Carbon–Carbon Bond Formation and Hydrogen Production in the Ketonization of Aldehydes

Article abstract of DOI:10.1002/cssc.201600654

Aldehydes possess relatively high chemical energy, which is the driving force for disproportionation reactions such as Cannizzaro and Tishchenko reactions. Generally, this energy is wasted if aldehydes are transformed into carboxylic acids with a sacrificial oxidant. Here, we describe a cascade reaction in which the surplus energy of the transformation is liberated as molecular hydrogen for the oxidation of heptanal to heptanoic acid by water, and the carboxylic acid is transformed into potentially industrially relevant symmetrical ketones by ketonic decarboxylation. The cascade reaction is catalyzed by monoclinic zirconium oxide (m-ZrO₂). The reaction mechanism has been studied through cross-coupling experiments between different aldehydes and acids, and the final symmetrical ketones are formed by a reaction pathway that involves the previously formed carboxylic acids. Isotopic studies indicate that the carboxylic acid can be formed by a hydride shift from the adsorbed aldehyde on the metal oxide surface in the absence of noble metals.

Full text of DOI:10.1002/cssc.201600654

DOI: 10.1002/cssc.201600654
Full Papers
Carbon–Carbon Bond Formation and Hydrogen Production
in the Ketonization of Aldehydes
[
a]
Lina M. Orozco, Michael Renz,* and Avelino Corma*
Aldehydes possess relatively high chemical energy, which is
the driving force for disproportionation reactions such as Can-
nizzaro and Tishchenko reactions. Generally, this energy is
wasted if aldehydes are transformed into carboxylic acids with
a sacrificial oxidant. Here, we describe a cascade reaction in
which the surplus energy of the transformation is liberated as
molecular hydrogen for the oxidation of heptanal to heptanoic
acid by water, and the carboxylic acid is transformed into po-
tentially industrially relevant symmetrical ketones by ketonic
decarboxylation. The cascade reaction is catalyzed by mono-
clinic zirconium oxide (m-ZrO ). The reaction mechanism has
2
been studied through cross-coupling experiments between dif-
ferent aldehydes and acids, and the final symmetrical ketones
are formed by a reaction pathway that involves the previously
formed carboxylic acids. Isotopic studies indicate that the car-
boxylic acid can be formed by a hydride shift from the ad-
sorbed aldehyde on the metal oxide surface in the absence of
noble metals.
Introduction
Carbon–carbon bond formation reactions are important trans-
formations for processing biomass-derived molecules to pro-
form an anion, which then attacks a second molecule (alde-
hyde or ketone). Water elimination provides the a,b-unsaturat-
ed carbonyl compound. On the other hand, the mechanism of
the ketonization of carboxylic acids was under debate until re-
[
1–4]
duce chemicals and fuels.
Biomass-based platform mole-
cules derived from pentoses and hexoses possess five or six
carbon atoms in most cases, whereas fuel molecules are typi-
cally in the ranges C to C for kerosene and C to C for
[15]
cently. A combination of experimental studies and density
functional theory calculations revealed that this reaction starts
with a double deprotonation of one acid molecule (Scheme 2).
A second molecule is dihydroxylated, and the electrophilicity
at the carbonyl carbon atom increases. Hence, in a similar way
as that for the aldol condensation, the negatively charged a-
carbon atom attacks the carbonyl carbon atom. The formed in-
termediate is then decarboxylated to provide the ketone
(Scheme 2).
9
15
9
20
diesel fuel. Therefore, reactions that involve CꢀC bond forma-
tion are required to produce kerosene and diesel from
[
6,7]
[8,9]
cellulose and hemicellulose. One strategy for the above in-
volves deoxygenation and decarboxylation to yield C –C ole-
4
6
[
7,10]
fins,
which are then oligomerized. However, it is possible to
take advantage of the reactivity introduced by oxygen func-
tionalities to form CꢀC bonds selectively. Hence, other routes
[
6,11,12]
involve aldol condensation reactions
of aldehydes and ke-
(also called ketoni-
The intrinsic reactivity of the molecules and the catalytic
procedures that have been developed make both reactions
valuable tools for synthetic organic chemistry to provide pre-
dictable products with high selectivities. A related reaction
that has received much less attention is the transformation of
two molecules of aldehyde into a ketone. Formally it can be
considered as an oxidation of the aldehyde to the correspond-
ing carboxylic acid [Eq. (1)] and subsequent ketonic decarboxy-
[
13–15]
tones and the ketonic decarboxylation
[
2]
zation) of carboxylic acids.
In the two last routes, it is the carbonyl group that introdu-
ces certain reactivity into the molecule. Carbonyl compounds
are CH-acidic at the a position, and the carbonyl carbon atom
can act as an electrophilic center. This reactivity has been dem-
onstrated to be the origin of the two reactions named above,
that is, aldol condensation and ketonization. The mechanism
of the aldol condensation (Scheme 1) is widely accepted and
starts with the deprotonation of an aldehyde or a ketone to
[
a] L. M. Orozco, Dr. M. Renz, Prof. Dr. A. Corma
Instituto de Tecnologꢀa Quꢀmica
Universitat Politꢁcnica de Valencia-Consejo Superior de Investigaciones
Cientꢀficas (UPV–CSIC)
[16]
lation of the latter (as described above; Scheme 2). However,
the mechanism is still under debate: disproportionations of
Av. de los Naranjos s/n
E-46022 Valencia (Spain)
E-mail: mrenz@itq.upv.es
the Cannizzaro or Tishchenko types may provide the acid (de-
[17–20]
rivative) for subsequent ketonic decarboxylation.
Alterna-
acorma@itq.upv.es
tively, aldol condensation products have been discussed as in-
termediates, which are then dehydrogenated and decarboxy-
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cssc.201600654.
[16,21–25]
lated.
Furthermore, this reaction is not restricted to alde-
ChemSusChem 2016, 9, 1 – 14
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Scheme 1. Reaction mechanism of the aldol condensation.
2
Scheme 2. Reaction mechanism for the ketonic decarboxylation on a ZrO surface, as described in Ref. [5].
hydes as starting materials, and primary alcohols can also be Results and Discussion
[
22,26–28]
converted to symmetrical ketones.
In this case, the alde-
hyde is supposed to be produced in situ and converted to the
ketone.
Catalyst characterization
Two high-surface-area materials, m-ZrO and t-ZrO , with mon-
Manifold materials have been employed in the transforma-
tion of aldehydes and alcohols in recent decades, for example,
CeO , Y/CeO , Pd/CeO , Co/CeO , CeO –MgO , CeO –Fe O ,
2
2
2
ꢀ1
oclinic and tetragonal phase, respectively (103 and 150 m g ,
respectively; Table 1, Entries 1 and 3), and one zirconium oxide,
2
2
2
2
2
2
2
2
3
ZrO , Ce Zr O , Cr/ZrO , Pd/m-ZrO , ZnO , MgO, ThO , La O ,
m-ZrO -A, with a low surface area and monoclinic phase
2
2
x
1ꢀx
2
2
2
2
2
2
3
2
ꢀ1
Cu/La Zr O ,
Fe O ,
Fe O –ZnO,
and
Sn–Ce–Rh
(4 m g ; Table 1, Entry 2) were used in the present work to-
gether with a cerium oxide material with redox properties and
2
2
7
3
4
2
3
[
16,17,20–22,24,25,27–35]
oxide.
Many of these materials involve
2
ꢀ1
a redox-active element such as Fe, Ce, or Cr or a noble metal
a high surface area of 114 m g (similar to m-ZrO ; see
2
2
ꢀ1
for alcohol dehydrogenation. In some cases, such as CeO , the
Table 1). For the high-surface-area (>100 m g ) materials, the
pore diameters were within the range 40 to 90 ꢁ, which corre-
sponds to mesopores (Table 1, Entries 1, 3, and 4). Their meso-
2
consumption of surface oxygen atoms for aldehyde oxidation
[
24,31,33]
has been discussed.
It seems that the redox functionality
is a requirement for a smooth reaction, although
redox-inactive metal oxides have also been reported
[
19]
as catalysts. Zirconium oxide has been reported as
[
19,20,36]
Table 1. Physical and morphological properties of the catalysts.
a catalyst for the ketonization of aldehydes
but
[
a]
[b]
[c]
[e]
has also been claimed to favor aldol condensation
Entry
Catalyst
SBET
Dp
[ꢁ]
Vp
Reducible
sites [wt%]
Acid sites
[mmolg
[
37]
2
ꢀ1
3
ꢀ1
[d]
ꢀ1
over the ketonization of aldehydes.
