Solvent effects in hydrodeoxygenation of furfural-acetone aldol condensation products over Pt/TiO2 catalyst

Article abstract of DOI:10.1016/j.apcata.2016.11.023

The solvent effects on hydrodeoxygenation (HDO) of 4-(2-furyl)-3-buten-2-one (F-Ac) over Pt/TiO₂ catalyst were investigated at T = 200 °C and P(H₂) = 50 bar. The initial reactant is the main product of aldol condensation between furfural and acetone, which constitutes a promising route for the production of bio-based chemicals and fuels. A sequence of experiments was performed using a selection of polar solvents with different chemical natures: protic (methanol, ethanol, 1-propanol, 2-propanol, 1-pentanol) and aprotic (acetone, tetrahydrofuran (THF), n,n-dimethylformamide (DMF)). In case of protic solvents, a good correlation was found between the polarity parameters and conversion. Consequently, the highest hydrogenation rate was observed when 2-propanol was used as a solvent. In contrast, the hydrogenation activity in presence of aprotic solvents was related rather to solvent-catalyst interactions. Thus, the initial hydrogenation rate declined in order Acetone > THF > DMF, i.e. in accordance with the increase in the nucleophilic donor number and solvent desorption energy. Regarding the product distribution, a complex mixture of intermediates was obtained, owing to the successive hydrogenation (aliphatic C[dbnd]C, furanic C[dbnd]C and ketonic C[dbnd]O bonds), ring opening (via C[sbnd]O hydrogenolysis) and deoxygenation reactions. Based on the proposed reaction scheme for the conversion of F-Ac into octane, the influence of the studied solvents over the cascade catalytic conversion is discussed. A significant formation of cyclic saturated compounds such as 2-propyl-tetrahydropyran and 2-methyl-1,6-dioxaspiro[4,4]nonane took place via undesirable side reactions of cyclization and isomerization. The best catalytic performance was found when using acetone and 2-propanol as solvents, achieving significant yields of 4-(2-tetrahydrofuryl)-butan-2-ol (28.5–40.4%) and linear alcohols (6.3–10.4%). The better performance of these solvents may be associated with a lower activation energy barrier for key intermediate products, due to their moderate interaction with the reactant and the catalyst. In case of methanol and DMF, undesired reactions between the reactant and the solvent took place, leading to a lower selectivity towards the targeted hydrodeoxygenated products.

Full text of DOI:10.1016/j.apcata.2016.11.023

Accepted Manuscript
Title: Solvent effects in hydrodeoxygenation of
furfural-acetone ALDOL condensation products over Pt/TiO2
catalyst
´
ˇ
ˇ
Author: Ruben Ramos Zdenek Tisler Oleg Kikhtyanin David
ˇ
Kubicka
PII:
S0926-860X(16)30562-2
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.11.023
APCATA 16071
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
26-9-2016
14-11-2016
15-11-2016
´
ˇ
ˇ
Please cite this article as: Ruben Ramos, Zdenek Tisler, Oleg Kikhtyanin,
ˇ
David Kubicka, Solvent effects in hydrodeoxygenation of furfural-acetone ALDOL
condensation products over Pt/TiO2 catalyst, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2016.11.023
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.

SOLVENT EFFECTS IN HYDRODEOXYGENATION OF FURFURAL-
ACETONE ALDOL CONDENSATION PRODUCTS OVER Pt/TiO2
CATALYST
Rubén Ramos1,*, Zdeněk Tišler1, Oleg Kikhtyanin1, David Kubička1,2
1Research Institute of Inorganic Chemistry, RENTECH-UniCRE, Chempark Litvínov, Záluží-Litvínov,
436 70, Czech Republic
2Technopark Kralupy, University of Chemistry and Technology Prague, Žižkova 7, 27801 Kralupy nad
Vltavou, Czech Republic
*Corresponding author: Ruben Ramos Velarde
Present address: Department of Chemistry, University of Liverpool, Crown Street, Liverpool
L69 7ZD, U.K.
e-mail: R.Ramos-Velarde@liverpool.ac.uk
Graphical abstract
1

Solvent effects
Solvent effects:
-Methanol -Acetone
-Ethanol
-1-propanol
-2-propanol
-Tetrahydrofuran
-Dimethylformamide
-1-pentanol
O
Hydrodeoxygenation
(P= 50 bar, T= 200 ºC)
O
O
CH3
Furfural-acetone
condensation products
Pt/TiO2
Highlights
 The hydrogenation rate and further transformations of furfural acetone condensate
product are discussed considering solvent effects.
 Important differences were observed among the investigated solvents, achieving the
best HDO activities when using 2-propanol and acetone.
 This work provides a better understanding of the role of a solvent, contributing to a
rational selection that would improve the overall conversion efficiency.
Abstract
The solvent effects on hydrodeoxygenation (HDO) of 4-(2-furyl)-3-buten-2-one (F-Ac) over
Pt/TiO2 catalyst were investigated at T=200 °C and P(H2)=50 bar. The initial reactant is the
main product of aldol condensation between furfural and acetone, which constitutes a
promising route for the production of bio-based chemicals and fuels. A sequence of
experiments was performed using a selection of polar solvents with different chemical
natures: protic (methanol, ethanol, 1-propanol, 2-propanol, 1-pentanol) and aprotic (acetone,
tetrahydrofuran (THF), n,n-dimethylformamide (DMF)). In case of protic solvents, a good
correlation was found between the polarity parameters and conversion. Consequently, the
highest hydrogenation rate was observed when 2-propanol was used as a solvent. In contrast,
the hydrogenation activity in presence of aprotic solvents was related rather to solvent-
catalyst interactions. Thus, the initial hydrogenation rate declined in order
2

Acetone>THF>DMF, i.e. in accordance with the increase in the nucleophilic donor number
and solvent desorption energy. Regarding the product distribution, a complex mixture of
intermediates was obtained, owing to the successive hydrogenation (aliphatic C=C, furanic
C=C and ketonic C=O bonds), ring opening (via C-O hydrogenolysis) and deoxygenation
reactions. Based on the proposed reaction scheme for the conversion of F-Ac into octane, the
influence of the studied solvents over the cascade catalytic conversion is discussed. A
significant formation of cyclic saturated compounds such as 2-propyl-tetrahydropyran and 2-
methyl-1,6-dioxaspiro[4,4]nonane took place via undesirable side reactions of cyclization and
isomerization. The best catalytic performance was found when using acetone and 2-propanol
as solvents, achieving significant yields of 4-(2-tetrahydrofuryl)-butan-2-ol (28.5-40.4 %) and
linear alcohols (6.3-10.4 %). The better performance of these solvents may be associated with
a lower activation energy barrier for key intermediate products, due to their moderate
interaction with the reactant and the catalyst. In case of methanol and DMF, undesired
reactions between the reactant and the solvent took place, leading to a lower selectivity
towards the targeted hydrodeoxygenated products.
Keywords: Furfural, aldol condensation, hydrodeoxygenation, solvent effects, platinum
titania catalyst.
1. Introduction
Biomass is a key renewable resource with the potential to satisfy at least partially the growing
energy demand and to provide alternative technologies to petroleum-based ones.
Hemicellulose is one of the major components (25-35 wt%) of non-food lignocellulose
biomass, which can be transformed into C5 and C6 sugars monomers via acid-catalyzed
hydrolysis [1,2]. This sugar fraction can be further processed to furanic compounds (via acid-
catalyzed dehydration), leading to the production of lignocellulosic platform molecules (e.g.
furfural and 5-hydroxymethylfurfural) that could play the role of basic building blocks in
future biorefineries [3,4]. Therefore, in order to produce valuable green chemicals and fuels
3

