Palladium catalyzed oxidation of biorenewable β-citronellol and geraniol for the synthesis of polyfunctionalized fragrances

Article abstract of DOI:10.1016/j.mcat.2021.111449

β-Citronellol and geraniol are bio-renewable compounds found in essential oils of various plants. Allylic oxidation of trisubstituted sterically highly encumbered carbon-carbon double bonds in their molecules with dioxygen can be efficiently catalyzed by the palladium acetate/p-benzoquinone combination under 5?10 atm of dioxygen pressure without the need for conventionally used metal co-catalysts. Both substrates gave allyl acetates in ca. 90% combined yield. The oxidation of geraniol proceeded with excellent site-selectivity involving only one of its two C[dbnd]C bonds. Geraniol gave both linear and branched allyl acetates in comparable yields, whereas the reaction of β-citronellol showed high regioselectivity for the corresponding branched product. Polyfunctionalized oxygenated terpenes produced from β-citronellol and geraniol are of a potential importance as ingredients for fragrance compositions.

Full text of DOI:10.1016/j.mcat.2021.111449

Molecular Catalysis 504 (2021) 111449
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.journals.elsevier.com/molecular-catalysis
Palladium catalyzed oxidation of biorenewable β-citronellol and geraniol
for the synthesis of polyfunctionalized fragrances
Aldino N. Venancioa, Luciano Meninib, Daiane N. Marondea, Elena V. Gusevskayac, Luciana
A. Parreiraa *
,
a
ˆ
Universidade Federal do Espírito Santo, Centro de Ciencias Exatas, Naturais e da Saúde, Departamento de Química e Física, 29500-000, Caixa Postal 16, Alegre, ES,
Brazil
b
˜
ˆ
Instituto Federal de Educaçao, Ciencia e Tecnologia do Espírito Santo, Campus de Alegre, 29500-000, Caixa Postal 47, Alegre, ES, Brazil
c Universidade Federal de Minas Gerais, Departamento de Química, 31270-901, Belo Horizonte, MG, Brazil
A R T I C L E I N F O
A B S T R A C T
Keywords:
β-Citronellol and geraniol are bio-renewable compounds found in essential oils of various plants. Allylic
oxidation of trisubstituted sterically highly encumbered carbon-carbon double bonds in their molecules with
dioxygen can be efficiently catalyzed by the palladium acetate/p-benzoquinone combination under 5ꢀ 10 atm of
dioxygen pressure without the need for conventionally used metal co-catalysts. Both substrates gave allyl ace-
Biorenewables
Geraniol
β-citronellol
Allylic oxidation
Palladium
tates in ca. 90% combined yield. The oxidation of geraniol proceeded with excellent site-selectivity involving
–
only one of its two C C bonds. Geraniol gave both linear and branched allyl acetates in comparable yields,
–
whereas the reaction of β-citronellol showed high regioselectivity for the corresponding branched product.
Polyfunctionalized oxygenated terpenes produced from β-citronellol and geraniol are of a potential importance
as ingredients for fragrance compositions.
1. Introduction
compounds, Cu(OAc)2, and special ligands able to keep palladium(0)
complexes in solutions [8–17].
Terpenoid compounds are commercially available in large amounts
from renewable biomass sources [1]. Due to pleasant organoleptic
properties, many of them have found direct applications in various
commercial products. Chemical transformations of natural terpenoids
into their derivatives with multiple oxygenated functions through cat-
alytic processes is a promising way for the production of new
value-added molecules of interest for the fine chemicals industry, mainly
to be exploited as drugs and new flavors [2–5].
BQ is an excellent reagent to regenerate palladium in Wacker type
oxidative processes; however, the reaction of its reduced form, p-hy-
droquinone (BQH2), with dioxygen at 1 atm is extremely slow without
co-catalysts, e.g., copper, manganese and cobalt salts [12,17]. In the
previous studies, we showed that dioxygen can effectively regenerate
BQ directly at 5–10 atm, without the need for redox auxiliaries [18].
Based on these results, we have further developed several efficient
processes to oxidize a number of biomass-based substrates with different
–
One of the processes used industrially for the oxidation of alkenes is
the well-known Wacker process which involves dioxygen as the final
oxidant and PdCl2 and CuCl2 as catalyst and co-catalyst, respectively [6,
7]. This catalytic system is very convenient to convert small terminal
olefins; however, with larger chain olefins or internal olefins the system
selectivity decreases. Such a problem occurs because of the substrate
isomerization caused by CuCl2 and the appearance of chlorinated
by-products caused by chloride ions, which are needed in a large amount
to keep palladium complexes in solutions. Several alternatives to replace
CuCl2 have been proposed, e.g., p-benzoquinone (BQ), heteropoly
types of the C C bonds. [3,19–23]. It has been found that the Pd/BQ
–
system works well for the allylic oxidation of trisubstituted and disub-
–
stituted endocyclic C C bonds [19–22]. 1-Octene was also readily
–
oxidized in this system, albeit with poor selectivity due to the significant
–
double bond isomerization [19]; whereas disubstituted terminal C
C
–
bonds in limonene and caryophyllene did not react at all under similar
conditions [19,22]. The reactivity of alkenes strongly depended on the
substrate structure and other functional groups present in the molecule.
–
–
For example, a tetrasubstituted endocyclic C C bond in β-ionone
reacted with palladium to give selectively an allylic oxidation product,
* Corresponding author.
E-mail addresses: lucianap.ufes@gmail.com, luciana.parreira@ufes.br (L.A. Parreira).
https://doi.org/10.1016/j.mcat.2021.111449
Received 8 October 2020; Received in revised form 26 January 2021; Accepted 1 February 2021
Available online 15 February 2021
2468-8231/© 2021 Elsevier B.V. All rights reserved.

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
–
whereas the other C C bond in the β-ionone molecule (the acyclic
(17) [M+ - HOAc - H2O - CH3], 111 (10), 109 (11), 95 (12), 93 (15), 84
(15), 81 (22), 79 (13), 72 (11), 71 (18), 69 (24), 68 (19), 67 (25), 55
(21), 44 (12), 43 (100) (first isomer); 154 (4) [M+ - HOAc], 139 (11))
[M+ꢀ - HOAc - CH3], 136 (1) [M+ - HOAc - H2O], 121 (17) [M+ - HOAc -
H2O - CH3], 111 (10), 109 (11), 95 (12), 93 (15), 84 (16), 81 (21), 79
(13), 72 (11), 71 (18), 69 (24), 68 (18), 67 (24), 55 (23), 44 (10), 43
(100) (second isomer).
