Ultrafast Spectroscopy of Hydroxy-Substituted Azobenzenes in Water

Article abstract of DOI:10.1002/chem.201501863

Ultrafast UV/Vis pump/probe experiments on ortho-, meta- and para-hydroxy-substituted azobenzenes (HO-ABs), as well as for sulfasalazine, an AB-based drug, were performed in aqueous solution. For meta-HO-AB, AB-like isomerisation behaviour can be observed, whereas, for ortho-HO-AB, fast proton transfer occurs, resulting in an excited keto species. For para-HO-AB, considerable keto/enol tautomerism proceeds in the ground state, so after excitation the trans-keto species isomerises into the cis form. Aided by TD-DFT calculations, insight is provided into different deactivation pathways for HO-AB, and reveals the role of hydroxy groups in the photochemistry of ABs, as well as their acetylation regarding sulfasalazine. Hydroxy groups are position-specific substituents for AB, which allow tuning of the timescale of thermal relaxation, as well as the amount and contribution of the keto species to photochemical processes.

Full text of DOI:10.1002/chem.201501863

DOI: 10.1002/chem.201501863
Full Paper
&
Photochemistry
Ultrafast Spectroscopy of Hydroxy-Substituted Azobenzenes in
Water
Sabrina Steinwand,[a] Thomas Halbritter,[b] Dominique Rastädter,[c] Juan Manuel Ortiz-
Sµnchez,[c] Irene Burghardt,*[c] Alexander Heckel,*[b] and Josef Wachtveitl*[a]
Abstract: Ultrafast UV/Vis pump/probe experiments on
ortho-, meta- and para-hydroxy-substituted azobenzenes
(HO-ABs), as well as for sulfasalazine, an AB-based drug,
were performed in aqueous solution. For meta-HO-AB, AB-
like isomerisation behaviour can be observed, whereas, for
ortho-HO-AB, fast proton transfer occurs, resulting in an ex-
cited keto species. For para-HO-AB, considerable keto/enol
tautomerism proceeds in the ground state, so after excita-
tion the trans-keto species isomerises into the cis form.
Aided by TD-DFT calculations, insight is provided into differ-
ent deactivation pathways for HO-AB, and reveals the role of
hydroxy groups in the photochemistry of ABs, as well as
their acetylation regarding sulfasalazine. Hydroxy groups are
position-specific substituents for AB, which allow tuning of
the timescale of thermal relaxation, as well as the amount
and contribution of the keto species to photochemical pro-
cesses.
Introduction
or cis isomers. Irradiation of the trans form with UV light typi-
cally generates the cis isomer, which, in turn, can be switched
back to the trans isomer with visible-light irradiation or
through a thermally controlled pathway.[5,6]
In recent years, photochromic molecules, which have the abili-
ty to undergo reversible structural changes after activation by
light, have become increasingly important for a wide field of
applications, ranging from regulation of biological processes to
data storage.[1–4] The photoisomers of these photoswitches
differ significantly and changes in their physical and chemical
properties can be used to control a variety of biological effects
or properties of a material with an external light signal. The
use of light has various advantages: no additional chemical
components have to be brought into the system, and there-
fore, possible side reactions and disturbance of the biological
system is avoided. Furthermore, irradiation can be performed
with a simple experimental setup and spatial and temporal
control allows an exact location and timing of the desired
effect.
This type of photoswitch is a powerful tool whenever a con-
formational change should be induced. In this way, AB is used
in polymers[7] and foldamers,[8] gold nanoparticles,[9] liquid crys-
tals[4] and even crystals for non-linear optics.[10] Another impor-
tant field is the biological application of ABs,[11] with contribu-
tions from our groups and the groups of Asanuma or Trauner.
In AB-modified peptides, a light-dependent modulation of sec-
ondary[12,13] and tertiary structural motifs[14] was demonstrated.
ABs have also been covalently introduced into DNA and
RNA[15] either as hybridisation photoswitches[16] or to modulate
the activity of ligand-gated ion channels.[17,18]
In our attempts to develop a light-responsive riboswitch, we
are interested in highly functionalised AB-containing small
molecules. Investigating the photoswitchability of the commer-
cially available sulfasalazine (1; Figure 1), a drug known for the
treatment of inflammatory bowel disease and rheumatoid ar-
thritis, we could only detect moderate switching behaviour in
anhydrous DMSO. In aqueous environments, however, trans–
cis isomerisation could not be observed by UV/Vis measure-
ments after illumination at l=365 nm. Acetylation of the hy-
droxy group (2) allows accumulation of the cis isomer, again
on a moderate timescale. A variety of studies have been per-
formed on HO-ABs regarding, for example, pH[19] and solvent
dependence,[20,21] as well as fast thermal cis–trans relaxa-
tion.[22,23] One possible reason for the fast relaxation rates is
the proposed keto/enol tautomerism of HO-AB, resulting in
a ketohydrazone in protic solvents.[24–29] Although ortho-HO-
ABs could undergo intramolecular proton transfer, para substi-
tution should give rise to an intermolecular mechanism
Among the best studied photochromic systems are deriva-
tives of azobenzene (AB), which can be present as either trans
[a] S. Steinwand, Prof. Dr. J. Wachtveitl
Institute of Physical and Theoretical Chemistry
Goethe-University Frankfurt, Max-von-Laue-Str. 7
60438 Frankfurt (Germany)
E-mail: wveitl@theochem.uni-frankfurt.de
[b] T. Halbritter, Prof. Dr. A. Heckel
Institute for Organic Chemistry and Chemical Biology
Goethe-University Frankfurt, Max-von-Laue-Str. 7
60438 Frankfurt (Germany)
[c] D. Rastädter, Dr. J. M. Ortiz-Sµnchez, Prof. Dr. I. Burghardt
Institute of Physical and Theoretical Chemistry
Goethe-University Frankfurt, Max-von-Laue-Str. 7
60438 Frankfurt (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501863.
Chem. Eur. J. 2015, 21, 15720 – 15731
15720
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Figure 1. Chemical structures of sulfasalazine (1) and its acetylated derivative
2, as well as the three hydroxy-substituted azobenzenes (HO-ABs; 3–5) in-
vestigated herein.
Figure 2. Steady-state UV/Vis spectra of 1 in (anhydrous) DMSO and PBS,
and 2 in PBS before (trans) and after illumination (photostationary state
(PSS)) at l=365 nm (top), as well as spectra of 3, 4 and 5 at thermal equilib-
rium in PBS (bottom).
through solvent molecules. However, no ultrafast measure-
ments have been performed so far, so the question arises as to
how HO-ABs isomerise (trans–cis) and if the contribution from
a keto species can be detected. Moreover, one needs to inves-
tigate if intra-/intermolecular proton transfer occurs in the
ground state or only after photoactivation. Therefore, we pres-
ent herein ultrafast spectroscopic studies of HO-ABs 3–5
(Figure 1), in addition to sulfasalazine before (1) and after ace-
tylation (2). Moreover, DFT and time-dependent (TD) DFT calcu-
lations have been carried out to shed light on the complex
photochemical scenario outlined above. Our theoretical studies
of the potential energy surfaces (PESs) of 3 and 5 for S0, S1 and
S2 have been carried out with an implicit water model, in
which the molecule is placed inside a cavity of the solvent.
Vis spectrum is also observed with an additional blueshifted
band at lꢀ320 nm. However, no spectrum of a PSS can be re-
corded due to acceleration of the thermal back reaction. In
a flash photolysis setup, this thermal relaxation of an aqueous
solution of 1 was measured after excitation at l=325 nm.