[m g
]
[cm g
]
]
Herein, we use monoclinic zirconium oxide, m-
1
2
3
4
m-ZrO2
103
87.9
0.24
<1
<1
1
170
(2.1 wt%)
20
0.3 wt%)
180
(2.2 wt%)
20
ZrO , as a bifunctional and nonreducible catalyst for
2
m-ZrO
t- ZrO
CeO
2
-A
4
150
114
202.7
39.3
47.2
0.00
0.14
0.15
the transformation of aldehydes into ketones. Hence,
heptanal was converted into 7-tridecanone with a se-
lectivity of up to 67%, and molecular hydrogen was
observed as a reaction product. Water increases the
selectivity by suppressing the formation of the aldol
condensation product. It is further demonstrated that
water acts as a reaction partner; one proton is used
to form molecular hydrogen, and the hydroxy group
is incorporated into the carboxylic acid. Hence, the
oxidation of aldehydes to carboxylic acids by water is
supported in the present manuscript by kinetic and
isotopic studies.
(
2
2
9.8
(
(
4628C)
2.2
(0.4 wt%)
1
7278C)
5
6
7
ZrO-MCM-41
Zr-MCM-41
670
737
26.2
0.17
0.35
0.11
–
–
–
–
–
–
29.3
–[f]
–
[f]
ZrSiO
4
[a] Specific surface area calculated by the BET method for the N
2
adsorption isotherm
at ꢀ1968C. [b] Pore size. [c] Pore volume. [d] From TPR measurements. [e] From TPD
measurements. [f] Not measurable.
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
3
ꢀ1
pore volumes are also high and vary from 0.14 to 0.24 cm g .
In contrast, the metal oxide with a low surface area, m-ZrO -A,
2
has almost no porosity (Table 1, Entry 2).
Other Zr-containing materials such as ZrO-MCM-41 and Zr-
MCM-41 were synthesized, and ZrSiO was used as a reference
4
for the ones above (Table 1, Entries 5–7). Zr-MCM-41 was ob-
tained by direct synthesis through the addition of the zirconi-
um precursor to the synthesis gel of the structured mesopo-
rous silicate. In contrast, ZrO-MCM-41 was prepared through
the impregnation of the precursor ZrO(NO ) onto the calcined
3
2
MCM-41 material. Both materials have isolated Zr sites coordi-
nated to up to four silyloxy groups or di- or trimetallic active
[38]
centers with Zr atoms connected by oxygen bridges. The
grafted material has all Zr sites on the surface, whereas the Zr
atoms in the direct-synthesis material should substitute Si
atoms of the wall. The mesoporous structures of the materials
were confirmed by X-ray diffraction (Figure S1), and the surface
2
ꢀ1
Figure 2. XRD patterns of m-ZrO -A, m-ZrO , t-ZrO , and CeO .
area was approximately 700 m g in both cases (Table 1, En-
tries 5 and 6). The amount of zirconium incorporated was de-
termined to be 1.6 and 2.8 wt% for ZrO-MCM-41 and Zr-MCM-
2
2
2
2
41, respectively, by inductively coupled plasma optical emis-
of acidic sites is equally as low as that for the low-surface-area
sion spectroscopy (ICP-OES). These isolated Zr centers possess
Lewis acidity, as has been shown by in situ IR spectroscopy of
m-ZrO -A (Table 1, Entries 2 and 4). The X-ray diffraction pat-
2
terns of the oxide materials are given in Figure 2. From the X-
[
38]
adsorbed cyclohexanone.
The acidity of the ZrO and CeO materials was determined
ray diffractograms, it can be concluded that m-ZrO is mono-
2
clinic and t-ZrO is purely tetragonal. However, the contamina-
2
2
2
by temperature-programmed desorption of ammonia (TPD-
tion of m-ZrO with the tetragonal phase cannot be excluded
2
NH ). The corresponding profiles are displayed in Figure 1, and
as all of the corresponding signals overlap, and the broadness
3
2
ꢀ1
those for the high-surface-area ZrO2 materials (>100 m g )
of the bands makes a differentiation impossible. CeO exhibit-
2
are very similar. Ammonia desorption occurs between 250 and
ed a cubic fluorite phase. High-surface-area materials, formed
by small crystallites, give broader peaks than samples with
larger particles and low surface areas, which exhibit sharper
signals.
3008C. Interestingly, dehydration was observed from 2008C
with increasing intensity up to 6008C. The dehydration is more
pronounced for the monoclinic material than for the tetrago-
nal one. For m-ZrO -A, the acidic sites were hardly detected
The temperature-programmed reduction (TPR) profile for
pure CeO2 showed two distinct regions. A low-temperature
region (300–6008C) with a maximum at approximately 4708C
2
(not shown), certainly because of the low surface area. The
NH -TPD profile of CeO shows a maximum between 300 and
3
2
4+
4508C. However, the MS detector indicates that the signal is
is attributed to the surface reduction of the Ce ions, and
[39]
not related to ammonia desorption but predominately to de-
hydration. Therefore, it has to be concluded that the amount
a high-temperature region (700–9008C) with a maximum at
approximately 7508C is probably related to the reduction of
[31]
the bulk oxide (Figure 3). The reducibility of ZrO is more
2
than one order of magnitude lower than that of CeO . This is
2
in good accordance with previous reports in which ZrO was
2
[14,31]
considered to be a nonreducible material.
In particular, m-
ZrO did not show any peaks in its TPR profile (Figure 3), and
2
the concentration of the reducible sites was less than 1 wt%
(
Table 1, Entry 1). The residual reducibility may result from the
remaining organic parts of the metal precursors.
The IR spectroscopic measurements of the thermally treated
high-surface-area zirconium oxides indicate that the surfaces
are highly hydrated at room temperature (Figure 4). Both
phases, tetragonal and monoclinic, were heated at 3008C
ꢀ
2
ꢀ3
under vacuum (10 to 10 Pa), and water desorption and two
different types of surface hydroxy groups became evident. Fur-
thermore, it can be concluded that Lewis acidic surface sites,
which were occupied by water before its desorption, are made
ꢀ
1
accessible. In the carbonyl region ( n˜ =1600 to 1400 cm ),
traces of organic residues from the metal precursors can be
detected.
3 2 2 2
Figure 1. Profiles of the TPD-NH of m-ZrO , t-ZrO , and CeO .
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
2 2 2 2
Figure 3. TPR profiles of CeO , m-ZrO , m-ZrO -A, and t-ZrO .
Figure 5. Conversion and product yields for the ketonization of heptanal
2
over m-ZrO in (a) the absence and (b) presence of water as a function of re-
action temperature. Reaction conditions: heptanal/water molar ratio 1:8
ꢀ1
ꢀ1
ꢀ1
(
5 mL of heptanal, 0.2 mLmin ), 0.204 mLmin water, 50 mLmin
2
N , and
catalyst (1 g).
yields of 67 and 37% in the presence and absence of water, re-
spectively. Interestingly, in the absence of water, the aldol con-
densation product was observed in approximately 23% yield,
whereas it was observed only in traces in the presence of
water. From this first survey, 4508C was selected as the stan-
dard reaction temperature for the study of the catalytic activi-
ties.
Figure 4. IR spectra of self-supporting wafers of t-ZrO
gassed at room temperature, 300, and 4008C.
2 2
and m-ZrO out-
Ketonization of heptanal over m-ZrO and related materials
2
Heptanal was chosen as a model substrate as the size of the
molecule facilitates the condensation of the products if the re-
action is performed in a fixed-bed continuous-flow reactor.
Heptanal is available as a biomass-derived platform molecule
through the thermal treatment of methyl ricinoleate or castor
The above catalytic results were achieved with m-ZrO ,
2
2
ꢀ1
which has a high surface area (103 m g , Table 1, Entry 1) and
a monoclinic structure. A material with the same crystalline
2
ꢀ1
phase but with a very low surface area (<5 m g , cf. Table 1,
[
40,41]
[19,31]
oil.
sence
Ketonizations of aldehydes in the presence
and ab-
Entry 2) was employed, that is, m-ZrO -A, and the conversion
2
[
20,24,27]
of water have been reported, and a beneficial
dropped to 15% in the absence of water and below 10% if
water was added to the feed (Table 2, Entry 2). If the initial re-
action rates are normalized to the catalyst surface areas, the
values obtained are much closer; therefore, the active sites
were the same, and the surface area was the main factor that
explains the activity differences.
effect of the water on the ketone selectivity has been ob-
served. Therefore, both conditions were assessed in our study.
