Selective hydrogenation of fatty acids to alcohols over highly dispersed ReOx/TiO2 catalyst the paper is dedicated to the living memory of Dr. Haldor Topsoe.

Article abstract of DOI:10.1016/j.jcat.2015.01.003

Production of fatty alcohols through selective hydrogenation of fatty acids was studied over a 4% ReO_x/TiO₂ catalyst. Stearic acid was hydrogenated to octadecanol at temperatures and pressures between 180-200°C and 2-4 MPa, with selectivity reaching 93%. A high yield of octadecanol was attributed to a strong adsorption of the acid compared to alcohol on the catalyst, which inhibits further alcohol transformation to alkanes. Low amounts (<7%) of alkanes (mainly octadecane) were formed during the conversion of stearic acid. However, it was found that the catalyst could be tuned for the production of alkanes. The reaction intermediates were octadecanal and stearyl stearate. Based on the reaction products analysis and catalyst characterization, a reaction mechanism and possible pathways were proposed.

Full text of DOI:10.1016/j.jcat.2015.01.003

Journal of Catalysis 328 (2015) 197–207
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Selective hydrogenation of fatty acids to alcohols over highly dispersed
ReO /TiO catalyst
x
2
a
a,1
a
b
b
Bartosz Rozmysłowicz , Alexey Kirilin , Atte Aho , Haresh Manyar , Christopher Hardacre ,
a,c
a
a,
⇑
Johan Wärnå , Tapio Salmi , Dmitry Yu. Murzin
a
Laboratory of Industrial Chemistry and Reaction Engineering, Process Chemistry Centre, Åbo Akademi University, FI-20500 Åbo-Turku, Finland
CenTACat, School of Chemistry and Chemical Engineering, Queen’s University of Belfast, UK
Umeå University, SE-901 87 Umeå, Sweden
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Production of fatty alcohols through selective hydrogenation of fatty acids was studied over a 4% ReOx/
Received 8 August 2014
Revised 5 November 2014
Accepted 3 January 2015
2
TiO catalyst. Stearic acid was hydrogenated to octadecanol at temperatures and pressures between
180–200 °C and 2–4 MPa, with selectivity reaching 93%. A high yield of octadecanol was attributed to
a strong adsorption of the acid compared to alcohol on the catalyst, which inhibits further alcohol trans-
formation to alkanes. Low amounts (<7%) of alkanes (mainly octadecane) were formed during the conver-
sion of stearic acid. However, it was found that the catalyst could be tuned for the production of alkanes.
The reaction intermediates were octadecanal and stearyl stearate. Based on the reaction products analy-
sis and catalyst characterization, a reaction mechanism and possible pathways were proposed.
Ó 2015 Elsevier Inc. All rights reserved.
The paper is dedicated to the living memory
of Dr. Haldor Topsøe.
Keywords:
Fatty acids
Fatty alcohols
Rhenium
Selective hydrogenation
Hydrocarbons
1
. Introduction
Ru–Sn/Al
hols (250 °C; 5.6 MPa) in methyl laurate hydrogenation. In addi-
tion, a change of the support in the Ru–Sn catalyst from Al to
TiO increases the activity and selectivity which is ascribed to
Ru–Ti metal–support interactions [6]. The same properties were
observed with a Pt/TiO catalyst [7] as hydrogenation of fatty acids
to alcohols was performed at relatively low temperatures (110–
150 °C) and hydrogen pressures (2 MPa) with selectivity 90–93%.
It is assumed that an enhanced activity is caused by the Ti3+ ion/
oxygen vacancy on the metal–support interface, which interacts
with the carbonyl oxygen and weakens C@O bond, promoting
hydrogenation of the carboxylic acid [6,7].
Rhenium catalysts were disclosed in the patent literature as
being capable of hydrogenating carbonyl compounds to alcohols
with high overall selectivity, without formation of ethers. Selective
hydrogenation of fatty acids to alcohols was also studied over rhe-
nium-based catalysts. Decanoic acid transformation has been per-
2 3
O [5] shows the highest selectivity towards fatty alco-
Fatty alcohols are non-ionic surfactants widely used as emulsifi-
2 3
O
ers, emollients and thickeners in alimentary and cosmetic indus-
tries [1]. Moreover, they can be substrates for the production of
other surface-active materials, such as alkylamines and alkylsulfat-
es. The current world production capacity of fatty alcohols reaches
2
2
3
.35 Mt/a, with an estimated global demand growth (2012–2017)
of 3.2%/a [2]. Fatty alcohols are mainly produced through catalytic
hydrogenation of fatty acids, methyl esters or wax esters. In tradi-
tional processes, the methyl esters are hydrogenated over Cu–Cr-
based [3] catalysts with a high selectivity towards alcohols. How-
ever, due to low reactivity of the carboxyl group, elevated tempera-
tures (200–300 °C) and hydrogen pressures (20–30 MPa) have to be
applied. Moreover, chromium catalysts are not environmentally
friendly due to a release of hazardous chromium compounds.
Noble metal catalysts such as Ru or Pt were reported to be
active in selective hydrogenation of fatty acids. Many kinds of
Ru-based catalysts have been studied [4], from which bimetallic
formed over Re
conversion and a moderate selectivity towards 1-decanol [8].
Moreover, it has been shown that an addition of OsO to form a
2 7
O at 10 MPa of hydrogen and 130 °C, giving a low
4
bimetallic Re–Os [8] catalyst increases not only the rate but also
the selectivity to 84%. Furthermore, Re and bimetallic Re–Sn
deposited over ZrO and Al O have been studied at 250 °C and
2 2 3
⇑
Corresponding author.
E-mail address: dmurzin@abo.fi (D.Yu. Murzin).
Dow Benelux B.V., P.O. Box 48, 4530 AA, Terneuzen, Netherlands.
1
http://dx.doi.org/10.1016/j.jcat.2015.01.003
021-9517/Ó 2015 Elsevier Inc. All rights reserved.