Negative absorption changes were recorded at l=365 nm,
corresponding to bleaching of the pp* band of trans-1, which
decays with t=(61.3Æ0.8) ms (see Figures S9 and S10 in the
Supporting Information). This fast thermal back reaction, as
well as the slight redshift with respect to pure AB (lmax(pp*)=
314 nm),[30] can be explained by a push–pull effect of the hy-
droxy and sulfonamide groups of compound 1; this is, howev-
er, weakened by the ÀM effect of the carboxy group. Acetyla-
tion results in a loss of this effect and decelerates the thermal
back rate to a timescale of hours. Moreover, the UV/Vis spec-
trum of the 2-trans isomer is about 32 nm hypsochromically
shifted with respect to 1. The pp* band of trans-2 can be seen
at l=328 nm with a weak np* contribution at l=435 nm.
This compound can be converted into the PSS spectrum with
an increased np* band at l=420 nm by illumination at l=
365 nm (Figure 2).
Results and Discussion
Steady-state investigations
Spectral characteristics of compound 1 are highly dependent
on the proticity of the solvent. Only about 4% of phosphate-
buffered saline (PBS) in a solution of compound 1 in DMSO
suffices to induce a considerable blueshift of the absorbance
maximum, resulting in a loss of switchability in the steady
state (data not shown), whereas isomerisation behaviour can
be detected in the range of seconds in DMSO. In detail, com-
pound 1 shows a broad and structured UV/Vis spectrum in an-
hydrous DMSO (Figure 2). The most intense band at l=
400 nm is assigned to the pp* band (S2 state) of the trans
isomer, whereas the np* band (S1 state) can barely be seen.
After illumination at l=365 nm, compound 1 shows sufficient-
ly good switching behaviour in anhydrous DMSO. Whereas the
absorption of the pp* band of trans-AB decreases, the pp*
band at l=327 nm and the np* band at l=536 nm of the cis
isomer increase. The PSS, however, contains only a small
amount of cis isomer. The thermal back reaction proceeds with
a lifetime of t=(24.6Æ0.7) s. Nevertheless, the addition of only
a small amount of PBS shifts the pp* band hypsochromically
by about 40 nm to l=360 nm; the np* contribution is now
seen at lꢀ450 nm. Similar to DMSO, in PBS a structured UV/
The UV/Vis spectrum of trans-4 shows a narrow pp* band at
l=319 nm with an np* band at l=420 nm; for the PSS, an in-
creased np* band at l=425 nm can be seen (data not shown).
The absorption bands of compound 3 are only slightly red-
shifted to l=324 nm (S2). However, in contrast to 4, an addi-
tional absorption band at lꢀ380 nm can be seen, which is as-
signed by quantum-chemical calculations to the keto species
of 3.
In the steady-state spectrum of 5, two pronounced absorp-
tion bands can be detected with absorption maxima at l=351
and 433 nm. Whereas the former band results from the pp*
transition, the intense band at l=433 nm overlaps with the
np* band. In addition to the considerable intensity at l=
433 nm, the two bands at l=351 and 433 nm are highly reac-
tive to changes in temperature; this results in different ratios
of these bands. At room temperature the ratio is A351/433 =3.3
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15721
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
(ꢀ14% keto), whereas it increases slightly to 3.7 at 48C and
drastically decreases to 1.8 at 808C. Therefore, there is an equi-
librium between the two species (keto and enol), which is
highly affected by temperature (see Figure S13 in the Support-
ing Information).
Time-resolved investigations
Time-resolved UV/Vis pump/probe experiments have been per-
formed in the femto- to picosecond timescale for all samples
in a solution of PBS/10% DMSO to detect the isomerisation dy-
namics and influence of the keto species after excitation at l=
320, 360 and 440 nm (see Figure 3, below, for transient absorp-
tion spectra of 1 and 2 after excitation at l=320 nm; see
Figure 6, below, for transient absorption spectra of 3–5 after
excitation at l=320 nm; and see Figure 10, below, for transi-
ent absorption spectrum of 5 after excitation at l=440 nm;
spectra after excitation at l=360 nm are not shown). A sum-
mary is given in Table 1.
Figure 3. Left: Transient absorption difference spectra of compounds 1 and
2 after excitation at l=320 nm; right: comparison of the absorption spectra
before excitation (black) with the transient absorption difference spectrum
(red) after 2 ns (inverse decay-associated spectra (DAS) of the longest life-
time). The dashed white lines indicate the wavelengths of the transients
shown in Figure 5. As for all transients, the delay time axis is linear for
t <1 ps and logarithmic for longer delay times.
Table 1. Lifetimes [ps] resulting from global lifetime analysis of time-re-
solved measurements of compounds 1–5 after excitation at l=320 and
440 nm.
Lifetime
lex
320 nm
440 nm
1
2
3
4
5
5
t1
t2
t3
t4
0.40
1.20
4.40
1
0.15
0.76
4.38
1
0.25
1.30
3.90
1
0.24
1.00
4.27
1
0.33
1.85
5.03
1
0.45
1.96
–
1
Sulfasalazine versus acetylated sulfasalazine
Independently of the excitation wavelength, compound
1 shows two instantaneously arising excited-state absorption
(ESA) bands at l=435 (ESA1) and 550 nm (ESA2), as well as
ground-state bleaching (GSB) below l=390 nm with lmin
=
368 nm (Figure 3).
However, the relative contribution of the ESA band to the
transient spectrum exhibits a clear dependence on the excita-
tion wavelength. The dataset can be adequately fitted with
four time constants. Their respective spectral contribution is vi-
sualised in the DAS in Figure 4. While the lifetime t1 =0.40 ps
describes the further increase of the amplitude of ESA1, a gen-
eral decrease in a broad positive band between l=390 and
600 nm can also be seen. The lifetimes t2 =1.20 ps and t3 =
4.40 ps describe the relaxation of ESA1 and ESA2. The DAS of t3
shows a sinusoidal shape with a high fit amplitude that de-
scribes not only the decay of ESA1, but also of the GSB. After
about 10 ps, compound 1 has relaxed to the ground state. The
spectrum reveals a weak product absorption band (PA) at l=
470 nm and a GSB below l=445 nm. The band of the GSB is
redshifted with respect to the absorption maximum of the
steady-state spectrum (Figure 3, right panel). This indicates
Figure 4. DAS after excitation of compounds 1 and 2 at l=320 nm for the
indicated time constants.
a return to the final spectrum beyond the 1 ns timescale,
which can be caused by slow conversion reactions.
Because the excitation at l=320 nm induces a transition
into the S2 state of 1, the broad positive absorption changes at
about 100 fs correspond to the S2!Sn ESA. With a lifetime of
0.40 ps, an ultrafast S2!S1 transition occurs, as indicated by
a decrease in the early broad ESA, which is partially compen-
sated for by an increase in the ESA1 and ESA2 bands corre-
sponding to S1!Sm,n transitions. The relaxation into the
ground state occurs biexponentially with 1.20 and 4.40 ps.
After about 10 ps, the (re-)population of the trans ground state
and the cis-1 ground state, which can be seen as the PA, is
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15722
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
complete (Figure 4). The spectral behaviour further supports
the assignment that the lifetime t3 describes the internal con-
version into the ground state. Vibrational cooling may contrib-
ute; however, no separate lifetime is necessary to describe it.
These results correspond well with earlier studies on ABs, for
example, np*-excited (4-aminomethyl)-phenyl-azobenzoic acid
(AMPB),[31–33], known for fast thermal back reactions in aqueous
solutions, or pp* excited push–pull ABs.[34]. Increased longer-
lived components have been observed for ABs in organic sol-
vents such as acetonitrile, ethanol or n-hexane.[35–37]
Until about 400 fs, a positive signal can be seen in the tran-
sients of 1 at l=393 nm, but due to spectral overlap of the
GSB with the ESA band, a negative signal results. Because no
absorption changes were detected in steady-state spectra,
photodeacetylation of 2 during the measurement, and there-
fore, the excitation of a mixture of 1 and 2 can be excluded.