–
–
–
disubstituted bond with the C O group in allylic position) remained
–
–
intact [21]. On the other hand, a tetrasubstituted endocyclic C C bond
in cis-jasmone was not involved in the oxidation, whereas its acyclic
–
–
disubstituted C C bond reacted to give allylic oxidation products [23].
In the present project, we directed our efforts to the oxidative func-
tionalization of trisubstituted acyclic carbon-carbon double bonds in
βꢀ citronellol and geraniol, two monoterpenic alkenyl alcohols.
βꢀ Citronellol and geraniol are both natural compounds available
from several essential oils extracted from plants, such as citronella, rose
and palmarosa oils [24,25]. These monoterpenes have found many ap-
plications as flavor components in detergents, drugs and perfumes and
are also used in insect repellents [24,25]. We do believe that oxy-
functionalization of the βꢀ citronellol and geraniol molecules can open
new possibilities for their commercial uses.
Compound 1b``(two isomers): MS (70 eV, EI): m/z (%): 136 (6) [M+ -
2HOAc], 121 (52) [M+ - 2HOAc - CH3], 93 (16), 84 (12), 81 (16), 79
(10), 69 (15), 68 (11), 67 (12), 55 (17), 43 (100) (first isomer); 136 (6)
[M+ - 2HOAc], 121 (50) [M+ꢀ - 2HOAc - CH3], 93 (16), 84 (12), 81 (16),
79 (10), 69 (15), 68 (11), 67 (12), 55 (17), 43 (100) (second isomer); 1H
3
◦
NMR (400 MHz, CDCl3, 25 C, Me4Si): δ = 0.90 (d, J = 6.0 Hz, 3 H;
C10H3), 1.70 (s, 3 H; C9H3), 2.03 (s, 3 H; OCOCH3), 2.05 (s, 3 H;
OCOCH3), 1.00–1.30 (m, 3 H; C2H2), 1.30–1.70 (m, 5 H; C3H, C4H2 and
C5H2), 4.07 (t, 3J = 8.0 Hz, 2 H; C1H2), 4.88 (s, 1 H; C8HH), 4.93 (s, 1 H;
C8HH), 5.12 ppm (t, 3J = 6.0 Hz, 1 H; C6H); 13C NMR (100 MHz, CDCl3,
25 ◦C, TMS): δ = 17.09 and 18.06 (C9), 19.32 (C10), 20.95 (OCOCH3),
21.18 (OCOCH3), 29.58 and 29.60 (C3), 29.80 and 29.82 (C5), 32.18 and
32.26 (C2), 35.31 and 35.33 (C4), 62.68 (C1), 77.50 and 77.52 (C6),
112.73 and 112.98 (C8), 143.06 (C7), 170.32 (OCOCH3), 171.17 ppm
(OCOCH3). Atom numbering is given in Scheme 1.
Herein, we report for the first time the palladium/p-benzoquinone
catalyzed oxidation of carbon-carbon double bonds in βꢀ citronellol and
geraniol molecules with dioxygen. The reactions produced poly-
oxygenated products with fragrance characteristics, that would be
difficult to obtain by conventional methods of organic synthesis.
The oxidative transformation of carbon-carbon double bonds in both
βꢀ citronellol and geraniol is a difficult task because the bonds are
trisubstituted and sterically highly encumbered. Internal alkenes which
do not contain strongly activating functionalities usually show low
Compound 1c: MS (70 eV, EI): m/z (%): 136 (11) [M+ - 2HOAc], 121
(62) [M+ - 2HOAc - CH3], 107 (11), 94 (11), 93 (21), 84 (17), 81 (21), 79
(15), 69 (10), 68 (11), 67 (14), 55 (18), 43 (100); 1H NMR (400 MHz,
CDCl3, 25 ◦C, Me4Si): δ = 0.92 (d, 3J = 6.0 Hz, 3 H; C10H3), 1.65 (s, 3 H;
C9H3), 2.06 (s, 3 H; OCOCH3), 2.08 (s, 3 H; OCOCH3), 1.00–1.30 (m, 2 H;
C2H2), 1.30–1.70 (m, 5 H; C3H, C4H2 and C5H2), 4.07 (t, 3J = 8.0 Hz, 2 H;
C1H2), 4.44 (s, 2 H; C8H2), 5.43 ppm (t, 3J = 6.6 Hz, 1 H; C6H); 13C NMR
(100 MHz, CDCl3, 25 ◦C, TMS): δ = 14.10 (C9), 19.42 (C10), 21.09
(OCOCH3), 21.29 (OCOCH3), 25.19 (C5), 29.40 (C2), 29.65 (C3), 36.43
(C4), 62.49 (C1), 70.15 (C8), 129.82 (C6), 130.08 (C7), 170.44
(OCOCH3), 171.21 ppm (OCOCH3). Atom numbering is given in Scheme
1.
reactivity in the oxidations of Wacker type and high tendency to the
–
isomerization of the C C bond [26]. Geraniol and βꢀ citronellol were
–
used as model molecules in many studies on the oxidation of alcohols by
dioxygen catalyzed by palladium [27–34]. In all these works, the
oxidation involved only the hydroxyl groups to give corresponding al-
–
–
dehydes, whereas the C C bonds in both geraniol and βꢀ citronellol
remained intact. We found only one report on the palladium catalyzed
oxidation of these substrates involving the olefinic moiety: oxidative
cyclization of βꢀ citronellol to give an 8-membered heterocyclic com-
pound using H2O2 as the final oxidant [35].
Compound 1d (rose oxide): major isomer: MS (70 eV, EI): m/z (%):
154 (9) [M+], 139 (100) [M+ - CH3], 85 (13), 83 (30), 69 (84), 67 (13);
minor isomer: 139 (100) [M+ - CH3], 85 (11), 83 (29), 69 (90), 67 (12).
