Oxidative degradation of fragrant aldehydes. Autoxidation by molecular oxygen

Article abstract of DOI:10.1016/j.tet.2013.01.034

The oxidative degradation of fragrant aldehydes by molecular oxygen has been investigated. The oxygen consumption was monitored and the bond dissociation energy (BDE) of the aldehyde C(O)-H bond were calculated by DFT method. The oxidation products were identified by GC/MS. The different pathways accounting for the oxidative degradation are discussed. The main product is the acid, beside the formate ester. Both oxidation products result from the Baeyer-Villiger reaction involving a peracid R(CO)OOH whereas minor products arise from the hydroperoxide ROOH intermediate derived either from the acyl peroxy radical, R(CO)OO or from the decarboxylation of the peracid RC(O)OOH.

Full text of DOI:10.1016/j.tet.2013.01.034

Tetrahedron 69 (2013) 2268e2275
Contents lists available at SciVerse ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Oxidative degradation of fragrant aldehydes. Autoxidation by
molecular oxygen
,
,
,
,
C. Marteau a b, F. Ruyffelaere a b, J.-M. Aubry a b, C. Penverne a b, D. Favier c,
V. Nardello-Rataj a
Universite Lille Nord de France, F-59000 Lille, France
,b,
*
a
ꢀ
ˇ
b
ꢀ
ꢀ
ꢀ
Universite Lille 1 and ENSCL, EA 4478 Chimie Moleculaire et Formulation, Cite Scientifique, Batiment C6, F-59655 Villeneuve d’Ascq Cedex, France
c International Flavors & Fragrances Research and Development, 61 rue de Villiers, 92523 Neuilly-sur-Seine Cedex, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The oxidative degradation of fragrant aldehydes by molecular oxygen has been investigated. The oxygen
consumption was monitored and the bond dissociation energy (BDE) of the aldehyde C(O)eH bond were
calculated by DFT method. The oxidation products were identified by GC/MS. The different pathways
accounting for the oxidative degradation are discussed. The main product is the acid, beside the formate
ester. Both oxidation products result from the BaeyereVilliger reaction involving a peracid R(CO)OOH
whereas minor products arise from the hydroperoxide ROOH intermediate derived either from the acyl
peroxy radical, R(CO)OOꢀ or from the decarboxylation of the peracid RC(O)OOH.
Received 25 October 2012
Received in revised form 3 January 2013
Accepted 9 January 2013
Available online 18 January 2013
Keywords:
Aldehyde
Oxidation
Fragrance
Peracid
Ó 2013 Elsevier Ltd. All rights reserved.
Hydroperoxide
Mechanism
1. Introduction
be clearly identified. From several years, the mechanism of air oxi-
dationofsometypicalfragrancemolecules, inparticularterpenesand
Oxidative degradation of organic compounds, involving either
ground state molecular dioxygen (3O2) or reactive oxygen species
(ROS), such as singlet oxygen 1O2, superoxide anion O2$ꢁ, hydro
(pero)xyl HOꢀ and HOOꢀ, alkoxy ROꢀ, peroxy ROOꢀ, acyl RCOꢀ and acyl
peroxy RC(O)OOꢀ radicals, is an important issue to account for the
long-term stabilityofmanyend-use products.1 Forperfumesuppliers
and the cosmetic industry in particular, the oxidation of fragrance
molecules is directly linked to the product safety, shelf-life, off notes
or colouration, which must be avoided. Moreover, during the last 20
years, fragrance contact allergy has also become a major problem due
to an increase of fragrance-based products and because the products
resulting from the oxidative degradation of fragrances, such as hy-
droperoxides are frequent causes of contact allergy.2 Studies have
proven that almost 1% of the population suffers from respiratory
threats caused by allergens in perfumes and fragrances.3 This has led
to a listof 26fragrance molecules, which have to be mentioned on the
packaging since 1999 when they are present in a formulation. For
these reasons, it appears crucial to understand the mechanisms in-
volved and responsible for the formation of allergens, which have to
terpenoids, such as limonene, geraniol, linalool, caryophyllene, have
been thoroughly investigated with a special focus at their skin sen-
sitizing properties through the formation of hydroperoxides.4
Fragrance molecules, typically terpenes and phenols but also al-
dehydes, are actuallyelectron-rich molecules likely to oxidize readily
upon contact air exposure.5 Their oxidation, occurring through a free
radical chain process, can be initiated by several external factors like
heat, oxygen, light but also by impurities often present in perfumes
like traces of metals, hydroperoxides, peracids, photosensitizers, etc.
as well as by interaction with the other ingredients of the formula-
tion, such as enzymes or bleaching agents.6 Preventing a complex
mixture of fragrances from oxidation in order to minimize the
presence of unfavourable by-products is a real challenge, which has
become nowadays a priority for the perfumer who must take care to
maintain both the safety and the performances of the product.
Among the diversity of fragrance molecules, aldehydes are the most
sensitive towards oxidation by 3O2 since the bond dissociation en-
ergy (BDE) of the aldehyde function is relatively low
(z89 kcal molꢁ1) leading to a radical RC(O)ꢀ through hydrogen ab-
straction.7 Aldehydesarepresentinmanyfragrance compositions. As
an example, 2-methylundecanal was the first synthetic aldehyde
molecule introduced into a perfume in 1921, i.e., the famous Chanel
nꢂ5. However, their degradation takes place rapidly involving
* Corresponding author. E-mail address: veronique.rataj@univ-lille1.fr
(V. Nardello-Rataj).
