Mechanistic and Synthetic Investigations on the Dual Selenium-π-Acid/Photoredox Catalysis in the Context of the Aerobic Dehydrogenative Lactonization of Alkenoic Acids

Article abstract of DOI:10.1021/acscatal.7b02729

The aerobic dehydrogenative lactonization of alkenoic acids facilitated by a cooperative nonmetallic catalyst pair is reported. The title procedure relies on the adjusted interplay of a photoredox and a selenium-π-acid catalyst, which allows for the regiocontrolled construction of five- and six-membered lactone rings in yields of up to 96%. Notable features of this method are pronounced efficiency and practicality, good functional group tolerance, and high sustainability, since ambient air and visible light are adequate for the clean conversion of alkenoic acids into their respective lactones. The title method has been used as a case study to elucidate the general mechanistic aspects of the dual selenium-π-acid/photoredox catalysis. On the basis of NMR spectroscopic, mass spectrometric, and computational investigations, a more detailed picture of the catalytic cycle is drawn and the potential role of trimeric selenonium cations as catalytically relevant species is discussed.

Full text of DOI:10.1021/acscatal.7b02729

Research Article
pubs.acs.org/acscatalysis
Mechanistic and Synthetic Investigations on the Dual
Selenium-π-Acid/Photoredox Catalysis in the Context of the Aerobic
Dehydrogenative Lactonization of Alkenoic Acids
Stefan Ortgies,† Rene Rieger,† Katharina Rode,† Konrad Koszinowski,† Jonas Kind,‡
Christina M. Thiele,‡ Julia Rehbein,§ and Alexander Breder*,†
†Institut fur Organische und Biomolekulare Chemie, Georg-August-Universitat, Tammannstrasse 2, 37077 Gottingen, Germany
̈
̈
̈
‡Clemens-Schopf-Institut fur Organische Chemie und Biochemie, Technische Universitat Darmstadt, Alarich-Weiss-Strasse 16,
̈
̈
̈
64287 Darmstadt, Germany
§Organische Chemie, Universitat Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
̈
S
* Supporting Information
ABSTRACT: The aerobic dehydrogenative lactonization of
alkenoic acids facilitated by a cooperative nonmetallic catalyst
pair is reported. The title procedure relies on the adjusted
interplay of a photoredox and a selenium-π-acid catalyst, which
allows for the regiocontrolled construction of five- and six-
membered lactone rings in yields of up to 96%. Notable fea-
tures of this method are pronounced efficiency and practi-
cality, good functional group tolerance, and high sustainability,
since ambient air and visible light are adequate for the clean
conversion of alkenoic acids into their respective lactones.
The title method has been used as a case study to elucidate
the general mechanistic aspects of the dual selenium-π-acid/
photoredox catalysis. On the basis of NMR spectroscopic,
mass spectrometric, and computational investigations, a more detailed picture of the catalytic cycle is drawn and the potential
role of trimeric selenonium cations as catalytically relevant species is discussed.
KEYWORDS: π-acid catalysis, selenium, alkenoic acids, oxidation, reaction mechanism
protocol that operates with air as a terminal and gratuitous
oxidant.8,9 This procedure was realized by the cooperative inter-
play of a diselane and a photoredox catalyst and allowed for the
oxidative coupling of alkenes with a broad variety of carboxylic
acids under aerial conditions. Cognate selenium-π-acid catalysis
methods have so far exclusively involved the use of customary
oxidants, such as iodine(III) reagents,10 N-haloimides,11 and
N-fluorobenzenesulfonimide (NFSI).12 On the basis of our
initial results, we became interested in two main questions.
(a) How does the dual selenium-π-acid/photoredox catalysis
regime proceed mechanistically? (b) Is it possible to efficiently
implement this catalysis manifold in the dehydrogenative aerial
cyclization of simple alkenoic acids as a new means to construct
lactone architectures?11,12a,13,14 Such an approach would ele-
gantly combine the redox-economical15 advantages of tradi-
tional oxidative lactonization reactions with the nonpollut-
ing chemical profile of air as a terminal oxidant (Scheme 1).
As a result of these considerations, we report herein the first
INTRODUCTION
■
The utilization of molecular oxygen as a terminal oxidant in
catalytic transformations is playing an increasingly important
role in modern chemical research.1 Throughout the last two
decades, the most far-reaching progress in the context of cata-
lytic redox functionalizations in which oxygen serves solely as
an oxidant without incorporating any of its O atoms into the
productshereafter termed abiotic oxidase reactionshas been
achieved with transition-metal catalysts.1,2 In particular, Pd and
Cu complexes have experienced widespread application in a
broad panoply of O2-reliant, oxidative methodologies, such as
alkene functionalizations, amine and alcohol oxidations, and
dehydrogenative coupling reactions of arenes.1,3 Although spo-
radic exceptions have been documented in the literature,4 the
majority of abiotic oxidase protocols operate at high oxygen
concentrations (e.g., neat O2 atmosphere). In certain cases it
has been even shown that the use of air as an oxidant instead of
neat O2 leads to significantly diminished product yields.5 Thus,
the design of new abiotic oxidase catalysis regimes that effi-
ciently operate under a gas atmosphere with reduced oxygen
concentrations (≤21 vol %) are highly sought after.6 Recently,
we have disclosed the first selenium-π-acid/photoredox catalysis7
Received: August 12, 2017
Revised: September 17, 2017
© XXXX American Chemical Society
7578
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
Scheme 1. Selenium-Catalyzed Oxidative Lactonizations of
Alkenoic Acids Facilitated by Classical Oxidants (Previous
Work) and Air (This Work)
Table 1. Optimization of the Se-Catalyzed C(sp2)−H
a
Lactonization
entry
loading x/y
solvent
atmosphere
yield (%)
1
10/5
10/5
10/5
10/5
10/5
5/5
1,4-dioxane
THF
air
air
air
air
air
air
air
air
air
air
air
air
air
argon
0
>5
63
95
>95
>95
80
50
64
52
0
2
3
acetone
DCE
4
5
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
6
7
4/5
8
3/5
9
5/2.5
5/0.3
0/5
10
11
12
13
14
5/0
0
b
5/5
0
5/5
<5
a
Yields were determined by 1H NMR spectroscopy using 1,3,5-
trimethoxybenzene or 1,4-dimethoxybenzene as an internal standard.
