication. This paper was pub-
lished on June 14, 2010, with several typographic errors in the
Introduction and Results sections. The corrected version was
reposted on June 16, 2010.
!
!
!
Fsample
ηsample
Astandard
Asample
sample
fl
standard
fl
Φ
¼ Φ
ð4Þ
Fstandard
ηstandard
Acknowledgment. We thank Professor Challa Vijaya
Kumar for insightful discussions. Work in the laboratories
of S.C.B. and J.G. was supported by the University of Con-
necticut. Work in the laboratory of H.A.F. is supported by a
grant from the National Science Foundation (MCB-0913022)
and by the University of Connecticut Research Foundation.
W.I. was supported by the University of Connecticut Depart-
ment of Chemistry NSF-REU program (CHE-0754580).
Computational Methods. Ground and excited states were
obtained using density functional theory (DFT) and time-
dependent DFT with the hybrid functional B3LYP and the
split-valence double-ζ basis set 6-31g**. The quantum chemistry
package Jaguar was used for all calculations.61
Ultrafast Time-Resolved Spectroscopic Methods. Transient
absorption spectra were recorded using a femtosecond laser
spectrometer system previously described.62 The transient experi-
ments were done at room temperature on AzoAMP-1 and
AzoAMP-2 ethanol/ether (1:1, v/v) and adjusted to an optical
density of ∼1.0 in a 2 mm path length quartz cuvette at the
excitation wavelength of 490 nm. The pump laser beam had an
energy of 1 μJ/pulse and a spot diameter of 1 mm corresponding to
an intensity of ∼3.2 ꢀ 1014 photons/pulse/cm2. Steady-state
absorption spectra were recorded after each experiment to check
the integrity of the samples. Surface Xplorer Pro (v.1.1.0.17)
software was used for chirp correction of the spectral data sets
and for the determination of the number of principal components
via single value decomposition (SVD). ASUfit version 3.0 software
was used for global fitting analysis. The goodness of fit was
checked from the values of the residuals matrix and chi square (χ2).
1
Supporting Information Available: H and13C NMR spectra
for all new compounds synthesized; absorption of AzoAMP-1
and AzoAMP-2 at room temperature and 77 K; changes in the
absorption spectrum of AzoAMP-1 upon irradiation; changes in
the absorption spectrum ofAzoAMP-2uponirradiation; thermal
recovery of trans-AzoAMP-2 after removal from light source;
titration of AzoAMP-2 with Zn2þ 1H NMR of AzoAMP-2
;
before and after irradiation; changes in the absorption spectrum
of AzoAMP-3 upon irradiation; additional calculated energy
surfaces for inversion, rotation, and concerted inversion path-
ways for azobenzene, AzoAMP-1, and AzoAMP-2; changes in
the absorption spectrum of 2,20-diaminoazobenzene upon irra-
diation; complete tables of X-ray data and fully labeled ORTEP
diagram; transient spectrum of ethanol/ether (1:1, v/v) recorded
at a delay time of 100 fs after laser excitation. This material is
available free of charge via the Internet at http://pubs.acs.org.
(61) Jaguar 7.5; Schrodinger, LLC: New York, NY, 2008.
(62) Ilagan, R. P.; Christensen, R. L.; Chapp, T. W.; Gibson, G. N.;
Pascher, T.; Polivka, T.; Frank, H. A. J. Phys. Chem. A 2005, 109, 3120–
3127.
J. Org. Chem. Vol. 75, No. 14, 2010 4827