Influence of W on the reduction behaviour and Br?nsted acidity of Ni/TiO2catalyst in the hydrogenation of levulinic acid to valeric acid: Pyridine adsorbed DRIFTS study

Article abstract of DOI:10.1016/j.apcata.2016.10.032

Effect of W on 20 wt%Ni/TiO₂catalyst is examined in the hydrogenation of levulinic acid (LA) to valeric acid at ambient H₂pressure. The interaction between W and Ni had a significant influence on the hydrogenation activity and product selectivity. The H₂-TPR (temperature programmed reduction) results emphasized a shift in T_maxto very high temperatures due to W species which are in close proximity to Ni particles. The N₂O decomposition measurements showed a decrease in N₂O uptake with the increase in ‘W’ loadings due to a high ratio of Ni²⁺/Ni⁰species at higher tungsten content. X-ray photoelectron spectra (XPS) demonstrated a shift in binding energy to higher owing to a strong interaction between W and Ni particles by the presence of ionic Ni at the near surface region. The ionic Ni species seems to be involved in the conversion of γ-valerolactone (GVL) to valeric acid (VA). Pyridine adsorbed infrared (IR) spectra revealed an enormous increase in surface Br?nsted acidity originated from tungsten interacted Ni/TiO₂are the sites responsible for ring opening of GVL to form VA.

Full text of DOI:10.1016/j.apcata.2016.10.032

Accepted Manuscript
Title: Influence of W on the reduction behaviour and Brønsted
acidity of Ni/TiO2 catalyst in the hydrogenation of levulinic
acid to valeric acid: Pyridine adsorbed DRIFTS study
Author: V. Vijay Kumar G. Naresh S. Deepa P. Ganga
Bhavani M. Nagaraju M. Sudhakar K.V.R. Chary J. Tardio
S.K. Bhargava A. Venugopal
PII:
S0926-860X(16)30539-7
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.10.032
APCATA 16048
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
27-8-2016
26-10-2016
31-10-2016
Please cite this article as: V.Vijay Kumar, G.Naresh, S.Deepa, P.Ganga Bhavani,
M.Nagaraju, M.Sudhakar, K.V.R.Chary, J.Tardio, S.K.Bhargava, A.Venugopal,
Influence of W on the reduction behaviour and Brønsted acidity of Ni/TiO2 catalyst in
the hydrogenation of levulinic acid to valeric acid: Pyridine adsorbed DRIFTS study,
Applied Catalysis A, General http://dx.doi.org/10.1016/j.apcata.2016.10.032
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Influence of W on the reduction behaviour and Brønsted acidity of Ni/TiO2 catalyst in
the hydrogenation of levulinic acid to valeric acid: Pyridine adsorbed DRIFTS study
V. Vijay Kumara,b, G. Naresha,b, S. Deepaa, P. Ganga Bhavania, M. Nagarajua, M. Sudhakara,
K.V.R. Charya, J. Tardiob, S. K. Bhargavab,* and A. Venugopala,*
a
Catalysis Laboratory, Inorganic and Physical Chemistry Division, CSIR-Indian Institute of
Chemical Technology, Hyderabad-500 007, Telangana, India.
b
Centre for Advanced Materials & Industrial Chemistry (CAMIC), School of Applied
Sciences, RMIT University, GPO BOX-2476, Melbourne 3001, Australia.
*Corresponding author, Email: akula@iict.res.in or suresh.bhargava@rmit.edu.au
Tel.: +91 40 27193165/+91 40 27193510; Fax: +91 40 27160921.

Graphical abstract

Highlights
 Strong interaction between W and Ni leads to increased Ni2+/Ni0 ratio on surface
 At ambient H2 pressure levulinic acid was converted to valeric acid over W-Ni/TiO2
 Lewis acid site in combination with Ni0 active for -valerolactone
 Brønsted acid sites close to nickel sites are responsible for valeric acid formation

Abstract
Effect of W on 20wt%Ni/TiO2 catalyst is examined in the hydrogenation of levulinic
acid (LA) to valeric acid at ambient H2 pressure. The interaction between W and Ni had a
significant influence on the hydrogenation activity and product selectivity. The H2-TPR
(temperature programmed reduction) results emphasized a shift in Tmax to very high
temperatures due to W species which are in close proximity to Ni particles. The N2O
decomposition measurements showed a decrease in N2O uptake with the increase in ‘W’
loadings due to a high ratio of Ni2+/Ni0 species at higher tungsten content. X-ray
photoelectron spectra (XPS) demonstrated a shift in binding energy to higher owing to a
strong interaction between W and Ni particles by the presence of ionic Ni at the near surface
region. The ionic Ni species seems to be involved in the conversion of -valerolactone (GVL)
to valeric acid (VA). Pyridine adsorbed infrared (IR) spectra revealed an enormous increase
in surface Brønsted acidity originated from tungsten interacted Ni/TiO2 are the sites
responsible for ring opening of GVL to form VA.
Key words: Levulinic acid, Valeric acid, -Valerolactone, DRIFTS, W-Ni/TiO2, Brønsted-
Lewis acidity.

1. Introduction
Development of catalysts for the production of fine chemicals and fuels additives
from biomass-derived compounds has become a topic of interest [1-3]. Inedible biomass
feedstock, namely lignocellulose is an attractive raw material for the synthesis of second-
generation biofuels and fuel additives [4]. A variety of multi-step processes to convert
inedible biomass into fuels include; hydrolysis, gasification and pyrolysis of biomass [5-6].
Of the many, production of levulinic acid (LA) has received significant interest due to the
potential conversion of LA into a range of fine chemicals and fuel additives [7-8]. Numerous
studies have been reported on LA hydrogenation to -valerolactone (GVL) over both
homogeneous and heterogeneous catalytic systems using noble and non-noble metal catalysts
in liquid phase as well as in the vapour phase [9-12]. Studies pertaining to GVL conversion to
other useful compounds such as valeric acid (VA), 1,4-pentanediol and 2-
methyltetrahydrofuran (MTHF) were also reported [13-18]. Among various compounds, the
high energy density liquid; valeric acid (VA) has been found to be a potential candidate in the
bio-fuel production as the valerate esters are being used as blending agents for both gasoline
and diesel [19]. Conversion of LA over non-noble metal based catalysts is a promising
approach when compared to the use of expensive noble metal catalysts [20]. From R&D
perspectives, catalytic processes with zero waste and reduction of multitude processes into a
single step are mandatory [21]. Exploration of base metal catalysts for the conversion of
levulinic acid to valeric acid is economically attractive. Thus, surface modification of Ni-
based catalysts for the single-step conversion of LA to VA is gaining momentum [22].
Lange et al. reported the synthesis of valerate esters in a multi-step process using
Pt/TiO2 catalyst for LA to GVL conversion followed by subsequent GVL conversion to VA
and esterification of VA to valerate ester over Pt/H-ZSM-5 catalyst [23]. In a systematic

