Deoxygenation of heptanoic acid to hexene over cobalt-based catalysts: A model study for α-olefin production from renewable fatty acid

Article abstract of DOI:10.1016/j.apcata.2020.117644

Deoxygenation of heptanoic acid, a model compound, over bimetallic cobalt (Co-Pt, Co-Au, Co-Pd, Co-Ru) supported silica catalysts, was examined for α-olefin production. The catalysts were prepared by conventional impregnation of the metal precursors on silica and characterized by XRF, TEM, H₂-TPR, acetic acid-TPD, and XANES. Catalytic testing was performed in a fixed-bed flow reactor under atmospheric H₂ pressure. Monometallic cobalt catalysts yielded mainly 1-hexene, but rapid deactivation was observed. Incorporation of 0.5percentwt secondary metal, particularly Pt, increases activity and stability under H_2. A relatively higher olefin/paraffin ratio can be obtained from the reaction over 5percentCo+0.5percentPt/SiO₂ when compared to that with higher Pt loading. The co-impregnation method offers Co-Pt catalysts with stability higher than that prepared by the sequential impregnation method. Over cobalt-based catalysts, the deoxygenation is proposed to proceed via reduction of heptanoic acid to heptanal that is an intermediate for decarbonylation to hexene; while other side reactions are suppressed.

Full text of DOI:10.1016/j.apcata.2020.117644

Journal Pre-proof
Deoxygenation of heptanoic acid to hexene over cobalt-based catalysts:
A model study for ␣-olefin production from renewable fatty acid
Ploynisa Phichitsurathaworn (Investigation) (Writing - original draft),
Kittisak Choojun (Writing - review and editing) (Supervision)
(Funding acquisition), Yingyot Poo-arporn (Investigation) (Data
curation), Tawan Sooknoi (Conceptualization) (Writing - review and
editing) (Supervision) (Funding acquisition)
PII:
S0926-860X(20)30237-4
DOI:
https://doi.org/10.1016/j.apcata.2020.117644
APCATA 117644
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
17 March 2020
12 May 2020
18 May 2020
Please cite this article as: Phichitsurathaworn P, Choojun K, Poo-arporn Y, Sooknoi T,
Deoxygenation of heptanoic acid to hexene over cobalt-based catalysts: A model study for
␣-olefin production from renewable fatty acid, Applied Catalysis A, General (2020),
doi: https://doi.org/10.1016/j.apcata.2020.117644

This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
© 2020 Published by Elsevier.

Deoxygenation of heptanoic acid to hexene over cobalt-based
catalysts: A model study for -olefin production from
renewable fatty acid
Ploynisa Phichitsurathaworn a, Kittisak Choojun a,b *, Yingyot Poo-arpornc,
and Tawan Sooknoi a,b**
aDepartment of Chemistry, Faculty of Science, King Mongkut’s Institute of Technology
Ladkrabang (KMITL), Bangkok 10520, Thailand
bCatalytic Chemistry Research Unit, Faculty of Science, King Mongkut’s Institute of
Technology Ladkrabang (KMITL), Bangkok 10520, Thailand
cSynchrotron Light Research Institute (Public Organization), 111 University Avenue, Muang
District, Nakhon Ratchasima 30000, Thailand
* Corresponding author email a: kittisak.ch@kmitl.ac.th,
** Corresponding author email b: kstawan@gmail.com
Tel.: + 66 8 1929 8288; Fax: + 66 2 326 4415
Graphical abstract

Highlights
 Strong feed adsorption leads to rapid deactivation over Co based catalyst.
 Incorporation of 0.5%wt Pt in 5%wtCo/SiO2, increases activity and stability.
 Higher olefin/paraffin ratio is obtained over Co-Pt/SiO2 under atmospheric H2.
 DeCO activity/stability w/o excessive hydrogenation can be tuned over Co-Pt/SiO2.
Abstract
Deoxygenation of heptanoic acid, a model compound, over bimetallic cobalt (Co-Pt,
Co-Au, Co-Pd, Co-Ru) supported silica catalysts, was examined for -olefin production. The
catalysts were prepared by conventional impregnation of the metal precursors on silica and
characterized by XRF, TEM, H2-TPR, acetic acid-TPD, and XANES. Catalytic testing was
performed in a fixed-bed flow reactor under atmospheric H2 pressure. Monometallic cobalt
2

