Photochemistry of a 9-Dithianyl-Pyronin Derivative: A Cornucopia of Reaction Intermediates Lead to Common Photoproducts

Article abstract of DOI:10.1002/cplu.202000370

Leaving groups attached to the meso-methyl position of many common dyes, such as xanthene, BODIPY, or pyronin derivatives, can be liberated upon irradiation with visible light. However, the course of phototransformations of such photoactivatable systems can be quite complex and the identification of reaction intermediates or even products is often neglected. This paper exemplifies the photochemistry of a 9-dithianyl-pyronin derivative, which undergoes an oxidative transformation at the meso-position to give a 3,6-diamino-9H-xanthen-9-one derivative, formic acid, and carbon monoxide as the main photoproducts. The course of this multi-photon multi-step reaction was studied under various conditions by steady-state and time-resolved optical spectroscopy, mass spectrometry and NMR spectroscopy to understand the effects of solvents and molecular oxygen on individual steps. Our analyses have revealed the existence of many intermediates and their interrelationships to provide a complete picture of the transformation, which can bring new inputs to a rational design of new photoactivatable pyronin or xanthene derivatives.

Full text of DOI:10.1002/cplu.202000370

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
1
2
3
4
5
6
7
8
9
Photochemistry of a 9-Dithianyl-Pyronin Derivative: A
Cornucopia of Reaction Intermediates Lead to Common
Photoproducts
Marek Martínek,[a, b] Jiří Váňa,[c] Peter Šebej,[b] Rafael Navrátil,[d] Tomáš Slanina,[a, b]
Lucie Ludvíková,[a, b] Jana Roithová,*[e] and Petr Klán*[a, b]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Leaving groups attached to the meso-methyl position of many
common dyes, such as xanthene, BODIPY, or pyronin deriva-
tives, can be liberated upon irradiation with visible light.
However, the course of phototransformations of such photo-
activatable systems can be quite complex and the identification
of reaction intermediates or even products is often neglected.
This paper exemplifies the photochemistry of a 9-dithianyl-
pyronin derivative, which undergoes an oxidative transforma-
tion at the meso-position to give a 3,6-diamino-9H-xanthen-9-
one derivative, formic acid, and carbon monoxide as the main
photoproducts. The course of this multi-photon multi-step
reaction was studied under various conditions by steady-state
and time-resolved optical spectroscopy, mass spectrometry and
NMR spectroscopy to understand the effects of solvents and
molecular oxygen on individual steps. Our analyses have
revealed the existence of many intermediates and their
interrelationships to provide a complete picture of the trans-
formation, which can bring new inputs to a rational design of
new photoactivatable pyronin or xanthene derivatives.
Introduction
organic synthesis, although arduous conditions or transition-
metal catalysis are usually required to make such reactions
synthetically useful.[1,2] The photochemical homolytic cleavage
of nonpolar covalent bonds is frequent in many primary
photoinitiated processes,[3] whereas the photochemical hetero-
lytic cleavage is less common.[4–6] Photochemical heterolysis of
some common polar bonds may require a specific conical-
intersection control to give destabilized carbocations in relation
to the stabilized excited-state surfaces.[7] For example, the
specific role of conical intersections could be used for explain-
ing the “meta effect”[8] or the photochemical heterolysis of
carbon-leaving group (CÀ LG) bonds in some photoactivatable
systems (or photoremovable protecting groups, PPGs), which
allow spatial and temporal control over the release of various
molecules of interest.[9]
In the past decade, several novel visible-light-absorbing
photoactivatable systems have been developed. The arguably
first non-metal-based PPG absorbing over 500 nm, (6-hydroxy-
3-oxo-3H-xanthen-9-yl)methyl group, was shown to undergo an
intramolecular CÀ LG bond cleavage (Scheme 1a) upon irradi-
ation with green light.[10] Later, several meso-methyl-BODIPY-
based PPGs (Scheme 1b) have been designed to release
LGs,[11–13] representing synthetically accessible and thermally
stable scaffolds with structurally easily adjustable optical and
photophysical properties. Many other transition-metal-free PPG
systems have been reported to date.[14–19]
The controlled heterolytic cleavage of nonpolar covalent bonds,
such as a CÀ C bond, belongs among the most desired tasks in
[a] Dr. M. Martínek, Dr. T. Slanina, Dr. L. Ludvíková, Prof. P. Klán
Department of Chemistry
Faculty of Science
Masaryk University
Kamenice 5, 625 00 Brno (Czech Republic)
E-mail: klan@sci.muni.cz
[b] Dr. M. Martínek, Dr. P. Šebej, Dr. T. Slanina, Dr. L. Ludvíková, Prof. P. Klán
RECETOX
Faculty of Science
Masaryk University
Kamenice 5, 625 00, Brno (Czech Republic)
[c] Dr. J. Váňa
Institute of Organic Chemistry and Technology
Faculty of Chemical Technology
University of Pardubice
Studentská 573, 532 10, Pardubice (Czech Republic)
[d] Dr. R. Navrátil
Department of Organic Chemistry
Faculty of Science
Charles University
Hlavova 2030/8, 128 43 Prague (Czech Republic)
[e] Prof. J. Roithová
Institute for Molecules and Materials
Radboud University
Heyendaalseweg 135, 6525 AJ Nijmegen (The Netherlands)
E-mail: j.roithova@science.ru.nl
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cplu.202000370
Irradiation of pyronin derivative 1, analogous to (6-hydroxy-
3-oxo-3H-xanthen-9-yl)methyl moiety (Scheme 1a), bearing a
dithian-2-yl group in the C9-position, with visible light (λirr =
592 nm) was reported by some of us to exhibit a rather
unexpected chemical transformation at the meso-methyl carbon
atom.[20] The only observed intermediate was meso-methoxy
pyronin derivative 2, successively converted into 3,6-di(piper-
This article is part of a Special Collection on “Chemistry in the Czech Re-
public”.
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA. This
is an open access article under the terms of the Creative Commons Attri-
bution Non-Commercial License, which permits use, distribution and re-
production in any medium, provided the original work is properly cited and
is not used for commercial purposes.
ChemPlusChem 2020, 85, 1–14
1
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
serve as an example of the mechanistic investigation of an
extremely complex chemical transformation.
1
2
3
4
5
Results and Discussion
6
7
8
9
Photodegradation of the 9-dithianyl-pyronin derivative 1 is a
complex reaction. Classical spectroscopy techniques allowed us
to intercept one major intermediate (2, Scheme 2)[20] but at the
same time, they indicated that the overall transformation is
more complicated. In this paper, we first discuss the photo-
physical properties of reactants and major products, and then
we demonstrate how optical, mass and NMR spectroscopies
were used to identify and monitor different intermediates
formed during the reaction. Finally, all data together with
density functional theory (DFT) computations are used to
unravel the reaction mechanisms.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 1. Examples of recently developed visible light absorbing photo-
activatable systems based on (a) (6-hydroxy-3-oxo-3H-xanthen-9-yl)methyl
moiety and (b) meso-methyl-BODIPY core.[10,12]
idin-1’-yl)-9H-xanthen-9-one 3 as the final isolable product
(Scheme 2). Although the mechanism of this phototransforma-
tion was not investigated in detail, it was suggested that a
photoinduced electron transfer from an electron-rich sulfur
atom to the electron-deficient pyronin core might be the first
step followed by a nucleophilic attack of methanol as a solvent
at the C9 position. Analogous photoinduced electron transfer
from the dithiane moiety to an excited photosensitizer was
demonstrated by Falvey, Kutateladze, and coworkers to be the
key step in the fragmentation of dithiane-carbonyl adducts.[21,22]
The second plausible explanation of the phototransformation of
1 is the release of an alkylthiolate as a LG (nucleofuge), similar
to the reaction observed for PPGs shown in Scheme 1.