0

1
98
B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
5
.6 MPa of hydrogen pressure in the hydrogenation of oleic acid.
2.6. Inductively Coupled Plasma–Optical Emission Spectroscopy (ICP–
OES)
The Re/Al catalyst afforded the highest selectivity to alcohols,
2 3
O
reaching 58% [9].
In this work, Re/TiO
2
was studied in the selective hydrogenation
The Re content of the catalyst was determined by Inductively
of stearic acid to 1-octadecanol, displaying high selectivity of 93%
under mild reaction conditions. Unusual properties of Re allow a
selective conversion of fatty acids suppressing the formation of
by-products. Moreover, the catalyst can be tuned for the produc-
tion of hydrocarbons.
Coupled Plasma (ICP) using an Optical Atomic-Emission Spectrom-
eter Optima 4300 DV. For the analysis of the Re content, the Re/
TiO
ture (2 ml
40% + 0.5 ml HNO
2
catalyst was dissolved in a microwave oven by an acid mix-
SO 96% + 2 ml PO 85% + 1 ml HF
– 65%).
H
2
4
–
H
3
4
–
–
3
2.7. Nitrogen physisorption
2
. Experimental
The specific surface area of the fresh catalyst was measured by
2
.1. Catalyst preparation
nitrogen physisorption using a Sorptometer 1900 (Carlo Erba
instruments) apparatus. Prior to the analysis, the sample was
degassed at 200 °C for 3 h. The specific surface area was calculated
by BET equation.
Titania P-90 and perrhenic acid (assay 39.4%) were supplied by
Nippon Aerosil and Johnson Matthey, respectively. The Re catalysts
were prepared by incipient wetness impregnation. An appropriate
mass of the perrhenic acid solution was diluted with deionised
water (18 MO) of volume equal to the pore volume of the support.
The solution was added to the support in three portions of equal
volumes with stirring after each addition, until the solution was
thoroughly mixed. The product was then dried at 120 °C for 12 h,
followed by calcination at 500 °C for 4 h.
2
.8. X-ray photoelectron spectroscopy (XPS)
The Re/TiO catalyst was reduced at atmospheric pressure and
2
at 0.5 MPa in hydrogen. The catalyst was transported under ace-
tone to a glove box where it was filtered and dried at room temper-
ature in an inert atmosphere. The spent catalyst was transported
after the reaction to the glove box submerged in reaction mixture.
At inert atmosphere, the catalyst was washed with acetone and
dried at room temperature. All catalysts were brought to the XPS
equipment in sealed containers. During the insertion to the XPS
equipment, the catalyst had less than 2-min contact with air.
2.2. Temperature-programmed reduction (TPR)
2
Prior to the analysis, 50 mg of non-reduced Re/TiO catalyst was
kept in an oven at 100 °C overnight. Temperature-programmed
reduction was performed in an Autochem Micromeritics 2910
A Perkin–Elmer PHI 5400 spectrometer with a Mg K
a X-ray
source operated at 14 kV and 200 W was used in the XPS analysis
of the samples. The pass energy of the analyser was 17.9 eV and the
energy step 0.05 eV. The binding energy calibration was based on
the Ti 2p3/2 peak at 458.8 eV. Fitting of the XPS data was performed
after removal of Shirley background using Voigt profile. In the peak
fitting procedure, the Re 4f5/2 and Re 4f7/2 intensity ratio was 0.75
and energy separation was 2.43 eV. The sensitivity factors used
in the quantitative analysis for O 1s, Ti 2p3/2, Re 4f and C 1s were
0.711, 1.334, 3.961 and 0.296, respectively.
2
using 5% H in Ar. The temperature ramp was 10 °C/min to 500 °C.
2.3. Temperature-programmed desorption (TPD)
Temperature-programmed desorption was performed in an
Autochem Micromeritics 2910. Ammonia and carbon dioxide
were used for the determination of the acid and base sites,
respectively. The sample was at first kept under a helium atmo-
sphere at 400 °C for 1 h then cooled to 100 °C, in the case of
NH
3
TPD, and 30 °C, in the case of CO
2
TPD. Afterwards, the adsor-
2.9. Kinetic experiments
bate was introduced for 30 min. Subsequently, the catalyst was
flushed with an inert gas for 30 min after which the analysis
was started. The temperature ramp was 10 °C/min up to 700 °C
for NH and 600 °C for CO where the sample was kept for 1 h.
3 2
2
Before the adsorption, the Re/TiO catalyst was reduced in situ
In all of the kinetic experiments, the reaction mixture was pre-
pared from 1 g (0.0035 mol) of stearic acid (Merck 97%) and dis-
solved in 100 ml of dodecane (Sigma–Aldrich 99%) at 60 °C. The
solution was injected to a heated pre-reactor at 90 °C. To remove
air, the pre-reactor was flushed with inert gas, which afterwards
was substituted with hydrogen.
with hydrogen at 400 °C for 2 h.
Experiments were performed in a high-pressure steel Parr reac-
tor at temperature range 180–220 °C and total over pressure was
2.4. CO chemisorption
2
–4 MPa in hydrogen (AGA, 5.0). The catalyst (0.1 g) was reduced
ex situ, under a hydrogen atmosphere (1 bar) for 2 h at 400 °C (tem-
perature was selected based on Re/TiO TPR – Fig. 2) with a heating
CO chemisorption was performed in an Autochem Micromeri-
tics 2910. The Re/TiO catalyst was reduced in situ with hydrogen
2
2
prior to the analysis at 400 °C for 2 h. The hydrogen excess was
removed by flushing the catalyst with helium at reduction temper-
ature for 30 min. Pulse chemisorption was performed in a quartz
glass reactor performed at 25 °C and glass reactor submerged in
a water bath for temperature control.
ramp of 2 °C/min. After the reduction, the catalyst was immedi-
ately transferred to the reactor. The air in the reactor was substi-
tuted first with inert gas and subsequently with 0.2 MPa of
hydrogen. Thereafter, the reactor was heated (5 °C/min) to the
desired temperature at which the reaction mixture was injected
together with hydrogen to achieve the desired pressure. The pro-
peller stirrer with a stirring rate set to 1200 rpm was used to
exclude external mass transfer limitations. The internal mass
transfer limitations can be excluded as a grain size of titania sup-
2.5. Transmission electron microscopy (TEM)
The rhenium particle-size distribution was measured by a JEM-
400 Plus Transmission Electron Microscope (voltage 120 kV). The
1
port was below 63
lm.
sample was reduced at 400 °C for 2 h before analysis. Histograms
In the experiment with the non-reduced Re/TiO , the catalyst
2
of the particle-size distribution were obtained by counting at least
was placed in the reactor at room temperature and flushed with
argon to remove air. Afterwards, temperature was raised to
1
00 particles on the micrographs for the sample.

