Hydrogenolysis of glycerol in an aqueous medium over Pt/WO3/zirconium phosphate catalysts studied by1H NMR spectroscopy

Article abstract of DOI:10.1039/d0nj05557c

Bifunctional Pt/WO₃/zirconium phosphate catalyzes the liquid-phase hydrogenolysis of glycerol in an aqueous medium.¹H NMR spectroscopy (solvent suppression pulse program) is employed to monitor this reaction. Propanediols (1,3 + 1,2-PDO) formed as the major product along with propanols (1- and 2-POs) as the minor product. A synergistic enhancement in glycerol conversion and selectivity to 1,3-PDO was observed when both Pt and WO₃were present in the catalyst. Avolcano-shapevariation of catalytic activity with W content was observed. A catalyst with 8 wt% W and 1 wt% Pt exhibited the highest selective hydrogenolysis performance (glycerol conversion = 92.3% and total PDOs selectivity = 45.9% and 1,3-PDO selectivity = 20.8% at 200 °C). Dispersed Pt in contact with polytungstate-type WO₃species was found to be the active catalytic site.¹H NMR spectroscopy is demonstrated as an attractive technique toquantifythe products of a glycerol hydrogenolysis reaction.

Full text of DOI:10.1039/d0nj05557c

NJC
View Article Online
View Journal | View Issue
PAPER
Hydrogenolysis of glycerol in an aqueous medium
over Pt/WO /zirconium phosphate catalysts
studied by H NMR spectroscopy†
3
Cite this: New J. Chem., 2021,
1
45, 5013
Susmita Bhowmik,
Nagasuresh Enjamuri and Srinivas Darbha
*
3
Bifunctional Pt/WO /zirconium phosphate catalyzes the liquid-phase hydrogenolysis of glycerol in an
1
aqueous medium. H NMR spectroscopy (solvent suppression pulse program) is employed to monitor
this reaction. Propanediols (1,3 + 1,2-PDO) formed as the major product along with propanols (1- and
2-POs) as the minor product. A synergistic enhancement in glycerol conversion and selectivity to 1,3-
PDO was observed when both Pt and WO were present in the catalyst. A volcano-shape variation of
3
Received 13th November 2020,
Accepted 10th February 2021
catalytic activity with W content was observed. A catalyst with 8 wt% W and 1 wt% Pt exhibited the
highest selective hydrogenolysis performance (glycerol conversion = 92.3% and total PDOs selectivity =
4
5.9% and 1,3-PDO selectivity = 20.8% at 200 1C). Dispersed Pt in contact with polytungstate-type WO
3
DOI: 10.1039/d0nj05557c
1
species was found to be the active catalytic site. H NMR spectroscopy is demonstrated as an attractive
rsc.li/njc
technique to quantify the products of a glycerol hydrogenolysis reaction.
forming ultra-pure glycerol reactant from bio-glycerol. Catalysts
with strong acidity are the best choice for 1,3-PDO formation.
1
Introduction
Hydrogenolysis of glycerol, a by-product of the modern But most of the acid catalysts are sensitive to aqueous medium
6
oleochemical industry, to propanediols (1,2- and 1,3-PDOs) is and result in poor yield of 1,3-PDO. All these limitations
of great importance, as these diols are the building-block demand efficient catalysts for 1,3-PDO formation from aqueous
7
chemicals of polymers, paints, cleaning products, carpets, glycerol. Among the reported solid catalysts, supported
1
,2
adhesives and personal care items. Considerable progress Ir–ReO
x x
and Pt–WO are the most studied ones as they are
in catalyst development for 1,2-PDO has already been more active and selective toward 1,3-PDO, due to adequate
3
,4
achieved. But the conversion of glycerol to 1,3-PDO is more synergy between the metal oxide and metal nanoparticles.
8
–11
challenging. The 21-OH group of glycerol is sterically hindered Tomishige and co-workers
by the two 11-OH groups, making it less accessible to the active supported on SiO catalyzes this reaction under mild conditions
sites in the catalyst. The activation energies for dehydration of (B120 1C, 80 bar, 12–24 h) and converts glycerol into 1,3-PDO
1- and 21-OH groups (based on neutral glycerol) are similar with good activity by the direct dehydroxylation approach.
found that ReO -modified Ir
x
2
1
ꢀ
1
ꢀ1
12
(
70.9 kcal mol for 21-OH and 73.2 kcal mol for 11-OH) and In another study, Arundhathi et al. disclosed a highly selective
boehmite-based Pt/WO /AlOOH catalyst which catalyzes the
x
for 21-OH and 194.8 kcal mol for the 11-OH group). These reaction at 180 1C and 50 bar H . Several groups extended the
ꢀ1
the proton affinities are also nearly the same (195.4 kcal mol
ꢀ
1
5
2
small differences in reactivity make it difficult to achieve high studies and probed the influence of the support, metal loading
1
3–20
selectivity to 1,3-PDO. A moderate temperature operation is and additives on the performance of a Pt–WO catalyst.
x
needed as high temperature leads to deep hydrogenolysis.
Water is contained in raw bio-glycerol and it is also formed different forms depending on the method of preparation.
Zirconium phosphates (ZrP) are layered materials present in
2
1
as a co-product in the glycerol hydrogenolysis reaction. Thus, it Their acidity, high thermal and chemical stability and
makes more sense to develop a catalytic process for glycerol in ion-exchange capacity (like in zeolites) make them attractive
2
2
23
an aqueous medium to avoid the expensive distillation step for in catalysis. Due to their high hydrothermal stability, they
have been used as catalysts and supports in biomass
2
4–30
transformation.
Pt supported on metal phosphates
Catalysis and Inorganic Chemistry Division, CSIR-National Chemical Laboratory,
Dr Homi Bhabha Road, Pune-411008, India. E-mail: d.srinivas@ncl.res.in;
Fax: +91-20-25902633; Tel: +91-20-25902018
exhibited high activity for the hydrogenolysis of glycerol to
1
2
- and 2-propanols (1- and 2-POs) with a yield of 97% at
20 1C and atmospheric pressure in a fixed-bed reactor. As
3
1
†
Electronic supplementary information (ESI) available: Catalytic activity calcula-
tion, XRD profiles and nitrogen physisorption data. See DOI: 10.1039/d0nj05557c the combination of Pt and W exhibits high selectivity toward
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 5013ꢀ5022 | 5013

View Article Online
Paper
NJC
ꢀ
1
1
,3-PDO, in this study, we have attempted this combination on 450 1C for 4 h (ramp rate = 2 1C min ). The sample obtained was
ZrP and probed the hydrogenolysis reaction of aqueous glycerol named as yWO /ZrP, where y was the nominal weight percentage
at a moderate temperature (200 1C). Additionally, we report the of tungsten in the sample.
