Highly selective and efficient solvent-free transformation of bio-derived levulinic acid to γ-valerolactone by Ru(II) arene catalyst precursors

Article abstract of DOI:10.1016/j.ica.2018.06.050

The selective and efficient solvent-free hydrogenation of bio-based levulinic acid (LA) to γ-valerolactone (GVL) was achieved with new pyridylimine ruthenium(II) complexes as catalyst precursors. The hydrogenation reactions were performed in the presence of formic acid as hydrogen source using a catalyst loading as low as 0.1 mol% with potassium hydroxide or triethylamine (Et₃N). 4-Hydroxyvaleric acid (HVA) was produced only when KOH was used, whereas reactions involving Et₃N were selective to GVL. At 150 °C, >96% LA conversions were achieved with 100% GVL selectivity. Recyclability of catalyst precursors was demonstrated by running three consecutive reactions where 100% conversion and selectivity was maintained. In-situ NMR studies show that hydrogen gas is formed by the decomposition of formic acid to carbon dioxide and hydrogen. Ru-hydride species have been detected, by ¹H NMR, and are believed to be the catalytically active species, and a mechanism of the reaction has been proposed.

Full text of DOI:10.1016/j.ica.2018.06.050

Accepted Manuscript
Research paper
Highly Selective and Efficient Solvent-free Transformation of Bio-derived Lev-
ulinic acid to γ-Valerolactone by Ru(II) Arene Catalyst Precursors
Novisi K. Oklu, Banothile C.E. Makhubela
PII:
S0020-1693(18)30831-4
https://doi.org/10.1016/j.ica.2018.06.050
ICA 18342
DOI:
Reference:
Inorganica Chimica Acta
To appear in:
Received Date:
Revised Date:
Accepted Date:
29 May 2018
27 June 2018
28 June 2018
Please cite this article as: N.K. Oklu, B.C.E. Makhubela, Highly Selective and Efficient Solvent-free Transformation
of Bio-derived Levulinic acid to γ-Valerolactone by Ru(II) Arene Catalyst Precursors, Inorganica Chimica Acta
(2018), doi: https://doi.org/10.1016/j.ica.2018.06.050
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.

Highly Selective and Efficient Solvent-free Transformation of
Bio-derived Levulinic acid to γ-Valerolactone by Ru(II) Arene
Catalyst Precursors
Novisi K. Oklua and Banothile C. E. Makhubelaa*
aDepartment of Chemistry, University of Johannesburg, Auckland Park Kingsway Campus
P.O. Box 524, South Africa, Johannesburg
*E-mail: bmakhubela@uj.ac.za
ABSTRACT
The selective and efficient solvent-free hydrogenation of bio-based levulinic acid (LA) to γ-
valerolactone (GVL) was achieved with new pyridylimine ruthenium(II) complexes as
catalyst precursors. The hydrogenation reactions were performed in the presence of formic
acid as hydrogen source using a catalyst loading as low as 0.1 mol% with potassium
hydroxide or triethylamine (Et3N). 4-Hydroxyvaleric acid (HVA) was produced only when
KOH was used, whereas reactions involving Et3N were selective to GVL. At 150 °C, >96 %
LA conversions were achieved with 100 % GVL selectivity. Recyclability of catalyst
precursors was demonstrated by running three consecutive reactions where 100 % conversion
and selectivity was maintained. In-situ NMR studies show that hydrogen gas is formed by the
decomposition of formic acid to carbon dioxide and hydrogen. Ru-hydride species have been
1
detected, by H NMR, and are believed to be the catalytically active species, and a
mechanism of the reaction has been proposed.
KEYWORDS
Solvent-free reactions; formic acid; hydrogen generation; levulinic acid; gamma-
valerolactone; arene ruthenium(II) complexes
ABBREVIATIONS
γ-Valerolactone: GVL
Levulinic acid: LA
4-Hydroxyvaleric acid: 4-HVA
Triethylamine: Et3N
Formic acid: FA
Hydroxymethylfurfural: HMF
1

1. INTRODUCTION
Liquid fuels as well as most chemicals and polymers that are currently used are derived from
petroleum. However, due to high demand for petroleum products, crude oil reserves are
heading towards rapid depletion [1]. The problems of environmental pollution and climate
impacts often linked to the burning of fossil resources (crude oil, natural gas and coal), are
also of concern [2]. Over the past 50 years, atmospheric carbon dioxide concentrations have
risen by about 30 % and this has been accompanied by a rise in other Green House Gases,
including CH4 and NO2. This has caused about 0.6 °C increase in the average temperature of
the earth’s surface [3]. As such, renewable alternative sources of energy and chemical
resources need to be developed to meet the growing demands [2]. One of such potential
alternative feedstock is biomass [4]. Non-edible, lignocellulosic or ‘wood-based’, biomass
has gained attention recently. It is a complex carbohydrate polymer, which is mainly made up
of cellulose (41 %), hemicellulose (28 %) and lignin (27 %) [4,5]. This lignocellulosic
biomass can be transformed into many chemicals including ethanol, n-butanol, sorbitol,
hydroxymethylfurfural, dimethylfuran and γ-valerolactone (GVL) [6,7].
GVL has gained much attention over the past few decades due to its attractive physical and
chemical characteristics as well as its unique fuel characteristics [8]. It has low toxic levels,
an acceptable and definitive smell which makes detection of leaks and spills easy, high flash
point (96 °C), high boiling (207 °C) point, low melting point (-31 °C), and above all, it is
obtained from a renewable source. It is also used as an additive in the food industry [9]. This
has made GVL to stand out as one of the most promising renewable platform molecules that
can be transformed into various chemicals, such are valeric acid, 1,4-pentanediol, 2-methyl
tetrahydrofuran and 2-butene. In turn, GVL serves as an intermediate in the production of a
broad range of biofuels as well as commodity and fine chemicals [7,10].
GVL can be produced from cellulose and sugars by first hydrolyzing waste cellulose to
glucose followed by glucose ↔ to fructose isomerization. Fructose can then be converted to
hydroxymethylfurfural (HMF) via dehydration. This is followed by rehydration of the HMF
to levulinic acid (LA). LA is further hydrogenated to GVL using metal catalysts with an
external hydrogen source, or formic acid (FA) as the source of hydrogen (Scheme 1). Using
FA as the source of hydrogen is more economical in this case because hydrolysis of glucose
2

