The detection of the precursors of the photorearranged products of 3-hydroxyflavones in selected solvents from UV-visible spectra: In situ

Article abstract of DOI:10.1039/c9pp00316a

Mechanistic studies relating to the photochemistry of 3-hydroxy-2-phenyl-4H-chromen-4-one (3HF) and 6-chloro-3-hydroxy-2-phenyl-4H-chromen-4-one (Cl-3HF) have been reinvestigated in selected solvents. The UV-visible spectra of the photoproduct(s) of 3HF and Cl-3HF have been computed in situ via subtracting the spectra of unreacted substrates, with acetonitrile (ACN) and methanol (MeOH) as solvents. These spectra turn out to be different from the spectra of the corresponding isolated photoproducts: 3-hydroxy-3-phenyl-indan-1,2-dione and 6-chloro-3-hydroxy-3-phenyl-indan-1,2-dione (referred to as dione). Analyses of the photoproduct(s) via GC-MS show the formation of a single detectable product, i.e., the corresponding dione. On the basis of some experimental observations, it is proposed that the primary photoproduct in situ is 2,3-epoxy-2-hydroxy-1-indanone (referred to as epoxide) instead of dione as reported in previous years. Earlier, epoxide has been proposed to be the intermediate in the mechanism for the formation of dione. This is the first report where the formation of epoxide has been directly detected in the selected solvents. On the other hand, both dione and epoxide (2?:?1) are shown to be formed with MeOH as solvent. The second important finding is that epoxide and dione interconvert in the dark, depending upon the environment. With ACN as solvent, pure dione in the dark is kinetically and partially converted to epoxide. With MeOH as solvent, epoxide is instantly and partially converted to dione until both are in equilibrium. However, a solution of dione in MeOH remains stable in the dark. The photoformation of epoxide is quantitative with ACN as solvent and it is sufficiently stable. It has been further observed that epoxide solutions of 3HF and Cl-3HF in ACN are quantitatively converted into 3-phenylisobenzofuran-1(3H)-one and 6-chloro-3-phenylisobenzofuran-1(3H)-one, i.e., the corresponding phthalides, through the loss of CO when kept in the dark for some days. A mechanism has been proposed where epoxide has been shown to give dione and/or phthalide via selective C-O or C-C bond cleavage in the oxiranyl ring, respectively. The selection of this cleavage depends mainly on the solvent system and the substituents in the parent flavones.

Full text of DOI:10.1039/c9pp00316a

Photochemical &
Photobiological Sciences
View Article Online
View Journal | View Issue
PAPER
The detection of the precursors of the
photorearranged products of 3-hydroxyflavones in
selected solvents from UV-visible spectra in situ†
Cite this: Photochem. Photobiol. Sci.,
2019, 18, 2912
Jyoti Tomar, Kulvir Kaur and Manisha Bansal
*
Mechanistic studies relating to the photochemistry of 3-hydroxy-2-phenyl-4H-chromen-4-one (3HF)
and 6-chloro-3-hydroxy-2-phenyl-4H-chromen-4-one (Cl-3HF) have been reinvestigated in selected
solvents. The UV-visible spectra of the photoproduct(s) of 3HF and Cl-3HF have been computed in situ
via subtracting the spectra of unreacted substrates, with acetonitrile (ACN) and methanol (MeOH) as sol-
vents. These spectra turn out to be different from the spectra of the corresponding isolated photo-
products: 3-hydroxy-3-phenyl-indan-1,2-dione and 6-chloro-3-hydroxy-3-phenyl-indan-1,2-dione
(referred to as dione ). Analyses of the photoproduct(s) via GC-MS show the formation of a single detect-
able product, i.e., the corresponding dione. On the basis of some experimental observations, it is pro-
posed that the primary photoproduct in situ is 2,3-epoxy-2-hydroxy-1-indanone (referred to as epoxide)
instead of dione as reported in previous years. Earlier, epoxide has been proposed to be the intermediate
in the mechanism for the formation of dione. This is the first report where the formation of epoxide has
been directly detected in the selected solvents. On the other hand, both dione and epoxide (2 : 1) are
shown to be formed with MeOH as solvent. The second important finding is that epoxide and dione inter-
convert in the dark, depending upon the environment. With ACN as solvent, pure dione in the dark is kine-
tically and partially converted to epoxide. With MeOH as solvent, epoxide is instantly and partially con-
verted to dione until both are in equilibrium. However, a solution of dione in MeOH remains stable in the
dark. The photoformation of epoxide is quantitative with ACN as solvent and it is sufficiently stable. It has
been further observed that epoxide solutions of 3HF and Cl-3HF in ACN are quantitatively converted into
3-phenylisobenzofuran-1(3H)-one and 6-chloro-3-phenylisobenzofuran-1(3H)-one, i.e., the corres-
ponding phthalides, through the loss of CO when kept in the dark for some days. A mechanism has been
proposed where epoxide has been shown to give dione and/or phthalide via selective C–O or C–C bond
cleavage in the oxiranyl ring, respectively. The selection of this cleavage depends mainly on the solvent
system and the substituents in the parent flavones.