from sustainable sources, the catalytic transformation of furan derivatives is regaining the
attention of different research groups in recent years [5,6].
Among the different conversion strategies, aldol condensation allows increasing carbon chain
length of furanic compounds by forming larger bio-based molecules, more appropriate for
upgrading to gasoline and diesel fuels. Aldol condensation is a well-known organic synthesis,
which generally takes place between carbonyl groups at low temperatures (25-70 ºC) and
atmospheric pressure. This route has been successfully used for the condensation of furfural
(C5) with acetone in presence of solid basic catalysts, yielding mostly 4-(2-furyl)-3-buten-2-
one (F-Ac) and 1,5-difuryl-1,4-pentadien-3-one (F-Ac-F) [7,8]. However, compared with
most petrochemical products, furfural-acetone condensation products are multifunctionalized
oxygenated compounds which require challenging chemical transformations i.e.
hydrogenation (C=C aliphatic, C=O ketone and C=C furanic ring bonds), ring opening (C-O
hydrogenolysis) and deoxygenation (via dehydration or C-OH hydrogenolysis) in order to
yield liquid transportation fuels. Hence, the overall conversion of these adducts into alkanes
may be performed by hydrodeoxygenation (HDO), generally at mild temperatures (150-300
ºC) and high hydrogen pressures (20-100 bar) [9,10].
The presence of heterogeneous catalysts in the HDO processing of furfural-acetone
condensation products favours the conversion rates (thermodynamic and kinetic
considerations) and allows tailoring the product distribution through the selection of the
catalyst features. In a previous work, we have investigated the influence of different supported
Pt catalysts on the HDO of furfural-acetone condensation products [11]. Similarly, different
studies from the literature have reported the performance of metal-based catalysts on this
process, exploring the effects of different parameters such as; nature of the metal phase [10],
catalyst morphology [12,13,14], reaction system [13,15,16] and operation conditions
[9,17,18]. Most of these studies have been carried out in the liquid phase, using different
kinds of solvents to dissolve the solid reactants and to dissipate the heat generated from
exothermic reactions. Thus, Dumesic and co-workers investigated the hydrogenation of
furfural-acetone adducts employing Pd/MgO-ZrO2 as catalyst in aqueous environment
[15,18]. From the same research group, another work investigated the activity of Pt/SiO2-
Al2O3 in dehydration of these compounds using tetrahydrofuran (THF) as solvent [16]. In the
same way, Lu et al. [14,17,19] carried out a series of experiments combining different
solvents such as isopropanol, ethanol, THF and cyclohexane. Similarly, Gordon et. al. [9]
compared the results obtained using acetic acid and methanol as solvents in deoxygenation
4

reactions under mild operating conditions. Additionally, HDO of furfural-acetone
condensation products has been carried out in hexane [10,12] and, also using ionic liquids
[20]. Nevertheless, due to the lack of systematic experimental data, the influence of the
reaction media (solvents) on the activity and selectivity in the HDO of furfural-acetone
condensation products has been scarcely discussed.
According to preceding works [21-24], it is expected that solvent effects have a strong
influence on hydrogenation reactions, which constitute an essential initial step in the HDO
processing. A general reaction scheme for F-Ac hydrogenation is shown in Figure 1. The
primary C=C aliphatic hydrogenation (4-(2-furyl)-butan-2-one; A) is followed by further
hydrogenation steps of C=C furanic ring bonds (4-(2-tetrahydrofuryl)-butan-2-one; B) or C=O
ketone bond (4-(2-furyl)-butan-2-ol; C). On the pathway to the completely hydrogenated
compound (4-(2-tetrahydrofuryl)-butan-2-ol; E), the partial hydrogenation of the latter yields
a reactive intermediate (4-(2-dihydrofuryl)-butan-2-ol; D), which may undergo different
reactions [11].
In spite of the progress in understanding solvent effects on hydrogenation reactions, much
work remains to be done to unravel the influence of solvents in multi-step catalytic
transformations, such as the HDO of F-Ac towards valuable alkanes, which further comprises:
ring opening reactions (on metal and possibly acid sites), dehydration/hydrogenolysis (most
likely on Brønsted acid sites), as well as different side reactions. Likewise, the regular use of
bifunctional catalysts in such processes may increase the complexity of solvent effects owing
to the likely interactions between the metal phases and the solvent [25].
In the present contribution, we aim to investigate the solvent effects on the HDO of 4-(2-
furyl)-3-buten-2-one (F-Ac) in terms of conversion rates and product distribution. Pt/TiO2
was selected as a reference catalyst due to the well-known hydrodeoxygenation activity of this
metal phase and the enhanced accessibility provided by the titania support. A sequence of
experiments was performed using a selection of polar solvents with different chemical nature:
protic (methanol, ethanol, 1-propanol, 2-propanol, 1-pentanol) and aprotic (acetone,
tetrahydrofuran, n,n-dimethylformamide). The initial hydrogenation rate and further
transformations through furan ring opening and deoxygenation steps, are discussed
considering the solvent-reactant and the solvent-catalyst interactions. To our knowledge, such
systematic investigation of solvent effects on HDO of furfural-acetone condensation products
has not been described previously. Therefore, this work may provide a better understanding of
5

