Synthesis, characterization and photochromism of 3′-butoxyflavylium derivatives

Article abstract of DOI:10.1016/j.jphotochem.2012.06.013

The compounds 3′-butoxy-7-methoxyflavylium and 3′-butoxy-7- hydroxyflavylium were synthesized and the respective equilibrium and rate constants determined by two complementary techniques, pH jumps and flash photolysis. An experimental strategy based on th

Full text of DOI:10.1016/j.jphotochem.2012.06.013

Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
Contents lists available at SciVerse ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Synthesis, characterization and photochromism of 3ꢀ-butoxyflavylium derivatives
Sandra Gago, Vesselin Petrov, A. Jorge Parola, Fernando Pina∗
REQUIMTE, Departamento de Química, Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Monte de Caparica, Portugal
a r t i c l e i n f o
a b s t r a c t
The compounds 3ꢀ-butoxy-7-methoxyflavylium and 3ꢀ-butoxy-7-hydroxyflavylium were synthesized
and the respective equilibrium and rate constants determined by two complementary techniques, pH
jumps and flash photolysis. An experimental strategy based on these two techniques allowed calcula-
tion of all the equilibrium and rate constants of the system carried out for the first time in flavylium
compounds lacking of the high cis–trans isomerization barrier. Irradiation of the trans-chalcone gives
rise to the formation of the cis-chalcone still during the lifetime of the flash, which disappears through
two parallel reactions: (i) one leading to the recovery of the trans-chalcone and the other, (ii) forming
flavylium cation via hemiketal. This last reaction is globally dependent on pH and at less acidic pH the
system reverts back to the trans-chalcone. The highest yield of colour production upon the flash takes
place in the pH range 2–3.5.
Article history:
Received 1 May 2012
Received in revised form 6 June 2012
Accepted 16 June 2012
Available online 10 July 2012
Keywords:
Flavylium compounds
Photochromic
Anthocyanins
Flash photolysis
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
at very acidic pH values. As the pH is raised, other species are formed
in a more or less complex kinetic process, the respective details
Photochromic compounds, which have the property of changing
their colour upon light absorption and revert thermally to the ini-
tial state, seem to be magic and always fascinating [1]. In recent
years, a new family of photochromic compounds derived from
2-phenyl-1-benzopyrylium, the so-called synthetic flavylium com-
pounds, have been reported in literature [2,3]. Synthetic flavylium
compounds possess the same basic structure of anthocyanins, the
ubiquitous colorants of most flowers and fruits and follow together
being very dependent on the substitution pattern of the flavylium
backbone.
When sufficiently acidic solutions containing flavylium cations
bearing hydroxyl substituents are made less acidic by addition of
base (pH jumps), the first observation is the appearance of a dif-
ferent colour due to the deprotonation of the hydroxyl group to
form a quinoidal base, see Eq. (1) where this is exemplified for
4ꢀ,7-dihydroxyflavylium. All anthocyanins possess hydroxyl groups
and deprotonation usually leads to the appearance of blue coloured
quinoidal bases.
(1)
with deoxyanthocyanins [4–7] styrylflavylium [8–12] and naph-
thoflavylium analogues [13,14], the same network of chemical
reactions. The flavylium cation (red in anthocyanins), is stable only
In competition with this reaction the hydration of the flavylium
cation by water addition at position 2 takes place, Eq. (2)
(2)
These two parallel reactions occur in different scales of time:
proton transfer in the micro-second and hydration in seconds to
sub-seconds, depending on pH. By consequence A appears as a
∗
Corresponding author. Tel.: +351 212948355; fax: +351 212948550.
E-mail address: fp@fct.unl.pt (F. Pina).
1010-6030/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jphotochem.2012.06.013

S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
55
kinetic product that usually fades with time, because it is not the
most stable species at the equilibrium.
The hemiketal (chromene) is involved in a tautometic process
that leads to the ring opening with formation of a cis-chalcone form.
This reaction usually occurs in the sub-second timescale.
In spite of the fact that practically all flavylium compounds fol-
low the general network of chemical reactions above described
there are many differences on the mole fraction distribution of
species at the equilibrium and pseudo-equilibrium as well as on the
rates of conversion between the species, which are very dependent
(3)
on the substitution pattern [2]. Many flavylium compounds proved
to be versatile photochromic systems with potential applications
in solution [2] or incorporated in polymers [20], gels [21] and ionic
Finally the cis-chalcone isomerizes and gives the trans-chalcone,
inthetimescaleofsecondsormanydaysdependingonthecis–trans
isomerization barrier.
(4)
liquids [22]; in particular, those possessing a high cis–trans isomer-
ization barrier have been claimed as models for optical memories
[23]. In this work, the effect of the substituents in position 3ꢀ in the
performance of the photochromic system is reported. Preliminary
results suggested that the apparent acidity of the flavylium cation
in these compounds is very low, meaning that it is possible to have
the photoactive trans-chalcone species at low pH values, enlarging
the operational pH range of the photochromic system [24]. Two
3ꢀ-butoxyflavylium derivatives (Scheme 1) were synthesized and
studied by conjugation of pH jumps and flash photolysis. The use
of both techniques allowed to calculate with good accuracy all the
rate and equilibrium constants of both flavylium chemical reac-
tion networks, Eqs. (1)–(5). In particular flash photolysis has been
employed to characterize the flavylium photochromism but it is
the first time that is used extensively in a strategy that allows the
calculation of all the rate and equilibrium constants of the network
for compounds lacking the cis–trans isomerization barrier.
The photochemistry arises from the photoinduced isomeriza-
tion of the trans-chalcone to the cis-chalcone, which at moderately
acidic pH values is spontaneously transformed into the flavylium
cation or the quinoidal base (if hydroxyl substituents are present)
[2,15].