catalysts yielded mainly 1-hexene, but rapid deactivation was observed. Incorporation of
0.5%wt secondary metal, particularly Pt, increases activity and stability under H2. A relatively
higher olefin/paraffin ratio can be obtained from the reaction over 5%Co+0.5%Pt/SiO2 when
compared to that with higher Pt loading. The co-impregnation method offers Co-Pt catalysts
with stability higher than that prepared by the sequential impregnation method. Over cobalt-
based catalysts, the deoxygenation is proposed to proceed via reduction of heptanoic acid to
heptanal that is an intermediate for decarbonylation to hexene; while other side reactions are
suppressed.
Keywords: Deoxygenation; Heptanoic acid; Cobalt; -olefins; bimetallic catalysts, Co-Pt
1. Introduction
Linear alpha olefins (LAOs) is one of the important intermediate chemicals widely used
as plasticizers, synthetic lubricants, surfactants, synthetic drilling fluids, detergents, and co-
monomers [1–3]. Industrially, linear alpha olefins are commonly produced by the
oligomerization of ethylene, which is derived from non-renewable feedstocks [4]. While
decarbonylation of fatty acid is an attractive approach for -olefin production from renewable
sources [5–7].
In general, conversion of fatty acids to olefinic or paraffinic hydrocarbons can be
proceeded via catalytic decarbonylation [6,8,9] or decarboxylation [6,10–12], respectively.
However, other side reactions, such as ketonization [13], and cracking [11] can be competitive.
Therefore, the design of the catalyst and reaction conditions are of particular importance.
Metals supported on silica catalysts, such as platinum [14–18], palladium [14,19,20] and
rhodium [11,21], are normally used in a semi-batch reactor. Murzin, et al. reported that the
activity of the supported metal on carbon catalysts was in the order of Pd > Pt > Rh > Ir > Ru
> Os > Ni for the deoxygenation of stearic acid at 300˚C and 6 bar of H2 [11]. However, under
3

this high pressure of H2, paraffin yields are higher than that of olefins, due to rapid
hydrogenation of the olefins primarily formed. In a similar manner, the conversion of fatty
acid over Pd/C, catalysts provides high paraffinic hydrocarbons in a continuous flow reactor
[5].
A severe deactivation was typically observed for the conversion of fatty acid over metal
catalysts [11,14,22]. This is mainly derived from the strong interaction of fatty acid over the
active metal surface. Therefore, high hydrogen pressure is generally used in the reaction system
[11,14,22,23]. This could prolong the catalyst stability, but the product selectivity is shifted to
paraffinic hydrocarbons. Thus, minimization of (i) the hydrogenation activity of the catalyst
and (ii) the feed interaction with the catalyst’s surface, is the key challenge in this reaction.
With this regard, non-noble oxophillic metals (Ni, Cu, Co, and Cr over SiO2 support)
are good candidates due to their limited hydrogenation activity, as compared to the noble metal
(Pt, Pd, etc.) [24]. Amongst those, cobalt exhibits high initial activity and selectivity to olefins.
However, large cobalt particles possess a strong interaction with a carboxylic acid, which leads
to successive adsorption of the feed on the active sites, resulting in a rapid deactivation.
In this work, we thus further investigate the effects of cobalt particle size and
incorporated secondary metals (Ru, Pd, Pt, and Au). This aims to modify the feed adsorption
vs. hydrogenation over the cobalt active site in the deoxygenation of heptanoic acid (a model
compound of fatty acids) to -olefins under atmospheric H2 flow. The interaction of the feed
on the active sites under H2 is evaluated by acetic acid-temperature-programmed desorption
(acetic acid-TPD). The electronic property of the cobalt active sites is studied by X-ray
absorption near edge structure (XANES) and temperature-programmed reduction (TPR).
Deoxygenation activity to the corresponding -olefins with minimal hydrogenation of the -
olefins produced, is deliberately tuned. Besides, methods of bimetallic catalyst preparation (co-
and sequential impregnation) are verified.
4

