Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

Renewable Energy 197 (2022) 40–49

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Fe-promoted Ni catalyst with extremely high loading and oxygen vacancy


for lipid deoxygenation into green diesel
Fei Wang a, *, Hui Xu a, Songyin Yu a, Hao Zhu a, Yuchan Du a, Zeng Zhang a, Chaoqun You b,
Xiaoxiang Jiang a, **, Jianchun Jiang b, c
a
School of Energy & Mechanical Engineering, Nanjing Normal University, Nanjing, 210023, PR China
b
Joint Laboratory of Advanced Biomedical Materials (NFU-UGENT), College of Chemical Engineering, Co-Innovation Center of Efficient Processing and Utilization of
Forest Resources, Nanjing Forestry University, Nanjing, 210037, PR China
c
Institute of Chemical Industry of Forest Products, Chinese Academy of Forestry, Nanjing, 210042, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: The catalytic performance of Ni catalyst in lipid deoxygenation into green diesel is commonly limited by the
Lipid hydrotreatment quantity of accessible active sites, the specific surface area, and the acidity of the catalyst. To cope with these
Second-generation bio-diesel problems, the Fe-promoted Ni catalyst on ZrO2 with 60 wt% Ni loading (60%Ni–Fe/ZrO2) is synthesized by the
Co-precipitation method
co-precipitation method. The deoxygenation of oleic acid over 60%Ni–Fe/ZrO2 catalyst attained 98.7% con­
Hydrocarbons
Hydrotreatment pathway
version and 91.8% alkane yield, much higher than that of 20%Ni–Fe/ZrO2-2 prepared with the impregnation
method (42.8% conversion and 16.2% alkane yield). The kinetics of octadecanol hydrotreatment revealed that
the activation energy over 60%Ni–Fe/ZrO2 was much lower than Ni–Fe/ZrO2-2 (78.48 kJ/mol vs.191.90 kJ/
mol). From the catalyst characterizations, the excellent performance of 60%Ni–Fe/ZrO2 was ascribed to its fine
Ni particles, abundant Ni sites, high surface area, and coordinated acidity. Moreover, the mechanism of lipid
deoxygenation over 60%Ni–Fe/ZrO2 was proposed.

1. Introduction diesel at any proportion [4].


During deoxygenation, as the intermediates the fatty acids would
As a kind of liquid fuel, diesel has been playing an important role in generate hydrocarbons via three possible pathways: decarboxylation
the industrial and transportation sectors. At present, diesel derives pri­ (DCX), decarbonylation (DCN) as well as hydrodeoxygenation (HDO).
marily from the thermal cracking of crude oil and Fischer-Tropsch The hydrocarbons with the same carbon number as fatty acids are
synthesis of CO and H2 [1,2]. The excessive use of nonrenewable generated via HDO, while the hydrocarbons with one carbon less are
crude oil and coal has led to a severe environmental pollution, therefore, produced via DCX and DCN [5,6]. The pathways closely depend on the
exploring a green and sustainable resource to produce diesel is a sig­ catalyst kind. By now the noble catalyst (Pd, Pt) and vulcanized mo­
nificant project worldwide. Cooking oil, inedible vegetable oil, and lybdenum catalyst are commonly used for green diesel production [7,8].
microalgae oil are considered promising substitutes for fossil fuels Unfortunately, the expensive cost of noble metals and environmental
because they are inedible and productive. These lipid resources can be issues from sulfur agents restrain respectively the development of both
converted into fatty acid methyl ester (FAME) via transesterification. catalysts. The molybdenum carbide was currently applied in lipid
The FAME, namely first-generation bio-diesel can replace partly the deoxygenation and exhibited decent catalytic performance [9–11].
traditional diesel but the percentage of FAME in diesel is very limited Whereas, the high temperature (above 700 ◦ C) under H2 during the
(less than 10 vol%) due to its low poor cold-flow properties and high carburization process are not beneficial for industrial application.
oxygenation content [3]. Through deoxygenation (hydrotreatment), the With cheap cost, sulfur-free, and relatively low reduction tempera­
lipids can be transformed into C15–C18 alkanes, namely ture, the Ni catalyst draws lots of attention [12]. However, the poor
second-generation biodiesel or green diesel, which can replace fossil dispersion and the agglomeration of Ni metal lead to deficient Ni active

* Corresponding author.
** Corresponding author.
E-mail addresses: feiwang@njnu.edu.cn (F. Wang), 62081@njnu.edu.cn (X. Jiang).

https://doi.org/10.1016/j.renene.2022.07.079
Received 17 February 2022; Received in revised form 27 April 2022; Accepted 14 July 2022
Available online 1 August 2022
0960-1481/© 2022 Published by Elsevier Ltd.
F. Wang et al. Renewable Energy 197 (2022) 40–49

