2-Hydroxyazobenzenes to tailor pH Pulses and oscillations with light

Article abstract of DOI:10.1002/chem.201000541

This paper evaluates the 2hydroxyazobenzene platform for tailoring proton concentration pulses and oscillations with monochromatic light. The easily prepared 2-hydroxyazobenzenes exhibit large absorptions in the near-UV range. Photoisomerization was investigated by UV/Vis absorption, ¹H NMR spectroscopy, and steady-state fluorescence emission. In the whole investigated series, the trans stereoisomer of the 2-hydroxyazobenzene motif provides the corresponding cis derivative with an action cross section in the 10 ³M^-1cm^-1 range. At the same time, photoisomerization is accompanied by a significant pK drop of the phenol group. According to the phenyl-substituent pattern, cis-to-trans thermal back-isomerization can be tuned in the 10 ms-100 s range. Up to 2 units of reversible pH drops or pH oscillations on the 10 s timescale have been obtained by appropriately tailoring single-wavelength illumination of 2-hydroxyazobenzene solutions.

Full text of DOI:10.1002/chem.201000541

DOI: 10.1002/chem.201000541
2-Hydroxyazobenzenes to Tailor pH Pulses and Oscillations with Light
Matthieu Emond,[a] Thomas Le Saux,[a] Sylvie Maurin,[a] Jean-Bernard Baudin,[a]
Raphael Plasson,[b] and Ludovic Jullien*[a]
Abstract: This paper evaluates the 2-
hydroxyazobenzene platform for tailor-
ing proton concentration pulses and os-
cillations with monochromatic light.
The easily prepared 2-hydroxyazoben-
zenes exhibit large absorptions in the
near-UV range. Photoisomerization
was investigated by UV/Vis absorption,
1H NMR spectroscopy, and steady-state
fluorescence emission. In the whole in-
vestigated series, the trans stereoisomer
of the 2-hydroxyazobenzene motif pro-
vides the corresponding cis derivative
with an action cross section in the
103 mꢀ1 cmꢀ1 range. At the same time,
photoisomerization is accompanied by
a significant pK drop of the phenol
group. According to the phenyl-sub-
stituent pattern, cis-to-trans thermal
back-isomerization can be tuned in the
10 ms–100 s range. Up to 2 units of re-
versible pH drops or pH oscillations on
the 10 s timescale have been obtained
by appropriately tailoring single-wave-
length illumination of 2-hydroxyazo-
benzene solutions.
Keywords:
acidity
·
azo
compounds
·
isomerization
·
photochromism · protonation
Introduction
whereas buffers are traditionally used to investigate thermo-
dynamic aspects of proton significance,[3] specific tools have
to be developed to interrogate kinetic aspects that require
to control the temporal variation of the proton concentra-
tion.
The proton is the essential ion that rules exchanges between
acidic and basic groups.[1] Its reaction with a molecule often
leads to major alterations of structure and reactivity, which
makes it a ubiquitous catalyst.[2] The investigation of the
role of the proton on the behavior of reactive systems is
well established in chemistry as well as in biology. However,
Photochemistry is particularly attractive here, as it can
provide a noninvasive road with spatial and temporal resolu-
tion. Proton photoactivation can rely on a photoacid, which
irreversibly photoreleases a proton once illuminated. This
situation is encountered in caged weak bases such as phos-
phates[4] and sulfates,[5] or in o-nitrobenzaldehydes that yield
a benzoic acid after photorearrangement.[6–8] Alternatively,
the proton may be only transiently photoreleased from a
photoacid, thereby making it possible to shape proton
pulses with light.[9] In particular, this feature opens a road to
extract rate constants and mechanisms of processes that in-
volve protons by analyzing the dependence of the system re-
sponse on the frequency of periodic oscillations of proton
concentration.[10,11] Interestingly, since proton exchanges are
fast[1] and often interfere with various chemical reactions, re-
versible photoacids could also be used to investigate, along
the same way, the kinetics and mechanisms of other process-
es that involve inorganic ions,[12] biomolecules,[13,14] or syn-
thetic systems.[15]
[a] M. Emond, Dr. T. Le Saux, S. Maurin, Prof. Dr. J.-B. Baudin,
Prof. Dr. L. Jullien
Ecole Normale Supꢀrieure, Dꢀpartement de Chimie
UMR CNRS-ENS-UPMC Paris 06 8640 Pasteur
24, rue Lhomond, 75231 Paris Cedex 05 (France)
Fax : (+33)1-4432-3325
E-mail: Ludovic.Jullien@ens.fr
[b] Dr. R. Plasson
Nordita, Roslagstullsbacken 23, 10691 Stockholm (Sweden)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201000541. It contains a simple
model to analyze the pH change resulting from illuminating a reversi-
ble photoacid; the organic syntheses; the derivation of the proton-ex-
change constants of the trans stereoisomers of the 2-hydroxyazoben-
zenes (Figure 1S); the evidence for the trans-to-cis photochemical iso-
merization of TH2Cl3 (Figures 2S and 3S); the kinetic analysis of the
photochemical behavior of 2-hydroxyazobenzenes: theory and experi-
mental results (Figures 4S–8S); the protocol for correcting the emis-
sion from the fluorescent pH reporter from the variation of the
sample absorbance (Figures 9S–12S); and the numerical simulation
protocol for proton concentration pulses (Figure 13S).
In relation to molecular engineering, three parameters are
relevant to photogenerate periodic pH oscillations (see the
Supporting Information): 1) the shift of the proton-exchange
thermodynamic constant upon photoactivation of the rever-
8822
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8822 – 8831

FULL PAPER
sible photoacid AH (pKAH ꢀpKAH); 2) the ratio of the aver-
in relation to the photocontrol of proton-exchange thermo-
dynamic constants. Provided the reported photochemical be-
havior is robust enough,[24] this approach could be attractive.
The alteration of the environment of an acidic group after
photoisomerization could also be used to generate proton
pulses with light. The corresponding alteration generally
provides only modest changes of proton-exchange con-
stants.[25–27] In contrast, from partition coefficient measure-
ments, Haberfield derived promising orders of magnitude
for the photoenhancement of the ionization constants of
two 2-hydroxyazobenzenes.[28] The best observed factor of
600 was explained by environmental changes of the acidic
proton upon illumination: in the trans stereoisomer, acidity
is diminished by the hydrogen bond between the phenol and
an azo nitrogen, whereas in the cis configuration, the hydro-
gen bond is lost and the intrinsic acidity of the substituted
phenol is restored (see Scheme 2; see below).[29] However,
Haberfieldꢂs purpose of investigating photocontrolled phase
transfer made it impossible to get any insight about the life-
time of the photoactivated acidic cis stereoisomer, thus pre-
venting the evaluation of the relevance of the system toward
the photogeneration of pH pulses or oscillations.