Received 31st July 2019,
Accepted 16th October 2019
DOI: 10.1039/c9pp00316a
rsc.li/pps
blue colours in flowers, fruits and leaves.4,5 The basic flavo-
noids have been classified in different families depending
1. Introduction
Flavonoids belong to a large group of abundant plant second- upon the presence of several phenolic hydroxyl functionalities
ary metabolites that can be found in vascular plants such as within the ring, including flavones and flavonols. Flavonols
ferns, conifers and flowering plants.1–3 Approximately, 4000 are characterized by the possession of a 3-hydroxypyran-4-one
varieties of flavonoids have been identified and many of these ring. 3-Hydroxy-2-phenyl-4H-chromen-4-one (3HF) is the sim-
are intense pigments, providing a spectrum of yellow, red and plest flavonol, and it has attracted strong interest due to its
peculiar photophysical and photochemical properties. Its
unique photophysical properties, involving double emission
peaks attributed to an excited-state intramolecular proton
transfer (ESIPT) reaction upon photo-excitation, which were
discovered by Sengupta and Kasha have been widely
studied.6–10 The wavelengths and intensities of the N* (the
excited state of the normal form) and T* (the excited state of
the tautomeric form) bands are sensitive to many physical pro-
perties of the environment. The photochemistry of 3HF has
Department of Chemistry, Punjabi University, Patiala, Punjab, 147002, India.
E-mail: jindal_manisha@yahoo.co.in; Tel: 9417812003
†Electronic supplementary information (ESI) available: 1H and 13C{1H} NMR, IR,
and GC-MS spectra of 3HF and 3HF phthalide; 1H and 13C{1H} NMR, HRMS, IR,
and GC-MS spectra of Cl-3HF and Cl-3HF phthalide; IR and GC-MS spectra of
3HF dione and Cl-3HF dione; IR spectrum of epoxide in the key spectral region
and GC-MS data from the photoproduct(s) of Cl-3HF photolysed (6 h) solution
in ACN. See DOI: 10.1039/c9pp00316a
2912 | Photochem. Photobiol. Sci., 2019, 18, 2912–2920
This journal is © The Royal Society of Chemistry and Owner Societies 2019

View Article Online
Photochemical & Photobiological Sciences
Paper
also been studied under aerobic and anaerobic conditions, (PerkinElmer, Spectrum Two). 1H and 13C{1H} NMR spectra
and in different solvents, over a long period of time.11–16 The were recorded with a Bruker Advance 400 MHz spectrometer
different routes for its photochemical reactions proposed in using TMS as an internal standard. HRMS data were obtained
the literature, i.e. photooxygenation and photorearrangement, using a XEVO G2-XS QTOF mass spectrometer with ion-source
and the pathways in different solvents have been well TOF MS ES+ apparatus (1.27e + 008).
summarized.17,18 Many reports on the photorearrangement
paths of 3HF, some of them contrasting, have been studied.
2.2. General procedure for the synthesis of the substrates
3HF and Cl-3HF
Matsuura and co-workers were the first to study this photoreac-
tion, where they identified 3-hydroxy-3-phenyl-indan-1,2-dione A previously reported two step method was used for the
(dione) as the photoproduct, formed through a 2,3-epoxy-2- synthesis of 3HF and Cl-3HF.19–21 The condensation of
hydroxy-1-indanone (epoxide) intermediate.11 There are some 0.001 moles of 2′-hydroxyacetophenone (0.12 mL) for 3HF or of
papers on the mechanism of formation with dione as the 5′-chloro-2′-hydroxyacetophenone (0.170 g) for Cl-3HF and
major product and 3-phenylisobenzofuran-1-(3H)-one (phtha- 0.001 moles (0.1 mL) of benzaldehayde was carried out in an
lide) as the minor photoproduct of 3HF, especially with aceto- ethanolic solution of sodium hydroxide at 5 1 °C in the first
nitrile (ACN) as solvent.16,17 Phthalide has been reported to be step. This reaction mixture was stirred for 6–7 h, and then the
formed as a secondary photoproduct of dione.
ice-cold mixture was neutralised with dilute HCl to obtain the
Here, we are reporting first time the detection of epoxide as chalcone. The corresponding chalcones were further oxidised
the precursor for the formation of dione during the photolysis from alkaline alcoholic solution, kept as an ice-cold mixture,
of 3HF and 6-chloro-3-hydroxy-2-phenyl-4H-chromen-4-one via the addition of 2 mL of 30% (w/v) H2O2 every 2 h under
(Cl-3HF), via computing the UV-visible spectra of the photopro- continuous stirring until the reaction was complete (about 8 h).
duct(s) in situ with ACN and methanol (MeOH) as solvents. Then, the final flavones were separated via neutralising the
Some simple experiments carried out after this detection are reaction mixture with dilute HCl and were crystallised from
likely to have significant implications for already reported EtOH : CHCl3 (1 : 1 v/v) to obtain yellow needles of both. All
mechanisms and for future studies of the photochemistry of spectroscopic data obtained here for 3HF is identical with the
various 3HF-type compounds.
reported data.20 This spectroscopic data has also been
obtained from Cl-3HF and is expectedly similar to that of 3HF.
3-Hydroxy-2-phenyl-4H-chromen-4-one (3HF). M.p: 171–172 °C;
FTIR (neat, cm−1) νmax: 3196 (OH), 1604 (CvO); GC-MS reten-
2. Experimental
2.1. General information
1
tion time (tr) = 18.25 min, m/z = 237 [M]+; H NMR (400 MHz,
DMSO-d6) δ: 9.65 (s, OH exchangeable with D2O), 8.23 (d, J =
All reagents used for the preparation of 3-hydroxy-2-phenyl-4H- 7.40 Hz, 2H), 8.13 (dd, J = 8.04, 1.12 Hz, 1H), 7.82–7.76 (m,
chromen-4-one (3HF) and 6-chloro-3-hydroxy-2-phenyl-4H- 2H), 7.60–7.56 (m, 2H), 7.53–7.46 (m, 2H); 13C{1H} NMR
chromen-4-one (Cl-3HF) were bought from Sigma-Aldrich, (100 MHz, DMSO-d6) δ: 172.9, 154.5, 145.1, 139.0, 133.7, 131.2,
USA, and were used as received without further purification. 129.8, 128.5, 127.6, 124.7, 124.4, 121.2, 118.4.