B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
199
2
20 °C. Reactants and hydrogen were injected at the same time and
Table 1
x 2
Properties of 4% ReO /TiO catalyst.
the stirring was started.
Surface
Pore
volume
(ml/g)
Basicity
mol/g)
Acidity
mol/g)
Average
particle
size (nm)
Metal
loading
(wt.%)
area
(
l
(
l
2.10. Product analysis
2
(
m /g)
TiO
Re/TiO
2
108
63
0.18
0.22
37
531
15
270
–
–
3.8
The analysis of the liquid phase was performed using gas chro-
matography (HP 5890) containing ZB-5 Inferno column (30 m,
2
0.9 ± 0.03
0
.32 mm, 0.1
lm). Samples for analysis were prepared by mixing
0
.1 ml of the reaction mixture with 0.1 ml of the silylation agent
(
BSTFA, Sigma–Aldrich), 0.1 ml of eicosane (Acros Organics, 99%)
three kinds of sites were found with maxima at 80 °C, 161 °C and
dissolved in dodecane (1 mg/ml) as an internal standard and 1 ml
of pyridine (Sigma–Aldrich, 99.9%). Thereafter, the samples were
left in the oven at 70 °C for 1 h to ensure a proper silylation before
injection to the GC. The same sample preparation procedure was
performed before the qualitative analysis in GC–MS (GC-6890N,
MS-5973) with the chromatographic column DB-PETRO (100 m,
2
348 °C (Fig. 1b), with total volume of CO adsorbed increasing over
14 times (Table 1).
3
.1.1. Oxidation state of rhenium
The TPR analysis shows that the reduction of rhenium starts at
around 200 °C and passes through two maxima at 278 °C and
30 °C (Fig. 2). In all of the experiments, the catalyst was reduced
0
.25 mm, 0.5
lm).
3
at 400 °C for 2 h. At the same time, the XPS analysis revealed that
3
. Results and discussion
the catalyst while being active was not fully reduced under these
7
+
6+
4+
conditions (Fig. 3b) and contains Re , Re and Re species, which
is in agreement with the previous studies on the reduction of sup-
3.1. Catalyst characterization
7
+
ported rhenium [11]. Re was reduced during the reaction to form
6
+
4+
Temperature-programmed desorption (TPD) of NH
3
and CO
2
Re and Re (Fig. 3c). Moreover, similar catalyst oxidation states
were found after the reaction independently on the reduction state
was used to determine the strength and the amount of acid and
basic sites. The NH TPD of non-impregnated TiO shows weak acid
sites with a desorption maximum at 140 °C. The Re/TiO catalyst
3
2
of the catalyst before an experiment (Re-oxide/TiO or reduced Re/
2
0
2
TiO ) (Table 2). Furthermore, there were no traces of metallic Re
2
showed two maxima at 198 °C and 368 °C, indicating two kinds
species in the fresh and spent catalysts suggesting that ReO is
x
of acid sites (Fig. 1a). The total volume of ammonia adsorbed over
the active component in the selective hydrogenation of fatty acids
the Re/TiO
ing a high increase of the acidity after the deposition of rhenium
Table 1). However, it has to be pointed out that the decomposition
of ammonia over rhenium can occur at temperatures around
50 °C, which could influence the generation of the second peak
10].
The TPD of CO
and Re/TiO catalysts. Pure TiO
low temperature (92 °C), while after the deposition of rhenium,
2
catalyst was 18 times higher than of bare TiO
2
, indicat-
to alcohols.
It is important to mention that rhenium has a high affinity to
oxygen, which increases with the metal dispersion [12]. A highly
dispersed catalyst, which is prepared in this work, can be oxidized
upon contact with air. To confirm this hypothesis, the reduced
sample was separated into two parts. The first part was kept under
an inert atmosphere, while the second one was exposed to air at
room temperature for 15 min. Thereafter, the catalysts were ana-
lysed with XPS. This experiment confirmed that even a short expo-
sure to air at room temperature can result in the partial oxidation
(
3
[
2
was performed similar to that of NH
3 2
on TiO
2
2
showed a desorption peak at a
6
+
4+
7+
of Re and/or Re to Re (Table 2).
3
2
2
2
2
2
1
1
1
1
1
0
0
0
0
0
.0
.8
.6
.4
.2
.0
.8
.6
.4
.2
.0
.8
.6
.4
.2
.0
(b)
3
.1.2. Particle size of rhenium
The rhenium particle-size distribution were measured by trans-
CO2
mission electron microscopy (TEM) technique, showing a very high
dispersion of rhenium (close to 100%) with an average particle size
calculated from the images being 0.9 ± 0.03 nm (Fig. 4). For com-
parison, CO chemisorption on the catalyst was also performed
TiO2
Re/TiO2
(a)
NH3
0
100 200 300 400 500 600 700
Temperature (°C)
0
100
200
300
400
500
600
700
Fig. 1. (a) NH
3
-TPD and (b) CO
2
-TPD of (red) Re/TiO
2
and (black) TiO
2
. (For the
Temperature (°C)
interpretation of the references to colour in this figure legend, the reader is referred
to the Web version of this article.)
x 2
Fig. 2. Temperature-programmed reduction (TPR) of 4% ReO /TiO catalyst.