b
DCE = 1,2-dichloroethane. Reaction was performed in the dark.
yields, respectively. We rationalized this outcome with an insuf-
ficient solubility of photocatalyst 3 in these solvents. Changing
to acetone as a reaction solvent resulted in a markedly improved
yield of 63% (entry 3). Even better results were obtained when
dichloroethane (95%, entry 4) and MeCN (>95%, entry 5)
were used as solvents. Subsequently, we investigated the effect
of reduced catalyst loadings (entries 6−12). In this series of
experiments, we found that the loading of (PhSe)2 could be
lowered to 5 mol % while still maintaining full conversion of
substrate 1a (entry 6). Smaller amounts of the diselane cata-
lyst led to incomplete conversions (entries 7 and 8). A similar
detrimental effect on the yield was observed when loadings of
less than 5 mol % of pyrylium salt 3 were applied (entries 9 and
10). In the absence of either catalyst no product formation
was observed (entries 11 and 12). Furthermore, no or insig-
nificant turnover was recorded when the title transformation
was attempted in the dark or under an atmosphere of
argon instead of air (entries 13 and 14). These prelim-
inary findings clearly indicate that the reaction mechanism is
most likely of cooperative nature and that air serves as an
indispensable reagent in these transformations. Upon identi-
fication of optimal reaction conditions (cf. Table 1, entry 6),
our investigations continued with the exploration of the scope
and limitations of the title procedure (Tables 2 and 3). In the
first part of this endeavor a series of β- and γ-alkenoic acids
1a−o, possessing di- and trisubstituted olefin entities, was con-
verted into the corresponding set of (dihydro)furanones 2a−o
via 5-endo- and 5-exo-trig cyclization, respectively (Table 2).
Lactones containing n-alkyl side chains in the 5-position of the
furanone core were obtained in yields ranging between 66 and
90% (entries 1−4). Products possessing benzylic substituents in
the same position were isolated in 51−67% yield (entries 5−7).
A useful feature of this cooperative selenium-π-acid/photo-
redox catalysis protocol is its pronounced tolerance of protic
functional groups, such as an additional carboxylic acid entity
(entry 8, 96% yield) or a tertiary alcohol (entry 9, 67% yield).
selenium-π-acid catalyzed dehydrogenative lactonization proto-
col under aerial conditions enabled by a photoactive pyrylium
catalyst in the context of mechanistic and computational inves-
tigations to elucidate the operative catalytic cycle.
RESULTS AND DISCUSSION
■
Our investigations commenced with the identification of opti-
mal reaction conditions for the desired oxidative ring closure.
In previous studies,8 we have demonstrated that 2,4,6-tris(4-
methoxyphenyl)pyrylium tetrafluoroborate (3) is particularly
suitable as a cocatalyst for oxidative esterification reactions
of alkenes due to its proper excited state redox potential of
+1.74 V vs SCE in MeCN.7,16 Such a value is just high enough
to allow for the oxidation of electron-neutral diaryl diselanes
17
red
(e.g., (PhSe)2; E1/2 = +1.35 V vs SCE in MeCN) by pyry-
lium salt 3 under photocatalytic conditions but at the same time
low enough to leave electron-neutral, nonconjugated alkenes
(E1/2 ≥ +1.80 V vs SCE in MeCN) unaffected.18 Thus, in
red
initial experiments (E)-hex-3-enoic acid (1a) was used as a
test substrate to identify appropriate reaction parameters for
its direct conversion into 5-ethylbutenolide (2a) (Table 1).
Initially, acid 1a was exposed to 10 mol % of (PhSe)2 and
5 mol % of pyrylium salt 3 in various solvents (0.1 M) at room
temperature under an atmosphere of air and irradiation at
465 nm (entries 1−5). When the reaction was performed in
ethereal solvents, such as 1,4-dioxane and THF (entries 1 and
2), formation of product 2a occurred in 0 and less than 5%
7579
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
Table 2. Scope of the Aerial Oxidative Lactonization of β
(in Red)- and γ-Alkenoic Acids (in Blue) 1a−o Through
5-endo-trig (in Red) and 5-exo-trig Cyclizations (in Blue),
Table 3. Scope of the Aerial Oxidative Lactonization of
δ- and ε-Alkenoic Acids 1p−z through 6- and 7-exo-trig
a
Cyclizations, Respectively
a
Respectively
a
Reaction conditions: alkenoic acids (1.0 mmol), (PhSe)2 (5 mol %),
pyrylium salt 3 (5 mol %), MeCN (0.1 M), 16−24 h. Yields cor-
respond to isolated products.
δ,ε-unsaturated carboxylic acids 1p−y via 6-exo-trig cyclization
(Table 3). Both simple hydrocarbon residues and function-
alized alkyl groups generally led to reasonable or good yields
ranging from 69 to 78% (entries 1−3). As was observed in the
experiments summarized in Table 2, the title method tolerates
substrates possessing protic (entry 2, 71%) or acid-labile (entry 3,
78%) functional groups, thus demonstrating the mildness
of the dual selenium-π-acid/photoredox catalysis. Alkenoic
acids containing substituted aromatic entities, such as 4-fluoro,
4-cyano, 4-trifluoromethyl, and 4-tert-butylphenyl groups,
generally led to good isolated yields of the respective lactones
2t−w (entries 5−8, 73−83%). The sterically more demanding
1-naphthyl group of alkenoic acid 1x led to a slightly diminished
yield of 57% (entry 9). The detrimental effect of steric encum-
brance was even more pronounced when acid 1y was subjected
to the title conditions (entry 10). The resulting lactone 2y,
which possesses a trisubstituted C−C double bond, was iso-
lated in merely 33% yield. Finally, we wanted to determine
whether cyclizations leading to ring sizes larger than six chain
members are feasible via the title procedure. Therefore, we
subjected petroselinic acid (1z) to the standard reaction condi-
tions. However, the corresponding oxepan-2-one derivative 2z
was isolated only in trace amounts (entry 11, 5%).