sites, thus limiting the catalytic performance of Ni catalyst [13,14]. Å) radiation source at 40 kV. The XRD patterns were recorded from
Interestingly, the Ni catalyst could be improved by other metals and 10.0◦ to 90◦ at a scan rate of 10.0◦ /min. The N2 physical sorption was
supports. The Cu [15,16], Sn [17], Mo [18], Pt [19], and Zn [20] have performed on Quanta NOVA2200e at − 196 ◦ C to detect the textural
been used to promote the activity of Ni catalyst for lipid hydrotreatment. properties of catalysts. Before analysis, the samples were degassed at
In our previous study, the NiCu/Al2O3 catalyst was used for green diesel 200 ◦ C for 10 h. Prodigy Plus ICP-OES was used to identify the metal
production. The Cu addition improved the dispersion of Ni sites and content in catalysts. Before analysis, the calcined catalysts were firstly
enhanced the electron density of Ni, leading to high catalytic activity in dissolved with HCl and HNO3 solution. X-ray photoelectron spectros­
lipid hydrotreatment [21]. The Al2O3, SiO2, zeolite, and carbon material copy (XPS) was carried out on ESCALAB Xi+ (Thermo Scientific) using
are commonly utilized as the supports of Ni catalyst. When ZrO2 was an Al Kα X-ray source. The XPS peaks were fitted by Avantage software.
used as support, the Ni catalyst showed a high catalytic performance due H2-TPR of catalysts was conducted on the BelCata-II (MicrotracBEL)
to its oxygen vacancy [22]. The Ni loading in the catalyst prepared with to analyze the reduction property of catalysts. About 0.1 g calcined
the wetness impregnation method is normally less than 25% because catalyst precursor was loaded into the reactor and pretreated at 300 ◦ C
when the Ni loading further increases, the agglomeration of Ni metal for 1 h in the 50 mL/min He. When the temperature was cooled down to
occurs and the extra Ni blocks the pores of catalysts. Therefore, 40 ◦ C, the gas was switched to 75 mL/min 10%H2/Ar. After the TCD
increasing Ni loading and simultaneously obtaining high dispersion of signal was stable, the reactor was heated to 800 ◦ C with a rate of 10 ◦ C/
the active site and surface area of the catalyst is an effective strategy to min. The H2 consumption was evaluated by TCD. Before the gas went
prepare the catalyst with high performance in lipid deoxygenation [23]. through the TCD, the water from catalyst reduction was removed by a
Christos et al. [24] prepared the Ni/Al2O3 catalyst with 60 wt% Ni cold trap containing a mixture of dry ice and acetone.
loading by co-precipitation method and the catalyst exhibited good NH3-TPD was conducted on the AutoChem II 2920 (Micromeritics) to
performance in sunflower oil hydrotreatment into green diesel. More­ determine the acidity of catalysts. The calcined solid was firstly reduced
over, the Zn [25] and Mo [18] addition promoted the catalytic perfor­ in the instrument at 400 ◦ C for 3 h in 60 mL/min 10%H2/Ar, followed by
mance of the co-precipitated Ni/Al2O3 catalyst. pretreatment at 450 ◦ C for 30 min in Ar flow. After that, the reactor was
So far, for the co-precipitated Ni catalysts only Al2O3 was used as purged with 50 mL/min 4% NH3/He for 1 h, and then swept with 100
support and the promotion metals were very limited (Zn or Mo). Herein, mL/min He for 1 h to remove the physically adsorbed NH3. Finally, the
by co-precipitation method the Ni–Fe/ZrO2 catalyst with extremely high catalyst was heated from 50 to 800 ◦ C, and the TCD was used to detect
Ni loading was synthesized to transform lipids into green diesel. The the desorbed NH3.
Ni–Fe interaction was deeply studied, and the equilibrium between Ni Acetic acid-TPD was also conducted on the AutoChem II 2920
sites and oxygen vacancy was investigated since the increase of Ni (Micromeritics) coupled with a mass spectrometer (MS, Pfeiffer, Ger­
loading resulted in the decrease of ZrO2 content. Moreover, the mech­ many) to study the process of organic acid hydrotreatment over cata­
anism of lipid deoxygenation and the kinetics of octadecanol hydro­ lysts. The calcined solid was firstly reduced in the instrument at 400 ◦ C
treatment over Ni–Fe/ZrO2 catalyst were also investigated. for 3 h in 60 mL/min 10%H2/Ar, followed by pretreatment at 450 ◦ C for
30 min in Ar flow. After that, the reactor was cooled down to 50 ◦ C, and
2. Experiment the gas was switched to 10% acetic acid/He for 1 h. And then the He was
used to sweep the acetic acid which was adsorbed physically on the
2.1. Catalyst preparation catalyst for 1 h. Finally, the catalyst was heated from 50 to 800 ◦ C, and
the MS was used to detect the desorbed CH3COOH, CO2, CO, CH4, and
Ni–Fe/ZrO2 catalysts with various Ni loading (20, 30, 40, 50, 60, and H2.
70 wt%) were synthesized with the co-precipitation method and the Ni/
Fe ratio was consistently fixed at 4:1. First of all, the Ni(NO3)2⋅6H2O, Fe 2.3. Lipid deoxygenation
(NO3)3⋅9H2O, and Zr(NO3)4⋅5H2O were dissolved in 200 mL of deion­
ized water to get solution A. 150 mL ammonia water with pH 8 was The lipid deoxygenation was carried out in a 100 mL batch reactor.
prepared and denoted as solution B. Then solution A was dropped slowly Before the deoxygenation reaction, the catalysts were reduced in a tube
into solution B in 2 h, and the pH of the mixture was maintained at 8 by furnace under 10%H2/N2 flow at 400 ◦ C for 3 h. Then the catalysts were
adding extra ammonia water during the mixing process. After dropping, passivated under 1%O2/N2 flow for 1 h at ambient temperature to avoid
the mixture was continuously stirred for 14 h. The solid was collected by the acute oxidization of catalyst in the air. After that, 0.25 g catalyst,
vacuum filtration and washed several times until the pH reached 7. After 20.0 g dodecane, and 2.0 g feedstock were transferred into the batch
that, the solid was dried in the oven at 105 ◦ C for 12 h, followed by reactor and the stirrer was turned on at 500 rpm. The reactor was purged
calcination in a furnace at 550 ◦ C for 3 h. Before lipid deoxygenation, the with H2 three times to sweep away the air and subsequently kept at 2.0
oxidized catalysts were reduced to obtain x%Ni–Fe/ZrO2 catalyst, in MPa. Then the reactor was heated electrically to the final temperature
which the x% meant the Ni loading. In addition, the Ni–Fe on Al2O3, and maintained for 3 h. After the reaction, the liquid product was
CeO2, Al2O3–ZrO2 (weight ratio 3:2), and Al2O3–CeO2 (weight ratio 3:2) separated from the catalyst by filtration, and the product absorbed in the
with 60 wt% Ni loading were also prepared with the same method. catalyst was rinsed with chloroform. After each reaction, the mass bal­
For comparison, the Ni–Fe/ZrO2 catalyst with 20 wt% Ni and 5 wt% ance (total mass after reaction/total mass before reaction) was calcu­
Fe loading was prepared by the wetness impregnation method. The ZrO2 lated and it was all above 96%. The vegetable oil (sunflower seed oil),
support was firstly synthesized with the above precipitation method microalgae oil, oleic acid (89 wt% C18:2 and 11 wt% C16:2), stearic acid
(specific surface area: 32 m2/g). Afterward, the 2.0 ZrO2, 2.6 g Ni (43 wt% C18:0 and 57 wt% C16:0), and octadecanol were respectively
(NO3)2⋅6H2O, and 1.0 g Fe(NO3)3⋅9H2O were mixed in 12.0 g deionized used as feedstock. The corresponding fatty acids of vegetable oil con­
water and stirred for 3 h. Then the water was removed by rotary tained 7.3 wt% C16:1 and 92.7 wt% C18:1. The composition of corre­
evaporator and dried in an oven at 105 ◦ C for 12 h. After calcination and sponding fatty acids in microalgae oil was as follows: 41.1 wt% C16: 1,
reduction, the 20 wt%Ni-5 wt.%Fe/ZrO2 catalyst was obtained and 49.7 wt% C22: 6, 4.2 wt% C14:1, 2.1% C18:1 and 2.9 wt% C20:1.
labeled as Ni–Fe/ZrO2-2. The composition of the product was identified by GC (Agilent 7890
A) and GC-MS (Agilent 8890-7000D). The capillary chromatographic
2.2. Catalyst characterization column HP-5MS (30 m × 0.25mm × 0.25 μm) was used along with a
temperature program: 50 ◦ C for 2 min; heating to 280 ◦ C with a ramp of
The phase of catalysts was identified by X-ray diffraction (XRD) 5 ◦ C/min and holding 1 min. The inlet temperature was 260 ◦ C and the
patterns on the D/max 2500/PC diffractometer with a Cu-Kα (λ = 1.54 split ratio was 50:1. The temperature of the MS source was set to 260 ◦ C

41
F. Wang et al. Renewable Energy 197 (2022) 40–49

and the scan range of MSD was 30–500 Da. Neicosane was used as the
internal standard to quantify each composition. Prior to GC and GC-MS
analysis, the product was methylated with TMAH to eliminate uncon­
verted fatty acids and triglycerides. Typically, when the phenolphtha­
lein was used as an indicator, TMAOH solution (25% in methanol) was
dropped into the sample, solved in the mixture of chloroform and
methanol. Based on the GC results, the conversion, yield of products,
HDO/DC, and cracking ratio were respectively calculated by equations
(1)–(4):

Conversion = (CF–CP) / CF × 100% (1)

CP represents the carbon moles of reactant in the product, and CF is


the carbon mole of feedstock, respectively.

Alkane (or fatty alcohols) yield = CA (or CO) / CF × 100% (2)

CA and CO are the carbon mole of alkanes and fatty alcohols in the
product, while CF represents the carbon mole of the initial feedstock.