the role of a solvent in this catalytic conversion, contributing to a rational solvent selection
that would improve the overall process efficiency.
2. Experimental
2.1. Chemicals
Methanol (Sigma Aldrich, >99.9 %), Ethanol (Sigma Aldrich, >99.8 %), 1-propanol (Sigma
Aldrich, >99.5 %), 2-propanol (Lach-Ner, >99.7 %), 1-pentanol (Sigma Aldrich, >99 %),
acetone (Sigma Aldrich, >99.5 %), tetrahydrofuran (Sigma Aldrich, >99.9 %) and n,n-
dimethylformamide (Sigma Aldrich, >99.8 %) were used without further treatment. The TiO2
support and tetraammine-platinum(II) nitrate were purchased from Eurosupport
Manufacturing Czechia and Sigma Aldrich, respectively, and used as received. Hydrogen was
supplied from an external reservoir (Linde, >99 %).
2.2. Catalyst preparation
The catalyst, Pt/TiO2, was prepared by incipient wetness impregnation using tetraammine-
platinum(II) nitrate (Pt(NH3)4(NO3)2) to obtain a nominal metal loading of 2 wt% (referred to
mass catalyst). The deposition of Pt was carried out by adding dropwise an aqueous solution
of the precursor to the support at ambient temperature. After metal incorporation, the catalyst
was dried at 120 ºC overnight and calcined at 500 ºC in air for 6 h (heating rate 1 ºC·min-1).
2.3. Catalyst characterization
The properties of Pt/TiO2 were analyzed by different physicochemical characterization
methods. The amount of incorporated Pt was determined by X-ray Fluorescence Spectroscopy
(XRF) using a Philips PW 1404 UniQuant apparatus. Textural properties were determined
from N2 physisorption isotherms at 77 K obtained by using a Quantachrome AUTOSORB
unit. Prior to the analyses, the sample was outgassed at 250 ºC for 3 h flowing N2. BET
equation and BJH method were used to calculate the specific surface area and the average
pore size, respectively.
Acid and base sites were evaluated by temperature programmed desorption (TPD) of
ammonia and carbon dioxide, respectively, employing a Micromeritics Autochem 2950
equipment. Prior to the TPD analyses, the sample (0.1 g) was reduced in situ under 10 vol%
H2/Ar (25 cm3·min-1) at 350 ºC for 60 min. Then, the catalyst was outgassed in flowing
6

helium (25 cm3·min-1) with a heating ramp of 10 °C·min-1 up to 500 °C. The NH3 TPD
measurement was started by cooling the sample down to 180 ºC and saturating it with
ammonia (25 cm3·min-1 of 10 vol% NH3/He) for 30 min. Subsequently, the physically
adsorbed NH3 was removed by flowing He (25 cm3·min-1) for 60 min. Finally, TPD was
performed heating the Pt/TiO2 catalyst in flowing He (25 cm3·min-1) with a rate of 15 °C·min-
1
up to 500 °C, and maintaining this temperature for 30 min. The CO2 TPD experiment was
initiated by cooling the sample to 50 °C followed by saturating it with a flow of 10 vol%
CO2/He (50 cm3·min-1) at 50 ºC for 30 min. Afterward, the physisorbed carbon dioxide was
removed by flowing He (25 cm3·min-1) at 50 ºC for 60 min. Finally, the chemically adsorbed
CO2 was desorbed by heating up the sample to 900 ºC (heating rate of 15 °C·min-1) in flowing
He (25 cm3·min-1) and maintaining this temperature for 30 min.
The metal dispersion of Pt in the Pt/TiO2 catalyst was determined by CO pulse chemisorption
using a Micromeritics Autochem 2950 equipment. After the reduction treatment, the sample
was degassed under He flow at 500 ºC for 1 h and then cooled down to 50 ºC. Subsequently, a
sequence of 5 vol% CO/He pulses (500 μL) was introduced in a flow of He (50 cm3·min-1)
and passed through the sample until the areas of the successive peaks remained constant. The
relation between the areas of the different chemisorbed CO pulses allowed evaluating the
metal dispersion as a function of metal loading and defined adsorption stoichiometry factor (1
CO/Pt atom).
Solvent interactions with Pt/TiO2 catalyst were investigated by temperature programmed
desorption (TPD) of the pre-adsorbed solvent. A sample of reduced Pt/TiO2 catalysts was
saturated at 50 ºC by pulse dosing of a gaseous stream of He (20 cm3·min-1) bubbled through
the solvent. Then, the adsorbed solvent was removed under a flow of He (25 cm3·min-1) by
increasing the temperature from 50 to 500 ºC. Different heating rates (10, 15 and 20 ºC·min-1)
were used during the TPD resulting in a variation in the temperature of the desorption peak
maximum. The desorption enthalpy (Ed) was then calculated from the slope of the linear
relationship between the heating rate and the temperature of the desorption peak maximum,
according to the method reported by Cvetanovic and Amenomiya [26].
2.4. Catalytic experiments
The catalytic conversion of 4-(2-furyl)-3-buten-2-one (Alfa Caesar, 98%) was studied in a
steel 300 ml batch reactor (Parr Instrument Co.). Temperature, pressure and stirring speed
were monitored and logged through a Parr 4848B acquisition interface. Prior to the
experiments, the catalyst (0.2 g) was reduced in situ by H2 flow (80 cm3·min-1) at 350 ºC for
7

90 min. After cooling down, a prepared solution of 1 g of 4-(2-furyl)-3-buten-2-one (F-Ac)
and 150 ml of the solvent was loaded into the reactor. Then, the system was heated up to 200
ºC and pressurized with H2 to 50 bar, under a slow stirring speed (100 rpm). The initial
reaction time was taken when the reactor reached the desired temperature and the stirring
speed was increased to 1000 rpm. The hydrogen consumed during the reaction (480 min) was
replenished from an external source to maintain a constant reaction pressure. Samples of
products were collected at different reaction times in order to study the progress of the
reaction. The identification of the resulting compounds was done by means of GC-MS
analyses using a Thermo Scientific ITQ 1100 apparatus. The quantification of the product
mixtures were analyzed by an Agilent 7890A GC unit (FID detectors) equipped with a HP-5
capillary column (30 m, 0.32 ID, 0.25 μm). To this end, standard reference compound like 4-
(2-furyl)-3-buten-2-one (Sigma Aldrich, >95 %) was used for calibration measurements,
assuming the same GC sensitivity for the rest of C8 furan/tetrahydrofuran derived compounds.
F-Ac conversion at time t (Xt) was calculated using Equation (1):
[C
]
-[C
]
F-Ac
F-Ac
t=0
X푡 =
t ∙100
(1)
[C
]
F-Ac
t=0
where [CF-Ac t=0 and [CF-Ac]t are the initial molar concentration and the molar concentration at
]
time t of F-Ac in solution, respectively. The yield of the specie i at time t ([Yi]t) was given by
Equation (2):
[C ] −[C ]
t
t=0 ∙ 100
(2)
푖
i
[Y푖]t =
[C
]
F−Ac t=0
where [Ci]t=0 and [Ci]t are the initial molar concentration and the molar concentration at time t
of the specie i, respectively.
3. Results and discussion
3.1. Catalyst characterization
The main physicochemical properties of Pt/TiO2, characterized by the instrumental techniques
described in section 2.3, are reported in Table 1.
The final Pt content in the catalysts was determined by XRF at to be 1.8 wt%, indicating an
almost complete incorporation of the metal phase on the support by the used impregnation
method. The nitrogen adsorption-desorption isotherm of Pt/TiO2 (Figure S1 in the Supporting
8