Considering the last-mentioned results as well as the ex-
planation that accounts for the photoinduced acidity, we
have envisioned that the 2-hydroxyazobenzene backbone
could provide a versatile platform for building a series of re-
versible photoacids appropriate for tailoring significant pH
pulses and oscillations. Numerous synthons are commercial-
ly available to achieve efficient syntheses of 2-hydroxyazo-
benzenes in one step. Changing starting materials should
provide an easy way to alter solubility of the targeted 2-hy-
droxyazobenzenes as well as both their photophysical and
acidic properties in the ground and photoactivated states.
Structural changes should be similarly expected to provide a
road to finely tune the kinetics for thermal return from the
photoactivated cis stereoisomer, which is known to be faster
than the corresponding azobenzene series.[29–31] Azobenzenes
are additionally well known for the robustness of their pho-
tochromism:[32,33] 2-hydroxyazobenzenes should exhibit no
significant fatigue upon multiple photocycles that lead to
proton pulses. Eventually, their ionization that accompanies
proton photorelease during the trans-to-cis photoisomeriza-
tion could further enlarge the broad field of relevance of
the azobenzene moiety, which has been extensively used to
photocontrol the behavior of small molecules as well as of
natural and synthetic macromolecules.[34]
*
age concentrations in the photoactivated and in the ground
states (AH*/AH); and 3) the relaxation time associated with
the deactivation of the AH* photoactivated state. Together
with the total photoacid concentration, the two first parame-
ters control the maximal amplitude of the pH oscillation
which can be obtained from periodically modulating the illu-
mination of the photoacid, whereas the third parameter con-
trols the effective amplitude which can be reached at a
given light modulation frequency. To get pH oscillations of
significant amplitude in a broad range of light modulation
frequencies, one has to favor a photoacid that exhibits a
large pKAH ꢀpKAH shift and a high rate of relaxation of the
*
photoactivated state, and to increase light intensity to reach
a large AH*/AH ratio.
From the first point, molecules such as phenols that ex-
hibit a pK value of their first electronic excited state that is
up to ten units lower than the pK of their ground state
could seem particularly attractive.[16,17] Indeed, they have
been fruitfully used to photogenerate transient out-of-equi-
librium states to investigate local buffering properties of var-
ious media. However, the short lifetime of singlet excited
states is a major limitation to the generation of significant
overall acidifications: powerful light sources would have to
be used with the risk of encountering detrimental side ef-
fects. In fact, the delivery of significant pH pulses requires
the lifetime of the photoactivated state AH* to be longer-
lived than an electronic excited state but still short-lived
enough to get a good temporal resolution.
Photocontrolled reorganization of the conjugated pathway
between donor- and acceptor-terminating groups that bear
acidic or basic functions is here a first strategy.[18–20] Howev-
er, the illumination of the already investigated molecules
led only to modest changes of pK. Photoisomerization of
substituted aromatic rings that transiently generate acidic
groups has also been examined,[21–23] although not explicitly
Abstract in French: Cet article ꢀtudie la plateforme 2-hy-
droxy-azobenzꢁne afin dꢂobtenir des sauts ou des oscillations
de concentration de proton sous irradiation monochromati-
que. Les 2-hydroxy-azobenzꢁnes sont facilement synthꢀtisꢀs
et prꢀsentent une bande dꢂabsorption intense dans le proche
UV. La photoisomꢀrisation de ces molꢀcules a ꢀtꢀ ꢀtudiꢀe par
1
spectroscopie UV-Visible, RMN H et par fluorimꢀtrie. Dans
toute la sꢀrie, lꢂisomꢁre trans est transformꢀ en isomꢁre cis
avec une section efficace de photoconversion de lꢂordre
103 mꢀ1 cmꢀ1. Cette photoisomꢀrisation sꢂaccompagne dꢂune
diminution significative du pKa du groupe phꢀnol. Suivant la
nature du substituant en para du noyau aniline, lꢂisomꢀrisa-
tion thermique cis/trans peut se produire ꢃ une ꢀchelle de
temps caractꢀristique comprise entre 10 ms et 100 s. Des sauts
de pH de presque deux unitꢀs et des oscillations ꢃ une pꢀrio-
de de lꢂordre de la dizaine de secondes ont ꢀtꢀ observꢀs en il-
luminant convenablement des solutions de 2-hydroxyazoben-
zꢁne ꢃ une seule longueur dꢂonde.
The present paper examines the 2-hydroxyazobenzene
platform in the generation of light-driven pH pulses and os-
cillations. The first section deals with the design and the syn-
theses of two series of 2-hydroxyazobenzene derivatives that
differ by the substitution pattern of the phenyl rings borne
by the nitrogen–nitrogen double bond. The second section
investigates the acid–base, photophysical, and photochemi-
cal properties of those 2-hydroxyazobenzenes. The third sec-
tion implements two 2-hydroxyazobenzenes in the photo-
generation of pulses and oscillations of proton concentra-
tion.
Chem. Eur. J. 2010, 16, 8822 – 8831
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8823

L. Jullien et al.
Results and Discussion
crease of conjugation that accompanies the deprotonation of
the phenol group, the absorption spectrum is redshifted in
Design and synthesis: In the present study, we were willing
to evaluate the robustness of the change of the proton-ex-
change thermodynamic constant as a function of the molec-
ular structure by altering the substituent pattern of the two
phenyl rings of the 2-hydroxyazobenzene motif. Thus we
generated two series of derivatives (Scheme 1): 1) starting
basic media (lmax
(THCl3)=435 nm at pH 11.1). From the whole collection of
absorption spectra, we extracted pKTH Cl =9.4. Despite the
presence of the additional -N=N- electroattractive group in
the ortho position of the phenol moiety, this value is signifi-
cantly larger than the pK=8.3 that we measured for the
parent 2,4,5-trichlorophenol in
ACHTUNGNRETN(NUG TH2Cl3)=347 nm at pH 7.2, whereas lmax-
AHCTUNGTRENNUNG
2
3
the same solvent. Such a large
pK shift probably originates
from a hydrogen bond between
the proton of the phenol group
and one nitrogen atom of the
azo motif (see Scheme 2). This
explanation is also supported
by the observation that the
chemical shift in the 1H NMR
spectrum of the proton borne
Scheme 1. Synthesis and structures of the 2-hydroxyazobenzene derivatives investigated in the present ac-
count.
by the hydroxy group in TH2Cl3
(d=14.6 ppm) is much larger
from the most promising compound investigated by Haber-
field,[28] we retained 2,4,5-trichlorophenol for the phenol
moiety in the first series: THRCl3 (R=NO2, SO3K, Br,
COOH, H, SMe, Et, Me, OMe). In contrast, we changed the
substituent borne at the para position of the other phenyl
ring with the expectation of affecting both the strength of
the hydrogen bond in the trans stereoisomer and the rate
constant for the thermal cis-to-trans isomerization. 2) In the
second series, we explored three other phenols (2,4-difluoro-,
2,4-dichloro-, and 4-nitrophenol) that bear electron-with-
drawing groups,[28] and which have in common the absence
of a substituent in the 6-position of the resulting 2-hydroxya-
zobenzenes.
