Effect of Methyl, Hydroxyl, and Chloro Substituents in Position 3 of 3′,4′,7-Trihydroxyflavylium: Stability, Kinetics, and Thermodynamics

Article abstract of DOI:10.1002/chem.201601564

The effect of methyl, hydroxyl, and chloride substituents in position 3 of the 3′,4′,7-trihydroxyflavylium core structure was studied. The stability, relative energy of each of chemical species (thermodynamics), and their rates of interconversion (kinetics) are very dependent on these substituents. By comparing the mole fraction distribution at equilibrium of the three multistate systems with the parent 3′,4′,7-trihydroxyflavylium, introduction of a methyl substituent in position 3 increases the mole fraction of hemiketal at the expense of the trans-chalcone and increases the hydration rate very significantly; a hydroxyl substituent in position 3 gives rise to a degradation process, as observed in anthocyanidins. In the case of 3-chloro-3′,4′,7-trihydroxyflavylium, a dramatic increase of the flavylium cation acidity was observed and a photochromic system can be operated upon irradiation of the respective trans-chalcone in 1 m HCl. According to the photochromic response of 3,3′,4′,7-tetrahydroxyflavylium and 3′,4′,7-trihydroxyflavylium, some requirements for a good photochromic performance are discussed.

Full text of DOI:10.1002/chem.201601564

DOI: 10.1002/chem.201601564
Full Paper
&
Substituent Effects
Effect of Methyl, Hydroxyl, and Chloro Substituents in Position 3
of 3’,4’,7-Trihydroxyflavylium: Stability, Kinetics, and
Thermodynamics
Alfonso Alejo-Armijo,[a] Sofꢀa Salido,[a] Joaquꢀn Altarejos,[a] A. Jorge Parola,*[b] Sandra Gago,[b]
Nuno Basꢀlio,[b] Luis Cabrita,[b] and Fernando Pina*[b]
Abstract: The effect of methyl, hydroxyl, and chloride sub-
stituents in position 3 of the 3’,4’,7-trihydroxyflavylium core
structure was studied. The stability, relative energy of each
of chemical species (thermodynamics), and their rates of in-
terconversion (kinetics) are very dependent on these sub-
stituents. By comparing the mole fraction distribution at
equilibrium of the three multistate systems with the parent
3’,4’,7-trihydroxyflavylium, introduction of a methyl substitu-
ent in position 3 increases the mole fraction of hemiketal at
the expense of the trans-chalcone and increases the hydra-
tion rate very significantly; a hydroxyl substituent in position
3 gives rise to a degradation process, as observed in antho-
cyanidins. In the case of 3-chloro-3’,4’,7-trihydroxyflavylium,
a dramatic increase of the flavylium cation acidity was ob-
served and a photochromic system can be operated upon ir-
radiation of the respective trans-chalcone in 1m HCl. Accord-
ing to the photochromic response of 3,3’,4’,7-tetrahydroxy-
flavylium and 3’,4’,7-trihydroxyflavylium, some requirements
for a good photochromic performance are discussed.
Introduction
proton transfer :
AHþ þ H2O Ð A þ H3Oþ ðKaÞ
ð1Þ
Flavylium derivatives constitute a family of compounds that in-
clude anthocyanins, anthocyanidins, deoxyanthocyanidins, and
numerous natural and synthetic compounds containing the fla-
vylium core.[1–4] Some details of anthocyanin chemistry could
be fully understood only after achievements obtained by the
study of synthetic flavylium analogues.[5,6] On the other hand,
flavylium derivatives can perform as photochromic sys-
tems,[1,7,8] both in solution and in more organized matrices,[9–13]
as well as models for the development of optical memo-
ries.[1,14]
hydration :
AHþ þ 2 H2O Ð B þ H3Oþ ðKhÞ
ð2Þ
ð3Þ
ð4Þ
tautomerization :
B Ð Cc ðKtÞ
isomerization :
Cc Ð Ct ðKiÞ
One interesting characteristic of flavylium derivatives is the
existence of a common sequence of chemical reactions, as
shown in Scheme 1 for malvidin-3-O-glucoside (oenin).[1,15–18]
The flavylium cation is only a stable species at very acidic
pH values. Increasing the pH leads to the formation of several
species (Scheme 1) through a series of chemical equilibria ac-
counted for in Equations (1)–(4):
The system can be simplified by considering a single acid–
base equilibrium involving species AH+ (acid) in equilibrium
with a conjugate base (CB), with [CB]=[A]+[B]+[Cc]+[Ct]
[Eqs. (5) and (6)].[1]
acid ꢀ base global equilibrium :
ð5Þ
AHþ þ H2O Ð CB þ H3Oþ ðK0 Þ
a
[a] A. Alejo-Armijo, Dr. S. Salido, Prof. J. Altarejos
Departamento de Quꢀmica Inorgꢁnica y Orgꢁnica
Facultad de Ciencias Experimentales, Universidad de Jaꢂn
23071 Jaꢂn (Spain)
K0 ¼ Ka þ Kh þ KhKt þ KhKtKi
ð6Þ
a
In some cases (e.g., a high cis–trans isomerization kinetic
barrier), it is possible to define a pseudo-equilibrium; a transi-
ent state in which all species except Ct (pseudo)equilibrate
[Eqs. (7) and (8), with [CB^]=[A]+[B]+[Cc]].
[b] Dr. A. J. Parola, Dr. S. Gago, Dr. N. Basꢀlio, Dr. L. Cabrita, Prof. F. Pina
LAQV, REQUIMTE, Departamento de Quꢀmica
Faculdade de CiÞncias e Tecnologia, Universidade NOVA de Lisboa
2829-516 Caparica (Portugal)
E-mail: ajp@fct.unl.pt
fp@fct.unl.pt
acid ꢀ base pseudo-equilibrium :
ð7Þ
Supporting information and ORCIDs from the authors for this article are
available on the WWW under http://dx.doi.org/10.1002/chem.201601564.
