Petroleum like biodiesel production by catalytic decarboxylation of oleic acid over Pd/Ce-ZrO2 under solvent-free condition

Article abstract of DOI:10.1016/j.apcata.2018.07.005

The Ce/Zr ratio of Pd/Ce-ZrO₂ catalysts was systematically changed in order to investigate the effect of oxygen vacancy concentration on their decarboxylation activityunder solvent-free conditions for potential sustainable petroleum like biodiesel production. Pd/Ce_0.5Zr_0.5O₂ exhibited the highest catalytic activity from all other tested catalysts because it contained the highest oxygen vacancy concentration and Pd dispersion, as shown by the X-ray photoelectron spectroscopy, Raman spectroscopy, and CO-chemisorption data. A catalyst deactivation study also showed that both the Pd dispersion and the oxygen vacancy concentration influences the catalytic activity.The catalyst deactivation was found to occur mainly due to Pd sintering, decreases in the BET surface area and Pd dispersion, and partially due to the loss of oxygen vacancies.

Full text of DOI:10.1016/j.apcata.2018.07.005

AppliedCatalysisA,General563(2018)163–169
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Petroleum like biodiesel production by catalytic decarboxylation of oleic
acid over Pd/Ce-ZrO2 under solvent-free condition
Jae-Oh Shima, Won-Jun Janga, Kyung-Won Jeona, Da-We Leea, Hyun-Suk Naa, Hak-Min Kima,
Yeol-Lim Leea, Seong-Yeun Yooa, Byong-Hun Jeonb, Hyun-Seog Roha,⁎, Chang Hyun Koc,⁎
a
Department of Environmental Engineering, Yonsei University, 1 Yonseidae-gil, Wonju, Gangwon, 26493, Republic of Korea
Department of Natural Resources and Environmental Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul, 04763, Republic of Korea
School of Applied Chemical Engineering, Chonnam National University, 77 Yongbong-ro, Gwangju, 61186, Republic of Korea
b
c
A R T I C L E I N F O
A B S T R A C T
Keywords:
The Ce/Zr ratio of Pd/Ce-ZrO2 catalysts was systematically changed in order to investigate the effect of oxygen
vacancy concentration on their decarboxylation activityunder solvent-free conditions for potential sustainable
petroleum like biodiesel production. Pd/Ce0.5Zr0.5O2 exhibited the highest catalytic activity from all other tested
catalysts because it contained the highest oxygen vacancy concentration and Pd dispersion, as shown by the X-
ray photoelectron spectroscopy, Raman spectroscopy, and CO-chemisorption data. A catalyst deactivation study
also showed that both the Pd dispersion and the oxygen vacancy concentration influences the catalytic
activity.The catalyst deactivation was found to occur mainly due to Pd sintering, decreases in the BET surface
area and Pd dispersion, and partially due to the loss of oxygen vacancies.
Oxygen vacancy
Sustainable biodiesel
Decarboxylation
Solvent-free
Ce/Zr ratio
1. Introduction
(R − COOH → R − H + CO2) to produce linear hydrocarbons that are
similar to conventional petroleum-derived hydrocarbons [2,11,12].
Extensive research has been carried out on the production of sus-
tainable biodiesel from renewable resources in response to accelerating
fossil fuel consumption and depletion, and to mitigate the effects of
greenhouse gas emissions from fossil fuel combustion [1–5]. These is-
sues have initiated studies on the conversion of biomass into biodiesel,
which is produced from renewable sources such as plants and micro-
algae, making it a sustainable and carbon-neutral transport fuel [6,7].
First-generation biodiesel is commonly produced by transester-
ification, but must be blended with fossil fuels because its physical and
chemical properties are less desirable than those of petroleum-derived
diesel fuel [2,8–10]. The oxygenated compounds in biodiesel precursor
must be eliminated to improve its physical and chemical properties. In
this viewpoint, hydrodeoxygenation (HDO) process has been developed
to overcome the drawbacks of first-generation biodiesel. The HDO
process has advantage that the oxygen contents can be eliminated in the
form of H2O, thus resultant product can retain carbon number and total
energy contents [11]. However, HDO process requires excess amount of
hydrogen that limits the economic feasibility of biodiesel production
via HDO process [3]. Thus, catalytic decarboxylation of fatty acids was
developed to overcome the above-mentioned limitation. The carbox-
ylate group (COO−) of fatty acids (ReCOOH) is removed as CO2
Several researchers have found that H2 is essential for the dec-
arboxylation reaction, and that the reaction pathway changes de-
pending on whether H2 or N2 atmospheres are used [2,13–15]. Immer
et al. found that oleic acid is mainly decarboxylated in the presence of
hydrogen, but inert conditions cause decarbonylation instead [2].
Arend et al. also reported that reaction condition is important for
deoxygenation reaction [13]. Thus, proper amounts of H2 should be
required for the decarboxylation reaction.
Currently, catalysts based on various metals such as Pd, Pt, and Ni
are used for the deoxygenation of fatty acids, and those based on pal-
ladium are the most effective catalysts for these reactions
[1,2,11,13,14,16–28]. Murzin and co-workers reported that Pd has the
highest catalytic activity in deoxygenation reactions at 300 °C [11].
They also reported that the catalytic activity of monometallic catalysts
with various active metals decreases in the following order: Pd >
Pt > Ni > Rh > Ir > Ru > Os [11]. Recent reports have shown
that Pd-based catalysts for fatty acid decarboxylation are highly sensi-
tive to the effects of the catalyst support, which plays a vital role in
controlling the decarboxylation reaction pathway [29]. Ford et al. re-
ported that a carbon-supported Pd catalyst was more active than SiO2-
and Al2O3-supported Pd catalysts [29]. In addition, Pd/C catalysts
⁎
Corresponding authors.
E-mail addresses: hsroh@yonsei.ac.kr (H.-S. Roh), chko@jnu.ac.kr (C.H. Ko).
https://doi.org/10.1016/j.apcata.2018.07.005
Received 21 March 2018; Received in revised form 22 June 2018; Accepted 2 July 2018
Availableonline08July2018
0926-860X/©2018ElsevierB.V.Allrightsreserved.

