Cu boosting the collaborative effect of Ni and H+in alloyed NiCu/saponite catalysts for hydrogenolysis of glycidol

Article abstract of DOI:10.1039/d1dt01189h

The effect of copper on various acid saponite supported Ni-Cu bimetallic catalysts, prepared with different Ni?:?Cu ratios, was studied for the liquid phase hydrogenolysis of glycidol on a batch reactor at 393 and 453 K. Characterization of the catalysts showed that Ni and Cu are in close contact as the XRD measurements evidenced the formation of an alloy. H₂chemisorption results revealed that the measured metallic area progressively decreased with an increase in the wt% of copper. In the presence of high metal activity (higher Ni wt%), the formation of 1,2-propanediol (1,2-PD) outweighed, while acid activity led to the formation of dimerization and oligomerization products. The addition of Cu and the increase of the reaction temperature decreased the diol formation but boosted the 1,3-PD/1,2-PD ratio. This could be explained by an improvement of the collaborative effect between the metal Ni and the H⁺of the saponite. Therefore, the presence of an appropriate amount of Cu allowed the control of the hydrogenation capacity of Ni and enhanced the collaborative effect of Ni and H⁺favouring the formation of 1,3-propanediol with respect to 1,2-propanediol.

Full text of DOI:10.1039/d1dt01189h

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
+
Cu boosting the collaborative effect of Ni and H in
alloyed NiCu/saponite catalysts for hydrogenolysis
of glycidol
Cite this: Dalton Trans., 2021, 50,
9198
a
b,c
a
Fiseha Bogale Gebretsadik, Javier Ruiz-Martinez, María Dolores González,
a
a
Pilar Salagre and Yolanda Cesteros
*
The effect of copper on various acid saponite supported Ni–Cu bimetallic catalysts, prepared with
different Ni : Cu ratios, was studied for the liquid phase hydrogenolysis of glycidol on a batch reactor at
393 and 453 K. Characterization of the catalysts showed that Ni and Cu are in close contact as the XRD
measurements evidenced the formation of an alloy. H chemisorption results revealed that the measured
2
metallic area progressively decreased with an increase in the wt% of copper. In the presence of high
metal activity (higher Ni wt%), the formation of 1,2-propanediol (1,2-PD) outweighed, while acid activity
led to the formation of dimerization and oligomerization products. The addition of Cu and the increase of
the reaction temperature decreased the diol formation but boosted the 1,3-PD/1,2-PD ratio. This could
+
Received 10th April 2021,
Accepted 26th May 2021
be explained by an improvement of the collaborative effect between the metal Ni and the H of the sapo-
nite. Therefore, the presence of an appropriate amount of Cu allowed the control of the hydrogenation
DOI: 10.1039/d1dt01189h
+
capacity of Ni and enhanced the collaborative effect of Ni and H favouring the formation of 1,3-propa-
rsc.li/dalton
nediol with respect to 1,2-propanediol.
1
3,14
light stability, elastic recovery and dyeability.
The hydroge-
1
. Introduction
nolysis of glycerol to 1,3-PD has been attempted by different
1
0,15–23
Biomass valorization is crucial for the sustainability of our research groups.
society and aids to solve the environmental impact related of 1,3-PD yield were obtained when expensive noble metal-
to the use of fossil fuels. One of the most explored valoriza- based catalysts were used under harsh reaction conditions and
tion processes at the academic and industrial levels is the con- long reaction times.
version of renewable oils into biodiesel. In addition, this Glycidol could be an alternative to glycerol for propanediol
The most interesting results in terms
1
,2
3
,4
1
0,12,18–23
chemical process also produces glycerol in a 10% yield. The production as it is a highly reactive molecule that could be
rapid growth of the biodiesel industry results in a glycerol over- readily obtained in one step from the reaction of glycerol and
2
4,25
supply, which negatively impacts the market value of this dimethyl carbonate
or from epichlorohydrin industry
5
–9
26
product. To make the biodiesel industry more economically wastes. Epoxides can undergo ring opening reactions in the
2
7–29
30–32
competitive, various innovative processes were developed to presence of acid solids
and supported metal catalysts.
Isomerization and hydrogenolysis of epoxides to aldehydes,
Selective hydrogenolysis of glycerol to 1,3-propanediol (1,3- ketones or alcohols using different supported metal catalysts
8
,10
transform glycerol into other added value chemicals.
3
0–32
PD) is a valorization route, especially attractive due to the high have been previously described.
cost of the current industrial processes. 1,3-PD is an impor- (THFA) was hydrogenolyzed to 1,5-pentanediol using Mo and
and Re modified Ir catalysts supported on
PPT), a polymer that displays excellent chemical resistance, SiO2 or C. Dumesic et al. reported the hydrogenolysis of
Tetrahydrofurfuryl alcohol
1
1
3
3–35
tant monomer in the synthesis of polypropylene terephthalate Re modified Rh
(
1
2
36
3
7
secondary C–O bonds of several cyclic ethers.
The
authors claimed that these reactions require a bifunctional
a
Universitat Rovira i Virgili, Departament de Química Física i Inorgànica, C/Marcel·lí catalyst. In this way, acidic O–H groups on ReO activated the
x
Domingo 1, 43007 Tarragona, Spain. E-mail: yolanda.cesteros@urv.cat
cyclic ether.
Saponite, Mx−y/n[(Mg6−yAl
b
Inorganic Chemistry and Catalysis, Debye Institute of Nanomaterials, Utrecht
Td
Oh
y
)] [(Si8−xAl
x
)
4 2
·O20(OH) ]·nH O,
University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands
c
is a trioctahedral clay of the smectite group containing an octa-
hedral (Oh) sheet of MgO between tetrahedral (Td) sheets of
KAUST Catalysis Center (KCC), King Abdullah University of Science and Technology
Bldg. 3, Level 4, Room 4275, Thuwal 23955-6900, Saudi Arabia
9198 | Dalton Trans., 2021, 50, 9198–9207
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
SiO
Paper
4
. Isomorphous substitution of Si by Al generates a negative ation. The samples were dried in an oven at 363 K overnight,
−1
layer charge, which should be compensated by interlayer calcined in static air from room temperature at 10 K min to
n+
hydrated cations (M ). When the interlayer cation is 723 K and maintained at this temperature for 5 h. Then, the
+
exchanged by H , the saponite has Brønsted acidity and can be samples were reduced in a tubular reactor under a hydrogen
flow (75 mL min− ) at 623 K for 6 h. The monometallic cata-
1
used as a support with acid properties.