The solvents used for the photochemical reactions were of
6-Chloro-3-hydroxy-2-phenyl-4H-chromen-4-one (Cl-3HF).
HPLC grade and were purchased from S. D. Fine and Loba M.p.: 168–169 °C; FTIR (neat, cm−1) νmax: 3273 (OH), 3057
Chemie. Photoreactions were carried out either in a 4 mL (C–H Ar), 1602 (CvO); GC-MS tr = 20.40 min, m/z = 273 [M]+;
tightly covered quartz cuvette or in a 50 mL Pyrex glass tube. A 1H NMR (400 MHz, CDCl3) δ: 8.25–8.22 (m, 3H), 7.65 (dd, J =
medium pressure mercury lamp (125 W) with a built-in filter 8.96, 2.52 Hz, IH), 7.57–7.49 (m, 4H), 7.00 (s, OH exchangeable
mainly emitting at around 366 nm was powered through a with D2O); 13C{1H} NMR (100 MHz, DMSO-d6) δ: 171.9, 153.0,
stabilised power supply. Although the emission band through 145.6, 139.2, 133.5, 130.9, 130.0, 128.9, 128.5, 127.6, 123.5,
the built-in filter is mainly centred at 366 nm, another glass 122.4, 120.9; HRMS calcd for C15H10O3Cl ([M + H]+): 273.0318,
filter with a cut-off of 348 nm was incorporated in between the found: 273.0316.
photolysis cell and the lamp, to eliminate any low intensity
emission bands below 348 nm. A special holder was used to
2.3. Characterisation data of the separated photoproducts
place the photolysis cell (at a distance of 6 cm from the lamp) There were two major photoproducts, dione and phthalide,
in front of the glass filter, so that maximum light falls on its formed in the photolysis of both 3HF and Cl-3HF with ACN
surface. A shutter was used in between the filter and the cell to and dichloromethane as solvents. These photoproducts were
interrupt the light whenever this was desired. The photolysis separated via passing the photolysed samples through a silica
cell was shielded from the heat of the lamp with a wooden gel column (60–120 mesh) using 5% ethylacetate in hexane
screen that had a window for the filter. Most of the experi- (v/v) as the eluting solvent. Both diones were separated and
ments were carried out at 20 2 °C. The progress of the reac- obtained in the form of white amorphous solids, and both
tion was monitored with a double beam UV/vis spectrophoto- phthalides were in the form of grey needles. The 3HF dione
meter (UV-1800, Shimadzu) equipped with UV-probe software was identified via comparing its GC-MS, IR and UV-visible
and GC-MS apparatus (GCMS QP-2010 plus, Shimadzu). spectra with the reported data.11,12,15 The Cl-3HF dione was
IR spectra were recorded using an FTIR spectrometer identified via comparing the GC-MS, IR and UV-visible spectra
This journal is © The Royal Society of Chemistry and Owner Societies 2019
Photochem. Photobiol. Sci., 2019, 18, 2912–2920 | 2913

View Article Online
Paper
Photochemical & Photobiological Sciences
with those of the 3HF dione. The data from both are similar,
as is expected. The 3HF phthalide was identified via compar-
ing its GC-MS, UV-visible, IR, 1H and 13C{1H} NMR spectra
with the reported data.12,17 The data from the Cl-3HF phtha-
lide are also similar to those from the 3HF phthalide.
3-Hydroxy-3-phenyl-indan-1,2-dione (3HF dione). IR (neat,
cm−1) νmax: 3423 (OH), 3063 (C–H Ar), 1763, 1722 (CvO);
GC-MS tr = 16.58 min, m/z = 238 [M]+.
3-Phenylisobenzofuran-1(3H)-one (3HF phthalide). IR (neat,
cm−1) νmax: 2925, 2853 (C–H aliphatic), 1756 (CvO); GC-MS
tr = 14.37 min; m/z = 210 [M]+; 1H NMR (400 MHz, DMSO-d6) δ:
7.93 (d, J = 7.6 Hz, 1H), 7.77 (dt, J = 7.56, 1.04 Hz, 1H), 7.64 (t,
J = 7.44 Hz, 1H), 7.48–7.39 (m, 4H), 7.33–7.31 (m, 2H), 6.74 (s,
1H); 13C{1H} NMR (100 MHz, DMSO-d6) δ: 169.9, 149.9, 136.8,
134.7, 129.5, 129.0, 128.9, 126.7, 125.0, 124.5, 123.2, 81.8; twin
peaks in the UV-visible spectrum at around 271 and 278 nm.
6-Chloro-3-hydroxy-3-phenyl-indan-1,2-dione (Cl-3HF dione).
IR (neat, cm−1) νmax: 3414 (OH), 3066 (C–H Ar), 1769, 1741
(CvO); GC-MS tr = 18.26 min, m/z = 272 [M]+.
6-Chloro-3-phenylisobenzofuran-1(3H)-one (Cl-3HF phtha-
lide). IR (neat, cm−1) νmax: 2957, 2927 (C–H aliphatic), 1749
(CvO); GC-MS tr = 16.49 min, m/z = 244 [M]+; 1H NMR
(400 MHz, DMSO-d6) δ: 8.00 (d, J = 1.84 Hz, 1H), 7.82 (dd, J =
8.16, 1.92 Hz, 1H), 7.50 (d, J = 8.24 Hz, 1H), 7.43–7.40 (m, 3H),
7.35–7.32 (m, 2H), 6.75 (s, 1H). 13C{1H} NMR (125 MHz,
CDCl3) δ: 169.0, 147.8, 135.9, 134.6, 129.5, 129.1, 128.6, 128.3,
127.5, 126.9, 125.5, 124.1, 82.6; HRMS calcd for C14H10O2Cl
([M + H]+): 245.0369, found: 245.0349; twin peaks in the UV-
visible spectrum at around 284 and 292 nm.