The investigated 2-hydroxyazobenzene derivatives have
Scheme 2. Mechanism accounting for the photochemical behavior of 2-
hydroxyazobenzenes.
been obtained in one step with good yields (50–80%) from
coupling the appropriate phenol in a basic medium with the
diazonium that results from treating the parent aniline with
an excess amount of acidic solution of sodium nitrite
(Scheme 1).[35–38]
than the one observed in the parent 2,4,5-trichlorophenol
(d=5.5 ppm) in deuterated chloroform.
Physicochemical properties
The latter interpretation was strengthened after reproduc-
ing the same experiment in the whole THRCl3 series (see
TH proton-exchange thermodynamic constants: We first ex-
amined the acid–base properties of the trans stereoisomers
TH of the present 2-hydroxyazobenzenes. To avoid any
problem of solubility in the whole series, water/acetonitrile
1:1 (v/v) was used as a solvent. The absorption spectrum of
the TH derivatives was recorded as a function of the proton
concentration in Britton–Robinson buffer[39] (acetic acid,
boric acid, phosphoric acid; 0.04m)/acetonitrile 1:1 (v/v). All
the solutions were protected from sunlight to avoid any
trans-to-cis photoisomerization (see below).
Table 1). The lmaxACTHUNGTRENN(UG THRCl3), lmaxCATHUGNTRNE(NUGN TRCl3), and pKTHRCl
3
values range between 347 and 416 nm, 426 and 494 nm, and
8.3 and 9.7, respectively. In particular, when the R substitu-
ent exhibits more and more electron-withdrawing properties,
pKTHRCl decreases. This trend was further quantified by
3
plotting pKTHRCl as a function of the Hammett coefficient
3
sR of the R substituent.[40–42] As shown in Figure 1a, the de-
pendence of pKTHRCl on sR is fairly linear. The negative
3
slope observed for pKTHRCl is in line with a weakening of
3
the electron density at the nitrogen site involved in the hy-
drogen bond in THRCl3 when R becomes more and more
electron withdrawing.[43]
Figure 1S in the Supporting Information displays repre-
sentative results obtained with TH2Cl3. In line with the in-
8824
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8822 – 8831

2-Hydroxyazobenzenes
FULL PAPER
Table 1. Maximal wavelengths of absorption and molar absorption coefficients for the acidic (lTmHax, e(lmTHax))
and basic (lTmax, e(lmT ax)) states, proton-exchange constants for the trans (pKTH) and the cis (pKCH) stereoiso-
mers, rate constants kDCH!TH and kCD!T associated with thermal cis-to-trans isomerization of the acidic CH and
basic C states for the investigated 2-hydroxyazobenzenes. Solvent: Britton–Robinson buffer[39] (acetic acid,
boric acid, phosphoric acid; 0.04m)/acetonitrile 1:1 (v/v), T=298 K.
drogen bond between the hy-
droxyl group of the phenol and
a nitrogen atom. However, in
contrast with the previous ob-
servations, the pK values of
THOMeCl2 (9.2), TH2Cl2 (8.7),
and TH2F2 (8.8) were signifi-
cantly smaller than the ones of
the parent 2,4-dichlorophenol
(9.4) and 2,4-difluorophenol
(10.1). This behavior could be
explained by a better phenol
conjugation with the electron-
withdrawing diazo double bond
in the second series of 2-hy-
droxyazobenzenes, which lacks
the chloro substituent at the 5-
phenol position. This explana-
tion would be supported by the
trend observed in UV/Vis ab-
sorption (Table 1): for a same
TH
max
TH
max
TH
l
[nm] (e(l
)
lmT ax [nm] (e(lmT ax
[mmꢀ1 cmꢀ1])
)
pKTH ꢁ0.1 pKCH ꢁ0.1 kDCH!TH ꢁ0.1 102 ꢃkDC!T
[mmꢀ1 cmꢀ1])
[minꢀ1
]
[minꢀ1
]
THOMeCl3
THMeCl3
THEtCl3
THSMeCl3
TH2Cl3
380 (20)
355 (22)
357 (19)
416 (8)
426 (4)
430 (5)
431 (4)
441 (5)
435 (5)
449 (5)
443 (4)
451 (6)
494 (10)
475 (13)
477 (10)
470 (5)
9.5
9.7
9.6
9.6
9.4
9.2
8.9
8.9
8.3
9.2
8.7
8.8
7.0
6.6
6.8
6.7
6.6
0.4
1.3
1.5
2.6
4.5
(3ꢁ3)[a]
(2ꢁ2)[a]
(2ꢁ2)[a]
(2ꢁ2)[a]
<2
347 (20)
[b]
[b]
[b]
THCOOHCl3 353 (16)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
[b]
[b]
[b]
THBrCl3
THSO3KCl3
THNO2Cl3
THOMeCl2
TH2Cl2
355 (12)
349 (22)
354 (22)
361 (22)
328 (20)
324 (17)
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
TH2F2
–
–
–
[a] From the fit with Equation (44) in the Supporting Information. [b] Thermal cis-to-trans isomerization took
place too rapidly to be measured with our protocol.
phenyl substituent, the removal of the 5-chloro substituent
significantly redshifts the maximal wavelength of absorption
of the phenolate T, in line with a larger conjugation pathway
in the second series of 2-hydroxyazobenzenes.
Preliminary photochemical observation: In a series of pre-
liminary experiments performed at 277 K (to avoid the ther-
mal cis-to-trans isomerization; see below), we illuminated
42 mm solutions of TH2Cl3 in Britton–Robinson buffer[39]
(acetic acid, boric acid, phosphoric acid; 0.04m)/acetonitrile
1:1 (v/v) at 365 nm for 5 min. We subsequently recorded and
compared their UV/Vis absorption spectra either immedi-
ately or ten minutes after turning off the UV light.
At pH 8.0 as well as pH 11.8 (at which points TH2Cl3
exists as its acidic and basic states, respectively), UV illumi-
nation caused a significant hypochromic effect (by about
50%) without major alteration of the absorption bands (Fig-
ures 2Sa and b in the Supporting Information). These obser-
vations were in line with a trans-to-cis photoisomerization
both for the acidic and basic states of TH2Cl3. Indeed, they
agreed with reported literature on 2-hydroxyazobenzenes in
which the marked difference of light absorption between
the trans and cis stereoisomers was interpreted as resulting
from the nonplanar configuration of the cis isomer.[29,44] The
preceding conclusions were also supported by 1H NMR
spectroscopic observations that showed that UV illumina-
tion of the trans stereoisomer generated the formation of a
new species with NMR spectroscopic peaks in line with the
corresponding cis derivative (Figure 3S in the Supporting In-
formation). We additionally noticed that the absorption
spectrum of the illuminated solution at the lowest pH fully
recovered when the solution was left unilluminated for ten
minutes at room temperature. In contrast, the illuminated
sample at the largest pH did not exhibit any significant evo-
lution after the same delay. We interpreted this behavior as
resulting from a thermal cis-to-trans isomerization; it oc-
Figure 1. a) Plot of pKTHRCl (squares) and pKCHRCl (triangles), and
3
3
D
b) log10 k
as a function of the Hammett coefficient sR of the
CHRCl3!THRCl3
R substituent. Markers: experimental data; solid line: linear fit. We
found: pKTHRCl =9.4–1.3sR, log10 k
[minꢀ1]=0.66+3.76sR.
