Visible light switching of a BF2-coordinated azo compound

Article abstract of DOI:10.1021/ja306030d

Here we report the synthesis and characterization of a BF₂-azo complex that can be induced to isomerize without the need of deleterious UV light. The complexation of the azo group with BF₂, coupled with the extended conjugation of the N=N π-electrons, increases the energy of the n-π* transitions and introduces new π-nonbonding (π_nb) to π* transitions that dominate the visible region. The well separated π_nb-π* transitions of the trans and cis isomers enable the efficient switching of the system by using only visible light. The complexation also leads to a slow cis → trans thermal relaxation rate (t_1/2 = 12.5 h). Theoretical calculations indicate that the absorption bands in the visible range can be tuned using different Lewis acids, opening the way to a conceptually new strategy for the manipulation of azo compounds using only visible light.

Full text of DOI:10.1021/ja306030d

Communication
pubs.acs.org/JACS
Visible Light Switching of a BF2‑Coordinated Azo Compound
Yin Yang, Russell P. Hughes, and Ivan Aprahamian*
Department of Chemistry, Dartmouth College, Hanover, New Hampshire 03755, United States
S
* Supporting Information
which is undesirable for certain applications (e.g., protein
probes,14 molecular machines6), side reactions as is the case
with protonation,15 and relatively low isomerization ratios as is
the case with the MLCT systems.12 Nonetheless, Herges,
Woolley and Nishihara have successfully developed a handful of
systems9,12c that can be activated by visible light in both
directions, without the accompanied enhancement in thermal
isomerization rate. This behavior results either from the
separation of the n−π* bands or interligand CTs.
ABSTRACT: Here we report the synthesis and character-
ization of a BF2−azo complex that can be induced to
isomerize without the need of deleterious UV light. The
complexation of the azo group with BF2, coupled with the
extended conjugation of the NN π-electrons, increases
the energy of the n−π* transitions and introduces new π-
nonbonding (πnb) to π* transitions that dominate the
visible region. The well separated πnb−π* transitions of the
trans and cis isomers enable the efficient switching of the
system by using only visible light. The complexation also
An interesting but a much less explored pathway to shifting
the π−π* transitions in azo compounds to the visible range is
by coordinating the azo group’s n-electrons with Lewis acids.
Some early reports have looked at this effect;16 however, none
to the best of our knowledge took advantage of it to make
switchable systems that can be fully modulated in the visible
range. Here we wish to report the photoswitching of a BF2−azo
complex (1), which demonstrates a highly efficient trans/cis
isomerization process that can be fully controlled using visible
light. The system also undergoes a slow cis → trans thermal
isomerization, the rate of which can be modulated using
molecular oxygen. On the basis of theoretical analysis, the
complexation of BF2 with the azo group’s nitrogen lone-pair
electrons, coupled with the extended π-system in 1, results in
the reversal of the position of the n−π* and πnb−π* transitions
on the energy scale (similar to the pseudo stilbene type
systems). This phenomenon leads to two isomers with
extraordinarily well separated πnb−π* bands that have large
molar extinction coefficients in the visible range.
leads to a slow cis → trans thermal relaxation rate (t1/2
=
12.5 h). Theoretical calculations indicate that the
absorption bands in the visible range can be tuned using
different Lewis acids, opening the way to a conceptually
new strategy for the manipulation of azo compounds using
only visible light.
zobenzene has been the focus of intense research1 since it
A
was first reported by Alfred Nobel in 1856.2 Because of its
unique photophysical properties, azobenzene has been used in
various applications ranging from industrial dyes,3 to actuators,4
nonlinear optical devices,5 molecular switches and machines,6
ion channel modulators,7 and so forth. The π-system of azo
compounds can be easily modified through the use of
substituents that lead to changes in its absorption profile and
more importantly its reversible trans/cis isomerization process.
