Ketone Formation from Carboxylic Acids by Ketonic Decarboxylation: The Exceptional Case of the Tertiary Carboxylic Acids

Article abstract of DOI:10.1002/chem.201702680

For the reaction mechanism of the ketonic decarboxylation of two carboxylic acids, a β-keto acid is favored as key intermediate in many experimental and theoretical studies. Hydrogen atoms in the α-position are an indispensable requirement for the substrates to react by following this mechanism. However, isolated observations with tertiary carboxylic acids are not consistent with it and these are revisited and discussed herein. The experimental results obtained with pivalic acid indicate that the ketonic decarboxylation does not occur with this substrate. Instead, it is consumed in alternative reactions such as disintegration into isobutene, carbon monoxide, and water (retro-Koch reaction). In addition, the carboxylic acid is isomerized or loses carbon atoms, which converts the tertiary carboxylic acid into carboxylic acids bearing α-proton atoms. Hence, the latter are suitable to react through the β-keto acid pathway. A second substrate, 2,2,5,5-tetramethyladipic acid, reacted by following the same retro-Koch pathway. The primary product was the monocarboxylic acid 2,2,5-trimethyl-4-hexenoic acid (and its double bond isomer), which might be further transformed into a cyclic enone or a lactone. The ketonic decarboxylation product, 2,2,5,5-tetramethylcyclopentanone was observed in traces (<0.2 % yield). Therefore, it can be concluded that the observed experimental results further support the proposed mechanism for the ketonic decarboxylation via the β-keto acid intermediate.

Full text of DOI:10.1002/chem.201702680

A Journal of
Accepted Article
Title: Ketone formation from carboxylic acids by ketonic
decarboxylation: the exceptional case of the tertiary carboxylic
acids
Authors: Borja Oliver-Tomas, Michael Renz, and Avelino Corma
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201702680
Link to VoR: http://dx.doi.org/10.1002/chem.201702680
Supported by

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
Ketone formation from carboxylic acids by ketonic
decarboxylation: the exceptional case of the tertiary carboxylic
acids
Borja Oliver-Tomas,[a] Michael Renz,*[a] and Avelino Corma*[a]
Abstract: For the reaction mechanism of the ketonic Introduction
decarboxylation of two carboxylic acids, a β-keto acid is favored
as key intermediate in many experimental and theoretical studies.
Hydrogen atoms in α-position are an indispensable requirement
for the substrates to react following this mechanism. However,
isolated literature observations with tertiary carboxylic acids are
not in accordance with it and these are revisited and discussed
herein.
The experimental results obtained with pivalic acid indicated that
the ketonic decarboxylation does not occur with this substrate.
Instead, it is consumed in alternative reactions such as
disintegration into iso-butene, carbon monoxide and water (retro-
Koch reaction). In addition, the carboxylic acid is isomerized or
loses carbon atoms which converts the tertiary carboxylic acid into
carboxylic acids bearing α-proton atoms. Hence, the latter are
suitable to react via the β-keto acid pathway.
The ketonic decarboxylation converts two molecules of carboxylic
acid into a ketone together with carbon dioxide and water, one
molecule each (eq. 1).[1,2] When regarding the concept of Green
Chemistry, the reaction can be considered as environmentally
benign since a high percentage of the substrate atoms are
recovered in the product and apart from the carboxylic acid no
further reagents are required. Furthermore, this reaction is
extremely suitable to remove selectively the oxygen content of the
chemical compounds since only one carbon atom is eliminated
together with three oxygen atoms out of four. This fact makes it
predestined for the up-grade of biomass-derived product
mixtures.[3–8] The last oxygen atom in the ketone can be removed
by hydrodeoxygenation via alcohol and olefin to produce linear
alkanes which might serve as fuels or lubricants.[9–12]
A second substrate, 2,2,5,5-tetramethyladipic acid, reacted
following the same retro-Koch pathway. The primary product was
the mono-carboxylic acid 2,2,5-trimethyl-4-hexenoic acid (and its
double bond isomer) which might be further transformed into a
cyclic enone or a lactone. The ketonic decarboxylation product,
2,2,5,5-tetramethylcyclopentanone was observed in traces (<
0.2% yield).
Therefore, it can be concluded that the observed experimental results
further support the proposed mechanism for the ketonic
decarboxylation via the β-keto acid intermediate.
In the biomass transformations an additional desired
characteristic is the formation of a carbon–carbon bond, joining
small molecules into molecules with more advantageous
properties for fuels and chemicals. Therefore, with the intended
change from fossil-based chemistry towards a biomass-based
one, the ketonic decarboxylation has become a highly trendy and
beneficial reaction.
The mechanism of the ketonic decarboxylation has been matter
of interest for a long period of time and has caused some trouble
in literature. As early as 1939 Neunhoeffer and Paschke
postulated a mechanism with a β-keto acid as intermediate
(Scheme 1).[13] For the formation of the latter, one molecule has
to be deprotonated in α-position and this in turn excludes
substrates without any α-hydrogen atoms from reacting. The
mechanistic proposal was based on the colorimetric detection of
the intermediate in an estimated concentration of 1% and it was
in accordance with all experimental observations from literature at
that time.
[a]
B. Oliver-Tomas, Dr. M. Renz, Prof. Dr. A. Corma,
Instituto de Tecnología Química,
Universitat Politècnica de Valencia – Consejo Superior de
Investigaciones Científicas (UPV - CSIC)
Av. de los Naranjos s/n, E-46022 Valencia, Spain
E-mail: acorma@itq.upv.es; mrenz@itq.upv.es
Scheme 1. Mechanism for the ketonic decarboxylation as proposed by
Neunhoeffer and Paschke.[13]
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
A particular significance was attributed to the reaction of
oxide.[30] Instead, oxides having a high lattice energy, such as
alumina, chromia, titania and zirconia, remain unaltered in
presence of carboxylic acids.[30] Hence, zirconium oxide has been
identified experimentally as a suitable catalyst out of several metal
oxides upon its catalytic activity[25,31] and its stability.[11,20] With this
material full conversion was achieved at 400 ºC and several
hundreds of grams per gram of metal oxide were passed through
the catalytic bed with intermediate reactivation steps.
In a theoretical DFT study with this material, again a mechanistic
pathway via the β-keto-acid has been favored kinetically, fully in
accordance with the initial work of Paschke and Neunhoeffer.
Hence, two molecules of carboxylic acid are adsorbed in a
different fashion, one by dehydroxylation and the other by double
tetramethyladipic
acid
derivatives.
Hence,
2,2,5,5-
tetramethyladipic acid (1), without any α-hydrogen atoms, could
not be converted into the cyclopentanone derivative 2 (eq.