0040-4020/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tet.2013.01.034

C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
2269
different types of radical species.8 From a general view point, the
reaction occurs through classical radical pathways already described
for current aldehydes but very little work report studies dealing with
fragrant aldehydes specifically used in perfumery.9
In the present work, we have investigated the oxidation of seven
fragrant aldehydes commonly used in fine perfumery. As a kinetic
approach, the oxidation rates have been determined by monitoring
the oxygen consumption. Bond dissociation energies (BDE) have
also been calculated in order to rationalize the oxidability of the
fragrant aldehydes. In a second step, the oxidation products have
been detected and identified by GC/MS. Combination of all this
information led us to propose a complete mechanism for the au-
toxidation as well as a rationalization of the structural effects on the
autoxidation of fragrant aldehydes.
ꢀ
RCHO þ ROOH/RðCOÞ þ ROꢀ þ H2O
(1c)
Denisov has shown that, in peroxide-free aldehydes, the tri-
molecular reaction 1b predominates when the aldehyde exhibits
a very low BDE of the C(O)eH bond, such as benzaldehyde and
unsaturated aldehydes whereas in other cases, the bimolecular
reaction 1a is usually the main pathway.7 However, according to
Peeters et al., the other pathway 1c might be important in the
presence of traces of peroxide since they reported that the disso-
ciation of peroxides can be accelerated by the abstraction of the
weak H-atom of the aldehyde group.14
Propagation
2. Results and discussion
ꢀ
ꢀ
3
RCO þ O2/RðCOÞOO
(2)
(3)
From a general point of view, oxidation of aldehydes by ground
state dioxygen (3O2) is known for a long time.11 As early as 1835,
Liebig observed that in the presence of air, aldehydes are converted
into their corresponding acids.12 Since that time, the reaction has
been widely studied with various aldehydes and the detailed
mechanism is now well established.12 Surprisingly, very little work
report studies dealing with fragrant aldehydes specifically used in
perfumery in spite of the importance of this phenomenon for the
stability of fragrance composition and for consumer health.9 The
autoxidation of seven fragrant aldehyde molecules, selected for their
structural features and their interest in perfumery (Scheme 1) has
thus been investigated. Except valeraldehyde 5, they all bear either
kp
/
,
RðCOÞOO, þ RCHO
RðCOÞOOH þ RðCOÞ
Termination
kt
/
2RðCOÞOO,
molecular products
(4)
During the propagation step, the acyl radical R(CO)ꢀ rapidly reacts
with an oxygen molecule to give the acyl peroxy radical R(CO)OOꢀ
(Eq. 2), which abstracts a hydrogen atom to another aldehyde mol-
ecule with a relatively high rate constant kp (z103 mol Lꢁ1 sꢁ1
)
alkyl or benzyl substituents in
order to study the impact of substitution on their oxidability.
a position of the aldehyde function in
compared to z1e10 mol Lꢁ1 sꢁ1 for alkyl peroxy radicals on alkene,8
giving the corresponding peracid R(CO)OOH and regenerating at the
same time the acyl radical R(CO)ꢀ (Eq. 3). In the termination step,
the acyl peroxy radical R(CO)OOꢀ can recombine leading to the
acyloxy radical R(CO)Oꢀ as an intermediate, which gives, after de-
carboxylation and oxygen addition, minor oxidation products, such
as alcohols, aldehydes or ketones (Eq. 4) (see Section 2.3 for details
and discussion). However, the main pathway providing the major
oxidation products, i.e., the carboxylic acids 11 and the formate es-
ters 12, corresponds to the addition of the peracid R(CO)OOH on the
starting aldehyde according to the BaeyereVilliger rearrangement
(Eq. 5). Besides the acid, an ester formate can also be formed (Eq. 6).15
O
O
O
O
H
H
H
H
bourgeonal, 1
florhydral, 2
lilial, 3
floralozone, 4
O
O
O
H
H
H
valeraldehyde, 5
2-methyl valeraldehyde, 6
2-ethyl hexanal, 7
Scheme 1. Chemical structures of the fragrant aldehydes.
Non-radical molecular reactions
2.1. Kinetic and thermodynamic approaches of the autoxi-
dation of fragrant aldehydes
RðCOÞOOH þ RCHO/2RCOOH
(5)
(6)
The autoxidation ofaldehydes RCHO 8is well known fora longtime.
It occurs through a free radical chain process involving a series of ele-
mentary reactions, i.e., initiation, propagation and termination steps.
As reported by many authors, initiation starts with abstraction of the
aldehydic hydrogen, induced by (i) heat plus O2 (reactions 1a and 1b),
(ii) transition metallic ions traces, (iii) UV light or (iv) traces of perox-
ides present in the medium.12,13 All these various initiation reactions
are represented through the generic Eq.1 that leads to the formation of
acyl radicals R(CO)ꢀ according to a global initiation rate constant Ri.
RðCOÞOOH þ RCHO/RCOOH þ ROCHO
Thedetermination of thepropagation rate constant(kp) (Eq. 3) can
be used to discuss the oxidability of organic compounds (RH) since it
is generally well correlated to the activation energy of the hydrogen
atom transfer reaction.16 kp is determined by measuring the rate of
oxygen consumption (ꢁd[O2]/dt) knowing the rate constants of the
termination (kt) and of the initiation rate (Ri) according to Eq. 7.17
Initiation
R0i :5½RHꢃkp
d½O2ꢃ
dt
pffiffiffiffiffiffiffi
ꢁ
¼
(7)
Ri
!
,
2kt
RCHO
RðCOÞ
(1)
Since rate constants of the termination and initiation steps for
ꢀ
RCHO þ O2/RðCOÞ þ HOOꢀ
(1a)
the fragrant aldehydes are not referenced, kp cannot be easily de-
termined. Hence, ꢁd[O2]/dt determined using the PetroOxy appa-
ratus can be used as an indicator of the oxidability since generally,
the ratio kp/(2kt)0.5 is used as an oxidability parameter.18 The seven
fragrant aldehydes have thus been oxidized at a concentration
ꢀ
2RCHO þ O2/2RðCOÞ þ H2O2
(1b)

2270
C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
equal to 2 mol Lꢁ1 in octane, an inert solvent, at 25 and 40 ꢂC, and
under an oxygen pressure of 300 kPa. The decrease of the pressure
reflects the oxygen consumption and hence, the oxidability of the
aldehydes could be monitored as a function of time. The experi-
ments were automatically stopped when the pressure of oxygen
had decreased down to half of the starting value after heating
(Pf¼Pmax/2). The evolution of the pressure was then converted to
a concentration of oxygen consumed per volume of the liquid phase
D[O2]t defined by Eq. 8.