AHþ þ H2O Ð CB^þ H3Oþ ðK^aÞ
Chem. Eur. J. 2016, 22, 1 – 12
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Scheme 1. The chemical reaction network established by malvidin-3-O-glucoside (oenin) in aqueous solution illustrates the multistate system characteristics
of flavylium cations.
K^a ¼ Ka þ Kh þ KhKt
ð8Þ
by means of the hydration of AH+ to give the hemiketal (B).
The hemiketal then tautomerizes to form the cis-chalcone (Cc),
which finally isomerizes to yield the trans-chalcone (Ct).
A few years ago, we introduced energy-level diagrams as
a very convenient way to visualize these multistate systems
(Scheme 2). It is easily constructed provided that all equilibri-
um constants in Scheme 1 [Eqs. (1)–(4)] were previously deter-
mined. A detailed explanation of its construction is reported in
reference [19].
The most practical way to study the flavylium-based chemi-
cal reaction network is to carry out direct pH jumps, which are
defined as the addition of base to equilibrated solutions of
AH+ at sufficiently low pH values at which it is the sole spe-
cies. In addition, reverse pH jumps, which are defined as the
addition of acid to solutions equilibrated at higher pH values
to give back AH+, are also very useful. In both cases, the pro-
cess is usually followed by UV/Vis spectrophotometry; when
the reactions take place on a timescale of seconds up to milli-
seconds, a stopped flow is required.
Following the energy-level diagram in Scheme 2, it is easy to
confirm that at low pH values the flavylium cation (AH+) is
the stable species. By raising the pH, for example, to pH=pKa,
50% of each AH+ and A are formed because proton transfer is
by far the fastest reaction of the multistate system. The AH+
and A species are in fast equilibrium and thus behave as
a single species in subsequent kinetic steps. Brouillard and
Dubois, who discovered that in acidic medium the quinoidal
base, A, did not hydrate, made an important breakthrough in
the comprehension of flavylium multistate kinetics.[18] In other
words, the evolution towards equilibrium is exclusively made
After a direct pH jump, base A is formed upon proton trans-
fer to the solvent, in a process with a rate constant of the
order of microseconds; a rate not measurable by stopped flow.
The next step is the disappearance of AH+/A through the hy-
dration of AH+ followed by tautomerization, occurring on the
subsecond timescale. When there is a cis–trans isomerization
Scheme 2. Energy-level diagram of the oenin multistate system in aqueous solution.[19]
Chem. Eur. J. 2016, 22, 1 – 12 www.chemeurj.org
&
&
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
barrier, after this process all species of the multistate are in
equilibrium, except Ct. In other words, once B is formed it im-
mediately gives Cc and at this point species AH+, A, B, and Cc
are in a metastable state that is called a pseudo-equilibrium
[Eqs. (7) and (8)]. Finally, from the pseudo-equilibrium, the
equilibrium is achieved upon the cis–trans isomerization of Cc
to yield Ct. This process may occur on a timescale from subsec-
onds to days and will control the global rate of the equilibra-
tion process. The kinetic expressions to account for the three
kinetic steps are given by Equations (9)–(11):[1,19]
k1 ¼ ka þ k ½Hþꢁ
ð9Þ
ꢀa
Scheme 3. From left to right: 3,3’,4’,7-tetrahydroxyflavylium (F1), 3’,4’,7-trihy-
droxy-3-methylflavylium (F2), 3-chloro-3’,4’,7-trihydroxyflavylium (F3),
3’,4’,5,7-tetrahydroxyflavylium (luteolinidin), and 3,3’,4’,5,7-pentahydroxyfla-
vylium (cyanidin).
½Hþꢁ
1
k2 ¼
k3 ¼
kh þ
k
½Hþꢁ
ð10Þ
ꢀh
½Hþꢁ þ Ka
1 þ Kt
KhKt
ki þ k
ð11Þ
ꢀi
al bridged structures present in the A-type oligocatechins from
appropriate flavylium salts, a study of the properties of the
starting flavylium cations is required.
½Hþꢁ þ Ka þ Kh þ KhKt
Tautomerization is generally faster than hydration, except for
very acidic pH values, as observed in reverse pH jumps or even
in direct pH jumps. In the former case, two kinetic processes
are observed given by Equations (12) and (13):[20]
Results and Discussion
½Hþꢁ
3,3’,4’,7-Tetrahydroxyflavylium (F1)
ð12Þ
ð13Þ
k4 ¼
kh þ k ½Hþꢁ
ꢀh
½Hþꢁ þ Ka
The stability of F1 in moderately acidic solutions was checked
according to the following sequence of events: 1) a direct pH
jump from a solution of the compound equilibrated at pH 1.15
to pH 4.40; 2) this solution was kept for 20 min (the experi-
ment was repeated for longer times); and 3) a reverse pH
jump to pH 1.17 (Figure 1a). The respective kinetics of flavyli-
um cation recovery was then monitored (Figure 1b). The kinet-
ic process reported in Figure 1b recovers only about 2% of
the initial flavylium. At the end of the kinetics reported in Fig-
ure 1b, 10% of the flavylium cation was not recovered, which
indicated that a decomposition process took place at pH 4.40.
This behavior, to a lesser extent, shows similarities with the
degradation of cyanidin (Scheme 3) recently reported (Fig-
ure 1c).[23] A direct pH jump from pH 0.8 to 4.6, waiting 20 min
at this pH, followed by a reverse pH jump back to pH 0.9, gave
rise to more than 95% degradation. In other words, less than
5% of flavylium cations were recovered after this process.