3
1
application of proton nuclear magnetic resonance ( H NMR)
Then, in the second-step, the required amount of tetraammine-
spectroscopy – solvent suppression pulse program for identification platinum(II) nitrate (Pt source) was dissolved in 10 ml of Milli-Q
and quantification of glycerol hydrogenolysis products. The water. yWO /ZrP (1 g) was added to it. The suspension was
3
product analysis is conventionally carried out by chromato- stirred at 40 1C for 4 h. Water in it was removed by using a
graphic methods. The NMR method is a simple and non- rotary evaporator. The solid formed was dried and calcined as
destructive tool and opens up an opportunity for this reaction described above. Then, it was reduced in a flow of hydrogen
ꢀ
1
ꢀ1
as a characterization tool. While our work was in progress, Luo (50 ml min ) at 350 1C for 2.5 h (ramp rate = 10 1C min ). This
3
2
et al. reported a study using a similar catalyst (Pt/7WO -ZrP) catalyst was labeled as xPt/yWO /ZrP. Here, x is the nominal
x
3
that resulted mainly in 1-PO (81% yield) in a continuous flow weight percentage of Pt and y is the weight percentage of W in
reaction conducted at a higher temperature (270 1C). Our study the catalyst.
reports the catalyst behavior at lower temperatures and we
observe formation of PDOs as a major product.
2.2 Catalyst characterization techniques
Structural characteristics of the catalyst samples were determined
by X-ray powder diffraction (XRD) performed using a PANalytical
X’Pert Pro diffractometer with a dual goniometer, X’celerator
2
Experimental section
solid-state detector module and Ni-filtered Cu-K
l = 0.15418 nm) operating at 40 kV and 30 mA. The diffraction
.1.1 ZrP. In a typical synthesis of ZrP by a hydrothermal data were collected between the incident angles of 2y = 51 and
a
radiation
2.1 Catalyst preparation
(
2
3
3
ꢀ1
method, 100 ml of Milli-Q water was taken in a beaker (250 ml 901 at a scan rate of 1.4161 min . Textural properties of the
capacity), and 3.9246 g of orthophosphoric acid (85 wt%, samples were deduced from nitrogen-physisorption studies con-
Thomas Baker) was added drop-wise with constant stirring. ducted on an Autosorb-1C Quantachrome, USA instrument. Prior
Then, 6.5572 g of zirconium isopropoxide (70 wt% solution in to nitrogen adsorption (at ꢀ196 1C), the catalyst samples were
1
-PO, Sigma-Aldrich) was added to it. The mixture was stirred pretreated by degassing at 350 1C for 3 h. The Brunauer–Emmett–
for 2 h at 25 1C and transferred into a Teflon-lined stainless Teller specific surface area (SBET) of the catalysts was calculated
steel autoclave (200 ml capacity). It was placed in an air-oven from six adsorption data points in the relative pressure (P/P0)
and aged statically at 80 1C for 24 h. The solid ZrP formed was region of 0.05 r P/P r 0.3 of the isotherm. Pore volume (PV) of
0
filtered, dried overnight at 110 1C, ground into a fine powder the catalyst sample was calculated following the Barrett–Joyner–
and used in further study.
.1.2 Pt/ZrP. ZrP-supported Pt catalysts (Pt = 1 and 2 wt%) catalyst samples was done by transmission electron microscopy
were prepared by wet impregnation method using (TEM) performed on
JEM-F200 multi-purpose electron
Halenda (BJH) method. Microstructural analysis of the
2
a
a
tetraammineplatinum(II) nitrate (99.99%, Sigma-Aldrich) as a microscope (JEOL Ltd) fitted with a 200 kV field emission gun.
Pt source. A required amount of the Pt source was dissolved in 0.001–0.005 g of the sample was suspended in 5 ml of 2-PO. It
10 ml of Milli-Q water. To it, ZrP (1 g, prepared as above) was was sonicated for 30 min at 25 1C and a drop of this
added and the suspension was stirred at 40 1C for 4 h. Then, suspension was placed on a copper grid (200 mesh, ICON
water in it was removed over a rotary evaporator. The solid Analytical) and allowed to dry before it was subjected to the
formed was recovered, dried overnight at 110 1C, and calcined TEM analysis.
ꢀ1
in a muffle furnace at 450 1C for 4 h (ramp rate = 2 1C min ). It
Fourier transform Raman (FT-Raman) spectra were recorded
was reduced in a flow of hydrogen (50 ml min ) at 350 1C for on a Horiba-Jobin-Yvon LabRAM HR 800 spectrometer. The
ꢀ1
ꢀ1
2
.5 h (ramp rate = 10 1C min ). The catalyst, thus obtained, spectra were acquired using a 632 nm (He–Ne) laser (20 mW)
was labeled as xPt/ZrP. Here ‘x’ refers to the nominal Pt content with an acquisition time of 10 s. The FT-Raman spectra were
ꢀ
1
(wt%) in the catalyst.
recorded in the wave number region of 200–1600 cm . Diffuse
2
.1.3 Pt/WO /ZrP. A series of bimetallic Pt/WO /ZrP catalysts reflectance UV-visible (DRUV-vis) spectroscopy measurements
x
3
(Pt = 1 and 2 wt%, and W = 2, 4, 8, 12 and 16 wt%) was prepared by were conducted on a Shimadzu UV-2700 spectrophotometer
a sequential wet impregnation method. Firstly, the W source was equipped with an integrating sphere attachment. The spectra
impregnated on ZrP and then, the Pt source was impregnated. were recorded in the wavelength range of 200 to 800 nm.
In the first-step, while stirring, a required amount of ammonium The Kubelka–Munk (K. M.) function (F(RN)) for infinitely
metatungstate ((NH
4
)
6
H
2
W
12
O
40ꢁnH
2
O – W source, Sigma-Aldrich) thick samples was used to convert reflectance measurements
was dissolved in 10 ml of Milli-Q water taken in a glass round- (RSample) into equivalent absorption spectra using the reflectance
bottom flask (100 ml capacity) placed in a temperature-controlled of BaSO as a reference standard (RStandard) and the following
4
oil bath maintained at 40 1C. To it, 1 g of ZrP powder was added equation:
and the suspension was stirred for 16 h while heating at 40 1C.