2
00
B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
7
+
6+
4+
ꢀ1 ꢀ1
Re4
Re
Re
from 0.07 ðmolstearic acid
g
h
Þ to 0.39 between 180 and 220 °C
f7/2
4f7/2
4f7/2
Re
(
Fig. 5b). Moreover, an increase in the hydrogen pressure enhances
3
2
1
0
the reaction rate. It was observed that, at 200 °C, an increasing the
hydrogen pressure from 2 to 3 to 4 MPa resulted in an increased
conversion of 40% and 70% after 30 min reaction time, respectively.
The apparent reaction order with respect to hydrogen was close to
unity.
(
a)
b)
non-reduced
x 2
The formation of the alcohol from the acid over ReO /TiO cat-
alyst occurs through an aldehyde intermediate, which is formed
at the beginning of the reaction and passes through a maximum.
Simultaneously, small amounts of hydrocarbons were observed.
Octadecane and octadecene were formed by dehydration and sub-
sequent hydrogenation (only octadecane) of octadecanol, while
heptadecane and heptadecene were the products of stearic acid
decarboxylation or octadecanal decarbonylation/dehydroformyla-
tion. Similar reaction pathways have been proposed for deoxygen-
ation of fatty acid over noble-metal-supported catalysts at higher
temperatures [14]. Furthermore, the formation of a wax ester (ste-
aryl stearate) was observed, which appears after ꢁ10% of stearic
acid conversion from the reaction of stearic acid with 1-octadeca-
nol. Based on the product distribution, the reaction pathway is pro-
posed (Scheme 1).
The selectivity to octadecanol increases with the conversion of
stearic acid reaching 88–93% at a high conversions (Fig. 6a and
b). The aldehyde formation is prominent at low conversion levels,
while ester formation influences the final selectivity at high con-
version level. The formation of alkane is visible during the whole
reaction. The selectivity to hydrocarbons is not strongly affected
by a change of the reaction conditions, in contrary to the selectivity
towards the aldehyde and the ester.
(
reduced
(
c)
spent
38
5
0
48
46
44
42
40
Binding energy (eV)
Fig. 3. XPS rhenium species for (a) non-reduced catalyst; (b) reduced catalyst at
00 °C for 2 h and (c) spent catalyst after reaction with pre-reduced catalyst.
4
The initial selectivity towards the aldehyde can be as high as
5% for the reaction at 220 °C and 2 MPa of total pressure. How-
Table 2
4
x 2
The area (%) for rhenium species in XPS analysis of 4% ReO /TiO catalyst.
ever, a decrease of the temperature at 2 MPa of total pressure
decreases the rate of aldehyde formation (Fig. 6c), therefore,
increasing the initial selectivity to the alcohol. Moreover, it can
be observed that the hydrogen pressure does not influence the
selectivity towards the aldehyde (Fig. 6d). On the contrary, the for-
mation of stearyl stearate at high conversion levels can be inhib-
ited by a high hydrogen pressure that decreases the selectivity to
ester from 3.5% to 1.5% (conversion = 80%) in experiments at
4
% ReO
x
/TiO
2
Peak area (%)
Re4+
Re6+
Re7+
Non-reduceda
0
47
73
58
49
31
36
27
37
30
69
17
0
5
21
Pre-reducedb
c
Spent (non-reduced)
d
Spent (pre-reduced)
Oxidized (pre-reduced)e
a
b
c
Fresh non-reduced catalyst.
Fresh pre-reduced catalyst at 400 °C for 2 h in H
Spent catalyst after reaction at 220 °C and 2 MPa of total pressure (without pre-
2
MPa and 4 MPa of total pressure, respectively. The decrease of
2
(0.1 MPa).
temperature had no influence on the selectivity towards the ester.
The selectivity to hydrocarbons remains constant below 5%
between 20% and 90% of the stearic acid conversion. During initial
conversion period, an increased formation of alkanes can be
observed (Fig. 7). Moreover, an increase of the hydrogen pressure
does not influence the formation rates of the hydrocarbons as well
as the distribution of hydrocarbon products. It was observed that
lower reaction temperatures decrease the selectivity to total
hydrocarbons. At 180 °C (2 MPa), no unsaturated products were
found. Furthermore, the decarboxylation/decarbonylation rates
decrease under these conditions, yielding less C17 hydrocarbons,
while the rate of octadecanol hydrogenation to octadecane remains
the same. In overall, the selectivity towards alkanes decreased to
reduction).
d
Spent catalyst after reaction at 220 °C and 2 MPa of total pressure (pre-reduced
in H at 400 °C and 0.1 MPa for 2 h).
Fresh catalyst after pre-reduction at 400 °C for 2 h in H (0.1 MPa) and sub-
2
sequent exposure on air at room temperature for 15 min.
2
e
and the calculated (with assumed stoichiometric factor CO/Re = 1)
dispersion was 35%. The differences between TEM and CO chemi-
sorption can be explained by the partial reduction of rhenium.
4
+
6+
0
Re and Re binds much weaker CO than Re since rhenium cat-
ions have less electrons to back-donate to the CO ligand [13]. This
results in a lower CO uptake and subsequently underestimation of
the dispersion.
x 2
the level around 4%. The selectivity of 4% ReO /TiO at different
temperatures and pressures can be seen in Table 3.
