Reductive amination of levulinic acid to different pyrrolidones on Ir/SiO2-SO3H: Elucidation of reaction mechanism

Article abstract of DOI:10.1016/j.cattod.2017.08.038

Levulinic acid (LA) transformation to different pyrrolidones via reductive amination was studied using an Ir/SiO₂-SO₃H catalyst in liquid phase. The effects of solvent, amine concentration, H₂ partial pressure, catalyst mass, and reuse were studied. The spent catalysts were evaluated by DRIFTS to obtain evidence of the interaction of levulinic acid with aniline on the catalyst surface. The sulfonic groups on SiO₂ improved the yield to pyrrolidones and avoided side reactions. A reaction mechanism is proposed where the reductive amination occurs between LA and amine towards an intermediate amine that is then cycled. The Gibbs free energy for the reaction mechanism was evaluated. Besides, the HOMO and LUMO of the amine reactants and intermediates using density functional theory (DFT) with the B3LYP-D3 method were theoretically determined to understand the rate-limiting and the cyclization steps.

Full text of DOI:10.1016/j.cattod.2017.08.038

Accepted Manuscript
Title: REDUCTIVE AMINATION OF LEVULINIC ACID
TO DIFFERENT PYRROLIDONES ON Ir/SiO -SO H:
2
3
ELUCIDATION OF REACTION MECHANISM
Authors: Jos e´ J. Mart ´ı nez, Leonardo Silva, Hugo A. Rojas,
Gustavo P. Romanelli, Lucas A. Santos, Teodorico C.
Ramalho, Mar ´ı a H. Brijaldo, Fabio B. Passos
PII:
S0920-5861(17)30560-6
DOI:
Reference:
http://dx.doi.org/10.1016/j.cattod.2017.08.038
CATTOD 10981
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
14-1-2017
6-6-2017
14-8-2017
Please cite this article as: Jos e´ J.Mart ´ı nez, Leonardo Silva, Hugo A.Rojas,
Gustavo P.Romanelli, Lucas A.Santos, Teodorico C.Ramalho, Mar ´ı a H.Brijaldo,
Fabio B.Passos, REDUCTIVE AMINATION OF LEVULINIC ACID TO
DIFFERENT PYRROLIDONES ON Ir/SiO2-SO3H: ELUCIDATION OF REACTION
MECHANISM, Catalysis Todayhttp://dx.doi.org/10.1016/j.cattod.2017.08.038
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.

REDUCTIVE AMINATION OF LEVULINIC ACID TO DIFFERENT PYRROLIDONES
2
ON Ir/SiO -SO
3
H: ELUCIDATION OF REACTION MECHANISM
1
1
1
José J. Martínez , Leonardo Silva , Hugo A. Rojas ,
2
3
3
Gustavo P. Romanelli , Lucas A. Santos , Teodorico C. Ramalho ,
4
4
María H. Brijaldo , Fabio B. Passos
1
Escuela de Ciencias Químicas, Facultad de Ciencias. Universidad Pedagógica y Tecnológica
de Colombia UPTC, Avenida Central del Norte, Tunja, Boyacá, Colombia.
2.
Centro de Investigación y Desarrollo en Ciencias Aplicadas “Dr. J.J. Ronco” (CINDECA),
Departamento de Química, Facultad de Ciencias Exactas, UNLP-CCT-CONICET. Calles 47 Nº
257, B1900 AJK, La Plata, Argentina.
3
Departamento de Química, Universidade Federal de Lavras, Campus Universitário, Lavras,
MG, 37200-000, Brazil.
4
Departamento de Engenharia Química e de Petróleo, Universidade Federal Fluminense, R.
Passos da Pátria, 156, Niterói, RJ, 24210-240, Brazil.
Corresponding authors: jose.martinez@uptc.edu.co; hugo.rojas@uptc.edu.co

Graphical abstract
Highlights



Ir/SiO
2
-SO
3
H showed a high yield to pyrrolidone (88 % at 24 h)
Sulfonic groups on SiO
2
improved the yield to pyrrolidones and avoided side reactions.
Reductive amination occurs between LA and amine towards an intermediate amine
that is then cycled.


Theoretical and experimental findings are in a good agreement with the mechanism
proposed.
The substituents amines alter the contribution HOMO – LUMO and consequently
modifying the yields obtained.

Abstract
Levulinic acid (LA) transformation to different pyrrolidones via reductive amination was
studied using an Ir/SiO
concentration, H partial pressure, catalyst mass, and reuse were studied. The spent catalysts
were evaluated by DRIFTS to obtain evidence of the interaction of levulinic acid with aniline
on the catalyst surface. The sulfonic groups on SiO improved the yield to pyrrolidones and
2
-SO
3
H catalyst in liquid phase. The effects of solvent, amine
2
2
avoided side reactions. A reaction mechanism is proposed where the reductive amination
occurs between LA and amine towards an intermediate amine that is then cycled. The Gibbs
free energy for the reaction mechanism was evaluated. Besides, the HOMO and LUMO of the
amine reactants and intermediates using density functional theory (DFT) with the B3LYP-D3
method were theoretically determined to understand the rate-limiting and the cyclization
steps.
Keywords: pyrrolidones, sulfonic groups, reductive amination
1
. Introduction
Levulinic acid (LA) is obtained by hydrolysis of cellulose, and its reductive amination is an
interesting process because pyrrolidones are produced. In several industrial processes,
pyrrolidones are important as solvents, surfactants, complexing agents, and as part of
pharmaceutical formulations [1, 2]. Recently, Zhu et al. [3] have demonstrated that the system
composed of N-methyl-2-pyrrolidone and C1–C4 carboxylic acid can be used as a solvent with
exceptional lignin solubility that outperforms conventional solvents and ionic liquids.
Therefore, reductive amination has attracted considerable interest due to its potential to
produce a variety of pyrrolidones [4]. Raney nickel catalysts or silica gel-supported nickel
catalysts were used to transform LA to 1,5-methyl-2-pyrrolidone by reductive amination using
high pressures of H
catalysts to obtain aryl-, alkyl-, and cycloalkylpyrrolidones with the aim of increasing the
catalytic activity for the reductive amination of LA. They used Pt and MoO coloaded TiO2,
which allows polarization of the carbonyl groups via Lewis acid−base interaction. However,
2
(1000 - 2000 Psi) [4,5]. Touchy et al. [7] studied different metal-supported
x

