Radical α–alkylation of ketones with unactivated alkenes under catalytic and sustainable industrial conditions

Article abstract of DOI:10.1016/j.apcata.2021.118021

The industrially–viable aerobic α–alkylation of cyclic and acyclic ketones with allyl and alkyl alkenes in the presence of catalytic amounts of Mn²⁺, under homo– and heterogeneous conditions, is achieved here. The substitution of organic peroxides by Mn²⁺ either as a simple soluble salt or supported in zeolites, in air, generates in–situ peracid radicals and circumvents the aggressiveness of current industrial protocols, to pave the way for the design of sustainable aerobic catalytic systems. Combined reactivity and mechanistic studies show that large cyclic ketones stabilize a radical in the α–position due to a higher polarizability, steric hindrance and no proximity effects. As a proof of concept, the gram–scale synthesis of the industrial fragrance exaltolide is carried out with the Mn²⁺ catalysts in air, which clearly improves any other previously reported method not only in safety and environmental terms but also in number of synthetic steps and overall yield.

Full text of DOI:10.1016/j.apcata.2021.118021

Applied Catalysis A, General 613 (2021) 118021
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Radical
α
–alkylation of ketones with unactivated alkenes under catalytic
and sustainable industrial conditions
Sergio Sanz-Navarroa , Francisco Garnes-Portoles , Carlos Lopez-Cruz ,
,
1
a,
1
b
´
´
b
a
a,
´
´
Estela Espinos-Ferri , Avelino Corma , Antonio Leyva-Perez *
a
`
`
Instituto de Tecnología Química (UPV–CSIC), Universidad Politecnica de Valencia, Consejo Superior de Investigaciones Científicas, Avda. de los Naranjos s/n, 46022,
Valencia, Spain
b
´
´
International Flavours & Fragrances Inc., Avda Felipe Klein 2, 12580, Benicarlo, Castellon, Spain
A R T I C L E I N F O
A B S T R A C T
Keywords:
The industrially–viable aerobic α–alkylation of cyclic and acyclic ketones with allyl and alkyl alkenes in the
presence of catalytic amounts of Mn2+, under homo– and heterogeneous conditions, is achieved here. The
substitution of organic peroxides by Mn2+ either as a simple soluble salt or supported in zeolites, in air, generates
in–situ peracid radicals and circumvents the aggressiveness of current industrial protocols, to pave the way for
the design of sustainable aerobic catalytic systems. Combined reactivity and mechanistic studies show that large
Catalysis
Aerobic sustainable industrial reactions
α
–alkylation
Manganese
Zeolites
cyclic ketones stabilize a radical in the
α–position due to a higher polarizability, steric hindrance and no
proximity effects. As a proof of concept, the gram–scale synthesis of the industrial fragrance exaltolide is carried
out with the Mn2+ catalysts in air, which clearly improves any other previously reported method not only in
safety and environmental terms but also in number of synthetic steps and overall yield.
1. Introduction
circumvented these by–reactions so far [4]. Here, we envisioned that the
higher polarizability, lack of proximity effects [5] and more severe steric
The
α–alkylation of ketones by coupling of deprotonated ketones
hindrance compared with shorter ketones [6] of large cyclic ketones
with nucleofuge–containing molecules, under strong basic conditions, is
a fundamental reaction in organic synthesis, in both industry and
academia [1]. However, this low atom–efficient approach requires
additional reagents to avoid competitive O–alkylations, which makes
them industrially unviable in many cases. Modern procedures based on a
combination of metal catalysts with spatially–directing amine additives
have allowed the use of alkenes as coupling partners, which avoid any
waste generation, however, the expensive catalysts/reagents employed
in these protocols make them difficult to implement at industrial scale
[2].
(≥C12) and, in general, long alkyl ketones, may allow to stabilize a
radical in the α position, in such a way that a simple metal catalyst could
rapidly perform the coupling with unactivated alkenes. If so, intra– and
intermolecular reactions will be significantly avoided, and considering
that α–alkylated (large cyclic) ketones are key intermediates in the fine
chemical industry, [7] this particular reactivity deserves attention.
We show here that large cyclic and alkyl ketones couple with unac-
tivated alkenes in the presence of catalytic amounts (1–2 mol%) of
isolated Mn2+ atoms in zeolites and the non–toxic and inexpensive Mn
(OAc)2⋅4H2O salt in the homogeneous phase, [8a–e] under air at 130 ◦C,
and using the biomass–derived compound hexanoic acid as a solvent
[8f].Notice that Mn2+ is used here in catalytic amounts, which differs
significantly of the well–known Mn3+–mediated reactions, which to
become catalytic requires high amounts (up to 50 mol%) of mixed Mn
compounds [8g]. The reaction protocol presented here is fully sustain-
able, industrially implementable, and does not show any toxicity and
flammability limitations, which significantly improve the current
Fig. 1 shows that a full atom–economy alternative to enolates and/or
complex metal systems consists in the use of α–radical ketones, which
will naturally couple with simple alkenes [3]. However, the generation
of the radical ketone generally triggers a plethora of other radical intra–
and inter–molecular reactions, such as dimerization, transannular re-
action, radical walking and different rearrangements, among others, and
only the combination of different metal salts has somewhat
* Corresponding author.
´
E-mail address: anleyva@itq.upv.es (A. Leyva-Perez).
1
These authors contributed equally to the work.
https://doi.org/10.1016/j.apcata.2021.118021
Received 2 November 2020; Received in revised form 31 December 2020; Accepted 25 January 2021
Available online 29 January 2021
0926-860X/© 2021 Elsevier B.V. All rights reserved.

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
room temperature and the solution dried with magnesium sulfate, to
give 900 mg of pure solid compound according to 1H NMR.
2.2.2. Synthesis of the ditert-butylperoxinitrite
Anhydrous ZnCl2 (20 mmol) was added to a solution of sodium trans-
hyponitrite hydrate (Na2N2O2⋅xH2O, 1.9 g, 18 mmol) and tert-butyl-
bromide (17 mL, 150 mmol) in Et2O (10 mL) and stirred for 1.5 h at 25
◦C. The inorganic salt was filtered off and the organic phase was washed
with H2O. The aqueous phase was extracted with Et2O and the combined
organic layers were washed with water, dried over anhydrous Na2SO4
and concentrated under reduced pressure at 25 ◦C (bath temperature) to
afford a solid crude product. Recrystallization from pentane provided
DTBHN as crystalline white long needles.
