Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
3~
REACTOR CONFIGURATION FOR
ALKYLENE O~IDE PROD~CTION PROCESS
(D#80,361-F)
CROSS-REFERENCES TO RELATED APPLICATIONS
This application is related to Canadian patent
application Serial No. 483,414, filed June 7, 1985, which is
concerned with an improved,method for making alkylene oxides
from alkenes using a non-acidic molybdenum catalyst, and
Canadian patent application Serial No. 484,477, filed June 19,
1~ 1985, which is concerned with unusually high molybdenum
recovery in methods for making alkylene oxides, both filed of
even date. This application is also related to Canadian patent
application Serial No. 483,290, filed June 19, 1985, which is
concerned with a method for making alkylene oxides with an
unusually low alkylene oligomer by-product production.
BACKGROU~D OF THE_INVE~TION
1. Field of the Invention
The invention relates to the catalytic production of
alk~lene oxides from olefinically unsaturated organic compounds
2~ and more particularly relates to the productions of alkylene
oxides using particular reactor configurations.
2. Other Related Method in the Field
.. . . . . . ...
It is well known that the epoxidation of olefins to
give various oxide compounds has long been an area of study by
those skilled in the art. It is equally well known that the
reactivities of the various olefins differs with the number of
substituents on the carbon atoms involved in the double bond.
Ethylene itself has the lowest relative rate of epoxidation,
with propylene and other alpha olefins being the next slowest.
Compounds of the formula R2C-CR2
-1-
~here R simply represents alkyl or other substituents may be
epoxidjzed fastest.
Of course, the production of ethylene oxide from
ethylene has long been known to be accomplished by reaction
with molecular oxygen over a silver catalyst. Numerous
patents have issued on various silver-catalyzed processes
or the production of ethylene oxide.
Unfortunately, the silver catalyst route is poor
for olefins other than ethylene. For a long time the com-
mercial production of propylene oxide could only be accom-
- plished via the cumbersome chlorohydrin process.
Another commercial process for the manufacture of
substituted oxides from alpha olefins such as propylene was
not discovered until U. S. Patent 3,351,635 taught that an
organic oxide compound could be made by reacting an olefini-
cally unsaturated compound with an organic h droperoxide in
the presence of a molybdenum, tungsten, titanium, columbium,
tantalum, rhenium, selenium, chromium, zirconium, tellurium
or uranium catalyst. U. S. Patent 3,350,422 teaches a simi-
lar process using a soluble vanadium catalyst. Molybdenum
is the preferred catalyst. A substantial excess of olefin
relative to the hydroperoxide is taught as the normal pro
cedure for -the .reaction. See also U. S. Patent 3,525,645
which teaches the 510w addition of organic hydroperoxide
to an excess of olefin as preferred.
However, even though this work was recognized as
extremely important in the development of a commercial
propylene oxide process that did not depend on the chloro~
hydrin route, it has been recognized that the molybdenum
process has a number of problems. For example, large
3~
quantities of the alcohol corresponding -to the peroxide used
were formed; if t-butyl hydroperoxide was used as a co-react-
ant, then a use or market for _-butyl alcohol had to be found.
Other troublesome by-products were the olefin oligomers.
If propylene was used, various propylene dimers, sometimes
called hexenes, would result. Besides being undesirable in
that the best use of propylene was not made, prcblems would -
also be caused in separating the desired propylene oxid~
from the product mix. In addition, the molybdenum catal~st
may not be stable or the recovery of the catalyst for recycle
may be poor.
A number of other methods for the production of
alkylene oxides from epoxidizin~ olefins (particularly
propylene) have been proposed. U. S. Patent 3,655,777 to
Sargenti reveals a process for epoxidizing propylene using a
molybdenum-containing epoxidation catalyst solution prepared
by heating molybdenum powder with a stream cont~ining unre-
acted tertiary butyl hydroperoxide used in the epoxidation
process as the oxidizing agent and polyhydric compounds.
~0 The polyhydric compounds are to have a molecular weight from
200 to 300 and are to be formed as a by-product in the epoxi-
dation process. A process for preparing propylene oxide by
direct oxidation of propylene with an organic hydroperoxide
in the presence of a catalyst ( such as molybdenum or vanad-
ium) is described in British Pa-tent 1,338,015 to Atlantic-
Richfield. The improvement therein resides in the inclusion
of a free radical inhibitor in the reaction mixture to help
eliminate the formation of Cs to C~ hydrocarbon by products
which must be removed by extractive distillation. Proposed
_3_
3~;
free radical inhibitors are tertiary butyl catechol and
2,6-di-t-butyl-4-methyl phenol.