Fig. 1 The electronic absorption spectra of 3HF in ACN photolysed
after 0, 10, 20, 30, 40, 50, and 60 min.
photolysed mixture spectra, absorbance beyond 300 nm is due
to reactant only. Pure reactant spectra after various photolysis
times were constructed, considering that absorbance at
355 nm of the photolysed mixture is due to reactant only. It
was noticed that the so-constructed spectra of the reactant
overlapped (from 330 to 390 nm) with the photolysed mixture
spectra at different times. The reactant also follows the Beer–
Lambert law at all wavelengths. Thereafter, the photoproduct
(s) spectra were obtained through subtracting the corres-
ponding constructed spectra of unreacted reactant from the
photolysed mixture spectra at different photolysis times, and
these are shown in Fig. 2. We tried this method in our earlier
paper also.22
The spectra of the photoproduct(s) shown in Fig. 2 show
one shoulder at around 220.2 nm, one peak with a λmax value
of 280.2 nm and one hump at around 325 nm, growing as the
photolysis time increases.
Similar experiments were carried out with Cl-3HF, and
Fig. 3 shows the photoproduct(s) spectra obtained from the
photolysed mixtures, as was done in the case of 3HF.
3. Results and Discussion
3.1. The detection of 2,3-epoxy-2-hydroxy-1-indanone (epoxide)
in situ during the photolysis of 3-hydroxy-2-phenyl-4H-chromen-
4-one (3HF) and 6-chloro-3-hydroxy-2-phenyl-4H-chromen-4-one
(Cl-3HF)
4 mL of 3HF solution (4 × 10−5 M) with acetonitrile (ACN) as
the solvent was taken in an air-tight quartz cuvette and this
was photolysed for different lengths of time.
Fig. 1 shows the electronic absorption spectra of the solu-
tions photolysed for various intervals of time. The spectral
changes in this figure show that the absorbance is continu-
ously increasing between 285 and 270 nm and decreasing
between 270 and 228 nm. There are three isobestic points, at
287.4, 264.5, and 224.8 nm. The decreasing absorbance at
higher wavelengths and the increasing absorbance at selected
wavelengths in the UV region clearly show that reactant is dis-
appearing and products, which absorb only at lower wave-
lengths in the UV region, are being formed.
3.1.1. UV-visible absorption spectra of photoproduct(s)
in situ. In a separate experiment, the photolysis product has
been analysed to be 3-phenyl-3-hydroxy-1,2-indandione
(dione). The UV-visible spectrum of dione shows that it does
not absorb beyond 300 nm, while 3HF has a peak at 339.4 nm
Fig. 2 The electronic absorption spectra of the photoproduct(s) after
subtracting the unreacted reactant spectra from the photolysed solution
and a shoulder at 354.2 nm. Therefore, it appears that in the spectra of 3HF in ACN after 10, 20, 30, 40, 50, and 60 min (cf. Fig. 1).
2914 | Photochem. Photobiol. Sci., 2019, 18, 2912–2920
This journal is © The Royal Society of Chemistry and Owner Societies 2019

View Article Online
Photochemical & Photobiological Sciences
Paper
intermediate from the (σ2 + π2) cycloaddition of the substrate
in its triplet state (Scheme 1).
In the same period, a similar rearrangement was reported
in the photolysis of 2,6-dimethyl-3-hydroxy-4H-pyrane-4-one.23
Here it was proposed that both 2 and 3 were formed, but the
former could not be isolated (Scheme 2).
Matsuura and co-workers did not verify whether dione (1) is
formed in one step, like that of 3, or in two steps. On the basis
of these reports it can be surmised that, in our case, it is most
likely that the primary product is epoxide (A).
3.2. The formation and conversion of the epoxide
intermediate
3.2.1. Detection of epoxide through IR spectroscopy. It has
been observed that epoxide formed in the photolysis of Cl-3HF
is relatively more stable than in the case of 3HF. It has a pink
colour in ACN that, upon evaporation, is immediately con-
verted to yellow at room temperature. However, when the
solvent (ACN) was distilled off at low temperature (10–12 °C)
under rotary evaporation, the pink colour could be retained for
Fig. 3 The electronic absorption spectra of the photoproduct(s) after
subtracting the unreacted reactant spectra from the photolysed solution
spectra of Cl-3HF in ACN after 10, 20, 30, 40, 50, and 60 min.
3.1.2. Analysis via GC-MS. 25 mL of 3HF solution in ACN some time. An attempt has been made to record its IR spec-
(8.40 × 10−4 M) in a Pyrex tube was photolysed for a total of trum immediately (Fig. 6). This shows a single absorption
8 h. GC-MS analyses of these photolysed solutions carried out band at 1732 cm−1 in the keto group region, instead of twin
after 3, 6 and 8 h of photolysis show that the amount of a bands at 1768 and 1741 cm−1 due to the expected dione
single product with a molecular mass of 238 g mol−1 grows as product. Also, the IR spectrum of epoxide is different from
the photolysis time increases. This product, which has the that of 6-chloro-3-phenylisobenzofuran-1(3H)-one (phthalide),
same mass but a different retention time (tr) to the reactant, where the bands observed due to C–H stretching at 2957, 2927
should be the photorearranged product. Representative GC-MS and 2872 cm−1 in phthalide are missing in the case of
analysis data from photolysed solutions of 3HF in ACN epoxide, as expected.
(Fig. 4a) and from a separated sample of pure dione (Fig. 4b)
are given in Fig. 4.