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
predominantly proceed via a decarboxylation route, whereas Pd/SiO2
catalysts primarily give deoxygenation via decarbonylation. Recently,
reducible oxides such as CeO2, TiO2, and ZrO2 have been used in
deoxygenation reactions [17,20,22,30]. According to the literature,
carboxylic acids first adsorb at oxygen defect sites in the oxides, fol-
lowed by elimination of the carboxyl group [20,22]. Thus, we have
assumed that the concentration of oxygen vacancies affects the dec-
arboxylation reaction.
by assuming the adsorption stoichiometry of one CO per Pd surface
atom (CO/Pds = 1) [36]. The palladium dispersion (D (%)), metallic
surface area (SPd (m2/g)), and active Pd particle size (nm) of the cata-
lysts were calculated using the following the equations:
V
W
s
× SF × MPd
s
D (%) =
× 100
× Wf × Vm, STP
V
s
× SF × NA × SC
× Wf × Vm, STP
SPd (m2/gPd) =
The incorporation of another cation, such as Zr4+, in the CeO2
lattice is known to enhance the creation of oxygen vacancies [31–34].
Our previous study also reveals that the insertion of Zr4+ in the CeO2
lattice enhances the redox properties and thermal resistance of CeO2,
and the Ce/Zr ratio is well known to affect the physical and chemical
properties of Ce-ZrO2 mixed oxide supports [17,30,31,35]. The main
purpose of our previous study is to increase the catalytic performance of
Ce-ZrO2 based catalysts by introducing small amount of hydrogen
(1 bar, 20 vol.%) to maintain and activate the active sites of catalyst
[17,30]. We designed a Pd-doped Ce-ZrO2 catalyst with various Ce/Zr
ratio for decarboxylation reactions to investigate the relationship be-
tween the concentration of oxygen vacancies and catalytic activity in a
decarboxylation reaction, which was carried out at 300 °C.
W
s
6 × MPd
Active Pd particle size (nm) =
W
× Wf × ρPd × NA × SPd
s
where Vs represents the volume of sorbed H2 (mL) under standard
temperature and pressure (STP) conditions, SF is the stoichiometry
factor (1), MPd is the gram molecular weight of Pd (106.42 g/mol), Ws
represents the sample weight (g), Wf is the weight fraction of Pd in the
catalyst (1%), Vm,STP is the molar volume of H2 (22,414 mL/mol) under
STP conditions, NA is the Avogadro’s number (6.023 × 1023/mol), SC
represents the cross sectional area of Pd (7.87 × 10−20 m2), and ρPd is
the density of Pd (12.02 g/cm3).
XPS spectra were obtained using a Kα spectrophotometer (VG
Multilab 2000) with a high-resolution monochromator. The pressure of
the analysis chamber was maintained at 6.8 × 10−9mbar. The detector
was used in constant energy mode with a pass energy of 100 eV for the
survey spectrum and 50 eV for the detailed scan. Binding energies were
calibrated against the C 1 s transition, which appeared at 284.6 eV.
Raman spectra were obtained using a LabRam Aramis (Horiba Jobin
Yvon) with excitation at 532 from a Nd-YAG laser. A 500 μm pinhole
was used for a spectral resolution grating, yielding approximately 1.5
cm−1 resolution. The concentration of oxygen vacancies (N) was cal-
culated using the spatial correlation model from the relationship be-
tween correlation length (L) and grain size (dg) [37,38]. The detailed
equation used was as follows:
Generally, deoxygenation reaction has been carried out in various
solvent such as dodecane and toluene. From an economic view point,
employment of auxiliary reagent could significantly raise the cost of the
total biodiesel processing. Therefore, we have designed deoxygenation
reaction under solvent free condition. The physical and chemical
properties of Pd-doped Ce-ZrO2 catalysts with various Ce/Zr ratios were
characterized using various techniques, including Brunauer-Emmett-
Teller (BET) surface area, X-ray diffraction (XRD), X-ray photoelectron
spectroscopy (XPS), Ramanspectroscopy, and CO-chemisorption. The
characterization data for each catalyst was correlated with its dec-
arboxylation activity under solvent-free conditions.
2. Experimental
51.8
dg (nm) =
(Γ−5)
2.1. Catalyst preparation
Ce-ZrO2supports were prepared by a co-precipitation method using
a solution of KOH (15 wt.%) as the precipitating agent, as reported in
our earlier work [30,32–35]. Stoichiometric quantities of zirconyl ni-
trate solution (20 wt.% ZrO2 basis, MEL Chemicals) and Ce(NO3)3∙6H2O
(99%, Aldrich) were combined in distilled water. The Ce/Zr atomic
ratios were 4:1, 1:1, and 1:4 for Ce0.8Zr0.2O2, Ce0.5Zr0.5O2, and
Ce0.2Zr0.8O2, respectively. The precipitates were aged at 80 °C for 3
days before they were washed five times with distilled water and then
air-dried for 1 day at 110 °C. The prepared Ce-ZrO2supports were cal-
cined in air atmosphere from room temperature to 500 °C at a heating
rate of 1 °C/min, after which the final temperature was maintained for
6 h. Pd (Pd loading: 1 wt.%) was loaded by an incipient wetness im-
pregnation method. Pd(NO3)2 (99.999%, Aldrich) was used as a pre-
cursor. The Pd-loaded Ce-ZrO2catalysts were also calcined at 500 °C for
6 h.
α
⎡
⎤
(dg−2α)3 + 4dg2α
2
3
L(nm) =
(
)
⎢
⎣
⎥
⎦
2dg
3
N =
4πL3
where Γ is the half-width at half-maximum (HWHM) of the Raman line
(around 450 cm−1) broadening, and α is the radius of CeO2 units
(0.34 nm), determined from universal constants [37,38]. Oxygen con-
tent was measured by elementary analysis using a Thermo Finnigan
FLASH EA-1112 Elemental Analyzer (EA).