These properties have been extensively studied leading to the
categorization of azobenzene derivatives into three groups:
azobenzene, aminobenzene, and pseudo stilbene type mole-
cules, based on the relative energetic order of their n−π* and
π−π* transitions.1 Recently, there has been a thrust8 to develop
azobenzene photoswitches that can be toggled between the
trans and cis states by using only visible light.9 The reason
behind this is that the use of UV light, which is required for the
trans → cis isomerization can lead to complications (i.e.,
photodamage)9a especially for in vivo applications (e.g., strong
scattering10 and apoptosis11). Various strategies,1,9 which rely
on the separation of the n−π* bands in the visible range,
pushing the π−π* transition to that region, or using metal-to-
ligand charge transfer (MLCT) to red-shift the activation
wavelength,12 have been used to accomplish this goal,
including: (i) the incorporation of electron-withdrawing or
-donating groups in the ortho and para positions, (ii) the use of
push−pull substituents, (iii) protonation, (iv) the use of
bridgehead derivatives, and (v) the incorporation of metal
complexes. In certain cases, these approaches have led to the
desired red-shift in the absorption profiles but they were also
accompanied by very fast cis → trans thermal isomerization,7c,13
Our interest in hydrazone-based switches17 has led us to the
development of BF2−hydrazone complexes18 that display
aggregation induced emission.19 As part of our efforts to
optimize the properties of these solid-state dyes, we
serendipitously discovered that the reaction of boron trifluoride
with hydrazone 3 (Scheme 1) yields 1 as the main product
(68%) instead of the expected BF2−hydrazone 2 (10%). On
the other hand, carrying out the reaction at elevated
temperatures (60 °C) instead of RT yields 2 as the
Scheme 1. Synthesis of Compounds 1 and 2
Received: June 27, 2012
Published: September 6, 2012
© 2012 American Chemical Society
15221
dx.doi.org/10.1021/ja306030d | J. Am. Chem. Soc. 2012, 134, 15221−15224

Journal of the American Chemical Society
Communication
predominant compound. Both of these complexes were
isolated, purified using chromatography, and characterized
using NMR spectroscopy, mass spectrometry and X-ray
crystallography (see SI).
Figure 2. (a) The UV/vis spectral changes upon the photo-
isomerization of 1 in deoxygenated CH2Cl2 (0.1 mM). The black
trace is of 1 equilibrated in the dark (mainly trans), which upon
irradiation at λ = 570 nm gives the cis PSS (blue trace), which when
irradiated at λ = 450 nm gives the trans PSS (red trace). (b) Multiple
isomerization cycles of 1 (0.1 mM) in CH2Cl2 (not deoxygenated)
after alternative irradiation at λ=570 (red trace) and 450 nm (black
trace).
Figure 1. ORTEP drawing (50% probability ellipsoids) of the crystal
structure of 1. The hydrogen atoms and disorder in the phenyl ring
have been removed for clarity.
Analysis of the crystallographic data (Figure 1 and Figure
S15) reveals that 1 crystallizes in the trans form with a N(2)
N(3) bond length of 1.278(3) Å that is almost identical to that
of azobenzene (average 1.25 Å).20 Moreover, the X-ray
structure shows that the coordination of BF2 with one of the
azo group nitrogens (N2) along with the quinolinyl nitrogen
(N1) forms a five-membered ring having B−N bond lengths of
1.632(3) and 1.568(3) Å, respectively.21 The 1H NMR
spectrum (Figure S10) of an equilibrated CH2Cl2 solution
(in the dark) shows two sets of signals in the ratio of 87:13
assigned to the trans and cis isomers, respectively. A full
exchanging during the isomerization process (Figure 2a). What
is remarkable is that the trans/cis isomerization can be activated
solely by the use of visible light and there is no need for UV
light. Furthermore, as shown in Figure 2b, the system shows
very good reversibility as no degradation of 1 was observed
during the entire photoisomerization studies that lasted for
more than a month.