these authors did not show evidence of this type of interaction. Wei et al. used transfer
hydrogenation methods [8] or a noncatalytic system [9] with the aim of diminishing the high
pressure or temperature employed in conventional hydrogenation processes. They found that
the solvent DMSO is critical for the high reactivity observed. The reason why DMSO is an
effective solvent is unclear. Andrioletti et al. [10] reported another example of metal-free
reductive amination of LA. However, the reaction required more forceful conditions including
higher temperatures (433 – 473 K) with the disadvantage of the formation of formamide as
byproduct of the condensation of the formic acid and amine used. Table 1 summarizes the yield
and recycle of catalysts reported in the literature to obtain 5-methyl-N-phenyl-2-pyrrolidone.
In this contribution, we studied the synthesis of some pyrrolidones from levulinic acid using
Ir/SiO
2
-SO H. The aim of this work is to determine the role of acid sites in the reductive
3
amination of levulinic acid. The reductive amination of levulinic acid requires nucleophilic
addition, which could be favored by the presence of Brönsted sites. To the best of our
knowledge, this is the first time an Ir/SiO
amination of levulinic acid.
2
-SO H catalyst has been used in the reductive
3
2. Experimental
2.1. Catalyst preparation
The Ir/SiO
2
-SO
3
H catalyst was previously prepared and characterized as described in
2
reference [14]. Briefly, a mixture of 5 g SiO
2
(Syloid, SBET = 283 m /g) in 100 mL of dry toluene
(
Panreac, 99.5%) and 3-mercaptopropyltrimethoxysilane (MPTMS; Sigma-Aldrich, 95%) (1.15
mL, 6.1 mmol) was refluxed for 24 h. The obtained solid was washed with toluene and dried at
63 K. Subsequently, mercaptopropyl groups were oxidized to sulfonic acid groups with excess
3
hydrogen peroxide (30%, Panreac) at room temperature for 24 h. A few drops of H
added during 12 h, and the solid was washed with acetone and dried at 393 K.
2
SO were
4
The solid SiO
2
-SO H was then impregnated with Ir-PVP colloids (PVP: polyvinylpyrrolidone)
3
at room temperature for 6 h under stirring and then dried under vacuum at 353 K. The synthesis
of Ir-PVP colloids was similar to that described in the literature [15].

2.2 Catalyst characterization
The Ir-PVP supported catalyst was characterized by X-ray photoelectron spectroscopy (XPS).
XPS data were obtained in a Thermo Scientific Escalab 250 XI spectrometer. The samples were
purged with helium (25 mL/min) for 1 h in the chamber of analysis. Measurements were
performed at room temperature with monochromatic Al Kα (hv = 1486.6 eV) radiation. The
analyzer was operated at 25 eV pass energy with a step size of 0.05 eV. The pressure in the
-
9
analytical chamber was 6.3 x 10 mBar. The C 1s signal (284.6 eV) was used as internal energy
reference in all the experiments. Core-level peak positions were determined after background
subtraction according to Shirley method using Avantage software. Peaks in a spectrum were
fitted by a combination of Gauss and Lorentz curves, and overlapping peaks were also
separated using this combination.
Iridium particle size was determined by transmission electron microscopy in a JEOL JEM-1011
analytical microscope. At least 200 particles were measured to obtain the particle size
3
2
distribution and the average particle size d
the number of particles with a diameter di.
p
defined as d = ∑nidi /∑nidi , where ni represents
p
Thermogravimetric analysis (TGA) was performed in SETARAM SA equipment, using a
.
-1
heating rate of 5 K min from room temperature to 1073 K in inert atmosphere.
The nature of acid sites was studied by pyridine (Pyr) adsorption followed by FTIR. Infrared
spectra were collected using Nicolet iS50 equipment with an in situ diffuse reflectance cell
(Harrick, Praying Mantis). For the determination of acid sites, a pretreatment at 423 K was
performed with a helium flow of 15 mL/min for 1 h to clean the surface of possible contaminants.
Subsequently, the samples were gradually cooled down to 303 K, and then pyridine adsorption
was performed for 1 h. After adsorption, the gas phase was removed by evacuation with a
helium flow (20 mL/min).
DRIFT spectra of the spent catalyst were collected at different times of reaction. The catalyst
was separated by filtration, dried, and placed in a Harrick cell. The cell was purged with helium
at a constant flow of 50 mL/min at 373 K before collecting the spectrum.