The absorption spectrum 3 h after a direct pH jump from
pH 1 to pH 3.2 is shown in Figure 2a. The corresponding spec-
trum after one month is given in Figure 2b. Acidification of
this last solution leads to spectrum Figure 2c, which is coinci-
dent with that of Figure 2b. Clearly, no flavylium is recovered
and the final product after one month at pH 3.2 is not depen-
dent on pH (in this range). Conversely, luteolinidin possessing
a hydroxyl substituent in position 5 (but not in position 3;
Scheme 3) is relatively stable.[24] This result corroborates that
the degradation process in anthocyanidins is due to the exis-
tence of the hydroxyl substituent in position 3, with a suggest-
ed mechanism involving the formation of a 3,4-diketone that
breaks down in two carboxylic acids.[23] By analogy, compound
F1 would expectedly degrade into 2,4- and 3,4-dihydroxyben-
zoic acid.
k5 ¼ k
ꢀt
In the latter case (less common), the equilibrium AH+/A/B is
established before ring opening/closure and Equation (14)
holds:
Kh
k6 ¼
kt þ kꢀt þ ktH½Hþꢁ
ð14Þ
½Hþꢁ þ Ka þ Kh
in which ktH accounts for the acidic catalysis of ring opening/
closure.[5,6]
When the cis–trans isomerization barrier is small, the system
is equivalent to three species in equilibrium, XQYQZ, and
Equations (15) and (16) can be deduced:[21]
kh
ki
Ka
Kt
þ
ƒ!
ƒƒƒ! ƒ! ƒ!
A
AH
B
Cc
Ct
ð15Þ
ð16Þ
ƒ
ƒƒƒ ƒ ƒ
k
½Hþꢁ
Hþ
k
ꢀh
ꢀi
½Hþꢁ
KhKtki þ k ½Hþꢁ
½HþꢁþKa
ꢀi
kobs
¼
kiKt
k
þ ½Hþꢁ
ꢀh
Herein, three new flavylium derivatives, 3,3’,4’,7-tetrahy-
droxyflavylium (F1), 3’,4’,7-trihydroxy-3-methylflavylium (F2),
and 3-chloro-3’,4’,7-trihydroxyflavylium (F3), were synthesized
and their stability, thermodynamics, and kinetics studied and
compared with similar flavylium derivatives reported in the lit-
erature (Scheme 3).
The synthesis of 3-substituted flavylium salts is of great in-
terest because their reaction with p-nucleophilic units (i.e., cat-
echin) could open up a route to prepare A-type proanthocya-
nidins or analogues.[22] To achieve the synthesis of such unusu-
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
ents in any position of the flavylium core redshifts the maxi-
mum of the flavylium cation; a phenomenon already described
for anthocyanins.[28] The red color is obtained only for F1. The
solubility of this compound is low in water or mixtures of
water and organic solvents, which prevents the use of NMR
spectroscopy and the study of other deprotonation equilibria
taking place at neutral and basic pH values.
Table 1. Absorption maximum of hydroxyl-substituted flavylium cations.
Cation
l [nm]
[8][a]
–
394
435
435
458
469
531
7-OH[25]
4’-OH[25]
4’,7-di-OH[26]
3’,4’,7-tri-OH[27]
3,3’,4’,7-tetra-OH[b]
[a] Flavylium with no substituents. [b] This work.
3’,4’,7-Trihydroxy-3-methylflavylium (F2)
The sequence of spectral variations accompanying direct pH
jumps for F2 (lmax =478 nm) are reported in Figure 3.
The quinoidal base (lmax =525 nm) is formed during the
mixing time of the base added to equilibrated solutions of fla-
vylium cation, at pH 1.0, as shown in Figure 3a; the spectra
were monitored by stopped flow 10 ms after mixing. The in-
flection point of the trace in the inset corresponds to the pKa
of the flavylium cation–quinoidal base acid–base equilibrium,
pKa =4.3. It is expected that the quinoidal base forms on posi-
tion 7 (as shown in Scheme 1 for oenin), which is known to be
more acidic than the 4’-position.[1]
Figure 1. a) Spectral variations of a 1.3ꢂ10ꢀ4 m solution of F1 in water/meth-
anol (1:1) upon a direct pH jump from pH 1.15 to pH 4.40 (standing 20 min)
followed by a reverse pH jump back to pH 1.17. b) Kinetics of flavylium re-
covery. At the end of this process, only 90% of flavylium cations were recov-
ered. c) Cyanidin spectral variations upon a direct pH jump from pH 0.8 to
pH 4.6 (standing 20 min) followed by a reverse pH jump back to pH 0.9.
The pH-dependent spectral variations at pseudo-equilibrium
are shown in Figure 3b and again the system behaves as
a single acid–base equilibrium involving AH+ and CB^ [Eq. (7)].
Finally, equilibrium is reached in a few hours, corresponding to
a single acid–base equilibrium involving AH+ and CB [eq. (5)].
Taking into account [see Eqs. (6) and (8)] that K’aꢀK^ =
a
KhKtKi =1.5ꢂ10ꢀ3 m, the fraction of Ct in CB can be calculated
as KhKtKi/K’a =0.6. Moreover, the ratio Ka/K’a predicts a fraction
of 0.05 of the base at equilibrium, which is in good agreement
with spectral variations and the data of the direct pH jumps
(Figure 3c). Thus, a fraction of 0.35 should be assigned to the
mole fractions of B and Cc.
From the initial state of the system represented in Figure 3a
containing only AH+ and A, the multistate system evolves to-
wards equilibrium through a complex kinetic process. The pH
jumps monitored over time allow three kinetic steps to be dis-
tinguished (Figure 4).
Figure 2. Absorption spectra of a 1.3ꢂ10ꢀ4 m solution of F1 in water/metha-
nol (1:1) after a direct pH jump to 3.2: a) after 3 h; b) after 1 month; c) the
same as that of b) at pH 1.
The faster processes were followed by stopped flow (Fig-
ure 4a), data for which could be fitted with biexponential laws.