3.2. Hydrogenation of stearic acid
3.2.1. Formation of alkanes
During the conversion of stearic acid, the formation of hydrocar-
bons is minimal and stays at a level of around 4–5%. However, after a
complete conversion of the acid, the rates of the alkane formation
increase rapidly. For the experiment at 220 °C and 2 MPa of hydro-
gen pressure, the generation activity of the main hydrocarbon prod-
x 2
The hydrogenation of stearic acid over ReO /TiO was per-
formed in the temperature range between 180 °C and 220 °C in
hydrogen at a total pressure between 2 and 4 MPa. The main prod-
uct was octadecanol (Fig. 5a). The experiments showed that the
product formation can be achieved with a high activity even at
temperatures below 200 °C. The initial activity of the catalyst (cal-
culated at 30 min for each reaction) at 2 MPa total pressure varied
ꢀ1
ꢀ1
uct (octadecane) increases from 0.01 ðmoloctadecane
g
h
Þ – during
Re
the conversion of fatty acids (60–80%) – to 0.05 – during the later
transformation of the alcohol (3–9%). This phenomenon can be

B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
201
(a)
(b)
Fig. 4. (a) TEM image of pre-reduced at 400 °C for 2 h 4% ReO
x
/TiO
2
2
catalyst; (b) size distribution of rhenium nanoparticles over TiO .
(b) 0
.40
(a) 100
0.35
8
6
4
2
0
0
0
0
0
0.30
0.25
0.20
0.15
0.10
Heptadecane
Octadecane
Octadecanol
Stearic acid
Heptadecene
Octadecene
Octadecanal
Stearyl-stearate
0.05
0.00
0
60
120
180
1
80°C 200°C 220°C 2 MPa 3 MPa 4 MPa
Time (min)
2 MPa
200°C
x 2 2
Fig. 5. (a) Reaction of stearic acid over 4% ReO /TiO catalyst at 220 °C in hydrogen at 2 MPa of total pressure; (b) activity of Re/TiO catalyst in stearic acid hydrogenation
ꢀ
1
ꢀ1
ðmolstearic acid
g
h
Þ calculated at t = 30 min.
Re
x 2
Scheme 1. Reaction pathways of stearic acid conversion over 4% ReO /TiO catalyst.

2
02
B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
(a) 100
(b) 100
9
8
7
0
0
0
9
8
7
6
5
0
0
0
0
0
o
2 MPa
3 MPa
4 MPa
180 C
2
2
o
00 C
o
20 C
60
50
0
20
40
60
80
100
0
20
40
60
80
100
Conversion of stearic acid (%)
Conversion of stearic acid (%)
(c) 50
(d) 50
4
0
4
3
2
1
0
0
0
0
0
o
1
2
2
80 C
30
20
10
0
2 MPa
3 MPa
4 MPa
o
00 C
o
20 C
0
20
40
60
80
100
0
20
40
60
80
100
Conversion of stearic acid (%)
Conversion of stearic acid (%)
x 2
Fig. 6. Selectivity (%) from conversion of stearic acid (%) over 4% ReO /TiO catalyst in hydrogen atmosphere to: (a) octadecanol and (c) octadecanal for reaction at 2 MPa of
total pressure with temperatures between 180 and 220 °C; (b) octadecanol and (d) octadecanal at 200 °C with total pressure from 2 to 4 MPa.
o
The ratio between the hydrogenation product (octadecane) and
decarboxylation/decarbonylation product (heptadecane) increases
with time as octadecanol is the main substrate for further transfor-
mation. However, a small amount of heptadecane was produced
due to the conversion of stearyl stearate, which is still present in
the system.
1
1
1
0
9
8
7
6
5
4
3
180 C - 2 MPa
2
o
00 C - 2 MPa
o
220 C - 2 MPa
o
2
2
00 C - 3 MPa
o
00 C - 4 MPa
3.2.2. Reduction of the catalyst
2
A reaction with a non-reduced Re/TiO catalyst was performed
2
at 220 °C in 2 MPa H . The conversion of stearic acids was signifi-
cantly lower than in the reaction with the pre-reduced catalyst
ꢀ
1
ꢀ1
(
Fig. 9a) with an initial activity of 0.06 ðmolstearic acid
g
h
Þ com-
Re
pared to 0.39 for pre-reduced catalyst at the same conditions.
Moreover, the selectivity to the main product for the non-reduced
catalyst is lower at all conversion levels, which is caused by an
increased formation of hydrocarbons in comparison with the
experiment with a pre-reduced catalyst (Fig. 9b).
0
20
40
60
80
100
Conversion of fatty acids (%)
The XPS measurements revealed that the oxidation state of rhe-
nium for both non-reduced and pre-reduced spent catalysts was
similar (Table 2), indicating that the reduction of rhenium for a
non-reduced catalyst occurred during the reaction. However, a
lower conversion in the experiment with the non-reduced catalyst
indicates that the gas-phase pre-reduction step is crucial for the
catalyst activity. It can be explained by a deficiency of the oxygen
vacancies in titania, which formation requires a higher reduction
temperature than used in the experiment [15]. During the pre-
Fig. 7. Selectivity to hydrocarbons (%) from conversion of stearic acid in hydrogen
atmosphere over 4% ReO /TiO catalyst for reactions at temperature range between
80–220 °C and 2–4 MPa of total pressure.
x
2
1
caused by a difference of the adsorption strength between fatty acid
and the alcohol. Since the acid is stronger adsorbed on the active
site, the alcohol cannot adsorb and react to form alkanes. In the
absence of stearic acid, the transformation of the alcohol takes
place, and thus, competitive adsorption is the main reason of the
unusually high selectivity of stearic acid conversion to octadecanol
reduction of the catalyst at 400 °C, TiO is partially reduced creat-
2
over ReO
x
/TiO
2
at medium temperatures and pressures.
ing oxygen vacancies needed for the reaction to occur through the
As all acid was converted, the transformation of octadecanol
took place. The main product of the octadecanol conversion is octa-
decane with a rising selectivity as the reaction proceeds (Fig. 8).
proposed mechanism, which is similar to the mechanisms sug-
gested previously for the alcohol formation over titania-supported
platinum [7] and ruthenium [6]. A shift in the binding energy of

B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
203
Table 3
Selectivity to products at different conversion levels in the reaction of stearic acid over 4% ReO
x
2
/TiO catalyst.
