Direct conversion of carboxylic acids (Cn) to alkenes (C2n - 1) over titanium oxide in absence of noble metals

Article abstract of DOI:10.1016/j.molcata.2016.01.019

Carbon-carbon bond formations and deoxygenation reactions are important for biomass up-grading. The classical ketonic decarboxylation of carboxylic acids provides symmetrical ketones with 2n + 1 carbon atoms and eliminates three oxygen atoms. Herein, this reaction is carried out with titanium oxide at 400°C, and an olefin with 2n + 1 carbon atoms is obtained instead of the ketone. For olefin formation hydrogen transfer reactions are required from suitable precursors to form aromatics and coke. Additional aldol condensation reactions increase further molecular weight in the product mixture. Hence, a combination of titanium oxide with a hydrodeoxygenation bed provides double amount of diesel fuel as the combination with zirconium oxide when reacting hexose-derived pentanoic acid.

Full text of DOI:10.1016/j.molcata.2016.01.019

Accepted Manuscript
Title: Direct conversion of carboxylic acids (C ) to alkenes
n
(
C2 ) over titanium oxide in absence of noble metals
n−1
Author: Borja Oliver-Tomas Michael Renz Avelino Corma
PII:
S1381-1169(16)30019-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2016.01.019
MOLCAA 9751
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Accepted date:
30-12-2015
18-1-2016
Please cite this article as: Borja Oliver-Tomas, Michael Renz, Avelino Corma,
Direct conversion of carboxylic acids (Cn) to alkenes (C2nminus1) over titanium
oxide in absence of noble metals, Journal of Molecular Catalysis A: Chemical
http://dx.doi.org/10.1016/j.molcata.2016.01.019
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Direct conversion of carboxylic acids (C ) to alkenes
n
(
C2 ) over titanium oxide in absence of noble metals
n–1
Borja Oliver-Tomas, Michael Renz*, Avelino Corma* acorma@itq.upv.es
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, Tel.: +34 96 387 7800, Fax: +34 96 387 9444.

Graphical Abstract

Highlights
TiO
2
2
provides the ketonic decarboxylation product at lower temperature than ZrO
2
TiO
transforms the ketone into the corresponding olefin by hydride transfer
Additional Carbon-carbon bonds increase diesel fraction from pentanoic acid

Abstract
Carbon-carbon bond formations and deoxygenation reactions are important for
biomass up-grading. The classical ketonic decarboxylation of carboxylic acids provides
symmetrical ketones with 2n+1 carbon atoms and eliminates three oxygen atoms.
Herein, this reaction is carried out with titanium oxide at 400 ºC, and an olefin with
2n+1 carbon atoms is obtained instead of the ketone. For olefin formation hydrogen
transfer reactions are required from suitable precursors to form aromatics and coke.
Additional aldol condensation reactions increase further molecular weight in the
product mixture. Hence, a combination of titanium oxide with a hydrodeoxygenation
bed provides double amount of diesel fuel as the combination with zirconium oxide
when reacting hexose-derived pentanoic acid.
Keywords: Aldol condensation; Carbon-carbon bond formation; Ketonic
decarboxylation; Hydride transfer; Second generation diesel fuel

Introduction
The connection of two carboxylic acids to a ketone, i.e. the ketonic decarboxylation or
ketonization of carboxylic acids, has experienced a revival during the last decade.[1, 2]
Before World War I this reaction was applied industrially for the synthesis of small
ketones, especially acetone. The change to a petroleum based economy has
generated competitive routes and for instance acetone is produced from cumene as
by-product of the Hock process. For this reason most industrial processes based on
the ketonic decarboxylation have been discontinued.
At the beginning of our century sustainability became an important issue in
chemistry.[3] New biorefinery concepts were designed based on renewable materials
such as lignocellulose since the latter does not compete with food production.[4]
However, the composition of these bio-based materials is diametrically opposed to the
composition of petroleum-based materials: the former are extremely oxygen-rich
whereas the latter are nearly oxygen free. This means that the bio-based raw materials
have to be deoxygenated to convert them into classical commercial products derived
from petroleum. As a second characteristic difference, lignocellulose is built-up from
small molecules from five to nine carbon atoms. When lignocellulose is degraded
without cleaving carbon-carbon bonds, mainly sugar molecules of five and six carbon
atoms are obtained. As a consequence, two or more of these five to six carbon-atom
molecules have to be connected together if molecules with a larger number of carbon
atoms are desired for their use as chemicals or fuels.[5-7]
The carbon-carbon bond formation in biomass derived products can be carried out
after complete hydrodeoxygenation of the feed or starting with oxygen-containing
platform molecules. Both strategies have been employed in proposals for the
production of fuels from biomass. Hence, hexose is completely deoxygenated to
butene which is then oligomerized.[8-11] Aldol condensation reactions[12, 13] and
hydroalkylations/alkylations[14-19] have also been employed to build up fuel
precursors from small molecules. Thereby, it can be seen that the latter
transformations can proceed with a high degree of selectivity (with respect to the
carbon atom number), and it has been proposed to carry out the carbon-carbon bond
formations with moderately oxygenated molecules.[20] The ketonic decarboxylation fits
very well this requirement since two carboxylic acid molecules with n carbon atoms are
converted into one molecule of ketone with 2n–1 carbon atoms, carbon dioxide and
water (Eq. 1).[1, 2] By-products are innocuous and solvent is not required. Hence, the
reaction also meets the demands of Green Chemistry.[3] The reaction is carried out in
the gas phase over a metal oxide catalyst. Different sites intervene in the reaction
mechanism: Lewis-acid metal centers and basic surface oxygen atoms. It has been