The use of Ni as a cheaper alternative to noble metals in lysts were named Ni/S and Cu/S. In the case of bimetallic cata-
the hydrogenolysis of biomass-derived polyols has attracted lysts, the nomenclature was NiCu(X : Y)/S where X and Y rep-
3
8–40
great attention.
Ni has high hydrogen activation capacity resent the relative wt% ratio of Ni and Cu, respectively. One
and could be used for the production of alcohols from epox- more sample (NiCu(6 : 1)*/S) was prepared at 40 wt% of total
ides. The mechanism of ring opening depends on several metal loading with a Ni : Cu wt% ratio of 6 : 1 for comparison.
3
7
factors, such as the presence of Brønsted acidity and metals.
The catalyst precursors were labelled with a P before the name
The control of the reaction pathway will be fundamental for of the catalyst. Several spent catalysts were characterized after
the production of 1,3-PD or 1,2-PD from glycidol. In previous the reaction and were named by adding –U to the nomencla-
works, we achieved a 29% yield of 1,3-PD from the hydrogen- ture of the catalyst.
ation of glycidol over an acid saponite supported Ni catalyst,
2
.2. Characterization of catalysts
which further increased to 46.1% over a ReOx modified sup-
4
1,42
ported Ni catalyst.
More recently, Sun et al. reported a yield The X-ray diffraction (XRD) patterns of the powdered catalyst
of 1,3-PD up to 50.4% using a carbon film encapsulated Co samples were recorded on a Siemens D5000 diffractometer
4
3
catalyst. The authors concluded that the formation of 1,3-PD equipped with a CuKα radiation (λ ≈ 1.5406 Å) source. The
was sensitive to the amount of adsorbed hydrogen, being measurements were done at 5–70° 2θ diffraction angles, with
favoured at moderate adsorption values. Taking into account an angular step of 0.05° and a rate of 3 s per step. The crystal-
this idea, the aim of this work was to delve into the study of line phases were identified by cross comparison of the diffrac-
the effect of adding a second metal with less hydrogen adsorp- togram of the sample with reference data from the
tion capacity, such as Cu, on the catalytic activity of supported International Center for Diffraction Data (ICDD, release 2004)
Ni–Cu bimetallic catalysts, prepared with different Ni : Cu files. TOPAS 4.1 software was used to fit the diffractograms
ratios, for the hydrogenolysis of glycidol. The Ni activity is using the Rietveld method and the crystal structure of each
4
4
strongly dependent on the Ni nanoparticle size. The addition phase. The average crystallite size was calculated as a mean
of Cu, as a second metal, can be a good option to decrease the value from all the diffracted peaks once the instrumental con-
Ni particle size since Cu can reduce the ensemble size of Ni by tribution to the peak width was removed. The cell parameters
4
5
the formation of an alloy. Cu is a less-active metal but can be calculated from the refined diffractogram were used to
4
6,47
an efficient catalyst for selective C–O bond hydrogenolysis.
examine the formation of the Ni–Cu alloy and to determine
Some selected catalysts will be also tested at higher reaction the stoichiometry of the constituent metals in the alloy.
temperatures to study the effect of this parameter on the cata-
2
N adsorption–desorption analysis of the catalyst precursors
lytic results.
and spent catalysts were performed using a Quadrasorb SI
surface analyser at liquid nitrogen temperature (77 K) to deter-
mine the BET area, pore volume and average pore size. Before
measurement, the samples were degassed overnight at 383 K.
The spent catalysts were degassed at a higher temperature
2
. Experimental
2
.1. Preparation of catalysts
(
573 K) for 3 h in order to remove the high boiling point
Several supported Ni–Cu catalysts, with different Ni : Cu ratios, solvent, sulfolane (b.p.: 558 K), used during the catalytic reac-
were prepared by the incipient wetness impregnation method tion that could not be removed by aqueous washing. The
using H-saponite (S), the synthesis of which was previously specific surface area was determined by applying the BET
4
8
reported, as a support. One H-saponite supported monome- method and the pore size distribution was estimated by the
tallic Ni catalyst and one H-saponite supported monometallic BJH method.
Cu catalyst were also prepared as references for comparison.
TEM characterization of the catalysts was done using a
For the catalyst preparation, 1 g of the support was introduced JEOL 1011 transmission electron microscope operating at an
into a round bottom flask. A calculated volume of 0.5 M nickel accelerating voltage of 100 kV and magnification of 200k. The
or copper nitrate solution in ethanol was poured uniformly catalyst powder (0.1 mg) was dispersed in ethanol (50 µL) by
over the support for preparing supported catalysts at 30 wt% using ultrasound. Then, the suspension was placed on a
Ni or Cu loading. For the Ni–Cu bimetallic catalysts, calculated carbon coated copper grid and air dried.
amounts of Ni(NO
catalyst at Ni : Cu wt% ratios of (9 : 1), (6 : 1), (3 : 1), (1 : 1) and lyst precursors was analyzed by using the Autochem AC2920
1 : 3), were dissolved in ethanol to prepare a 0.5 M solution Micrometric apparatus. The sample calcined beforehand
3
)
2
and Cu(NO
3
)
2
, for preparing a Ni–Cu
The temperature-programmed reduction profile of the cata-
(
with respect to the total moles of the metal precursors. The (100 mg) was introduced into the sample tube equipped with a
solution was transferred uniformly into a round bottom flask thermal conductivity detector. The sample was heated from
with the support. The solvent was removed by rotary evapor- room temperature to 1173 K at a rate of 10 K min− under a
1
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 9198–9207 | 9199

View Article Online
Paper
Dalton Transactions
flow of 5% H
2
in argon (50 mL min− ). The amount of hydro- unknown products. The conversion and selectivity of the reac-
1
gen consumed was monitored using a TCD detector. The tions were calculated according to the following equations:
measured metallic area was calculated by applying the chemi-
ðmoles of glycidol t0 ꢀ moles of glycidol tf Þ
sorption factor of copper for Cu/S and the chemisorption
factor of nickel for Ni/S and NiCu catalysts taking into account
the low hydrogen adsorption capacity of copper.