Stein, et al. in U. S. Patent 3,~49,451 have im~
proved upon the Kollar process of U. S. Paterts 3,350,422
and 3,351,635 by requiring a close control of the reaction
temperature, between 90-200C and autogeneous pressures,
among other parameters. The primary benefits seem -to be im-
proved yields and reduced side reactions. The molybdenum-
catalyzed epoxidation of alpha olefins and alpha substituted
olefins with relatively less stable hydroperoxides may be
accomplished according to IJ. S. PatPnt 3,862,961 to Sheng,
et al. by employing a critical amount of a stabilizing agent
consisting of a C3 to Cg secondary or tertiary mcnohydric
alcohol. The preferred alcohol seems to be terti2ry butyl
alcohol. Citric acid is used to minimize the iron-catalyzed
decomposition of the organic hydroperoxide withou-t adversely
affecting the reaction between the hydroperoxidé and the ole-
fin in a similar oxirane producing process taught by Herzog
in U. S. Patent 3,928,393. The inventors in U. S. Patent
4,217,287 discovered that if barium oxide is present iIl the
reaction mixture, the catalytic epoxida-tion of olefins wi~h
organic hydroperoxides can be successfully carried out with
good selectivity to the epoxide based on hydroperoxide con-
verted when a relatively low olefin to hydroperoY~ide mole
ratio is used. The alpha-olefinically unsaturated compound
must be added incrementally to the organic hydroperoxlde to
provide an excess of hydroperoxide that is effective.
Selective epoxidation of olefins with cumene hydro-
peroxide (CHP) can be accomplished at high CHP to olefin
_~_
ratios if barium oxide is present ~7ith the molybdenum cat-
alyst as reported by Wu and Swift in "Selective Olefin Epoxi-
dation at High Hydroperoxide to Olefin Ratios," Journal of
Catalysis, Vol. 43, 380-383 (1976).
Catalysts other than molybdenum have been tried.
Copper polyphthalocyanine which has been acti~-ated by contact
with an aromatic heterocyclic amine is an effective catal~st
for the oxidation of certain aliphatic and allcyclic compounds
(propylene, for instance) as discovered by ~rownstein, et al.
described in U. S. Patent 4,028,423.
Various methods for preparing molybdenum catalysts
useful in these olefin epoxidation methods are described in
the following patents: U. S. 3,362,972 to Kollar; U. S.
3,480,563 to Bonetti, et al.i U. S. 3,578,690 to Becker;
U. S. 3,953,362 and U. S. 4,009,122 both to Lines, et al.
More pertinent to the subject discovery are those
patents which address schemes for separating propylene oxide
from the other by-products produced. These patents demon-
strate a high concern for separating out the useful propylene
oxide from the close boiling hexene oligomers. It would be
a great progression in the art if a method could be devised
where the oligomer by-products would be produced not at all
or in such low proportions that a separate separation step
would not be necessary as in these patents.
U. S. Patent 3,464,897 addresses the separation of
propylene oxide from other hydrocarbons having boiling points
close to propylene oxide by distilling the mixture in the
presence of an open chain or cyclic paraffin containing from
8 to 12 carbon atoms. Similarly, propylene oxide can be sep-
arated from water using idenkical entrainers as disclosed in
~5--
~%'7~33~3~
6~878-28
U.S. Patent 3,607,669. Propylene oxide is purified from lts
by-products by fractiona~ion in the presen~e of a hydrocarbon
having from 8 to 20 carbon atoms according ~o U.S. Patent
3,843,~88. Additionally, U.S. Patent 3,90g,366 ~eaches that
propylene oxide may be purified with respeck to contamina~iny
para~finic and olefinic hydrocarbons by extractive distillation
in the presence of an aroma~ic hydrocarbon having from 6 to 12
carbon atoms.
SUM~ARY OF THE lNVENTION
lQ According to one aspec~ of the present invention
there is provided a method of preparlng an alkylene oxide
compound comprlsing reacting an olePinically unsatura~ed
compound in the presence of a molybdenum catalyst with an
organic hydroperoxide in a series of continuous s~irred tank
reactors in more than one stage in which the mole of ole~in to
hydroperoxide at no poln~ in any of said reactors exceeds 3.0~1
and maintaining more than 60 wt.~ of polar components in the
reaction medium in each of said reactors.
According to a further aspect of the present
invention there is provided a method of preparing an alkylene
oxide compound comprising a) reacting an olefinically
unsaturated compound wlth an organic hydroperoxide in the
presence of a molybdenum catalyst~ in a series o~ continuous
stirred tank reactors in more than one stage to give an
intermediate reaction mixture, and b) further reactlng the
intermediate reaction mixture from the series of continuous
stirred tank reactors in a second reactor to give an alkylene
oxide reaction product, and maintaining more than 60 wt.~ of
polar components in the reaction medium in each of said
reactors.