Compound 1e: MS (70 eV, EI): m/z (%): 152 (26) [M+ - HOAc], 137
(100) [M+ - HOAc - CH3], 109 (12), 99 (10), 97 (12), 93 (11), 83 (33), 81
2. Experimental
The reagents and solvents were used without further treatment,
except for p-benzoquinone, which was purified by re-sublimation.
(±)-βꢀ Citronellol (3,7-dimethyl-6-en-1-ol) and geraniol (trans-3,7-
dimethyl-2,6-octadien-1-ol) were from Sigma-Aldrich. To perform the
reactions, two types of reactors were used: a three necks round bottom
glass flask with a rubber septum for taking aliquots (at 1 atm) or a 100
mL 316 stainless steel autoclave with a sampler for periodic removal of
aliquots (at 5–10 atm). The glass reactor was supplied by a burette to
monitor the dioxygen uptake. The reactors were put under magnetic
stirring in a thermal bath with fixed temperatures between 70 and 90 ◦C.
Periodically, the samples from reaction mixtures were taken. The reac-
tion development was followed by gas chromatography (GC) on a Shi-
madzu equipment (GC2010-Plus) using a Rtx®-5MS capillary column
and FID detector. The amounts of the reacted substrate were considered
to calculate conversions and selectivities. Dodecane was used as internal
standard for the control of a mass balance (0.10 M, added at the reaction
beginning together with the substrate). The relative order of elution of
the products derived from citronellol (1a) was as follows: 1d, 1a, 1b`,
1e, 1b``, 1c. and those derived from geraniol (2a): 2a, 2b`, 2c`, 2b``,
2c``.
1
(17), 79 (16), 71 (12), 69 (51), 67 (16), 55 (66); H NMR (400 MHz,
CDCl3, 25 ◦C, Me4Si): δ = 3.40–3.50 (m, 1 H; C1HH), 3.90–4.10 (m, 2 H;
C1HH and C5H), 4.44 (s, 2 H; C8H2), 5.50–5.60 ppm (m, 1 H; C6H); 13
C
NMR (100 MHz, CDCl3, 25 ◦C, TMS): δ = 67.59 (C1), 68.92 (C8), 73.59
(C5), 126.13 ppm (C6). Atom numbering is given in Scheme 2.
Compound 2b`: MS (70 eV, EI): m/z (%): 119 (38) [M+ - HOAc –
H2O- CH3], 93 (13), 91 (9), 84 (16), 81 (9), 79 (10), 71 (13), 69 (21), 68
(11), 67 (22); 44 (38), 43 (100).
Compound 2b``: MS (70 eV, EI): m/z (%): 134 (14) [M+ - 2HOAc],
119 (38) [M+ - 2HOAc - CH3], 105 (10), 93 (13), 91 (12), 84 (35), 81
(11), 79 (11), 68 (11), 67 (16); 43 (100); 1H NMR (400 MHz, CDCl3, 25
◦C, Me4Si): δ = 1.67 (s, 3 H; C9H3 or C10H3), 1.68 (s, 3 H; C9H3 or C10H3),
2.01 (s, 3 H; OCOCH3), 2.03 (s, 3 H; OCOCH3), 1.70–2.10 (m, 4 H; C4H2
and C5H2), 4.57 (d, 3J = 7.0 Hz, 2 H; C1H2), 4.86 (s, 1 H; C8HH), 4.91(s, 1
H; C8HH), 5.10 (t, 3J = 6.6 Hz, 1 H; C6H); 5.33 ppm (t, 3J = 7.0 Hz, 1 H;
C2H); 13C NMR (100 MHz, CDCl3, 25 ◦C, TMS): δ = 16.39 (C10), 19.33
(C9), 21.03 (OCOCH3), 21.04 (OCOCH3), 30.50 (C5), 34.96 (C4), 61.31
(C1), 77.50 (C6), 112.00 (C8), 118.38 (C2), 141.42 (C7), 142.74 (C3),
171.20 (OCOCH3), 171.21 ppm (OCOCH3). Atom numbering is given in
Scheme 3.
Reaction products were separated by neutralizing reaction solutions
(total volume 12 mL) with NaHCO3 (18 g) and extracting twice with 20
mL of diethyl ether. The separation of individual products from the
mixture was realized by a silica gel column chromatography (silica gel
60, eluents: dichloromethane and hexane). The products were identified
by GC-MS (Shimadzu QP2010-PLUS equipment, 70 eV) and NMR
spectroscopy (Bruker 400 MHz equipment, CDCl3, TMS).
Compound 2c` (8-geranyl acetate): MS (70 eV, EI): m/z (%): 119 (5)
[M+ - HOAc – H2O- CH3], 93 (5), 84 (23), 69 (15), 68 (21), 67 (25); 44
(22), 43 (100).
Compound 2c`` (1,8-geranyl diacetate): MS (70 eV, EI): m/z (%): 134
(15) [M+ - 2HOAc], 119 (16) [M+ꢀ - 2HOAc - CH3], 85 (13), 84 (24), 68
(11), 67 (14), 43 (100); 1H NMR (400 MHz, CDCl3, 25 ◦C, Me4Si): δ =
1.62 (s, 3 H; C9H3), 1.67 (s, 3 H; C10H3), 2.01 (s, 3 H; OCOCH3), 2.03 (s, 3
Compound 1b`(two isomers): MS (70 eV, EI): m/z (%): 154 (4) [M+ -
HOAc], 139 (11) [M+- HOAc - CH3], 136 (1) [M+ - HOAc - H2O], 121
2

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
Scheme 1. Transformation of β-citronellol (1a).
Scheme 2. Minor products formed from β-citronellol (1a).
Scheme 3. Transformation of geraniol (2a).
H; OCOCH3), 1.70–2.10 (m, 4 H; C4H2 and C5H2), 4.44 (br.s, 2 H; C8H2),
3. Results and discussion
4.57 (d, 3J = 7.0 Hz, 2 H; C1H2), 5.42 ppm (br.t, 2 H; C2H and C6H); 13
C
NMR (100 MHz, CDCl3, 25 ◦C, TMS): δ = 14.15 (C9), 16.41 (C10), 21.03
(OCOCH3), 21.04 (OCOCH3), 25.80 (C5), 38.79 (C4), 61.24 (C1), 70.05
(C8), 122.17 (C2), 129.06 (C6), 130.38 (C7), 141.54 (C3), 171.07
(OCOCH3), 171.06 ppm (OCOCH3). Atom numbering is given in Scheme
3.