a
Reaction conditions (unless indicated otherwise): alkenoic acid
(1.0 mmol), (PhSe)2 (5 mol %), pyrylium salt 3 (5 mol %), MeCN
(0.1 M), 16−24 h. Yields correspond to isolated products. Catalyst
loading (PhSe)2 (10 mol %), reaction time 72 h. Reaction time 40 h.
b
c
In the case of substrates with an alkene moiety surrounded by a
sterically demanding periphery, we observed mixed results in
terms of isolated yield. While lactone 2l (R1 = i-Pr) was
obtained in a yield of 73% (entry 12), its phenyl-substituted
analogue 2m was obtained in a significantly lowered yield of
only 32% (entry 13). This finding is congruent with an observa-
tion we had made in a previous study on intermolecular oxida-
tive esterifications,8 in which the nonaromatic C−C double
bond of a styrene unit was found to be less reactive than regular
alkenes. The effect of sterics was even more dominant in the
case of a trisubstituted olefin (entry 14). Conversion of sub-
strate 1n required an increased loading of the selenium catalyst
(10 mol %) and an extended reaction time of 72 h in order to
reach an isolated yield of 43%. Formation of bicyclic lactone
2o was accomplished in the course of 40 h with a moderate
yield of 40% (entry 12). Next, we investigated the synthesis of
six-membered lactones 2p−y derived from the corresponding
An important aspect associated with the use of either air or
pure O2 as a reagent in abiotic oxidase reactions is the potential
hazard of spontaneous solvent combustion caused by molec-
ular oxygen in medium to high concentrations (21−100 vol %).
7580
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
As was recently demonstrated by Stahl and co-workers,6 such
risks can be eliminated if the corresponding abiotic oxidase
reactions are performed at oxygen concentrations below solvent-
dependent threshold levels (i.e., LOC values; LOC = limiting
oxygen concentration). Thus, we wanted to evaluate whether
the dual selenium-π-acid/photoredox catalysis will still lead to
useful results when it is conducted at very low oxygen con-
centrations. Therefore, we subjected alkenoic acid 1a to the
standard title procedure using an atmosphere of 6 vol % of O2
in argon (eq 1). Under these modified conditions, lactone 2a
by the detection of an 18O-labeled dichlorophosphate anion
(Table 4, entry 3), which we believe is derived from its con-
jugated acid. (b) No 18O incorporation was recorded for the
lactone product 2a, thus suggesting that the dual selenium-π-
acid/photoredox catalysis regime generally emulates the reac-
tivity profile of oxidases and not that of oxygenases (i.e., incor-
poration of at least one O2-derived oxygen atom into the
product). In combination with the fact that no H2O2 was
detected at any instance throughout the reaction, it stands to
reason that molecular oxygen serves as a four-electron acceptor
in this dual catalytic manifold. It remains unclear at this point,
however, whether H2O2 is formed in very small quantities and
directly reduced to water within the catalytic cycle or whether
an off-cycle disproportionation of transiently formed hydrogen
superoxide (HO2) is taking place.21
Analysis of Reaction Intermediates and the Kinetic
Profile. On the basis of interval irradiation experiments pre-
viously performed in our group, we hypothesized that the
selenium-π-acid catalysis presumably does not proceed via a
radical chain mechanism.8a To substantiate this conjecture, we
determined the quantum yield of the title reaction by chemical
actinometry.22 As a result of these experiments, a quantum yield
of 14% was recorded, an outcome that is congruent with our
initial postulate that a long-lived radical chain mechanism is
most likely not operative.
was isolated in 90% yield after 16 h. This result clearly demon-
strates that our Se-reliant abiotic oxidase catalysis regime performs
potently in the dehydrogenative cyclization of alkenoic acids
even at very low O2 concentrations.
Mechanistic Investigations: The Role of O2. Due to the
efficiency and robustness of the oxidative lactonization of
alkenoic acids, we decided to use this title procedure as a repre-
sentative case study to investigate the general mechanistic fea-
tures of the dual selenium-π-acid/photoredox catalysis. At the
outset of this endeavor, we wanted to determine whether
oxygen serves as a two- or four-electron acceptor. In the former
case, H2O2 would be formed in stoichiometric quantities, while
in the latter case, 1 equiv of H2O would be generated as a
byproduct. Consequently, we initially performed qualitative
peroxide tests of several reaction solutions at various stages of
reaction progress. At no instance were we able to detect any
quantities of H2O2.19 Therefore, the reaction mixture was
analyzed by ESI mass spectrometry for the detection of water
that had unambiguously arisen from O2. Since H2O itself is
difficult to detect by mass spectrometry, we had to devise an
indirect proof for the formation of water. Consequently, we
initially added 1 equiv of H218O to solutions of different acid
chlorides in order to test whether they could be readily con-
verted into the corresponding 18O-labeled acids. To this end,
POCl3 proved to be very suitable and smoothly under-
went monohydrolysis to give HPO(18O)Cl2 under both argon
and 16O2 atmospheres (Table 4, entries 1 and 2, respectively).
In the next step, the kinetic profiles of substrate 1a (blue
circles), product 2a (black circles), and selenolactonization
1
intermediate 4a (green circles) were monitored by H NMR
spectroscopy by drawing samples from the reaction mixture
at the respective time points (Figure 1).23 Under standard
conditions (cf. Table 2), an initiation phase (approximately
90 min) for the formation of lactone 2a was observed, while the
diselane catalyst and alkenoic acid 1a were rapidly converted
into intermediate 4a until the former was fully consumed. Fur-
thermore, the rate of product formation reached its maximum
at the point when the concentration of the diselane approached
0 M (180 min), whereas the rate of substrate consumption
virtually remained invariant until 90% of conversion (480 min).