In summary, significant differences between the S0!S2 tran-
sitions of compounds 1 and 2 could be observed and attribut-
ed to their spectral differences. However, they show spectral
similarities for the S1!Sn transitions and comparable dynamics
on the sub-nanosecond timescale. Therefore, we suggest a sim-
ilar electronic structure for the two compounds in the first ex-
cited state. Consequently, there is no effect of acetylation on
the isomerisation behaviour on the ultrafast timescale.
Excitation of trans-2 under the same conditions shows
roughly comparable spectral and dynamic behaviour. However,
the ESA1 band for 2 is significantly broadened in comparison
to that of 1, due to an additional band at l=393 nm (A393/437
=
0.65). Contributions at l=570 nm can be observed in addition
to the band at l=550 nm, which, however, show the same dy-
namic features and are therefore considered as a broadened
ESA2 band. The increase of the ESA1 band for 2 can be seen in
the DAS of lifetime t1, especially for the band at l=393 nm
and only slightly at l=460 nm. After about 8 ps, ESA1 is al-
ready relaxed enough to detect the GSB below l=361 nm
again. The PA, which can be seen after about 20 ps, is located
at l=428 nm and a GSB below l=400 nm is left. These contri-
butions, as well as the blueshifts of about 40 nm with respect
to 1, are in agreement with the steady-state spectra. The PA is
more pronounced for 2, which should mainly be due to the
weaker overlap of the GSB with the PA (Figure 4). By compar-
ing the ESA1 bands of 1 and 2, the absorbance changes at
lꢀ437 nm are similar (Figure 5). For 2, the contribution of the
lifetime t2 to the excited-state decay is higher than that for 1,
so the absorbance changes faster between 0.5 and 3 ps.
ortho-/meta-/para-HO-AB
After excitation of the pp* band of trans-4 at lex =320 nm
(Figure 6), two instantaneous ESA bands at l=393 (ESA1) and
510 nm (ESA2) are induced (A393/510 =2.9). Due to the same dy-
namic features of the contributions at l=550 nm as those at
l=510 nm, it is considered as a broadened ESA2 band. After
about 3 ps, ESA1 has already decayed enough to allow detec-
tion of the GSB below l=361 nm. Both ESA bands are com-
In the transient absorption spectrum (Figure 3), an additional
absorption band at l=393 nm can be seen for compound 2,
which seems to be in contrast to 1; by comparing the transi-
ents of 1 and 2 at l=393 nm (Figure 5), however, they show
comparable decay kinetics. Therefore, there is also an ESA
band for compound 1 with a spectral overlap with its GSB.
Figure 6. Left: Transient absorption difference spectra of compounds 3, 4
and 5 after excitation at l=320 nm; right: comparison of the absorption
spectra before excitation (black) with the transient absorption difference
spectrum (red) after 2 ns (inverse DAS of the longest lifetime); the time slice
for 3 is solvent-corrected. The dashed white lines indicate the wavelengths
of the transients shown in Figure 8.
Figure 5. Comparison of transients of compounds 1 and 2 at l=393 and
437 nm normalised at the absorbance maximum of each sample
(lꢀ435 nm).
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15723
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
pletely relaxed to the ground state after about 15 ps for 4, and
a PA at l=430 nm and GSB below l=400 nm remain.
photolysis measurements beginning at 45 ms. With a lifetime of
t=3.2 ms, a negative absorption change at l=329 nm de-
creases (see Figures S11 and S12 in the Supporting Informa-
tion).
Excitation of 5 also induces two slightly redshifted ESA
bands at l=400 (ESA1) and 540 nm (ESA2), with a ratio of
A400/540 = 3.1. Although the ESA1 band of 5 seems narrower
due to spectral overlap with the GSB at l=363 nm, the spec-
tral behaviour is comparable to the isomerisation behaviour of
4. After about 15 ps, the ESA bands have completely decayed
to the ground state and only a GSB below l=425 nm is left,
which is slightly redshifted relative to the steady-state spec-
trum. The PA at l=460 nm is spectrally comparable to that of
1, yet it is more pronounced. After excitation of 3 at l=
320 nm, a GSB signal below l=375 nm and two ESA bands
(ESA1 at l=416 nm and a weaker pronounced band ESA2 at
l=510 nm) arise instantaneously after excitation. In contrast
to the other HO-ABs, a third ESA band (ESA3) between l=470
and 620 nm with lmax =550 nm is populated with t1: the re-
spective decay associated spectrum (Figure 7) shows a transi-
Similar to 4 and 5, a fast S2!S1 transition occurs on a time-
scale faster than the temporal resolution of the setup, conse-
quently ESA1 and ESA2 arise “instantaneously”. These bands are
assigned to the S1!Sn transition of trans-3, which decays biex-
ponentially with t2 and t3 through a conical intersection back
to the trans-ground state and the ground state of the cis
isomer. However, the lifetime t1 describes the transition from
the ESA1 band to the S1 of the keto form (ESA3), which then
decays with t2. Therefore, either an excited-state intramolecular
proton transfer (ESIPT) occurs from the hydroxyl group (enol)
to the nitrogen of the azo group (keto), resulting in a hydra-
zone structure, or a S2!S1 transition of the keto-3 takes place.
For the latter possibility, an ESA of the S2 state would need to
overlap with the ESA band of enol-3, for which there is no indi-
cation. Although, regarding GSB2, not only was trans-3 excited,
but also the keto form; this is in agreement with results from
theoretical calculations, which indicate the coexistence of both
species in the ground state. This results in ESA3 in the first
place, which is then enhanced by the ESA1!ESA3 transition
(Figure 8).
Figure 8. Comparison of transients of compounds 3, 4 and 5 at l=403 and
550 nm, normalised at the absorbance maximum of each sample
(lꢀ400 nm); solvent correction for 3 was performed as described in the Ex-
perimental Section.
Figure 7. DAS after excitation of compounds 3, 4 and 5 at l=320 nm corre-
sponding to the indicated time constants.
tion from ESA1 to ESA3. Therefore, ESA1 decays with three time
constants (t1–t3), whereas the decay of ESA2 is described by
two time constants (t2 and t3). The ESA3 band decays exclu-
sively with a time constant of t2 =1.26 ps. After about 10 ps,
ESA1 and ESA2 have decayed to the ground state and also the
GSB below l=365 nm decreases. However, a second GSB2 at
l=380 nm remains unchanged. Only small changes, barely
above the signal-to-noise ratio, can be observed in the range
of ESA3 and the GSB2 region after 10 ps. One positive PA at l=
515 nm remains after decay of the excited-state bands at
about 10 ps until at least 1.5 ns, but cannot be seen in flash
Although for excitation at lex =320 nm the ratio of A410/550 is
1.9 and slightly smaller for lex =440 nm (1.3), at which the np*
band is excited, for lex =360 nm the ESA3 band is more pro-
nounced (<1.0) due to increased direct keto- or decreased
trans-3 excitation.
The transients of 4 and 5 at l=403 and 500 nm (Figure 8)
show similar dynamics. Until about 2 ps, nearly no differences
can be observed, but afterwards 5 relaxes slightly slower. How-
ever, compound 3 differs not only in the contributions of ESA1,
but also in the ESA2/3 bands. At l=403 nm, the positive ab-
sorption change decays significantly faster with a pronounced
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15724
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
component on the 100 fs timescale. Nevertheless, complete re-
laxation to the ground state occurs on a comparable timescale
(ꢀ10 ps) as those for 4 and 5. At l=550 nm, a slight increase
in the positive signal can be seen before 350 fs, whereas 4 and
5 gain full intensity after 180 fs. Moreover, compound 3 exhib-
its a significantly higher intensity, which already reduces below
a delay time of 1 ps. After that time, the decay is very steep
and remains unchanged after about 6 ps, leaving a small posi-
tive signal. This positive absorption arises, on one hand, from
trans–cis isomerisation, showing the np* band of the cis-3,
whereas, on the other hand, it results from an additional path-
way via the keto form and can either be assigned to keto-3
(trans- (TK) or cis-keto (CK)) in the ground or a triplet states,
due to its redshift and longevity, or alternatively to the np*
band of cis-3. However, due to the spectral proximity of the
isomers involved, a statement about the product of the addi-
tional pathway cannot be made unambiguously.