D
CHRCl3!THRCl3
3
The experiments performed in the second series of 2-hy-
droxyazobenzenes brought additional information to the
preceding picture (Table 1). As in the first series, the obser-
vation pKTHOMeCl >pKTH Cl supports the presence of a hy-
2
2
2
Chem. Eur. J. 2010, 16, 8822 – 8831
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8825

L. Jullien et al.
curred rather fast for the acidic TH2Cl3 state but much
slower for the basic THCl3 form.[29]
to-trans isomerization of the acidic CH and basic C states,
respectively (see Figure 4S in the Supporting Information).
As shown in Table 1, whereas our experimental conditions
were adapted to measure kDCH!TH with good precision, kDC!T
was too low to be reliably evaluated. The value of kCDH!TH
strongly depends on the phenyl substituent R. The slowest
thermal cis-to-trans relaxations are observed for the sub-
stituents that exhibit the strongest electron-donating proper-
ties. The corresponding free-energy linear relationship with
the Hammett parameters is displayed in Figure 1b. By ex-
tending the observed linear dependence to the whole series,
this study suggests that changing the THRCl3 phenyl sub-
stituent is an efficient way to tune the relaxation time asso-
ciated with the thermal cis-to-trans isomerization over five
orders of magnitude: the 100 s range is observed with the
slowest CHOMeCl3, whereas 10 ms can be extrapolated for
the most reactive CHNO2Cl3.
We were unable to reliably collect the rate constants asso-
ciated with the thermal cis-to-trans isomerization in the
second series of 2-hydroxyazobenzenes: even at the highest
pH, thermal return of the photogenerated CH is already
fast. For instance, it occurs in 7 min in the basic state of
CHCl2, whereas we did not observe any relaxation beyond
the hour timescale for CHCl3. In particular, the latter obser-
vation may suggest conjugation between the phenol and the
diazo bond to be significant along the way from the cis to
the trans stereoisomer since the second series lacks a 5-
phenol substituent. The phenyl-substituent effect observed
in the first series was again present: at pH 10.4, the respec-
tive relaxation times of CHOMeCl2 and CH2Cl2 were 40 and
4 s. In addition, 2,4-difluorophenol yielded a faster thermal
return than 2,4-dichlorophenol: at pH 11.7, we found 420
and 10 s for the relaxation times of CH2Cl2 and CH2F2, re-
spectively.
After the series of preliminary experiments, we adopted
the mechanism shown in Scheme 2 to account for the photo-
chemical behavior of 2-hydroxyazobenzenes: we consider
the interconversion between four states: TH, T, CH, and C,
which correspond to the acidic and basic conjugates of the
trans and cis stereoisomers, respectively. The trans-to-cis
conversion is considered to be driven by light only in view
of the larger thermodynamic stability of the trans stereoiso-
mer, whereas cis-to-trans isomerization can be thermally as
well as photochemically driven. In contrast, proton exchang-
es occur under thermal control only.
CH proton-exchange thermodynamic constants: cis-Azoben-
zene stereoisomers often live long enough to be purified by
chromatography.[32,45] In the present 2-hydroxyazobenzene
series, the fast cis-to-trans thermal isomerization forbids
both the purification and the application of the titration pro-
tocol at equilibrium to measure the CH proton-exchange
thermodynamic constants. In contrast, the analysis at differ-
ent proton concentrations of the kinetics of thermally driven
cis-to-trans isomerization gives access to the sought-after
thermodynamic constants (see Figure 4S in the Supporting
Information).
We illuminated TH solutions at their wavelength of maxi-
mal absorption at various proton concentrations to photo-
generate the cis stereoisomer CH. We subsequently ana-
lyzed in the dark the temporal relaxation of the solution ab-
sorbance to derive pKCH. Table 1 displays the results for the
2-hydroxyazobenzenes for which the lifetime of the cis ste-
reoisomer was large enough to apply the present protocol.
The measured pKCH range between 6.6 and 7.0, much below
the 8.3 one obtained for the 2,4,5-trichlorophenol in Brit-
ton–Robinson buffer[39] (acetic acid, boric acid, phosphoric
acid; 0.04m)/acetonitrile 1:1 (v/v) (but close to its value in
water, pK=6.9[46]). Since cis-2-hydroxyazobenzenes cannot
form the hydrogen bond present in the trans stereoisomers,
the latter observation is significant for the evaluation of the
relative contributions of the factors that govern the acidity
of the 2-hydroxyazobenzenes in the TH2Cl3 series: The 1.0–
1.5-unit pK decrease of CH with regard to the parent
phenol probably evaluates the electron-withdrawing effect
of the diazo double bond, whereas the 2.5-unit
(pKTHꢀpKCH) difference reflects the strength of the hydro-
gen bond.
The present analysis of the kinetics of thermal cis-to-trans
isomerization shows that relaxation times are noticeably
lower than the ones of the corresponding azobenzenes, thus
making the 2-hydroxyazobenzene backbone an attractive
photochromic platform for “fast” applications. In particular,
photoswitching proton-exchange constants will not require
two-color illumination to be kinetically efficient.
Kinetics of the photochemically driven trans-to-cis and cis-to-
trans isomerization: The rate constants associated with azo-
benzene photochemical reactions have been often measured
by recording the sample absorbance after illumination for
various times.[32,45] In contrast, the present 2-hydroxyazoben-
zene series exhibits a thermal cis-to-trans isomerization that
is too rapid to apply the same protocol. We recently imple-
mented a protocol relying on continuous illumination to
face a similar situation that involved fast photochemical
events.[47–49] Here we also adopted continuous illumination
and a fluorescent probe both to generate the photochemical
reactions and to report on their extent by the inner filter
effect. After a series of preliminary experiments (see Figur-
es 5S and 6S in the Supporting Information), we adopted
the strongly fluorescent rhodamine B probe, which is soluble
Kinetics of thermal cis-to-trans isomerization: With our in-
strumental setup, we were able to measure the relaxation
time associated with the thermal CH cis-to-trans isomeriza-
tion when it was higher than a few seconds.
In the TH2Cl3 series, the associated rate constants were
shown to strongly depend on the proton concentration with
half-times ranging between minutes (under acidic condi-
tions) and beyond several hours (at the lowest proton con-
centrations). The analysis of this dependence yielded the
rate constants kCDH!TH and kCD!T associated with thermal cis-
8826
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8822 – 8831

2-Hydroxyazobenzenes
FULL PAPER
in acetonitrile/water 1:1 (v/v); it weakly absorbs in and
strongly emits beyond the absorption wavelength range of
TH/CH (to report on TH/CH absorbance and to avoid any
absorption of the fluorescence emission by TH/CH; see Fig-
ures 2Sa and b in the Supporting Information). Moreover,
rhodamine B is not sensitive to pH in the investigated range
(to identically report on sample absorbance whatever the
pH change during the evolution). Thus we submitted mix-
tures that initially contained the trans stereoisomer TH (typ-
ically at 5 mm concentration) and 1 mm rhodamine B to con-
tinuous illuminations while recording the intensity of the
rhodamine B fluorescence emission. As expected from the
formation of the less-absorbing cis stereoisomer, which
leads to an increase in the light intensity experienced by the
rhodamine B, the fluorescence intensity increases as a func-
tion of time (see Figure 7Sa in the Supporting Information).