Both phases of ZrO that are stable at low temperature, that
2
is, m-ZrO and t-ZrO , were considered in the present study.
2
2
The results obtained were also compared with those for
cerium oxide, which is a well-known catalyst for the ketoniza-
If the Zr sites were incorporated in silica MCM-41 either by
postsynthetic modification (ZrO-MCM-41) or by a direct-
synthesis procedure (Zr-MCM-41), the ketonization reaction did
not occur (Table 2, Entries 3 and 4). The aldehyde conversion
was rather low, and the main products were condensation
products. These results indicate that various neighboring sites
[
31,42]
tion of aldehydes.
Before the catalyst screening, the reaction temperature was
selected with m-ZrO as the catalyst (Figure 5). The reaction
2
temperature had to be increased to 4008C to enable a conver-
sion of 20 to 30%, in the presence or absence of water, and
a further increase to 4508C provided almost full conversion.
The corresponding ketone, 7-tridecanone, was obtained in
with well-defined geometrical arrangements on the ZrO sur-
2
face should be involved in the ketonization. On the other
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[
a]
Table 2. Heptanal conversion and principal product yields over different catalysts for the ketonization of heptanal.
[
b]
Entry
Catalyst
Conv.
%]
Yields [%]
13 ketone Aldol cond. prod.
[c]
[d]
[e]
[f]
[g]
[
C
Aldol cond. isomer
Ketone fragments
Others
2H
2
CO
2
CO
1
2
3
4
5
6
7
m-ZrO
m-ZrO
2
2
90 (95) 67 (37)
8 (15) 1 (1)
1 (23)
1 (1)
1 (1)
0 (1)
10 (2)
(0)
5 (15)
2 (3)
3 (2)
1 (1)
8 (5)
(1)
13 (4)
0 (1)
8 (2)
0 (1)
4 (14)
(1)
4 (16)
4 (9)
54 (48)
4 (20)
27 (73)
(4)
69 (42) 65 (33) 2 (5)
-A
2 (2)
0 (0)
1 (1)
4 (4)
1 (3)
2 (4)
0 (1)
4 (37)
2 (4)
ZrO-MCM-41 68 (58) 2 (5)
Zr-MCM-41 6 (27) 1 (4)
t-ZrO 65 (99) 16 (5)
ZrSiO (6) (0)
CeO 87 (96) 66 (54)
2
16 (16) 23 (20) 9 (5)
(0) (2) (0)
68 (45) 54 (29) 3 (14)
4
2
1 (3)
1 (13)
12 (10)
7 (16)
-
1
[
a] Reaction conditions: the catalyst (1.0 g) was placed in a fixed bed, and heptanal (5 mL, 0.2 mLmin ) was reacted in the gas phase in a continuous-flow
-
1
ꢀ1
reactor at 4508C under a nitrogen flow (50 mLmin ), the heptanal/water molar ratio of the feed was 1:8 (0.204 mLmin water). The numbers in parenthe-
ses indicate conversion and yield without water cofeed. [b] Yield calculated with respect to the aldehyde substrate with consideration of the reaction stoi-
chiometry: two aldehyde molecules form one ketone molecule or one molecule of the other products. [c] Yield for the aldol condensation product.
8
[d] Yield for isomerized dimers with the same mass as the aldol condensation product but different GC retention times. [e] C –C11 ketone fragmentation
products; probably derived from the ketone product. [f] Other products observed in the liquid phase and quantified by GC. For more information on this
group, see Table S1. [g] Yield of two equivalents of molecular hydrogen per two equivalents of aldehyde substrate.
hand, if the crystallographic phase of ZrO was changed from
Reaction stoichiometry
2
monoclinic to tetragonal with the high surface area maintained
2
ꢀ1
(
150 m g , Table 1, Entry 3), the conversion remained com-
The gaseous products produced during the ketonization reac-
tion, that is, H and CO , were monitored and quantified. For
plete under our reaction conditions (Table 2, Entry 5), but the
yield of the desired ketone, 7-tridecanone, dropped from 37
2
2
m-ZrO and CeO , two equivalents of H and one equivalent of
2
2
2
(entry 1) to 5% (entry 5) without a water cofeed and from 67
CO were obtained per equivalent of ketone, regardless of the
2
to 16% in the presence of water. The origin of the different
catalytic activities is probably the different coordination
spheres of the Zr centers in the monoclinic and tetragonal
phases: in the first case, the Zr atoms are surrounded by seven
oxygen atoms, and those in the second case are surrounded
cofeeding of water or not (Table 2, Entries 1 and 7).
For m-ZrO , the reaction was run for a prolonged time on
2
stream until almost 100 g of substrate was processed per gram
of catalyst (Figure 6). With this amount of substrate passed,
a steady state was achieved. Therefore, under these conditions,
[
43]
by eight. From this comparison, it can be seen that several
sites are involved in the ketonization and also that the geo-
metrical arrangement is important. It can then be stated that
the crystalline phase is decisive for the catalytic activity of ZrO2
and that the monoclinic phase is the most active one.
A comparison with a zirconium silicate indicates again that
the reactions observed are not attributable to the simple pres-
ence of Zr sites nor to thermal transformations. The conversion
obtained with this material was as low as 6% (Table 2, Entry 6).
No ketone was observed, and only traces of the condensation
products were detected.
the ZrO can be considered as a stable catalyst that does not
2
intervene in the stoichiometry.
Furthermore, the results in Figure 6 indicate that the catalyt-
ic activity is stable with time on stream. During this time on
stream, the yield of the three main products, that is, 7-trideca-
none, H , and CO , remained constant, and the stoichiometry
2
2
corresponds to the conversion of two molecules of aldehyde
into one molecule of ketone, one molecule of CO , and two
2
molecules of H [Eq. (2)].
2
The results obtained with m-ZrO were also compared with
2
those for CeO , which is a classical catalyst for the ketonization
2
of aldehydes, as mentioned before. If the reaction was per-
formed with a water cofeed, the yield of the ketone was practi-
cally the same for the two catalysts (Table 2, Entries 1 and 7).
However, without a water cofeed, the yield with CeO2 was
5
4% versus 37% with m-ZrO . This comparison demonstrated
2
that m-ZrO is a suitable catalyst for the ketonization of alde-
2
hydes and that the catalytic performance with a water cofeed
is at the same level as that of the best catalyst reported to
date. The fact that the reaction catalyzed by CeO is less influ-
2
enced by water than that catalyzed by m-ZrO2 will be ex-
2
Figure 6. Catalytic performance of m-ZrO with time on stream in the ketoni-
plained later.
zation of heptanal. Yield calculated per two equivalents of aldehyde; the sto-
ichiometry is shown in Eq. 2. Reaction conditions: heptanal/water molar
ꢀ1
ꢀ1
ꢀ1
ratio 1:8, 0.2 mLmin heptanal, 0.204 mLmin water, 50 mLmin
2
N ,
4508C. During the reaction 104 g of heptanal were passed per g of catalyst.
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
The atomic balance of Equation (2) can be achieved by
adding one water molecule as a substrate. Then, if water is re-
quired as a reactant for the process to occur, the question is
how can the reaction proceed in the absence of water? The re-
sults in Table 2 indicate that water is formed during reactions
such as aldol condensations. Consequently, even if water is not
added to the feed, some water will always be present during
the reaction. Therefore, there are two issues that require fur-
ther investigation, namely, the role of water in the reaction
Figure 8. Yields of ketone at different contact times for the ketonization of
4
heptanoic acid at a partial pressure of 10 Pa with m-ZrO
2
in the absence
and presence of water. Reaction conditions: heptanoic acid (5 mL, 0.041,
ꢀ
1
ꢀ1
0.064, 0.128, and 0.213 mLmin ), N
2
(71, 111, 222, and 365 mLmin ), hepta-
noic acid/water molar ratio 1:8, water (0.041, 0.065, 0.130, and
ꢀ
1
ꢀ1
0
2
.216 mLmin ), N (15, 23, 46, and 77 mLmin ), catalyst (15 mg), 4508C.
and the formation of 2 mol of H per mol of ketone formed.
2
Water also influences the rate of accompanying reactions
such as the aldol condensation between two aldehyde mole-
cules (Scheme 3). In the absence of water, aldol condensation
was observed as a competitive reaction pathway that is de-
creased considerably in the presence of water. Aldol condensa-
tion is an equilibrium reaction in which water is coproduced
Role of water on product selectivity
Water improves the selectivity towards the ketone product
and is consumed for the formation of ketone. Therefore, one
could expect that water would enhance the reaction rate for
the formation of the ketone. To check this hypothesis, kinetic
measurements were performed in the absence of water and in
the presence of increasing amounts of water. Surprisingly, no
beneficial effect on the reaction kinetics was detected
(
Scheme 3). Therefore, the addition of water might shift the
aldol equilibrium towards the substrate, that is, the aldehyde,
which can follow the desired ketonization pathway. As the ke-
tonization reaction is irreversible, the product selectivity is
shifted successively towards the ketone.