Temperature T
°C)
Total pressure Selectivity at 20% conversion
Selectivity at 80% conversion
Selectivity to hydrocarbons at conversions
between 20% and 80%
(
(MPa)
Octadecanol Octadecanal Stearyl
Octadecanol Octadecanal Stearyl
stearate
stearate
180
200
220
200
200
2
2
2
3
4
83
75
64
75
75
13
20
30
20
20
–
–
–
–
–
–
–
–
4
5
5
5
5
90
88
92
92
1.5
3.5
1.5
1.5
3.5
3.5
1.5
1.5
9
8
8
8
8
8
0
9
8
7
6
5
mation of alkanes as the adsorption of alcohol on the catalyst
increased. A similar effect was observed for hydrogenation of CO
over ReO
.3. Kinetic analysis
The reaction network for hydrogenation of stearic acid is dis-
x 2
/TiO at 250 °C and 2.1 MPa [15].
3
played in Scheme 2. For the development of the kinetic model, it
was assumed that all organic molecules are adsorbed on the same
type of sites. The reaction order with respect to hydrogen is almost
unity, which is consistent with low hydrogen coverage and its poor
adsorption on rhenium compared to such metals as platinum. It
was thus reasonable to assume that hydrogen from the fluid phase
takes part in the reaction. Assumption of a pairwise addition of
weakly adsorbed hydrogen would result in the same kinetic
expression, while a concept of hydrogen adsorption with dissocia-
tion and stepwise addition of hydrogen atoms with the first step
being the rate limiting in the spirit of the Horiuti–Polanyi mecha-
nism inevitably results in the reaction order towards hydrogen
equal to 0.5. Based on kinetic regularities, generation of an inter-
mediate complex on the surface consisting of stearic acid and
hydrogen can be considered as the rate-limiting step in the route
from stearic acid to octadecanal while a further rearrangement of
the surface complex to the aldehyde and water is fast. An analo-
gous sequence of steps was supposed for hydrogenation of octade-
0
2
4
6
8
10
Conversion of ocatadecanol (%)
Fig. 8. Selectivity to C18 hydrocarbons from conversion of octadecanol in reaction of
stearic acid over 4% ReO
stearic acid conversion.
x 2
/TiO at 220 °C and 2 MPa of total pressure – after full
rhenium species for the reduced catalyst can be ascribed to sub-
stantial interactions between rhenium and Ti3 . The Re–O–Ti
+
3+
sites are responsible for the formation of alcohol on which the car-
bonyl group of the acid is stronger adsorbed, thus inhibiting fur-
ther transformations of the alcohol to hydrocarbons. A decrease
in the titania oxygen vacancies on the interface with rhenium
nanoparticles leads to a decrease of the number of sites on which
adsorption of the acid is strong. Therefore, the alcohol can be more
easily transformed to alkanes when the acid is still present in the
system. This explains both lower conversion of fatty acids as a
number of active sites for the reaction decreased and a higher for-
canal to octadecanol. Transformations of the latter to
a
corresponding alkane require first dehydration to octadecene with
subsequent hydrogenation.
Similar type of hydrogenation steps was considered in the liter-
ature [16]. Such sequence of steps can account for reaction orders
in organic compounds ranging from zero to unity.
The reaction mechanism is comprised of four reaction routes
(Schemes 2 and 3). In Schemes 2 and 3, SA is stearic acid, AD, AC,
(
a)
1
non-reduced catalyst
pre-reduced catalyst
(b) 100
100
80
60
40
20
0
00
8
6
4
2
0
0
0
0
0
8
6
4
2
0
0
0
0
0
non-reduced catalyst
pre-reduced catalyst
0
60
120
180
240
300
360
0
20
40
60
80
100
Time (min)
Conversion of stearic acid (%)
Fig. 9. Comparison of reactions with pre-reduced and non-reduced catalysts at 220 °C and 2 MPa of total pressure: (a) conversion of stearic acid in time; (b) selectivity to
octadecanol and hydrocarbons from conversion of stearic acid.

2
04
B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
Table 4
Values of the estimated kinetic parameters.
*
Parameter Units
Value
Relative standard
error (%)
0
0
0
0
ꢀ1
ꢀ1
ꢀ1
4.3 ꢂ 10ꢀ3 ± 0.610
ꢀ3
k
k
k
k
E
E
E
E
1
2
3
4
bar min
6.9
12.8
20.1
20.2
1.4
31.1
28.9
31.6
11.0
ꢀ1
bar min
0.01 ± 0.0026
ꢀ1
ꢀ3
ꢀ3
ꢀ5
min
0.424 ꢂ 10 ± 0.17 10
ꢀ1
ꢀ1
ꢀ5
(mol%) min
kJ/mol
kJ/mol
kJ/mol
kJ/mol
0.382 ꢂ 10 ± 0.14 10
Scheme 2. Reaction network.
1
2
3
4
106 ± 3
39.3 ± 22
70.6 ± 40
86.3 ± 54
0.025 ± 0.004
OE, OD and SS are, respectively, octadecanal, octadecanol, octade-
ꢀ
1
⁄
K
1
(mol%)
cene, octadecane and stearyl stearate. Sign in Scheme 2 denotes
*
the active sites on the metal surface, while steps R1, etc. are reac-
tion steps, while steps A1, etc. correspond to adsorption. The fast
Values are given with 95% confidence intervals.
0
steps are denoted as, for example, R1 .
along the four routes N(1)–N(4) are given. These numbers are
selected in a way, that the overall chemical equations do not con-
tain intermediates.
Since octadecene was not present in the reaction mixture in
experiments with the rhenium catalyst, its coverage can be consid-
ered low and further hydrogenation fast. Based on quasi-equilib-
rium adsorption of reactants and products, the following
expressions can be written, respectively, for active sites
The schemes are based on the theory of complex reactions
developed by Horiuti and extended by Temkin [17–19].