Information) exhibited a profile characteristic of mesoporous materials, with a noticeable
adsorption at high relative pressures (P/P0>0.8). Based on the experimental data provided by
this isotherm, BET surface area (95 m2·g-1) and average pore size (30.4 nm) were determined.
The obtained textural properties provide an appropriate specific surface area for the dispersion
of the metal phase, as well as an extensive pore size, capable of accommodating the F-Ac
molecules into the mesopore system without steric limitations.
Figure 2 illustrates the temperature programmed desorption curves obtained for the desorption
of NH3 (Figure 2a) and CO2 (Figure 2b). The integration of the NH3 TPD curve revealed
presence of acidic sites (0.134 meq NH3·g-1) of weak strength (maximum desorption at 234
ºC). On the contrary, the total amount of basic sites, calculated from the CO2 TPD curve, was
considerably higher (0.416 meq CO2·g-1). This value was estimated without considering the
CO2 desorbed above 500 ºC, since, according to some authors [27,28] and, also in agreement
to the performed blank test, the TPD signal obtained at those temperatures is associated with
residual carbonates not removed during the calcination or outgassing steps (500 ºC). Thus,
only the chemisorbed CO2 in form of bicarbonate anions of weak basic strength (90-110 ºC)
and bidentate species (190-220 ºC) of medium strength, were taken into account for the
quantification of the catalyst basic properties.
Finally, Pt dispersion (15.6%) and average Pt particle size (7.2 nm) were estimated by CO
pulse chemisorption measurements (Figure S2). These features may be crucial for catalyst
performance, since the activation of the intermediate compounds over metal centers is
strongly influenced by the dispersion degree and particle size of the metal phase [29,30].
3.2. Conversion and initial hydrogenation rate
Following the experimental procedure defined in the section 2.4, the effects arising from the
use of different solvents in the HDO of F-Ac were evaluated in terms of the reactant
conversion and the initial hydrogenation rate. Control experiments without catalysts and with
non-impregnated TiO2 were performed in acetone, showing a negligible conversion (< 5 %)
and suggesting that it is the presence of Pt0 particles which allows for H2 uptake on the
surface and catalyses F-Ac hydrogenation. Likewise, catalyst particle size (<30 μm) and
vigorous stirring (1000 rpm) were chosen to avoid mass transfer limitations and internal
9

diffusion resistance. The first hydrogenation step usually corresponds to the fast
hydrogenation of the aliphatic C=C bond, which is favoured by thermodynamic conditions
and the absence of steric limitations (unlike the case of the furanic ring bonds) [12,15,16]. An
overview of the conversion results obtained from the performed experiments is represented in
Figure 3, where the studied polar solvents were grouped according to their chemical nature
into (a) protic and (b) aprotic.
Figure 3a shows a clear difference between the conversion rates of experiments carried out in
methanol and 1-pentanol, and in the rest of the studied alcohols (ethanol, 1-propanol and 2-
propanol). The latter group of solvents showed higher hydrogenation rates, achieving a full
conversion of F-Ac already after 180 min of reaction. However, the hydrogenation rates of F-
Ac when using 1-pentanol and methanol were substantially lower, leading to an incomplete
conversion even after 480 min of reaction (97 and 92.5 %, respectively). Similar behaviour
was observed in Figure 3b for aprotic solvents, where in acetone and tetrahydrofuran (THF) a
complete conversion was reached at 240 min, whereas, due to a considerably lower
hydrogenation rate, the final conversion (480 min) in n,n-dimethylformamide (DMF) was
around 93 %.
The factors responsible for these variations may in principle originate from different
phenomena, such as: interaction of the solvent with the reactant, competitive adsorption on
the catalyst surface, or reactant and hydrogen solubilities. The effects related to the solubility
of the reactant can be considered negligible due to the high reaction temperature (200 °C) and
the excess of solvent used (1g of F-Ac per 150 ml solvent), ensuring a complete solvation of
the reactant in all performed experiments. On the other hand, although in some cases solvent
effects have been attributed to different hydrogen solubility measurements [31], most of the
authors agree that a clear correlation cannot be established between hydrogenation activity
and hydrogen solubility [22,24,32,33]. In this respect, it is worth to note that accurate data of
hydrogen solubility under the operating conditions are difficult to obtain, and a limited
information is available in the literature to correlate the obtained results for the whole set of
studied solvents.
Concerning the influence of the solvent-reactant interactions, Table 2 summarizes polarity
parameters of each solvent and the initial hydrogenation rates exhibited in them. The dipole
moment (μ) provides information about the structure of the molecule, whereas dielectric
constant (εr) is a useful macroscopic measurement of the solvent polarity. The initial
10

hydrogenation rates were determined from the change in the molar concentration of F-Ac
during the first 30 min.
The experimental results presented in Table 2 do not show any correlation between the initial
hydrogenation rates and the polarity parameters when the complete series of the tested
solvents is considered. For instance, the highest hydrogenation activity was observed in
acetone and 2-propanol in spite of the significant differences in their dipole moment (1.66 and
2.88 D, respectively). In contrast, a satisfactory correlation was found between the solvent
polarity and the initial hydrogenation rate when considering only protic solvents. In general,
the initial hydrogenation rate of F-Ac decreased as the polarity of the alcohol solvents
increased. Thus, the variation in the dielectric constant values (εr) fitted reasonably well with
the estimated initial hydrogenation rates. A linear correlation was found (Figure 4) with the
exception of 1-pentanol, which despite having a relatively low value of εr (13.9), presented a
low initial hydrogenation rate (2.44·10-4 mmol·gcat-1·min-1). The trend showed by most of the
studied alcohols may be explained considering the hydrogen bond donors nature of protic
solvents. This fact may be responsible for stronger bonding between the hydroxyl group of the
alcohols and the reactant (F-Ac). Thus, as the polarity increases the solvent-reactant
interaction becomes more important, improving the solvation effects and hindering the
adsorption of F-Ac on the catalytic surface [24,33]. The unusually lower hydrogenation
activity in 1-pentanol corresponds well with similar results obtained in hydrogenation of 1-
phenyl-1,2-propanedione [22]. This behaviour may be related with the higher bulkiness of 1-
pentanol, which can hinder the adsorption of the molecules, leading to a lower reaction rate
[34]. Nevertheless, in this particular case probably other effects such as hydrogen solubility or
adsorption strength may be involved as well.
Regarding the polar aprotic solvents, the initial hydrogenation rate pattern followed the trend:
Acetone>THF>DMF, differing from the polarity correlation. Thus, no relationship between
the initial hydrogenation rate and polarity parameters was found for the series of polar aprotic
solvents. In particular, the highest initial hydrogenation rate was observed in acetone, which
has an intermediate value of both μ and εr. Therefore, even though polarity parameters may
play a significant role in the liquid phase hydrogenation of F-Ac, other physicochemical
phenomena must be considered in order to explain the observed behaviour with polar aprotic
solvents. In this respect, solvent adsorption on Pt/TiO2 surface may have a strong influence on
the catalytic activity. Despite the polar aprotic solvents are unable to act as hydrogen donors,
they are good electron pair donor solvents. Thus, in case of working with bifunctional
11