Consequently, isomerisation behaviour similar to that of un-
substituted AB could be observed for compounds 4 and 5,
whereas compound 3 showed
energy barrier of 0.2 eV. This indicates that both species might
coexist in solution. The trans–cis isomerisation around the C2-
N1-N2-C3 bond of the TE species requires surmounting an
energy barrier of 2 eV, resulting in a cis-enol (CE) conformation
that is 0.9 eV higher in energy than that of TE. Conversely,
trans–cis isomerisation of the keto tautomers involves a some-
what lower barrier (1.4 eV) and leads to a CK form at 0.8 eV
above TK. In addition, we have explored internal rotation
around the N2-N1-C2-C1 dihedral angle, following a procedure
reported by Cui et al.,[38] in search of additional deactivation
paths between the ground and excited electronic states, lead-
ing to an alternate TK species (labelled as TK1), 0.4 eV above
TK, with an energy barrier of 1.6 eV. These results are summar-
ised in Figure 9a–d. Turning to the S1 and S2 potential surfaces
calculated by TD-DFT (Figure 9b), in S1, TE is separated from
TK by an energy barrier of 0.3 eV. In S2, there is no energy mini-
mum for TE, which suggests a barrier-less process starting
from the S2 Franck–Condon geometry. As illustrated in Fig-
ure 9a and b, two deactivation pathways can be envisaged,
an additional excited-state band
at l=550 nm, which was attrib-
uted to the keto species. This
state is populated by the enol
species with a lifetime of 250 fs,
which corresponds to an ESIPT
process. Additionally, isomerisa-
tion to cis-3 occurs, due to addi-
tional deactivation paths; how-
ever, its efficiency is reduced.
According to experimental ob-
servations, the photochemical
activity of 3 is not dependent on
the polarity of the solvent. Very
similar spectroscopic signals are
measured in different solvents,
such as DMSO or water. Apart
from trans–cis isomerisation and
other geometrical deformations
in AB derivatives, ESIPT is a well-
known deactivation mechanism
for many photochromic species.
Because the ortho substitution
of the hydroxy group allows an
ESIPT process without participa-
tion of solvent molecules, we
consider the O1ÀH1 distance to
be one of the most relevant co-
ordinates for understanding the
photochemistry of compound 3.
Exploration of the electronic
ground state (S0) at the DFT
Figure 9. Photochemical landscape of 3, as inferred from our TD-DFT results. Energies (in eV) in the ground state
level leads to a planar trans-enol (S0, in black), first singlet excited state (S1, in dark grey) and second singlet excited state (S2, in light grey) are refer-
enced to the TE minimum in S0. Energy profiles in S0 and S1 are fully relaxed along the corresponding reaction co-
(TE) conformation of 3 as the
ordinate, whereas the energy profile in S2 corresponds to vertical energies from S1. The grey dashed arrow indi-
most stable tautomer. The corre-
cates connecting points between panels b) and d). The black arrow corresponds to the starting photoexcitation
sponding TK tautomer lies only
(Franck–Condon) point of 3. Dashed lines indicate interpolation in non-adiabatic coupling regions that cannot be
0.1 eV above TE, separated by an described by TD-DFT.
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15725
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
namely, ESIPT (TE–TK conversion) and trans–cis isomerisation in
the enol form (TE–CE conversion). ESIPT is favoured by the
proximity of the S1 and S2 PESs, which run parallel to each
other along the proton transfer coordinate, showing an energy
gap of <0.1 eV. This is indicative of an efficient, ultrafast deac-
tivation path between S2 and S1. By contrast, the PES cut along
the trans–cis isomerisation coordinate (Figure 9a) shows an S2
barrier of 0.6 eV and an S2–S1 energy gap of around 1.0 eV
(which could decrease if additional coordinates, leading to an
S1–S2 conical intersection, were included). Transitions between
S1 and S0 are seen to be mediated by the torsional coordinates
C2-N1-N2-C3 and N2-N1-C2-C1 (Figure 9a, c, and d). Assuming
that TE–TK conversion is favoured in the initial S2–S1 dynamics
(see above), subsequent trans–cis isomerisation of the TK spe-
cies encounters an energy gap of 0.6 eV between S1 and S0
(Figure 9d), whereas torsion around the N1-N2-C2-C3 angle re-
sults in a smaller gap of 0.3 eV (Figure 9c). In combination,
these two torsional coordinates could indicate a hula-twist
mechanism.[39] Although TD-DFT is not able to provide a reliable
description of the non-adiabatic coupling region, and specifi-
cally of the conical intersection topology,[40] related studies
have shown that a qualitatively correct picture can be ob-
tained for the photochemical pathway leading on the S1 state
towards the intersection, and its continuation on the S0
state.[41–43] The energy gap of 0.3 eV is small enough to consid-
er the presence of a conical intersection between these two
states, and therefore, an ultrafast and efficient deactivation
path.
Figure 10. Left: Transient absorption difference spectrum of compound 5
after excitation at l=440 nm; right: comparison of the absorption spectrum
before excitation (black) with the transient absorption spectrum (red) after
2 ns (inverse DAS of longest lifetime). The dashed white lines indicate the
wavelengths of the transients shown in Figure 11.
completely different behaviour than that with lex =320 nm
(Figure 10).
Excitation at lex =440 nm clearly addresses a different ab-
sorption band (Figure 2). Instantaneously, after excitation,
three ESA bands arise at l=383 nm (ESA1) and above l=
480 nm, at which two ESA bands can be distinguished at l=
510 nm (ESA2) and a less pronounced band at l=570 nm
(ESA3). Although ESA1 and ESA3 decay quickly with a lifetime of
t1, the decays of the GSB ESA2 occur on a longer timescale.
With a lifetime of t2, a decrease in the GSB band below l=
400 nm and at l=452 nm is described, in addition to the
decay of the ESA2 band and a positive signal at l=400 nm,
which is, however, overlaid by the negative signals. After
about 10 ps, the excited states have completely decayed to
the ground state and leave a positive absorption band at l=
510 nm and a GSB below l=490 nm (lmin ꢀ430 nm). The spec-
tral and dynamic characteristics can neither be explained by
excitation of trans-5 nor of cis-5; thus, a different species has
to be responsible for this behaviour. Because this species can
indeed be seen in the steady-state spectrum (Figure 2), we
suggest that 5 is partially present in its keto form, which is not
unknown for HO-ABs in protic solvents.[27]
From the above discussion, the overall scenario of the pho-
tochemistry of 3 would involve the following sequence of
events: initial excitation of the TE tautomer to S2 is likely to be
followed by ultrafast ESIPT on the coupled S2 and S1 surfaces,
generating the TK species on S1 (Figure 9b). Although trans–cis
isomerisation of the TE species through S2–S1–S0 cannot be ex-
cluded, this path is energetically less favoured (Figure 9a).
Next, the TK species undergoes passage from S1 to S0 through
the CNNC and/or NNCC torsional coordinates (Figure 9c and
d), presumably leading to a mixture of CK and TK species. The
latter is more stable by 0.8 eV than the former.
