Roles of supports on reducibility and activities of Cu3P catalysts for deoxygenation of oleic acid: In situ XRD and XAS studies

Article abstract of DOI:10.1016/j.mcat.2021.111425

This work demonstrates for the first time that SiO₂ and ultra-stable zeolite Y (USY) supports play significant roles in the reducibility of Cu₂P₂O₇ to form Cu₃P, which consequently affects the selectivity of oleic acid deoxygenation. The formation of supported Cu₃P nanoparticles during hydrogen reduction of Cu₂P₂O₇ was carefully investigated by in situ X-ray diffraction (in situ XRD), and in situ X-ray absorption spectroscopy (in situ XAS). The results indicate that the transformation of Cu₂P₂O₇ to Cu₃P occurs through several steps. In the first step, all supported Cu₂P₂O₇ precursors are reduced to metallic Cu. Then, copper particles on SiO₂ support react with phosphorus compounds and directly transform to Cu₃P. On the other hand, copper particles on USY support partially transform to CuP₂ and Cu(OH)₂ before all converting to Cu₃P. Despite multi-step transformations, Cu₂P₂O₇/USY exhibits the lowest onset reduction temperature and provides Cu₃P with a small particle size. The deoxygenation of oleic acid over Cu₃P supported catalysts reaches nearly 100 % conversion. Both catalysts favor cyclization and aromatization to form cyclic and aromatic compounds. Cu₃P/SiO₂ achieves higher dodecylbenzene yield (46 %) than Cu₃P/USY (33 %). A proposed mechanism consists of hydrogenation of oleic acid and deoxygenation, then followed by cracking, cyclization, aromatization, and alkyl rearrangement.

Full text of DOI:10.1016/j.mcat.2021.111425

Molecular Catalysis xxx (xxxx) xxx
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.journals.elsevier.com/molecular-catalysis
Roles of supports on reducibility and activities of Cu3P catalysts for
deoxygenation of oleic acid: In situ XRD and XAS studies
Nopparuj Kochaputia, Pongtanawat Khemthongb c,*, Panita Kasamechonchungb,
,
,
,c
Teera Butbureeb c, Wanwisa Limphiratd, Yingyot Poo-arpornd, Sanchai Kuboonb
,
,
,
Kajornsak Faungnawakijb c, Chanapa Kongmarka c,**
a Department of Materials Science, Faculty of Science, Kasetsart University, Bangkok, 10900, Thailand
b National Nanotechnology Center (NANOTEC), National Science and Technology Development Agency (NSTDA), Pathum Thani, 12120, Thailand
c Research Network of NANOTEC – KU on NanoCatalysts and NanoMaterials for Sustainable Energy and Environment, Thailand
d Synchrotron Light Research Institute, 111 University Avenue, Muang, Nakhon Ratchasima, 30000, Thailand
A R T I C L E I N F O
A B S T R A C T
Keywords:
Deoxygenation
Oleic acid
Cu3P
This work demonstrates for the first time that SiO2 and ultra-stable zeolite Y (USY) supports play significant roles
in the reducibility of Cu2P2O7 to form Cu3P, which consequently affects the selectivity of oleic acid deoxygen-
ation. The formation of supported Cu3P nanoparticles during hydrogen reduction of Cu2P2O7 was carefully
investigated by in situ X-ray diffraction (in situ XRD), and in situ X-ray absorption spectroscopy (in situ XAS). The
results indicate that the transformation of Cu2P2O7 to Cu3P occurs through several steps. In the first step, all
supported Cu2P2O7 precursors are reduced to metallic Cu. Then, copper particles on SiO2 support react with
phosphorus compounds and directly transform to Cu3P. On the other hand, copper particles on USY support
partially transform to CuP2 and Cu(OH)2 before all converting to Cu3P. Despite multi-step transformations,
Cu2P2O7/USY exhibits the lowest onset reduction temperature and provides Cu3P with a small particle size. The
deoxygenation of oleic acid over Cu3P supported catalysts reaches nearly 100 % conversion. Both catalysts favor
cyclization and aromatization to form cyclic and aromatic compounds. Cu3P/SiO2 achieves higher dode-
cylbenzene yield (46 %) than Cu3P/USY (33 %). A proposed mechanism consists of hydrogenation of oleic acid
and deoxygenation, then followed by cracking, cyclization, aromatization, and alkyl rearrangement.
In situ XRD
In situ XAS
´
1. Introduction
electrochemical applications. Besides, Sotak et al. reported the potential
use of carbon supported Cu3P catalysts for hydrogenolysis of
Deoxygenation (DO) of oxygenated compounds over transition metal
phosphides have received considerable attention in recent years owing
to their excellent catalytic activity and stability under hydrotreatment
conditions [1,2]. Various supported metal phosphides including WP,
MoP, Ni2P, Co2P and Cu3P have been used for removing oxygen atoms
from oxygenated hydrocarbons [3–7]. Among these metal phosphides,
Cu3P seems to be less explored. Recently, it was reported that Cu3P
shows metal-like catalytic behaviors with high catalytic activity and
excellent hydrogen adsorption-desorption ability [7,8]. Cu3P has also
been used as an electrocatalyst for the hydrogen evolution reaction
(HER) under both acidic and basic conditions [9,10]. Its high electrical
conductivity and chemical stability are also beneficial for
biomass-derived alcohols such as glucose, sorbitol, and xylitol [11].
Indeed, supported Cu3P catalysts have been shown to have a good cat-
alytic activity for fatty acid deoxygenation [7].
Great efforts have been dedicated for preparation of Cu3P with
desired properties using various synthetic approaches such as colloidal
methods [12,13], solvothermal synthesis [14,15], solid-state reaction
[14,16], and reduction methods in H2 atmosphere [11]. Well-dispersed
metal phosphide nanoparticles on supports are desirable for catalytic
applications. In general, hydrogen reduction of metal precursors (e.g.,
metal phosphates/phosphites/hypophosphites or a mixture of metal
salt/oxide and phosphorus compounds) deposited on the pore walls of
supports by aqueous impregnation would allow the formation of metal
* Corresponding author at: National Nanotechnology Center (NANOTEC), National Science and Technology Development Agency (NSTDA), Pathum Thani, 12120,
Thailand.