Eqs. (1)–(4) of the flavylium network are equivalent to a global
acid–base equilibrium defined by Eqs. (5) and (6), where CB refers
to all species formed from AH+, Eq. (7). Under these conditions, the
apparent acidity constant pKꢀa relates with the equilibrium con-
stants of equilibria (1)–(4) through Eq. (8) [16–18].
AH+ + 2H2O ꢀ CB + H3O+
(5)
(6)
[CB][H3O+]
Kaꢀ
=
[AH+]
[CB] = [Ct] + [Cc] + [B] + [A]
Kaꢀ = Ka + Kh + KhKt + KhKtKi
(7)
(8)
It is very useful to define a pseudo-equilibrium, a transient state
that is obtained when all the species are in (pseudo)equilibrium
except trans-chalcone due to its slower formation in comparison
with all the other kinetic steps, Eqs. (9)–(11).
2. Experimental
2.1. General
AH+ + 2H2O ꢀ CB∧ + H3O+
[CB∧][H3O+]
(9)
All reagents and solvents were of analytical grade. 1H and
2D 1H COSY NMR spectra were recorded at 400.13 MHz with a
Bruker AMX400. The flavylium salts were dissolved in CD3OD,
acidified with DCl or basified with NaOD. The assignments of
the several protons were based on the COSY experiments. Ele-
mental analyses (EA) were performed in a Thermofinnigan Flash
EA 112 series. Mass spectra were run on an Applied Biosystems
Voyager-DETM PRO. Spectroscopic experiments were carried out
in buffered [25] water/ethanol, 70:30 (v/v) solvent mixture. pH
values were adjusted by the addition of a 0.1 M NaOH aqueous
solution or a 0.1 M HCl aqueous solution, and measured with a
MeterLab pHM240 pH meter from Radiometer Copenhagen. For
solutions with HCl concentration above 0.1 M, pH was calculated
from −log [HCl]. UV/Vis absorption spectra were recorded with a
Varian-Cary 100 Bio spectrophotometer or in a Shimadzu VC2501-
PC. The stopped flow experiments were conducted in an Applied
Ka∧
=
(10)
[AH+]
[CB∧] = [Cc] + [B] + [A]
Ka∧ = Ka + Kh + KhKt
(11)
(12)
The determination of the thermodynamic and kinetic constants
of the network of chemical reactions is carried out by changing the
equilibrium conditions from one state to another and follow the
relaxation of the system, generally by UV–vis or 1H NMR. In some
cases, as in the present compounds, the system is photochromic and
a great profit can be taken from transient absorption measurements
(flash photolysis). Once the equilibrium constants determined, an
energy level diagram (as the ones reported in Schemes 2 and 3) can
be obtained as well as the pH dependent mole fraction distribution
of species according to Eq. (13) [19].
[H+]
[H+] + Kaꢀ
Kaꢀ
Kh
ꢀAH+ =
;
ꢀA
=
;
ꢀB
=
;
H3CO
HO
O
[H+] + Kaꢀ
[H+] + Kaꢀ
O
+
+
O
O
KhKt
[H+] + Kaꢀ
KhKtKi
[H+] + Kaꢀ
ꢀCc
=
;
ꢀCt
=
(13)
Scheme 1. Structures of 3ꢀ-butoxy-7-methoxyflavylium and 3ꢀ-butoxy-7-
hydroxyflavylium.

56
S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
Photophysics SX20 stopped-flow spectrometer provided with a
PDA.1/UV photodiode array detector. Irradiation experiments were
carried out on a spectrofluorimeter Spex Fluorolog 1681 at the
wavelength 377 nm without slits. Light intensity was measured
by ferrioxalate actinometry [26]. The flash photolysis experiments
were performed on a Varian Cary 5000 spectrophotometer with
a Harrick FiberMate attached to the CUV-ALL-UV 4-way cuvette
holder compartment (Ocean Optics) on the external side of the
sample holder in order to perform light excitation perpendicular
to the analyzing beam and with the sample compartment shielded
with black cardboard and black tape. As a pulsed light source a com-
mercially available Achiever 630AF camera flash was used, placed
in close contact with the sample holder. The excitation was made
with the white light of the camera flash with a time resolution of
ca. 0.05 s. Further details are described elsewhere [27].
2.4. trans-Chalcone of 3ꢀ-butoxy-7-hydroxyflavylium
hydrogensulphate
Dissolving 3ꢀ-butoxy-7-hydroxyflavylium hydrogensulphate in
CD3OD with a drop of D2O and allowing equilibration overnight, in
the dark, leads to a mixture containing >80% Ct. 1H RMN (CD3OD,
pD ≈ 2.5, 400.13 MHz, 298 K) ı (ppm): 8.00 (1H, H4, d, 3J = 15.6 Hz),
7.58 (1H, H5ꢀ, dd, 3J = 3.8 Hz), 7.56 (1H, H3, d, 3J = 15.6 Hz), 7.52 (1H,
H5, d, 3J = 8.4 Hz), 7.48 (1H, H2ꢀ, s), 7.45 (1H, H6ꢀ, d, 3J = 8.0 Hz), 7.18
(1H, H4ꢀ, dd, 3J = 8.4 Hz, 4J = 2.0 Hz), 6.37 (2H, H6 + H8, m), 4.08 (2H,
t, R4), 1.82 (2H, m, R3), 1.57 (2H, m, R2), 1.04 (3H, m, R1).
2.5. Synthesis of 3ꢀ-butoxy-7-methoxyflavylium
hydrogensulphate
2.2. Synthesis of 3ꢀ-butoxyacetophenone (this procedure was
adapted from Refs. [28,29])
A mixture of 3ꢀ-hydroxyacetophenone (1.0 g, 7.3 mmol), 1.5
equivalents of chlorobutane (1.01 g, 10.9 mmol), 3 equivalents of
potassium carbonate (3.03 g, 21.9 mmol) and a catalytic amount of
potassium iodide (0.10 g) in dry DMF (20 mL) was heated at 90 ◦C
overnight, under inert atmosphere. In the following day the reac-
tion was monitored by TLC (ethyl acetate:hexane, 2:8) and addition
of chlorobutane was needed. After completion of the reaction, the
mixture was cooled to room temperature, filtered and the result-
ing solution combined with water (20 mL) and extracted with ethyl
acetate (4 × 20 mL). The organic phase was washed with water
(3 × 15 mL), dried over anhydrous sodium sulphate, filtered and the
solvent removed on a rotary evaporator. A yellow oil was obtained
and purified by column chromatography on silica gel (Yield: 1.26 g,
90%).