Conversion ¼
moles of glycidol t0
ꢁ
100%
The surface acidity of the catalyst precursors was deter-
mined from the thermogravimetric analyses (TGA) of their
cyclohexylamine (CHA) treated samples by following the pro-
Selectivity to xð%Þ ¼
moles x ꢁ carbon in x
ꢁ 100%
moles of glycidol converted ꢁ carbon in glycidol
4
9
tocol described by Mokaya et al. The total acidity was equi-
valent to the amount of cyclohexylamine desorbed in the
temperature range of 473–923 K in which the sample
without cyclohexylamine treatment did not show any signifi-
3
3
. Results and discussion
.1. Characterization of catalytic precursors
4
9
cant mass loss. The samples without cyclohexylamine treat-
ment were analysed in the same way and used as a blank to
account for the potential mass loss unrelated to the basic The reducibility of the supported metal oxides was assessed by
probe molecule. The analysis was conducted in a Labsys temperature-programmed reduction. Fig.
shows the
Setaram TGA microbalance equipped with a programmable reduction profiles of the catalyst precursors. The reducibility of
temperature furnace. Each sample was heated under a N
a metal oxide depends on many factors, such as the nature of
flow (80 cm min ) from 303 to 1173 K at a rate of 10 K the supported metal, its particle size and its interaction with
1
2
3
−1
−
1
50
min . In all experiments, the amount of sample used was the support. The monometallic Ni catalyst precursor (PNi/S)
around 100 mg.
displayed two reduction peaks: one well-defined peak centered
The amount of organic material deposited on the spent at 770 K and a broad shoulder between 820 and1000 K. The
catalyst was determined by TG analysis. In the analysis, the lower temperature peak was due to the smaller NiO particles
spent catalyst was dried at 573 K under vacuum for 3 h in whereas the broad peak was assigned to either more aggre-
order to remove the adsorbed sulfolane solvent used in the gated NiO or NiO particles, which had a strong interaction
catalytic reaction. About 100 mg of the dried samples was put with the support. Similarly, PCu/S (Fig. 1) exhibited the main
in the sample holder and heated from 303 to 1173 K under an reduction peak at 520 K with shoulders between 390 and 470
3
−1
−1
air flow (80 cm min ) at a heating rate of 10 K min . The and at 557 K. The different contributions can be attributed to
same temperature range, as in the acidity determination CuO particles with different sizes and support interactions.
(523–923 K), was used to determine the percentage of the The presence of Cu and an increase in its wt% ratio progress-
organic product.
ively shifted the main reduction peak to lower reduction temp-
eratures. For PNiCu(3 : 1)/S, another small peak, which was
shifted to even lower reduction temperatures with respect to
2
.3. Catalytic activity
Hydrogenolysis of glycidol was carried out in a 50 mL stainless the monometallic Cu catalyst, appeared at 490 K. This peak
steel autoclave equipped with an electronic temperature con- was the main reduction peak for the samples PNiCu(1 : 1)/S
troller and a mechanical stirrer. Glycidol (3.77 mL) was dis-
solved in sulfolane (30 mL) and purged with nitrogen. A
freshly reduced catalyst (1 g) was carefully transferred to the
reactor under a positive nitrogen pressure to avoid contact
with the atmosphere. The reactor was sealed and purged three
2 2
times at 4 MPa N pressure and three times at 1 MPa H
pressure. After performing a leakage test, the temperature of
the reaction was set at 393 or 453 K, and 1 MPa H was fed
2
with stirring at 600 rpm. When the temperature in the elec-
tronic control reached the final temperature of the reaction,
the hydrogen pressure was adjusted to 5 MPa and the reaction
was started (t
while feeding hydrogen on demand. At the end of the reaction
tf), the reaction mixture was cooled and the gaseous reaction
0
). The reaction was allowed to run for 1 or 4 h
(
products were collected in a gas bag. The reaction products
were separated from the catalyst by centrifugation and the
liquid products were analysed on a Shimadzu GC-2010 chro-
matograph using a SupraWAX-280 capillary column, 1-butanol
as the internal standard and an FID detector. Some liquid
samples were subjected to GC-MS analysis to identify the Fig. 1 TPR profiles of all the catalyst precursors.
9200 | Dalton Trans., 2021, 50, 9198–9207
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
and PNiCu(1 : 3)/S. As we will discuss later, this peak could be
related to the formation of an alloy. The PNiCu(1 : 1)/S and
PNiCu(1 : 3)/S samples also showed peaks that shifted to lower
reduction temperatures compared to the monometallic PCu/S
catalyst precursor. This could be attributed to the higher redu-
cibility of the smaller copper oxide particles obtained from the
5
1,52
bimetallic catalysts.
The N -adsorption results of the support and the catalyst
2
precursors are presented in Table 1. The support (S) had a
2
−1
3
specific BET area of 196 m g and a pore volume of 0.21 cm
−
1
g
. The specific BET area and pore volume values decreased
2
−1
3
−1
to 122–112 m g and 0.17–0.12 cm g , respectively, for the
supported catalyst precursors prepared at 30 wt% total metal
loading. The value of the specific BET area decreased further
to 77 m g for the catalyst prepared at 40 wt% total metal
loading. This could be attributed to the coverage of part of the
support pore volume by the metal oxide particles. The decrease
of the pore volume was more significant when the wt% of Ni
loading increased. A similar result was reported by Obregón
2
−1
Fig. 2 XRD patterns of the catalysts.
5
3
and coworkers. They claimed that Ni is preferentially de-
posited in bigger pores while Cu takes up the external surface.
The total surface acidity of all the catalyst precursors ing the significant amounts of the NiO phase. The XRD
−
1
(
1.03–1.21 mEq CHA g ) was lower than the acidity of the pattern of Cu/S showed reflections at 43.3° and 50.4°, 2θ,
−
1
support alone (1.33 mEq CHA g ), as shown in Table 1. The which are assigned to the (111) and (200) planes of metallic
decrease in the acidity of the catalyst precursors could be Cu.
explained by the coverage of part of the acid sites with the
In the case of the Ni–Cu bimetallic catalysts, as illustrated
impregnated metal, as confirmed by the lower surface area of in Fig. 2b–f, the position of the reflection peaks shifted,
the catalyst precursors compared to the support (Table 1). It depending on the wt% ratios of Ni and Cu to lower 2θ angles
was also observed that the surface acidity of the catalyst precur- with respect to the peaks of the supported monometallic Ni
sor decreased at higher wt% Cu loading. This could be related catalyst. Accordingly, a slightly comparable shift in the posi-
to the presence of a higher amount of CuO on the external tion of the Ni peak at 44.5°, 2θ was observed for the NiCu
5
3
surface.