** Corresponding author at: Department of Materials Science, Faculty of Science, Kasetsart University, Bangkok, 10900, Thailand.
E-mail addresses: pongtanawat@nanotec.or.th (P. Khemthong), chanapa.k@ku.th (C. Kongmark).
https://doi.org/10.1016/j.mcat.2021.111425
Received 30 September 2020; Received in revised form 2 January 2021; Accepted 18 January 2021
2468-8231/© 2021 Elsevier B.V. All rights reserved.
Please cite this article as: Nopparuj Kochaputi, Molecular Catalysis, https://doi.org/10.1016/j.mcat.2021.111425

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
phosphide particles dispersed on a support material [1,4]. Typically, this
precursors. Finally, those calcined samples were further reduced under
hydrogen atmosphere at 650 ◦C for 5 h to obtain Cu3P/SiO2 and Cu3P/
USY catalysts.
◦
method requires high temperatures (~ 500 – 1000 C) and hydrogen
atmosphere for the reduction of the metal precursors. Although
hydrogen reduction is an important process for the production of highly
active metal phosphides, only a few studies have been carried out for
FeP, Ni2P, Ni3P, Ni12P5, CoP, MoP and WP [1,4,17]. Their results suggest
that the reduction of phosphate or phosphite compounds mostly occurs
through the formation of metal particles which later react with phos-
phorus species and transform to metal phosphide particles. However,
metal phosphide particles can also be formed by other routes. For
example, proved by in situ X-ray diffraction (XRD) and X-ray absorption
spectroscopy (XAS) techniques, Berhault et al. reported that ammonium
nickel phosphate NiNH4PO4⋅H2O can be reduced to nickel phosphide via
an amorphous intermediate phase [18]. Time-resolved XRD results re-
ported by Rodriguez et al. illustrated the formation of Ni2P/SiO2 catalyst
through Ni and Ni12P5 intermediates which were not detected in the
unsupported catalyst [19]. It is well known that the catalyst support
plays an important role on the reduction of metal precursors, the
structure of active phases and the catalytic activity [20]. To the best of
our knowledge, there are no reports on the formation mechanism of
supported Cu3P catalysts using hydrogen reduction method. Details of
these reduction pathways remain unclear. Therefore, a comprehensive
study on the formation mechanism of Cu3P nanoparticles on various
supports during hydrogen reduction process is essential for the optimi-
zation of this material for catalytic applications.
2.2. Characterization of the supported Cu3P catalysts
The structural evolution during the hydrogen reduction of Cu2P2O7
to Cu3P was studied by ex situ and in situ XRD (D8 ADVANCE, Bruker,
Ltd.) using Cu Kα radiation with Ni filter, operated at 40 kV and 40 mA,
in the 2θ range of 10 – 80◦. The supported Cu2P2O7 precursor was put on
a Pt-Rh alloy plate in a heating cell chamber which was later heated
from 50 ◦C to 650 ◦C with a heating rate of 0.1 ◦C sꢀ 1 under hydrogen
flow (20 mL minꢀ 1) and hold at 650 ◦C for 45 min. The diffraction
◦
pattern was acquired at a regular interval (every 50 C) by the same
powder diffractometer that was used at room temperature measure-
ments. LaB6 was used as a standard to determine the instrumental res-
olution of the X-ray diffractometer.
The morphology and dispersion of Cu3P catalyst on different sup-
ports were determined by scanning electron microscopy (HITACHI
SU5000 FE-SEM) in backscattered electron (BSE) mode with an accel-
eration voltage of 10 kV.
Temperature-programmed reduction profiles were performed on a
TPR automated chemisorption analyzer (ChemBET Pulsar, Quantach-
rome). Each 100 mg of calcined samples was pretreated under helium
atmosphere at 120 ◦C for 1 h. Subsequently, the samples were then
As demonstrated in our previous report [7], the catalyst supports
(USY, Al2O3, and SiO2) significantly influence the activity and product
selectivity of the Cu3P catalyst. Therefore, in the present work, we aim to
further clarify the roles of the supports on the catalytic behavior of the
Cu3P. Based on the assumption that a different support could induce the
formation of different copper species during hydrogen reduction pro-
cess, we carefully investigate those phase transformation processes.
Thus, in this study, we prepared copper pyrophosphate precursor
(Cu2P2O7) supported on SiO2 and ultra-stable zeolite Y (USY) by
impregnation method using metal salts and ammonium phosphate. The
formation of copper phosphide (Cu3P) by hydrogen reduction of
Cu2P2O7 was monitored by in situ X-ray diffraction (XRD) and X-ray
absorption spectroscopy (XAS) techniques. The combination of in situ
XRD and in situ XAS provided information of both short-range and long-
range structural modifications, which allow us to understand crystalli-
zation processes and effects of the catalyst supports on the formation
mechanism. This is regarded as one of the most powerful methods to
determine the nature of intermediate phases formed during the synthesis
and to optimize the conditions for the production of supported Cu3P
nanoparticles. The catalytic performance was demonstrated for oleic
deoxygenation (DO) reaction. We believe that this insightful investiga-
tion would reveal the evolving intrinsic activities and materials prop-
erties as a function of catalyst supports. This will be essential for
understanding reduction behaviors and catalytic mechanisms of the
supported catalysts.
reduced under H2/Ar mixture (H2/Ar = 1.5; total flow 30 cm3 minꢀ 1
)
and heated from 100 to 800 ◦C with a heating rate of 5 ◦C minꢀ 1
.