produces FA in equimolar amounts to LA. In the presence of a catalyst, FA decomposes into
CO2 and H2 and hence can be used as an internal source of hydrogen (Scheme 1) [11].
Scheme 1: Production of GVL from LA using FA or molecular hydrogen [5-9].
Recently, much effort has been put into the development of heterogeneous catalytic systems,
such as Ru, Ni, Rh, Ir, Pt and Pd on carbon and oxides, for the conversion of LA to GVL
[12–14]. However, it has been shown that for the hydrogenation of bio-sourced molecules,
including the conversion of LA to GVL, ruthenium metal particles supported on various
carbons and oxides, such as 5 % Ru/C, 5 % Ru/SiO2 and 5 % Ru/Al2O3, were the most
efficient catalysts to achieve a fast and selective conversion [15,16]. With regards to
ruthenium metal particles supported on oxides, Luo and co-workers studied the influence that
supports of different acidities (Nb2O5, TiO2, H-ZSM5 and H-β) have on the activity and
selectivity of the hydrogenation of LA using ruthenium, and found that supports with higher
amounts of Lewis acid sites gave higher conversions [17].
Even though heterogeneous catalysts have the advantages of easy separation and recycling of
the catalyst, the homogeneous catalysts have very efficient activities and are also highly
selective under relatively mild conditions [9,12]. In the transformation of LA to GVL, water-
soluble homogeneous catalysts have been increasingly attractive. Especially because GVL
3

does not form an azeotropic mixture with water, hence the catalyst can be recovered for
reuse, through distillation [18].
Furthermore, studies have shown that homogeneous catalysts have higher activities in the
decomposition of FA to produce hydrogen gas [19]. Some of the metal complexes used in
this regard include those with Pd [19], Ir [20,21], Fe [22], Rh [23] and Ru [24] metal centers.
Deng et. al., reported the use of FA as a source of hydrogen for the conversion of LA to GVL
in the presence of a mixture of RuCl3 (0.2 mol%) and PPh3 (0.6 mol%) catalyst. They
proposed that the formic acid is decomposed to hydrogen gas which is then used for the
hydrogenation of the LA [25] as opposed to going through the transfer hydrogenation process
with Shvo catalysts proposed by Horvath et al. [26] Recently, we reported the transformation
of LA to GVL with FA as a source of hydrogen in the presence of pyrazolylphophinite and
pyrazolylphosphite-ruthenium (II) complexes bearing p-cymene auxiliary ligand. In situ
1HNMR studies showed that the reaction proceeds by the production of molecular hydrogen
and carbon dioxide gases from FA [27]. Hence, it can be deduced that homogeneous
ruthenium catalysts have the added advantage of both facilitating the production of hydrogen
gas from FA as well as converting LA to GVL.
While much research has been carried out on the development of water-soluble phosphines,
hydrophilic nitrogen ligands have received relatively little attention [28]. It has however been
reported that nitrogen donor ligands, such as Schiff bases in particular, are able to stabilize
many different metals in different oxidation states and as such control the activity of metals
in various transformations [29,30]. A recent study reported the use of half-sandwich
ruthenium(II) complexes with Schiff base ligands in the transfer hydrogenation of ketones to
alcohols. Sodium formate was employed as the hydrogen source for the reaction which
resulted in yields greater than 95 % [31a]. In view of this, ruthenium hydrogenation catalyst
systems, especially those derived from nitrogen donor ligands, may be of interest in the
conversion of LA to GVL. In this work, we report the synthesis of new Schiff base N^N
donor ligands and their use in the solvent-free hydrogenation of LA to GVL, using FA as a
source of hydrogen. The highly selective synthesis of GVL was achieved by using
pyridylimine ruthenium(II) complexes as catalyst precursors. The use of FA (which is
produced as a by-product in the manufacture of LA) [31b] as the source of hydrogen and the
absence of solvent, make this an appealing and green way for GVL production.
4

2. MATERIALS AND METHODS
2.1 General information
4-aminobenzoic acid (99 %), 2-pyridinecarboxylic acid (99 %), 2-quinolinecarboxaldehyde
(97 %), 2-aminopyridine (98 %), 4-formylbenzoic acid (97 %), 4-(aminomethyl)-benzoic
acid (97 %), dichloro(p-cymene)ruthenium(II)dimer, 1,5-cyclooctadiene (COD) (98 %),
RuCl3·xH2O, levulinic acid (97 %) and formic acid (95 %) were all purchased from Sigma-
Aldrich and were used as supplied. All solvents used were of analytical grade and were dried
using MBRAUN SPS-800 solvent drying system. L1 [33] and L2 [34] were prepared
according to literature methods. The metal precursor [RuCl2(COD)]n was synthesized
1
following reported literature method [32]. H NMR (400 MHz) and 13C{1H} NMR (100
MHz) spectra were recorded on a Bruker-400 MHz spectrometer and values were reported
relative to tetramethylsilane (δ 0.0) as internal standard. FT-IR spectra were recorded using a
Perkin
Elmer
FT-IR
Spectrum
BX-ATR.
Elemental
analyses
were
performed on a Thermo Scientific FLASH 2000 CHNS-O analyzer. HR-MS (ESI) spectra
were recorded on a Waters Synapt G2 spectrometer. Melting points were determined using a
Gallenkamp digital melting point apparatus. All hydrogenation reactions were performed in
PPV-CTRO1-CE high pressure reactor vessels fitted into a high pressure autoclave reactor
with-in built stirring, heating and cooling systems.
2.2 Synthesis of (E)-4-((quinolin-2-ylmethylene)amino)benzoic acid (L3)
2-Quinolinecarboxaldehyde (0.510 g, 3.24 mmol) was dissolved in hot dry ethanol (15 ml),
followed by the addition of 4 drops of formic acid. 4-Aminobenzoic acid (0.231 g, 1.68
mmol) was then added. The solution formed was refluxed for 2 h after which the solvent was
removed by using a rotary evaporator to afford a brown solid. Dry dichloromethane (15 ml)
was then added to the solid to form a suspension. The suspension was filtered and the residue
washed with dry dichloromethane (10 ml) to afford a pale brown solid which was dried
overnight under vacuum. Yield: 0.328g (70.50 %); M.p: 210 - 213 °C; FT-IR (νmax/cm-1):
2892 b (OH), 1708 s (C=O), 1597 s (HC=N), 1566 w (C=N); 1H NMR (400 MHz, DMSO-d6,
3
25 °C, δ, ppm): 10.10 (s, 1H, -OH), 8.70 (s, 1H, imine-CH), 8.50 (d, JHH = 8.80 Hz, 1H,
3
3
quin-CH), 8.30 (d, JHH = 8.80 Hz, 1H, quin-CH), 8.10 (d, JHH = 6.40 Hz, 1H, quin-CH),
3
3
8.08 (d, JHH = 8.00 Hz, 1H, quin-CH), 8.03 (d, JHH = 8.40 Hz, 2H, aromatic-CH), 7.80 (t,
3
3
3JHH = 6.80 Hz, 1H, quin-CH), 7.70 (t, JHH = 7.20 Hz, 1H, quin-CH), 7.40 (d, JHH = 8.40
Hz, 2H, aromatic-CH); 13C{1H} NMR (100 MHz, DMSO-d6, 25 °C, δ, ppm): 166.92 (-
5