ðPmax ꢁ PtÞ Vtot ꢁ Vald
D
½O2ꢃt
¼
ꢄ
(8)
RT
Vald
where Vtot and Vald are the volumes of the cell and aldehyde solu-
tion, respectively, Pmax is the maximum pressure obtained a few
minutes after heating the cell and Pt is the pressure measured at
a given time t (Fig. 1). It is noteworthy that rigorously, the equi-
librium between the liquid and the gas phases expressed through
the Henry coefficient should be taken into account. However, in our
study, the oxygen rate consumption was determined by consider-
ing the initial decrease of the pressure, which represents about 10%
of Pi. Hence, we can assume that in these conditions, the oxidation
rates of the different aldehydes can be compared with confidence.
Fig. 2. Kinetics of oxygen consumption at 40 ꢂC for bourgeonal 1, florhydral 2, lilial 3,
floralozone 4, valeraldehyde 5, 2-methyl valeraldehyde 6 and 2-ethyl valeraldehyde 7
measured with the PetroOxy apparatus. Conditions: P(O2)0¼300 kPa, T¼40 ꢂC,
[aldehyde]¼2 mol Lꢁ1 in octane.
Table 1
Rates of oxygen consumption, Rox, for bourgeonal 1, florhydral 2, lilial 3, floralozone
4, valeraldehyde 5, 2-methyl valeraldehyde 6 and 2-ethyl valeraldehyde 7 measured
with the PetroOxy apparatus. Conditions: P(O2)0¼300 kPa, [aldehyde]¼2 mol Lꢁ1 in
octane, T¼25 and 40 ꢂC
Aldehyde
BDEa
ꢁd[O2]/dt
(kcal molꢁ1
)
(ꢄ103 mol Lꢁ1 minꢁ1
)
25 ꢂC
40 ꢂC
Bourgeonal 1
Florhydral 2
Lilial 3
Floralozone 4
Valeraldehyde 5
2-Methyl valeraldehyde 6
2-Ethyl hexanal 7
93.8
93.7
93.3
93.1
93.4
93.2
92.2
0.6
2.5
7.8
1.7
5.9
16.5
12.1
19.3
30.2
39.0
5.4
15.2
24.1
32.0
a
Calculated with the DFT method B3LYP 6-311G (2d, 2p) in heptane.
aldehydes 5, 6 and 7 as well as lilial 3 and floralozone 4 are the most
oxidable aldehydes whereas bourgeonal 1 oxidizes but to a much
lower extent and florhydral 2 exhibits an intermediate behaviour.
As depicted on Fig. 2, alkyl aldehydes 5, 6 and 7 are more rapidly
oxidized than aldehydes bearing an aromatic ring. Indeed, the
phenyl group has an electro-withdrawing effect, which destabilizes
the radical formed by abstraction of the aldehydic hydrogen. As
a consequence, the BDE value increases of 0.1e0.4 kcal molꢁ1 for
the corresponding couples 5/1 and 6/3 and the oxidation rate de-
creases accordingly (see Table 1). The presence of an alkyl group in
Fig. 1. Monitoring of the oxygen consumption with the Petroxy apparatus during the
autoxidation of bourgeonal 1 (2 mol Lꢁ1) at 40 ꢂC.
Because aldehydes are highly reactive towards oxygen, oxygen
rate consumption of the seven aldehydes could be determined
precisely at a relatively low temperature (40 ꢂC) within a reason-
able time. The oxidation of the seven fragrant aldehydes has also
been monitored at 25 ꢂC to check whether the mechanism of oxi-
dation depends on the temperature. From the linear part of the
curves shown in Fig. 2, the rate of oxidation (ꢁd[O2]/dt) is calcu-
lated from the slope for each aldehyde at the two investigated
temperatures. Results are reported in Table 1 and oxygen con-
sumption as a function of time at 40 ꢂC is shown on Fig. 2.
The characteristic value of the bond dissociation energy (BDE) of
the aldehydic hydrogen defined as the energy, which is necessary to
break the bond between the carbonyl group and the aldehydic
hydrogen, has been calculated using the DFT method B3LYP 6-311G
(2d, 2p) in heptane.10 This thermodynamic data is generally cor-
related to the oxidability of compounds and particularly to the
propagation rate constant.19 Actually, the lower the BDE value, the
higher the propagation rate constant and the oxidation rate are.
Despite similar BDE values for benzylic and aldehydic bonds, we
have only considered the second one for aromatic aldehydes be-
cause no oxidation product corresponding to hydrogen abstraction
from the benzylic hydrogens was detected.
a
position of the aldehyde function clearly speeds up the oxidation
if we compare ꢁd[O2]/dt of 3 versus 1 and 6 versus 5 (see Table 1).