Then, water in it was removed over a rotary evaporator. The solid
formed was dried in an oven overnight at 110 1C and calcined at
2
ð1 ꢀ R1Þ
aðabsorption coefficientÞ
Sðscattering coefficientÞ
FðR1Þ ¼
¼
2
R1
5014 | New J. Chem., 2021, 45, 5013ꢀ5022
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
where,
(Fig. 1). In this program, a low power continuous wave (CW)
2
irradiation (presat, pl9) on the solvent (H O) was applied before
RSample
1
R1 ¼
:
the first 301 H pulse (p1) during the relaxation delay (d1).
RStandard
1
A normal method of H NMR analysis (using standard zg30
program) is not suitable as the solvent signal is much more
intense than the spectral features of glycerol and product
2.3 Reaction procedure: hydrogenolysis of glycerol
The hydrogenolysis reaction of glycerol was conducted in a molecules and it thereby, presents a poor quality (low intensity)
00 ml batch stainless steel pressure reactor. In a typical spectrum that is not suitable for analysis and quantification.
1
1
reaction, 0.1 g of glycerol (Sigma-Aldrich) and 3 g of Milli-Q Representative H NMR spectra of an aqueous solution of
water were taken in the reactor. 0.1 g of the catalyst was added glycerol before (c and d) and after the hydrogenolysis reaction
to it. The reactor was sealed, purged with nitrogen to expel out (e and f) over the 1Pt-8WO /ZrP catalyst with normal (a) and
x
the air present inside the reactor and then, pressurized with solvent-suppression (b) pulse sequences are shown in Fig. 1.
1
hydrogen to 40 bar. The temperature was raised to 200 1C and The assignments of the H NMR signals of glycerol and its
the reaction was conducted for 6 to 12 h while stirring products are presented in Table 1. The signals due to 1,2- and
1
(
500 rpm). Later, the reactor was cooled down to 25 1C. The 1,3-PDOs were clearly separated in the H NMR spectrum
gas was vented out and the reaction mixture was collected. enabling quantification of their individual selectivity. The
About 1 ml of Milli-Q water was added additionally to the following parameters were employed in acquiring the
1
reactor vessel and any left out content was taken out and
H NMR spectra: p1 = 13.24 ms (for 901 high power pulse),
combined with the other amount. The catalyst was separated plw1 = 11.00 W, plw9 = 0.00005617 W (only in the case of zgpr),
by centrifugation/filtration. Then, the liquid was passed spectral width (SWH) = 8196.722 Hz, acquisition time (AQ) =
s
through an MS Nylon Syringe Filter (diameter = 33 mm and 3.9976959 s, receiver gain (RG) = 43.4783 number of scans
pore size = 0.22 mm) to eliminate any catalyst particles in it and (NS) = 16 (for zg30) and 32 (for zgpr), dummy scans (DS) = 2,
1
then, analyzed by H NMR spectroscopy. For comparison a display number of points used to define FID (TD) = 65536,
few products were analyzed also by high performance liquid relaxation delay (d1) = 1.0 s, DW = 61.0 ms, and DE = 12.99 ms
chromatography (HPLC). Each reaction and analyses were (for zg30) and 10.16 ms (for zgpr). Product selectivity (mol%) was
1
repeated thrice to confirm the reproducibility.
determined from the areas of the H NMR signals of the
product components. In this calculation the NMR signals of
2.4 Product analysis
glycerol appearing at 3.65–3.77 ppm (due to –CH (OH); 1H;
.4.1 H NMR. To date chromatographic techniques are multiplet), 1,3-PDO at 1.68–1.72 ppm (–CH –; 2H; multiplet),
c
1
2
b
used to monitor the reaction of glycerol hydrogenolysis. We 1,2-PDO at 1.04–1.05 ppm (–CH
d
; 3H; doublet), 1-PO at 0.79–
a
; 3H; triplet) and 2-PO at 1.08–1.09 ppm (–2CH ;
1
report, here, a H NMR spectroscopy method to monitor and 0.82 ppm (–CH
d
1
quantify this reaction. The H NMR method is simple, quick 6H; doublet) were considered (Fig. 1 and Table 1). The follow-
and non-destructive to analyze the products of this reaction. It ing equations were used to estimate the percentage glycerol
is used for quantitative analysis because of the fact that the conversion, distribution of chemical component and compo-
1
intensity of the NMR signal is proportional to the content as nent selectivity from the H NMR spectra. An example of the
well as the number of proton nuclei of the product molecule. calculation for 1Pt-8WO /ZrP is reported in the ESI† (S1).
3
The error in estimation by this method is below 0.5%. The
chromatographic methods require determination of response
Distribution of component Xi ðmol%Þ
Area of the peak of Xi ðnormalized to one protonÞ
factors of each component which the NMR method doesn’t
demand. As the glycerol hydrogenolysis reaction is conducted
in an aqueous medium, the GC method requires a special
column for the analysis. NMR was applied for the quantitative
analysis of glycerol in drug injections and several natural
products such as food products, plant and herbal remedies
¼
ꢂ 100
n
P
Area of peak of Xi ðnormalized to one protonÞ
i¼1
Glycerol conversion (mol%) = 100 ꢀ Percentage distribution of
3
4–36
37
1
glycerol in the reaction mixture
and bio-fluids.
Vila et al. used standard H NMR and Zyu
3
8
et al. employed diffusion-ordered spectroscopy (DOSY), a
pseudo two-dimensional (2D) NMR technique and 1D selective
total correlation spectroscopy (TOCSY) to identify the glycerol
hydrogenolysis products.
Selectivity of product Xi ðmol%Þ
Percentage distribution of Xi in the product
¼
ꢂ 100
Percentage of glycerol conversion
1
In the present study, the H NMR spectra were recorded on a
Bruker AV400 spectrometer. To about 0.025 g of the hydroge-
nolysis product, 0.6 ml of D O (for external locking purpose;
2.4.2 HPLC. A PerkinElmer (Series 200) HPLC equipped
2
Sigma-Aldrich) was added and used in the analysis. A simple with a refractive index detector (PerkinElmer Series 400) was
and robust solvent (H O)-suppression program (zgpr – 1D water used. Product separation was achieved using a REZEX ROA
pre-saturation) that enabled acquisition of high quality (H Organic Acid) column. 0.005 M H
2
+
2
SO
4
was used as the
1
ꢀ1
H NMR spectra of the reactant and products was adopted mobile phase at a flow rate of 0.5 ml min . The column
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 5013ꢀ5022 | 5015

View Article Online
Paper
NJC
1
1
Fig. 1 H NMR pulse programs for (a) standard (zg30) and (b) solvent suppression (zgpr) experiments. Standard H NMR spectra of aqueous glycerol
1
solution: (c) before and (e) after the hydrogenolysis reaction. Solvent suppressed H NMR spectra of the above solution: (d) before and (f) after the
reaction. Reaction conditions: glycerol (0.1 g), water (3 g), catalyst – 1Pt/8WO
00 1C and reaction time = 12 h.
