An improved catalytic system for the reduction of levulinic acid to γ-valerolactone

Article abstract of DOI:10.1039/c4cy00719k

An improved bidentate phosphine-modified recyclable catalytic system was developed for the selective conversion of biomass-derived levulinic acid into γ-valerolactone with a TOF of 21233 h^-1 in solvent-, chlorine- and additive-free reaction environments. the Partner Organisations 2014.

Full text of DOI:10.1039/c4cy00719k

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
COMMUNICATION
An improved catalytic system for the reduction of
levulinic acid to γ-valerolactone†
Cite this: Catal. Sci. Technol., 2014,
4, 2908
a
b
b
a
*
József M. Tukacs, Márton Novák, Gábor Dibó and László T. Mika
Received 3rd June 2014,
Accepted 3rd July 2014
DOI: 10.1039/c4cy00719k
www.rsc.org/catalysis
An improved bidentate phosphine-modified recyclable catalytic
system was developed for the selective conversion of biomass-
derived levulinic acid into γ-valerolactone with a TOF of 21 233 h−1
in solvent-, chlorine- and additive-free reaction environments.
are relatively low. Thus, high temperature and pressure as
well as a co-catalyst and/or a suitable solvent are necessary to
achieve a satisfactory conversion.13 On the other hand, homo-
geneous catalysts exhibit higher activity and selectivity. Conse-
quently, the development of a simple, highly efficient catalyst
for the reduction of LA to GVL is of utmost importance.14
Leitner and co-workers reported the application of Ru/triphos
catalysts for the efficient and selective production of GVL,
1,4-pentanediol and methyltetrahydrofuran in the presence
of additives such as NH4PF6.15 The pincer ligand-modified
iridium complexes were also found to be active catalysts
(TOF = 1480 h−1) for the hydrogenation of LA in the presence
of a base (e.g. NaOH or KOH) in alcohols (e.g. EtOH,
iPrOH).16
Currently, fossil resources provide more than 95% of our
energy needs and feedstock in the chemical industry. Their
replacement is one of the most challenging tasks faced by
this generation. Especially, the production of carbon-based
chemicals from sustainable resources has become a key issue
for the chemical industry. Intensive research activities on bio-
mass conversion have led to the identification of unique plat-
form molecules such as 5-hydroxymethylfurfural (5-HMF)1
and γ-valerolactone (GVL).2 These small molecules can
replace the currently used fossil-based chemicals or serve as
renewable feedstock for their production. Due to its outstand-
ing physical and chemical properties, GVL was suggested as a
sustainable liquid by Horváth et al. in 2008.2,3 Subsequently,
GVL has been used for the production of transportation
fuels,4 butane isomers,4b octane boosters,2 alkanes,5
2-methyltetrahydrofuran,5 pentane-1,4-diol,5,15a ionic liquids,6
adipic acid,7 alkyl valerates,6a polymers,8 and pentanoic acid9
and as a solvent.10
Recently, similarly active (TOF = 2160 h−1) iridium com-
plexes with 2,2′-bipyridine derivatives have been reported for
the reduction of LA under mild reaction conditions in water
using [LA] = 1 mol dm−3 initial concentration;17 n.b.: in spite
of the high activity, the final product concentration was less
than 1 M and the water-removal step resulted in low energy
Obviously, the most effective protocol to manufacture GVL
is the selective hydrogenation of levulinic acid (LA).11 LA is
another platform molecule which can be obtained via 5-HMF
by acid-catalysed dehydration of hexoses, a major component
derived from biomass (Fig. 1). It is important to note that GVL
can be used as a green solvent for the hydrolysis of fructose, the
precursor of 5-HMF.12 Although, heterogeneous catalysts can
be easily removed from the reaction mixtures, their activities
a Department of Chemical and Environmental Process Engineering,
Budapest University of Technology and Economics, Budapest, Hungary.
E-mail: laszlo.t.mika@mail.bme.hu; Fax: +36 1 463 3197; Tel: +36 1 463 1263
b Institute of Chemistry, Eötvös Loránd University, Budapest, Hungary
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4cy00719k
Fig. 1 Carbohydrate conversion.