The evolution of the valence state, geometry, and Cu species during
hydrogen reduction was monitored by in situ XAS. Cu K-edge XAS ex-
periments were performed in transmission mode, at the Time-resolved
XAS beamline (BL2.2) of the Synchrotron Light Research Institute
(SLRI) in Thailand. A bent crystal Si (111) in the energy dispersive
monochromator was used to focus a polychromatic X-ray beam onto the
sample. The X-rays pass through the sample and diverge towards a
position-sensitive detector (NMOS-linear image sensor). The experi-
ments were carried out in a specially designed high temperature cell
[21]. Each supported Cu2P2O7 precursor was pressed into a pellet and
loaded into the cell. Then it was reduced in 75 %H2 in N2 (total flow rate
of 20 mL minꢀ 1) at 650 ◦C for 30 min with a heating rate of 2 ◦C minꢀ 1
.
The obtained XAS data were processed using the Athena graphical
interface of the IFEFFIT program suite [22]. A combination of principal
component analysis (PCA) and linear combination fit (LCF) analysis
were applied on the normalized XANES spectra in the ꢀ 20 eV < E0 < 90
eV range.
2.3. Catalytic deoxygenation of oleic acid
Oleic acid was used as a model compound for deoxygenation (DO)
reaction. The reaction was carried out in a Parr batch reactor. Approx-
imately 1 g of supported metal phosphide catalyst and 60 mL of a 5 wt%
solution of oleic acid in dodecane were loaded into the reactor. Prior to
the reaction, the reactor was purged with N2, then heated to 340 ◦C and
held at this temperature for 6 h. The liquid products were analyzed by a
gas chromatograph equipped with a mass spectrometer (GC–MS). The
DB-1HT was used as a capillary column. In addition, the gas-phase
products were not further analyzed because methanation and water
gas shift reaction were involved as the main gas phase reaction. The
conversion of oleic acid was calculated according to the following
equations.
2. Experimental section
2.1. Catalyst preparation
Synthesis of the supported Cu3P catalysts was based on hydrogen
reduction of supported copper pyrophosphate (Cu2P2O7) precursors.
First, the supported Cu2P2O7 precursors (10 wt.% Cu) were prepared by
incipient wetness impregnation method using an aqueous solution
containing Cu(NO3)2 and (NH4)2HPO4 with the Cu/P molar ratio of 2:1
under constant stirring. A few drops of nitric acid were added to dissolve
some precipitates until a clear light blue solution was obtained. Then,
the prepared solution was impregnated on SiO2 or ultra-stable zeolite Y
(USY) supports. These impregnated samples were dried at 80 ◦C for 12 h
and calcined under air atmosphere at 450 ◦C for 3 h, resulting in light
blue powders which were referred as Cu2P2O7/SiO2 and Cu2P2O7/USY
initial m ole of oleic acid ꢀ final m ole of oleic acid
conversion (%) =
initial m ole of oleic acid
× 100
(1)
The selective yield (Y) of products were calculated based on carbon
mass balance [23]
2

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
Fig. 3. H2-TPR profiles of Cu2P2O7 precursor on SiO2 and USY supports.
space group (PDF no. 00-044-0182) at 28.3◦, 30.2◦ and 30.5◦. The low
intensity of Cu2P2O7 lines is observed due to the low metal concentra-
tion. After the reduction in hydrogen atmosphere at 650 ◦C for 5 h, all
catalysts show the characteristic diffraction peaks of Cu3P hexagonal
structure with P63cm space group (PDF no. 01-071-2261) at 36.0◦,
39.1◦, 41.6◦, 45.1◦, 46.2◦ and 47.3◦, as illustrated in Fig. 1b. There is no
characteristic peaks of impurities observed in Cu3P/USY catalysts. In the
XRD pattern of Cu3P/SiO2 catalyst, a small additional peak is noticed at
24.1◦ which can be attributed to SiP2O7 phase (PDF no. 01-070-2245).
Tiberius et al. also found the formation of SiP2O7 during the thermal
dehydration of Ca(H2PO4)2⋅H2O–SiO2 system [24]. It was suggested that
H3PO4 occurred during the reaction reacted with the SiO2 and yielded
SiP2O7. The crystallite sizes of Cu3P catalysts, estimated from the XRD
peak width by using the Debye Scherrer equation, are 80.4 and 42.1 nm
for SiO2 and USY supports, respectively. The Cu3P catalyst particles on
USY support are about half the size of those on SiO2 support.
Fig. 1. XRD patterns of (a) calcined catalysts and (b) reduced catalysts
with supports.
The morphologies of supported Cu3P catalysts are illustrated in
Fig. 2. Backscattered electron images show specimen contrast caused by
variation of atomic weights. The brighter areas are Cu3P particles
deposited on each support. Note that Cu3P particles are well dispersed
on the USY support (Fig. 2b) with a homogeneous particle size of less
than 100 nm, while those on SiO2 support (Fig. 2a) form agglomerates
with diameters ranging from 50 to 500 nm.
(
)
ni × ai
Yi (m ol %) =
× 100
(2)
nOleic acid × aOleic acid
Where ni and ai represent the mole and carbon atom number of product i.
nOleic acid and aOleic acid represent the mole and carbon atom number of
oleic acid. The qualitative and quantitative analyses were conducted on
HDO and DCO/DCO2 products. The unspecified liquid products could be
referred to as cyclic or aromatic compounds and polymerized products.
3.2. Reduction behaviors
3. Results and discussion
Hydrogen temperature-programmed reduction (H2-TPR) experi-
ments were performed to investigate the reduction behavior of Cu2P2O7
supported on SiO2 and USY. Fig. 3 shows the evolution of hydrogen
consumption during the reduction process. To the best of our knowl-
edge, our present work is the first report on TPR studies for copper
phosphide catalysts which are synthesized by reduction of phosphate
3.1. Structures and morphologies
The crystal structure of the supports and the calcined catalysts were
confirmed by means of ex situ XRD and presented in Fig. 1a. All samples
exhibit characteristic lines of α-Cu2P2O7 monoclinic structure with C2/c
Fig. 2. Back scattered SEM images of Cu3P catalyst on (a) SiO2 and (b) USY supports.