For instance, a pH jump from 1.0 to 3.6 is fitted with kobs1
=
0.42 sꢀ1, which is compatible with the hydration process [at
this pH value, hydration is faster than tautomerization, due to
its direct dependence on proton concentration; Eq. (10)] and
The number and position of hydroxyl substituents in the fla-
vylium core accounts for the different colors of the com-
pounds. Table 1 shows that introduction of hydroxyl substitu-
k
obs2 =0.029 sꢀ1; this can be assigned to the tautomerization
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 4. a) Trace of the absorbance at l=473 nm of F2 in water/ethanol
(1:1) upon a direct pH jump from 1.0 to 3.6, 5.6ꢂ10ꢀ5 m, followed by
stopped flow; the trace can be fitted with a biexponential law with rate con-
stants of 0.42 sꢀ1 and 0.029 sꢀ1, which can be assigned to hydration and tau-
tomerization, respectively. b) The same measurements as those in a) for
pH 1.0 to 3.7, 6.5ꢂ10ꢀ5 m, obtained on a common spectrophotometer; fit-
ting was achieved with a monoexponential law with a rate constant of
4.8ꢂ10ꢀ4 sꢀ1, which was assigned to isomerization.
achieved with the rate and equilibrium constants reported in
Table 2 with excellent agreement.
Direct pH jumps are conveniently complemented by reverse
pH jumps (Figure 6). The trace for flavylium formation shows
two kinetic processes. The species constituting CB at pH 6 are
in equilibrium and after the pH jumps back to 1.7, for example,
A is immediately transformed into AH+. This step occurs
during the mixing time of the stopped flow. The first kinetic
step observed in Figure 6a corresponds to the formation of
AH+ from B. The observed rate constant is in very good agree-
ment with the value expected from Equation (12). The last step
Figure 3. The pH-dependent spectral variations of F2 in ethanol/water (1:1)
upon pH jumps from a stock solution at pH 1.2 to higher pH values:
a) 10 ms after a direct pH jump, followed by stopped flow (pKa =4.3,
5.6ꢂ10ꢀ5 m); b) at pseudo-equilibrium (pK^a =3.2, 5.3ꢂ10ꢀ5 m); c) at equilibri-
um (pK’a =2.7, 6.5ꢂ10ꢀ5 m).
equilibrium. On a slower timescale, followed by using a spectro-
photometer, a third step was observed and attributed to cis–
trans isomerization, with kobs3 =4.8ꢂ10ꢀ4 sꢀ1
.
can be assigned to ring closure [k ; Eq. (13)] because there is
ꢀt
A series of direct pH jumps to several pH values permits the
hydration (kobs1) and tautomerization (kobs2) processes to be
fitted with Equations (10) and (14), respectively (Figure 5a).
The last step to reach equilibrium is controlled by isomeriza-
tion and can be accounted for by Equation (11). Fitting was
no equilibrium between B and Cc. In other words, once B is
formed, it disappears to give AH+. The amplitude of the expo-
nentials permits Kt =0.68 to be calculated. On the other hand,
the tautomerization reaction is catalyzed by acids and
bases.[5,6] Although the equilibrium constant is the same, the
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
*
Figure 5. a) The pH dependence of hydration [ ; Eq. (10)] and tautomeriza-
tion [ ; Eq. (14)]; fitting was achieved for kh =0.3 sꢀ1, kꢀh =600mꢀ1 sꢀ1
,
*
pKa =4.3, kHt =40mꢀ1
ꢀt. b) Trace of the AH+ (l=473 nm) and CB (l=345 nm) absorption
changes that correspond to the slowest kinetic step of a direct pH jump to
3.9; fitting of the inset was achieved with Equation (11) for K^a =10ꢀ3.2
Kt =0.7, Kh =5.0ꢂ10ꢀ4 mꢀ1 ꢀ1, ki =7.0ꢂ10ꢀ4 sꢀ1, kꢀi =2.0ꢂ10ꢀ4 sꢀ1; these re-
s
ꢀ1; there is not enough accuracy to determine kt and
Figure 6. a) Reverse pH jump from pH 6.0 to 1.7; trace for flavylium cation
formation followed by stopped flow at l=473 nm (2.8ꢂ10ꢀ5 m). The ratio
between the amplitudes of Cc and B permits the equilibrium constant to be
k
calculated: Kt =0.68. Fitting was achieved for k1 =20 sꢀ1 and k2 =1.26 sꢀ1
.
,
b) Reverse pH jump from pH 5.4 to 1.2 (5.3ꢂ10ꢀ5 m); line corresponds to the
final spectrum at pH 1.2. The system evolves to AH+ according to a monoex-
ponential law with a rate constant equal to 2.7ꢂ10ꢀ4 sꢀ1. The ratio of the ex-
ponential amplitudes and final absorbance leads to a mole fraction of
Ct=0.6 at pH 5, which is in agreement with the fraction obtained from the
data shown in Figure 3c.
s
sults are in good agreement with data from other experiments, see Table 2.
Table 2. Equilibrium constants in water/methanol (1:1).[a]
pK’a pK^
pKa Kh [m]
Kt
Ki
a
3,3’,4’,7-tetra-OH (F1)
3’,4’,7-tri-OH[27]
3-Me-3’,4’,7-tri-OH (F2) 2.7
2.5
3.1
2.88 2.9
–
–
–
4.1 7.6ꢂ10ꢀ7 0.29 3.5ꢂ103
Table 3. Rate constants in water/methanol (1:1).[a]
4.3
3.2
4.3 5.0ꢂ10ꢀ4 0.7
3.5
kh
[sꢀ1
k
kt
k
ki
[sꢀ1
k
ꢀh
ꢀt
ꢀi
[a] Estimated error 20%.
]
[mꢀ1
s
ꢀ1] [sꢀ1
]
[sꢀ1
]
]
[sꢀ1
]
3,3’,4’,7-tetra-OH (F1) 0.0027 100
–
–
0.8
–
0.7
–
3’,4’,7-tri-OH[27]
3-Me-3’,4’,7-tri-OH
(F2)[c]
0.013 1.7ꢂ104 0.23
2.0ꢂ10ꢀ4
0.3
600
0.007[b] 0.01[b] 7ꢂ10ꢀ4 2ꢂ10ꢀ4
rate constants increase with increasing pH, and thus, the
values reported in Table 3 refer to pH 1.7. From Equation (14),
it is possible to calculate ktH =40mꢀ1 ꢀ1; the constant accounts
s
for acidic catalysis.