The photolysis of 3HF has also been carried with methanol
(MeOH) as solvent, in a similar fashion to that of with
It is clear from Fig. 4a that a single detectable product, ACN experiments. Fig. 7 shows the photoproduct(s) spectra
dione, is formed during photolysis. The retention time, mole- obtained in situ via subtracting the unreacted 3HF from the
cular mass and fragmentation pattern of this product are iden- photolysed solution.
tical to pure dione (Fig. 4b). The UV-visible spectrum of dione
There are two λmax values at around 245 and 290 nm, which
shows two peaks at 246.7 and 291.1 nm. On the other hand, are close to those obtained from the spectrum of pure dione in
the subtracted photoproduct spectra (Fig. 2) obtained in situ MeOH (Fig. 5). However, the finer details of these spectra have
show peaks not only at different positions than those of dione been compared with pure dione in MeOH and epoxide in ACN.
but with a minimum around λ245 nm, whereas dione has a It is observed that these subtracted spectra correspond to mix-
maximum. This indicates that, in situ, there is some other tures of epoxide and dione. The ratio of the concentration of
product from which dione is formed during GC-MS analysis.
To look into this, conversion spectral changes were ana- from the results obtained in ACN (Fig. 2). As explained above,
lysed via a set of four spectra in ACN (Fig. 5).
the primary product in ACN is epoxide; it appears that in
dione to epoxide is about 2. Clearly, this result is different
On comparing spectrum II with the spectrum of the separ- MeOH, the primarily formed epoxide is instantly partially con-
ated sample of dione (spectrum IV), it is apparent that upon verted to dione. In another experiment, the effects of adding
removing the solvent, the primary product is partly changed to MeOH to a solution of epoxide in ACN were studied. A solution
dione. Spectrum IV becomes more interesting because dione (30 mL) of 3HF with ACN as solvent (0.7 × 10−6 M) was photo-
formed partly in spectrum II appears to go back to the product lysed for 110 min. The UV spectrum of the photolysed solution
seen in spectrum I. Obviously, during recording the GC-MS shown in Fig. 8 shows epoxide only, i.e., the reactant appears
data, a stable dione product is formed from the primary to be quantitatively converted into the product (epoxide).
product. Now the question arises; which is the primary product? When MeOH was added to this photolysed solution, the initial
The photorearrangement of 3HF was studied first in the UV spectrum instantly changed and shows two shoulders at
1970s by Matsuura and co-workers in benzene : isopropyl about 246 and 295 nm, which correspond to the absorption
alcohol under N2.11 The only photoproduct obtained under bands of dione. This shows that upon adding MeOH, epoxide
these conditions was dione, and they proposed the formation is instantly partially changed to dione. This supports the
of this product via a 2,3-epoxy-2-hydroxy-1-indanone (epoxide) above-mentioned photolysis results with MeOH as solvent.
This journal is © The Royal Society of Chemistry and Owner Societies 2019
Photochem. Photobiol. Sci., 2019, 18, 2912–2920 | 2915

View Article Online
Paper
Photochemical & Photobiological Sciences
Fig. 4 GC-MS analysis of the photoproduct(s) of 3HF: (a) photolysed solution (6 h) of 3HF with ACN as solvent (top panel). Second and third panels
are for the first component and fourth and fifth panels are for second component of top panel; and (b) a separated sample of 3HF dione.
M. Kubinyi and co-workers in recent reports have carried another route involving T1–S0 conical intersection, which they
out theoretical studies on photoreactions of 3HF.18 They have have not studied.
suggested the formation of an intermediate with a 1-indanone
3.2.2. The interconversion of epoxide and dione. As
skeleton (B) instead of an epoxide intermediate in the photo- suggested by the results of Fig. 5, epoxide formed during the
rearrangement reaction of 3HF to form dione (Scheme 3). photolysis of 3HF is reversibly changed into dione. Therefore,
However, the stability of the intermediate in ACN as reported the stability of dione was studied with ACN as solvent. It has
in the present work suggests that it cannot be a biradical (B) been observed that the electronic absorption spectrum of
rather than epoxide, as suggested by M. Kubinyi and co- dione changes continuously, even if it is kept in the dark, as
workers. However, they admit that epoxide may be formed via shown in Fig. 9a. The changes in the spectra, i.e., the decreas-
2916 | Photochem. Photobiol. Sci., 2019, 18, 2912–2920
This journal is © The Royal Society of Chemistry and Owner Societies 2019

View Article Online
Photochemical & Photobiological Sciences
Paper
Fig. 5 The electronic absorption spectra of photolysed samples of 3HF
under various conditions and of dione: (I) a freshly photolysed solution;
(II) after evaporating ACN from solution (I) and then adding the same
amount of ACN; (III) when solution (II) was kept overnight; and (IV) the
spectrum of a separated sample of 3HF dione.
Fig. 7 The electronic absorption spectra of the photoproduct(s) after
subtracting the unreacted reactant from photolysed solutions of 3HF in
MeOH after 10, 20, 30, 40, and 50 min.
Scheme 1 The photorearrangement of 3HF as proposed by Matsuura
et al.
Fig. 8 The effect of adding MeOH into a solution of the primary photo-
product (epoxide) in ACN: (a) is the photolysed solution spectrum of
3HF in ACN and (b) is the spectrum of (a) after MeOH (0.5 mL to a total
of 4 mL of (a)) was added.
Scheme 2 The photoreaction of 2,6-dimethyl-3-hydroxy-4H-pyrane-
4-one in MeOH.