3

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
Fig. 5. Reduction mechanisms of Cu2P2O7 precursors on SiO2 and USY supports
to Cu3P catalysts deduced from in situ XRD data.
Cu2P2O7 and corresponding supports are observed at room temperature.
◦
When Cu2P2O7/SiO2 is heated up to 300 C, the intensity of Cu2P2O7
lines slightly increases due to a higher degree of crystallinity. At 350 ◦C,
the characteristic lines of Cu are noticed (Fm-3m, PDF no. 01-089-2838).
◦
Cu lines become stronger and reach the maximum intensity at 400 C
while the Cu2P2O7 precursorvanishes, indicating the complete reduction
of Cu2P2O7 to metallic Cu. At 500 ◦C, the diffraction peaks attributed to
Cu3P (P-3m1 with PDF no. 03-065-1973) emerge while those of Cu
gradually disappear, suggesting the transformation from metallic Cu to
Cu3P structure. The pure crystalline phase of Cu3P (P-3m1) is success-
fully obtained at 650 ◦C. This high temperature trigonal form of Cu3P is
later stabilized in hexagonal polymorph (P63cm) at room temperature
[26,27], which is in line with the ex situ XRD studies. In addition, a small
characteristic peak of SiP2O7 is observed (2θ ~ 23.9◦) in the diffraction
patterns of silica-supported catalyst at temperatures above 300 ◦C. The
formation of SiP2O7 can occur on the surface of silica-supported metal
phosphide catalyst with an excess phosphorus or metal phosphide con-
tent [24]. This is in accordance with the ex situ XRD results. The
diffraction peaks of Pt-Rh alloy contributed from the sample holder can
also be observed at high temperature due to the sintering shrinkage of
the sample.
Fig. 4. XRD patterns collected during hydrogen reduction of Cu2P2O7 precursor
on (a) SiO2 and (b) USY supports.
precursors on different supports. A few H2-TPR studies were conducted
on other metal phosphide catalysts (e.g., FeP, Ni2P, CoP, MoP and WP)
[1,17,25]. Although each catalyst shows a specific TPR profile
(composed of a single peak or several overlapping peaks), it was sug-
gested that most of metal phosphides follow a similar reduction process.
The partly or completely reduced transition metal promotes the reduc-
tion of phosphite or phosphate components to volatile phosphorus
compounds such as phosphine (PH3), H3PO4, H3PO3 or H3PO2 that react
with reduced metals to form transition metal phosphides. As shown in
Fig. 3, the TPR profiles of Cu3P/SiO2 catalyst consist of a strong
hydrogen uptake peak at 530 ◦C and a weak broad peak at around
690ꢀ 760 ◦C, while the Cu3P/USY catalyst shows three unresolved peaks
at 360, 403 and 445 ◦C and a broad peak at 690 ◦C. The hydrogen
reduction process of the catalyst supported on USY seems to be more
complex and takes place at lower temperature than that on SiO2 support.
The hydrogen reduction mechanism of Cu2P2O7 to Cu3P on catalyst
supports is still unclear. Therefore, the structural evolution behind these
reduction processes were further studied in detail by in situ XRD and in
situ XAS.
The structural evolution during the reduction of Cu2P2O7 on USY
3.3. Formation mechanisms of supported Cu3P catalysts
In situ high-temperature XRD patterns collected during hydrogen
reduction of Cu2P2O7 precursor on SiO2 and USY supports from room
temperature to 650 ◦C are illustrated in Fig. 4a and b, respectively. The
Cu2P2O7/SiO2 and Cu2P2O7/USY precursors exhibit different reduction
behaviors. At the beginning of the reaction, a crystalline phase of
Fig. 6. XANES spectra of Cu references, Cu2P2O7 precursors and Cu3P catalysts
on SiO2 and USY supports.
4

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
Fig. 8. Composition of the Cu species at different reaction times calculated by
linear combination fitting (LCF) of the XANES spectra recorded during
hydrogen reduction of Cu2P2O7/SiO2.
feature at 9001 eV and a similar oscillatory structure. No evidence of the
characteristic features from either Cu2O or CuO is observed in XANES
spectra of the precursors and the catalysts, which would affirm the
absence of copper oxide traces in all of our samples. Regarding to the
supported Cu3P catalysts, they exhibit identical edge energy and oscil-
latory structure. Their absorption edge energy significantly shifts to
lower energy (~ 8981 eV) due to a decrease in the oxidation states of Cu
from Cu+2 to Cu+1 after hydrogen reduction. Comparing XANES spectra
of Cu3P and Cu2O standard at the absorption edge region, Cu2O with
linear 2-coordinate Cu(I) complex shows a shoulder feature at 8983 eV
corresponding to 1s → 4pxy electronic transition which is not observed in
the 4-coordinate tetrahedral Cu(I) complex of Cu3P [29–31]. Therefore,
these XAS results would confirm Cu2P2O7 structure for the calcination
step and Cu3P structure for the reduction step, as previously shown in
the XRD results.
Fig. 7. Stacked plots of XANES spectra of (a) Cu2P2O7/SiO2 and (b) Cu2P2O7/
USY precursors during hydrogen reduction.
(Fig. 4b) appears differently from that on SiO2 support. By increasing the
temperature to 300 ◦C, the characteristic peaks of an intermediate
metallic Cu phase emerge and grow at the expense of Cu2P2O7 which
gradually disappears at 500 ◦C. This result is in good agreement with the
H2-TPR profiles which show that Cu2P2O7 on USY support is reduced to
metallic Cu at lower temperatures than that on SiO2 support. Afterward,
the metallic Cu possibly reacts with hydrogen, water and the remaining
phosphorus compounds, such as gaseous PH3 or H3PO4 [1,4,17,24,25]
that cannot be detected by XRD, to form intermediate phases including
CuP2 (P21/c, PDF no. 00-055-0400) and Cu(OH)2 (Cmcm, PDF no.