[a] Estimated error 20%. [b] At pH 1.7 under acidic catalysis. [c] Water/eth-
anol (1:1).
A global fitting of the data allows all of the rate and equilib-
rium constants of the multiequilibria to be calculated (Tables 2
and 3).
fraction distribution of the CB species, leading to stabilization
of hemiketal B, see compounds 1 and 2. Differently, the addi-
tion of hydroxyl substituents to the flavylium structure de-
creases species B at the expense of Ct, see compounds 3 and
F2.
The mole fraction distribution for F2 is shown in Scheme 4,
in comparison with three other compounds: unsubstituted fla-
vylium (1), 3-methylflavylium (2), and 4’,7-dihydroxy-3-methyl-
flavylium (3). The introduction of a methyl group in position 3
of the flavylium core leads to a dramatic change to the mole
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 4. Mole fraction distribution of flavylium compound[1] (1; water); 3-methylflavylium[29] (2; water); 4’,7-dihydroxy-3-methylflavylium[29] (3; water); and
F2 (this work, ethanol/water (1:1)).
3-Chloro-3’,4’,7-trihydroxyflavylium (F3)
Equilibrated solutions of F3 at acidic pH values (Figure 7)
reveal a dramatic shift of the global acidity of AH+ towards ex-
tremely low pK’a values (pK’a ꢂꢀ0.65). The increase of AH+
acidity can be explained by the high electronegativity of the
chlorine substituent in position 3 (ꢀI inductive effect), render-
ing the system more reactive to hydration and deprotonation.
In the case of F1, in which hydroxyl is present in position 3, al-
though oxygen is similar in electronegativity to chlorine (ꢀI),
the hydroxyl group is also an electron p donor (+M mesomer-
ic effect). The overall effect reflects in the pK’a values being
similar to those of 3’,4’,7-trihydroxyflavylium and F2 (Table 2),
in which the 3-substituent is hydrogen and a methyl group, re-
spectively (slight +I inductive effect).
In the case of F3, it is very difficult to carry out rigorous ex-
periments because the flavylium cation should be stored at
Figure 7. Spectral variations of equilibrated solutions of 5.6ꢂ10ꢀ5 m F3 in
ethanol/water (1:1); pK’a ꢂꢀ0.65.
[HCl]>10m. This extreme acidity makes it very difficult to con-
trol direct pH jumps. Moreover, at higher pH values decompo-
sition is observed. We have measured the absorption spectra
under very acidic conditions and the system is compatible
with the equilibrium between flavylium cation, quinoidal base,
and Ct. Photochemical experiments were carried out in etha-
nol/water (1:1) at pH 3.9 and in ethanol/1m HCl (1:1; Figure 8).
In the first case, the proton concentration is not sufficient to
reach the thermodynamic level of the flavylium cation, and
thus, only the disappearance of Ct (at lꢂ380 nm) is detected
(Figure 8a). In the second case, the appearance of the flavyli-
um cation is observed (Figure 8b). Two interesting aspects of
this photochromic system should be highlighted: 1) it works at
extremely high proton concentrations; and 2) due to the high
proton concentration, the thermal back reaction is very fast
and as a consequence the photochromic system is reversible
(Figure 8b, inset). In spite of the harsh acidic conditions, the
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Figure 9. Flash photolysis of 5.6ꢂ10ꢀ5 m F3 in ethanol/1m HCl (1:1) at:
a) l=500 nm (AH+ absorption), k=2.8 sꢀ1; and b) l=370 nm (Ct absorp-
tion).
Figure 8. Irradiation of F3 at l=370 nm in: a) ethanol/water (1:1), pH 3.9,
5.6ꢂ10ꢀ5 m, 22 min total irradiation time; b) ethanol/1m HCl (1:1),
5.6ꢂ10ꢀ5 m, 10 min total irradiation time; inset: thermal recovery.
I0(l=370 nm)=2.8ꢂ10ꢀ7 Einsteinminꢀ1
.
system measured by the maximum amount of color (AH+ and
A) reached by the system is illustrated in Figure 10. In the case
of F2 (Figure 10a), Cc and B have similar energies to the flavyli-
um cation at pH 4.1, and thus, the appearance of the colored
AH+ and A species is limited. A better performance can be
achieved at lower pH values, but in that case the initial fraction
of flavylium cations increases, and thus, the color contrast is
compromised. In the second case, in which the 3-methyl
group is lacking, the situation is much more favorable from
the point of view of the color appearance (Figure 10b). At
pH 4.6, the flavylium cation is much lower in energy than B
and Cc, and thus, more color can grow upon irradiation. How-
ever, the quantum yield for flavylium formation in the case of
3’,4’,7-trihydroxyflavylium (F=0.02 at pH 4.6) is lower than the
value of F=0.05 observed for F2 at pH 4.1. Moreover, at this
pH, species B, Cc, and Ct are at roughly the same energy level,
and thus, the amount of flavylium formed is only one-third of
the total disappearance of Ct and the real quantum yield
should be around 0.15. This is explained by comparing the
rates of the primary photoproduct disappearance in both
cases. In F2, the ratio between the forward and backward reac-
tions is >85 and there is practically no back reaction at this
stage of the process, whereas in 3’,4’,7-trihydroxyflavylium that
ratio is only 1.4 and a significant fraction of Cc gives back Ct
immediately after the flash.
system loses less than 5% efficiency after 6 irradiation cycles
(see the Supporting Information).
The results of flash photolysis experiments are reported in
Figure 9. The trace corresponding to the formation of AH+ fits
to a monoexponential process with a rate constant of 2.8 sꢀ1
,
which indicates that the tautomerization of Cc gives B and the
subsequent reaction of B gives AH+. It is notable that the lack
of recovery of the Ct absorption after the flash shows that
isomerization (backward reaction) is much slower than the two
forward processes to give the flavylium cation. However, isom-
erization (thermal back reaction) is faster than the other flavyli-
um systems reported to date.[1] The color contrast of the pho-
tochromic system is relatively small, but it serves to illustrate
the possibility of having a photochromic system operating at
such high concentrations of proton.