2.3. Catalytic reaction
Decarboxylation reactions were carried out in an autoclave reactor
(100 mL capacity) operating in batch mode. The temperature was
measured using a K-type thermocouple. In a typical batch experiment,
27.5 g of oleic acid and 1.35 g of catalyst (reactant/catalyst = 20/1 w/
w) were placed in the reactor. The Pd/Ce-ZrO2 catalysts were reduced
in 10% H2/N2 atmosphere under a pressure of 1 atm at 200 °C (heating
rate: 2.9 °C/min) for 2 h using a tubular furnace. The inlet reduction gas
flow rate was 100 mL/min.Passivation process using 2% O2/N2 at room
temperature for 12 h was followed before the exposure to air to avoid
explosive oxidation of reduced Pd/Ce-ZrO2 catalyst. After oleic acid and
the reduced Pd/Ce-ZrO2 catalyst were loaded into the reactor, the re-
actor was flushed with nitrogen to remove any remaining oxygen. Then,
the reactor was purged with pure hydrogen, and the pressure was in-
creased to 20 bar. The reactor was heated from room temperature to
300 °C at a heating rate of 4.5 °C/min, and the reaction temperature was
2.2. Characterization
The BET surface area of the catalysts was measured using a nitrogen
adsorption technique on an ASAP 2010 (Micromeritics)accelerated
surface area and porosimetry instrument. Before analysis, the samples
were degassed for 12 h at 110 °C under a vacuum less than 0.5 mm Hg.
XRD was carried out using a Rigaku D/MAX–IIIC diffractometer oper-
ated at 40 kV and 100 mA with Ni-filtered CueKα radiation. Lattice
parameters were calculated from XRD peaks using Bragg’s law. CO-
chemisorption was conducted in an AutoChem 2920 instrument
(Micromeritics). Temperature-programmed desorption of ammonia
(NH3−TPD) was carried out to evaluate the total acidity of the catalysts
using an Autochem 2920 instrument. The Pd dispersion was calculated
164

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
maintained for 3 h. A multi-blade impeller mixed the liquid reactant
and solid catalyst and the stirring speed was fixed at 300 rpm during the
reaction. The reactor was subsequently cooled down to room tem-
perature. The liquid products were collected after filtering the solid
phase catalysts from the mixture. The used catalyst was collected,
washed with 100 mL of toluene, and air-dried for 1 day at room tem-
perature to test the stability of the Pd/Ce0.5Zr0.5O2 catalyst. The liquid
products were analyzed using gas chromatography (HP 6890 N) with a
flame ionization detector and a capillary column (HP-5, 30 m). The
product gas from the batch reactor was sampled to 0.5 L Tedlar bag
(CEL Scientific Corp.) and it was analyzed by an online micro-gas
chromatograph (Agilent 3000) equipped with molecular sieve and
PLOT U columns. The conversion of oleic acid and the selectivity of Ci
products were calculated using the following equations (n : mole of
compound, i : carbon number).
nOleic acid, in−nOleic acid, out
Oleic acid conversion (%) =
Ci selectivity (%) =
× 100
nOleic acid, in
Fig. 1. XRD patterns ofPd/Ce-ZrO2 catalysts with various Ce/Zr ratios.
nC
i
relatively small atomic size of Zr4+ (0.84 Å) compared to the atomic
size of Ce4+ (0.97 Å) [32]. In addition, the reported lattice parameter of
pure CeO2 is 0.5409 nm [41]. According to the literature, the structure
of the Ce-ZrO2 support depends on its composition [39]. When the Ce
content is below 50%, the Ce-ZrO2 supports adopt a tetragonal structure
(Pd/Ce0.2Zr0.8O2), whereas a cubic Ce-ZrO2 structure is formed when
the Ce content is above 50% (Pd/Ce0.8Zr0.2O2 and Pd/Ce0.5Zr0.5O2)
[39]. No peaks corresponding to Pd metal were detected in any of the
catalysts because the size of the Pd metal particles was less than 3 nm in
all catalysts (highly dispersed), which is below the limit for X-ray line
broadening analysis [42]. An XPS survey scan (Fig. S1), carried out to
check for the presence of Pd species on the catalyst surface, confirmed
the presence of metallic Pd species on the catalyst surface. Thus, the Pd
nanoparticles were finely distributed on the Ce-ZrO2 supports.
× 100
nOleic acid,in−nOleic acid,out
3. Results and discussion
3.1. Catalyst characterization
Table 1 summarizes the characteristics of Pd/Ce-ZrO2 catalysts with
various Ce/Zr ratios. The BET surface area decreased with increasing Ce
content: a phenomenon that has been well documented in related lit-
erature [30,39]. As a result, the Pd/Ce0.8Zr0.2O2 catalyst showed the
lowest BET surface area (45 m2/g). The BET surface area of the Pd/Ce-
ZrO2 catalysts with various Ce/Zr ratios decreased in the following
order:
Pd/Ce0.2Zr0.8O2
(197 m2/g) > Pd/Ce0.5Zr0.5O2
(122 m2/
The O 1 s spectra of Pd/Ce-ZrO2 catalysts with various Ce/Zr ratios
are shown in Fig. 2. Three types of oxygen species were detected in the
XPS spectra, labeled as OL, OD, and OH, respectively [38]. The OL, OD,
g) > Pd/Ce0.8Zr0.2O2 (45 m2/g).
The Pd dispersion of the prepared catalysts was calculated from
pulse CO chemisorption results. The Pd dispersion value of Pd/Ce-ZrO2
catalysts with various Ce/Zr ratios is also summarized in Table 1. The
Pd dispersion values ranged from 44% to 67%, with the Pd/Ce0.5Zr0.5O2
catalyst exhibiting the highest value (67%) and the Pd/Ce0.2Zr0.8O2
catalyst having the lowest (44%). As a result, the Pd/Ce0.5Zr0.5O2 cat-
alyst had the smallest active Pd particle size (1.7 nm) while the Pd/
Ce0.2Zr0.8O2 catalyst had the largest (2.6 nm). Simakova et al. reported
that Pd catalysts with different Pd dispersion could show different
catalytic activity and product distributions in the deoxygenation of
mixtures of stearic acid and palmitic acid [26]. Pd catalysts with a high
metal dispersion have higher catalytic activity than those with low Pd
dispersion.