COOH), 162.55 (imine-C=N), 154.19 (aromatic-C-N), 154.01, 147.38 (quin-C=N), 137.18
(quin-CH), 130.67 (aromatic-CH), 130.37 (quin-CH), 129.30 (quin-CH), 128.92, 128.58
(aromatic-C(COOH)), 128.20 (quin-CH), 128.14 (quin-CH), 121.23 (aromatic-CH), 118.28
(quin-CH); CHN-calculated: (73.90 % C, 4.38 % H, 10.14 % N), CHN-obtained: (73.47 % C,
4.24 % H, 9.81 % N); HR-MS (ESI+) C17H13N2O2 calculated, m/z = 277.0977 [M+H]+, found,
m/z = 277.0967 [M+H]+; Solubility: DMSO, hot methanol and ethanol.
2.3 Synthesis of [Ru(p-cymene)Cl(L1)]Cl (1)
A mixture of L1 (122.16 mg, 0.540 mmol) and [Ru(p-cymene)Cl2]2 (165.36 mg, 0.270
mmol) in dry methanol (20 mL) was transferred into a schlenk tube. The tube was deaerated
with nitrogen gas and the solution stirred at room temperature for 20 h. After 20 h, the
solvent was removed by rotary evaporation to afford an orange solid product which was dried
overnight in vacuo. Yield: ( 270 mg, 93.9 %); M.p: decomposes without melting (onset at
1
180 °C); FT-IR (νmax/cm-1): 3380 b (OH), 1702 s (C=O), 1600 s (HC=N); H NMR (400
3
MHz, MeOD-d4, 25 °C, δ, ppm): 9.53 (d, JHH = 5.60 Hz, 1H, pyr-CH), 8.84 (s, 1H, imine-
3
CH), 8.31 – 8.25 (m, 4H, pyr-CH, pyr-CH, aromatic-CH), 7.89 (d, JHH = 8.40 Hz, 2H,
3
3
aromatic-CH), 7.86 (t, JHH = 5.60 Hz, 1H, pyr-CH), 6.00 (d, JHH = 6.00 Hz, 1H, p-cym-
CH), 5.69 (d, 3JHH = 6.00 Hz, 1H, p-cym-CH), 5.64 (d, 3JHH = 6.00 Hz, 1H, p-cym-CH), 5.50
(d, 3JHH = 6.40 Hz, 1H, p-cym-CH), 2.60 (m, 3JHH = 7.20 Hz, 1H, p-cym-CH(CH3)2), 2.25 (s,
3
3H, p-cym-CH3), 1.09 (d, JHH = 6.80 Hz, 6H, p-cym-CH3); 13C{1H} NMR (100 MHz,
DMSO-d6, 25 °C, δ, ppm): 206.41 (-COOH), 168.94 (imine-C=N), 166.48 (pyridyl-C=N),
156.15 (pyr-CH), 154.72, 154.38 (aromatic-C(COOH)), 139.94 (pyr-CH), 130.64 (aromatic-
CH), 130.50 (pyr-CH), 129.11 (pyr-CH), 122.76 (aromatic-CH), 105.24 (p-cym-C(CH3)),
103.71, 86.67 (p-cym-CH), 86.00 (p-cym-CH), 84.87 (p-cym-CH, p-cym-CH), 30.45 ((p-
cym-C(CH3)2)), 21.67 (p-cym-CH3, p-cym-CH3), 18.26 (p-cym-CH3); CHN-calculated:
(51.88 % C, 4.56 % H, 5.26 % N), CHN-obtained: (52.26 % C, 4.53 % H, 4.96 % N); HR-
MS (ESI+) [C23H24ClN2O2Ru]+ calculated, m/z = 497.0570 [M]+, found, m/z = 497.0570
[M]+; Solubility: water, methanol, ethanol.
2.4 Synthesis of [Ru(p-cymene)Cl(L2)]Cl (2)
The synthesis was performed as for 1 by using L2 (129.60 mg, 0.540 mmol) and [Ru(p-
cymene)Cl2]2 (165.36 mg, 0.270 mmol). Yield: (0.237 g, 80.34 %); M.p: decomposes without
melting (onset at 175 °C); FT-IR (νmax/cm-1): 3385 b (OH), 1705 s (C=O), 1611 s (HC=N),
6