On the design of photochromic systems
To rationalize the photochromic behavior of F2 and compare it
with the analogue lacking the methyl group in position 3, it is
very convenient to draw an energy-level diagram (Scheme 5).
Irradiation of Ct leads to Cc, which spontaneously equili-
brates with B and AH+/A. The efficiency of the photochromic
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Figure 10. a) Spectral variations of F2 (5.6ꢂ10ꢀ5 m) upon irradiation at
l=340 nm, ethanol/water (1:1), pH 4.1, F=0.05, 13 min total irradiation
time; I0(l=340 nm)=2.0ꢂ10ꢀ7 Einsteinminꢀ1. b) Spectral variations of
3’,4’,7-trihydroxyflavylium (3.5ꢂ10ꢀ5 m), pH 4.6, F=0.02, 11 min total irradia-
Scheme 5. Energy-level diagram for F2 (top) and 3’,4’,7-trihydroxyflavylium
(bottom); the light excitation arrows are not to scale.
tion time; I0(l=366 nm)=1.9ꢂ10ꢀ6 Einsteinminꢀ1
.
The best performance for a photochromic system based on
the flavylium reaction network is reached when B and Cc pos- Experimental Section
sess higher energy levels relative to that of Ct, along with a fa-
General
vorable ratio between the forward/backward reactions to pre-
All solvents and chemicals employed for synthesis and the prepara-
tion of samples were of reagent or spectrophotometric grade and
were used as received; Millipore-grade water was used. NMR spec-
tra were run on a Bruker AMX 400 instrument operating at 400.13
(1H) and 100.00 MHz (13C). Elemental analyses were performed on
a Thermofinnigan Flash EA 112 series instrument.
vent the recovery of Ct from Cc.
Conclusions
The study of the thermodynamics and kinetics of flavylium de-
rivatives by constructing energy-level diagrams is an essential
tool to rationalize the overall behavior of these compounds in
solution. The stability, pH-dependent mole fraction distribution
of species, and possible applications as photochromic systems,
molecular-level optical memories, and reagents for the synthe-
sis of higher-level flavonoids are strongly dependent on the
nature and position of the substituents. The introduction of
three different substituents, hydroxyl, methyl, and chlorine, in
the 3-position of 3’,4’,7-trihydroxyflavylium leads to dramatical-
ly different behavior of the thermal and photochemical reactiv-
ity in alcoholic aqueous solutions showing that for synthetic
flavylium compounds there is still a great field of progression
and systematization to optimize the synthesis according to
achievements required for practical applications of these com-
pounds.
Synthesis of 2,3’,4’-triacetoxyacetophenone
This compound was prepared according to a procedure reported
in the literature for a similar compound, 2,4’-diacetoxyacetophe-
none.[30] Potassium acetate (14.3 mmol, 1.4 g) was added to
a
stirred solution of 2-chloro-3’,4’-dihydroxyacetophenone
(10.7 mmol, 2.0 g) in acetic anhydride (20 mL). The reaction mixture
was heated at 1008C overnight. Then, crushed ice was added and
a white solid precipitated. The solid was filtered off and carefully
washed with water and dried, yielding 2,3’,4’-triacetoxyacetophe-
none (2.187 g, 7.4 mmol, 69%). 1H NMR (CDCl3): d=7.84 (brd,
3J(H5’,H6’)=8.3 Hz, 1H; H6’), 7.79 (brs, 1H; H2’), 7.36 (d,
3J(H5’,H6’)=8.3 Hz, 1H; H5’), 5.30 (s, 2H; -CO-CH2-OAc), 2.34 (s, 6H;
OAc), 2.24 ppm (s, 3H; OAc).
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
from neutralization). The final pH of the solutions was measured in
a Crison basic 20 + pH meter. UV/Vis absorption spectra were re-
corded on Varian Cary 100 Bio or Varian Cary 5000 spectrophotom-
eters. The stopped-flow experiments were conducted in an Ap-
plied Photophysics SX20 stopped-flow spectrometer provided with
a PDA.1/UV photodiode array detector.
Synthesis of F1
This compound was prepared by condensation of 2,4-dihydroxy-
benzaldehyde (4.0 mmol, 0.552 g) and 2,3’,4’-triacetoxyacetophe-
none (4.0 mmol, 1.179 g). The reagents were dissolved in acetic
acid (5 mL), and 98% H2SO4 (1.2 mL) was added. The reaction mix-
ture was stirred overnight. Then, diethyl ether was added and
a red solid precipitated. The solid was filtered off and carefully
washed with diethyl ether and dried. The solid was recrystallized
by dissolving it in EtOH/H2O containing a few drops of HClO4
Irradiation experiments
1
Irradiation experiments were carried out on a spectrofluorimeter
Spex Fluorolog 1681 instrument at wavelengths of 370 (I0 =2.8ꢂ
10ꢀ7 Einsteinminꢀ1) or 340 nm (I0 =2.0ꢂ10ꢀ7 Einsteinminꢀ1). In the
case of 3’,4’,7-trihydroxyflavylium, the photochemical experiments
were conducted with a Xe–Hg lamp with the irradiation wave-
length isolated by using a bandpass filter (Oriel, 366 nm, I0 =1.9ꢂ
10ꢀ6 Einsteinminꢀ1). Light intensity was measured by ferrioxalate
actinometry.[32] The flash photolysis experiments were performed
on a Varian Cary 5000 spectrophotometer with a Harrick FiberMate
attached to the CUV-ALL-UV four-way cuvette holder compartment
(Ocean Optics) on the external side of the sample holder to per-
form light excitation perpendicular to the analyzing beam with the
sample compartment shielded with black cardboard and black
tape. As a pulsed light source, a commercially available Achiever
630AF camera flash was used, placed in close contact with the
sample holder. Excitation was made with the white light of the
camera flash with a time resolution of about 0.05 s. Further details
were previously described.[33]
(70%), yielding F1 (0.592 g, 1.60 mmol, 40%). H NMR (DCl/CD3CN,
pDꢂ1.0): d=8.47 (s, 1H; H4), 8.32 (d, 3J(H5,H6)=8.9 Hz, 1H; H5),
3
8.20 (brs, 1H; H2’), 7.96 (brd, J(H5’,H6’)=8.7 Hz, 1H; H6’), 7.44 (brs,
1H; H8), 7.34 (brd, 3J(H5,H6)=8.9 Hz, 1H; H6), 7.05 ppm (d,
3J(H5’,H6’)=8.7 Hz, 1H; H5’); elemental analysis calcd (%) for
C15H11O9Cl: C 48.60, H 2.99; found: C 48.55, H 3.04.