This compound was prepared as described above but
using 2-hydroxy-4-methoxybenzaldehyde instead of 2,4-
dihydroxybenzaldehyde. Yield: 0.87 g, 83%, dark orange solid.
1H RMN (CD3OD + DCl, 400.13 MHz, 298 K) ı (ppm): 9.35 (1H,
H4, d, 3J = 8.4 Hz), 8.55 (1H, H3, d, 3J = 8.4 Hz), 8.28 (1H, H5, d,
3J = 9.2 Hz), 8.07 (1H, H6ꢀ, d, 3J = 8.0 Hz), 7.93 (1H, H2ꢀ, s), 7.90
(1H, H8, s); 7.69 (1H, H5ꢀ, dd, 3J = 10.0 Hz), 7.60 (1H, H6, dd,
3J = 9.2 Hz, 4J = 1.6 Hz), 7.40 (2H, H4ꢀ, dd, 3J = 8.4 Hz, 4J = 1.6 Hz),
4.18 (2H, t, R4), 1.83 (2H, m, R3), 1.56 (2H, m, R2), 1.02 (3H, t,
R1); MALDI-TOF/MS: m/z (%): calcd for C20H21O3+: 309.15; found:
309.11 [M+] (100%). The compound was recrystallized in the
perchlorate form for elemental analysis purpose, by dissolving
in the minimum amount of ethanol and adding perchloric acid
to precipitate the product. The red solid obtained was filtered off
and washed several times with diethyl ether. Elemental analysis
(%) calcd for C20H21O3·ClO4: C 58.76, H 5.18; found: C 58.89, H
5.25.
1H RMN (CDCl3, 400.13 MHz, 298 K) ı (ppm): 7.50 (1H, d,
3J = 7.3 Hz), 7.47 (1H, s), 7.34 (1H, t, 3J = 15 Hz), 7.09 (1H, d,
3J = 6.8 Hz), 3.99 (2H, t, R4), 2.58 (3H, s, Rꢀ1); 1.77 (2H, q, R3), 1.49
(2H, q, R2), 0.97 (3H, t, R1).
2.6. trans-Chalcone of 3ꢀ-butoxy-7-methoxyflavylium
2.3. Synthesis of 3ꢀ-butoxy-7-hydroxyflavylium
hydrogensulphate
Dissolving 3ꢀ-butoxy-7-hydroxyflavylium hydrogensulphate in
CD3OD with a drop of D2O and allowing equilibration overnight, in
the dark, leads to a mixture containing >98% Ct. 1H RMN (CD3OD,
pD ≈ 3.5, 400.13 MHz, 298 K) ı (ppm): 7.99 (1H, H4, d, 3J = 15.8 Hz),
7.60 (1H, H5, d, 3J = 8.8 Hz), 7.58 (1H, H3, d, 3J = 15.8 Hz), 7.57 (1H,
H5ꢀ, dd, 3J = 8.0 Hz), 7.48 (1H, H6ꢀ, d, 3J = 8.0 Hz), 7.43 (1H, H2ꢀ, s),
7.20 (1H, H4ꢀ, dd, 3J = 7.6 Hz), 6.55 (1H, H6, dd, 3J = 8.8 Hz), 6.50 (1H,
H8, d), 4.06 (2H, t, R4), 3.82 (1H, s, OCH3), 1.76 (2H, m, R3), 1.48
(2H, m, R2), 0.96 (3H, t, R1).
A solution of 3ꢀ-hydroxyacetophenone (0.5 g, 2.6 mmol) with
one equivalent of 2,4-dihydroxybenzaldehyde (0.36 g) in a mixture
of glacial acetic acid/concentrated sulphuric acid (4 mL:1 mL) was
stirred overnight at room temperature. Diethyl ether was added
to the red solution and a precipitate was formed. This dark yel-
low solid was filtered, washed several times with diethyl ether
and dried under vacuum. Yield: 0.70 g, 69%. 1H RMN (CD3OD + DCl,
400.13 MHz, 298 K) ı (ppm): 9.33 (1H, H4, d, 3J = 8.4 Hz), 8.51 (1H,
H3, d, 3J = 8.4 Hz), 8.29 (1H, H5, d, 3J = 9.2 Hz), 8.06 (1H, H6ꢀ, d,
3J = 7.6 Hz), 7.95 (1H, H2ꢀ, s), 7.67 (1H, H8, s); 7.64 (1H, H5ꢀ, t,
3J = 8.0 Hz), 7.53 (1H, H6, dd, 3J = 9.0 Hz, 4J = 2.2 Hz), 7.40 (2H, H4ꢀ, dd,
3. Results and discussion
3.1. 3ꢀ-Butoxy-7-methoxyflavylium
3J = 8.0 Hz, 4J = 2.4 Hz), 4.18 (2H, t, R4), 1.83 (2H, m, R3), 1.56 (2H, m,
The absorption spectra of equilibrated solutions of the com-
pound 3ꢀ-butoxy-7-methoxyflavylium in ethanol:water (30:70), in
the dark, are shown in Fig. 1A for the pH range 2 M (HCl) < pH < 4.97
and in Fig. 1B for 3.30 < pH < 10.75. At lower pH values, the
+
R2), 1.02 (3H, t, R1); MALDI-TOF/MS: m/z (%): calcd for C19H19O3
:
295.13; found: 295.11 [M+] (100%); Elemental analysis (%) calcd for
19H19O3·HSO4: C 58.15, H 5.14; found: C 58.28, H 5.17.
