Pd-catalyzed decarboxylation of glutamic acid and pyroglutamic acid to bio-based 2-pyrrolidone

Article abstract of DOI:10.1039/c4gc02194k

In order to recycle nitrogen from nitrogen-rich waste streams, particularly protein waste, we studied the decarboxylation of pyroglutamic acid and glutamic acid in a one-pot reaction to bio-based 2-pyrrolidone. After the screening of a wide range of supported Pd and Pt catalysts, 5 wt% Pd/Al₂O₃ displayed the highest yield (70%) and selectivity (81%) for the decarboxylation of pyroglutamic acid in water at 250°C and under an inert atmosphere. Side products originate from consecutive reactions of 2-pyrrolidone; different reaction pathways are proposed to explain the presence of degradation products like propionic acid, γ-hydroxybutyric acid, γ-butyrolactone and methylamine. An extensive study of the reaction parameters was performed to check their influence on selectivity and conversion. This heterogeneous catalytic system was successfully extended to the conversion of glutamic acid.

Full text of DOI:10.1039/c4gc02194k

View Article Online
View Journal
Green
Chemistry
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: F. De Schouwer,
L. Claes, N. Claes, S. Bals, J. Degrève and D. E. De Vos, Green Chem., 2015, DOI: 10.1039/C4GC02194K.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/greenchem

Page 1 of 9
Green Chemistry
View Article Online
DOI: 10.1039/C4GC02194K
Green Chemistry
RSCPublishing
Article
Pd-catalyzed Decarboxylation of Glutamic Acid and
Pyroglutamic Acid to Bio-based 2-Pyrrolidone
Cite this: DOI: 10.1039/x0xx00000x
a
a
b
b
c
Free De Schouwer, Laurens Claes , Nathalie Claes , Sara Bals , Jan Degrève and
Dirk E. De Vos
a
DOI: 10.1039/x0xx00000x
In order to recycle nitrogen from nitrogenꢀrich waste streams, particularly protein waste, we
studied the decarboxylation of pyroglutamic acid and glutamic acid in a oneꢀpot reaction to
bioꢀbased 2ꢀpyrrolidone. After screening of a wide range of supported Pd and Pt catalysts, 5
wt% Pd/Al O displayed the highest yield (70%) and selectivity (81%) for the decarboxylation
www.rsc.org/
2
3
of pyroglutamic acid in water at 250°C and under inert atmosphere. Side products originate
from consecutive reactions of 2ꢀpyrrolidone; different reaction pathways are proposed to
explain the presence of degradation products like propionic acid, γꢀhydroxybutyric acid, γꢀ
butyrolactone and methylamine. An extensive study of the reaction parameters was performed
to check their influence on selectivity and conversion. This heterogeneous catalytic system was
successfully extended to the conversion of glutamic acid.
Introduction
Therefore, recycling of nitrogen from protein residuals to
chemicals could result in a more efficient closure of the Nꢀ
Nitrogen is a biologically essential element for the synthesis of cycle, and in a strong decrease in energy and environmental
nucleotide and amino acid building blocks in living organisms, impact.1
as well as an essential constituent of many commodity
chemicals (solvents, bases, polymers etc.). Currently, the in plant biomass. It is therefore highly attractive for producing
HaberꢀBosch process is the main entry of nitrogen into bioꢀ industrially relevant bioꢀbased chemicals, for instance 3ꢀ
Glutamic acid is the most abundant amino acid constituent
1
1
1
3
14
based and fossilꢀderived molecules: ammonia is a major cyanopropionate , which is an intermediate to acrylonitrile
1
5
resource for plant fertilizer formulation and nitrogenꢀcontaining and succinonitrile by chemocatalysis, or γꢀaminobutyric acid
1
16
organic chemicals production. The increasing demand to via enzymatic decarboxylation. The latter can be converted
1
7
sparingly use energy and resources is forcing the exploitation of further to pyrrolidones. In addition, from pyroglutamic acid,
alternative and sustainable sources of energy, fuels and which is the cyclic condensation product, one can obtain
2
18
chemicals. Biomass constituents containing C, H and O in pyroglutaminol and prolinol by Ruꢀcatalyzed hydrogenation,
waste streams, particularly carbohydrates and lipids, have been and succinimide by oxidative decarboxylation on Ag
1
9
extensively studied as a potential renewable resource for catalysts. The latter reaction, however, uses potassium
3
ꢀ5
chemicals production.
protein fraction has received only little attention so far.
Nevertheless, exemplary biomass streams, like dried distiller decarboxylation of pyroglutamic acid to 2ꢀpyrrolidone, a
Meanwhile, the nitrogenꢀcontaining persulfate as oxidant, producing a large amount of salt waste.
This inspired us to develop a chemocatalytic system for the
7
,8
grains with solubles (DDGS) from wheat or maize, sugarbeet valuable industrial solvent and
a
precursor for
nonꢀreprotoxic ꢀalkyl
from castor oil production contain up to 20ꢀ40 wt% of pyrrolidones. Pyroglutamic acid is formed very fast at
Nꢀ
9
17
and sugarcane vinasse, poultry feather meal or press cakes vinylpyrrolidone
and
new
N
1
0
20
1
1
21
proteins; the amino acid constituents can be chemically elevated temperatures (> 120°C), and might be present in
1
2
modified into amines, amides, nitriles etc. Until now, protein biomass streams after highꢀtemperature processing. Direct
waste is mainly processed into animal feed. However, the Nꢀ decarboxylation of pyroglutamic acid to 2ꢀpyrrolidone using
cycle is rather inefficient because a large fraction of the nonꢀ metal catalysts, in a reaction similar to the known fatty acid
essential nitrogen is excreted by animals and lost to the soil decarboxylation to paraffinic hydrocarbons,22ꢀ26 can offer a
with additional production of the greenhouse gas N O. great bioꢀbased alternative to the current petrochemical
2
This journal is © The Royal Society of Chemistry 2012
Green. Chem., 2015, 00, 1-3 | 1

Green Chemistry
Page 2 of 9
View Article Online
ARTICLE
DOI: 10.1039/C4GC02194K
Green Chemistry
pathway to pyrrolidones. A reported bioꢀbased method for the Catalyst characterization
production of pyrrolidones from glutamic acid involves a twoꢀ
Powder Xꢀray diffraction (XRD) measurements were collected
on a STOE STADI P diffractometer equipped with an image
plate detector using CuKα radiation and DebyeꢀScherrer
geometry. Nitrogen physisorption measurements were
performed by using a Micromeritics 3Flex surface analyzer at
step process: an enzymatic decarboxylation to γꢀaminobutyric
1
7
acid followed by cyclization at elevated temperatures.