XRD patterns of the various Pd/Ce-ZrO2 catalysts are shown in
Fig. 1, and the lattice parameters of the Pd/Ce-ZrO2 catalysts with
various Ce/Zr ratios are summarized in Table 1. The diffractogram of
Pd/Ce0.8Zr0.2O2 shows peaks corresponding to cubic CeO2 [30,31,40].
The main peaks shifted to higher angles and the lattice parameter of
CeO2 decreased slightly with increasing in Zr content because of the
and OH peaks at 528.7
0.4, 530.1
0.4, and 531.7
0.2 eV cor-
respond to the lattice oxygen in Ce-ZrO2 mixed oxide, oxygen va-
cancies, and adsorbed oxygen species from hydroxyl or water species on
the metal surface, respectively [38]. When the Zr contents were in-
creased in the Ce-ZrO2 mixed oxide, all peaks shifted to higher binding
energy (BE) because Zr4+ is more electronegative than Ce4+ [43]. The
quantitative XPS elemental analysis of the O 1 s peak of the various Pd/
Ce-ZrO2 catalysts are summarized in Table 1. The Pd/Ce0.5Zr0.5O2
catalyst had the highest surface defect oxygen ratio of the prepared
catalysts, of ∼46.2%. Our hypothesis that the concentration of oxygen
vacancies affects the decarboxylation reaction is based on the report by
Peng et al. that carboxylic acids are reduced to aldehydes by the ad-
sorption of the oxygen atoms of the carboxylic acid at the oxygen va-
cancies on the surfaces of reducible oxides (CeO2, TiO2, ZrO2), and the
fact that adsorbed carboxylic acids can be converted to carboxylates
followed by reaction with activated H2, which is activated through
dissociative adsorption at oxygen defect sites [20]. Thus, we would
Table 1
Characteristics of Pd/Ce-ZrO2 catalysts with varying Ce/Zr ratio.
Catalysts
BET S.A.
(m2/g)a
Pd dispersion
(%)b
Active Pd particle size (nm)b
Lattice parameter
(nm)c
OD
Concentration of oxygen vacancies (cm−3
)
/(OL+OD+OH
)
Pd/Ce0.8Zr0.2O2
Pd/Ce0.5Zr0.5O2
Pd/Ce0.2Zr0.8O2
45
122
197
60
67
44
1.9
1.7
2.6
0.5385
0.5327
0.5268
34.2%
46.2%
29.6%
4.63 × 1021
6.06 × 1021
N.A.d
a
Estimated from N2 adsorption at −196 °C.
Estimated from CO-chemisorption.
Lattice parameter calculated from XRD peaks by Bragg’s law.
Not available due to very broad and weak Raman spectra.
b
c
d
165

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
supply of oxygen vacancies [49].
3.2. Reaction results
The compositions of the reaction products obtained using the pre-
pared catalysts are shown in Table 2. Oleic acid residue was not de-
tected in any of the products, meaning that all the catalysts showed
100% oleic acid conversion. Stearic acid can be formed from hydro-
genation of oleic acid, and was the most prevalent single product of the
studied reactions [3,51]. Interestingly, the percentage of C17 hydro-
carbons (saturated and unsaturated) was highest in the case of Pd/
Ce0.5Zr0.5O2catalyst. These trends are well matched with the con-
centration of oxygen defects in the catalysts, as calculated by XPS and
Raman spectroscopy. Thus, we conclude that the decarboxylation re-
action was affected by the oxygen vacancy concentration of the cata-
lysts. Previous reports have found that carboxylic groups that adsorb at
the oxygen vacancy sites of nickel catalysts are converted to ketene
intermediates (R − COOH → R − COO → R1−C = C]O), which are
hydrogenated to form aldehydes [20,22]. Aldehydes can be converted
to both alcohols and linear hydrocarbons, but neither of these com-
pounds was detected in this study. Furthermore, no C18 hydrocarbons
were detected, which suggests that no hydrodeoxygenation reactions
occurred [14,52,53].
In order to confirm the main reaction pathway, micro-GC analysis of
the gaseous products was carried out. As shown in Table 3, the main
constituent of the product gases was CO2. A decarboxylation reaction
leads to the removal of a carboxyl group (−COO) from a fatty acid by
releasing CO2. This clearly indicates that the decarboxylation reaction
was the main reaction pathway. In addition, the CO2 percentage of the
product gas was consistent with the extent of decarboxylation. A small
amount of CO, which was produced via a decarbonylation reaction, was
also presents as part of the product gases. Fig. 4 also shows the per-
centage of saturated C17 hydrocarbons (heptadecane) and unsaturated
C17 hydrocarbons (8-heptadecene). Saturated C17 hydrocarbon was the
predominant component of the C17 hydrocarbon products. This is in
agreement with the results from the micro-GC analysis, indicating that
most of the oleic acid was converted via decarboxylation. However, we
could not determine the other compounds due to the technical limita-
tions connected to micro-GC analysis. It is possible that these by-
products are more shorter hydrocarbons (C4∼C8) and oxygenate com-
pounds (due to cracking reaction) [3,13]. The results showed that the
Pd/Ce0.2Zr0.8O2 catalyst exhibited 3 to 8 times higher amount of by-
products than the other two catalysts. In other to find out the reason of
the excessive cracking reaction, an NH3-TPD analysis was carried out.
As shown in Fig. S2, the shapes of the NH3-TPD patterns of the prepared
catalysts were similar in the temperature range between 100 °C and
500 °C. However, a strong acidic site was detected at 625 °C in the case
of the Pd/Ce0.2Zr0.8O2 catalyst. Moreover, the Pd/Ce0.2Zr0.8O2 catalyst
had a higher acidity (701 μmol NH3/gcat) than both the Pd/Ce0.5Zr0.5O2
catalyst (572 μmol NH3/gcat) and Pd/Ce0.8Zr0.2O2 catalysts (401 μmol
NH3/gcat). According to the literature, a strong acidity of a catalyst
could induce a cracking reaction [3]. Therefore, it can be assumed that
the higher byproduct content in the reaction with the Pd/Ce0.2Zr0.8O2
catalyst was due to the strong acidity of the Pd/Ce0.2Zr0.8O2 catalyst.