~%~
6~7g-~8
According to ano~her aspect of ~he present invention
there is provided a method for preparing an alkylene oxide
compound comprising a) reacting an olefinically unsaturated
compound wi~h and organic hydroperoxide in the presence of a
non-acid molybdenum catalys~, in a series of continuous stirred
tank reactors in more than one stage to give an intermediate
reaction mixture, and b) subsequently reac~ing the intermediate
reaction mixture from the series of continuous stirred tank
reactors in a plug flow reactor at a higher temperature than
that used in the continuous stirred tank reactor to give an
alkylene oxide reaction product containing 5 ppm or less olefin
oligomer by-product, and maintaining more than 60 wt.% of polar
components in the reaction medium in each of said reactors.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
It has been discovered that the process of producing
alkylene oxide compounds, especially propylene oxide, from
olefinically unsaturated compounds, such as propylene, together
with organic hydroperoxides over a molybdenum ca~alyst can be
improved by reducing the mole ratio of olefinic substrate to
organic peroxide. Extremely low productions of olefin
oligomers, the most troublesome by-products, can be achleved
with this technique. Surprisingly, high alkylene oxide
concentrations, selectivities and yields may also be achieved
together with high hydroperoxide conversions and high
molybdenum catalyst recoveries, all simultaneously
It has also been discovered that a series of
continuous stirred tank reactors (abbreviated CSTR) may be
employed to implement a system whereby the mole ratio of olefin
6a
.. 2~
to hydropero~ide is maintained low. Since the contents of 2
CSTR are continuously stirred, the organic hydroperoxide
only "sees" a very small proportion of olefinically unsat-
urated compound at any one time. Maintenance of a low molar
ratio of olefin to hydroperoxide is furt~er aided by option-
al s-taged addition of olefin to a staged series of reactors.
By staged addition is meant the injection of olefin into the
reaction mixture contained in the staged series of reactors
at more than one point along the staged series of reactors.
By this technique, buildup of olefin at any one point in the
staged series of reactors, rela-tive to the hydroperoxide, is
mini mized.
Normally, the ratio of olefinic substrate to or~
ganic peroxide is thought to be variable over the range of
from about 2 1 to 20:1, expressed as a mole ratio. A mole
ratio of olefin to hydroperoxide of less than 2:1 has been
thought to be unsuitable. In this invention, the mole ratio
of olefin to hydroperoxide should never exceed 3.0:1. The
broad range, expressed in terms of the overall contents of
the reactor, for ~he mole ratio of olefinic substrate to
organic peroxide of this invention is from 0.9:1 to 3.0:1,
and preferably from 1.8:1 to 0.9:1. Most preferably, the
mole ratio of olefin to hydroperoxide is 1.05:~ to 1.35:1.
of course, the local, instantaneous ratios at any place in
the staged series of CST~s will be appreciably lower than
these ratios.
One of the particularly preferred embodiments of
this invention involves the use of a staged series of CSTRs
in series with a tubular or plug flow reactor (abbreviated
PFR). In one preferred version, the CSTR ls placed in fron-t
~7B~
of the PFR in the scheme of reac-tant flow; i.e., the reac-
tants encounter the CSTR first and are subsequently transfer-
red to the PFR. This reactor con~iguration provides good
temperature control of a highly exothermic reactlon during
the initial stage of the epoxidation reaction ar.d good hydro-
peroxide conversions in the latter stage. In addition to
the above mentioned benefits, the formation of olefin
dimers, which are particularly undesirable by-products
because they are difficult to remove, is minimized.
Another of the preferred embodiments include oper-
ating conditions which allow for a relatively cool reaction
to take place in a CSTR followed by a PFR operating at rel-
atively severe temperatures, and preferably, a feed ratio of
olefin to hydroperoxide that is low, with high catalyst con-
centrations. The CSTR may be operated at a temperature in
the range of about 70 to 115C, 90 to 115C being preferred,
with 100-110C as the most preferred reaction temperature
range. The PFR should be operated at a higher temperature,
from over 115C to 150C, with 120`-140C as the most pre-
ferred range. The effluent from the CSTR may be termed anintermediate reaction mixture since there is more reacting
still to occur. The residence time of the reactants in each
reactor is left to the operator although it is preferred
that the reactants stay in each reactor approximately the
same length of timeO T~pically, the residence times may be
0.5 to 4.0 hours, preferably l.0 to 2.0 hours per reactor.
Surprisingly, the same temperatures, 120-i40C,
which normally produce large amounts of alkylene dimer, par-
ticularly propylene dimer, yield only small amounts of alkyl-
ene dimer when employed with a PFR, in series after a CSTR.
3~5
However, it is also con~emplat~d that somewhatdifferent reactor configurations may also be particularly
effective. For example, the PFR might be placed before the
CSTR in sequence. Or a series of continuous s~irxed tank
reactors may prove advantageous. Staged introduction of -the
olefinically unsaturated reactant may be inco~porated in any
of these con~igurations.