proposed that the carboxylic acid form
decarboxylated.[21] Interestingly, many different oxides have been reported as suitable
materials for this transformation: CeO ,[21-24] ZrO ,[21, 25, 26] MgO,[27, 28]
MnO ,[26, 29] Mn ,[26] Fe ,[30, 31] Fe ,[32] Cr ,[33, 34] Al ,[21, 26]
ThO ,[35, 36] Bi ,[26] TiO .[37, 38]. In many cases the desired ketone product has
a
beta-keto acid which is then
2
2
2
O
3 4
O
2 3
O
3 4
2
O
3
2 3
O
2
O
2 3
2
been obtained almost quantitatively.
Herein we report on the use of titanium dioxide as catalyst for the ketonic
decarboxylation. Although this material has been employed before as catalyst and
support for this reaction,[37-41] it will be described here that the reaction outcome at
selected reaction conditions, is very different from the performance of other materials
2 2
such as ZrO . In fact, on TiO an olefin was observed as main product instead of the
typical ketone obtained when working at 400 ºC. The mechanism of the olefin formation
has been studied here and it will be presented that a hydride shift is involved in
absence of any noble metal. Furthermore, changes in the hydrogenated product
mixture induced by aldol condensations, which occur under the ketonic decarboxylation
2
conditions, will be discussed in comparison with ZrO .
Experimental Part
General
Decanoic acid, pentanoic acid, 2-undecanone, 2-pentanone, cyclopentanone, 4-
methyl-1-cyclohexene, -terpinene, limonene and nitrobenzene were supplied from
Acros, Aldrich and Takasago, and used as received. Titanium oxide (surface area 151
2
m /g) and monoclinic zirconium oxide were received from ChemPur (Germany) as
2
pellets and a titanium oxide sample from Evonik (P25; surface area 55 m /g) in powder
form. They were pelletized (if necessary), crushed and sieved, and the 0.4 – 0.8 mm
fraction employed as catalysts. All reactions were carried out in a tubular fixed-bed
continuous flow reactor. The set-up has been described in a previous work.[21] Briefly,
the reagents were placed in a syringe at room temperature except for decanoic acid for
which the syringe was heated to 45 ºC. The reactor was loaded with 2.50 g of catalyst
–
1
diluted with silicon carbide. Typically, a feed rate of 0.15 mL min was employed and
–1
nitrogen or hydrogen gas was used as carrier gas (50 or 100 mL min ). The reactions
were carried out at atmospheric pressure and at temperatures between 300 and 450
ºC. An ice bath or acetone plus dry ice bath were used to condense the products.

Different gas chromatographs were employed for analyzing the products off line. The
liquid products were analyzed with an Agilent 7890A with split injection equipped with a
HP-5 column (30 m x 0.320 mm x 0.25 µm) and for simulated distillation a Bruker 450-
GC gas chromatograph equipped with an Agilent J&W GC Column CP-SimDist (10 m x
0.53 mm x 2.65 µm), both of them with a FI detector. Permanent gases were analyzed
off-line on a GC apparatus RGA analyzer equipped with three channels. Hydrogen was
analyzed on a 2 m molecular sieve 5 Å column and TCD detector. Permanent gases
such as CO and CO
quantified by a TCD detector. Low molecular weight hydrocarbons were analyzed on a
0 m Plot/Al column with a FI detector. The reaction products were identified by
GC-MS with an Agilent Technologies 6890N gas chromatograph using a HP-5 column
30 m x 0.320 mm x 0.25 µm) coupled with a mass selective detector Agilent 5973N.
2
were separated on a 2.5 m molecular sieve 13X column and
5
2 3
O
(
Catalyst characterization
2
The X-ray diffraction measurements to analyze the phase of TiO were performed on a
PANalytical CUBIX-PRO diffractometer equipped with a PW3050 goniometer (Cu Kα
radiation) provided with a variable divergence slit. Nitrogen physisorption isotherms
were obtained using a Micromeritics ASAP 2420 analyzer. The metal oxide was out-
gassed in vacuum at 400 ºC for 12 h prior to the analysis. The Brunauer–Emmett–
Teller (BET) method was used to calculate the surface area in the range of relative
pressures between 0.01 and 0.20 Pa and the Barrett-Joyner-Halenda model for the
determination of the pore volumes. Infrared (IR) measurements were performed with a
Nicolet 710 FTIR spectrometer using vacuum cells. For the measurement, the metal
oxide was pressed into a self-supporting wafer and was activated overnight at 400 ºC
–4
–5
and in vacuum of 10 to 10 Torr in the IR chamber. Pyridine was allowed to enter the
chamber and excess pyridine was desorbed at room temperature applying vaccum.
Then pyridine was desorbed at increasing temperature at 150, 250 and 350 ºC for 60
min at each temperature. After each desorption step the sample was cooled down and
an IR spectrum recorded. The thermogravimetric (TG) analysis was obtained on a
Mettler Toledo TGA/SDTA 851 instrument. The sample (approximately 5–10 mg) was
placed in an alumina crucible on a precision balance with a S type thermocouple. The
sample was heated at a rate of 10 K per minute from 30 ºC to the final temperature in a
controlled atmosphere.
Synthesis and catalytic reaction of the aldol condensation product of cyclopentanone
(
2-cyclopentylidenecyclopentanone)