Fig. 5 displays the pressure difference with time on stream over the
Pd/Ce0.5Zr0.5O2catalyst, from which the reaction step can be defined.
At temperatures up to 100 °C, the initial pressure increase was caused
by the thermal cracking of oleic acid and simply the expansion of H2.
However, when the reaction temperature increased further, to 300 °C, a
dramatic pressure drop was observed. The values in Table 2 show that
hydrogen is consumed by the conversion of oleic acid
(CH3(CH2)7CH = CH(CH2)7COOH) to stearic acid (CH3(CH2)16COOH).
Thus, the hydrogenation reaction occurred before decarboxylation re-
action, which occur on the Pd surface. The decarboxylation reaction
over Ce0.5Zr0.5O2 support without Pd was conducted to check the re-
action step. Despite the increasing reaction temperature, a pressure
Fig. 2. O 1sXPS spectra of Pd/Ce-ZrO2 catalysts with various Ce/Zr ratios.
expect that concentration of oxygen vacancies affects the decarbox-
ylation reaction, and expect that the Pd/Ce0.5Zr0.5O2 catalyst should
have a higher activity than the others.
Raman spectroscopy analysis was carried out (Fig. 3) to confirm the
structure and concentration of oxygen vacancies. Except for the Pd/
Ce0.2Zr0.8O2 catalyst, the catalysts showed a single Raman band at
448–462 cm−1, which was attributed to the F2g vibration mode of the
fluorite structure of CeO2 [38,43–48]. This peak shifted to lower wave
number with the addition of Zr, since the addition of the smaller Zr4+
(0.84 Å) caused local structure distortions in the CeO2 lattice [38]. This
result agreed with measurements of the lattice parameters. When the
Ce/Zr ratio was decreased to 1/4 (as in the Pd/Ce0.2Zr0.8O2 catalyst),
this peak became weak and disappeared. This change indicates the
disappearance of the cubic fluorite structure phase [49,50]. The con-
centration of oxygen vacancies can be calculated from the Raman band
at 448–462 cm−1, and Table 1 shows the calculated concentrations of
oxygen vacancies for the various Pd/Ce-ZrO2 catalysts [37,38]. The Pd/
Ce0.8Zr0.2O2 catalyst had an oxygen vacancy concentration of
4.63 × 1021 cm−3. An increase in the oxygen vacancy concentration
was observed when the Zr content increased to 50%, and the Pd/
Ce0.5Zr0.5O2catalyst exhibited the highest concentration of oxygen va-
cancies. Calculation of the oxygen vacancy concentration of Pd/
Ce0.2Zr0.8O2 catalyst was not possible because of the broad and weak
Raman spectra. However, we can infer from previous XPS quantitative
elemental analyses that the oxygen vacancy concentration of the Pd/
Ce0.2Zr0.8O2 catalyst was lower than those of the other catalysts. This
difference may be due to the decreased Ce content providing a limited
Fig. 3. Raman spectra of Pd/Ce-ZrO2 catalysts with various Ce/Zr ratios.
166

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
Table 2
Reaction results over Pd/Ce-ZrO2 catalysts with various Ce/Zr ratios.
Catalysts
Oleic acid
conversion
Stearic acid
selectivity
C17 hydrocarbons
selectivity
C9–C16 hydrocarbons
Oxygen
contents
Oxygen removal
Liquid yieldc
(liquid
selectivitya
efficiencyb
amount)
Pd/Ce0.8Zr0.2O2 100%
Pd/Ce0.5Zr0.5O2 100%
Pd/Ce0.2Zr0.8O2 100%
76%
63%
91%
22%
34%
7%
2%
3%
2%
9.4%
9.1%
10.1%
18.3%
20.9%
12.1%
83.9%
(23.1 g)
77.0%
(21.2 g)
88.1%
(24.2 g)
*Reaction condition: 300 °C, reactant/catalyst = 20/1, 20 bar, H2 condition.
a
Saturated hydrocarbons + unsaturated hydrocarbons.
Oxygen contents of oleic acid (reactant) = 11.5%.
Liquid yield expressed as percent (%) is calculated by measuring initial weight of reactant(27.5 g)overweight of liquid product.
b
c
Table 3
Micro-GC analysis results over Pd/Ce-ZrO2 catalysts with various Ce/Zr ratios.
Catalysts
CO2
H2
CO
CH4
Others
Total
Pd/Ce0.8Zr0.2O2
Pd/Ce0.5Zr0.5O2
Pd/Ce0.2Zr0.8O2
68.6%
75.0%
56.5%
19.5%
15.0%
20.2%
6.6%
7.6%
8.0%
1.1%
0.8%
2.4%
4.2%
1.6%
12.9%
100.0%
100.0%
100.0%
*Micro-GC analysis results is volume percent.
Fig. 5. Pressure differences with time on stream over a Pd/Ce0.5Zr0.5O2catalyst.
(Reaction condition: 300 °C, reactant/catalyst = 20/1, 20 bar, H2 condition).
Fig. 4. Percentage of saturated C17 hydrocarbon and unsaturated C17 hydro-
carbon. (Reaction condition: 300 °C, reactant/catalyst = 20/1, 20 bar, H2 con-
dition).
drop was not observed (Fig. S3). In the case of bare Ce0.5Zr0.5O2 sup-
port, the pressure increased with the increase in reaction temperature
and time. Thus, it can be concluded that the hydrogenation of the oleic
acid occurred only on the Pd surface. After the hydrogenation reaction,
stearic acid can be converted to heptadecane. After heating was com-
plete, the pressure gradually increased (5.4 bar (0 h) → 8.0 bar (3 h)).
One mole of carbon dioxide was generated in the decarboxylation re-
action, which is a likely cause of the pressure increase. In addition, a
small number of shorter hydrocarbons (C9∼C16) and byproducts were
detected, which were likely formed from the cracking of reactant [3].