Stacked plots of time resolved Cu K-edge XANES spectra of Cu2P2O7
on SiO2 and USY supports during hydrogen reduction from room tem-
◦
perature to 650 C are given in Fig. 7a and 7b, respectively. The two
samples exhibit dissimilar structural evolution. For Cu2P2O7/SiO2,
principal component analysis indicates three different species involved
in this reaction, including Cu2P2O7, Cu0 and Cu3P. Linear combination
fit (LCF) analysis of XANES spectra provides the evolution of the relative
abundances of copper species during hydrogen reduction of Cu2P2O7/
SiO2 (shown in Fig. 8). Cu2P2O7 is reduced to metallic Cu at tempera-
tures above 310 ◦C. This is shown in a gradual decrease of the absorption
edge from ~ 8987 to ~ 8979 eV, together with a decrease in white line
intensity. The maximum amount of Cu0 intermediate phase is observed
at 360 ◦C. Further heating causes a shift in absorption edge energy from
~ 8979 to ~ 8981 eV, a decrease in white line intensity, and an alter-
ation of oscillatory feature above the edge to become Cu3P. It is found
that Cu3P forms at the expense of Cu0 which diminishes at 620 ◦C.
In contrast, XANES spectra of Cu2P2O7/USY evolve differently. As
illustrated in Fig. 7b, several trends can be seen. The structural trans-
formation from Cu2P2O7 to Cu3P seems to proceed through several re-
action steps, in accordance with in situ XRD results. Although
quantitative analysis could not be performed due to a lack of reference
spectra of metastable intermediate phases, a close inspection of XANES
spectra reveals the first phase transformation at temperatures above 220
◦C. The white line intensity gradually decreases, a characteristic shoul-
der of Cu0 emerges at ~ 8981 eV, and the edge progressively shifts to ~
8979 eV (at 450 ◦C). This modification of spectral features indicates the
reduction of Cu2P2O7 (Cu2+) to Cu0 metal which takes place at a tem-
perature lower than that of Cu2P2O7/SiO2 precursor. This result is in
good agreement with H2-TPR and in situ XRD analyses. When the tem-
perature increases from 450 to 650 ◦C, the spectral features evolve to-
ward the Cu3P structure through several intermediate phases, as
different trends of oscillation alteration are observed. In general, the
◦
00-035-0505) at 550 C. These intermediate phases later transform to
Cu3P (P-3m1) upon heating to 650 ◦C. The Cu3P on USY support is later
stabilized in the most stable polymorphic form, i.e., a hexagonal struc-
ture with P63cm space group at room temperature, as previously evi-
denced by ex situ XRD studies [26,27]. The reduction mechanisms of
Cu2P2O7 to Cu3P catalyst on SiO2 and USY supports deduced from XRD
results are summarized in Fig. 5.
Furthermore, the evolution of the valence state and local atomic
structure of copper in the catalysts during hydrogen reduction was
studied by in situ XAS. Fig. 6 presents Cu K-edge XANES (X-ray ab-
sorption near edge structure) spectra of supported Cu2P2O7 precursors
and Cu3P catalysts at room temperature together with the spectra of
reference materials, i.e., Cu foil, Cu2O and CuO. XANES spectrum of each
reference material exhibits specific features, i.e., edge energy, pre-peak,
white line peak and oscillation shape, because they are sensitive to the
oxidation state and local structure around the copper atoms. Cu0, Cu+
and Cu2+ exhibit absorption edges at 8979, 8981 and 8984 eV, respec-
tively [28]. The absorption edge of supported Cu2P2O7 samples correlate
well with that of CuO, suggesting the presence of Cu(II) ions. Note that
the dissimilarity in shoulder feature at 8986 eV (attributed to 1s → 4pxy
electronic transition) between CuO and Cu2P2O7 precursor is related to
the difference in copper ion coordination and covalency of the equato-
rial ligands [29]. All supported Cu2P2O7 precursors exhibit a white line
5

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
catalyst. From the in situ XAS results that were collected every 10 ◦C, the
onset reduction temperature of Cu2P2O7/USY (~ 230 ◦C) is much lower
than that of Cu2P2O7/SiO2 (~ 320 ◦C). This can be explained by the fact
that Cu2P2O7 might exhibit a stronger interaction with SiO2 support, as
small amounts of SiP2O7 is also formed during the reaction. Moreover,
the ultra-stable zeolite Y (USY) support is known to have higher acidity
than SiO2 support, which may result in a high efficiency of spilled over
hydrogen diffusion enhancing the reduction of Cu2P2O7 [32]. The
Cu2P2O7/USY is therefore more easily reduced than the other support,
possibly enabling the formation of Cu3P with smaller particle size.
Table 1
The %Conversion of oleic acid and the %Yield of products measured at 340 ◦C
for 6 h.
Percentage
Cu3P/USY
Cu3P/SiO2
Conversion
99.99
99.99
Yield
- Heptadecane (C17H36
)
–
–
- Octadecane (C18H38
)
–
–
- Dodecylbenzene (C18H30
)
32.84
8.31
5.01
53
46.13
4.55
–
- Dodecylcyclohexane (C18H36
)
- Heptylcyclopentane (C12H24
- Others
)
49
3.4. Catalytic deoxygenation of oleic acid
structural evolutions of all supported catalysts deduced from in situ XAS
spectra are consistent with in situ XRD results. A slight discrepancy in
phase transformation temperatures observed by both techniques is
possibly related to the different increments of temperature. As XRD
measurement required relatively long data collection times, diffraction
Conversion of oleic acid and deoxygenation (DO) yields of supported
Cu3P catalysts are listed in Table 1. At the collecting time of 6 h, each
catalyst reaches a complete conversion. Both Cu3P/SiO2 and Cu3P/USY
catalysts exhibit similar catalytic behavior. Oleic acid is transformed to
cyclic and aromatic compounds through cyclization and aromatization.