This result was observed by Everett and McDonough during an
autoxidation of aldehydes in the dark at 29 ꢂC.20 Indeed, the methyl
group has an electro-donating effect, which decreases the BDE of
0.2 and 0.5 kcal molꢁ1 for the couples 5/6 and 1/3, respectively, and
stabilizes the radical formed and thus favouring the autoxidation of
the molecule. The substitution of a methyl group by an ethyl group
in
position of this alkyl group plays a more important role. Thus lilial 3
with a methyl group in position exhibits both a lower BDE and
a much larger ꢁd[O2]/dt than florhydral 2, which has a methyl
group in position. Unexpectedly, an additional alkyl group in
position slows down the oxidation (compare 3 and 4 in Table 1) in
a
position has a significant effect on ꢁd[O2]/dt (6 vs 7) and the
a
b
a
spite of the decrease of the BDE value. This results from the for-
mation of the ROOꢀ intermediate (see below). From these results,
we can conclude that the oxidability of the fragrant aldehydes
strongly depends on their chemical structure and slightly on the
BDE value of the aldehyde function. The seven aldehydes
As shown in Table 1, the order of oxidability of the fragrant al-
dehydes does not depend on the temperature. The three alkyl

C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
2271
understudy can thus be classified from the more sensitive to
2.2. Oxidation products formed during the autoxidation of
the fragrant aldehydes
the more stable molecule against oxidation as follows: 2-ethyl
hexanal 7>2-methyl valeraldehyde 6>valeraldehyde 5>>lilial
3>floralozone 4>florhydral 2>>bourgeonal 1. It is noteworthy
that, although the range of BDE values calculated for the seven
aldehydes is rather narrow, they provide a coherent tendency in
good agreement with kinetics results. Therefore the BDE of the
R(CO)eH bond appears to be a significant parameter to account for
the formation of R(CO)ꢀ radical besides the nature and the number
The oxidation products formed during the autoxidation process
in octane at 40 ꢂC under oxygen pressure in the PetroOxy apparatus
have been analyzed and identified by GC/MS. The chemical struc-
tures of the detected products and their relative proportions are
reported in Table 2.
As reported in Table 2, the nature and the distribution of the
oxidation products resulting from the autoxidation process, i.e.,
of a-substituents that also have a strong influence.
Table 2
Oxidation products formed during the autoxidation of bourꢁge1onal 1, florhydral 2, lilial 3, floralozone 4, valeraldehyde 5, 2-methyl valeraldehyde 6 and 2-ethyl valeraldehyde 7.
Conditions: P(O2)0¼300 kPa, T¼40 ꢂC, [aldehyde]¼2 mol L in octane
Fragrances
O
O
O
O
O
O
O
H
H
H
H
H
H
H
1
2
3
4
5
6
7
Reaction time (min)a
140
55
22
28
15
8
6
Major Oxidation products (relative %)b
Acids 11 from pathway 3a (first generation)
O
O
O
O
O
O
O
OH
OH
OH
OH
OH
OH
OH
95.6
99.0
99.1
75.0
100
92.5
81.1
Formate esters 12 from pathway 3b (first generation)
O
H
O
H
O
H
O
H
nd
nd
O
O
O
O
< 1
22.0
7.5
18.9
Minor Oxidation productsb
Products of first generation
Alcohols 17 from pathway 5a
OH
OH
OH
OH
Aldehydes 15 from pathway 5b
H
H
H
nd
O
O
O
Alcohols 14 from pathway 4
OH
OH
OH
OH
Products of second generation
Acids 11' from pathway 3a
O
OH
nd
nd
nd
Alcohols 17' from pathway 5a
nd
OH
O
OH
nd
Aldehydes 15' from pathway 5b
O
O
O
H
H
H
nd
a time at which P(O2)t = P(O2)0/2. b detected and identified by GC/MS. nd = not detected. For pathway numbers, see Scheme 5.

2272
C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
acids, alcohols, aldehydes, ketones and formates are slightly dif-
ferent from one aldehyde to the other, especially if we compare
aliphatic and aromatic aldehydes.
oxidation products slightly differ from one aldehyde to the other.
Under our conditions (i.e., after consumption of half of the initial
P(O2)0), less than 1% of these products could be detected for flo-
rhydral 2 and lilial 3, about 3% for floralozone 4 and 4.4% for
bourgeonal 1, which provide less carboxylic acid that 2 and 3 and
no formate ester compared to 4. This finding is in correlation with
the fact that bourgeonal 1 oxidizes much slower than the other
aldehydes and that floralozone 4 oxidizes slower than lilial 3 in
spite of a slightly lower BDE. With regard to the nature of the minor
oxidation products, they all result from various cleavages providing
shorter chain compounds. It is noteworthy that the shorter alde-
hydes formed can further oxidized, mainly according to the
BaeyereVilliger or to the Hock rearrangements, as the starting al-
dehyde leading to same type of oxidation products of second
generation. In the case of bourgeonal 1, for example, the shorter
carboxylic acid 110 appearing among the minor oxidation products
was detected in significant amounts (z3%).
However, in all cases, the main product is the acid RCOOH 11.
Indeed, as depicted in Scheme 2, addition of the peracid R(CO)OOH
9 on the carbonyl group of the starting aldehyde RCHO 8 leads to
a peracid-aldehyde adduct 10 according to a reversible reaction, the
so-called BaeyereVilliger mechanism.15 This adduct undergoes an
acid-catalyzed rearrangement by migration of the aldehydic hy-
drogen (pathway a) providing two molecules of acid 11 or by
transposition of the substituent R (pathway b) giving one molecule
of acid 11 and one molecule of formate ester 12.
O
R
C OH
(a)
Acid 11
H
R
C
O
O
C
O
C
O
C
O
C
H
C
R
and
H
H+
O2
R
H
R
OOH
R
OH
R
O
O
O
H
+
slow
Acid 11
Peracid 9
Aldehyde 8
RO
C
O
2.3. Mechanism of autoxidative degradation of fragrant
aldehydes
Peracid-aldehyde adduct 10
(b)
Formate
ester 12
Scheme 2. BaeyereVilliger rearrangement of the peracid-aldehyde adduct 10.15
Many unstable intermediates are involved in the autoxidative
degradation of fragrant aldehydes. They account for the formation
of all the oxidation products, the nature of which depends on the
aldehyde structure.