The interaction of β-citronellol (1a) and geraniol (2a) molecules with
palladium can proceed through the OH group and through the olefinic
–
–
moieties - the internal trisubstituted C C bonds. In most of previous
works it was reported that palladium promoted in these substrates the
oxidation of the OH groups to form corresponding aldehydes, whereas
–
the C C bonds stayed intact [27–34]. In the present study, we have
–
applied a palladium/p-benzoquinone catalytic combination aiming to
–
–
perform the Wacker type oxidation of the C C bonds in β-citronellol
3

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
and geraniol molecules with dioxygen.
isomers with virtually the same mass spectra. Two GC distinguishable
peaks were observed for both 1b′ and 1b′′. Considering that new
asymmetric center (C-6) appeared when products 1b′ and 1b′′ were
formed and that a racemic mixture of (3R) and (3S)-β-citronellol was
employed as the starting material, the two GC peaks observed should
belong to two enantiomeric pairs of diastereoisomers with different
configurations at asymmetric carbons C-3 and C-6. The NMR spectra of
the isomers of major acetate 1b′′ were slightly different. A geometry at
the C-6 – C-7 double bond in diacetate 1c is predominantly trans (ca.
95% trans). Compounds 1b′′ and 1c have been reported previously as
the products of the high temperature pyrolysis of 3,7-dimethylocta-
1,6,7-tri acetate [36].
3.1. Oxidation of β-citronellol
We started the catalytic studies monitoring the behavior of β-citro-
nellol in the solutions of acetic acid under elevated temperatures
without the addition the palladium salt. Partial esterification of
β-citronellol with acetic acid occurred in the reaction media: nearly 70%
of β-citronellol was converted in citronellyl acetate for 5 h at 80 ◦C
(Table 1, run 1). To calculate conversion and selectivity values, cit-
ronellyl acetate was counted together with non-converted substrate
because our aim was to evaluate the oxidative transformations of
β-citronellol.
The oxidation of β-citronellol by BQ gave acetates 1b and 1c with
80% total selectivity. In addition, two minor products 1d and 1e were
identified in the reaction mixtures, both originated from the cyclization
of the β-citronellol molecule (Scheme 2). Product 1d is known as rose
oxide, a natural compound occurring in essential oils from several rose
species and commonly used in fine perfumery [24]. Considering its
structure, rose oxide 1d arises from the allylic oxidation of β-citronellol,
similarly to 1b and 1c; however, the internal OH group rather than
acetic acid is involved in its formation. Compound 1e detected at
advanced reaction times arises from the oxidation of rose oxide 1d at the
allylic position, i.e., it is the product of the double-oxidation of β-citro-
nellol. As compound 1e was presented in the isolated product mixtures
in low concentrations only, its structure was suggested considering
characteristic NMR signals, mass spectrum and GC retention time. It is
important that all isolated products and also their mixtures show a nice
woody-floral scent and therefore can be applied directly in flavor
compositions.
The reaction of β-citronellol with 1.5 equivalents of BQ in the pres-
–
ence of 5 mol% of Pd(OAc) was found to give the products of the C
C
–
2
bond oxidation (Table 1, run 2). The process is stoichiometric in BQ
which acts as the final oxidant. No consumption of dioxygen was
detected during the reaction indicating that BQH2, formed at the
reduction of BQ, was not regenerated. The accumulation of BQH2 in
reaction mixtures could be accompanied by GC analyses.
The structure of main products depicted in Scheme 1 were confirmed
by NMR spectroscopy after their isolation from the reaction mixtures. All
–
main products were originated by the allylic oxidation of the C C bond.
–
Monoacetate 1b′ and diacetate 1b′′ were formed from β-citronellol and
citronellyl acetate, respectively, with the acetoxy group being incorpo-
rated at olefinic carbon C-6 in original molecules. Diacetate 1b′′ could
be also formed from monoacetate 1b′ through the esterification of its
hydroxyl group. Diacetate 1c is the linear isomer of branched allyl ac-
etate 1b′′, with the acetoxy group being incorporated into the terminal
allylic position at C-8.
Under 5 atm of dioxygen, the oxidation of β-citronellol readily pro-
ceeded with the use of BQ in catalytic amounts (Table 1, runs 3 and 4).
The formation of palladium metal was not observed during the reaction
indicating the successful regeneration of BQH2 directly with dioxygen.
The change in dioxygen pressure did not influence the conversion rate:
in two pairs of experiments, the reactions at 5 and 10 atm of dioxygen
showed approximately the same kinetics (Table 1, run 3 vs run 5; run 4
vs run 7). This suggests that the regeneration of BQH2 does not control
kinetically the whole catalytic reaction at the conditions employed in
these experiments. Although the reaction proceeded faster with higher
concentrations of BQ (Table 1, run 3 vs run 4; run 5 vs run 7), it is un-
likely that the oxidation of BQH2 is a slowest reaction step. The reactions
with 0.25 equivalents of BQ were slower than those with 0.5 equivalents
of BQ at early reaction stage already, when the interaction of BQH2 with
O2 could not limit yet the conversion of βꢀ citronellol. Moreover, no GC
peak characteristic for BQH2 was registered during the reactions. It is
well documented for related systems that BQ can coordinate to palla-
dium [12]; therefore, it appeared that BQ not only participates in the
oxidation of reduced Pd complexes but also can influence the reactivity
of Pd towards βꢀ citronellol. In addition, it was found that the Pd con-
centration positively affected the reaction rate (Table 1, cf. runs 5 and
6). Thus, we propose that the oxidation of βꢀ citronellol with Pd(II)
species is the slowest part of the overall process.
Each of acetates 1b′ and 1b′′ was formed as ca. 1:1 mixture of two
Table 1
Oxidation of β-citronellol (1a) with dioxygena.