5

2. Experimental procedure
Silica (99.0%; Fluka) was calcined before use, in a muffle furnace at 600˚C for 1 h with
a heating rate of 5˚C/ min. Co/SiO2 (2, 5, 10, 15%wt) and 5%Co-X%M/SiO2 (X = 0.25, 0.5,
0.75 and M = Pt, Pd, Au, Ru) were prepared by impregnation (IMP) and co-impregnation (CIP).
For IMP, solution of cobalt precursor (Co(NO3)2.6H2O; Carlo) or a mixture of Co and M
precursors (Pd(OAc)2, [(p-cymene)RuCl2]2, AuCl3.3H2O; Sigma-Aldrich, Pt(NH3)4(NO3)2;
Ramken) was sprayed on the calcined silica. After that, the solid was dried in an oven at 70˚C
for 24 h and then calcined in a tube furnace at 600˚C for 1 h with a heating rate of 5˚C/min.
The sequential impregnation method was used to prepare 5%Co+0.5%Pt/SiO2 (SIP). Briefly,
the 5%Co/SiO2 (IMP) was repeatedly impregnated by the solution of Pt(NH3)4(NO3)2. The
solid was dried in an oven at 70˚C for 6 h and calcined at 600˚C for 1 h.
Specific surface area, chemical composition, and morphology of the catalyst were
determined by nitrogen gas adsorption, wavelength-dispersive X-ray fluorescence
spectrometry (WD-XRF), and transmission electron microscopic (TEM), respectively. The
reducibility of the incorporated metals was examined by a temperature-programmed reduction
(H2-TPR), equipped with a thermal conductivity detector (VICI). The H2 consumption was
calculated using CuO as a standard. The adsorption of acetic acid on the catalysts was studied
by temperature-programmed desorption (acetic acid-TPD) under He and H2. A portion of
samples (~20 mg) was flushed with He at 300˚C for 1 h (50 mL/min) to remove the absorbed
moistures, and then treated with acetic acid vapor at 140˚C for 30 min using a bubble flow
vapor generator (glacial acetic acid) at room temperature. After the adsorption, the excess
acetic acid was flushed with He or H2 at 140˚C. The acetic acid-TPD signal was then detected
from 140˚C to 900˚C under He or H2. In situ XANES experiments were performed at BL2.2:
TR XAS, Synchrotron Light Research Institute (SLRI), Thailand. An energy-dispersive
monochromator and position-sensitive detector were employed to perform XANES
6

measurement [25]. The Co K-edge XANES was collected using the integration time of 2000
ms with an average of 10 scans. The oxidation state of cobalt composition was estimated by
Linear Combination Fitting (LCF) of XANES spectra over Athena software using Co foil,
CoO, and Co3O4 as a standard. A thermogravimetric analyzer (Pyris) was used to quantify coke
content on the used catalysts.
Deoxygenation of heptanoic acid was investigated at atmospheric pressure in a fixed-
bed flow reactor made with a quartz tube (6.3 mm O.D.). The catalyst was primarily activated
at 600˚C (heating rate of 5˚C/min) for 1 h under the stream of air (100 mL/min). N2 was purged
for half an hour, and then the gas stream was switched to H2 (100 mL/min) at the same
temperature for 3 h. After reduction, the furnace was cooled down to the designed reaction
temperature (350 – 425˚C) and solution of reactant (10%wt./wt. heptanoic acid (Carlo) in
octane (Sigma-Aldrich)) was fed to the reactor by a syringe-pump. The reaction was carried
out for 6 h on stream under the stream of H2 (100 mL/min). The gas sample was collected in a
gas sampling loop, then periodically injected into GC column (MXT-1, 60 m length, 0.25 mm
internal diameter, 0.5 µm film thickness) equipped with a flame ionized detector (FID).
7

3. Results & discussion
3.1 Catalyst characterization
Metal loadings and surface area of the catalysts are summarized in Table 1. The
observed metal loadings are in agreement with the desired catalyst preparation. Comparing
with SiO2 support (319 m2/g), a proportional decrease in surface area (226-286 m2/g) is
observed with the increase in metal loadings, presumably due to a partial surface/pore blockage
by the loaded metal species.
Table 1 Metal loadings, surface area, H2 consumptions and metal dispersion of the catalysts.
Metalb
conten
t
H2
Dispersio
nc
Elemental
analysis (XRF)
SBET
consumptio
n
Enatr
y
Catalysts
Co Secondar
(%wt. y metal
mmol/g
(%wt.
)
(%)
(m2/g
)
)
(%wt.)
SiO2
-
-
-
-
319
286
277
-
1
2
3
2%Co/SiO2
5%Co/SiO2
1.9
4.8
0.5
1.3
2.4
5.5
22.5
6.7
2.4
2.4
10%Co/SiO2
15%Co/SiO2
9.2
-
-
236
225
2.2
3.2
9.8
4
5
13.7
14.1
5%Co+0.5%Pd/SiO
4.4
4.6
4.5
4.4
0.6
252
254
251
253
1.1
1.3
1.4
1.3
4.7
5.6
6.0
5.6
n/a
n/a
6
7
8
9
2
5%Co+0.5%Au/SiO
0.5a
0.7a
0.5
2
n/a
n/a
n/a
n/a
5%Co+0.5%Ru/SiO
2
5%Co+0.5%Pt/SiO2
(CIP)
5%Co+0.5%Pt/SiO2
4.9
4.6
0.5
0.2
227
249
1.3
1.2
5.8
5.3
10
11
(SIP)
5%Co+0.25%Pt/Si
O2
8