The isomerization process was also studied extensively using
NMR spectroscopy (Figures S5−S10). The 2D heteronuclear
1H−19F NOESY spectrum of 1 (trans dominant) shows (Figure
S8) as expected a correlation between the BF2 fluorine atoms
and protons H11 and H1 (Scheme 2). On the other hand, only
one correlation is observed for the cis isomer, between the BF2
fluorine atoms and proton H11′ (Figure S7). This indicates
that the phenyl ring is moving away from the BF2 group upon
1
assignment of the H NMR signals using COSY and NOESY
spectroscopies (Figures S4 and S5) revealed that the well
resolved signals at 8.42 and 8.33 ppm belong to proton H7 of
the cis and trans isomers of 1 (Scheme 2), respectively. These
1
isomerization. The H NMR spectrum of a 1:1 mixture of 1
Scheme 2. Visible Light-Induced Trans/Cis Isomerization of
1
does not change when pyridine is added to the solution (Figure
S9). This indicates that the boron atom is not available for
coordination with pyridine, that is, it is tetracoordinated in both
isomers. These experiments rule out the light-induced B−N
bond cleavage observed in other systems,21 and show that the
observed phenomenon in 1 is indeed the trans/cis isomerization
process.
The photoswitching performance of 1 in deoxygenated
dichloromethane was evaluated by measuring its photosta-
tionary state (PSS) and photoisomerization quantum yield
(Figures S10−S13). The irradiation of a trans-dominant NMR
sample of 1 at 570 nm for 5 min yields 97% of the cis isomer at
the PSS, as determined by the signal intensity of protons H7
and H7′. Using the change in integration of protons H7 and
H7′ as a function of irradiation time and knowing the light
intensity at 570 nm (see SI), the quantum yield for the trans →
cis isomerization (Φtrans→cis) was calculated to be 48 6%.23
Similarly, irradiation at 450 nm for 30 min produced 80% of the
signals were subsequently used as probes to quantitatively
follow the trans/cis isomerization process in 1. The 19F NMR
spectrum also showed two quartets at −142.72 and −150.60
ppm that were assigned to the trans and cis isomers,
respectively.
The photoisomerization of 1 was studied extensively by UV/
vis (Figure 2) spectroscopy. When stored in the dark, 1 adopts
its thermodynamically stable trans form that has an absorption
maximum (λmax) at 530 nm (ε = 8026 M−1 cm−1). Upon
irradiation at 570 nm, the cis form (λmax = 480 nm; ε = 7792
M−1 cm−1) becomes dominant, accompanied by a sharp color
change of the solution from bright purple to light orange
(Scheme 2).22 The process is also accompanied by changes in
the intensity of bands at higher energies (λmax = 340 and 264
nm). Irradiation at 450 nm drives the system back to its trans
form. The isosbestic points (λ = 499, 399, 330, and 257 nm) in
the UV/vis spectra demonstrate that only two species are
trans isomer at the PSS, with a Φcis→trans of 67
8%. The
thermal cis→trans isomerization of 1 was monitored using UV−
vis spectroscopy (Figure S14), and the half-life at 294 K was
determined to be 12.5 h, which is very long compared to
pseudo stilbene type azo systems that are in the millisecond
domain!1 Significantly, we found that the rate of thermal cis →
trans isomerization is sensitive to the presence of oxygen in the
solvent: when regular dichloromethane (not deoxygenated)
was used for the measurements, the half-life went down to 30
min. Such a dependency of the isomerization rate on oxygen
15222
dx.doi.org/10.1021/ja306030d | J. Am. Chem. Soc. 2012, 134, 15221−15224

Journal of the American Chemical Society
Communication
concentration is atypical for azo compounds.1 The nature of the
effect and its consequence on the isomerization mechanism in
our system is under investigation.
energy πnb molecular orbital which serves as the HOMO
(Figures S37−S41). This in turns leads to the strong absorption
band in the visible range that enables the manipulation of 1
using only visible light. Significantly, the BF2 group is not
unique in promoting this effect. As a proof of concept, we
replaced it with Na+ (in silico), which also resulted in red-shifted
UV/vis absorption spectra (Figures S33 and S34). Intriguingly,
the trans and cis isomers have a λmax of 466 and 472 nm,
respectively, implying that the absorption profile of the system
can be manipulated using different Lewis acids! Moreover, the
calculations predict that replacing the CN group with π-
electron donating groups (Figures S35 and S36) and/or
substituting the phenyl ring with such groups (based on the
HOMO in Figure S29) will lead to a red shift in the absorption
band, opening the door for further manipulations of the
photophysical properties of the azocompound.