Therefore, at l=440 nm, predominantly keto-5 is excited
and results in a GSB and ESA band at l=510 nm (which corre-
sponds to the S1!Sn transition). However, this cannot explain
ESA1 and ESA3, which we attribute to an additional minor exci-
tation of trans-5 in the np* band (Figure 11). Due to the low
extinction coefficient of the np* band, the dynamics of the
As mentioned above, the TK species can easily revert to the
initial TE reactant by enol/keto tautomerism in S0; thus closing
the photocycle, as observed experimentally. Conversely, the ini-
tial photoexcited state should more accurately be described as
a mixture of TE and TK, and the above discussion can be
straightforwardly adapted to
that perspective.
Excitation wavelength depend-
ence of para-HO-AB
Although
wavelength-depen-
and
dent excitation of
1
3
showed mainly differences in the
intensity of the absorbance
changes in the ESA band at
lꢀ550 nm, excitation of com-
Figure 11. Comparison of transients of compound 5 at l=452 and 497 nm (left) and DAS of 5 (right) after excita-
pound 5 at lex =440 nm leads to tion at l=440 nm.
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15726
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
keto species dominate. Nevertheless, trans–cis isomerisation
does occur. The resulting cis-np* band at l=510 nm, however,
is too intense, relative to the small GSB of trans-5 at l=
358 nm (DAS(t4): for lex =440 nm: A520/361 =0.47; for lex =
320 nm: A520/361 =0.03), so we propose additional formation of
the cis-5 isomer (CE or CK) from the excited keto form through
a conical intersection with lifetime t2.
ed in water on that timescale due to a fast back reaction.
Hence, the solvent is expected to play a key role in the deacti-
vation paths of compound 5. Although the para substitution
of the hydroxy group rules out an ESIPT process, an intermo-
lecular proton transfer process mediated by solvent molecules
is possible. To model this scenario, we have included two
types of models in our electronic structure calculations of 5:
The first model includes a chain of seven water molecules that
connect the azo bond with the hydroxy group of AB (Fig-
ure 12B). The second, simpler model includes two water mole-
cules at a hydrogen-bonding distance to the azo and hydroxy
groups, respectively. The first water model was used to identify
possible proton transfer along the water chain from the hy-
droxy group to the azo group through a concerted mecha-
nism. In these models, the optimised ground-state structure
again has a planar form for TE and TK, of which TE is slightly
more stable than TK by 0.1 eV.
In summary, it is possible to address both species of com-
pound 5 (keto and enol) independently by changing the exci-
tation wavelength. Because at room temperature there is still
86% enol species, excitation at l=440 nm results in a mixture
of excitation of the keto (pp* band) and enol (np* band). It
should be noted that, by changing the temperature, however,
the thermal equilibrium, and therefore, ratio of excited keto
and enol species can be shifted significantly (Figure S13 in the
Supporting Information).
As opposed to compound 3, the photochemistry of com-
pound 5 is strongly dependent on the physical properties of
the solvent. Whereas the isomerisation of 5 is observable on
a timescale of seconds in DMSO, no conversion can be detect-
For 5, the barrier between 5-TE and 5-TK strongly depends
on the solvent. For the simple water chain model shown in
Figure 12, proton transfer across the chain is feasible, but in-
Figure 12. A) Photochemical landscape of compound 5, as inferred from our electronic structure results. Energies (in eV) in the ground state (S0, in black), first
singlet excited state (S1, in dark grey) and second singlet excited state (S2, in light grey) are referenced to the TE minimum in S0. Energy profiles in S0 and S1
are fully relaxed along the corresponding reaction coordinate, whereas the energy profile in S2 corresponds to vertical energies from S1. The grey dashed
arrow indicates connecting points between panels a)–d). The black arrow indicates the starting photoexcitation (Franck–Condon) point of 5. As in Figure 9,
dashed lines indicate interpolation in non-adiabatic coupling regions that cannot be described by TD-DFT. B) The reaction coordinates for 5 and the water
chain model.
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15727
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
volves an artificially high barrier of about 2 eV (Figure 12b).
This barrier could be overcome by supportive OÀH frequencies
(see Figure S14 in the Supporting Information). Frequency cal-
culations for S0, S1 and S2 show a marked interplay of modes
involving water molecules and the azo and hydroxy groups;
this supports possible proton transfer from the hydroxy group
to the azo group through the water chain. The relevant fre-
quencies are around 3000 to 3500 cmÀ1 (0.3–0.5 eV) for S0 and
can be accessed by participation of the first solvent shell (see
Figure S13 in the Supporting Information). The trans–cis iso-
merisation of TE in the S0 ground state is hindered by a high
barrier of about 1.9 eV, leading to a stable CE conformer at
0.6 eV. The trans–cis isomerisation of the TK form exhibits
a lower barrier of around 0.9 eV, leading to the CK conforma-
tion at 0.5 eV. The reverse cis–trans isomerisation therefore has
a reduced barrier of around 0.4 eV. Internal rotation around
the N2-N1-C2-C1 dihedral angle is also a possible pathway for 5-
TK, with a barrier height of 1.4 eV, which is lower than that for
3-TK. For S2 and S1, TD-DFT calculations also show a high barri-
er between TE and TK, although this is somewhat reduced
compared with the ground
(1) showed no apparent photoswitching behaviour in PBS. In
anhydrous DMSO photoswitching can be achieved. However,
this effect is prevented by the addition of only 4% PBS to this
solution.
Acetylation (to form 2) afforded a photoswitchable deriva-
tive in PBS. This observation suggested that (photo)tautomer-
ism might be the reason for this behaviour. In pump–probe ex-
periments on the ultrafast timescale, we found no significant
difference in the time-dependent behaviour of compounds
1 and 2, which showed that the relevant processes must have
occurred on a longer timescale than 2 ns. We confirmed that
sulfasalazine in the ground state was present predominantly as
the tautomer shown in Figure 1. After photoexcitation, tauto-
merism in which the resulting keto form had an NÀN single
bond occurred to a significant degree, leading to fast relaxa-
tion to the TE ground state.
To investigate this effect further, we performed a compara-
tive study on the three HO-ABs 3, 4 and 5 and revealed molec-
ular details of their photodynamics (Figure 13). The influence
of the hydroxy group is highly position specific and modulates
state. Possible state crossing be-
tween S2 and S1 indicates that
solvent-mediated passage to S1
through proton transfer is, in
principle, feasible. On the S1 PES,
trans–cis isomerisation proceeds
in a similar fashion for the TK
and TE forms, possibly involving
participation of the NNCC tor-
sion, as discussed above.
Deactivation pathway of 5
Following photoexcitation to S2,
compound 5-TE preferentially
decays by trans–cis isomerisation
to CE (Figure 12Aa), due to the
high barrier of competing enol/
keto tautomerism (Figure 12Ab).
As mentioned above, this barrier
might be significantly lowered,
Figure 13. Reaction pathways of HO-ABs in a simplified potential energy scheme. Black arrows indicate the iso-
merisation mechanism of compounds 1–5. Grey arrows show additional pathways for 3 (dark) and 5 (light). Chem-
ical structures of the respective species are shown for 5 as an example (bottom).
though, in an improved elec-
tronic structure description, such
that both isomerisation and tau-
tomerism would be feasible.