Addi-tionally, it is contemplated that the temper-
ature se~uence may be changed from that recom~"ended above
with positive results. For example, the higher reaction
temperature may occur first followed by a relatively cooler
reactor region. Or a series of hotter regions inierspersed
~ith cooler ones may be effective.
Continuous stirred tank and tubular plug flow re-
actors are well known in the art. It is expected that theinventive method would work well with any of -the two tyoes
available.
The other advantages to the preferred embodiments
of this invention include much higher oxidè concentrations
(29-32~), higher oxide selectivities (96-99~%) and hydro-
peroxide conversions ~96-99%) and oxide yields (94-98%) than
are described as possible in the current literature or
patent art. Further, olefin oligomer by-product contents in
the crude alkylene oxide reaction product stream as low as 5
ppm and lower can be achieved. Molybdenum catalyst contents
in the product stream, also called recoverable molybdenum,
may be 85% or higher relative to -the charged catalyst
proportion.
Further, a molybdenum alcohol catalyst which COIl-
tains no free or excess carboxylic acid is employed whereas
3~
the catalyst cited in commercial process descrip-tions is
derived from an acid reactant and contains excess acid.
For example, the preferred catalysts are molybdenum 2-ethyl-
hexanol, molybdenum ethylene glycol or molybderum propylene
glycol complexes whereas the commercial catalyst is a mo-
lybdenum naphthenate or molybdenum octoate derived from
naphthenic acid or 2-e.thyl hexanoic acid. These catalysts
inherently contain a substantial amoun-t of lre-e (excess)
acid (up to 50-70 wt.% acid).
Details of Reac~ants and Catalysts
The method of this invention could be used to ep-
oxidize any olefinically unsaturated compound such as sub-
stituted and unsubstituted aliphatic and alicyclic olefins
which may be hydrocarbons, esters, alcohols, ketones, ethers
and the like. It is expected that the process would be par-
ticularly useful in epoxidizing compounds having 2 to 30
carbon atoms and at least one double bond situated in the
alpha position of a chain or internally. Representative
compounds include ethylene, propylene, normal butylene,
isobutylene, pentenes, methyl pentenes, hexenes, octenes,
dodecenes, cyclohexene, substituted cyclohexenes, butadiene,
styrene, substituted styrenes, vinyl toluene, vinyl cyclo-
hexene, phenyl cyclohexenes and the like. Olefins having
substituents containing halogens, oxygen, sulfur and the
like may be used. In general, all olefinic materials epoxi-
dized by previous methods could be used in connection with
this process including olefinically unsaturated polymers.
The invention will probably find its greatest
utility in the epoxidation of primary or alpha olefins. How-
ever, it is propylene that may be epoxidized particularly
-lQ-
7~
advantageously by the inventive technique, and it is this
olefin that is par-ticularly preferred.
Any of the hydroperoxldes used in the previously
described prior epoxidation methods may be used effectively
in this invention. Suitable organic hydroperoxide reactants
may have the formula ROOH where R is an organic radical.
Preferably, R is a substituted or unsubstituted alkyl, cyclo~
alkyl, aralkyl, aralkenyl, hydroxy aralkyl, cycloalkenyl,
hydroxy cycloalkyl and the like having about 3 to 20 carbon
atoms. R may also be a heterocyclic radical.
However, it has been discovered that t-butyl hydro-
peroxide ~TB~P~ gives better resulks to propylere oxide than
do other hydroperoxides, such as cumene hydroperoxides (CHP)
even though both show improvements when a two-stage reaction
scheme is used. Thus, the most preferred hydroperoxide for
the method of this invention is TB~P in a solution of t-butyl
alcohol (TBA). The weight ratio of the mixture should be
40-80% TBHP with the balance being TBA and other minor species.
A TBHP ~oncentration of 68-~0% is particularly preferred~
Catalysts suitable for the epoxidation method of
this invention include any molybdenum complexes with no free
or excess carboxylic acid present. If an acidic catalyst is
used, as in complexes of molybdenum with 2-ethyl hexanoic
acid, the excess acid should be removed, such as by distil-
lation with a higher boiling paraffin such as a Cl~ (hexa-
decane, for example). A "non-acidic" catalys-t i5 defined
as one having an acid number no higher than 5Q due to free
or excess carboxylic acid.