Scheme 3 The structure of intermediate B proposed by M. Kubinyi
et al. from theoretical studies.
ing absorbance at around 245 nm and the increasing absor-
bance at around 280 nm, suggest that dione is kinetically con-
verting to epoxide. Unchanged dione spectra over time could
be constructed (Fig. 9b), and upon subtracting these spectra
from the corresponding spectra in Fig. 9a, the converted
product spectra (Fig. 9c) are obtained. These spectra are identi-
cal to those of epoxide as observed for the product formed
Fig. 6 The IR spectrum of 4-chloro-2,3-epoxy-2-hydroxy-1-indanone
(epoxide).
This journal is © The Royal Society of Chemistry and Owner Societies 2019
Photochem. Photobiol. Sci., 2019, 18, 2912–2920 | 2917

View Article Online
Paper
Photochemical & Photobiological Sciences
Fig. 9 (a) The changing spectra of dione in the dark with ACN as solvent, (b) the spectra of unchanged dione after different time intervals, and
(c) the remaining spectra upon subtracting the spectra of (b) from the corresponding spectra of (a).
in situ. This conversion is also temperature dependent. Similar conditions, and in different solvent systems.11–16 A common
results are also obtained with Cl-3HF. However, there are no feature is that dione is the photorearranged product, along
changes in the dione spectra of 3HF with MeOH as solvent. with the formation of phthalide. Phthalide has been shown to
Detailed kinetics studies of these processes are being carried be formed from dione following long-term photolysis. Here,
out and they will be reported later on.
after the detection of epoxide as the primary product, it has
Similar processes have been reported in the rearrangement been observed that phthalide is formed directly from epoxide
of an oxiranyl carbinyl radical (C) into an allyloxy radical (D) instead of through dione. As explained above, dione in ACN is
via C–O bond cleavage and/or in the formation of E via C–C converted to epoxide kinetically. It has been found that if
bond cleavage, depending upon the natures of R3 and R4 epoxide solutions of both 3HF and Cl-3HF in ACN are kept in
(Scheme 4).24
the dark for some days, the corresponding epoxides are quanti-
Evidence for reversibility between C and D has been tatively converted to the corresponding phthalides. Phthalides
reported in different cases. The mechanistic and kinetic have been identified via their 1H NMR, 13C{1H} NMR, IR,
aspects of this mechanism have been widely studied.25–29
GC-MS and UV-visible spectra. It has also been verified that a
S. Tommasini and co-workers have studied the photoreac- solution of dione in MeOH remains stable when placed in the
tions of 3HF with ACN as solvent.14 They analysed the photo- dark, unlike with ACN as solvent (see the above discussion). It
lysed solutions after selected irradiation times via reversed- is neither converted to epoxide nor decomposed to phthalide,
phase liquid chromatography. Multistep gradient elution was as revealed by the UV-visible spectrum and GC-MS data. On
carried out using MeOH/H2O (with 10% glacial acetic acid) as the basis of the above-mentioned observations, results, and
solvent. The amounts of a number of products are shown in detection of epoxide as the primary product, the following
the chromatogram to grow with increasing irradiation time. basic steps shown in Scheme 5 are proposed.
The absorption spectra of four products appearing at different
The formation of epoxide has been earlier suggested to
retention times have been recorded but no interpretation of occur via (σ2 + π2) cycloaddition of the substrate in its triplet
these spectra has been made. With the help of our findings state.11 In the present work, we have inferred that the for-
here, these results can be easily analysed. First, two of these mation, stability and further conversion of epoxide into dione,
spectra should be due to dione and two should be due to phthalide and other minor products depend mainly upon the
epoxide. Under the solvent system conditions used, the solvent system and the substrate structure. The role of ACN in
epoxide formed in the primary reaction is partly converted to the formation of epoxide and the interconversion of dione in
dione, along with the formation of many side products.
this solvent appears to be significant.
3.2.3. The formation of 3-phenylisobenzofuran-1(3H)-one
In a report by R. Diller and co-workers, they obtained mid-
(3HF phthalide) and 6-chloro-3-phenylisobenzofuran-1-(3H)- IR vibrational spectra of 3HF excited states in solution using
one (Cl-3HF phthalide). There are a number of reports on the acetonitrile-d3.30 On the basis of vibrational frequencies, struc-
photochemical reactions of 3HF under aerobic and anaerobic tural information about the intermediate states was obtained.
Scheme 4 The reversibility and modes of the ring opening of the oxiranyl carbinyl radical (C).
2918 | Photochem. Photobiol. Sci., 2019, 18, 2912–2920
This journal is © The Royal Society of Chemistry and Owner Societies 2019

View Article Online
Photochemical & Photobiological Sciences
Paper
epoxide is kept. The conversion of epoxide into dione with
MeOH as solvent is favoured because of the hydroxyl group in
MeOH, which helps in the transfer of H from OH to the O
atom of the oxiranyl ring of epoxide during C–O bond
cleavage.
The detection of epoxide as the primary product, especially
in ACN, here will definitely help in explaining the mechanisms
of the photorearrangement reactions of 3HF-type molecules
that have already been studied or are being studied.
4. Conclusions
Scheme 5 The proposed steps for the formation of dione and phtha-
lide in ACN from the epoxide formed by the photolysis of 3HF.
Epoxide as a precursor for the formation of dione has been
detected in the photorearrangements of 3HF and Cl-3HF from
the photoproduct spectra in situ. It has been found that
epoxide and dione are interconvertible, depending upon the
environment. Phthalide has been shown to be formed from
epoxide in the solution phase in the dark instead of from the
secondary photolysis of dione, in contrast to some reports
from previous years. Thus, it can be said that epoxide is the
photoproduct in the photorearrangement reactions of 3HF
molecules, at least in the selected solvents. To date, in all
reports from previous years, dione and phthalide have been
reported as the photoproducts in these reactions; however,
they actually appear to be formed from the epoxide photo-
product in the dark, especially with ACN as solvent.