Run
[BQ]
(M)
P
T
(oC)
Time
(h)
Cb
Selectivityb (%)
TOFc
(atm)
(%)
1b (1b′′
′/1b′′)
1c
1d
(hꢀ 1
)
1d
2
none
0.30
5
1
80
80
5
0
0
0
0
–
3
41
74
71 (62/
38)
4
25
17
5.0
12
14
66 (26/
74)
3
4
0.05
0.10
5
5
80
80
3
7
43
90
65 (37/
63)
10
18
25
13
4.4
6.0
67 (7/93)
69 (32/
68)
3
7
61
10
13
18
11
100
73 (10/
90)
5
0.05
10
80
3
9
45
99
70 (30/
70)
8
22
10
4.0
16
70 (1/99)
73 (3/98)
68 (40/
60)
6e
7
0.05
0.10
10
10
80
80
12
3
78
17
11
17
10
21
13
4.0
5.6
61
Branched allyl acetates 1b′ and 1b′′ were the major products of this
reaction, being formed with ca. 70% selectivity together with ca. 15% of
minor linear allyl acetate 1c. The relative contribution of diacetate 1b′′
as compared to monoacetate 1b′ increased during the process because of
the acetoxylation of the OH group in both βꢀ citronellol and mono-
acetate 1b′. At advanced reaction times, none or only small concentra-
tions of monoacetate 1b′ were detected. Monoacetate corresponding to
minor diacetate 1c was not formed in detectable concentrations.
Rose oxide 1d, formed by the oxidative cyclization of βꢀ citronellol,
was observed in higher proportions at early reaction stages, when
βꢀ citronellol concentrations were still high. At advanced reaction times,
the selectivity for 1d gradually decreased because of the esterification of
the βꢀ citronellol hydroxy group, a key player in the formation of 1d.
The molecule of citronellyl acetate did not undergo the cyclization.
7
100
69 (20/
80)
8
0.05
0.10
10
10
90
90
3
7
3
78
90
95
68 (7/93)
72 (7/93)
70 (7/93)
15
16
17
14
7
7.2
9.0
9
a
8
Conditions: [β-citronellol] = 0.20 M, [Pd(OAc)2] = 0.01 M, gas phase – O2,
solvent - acetic acid (glacial).
b
Conversions (C) and selectivities are based on β-citronellol reacted; cit-
ronellyl acetate was counted together with non-converted β-citronellol.
c
Turnover frequency (TOF): average rate of the conversion of β-citronellol
and citronellyl acetate per mol of Pd during the first reaction hour.
d
Without Pd(OAc)2; ca. 70% of 1a was transformed in citronellyl acetate for 5
h.
e
[Pd(OAc)2] = 0.005 M.
4

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
The oxidation of βꢀ citronellol promoted by Pd(OAc)2/BQ proceeds
faster and still selectively at 90 ◦C to give diacetates 1b′′ and 1c with
approximately 90% selectivity (Table 1, runs 8 and 9). It is worthwhile
to note an excellent regioselectivity of allylic oxidation: the OAc group is
incorporated with high preference (ca. 80%) at carbon C-6, the origi-
nally olefinic position in βꢀ citronellol, to give branched allyl products
1b′ and 1b′′.
genus, widespread through tropical and subtropical regions in Africa
and Asia [37]. No information was found about the natural occurrence
of compounds 2b ’and 2b”; however, both of them were mentioned as
synthetic intermediates [38,39]. Compound 2b’ was reported as an in-
termediate in the synthesis of enantiomeric aleuriaxanthines [38],
whereas compound 2b’’ in the synthesis antimalarials by photooxida-
tion of geraniol [39].
–
It should be mentioned that the trisubstituted C C bonds in the
–
geraniol and citronellol molecules reacted more slowly than disubsti-
3.2. Oxidation of geraniol
–
tuted C C bonds, either endocyclic or acyclic, under the same condi-
–
tions (in cyclohexene and cis-jasmone, respectively, studied in our
previous works [19,23]). No oxidative cyclization of geraniol (which
would give the product analogous to rose oxide 1d formed from
β-citronellol) was detected al all; probably, because of trans configura-
tion at the C-2 – C-3 bond in the geraniol molecule and, therefore, spatial
difficulty for the OH group to participate in the reaction at the C-6 – C-7
double bond. Also, differently from β-citronellol, geraniol gave small
amounts of geranial, the product resulted from the oxidation of its OH
group. Geranial was formed mainly at the reaction beginning when
significant amounts of geraniol were yet in its non-acetylated original
form. Because of the fast geraniol esterification, the contribution of
geranial gradually decreased with time, its selectivity being usually less
than 5% at the end of the process.
Geraniol is a primary allylic alcohol having the same structural
–
fragment, a trisubstituted C C bond, as βꢀ citronellol, and additionally
–
–
–
a second C C bond also trisubstituted, in γ position relatively to the first
one. All previously published studies regarding the Pd catalyzed
oxidation of geraniol report only the oxidation of its hydroxyl group into
the aldehyde moiety.
The reactions of geraniol (2a) in the solutions of HOAc containing Pd
(OAc)2 and BQ, both stoichiometric and catalytic regarding to BQ, gave
two main products numbered as 2b and 2c (Table 2). Both 2b and 2c
–
were formed at the oxidation of only one C C bond in the geraniol
–
–
molecule, whereas the other C C bond remained unreacted (Scheme 3).
–
In the experiment without the Pd salt added, no oxidation of geraniol
was registered at all (Table 2, run 1). Similarly to β-citronellol, geraniol
was gradually converted into geranyl acetate because of the esterifica-
tion of its OH group. To calculate conversion and selectivity values,
geranyl acetate was counted together with geraniol as non-converted
substrate because our aim was to follow specifically the oxidative
transformations.
The regioselectivity at the oxidation of geraniol was only slightly
influenced by the reaction variables, with the 2b/2c ratio varying be-
tween 4/6 and 6/4 at 70ꢀ 80 ◦C (Table 2). At higher temperature, more
sterically demanded branched isomer 2b, with the acetoxy group at the
internal C-6 carbon, was slightly more favorable: the 2b/2c ratio was 7/
3 at 90 ◦C (Table 2, run 10). The regioselectivity tendency in the
oxidation of geraniol differs significantly from that observed for
β-citronellol, where branched allyl acetate 1b (ca. 80%) was highly
preferred under all tested conditions (Table 1).