C

S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
57
Fig. 1. Spectral variations of 3ꢀ-butoxy-7-methoxyflavylium1.19 × 10−5 M in ethanol:water (30:70) upon direct pH jumps from a mother solution in 0.24 M HCl (ethanol:water,
30:70) to several pH values, after equilibration in the dark overnight. (A) 2 M HCl < pH < 4.97; inset: fitting of the absorption at 440 nm yields pKꢀa = 0.91. (B) 3.30 < pH < 10.75;
inset: simultaneous fitting of the absorbance at 377 nm (᭹) and 453 nm (ꢁ) leads to pKCt = 8.97. (C) Spectral variations over time upon at pH jump to pH 1.38; inset: fitting
the kinetic trace at 440 nm as a monoexponential decay leads to kobs = 2.7( 0.2) × 10−4 s−1. (D) pH dependence of kobs determined from direct pH jumps (ꢁ) and from flash
photolysis (᭹); fitting with Eq. (14) was achieved with 1.5( 0.1) × 10−6/([H+] + 3.4( 0.2) × 10−5) + 1.25( 0.1) × 10−5
.
spectral variations are compatible with an apparent acid–base type
equilibrium, as reported in Eq. (5), involving the flavylium cation,
AH+, and trans-chalcone, Ct, with pKꢀa = 0.9( 0.1). The assign-
ment of the species with ꢁmax = 377 nm as the trans-chalcone
was confirmed by 1H NMR at pD = 3.5, through the large value
of 15.8 Hz for the coupling constant between protons H3 and
H4 (see Section 2) [30]. At higher pH values, Ct deprotonates
to give unprotonated trans-chalcone, Ct−, with pKCt = 8.97( 0.05),
Fig. 1B.
Direct pH jumps from solutions of flavylium cation in 0.24 M
HCl (ethanol:water, 30:70) to higher pH values, as shown in Fig. 1C,
lead to spectral variations compatible with a first-order kinetic pro-
cess where AH+ disappears to give Ct. The rate of this process as
a function of pH is shown in Fig. 1D, white circles. The shape of
the curve is characteristic of a kinetic process controlled by the
cis–trans isomerization [31]. In the same figure, the rate constants
obtained from flash photolysis experiments where the recovery of
Ct is measured after flash irradiation (see below) are also reported
(black circles).
KhKt
[H+] + Kh(1 + Kt)
kobs
=
ki + k
(14)
−i
Considering Eqs. (2)–(4), the rate constant of trans-chalcone for-
mation is given by Eq. (14) [2], used to fit the experimental data in
Fig. 1D. The fitting was achieved for the values defined through Eqs.
(15)–(17).
KhKtki = 1.5( 0.1) × 10−6 M−1 s−1
Kh(1 + Kt) = 3.4( 0.2) × 10−5 M−1
(15)
(16)
(17)
k
= 1.25( 0.1) × 10−5 s−1
−i

58
S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
From the value of Kꢀa = 10−0.91 (Fig. 1A) and subtracting Eq. (16)
from Eq. (8), Eq. (18) is obtained (Eq. (8) for this compound has no
Ka since there are no OH groups to form quinoidal base).
KhKtKi = 0.12( 0.02) M−1
(18)
The ratio KhKtKi/Kꢀa = 0.99 corresponds to the maximum of the
mole fraction of Ct at the equilibrium confirming that [CB] ≈ [Ct]
and the contribution of the other basic species is negligible. The
ratio between Eqs. (15) and (18) confirms the value of k
.
−i
3.2. Photochemistry
Scheme 2. Energy level diagram and rate constants for the compound 3ꢀ-butoxy-
7-methoxyflavylium in ethanol:water, 30:70, at 294 K.
Similarly to other flavylium chemical reaction networks,
3ꢀ-butoxy-7-methoxyflavylium is photochromic. The spectral vari-
ations upon continuous irradiation at 377 nm of the respective
trans-chalcone at pH 2.0 is shown in Fig. 2. Flavylium cation is
formed with a quantum yield of 0.1, and the system recovers ther-
mally with a rate constant of 6( 0.05) × 10−4 s−1 (Fig. 2B).
the pH dependence of the rate constants for the faster and for the
slower steps, respectively.
The faster process can thus be assigned to the competition
between two parallel reactions: a forward reaction to give AH+
and a back reaction to recover Ct. Regarding the formation of AH+,
a sequence of tautomerization and dehydration reactions should
occur and the question is which of them is the rate determining step
for flavylium formation after the flash. Anticipating the quantita-
tive energy level diagram that could be drawn upon determination
of the equilibrium (and kinetic) constants [33], the flash photolysis
behaviour can be rationalized by Scheme 2.
The Cc species formed upon flash irradiation disappears via
routes a plus b in Scheme 2 (parallel reactions). While the rate via
route a is approximately equal to ki + k−i ≈ ki (ki > >k−i), the rate via
route b exhibits two potential dependences on pH: (i) the hydra-
tion reaction through the term k−h[H+] and (ii) the tautomerization
reaction due to the possibility of basic and acid catalysis, an effect
previously reported by McClelland and co-workers and already
observed in many flavylium systems [2,34]. Low pH values tend
to favour formation of flavylium cation while at higher pH values
Ct is completely recovered through route a.