2908 | Catal. Sci. Technol., 2014, 4, 2908–2912
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Communication
efficiency of the process. We demonstrated that LA can be
selectively and quantitatively reduced to GVL in the presence
of a catalyst in situ generated from Ru(acac)3 and 10 eq. of
tertiary sulfonated phosphines without using any solvent or
additive with TOF = 1440–3538 s−1 18
It should be noted that
.
this system was also found to be active under continuous
conditions.19
Scheme 1 Bidentate phosphine ligands.
During the decomposition of carbohydrates, LA and
formic acid are formed in equimolar amounts. Thus, formic
acid can be used as a hydrogen source for the reduction of
LA and spontaneous lactone formation to GVL. Horváth and
co-workers reported that LA was converted to GVL by using
a complex catalyst [(η6-C6Me6)Ru(bpy)(H2O)][SO4] at 70 °C.5
Recently, GVL has been quantitatively prepared via transfer
hydrogenation by using HCOOH in the presence of Shvo-type
catalysts such as {[2,5-Ph2-3,4-Ar2-(η5-C5O)]2H}Ru2(CO)4(μ-H)]
[Ar = p-MeOPh, p-MePh, and Ph] at 100 °C without any
additive.20,21 Deng et al. reported a RuCl3/PPh3 catalytic sys-
tem for the transfer hydrogenation of LA to GVL; however,
this required the addition of a base (e.g. pyridine, LiOH, and
NEt3), resulting in GVL with a yield of 16–95%.22
the hydrogenation activity of the Ru catalysts can be sig-
nificantly enhanced by the use of chelating-type bidentate
Ph2P(CH2)nPPh2 (n = 1–3) ligands.23
Thus, we compared the activity of a Ru catalyst modified
with Ph2P(CH2)4PPh2 (Scheme 1, DPPB), a Bu-DPPDS ana-
logue, for the reduction of LA. Quantitative conversion of LA
to GVL for 1 h (TOF = 6370 h−1) under identical conditions
was achieved (Table 1, entry 3). It was shown that the cata-
lytic activity was strongly influenced by the phosphine con-
tent of the reaction mixture. Furthermore, the activity was
also influenced by the number of methylene spacers between
the phosphorus atoms of the bidentate ligands, i.e. the size
of the chelate ring of the active form of the catalyst. Conse-
quently, hydrogenation was repeated at a lower catalyst con-
centration in the presence of various bidentate ligands
(Table 1, entries 4–8). It was shown that the activity was negli-
gible when DPPE was applied, whereas it was similar for the
Bu-DPPDS- and DPPP-modified systems. In the case of DPPB,
quantitative formation of GVL was detected within 1 h
(Table 1, entry 6), exhibiting an outstanding activity (TOF =
7077 h−1). By increasing the number of methylene spacers in
the ligand, i.e. the size of the chelate ring on the catalytically
active species generated in situ from Ru(acac)3 and ligands,
the activity decreased. DPPPe showed a similar activity to
Bu-DPPDS. The activity was below 2000 h−1 in the case of
Here it is proposed that the activity of the Ru-based cata-
lytic systems can be increased by the application of bidentate
phosphine ligands, retaining the environmentally benign
benefits of a solvent-free system.