5%Co+0.75%Pt/Si
O2
n/a
4.6
0.7
226
1.3
5.7
12
c
a SEM-EDX bCalculated from H2 Consumption, Calculated
from H2 Chemisorption
Figure 1 Temperature programmed reduction profiles of catalysts.
The reducibility of the catalysts can be deduced from the H2-TPR, as shown in Figure
1. The hydrogen consumption is also collected in Table 1. For all cobalt catalysts (Figure 1, a-
d), the only peak at ~340˚C is observed and attributed to the reduction of Co3O4 to CoO species
and CoO to metallic Co [26]. When cobalt loading is increased from 2% to 15%wt., the H2
consumption is increased proportionally (Table 1, entry 2-5 and Figure S1 supporting
information). The higher reduction temperature (320-340˚C) for 10% and 15 %wt. Co/SiO2,
9

compared with those for 2% and 5%wt. Co/SiO2 (300-320˚C), suggests a larger cobalt particle
at high loadings [27,28], as confirmed by TEM (Figure 2). It can be seen that the average cobalt
particle size is approximately 18, 27, 41, and 52 nm for 2%, 5%, 10%, and 15%wt. loadings,
respectively.
10

Figure 2 TEM images and particle size distribution histograms of reduced catalysts:
a) 2%Co/SiO2, b) 5%Co/SiO2, c) 10%Co/SiO2, d) 15%Co/SiO2.
11

Upon incorporation of the secondary metals, the H2 consumption of bimetallic Co
catalysts (1.1-1.4 mmol/g) is similar to that of monometallic Co (~1.3 mmol/g). This is because
the only small amount of the secondary metals is incorporated (~0.5 %wt.). However, the
reduction peaks of the bimetallic catalysts (Figure 1, e-h) generally shift to lower temperatures,
as compared to the monometallic cobalt catalyst. The first reduction at ~120˚C is assigned to
the reduction of Co-Pt mixed oxide, presumably forming a bimetallic Co-Pt. While, the second
and third peaks (130 and 240˚C) suggest the reduction of Co3+ to Co2+ and Co2+ to Co0,
respectively [29]. The presence of Pt facilitates the reduction of nearby cobalt oxide particles
presumably via hydrogen spilled-over [30,31]. Hence, the incorporated Pt significantly
decreases the reduction temperature of cobalt, as observed.
Similarly, 5%Co+0.5%Pd/SiO2 contains three reduction peaks at 115, 135, and 280˚C
(Figure 1f) [32]. Even though Co-Pd mixed oxide is reduced at a lower temperature compared
with Co-Pt, the reduction of the remaining cobalt species in this sample requires a higher
temperature. This indicates the lower activity of Pd for H2-spilled-over, as compared to the Pt.
Unlike 5%Co+0.5%Pt/SiO2 and 5%Co+0.5%Pd/SiO2, 5%Co+0.5%Ru/SiO2 shows two
reduction peaks at 155 and 245˚C, referring to the reduction of the Co-Ru mixed oxide to
bimetallic Co-Ru and the remaining Co2+ to Co0, respectively [33]. On the other hand,
5%Co+0.5%Au/SiO2 shows only a single reduction peak at 300˚C. It is suggested that Au may
well promote the reduction of Co3+ to metallic Co in a single step.
12

3.2 Catalysts activity
Effect of temperature over 5%Co/SiO2 catalysts
The conversion of heptanoic acid using 5%Co/SiO2 as a catalyst is enhanced (total yield
from 22% to 56%) upon increasing the reaction temperature (from 350˚C to 425˚C), as shown
in Figure 3. 1-Hexene can be observed as the main product; while, minors include heptanal,
7-tridecanone and other lighter hydrocarbons. Yields of all products are also increased with
temperature, except for heptanal that may well undergo decarbonylation to hexene at high
temperature [33,34]. Unlike other report on transfer hydrogenation of fatty acid over cobalt
catalysts [35], heptanol was not detected. This was presumably because a higher temperature
was used in this study; while other report in the liquid phase was tested at relatively lower
temperature (~200˚C) in the presence of isopropanol as hydrogen donor. However, at 425˚C
hexene was decreased; while, methane and 7-tridecanone were notably enhanced. This suggests
that, at high temperatures, hydrogenolysis of the light hydrocarbons and ketonization of
heptanoic acid could be promoted over the cobalt catalyst [36–38]. Besides, methane could be
formed by the hydrogenation of CO and CO2 produced by decarbonylation and decarboxylation
of heptanoic acid, respectively [33,36]. In a different manner from the reaction over noble metal
catalysts i.e. Pd/SiO2 [39] and PdSn/SiO2 [40], higher selectivity of linear -olefin were
obtained in this study without iso-olefins. This could attribute to the weaker adsorption of
olefinic product over Co, as compared to that over Pd. Hence further hydrogenation and/or
double bond/skeleton isomerization of the initially formed linear olefins are suppressed in this
work.
13