More information regarding the system was obtained through
flash photolysis experiments carried out at different pH values,
as exemplified in Fig. 3 for pH 3.19 and pH 5.6. At pH 3.19, the
flavylium absorption measured at 444 nm changes according to a
bi-exponential law, the faster step accounting for an increase in
absorbance immediately after the flash corresponding to the for-
mation of AH+ from Cc via B, and the slower one consisting in a
decrease on absorbance due to the thermal recovery of the sys-
tem to give back Ct, the thermodynamically stable species at this
pH value. On the other hand, the absorption changes of the trans-
chalcone followed at 377 nm (Fig. 3A) show an initial bleaching
occurring during the flash lifetime, an indication that Ct was con-
sumed to give Cc [32], followed by a recovery of Ct also taking
place through two kinetic steps, the faster corresponding to the
formation of Ct from Cc and the slower consisting in the recovery
of Ct from AH+. At pH 5.6, Fig. 3B, the formation of AH+ is almost
imperceptible (top) and the trace at 377 nm (bottom) is nearly a
monoexponential law. The values of the two rate constants of both
bi-exponential processes, measured at 444 nm and 377 nm, are the
same within experimental error (ca. 10%), see Figs. 4A and 1D for
The rate constants of the faster process determined from
flash photolysis experiments carried out at several pH values are
Fig. 2. (A) Spectral variations accompanying the irradiation at 377 nm of 3ꢀ-butoxy-7-methoxyflavylium at pH 2.0 (2 mL of solution, 2.06 × 10−5 M, ethanol:water, 30:70) for
the following irradiation times: 0, 0.67, 1.33, 2, 3, 4, 5, 6, 7, 9, 11, 13, 15 min (I0 = 2.19 × 10−7 Einstein min−1). (B) Thermal recovery of the irradiated solution at 294 K; inset:
fitting of the absorbance at 438 nm as a monoexponential decay leads to kobs = 6.0( 0.05) × 10−4 s−1
.

S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
59
Fig. 3. Flash photolysis of 3ꢀ-butoxy-7-methoxyflavylium 2.74 × 10−5 M in ethanol:water, 30:70 at pH 3.19 (A) and at pH 5.6 (B).
presented in Fig. 4. Inspection of these plots suggests a change of
regime on the pH dependence of route b (the contribution of route
a for kobs is pH independent) with kobs increasing with increas-
ing pH until pH ≈ 3.7 and decreasing thereon. At low pH values,
the hydration is expected to be relatively fast (it is proportional to
[H+]) and the kinetics is controlled by the tautomerization reaction.
The small increase of the tautomerization rate can be attributed
to a basic catalysis [34] (no acidic catalysis was detected in this
pH range, otherwise an increase of the rate would be expected at
low pH values). However, at higher pH values the hydration rate
decreases and becomes the rate-determining step. In conclusion,
the rate constants for the branch controlled by the tautomerization
reaction should follow Eq. (19) while the branch controlled by the
hydration reaction should follow Eq. (20). In Eq. (19), it was consid-
ered that since the hydration is fast, once B is formed it immediately
transforms into AH+ and no reversibility from B to Cc was included.
On the other hand, in Eq. (20) it was considered that B and Cc have
time to (pseudo) equilibrate before the hydration occurs [2].
kobs(Taut) = k + k + kOH[OH−]
(19)
(20)
−t
i
−t
kiKt
k
−h[H+]
kobs(Hyd)
=
+ kh +
1 + Kt
1 + Kt
Fitting of the data in Fig. 4A was achieved by means of Eq. (19) with
ki + k−t = 0.4 s−1 and kOH = 2.5( 0.3) × 109 M−1 s−1 (tautomeriza-
−t
tion control) and Eq. (20) with kiKt/(1 + Kt) + kh = 0.085 s−1 and
k
−h = 2.0( 0.3) × 103 M−1 s−1 (hydration control).
In order to confirm this interpretation, the pH dependent mole
fraction of flavylium cation formed after the first kinetic step
upon flash irradiation was calculated using the kinetic traces
at 377 and 444 nm for each pH (Fig. 3) and are shown in
Fig. 4B. The mole fraction of flavylium cation was calculated
Fig. 4. (A) Observed rate constants for the faster process upon flash irradiation as
a function of pH; fitting for the tautomerization control range (pH < 3.7)
attained through Eq. (19) with 0.4( 0.1) + 2.5( 0.3) × 109[OH−] and for the hydration control range (pH > 3.7) attained through Eq. (20) with 0.04 + 2.0( 0.2) × 103[H+].
(B) Mole fraction of the flavylium cation after the first kinetic step upon flash irradiation as
a function of pH; fitting achieved with Eq. (21) for
(0.29( 0.1) + 2.5( 0.3) × 109[OH−])/(0.4( 0.1) + 2.5( 0.3) × 109[OH−]).

60
S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
Fig. 5. Spectral variations of 3ꢀ-butoxy-7-hydroxyflavylium 1.1 × 10−5 M in ethanol:water (30:70) upon direct pH jumps from a mother solution in 0.24 M HCl (ethanol:water,
30:70) to several pH values, after equilibration in the dark overnight. (A) 2 M HCl < pH < 2.16; inset: fitting of the absorption at 444 nm yields pKꢀa = 1.16( 0.08). (B)
2.76 < pH < 12.55. (C) Simultaneous fitting of the absorbance at 495 nm (᭹) and 377 nm (ꢁ) over the pH range in (B) leads to pKCt = 7.95( 0.1) and pKCt− = 10.45( 0.1);
traced lines correspond to the mole fraction distribution of the trans-chalcone species.
as ꢀAH+ = [AH+]/[Ct]0 with [AH+] = A444/ε444(AH+) and [Ct]0
=
3.3. 3ꢀ-Butoxy-7-hydroxyflavylium
A0377/ε377(Ct), considering b = 1 cm. The value of A444 was obtained
from the respective kinetic trace after the first kinetic pro-
cess, the value of A0377 from the respective kinetic trace at t = 0
(immediately after the flash), ε444(AH+) = 24,740 M−1 cm−1 and
ε377(Ct) = 17,870 M−1 cm−1 (Fig. 1A).
The pH dependent spectral variations of dark equilibrated solu-
tions of 3ꢀ-butoxy-7-hydroxyflavylium (ethanol:water, 30:70) are
shown in Fig. 5A for a more acidic region (2 M HCl < pH < 2.16) and
in Fig. 5B for a larger pH range (2.8 < pH < 12.6).