3
/ZrP (0.1 g), hydrogen pressure at 25 1C = 40 bar, reaction temperature =
2
temperature was maintained at 60 1C throughout the HPLC Product identification was carried out by using standard
analysis and a typical analysis run lasted for about 40 min. samples. The retention times of the HPLC peaks of the reactant
1
Table 1 Assignment of H NMR peaks of glycerol and its hydrogenolysis products
Compound
Glycerol
Molecular structure
Proton chemical shift (ppm)
Spectral pattern
H
H
H
a
: 3.54–3.63
: 3.54–3.63
: 3.65–3.77
Doublet of doublet
Doublet of doublet
Multiplet
b
c
H
H
a
: 3.60
: 1.68–1.72
Triplet
Multiplet
b
1
1
1
2
,3-Propanediol (1,3-PDO)
,2-Propanediol (1,2-PDO)
-Propanol (1-PO)
H
H
H
H
a
: 3.44–3.52
: 3.44–3.52
: 3.86
: 1.04–1.05
Doublet of doublet
Doublet of doublet
Multiplet
b
c
d
Doublet
H
H
H
H
a
: 3.59
: 3.59
: 1.59
: 0.79–0.82
Doublet of triplet
Doublet of triplet
Multiplet
b
c
d
Triplet
H
H
a
: 1.08–1.09
: 4.0
Doublet
Septet
b
-Propanol (2-PO)
5016 | New J. Chem., 2021, 45, 5013ꢀ5022
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
and products were as follows: glycerol (16.9 min), ethylene 1Pt/8WO
3
/ZrP and 2Pt/8WO
3
/ZrP and it was smaller for mono-
glycol (EG, 19.9 min), 1,2-propanediol (1,2-PDO, 21.2 min), metallic Pt/ZrP samples.
,3-propanediol (1,3-PDO, 21.6 min), ethanol (26.4 min), 2- The N2 adsorption–desorption isotherms of ZrP samples
propanol (2-PO, 29.0 min) and 1-propanol (1-PO, 32.5 min). (ESI,† S3) were of type II behavior, characteristic of layered
1
4
0
In most cases, the peaks of diols (1,2- and 1,3-PDO) overlapped materials. These isotherms should not be confused with type
and appeared as a partially split peak with maximum at around IV isotherms as they don’t have a plateau at high relative
2
1
1.4 min. Hence, we didn’t quantify separately the selectivity of pressure (P/P
,2- and 1,3-PDO from the HPLC analysis. It is reported only as at relative pressures where multi-layers and capillary condensation
values in the range of 0.2 to
were used to calculate glycerol conversion and product selec- 0.4 depending on the catalyst composition and represent the
0 3
) values. Hysteresis loops (H type) were observed
total diol (1,2- + 1,3-PDO) selectivity. The following formulae occur. These loops closed at P/P
0
tivity (wt%) from the HPLC analysis.
solids with wide distribution of pore sizes. This pseudo-type II
character of the physisorption isotherms corresponded to the
meta-stability of the adsorbed multilayer and delayed capillary
condensation due to a low-degree of pore curvature and non-
Glycerol conversionð%Þ ¼
ðAreaof glycerolpeakatzero time ꢀ Areaof glycerolpeakattimetÞ
rigidity of the aggregate structure. The shape of the H
3
loop was
Areaof glycerolpeakatzerotime
linked to the non-rigid nature of the adsorbent and slit-shaped
ꢂ 100
41
2
ꢀ1
pores. The specific surface area (SBET = 13–29 m g ) and
ꢀ1
pore volume (0.04–0.10 ml g ) of the samples were found to be
low (ESI,† S4) corresponding to the amorphous nature of the
ZrP samples. The average pore radius of the samples was in the
range of 1.7 to 2.1 nm.
Product selectivity ð%Þ
Area of product peak at time t
¼
ꢂ 100
Sum of the areas of all products peaks at time t
DRUV-vis spectroscopy is a useful technique to probe the
electronic structure and domain size of transition metal oxides.
‘
(
‘Neat’’ ZrP showed a weak band in the DRUV-vis spectrum
Fig. 3(a)) with maximum at 278 nm due to ligand-to-metal
charge transfer (LMCT) transition (phosphate oxygen ions to
3
Results and discussion
2
ꢀ
4+
zirconium cations; O - Zr ). This band for Pt–WO /ZrO
2
was reported to occur at 230 nm. Phosphate ions (instead of
3
.1 Catalyst characterization
‘Neat’’ ZrP showed two broad XRD peaks in the 2y ranges of
0–401 and 40–701 asserting its amorphous structure (ESI,†
x
4
2
‘
1
2
ꢀ
neat O ) in the structure could be the reason for the red-shift
of the LMCT band in the ZrP samples. 8WO /ZrP and 2Pt/
WO /ZrP showed absorptions (spread over the spectral range
3
1
3
S2). A tungsten loaded sample (8WO /ZrP) showed additional
3
8
3
sharp peaks at 2y = 23.2, 23.8, 26.6, 28.8, 33.5 and 34.01
of 200–450 nm) that could be deconvoluted into three bands
with peak maximum at 220, 260 and 340 nm (Fig. 3(a)). These
bands were attributed to oxygen (O2p) to tungsten (W5d) charge
transfer transitions in the tungsten oxide species of the
assignable to (002), (020), (120), (112), (022) and (202) planes,
3
9
respectively, of monoclinic (I) g-WO
3
phase (COD 2106382).
2Pt/ZrP exhibited a pattern similar to ‘‘neat’’ ZrP. No peaks
due to metallic Pt were detected implying that Pt crystallites in
this sample are highly dispersed on the ZrP surface and their
crystal size is below the X-ray detection limit. The bimetallic
4
3
44
samples. Guntida et al., in their studies on SiO2 and
Al O -supported tungsten found three types of tungsten species
2
3
3
viz., tetrahedral WO , octahedral polytungstate and crystalline
composition, 2Pt/8WO /ZrP depicted a profile containing the
3
WO species. The two high energy UV bands at 220 and 260 nm
3
3
XRD patterns of both ZrP and g-WO and additional peaks
may be assigned to the tungsten species with distorted tetra-
hedral and octahedral coordination environments, respectively,
(at 2y = 39.6, 46.1, 67.4 and 81.51 due to the (111), (200), (220)
and (311) planes, respectively) of metallic Pt having a cubic
close-packed structure (with space group of Fm3%m; JCPDS No.