In order to find a plausible explanation for the observed initia-
tion phase, a series of control experiments was performed
(Table 5). In the first experiment, selenolactone 4a was
exposed to 5 mol % of both pyrylium salt 3 and (PhSe)2. Under
these conditions, no product formation was recorded in the
course of 16 h (entry 1). An equal outcome was observed when
the same transformation was attempted in the absence of
photocatalyst 3 (entry 2). However, selenolactonization inter-
mediate 4a was readily converted into product 2a (74%) in
the sole presence of photocatalyst 3 (entry 3). Furthermore,
when compound 4a was used in substoichiometric amounts
(10 mol %) relative to alkenoic acid 1a and in the presence of
5 mol % of pyrylium salt 3 (entry 4), an overall yield of 89%
was achieved. The combination of these findings strongly
suggests that (PhSe)2 exerts a pronounced inhibitory effect on
the conversion of intermediate 4a into product 2a. A plausible
rationale for this adverse influence of (PhSe)2 could be a higher
Table 4. 18O Incorporation (%) upon Hydrolysis of POCl3 in
a b
,
CH3CN
18O incorp 18O incorp
av 18O
incorp
entry
conditions
[P]−
[P]2H−
1
2
3
H218O/Ar
94
84
94
84
73
94
84
73
H218O/16O2
1a, Ph2Se2, 3+ 465 nm/18O2
72
a
b
Determined by ESI mass spectrometry. [P] = PO(18O)Cl2.
With negative ion mode ESI mass spectrometry the result-
ing anions PO(18O)Cl2− and (PO(18O)Cl2)2H− were detected,
which showed the predominant incorporation of the 18O
label.20 Subsequently, hex-3-enoic acid (1a) was oxidatively
lactonized under standard conditions using an 18O2 atmosphere
(Table 4, entry 3; eq 2). Upon completion of the reaction and
treatment of the crude mixture with 0.1 equiv of POCl3, mass
spectrometric analysis of the resulting solution revealed two key
findings. (a) The proposed formation of water could be verified
7581
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
varied between 2.5 and 20 mol % at a constant light intensity of
1900 μW/cm2. The rate dependence of species 4a was deter-
mined to be 0.63 0.01. Similarly, variation of the light inten-
sity from 1900 to 4223 μW/cm2 at a constant concentration of
compound 4a revealed a rate order of approximately 1. Fur-
thermore, no significant decrease of the relative 1H NMR signal
intensities of photocatalyst 3 could be detected during irradia-
tion. This observation supports the hypothesis that the regen-
eration of photocatalyst 3 by oxygen occurs rapidly enough
to avoid the accumulation of the reduced form of catalyst 3
to detectable quantities. From the sum of these findings we
conclude that both intermediate 4a and photocatalyst 3 are
involved in the rate-limiting step. This scenario can be ratio-
nalized by a mechanism in which a slow SET between inter-
mediate 4a and photocatalyst 3 is followed by a fast fragmen-
tation of the resulting radical cation 4a•+ to eventually result in
the formation of product 2a (Scheme 2). At this point, it was
Scheme 2. Mechanistic Hypotheses on the Conversion of
Intermediate 4a into Product 2a through Direct Elimination
(Scenario a) or a Radical Recombination Mechanism
(Scenario b)
Figure 1. Kinetic profiles of hex-3-enoic acid (1a), (PhSe)2,
selenenylated lactone 4a, and product 2a. Data points were measured
with 1H NMR spectroscopy of samples drawn from the reaction
mixture at the respective time points.
Table 5. Mechanistic Control Experiments To Determine
a
the Catalytic Function of Intermediate 4a
not clear from the NMR experiments whether the proposed
fragmentation would occur spontaneously via deprotodesele-
nenylation of 4a•+ to give product 2a and PhSe• (Scheme 2,
scenario a) or through dimerization of 4a•+ to transiently give
dication [4a-4a]2+ followed by sequential elimination and
formation of product 2a and (PhSe)2 (scenario b).26
entry (PhSe)2 (mol %) photocat. 3 (mol %) acid 1a (equiv) yield (%)
1
2
3
4
5
5
0
0
5
0
5
5
0
0
0
0
0
74
89
b
To find reliable evidence for either of the two proposed
scenarios, we devised experimental conditions that would allow
for the direct detection of cationic intermediates resulting from
a single-electron oxidation of selenides of type 4 by mass
spectrometry. More specifically, we used selenenylated lactone
10
a
All control experiments were performed on a 1 mmol scale. Yields
b
refer to isolated compounds. Alkenoic acid 1a was used as the
limiting reagent (1 mmol), and lactone 4a was used in 10 mol %. The
denoted yield refers to the maximally achievable amount of product
(i.e, 1.1 mmol).
27
4a(o‑anisyl) as our model substrate and treated it with nitrosyl
tetrafluoroborate, which served as a single-electron oxidant to
emulate the role of the photocatalyst. Solutions of these two
reactants were subsequently analyzed by ESI-MS (Figure 2).
In these experiments, we were able to detect the cationic dis-
elenonium species Ib and II. We speculate that cation Ib is
derived from the reaction of dication Ia (i.e., the analogue of
[4a-4a]2+; Scheme 2) with residual water present in the sample
or the mass spectrometer. Collision-induced dissociation of
mass-selected species Ib revealed that this cation fragments to
directly give cation II by loss of 1 equiv of product 2a and
water. Although these results do not disprove the possibility of a
direct elimination of PhSe• from intermediate 4a•+ (scenario a),
they strongly suggest that the formation of dication Ia followed
by sequential elimination of 2 equiv of product 2a under regular
catalytic conditions is a feasible process. Thus, the detected
intermediates Ib and II are in full agreement with our mech-
anistic hypothesis b depicted in Scheme 2. Moreover, we
detected a trimeric selenonium ion [(o-anisyl-Se)3]+ (Figure 2).
rate of single-electron transfer between excited pyrylium salt 3
and the diselane in comparison with the cognate SET process
between catalyst 3 and intermediate 4a. Moreover, the reaction
outcome depicted in Table 5, entry 4, corroborates the hypo-
thesis that compound 4a serves as an integral constituent of the
operative catalytic cycle.