(9 : 1)/S and NiCu/(6 : 1) samples. The catalyst NiCu(3 : 1)/S
showed a further shift and the reflection appeared broader.
This could be related to the formation of a Ni–Cu alloy, as pre-
3
.2. Characterization of fresh catalysts
45
viously reported. For the NiCu(1 : 1)/S and NiCu(1 : 3)/S cata-
The XRD patterns of the catalysts presented the characteristic lysts, the peak shift was even more pronounced. Unlike in pre-
4
5,47
saponite reflections at 19.4°, 35.6°, 60.5°, 2θ, attributed to the vious reports
110, 020), (201) and (060) planes, respectively, together with with Cu irrespective of the initial metal ratio, in our case, the
where the Ni–Cu alloy was more enriched
(
metallic phases (Fig. 2). For the monometallic Ni catalyst, Ni/S, stoichiometry of the alloy formation was governed by the
the reflections appeared at 44.5° and 51.9°, 2θ, corresponding atomic ratios of Ni and Cu in the samples. This might be the
to the (111) and (200) planes of metallic nickel without observ- result of the calcination procedure at 723 K for 5 h that favored
good mixing of the metals. In catalysts NiCu(1 : 1)/S and NiCu
(
1 : 3)/S, especially in the latter sample, some amount of metal-
Table 1 Characterization results of the support and catalyst precursors
lic Cu, which did not form an alloy, was also present.
The cell parameters of the samples were plotted against the
atomic ratios of Ni, as shown in Fig. 3. The cell parameter
clearly showed a tendency to decrease as the Ni content of the
alloy increased, as expected from Vegard’s law. As Ni and Cu
have the same crystal structure, a continuous solid solution
between the two metals can be expected. For catalysts NiCu
a
Support and
catalyst
precursors
Average
pore radius
(Å)
Pore
Acidity
−3
BET
volume
(×10 eq.
CHA per g)
(m2 g−1)
(cm3 g−1)
5
4
S
196
126
129
127
116
116
114
112
77
21
21
25
23
29
27
30
29
30
0.21
0.13
0.13
0.14
0.17
0.17
0.17
0.18
0.15
1.33
1.21
1.22
1.22
1.19
1.07
1.03
1.05
0.81
PNi/S
PNiCu(9 : 1)/S
PNiCu(6 : 1)/S
PNiCu(3 : 1)/S
PNiCu(1 : 1)/S
PNiCu(1 : 3)/S
PCu/S
(
9 : 1)/S, NiCu(6 : 1)/S, NiCu(3 : 1)/S and NiCu(1 : 1)/S, there was
a good agreement between the Vegard’s law and the cell
atomic ratios of Ni and Cu in the sample and the stoichio-
metry of the alloy were almost identical. In sample NiCu(1 : 3)/
S, probably the surface was enriched with Cu, as confirmed by
the deviation of its cell parameter from what would be
PNiCu(6 : 1)*/S
a
Calculated from TGA of adsorbed cyclohexylamine.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 9198–9207 | 9201

View Article Online
Paper
Dalton Transactions
Fig. 3 Cell parameters calculated from the XRD patterns as a function
of Ni content and Vegard’s curve.
expected from Vegard’s law. This might be due to thermo-
dynamic reasons whereby the lower surface free energy of Cu,
when compared to Ni, led to the enrichment of Cu in the exter-
5
5
nal surface of the bimetallic catalyst. The average crystallite
Fig. 4 TEM images of the (A) Ni/S (300k), (B) NiCu(3 : 1)/S (300k), (C)
sizes of Ni, Cu or the Ni–Cu alloy of the catalysts were deter- NiCu(1 : 1)/S (500k), (D) NiCu(1 : 3)/S (300k) (E) Cu/S (80k) catalysts.
mined from their XRD patterns by applying the Scherrer
equation (Table 2). The monometallic Ni and Cu catalysts
showed particle sizes of 8 and 62 nm, respectively. The NiCu ively, showed particles with good size homogeneity. The
(9 : 1)/S, NiCu(6 : 1)/S and NiCu(3 : 1)/S samples had the Ni–Cu average particle size of Ni/S (15 nm) was much smaller than
alloy particle sizes of 12, 11 and 9 nm, respectively. For the that of Cu/S (95 nm). This is in agreement with the crystallite
other two samples, NiCu(1 : 1)/S and NiCu(1 : 3)/S, two types of size calculated from the XRD patterns. The NiCu(3 : 1)/S cata-
particles, namely, metallic Cu and Ni–Cu alloy were observed. lyst, in Fig. 4B, had a particle size distribution similar to that
The average crystallite sizes of the alloy particles were 11 and 7 of the monometallic Ni/S. For the other two samples, NiCu
and the Cu crystallite sizes were 18 and 41 nm, correspond- (1 : 1)/S and NiCu(1 : 3)/S, a more heterogeneous particle size
ingly, for NiCu(1 : 1)/S and NiCu(1 : 3)/S. These results suggest distribution was observed(Fig. 4C and D, respectively). Small
that the formation of a Ni–Cu alloy helps in the dispersion of particles, which might correspond to the alloys, and bigger
Cu.
particles of metallic copper were observed. This was in agree-
Representative TEM images of some of the catalysts are dis- ment with the XRD pattern discussed previously that showed
played in Fig. 4. TEM micrographs gave direct evidence of the two distinct peaks for these two samples corresponding to Ni–
degree of dispersion and homogeneity of the metal particles. Cu alloy and metallic Cu.
The monometallic Ni/S and Cu/S, in Fig. 4A and E, respect-
The metallic area of fresh catalysts, determined from the
hydrogen chemisorption analysis, is shown in Table 2. Ni/S
2
−1
showed the highest metallic area (7.5 m
g ). Interestingly,
the measured metallic area decreased for the bimetallic cata-
lysts when increasing the wt% of Cu loading. Thus, the NiCu
(1 : 1)/S and NiCu(1 : 3)/S catalysts had metallic areas of 0.4 and
Table 2 Characterization of fresh catalysts
b
Metal crystallite
size XRD (nm)
Particle size
TEM (nm)
Metallic area
2 −1
2
−1
Catalysts
(m g sample)
0.1 m g , respectively. The Cu/S catalyst did not show an
appreciable amount of hydrogen strongly chemisorbed. These
results were in agreement with the low hydrogen chemisorp-
tion capacity reported for Cu.