It has been shown that higher conversions into the acid 11 are
obtained in polar and protic solvents like aliphatic alcohols due to
their ability to create hydrogen bonds.21 The preferred pathway de-
pends on the ability of the R group bound to the aldehydic carbon to
migrate, i.e., H>>tert-alkyl>cyclohexyl>sec-alkyl>benzyl>phenyl>
n-alkyl>methyl.22 The acid 11 is always the major product because
the H atom readily migrates and pathway (a) is always largely pre-
dominant. Since the oxidation of aldehydes by a peracid provides the
same products in same relative ratios as those obtained by autoxi-
dation, Lehtinen et al. have established some rules for the substituent
effect on the selectivity of the BaeyereVilliger reaction on the basis of
the oxidation of more than ten model aldehydes by m-chloro-
perbenzoïc acid (m-CPBA).23 The authors have shown that a sub-
During the propagation step, acyl peroxy radicals R(CO)OOꢀ are
generated (Eq. 2). Their recombination in the termination step can
lead to the formation of alkyl radicals Rꢀ, after thermal cleavage,
which rapidly react with oxygen to form alkyl peroxy radicals ROOꢀ
(Scheme 3). It is noteworthy that a minor pathway based on the
recombination between the acyl peroxy R(CO)OOꢀ and the alkyl
peroxy ROOꢀ radicals can also take place providing ketones, alde-
hydes, alcohols and acids.
stitution in
a position of the aldehyde function is required for the
formation of the formate ester HC(O)OR 12.
Thus, aliphatic aldehydes bearing no substituent in
a position,
Scheme 3. Detailed mechanism of the termination Reaction 4 involving two acyl
peroxy radicals R(CO)OOꢀ. In the first step it produces oxygen and an acyloxy radical
R(CO)Oꢀ, which gives in a second step CO2 and an alkyl radical Rꢀ after thermal
cleavage.
such as valeraldehyde 5 are quantitatively oxidized into the cor-
responding acid because the primary alkyl group does not migrate
easily in the peracid-aldehyde adduct 10. However, when an alkyl
group is present in a position as in the case of aldehydes 6 and 7, the
The alkyl peroxy radical ROOꢀ formed during the recombination
and decomposition of the acyl peroxy radical R(CO)OOꢀ can abstract
a hydrogen from a surrounding aldehyde leading to another neutral
unstable intermediate, i.e., a hydroperoxide ROOH 13. In acidic
medium this hydroperoxide can undergo a Hock rearrangement,
which can also account for the formation of alcohols, ketones and
aldehydes.21 Indeed, acidic attack can induce a heterolytic cleavage
of the OeO bond giving an oxocarbenium ion, which reacts with
H2O providing a hemiacetal, which is cleaved into the alcohol and
the corresponding carbonyl derivative in acidic conditions as
depicted in Scheme 4.
carboxylic acid 11 is still the major product but a significant amount
of formate ester 12 resulting from a migration of the sec-alkyl
groups as depicted in Scheme 2 is also detected, up to z19% in the
case of 7. This formate ester 12 is actually notably favoured in the
case of the ethyl group compared to methyl group in agreement
with Lehtinen et al.23
As for aliphatic aldehydes, the main product for the aromatic
aldehydes is the corresponding acid ranging from 75% for flor-
alozone 4 to 99% for lilial 3. In the case of 4, the formate ester
derivate is formed in an important amount, about 22%, because the
presence of two methyl groups in
a position provides a tert-alkyl
group, which easily migrates during the rearrangement of the
aldehyde-peracid adduct 10. As far as lilial 3 is concerned, the mi-
gration of the sec-alkyl group is a minor pathway and acid is
preferentially formed (z99%).
Besides the acid 11 and the formate 12 expected from the
BaeyereVilliger rearrangement, several other minor products, such
as alcohols, ketones, aldehydes and acids resulting from further
decomposition of the peracid 9 and corresponding to a loss of one
or two carbons have also been detected for the aromatic fragrant
aldehydes 1 to 4 (Table 2). It is noteworthy that depending on their
The major pathways that can take place during the autoxidation
of fragrant aldehydes as well as the different possible resulting
oxidation products (i.e., major and minor oxidation products of first
and second generations) are summarized in Scheme 5.
Most of the initial radicals R(CO)ꢀ reacts quickly with oxygen to
form acyl peroxy radicals R(CO)OOꢀ, which propagate the autoxi-
dation by abstracting aldehydic hydrogen giving a peracid R(CO)
OOH 9 as the main intermediate product. This peracid reacts with
the aldehyde according to the BaeyereVilliger reaction giving the
major oxidation product, i.e., a carboxylic acid RCOOH 11 alone
(Scheme 2, pathway a) when the aldehyde is linear or in mixture
with an ester formate 12 (Scheme 2, pathway b) in the case of
chemical structure, i.e., on the substituents in
a position of the al-
dehyde function, the amount and the distribution of these minor

C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
2273
in radical concentration and could be only observed for aliphatic
and alicyclic acyl radicals.27
R3 R1
R3 R1
R3 R1
H
H
H
H abstraction
R4
C
H
C
R2
O O H
R4
C
H
C
R2
O O
R4
C
H
C
R2
O O
The alkyl radical Rꢀ formed can react according to two main
routes. The first one consists in trapping oxygen according to
a diffusion controlled process providing an alkyl peroxy radical
ROOꢀ, which abstracts one hydrogen atom giving a hydroperoxide
ROOH 13 as an intermediate oxidation product. This hydroperoxide
can then further decompose in acidic media into shorter alcohols,
ketones or aldehydes according to the Hock rearrangement
(Scheme 4).28 These first generation oxidation products are likely to
undergo the same oxidative degradation process as the initial al-
dehyde providing oxidation products of second generation. How-
ever, these minor products are at the limit of the GC/MS detection
and they could not always be detected. It is noteworthy that the
formation of radicals like ROOꢀ, which are less reactive towards
aldehydic hydrogen than R(CO)OOꢀ is in agreement with a lower
rate of oxygen consumption for bourgeonal 1. According to Denisov
et al., acyl peroxy radicals are more active in this chain propagation
reactions compared to peroxy radicals since the BDE of the
resulting peracid is higher than that of the hydroperoxide (z92.6
and 87.3 kcal molꢁ1, respectively).7 In this way, the difference in the
BDE values between the initial step (aldehyde and radicals) and the
final step (carbonyl radical and peracid or hydroperoxide) is lower
and thus, the process is favoured. This unexpected decrease of the
Hydroperoxide 13
Peroxyl radical
- H2O
c)
a)
b)
R3
C
R3 R1
C
H
R1
C
R2
R3
R4
C
O
R1
R4
C O R2
O
O
C
H
R4
H
R2
+ H2O
3 OH
+ H2O
+ H2O
3 R1
R1
C
R2
R3
C
H
R
C
H
R
C
H
HO
R4
R4
C
R2
O
R1
R4
C O R2
OH
H
H
H
C
R1
C
R2
R3
C
R3
C
H
O
O
O
R4
C
R4
R2
R1
H
Aldehyde 15 if R2 or R1 = H or Ketone 16
R3
C
HO R1
HO R2
HO
R4
H
Alcohol 17
Scheme 4. Hock rearrangement of hydroperoxides in acidic medium.