Firstly, we attempted to reduce LA by Ru(III)-acetylacetonate
in the absence of phosphine ligands. Expectedly, no conver-
sion was detected after 2 h (Table 1, entry 1). Sulfonation of
nBuP(C6H5)2, a malodorous, colorless liquid, resulted in the
formation of the solid nBuP(C6H4-m-SO3Na)2 (Bu-DPPDS)
ligand with similar electronic and steric properties to
nBuP(C6H5)2. For the selective conversion of LA to GVL, the
highest activity of the Ru catalyst modified with a monodentate
phosphine ligand was achieved by applying Bu-DPPDS under
100 bar H2 at 140 °C (Table 1, entry 2).18 It was revealed that
Table 1 Hydrogenation of levulinic acid in the presence of phosphine ligands
Entry
[Ru] × 103 (mol dm−3
)
Ligand
[Ligand] × 103 (mol dm−3
)
T (°C)
P(H2) (bar)
Time (h)
Yield (%)
TON
TOF (h−1
)
1
1.55
1.55
1.55
—
—
140
140
140
140
140
140
140
140
140
140
140
140
140
140
160
120
100
100
100
100
100
100
100
100
100
100
75
50
25
10
10
100
100
100
2.0
1.8
1.0
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
1.8
4.5
0.6
1.8
1.8
0
0
0
3538
6370
125
2a
3a
4
Bu-DPPDS
DPPB
DPPE
DPPPr
DPPB
DPPPe
DPPH
( )-BINAP
DPPB
DPPB
DPPB
DPPB
DPPB
DPPB
DPPB
DPPB
15.5
15.5
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
7.75
>99.9
>99.9
1.8
6370
6370
226
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
0.775
5
6
7
8
57.5
>99.9
51.2
25.3
98.6
>99.9
>99.9
50.6
26.3
70
7325
12 740
6522
3223
12 561
12 740
12 740
6446
3350
8918
12 740
4548
815
4069
7077
3623
1790
6978
7077
7077
3581
1861
1981
21 233
2526
453
9
10
11
12
13
14
15
16
17
>99.9
35.7
6.4
a
Conditions from ref. 18. Reaction conditions: 30 mL (293.02 mmol) of LA, Ru precursor: Ru(acac)3, ligand/Ru = 10. TON is defined as moles
of substrate converted per mole of Ru.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2908–2912 | 2909

View Article Online
Communication
Catalysis Science & Technology
DPPH. Due to the six CH2 units, DPPH could act as a mono-
dentate phosphine. To achieve further improvement in
the TOF by optimizing the ligand concentration in the reac-
tion mixture, we investigated the reduction of LA by using a
Ru/DPPB catalyst in the concentration range of [DPPB] =
0.77 × 10−3–15.4 × 10−3 at a constant Ru content of 0.775 ×
10−3 mol dm−3 (8 × 10−3 mol%). Expectedly, the ligand con-
tent had a strong influence on the formation of GVL as
shown in Fig. 2. When an equimolar amount or 2.5-fold
excess of the ligand was used, 4.5 and 12% conversions were
detected. A significant increase in the rate were achieved at
DPPB/Ru ratios between 5 and 10 followed by a slight
decrease at higher ratios. The concentration dependence
curve shows a maximum at [DPPB] = 7.75 1.0 × 10−3 mol dm−3,
which corresponds to a ligand/Ru ratio of 10 1.0. This is in
good agreement with our previous report for Bu-DPPDS18 and
ligand excess favored reduction of carbonyl groups.24 The
ligand exchange equilibrium between the catalyst species
and the free ligand—depending on the pressure, tempera-
ture, etc.—determines the optimal ligand concentration in
the reaction mixture. Furthermore, due to the presence of a
bidentate ligand, the reversible arm-off dissociation cannot
be excluded.25 The GC and NMR measurements of the con-
version of 293.02 mmol of LA in the presence of the catalyst
in situ formed from Ru(acac)3 (0.023 mmol) and DPPB
(0.23 mmol) proved that no by-product, including over-
hydrogenated products such as 2-methyltetrahydrofuran,
could be detected. The 31P-NMR spectrum of the final reac-
tion mixture did not indicate the formation of phosphine
oxide and any sign of decomposition of the ligand. These
types of ligands are solid, thus no sulfonation was necessary.
Finally, the binaphthyl derivative ( )-BINAP, which could also
form a seven-member chelate ring with a Ru center similarly
to DPPB, was also investigated. However, the activities were
similar to those obtained using the less expensive DPPB
(Table 1, entry 6).
Ru/DPPB system (Table 1, entries 6, 10–14). As was shown
for the Ru/DPPDS system,18 the conversion rate was not
affected by the reduction of the pressure to 50 bar for 1.8 h;
however, upon further decreasing the pressure (25 and
10 bar), the reaction rates decreased significantly.