Ni/S
8
12
11
9
11 (18)
7 (41)
62
—
15
—
—
14
13
18
95
—
7.5
3.4
2.8
2.4
0.4
0.2
0.1
4.3
a
NiCu(9 : 1)/S
NiCu(6 : 1)/S
NiCu(3 : 1)/S
NiCu(1 : 1)/S
NiCu(1 : 3)/S
Cu/S
a
56
a
a
a
a
a
a
3.3. Catalytic activity
a
Ni–Cu(6 : 1)*/S
The effect of the type of metal (Ni or Cu) and the relative com-
position of the two metals in the Ni–Cu bimetallic catalysts on
the product distribution of glycidol hydrogenolysis were
studied at 393 K for 4 h. The major products of the reaction
a
Particle size of the Ni–Cu alloy. The values between parentheses
b
correspond to the Cu particle size.
chemisorption.
Measured by hydrogen
9202 | Dalton Trans., 2021, 50, 9198–9207
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Fig. 5 Catalytic activity results for the hydrogenolysis of glycidol at
393 K for 4 h.
Scheme 1 A simplified reaction pathway for acid, Ni/Cu and Ni/Cu.
whereas conversion decreased further to 56–34% for NiCu
1 : 1)/S, NiCu(1 : 3)/S and Cu/S as the wt% ratio of Cu
(
identified from GC-MC analysis include 1,3-PD, 1,2-PD, propa- increased. These conversion values are in good agreement
nol and cyclic dimerization products such as dioxane and diox- with the measured metallic area, taking into account the
olane derivatives. The proposed acid, metal, and acid/metal similar acidity of the catalysts, and could be explained by the
5
7
catalyzed reaction pathways leading to each of the products higher hydrogenating capacity of Ni compared to Cu at this
identified from the GC-MS analysis are illustrated in Scheme 1. comparatively lower reaction temperature. Regarding the
The metal and acid sites promote the formation of 1,3-PD; product distribution of the catalysts, it was observed that the
therefore, we foresee that close contact between both the active total diol selectivity decreased and the propanol and dioxane
sites will be crucial. The acid catalyzed isomerization reaction and dioxolane selectivity values increased upon increasing the
products, reported by other authors on differently substituted wt% ratio of Cu in the catalysts (Table 3 and Fig. 5). We
2
7,28
oxiranes,
were not detected. This might be because the propose that diols might be formed mainly by nickel catalyzed
acid catalyzed isomerization reaction products of glycidol, the hydrogenation, thus favoring the 1,2-PD formation. However,
bifunctional 3-hydroxypropionaldehyde or acetol, were the 1,3-PD/1,2-PD ratio increased with the increase of Cu. This
unstable under our reaction conditions and could undergo suggests that it is necessary to modify the activity of Ni in
hydrogenation, condensation or oligomerization reactions. order to promote the collaborative effect between the acid and
The formation of high molecular weight oligomerization reac- metal sites. The acid activated oxirane ring of glycidol could
tion products, deposited on the spent catalysts, was confirmed favor the cycle opening by the cleavage of the more hindered
for some catalysts by temperature programmed oxidation- C2, as has been commented in the literature for other cyclic
2
7
thermogravimetric analysis as commented below.
ethers, thus favoring the formation of 1,3-PD. The high total
The catalytic activity results of the catalysts tested at 393 K diol selectivity values of the catalysts with high Ni wt% ratio
for 4 h are shown in Table 3 and some of them are represented confirmed the higher hydrogenating capacity of Ni and was in
in Fig. 5. The Ni/S catalyst showed full conversion. However, agreement with their higher measured metallic area values. In
the conversion values decreased when increasing the wt% of the case of Ni/S catalyst, the lowest 1,3-PD/1,2-PD ratio (0.17)
2
−1
Cu loading in the catalysts. Thus, NiCu(9 : 1)/S, NiCu(6 : 1)/S might indicate that its higher metallic area (7.5 m g of cata-
and NiCu(3 : 1)/S exhibited a comparable conversion (78–73%) lyst) could lead to direct hydrogenation without the involve-
Table 3 Catalytic activity of all the catalysts at 393 K for 4 h
Selectivity (%)
Conversion
(%)
1,2-
PD
1,3-
PD
1,2-PD + 1,3-
PD
Dioxane/
Propanol dioxolane
1,3-PD/1,2-PD
Others ratio
1,3-PD yield
(%)
Catalyst
Ni/S
100
78
55
48
43
35
28
9
13
23
22
20
17
15
5
91
78
70
63
52
42
14
82
3
6
—
8
6
8
7
5
7
7
5
4
0.17
0.42
0.46
0.47
0.49
0.53
0.56
0.19
13
18
17
15
10
6
NiCu(9 : 1)/S 78
NiCu(6 : 1)/S 77
NiCu(3 : 1)/S 73
NiCu(1 : 1)/S 56
NiCu(1 : 3)/S 37
12
13
19
23
34
9
11
18
21
28
47
5
Cu/S
NiCu(6 : 1)
34
70
2
9
69
13
*
/S
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 9198–9207 | 9203

View Article Online
Paper
Dalton Transactions
ment of acid sites. This could favor the hydrogenolysis of the Cu, a less active metal, to control the formation of cracking
less sterically hindered primary C–O bond leading to the for- gaseous products and the deep hydrogenation to carbonaceous
mation of 1,2-PD. In the other catalysts, in which the metallic materials that can be produced at higher reaction tempera-
areas were moderate or low, the contribution of such direct tures. The consequences of the selectivity of 1,3-PD and 1,2-PD
reactions could be lower and hydrogenolysis, through acid acti- were also studied (Table 4). The types of products that were
vation, favored 1,3-PD formation.
