Selective hydrogenation of maleic anhydride to γ-butyrolactone over Pd/Al2O3 catalyst using supercritical CO2 as solvent

Article abstract of DOI:10.1039/b111047k

A selective hydrogenation of maleic anhydride to either γ-butyrolactone or succinic anhydride over simple Pd/Al₂O₃ catalyst under supercritical CO₂ medium is described for the first time which has considerable promise for both lab-scale as well as industrial selective hydrogenations of low vapor pressure compounds without employing environmentally harmful organic solvents.

Full text of DOI:10.1039/b111047k

Selective hydrogenation of maleic anhydride to g-butyrolactone over
Pd/Al O catalyst using supercritical CO as solvent
2
3
2
Unnikrishnan R. Pillai and Endalkachew Sahle-Demessie*
National Risk Management Research Laboratory (NRMRL), Sustainable Technology Division, MS 443,
United States Environmental Protection Agency (US-EPA), Cincinnati Ohio-45268, USA.
E-mail: Sahle-Demessie.Endalkachew@epa.gov; Fax: 001-513-569-7677
Received (in Corvallis, OR, USA) 21st November 2001, Accepted 4th January 2002
First published as an Advance Article on the web 1st February 2002
A selective hydrogenation of maleic anhydride to either g-
butyrolactone or succinic anhydride over simple Pd/Al
catalyst under supercritical CO medium is described for the
first time which has considerable promise for both lab-scale
as well as industrial selective hydrogenations of low vapor
pressure compounds without employing environmentally
harmful organic solvents.
hydrogenation catalysts, but are not known to be effective as
1
5
2
O
3
partial/selective hydrogenation catalysts like Pt or Rh.
2
Nevertheless, selectivities different from those obtained under
normal liquid phase hydrogenations in organic solvents are
reported here under supercritical conditions.
2 3
A 1% Pd/Al O (w/w) catalyst was prepared by wet
impregnation of alumina pellets (1400 m, Aldrich) with a
solution (0.02 M) of palladium chloride (Aldrich). After the
impregnation step, the catalyst was dried at 110 °C over night
and then calcined in air at 450 °C for 5 h followed by reduction
in flowing hydrogen at 450 °C for 5 h.
Hydrogenation of maleic anhydride (MA) is an important
industrial reaction as all its products viz., succinic anhydride
(
SAH), succinic acid (SA), g-butyrolactone (GBL) and tetra-
hydrofuran (THF), are commodity chemicals of considerable
industrial importance.1 GBL is currently one of the most
valuable alternatives to the environmentally harmful chlorin-
ated solvents, which have been widely used in the polymer and
paint industries. Hydrogenation of MA is the most direct way to
produce GBL and it does not involve the use of hazardous
materials. GBL is currently being manufactured by the vapor
phase hydrogenation of MA using reduced copper chromite
catalysts containing some physical and chemical promoters.2
Cr-containing catalysts are getting increasingly difficult to use
owing to their toxicity. Hydrogenation of MA over different
noble metal and Cu based catalysts both in vapor and liquid
Isothermal hydrogenation of MA was conducted in a 500 mL
stainless steel batch reactor using 2.45 g MA (25 mmol) and 0.5
g catalyst loaded into a spinning dynamic basket. The reactor
was then heated to the desired temperature while being filled
with known amounts of hydrogen and carbon dioxide based on
the set pressure. After the reaction, the reactor was cooled to
room temperature and vented slowly. The products were
collected and analyzed by a Hewlett-Packard 6890 gas
chromatograph using a HP-5 5% phenyl methyl siloxane
capillary column (30 m 3 320 mm 3 0.25 mm) and a quadruple
mass filter equipped HP 5973 mass selective detector. Quantifi-
cation of the products was obtained using a multi-point
calibration curve for each product.
3
phases have been reported by many research groups. Recent
patent literature suggests that Al-containing catalysts may
represent an interesting and promising alternative to the
chromite catalysts.4 A recent study reports a two-stage liquid
phase hydrogenation of MA to GBL over supported Pd catalyst
The results of the hydrogenation experiments in Sc-CO
2
medium are shown in Table 1. MA conversion and product
yields are also compared with the conventional liquid phase
reaction under similar conditions, using commonly employed
solvents such as polyethylene glycol (PEG), ethylene glycol
dimethyl ether (EGDME), acetone as well as with the reaction
a–c
followed by a homogeneous ruthenium catalyst system consist-
3 3
ing of Ru(acac) , P(octyl)
and toluene p-sulfonic acid.5
a
However, the activity is low besides the fact that homogeneous
catalytic systems suffer from the problem of catalyst separation
from the reaction mixture. Another drawback with the reported
efficient catalytic transformation of MA to GBL is the use of
expensive solvents like ethylene glycol dimethyl ether and
polyethylene glycol dimethyl ether.5
under N
80% yield to GBL was obtained at 200 °C and at a pressure of
2.1 MPa H and 12 MPa CO (Scheme 1). The high conversions
2
pressure. A 100% MA conversion with approximately
2
2
( ~ 100%) obtained irrespective of the variations in pressure and
temperature or the type the reaction medium could be the result
of Pd being a good hydrogenation catalyst and the high catalyst-
to-substrate ratio. Product selectivity, however, varied sig-
nificantly with changes in temperature, pressure as well as the
reaction medium (Table 1). It was seen that GBL selectivity
b
In the past decade, supercritical carbon dioxide (Sc-CO
been increasingly used as an environmentally friendly reaction
medium. The ability of the Sc-CO to dissolve many of reactive
gases like H , O and also a variety of organic compounds
2
) has
2
2
2
2
increased with increase in CO pressure. Temperature was
facilitates oxidation and hydrogenation reactions in this me-
dium, thereby eliminating interphase mass transfer limitations
critical in obtaining the desired GBL yield or selectivity. GBL
yield/selectivity increased with increase in reaction tem-
perature. A high GBL yield could not be obtained at 150 °C
even at pressures higher than 12 MPa. It is obvious from the
results that the desired product selectivity can be obtained by
controlling the temperature and pressure. On the other hand,
hydrogenation in the presence of organic solvents produced the
double bond hydrogenated product, SAH, selectively. In the
case of liquid phase reaction, there is an additional requirement
6
and enhancing the reaction rate. The fact that slight changes in
temperature and pressure near the critical point can change the
activity and selectivity significantly provides remarkable poten-
tial for its use as a reaction medium. Several recent publications
have demonstrated the potential of Sc-CO
reaction medium for a variety of synthetic transformations.
However, relatively few studies have been reported on
2
as an alternative
6,7
hydrogenation reactions in supercritical fluid media.8
has been reported to have a fairly good solubility in liquid CO
7.5%).14 It is expected to show a much higher solubility at
higher temperatures and pressures. Therefore we decided to
investigate the hydrogenation of MA in Sc-CO medium. This
–13
MA
2
(
2
communication presents the first example of organic solvent-
free hydrogenation of a low vapor pressure compound under
supercritical conditions over a simple alumina supported
palladium catalyst. Supported Pd catalysts are highly efficient
Scheme 1
4
22
CHEM. COMMUN., 2002, 422–423
This journal is © The Royal Society of Chemistry 2002