In this work, a thorough mechanistic study of the photo-
chemistry of pyronin 1 using several advanced steady-state and
transient spectroscopic and analytical techniques as well as
quantum-chemical calculations is presented. Several short-lived
intermediates were determined in the individual photochemical
steps, and the roles of oxygen and solvent molecules in the
reaction mechanism were established. Because the dithiane
Major Photoproducts
The photochemistry (λirr =592 nm, Figure S55 in the Supporting
Information) of 1 in methanol leads to the final major product,
3,6-diamino-9H-xanthen-9-one derivative 3 (>80% chemical
yield, Scheme 2), which can be isolated.[20] The reaction
proceeds further under exhaustive irradiation by white LEDs
(3×100 W, Figure S58) resulting in the decomposition of
1
molecule 3. H NMR analyses performed in this work revealed
that the formation of 3 is accompanied by the formation of
methyl formate (4, a signal at δ=8.1 ppm; Figures S60 and S63)
as a co-product in up to ~60% chemical yield. The reaction was
carried out in methanol, therefore, the initially formed com-
pound could either be methyl formate or formic acid that reacts
with methanol. Indeed, esterification takes place readily upon
the addition of formic acid in methanol under identical
conditions (Figure S76), however, it was impossible to prove or
refute the intermediacy of formic acid due to overall low
concentrations of intermediates during the irradiation experi-
ments (typically c(1) ~10À 5 moldmÀ 3). We also searched for
scaffold is
a
versatile protecting group for carbonyl
compounds,[23–25] our study of the photochemistry of 1 can
Scheme 2. Yellow-light induced transformation of pyronin-dithiane derivative 1.
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
2
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
gaseous photoproducts using GC-MS headspace analysis, and
we observed carbon monoxide (CO) evolution (~9% yield; CO
was also identified by IR gas-phase spectroscopy and a
reduction gas analyzer detector[26,27]). Experiments performed in
CH3OH/D218O (99:1, v/v) revealed that the 18O isotope is
incorporated into C18O, suggesting that it occurred before the
decarbonylation step took place.
sensitizer (xanthone: ΦISC ~1),[33] was irradiated at λexc =375 nm,
the wavelength at which 3 absorbs >80% of light, resulting in
an identical mixture of photoproducts (Figure S34). In more
thoroughly degassed methanol (6 freeze-pump-thaw cycles),
the degradation of 1 was slower by 1–2 orders of magnitude
compared to that observed in aerated solutions. Besides that,
products 2 or 3 were not formed (Figure S35). This suggests
that there is at least one principal reaction step that critically
relies on the presence of oxygen, such as Type I or II
oxygenations.[34–36] We estimated that the molar amount of
residual oxygen in this sample is lower than that of the starting
material (see later for details). It was also found that the
solution acidity increased upon exhaustive irradiation (the initial
glass electrode readout of 6.5 dropped to 3.5), therefore, the
course of the reaction could also be affected by a change in pH.
1
2
3
4
5
6
7
8
9
Photophysical Properties of Compounds 1and 3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Compound 1 exhibits a strong absorption at λmax(abs)=585 nm
(ɛmax =62000 molÀ 1 dm3 cmÀ 1) and 590 nm in methanol and
water, respectively, and a relatively strong emission at λmax(em)
=608 nm (Φf =(14.7�0.1)%;[20] τf =(1.35�0.02) ns) in methanol
(Figure S22). The excitation spectra match the absorption
spectra. The absorption spectrum of compound 1 in an
aqueous solution (c>1×10À 5 moldmÀ 3) showing an additional
band at 547 nm (Figure S23) suggests that the chromophore
aggregates[28,29] due to its hydrophobic and relatively rigid
structure. No aggregates were observed upon the addition of
10% of an organic co-solvent such as DMSO. The main product
3 (Scheme 2) features a major absorption band at λmax(abs)=
382 nm (ɛmax =33900 molÀ 1 dm3 cmÀ 1) and a strong emission
band at λmax(em)=448 nm in methanol.[20] Compound 3 is not
soluble in water even in the concentration range commonly
used for absorption spectroscopy measurements (10À 5–
10À 4 moldmÀ 3); the addition of water to methanol solutions of
3 leads to its precipitation.
Time-Resolved Experiments
Femtosecond pump-probe experiments were carried out to
characterize the major transients formed upon excitation of 1.
The transient spectra of 1 in methanol (c=5×10À 4 moldmÀ 3)
were recorded using 100-fs steps up to a 10 ps delay after the
pump pulse (λexc =280 nm). Fitting with a single exponential
decay resulted in the spectra of two species with λmax =465 nm
(lifetime τ=(1.1�0.1) ns) and λmax =425 nm (τ~3 ns; Figure 1),
which were assigned to the singlet (11*) and triplet (31*) excited
states, respectively, because their spectral features were similar
to those of analogous singlet- and triplet-excited Si-rhodamines
or phenyl and thienyl rhodamines.[37–39]
The signal of 31* in methanol (λex =266 nm, c(1)=1.3×
10À 5 moldmÀ 3) was not detected within the resolution of our
nanosecond laser flash photolysis setup (>5 ns), because the
lifetime of 31* is too short (see above). Instead, a new long-lived
Effects of Dioxygen
The major deactivation pathways of singlet-excited xanthene
and pyronin dyes are fluorescence and radiationless decays.[30]
Only derivatives bearing heavy atoms, such as rose bengal,
undergo efficient intersystem crossing.[31,32] We performed a
series of experiments in the presence of various triplet
quenchers and sensitizers to determine the multiplicity of the
productive excited state of 1, that is, the species responsible for
its initial photochemical step. The photodegradation efficiency
of 1 in methanol was found to be dependent on the
concentration of oxygen, which may act both as a triplet
quencher and as an oxidant. Upon irradiation at λexc =592 nm,
the observed rate of disappearance of 1, monitored by
absorption spectroscopy, increased three-fold in a deoxygen-
ated solution (3 freeze-pump-thaw cycles;
coxygen !5.1×
10À 8 moldmÀ 3) when compared to that found for an aerated
solution (Table S1 in the Supporting Information), which implies
that oxygen acts as a triplet state quencher. Because the
reaction was not quenched completely even in samples purged
with oxygen, we concluded that the lifetime of the productive
triplet state was too short to compete with a bimolecular triplet
energy transfer[9] or other alternative reaction pathways are
available. To further verify the triplet-state origin of the major
reaction step, a methanol solution of 1 (c=3×10À 5 moldmÀ 3)
and xanthen-9-one 3 (c=3×10À 5 moldmÀ 3), used as a triplet
Figure 1. Deconvoluted excited-state absorption spectra of 11* (black line)
and 31* (red line) resulting from a global analysis of the transient absorption
spectra measured at short delays (0–10 ps) after the pump pulse
(λexc =280 nm) in methanol (c(1)=5×10À 4 moldmÀ 3). Note: A broad negative
band between 530 and 620 nm is the ground-state bleach of 1; the peak at
~560 nm is an artefact from the laser pulse.
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
3
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
species (τ in the order of ms) was observed in the range of 560–
600 nm, and it was attributed to the first, relatively long-lived
intermediate formed directly from 31*.
photodissociation spectroscopy (Figures S1–S4 in the Support-
ing Information).[44–46] At short photolysis times, we detected the
first low-intensity signal at m/z 497 that corresponds to an
adduct of 1+32 mass units. It was assigned either to an adduct
of 1 with methanol (6 or 7) or a product of oxidation of 1, such
as sulfone 8. Upon prolonged irradiation, major signals
corresponding to 2 (m/z 377), 5 (m/z 407), and 9H+ (m/z 439)
and minor signals assigned to 10 (m/z 347) and product 3H+
(m/z 363) were found. Upon irradiation of 1 in CH3CN/H2O (9:1,
v/v), major signals corresponding to aldehyde 11 (m/z 375) and
its hydrate 12 (m/z 393) were observed, whereas photolysis of 1
in CH3OH/H2O enabled the detection of intermediates 13 and
14. Many of the detected species were formed by reactions of
intermediates with methanol or water molecules, which was
confirmed by performing the reactions in isotopically labeled
solvents (see the m/z signal changes in Figure 3).