Figure 3 The effect of the reaction temperature on the conversion of heptanoic acid (total yields)
and yield of products from the deoxygenation of heptanoic acid over 5%Co/SiO2. (Conditions:
10%wt. heptanoic acid in octane, contact time 20 gh.mol-1, activation temperature 550˚C and reduction
temperature 500˚C, reaction temperature 350 – 425˚C, 1 atm, and 100 mL/min of H2. The activity is at 30 min
on stream.)
14

Effect of cobalt loading on the silica support
With the same contact time based on cobalt, the conversion at 1 h on stream is decreased
(from 84.7% to 49.0%) when cobalt loading is increased (from 5%wt. to 15%wt.), as shown in
Table 2, entry 2-4, and Figure 4. This is presumably due to a lower dispersion of cobalt particles
at higher loading, as seen from Table 1 (entry 2-5) and Figure 2. However, 2%Co/SiO2 gives
lower conversion despites its higher metal dispersion (~22% dispersion). This could be derived
from the severe deactivation of this catalyst, as observed in Figure S2 (supporting information).
The hexene/hexane ratio is generally improved (21.1 to 53.0, Table 2) with cobalt loading (5%-
15%), possibly due to the suppression of hydrogenation activity when Co dispersion is
decreased (from 6.7% to 2.4 %). It is also seen that over 15%Co/SiO2, the yield of 7-
tridecanone is notably high (Figure 4). This is presumably because, at high Co loading, some
of the non-reduced cobalt oxides may well be retained even after reduction at 600˚C. In line
with this view, a high reduction temperature (>600˚C) was particularly observed in this sample
(Figure S3 in supporting information). Such non-reduced cobalt oxide species presenting in
15%Co/SiO2 can readily promote the ketonization of heptanoic acid to 7-tridecanone, as
observed [38].
15

Table 2 Deoxygenation of heptanoic acid and product distribution of various catalysts
Contact
time
%Yield
(g.h/mol
)
Entr
Catalysts
y
13.
3 4
2.
4
22.
6
29.
46.
2%Co/SiO2
1.64 82 57.2
1.62 33 84.7
1.3 0.0
2.2 2.2
1.4 4.8
0.5 0.6
4.2 6.5
3.5 7.7
0.6 5.0
1
2
3
4
5
6
7
8
19.
1 4
3.
6
21.
1
5%Co/SiO2
3.
5
38.
4
53.
7
10%Co/SiO2
15%Co/SiO2
5%Co+0.5%Pt/SiO2
1.65 17 78.6 9.5
1.63 11 49.0 6.1
1.
9
10.
53.
0
26.
5
2.6
8
100. 11.
1.69 34
17. 14. 2.
54.
0.0 0.0
3.1 7.5
3.0
0
2 4
9
3 2
2.
5%Co+0.5%Au/Si
O2
13.
2 1
25.
7
39.
1.69 34 69.1
1.5 1.8
3.3 1.5
1.0 1.2
8
5%Co+0.5%Pd/SiO
2. 15.
3 3
11.
0
36.
6
1.69 34 75.7 9.1
1.68 34 57.5 8.7
7.6
2
5%Co+0.5%Ru/Si
O2
2.
1
35.
9
37.
2
3.0 4.3
2.1 0.0
3.2 0.0
2.9 1.5
2.9 0.0
5%Co+0.5%Pt/SiO2
100. 16.
1.69 34
15. 11. 2.
52.
3.5
0.1
4.8
3.0
9
0
2 5
0
6 6
(c-SIP)
5%Co+0.5%Pt/SiO2
100.
0
71. 14. 0.
1.69 34
3.3
14.
10
11
12
6.9
52.
9
6 0
11. 10. 2.
4 6
17. 14. 2.
(nc-SIP)
5%Co+0.25%Pt/Si
O2
1.70 34 95.7
1.68 34 100
9 4
0
5%Co+0.75%Pt/Si
O2
11.
4 2
52.
4
1 0
(Condition: 10%wt. heptanoic acid in octane, reaction temperature 400˚C,
activation temperature 600˚C and reduction temperature 600˚C, 1 atm, and
16