2.3 Catalytic tests
The liquid phase reductive amination reaction was carried out in a batch-type reactor at
constant stirring rate (1000 rpm) using 40 mL of 0.125 M aniline solution and 0.250 M levulinic
acid solution, in ethyl acetate as solvent, using 0.1 g of Ir/SiO
temperature of 373 K. The effect of H pressure on the reductive amination of levulinic acid was
studied at a H pressure ranging from 250 to 750 Psi. The Weisz–Prater criterion was employed
2
-SO H catalyst, at a reaction
3
2
2
to assess the influence of the intraparticle diffusion resistance criterion. Reaction products
were analyzed in a gas chromatograph coupled to a mass spectrometer (GC-MS) Varian 3800-
Saturn 2000 furnished with a β-Dex column, helium as carrier, and constant temperature at
393 K. The conversion of substrate (levulinic acid) and selectivity to pyrrolidone were
determined using Eqs 1 and 2. The conversion of reactant (α) and product yield (Y) and
selectivity were defined as follows:
퐶0−퐶푖
α (%) =
1ꢀꢀ (Eq. 1)
퐶0
퐶푝푡표푑
S (%) = ∑
1ꢀꢀ (Eq. 2)
퐶푝푡표푠
푌 = 푆 ∗ 훼 (Eq. 3)
where C
0
is the initial concentration, C is the concentration at time i, Cptod is the concentration
i
of the desired product, and Cptos is the concentration of the obtained products. YN=C and YN-C
are yield to imine or yield to pyrrolidone, respectively.
Different pyrrolidones were synthesized by changing the amine molecule. Its identification was
performed by comparison of the m/z spectra, because they are not new compounds, except for
pyrrolidone derived from furfurylamine.
(i) 5-Methyl-N-phenyl-2-pyrrolidone: Yield: 63%; m/z calc.: 175.2. Found: 175.2.
(
ii) 5-Methyl-1-p-tolyl-pyrrolidin-2-one: Yield: 43%; m/z calc.: 189.257. Found: 189.3.
iii) 1-(4-Chlorophenyl)-5-methylpyrrolidin-2-one: Yield: 31%; m/z calc.: 209.675. Found: 210.
(

(
(
(
iv) 5-Methyl-1-(3-nitro-phenyl)-pyrrolidin-2-one: Yield: 3%; m/z calc.: 220.228. Found: 220.3.
v) N-benzyl-5-methylpyrrolidin-2-one: Yield: 31% (lit. [16])
vi) 1-(4-Aminophenyl)-5-methylpyrrolidin-2-one: Yield: 46%; (lit. [12]); m/z calc.: 191.22642. Found:
1
91.23.
vii) 1-(Furan-2-ylmethyl)-5-methyl-pyrrolidin-2-one: Yield: 39%; m/z calc.: 179.216. Found: 179.5.
(
2.4 Computational Details
2.4.1 Determination of HOMO and LUMO
The geometries of the molecules in the gas phase were optimized at the B3LYP/6-31++** level
with Grimme’s D3 empirical dispersion correction using the Gaussian 09 software [17], in order
to obtain the HOMO and LUMO energies of the anime reactants and intermediates. The
structures were minimum on potential energy surface, and their harmonic vibrational
frequencies were positive (see supplementary material). Molecule visualizations at their
optimized geometries and ionization potential surfaces were performed using the MacMolPlt
program package.
2.4.2 Thermodynamic analysis
The optimized geometries of each reaction step and their correspondent Gibbs free energies in
ethyl acetate at 363 K were calculated at the B3LYP/6-31++** level with Grimme’s D3 empirical
dispersion correction [18], using the Gaussian 09 software [17]. The solvation environment was
simulated using/employing the polarizable continuum model (PCM) [19].
The transition state calculations of the hydrogenation step were carried out at B3LYP-
D3/lanL2DZ level. The iridium nanoparticles were modeled using the Ir unit cell coordinates
imported from materialsproject.org [ID: mp-101].
2.5. Reuse of catalyst
The catalytic activity of the reductive amination of levulinic acid with aniline was studied for
cycles. After a typical reaction, the Ir/SiO -SO H catalyst was filtered with a Whatman®
cellulose filter paper – (Sigma-Aldrich), and dried at 373 K for 1 h.
4
2
3

3. Results and discussion
3.1 Catalyst characterization
Fig. 1 displays typical TEM images and the corresponding particle size distribution histograms
for Ir/SiO and Ir/SiO -SO H catalysts. The average particle size d was 1.7 nm for Ir/SiO , while
for Ir/SiO -SO H it was 2.3 nm. The particle size distribution was broader on Ir/SiO -SO
2
2
3
p
2
2
3
2
3
H
support, possibly due to some interaction between the sulfonic groups and the PVP used as
stabilizing on Ir colloids. This phenomenon does not occur when metal particles are deposited
without surfactant as in the case of Ru/SiO
2
-SO H [20].
3
Fig. 2 show the deconvoluted XP spectra of Ir/SiO
2
and Ir/SiO
2
-SO H in the region of Ir 4f. The
3
binding energies (BE) of Ir 4f7/2 and 4f5/2 peaks are summarized in Table 2. Both contributions
0
of Ir 4f are associated with Ir atoms and partially oxidized IrO
2
, respectively. BE values are
0
slightly smaller than the corresponding values of the bulk for Ir . In both cases (Ir/SiO
2
and
Ir/SiO
14], which makes the Ir particles slightly negatively charged [21]. This phenomenon is most
evident in the Ir/SiO catalyst. Thus, it seems that the sulfonic groups have some interaction
with PVP, altering the chemical environment of the metal particles. In addition, peaks
assignable to IrO were also slightly smaller than the corresponding bulk values.
2
-SO
3
H), the corresponding values can be attributed to electron donation from the PVP
[
2
2
Thermogravimetric analyses (TGA) were performed to corroborate the temperature of
decomposition of organic residues. Fig. 3 shows the TGA of Ir/SiO -SO H catalyst. The loss of
O can be observed at 393 K. However, the loss of organic residues (associated with MPTMS,
2
3
H
2
sulfonic groups, and PVP) can be seen from 550 K to 820 K, confirming that the catalyst
structure is preserved at the reaction temperatures.
In order to determine the acid site nature of catalysts, pyridine-FTIR analysis was performed
-
1
-1
-1
(Fig. 4). In SiO
2
-SO
3
H three bands at 1490 cm , 1503 cm and 1531 cm can be observed. The
-1
-1
bands at 1490 cm and 1531 cm are associated with typical Brönsted sites. The band at 1503