In conclusion, we have shown that the coordination of BF2
with an azo group’s nitrogen lone-pair, coupled with the
extended conjugation of the NN π-electrons, leads to a
reconfiguration of its electronic structure, and a reversal of the
n−π* and πnb−π* transitions on the energy scale. Significantly,
unusually well separated and strong πnb−π* transitions are
observed for the trans and cis isomers in the visible range, which
enable the switching of the system using only visible light. The
high photoconversion, and photoisomerization quantum yields
(which are higher than in azobenzene)1 combined with the
slow thermal relaxation and tunability of the photophysical
properties make this conceptually new Lewis acid complexation
strategy a viable pathway to the development of visible-light
actuated functional materials.4
The coordination of BF2 with the n-electrons of the azo
nitrogen in 1 is expected to reverse the position of the n−π*
and π−π* transitions on the energy scale, as with pseudo
stilbene type azobenzenes.1 Recent reports24 have indicated,
however, that this energy reversal might not be a universal one
in these systems. To understand the effect of Lewis acid
coordination on the photophysical properties of azo com-
pounds, we conducted computational modeling of the trans and
cis isomers of 1. Structures were optimized by density
functional theory (DFT) using the B3LYP hybrid func-
tional25,26 and the 6-311++G** basis set,27 as implemented
in Jaguar;28 this combination of method and basis set is
appropriate for such systems.9c The optimized structure of
trans-1 matches well with its crystal structure (Figures S17 and
S18). Calculations of the UV/vis spectra of the optimized
structures were carried out in ADF29 using time-dependent
DFT (TDDFT)30 using the B3LYP functional and a triple-ζ
basis with two added polarization functions (TZ2P). These
calculations were also successful in predicting the UV/vis
spectra of the cis and trans isomers of 1 (Figures S31 and S32),
and show that the absorption bands in the visible range (Figure
3, Table S2 and Figures S37−S41) stem from π-nonbonding to
ASSOCIATED CONTENT
* Supporting Information
■
S
Experimental procedures, NMR spectra of key compounds,
photoisomerization studies, kinetic measurements, computa-
tional details and X-ray crystallography data. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
ivan.aprahamian@dartmouth.edu
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by Dartmouth College and the Burke
Research Initiation Award. R.P.H. thanks the National Science
Foundation for an operating grant that made possible the
purchase of hardware and software used for the DFT
calculations. The authors thank Dr. Richard Staples (Michigan
State University) for the X-ray analysis, and Sahag Voskian for
help with the table of contents graphic.
Figure 3. The calculated (B3LYP/6-311++G**) molecular orbital
energy levels, transition energies and oscillator strengths of the n−π*
and πnb−π* transitions of the trans isomer of 1.
π*-antibonding transitions (HOMO → LUMO). Moreover,
the calculations predict correctly the separation between the cis
(λmax = 482 nm) and trans bands (λmax = 510 nm). Another set
of πnb−π* transitions (HOMO to LUMO+1) is also predicted
at a higher energy level (λmax = 376 and 353 nm for trans and
cis, respectively) where a smaller band is clearly visible in the
UV/vis spectrum (Figure 2a), whereas the n−π* transition is,
as predicted, at an even shorter wavelength (λmax = 338 nm for
trans).31 Relative to azobenzenes, binding to BF2 drastically
lowers the energy of the n-electrons, while additional
conjugation in the N-C-C-N-N skeleton provides a higher
REFERENCES
■
(1) (a) Rau, H. In Photochemistry and Photophysics; Rabek, J. F., Ed.;
CRC Press: Boca Taton, FL, 1990; Vol. 2, pp119−141. (b) Knoll, H.
CRC Handbook of Organic Photochemistry and Photobiology, 2nd ed;
CRC Press: Boca Taton, FL, 1996. (c) Bandara, H. M. D.; Burdette, S.
C. Chem. Soc. Rev. 2012, 41, 1809.
(2) Noble, A. Justus Liebigs Ann. Chem. 1856, 98, 253.
(3) Venkataraman, K. The Chemistry of Synthetic Dyes; Academic
Press: New York, 1956.