From the kinetic profile depicted in Figure 1 we furthermore
concluded that the rate-determining step must take place after
the formation of intermediate 4a due to its accumulation at the
beginning of the reaction. Consequently, its generation must
occur very rapidly from a selenenylating source in low concen-
trations (either (PhSe)2 or alternative selenium intermediates
not visible by 1H NMR). Initial-rate experiments executed with
in situ irradiation NMR24 under conditions analogous to those
described in Table 5 (entry 4) were used to investigate the rate
dependence of the concentration of intermediate 4a and the
light intensity.25 Accordingly, the initial concentration of 4a was
7582
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
Figure 2. Positive ion mode ESI mass spectrum of the products formed upon the reaction of the lactonization intermediate 4a(o‑anisyl) (10 mM) with
[NO][BF4] in dichloromethane (Ar = o-anisyl, R = 5-ethyltetrahydrofuran-2-on-4-yl). The insets show the measured (black) and simulated (blue)
isotope patterns of the observed ions.
We surmised that this species is derived from the oxidation of
the neutral diselane (o-anisyl-Se)2. The formation of structur-
ally related trimeric organochalcogen cations by means of
single-electron oxidations of disulfides and diselanes had been
reported previously, inter alia, by Geiger et al. and Laitinen
et al, respectively.28,29 To further substantiate our hypothesis
that a similar process is operative in our case as well, we
added [NO][BF4] to a solution of (o-anisyl-Se)2 in dichloro-
methane and analyzed the reaction mixture by ESI-MS.
Under these conditions, we again detected the same trimeric
cation [(o-anisyl-Se)3]+ along with other oligomeric selenonium
species. These findings are in good accordance with literature
reports on the tendency of selenium species to form cationic
chains and rings.30 In conclusion, the results of these ESI-MS
experiments suggest that (a) diselanes are most likely
regenerated under oxidative conditions from intermediates
of type 4, potentially through the mechanism illustrated in
exposed to [NO][BF4] under otherwise unaltered conditions,
which led to the clean formation of product 2a and (o-anisyl-
Se)2 in 98 and 26% yields, respectively (eq 4). These results
firmly indicate that conditions under which triselenonium
cations are formed lead to the same set of intermediates and
products that was formed in the case of the photocatalytic
reaction. It therefore stands to reason that the photocatalytic
process may also proceed through the intermediacy of trimeric
selenonium cations, although further experimentation will be
needed to fully confirm or confute this hypothesis.
In order to find theoretical arguments that may support the
existence of triselenonium ions, we compared the relative sta-
bilities of the monomeric selenenium cation (PhSe+) and its
+
trimeric analogue (Ph3Se3 ) by computational methods. In so
doing, we wanted to estimate whether one of these species is
energetically strongly favored and thus more likely to be formed
and involved in the catalytic cycle. Computations were run with
the Gaussian 09.D01 suite of programs32 using the M06-2X/6-
311+G** level of theory.33,34 As can be taken from Scheme 3
(top), the monomeric selenenium cation is 98.2 kcal/mol
higher in energy in comparison with its trimeric analogue.
The same trend was observed when we compared the ener-
+
Scheme 2 (scenario b), and (b) triselenonium species (ArSe)3
and related polyselenonium cations may likely be operative
intermediates in the catalytic cycle. It is important to point out
that, in contrast to sporadic reports on olefin functionalizations
mediated by trisulfonium cations [(ArS)3]+,28,31 cognate tri-
selenonium cations have so far not been generally considered as
catalytically active intermediates in analogous SET oxidation
processes on the basis of experimental results but rather the corre-
sponding monomeric selenenium electrophiles (RSeY).9a,42
To validate the hypothesis that triselenonium cations may
lead to the very same intermediates and products that were
observed under photocatalytic conditions (i.e., intermediates 4
and products 2; cf. Figure 1), we performed two independent
stoichiometric control experiments in which first hex-3-enoic
acid (1a) was reacted with 0.5 equiv of (o-anisyl-Se)2 and
1.0 equiv of [NO][BF4] as the terminal oxidant (eq 3). As a
result, lactone 4a(o‑anisyl) was obtained in an isolated yield of
86%. In a second experiment, lactone 4a(o‑anisyl) was again
−
gies of the cationic Se species in the presence of a BF4 coun-
terion (ΔRESCF = −18.1 kcal/mol; Scheme 3, bottom) or with
an implicit solvent model taken into account (PCM,35 aceto-
nitrile, ΔRESCF = −69.9 kcal/mol). From these calculations, we
conclude that monomeric selenenium cations (PhSe+) are pre-
sumably too high in energy to be formed under the title
Scheme 3. Computational Comparison of the Relative
Energies of PhSe+ and [(PhSe)3]+ in the Absence and
−
Presence of a BF4 Counterion
7583
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
Scheme 4. Mechanistic Hypothesis for the Cooperative Selenium-π-Acid/Photoredox-Catalyzed Oxidative Lactonization of
Alkenoic Acids 1
conditions. In analogy to reports by Geiger et al. and Laitinen et al.
on the synthesis of trimeric chalcogenonium ions (chalcogen =
S, Se) via single-electron oxidations,28,29 we believe that the
formation of oligomeric selenonium ions (PhnSen+; n ≥ 3) under
the conditions delineated in Tables 2 and 3 is a very likely
scenario. Furthermore, we were able to show that the transfer of
a PhSe residue from these trimeric selenonium ions onto an
enoic acid substrate is thermodynamically rather facile (ΔG =
+2.2 kcal/mol, Scheme 3 (middle)). It remains unclear at this
point, however, to what extent oligomeric selenonium ions may
play a significant role in related catalytic olefin oxidations in the
presence of single-electron oxidants that release strongly coordi-
nating counterions. In view of reports, inter alia, by Poleschner
and Seppelt36 as well as Fujihara et al.,37,38 who unanimously
suggest that isolated selenenium ions (ArSe+) usually require
strongly Lewis basic neighboring groups, it seems plausible to
suspect that oligomeric selenonium ions may be involved at least
in those catalytic regimes in which firmly coordinating counter-
ions or Lewis basic additives or solvents are absent.
visible light. This sequence of concatenary electron-transfer
processes from the alkene to O2 eventually completes the entire
catalytic cycle.