2 3 2
Table 1 Maleic anhydride hydrogenation over 1% Pd/Al O under supercritical CO
mediuma
Yield (%)b
GBL
CO
pressure/MPa
2
Temperature/°C
Organic solvent
Duration/h
Conversion (%)
SAH + SA
200
200
200
200
200
200
200
180
150
200
200
200
12
No solvent
No solvent
No solvent
No solvent
No solvent
No solvent
No solvent
No solvent
No solvent
2
2
2
1
2
2
2
2
2
2
2
2
100
100
100
100
100
100
100
100
100
100
100
100
~ 80c
56
42
35
48
30
30
28
5
ND
44
58
8.3
8.3
8.3
3.5
2.1
12 (N
12
12
--
d
65
52
70
e
2
)
60
63
95
78
80
96
f
PEG (10 mL)
EGDME (10 mL)g
Acetone (10 mL)
2
7
2
--
--
a
Reaction conditions: MA = 25 mmol, catalyst = 0.5 g, H
2
= 2.1 MPa, stirrer = 350 rpm, duration = 2 h, ND = below the method detection limit of
b
c
GC-MS. GC yield. Exact values could not be ascertained above 80% as the GC-MS signals for SAH and SA at these concentrations were close to their
method detection limits. 0.25 g catalyst. Under N
2
pressure. f PEG = polyethylene glycol. g EDGME = ethylene glycol dimethyl ether.
d
e
of removing the high boiling solvents (PEG, EGDME) from the
reaction mixture. The gas phase reaction under pressurized
nitrogen, at 200 °C, yielded better GBL selectivity than the
organic liquid phase reactions, but much lower than the
desired end by a simple maneuver of the reaction temperature
and pressure. A more systematic investigation is underway to
determine the phase behavior of the mixture and the reaction
kinetics to understand the physicochemical processes, which
will help in addressing issues like reactor design, process
development and product separation.
U. R. Pillai is a postgraduate research participant at the
National Risk Management Research Laboratory administered
by the Oak Ridge Institute for Science and Education through an
interagency agreement between the US Department of Energy
and the US Environmental Protection Agency. The authors are
grateful to Douglas Young for conducting phase equilibria
calculations.
selectivity obtained under Sc-CO
temperature. However, carrying out the hydrogenation reaction
under N medium poses an explosion hazard due to the
possibility of sudden temperature shoot-out and therefore is not
very safe. This, however, is not the case with Sc-CO medium
as it possesses high heat capacity and hence can act as a good
heat sink. The unusual GBL selectivity obtained over Pd/Al
may be a consequence of the enhanced reactant-product
2
medium at the same
2
2
2 3
O
solubility as well as different reaction energetics in Sc-CO
which require further studies in detail.
2
,
Phase behavior experiments of MA in Sc-CO
ducted at 150 °C using a 75 mL stirred high pressure view cell
containing 0.5 g MA, in the presence of a constant H pressure
of 2.1 MPa and varying CO pressure in the range 5.5 to 14
MPa. MA showed no miscibility at 5.5 MPa. However, the
miscibility increased with increase in CO pressure and showed
complete miscibility at 11 MPa CO pressure (a homogeneous
2
were con-
Notes and references
2
1
(a) N. Harris and M. W. Tuck, Hydrocarbon Process, 1990, 69, 79; (b)
A. M. Brownstein, Chemtech, 1991, 21, 506.
2
2
G. L. Castiglioni, A. Vaccari, G. Fierro, M. Inversi, M. L. Jacono, G.
Minelli, I. Pettiti, P. Porta and M. Gazzano, Appl. Catal. A. General,
2
2
1
995, 123, 123.
and uniform single phase was visible). The solubility experi-
ments could not be conducted at 200 °C due to the temperature
limitation of the pressure view cell. Nevertheless, the above
studies tend us to believe that MA should be completely
3 (a) M. Messori and A. Vaccari, J. Catal., 1994, 150, 177; (b) G. L.
Castiglioni, M. Ferrari, A. Guercio, A. Vaccari, R. Lancia and C.
Fumagalli, Catal.Today, 1996, 27, 181.
4
(a) W. De. Thomas, P. D. Taylor and H. F. Tomfohrde III, US Pat.,
149836, 1992; (b) D. S. Thakur, B. R. Roberts, T. J. Sullivan and A. L.
5
miscible with Sc-CO
in the hydrogenation studies. Phase behavior calculation of a
mixture of CO , H , and MA, conducted using the Peng-
2
at the experimental conditions employed
Vichek, US Pat., 5155086, 1992; (c) R. W. Wegman and D. R. Bryant,
US Pat. 5191091, 1993.
2
2
5
(a) Y. Hara, H. Kusaka, H. Inagaki, K. Takahashi and K. Wada, J.
Catal., 2000, 194, 188; (b) M. W. M. Tick, M. A. Wood and A. G. Hiles,
PCT Int. Pat., Appl. No. WO 9743234 1997.
Robinson equation of state subroutine built in to ASPEN plus
software, showed that the reactor composition is a single phase
at 200 °C. The loss of any product during venting of the reaction
vessel was tested by passing the cooled reactor vent through a
chilled methanol trap and later analyzing the methanol solution.
However, no reactant or products were detected by GC, which
suggests no appreciable volatile compound loss during the
venting of the autoclave. Also, no significant leaching of Pd by
6 A. Baiker, Chem. Rev., 1999, 99, 453.
7 (a) E. Sahle-Demessie, M. A. Gonzalez, J. Enriquez and Q. Zhao, Ind.
Eng. Chem. Res., 2000, 39, 4858; (b) O. Krocher, R. A. Koppel and A.
Baiker, Chem. Commun., 1996, 1497; (c) M. G. Hitzler, F. R. Smail, S.
K. Ross and M. Poliakoff, Chem. Commun., 1998, 359.
8
B. Minder, T. Mallat, K. H. Pickel, K. Steiner and A. Baiker, Catal.Lett.,
995, 34, 1.
A. Bertucco, P. Canu and L. Devetta, Ind. Eng. Chem. Res., 1997, 36,
626.
1
2
Sc-CO was evident as a regenerated catalyst (calcination and
9
reduction of the spent catalyst) showed almost similar activity
and GBL selectivity. Some studies have reported the potential
2
1
0 (a) O. Krocher, R. A. Koppel, M. Fobra and A. Baiker, J. Catal., 1998,
178, 284; (b) C. A. Thomas, R. J. Bonilla, Y. Huang and P. G. Jessop,
Can. J. Chem., 2001, 79, 719.
deactivation of hydrogenation catalysts due to the formation of
carbon monoxide or metal formate.8
,10
In this study no
significant deactivation of the Pd sites was observed for the two-
11 D. Chouchi, D. Gourgouillon, M. Courel, J. Vital and M. Nunes da
Ponte, Ind. Eng. Chem. Res., 2001, 40, 2551.
12 (a) M. G. Hitzler and M. Poliakoff, Chem. Commun., 1997, 1667; (b) M.
G. Hitzler, F. R. Smail, S. K. Ross and M. Poliakoff, Org. Proc. Res.
Dev., 1998, 2, 137.
cycle two hour run tests. Similarly, other researchers observed
little deactivation after an extended period of tests.13
In conclusion, this study demonstrates that selective hydro-
genation of a low vapor pressure compound like MA can be
successfully carried out in Sc-CO medium using a simple
2
1
3 V. Arunajatesan, B. Balasubramaniam, K. W. Hutchenson and F. E.
Herkes, Chem. Eng.Sci., 2001, 56, 1363.
supported metal catalyst thereby accomplishing the goal of
green chemistry cost effectively. It further demonstrates the
1
4 A. W. Francis, J. Phys. Chem., 1954, 58, 1099.
15 R. F. Heck, Palladium Reagents in Organic Synthesis, Academic Press,
potential of Sc-CO
2
medium for tuning the reaction to the
New York, 1985.
CHEM. COMMUN., 2002, 422–423
423