The investigation of the effect of reaction temperature
(Fig. 3, Table 1, entries 6, 15–17) showed that the conversions
were negligible at 100 °C and modest at 120 °C. By increasing
the temperature to 160 °C, full conversion was achieved
after 0.6 h, corresponding to TOF = 21 233 h−1. To our best
knowledge, so far it is the highest TOF value obtained for
the selective homogeneous hydrogenation of LA to GVL.
The recyclability of the Ru(acac)3/DPPB catalyst was
investigated under batch conditions using [Ru] = 1.55 ×
10−3 mol dm−3 (0.16 mol%) and [DPPB] = 0.15 mol dm−3
under 100 bar of H2 at 140 °C.18 Full conversion of LA to GVL
was detected in the first run resulting in a light orange solu-
tion after 30 min (ESI,† Picture S1). The volatile compounds
were removed by vacuum transfer (4–6 mmHg) using a 20 cm
column resulting in a clear solution. The orange glue-like res-
idue (ESI† Picture S2) was completely dissolved in 30 mL
(293.02 mmol) of LA. Expectedly no phosphorus compounds
were detected in the distillate. The 1H-NMR and GC
analysis proved that the distillate was pure GVL. The light
orange solution (ESI,† Picture S3) was transferred to the reac-
tor and then pressurized to 100 bar H2. The reaction was
started by increasing the temperature up to 140 °C. A signifi-
cant pressure drop indicated the reduction process. After a
similar work-up of the initial run, the second yield was
>99.9% GVL and was unchanged in the next 8 runs (ESI,†
Fig. S1, Table S1, and Picture S4). The water content of the
combined 10 distillates was removed by vacuum distillation
(5 mmHg) using 10 cm of Sulzer EX-20 package at 40 °C, resulting
in 247.5 g (2.47 mol) of colorless GVL with a water content of
98 ppm and 36.3 g of water. This result showed that the
improved Ru/BDPP catalytic system was stable enough with
the catalyst and ligand amounts and reaction conditions
applied here for successful recycling including the low-
The effect of hydrogen pressure on the GVL formation
was also determined in the range of 10–100 bar using the
Fig. 2 Turnover frequencies (TOF50% = moles of LA per mole of Ru
per hour at 50% conversion) at different DDPB concentrations.
Reaction conditions: [Ru] = 0.775 × 10−3 mol dm−3 (8 × 10−3 mol%),
100 bar H2, 140 °C.
Fig.
3 Conversion of LA to GVL at different temperatures.
Conditions: [Ru] = 0.775 × 10−3 mol dm−3 (8 × 10−3 mol%), [DPPB] =
7.75 × 10−3 mol dm−3, 100 bar H2.
2910 | Catal. Sci. Technol., 2014, 4, 2908–2912
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Communication
pressure separation step. Since 100% conversion was
achieved, it cannot be stated whether the catalyst lost any
activity or not during the course of the reactions.26
straw-yellow organic phase was dried over MgSO4. The sol-
vent was removed under reduced pressure at 60 °C resulting
in ca. 1.3 mL (1.57 g) of a dark brown liquid as the product.
Yield: 51%.
In order to investigate the applicability of the catalyst to
the conversion of LA derived directly from the dehydration
of carbohydrates, 4 g (22.2 mmol) of D-fructose was dissolved
in 80 mL of 2 M H2SO4 and heated at 170 °C for 8 h.27 After
our published work-up procedure was performed,26 ca.