2.2.3. Impregnation of Mn(OAc)2 in zeolites
Fig. 1. Feasibility of the full–atom economic radical α–alkylation of ketones
To a beaker with 2 g of the corresponding zeolite, a solution of Mn
(OAc)2 (1.82 mmol, 315 mg, 5% wt of Mn) in 2 mL of H2O (0.9 M) was
added. This mixture, was mixed with a spatula until obtaining a ho-
mogeneous paste, this paste was heated at 100 ◦C in an oven for 24 h.
with alkenes. Color code: Positive (green), neutral (orange), negative (red).
industrial process based on sub–stoichiometric amounts of the radical
initiator ditert–butyl peroxide (DTBP). A comparative mechanism study
will show that a four–step, one–electron transfer chain occurs during the
catalytic process, from Mn2+ to molecular oxygen then to the organic
acid and finally to the ketone, thus circumventing the much more en-
ergetic direct organic peroxide abstraction. These findings have led to a
new synthesis of the multi–ton industrial fragrance exaltolide, based on
the aerobic radical procedure.
2.2.4. Exchange of Mn(OAc)2 in zeolites
In a round bottom flask with 2 g of the corresponding zeolite, a 0.7 M
solution of Mn(OAc)2 in H2O (5 mL) was added. This mixture was stirred
◦
for 16 h at 25 C, then vacuum filtered and washed with water. Once
washed, the solid was transferred to a beaker and heated at 60 ◦C for 16
h. The Mn loading of the solid was determined by inductively coupled
plasma-atomic emission spectroscopy (ICP-AES) technique, after disag-
gregation of the solid in aqueous acid mixture.
2. Materials and Methods
2.1. General
2.3. Catalytic studies
Reagents and solvents were obtained from commercial sources
(Aldrich) and were used without further purification otherwise indi-
cated. All the products obtained were characterised by GC–MS, 1H, 13C-
NMR, and DEPT. When available, the characterisation given in the
literature was used for comparison. Gas chromatographic analyses were
performed in an instrument equipped with a 25 m capillary column of
5% phenylmethylsilicone. GC/MS analyses were performed on a spec-
trometer equipped with the same column as the GC and operated under
2.3.1. General reaction procedure for the Mn2+-catalyzed reaction
The corresponding amounts of Mn(OAc)2⋅4H2O (for heterogeneous
catalysis, the appropriate amount of solid catalyst to have 2 mol% of
Mn2+ in the reaction, typically 130 mg, was added) and CDDK 2 (2.16 g,
12 mmol) were placed in a 6 mL vial equipped with a magnetic stirrer.
AcOH (1 mL) or hexanoic acid (1ꢀ 2 mL), and allyl alcohol 1a (88
μL, 1.2
mmol) or allyl acetate 1b (128 L, 1.2 mmol) were then added and the
μ
1
the same conditions. H, 13C, and DEPT were recorded in a 300 MHz
vial was sealed, connected to a manometer. Oxygen was introduced
through a valve to reach 10 bar and liberated to purge the vial. This
operation was repeated twice to finally leave 10 bar oxygen atmosphere
(c.a. 3 mmol, in principle none of the reactants and products ketone,
alcohol or ester and carboxylic acid have a dangerous explosion index
under molecular oxygen). The vial was placed in an oil-bath at 80 or 110
◦C and stirred for the required time. After that, the vial was cooled,
remaining oxygen liberated, and the mixture was analyzed by GC after
dilution in dichloromethane (previous filtration for solid catalysts).
instrument using the appropriate solvent containing TMS as an internal
standard. X-ray powder diffraction measurements were carried out,
before and after catalysis, in a powder diffractometer using Cu Kα ra-
diation (λ =1.54056 Å). Fourier-transformed infrared spectra (FT-IR) of
the solids were recorded on a Jasco instrument by direct pressure on the
measurement window. High-resolution field emission scanning electron
microscopy (HR-FESEM) was performed in a Zeiss Ultra 55 model, by
impregnating dispersions of the zeolites on the FESEM sampleholder and
leaving to evaporate the solvent. X-ray photoelectron spectroscopy
(XPS) measurements were performed by sticking, without sieving, the
zeolite onto a molybdenum plate with scotch type film, followed by air
drying. Measurements were performed using non–monochromatic Mg
KR (1253.6 eV) X–ray source working at 50 W.
2.3.2. Reaction procedure for best results in Table 1
Mn(OAc)2⋅4H2O (2.8 mg, 1 mol%) and CDDK 2 (430 mg, 2.4 mmol
for 2:1, entry 15; 860 mg, 4.8 mmol for 4:1, entry 16) were placed in a 6
mL vial equipped with a magnetic stirrer. AcOH (0.5 mL) and allyl ac-
etate 1b (128 μL, 1.2 mmol) were then added and the vial was sealed,
connected to a manometer. Oxygen was introduced through a valve to
reach 10 bar and liberated to purge the vial. This operation was repeated
twice to finally leave 10 bar oxygen atmosphere (c.a. 3 mmol, in prin-
ciple none of the reactants and products ketone, alcohol or ester and
carboxylic acid have a dangerous explosion index under molecular ox-
ygen). The vial was placed in an oil-bath at 130 ◦C and stirred for 3.5 h.
After that, the vial was cooled, remaining oxygen liberated, and the
mixture was analyzed by GC after dilution in dichloromethane.
2.2. Synthesis of catalysts
2.2.1. Synthesis of the ditert-butylperester
Ditertbutylperoxide DTBP (900 mg, 10 mmol) and pyridine (800 mg,
10 mmol) are dissolved in pentane (10 mL) in a 50 mL round-bottomed
flask equipped with a stirring magnetic bar, and the solution cooled at
-10 ◦C. Then, a solution of oxalyl chloride (530 mg, 4 mmol) in pentane
(5 mL) is added dropwise. The pyridinium chloride was filtered and
washed with pentane. The combined filtrate was cooled in a dry ice-
acetone mixture and allowed to stand about 15 min to precipitate the
perester as fine white crystals. The perester was collected quickly on a
sintered glass filter, dissolved in a minimal amount of pure pentane at
2.3.3. Catalyst leaching study
Two identical reactions were set up, as indicated in the general
procedure with Mn2+-heterogeneous catalysts. After 15 min of reaction,
an aliquot (50 μL) of the two reactions was taken, and one of them was
2

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
Table 1
Catalytic results for the radical coupling of CDDK 2 with allyl derivatives 1a–d under air.
