CHEMSUSCHEM
FULL PAPERS
slightly lower selectivities. Ru was a more active metal for the
pled by broad band decoupling. HPLC was performed by using
a Thermo Fisher Scientific autosampler and pump (eluent: 12 mm
conversion of itaconic acid but led to lower MAA selectivity (<
ꢁ1
H SO , pump flow: 0.6 mLmin ), with an ROA Organic acid 150ꢃ
25%). Compared to benchmark reaction procedures that are
2
4
7
.8 mm column (column temperature: 408C) and an injection
performed in near-critical and supercritical water conditions,
MAA can be obtained at temperatures of 200–2508C without
any external added pressure. Moreover, an intensive optimisa-
tion of the reaction parameters was completed to attain better
conversion and selectivity. Two different purification tech-
niques were tested, and MAA was isolated in ꢀ50% yield.
Citric acid, a widely available bio-based feedstock, was suc-
cessfully used as a substrate, and the one-pot dehydration–
double decarboxylation pathway resulted in the formation of
MAA with 41% selectivity by using a Pd/Al O catalyst. A two-
volume of 50 mL. UV detection was performed at 210 nm, and the
quantification of compounds was achieved by calibration with
phthalic acid as an internal standard and checked by calibration
from pure products. Retention times [min] were: citric acid (4.3),
pyruvic acid (5.5), citraconic acid (6.4), itaconic acid (7.3), mesaconic
acid (10.3), MAA (13.1), crotonic acid (14.5), phthalic acid (16.3, in-
ternal standard).
Typical experimental procedure for the decarboxylation of
itaconic acid
2
3
step procedure via the isolation of aconitic acid did not lead to
better results.
In a typical experiment, itaconic acid (0.4 g, 3.07 mmol) and the
solid transition-metal catalyst (2.5 mol% of metal) were weighed in
air and placed in a glass liner with a stirrer bar. The liner was intro-
duced into the reactor, and the atmosphere was purged twice
with Ar. The basic aqueous solution was prepared in advance at
the appropriate concentration (NaOH, 0.15m) and degassed with
Ar by bubbling gas through the solution for 1 h. The basic solution
(20 mL) was added to the reactor with a syringe. The reactor was
closed, purged three times with Ar, stirred (1000 rpm) and heating
began (typically heating up to 2508C was achieved in 30 min). The
internal pressure reached 38 bar. The recorded reaction time start-
ed when the desired temperature was reached. After the allocated
reaction time, the reactor was allowed to cool to RT (1–1.5 h) and
the gas pressure was released. The crude solution was purified by
micro-filtration (Millipore micro-filters, 0.20 mm), the pH was mea-
sured and an aliquot of the crude reaction mixture was diluted
with MilliQ-grade water prior to analysis by HPLC.
Hence, the production of bio-based MAA, a precursor of bio-
based poly(methyl methacrylate), can be achieved in one step
in modest to good selectivity under relatively mild conditions
from substrates available from biomass fermentation or raw
fruit juices. Even if the overall molecular mass balance appears
detrimental, for instance, there is a 55% molecular mass loss
from citric acid to MAA, it is possible that this process could
be economically viable. Firstly, it takes place in an aqueous
medium and, therefore, is potentially compatible with the
nature of fermentation broths obtained directly from the pro-
duction of itaconic acid and citric acid. If these broths are used
as the starting material, the purification of the aforementioned
substrates before decarboxylation will be obsolete. Secondly,
the solid catalysts that we employed are known to be recycla-
ble and can be used in continuous processes, which will con-
stitute an added benefit to the process. Both implementations
of our decarboxylative synthesis of MAA, together with an eco-
nomic feasibility study of large-scale production, are currently
under investigation in our laboratory and will be reported in
due course.
Purification of crude reaction mixtures and isolation of MAA
With the aim to obtain an isolated yield from our metal-catalysed
decarboxylation procedure, we performed two series of six identi-
cal reactions in our parallel reactors setup with itaconic acid (0.6 g)
as the substrate, one equivalent of base in water (30 mL) and Pt/
Al O as the catalyst in each reactor. The purification of the two
2
3
series of reaction was performed by distillation and solvent extrac-
tion.
Experimental Section
MilliQ-grade water was used for all reactions. Sodium hydroxide
Distillation followed by extraction: After the six identical itaconic
acid decarboxylation reactions were performed (6ꢃ0.6 g of itacon-
ic acid in 30 mL of 0.15m NaOH solution, Pt/Al O catalyst
(
98%), calcium hydroxide (96%), itaconic acid (2-methylenesuccinic
acid, 99%), citric acid (2-hydroxypropane-1,2,3-tricarboxylic acid,
2
3
9
9
9%), trans-aconitic acid ((E)-prop-1-ene-1,2,3-tricarboxylic acid,
(2.5 mol% Pt)), the crude mixtures were filtered (micro-filtration
0.2 mm) and combined together to form the solution Crude1
(175 mL). HPLC analysis revealed the following composition: 60%
of MAA, 3% of itaconic acid isomers and 37% of byproducts. The
solution Crude1 (175 mL), was subjected to azeotropic vacuum dis-
tillation (20 mbar) to give a main fraction of 107 mL (20 mbar,
vapour temperature of 548C). This fraction was analysed by HPLC
and showed the following composition: 99.0% MAA, 0.7% crotonic
acid, 0.2% pyruvic acid and 0.1% other compounds. Extraction
8%) and CDCl (99.8 atom% D) were purchased from Sigma–Al-
3
drich. The catalysts had the following specifications and origins:
Pd/C (5 wt%, Sigma–Aldrich), Pd/Al O3 (10 wt%, Sigma–Aldrich),
Pd/BaSO (10 wt%, Fluka), Pearlman’s catalyst (20 wt% Pd(OH) /C,
Sigma–Aldrich), Ru/C (5 wt%, Sigma–Aldrich), Pt/C (10 wt%,
Sigma–Aldrich), Pt/Al O (5 wt%, Sigma–Aldrich). Ar gas was sup-
plied by Linde Gas Benelux (Instrument argon 5.0). Unless stated
otherwise, all chemicals were used as received and used without
any further purification.
2
4
2
2
3
with Et O (3ꢃ50 mL) followed by drying the combined organic
2
layers over MgSO and evaporation to dryness led to the recovery
of a colourless oil. NMR spectroscopy showed the presence of pure
MAA (1.22 g, 51% isolated yield, >99.0% purity (HPLC)).
4
High-pressure reactions were performed in a parallel Parr system
reactor (Series 5000, 6ꢃ75 mL, Hastelloy C-276) equipped with
glass liners and glass-coated magnetic stirring bars at 1000 rpm.
NMR spectra were recorded by using a Bruker Avance III spectrom-
Extraction followed by distillation: Another set of six identical ita-
conic acid decarboxylation reactions was performed (6ꢃ0.6 g of
itaconic acid in 30 mL of 0.15m NaOH solution, Pt/Al O catalyst
1
13
1
eter operated at 400.17 MHz ( H) and 100.62 MHz ( C). H NMR
chemical shifts are quoted in parts per million (ppm) referenced to
2
3
13
the appropriate solvent peak. C NMR spectra were fully decou-
(2.5 mol% Pt)). The crude mixtures were filtered (micro-filtration
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2712 – 2720 2719