1473 w (C=N); 1H NMR (400 MHz, DMSO-d6, 25 °C, δ, ppm): 13.133 (s, 1H, -COOH), 9.55
3
(d, JHH = 5.20 Hz, 1H, pyr-CH), 8.51 (s, 1H, imine-CH), 8.24 – 8.19 (m, 2H, pyr-CH, pyr-
CH), 8.02 (d, 3JHH = 8.0 Hz, 2H, aromatic-CH), 7.79 (t, 3JHH = 6.00 Hz, 1H, pyr-CH), 7.61 (d,
3
3
3JHH = 8.00 Hz, 2H, aromatic-CH), 6.25 (d, JHH = 6.00 Hz, 1H, p-cym-CH), 6.06 (d, JHH
=
5.60 Hz, 1H, p-cym-CH), 5.90 (d,3JHH = 6.8 Hz, 1H, p-cym-CH), 5.88 (d,3JHH = 6.4 Hz, 1H,
p-cym-CH), 5.80 (d, 3JHH = 16.00 Hz, 1H, -CH2), 5.56 (d, 3JHH = 16 Hz, 1H, -CH2), 2.49 (m,
3
1H, p-cym-CH(CH3)2), 2.08 (s, 3H, p-cym-CH3), 0.98 (d, JHH = 6.80 Hz, 3H, p-cym-CH3),
3
13
0.90 (d, JHH = 6.80 Hz, 3H, p-cym-CH3); C{1H} NMR (100 MHz, DMSO-d6, 25 °C, δ,
ppm): 206.41 (-COOH), 168.39 (imine-C=N), 167.46 (-C(CH2)), 156.50 (pyr-CH), 154.91
(pyridyl-C=N), 139.49 (pyr-CH), 131.48 (aromatic-C(COOH)), 130.62 (aromatic-CH),
130.31 (aromatic-CH), 129.87 (pyr-CH), 128.84 (pyr-CH), 105.08 (p-cym-C(CH3)), 103.76,
87.97 (p-cym-CH), 84.96 (p-cym-CH), 84.83 (p-cym-CH), 84.73 (p-cym-CH), 68.71 (-CH2),
30.89 (p-cym-C(CH3)2), 22.51 (p-cym-CH3), 21.94 (p-cym-CH3), 18.80 (p-cym-CH3); CHN-
calculated: (52.75 % C, 4.80 % H, 5.13 % N), CHN-obtained: (53.06 % C, 4.65 % H, 5.28 %
N); HR-MS (ESI+) [C24H26ClN2O2Ru]+ calculated, m/z = 511.0726 [M]+, found, m/z =
511.0737 [M]+; Solubility: water, methanol, ethanol.
2.5 Synthesis of [Ru(p-cymene)Cl(L3)]Cl (3)
Complex 3 was prepared following a procedure similar to that reported for 1 by using L3
(99.48 mg, 0.360 mmol) and [Ru(p-cymene)Cl2]2 (110.24 mg, 0.180 mmol). Yield: (193 mg,
91.90 %); M.p: decomposes without melting (onset at 187 °C); FT-IR (νmax/cm-1): 3387 b
(OH), 1706 s (C=O), 1593 s (HC=N), 1513 w (C=N); 1H NMR (400 MHz, MeOD-d4, 25 °C,
δ, ppm): 9.09 (s, 1H, imine-CH), 8.85 (d, 3JHH = 8.40 Hz, 1H, quin-CH), 8.82 (d, 3JHH = 8.80
3
3
Hz, 1H, quin-CH), 8.33 (d, JHH = 8.40 Hz, 1H, quin-CH), 8.32 (d, JHH = 8.40 Hz, 2H,
aromatic-CH), 8.28 (d, 3JHH = 8.40 Hz, 1H, quin-CH), 8.20 (t, 3JHH = 7.60 Hz, 1H, quin-CH),
3
3
8.06 (d, JHH = 8.80 Hz, 2H, aromatic-CH), 8.01 (t, JHH = 7.20 Hz, 1H, quin-CH), 6.03 (d,
3
3
3JHH = 6.40 Hz, 1H, p-cym-CH), 5.89 (d, JHH = 6.00 Hz, 1H, p-cym-CH), 5.77 (d, JHH
=
6.40 Hz, 1H, p-cym-CH), 5.32 (d, 3JHH = 6.00 Hz, 1H, p-cym-CH), 2.38 (m, 3JHH = 7.20 Hz,
3
1H, p-cym-CH(CH3)2), 2.29 (s, 3H, p-cym-CH3), 0.99 (d, JHH = 6.80 Hz, 3H, p-cym-CH3),
3
13
0.84 (d, JHH = 6.80 Hz, 3H, p-cym-CH3); C{1H} NMR (100 MHz, DMSO-d6, 25 °C, δ,
ppm): 206.92 (-COOH), 170.34 (imine-C=N), 166.98 (pyridyl-C=N), 156.09, 155.55
(aromatic-C(COOH)), 148.92, 141.37 (quin-CH), 133.93 (quin-CH), 132.51, 131.27
(aromatic-CH), 130.99 (quin-CH), 129.80 (quin-CH, quin-CH), 125.39 (quin-CH), 123.23
7

(aromatic-CH), 106.01 (p-cym-C(CH3)), 105.54, 87.01 (p-cym-CH), 86.82 p-cym-CH), 85.98
(p-cym-CH), 85.09 (p-cym-CH), 30.95 (p-cym-C(CH3)2), 22.51 (p-cym-CH3), 21.43 (p-cym-
CH3), 18.78 (p-cym-CH3); CHN-calculated: (55.67 % C, 4.50 % H, 4.81 % N), CHN-
obtained: (56.00 % C, 4.56 % H, 4.75 % N); HR-MS (ESI+) [C27H26ClN2O2Ru]+ calculated,
m/z = 547.0726 [M]+, found, m/z = 547.0725 [M]+; Solubility: water, methanol, ethanol.
2.6 Synthesis of [Ru(COD)Cl2(L1)] (4)
A mixture of L1 (81.4 mg, 0.36 mmol) and the ruthenium polymer (100 mg, 0.36 mmol) in
dry ethanol (10 mL) was transferred into a schlenk tube. The tube was deaerated with
nitrogen gas and the mixture refluxed with stirring for 2 h. The solution was then reduced to
about a third of the original volume and kept overnight at 4 °C to allow precipitation. The
resulting dark solution was carefully decanted off to afford a brown precipitate. Cold ethanol
(10 mL) was added to the precipitate and then filtered off. Further washing of the brown
precipitate was done with cold ethanol (10 mL) followed by diethyl ether (10 mL). The
brown precipitate was dissolved in 5 mL chloroform followed by the addition of excess n-
hexane (20 mL) to induce precipitation of the product. Finally, the precipitated product was
obtained by filtration and washed with n-hexane (10 mL) to afford a brown solid product
which was dried overnight under vacuum. Yield: (158 mg, 87.00 %); M.p: decomposes
without melting (onset at 260 °C); FT-IR (νmax/cm-1): 3035 b (OH), 1721 s (C=O), 1598 s
1
(HC=N), 1497 w (C=N); H NMR (400 MHz, DMSO-d6, 25 °C, δ, ppm): 13.10 (s, 1H, -
3
COOH), 8.90 (s, 1H, imine-CH), 8.30 (d, JHH = 5.20 Hz, 1H, pyr-CH), 8.28 (d, 3JHH = 6.80
3
3
Hz, 1H, pyr-CH), 8.23 (t, JHH = 7.20 Hz, 1H, pyr-CH), 8.00 (d, JHH = 8.00 Hz, 2H,
3
3
aromatic-CH), 7.70 (t, JHH = 6.00 Hz, 1H, pyr-CH), 7.30 (d, JHH = 8.40 Hz, 2H, aromatic-
CH), 4.50 (s, 2H, cod-CH), 3.80 (s, 2H, cod-CH), 2.00 (m, 3H, cod-CH2), 1.90 (m, 3H, cod-
CH2); 13C{1H} NMR (100 MHz, DMSO-d6, 25 °C, δ, ppm): 170.79 (imine-C=N), 166.71
(pyridyl-C=N), 156.17, 152.50 (aromatic-C(COOH)), 150.91 (pyr-CH), 139.13 (pyr-CH),
130.30 (aromatic-CH), 129.78 (pyr-CH), 128.67 (pyr-CH), 121.03 (aromatic-CH), 90.20
(COD-CH), 89.73 (COD-CH), 29.33 (COD-CH2), 28.93 (COD-CH2); CHN-calculated:
(49.81 % C, 4.38 % H, 5.53 % N), CHN-obtained: (49.33 % C, 4.39 % H, 5.16 % N); HR-
MS (ESI+) C21H21N2O2Ru calculated, m/z = 435.0725, found, m/z = 435.0639 [M-H-2Cl]+;
Solubility: DMSO.
8