Entry
Catalyst (mol%)
Alkene
1a
R2
O2 (atm)
T (ºC)
3 (%)
1a
(tBuO)2 (DTBP, 5)
(tBuO)2 (15)
H2O2 (15)
11
15
<5
9
2a
140
3a
4a
(BnO)2 (15)
(tBuO)2 (15)
(tBuO)2 (30)
(tBuO)2 (30)
(tBuO)2 (60)
(tBuCOO)2 (15)
(tBuON)2 (15)
V2O5 (5)
5a
19
25
14
21
7
6a
160
7a
1b
8a
9a
45
60
10a
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
9
<5
<5
<5
<5
<5
<5
<5
<5
<5
<5
8
FeCl2 (5)
Air
FeCl3 (5)
Me
CuCl (5)
CuCl2 (5)
1a
CoCl2 (5)
Co(OAc)2 (5)
Pd(OAc)2 (5)
ZnO (5)
110
CeO2 (5)
Mn(OAc)2 (5)
Mn(OAc)2⋅4H2O (5)
9
1b
1a
13
15
37
41
52
55
58
16
20
27
32
<5
<5
<5
<5
<5
1b
4
Mn(OAc)2⋅4H2O (2)
1c
1d
4 þ 20 N2
MnO
n–Hex
MnCl2
4
1c
Mn2O3
130
MnO2
FeCl3 (5)
Fe(acac)3 (5)
CuCl (5)
Cu(OAc)2 (5)
Co(OAc)2 (5)
1b
Me
4
a
Without AcOH solvent, increasing O2 pressure has no effect on the reaction.
filtered in hot through a 20
μ
m nylon filter to remove all the solids. The
2.4. Studies on the DTBP-mediated process
resulting solution was reacted under exactly the same conditions and
aliquots of both reactions were taken at 30, 60, 120 and 180 min. All
2.4.1. g-scale experimental procedure with DTBP radical initiator and
isolation of the products
aliquots were diluted with EtOAc (1 mL), filtered with a 20
μm nylon
filter and analyzed by GC after adding n-dodecane (22
μ
L, 0.1 mmol) as
CDDK 2 (2.7 g, 15 mmol), allyl alcohol 1a or allyl acetate 1b (1.1 or
1.6 mL respectively, 1 equivalent) and peroxide DTBP (0.3 and 1.2 mL,
15 and 60 mol% respectively) are placed in a two-necked 25 mL round-
bottomed flask equipped with a magnetic stir bar and a condenser, and
heated in an oil bath at 140 ◦C or 160 ◦C, respectively, for 2.5 h. After
that time the mixture is cooled, an aliquot is taken for GC analysis, and
the crude is dissolved in CH2Cl2 (2 mL) and directly submitted to flash
column chromatography on silica gel, eluting with a mixture of hexane
and ethyl acetate. Rf (ketoalcohol, 30 % AcOEt) = 0.23. Rf (ketoester,
5% AcOEt) = 0.26. Yield of ketoalcohol: 240 mg (7%). Yield of
ketoester: 1.0 g (23 %).
an external standard.
2.3.4. Recovery and reuse of Mn2+-NaBeta
The
α-alkylation reaction of 2 with 1b was performed using the
general method and, after 4 h reaction time, the reaction was stopped
and the zeolite separated from the solution using a centrifuge at 6000
rpm. After separation, the zeolite was cleaned with EtOAc (5 mL) and
separated again in the centrifuge. The process was repeated twice. Then,
the zeolite was left to dry overnight, weighed and used again in reaction,
adjusting the mass of substrates and volume of hexanoic acid to keep the
same final concentration.
2.4.2. Experimental procedure for 4:1 ketone:alcohol ratio with DTBP
peroxide catalyst
CDDK 2 (10.94 g, 60 mmol) was placed in a two-necked 25 mL
round-bottomed flask equipped with a magnetic stir bar, a pressure-
3

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
compensated addition funnel and a condenser, and heated in an oil bath
at 140 ◦C. A solution of tert-butyl peroxide DTBP (0.44 mL, 15 mol%) in
allyl alcohol 1a (1.1 mL, 15 mmol) was poured into the addition funnel
and added dropwise during 6 h. After that time, the mixture was left to
stir at 140 ◦C for additional 3 h and then the excess CDDK was distilled
off (107 ◦C, 13 mmHg, 95 % recovering). Then, phosphoric acid (85 wt
%, 0.01 mL, 0.1 mol%) was added and the solution was heated at 120 ◦C
while distilling off the so-formed water (c.a. 30 min). After that, the
product 3a was distilled off and weighted (111 ◦C, 1 mmHg, 2.36 g, 17 %
yield).
Kinetic experiments confirm that the biomass–derived derivatives 1c
and hexanoic acid (solvent) are equally effective that the methyl–sub-
stituted partners under reaction conditions (Figures S2 and S3). These
biomass–derived reagents are not only more sustainable but also
circumvent the use of highly flammable and volatile C2–4 reagents,
whose flammability limits are below 3% in molecular oxygen, according
to literature [10a] and to our own measurements (Figure S4). Longer
alkyl organic acids also show a satisfactory reactivity, producing a clear
benefit in product yield compared to acetic acid and without any sig-
nificant erosion on selectivity (Table S3). These results are explained by
a better solubility of 2 in the longer alkyl chain solvents, since 2 does not
completely dissolve in acetic acid even under heating conditions.
Indeed, a high hydrophobic substrate such as 1–dodecene 1d engages
well in the coupling with the biomass–derived solvent, giving the
highest yield (58 %, entry 29).
