Chemistry Reference
In-Depth Information
H
2
O
2
epoxide
H
2
O
RCO
2
H
H
2
O
2
sulphuric
acid loop
solvent loop
distillation
reconcentration
RCO
3
H
H
2
O
Fig. 11.1
8
Peracid-based continuous
epoxidation process.
olefin
substrate is auto-oxidised by oxygen to a hydroper-
oxide, which then epoxidises the propylene in the
presence of a molybdenum- or titanium-based cat-
alyst:
dised partly to CO
2
and water, and the target here is
to minimise the amount needed for the reaction
rather than to recycle the residue.
Replacement of organic peracids by catalytic
methods is not easy to achieve here. The size of the
substrate molecules and the viscosity of the oils
do not lend themselves to heterogeneous liquid-
phase catalysis. The oils often are used for food
contact applications, so homogeneous metal-
catalysed systems must avoid any introduction of
metals into the product. In fact, there is no great
driving force to change on waste minimisation
grounds. It is, however, noteworthy that another
sizeable application—the production of epoxysuc-
cinic (and tartaric) acid from maleic anhydride
[162]—does use tungstate catalysis successfully to
make products with food-related applications (Fig.
11.19). In one disclosure [163], epoxysuccinic acid is
produced from maleic acid and H
2
O
2
, using Na
2
WO
4
at pH 4.8, for 6 h at 90°C. A glucamine chelating
resin is used to remove and recycle the tungstate,
with 99% efficiency.
There are currently many smaller epoxidation
applications using a variety of methods, but infor-
mation on these is often not made public. Certainly
metal-peroxo catalytic chemistry, including the
Ishii-Venturello system, is now in routine use along-
side traditional peracid technology. A recent practi-
cal method from Noyori's group embodies much of
what has been learned to date about the 'simple'
tungstate system [89b,164].
Homogeneous catalytic methods typically use two-
phase liquid systems with a phase-transfer catalyst.
The organic phase used to be a chlorinated hydro-
carbon, but solvents such as toluene and propylene
carbonate are now established in this role. Single-
RH + O
2
Æ ROOH
ROOH + CH
3
CH=CH
2
Æ epoxide + ROH
Matching demand for the co-product with that for
propylene oxide is a continuing challenge.
Despite their drawbacks, there has been investment
in both routes recently to raise capacity. Epichlorhy-
drin is produced mainly from allyl chloride by the
chlorhydrin route. As shown below, however, small
molecules can be epoxidised effectively by H
2
O
2
and
titanium silicalite, and there is keen interest at present
in the prospective manufacture of propylene oxide
and epichlorhydrin in this way.
The largest established application for H
2
O
2
in
epoxidation is for unsaturated natural oils such as
soybean, the epoxidised forms (ESBO, etc.) being
used chiefly as stabilising plasticisers for poly(vinyl
chloride) (ca. 100 kt year
-1
). The oil itself acts as
solvent, although a volatile hydrocarbon such as
hexane may be added to facilitate processing. An
amount of carboxylic acid is added (typically 0.14
units per unit of oil for formic and 0.2 for acetic, the
latter also requiring about 0.01 units of H
2
SO
4
cata-
lyst) and the H
2
O
2
, typically at 70% strength, is
added gradually to this mixture at 60°C. Complete
reaction to ESBO takes 10-15 h, giving an oxirane
content of 6.5-7.1% (80-85% of theory). Small
amounts of olefin remain unreacted and some
undergoes ring-opening to the diol or its monoester.
This process produces very little waste: residual car-
boxylic acid is separated by washing and can, in the
case of acetic acid, be recycled. Formic acid is oxi-