formed at 453 K were similar to those obtained at 393 K. The
In the case of propanol, for the Ni/S catalyst, the selectivity most important difference was observed in product distri-
was 3% and for other catalysts, the selectivity increased pro- bution. The diol selectivity decreased and the selectivity to pro-
gressively when the measured metallic areas decreased and panol and dioxane or dioxolane increased for all catalysts
the wt% of Cu in the catalysts increased. This might indicate when the reaction temperature was increased from 393 to
that the main route for propanol formation was not via hydro- 453 K (Tables 3 and 4). However, an increase in the 1,3-PD/1,2-
genolysis, as could be expected from the lower activity of Cu PD ratio was also observed for the catalysts used at 453 K com-
and the lower degree of hydrogenolysis expected at such lower pared to those tested at 393 K. In general, the selectivity to
reaction temperature (393 K) for Ni/S (Scheme 1). In the litera- “others” was more significant when the reactions were carried
ture, it was suggested that Cu has a deoxygenation effect on out at higher temperatures and with catalysts of high Cu wt%
3
2
epoxide rings at similar reaction conditions as those used in ratio. All catalysts displayed higher conversion values at 453 K
our experiments. Hence, propanol could be the product of the than at 393 K. The increase of activity was higher for the cata-
deoxygenation of glycidol to 3-propen-1-ol followed by hydro- lysts with higher wt% of Cu. The conversion values were com-
genation. The dioxane or dioxolane selectivity increased with a parable for Ni/S, NiCu(9 : 1)/S, NiCu(6 : 1)/S and NiCu(3 : 1)/S
decrease in the metallic area. This suggests that when the catalysts (99–92%) (Table 4). The activity dropped to 82 and
hydrogenolysis activity is lower, the contribution of acid cata- 75%, respectively, for NiCu(1 : 3)/S and Cu/S. The diol selecti-
lyzed reaction is higher and favors the intermolecular oligo- vity decreased and the propanol selectivity increased at higher
merization and dimerization of the reaction products.
reaction temperatures. This could be explained by over hydro-
In order to establish the role of acidity in the selective for- genolysis of the diols to propanol. Similar results were
mation of 1,3-PD, the catalytic activity of the NiCu(6 : 1)*/S reported for diols over hydrogenolysis for supported Ni cata-
4
2
catalyst with a higher total metal loading (40 wt%) was deter- lysts at comparable reaction temperatures. In contrast, for
mined at 393 K for 4 h. This catalyst showed the lowest surface catalysts with a lower metallic area and higher Cu wt% ratio,
−
1
32
acidity (0.81 mEq CHA g ) among all the catalysts tested, the Cu catalyzed deoxygenation pathway resulted, initially, in
probably due to the coverage of the acid sites by metal sites the formation of acrolein, which upon hydrogenation gave
2
−1
and its higher metallic area (4.3 m g of catalyst) compared propanol.
2
−1
to NiCu(6 : 1)/S (2.8 m g of catalyst). This catalyst showed a
Dioxane and dioxolane were not observed for the Ni/S cata-
higher total diol selectivity and lower propanol and dioxane or lyst at both 393 and 453 K suggesting that the acid catalyzed
dioxolane selectivity compared to the other bimetallic cata- reactions were less important for the catalyst without copper
lysts. In addition, the 1,3-PD/1,2-PD ratio was the lowest (0.19) due to the higher activity of Ni. In contrast, the dioxane and
of all the bimetallic catalysts. This might suggest that the low dioxolane amounts were higher for the catalysts with a lower
acidity of the catalyst might decrease the probability of acid metallic area (Table 4). A thermogravimetric-temperature pro-
activation necessary for the formation of 1,3-PD. The reaction grammed oxidation analysis (TG-TPO) of the spent catalysts
could be dominated by direct hydrogenolysis leading to the revealed a higher percentage of carbon deposition materials
cleavage of the less sterically hindered primary C–O bond and, when the reaction was performed at 453 K and when the cata-
therefore, to the formation of 1,2-PD.
lysts with a higher amount of Cu were used (Table 5). This
The effect of reaction temperature was studied by testing might suggest that the “other” products in this reaction were
several of the Ni–Cu bimetallic catalysts at different relative high molecular weight oligomerization products that could
wt% of Ni and Cu at 453 K in order to evaluate the effect of not be identified in the GC-MS analysis.
Table 4 Catalytic activity of the selected catalysts at 453 K for 1 h
Selectivity (%)
Conversion
(%)
Dioxane/
dioxolane
1,3-PD/1,2-PD
ratio
1,3-PD yield
(%)
Catalyst
1,2-PD
1,3-PD
1,2-PD + 1,3-PD
Propanol
Others
Ni/S
99
96
94
92
82
75
95
57
41
37
32
18
5
23
22
21
19
13
4
80
63
58
50
31
9
14
16
21
22
32
45
21
—
15
15
18
26
31
4
6
6
6
10
11
15
5
0.40
0.54
0.57
0.59
0.72
0.80
0.44
23
21
20
17
11
3
NiCu(9 : 1)/S
NiCu(6 : 1)/S
NiCu(3 : 1)/S
NiCu(1 : 3)/S
Cu/S
NiCu(6 : 1)*/S
48
21
70
20
9204 | Dalton Trans., 2021, 50, 9198–9207
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Table 5
results of the spent catalysts
N
2
physisorption and temperature programmed oxidation tested at a higher reaction temperature (453 K) exhibited a
lower BET area and pore volume. These results could be
explained considering the differences in the residual carbon
compounds deposited on the catalyst surface during the reac-
tion. These differences were supported by the quantification of
coke formation by the temperature programmed oxidation-
thermogravimetric analysis of the spent catalysts (Table 5).
The results showed a lower percentage of the deposited
organic compounds for the samples with lower Cu wt% and
for all catalysts used at lower reaction temperatures.