1.3 mL (1.57 g) of LA was obtained as a dark brownish-yellow
solution (ESI,† Picture S5). This represents a slightly higher
yield of LA than that obtained from the microwave-assisted
dehydration of fructose (400 mg).26 However, humin forma-
tion cannot be excluded from the acid-catalysed dehydration
process as described by Horváth et al.12 In 1 mL of this solu-
tion, 6.2 mg of Ru(acac)3 and 66.5 mg of DPPB was dissolved,
after which the solution was transferred to a 10 mL Parr HP
reactor. The mixture was pressurized to 100 bar of H2 and
Conclusions
We have demonstrated that the catalyst in situ generated
from Ru(III)-acetylacetonate and 1,4-bis(diphenylphosphino)
butane (DPPB) can be used for the efficient conversion of
biomass-derived levulinic acid to gamma-valerolactone with a
representative TOF of 21 233 h−1. The catalyst can be recycled
for ten consecutive runs while full conversion of LA was
achieved. The maximum hydrogenation rate was achieved
by applying a 10-fold excess of the DPPB ligand to the
Ru precursor.
1
heated to 140 °C. By using H NMR and GC monitoring, it
Acknowledgements
was found that complete conversion was achieved after 1 h
(ESI,† Fig. S2 and S3). To summarize, an improved catalytic
system was developed and successfully used for the hydroge-
nation of “real” bio-derived levulinic acid.
The authors would like to thank Prof. László Kollár (University
of Pécs, Pécs, Hungary) for helpful discussion and sugges-
tions. This work was funded by the Budapest University
of Technology and Economics under project number
KMR_12-1-2012-0066. The authors are grateful to the support
of János Bólyai Research Fellowship of the Hungarian
Academy of Sciences and the scholarship of Campus Hungary
Program.
Experimental
Levulinic acid, 1,2-bis(diphenylphosphino)ethane (DPPE), 1,3-
bis(diphenylphosphino)propane (DPPP), 1,4-bis(diphenylphosphino)
butane (DPPB), 1,5-bis(diphenylphosphino)pentane (DPPPe),
1,6-bis(diphenylphosphino)hexane (DPPH), D-fructose, and
( )-BINAP were purchased from Sigma-Aldrich and used as
received. GC analyses were performed on an Agilent 6890N
instrument with an HP-INNOWax capillary column (15 m ×
0.25 mm × 0.25 μm) using H2 as a carrier gas. For the analy-
sis, 10 μL of the reaction mixture was added to 1 mL of
methylene chloride followed by the addition of 10 μL of
toluene as an internal standard.
In a typical hydrogenation experiment, the 120 mL Parr HP
reactor was charged with 34.02 g (30 mL, 293.02 mmol) of
levulinic acid followed by the addition of Ru(acac)3 and the
corresponding phosphine ligand, resulting in a light red solu-
tion. The reaction mixture was pressurized to the desired
values and heated up to 140 °C. Samples were taken for off-
line GC analysis via a dip-leg into a sample holder. After the
given reaction time, the autoclave was cooled to ambient tem-
perature and stirring was stopped.
Notes and references
1 P. Gallezot, Chem. Soc. Rev., 2012, 41, 1538.
2 I. T. Horváth, H. Mehdi, V. Fábos, L. Boda and L. T. Mika,
Green Chem., 2008, 10, 238.
3 V. Fábos, G. Koczó, H. Mehdi, L. Boda and I. T. Horváth,
Energy Environ. Sci., 2009, 2, 767.
4 (a) J. Q. Bond, D. Martin Alonso, R. M. West and J. A. Dumesic,
Langmuir, 2010, 26, 16291; (b) J. Q. Bond, J. A. Dumesic,
D. M. Alonso, D. Wang and R. M. West, Science, 2010,
327, 1110.
5 H. Mehdi, V. Fábos, R. Tuba, A. Bodor, L. T. Mika and
I. T. Horváth, Top. Catal., 2008, 48, 49.
6 (a) D. Fegyverneki, L. Orha, G. Láng and I. T. Horváth,
Tetrahedron, 2010, 66, 1078; (b) A. Strádi, M. Molnár,
M. Óvári, G. Dibó, F. U. Richter and L. T. Mika, Green Chem.,
2013, 15, 1857.