study, Weckhuysen et al. examined the zeolite supported Ru catalyst for LA to VA with
45.8% yield using dioxane as solvent at 200 °C and 40 bar H2 pressure [24]. Fu et al. tested
the LA hydrogenation over Ru/SBA-SO3H catalyst in ethanol resulted in the co-production of
VA and ethyl valerate with a total yield of 94% at 240 °C and 40 bar H2 pressure [25].
Kenichi et al. developed direct and selective catalytic route for VA from LA over Pt/HMFI
catalyst at 200 °C and 8 bar H2 pressure [26]. Shell laboratory obtained a patent on the
conversion of LA to GVL and subsequent GVL conversion to VA with 70% yield over the
Pt/H-ZSM-5/SiO2 catalyst [27]. An excellent paper reported from Dumesic group on the
hydrogenation of LA to VA and 5-nonanone through GVL over a combination of Ru/C (LA
to GVL) and Pd/Nb2O5 (GVL to VA and 5-nonanone) catalysts [28]. Most of these reports
were on supported noble metal catalysts at slightly elevated H2 pressures.
To the best of our knowledge, very few reports are published on single-step
conversion of LA to VA in the vapour phase at ambient H2 pressure and the reaction
temperature maintained in this study otherwise is also very close to the reported studies on
the conversion of LA in liquid and vapour phase. Earlier we have reported the hydrogenation
of LA to GVL over supported Ni and Ru catalysts [29-31]. The 20wt%Ni/TiO2 has been
modified using several promoters such as Fe, Co, Cu, Bi, W and Sn, which showed promising
activity towards valeric acid in the vapour phase hydrogenation of aqueous LA (Table S1).
Among them, the W modified 2wt%W-20wt%Ni/TiO2 has demonstrated interesting results.
Hence, a detailed study on W modified Ni/TiO2 catalysts are investigated. The addition of W
to Ni/TiO2 drastically modified the reduction behaviour of Ni species and surface acidity
which in turn influenced the hydrogenation activity to produce GVL along with valeric acid.
The reduction behaviour of W modified Ni/TiO2 catalysts are investigated by H2-TPR
analysis. Nature of acid sites on W modified Ni/TiO2 catalyst is rationalised using pyridine
adsorbed DRIFTS and the Ni dispersion is estimated by N2O titration. Long term stability

tests and post characterization studies were also conducted on the optimized catalyst
composition.
2. Experimental
2.1 Catalyst Preparation
The Ni supported on TiO2 (P25 Degussa BET surface area 56 m2/g) catalyst was
prepared by an incipient wet impregnation (IWI) method. In a typical procedure, required
amount of aqueous Ni(NO3)3· 6H2O solution mixed with TiO2 (to give 20wt%Ni) under
constant stirring at 100 °C until the water evaporated, dried at 120 °C for 12 h subsequently
calcined in static air at 500 C for 4 h. The W modified 20wt%Ni/TiO2 catalysts with
different loadings (1.0 - 2.5wt%W) were impregnated on calcined 20wt%Ni/TiO2 by using
(NH4)10W12O41.xH2O precursor. Then the samples were dried at 120 C followed by
calcination in static air at 500 C for 4 h. For comparison, two different samples were
prepared namely (i) a 1.5wt%W/TiO2 was prepared by IWI in a similar protocol (ii)
mechanical mixing of 1.5wt%W/TiO2 and 98.5wt% of 20wt%Ni/TiO2 (calcined) followed by
calcination in static air at 500 C for 4 h. The elemental analysis (ICP-MS) of these samples
is reported in Table S2. The metal contents in all the catalysts were denoted by their weight
percentage.
2.2 Characterization of catalysts
Powder XRD patterns of the catalysts were recorded with Bruker D4 Endeavor wide
angle X-ray diffractometer (M/s. Bruker Corporation, Germany) using Ni-filtered Cu Kα
radiation (λ = 0.15418 nm) with a scan rate of 2° min-1 in the 2θ range of 10-90° at 40 kV and
20 mA. The crystallite size (D) was measured by applying Debye-Scherrer equation w.r.t. Ni
(111) plane. The acid site distribution on 20wt%Ni/TiO2 (fresh) and 20wt%Ni-
1.5wt%W/TiO2 (fresh and used) catalysts were recorded on Micromeritics (Auto Chem 2910)
by temperature programmed desorption (TPD) of NH3. In a typical method, about 0.15-0.2 g