100 mL/min of H2. The activity is at 1 h on stream. (5%Co+0.5%Pt/SiO2 (nc-
SIP) is not in situ calcined).
Figure 4. Effect of Co loading (wt%) of cobalt catalysts on deoxygenation of heptanoic acid.
(Condition: 10%wt. heptanoic acid in octane, reaction temperature 400˚C, activation temperature 600˚C
and reduction temperature 600˚C, 1 atm, and 100 mL/min of H2. The activity is at 1 h on stream)
Catalyst deactivation
Although 5%Co/SiO2 provided high activity amongst other loadings, the deactivation
was notably observed. All the product yields are decreased, except for 7-tridecanone, as shown
in Figure 5. The cause of catalyst deactivation could be the results of (i) strongly adsorbed
heptanoic acid or (ii) coke formation on the Co surface. To verify the nature of adsorbed
species, TGA of the used catalysts shows only major weight loss ~10% at 330˚C, in addition
to ~2%wt. absorbed moisture (50-100˚C), as depicted in Figure S4 (see supporting
information). This low temperature weight loss may be due to the decomposition of adsorbed
heptanoic acid on the cobalt surface rather than the coke deposition. To validate this
assumption, the heptanoic acid was paused after 6 h on stream. Subsequently, the catalyst bed
17

was flushed with steam at the reaction temperature (400˚C) for another 6 h (6-12 h on stream
in Figure 5) before the feed was re-introduced. It can be seen that, from 12-15 h on stream, the
activity is recovered with similar product distribution. This experiment clearly shows that
steaming can displace the feed that is adsorbed on the surface of catalysts. If the coke deposit
is the cause of the catalyst deactivation, the activity shall not regain after steaming.
Accordingly, the observed deactivation appears to derive from the strong adsorption of
heptanoic acid on the cobalt active site. In a supportive manner, a significantly stronger TPD
signal was observed (Figure S5) over 5%Co/SiO2, as compared to the parent support, after
these samples was pre-adsorbed with acetic acid (a smaller carboxylic acid probe). Such
signals are derived from ketonization of the chemisorbed acetic acid as reported previously
[41]. This suggests that the presence of Co promotes both chemisorption and ketonization of
carboxylic acid at elevated temperature.
Figure 5 Conversion of heptanoic acid and yield of products from the deoxygenation of
heptanoic acid over 10%Co/SiO2. (Conditions: 10%wt. heptanoic acid in octane, contact time 17
18

gh/mol, activation temperature 600˚C and reduction temperature 600˚C, reaction temperature 400˚C, 1
atm, and 100 mL/min of H2. ; () conversion ; (×) C1 ; () C2 ; () C3 ; () C4 ; (-) C5 ; (.) Hexene ; (□)
Hexane ; () iso-C6 ; (+) Heptene ; () Heptanal ; () 7-tridecanone)
19

Effect of secondary metal incorporated into 5%cobalt catalysts
To improve catalyst stability, 0.5%wt. secondary metals (Pt, Pd, Ru, and Au) were
incorporated in 5%Co/SiO2. As seen from Figure 6, the stability 5%Co+0.5%Au/SiO2 is not
modified because desorption/decomposition of the carboxylic acid is not facilitated even under
H2 (see Figure 7c), as compared to that observed with 5%Co/SiO2 (Figure 7a). It is worth noting
here that the peak area of acetic acid-TPD would refer to the amount of chemisorbed acetic
acid on the catalyst. While the peak temperature would suggest the desorption/ decomposition
of such chemisorbed acetic acid. Accordingly, the catalyst exhibiting signal at lower
temperature and/or with smaller peak area, would be generally less adsorptive for carboxylic
acid. Accordingly, similar product distribution to 5%Co/SiO2 was observed over
5%Co+0.5%Au/SiO2 (Table 2, entry 6). The stability of 5%Co+0.5%Pd/SiO2 is not improved,
despite a facile H2 spilled-over from Pd to Co surface, as implied by the lower reduction
temperature for this sample (Figure 1). This could be the result of the strong interaction with
the carboxylic acid over Co-Pd/SiO2, as deduced from the high decomposition temperature of
the adsorbed acetic acid (Figure 7d). A noticeable yield to heptanal was obtained over this
catalyst (Table 2, entry 7), suggesting that the bimetallic Co-Pd also promotes the reduction of
heptanoic acid to heptanal. Although, in the case of 5%Co+0.5%Ru/SiO2, the high selectivity
of hexene was obtained (Table 2, entry 8), the catalyst is rapidly deactivated, possibly due to
the CO poisoning (that is produced from the decarbonylation of heptanal), as reported in the
literature [42,43].
20