-1
cm is characteristic of Lewis acid sites [25, 26]. The Pyr-FTIR spectrum of Ir/ SiO
2
-SO
3
H only
shows the bands associated with Brönsted sites. Ir/SiO
sites.
2
catalyst does not present Brönsted
3.2 Catalytic tests
Levulinic acid with aniline was studied as test reaction, and blank experiments were carried
out to study the effect of the sulfonic groups (Table 3). Fig. 5 shows the conversion level as a
function of reaction time in the reductive amination of levulinic acid. The plateau only seems
to be reached in the test with Ir/SiO
2
-SO H catalyst with a conversion near 90% at 24 h. The
3
Ir/SiO catalysts shows the lowest conversion.
2
With regard to selectivity, the hydrogenation of both carbonyl groups of levulinic acid was not
observed in any catalyst. When SiO -SO H was used as catalyst, imine was the only product
observed. By using the Ir/SiO catalyst, the reaction proceeds to pyrrolidone with a lower yield
at the time of reaction studied. The pyrrolidone yield increased when the Ir/SiO -SO H catalyst
2
3
2
2
3
was used. Due to this fact, pyrrolidone was the only product obtained in the Ir catalysts; the
activity was compared at 8 h, and it was expressed as yield to imine (YC=N (%)) or yield to
pyrrolidone (YN-C (%)). Table 3 summarizes the catalytic data obtained at 8 h of reaction.
Fig. 6 shows the evolution of levulinic acid and aniline concentration and their products in time
using Ir/SiO
2
-SO H catalyst. It is noted that imine formation takes place mainly during the first
3
2
h of reaction and it is not detected at high/long reaction times. Thus, the hydrogenation is
fast. Pyrrolidone is the main product of reaction, and other products were not observed.
The best reaction conditions for the reductive amination of levulinic acid with aniline were:
constant stirring rate (1000 rpm) using 40 mL of 0.125 M aniline solution and 0.250 M levulinic
acid solution, with ethyl acetate as solvent, at a reaction temperature of 373 K, with a hydrogen
pressure of 500 psi, using 0.1 g of Ir/SiO
2
-SO H catalyst.
3
3.2.1 Effect of catalyst mass

The effect of catalyst loading on the yield to pyrrolidone in the catalytic reductive amination of
LA was studied. The results are listed in Table 4, and they lead us to believe that there is a linear
trend in the yield to pyrrolidone with an increase of the catalyst mass used. This suggests an
absence of external mass transfer resistance. The Weisz–Prater criterion was employed to
assess the influence of intraparticle diffusion resistance. Table 4 shows the Weisz- Prater
values at distinct mass used. In all cases, the value of the Weisz-Prater (NWP) criterion
calculated was less than 1. Hence, there was no intraparticle diffusion resistance.
3
.2.2 Effect of amine concentration and H partial pressure
2
The role of sulfonic groups favors the adsorption of LA and could also allow adsorption of the
amine molecule, causing a possible decrease of the yield to pyrrolidone. Thus, the surface
coverage by amines inhibits the adsorption of LA step, and this could cause the formation of an
amine salt (-SO
3
H + RNH
2
= SO
3
-NRH
+). The above-mentioned can be observed in Table 5, as
3
the highest aniline: LA ratio inhibits the formation of pyrrolidone. This fact was observed by
-
1
the band at 1602 cm in the IR spectra of the spent catalysts (see Fig. 8).
The effect of H pressure on the reductive amination of levulinic acid was studied in the range
2
of 250 to 750 psi at constant initial concentration. The results of different partial pressures of
hydrogen are listed in Table 6. An increase in hydrogen pressure caused a decrease in the
conversion and yield of pyrrolidone. This could be explained by a small interaction of amine
molecules on the catalyst surface with an increase in H pressure.
2
3.2.3 Effect of solvent
We also studied the effect of solvent, and the results are summarized in Table 7. Pyrrolidone is
the only product when the solvent reaction is acetonitrile and ethyl acetate. In ethanol, a lower
yield to pyrrolidone is observed because of the competition of the reductive amination with the
esterification between LA and ethanol (63 %), producing a compound with m/z =144.2.
3.2.4. Plausible mechanism of the reductive amination of levulinic acid

In line with the above findings, we propose a plausible mechanism (Fig. 7) for the reductive
amination of levulinic acid with aniline over Ir/SiO
assumptions:
2
-SO H catalyst based on the following
3
1
. The sulfonic groups interact with the carboxylic group of levulinic acid. An effective
inductive effect on the CO group at that distance is not possible. The COOH-SO H- interaction
3
only allows the carbonyl group C2 to be available for the nucleophilic addition and formation
of imine.
2. The second step includes the C=C bond hydrogenation of imine by hydrogen on iridium
nanoparticles to obtain the amine intermediate. This step results in the formation of the amine
intermediate.
3
. In the third step, the cyclization of the amine intermediate with the carboxyl group of
levulinic acid occurs. This step also involves a nucleophilic attack of the amine intermediate
with the carboxylate adsorbed on -SO H groups, this reaction being spontaneous. This step
3
occurs easily between the nitrogen atom (HOMO) and the carboxyl group (LUMO). The
substituents alter the contribution HOMO – LUMO and consequently modify the yields
obtained.
To confirm this plausible mechanism, DRIFT spectra of spent catalysts at different times of
reaction and theoretical calculations were obtained.
-1
Fig. 8a and 8b show the DRIFT spectra in the region of 1900 to 1590 cm of aniline and LA
-1
adsorbed on the catalyst surface, respectively. In Fig. 8a, the band at 1602 cm can be ascribed
to the amine group scissoring vibration in anilinium sulfate, which has been previously
-
1
-1
assigned [26]. In Fig. 8b the band has two shoulders, one at 1737 cm and other at 1726 cm ,
which could be ascribed to the interaction of COOH groups and the carbonyl group of levulinic
-
1
acid with sulfonic groups. The band at 1737 cm could be attributed to C=O bond stretching
and angular deformation of the C-O-H bond of the COOH group, which are described by the
-1
-1
band at 1692 cm in the theoretical spectrum (Fig. 9a). The other band at 1760 cm is ascribed
-
1
to C=O bond stretching of the carbonyl group, in agreement with the band at 1764 cm in the
theoretical spectra of levulinic acid adsorbed on Ir/SiO -SO H (Fig. 9b).
2
3