Fig. 6 shows the selectivities for stearic acid and C17 hydrocarbons as a
function of time on stream over the Pd/Ce0.5Zr0.5O2 catalyst. Interest-
ingly, oleic acid was totally converted during temperature rising pro-
cess. This result is well reflected in the pressure difference with time on
stream results shown in Fig. 5. As the initial reaction temperature
reached 300 °C, most of the oleic acid is converted to stearic acid. Al-
though the selectivity for the stearic acid at t = 0 h was relatively high,
it decreased with increasing the time on stream (89%, 0 h → 63%, 3 h).
It can be said that the stearic acid was firstly formed before the reaction
Fig. 6. Changes in selectivities for stearic acid and C17 hydrocarbons observed
as a function of time on stream in reaction utilizing the Pd/Ce0.5Zr0.5O2catalyst.
and it was decarboxylated/decarbonylated to afford saturated/un-
saturated hydrocarbons.
Elemental analysis was carried out to check the remaining oxygen
content of the products, and the results are shown in Table 2 along with
the oxygen removal efficiency of the various Pd/Ce-ZrO2 catalysts. The
167

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
Fig. 7. Relation between the characteristics of catalysts and reaction results.
Fig. 8. O 1s XPS spectra of used Pd/Ce0.5Zr0.5O2catalysts.
(Reaction condition: 300 °C, reactant/catalyst = 20/1, 20 bar, H2 condition).
spectrum, as for the fresh Pd/Ce0.5Zr0.5O2 catalyst. The positions of the
deconvoluted peaks are similar to those of the fresh Pd/Ce0.5Zr0.5O2
catalyst. The quantitative XPS elemental analyses of the O 1 s peak of
the used Pd/Ce0.5Zr0.5O2 catalysts are shown in Table 4. The con-
centration of surface oxygen defects of the used Pd/Ce0.5Zr0.5O2 cata-
lysts decreased as the number of uses increased (46.2% → 34.9% →
30.4%). This result clearly shows that the oxygen vacancy concentra-
tion is related to the activity of the catalyst for the decarboxylation
reaction. Especially, a dramatic reduction of BET surface area (122 m2/
g → 42 m2/g → 24 m2/g) and Pd dispersion (67% → 5% (−93%) → 2%
(−97%)) are a main factor in catalyst deactivation. We suggest that the
oxygen vacancies are blocked by strongly adsorbed carboxylic groups,
thereby decreasing the concentration of oxygen vacancies. Table 4
compares the characteristics of the fresh and used Pd/Ce0.5Zr0.5O2
catalysts. The BET surface area of Pd/Ce0.5Zr0.5O2 catalyst decreased by
65% and 80% after each subsequent reaction. The Pd dispersion of Pd/
Ce0.5Zr0.5O2 catalyst also decreased to 5% and 2% after recycling. The
CO-chemisorption data indicates that significant Pd sintering is also
observed. Thus, we conclude that the causes of catalyst deactivation are
mainlyPd sintering, decreases in BET surface area and Pd dispersion,
and the partial reduction in oxygen vacancy concentration.
In this study, the Pd/Ce-ZrO2 catalyst activity was affected by the
Ce/Zr ratio. The Ce/Zr ratio affects the formation of oxygen vacancy
concentration, which is the effect on the decarboxylation reaction. We
also found that the decarboxylation reaction was the main reaction
pathway in this study. Thus, we propose the reaction step in this re-
action is as follows. First, oleic acid is hydrogenated to stearic acid.
Second, stearic acid is dissociated with carboxylate compounds
(ReCOO) and hydrogen, which occurred at the oxygen vacancy sites
and Pd surface. Thus, the oxygen vacancy concentration and Pd dis-
persion (active Pd particle size) have a strong influence on this reaction.
After adsorption, CeC scission occurs and the resultant products are
hydrogenated to produce linear hydrocarbons and carbon dioxide.
Pd/Ce0.5Zr0.5O2catalyst exhibited the highest oxygen removal effi-
ciency among the prepared catalysts. According to previous reports, the
oxygen removal efficiency and liquid product yield are related to the
decarboxylation reaction. The oxygen contained in fatty acids is re-
moved during decarboxylation in the form of carbon dioxide [3,12,54].
The liquid product yields of the various Pd/Ce-ZrO2 catalysts are shown
in Table 2. The Pd/Ce0.5Zr0.5O2catalyst also exhibited the lowest liquid
product yield among the tested catalysts. The liquid product yield in-
creased in the order: Pd/Ce0.5Zr0.5O2 (77.0%) < Pd/Ce0.8Zr0.2O2
(83.9%) < Pd/Ce0.2Zr0.8O2 (88.1%). The selectivity for C17 hydro-
carbons, liquid product yield, and oxygen removal efficiency are com-
pared for each catalyst in Fig. 7. The decarboxylation reaction (se-
lectivity for C17 hydrocarbons) influences the liquid product yield and
oxygen removal efficiency. In addition, catalytic performance was
strongly affected by the oxygen vacancy concentration and Pd disper-
sion. As shown in Table 3, the sum of CO2, H2 (just residue), and CO
was almost 95%, except for the case of the Pd/Ce0.2Zr0.8O2 catalyst,
which contained 12.9% thermal and catalytic cracking-derived com-
pounds. Thus, it can be concluded that the gas yield was almost equal to
the oxygen removal efficiency. Table S1 shows the differences between
these two factors. Note, that small differences may occur due to ana-
lytical errors. Therefore, it was concluded that the oxygen removal ef-
ficiency was in accordance with the extent of decarboxylation.
A
cycling test was conducted to check the stability of Pd/
Ce0.5Zr0.5O2 catalyst (Table 4). Interestingly, all the used catalysts
showed 100% oleic acid conversion. However, their selectivity for C17
compounds decreased as the number of usage cycles increased, and the
stearic acid selectivity increased as the C17 selectivity decreased. BET,
CO-chemisorption, and XPS analyses of the recycled catalysts were
conducted to determine the cause of the observed deactivation (Table 4
and Fig. 8). Fig. 8 shows the O 1 s spectra of the used Pd/Ce0.5Zr0.5O2
catalyst. Three kinds of oxygen species were detected in the XPS
Table 4
Characteristics and reaction results of fresh and used Pd/Ce0.5Zr0.5O2 catalysts.