Structures of all ions were assigned based on their
fragmentation patterns and based on their vibrational spectra
(e.g., Figure 4, see also Supporting Information). We measured
infrared spectra of mass selected ions by helium tagging
photodissociation approach.[47] The experimental helium tag-
ging infrared photodissociation (IRPD) spectra are usually well
compared with theoretical IR spectra of ions in the gas phase
allowing their assignment based on a comparison. One example
is shown in Figure 4 where the experimental IRPD spectrum of
ions at m/z 375 matches with the theoretical spectrum of
aldehyde 11 allowing the assignment.
Having established the structures of the detected intermedi-
ates, we also investigated photofragmentation spectra of
compounds 2, 5, 11, 13, and 14 in the visible part of the
spectrum (430–700 nm).[48] Upon absorption of a photon, the
mass-selected ions in the gas phase fragment. The absorption
spectra of the mass-selected ions are determined from the yield
of this photofragmentation as a function of the photon wave-
length (Figures 5 and S2). The absorption band of isolated ion 1
was found hypsochromically shifted by ~40 nm compared to
that of 1 in a methanol solution (Figure S2), and we expected to
find similar media effects for all other investigated xanthene
derivatives. Their absorption spectra at least partly overlap with
that of 1. Because the intermediates were formed during the
course of irradiation, they could also be excited except for
intermediate 2, the spectrum of which is hypsochromically
shifted outside the wavelength range of an irradiation source.
Interestingly, the photofragmentation of 11, 13, and 14 in the
gas phase resulted in decarbonylation; for example, carboxylate
13 gives directly the final product 3, and methyl ester 14
photofragments to 2. These phototransformations are unim-
olecular processes and do not reflect solvent effects. Therefore,
they are only remotely useful for the interpretation of chemical
processes occurring in protic solvents.
1
2
3
4
5
6
7
8
9
Detection and Assignment of Reaction Intermediates
Monitoring the 1!2!3 transformation by steady-state UV-vis
spectroscopy suggested that the reactions are not elemental
(Figure S24).[20] The spectra do not exhibit clear isosbestic points
even at low concentrations of 1. Thus, other intermediates were
most likely formed in low concentrations, which hampered their
immediate detection. Electrospray ionization mass spectrometry
is an ideal tool to detect such low abundant species.[40–42] Hence,
we irradiated methanol, methanol/water, and acetonitrile/water
solutions of 1 and analyzed the detected ions (λirr =592 nm,
Figure 2).
The structures of all relevant ions were determined based
on their IR characteristics and other supporting methods, such
as collision-induced dissociation experiments or photofragmen-
tation pattern.[43–46] For clarity of the following text, these
assignments, except for one example, are discussed in the
chapter “Going Deeper: The Structural Assignment of Detected
Ions” in Experimental Section, and the remaining details are
provided in Supporting Information.
The detected dominant signals corresponded to inherently
charged species bearing a conjugated xanthene backbone as
those of reactant 1 (m/z 465) and intermediate 2 (m/z 377).
Charge-neutral molecules, such as 3, could also be detected in
their protonated forms (i.e., 3H+ at m/z 363), albeit as less
intense signals. Besides, the m/z signals of several additional
intermediates were observed (Figure 3; except formate 4, see
the text above). Their abundances varied depending on the
reaction time and the type of solvent. For example, Figure 2
shows intense signals of intermediate 5 and weak signals of
compounds 6–9. Their structures were assigned based on
collision-induced dissociation (CID) experiments (Figures S7 and
S8),[43] isotopic labeling (Figures S9, S43, and S44), and infrared
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Furthermore, we investigated which intermediates are
formed directly upon irradiation and which are formed in the
subsequent dark processes (Figures S42 and S43). The photo-
degradation of 1 was found to correlate with the increase in
signal intensities of methoxy intermediate 2, hemiacetal 5, and
a methanol adduct of hemiacetal 9H+. These results showed
that 9H+ is likely formed from 5 because the growth of the 9H+
Figure 2. A representative MS spectrum collected after 1 h of irradiation of 1
(c=10À 5 moldmÀ 3) in methanol with LEDs (λirr =592 nm, for the setup see
Figures S54 and S55).
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
4
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 3. Species detected by electrospray ionization mass spectrometry upon irradiation of 1 in indicated solvents. The values refer to the corresponding m/z
signals of intermediates in given isotopically labeled solvents.
signal was always slightly delayed. The intensity of the signal at
m/z 497, tentatively assigned above to an adduct of 1 with
methanol (6 or 7) or sulfone 8, formed by the oxidation of 1,
increased only initially and then slightly decreased during the
first dark period (Figure S41). Compound 6 is most probably
metastable and should not be responsible for a persistent
signal, whereas intermediate 7 should react further upon
irradiation (see later). Therefore, possibly photochemically less
reactive sulfone 8 was assigned to the m/z 497 signal.
NMR Spectroscopy
Figure 4. (a) Helium tagging IRPD spectrum of mass selected ion at m/z 375
measured at 3 K. (b) Theoretical infrared spectrum of 11 calculated at the
B3LYP-D3/6-311+G** level of theory. Theoretical vibrational frequencies
were scaled by 0.975. Gaussian broadening (FWHM of 5 cmÀ 1) was applied to
the theoretical infrared spectra for better comparison with the experimental
infrared photodissociation spectra.
The changes in concentrations of 1–5 and 7 upon irradiation of a
CD3OD solution of 1 in an NMR tube (containing a sealed capillary
with an internal standard) by white LEDs (3×100 W; Figure S58)
1
were analyzed by H NMR spectroscopy to follow the reaction
kinetics and to determine the interrelationships of the detected
intermediates. The experiments were performed under argon, air,
or oxygen atmosphere to study the effects of oxygen on the
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
5
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 5. Visible photofragmentation spectra of indicated mass-selected ions
(right), and the corresponding photofragmentation mass spectra at the
indicated irradiation wavelength (left).
individual reaction steps (Figures 6a–c). It should be emphasized
that the concentrations of 1 in NMR experiments were two orders
of magnitude higher (c=4×10À 3 moldmÀ 3) than those used in MS
and UV-vis (c~10À 5) experiments described above. NMR analyses
enabled a more reliable detection of two compounds, thioacetal 7
and ketone 3, for which the mass spectrometry analysis was not
sufficiently sensitive.
At first glance, the occurrence of all species detected by NMR
(Figure 6a) was found in the following order: 1!7!5!2. The
final product 3 was formed already in the first minutes of
irradiation, whereas formate 4 appeared later. The maximum
chemical yields of intermediates 2 and 5 as well as final products
3 and 4 reached 60–65% in aerated solutions, that of 7 was
~30%. The only compound the concentration of which remained
constant even upon exhaustive irradiation was 4. The chemical
yield of CO photoproduction (headspace GC/MS) was 8%.
The initial fast formation of 3 in aerated samples (ca. 30% yield
maximum after 0.5 h; Figure 6a), attributed to direct oxidation of 1
and its dark decomposition to 3 (see later), is followed by the
subsequent slower photochemical formation of 3 to give a
maximum (~60%) yield after 70 h (see an inset of Figure 6a). In
contrast, the yields up to 50% were observed in oxygen-purged
samples (Figure 6b) after 0.5 h, however, no significant production
of 3 was detected at longer irradiation times. In the case of argon-
purged solutions (Figure 6c), only an inefficient formation of 3
occurred, and its kinetics was similar to that of the second (slow)
process observed in oxygenated solutions.
Figure 6. Time evolution of the integrated 1H NMR signals in irradiated
solutions of 1 in methanol-d4 which were saturated with (a) air, (b) oxygen,
and (c) argon (data points are connected with straight lines for clarity). The
insets represent longer time scales.
without significantly affecting the maximum chemical yields of
both 5 and 2 (Figure 6b) may have a plausible explanation. Our
experiments suggested thattriplet-excited compound
1
is
quenched by oxygen only partially and the overall rate of the
disappearance of 1 is much faster at higher O2 concentrations,
which implies that there is an alternative pathway for the
conversion of 1 into 5. The oxidation of sulfur of the dithiane
moiety may result in the formation of sulfones (see analysis of the
ions detected by ESI-MS), which could also be converted into
common intermediate 5. The synchronized formation of other
intermediates follows the same order of appearances observed in
Figure 6a. The formation of formate proceeded concomitantly
with the production of 2 and preceded the formation of 3. The
chemical yield of CO (headspace GC/MS) was 9%.