2b)[14,15] whereas the corresponding one was obtained from the
3,3,4,4-isomer (eq. 2a).[14] However, in 1962 the transformation of
2,2,5,5-tetramethyladipic acid into the corresponding ketone has
been reported in 52% to 72% yield in the presence of KF or
BaO.[16] These results were clearly incompatible with the widely
accepted β-keto acid mechanism and the authors proposed a
carbon centered anion as intermediate instead (Scheme 2). As
this intermediate would deprotonate immediately other carboxylic
acids and produce formally hydrodecarboxylation products (cf.
Scheme 2, alternative pathway) which are not observed in general, deprotonation (Scheme 3). Then, the resulting dianionic species
this mechanism was modified and the decarboxylation and
carbon-carbon bond formation were proposed to occur in a
concerted fashion.[1]
attacks the acylium fragment nucleophilically forming a carbon–
carbon bond and therewith the β-keto-acid. The latter is
decarboxylated and the resulting enolate protonated. Desorption
of all products closes the catalytic cycle. A very similar
mechanism via the β-keto-acid has been described for rutile
titania by DFT calculation differing only in the moment of the
second deprotonation.[27] Essential reaction steps such as
carbon-carbon bond formation to the β-keto-acid followed by
decarboxylation are identical. The carboxylic acid anhydride may
also be an intermediate, since both precursors for the β-keto-acid
can be formed from one molecule of acid anhydride. Therefore, in
this case the principal kinetic effect should be: to carry out the
reaction under “less humid conditions” avoiding water desorption
within the catalytic cycle.[22] The latter turned out to be highly
endothermic on zirconium oxide.[25]
The reaction via the β-keto-acid intermediate was favored over
the concerted mechanism or radical reactions for substrates
without steric impediments in α-position.[25] It was further assured,
experimentally and theoretically, that the same mechanism also
applies for substrates with an additional methyl substituent in α-
position.[26] This additional group has a strong retarding effect on
the reaction kinetics, for instance one additional methyl group
lowers the reaction rate by a factor of 30 (in the temperature range
from 350 to 400 °C).[26] In addition, it was concluded that the
substrates with a higher steric demand react preferentially at
edges and corners of the catalyst crystals.[26]
The only result in literature which is not in accordance with the
mechanism via a β-keto-acid is the transformation of 2,2,5,5-
tetramethyladipic acid (1) into 2,2,5,5-tetramethylcyclopentanone
(2) mentioned before (eq. 2b).[16] The solution to this
inconsistency seems to be the last missing puzzle piece for an
entire and comprehensive picture for the mechanism of the
ketonic decarboxylation. Therefore, we revisited this reaction
employing the catalyst used in the combined theoretical and
experimental study, i.e., zirconium oxide, and carried out the
reaction under the best conditions used in the 1962-publication,
i.e., in presence of barium oxide. In addition, we selected pivalic
acid (2,2-dimethylpropanoic acid, 3) as a model substrate. This
molecule can be reacted under the same reaction conditions in
the gas phase as valeric and 2-methylbutyric acid whereas
carboxylic acid 1 cannot be volatized easily. Moreover, pivalic
acid (3) has the same number of carbon atoms as both carboxylic
acids used before for the theoretical/experimental study so that
Scheme 2. Mechanism for the ketonic decarboxylation as proposed by Rand et
al.[16] and an alternative pathway from the carbanion to the
hydrodecarboxylation product.
As already mentioned, with the increasing interest due the
relevance to biomass transformation, the ketonic decarboxylation
and its mechanism has become matter of interest again. This is
manifested in the number of catalytic studies appeared
recently[17–21] and the experimental and theoretical evaluations of
the mechanism.[22–29] For the theoretical examination it was
mandatory to “fix” the reaction centers on a crystalline surface to
get reliable information and to avoid excessive freedom for
reactants and catalysts/catalytic sites. Furthermore, it was of
paramount interest that the catalyst was stable under the reaction
conditions and the geometry unchanged after the transformation.
This condition excludes oxides with low lattice energy, which form
bulk carboxylates, such as lead, bismuth, zinc or magnesium
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
Scheme 3. Kinetically favored pathway by DFT calculations for the ketonic decarboxylation. Two molecules adsorb in a different fashion, namely by
dehydroxylation and double deprotonation. A nucleophilic attack forms a carbon–carbon bond and subsequent decarboxylation and protonation the ketone
product. Reproduced from Ref.[32]
similar physical properties (e.g. boiling point, vapor pressure, etc.)
can be expected with respect to the reaction temperature. It has
been reported also for pivalic acid (3) that the corresponding
ketone 2,2,4,4-tetramethyl-3-pentanone (4) was not observed
under standard reaction conditions.[13] With these proposed
experiments, the present study will help to eliminate these last
obstacles for a comprehensive understanding of the ketonic
decarboxylation.
However, the symmetrical ketonic decarboxylation product, i.e.,
2,2,4,4-tetramethyl-3-pentanone (4, eq. 3), was not observed.
This was assured unambiguously by comparison with an
authentic sample. Instead, more than 15 different products were
detected and identified (see Table 1).
Results and Discussion
Pivalic acid (3) was chosen as test molecule and reacted under
standard conditions for the ketonic decarboxylation in the 400 −
450 ºC temperature range. In general, most ketonizations are
carried out even below 400 ºC.[1,2] As catalyst zirconium oxide[33,34]
was employed first, which has turned out to be an active and
stable catalyst for the reaction with valeric acid, i.e., the carboxylic
acid with linear chain and also five carbon atoms.[35] No
conversion was observed under these conditions (eq. 3). The
result should be generalized and, therefore, other classical oxides
were tested for this reaction such as barium oxide (supported on
zirconium oxide) and magnesium oxide.[36] But, again, no
conversion was observed under the standard conditions (eq. 3).
These results were in accordance with literature reports since also
calcium pivalate has been resistant to the conversion into a
ketone,[13] and neither was the symmetrical ketone observed in
cross coupling reactions.[25]
At 550 ºC, the most abundant gas products detected were CO2
with 20% yield, CO with 14% yield, isobutene (12%) and H2 (10%).
Isobutane and methane (~3%) were also observed in the gas
phase, but in lower yield. In the liquid phase, a large variety of
carbonyl compounds (5-12) was obtained which were identified
by GC-MS and some confirmed by comparison with the authentic
sample (compounds 5, 6, 10 and 11). Most of them have a lower
molecular weight than the hypothetical symmetrical ketonic
decarboxylation product.