Catalysts
Oleic acid
conversion
Stearic acid
selectivity
C17 hydrocarbons
selectivity
BET S.A.
(m2/g)a
Pd dispersionb Active Pd particle size
(nm)b
OD
/(OL+OD+OH
)
(Fresh) Pd/Ce0.8Zr0.2O2 100%
63%
81%
94%
34%
17%
5%
122
67%
1.7
24.9
72.3
46.2%
34.9%
30.4%
(1st) Pd/Ce0.5Zr0.5O2
(2nd) Pd/Ce0.2Zr0.8O2
100%
100%
42 (−65%) 5% (−93%)
24 (−80%) 2% (−97%)
*Reaction condition: 300 °C, reactant/catalyst = 20/1, 20 bar, H2 condition.
a
Estimated from N2 adsorption at −196 °C.
Estimated from CO-chemisorption.
b
168

J.-O. Shim et al.
AppliedCatalysisA,General563(2018)163–169
Finally, both are desorbed.
[18] Y. Wang, J. Wu, S. Wang, RSC Adv. 3 (2013) 12635–12640.
[19] V.A. Yakovlev, S.A. Khromova, O.V. Sherstyuk, V.O. Dundich, D.Y. Ermakov,
V.M. Novopashina, M. Yu, O. Lebedev, V.N. Bulavchenko, Parmon, Catal. Today
144 (2009) 362–366.
4. Conclusions
[20] B. Peng, X. Yuan, C. Zhao, J.A. Lercher, J. Am. Chem. Soc. 134 (2012) 9400–9405.
[21] E. Meller, U. Green, Z. Aizenshtat, Y. Sasson, Fuel 133 (2014) 89–95.
[22] B. Peng, C. Zhao, S. Kasakov, S. Foraita, J.A. Lercher, Chem. Eur. J. 19 (2013)
4732–4741.
The catalytic performance of Pd/Ce-ZrO2 catalysts depends on the
Ce/Zr ratio. The formation of oxygen vacancy concentration was also
affected by the Ce/Zr ratio. Moreover, the oxygen vacancy concentra-
tion has a strong effect on the decarboxylation reaction. The Pd/
Ce0.5Zr0.5O2 catalyst shows the highest selectivity for C17 compounds
and oxygen removal efficiency among the prepared catalysts. The
highest catalytic activity is found with the catalyst with the highest
oxygen vacancy concentration. The Pd dispersion also has a strong ef-
fect on the catalytic activity. The causes of catalyst deactivation are
mainly Pd sintering, a decrease of BET surface area and Pd dispersion,
and partially the loss of oxygen vacancies. Additionally, the reaction
proceeded mainly via a decarboxylation reaction pathway.
[23] P. Mäki-Arvela, I. Kubičková, M. Snåre, K. Eränen, D.Y. Murzin, Energy Fuels 21
(2007) 30–41.
[24] S. Lestari, I. Simakova, A. Tokarev, P. Mäki-Arvela, K. Eränen, D.Y. Murzin, Catal.
Lett. 122 (2008) 247–251.
[25] S. Lestari, P. Mäki-Arvela, K. Eränen, J. Beltramini, G.Q. Max Lu, D.Y. Murzin,
Catal. Lett. 130 (2009) 48–51.
[26] I. Simakova, O. Simakova, P. Mäki-Arvela, A. Simakov, M. Estrada, D.Y. Murzin,
Appl. Catal. A: Gen. 355 (2009) 100–108.
[27] S. Lestari, P. Mäki-Arvela, H. Bernas, O. Simakova, R. Sjöholm, J. Beltramini,
G.Q. Max Lu, J. Myllyoja, I. Simakova, D.Y. Murzin, Energy Fuels 23 (2009)
3842–3845.
[28] I. Simakova, B. Rozmysłowicz, O. Simakova, P. Mäki-Arvela, A. Simakov,
D.Y. Murzin, Top. Catal. 54 (2011) 460–466.
[29] J.P. Ford, J.G. Immer, H.H. Lamb, Top. Catal. 55 (2012) 175–184.
[30] J.-O. Shim, D.-W. Jeong, W.-J. Jang, K.-W. Jeon, B.-H. Jeon, S.Y. Cho, H.-S. Roh, J.-
G. Na, C.H. Ko, Y.-K. Oh, S.S. Han, Renew. Energy 65 (2014) 36–40.
[31] J.-O. Shim, Y.J. Hong, H.-S. Na, W.-J. Jang, Y.C. Kang, H.-S. Roh, ACS Appl. Mater.
Interfaces 8 (2016) 17239–17244.
[32] W.-J. Jang, H.-M. Kim, J.-O. Shim, S.-Y. Yoo, K.-W. Jeon, H.-S. Na, Y.-L. Lee, D.-
W. Jeong, J.W. Bae, I.W. Nah, H.-S. Roh, Green Chem. 20 (2018) 1621–1633.
[33] W.-J. Jang, D.-W. Jeong, J.-O. Shim, H.-S. Roh, I.H. Son, S.J. Lee, Int. J. Hydrogen
Energy 38 (2013) 4508–4512.
Acknowledgments
This work was supported by the National Research Foundation of
Korea (NRF) grant funded by the Korea government (MSIP)
(2017R1A2B4007145).
[34] H.-S. Roh, I.-H. Eum, D.-W. Jeong, Renew. Energy 42 (2012) 212–216.
[35] W.-J. Jang, D.-W. Jeong, J.-O. Shim, H.-M. Kim, W.-B. Han, J.W. Bae, H.-S. Roh,
Renew. Energy 79 (2015) 91–95.
[36] R. Pellegrini, G. Leofanti, G. Agostini, E. Groppo, M. Rivallan, C. Lamberti,
Langmuir 25 (2009) 6476–6485.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2018.07.005.
[37] P. Trogadas, J. Parrondo, V. Ramani, ACS Appl. Mater. Interfaces 4 (2012)
5098–5102.