2.4.3. High-scale experimental procedure for 1:1 ketone: alcohol ratio with
DTBP radical initiator
CDDK 2 (18.2 g, 100 mmol) was placed in a two-necked 50 mL
round-bottomed flask equipped with a magnetic stir bar, a pressure-
compensated addition funnel and a condenser fitted with a nitrogen
balloon, and then heated in an oil bath at 140 ◦C. A solution of tert-butyl
peroxide DTBP (10.8 mL, 30 mol%) in allyl alcohol 1a (7.3 mL, 100
mmol) was poured into the addition funnel during 30 min. After that
time the mixture was left to stir at 140 ◦C for additional 6 h and then the
corresponding acid solid catalyst (2.0 g) or phosphoric acid (85 wt%,
0.05 mL, 0.1 mol%) was added to the solution, heated at 120 ◦C. After
that, the mixture was analyzed by 1H nuclear magnetic resonance
(NMR) spectroscopy and also by gas chromatography coupled to mass
spectrometry (GC–MS) after diluting in CDCl3, using prepared standards
as a calibrate. The calculated 3a (based on ketone and not in allyl
alcohol) yield is 7%.
3.2. Radical mechanism of CDDK 2 α–alkylation
The industrial process for the alkylation of CDDK 2 with 1a–b based
on DTBP (radical initiator) has, to our knowledge, only been described
in patents but not in academic reports, [4a] and the mechanism is not yet
understood. This lack of knowledge about a challenging but applied
organic reaction is surprising. Thus, the mechanisms for both the DTBP
radical initiator and the Mn2+ catalyst will be studied here, taking as a
reaction model the alkylation of CDDK 2 with 1a or 1b, in order to
stablish parallelisms and differences.
3.2.1. Radical mechanism with the DTBP initiator
2.4.4. Reproduction of industrial reaction conditions
The coupling between CDDK 2 and 1a initiated by DTBP proceeds
with a first–order kinetics in the initial 1–2 h of reaction, and then
rapidly leverages, regardless the amounts of reagent used (Figure S5).
The rate equation obtained with different amounts of each reactant
(ketone 2, alkene 1a and DTBP) and different temperatures, is r0=kapp
[2][DTBP] (kapp= an apparent rate constant), which suggests that the
CDDK 2 (50.0 g, 274.7 mmol) was placed in a two-necked 100 mL
round-bottomed flask equipped with a magnetic stir bar, a pressure-
compensated addition funnel and a condenser, and heated in an oil
bath at 140 ◦C. A solution of tert-butyl peroxide DTBP (4.4 mL, 30 mol%)
in allyl alcohol 1a (5.5 mL, 75 mmol) was poured into the addition
funnel and added dropwise during 6 h. After that time, the mixture was
left to stir at 140 ◦C for additional 3 h and then the excess CDDK 2 was
distilled off through a 10 cm long Vigreaux column (107 ◦C, 13 mmHg,
40.3 g, >99 % recovering). Then, the corresponding acid solid catalyst
(5.0 g) or phosphoric acid (85 wt%, 0.05 mL, 0.1 mol%) was added and
the solution was heated at 120 ◦C while distilling off the so-formed water
(c.a. 30 min). After that, product 3a was distilled off in pure form (111
◦C, 1 mmHg, 10.8 g, 17 % yield).
abstraction of the
α hydrogen atom in 2 is the rate–determining step
(rds) of the reaction. The fact that thermally sensitive but much more
reactive peroxy radicals such as (tBuCOO)2 and (tBuON)2 gave signifi-
cant amounts of 3a (entries 9 and 10 in Table 1), supports this
hypothesis.
A combined analysis of the liquid and gas phase during reaction by
GC and 1H NMR shows that the conversion of 1a is much faster than the
formation of 3a, that DTBP steadily decomposes to methane, ethane and
acetone after 1 h reaction time, and that product 3a also decomposes
under reaction conditions, as assessed with neat 3a (isolated by col-
umn–chromatography, Figures S6 and S7). Control experiments show
that the rate of degradation of 3a is just slightly lower than its formation
(~15 % hꢀ 1) in the presence of the peroxide, which explains the sudden
stopping of the reaction. Indeed, if more DTBP is added when the re-
action stops, the yield of 3a does not increase beyond an additional 1%.
Product 3a does not degrade (<5% after 2 h) without DTBP. These re-
sults strongly suggest that 1a and 3a decompose during reaction by the
aggressiveness of the radicals, thus the formation of 3a by the
DTBP–mediated reaction is intrinsically limited. [10b] This degradation
is directly related to the presence of more acid hydrogen atoms in the
product than in the reactants (see Figure S8 for pKas), which explains
the better results obtained with allyl derivatives 1b–d respect to allyl
alcohol 1a. The use of reagents that may alleviate radical degradation
such as bases, organic molecules (N–hydroxyphthalimide, NHPI),
radical trapping metals (copper), acetone and hydrophobic solvents do
not produce significant improvements (Table S8 and Figures S9–S11).
Despite the use of allyl acetate 1b allows to increase the amount of
peroxide without suffering severe decomposition, even after addition in
twice (Table S5), and GC and NMR measurements confirm the better
stability and isolation (Figure S12) of product 3b under reaction con-
ditions (Figure S13), degradation still occurs and none of the palliative
methods tested with 1a worked with 1b (Figure S14).
3. Results and Discussion
3.1. Catalytic radical
α–alkylation of cyclododecanone (CDDK) with
biomass–derived reagents
Table 1 shows the coupling reaction between allyl alcohol de-
rivatives 1a–d and cyclododecanone 2 (CDDK) under open air condi-
tions, at different reaction temperatures. From >20 radical organic and
metal catalysts tested, [9] only DTBP (entries 5–8) and Mn2+ salts, either
in anhydrous or aqueous form (i.e. entries 21–23), gave >10 % yields of
the alkylated products 3a–b. In particular, the Mn(OAc)2 catalyst ach-
ieves ca. 40 % yield under 4 atmospheres of O2 (entries 24–26), and ca.
55 % yield when the biomass–derived reagent allyl hexanoate (1c, R1=
n–hex) and hexanoic acid solvent were used, either under pure oxygen
atmospheres (entry 27) or simulated air (entry 28, see optimization tests
in Tables S1–S2 and Figure S1). Other Mn catalysts were also reasonable
effective (entries 30–33), but less active than the corresponding acetate
salt. The beneficial role of acetates to generate the active radicals under
aerobic atmosphere will be commented ahead. Although product yields
do not exceed 60 %, these results are extraordinary if one considers the
number of synthetic steps and waste generation saved, which is
dramatically reflected in the final price of the product (see ahead for the
fragrance exaltolide).