Figure 6 Catalytic performance on deoxygenation of various cobalt/metal bimetallic supported
on silica catalysts. (Condition 10%wt. heptanoic acid in octane, contact time 34 gh/mol, reaction
temperature 400˚C, activation temperature 600˚C and reduction temperature 600˚C, 1 atm, and 100
mL/min of H2)
Figure 7 Acetic acid decomposition of 5%Co/SiO2 and 5%Co/metal bimetallic catalysts:
helium atmosphere (solid lines) and hydrogen atmosphere (square dot).
21

In sharp contrast, Co-Pt/SiO2 provides an improved stability, presumably because Pt
can facilitate desorption of the heptanoic acid from Pt-Co surface, as seen from acetic acid-
TPD in Figure 7b. It is clear that no acetic acid desorption/decomposition was detected at a
temperature higher than 300˚C, over Co-Pt/SiO2 under H2, while some adsorbed species were
retained over other catalysts. This suggests that H2 can assist the desorption of carboxylic acid
more efficiently over Co-Pt/SiO2, possibly due to an improved H2 spilled-over on the catalyst
surface. In line with this view, the in situ TR-XANES measurement of Co-Pt catalysts under
H2 from 50 to 600˚C revealed that the Co species could be more readily reduced, as compared
to 5%Co/SiO2 catalyst. It can be seen from the linear combination fitting of XANES spectra in
Figure 8 that the metallic Co species could be formed in 5%Co+0.5%Pt/SiO2 at a relatively
much lower temperature (~100˚C), as compared to 5%Co/SiO2 catalyst (~200˚C). The
complete reduction of 5%Co+0.5%Pt/SiO2 can be observed at 300 ˚C; while, less than 15%
metallic Co is present in 5%Co/SiO2 at this temperature (Figure 8). This emphasizes that the
addition of Pt clearly facilitates the reduction of Co species. Vajda, et al. reported that the
addition of 25%wt. Pt into Co would result in a higher fraction of CoO in the catalyst, as
compared to the bulk Co oxides [44]. Similarly, ~20%CoO was observed in
5%Co+0.5%Pt/SiO2, but Co3O4 was solely contained in 5%Co/SiO2. In consistence with this
result, the reduction peak (TPR) of 5%Co+0.5%Pt/SiO2 shifts to the low temperature (Figure
1e).
22

Figure 8 Fraction of Co components from linear combination fitting of a) 5%Co/SiO2 and b)
5%Co+0.5%Pt/SiO2 catalysts under reduction condition from 50 to 600˚C using CoO and
Co3O4 bulk standards.
23

Even though 5%Co+0.5%Pt/SiO2 shows similar hexene selectivity (~54%, Table S1,
entry 5) as compared to 5%Co/SiO2, hexane and iso-C6 selectivities are particularly increased
(from 2.6 and 2.5 to 17.9 and 14.3%, respectively). This could be attributed to the fact that the
incorporated Pt could promote hydrogenation of hexene to hexane, decarboxylation of
heptanoic acid to hexane, and also isomerization of hexane to iso-C6 [16]. In accordance,
hexane and iso-C6 are obtained only at high contact time (Figure 9). While hexene, heptanal,
and 7-tridecanone are observed initially (≤12 gh/mol). The major product, hexene, can be
produced via the decarbonylation of heptanal that is primarily formed via the reduction of
heptanoic acid [45]. While 7-tridecanone can be alternatively produced via ketonization of
heptanoic acid. The C1-C5 hydrocarbons, minor products, were derived from cracking of the
C6-C7 obtained initially. The overall reaction pathway for the deoxygenation of heptanoic acid
over 5%Co+0.5%Pt/SiO2 catalyst can be proposed in Figure S6.
Figure 9 a) Contact time using 5%Co+0.5%Pt/SiO2 as a catalyst for the heptanoic acid
conversion b) Zoom in other product yields. (Condition: 10%wt. heptanoic acid in octane, reaction
temperature 400˚C, activation temperature 600˚C and reduction temperature 600˚C, 1 atm, and 100 mL/min of
H2. ; () conversion ; (×) C1 ; () C2 ; () C3 ; () C4 ; () C5 ; () Hexene ; (.) Hexane ; (□) iso-C6 ; (+)
Heptene ; () Heptanal ; () 7-tridecanone)
24