The reaction kinetics in a solution saturated by oxygen was
apparently faster (Figure 6b, the methanol solution was purged
with oxygen, coxygen =10.2×10À 3 moldmÀ 3).[49] A two-fold faster
decomposition of 1 and the formation of 5 (and subsequently 2)
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
6
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
Both the degradation of 1 and the formation of 5 in solutions
opening of 1,3-dioxolane rings attached to 4-coumarinyl[55] or 2-
(2-hydroxy-4-methoxyphenyl)[56] cages has also been reported.
However, the former study showed that six-membered 4-
coumarinyl-1,3-dioxanes are photochemically inert because of
the inherently higher thermodynamic stability of a six-mem-
bered ring.[55]
1
2
3
4
5
6
7
8
9
of 1 purged with argon (coxygen ~5.1×10À 8 moldmÀ 3; Figure 6c)
were faster compared to rates observed in aerated solutions. We
hypothesized that it is related to the fact that the triplet-excited
state of 1 is not fully quenched by oxygen and that oxygen is not
required in the reaction leading to 5 (Figure 6c). The NMR signals
of 5 were weaker and methoxy derivative 2 was formed faster
than in aerated/oxygenated solutions (Figures 6a and 6b). We thus
imply that the excited state of 5 preceding the formation of 2 is
also a triplet. Lower overall maximum yields of intermediates 2
and especially 3 suggest that oxygen also acts as an oxidant
during their formation. The oxygen level in argon-purged samples
might be low enough to preclude quenching of the triplet states
but still large enough (1 equiv. at least) to oxidize the intermedi-
ates. Formate 4 was not initially formed in the absence of oxygen.
The chemical yield (headspace GC/MS) of CO was slightly higher
(13%) compared to those of aerated and oxygenated samples
(8%).
The presence of oxygen in solution partially quenches the
3
triplet state 1*, but higher concentrations of dissolved O2 lead
to dithiane oxidation. The oxidation products may undergo
further reactions or eventually give the same final product 3.
Cationic pyronin chromophores, such as 1, are known to be in
equilibrium with neutral xanthydrol 16 in the presence of
water[57] (Scheme 4) or with their methoxy analogs in methanol.
Only some of them were detected by our analytical techniques
(e.g., 6, 9, 17). The absorption spectra of such species are largely
hypsochromically shifted (λmax ~460 nm),[57] thus they cannot be
directly excited at the irradiation wavelength used in our
experiments. The equilibrium constant for pyronin-based (=P)
xanthydrol 16 ([POH]/[P+][OHÀ ]) of À 2.6[57] suggests that their
participation in any chemical transformation is minimal. This
was evidenced in an experiment, in which the light source was
periodically turned off and on (Figure S41). The intensities of
the MS signals of major intermediates 11, 7, 5 remained
constant during the light-off phase, therefore, their trans-
formations were initiated photochemically.
In contrast, the changes in the signal intensities of
intermediates lacking the xanthene-core conjugation, such as 6,
9, and 17 (Figure S41), were found photochemically inert. The
photochemistry of 1 in methanol solution containing furfur-
ylalcohol (0.1%vol.) provided no evidence of the presence of
singlet oxygen trapping products. Therefore, we suggest that
the formation of singlet oxygen is negligible and does not
interfere with the reaction course. Cyclic thioethers readily
undergo single-electron-transfer oxidation due to their low
oxidation potential, which was utilized in the photochemical
ring-opening of 1,3-dithiane to give carbonyl compounds in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
We paid a special attention to the formation of aldehyde 11
as the last identified intermediate on the pathway leading to 2,
particularly because NMR experiments did not provide any
evidence of its presence in the spectral region typical for
aldehydes. On the other hand, the formation of 11 and its
hydrate 12 was unambiguously detected by ESI-MS, particularly
in the mixture of CH3CN/H2O (9:1, v/v; Figure S9).
Mechanism
We identified major intermediates (MS) and monitored their
concentrations (NMR) in the course the photoreaction of 1
carried out under various conditions. Our experiments sug-
gested the order of their appearances, 1!7!5(!11)!2, and
their maximum concentrations and the absence of other
(photo)products gave us the first evidence about their
interrelationships. Scheme 3 shows the proposed overall mech-
anism based on the results from all experimental and computa-
tional studies. The scheme depicts structures of intermediates
detected by our analytical methods (black), plausible short-lived
species (red), and possible but undetected intermediates (blue).
presence
of
oxygen
via
dithianyl
radical
cation
intermediates.[58–60] Such a radical cation undergoes either
single-electron transfer to form dithioacetal dication to trigger
the CÀ S bond cleavage and subsequent hydrolysis to give
ketone,[61] or it can be attacked by air oxygen or superoxide
radical anion to form persulfides.[58,62] Dithiane oxidation was a
significant pathway observed only in oxygen saturated solu-
tions, however, we observed no EPR signal in the irradiated
mixtures (see Supporting Information for more details), thus we
have not further investigated this process.
3
The first reaction step of triplet excited 1*[39,50,51] is the CÀ S
bond heterolytic cleavage resulting in the dithiane ring open-
ing. The formation of a carbocation intermediate followed by
the nucleophilic attack of a solvent, such as methanol, to give 7
would be a plausible pathway analogous to the reactions
observed for meso-methyl BODIPY[11–13] or xanthene[10] deriva-
tives (Scheme 1). The concentration of intermediate 7 was
always found to be very low, suggesting that the photo-
chemical efficiency of its subsequent photodegradation is very
high. If undetected hemithioacetal 15 is formed in the presence
(or traces) of water instead, it must be promptly converted into
a different species, such as 5 or 9. Indeed, the hydrolysis of
hemithioacetals is known to be diffusion-controlled under acid
or base catalysis.[52,53] Thiolate as conjugate base of a moder-
ately strong acid is a mediocre leaving group (although better
than alcoholates), which can be liberated, for example, from a
4-hydroxyphenacyl PPG upon irradiation.[54] The photochemical
Additional photon is needed for the phototransformation of
7 into hemiacetal 5, which is analogous to the conversion of 31*
into 10. Compound 5 was always observed as one of the
principal intermediates and its maximum concentration typi-
cally reached 60% (NMR). Its further transformation to hydrate
12, which is in equilibrium with aldehyde 11, evidently took
place via the triplet excited state (see quenching with oxygen
above; Scheme 3), but we should also not exclude a dark
hydrolysis as a minor pathway.
The most difficult part of this investigation was to clarify the
mechanism of the conversion of 11 to 2 (and subsequently 3).
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
7
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 3. The proposed mechanism of the phototransformation of 1 in a methanol/water mixture, showing the intermediates detected by our analytical
methods (black), plausible short-lived species (red), and possible but undetected intermediates (blue).
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
8
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 5. Formation of a putative α-lactone intermediate 19 and subse-
quent CO liberation.
Scheme 4. Equilibrium between 1 and its corresponding xanthydrol 16 in an
aqueous solution, and 17 as the product of oxidation of 1 or 16.
irradiation of 1 in methanol, and thus they were synthesized
independently to elucidate their roles in the reaction mecha-
nism (Scheme S2). Upon the irradiation of 13 and 14 under the
reaction conditions identical to those used in the photolysis of
1, their photodegradation quantum efficiency was at least
100 times lower than that of 1 (Figures S38 and S40). The
compounds were almost unreactive even in the mixture with
other intermediates formed upon irradiation of a solution of 1
(Figure S39). Therefore, we excluded both 13 and 14 as key
intermediates of the 11!3 reaction; their contribution as a
source of CO has to be insignificant. Direct photodecarbonyla-
tion of aldehyde 11 was also considered but this hypothesis
suffered from two drawbacks. The formation of C9 radical was
found to be endothermic by 50 kcalmolÀ 1 (M06L-D3/def2TZVP,
PCM (MeOH); Figure S19), and the transition-state energy of the
CO dissociation via the direct homolytic CÀ C bond cleavage
should also be very high as we could not localize the transition
state in our DFT calculations. However, the products are
~89 kcalmolÀ 1 higher in energy than aldehyde 11; Figure S20).