The main product is dodecylbenzene while dodecylcylclohexane is a
minor product. Cu3P/SiO2 achieves higher dodecylbenzene yield (46 %)
than Cu3P/USY (33 %). The lower selectivity of Cu3P/USY is possibly
due to the higher acidity of USY support [23,33,34] that may favor
cracking reaction, giving heptylcyclopentane as a by-product. In this
case, support appears to exert influence over the product selectivity but
not the reaction mechanism of the Cu3P catalyst.
◦
patterns were recorded every 50 C, while time-resolved XAS spectra
were collected every 10 ◦C. An alternative explanation for this
discrepancy may be the ability of XAS to obtain local structural infor-
mation on both the crystalline and amorphous nanoparticles. It seems
likely that XAS can probe amorphous particles or very small nano-
particles which may occur before the formation of nanocrystalline
phases detected by XRD.
In summary, in situ XRD and in situ XAS give complementary struc-
tural information. In situ XRD provides insights into the crystal structure
transformation process while in situ XAS provides the information on the
evolution of the valence states and the local atomic structures of Cu
during hydrogen reduction of Cu2P2O7 to Cu3P nanoparticles on SiO2
and USY supports. Both techniques revealed that the formation mech-
anism of Cu3P consists of multiple reaction steps. The Cu2P2O7 precursor
is first reduced to metallic Cu which can facilitate the dissociation of
hydrogen molecules to hydrogen atoms. The hydrogen spillover may
occur and assist the reduction of phosphate to volatile phosphorus
compounds, i.e., PH3 and H3PO4 [1,4,17,25] which later react with
metallic Cu to form Cu3P on SiO2 support. In contrast, the reduced Cu
particles on USY support partially transform to CuP2 and Cu(OH)2 before
all converting to Cu3P. It is also found that the nature of catalyst sup-
ports exerts a significant influence on the reduction behavior of Cu2P2O7
It is worth noting that Cu3P catalyst provides different reaction
mechanism compared to other metal phosphide catalysts, such as CoP,
Ni2P, WP, MoP, FeP, which have been reported to favor deoxygenation
of oxygenated compounds to alkanes [1–7]. Additionally, different
supports may possess different structure, different Lewis acid sites
¨
(metal sites) and Bronsted sites (P–OH sites) of metal phosphides that
may affect their intrinsic activities [2,33,34]. Therefore, the supported
Cu3P catalysts can have potential applications in the valorization of
vegetable oil or biofuel for the production of cyclic and aromatic
compounds.
According to the analysis of catalytic results in Table 1, a plausible
mechanism for the deoxygenation of oleic acid over supported Cu3P
catalysts could be summarized in Scheme 1. This proposed mechanism is
based on the surface reaction. It can be seen that supported Cu3P
Scheme 1. Proposed mechanism for the deoxygenation of oleic acid over supported Cu3P.
6

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
phosphides and their activity in the hydrodeoxygeblenation of a biofuel model
compound, J. Catal. 294 (2012) 184–198, https://doi.org/10.1016/j.
jcat.2012.07.021.
catalysts favor the hydrogenation of oleic acid to form stearic acid. Then,
there are two further main reaction pathways. The first one starts with
cracking of stearic acid, cyclization, aromatization, and alkyl rear-
rangement. This pathway gives dodecylbenzene as a final product. The
second one starts with the removal of H2O, CO and CO2 from stearic
acid, followed by cracking, cyclization and alkyl rearrangement. This
pathway could provide dodecyl cyclohexane as the final product.
´
´
[2] M. Peroni, I. Lee, X. Huang, E. Barath, O.Y. Gutierrez, J.A. Lercher, Deoxygenation
of palmitic acid on unsupported transition-metal phosphides, ACS Catal. 7 (2017)
6331–6341, https://doi.org/10.1021/acscatal.7b01294.
[3] H.Y. Zhao, D. Li, P. Bui, S.T. Oyama, Hydrodeoxygenation of guaiacol as model
compound for pyrolysis oil on transition metal phosphide hydroprocessing
catalysts, Appl. Catal. A Gen. 391 (2011) 305–310, https://doi.org/10.1016/j.
apcata.2010.07.039.
[4] R. Prins, M.E. Bussell, Metal phosphides: preparation, characterization and
catalytic reactivity, Catal. Lett. 142 (2012) 1413–1436, https://doi.org/10.1007/
s10562-012-0929-7.
4. Conclusion
´
´
´
[5] J.A. Cecilia, A. Infantes-Molina, E. Rodríguez-Castellon, A. Jimenez-Lopez, S.
T. Oyama, Oxygen-removal of dibenzofuran as a model compound in biomass
derived bio-oil on nickel phosphide catalysts: role of phosphorus, Appl. Catal. B
136 (2013) 140–149, https://doi.org/10.1016/j.apcatb.2013.01.047.
[6] J.-S. Moon, E.-G. Kim, Y.-K. Lee, Active sites of Ni2P/SiO2 catalyst for
hydrodeoxygenation of guaiacol: a joint XAFS and DFT study, J. Catal. 311 (2014)
144–152, https://doi.org/10.1016/j.jcat.2013.11.023.
Cu3P catalysts supported on SiO2 and ultra-stable zeolite Y (USY)
were successfully prepared by hydrogen reduction of copper pyro-
phosphate precursor (Cu2P2O7). The phase transition was monitored by
H2-TPR, in situ XRD and XAS. The combination of these techniques was
proven to be a powerful tool to understand the formation mechanism of
supported metal phosphide nanoparticles and the influence of the sup-
port materials on the reaction mechanism. Several phase trans-
formations were observed. The Cu2P2O7 precursor was reduced to
metallic Cu, which later reacted with volatile phosphorus compounds
and directly transformed to Cu3P particles on SiO2 support, whereas the
reduced Cu on USY support evolved through CuP2 and Cu(OH)2 in-
termediates before transforming to Cu3P. The reduction of Cu2P2O7 to
Cu3P on USY seemed to be more facile compared to that on SiO2,
although it proceeded through multiple steps. It exhibited the lowest
onset reduction temperature and provided Cu3P having a smaller par-
ticle size. The Cu3P supported catalysts were demonstrated to achieve a
high conversion of almost 100 % for deoxygenation of oleic acid. Both
Cu3P/SiO2 and Cu3P /USY catalysts favored cyclization and aromati-
zation to form cyclic and aromatic compounds. Cu3P/SiO2 achieved
higher dodecylbenzene yield (46 %) than Cu3P/USY (33 %). The catalyst
support appeared to exert influence over the product selectivity but not
the reaction mechanism of the Cu3P catalyst. Cu3P could be considered
as a promising catalyst for the valorization of vegetable oil or biofuel for
the production of cyclic and aromatic compounds. The information
obtained from this study could be used to optimize conditions for the
production of supported Cu3P nanopartcles for catalytic applications.