Likewise, a photoexcited 5-TK species would preferentially
adopt the trans–cis isomerisation channel towards CK (see Fig-
ure 12Ad). Alternatively, the N2-N1-C2-C1 torsion is feasible (see
Figure 12Ac). In the electronic ground state (S0), the barrier for
CK-TK back transfer is lower than that of the tautomerism bar-
rier.
the individual PESs differently. The meta-substituted derivative
4 showed typical AB-like photochemical behaviour due to the
absence of prominent tautomerism pathways. The cis isomer
was stable for several hours. The ortho-substituted derivative 3
reacted along two pathways upon excitation: Part of the en-
semble of molecules showed normal AB photochemistry with
a thermal back reaction to the TE ground state after 3 ms. Be-
cause the two additional reaction coordinates (C2-N1-N2-C3 and
N2-N1-C2-C1), resulting in TK and TK1, showed only small energy
differences between S1 and S0, this could enable isomerisation
through a hula-twist mechanism for ortho substitution. Anoth-
Conclusion
In the search for a photoswitchable riboswitch ligand, we real-
ised that the commercially available antipyretic sulfasalazine
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15728
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
corded by using a Specord S600 instrument with a Peltier cooled
cell holder with stirrer and external heat exchange.
er deactivation pathway through trans–cis isomerisation along
the CNNC torsion on the S1 state was suggested. The other
part of the excited ensemble performed tautomerism in about
250 fs by OH bond stretching (ꢀ0.3 Š) and then relaxation to
the ground state in 10 ps. The behaviour of para-substituted
compound 5 (which cannot tautomerise intramolecularly) was
more complex: In the ground state, a temperature-dependent
equilibrium of keto and enol species could be observed, as
supported by theoretical calculations; hence, these two species
could be selectively excited. Excitation of the enol form again
afforded AB-like behaviour with relaxation to the trans form
within milliseconds. Excitation of the keto form, however, re-
sulted in completely different photochemistry; the transient
absorption spectrum reflected the conversion of the excited
keto-5 into the cis form.
ABs are important photoswitches that have now been used
for several decades to control many important processes in
chemical biology and material sciences with light. Knowledge
of the pitfalls by a deeper mechanistic understanding of this
system will help to make this endeavour an even greater suc-
cess. With this study, we wanted to contribute to the under-
standing of why certain AB show no apparent switching on
the timescale of seconds and suggest that, for example, simple
acetylation can afford photoswitching complex AB derivatives,
which could then be good candidates for the development of
the corresponding aptamer and later a light-responsive ribo-
switch.
Illumination procedure
For the illumination of the HO-ABs, a Thorlabs light-emitting diode
(LED) Driver DC4100 and DC4100-HUB was used. UV light for
trans–cis isomerisation was provided by a l=365 nm LED and visi-
ble light for cis–trans isomerisation by a l=420 nm LED (Thorlabs
M365L2 and M420L2, 300 mW).
Time-resolved UV/Vis pump–probe setup
Time-resolved measurements were performed in a self-built UV/Vis
pump–probe setup. For the generation of ultrashort laser pulses
(150 fs), a Ti:sapphire femtosecond laser system from Clark, MXR-
CPA-2001, was used with a central wavelength of l=775 nm and
a repetition rate of 1 kHz. Excitation pulses were generated in
a non-collinear optical parametric amplifier (NOPA) and a subse-
quent sum frequency mixing process. For an excitation wavelength
of l=320 nm, pulses with a wavelength of l=545 nm were mixed
with a l=775 nm fundamental pulse in a BBO crystal (q=368); for
lmax =360 nm a NOPA pulse of l=672 nm and a q=30.28 BBO
crystal were used and for lmax =440 nm a NOPA pulse of 1018 nm
and a q=268 BBO crystal were used, resulting in an energy of
about 50 nJ. For the probe pulse, a fundamental pulse was guided
into a 5 mm CaF2 window to generate white light, which was fo-
cused together with the excitation pulse into the sample. Polarisa-
tion between the pump and probe pulse was adjusted to an angle
of 54.78 between the two pulses to eliminate anisotropic effects.
The cuvette was rotated in the x,y plane and illuminated by a l=
420 nm LED during measurements to provide the initial state of
the sample. The averaged temporal resolution was in the range of
about 220 fs. Difference spectra were calculated by blocking every
second excitation pulse. Transient absorption data were analysed
by global lifetime analysis (GLA)[44] by using up to four exponential
functions. The coherent artefact contribution around time zero was
approximated with a Gaussian function and its first and second de-
Experimental Section
Sample preparation
Compounds 2, 3, 4, 5 were prepared by using reported proce-
dures. For comparability, all samples were dissolved in DMSO and
diluted with PBS (pH 7.4) to a content of 10% DMSO (ROTIPURAN
ꢁ99.8%, p.a.) in solution. Although most of the samples are solu-
ble to an acceptable level, DMSO addition was necessary for the
solubility of 3. For steady-state measurements, about 50 nmol
(35 mm, 1.5 mL) was necessary and for flash photolysis about
5 nmol (35 mm, 150 mL) was necessary, whereas for time-resolved
studies 90 nmol (350 mm, 250 mL) was used, except for 3. Due to its
low solubility in water, only 12.5 nmol (50 mm, 250 mL) could be
prepared for 3. Steady-state measurements were performed in
a 1”1 cm quartz fluorescence cuvette equipped with a 4 mm stir-
ring bar, which allowed simultaneous illumination during measure-
ments. Time-resolved measurements were performed in a 1 mm
quartz cuvette. For flash photolysis experiments, a 1”0.2 cm
quartz fluorescence cuvette was used.
rivatives.[45] The analysis was performed by using OPTIMUS.[46]
.
Time-resolved spectra were solvent-corrected by subtracting the
scaled transient data of the solvent.
Flash photolysis
An excitation pulse with a wavelength of l=320 nm (pulse width:
20 ns, intensity: 0.7 mJcmÀ2) was obtained by second-harmonic
generation of the pulses of an optical parametric oscillator (OPO;
GWU-Lasertechnik Vertriebsges. mbH), which was pumped by
a Nd:YAG laser (Spitfire 600 Innolas Laser GmbH). For broadband
detection, a charge-coupled device (CCD) camera (PI-MAX3, Prince-
ton Instruments) and a spectrograph (Acton SP2150, Princeton In-
struments) were used. For difference spectra, every second excita-
tion pulse was blocked.
Experimental setup
Theoretical calculations
NMR spectra were recorded on Bruker AV 500 MHz spectrometer
by using [D6]DMSO as a solvent. HRMS spectra were recorded by
using a Thermo Scientific MALDI LTQ Orbitrap instrument.
DFT methods were employed to explore the topology of the
ground (S0), first singlet excited (S1) and second singlet excited (S2)
electronic states. Ground-state DFT calculations were performed in
S0, whereas TD-DFT[47] was used in S1 and S2. TD-DFT has been suc-
cessfully applied to study other related photochromic and ESIPT
systems. The three-parameter hybrid functional of Becke with the
correlation functional of Lee, Yang and Parr (B3-LYP) has been
used.[48,49] The Weigend and Ahlrichs def2-TZVP basis set was con-
Steady-state absorption measurements
Absorption spectra were recorded by using a Specord S100 UV/Vis
spectrometer. Temperature-dependent absorption spectra were re-
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15729
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
sidered,[50] which included polarisation functions for all atoms. All
calculations were carried out by using the Gaussian 09 package.[51]
For the ground electronic state, the geometries of all stationary
points (minima and transition states) were localised upon direct
optimisation. In the first excited electronic state, the minima were
also directly minimised, whereas the transition states were located
as maxima of reaction coordinates. The second singlet excited
state was constructed as vertical single-point calculations from the
relaxed structures of S1. For the reaction coordinates, the O1ÀH1
distance was used for the intramolecular proton transfer process in
compound 3, whereas the H1ÀO2 and H3ÀN2 distances were used
for compound 5. For both molecules, the C2-N1-N2-C3 and N2-N1-C2-
C1 dihedral angles were used for internal rotations (see Figure 9 for
atom numbering). The calculations were first carried out for the
isolated molecule (gas phase) and later with solvent by using the
polarisable continuum model (PCM),[52–54] in which the HO-ABs
were placed inside a cavity of implicit water solvent. All optimisa-
tions and reaction coordinates were repeated inside the solvent
cavity. For compound 5, explicit water molecules were used, as ob-
tained from a molecular dynamics trajectory of 5 in a cubic box of
TIP3P water molecules[55] by using the Amber 10 package[56] and
Generalized Amber Force Field.[57]. After 10 ns of equilibration at
a constant pressure and temperature of 1 atm and 300 K, respec-
tively, we manually selected the smallest cluster of water molecules
that connected the ÀOH donor and ÀN acceptor groups of com-
pound 5. Stationary points were identified through frequency cal-
culations in the ground and first singlet excited electronic states.