of the sample was reduced at 500 °C for 1 h in 4.97% H2/Ar (v/v) at a flow rate of 30 mL
min−1. After reductive pretreatment, the sample was cooled to 100 °C followed by saturation
with NH3 gas and subsequently flushed in helium flow at 100 °C for 1 h to remove the
physisorbed NH3. The desorption of NH3 was then carried out by heating the sample up to
600 °C at a ramping rate of 5 °C min-1. The inductively coupled plasma-mass spectroscopy
(ICP-MS) analysis was done using Thermo Jarrell Ash model IRIS Intrepid II XDL, USA
instrument to determine the elemental composition of the catalysts. In the current study, prior
to the analysis, ~0.05 g of the catalyst sample was added to the mixture of aqua regia and
excess distilled water. Again 10 mL of the above solution was further diluted with distilled
water and was used for the ICP analysis.The metal surface area of W loaded Ni/TiO2
catalysts was measured by N2O decomposition experiments carried out at 80 °C. In a typical
method, about 0.05 g of the catalyst was reduced in H2 (4.97% H2/Ar) atmosphere 500 °C for
1 h followed by cooling down the reactor to 80 °C in helium flow (30 mL min-1). The pulses
(maximum of 3-4) of N2O are then injected onto the catalyst sample. The uptake of N2O is
monitored using GC equipped with a TCD detector and Porapak N column. The experimental
details on H2-TPR, XPS, TEM, BET-SA, and pyridine adsorbed DRIFT spectra of the
catalysts used in this study are similar to what was reported in our earlier papers [29-31] and
are also given in detail in the experimental section of ESI.
2.3 Activity measurements
Vapour phase hydrogenation of levulinic acid was carried out in a fixed-bed quartz
reactor (i.d. = 12 mm, length = 420 mm) in down flow mode. Prior to the reaction, the
catalyst sample (~0.2 to 0.3 g -18/+25 BSS sieved particles) were reduced at 500 °C for 4 h in
4.97% H2 (balance Ar) at a flow rate of 30 mL min-1. After reduction, the catalyst bed was
brought to the desired reaction temperature and the 10wt% aqueous LA was fed using an
HPLC pump along with H2 at a flow rate of 30 mL min-1. The gas flow rates were monitored

using mass flow controllers (Alicat Scientific) to regulate the flow. The product mixture was
collected in an ice cold trap for each 1 h and analyzed by GC-MS (QP5050A Shimadzu) and
gas chromatograph (Varian CP-3800 GC equipped with ZB capillary wax column and FID
detector). The carbon mass balance was >98% based on inlet and outlet carbon unless
otherwise stated. The gaseous products were analyzed by connecting the reactor outlet to a
six-way sampling valve interfaced to a GC with TCD (Agilent 7820A).
3. Results
3.1 Powder X-ray diffraction (XRD)
The powder XRD patterns of 20wt%Ni/TiO2 and W modified Ni/TiO2 samples are
reported in Figure S1. The diffraction lines at 2 = 37.2, 43.2, 62.8 and 75.4° due to NiO
phase with face centred cubic geometry (ICDD # 78-0429) is observed. Diffraction lines at
2 = 19.2, 31.0 and 65.8° are due to NiWO4 that has monoclinic geometry (ICDD # 72-0480)
corresponding to (100), (111) and (202) planes respectively. The diffraction peaks at 2 =
14.2, 28.8 and 32.6° are attributed to the presence of different mono and polymeric WxOy
species (such as WO3 and W18O49). The remaining diffraction lines are due to the TiO2
support both anatase and rutile; ICDD # 78-2486 and ICDD # 87-0710 respectively.
In the reduced samples, metallic Ni emerged as the main phase (ICDD # 04-0850) at
2 = 44.5, 52.0 and 76.5° (Figure 1). The diffraction lines due to NiWO4 are also found at 2
= 19.2, 31.0 and 65.8° (ICDD # 72-0480). Reduction of W-Ni/TiO2 samples led to strong
diffraction lines due to NiWO4 and WO2 (ICDD # 02-0414) species. From the XRD spectra
(inset of Figure 1), it is also observed that after the reduction at 500 °C, the W loaded
Ni/TiO2 samples showed the presence of ionic Ni (NiO; ICDD # 78-0429) along with
metallic Ni, whereas Ni/TiO2 catalyst showed only metallic Ni.

The WxOy species are reduced to various WxOy species (W18O49, W20O56 and W20O58)
along with WO2, whereas the 1.5wt%W/TiO2 sample has shown (Figure S2) WO3 along with
TiO2 phase. These results indicate that WOx species are not reduced and its existence is
different when impregnated on 20wt%Ni/TiO2. Reduction of WxOy species seems to occur
while WxOy species are in close contact with nickel particles by the hydrogen activation on
Ni surface. At lower W loadings (1.0 and 1.5wt.%), the diffraction lines of WxOy species are
absent. Interestingly, the 1.5wt%W loaded 20wt%Ni/TiO2 sample showed more prominent
and sharp diffraction line due to NiWO4 phase and above this loading; the intensity of
diffraction lines due to WO2 species are strong. The XRD spectra also revealed the presence
of WO2 species at higher loading (W = 2.5wt%) after reduction of 2.5wt%W-20wt%Ni/TiO2
sample at 500 °C for 4 h (Figure S3).
3.2 Physico chemical characteristics
The Physico chemical characteristics of the TiO2 supported Ni catalysts are compiled
in Table 1. The N2-sorption analysis showed a small decrease in the surface area of the
Ni/TiO2 catalyst after W addition and the similar trend is observed as W loading is increased
from 1 to 2.5wt.%. TPD of NH3 results showed a significant increase in the total acidity after
W modification to the Ni/TiO2 catalyst. Similarly, a higher number of acid site distribution is
observed for the 2.5wt.% W loaded Ni/TiO2 catalyst. The results of NH3-TPD analysis results
are also in good agreement with the results obtained from the pyridine adsorbed DRIFTS
analysis; where significantly higher BAS/LAS ratio is observed for the W modified Ni/TiO2
catalyst compared to that of un-promoted one. The crystallite size of Ni is also calculated
(Table 1) for all the W loaded Ni/TiO2 catalysts and the results showed a small increase in the
Ni crystallite size after W modification to the Ni/TiO2 catalysts and the results are consistent
with the results obtained from TEM analysis of these catalysts (increase in average particle
size of Ni after W addition to the Ni/TiO2 catalyst).