Products 2b and 2c appeared in the reaction mixtures as corre-
sponding monoacetates 2b′ and 2c′, which contain the hydroxy group
from the original geraniol, and diacetates 2b′′ and 2c′′, in which the
alcohol moiety was acetoxylated. At the reactions, monoacetates 2b′
and 2c′ were gradually converted into diacetates 2b′′ and 2c′′. In the
final reaction mixtures, diacetates 2b′′ and 2c′′ were usually the only
two products of allylic oxidation. In products 2b′ and 2b′′, the acetoxy
group is attached to the originally olefinic C-6 atom, while products 2c′
and 2c′′ - to the terminal originally allylic C-8 atom. Compounds 2c′ and
2c′′ occur in nature being isolated from flavoring plants of the Premna
The kinetic trends of the geraniol oxidation were analogous to those
observed with β-citronellol. It appears that the reaction of BQH2 with
dioxygen also did not control the rate of the overall catalytic process.
Variations in dioxygen pressure did not influence the reaction: in two
sets of the experiments, the reactions at 5 and 10 atm showed similar
kinetics (Table 2, run 3 vs run 5; run 4 vs run 8). Similarly to
Table 2
Oxidation of geraniol (2a) with dioxygena.
Run
[BQ]
P
T
Time
Conversionb
Selectivityb (%)
TOFc
(M)
(atm)
(oC)
(h)
3
(%)
0
2b (2b′/2b′′)
–
2c (2c′/2c′′)
–
(hꢀ 1
)
1d
2
0.20
0.30
5
1
80
–
80
3
64
96
42
98
58
98
38
100
100
43
94
55
99
50
67
100
100
80
38 (60/40)
39 (24/76)
50 (20/80)
58 (0/100)
43 (33/67)
52 (5/95)
54 (8/92)
60 (0/100)
51 (15/85)
48 (19/81)
56 (0/100)
46 (10/90)
53 (0/100)
41 (20/80)
60 (0/100)
67 (0/100)
62 (0/100)
45 (7/93)
60 (70/30)
60 (30/70)
45 (32/68)
38 (0/100)
49 (43/57)
44 (9/91)
36 (14/86)
36 (0/100)
36 (12/88)
42 (30/70)
40 (0/100)
46 (29/71)
42 (0/100)
51 (30/70)
32 (0/100)
28 (0/100)
32 (0/100)
27 (0/100)
5.2
10
3
3
4
5
0.05
0.10
0.05
5
80
80
80
4.0
5.6
4.0
12
3
5
7
10
3
12
7
6e
7
0.05
0.07
10
10
80
80
3.4
4.6
3
7
8
0.10
10
80
3
5.6
7
9
0.05
0.05
10
10
70
90
7
2.0
8.0
10
3
7
11f
0.05
0.05
10
10
80
80
7
2.0
8.0
12g
7
a
Conditions: [geraniol] = 0.20 M, [Pd(OAc)2] = 0.01 M, gas phase – O2, solvent - acetic acid (glacial).
b
c
d
e
f
Conversions (C) and selectivities are based on geraniol reacted; citronellyl acetate was counted together with non-converted geraniol.
Turnover frequency (TOF): average rate of the conversion of geraniol and geranyl acetate per mol of Pd during the first reaction hour.
Without Pd(OAc)2; 98% of 2a was transformed in geranyl acetate for 3 h.
[Pd(OAc)2] = 0.02 M.
[geraniol] = 0.10 M.
[geraniol] = 0.40 M.
g
5

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
β-citronellol, the process was positively affected by the concentration of
BQ (Table 2, run 3 vs run 4; run 5 vs run 7 and run 8). Nevertheless, the
re-oxidation of BQH2 unlikely is a rate-limiting step as no GC peak
characteristic for BQH2 was registered during the reactions. Further-
more, a strong acceleration effect of Pd(OAc)2 (Table 2, cf. runs 5 and 6)
and positive order in the concentration of geraniol (Table 2, cf. runs 5,
11 and 12) suggest that the oxidation of geraniol with Pd (II) species is
the slowest part of the overall catalytic process. As discussed above, it
appears that the coordination of BQ to palladium can positively change
its reactivity, in agreement with the data obtained previously in related
processes [9,12,17,23].
the C2-C3 double bond in the geraniol molecule. All products derived
from geraniol and also their mixture show a nice woody-floral scent and
therefore can be applied directly in fragrance compositions.
3.3. Mechanistic considerations
The suggested mechanistic schemes for the reactions of β-citronellol
and geraniol are shown in Schemes 4 and 5. There are two reaction
routes suggested for the allylic oxidation of alkenes: via palladium η3
-
allyls formed by the removal of allylic hydrogen and via σ-alkyl palla-
dium complexes formed by oxypalladation, which is the direct attack of
Thus, geraniol can be oxidized by dioxygen (5–10 atm) with up to
95% selectivity for allylic oxidation products in the presence of Pd
the nucleophile (acetic acid in our case) on the
η
2-coordinated on
palladium olefin [26]. In principle, products 1b and 2b, in which the
acetate is attached to the originally olefinic C-6 atom, could be origi-
nated by any of two alternatives. However, products 1c, 1d, 1e and 2c
cannot be formed by the oxypalladation pathway as they arise from the
oxidation of β-citronellol and geraniol at the originally allylic positions
(at C-5 or C-8). These products are more probable formed via palladium
(OAc)2 and BQ, both in catalytic amounts. It is worthwhile underlining
–
that only one of the C C bonds in geraniol undergoes the oxidation. We
–
suppose that the C2-C3 double bonds in geraniol and geranyl acetate can
be deactivated toward the interaction with palladium by the presence of
electron-withdrawing OH and OAc groups, respectively, in allylic posi-
tions. The data from our previous works on the reactivity of two other
η
3-allyls depicted in Schemes 4 and 5. Moreover, as discussed in [26],
hydroxyolefins,
α
-bisabolol and linalool, with the same catalytic system
–
the mechanism involving allylic intermediates is more expected for
moderately basic nucleophiles, like acetates, because they can assist at
support this suggestion. The terminal monosubstituted C C bond in the
–
linalool molecule with a hydroxyl group in the allylic position was not
the allylic hydrogen removal to form palladium
η
3-allyls. Therefore, we
involved in the reaction with palladium [3]. On the other hand, the
–
–
believe that the oxidations of β-citronellol and geraniol in solutions of
trisubstituted C C bonds in linalool [3] and also in
α-bisabolol [20],
acetic acid promoted by Pd(OAc)2 involve more feasibly palladium
both having a hydroxyl group at a longer distance (in γ-positions),
readily reacted with Pd(II) to give allylic oxidation products. In addition,
we suppose that a steric effect can also contribute to the deactivation of
η
η
3-allyls as intermediates.