At a glance, approximately a yield of 20% of ketones was detected
and a superior amount of carbon dioxide, which is in accordance
with the hypothesis that the ketones were obtained by ketonic
decarboxylation. In addition, 12 to 14% of carbon monoxide and
isobutene were obtained (Table 1). The similar yields of the latter
two products suggest that they are formed from a retro-Koch type
reaction (eq. 4). The ability of pivalic acid to undergo a retro-Koch
type reaction was further supported when it was reacted over SiO2,
FeSO4/SiO2, and Fe2O3/SiO2 at 500 ºC: both products were
observed with approximately 90% selectivity in each case (Supp.
Inf., Figure S1).
The reaction set-up for the gas phase reaction over zirconium
oxide allowed to raise the temperature further. This was done to
force the reaction conditions and to facilitate the ketonic
decarboxylation via kinetically less favored pathways (if these
existed) or to identify and allow alternative reaction pathways
which cannot be accessed at standard temperature. Hence, when
increasing reaction temperature to 500 ºC a low conversion of
11% was observed, which reached already 63% at 550 ºC.
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
Table 1. Yields (in mol%) of products observed in the reaction of pivalic
acid (3) at 500 ºC and at 550 ºC over zirconium oxide. Yields of products in
the gas phase are estimated by determination of their concentration in the
gaseous effluent and the concentration of N2 and the total volume of N2,
employed during the experiment as carrier gas. Products in the organic
phase (condensed with an ice bath at the exit of the reactor) were
quantified by GC using dodecane as external standard and are related to
one equivalent of acid 3 for the gaseous products and to two equivalents
for the ketone products (For formulas see Supp. Inf.).
For the ketone products, formation via ketonic decarboxylation
can be discussed, even through the β-keto acid mechanism since
all ketone compounds 6–12 possess an α-hydrogen atom with
respect to the carbonyl group. In any case, pivalic acid (3) has to
undergo rearrangement or fragmentation reaction prior to the
carbon-carbon bond formation, i.e., before reacting with a second,
unchanged pivalic acid molecule.
For establishing the reaction network for the product formation the
same surface species, with the exception of the doubly
deprotonated carboxylic acid as before in the theoretical and
experimental study, should be taken into account, i.e., a
deprotonated one and a dehydroxylated one (eq. 5).
[a]
Catalyst
ZrO2
500
11
ZrO2
550
63
ThO2
490
79
Reaction temperature/ºC
conversion/%
yield/%
yield/%
yield/%
products in the gas phase
carbon dioxide
1.6
1.5
0.0
1.3
0.3
0.2
0.1
0.3
20.1
13.7
9.8
20.0
2.7
carbon monoxide
hydrogen
0.3
iso-butene
11.8
2.0
(2.6)[b]
(5.0)[c]
0.6
When dehydroxylating pivalic acid (3) the resulting carbocationic
species can decompose easily into isobutene (releasing one
proton to the surface) and carbon monoxide in a retro-Koch-type
reaction (eq. 4), providing both molecules in a equimolar ratio as
observed in the reaction over zirconium oxide at 550 ºC or over
other materials at 500 ºC (Supp. Inf., Figure S1). This
fragmentation of pivalic acid (3) has been observed before in the
presence of a strong Brönsted acid (HBr).[38] In addition, pivalic
acid (3) has been used for transcarboxylation reactions together
with tertiary alkanes catalyzed by sulfuric acid.[39] Again, cationic
intermediates are proposed and the transfer of a CO fragment.
On the zirconium oxide surface strong Brönsted acid sites are not
present, only Lewis acidic zirconium sites. This might be one of
the reasons why such a high reaction temperature is required.
However, the bond rupture may also be “assisted” by the surface:
for instance by basic sites, which are responsible for the proton
abstraction in carboxylic acids in α-position, and which might
stabilize the tert.-butyl cation or remove a proton from a methyl
group to form the olefin. In light of the experimental data,
combined with the theoretical knowledge on the surface activity
and the literature references on the reactivity of pivalic acid (3), it
seems reasonable to propose that one of the reaction pathways
for the pivalic acid (3) on zirconium oxide at 550 ºC is the
disintegration via the retro-Koch reaction as described in eq. 4.
This means that isobutene, carbon monoxide and water are
produced in equimolar amounts from one molecule of pivalic acid
(3).
iso-butane
methane
2.9
propylene
1.0
0.03
0.3
others
0.3
products in the condensed phase
2,2,4,4-tetramethyl-3-pentanone (4)
2,2-dimethylpropanal (5)
3,3-dimethylbutanone (6)
2,2-dimethyl-3-pentanone (7)
2,2,5-trimethyl-3-hexanone (8)
2,6-dimethyl-4-heptanone (9)
4-methyl-2-pentanone (10)
3-methylbutanone (11)
3-methyl-3-butenone (12)
(total ketones and aldehydes)
3-methylbutyric acid (13)
others
0.00
0.33
0.00
0.00
0.83
0.09
0.00
0.00
1.78
(3.03)
0.09
1.83
0.00
4.14
6.86
2.09
2.56
0.68
0.52
0.70
5.25
(22.8)
0.53
1.95
0
3
9
0.7
15
[d]
[d]
[d]
[d]
(27)
[d]
20
[a] From literature [37]. [b] Indicated as butanes. [c] Indicated as butenes. [d]
Not specified.
Isobutane was detected in the gas phase with 2% yield (Table 1).
Formally, this is the decarboxylation product of pivalic acid (3), i.e.,
it is co-produced with carbon dioxide (eq. 6). Hence, this is a
plausible formation under the present reaction conditions and
may be induced by deprotonation of the acid and further promoted
by surface protons which might stabilize the leaving group. Once
formed, isobutane may be fragmented into propene and methane
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
(eq. 6) which both were detected in the gas phase in 1.0% and
2.9% yield, respectively (Table 1). Such fragmentation has been
observed at high temperature over solid materials.[40–42] However,
having in mind the product ratios, it becomes evident that further
sources (precursors) for the methane formation must exist.
gaseous effluent of the reaction at elevated temperature (550 ºC)
and with nitrogen as carrier gas (Table 1).
The reaction of pivalic acid (3) over a metal oxide has been
carried out before at high temperatures (490 ºC) over thorium
oxide with similar results.[37] The main products at a conversion of
79% were t-butyl isobutyl ketone (8) and t-butyl methyl ketone (6)
with selectivities of 19% and 11%, respectively (Table 1). Also in
this case, the symmetrical ketonic decarboxylation product, i.e.,
2,2,4,4-tetramethyl-3-pentanone (4) was not identified in the
product mixture. This means that similar reaction pathways may
apply here as for the ZrO2 case, although a different interpretation
has been originally provided by the Authors, possibly due to the
fact that the most indicative products, i.e., iso-butene and carbon
monoxide, have been observed in lower concentration.