Into a 100 mL round bottom flask equipped with reflux condenser and magnetic stirring
bar cyclopentanone (50 mL) was introduced and an aqueous NaOH solution (10 wt.%,
5
mL). The reaction mixture was stirred and heated to reflux temperature. After 20 h
the mixture was cooled down to room temperature and the organic phase was
separated from aqueous phase by decanting. The crude product was distilled under
reduced pressure and 2-cyclopentylidenecyclopentanone was isolated (97% purity,
43% yield).
The condensation product (97%) was passed through the fixed-bed continuous-flow
reactor (described above) filled with 2.50 g of TiO
2
at 425 ºC, together with nitrogen as
–
1
–1
–1
carrier gas (100 mL min ; F = 9.09 g h ; WHSV = 3.64 h ).
Isolation of the olefin product and catalytic transformation of the latter
2
The product fractions obtained when passing 2-undecanone over TiO were jointed
and the mixture (38 g) was distilled at reduced pressure (5 mbar). The distillate was
collected in different fractions obtaining a 17% yield for the C11 olefinic fraction.
This product fraction was used as feed and passed through the fixed-bed reactor filled
–1
with 2.50 g of TiO
2
, together with nitrogen as carrier gas (100 mL min ). The flow rate
–1
of the organic liquid was 6.66 g h and the reaction temperature 425 ºC (WHSV = 2.66
–
1
h ).
Results and Discussion
Catalyst characterization
2
The crystalline phase and textural properties of the TiO catalyst (ChemPur) were
evaluated by X-ray diffraction and nitrogen adsorption. The X-ray diffraction pattern
shows that the titanium oxide was crystallized as anatase phase (Figure 1). Broad
signals indicate that the particles are very small. This was confirmed by measuring the
2
surface area by nitrogen adsorption. A surface area of 151 m /g was determined which
is quite high for a metal oxide. Pore diameters of 79 Å (BJH) measured are typical for
the interparticle void between the oxide particles.
The material possessed considerable Lewis acidity. This was evidenced by in-situ IR
spectroscopy with pyridine as a probe molecule (Figure 2). The typical band at 1450
–1
cm indicates strong Lewis acidity since it is still detected even after pyridine
desorption at 350 ºC in vacuum. The sample did not show any Brönsted acidity (cf.
–1
Figure 2, band lacking at 1550 cm ).
From these analyses it can be concluded that the titanium oxide consisted of small
nanoparticles with a high surface area and strong Lewis acid sites.

Catalytic performance of TiO
2
Titanium oxide was placed in a fixed bed continuous flow reactor to assess its catalytic
activity for the ketonic decarboxylation. Hence, decanoic acid was selected as
substrate and passed through the catalytic bed in the temperature range of 300 to 450
ºC (Figure 3). It can be seen that the catalytic performance differs significantly from the
zirconium oxide. Considerable catalytic activity was already observed at 300 ºC with
the TiO
temperatures the ketone was the only product observed on TiO
50 ºC and higher, the conversion was complete but a second product was obtained: a
2
catalyst. At 325 ºC conversion was already higher than 50%. At both
2
. At a temperature of
3
mixture of nonadecene isomers, i.e. molecules with the same number of carbon atoms
but with a double bond instead of the carbonyl functional group. In a working
hypothesis it was proposed that the olefin was produced from the ketone by reduction
to the corresponding alcohol and subsequent dehydration (Eq. 2). The formation of the
olefin increased with temperature from 325 to 400 ºC. At 400 ºC a maximum yield of
approximately 35% was observed (Figure 3). At higher temperature the yield
decreased slightly. The maximum yield for the ketone was observed at 350 ºC with
approximately 70%. At higher temperature the yield decreased and at 400 ºC almost
no ketone was observed anymore (Figure 3). Furthermore, it was observed that one
condition for this outstanding reactivity of the titanium oxide was the high surface area.
2
2
When a sample (P25) with a surface area of 55 m /g instead of 151 m /g was
employed, the selectivity towards the ketone below 400 ºC was 90% and the olefin
yield reached only 25% at 425 ºC (See Supplementary Information, Figure S1).
2
When ZrO was employed as catalyst, at the same temperature, i.e. 400 ºC, the
maximum yield of ketone (approx. 95%) was observed. At higher temperatures the
selectivity decreased slightly and below 400 ºC the conversion was clearly lower with
ZrO
TiO
with TiO
2
than with TiO
, the C19-olefin was not observed in significant amounts. The formation of the olefin
as catalyst was unexpected since this product has not been reported before
2 2
. However, with ZrO , in contrast to the transformation catalyzed by
2
2
in the literature on a metal oxide, as far as we know. Highly selective transformation to
the ketone[42, 43] or water and tar producing reactions have been described at above
350 ºC.[39] In any case, the different reaction outcome was interesting and its origin
should be elucidated.