(Figure 7), but the ketone formation was fastest in the absence
To check this hypothesis, the aldol condensation product
of water. Interestingly, a similar negative effect of water on the
reaction rate was observed for the ketonic decarboxylation of
carboxylic acid (Figure 8). From these results, as a first approxi-
(
99% purity) and water were submitted to the reaction condi-
tions. To avoid the masking of the results by the direct trans-
formation of the aldol condensation product into the ketoniza-
mation, we could assume that H O competes with the alde-
2
[6,11,12]
tion product,
the reaction was performed in the presence
hyde for adsorption to surface sites, which would decrease the
rate of the reaction, or retards the reaction rate at a later
stage.
of an aldehyde with a different chain length, that is, hexanal.
Then, if the aldol condensation is reversible under the em-
ployed reaction conditions, the cross-coupling ketonization
Figure 7. Initial rates for the ketonization of heptanal at a partial pressure of
4
1
1
0
1
6
1
.5ꢂ10 Pa with m-ZrO
2
in the presence of different amounts of water (0,
4
.5, 4.5, and 7.5ꢂ10 Pa). Reaction conditions: heptanal (5 mL,
.147 mLmin ), heptanal/water molar ratio 1:1 (0.019 mLmin water,
18 mLmin
7 mLmin
6 mLmin
ꢀ
1
ꢀ1
ꢀ
1
ꢀ1
N
2
, 6, 15, 23 mg of catalyst), 1:3 (0.056 mLmin water,
Scheme 3. Two alternative reaction pathways for heptanal: aldol condensa-
tion and ketonization. As the aldol condensation is an equilibrium reaction,
the presence of additional water may shift this reaction towards the alde-
hyde, which can then react through ketonization. Hence, water improves
the yield of the ketone.
ꢀ
1
ꢀ1
N
N
2
, 8, 15, 23 of mg catalyst), and 1:5 (0.094 mLmin water,
ꢀ
1
2
, 15, 30, 50 mg of catalyst), 4508C; without water, the reac-
ꢀ1
tion was performed with 5, 16, and 22 mg of catalyst and 144 mLmin
For details, see the Kinetics study section.
2
N .
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 4. Product distribution obtained for the cross-coupling of hexanal with the aldol condensation product of heptanal or heptanal. Cond. I: hexanal/2-
pentyl-2-nonenal/water molar ratio 4:1:8; conversion hexanal 79.5%; conversion aldol condensation product 48.7%. Cond. II: hexanal/heptanal/water molar
ratio 8:1:9 (for this experiment, the 50% conversion of the aldol condensation product was considered, and the initial hexanal/heptanal ratio was changed
2
from 4:1 to 8:1); conversion hexanal 73.3%; conversion heptanal 69.8%. Reaction conditions: m-ZrO (1 g), 4508C. Cond. I: reaction mixture (5 mL,
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
0
.167 mLmin ), 0.041 mLmin water, 185 mLmin
N
2
. Cond. II: reaction mixture (5 mL, 0.140 mLmin ), 0.021 mLmin water, 213 mLmin
2
N .
product of hexanal and heptanal, that is, 6-dodecanone,
should be observed. This was indeed the case, and 6.5% of
6-dodecane was observed, whereas 7-tridecanone only formed
in 2.3%. Interestingly, the amounts for all three ketoniza-
tion products were very similar to those obtained for the
reaction of a comparable mixture of hexanal and heptanal
(cf. cond. II, Scheme 4), that is, 7.8% of 6-dodecane and 2.2%
of 7-tridecanone. This result clearly indicates that the role of
water is to suppress byproducts by shifting the equilibria of
side reactions towards the substrate. Then, as CeO gives less
2
aldol condensation than m-ZrO , a smaller effect of the added
2
water can be expected for the former (cf. above). The cross-
coupling experiment also proved unequivocally that the aldol
condensation product was not an intermediate towards the
ketonization product, as the yield of 7-tridecanone would have
been higher than the yield of 6-dodecanone if this were the
case.
Figure 9. Yield versus conversion of 7-tridecanone, 2-pentyl-2-nonenal (aldol
condensation product), and heptanoic acid. Reaction conditions: heptanal/
ꢀ
1
4
water ratio 1:5, heptanal (5 mL, 0.147 mLmin , 10 Pa partial pressure),
ꢀ
1
ꢀ1
0
4
.094 mLmin water, 16 mLmin
508C.
2
N , m-ZrO2 (0.1, 0.2, 0.4, 0.6, 0.8, and 1 g),
stable product. At conversions below 5%, the yield of hepta-
noic acid is even higher than the yields of 7-tridecanone and
the aldol condensation product.
Mechanism of the ketonization reaction
The reaction mechanism for the ketonization of aldehydes is
[
21]
still under debate. In our case, we suggested above that the
aldol condensation product can be excluded as an intermedi-
ate for the formation of the ketone. Now, a possible reaction
pathway that involves the oxidation of the aldehyde followed
by the ketonic decarboxylation of the corresponding carboxylic
acid derivate will be considered.
Nevertheless, the detection of the carboxylic acid as a pri-
mary unstable product is not definitive proof for its participa-
tion in the ketonization of aldehydes. To demonstrate this par-
ticipation unequivocally, the cross-coupling of heptanal was
performed in the presence of the same molar amount of hexa-
noic acid under ketonization conditions (Scheme 5). If the keto-
nization of aldehydes does not proceed through the formation
of the corresponding acid, one could expect the hexanoic acid
to react through the ketonic decarboxylation to 6-undecanone,
whereas heptanal should be converted into 7-tridecanone as
the main product by aldehyde ketonization. In other words,
Along this line, an experimental observation was the pres-
ence of heptanoic acid in the product mixture in low yield (1–
2
%) for the reaction in the presence of water (Figure 9). Fur-
thermore, if the yield is plotted versus conversion, the shape
of the curve indicates that heptanoic acid is a primary and un-
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
It is remarkable that water acts as terminal oxidant and hy-
drogen is produced. Clearly, the driving force for this oxidation
is not the oxidation power of water but the high enthalpy of
[44]
formation of the aldehyde.
The latter is also the driving
Scheme 5. Product distribution obtained for the cross-coupling of aldehyde
with a carboxylic acid with a different number of carbon atoms compared
with the cross-coupling of two carboxylic acids with the corresponding
chain lengths. Reaction conditions: reaction mixture 1: heptanal (R=H) and
hexanoic acid (R=OH), molar ratio 1:1; reaction mixture 2: heptanoic acid
force for the disproportionation reactions of aldehydes, that is,
the Cannizzaro and Tishchenko reactions. Water is a mild oxi-
dant and is not active at low temperatures (below steam-re-
forming temperatures). However, thermodynamic calculations
have shown that the transformation depicted in Scheme 6
ꢀ
1
and hexanoic acid, molar ratio 1:1; reaction mixtures (5 mL, 0.2 mLmin ),
ꢀ
1
50 mLmin
2
N , 1 g m-ZrO2, 4508C. Water was not fed. Conversion>99%.
[24]
(
balanced equation) is favorable for propanal.
Having confirmed the participation of the corresponding
carboxylic acid during the ketonization of an aldehyde by
direct detection and a cross-coupling experiment, we studied
the mechanism of the oxidation at a molecular level.
both products will be generated by self-condensation reac-
tions. However, if the carboxylic acid is formed as an inter-
mediate for the ketonization reaction from the aldehyde, an
additional product should be observed, that is, the cross-
coupling product of heptanoic and hexanoic acids, 6-dodeca-
none. The results obtained show the formation of 6-dodeca-
none in 16% yield (Scheme 5). Nevertheless, the 16% yield
was lower than the yield obtained (48%) if both carboxylic
acids, instead of acid and aldehyde, were reacted (Scheme 5).
This result could be explained by considering that the ketonic
decarboxylation is faster than the oxidation of the aldehyde to
the carboxylic acid under the present reaction conditions (cf.
Kinetic study). Hence, the initial carboxylic acid is consumed
rapidly, and a low amount remains for the cross-coupling reac-
tion at a later stage of the reaction.