catalysts, the formation of strong complexes EPD/EPA (electron pair donor/electron pair
acceptor) may occur through an electron transfer between the aprotic solvent (n-EPD) and the
metal phase (υ-EPA) [35]. Consequently, a competitive solvent-reactant adsorption may arise,
blocking partially the access of F-Ac molecules to the active sites and hindering the reaction
progress. Figure 5a shows the hydrogenation rates as a function of the nucleophilic donor
number (DN), which represents a quantitative measure of the electron pair donor ability [36].
The initial hydrogenation rates showed a continuous decrease as DN of the polar aprotic
solvents increased. For example, the best initial hydrogenation rate exhibited in acetone
corresponded with the lowest DN (71 kJ·mol-1), whereas the limited hydrogenation activity
showed by DMF is related to the highest DN (111.2 kJ·mol-1). These results may also be
described as a coordinative Lewis base/acid interaction between the solvent (Lewis base) and
the catalyst. Consequently, considering the accepted Lewis acid character of the Pt/TiO2
catalyst (Figure 2a), a series of chemisorption experiments was performed in order to confirm
the strength of the solvent-catalyst interactions. Thus, following the method reported by
Amenomiya et. al [26], the desorption enthalpy was calculated for each solvent and plotted as
a function of the initial hydrogenation rate (Figure 5b). The maximum desorption enthalpy
was found for DMF, showing a value of 263 kJ·mol-1. On the contrary, acetone and THF
presented considerably lower desorption enthalpies of 108 and 150 kJ·mol-1, respectively.
These results were consistent with DN data for polar aprotic solvents [36], showing a close
correlation, and consequently, a very similar decreasing tendency of the initial hydrogenation
rates (r0) as energy of desorption (Ed) increased (Figure 5a and b). These facts suggest that
DMF interacts much stronger than THF and acetone with Pt/TiO2 surface, promoting a
competitive solvent/reactant adsorption on the catalyst active sites. Similarly, the higher
hydrogenation rate observed in acetone would be related to the weaker adsorption of this
solvent on Pt/TiO2. In summary, we can conclude that solvent adsorption strength played an
essential role in determining the F-Ac initial hydrogenation rate when polar aprotic solvents
were used. On contrary, in the case of polar protic solvents the desorption enthalpies are not a
so useful parameter due to the existence of other larger interactions like hydrogen bonding
[35].
Obviously, it would have been interesting to complete the study using some apolar solvents,
which are expected to show very weak interactions with both the reactant and the catalyst.
However, from our own experience, 4-(2-furyl)-3-buten-2-one is nearly immiscible in both
12

apolar solvents and in water, limiting the F-Ac concentration in the vicinity of the accessible
active sites and inducing mass transfer limitations.
3.3. Hydrodeoxygenation products
The experimental data allowed elucidating that the transformation of F-Ac on bifunctional
catalysts started with the initial hydrogenation of aliphatic C=C bond followed by further
hydrogenation steps catalyzed by metals and continued by subsequent ring opening and
deoxygenation reactions on acid and metal sites, respectively. According to a previous study
[37], the further hydrogenation of the carbonyl group is more demanding since it is limited by
the presence of the furan ring, which causes significant steric hindrances. In this respect, the
hydrogenation of the furan ring is strongly influenced by the reactant adsorption mode on the
metal sites, which is closely related to the metal dispersion and metal particle size [38].
Regarding the ring-opening and deoxygenation reactions, they typically require the presence
of strong acid sites that facilitate C-O hydrogenolysis and dehydrations steps [39]. Figure 6
displays an example of the evolution of the yields of the main products from the HDO of F-
Ac when using acetone as a solvent. Based on the information provided by these kinetic
curves, a reaction scheme was proposed in a previous work, demonstrating that a very
complex mixture of products related through different catalytic transformations is formed
[11].
Figure 6 clearly shows that the first hydrogenation step of C=C aliphatic bond, forming 4-(2-
furyl)-butan-2-one (A), is followed by formation of further hydrogenated products such as: 4-
(2-tetrahydrofuryl)-butan-2-one (B), 4-(2-furyl)-butan-2-ol (C) and 4-(2-dihydrofuryl)-butan-
2-ol (D). The presence of a maximum in the yield of those compounds reveals their
intermediate character on the pathway towards the fully hydrogenated molecule 4-(2-
tetrahydrofuryl)-butan-2-ol (E), which shows a continuous increase as the reaction proceeds.
A similar trend is observed in the case of 2-butyl-dihydrofuran (G) and 2-butyl-
tetedrahydrofuran (H), which are formed most probably by dehydration steps and subsequent
ring hydrogenation of the primary products. The continuously increasing yield of these
compounds (E, G and H) indicates a limitation in the subsequent ring opening steps on the
conversion pathway of F-Ac to linear alcohols over Pt/TiO2 catalyst. This fact is likely
associated with the absence of stronger acid sites (mainly Brønsted), which, in combination
with metal sites, are mostly responsible for ring-opening reactions typically through C-O
hydrogenolysis. Nevertheless, in case of acetone as a solvent, a significant formation of
13

octanediols and octanols (J) took place (≈10 %), suggesting that ring opening takes place
though to a limited extent.
The use of Pt/TiO2 as a catalyst for the study of the solvent effects is justified by its multi-
functional properties (metal sites and weak acidity/basicity) as well as by the absence of
stronger basic/acid sites, which usually, under the selected operation conditions, promote side
reactions with the solvent [11,40]. This selection was based on our previous work [11], in
which Pt/TiO2 catalyst showed the best conversion rate of furfural-acetone adducts and the
lowest formation of solvent-reactants condensates. Thus, the use of weakly basic/acid support
can be a suitable strategy to modify the catalyst performance, as long as the basicity/acidity is
not strong enough to catalyze undesired reaction steps like aldol condensation, etherification
or acetalization.
The above presented differences in F-Ac conversion rates could be also applied when
discussing the concentration of intermediate compounds derived from the tandem reaction
pathway of the studied HDO conversion [11,12]. Thus, solvent-reactant and solvent-catalyst
interactions may affect the kinetics of each step, leading to major changes as the reaction
proceeds. Figure 7a shows the yields of the main hydrogenated products at the end of the
reaction time (480 min). Thereby, important differences in the product distribution may be
appreciated between the tested solvents. For example, the yield of A varies from 0 % in case
of 2-propanol to 60.5 % for methanol. Similar deviations can be found in the production of B,
C and E showing their maximum yields in ethanol (16 %), THF (47.8 %) and 2-propanol
(40.4 %), respectively. However, these variations may be considerably reduced if we refer to
the results in terms of selectivity. Since the reaction time needed to reach full conversion of
the initial reactant (F-Ac) is very different in each experiment depending on the solvent used,
the yields of the hydrogenated products showed in Figure 7a could be strongly influenced by
the conversion degree of F-Ac. In this respect, Figure 7b represents the product distribution
obtained with each solvent when considering the same conversion degree of F-Ac (i.e. 70 %).
Thus, for instance, the yield of A varies within a much narrower range (50-61 %). Likewise,
the yield variations of B, C, D and E are substantially reduced when the conversion of F-Ac is
the same for each solvent. Hence, the results presented in Figure 7 suggest that the yields of
the products are strongly dependent on the conversion rates showed by each solvent, having a
deeper impact due to the cascade reaction of the studied HDO conversion.
14