When the yield of the two products was depicted versus conversion (Figure S3), it can
be concluded from the shape of the graph that the ketone is a primary unstable product
and the olefin a secondary product. This result supports the proposed formation of the
olefin from the ketone as depicted in Eq 2. For further evidence, a ketone, namely 2-
undecanone, was submitted to the reaction conditions. Also in this case the olefin was
observed for a time on stream of 2.8 h (Figure 4). During this time the yield was quite
constant at approx. 25%. However, conversion of the ketone decreased slightly from
full conversion to approximately 90%. When the olefin produced from 2-undecanone,
was purified by distillation (90% purity) and submitted to the reaction conditions, almost
no change was observed (conversion < 2%). This means that the olefin is relatively
stable under reaction conditions and does not undergo neither hydrogenation nor
oligomerization reactions. Interestingly, the yield of the olefin from the ketone was not
influenced by the carrier gas: an identical yield was observed with either nitrogen or
hydrogen as carrier gas (Figure 4). The rest of the products were lower molecular
weight olefins (C6 to C10; < 10% yield; for more detail see Figure S4) and higher
molecular weight products (40 to 50%).
In summary, experimental results confirm titanium oxide as a ketonic decarboxylation
catalyst. However, under certain reaction conditions (higher reaction temperature) the
ketone is transformed into an olefin. This reaction can be depicted as a cascade
reaction, i.e. as reduction of the ketone to the corresponding alcohol and subsequent
dehydration (Eq. 2). However, for such a process, a reductant or a hydrogen source is
missing. This issue will be further studied in the next section.
Hydrogen transfer reactions with titanium oxide
With the aim to reveal the nature of the reductant, the evolution of the reaction was
followed for a longer time on stream. Hence, in Figure 5 it can be seen that the
formation of the olefin is quite constant for 340 min involving a small decrease. On the
other hand, the supposed intermediary ketone is observed in increasing amounts
(faster than the olefin disappears) and after 230 min the yield is very similar to the
olefin. After 340 min, the catalyst was treated with air at 450 ºC but the tendencies in
the product composition continued for both products and a recovery of the initial
catalytic activity for the olefin formation was not observed, when the catalyst was used
again.
In the long term experiment more than 35 g of carboxylic acid was processed per g of
2
catalyst and 2 equivalents (in mol) of olefin were obtained per mol of TiO . This

indicated that a surface modification of the catalyst or changes in the catalyst bulk
composition were not the origin for ketone reduction. After the reaction, the amount of
coke was measured by TG to be 7.6 wt% of the catalyst. It cannot be excluded that
there was some contribution of the coke to the reduction of the intermediate ketone but,
again, the amount of coke is too low to count for the ketone reduced to provide the
olefin. Therefore, it was concluded that a hydrogen source has to be formed as a side
product. As a starting hypothesis it was speculated that hydrogen transfer reactions
occur and that cyclohexane rings, saturated or partly unsaturated, act as hydrogen
source. In order to check this hypothesis, first dehydrogenations were studied under
reaction conditions. Hence, potential hydrogen-donor model compounds such as
methyl-cyclohexene, -terpinene and limonene were selected. In literature, it has been
2
shown recently that the latter can be dehydrogenated with TiO to the corresponding
product, i.e. para-cymene, which was observed in 90% yield.[44] However, it has been
further reported that the catalytic activity depended not only on the chemical
composition of the catalyst, but also on the properties and the amount of the catalytic
2
sites. Therefore, the activity of the TiO used in the present study had to be confirmed
for this dehydrogenation reaction. Indeed, under the present reaction conditions, the
conversion of limonene and -terpinene was complete and para-cymene was obtained
in 65 and 68% yield, respectively (Table 1, entries 1 and 2). Furthermore,
dehydrogenation was slightly higher since approximately 5% of side products were
probably formed from para-cymene by Lewis-acid catalyst isomerization reactions.
If the substrate is less prone to aromatization, for instance in the case of cyclohexene
derivatives with only one double bond, isomerization of the carbon skeleton competes
even more efficiently. Hence, with 4-methyl-1-cyclohexene (1) the corresponding
dehydrogenation product, i.e. toluene (2), was observed with a yield of 15% (Eq. 3).
With the same skeleton, double bond isomerization (compound 3) and hydrogenation
(compound 4) was observed. Other products involved carbon skeleton isomerization,
for instance compounds 5, 6 and 7.
(
3)
1
2
3
4
5
6
7
Other
25.6
Yield/%: 15.4
20.2
12.1
2.3
2.6
9.7