Fitting the data for the more acidic region in Fig. 5A leads
to pKꢀa = 1.16( 0.08). The data is compatible with an equilib-
rium between the flavylium cation and a species exhibiting an
absorption band which shape and position is characteristic of the
trans-chalcone, Ct. This interpretation was corroborated by 1H
NMR where Ct is clearly identified by the large coupling constant
(15.6 Hz) between protons H3 and H4 (see Section 2).
Fitting of the data in Fig. 4B was made with Eq. (21) that accounts
for the fraction of flavylium cation obtained under tautomerization
control. In the case of the hydration control, the system essentially
reverts back to Ct and the separation between the slow process
upon flash irradiation and the thermal recovery cannot be made
with sufficient accuracy from the traces at 377 nm.
In Fig. 5B, the spectral variations show red-shits compatible with
successive formation of unprotonated trans-chalcones, Ct− and
Ct2−. Simultaneous fitting of the absorbances at 377 and 495 nm
leads to pKCt = 7.95( 0.1) and pKCt− = 10.45( 0.1), Fig. 5C.
Unlike the previous flavylium, 3ꢀ-butoxy-7-hydroxyflavylium
has a OH group in position 7 allowing the formation of a
quinoidal base upon deprotonation. Determination of the pKa was
achieved by means of stopped-flow experiments, Fig. 6. Solutions
of flavylium cation (0.24 M HCl in ethanol:water, 30:70) were
submitted to pH jumps to higher pH values and the absorbance
measured immediately after (dead time of the stopped-flow lower
than 10 ms). The absorbance variations as a function of pH are com-
patible with an acid–base equilibrium between flavylium cation
and quinoidal base, Eq. (1), with pKa = 4.6( 0.1), see inset in Fig. 6A.
The kinetics of direct pH jumps is exemplified for final pH val-
ues of 6.75 and 12.55 in Fig. 6B and C, respectively. In Fig. 6B, the
quinoidal base evolves to the trans-chalcone Ct, while in Fig. 6C
the same quinoidal base leads to the ionized trans-chalcone Ct2−
(ꢁmax = 495 nm). This procedure was extended to cover the pH
k
+ kOH[OH−]
−t
−t
ꢀAH+ =
(21)
k + k + kOH[OH−]
−t
i
−t
OH
The fitting was achieved using the values of k = 2.5 ×
−t
109 M−1 s−1 and ki + k−t = 0.4 s−1 determined above, allowing to
obtain k−t = 0.29 s−1 and consequently ki = 0.4 − 0.29 = 0.11 s−1
Substitution of ki in Eq. (15) permits to obtain
.
KhKt = 1.36 × 10−5 M−1
(22)
Substitution of Eq. (22) in Eq. (16) gives Kh = 2.0 × 10−5 M−1 and
from Eq. (22), Kt = 0.69, leading to kt = 0.2 s−1; Eqs. (18) and (22)
allow to obtain Ki = 8.8 × 104.
In conclusion, the data obtained from direct pH jumps and flash
photolysis can be fitted with the same set of constants presented
in Table 1.
Table 1
Thermodynamic and kinetic constants of 3ꢀ-butoxy-7-methoxyflavylium.a
(M−1 s−1
)
OH
−t
pKꢀa
Kh (M−1
)
Kt
Ki
ki (s−1
)
k
−i
(s−1
)
kt (s−1
)
k
(s−1
−t
)
kh (s−1
)
k
−h
(M−1 s−1
)
k
0.9
2.0 × 10−5
0.69
8.8 × 103
0.11
1.25 × 10−5
0.2
0.29
0.04
2.05 × 103
2.5 × 109
a
Estimated error is about 10%.
Table 2
Rate and equilibrium constants of 3ꢀ-butoxy-7-hydroxyflavylium obtained from Figs. 5 and 6.
(1)
(2)
(2)
(2)
(2)
pKꢀa
pKa
KhKtKi(2)
kh
0.023 s−1
k
KhKtki(2)
Ktki/k
−h
−i
1.2
4.6
0.063 M−1
5.7 × 10−5 s−1
4.0 × 10−6 s−1 M−1
1.9 × 10−4
M

S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
61
Fig. 6. (A) Spectral variations of 3ꢀ-butoxy-7-hydroxyflavylium 5.43 × 10−5 M in ethanol:water (30:70) upon direct pH jumps from a mother solution in 0.24 M HCl
(ethanol:water, 30:70) to several pH values in the range 2.86 < pH < 6.98, followed by stopped-flow (dead time ca. 10 ms); inset: fitting of the absorption at 484 nm
yields pKa = 4.6( 0.1). (B) Spectral variations over time upon at pH jump to pH 6.75; inset: fitting the kinetic trace at 378 nm as a monoexponential decay leads to
kobs = 4.53( 0.05) × 10−4 s−1. (C) Same as in B to pH 6.75; inset: fitting the kinetic trace at 495 nm as a monoexponential decay leads to kobs = 4.0( 0.1) x10−4 s−1. (D)
pH Dependence of kobs determined from direct pH jumps; fitting with Eq. (23) for pKa = 4.6( 0.1), KhKtki = 4.0 × 10−6 s−1 M, Ktki/k−h = 1.9 × 10−4 M, k−i = 5.7 × 10−5 s−1 and
k
OH = 1.5( 0.4) s−1 M−1
.
range 1–12, the respective observed rate constants being reported
in Fig. 6D. A bell-shaped curve is obtained in acidic medium and
a linear dependence with [OH−] concentration occurs in basic
medium, following Eq. (23).
cis–trans isomerization while at higher pH values the hydration
is the rate-determining step. In basic medium, the direct attack
of OH− to the base leads to the ionized cis-chalcones that iso-
merize to form the respective ionized trans-chalcones. Fitting
of the data in Fig. 6D was obtained for the constants reported
in Table 2.