These experimental results with pivalic acid (3) over zirconium
oxide, and also with thorium oxide, indicate that the ketonic
decarboxylation with two molecules of the substrate does not
occur, not even under forced reaction conditions. Instead, the
substrate is consumed in alternative surface-catalyzed reactions
such as base-induced disintegration into iso-butene, carbon
monoxide and water (retro-Koch reaction). This reactivity follows
the prediction of the DFT calculations that show high barriers for
alternative ketonic decarboxylation reactions without abstraction
of α-hydrogen atoms. Activation energies are so high that carbon-
carbon bond scission can compete.
The second critical reaction for the ketonic decarboxylation
reported in literature with respect to the β-keto-acid mechanism
was the transformation of carboxylic acid 1. Therefore, also this
reaction was revisited. As this acid cannot be volatilized without
decomposition, the reaction had to be carried out under different
conditions, i.e., with a different reaction setup. Hence, the acid
was treated in a batch reactor, from which lower-boiling-point
products were distilled off, as it has been employed in the
transformation of the unsubstituted adipic acid into
cyclopentanone.[44,45]
For the liquid carbonyl compounds, formation by ketonic
decarboxylation was assumed as a working hypothesis. The
second product of this reaction, carbon dioxide, is present in the
gaseous effluent in sufficient yield, which is in accordance with
the hypothesis. However, for enabling the reaction via the β-keto
acid pathway, at least part of the pivalic acid (3) has to undergo
modifications in its carbon skeleton: either undergo
rearrangement reactions or lose carbon atoms. This is clearly in
line with the experimental results since the symmetrical ketone 4
obtained from two molecules of carboxylic acid 3 was not
observed. Very clear evidence for a rearrangement of the carbon
skeleton is the formation of ketone 9, i.e., 2,6-dimethylbutanone,
which is observed in 0.7% yield. This product can be formed from
two molecules of 3-methylbutyric acid (13), which was detected in
the liquid product mixture in traces (Table 1), via the established
β-keto-acid mechanism since this substrate contains two α-
hydrogen atoms. Consequently, for the formation of ketone 9 the
reaction sequence depicted in eq. 7 was assumed involving
skeleton rearrangement of the carboxylic acid 3 into carboxylic
acid 13 followed by ketonic decarboxylation via the β-keto-acid
mechanism. Probably, the rearrangement occurs in one of the
surface stabilized species, further supported by other surface
functionalities, but for a more detailed description of the
mechanism at molecular level theoretical calculations should be
carried out.
When 2,2,5,5-tetramethyladipic acid (1) was contacted with
zirconium oxide (10 wt%) or with barium oxide (10 mol%) at 350
to 370 ºC, or heated in absence of any catalyst, gas evolution was
observed (Figure 1). Thereby, it can be seen that the thermal
reaction proceeded smoothly, but barium oxide, and even more
zirconium dioxide, accelerated the reaction rate. The analysis of
the gas composition indicated immediately that the reaction
occurred was not the standard ketonic decarboxylation since
carbon dioxide was obtained only as a minor product together with
more than 80% yield of carbon monoxide in each case (Table 2).
Having in mind this carbon skeleton rearrangement, also the
explanation of the formation of ketone 8 is straightforward: it is the
cross-coupling product of the rearranged acid and the original one.
In the other ketone products carbon atoms are missing indicating
fragmentation reactions either at the substrate stage or in the
product (formed with the rearranged acid). If it is assumed that
these reactions occur already in the substrate, again the ketonic
decarboxylation may occur since all the ketones involve α-
hydrogen atoms.
Apart from the ketone product one aldehyde was observed:
pivaldehyde (5). The latter can be produced by hydrogenation in
an analogous way as reported for benzoic acid which is
hydrogenated to benzaldehyde over zirconium oxide.[43] The
aldehyde was also observed experimentally with a 32% yield at
400 ºC when pivalic acid (3) was fed together with hydrogen as
carrier gas. The required hydrogen was also observed in the
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
When establishing the reaction network, 2,2,5-trimethyl-4-
hexenoic acid (14), and its double bond isomer 15, can be
assumed to be primary products. These molecules can be formed
by a retro-Koch-type reaction (eq. 8), in an analogous manner to
the reaction proposed above for the pivalic acid (3) transformation
into carbon monoxide and iso-butene. This assumption is further
supported by the detection of carbon monoxide in large amounts
in the gas phase effluent.
In the present case the selectivity towards this reaction pathway
was considerably higher with 75%, when considering ketone and
lactone as secondary, 2,2,5-trimethyl-4-hexenoic acid (14)-
derived products. Hence, the lactone 17 can be obtained by a
simple intramolecular addition of the carboxylic acid to the double
bond (eq. 9).[46] This cyclization has been described in literature
and was utilized for the synthesis of an authentic sample. The
ketone 16 can be formed by an intramolecular attack of the double
bond onto the carbonyl group (eq. 10), although at present the
mechanism has not been explored in detail. This ketone has also
been identified by comparison with an authentic sample.
Figure 1. Yield of the gas evolution during the thermal treatment of 2,2,5,5-
tetramethyladipic acid (1) and in the presence of ZrO2 (10 wt%) or BaO (10
mol%) at 350 – 370 ºC. The composition of the gas is specified in Table 2.
An organic distillate of approximately 45 to 50 wt% was obtained
and 5 to 10 wt% of water whereas the gaseous phase accounted
for more or less 10 wt% in each case. In the distillate, the ketonic
decarboxylation
product,
i.e.,
ketone
2,2,5,5-
tetramethylcyclopentanone (2), was not found by comparison with
an authentic sample in the GC analysis when carrying out the
reaction with BaO as catalyst and in less than 0.2% yield with ZrO2
as catalyst or without any catalyst. A six-membered ring ketone,
namely 3,6,6-trimethylcyclohex-2-en-1-one (16) was obtained
instead,
tetramethylvalerolactone
together
with
a
(17),
lactone,
and
namely
two unsaturated
α,α,δ,δ-
monocarboxylic acid isomers as main products for the catalyzed
reactions, 2,2,5-trimethyl-4-hexenoic acid (14, main isomer) and
2,2,5-trimethyl-5-hexenoic acid (15, Scheme 4; for product ratios
see Table 2).
The actual ketonic decarboxylation product, the ketone 2, was not
detected as one of the main products, but in traces in two cases.