Light alkane yield (Cracking degree) = C13-14 / CF × 100% (3)

C13-14 is the total carbon moles of light alkanes (carbon numbers 13


and 14), which were produced from the cracking reaction. CF represents
the carbon mole of the initial feedstock. Fig. 1. XRD spectra of various Ni–Fe catalysts.
HDO/DC = (C16 + C18) / (C15 + C17) × 100% (4)
and 60%Ni–Fe/Al2O3, respectively. Notably, the m-ZrO2 (PDF 37–1484)
C15, C16, C17, and C18 are respectively the relative carbon moles of was detected in Ni–Fe/ZrO2-2, while the t-ZrO2 (PDF 34–1084) was
pentadecane, hexadecane, heptadecane, and octodecane in the present in 30%Ni–Fe/ZrO2 and 60%Ni–Fe/ZrO2. The pure ZrO2 support
products. was synthesized by the same method with the 60%Ni–Fe/ZrO2 catalyst,
but its XRD spectrum only showed the m-ZrO2 phase (Fig. S1). This
2.4. Kinetics of octadecanol deoxygenation over Ni–Fe catalysts indicated that the presence of Ni and Fe led to the m-ZrO2 transforming
to t-ZrO2. Based on the diffraction peak at 43.3◦ , NiO crystalline size was
The kinetics of octadecanol deoxygenation over 60%Ni–Fe/ZrO2 and calculated with the Scherrer equation (Table 1). The NiO crystalline size
Ni–Fe/ZrO2-2 catalysts were carried out to analyze the intrinsic activity of 60%Ni–Fe/ZrO2 was only 5.6 nm. In contrast, the Ni–Fe/ZrO2-2
of both catalysts. The elementary reaction from octadecanol to alkane prepared with the wetness impregnation method had a crystalline size of
was the rate-determining step in the lipid deoxygenation, therefore, the 21.1 nm, although its Ni loading was only 20%. This indicated that the
octadecanol was used as the substrate in the kinetics study. Before the co-precipitation method formed the fine NiO crystalline grain. The NiO
kinetics experiment, the diffusion limitation from the particle size of the grain size of 60%Ni–Fe/Al2O3 and 60%Ni–Fe/CeO2 was respectively
catalysts and the agitation speed was verified and the 0.1–0.15 mm 7.7 nm and 6.5 nm. Moreover, the NiO crystalline size of 60%Ni/ZrO2
(particle size of catalyst) and 500 rpm (agitation speed) were adopted. was 14.1 nm, much higher than that of 60%Ni–Fe/ZrO2. It revealed that
The octadecanol deoxygenation over each catalyst was severally con­ the Fe addition promoted the dispersion of Ni and the decrease of NiO
ducted at three reaction temperatures. Each reaction lasted 300 min, grain size.
and the samples were taken per 30 min during the reaction. The con­
version from octadecanol to alkane conformed to the equation (5) of 3.1.2. Nitrogen physical sorption
first-order reaction [26]: The nitrogen physical sorption was used to determine the textural
characters of catalysts. All catalysts exhibited typical Ⅳ isotherm with
ln (1/(1-x)) = kT⋅t (5) an H1 hysteresis loop, which was associated with capillary condensation
in mesopores (Fig. 2) [27]. From Table 1, the surface area and pore
where the x, kT, and the t represent the conversion, reaction rate, and the
volume of 60%Ni–Fe/ZrO2 catalyst were respectively 93.7 m2/g and
reaction time (s), respectively.
0.22 cm3/g, much higher than that of Ni–Fe/ZrO2-2 with 20% Ni
From the Arrhenius equation (6), the activation energy can be
calculated:
Table 1
ln kT = -Ea / (R⋅T) + ln A (6)
The textural properties and NiO crystalline size of catalysts.
where the Ea, T, and A were respectively activation energy, temperature Samples N2 physical sorption NiO crystalline
(K), and the pre-exponential factor. size (nm)
Surface area Pore volume Pore size
(m2/g) (cm3/g) (nm)
3. Results and discussion 60%Ni–Fe/ 144.9 0.26 5.6 7.7
Al2O3
3.1. Catalyst characterization 60%Ni–Fe/CeO2 70.9 0.20 8.4 6.5
60%Ni–Fe/ZrO2 93.7 0.22 7.1 5.6
Ni–Fe/ZrO2-2 23.3 0.07 8.1 21.1
3.1.1. XRD patterns
60%Ni/ZrO2 47.3 0.15 6.3 14.1
The XRD patterns of Ni–Fe catalysts were shown in Fig. 1. All cata­ 15%Fe/ZrO2 99.6 0.16 3.7 -
lysts displayed the typical diffraction peaks at 37.2◦ (111), 43.3◦ (200), 60%Ni–Fe/ 122.8 0.24 5.5 6.5
62.9◦ (220), 75.4◦ (311), and 79.4◦ (222), indexed to the NiO phase Al2O3–ZrO2
(PDF 47–1049) [17]. In addition, the CeO2 (PDF 34–0394) and Al2O3 60%Ni–Fe/ 129.4 0.26 5.9 6.0
Al2O3–CeO2
(PDF 50–1496) were also identified in the spectrum of 60%Ni–Fe/CeO2

42
F. Wang et al. Renewable Energy 197 (2022) 40–49

Fig. 2. Nitrogen sorption isotherms (left) and pore size distribution (right) of catalysts.

loading. This manifested the Ni–Fe/ZrO2 catalyst from the and supports.
co-precipitation method has a higher surface area than that from the
impregnation method. In addition, the surface area of 60%Ni/ZrO2 and 3.1.4. NH3-TPD of catalysts
15%Fe/ZrO2 were severally 47.3 and 99.6 m2/g, much lower than that NH3-TPD was conducted to determine the acidity of catalysts and the
of 60%Ni–Fe/ZrO2, demonstrating the Fe addition increased the surface curves were shown in Fig. 4, from which the acidity could be divided
areas of catalyst. Moreover, when the Al2O3 acted as support, the surface into weak acidity (<400 ◦ C, Lewis acidity) and strong acidity (>400 ◦ C,
area attained the maximum (144.9 m2/g) and the pore size was only 5.6 Brønsted acidity) [29]. The areas of weak acidity and strong acidity were
nm, which was less than 7.1 nm of 60%Ni–Fe/ZrO2. respectively calculated by integrating the background line between the
curves from 50 to 400 ◦ C and 400–800 ◦ C. Based on the total acidity
3.1.3. H2-TPR of catalysts amount and the percentages of both acidities, the amounts of weak
The H2-TPR of calcined catalysts was conducted to analyze the acidity and strong acidity were identified, which was displayed in
reduction property of catalysts, and the corresponding curves were Table 2. It was worthy to note that a peak of medium acidity at 250 ◦ C
shown in Fig. 3. The H2-TPR of 60%Ni/ZrO2 and 15%Fe/ZrO2 (Fig. S2) appeared when the ZrO2 was used as the support and the weak acidity
displayed that the reduction peak of NiO centered at 438 ◦ C, and the iron was primary. For the 60%Ni–Fe on three supports, the amount of total
oxide had three reduction peaks at 384, 550, and 700 ◦ C. In the H2-TPR acidity in 60%Ni–Fe/Al2O3 reached up to 6.30 mmolNH3/g, much higher
curve of 60%Ni–Fe/ZrO2, a small peak at 350 ◦ C and a strong, board than that in 60%Ni–Fe/ZrO2 (4.55) and 60%Ni–Fe/CeO2 (3.42). How­
peak at 500 ◦ C were formed. The former was assigned to the reduction of ever, the ratio of weak acidity/strong acidity (W/S ratio) in 60%
Fe2O3 to FeO, while the latter was attributed to the reduction of NiO and Ni–Fe/ZrO2 was 5.5, much higher than the other catalysts (1.9 and 2.8).
the reduction of FeO to Fe [28]. Among the three supports, the CeO2 Moreover, both values of the total acidity amount and W/S ratio in 60%
supported 60%Ni–Fe had a very similar reduction property with ZrO2, in Ni–Fe/ZrO2 were much higher than that in the Ni–Fe/ZrO2-2 catalyst
contrast, the reduction temperature of 60%Ni–Fe/Al2O3 was the highest synthesized by the impregnation method.
(650 ◦ C). It can be inferred that compared to the Al2O3 support, the CeO2
and ZrO2 facilitated the reduction of nickel oxide and iron oxide. As for 3.1.5. XPS of catalysts
the Ni–Fe/ZrO2-2 catalyst, the reduction peaks were rather lower, In the XPS survey spectra (Fig. S3), the Ni, Fe, and the metals of
indicating the Ni–Fe/ZrO2 catalyst, prepared by the wetness impreg­ supports (Zr, Ce, and Al) were present in the corresponding catalysts.
nation method, was easier to be reduced than that prepared with the The high-resolution O 1s XPS spectra (Fig. 5) can be fitted into two peaks
co-precipitation method, because of the weak interaction between Ni–Fe at 529.9 eV and 531.8 eV, which were severally due to the lattice oxygen