The spectra collected in the case of the spent catalyst at 15 – 90 min of reaction (Fig. 8c -8f)
showed a shift of the adsorption bands of levulinic acid to lower wavenumbers, so the blue shift
indicates a modification in the structure of levulinic acid as a consequence of aniline addition.
-
1
In this spectrum, a band at 1641 cm related to C-C bond stretching of the aromatic ring (which
-
1
is also observed in the experimental spectra) is evidenced, while the band at 1697 cm is
characteristic of C=O bond stretching and angular deformation of the C-O-H bond of the
COOH group. This was confirmed with the DRIFT theoretical spectrum of the step II product
(Fig. 9c).
Comparing the data of the levulinic acid adsorbed on Ir/SiO
2
-SO
3
H (8b) and Ir/SiO
2
(Fig. 10a)
,
there is a blue shift when the interactions with the SO H group take place. In Ir/SiO
3
2
(Fig. 10a),
the proportion of the bands of LA adsorption change, a band at 1750 cm-1 being more
predominant and possibly associated with the -COOH group vibrations (theoretical data
-
1
points of the -COOH group vibrations of levulinic acid occur at 1760 cm ), which means that
in Ir/SiO more -COOH groups are available (i.e., not adsorbed). With progress of the reaction,
2
-
1
the bands at 1750 cm (Fig. 10 b-e) decrease due to formation of the product (Fig. 9d).
In order to shed some more light on the experimental data, thermodynamic study about the
stability of the intermediates was performed. In fact, the theoretical calculations emphasize the
possibility of the proposed mechanism. Fig. 11 shows the Gibbs free energy of the mentioned
-1
steps, and the global ΔG is equal to -29.74 kcal.mol . All calculated structures of each step are
depicted in Fig. 12. In the first step (I), when imine formation occurs, the Gibbs free energy is
-
1
5
.15 kcal.mol . Despite this, the following steps are all spontaneous. The hydrogenation step
-
1
(
II) results in a stabilization of -21.12 kcal.mol . This step is the most crucial to the whole
reaction, with the lowest ΔG among the three steps. Here, the proposed mechanism suggests
an interaction between the hydrogens and the iridium nanoparticles. In this sense, we
calculated the transition state of step II with and without Ir atoms. In fact, the stabilization of
the transition state in step II deserves some attention. Our calculations have shown that the
difference between the activation energies with (GII-TS-Ir) and without (GII-TS) iridium
-1
nanoparticles is -57.90 kcal.mol (Fig. 13). Thus, the contribution of iridium nanoparticles to

active hydrogen availability is very important. The third and final step (III) leads to the
-
1
products with ΔGIII = -13.77 kcal.mol .
3.2.5. Effect of amine substituent
The reductive amination of LA was carried out under optimized conditions, using a variety of
aromatic amines and furfurylamine. The highest conversions were obtained with electron-
donor substituents (entries ii and vi); however, by using electron-withdrawing groups (entry
iv), the reaction is not favored (Table 8).
In order to clarify these phenomena, we correlated the activity and the molecular orbital
properties of amine reactants in the first step of the proposed mechanism. This step is the most
crucial to the mechanism, as our thermodynamic analysis has shown. Furthermore, we
theoretically studied the HOMO and LUMO of amine intermediates formed to understand the
cyclization step.
To compare the catalytic performance between all amine reactants, the electronic chemical
potential (μ) was calculated using the HOMO – LUMO energies defined by Parr and Pearson
[
27] considering the equation: μ=-½ (I+A) = ½ (EHOMO +ELUMO); where A is electron affinity (A),
and I is ionization energy. The data of Table 8 show a relation between the electronic chemical
potential of the amine intermediates and the yield (%) to pyrrolidones. Electron-withdrawing
groups decrease the HOMO energies, whereas they are increased by the presence of electron-
donor groups. In fact, the substituent effect will affect the nucleophilic attack of the amine
nitrogen on the carboxyl group, i.e., a higher electronegativity of amine reactants favors the first
step effectiveness (see Fig. 14).
The pictorial view of the HOMO and LUMO in the amine intermediates (compounds i-vii) is
depicted in Fig. 15. The HOMO shows a contribution from the nitrogen and benzene ring (or
furan ring) and a very small contribution of the methyl group, whereas the LUMO corresponds
mainly to the benzene ring and the carboxyl group. The cyclization step occurs easily between
the nitrogen atom and the carboxyl group due to the interaction HOMO – LUMO, respectively.