An authentic sample was synthesized by an alternative pathway,
namely by complete methylation of cyclopentanone in the α-
positions. With the substance in hand, the compound was
identified unambiguously in the product mixtures of the thermal
and the ZrO2 catalyzed reactions, with a very low yield in both
cases, below 0.2%. When the reaction was carried out in the
presence of BaO, ketone 2 was not observed at all. These results
are clearly opposed to the ones of Rand et al.[16]
Already in 1966, Eberson proposed that the catalytic
transformation of 2,2,5,5-tetramethyladipic acid (1) produced
3,6,6-trimethylcyclohex-2-en-1-one (16), instead of 2,2,5,5-
cyclopentanone (2).[15] An identical melting point of the 2,4-
dinitrophenylhydrazone was obtained as for the compound
isolated by Rand et al., and a similar refractive index. Both
analytical data were used by Rand for the identification of the
product.[16] Our results further support those of Eberson and the
mono-acids 14 and 15, isolated and characterized herein, further
explain the formation of this ketone.
Scheme 4. Product mixture observed for the treatment of dicarboxylic acid 1 in
the presence of ZrO2, BaO or without any catalyst. Formation of 2,2,5-trimethyl-
4-hexenoic acid (14), 2,2,5-trimethyl-5-hexenoic acid (15), 3,6,6-
trimethylcyclohex-2-en-1-one (16), α,α,δ,δ-tetramethylvalerolactone (17) were
detected. For product ratios see Table 2.
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
instead of the isomer 15, the five-membered cycle is formed
(connecting carbon atoms 1 and 5, instead of 1 and 6).
Hydrogenation of the double bond of the primary product 2,2,5,5-
tetramethylcyclopent-3-en-1-one, or reduction of the intermediate
carbo cation, provides ketone 2. It is clear that both pathways are
not suitable to form selectively ketone 2 from diacid 1, but they
may explain well the formation of the compound in traces.
Table 2. Results for the thermal treatment of 2,2,5,5-tetramethyladipic acid
(1) and in the presence of ZrO2 (10 wt%) or BaO (10 mol%) at 350 – 370
ºC. For the structure of the organic compounds see Scheme 4.
thermal
ZrO2
BaO
yield of products
gaseous products/wt%[a]
organic distillate/wt%
water/wt%
6.6
26.0
4.4
11.8
60.5
5.3
8.0
51.2
8.3
Conclusions
The attempted ketonic decarboxylation of α,α-disubstituted
carboxylic acids has been revisited employing 2,2,5,5-
tetramethyladipic acid (1) and pivalic acid (3) as starting materials.
For both substrates it was found that the preferred reaction
pathway was a retro-Koch reaction in which the carboxylic acid
group is converted into carbon monoxide and water, producing
the olefin as main product. The primary retro-Koch product
obtained from diacid 1 helps to understand controversial literature
reports on the reaction since its identification allows the proposal
of reasonable reaction mechanism to alternative products. The
typical symmetrical ketone products were not obtained, except for
two cases when the cyclic ketone 2 was detected in traces which,
probably, was formed by an alternative pathway involving
demethylation and methylation reactions or cyclization of the
primary product followed by hydrogenation. Therefore, it can be
concluded that all experimental observations were in accordance
with the proposed mechanism for the ketonic decarboxylation, i.e.,
the formation of the β-keto acid and subsequent decarboxylation,
which excludes α,α-disubstituted carboxylic acids as potential
substrates. As a consequence, the β-keto acid mechanism can
be regarded as universal mechanism of the ketonic
decarboxylation of carboxylic acids.
residue/wt%
51.6
88.4
21.7
99.3
29.7
97.2
weight balance/%
gas phase yield
carbon monoxide/mol%
carbon dioxide/mol%
36.0
7.0
60.1
8.7
60.6
1.5
composition of the organic mixture
substrate 1/%
ketone 2/%
9.70[b]
0.44
6.10
47.8
0
0.59
0.34
57.1
12.2
17.5
12.3
5
n. d.[c]
35
acid 14 and 15/%
ketone 16/%
lactone 17/%
others/%
20
Experimental Section
15
General.- Reagents and solvents were purchased from standard chemical
suppliers as stated: dimethyl sulfoxide (Aldrich, 99%), dichloroethane
(Aldrich, 98%), diethyl ether (Scharlau, reagent grade), pentane (Aldrich,
98%), acetic acid (Aldrich, glacial), cyclopentanone (Acros, 99%), 4,4-
dimethyl-2-cyclohexen-1-one (Acros, 97%), MeMgBr in diethyl ether (3 M,
Aldrich), methyl iodide (Aldrich, 99%), pyridinium chlorochromate (Fluka
98%), pivalic acid (3, Aldrich, 99%), adipic acid (Acros, 99%), silver triflate
(Alfa Aesar, 98%), KOH (Scharlau, reagent grade), H2SO4 (Aldrich, 98%),
HCl (Aldrich, 37%), hydrogen peroxide (35%, Aldrich), FeSO4 · 7 H2O
(Aldrich, 99%), Fe(NO3)3 · 9 H2O (Aldrich, 98%), NaHCO3 (Acros, 99%),
Ba(NO3)2 (Probus, pure grade), BaO (Aldrich, 99.99%), NaOH (Scharlau,
synthesis grade). Monoclinic zirconium oxide (m-ZrO2) was obtained as
pellets from ChemPur, Germany, with a surface area of 100 m2 g–1, MgO
was received as a powder from Riedel-de-Haën, Germany, with a surface
area of 53 m2 g–1 and SiO2 was obtained as pellets from ChemPur,
36.0
25
[a] estimated from volume and composition. [b] 9.11% was recovered as
anhydride. [c] Compound was not detected in GC and GC-MS by
comparison with the authentic sample.
In view of these experimental results and literature reports, it can
be concluded that it is highly unlikely that carboxylic acid 1 is
converted directly into the corresponding ketone in a selective
manner under classical ketonic decarboxylation conditions.
Traces may be produced by alternative mechanisms. Possible
precursors are 2,2,5-trimethylcyclopentanone or 2,2,5,5-
tetramethylcyclopent-3-en-1-one which have been observed both
in the product mixture. The former can be produced by
demethylation of the starting material 1, classical ketonic
decarboxylation involving the β-keto acid, and re-methylation. The
second alternative pathways comprises the same reaction steps
as the formation of ketone 16. When reacting carboxylic acid 14
Germany, with a surface area of 240 m2 g–1
.
The organic liquids and solids were analyzed on a Agilent 7890A
apparatus equipped with a HP-5 column (30 m x 0.320 mm x 0.25 µm) and
the substances identified on a GC-MS apparatus Agilent 6890N, equipped
with the same column and a mass selective detector Agilent Technologies
5973 Network. The gases were analyzed on
a Varian 3800 gas
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
chromatograph equipped with three columns and three detectors. The first
column was a Ultimetal Molsieve 5 Å 80-100 Mesh (1.5 m x 1/16” x 1 mm)
connected to a thermal conductivity detector (TCD) for hydrogen analysis.