CONCLUSIONS
■
In summary, we have disclosed a novel selenium-π-acid based
abiotic oxidase catalysis protocol for the expedient and econom-
ically sound synthesis of 2-furanone and 2-pyranone derivatives
through oxidative lactonization of the respective β-, γ-, and
δ-alkenoic acids. The title procedure is characterized by very
mild reaction conditions, good to high yields of up to 96%,
excellent catalytic performance even at reduced oxygen concen-
trations of about 6 vol %, and the general use of ambient air as a
gratuitous and ecologically compatible oxidant. Moreover, on
the basis of mass spectrometric investigations we were able to
show that O2 serves as a four-electron acceptor and that water
is formed as the only stoichiometric byproduct in the title
procedure. Ex situ and in situ irradiadiation NMR experiments
were used to identify a slow SET between intermediate 4 and
excited photocatalyst 3+* as the rate-limiting step in the cata-
lytic cycle. Additional mechanistic studies by both ESI-MS and
computational methods have furthermore revealed that trimeric
selenonium cations are likely to play a crucial role as inter-
mediates within the catalytic cycle. Given the pronounced
efficiency and practicality of the title method and the simplicity
by which the reaction can be conducted, we believe that the
cooperative selenium-π-acid/photoredox catalyzed oxidative
lactonization serves as a valuable complement to other contem-
poraneous abiotic oxidase strategies. The further development
of the title method into an enantioselective process and its
implementation into a macrolactonization strategy are currently
ongoing in our laboratories.
On the basis of all mechanistic investigations described
above, we propose the following catalytic scenario (Scheme 4).
Upon single-electron transfer from the diselane catalyst to the
•+
excited photocatalyst 3+*, the resulting radical cation (R2Se2
)
is then converted into trimeric selenonium cation 5.29,39−41
Species 5 subsequently transfers an arylselenenium group onto
alkenoic acid 1. The resulting seleniranium ion 6 undergoes
deprotonation and cyclization to give rise to selenenylated
intermediate 4. Compound 4 then undergoes a second and rate-
determining single-electron-transfer process, transiently leading
to radical cation 4•+. A sequence consisting of the dimerization of
species 4•+ to give dication Ia, followed by sequential elimina-
tion of lactone 2, completes the selenocatalytic cycle. Simulta-
neously, the reduced photocatalyst 3• is oxidized by O2 to
−
generate pyrylium ion 3+ and a superoxide (O2 ) counterion.
ASSOCIATED CONTENT
* Supporting Information
■
S
While the latter ion is protonated and presumably undergo-
ing further conversion to give H2O and O2,21 pyrylium salt 3+
reenters the photocatalytic regime through irradiation with
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.7b02729.
7584
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Experimental procedures, full spectroscopic and MS data
Research Article
(8) (a) Ortgies, S.; Depken, C.; Breder, A. Org. Lett. 2016, 18, 2856−
2859. (b) Leisering, S.; Riano, I.; Depken, C.; Gross, L.; Weber, M.;
Lentz, D.; Zimmer, R.; Stark, C. B. W.; Breder, A.; Christmann, M.
Org. Lett. 2017, 19, 1478−1481.
for all new compounds, kinetic and computational data,
and NMR spectra (PDF)
(9) For seminal reports on stoichiometric selenenylation and
deselenenylation reactions of alkenes using O2 as the terminal oxidant,
see: (a) Pandey, G.; Soma Sekhar, B. B. V. J. Org. Chem. 1992, 57,
4019−4023. (b) Pandey, G.; Soma Sekhar, B. B. V. J. Org. Chem. 1994,
59, 7367−7372. (c) Pandey, G.; Gadre, S. R. Acc. Chem. Res. 2004, 37,
201−210. For a photosensitized selenenylation of alkenes using
NaNO2 as an oxidant, see: Bushan, K. M.; Gopal, V. R.; Reddy, A. M.;
Rao, V. J. J. Photochem. Photobiol., A 1998, 117, 99−104.
(10) (a) Shahzad, S. A.; Venin, C.; Wirth, T. Eur. J. Org. Chem. 2010,
2010, 3465−3472. (b) Singh, F. V.; Wirth, T. Org. Lett. 2011, 13,
6504−6507.
AUTHOR INFORMATION
Corresponding Author
*E-mail for A.B.: abreder@gwdg.de.
ORCID
Konrad Koszinowski: 0000-0001-7352-5789
Christina M. Thiele: 0000-0001-7876-536X
Alexander Breder: 0000-0003-4899-5919
■
Notes
The authors declare no competing financial interest.
(11) (a) Mellegaard, S. R.; Tunge, J. A. J. Org. Chem. 2004, 69,
8979−8981. (b) Mellegaard-Waetzig, S. R.; Wang, C.; Tunge, J. A.
Tetrahedron 2006, 62, 7191−7198.
ACKNOWLEDGMENTS
■
(12) (a) Trenner, J.; Depken, C.; Weber, T.; Breder, A. Angew.
This work was financially supported by the German Research
Council (DFG, Emmy Noether Fellowship to A.B. (BR 4907/
1-1) and J.R. (RE 3630)), the Fonds der Chemischen Industrie
(FCI, Chemiefonds Fellowship to S.O.), and the state of Lower
Saxony (Georg-Christoph-Lichtenberg Fellowship to K.R.). We
thank Prof. Dr. Lutz Ackermann (University of Gottingen,
Gottingen, Germany) for the generous donation of solvents
Chem., Int. Ed. 2013, 52, 8952−8956. (b) Kratzschmar, F.; Kaßel, M.;
̈
Delony, D.; Breder, A. Chem. - Eur. J. 2015, 21, 7030−7034.
(c) Ortgies, S.; Breder, A. Org. Lett. 2015, 17, 2748−2751. (d) Deng,
Z.; Wei, J.; Liao, L.; Huang, H.; Zhao, X. Org. Lett. 2015, 17, 1834−
1837. (e) Kawamata, Y.; Hashimoto, T.; Maruoka, K. J. Am. Chem. Soc.
2016, 138, 5206−5209.
̈
̈
(13) For an early representative example of a Pd-catalyzed oxidase
type lactonization, see: (a) Trend, R. M.; Ramtohul, Y. K.; Stoltz, B.