Fig. 3. H2-TPR curves of Ni–Fe catalysts on various supports. Fig. 4. NH3-TPD curves of various Ni–Fe catalysts.

43
F. Wang et al. Renewable Energy 197 (2022) 40–49

Table 2 3.2. Catalytic deoxygenation of lipids


The distribution of various acid sites.
Catalyst Total acidity amount Amount of two W/S 3.2.1. Catalyst preparation method and catalyst support
(mmolNH3/g) acidities ratio The oleic acid deoxygenation over 60%Ni–Fe/ZrO2, 60%Ni–Fe/
(mmolNH3/g) Al2O3, 60%Ni–Fe/CeO2, 60%Ni/ZrO2, and Ni–Fe/ZrO2-2 was per­
Weak Strong formed. The GC spectrum of liquid product from catalytic deoxygen­
Ni–Fe/ZrO2-2 3.14 2.40 0.74 3.2
ation of oleic acid was shown in Fig. S4. The product mainly contained
60%Ni–Fe/ 4.55 3.85 0.70 5.5 alkanes, fatty alcohols (hexadecanol and octadecanol), and fatty acids
ZrO2 (octadecanoic acid, oleic acid, palmitic acid, hexadecenoic acid) which
60%Ni–Fe/ 6.30 4.10 2.21 1.9 were considered unconverted substrates. Among alkanes, the C15, C16,
Al2O3
C17, and C18 were generated from oleic acid hydrotreatment via HDO
60%Ni–Fe/ 3.42 2.51 0.91 2.8
CeO2 and DCX/DCN routes, while the light alkanes with carbon number less
than 15 were originated from the cracking reaction. The catalytic per­
formance over 5 catalysts was displayed in Fig. 7. Under the reaction
(OL) and the defect oxygen (OD) [30,31]. Based on the integral areas of conditions, the deoxygenation of oleic acid over 60%Ni–Fe/ZrO2
the fitted peaks, the percentage was calculated in Table 3. When the
Al2O3 was used as the support, the OD percentage was maximum
(55.5%), demonstrating the defect oxygen was easily formed in Al2O3 Table 3
among the three supports. Compared to 60%Ni–Fe/ZrO2, Ni–Fe/ZrO2-2 The ratio of different states of O and Fe in catalysts.
catalyst had a little more OD. In addition, the peaks of OL and OD shifted Catalyst Ratio of O (%) Ratio of Ni (%)
to 529.5 and 531.2 eV, respectively. In the Ni 2p spectra, the doublet OL OD Ni II Ni III Satellite 1 Satellite 2
peaks at 854.7 and 872.5 eV represented the Ni2+, while the doublet
Ni–Fe/ZrO2-2 64.0 36.0 25 29 43 3
peaks at 856.7 and 873.9 eV revealed the presence of Ni3+ [16,32]. 60%Ni–Fe/ZrO2 70.3 29.7 29 26 39 6
Notably, in the 60%Ni–Fe/ZrO2, the percentage of Ni2+ was higher than 60%Ni–Fe/Al2O3 44.5 55.5 31 25 36 8
Ni3+, however, it was reversed in the Ni–Fe/ZrO2-2. In the Fe 2p spectra, 60%Ni–Fe/CeO2 71.4 28.6 27 29 31 13
the binding energy at 711.8 and 724.7 eV with a satellite peak at 718.7
eV were assigned to Fe3+ [33,34], while the binding energy at 704.2 eV
corresponded to the pre-peak of Fe 2p3/2 [35].

3.1.6. Acetic acid-TPD


The acetic acid-TPD was used to study the hydrotreatment process
over 60%Ni–Fe/ZrO2 and Ni–Fe/ZrO2-2 catalyst, and the curves were
shown in Fig. 6. The MS signals of 43, 44, 28 (deducting the signal from
CO2), 15, and 2 represented respectively the CH3COOH, CO2, CO, CH4,
and H2. Notably, from 50 to 550 ◦ C only very little acetic acid was
desorbed at about 250 ◦ C for both catalysts. This result was different
from the acetic acid-TPD over Cu–Ni alloy catalyst, where the obvious
desorption of acetic acid happened from 100 to 400 ◦ C [36]. It can be
deduced that the acetic acid was easily transformed into other com­
pounds over the Ni–Fe catalysts. For the 60%Ni–Fe/ZrO2 catalyst, when
the temperature reached 220 ◦ C, the CO and CO2 signal became strong
due to the DCN and DCX reaction. The CO2 signal had three peaks at
290, 360, and 420 ◦ C, and but for the CO signal only one broad desorbed
peak from 250 to 380 ◦ C appeared. Meanwhile, a strong H2 signal and a
moderate CH4 signal also came out, which were formed from the com­
bination of H* and CH3* [37,38]. However, in the acetic acid-TPD over
Ni–Fe/ZrO2-2, the peak of CO2 at about 360 ◦ C disappeared, while a
Fig. 6. The MS signal during acetic acid-TPD over 60%Ni–Fe/ZrO2 and Ni–Fe/
strong CO peak at 470 ◦ C emerged, indicating the DCX became weaker
ZrO2-2.
and the DCN route got higher than that over 60%Ni–Fe/ZrO2.

Fig. 5. High-resolution O 1s XPS (left) and Fe 2p (right) of calcined catalysts.