It is especially preferred that the catalyst be
molybdenum complexes of 2-ethyl-1-hexanol or molybdenum
~%~33~5
complexes of glycols. A detailed account of the pre-Eerred
preparation of these complexes may be found in co-pending
Canadian patent applications Serial Nos. 483,719, filed June
12, 1985 and ~83,634, filed June 11, 1985. E~or the purposes of
the instant applica~ion, the complexes are generally made by
reacting a molybdenum compound, such as M003 with
~-ethyl-l-hexanol in the presence of NH40H and heat over a
period of time. The mole ratios of 2-ethyl-1-hexanol to gram
atoms of molybdenum should range from 10:1 to 55:1. The ratio
of moles of ~H40H to gram atoms of molybdenum should range from
1:1 to 5:1. These ratios of reactants and reaction
temperatures of 150 to 185C accompanied by removal of water
will afford a surprisingly high molybdenum content (2 to 7
wt.~) catalyst complex which is stable upon standing.
Alternatively, the molybdenum 2-ethyl-1-hexanol catalyst can
also be made by reaction of ammonia heptamolybdate (AHM) with
2-ethyl-1-hexanol and water in the right proportions. The mole
ratios of 2-ethyl-1-hexanol to gram atoms molybdenum in AHM
should range from 7-13:1. The amount of water used initially
should be in the range of 1-3:1 moles water/molybdenum~ The
three reactant mixture should be heated at 125-185C for 5-8
hours, affording a molybdenum catalyst with 5-10% molybdenum
content. The molybdenum glycol complexes are made by digesting
ammonium heptamolybdate or ammonium dimolybdate with ethylene
or propylene glycol (mole ratio of glycol to gram atoms of
molybdenum 8:1 to 16:1) for approximately one to two hours at
about 90-130C and then pulling a vacuum and stripping off
water and glycol to leave a clear catalyst bottoms product
amounting to 75 to 95~ of the total charge to the catalyst
-12-
~7~
preparation. The molybdenum contents of these glycol catalysts
range from about 10 to 16~. Canadian patent application Serial
No. 483,634, filed ~une 11, 1985 provides more details on the
synthesis of these catalysts.
Other Reaction Conditions
The reaction should be carried out in liquid phase
under autogenous pressure which should not exceed about 800
psig. The pressure is preferably kept in the range of 200 to
800 psig.
The catalyst concentrations in the method of this
invention should be in the range of 100 to 1000 ppm (0.01 to
0.10 wt.~) based on the total reactant charge. A preferred
range is 200 to 600 ppm. Generally, about 250 to 500 ppm is
the most preferred level. These catalyst levels are higher
than those presently used in prior methods, which tend to run
from 50 to 200 ppm.
The epoxidation reaction of this invention can be
carried out in the presence of a solvent and it is preferred
that one be used. Ideally, the solvent should be inert in the
reaction and have the same carbon s~eleton as the hydroperoxide
used to minimize solvent separation problems. For example, if
TBHP is used as the hydroperoxide, TBA is preferably the
solvent. Of course, the TBA in the TB~P solution may be
enough to serve as the solvent in the reaction.
It is preferred that the reaction mixture contain
very little water, between zero and 0.5 wt.~.
Preferably, the reaction should be carried out to
achieve as high a hydroperoxide conversion as possible,
typically 96 to 99~, consistent with reasonable oxide se-
-13-
3~)5i
lectivities, ~ypically also 96 to 99% for the method of this
invention. For both of these values to be simultaneously so
high is very unusual. While prior me~hods concerning propyl-
ene epoxidation have accomplished hydroperoxide conversions
o 98 to 99%, the propylene oxide selectivity b2sis TsHp con-
sumed runs only about 90 to 91%. As a result, the yield or
utilization to the oxide for the inventive process runs about
94 to 98% as compare`d with 8~ to 90% for prior ?rocesses~
Although such differences seem small, they may provide the
distinction between a profitable process and a 'otally un-
acceptable, wasteful one. When the plant size is several
hundred million pounds of product, the 4-8% differences in
yields are very significant.
For a batch mode, the reaction procedure generally
lS begins by charging the olefin to the CSTR, if the CSTR is
first in the configuration. Next, the hydroperoxide, sol-
vent and catalyst may be added and the contents heated to
the desired reaction temperature. Alternatively, the olefin
reactant may be heated to at or near the preferred reaction
temperature. Further heat may be provided by the exotherm
of the reaction. The reaction is then allowed to proceed
for the desired amount of time before transfer to the PFR
for the desired time. The mixture finally is cooled do~m
and the oxide recovered. In a continuous mode, the re-
actants are run through the chosen reaction configurationcontinuously, with the residence times for each reactor ad-
justed as desired.
Generally, for the method of this invention, the
oxide concentration runs from abou-t 29-32% which is quite a
bit higher than the oxide concentrations possible in prior
~7~
methods, usually in the neighborhood of 12-16%.