(4) Bleg
́
er, D.; Yu, Z.; Hecht, S. Chem. Commun. 2011, 47, 12260.
15223
dx.doi.org/10.1021/ja306030d | J. Am. Chem. Soc. 2012, 134, 15221−15224

Journal of the American Chemical Society
Communication
(5) Delaire, J. A.; Nakatani, K. Chem. Rev. 2000, 100, 1817.
(6) (a) Kay, E. R.; Leigh, D. A.; Zerbetto, F. Angew. Chem., Int. Ed.
2007, 46, 72. (b) Balzani, V.; Credi, A.; Venturi, M. Molecular Devices
and Machines; Wiley-VCH: Weinheim, 2008.
(7) (a) Banghart, M.; Borges, K.; Isacoff, E.; Trauner, D.; Kramer, R.
H. Nat. Neurosci. 2004, 7, 1381. (b) Beharry, A. A.; Woolley, G. A.
Chem. Soc. Rev. 2011, 40, 4422. (c) Mourot, A.; Fehrentz, T.; Feuvre,
Y. L.; Smith, C. M.; Herold, C.; Dalkara, D.; Nagy, F.; Trauner, D.;
Kramer, R. H. Nat. Methods 2012, 9, 396.
Boni, L.; Toro, C.; Masunov, A. E.; Hernandez, F. E. J. Phys. Chem. A
2008, 112, 3886.
(25) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
(26) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 1372. (b) Becke, A. D.
J. Chem. Phys. 1993, 98, 5648.
(27) (a) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem.
Phys. 1980, 72, 650. (b) Clark, T.; Chandrasekhar, J.; Spitznagel, G.
W.; Schleyer, P. v. R. J. Comput. Chem. 1983, 4, 294. (c) Frisch, M. J.;
Pople, J. A.; Binkley, J. S. J. Chem. Phys. 1984, 80, 3265.
(28) Jaguar, version 7.7; Schrodinger, LLC: New York, NY, 2010.
̈
(8) Wegner, H. A. Angew. Chem., Int. Ed. 2012, 51, 4787.
(29) (a) ADF 2012; Vrije Universiteit: Amsterdam, The Nether-
lands, 2012. Available at http://www.scm.com. (b) Guerra, C. F.;
Snijders, J. G.; Velde, G. t.; Baerends, E. J. Theor. Chem. Acc. 1998, 99,
391. (c) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca
Guerra, C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J.
Comput. Chem. 2001, 22, 931.
(30) van Gisbergen, S. J. A.; Snijders, J. G.; Baerends, E. J. Comput.
Phys. Commun. 1999, 118, 119.
(31) It is difficult to assign the n−π* transition for the cis isomer
because of the mixed character of the bands at higher energy.
(9) (a) Siewertsen, R.; Neumann, H.; Buchheim-Stehn, B.; Herges,
R.; Nather, C.; Renth, F.; Temps, F. J. Am. Chem. Soc. 2009, 131,
15594. (b) Beharry, A. A.; Sadovski, O.; Woolley, G. A. J. Am. Chem.
Soc. 2011, 133, 19684. (c) Venkataramani, S.; Jana, U.; Dommaschk,
M.; Sonnichsen, F. D.; Tuczek, F.; Herges, R. Sciecne 2011, 331, 445.
̈
(10) Cheong, W. F.; Prahl, S. A.; Welch, A. J. IEEE J. Quantum
Electron. 1990, 26, 2166.
(11) (a) Tamai, T. K.; Vardhanabhuti, V.; Foulkes, N. S.; Whitmore,
D. Curr. Biol. 2004, 14, R104. (b) Banerjee, G.; Gupta, N.; Kapoor, A.;
Raman, G. Cancer Lett. 2005, 223, 275.
(12) (a) Kurihara, M.; Hirooka, A.; Kume, S.; Sugimoto, M.;
Nishihara, H. J. Am. Chem. Soc. 2002, 124, 8800. (b) Nihei, M.;
Kurihara, M.; Mizutani, J.; Nishihara, H. J. Am. Chem. Soc. 2003, 125,
2964. (c) Sakamoto, R.; Murata, M.; Sampei, H.; Sugimoto, M.;
Nishihara, H. Chem. Commun. 2005, 1215. (d) Kume, S.; Nishihara, H.