Albeit with the lowest band gap, when the substituent is m-nitroaniline, the carboxyl group
does not contribute to the LUMO, and the yield to pyrrolidone is much lower (Table 7).
3.2.6. Reuse of catalyst
The catalyst reuse was studied in 4 cycles (Fig. 16). In this study, the catalyst was recovered,
and not regenerated, to evaluate its behavior with the recycles. Thus, a progressive use of the
catalyst in this reaction generates a loss of catalytic activity, due to saturation of -SO H sites
3
by the amine. The interaction of amine molecules with sulfonic groups allows the saturation
of these sites. If the catalyst is not regenerated, amines adsorption is responsible for
deactivation.
4. Conclusions
Levulinic acid transformation via reductive amination to pyrrolidones was studied using a
bifunctional metallic-acidic catalyst in a liquid phase. The catalytic test of this reaction showed
a yield to pyrrolidone higher than 60%, with a selectivity of 100%, using aniline as initial
substrate and ethyl acetate as solvent. Although the max yield reported in this reaction is near
90 %, reached in 20 h with Pt-MoOx/TiO
2
, with our Ir/SiO
2
-SO H catalyst it is 88 % at 24 h.
3
We propose a reaction mechanism where the carboxyl group is adsorbed on sulfonic groups
favoring the formation of imine, which is then hydrogenated. This mechanism was
corroborated with theoretical thermodynamic calculations and experimental results. The
electronegativity (μ) of amine intermediates determines the cyclization step, mainly due to the
interaction HOMO – LUMO between the nitrogen atom and the carboxyl group, respectively.
Acknowledgments
The authors thank Colciencias for financial support under the project 110965843004. Contract:
Colciencias-UPTC 047-2015 and the project European Research Area Network; ERANet LAC
(ref. ELAC2014/BEE-0341).

References
[
1]
L.E. Manzer, Production of 5-methyl-N-(methyl aryl)-2-pyrrolidone, 5-methyl-N-
methyl cycloalkyl)-2-pyrrolidone and 5-methyl-N-alkyl-2-pyrrolidone by reductive
amination of levulinic acid with cyano compounds, US 6,841,520 B2, 2005.
(
[
[
2] L.E. Manzer, F.E. Herkes, Production of 5-methyl-1hydrocarbyl-2-pyrrolidone by
reductive amination of levulinic acid, US2004/0192933, 2004.
3] L. Mu, Y. Shi, L. Chen, T. Ji, R. Yuan, H. Wang, J. Zhu, [N-Methyl-2-pyrrolidone][C1-C4
Carboxylic Acid]: A Novel Solvent System with Exceptional Lignin Solubility, Chem.
Commun. 51 (2015) 13554–13557.
[
4] J.D. Vidal, M.J. Climent, P. Concepcion, A. Corma, S. Iborra, M.J. Sabater, Chemicals
from Biomass: Chemoselective Reductive Amination of Ethyl Levulinate with Amines,
ACS Catal. 5 (2015) 5812–5821.
[
[
5] F. Robert, S. William, Z. Blossom, 1,5-Dimethyl-2-Pyrrolidone, Org. Synth. 27 (1947) 28.
6] W.C. Wilbur L. Shilling, Making lactams by the vapor phase reductive amination of oxo
carboxylic acid compounds, US Patent 3235562, 1966.
[
[
7]
A.S. Touchy, S.M. a H. Siddiki, K. Kon, K. Shimizu, Heterogeneous Pt Catalysts for
Reductive Amination of Levulinic Acid to Pyrrolidones, ACS Catal. 2 (2014) 3045–3050.
8] Y. Wei, C. Wang, X. Jiang, D. Xue, J. Li, J. Xiao, Highly efficient transformation of
levulinic acid into pyrrolidinones by iridium catalysed transfer hydrogenation, Chem.
Commun. (2013) 5408–5410.
[
[
[
9] Y. Wei, C. Wang, X. Jiang, D. Xue, Z.-T. Liu, J. Xiao, Catalyst-free transformation of
levulinic acid into pyrrolidinones with formic acid, Green Chem. 16 (2014) 1093–1096.
10] a. Ledoux, L. Sandjong Kuigwa, E. Framery, B. Andrioletti, A highly sustainable route to
pyrrolidone derivatives – direct access to biosourced solvents, Green Chem. (2015).
11] Y. Huang, J. Dai, X. Deng, Y. Qu, Q. Guo, Y. Fu, Ruthenium-Catalyzed Conversion of
Levulinic Acid to Pyrrolidines by Reductive Amination, (2011) 1578–1581.

[
[
[
12] Y. Ogiwara, T. Uchiyama, N. Sakai, Reductive Amination/Cyclization of Keto Acids
Using a Hydrosilane for Selective Production of Lactams versus Cyclic Amines by
Switching of the Indium Catalyst, Angew. Chemie - Int. Ed. (2015) 1864–1867.
13] C. Ortiz-Cervantes, M. Flores-Alamo, J.J. García, Synthesis of pyrrolidones and
quinolines from the known biomass feedstock levulinic acid and amines, Tetrahedron
Lett. 57 (2016) 766–771.
14] J.J. Martínez, E. Nope, H. Rojas, M.H. Brijaldo, F. Passos, G. Romanelli, Reductive
amination of furfural over Me/SiO2-SO3H (Me: Pt, Ir, Au) catalysts, J. Mol. Catal. A
Chem. 392 (2014) 235–240.
[
[
[
15] W. Yu, M. Liu, H. Liu, J. Zheng, Preparation of Polymer-Stabilized Noble Metal Colloids,
J. Colloid Interface Sci. 210 (1999) 218–221.
16] R. Lukeš, Z. Koblicová, K. Bláha, Über die reaktion der angelicalactone mit aminen,
Collect. Czechoslov. Chem. Commun. 28 (1963) 2182–2198.
17] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, G.A.J.R. Cheeseman, G. Scalmani,
V. Barone, B. Mennucci, A.F. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian,
M.E. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, Y.H.K. Toyota, R. Fukuda,
J. Hasegawa, M. Ishida, T. Nakajima, J.E.P.O. Kitao, H. Nakai, T. Vreven, J.A.
Montgomery, V.N.F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, A.R.
Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, J.M.M.J.C. Burant, S.S. Iyengar,
J. Tomasi, M. Cossi, N. Rega, J.J.M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo,
R.C.R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, V.G.C. Pomelli, J.W.
Ochterski, R.L. Martin, K. Morokuma, S.D. Zakrzewski, G.A. Voth, P. Salvador, J.J.
Dannenberg, A.A.D. Daniels, Farkas, J.B. Foresman, J. V. Ortiz, J. Cioslowski, D.J. Fox.,
Gaussian Inc. Gaussian Inc. Wallingford CT, (2009).
[
[
[
18] C. Steffen, K. Thomas, U. Huniar, A. Hellweg, O. Rubner, A. Schroer, TmoleX--a
graphical user interface for TURBOMOLE., J. Comput. Chem. 31 (2010) 2967–2970.
19] J. Tomasi, B. Mennucci, R. Cammi, Quantum Mechanical Continuum Solvation Models,
Chem. Rev. 105 (2005) 2999–3093.
20] W. Zhu, H. Yang, J. Chen, C. Chen, L. Guo, H. Gan, X. Zhao, Z. Hou, Efficient