44
F. Wang et al. Renewable Energy 197 (2022) 40–49

catalyst attained 98.7% conversion and 91.8% alkane yield (light alkane 3.2.2. Ni content in Ni–Fe/ZrO2 catalysts
yield 2.0%). In addition, the HDO/DC ratio was 0.13, which indicated By using the co-precipitation method, the Ni–Fe/ZrO2 catalysts were
the DCX/DCN was predominant among three hydrotreatment routes. endowed with large Ni loading, high Ni dispersion, and surface area.
The hydrotreatment route depended highly on the catalyst type, among Hence, the Ni loading (from 20 wt% to 70 wt%) of catalysts was
which the Ni catalyst preferred to DCX/DCN pathway [36]. Notably, the investigated in this part. From the XRD patterns (Fig. S3), as the Ni
60%Ni/ZrO2 only got 44.4% conversion and 33.2% alkane yield, man­ content increased, the NiO diffraction peaks became stronger while the
ifesting the Fe promoted greatly the activity of Ni catalyst. However, the peaks of t-ZrO2 varied in the opposite trend. As the Ni loading rose from
improvement of activity was not attributed to the catalytic activity of Fe 20 to 70 wt%, the average particle size of NiO was increased from 4.3 nm
metal because 15%Fe/ZrO2 only got 7.6% conversion of oleic acid to 6.5 nm due to agglomeration of partial Ni (Table 4). Fortunately, even
(Table S1). It can be deduced that the interaction between Fe and Ni though the Ni loading was 70 wt%, the NiO particle size was still much
enhanced the catalytic activity. From the above characterization, the lower than that of Ni–Fe/ZrO2-2 (6.5 nm vs. 21.1 nm). The N2 physical
presence of Fe produced fine NiO particles and enlarged the surface area sorption in Table 4 demonstrated the augment of Ni loading from 20 to
of the catalyst, exposing more Ni sites for the deoxygenation reaction. 50 wt% led to the decrease of surface area from 135.4 to 99.8 m2/g. But
Moreover, the Fe catalyst had a very high hydrogen enrichment capac­ the surface area of the catalyst barely changed with the further increase
ity, and the hydrogen upon the surface was easily dissociated [39]. of Ni loading.
Therefore, the Fe site on 60%Ni–Fe/ZrO2 might influence the catalytic The catalytic deoxygenation of oleic acid as the function of Ni
performance by facilitating the hydrogenation of fatty acid into alde­ loading of Ni–Fe/ZrO2 catalysts was displayed in Fig. 8. It was clear to
hyde/alcohol. The cracking degree over 60%Ni/ZrO2 was 5.1, much find that all catalysts achieved a conversion above 90%, revealing the
higher than 60%Ni–Fe/ZrO2. The strong cracking ability of Ni catalyst in catalysts had a high catalytic activity. As the Ni loading increased from
lipid deoxygenation was found in the previous report [19], and the 20 to 60 wt%, the alkane yield augmented continuously from 36.5% to
excess cracking reaction easily resulted in catalyst deactivation [17]. It 91.8% with the reduction of alcohol yield from 53.9% to 6.9%. How­
was worthy to note that the hydrotreatment of oleic acid over ever, the further increase of Ni loading resulted in the decrease of alkane
NiFe/ZrO2-2 catalyst only obtained 42.8% conversion and 16.2 alkane yield to 20.6%. As the Ni loading rose from 20 to 60 wt%, more Ni sites
yield, much lower than that of 60%Ni–Fe/ZrO2. In addition, the reaction were generated, thus increasing alkane yield. The further increase of Ni
rates of oleic acid deoxygenation over both catalysts were determined at loading led to the Ni agglomeration and fewer Ni sites, which decreased
low conversion, as listed in Table S2. The reaction rate over 60% alkane yield. Moreover, the increase of Ni loading elevated the HDO/DC
Ni–Fe/ZrO2 was greatly higher than the value over NiFe/ZrO2-2 catalyst
(0.0071 vs. 0.0008 mol/(gcat⋅h)), which was closely related to its high
Table 4
surface area, abundant Ni sites, and coordinated acidity.
The textural properties and NiO particle size of catalysts with different Ni
For 60%Ni–Fe on three supports, the catalytic performance was in loading.
the order of ZrO2 > Al2O3> CeO2, in which the conversion and alkane
Samples N2 physical sorption NiO particle
yield over 60%Ni–Fe/Al2O3 were respectively 97.7% and 83.6%. The size (nm)
difference of activity between 60%Ni–Fe/Al2O3 and 60%Ni–Fe/ZrO2 Surface area Pore volume Pore size
(m2/g) (cm3/g) (nm)
can be ascribed to the lower NiO grain size (Table 1) and better reduc­
tion property (Fig. 3) of the later catalyst. Interestingly, the conversion 20%Ni–Fe/ 135.4 0.18 4.6 4.3
over Ni–Fe/CeO2 reached 92.3%, however, the alkane yield was only ZrO2
30%Ni–Fe/ 121.4 0.18 4.9 4.6
19.6% with the alcohol yield of 72.6%. It revealed this catalyst possessed ZrO2
a very high selectivity for intermediate fatty alcohol instead of alkanes 40%Ni–Fe/ 100.4 0.19 5.0 4.7
owing to its pore property and acidity. Moreover, the catalytic activity of ZrO2
60%Ni–Fe on oxide-mixture supports (Al2O3–ZrO2 and Al2O3–CeO2) 50%Ni–Fe/ 99.8 0.20 6.2 5.0
ZrO2
were also tested and both catalysts showed much lower performance
60%Ni–Fe/ 93.7 0.22 7.1 5.6
than 60%Ni–Fe/ZrO2 (Table S3). Herein, the ZrO2 was deemed as the ZrO2
optimal support for 60%Ni–Fe to produce alkanes from deoxygenation 70%Ni–Fe/ 90.1 0.19 7.2 6.5
of oleic acid. ZrO2

Fig. 7. Deoxygenation of oleic acid over 60%Ni–Fe catalysts on various supports


(Reaction condition: temperature 240 ◦ C, initial hydrogen pressure 2.0 MPa, time 3 h).

45
F. Wang et al. Renewable Energy 197 (2022) 40–49

Fig. 8. The effect of Ni loading on catalytic deoxygenation of oleic acid


(Reaction condition: temperature 240 ◦ C, initial hydrogen pressure 2.0 MPa, time 3 h).