Synthesis of F2
This compound was prepared by condensation of 2,4-dihydroxy-
benzaldehyde (2.0 mmol, 0.276 g) and 3’,4’-dihydroxypropiophe-
none (2.0 mmol, 0.332 g). The reagents were dissolved in acetic
acid (2.5 mL), and 98% H2SO4 (0.6 mL) was added. The reaction
mixture was stirred overnight. Then, diethyl ether was added and
a red solid precipitated. The solid was filtered off and carefully
washed with diethyl ether and dried, yielding F2 (0.517 g,
0.77 mmol, 38%). 1H NMR (DCl/CD3CN, pDꢂ1.0) d=8.94 (s, 1H;
3
4
H4), 8.07 (d, J(H5,H6)=9.0 Hz, 1H; H5), 7.75 (d, J(H2’,H6’)=1.7 Hz,
3
4
1H; H2’), 7.69 (dd, J(H5’,H6’)=8.7 Hz, J(H2’,H6’)=1.7 Hz, 1H; H6’),
7.60 (d, 4J(H6,H8)=1.8 Hz, 1H; H8), 7.51 (dd, 3J(H5,H6)=8.9 Hz,
4J(H6,H8)=1.8 Hz, 1H; H6), 7.04 (d, 3J(H5’,H6’)=8.7 Hz, 1H; H5’),
2.72 ppm (s, 3H; 3-CH3); 13C NMR (DCl/CD3CN, pDꢂ1.0): d=173.0
(C2), 170.0 (C7), 160.2 (C9), 157.3 (C4), 154.5 (C4’), 147.6 (C3’), 133.2
(C5), 126.7 (C6’), 126.1 (C3), 123.0 (C6), 122.9 (C1’), 120.2 (C10), 118.3
(C2’), 117.2 (C5’), 102.8 (C8), 20.3 ppm (CH3); elemental analysis calcd
(%) for C32H30O14S: C 57.31, H 4.51, S 4.78; found: C 56.99, H 4.64, S
4.68.
Acknowledgements
Financial support from the Fundażo para a CiÞncia e Tecnolo-
gia (Portugal) through projects PTDC/QEQ-QFI/1971/2014, UID/
QUI/50006/2013 and RECI/BBB-BQB/0230/2012 (co-financed by
the ERDF under the PT2020 Partnership Agreement POCI-01-
0145-FEDER-007265). A.A.A. thanks the University of Jaꢃn for
granting him with a predoc fellowship. N.B. and S.G. acknowl-
edge postdoc grants SFRH/BPD/84805/2012, and SFRH/BPD/
111168/2015, respectively.
Synthesis of F3
This compound was prepared by condensation of 2,4-dihydroxy-
benzaldehyde (2.0 mmol, 0.276 g) and 2-chloro-3’,4’-dihydroxyace-
tophenone (2.0 mmol, 0.373 g). The reagents were dissolved in
EtOH (20 mL). The solution was saturated with dry hydrogen chlo-
ride for 1 h. The reaction mixture was stirred overnight. Then, the
solvent was removed. The red solid that appeared was carefully
washed with diethyl ether and dried, yielding F3 (0.572 g,
Keywords: dyes/pigments
substituent effects · thermodynamics
· kinetics · photochromism ·
[1] F. Pina, M. J. Melo, C. A. T. Laia, A. J. Parola, J. C. Lima, Chem. Soc. Rev.
2012, 41, 869–908.
[2] J. G. Sweeny, G. A. Iacobucci, J. Agric. Food Chem. 1983, 31, 531–533.
[3] P. Coggon, G. A. Moss, H. N. Graham, G. W. Sanderson, J. Agric. Food
Chem. 1973, 21, 727–733.
1
1.5 mmol, 77%). H NMR (DCl/CD3CN, pDꢂ1.0): d=9.07 (s, 1H; H4),
3
8.03 (d, J(H5,H6)=8.9 Hz, 1H; H5), 7.94 (m, 2H; H2’, H6’), 7.53 (brs,
1H; H8), 7.50 (dd, 3J(H5,H6)=8.9 Hz, 4J(H6,H8)=2.2 Hz, 1H; H6),
3
7.14 ppm (d, J(H5’,H6’)=8.5 Hz, 1H; H5’); 13C NMR (DCl/CD3CN, pD
[4] R. Brouillard, G. A. Iacobucci, J. G. Sweeny, J. Am. Chem. Soc. 1982, 104,
7585–7590.
ꢂ1.0): d=168.7 (C7), 167.5 (C2), 158.0 (C9), 153.7 (C4), 152.6 (C4’),
144.5 (C3’), 131.6 (C5), 126.2 (C6’), 122.2 (C6), 120.9 (C3), 119.7 (C1’),
118.3 (C10), 116.8 (C2’), 115.7 (C5’), 101.7 ppm (C8); elemental analysis
calcd (%) for C17H16O5Cl2: C 55.00, H 4.34; found: C 55.34, H 4.01.
[5] R. A. McClelland, S. Gedge, J. Am. Chem. Soc. 1980, 102, 5838–5848.