After a typical delay of a few seconds to a few tens of sec-
onds, it subsequently reaches a plateau associated with the
photosteady state. We satisfactorily fit the corresponding
evolution to extract functions of the rate constants associat-
ed with the trans-to-cis and cis-to-trans isomerization for a
particular light intensity at a given pH. Reproducing this ex-
periment for various light intensities and at various proton
concentrations eventually provides access to the photochem-
ical features of the acidic and basic states at a given excita-
tion wavelength, eTHFTH!CH and eCHFCH!TH, and eTFT!C
and eCFC!T, in which e and F refer to molar absorption co-
efficients and quantum yields associated with the photoiso-
merization (see the Supporting Information). As eTH and eT
can be directly measured from the absorption spectra of the
thermodynamically stable trans stereoisomer, and consider-
ing that eCH and eC can be roughly estimated (see ref. [44]
Pulses of proton concentration: Our glass electrode was not
sensitive to UV light in the range of investigated powers:
We observed no response to illumination in water/acetoni-
trile 1:1 (v/v). In contrast, we observed up to 2 pH-unit
drops in solutions that contained either TH2Cl3 or
THSO3KCl3 at 200–300 mm concentration (Figure 2a) when
and Figure 2S in the Supporting Information), FTH!CH
FCH!TH, FC!T, and FT!C can be additionally extracted. For
TH2Cl3 at 347 nm, we found eTH(347)FTH!CH(347)=
630mꢀ1 cmꢀ1, eCH (347)=140mꢀ1 cmꢀ1, eC
(347) FCH!TH (347)
FC!T (347)FT!C
(347)=260mꢀ1 cmꢀ1, eT
from which we subsequently derived FTH!CH
FCH!TH (347)ꢂ18%, and FT!C
(347)ꢂ2%, FC!T
by eTH eCH
(347)=2ꢃ104 mꢀ1 cmꢀ1,
(347)=4.8ꢃ
,
G
ACHTUNGTRENNUNG
Figure 2. Proton concentration pulses upon transient illumination a) as
measured with the pH meter in 6 mL of 220 mm TH2Cl3 (or 330 mm
THSO3KCl3; inset) submitted to a 30 s-long (respectively, 50) 365 nm
E
E
ACHTUNGTRENNUNG
G
G
E
;
AHCTUNGTRENNUNG
pulse delivering 1.2ꢃ10ꢀ5 Einsteinsꢀ1 (with
a
flux of 1.7ꢃ
10ꢀ6 Einsteinsꢀ1 cmꢀ2) at the sample; b) as measured with the fluorescein
reporter in 105 mm TH2Cl3 (2 mL), 10 nm rhodamine B, and 10 nm fluores-
cein submitted to a 150 s-long pulse delivering 1.9ꢃ10ꢀ7 Einsteinsꢀ1 (with
a flux of 1.9ꢃ10ꢀ7 Einsteinsꢀ1 cmꢀ2) at the sample. Solid line: Fit result-
ing from numerical simulations (see the Supporting Information). Sol-
vent: degassed water/acetonitrile 1:1 (v/v). T=298 K.
G
R
ACHTUNGTRENNUNG
using
R
ACHTUNGTRENNUNG
103 mꢀ1 cmꢀ1, eC
103 mꢀ1 cmꢀ1.
N
ACHTUNGTRENNUNG
Pulses and oscillations of proton concentration: Equipped
with the preceding derivation of thermodynamic and kinetic
constants, we were able to devise a series of experiments
aimed at observing both pulses and oscillations of proton
concentration. We submitted degassed solutions of 2-hy-
droxyazobenzene in water/acetonitrile 1:1 (v/v) to various
sequences of illumination. The photogenerated variations of
proton concentration were detected by two different means.
We first relied on a glass electrode by observing light-in-
duced drops of the pH-meter indication. We also used a flu-
orescent reporter of the proton concentration to facilitate
the measurements by relying on optical detection only.
they were submitted to a strong UV illumination. As expect-
ed from the photogeneration of the more acidic cis stereo-
isomer in the presence of UV light, the proton concentra-
tion increases and then stabilizes when the steady state is
reached. In contrast, turning off the light results in the relax-
ation of the proton concentration to the initial unilluminat-
ed state. The relaxation timescale is about ten times faster
for THSO3KCl3 than for TH2Cl3, as expected from a consid-
eration of the increasing value of the thermal cis-to-trans
isomerization rate constant when the substituent group R is
more and more electron-withdrawing.
Chem. Eur. J. 2010, 16, 8822 – 8831
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8827

L. Jullien et al.
To circumvent any limitation associated with the one- to
ten-second response time of the glass electrode, we alterna-
tively relied on a fluorescent reporter, which can be used
down to the millisecond timescale to analyze the temporal
variations of proton concentration. Considering the investi-
gated pH range, we chose fluorescein, which exhibits
pKFlu =7.6 in the considered solvent. Since fluorescein emis-
sion depends not only on proton concentration but also on
the intensity of the exciting source, we also added the pH-
insensitive rhodamine B to report on the variation of the
sample absorbance associated with the trans-to-cis isomeri-
zation of the 2-hydroxyazobenzenes (see above). Therefore
the ratio of the fluorescence emission from fluorescein and
from rhodamine B gives access to the desired proton con-
centration after calibration (see Figures 9S and 10S in the
Supporting Information). Figure 2b displays the typical pH
drop observed upon submitting around 100 mm TH2Cl3 to a
365 nm pulse generated with the fluorometer light source.
As anticipated from the lower illumination power and the
lower TH2Cl3 concentration, the pH drop in Figure 2b is
smaller than the one in Figure 2a.
been deduced from analysis of the fluorescence variations of
fluorescein and rhodamine B as described above and in the
Supporting Information (see Figure 11S).