A significant advantage of this invention is that
there is very little propylene dirner (hexenes) content in
the reactor effluent (when propylene is reactea). It typi-
cally runs less than 5 ppm in the total reactor productstream. This level of dimer would very likely eliminate the
high capital cost for the typical three tower e~tractive dis-
tillation module typically used to remove propylene dimer
from commercial product streams, as well as save money on
operating costs for this unit. Further, no hydrocarbon en-
trainer is needed as suggested by the prior ar_. Instead,
the low dimer level of this invention (e~uiv21ent to less
than 20 ppm on a "pure" oxide basis; that is, after the UIl-
reacted products are removed) allows an operator to leave
the dimer in the oxide where it is virtually trouble ~ree at
these low levels.
The process of -this invention may be preferably
used in a continuous mode.
At olefin/hydroperoxide ratios of 1.2:1 to 1.8:1,
the preferred temperature in the one or more CSTRs in series
is 70-115 for 1.0-2.0 hours and the preferred temperature
is 120-130C for 1.0-2.0 hours at catalyst Ievels of
200-600 ppm, preferably 300 to 500 ppm. If the olefin/hydro-
peroxide level is relatively low, about 1:1 to 1.2:1, the
preferred -temperature in the one or more CSTRs in series is
90-120C for 1.0-2.0 hours and the preferred PFR temperature
is 125-140C for 1.0-2.0 hours at 300 to 500 ppm catalyst
levels without any appreciab1e effects on selectivity, etc.
The method of this invention is illustrated, but
not limited, by the following examples.
-15-
~2~
FXAMPLE
To a 300 ml Hastelloy C agitated au~oclave followed
by a 200 ml 316 stainless steel tubular reac~or was fed
56.9 g/hr of propylene and 126.7 g/hr of a T~EP/TBA/ca-talyst
mixture (which analyzed as 71.0 ~t.% TBHP, 28.8 wt.% TBA,
0.2 wt.% H2O, 695 ppm molybdenum complex of 2-e hyl-l-hexanol
catalyst). The CSTR was operated at llOC, the EFR at 120C.
The product from the said reactor system analyzed as follows:
Propylene, wt.% 9.83
Propylene oxide, wt.% 29.81
TBA, wt.% 56.18
TBHP, wt.% 2.64
Recoverable molybdenum catalyst, ppm 445
Propylene dimer, ppm ~pure PO basis) 10
Propylene dimer, ppm (crude product basis) <5
NOTE: Propylene/TBHP molar ratio was 1.355/1.
-16-
3~5
EXAM LE 2
To the same reactor systern as above was fed
50.0 g/hr propylene and 128.2 g/hr of TB~P/~A/catalyst
(72.4 wt.% TBHP, 27.4 wt.% TBA, 0.2 wt.% H~, 51~ ppm
molybdenum complex of 2-ethyl-1-hexanol catalysi). The
CSTR was operated at 110C, the PFR at 130C. The product
analyzed as follows:
Propylene, wt.% 5.22
Propylene oxide, wt.~ 31.40
TBA, wt.% 58.61
TBHP, wt.% 2.63
Recoverable molybdenum catalyst, ppm 445
Propylene dimer, ppm (pure PO basis) 10
Propylene dimer, ppm (crude product b~sis) ~S
NOTE: Propylene/TBHP molar ratio was 1.154/1.
In contrast to the above results, Ex2mple 3, con-
ducted at 120C for two hour reaction time in a s-tirred auto-
clave in the batch mode, resulted in a substantial make of
the undesired by-product, propylene dimer (94 ppm pure PO
basis), which co-distills with propylene oxide upon sep-
aration/pur~fication.
-17-
3~i
EXAMPLE 3
To a 300 ml 316 stainless steel autoclave were
charged 48.lg propylene (1.145 moles) followed by a solution
containing 129.21g of a t-~utyl hydroperoxide solution (con~
taining 61.50% TBHP, 38.30% _~butyl alcohol, ~d 0.2% H2O)
and 1.29g of molybdenum 2-ethyl hexanol (6.50% molybdenum).
The clave was stirred and heated to 120C for 2.0 hours,
cooled and sampled under pressure to GLC. ~;~e recovered
total product was 177.2g. Total liquid product after
flashing propylene was 150.4g. The liquid product con-
tained 1.03% TBHP unreacted (T~P conversion = 98.0%) and
594 ppm molybdenum (essentially quantitative molybdenum re~
covery). GLC indication was that 6.86% propylene was un-
reacted with PO make at 26.56% (selectivity moles PO/moles
TBHP reacted = 93.7%). Yield moles PO/moles TBHP fed =
91.9%. Propylene dimer (basis pure PO) was analyzed for
and found to be 94 ppm.