However, high enzyme activity requires a rather strict pH
control and sufficient supply of an expensive pyridoxal
phosphate coꢀfactor.
7
7 K. Before the measurement, the 100 mg samples were
This paper discloses
a
chemocatalytic route to
outgassed at 423 K for 6 h under vacuum. High angle annular
dark field scanning transmission electron microscopy
decarboxylate pyroglutamic acid or glutamic acid in a oneꢀpot
reaction to bioꢀbased 2ꢀpyrrolidone with supported Pd catalysts
in water, at 250°C and under inert atmosphere. Both the solvent
and the substrate are radically different from those of the
known fatty acid decarboxylation, which made us decide to
screen a much wider range of supports and metals than the most
performant reported catalysts. In addition, both pH and
temperature will be varied to check their influence on
selectivity and conversion.
(
HAADFꢀSTEM) was performed for Pd loaded Al O , SiO ,
2 3 2
ZrO and MgAl O using a FEI Tecnai operated at 200kV. In
2
2
4
case of the sample where Pd is deposited on ZrO , the contrast
2
in the HAADFꢀSTEM images was too small and energy
dispersive Xꢀray (EDX) spectroscopy was applied. To
determine acidic and basic properties of the different catalysts,
5
0 mg of the different catalysts were suspended in 3 ml H O,
2
stirred for
6 h after which the pH was measured.
Characterization data are available in the Supporting
Information.
Experimental
Support synthesis
Reaction
Besides carbon, various oxide powders were used as supporting All reactions were performed in a 50 ml highꢀpressure Parr
materials. High surface area (HSA) spinelꢀtype materials, HSAꢀ batch reactor (Series 5500; Type SSꢀ316; Model 4848
MgAl O , HSAꢀBaAl O and HSAꢀCaAl O , were synthesized temperature controller). In a typical reaction, the reactor was
2
4
2
4
2
4
2
7,28
according to literature procedures.
An aqueous solution of charged with the substrate (4 mmol), water (20 ml) and the
respectively Mg(NO ) .6H O, Ba(NO ) or Ca(NO ) .4H O noble metal catalyst (metal/substrate = 4 mol%). The reactor
3
2
2
3 2
3+
3 2
2
2+
together with Al(NO ) .9H O (M /Al = 0.5) was mixed with was sealed, purged 6 times with N while stirring and finally
3
3
2
2
poly(vinyl alcohol), with a molar ratio of metal ions to alcohol pressurized with 6 bar N2 at room temperature. Next, the
groups of 1. The mixture was stirred at 50°C till complete reactor was heated to 250°C (except for the temperature
dissolution. The pH was adjusted to 10 with ammonia (25 wt% variation experiments). During the heating step (20 min), the
in water). Next, the suspension was stirred for 3 h and aged pressure increased to 40 bar at 250°C. After 6 h, heating was
overnight. The precipitate was washed thoroughly with water stopped and the reactor was allowed to cool down. The solid
and lyophilized afterwards, leading to a fine powder. Finally, catalyst was removed by centrifugation and the reaction
the spinels were obtained after calcination at 800°C (8 h, mixture was analyzed by NMR and HPLC. For timeꢀcontrolled
3
°C/min).
experiments, the catalyst loading was reduced (metal/substrate
2 mol%) to verify initial reaction products in more detail;
time recording started when 250°C was reached and the
The metal loading of both commercial and selfꢀprepared reaction was terminated by fast cooling with water and ice.
catalysts was 5 wt%. Pdꢀbased catalysts were synthesized by
To study the effect of the pH on the conversion of
impregnating the supports with an aqueous solution of PdCl2 pyroglutamic acid and glutamic acid, a Teflon liner was loaded
and a minimal amount of HCl to ensure precursor dissolution. with the pyroglutamic acid or glutamic acid solution (0.2
in
Impregnated supports were dried overnight at 60°C. Finally, the water (15 ml)), adjusted to a certain pH with H PO or NaOH,
=
Catalyst synthesis
M
3
4
catalysts were calcined at 400°C (30 min, 2°C/min, 100 ml/min and the solid noble metal catalyst (metal/substrate = 4 mol%).
O ) and reduced at 300°C (1 h, 2°C/min, 100 ml/min H ) in a The Teflon liner was inserted into the reactor and the standard
2
2
quartz Uꢀtube. Pd/Al O was also prepared by incipient wetness reaction procedure was applied.
2
3
impregnation with a Pd(NH ) Cl .H O precursor to verify the
effect of the Pd precursor, but similar results were obtained as
3
4
2
2
Product analysis and identification
1
with PdCl . Pt/Al O was produced by incipient wetness Reaction mixtures were analyzed by HꢀNMR spectroscopy to
2
2
3
impregnation with a Pt(NH ) Cl .H O precursor and dried quantify the product yields as well as by HPLC for conversion
3
4
2
2
1
overnight at 60°C. Next, the material was calcined at 400°C (1 determination. In order to determine product yields with Hꢀ
h, 2°C/min, 100 ml/min O ) and reduced at 400°C (1 h, NMR, DMSO (δ = 2.7 ppm, s, 6H) was added as an external
2
2
°C/min, 100 ml/min H ) in a quartz Uꢀtube.