It was predicted that 3HF has two structures in the ground
state: one possessing intact intramolecular hydrogen bonds,
and another in the form of a 3HF-ACN aggregate with inserted
ACN. Aggregation weakens the intramolecular hydrogen
bonding in 3HF. It has also been reported that a small pro-
portion of 3HF exists in the anionic form.31 The anionic form
proportion is relatively higher with MeOH as solvent, com-
pared to ACN. The formation, stability, etc. of epoxide in ACN
could be due to these properties of 3HF, i.e. special kinds of
solute-solvent interactions.
The quality of 3HF photoreactions also depends upon the
substitution at the phenyl ring in the second position. In a
report on the photoreactions of 4′-diethylamino-3-hydroxy-
flavone, it was found that it was oxidised into O-4-diethyl-
aminobenzoyl salicylic acid differently from 3HF, which under-
goes photorearrangement with ACN as solvent.16
Conflicts of interest
The authors have no competing interests to declare.
We have also reported the photoreactions of 2-(furan-2-yl)-
3-hydroxy-4H-chromen-4-one (FHC) and 3-hydroxy-2-(thio-
phene2-yl)-4H-chromen-4-one (THC) with ACN as solvent.22 In
those cases, it was observed that the photochemical reactions
of two similar looking flavones, FHC and THC, are different in
the same solvent (ACN), while the same flavone behaves differ-
ently in different solvents. FHC in aerated ACN gives a dimer
(an oxidised product) and its dione upon photolysis, while in
de-oxygenated ACN, its dione and phthalide are formed. In the
case of THC, only dimer is formed in aerated ACN. It has been
explained that this kind of flavone behaviour is because of
complexity added by excited state intramolecular charge trans-
fer (ESICT) to the already complex behaviour resulting from
ESIPT reactions in excited states.
Furthermore, the results of the present work suggest that
phthalide is formed via the decomposition of epoxide, instead
of dione as reported earlier.12,17 It is proposed that the for-
mation of dione from epoxide takes place via C–O bond clea-
vage, and the formation of phthalide takes place via the C–C
bond cleavage of the oxiranyl ring of epoxide along the same
lines as suggested in the case of oxiranyl carbinyl radical C
(Scheme 4).24 These selective bond cleavages depend upon the
nature of the substituents in the rings of the substrate and,
most importantly, on the nature of the solvent in which the
Acknowledgements
We acknowledge the regional Sophisticated Instrumentation
Centre, Panjab University, Chandigarh and HRMS Lab, IIT
Ropar, for recording the NMR spectra and HRMS respectively.
We also acknowledge the Head, Department of Chemistry,
Punjabi University, Patiala for providing the facilities. Our
heartiest thanks are due to Prof. W. R. Bansal (Retd,
Department of Chemistry, Punjabi University, Patiala) for
helpful discussion.
References
1 G. D. Carlo, N. Mascolo, A. A. Izzo and F. Capasso,
Flavonoids: Old and new aspects of a class of natural thera-
peutic drugs, Life Sci., 1999, 65, 337–353.
2 L. Rackova, S. Firakova, D. Kostalova, M. Stefek, E. Sturdik
and M. Majekova, Oxidation of liposomal membrane sup-
pressed by flavonoids: Quantitative structure-activity
relationship, Bioorg. Med. Chem., 2005, 13, 6477–6484.
3 E. Middleton, C. Kandaswami and T. C. Theoharides, The
effects of plant flavonoids on mammalian cells:
This journal is © The Royal Society of Chemistry and Owner Societies 2019
Photochem. Photobiol. Sci., 2019, 18, 2912–2920 | 2919

View Article Online
Paper
Photochemical & Photobiological Sciences
Implications for inflammation, heart disease, and cancer, 18 Z. Szakacs, M. Kallay and M. Kubinyi, Theoretical study on
Pharmacol. Rev., 2000, 52, 673–751.
the photooxygenation and photorearrangement reactions
4 R. J. Nijveldt, E. van Nood, D. E. van Hoorn, P. G. Boelens,
of 3-hydroxyflavone, RSC Adv., 2017, 7, 32185–32192.
K. van Norren and P. A. van Leeuwen, Flavonoids: a review 19 M. Wera, V. G. Pivovarenko, A. Sikorski, T. Lis and and
of probable mechanisms of action and potential appli-
cations, Am. J. Clin. Nutr., 2001, 74, 418–425.
J. Blazejowski, 2-(Furan-2-yl)-3-hydroxy-4H-chromen-4-one,
Acta Cryst. E, 2011, 67, 0266.
5 J. B. Harborne and C. A. Williams, Advances in flavonoid 20 D. Zhang, X. Ji, R. Gao, H. Wang, S. Meng, Z. Zhong, Y. Li,
research since 1992, Phytochemistry, 2000, 55, 481–504.
6 P. K. Sengupta and M. Kasha, Excited state proton transfer
spectroscopy of 3-hydroxyflavone and quercetin, Chem.
Phys. Lett., 1979, 68, 382–385.
J. Jiang and Z. Li, Synthesis and antiviral activities of a
novel class of thioflavone and flavonoid analogues, Acta
Pharm. Sin. B, 2012, 2, 575–580.
21 M. Bansal and R. Kaur, Electromeric effect of substitution
at 6th position in 2-(Furan-2-yl)-3-hydroxy-4H-chromen-4-
one (FHC) on the absorption and emission spectra,
J. Chem. Sci., 2015, 127, 405–412.
7 A. J. G. Strandjord, D. E. Smith and P. F. Barbara, Structural
Effects
on
the
Proton-Transfer
Kinetics
of
3-Hydroxyflavones, J. Phys. Chem., 1985, 89, 2362–2366.
