Hydrogenation of levulinic acid to valeric acid over platinum-tungsten catalysts supported on γ-Al2O3

Article abstract of DOI:10.1039/c9nj04056k

To address the challenges associated with fossil fuel depletion, exploration of the sustainable energy resources is one of the most pursued research areas in this century. Valeric acid (VA) is considered as a valuable platform molecule for the production of biofuels and chemical intermediates. However, the production of VA from levulinic acid (LA) has become one of the most desirable reactions. The aim of this work was to investigate the amount of acidity, Pt loading and experimental conditions for the hydrogenation of LA into VA operated at 0.1 MPa H₂ pressure and to elucidate the relationship between the active sites and the catalytic performance. The Pt-WO₃ catalysts supported on γ-Al₂O₃ were synthesized by a simple wet impregnation method with various Pt loadings from 0.5 to 3 wt% with constant 10 wt% of WO₃ and calcination at 500 °C for 4 h under air. The catalysts were characterized by powder X-ray diffraction (XRD), scanning electron microscopy-energy dispersive X-ray spectroscopy (SEM-EDS), N₂ adsorption-desorption, temperature programmed desorption of ammonia (NH₃-TPD), pyridine Fourier transform infrared spectroscopy (Pyr.FT-IR) and CO-chemisorption studies. Under optimized reaction conditions, the 2Pt-10WO₃/γ-Al₂O₃ catalyst exhibited the highest selectivity to VA (58%) with 91% conversion of LA. This is due to the availability of a huge number of acidic and Pt active sites on the catalyst surface. In addition, the catalytic activity, reaction parameters and stability of the catalyst are demonstrated clearly.

Full text of DOI:10.1039/c9nj04056k

NJC
View Article Online
View Journal | View Issue
PAPER
Hydrogenation of levulinic acid to valeric acid
over platinum–tungsten catalysts supported
on c-Al2O3
Cite this: New J. Chem., 2019,
43, 18003
Ponnala Bhanuchander,a Shanthi Priya Samudrala, *b Balla Putrakumar,c
Perupogu Vijayanand,d Beepala Sateesh Kumara and Komandur V. R. Chary*a
To address the challenges associated with fossil fuel depletion, exploration of the sustainable energy
resources is one of the most pursued research areas in this century. Valeric acid (VA) is considered as a
valuable platform molecule for the production of biofuels and chemical intermediates. However, the
production of VA from levulinic acid (LA) has become one of the most desirable reactions. The aim of
this work was to investigate the amount of acidity, Pt loading and experimental conditions for the
hydrogenation of LA into VA operated at 0.1 MPa H2 pressure and to elucidate the relationship between
the active sites and the catalytic performance. The Pt–WO3 catalysts supported on g-Al2O3 were synthesized
by a simple wet impregnation method with various Pt loadings from 0.5 to 3 wt% with constant 10 wt% of
WO3 and calcination at 500 1C for 4 h under air. The catalysts were characterized by powder X-ray diffraction
(XRD), scanning electron microscopy-energy dispersive X-ray spectroscopy (SEM-EDS), N2 adsorption–
desorption, temperature programmed desorption of ammonia (NH3-TPD), pyridine Fourier transform infrared
spectroscopy (Pyr.FT-IR) and CO-chemisorption studies. Under optimized reaction conditions, the
2Pt–10WO3/g-Al2O3 catalyst exhibited the highest selectivity to VA (58%) with 91% conversion of LA.
This is due to the availability of a huge number of acidic and Pt active sites on the catalyst surface. In
addition, the catalytic activity, reaction parameters and stability of the catalyst are demonstrated clearly.
Received 5th August 2019,
Accepted 21st October 2019
DOI: 10.1039/c9nj04056k
rsc.li/njc
Therefore, the synthesis of valeric acid from levulinic acid has
become one of the most desirable reactions during the past
1. Introduction
To address the challenges associated with fossil fuel depletion, decade.
exploration of sustainable energy resources is one of the most
In a pioneering research study,17 solid base catalysts were
pursued research areas in this century.1–6 Biomass technology found to be the most active for the hydrogenation of LA to GVL.
has become unique among the various available sustainable Nevertheless, limited reports exist for the direct conversion of
energy strategies such as solar, wind, geothermal, etc. Levulinic LA to VA. Lange et al.18 studied the multi-step synthesis of
acid (LA) is recognized by the U.S. Department of Energy as one valeric acid from LA using a Pt/TiO2 catalyst. Weckhuysen
of the most valuable chemicals from lignocellulose and its et al.16 explored the Ru catalyst on supported zeolite for the
derivatives are important biomass building block molecules conversion of levulinic acid to valeric acid and obtained 45.8%
to obtain products ranging from cosmetics to fuel additives.7–11 yield by employing dioxane as solvent at 200 1C and 40 bar H2
Valeric acid (VA) is considered as a valuable platform molecule pressure. Kon et al.19 explored the catalytic hydrogenation of LA
to produce biofuels and chemical intermediates.12–16 Moreover, VA to VA over a Pt/HMFl catalyst at 200 1C and 8 bar H2 pressure.
is considered as a potential solvent for biomass conversion. Fu et al.20 reported levulinic acid hydrogenation on a Ru/SBA–
SO3H catalyst in ethanol that resulted in the co-production
of valeric acid and ethyl valerate with 94% yield at 240 1C and
40 bar H2 pressure. Although the catalytic systems demonstrate
direct VA synthesis, the present procedures are neither competent
nor cost-effective.