[7] N. Kochaputi, C. Kongmark, P. Khemthong, T. Butburee, S. Kuboon,
A. Worayingyong, K. Faungnawakij, Catalytic behaviors of supported Cu, Ni, and
Co phosphide catalysts for deoxygenation of oleic acid, Catalysts 9 (2019) 715,
https://doi.org/10.3390/catal9090715.
[8] R. Wang, X.-Y. Dong, J. Du, J.-Y. Zhao, S.-Q. Zang, MOF-derived bifunctional Cu3P
nanoparticles coated by a N, P-codoped carbon shell for hydrogen evolution and
oxygen reduction, Adv. Mater. 30 (2018) 1703711, https://doi.org/10.1002/
adma.201703711.
[9] L. Ma, X. Shen, H. Zhou, J. Zhu, C. Xi, Z. Ji, L. Kong, Synthesis of Cu3P nanocubes
and their excellent electrocatalytic efficiency for the hydrogen evolution reaction
in acidic solution, RSC Adv. 6 (2016) 9672–9677, https://doi.org/10.1039/
C5RA24427G.
[10] C.-C. Hou, Q.-Q. Chen, C.-J. Wang, F. Liang, Z. Lin, W.-F. Fu, Y. Chen, Self-
supported cedarlike semimetallic Cu3P nanoarrays as a 3D high-performance Janus
electrode for both oxygen and hydrogen evolution under basic conditions, ACS
Appl. Mater. Interfaces 8 (2016) 23037–23048, https://doi.org/10.1021/
acsami.6b06251.
´
[11] T. Sotak, T. Schmidt, M. Hronec, Hydrogenolysis of polyalcohols in the presence of
metal phosphide catalysts, Appl. Catal. A Gen. 459 (2013) 26–33, https://doi.org/
10.1016/j.apcata.2013.04.006.
[12] L. De Trizio, A. Figuerola, L. Manna, A. Genovese, C. George, R. Brescia, Z. Saghi,
R. Simonutti, M. Van Huis, A. Falqui, Size-tunable, hexagonal plate-like Cu3P and
janus-like Cu-Cu3P nanocrystals, ACS Nano 6 (2012) 32–41, https://doi.org/
10.1021/nn203702r.
`
´
[13] S. Carenco, D. Portehault, C. Boissiere, N. Mezailles, C. Sanchez, Exploring
nanoscaled matter from speciation to phase diagrams: metal phosphide
nanoparticles as a case of study, Adv. Mater. 26 (2014) 371–390, https://doi.org/
10.1002/adma.201303198.
[14] M.-P. Bichat, T. Politova, H. Pfeiffer, F. Tancret, L. Monconduit, J.-L. Pascal,
T. Brousse, F. Favier, Cu3P as anode material for lithium ion battery: powder
morphology and electrochemical performances, J. Power Sources 136 (2004)
80–87, https://doi.org/10.1016/j.jpowsour.2004.05.024.
CRediT authorship contribution statement
Nopparuj Kochaputi: Writing - original draft. Pongtanawat
Khemthong: Conceptualization, Writing - review & editing. Teera
Butburee: Writing - review & editing. Sanchai Kuboon: Writing - re-
view & editing. Kajornsak Faungnawakij : Writing - review & editing.
Chanapa Kongmark: Conceptualization, Writing - review & editing.
[15] S. Liu, Y. Qian, L. Xu, Synthesis and characterization of hollow spherical copper
phosphide (Cu3P) nanopowders, Solid State Commun. 149 (2009) 438–440,
https://doi.org/10.1016/j.ssc.2008.12.046.
[16] H. Pfeiffer, F. Tancret, T. Brousse, Synthesis, characterization and electrochemical
properties of copper phosphide (Cu3P) thick films prepared by solid-state reaction
at low temperature: a probable anode for lithium ion batteries, Electrochim. Acta
50 (2005) 4763–4770, https://doi.org/10.1016/j.electacta.2005.02.024.
[17] C. Stinner, Z. Tang, M. Haouas, Th. Weber, R. Prins, Preparation and 31P NMR
characterization of nickel phosphides on silica, J. Catal. 208 (2002) 456–466,
https://doi.org/10.1006/jcat.2002.3577.
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgements
´
[18] G. Berhault, P. Afanasiev, H. Loboue, C. Geantet, T. Cseri, C. Pichon, C. Guillot-
Deudon, A. Lafond, In situ XRD, XAS, and magnetic susceptibility study of the
reduction of ammonium nickel phosphate NiNH4PO4⋅H2O into nickel phosphide,
Inorg. Chem. 48 (2009) 2985–2992, https://doi.org/10.1021/ic802074k.
[19] J.A. Rodriguez, J.-Y. Kim, J.C. Hanson, S.J. Sawhill, M.E. Bussell, Physical and
chemical properties of MoP, Ni2P, and MoNiP hydrodesulfurization catalysts: time-
resolved X-ray diffraction, density functional, and hydrodesulfurization activity
studies, J. Phys. Chem. B 107 (2003) 6276–6285, https://doi.org/10.1021/
jp022639q.