ratio, indicating the HDO route was enhanced over larger Ni particles. increased, the conversion kept stable. The increase of reaction temper­
Among 6 catalysts, the cracking degree over 60%Ni–Fe/ZrO2 was min­ ature from 200 to 260 ◦ C greatly improved the alkane yield. At the
imum due to its highest number of Ni sites. temperature of 200 and 220 ◦ C, the alcohol dominated in the products,
revealing that the rate of alcohol formation was higher than that of
3.2.3. Feedstock and deoxygenation conditions alkane formation from alcohol. Moreover, with the rise in temperature,
From the above parts, the 60%Ni–Fe/ZrO2 was considered as the the HDO/DC ratio decreased gradually from 0.32 to 0.10, as evidence of
optimal catalyst for the deoxygenation of oleic acid into green diesel. enhancement of the DCX/DCN pathway. The same phenomenon was
Since the vegetable oil and microalgae oil were two kinds of promising also found in relative studies [5,41]. The influence of initial H2 pressure
feedstocks for green diesel production, the catalytic deoxygenation of on the catalytic performance was shown in Fig. 9 (right). It was worthy
both triglycerides and the intermediates (stearic acid and octadecanol) to note that the conversion was merely zero in the absence of H2, and
was also carried out (Table 5). The deoxygenation of stearic acid ob­ when the H2 pressure was 1.0 MPa the conversion attained 55.8%. With
tained a similar conversion to oleic acid, but the alkane yield was higher the further increase of H2 pressure, the conversion of vegetable oil was
(99.1% vs. 91.8%). During oleic acid hydrotreatment, the first step was greatly elevated to 97.98% at the 3.0 MPa. This indicated that the H2
the hydrogenation of oleic acid into stearic acid, but this step can’t was necessary for the deoxygenation of vegetable oil even though the
explain the higher alkane yield from the hydrogenation of stearic acid DCX/DCN route dominated over 60%Ni–Fe/ZrO2 catalyst.
because the hydrogenation rate was highly rapid. From the previous Based on the above results, the mechanism for triglyceride hydro­
studies, the unsaturated lipid triggered readily the coking reaction that treatment was proposed as displayed in Fig. 10. During the deoxygen­
weakened the catalytic activity [10,40]. However, for the deoxygen­ ation of triglyceride, the first step was the β elimination of triglyceride to
ation of triglyceride (vegetable oil and microalgae), the alcohol yield generate unsaturated fatty acid with a very high reaction rate, followed
was high at the cost of alkane yield. During the hydrotreatment of tri­ by the hydrogenation reaction producing saturated fatty acid. After that,
glyceride, the oleic acid existed as an intermediate, which derived from the saturated fatty acid could be either converted into alkane with one
β elimination of triglyceride, as shown in Fig. 10. This reaction occurred carbon less via the DCX route or transformed into aliphatic aldehyde
rapidly that made merely a little impact on the hydrotreatment [6]. The followed by the DCN route to yield the same alkane. However, the result
low alkane yield in triglyceride deoxygenation was closely related to in Fig. 9 (right) revealed that H2 was necessary for the production of
that the large molecular triglyceride was not easily catalyzed by the alkane, which indicated that herein the pentadecane and heptadecane
active sites in the pores of catalysts because of the steric hindrance. (with one carbon number less than corresponding fatty acids) were
The reaction temperature and initial hydrogen pressure in deoxy­ mainly derived from the DCN, rather than DCX route. This conclusion
genation of vegetable oil over 60%Ni–Fe/ZrO2 were investigated was different from previous reports, where the DCX played a dominating
(Fig. 9). It was obvious that the temperature played an important role in role in heptadecane production during lipid hydrotreatment [42–44].
the reaction. As the temperature rose from 200 to 240 ◦ C, the conversion Sarkar et al. [45] studied the competition of DCX and DCN route of fatty
increased from 10.3 to 98.7%. But when the temperature further acid over Pd catalyst with the DFT, and proved that the initial Cα-H
activation was stronger than the initial OH activation leading to the
preference of DCN over the DCX pathway. Aliphatic alcohol from the
Table 5 hydrogenation of olefin alcohol (tautomer of aliphatic aldehyde) could
The deoxygenation of various feedstocks over 60%Ni–Fe/ZrO2 catalysta.
be transformed into hexadecane and octadecane (with the same carbon
Substrate Conversion Alkane Alcohol Cracking HDO/ number as corresponding fatty acids) via the HDO route. Herein, the
(%) yield (%) yield (%) degree (%) DC
pentadecane and heptadecane from DCN were predominant in alkanes,
Oleic acid 98.7 ± 0.24 91.8 ± 6.9 ± 2.0 ± 0.56 0.13 but the aliphatic alcohol took a high proportion in products at the low
1.83 2.05 ± 0.01 reaction temperatures. It can be inferred that aliphatic alcohol could
Stearic acid 99.1 ± 0.16 99.1 ± 0 0.0 3.4 ± 0.19 0.14
form aliphatic aldehyde through dehydration and then produce alkanes
± 0.01
Octadecanol 100 100 - 2.3 ± 0.02 0.26 via the DCN route.
± 0.02
Vegetable 96.8 ± 0.44 79.2 ± 17.5 ± 3.9 ± 0.16 0.31 3.2.4. Kinetics of octadecanol hydrotreatment over 60%Ni–Fe/ZrO2 and
oil 1.02 1.07 ± 0.02
Ni–Fe/ZrO2-2
Microalgae 92.0 ± 1.64 77.0 ± 14.9 ± 1.2 ± 0.46 0.23
oil 2.60 0.96 ± 0.01 In the kinetics study, the octadecanol hydrotreatment was carried
a
out over 60%Ni–Fe/ZrO2 at 200, 220, and 240 ◦ C, while the reaction
Reaction condition: temperature 240 ◦ C, initial hydrogen pressure 2.0 MPa, over Ni–Fe/ZrO2-2 at 220, 240, and 260 ◦ C to ensure a low conversion.
time 3 h.

46
F. Wang et al. Renewable Energy 197 (2022) 40–49

Fig. 9. The catalytic deoxygenation of vegetable oil as the function of reaction temperature (left) and H2 pressure (right).

Fig. 10. The proposed mechanism for hydrotreatment of triglyceride.

From the first-order reaction equation, the kT over two catalysts at former catalyst had much higher catalytic activity than the latter.
various temperatures was identified, as shown in Fig. 11 (left). Notably, Moreover, as the reaction temperature rose, the k over both catalysts
at 240 ◦ C the k (4.30 × 10− 5 s− 1) over 60%Ni–Fe/ZrO2 was extremely was increased. By using the Arrhenius equation, the activation energy of
higher than that of 60%Ni–Fe/ZrO2-2 (6.35 × 10− 6 s− 1), indicating the octadecanol hydrotreatment over both catalysts was calculated. The

Fig. 11. Reaction rates at different temperatures (left) and Arrhenius plots (right).