The formation of a carboxylic acid (derivative) during ketoni-
[
20,23,25,31]
zation by disproportionation, that is, by Cannizzaro
or
[17,18,21,45]
Tishchenko
reactions, has been discussed (Scheme 6)
previously. There, the alcohol could be dehydrogenated to
form one of the two aldehydes and to produce the H indicat-
2
ed in Equation (3). Hence, the global reaction equation (cf.
Scheme 6) has exactly the same stoichiometry as Equation (3).
In any case, the important result from the experiment was
the appearance of the cross-coupling product, which can only
be formed from the two corresponding carboxylic acids, that
is, hexanoic and heptanoic acid. Therefore, it can be concluded
that heptanal is oxidized to heptanoic acid under the ketoniza-
tion conditions and that the carboxylic acid is a reaction inter-
mediate.
The formation of the carboxylic acid intermediate is also in
accordance with the empirical reaction equation obtained
from the long-term experiment [Eq. (2)]. Hence, two aldehyde
molecules are oxidized with two molecules of water to afford
Scheme 6. The oxidation of an aldehyde into a carboxylic acid by the hypo-
thetic reaction pathway involving the disproportionation of the aldehyde
through a Tishchenko or Cannizzaro reaction with the subsequent dehydro-
genation of the produced alcohol. Two molecules of aldehyde are con-
sumed, and one of these is recovered so that one molecule of aldehyde is
oxidized to one carboxylic acid molecule and one equivalent of hydrogen is
produced. Additionally, one water molecule is consumed. The balanced reac-
tion equation coincides with Equation (3).
carboxylic acid and H [Eq. (3)]. In the next step, the two mole-
2
cules of carboxylic acid can react through the classical ketonic
decarboxylation reaction to produce the ketone, carbon diox-
ide, and water [Eq. (4)]. If both reactions are combined, the ob-
tained Equation (5) corresponds exactly with the molecular sto-
ichiometry found experimentally.
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Two reactions were designed to assess whether or not a dis-
proportionation of the aldehyde occurs during its ketonization.
First, the ester itself was passed through the catalytic fixed-bed
reactor, and the product mixture was compared to the compo-
sition obtained during the ketonization of the aldehyde. From
the ester, 7-tridecanone was observed in 31% yield (Scheme 7)
2
Scheme 8. Product distribution obtained for heptanol with m-ZrO under
the reaction conditions for the ketonization of heptanal. Reaction condi-
ꢀ
1
4
tions: heptanol (5 mL, 0.2 mLmin , 4.1ꢂ10 Pa partial pressure),
ꢀ
1
50 mLmin
2
N , catalyst (1 g), 4508C.
Hence, this demonstrates that the catalyst is active for the de-
hydrogenation of alcohols.
In summary, we can say that the disproportionation of the
aldehyde during the ketonization can be excluded. Side prod-
ucts that could indicate the presence of the alcohol during the
reaction, for example, the olefin formed by dehydration or the
alcohol itself, were not observed. Therefore, the direct dehy-
drogenation of the aldehyde on the ZrO was considered as
2
a real possibility.
Although ZrO is regarded as a nonreducible catalyst, the
2
hydrogenation of benzoic acid to benzaldehyde has been re-
Scheme 7. Product distribution obtained for heptyl heptanoate with m-ZrO
2
ported to occur, and the coordination of hydride species to
under the reaction conditions for the ketonization of aldehydes. Reaction
[19,46]
the Zr centers has been proposed by Yokoyama.
Interest-
ꢀ1
4
conditions: heptyl heptanoate (5 mL, 0.2 mLmin , 2.7ꢂ10 Pa partial pres-
ꢀ1
ingly, the hydrogenation of benzoic acid is exactly the reverse
reaction of Equation (3). Hence, from these results, it can be
concluded that it is not unreasonable to propose the direct de-
sure), 50 mLmin
2
N , catalyst (1 g), 4508C.
hydrogenation of an aldehyde on a ZrO surface.
2
together with CO . However, 1-heptanol and 1-heptene were
To further explore a possible direct dehydrogenation step,
the deuterated aldehyde [see Eq. (6)] was employed as a sub-
strate. In this case, HD was expected to form as a product
from the deuterium atom from the aldehyde and a proton
from water or from the surface. Experimentally, the ketoniza-
tion proceeded smoothly, and the ketone product was ob-
tained together with a mixture of HD and H2 [Eq. (6)]. The
former was the predicted product, whereas the later was
2
also observed in 13 and 12% yield, respectively (Scheme 7), to-
gether with a considerable amount of coke deposited on the
catalyst. As 1-heptanol and 1-heptene were not obtained from
the aldehyde as substrate, the ester can be excluded as an in-
termediate in the ketonization of aldehydes. The dehydrogena-
tion of the alcohol to the aldehyde is probably rather slow,
that is, it is not facilitated effectively by the catalyst and, conse-
quently, this compound accumulates in the reaction medium
or is consumed in side reactions such as dehydration to give
the corresponding olefin. Consequently, both the alcohol and
the olefin are observed.
[19,46]
rather unexpected. However, Yokoyama et al.
indicated
that molecular hydrogen (and therefore HD, once formed and
desorbed) can be cleaved heterolytically and adsorbed as
proton and hydride species. Furthermore, the deuteration of
the surface of ZrO by D has been observed by FTIR spectros-
In an additional experiment, we performed the reaction with
the alcohol (1-heptanol) by feeding it separately to the reactor.
The product mixture contained two main products: the olefin
2
2
[47]
copy.
Consequently, H/D exchange with surface hydroxy
groups is possible for D ; therefore, H/D exchange between
2
1
-heptene (36%) and the ketonization product 7-tridecanone
HD and water could also be expected in the presence of ZrO2.
This is exactly what was observed when deuterated water was
passed together with a hydrogen flow: H , D , and HD were
(23%, Scheme 8) at a conversion of 78%. The aldehyde was
only observed in traces. The composition of the mixture con-
firmed that the alcohol is dehydrated to the terminal olefin
under the reaction conditions. As the latter was not observed
during the ketonization of aldehydes, the alcohol can also be
excluded as a reaction intermediate. On the other hand, it can
be concluded that the alcohol is dehydrogenated and that the
aldehyde is converted rapidly to the ketonization product.
2
2
formed [Eq. (7)]. This result demonstrates unequivocally that
surface hydride species are formed under the reaction condi-
tions of the ketonization of aldehydes and this, in turn, ex-
plains how a direct dehydrogenation of an aldehyde molecule
adsorbed onto a ZrO surface can occur, namely, by the expul-
2
sion of a hydride species, which is transferred to the surface. A
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
hydride shift directly from the trigonal planar carbon atom of
the aldehyde is unlikely and should instead occur from a tetra-
hedral carbon atom. In a tetrahedral configuration, the alde-
hyde hydrate can be considered as an analog of a secondary
alcohol, and the mechanism of the hydride shift is analogous
ylate is already prepared to react with a second molecule in
the ketonic decarboxylation.
Alternatively, the carboxylic acid can be desorbed, and the
surface hydroxy group can be restored. Therefore, a water mol-
ecule has to be adsorbed and dissociated to provide the origi-
nal surface hydroxy group and the proton required for the pro-
tonation of the carboxylate. It has been proposed previously
that water can replenish surface oxygen atoms on a cerium
to that for the Meerwein–Ponndorf–Verley reaction. ZrO has
2
been reported as an active catalyst for this reaction in several
[
36,48,49]
cases.
The proper configuration can be achieved
[35]
through the reaction of the aldehyde with a surface hydroxy
group. The presence of surface hydroxy groups has been de-
tected by IR spectroscopy (Figure 4), even after treatment of
the oxide at 4008C in high vacuum. These hydroxy groups are
probably related to defect sites on the surface.
oxide surface in the ketonization of aldehydes. A certain hy-
dration and dehydration ability of the ZrO material was de-
2
tected in the TPD between 200 and 6008C (Figure 1). With the
displacement of the carboxylic acid by a water molecule, the
catalytic cycle for the aldehyde oxidation by dehydrogenation
is closed (Scheme 9), and the reaction equation with all reac-
tants and products involved in the mechanistic cycle has the
same stoichiometry as Equation (3), which is in accordance
with empirical Equation (2). Therefore, it can be concluded that
this mechanistic cycle is supported by the empirical equation
for the long-term experiment.
The reaction of water, adsorbed onto the surface and disso-
ciated, with the aldehyde instead of the surface hydroxyl
group is less likely, as the addition of water to the feed should
accelerate the reaction rate in this case. However, rate retarda-
tion was observed instead (Figure 7).