Major products 1b/2b and 1c/2c arise from the terminal palladium
3-allyls depicted in Schemes 4 and 5, which are formed by the
Scheme 4. Suggested mechanism for the β-citronellol oxidation (1a).
6

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
Scheme 5. Suggested mechanism for the geraniol oxidation (2a).
Markovnikov-type hydrogen abstraction, i.e., the abstraction from the compositions.
–
carbon allylic to the more substituted end of the C C bond (C-8). The
–
nucleophilic attack of HOAc on the terminal palladium
η
3-allyls could
CRediT authorship contribution statement
proceed at either C-6 or C-8 carbons and would give two types of allylic
products. Linear allyl acetates 1c and 2c result from the attachment of
the acetoxy group to sterically more accessible carbon C-8, whereas
branched allyl acetates 1b and 2b - to more alkyl substituted carbon C-6.
The regioselectivity of the reactions with β-citronellol and geraniol
differs significantly. Whereas geraniol gave both branched and linear
allyl acetates in comparable concentrations, β-citronellol showed a high
preference for more sterically demanded branched isomer 1b. The dif-
ference in regioselectivity could be related to the of the existence of
Aldino N. Venancio: Conceptualization, Investigation, Validation,
Writing - original draft. Luciano Menini: Methodology, Project
administration, Writing - review & editing. Daiane N. Maronde:
Methodology, Investigation. Elena V. Gusevskaya: Conceptualization,
Writing - review & editing. Luciana A. Parreira: Conceptualization,
Supervision, Project administration, Writing - review & editing, Funding
acquisition.
–
second C C bond in the geraniol molecule that can coordinate to
–
Declaration of Competing Interest
palladium acting as a directing group, changing the accessibility of the
C-6 and C-8 carbon atoms in palladium
toxy group.
η
3-allyls for the attacking ace-
The authors report no declarations of interest.
Acknowledgments
The products which could be originated from an internal
η
3-allylic
intermediate (not depicted in Scheme 5) were not found in appreciable
concentrations at the oxidation of geraniol. However, in the case of
β-citronellol this alternative partway was responsible for nearly 15% of
the final products. Products 1d and 1e seem to arise from the internal
We acknowledge a financial support from FAPEMIG, FAPES, CNPq,
´
CAPES and INCT Catalise.
palladium
η
3-allylic intermediate formed form β-citronellol by the
Appendix A. Supplementary data
removal of the hydrogen atom from the allylic C-5 carbon (Scheme 4).
Then, the molecule cyclization occurs via the internal OH group attack
on carbon C-5 to give rose oxide 1d. A subsequent oxidation of rose
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.mcat.2021.111449.
oxide involving the terminal palladium
η
3-allyl (Scheme 4) gives ter-
minal allylic acetate 1e. As it was mentioned above, the analogous
cyclization of the geraniol molecule is not favorable because of the trans
arrangement at the C-2 - C-3 bond and, therefore, a spatial difficulty for
the OH group to participate at the transformation of the C-6 - C-7 bond.
References
[1] J. Bohlmann, C.I. Keeling, Plant J. 54 (2008) 656–669, https://doi-org.ez43.
periodicos.capes.gov.br/10.1111/j.1365-313X.2008.03449.x.
[2] M. Ceborska, Chem. Phys. Lett. 651 (2016) 192–197, https://doi.org/10.1016/j.
cplett.2016.03.051.
4. Conclusions
[3] L.A. Parreira, A.F. Azevedo, L. Menini, E.V. Gusevskaya, J. Mol. Catal. A Chem. 426
(2017) 429–434, https://doi.org/10.1016/j.molcata.2016.07.033.
β-Citronellol and geraniol are biorenewable natural products
commonly used in commercial cosmetic products and insect repellents.
Aiming to upgrade these substrates, we performed their palladium
catalyzed allylic oxidation with dioxygen, for the first time to our
knowledge. The oxidation of trisubstituted and sterically highly hin-
`
¨
[4] E.V. Gusevskaya, J. Jimenez-Pinto, A. Borner, ChemCatChem 6 (2014), https://doi.
org/10.1002/cctc.201490007, 382ꢀ 311.
[5] E.V. Gusevskaya, ChemCatChem 6 (2014) 1506–1515, https://doi.org/10.1002/
cctc.201400052.
[6] J. Smidt, W. Hafner, R. Jira, J. Sedlmeier, R. Sieber, R. Ruttinger, H. Kojer, Angew.
Chem. 71 (1959) 176–182, https://doi.org/10.1002/ange.19590710503.
[7] R. Jira, Angew. Chem. Int. Ed. 48 (2009) 9034–9037, https://doi.org/10.1002/
anie.200903992.
–
dered C C bonds in the β-citronellol and geraniol molecules gave allyl
–
acetates in ca. 90% combined yield. The reactions were efficiently
catalyzed by the palladium acetate/p-benzoquinone combination under
5ꢀ 10 atm of dioxygen pressure without the need for conventionally
used metal co-catalysts.
[8] K.I. Matveev, Kinet. Catal. (Engl. Transl.) 18 (1977) 716–727. https://www.osti.
gov/etdeweb/biblio/5641220.
¨
[9] J.-E. Backvall, A. Gogoll, Tetrahedron Lett. 29 (1988) 2243–2246, https://doi.org/
10.1016/S0040-4039(00)86722-2.
[10] S.S. Stahl, Angew. Chem. Int. Ed. 43 (2004) 3400–3420, https://doi.org/10.1002/
anie.200300630.