To prepare “real” bio-based levulinic acid, 4 g of D-fructose
was dissolved in 80 mL of 2 M H2SO4. The solution was trans-
ferred to a 120 mL Parr HP HC reactor equipped with a PID
temperature controller (Parr 4843), safety relief valve, sam-
pling line and manometer. When the temperature reached
170 °C, the pressure increased to 18 bar. After 8 h, the mix-
ture was cooled and the dark brownish-black reaction mix-
ture was filtered through a glass filter. The black solid was
washed with 3 × 30 mL of distilled water and with 3 × 30 mL
of ethyl acetate. The aqueous phase was separated and
extracted with 4 × 100 mL of ethyl acetate. The combined
7 P. K. Wong, C. Li and L. Stubbs, Patent Appl., WO 2012/
134397A1, 2012.
8 M. Chalid, H. J. Heeres and A. A. Broekhuis, J. Appl. Polym.
Sci., 2011, 123, 3556.
9 J.-P. Lange, R. Price, P. M. Ayoub, J. Louis, L. Petrus,
L. Clarke and H. Gosselink, Angew. Chem., Int. Ed., 2010,
49, 4479.
10 (a) I. T. Horváth, Green Chem., 2008, 10, 1024; (b) L. Qui and
I. T. Horváth, ACS Catal., 2012, 2, 2247.
11 B. Girisuta, L. P. B. M. Janssen and H. J. Heeres, Chem. Eng.
Res. Des., 2006, 84, 339.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2908–2912 | 2911

View Article Online
Communication
Catalysis Science & Technology
12 L. Qui, G. Akien and I. T. Horváth, Chem. Commun., 2012,
48, 5850.
19 J. M. Tukacs, R. V. Jones, F. Darvas, G. Dibó, G. Lezsák and
L. T. Mika, RSC Adv., 2013, 3, 16283.
13 W. R. H. Wright and R. Palkovits, ChemSusChem, 2012,
5, 1657.
20 V. Fábos, I. T. Horváth, N. Kaposy and H. Mehdi, Hungarian
Patent Application P0900276, 2009.
14 J. G. de Vries, P. J. Deuss and K. Barta, Catal. Sci. Technol.,
2014, 4, 1174.
21 V. Fábos, L. T. Mika and I. T. Horváth, Organometallics,
2014, 33, 181.
15 (a) F. M. A. Geilen, B. Engendahl, A. Harwardt,
W. Marquardt, J. Klankermayer and W. Leitner, Angew.
Chem., Int. Ed., 2010, 49, 5510; (b) F. M. A. Geilen,
B. Engendahl, M. Hölscher, J. Klankermayer and W. Leitner,
J. Am. Chem. Soc., 2011, 133, 14349.
16 W. Li, J.-H. Xie, H. Lin and Q.-L. Zhou, Green Chem., 2012,
14, 2388.
17 J. Deng, Y. Wang, T. Pan, Q. Xu, Q.-X. Guo and Y. Fu,
ChemSusChem, 2013, 6, 1163.
22 L. Deng, J. Li, D.-M. Lai, Y. Fu and Q.-X. Guo, Angew. Chem.,
Int. Ed., 2009, 48, 6529.
23 O. Kröcher, R. A. Köppel and A. Baiker, Chem. Commun.,
1997, 453.
24 A. Bényei and F. Joó, J. Mol. Catal., 1990, 58, 151.
25 G. Kiss and I. T. Horváth, Organometallics, 1991, 10, 3798.
26 Homogeneous Catalysts: Activity - Stability – Deactivation,
ed. J. C. Chadwick, R. Duchateau, Z. Freixa and P. W. N. M.
van Leeuwen, Wiley-VCH, Boschstr. 12, 69469 Weinheim,
Germany, 2011, pp. 2-50.
18 J. M. Tukacs, D. Király, A. Strádi, G. Novodárszki, Z. Eke,
G. Dibó, T. Kégl and L. T. Mika, Green Chem., 2012,
14, 2057.
27 Á. Szabolcs, M. Molnár, G. Dibó and L. T. Mika, Green
Chem., 2013, 15, 439.
2912 | Catal. Sci. Technol., 2014, 4, 2908–2912
This journal is © The Royal Society of Chemistry 2014