From all of the above, the following mechanism for the de-
hydrogenation of heptanal can be proposed (Scheme 9): The
aldehyde is adsorbed onto a surface hydroxy group by the nu-
cleophilic attack of this group onto the carbon atom to form
the required tetrahedral geometry of the carbon atom in the
aldehyde. Then, the hydride is transferred to a suitable, Lewis
acidic Zr atom. The fusion of two different hydrogen species,
that is, one coordinated to a Lewis acid center and the other
In summary, it can be concluded that the experimental re-
sults support the ketonization of aldehydes by dehydrogena-
tion of the aldehyde to the carboxylic acid on the ZrO surface.
2
In this reaction, surface hydride species are involved. However,
this proposed mechanism should be further confirmed by
spectroscopic means, theoretical DFT calculations, and kinetic
studies.
to a basic oxygen atom, will release H . The remaining carbox-
2
Kinetic study
It has been shown above that the ketonization of aldehydes
consists of two separate reactions, that is, the dehydrogena-
tion of the aldehydes to carboxylic acids, and the subsequent
ketonic decarboxylation of these acids to the ketone product.
If the reaction rates for the formation of the ketone from the
aldehyde and from the carboxylic acid are compared (Fig-
ure S2), it can be seen that ketone formation from the carbox-
ylic acid is straightforward and 2.5 times faster than that from
the aldehyde. From these results, it can be concluded that the
carboxylic acid consumption is faster than its formation from
the aldehyde. Accordingly, it was observed that the corre-
sponding carboxylic acid was a primary unstable product in
the ketonization of aldehydes (Figure 9). Furthermore, in the
cross-coupling experiment with an aldehyde together with
a carboxylic acid (Scheme 5), different product ratios were ob-
served for symmetrical and cross-coupling products than those
for the reaction of two carboxylic acids. This phenomenon is
well in line with the explanation that the carboxylic acid is
more reactive than the aldehyde; the former is consumed rap-
idly through self-condensation in the presence of an aldehyde,
and a smaller amount remains available for cross-coupling.
If the slow first reaction step, that is, the dehydrogenation, is
divided into its single steps [Eqs. (S1) to (S3), Supporting Infor-
Scheme 9. Proposed mechanism for the dehydrogenation of heptanal
(R=C
6
H13) to heptanoic acid. A hydride species is transferred from the alde-
hyde to the surface, and H
2
is formed through the recombination of the sur-
face hydride species with a surface proton. Water acts as a hydrogen and
oxygen source.
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
mation], two of the single steps are quite fast equilibrium pro-
cesses, that is, the adsorption of the aldehyde and the desorp-
tion of the reaction product, as is water adsorption and disso-
ciation. Even the recombination of the surface hydride species
and a surface proton is an equilibrium, as was reported above.
Therefore, it can be concluded that the rate-determining step
is the hydride transfer from the adsorbed aldehyde to the cata-
lyst surface [Eq. (S2)], and the rate of the latter depends on the
concentration of such surface species. Water has been consid-
ered as a competitor for aldehyde adsorption to the active
site, that is, the surface hydroxy group (Scheme 10).
Figure 10. Comparison of the experimental and calculated initial reaction
rates in heptanal ketonization catalyzed with m-ZrO
2
in the presence of
4
water (1.5ꢂ10 Pa) as a function of heptanal partial pressure (1.5, 3, 4.5, and
4
6
ꢂ10 Pa). The calculated graph was obtained with the Excel Solver pro-
ꢀ
1
gram. Reaction conditions: heptanal (5 mL), water (0.022 mLmin ), heptanal
4
ꢀ1
partial pressure 1.5ꢂ10 Pa (0.636 mL water, 0.169 mLmin heptanal, 6, 15,
ꢀ
1
4
2
0
4
3 mg of catalyst, 136 mLmin
2
N ), 3ꢂ10 Pa (0.319 mL water,
ꢀ
1
ꢀ1
.350 mLmin heptanal, 12, 25, 40 mg of catalyst, 110 mLmin
.5ꢂ10 Pa (0.213 mL water, 0.518 mLmin heptanal, 15, 25, 40 mg of cata-
N ), and 6ꢂ10 Pa (0.159 mL water, 0.69 mLmin heptanal,
2
0, 30, 45 mg of catalyst, 50 mLmin
N
2
),
Scheme 10. Proposed competitive adsorption of water to the active m-ZrO
2
4
ꢀ1
surface hydroxy sites.
ꢀ
1
4
ꢀ1
lyst, 80 mLmin
2
ꢀ
1
2
N ).
Then, as the number of surface hydroxy groups is constant
on the catalyst (it is assumed that surface hydroxy group resist-
ant to desorption at 4508C are related to surface defects), the
reaction rate has to be proportional to the concentration of al-
dehyde in the gas phase. As a final rate equation, Equation (8)
has been deduced (for more detail, see Supporting Informa-
tion):
tion product at an advanced stage of the reaction. Further-
more, the effect of water on the reaction kinetics of the ketoni-
zation of carboxylic acids, that is, the second reaction of the
ketonization of aldehydes, was clearly rate-retarding (Figure 8).
We have also proposed that the water competes with the alde-
hyde for adsorption to the surface hydroxy groups. A clear
rate-retarding effect of water on the aldehyde ketonization has
been measured (Figure 7), but there are probably at least three
different ways by which the water concentration affects the
rate equation, and this issue will not be treated here.
k½RCHOꢁ
ꢀrRCHO ¼ 1
ð8Þ
þ K ½RCHOꢁ þ K ½H Oꢁ
1
4
2
The rate equation was evaluated by the initial-rates method.
The initial rates of ketone formation from heptanal were mea-
sured at low conversion (<12%) as the reactant concentration,
that is, its partial pressure in the feed, was varied and the
The activation energy was determined in the temperature
range 410 to 4508C for m-ZrO . The logarithm of the ketone
2
formation rate versus the inverse reaction temperature is
4
ꢀ1
water partial pressure was maintained at 1.5ꢂ10 Pa
shown in Figure S5. From these data, a value of 193 kJmol
(
Figure 10). At the same time, the rate equation was also evalu-
was calculated for the activation energy. This value is consider-
ably higher than that for the ketonic decarboxylation of car-
ated as the reactant concentration was maintained at 1.5ꢂ
4
ꢀ1
1
0 Pa and the water partial pressure was varied in the feed
boxylic acids. For instance, a value of 116 kJmol was report-
[50]
(Figure 7). The Solver tool of the Excel software was used to
ed for the ketonic decarboxylation of pentanoic acid. Both
values are in line with the different reaction-controlling steps
for these reactions, as discussed before.
obtain a good fit of Equation (8) to the experimental data (cf.
Figures 7 and 10). In addition, in a different presentation, that
is, the inverse of the ketone formation rate over the inverse of
the aldehyde concentration, a straight line could be fitted to
all experimental points for the variation of the aldehyde con-
centration (Figure S3) or the water concentration (Figure S4).
Both graphical solutions support the deduced equation and,
therefore, the hypothesis that the dehydrogenation of the al-
dehyde coordinated to a surface hydroxy group is the rate-de-
termining step.
Conclusions
Monoclinic zirconium oxide (m-ZrO ) with a high surface area
2
2
ꢀ1
(>100 m g ) is an active catalyst for the transformation of al-
dehydes (heptanal) into symmetrical ketones (7-tridecanone).
Although ZrO is a nonreducible metal oxide, the material is
2
The role of the water on the kinetics of the reaction is con-
sidered quantitatively only at low conversion in the present
paper. As already mentioned, the role of the water is complex,
as it influences the overall reaction in at least in two or three
different ways. Water affects the aldol condensation equilibri-
um and provides additional aldehyde from the aldol condensa-
suitable for this transformation, and a yield of 66% was ach-
ieved.
The reaction mass balance shows that water is consumed
and hydrogen is formed. Moreover, carbon dioxide is formed
during the transformation of the aldehyde into the ketone. In
this reaction, no sacrificial oxidant was necessary, as water
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
plays this role by providing hydrogen and oxygen atoms at
the same time.
vacuum (0.1 to 1 Pa) at room temperature. The IR spectra were re-
corded at room temperature and also after the treatment of the
samples at 300 and 4008C under vacuum.
Water has a beneficial effect on product selectivity for m-
ZrO as the catalyst. This has been explained by a shift in the
2
composition for the aldol condensation equilibrium. Thus, the
aldol condensation byproduct is minimized, and the liberated
aldehyde can react towards the desired product.