BET
Pore volume
% organic deposition
(TPO)
(m2 g−1)
3
−1
Used catalyst
(cm g
)
a
Ni/S–U
157
155
156
155
153
134
131
124
110
130
118
101
0.20
0.20
0.20
0.19
0.20
0.17
0.17
0.16
0.15
0.14
0.16
0.15
n.d
1
1
1
3
4
2
NiCu(9 : 1)/S–U
NiCu(9 : 1)/S–U
NiCu(6 : 1)/S–U
NiCu(6 : 1)/S–U
NiCu(3 : 1)/S–U
NiCu(1 : 1)/S–U
NiCu(1 : 3)/S–U
NiCu(1 : 3)/S–U
Cu/S–U
a
a
a
2
7
a
2
9
1
a
Cu/S–U
4. Conclusions
a
Ni–Cu(6 : 1)*/S–U
Acid saponite supported Ni–Cu bifunctional catalysts were
active for the hydrogenolysis of glycidol. We investigated the
activity of the materials with different Ni–Cu ratios and reac-
tion temperatures. Characterization of the catalysts revealed
the formation of a Ni–Cu alloy and a decrease in the metallic
area, as measured by hydrogen chemisorption, with the
addition of Cu. This indicates that Ni and Cu are in close
contact. The presence of Cu moderated the hydrogenation
activity of Ni, decreasing the yield of diols, but increasing the
selectivity towards 1,3-PD. These results suggest that Cu is
a
The results correspond to the catalysts used at 453 K for 1 h whereas
the rest of the results are for the catalysts used at 393 K for 4 h.
The 1,3-PD/1,2-PD ratio increased when increasing the reac-
tion temperature for all catalysts (Table 4). For example, the
1
0
in the mechanism leading to the formation of diols when the
reaction temperature was increased. As we proposed earlier,
the higher acid activity at 453 K could favor hydrogenolysis
,3-PD/1,2-PD ratio drastically increased from 0.17 at 393 K to
.40 at 453 K for Ni/S. This might suggest that there was a shift
+
required to modulate Ni activity and then promote the H -
+
metal collaborative mechanism. At higher reaction tempera-
tures, the acid activation was improved, catalyzed reactions
were promoted and, thus, the 1,3-PD/1,2-PD ratio increased.
However, this increase in the reaction temperature with high
Cu loading decreased diol selectivity significantly and the
selectivity for Cu catalyzed deoxygenation and acid catalyzed
polymerization and oligomerization reaction products
increased.
through collaborative action between H and metal sites favor-
ing the formation of 1,3-PD. This ratio increased further when
the wt% ratio of Cu increased suggesting that the acid cata-
lyzed route is even more important for a lower metallic area,
which is responsible for the direct hydrogenolysis route.
A similar trend in conversion and selectivity was observed
at higher reaction temperatures when NiCu(6 : 1)*/S, with
lower surface acidity and higher metallic area, was checked for
comparison (Table 4). However, the diol selectivity was higher
and the dioxane/dioxolane selectivity was lower for this catalyst
compared to that of NiCu(6 : 1)/S. This was in agreement with
the higher metallic area and lower surface acidity of the
former sample. Although the 1,3-PD/1,2-PD ratio increased for
NiCu(6 : 1)*/S at higher reaction temperatures, the value was
lower than that of NiCu(6 : 1)/S indicating that the collaborative
mechanism was less important for the former catalyst. This
could be due to the combined effect of the higher metallic
area and lower surface acidity of this catalyst.
Author contributions
Fiseha Bogale Gebretsadik: investigation, visualization, and
writing-original draft; Javier Ruíz-Martínez: validation and
supervision; María Dolores González: conceptualization, vali-
dation and writing-original draft; Pilar Salagre: conceptualiz-
ation, resources and supervision; Yolanda Cesteros: conceptu-
alization, resources, supervision, writing-original draft,
writing-review and editing.
3
.4. Characterization of spent catalysts
The XRD pattern of the spent catalysts did not show significant
variations with respect to those of the fresh catalysts. The N
Conflicts of interest
2
-
There are no conflicts to declare.
adsorption results of the spent catalysts are presented in
Table 5. Comparing these results with the corresponding
values obtained for the catalyst precursors (Table 1), the spent
catalysts, in general, had higher BET areas and pore volumes
Acknowledgements
than their corresponding catalyst precursors. This could be The authors are grateful for the financial support from the
due to the higher metal dispersion of the spent catalyst, which Ministerio de Economia y Competividad of Spain and FEDER
was also reported for a spent Cu catalyst supported on hector- funds (CTQ2011-24610) and the recognition from the
5
8
ite. However, the spent catalysts with a higher Cu wt% and Generalitat de Catalunya (2017 SGR 798).
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 9198–9207 | 9205

View Article Online
Paper
Dalton Transactions
2
6 D. Cespi, R. Cucciniello, M. Ricciardi, C. Capacchione,
I. Vassura, F. Passarini and A. Proto, Green Chem., 2016, 18,
4559–4570.
Notes and references
1
A. J. Ragauskas, C. K. Williams, B. H. Davison, G. Britovsek,
J. Cairney, C. A. Eckert, W. J. Frederick Jr., J. P. Hallett, 27 A. Fási, I. Pálinkó and I. Kiricsi, J. Catal., 1999, 188, 385–
D. J. Leak, C. L. Liotta, J. R. Mielenz, R. Murphy, 392.
R. Templer and T. Tschaplinski, Science, 2006, 311, 484– 28 A. Fási, I. Pálinkó and I. Kiricsi, Stud. Surf. Sci. Catal., 2000,
89. 130, 839–844.
S. Fernando, S. Adhikari, C. Chandrapal and N. Murali, 29 I. Salla, O. Bergadà, P. Salagre, Y. Cesteros, F. Medina,
4
2
3
4
5
6
7
Energy Fuels, 2006, 20, 1727–1737.
J. ten Dam and U. Hanefeld, ChemSusChem, 2011, 4, 1017–
J. E. Sueiras and T. Montanari, J. Catal., 2005, 232, 239–
245.
1
034.
A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
411–2502.
G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006,
06, 4044–4098.
30 A. Fási, I. Pálinkó, K. Hernadi and I. Kiricsi, Catal. Lett.,
2002, 81, 237–240.
31 M. Bartók, A. Fási and F. Notheisz, J. Catal., 1998, 175, 40–
47.
32 M. Bartók, A. Fási and F. Notheisz, J. Mol. Catal. A: Chem.,
1998, 135, 307–316.
33 S. Koso, I. Furikado, A. Shimao, T. Miyazawa, K. Kunimori
and K. Tomishige, Chem. Commun., 2009, 2035–
2037.
2
1
E. van Ryneveld, A. S. Mahomed, P. S. van Heerden and
H. B. Friedrich, Catal. Lett., 2011, 141, 958–967.
M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi and
C. D. Pina, Angew. Chem., Int. Ed., 2007, 46, 4434–
4
440.