References
[38] J.-O. Shim, H.-S. Na, A. Jha, W.-J. Jang, D.-W. Jeong, I.W. Nah, B.-H. Jeon, H.-
S. Roh, Chem. Eng. J. 306 (2016) 908–915.
[39] D.-W. Jeong, H.S. Potdar, J.-O. Shim, W.-J. Jang, H.-S. Roh, Int. J. Hydrogen Energy
[1] J.P. Ford, N. Thapaliya, M.J. Kelly, W.L. Roberts, H.H. Lamb, Energy Fuels 27
(2013) 7489–7496.
[2] J.G. Immer, M.J. Kelly, H.H. Lamb, Appl. Catal. A: Gen. 375 (2010) 134–139.
[3] J.-O. Shim, D.-W. Jeong, W.-J. Jang, K.-W. Jeon, S.-H. Kim, B.-H. Jeon, H.-S. Roh,
J.-G. Na, Y.-K. Oh, S.S. Han, C.H. Ko, Catal. Commun. 67 (2015) 16–20.
[4] S. Dutta, B. Saha, ACS Catal. 7 (2017) 5491–5499.
38 (2013) 4502–4507.
[40] Z. Sun, D.Y. Lu, F.N. Ridha, R.W. Hughes, D. Filippou, Appl. Energy 195 (2017)
303–315.
[41] L. Qi, Q. Yu, Y. Dai, C. Tang, L. Liu, H. Zhang, F. Gao, L. Dong, Y. Chen, Appl. Catal.
B: Environ. 119−120 (2012) 308–320.
[42] H.-S. Roh, H.S. Potdar, K.-W. Jun, Catal. Today 93−95 (2004) 39–44.
[43] W. Chen, R. Ran, D. Weng, X. Wu, J. Zhong, A. Zhu, S. Han, J. Mater. Sci. 50 (2015)
6268–6276.
[44] M. Piumetti, S. Bensaid, T. Andana, N. Russo, R. Pirone, D. Fino, Appl. Catal. B:
Environ. 205 (2017) 455–468.
[45] F. Morfin, T.-S. Nguyen, J.-L. Rousset, L. Piccolo, Appl. Catal. B: Environ. 197
(2016) 2–13.
[46] Z.-Y. Pu, J.-Q. Lu, M.-F. Luo, Y.-L. Xie, J. Phys. Chem. C 111 (2007) 18695–18702.
[47] M. Shen, L. Lv, J. Wang, J. Zhu, Y. Huang, J. Wang, Chem. Eng. J. 255 (2014)
40–48.
[48] B. Mallesham, P. Sudarsanam, B.V.S. Reddy, B.M. Reddy, Appl. Catal. B: Environ.
181 (2016) 47–57.
[49] L. Yang, X. Yang, R. Zhou, J. Phys. Chem. C 120 (2016) 2712–2723.
[50] T.H. Vuong, J. Radnik, E. Kondratenko, M. Schneider, U. Armbruster, A. Brückner,
Appl. Catal. B: Environ. 197 (2016) 159–167.
[51] A. Dragu, S. Kinayyigit, E.J. García-Suárez, M. Florea, E. Stepan, S. Velea, L. Tanase,
V. Collière, K. Philippot, P. Granger, V.I. Parvulescu, Appl. Catal. A: Gen. 504
(2015) 81–91.
[52] K.A. Roger, Y. Zheng, ChemSusChem 9 (2016) 1750–1772.
[53] M. Snåre, I. Kubičková, P. Mäki-Arvela, D. Chichova, K. Eränen, D.Y. Murzin, Fuel
87 (2008) 933–945.
[5] P.M. Mortensen, J.-D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Appl.
Catal. A: Gen. 407 (2011) 1–19.
[6] G. Bauer, S. Lima, J. Chenevard, M. Sugnaux, F. Fischer, ACS Sustain. Chem. Eng. 5
(2017) 1931–1937.
[7] L. Yang, K.L. Tate, J.B. Jasinski, M.A. Carreon, ACS Catal. 5 (2015) 6497–6502.
[8] B. Smith, H.G. Greenwell, A. Whiting, Energy Environ. Sci. 2 (2009) 262–271.
[9] G. Knothe, Fuel Process. Technol. 86 (2005) 1059–1070.
[10] M.J. Ramos, C.M. Fernandez, A. Casas, L. Rodriguez, A. Perez, Bioresour. Technol.
100 (2009) 261–268.
[11] M. Snåre, I. Kubičková, P. Mäki-Arvela, K. Eränen, D.Y. Murzin, Ind. Eng. Chem.
Res. 45 (2006) 5708–5715.
[12] J.-G. Na, J.K. Han, Y.-K. Oh, J.-H. Park, T.S. Jung, S.S. Han, H.C. Yoon, S.H. Chung,
J.-N. Kim, C.H. Ko, Catal. Today 185 (2012) 313–317.
[13] M. Arend, T. Nonnen, W.F. Hoelderich, J. Fischer, J. Groos, Appl. Catal. A: Gen. 399
(2011) 198–204.
[14] M. Snåre, I. Kubičková, P. Mäki-Arvela, K. Eränen, J. Wärnå, D.Y. Murzin, Chem.
Eng. J. 134 (2007) 29–34.
[15] M. Ahmadi, E.E. Macias, J.B. Jasinski, P. Ratnasamy, M.A. Carreon, J. Mol. Catal. A:
Chem. 386 (2014) 14–19.
[16] H.-S. Roh, I.-H. Eum, D.-W. Jeong, B.E. Yi, J.-G. Na, C.H. Ko, Catal. Today 164
(2011) 457–460.
[54] J.-G. Na, B.E. Yi, J.N. Kim, K.B. Yi, S.-Y. Park, J.-H. Park, J.-N. Kim, C.H. Ko, Catal.
[17] J.-O. Shim, D.-W. Jeong, W.-J. Jang, K.-W. Jeon, B.-H. Jeon, S.-H. Kim, H.-S. Roh,
J.-G. Na, S.S. Han, C.H. Ko, J. Nanosci. Nanotechnol. 16 (2016) 4587–4592.
Today 156 (2010) 44–48.
169