In addition, the direct decarbonylation of 11 should give
pyronin 10 formed by the reduction of pyronin C9 radical
(Figure S19). Because 10[79] is unreactive under the given
conditions as evidenced by the photolysis of an authentic
sample of 10 by 592-nm light in this work, it should accumulate
in the reaction mixture, which was not observed. Therefore, the
decarbonylation of 11 was rejected as the major pathway.
This transformation involves bond umpolung, requires the
presence of oxygen, and consists of several steps (Figure S12).
However, we were not able to detect any additional associated
intermediates. The CÀ C bond is cleaved and the overall reaction
is oxidation. Oxygen (O2) has a dual effect on the reaction; it
slows down some of the individual reaction steps but is
required as a reactant (oxidant). We proposed nine possible
reaction mechanisms for this transformation (Figures S12–S21).
Quantum chemical calculations at the M06L-D3/def2TZVP PCM
(MeOH) level of theory[63–66] allowed us to rule out five proposed
mechanisms and one was excluded based on the absence of
suggested photoproducts (Supporting Information).
Finally, the three remaining plausible reaction mechanisms
of the formation of 2 from aldehyde 11 are: (A) the formation of
pyronin-9-carboxylic acid 13 via oxidation of aldehyde 11 and
its photochemical conversion to 3, (B)
a single-electron
reduction of 11 and subsequent oxidation, and (C) a solvent-
assisted Dakin-like reaction[67] to give 2 as the key intermediate
on the way to 3 (Scheme 3).
In pathway A, carboxylic acid 13 could simply be formed by
a dark autoxidation of aldehyde 11 in the presence of
oxygen[68,69] via radical photoinitiation (Scheme 3).[70] The result-
ing acyl radical could efficiently couple with ground-state
oxygen to form peroxy acid 18,[71,72] which would react with
another molecule of aldehyde and rearrange to give a Criegee-
like intermediate, which disproportionates into two molecules
of acid 13.[73,74] Our previous work suggested that pyronin-9-
carboxylates, analogous to xanthene-9-carboxylic acid,[75] could
release carbon monoxide (CO) upon irradiation via a putative α-
lactone intermediate (19, Scheme 5), observed, for example, in
atmospheric photochemistry of methacrolein.[76,77] Another
example is the 9-acridinecarboxaldehyde oxidation to give 9-
acridinecarboxylic acid and 9(10H)-acridinone, observed as a
minor byproduct.[78]
The second plausible reaction, pathway B, involves pyronin-
9-peroxyradical-9-carbaldehyde 20 as the key intermediate and
starts with the photoreduction of the triplet excited state of
electron-deficient aldehyde 11 in the presence of an electron
donor (Scheme 3), such as a sulfur-containing dithiane frag-
ment, observed in structurally similar rhodamines.[80,81] The
*
*
reaction gives pyronin 11 and thiyl radicals.[82] Intermediate 11
could react with triplet ground-state oxygen to form 20,[83]
which would further fragment to an acyl radical[84] and CO. An
analogous reaction was proposed for methacrolein undergoing
oxidation with atmospheric oxygen.[76,77] This pathway is the
most probable decarbonylation step, although it is still a minor
process in the overall reaction.
Although CO was detected as a reaction side-product in less
than 10% yield, we searched for the presence of carboxylic acid
13 and its methyl ester 14 in irradiated samples by ESI-MS.
Small amounts of both compounds were detected upon
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
9
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
The most important evidence in our investigation was that
dissociation to give product 3 and formic acid (pathway Ca,
Scheme 6; note that energies are given relative to the energy of
23) with an activation energy of 22.8 kcalmolÀ 1 (green lines;
Figure S15). Alternatively, 21 can rearrange to form peroxy
anion 23, which lies 7.5 kcalmolÀ 1 higher in energy than 21. The
structure of this hydroperoxy intermediate 23 has not been
determined experimentally (either NMR or ESI-MS) in this work,
although it figures as the key reactant in three reaction
pathways. It can undergo a 1,2-aryl migration (pathway Cb,
Scheme 6) to give an undetected 9-formyloxypyronin 24 in a
Dakin-like process, a hydrogen migration (pathway Cc) to form
carboxylic acid 13 or a proton shuttle via 25 to produce
dioxetane 26 (pathway Cd, Scheme 6). All these reaction path-
ways possess very similar activation energies. Pathways Cb and
Cc require the participation of a protic solvent molecule in the
transition state (e.g., water or methanol; the energies in
Scheme 6 are calculated with a methanol assistance). The
hydrogen migration (pathway Cc) is slightly energetically
preferred to the Dakin rearrangement (Cb; Figure S15). Pathway
1
2
3
4
5
6
7
8
9
methyl formate (as we demonstrated above, formic acid can be
rapidly esterified in methanol) was produced in high chemical
yields (~60%) along with high amounts of successive methoxy
intermediate 2 and ketone 3. We propose that the third
reaction mechanism of the 11!3 transformation involves a
water- or methanol-assisted Dakin-like reaction in the presence
of oxygen (pathway C, Scheme 3).[67] The original Dakin method
uses an excess of hydrogen peroxide and sodium hydroxide at
elevated temperatures. Interestingly, flavin also catalyzes Dakin
oxidation of aromatic aldehydes in the presence of oxygen or
hydrogen peroxide as sacrificial oxidants under mild
conditions.[85,86]
These examples have led us to suggest the most plausible
mechanism of this reaction sequence. A tentative 9-hydro-
peroxy intermediate (21), formed by the reaction of pyronin
radical 22 with oxygen (O2) and an H-atom source with a weak
CÀ H bond ([H]), such as thiols formed upon dithiane ring
opening (Scheme 3), at the meso-carbon, can undergo a direct
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 6. The formation of 9-hydroperoxy intermediate 21 followed by: (Ca) the direct rearrangement of 21 into 3 and formic acid, (Cb) Dakin-like reaction
of aldehyde 11, (Cc) 1,2-hydrogen migration leading to acid 13, and (Cd) the formation of dioxetane 26. The Gibbs energies of the corresponding structures
and transition states were calculated at 298 K in methanol solution (see also Figure S15) and are given relative to the energy of 23 (G298MeOH =À 951.054599
Hartree).
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
10
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
the reaction with an alcohol. However, the signal at m/z 497 was
also present and its intensity remained relatively constant during
the whole course of irradiation. Collision-induced dissociation of
ions at m/z 497 showed a dominant elimination pathway of
64 mass units, which correspond to SO2 (Figure S7). Therefore, the
structure of this ion most probably corresponds to oxidized 1 with
one sulfide group transformed to sulfone 8.[20] Ions appearing in the
beginning of the reaction corresponding to the adduct with alcohol
could either be the products of an attack of methanolate at the C9-
carbon atom of xanthene forming 6 (detected as protonated ions;
this species should be photochemically inactive)[88] or a product of
the ring opening (7).[20] Either 6H+ or 7 appear at low abundances
for a short reaction time, therefore, we could not further character-
ize them.
Cd involving dioxetane 26 and leading to product 3 proceeds
with the same activation energy as that of Cc. A stepwise Dakin-
like dissociation to give formyloxy intermediate 24 must be
assisted by a proton-shuttle molecule, such as protic methanol
or water. The calculated relative energies of the corresponding
intermediates and transition states were found to be close to
each other (within ~2 kcalmolÀ 1) for methanol- and water-
assisted pathways (Figure S15).
The nucleophilic substitution of the formyloxy group at the
C9 position in 24 with methanol results in the formation of
methoxypyronin 2.[20] Both SN1-like or addition-elimination
mechanisms are possible. Intermediate 2 was found as one of
the most persistent intermediates, the concentration of which
reached ~40% yield (1H NMR) in a typical reaction setup.