The elementary steps above can be described by four reaction
routes, i.e. sets of stoichiometric numbers of steps. Elementary
reactions are grouped in steps, and chemical equations of steps
contain reactants and intermediate products. A set of stoichiome-
tric numbers of steps is defined as a reaction route. Routes are
essentially different, and it is impossible to obtain one route
through multiplication of another route by a number, even in cases
when their respective overall equations are identical. The number
of basic routes, P, is determined by P = S + W ꢀ I, where S is the
number of steps, W is the number of balance (link) equations
and I is the number of intermediates. Such balance (link) equations
can correspond to the total coverage of sites equal to unity. On the
right hand side of equations for the steps, stoichiometric numbers
h
SA ¼ K
1
C
SA
h
V
;
h
AD ¼ K
2
C
AD
h
V
;
hAC ¼ K
3 AC V
C h ;
h
OD ¼ K
4
C
OD
h
V
ð1Þ
where K , etc. are the equilibrium constant of respective equilib-
rium steps A1–A4, CSA, etc., are concentrations and hV is the cover-
age of the vacant sites.
1
Scheme 3. Detailed reaction mechanism.

B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
205
(
a) 100
(b)100
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
9
8
7
6
5
4
3
2
0
0
0
0
0
0
0
0
10
0
0
50
100
150
200
250
300
350
0
50
100 150 200 250 300 350 400
time (min)
time (min)
c) 1
00
(
(d)100
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
9
8
7
6
5
0
0
0
0
0
40
30
20
10
0
0
50
100
150
200
250
300
350
0
50
100
150
200
250
300
350
time (min)
time (min)
100
(
e)
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
0
0
50
100 150 200 250 300 350
time (min)
Fig. 10. Comparison for experimentally observed data and model predictions for stearic acid, aldehyde and alcohol: (a) 180 °C, 2 MPa; (b) 200 °C, 2 MPa; (c) 220 °C, 2 MPa; (d)
2
00 °C, 3 MPa and (e) 200 °C, 4 MPa.
The reaction rate r along the first route N(1) is defined through
step R1 which can be considered as the one limiting the rate in this
route
ð1Þ
From the balance equations for different types of sites, coverage
of vacant sites can be easily computed
1
h
V
¼ 1
ð2Þ
1 1 SA H2
k K C C
ð1Þ
þ K
1
C
SA þ K
2
C
AD þ K
3
C
AC þ K
4
C
OD
r
¼ 1
ð3Þ
þ K
1
C
SA þ K
2
C
AD þ K
3
C
AC þ K
4 OD
C

2
06
B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
The rates along the second and the third routes are given in a
The values of activation energies for selective hydrogenation of
similar fashion
acids to alcohols are not readily available for Re-based catalysts,
thus a comparison can be made only with other types of catalysts,
such as homogeneous catalytic hydrogenation with Cu oleates giv-
ing 56 kJ/mol [21].
The kinetic model, presented above, was developed for the case
of first order in hydrogen, which implies weak adsorption of hydro-
gen on the catalyst surface. It can be thus probably applied for sim-
ilar catalytic systems as applied in the current work when, for
example, oxophilic metals are used on reducible oxides or even
non-reducible oxides. For reducible metals, such as Pt or a combi-
2 2 AD
k K C C
H2
ð2Þ
r
¼ 1
ð4Þ
ð5Þ
þ K
þ K
1
C
SA þ K
2
C
AD þ K
3
C
C
AC þ K
AC þ K
4
C
C
OD
3 3 AC
k K C
ð3Þ
r
¼ 1
1
C
SA þ K
2
C
AD þ K
3
4
OD
The route N( is considered as a non-catalytic, therefore
4)
ð4Þ
r
¼ k
4
C
AC
C
SA
ð6Þ
nation of reducible and oxophilic metals, such as ReOx–Pd/SiO
2
Initially, this step was considered as reversible; however, dur-
ing the numerical data fitting, it turned out that the backwards
reaction in the studied domain can be neglected.
catalyst utilized for hydrogenation of saturated fatty acids [22],
some modifications of the kinetic model might be needed to
address changes in hydrogen adsorption behaviour and possible
different kinetic regularities.
Finally, the generation rates of compounds can be written as:
dCSA
qBdt
dCAD
qBdt
dCAC
qBdt
ꢀ
¼ rð1Þ þ rð4Þ
;
¼ rð1Þ ꢀ rð2Þ
;
¼ rð2Þ ꢀ rð3Þ ꢀ rð4Þ
;
4. Conclusions
dCOD
qBdt
dCSS
dt
ð3Þ
¼ rð4Þ
¼
r
;
ð7Þ
B
q is the mass-of-catalyst-to-liquid-volume
x 2
Highly dispersed ReO /TiO was applied as a catalyst for the
selective hydrogenation of fatty acids to alcohols. The selectivity
to octadecane up to 93% was achieved in the temperature range
between 180–220 °C and 2–4 MPa of total pressure.
where t is time and
ratio.
The temperature dependence of the rate constants was
assumed according to the Arrhenius equation, while negligible
influence of temperature on KSA was supposed. The kinetic model-
ling was performed for all reaction rates and reaction sets together.
For the parameter estimation instead of absolute values of product
concentrations, their molar percentage was used, and a set of dif-
ferential equations describing the changes in the concentrations
profiles of the reagents and products with time was solved by
means of ModEst software [20]. Using Levenberg–Marquardt sim-
plex method, the target function, which was defined as incompli-
ance between the experimental and calculated values of
concentrations, was used to solve the system. The sum of the resid-
ual squares between the model and the experimental data was
minimized using the following objective function:
It was demonstrated that with a decrease of temperature and an
increase of pressure, a higher selectivity to the main product is
achieved. A high yield of octadecanol was obtained due to a strong
adsorption of the fatty acid compared to the fatty alcohol, which
inhibits a further transformation of the alcohol to alkanes. The
hydrocarbon can be treated as the only by-product of the reaction,
since stearyl stearate, which is also formed, can be further trans-
formed to the desired products. Moreover, with pre-reduced 4%
x 2
ReO /TiO , the selectivity to hydrocarbons is around 5%. However,
after a complete conversion of stearic acid, hydrocarbons were
mainly formed with octadecane as the main product.