The transformation of geraniol proceeded with excellent site-
–
[11] T. Mitsudome, T. Umetani, N. Nosaka, K. Mori, T. Mizugaki, K. Ebitani, K. Kaneda,
Angew. Chem. Int. Ed. 45 (2006) 481–485, https://doi.org/10.1002/
anie.200502886.
selectivity, with only one C C bond participating in the oxidative
–
process. Geraniol gave linear as well as branched allyl acetates in
comparable concentrations, whereas the reaction of β-citronellol
showed high regioselectivity for corresponding branched products. The
work reported herein represents one of a few Wacker-type oxidative
transformations of non-activated acyclic internal alkenes published so
far. Polyfunctionalized oxygenated terpenes produced from β-citronellol
and geraniol are of a potential interest as ingredients for fragrance
¨
[12] J. Piera, J.-E. Backvall, Angew. Chem. Int. Ed. 47 (2008) 3506–3523, https://doi.
org/10.1002/anie.200700604.
[13] M.S. Sigman, E.W. Werner, Acc. Chem. Res. 45 (2012) 874–884, https://doi.org/
10.1021/ar200236v.
[14] D. Wang, A.B. Weinstein, P.B. White, S.S. Stahl, Chem. Rev. 118 (2018)
2636–2679, https://doi.org/10.1021/acs.chemrev.7b00334.
[15] M.G. Speziali, V.V. Costa, P.A. Robles-Dutenhefner, E.V. Gusevskaya,
Organometallics 28 (2009) 3186–3192, https://doi.org/10.1021/om8011213.
7

A.N. Venancio et al.
Molecular Catalysis 504 (2021) 111449
[16] L.A. Parreira, A.M. da Cunha, L. Menini, E.V. Gusevskaya, RSC Adv. 5 (2015),
https://doi.org/10.1039/c5ra10563, 56987–56987.
[28] M. Burgener, T. Tyszewski, D. Ferri, T. Mallat, A. Baiker, Appl. Catal. A Gen. 299
(2006) 66–72, https://doi.org/10.1016/j.apcata.2005.10.021.
[29] N. Kakiuchi, Y. Maeda, T. Nishimura, S. Uemura, J. Org. Chem. 66 (2001)
6620–6625, https://doi.org/10.1021/jo010338r.
[17] A. Vasseur, J. Muzart, J.L. Bras, Eur. J. Org. Chem. (2015) 4053–4069, https://doi.
org/10.1002/ejoc.201500080.
˜
[18] A.C. Bueno, B.B.N.S. Brandao, E.V. Gusevskaya, Appl. Catal. A Gen. 351 (2008)
[30] M.J. Schultz, S.S. Hamilton, D.R. Jensen, M.S. Sigman, J. Org. Chem. 70 (2005)
3343–3352, https://doi.org/10.1021/jo0482211.
226–230, https://doi.org/10.1016/j.apcata.2008.09.023.
´
[19] A.C. Bueno, A.O. de Souza, E.V. Gusevskaya, ChemCatChem 4 (2012) 1382–1388,
[31] C.M.A. Parlett, L.J. Durndell, K. Wilson, D.W. Bruce, N.S. Hondow, A.F. Lee, Catal.
Commun. 44 (2014) 40–45, https://doi.org/10.1016/j.catcom.2013.07.005.
[32] A. Vasseur, R. Membrat, D. Gatineau, A. Tenaglia, D. Nuel, L. Giordano,
ChemCatChem 9 (2017) 728–732, https://doi.org/10.1002/cctc.201601261.
[33] T. Nishimura, T. Onoue, K. Ohe, S. Uemura, J. Org. Chem. 64 (1999) 6750–6755,
https://doi.org/10.1021/jo9906734.
https://doi.org/10.1002/cctc.201200066.
[20] L.A. Parreira, L. Menini, E.V. Gusevskaya, Catal. Sci. Technol. 4 (2014) 2016–2022,
https://doi.org/10.1039/C4CY00260A.
[21] M. dos Santos Costa, A.L.P. de Meireles, E.V. Gusevskaya, Asian J. Org. Chem. 6
(2017) 1628–1634, https://doi.org/10.1002/ajoc.201700337.
[22] A. de Camargo Faria, T.A. dos Santos Wanderley, E.E. Alberto, E.V. Gusevskaya,
Appl. Catal. A Gen. 548 (2017) 33–38, https://doi.org/10.1016/j.
apcata.2017.07.042.
[34] D.M. Carari, M.J. da Silva, Catal. Lett. 142 (2012) 251–258, https://doi.org/
10.1007/s10562-011-0754-4.
[35] M.J. da Silva, J.A. Villarreal, Catal. Lett. 147 (2017) 1646–1653, https://doi.org/
10.1007/s10562-017-2063-z.
[23] M.S. Costa, A.C. Faria, E.V. Gusevskaya, Appl. Catal. A Gen. 584 (2019) 117171,
https://doi.org/10.1016/j.apcata.2019.117171.
[36] E.H. Eschinasi, J. Org. Chem. 35 (1970) 1097–1100, https://doi.org/10.1021/
jo00829a051.
[24] E. Breitmaier, Terpenes. Flavors, Fragrances, Pharmaca, Pheromones, Wiley-VCH,
Weinheim, 2006.
[37] H. Sudo, A. Takushi, E. Hirata, T. Ide, H. Otsuka, Y. Takeda, Phytochemistry 52
(1999) 1495–1499, https://doi.org/10.1016/S0031-9422(99)00218-6.
[38] W. Eschenmoser, P. Uebelhart, C.H. Eugster, Helv. Chim. Acta 66 (1983) 82–91,
https://doi.org/10.1002/hlca.19830660110.
[25] W. Chen, A.M. Viljoen, S. Afr. J. Bot. 76 (2010) 643–651, https://doi.org/10.1016/
j.sajb.2010.05.008.
[26] S. Mann, L. Benhamou, T. Sheppard, Synthesis 47 (2015) 3079–3117, https://doi.
org/10.1055/s-0035-1560465.
[39] C. Singh, N. Guptaa, S.K. Puri, Bioorg. Med. Chem. Lett. 12 (2002) 1913–1916,
https://doi.org/10.1016/S0960-894X(02)00320-7.
[27] K. Kaneda, Y. Fujie, K. Ebitani, Tetrahedron Lett. 38 (1997) 9023–9026, https://
doi.org/10.1016/S0040-4039(97)10411-7.
8