2.7 General procedure for hydrogenation reactions
Levulinic acid (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol / 0.1 mol %), and base
(Et3N/KOH) (20 mmol) were added to an autoclave reactor. The mixture was heated to the
desired temperature after purging four times with nitrogen gas. The mixture was then left to
stir for the required length of time. At the end of the reaction, the reactor vessel was cooled
and the gas generated released. A sample of the mixture was then analysed by 1H NMR. All
hydrogenation reactions were carried out in triplicates.
3. RESULTS AND DISCUSSION
3.1 Synthesis and characterisation of ligands L1 to L3
1
The ligands were prepared as shown in (Scheme 2) and characterised using H NMR,
1
13C{1H} NMR, elemental analysis (CHN) and mass spectroscopy. The H NMR spectrum of
the compounds show characteristic signals for the imine proton between 8.20 and 8.60 ppm.
This was further corroborated by absorption bands between 1590 and 1690 cm-1 in the
infrared spectra of L1 to L3. The IR spectra of the ligands also show absorption bands
between 1560 and 1590 cm-1 which correspond to the C=N stretching frequency of the
pyridyl moiety.
Scheme 2: Outline of the syntheses of ligands L1 to L3.
3.2 Synthesis and characterisation of complexes 1 to 3
Complexes 1 – 3 were synthesized by the reaction of the appropriate ligands (L1 to L3) with
[RuCl2(p-cymene)]2 at room temperature for 20 h. Complex 4 was synthesized by refluxing
L1 with [RuCl2(1,5-cyclooctadiene)]n in methanol for 24 h (Scheme 3). All complexes were
9

1
13
characterised using H and C{1H} NMR spectroscopy, elemental analysis (CHN), infrared
1
spectroscopy and mass spectrometry. The H NMR spectrum of the complexes show
significant shifts of the signal for the imine protons as well as those adjacent to the pyridyl
nitrogen. The shifts in these proton signals serve as a confirmation that both nitrogen atoms
have coordinated to the ruthenium centre by displacing a chloride which serves as a counter
ion to the complex. The signals of the four aromatic protons of the p-cymene moiety in
complexes 1–3 were expected to appear as two doublets if the p-cymene sits on a
perpendicular plane to the metal. However, the signals appeared as four separate peaks each
appearing as a doublet. This could be due to the fact that the p-cymene moiety loses its
symmetry upon coordination of the N^N donor ligand to the metal. A similar trend has been
reported in literature, where an imine and amine nitrogen atoms were coordinated to a
ruthenium with a p-cymene ligand [35].
Scheme 3: Outline for the syntheses of complexes 1 to 4.
In addition, the methylene protons adjacent to the imine nitrogen in complex 3 have split into
two separate peaks appearing at 5.80 ppm and 5.56 ppm. This can also be attributed to loss of
symmetry. The [M]+ peaks, which correspond to the cationic part of the complexes, were
10

observed in the mass spectra of complexes 1 to 3. In the mass spectrum of complex 4, a peak
is observed at m/z = 435.0639 [C21H21N2O2Ru]+, which corresponds to the mass of 4 minus
hydrochloric acid and a chloride ligand. This indicates that, unlike 1-3, this compound
fragments easily under the conditions that the mass spectra were collected.
3.3 Hydrogenation of LA with formic acid
The reactions were performed under solvent-free conditions by reacting levulinic acid (20
mmol), formic acid (20 mmol), catalyst (0.02 mmol), and KOH (20 mmol) in an autoclave
reactor at 150 °C for 12 h (Table 1). Catalyst precursor 4 gave the highest conversion (96 %)
followed by 3 (93 %), 2 (59 %) and 1 (45 %) (Table 1, entries 1 – 4). The intermediate, 4-
HVA, was also produced in addition to GVL, with pre-catalysts 1, 2, 3 and 4 giving 4-HVA
selectivities of 66 %, 34 %, 58 % and 47 % respectively. The reactions were repeated with a
reduced amount of KOH (2 mmol) which resulted in reduced conversions (Table 1, entries 5
– 8). Pre-catalyst 3 recorded the highest conversion (42 %) followed by 2 (24 %) as well as 4
and 1 with conversions of 20 % and 14 % respectively. Interestingly, when the amount of
KOH was reduced to 2 mmol, the only product detected was GVL. This shows that the pre-
catalysts are selective to only GVL at lower amounts of KOH. However, higher conversion of
LA can be achieved at higher amounts of KOH but with the production of both GVL and the
4-HVA intermediate.
Table 1: Hydrogenation of LA using formic acid and KOH.
LA Conversion (%)b
GVL Selectivity (%)b
4-HVA Selectivity (%)b
Entry Cat.c
1
2
45
59
93
96
14
24
42
34
66
66
34
58
47
0
1
2
3
4
1
2
3
3
42
4
53
5a
6a
7a
100
100
100
0
0
11

8a
20
100
0
4
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), KOH (20 mmol), 12 h, 150
1
°C.[25] [a] 2 mmol of KOH was added. [b] Conversion and Selectivity determined by H NMR
spectroscopy.[c] Cat. = catalyst precursor; Average error estimates: ±0.44 (1), ±0.49 (2), ±0.53
(3), ±0.57 (4).
When Et3N was used (Table 2, entries 1 – 4), all pre-catalysts gave conversions above 96 %
with 3 and 4 giving the highest conversion (97 %). The reaction was also repeated with less
amount of Et3N (2 mmol) (Table 2, entries 5 – 8) but the conversions observed did not differ
much from when 20 mmol of Et3N was used. Pre-catalyst 2 only showed the lowest
conversion (36 %) when small amount of Et3N was used. Pre-catalyst 3 gave the highest
conversion (98 %) followed by 4 and 1 with 93 % conversions each. This shows that the pre-
catalysts only require small amounts of Et3N in performing the transformation of levulinic
acid to GVL selectively.
Table 2: Hydrogenation of LA using formic acid and Et3N.
LA Conversion (%)e
GVL Selectivity (%)e
4-HVA Selectivity (%)e
Entry Cat.f
1
2
96
96
97
97
93
36
98
93
100
100
100
100
100
100
100
100
0
0
0
0
0
0
0
0
1
2
3
4
1
2
3
4
3
4
5d
6d
7d
8d
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (20 mmol), 12 h, 150
1
°C.[25] [d] 2 mmol of Et3N was added. [e] Conversion and Selectivity determined by H NMR
spectroscopy. [f] Cat. = catalyst precursor; Average error estimates: ±0.52 (1), ±0.49 (2), ±0.37
(3), ±0.47 (4).
12