- CO
x 2 (termination)
R
R3 R1
O
O
O
O
H
O
propagation
initiation
R
C
R
C
R C
R4
C
H
C
R2
C
R
C
R
R OOH
R
OO
O
OO
OOH
CO2
O2
HO
O2
Alkyl
Acyl peroxy
radical
Peracid 9
Hydroperoxide 13
Aldehyde 1-7
Acyl
radical
Acyl oxy
radical
Peroxyl
radical
radical
Hock
Baeyer Villiger
rearrangement
(Scheme 3)
O
H
rearrangement
H
R
C
(Scheme 5)
Cage
reaction
5a)
5b) or 5c)
3b)
4)
3a)
OH
R3 CH R4
R3
C
H
O
O
O
H
R
C
R
O
C
R
OH
R4
C R1 (or R2)
OH
Acid 11
Ester formate 12
Alcohol 14
Alcohol 17
Aldehyde 15 if R1= R2 = H
or Ketone 16
Scheme 5. Overview of the different mechanisms and potential oxidation products involved in the autoxidation of fragrant aldehydes.
substituted alkyl aldehydes. According to an alternative minor
pathway, peracid 9 may also suffer a homolytic cleavage giving HOꢀ
and the acyloxy radical R(CO)Oꢀ, which may also be generated
through the termination reaction (6) involving two acyl peroxy
radical R(CO)OOꢀ.24 Then, R(CO)Oꢀ easily looses CO2 providing
the alkyl radical Rꢀ because this step is energetically favourable
propagation rate constant has been observed by Zaikov et al. who
noticed that the oxidation rate for benzaldehyde was lower in the
presence of tetralin and its hydroperoxides due to the participation
of the peroxy radicals in the oxidation chain.8 The formation of
alkyl radicals decreases the oxidability of the compounds and leads
to the formation of other products.
Another pathway explaining the formation of alcohols 14 is the
reaction between Rꢀ and the very reactive HOꢀ radical according to
a cage reaction since they are formed simultaneously according to
the process:
(kz1010
s
ꢁ1).25,26
Indeed, aliphatic peracids are unstable and can loss a hydroxyl
radical HOꢀ leading to the acyloxy radical, which then de-
carboxylates to give an alkyl radical Rꢀ and CO2. It is noteworthy
that, as reported by Vinogradov and Nikishin,27 a loss of carbon
monoxide from the acyl radical RCOꢀ at the beginning of the au-
toxidation process can also take place, leading to the formation of
the alkyl radical Rꢀ but this pathway cannot be considered as sig-
nificant under our conditions. Indeed, this reaction is favoured
when there is no more oxygen in the medium leading to an increase
cage
RðCOÞOOH/R, þ HO, þ CO2
ROH
(9)
reaction
Finally, a pathway consisting in the homolytic cleavage of hy-
droperoxide leads to the very reactive alkoxy radicals ROꢀ, which
form alcohols 14 after hydrogen abstraction.

2274
C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
3. Conclusion
methylpropanal (lilialÒ) 3, 2-2-dimethyl-3-(2-ethylphenyl) propanal
(floralozoneÒ) 4 were provided by IFF. Pentanal (valeraldehyde) 5 and
2-methyl-pentanal (2-methyl-valeraldehyde) 6 and 2-ethyl hexanal 7
were from SigmaeAldrich. All products were purified on basic alu-
minium oxide (solvent: AcOEt/petrol ether¼20/80) and distilled un-
der reduced pressure (10ꢁ2 bar)beforeuseinorder toremovepossible
traces of antioxidants.
Autoxidation of fragrant aldehydes results from a complex
process depending on many intrinsic parameters, e.g., aldehyde
structure, solvent, light, temperature, etc. and leading to a wide
variety of oxidation products. The measurement of the rate of ox-
ygen consumption clearly shows that the linear aldehydes are the
more sensitive towards oxidation and that an alkyl substitution in
a
position of the aldehydic function speeds up the autoxidation
4.2. Gas chromatographyemass spectrometry
process. Identification of the products by GC/MS led us to propose
a series of pathways, which explain the formation of major but also
minor products, of first and second generation. The two main non-
radical intermediates are the peracid RC(O)OOH 9 and the hydro-
peroxide ROOH 13 formed by the reaction of dioxygen with the
alkoxy radical RC(O)ꢀ and the alkyl radical Rꢀ, respectively. The first
one reacts with the aldehyde according to the BaeyereVilliger
mechanism forming acid or formate ester while the second one is
subject to the Hock’s rearrangement in acidic media forming many
minor products like alcohols, ketones and shorter aldehydes.