Dalton Trans. 2008, 3260. (e) Sakamoto, R.; Kume, S.; Sugimoto, M.;
Nishihara, H. Chem.Eur. J. 2009, 15, 1419.
(13) (a) Sadovsky, O.; Beharry, A. A.; Zhang, F.; Woolley, G. A.
Angew. Chem., Int. Ed. 2009, 48, 1484. (b) Mourot, A.; Kienzler, M. A.;
Banghart, M. R.; Fehrentz, T.; Huber, F. M. E.; Stein, M.; Kramer, R.
H.; Trauner, D. ACS Chem. Neurosci. 2011, 2, 536.
(14) Kim, Y.; Phillips, J. A.; Liu, H.; Kang, H.; Tan, W. Proc. Natl.
Acad. Sci. U.S.A. 2009, 106, 6489.
(15) (a) Lewis, G. E. J. Org. Chem. 1960, 25, 2193. (b) Badger, G. M.;
Joshua, C. P.; Lewis, G. E. Aust. J. Chem. 1965, 18, 1639.
(16) (a) Shuba, R. J.; Zenchelsky, S. T. J. Am. Chem. Soc. 1960, 82,
4136. (b) Gutmann, V.; Steininger, A. Monatsh. Chem. 1965, 4, 1173.
(17) (a) Landge, S. M.; Aprahamian, I. J. Am. Chem. Soc. 2009, 131,
18269. (b) Su, X.; Aprahamian, I. Org. Lett. 2011, 31, 30. (c) Su, X.;
Robbins, T. F.; Aprahamian, I. Angew. Chem., Int. Ed. 2011, 50, 1841.
(d) Landge, S. M.; Tkatchouk, K.; Benitez, D.; Lanfranchi, D. A.;
Elhabiri, M; Goddard, W. A., III.; Aprahamian, I. J. Am. Chem. Soc.
2011, 133, 9812. (e) Su, X.; Lessing, T. Beilstein J. Org. Chem. 2012, 8,
872. (f) Ray, D.; Foy, J. T.; Hughes, R. P.; Aprahamian, I. Nat. Chem.
2012, 4, 757.
(18) Yang, Y.; Su, X.; Carroll, C. N.; Aprahamian, I. Chem. Sci. 2012,
3, 610.
(19) Hong, Y.; Lam, J. W. Y.; Tang, B. Z. Chem. Soc. Rev. 2011, 40,
5361.
(20) Allmann, R. In The Chemistry of the Hydrazo, Azo and Azoxy
Groups; Patai, S., Ed.; John Wiley & Sons: London, 1975; Chapter 2, p
43.
(21) These B−N bonds clamp the system in position, and hence,
only the NN bond can undergo isomerization.
(22) In structurally related systems, no light-induced B−N bond
cleavage is observed when the bond length is 1.638 Å or shorter.
Hence, and based on the B−N bond lengths in 1, we can rule out this
process in this system. This conclusion is supported by the NMR
spectroscopy and calculation data. For more details see: (a) Yoshino,
J.; Kano, N.; Kawashima, T. Chem. Commun. 2007, 559. (b) Yoshino,
J.; Furuta, A.; Kambe, T.; Itoi, H.; Kano, N.; Kawashima, T.; Ito, Y.;
Asashima, M. Chem.Eur. J. 2010, 16, 5026.
(23) Bandara, H. M. D.; Friss, T. R.; Enriquez, M. M.; Isley, W.;
Incarvito, C.; Frank, H. A.; Gascon, J.; Burdette, S. C. J. Org. Chem.
2010, 75, 4817.
(24) (a) Crecca, C.; Roitberg, A. J. Phys. Chem. A 2006, 110, 8188.
(b) Poprawa-Smoluch, M.; Baggermann, J.; Zhang, H.; Maas, H. P. A.;
De Cola, L.; Brouwer, M. J. Phys. Chem. A 2006, 110, 11926. (c) De
15224
dx.doi.org/10.1021/ja306030d | J. Am. Chem. Soc. 2012, 134, 15221−15224