standard to the reaction mixture. The NMR samples were
2
prepared by diluting 300 µl of the reaction mixture containing
1
external standard with 300 µl of D O. HꢀNMR spectra were
2
recorded on a Bruker Avance 400 MHz spectrometer equipped
2
| Green. Chem., 2015, 00, 1-3
This journal is © The Royal Society of Chemistry 2012

Page 3 of 9
Journal Name
Green Chemistry
View Article Online
ART94ICLE
DOI: 10.1039/C4GC021 K
1
31
32
with a BBI 5 mm probe. The broad water signal in the HꢀNMR arylcarboxylic acids and very recently also itaconic acid
spectra was suppressed by applying an adapted pulse program: have been successfully decarboxylated in water with Pdꢀ and
p1 8 µs; pl1 ꢀ1 db; pl9 50 db; o1P on the resonance signal of Ptꢀbased catalysts at 250°C and 40ꢀ50 bar of N and steam. As
2
1
water, determined from the previous HꢀNMR measurement: ds carbonꢀsupported Pd and Pt were demonstrated to be the best
; ns 32; d1 5 s; aq 2.55 s; sw 16. The conversion of catalysts for decarboxylations in organic solvents as well as in
we decided in an initial experiment to expose
quantification by HꢀNMR was inaccurate as a result of pyroglutamic acid to a Pd/C catalyst in water at 250°C under 40
2
5
2
2
2,23,32
pyroglutamic acid was determined by HPLC, because water,
1
deuterium exchange with the acidic proton on the
αꢀcarbon and, bar of N . After 6 h, quantitative analysis showed that 2ꢀ
2
hence, a decreased NMR signal at δ = 4.1ꢀ4.2 ppm (t,
1
J
= 8 Hz, pyrrolidone was formed with a yield of 56% at a pyroglutamic
H). The conversion of pyroglutamic acid was directly acid conversion of 83%. Ford et al. showed that, compared to
calculated from HPLC output data recorded on an Agilent SL Pd/C, Pd/Al O3 has an even higher initial decarboxylation
2
1
2
1
200 binary system equipped with a ZORBAX ECLIPSE Plus activity for the fatty acids; however, the catalyst deactivated
3
3
C18 column (4.6 x 250 mm, 5 µm particles) and a UV detector rapidly in organic solvents. Therefore, pyroglutamic acid was
212 nm). The mobile phase was a combination of H O/acetic also exposed to a Pd/Al O catalyst under identical conditions
(
2
2
3
acid (1 v%) and acetonitrile with a gradient elution profile. as for the Pd/C catalyst. This yielded already 70% of 2ꢀ
Product identification was also performed with gas pyrrolidone at a pyroglutamic acid conversion of 85%, which
chromatography coupled to a mass spectrometer (GCꢀMS), confirmed the excellent decarboxylation activity.
using an Agilent 6890 GC, provided with a HPꢀ5ms column
Next, cyclopentanecarboxylic acid was reacted under the
and coupled to a 5973 MSD mass spectrometer, and same conditions as pyroglutamic acid with the Pd/Al O
2
3
1
13
15
additionally by 2D Hꢀ CꢀNMR and NꢀNMR spectroscopy, catalyst. However, only 8% of cyclopentanecarboxylic acid was
respectively on a Bruker Avance 400 MHz spectrometer converted to cyclopentane. This suggests that the amide group
equipped with a BBI 5 mm probe and a Bruker Avance 2 plus in pyroglutamic acid, which is electronꢀwithdrawing relative to
6
00 MHz spectrometer equipped with a BBO 5 mm probe. The the carboxylic acid, might be required for
a
high
composition of the gas phase after reaction (with Pd/ZrO ) was decarboxylation rate. The amide group might also enhance the
2
analyzed by GC, using an Interscience TraceGC equipped with adsorption of pyroglutamic acid on the catalytic sites compared
a Hayesep Q and a Molsieve 5A column connected to a thermal to cyclopentanecarboxylic acid and consequently induce a
conductivity detector (TCD) and a Rtxꢀ1 column connected to a higher decarboxylation rate.
flame ionization detector (FID). Thermodynamic calculations
of side reactions were performed with Aspen Plus software; a
®
Side product formation
detailed description is given in the Supporting Information.
In the decarboxylation of pyroglutamic acid on Pd/C or
Pd/Al O , besides 2ꢀpyrrolidone, several side products were
identified by GCꢀMS and HꢀNMR (Supporting Information):
propionic acid (PA), γꢀhydroxybutyric acid (GHB), γꢀ
butyrolactone (GBL) and pyrrolidine, with yields of 12%, 4%,
2
3
1
Results and discussion
Initial experiments with Pd/C and Pd/Al O
2
3
1
13
2
% and < 1% respectively. In addition, 2D coupled Hꢀ Cꢀ
In the context of biomass deoxygenation for the production of
biodiesel, direct decarboxylation with metallic heterogeneous
catalysts has already been investigated for the conversion of
fatty acids to hydrocarbons under inert atmosphere, or in the
1
5
NMR and NꢀNMR measurements confirmed the formation of
another side product, of which the spectroscopic signatures
agree with those of methylamine. The 2D coupled Hꢀ CꢀNMR
1
13
1
3
2
4,26
measurements showed a coupled signal between a C atom at δ
presence of a minor H pressure.
With stearic acid as a
2
1
15
=
33 ppm and a H atom at
measurements on the product mixture showed besides the
expected signal at = ꢀ258 ppm, which can be attributed to the
lactam, another clear signal at = ꢀ361 ppm, which is
consistent with the formation of an alkylamine like
δ = 2.4 ppm. The NꢀNMR
model reactant, Snåre et al. have tested the decarboxylation
activity of Ni, NiMoO , Ru, Pd, Pt, Ir, Os and Rh supported on
x
δ
silica, alumina, active carbon and several alloys. Pd and Pt
supported on carbon were identified as the most promising
δ
2
5
catalysts for this reaction at 300°C.
decarboxylation of fatty acids is mostly performed in highꢀ
boiling organic media, like ꢀdodecane; reports on this reaction
in aqueous media are scarce. However, according to the
The direct
3
4
methylamine. In principle, one could also expect the further
decarboxylation of propionic acid to ethane, although
decarboxylation of free carboxylic acids, like fatty acids, would
require temperatures above 250°C. To check whether or not
ethane is formed, propionic acid was treated under identical
reaction conditions. Only low conversion of propionic acid (<
n
2
9
principles of green chemistry, water is highly preferable for
the decarboxylation of (pyro)glutamic acid, which additionally
will often be available as aqueous solutions. Until now,
hydrothermal decarboxylation of fatty acids has been reported
2
%) was observed. After gasꢀphase analysis of the reaction with
3
0
pyroglutamic acid, only traces of ethane were found. Therefore,
2,23 the formation of ethane as a consequence of propionic acid
under nearꢀ or supercritical water conditions. Pt/C and Pd/C
2
are able to catalyze this reaction at temperatures > 330°C.