Synthesis of 4
Nitrosobenzene (0.50 g, 4.67 mmol) was dissolved in acetic acid
(30 mL). 3-Aminophenol (0.51 g, 4.67 mmol) was added to the
green solution and the mixture was stirred for 24 h under an
argon atmosphere. The solvent was removed under pressure and
the product was isolated by column chromatography (CH/EE 9:1)
1
to yield 4 (0.67 g, 3.41 mmol, 73%). H NMR (500 MHz, [D6]DMSO):
d=9.87 (s, 1H; -OH), 7.88–7.86 (m, 2H; Har), 7.60–7.55 (m, 3H; Har),
7.40–7.39 (m, 2H; Har), 7.27–7.26 (m, 1H; Har), 6.99–6.96 ppm (m,
1H; Har); 13C NMR (125.8 MHz, [D6]DMSO): d=158.3, 153.2, 151.9,
131.5, 130.2, 129.5, 122.5, 118.8, 115.5, 107.2 ppm; HRMS (MALDI):
m/z calcd for C12H10N2O [M]+: 199.08659; found: 199.08651 (Dm
0.00008, error 0.40 ppm).
Synthesis of 5
Nitrosobenzene (0.50 g, 4.67 mmol) was dissolved in acetic acid
(30 mL). 4-Aminophenol (0.51 g, 4.67 mmol) was added to the
green solution and the mixture was stirred for 24 h under an
argon atmosphere. The solvent was removed under pressure and
the product was isolated by column chromatography (CH/EE 9:1)
1
to yield 5 (0.73 g, 3.69 mmol, 79%). H NMR (500 MHz, [D6]DMSO):
d=10.31 (s, 1H; -OH), 7.82–7.80 (m, 4H; Har), 7.57–7.54 (m, 2H;
Har), 7.50–7.49 (m, 1H; Har), 6.96–6.94 ppm (m, 2H; Har); 13C NMR
(125.8 MHz, [D6]DMSO): d=161.0, 152.1, 145.2, 130.5, 129.3, 124.9,
122.1, 115.9 ppm; HRMS (MALDI): m/z calcd for C12H10N2O [M]+:
199.08659; found: 199.08667 (Dm 0.00008, error 0.40 ppm).
Acknowledgements
Synthesis of 2
Sulfasalazine (0.50 g, 1.26 mmol) was added to acetic anhydride
(3 mL) and the reaction mixture was heated for 3 h. The solvent
was removed under pressure and the product was isolated by
column chromatography (cyclohexane (CH)/ethyl acetate (EE) 1:1)
We thank the Deutsche Forschungsgemeinschaft (DFG) for fi-
nancial support through SFB 902 “Molecular Principles of RNA-
Based Regulation”.
1
to yield 2 (0.27 g, 0.56 mmol, 45%). H NMR (500 MHz, [D6]DMSO):
d=13.55 (s, 1H; -COOH), 8.71–8.69 (m, 1H; Har), 8.46–8.45 (d, J=
2.4 Hz, 1H; Har), 8.29–8.22 (m, 3H; Har), 8.16–8.11 (m, 3H; Har), 7.79–
7.78 (d, J=7.8 Hz, 1H; Har), 7.66–7.63 (m, 1H; Har), 7.51–7,49 (d, J=
8.6 Hz, 1H; Har), 2.31 (s, 3H; -CH3), 1.81 ppm (s, 3H, -CH3); 13C NMR
(75.5 MHz, [D6]DMSO): d=169.6, 169.0, 164.9, 154.4, 152.8, 150.0,
149.4, 149.2, 140.2, 139.8, 130.1, 128.1, 126.0, 125.50, 125.45,
125.43, 125.3, 123.0, 24.4, 20.8 ppm; HRMS (MALDI): m/z calcd for
C22H18N4O7SK [M+K]+: 521.05278; found: 521.05222 (Dm 0.00056,
error 0.11 ppm).
Keywords: azobenzenes · tautomerism · photochemistry ·
time-resolved spectroscopy · water chemistry
´
[1] W. Szymanski, J. M. Beierle, H. A. V. Kistemaker, W. A. Velema, B. L. Ferin-
ga, Chem. Rev. 2013, 113, 6114–6178.
[2] C. Brieke, F. Rohrbach, A. Gottschalk, G. Mayer, A. Heckel, Angew. Chem.
Int. Ed. 2012, 51, 8446–8476.
[3] G. Mayer, A. Heckel, Angew. Chem. Int. Ed. 2006, 45, 4900–4921; Angew.
Chem. 2006, 118, 5020–5042.
[4] Y. Zhao, T. Ikeda, Smart Light-Responsive Materials: Azobenzene-Contain-
ing Polymers and Liquid Crystals, Wiley, 2009.
[5] H. M. D. Bandara, S. C. Burdette, Chem. Soc. Rev. 2012, 41, 1809–1825.
[6] A. A. Beharry, G. A. Woolley, Chem. Soc. Rev. 2011, 40, 4422–4437.
[7] G. S. Kumar, D. C. Neckers, Chem. Rev. 1989, 89, 1915–1925.
[8] Z. Yu, S. Hecht, Angew. Chem. Int. Ed. 2013, 52, 13740–13744.
[9] A. Kçhntopp, A. Dabrowski, M. Malicki, F. Temps, Chem. Commun. 2014,
50, 10105–10107.
[10] S. K. Yesodha, C. K. S. Pillai, N. Tsutsumi, Prog. Polym. Sci. 2004, 29, 45–
74.
[11] J. Wachtveitl, A. Zumbusch, Chembiochem 2011, 12, 1169–1170.
[12] J. Bredenbeck, J. Helbing, A. Sieg, T. Schrader, W. Zinth, C. Renner, R.
Behrendt, L. Moroder, J. Wachtveitl, P. Hamm, Proc. Natl. Acad. Sci. USA
2003, 100, 6452–6457.
[13] S. Spçrlein, H. Carstens, H. Satzger, C. Renner, R. Behrendt, L. Moroder,
P. Tavan, W. Zinth, J. Wachtveitl, Proc. Natl. Acad. Sci. USA 2002, 99,
7998–8002.
[14] U. Kusebauch, S. A. Cadamuro, H.-J. Musiol, M. O. Lenz, J. Wachtveitl, L.
Moroder, C. Renner, Angew. Chem. Int. Ed. 2006, 45, 7015–7018; Angew.
Chem. 2006, 118, 7170–7173.
Synthesis of 3
Nitrosobenzene (0.50 g, 4.67 mmol) was dissolved in acetic acid
(30 mL). 2-Aminophenol (0.51 g, 4.67 mmol) was added to the
green solution and the mixture was stirred for 24 h under an
argon atmosphere. The solvent was removed under reduced pres-
sure and the product was isolated by column chromatography
1
(CH/EE 40:1) to yield 3 (0.47 g, 2.39 mmol, 51%). H NMR (500 MHz,
[D6]DMSO): d=11.18 (s, 1H; -OH), 7.99–7.97 (m, 2H; Har), 7.77–7.75
(dd, J=8.05, 1H; Har), 7.61–7.55 (m, 3H; Har), 7.44–7.41 (m, 1H; Har),
7.08–7.06 (m, 1H; Har), 7.03–6.99 ppm (m, 1H; Har); 13C NMR
(125.8 MHz, [D6]DMSO): d=154.4, 151.4, 138.3, 133.6, 131.3, 129.4,
123.2, 122.6, 119.8, 118.2 ppm; HRMS (MALDI): m/z calcd for
C12H10N2O [M]+: 199.08659; found: 199.08654 (Dm 0.00005, error
0.25 ppm).