Although the yield of the dehydrogenation product was low in the case of cyclohexene
, all three test reactions demonstrate unambiguously that dehydrogenation reactions
1
are possible under the employed reaction conditions, even in absence of metal
catalysts based on precious metals, Cu or Ni. As next step the hydrogen transfer
should be confirmed. As the reduction product of the ketone, i.e. the corresponding
alcohol, was supposed to be unstable under reaction conditions and dehydrated to the
olefin, a different substrate was selected, namely nitrobenzene. The latter can be
hydrogenated easily with formation of two molecules of water. Hence, mixtures of
limonene with nitrobenzene and -terpinene with nitrobenzene in a 3 to 1 molar ratio
were contacted with titanium oxide at 425 ºC. The dehydrogenation of the
monoterpene worked as before and para-cymene was obtained in a yield of 64 to 76%
(
Scheme 1). However, additionally the hydrogen evolved from the molecule was
utilized to reduce nitrobenzene to aniline. This means that the hydrogen transfer
reaction worked well and hydrogen loss, if any, was minimal.
Once having demonstrated that hydrogen transfer reactions are possible with titanium
oxide at elevated temperatures, it should be shown that these reactions include the
reduction of a ketone to the corresponding alcohol. Hence, 2-pentanone as model
compound of a ketone and in a second experiment pentanoic acid were contacted to
titanium oxide in presence of alpha-terpinene. From equations 4 and 5 it can be seen
that the yield of para-cymene increased significantly from 68% (Table 1, entry 2) when
the monoterpene is reacted alone, to over 80%. On the other hand, the corresponding
olefin was observed in 28 and 37%, respectively. In the case of the olefin, the
improvement for the yield was not as clear as it was for the aromatization. However,
the beneficial effect might be masked here by the low selectivity and the formation of
side products from the alcohol.
–1
–1
2
Reaction conditions: 2.50 g of TiO , F = 7.35 g h , WHSV = 2.94 h , 425 ºC, 100 mL
–1
2
min of N .

–1
–1
2
Reaction condtions: 2.50 g of TiO , F = 8.06 g h , WHSV = 3.22 h , 425 ºC, 100 mL
–1
2
min of N .
In summary, the experimental results demonstrated that hydrogen transfer reactions
can occur under reaction conditions and that it is very likely that the ketone is
hydrogenated to the alcohol which is then dehydrated to the olefin. The missing piece
in the puzzle, however, is still the hydrogen donor molecule. This issue should be
strongly related to the observation that significant amounts of higher molecular weight
molecules (cf. Figure S4), i.e. molecules with higher molecular weight than the primary
ketone product, are formed during the reaction. It is supposed that both issues are
related to aldol condensation products of the ketonic decarboxylation product, i.e. the
ketone.
The impact of aldol condensation reactions
Titanium oxide is known to catalyze aldol condensations, for instance of acetaldehyde
or acetone, at 250 ºC and below.[45-47] As a consequence, it has to be assumed that
the primary reaction product of the ketonic decarboxylation, i.e. the ketone which is
prone to aldol condensation, can further react under our reaction conditions. Hence,
aldol condensations and several other side-reactions (carbon skeleton isomerizations
as for instance observed in the transformation of 4-methyl-1-cyclohexene, 1, reported
before) generate a complex product mixture at higher reaction temperature (> 400 ºC).
With the aim to identify different reaction pathways, a ketone with lower molecular
weight, i.e. 2-pentanone, was selected and submitted to reaction conditions at different
W/F contact times. With this substrate it was possible to identify and characterize
products built from two or three substrate molecules (Figure 6).
At lower conversion (approx. 30%), the aldol condensation product (unsaturated C10
-
ketone; Scheme 2) was obtained with higher selectivity and yield (almost 10% yield). At

higher conversion the yield decreased to below 3% indicating that this ketone was a
primary, unstable product. Therefore, the unsaturated C10-ketone can undergo a
second aldol condensation with subsequent cyclization and dehydration to yield a
15
C H24 aromatic isomer, and can also be hydrogenated to the saturated C10-ketone.
This again supports clearly that hydrogen transfer reactions occur under the reaction
conditions, being in this case the double bond of an α,β-unsaturated ketone the
hydrogen receptor. Additionally, the hydrogenated condensation product might loose
7 8
two or three carbon atoms by the McLafferty rearrangement and give C - and C -
ketones. Furthermore, as observed for the original ketonic decarboxylation reaction, all
saturated ketones were converted partly into the corresponding olefins.
The products of interest for the hydrogen production are precursors of the C10
aromatics and phenols (Figure 6 and Scheme 2). These can be formed from the
unsaturated ketone under liberation of one or two equivalents of molecular hydrogen,
respectively. These products clearly identified a reaction pathway to produce hydrogen
(equivalents) during their formations.
Another interesting observation was made when employing the aldol condensation
product of cyclopentanone as substrate. In this case, skeleton rearrangement and ring
expansion occurred (Eq. 6), transforming a five-membered ring into a six-membered
ring molecule. Similar skeleton rearrangements were mentioned before. Hence, after
this ring expansion, smaller aliphatic rings may also serve as hydrogen source.
(
6)
From this study on the reactivity of ketones and aldol condensation products it can be
concluded unambiguously that hydrogen transfer reactions occur with titanium oxide
without any precious metal. In the case of the α,β-unsaturated ketone (aldol
condensation product) the direct hydrogenation product, i.e. the saturated ketone, was
detected. As hydrogen donor, aldol condensation products, after cyclization, are
proposed. Products such as C10 aromatics and phenols were identified which can be
obtained from the aldol condensation product, together with one and two equivalents of
hydrogen, respectively. Then, it becomes clear that the olefins, observed in the ketonic
decarboxylation of carboxylic acids, are formed in a cascade reaction from the primary