Table 3 summarizes the final yields (after 480 min) of the products from HDO of F-Ac in
experiments using different solvents. As discussed above, it is important to mention that these
results are primarily affected by the different hydrogenation rates showed by each solvent.
The results derived from the use of DMF were not included in this table owing to the totally
different product distribution obtained in this solvent media. This behaviour will be discussed
in detail separately hereafter.
The catalytic performance of Pt/TiO2 varied significantly depending on the applied solvent. In
case of methanol, the reaction practically seized after the primary hydrogenation (A=60.5 %)
and did not proceed further. As it has been previously reported [23,41], the smaller and more
reactive is the alcohol the stronger is the hydrogen bonding between the carbonyl oxygen of
the reactant and the hydroxy group of the solvent. Thus, in case of methanol the interaction
was the most pronounced among the used solvents, reducing thus the preference towards C=O
hydrogenation intermediate (C), which constitutes the preferred pathway in the following
hydrogenation steps. The hydrogenation of the C=C furanic bonds (yielding B) is relatively
limited in F-Ac due to the aromatic nature of the furan ring and also owing to the presence of
the C=O bond in the same adsorption plane [42]. Thus, a selective hydrogenation of the C=O
bond (C) took usually preference over the hydrogenation of furanic C=C (B) in most of the
performed experiments. On the other hand, a significant yield (19.3 %) of heavier products
(molecular mass over 146 g·mol-1) were found in case of using methanol. This fact is in
agreement with previous works reporting that shorter chain alcohols favour side reactions like
acetalization [33] or etherification [41,43] between the reactant and the solvent. The formed
acetal and ether intermediates may undergo further conversions at longer reaction times
and/or higher temperatures, producing equally interesting hydrodeoxygenated by-products
[44,45].
Similarly to methanol, the experiment carried out with 1-pentanol showed a high selectivity to
the primary hydrogenated product (A=57 %), although in this case a significant
hydrogenation of the C=O bond (C=28 %) was also observed. This fact can be attributed to
the less effective hydrogen bonding between the solvent OH group and the reactant, due to the
larger molecular size of 1-pentanol. In any case, according to these results, both protic
solvents showed a limited HDO activity due to different causes namely strong reactant-
alcohol interaction (methanol) and higher bulkiness (1-pentanol) which retarded F-Ac
adsorption on the catalyst [34,41].
15

The hydrogenation degree of F-Ac was gradually enhanced when using ethanol, 1-propanol
and 2-propanol, reaching yields of E equal to 7.1, 15.2 and 40.4 % after 480 min,
respectively. The better performance of 2-propanol over the rest of alcohols agrees well with
the higher hydrogenation activity observed when secondary alcohols are used as solvents
instead of primary alcohols [24,46]. This improvement in the hydrogenation ability has been
attributed to an additional reduction mechanism involving H2 transfer from the alcohol
hydroxy group on the metal surface [24] or to the more effective hydrogen donor via inter-
hydride transfer on Lewis acid catalysts [47,48]. The noticeable production of B in ethanol
(16 %) is probably related to the mentioned stronger hydrogen bonding (between the ketone
group of the reactant and the solvent OH group) as the alcohol molecular size is reduced,
preventing thereby the formation of further hydrodeoxygenated intermediates. Likewise, an
important yield of 2-methyl-1,6-dioxaspiro[4,4]nonane (F) was observed, especially with 1-
propanol (19.6 %). The formation of this compound occurs via cyclization of D due to the
interaction between the OH group and the partially hydrogenated furanic ring [49].
Dehydration also took place resulting in the elimination of hydroxyl groups from C, D and E,
and consequently yielding products such as: 2-butyl-dihydrofuran (G), and 2-butyl-
tetrahydrofuran (H) and 2-propyl-tetrahydropyran (I). Similarly to the production of F, the
formation of I is associated with fast hydrogenation and isomerization of the partially
hydrogenated furanic ring (G) [50]. The concentration of I was quite significant in 1-propanol
(17.3 %) and 2-propanol (31.7 %), suggesting
a
pronounced extent of
isomerization/cyclization steps when using these solvents. Nevertheless, both F and I
constitute an important fraction of cyclic saturated products formed by undesirable side
reactions that are difficult to transform into linear hydrocarbons due to the limited ring
opening activity of the Pt/TiO2 catalyst.
In case of aprotic solvents, pronounced differences were found between the product
distribution showed by acetone and THF. In case of THF, along with aliphatic bond
hydrogenation (A=23 %), a prominent yield of C was observed (47.8 %). Equally, in spite of
the low production of partially deoxygenated products (G and H), a remarkable formation of I
was obtained (13.3 %). This fact is likely a consequence of fast isomerization/cyclization
reactions when using THF as a solvent. On the other hand, the HDO of F-Ac in acetone
presents the highest production of both linear alcohols (J=10.3 %) and partially deoxygenated
products (G+H+I=36.2 %). It also needs to be mentioned that a significant extent of acetone
self-condensation was observed, yielding mesityl oxide (0.01 g·cm-3) and methyl isobutyl
16

ketone (0.03 g·cm-3). Moreover, partial hydrogenation of acetone took place (<5%) forming
2-propanol in the reaction media. Both acetone and 2-propanol showed the highest yield of
products found at the end of the HDO cascade reaction network (J, G, H, I,), which is in
agreement with the higher conversion rates observed with these solvents. Therefore, the better
performance of these solvents may be associated with a lower activation energy barrier for the
key intermediate products, leading to enhanced reactions rates. Nevertheless, unlike in 2-
propanol in which only octanediols were present, the use of acetone favoured the production
of G and H (15.3 %) which are the direct precursors for the formation of octanol. Its
concentration in the final product mixture was 3.2 %.
The product distribution observed when using DMF differed completely from the rest of the
tested solvents, showing a high selectivity (>80 %) to the formation of two isomers of 4-(2-
furyl)-3-buten-2-dimethylamine. These compounds were likely formed via thermal
decomposition of DMF into carbon dioxide and dimethylamine (Figure 8a) [51], followed by
the subsequent ketone amine addition between F-Ac and dimethylamine (Figure 8b). This
reaction involves the formation of the corresponding hemiaminal and enamine intermediates
which are rapidly converted by dehydration and hydrogenation steps, respectively. Thus,
despite the high conversion of F-Ac achieved with DMF (93 %), the negligible yield of the
targeted hydrodeoxygenated products discourages the use of this solvent. This fact illustrates
the importance of a proper solvent selection not only with respect to conversion rates but also
in terms of supressing undesirable side reactions.
Conclusions
Solvents influence the catalytic hydrodeoxygenation of 4-(2-furyl)-3-buten-2-one (F-Ac) by
altering solvent-reactant and solvent-catalyst interactions. The use of Pt/TiO2 as a catalyst
allowed an almost complete conversion of F-Ac (>92 %) with all the tested solvents after 480
min under P(H2)=50 bar and at T=200 °C. The solvent polarity of protic solvents correlated
well with the observed hydrogenation activity, indicating the importance of solvent-reactant
interactions. Specifically, F-Ac conversion increased with decreasing protic solvent polarity.
On the other hand, a strong interaction between the catalysts and the solvents could inhibit
hydrogenation rate in case of aprotic polar solvents.
The differences in conversion rates have a strong influence on the obtained product
distribution due to the complex network of reactions involved in the studied transformation.
17