4

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
With the above results in hand, Fig. 2 shows the proposed mechanism
for the radical addition of alkenes to CDDK 2 initiated by DTBP, which
consists in peroxide homolytic fragmentation, H abstraction from 2 and
final coupling with the alkene. The O–centered radical of DTBP de-
composes to a C–centered radical and acetone, which consumes the
DTBP and significantly degrades reagents and product (Figure S15).
the anionic network of the zeolite may further stabilize Mn2+ cationic
species, avoiding leaching [13]. Besides, zeolites are active matrixes for
radical reactions of ketones and aldol–type transformations [13a–f].
The representative zeolites Y and Beta, in Na, NH4 and H ionic forms,
were tested as supports for Mn2+, since these 3D zeolites present a pore
opening of ~7 Å, wide enough to diffuse the reagents and products of
the reaction (see Figure S18 for zeolite details) [13]. Recalling that other
Mn2+ species are also catalytically active (see entries 30–33 in Table 1
above), the catalytic Mn2+ species were supported here not only by
impregnation but also by aqueous cationic exchange of Mn(OAc)4⋅4H2O
with the zeolite cations, since the latter enables a stronger ionic inter-
action between the zeolite and Mn2+, at expenses of removing acetate
ligands. The loading of Mn2+ supported by ionic exchange depends on
the zeolite employed and ranges in 1.7–2.2 wt% (Figure S18) according
to inductively coupled plasma atomic emission spectroscopy measure-
ments (ICP–AES), after disaggregating the solid in concentrated HF so-
lutions. With this result in hand, the impregnation of the zeolites was
carried out with similar amounts of Mn(OAc)4⋅4H2O (~1.5 wt%), in
order to have comparative results. The comparative characterization of
Mn2+–NaBeta with the pristine zeolite by powder X–ray diffraction
(XRD, Figure S19), Fourier–transformed infrared spectra (IR,
Figure S20) and high–resolution field emission scanning electron mi-
croscopy (HR–FESEM, Figure S21) shows that the microstructure re-
mains unaltered after Mn2+ incorporation. Besides, Mn is hardly
detected in the energy dispersive X–ray spectra of the sample during
FESEM measurements, thus indicating a good incorporation and
dispersion of Mn within the zeolite framework. X–ray photoelectron
microscopy (XPS, Figure S22) shows a binding energy for Mn2+ around
645 eV, attributable to [MnO4], [11c] in other words, the Mn2+ sites are
surrounding by multiple hydroxyl groups of the zeolite walls, as it may
be expected. These results, together, confirm the incorporation of Mn2+
in cationic positions within the zeolite.
3.2.2. Radical mechanism with the Mn2+/O2 catalyst
The rate equation obtained with different amounts of each reactant
(2, 1b, Mn(OAc)2⋅4H2O, O2 and acid solvent) for the Mn2+–catalyzed
process is r0=kapp [2][Mn][O2][acid solvent] (kapp= an apparent rate
constant), thus resembling that of DTBP but changing the organic
peroxide by the combination of Mn2+ catalyst, air and the organic acid
solvent (Figure S16). The use of stoichiometric amounts of Mn
(OAc)2⋅4H2O under N2 did not give any conversion of 2, which indicates
that Mn2+ does not transfer electrons to 2 unless O2 is present, and when
–
so, CO2 extrusion and C C breaking of 2 are also observed (Figure S17).
These results support the formation of peroxy radicals during the
Mn2+–catalyzed reaction. [11a,b] Indeed, if the catalytic Mn2+/O2
system is replaced by peracetic acid, the reaction proceeds with a similar
first order kinetics. No other solvents rather than alkyl carboxylic acids
are active for the reaction (see Figure S16). These results, together,
indicate that peracid species are catalytically formed from Mn2+, O2 and
the solvent during reaction, and that this peracid is the true active
radical initiator of the coupling, which nicely explains the equation rate
obtained. Although the transformation of Mn2+ acetate to Mn oxide
during reaction is a possibility, [11c] the fact that MnO is the less active
of all Mn catalysts (Table 1, entry 30) and that the reaction is performed
in a high excess of AcOH or other carboxylic acid (see also ahead),
strongly suggests that the catalytic Mn site is in acetate form. The
mechanism proposed for the Mn2+–catalyzed coupling of 2 with 1b is
shown in Fig. 3, above.
The catalytic results in Table 2 show that all the zeolites worked well
as a catalyst for the coupling of CDDK 2 with allyl acetate 1b, with
general yields of 3b ca. 20–40 % for both the impregnated and the
exchanged zeolites, except for zeolite HBeta. In general, the non–acid
zeolites worked better than the acid ones, and remarkably, the
exchanged zeolites worked better than the impregnated zeolites,
particularly Mn2+–NaBeta, which gave a 48 % yield of 3b, a better yield
than any homogeneous catalyst tested. These results strongly suggest
that the zeolite framework can isolate Mn cations by ion exchange,
stabilizing Mn2+ to efficiently catalyze the radical reaction [14]. Fig. 3
shows the hot filtration test for the reaction with Mn2+–NaBeta as a
catalyst, and the catalytic activity coming from species in solution is 14
% of the total (4% out of 28 % after filtration). ICP–AES analyses showed
a loss of 0.19 Mn wt% in the solid after use and a 0.14 Mn wt% in the
reaction filtrates, which match the percentage of catalytic activity in
solution. In addition, Fig. 3 also shows that Mn2+–NaBeta can be recy-
cled three times but with a continuous decrease of catalytic activity
throughout the results, due to the systematic loss of Mn, as confirmed by
ICP–AES analyses. Nevertheless, although the solid catalyst presents
some leaching and the consequent loss of activity throughout the reuses,
these results open the way to design solid catalysts for radical–catalyzed
aerobic transformations in acidic organic media, and bring a rare
3.3. Mn2+–zeolites as recyclable solid catalysts
The transformation of homogeneous into heterogeneous catalytic
process is of interest not only from an environmental but also from an
industrial point of view, since the solid catalysts can be recovered,
recycled and ultimately engineered in continuous flow. [12a–c] Radical
catalysts dissolved in acetic acid are very difficult to translate into
recyclable solids, since the extreme polarity of the reaction medium
favors a rapid leaching of the catalytic species. [12d] However, the use
of longer alkyl organic acids instead of acetic acid as a solvent enables
here the possibility of having metal supported catalysts without leach-
ing. Table 2 shows the results for different Mn2+–supported solid cata-
lysts under the optimized conditions for the homogenous system. Mn
(OAc)4⋅4H2O is still active when supported by impregnation on silica,
however, the catalytic activity corresponds exclusively to Mn2+ species
leached into the solution, as it may be expected from a low metal-
–support interaction. This result indicates that the Mn2+ catalytic spe-
cies must be more tightly anchored on the solid surface in order to have a
recyclable catalyst, otherwise leaching will occur. Thus, zeolites were
envisioned as potential support candidates, since the silico–alumina
structure should not interfere in the radical process, as silica does, but
Fig. 2. Proposed reaction mechanisms for the
α
–alkylation of 2 with the radical initiator DTBP and Mn2+/O2/alkyl acid catalysts.