47
F. Wang et al. Renewable Energy 197 (2022) 40–49

activation energy over 60%Ni–Fe/ZrO2 and 60%Ni–Fe/ZrO2-2 was [3] E.-M. Ryymin, M.L. Honkela, T.-R. Viljava, A.O.I. Krause, Insight to sulfur species
in the hydrodeoxygenation of aliphatic esters over sulfided NiMo/γ-Al2O3 catalyst,
respectively 78.48 and 191.90 kJ/mol. This manifested the energy
Appl. Catal. Gen. 358 (2009) 42–48.
barrier for the reaction of octadecanol hydrotreatment over 60%Ni–Fe/ [4] X. Yao, T.J. Strathmann, Y. Li, L.E. Cronmiller, H. Ma, J. Zhang, Catalytic
ZrO2 was much lower than 60%Ni–Fe/ZrO2-2. From the above- hydrothermal deoxygenation of lipids and fatty acids to diesel-like hydrocarbons: a
mentioned analysis, the reduction of activation energy over 60% review, Green Chem. 23 (2021) 1114–1129.
[5] F. Wang, J. Jiang, K. Wang, Q. Zhai, H. Sun, P. Liu, J. Feng, H. Xia, J. Ye, Z. Li, F. Li,
Ni–Fe/ZrO2 was ascribed to the increase of active sites, the improvement J. Xu, Activated carbon supported molybdenum and tungsten carbides for
of surface area, optimization of acidity, and the weakening of the DCN hydrotreatment of fatty acids into green diesel, Fuel 228 (2018) 103–111.
pathway. [6] R.W. Gosselink, S.A. Hollak, S.W. Chang, J. van Haveren, K.P. de Jong, J.H. Bitter,
D.S. van Es, Reaction pathways for the deoxygenation of vegetable oils and related
model compounds, ChemSusChem 6 (2013) 1576–1594.
4. Conclusion [7] N. Chen, N. Wang, Y. Ren, H. Tominaga, E.W. Qian, Effect of surface modification
with silica on the structure and activity of Pt/ZSM-22@SiO2 catalysts in
hydrodeoxygenation of methyl palmitate, J. Catal. 345 (2017) 124–134.
To improve the catalytic performance of the Ni catalyst in lipid [8] T. Burimsitthigul, B. Yoosuk, C. Ngamcharussrivichai, P. Prasassarakich,
deoxygenation, the Fe-promoted Ni catalyst with 60 wt% Ni loading Hydrocarbon biofuel from hydrotreating of palm oil over unsupported Ni–Mo
(60%Ni–Fe/ZrO2) was prepared with the co-precipitation method. In the sulfide catalysts, Renew. Energy 163 (2021) 1648–1659.
[9] F. Wang, W. Zhang, J. Jiang, J. Xu, Q. Zhai, L. Wei, F. Long, C. Liu, P. Liu, W. Tan,
deoxygenation of oleic acid, the reaction over 60%Ni–Fe/ZrO2 catalyst D. He, Nitrogen-rich carbon-supported ultrafine MoC nanoparticles for the
reached 98.7% conversion and 91.8% alkane yield, however, the reac­ hydrotreatment of oleic acid into diesel-like hydrocarbons, Chem. Eng. J. 382
tion over 20%Ni–Fe/ZrO2-2 prepared with impregnation method only (2019) 122464.
[10] F. Wang, J. Jiang, K. Wang, Q. Zhai, F. Long, P. Liu, J. Feng, H. Xia, J. Ye, J. Li,
got 42.8% conversion and 16.2% alkane yield. The activation energy of J. Xu, Hydrotreatment of lipid model for diesel-like alkane using nitrogen-doped
octadecanol hydrotreatment over 60%Ni–Fe/ZrO2 catalyst was only mesoporous carbon-supported molybdenum carbide, Appl. Catal. B Environ. 242
78.48 kJ/mol, much lower than that of 20%Ni–Fe/ZrO2-2 (191.90 kJ/ (2019) 150–160.
[11] C.-C. Tran, D. Akmach, S. Kaliaguine, Hydrodeoxygenation of vegetable oils over
mol). The characterizations of the catalysts revealed that 60%Ni–Fe/
biochar supported bimetallic carbides for producing renewable diesel under mild
ZrO2 catalyst obtained smaller Ni particle size, more Ni sites, a higher conditions, Green Chem. 22 (2020) 6424–6436.
surface area of the catalyst, and coordinated acidity compared to the [12] S. Lycourghiotis, E. Kordouli, C. Kordulis, K. Bourikas, Transformation of residual
fatty raw materials into third generation green diesel over a nickel catalyst
20%Ni–Fe/ZrO2-2. The acetic acid-TPD indicated the difference in
supported on mineral palygorskite, Renew. Energy 180 (2021) 773–786.
hydrotreatment route between both catalysts as the HDN route over [13] S. Lycourghiotis, E. Kordouli, L. Sygellou, K. Bourikas, C. Kordulis, Nickel catalysts
20%Ni–Fe/ZrO2-2 was stronger than 60%Ni–Fe/ZrO2. supported on palygorskite for transformation of waste cooking oils into green
diesel, Appl. Catal. B Environ. 259 (2019) 118059.
[14] I. Hachemi, N. Kumar, P. Mäki-Arvela, J. Roine, M. Peurla, J. Hemming, J. Salonen,
CRediT authorship contribution statement D.Y. Murzin, Sulfur-free Ni catalyst for production of green diesel by
hydrodeoxygenation, J. Catal. 347 (2017) 205–221.
[15] Z. Zhang, H. Chen, C. Wang, K. Chen, X. Lu, P. Ouyang, J. Fu, Efficient and stable
Fei Wang: Conceptualization, Methodology, Formal analysis, Vali­
Cu-Ni/ZrO2 catalysts for in situ hydrogenation and deoxygenation of oleic acid into
dation, Writing – original draft. Hui Xu: Validation, Methodology, heptadecane using methanol as a hydrogen donor, Fuel 230 (2018) 211–217.
Formal analysis, Investigation, Writing. Songyin Yu: Formal analysis, [16] C. Denk, S. Foraita, L. Kovarik, K. Stoerzinger, Y. Liu, E. Baráth, J.A. Lercher, Rate
Investigation, Methodology. Hao Zhu: Conceptualization, Methodol­ enhancement by Cu in NixCu1− x/ZrO2 bimetallic catalysts for
hydrodeoxygenation of stearic acid, Catal. Sci. Technol. 9 (2019) 2620–2629.
ogy, Formal analysis. Yuchan Du: Conceptualization, Methodology, [17] R. Loe, E. Santillan-Jimenez, T. Morgan, L. Sewell, Y. Ji, S. Jones, M.A. Isaacs, A.
Formal analysis. Zeng Zhang: Conceptualization, Methodology, Formal F. Lee, M. Crocker, Effect of Cu and Sn promotion on the catalytic deoxygenation of
analysis. Chaoqun You: Conceptualization, Review. Xiaoxiang Jiang: model and algal lipids to fuel-like hydrocarbons over supported Ni catalysts, Appl.
Catal. B Environ. 191 (2016) 147–156.
Conceptualization, Methodology, Review. Jianchun Jiang: Conceptu­ [18] E. Kordouli, B. Pawelec, K. Bourikas, C. Kordulis, J.L.G. Fierro, A. Lycourghiotis,
alization, Methodology, Review. Mo promoted Ni-Al2O3 co-precipitated catalysts for green diesel production, Appl.
Catal. B Environ. 229 (2018) 139–154.
[19] R. Loe, K. Huff, M. Walli, T. Morgan, D. Qian, R. Pace, Y. Song, M. Isaacs,
Declaration of competing interest E. Santillan-Jimenez, M. Crocker, Effect of Pt promotion on the Ni-catalyzed
deoxygenation of tristearin to fuel-like hydrocarbons, Catalysts 9 (2019) 200–219.
[20] L. Chen, Z. Feng, L. Guangci, L. Xuebing, Effect of Zn/Al ratio of Ni/ZnO-Al2O3
The authors declare that they have no known competing financial catalysts on the catalytic deoxygenation of oleic acid into alkane, Appl. Catal. Gen.
interests or personal relationships that could have appeared to influence 529 (2017) 175–184.
the work reported in this paper. [21] X. Cao, F. Long, F. Wang, J. Zhao, J. Xu, J. Jiang, Chemoselective decarboxylation
of higher aliphatic esters to diesel-range alkanes over the NiCu/Al2O3 bifunctional
catalyst under mild reaction conditions, Renew. Energy 180 (2021) 1–13.
Acknowledgment [22] S. Foraita, Y. Liu, G.L. Haller, E. Baráth, C. Zhao, J.A. Lercher, Controlling
hydrodeoxygenation of stearic acid to n-heptadecane and n-octadecane by
adjusting the chemical properties of Ni/SiO2-ZrO2 Catalyst, ChemCatChem 9
This work was supported financially by the Natural Science Foun­ (2017) 195–203.
dation of Jiangsu Province of China (BK20210566), and the General [23] P. Munnik, P.E. de Jongh, K.P. de Jong, Recent developments in the synthesis of
Project of Education Department of Jiangsu Province (21KJB480003). supported catalysts, Chem. Rev. 115 (2015) 6687–6718.
[24] M. Gousi, C. Andriopoulou, K. Bourikas, S. Ladas, M. Sotiriou, C. Kordulis,
The authors would like to thank Feng Wei from Shiyanjia Lab (www.shiy A. Lycourghiotis, Green diesel production over nickel-alumina co-precipitated
anjia.com) for the H2-TPR analysis. catalysts, Appl. Catal. Gen. 536 (2017) 45–56.
[25] G. Zafeiropoulos, N. Nikolopoulos, E. Kordouli, L. Sygellou, K. Bourikas,
C. Kordulis, A. Lycourghiotis, Developing nickel-Zirconia Co-precipitated catalysts
Appendix A. Supplementary data for production of green diesel, Catalysts 9 (2019) 210.
[26] P. Kumar, S.R. Yenumala, S.K. Maity, D. Shee, Kinetics of hydrodeoxygenation of
Supplementary data to this article can be found online at https://doi. stearic acid using supported nickel catalysts: effects of supports, Appl. Catal. Gen.
471 (2014) 28–38.
org/10.1016/j.renene.2022.07.079.
[27] F. Wang, J. Xu, J. Jiang, P. Liu, F. Li, J. Ye, M. Zhou, Hydrotreatment of vegetable
oil for green diesel over activated carbon supported molybdenum carbide catalyst,
References Fuel 216 (2018) 738–746.
[28] N. Katabathini, I.H.A. El Maksod, M. Mokhtar, Cu, Fe and Mn oxides intercalated
SiO2 pillared magadiite and ilerite catalysts for NO decomposition, Appl. Catal.
[1] H. Schulz, Short History and Present Trends of Fischer-Tropsch Synthesis, Applied
Gen. (2021) 616.
Catalysis A: General, 1999, pp. 3–12.
[29] Y. Duan, Y. Wu, Q. Zhang, R. Ding, Y. Chen, J. Liu, M. Yang, Towards conversion of
[2] M.R. Nielsena, A.B. Mossab, A.S. Bjørnlundab, X. Liucd, A. Knop-Gerickeef, A.
octanoic acid to liquid hydrocarbon via hydrodeoxygenation over Mo promoter
Y. Klyushineg, J.-D. Grunwaldthi, T.L. Sheppardhi, D.E. Doronkinhi, A. Ziminahi,
nickel-based catalyst, J. Mol. Catal. Chem. 398 (2015) 72–78.
T.E.L. Smitshuysenb, C.D. Damsgaardab, J.B. Wagnera, T.W. Hansena, Reduction
[30] B. Xu, Q. Zhang, S. Yuan, M. Zhang, T. Ohno, Morphology control and
and carburization of iron oxides for Fischer–Tropsch synthesis, J. Energy Chem. 51
characterization of broom-like porous CeO2, Chem. Eng. J. 260 (2015) 126–132.
(2020) 48–61.