Thus, important differences were observed among the investigated solvents, achieving the
best HDO activities when using 2-propanol and acetone. Despite the formation of stable
cyclic intermediates such as 2-methyl-1,6-dioxaspiro[4,4]nonane (F) and 2-propyl-
tetrahydropyran (I), the use of these solvents led to a significant production of 4-(2-
tetrahydrofuryl)-butan-2-ol (28.5-40.4 %) and linear alcohols (6.3-10.4 %). The production of
the latter demonstrated that, besides hydrogenation reactions, ring opening took place when
using acetone and 2-propanol as solvents. Likewise, undesirable condensation reactions
between the solvent and the reactant were also observed with methanol (etherification) and
DMF (ketone-amine addition). The results indicate that in addition to the development and
design of the catalysts, the appropriate selection of the solvent media is essential in order to
optimise the HDO of furan derivatives compounds.
Acknowledgments
This publication is a result of the project no. P106/12/G015 supported by the Czech Science
Foundation which is being carried out in the UniCRE center (CZ.1.05/2.1.00/03.0071) whose
infrastructure was supported by the European Regional Development Fund and the state
budget of the Czech Republic.
18

References
[1] R.W. Torget, J.S. Kim, Y.Y. Lee, Ind. Eng. Chem. Res. 39 (2000) 2817-2825.
[2] K. Karimi, S. Kheradmandinia, M.J. Taherzadeh, Biomass Bioenergy 30 (2006) 247-
253.
[3] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933-1937.
[4] J.J. Bozell, G.R Petersen, Green Chemistry 12 (2010) 539-554.
[5] A.S. Dias, S. Lima, M. Pillinger, A.A. Valente, Advances in Synthetic Chemistry 8
(2010) 167-186.
[6] K.J Zeitsch, The chemistry and technology of furfural and its many by-products. Vol.
13. Elsevier, 2000.
[7] G.W. Huber, J.N. Chheda, C.J. Barrett, J.A. Dumesic, J. A., Science 308 (2005) 1446-
1450.
[8] L. Hora, V. Kelbichová, O. Kikhtyanin, O. Bortnovskiy, D. Kubička, Catal. Today 223
(2014) 138-147.
[9] A.D. Sutton, F.D., Waldie, R. Wu, M. Schlaf, A. Louis III, J.C. Gordon, Nat. Chem. 5
(2013) 428-432.
[10] L. Faba, E. Díaz, S. Ordónez, Appl. Catal. B: Env. 160 (2014) 436-444.
[11] R. Ramos, Z. Tišler, O. Kikhtyanin, D. Kubička, Catal, Sci. Tech. 6 (2016), 1829-1841.
[12] L. Faba, E. Díaz, S. Ordóñez, Catal, Sci. Tech. 5 (2015) 1473-1484.
[13] L. Faba, E. Díaz, S. Ordóñez, Chem. Sus. Chem, 7 (2014) 2816-2820.
[14] Q.N. Xia, Q. Cuan, X.H. Liu, X.Q. Gong, G.Z. Lu, Y.Q. Wang, Ang. Chem. Int. Ed. 53
(2014) 9755-9760.
[15] C.J. Barrett, J.N. Chheda, G.W. Huber, J.A. Dumesic, Appl. Catal B: Env 66 (2006)
111-118.
[16] R. Xing, A.V. Subrahmanyam, H. Olcay, W. Qi, G.P. van Walsum, H. Pendse, G.W.
Huber, Green Chem., 12 (2010), 1933-1946.
[17] W. Xu, Q. Xia, Y. Zhang, Y. Guo, Y. Wang, G. Lu, Chem. Sus. Chem. 4 (2011) 1758-
1761.
[18] J.N. Chheda, J.A. Dumesic, Catal. Today 123 (2007) 59-70.
[19] W. Xu, X. Liu, J. Ren, P. Zhang, Y. Wang, Y. Guo, G. Lu, Catal. Comm. 11 (2010)
721-726.
[20] K.L. Luska, J. Julis, E. Stavitski, D.N. Zakharov, A. Adams, W. Leitner, Chem. Sci. 5
(2014) 4895-4905.
[21] P. Mäki-Arvela, L.P. Tiainen, A.K. Neyestanaki, R. Sjöholm, T.K. Rantakylä, E. Laine,
T. Salmi, D.Y. Murzin, Appl. Catal. A: Gen. 237 (2002) 181-200.
[22] E. Toukoniitty, P. Mäki-Arvela, J. Kuusisto, V. Nieminen, J. Päivärinta, M. Hotokka, T.
Salmi, D. Murzin, J. Mol. Cat. A: Chem. 192 (2003) 135-151.
[23] J. Hájek, N. Kumar, P. Mäki-Arvela, T. Salmi, D.Y. Murzin, J. Mol. Cat. A: Chem. 217
(2004) 145-154.
19