hydrogenolysis of cellulose into sorbitol catalyzed by a bifunctional catalyst, Green
Chem. 16 (2014) 1534–1542.
[
[
[
21] M.J. Sharif, P. Maity, S. Yamazoe, T. Tsukuda, Selective Hydrogenation of
Nitroaromatics by Colloidal Iridium Nanoparticles, Chem. Lett. 42 (2013) 1023.
22] S.J. Morgan, R.H. Williams, J.M. Mooney, An XPS study of thin Pt and Ir silicide
overlayer formation on Si(100)2 X 1 surfaces, Appl. Surf. Sci. 56–58 (1992) 493–500.
23] A.K. Shukla, A.M. Kannan, M.S. Hegde, J. Gopalakrishnan, Effect of counter cations on
electrocatalytic activity of oxide pyrochlores towards oxygen reduction/evolution in
alkaline medium: an electrochemical and spectroscopic study, J. Power Sources. 35
(1991) 163–173.
[
24] P.P. Upare, J.-W. Yoon, M.Y. Kim, H.-Y. Kang, D.W. Hwang, Y.K. Hwang, H.H. Kung,
J.-S. Chang, Chemical conversion of biomass-derived hexose sugars to levulinic acid over
sulfonic acid-functionalized graphene oxide catalysts, Green Chem. 15 (2013) 2935.
[
[
[
25] S. Hamoudi, S. Kaliaguine, Sulfonic acid-functionalized periodic mesoporous
organosilicas, Stud. Surf. Sci. Catal. 146 (2002) 473–476.
26] P.L. Anto, R.J. Anto, H.T. Varghese, C.Y. Panicker, D. Philip, a. G. Brolo, FT-IR, FT-
Raman and SERS spectra of anilinium sulfate, J. Raman Spectrosc. 40 (2009) 1810–1815.
27] R.G. Parr, R.G. Pearson, Absolute hardness: companion parameter to absolute
electronegativity, J. Am. Chem. Soc. 105 (1983) 7512–7516.

8
6
4
2
0
0
0
0
0
a.
b.
dp= 1.7 nm
1
2
3
4
Particle size (nm)
4
3
2
1
0
0
0
0
0
dp = 2.3 nm
1
2
3
4
5
Particle size (nm)
Fig. 1. Typical TEM images and particle size distribution of Ir catalysts (a) Ir/SiO
Ir/SiO -SO H.
2
and (b)
2
3
Ir/SiO -SO H
2
3
Ir/SiO2
68
66
64
62
60
58
56
Binding energy (e. V)
Fig. 2. XPS spectra of Ir/SiO
2
-SO
3
H and Ir/SiO
2
catalysts

100
-
H O: -2.99%
2
95
90
85
80
-
Organic
Residues: -14.7%
400
600
800
1000
Temperature (K)
Fig. 3. Thermogram of Ir/SiO
2
-SO H catalyst
3
B
L
0,6
0,4
0,2
0,0
c
b
a
1540
1520
1500
1480
-
1
Wavenumber (cm )
Fig. 4. FTIR-Pyr of: (a). SiO
2
-SO
3
H (b). Ir/SiO
2
and (c). Ir/SiO
2
-SO H, where B and L are
3
Brönsted and Lewis sites, respectively.

100
c
b
75
50
25
0
a
0
6
12
18
24
Time (h)
Fig. 5. Conversion level as a function of reaction time in the reductive amination of levulinic
acid using (a) SiO SO H and (b) Ir/SiO , and (c) IrSiO -SO H. Reaction conditions: 0.125
, 0.1 g of catalyst, ethyl acetate as
2
3
2
2
3
moles of LA and 0.25 moles of aniline at 373 K, 500 psi of H
solvent.
2
0.028
0.021
0.014
0.007
0.000
Aniline
Levulinic Acid
Pyrrolidone
Imine
0
5
10
15
20
25
Time (h)
Fig. 6. Catalytic activity of the reductive amination of levulinic acid with aniline, using Ir/SiO
SO H catalyst. Reaction conditions: 0.125 moles of LA and 0.25 moles of aniline at 373 K, 500
psi of H , 0.1 g of catalyst, ethyl acetate as solvent.
2
-
3
2

Fig. 7. Plausible mechanism for the reductive amination of levulinic acid with aniline over
Ir/SiO
2
-SO H catalyst.
3
1737
1726
1697
1760
1641
1602
f
e
d
c
b
a
1
900 1850 1800 1750 1700 1650 1600
-
1
Wavenumber (cm )
Fig. 8. DRIFT spectra of Ir-SiO
adsorption of levulinic acid on the catalyst surface. Spectra a and b are considered as time zero.
c), (d), (e), and (f): DRIFT spectra of the spent catalyst at 15, 30, 60 and 90 min of reaction,
respectively.
2
at different reaction times. (a): Adsorption of aniline, and (b):
(

1716
d
c
1697
1641
1692
1764
b
a
1
760
1
781
1
900 1850 1800 1750 1700 1650 1600
-
1
Wavenumber (cm )
Fig. 9. Theoretical IR (a) levulinic acid, (b) levulinic acid adsorbed on Ir/SiO
2
-SO
3
H, (c) step
II product (hydrogenation product interacting with -SO
reaction.
3
H groups), and (d) final product of
e
2
d
c
b
0
a
1
900 1850 1800 1750 1700 1650 1600
-
1
Wavenumber (cm )
Fig. 10. DRIFT spectra of Ir-SiO at different reaction times. (a) Adsorption of levulinic acid on
2
the catalyst surface. Spectra a and b are considered as time zero. (c), (d), (e), and (f): DRIFT
spectra of the spent catalyst at 15, 30, 60 and 90 min of reaction, respectively.

Fig. 11. Thermodynamic details following the three steps proposed for the reaction mechanism
from the reactants (R) to the products (P).
I
III
I
Fig. 12. Theoretical geometries of the first (I), second (II) and third (III) step products. Color
code: carbon in gray, hydrogen in white, nitrogen in blue, oxygen in red, and sulfur in yellow.

Fig. 13. Comparison between step II transition states with (GII-ts-Ir) and without (GII-ts) iridium
nanoparticles. Color code: carbon in gray, hydrogen in white, nitrogen in blue, oxygen in red,
and sulfur in yellow. The IR nanoparticles are represented by the blue and dark gray tubes.
Fig. 14. Relation between the electronic chemical potential of the amine reactants and the yield
(%) to pyrrolidones. Y = -4.6992 + 0.0396X, R² = 0.9472.

Fig. 15. Pictorial view of the HOMO and LUMO orbitals of amine intermediates. * Amine
used/amine intermediate: (i) Aniline / 4-phenylazanylpentanoic acid; (ii) p-toluidine / 4-[(4-
methylphenyl)amino]
pentanoic
acid;
(iii)
p-chloroaniline
/
4-[(4-
chlorophenyl)amino]pentanoic acid; (iv) m-nitroaniline / 4-[(4-nitrophenyl)amino]pentanoic
acid; (v) benzylamine / 4-[(phenylmethyl)amino]pentanoic acid; (vi) p-aminophenol/4-[(4-
hydroxyphenyl)amino]pentanoic
ylmethylamino)pentanoic acid.
acid;
(vii)
furfurylamine
/4-(furan-2-

Fig. 16. Catalyst recycles at optimized reaction conditions

Table 1. Yield and recycle of catalysts to obtain 5-methyl-N-phenyl-2-pyrrolidone.
LA
:A Catalyst Solvent Y (%) Time (h) Recycles
mmol)
Ref.
(
1
:1
Pt-MoOx/TiO
2
No solvent 90
20
1
Four cycles
[7]
3
.3:8.6
Iridicycle catalyst No solvent 91
Not evaluated [8]
No catalyst [9]
1
1
:3
:1
No catalyst
t-Bu
PHBF
In(OAc)
and PhSiH
[Ru
DMSO
34
12
12
3
Toluene
>70
Homogeneous [11]
catalyst
4
1
:1
3
Toluene
94
24
12
Homogeneous [12]
catalyst
3
2
:1
3
(CO)12
]
No solvent 93
Homogeneous [13]
catalyst

Table 2. Binding energies of Ir and IrO
2
in the solids studied and in Ir-bulk reported in the
literature.
0
0
Ir Ir4f7/2
Ir 4f7/2
IrO
2
4f7/2
IrO 4f7/2
2
Ir/SiO
Ir/SiO
2
2
59.3
62.5
61.7
64.0
63.9
65.9
-SO
3
H
60.2
63.1
62.5
Ir-Bulk
60.8 [22]
63.8
62.9 [23]
Table 3. Catalytic data obtained at 8 h of reaction in the reductive amination of levulinic acid
with aniline, where YC=N and YN-C are the yield to imine and amine, respectively. Reaction
conditions: 0.125 moles of LA and 0.25 moles of aniline at 373 K, 500 psi of H
ethyl acetate as solvent.
2
, 0.1 g of catalyst,
Catalyst
Conversion (%)
Y
C=N (%) YN-C (%)
Blank
0
0
0
SiO
2
-SO
3
H
50
6
50
0
0
Ir/SiO
2
6
2
Ir/SiO -SO
3
H
63
0
63

Table 4. Effect of catalyst mass on the yield to pyrrolidone in the reductive amination of
levulinic acid with aniline. Reaction conditions: 0.125 moles of LA and 0.25 moles of aniline at
373 K, 500 psi of H
2
, 0.1 g of catalyst, ethyl acetate as solvent.
3
Catalyst mass (g)
r
0
(mol/m . s)
N
WP
0.01
0.05
0.1
0.0014
0.0011
0.0010
2.3x10-6
1.7x10-6
1.2x10-6
Table 5. Effect of aniline:LA molar ratio on pyrrolidone yield at 8 h of reaction. Reaction
conditions: 373 K, 500 psi of H , 0.1 g of catalyst, ethyl acetate as solvent.
2
Molar ratio of α (%),
Y
C=N (%)
YN-C (%)
aniline: LA
0.125 : 0.250
0.250 : 0.250
0.250 : 0.125
63
18
4
-
-
-
63
18
4