[6] R. A. McClelland, G. H. McGall, J. Org. Chem. 1982, 47, 3730–3736.
[7] M. J. Melo, A. J. Parola, J. C. Lima, F. Pina, in Chromic Materials Phenom-
ena and Their Technological Applications (Ed.: P. R. Somani), Applied Sci-
ence Innovations Private Limited, 2010, pp. 537–576.
[8] F. Pina, A. J. Parola, R. Gomes, M. Maestri, V. Balzani, in Molecular
Switches, Vol. 1 (Eds.: B. L. Feringa, W. R. Browne), Wiley-VCH, Weinheim,
2011, pp. 181–226.
[9] N. Jord¼o, R. Gavara, A. J. Parola, Macromolecules 2013, 46, 9055–9063.
[10] S. Gago, I. M. Fonseca, A. J. Parola, Microporous Mesoporous Mater.
2013, 180, 40–47.
[11] Y. Kawano, T. Takenouchi, Y. Kohno, R. Matsushima, M. Shibata, J. Jap.
Soc. Colour Mater. 2010, 83, 289–294.
Thermodynamic and kinetic studies
The pH jumps were carried out by adding a stock solution of flavy-
lium salt in 1:1 ROH/HCl 0.2m (1 mL; R=Me or Et) to a 3 mL
quartz cuvette containing a solution of 1:1 ROH/NaOH 0.2m
(1 mL), ROH (0.5 mL), and universal buffer of Theorell and Stenha-
gen (0.5 mL)[31] at the desired final pH. This defined the ionic
strength as 0.1m (controlled by the NaCl concentration resulting
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
10
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[12] Y. Kohno, M. Ito, M. Kurata, S. Ikoma, M. Shibata, J. Photochem. Photo-
biol. A 2011, 218, 87–92.
[13] T. Takenouchi, Y. Kawano, Y. Takeuchi, Y. Kohno, K. Takagi, M. Shibata, J.
Jap. Soc. Colour Mater. 2011, 84, 308–312.
[14] F. Pina, M. Maestri, V. Balzani, in Handbook of Photochemistry and Photo-
biology, Vol. 3 (Ed.: H. S. Nalwa), American Scientific Publishers, 2003,
pp. 411–449.
[15] F. Pina, J. Chem. Soc. Faraday Trans. 1998, 94, 2109–2116.
[16] R. Brouillard, B. Delaporte, J. Am. Chem. Soc. 1977, 99, 8461–8468.
[17] R. Brouillard, B. Delaporte, J. E. Dubois, J. Am. Chem. Soc. 1978, 100,
6202–6205.
[26] P. Figueiredo, J. C. Lima, H. Santos, M. C. Wigand, R. Brouillard, A. L. Mac-
anita, F. Pina, J. Am. Chem. Soc. 1994, 116, 1249–1254.
[27] N. Basꢀlio, F. Pina, ChemPhysChem 2014, 15, 2295–2302.
[28] A. L. MaÅanita, F. Pina, J. C. Lima, BiotecnoI 1997, 56, 4–13.
[29] A. Roque, C. Lodeiro, F. Pina, M. Maestri, R. Ballardini, V. Balzani, Eur. J.
Org. Chem. 2002, 2669–2709.
[30] A. Robertson, R. Robinson, J. Chem. Soc. 1928, 1460–1472.
[31] F. W. Kꢄster, A. Thiel, E. Brꢄckner, Tabelle per Le Analisi Chimiche e Chimi-
co-fisiche, Hoepli, Milano, 1985, 157–160; the universal buffer used was
prepared in the following way: 85% (w/w) phosphoric acid (2.3 cm3),
monohydrated citric acid (7.00 g), and boric acid (3.54 g) were dissolved
in water; 1m NaOH (343 mL) was then added, and the solution was di-
luted to 1 dm3 with water.
[18] R. Brouillard, J. Dubois, J. Am. Chem. Soc. 1977, 99, 1359–1364.
[19] F. Pina, in Recent Advances Polyphenol Research, Vol. 4 (Eds.: A. Romani,
V. Lattanzio, S. Quideau), Wiley, Chichester, 2014, pp. 341–370.
[20] F. Pina, Dyes Pigments 2014, 102, 308–314.
[32] C. G. Hatchard, C. A. Parker, Proc. R. Soc. A Math. Phys. Eng. Sci. 1956,
235, 518–536.
[21] F. Pina, J. Agric. Food Chem. 2014, 62, 6885–6897.
[22] G. A. Kraus, Y. Yuan, A. Kempema, Molecules 2009, 14, 807–815.
[23] L. Cabrita, V. Petrov, F. Pina, RSC Adv. 2014, 4, 18939–18944.
[24] M. J. Melo, S. Moura, A. Roque, M. Maestri, F. Pina, J. Photochem. Photo-
biol. A 2000, 135, 33–39.
[33] M. Maestri, R. Ballardini, F. J. S. Pina, M. J. Melo, J. Chem. Educ. 1997, 74,
1314–1316.
Received: April 4, 2016
[25] F. Pina, M. J. Melo, A. J. Parola, M. Maestri, V. Balzani, Chem. Eur. J. 1998,
4, 2001–2007.
Published online on && &&, 0000
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
11
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
FULL PAPER
Photochromism in 1m HCl: Changing
the substituent in position 3 of the fla-
vylium core structure leads to dramatic
changes to the thermodynamics and ki-
netics of the multistate molecular net-
work (see figure). In the case of chlorine,
the photochromic system operates
under extremely acidic conditions.
&
Substituent Effects
A. Alejo-Armijo, S. Salido, J. Altarejos,
A. J. Parola,* S. Gago, N. Basꢀlio,
L. Cabrita, F. Pina*
&& – &&
Effect of Methyl, Hydroxyl, and Chloro
Substituents in Position 3 of 3’,4’,7-
Trihydroxyflavylium: Stability, Kinetics,
and Thermodynamics
&
&
Chem. Eur. J. 2016, 22, 1 – 12
www.chemeurj.org
12
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!