8 S. Protti and A. Mezzetti, Any colour you like. Excited state 22 K. Kaur, R. Kaur, J. Tomar and M. Bansal, Photoreactions
and ground state proton transfer in flavonols and appli-
cations, Photochemistry, 2012, 40, 295–322.
9 V. I. Tomin, A. P. Demchenko and P. T. Chou,
Thermodynamic vs. kinetic control of excited-state proton
of
2-(furan-2-yl)-3-hydroxy-4H-chromen-4-one
and
3-hydroxy-2-(thiophene-2-yl)-4H-chromen-4-one using cyclo-
hexane and acetonitrile as solvents, Photochem. Photobiol.
Sci., 2017, 16, 1311–1319.
transfer reactions, J. Photochem. Photobiol., C, 2015, 22, 1–18. 23 M. Shiozaki and T. Hiraoka, Photochemistry of β-hydroxy-
10 A. C. Sedgwick, L. Wu, H. H. Han, S. D. Bull, X. P. He,
γ-pyrone. A new synthesis of 3-methylcyclopent-2-en-2-ol-1-
one from Maltol, Tetrahedron Lett., 1972, 46, 4655–4658.
T. D. James, J. L. Sessler, B. Z. Tang, H. Tian and J. Yoon,
Excited-state intramolecular proton-transfer (ESIPT) based 24 J. Amaudrut and O. Wiest, Epoxide Formation by Ring
fluorescence sensors and imaging agents, Chem. Soc. Rev.,
2018, 47, 8842–8880.
Closure of the Cinnamyloxy Radical, Org. Lett., 2000, 2,
1251–1254 (and the references cited there in).
11 T. Matsuura, T. Takemoto and R. Nakashima, 25 F. E. Ziegler and A. K. Petersen, Tandem Fragmentation of
Photoinduced Reactions—LXXI*1: Photorearrangement of
3-hydroxyflavones to 3-Aryl-3-hydroxy-1,2-indandiones,
Tetrahedron, 1973, 29, 3337–3340.
Cyclopropylcarbinyl/Oxiranylcarbinyl Radicals. On the
Reversibility of Oxiranylcarbinyl/Allyloxyl Radical
Formation, J. Org. Chem., 1994, 59, 2707–2714.
12 I. Yokoe, K. Higuchi, Y. Shirataki and M. Komatsu, 26 D. J. Pasto, F. Cottard and C. Picconatto, The C3H5O
Photochemistry of Flavonoids. III. Photorearrangement of
Flavonols, Chem. Pharm. Bull., 1981, 29, 894–898.
13 W. E. Brewer, S. L. Studer, M. Standiford and P. T. Chou,
Dynamics of the Triplet State and the Reverse Proton
Radical System. Formation and Reactions of the Allyloxy
and Oxiranylmethyl Radicals Generated by the Photolysis
of the Allyl and 4,5-Epoxy-2-methyl-2-pentyl 4-nitrobenzene-
sulfenates, J. Org. Chem., 1994, 59, 7172–7177.
Transfer of 3-hydroxyflavone, J. Phys. Chem., 1989, 93, 27 F. E. Ziegler and A. K. Petersen, Allyloxy Radicals are
6088–6094.
Formed Reversibly from Oxiranylcarbinyl Radicals: A
14 M. L. Calabro, S. Tommasini, D. Raneri, P. Donato,
Kinetic Study, J. Org. Chem., 1995, 60, 2666–2667.
P. Ficarra and R. Ficarra, Selective reversed-phase liquid 28 D. J. Pasto, Ab Initio, Theoretical Studies on the Ring-
chromatography method for the kinetic investigation of
3-hydroxyflavone photostability, J. Chromatogr. B, 2004,
800, 245–251.
Opening Modes of the Oxiranyl-, Aziridinyl-, Oxaziridinyl-,
and Thiaranylmethyl Radical Systems, J. Org. Chem., 1996,
61, 252–256.
15 S. Protti, A. Mezzetti, C. Lapouge and J. P. Cornard, 29 V. Krishnamurthy and V. H. Rawal, Kinetics of the
Photochemistry of metal complexes of 3-hydroxyflavone:
towards a better understanding of the influence of solar
Oxiranylcarbinyl Radical Rearrangement, J. Org. Chem.,
1997, 62, 1572–1573.
light on the metal–soil organic matter interactions, 30 K. Chevalier, M. M. N. Wolf, A. Funk, M. Andres,
Photochem. Photobiol. Sci., 2008, 7, 109–119.
M. Gerhards and R. Diller, Transient IR spectroscopy and
ab initio calculations on ESIPT in 3-hydroxyflavone solvated
in acetonitrile, Phys. Chem. Chem. Phys., 2012, 14, 15007–
15020.
16 Z. Szakacs, M. Bojtar, L. Drahos, D. Hessz, M. Kallay,
T. Vidoczy, I. Bitter and M. Kubinyi, The kinetics and mecha-
nism of photooxygenation of 4′-diethylamino-3-hydroxy-
flavone, Photochem. Photobiol. Sci., 2016, 15, 219–227.
17 S. Protti and A. Mezzetti, Solvent effects on the photophy-
sics and photoreactivity of 3-hydroxyflavone: A combined
spectroscopic and kinetic study, J. Mol. Liq., 2015, 205,
110–114.
31 R. Kaur, K. Kaur and M. Bansal, Comparative Studies in
the Fluorescence and Excitation Spectra of 3-hydroxy-
flavone and 2(2′-heteroaryl)-3-hydroxy-4H-chromen-4-ones,
J. Appl. Spectrosc., 2017, 84, 219–225 (and the references
cited there in).
2920 | Photochem. Photobiol. Sci., 2019, 18, 2912–2920
This journal is © The Royal Society of Chemistry and Owner Societies 2019