The second column was a Molsieve 13X 80-100 Mesh (1.2 m x 1/16” x 1
mm) for other permanent gases also connected to a TCD and the third
column an Al2O3 MAPD (25 m x 0.32 mm x 5 µm) for hydrocarbon
separation with subsequent quantification with a flame ionization detector
(FID). For the exact calculation of conversion and yields see Supp. Inf.
Isolation and identification of 2,2,5-trimethylhex-4-enoic acid (14) from the
product mixture[51].- The liquid organic product (2.50 g) was dissolved in
dichloromethane and the solution was extracted with a saturated aqueous
NaHCO3 solution (2 x 25 mL). The combined aqueous solutions were
acidified with hydrochloric acid to
a pH of 2 and extracted with
dichloromethane (25 mL). The solvent was evaporated and 2,2,5-
trimethylhex-4-enoic acid (14, 0.175 g, 1.12 mmol, corresponding to 7 wt%
of the crude product mixture) obtained as colorless crystals.
Synthesis of 2,2,5,5-tetramethyladipic acid[47] (1).- In a beaker with
mechanical stirring and cooled to 0 ºC, water (600 mL), sulfuric acid (7.5
mL) and pivalic acid (3, 51.0 g, 500 mmol) were placed. During 15 min
hydrogen peroxide (35%, 43 mL, 500 mmol), and a solution of FeSO4 · 7
H2O (139 g, 500 mmol) and sulfuric acid (27.5 mL) in 288 mL of distilled
water were added. The mixture was concentrated to 250 mL, the solids
collected by filtration, washed with cold distilled water, and dried in vacuum.
The obtained solid was recrystallized twice from acetic acid (1 mL for each
g of solid) and 2,2,5,5-tetramethyladipic acid obtained as colorless crystals
(12.2 g, 60.3 mmol, 24% yield). The product was dissolved and analyzed
by inductively coupled plasma optical emission spectrometry (ICP OES)
and an iron content from 2421 to 5450 ppm (2.42 to 5.45 mg/g product)
determined.
1H NMR (300 MHz, CDCl3):[51] δ = 11.8 (s, 1H), 5.12 (tdq, J=7.3, 2.8, 1.4
Hz, 1H), 2.25 (dt, J=7.6, 1.2 Hz, 2H), 1.71 (d, J=1.3 Hz, 3H), 1.61 (d, J=1.4
Hz, 3H), 1.18 (s, 6H).- 13C NMR (75 MHz, CDCl3):[51] δ = 184.6, 134.6,
119.6, 42.7, 38.4, 26.0, 24.5, 17.9.- MS: m/z (%): 156 (23) [M+], 27 (6),
41(50), 55 (9), 69 (100), 88 (43).
Synthesis of 3,6,6-trimethylcyclohex-2-en-1-one[52] (16).- 3,6,6-
trimethylcyclohex-2-en-1-one (16) was synthesized in two steps. Under
nitrogen atmosphere and at – 78 ºC a solution of 4,4-dimethyl-2-
cyclohexen-1-one (2.00 g, 16.1 mmol) in 20 mL anhydrous diethyl ether
was placed in a round-bottom flask and stirred magnetically. The Grignard
reagent MeMgBr in diethyl ether (3 M, 5.7 mL, 17.2 mmol) was added
dropwise with a syringe and the mixture was allowed to warm to room-
temperature. After two hours distilled water (10 mL) was added dropwise.
The organic phase was washed with water (20 mL), dried with MgSO4 and
the solvent evaporated. The liquid obtained was a mixture of 84 to 16 of
product (1,4,4-trimethylcyclohex-2-en-1-ol) to substrate that was used as
such for the next step. In a round-bottom flask a solution of pyridinium
chlorochromate (4.30 g, 20.0 mmol) in dichloromethane (30 mL) were
placed and a solution of the product mixture obtained in the previous step
(1.4 g, 8.21 mmol of 1,4,4-trimethylcyclohex-2-en-1-ol) in dichloromethane
(10 mL) added. After stirring magnetically during two hours diethyl ether
was added (40 mL). By this, two phases were obtained, one ethereal and
another black viscous one. The latter is extracted with diethyl ether (20
mL). The combined organic phases were extracted with a 5% aqueous
sodium hydroxide solution (2 x 100 mL), with a 5% aqueous hydrogen
chloride solution (100 mL), and a saturated aqueous NaHCO3 (2 x 50 mL).
After drying with MgSO4 the solvent was evaporated and 3,6,6-
trimethylcyclohex-2-en-1-one (5, 1.11 g, 6.51 mmol, 79% yield) obtained
as a colorless liquid.
Reaction of 2,2,5,5-tetramethyladipic acid (1) in the presence of ZrO2, BaO
or without any catalyst (thermal).- With the aim to assure that the reaction
conditions would work well for the ketonic decarboxylation, the reaction
was carried out with unsubstituted adipic acid. A literature procedure was
selected and tested with adipic acid as substrate.[45] At a reaction
temperature of 350 ºC 85% of cyclopentanone and the corresponding
amount of carbon dioxide were obtained after 46 min of reaction time.
Apparatus and conditions were employed for the transformation of 2,2,5,5-
tetramethyladipic acid (1). In a round-bottomed flask, placed on a heating
mantle and connected to a gas burette, 2,2,5,5-tetramethyladipic acid
(10.0 g, 49.4 mmol) and optionally BaO (0.769 g, 5.02 mmol, 5.02 equiv.)
or ZrO2 (1.00 g, 10 wt%) were filled and the mixture heated to 350 – 370
ºC (internal temperature) under magnetic stirring. The gas evolution was
monitored (Figure 1), the gas was analyzed offline as indicated above and
the compositions are summarized in Table 2. A mainly liquid product
mixture and water were obtained as distillate and a solid dark-colored
residue. The composition of the product mixture is also summarized in
Table 2. 2,2,5,5-cyclopentanone was observed only in traces (< 0.2%
yield) as it was proven unambiguously by comparison with an authentic
sample (synthesis see below). 2,2,5-trimethylhex-4-enoic acid was
isolated in satisfactory purity from the crude product mixture to confirm its
structure by NMR and comparison with literature data (see below). 3,6,6-
trimethyl-2-cyclohexen-1-one (16) and the lactone 3,3,6,6-tetramethyl-2H-
pyran-2-one (17) were synthesized by a literature procedure and the data
of the authentic sample matched perfectly with the substance in the crude
reaction mixture (see below).