Structural parameters play an important role in the favoured
pathways and as a consequence, in the nature of the oxidation
products. As far as linear aldehydes are concerned, the oxidative
degradation is limited in terms of minor products, which essentially
result from the Hock rearrangement. In addition, no substituent in
A Thermofisher GC Trace equipped with an AI 3000 injector
connected to Thermofisher DSQ II simple quadrupole detector was
used for the GCeMS analysis. Compound separation was achieved
on a 30 m, DB5MS with 0.25 mm i.d. and 0.25 mm film thickness gas
chromatographic column (J&W Scientific, Folsom, CA, USA). Carrier
gas (ultra-pure helium) flow rate is 1.0 ml/min and the injector, the
transfer line and the ions source were maintained at 250, 250 and
220 ꢂC, respectively. The MS detector was used in the EI mode with
an ionizationvoltage of 70eV. Thecolumnwas held at 50 ꢂC for 5 min
and then programmed at 10 ꢂC minꢁ1 to 250 ꢂC and maintained for
10 min. The compounds were injected in the Split mode with a ratio
of 12. The NIST 2008 was used to identify the chemical compounds.
4.3. Autoxidation of aldehydes by oxygen
a
position contributes to form quantitatively the corresponding acid
whereas longer chain length and hindrance of a substituent in
position favour the formation of the formate ester. On the other
5 mL of a solution of aldehyde (2 mol Lꢁ1) in octane were oxi-
dized at 25 and 40 ꢂC under an atmosphere of pure dioxygen
(300 kPa) in the cell of a PetroOxy apparatus (Petrotest) (Scheme 6).
a
hand, numerous minor products are detected in the case of aromatic
aldehydes resulting from the Hock rearrangement and a cage re-
action between the hydroxyl radical HOꢀ and the alkyl radical. Theses
two pathways are significant when there is no substituent between
the aldehydic function and the aromatic ring. As in the case of linear
aldehyde, formate esters are also more important when alkyl sub-
stituents are present in
a position.
Monitoring the autoxidation of a fragrance molecule in a reactor
filled with pure oxygen under pressure is a relevant method to
mimic the behaviour of fragrances during the crash test, which is
commonly used as a practical test to assess the perfume stability
under defined and drastic conditions (e.g., 2 weeks at 60 ꢂC under
sunlight). This study shows that under such accelerated conditions,
the oxidation pathway of degradation is not modified and reactions
at this temperature is representative of reactions taking place at
room temperature, for which the decarbonylation is a minor pro-
cess. The method consisting in monitoring the oxygen consumption
can also be applied to other odorant organic compounds, such as
terpenes, which are also very sensitive towards oxidation. Com-
parison of oxidation rates of various substrates provides a quanti-
tative indicator of the oxidability. This study is representative for
the autoxidation of organic molecules during the shelf-life of
a product and it is important to notice that the matrix in which the
perfume is introduced can also lead to different reactivity and re-
actions. Finally, other aldehydes would be of great interest re-
garding their stability towards oxidative degradation. For instance,
lyral is known to be one of the major allergenic fragrance com-
pounds and oxidation products or peroxo intermediates are prob-
ably involved in this phenomenon. However, it has not been
investigated in the present work since the additional olefinic
function would lead to parasitic reactions, which would decrease
the apparent oxidizability of the aldehyde function.29
Scheme 6. PetroOxy apparatus for measurement of oxygen consumption during the
autoxidation process (from Petrotest).
The PetroOxy apparatus was equipped with an inox cell (7, dim
20ꢄ40ꢄ26 cm) corresponding to a total volume of 25 mL in which
the sample was introduced (the recommended volume is 5 mL) at
ambient temperature. The cell was then closed by a screw cap (3)
and a safety hood (2), which was locked by a latch (1 and 8). The gas
was removed of the cell by the extraction gas connection (4) and
replaced by only pure dioxygen, which was injected through the
gas alimentation (5) at the pressure indicated on the interface
4. Experimental section
4.1. Materials
3-(4-tert-Butylphenyl)propanal (bourgeonalÒ) 1, 3-(3-isopro-
pylphenyl)butanal
(florhydralÒ)
2,
3-(4-tert-butylphenyl)-2-

C. Marteau et al. / Tetrahedron 69 (2013) 2268e2275
2275
€
€
4. (a) Backtorp, C.; Hagvall, L.; Borje, A.; Karlberg, A.-T.; Norrby, P.-O.; Nyman, G. J.
Chem. Theory Comput. 2008, 4, 101e106; (b) Christensson, J. B.; Johansson, S.;
screen (9) (300 kPa). The cell was then heated up to the tempera-
ture set (25 or 40 ꢂC) and the pressure was monitored every minute
by the pressure sensor (6). The experiment was ended when the
pressure reached 50% of the maximum pressure. The pressure de-
crease was converted into a concentration of oxygen per volumes of
the aldehyde solution. The products formed were analyzed by
GC/MS.
€
Hagvall, L.; Jonsson, C.; Borje, A.; Karlberg, A.-T. Contact Dermatitis 2008, 59,
344e352; (c) Hagvall, L.; Backtorp, C.; Svensson, S.; Nyman, G.; Borje, A.;
Karlberg, A.-T. Chem. Res. Toxicol. 2007, 20, 807e814; (d) Skold, M.; Borje, A.;
Harambasic, E.; Karlberg, A.-T. Chem. Res. Toxicol. 2004, 17, 1697e1705; (e)
Skold, M.; Borje, A.; Matura, M.; Karlberg, A.-T. Contact Dermatitis 2002, 46,
267e272.