As the Pt loading is increased (0.25-0.75%wt.), C1-C5 and 7-tridecanone yields were
suppressed, as compared to the 5%Co/SiO2 (Table 2, entry 2, 5, 11, 12, and Figure 10),
presumably due to the increased electron density of bimetallic Co-Pt, which minimizes the
cracking and ketonization activity. In addition, the stability is improved with the Pt loading, as
seen in Figure 10. This is presumably because the increased Pt content in the bimetallic Co-Pt
surface would increase the H2 spilled-over, which assists the desorption of heptanoic acid from
the bimetallic surface. In a supportive manner, the reduction of the bimetallic catalysts was
shifted to lower temperature along with the increase in Pt loading (see Figure S7). However,
as Pt loading is increased to 0.75%wt., hexane and iso-C6 yields are slightly increased, causing
the decrease in hexene/hexane ratio as Pt readily promotes decarbonylation and hydrogenation
(Table 2, entry 11, 5 and 12).
Figure 10. Effect of Pt loading (wt%) over 5%Co/SiO2 for deoxygenation of heptanoic acid.
(Condition: 10%wt. heptanoic acid in octane, contact time 34 gh/mol, reaction temperature 400˚C,
activation temperature 600˚C and reduction temperature 600˚C 1 atm, and 100 mL/min of H2)
25

Figure 11. Catalytic performance of deoxygenation of heptanoic acid using Co-Pt bimetallic
catalysts. (Condition: 10%wt. heptanoic acid in octane, contact time 34 gh/mol, reaction temperature
400˚C, activation temperature 600˚C and reduction temperature 600˚C 1 atm, and 100 mL/min of H2,
(5%Co+0.5%Pt/SiO2 (nc-SIP) is not in situ calcined))
26

When the 5%Co+0.5%Pt/SiO2 was prepared by sequential impregnation (SIP), lower
stability with similar product distribution was observed, as compared to the catalyst prepared
by co-impregnation. (Figure 11, Table 2, entry 5 and 9). Such lower stability might be the result
of the lower fraction of bimetallic Co-Pt prepared by the sequential impregnation method (SIP),
as compared to the co-impregnation (CIP). This is deduced from an early reduction temperature
of Pt species in 5%Co+0.5%Pt/SiO2 (c-SIP), while the reduction of Co2+ to Co0 is observed at
a higher temperature. (Figure S7). Such phenomenon suggests a disaggregation of Pt from the
Co particles upon reduction, presumably due to poor mixing of Pt and Co precursors during
SIP preparation. The disaggregation of the Pt particle becomes more severe when the catalyst
is reduced without calcination. The H2-TPR of 5%Co+0.5%Pt/SiO2 (nc-SIP) shows a small
peak at 128˚C indicating the reduction of platinum (II) nitrate and higher reduction temperature
peaks of Co3+ to Co2+ and Co2+ to Co0 at 178 and 314˚C (Figure S7). Accordingly, a high
activity (Figure 11) with exceptionally high hexane selectivity (Table 2, entry 10) was observed
in the non-calcined catalyst, 5%Co+0.5%Pt/SiO2 (nc-SIP). This is because the isolated Pt
successively promotes the decarboxylation of heptanoic acid yielding hexane and its isomers.
27

4. Conclusion
-Olefin (1-hexene) can be selectively produced from deoxygenation of heptanoic acid,
a fatty acid model compound, over supported cobalt-based catalysts. Cobalt particle size plays
a significant role in the catalytic activity (optimum at ~30 nm), despites that deactivation is
observed for all catalysts. The strong feed adsorption on the Co active sites leads to the catalyst
deactivation. The incorporation of secondary metals, namely Au, Pd, Ru, facilitates the
reduction of cobalt oxides. However, the catalyst stability is not readily improved even under
H2. In sharp contrast, bimetallic 5%Co+0.5%Pt/SiO2 provides a higher stability presumably
due to a weaker interaction of the carboxylic acid with the Co-Pt bimetallic surface. The in situ
TR-XANES shows a lower edge energy of Co when Pt is incorporated. This indicates a higher
electron density of Co-Pt bimetallic catalyst, leading to a lower reduction temperature (H2-
TPR) and a facile desorption of carboxylic acid (Acetic acid-TPD). The 5%Co+0.5%Pt/SiO2
shows hexene selectivity, similar to 5%Co/SiO2; however, hexane and iso-C6 are particularly
increased, especially upon the increase in contact time and Pt content. An appreciated
decarbonylation activity without excessive hydrogenation of the olefin produced, results in
selective conversion of fatty acid to a desirable -olefin. The preparation of Co-Pt catalysts
by co-impregnation (CIP) provides a higher -olefin selectivity, as compared to that by
sequential impregnation (SIP).
CRediT author statement
Ploynisa Phichitsurathawor: Investigation, Writing - Original Draft
Kittisak Choojun: Writing - Review & Editing, Supervision, Funding acquisition
Yingyot Poo-arporn: Investigation, Data Curation
Tawan Sooknoi: Conceptualization, Writing - Review & Editing, Supervision, Funding acquisition
28