Catalytic tests in a fixed-bed, continuous-flow reactor
The transformations of heptanal and alternative reactions (hexanal
aldol condensation product, hexanal–heptanal mixture, heptanal–
heptanoic acid mixture, heptanoic–hexanoic acid mixture, heptyl
heptanoate and heptanol) were performed in a tubular stainless-
steel reactor. The catalyst (1.0 g, pellets 0.4–0.8 mm) was diluted
with silicon carbide (2.0 g), placed as a fixed bed in a stainless-steel
tube (0.4 cm internal diameter and 20 cm length), and calcined at
Concerning the reaction mechanism, the carboxylic acid has
been identified as a reaction intermediate. It is proposed that,
in a key-step, the aldehyde is adsorbed onto the metal oxide
surface and a hydride species is transferred to the surface.
These findings are interesting and may help with catalyst im-
provement, for instance, through the incorporation of metal
sites that can facilitate this hydride transfer.
ꢀ1
4
508C for 2 h in air (50 mLmin ). The reactor was heated to differ-
ent reaction temperatures between 300 to 4508C. Heptanal
ꢀ
1
(
0.2 mLmin ) and water were fed separately into the reactor with
ꢀ1
a molar ratio of 1:8 together with a nitrogen flow of 50 mLmin
Experimental Section
at ambient pressure. Heptanal, the reaction mixtures, and the sub-
strates mentioned before (5 mL aliquots) were passed through the
General
ꢀ1
reactor at 4508C at a rate of 0.2, 0.167, or 0.140 mLmin together
ꢀ1
with a gas flow of 50, 185, or 213 mLmin at ambient pressure, as
stated in the corresponding figures. The liquid product was con-
densed with an ice bath and analyzed offline by GC with an Agi-
lent 7890A apparatus equipped with an HP-5 column (30 m,
Heptanal was purchased from Aldrich and distilled under reduced
pressure before use. Water was employed in deionized form. m-
ZrO and t-ZrO were purchased from ChemPur, Germany, as pel-
2
2
lets, and m-ZrO -A was obtained from Aldrich as a powder. Cerium
2
0
.32 mm, 0.25 mm) and a flame ionization detector (FID), and n-do-
oxide (nanopowder) was received from Rhodia. Heptanoic acid,
heptanal, and p-toluenesulfonic acid monohydrate were supplied
by Sigma–Aldrich.
decane (Aldrich) was used as an external standard. The gaseous
products were trapped in a gas burette and analyzed by GC with
a Varian 450 instrument in a “refinery gas analyzer” configuration
with three channels. Hydrogen was analyzed with a thermal con-
ductivity detector after separation with a 2 m molecular sieves
Catalyst characterization
(
5 ꢁ) column. Permanent gases such as CO and CO were separat-
2
The X-ray diffraction measurements of m-ZrO , t-ZrO , m-ZrO -A,
2
2
2
ed with a 2.5 m molecular sieves (13X) column and quantified by
a thermal conductivity detector. Low-molecular-weight hydrocar-
bons were separated with a 50 m Plot/Al O column and quantified
and CeO2 were performed to confirm the crystallinities of the
active phases. The analyses were performed with a PANalytical
CUBIX-PRO diffractometer equipped with a PW3050 goniometer
2
3
with a flame ionization detector. 7-Tridecanone was distilled from
the reaction mixture and identified by mass spectroscopy and
(
CuK radiation) with a variable divergence slit. Nitrogen physisorp-
a
13
1
tion isotherms were obtained with a Micromeritics ASAP 2420 ana-
lyzer. The catalysts were outgassed in vacuum at 2008C before the
analysis until the static pressure remained less than 70 Pa. The BET
method was used to calculate the surface area in the relative pres-
C NMR and H NMR spectroscopy (see Supporting Information).
To test the catalyst stability, 15 portions of heptanal (7.33 g) and
water (9.2 g), molar ratio 1:8, were fed consecutively into the reac-
tor (1.0 g m-ZrO ) at 4508C without intermediate catalyst calcina-
2
sure range 1–20 Pa. The TPR of the CeO , m-ZrO t-ZrO , and m-
2
2,
2
tions. The liquid and gas phases were analyzed by GC, as men-
tioned above.
ZrO -A catalysts was performed with a conventional flow apparatus
2
(
Autochem 2910, Micromeritics). A sample (0.3 g) was pretreated in
an O (2%)/He flow at 5508C for 1 h, cooled in a He flow to room
2
temperature, and purged with Ar. The sample was then reduced in
a H2 (10%)/Ar flow. The temperature was increased from room
temperature to 9508C at a constant heating rate of 10 Kmin and
Isotopically labeled experiments in a fixed-bed, continuous-
flow reactor
ꢀ
1
held for 5 min. The water produced during the reduction was re-
The transformation of [1-D]heptanal was performed as described
moved with a frozen n-propanol trap, and the amount of H con-
sumed was monitored with a thermal conductivity detector (TCD).
above at 4508C over m-ZrO (1.0 g, pellets 0.4–0.8 mm). [1-D]Hep-
tanal (2 mL, 0.09 mLmin ) and water (1.275 mL, 0.057 mLmin )
2
2
ꢀ
1
ꢀ1
were fed separately into the reactor in a molar ratio of 1:5 with a ni-
trogen flow of 10 mLmin at ambient pressure. The liquid product
The Lewis acidities of the CeO , m-ZrO t-ZrO , and m-ZrO -A cata-
ꢀ1
2
2,
2
2
lysts were characterized by NH -TPD. The analysis was performed
3
was condensed with an ice bath and analyzed offline by GC. The
gaseous products were trapped in a gas burette and analyzed by
GC and MS with a OmniStar/ThermoStar mass spectrometer (soft-
ware: Quadera QMG220 version 4.40). The conversion of [1-D]hep-
tanal was 89%, and the corresponding ketone, 7-tridecanone, was
obtained in 61% yield. The experimental molar yields of HD and H2
were 46 and 26%, respectively. In comparison, the experimental
global amount of H2 obtained for the non-deuterated heptanal
was 58%.
with an Autochem II chemisorption analyzer. The samples (0.1 g)
were pretreated in an O2 flow at 4508C for 30 min, cooled to
1
1
1
768C in an Ar flow, and saturated with NH in He at a flow rate of
3
1
ꢀ
0 mLmin . Desorption was performed by heating the sample at
ꢀ
1
0 Kmin from 176 to 6008C. The TPD profiles of the catalysts
were recorded with a TCD detector, and the compounds desorbed
were identified with an OmniStar mass spectrometer (Balzers In-
struments). The IR spectroscopy measurements of the catalysts
were performed with a Nicolet iS10 spectrometer with a vacuum
cell. The solids were employed as self-supported wafers of 1 cm di-
ameter and 15–25 mg weight and were degassed for 1 h under
To prove H/D exchange over m-ZrO2, an amount of H2
ꢀ1
(14 mLmin ) similar to that produced during the reaction was fed
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
12
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
ꢀ
1
in the presence of D O (1.4 mL, 0.057 mLmin ) and N2
(
[12] E. F. Iliopoulou, Curr. Org. Synth. 2010, 7, 587–598.
2
ꢀ1
[
13] M. Gli n´ ski, G. Zalewski, E. Burno, A. Jerzak, Appl. Catal. A 2014, 470,
78–284.
14] T. N. Pham, T. Sooknoi, S. P. Crossley, D. E. Resasco, ACS Catal. 2013, 3,
456–2473.
10 mLmin ). For H , DH, and D yields of 14, 25, and 19% were
2 2
obtained, respectively. The same experiment was performed over
SiC, and H/D exchange was not observed.
2
[
2
[
[
15] M. Renz, Eur. J. Org. Chem. 2005, 979–988.
16] R. S. Murthy, P. Patnaik, P. Sidheswaran, M. Jayamani, J. Catal. 1988, 109,
Kinetic measurements of initial rates in a fixed-bed, continu-
ous-flow reactor.
298–302.
[17] F. Hayashi, M. Tanaka, D. Lin, M. Iwamoto, J. Catal. 2014, 316, 112–120.
[18] H. Idriss, E. G. Seebauer, J. Mol. Catal. A 2000, 152, 201–212.
[19] T. Yokoyama, N. Fujita, T. Maki, Appl. Catal. A 1995, 125, 159–167.
[20] S. M. Kim, M. E. Lee, J.-W. Choi, D. J. Suh, Y.-W. Suh, Catal. Commun.
2011, 16, 108–113.
The kinetic measurements were performed in a tubular stainless-
steel reactor at 4508C under ambient pressure with 0.2–0.4 mm
pellets of m-ZrO . Adequate N and heptanal flow rates were se-
2
2
lected to avoid heat- and mass-transfer limitations. To determine
the kinetic expression of the reaction, the initial rates of formation
of the principal product (7-tridecanone) were calculated by keep-
ing the heptanal partial pressure and total flow constant as the
contact time (W/F) was varied by employing different amounts of
catalyst. The heptanal conversions were less than 12%.