34 S. Koso, N. Ueda, Y. Shinmi, K. Okumura, T. Kizuka and
K. Tomishige, J. Catal., 2009, 267, 89–92.
8
9
C. H. C. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. M. Lu,
Chem. Soc. Rev., 2008, 37, 527–549.
A. Behr, J. Eilting, K. Irawadi, J. Leschinski and F. Lindner,
Green Chem., 2008, 10, 13–30.
35 K. Chen, S. Koso, T. Kubota, Y. Nakagawa and
K. Tomishige, ChemCatChem, 2010, 2, 547–555.
36 K. Chen, K. Mori, H. Watanabe, Y. Nakagawa and
K. Tomishige, J. Catal., 2012, 294, 171–183.
37 M. Chia, Y. J. Pagán-Torres, D. D. Hibbitts, Q. Tan,
H. N. Pham, A. K. Datye, M. Neurock, R. J. Davis and
J. A. Dumesic, J. Am. Chem. Soc., 2011, 133, 12675–12689.
1
0 L. Z. Qin, M. J. Song and C. L. Chen, Green Chem., 2010, 12,
466–1472.
1
1
1
1 G. A. Kraus, Clean: Soil, Air, Water, 2008, 36, 648–651.
2 J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion,
C. Pinel and C. Rosier, Green Chem., 2004, 6, 359– 38 L. Ye, X. Duan, H. Lin and Y. Yuan, Catal. Today, 2012, 183,
3
61.
65–71.
1
1
3 P. N. Caley and R. C. Everett, US Patent 3350871, 1967.
4 D. Zimmerman and R. B. Isaacson, US Patent 3814725,
39 U. Saxena, N. Dwivedi and S. R. Vidyarthi, Ind. Eng. Chem.
Res., 2005, 44, 1466–1473.
1
974.
40 L. Zhou, A. Wang, C. Li, M. Zheng and T. Zhang,
ChemSusChem, 2012, 5, 932–938.
1
1
1
1
1
2
2
2
5 M. Akiyama, S. Sato, R. Takahashi, K. Inui and M. Yokota,
Appl. Catal., A, 2009, 371, 60–66.
41 F. B. Gebretsadik, J. Ruiz-Martinez, P. Salagre and
Y. Cesteros, Appl. Catal., A, 2017, 538, 91–98.
42 F. B. Gebretsadik, J. Llorca, P. Salagre and Y. Cesteros,
ChemCatChem, 2017, 9, 3670–3680.
43 Y. Sun, Z. Cai, X. Li, P. Chen and Z. Hou, Catal. Sci.
Technol., 2019, 9, 5022–5030.
6 I. Gandarias, P. L. Arias, J. Requies, M. B. Güemez and
J. L. G. Fierro, Appl. Catal., B, 2010, 97, 248–246.
7 Y. Nakagawa, M. Tamura and K. Tomishige, J. Mater. Chem.
A, 2014, 2, 6688–6702.
8 L. Gong, Y. Lu, Y. Ding, R. Lin, J. Li, W. Dong, T. Wang and
W. Chen, Appl. Catal., A, 2010, 390, 119–126.
44 V. Ponec, Appl. Catal., A, 2001, 222, 31–45.
9 Y. Nakagawa, Y. Shinmi, S. Koso and K. Tomishige, 45 I. Gandarias, S. G. Fernández, M. El Doukkali, J. Requies
J. Catal., 2010, 272, 191–194. and P. L. Arias, Top. Catal., 2013, 56, 995–1007.
0 J. Wang, M. Yang and A. Wang, Chin. J. Catal., 2020, 41, 46 S. Sato, M. Akiyama, R. Takahashi, T. Hara, K. Inui and
311–1319. M. Yokota, Appl. Catal., A, 2008, 347, 186–191.
1 Y. Liang, G. Shi and K. Jin, Catal. Lett., 2020, 150, 2365– 47 P. Li, J. Liu, N. Nag and P. A. Crozier, J. Catal., 2009, 262,
376. 73–82.
2 W. Zhou, Y. Li, X. Wang, D. Yao, Y. Wang, S. Huang, 48 F. B. Gebretsadik, D. Mance, M. Baldus, P. Salagre and
1
2
W. Li, Y. Zhao, S. Wang and X. Ma, J. Catal., 2020, 388,
54–163.
3 Z. Xi, Z. Hong, F. Huang, Z. Zhu, W. Jia and J. Li, Catalysts,
020, 10, 312.
4 J. W. Yoo, Z. Mouloungui and A. Gaset, US Patent 6316641,
001.
Y. Cesteros, Appl. Clay Sci., 2015, 114, 20–30.
49 R. Mokaya, W. Jones, S. Moreno and G. Poncelet, Catal.
Lett., 1997, 49, 87–94.
50 A. Carrero, J. A. Calles and A. J. Vizcaíno, Chem. Eng. J.,
2010, 163, 395–402.
1
2
2
2
2
2
51 J.-P. Shen and C. Song, Catal. Today, 2002, 77, 89–98.
5 R. Bai, H. Zhang, F. Mei, S. Wang, T. Li, Y. Gu and G. Li, 52 F. E. López-Suárez, A. Bueno-López and M. J. Illán-Gómez,
Green Chem., 2013, 15, 2929–2934.
Appl. Catal., B, 2008, 84, 651–658.
9206 | Dalton Trans., 2021, 50, 9198–9207
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
5
3 I. Obregón, E. Corro, U. Izquierdo, J. Requies and 56 J. Batista, A. Pintar, J. P. Gomilšek, A. Kodre and
P. L. Arias, Chin. J. Catal., 2014, 35, 656–
62.
4 A. West, Solid State Chemistry and its Application, John
F. Bornette, Appl. Catal., A, 2001, 217, 55–68.
57 B. C. Miranda, R. J. Chimentão, J. Szanyi, A. H. Braga,
J. B. O. Santos, F. Gispert-Guirado, J. Llorca and F. Medina,
Appl. Catal., B, 2015, 166–167, 166–180.
6
5
Wiley & Sons, New York, 1985.
5
5 A. Kitla, O. V. Safonova and K. Föttinger, Catal. Lett., 2013, 58 T. Sánchez, P. Salagre, Y. Cesteros and A. Bueno-López,
43, 517–530. Chem. Eng. J., 2012, 179, 302–311.
1
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 9198–9207 | 9207