3.3 H2-Temperature programmed reduction (TPR)
The TPR patterns of 1.5wt%W/TiO2 (IWI) sample did not show any reduction signals
indicating the inert behaviour of tungsten oxide species even up to 1000 C (Figure 2a). It
should be noted that WO3 species require high temperatures for reduction when deposited on
TiO2 alone i.e. in the absence of Ni [32]. In contrast, the 20wt%Ni/TiO2 with W loadings of
0.0, 1.0, 1.5, 2.0 and 2.5wt.% showed two reduction peaks. The low-temperature reduction
peak around 452 C (Figure 2b) is due to Ni2+ to Ni0 (dispersed NiO) and the high-
temperature signal at 648 C is ascribed mostly to the reduction of large clusters of NiO
particles and also due to the reduction of NiO particles interacted with tungsten to some
extent [33]. Both of these reduction peaks are shifted to high temperatures upon increasing
the W loading on 20wt%Ni/TiO2 (Figure 2c-2f). The reduction peak shift towards high
temperatures with the increase in W loading is probably due to an interaction between nickel
and tungsten oxide particles. The H2 uptakes are marginally increased upon doping of W to
20wt%Ni/TiO2 catalyst (Table S2). This increase is probably due to hydrogen activation on
Ni site which is catalysing the reduction of tungsten oxide species to some extent as the
1.5wt%W/TiO2 alone did not show any reduction signal. From these results it can be inferred
that a reduction temperature of 500 C is insufficient to convert the total Ni2+ to metallic Ni
and a minimum of temperature >750 C is required to obtain the metallic Ni for the W loaded
samples (Figure 2b-2f). Therefore, the W modified 20wt%Ni/TiO2 samples reduced at 500
C demonstrated the existence of both metallic and ionic Ni particles (inset Figure 1). The
shift in reduction signal to very high temperatures is originated from the strongly interacted
NiO with W oxide particles which is also further supported by the UV-DR spectra (Figure
S4) of the 20wt%Ni/TiO2 and 1.5wt%-20wt%Ni/TiO2 catalysts.
3.4 X-ray photo electron spectroscopy (XPS)

The X-ray photoelectron spectra of reduced 20wt%Ni/TiO2 and 1.5wt%W-
20wt%Ni/TiO2 samples are reported in Figure 3. The Ni 2p signals of 20wt%Ni/TiO2 showed
nickel in both Ni0 and Ni2+ species in the near surface region with binding energy (BE) of
852.1 eV (Ni0) and Ni2+ signal at 855.4 eV respectively [33]. No significant shift (in BE) is
observed for metallic Ni over 20wt%Ni/TiO2 however, after W modification the BE of ionic
Ni is shifted to higher BE in the 1.5wt%W-20wt%Ni/TiO2 compared to 20wt%Ni/TiO2
sample, thus emphasizing the interaction of ionic Ni with tungsten particles on the catalyst
surface. A high ratio of ionic Ni is observed from the TPR analysis wherein a drifting of
reduction signals towards high temperatures suggesting the interaction between Ni and W.
These results are in good agreement with earlier reports on a strong interaction between W
and Ni particles [34].
3.5 Pyridine adsorbed DRIFTS analysis
The nature of acid sites is investigated by pyridine adsorbed DRIFT spectroscopy
(Figure 4). The vibrational bands due to pyridine-surface acid site interactions are
distinguished by a peak fitting using the Gaussian function. These spectra show the formation
of both protonated pyridine (1540 and 1636 cm-1) and co-ordinated pyridine (at 1620 and
1450 cm-1) in the presence and the absence of W in 20wt%Ni/TiO2 catalyst. The other bands
present in the DRIFT spectra correspond to both protonated and coordinated pyridine at 1489
cm-1 and 1513 cm-1. The bands present in 1580-1600 cm-1 region is due to a weak interaction
of pyridine with the acid sites present on the catalyst surface [35-36]. The characteristic
pyridine adsorbed IR bands attributed to Lewis acid (1448 cm-1) and Brønsted acid sites at
1539 cm-1 (quantified DRIFT spectra is obtained using Kubelka-Munk correction) are used to
measure the relative distribution of the acid sites on 20wt%Ni/TiO2 and 1.5wt%W-
20wt%Ni/TiO2 catalysts. The spectra clearly demonstrated a high proportion of Brønsted to
Lewis acid sites over 1.5wt%W-20wt%Ni/TiO2 compared to 20wt%Ni/TiO2 catalyst.

3.6 Measurements of the catalytic performance
The vapour phase hydrogenation of LA carried out over various W loaded (0.0 to 2.5
wt%) 20wt%Ni/TiO2 catalysts and the data is reported in Table 2. Conversion of LA
decreased from 92.3 to ~73.0% upon doping of W to 20wt%Ni/TiO2 and a marginal
difference even at higher tungsten loadings is found. Interestingly, the rate of angelica lactone
is increased with increase in tungsten loading. However, the rate of VA increased up to a
loading of 1.5wt%W and beyond this, it is slightly decreased. It is also found that the rate of
VA formation is at the cost of GVL. The enhanced rate of angelica lactone in the W loaded
20wt%Ni/TiO2 catalysts may be due to a higher number of acid sites contributed from W-
TiO2 interacted species. Formation of other products such as MTHF, C4, and/or C5-alcohols
are absent in the absence of tungsten. On the contrary, yields of these products are increased
with increase in W loading. The addition of tungsten to 20wt%Ni/TiO2 seems to enhance the
Brønsted acid sites which are identified as active sites for the ring opening of GVL to form
valeric acid [26].
The 1.5wt%W/TiO2 catalyst showed about 5% LA conversion with 92% angelica
lactone selectivity (Table 3). Products such as VA and GVL are absent on the 1.5wt%W/TiO2
catalyst. These results thus explain that 1.5wt%W/TiO2 sample seems to dehydrate the LA
and the active sites present on 1.5wt%W/TiO2 are inactive for the hydrogenation of angelica
lactone into either GVL or other products. Modification of surface Ni by W drastically
changed the product distribution (Table 2 and Table 3). The rate of GVL is reduced
consequently the VA rate is improved in the presence of tungsten with an optimum tungsten
loading of 1.5wt% demonstrated 48.5% selectivity of VA.
In the absence of Ni, the 1.5wt%W/TiO2 demonstrated angelica lactone as the major
product. These results thus indicate the involvement of Ni sites that are crucial for the
hydrogenation particularly, the reduction of AL to GVL on metal sites and the ring opening

of GVL over Brønsted acid sites. To further understand the role of the W and the NiO-WOx
interacted species on LA conversion and on the product selectivity; the reaction is carried out
using mixed oxides composed of 20wt%Ni/TiO2 + 1.5wt%W/TiO2 prepared by mechanical
mixing. The mechanically mixed catalyst showed selective GVL formation and very poor VA
yields (Table 3). Therefore, it can be concluded that Ni interacted with W are prone to ring
opening of GVL to form VA on Brønsted acid sites.
To gain an insight into the reaction mechanism, catalytic activities were evaluated
using aqueous GVL (Table 4). The rate of VA over 20wt%Ni/TiO2 catalyst is an order of
magnitude lower than on 1.5wt%W-20wt%Ni/TiO2. The selectivity of VA was considerably
good (>58%) over 20wt%Ni-1.5wt%W/TiO2 with a 60% conversion of GVL. However, the
conversion of GVL over both the catalysts was lower compared to LA conversion under
identical experimental conditions (Table 4). Using angelica lactone as a substrate, the
20wt%Ni/TiO2 exhibited better conversion (~68.8%) compared to 1.5wt%W-20wt%Ni/TiO2
catalyst (~44.6%). The rate of GVL is almost twice over 20wt%Ni/TiO2 than on 1.5wt%W-
20wt%Ni/TiO2 (Table 4). Whereas GVL as a substrate, the hydrogenation rate of -keto group
(in GVL) to MTHF is slightly higher in the absence of tungsten. The hydrogenolysis rate (i.e.
ring opening via C-O cleavage) of GVL to diols is low over 1.5wt%W-20wt%Ni/TiO2
catalyst emphasizing that metallic Ni sites influencing GVL ring to a large extent.
3.7 Time on stream activity results
Catalyst longevity test was conducted (60 h) to investigate the stability of the
1.5wt%W-20wt%Ni/TiO2 for VA formation (Table 5). The results of this analysis showed no
significant loss in activity even after 60 h of the reaction.
The CHNS analysis of the used 1.5wt%W-20wt%Ni/TiO2 catalyst showed a small
amount of coke deposition (Table 6). A comparison between XRD analysis of reduced (at
500 C for 4 h) and used 1.5wt%W-20wt%Ni/TiO2 catalyst (recovered after 60 h of

continuous operation) showed no significant changes in the intensity of diffraction lines due
to metallic Ni and/or TiO2 except the formation of polymeric WxOy species only in the used
sample (Figure 5). The Ni crystallite size measured by using Debye-Scherrer equation for
both the reduced and used forms of 1.5wt%W-20wt%Ni/TiO2 catalyst is found to be ca. 22.7
nm which is very close to the Ni particle size (ca. ~25.0 nm) measured by TEM analysis of
the 1.5wt%W-20wt%Ni/TiO2 (reduced and used) catalyst (Table 6).
Figure 6 shows the Ni 2p spectra of both the reduced (at 500 C for 4 h) and used
(recovered after 60 h of continuous operation) forms of the 1.5wt%W-20wt%Ni/TiO2 sample.
It shows that Ni is present in both metallic (2p3/2) at a BE of 851.8 eV (satellite peak at 862.8
eV) and oxidic form (2p3/2) with the binding energy of 856.7 eV in both the reduced and used
samples. No significant changes were observed in peak positions due to binding energies of
metallic and ionic Ni present 1.5wt%W-20wt%Ni/TiO2 sample exemplifying the stability of
the catalyst. The XPS of W4f spectra showed the presence of W in +6 oxidation state with
corresponding BE at 37.5 eV and 35.6 eV ascribed to W4f5/2 and W4f7/2 respectively (Figure
7) [34]. The emergence of satellite peak is an indication WxOy species, observed only in the
case of used catalyst. The surface characteristics are in good agreement with the bulk
properties of the used 1.5wt%W-20wt%Ni/TiO2 catalysts revealed by XRD analysis that
showed polymeric WxOy species (Figure 5).
The TPD of NH3 studies revealed a marginal decrease in acidity of the used
1.5wt%W-20wt%Ni/TiO2 compared to its reduced form. The decrease in acidity was most
likely due to the irreversible chemisorption of a fouling agent such as angelica lactone (one of
the intermediates in the reaction sequence) leads to deposition of carbon on the catalyst
surface during the course of the reaction (Table 6). The ICP-MS analysis of fresh and used
samples revealed that there is no leaching of either Ni or W species even after 60 h of

continuous operation thus suggesting the stability of the catalyst under the reaction conditions
maintained in this study.
4 Discussion
According to the results compiled in Table 2, the drastic fall in the rate of GVL can be
explained by a decrease in surface Ni metal sites by the interaction with tungsten which is
confirmed from N2O decomposition measurements (Table 2). This phenomenon is further
supported by the H2-TPR analysis of the tungsten loaded 20wt%Ni/TiO2 catalysts (Figure 2).
Though the rate of VA is increased up to 1.5wt%W; the GVL rate is further decreased at
higher W loadings. This indicates that at higher loadings of tungsten; conversion of GVL
either to MTHF or other products occurs. In the comparative analysis, the 20wt%Ni/TiO2
demonstrated very low VA and a higher GVL rate than the 1.5wt%W-20wt%Ni/TiO2 catalyst
(Table 2).
Further increase in W loading led to a marginal decrease in VA selectivity with the
formation of other products such as MTHF and 1, 4-diol. At higher W loaded (2.0 and
2.5wt.%) Ni/TiO2 catalysts, an increase in the Brønsted acidity is expected (Figure 4 and
Table 1), at the same time formation of WO2 species are higher compared to lower W loading
samples (confirmed by X-ray line broadening studies and XPS analysis of the 2.5wt%W-
20wt%Ni/TiO2 catalyst; Figure S3 and Figure S5). It is also found that the addition of W to
Ni/TiO2 led to a decrease in GVL selectivity which confirms the ring opening of GVL
occurring on the Brønsted acid sites to form VA. Therefore, from the aforementioned results,
it is clear that the influence of W loading is effective at W = 1.5wt% for the optimized VA
yield under the reaction conditions maintained in this investigation.
Reduction of the tungsten modified 20wt%Ni/TiO2 samples at 500 C results in the
formation of both metallic and ionic Ni species. The TPR analysis showed the increased
interaction between nickel and tungsten particles caused the difficulty in the reduction of

nickel oxide species as indicated a dramatic shift in Tmax to very high temperatures (Figure 2).
We believe that metallic Ni seems to be more reactive than the ionic Ni and the Ni sites that
have closely interacted with W-TiO2 particles may be involved in the further conversion of
GVL to VA. This effect is clearly seen as all the samples were reductively pre-treated at 500
C prior to the reaction and the TPR patterns demonstrated that upon increasing W loading;
the peak intensity of high-temperature reduction signal attributed to ionic Ni is increased. On
the other hand, Ni particles interacted with TiO2 are selective for the formation of GVL
through the hydrogenation of angelica lactone.
The overall VA rate is higher by an order of magnitude in presence of W. The
Ni/TiO2 is selective for the conversion of LA to GVL and GVL as a substrate the ring
opening of GVL occurred to form diols rather than VA (Table 2 and Table 4). Quiet contrast
to this, the rate of VA is higher by an order of magnitude over 1.5wt%W-20wt%Ni/TiO2 with
GVL as a substrate compared to 20wt%Ni/TiO2. Therefore formation of VA has occurred in
presence of W adjacent to Ni site on the catalyst surface.
Based on the results obtained and from earlier reports [29, 31], the hydrogenation of
AL to GVL is rapid relative to ring opening of GVL to form other products such as VA and
MTHF (Scheme 2). The LA stream that contained large amount of H2O (90wt.%) tends to
influence the surface ionic Ni species rather than metallic Ni during the course of the
reaction. One would expect two possibilities wherein (i) first case the metallic Ni reacts with
H2O to form NiO and H2 and the second case (ii) is the reaction between WOx-NiO and H2O
leads to the formation of enhanced Brønsted acid sites such as Wx(OH)y-Ni(OH)2. It is
unlikely that the metallic Ni reacts with H2O to form H2 and NiO (reversal of H2-TPR; first
case). It has been observed that the reduced 1.5wt%W-20wt%Ni/TiO2 sample showed the
presence of Ni2+ species due to a strong interaction between Ni and W particles (Figure 1 in
set). These species seems to be stabilising the ionic Ni. Hence, the first case may be ruled out

where in the reduced 20wt%Ni/TiO2 exhibited diffraction lines attributed to metallic Ni. The
second case can be considered as the Ni particles in close contact with W particles are highly
populated with ionic Ni than the Ni0 which is higher on 20wt%Ni/TiO2 (Table S3). We
believe that either WxOy or a combination of WxOy-NiO interacted species behave as
Brønsted acid sites rather than basic sites that are contributing to the enhanced VA
selectivity.
The catalyst stability test showed no significant loss in the activity (towards VA) over
the 1.5wt%W-20wt%Ni/TiO2 catalyst even after 60 h of reaction time. The used catalyst
(after 60 h of the reaction time) is also characterized by XRD, CHNS, XPS, ICP-MS, TEM,
and TPD of NH3 analyses and compared with that of reduced 1.5wt%W-20wt%Ni/TiO2
catalyst (Table 6). The results of this analysis showed no significant differences between the
fresh (reduced at 500 C) and used (after 60 h) 1.5wt%W-20wt%Ni/TiO2 catalyst except a
small loss in acidity of the catalyst. This shows the 1.5wt%W-20wt%Ni/TiO2 catalyst is
stable under the experimental conditions used to produce the VA from aqueous LA.
5 Conclusions
The W modified Ni/TiO2 catalyst prepared by simple impregnation method has been
identified as a promising catalyst for single step synthesis of valeric acid (an important
intermediate in the production of valeric biofuels) from biomass derived levulinic acid at
ambient H2 pressure and moderate reaction temperatures. The bulk and surface properties of
the 20wt%Ni/TiO2 catalyst is significantly altered by the addition of W which is corroborated
by H2-TPR, XPS and N2O titration. H2-TPR results showed the interaction between W and Ni
particles and the interaction became stronger with the increase in W loadings. It was found
that the synergism between Ni and W is virtuous at W loading of 1.5wt% on 20wt%Ni/TiO2.
The N2O decomposition measurements revealed that higher the Ni metal surface area; higher

the selectivity of -valerolactone. On the contrary, a drastic decrease in surface metallic Ni
due to interaction with W resulted in the formation active sites for increased valeric acid
selectivity in parenthesis a decrease in the selectivity of GVL is observed. The pyridine
adsorbed DRIFT spectroscopic studies demonstrated a dramatic increase in the Brønsted
acidity of 20wt%Ni/TiO2 by the addition of tungsten. 20wt%Ni/TiO2 alone was selective for
the hydrogenation of levulinic acid to -valerolactone whereas the Ni interacted with W are
the active species for formation of valeric acid. The enhanced Brønsted acid sites on W
modified 20wt%Ni/TiO2 catalyst was responsible for the formation of valeric acid through
the ring opening of -valerolactone. The 1.5wt%W-20wt%Ni/TiO2 catalyst was also found to
be stable up to 60 h of continuous operation. Finally, it can be concluded that a combination
of metallic Ni and the Ni interacted with W species are active and selective for the single-step
conversion of levulinic acid to valeric acid.
Acknowledgements
The authors VVK and GN thank CSIR-IICT Hyderabad and RMIT Australia for the
award of fellowships. SD, KGB thank DST for PhD fellowship through INSPIRE program.
All authors acknowledge Frank Antolasic, RMMF staff, RMIT, Australia for the XPS
analysis and other characterization.

References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
D. M. Alonso, J. Q. Bond, J. A. Dumesic, Green Chem. 12 (2010) 1493-1513.
A. V. Bridgwater, Catal. Today 29 (1996) 285-295.
G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044-4098.
D. M. Alonso, S. G. Wettstein, J. A. Dumesic, Green Chem. 15 (2013) 584-595.
N. A. S. Ramli, N. A. S. Amin, Appl. Catal. B: Environ. 163 (2015) 487-498.
M. Stocker, Angew. Chem. Int. Ed. 47 (2008) 9200-9211.
Y. Guo, K. Li, X. Yu, J. H. Clark, Appl. Catal. B: Environ. 81 (2008) 182-191.
J. J. Bozell, L. Moens, D. C. Elliott, Y. Wang, G. G. Neuenscwander, S. W.
Fitzpatrick, R. J. Bilski, J. L. Jarnefeld, Resour. Conserv. Recy. 28 (2000) 227-239.
[9] M. Chalid, A. A. Broekhuis, H. J. Heeres, J. Mol. Catal. A: Chem. 341 (2011) 14-21.
[10] M. Chia, J. A. Dumesic, Chem. Commun. 47 (2011) 12233-12235.
[11] M. Sudhakar, M. L. Kantam, V. S. Jaya, R. Kishore, K. V. R. Chary, A. Venugopal,
Catal. Commun. 50 (2014) 101-104.
[12] S. G. Wettstein, J. Q. Bond, D. M. Alonso, H. N. Pham, A. K. Datye, J. A. Dumesic,
Appl. Catal. B: Environ. 117-118 (2012) 321-329.
[13] J. Xin, D. Yan, O. Ayodele, Z. Zhang, X. Lu, S. Zhang, Green Chem. 17 (2015) 1065-
1070.
[14] X. L. Du, Q. Y. Bi, Y. M. Liu, Y. Cao, H. Y. He, K. N. Fan, Green Chem. 14 (2012)
935-939.
[15] J. Q. Bond, D. Martin Alonso, R. M. West, J. A. Dumesic, Langmuir 26 (2010)
16291-16298.
[16] C. E. Chan-Thaw, M. Marelli, R. Psaro, N. Ravasio, F. Zaccheria, RSC Adv. 3 (2013)
1302-1306.

[17] W. Luo, P. C. A. Bruijnincx, B. M. Weckhuysen, J. Catal. 320 (2014) 33-41.
[18] M. Li, G. Li, N. Li, A. Wang, W. Dong, X. Wang, Y. Cong, Chem. Commun. 50
(2014) 1414-1416.
[19] L. Xin, Z. Zhang, J. Qi, D. J. Chadderdon, Y. Qiu, K. M. Warsko, W. Li,
ChemSusChem 6 (2013) 674-686.
[20] P. P. Upare, J. Lee, Y. K. Hwang, D. W. Hwang, J. Lee, S. B. Halligudi, J. Hwang, J.
Chang, ChemSusChem 2011, 4, 1749-1752.
[21] Y. Qiu, L. Xin, D. J. Chadderdon, J. Qi, C. Liang, W. Li, Green Chem. 16 (2014)
1305-1315.
[22] P. Sun, G. Gao, Z. Zhao, C. Xia, F. Li, Appl. Catal. B: Environ. 189 (2016) 19-25.
[23] J. P. Lange, R. Price, P. M. Ayoub, J. Louis, L. Petrus, L. Clarke, H. Gosselink,
Angew. Chem. Int. Ed. 49 (2010) 4479-4483.
[24] W. Luo, U. Deka, A. M. Beale, E. R. H. van Eck, P. C. A. Bruijnincx, B. M.
Weckhuysen, J. Catal. 301 (2013) 175-186.
[25] T. Pan, J. Deng, Q. Xu, Y. Xu, Q. X. Guo, Y. Fu, Green Chem. 15 (2013) 2967-2974.
[26] K. Kon, W. Onodera, K. i. Shimizu, Catal. Sci. Technol. 4 (2014) 3227-3234.
[27] J. P. Lange, US Patent 2011/0112326, Shell Int. B.V. 2011.
[28] J. C. Serrano-Ruiz, D. Wang, J. A. Dumesic, Green Chem. 12 (2010) 574-577.
[29] V. V. Kumar, G. Naresh, M. Sudhakar, J. Tardio, S. K. Bhargava, A. Venugopal,
Appl. Catal. A: Gen. 505 (2015) 217-223.
[30] M. Sudhakar, V. V. Kumar, G. Naresh, M. L. Kantam, S. K. Bhargava, A. Venugopal,
Appl. Catal. B: Environ. 180 (2016) 113-120.
[31] V. V. Kumar, G. Naresh, M. Sudhakar, S. K. Bhargava, J. Tardio, V. Karnakar, A. H
Padmasri, A. Venugopal, RSC Adv. 6 (2016) 9872-9879.

[32] M. I. Zaki, N. E. Fouad, S. A. A. Mansour, A. I. Muftah, Thermochemica Acta 523
(2011) 90-96.
[33] W. Lin, H. Cheng, L. He, Y. Yu, F. Zhao, J. Catal. 303 (2013) 110-116.
[34] S. V. Green, A. Kuzmin, J. Purans, C. G. Granqvist, G. A. Niklasson, Thin Solid
Films 519 (2011) 2062-2066.
[35] N. Scotti, M. Dangate, A. Gervasini, C. Evangelisti, N. Ravasio, F. Zaccheria, ACS
Catal. 4 (2014) 2818-2826.
[36] H. Knozinger, P. Ratnasamy, Catal. Rev. 17 (1978) 31-70.

Figure captions:
Table 1: Physico chemical characteristics of the TiO2 supported Ni catalysts
Table 2: Hydrogenation of LA over W modified 20wt%Ni/TiO2 catalysts prepared by co-
impregnation of 20wt%Ni and varied W loadings on TiO2. Reaction temperature = 270 °C;
catalyst wt ~0.25 g; 10wt% aqueous LA; GHSV = 4.56 mL g-1 s-1
Table 3: Effect of method of preparation of W-Ni/TiO2 catalysts in the hydrogenation of LA
Reaction temperature = 270 °C; 10 wt% aqueous LA; catalyst wt: ~ 0.25 g
Table 4: Hydrogenation of different feeds over 20wt%Ni/TiO2 and 1.5wt%W-20wt%Ni/TiO2
catalysts, reaction temperature: 270 °C; 10wt% aqueous (LA and/or GVL) feed and 4wt%
aqueous AL, H2 flow rate 30 mL min-1 (GHSV = 4.56 mL g-1 s-1); Catalyst: ~ 0.25 g
Table 5: Time on stream study of LA hydrogenation over the 1.5wt%W-20wt%Ni/TiO2
catalyst Reaction conditions: 270 °C, 10% aqueous LA, GHSV=4.56 mL g-1 s-1
Table 6: Physicochemical properties of the reduced and used 1.5wt%W-20wt%Ni/TiO2
catalyst
Scheme 1: Catalytic hydrogenation of levulinic acid to valeric acid
Scheme 2: Reaction pathways during LA hydrogenation over W-Ni/TiO2 catalysts
Figure 1: Powder XRD patterns of the reduced 20wt%Ni-W/TiO2 with W loadings of (a) 0.0
(b) 1.0 (c) 1.5 (d) 2.0 and (e) 2.5wt% catalysts
Figure 2: H2-TPR profiles of the (a) 1.5wt%W/TiO2 and 20%Ni/TiO2 loaded with W (b) 0.0
(c) 1.0 (d) 1.5 (e) 2.0 (f) 2.5wt% catalysts
Figure 3: Ni 2p XPS spectra of fresh a) 20wt%Ni/TiO2 and b) 1.5wt%W-20wt%Ni/TiO2
catalysts
Figure 4: Pyridine adsorbed DRIFT spectra of a) 20wt%Ni/TiO2 and b) 1.5wt%W-
20wt%Ni/TiO2 catalysts

Figure 5: Powder XRD patterns of a) reduced and b) used 1.5wt%W-20wt%Ni/TiO2 catalysts
Figure 6: Ni 2p XPS of reduced and used 1.5wt%W-20wt%Ni/TiO2 catalysts
Figure 7: XPS spectra of W 4f of the reduced and used 1.5wt%W-20wt%Ni/TiO2 catalyst

Tables
Table 1: Physico chemical characteristics of the TiO2 supported Ni catalysts
W-Ni/TiO2
(20wt% Ni)
0.0wt%W
1.0wt%W
1.5wt%W
2.0wt%W
2.5wt%W
BET-SA
(m2/g)a
42.1
41.8
40.5
NH3 uptake
BAS/LAS
Crystallite/Average particle size (nm)
(µmol g-1)b
ratioc
XRDd
21.5
22.1
22.7
22.5
22.8
TEMe
41.2
-
64.8
-
0.2
-
1.8
-
22.4±1.6
-
24.8±1.1
39.6
36.1
-
-
76.8
-
a Obtained from N2 sorption analysis;
b Obtained from TPD of NH3;
c Calculated using quantified pyridine adsorbed DRIFTS curves;
d Crystallite size of Ni calculated form XRD patterns using Scherrer equation;
e Average particle size calculated from TEM images.

Table 2: Hydrogenation of LA over W modified 20wt%Ni/TiO2 catalysts prepared by co-impregnation of
20wt%Ni and varied W loadings on TiO2. Reaction temperature = 270 °C; Catalyst wt ~0.25 g; 10wt% aqueous
LA; GHSV = 4.56 mL g-1 s-1
W loading in
W-20wt%Ni/TiO2
N2O uptake
XLA
(%)
92.3 0.6
74.2 29.3 52.0
73.5 48.5 36.0
74.3 39.0 37.5
68.0 35.0 32.5
Selectivity (%)
a(TOF s-1 )ˣ 10-3
Othersb
-
VA
GVL
AL
-1
VA
GVL
AL
0.9
13.0 5.7
11.3 4.2
13.5 10.0
18.0 14.5
(mol gcat
237.0
)
0.0
1.0
1.5
2.0
2.5
98.5
0.04 7.3
2.56 4.5
5.83 4.3
4.24 4.0
3.66 3.4
0.06
1.13
1.36
1.47
1.88
162.0
116.8
130.5
124.2
a TOF = Rate / N2O uptake;
b Others include MTHF and C5-alcohols .

Table 3: Effect of method of preparation of W-Ni/TiO2 catalysts in the hydrogenation of LA
Reaction temperature = 270 °C; 10 wt% aqueous LA; Catalyst: ~ 0.25 g
Rate (mol s-1 g-1) ˣ 10-8
Catalyst
XLA
(%)
Selectivity (%)
VA
2.4
0.0
GVL
90.0
0.0
AL
1.0
92.1
Othersa
6.6
VA
1.5
0.0
GVL
56.3
0.0
AL
0.62
8.81
(20%Ni/TiO2+1.5wt%W/TiO2)-MMb
1.5wt%W/TiO2
32.7
5.0
7.9
a Others include MTHF and C5-alcohols;
b Catalysts prepared by mechanical mixing of Ni/TiO2 and W/TiO2.

Table 4: Hydrogenation of different feeds over 20wt%Ni/TiO2 and 1.5wt%W-20wt%Ni/TiO2 catalysts, reaction
temperature: 270 °C; 10wt% aqueous (LA and/or GVL) feed and 4wt% aqueous AL, H2 flow rate 30 mL min-1
(GHSV = 4.56 mL g-1 s-1); Catalyst: ~ 0.25 g.
Feed Catalyst
Conversion
(%)
Selectivity (%)
VA GVL AL MTHF1,4-Diol 2-pentanol VA
Rate (mol s-1 g-1) ˣ 10-8
GVL
-
-
114.5
64.6
AL
0.0
0.0
-
GVL 20wt%Ni/TiO2
GVL 20wt%Ni-1.5%W/TiO2
35.3
60.8
68.8
44.6
10.0
58.6
-
-
0.0 19.7 41.5
0.0 13.3 19.2
28.8
8.9
6.75
68.1
13.0
20.6
AL
AL
20wt%Ni/TiO2
20wt%Ni-1.5%W/TiO2
10.0 88.8
24.2 75.8
-
-
1.2
0.0
0.0
0.0
0.0
0.0
-