This work was co-supported by Kasetsart University Research and
Development Institute (KURDI) and National Nanotechnology Center
(NANOTEC) through its program of Research Network NANOTEC
(RNN). This work was also supported by National Science and Tech-
nology Development Agency (NSTDA) via CAS-NSTDA Project
(P1952712) i. We also would like to thank Prof. Gopinathan Sankar and
Attera Worayingyong for insightful advice and discussions. We are
grateful to Prof. S. Seraphin (Professional Authorship Center, NSTDA)
for fruitful discussions on manuscript editing. All authors also thank Mr.
Chakrit Yimsukanan for his support.
[20] H. Shi, J. Chen, Y. Yang, S. Tian, Catalytic deoxygenation of methyl laurate as a
model compound to hydrocarbons on nickel phosphide catalysts: remarkable
support effect, J. Catal. 208 (2002) 456–466, https://doi.org/10.1006/
jcat.2002.3577.
[21] Y. Poo-arporn, P. Chirawatkul, W. Saengsui, S. Chotiwan, S. Kityakarn,
S. Klinkhieo, J. Hormes, P. Songsiriritthigul, Time-resolved XAS (Bonn-SUT-SLRI)
beamline at SLRI, J. Synchrotron Rad. 19 (2012) 937–943, https://doi.org/
10.1107/S090904951204109X.
[22] B. Ravel, M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray
absorption spectroscopy using IFEFFIT, J. Synchrotron Rad. 12 (2005) 537–541,
https://doi.org/10.1107/S0909049505012719.
[23] H. Zuo, Q. Liu, T. Wang, L. Ma, Q. Zhang, Q. Zhang, Hydrodeoxygenation of methyl
palmitate over supported Ni catalysts for diesel-like fuel production, Energy Fuels
26 (2012) 3747–3755, https://doi.org/10.1021/ef300063b.
References
[1] P. Bui, J.A. Cecilia, S.T. Oyama, A. Takagaki, A. Infantes-Molina, H. Zhao, D. Li,
´
´
E. Rodríguez-Castellon, A.J. Lopez, Studies of the synthesis of transition metal
7

N. Kochaputi et al.
Molecular Catalysis xxx (xxxx) xxx
[24] T.C. Vaimakis, A.T. Sdoukos, The thermal dehydration of the Ca(H2PO4)2⋅H2O-
SiO2 system. Part 1. Mechanism, Thermochim. Acta 277 (1996) 107–120, https://
doi.org/10.1016/0040-6031(95)02751-3.
[30] S. Pithakratanayothin, R. Tongsri, T. Chaisuwan, S. Wongkasemjit, P. Khemthong,
S. Limpijumnong, P. Pharanchai, K. Malaicharoen, Discovery of mono(u-oxo)
dicopper and bis(u-oxo)dicopper in ordered Cu incorporated in SBA-15 via sol-gel
process from silatrane at room temperature: an in situ XAS investigation,
Microporous Mesoporous Mater. 301 (2020) 110225, https://doi.org/10.1016/j.
micromeso.2020.110225.
[25] X. Lan, E.J.M. Hensen, T. Weber, Silica-supported Ni2P: Effect of preparation
conditions on structure and catalytic performance in thiophene
hydrodesulfurization (HDS), Catal. Today 292 (2017) 121–132, https://doi.org/
10.1016/j.cattod.2016.12.040.
[31] P. Khemthong, P. Photai, N. Grisdanurak, Structural properties of CuO/TiO2
nanorod in relation to their catalytic activity for simultaneous hydrogen
production under solar light, Int. J. Hydrogen Energy 38 (2013) 15992–16001,
https://doi.org/10.1016/j.ijhydene.2013.10.065.
¨
[26] H. Schlenqer, H. Jacobs, R. Juza, Ternare phasen des lithiums mit kupfer und
phosphor, Z. Anorg. Allg. Chem. 385 (1971) 177–201, https://doi.org/10.1002/
zaac.19713850302.
[27] J.-H. Chen, K.H. Whitmire, A structural survey of the binary transition metal
phosphides and arsenides of the d-block elements, Coord. Chem. Rev. 355 (2018)
271–327, https://doi.org/10.1016/j.ccr.2017.08.029.
[32] P.S.F. Mendes, G. Lapisardi, C. Bouchy, M. Rivallan, J.M. Silva, M.F. Ribeiro,
Hydrogenating activity of Pt/zeolite catalysts focusing acid support and metal
dispersion influence, Appl. Catal. A Gen. 504 (2015) 17–28, https://doi.org/
10.1016/j.apcata.2015.03.027.
¨
[28] M. Staniuk, D. Zindel, W. van Beek, O. Hirsch, N. Kranzlin, M. Niederberger,
D. Koziej, Matching the organic and inorganic counterparts during nucleation and
growth of copper-based nanoparticles – in situ spectroscopic studies,
CrystEngComm 17 (2015) 6962–6971, https://doi.org/10.1039/c5ce00454c.
[29] L.-S. Kau, D.J. Spira-Solomon, J.E. Penner-Hahn, K.O. Hodgson, E.I. Solomon, X-
ray absorption edge determination of the oxidation state and coordination number
of copper: application to the type 3 site in Rhus vernicifera Lacease and its reaction
with oxygen, J. Am. Chem. Soc. 109 (1987) 6433–6442, https://doi.org/10.1021/
ja00255a032.
[33] M. Breysse, P. Afanasiev, C. Geantet, M. Vrinat, Overview of support effects in
hydrotreating catalysts, Catal. Today 86 (2003) 5–16, https://doi.org/10.1016/
S0920-5861(03)00400-0.
[34] Y. Zheng, J. Wang, C. Liu, Y. Lu, X. Lin, W. Li, Z. Zheng, Efficient and stable Ni-Cu
catalysts for ex situ catalytic pyrolysis vapor upgrading of oleic acid into
hydrocarbon: Effect of catalyst support, process parameters and Ni-to-Cu mixed
ratio, Renew. Energy 154 (2020) 797–812, https://doi.org/10.1016/j.
renene.2020.03.058.
8