€
€
€
€
€
€
5. Schultz, H. W.; Day, E. A.; Libbey, L. M. In Chemistry and Technology of Flavors
and Fragrances; Rowe, D., Ed.; Blackwell: 2004.
6. Maslow, S.; Bluymberg, E. Russ. Chem. Rev. 1976, 45, 155e167.
7. Denisov, E. T.; Afanas’ev, I. B. Oxidation and Antioxidants in Organic Chemistry
and Biology; Taylor & Francis: New York, NY, 2005.
4.4. Bond dissociation energy (BDE) calculation10
8. Zaikov, G. E.; Howard, J. A.; Ingold, K. U. Can. J. Chem. 1969, 47, 3017e3029.
9. The Chemistry of Fragrances: From Perfumer to Consumer; Sell, C. S., Ed.; 2006
Cambridge, UK.
All the calculations were performed by using Gaussian 03
packages. The geometries of all the parent molecules were firstly
optimized by using the PM3 method and then the DFT one by using
the B3LYP/6-311G (d,p) basis set. Orbitals from this method were
used as inputs to the final RB3LYP/6-311Gþþ (2d, 2p) calculation.
The conformer with the lowest electronic energy was used in this
work. For radicals, the optimization also used the PM3 step plus the
final UB3LYP/6-311G (d, p) method. The orbitals were then used as
inputs to the final UB3LYP/6-311Gþþ (2d, 2p) calculation. Calcu-
lations were performed in heptane as a solvent.
10. Uchimaru, T.; Chandra, A. K.; Sugie, M. Recent Res. Dev. Org. Chem. 2001, 5,
165e179.
ꢀ ꢀ
11. Sajus, L.; Seree de Roch, I. Compr. Chem. Kinet. 1980, 16, 89e124.
12. McNesby, J.; Heller, C. Chem. Rev. 1954, 54, 325e329.
13. (a) Blacet, F. E. J. Phys. Colloid Chem. 1948, 52, 534e545; (b) Chatgilialoglu, C.;
Crich, D.; Komatsu, M.; Ryu, I. Chem. Rev. 1999, 99, 1991e2069.
14. Hermans, I.; Jacobs, P. A.; Peeters, J. Chem.dEur. J. 2006, 12, 4229e4240.
15. Krow, G. R. In 5.1 The BaeyereVilliger Reaction; Trost, B. M., Fleming, I., Eds.;
Comprehensive Organic Synthesis, Selectivity, Strategy and Efficiency in
Modern Organic Chemistry; 1991; Vol. 7.
16. Roberts, B. P.; Steel, A. J. J. Chem. Soc., Perkin Trans. 1 1994, 2, 2155e2162.
17. Pisarenko, L. M.; Krugovov, D. A.; Shchegolikhin, A. N.; Kasaikina, O. T. Russ.
Chem. Bull. 2008, 57, 83e89.
18. Niki, E.; Saito, T.; Kawakami, A.; Kamiya, Y. J. Biol. Chem. 1984, 259, 4177e4182.
19. (a) Pratt, D. A.; Mills, J. H.; Porter, N. A. J. Am. Chem. Soc. 2003, 125, 5801e5810;
(b) Korcek, S.; Chenier, J. H. B.; Howard, J. A.; Ingold, K. U. Can. J. Chem. 1972, 50,
2285e2297.
20. McDonough, E. G. Columbia University, Dissertation, New York, 1932.
21. Lehtinen, C.; Nevalainen, V.; Brunow, G. Tetrahedron 2001, 57, 4741e4751.
22. Li, J. J. Name Reactions; Springer: Berlin, 2009.
23. Lehtinen, C.; Nevalainen, V.; Brunow, G. Tetrahedron 2000, 56, 9375e9382.
24. Sorba, J.; Lefort, D.; Fossey, J.; Grisel, F.; Sanderson, W. R. J. Chem. Soc., Perkin
Trans. 2 1992, 23, 1565e1569.
25. Kaptein, R. In Adv. Free-radical Chemistry; Williams, G. H., Ed.; Elek Science:
London, 1975; vol. 5, pp 319e380.
26. Hermans, I.; Jacobs, P.; Peeters, J. Chem.dEur. J. 2007, 13, 754e761.
27. Vinogradov, M. G.; Nikishin, G. I. Russ. Chem. Rev. 1971, 40, 916e932.
28. Wang, D. Z. Comprehensive Organic Name Reactions and Reagents; J. Wiley &
Sons: New York, NY, 2009.
29. Wentzel, B. B.; Alsters, P. L.; Feiters, M. C.; Nolte, J. M. R. J. Org. Chem. 2004, 69,
3453e3464.
Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.tet.2013.01.034.
References and notes
1. (a) Nicoll, S. Cosmet. Toiletries 1999, 114, 59e63; (b) Geria, N. Cosmet. Toiletries
1990, 105, 41e45.
€
2. (a) Christensson, J. B.; Forsstrom, P.; Wennberg, A.-M.; Karlberg, A. T.; Matura, M.
€
€
Contact Dermatitis 2009, 60, 32e40; (b) Karlberg, A.-T.; Bergstrom, M. A.; Borje,
A.; Luthman, K.; Lars, J.; Nilsson, G. Chem. Res. Toxicol. 2008, 21, 53e69.
ꢀ
3. (a) Elberling, J.; Linneberg, A.; Mosbech, H.; Dirksen, A.; Menne, T.; Nielsen, N. H.;
Madsen, F.; Frølund, L.; Duus Johansen, J. Contact Dermatitis 2005, 52, 65e72; (b)
Baur, X.; Schneider, E. M.; Wieners, D.; Czuppon, A. B. Allergy 1999, 54,1334e1335;
(c) Bridges, B. Flavour Fragrance J. 2002, 17, 361e371.