However, these conditions will probably be too severe to
preserve other functionalities in amino acids. In addition,
decarboxylation can be neglected.
To clarify whether propionic acid is produced from either
pyroglutamic acid or from 2ꢀpyrrolidone, the yields of both
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2015, 00, 1-3 | 3

Green Chemistry
Page 4 of 9
View Article Online
ARTICLE
DOI: 10.1039/C4GC02194K
Green Chemistry
propionic acid and 2ꢀpyrrolidone were measured as a function atmosphere over the Pd catalyst. Due to the excess of water and
of time (Figure 1). Immediately after heating the reactor up to the acidic conditions, this imine is readily hydrolyzed to form
2
2
50°C, small amounts of propionic acid were detected, next to an aldehyde and ammonia. This aldehyde is probably rather
ꢀpyrrolidone, suggesting that propionic acid is at least partially present as a reactive intermediate, activated at the catalyst
3
7
formed directly from pyroglutamic acid. However, after 2 h one surface by the formation of a PdꢀC or PdꢀO bond. The
observes a decrease of the 2ꢀpyrrolidone selectivity and an presence of such an intermediate provides access to two of the
increase of the propionic acid selectivity, which indicates that observed side products: on the one hand it could react to
propionic acid is also formed in a consecutive reaction from 2ꢀ propionic acid and CO by decarbonylation (CꢀC cleavage). On
pyrrolidone. After the first 2 h, the rate of pyroglutamic acid the other hand it could be hydrogenated to the corresponding
conversion started to slow down and the formation of propionic alcohol, GHB. Subsequently, GHB can undergo cyclization to
1
7
acid was significant. Minor amounts of GHB, GBL and form GBL and H O, as expected at this high temperature; or
2
pyrrolidine were only observed upon prolonged heating, the alcohol group can be hydrogenolyzed in the presence of H2
indicating that they are produced by consecutive reactions as to form butyric acid. Many of these side products could further
well. To confirm the origin of these side products, 2ꢀ degrade to lower, volatile hydrocarbons, CO and H . This
2
pyrrolidone was subjected to identical reaction conditions as would explain the increasing mass imbalance and thus a
pyroglutamic acid. Propionic acid, methylamine, GHB, GBL stronger decrease in the selectivity of 2ꢀpyrrolidone in time.
and pyrrolidine but also butyric acid were observed in the After 2 h, only 7 mol% of the converted pyroglutamic acid
reaction of 2ꢀpyrrolidone, confirming that most side products could not be accounted for, at a conversion of 60%; this
originate from 2ꢀpyrrolidone rather than directly from increased to 18 mol% after 8 h, at a conversion of 93%.
pyroglutamic acid. Remarkably, in the decarboxylation of
pyroglutamic acid, the selectivity for 2ꢀpyrrolidone increased at
short reaction times, reaching a maximum after 1ꢀ2 h. This
might be explained by the pH change during the early stage of
the reaction: the pH increased from 1.8 in the initial
pyroglutamic acid solution, to 5 at 85% conversion, because
basic amine byꢀproducts are formed.
Scheme 1. Possible reaction pathways towards the observed side products in the
decarboxylation of pyroglutamic acid; acidꢀ or baseꢀcatalyzed reactions are
marked using brackets; LA: Lewis acid.
Furthermore, aqueous reforming reactions at these temperatures
and with these catalysts are often accompanied by the waterꢀgas
3
7,38
shift (WGS) reaction.
reforming reaction, can react with water on the metal catalysts
to generate H and CO
In this way CO, from the aqueous
Figure 1. Pd/Al
of time. Conditions: pyroglutamic acid (0.2
Pd), 250°C, 6 bar N
2
O
3
catalyzed decarboxylation of pyroglutamic acid as a function
M
in 20 ml water), Pd/Al (2 mol%
2 3
O
2
.
2
:
2
It is important to note that noble metals (Pt, Ru, Rh, Pd etc.)
are able to catalyze aqueous reforming in the same temperature
range as decarboxylation (200°Cꢀ250°C). During aqueous
reforming, oxygenated molecules, in particular alcohols and
COꢀꢀ ꢁ ꢀꢀH
ꢂ
Oꢀꢀ ⇋ ꢀꢀCO
ꢂ
ꢀꢀꢁꢀꢀH
ꢂ
(1)
The production of H2 in the waterꢀgas shift reaction and
aqueous reforming can thus contribute to the formation of
butyric acid and can also explain the reduction of 2ꢀpyrrolidone
to pyrrolidine (Scheme 1). This was confirmed by adding 20
3
5,36
aldehydes,
are reformed to lower hydrocarbons, CO and H2
in a complex network of dehydration, dehydrogenation,
v% of H to the reaction atmosphere at room temperature,
3
7
2
hydrogenation, CꢀC cleavage and CꢀO cleavage reactions.
which yielded 9% of pyrrolidine after 6 h, compared to < 1% in
the absence of H . By gasꢀphase analysis, besides CO , a
Bearing this in mind, a reaction scheme explaining the byꢀ
product formation can be proposed (Scheme 1). First, the
lactam product is hydrolyzed to a γꢀamino acid; this step is
catalyzed by a Lewis acid or by a base. The resulting amine
group is easily dehydrogenated to an imine under inert
2
2
significant amount of H was detected. The quantity of CO was
2
below the detection limit of the detector, due to the position of
the waterꢀgas shift equilibrium (1) towards CO2 and H2.