Finally, 2 is slowly converted into ketone 3 in the presence of
water, even in the absence of light, as reported before.[20]
Conclusions
This paper presents a thorough mechanistic study of the
photochemistry of 9-dithianyl-pyronin 1 in methanol using
advanced steady-state and transient spectroscopic and analyt-
ical techniques as well as quantum-chemical calculations.
Several photoreactive intermediates were determined, and the
roles of dioxygen and nucleophilic protic solvents in the
reaction mechanism were established. The photoreaction was
shown to proceed through pyronin-9-carbaldehyde 11 as the
key intermediate and we suggested three plausible pathways
for its conversion to the final photoproducts, xanthen-9-one
derivative 3 and formate 4: (A) the formation of pyronin-9-
carboxylic acid 13 via oxidation of aldehyde 11 and its
photochemical conversion into xanthen-9-one 3, (B) single-
electron reduction of 11 and its subsequent oxidation, and (C)
either the direct dissociation of oxidized intermediate to formic
acid and xanthen-9-one 3 or a solvent-assisted Dakin-like
reaction to give methoxy derivative 2 as an intermediate on the
way to 3. The photochemistry of compound 1 induced by
visible-light irradiation at 592 nm in the presence of oxygen
involves several redox processes and a CÀ C bond cleavage (~
80–90 kcalmolÀ 1).[87] Despite the complexity of the overall
mechanism, we show that although the xanthen-9-yl moiety is
not a suitable photolabile protecting group for aldehydes in
this particular case, our mechanistic investigations provided
many useful and sometimes surprising insights into the (photo)
reactivity of substituted xanthene/pyronin dyes, the photo-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Ion at m/z 407 (Figure 3): The signal at m/z 407 corresponds to
hemiacetal 5 (Figure 2). The assignment was based on the agree-
ment of the infrared photodissociation spectrum of this ion and the
calculated IR spectrum of 5 (Figure S48). Further evidence for the
structural assignment of 5 was based on the results from irradiation
experiments performed in methanol-d4 or D218O (CH3CN/D218O
mixture, 99:1, v/v), in which the corresponding MS signals of
deuterium- and 18O-labeled structures, respectively, were detected
(Figures 3 and S9). The main dissociation pathway of 5 under CID
conditions was found to be the loss of methanol (32 mass units,
Figure S7) with the concomitant formation of a new species at m/
z 375 (later assigned to aldehyde 11, Figure 4). Similarly, 1-d yields
5-d (m/z 408) upon irradiation and 5-d loses methanol (32 mass
units, Figure S10) in CID to form aldehyde 11-d (m/z 376). The
aldehyde 11-d loses CO in the MS experiment to form 9-(2H)-
pyronin 10-d (m/z 348, Figure S10). In addition, the photolysis of 1
in CH318OH gave the expected analogous species at m/z 409,
whereas the irradiation of 1 in ethanol and propan-2-ol led to the
formation of analogous intermediates at m/z 421 and 435,
respectively. Finally, the photolysis of 1 in CD3OD gives the ion at
m/z 411 that undergoes equilibration after dilution of the mixture
1
13
with CH3OH (Figure S44). We also analyzed the HÀ C HSQC NMR
1
spectra of 1 irradiated in CD3OD, and the observed H signal at δ=
6.47 ppm (HOÀ CHRÀ OCH3, R=pyronin; Figure S64) was correlated
with the 13C signal at δ=93.5 ppm (HOÀ CHRÀ OCH3; Figure S64);
both signals are in the typical range expected for hemiacetals. The
maximum chemical yield of 5 was ~60% (based on a quantitative
NMR analysis).
Ion at m/z 391 (Figure 3): These ions correspond to either
carboxylic acid 13 or formyloxy intermediate 24. Two isomers of 13,
carboxylic acid 13a and zwitterion 13b, which could exist as an α-
lactone (see Scheme 5), can be proposed (Figure 7). Helium tagging
infrared photodissociation (IRPD) spectrum of this ion (m/z 391)
does not show any bands in the range of 1700–2000 cmÀ 1. The lack
of a clear carbonyl stretching band suggests that the detected ions
do neither correspond to zwitterion 13b nor to formyloxy
intermediate 24. Also, all other ions studied by IRPD spectroscopy
(Figures 5, S2, and S3) have a conjugated xanthene backbone,
which gives well-resolved spectra in the 1400–1700 cmÀ 1 range.[89]
Here, the spectrum is more convoluted which can be associated
with the protonation of one of the piperidine moieties. Hence, the
IRPD spectrum represents ion 13 in the form of zwitterion 13b.
Zwitterions in the gas phase lie considerably higher in energy than
their neutral forms (see the relative energies in Figure 7); however,
it has been reported earlier that protic solvents support electro-
spray ionization transfer of solution-favored protonated forms.[90]
Ion 13 predominantly eliminates CO in the collision-induced
dissociation, followed by the CO2 elimination. This is again
consistent with the structure of zwitterion 13b.
bleaching of which can be
applications.
a serious obstacle in their
Experimental Section
Detailed Information on the Structural Assignment of
Detected Ions
Ion at m/z 497 (Figure 3): The first intermediate on the degradation
pathway of 1 is the ion at m/z 497. This ion is detected right in the
beginning of the reaction. Formally, it can be formed by addition of
methanol or two oxygen atoms to 1. If the reaction proceeds in
ethanol, CD3OD or propan-2-ol, we observed a mass shift of these
adducts to m/z 511, m/z 501, and m/z 525, respectively, confirming
Ion at m/z 375and m/z 393 (Figure 3): Species at m/z 375 and m/
z 393 were detected upon irradiation of 1 in a CH3CN/H2O mixture,
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
11
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
with CH3OH (Figure S43). The main dissociation pathway in CID and
1
2
3
4
5
in the photodissociation spectrum is the loss of the methyl group
(Figure S7 and Figure 4, respectively). The IRPD spectrum provided
a good agreement with the predicted one (Figure S50). We also
assigned the 1H NMR signals at δ=8.2 ppm to two aromatic
hydrogens which are spatially closest to those of the methoxy
group (Figure S63) and we estimated the yield of its formation
(50%, 1H NMR).
6
7
8
9
Acknowledgements
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Support for this work was provided by the Czech Science
Foundation (GA18-12477S). We also thank the CETOCOEN
EXCELLENCE Teaming 2 project (supported by the Czech Ministry
of Education, Youth and Sports: CZ.02.1.01/0.0/0.0/17_043/
0009632 and EU H2020: 857560) and the RECETOX research
infrastructure (LM2018121) (P.K.). The results from the CLIO were
obtained thanks to the funding from the European Union’s
Seventh Framework Programme (FP7/2007-2013) under the grant
agreement No. 226716 (J. R.). The CLIO staff, particularly Vincent
Steinmetz, is acknowledged for their help and assistance. In
addition, we would like to thank Jiří Huzlík (CDV, v.v.i., Brno),
Zdeněk Moravec (Masaryk University) and Lucie Muchová for
consultations on the quantification of CO, Juraj Jašík for his help
with ISORI MS experiments and technical assistance (both from
the Charles University, Prague), Ján Tarábek (IOCB, Prague) for
EPR analysis. We would also like to acknowledge Tatsiana Rusina
and Branislav Vrana for the head-space GC-MS analysis, Miroslava
Bittová for the HRMS analysis, Peter Horváth for the measurement
of fluorescence quantum yield and Luboš Jílek (all from Masaryk
University) for his assistance with light sources.