48
F. Wang et al. Renewable Energy 197 (2022) 40–49

[31] K.-W. Jeon, J.-O. Shim, J.-W. Cho, W.-J. Jang, H.-S. Na, H.-M. Kim, Y.-L. Lee, B.- [38] M.S. Hofman, E.V. Scoullos, J.P. Robbins, L. Ezeonu, D.V. Potapenko, X. Yang, S.
H. Jeon, J.W. Bae, H.-S. Roh, Synthesis and characterization of Pt-, Pd-, and Ru- G. Podkolzin, B.E. Koel, Acetic acid adsorption and reactions on Ni(110), Langmuir
promoted Ni–Ce0.6Zr0.4O2 catalysts for efficient biodiesel production by 36 (2020) 8705–8715.
deoxygenation of oleic acid, Fuel 236 (2019) 928–933. [39] Y. Sun, Q. Ma, Q. Ge, J. Sun, Tunable synthesis of ethanol or methyl acetate via
[32] P. Reangchim, T. Saelee, V. Itthibenchapong, A. Junkaew, N. Chanlek, A. Eiad-ua, dimethyl oxalate hydrogenation on confined iron catalysts, ACS Catal. 11 (2021)
N. Kungwan, K. Faungnawakij, Role of Sn promoter in Ni/Al2O3 catalyst for the 4908–4919.
deoxygenation of stearic acid and coke formation: experimental and theoretical [40] S.A.W. Hollak, R.W. Gosselink, D.S. van Es, J.H. Bitter, Comparison of tungsten and
studies, Catal. Sci. Technol. 9 (2019) 3361–3372. molybdenum carbide catalysts for the hydrodeoxygenation of oleic acid, ACS Catal.
[33] Z. Ma, C. Zhou, D. Wang, Y. Wang, W. He, Y. Tan, Q. Liu, Co-precipitated Fe-Zr 3 (2013) 2837–2844.
catalysts for the Fischer-Tropsch synthesis of lower olefins (C2O ~ C4O): [41] C. Miao, O. Marin-Flores, S.D. Davidson, T. Li, T. Dong, D. Gao, Y. Wang,
synergistic effects of Fe and Zr, J. Catal. 378 (2019) 209–219. M. Garcia-Pérez, S. Chen, Hydrothermal catalytic deoxygenation of palmitic acid
[34] Y. Tian, A. Huang, Z. Wang, M. Wang, Q. Wu, Y. Shen, Q. Zhu, Y. Fu, M. Wen, Two- over nickel catalyst, Fuel 166 (2016) 302–308.
dimensional hetero-nanostructured electrocatalyst of Ni/NiFe-layered double [42] L. Boda, G. Onyestyák, H. Solt, F. Lónyi, J. Valyon, A. Thernesz, Catalytic
oxide for highly efficient hydrogen evolution reaction in alkaline medium, Chem. hydroconversion of tricaprylin and caprylic acid as model reaction for biofuel
Eng. J. 426 (2021). production from triglycerides, Appl. Catal. Gen. 374 (2010) 158–169.
[35] Z. Wu, Z. Zou, J. Huang, F. Gao, NiFe2O4 nanoparticles/NiFe layered double- [43] J. Liang, Z. Zhang, K. Wu, Y. Shi, W. Pu, M. Yang, Y. Wu, Improved conversion of
hydroxide nanosheet heterostructure array for efficient overall water splitting at stearic acid to diesel-like hydrocarbons by carbon nanotubes-supported CuCo
large current densities, ACS Appl. Mater. Interfaces 10 (2018) 26283–26292. catalysts, Fuel Process. Technol. 188 (2019) 153–163.
[36] Z. Zhang, Q. Yang, H. Chen, K. Chen, X. Lu, P. Ouyang, J. Fu, J.G. Chen, In situ [44] S. Xing, Y. Liu, X. Liu, M. Li, J. Fu, P. Liu, P. Lv, Z. Wang, Solvent-free
hydrogenation and decarboxylation of oleic acid into heptadecane over a Cu–Ni hydrodeoxygenation of bio-lipids into renewable alkanes over NiW bimetallic
alloy catalyst using methanol as a hydrogen carrier, Green Chem. 20 (2018) catalyst under mild conditions, Appl. Catal. B Environ. 269 (2020).
197–205. [45] C. Sarkar, S.C. Shit, D.Q. Dao, J. Lee, N.H. Tran, R. Singuru, K. An, D.N. Nguyen, Q.
[37] H. Chen, Y. Wu, S. Qi, Y. Chen, M. Yang, Deoxygenation of octanoic acid catalyzed V. Le, P.N. Amaniampong, A. Drif, F. Jerome, P.T. Huyen, T.T.N. Phan, D.V.N. Vo,
by hollow spherical Ni/ZrO2, Appl. Catal. Gen. 529 (2017) 79–90. N. Thanh Binh, Q.T. Trinh, M.P. Sherburne, J. Mondal, An efficient hydrogenation
catalytic model hosted in a stable hyper-crosslinked porous-organic-polymer: from
fatty acid to bio-based alkane diesel synthesis, Green Chem. 22 (2020) 2049–2068.

49

You might also like