65-2868) (ESI,† S2). TEM studies validated the highly dispersed
nature of Pt particles on ZrP (Fig. 2). The average particle size of
Pt (determined from 100 random particles) was 1.5–2.0 nm for
Fig. 3 (a) Diffuse reflectance UV-visible spectra of ZrP, 8WO
3
/ZrP,
/ZrP
/ZrP under normalized conditions. Spectral bands due to
(220 nm), polytungstate clusters (260 nm) and
3
(340 nm) are marked.
2
3 3
Pt/ZrP and 2Pt/8WO /ZrP. (b) Comparative spectra of 2Pt/4WO
and 2Pt/8WO
3
isolated monomeric WO
crystalline WO
3
Fig. 2 TEM images of 1Pt/8WO
3
/ZrP (left) and 2Pt/8WO
3
/ZrP (right).
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 5013ꢀ5022 | 5017

View Article Online
Paper
NJC
while the lower energy band at 340 nm corresponded to the
crystalline WO species. Pt/ZrP showed a broad background
3
absorption in the measured spectral range due to Pt being in
the metallic state.
The surface densities of W atoms for 4 and 8 wt% W catalysts
ꢀ
2
(
calculated using SBET values) were 4.5 and 14.6 W nm ,
4
5
respectively. Barton et al. demonstrated the tungsten density
of 0–4 W nm as the sub-monolayer region (where isolated
ꢀ
2
6
+
four-coordinate
W
centers tetrahedrally coordinated to
oxygen showing the highest absorption edge energy value of
4
4
.89 eV and difficult to reduce with hydrogen species exist),
–8 W nm as the polytungstate growth region (where isolated
ꢀ
2
polytungstate clusters containing tungsten oxide octahedra
bonded through corner and edges, showing intermediate
absorption edge energy values of about 3.54 eV and easier to
ꢀ
2
reduce species exist), and 48 W nm as the polytungstate/
crystalline WO coexistence region (where WO species in an
extended three-dimensional crystalline network of distorted
tungsten oxide octahedra bonded to six neighboring tungsten Fig. 4 FT-Raman spectra of xPt/yWO /ZrP (x = 1 and y = 2, 8, 12 and 16).
3
x
3
oxide octahedra showing lowest absorption edge energies of Raman bands of ZrP and polytungstate/crystalline WO
3
are marked.
2
.59 eV and the easiest to reduce exist). Thus, in the ZrP
ꢀ2
samples with 4 wt% W (surface density = 4.5 W nm ), the
tungsten oxide species are mainly in polytungstate form, and in
the samples with 8 wt% W (surface density = 14.6 W nm ),
polytungstate combined with crystalline WO are the prevalent
ꢀ
2
3
W species. Comparative DR UV-vis spectra under normalized
conditions (Fig. 3(b)) have indeed revealed this conclusion by
exhibiting more intense 260 and 340 nm bands in 2Pt-8WO3/
ZrP than in 2Pt-4WO /ZrP. XRD studies (ESI,† S2) have also
3
authenticated this conclusion showing the characteristic peaks
for the presence of crystalline, monoclinic WO
WO /ZrP and 2Pt-8WO /ZrP samples.
FT-Raman spectroscopy is a valuable technique for unraveling
the molecular structure of metal oxides. ‘‘Neat’’ ZrP showed a
3
species in
8
3
3
ꢀ1
characteristic Raman band at 1055 cm attributable to P–O
2
ꢀ
45
vibrations in the orthophosphate group (HPO4 ). On loading
Pt, a shift in the position of this vibrational band to a higher Fig. 5 FT-Raman spectra of 1Pt/8WO
value (1072 cm ) was observed, indicating strong interaction
of supported Pt particles with the orthophosphate moiety of
3 3
/ZrP and 2Pt/8WO /ZrP.
ꢀ
1
ZrP. A weak, additional vibrational feature was observed at With increasing tungsten amount (from 2 to 16 wt%), a growth
ꢀ1
ꢀ1
2
30 cm for 2Pt/ZrP corresponding to the support. FT-Raman in the intensity of the peaks at 809, 686 and 248 cm
3
spectra of some representative xPt/yWO /ZrP samples are corresponding to W–O stretching, W–O bending and W–O–W
pictured in Fig. 4 and 5. In these samples, additional spectral deformation modes, respectively, of crystalline WO was
3
2
0
features due to the presence of tungsten oxide were observed. observed (Fig. 4). In conformity with DRUV-vis and XRD
The position and intensity of these Raman bands varied results, FT-Raman spectroscopy concludes that the catalyst
with the W content. 1Pt/2WO
3
/ZrP showed an intense band at with 8 wt% W contains both polytungstate and crystalline
ꢀ
1
ꢀ1
1013 cm and a weak band at 428 cm
.
3
WO species, with the former being more abundant than the
These have been assigned to the symmetric stretching mode latter type tungsten oxide species. The samples with 12 and
of terminal WQO and W–O bonds, respectively, present in 16 wt% W contained more proportion of crystalline WO
3
ꢀ
1
isolated, tetrahedral monotungstate and octahedral polytung- species (intense Raman features at 809 and 686 cm ). A shift
state cluster species and at the surface of WO crystals. A shift in the Raman band position to lower wave numbers with
increasing W content was detected confirming the growth in
species. Also Pt content had an
3 3
interaction of WO , withdraws electron density from the ortho- effect on the type of WO species in the bimetallic catalysts
4
5
3
ꢀ1
in the position of the Raman band of the support to 1078 cm
due to interaction of WO with the support was noted. Strong the grain size of the WO
4
6
3
3
phosphate moiety, strengthens the P–O bond and thereby (Fig. 5). While maintaining the W content at 8 wt%, the Pt
leads to a high energy shift in the Raman band position. content was varied from 1 to 2 wt%. It was found from the
5018 | New J. Chem., 2021, 45, 5013ꢀ5022
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
A further reaction of 1,2-PDO gives 1- and 2-propanols (1- and 2-
POs; 4 and 5) and that of 1,3-PDO gives 1-PO (6). On the other
hand, the C–C bond hydrogenolysis of glycerol yields ethylene
glycol (EG) (3), and its further reaction forms ethanol and
methanol (7). All these hydrogenolysis products of glycerol
are liquids under ambient conditions. Reduction of mono
alcohols yields gaseous (C1–C3) hydrocarbon products. The
yield of the liquid products in the reactions over ZrP catalysts
was nearly 98% of the theoretical value suggesting that over-
reduction to hydrocarbons (to gaseous products) was insignificant
at our reaction conditions and hence, they were neglected.