1H NMR (300 MHz, CDCl3):[53] δ = 5.74 (h, J=1.5 Hz, 1H), 2.27 (tdd, J=6.2,
1.7, 0.9 Hz, 2H), 1.90 (q, J=1.0 Hz, 3H), 1.78 (t, J=6.1 Hz, 2H), 1.06 (s,
6H).- 13C NMR (75 MHz, CDCl3):[53] δ = 204.4, 160.2, 125.0, 40.1, 36.0,
28.4, 24.1, 24.0.- MS: m/z (%): 138 (17) [M+], 39 (18), 54 (21), 67 (6), 82
(100), 110 (14).
Identification of 3,3,6,6-tetramethyl-2H-pyran-2-one (17).- The lactone
3,3,6,6-tetramethyl-2H-pyran-2-one (17) was identified in the crude
reaction mixture (organic liquid) by 13C NMR by comparison with an
authentic sample.[46] In a round-bottom flask, equipped with a magnetic
stirring bar and heated to 83 ºC, a solution of 2,2,5-trimethylhex-4-enoic
acid (0.102 g; 92% purity, GC, 0.654 mmol) in dichloroethane (4 mL) was
placed. Silver triflate (8.0 mg, 0.03 mmol) was added and the mixture
stirred for 15 h. The crude reaction mixture was extracted three times with
a saturated aqueous sodium bicarbonate solution, dried with MgSO4, and
the solvent rota-evaporated. A slightly yellow solid (17, 90 mg, 0,562 mmol;
94% yield) was obtained with 97% (GC) purity.
Synthesis of 2,2,5,5-tetramethylcyclopentanone[48] (2).- In a round-bottom
flask, equipped with reflux condenser, dimethyl sulfoxide (19 mL) was
placed and heated to 50 ºC. Cyclopentanone (0.846 g, 10.1 mmol), methyl
iodide (5.0 mL, 80.3 mmol) and potassium hydroxide (11,2 g, 200 mmol)
were added and the mixture stirred magnetically for one hour. The mixture
was cooled to 0 ºC and extracted with pentane (3 x 7 mL). The combined
three organic phases were washed with distilled water (3 x), dried with
MgSO4 and the solvent removed by distillation. The colorless oil obtained
as
crude
reaction
product
consisted
in
pure
2,2,5,5-
13C NMR (75 MHz, CDCl3):[46] δ = 177.7, 82.5, 37.2, 32.0, 31.5, 28.9, 27.6.-
MS: m/z (%): 41 (34), 56 (100), 70 (18), 95 (11), 113 (17), 141 (9).
tetramethylcyclopentanone (2, 0.879 g, 6.28 mmol, 62% yield).
1H NMR (300 MHz, CDCl3):[49] δ = 1.68 (s, 4H), 0.95 (s, 12H).- 13C NMR
(75 MHz, CDCl3):[50] δ = 226.7, 45.1, 34.7, 24.8.- MS: m/z (%): 140 (33)
[M+], 27 (7), 41 (39), 56 (100), 69 (20), 97 (5).
Catalyst preparation for fixed-bed continuous flow reactions.- m-ZrO2 and
SiO2 were crushed and sieved, employing the 0.4 – 0.8 mm fraction as
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
catalyst. MgO was pelletized, crushed and sieved and the 0.4 – 0.8 mm
fraction was used. BaO(3 wt%)/ZrO2 was obtained by the incipient wetness
impregnation method: an aqueous solution of Ba(NO3)2 was prepared with
the corresponding concentration to achieve the desired BaO content in the
solid. The m-ZrO2 (0.4 – 0.8 mm) was impregnated, dried at room
temperature in vacuum during 2 h and afterwards overnight in an oven at
100 °C. The material was calcined in-situ in the reactor under air flow
(atmospheric pressure) at 550 °C during 2 h (heating rate of 3 K/min).
[14]
[15]
[16]
E. H. Farmer, J. Kracovski, J. Chem. Soc. 1927, 680–685.
L. Eberson, Tetrahedron Lett. 1966, 7, 223–226.
L. Rand, W. Wagner, P. O. Warner, L. Robert Kovac, J. Org. Chem.
1962, 27, 1034–1035.
[17]
[18]
[19]
N. Aranda-Pérez, M. P. Ruiz, J. Echave, J. Faria, Appl. Catal. A
Gen. 2017, 531, 106–118.
J. A. Bennett, C. M. A. Parlett, M. A. Isaacs, L. J. Durndell, L. Olivi,
A. F. Lee, K. Wilson, ChemCatChem 2017, 1648–1654.
S. D. Stefanidis, S. A. Karakoulia, K. G. Kalogiannis, E. F.
Iliopoulou, A. Delimitis, H. Yiannoulakis, T. Zampetakis, A. A.
Lappas, K. S. Triantafyllidis, Appl. Catal. B Environ. 2016, 196, 155–
173.
Reaction of pivalic acid (3) in the presence of ZrO2, BaO/ZrO2, MgO, SiO2,
Fe(SO4)/SiO2 or Fe2O3/SiO2.- For the preparation of Fe(SO4)/SiO2 or
Fe2O3/SiO2 see Supp. Inf. The reactions were carried out in a tubular fixed-
bed continuous flow reactor described in a previous work.[25] The reactor
was loaded with 2.50 g of catalyst (0.4 – 0.8 mm) diluted with silicon
carbide. Pivalic acid was heated to 45 ºC inside a syringe and fed at a rate
of 0.15 mL min–1 (8.5 g h–1, WHSV = 3.4 h–1). 50 mL min–1 of nitrogen gas
was used as carrier gas, or alternatively 50 mL min–1 of hydrogen in the
case of ZrO2. The reactions were carried out at atmospheric pressure and
a temperature between 400 and 550 ºC. The products streams were
condensed and analyzed off-line.
[20]
[21]
[22]
Y. Lee, J.-W. Choi, D. J. Suh, J.-M. Ha, C.-H. Lee, Appl. Catal. A
Gen. 2015, 506, 288–293.
R. A. L. Baylon, J. Sun, K. J. Martin, P. Venkitasubramanian, Y.
Wang, Chem. Commun. 2016, 52, 4975–4978.
Y. Woo, Y. Lee, J.-W. Choi, D. J. Suh, C.-H. Lee, J.-M. Ha, M.-J.
Park, Ind. Eng. Chem. Res. 2017, 56, 872–880.
A. V Ignatchenko, J. Phys. Chem. C 2011, 115, 16012–16018.
A. V Ignatchenko, E. I. Kozliak, ACS Catal. 2012, 2, 1555–1562.
A. Pulido, B. Oliver-Tomas, M. Renz, M. Boronat, A. Corma,
ChemSusChem 2013, 6, 141–151.
[23]
[24]
[25]
Acknowledgements
[26]
B. Oliver-Tomas, F. Gonell, A. Pulido, M. Renz, M. Boronat, Catal.
Sci. Technol. 2016, 6, 5561–5566.