The results from these preliminary screening show that, for the selected conditions, the pre-
catalysts perform better with organic bases than with inorganic bases, possibly due to lack of
solubility of the KOH. Wang et.al., reported 99 % yield with 1 equivalent Et3N and 2
equivalent FA in 16 h, when ruthenium catalysts bearing dipyridyl-amine ligands were used
to synthesize GVL from LA and formic acid [36]. Our catalysts require only 0.1 equivalent
Et3N together with 1 equivalent FA to achieve yields from 96 – 97 % in 12 h. The reactions
were also carried out without the presence of a base [37a] and the conversions of LA for each
catalyst was poor, indicating that the base is necessary for the reaction. Also, when the
reaction was repeated without catalyst [37a], only 10 % conversion of LA was observed. This
confirms that the hydrogenation reactions are being catalyzed by the ruthenium. We chose
catalytic amounts of base (0.1 equivalent Et3N) over stoichiometric amounts (1 equivalent) as
the optimal amount of base for the hydrogenation reactions and this was used for all
subsequent reactions.
3.4 Conversion as a function of time
The reactions were performed using 0.1 equivalents Et3N for 2 h, 4 h, 8 h, 12 h and 16 h. All
pre-catalysts have an induction period of 2 to 4 h (Figure 1). The conversions with 1, 3 and 4
gradually increased after 4 h to greater than 90 % in 16 h, with 4 being the highest (99 %).
Pre-catalyst 2, however, has a shorter induction period and recorded a conversion of 65%
even after16 h.
13

98
92
99
97
100
90
80
70
60
50
40
30
20
10
0
93
92
84
60
65
1
3
2
4
56
36
28
36
23
18
20
18
14
8
25
0
5
10
15
20
Time (h)
Figure 1: Time dependence studies of the hydrogenation of LA using formic acid.
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (2 mmol) and
150 °C. ; Average error estimates: ±0.24 (1), ±0.49 (2), ±0.33 (3), ±0.42 (4).
The low activity of 2 can be explained by comparing its structural difference to 1. The only
difference between the two compounds is the added methylene group between the imine
nitrogen and the aromatic group in 3. It is possible that the degree of freedom about the
methylene group may cause flexibility, therefore allowing the aromatic group to fold over
and prevent easy access to the ruthenium active center. This steric bulk around the ruthenium
center could hinder the coordination of the formate to the metal center, which is the first step
of the reaction. It was interesting to observe 100 % selectivity to GVL regardless of how long
the reaction had proceeded for all the catalyst precursors (Figure 2). From the above results,
16 h was taken as the optimum time for further reactions.
14

Figure 2: 1H NMR spectra of hydrogenation reactions with 4
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (2 mmol),
150 °C.
3.5 Effect of temperature
Varying the temperature from 125 °C to 175 °C (Figure 3) resulted in increased conversion
for all pre-catalysts, while the selectivity was maintained at 100 % GVL. At 125 °C, all pre-
catalysts gave conversions below 30 %, with the exception of 3 which gave a conversion of
88 %. At 175 °C, 3 and 4 gave complete conversions of LA to afford GVL. Thus, pre-catalyst
3 is the best performing catalyst since it maintains activity even at a lower temperature of 125
°C.
15

100
90
80
70
60
50
40
30
20
10
0
125 °C
150 °C
175 °C
1
2
3
4
Catalyst Precursor
Figure 3: Effects of temperature on the hydrogenation of LA using formic acid.
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (2 mmol),
16 h.
3.6 Mercury poisoning tests
Homogeneity test was performed for pre-catalysts 3 and 4 in the hydrogenation of LA to
GVL for 16 h at 175 °C using metallic mercury, where 2 mg of metallic mercury was
dropped in the reaction vessel. This was done to poison any nanoparticles that might be
formed, during the reaction, and subsequently promote the conversion. Our experiments show
that there is no significant change in conversion for both pre-catalysts in the presence of
mercury. LA conversions with 3 and 4 dropped from 100 % to 98 % and 97 % respectively
(Figure 4), thus confirming that the active species generated from the pre-catalysts, are of a
homogeneous nature.
16

100
90
80
70
60
50
40
30
20
10
0
without Hg(0)
with Hg(0)
3
4
Catalyst Precursor
Figure 4: Mercury drop test of 3 and 4 in the hydrogenation of LA using formic.
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (2 mmol),
Hg(0) (2 mg), 16 h, 175 °C.
3.7 Recyclability of 3 and 4
Recyclability of 3 and 4 (which were the best performing catalyst precursors) was
investigated at 175 °C for 16 h (Figure 5). At the end of the reaction, the crude mixture was
dissolved in ethanol and transferred into a Schlenk tube. The resulting solution was
evaporated off under vacuum at 90 °C, leaving behind the catalyst [26]. Ethanol was then
used to wash the catalyst back into the autoclave reactor, followed by drying in a vacuum
oven at 40 °C. After the ethanol was completely removed, the reactor was recharged with
LA, FA and Et3N. The mixture was then heated at 175 °C for 16 h. The procedure was
repeated until the fifth run. For both 3 and 4, 100 % conversion of LA and 100 % selectivity
to GVL was maintained until the third run. However, for 3, the conversion reduced to 75 %
and 22 % in the fourth and fifth runs respectively, whereas that of 4 was reduced to 97 % and
61 % in the fourth and fifth runs respectively. This effectively indicates that for both pre-
catalysts, deactivation occurs after the fourth run with the rate of deactivation occurring faster
for 3 than 4. It is well known that ruthenium-arene complexes bearing chloride ligands
readily undergo chloride ligand substitution with aqua ligands in the presence of water.
Therefore, deactivation of the pre-catalyst may be due the formation of inactive oxo-bridging
17

species [35,37b], which can easily be accessed from pre-formed ruthenium aqua complexes,
where the water is derived from the cyclization of 4-HVA to GVL.
100
90
80
70
60
50
3
40
4
30
20
10
0
run 1
run 2
run 3
run 4
run 5
Number of runs
Figure 5: Recyclability test of 3 and 4 in the hydrogenation LA using formic.
Conditions: LA (20 mmol), formic acid (20 mmol), catalyst (0.02 mmol), Et3N (2 mmol), 16
h, 175 °C.
3.8 In-situ NMR studies
The catalytic reaction was monitored by performing a small scale reaction in MeOD-d4 in a J
Young NMR tube. Levulinic acid, formic acid, Et3N and pre-catalyst 1 were loaded into the J
1
Young NMR tube and heated in an oil bath at 125 °C. After 30 min, the H NMR spectrum
showed the formation of hydrogen gas as a result of decomposition of formic acid [37b]. The
chemical shift at 8.12 ppm, which was from the formic acid, disappeared completely after 30
min, with a simultaneous appearance of a new singlet at 4.47 ppm which corresponds to
hydrogen gas. Deng et al., also reported the formation of hydrogen gas after 20 min, when
they reacted in situ generated active catalytic species from RuCl3·3H2O and PPh3 with LA
and FA in the presence of Et3N, albeit without 1H NMR evidence [25]. 1H NMR evidence of
the decomposition of formic acid has, however, been reported with Pd [19] and Ru [27]
catalysts. In addition to the formation of hydrogen gas, there was cleavage of the Ru – imine
nitrogen bond [37c]. There was disappearance of the signals of the pre-catalyst 1 and the
18