Figure 7. (a) Helium tagging infrared photodissociation spectrum of mass-
selected ions at m/z 391. (b) Theoretical infrared spectra of carboxylic acid
13a, its zwitterion 13b, and formyloxy derivative 24 calculated at the B3LYP-
D3/6-311+G** level of theory. Theoretical vibrational frequencies were
scaled by 0.975. Gaussian broadening (FWHM of 5 cmÀ 1) was applied to the
theoretical infrared spectra for better comparison with the experimental
infrared photodissociation spectra.
and we assigned them to pyronin-9-carbaldehyde 11 and pyronin-
9-yl dihydrate 12, respectively (Figure 3). We did not detect these
species in a methanolic solution with low water content, most likely
because the equilibrium between 8 and 11 and their corresponding
methoxy hemiacetals and acetals was shifted strongly toward the
latter. The IRPD spectrum of the ion at m/z 375 is in close
agreement with the calculated IR spectrum of 11 (Figure 4). The
CID spectrum shows the elimination of CO to form ions at m/z 347,
presumably 9-H-pyronin 10 (Figure S8). We have also detected a
hydrate of 11, ion 12 (m/z 393). Hydrate 12 loses a molecule of
water upon collisions to form 11.
Conflict of Interest
The authors declare no conflict of interest.
Ion at m/z 439 (Figure 3): The intermediate at m/z 439 was assigned
to an adduct of 5 and methanol, 9 (Figure 3), based on the
agreement of its IRMPD spectrum with the calculated one (Fig-
ure S45). The experiments performed under the same conditions in
CD3OD provided the ion at m/z 447, whereas an analogous adduct
at m/z 469 was found in ethanol and the ion at m/z 440 upon
irradiation of 1-d in methanol. We did not observe any distinct NMR
signals of these species; however, the reaction pathway might be
different under different experimental conditions. Because the
concentration of the starting material 1 was about two orders of
magnitude higher in NMR experiment solutions, the probability of
bimolecular processes is much higher but, at the same time, an
internal optical filter effect is more pronounced.
Keywords: DFT calculations · dyes/pigments · photochemistry ·
reaction mechanisms · spectroscopy
[1] C.-H. Jun, Chem. Soc. Rev. 2004, 33, 610–618.
[2] K. Ruhland, Eur. J. Org. Chem. 2012, 2012, 2683–2706.
[3] P. Klán, J. Wirz, Photochemistry of organic compounds: From concepts to
practice, 1st ed., John Wiley & Sons Ltd.: Chichester 2009.
[4] P. Wan, S. Muralidharan, J. Am. Chem. Soc. 1988, 110, 4336–4345.
[5] P. K. Das, Chem. Rev. 1993, 93, 119–144.
[6] M. Freccero, M. Fagnoni, A. Albini, J. Am. Chem. Soc. 2003, 125, 13182–
13190.
[7] A. T. Buck, C. L. Beck, A. H. Winter, J. Am. Chem. Soc. 2014, 136, 8933–
8940.
Ion at m/z 377 (Figure 3): The ion at m/z 377 was previously
assigned to 9-methoxy pyronin 2,[20] the longest-lived intermediate
observed in the UV-vis spectra. Upon irradiation of 1 in CD3OD and
CH318OH, the signal at m/z 377 mass-shifted to m/z 380 and 379,
respectively (Figure 3). In ethanol, an analogous species at m/z 391
was observed. No incorporation of 18O into the molecule of 2 was
observed when 1 was irradiated in a CH3CN/D218O solution. The
OCD3 group in this intermediate, which was generated in CD3OD,
was exchanged for the OCH3 group after dilution of the solution
[8] H. E. Zimmerman, J. Am. Chem. Soc. 1995, 117, 8988–8991.
[9] P. Klán, T. Šolomek, C. G. Bochet, A. Blanc, R. Givens, M. Rubina, V. Popik,
A. Kostikov, J. Wirz, Chem. Rev. 2012, 113, 119–191.
[10] P. Šebej, J. Wintner, P. Müller, T. Slanina, J. Al Anshori, L. A. P. Antony, P.
Klán, J. Wirz, J. Org. Chem. 2012, 78, 1833–1843.
[11] E. Palao, T. Slanina, L. Muchová, T. Šolomek, L. Vítek, P. Klán, J. Am.
Chem. Soc. 2015, 138, 126–133.
[12] T. Slanina, P. Shrestha, E. Palao, D. Kand, J. A. Peterson, A. S. Dutton, N.
Rubinstein, R. Weinstain, A. H. Winter, P. Klán, J. Am. Chem. Soc. 2017,
139, 15168–15175.
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
12
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/cplu.202000370
ChemPlusChem
[13] J. A. Peterson, C. Wijesooriya, E. J. Gehrmann, K. M. Mahoney, P. P.
Goswami, T. R. Albright, A. Syed, A. S. Dutton, E. A. Smith, A. H. Winter, J.
Am. Chem. Soc. 2018, 140, 7343–7346.
[14] H. Yu, J. Li, D. Wu, Z. Qiu, Y. Zhang, Chem. Soc. Rev. 2010, 39, 464–473.
[15] P. Wang, Asian J. Org. Chem. 2013, 2, 452–464.
[16] A. P. Gorka, R. R. Nani, M. J. Schnermann, Acc. Chem. Res. 2018, 51,
3226–3235.
[17] P. Hu, K. Berning, Y.-W. Lam, I. H.-M. Ng, C.-C. Yeung, M. H.-W. Lam, J.
Org. Chem. 2018, 83, 12998–13010.
[18] X. Wang, J. A. Kalow, Org. Lett. 2018, 20, 1716–1719.
[19] A. Y. Vorobev, A. E. Moskalensky, Comput. Struct. Biotechnol. J. 2020, 18,
27–34.
[20] P. Štacko, P. Šebej, A. T. Veetil, P. Klán, Org. Lett. 2012, 14, 4918–4921.
[21] P. Vath, D. E. Falvey, L. A. Barnhurst, A. G. Kutateladze, J. Org. Chem.
2001, 66, 2887–2890.
[22] W. A. McHale, A. G. Kutateladze, J. Org. Chem. 1998, 63, 9924–9931.
[23] P. G. M. Wuts, Greene’s Protective Groups in Organic Synthesis, John Wiley
& Sons, Inc. 2014.
[24] M. Yus, C. Nájera, F. Foubelo, Tetrahedron 2003, 59, 6147–6212.
[25] T. E. Burghardt, J. Sulfur Chem. 2005, 26, 411–427.
[26] D. K. Stevenson, L. K. Kwong, H. J. Vreman, Clin. Chem. 1984, 30, 1382–
1386.
[27] H. J. Vreman, D. K. Stevenson, Anal. Biochem. 1988, 168, 31–38.
[28] M. Arık, K. Meral, Y. Onganer, J. Lumin. 2009, 129, 599–604.
[29] O. Valdes-Aguilera, D. C. Neckers, Acc. Chem. Res. 2002, 22, 171–177.
[30] G. R. Fleming, A. W. E. Knight, J. M. Morris, R. J. S. Morrison, G. W.
Robinson, J. Am. Chem. Soc. 1977, 99, 4306–4311.
[31] P. Murasecco-Suardi, E. Gassmann, A. M. Braun, E. Oliveros, Helv. Chim.
Acta 1987, 70, 1760–1773.
[32] L. Ludvíková, P. Friš, D. Heger, P. Šebej, J. Wirz, P. Klán, Phys. Chem.
Chem. Phys. 2016, 18, 16266–16273.
[33] J. C. Scaiano, J. Am. Chem. Soc. 1980, 102, 7747–7753.
[34] C. S. Foote, Science 1968, 162, 963–970.
[35] C. S. Foote, ACS Symp. Ser. 1987, 339, 22–38.
[36] K. Gollnick, Adv. Photochem., Vol. 6, John Wiley & Sons 1968, pp. 1–122.
[37] S. Kim, M. Fujitsuka, M. Miyata, T. Majima, Phys. Chem. Chem. Phys. 2016,
18, 2097–2103.
[38] R. P. Sabatini, M. F. Mark, D. J. Mark, M. W. Kryman, J. E. Hill, W. W.
Brennessel, M. R. Detty, R. Eisenberg, D. W. McCamant, Photochem.
Photobiol. Sci. 2016, 15, 1417–1432.
[61] M. Kamata, H. Otogawa, E. Hasegawa, Tetrahedron Lett. 1991, 32, 7421–
7424.
1
2
3
4
5
6
7
8
9
[62] G. Oksdath-Mansilla, V. Hajj, D. M. Andrada, J. E. Argüello, J. Bonin, M.