[
21] Y. Kamimura, S. Sato, R. Takahashi, T. Sodesawa, T. Akashi, Appl. Catal. A
2003, 252, 399–410.
[22] S. Sato, R. Takahashi, T. Sodesawa, K. Matsumoto, Y. Kamimura, J. Catal.
999, 184, 180–188.
1
[
[
23] V. I. Komarewsky, J. R. Coley, Adv. Catal. 1956, 8, 207–217.
24] J. B. Claridge, M. L. H. Green, S. C. Tsang, A. P. E. York, J. Chem. Soc. Fara-
day Trans. 1993, 89, 1089–1094.
The initial rates were calculated for partial pressures of 1.5, 3, 4.5,
[25] S. Ananthan, N. Venkatasubramanian, C. N. Pillai, J. Catal. 1984, 89,
89–497.
[26] Y. Wang, B. H. Davis, Appl. Catal. A 1999, 180, 277–285.
4
4
and 6ꢂ10 Pa for the heptanal feed (5 mL) with a flow of 0.169,
ꢀ
1
0
.350, 0.518, and 0.690 mLmin respectively. In this series, the ni-
ꢀ
1
[27] J. Wrzyszcz, H. Grabowska, R. Klimkiewicz, L. Syper, Appl. Catal. A 1999,
trogen flow was kept constant (144 mLmin ).
1
85, 153–156.
To test the effect of the addition of water to the reaction on the in-
[28] H. Teterycz, R. Klimkiewicz, B. W. Licznerski, Appl. Catal. A 2001, 214,
243–249.
[29] J. Bussi, S. Parodi, B. Irigaray, R. Kieffer, Appl. Catal. A 1998, 172, 117–
4
itial rate, a heptanal partial pressure of 1.5ꢂ10 Pa was maintained
ꢀ
1
at a heptanal flow of 0.147 mLmin , and the heptanal/water
molar ratio was varied from 1:1 to 1:3 and 1:5. Kinetic modeling to
fit the experimental data with the proposed rate expression was
performed with the Excel Solver program. The activation energy
for the heptanal reaction was determined in the temperature
129.
[
30] S. Foraita, J. L. Fulton, Z. A. Chase, A. Vjunov, P. Xu, E. Barꢃth, D. M. Ca-
maioni, C. Zhao, J. A. Lercher, Chem. Eur. J. 2015, 21, 2423–2434.
31] A. Gangadharan, M. Shen, T. Sooknoi, D. E. Resasco, R. G. Mallinson,
Appl. Catal. A 2010, 385, 80–91.
[
ꢀ1
range 410 to 4508C with a heptanal flow of 0.147 mLmin , a N
[32] M. Kobune, S. Sato, R. Takahashi, J. Mol. Catal. A 2008, 279, 10–19.
[33] N. Plint, D. Ghavalas, T. Vally, V. D. Sokolovski, N. J. Coville, Catal. Today
2
ꢀ1
flow of 144 mLmin , and m-ZrO (23 mg).
2
1
999, 49, 71–77.
[
34] J. Wrzyszcz, H. Grabowska, R. Klimkiewicz, L. Syper, Catal. Lett. 1998, 54,
55–58.
Acknowledgements
[
[
[
35] T. S. Hendren, K. M. Dooley, Catal. Today 2003, 85, 333–351.
36] K. Tanabe, T. Yamaguchi, Catal. Today 1994, 20, 185–197.
37] S. Wan, T. Pham, S. Zhang, L. Lobban, D. Resasco, R. Mallinson, AIChE J.
The authors thank MINECO (CTQ2015-67591-P, Consolider Ingenio
010-MULTICAT and Severo Ochoa program, SEV-2012-0267) and
2
2
013, 59, 2275–2285.
Generalitat Valenciana (PROMETEO II/2013/011 Project) for fund-
ing this work. L.M.O. is grateful to the COLCIENCIAS institute for
a Francisco-Jose-de-Caldas (512/2010) doctoral fellowship.
[38] A. Corma, L. M. Orozco, M. Renz, New J. Chem. 2013, 37, 3496–3502.
[
[
39] J.-R. Kim, W.-J. Myeong, S.-K. Ihm, Appl. Catal. B 2007, 71, 57–63.
40] T. J. Farmer, M. Mascal in Introduction to Chemicals from Biomass (Eds.: J.
Clark, F. Deswarte), Wiley, Chichester, 2015, pp. 89–155.
41] H.-B. Hu, K.-W. Park, Y.-M. Kim, J.-S. Hong, W.-H. Kim, B.-K. Hur, J.-W.
Yang, J. Ind. Eng. Chem. 2000, 6, 238–241.
[
[
[
Keywords: aldehydes
·
biomass
·
carboxylic acids
·
42] H. Idriss, C. Diagne, J. P. Hindermann, A. Kiennemann, M. A. Barteau, J.
Catal. 1995, 155, 219–237.
heterogeneous catalysis · ketones
43] M. V. Ganduglia-Pirovano, A. Hofmann, J. Sauer, Surf. Sci. Rep. 2007, 62,
219–270.
[
[
[
1] M. J. Climent, A. Corma, S. Iborra, Green Chem. 2014, 16, 516–547.
2] D. Y. Murzin, I. L. Simakova, Catal. Ind. 2011, 3, 218–249.
[44] G. da Silva, J. W. Bozzelli, J. Phys. Chem. A 2006, 110, 13058–13067.
[45] S. Y. Onozawa, T. Sakakura, M. Tanaka, M. Shiro, Tetrahedron 1996, 52,
4291–4302.
3] J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007, 46,
164–7183; Angew. Chem. 2007, 119, 7298–7318.
7
[
[
4] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
5] A. Pulido, B. Oliver-Tomas, M. Renz, M. Boronat, A. Corma, ChemSu-
sChem 2013, 6, 141–151.
[46] T. Yokoyama, T. Setoyama, N. Fujita, M. Nakajima, T. Maki, K. Fujii, Appl.
Catal. A 1992, 88, 149–161.
[47] T. Merle-Mꢄjean, P. Barberis, S. B. Othmane, F. Nardou, P. E. Quintard, J.
Eur. Ceram. Soc. 1998, 18, 1579–1586.
[48] C. Battilocchio, J. M. Hawkins, S. V. Ley, Org. Lett. 2013, 15, 2278–2281.
[49] T. Komanoya, K. Nakajima, M. Kitano, M. Hara, J. Phys. Chem. C 2015,
119, 26540–26546.
[50] B. Oliver-Tomꢃs, F. Gonell, A. Pulido, M. Renz, M. Boronat, Catal. Sci.
Technol. 2016, 6, 5561–5566.
[
[
[
[
6] G. W. Huber, J. N. Chheda, C. J. Barrett, J. A. Dumesic, Science 2005, 308,
1
446–1450.
7] J. Q. Bond, D. M. Alonso, D. Wang, R. M. West, J. A. Dumesic, Science
010, 327, 1110 – 1114.
8] A. Corma, O. de la Torre, M. Renz, N. Villandier, Angew. Chem. Int. Ed.
011, 50, 2375–2378; Angew. Chem. 2011, 123, 2423–2426.
9] A. Corma, O. de la Torre, M. Renz, Energy Environ. Sci. 2012, 5, 6328–
344.
2
2
6
[
[
10] B. G. Harvey, H. A. Meylemans, Green Chem. 2014, 16, 770–776.
11] R.-J. van Putten, J. C. van der Waal, E. de Jong, C. B. Rasrendra, H. J.
Heeres, J. G. de Vries, Chem. Rev. 2013, 113, 1499–1597.
Received: May 16, 2016
Published online on && &&, 0000
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
13
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
L. M. Orozco, M. Renz,* A. Corma*
&& – &&
Carbon–Carbon Bond Formation and
Hydrogen Production in the
Ketonization of Aldehydes
The key to ketones: By isolating molec-
ular hydrogen as a coproduct in the ke-
tonization of aldehydes, its sacrificial ox-
idation to water is avoided, and the
energy liberated in the oxidation of an
aldehyde into a carboxylic acid can be
stored. The hydrogen source is water. It
is confirmed that carboxylic acids are in-
termediates in the ketonization of alde-
hydes, and the second step is the classi-
cal decarboxylation of carboxylic acids.
&
ChemSusChem 2016, 9, 1 – 14
www.chemsuschem.org
14
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!