20`
-18-
~2~3~3~
~ O ~
c~ ~ u~ o u~ U) o o o o u~
~ ,. ~ ,, ~ ~ ," ~ ~ ~ ~ ~ ~, C~
~ ~ ., ~ ;~ ~ ,,
P~ )~ K ~1 ''
. ,~
o ~ ~o ~ ~ u~ ~ ~ ~ o ~ ~4 2
o
~D O C~ CO ~ ~ ~ 1-- CO ~ ~ ~ ~
o o K ~ oo C~ o r~ o
. `J r- Lr) ,~ ~ u~ C~l ~ ~ ~ ~ CO
~ o r~ ~ co r- CO co ~ ~ 0~ 0~ co
c~ E~
.. 3
~,1
~3 ¢ ~ ~ ~ ~ ~ 1_ ~J O 0~ O O
~3 ~ o a~ ~ ~ CS~ Cr~ ~ o g o G~ O g
¢ -! ~1 ~ X O ~rl ~1 ~ ~
¢ ~ ~ ~1 0 U~ ~
P~ h
¢ ~ c~ ~ O O ~ O C~
~ ~ ~ ~ 1 ~O 11~ G` ~ I` '~ \D 1- L~7 ~ 00 CO C~
H O O CJ ~rl ~ 10 ~ ~0 ~ ~ In 1~ I ~1 ~ ~ CO ~J
~ ~1 h ~ X ~rl 00 ~ a~ cn a~ a~ ~ Cil O Cl~ O a~
,~
~0; 41
~; ~ ~ D JJ
. ~ V~ ¢ C ~ h Ul ~ `:t ~ O `J O ~0 CO ~ U') 0~ ~ ~`J Cl
:C :~ ~ E E ~ ~ r~ ~ ~1 ~ Ln ~ ~ ~ ~ ~1 u~ h
E- p~ ,~
~ ~ ~ O ~ ~
:~ ,~~o o ~o u~ ~o u~ o o~ o
~ ~ ~ ~ ~~ ~o o In ~ ~ o cn ~o ~O ~ ~ Q~
P O ~ ~rl ~ ~ ~ ~J ~ ~ ~ ~ ~ u~ ~ ~ ~J
; ' ~ O g,,, ,~ O '`
¢ E E ~ C
P.l ", ~ u E
;OZ :~ o o ~ o o CO ~ ~ o ~ ~ o ~ C) ~
h -- '^ O "~ E3 G 'Cl C.l
E ~ 4.
C u In ~ U Ç4 o~
O OOOOOOOOOOO W~O~rl
ri o ~ u ~ ~1 ?
. ~ ~ ,ul E O ~ ?~
C o ~ ,a
~ d ~ u
C . ~ X ~ C u
o o o n ~ In o o o ~ u~ u~ ~ ~ o ~ s~
E o ~ ~`I '`J ~ '~ '' ~ '' ~ ~, ~ ~4
~ ~ O ~ Cl GJ
a ~ ¢ ~ 3 C
u 1~ O --~ ~ O ~ c~ O -~ ~ O ~1
v o o o o o o o o o o o o 3 3 ~ v C
~ oooooooooooo v~
,1 ~ h O
IJ X v v u ~ d ~ cJ
X ~ 0 ~ c~J ~ ;l u~
~71~
a~ ~
al v,
O ~r- oco ~ ~o r_ o~
U ~~ )~ ~ ~O ~ ~ _, ~,
O ~ ~00 CO GO 0~ O G~
c~ K
o ~-
n ~ ~I r~ ~ ~ ~ ,~ G~ `J r~
~ o I` ~ ~ r~ o
U~
O O ~ ~ U ~ ~ 0~ O ~ ~0
P~ ~ o . . . . . .
cq O CJ .,~ oo~ o ~ 1
? ~ ~ X
0 3 ~ ~ O u~
~ ~1
~ ~n
:~ ~ Q~ ?, a~ ~a ~ ~ o o In ~ o~ o
E~ ~ ? ~ ~ ,~ u~
.. 3 ,~ o
r~ ~i C X rl 0~ ~ ~ ~i 00 ~ r~, ~ ~1
~; ~ C ~ ~i O ~ ~ a~
X
E~
P~ o ~ o ~
~ ~ ~ 4X
3 U14 ~ h ~q ~
O n~E~ ~ ~i ~ o o o ~ `J o o o o
~r~
~ ~1 o
h , ~i
~1 ~ ~1 ~pe o o c~ o
r~i 1~ ~J ~ ~ co
X ~ . O ~ri ~ r-l r-~ ~ O O O
, a~ ~ ~ P~ ~1 o ~ ~ ~) ~ c~ ~ ~ ~ ~i
~ C~ C
U~ ~,i
po; ~ l
:s ,
E~ ~ ~ p o ~ ~ ~ ~ o 03
cJ X ~ ~ o o o~ o
,i ~ ~ ~ o
?~ ~ X ~ ,~ , i ~ o ,~
I 3 o
3 ,C~j '~
F~ o ~ ~J ~ o o o o ~ n ~ n n n n
Z ~:1 ~ ~ i ~i ~ r i r1 ~ r-i r1 ~ r~
u~i E-i
~i ~
a o
rl . ~
~3 O . P; U ~0 O O O O O O O O O O O
~ ,_ ~ ~ , i ~I r i ~ r~i r~
r~
.