We first explored conditions of weak illumination modula-
tion to remain in the regime of first-order perturbation theo-
retically calculated in the Supporting Information (Fig-
ure 3a). As predicted, the proton concentration exhibits a si-
nusoidal modulation at the excitation frequency: for a 40%
210 mHz modulation of the 365 nm UV light around the
7.2ꢃ10ꢀ8 Einsteinsꢀ1 average value, we observed a 11%
modulation of the proton concentration delayed by f=
0.4 rad with regard to light modulation. Using Equation (23)
in the Supporting Information, we retrieved from the pre-
ceding observations the sum of the rate constants associated
with the trans-to-cis and cis-to-trans THSO3KCl3 isomeriza-
tions: khTvH!C +kChv!TH +kCD!TH =0.5 sꢀ1. This estimate is in line
with the corresponding value that can be directly extracted
The preceding pH drops have been evaluated for their
consistency with the thermodynamic, kinetic, and photo-
chemical properties evaluated above. Taking into account all
the chemical reactions of the system, the temporal evolution
of the chemical concentrations was simulated using the
known set of parameters and the experimental curves were
fitted by adjusting the remaining parameters (see Support-
ing Information). Starting from the simulation of the behav-
ior observed by using fluorescein to report on pH, we ob-
tained a satisfactory fit of the experimental data (see Fig-
ure 2b) by considering that the illuminated solution con-
tained a minor (70 mm) weak base contaminant. It was sug-
gested that it was made of residual carbonates from
analyzing ionization constants as well as the rate constant
associated with an acidic state decomposition that we attrib-
uted to H2CO3 (see the Supporting Information). In particu-
lar, the resulting set of fitting parameters was also appropri-
ate for fitting the amplitude of the photogenerated pH
change displayed in Figure 2a. Thus the numerical simula-
tions showed that the observed pH drops were consistent
with all the kinetic and thermodynamic parameters experi-
mentally determined in the present work or available in the
literature. However, they also suggested that, as anticipated,
the presence of buffering species (here residual carbonates)
may dampen the photogenerated pH drop from illuminating
2-hydroxybenzene derivatives.
Oscillations of proton concentration: After generating pulses
of proton concentration, we were interested in evaluating
whether modulations of the UV light source can lead to the
modulation of pH. For this series of experiments, we re-
tained the highly soluble THSO3KCl3 2-hydroxyazobenzene.
Indeed, this derivative was expected to exhibit one of the
fastest thermal returns of its cis stereoisomer, thereby allow-
ing us to perform significant pH modulation at the highest
frequencies. The pH values and proton concentrations have
Figure 3. Oscillations of proton concentration (solid line) as measured
with the fluorescein reporter upon applying a periodic sinusoidal modula-
tion of the UV illumination (dotted line), IE =(IE)(1+bsinwt), to a solu-
tion that contained 300 mm THSO3KCl3, 10 nm rhodamine B, and 10 nm
fluorescein. a) IE =7.2ꢃ10ꢀ8 Einsteinsꢀ1
,
b=0.4, and w=210 mHz;
b) IE =1.6ꢃ10ꢀ7 Einsteinsꢀ1, b=0.9, and w=210 mHz. Solvent: degassed
water/acetonitrile 1:1 (v/v). T=298 K.
8828
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8822 – 8831

2-Hydroxyazobenzenes
FULL PAPER
either by bubbling N2 for 10 min (when the pH meter was used to report
on pH pulses) or by freeze–thaw cycling three times under vacuum
(when fluorescence was used to report on pH pulses and oscillations).
from the timescale for THSO3KCl3 photoisomerization:
With Equation (51) in the Supporting Information and the
hv
TH!C
data shown in the inset of Figure 2a, we derived: k
+
pH measurements: By assimilating activity and concentration, the proton
concentration in acetonitrile/water 1:1 (v/v) solutions was directly mea-
sured after calibration of the pH meter (Standard pH meter PHM210,
Radiometer Analytical equipped with Radiometer Analytical electrodes)
with various buffers (acetic acid, phosphoric acid, and boric acid; 50 mm
concentration in both the acidic and the basic states), for which the ioni-
zation constants in acetonitrile/water 1:1 (v/v) solutions have been report-
ed.[51–53]
khCv!TH +kDC!TH =0.3 sꢀ1.
Figure 3b shows that much larger pH modulations can be
obtained upon applying larger sinusoidal modulations of the
UV light intensity: In a degassed 300 mm THSO3KCl3 solu-
tion, pH periodically varied between 7.0 and 7.4 when the
light intensity was sinusoidally modulated at 210 mHz by
90% around the average value of 1.6ꢃ10ꢀ7 Einsteinsꢀ1 (see
also Figure 11S in the Supporting Information). Moreover,
other periodical patterns can be applied to the sample. As
an example, Figure 12S displays the pH evolution that can
be obtained from applying a triangular periodical illumina-
tion.
UV/Vis absorption and steady-state emission spectroscopies: UV/Vis ab-
sorption spectra were recorded using a Uvikon-940 spectrophotometer
(Kontron, Zꢄrich, Switzerland). Corrected fluorescence spectra were ac-
quired using a LPS 220 spectrofluorometer (PTI, Monmouth Junction,
NJ). The quartz cuvettes (Hellma) were 1ꢃ1 cm. The temperature of the
holders was maintained by using a thermostat of circulating baths (Poly-
stat 34-R2, Fisher Bioblock Scientific, Illkirch, France), and the tempera-
ture was directly measured in the cuvettes using a type K thermocouple
connected to a ST-610B digital pyrometer (Stafford Instruments, Stafford,
UK).
Conclusion
Stopped-flow experiments were performed using a RX2000 rapid kinetic
stopped-flow accessory (Applied Photophysics, Leatherhead, UK) adapt-
ed to the LPS 220 spectrofluorometer. In this setup, two 100 mL solutions
were mixed with typical dead times of 100 ms and the fluorescence inten-
sity was recorded over time at 3 Hz.
This study demonstrates that the 2-hydroxyazobenzene
series provides a versatile platform for the design of reversi-
ble photoacids to generate significant pH pulses and oscilla-
tions with monochromatic light. The syntheses are easy and
numerous synthons will be available to tune solubility as
well as to graft the 2-hydroxyazobenzene moiety onto vari-
ous molecular backbones. The photoisomerization behavior
associated with phenol ionization is robust and capable of
photoincreasing proton concentration by typically two
orders of magnitude in neutral media. The Gibbs free-
energy relationships we derived for the protonation con-
stants of the trans and cis stereoisomers as well as for the
rate constant for thermal cis-to-trans isomerization of 2-hy-
droxyazobenzenes make it possible to predict which substi-
tution patterns at the phenyl and phenol rings should be
chosen to observe a given amplitude of pH change under
specific illumination conditions. In particular, we showed
that analytic calculations and numerical simulations satisfac-
torily accounted for our results.
In addition to the preceding remarks that focus on the
photogeneration of protons, this study can also be seen as
providing a generic photochromic unit, the charge of which
will change upon illumination in a wide pH range around
neutrality. Considering the wide use of the azobenzene
moiety as a photoswitch, this alternative perspective should
attract attention given the major structural change that
occurs upon illumination of the 2-hydroxyazobenzene back-
bone.
Irradiation experiments: Different protocols have been used to perform
the irradiations.
Protocol 1: For the preliminary experiment to evidence the trans-to-cis
photoisomerization of TH2Cl3 (Figure 2S in the Supporting Information),
we put a bench-top UV lamp (6 W, 365 nm UV lamp; Fisher Bioblock)
above a 1ꢃ1 cm quartz cuvette that contained the solution (2 mL) that
was subsequently transferred into the UV/Vis spectrometer.
Protocol 2: The experiments to acquire the photochemical properties of
the investigated 2-hydroxyazobenzenes were performed using the
stopped-flow apparatus (RX.2000 Stopped-Flow Mixing Accessory, Ap-
plied Photophysics) upon illuminating the contents of the 1ꢃ0.2 cm
quartz cuvette (ꢂ200 mL) with the 75 W xenon lamp of the Photon Tech-
nology International LPS 220 spectrofluorometer at several slit widths to
cover a significant range of incident light intensities.