Noble metal-based catalysts have attracted significant considera-
tion due to their wide applications as catalysts in reactions such as
hydrogenation, dehydrogenation and oxidation. The supported
noble metal catalysts (e.g., Ru, Rh, Pd, Pt and Au) are the most
a Catalysis and Fine Chemicals Division, CSIR-Indian Institute of Chemical
Technology, Uppal Road, Hyderabad 500007, India. E-mail: kvrchary@iict.res.in,
charykvr@gmail.com
b Department of Chemical Engineering, Monash University, Melbourne, 3800,
Australia. E-mail: priya.shanthipriya@monash.edu
c State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics,
Chinese Academy of Sciences, Dalian, China
d Centre for Environmental Engineering & Fossil Fuels Division, CSIR-Indian
Institute of Chemical Technology, Uppal Road, Hyderabad 500007, India
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 18003--18011 | 18003

View Article Online
Paper
NJC
Scheme 1 Selective hydrogenation of LA to VA over Pt–WO3/g-Al2O3.
Fig. 1 XRD patterns of 10WO3/g-Al2O3 and various Pt-10WO3/g-Al2O3
catalysts.
commonly used catalysts to produce GVL from LA in the
continuous and batch reactors. Several results indicate that
noble metal-based catalysts have shown higher activity for
hydrogenation of LA to GVL in both homogeneous and hetero-
geneous systems. But, the enormously high cost and rare
resources limit their large-scale applications. Hence, it is necessary
to expand the exploitation of noble metals. Various efforts have
been devoted to improving the catalytic efficiency by dispersing
noble metal catalysts onto an appropriate support with high
surface area, such as metallic oxides. Among the various support
materials, Al2O3 based materials have received significant research
interest due to their decent electronic properties, good surface area
and excellent thermal stability.
The aim of this work was to investigate the impact of acidity,
Pt loading and experimental conditions on LA conversion into VA
over 2Pt–10WO3/g-Al2O3 at 0.1 MPa H2 pressure and to elucidate
the relationship between the active sites and catalytic performance.
Characterization techniques including N2 adsorption–desorption
isotherms, temperature-programmed desorption of NH3
(NH3-TPD), powder X-ray diffraction (XRD), UV-diffuse reflectance
spectra, scanning electron microscopy-energy dispersive X-ray
spectroscopy (SEM-EDS), CHNS analysis and transmission electron
microscopy were employed to comprehend the development of the
catalyst structure and the physiochemical properties. New insights
into the parameters that govern the catalytic activity, selectivity and
stability are highlighted (Scheme 1).
(220) lattice planes of Pt (ref. pattern. 00-001-1311).21 These peaks
are gradually sharped from sample 0.5Pt–10WO3/Al to 3Pt–
10WO3/Al, revealing that an increase in the Pt amount
increased the Pt particle size.
2.1.2 CO-chemisorptions. The platinum dispersion, metal
surface area and average particle size of XPt–10WO3/g-Al2O3
catalysts with different platinum loadings (0.5–3 wt%) were
measured using irreversible CO-chemisorption and the results
are presented in Table 1. These results demonstrated that up to
2 wt% of platinum content, the CO uptake value was increased
(from 4.25 to 5.79) and with a further increase of Pt content
beyond 2 wt%, the CO-uptake value was decreased (3.25).
The metal surface area of the 10WO3/g-Al2O3 catalyst with
different platinum loadings was found to be in the range of
0.28–0.40 m2 gꢀ1 whereas the particle size varied between
8.20 and 22.95 nm. With the increase of platinum loading from
0.5 to 3 wt% on the WO3/g-Al2O3 catalyst a decrease in platinum
dispersion was observed. These effects are probably due to the
agglomeration of platinum particles over the support. The
introduction of tungstate particles into the catalyst led to a
decrease in the dispersion of Pt.
2.1.3 BET surface area and pore size distribution studies.
The BET surface area, pore volume and average pore diameter
of pure g-Al2O3, 10WO3/g-Al2O3 and XPt–10WO3/g-Al2O3 catalysts
with different platinum loadings are presented in Table 2. The
pure g-Al2O3 and 10WO3/g-Al2O3 samples had a BET surface area
of 235 m2 gꢀ1 and 205 m2 gꢀ1 respectively. After impregnation of
WO3 and Pt on the support, the BET surface area of Pt with
2. Results and discussion
2.1 Catalyst characterization
2.1.1 X-ray diffraction (XRD) analysis. Fig. 1 presents the
XRD patterns of Pt–10WO3/Al2O3 catalysts with various Pt loadings
(0.5 to 3 wt%). The XRD pattern of 10WO3/Al2O3 exhibited three
broad diffraction peaks at 39.51, 45.81 and 66.71, which correspond
to the (222), (400) and (440) cubic crystalline phase of Al2O3 (ref.
pattern. 00-001-1303), and did not show any diffraction peaks due
to WO3, indicating that tungsten species mainly existed in the
amorphous structure on the g-Al2O3 support. After Pt loading, XRD
patterns showed three sharp diffraction peaks at around 39.71,
Table 1 Results of dispersion, CO uptake, metal area, and average particle
size of 2Pt–10WO3/g-Al2O3 catalysts
Catalyst
(a)
(b)
(c)
(d)
8.20
14.25
18.29
22.95
0.5Pt–10WO3/g-Al2O3
1Pt–10WO3/g-Al2O3
2Pt–10WO3/g-Al2O3
3Pt–10WO3/g-Al2O3
11.37
10.07
8.22
4.25
5.60
5.79
3.25
0.32
0.38
0.40
0.28
6.74
a – dispersion (%), b – CO uptake (mmol gꢀ1), c – metal surface area
cat, d – particle size (nm).
46.31 and 67.31, which can be attributed to the (111), (200) and (m2 gꢀ1
)
18004 | New J. Chem., 2019, 43, 18003--18011 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
Table 2 Physicochemical properties g-Al2O3 and g-Al2O3 supported
catalysts
Catalyst
(a)
(b)
(c)
Pure g-Al2O3
10WO3/g-Al2O3
0.5Pt–10WO3/g-Al2O3
1Pt–10WO3/g-Al2O3
2Pt–10WO3/g-Al2O3
3Pt–10WO3/g-Al2O3
235
205
195
172
167
153
6.2
5.2
4.6
4.3
4.0
3.8
0.45
0.32
0.23
0.18
0.16
0.12
a – BET surface area (m2 gꢀ1), b – pore diameter (nm), c – pore volume
(cc gꢀ1).
various loadings (0.5 to 3 wt%) decreased from 195 to 153 m2 gꢀ1
as compared to pure g-Al2O3 which could be due to the blocking of
g-Al2O3 pores by the deposition of Pt and tungstate species. The
results are shown in Fig. 2, and it was observed that the pore volume
and pore diameter of g-Al2O3 supported catalysts decreased from
0.32 to 0.12 cc gꢀ1 and from 5.2 to 3.8 nm respectively. These results
are in accordance with the previous reports.22
N2 adsorption–desorption measurement results are presented
in Fig. 2. The isotherm of 10WO3/g-Al2O3 shows a sharp uptake at
low relative pressure (P/P0 o 0.015) but a hysteresis loop at high
relative pressure (P/P0 4 0.45), which confirms the existence of
abundant micropore and meso/macropore structures in the
materials. The Brunauer–Emmett–Teller (BET) surface area of
10WO3/g-Al2O3 was found to be 190.7 m2 gꢀ1 with a pore volume
of 0.238 cc gꢀ1 and an average pore diameter of 53.2 Å.
2.1.4 SEM-EDX. The elemental mapping (Al, O, W and Pt
elements) and the energy dispersive X-ray spectroscopy (EDX)
analysis of the 2Pt–10WO3/g-Al2O3 catalyst have been conducted
to further explore the composition and elemental distribution of
the catalysts. EDX elemental mapping images indicate the
existence of Al, W, O and Pt elements on the surfaces of the
2Pt–10WO3/Al2O3 catalyst. Al and O elements are extremely
homogeneously distributed along with the catalyst, as observed
Fig. 3 (a) SEM and (b) EDX results and the corresponding elemental
mapping images of 2Pt–10WO3/g-Al2O3 catalysts.
in Fig. 3. W and Pt mapping images reveal separated and well
distributed W and Pt particles over the support. These results
demonstrate that W and Pt have been successfully and uniformly
incorporated into the alumina framework.
2.1.5 Transmission electron microscopy. The morphology
of the as-synthesized materials was investigated by transmission
electron microscopy (TEM). The TEM image of 10WO3/g-Al2O3
reveals that the composite material has a rod-like structure with
different sizes. After impregnation of platinum the rod became
smaller and wider. The results are shown in Fig. 4.
2.1.6 UV-visible diffusion reflectance spectra. The UV–vis
diffused reflectance spectra of the various Pt–10WO3/g-Al2O3
catalysts were recorded in the range of 200 to 600 nm and the
results are shown in Fig. 5. The UV-vis diffused reflectance
spectra of the 2Pt–10WO3/g-Al2O3 sample show one intensive
absorption band at 211.6 nm and two weak and broad absorption
bands at 270.3 and 374.4 nm. The main absorption band
corresponds to the ligand to metal charge transfer in isolated
Fig. 2 BET isotherms of 10WO3/g-Al2O3 (inset figure) and various 10WO3/
g-Al2O3 supported platinum catalysts.
Fig. 4 TEM images of (a) 10WO3/g-Al2O3 and (b) 2Pt-10WO3/g-Al2O3
catalysts.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 18003--18011 | 18005

View Article Online
Paper
NJC
Fig. 5 UV-DRS patterns of the various 10WO3/g-Al2O3 supported platinum
catalysts. (a) 0.5Pt–10WO3/g-Al2O3, (b) 1Pt–10WO3/g-Al2O3, (c) 2Pt–10WO3/
g-Al2O3 and (d) 3Pt–10WO3/g-Al2O3.
tetrahedral species. The weak charge transfer bands around
270 nm and 375 nm are due to the charge transfer of O2ꢀ to W6+ in
an octahedrally coordinated tungsten oxide species and the
presence of bulk WO3 crystallites respectively.23 With the increase
in Pt loading, the intensity of these charge transfer bands gradually
weakened and slightly shifted to a higher wavelength range.
2.1.7 Temperature programmed reduction. The temperature
programmed reduction (TPR) pattern of XPt–10WO3/g-Al2O3
catalysts with various platinum loadings (0.5 to 3 wt%) is shown
in Fig. 6 and the H2 uptake values are displayed in Table 3. All the
catalysts have shown the H2-TPR profile within the temperature
range of 50–800 1C. The bulk WO3 showed three main peaks in
the H2-TPR profile with maxima at 630, 790, and 910 1C in
accordance with the previous reports.24,25
Fig. 6 H2 TPR profiles of various 2Pt–WO3/g-Al2O3 catalysts. (a) g-Al2O3,
(b) 10WO3/g-Al2O3, (c) 0.5Pt–10WO3/g-Al2O3, (d) 1Pt–10WO3/g-Al2O3,
(e) 2Pt–10WO3/g-Al2O3 and (f) 3Pt–10WO3/g-Al2O3.
Table
3
Results of temperature programmed reduction of various
Pt–WO3/g-Al2O3 catalysts
Pt loading (wt%)
(a)
(b)
(c)
(d)
67.36
93.6
124.9
94.7
(e)
99.86
177.4
195.3
141.3
0.5Pt–10WO3/g-Al2O3
1Pt–10WO3/g-Al2O3
2Pt–10WO3/g-Al2O3
3Pt–10WO3/g-Al2O3
340
338
348
345
32.5
83.8
70.4
56.6
643
630
680
665
a and c – Tmax, b and d – H2 uptake (mmol gꢀ1), e – total H2 uptake (mmol gꢀ1).
The TPR analysis of various Pt–10WO3/g-Al2O3 samples
exhibited reduction peaks at 640 1C and 790 1C which indicate
the presence of WO3 species on the g-Al2O3 support.26,27 The
reduction peak at 640 1C is associated with the partial reduction
of WO3 to tungsten-oxide (WO2.9), whereas the simultaneous
reduction of WO2.9 oxide and of the remaining WO3 to WO2
phase occurs at approximately 790 1C. Another reduction peak
at 355 1C indicates reduction of Pt to the metallic phase (Pt2+ to
Pt0) for the Pt–WO3 catalyst supported on g-Al2O3.28 Therefore,
the reduction peaks observed must be due to the reduction of
Pt and WO3 only and the results are presented in Fig. 6.
2.1.8 Temperature programmed desorption of ammonia.
The TPD of ammonia is used to determine the strength of
acidic sites on a catalyst surface. The total acidity of the catalyst
and the results obtained are presented in Table 4. In general,
Table 4 Effect of WO3 loading on hydrogenation of levulinic acid to
valeric acid
Selectivity (%)
WO3 loading (%)
(a)
(b)
(c)
(d)
(e)
(f)
2Pt–5WO3/g-Al2O3
2Pt–10WO3/g-Al2O3
2Pt–15WO3/g-Al2O3
2Pt–20WO3/g-Al2O3
85
91
88
77
55
27
22
19
41
58
49
47
2
10
19
22
2
5
10
12
170
190
185
170
a – LA conversion (%), b – GVL, c – VA, d – AL, e – others, f – total acidity
(mmol gꢀ1). Reaction conditions: pure levulinic acid, catalyst weight
0.5 g, 2Pt–YWO3/g-Al2O3 catalyst (Y = 5, 10, 15 and 20 wt%), reaction
temperature 270 1C, WHSV 1.14 hꢀ1 (with respect to LA), and H2 flow
60 mL minꢀ1. GVL: g-valerolactone, VA: valeric acid, AL: angelica
lactone, others includes MTHF, C4-alcohols, and C5-alcohols.
the NH3-TPD peaks are classified into three different acid sites sample shows two low-temperature peaks in the range of
based on their strength: weak acid sites at 150–300 1C, medium 70–120 1C and 180–300 1C and a high-temperature broad peak
acid sites at 300–400 1C and strong acid sites at 450–650 1C.29 in the range of 460–640 1C. The low temperature peaks are
Fig. 7 presents NH3-TPD profiles showing the presence of only ascribed to physisorbed and weakly bound ammonia, whereas
two types of peaks, which are at low and high temperatures, over the the high temperature peak corresponds to ammonia adsorbed
catalyst surface. The NH3-TPD curve of the Pt-WO3/g-Al2O3 on the strong acidic sites, as shown in Fig. 7.
18006 | New J. Chem., 2019, 43, 18003--18011 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
Fig. 7 NH3-TPD profiles of various 2Pt–XWO3/g-Al2O3 catalysts: (a) 2Pt–5WO3/
g-Al2O3, (b) 2Pt–10WO3/g-Al2O3, (c) 2Pt–15WO3/g-Al2O3 and (d) 2Pt–20WO3/
g-Al2O3.
Fig. 8 Pyridine adsorbed FT-IR patterns of WO3/Al2O3 and various
Pt–WO3/Al2O3 catalysts. (a) 10WO3/g-Al2O3, (b) 0.5Pt–10WO3/g-Al2O3,
(c) 1Pt–10WO3/g-Al2O3, (d) 2Pt–10WO3/g-Al2O3 and (e) 3Pt–10WO3/g-Al2O3.
This result suggests that the introduction of WO3 led to the
increased acid strength of the catalysts. The acidity of the catalyst
mainly depends on the amount of the active phase of 2Pt–10WO3
present on alumina. Further, with an increase in the WO3 loading
the total amount of acid sites increased steadily, reached a max-
imum over 2Pt–10WO3/g-Al2O3, and then decreased marginally.
2.1.9 Pyridine adsorbed FTIR. As NH3-TPD is a useful
technique to determine the acidic strength of catalysts, the pyridine
adsorbed FTIR technique is another such method used to find the
nature of acidic sites present in the catalyst. The pyridine adsorbed
FTIR spectra of WO3/g-Al2O3 and that of Pt–WO3/g-Al2O3 catalysts
with 0.5–3 wt% loading are illustrated in Fig. 8. All bands contained
in the spectra at 1551 cmꢀ1 are due to the presence of pyridinium
ions on Brønsted acid sites,30,31 at 1455 cmꢀ1 characteristic of Lewis
sites and at 1490 cmꢀ1 associated with both Brønsted and Lewis
sites. It can be noticed from Fig. 8 that all the catalysts contain
Lewis sites, Brønsted sites and Lewis–Brønsted sites in various
proportions depending upon the loading of WO3 on the support.
conversion decreased and the selectivity of VA decreased to
47%. This indicates that 10 wt% WO3 loading was found to be
the optimized loading for selective conversion of LA to VA.
These results are in good conformity with NH3-TPD studies.
2.2.2 Effect of platinum loading. The influence of platinum
loading on the hydrogenation of LA to VA over Pt–WO3/g-Al2O3
catalysts was studied. As the Pt content increases from 0.5 to
2 wt%, the LA conversion also increases from 75% to 91%. As
shown in Table 5, the selectivity of VA increased with the
increase of Pt loading and reached the highest (58%) at 2 wt%
loading. However, a further increase in Pt loading to 3 wt% led
to a slight decrease in the selectivity of VA and the results are
presented in Fig. 9. This may be due to agglomeration of
platinum particles on the support which blocked the acid sites
and reduced the dispersion of Pt as evidenced by XRD, TEM and
elemental analysis results. The catalytic activity results are in good
correlation with the results obtained from acidity measurements
2.2 Catalytic activity
2.2.1 Effect of WO3 loading. The influence of WO3 loading
(5 to 20 wt%) on the conversion of levulinic acid over Pt–WO3/
g-Al2O3 was studied and the results obtained are shown in
Table 4. As the W content increased from 5 to 10 wt%, the
selectivity of GVL decreased and simultaneously the selectivity
of VA and the conversion of levulinic acid increased initially
and decreased later on with increasing WO3 loading beyond
10 wt%. The rise in the selectivity of VA with the inclusion of
WO3 into the catalyst can be endorsed to the generation of
Brønsted acid sites by WO3 species that has facilitated the ring
opening of GVL to form VA.4,9
Table 5 Effect of Pt loading on the hydrogenation of levulinic acid to
valeric acid
Selectivity (%)
Pt loading (wt%)
(a)
(b)
(c)
(d)
(e)
0.5Pt–10WO3/g-Al2O3
1Pt–10WO3/g-Al2O3
2Pt–10WO3/g-Al2O3
3Pt–10WO3/g-Al2O3
75
88
91
85
58
43
27
20
36
45
58
48
2
7
10
17
4
5
5
10
a – LA conversion (%), b – GVL, c – VA, d – AL, e – others. Reaction
conditions: pure levulinic acid, weight of catalyst 0.5 g, XPt–10WO3/
g-Al2O3 catalyst (X = 0.5 to 3.0 wt%), reaction temperature 270 1C, WHSV
1.14 hꢀ1 (with respect to LA), and H2 flow 60 mL minꢀ1. GVL:
g-valerolactone, VA: valeric acid, AL: angelica lactone, others includes
Moreover, the selectivity of VA was enhanced with the
increase of WO3 loading from 5 to 10 wt%. But on further
increasing the WO3 loading to 15 wt% and 20 wt%, the LA MTHF, C4-alcohols, and C5-alcohols.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 18003--18011 | 18007

View Article Online
Paper
NJC
Fig. 10 Effect of WHSV on the hydrogenation of levulinic acid over
2Pt–10WO3/g-Al2O3 catalysts. Reaction conditions: pure levulinic acid,
weight of catalyst 0.5 g, XPt-10WO3/g-Al2O3, reaction temperature 270 1C,
WHSV (1.14, 2.28, 3.42 and 4.56 hꢀ1 (with respect to LA)), and H2 flow
60 mL minꢀ1. GVL: g-valerolactone, VA: valeric acid, AL: angelica lactone,
others includes MTHF, C4-alcohols, and C5-alcohols.
Fig. 9 Effect of platinum loading on the hydrogenation of levulinic acid
over 2Pt–10WO3/g-Al2O3 catalysts.
by the NH3-TPD method. The total carbon content in used
catalysts was determined by CHNS analysis and the results are
presented in Table 5.
Table 6 Effect of reaction temperature on hydrogenation of LA to VA
2.2.3 Effect of weight hour space velocity (WHSV). The
influence of weight hour space velocity (WHSV) was investi-
gated for levulinic acid hydrogenation with different flow rates
of reaction feed (pure LA solution, density = 1.14 g mLꢀ1) from
Selectivity (%)
(a)
(b)
(c)
(d)
(e)
(f)
230
250
270
290
310
78
87
91
93
95
58
46
27
25
19
36
49
58
49
45
1
3
10
16
21
5
2
5
10
15
0.5 mL hꢀ1 (WHSV = 1.14 hꢀ1) to 2.0 mL hꢀ1 (WHSV = 4.57 hꢀ1
)
by keeping constant the weight of the catalyst. It is observed that
the LA conversion and VA selectivity decreased from 91% to 64%
and 58% to 47%, respectively, with an increase in WHSV from
1.14 hꢀ1 to 4.57 hꢀ1. In contrast, the selectivity of other products
such as GVL, AL, MTHF and others was enhanced. These results
demonstrate that maximum VA selectivity was obtained at lower
WHSV owing to the longer contact time. Since the available
number of active sites remains unchanged the decrease in LA
conversion and selectivity of VA at higher WHSV (high feed flow
rate) was as expected. WHSV values were calculated using the
following formula and the results are presented in Fig. 10:32
a – temperature (1C), b – LA conversion (%), c – GVL, d – VA, e – AL,
f – others. Reaction conditions: pure levulinic acid, catalyst weight
0.5 g, reaction temperature (230, 250, 270, 290 and 310 1C), WHSV
1.14 hꢀ1, and H2 flow 60 mL minꢀ1. GVL: g-valerolactone, VA: valeric
acid, AL: angelica lactone, others includes MTHF, C4-alcohols, and
C5-alcohols.
The catalytic activity mainly depends on the number of
active sites located on the support. The conversion of LA to
VA involves two consecutive steps: one is hydrogenation of LA
to GVL and the other is ring opening of GVL followed by
dehydration to VA. Hence, a catalyst used for the conversion
WHSV = mass of flow (g hꢀ1)/weight of the catalyst (g)
where mass of flow = feed flow rate (mL hꢀ1) ꢁ density of feed of LA to VA should possess both acidic and metallic sites to
flow (g mLꢀ1).
enable dehydration and hydrogenation steps respectively. In
2.2.4 Effect of reaction temperature. The influence of the present study, g-Al2O3 supported platinum–tungsten catalysts
reaction temperature on hydrogenation of levulinic acid over demonstrated superior catalytic performance due to the presence
the 2Pt–10WO3/g-Al2O3 catalyst in the range of 230–310 1C was of efficient metallic and acidic sites required for the conversion of
investigated and the results are presented in Table 6. The LA to VA. The active sites on Pt facilitated the hydrogenation of
levulinic acid (LA) conversion increased from 78% to 95% as LA to GVL and the acidic sites generated by incorporation of WO3
the temperature increased from 230 1C to 310 1C whereas the species in the catalyst facilitated the ring opening of GVL
selectivity of valeric acid (VA) reached 58% at 270 1C followed by followed by dehydration to VA.
a slight decrease at higher temperature. The optimal reaction
temperature to perform hydrogenation of LA to VA selectively (TOS) analysis up to 60 hours of reaction time was investigated
over a 2 wt% Pt–10WO3/Al2O3 catalyst is 270 1C. over 2Pt–10WO3/g-Al2O3 catalysts and the results are presented
2.2.5 Time-on-stream study. The effect of time on stream
18008 | New J. Chem., 2019, 43, 18003--18011 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
Fig. 12 (A) XRD, (B) TEM and (C) TPD images of the fresh and used
2Pt–10WO3/g-Al2O3 catalysts.
Fig. 11 Time on stream studies over 2Pt–10WO3/g-Al2O3 catalysts.
Table 7 Studies of the spent catalyst 2Pt–10WO3/Al2O3
(Table 7) due to the formation of carbon on surface acid sites
and blockage of mesopores. There were not many changes as
observed from the TEM images and XRD patterns of the spent
2Pt–10WO3/g-Al2O3 catalyst compared to the fresh catalyst as
shown in Fig. 12.
Catalyst
(a)
(b)
(c)
(d)
(e)
Fresh
Spent
91
88
58
53
27
25
167
155
190
188
a – LA conversion (%), b – VA (%), c – GVL (%), d – BET surface area (m2 gꢀ1),
e – acidity (mmol gꢀ1).
3. Experimental section
3.1 Catalyst preparation
in Fig. 11. At 5 h, the highest LA conversion and selectivity of VA
were 91% and 58% respectively. However, the maximum con-
version and selectivity remained stable till 25 h, and subse-
quently decreased with time. The time on stream results reveal
that such deactivation of the catalyst could be attributed to the
blocking of the active acid sites by carbon deposition as
revealed by CHNS analysis presented in Table 7.
2.2.6 Spent catalyst studies. The spent catalyst was regen-
erated by activating the catalyst in air at 300 1C for 3 h followed
by reduction in H2 flow at 300 1C for 3 h to remove the coke
deposited on the surface of the catalyst.33,34 The best catalyst
2Pt–10WO3/g-Al2O3 was reused for levulinic acid hydrogenation
in order to observe the catalytic performance and to study the
possible reasons for deactivation. The spent catalyst was char-
acterized by various techniques such as XRD, TEM, BET surface
area, CHNS analysis and NH3-TPD. The results are shown in
Table 7; the conversion of LA and selectivity of VA were slightly
decreased over the used catalyst when compared to that of the
fresh catalyst. By using CHNS analysis the amount of carbon
deposited on the spent catalyst was deduced (Table 8) and it
was noticed that the acidity and BET surface area of the spent
catalyst were slightly decreased compared to the fresh catalyst
Ammonium metatungstate hydrate ((NH4)6(H2W12O40)ꢂH2O),
chloro platinic acid hexahydrate (H2PtCl6ꢂ6H2O) and g-Al2O3
(Engelhard Corporation; BET surface area 214 m2
g
ꢀ1) and
deionised water were used. All chemicals were used as received.
A series of Pt–WO3 supported g-Al2O3 catalysts were synthesized by
the impregnation method. In a typical synthesis, g-Al2O3 was first
impregnated with tungsten precursor ((NH4)6(H2W12O40)ꢂH2O)
solution, followed by the introduction of Pt precursor
(H2PtCl6ꢂ6H2O) solution. The resulting solids were dried at
110 1C and calcined at 500 1C for 4 h in air after each impregnation
step. The loading of WO3 and Pt was 10 wt% and 0.5–3 wt%,
respectively. The final catalysts were labeled as XPt–YWO3/g-Al2O3,
where X refers to Pt loading and Y refers to WO3 loading.
3.2 Catalyst characterization
The X-ray diffraction (XRD) of catalysts was performed on a
Rigaku miniflex X-ray diffractometer The measurements were
taken at a scan rate of 21 minꢀ1 using Ni filtered Cu Ka radiation
(l = 0.15406 nm) from 2y = 2 to 651, with a beam voltage of 30 kV
and a beam current of 15 mA.
The surface area and pore size distribution studies of the
calcined catalysts were carried out by the multipoint BET method
using N2 adsorption–desorption experiments at ꢀ196 1C on
Autosorb 1 (Quanta Chrome Instruments).
Table 8 Elemental analysis by CHNS
Catalyst
N (%)
C (%)
H (%)
S (%)
The transmission electron microscopy (TEM) images were
taken on a Hitachi HT7700 microscope operated at 100 kV. The
sample grid was prepared by ultrasonically dispersing the
2Pt/10W–Al
Used 2Pt/10W–Al
—
—
—
5.20
0.80
1.24
0.35
0.55
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 18003--18011 | 18009

View Article Online
Paper
NJC
sample into ethanol, depositing droplets onto a carbon-coated temperature programmed reduction, and elemental mapping
copper grid, followed by drying in air. analyses showed the presence of tungsten in the synthesized
The FTIR spectra of pyridine adsorbed samples were catalysts. The temperature programmed desorption of ammonia
recorded on an IR (model: GC-FT-IR Nicolet 670) spectrometer and ex situ pyridine-FT-IR studies demonstrated the acidity
by the KBr disc method to elucidate the Brønsted and Lewis of the catalysts. Under optimized reaction conditions, the
acid sites. Initially the catalysts were activated in N2 flow at 2Pt–10WO3/g-Al2O3 catalyst exhibited the highest selectivity (58%)
300 1C for 1 h to remove moisture and cooled to room to valeric acid with 91% conversion of LA. The greater catalytic
temperature. This was followed by pyridine adsorption on the performance of the catalyst is attributed to the enhanced acidity
activated catalysts at 120 1C.
due to the incorporation of tungsten oxide, high surface area,
TPD experiments were carried out on an Auto Chem greater dispersion of Pt and the synergistic interaction between
2910 (Micromeritics, USA) instrument. In a typical experiment, active species and the support. The 2Pt–10WO3/g-Al2O3 catalyst
100 mg of dried catalyst was pretreated in high purity (99.995%) showed quite stable catalytic activity during the 60 h time on
helium (50 mL minꢀ1) at 200 1C for 1 h. After pretreatment, the stream. The slight decrease in the selectivity to VA is due to the
sample was saturated with 10% NH3–He (50 mL minꢀ1) at 80 1C deposition of carbonaceous species over the catalyst surface.
for 1 h and subsequently flushed with He (50 mL minꢀ1) at
80 1C for 30 min to remove physisorbed ammonia. TPD analysis
Conflicts of interest
was carried out from ambient temperature to 800 1C at a
heating rate of 10 1C minꢀ1
.
There are no conflicts to declare.
Diffusion reflectance spectra were recorded at room temperature
in the UV-visible region 200–900 nm at a split width of 1.5 nm and a
scan speed of 400 nm per minute using a GBC Cintra10e UV-visible Acknowledgements
spectrometer for the catalysts. About 15 mg of the catalyst sample
and dried KBr was used to make pellets by uniform mixing.
PB thanks University Grants Commission (UGC), New Delhi, for
the award of a Senior Research Fellowship and KVR thanks
CSIR, New Dehi, for financial support under the Emeritus
Scientist scheme.
3.3 Catalyst testing
The vapor hydrogenation reaction was carried out in a conventional
down flow fixed-bed quartz reactor (40 cm length, 9 mm i.d.) under
ambient pressure. Approximately, 0.5 g of the catalyst was mixed
with an equal amount of glass beads supported on the ceramic wool
bed in the reactor. The catalysts were reduced at 350 1C for 2 h in a
stream of H2. After cooling down to the reaction temperature
(270 1C), hydrogen (60 mL minꢀ1) and LA were introduced into
the reactor. The reaction products were condensed in an ice-water
trap and collected hourly for analysis on a gas chromatograph
GC-2014 (Shimadzu) equipped with a DB-wax 123-7033 (Agilent)
capillary column (0.32 mm i.d., 30 m long) and equipped with a
flame ionization detector. The total carbon content in used
References
1 B. Putrakumar, N. Nagaraju, V. Pavan Kumar and K. V. R.
Chary, Catal. Today, 2014, 7, 014.
2 X. Shaodan, Y. Deqing, Y. Tao and T. Panpan, RSC Adv.,
2017, 7, 1026.
3 N. P. Jindrayani, K. Alfin, E. S. Felycia, L. Shi-Yow, J. Yi-Hsu
and I. Suryadi, RSC Adv., 2015, 5, 41285.
4 V. V. Kumar, G. Naresh, S. Deepa, P. G. Bhavani, M. Nagaraju,
M. Sudhakar, K. V. R. Chary, J. Tardio, S. K. Bhargava and
A. Venugopal, Appl. Catal., A, 2016, 10, 032.
catalysts was determined by using
a CHNS Analyzer-
5 B. Kamm, P. R. Gruber and M. Kamm, Chem. Rev., 2006,
106, 4044.
6 J. J. Bozell, Science, 2010, 329, 522–523.
7 K. Kon, W. Onodera and K. I. Shimizu, Catal. Sci. Technol.,
2014, 4, 3227–3234.
8 Y. Qiu, L. Xin, D. J. Chadderdon, J. Qi, C. Liang and W. Li,
Green Chem., 2014, 16, 1305–1315.
ELEMENTAR Vario micro cube model. The overall carbon mass
balance was found to be 498% unless otherwise stated. The
conversion of levulinic acid and selectivity of products were
calculated as follows:
moles of LA ðinÞ ꢀ moles of LA ðoutÞ
Conversion ð%Þ ¼
ꢁ 100
moles of LA ðinÞ
9 D. J. Hayes, S. Fitzpatrick, M. H. B. Hayes and J. R. H. Ross, in
Biorefineries: Industrial Processes and Products, ed. B. Kamm,
P. R. Gruber, M. Kamm, Wiley-VCH, Weinheim, 2006, vol. 2,
pp.139–164.
moles of one product
moles of all products
Selectivity ð%Þ ¼
ꢁ 100
10 I. T. Horvath, H. Mehdi, V. Fabos, L. Boda and L. T. Mika,
Green Chem., 2008, 10, 238–242.
4. Conclusions
In conclusion, tungsten oxide incorporated platinum catalysts 11 S. V. D. Vyver, J. Thomas, J. Geboers, S. Keyzer, M. Smet,
supported on g-Al2O3 were successfully synthesized and applied
to the hydrogenation of levulinic acid to valeric acid in flow
W. Dehaen, P. A. Jacobsa and B. F. Sels, Energy Environ. Sci.,
2011, 4, 3601.
reactors at atmospheric pressure. SEM and TEM techniques 12 J. Xin, D. Yan, O. Ayodele, Z. Zhang, X. Lu and S. Zhang,
revealed the morphological aspects of the catalysts. X-ray diffraction, Green Chem., 2015, 17, 1065–1070.
18010 | New J. Chem., 2019, 43, 18003--18011 This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019

View Article Online
NJC
Paper
13 X. L. Du, Q. Y. Bi, Y. M. Liu, Y. Cao, H. Y. He and K. N. Fan, 24 E. P. Parry, J. Catal., 1963, 2, 371–379.
Green Chem., 2012, 14, 935–939.
14 J. Q. Bond, D. M. Alonso, R. M. West and J. A. Dumesic,
Langmuir, 2010, 26, 16291–16298.
15 C. E. Chan-Thaw, M. Marelli, R. Psaro, N. Ravasio and
F. Zaccheria, RSC Adv., 2013, 3, 1302–1306.
25 J. A. Horsley, I. E. Wachs, J. M. Brown, G. H. Via and F. D.
Hardcastle, J. Phys. Chem., 1987, 91, 4014–4020.
26 S. S. Priya, V. P. Kumar, M. L. Kantam, S. K. Bhargava,
A. Srikanth and K. V. R. Chary, Ind. Eng. Chem. Res., 2015,
54(37), 9104–9115.
16 W. Luo, P. C. A. Bruijnincx and B. M. Weckhuysen, J. Catal., 27 J. A. Horsley, I. E. Wachs, J. M. Brown, G. H. Via and F. D.
2014, 320, 33–41. Hardcastle, J. Phys. Chem., 1987, 91, 4014–4020.
17 N. P. Jindrayani, K. Alfin, E. S. Felycia, L. Shi-Yow, J. Yi-Hsu 28 A. D. Lucas, J. L. Valverde, P. Canizares and L. Rodriguez,
and I. Suryadi, RSC Adv., 2015, 5, 41285. Appl. Catal., A, 1998, 172, 165–176.
18 J. P. Lange, R. Price, P. M. Ayoub, J. Louis, L. Petrus, 29 X. L. Yang, W. L. Dai, R. Gao and K. Fan, J. Catal., 2007, 249,
L. Clarke and K. H. Gosselin, Angew. Chem., Int. Ed., 2010,
49, 4479–4483.
19 K. Kon, W. Onodera and K. I. Shimizu, Catal. Sci. Technol.,
2014, 4, 3227–3234.
20 T. Pan, J. Deng, Q. Xu, Y. Xu, Q. X. Guo and Y. Fu, Green
Chem., 2013, 15, 2967–2974.
278–288.
30 S. S. Priya, V. P. Kumar, M. L. Kantam, S. K. Bhargava and
K. V. R. Chary, RSC Adv., 2014, 4, 51893.
31 S. S. Priya, P. Bhanuchander, V. P. Kumar, D. Deepa, P. R.
Selvakannan, M. L. Kantam, S. K. Bhargava and K. V. R. Chary,
ACS Sustainable Chem. Eng., 2016, 4, 1212–1222.
21 P. Bhanuchander, S. S. Priya, V. P. Kumar, S. K. Hussain, 32 S. S. Priya, P. Bhanuchander, V. P. Kumar, S. K. Bhargava
N. P. Rajan, S. K. Bhargava and K. V. R. Chary, Catal. Lett.,
2017, 147, 845–855.
and K. V. R. Chary, Ind. Eng. Chem. Res., 2016, 55,
4461–4472.
22 J. S. Valente, M. S. Cantu, J. G. H. Cortez, R. Montiel, X. Bokhimi 33 S. S. Priya, V. P. Kumar, M. L. Kantam, S. K. Bhargava and
and E. L. Salinas, J. Phys. Chem., 2007, 111, 642–651. K. V. R. Chary, RSC Adv., 2014, 4, 51893–51903.
23 P. S. Dutta and A. Khanna, ECS J. Solid State Sci. Technol., 34 S. S. Priya, V. P. Kumar, M. L. Kantam, S. K. Bhargava and
2013, 2, R3153–R3167. K. V. R. Chary, Catal. Lett., 2014, 144, 2129–2143.
This journal is ©The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2019 New J. Chem., 2019, 43, 18003--18011 | 18011