appearance of new peaks upfield. Also, the aromatic protons of the p-cymene which
originally appeared as four separate peaks now coalesced at around 5.55 – 5.59 ppm. This
bond cleavage could be to allow formate ion to coordinate to the metal center as a first step in
formic acid decomposition. Studies have shown that the catalytic decomposition of formic
acid in the absence of tertiary amines is very slow [38]. Another small scale reaction was
performed by loading formic acid, Et3N and pre-catalyst 1 into a J Young NMR tube, and this
was heated in an oil bath at 125 °C. Again, the 1H NMR spectrum [37d] showed the complete
disappearance of the chemical shift of the formic acid which initially appears as a singlet at
8.10 ppm. A new singlet at 4.45 ppm also appeared, corresponding to hydrogen gas.
Additionally, the same peak patterns that indicate the cleavage of the Ru-imine nitrogen bond
were observed. This goes on to support that the hydrogenation reaction proceeds by the
ruthenium catalyzed decomposition of formic acid to generate hydrogen gas. Evidence of Ru-
hydride species formation was observed in the negative region of the 1H NMR (Scheme 4).
3.9 Proposed reaction mechanism
The proposed mechanism for the hydrogenation reaction is shown in Scheme 4. Formate ion
is first formed by the deprotonation of formic acid by Et3N. In Pathway A, the formate ion
coordinates to the metal center of pre-catalyst 1 after the Ru–imine nitrogen bond cleavage.
This is the first step in the Ru-catalyzed formic acid decomposition to form carbon dioxide
and hydrogen gas. Molecular hydrogen then coordinates to the metal centre as shown in 1(I)
which can quickly form a Ru-hydride species 1(III), with the expulsion of HCl. Ru-assisted
transfer of hydrogen to the ketone end of levulinic acid, to from 1(VI) is followed by the
formation of intermediate 1(V). After 4-HVA and CO2 are released a second molecule of
formate coordinates to the coordinatively unsaturated species 1(VI), this then produces
intermediate 1(II) again and leads to regeneration of the active species 1(III). 4-HVA
undergoes lactonization to form GVL with the loss of a water molecule. In Pathway B, an
anion metathesis between the formate ion and the chloride ion on the ruthenium center of pre-
catalyst 1 occurs. Once the metathesis has taken place, triethylammonium chloride is released
and intermediate 1(II) is formed. A hydride is abstracted from the formate which is
coordinated to ruthenium via its charged oxygen atom. This results in the active species
1(III) which then undergoes the same reaction steps as in Pathways A.
19

Scheme 4: Proposed catalytic mechanism.
3.10 Aqueous media hydrogenation of LA
The process of transformation of cellulose to levulinic acid results in the formation of an
aqueous mixture of formic acid and levulinic acid (Scheme 1) [39]. 15 wt% aqueous solution
of glucose was converted to LA (42 wt% ) and formic acid (17 wt%), by acid catalyzed
dehydration with 0.8 M HCl at 220 °C [25]. This was done with the aim of achieving direct
conversion of glucose to GVL in aqueous media. Once the LA and formic acid has formed, it
would be easy to then add the hydrogenation catalyst without having to remove water. For
this to be practical, the pre-catalyst would have to be effective in aqueous media. As such,
20

we then conducted the hydrogenation reactions in 50 wt% distilled water at 125 °C for 16 h
[37e]. Conversions of levulinic acid decreased as compared to when the reactions were
performed solvent-free. A further decrease in conversions was observed when the
temperature was increased to 175 °C (Figure 6).
100
90
80
70
60
50
40
30
20
10
0
Solvent-free (125 °C)
125 °C (50 wt% water)
175 °C (50 wt% water)
1
2
3
4
Catalyst Precursors
Figure 6: Aqueous hydrogenation of LA using formic acid.
These reductions in conversions may be due to the aquation of the chloride leaving groups on
the pre-catalysts which could then be followed by the formation of inactive oxo-bridging
species [35,37b]. This aquation of the chloride ligand may further be enhanced at elevated
temperatures as is evidence from when the reactions were performed at 175 °C [37b]. Since
pre-catalysts 1, 2 and 3 are soluble in water, it is unlikely that the reduction in conversions
might be due to solubility issues.
4. CONCLUSIONS
We have shown that easily prepared (in one step) pyridylimine ligands can be used to
synthesize new ruthenium(II) complexes, that are active pre-catalysts for the solvent-free
selective hydrogenation of levulinic acid (LA) to γ-valerolactone (GVL) by using formic acid
(FA) as the hydrogen source. This is practical since the source of hydrogen, FA, is a by-
product in the production of LA. The hydrogen gas was derived from ruthenium-catalysed
decomposition of FA and a mechanism for the catalytic reaction has been proposed. The
21