M. J. Am. Chem. Soc. 2005, 127, 17778−17788. For representative
examples of asymmetric halolactonizations, see: (b) Tan, C. K.; Zhou,
L.; Yeung, Y.-Y. Org. Lett. 2011, 13, 2738−2741. (c) Chen, J.; Zhou,
L.; Tan, C. K.; Yeung, Y.-Y. J. Org. Chem. 2012, 77, 999−1009.
(14) For representative examples, see: (a) Back, T. G.; Dyck, B. P.
Chem. Commun. 1996, 22, 2567−2568. (b) Tiecco, M.; Testaferri, L.;
Santi, C.; Tomassini, C.; Marini, F.; Bagnoli, L.; Temperini, A. Chem. -
Eur. J. 2002, 8, 1118−1124. (c) Browne, D. M.; Niyomura, O.; Wirth,
T. Org. Lett. 2007, 9, 3169−3171. (d) Alberto, E. E.; Braga, A. L.;
Detty, M. R. Tetrahedron 2012, 68, 10476−10481.
and technical equipment. We thank Prof. Dr. Franc Meyer
(University of Gottingen, Gottingen, Germany) for the donation
̈
̈
of 18O2. J.K. and C.M.T. thank the Adolf-Messer foundation and
the German Research Council (DFG, TH1115/9-1) for funding.
REFERENCES
■
(1) (a) Ryland, B. L.; Stahl, S. S. Angew. Chem., Int. Ed. 2014, 53,
8824−8838. (b) Wu, Y.; Wang, J.; Mao, F.; Kwong, F. Y. Chem. - Asian
J. 2014, 9, 26−47. (c) Cao, Q.; Dornan, L. M.; Rogan, L.; Hughes, N.
L.; Muldoon, M. J. Chem. Commun. 2014, 50, 4524−4543.
(d) McCann, S. D.; Stahl, S. S. Acc. Chem. Res. 2015, 48, 1756−
1766. (e) Wendlandt, A. E.; Stahl, S. S. Angew. Chem., Int. Ed. 2015,
54, 14638−14658. (f) Liu, C.; Yuan, J.; Gao, M.; Tang, S.; Li, W.; Shi,
R.; Lei, A. Chem. Rev. 2015, 115, 12138−12204.
(2) (a) Stahl, S. S. Angew. Chem., Int. Ed. 2004, 43, 3400−3420.
(b) Stahl, S. S. Science 2005, 309, 1824−1826. (c) McDonald, R. I.;
Liu, G.; Stahl, S. S. Chem. Rev. 2011, 111, 2981−3019. (d) Wendlandt,
A. E.; Suess, A. M.; Stahl, S. S. Angew. Chem., Int. Ed. 2011, 50, 11062−
11087. (e) Campbell, A. N.; Stahl, S. S. Acc. Chem. Res. 2012, 45, 851−
863.
(3) For recent representative examples of chemical oxidase reactions
catalyzed by transition-metal compounds other than Cu and Pd
complexes, see the following. Rh: (a) Fabry, D. C.; Zoller, J.; Raja, S.;
Rueping, M. Angew. Chem., Int. Ed. 2014, 53, 10228−10231. Ru:
(b) Fabry, D. C.; Ronge, M. A.; Zoller, J.; Rueping, M. Angew. Chem.,
Int. Ed. 2015, 54, 2801−2805. Co: (c) Mei, R.; Wang, H.; Warratz, S.;
Macgregor, S. A.; Ackermann, L. Chem. - Eur. J. 2016, 22, 6759−6763.
(d) Bechtoldt, A.; Tirler, C.; Raghuvanshi, K.; Warratz, S.; Kornhaaß,
C.; Ackermann, L. Angew. Chem., Int. Ed. 2016, 55, 264−267.
(4) For representative examples, see: (a) Chen, X.; Hao, X.-S.;
Goodhue, C. E.; Yu, J.-Q. J. Am. Chem. Soc. 2006, 128, 6790−6791.
(b) Gao, Y.; Wang, G.; Chen, L.; Xu, P.; Zhao, Y.; Zhou, Y.; Han, L.-B.
J. Am. Chem. Soc. 2009, 131, 7956−7957. (c) Miyasaka, M.; Hirano,
K.; Satoh, T.; Kowalczyk, R.; Bolm, C.; Miura, M. Org. Lett. 2011, 13,
359−361. (d) Pattillo, C. C.; Strambeanu, I. I.; Calleja, P.; Vermeulen,
N. A.; Mizuno, T.; White, M. C. J. Am. Chem. Soc. 2016, 138, 1265−
1272.
(15) Burns, N. Z.; Baran, P. S.; Hoffmann, R. W. Angew. Chem., Int.
Ed. 2009, 48, 2854−2867.
(16) Martiny, M.; Steckhan, E.; Esch, T. Chem. Ber. 1993, 126,
1671−1682. In this reference, the reduction potential of 3+* was
reported as 1.98 V vs NHE. For a more direct comparison, we have
presented the corresponding reduction potential of 1.74 V vs SCE.
(17) Kunai, A.; Harada, J.; Izumi, J.; Tachihara, H.; Sasaki, K.
Electrochim. Acta 1983, 28, 1361−1366.
(18) (a) Hamilton, D. S.; Nicewicz, D. A. J. Am. Chem. Soc. 2012,
134, 18577−18580. (b) Shono, T. Electroorganic Chemistry as a New
Tool in Organic Synthesis; Springer Verlag: Heidelberg, 1984.
(19) To exclude the possibility that H2O2 might decompose in the
mere presence of the reaction constituents, we prepared control
solutions to which hydrogen peroxide was added. In each of these
experiments, even at low concentrations of hydrogen peroxide, we
were able to detect H2O2, thus indicating that hydrogen peroxide is
not present in measureable amounts during regular catalytic
conditions. For a detailed description of these control experiments,
please refer to the Supporting Information.
(20) For a comprehensive description of the labeling experiments,
please refer to the Supporting Information.
(21) (a) Bielski, B. H. J.; Allen, A. O. J. Phys. Chem. 1977, 81, 1048−
1050. (b) Li, Y.; Zhong, G.; Yu, H.; Wang, H.; Peng, F. Phys. Chem.