As in the case of the previous compound, irradiation of the
Ct species of 3ꢀ-butoxy-7-hydroxyflavylium leads to formation
of flavylium cation, in this case with ˚ = 0.05, Fig. 7A. The
thermal recovery from Ct back to the flavylium occurs with
kobs = 4( 0.1) × 10−4 s−1 at pH 2.0.
([H+]/([H+] + Ka))KhKtki + k−i[H+]
kobs
+ kOH[OH−]
(23)
((Ktki)/k−h) + [H+]
The bell shaped part of Eq. (23) is obtained considering the kinetic
scheme shown in Eq. (24).
k
k
/(1+K )
t
−i
−h
Ct
ꢀ
Cc ꢀ B
ꢀ
AH+ ꢀ A + H+
(24)
+
+
(K /(1+K ))k
(([H ])/([H +K ]))k
t
t
a
i
h
In this scheme, the equilibrium between B and Cc is considered
to be much faster than the other processes, except the acid–base
equilibrium AH+/A, which is the fastest process of the network.
At lower pH values, the system is kinetically controlled by the
3.4. Flash photolysis
The flash photolysis traces represented in Fig.
the formation and disappearance of flavylium cation after the
8 regard

62
S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
Fig. 7. (A) Spectral variations accompanying the irradiation at 377 nm of 3ꢀ-butoxy-7-hydroxyflavylium at pH 2.0 (2 mL of solution, 1.91 × 10−5 M, ethanol:water, 30:70)
for the following irradiation times: 0, 0.5, 1, 2.5, 5, 7, 9, 11 min (I0 = 2.19 × 10−7 Einstein min−1). (B) Thermal recovery of the irradiated solution at 294 K; inset: fitting of the
absorbance at 438 nm as a monoexponential decay leads to kobs = 4.0( 0.1) × 10−4 s−1
.
Fig. 8. Flash photolysis of 3ꢀ-butoxy-7-hydroxyflavylium 3.3 × 10−5 M in ethanol:water, 30:70 at pH 4.09 (A) and at pH 6.2 (B).
flash (444 nm) and the bleaching and recovery of the trans-
When the kinetics is controlled by the tautomerization Eqs. (19)
and (21) are still valid, but when hydration is the rate-determining
step equilibration of flavylium cation with quinoidal base intro-
duces reversibility through the hydration constant and Eq. (20)
should be substituted by Eq. (25)
chalcone (377 nm). At pH 4.09, both traces are fitted with
two exponentials with the same lifetimes, kobs1 = 0.4( 5%) s−1
and kobs2 = 0.005( 5%) s−1, while at pH 6.2 the data is suit-
ably fitted by a single exponential with kobs = 0.2 s−1. The rate
constants for the faster process are plotted as
a function
kiKt
[H+]
kobs1
=
+ kh
+ k−h[H+]
(25)
of pH in Fig. 9A while those of the slower process fol-
low the bell shaped pH dependence shown in Fig. 6D. The
behaviour is qualitatively similar to that of the previous flavylium
compound.
[H+] + Ka
1 + Kt
As in the previous example of the methoxyflavylium analogous
the flash photolysis allows to obtain all the rate and equilibrium
Table 3
Rate and equilibrium constants of 3ꢀ-butoxy-7-hydroxyflavylium.
pKa
Kh (M−1
)
Kt
Ki
kh (s−1
)
k
−h
(s−1
)
kt (s−1)
k
−t
(s−1
)
ki (s−1
)
k
(s−1
)
−i
4.6
1.9 × 10−5
0.16
1.4 × 104
0.02
1.1 × 103
0.07
0.31
0.89
6.4 × 10−5

S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
63
Fig. 9. (A) Observed rate constants for the faster process upon flash irradiation of solutions of 3ꢀ-butoxy-7-hydroxyflavylium as a function of pH; fitting for the tau-
tomerization control range (pH < 3.5) attained through Eq. (19) with 1.2 + 1.0 × 1010[OH−] and for the hydration control range (pH > 3.5) attained through Eq. (25) with
0.17 + (0.023[H+])/([H+] + 2.5( 0.3) × 10−5) + 1.0( 0.5) × 1010[H+]. (B) Mole fraction of the flavylium cation after the first kinetic step upon flash irradiation as a function of
pH; fitting achieved with Eq. (21) for (0.31 + 1.0( 0.5) × 1010[OH−])/(1.2 + 1.0 × 1010[OH−]).
could result from the low isomerization rate or from the small frac-
tion of Cc available to isomerize, as reported for 3ꢀ-butoxyflavylium
derivatives.
Acknowledgments
This work was supported by Fundac¸ ão para a Ciência e Tecnolo-
gia, National NMR Network, project PTDC/QUI-QUI/117996/2010,
Post-doc grant SFRH/BPD/18214/2004 (VP).
Scheme 3. Energy level diagram and rate constants for the compound 3ꢀ-butoxy-
7-hydroxyflavylium in ethanol:water, 30:70, at 294 K.
References
[1] J.C. Crano, R.J. Guglielmetti (Eds.), Organic Photochromic and Thermochromic
constants of the system. Regarding the tautomerization control
Compounds, vol. 1, Plenum Press, NY, 1999.
the limits of Eq. (19) at lower pH values and (21) are respectively
[2] F. Pina, M.J. Melo, C.A.T. Laia, A.J. Parola, J.C. Lima, Chemical Society Reviews 41
(2012) 869–908.
[3] F. Pina, M.J. Melo, A.J. Parola, M. Maestri, V. Balzani, Chemistry A: European
Journal 4 (1998) 2001–2007.
ki + k−t = 1.2 s−1 and k−t/(ki + k−t) = 0.26, leading to ki = 0.89 s−1 and
to k−t = 0.31 s−1
.