4
| Green. Chem., 2015, 00, 1-3
This journal is © The Royal Society of Chemistry 2012

Page 5 of 9
Green Chemistry
View Article Online
DOI: 10.1039/C4GC02194K
a
Table 2. Screening of supports for Pd and Ptꢀbased catalysts for the decarboxylation of pyroglutamic acid
Selectivity
-pyrrolidone (%)
Selectivity
propionic acid (%)
Selectivity
others (%)
Catalyst
Conversion (%)
2
Pd/C
Pt/C
82
95
86
85
89
60
69
83
72
99
68
37
81
56
65
81
72
69
70
54
15
32
11
17
20
11
12
14
13
36
17
31
8
2 3
Pd/Al O
Pt/Al
Pd/ZrO
Pd/SiO
Pd/BaSO
Pd/CaAl
Pd/BaAl
Pd/MgAl
2
O
3
27
15
7
2
2
4
16
17
17
10
2
O
4
2
O
4
2
O4
a
Conditions: pyroglutamic acid (0.2
2 4
M in 20 ml water), catalyst (4 mol% Pd or Pt), 250°C, 6 bar N , 6 h. Others: identified C ꢀcompounds: GBL, GHB, butyric acid and pyrrolidine
and nonꢀidentified compounds: lower molecular weight compounds.
In both reactions the corresponding Cnꢀ1 acid was the main
product. Moreover, the yield of the corresponding Cnꢀ1 acid
increases with the ring size of the lactam.
Combining the previous reactions, viz. the degradation of 2ꢀ
pyrrolidone to propionic acid together with the waterꢀgas shift
reaction, one could obtain:
Catalyst screening
Pyrrolidone ꢁ 3H O → ꢀNH ꢁ PA ꢁ CO ꢁ 2H
(2)
However, next to the waterꢀgas shift reaction, part of the CO
ꢂ
ꢃ
ꢂ
ꢂ
As mentioned before, carbon has often been found to be the
best support for the decarboxylation of fatty acids, due to its
hydrophobic character, high surface area and thus suitable
adsorptive properties.
performed with smaller and more polar reactants, like
3
9
could also be hydrogenated to formaldehyde, which can react
with NH3 and H2 to methylamine. The degradation of 2ꢀ
pyrrolidone to propionic acid could then be rewritten as:
33
However, when the reaction is
Pyrrolidone ꢁ H O → ꢀPA ꢁ H C = NH
(3) pyroglutamic acid, and water is selected as solvent, the optimal
support choice could be quite different. Furthermore, supports
like Al O have a considerably higher isoelectric point than
ꢂ
ꢂ
with H C=NH as the imine derived from formaldehyde.
2
2
3
Combining twice reaction (3) and once reaction (2), thereby
including imine hydrogenation, provides the following overall
reaction:
carbon, enabling
a
Coulombic attraction between the
carboxylate group and the Al O support in the relevant pH
2
3
range (1.8ꢀ5), which would increase the coverage of the surface
with the reactant. A strong improvement was already observed
in the initial experiments using Pd/Al O . Therefore, also other
3
ꢀPyrrolidone ꢁ 5H O → ꢀ3ꢀPA ꢁ NH ꢁ CO ꢁ 2CH NH
ꢂ ꢃ ꢂ ꢃ ꢂ
(
4)
2
3
Pdꢀbased catalysts were screened to verify the influence of the
support on the activity and the selectivity in the aqueous
decarboxylation of pyroglutamic acid. Neutral or even basic
supports like BaSO and high surface area spinels of MgAl O ,
While equation (4) does not intend to be representative of the
precise stoichiometry of the overall process, it nevertheless
clearly shows that the degradation of 2ꢀpyrrolidone under inert
atmosphere, with formation of propionic acid, can be
accompanied by hydrogen formation, which is needed to
4
2
4
BaAl O and CaAl O , as well as acidic ZrO and SiO supports
2
4
2
4
2
2
were tested (Table 2). pH measurements of suspensions of the
Pd loaded supports in water demonstrated that the catalysts can
influence the acidity or basicity of the system (Supporting
Information). Pd/ZrO (yielding a pH of 3.7) and Pd/MgAl O
4
explain the formation of
a
reduced compound like
methylamine. In order to substantiate the feasibility of reactions
(
(
2) and (4), thermodynamic calculations were performed
Supporting Information). According to these calculations, both
2
2
(
pH 8.9) turned out to be the most acidic and basic catalysts
for reactions (2) and (4), the thermodynamic equilibrium is
strongly in favor of the products. Furthermore, ꢀvalerolactam
C ) and ꢀcaprolactam (C ) were subjected to the same
respectively. While Pd/BaAl O , Pd/CaAl O and Pd/BaSO are
2
4
2
4
4
δ
rather neutral or slightly basic materials, Pd/Al O , Pd/C and
2
3
(
ε
5
6
Pd/SiO show more acidic properties.
2
conditions as 2ꢀpyrrolidone (Table 1).
Blank reactions without catalyst, or using metalꢀfree carbon
or Al O gave almost no conversion; only Al O yielded 2ꢀ
a
Table 1. Catalytic degradation of different lactams by Pd/C catalysts.
2
3
2
3
pyrrolidone (3%), indicating that the alumina support is very
slightly active. On all metalꢀloaded catalysts, however,
conversions > 60% were reached within 6 h. The three main
reaction products were 2ꢀpyrrolidone, propionic acid and
methylamine, with only very small amounts of GBL, GHB,
pyrrolidine and sometimes butyric acid. Yields and selectivities
Lactam
Main product
Propionic acid
Yield (%)
2
-Pyrrolidone
20
29
44
δ-Valerolactam
Butyric acid
ε-Caprolactam
Conditions: lactam (0.2
Pentanoic acid
a
2
M in 20 ml water), Pd/C (4 mol% Pd), 250°C, 6 bar N , 6 h.
This journal is © The Royal Society of Chemistry 2012
Green. Chem., 2015, 00, 1-3 | 5

Green Chemistry
Page 6 of 9
View Article Online
ARTICLE
DOI: 10.1039/C4GC02194K
Green Chemistry
for 2ꢀpyrrolidone vary significantly among the different conversion using Pd/SiO in comparison with other supported
2
catalysts (Table 2). For example, Pd/MgAl O showed full catalysts.
2
4
conversion (> 99%), but the consecutive reaction to propionic
Le Nôtre et al. reported high selectivity in the aqueous
acid was also prominent: the yields of 2ꢀpyrrolidone and decarboxylation of itaconic acid to methacrylic acid using a 5
3
2
propionic acid were 54% and 36%, respectively. Similar effects wt% Pt/Al O3 catalyst. Therefore, Ptꢀbased catalysts were
2
were observed for Pd/ZrO catalysts, with a 2ꢀpyrrolidone and a evaluated in the decarboxylation of pyroglutamic acid as well
2
propionic acid yield of 57% and 20%, respectively. This could (Table 2). Although Pt/Al O and Pt/C were highly active, they
2
3
be attributed to the acidꢀbase properties of both catalysts: side catalyzed the consecutive reactions to propionic acid and
product formation is initiated by hydrolysis of the lactam to γꢀ butyric acid to a much larger extent than Pd. Ptꢀbased catalysts
aminobutyric acid, which is most likely catalyzed by Lewis are significantly more active for the aqueousꢀphase reforming
3
7ꢀ38
acid sites or by base (Scheme 1). Furthermore, catalysts with a reactions.
This also explains the more pronounced mass
comparable effect on the pH, e.g. Pd/SiO2, Pd/Al O , imbalance compared to Pdꢀbased catalysts: 9 mol% (Pd/C) vs.
2
3
Pd/BaSO , Pd/BaAl O and Pd/CaAl O4 show a similar 24 mol% (Pt/C), and 6 mol% (Pd/Al O ) vs. 16 mol%
4
2
4
2
2
3
selectivity for 2ꢀpyrrolidone and propionic acid as well. The (Pt/Al O ). The generation of more CO and H2 is also
2
3
2
highest selectivity to 2ꢀpyrrolidone was obtained with the evidenced by a stronger pressure increase during the reaction.
Al O ꢀ and SiO ꢀsupported catalysts. Decarboxylation with Consequently, the selectivity to 2ꢀpyrrolidone decreased to 37%
2
3
2
Pd/Al O yielded 70% 2ꢀpyrrolidone and only 9% propionic and 56% for Pt/C and Pt/Al O , respectively. Because the
2
3
2
3
acid; for Pd/SiO the yield was somewhat lower at 48%, but highest yield and selectivity were obtained with Pd/Al O , this
2
2
3
with only 5% of propionic acid, whereby the selectivity to 2ꢀ catalyst was used in further experiments. In a recycling
pyrrolidone remained high. Next to the assessment of the acidꢀ experiment, the selectivity to 2ꢀpyrrolidone was still 79% and
base properties of the catalysts, HAADFꢀSTEM imaging was the conversion decreased slightly to 76%. This small loss of
performed in combination with EDX on the Pd/Al O , Pd/SiO
activity could probably be attributed to sintering of the small
Pd/ZrO and Pd/MgAl O catalysts (Figure 2), in order to verify palladium particles at this high temperature.
2
3
2,
2
2
4
how the Pd particle size might influence reaction selectivity and
conversion.
Effects of temperature and pH; glutamic acid conversion
Low conversions (< 10%) were observed when the reaction was
performed below 200°C (Figure 3). Above 225°C significantly
higher conversions were measured and the highest yield of 2ꢀ
pyrrolidone (70%) was obtained at 250°C. The highest
selectivity to 2ꢀpyrrolidone (90%) was reached at 225°C, but it
dropped significantly to 57% at 275°C. This can be explained
by an enhanced aqueous reforming of 2ꢀpyrrolidone and other
products (in combination with the waterꢀgas shift reaction) to
volatile hydrocarbons, CO, CO and H ; at 275°C, the amount
2
2
of nonꢀidentified products was 22% based on the initial molar
input of pyroglutamic acid.
Figure 2. HAADFꢀSTEM pictures of a) Pd/SiO
2 2 4 2
, b) Pd/MgAl O , c) Pd/ZrO and
d) Pd/Al . Pd clusters can be seen as the highꢀintensity particles in the figure,
2
O
3
because of the chemical sensitivity of HAADFꢀSTEM.
Figure 3. Catalytic decarboxylation of pyroglutamic acid at different
The average particle diameter of the Pd clusters on Pd/SiO (8.8
2
temperatures using Pd/Al
and pyrrolidine; nonꢀidentified compounds are expected to include lower
in 20 ml
2 3 4
O . Identified C ꢀcompounds: GBL, GHB, butyric acid
nm) was clearly larger than for Pd/Al O (3.7 nm), Pd/ZrO
2
2
3
(
5.6 nm) and Pd/MgAl O (6.6 nm); these numbers correspond molecular weight compounds. Conditions: pyroglutamic acid (0.2
M
2
4
2 3 2
water), Pd/Al O (4 mol% Pd), 6 bar N , 6 h.
to rather broad particle size distributions (Supporting
Information). This could explain the significantly lower
6
| Green. Chem., 2015, 00, 1-3
This journal is © The Royal Society of Chemistry 2012

Page 7 of 9
Journal Name
Green Chemistry
View Article Online
ART94ICLE
DOI: 10.1039/C4GC021 K
An attempt was made to use glutamic acid, instead of
pyroglutamic acid, as the direct precursor to 2ꢀpyrrolidone. A
oneꢀpot process is proposed: in a first nonꢀcatalytic step,
glutamic acid is converted to pyroglutamic acid; the
pyroglutamic acid intermediate can then be decarboxylated in
the presence of the Pdꢀbased catalyst. Under standard reaction
conditions, the conversion of glutamic acid was nearly
complete after 6 h, whereas the conversion of pyroglutamic
acid was only 85% under identical conditions. Moreover, the
selectivity to 2ꢀpyrrolidone decreased to 63% compared to 81%
when starting from pyroglutamic acid (Figure 4).
Figure 5. Decarboxylation of pyroglutamic acid at different pH values of the
initial reaction solution adjusted with NaOH or H PO . Conditions: pyroglutamic
acid (0.2 in 20 ml water), Pd/Al (4 mol% Pd), 250°C, 6 bar N , 6 h. The
3
4
M
2
O
3
2
dotted vertical line represents a reaction without addition of acid or base.
The optimum for the H PO concentration was between 0.075
3
4
M
and 0.1 M, where almost the same yield and selectivity for 2ꢀ
pyrrolidone, 72% and 78% respectively, were obtained as when
starting from pyroglutamic acid. It is interesting to compare this
effect of acidification with the data on pyroglutamic acid
conversion of Figure 5. As explained before, when
pyroglutamic acid originates from glutamic acid in water, the
pH is slightly higher than that of a solution of genuine
pyroglutamic acid; by adding H PO to glutamic acid, the pH is
3
4
Figure
Conditions: pyroglutamic acid or glutamic acid (0.2
4 mol% Pd), 250°C, 6 bar N , 6 h. Conversions: 85% for pyroglutamic acid and
99% for glutamic acid.
4
.
Catalytic decarboxylation of different reactants with Pd/Al
2 3
O .
O
2 3
adjusted to the optimal conditions for decarboxylation of the
pyroglutamic acid intermediate, as shown in Figure 5.
M
in H O (20 ml)), Pd/Al
2
(
>
2
These differences in conversion and selectivity are probably the
result of a difference in pH of the reaction solution upon
pyroglutamic acid formation. In fact, the pH of an aqueous
pyroglutamic acid solution is around 1.8, whereas a slightly less
acidic pH is obtained at room temperature when pyroglutamic
acid is first formed from glutamic acid. These observations
suggest an optimal starting pH for achieving the highest yield
of 2ꢀpyrrolidone. Therefore the pH in the decarboxylation of
pyroglutamic acid was varied by the addition of H PO4 or
3
NaOH; however, any addition of acid or base led to poorer
results (Figure 5). Adding NaOH strongly reduced conversion
and selectivity to 2ꢀpyrrolidone. In contrast, the formation of
both 2ꢀpyrrolidone and propionic acid was suppressed in
acidified conditions. The consecutive reaction to propionic acid
Figure 6. Catalytic decarboxylation of glutamic acid in acidic conditions by
is more pronounced when the initial pH is above 4 (pH 8.2 after adjusting the pH with H
3
PO
4
. Conditions: pyroglutamic acid (0.2
, 6 h.
M in 20 ml
water), Pd/Al
2
O
3
(4 mol% Pd), 250°C, 6 bar N
2
reaction); hydrolysis of the lactam can be enhanced in more
basic conditions, initiating the different side reactions.
While the addition of acid or base has only adverse effects
Conclusion
in the case of pyroglutamic acid, H PO increased the yield and The aqueous decarboxylation of pyroglutamic acid and
3
4
selectivity to 2ꢀpyrrolidone when starting from glutamic acid glutamic acid to bioꢀbased 2ꢀpyrrolidone was successfully
Figure 6). The degradation to propionic acid was strongly performed using Pdꢀbased catalysts. Besides the conventional
(
suppressed in acidic conditions.
Pd/C, a broad series of supported Pd catalysts was screened for
the decarboxylation of pyroglutamic acid. For almost all
catalysts high conversions (> 80%) were obtained. Reactions
with Pd/Al O and Pd/SiO resulted in the highest selectivity
2
3
2
This journal is © The Royal Society of Chemistry 2012
Green Chem., 2015, 00, 1-3 | 7

Green Chemistry
Page 8 of 9
View Article Online
ARTICLE
DOI: 10.1039/C4GC02194K
Green Chemistry
(
81%) after 6 h at 250°C. With a yield of 70%, Pd/Al O is the 14. J. Le Nôtre, E. L. Scott, M. C. R. Franssen and J. P. M. Sanders,
2 3
most performant material for the production of 2ꢀpyrrolidone
from pyroglutamic acid. Optimization of the reaction 15. T. M. Lammens, J. Le Nôtre, M. C. Franssen, E. L. Scott and J. P.
parameters, like temperature and pH, for pyroglutamic acid
Sanders, ChemSusChem, 2011, , 785ꢀ791.
resulted in a selectivity to 2ꢀpyrrolidone of 90% at 225°C 16. T. M. Lammens, D. De Biase, M. C. R. Franssen, E. L. Scott and J. P.
without any pH adjustment, however at moderate conversion.
M. Sanders, Green Chem, 2009, 11, 1562–1567.
Besides pyroglutamic acid, glutamic acid was successfully 17. T. M. Lammens, M. C. R. Franssen, E. L. Scott and J. P. M. Sanders,
converted to 2ꢀpyrrolidone in a oneꢀpot reaction. The addition
Green Chem, 2010, 12, 1430ꢀ1436.
of 0.1 H PO yielded 72% of 2ꢀpyrrolidone at 250°C after 6 18. J.E. Holladay, US Pat. 7049446B2, 2004.
h, with 78% selectivity. The desired bioꢀbased lactam can thus 19. J. Deng, Q.ꢀG. Zhang, T. Pan, Q. Xu, Q.ꢀX. Guo and Y. Fu, RSC
be obtained from either pyroglutamic acid or glutamic acid.
Adv., 2014, , 27541ꢀ27544.
This is a valuable result as glutamic acid is massively available 20. B. Vandeputte, K. Moonen and P. Roose, WO 2013107822A1, 2013.
Green Chem, 2011, 13, 807ꢀ809.
4
M
3
4
4
1
1
40
in biomass waste streams or by fermentative production.
21. Y. Teng, E. L. Scott and J. P. M. Sanders, J. Chem. Technol.
Biotechnol., 2012, 87, 1458ꢀ1465.
Acknowledgements
22. J. Fu, X. Lu and P. E. Savage, ChemSusChem, 2011,
4
, 481ꢀ486.
3. J. Fu, X. Lu and P. E. Savage, Energy Environ. Sci., 2010, , 311ꢀ317
4. S. Lestari, P. MäkiꢀArvela, J. Beltramini, G. Q. Lu and D. Y. Murzin,
ChemSusChem, 2009, , 1109ꢀ1119.
2
2
3
F.D.S. and L.C. thank IWT for doctoral fellowships. D.D.V. is
grateful to KU Leuven for longꢀterm structural Methusalem
funding, to Belspo (IAPꢀPAI 7/05) and to FWO for research
project funding. S.B. and N.C. acknowledge funding by the
European Research Council (ERC Starting Grant # 335078ꢀ
COLOURATOMS). We also thank Karel Duerinckx for
assistance with NMR measurements.
2
2
2
5. M. Snåre, I. Kubičková, P. MäkiꢀArvela, K. Eränen and D. Y.
Murzin, Ind. Eng. Chem. Res., 2006, 45, 5708ꢀ5715.
6. R. W. Gosselink, S. A. Hollak, S. W. Chang, J. van Haveren, K. P. de
Jong, J. H. Bitter and D. S. van Es, ChemSusChem, 2013,
594.
6, 1576ꢀ
1
2
2
7. J. J. Guo, H. Lou, H. Zhao, X. G. Wang and X. M. Zheng, Mater.
Lett., 2004, 58, 1920ꢀ1923.
Notes and references
a
Centre for Surface Chemistry and Catalysis, Department of Microbial
8. P. G. N. Mertens, S. L. F. Corthals, X. Ye, H. Poelman, P. A. Jacobs,
B. F. Sels, I. F. J. Vankelecom and D. E. De Vos, J. Mol. Catal. A:
Chem., 2009, 313, 14ꢀ21.
and Molecular Systems, KU Leuven ꢀ University of Leuven, Kasteelpark
Arenberg 23 post box 2461, 3001 Heverlee, Belgium.
b
EMAT, University of Antwerp, Groenenborgerlaan 171, 2020 Antwerp,
Belgium
29. P. T. Anastas and J. C. Warner, Green Chemistry: Theory and
c
Department of Chemical Engineering, KU Leuven ꢀ University of
Practice. Oxford University Press, Oxford, 1998.
Leuven, Willem de Croylaan 46, post box 2423, 3001 Heverlee, Belgium
Electronic Supplementary Information (ESI) available: Catalyst
characterization and NMRꢀdata. See DOI: 10.1039/b000000x/
3
0. J. Fu, F. Shi, L. T. Thompson, X. Lu and P. E. Savage, ACS Catal.
2011, , 227ꢀ231.
31. S. Matsubara, Y. Yokota and K. Oshima, Org. Lett., 2004,
073.
32. J. Le Nôtre, S. C. M. Witteꢀvan Dijk, J. van Haveren, E. L. Scott and
J. P. M. Sanders, ChemSusChem, 2014, , 2712ꢀ2720.
E. Scott, F. Peter and J. Sanders, Appl. Microbiol. Biotechnol., 2007, 33. J. P. Ford, J. G. Immer and H. H. Lamb, Top. Catal., 2012, 55, 175ꢀ
,
1
6
, 2071ꢀ
2
1
2
.
.
Y. X. Huo, D. G. Wernick and J. C. Liao, Curr. Opin. Biotechnol.
012, 23, 406ꢀ413.
,
2
7
75, 751ꢀ762.
184.
3
4
5
7
.
.
.
.
G. W. Huber and A. Corma, Angew. Chem., 2007, 46, 7184ꢀ7201.
A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411ꢀ2502.
J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539ꢀ554.
34. R. M. Silverstein, F. X. Webster and J. D. Kiemle, Spectrometric
Identification of organic spectra, 7 edn., John Wiley & Sons, Inc.,
New York, 2005.
C. O. Tuck, E. Pérez, I. T. Horváth, R. A. Sheldon and M. Poliakoff, 35. M. Mavrikakis and M. A. Barteau, J. Mol. Catal. A: Chem., 1998,
Science, 2012, 337, 695ꢀ699.
131, 135ꢀ147.
R. L. Belyea, K. D. Rausch and M. E. Tumbleson, Bioresour. 36. G. W. Huber and J. A. Dumesic, Catal. Today, 2006, 111, 119ꢀ132.
8
.
.
Technol., 2004, 94, 293ꢀ298.
37. R. R. Davda, J. W. Shabaker, G. W. Huber, R. D. Cortright and J. A.
9
1
P. G. Dalev, Bioresour. Technol., 1994, 48, 265ꢀ267.
Dumesic, Appl. Catal., B, 2003, 43, 13ꢀ26.
0. C. Martín, A. Moure, G. Martín, E. Carrillo, H. Domínguez and J. C. 38. J. W. Shabaker, G. W. Huber, R. R. Davda, R. D. Cortright and J. A.
Parajó, Biomass Bioenergy, 2010, 34, 533ꢀ538.
Dumesic, Catal. Lett., 2003, 88, 1ꢀ8.
1. T. M. Lammens, M. C. R. Franssen, E. L. Scott and J. P. M. Sanders, 39. A. Gotti and R. Prins, J. Catal., 1998, 175, 302ꢀ311.
1
1
1
Biomass Bioenergy, 2012, 44, 168ꢀ181.
40. W. Leuchtenberger, K. Huthmacher and K. Drauz, Appl. Microbiol.
Biotechnol., 2005, 69, 1ꢀ8.
2. J. Sanders, E. Scott, R. Weusthuis and H. Mooibroek, Macromol.
Biosci., 2007, 7, 105ꢀ117.
3. L. Claes, R. Matthessen, I. Rombouts, I. Stassen, T. De
Baerdemaeker, D. Depla, J. A. Delcour, B. Lagrain and D. E. De
Vos, ChemSusChem, 2015, 8, 345ꢀ352.
8
| Green. Chem., 2015, 00, 1-3
This journal is © The Royal Society of Chemistry 2012

Page 9 of 9
Green Chemistry
View Article Online
DOI: 10.1039/C4GC02194K
Table of contents entry
A palladium based catalytic system was developed to decarboxylate glutamic acid and
pyroglutamic acid to bio-based 2-pyrrolidone, in aqueous media at 250°C and inert
atmosphere.