1
Unlike chromatographic techniques, H NMR spectroscopy
10
Scheme 1 C–O and C–C bond hydrogenolysis products of glycerol.
avoids the need of determining response factors of reactant and
1
product molecules. As solvent-water shows a dominant H NMR
signal causing trouble in the product analysis (as the peaks of
the products appear less intense than the solvent peak), we
used one-dimensional, solvent-suppression H pulse-program
in the analysis (Section 2.4.2; Fig. 1). Employing this technique,
the solvent signal was saturated and its intensity was sup-
pressed. Then, the peaks of glycerol and the products became
prominent and quantifiable. The spectral features of 1,2- and
FT-Raman spectral intensity (I809/I1071) and band positions that
the relative amount of crystalline WO (as against the poly-
tungstate species) and grain size of tungsten species were
higher in 2Pt/8WO /ZrP than in 1Pt/8WO /ZrP. Thus, the
amounts of Pt and W control the interaction and structure of
metal species present on the ZrP surface.
3
1
3
3
1
,3-PDOs were resolved (see the signals in the regions of 1.04–
3
.2. Catalytic activity
1
.05 ppm for 1,2-PDO and 1.68–1.72 ppm for 1,3-PDO; Fig. 1).
1
Glycerol can undergo hydrogenolysis at its C–O and C–C bond
H NMR enabled determination of their selectivity (Section
positions (Scheme 1; eqn (1)–(7)). While the hydrogenolysis of 2.4.2). The NMR spectroscopy confirmed the formation of 1,2-
the primary C–O bond yields 1,2-propanediol (1,2-PDO; 1), that and 1,3-PDOs and 1- and 2-POs in the reaction. Glycerol
of the secondary C–O bond produces 1,3-propanediol (1,3-PDO; 2). conversion and product selectivity variation trends of the
1
catalysts determined by HPLC and H NMR were nearly the
1
same (Fig. 6) and confirm that H NMR spectroscopy is a
convenient and alternative analytical method to monitor the
glycerol hydrogenolysis reaction.
Control experiments conducted at 200 1C and 40 bar H
2 h in the presence of ‘neat’’ ZrP and 8WO /ZrP showed no
conversion of glycerol. 1Pt/ZrP enabled glycerol conversion of
7.5% confirming that Pt particles are the catalytic centers
Table 2). PDOs (1,3-PDO = 16.7% and 1,2-PDO = 55.6%) and
-PO (22.2%) formed in higher amounts while 2-PO (5.5%) was
2
for
1
3
3
(
1
the minor product. The results reveal that, over this catalyst,
C–O bond hydrogenolysis (1 and 2) is the preferred reaction
pathway rather than C–C bond hydrogenolysis (3). 1-PO is
formed from both 1,2- and 1,3-PDOs (5 and 6), whereas 2-PO
is formed only from 1,2-PDO (Scheme 1; reaction 4). The
conversion of glycerol was higher (69.2%) when 2Pt/ZrP was
used as the catalyst. However, the intrinsic catalytic activity
Fig. 6 Correlation plot of glycerol conversion and PDO selectivity
obtained from HPLC and H NMR analyses. The R values for glycerol
1
2
conversion and PDO selectivity were 0.895 and 0.842, respectively.
a
Table 2 Hydrogenolysis of glycerol over mono and bimetallic ZrP-supported Pt catalysts
1
Product selectivity (%, H NMR)
1
; 3-PDO
; 2-PDO
ꢀ
1
Catalyst
Glycerol conversion (%)
1,3-PDO
1,2-PDO
1-PO
2-PO
TOF (h )
1
1
2
1
2
Pt/ZrP
Pt/ZrP
Pt/8WO
Pt/8WO
37.5
69.2
92.3
68.4
16.7
11.1
20.9
23.1
55.6
66.7
25.0
38.5
22.2
22.2
47.2
34.6
5.5
0
6.9
3.8
0.3
0.2
0.8
0.6
7
6
16
6
3
/ZrP
/ZrP
3
a
Reaction conditions: Glycerol = 0.1 g, water = 3 g, catalyst = 0.1 g, hydrogen pressure = 40 bar, reaction temperature = 200 1C and reaction time =
1
2 h. 1,2-PDO = 1,2-propanediol, 1,3-PDO = 1,3-propanediol, 1-PO = 1-propanol and 2-PO = 2-propanol. Turnover frequency (TOF) = moles of
glycerol converted per mole of Pt per hour.
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 5013ꢀ5022 | 5019

View Article Online
Paper
NJC
Fig. 7 Comparative glycerol hydrogenolysis activity of 1Pt/ZrP and 1Pt/
8
3
WO /ZrP. Reaction conditions: same as in Table 2.
Fig. 8 Variation of TOF as a function of W loading in the hydrogenolysis
of glycerol over Pt/yWO
in Table 3.
3
/ZrP, y = 2–16 wt%. Reaction conditions: same as
(turnover frequency; TOF = moles of glycerol converted per
mole of nominal Pt per hour) was about the same for both structures of Pt and WO
1
3
are the probable cause for the lower
Pt/ZrP and 2Pt/ZrP (Table 2). A higher amount of active Pt with catalytic activity of 2Pt/8WO /ZrP than 1Pt/8WO /ZrP.
3
3
similar dispersion as in 1Pt/ZrP is the reason for the higher
Tungsten content in the bimetallic catalyst had a pro-
glycerol conversion over 2Pt/ZrP. In the reaction over a nounced effect on glycerol conversion and 1,3-PDO selectivity.
bimetallic 1Pt/8WO /ZrP catalyst, a considerable increase in Having found that 1Pt/8WO /ZrP is the active catalyst (TOF =
glycerol conversion (from 37.5% for 1Pt/ZrP to 92.3% for 16 h ), we varied the content of W in the catalyst from 2 to
Pt/8WO /ZrP) and 1,3-PDO selectivity (from 16.7 for 1Pt/ZrP 16 wt% (Table 3). Catalytic activity (TOF) increased with
to 20.9% for 1Pt/8WO
3
3
ꢀ
1
1
3
3
/ZrP) was detected (Fig. 7). TOF of Pt increasing W content up to 8 wt% and then, decreased
ꢀ
1
ꢀ1
increased from 7 h (for 1Pt/ZrP) to 16 h (for 1Pt/8WO
3
/ZrP), gradually above that. Fig. 8 demonstrates a volcano-shape
inferring synergistic interaction between Pt and W species. The variation of catalytic activity (TOF) with W content in the
reaction to 1-PO was favored. The selectivity of total PDOs catalyst. 1Pt/8WO /ZrP was found to be the most active catalyst
3
decreased from 72.3 to 45.9% and the selectivity of 1-PO with high selectivity to 1,3-PDO (Table 3). Highly dispersed
increased from 22.2 to 47.2% (Table 2). Extent of over- nanoparticles of Pt in contact with polytungstate-like W-species
hydrogenolysis to 1-PO was less prevalent on 2Pt/8WO
than on 1Pt/8WO /ZrP. FT-Raman study revealed a higher volcano-shape variations were reported also by others.
amount of crystalline WO species (more intense 809 and WO acts as a promoter to the metal. Pt-(WO -H with dis-
/ZrP than in 1Pt/8WO /ZrP. The persed Pt metal and strong Brønsted acid adjacent to each
latter catalyst contained a higher proportion of polytungstate other is the active catalytic center. Bhanuchander et al.
3 3
/ZrP are responsible for the highest activity of 1Pt/8WO /ZrP. Such
1
5,47–49
3
3
3
x n
)
ꢀ
1
6
86 cm bands) in 2Pt/8WO
3
3
2
0
31
species. These differences in the WO structures must be the using several metal phosphate (including ZrP)-supported Pt
3
cause for the difference in the hydrogenolysis selectivity catalysts found 100% glycerol conversion with 97% selectivity
observed over these bimetallic catalysts. The TOF values of to POs at 220 1C and atmospheric pressure. In our study, we
ꢀ
1
1Pt/ZrP and 2Pt/ZrP were nearly the same (7 and 6 h
,
3
show that by promoting with WO , the selectivity to diols
respectively), but 1Pt/8WO
3
/ZrP showed much higher TOF (in particular to 1,3-PDO) could be enhanced while suppressing
ꢀ1
ꢀ1
(16 h ) than 2Pt/8WO
3
/ZrP (6 h ) indicating the influence the C–C bond cleavage and other product formation (Table 2).
of WO
3
structure on the activity of Pt. Furthermore, the particle Although, high conversion of glycerol is achieved, the for-
size of Pt is slightly bigger on 2Pt/8WO /ZrP (2 nm) than on mation rate of 1,3-PDO over 1Pt/8WO /ZrP is lower than the
3
3
1
2,13,15,42
1
Pt/8WO /ZrP (1.5 nm, Fig. 2). These differences in the reported systems.
The reported catalysts contained a
3
a
Table 3 Effect of tungsten composition on the hydrogenolysis of glycerol over Pt/WO
Product selectivity (%)
3
/ZrP catalysts
1
; 3-PDO
; 2-PDO
ꢀ
1
Catalyst
Glycerol conversion (%)
1,3-PDO
1,2-PDO
1-PO
2-PO
TOF (h )
1
1
1
1
1
1
2
Pt/2WO
Pt/4WO
Pt/8WO
Pt/12WO
Pt/16WO
3
3
3
/ZrP
/ZrP
/ZrP
74.5
80.6
86.3
72.7
58.1
100
17.1
12.0
20.8
18.8
18.7
15.4
40.0
32.0
25.0
37.5
34.4
25.6
40.0
48.0
47.2
37.5
40.6
51.3
2.9
8.0
6.9
6.2
6.3
7.7
0.4
0.4
0.8
0.5
0.5
0.6
13
14
15
13
10
9
3
/ZrP
/ZrP
3
3
Pt/4WO /ZrP
a
Reaction conditions: same as in Table 2.
5020 | New J. Chem., 2021, 45, 5013ꢀ5022
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021

View Article Online
NJC
Paper
higher amount of Pt (4–8 wt%) in their composition. To achieve
higher productivity rate of 1,3-PDO, it is essential to enhance
9 Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa and
K. Tomishige, Appl. Catal., B, 2011, 105, 117–127.
the surface area (by altering the synthesis procedure) 10 Y. Nakagawa, X. Ning, Y. Amada and K. Tomishige, Appl.
and strong acidity of the catalyst, and also to suppress the
Catal., A, 2012, 433–434, 128–134.
over-hydrogenolysis activity (by introducing additives), which 11 L. Liu, S. Kawakami, Y. Nakagawa, M. Tamura and K. Tomishige,
will be the study of our future work.
Appl. Catal., B, 2019, 256(Article No. 117775), 1–13.
2 R. Arundhathi, T. Mizugaki, T. Mitsudome, K. Jitsukawa
and K. Kaneda, ChemSusChem, 2013, 6, 1345–1347.
1
1
1
1
1
1
1
1
4
Conclusions
´
´
´˜
3 S. Garcıa-Fernandez, I. Gandarias, Y. Tejido-Nunez,
J. Requies and P. L. Arias, ChemCatChem, 2017, 9, 4508–4519.
4 S. Zhu, X. Gao, Y. Zhu and Y. Li, J. Mol. Catal. A: Chem.,
3
Bimetallic Pt/WO /ZrP catalysts (prepared by the sequential wet
impregnation method) exhibited high catalytic performance for
the hydrogenolysis of glycerol in an aqueous medium. They
gave mainly PDOs rather than POs at moderate temperatures
2015, 398, 391–398.
5 N. Lei, X. Zhao, B. Hou, M. Yang, M. Zhou, F. Liu, A. Wang
and T. Zhang, ChemCatChem, 2019, 11, 3903–3912.
6 S. Zhu, X. Gao, Y. Zhu, J. Cui, H. Zheng and Y. Li, Appl.
Catal., B, 2014, 158–159, 391–399.
7 S. Zhu, X. Gao, Y. Zhu, Y. Zhu, X. Xiang, C. Hu and Y. Li,
Appl. Catal., B, 2013, 140–141, 60–67.
8 N. Lei, Z. Miao, F. Liu, H. Wang, X. Pan, A. Wang and
T. Zhang, Chin. J. Catal., 2020, 41, 1261–1267.
9 S. S. Priya, V. P. Kumar, M. L. Kantam, S. K. Bhargava,
A. Srikanth and K. V. R. Chary, Ind. Eng. Chem. Res., 2015,
(
200 1C). Co-presence of Pt and W in the catalyst composition
led to synergistic enhancement in the catalytic activity and
,3-PDO selectivity. WO promoted the product selectivity of Pt
1
3
to diols (in particular to 1,3-PDO) rather than to other products.
A volcano-shape variation of catalytic activity with the W content
was found. A ZrP catalyst with 1 wt% Pt and 8 wt% W showed
the best selective catalytic performance. Highly dispersed Pt
x
nanoparticles in contact with polytungstate-like WO species
1
were active centers for the hydrogenolysis reaction. H NMR
spectroscopy was demonstrated as an alternative tool to
chromatography to quantify the hydrogenolysis products.
54, 9104–9115.
2
0 W. Zhou, Y. Li, X. Wang, D. Yao, Y. Wang, S. Huang, W. Li,
Y. Zhao, S. Wang and X. Ma, J. Catal., 2020, 388, 154–163.
1 A. Clearfield and D. S. Thakur, Appl. Catal., 1986, 26, 1–26.
2 M. Pica, Catalysts, 2017, 7(Article No. 190), 1–18.
3 Y. Kamiya, S. Sakata, Y. Yoshinaga, R. Ohnishi and
T. Okuhara, Catal. Lett., 2004, 94, 45–47.
2
2
2
Conflicts of interest
There are no conflicts to declare.
2
2
2
2
2
4 N. Li, G. A. Tompsett and G. W. Huber, ChemSusChem, 2010,
3
, 1154–1157.
5 Y. T. Kim, J. A. Dumesic and G. W. Huber, J. Catal., 2013,
04, 72–85.
Acknowledgements
3
S. D. acknowledges the Council of Scientific and Industrial
Research (CSIR), New Delhi for the award of a CSIR-Bhatnagar
Fellowship (SSB 000926; P90807). The authors thank Dr Mrs
Banu Marimuthu, Mr Sreejith Sreekantan, Ms Arun Arunima
B. Kirali and Dr Uday Kiran Marelli for the support.
6 G.-H. Han, M. W. Lee, S. Park, H. J. Kim, J.-P. Ahn and M.-
G. Seo, J. Catal., 2019, 377, 343–357.
7 C. Zhu, Q. Liu, D. Li, H. Wang, C. Zhang, C. Cui, L. Chen,
C. Cai and L. Ma, ACS Omega, 2018, 3, 7407–7417.
8 Q.-Y. Liu, Y.-H. Liao, T.-J. Wang, C.-L. Cai, Q. Zhang,
N. Tsubaki and L.-L. Ma, Ind. Eng. Chem. Res., 2014, 53,
1
2655–12664.
Notes and references
2
9 G. Li, N. Li, Z. Wang, C. Li, A. Wang, X. Wang, Y. Cong and
1
2
3
4
5
C.-H. Zhou, J. N. Beltramini, Y.-X. Fana and G. Q. Lu, Chem.
Soc. Rev., 2008, 37, 527–549.
J. ten Dam and U. Hanefeld, ChemSusChem, 2011, 4,
T. Zhang, ChemSusChem, 2012, 5, 1958–1966.
30 M. Wang, H. Yang, Y. Xie, X. Wu, C. Chen, W. Ma, Q. Dong
and Z. Hou, RSC Adv., 2016, 6, 29769–29778.
31 P. Bhanuchander, S. S. Priya, V. P. Kumar, S. Hussain,
N. P. Rajan, S. K. Bhargava and K. V. R. Chary, Catal. Lett.,
2017, 147, 845–855.
1
017–1034.
D. Sun, Y. Yamada, S. Sato and W. Ueda, Appl. Catal., B,
016, 193, 75–92.
2
M. R. Nanda, Z. Yuan, W. Qin and C. Xu, Catal. Rev.: Sci. 32 R. Luo, X. Zhao, H. Gong, W. Qian, D. Li, M. Chen, K. Cui,
Eng., 2016, 58, 309–336.
J. Wang and Z. Hou, Energy Fuels, 2020, 34, 8707–8717.
M. R. Nimlos, S. J. Blanksby, X. Qian, M. E. Himmel and 33 T.-Z. Ren, Z.-Y. Yuan and B.-L. Su, Chem. Commun., 2004,
D. K. Johnson, J. Phys. Chem. A, 2006, 110, 6145–6156.
2730–2731.
J. Oh, S. Dash and H. Lee, Green Chem., 2011, 13, 2004–2007. 34 J. Lu, P. Wang, Q. Wang, Y. Wang and M. Jiang, Molecules,
6
7
S. Bhowmik and S. Darbha, Catal. Rev.: Sci. Eng., 2020, DOI:
0.1080/01614940.2020.1794737.
Y. Nakagawa, Y. Shinmi, S. Koso and K. Tomishige, J. Catal.,
010, 272, 191–194.
2018, 23, 1177.
1
35 M. Hohmann, V. Koospal, C. Bauer-Christoph, N. Christoph,
H. Wachter, B. Diehl and U. Holzgrabe, J. Agric. Food Chem.,
2015, 63, 4112–4119.
8
2
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021
New J. Chem., 2021, 45, 5013ꢀ5022 | 5021

View Article Online
Paper
NJC
3
3
3
6 C. Simmler, J. G. Napolitano, J. B. McAlpine, S. N. Chen and 43 K. Narasimha Rao, A. Sridhar, A. F. Lee, S. J. Tavener,
G. F. Pauli, Curr. Opin. Biotechnol., 2014, 25, 51–59. N. A. Young and K. Wilson, Green Chem., 2006, 8, 790–797.
7 F. Vila, M. L. Granados and R. Mariscal, Catal. Sci. Technol., 44 A. Guntida, K. Suriye, J. Panpranot and P. Praserthdam,
017, 7, 3119–3127. Catal. Today, 2020, 358, 354–369.
8 Z. Lyu, F. Yue, X. Yan, J. Shan, D. Xiang, C. M. Pedersen, 45 D. G. Barton, M. Shtein, R. D. Wilson, S. L. Soled and
2
C. Li, Y. Wang and Y. Qiao, Fuel Proc. Technol., 2018, 171,
17–123.
9 K. Thummavichai, N. Wang, F. Xu, G. Rance, Y. Xia and
Y. Zhu, Roy. Soc. Open Sci., 2018, 5(Art No. 171932), 1–10.
0 R. M. Barrer, Pure Appl. Chem., 1989, 61, 1903–1912.
E. Iglesia, J. Phys. Chem. B, 1999, 103, 630–640.
46 S. E. Hobsley, D. V. Nowell and D. T. Stewart, Spectrochim.
Acta, 1974, 30A, 535–541.
47 T. Aihara, H. Miura and T. Shishido, Catal. Today, 2020, 352,
73–79.
1
3
4
4
1 K. S. W. Sing and R. T. Williams, Adsorption Sci. Technol., 48 S. Feng, A. Nagao, T. Aihara, H. Miura and T. Shishido,
004, 22, 773–782. Catal. Today, 2018, 303, 207–212.
2 Y. Fan, S. Cheng, H. Wang, J. Tian, S. Xie, Y. Pei, M. Qiao 49 K. Tomishige, Y. Nakagawa and M. Tamura, Green Chem.,
and B. Zong, Appl. Catal., B, 2017, 217, 331–341. 2017, 19, 2876–2924.
2
4
5022 | New J. Chem., 2021, 45, 5013ꢀ5022
This journal is The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2021