Protocol 3: The experiments devoted to the generation of pH pulses and
oscillations were performed by following two protocols. 1) When the pH
variation was reported by a pH meter, the experiment was done on sam-
ples (6 mL) in a small beaker submitted to a pulse of light from a Hama-
matsu LC8L9588 UV lamp. 2) When the pH variation was followed by
fluorescence spectroscopy, the experiment was done on samples (2 mL)
in 1ꢃ1 cm quartz cuvettes continuously subjected to a weak illumination
at 500 nm and additionally subjected to either pulse or periodic oscilla-
tions of the light from a UV-emitting diode (Nichia chip UV LED
NCSU033A(T), 0–700 mA), either by turning the on/off cycle or by using
an Agilent 33220A, 20 MHz function, arbitrary waveform generator.
In all cases, the incident-light intensities were calibrated by determining
the kinetics of photoconversion of the a-(4-dimethylaminophenyl)-N-
phenylnitrone into 3-(4-dimethylaminophenyl)-2-phenyloxaziridine in ab-
solute ethanol as described in ref. [54]. During the present series of ex-
periments, the typical photon flux at the sample was in the 10ꢀ5
10ꢀ10 Einsteinsꢀ1 range.
–
Data processing: The evolutions of the TH absorbance as a function of
pH were analyzed using the SPECFIT/32 Global Analysis System (Ver-
sion 3.0 for 32-bit Windows systems) to extract their pK.[55] The other
data have been processed and fitted with Igor Pro 6 (WaveMetrics, Lake
Oswego, OR).
Experimental Section
Reagents and solutions: Fluorescein, rhodamine B, and acetonitrile were
purchased from Sigma–Aldrich, TCI Europe, and Fischer Scientific. Hy-
drochloric acid (5m) was prepared from a commercial 37% solution.
Sodium hydroxide solution (9.2m) was commercially available. All solu-
tions were prepared with water purified using a Direct-Q 5 instrument
(Millipore, Billerica, MA). The Britton–Robinson buffers were prepared
according to Frugoni.[39] The water/acetonitrile solutions were degassed
Computer simulation: Simultaneously computing very slow and very fast
chemical reactions as in the present system requires an algorithm capable
of dealing with such a “stiff” set of ordinary differential equations
(ODE). Classical all-purpose algorithms (for instance, the Runge–Kutta
methods) would lead the simulation to scale on the fastest reactions, and
Chem. Eur. J. 2010, 16, 8822 – 8831
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8829

L. Jullien et al.
thus advance the computation by very small time steps, even when it is
not necessary.[56] The time evolution was thus computed using an implicit
method, the Bulirsch–Stoer algorithm.[57] For each run with a given set of
normalized parameters p, it was possible to evaluate the distance d(p) be-
[22] R. Haag, J. Wirz, P. J. Wagner, Helv. Chim. Acta 1977, 60, 2595–
2607.
[23] Y. Eichen, J.-M. Lehn, M. Scherl, D. Haarer, R. Casalegno, A.
Corval, K. Kuldovac, H. P. Trommsdorff, J. Chem. Soc. Chem.
Commun. 1995, 713–714.
tween the simulated evolution pH
simACHTUNGRTEN(NUNG p,t) with the experimental data
pHexpA(p,t) by using Equation (1), for all the experimental values of t,
taken at regular time intervals:
CHTUNGTRENNUNG
[24] For a report that illustrates the lack of robustness of a closely relat-
ed photochemistry by shifting the maximum wavelength absorption
to the red, see: I. Aujard, C. Benbrahim, M. Gouget, O. Ruel, J.-B.
Baudin, P. Neveu, L. Jullien, Chem. Eur. J. 2006, 12, 6865–6879.
[25] L. Jullien, A. Lemarchand, S. Charier, O. Ruel, J.-B. Baudin, J. Phys.
Chem. B 2003, 107, 9905–9917.
tend
X
2
dðpÞ ¼
½pHsimðp; tÞ ꢀ pHexpðp; tÞꢃ
ð1Þ
t¼0
[26] T. Matsuhira, K. Tsuchihashi, H. Yamamoto, T. Okamura, N.
Ueyama, Org. Biomol. Chem. 2008, 6, 3118–3126.
To automatically adjust some parameters to fit the experimental curve,
the function d(p) was minimized using a simplex algorithm.[58]
[27] R. S. Stoll, M. V. Peters, A. Kuhn, S. Heiles, R. Goddard, M. Bꢄhl,
C. M. Thiele, S. Hecht, J. Am. Chem. Soc. 2009, 131, 357–367.
[28] P. Haberfield, J. Am. Chem. Soc. 1987, 109, 6177–6178.
[29] G. Wettermark, M. E. Langmuir, D. G. Anderson, J. Am. Chem.
Soc. 1965, 87, 476–481.
[30] E. Fischer, Y. F. Frei, J. Chem. Soc. 1959, 3159–3163.
[31] G. Gabor, E. Fischer, J. Phys. Chem. 1962, 66, 2478–2481.
[32] J. Griffiths, Chem. Soc. Rev. 1972, 1, 481–493.
The simulation program was implemented in C (GCC, GNU Scientific
Library) using the Burlisch–Stoer algorithm for the ODE numerical inte-
gration and the simplex algorithm for the parameter optimization avail-
able in the Gnu Scientific Library (GSL).[59] Fully documented source
code is available as Supporting Information (see file phjump.c). The sim-
ulations were computed using a Lenovo R61 laptop, with an Intel Core2-
Duo Mobile Processor T7250. One typical simulation run takes a couple
of seconds of computation, and one parameter optimization process
takes a couple of minutes.
[33] S. Kobatake, M. Irie, Annu. Rep. Prog. Chem. Sect. C 2003, 99, 277–
313.
[34] G. Mayer, A. Heckel, Angew. Chem. 2006, 118, 5020–5042; Angew.
Chem. Int. Ed. 2006, 45, 4900–4921.
[35] H. T. Clarke, W. R. Kirner, Org. Synth. Coll. Vol. 1941, 1, 374–377.
[36] A. Vogel, Textbook of Practical Organic Chemistry, 4th ed., Long-
man, Mꢄnchen, 1978, pp. 714–715.
[37] S. Darwish, H. M. Fahmy, M. A. Abdel Aziz, A. A. Et Maghraby, J.
Chem. Soc. Perkin Trans. 2 1981, 344–349.
Acknowledgements
We are thankful to Axel Brandenburg for useful discussions. We also
thank for their support Nordita (concerning the computer simulations),
the Pierre-Gilles de Gennes Foundation (for a grant) and the European
program COST “System Chemistry” CM0703 (for its support).
[38] T. Kunitake, Y. Okahata, M. Shimomura, S. Yasunami, K. Takarabe,
J. Am. Chem. Soc. 1981, 103, 5401–5413.
[39] C. Frugoni, Gazz. Chim. Ital. 1957, 87, 403–407.
[40] H. H. Jaffꢀ, Chem. Rev. 1953, 53, 191–261.