efficiency of these reactions under solvent-free conditions makes the current hydrogenation
reactions green. In addition to this, the catalysts can be recycled up to three times while
maintaining 100 % LA conversion and GVL selectivity.
ACKNOWLEDGEMENTS
This work was supported by the University of Johannesburg’ Centre for Synthesis and
Catalysis, the National Research Foundation of South Africa (NRF) (Grant Numbers: 99269
and 105556) and SASOL LTD. The use of UJ Spectrum’s NMR facilities is also
acknowledged.
DECLARATIONS OF INTEREST: We declare no conflict of interest in financial support
for this work
REFERENCES
[1] J.J. Bozell, Science 329 (2010) 522–523.
[2] Y. Gong, L. Lin, Z. Yan, BioResources 6 (2011) 686–699.
[3] G.M. Joselin Herbert, A. Unni Krishnan, Renew. Sustain. Energy Rev. 59 (2016) 292–
308.
[4] (a) K.C. Wu, in:, B. Trewyn (Ed.), Heterog. Catal. Today’s Challenges Synth. Charact.
Appl., The Royal Society of Chemistry, 2015, pp. 1–27. (b) B.C.E. Makhubela, J.
Darkwa, Johnson Matthey Technol. Rev. 62 (2018) 4-31.
[5] X. Zhuang, W. Wang, Q. Yu, W. Qi, Q. Wang, X. Tan, G. Zhou, Z. Yuan, Bioresour.
Technol. 199 (2015) 68–75.
[6] M. Mascal, E.B. Nikitin, Green Chem. 12 (2010) 370–373.
[7] S. Dutta, S. Pal, Biomass and Bioenergy 62 (2014) 182–197.
[8] K. Yan, Y. Yang, J. Chai, Y. Lu, Appl. Catal. B Environ. 179 (2015) 292–304.
[9] I.T. Horváth, Green Chem. 10 (2008) 1024–1028.
[10] P. Gallezot, Chem. Soc. Rev. 41 (2012) 1538–1558.
[11] M. Wachala, J. Grams, W. Kwapinski, A.M. Ruppert, Int. J. Hydrogen Energy 41
(2016) 8688–8695.
[12] M.J. Climent, A. Corma, S. Iborra, Green Chem. 16 (2014) 516–547.
[13] L.E. Manzer, Appl. Catal. A Gen. 272 (2004) 249–256.
[14] M.G. Al-shaal, W.R.H. Wright, R. Palkovits, Green Chem. 14 (2012) 1260–1263.
22

[15] C. Michel, P. Gallezot, ACS Catal. 5 (2015) 4130–4132.
[16] A.M. Hengne, N.S. Biradar, C. V. Rode, Catal. Letters 142 (2012) 779–787.
[17] 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.
[18] J. Deng, Y. Wang, T. Pan, Q. Xu, Q.X. Guo, Y. Fu, ChemSusChem 6 (2013) 1163–
1167.
[19] C. Ortiz-Cervantes, M. Flores-Alamo, J.J. García, ACS Catal. 5 (2015) 1424–1431.
[20] J.H. Barnard, C. Wang, N.G. Berry, J. Xiao, Chem. Sci. 4 (2013) 1234–1244.
[21] W. Li, J.-H. Xie, H. Lin, Q.-L. Zhou, Green Chem. 14 (2012) 2388–2390.
[22] P.D. Gabor Laurenczy, Céline Fellay, J. Am. Chem. Soc. 132 (2010) 8924–8934.
[23] S. Fukuzumi, T. Kobayashi, T. Suenobu, ChemSusChem 1 (2008) 827–834.
[24] A. Boddien, B. Loges, H. Junge, F. Gärtner, J.R. Noyes, M. Beller, Adv. Synth. Catal.
351 (2009) 2517–2520.
[25] L. Deng, J. Li, D. Lai, Y. Fu, Q. Guo, Angew. Chemie - Int. Ed. 48 (2009) 6529–6532.
[26] V. Fábos, L.T. Mika, I.T. Horváth, Organometallics 33 (2014) 181–187.
[27] G. Amenuvor, B.C.E. Makhubela, J. Darkwa, ACS Sustain. Chem. Eng. 4 (2016)
6010–6018.
[28] G.J.P. Britovsek, G.Y.Y. Woo, N. Assavathorn, J. Organomet. Chem. 679 (2003) 110–
115.
[29] G. Venkatachalam, R. Ramesh, Inorg. Chem. Commun. 9 (2006) 703–707.
[30] T. Opstal, F. Verpoort, Angew. Chemie - Int. Ed. 42 (2003) 2876–2879.
[31] (a) S. Azizi Talouki, G. Grivani, P. Crochet, V. Cadierno, Inorganica Chim. Acta
(2016) 1–7. (b) Caserta Production Plant: Innovation and R&D’, GFBIochemicals,
Milan, Italy: http://www.gfbiochemicals.com/company/#caserta-plant (Accessed on 12
April 2018)
[32] M.O. Albers, T.V. Ashworth, H.E. Oosthuizen, E. Singleton, J.S. Merola, R.T.
Kacmarcik, Inorganic Syntheses, John Wiley and Sons, 1989.
[33] B.P. Buffin, E.B. Fonger, A. Kundu, Inorganica Chim. Acta 355 (2003) 340–346.
[34] B. Roberts, S.J. Harris, M.T.W. Hearn, R.I. Boysen, Chemical Compound Arrays
Useful for Detecting Bacterial Peptides, WO2011140599 A1, 2011.
[35] C. Gaiddon, W.H. Ang, M.J. Chow, C. Licona, D.Y.Q. Wong, G. Pastorin, J. Med.
Chem. 57 (2014) 6043–6059.
[36] S. Wang, V. Dorcet, T. Roisnel, C. Bruneau, C. Fischmeister, Organometallics 36
(2017) 708–713.
23

[37] (a) See electronic annex Table S1 (b) É.A. Enyedy, G.M. Bognár, T. Kiss, M. Hanif,
C.G. Hartinger, J. Organomet. Chem. 734 (2013) 38–44. (c) See electronic annex l
Figure S1. (d) See electronic annex Figure S2. (e) See electronic annex Table S2
[38] T. Zell, B. Butschke, Y. Ben-David, D. Milstein, Chem. - A Eur. J. 19 (2013) 8068–
8072.
[39] D. Garce, D. Eva, S. Ordo, Ind. Eng. Chem. Res. 56 (2017) 5221–5230.
24

Highly Selective and Efficient Solvent-free Transformation of
Bio-derived Levulinic acid to γ-Valerolactone by Ruthenium
Arene Catalyst Precursors
Novisi K. Oklua and Banothile C. E. Makhubelaa*
aDepartment of Chemistry, University of Johannesburg, Auckland Park Kingsway Campus
P.O. Box 524, South Africa, Johannesburg
*E-mail: bmakhubela@uj.ac.za
GRAPHICAL ABSTRACT
25

Highly Selective and Efficient Solvent-free Transformation of
Bio-derived Levulinic acid to γ-Valerolactone by Ruthenium
Arene Catalyst Precursors
Novisi K. Oklua and Banothile C. E. Makhubelaa*
aDepartment of Chemistry, University of Johannesburg, Auckland Park Kingsway Campus
P.O. Box 524, South Africa, Johannesburg
*E-mail: bmakhubela@uj.ac.za
HIGHLIGHTS
1. A highly selective method of converting bio-based levulinic acid to γ-valerolactone
using formic acid as a hydrogen carrier.
2. Hydrogenation of levulinic acid resulted in 100% conversions and selectivity to γ-
valerolactone, and in situ NMR characterization has aided the proposal of a plausible
mechanism of the reaction.
3. Ruthenium pre-catalysts catalyze the hydrogenation reactions without the need of a
solvent and are also recyclable, thereby making the entire reaction conditions greener.
26