5

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
Fig. 3. Hot filtration test (left) and reuses (right) of Mn2+–NaBeta catalyst during the
α–alkylation of CDDK 2 with allyl acetate 1b under optimized reaction
conditions (see Table 2). Error bars account for 5% uncertainty.
Table 2
Results for the catalytic radical coupling of allyl acetate 1b with CDDK 2 under aerobic conditions.
Entry
Catalyst
Synthetic method
3b (%)
1
Mn(OAc)2–SiO2
Mn(OAc)2–NaY
Mn2+–NaY
Impregnation
Impregnation
Ionic exchange
Impregnation
Ionic exchange
Impregnation
Ionic exchange
Impregnation
Ionic exchange
Impregnation
Ionic exchange
30a
29
21
21
38
30
47
27
43
<5
17
2
3
4
Mn(OAc)2–HY
5
Mn2+–HY
6
Mn(OAc)2–NaBeta
Mn2þ–NaBeta
Mn(OAc)2–NH4Beta
Mn2+–NH4Beta
Mn(OAc)2–HBeta
Mn2+–HBeta
7
8
9
10
11
a
Complete leaching of Mn observed.
Fig. 4. Yields of the radical α–alkylation of cyclic and
acyclic ketones with allyl hexanoate 1c and 1–dodecene
1d under the indicated conditions. Black points are the
reported electron affinities for cyclic ketones in
Kcal⋅mol–1 (aprox. + 2.0 for acyclic ketones), with the
last two data extrapolated (dash line).16 Green points
are the experimental yields for cycle ketones (circles for
reaction with 1c and squares for reaction with 1d) and
blue points are the experimental yields for linear ke-
tones (same coding). Lines are a guide to the eye.
6

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
Fig. 5. Synthesis of exaltolide 5 from CDDK 2 by different routes. a Overall isolated yield at >0.5 g–scale.
example of solid–catalyzed
α
–alkylation of zeolites with simple alkenes
manufactured at multi–ton amounts from 3a with DTBP as a sub-
–stoichiometric reagent. Fig. 5 shows the overall isolated yields obtained
for exaltolide 5, in our hands, for a) the current industrial DTBP proc-
ess,4a b) the aerobically Mn2+–catalyzed procedure, and c) other alter-
native processes reported in the literature (see details for each procedure
in Figures S23–27) [7a,18,19]. All the routes have in common the syn-
[15]. Characterization of the used zeolite shows that the microstructure
is preserved (see Figures S19-S20).
3.4. The uniqueness of large cyclic and alkyl ketones for the radical
α
–alkylation
thesis of intermediate α–alkylated cyclododecanones 3 and cyclopyrane
4, which is in turn converted to exaltolide 5 [20].
Molecular polarization is a well–known effect in ketones, clearly
The highest yield of exaltolide 5 (24 %) is obtained by the Mn2+/air
route described here which, besides, is the shortest route (4 steps,
Figure S23). The current industrial route [4a] gives a lower yield (17 %,
Figure S24) than the Mn2+/air route and uses the low sustainable and
more expensive organic peroxide in sub–stoichiometric amounts, which
translates in a calculated economic increase of ≈1 million euros for an
annual production of 500 tons [21]. The ketal route [19] makes use of
the intermediate ketal 6 to achieve exaltolide 5 after several steps with
expensive reagents (Figures S25–26) and to give the lowest yield of the
synthetic pathways tested (6%). The enamine route [18] has not been
reported before, as far as we know, for the synthesis of 5, but just for
intermediates 7 and 8. We have here completed the synthesis of 5 to
give, in our hands, a respectable yield (19 %, Figure S27), however, the
complete route requires 7 steps with expensive reagents. These results
showcase the effectiveness of the Mn2+–catalyzed aerobic process,
beyond the intrinsic environmental and practical advantages associated
to biomass–derived reagents and potential solid catalysts.
reflected not only into the physicochemical properties (boiling point,
refractive, index,… ) but also in the spectroscopic values (carbonyl NMR
and IR shifts,… ) which, however, has barely been correlated with
reactivity [5]. Fig. 4 shows the yield obtained for the radical coupling of
different cyclic and acyclic ketones with both allyl hexanoate 1c and
1–dodecene 1d, under the optimized conditions in Table 1 (2 mol% of
Mn(OAc)4⋅4H2O, 130 ◦C) vs. the number of carbon atoms (n) in the
cycle/chain. It can be seen that, regardless the type of ketone and alkene
employed, yields of alkylated ketone parallel their electronic affinity,
which is an accepted parameter to measure the molecular polarization
[16].
The higher radical αꢀ alkylation observed for ketones of increasing
length means that the reaction depends linearly on the ability of the
alkyl moiety to stabilize the radical by inductive effects, except for
medium–size cycles up to cyclodecanone (≤C10), where full conversion
is observed but without any alkylation product [17]. GC and NMR an-
alyses of the reaction mixture of cyclodecanone (C10) show the forma-
tion of no less than 10 new products, from which >70 % corresponds to
decaline products, together with some dimerization and minor uniden-
tified products. As depicted in Fig. 4, the decaline mixture is produced
by rapid radical rearrangement and transannular reactions, [6] which
4. Conclusions
The radical α–alkylation of ketones with unactivated alkenes pro-
ceeds with earth–abundant, cheap and non–toxic Mn2+ catalysts by a
biomass–based industrial aerobic process, to enable the multi–ton in-
dustrial fragrance exaltolide 5 in significantly higher yields than any
other reported method, including the current industrial route. A com-
bination of high polarizability, high steric hindrance for dimerization
and lack of proximity effects, makes large cyclic ketones particularly
explains the lack of
α–alkylation for medium–size cycles. Thus, it can be
said that the radical
α
–alkylation of ketones can be predicted on the
basis of the ketone structure, which is a sound starting point for the
design of new reactions.