[41] C. A. Reichardt, Solvents and Solvent Effects in Organic Chemistry,
3rd ed., VCH, Weinheim, 2003.
[42] M. B. Smith, J. March, Advanced Organic Chemistry, 5th ed., Wiley
Interscience, New York, 2001.
[43] The negative slope is also in agreement with an increase of the sta-
bilization of the conjugate base TRCl3 when R becomes more and
more electron withdrawing. However the slope value that is larger
than 1 suggests that the position subjected to the R influence is
close to the substituted phenyl ring.
[44] G. Gabor, Y. Frei, D. Gegiou, M. Kaganowitch, E. Fischer, Isr. J.
Chem. 1967, 7, 193–221.
[45] H. Rau, Photochromism: Molecules and Systems (Eds.: H. Dꢄrr, H.
Bouas-Laurent), Elsevier, Amsterdam, 1990.
[46] D. Stefanidis, S. Cho, S. Dhe-Paganon, W. P. Jencks, J. Am. Chem.
Soc. 1993, 115, 1650–1656.
[47] N. Gagey, P. Neveu, C. Benbrahim, B. Goetz, I. Aujard, J.-B.
Baudin, L. Jullien, J. Am. Chem. Soc. 2007, 129, 9986–9998.
[48] N. Gagey, P. Neveu, L. Jullien, Angew. Chem. 2007, 119, 2519–2521;
Angew. Chem. Int. Ed. 2007, 46, 2467–2469.
[49] N. Gagey, M. Emond, P. Neveu, C. Benbrahim, B. Goetz, I. Aujard,
J.-B. Baudin, L. Jullien, Org. Lett. 2008, 10, 2341–2344.
[50] It was not possible to directly retrieve an accurate value for eT-
[1] M. Eigen, Angew. Chem. 1963, 75, 489–508; Angew. Chem. Int. Ed.
Engl. 1964, 3, 1–19.
[2] W. P. Jencks, Catalysis in Chemistry and Enzymology, Dover Publi-
cations, New York, 1987.
[3] R. N. Goldberg, N. Kishore, R. M. Lennen, J. Phys. Chem. Ref. Data
2002, 31, 231–370.
[4] S. Khan, F. Castellano, J. L. Spudich, J. A. McCray, R. S. Goody,
G. P. Reid, D. R. Trentham, Biophys. J. 1993, 65, 2368–2382.
[5] A. Barth, J. E. T. Corrie, Biophys. J. 2002, 83, 2864–2871.
[6] M. V. Georges, J. C. Scaiano, J. Phys. Chem. 1980, 84, 492–495.
[7] K. Janko, J. Reichert, Biochim. Biophys. Acta 1987, 905, 409–416.
[8] G. Bonetti, A. Vecli, C. Viappiani, Chem. Phys. Lett. 1997, 269, 268–
273.
[9] For a system capable of photodelivering pulses of the hydroxide
anion, see: M. Irie, J. Am. Chem. Soc. 1983, 105, 2078–2079.
[10] M. Eigen, L. de Mayer, Relaxation Methods in Techniques of Or-
ganic Chemistry, Vol. VIII, 2nd ed. (Eds.: S. L. Friess, E. S. Lewis, A.
Weissberger) Wiley, New York, 1963, pp. 895–1054.
[11] D. Baurecht, U. P. Fringeli, Rev. Sci. Instrum. 2001, 72, 3782–3792.
[12] K. Kurin-Csçrgei, I. R. Epstein, M. Orbꢅn, Nature 2005, 433, 139–
142.
[13] D. Liu, S. Balasubramanian, Angew. Chem. 2003, 115, 5912–5914;
Angew. Chem. Int. Ed. 2003, 42, 5734–5736.
ACHUTNGRENU(NG 347)FT!CACHTUNTGREN(NUGN 347) from the preceding data treatment. We were only
[14] T. Liedl, F. C. Simmel, Nano Lett. 2005, 5, 1894–1898.
[15] J. R. Howse, P. Topham, C. J. Crook, A. J. Gleeson, W. Bras,
R. A. L. Jones, A. J. Ryan, Nano Lett. 2006, 6, 73–77.
[16] M. Gutman, D. Huppert, E. Pines, J. Am. Chem. Soc. 1981, 103,
3709–3713.
[17] L. M. Tolbert, K. M. Solntsev, Acc. Chem. Res. 2002, 35, 19–27.
[18] U. Steiner, M. H. Abdel-Kader, P. Fischer, H. E. A. Kramer, J. Am.
Chem. Soc. 1978, 100, 3190–3197.
[19] S. H. Kawai, S. L. Gilat, J.-M. Lehn, Eur. J. Org. Chem. 1999, 2359–
2366.
[20] Y. Odo, K. Matsuda, M. Irie, Chem. Eur. J. 2006, 12, 4283–4288.
[21] G. Wettermark, J. Phys. Chem. 1962, 66, 2560–2562.
able to derive an upper limit equal to 500mꢀ1 cmꢀ1 (see Figure 8S in
the Supporting Information). In fact, we used the relative propor-
tions in C and T obtained by NMR spectroscopy (see Figure 3S) and
the more precisely derived value of eC
ACHUTNGTREN(GUN 347)FC!TACHUTNGTREN(NUGN 347) to get a more
precise value of eT(347)FT!C(347).
A
ACHTUNGTRENNUNG
[51] S. Espinosa, E. Bosch, M. Rosꢀs, Anal. Chem. 2000, 72, 5193–5200.
[52] J. Barbosa, V. Sanz-Nebot, J. Pharm. Biomed. Anal. 1992, 10, 1047–
1051.
[53] J. Barbosa, V. Sanz-Nebot, Anal. Chim. Acta 1991, 244, 183–191.
[54] P. F. Wang, L. Jullien, B. Valeur, J.-S. Filhol, J. Canceill, J.-M. Lehn,
New J. Chem. 1996, 20, 895–907.
8830
www.chemeurj.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 8822 – 8831

2-Hydroxyazobenzenes
FULL PAPER
[55] H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuberbꢄhler, Talanta
1985, 32, 95–101; H. Gampp, M. Maeder, C. J. Meyer, A. D. Zuber-
bꢄhler, Talanta 1985, 32, 257–264; H. Gampp, M. Maeder, C. J.
Meyer, A. D. Zuberbꢄhler, Talanta 1985, 32, 1133–1139; H. Gampp,
M. Maeder, C. J. Meyer, A. D. Zuberbꢄhler, Talanta 1986, 33, 943–
951.
[57] G. Bader, P. Deuflhard, Numer. Math. 1983, 41, 373–398.
[58] J. A. Nelder, R. Mead, Comput. J. 1965, 7, 308–313.
[59] Gnu Scientific Library, http://www.gnu.org/software/gsl/.
[56] W. H. Press, S. A. Teukolsky, W. T. Vetterling, B. P. Flannery, Numer-
ical Recipes in C: The Art of Scientific Computing, 2nd ed., Cam-
bridge University Press, New York, 1992.
Received: March 2, 2010
Published online: June 25, 2010
Chem. Eur. J. 2010, 16, 8822 – 8831
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
8831