The authors thank MINECO (CTQ2015-67591-P and Severo
Ochoa program, SEV-2016-0683) and Generalitat Valenciana
(PROMETEO II/2013/011 Project) for funding this work. B. O.-T.
is grateful to the CSIC (JAE program) for his PhD fellowship.
[27]
[28]
[29]
[30]
S. Wang, E. Iglesia, J. Catal. 2017, 345, 183–206.
S. Tosoni, G. Pacchioni, J. Catal. 2016, 344, 465–473.
G. Pacchioni, ACS Catal. 2014, 4, 2874–2888.
R. Pestman, R. M. Koster, A. van Duijne, J. A. Z. Pieterse, V.
Ponec, J. Catal. 1997, 168, 265–272.
Keywords: carbon-carbon bond scission • carbon monoxide •
cyclization • lactonization • retro-Koch reaction
[31]
[32]
K. Parida, H. K. Mishra, J. Mol. Catal. A Chem. 1999, 139, 73–80.
L. M. Orozco, M. Renz, A. Corma, ChemSusChem 2016, 9, 2430–
2442.
[1]
[2]
M. Renz, European J. Org. Chem. 2005, 979–988.
T. N. Pham, T. Sooknoi, S. P. Crossley, D. E. Resasco, ACS Catal.
2013, 3, 2456–2473.
[33]
[34]
[35]
Ketones, 1982, JP 57197237.
K. Matsuoka, K. Tagawa, Ketones, 1985, JP 61207354.
A. Corma, B. Oliver-Tomas, M. Renz, I. L. Simakova, J. Mol. Catal.
A Chem. 2014, 388–389, 116–122.
[3]
C.-L. Cai, Q.-Y. Liu, T.-J. Wang, Q. Zhang, L.-L. Ma, N. Shi, J. Tan,
K. Li, Chem. Ind. For. Prod. 2015, 35, DOI 10.3969/j.issn.0253-
2417.2015.06.025.
[36]
[37]
M. Renz, European J. Org. Chem. 2005, 2005, 979–988.
A. L. Miller, N. C. Cook, F. C. Whitmore, J. Am. Chem. Soc. 1950,
72, 2732–2735.
[4]
[5]
K. Wu, Y. Wu, Y. Chen, H. Chen, J. Wang, M. Yang,
ChemSusChem 2016, 9, 1355–1385.
D. E. Resasco, B. Wang, S. Crossley, Catal. Sci. Technol. 2016, 6,
2543–2559.
[38]
[39]
J. T. D. Cross, V. R. Stimson, Chem. Commun. 1966, 350–351.
V. Lazzeri, R. Jalal, R. Poinas, R. Gallo, S. Delavarenne, New J.
Chem. n.d., 16, 521–524.
[6]
[7]
[8]
E. F. Iliopoulou, Curr. Org. Synth. 2010, 7, 587–598.
D. Y. Murzin, I. L. Simakova, Catal. Ind. 2011, 3, 218–249.
M. J. Climent, A. Corma, S. Iborra, Green Chem. 2014, 16, 516–
547.
[40]
[41]
G. Wang, H. Wang, H. Zhang, Q. Zhu, C. Li, H. Shan,
ChemCatChem 2016, 8, 3137–3145.
D. P. Estes, G. Siddiqi, F. Allouche, K. V Kovtunov, O. V Safonova,
A. L. Trigub, I. V Koptyug, C. Copéret, J. Am. Chem. Soc. 2016,
138, 14987–14997.
[9]
I. L. Simakova, D. Y. Murzin, J. Energy Chem. 2016, 25, 208–224.
J. C. Serrano-Ruiz, D. Wang, J. A. Dumesic, Green Chem. 2010,
12, 574–577.
[10]
[42]
[43]
[44]
[45]
U. Rodemerck, S. Sokolov, M. Stoyanova, U. Bentrup, D. Linke, E.
V. Kondratenko, J. Catal. 2016, 338, 174–183.
T. Yokoyama, T. Setoyama, N. Fujita, M. Nakajima, T. Maki, K. Fujii,
Appl. Catal. A Gen. 1992, 88, 149–161.
[11]
[12]
[13]
A. Corma, B. Oliver-Tomas, M. Renz, I. L. Simakova, J. Mol. Catal.
A Chem. 2014, 388–389, 116–122.
A. Corma, M. Renz, C. Schaverien, ChemSusChem 2008, 1, 739–
741.
J. F. Thorpe, G. A. R. Kon, in Org. Synth., John Wiley & Sons, Inc.,
2003.
O. Neunhoeffer, P. Paschke, Berichte der Dtsch. Chem.
Gesellschaft (A B Ser. 1939, 72, 919–929.
M. Renz, A. Corma, European J. Org. Chem. 2004, 2004, 2036–
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
2039.
2002, 85, 1523–1545.
[46]
[47]
C.-G. Yang, N. W. Reich, Z. Shi, C. He, Org. Lett. 2005, 7, 4553–
[50]
[51]
J. B. Stothers, C. T. Tan, Can. J. Chem. 1974, 52, 308–314.
K. Takenaka, S. C. Mohanta, M. L. Patil, C. V. L. Rao, S. Takizawa,
T. Suzuki, H. Sasai, Org. Lett. 2010, 12, 3480–3483.
4556.
D. D. Coffman, E. L. Jenner, R. D. Lipscomb, J. Am. Chem. Soc.
1958, 80, 2864–2872.
[52]
[53]
W. G. Dauben, D. M. Michno, J. Org. Chem. 1977, 42, 682–685.
J.-M. Vatèle, Tetrahedron 2010, 66, 904–912.
[48]
[49]
E. Langhals, H. Langhals, Tetrahedron Lett. 1990, 31, 859–862.
H. Giera, R. Huisgen, E. Langhals, K. Polborn, Helv. Chim. Acta
This article is protected by copyright. All rights reserved.

10.1002/chem.201702680
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Carbon-carbon bond scission instead
of carbon-carbon bond formation:
when tertiary carboxylic acids are
reacted, a retro-Koch reaction occurs
instead of ketonic decarboxylation.
For the latter, experimental results
further support a mechanism via the
β-keto acid intermediate. The primary
retro-Koch product helps to resolve
controversial reports on the reaction
since it allows the proposal of
Borja Oliver-Tomas, Michael Renz,* and
Avelino Corma*
Page No. – Page No.
Ketone formation from carboxylic
acids by ketonic decarboxylation: the
exceptional case of the tertiary
carboxylic acids
reasonable pathways to secondary,
obtained products.
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
((Insert TOC Graphic here; max. width: 11.5 cm; max. height: 2.5 cm))
Page No. – Page No.
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.