3.5. Synthesis of exaltolide
suitable to engage in this full–atom economic radical α–alkylation re-
action. A simple solid catalyst based on Mn2+–supported zeolites bring
Exaltolide 5 is a widely–used marketed fragrance molecule, currently
7

S. Sanz-Navarro et al.
Applied Catalysis A, General 613 (2021) 118021
[7] (a) N. Hanaki, K. Ishihara, M. Kaino, Y. Naruse, H. Yamamoto, Tetrahedron 52
(1996) 7297–7320;
promising results for the design of a process based on heterogeneous
catalysts. The consideration of perhaps old but sound organic concepts
to explain cyclic ketone reactivity poses a new scenario for the appli-
cation of ketones in industrial processes.
S. Lambrecht, W. Marks, H.-J. Topp, N. Richter, W. Kuhn, US 2005/0085536 A1
(2005). (d) X. Shen, T.T. Nguyen, M.J. Koh, D. Xu, A.W.H. Speed, R.R. Schrock, A.
H. Hoveyda, Nature 541 (2017) 380–386.
[8] (a) W. Hui, C. Isaac, R. Torben, K. Nikolaos, A. Lutz, Nat. Catal. 1 (2018)
993–1001. For Mn–catalyzed alkene reactions see:;
Declaration of Competing Interest
(b) V. Papa, Y. Cao, A. Spannenberg, K. Junge, M. Beller, Nat. Catal. 3 (2020)
135–142;
(c) T. Liu, Y. Yang, C. Wang, Angew. Chem., Int. Ed 59 (2020) 14256–14260;
(d) M. Pena–Lopez, P. Piehl, S. Elangovan, H. Neumann, M. Beller, Angew. Chem.,
Int. Ed 55 (2016) 14967–14971. For Mn–catalyzed ketone alkylation reactions see:;
(e) M. Milan, M. Bietti, M. Costas, ACS Cent. Sci. 3 (2017) 196–204. For
Mn–catalyzed peroxide–mediated reactions see:;
The authors have no competing interests to declare.
CRediT authorship contribution statement
(f) A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502. For a
representative review on biomass see:;
Sergio Sanz-Navarro: Investigation, Methodology. Francisco
´
´
Garnes-Portoles: Investigation, Methodology. Carlos Lopez-Cruz: Su-
(g) U. Linker, B. Kersten, T. Linker, Tetrahedron 51 (1995) 9917–9926. For
catalytic Mn3+ see:.
´
pervision, Validation. Estela Espinos-Ferri: Supervision, Project
[9] A. Kockritz, R. Eckelt, O. Koch, P. Esser, J. Panten, W. Baumann, H. Atia, Eur. J.
Inorg. Chem. 2018 (2018) 2776–2784.
administration, Funding acquisition. Avelino Corma: Supervision,
´
Validation, Funding acquisition. Antonio Leyva-Perez: Supervision,
[10] (a) P.M. Osterberg, J.K. Niemeier, C.J. Welch, J.M. Hawkins, J.R. Martinelli, T.
E. Johnson, T.W. Root, S.S. Stahl, Org. Process Res. Dev. 19 (2015) 1537–1543;
(b) M. Martin–Martinez, R.S. Ribeiro, B.F. Machado, P. Serp, S. Morales–Torres, A.
M.T. Silva, J.L. Figueiredo, J.L. Faria, H.T. Gomes, ChemCatChem 8 (2016)
2068–2078.
Project administration, Writing - original draft, Writing - review &
editing.
[11] (a) X. Zhao, T. Zhang, Y. Zhou, D. Liu, J. Mol. Catal. A Chem. 271 (2007) 246–252;
(b) G. Luo, X. Lu, X. Wang, S. Yan, X. Gao, J. Xu, H. Ma, Y. Jiao, F. Li, J. Chen RSC
Adv. 5 (2015) 94164–94170;
Acknowledgments
ITQ thanks IFF for its continuous support and collaboration. This
work was also supported by the MINECO (Spain) (Projects
(c) E. Karakhanov, A. Maximov, A. Zolotukhina, V. Vinokurov, E. Ivanov,
A. Glotov, Catalysts 10 (2020) 7.
[12] (a) L. Liu, A. Corma, Chem. Rev. 118 (2018) 4981–5079. For recent reviews see;
(b) J.R. Cabrero–Antonino, R. Adam, V. Papa, M. Beller, Nat. Commun. 361
(2019) 185–191;
CTQ2017–86735–P
and
Excellence
Unit
“Severo
Ochoa”
SEV–2016–0683). S. S.–N. thanks MINECO the concession of a FPI
contract associated to the above CTQ project, and F. G.–P. thanks ITQ for
the concession of a contract.
c) M. Viciano–Chumillas, M. Mon, J. Ferrando–Soria, A. Corma, A. Leyva–Perez,
́
D. Armentano, E. Pardo, Acc. Chem. Res. 53 (2020) 520–531;
(d) P. Kalck, C. Le Berre, P. Serp, Chem. Rev. 402 (2020), 213078. For reactions in
acetic acid see.
Appendix A. Supplementary data
[13] (a) H.Y. Luo, L. Bui, W.R. Gunther, E. Min, Y. Roman–Leshkov, ACS Catal. 2
(2012) 2695–2699;
(b) P. Wu, Y. Kubota, T. Yokoi, ACS Catal. 4 (2014) 23–30;
(c) S. Van de Vyver, Y. Roman–Leshkov, Angew. Chem., Int. Ed 54 (2015)
12554–12561;
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2021.118021.