ketone product by hydrogenation and dehydration. On the other hand, it has been
demonstrated that titanium oxide possesses is a high active catalyst for aldol
condensations, under the present reaction conditions. This property is clearly
responsible for further carbon–carbon bond formations.
2 2
Advantages of TiO versus ZrO for the production of fuels by ketonization
The additional carbon-carbon bond formation activity should be interesting for the
formation of fuels from biomass platform molecules since an increase in the molecular
weight of the products is desired. Hexose-derived pentanoic acid was transformed
before into nonane as kerosene component.[48] We achieved that with a double
catalytic bed involving ZrO
2 2 3
as ketonic decarboxylation catalyst and Pt/Al O for
subsequent hydrodeoxygenation. Now, the ketonic decarboxylation should be
combined with subsequent aldol condensation activity before hydrodeoxygenation, i.e.
2 2
ZrO substituted by TiO in the first catalytic bed. When the reaction was carried out
with the two catalytic beds, simulated distillation of the product (Figure 7) after the two
step process, shows nicely that the diesel yield is approximately doubled compared
2
with ZrO . Nevertheless, the gasoline fraction increases, probably due to an increased
activity of the metal oxide for fragmentation reactions but also to the introduction of
branching in the molecules which also facilitates fragmentations. In Figure 8, it can be
clearly seen that the substitution of ZrO
when diesel fuel is targeted.
2 2
by TiO in the two step process is favorable
The substitution has an additional effect on the product composition. Hence, when ZrO
was employed the main product was nonane with a selectivity of approx. 75%. In
contrast, with TiO not only alkanes were obtained but also 5 to 10% of aromatic
2
2
compounds (Figure 9). Furthermore, under the same reaction conditions, the
hydrodeoxygenation was not complete and up to 5% of oxygenated compounds were
detected.
Conclusions
The catalytic activity of TiO
decarboxylation. It has been found that the catalytic activity of TiO
temperature than with ZrO . With TiO , several reactions, i.e. ketonic decarboxylation,
reduction and dehydration, occur in a cascade mode. Thereby, at 400 ºC the main
product observed was an olefin instead of the ketone obtained with ZrO . The origin of
2
has been compared with ZrO
2
for the ketonic
2
starts at lower
2
2
2
the reductant could not be clearly identified, but dehydrogenations–hydrogen transfer
from hydrogen donating molecules are very likely since aromatics (approx. 11%) and
0.22 wt% coke were formed. Precursors for the hydrogen production can be formed by

aldol condensation reactions. The latter reactions also contribute to carbon-carbon
bond formations which has a positive effect on the average molecular weight of the
product mixture after hydrodeoxygenation.
Acknowledgement
The authors thank MINECO (MAT2011-28009, Consolider Ingenio 2010-MULTICAT,
CSD2009-00050 and Severo Ochoa program, SEV-2012-0267), Generalitat
Valenciana (PROMETEOII/2013/011 Project), and the Spanish National Research
Council (CSIC, Es 2010RU0108). B. O.-T. is grateful to the CSIC (JAE Programme) for
his PhD fellowship.