Robert, A. B. Peñéñory, J. Org. Chem. 2015, 80, 2733–2739.
[63] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H.
Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G.
Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.
Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C.
Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G.
Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D.
Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox,
Gaussian 09, revision D.01; Gaussian, Inc.: Wallingford, CT 2009.
[64] Y. Zhao, D. G. Truhlar, J. Chem. Phys. 2006, 125, 194101.
[65] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297.
[66] S. Grimme, J. Antony, S. Ehrlich, H. Krieg, J. Chem. Phys. 2010, 132,
154104.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[67] H. D. Dakin, J. Am. Chem. Soc. 1909, 42, 477–498.
[68] H. L. J. Bäckström, J. Am. Chem. Soc. 1927, 49, 1460–1472.
[69] M. Sankar, E. Nowicka, E. Carter, D. M. Murphy, D. W. Knight, D. Bethell,
G. J. Hutchings, Nat. Commun. 2014, 5.
[70] J. R. McNesby, C. A. Heller, Chem. Rev. 1954, 54, 325–346.
[71] C. A. McDowell, L. K. Sharples, Can. J. Chem. 1958, 36, 251–257.
[72] M.-Y. Ngai, A. Banerjee, Z. Lei, Synthesis 2018, 51, 303–333.
[73] L. Vanoye, A. Favre-Réguillon, A. Aloui, R. Philippe, C. de Bellefon, RSC
Adv. 2013, 3, 18931.
[74] M. Hajimohammadi, N. Safari, H. Mofakham, A. Shaabani, Tetrahedron
Lett. 2010, 51, 4061–4065.
[75] L. A. P. Antony, T. Slanina, P. Šebej, T. Šolomek, P. Klán, Org. Lett. 2013,
15, 4552–4555.
[76] J. D. Crounse, H. C. Knap, K. B. Ørnsø, S. Jørgensen, F. Paulot, H. G.
Kjaergaard, P. O. Wennberg, J. Phys. Chem. A 2012, 116, 5756–5762.
[77] H. G. Kjaergaard, H. C. Knap, K. B. Ørnsø, S. Jørgensen, J. D. Crounse, F.
Paulot, P. O. Wennberg, J. Phys. Chem. A 2012, 116, 5763–5768.
[78] W.-U. Palm, Photochem. Photobiol. Sci. 2018, 17, 964–974.
[79] A. D. Britt, W. B. Moniz, J. Org. Chem. 1973, 38, 1057–1059.
[80] P. C. Beaumont, D. G. Johnson, B. J. Parsons, J. Photochem. Photobiol. A
1997, 107, 175–183.
[39] V. E. Korobov, V. V. Shubin, A. K. Chibisov, Chem. Phys. Lett. 1977, 45,
498–501.
[40] M. N. Eberlin, Eur. J. Mass Spectrom. 2017, 13, 19–28.
[41] P. Chen, Angew. Chem. Int. Ed. 2003, 42, 2832–2847; Angew. Chem.
2003, 115, 2938–2954.
[42] L. Jašíková, M. Anania, S. Hybelbauerová, J. Roithová, J. Am. Chem. Soc.
2015, 137, 13647–13657.
[43] L. Sleno, D. A. Volmer, J. Mass Spectrom. 2004, 39, 1091–1112.
[44] L. Jašíková, J. Roithová, Chem. Eur. J. 2018, 24, 3374–3390.
[45] J. Roithová, Chem. Soc. Rev. 2012, 41, 547–559.
[46] L. MacAleese, P. Maître, Mass Spectrom. Rev. 2007, 26, 583–605.
[47] J. Roithová, A. Gray, E. Andris, J. Jašík, D. Gerlich, Acc. Chem. Res. 2016,
49, 223–230.
[48] J. Zelenka, J. Roithová, ChemBioChem 2020, (in print).
[49] M. Montalti, A. Credi, L. Prodi, M. T. Gandolfi, Handbook of Photo-
chemistry, 3rd ed., CRC Press Boca Raton 2006.
[50] Y. Onganer, E. L. Quitevis, J. Phys. Chem. 1992, 96, 7996–8001.
[51] X.-F. Zhang, J. Zhang, X. Lu, J. Fluoresc. 2015, 25, 1151–1158.
[52] R. E. Barnett, W. P. Jencks, J. Am. Chem. Soc. 1967, 89, 5963–5964.
[53] R. E. Barnett, W. P. Jencks, J. Am. Chem. Soc. 1969, 91, 6758–6765.
[54] A. Specht, S. Loudwig, L. Peng, M. Goeldner, Tetrahedron Lett. 2002, 43,
8947–8950.
[55] W. Lin, D. S. Lawrence, J. Org. Chem. 2002, 67, 2723–2726.
[56] A. P. Kostikov, V. V. Popik, Org. Lett. 2008, 10, 5277–5280.
[57] M. El Baraka, M. Deumié, P. Viallet, T. J. Lampidis, J. Photochem.
Photobiol. A 1991, 56, 295–311.
[81] T. J. Chozinski, L. A. Gagnon, J. C. Vaughan, FEBS Lett. 2014, 588, 3603–
3612.
[82] S. van de Linde, I. Krstić, T. Prisner, S. Doose, M. Heilemann, M. Sauer,
Photochem. Photobiol. Sci. 2011, 10, 499–506.
[83] H. Görner, Photochem. Photobiol. Sci. 2008, 7, 371.
[84] C. Chatgilialoglu, D. Crich, M. Komatsu, I. Ryu, Chem. Rev. 1999, 99,
1991–2070.
[85] J. Zelenka, R. Cibulka, J. Roithová, Angew. Chem. Int. Ed. 2019, 58,
15412–15420.
[86] S. Chen, M. S. Hossain, F. W. Foss, Org. Lett. 2012, 14, 2806–2809.
[87] Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC
Press, Boca Raton 2007.
[88] J. R. R. Majjigapu, A. N. Kurchan, R. Kottani, T. P. Gustafson, A. G.
Kutateladze, J. Am. Chem. Soc. 2005, 127, 12458–12459.
[89] Note that all IRPD spectra reported in this work contain more bands
than those predicted by the theoretical calculations for individual ions.
This is most likely associated with different conformations of piperidine
groups with respect to the xanthene backbone and with respect to
each other.
[90] E. Bodo, A. Ciavardini, A. Giardini, A. Paladini, S. Piccirillo, F. Rondino, D.
Scuderi, Chem. Phys. 2012, 398, 124–128.
[58] E. Baciocchi, C. Crescenzi, O. Lanzalunga, Tetrahedron 1997, 53, 4469–
4478.
[59] C. J. Regan, D. P. Walton, O. S. Shafaat, D. A. Dougherty, J. Am. Chem.
Soc. 2017, 139, 4729–4736.
[60] M. Bettoni, T. Del Giacco, M. Stradiotto, F. Elisei, J. Org. Chem. 2015, 80,
8001–8008.
Manuscript received: May 12, 2020
Revised manuscript received: July 12, 2020
ChemPlusChem 2020, 85, 1–14
www.chempluschem.org
13
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA
��
These are not the final page numbers!

FULL PAPERS
1
2
3
4
5
6
Dr. M. Martínek, Dr. J. Váňa, Dr. P.
Šebej, Dr. R. Navrátil, Dr. T. Slanina,
Dr. L. Ludvíková, Prof. J. Roithová*,
Prof. P. Klán*
7
8
9
1 – 14
Photochemistry of a 9-Dithianyl-
Pyronin Derivative: A Cornucopia
of Reaction Intermediates Lead to
Common Photoproducts
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
A hunt for intermediates: The photo-
chemistry of a 9-dithianyl-pyronin de-
rivative, which undergoes an
oxidative transformation at the meso-
position to give a 3,6-diamino-9H-
xanthen-9-one derivative, formic acid,
and carbon monoxide as the main
photoproducts, is presented. The
identification of many intermediates
and their interrelationships is
presented and was used to provide a
complete picture of the transforma-
tion that brings new inputs to a
rational design of new photoactivat-
able pyronin or xanthene derivatives.