Chem. Phys. 2015, 17, 21950−21959.
(22) For this part of the investigations, we followed a procedure
reported by Glorius et al., which represents a modified version of an
earlier report by Melchiorre et al. For details, see: (a) Wozniak, Ł.;
Murphy, J. J.; Melchiorre, P. J. Am. Chem. Soc. 2015, 137, 5678−5681.
(b) Tlahuext-Aca, A.; Hopkinson, M. N.; Sahoo, B.; Glorius, F. Chem.
Sci. 2016, 7, 89−93.
(5) Yoo, K. S.; O’Neill, J.; Sakaguchi, S.; Giles, R.; Lee, J. H.; Jung, K.
W. J. Org. Chem. 2010, 75, 95−101.
(6) Osterberg, P. M.; Niemeier, J. K.; Welch, C. J.; Hawkins, J. M.;
Martinelli, J. R.; Johnson, T. E.; Root, T. W.; Stahl, S. S. Org. Process
Res. Dev. 2015, 19, 1537−1543.
(23) For related mechanistic investigations, see: Denmark, S. E.;
Edwards, M. G. J. Org. Chem. 2006, 71, 7293−7306.
(7) Ortgies, S.; Breder, A. ACS Catal. 2017, 7, 5828−5840.
7585
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586

ACS Catalysis
Research Article
(24) (a) Kuprov, I.; Hore, P. J. J. Magn. Reson. 2004, 171, 171−175.
(b) Feldmeier, C.; Bartling, H.; Riedle, E.; Gschwind, R. M. J. Magn.
Reson. 2013, 232, 39−44. (c) Kind, J.; Kaltschnee, L.; Leyendecker,
M.; Thiele, C. M. Chem. Commun. 2016, 52, 12506−12509.
(25) We chose to use compound 4a as a precatalyst in the initial-rate
experiments, since no initiation phase was observed under these
conditions.
(26) For literature examples of diselenonium dications, see:
(a) Jouikov, V. V.; Fattakhova, D. S. J. Organomet. Chem. 2000, 613,
220−230. (b) Evans, D. H.; Gruhn, N. E.; Jin, J.; Li, B.; Lorance, E.;
Okumura, N.; Macías-Ruvalcaba, N. A.; Zakai, U. I.; Zhang, S.-Z.;
Block, E.; Glass, R. S. J. Org. Chem. 2010, 75, 1997−2009. (c) Zhang,
S.; Wang, X.; Sui, Y.; Wang, X. J. Am. Chem. Soc. 2014, 136, 14666−
14669.
(27) We used the anisyl analogue of compound 4a, since its redox
potential was suitable in order to oxidize 4a(o‑anisyl) with NOBF4. The
requisite diselane for the synthesis of 4a(o‑anisyl) was prepared according
to a literature procedure: Cresswell, A. J.; Eey, S. T.-C.; Denmark, S. E.
Nat. Chem. 2015, 7, 146−152.
(28) Lam, K.; Geiger, W. E. J. Org. Chem. 2013, 78, 8020−8027.
(29) Laitinen, R.; Steudel, R.; Weiss, R. J. Chem. Soc., Dalton Trans.
1986, 1095−1100.
(30) Engesser, T. A.; Lichtenthaler, M. R.; Schleep, M.; Krossing, I.
Chem. Soc. Rev. 2016, 45, 789−899.
(31) Aida, F.; Oyaizu, K. Chem. Lett. 2016, 45, 102−109.
(32) For Gaussian D01, the full citation is provided in the Supporting
Information.
(33) Zhao, Y.; Truhlar, D. G. Theor. Chem. Acc. 2008, 120, 215−241.
(34) (a) McLean, A. D.; Chandler, G. S. J. Chem. Phys. 1980, 72,
5639−5648. (b) Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J.
Chem. Phys. 1980, 72, 650−654.
(35) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999−3094.
(36) Poleschner, H.; Seppelt, K. Chem. - Eur. J. 2004, 10, 6565−6574.
(37) Fujihara, H.; Mima, H.; Furukawa, N. J. Am. Chem. Soc. 1995,
117, 10153−10154.
(38) For selected recent examples, see: (a) Matsumoto, K.; Sanada,
T.; Shimazaki, H.; Shimada, K.; Hagiwara, S.; Fujie, S.; Ashikari, Y.;
Suga, S.; Kashimura, S.; Yoshida, J. Asian J. Org. Chem. 2013, 2, 325−
329. (b) Matsumoto, K.; Fujie, S.; Suga, S.; Nokami, T.; Yoshida, J.
Chem. Commun. 2009, 5448−5450.
(39) A potential pathway by which cation 5 might be formed is
through the rapid recombination of diselenonium radical cations
(Ar2Se2•+) with selenenyl radicals (ArSe•). The latter species were
reported to permanently form under irradiation even with visible
light.40 Furthermore, radical recombinations of Se species were
reported to occur at rates close to diffusion control.41 Alternatively,
a dimerization of Ar2Se2•+ followed by the loss of an ArSe group could
•+
be envisioned. Such processes have been reported for Ar2S2 radical
cations.28 It remains uncertain at this point, however, whether
additional oligomeric selenonium cations ((ArnSen)+; n > 3) are
formed and involved in the catalytic cycle. Respective investigations to
answer this question and to elucidate the mechanism by which trimeric
selenonium intermediates are actually formed are ongoing in our
laboratories and will be reported in due course.
(40) Ji, S.; Cao, W.; Yu, Y.; Xu, H. Angew. Chem. 2014, 126, 6899.
(41) (a) Holm, A. H.; Yusta, L.; Carlqvist, P.; Brinck, T.; Daasbjerg,
K. J. Am. Chem. Soc. 2003, 125, 2148. (b) Andrieux, C. P.; Hapiot, P.;
́
Pinson, J.; Saveant, J.-M. J. Am. Chem. Soc. 1993, 115, 7783.
(42) For a merely speculative discussion on the potential participation
of triselenonium ions in stoichiometric selenofunctionalizations of
alkenes, see ref 9c.
7586
DOI: 10.1021/acscatal.7b02729
ACS Catal. 2017, 7, 7578−7586