References
[
[
1] M. Renz, Eur. J. Org. Chem., (2005) 979-988.
2] T.N. Pham, T. Sooknoi, S.P. Crossley, D.E. Resasco, ACS Catal., 3 (2013) 2456-
2
473.
[
[
3] P. Anastas, N. Eghbali, Chem. Soc. Rev., 39 (2010) 301-312.
4] J.H. Clark, R. Luque, A.S. Matharu, Annu. Rev. Chem. Biomol. Eng., 3 (2012) 183-
2
07.
[
[
[
[
5] G.W. Huber, S. Iborra, A. Corma, Chem. Rev., 106 (2006) 4044-4098.
6] A. Corma, S. Iborra, A. Velty, Chem. Rev., 107 (2007) 2411-2502.
7] M.J. Climent, A. Corma, S. Iborra, Green Chem., 16 (2014) 516-547.
8] G.W. Huber, J.N. Chheda, C.J. Barrett, J.A. Dumesic, Science, 308 (2005) 1446-
1
450.
[
[
9] D.Y. Murzin, I.L. Simakova, Catal. Ind., 3 (2011) 218-249.
10] S. Sankaranarayanapillai, S. Sreekumar, J. Gomes, A. Grippo, G.E. Arab, M.
Head-Gordon, F.D. Toste, A.T. Bell, Angew. Chem. Int. Ed., 54 (2015) 4673-4677.
11] B.M. Antunes, A.E. Rodrigues, Z. Lin, I. Portugal, C.M. Silva, Fuel Process.
Technol., 138 (2015) 86-99.
12] J.Q. Bond, D.M. Alonso, D. Wang, R.M. West, J.A. Dumesic, Science, 327 (2010)
110-1114.
13] Y.T. Kim, J.P. Chada, Z. Xu, Y.J. Pagan-Torres, D.C. Rosenfeld, W.L. Winniford,
E. Schmidt, G.W. Huber, J. Catal., 323 (2015) 33-44.
14] A. Corma, O. de la Torre, M. Renz, N. Villandier, Angew. Chem. Int. Ed., 50 (2011)
375-2378.
[
[
1
[
[
2
[
[
15] A. Corma, O. de la Torre, M. Renz, Energ. Environ. Sci., 5 (2012) 6328-6344.
16] I. Yati, M. Yeom, J.W. Choi, H. Choo, D.J. Suh, J.M. Ha, Appl. Catal. A: Gen., 495
(
[
3
2015) 200-205.
17] M.H. Chávez-Sifontes, M.E. Domine, S. Valencia, Catal. Today, 257 (2015) 305-
17.
[
[
18] M. Balakrishnan, E.R. Sacia, A.T. Bell, ChemSusChem, 7 (2014) 1078-1085.
19] G. Li, N. Li, X. Wang, X. Sheng, S. Li, A. Wang, Y. Cong, X. Wang, T. Zhang,
Energy Fuels, 28 (2014) 5112-5118.
[
[
20] D.E. Resasco, J. Phys. Chem. Lett., 2 (2011) 2294-2295.
21] A. Pulido, B. Oliver-Tomas, M. Renz, M. Boronat, A. Corma, ChemSusChem, 6
(
[
(
2013) 141-151.
22] O. Nagashima, S. Sato, R. Takahashi, T. Sodesawa, J. Mol. Catal. A: Chem., 227
2005) 231-239.
[
[
[
[
23] R.W. Snell, B.H. Shanks, ACS Catal., 4 (2014) 512-518.
24] R.W. Snell, B.H. Shanks, ACS Catal., 3 (2013) 783-789.
25] K. Parida, H.K. Mishra, J. Mol. Catal. A: Chem., 139 (1999) 73-80.
26] R. Pestman, R.M. Koster, A. van Duijne, J.A.Z. Pieterse, V. Ponec, J. Catal., 168
(1997) 265-272.
[
[
27] A. Corma, M. Renz, C. Schaverien, ChemSusChem, 1 (2008) 739-741.
28] G.A.H. Mekhemer, S.A. Halawy, M.A. Mohamed, M.I. Zaki, J. Catal., 230 (2005)
1
09-122.
[
[
29] G.P. Hussmann, in, Amoco Corporation, 1988.
30] R. Pestman, R.M. Koster, E. Boellaard, A.M. van der Kraan, V. Ponec, J. Catal.,
1
[
74 (1998) 142-152.
31] E. Karimi, I.F. Teixeira, L.P. Ribeiro, A. Gomez, R.M. Lago, G. Penner, S.W. Kycia,
M. Schlaf, Catal. Today, 190 (2012) 73-88.
32] E.J. Grootendorst, R. Pestman, R.M. Koster, V. Ponec, J. Catal., 148 (1994) 261-
69.
[
2
[
[
33] J.C. Kuriacose, R. Swaminathan, J. Catal., 14 (1969) 348-354.
34] H. Bayahia, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. B Environ., 165
(2015) 253-259.

[
[
[
[
35] S.S. Kistler, S. Swann, E.G. Appel, Ind. Eng. Chem., 26 (1934) 388-391.
36] A.L. Miller, N.C. Cook, F.C. Whitmore, J. Am. Chem. Soc., 72 (1950) 2732-2735.
37] K.S. Kim, M.A. Barteau, J. Catal., 125 (1990) 353-375.
38] R. Pestman, A. van Duijne, J.A.Z. Pieterse, V. Ponec, J. Mol. Catal. A: Chem., 103
(1995) 175-180.
[
[
[
39] M. Gliński, J. Kijeński, A. Jakubowski, Appl. Catal. A: Gen., 128 (1995) 209-217.
40] M. Gliński, J. Kijeński, Appl. Catal. A: Gen., 190 (2000) 87-91.
41] M. Gliński, G. Zalewski, E. Burno, A. Jerzak, Appl. Catal. A: Gen., 470 (2014) 278-
2
84.
[
[
42] R. Martinez, M.C. Huff, M.A. Barteau, J. Catal., 222 (2004) 404-409.
43] F. Gonzalez, G. Munuera, J.A. Prieto, J. Chem. Soc., Faraday Trans., 74 (1978)
1
[
517-1529.
44] M. Kamitsou, G.D. Panagiotou, K.S. Triantafyllidis, K. Bourikas, A. Lycourghiotis,
C. Kordulis, Appl. Catal. A: Gen., 474 (2014) 224-229.
[
[
[
[
45] S. Luo, J.L. Falconer, J. Catal., 185 (1999) 393-407.
46] J.E. Rekoske, M.A. Barteau, Langmuir, 15 (1999) 2061-2070.
47] J.E. Rekoske, M.A. Barteau, Ind. Eng. Chem. Res., 50 (2011) 41-51.
48] A. Corma, B. Oliver-Tomas, M. Renz, I.L. Simakova, J. Mol. Catal. A: Chem., 388-
389 (2014) 116-122.