[24] N.M. Bertero, A.F. Trasarti, C.R. Apesteguía, A.J. Marchi, Appl. Catal. A: Gen. 394
(2011) 228-238.
[25] P. Mäki-Arvela, J. Hajek, T. Salmi, D.Y. Murzin, Appl. Catal. A: Gen. 292 (2005) 1-49.
[26] R.J. Cvetanovic, Y. Amenomiya, Advan. Catal. Relat. Subj. 17 (1967) 103-149.
[27] M. Di Serio, M. Ledda, M. Cozzolino, G. Minutillo, R. Tesser, E. Santacesaria, Ind.
Eng. Chem. Res. 45 (2006) 3009-3014.
[28] M. Bolognini, F. Cavani, D. Scagliarini, C. Flego, C. Perego, M. Saba, Catal. Today, 75
(2002) 103-111.
[29] B. Coq, A. Tijani, F. Figuéras, F. J. Mol. Catal., 68 (1991) 331-345.
[30] A. Giroir-Fendler, D. Richard, P. Gallezot, Catal. Lett., 5 (1990) 175-181.
[31] P.A. Rautanen, J.R. Aittamaa, A.O.I. Krause, Ind. Eng. Chem. Res. 39 (2000) 4032-
4039.
[32] A. Drelinkiewicza, A. Waksmundzka, W. Makowski, J.W. Sobczak, A. Król, A.
Zieba, Catal. Lett. 94 (2004) 143-156.
[33] J. He, C. Zhao, J.A. Lercher, J. Catal. 309 (2014) 362-375.
[34] I. Kun, G. Szöllösi, M. Bartók, J. Mol. Catal. A: Chem. 169 (2001) 235-246.
[35] C. Reichardt, Solvent effects in organic chemistry. Verlag Chemie (1979).
[36] V. Gutmann, Electrochim. Acta 21 (1976) 661-670.
[37] H. Wan, A. Vitter, R.V. Chaudhari, B. Subramaniam, J. Catal. 309 (2014) 174-184.
[38] W. Zang, G. Li, L. Wang, X. Zhang, Catal. Sci. Tech. 5 (2015) 2532-2553.
[39] C.R. Waidmann, A.W. Pierpont, E.R. Batista, J.C. Gordon, R.L. Martin, R.M. West, R.
Wu, Catal. Sci. Tech., 3 (2013) 106-115.
[40] P. Panagiotopoulou, D.G. Vlachos, Appl. Catal. A: Gen. 480 (2014) 17-24.
[41] P. Panagiotopoulou, N. Martin, D.G. Vlachos, J. Mol. Catal. A: Chem. 392 (2014) 223-
228.
[42] R. Rao, A. Dandekar, R.T.K. Baker, M.A. Vannice, J. Catal. 171 (1997) 406-419.
[43] X. Hu, R.J. Westerhof, L. Wu, D. Dong, C.Z. Li, Green Chem. 17 (2015) 219-224.
[44] J. Luo, M. Monai, H. Yun, L. Arroyo-Ramírez, C. Wang, C.B. Murray, P. Fornasiero,
R.J. Gorte, Catal. Lett., 146 (2016) 711-717.
[45] J. Luo, L. Arroyo‐ Ramírez, R.J. Gorte, D. Tzoulaki, D.G. Vlachos. AIChE Journal, 61
(2015) 590-597.
[46] J.I. Di Cosimo, A. Acosta, C.R. Apesteguía, J. Mol. Catal. A: Chem. 234 (2005) 111-
120.
[47] M.A. Aramendıa, V. Borau, C. Jiménez, J.M. Marinas, J.R. Ruiz, F.J. Urbano, J. Mol.
́
Catal. A: Chem. 171 (2001) 153-158.
[48] A. Martin, U. Armbruster, I. Gandarias, P.L. Arias, Eur. J. Lipid Sci. Tech.115 (2013)
9-27.
[49] K. Alexander, L.S. Hafner, L.E Schniepp, J. Am. Chem. Soc. 73 (1951) 2725-2726.
[50] A.A. Anderson, S.P. Simonyan, E. Lukevics, Chem. Het. Comp. 34 (1998) 1406-1408.
20

[51] J. Muzart, Tetrahedron, 65 (2009) 8313-8323.
21

Figure captions
Figure 1
Reaction scheme for the hydrogenation of 4-(2-furyl)-3-buten-2-one over bifunctional
catalysts; (A) 4-(2-furyl)-butan-2-one, (B) 4-(2-tetrahydrofuryl)-butan-2-one, (C) 4-(2-furyl)-
butan-2-ol , (D) 4-(2-dihydrofuryl)-butan-2-ol, (E) 4-(2-tetrahydrofuryl)-butan-2-ol.
Figure 2
TPD curves of Pt/TiO2 catalyst; (a) ammonia TPD and (b) carbon dioxide TPD.
Figure 3
22

Solvent effects on the conversion of 4-(2-furyl)-3-buten-2-one over Pt/TiO2 catalyst; (a) protic
and
(b)
aprotic
solvents.
Figure 4
Initial hydrogenation rates as a function of solvent dielectric constants of polar protic
solvents.
Figure 5
23

Initial hydrogenation rate of F-Ac in aprotic solvents as a function of: (a) nucleophilic donor
number [36] and (b) solvent desorption energy (calculated by TPD).
Figure 6
Evolution of the yields of the main products from the HDO of 4-(2-furyl)-3-buten-2-one over
Pt/TiO2
using
acetone
as
solvent.
Figure 7
24

Yields of the hydrogenated products from HDO of F-Ac at (a) the end of the reaction (480
min) and (b) considering the same conversion degree of F-Ac (70 %).
Figure 8
Reaction scheme for (a) thermal decomposition of n,n-dimethylformamide and (b) addition of
4-(2-furyl)-3-buten-2-one and dimethylamine.
25

Table 1. Physicochemical properties of Pt/TiO2 catalyst.
Pt content
BET surface
Pore size
Amount of acid sites
NH3 desorption maximum (ºC)
Amount of base sites
CO2 desorption maximum (ºC)
Pt dispersion
Average Pt particle size
(wt%)
1.8
95
30.4
0.134
234
0.416
105/194
15.6
7.2
(m2·g-1)
(nm)
(meq NH3∙g-1)
(meq CO2∙g-1)
(%)
(nm)
26

Table 2. Solvent polarity parameters of the studied solvents (Dipole moment (μ), dielectric
constant (εr)) and initial hydrogenation rates of HDO conversion of F-Ac over Pt/TiO2.
μ
εr
(at 25 ºC) (mmol·gcat-1·min-1)
r0 x 10-4
Solvent
Debye
Methanol
Ethanol
1.70
1.69
1.68
1.66
1.70
1.63
2.88
3.86
32.7
24.3
20.1
18.3
13.9
7.6
0.80
4.31
7.77
8.08
2.44
1.72
8.70
0.35
1-propanol
2-propanol
1-pentanol
THF
Acetone
DMF
20.5
36.7
27

Table 3. Yields of F-Ac hydrodeoxygenation products after 480 min (T=200 ºC; P(H2)=50
bar).
OH
OH
OH
O
O
O
O
O
O
O
+
+
+
+
+
O
O
B
C
E
D
I
A
OH
O
O
O
O
O
F-Ac
OH
+
+
OH
+
H
G
F
J
Yield (mol%)
Conv.
(mol%)
92.5
Solvent
A
B
C
D
E
F
0
7.2
G+H
I
J
Methanol
Ethanol
1-Propanol
60.5
34.8
5.2
0
57
22.9
4.5
16
1.8
1.6
1.6
3.4
6.8
1.4
11.3
7.2
9
1.7
4.7
0
0
6.5
9.7
0
0
0
0
0
17.3
31.7
0
0
2
0
6.3
0
100
100
14.8
23.8
2.6
28
47.8
7.1
15.2
40.4
1
19.6
8.2
2-Propanol 100
1-Pentanol
THF
92
100
2.5
3.1
4.6
13.3
0
10.4
Acetone
100
2
1.7
6.1
4.5
28.5
10.3
15.3
20.9
28