(d) T. Ennaert, J. Van Aelst, J. Dijkmans, R. De Clercq, W. Schutyser, M. Dusselier,
D. Verboekend, B.F. Sels, Chem. Soc. Rev. 45 (2016) 584–611;
(e) Y. Wang, J.D. Lewis, Y. Roman–Leshkov, ACS Catal. 6 (2016) 2739–2744;
(f) J.D. Lewis, M. Ha, H. Luo, A. Faucher, V.K. Michaelis, Y. Roman–Leshkov, ACS
Catal. 8 (2018) 3076–3086;
References
[1] (a) F. Huang, Z. Liu, Z. Yu, Angew. Chem., Int. Ed 55 (2016) 862–875;
(b) B. Maji, M.K. Barman, Synthesis 49 (2017) 3377–3393;
(c) P. Piehl, R. Amuso, E. Alberico, H. Junge, B. Gabriele, H. Neumann, M. Beller,
Chem. Eur. J. 26 (2020) 6050–6055.
(g) A. Rodriguez–Fernandez, J. Di Iorio, C. Paris, M. Boronat, A. Corma,
Y. Roman–Leshkov, M. Moliner, Chem. Sci. 11 (2020) 10225–10235;
´
´
(h) M.A. Rivero–Crespo, J. Oliver–Meseguer, K. Kapłonska, P. Kustrowski,
[2] (a) M. Nakamura, T. Hatakeyama, E. Nakamura, J. Am. Chem. Soc. 126 (2004)
11820–11825;
´
E. Pardo, J.P. Ceron–Carrasco, A. Leyva–Perez, Chem. Sci. 11 (2020) 8113–8124;
´
´
(i) M.A. Rivero–Crespo, M. Tejeda–Serrano, H. Perez–Sanchez, J.
(b) F. Mo, G. Dong, Science 345 (2014) 68–72;
´
´
P. Ceron–Carrasco, A. Leyva–Perez, Angew. Chem. Int. Ed 59 (2020) 3846–3849;
(c) Y. Dang, S. Qu, Y. Tao, X. Deng, Z.-X. Wang, J. Am. Chem. Soc. 137 (2015)
6279–6291;
(j) M. Mon, R. Bruno, S. Sanz–Navarro, C. Negro, J. Ferrando–Soria, L. Bartella,
`
´
L. Di Donna, M. Prejano, T. Marino, A. Leyva–Perez, D. Armentano, E. Pardo, Nat.
Commun. (2020) 3080, https://doi.org/10.1038/s41467–020–16699–3. For other
related micro–structured materials, such as MOFs, see:.
(d) Z. Huang, H.N. Lim, F. Mo, M.C. Young, G. Dong, Chem. Soc. Rev. 44 (2015)
7764–7786;
(e) D. Xing, G. Dong, J. Am. Chem. Soc. 139 (2017) 13664–13667;
(f) X. Li, H. Wu, Y. Lang, G. Huang, Catal. Sci. Technol. 8 (2018) 2417–2426;
(g) D. Xing, X. Qi, D. Marchant, P. Liu, G. Dong, Angew. Chem., Int. Ed 58 (2019)
4366–4370;
´
´
[14] (a) P. Rubio–Marques, M.A. Rivero–Crespo, A. Leyva–Perez, A. Corma, J. Am.
Chem. Soc. 137 (2015) 11832–11837;
(b) I.C. Gerber, P. Serp, Chem. Rev. 120 (2020) 1250–1349;
´
(c) M. Tejeda-Serrano, S. Sanz–Navarro, F. Blake, A. Leyva-Perez Synthesis 52
(h) H.N. Lim, D. Xing, G. Dong, Synlett 30 (2019) 674–684.
[3] (a) A. Mastracchio, A.A. Warkentin, A.M. Walji, D.W.C. MacMillan, Proc. Natl.
Acad. Sci. 107 (2010) 20648–20651;
(2020) 2031–2037.
[15] G. Meng, M. Patel, F. Luo, Q. Li, C. Flach, R. Mendelsohn, E. Garfunkel, H. He,
M. Szostak, Chem. Comm. 55 (2019) 5379–5382.
[16] (b) A.H. Zimmerman, R.L. Jackson, B.K. Janousek, J. Brauman, J. Am. Chem. Soc.
100 (1978) 4674–4676.
(b) F. Lima, U.K. Sharma, L. Grunenberg, D. Saha, S. Johannsen, J. Sedelmeier, E.
V. Van der Eycken, S.V. Ley Angew, Chem., Int. Ed 56 (2017) 15136–15140.
[4] (b) T. Iwahama, S. Sakaguchi, Y. Ishii, Chem. Commun. (2000) 2317–2318;
R. Hopp, K. Bauer, USA patent 3856815, (1974).
´
[17] J.S. Sharley, A.M. Collado–Perez, E. Espinos–Ferri, A. Fernandez–Miranda, I.
R. Baxendale, Tetrahedron 72 (2016) 2947–2954.
[5] (a) J. Applequist, J.R. Carl, K.-K. Fung, J. Am. Chem. Soc. 94 (1972) 2952–2960;
(b) I.A. Blair, J.H. Bowie, V.C.J. Trenerry, Chem. Soc., Chem. Commun. (1979)
230–231;
[18] J.R. Mahajan, H. de Araujo, Synthesis 1980 (1980) 64–66.
[19] P.G. Gassman, S.J. Burns, K.B. Pfisterl, J. Org. Chem. 58 (1993) 1449–1457.
´
[20] Y.N. Ogibin, A.O. Terentev, G.L. Nikishin, Russ. Chem. Bull. 47 (1998) 1166–1169.
(c) K.A. Bode, J. Applequist, J. Phys. Chem. 100 (1996) 17820–17824;
´
´
´
[21] Corma A., Leyva–Perez, A., Espinos–Ferri, E., Lopez–Cruz, C., European patent
¨
(d) D.A. Walthall, J.M. Karty, B. Romer, O. Ursini, J.I. Brauman, J. Phys. Chem. A
number 16382632.4–1451 (2016) and USA patent 15833455 (2017).
109 (2005) 8785–8793.
´
[6] (a) M. Szostak, L. Yao, J. Aube, J. Am. Chem. Soc. 132 (2010) 2078–2084;
(b) E. Reyes, U. Uria, L. Carrillo, J.L. Vicario, Tetrahedron 70 (2014) 9461–9484.
8