Figure Captions
2
0
25
30
35
40
45
50
55
60
65
70
75
80
2
q/degree
2
Figure 1. X-ray diffraction pattern of TiO .

1
50 ºC
50 ºC
2
3
50 ºC
1
800
1750
1700
1650
1600
1550
1500
1450
1400
wavenumber/cm–1
Figure 2. IR spectra of the pyridine adsorption onto TiO
adsorbed at room temperature and desorbed at increasing temperatures (150, 250 and 350 ºC) for
0 min at each temperature.
2
. Experimental conditions: Pyridine was
6

1
00
90
80
70
60
50
40
30
20
10
0
3
00
325
350
375
400
425
450
T/ºC
Figure 3. Transformation of decanoic acid over TiO
2
(continuous line) and ZrO
2
(dashed line) in the
temperature range from 300 to 450 ºC. 
-nonadecanone yield,  nonadecene
–
1
–1
2
yield (only TiO ). Reaction conditions: 2.50 g of metal oxide, WHSV = 3.17 h , F = 7.92 g h , 50 mL
–
1
2
min of N .

1
00
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
100 120 140 160 180
t/min
2
Figure 4. Transformation of 2-undecanone under ketonization conditions in the presence of TiO .
–
1
–1
Reaction conditions: 2.50 g of TiO
2
, F = 7.47 g h , WHSV = 2.99 h , 425 ºC, carrier gas N
2
or H
2
50
–
1
mL min .  conversion (N
2
), ▲ conversion (H ),  undecene yield (H
2
2
2
).

1
00
90
80
70
60
50
40
30
20
10
0
0
100
200
300
400
500
600
700
800
t/min
Figure 5. Transformation of decanoic acid over TiO with time on stream at 400 ºC. Reaction
2
–
1
–1 –1
conditions: 2.50 g of TiO
catalyst was re-activated by in-situ calcination (air 50 mL min , 450 ºC, 2 h) and the reaction
continued. ▲ conversion,  -nonadecanone yield.
2
, F = 7.92 g h , WHSV = 3.17 h , N 50 mL min . After 340 min, the
2
–1

1
0
9
8
7
6
5
4
3
2
1
0
C10-ketone (unsat, aldol cond.)
C10–C12 phenols
C7-ketones
C5-olefins
cyclic condensation products
higher weight products (>C15)
C10-ketone (sat)
C15-aromatics
C10–C12 aromatics
C6-olefins
C8-ketones
C8-olefins
C7-olefins
3
0
40
50
60
70
Conversion/%
2
Figure 6. Yield over conversion for the transformation of 2-pentanone over TiO . Reaction
–
–
1
–1
conditions: 0.50 g of TiO
.
2
, F = 7.28 to 72.8 g h , WHSV from 14.6 to 145 h , 425 ºC, N
2
100 mL min
1

4
4
3
3
2
2
1
1
50
00
50
00
50
00
50
00
Diesel
Kerosene
5
0
0
0
10 20 30 40 50 60 70 80 90 100
wt.% distilled
Figure 7. Simulated distillation of the product mixtures of the reactions of pentanoic acid over TiO
continuous line) and ZrO (dashed line) after subsequent hydrodeoxygenation. Horizontal lines
indicate temperature limits for the boiling point ranges of kerosene (150 to 275 ºC) and diesel fuel
171 to 359 ºC). Reaction conditions: 2.50 g of metal oxide and 1.5 g of 2% Pt/Al were employed
–1
2
(
2
(
2 3
O
–
1
–1
2 2 2
as catalysts, F = 8.37 g h , WHSV = 3.35 h (with respect to TiO or ZrO ), 400 ºC, H 100 mL min
–1
2 2
(TiO ) and 470 mL min (ZrO ).

1
00
90
80
70
60
50
40
30
20
10
0
Gasoline
Kerosene
Diesel
Heavy diesel
Figure 8. Yields of different fractions of fuels when combining either TiO
hydrodeoxygenation treatment (Pt/Al ). Maximum values are stated for kerosene and diesel fuel
although both yield can not be achieved at the same time.
2 2
(blue) or ZrO (red) with a
2
O
3

1
00
90
80
70
60
50
40
30
20
10
0
0
100
200
300
400
500
600
700
t/min
Figure 9. Composition of the liquid product when pentanoic acid was treated in a two-bed process
with TiO and Pt/Al under hydrogen atmosphere. Alkanes and cycloalkanes , X aromatics, ▲
oxygenated. Reaction conditions: 2.50 g of TiO
2
2 3
O
–
1
2
and 1.5 g of 2% Pt/Al O , F = 8.37 g h , WHSV =
2 3
–
1
–1
3
2 2
.35 h (with respect to TiO ), 400 ºC, H 100 mL min .

2
Scheme 1. Hydrogen transfer from limonene or -terpinene to nitrobenzene catalyzed by TiO at
–1
–1
–1
4
2 2
25 ºC. Reaction conditions: 2.50 g of TiO , F = 8.19 g h , WHSV = 3.28 h , N 100 mL min .

Scheme 2. Proposed reaction pathways for the formation of the products observed when 2-
pentanone was contacted with TiO . Yields are displayed in Figure 6.
2