The fitting of Fig. 9A is well defined and according to Eq. (25)
should be given by kiKt/(1 + Kt) = 0.17, which allows to calculated
Kt = 0.23( 0.05) substitution of Kt in 3rd column of Table 2 leads
to KhKi = 0.27 M−1. Substitution of ki in 6th column of Table 2 gives
KhKt = 4.4 × 10−6 M−1 and Kh = 1.9 × 10−5 M−1 and Ki = 1.4 × 104 and
[4] J.G. Sweeny, G.A. Iacobucci, Journal of Agricultural and Food Chemistry 31
(1983) 531–533.
[5] P. Coggon, G.A. Moss, H.N. Graham, G.W. Sanderson, Journal of Agricultural and
Food Chemistry 21 (1973) 727–733.
[6] R. Brouillard, G.A. Iacobucci, J.G. Sweeny, Journal of the American Chemical
Society 104 (1982) 7585–7590.
[7] M.J. Melo, S. Moura, A. Roque, M. Maestri, F. Pina, Journal of Photochemistry
and Photobiology A: Chemistry 135 (2000) 33–39.
thus k and kt are obtained.
−i
A reasonable fitting of the fraction of flavylium cation formed
[8] L. Jurd, USA Patent 3301683.
[9] J.S. Buck, I.M. Heilbron, Journal of the Chemical Society 123 (1923) 2521–2531.
[10] D. Amic, N. Trinajstic, D.D. Amic, Journal of the Chemical Society: Perkin Trans-
actions 2 (11) (1992) 1933–1938.
[11] R. Gomes, A.M. Diniz, A. Jesus, A.J. Parola, F. Pina, Dyes and Pigments 81 (2009)
69–79.
[12] A.M. Diniz, R. Gomes, A.J. Parola, C.A.T. Laia, F. Pina, Journal of Physical Chem-
istry B 113 (2009) 719–727.
[13] M. Elhabiri, P. Figueiredo, F. George, J.-P. Cornard, A. Fougerousse, J.-C. Martin,
R. Brouillard, Canadian Journal of Chemistry 74 (1996) 697–706.
[14] R. Gavara, V. Petrov, F. Pina, Photochemical and Photobiological Sciences 9
(2010) 298–303.
[15] H. Horiuchi, A. Tsukamoto, T. Okajima, H. Shirase, T. Okutsu, R. Matsushima, H.
Hiratsuka, Journal of Photochemistry and Photobiology 205 (2009) 203–209.
[16] R. Brouillard, B. Delaporte, Journal of the American Chemical Society 99 (1977)
8461–8468.
[17] R. Brouilard, B. Delaporte, J.-E. Dubois, Journal of the American Chemical Society
100 (1978) 6202–6205.
after the flash using Eq. (21) and the data from Table 3 is shown in
Fig. 9B.
The energy level diagram for this compound is shown in
Scheme 3.
4. Conclusions
Flash photolysis is an excellent technique to complement the pH
jumps commonly used in the study of the photochromic flavylium
compounds network of chemical reactions. Conjugation of the two
techniques allows calculation of all the kinetic and thermody-
namic constants of the system and this information is conveniently
presented in an energy level diagram. The rate of trans-chalcone
appearance from flavylium cation upon a pH jump is not only a
function of the cis–trans isomerization barrier, but also of the frac-
tion of cis-chalcone available to isomerize; slow appearance of Ct
[18] F. Pina, Journal of the Chemical Society: Faraday Transactions 94 (1998)
2109–2116;
(Correction in Scheme 1 see) F. Pina, Journal of the Chemical Society: Faraday
Transactions 94 (1998) 3781.

64
S. Gago et al. / Journal of Photochemistry and Photobiology A: Chemistry 244 (2012) 54–64
[19] F. Pina, M.J. Melo, M. Maestri, R. Ballardini, V. Balzani, Journal of the American
Chemical Society 119 (1997) 5556–5561.
[26] C.G. Hatchard, C.A. Parker, Proceedings of the Royal Society of London Series A
235 (1956) 518–536.
[20] F. Galindo, J.C. Lima, S.V. Luis, A.J. Parola, F. Pina, Advanced Functional Materials
15 (2005) 541–545.
[27] M. Maestri, R. Ballardini, F. Pina, M.J. Melo, Journal of Chemical Education 74
(1997) 1314–1316.
[21] F. Pina, T.A. Hatton, Langmuir 24 (2008) 2356–2364.
[22] F. Pina, J.C. Lima, A.J. Parola, C.A.M. Afonso, Angewandte Chemie: International
Edition 43 (2004) 1525–1527.
[23] F. Pina, M.J. Melo, M. Maestri, P. Passaniti, V. Balzani, Journal of the American
Chemical Society 122 (2000) 4496–4498.
[24] According to our previous experience this is a convenient situation to a good
performance of the photochromic system.
[25] F.W. Küster, A. Thiel, Tabelle per le Analisi Chimiche e Chimico-Fisiche, 12th
ed., 1982, Hoepli, Milano, pp. 157–160. The universal buffer used was prepared
in the following way: 2.3 cm3 of 85% (w/w) phosphoric acid, 7.00 g of mono-
hydrated citric acid, and 3.54 g of boric acid are dissolved in water; 343 mL of
1 M NaOH is then added, and the solution diluted to 1 dm3 with water.
[28] A. Ravikrishnan, P. Sudhakara, P. Kannan, Polymer Degradation and Stability 93
(2008) 1564–1570.
[29] K. Takeuchi, M.R. Jirousek, M. Paal, G. Ruhter, T. Schotten, US2004009976-A1
(2004).
[30] A. Roque, J.C. Lima, A.J. Parola, F. Pina, Photochemical and Photobiological Sci-
ences 6 (2007) 381–385.
[31] F. Pina, V. Petrov, C.A.T. Laia, Dyes and Pigments 92 (2011) 877–889.
[32] In flavylium systems, the Cc species usually shows a smaller mole absorption
coefficient than Ct.
[33] Considering DG = −RTln K.
[34] R.A. McClelland, S. Gedge, Journal of the American Chemical Society 102 (1980)
5838–5848.