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15730
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[15] T. Goldau, K. Murayama, C. Brieke, S. Steinwand, P. Mondal, M. Biswas, I.
Burghardt, J. Wachtveitl, H. Asanuma, A. Heckel, Chem. Eur. J. 2015, 21,
2845–2854.
[16] H. Ito, X. Liang, H. Nishioka, H. Asanuma, Org. Biomol. Chem. 2010, 8,
5519–5524.
[17] M. A. Kienzler, A. Reiner, E. Trautman, S. Yoo, D. Trauner, E. Y. Isacoff, J.
Am. Chem. Soc. 2013, 135, 17683–17686.
[18] R. H. Kramer, D. L. Fortin, D. Trauner, Curr. Opin. Neurobiol. 2009, 19,
544–552.
[42] J. M. Ortiz-Sµnchez, R. Gelabert, M. Moreno, J. M. Lluch, J. M. Anglada,
J. M. Bofill, Chem. Eur. J. 2010, 16, 6693–6703.
[43] J. M. Ortiz-Sµnchez, R. Gelabert, M. Moreno, J. M. Lluch, ChemPhysChem
2010, 11, 3696–3703.
[44] I. H. M. van Stokkum, D. S. Larsen, R. van Grondelle, Biochim. Biophys.
Acta Bioenerg. 2004, 1657, 82–104.
[45] S. A. Kovalenko, A. L. Dobryakov, J. Ruthmann, N. P. Ernsting, Phys. Rev.
A 1999, 59, 2369–2384.
[46] C. Slavov, H. Hartmann, J. Wachtveitl, Anal. Chem. 2015, 87, 2328–2336.
[47] R. Bauernschmitt, R. Ahlrichs, Chem. Phys. Lett. 1996, 256, 454–464.
[48] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[19] N. J. Dunn, W. H. Humphries, A. R. Offenbacher, T. L. King, J. A. Gray, J.
Phys. Chem. A 2009, 113, 13144–13151.
´
[20] J. Dokic, M. Gothe, J. Wirth, M. V. Peters, J. Schwarz, S. Hecht, P. Saal-
[49] C. Lee, W. Yang, R. Parr, Phys. Rev. 1988, 37, 785–789.
[50] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297–3305.
[51] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant,
J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G.
Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, V.
Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev,
A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Moroku-
ma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dap-
prich, A. D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford,
J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayak-
kara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C.
Gonzalez, J. A. Pople, Gaussian 09, Revision D.01, Gaussian, Inc., Walling-
ford CT, 2009.
frank, J. Phys. Chem. A 2009, 113, 6763–6773.
[21] M. Kojima, S. Nebashi, K. Ogawa, N. Kurita, J. Phys. Org. Chem. 2005, 18,
994–1000.
[22] J. Garcia-Amorós, M. Díaz-Lobo, S. Nonell, D. Velasco, Angew. Chem. Int.
Ed. 2012, 51, 12820–12823.
[23] J. Garcia-Amorós, A. Sµnchez-Ferrer, W. A. Massad, S. Nonell, D. Velasco,
Phys. Chem. Chem. Phys. 2010, 12, 13238–13242.
[24] L. Antonov, V. Deneva, S. Simeonov, V. Kurteva, D. Nedeltcheva, J. Wirz,
Angew. Chem. Int. Ed. 2009, 48, 7875–7878; Angew. Chem. 2009, 121,
8015–8018.
[25] J.-M. SavØant, C. Tard, J. Am. Chem. Soc. 2014, 136, 8907–8910.
[26] G. Gabor, E. Fischer, J. Phys. Chem. 1962, 66, 2478–2481.
[27] J. García-Amorós, D. Velasco, Beilstein J. Org. Chem. 2012, 8, 1003–1017.
[28] B. R. Hsieh, D. DØsilets, P. M. Kazmaier, Dye. Pigment. 1990, 14, 165–189.
[29] M. Sahu, M. B. Saha, S. C. Bera, Photochem. Photobiol. 1995, 89, 19–24.
[30] J. Griffiths, Chem. Soc. Rev. 1972, 1, 481–493.
[31] C. Renner, L. Moroder, ChemBioChem 2006, 7, 868–878.
[32] H. Satzger, C. Root, C. Renner, R. Behrendt, L. Moroder, J. Wachtveitl, W.
Zinth, Chem. Phys. Lett. 2004, 396, 191–197.
[33] J. Wachtveitl, S. Spçrlein, H. Satzger, B. Fonrobert, C. Renner, R. Beh-
rendt, D. Oesterhelt, L. Moroder, W. Zinth, Biophys. J. 2004, 86, 2350–
2362.
[52] S. Miertus, E. Scroco, J. Tomasi, J. Chem. Phys. 1981, 55, 117–129.
[53] R. Improta, V. Barone, G. Scalmani, J. Frisch, J. Chem. Phys. 2006, 125,
0541031–0541039.
[54] R. Improta, G. Scalmani, M. J. Frisch, V. Barone, J. Chem. Phys. 2007, 127,
07450401–07450409.
[34] J. Bahrenburg, K. Rçttger, R. Siewertsen, F. Renth, F. Temps, Photochem.
Photobiol. Sci. 2012, 11, 1210–1219.
[55] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, M. L. Klein,
J. Chem. Phys. 1983, 79, 926–935.
[35] I. K. Lednev, T.-Q. Ye, R. E. Hester, J. N. Moore, J. Phys. Chem. 1996, 100,
13338–13341.
[36] H. Satzger, C. Root, M. Braun, J. Phys. Chem. A 2004, 108, 6265–6271.
[37] M. Quick, A. L. Dobryakov, M. Gerecke, C. Richter, F. Berndt, I. N. Ioffe,
A. A. Granovsky, R. Mahrwald, N. P. Ernsting, S. A. Kovalenko, J. Phys.
Chem. B 2014, 118, 8756–8771.
[38] G. Cui, P.-J. Guan, W.-H. Fang, J. Phys. Chem. A 2014, 118, 4732–4739.
[39] D. Sampedro Ruiz, A. Cembran, M. Garavelli, M. Olivucci, W. Fuß, Photo-
chem. Photobiol. 2002, 76, 622–633.
[56] D. A. Case, T. A. Darden, T. E. Cheatham, III, C. L. Simmerling, J. Wang,
R. E. Duke, R. Luo, M. Crowley, R. C. Walker, W. Zhang, K. M. Merz, B.
Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvµry, K. F. Wong, F. Paesa-
ni, J. Vanicek, X. Wu, S. R. Brozell, T. Steinbrecher, H. Gohlke, L. Yang, C.
Tan, J. Mongan, V. Hornak, G. Cui, D. H. Mathews, M. G. Seetin, C. Sagui,
V. Babin, P. A. Kollman, AMBER 10, University of California, San Francisco,
2010.
[57] J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman, D. A. Case, J. Comput.
Chem. 2004, 25, 1157–1174.
[40] B. G. Levine, C. Ko, J. Quenneville, T. J. Martinez, Mol. Phys. 2006, 104,
1039–1051.
Received: May 12, 2015
Published online on September 8, 2015
[41] J. M. Ortiz-Sµnchez, R. Gelabert, M. Moreno, J. M. Lluch, J. Chem. Phys.
2008, 129, 21430801–21430811.
Chem. Eur. J. 2015, 21, 15720 – 15731
www.chemeurj.org
15731
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim