Electron transfer-initiated epoxidation and isomerization chain reactions of β-caryophyllene

Article abstract of DOI:10.1002/chem.201404711

The abundant sesquiterpene b-caryophyllene can be epoxidized by molecular oxygen in the absence of any catalyst. In polar aprotic solvents, the reaction proceeds smoothly with epoxide selectivities exceeding 70%. A mechanistic study has been performed and the possible involvement of free radical, spin inversion, and electron transfer mechanisms is evaluated using experimental and computational methods. The experimental data-including a detailed reaction product analysis, studies on reaction parameters, solvent effects, additives and an electrochemical investigation-all support that the spontaneous epoxidation of b-caryophyllene constitutes a rare case of unsensitized electron transfer from an olefin to triplet oxygen under mild conditions (80 8C, 1 bar O2). As initiation of the oxygenation reaction, the formation of a caryophyllene-derived radical cation via electron transfer is proposed. This radical cation reacts with triplet oxygen to a dioxetane via a chain mechanism with chain lengths exceeding 100 under optimized conditions. The dioxetane then acts as an in situ-formed epoxidizing agent. Under nitrogen atmosphere, the presence of a one-electron acceptor leads to the selective isomerization of b -caryophyllene to isocaryophyllene. Observations indicate that this isomerization reaction is a novel and elegant synthetic pathway to isocaryophyllene.

Full text of DOI:10.1002/chem.201404711

DOI: 10.1002/chem.201404711
Full Paper
&
Sesquiterpenes |Hot Paper|
Electron Transfer-Initiated Epoxidation and Isomerization Chain
Reactions of b-Caryophyllene
Bart Steenackers,[a] Nicolꢀ Campagnol,[b] Jan Fransaer,[b] Ive Hermans,[c] and Dirk De Vos*[a]
Abstract: The abundant sesquiterpene b-caryophyllene can
be epoxidized by molecular oxygen in the absence of any
catalyst. In polar aprotic solvents, the reaction proceeds
smoothly with epoxide selectivities exceeding 70%. A mech-
anistic study has been performed and the possible involve-
ment of free radical, spin inversion, and electron transfer
mechanisms is evaluated using experimental and computa-
tional methods. The experimental data—including a detailed
reaction product analysis, studies on reaction parameters,
solvent effects, additives and an electrochemical investiga-
tion—all support that the spontaneous epoxidation of b-car-
yophyllene constitutes a rare case of unsensitized electron
transfer from an olefin to triplet oxygen under mild condi-
tions (808C, 1 bar O2). As initiation of the oxygenation reac-
tion, the formation of a caryophyllene-derived radical cation
via electron transfer is proposed. This radical cation reacts
with triplet oxygen to a dioxetane via a chain mechanism
with chain lengths exceeding 100 under optimized condi-
tions. The dioxetane then acts as an in situ-formed epoxidiz-
ing agent. Under nitrogen atmosphere, the presence of
a one-electron acceptor leads to the selective isomerization
of b-caryophyllene to isocaryophyllene. Observations indi-
cate that this isomerization reaction is a novel and elegant
synthetic pathway to isocaryophyllene.
Introduction
of oxygenated sesquiterpenoid compounds. Caryophyllene
oxide—the corresponding endocyclic epoxide of b-caryophyl-
lene—and other oxygenated derivatives have strong antibacte-
rial and antifungal activities.[2c,5] The reactivity of b-caryophyl-
lene and the selectivity of chemical transformations of b-caryo-
phyllene are often attributed to the strain of its trans-cyclono-
nene ring.[6] The chemistry of b-caryophyllene and caryophyl-
lene oxide and especially their rearrangement to other bi- and
tricyclic structures in (super)acid medium have been the sub-
ject of many studies.[7] However, little information is available
concerning the spontaneous oxidation of b-caryophyllene. A
common observation is that upon oxidative degradation of b-
caryophyllene, levels of caryophyllene oxide rise.[8] In a recent
study by Skçld, neat b-caryophyllene was exposed to air at
room temperature during several months.[9] At full conversion
of b-caryophyllene, caryophyllene oxide was formed with a se-
lectivity of 40%.
b-Caryophyllene (bC) is a very abundant naturally occurring
sesquiterpene. This semivolatile compound possesses a unique
bicyclic structure composed of a fused cyclobutane and trans-
cyclononene ring. It can be found in the essential oils of herbs
and spices such as clove, hop, cannabis, cinnamon, and black
pepper. The abundance and specific reactivity of b-caryophyl-
lene make it the largest sesquiterpenoid contributor to forma-
tion of secondary organic aerosols.[1] The use of b-caryophyl-
lene as a natural anti-inflammatory or antibacterial compound
is being explored,[2] and it is currently employed as an aroma-
tizing agent in foods (e.g. chewing gum can contain up to
500 ppm of b-caryophyllene) and cosmetics. On the European
market, 30% of the available cosmetics contain b-caryophyl-
lene.[3] The total annual consumption of b-caryophyllene
amounts to up to 5 tons.[4]
b-Caryophyllene is not only a valuable natural product in its
own right, but it also serves as intermediate for the synthesis
In our attempt to synthesize oxygenated caryophyllene-de-
rived sesquiterpenoids, the spontaneous oxidation of b-caryo-
phyllene in the presence of molecular oxygen as investigated.
We found that it is epoxidized under mild conditions in the
presence of molecular oxygen in polar aprotic solvents with
a selectivity of up to 70% without the addition of any catalyst
or co-reactant. According to the widely accepted Twigg epoxi-
dation mechanism, epoxidation of olefins involves stoichiomet-
ric coproduction of alkoxy radical-derived products—typically
alcohols—resulting in a maximal theoretical epoxide selectivity
of 50%.[10] Furthermore, we found that the reaction of b-caryo-
phyllene in the presence of dioxygen can be steered towards
isomerization to isocaryophyllene instead of oxygenation by
the addition of certain types of initiators and/or altering the re-
[a] B. Steenackers, Prof. D. De Vos
Centre for Surface Chemistry and Catalysis, KU Leuven
Kasteelpark Arenberg 23, 3001 Leuven (Belgium)
Fax: (+32)6321998
E-mail: dirk.devos@biw.kuleuven.be
[b] N. Campagnol, Prof. J. Fransaer
Department of Materials Engineering, KU Leuven
Kasteelpark Arenberg 23, 3001 Leuven (Belgium)
[c] Prof. I. Hermans
Department of Chemistry & Department of
Chemical and Biological Engineering
The University of Wisconsin-Madison
1101 University Avenue, 53706 Madison WI (USA)
Chem. Eur. J. 2015, 21, 2146 – 2156
2146
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
action conditions. Given the high selectivity of the reaction of
b-caryophyllene in the presence of molecular oxygen, we at-
tempted to elucidate the mechanism that lies at the base of
these unusual observations. With both experimental and theo-
retical means, the possible involvement of a free-radical mech-
anism, a spin-forbidden pathway or an electron transfer mech-
anism was studied. The results strongly support the formation
of the radical cation of b-caryophyllene via an electron transfer
from b-caryophyllene to dioxygen as a common initiating step
for both its isomerization and epoxidation.
Results
Quantitative analysis of reaction products
The reaction of bC in the presence of molecular oxygen was
1
monitored with GC-FID, GC-MS, and H NMR spectroscopy. The
concentration profiles of the reaction products and bC as
a function of reaction time (acquired in acetonitrile at 808C
and 1 bar O2) are depicted in Figure 1. 1,4-di-tert-butylbenzene
(DTB) was used as internal standard.
A small amount of 2,6-di-tert-butyl-p-cresol (or butylated hy-
droxytoluene, BHT) was added to every sample to ensure no
further reaction would take place after the sample was taken.
Reduction of the samples with an excess of trimethylphos-
phine prior to GC analysis did not result in the detection of
other products (hydroperoxides are smoothly reduced to their
respective alcohols upon reaction with trimethylphosphine). As
can be seen from Figure 1, a lag time precedes the formation
of oxidation products. In Scheme 1, the structures of the de-
tected products are given. Among the reaction products, b-car-
yophyllene oxide (bCO) was the dominant product, whereas
isocaryophyllene oxide (iCO)—the geometrical isomer of
bCO—was formed in minor amounts. A constant bCO/iCO
ratio of around 10:1 was observed. At prolonged reaction
times, caryophyllene-derived diepoxides were formed. In addi-
tion to epoxides, minor amounts (<5%) of isocaryophyllene
(iC)—the geometrical isomer of bC itself—and a caryophyl-
lene-derived dicarbonyl product (CDC) were formed.
Figure 1. Spontaneous oxidation of b-caryophyllene. Concentration profiles
of A) b-caryophyllene (bC), b-caryophyllene oxide (bCO) and isocaryophyl-
lene oxide (iCO) and B) caryophyllene-derived dicarbonyl products (CDC),
caryophyllene-derived diepoxides (CDE), and isocaryophyllene (iC). Reaction
conditions: [bC]0 =0.1m, [DTB]=0.01m, 1 bar O2, 808C, solvent=MeCN.
plotted as a function of the b-caryophyllene conversion in
Figure 2. Conversion and epoxide selectivity values were con-
The epoxide selectivity and mass balance, as determined by
GC-FID, of the oxygenation reaction of b-caryophyllene are
1
firmed by quantification with H NMR spectroscopy.
The total epoxide selectivity, which comprises the combined
yield of bCO, iCO, and CDE, increases steeply as the conversion
increases, whereas, upon approaching full conversion, the se-
lectivity slightly decreases. After the initial stage of the reac-
tion, the epoxide selectivity exceeds 65%, even at full conver-
sion.
A maximum epoxide selectivity of 73% was attained at 27%
conversion. The selectivity for iC varies between 1% and 2%
during the reaction, leading to an iC yield of 1.5% at full con-
version. The mass balance as determined by GC-FID decreased
as the reaction proceeded, which we attribute to the formation
of nonvolatile products. These nonvolatile products could not
1
be identified with H NMR spectroscopy and are most likely oli-
gomerization products.
The oxygenation of iC was attempted under the same condi-
tions; however, only 7% conversion was achieved after 24 h
Scheme 1. Detected spontaneous oxidation products of b-caryophyllene.
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2147
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Table 1. Influence of oxygen pressure and b-caryophyllene concentration
on reaction rate and selectivity.
[b]
[c]
Entry
[bC]0 [m]
pO2 [bar]
X[a] [%]
(t [h])
Sepox [%]
Sepox/Sisom
1
2
3
4
5
6
0.1
0.1
0.1
0.1
0.02
0.25
1
0[d]
0.2
10
1
22 (4.5)
70
—
67
71
76
69
52:1
—
20:1
223:1
68:1
49:1
<0.5 (10)
19 (6.2)
21 (2.0)
19 (14)
1
20 (2.5)
Reaction conditions, unless stated otherwise: [bC]0 =0.1m, [DTB]=0.01m,
1 bar O2, 808C, solvent=MeCN. [a] X=conversion; [b] Sepox is the total se-
lectivity for epoxides including bCO, iCO, and CDE; [c] Sisom is the selectivi-
ty for iC; [d] the solution was saturated with N2 prior to reaction.
Figure 2. Total epoxide selectivity (Sepox), selectivity for isocaryophyllene
(Sisom) and mass balance of b-caryophyllene oxygenation as a function of b-
caryophyllene conversion. Reaction conditions: [bC]0 =0.1m, [DTB]=0.01m,
1 bar O2, 808C, solvent=MeCN.
higher than 95%, even at higher conversion levels and an in-
creased epoxide selectivity was observed. If a higher initial
concentration of bC was used (Table 1, entry 6), no induction
phase was observed, the reaction was complete after 15 h, and
a mass balance <75% was obtained.
with a selectivity for epoxides of 43%. Two additional control
experiments were performed. First, bC was purified and re-iso-
lated by silica column chromatography and subjected to the
same reaction conditions; identical observations were made as
for the unpurified commercial bC. Second, limonene and va-
lencene—two terpenic components with both an endocyclic
trisubstituted C=C and exocyclic C=C bond—were subjected
to the reaction conditions but conversion levels remained
below 5% for both components after 20 h; no unusually high
epoxide selectivities were observed.
The effect of additives
To gain insight into the mechanism of the oxygenation reac-
tion of bC, the reaction was performed in the presence of dif-
ferent additives (Table 2). First, BHT, 1,2,4-trimethoxybenzene
(TMB), and 1,4-diazabicyclooctane (DABCO) were tested. Upon
addition of BHT (Table 2, entry 2), the reaction was effectively
inhibited and even after 24 h, no product formation was de-
tected. In the presence of TMB (Table 2, entry 3), the lag phase
increased significantly but at prolonged reaction time, the re-
action proceeded with a rate and selectivity comparable to
that of the reference reaction. The additive DABCO (Table 2,
entry 4) resulted in an overall slower reaction but no observa-
ble lengthening of the induction phase. Next, the effect of
a number of commonly used oxidation initiators was tested;
2,2’-azobis(2-methylpropionitrile) (AIBN), N-hydroxyphthalimide
(NHPI), ceric ammonium nitrate (CAN), and Co(acac)3 were se-
lected. When AIBN was added to the reaction mixture (Table 2,
entry 5), no induction phase could be observed and the reac-
tion proceeded significantly faster albeit with comparable se-
lectivity compared to the reference reaction. A remarkable fea-
ture of the reaction is that iC was the main product when
NHPI or CAN were used as additives in MeCN (Table 2, entries 6
and 8–10).
Variation of reaction parameters
The influence of the reaction conditions on the rate and selec-
tivity of the oxygenation reaction of bC was evaluated. At
a temperature below 708C, no conversion could be detected
whereas no induction time was observed when the oxygena-
tion was performed at 1008C. Irradiation of the glass reactor
with a 150 W halogen lamp or performing the oxygenation in
the absence of light did not result in different reaction rates or
selectivities compared to the standard conditions.
The influences of oxygen pressure and concentration of bC
on the conversion and selectivity of the reaction are given in
Table 1. When the standard reaction conditions were applied
(Table 1, entry 1), a conversion level of about 20% was attained
after 4.5 h with a corresponding selectivity for epoxides of
70% and a selectivity for iC of 1.5%. No reaction took place if
the solution was saturated with N2 (Table 1, entry 2). If the re-
action was performed under air instead of pure dioxygen
(Table 1, entry 3), the reaction rate decreased and the selectivi-
ty for iC increased to 3.5%. No induction period could be ob-
served when the reaction was performed in a stainless steel
autoclave at elevated oxygen pressure (Table 1, entry 4), and
only a trace of iC (<0.5%) was formed. If the initial concentra-
tion of bC was lowered (Table 1, entry 5), the reaction proceed-
ed significantly more slowly; the mass balance remained
The reaction with CAN was very fast but was unselective, as
evidenced by the low mass balance value. Performing the reac-
tion with CAN under N2 atmosphere, however, resulted in
a very selective reaction towards iC with a mass balance ex-
ceeding 95%, even at high conversion levels, with 95% iC
yield. Irrespective of the composition of the atmosphere, the
yellow-orange-colored CAN solution became colorless over the
course of 5 min, indicating the rapid consumption of the
cerium salt.
When NHPI was used under an N2 atmosphere, no products
were formed. Strikingly, when NHPI was added in DMAc
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2148
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Solvent effects
Table 2. Effect of additives on the conversion, selectivity and mass bal-
ance of the reaction of b-caryophyllene with molecular oxygen.
The oxygenation of bC was performed in different solvents in
order to gain further insight into the nature of the reaction
(Table 3). Three classes of solvents were tested: Heptane, tolu-
Entry Additive
(concentration
t [h] X [%]
Sepox [%] Sisom [%] Mass
balance
[%]
[mol%])
1[a]
2
none
3
10
3
10
3
13
67
<0.5
<0.5
1
68
70
—
—
60
1
1
—
—
1
97
84
>99
>99
98
Table 3. Solvent effects on the oxygenation of b-caryophyllene.
BHT (2)
Entry
Solvent (type)
X(t=5h) [%]
Sepox [%]
Sepo/Sisom
3
TMB (20)
10
3
10
1
3
1
3
1
3
1
18
<0.5
5
44
78
19
98
2
26
73
—
71
72
71
12
15
72
1
—
1
<0.5
<0.5
78
64
5
91
>99
98
89
84
>99
81
>99
>99
69
Apolar
Heptane
Toluene
4
DABCO (20)
AIBN (2)
1
2
3
<0.5
<0.5
<0.5
—
—
—
—
—
—
5
Iodobenzene
6
NHPI (10)
Protic Polar
Water
MeOH
PrOH
NMAc
4
5
6
7
<0.5
<0.5
<0.5
<0.5
—
—
—
—
—
—
—
—
7[b]
8
NHPI (10)
78
29
6
34
CAN (10)
56
3
1
3
3
10
3
10
1
3
3
87
63
>99
25
71
2
3
27
61
8
50
60
>99
<0.5
<0.5
28
32
96
95
91
96
2
<0.5
2
1
<0.5
<0.5
<0.5[e]
<0.5[e]
—
65
>99
98
98
98
98
98
93
81
95
79
>99
>99
>99
>99
9[c]
10[c]
11
12
13[d]
14[d]
15
CAN (10)
<0.5
<0.5
<0.5
<0.5
19
27
70
66
35
Aprotic Polar
MeCN
EtOAc
DMF
DMAc
8
9
10
11
12
27
21
<0.5
30
72
43
—
81
75
52:1
49:1
—
CAN (0.5)
Co(acac)3 (2)
Bu4N(PF6) (10)
DCA (0.1)
>500:1
>500:1
NMP
83
Reaction conditions: [bC]0 =0.1m, [DTB]=0.01m, 1 bar O2, 808C.
6
1
3
3
49
Rose Bengal (0.5)
<0.5[e]
<0.5[e]
—
ene, and iodobenzene as apolar solvents; water, methanol
(MeOH), propanol (PrOH), and N-methylacetamide (NMAc) as
protic polar solvents; and acetonitrile (MeCN), ethyl acetate
(EtOAc), N,N-dimethylformamide (DMF), N,N-dimethylacet-
amide (DMAc), and N-methylmorpholine (NMP) as aprotic
polar solvents.
KO2 (20)
18-crown-6 (20)
10
—
—
Reaction conditions, unless otherwise stated: [bC]0 =0.1m, [DTB]=0.01m,
1 bar O2, 808C, solvent=MeCN; amounts of additives are given relative
to the concentration of b-caryophyllene). [a] Reference; [b] Solvent=
DMAc; [c] 1 bar N2; [d] RT, hn; [e] allylic hydroperoxides—typical singlet
oxygenation products—were formed.
For none of the tested apolar (Table 3, entries 1–2), halogen-
ated apolar (entry 3), or protic polar solvents (entries 4–7),
could product formation be observed after 5 h. Even after
24 h, the conversion level did not exceed 5% in these solvents.
Performing the reaction in aprotic polar solvents did result in
an active system (Table 3, entries 8, 9, 11, and 12). Interestingly,
using an 80:20 MeCN/H2O mixture instead of neat MeCN as
the reaction medium rendered the system inert. The reaction
proceeded more slowly in EtOAc than in MeCN and resulted in
lower selectivities for both epoxides and iC after 5 h (Table 3,
entry 9). The iC and epoxide selectivity in EtOAc did increase
up to 1% and 70%, respectively, when the conversion exceed-
ed 50%. Neither with MeCN nor EtOAc could solvent-degrada-
tion products be detected over the course of the reaction.
Among the tested amide solvents, no reaction took place in
DMF (Table 3, entry 10), whereas in both DMAc and NMP the
spontaneous epoxidation of bC proceeded smoothly and se-
lectively (Table 3, entries 11 and 12). With epoxide selectivities
of 81% and 75% after 5 h in DMAc and NMP respectively, the
reaction in these amide solvents was significantly more selec-
tive than that in the non-amide polar aprotic solvents. Further-
more, only traces of iC (<0.05%) could be detected in these
solvents and when the additive NHPI was used in DMAc
(Table 2, entry 7), only minor amounts of iC were formed. In
(Table 2, entry 7; see also next section), epoxidation became
the main reaction pathway again. The use of Co(acac)3
(Table 2, entry 11) resulted in the inhibition of the oxygenation
reaction. A very short induction phase (<5 min) and a signifi-
cant rate enhancement were observed if the ammonium salt
Bu4NPF6 (Table 2, entry 12) was used as additive. The selectivity
of the reaction remained however unaffected. The photo-oxy-
genation of bC in the presence of 9,10-dicyanoanthracene
(DCA; Table 2, entry 13) yielded epoxides as the main oxidation
products. Photo-oxygenation in the presence of methylene
blue (Table 2, entry 14) yielded allylic hydroperoxides, as re-
ported in the literature,[11] and no bCO or iC were detected. Fi-
nally, the effect of the superoxide anion on the reaction was
tested by addition of a premixed solution containing KO2 and
a crown ether to the reaction mixture (Table 2, entry 15). The
reaction was completely inhibited in the presence of superox-
ide anion; moreover, when KO2 was added to an ongoing reac-
tion (Table 2, entry 8 after 1 h) the reaction was effectively
stopped and no further products were formed.
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2149
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
DMAc and NMP, solvent degradation did occur. NMAc was the
main degradation product formed from DMAc, whereas 1-
methyl-2,5-pyrrolidinedione and 2-pyrrolidinone were formed
if NMP was used as the solvent.
tion rate and a lower selectivity for the primary hydroperoxide
oxidation products. Secondary products associated with this
abstraction pathway are alcohols and ketones. Addition of
a peroxy radical to an olefin results in a radical peroxo adduct.
This adduct reacts further via the so-called Twigg mechanism
by homolytic cleavage of the peroxo bond into an epoxide
and an alkoxy radical [Scheme 2, reaction (7)].[10] Formation of
epoxides via known autoxidation mechanisms therefore
cannot occur with a selectivity that is higher than 50%. The re-
sidual alkoxy radical can re-enter the propagation chain via re-
action (2). Recombination of two radical species results in the
formation of non-radical products (NRP) which effectively ter-
minates the chain propagation [Scheme 2, reactions (8)
and (9)].
Discussion
Free-radical autoxidation
The epoxidation of bC is inhibited by radical inhibitors such as
BHT and is accelerated in the presence of the typical radical ini-
tiator AIBN. This clearly indicates the involvement of radical
species in the formation of bCO. The plausibility of a “classic”
autoxidation mechanism was therefore verified. Autoxidation
of olefins is performed in most cases at elevated temperatures
and oxygen pressures and the use of transition metals as cata-
lysts is common practice.[12] Olefin autoxidation proceeds via
a free-radical mechanism (Scheme 2). During initiation, precur-
The abstraction-to-addition ratio has been found to be very
structure dependent. Cyclohexene is a well-known example of
an olefin which reacts primarily via abstraction, whereas addi-
tion is the main pathway for the autoxidation of cyclooc-
tene.[13] In the present case of spontaneous bC oxidation, no al-
lylic functionalized reaction products were detected and
sample reduction with trimethylphosphine prior to GC analysis
did not result in a different product composition. These find-
ings indicate that H-atom abstraction is not a major pathway
and that hydroperoxides are not formed in our system. Howev-
er, even if bC would react exclusively via the addition pathway,
the observed epoxide selectivity in our system still cannot be
rationalized by the reactions depicted in Scheme 2.
We subsequently considered the possibility of an alternative
epoxidation pathway. To achieve an epoxide selectivity above
50%, this new route would have to avoid co-formation of
other products. Therefore an epoxidation mechanism was sug-
gested in which the peroxy radical carries a catalytic role
(Scheme 3). Initially, a peroxy radical adds to the C=C double
bond, forming the radical peroxo adduct [Scheme 2, reac-
tion (6)]. Dioxygen may add to this peroxo adduct, forming
a peroxoperoxy radical [Scheme 3, reaction (10)], which may in
turn convert bC into bCO according to the known Twigg
mechanism, yielding a residual oxoperoxy radical [Scheme 3,
reactions (11) and (12)]. This oxoperoxy radical could then de-
compose into bCO upon the release of the original peroxy
fragment [Scheme 3, reaction (13)]. There are, however, two
difficulties with this mechanism. First of all, it is known for
monoterpenes that reaction (10) is only able to kinetically com-
pete with reaction (7) at higher O2 pressures. Although it is
possible that the size and configuration of the trans-cyclono-
nene ring of bC stabilize the radical peroxo adduct, making re-
Scheme 2. Free-radical autoxidation mechanisms.
sors of free radicals are formed and little or no substrate is
converted [Scheme 2, reaction (1)], which typically leads to an
induction phase. In the propagation, chain-carrying peroxy rad-
icals are formed [Scheme 2, reactions (1)–(5)]. These peroxy
radicals react either via H-atom abstraction [Scheme 2, reac-
tion (5)], or alternatively they react via addition to the C=C
double bond [Scheme 2, reac-
tion (6)]. H-atom abstraction by
peroxy radicals at the weaker al-
lylic CÀH bond leads to the for-
mation of allylic hydroperoxides.
These species can decompose
homolytically at the elevated re-
action temperature, resulting in
the formation of more radicals.
This subsequently causes an ex-
ponential increase of the reac- Scheme 3. Proposed free-radical epoxidation mechanism.
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2150
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
action (10) more favorable,[14] there is still the difficult subse-
quent unimolecular step leading to bCO. We therefore compu-
tationally characterized the release of the peroxy fragment and
subsequent formation of bCO from the peroxo adduct
[Scheme 3, reaction (13)]. Unfortunately DFT predicts an activa-
tion energy larger than 25 kcalmolÀ1, in line with the formation
of the strained oxirane ring. This implies that this reaction
cannot kinetically compete with the bimolecular H-atom ab-
straction by the oxoperoxy radical, featuring an estimated bar-
rier of only 5–10 kcalmolÀ1. This analysis evidences that epox-
ide formation via the steps in Scheme 3 is unlikely. No other
free-radical mechanisms were identified that could account for
the high epoxide selectivity in the reaction of bC with dioxy-
gen. Other mechanistic possibilities were therefore explored.
The observed solvent effects do not correspond with the for-
mation of either a diradicaloid species or a zwitterionic inter-
mediate. The inhibition of the reaction in polar protic solvents
like MeOH and PrOH seems to exclude a zwitterionic inter-
mediate, whereas the rate-accelerating effect of polar aprotic
solvents like MeCN and DMAc is not in line with the formation
of a neutral diradicaloid species. Iodine-containing solvents
can be used to detect the involvement of a spin inversion step
by their accelerating effect on the reaction due to the heavy
atom effect.[15c,17] When iodobenzene was used as solvent how-
ever, no reaction took place. Spin inversion usually requires
input of thermal energy or photons. Irradiating the transparent
reaction vessel with visible or UV light did not, however, have
an effect on the reaction.
Electron transfer
Spin inversion mechanism
Electron transfer (ET) may occur between an organic com-
pound and a suitable electron acceptor interacting with each
other in a charge-transfer complex.[18] Reported reduction po-
tential values for dioxygen to the superoxide anion in aprotic
solvents range from À0.50 V to À0.65 V vs. standard hydrogen
electrode (SHE), making it a mediocre electron acceptor.[19]
Electron transfer to triplet oxygen is feasible for electron-rich
compounds with sufficiently low ionization potentials. The
electron transfer results in the formation of the radical cation–
superoxide anion pair. Possible reactions following the actual
electron transfer are summarized in Scheme 5. The ion radical
The direct addition of triplet dioxygen to organic substrates in
their singlet ground state, for example, via CÀH insertion or
[2+2] cycloaddition, is spin-forbidden. Examples of olefins that
are oxygenated by direct reaction with molecular triplet
oxygen are however known in the literature,[15] but, to our
knowledge, there is no example of a nonconjugated trisubsti-
tuted olefin that displays this remarkable reactivity. Direct reac-
tion of olefins with triplet oxygen is initiated by the formation
of a charge-transfer complex. Two hypothetical mechanisms
have been proposed for the formation of a ground-state oxy-
genated product starting from this charge-transfer complex
(Scheme 4; adapted from ref. [16]). The first is the recombina-
Scheme 5. Olefin (A) oxygenation via electron-transfer to triplet dioxygen.
Scheme 4. Direct reaction of an olefin A with triplet dioxygen.
pair may recombine into a zwitterionic singlet (1Z) or diradica-
loid singlet (1D) intermediate, resulting in the oxygenation of
the organic substrate via a non-chain mechanism. If the organ-
ic radical cation is sufficiently stable, it may diffuse out of the
solvent cage and initiate a chain reaction.
tion of dioxygen with the olefin resulting in the formation of
a triplet diradicaloid species (3D). This diradicaloid species may
then undergo intersystem crossing (ISC) through spin inver-
sion, yielding the ground-state oxygenation product. The
second hypothesis proposes that spin inversion precedes the
collapse of the charge-transfer complex with consequent for-
mation of singlet oxygen. Both members of the charge-transfer
complex are now in a singlet state and may react to form the
oxygenation product, typically via a zwitterionic singlet inter-
mediate (1Z).
The endoperoxidation of ergosteryl acetate—a polycyclic
conjugated diene—and the conversion of adamantylideneada-
mantane to its dioxetane are well-known examples of oxygen-
ations of olefins operating via ET.[20] In both cases, the addition
of a potent electron acceptor (EA), such as tris(4-bromophen-
yl)aminium hexachloroantimonate, is required to initiate the
reaction. There are accounts of olefin oxygenation in the litera-
ture in which ET to triplet oxygen itself is the initiating step; in
Although the reaction of bC with dioxygen is inhibited by
DABCO—a potent quencher of singlet oxygen—most of the
experimental data do not support a spin inversion mechanism.
In the presence of singlet oxygen, no epoxides were formed.
3
this case, O2 is the EA. For instance, Correa et al. reported that
the uncatalyzed oxidation of trans-stilbene and trans-b-me-
thoxystyrene at a temperature in the range of 100–1258C and
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2151
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
an O2 pressure higher than 25 bar yields oxidative C=C cleav-
age products due to dioxetane decomposition and minor
amounts of epoxides.[21] The autoxidation is initiated by elec-
tron transfer to dioxygen and a strong correlation was found
between the one-electron oxidation potential of the olefin and
the observed oxidation rate. The oxygenation of conjugated
dienes, adamantylidene derivatives and common olefins under
harsh reaction conditions occurs via a common chain mecha-
nism.[19] The formed radical cation reacts with triplet oxygen,
giving the peroxy radical cation. This peroxy radical cation
then abstracts an electron from an unreacted olefin, forming
the oxygenated product—either a dioxetane or an endoperox-
ide—and regenerating the radical cation. These reaction steps
are given for bC in Scheme 6.
tions of dienes and diarylcyclopropanes.[19] The formation of
charged species is energetically unfavorable in an apolar
medium, explaining the inhibitory effect of the solvents hep-
tane and toluene and rationalizing the accelerating effect of
salts. The oxidation reaction only occurs in aprotic polar sol-
vents such as MeCN. The beneficial effect of these solvents is
twofold. First, according to the Marcus theory,[26] the activation
energy of electron transfer is lowered in solvents with high po-
larizability and second, it is well known that these solvents fa-
cilitate the dissociation of the radical cation–superoxide anion
pair and thus shift the electron transfer equilibrium to the
right.[18] There is no straightforward rationalization for the in-
hibition of the reaction in DMF, but a strong rate-retarding
effect of DMF was also observed in the work of Bartlett on the
spontaneous epoxidation of a norbornylene derivative.[22] Diffu-
sion of the radical cation out of the solvent cage leads to
a chain mechanism, which is consistent with the conversion
profile as a function of the reaction time, with a clear induction
phase. The induction phase was shortened significantly if the
bC concentration or the dioxygen pressure were increased.
The fact that no reaction took place if the reaction vessel was
purged with N2 is strong evidence for dioxygen acting as the
EA.
To our knowledge, only one account of selective epoxidation
via electron transfer to dioxygen is known. The spontaneous
epoxidation of a tetrasubstituted norbornylene derivative was
reported by Bartlett and Banavali.[22] This strained olefin was
epoxidized in 70% yield at room temperature and ambient
pressure of O2. The only reported side product was a diketone
product. Electron transfer to dioxygen was suggested as initiat-
ing step; however, no chain mechanism seemed to operate.
Lastly, we would like to mention the selective epoxidation of
2,3-dimethyl-2-butene at 5 bar O2 and 1008C, which was re-
ported in the patent literature.[23] Even at 80% conversion, the
selectivity for tetramethyloxirane exceeded 70%, with copro-
duction of minor amounts of an allylic alcohol. Unfortunately,
no further mechanistic information is available.
The susceptibility of bC to electron transfer was evaluated
by performing an electrochemical study. The cyclic voltammo-
grams of bC and iC in MeCN at 308C and ambient atmosphere
are compared in Figure 3. The one-electron oxidation of bC ap-
peared to be irreversible in all of the conditions tested. The
maximum current for bC was at 1.3 V vs. Ag/AgI. For reference,
trans-stilbene, which readily undergoes one-electron oxidation,
has a one-electron oxidation potential of 1.6 V vs. Ag/AgI.[21]
The low oxidation potential of bC is reflected in the relatively
mild reaction conditions required for the spontaneous oxygen-
ation of bC when compared to those used in the oxidation of
the substrates in the work of Correa.[21] Furthermore, the oxida-
tion peak of iC appeared at a potential about 0.2 V higher
In the present case of spontaneous epoxidation of bC, the
experimental data indeed seem to support an electron transfer
mechanism. TMB is an efficient one-electron donor and its ad-
dition to the reaction mixture results in a significantly longer
induction phase. BHT and other para-activated phenols effi-
ciently quench free radicals via H-atom transfer and they can
consequently quench radical cations by proton-coupled elec-
tron transfer.[24] DABCO has been suggested to act as an elec-
tron shuttle between the radical cation and the superoxide
anion, restoring the reactants to their neutral state and dissi-
pating the energy as heat.[25] The addition of CAN - a one-elec-
tron acceptor - causes a clear rate-enhancing effect. Further-
more, the observed discoloration of the bright orange CAN is
indicative for the one-electron reduction of CeIV to CeIII. Other
additives known in the literature to act as initiators for electron
transfer chain reactions such as CoIII-based salts and tris(4-bro-
mophenyl)aminium hexachloroantimonate were tested as well,
but with little success.[20b] Co(OAc)3 was found to be insoluble
under the conditions reported in Table 2 (not shown). The in-
hibitory effect of Co(acac)3 may be explained by its reduction
to CoII and the consequent quenching of peroxo radicals by
Haber–Weiss type chemistry. The addition of tris(4-bromophe-
nyl)aminium hexachloroantimonate to the reaction mixture re-
sulted in the formation of a mixture of bC-derived skeletal iso-
merization and oxidation products, but no epoxides were de-
tected. Formation of epoxides during the DCA-sensitized
photo-oxygenation of bC is further evidence for the involve-
ment an electron transfer step. It is well known that DCA acts
as a one-electron acceptor in type I photosensitized oxygena-
Figure 3. CV curves for b-caryophyllene (bC) and Isocaryophyllene (iC). Ex-
perimental conditions: Substrate (5 mm), Bu4NPF6 (0.1m), solvent=MeCN,
308C, ambient atmosphere, scan rate=50 mVsÀ1, Pt working and counter
electrode.
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2152
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
than for bC. We suggest therefore that the lower oxidation
rate of iC in the same experimental conditions can be attribut-
ed to its higher oxidation potential.
our hypothesis on the formation of bCO via a chain mecha-
nism and not via the direct recombination of the radical cation
of bC with superoxide anion. Attempts to perform the oxygen-
ation in protic solvents all failed and even when 10 vol% of
water was added to MeCN, no oxidation products were detect-
ed. Nucleophilic attack of these protic solvents is suspected to
quench the chain-carrying radical cation and results in termina-
tion of the propagation chain. The dioxetane product CDO
was not detected directly by GC-MS but its thermal decompo-
sition product CDC was. It is indeed very likely that CDO de-
composes in the injection port to form CDC. With 1H NMR
spectroscopy, a small signal with a shift of 4.5 ppm was ob-
served, but the low concentration of suspected CDO allows
only a tentative identification.
The synthesis of the observed reaction products under the
standard reaction conditions was attempted electrochemically
in potentiostatic mode (Table 4). A potential interval of 1.7 V
Table 4. Electrochemical conversion of b-caryophyllene.
Entry
E vs. Ag/AgI
t [h]
Atmosphere
X
Sepox
Sisom
[V]
[%]
[%]
[%]
1
2
1.7
1.7
1.5
0.5
1 bar O2
1 bar N2
>99
13
15
<0.5
<0.5
97
Reaction conditions: [bC]0 =5 mm, Bu4NPF6 (0.1m), solvent=MeCN, 808C,
Isomerization of bC+· leads to the less strained cis-cyclono-
nene configuration. Addition of dioxygen to this stabilized rad-
ical cation results in the formation of iCDO whereas transfer of
an electron from bC gives iC and a new molecule of bC+·
(Scheme 6). Isomerization of bC to iC becomes the dominant
pathway if CAN is added to the reaction mixture under N2 at-
mosphere. In fact, the isomerization of bC in the presence of
CAN under N2 atmosphere constitutes an elegant and high-
yielding synthesis route to iC (see Table 2, entry 9) when com-
pared to the currently known routes which require light or
toxic metals.[27] When only 0.005 molar equivalents of CAN
were added under N2, a conversion of 71% was attained after
10 h, suggesting a chain length of at least 140.[28]
Isomerization of bC is also the main reaction when NHPI is
used in MeCN under an oxygen atmosphere. It is known that
NHPI is converted to the corresponding phthalimide N-oxyl
radical (PINO) via in situ H abstraction by dioxygen.[29] In most
of the cases studied to date, PINO functions as an H-atom
shuttle between carbon-centered radicals and peroxy radicals,
enhancing the rate of hydrocarbon autoxidation.[30] However,
besides this H-atom shuttling function, PINO is also capable of
accepting electrons, as was proven recently in a study on elec-
tron-transfer-catalyzed amine demethylation.[31] The NHPI-
mediated isomerization of bC to iC in MeCN then can be ex-
plained by an electron transfer
0.5 h, Pt working and counter electrode.
was applied to a solution of bC in MeCN at 808C. If the solu-
tion was saturated with dioxygen, full conversion was attained
after 1.5 h, with bCO and CDC as the main products detected
by GC-MS. However, the mass balance did not exceed 20%.
When the potentiostatic experiment was performed under N2
atmosphere, 13% of bC was converted selectively to iC with
a complete mass balance; no other products were detected.
The combined information extracted from the effects of addi-
tives, the solvent effects, the variation of reaction conditions
and the electrochemical study provides compelling evidence
for an electron transfer from bC to dioxygen as initiating step
and the subsequent formation of bCO via a chain mechanism.
For the formation of bCO, we propose a two-step mecha-
nism in which initially, a bC-derived dioxetane (CDO or iCDO) is
formed (Scheme 6). The radical cation of bC (bC+·) reacts with
dioxygen to form a peroxy radical cation. This peroxy radical
cation then may abstract an electron from bC, resulting in cyc-
lization to CDO and regeneration of bC+·.
The oxygenation of bC is inhibited by addition of KO2. We
attribute this to quenching of bC+· with superoxide, effectively
terminating the propagation chain (Scheme 5). This supports
from bC to PINO, followed by
isomerization of the radical
cation and an electron transfer
from a fresh molecule of bC to
the isomerized radical cation
(Scheme 6, left). The high epox-
ide selectivity, even at low
oxygen pressures, seems to
indicate that the consecutive
addition of dioxygen to bC+·
and formation of CDO are kineti-
cally more favorable than the
isomerization of the radical
cation. The lack of isomerization
in DMAc and NMP can then be
attributed to the electron trans-
fer properties of these sol-
Scheme 6. Isomerization and oxygenation of the b-caryophyllene-derived radical cation.
Chem. Eur. J. 2015, 21, 2146 – 2156 www.chemeurj.org
vents.[32]
2153
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 7. Proposed route for the selective formation of b-caryophyllene oxide.
The steep increase of the epoxide selectivity with conversion
suggests the initial formation of an epoxidizing species; epoxi-
dation becomes kinetically relevant only after the concentra-
tion of this epoxidizing species reaches a sufficient level. We
therefore suggest that the ET products CDO or iCDO react in
a second step with another molecule of bC forming two mole-
cules of bCO and/or iCO; the proposed mechanism is given in
Scheme 7. There is little literature available on the reaction of
dioxetanes with olefins. Adam et al. studied the reaction of dis-
ubstituted dioxetanes with substituted ethylenic compo-
nents.[33] The obtained products depend on the employed
olefin and solvent. The use of electron-poor olefins mainly
gave 1,6-cycloaddition products or ene-type products. If elec-
tron-rich olefins were employed, minor amounts of epoxides
were also detected. We propose an epoxidation mechanism
according to the original hypothesis by Adam and co-work-
ers.[33] CDO reacts with bC to form a zwitterionic 1,6-dipolar
transition state that may rearrange to a 1,4-dipole. Intramolec-
ular nucleophilic attack of this 1,4-dipole results in the forma-
tion of two molecules of bCO. The facilitating effect of aprotic
polar solvents is consistent with the proposed zwitterionic
transition state. Although epoxides were only minor products
in the report by Adam et al., the strained configuration of bC
may provide an additional driving force for epoxidation.[34] iCO
formation may then be explained by either the 1,6-cycloaddi-
tion by iCDO instead of CDO to bC, or by rotation of the R1ÀR2
axis. The observed high bCO/iCO ratio points however to the
unfavorable formation of iCO and is consistent with the pro-
posed high barrier for isomerization.
lene. The high epoxide selectivity and the absence of allylic ab-
straction products are not consistent with any of the known
free-radical autoxidation mechanisms and a hypothetical cata-
lytic function of a peroxy radical was found to be improbable
by computational methods. The observed solvent effects and
the lack of epoxide formation in the presence of singlet
oxygen excluded a spin inversion mechanism. An electron
transfer from b-caryophyllene to dioxygen was proposed as
the initiating step, forming the radical cation of b-caryophyl-
lene and superoxide anion. This mechanism is supported by
(i) the facilitating effect of aprotic polar solvents, (ii) the inhibi-
tory effect of diazabicyclooctane (a quencher of charge-transfer
complexes), 1,2,4-trimethoxybenzene (a quencher of radical
cations) and 2,6-di-tert-butyl-p-cresol (a radical inhibitor),
(iii) shortening of the induction phase upon increase of the b-
caryophyllene or dioxygen concentration in the solution,
(iv) the accelerating effect of one-electron acceptors, (v) the
epoxidation of bC during type I photosensitized oxidation, and
(vi) the low oxidation potential of b-caryophyllene. The radical
cation carries the propagation chain, in which addition of trip-
let oxygen results in the formation of the caryophyllene-de-
rived dioxetane. Inhibition of the oxygenation reaction by nu-
cleophilic protic solvents, cobalt salts and superoxide, as well
as the detection of the decomposition product of the dioxe-
tane, further support this mechanism. In a final step, the dioxe-
tane reacts with b-caryophyllene via a dipolar 1,6-cycloaddi-
tion, forming two equivalents of the epoxide. Isomerization of
b-caryophyllene to isocaryophyllene is unfavorable and only
becomes a dominant route if the radical cation is formed in
the absence of oxygen via the addition of one-electron oxi-
dants. Future work will be devoted to a thorough EPR spectro-
scopic study of this system and possible synthetic applications
of these findings.
Conclusions
The epoxidation of b-caryophyllene—an abundant sesquiter-
pene—in the presence of molecular oxygen without the addi-
tion of any catalyst was studied. In aprotic polar solvents, the
reaction proceeded spontaneously with selectivities for epox-
ides exceeding 70%. The isomerization product isocaryophyl-
lene was formed in minor amounts during the reaction but it
became the main product if N-hydroxyphthalimide was added
to the reaction mixture, or if ceric ammonium nitrate was used
as an additive under N2 atmosphere. The latter system pro-
vides a novel, elegant, and high-yielding route to isocaryophyl-
Experimental Section
All of the used chemicals were acquired commercially and were of
highest purity available. Solvents were dried over molecular sieves
before use. bC (1 g, 4.9 mmol) was purified by flash chromatogra-
phy with heptane as the eluent over SiO2 gel (50 g, 220–440 mesh)
and re-isolated by removing the solvent under reduced pressure.
Oxygenation reactions at ambient pressure and below 858C were
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2154
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[4] G. A. Burdock, Fenaroli’s Handbook of Flavor Ingredients, 5th ed., CRC
Press, Florida, 2005.
performed in a glass vial. Reactions at elevated pressure and/or
temperature were performed in stainless steel autoclave
a
[5] a) A. Sonboli, B. Babakhani, A. R. Mehrabian, Z. Naturfosch. C 2006, 61,
160–164; b) D. P. Yang, L. Michel, J. P. Chaumont, J. Millet-Clerc, Myco-
pathologia 1999, 148, 79–82.
[6] I. G. Collado, J. R. Hanson, A. J. Macias-Sanchez, Nat. Prod. Rep. 1998, 15,
187–204.
[7] a) L. E. Tatarova, D. V. Korchagina, V. A. Barkhash, Russ. J. Org. Chem.
2002, 38, 519–524; b) A. V. Tkachev, V. I. Mamatyuk, Z. V. Dubovenko,
Zh. Org. Khim. 1990, 26, 1698–1706; c) L. Fitjer, A. Malich, C. Paschke, S.
Kluge, R. Gerke, B. Rissom, J. Weiser, M. Noltemeyer, J. Am. Chem. Soc.
1995, 117, 9180–9189; d) G. A. Nisnevich, D. V. Korchagina, V. I. Makal-
skii, Z. V. Dubovenko, V. A. Barkhash, Zh. Org. Khim. 1993, 29, 524–541.
[8] a) K. C. Lam, R. T. Foster, M. L. Deinzer, J. Agric. Food Chem. 1986, 34,
763–770; b) R. T. Foster, J. Am. Soc. Brew. Chem. 1985, 43, 127–135.
[9] M. Skold, A. T. Karlberg, M. Matura, A. Borje, Food Chem. Toxicol. 2006,
44, 538–545.
[10] a) G. H. Twigg, Chem. Ind.(London) 1962, 4–11; b) T. H. Twigg, Angew.
Chem. 1965, 77, 868–869.
[11] K. H. Schulte-Elte, G. Ohloff, Helv. Chim. Acta 1968, 51, 494–495.
[12] a) M. T. Musser in Ullmann’s Encyclopedia of Industrial Chemistry, 6th ed.,
Wiley-VCH Verlag GmbH & Co., 2000, pp. 5345–5355; b) J.-M. Bre-
geault, Dalton Trans. 2003, 3289–3302.
[13] a) U. Neuenschwander, I. Hermans, J. Org. Chem. 2011, 76, 10236–
10240; b) F. R. Mayo, Acc. Chem. Res. 1968, 1, 193–201.
[14] a) U. Neuenschwander, F. Guignard, I. Hermans, Chemsuschem 2010, 3,
75–84; b) U. Neuenschwander, I. Hermans, Phys. Chem. Chem. Phys.
2010, 12, 10542–10549; c) U. Neuenschwander, E. Meier, I. Hermans,
ChemSusChem 2011, 4, 1613–1621.
equipped with a manometer and sampling valve. Autoclaves were
passivated with saturated sodium pyrophosphate in between dif-
ferent reactions. In a typical oxidation procedure, b-caryophyllene
(0.060 g, 0.30 mmol) and 1,4-di-tert-butylbenzene (0.006 g,
0.03 mmol) were added to acetonitrile (3 mL) in a glass vial with
a septum. The solution was purged with O2 and a balloon of
oxygen was attached to the glass vial with a needle. The solution
was heated to 808C with stirring. We emphasize that the experi-
ments were performed above the flash point of acetonitrile and
appropriate safety precautions had to be taken, especially during
sampling. At regular intervals, aliquots of the solution were taken,
a small quantity of BHT was added and the samples were analyzed
1
with GC-FID, GC-MS, and H NMR spectroscopy. Concentration pro-
files were determined by GC-FID and concentrations were quanti-
fied versus 1,4-di-tert-butylbenzene as an internal standard. Prod-
1
ucts were identified by GC-MS and H NMR spectroscopy; the ob-
tained spectra were compared and matched with those reported
in the literature. In a preparative isomerization procedure, bC
(0.2 g, 1.0 mmol) and CAN (0.055 g; 0.1 mmol) were added to ace-
tonitrile (10 mL) and the mixture was stirred at 808C and under
1 bar of oxygen for 2 h. After reaction, the mixture was extracted
n-heptane (2ꢂ20 mL). The combined heptane layers were washed
with brine (20 mL) and dried over MgSO4 and the solvent was re-
moved by evaporation under reduced pressure. The residual color-
less oil was purified by flash chromatography on silica gel with n-
heptane (100 mL), yielding isocaryophyllene (0.17 g, 85% yield,
98% GC purity). Calculations were performed at the B3LYP/6-
31G(d,p) level of theory using the Gaussian09 software package.[35]
Cyclic voltammetries and potentiostatic experiments were run in
a small electrochemical cell under controlled temperature and at-
mosphere. A platinum-coated silicon wafer was used as working
electrode, while a platinum coil and a home-made Ag/AgI (3m KCl)
electrode were used as counter and reference, respectively. All the
[15] a) C. H. Chou, W. S. Trahanovsky, J. Org. Chem. 1995, 60, 5449–5451;
b) K. B. Clark, J. A. Howard, A. R. Oyler, J. Am. Chem. Soc. 1997, 119,
9560–9561; c) M. Seip, H. D. Brauer, J. Am. Chem. Soc. 1992, 114, 4486–
4490; d) N. J. Turro, V. Ramamurthy, K. C. Liu, A. Krebs, R. Kemper, J. Am.
Chem. Soc. 1976, 98, 6758–6761; e) O. A. Kholdeeva, I. D. Ivanchikova,
O. V. Zalomaeva, A. B. Sorokin, I. Y. Skobelev, E. P. Talsi, J. Phys. Chem. B
2011, 115, 11971–11983.
[16] N. J. Turro, Tetrahedron 1985, 41, 2089–2098.
[17] a) P. D. Bartlett, R. E. McCluney, J. Org. Chem. 1983, 48, 4165–4168;
b) J. C. Koziar, D. O. Cowan, Acc. Chem. Res. 1978, 11, 334–341.
[18] Z. V. Todres, Ion-Radical Organic Chemistry. Principles and Applications,
2nd ed., CRC Press, Boca Ranton, 2008.
experiments were controlled by
EG&G 273.
a
galvanostat/potentiostat
[19] The Chemistry of Peroxides (Ed.: Z. Rappoport), Wiley, New York, 1983.
[20] a) D. H. R. Barton, R. K. Haynes, G. Leclerc, P. D. Magnus, I. D. Menzies, J.
Chem. Soc. Perkin Trans. 1 1975, 2055–2065; b) R. Tang, H. J. Yue, J. F.
Wolf, F. Mares, J. Am. Chem. Soc. 1978, 100, 5248–5249; c) E. L. Clennan,
W. Simmons, C. W. Almgren, J. Am. Chem. Soc. 1981, 103, 2098–2099;
d) S. F. Nelsen, R. Akaba, J. Am. Chem. Soc. 1981, 103, 2096–2097.
[21] P. E. Correa, G. Hardy, D. P. Riley, J. Org. Chem. 1988, 53, 1695–1702.
[22] P. D. Bartlett, R. Banavali, J. Org. Chem. 1991, 56, 6043–6050.
[23] K. G. Wilms, H. Waldmann, J. Grolig, EP 0074009 (B1), 1982, European
Patent Office: Munich, Germany.
Acknowledgements
B.S. is grateful to IWT, Flanders, for providing a research grant.
D.D.V. thanks Belspo (IAP 7/05) and KULeuven (Methusalem
grant CASAS) for support. N.C., J.F. and D.D.V. are thankful to
IWT for the funding of the MOFShape project.
[24] I. J. Rhile, T. F. Markle, H. Nagao, A. G. DiPasquale, O. P. Lam, M. A. Lock-
wood, K. Rotter, J. M. Mayer, J. Am. Chem. Soc. 2006, 128, 6075–6088.
[25] J. E. Pickett, ACS symp. ser. 1985, 280, 313.
[26] a) R. A. Marcus, J. Chem. Phys. 1956, 24, 966–978; b) C. J. Schlesener, C.
Amatore, J. K. Kochi, J. Phys. Chem. 1986, 90, 3747–3756.
Keywords: epoxidation · isomerization · radical ions · reaction
mechanisms · sesquiterpenes
[27] a) K. Gollnick, G. Schade, Tetrahedron Lett. 1968, 9, 689–695; b) S. Ra-
chlin, US 3621070 (A), 1971, Patent and Trademark Office: Washington,
DC.
[28] CAN is the only electron acceptor present in the system and CAN has
no known epoxidation activity. Furthermore, no dioxygen is present, ex-
cluding reoxidation of cerium. It is therefore reasonable to assume that
the minimum chain length is the ratio of the amount of epoxide
formed to the amount of CAN added.
[1] a) T. Hoffmann, J. R. Odum, F. Bowman, D. Collins, D. Klockow, R. C.
Flagan, J. H. Seinfeld, J. Atmos. Chem. 1997, 26, 189–222; b) T. E. Klein-
dienst, M. Jaoui, M. Lewandowski, J. H. Offenberg, C. W. Lewis, P. V.
Bhave, E. O. Edney, Atmos. Environ. 2007, 41, 8288–8300; c) M. Jaoui,
T. E. Kleindienst, K. S. Docherty, M. Lewandowski, J. H. Offenberg, Envi-
ron. Chem. 2013, 10, 178–193.
[2] a) J. Gertsch, M. Leonti, S. Raduner, I. Racz, J. Z. Chen, X. Q. Xie, K. H. Alt-
mann, M. Karsak, A. Zimmer, Proc. Natl. Acad. Sci. USA 2008, 105, 9099–
9104; b) H. Muroi, I. Kubo, J. Agric. Food Chem. 1993, 41, 1102–1105;
c) Y. S. Kim, S. J. Park, E. J. Lee, R. M. Cerbo, S. M. Lee, C. H. Ryu, G. S.
Kim, J. O. Kim, Y. L. Ha, J. Food Sci. 2008, 73, C540–C545.
[29] Y. Ishii, S. Sakaguchi, T. Iwahama, Adv. Synth. Catal. 2001, 343, 393–427.
[30] I. Hermans, P. Jacobs, J. Peeters, Phys. Chem. Chem. Phys. 2007, 9, 686–
690.
[31] a) E. Baciocchi, M. Bietti, M. Di Fusco, O. Lanzalunga, D. Raponi, J. Org.
Chem. 2009, 74, 5576–5583; b) E. Baciocchi, M. Bietti, M. F. Gerini, O.
Lanzalunga, J. Org. Chem. 2005, 70, 5144–5149.
[3] S. C. Rastogi, J. P. Lepoittevin, J. D. Johansen, P. J. Frosch, T. Menne, M.
Bruze, B. Dreier, K. E. Andersen, I. R. White, Contact Dermatitis 1998, 39,
293–303.
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2155
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[32] The use of NHPI in DMAc did not lead to isomerization but to a more
selective epoxidation and the demethylation product NMAc was de-
tected after reaction by GC-MS. These striking results can be rational-
ized by the demethylating properties of PINO; in DMAc, the solvent
itself is a one-electron donor and undergoes demethylation via ET to
PINO. This reaction of DMAc with PINO renders ET from bC to PINO ki-
netically uncompetitive. This has two consequences: (1) No viable iso-
merization route is present and (2) the presence of the methyl radical—
formed via the demethylation of DMAc—opens up an additional route
for bC epoxidation via reactions (4), (6), and (7) in Scheme 2, resulting
in a higher epoxide selectivity. An analogous role is suspected for the
solvent NMP. The high epoxide selectivity, even at low oxygen pressure,
seems to indicate that the concurrent addition of dioxygen to the gen-
erated radical cation and formation of CDO are kinetically more favora-
ble than the isomerization of the radical cation. Additionally, the higher
Sepox/Sisom ratio in DMAc compared to MeCN can be rationalized in
terms of the EA properties of DMAc and the apparent high barrier for
isomerization. The quenching of the radical cation by DMAc prevents
its rearrangement to its more stable cis-cyclononene configuration,
whereas in MeCN the solvent has no EA-properties and a few percent
of iC can form.
[34] a) K. J. Shea, J. S. Kim, J. Am. Chem. Soc. 1992, 114, 3044–3051; b) B.
Steenackers, A. Neirinckx, L. De Cooman, I. Hermans, D. De Vos, Chem-
PhysChem 2014, 15, 966–973.
[35] Gaussian 09, Revision A.02, G. W. T. M. J. Frisch, H. B. Schlegel, G. E. Scu-
seria, 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. Iz-
maylov, 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. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, 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. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
Received: August 3, 2014
[33] W. Adam, S. Andler, M. Heil, Angew. Chem. Int. Ed. Engl. 1991, 30, 1365–
1366; Angew. Chem. 1991, 103, 1395–1396.
Published online on November 27, 2014
Chem. Eur. J. 2015, 21, 2146 – 2156
www.chemeurj.org
2156
ꢁ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim