Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the Hydrocyclisation of Levulinic Acid to Γ-Valerolactone Using Ni/SiO2–Al2O3 Catalyst

Article abstract of DOI:10.1007/s10562-018-2618-7

Abstract: Commercial 65% Ni/SiO₂–Al₂O₃ (Ni/SA) catalyst was investigated for hydrocyclisation of levulinic acid (LA) to γ-valerolactone (Gvl) in presence of different hydrogen sources such as molecular hydrogen, isopropyl alcohol (IPA), and formic acid. At optimized reaction condition (200?°C, 10?bar H₂ for 30?min), the Ni/SA catalyst showed 100% yield of Gvl using molecular H₂ in tetrahydrofuran (THF) medium. The catalyst also exhibited 99% yield of Gvl in IPA through catalytic transfer hydrocyclisation of LA at 200?°C in 15?min. Further, the hydrocyclisation was successfully demonstrated in continuous mode using molecular hydrogen for 20?h time-on-stream which showed 98–99% conversion of LA with 100% selectivity of Gvl at optimized reaction condition in THF medium. Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-018-2618-7

Catalysis Letters
https://doi.org/10.1007/s10562-018-2618-7
Screening of Solvents, Hydrogen Source, and Investigation
of Reaction Mechanism for the Hydrocyclisation of Levulinic Acid
to γ-Valerolactone Using Ni/SiO2–Al2O3 Catalyst
Sreedhar Gundekari1,2 · Kannan Srinivasan1,2
Received: 8 August 2018 / Accepted: 18 November 2018
© Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract
Commercial 65% Ni/SiO2–Al2O3 (Ni/SA) catalyst was investigated for hydrocyclisation of levulinic acid (LA) to
γ-valerolactone (Gvl) in presence of different hydrogen sources such as molecular hydrogen, isopropyl alcohol (IPA), and
formic acid. At optimized reaction condition (200 °C, 10 bar H2 for 30 min), the Ni/SA catalyst showed 100% yield of Gvl
using molecular H2 in tetrahydrofuran (THF) medium. The catalyst also exhibited 99% yield of Gvl in IPA through catalytic
transfer hydrocyclisation of LA at 200 °C in 15 min. Further, the hydrocyclisation was successfully demonstrated in continu-
ous mode using molecular hydrogen for 20 h time-on-stream which showed 98–99% conversion of LA with 100% selectivity
of Gvl at optimized reaction condition in THF medium.
Graphical Abstract
Keywords Levulinic acid · γ-Valerolactone · Hydrocyclisation · Transfer hydrocyclisation · Nickel · Reusability · Reaction
mechanism · Continuous production
1 Introduction
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10562-018-2618-7) contains
supplementary material, which is available to authorized users.
Biomass-derived primary building blocks contain more
oxygen functionality, and for the conversion of these mol-
ecules to fuel and polymer intermediates mainly involves
Extended author information available on the last page of the article
Vol.:(0123456789)
1 3

S. Gundekari, K. Srinivasan
hydrogenation and/or hydrodeoxygenation reactions [1–6].
Levulinic acid (LA), is a keto carboxy functional molecule
that can easily be obtained through acid hydrolysis of lig-
nocellulosic biomass, and is of utmost industrial impor-
tance owing to the facile transformation of its functional
groups resulting in a gamut of value added products [7–9].
The hydrogenation reaction is extensively studied in the
case of LA, the building block in presence of H2 environ-
ment primarily ends with three useful products such as
γ-valerolactone (Gvl), 1,4-pentanediol (1,4-PDO), and
2-methyltetrahydrofuran (2-MTHF). Depending on the cata-
lytic system, reaction conditions, and solvents these products
may vary [10–12]. All the three hydrogenated products of
LA are industrially relevant, particularly Gvl, a sustainable
feedstock for energy and carbon-based chemicals (Fig. 1 in
the ESI) [13, 14]. Gvl itself acts as a greener solvent for
many organic transformations and is also reported as a good
medium for the conversion of lignocellulose components
[15–22].
reported using molecular H2 under batch and/or continuous
reaction conditions [30–40].
The Ni-based catalysts are also explored for catalytic
transfer hydrocyclisation (CTH) of LA to Gvl using vari-
ous alcohols as the hydrogen source. Rode reserch group
reported that 99% yield of Gvl using Ni/montmorillonite
(MMT) catalyst at 200 °C in 1 h in isopropyl alcohol-IPA
(H2 source) under N2 atmosphere [41]. Rong et al. screened
various supports for Ni catalyst such as SiO2, Al2O3, TiO2,
ZrO2, ZnO, and MgO, in which MgO supported Ni cata-
lyst showed maximum selectivity (93%) of Gvl at 150 °C,
10 bar H2 in 2 h in propanol medium [42]. Simple Raney-
Ni catalyst also exhibited conversion of alkyl levulinates to
Gvl with 94–99% yield under milder reaction conditions in
IPA medium presence of argon atmosphere [43, 44]. Obr-
egon et al. claimed Ni–Cu/Al2O3 as an efficient non-noble
bi-metallic catalyst for LA conversion towards two hydro-
genated products namely Gvl and 2-MTHF by systematic
tuning of metals, hydrogen source, and solvents wherein
they obtained these products in high selectivity [45, 46].
A few researchers have reported hydrocyclisation of LA
to Gvl using formic acid (FA) as a hydrogen source using
Ni metal based catalysts. The use of formic acid (FA) is
beneficial as it is the by-product obtained while preparing
LA from cellulose and is also safer than molecular H2. The
Ni-metal based catalysts such as Ni/Al2O3, Ni/MgO, Ni/
SiO2, Ni–Cu/SiO2 and Ag–Ni/ZrO2 were reported for this
conversion using FA source [47–50]. Rode and his research
group reported 99% yield of Gvl at 220 °C in 5 h using
10% Ag–20% Ni/ZrO2 catalyst in aqueous FA medium under
N2 atmosphere [49]. Upare et al. reported nickel-promoted
copper catalyst on silica nano composite support [(Ni(20)
Cu(60)–SiO2)] at 265 °C in 1,4-dioxane medium under N2
atmosphere and the catalyst exhibited 96% Gvl yield [50].
The reported Ni-based catalysts have some drawbacks
which includes: the requirement of co-metals, use of two
hydrogen source (alcohol+H2 or FA+H2), longer reaction
time, lesser selectivity towards product, necessity of inert
atmosphere for CTH, and higher energy consumption for the
reactivation of catalysts. Considering the above mentioned,
herein we introduce and successfully demonstrate commer-
cial Ni/SiO2–Al2O3 (Ni/SA) as an active catalyst for the con-
version of LA to Gvl in shorter reaction time. The catalyst
also shows remarkable catalytic activity towards CTH of LA
to Gvl using IPA as hydrogen source.
Palkovits, Wilson, Barbaro, and Lin reviewed the cata-
lytic hydrocyclisation of LA and its esters to Gvl using het-
erogeneous as well as homogeneous catalysts in both batch
and continuous flow reactors [12, 23–25]. Noble metals like
Ru, Pt, and Pd showed good catalytic activity towards the
LA to Gvl conversion. However, the noble metals are expen-
sive (because of minimum availability) which preclude their
use for bulk scale production of Gvl. Hence, many research
groups are working on non-noble metals like Ni and Cu. The
Ni-based catalysts are reported for this conversion using var-
ious hydrogen sources like alcohols (methanol, ethanol and
iso propyl alcohol etc), formic acid and molecular hydrogen.
In presence of molecular hydrogen (H2), various Ni-based
catalysts were reported: Chen research group found an mag-
netic recyclable (Ni4.59Cu1Mg1.58Al1.96Fe0.70) catalyst that
showed 98.1% yield of Gvl at 145 °C, 20 bar H2 for 3 h in
methanol and for recyclability of catalyst the material was
reactivated at 500 °C in H2 and demonstrated successfully
for five reaction recycles [26]. Pinto et al. screened various
bimetallic supported catalysts such as Ni–Mo/C, Cu–Mo/C,
Zn–Mo/C, Fe–Mo/C in which they claimed Ni–Mo/C is an
efficient catalyst that yielded 100% Gvl at 200 °C, 100 bar H2
and 2 h of reaction time in 1,4-dioxane medium [27]. Kantam
research group discussed Cu/Ni hydrotalcite-derived catalytic
system, wherein they achieved 100% yield of Gvl from LA
at 140 °C, 30 bar H2 in 3 h using dioxane solvent [28]. The
Grunwaldt research group reported Ni supported on Al2O3
(prepared by wet impregnation) towards this conversion that
was showed 57% yield of Gvl in aqueous medium at 200 °C,
50 bar H2 in 4 h [29]. Furthermore, Ni-based catalysts such
as Ni/Al2O3 (in dioxane solvent/vapour phase), Ni–MoOx/C,
Ni–Cu–Fe/SBA-15, 45Ni–5Cu–ZrO2 (76% yield of Gvl at
200 °C for 30 min), Ni–Sn/AlOH, Raney-Ni, Ni/TiO2, Ni/
ZrO2, Ni/H-ZSM-5 and Ni/mesoporous silica were also
2 Experimental Section
2.1 Materials
Levulinic acid (98%), γ-valerolactone (99%), methyl levuli-
nate (≥ 98%), reduced Ni/SiO2–Al2O3 powder (~ 65 wt%;
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
175 m2/g surface area; PubChem Substance ID: 24852528),
3 Results and Discussion
butyl levulinate (98%) were purchased from Sigma-Aldrich.
Isopropyl alcohol, formic acid, tetrahydrofuran, NiAl alloy
were procured from S.D. Fine Chemicals Limited, India.
3.1 Hydrocyclisation and CTH of Levulinic Acid (LA)
to γ‑Valerolactone (Gvl)
2.2 Batch Hydrocyclisation/CTH Experiments
and Analysis
3.1.1 Catalyst Screening
Initially, LA to Gvl conversion was demonstrated using two
types of H2 sources namely molecular H2 and alcohol (here
isopropyl alcohol-IPA). The reaction was studied in the
absence of a catalyst (blank) in three solvents (H2O, THF
and IPA) at Table 1 mentioned reaction conditions. The
blank reactions showed < 5% conversion of LA (Table 1,
entry 1–3). Further, two commercial catalysts, i.e. Ni/
SiO2–Al2O3 (Ni/SA) and NiAl alloy were screened for the
reaction. Sels research group used Ni/SA catalyst for selec-
tive preparation of alkylcyclohexanols from alkylguaiacols
and claimed highest selectivity for cyclohexanols [51].
Though activated Ni (Raney-Ni) from NiAl alloy is used for
hydrogenation reactions, being pyrophoric in nature and in
lieu of safety and handling issues, reactions were performed
using ‘as received’ NiAl alloy to assess its performance.
In THF medium both of the catalysts (NiAl alloy and Ni/
SA) at 200 °C for 30 min in the absence of H2 showed no
conversion of LA and THF (Table 1, entry 4 and 5). Fur-
thermore, these two catalytic experiments were conducted
in presence of H2 (15 bar) under same reaction conditions,
and the Ni/SA showed remarkable catalytic activity which
yielded 100% Gvl. The NiAl alloy gave only 7% conversion
of LA with 100% selectivity of Gvl (Table 1, entry 6 and 7).
The same experiment was monitored using water as a sol-
vent and observed that Ni/SA catalyst gave 20% conversion
of LA and here NiAl alloy showed higher catalytic activity
and exhibited 70% LA conversion (Table 1, entry 8 and 9).
Further, the same catalysts were tried for this reaction using
IPA as H-source. In this case, Ni/SA showed 100% Gvl yield
The batch hydrocyclisation/CTH reaction procedure, work-
up, schematic presentation of automated programmable
batch reactor set-up, and the analysis methodology of the
obtained products are mentioned in the ESI (Text 1 and
Fig. 2).
2.3 Continuous Process Hydrocyclisation
Experiments and Analysis
The continuous hydrocyclisation reaction procedure, work-
up, schematic presentation of automated programmable
fixed-bed reactor set-up, and the analysis methodology of
the obtained products are mentioned in the ESI (Text 2 and
Fig. 3).
2.4 Catalyst Characterization
The details of characterization of catalysts using various
physicochemical techniques such as PXRD, BET specific
surface area measurements, TEM, SEM, and NH3-TPD are
mentioned in the ESI (Text 3).
Table 1 Screening of Ni
Entry Catalyst
Solvent Temp. (°C) H2 (bar) Catalyst (mg) LA conv. (%) Gvl yield (%)
catalysts for hydrocyclisation
and CTH of LA to Gvl
1
2
3
4
Blank
Blank
Blank
Ni/SA
H2O
THF
IPA
200
200
200
200
200
200
200
200
200
200
200
15
15
–
–
–
–
3
5
–
0
2
4
–
0
THF
–
110
140
110
140
110
140
110
140
5
6
7
8
9
10
11
NiAl alloy THF
Ni/SA THF
NiAl alloy THF
Ni/SA H2O
NiAl alloy H2O
Ni/SA IPA
NiAl alloy IPA
–
0
0
15
15
15
15
–
100
7
20
70
100
3
100
7
20
70
100
3
–
1 g (8.6 mmol) of LA in 40 mL solvent, 200 °C for 30 min. Ni/SiO2–Al2O3 (Ni/SA)
1 3

S. Gundekari, K. Srinivasan
at 200 °C in 30 min whereas the NiAl alloy catalyst showed
only 3% conversion of LA (Table 1, entry 10 and 11). From
these initial results, we conclude that Ni/SA showed good
catalytic activity for hydrocyclisation of LA to Gvl in THF
solvent using molecular H2 and also the catalyst is active for
catalytic transfer hydrocyclisation (CTH) of LA to Gvl using
IPA as a hydrogen source and/or solvent. Subsequently, the
reaction optimization parameters for hydrocyclisation of LA
in THF medium and CTH of LA in IPA medium were stud-
ied independently using Ni/SA catalyst.
with 100% selectivity of Gvl which means 10 bar H2 is suf-
ficient for the reaction (Fig. 1B). Further, varied the wt%
of catalyst w.r.t LA from 1.75 to 5.25 at 200 °C, 10 bar H2
for 30 min and resulted that 3.5 wt% of catalyst (35 mg of
Ni/SA) catalyst is sufficient for complete conversion of LA
to Gvl (Fig. 1C). Finally, the reaction progress was moni-
tored at different time intervals from 5 to 30 min at 200 °C,
10 bar H2, 35 mg Ni/SA catalyst (Fig. 1D). The reaction
time experiments revealed that 30 min reaction time is suf-
ficient for the complete conversion of LA to Gvl. From the
above mentioned experiments, the final optimized reaction
condition is found as: 200 °C, 10 bar H2, 35 mg of Ni/SA,
and 30 min.
3.1.2 Hydrocyclisation of LA to Gvl Using Ni/SA Catalyst
in THF Medium
For LA to Gvl conversion in THF, the reaction tempera-
ture was varied from 25 to 200 °C at 15 bar H2 for 30 min
and observed that at < 200 °C the catalyst showed < 10%
conversion of LA and at 200 °C showed 100% yield of Gvl
(Fig. 1A). It indicates that the catalyst is active at this tem-
perature only, and most of the reported Ni-based catalysts
for this conversion is also claimed ≥ 200 °C temperature.
Further, varied the H2 pressure from 5 to 20 bar at 200 °C
for 30 min and it revealed that except for 5 bar in all the
cases (10, 15 and 20 bar) observed>99% conversion of LA
3.1.3 CTH of LA to Gvl Using Ni/SA Catalyst in Alcohols
as Hydrogen Source
Initially, varied the temperature from 25 to 200 °C for 30 min
in IPA (40 mL) medium and observed that in this case also
200 °C is optimum towards the reaction that showed 99% yield
of Gvl (Fig. 2A). Further varied the alcoholic hydrogen source
such as ethanol, 1-propanol, IPA, and 1-butanol which resulted
that except IPA, the other alcohols gave less than 5% conver-
sion of LA (Fig. 2B). After the reaction and upon cooling to
Fig. 1 Preparation of Gvl from LA (1 g, 8.6 mmol) using Ni/SA cata-
lyst in THF medium (40 mL). A Temperature variation (25–200 °C)
at 110 mg (11 wt% w.r.t LA) of catalyst, 15 bar H2 for 30 min. B
Pressure variation (5–20 bar H2) at 110 mg (11 wt% w.r.t LA) of cata-
lyst, 200 °C for 30 min. C Catalyst amount variation [17.5–52.5 mg
(1.75–5.25 wt% w.r.t LA)] at 200 °C, 10 bar H2 for 30 min. D Reac-
tion time variation (5–30 min) at 35 mg (3.5 wt% w.r.t LA) of cata-
lyst, 200 °C, 10 bar H2
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
Fig. 2 Preparation of Gvl from LA (1 g, 8.6 mmol) using Ni/SA
catalyst in alcohol medium (hydrogen source); 40 mL. A Tempera-
ture variation (25–200 °C) at 110 mg (11 wt%) of catalyst, for 30 min
in IPA. B Alcohol variation at 110 mg (11 wt%) of catalyst, 200 °C
for 30 min. C Catalyst amount variation [25–110 mg (1.5–7 wt%
w.r.t LA)] at 200 °C for 30 min in IPAD. D Reaction time variation
(5–20 min) at 50 mg (5 wt% w.r.t LA) of catalyst, 200 °C in IPA
room temperature the reaction using IPA showed a pressure of
9 bar, presumably generated through dehydrogenation of IPA.
In other words, Ni/SA catalyst is active for both dehydrogena-
tion of IPA to acetone and hydrogen and, hydrocyclisation of
LA to Gvl using the formed hydrogen from IPA. In addition,
Ni-based catalysts are well-known for dehydrogenation of sec-
ondary alcohols for the production of ketones and hydrogen,
and IPA is one of the largely studied transfer hydrogenation
H-source owing to favourable thermodynamic dehydrogena-
tion pathway [23, 52, 53].
After each cycle, the catalyst was easily recovered from the
product mixture by simple centrifugation and used for the
next cycle without any pre-treatment or solvent washing.
The Ni/SA catalyst showed decreasing catalytic activity in
THF medium where 1st, 2nd, and 3rd reaction cycles gave
the 100, 60, and 42% conversion of LA respectively. The
ICP of catalysts Ni/SA-A (in THF—1st cycle), Ni/SA-B
(in THF—2nd cycle) and Ni/SA-C (in THF—3rd cycle)
revealed 59, 52, and 47 wt% of Ni in the support owing to
leaching of metal during the reaction. The decrease in the
conversion of LA cycle to cycle is proportional to a decrease
of Ni wt% in Ni/SA catalyst, the increasing number of reac-
tion cycles observed a decreasing amount of Ni wt% w.r.t
LA, and accordingly a decrease in the conversion of LA
(Fig. 3A). The instability of Ni/SA catalyst in THF medium
is probably due to direct interaction of LA (as neat or at very
high concentration) with catalyst in liquid phase during the
reaction (leading to leaching of Ni metal due to acidity) as
THF would presumably present largely in the vapour phase
(a pressure increase was seen from 10 bar to 60 bar) owing
to its lesser boiling point (66 °C).
Further, varied the Ni/SA catalyst w.r.t LA from 1.59 to
7 at 200 °C in IPA medium and showed 5 wt% (50 mg of Ni/
SA) catalyst is sufficient for complete conversion of LA to Gvl
(Fig. 2C). Finally, the reaction time varied from 5 to 30 min at
200 °C, 50 mg of Ni/SA in IPA medium indicated that 15 min
time is sufficient for the reaction (Fig. 2D). From the above
mentioned experiments, the final optimized reaction condition
is found as: 200 °C, IPA medium, 50 mg of Ni/SA, and 15 min.
3.1.4 Recyclability of Ni/SA Catalyst for LA to Gvl
Conversion in THF and IPA (Hydrogen Source)
Medium
In contrast to the above results, the Ni/SA catalyst showed
good recyclability in IPA medium showing 98–99% conver-
sion of LA for monitored three reaction cycles. The ICP
analysis of spent catalysts in IPA medium revealed that there
Under the optimized reaction conditions, the recyclability of
Ni/SA catalyst was studied in both medium (THF and IPA).
1 3

S. Gundekari, K. Srinivasan
Fig. 3 Recyclability of Ni/SA catalyst. A In THF medium: 1 g of LA in 40 mL solvent, 70 mg of Ni/SA, 200 °C, 10 bar H2 for 30 min. B In IPA
medium: 1 g of LA in 40 mL solvent, 100 mg of Ni/SA, 200 °C for 30 min
was no leaching of Ni during reuse/recycling of the catalyst
such as Ni/SA-D (in IPA—1st cycle), Ni/SA-E (in IPA—2nd
cycle) and Ni/SA-F (in IPA—3rd cycle), which is the reason
for the stable catalytic activity in contrast to THF medium
(Fig. 3B).
medium (THF and IPA) the conversion of LA to Gvl was too
fast to observe either of the two reported intermediates, thus
precluding the identification of the mechanistic pathway.
To identify the mechanistic pathway, we employed methyl
levulinate (ML) and butyl levulinate (BL) as the substrates
instead of LA under our reaction conditions in both medium
(THF and IPA). If the reaction followed the path a, the con-
version of ML/BL to Gvl would involve transformation of
ML/BL to γ-hydroxy methyl/butyl pentanoate intermediate
which would then convert to Gvl by intramolecular trans-
esterification. The ML reaction product mixture of THF and
IPA (Table 2, entry 1 and 2) contains 3 and 11% yield of
un-cyclised γ-hydroxy methyl pentanoate. In the case of BL,
reaction product mixture of THF and IPA (Table 2, entry 3
and 4) observed 2 and 68% yield of un-cyclised intermediate
(γ-hydroxy butyl pentanoate). The yield (%) of γ-hydroxy
methyl pentanoate intermediate is lesser than Gvl product
but in the case of butyl levulinate reaction in IPA medium
observed high yield (68%) of γ-hydroxy butyl pentanoate
intermediate and which is higher than yield of Gvl (Table 2,
entry 4). The bulkiness of the butyl group (γ-hydroxy butyl
pentanoate) renders slower participation of intra-molecular
trans-esterification as compared to methyl group (γ-hydroxy
methyl pentanoate) in the formation of Gvl which might be
the reason for the observation of higher amount of γ-hydroxy
butyl intermediate in the product mixture.
3.1.5 Hydrocyclisation of LA to Gvl Using Ni/SA Catalyst
in Presence of Formic Acid as Hydrogen Source
We also made an attempt to convert LA to Gvl using Ni/
SA catalyst and formic acid (FA) as hydrogen source. FA is
more appropriate hydrogen source as compared to H2 and
alcohols as, FA is the by-product obtained while prepar-
ing LA from cellulose and thus, it is possible to integrate
it with Gvl preparation by using as obtained product mix-
ture (LA+FA stream). Use of FA as hydrogen source reac-
tions were studied in aqueous medium as the medium is
commonly used for LA and FA preparation from cellulose.
The reaction was conducted with LA (8.62 mmol) and FA
(86.2 mmol) in 40 mL of H2O, 10 wt% of Ni/SA (100 mg)
w.r.t LA, 200 °C for 10 h neither H2 nor inert gas. In this
case also, the catalyst showed good conversion of LA (70%)
with 100% selectivity of Gvl. From the above results, we
conclude that the Ni/SA catalyst is active for both reactions
such as decomposition of FA to H2 and simultaneous hydro-
cyclisation of LA to Gvl using formed H2 from FA.
The above-mentioned reaction intermediates were
analysed by GC–MS (Fig. 4) and 13C-NMR (Fig. 5). The
GC–MS profile of product mixture (Table 2, entry 1 and 2)
showed a peak at retention time of 8.4 min which belongs
to γ-hydroxy methyl pentanoate and it showed molecular
weight of 132 with 117, 101, 88, 85, 57, 55, and 45 fragmen-
tation peaks in the MS profile in which 101 corresponds to
acyl carbocation fragment and is an evidence for γ-hydroxy
methyl pentanoate (Fig. 4A). The product mixtures (Table 2,
entry 3 and 4) showed a peak at 11.6 min retention time
which corresponds to γ-hydroxy butyl pentanoate and it
showed molecular weight of 174 with 159, 130, 101 etc.,
3.1.6 Proposed Mechanism for the Conversion of LA to Gvl
Using Ni/SA Catalyst
Galletti et al. proposed two mechanistic pathways for the
formation of Gvl from LA. The path a involves the hydro-
genation of LA to γ-hydroxy pentanoic acid followed by
intra-molecular esterification to obtain Gvl, and path b in
which the enol form of LA and its cyclisation (intra-molec-
ular esterification) to α-angelica lactone (α-AL) followed
by hydrogenation to Gvl (Fig. 4 in the ESI) [54]. Using Ni/
SA material, under reaction conditions in presence of both
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
Table 2 Mechanistic study for
Entry Reactant Solvent p. H2 (bar) t (min) Conv. (%) γ-OH (MP/ Al yield (%) Gvl yield (%)
Gvl formation
BP) yield
(%)
1
2
3
ML
ML
BL
BL
ML
ML
THF
IPA
THF
IPA
THF
THF
10
–
10
–
–
10
15
15
30
30
30
c.p.
10
66
7
86
–
3
11
2
68
–
8
–
–
–
–
–
–
7
55
5
18
–
40
4
5
6a
48
8.62 mmol of reactants (ML and BL) in 40 mL solvent, 10 wt% of Ni/SA catalyst (100 mg), 200 °C,
15 min
MP methyl pentanoate, BP butyl pentanoate, Al angelica lactone
aContinuous process (c.p): 200 °C, 2 mL/min for H2 gas flow, and 1.5 mL/min of LA (5 wt%) feed in
THF))
Fig. 4 GC–MS profile of product mixture (IPA medium). a Table 2, entry 2 product mixture. b Table 2, entry 4 product mixture
Fig. 5 13C-NMR spectra of product mixture (IPA medium). A Table 2, entry 2 product mixture. B Table 2, entry 4 product mixture
1 3

S. Gundekari, K. Srinivasan
fragmentation peaks in MS profile, and here also with the
101 fragment of acyl carbocation fragment (Fig. 4B). In IPA
medium, the formation of reaction intermediate is higher
than in THF, and thus we analysed the IPA product mix-
ture by 13C-NMR (Table 2, entry 2 and 4). The 13C-NMR
spectrum showed the hydroxy containing carbon peak of
γ-hydroxy methyl pentanoate (Fig. 5A) and γ-hydroxy butyl
pentanoate (Fig. 5B) at 67.1 and 65.4 δ ppm respectively.
This further confirms the reaction takes place through path
a mechanism.
intense peaks which suggest that during the reaction the Ni
particles agglomerated to larger size. The recovered Ni/SA
after each cycle (three cycles) in both medium have same
PXRD patterns without any other diffraction/peaks (Fig. 6A,
B).
The SEM micrograph of the catalysts are provided in
Fig. 7. The SEM images of Ni/SA catalyst appeared as
spherical and plate-like morphologies with regular arrange-
ment (Fig. 7A). The recovered catalyst after 3rd cycle (Ni/
SA-F) also showed similar morphology as that of the fresh
catalyst (Ni/SA), which indicate that during the reaction the
structural morphology of the material/support is not affected
under the reaction conditions by reactants/products/interme-
diates (Fig. 7B). Moreover, the scanning electron micros-
copy-energy dispersive X-ray (SEM-EDX) of Ni/SA showed
59.7% of Ni content and the 3rd recycled catalyst (Ni/SA-F)
also showed almost similar content of Ni (57.3%) (Fig. 7A,
B). The experiments conducted in IPA medium, metal leach-
ing was not observed, and as confirmed by ICP-OES.
The TEM images of catalysts are given in Fig. 8. The
fresh Ni/SA catalyst showed that Ni particles are uniformly
distributed on the external surface of SA support with an
average particle size of 10–20 nm (Fig. 8A, B). The used
and recycled catalyst (Ni/SA-C and Ni/SA-F) also showed
Ni particles in which some of the particles of Ni are larger
in size (>80 nm for Ni/SA-C; 20–35 nm for Ni/SA-F) which
suggest that during the reaction, the Ni particles undergo
agglomeration leading to larger size (Fig. 8C, D). The size
difference of Ni particles was observed in 1st cycle itself,
wherein the PXRD of fresh catalyst (Ni/SA), the Ni shows
broader, but the PXRD of 1st cycle recovered materials (Ni/
SA-A and Ni/SA-D) showed sharp peak which maintained
even after 3rd cycle recovered catalyst (Ni/SA-C and Ni/
SA-F) (Fig. 6).
This observation supported that the path a mechanism is
operating in LA to Gvl conversion over this catalyst. Further,
the LA in presence of Ni/SA material and in the absence of
H2 did not show the formation of α-angelica lactone (LA
remained unconverted), which would be an expected inter-
mediate if the reaction followed path b mechanism (Table 2,
entry 5).
3.1.7 Scale-Up Study for LA to Gvl Conversion Using Ni/SA
Catalyst in IPA Medium (Hydrogen Source)
The catalytic activity of Ni/SA was successfully demon-
strated at 10 g scale of LA (5 wt% of LA in IPA) under
similar reaction conditions which was used for 1 g LA
(200 °C, 0.5 g of Ni/SA catalyst, and 30 min) and achieved
8.5 g (99%) isolated yield of Gvl. After the reaction, Gvl
was recovered from product mixture by simple distillation
using rota-evaporator.
3.2 Characterization of Catalysts (Before and After
Reaction)
The PXRD data of Ni/SA catalyst mentioned in Fig. 6,
showed broad peaks at 2θ values of 44.49° (111) and 51.8°
(200) attributed to metallic Ni, and it matched well matched
with the JCPDS file of Ni (Card No.: 00-004-0850). After
the reaction in both medium (THF and IPA), the Ni dif-
fraction (44.49° and 51.8°) patterns observed as sharp and
The temperature programmed desorption of ammonia
(NH3-TPD) profile of Ni/SA is shown in Fig. 5 in the ESI.
The acid site distributions are mainly classified by tem-
perature range as weak (<250 °C), medium (250–400 °C)
Fig. 6 PXRD of catalysts before and after reaction. A Ni/SA, Ni/SA-A, Ni/SA-B, Ni/SA-C. B Ni/SA, Ni/SA-D, Ni/SA-E, Ni/SA-F
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
Fig. 7 SEM images and EDX profile of catalysts: A Ni/SA, B Ni/SA-F
and strong (>400 °C) acidic sites [55]. The Ni/SA catalyst
showed two desorption peaks at around 285 and 380 °C
indicating the presence of moderate acidic sites. The acid-
ity of the support is useful for the cyclisation of γ-hydroxy
pentatonic acid intermediate to Gvl by the intra-molecular
esterification reaction.
CTH reaction of LA to Gvl using IPA as H-source/medium
wherein the surface area, porosity, and particle size of the
catalyst did not vary significantly while on the contrary in
THF medium along with molecular H2, the used catalyst
showed significant textural variation besides leaching. Thus,
the higher activity and reusability of the catalyst in IPA
medium is rationalized based on metal leaching resistance,
and insignificant changes in the textural properties.
The textural properties (surface area, pore size, and
pore volume) of catalysts are summarized in Table 3. The
BET specific surface area of the Ni/SA was measured as
146 m2/g with pore volume (0.25 cm3/g), and pore diameter
(7.4 nm). The N2 adsorption–desorption isotherms of the Ni/
SA catalyst showed characteristic type-IV isotherm which
was attributed to the capillary condensation of pores with
H1-type hysteresis according to IUPAC classification, and
the pore size indicated in the mesoporous range (Figs. 9A
and 10A). In the case of the used catalyst (Ni/SA-F), a slight
decrease in surface area was observed (121 m2/g) and almost
maintained the pore volume and diameter with characteristic
type-IV isotherm. A sharp decrease in the BET surface area
was observed (80 m2/g from 146 m2/g) for THF medium
used Ni/SA-C catalyst and maintained the pore volume with
small increment in pore diameter (Figs. 9B, C and 10B, C).
The decrease in surface area for the used catalysts (Ni/SA-C
and Ni/SA-F) is probably due to agglomeration of Ni parti-
cles during the reaction (Fig. 8C, D).
3.3 Continuous Production of Gvl from LA Using Ni/
SA Catalyst in THF Medium
Continuous vapor phase conversion of LA to Gvl was suc-
cessfully demonstrated using Ni/SA catalyst in THF medium
(5 wt%). Initially, the powder catalyst was made in to sieves
using Na-CMC as binder at 5 wt% loading. The sieved cata-
lyst was kept as two beds inside the reactor, where graphi-
cal presentation of reactor column and reaction procedure
are mentioned in Fig. 11 and ESI (Text 2). For this study,
several parameters were varied which are mentioned in the
Fig. 12A–D. Liquid feed flow was varied from 1.5 to 2.5 mL/
min by an increment of 0.5 mL/min; temperature from 150
to 250 by an increment of 50 °C; reactor pressure (H2) from
5 to 20 by increment of 5 bar H2; flow rate of H2 feed 100,
25, and 2 mL/min. After screening these parameters, the
optimized reaction condition is 200 °C, 10 bar H2 and 2 mL/
min for H2 gas flow, and under these conditions achieved
From the catalyst characterization results it may be con-
cluded that; the Ni/SA showed good catalytic activity for
1 3

S. Gundekari, K. Srinivasan
Fig. 8 TEM images of catalysts
(before and after reaction): A
and B Ni/SA, C Ni/SA-F, D Ni/
SA-C
Table 3 Textural properties of
Catalyst
BET surface area (m2/g) BJH desorption cumulative pore
volume (cm3/g)
Desorption average
pore diameter (nm)
catalysts
Ni/SA
Ni/SA-C
Ni/SA-F
146
80
121
0.25
0.26
0.25
7.0
10.3
6.6
Fig. 10 Pore size distribution of the catalysts: A Ni/SA, B Ni/SA-F,
C Ni/SA-C
Fig. 9 N2 adsorption–desorption isotherms of catalysts: A Ni/SA, B
Ni/SA-F, C Ni/SA-C
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
98–99% conversion of LA (collected after every 4 h and ana-
lysed) with 100% selectivity of Gvl for 20 h time-on-stream.
Interestingly, the same amount of catalyst taken in a
single-bed showed 90–92% conversion of LA on time-on-
stream; this suggests incomplete interaction of LA mol-
ecules with Ni(0) in onetime/single-bed. However, in two-
bed catalytic system, the missed/escaped LA molecules
from first-bed would interact in the second-bed, thereby
rendering > 98% conversion of LA. The continuous two-
bed catalytic system also followed same mechanistic
pathway as observed under batch conditions involving
hydrogenation followed by intra-molecular esterification
(path-a). While carrying out under continuous phase also,
we observed γ-hydroxy methyl pentanoate intermediate
(8%) when methyl levulinate is used as reactant (Table 2,
entry 6).
Fig. 11 Graphical presenta-
tion for continuous production
of Gvl from LA using Ni/SA
catalyst
Fig. 12 Parameter variation study for continuous production of Gvl
through hydrocyclisation LA using Ni/SA catalyst. A Liquid feed
flow rate variation (2.5–1.5 mL/min) at 200 °C, H2 pressure (25 bar),
and flow rate of H2 feed (100 mL/min). B Temperature variation
(250–150 °C) at flow rate of liquid feed (1.5 mL/min), H2 pressure
(25 bar), and flow rate of H2 feed (100 mL/min). C Reactor H2 pres-
sure variation (20–5 bar) at 200 °C, flow rate of liquid feed (1.5 mL/
min), and flow rate of H2 feed (100 mL/min). D H2 feed flow rate
variation (100–2 mL/min) at 200 °C, flow rate of liquid feed (1.5 mL/
min), and H2 pressure (10 bar)
1 3

S. Gundekari, K. Srinivasan
7. Ghorpade V, Hanna M (1997) Industrial applications for lev-
ulinic acid. In: Campbell GM, Webb C, McKee SL (eds) Cere-
als: novel uses and processes. Springer, Boston, pp 49–55
8. Yan K, Jarvis C, Gu J, Yan Y (2015) Renew Sustain Energy Rev
51:986–997
4 Conclusions
In this work, we have demonstrated successfully the use of
commercial Ni/SiO2–Al2O3 (Ni/SA) as an active catalyst
for the conversion of levulinic acid to γ-valerolactone via
catalytic hydrocyclisation (in tetrahydrofuran medium) using
molecular hydrogen and catalytic transfer hydrocyclisation
using isopropyl alcohol (IPA) as hydrogen source. The cata-
lyst showed excellent activity in both cases with maximum
yield of Gvl (99–100%) under optimised reaction conditions.
The catalyst was found to be stable while using in IPA and
maintained stable conversion and yield for several reaction
cycles. The process was successfully scaled to 10 g of LA
at 5 wt% in IPA and achieved 99% Gvl yield. The nature
of the catalyst (Ni/SA) before and after the reaction was
characterized using PXRD, SEM, TEM, and N2 physical
adsorption techniques. Mechanistic studies through control
experiments using LA esters as reactant showed the forma-
tion of γ-hydroxy alkyl pentanoate intermediates during
the course of the reaction. The intermediates formed were
analyzed by GC-MS and 13C-NMR, which confirmed that
the reaction occurs through hydrogenation of LA followed
by cyclization (intra-molecular esterification) to yield Gvl.
The catalyst was also applied for the continuous produc-
tion of Gvl in THF medium which showed 97–99% yield
up to 20 h time-on-stream. The commercial accessibility
of the catalyst, shorter reaction time and excellent activity
even at higher scale make the catalytic process industrially
attractive.
9. Rackemann DW, Doherty WOS (2011) Biofuels Bioprod Biore-
fin 5:198–214
10. Upare PP, Lee J-M, Hwang YK, Hwang DW, Lee J-H, Halligudi
SB, Hwang J-S, Chang J-S (2011) ChemSusChem 4:1749–1752
11. Mizugaki T, Nagatsu Y, Togo K, Maeno Z, Mitsudome T, Jit-
sukawa K, Kaneda K (2015) Green Chem 17:5136–5139
12. Wright WRH, Palkovits R (2012) ChemSusChem 5:1657–1667
13. Alonso DM, Wettstein SG, Dumesic JA (2013) Green Chem
15:584–595
14. Horváth IT, Mehdi H, Fábos V, Boda L, Mika LT (2008) Green
Chem 10:238–242
15. Alonso DM, Wettstein SG, Mellmer MA, Gurbuz EI, Dumesic
JA (2013) Energy Environ Sci 6:76–80
16. Horváth IT (2008) Green Chem 10:1024–1028
17. Duan Z-Q, Hu F (2012) Green Chem 14:1581–1583
18. Stradi A, Molnar M, Ovari M, Dibo G, Richter FU, Mika LT
(2013) Green Chem 15:1857–1862
19. Strappaveccia G, Ismalaj E, Petrucci C, Lanari D, Marroc-
chi A, Drees M, Facchetti A, Vaccaro L (2015) Green Chem
17:365–372
20. Strappaveccia G, Luciani L, Bartollini E, Marrocchi A, Pizzo F,
Vaccaro L (2015) Green Chem 17:1071–1076
21. Pongrácz P, Bartal B, Kollár L, Mika LT (2017) J Organomet
Chem 847:140–145
22. Fegyverneki D, Orha L, Láng G, Horváth IT (2010) Tetrahedron
66:1078–1081
23. Osatiashtiani A, Lee AF, Wilson K (2017) J Chem Technol Bio-
technol 92:1125–1135
24. Liguori F, Moreno-Marrodan C, Barbaro P (2015) ACS Catal
5:1882–1894
25. Tang X, Zeng X, Li Z, Hu L, Sun Y, Liu S, Lei T, Lin L (2014)
Renew Sustain Energy Rev 40:608–620
26. Zhang J, Chen J, Guo Y, Chen L (2015) ACS Sustain Chem Eng
3:1708–1714
Acknowledgements CSIR-CSMCRI Communication No. 133/2018.
S.G. thanks CSIR, New Delhi, for a Senior Research Fellowship. The
authors thank CSIR, New Delhi for financial support under the projects
OLP-0031, CSC-0123, and MLP-0028. The authors thank Analyti-
cal Division & Centralized Instrumental Facilities of this institute for
analytical support.
27. Pinto BP, Fortuna ALL, Cardoso CP, Mota CJA (2017) Catal Lett
147:751–757
28. Gupta SSR, Kantam ML (2018) Catal Today 309:189–194
29. Hengst K, Schubert M, Carvalho HWP, Lu C, Kleist W, Grun-
waldt J-D (2015) Appl Catal A 502:18–26
30. Fu J, Sheng D, Lu X (2016) Catalysts 6:6
31. Shimizu K-i, Kanno S, Kon K (2014) Green Chem 16:3899–3903
32. Rodiansono, Astuti MD, Hara T, Ichikuni N, Shimazu S (2016)
Catal Sci Technol 6:2955–2961
Compliance with Ethical Standards
Conflict of interest There are no conflicts of interest to declare.
33. Rong Z, Sun Z, Wang L, Lv J, Wang Y, Wang Y (2014) Catal Lett
144:1766–1771
34. Kumar VV, Naresh G, Sudhakar M, Tardio J, Bhargava SK, Venu-
gopal A (2015) Appl Catal A 505:217–223
35. Kumar VV, Naresh G, Sudhakar M, Anjaneyulu C, Bhargava SK,
Tardio J, Reddy VK, Padmasri AH, Venugopal A (2016) RSC Adv
6:9872–9879
References
1. Bozell JJ, Petersen GR (2010) Green Chem 12:539–554
2. Mariscal R, Maireles-Torres P, Ojeda M, Sadaba I, Lopez GM
(2016) Energy Environ Sci 9:1144–1189
36. Varkolu M, Velpula V, Ganji S, Burri DR, Kamaraju SRR (2015)
RSC Adv 5:57201–57210
37. Mohan V, Raghavendra C, Pramod CV, Raju BD, Rama Rao KS
(2014) RSC Adv 4:9660–9668
3. Nakagawa Y, Tamura M, Tomishige K (2013) ACS Catal
3:2655–2668
38. Mohan V, Venkateshwarlu V, Pramod CV, Raju BD, Rao KSR
(2014) Catal Sci Technol 4:1253–1259
4. Primo A, Concepcion P, Corma A (2011) Chem Commun
47:3613–3615
39. Wang L, Han YX, Gao SS, Luo Y, Liu L (2018) React Kinet Mech
Catal 124:389–399
5. Sánchez MA, Torres GC, Mazzieri VA, Pieck CL (2017) J Chem
Technol Biotechnol 92:27–42
40. Jones DR, Iqbal S, Thomas L, Ishikawa S, Reece C, Miedziak
PJ, Morgan DJ, Bartley JK, Willock DJ, Ueda W, Hutchings GJ
(2018) Catal Struct React 4:12–23
6. Yadav M, Mishra DK, Hwang J-S (2012) Appl Catal A
425–426:110–116
1 3

Screening of Solvents, Hydrogen Source, and Investigation of Reaction Mechanism for the…
41. Hengne AM, Kadu BS, Biradar NS, Chikate RC, Rode CV (2016)
RSC Adv 6:59753–59761
49. Hengne AM, Malawadkar AV, Biradar NS, Rode CV (2014) RSC
Adv 4:9730–9736
42. Lv J, Rong Z, Wang Y, Xiu J, Wang Y, Qu J (2015) RSC Adv
5:72037–72045
50. Upare PP, Jeong M-G, Hwang YK, Kim DH, Kim YD, Hwang
DW, Lee UH, Chang J-S (2015) Appl Catal A 491:127–135
51. Schutyser W, Van den Bossche G, Raaffels A, Van den Bosch S,
Koelewijn S-F, Renders T, Sels BF (2016) ACS Sustain Chem
Eng 4:5336–5346
43. Geboers J, Wang X, de Carvalho AB, Rinaldi R (2014) J Mol
Catal A 388–389:106–115
44. Yang Z, Huang YB, Guo QX, Fu Y (2013) Chem Commun
49:5328–5330
52. Shimizu K-i, Kon K, Shimura K, Hakim SSMA (2013) J Catal
300:242–250
45. Obregon I, Gandarias I, Miletic N, Ocio A, Arias PL (2015)
ChemSusChem 8:3483–3488
53. Mears DE, Boudart M (1966) AlChE J 12:313–321
54. Galletti AMR, Antonetti C, De Luise V, Martinelli M (2012)
Green Chem 14:688–694
46. Obregon I, Gandarias I, Al-Shaal MG, Mevissen C, Arias PL,
Palkovits R (2016) ChemSusChem 9:2488–2495
47. Varkolu M, Velpula V, Burri DR, Kamaraju SRR (2016) New J
Chem 40:3261–3267
55. Boréave A, Auroux A, Guimon C (1997) Microporous Mater
11:275–291
48. Varkolu M, Burri DR, Kamaraju SRR, Jonnalagadda SB, van Zyl
WE (2017) Chem Eng Technol 40:719–726
Affiliations
Sreedhar Gundekari1,2 · Kannan Srinivasan1,2
2
* Kannan Srinivasan
Academy of Scientific and Innovative Research,
skannan@csmcri.res.in; kanhem1@yahoo.com
CSIR-Central Salt and Marine Chemicals Research Institute,
GB Marg, Bhavnagar 364 002, India
1
Inorganic Materials and Catalysis Division, CSIR-Central
Salt and Marine Chemicals Research Institute, Council
of Scientific and Industrial Research (CSIR), GB Marg,
Bhavnagar 364 002, India
1 3