The XPS studies confirm that the oxidation state of rhenium was
Re4+ and Re6+ during the experiments. In experiments with a non-
reduced rhenium catalyst, a lower conversion was obtained with a
higher formation of hydrocarbons, which was caused by a lower
amount of oxygen vacancies on the interface between rhenium
and titania.
XX
2
2
Q ¼ kxexp ꢀ xestk ¼
ðxexp;it ꢀ xest;it
Þ
ð8Þ
t
i
where xexp is the experimental value, xest denotes the predictions
given by the model, i is the component index and t is the time value.
The quality of the fit and accuracy of the model description was
The reaction kinetics was modelled based on the mechanistic
considerations for stearic acid hydrogenation. The calculated
results were compared with experimental data through numerical
data fitting, showing a very good correspondence was obtained
(with the degree of explanation close to 99.6%).
2
defined by the degree of explanation R ; which reflects comparison
between the residuals given by the model to the residuals of the
simplest model one may think of, i.e., the average value of all the
2
data points. The R value is given by the expression
Acknowledgments
2
ðymodel ꢀ yexperimentÞ
2
R ¼ 100
ð9Þ
This work is a part of activities of Åbo Akademi Process Chem-
istry Centre (PCC). The authors would like to acknowledge Markus
Peurla from Laboratory of Electron Microscopy at University of
Turku for performing TEM analysis and Sten Linholm from Labora-
tory of Analytical Chemistry at Åbo Akademi for performing ICP–
OES analysis. Moreover, financial support from Åbo Akademi Uni-
versity Endowment is gratefully acknowledged and EPSRC as part
of the CASTech grant (EP/G011397/1).
2
ðymodel ꢀ yꢀ experiment
Þ
Preliminary calculations demonstrated that the values of equilib-
rium constants K and K are small enough to neglect them. Since
the concentration of aldehyde is also low, the term K
neglected. Thus the values of k and k in fact are lumped ones, con-
taining also the value of K and K3, respectively.
3
4
2
CAD was
2
3
2
The values of the calculated frequency factors, activation energy
and adsorption constants as well as the estimated relative standard
errors (in%) of the tested reaction mechanism are presented in
References
Table 4. The values of pre-exponential constants for k
include the catalyst concentration.
1
–k
3
also
[1] D. Myers, Surfactant Science and Technology, John Wiley & Sons, New York,
2005.
[
2] C. Boswell, Detergent Alcohols Growing Capacity Will Continue to Drive down
Operating Rates, 2013.
Fig. 10 displays as, an example, an excellent comparison for
experimentally observed data and model predictions for reactants
in the studied range of pressures, concentrations and temperature
with the degree of explanation 99.72%. The standard errors of all
parameters are acceptably low.
[3] T. Turek, D.L. Trimm, N.W. Cant, Catal. Rev. 36 (1994) 645–683.
[
4] J. Carnahan, T. Ford, W. Gresham, W. Grigsby, G. Hager, J. Am. Chem. Soc. 77
1955) 3766–3768.
(
[
5] M. Toba, S. Tanaka, S. Niwa, F. Mizukkami, Z. Koppany, L. Guczi, K.-Y. Cheah, T.-
S. Tang, Appl. Catal. A: Gen. 189 (1999) 243–250.

B. Rozmysłowicz et al. / Journal of Catalysis 328 (2015) 197–207
207
[
[
[
[
6] M. Mendes, O. Santos, E. Jordão, A. Silva, Appl. Catal. A: Gen. 217 (2001) 253–
62.
7] H.G. Manyar, C. Paun, R. Pilus, D.W. Rooney, J.M. Thompson, C. Hardacre, Chem.
Commun. 46 (2010) 6279–6281.
[13] M.S. Nacheff, L.S. Kraus, M. Ichikawa, B.M. Hoffman, J.B. Butt, W.M. Sachtler, J.
Catal. 272 (1987) 263–272.
[14] B. Rozmysłowicz, P. Mäki-Arvela, A. Tokarev, A.-R. Leino, K. Eränen, D.Y.
Murzin, Ind. Eng. Chem. Res. 51 (2012) 8922–8927.
[15] M. Komiyama, J. Sato, K. Yamamoto, Y. Ogino, Langmuir 3 (1987) 845–851.
[16] M.I. Temkin, D.Yu. Murzin, N.V. Kul’kova, Dokl. AN USSR 303 (1988) 659–662.
[17] J. Horiuti, J. Res. Inst. Catal. Hokkaido Univ. 5 (1957) 1–26.
[18] J. Horiuti, T. Nakamura, Z. Phys. Chem., Neue Folge 11 (1957) 358–365.
[19] M.I. Temkin, Dokl. AN USSR 152 (1963) 156–159.
2
8] K. Yoshino, Y. Kajiwara, N. Takaishi, Y. Inamoto, J. Tsuji, J. Am. Oil Chem. Soc. 67
(
1990) 21–24.
9] T.-S. Tang, K.-Y. Cheah, F. Mizukami, S. Niwa, M. Toba, Y.-M. Choo, J. Am. Oil
Chem. Soc. 70 (1993) 601–605.
[
[
10] J.P. McGeer, H.S. Taylor, J. Am. Chem. Soc. 73 (1951) 2743–2751
11] A. Fung, P. Tooley, M.J. Kelley, D.C. Koningsberger, B.C. Gates, J. Phys. Chem. 95
[20] H. Haario, ModEst 6.0 – User Guide, 2010.
(
1991) 225–234.
[21] J.D. Richter, P.J. van den Berg, J. Am. Oil Chem. Soc. 46 (1969) 158–162.
[22] Y. Takeda, Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 2 (2012) 2221–2223.
[
12] J. Okal, Appl. Catal. A: Gen. 287 (2005) 214–220.