~ r-i O ~
O ~ ~ O 0~ D ~ ~ O co
. ~,i u ~ n .n n n n n n n ~ ~ ~
ooooooooooo
3 P:l K ~ o o o o o o o o o o o
,~
~ o
Z r1 r1 r~ r i
.
t~
When reseaLch into this area was first be~1n, it
was discovered ~hat results irnproved dramatically when dry
TBHP (less than 0.4 wt.% H20) was used instead of the com-
mercially available TsHp~ ~here~ore, it is preferred that
the mixture of reactants contains very little water, 0.5% or
less.
The high olefin/hydroperoxide mole ra'io was chosen
initially because of -the repeated mention in pat~nts and the
literature that selectivities are lower when propylene/TsHp
mole ratios are too low. ~owever, when low propylene/TBHP
mole ratios were used, it was surprisingly disco~ered that
propylene oxide selectivities were actually his:~er rather
than lower, as expected. In addition, it was also surpris-
ingly discovered that the molybdenum recoveries increased
upon reduction of the propylene/TBHP mole ratio.
Using this new technique, different catalysts were
screened. Generally, the various dilute ethylene and propy-
lene glycol molybdenum complexes and those similarly dilute
catalysts derived from ethylene or propylene carbonate were
less active or selective than the several low acidity molyb-
den~um octoate catalysts made by distilling out the excess
acid from standard molybdenum octoate preparations. This
involved heating 2-ethyl hexanoic acid with molybdenum tri-
oxide or ammonium heptamolybdate. However, the molybdenum
2-ethyl-1-hexanol catalysts and concentrated (10-16% molyb-
denum content) molybdenum ethylene or propylene glycol ca-t-
alysts are the most preferred. The former preferred cat~
alysts are made from MoO3, 2-ethyl-1-hexanol and ammonium
hydroxide or from ammonium heptamolybdate and 2-ethyl-
1-hexanol. These latter preferred catalysts are made ~rom
-21-
7B3~353
ammonium heptamolybdate or ammonium dimolybdate and propyl-
ene or ethylene glycol
Propylène dimer, as noted, is an objectionable by-
product because it co-distills with propylene oxide and is
best separated from propylene oxide by a costly extractive
distilla-tion. The cost of the extractive towers as well as
the utilities cost to operate such a purification unit is
very high. The reactor configuration, low reaction temper~
atures and low propylene/TBHP mole ratios are the major
~actors in the low propylene dimer proportions. The ex-
` amples of Table I where the conventional propylene oxide
process is used reveals that the propylene dime- levels seen
in Table II are surprisingly low. The low propylene dimer
content of the resultant propylene oxide product may be left
in the propylene oxide product without adverse effect or
costly distillation.
The examples of Table II demonstrate the two-step
reactor scheme which permits extremely low propylene to TBHP
ratios (0.88:1) and yet produces little or no propylene dimer.
These examples further point out the unusual characteristics
of the instant invention.
The inventive process provides much higher concen-
trations of propylene oxide in the reac-tor effluent (29-32%)
than current commercial processes (13-15%) and u-tilizes much
less propylene due to the lower propylene/TBHP mole ratio.
Thus the reactors and other equipment (distillation towers)
are much reduced in size as well. In this process, 4 to 16%
of the propylene is unreac-ted in the reactor effluent as com-
pared with a~out 35 to 55% unreacted propylene in prior art
processes. Cur yields (moles of propylene oxide formed per
-22-
~æ~
mole of TB~P consumed) do not drop as we lower the propyl-
ene/TBHP mole ratio as expected or projected from the liter-
ature. Surprisingly, increased selectivities a~d conversions
were observed. Perhaps this is because the media is more
polar rendering the molybdenum catalyst more active, soluble
and stable than in a largely propylene media. The latter
situation would prevail if the epoxidation is conducted at
higher propylene to TBHP mole ratios. However, the inven-
tion should not be limited by any such theory.
It is surprising that selectivities ~o propylene
oxide are at least 96%, concentrations of prop-lene oxide in
the crude product stream can be at least 29%, ~ields to pro-
pylene oxide are at least 94%, hydroperoxide conversions are
at least 96% and propylene dimer contents are 5 ppm or less,
lS all simultaneously, using the method of this invention.
Examples 27 through 36 of Table III show that
propylene oxide may be formed with a different organic hydro-
peroxide, cumene hydroperoxide (C~). Although the previous
examples using TBHP give better results than the experiments
using CHP, the advantage of a two-step reaction scheme is
demonstrated. Compare Examples 27 and 37 where the CHP con-
version is only 66% for a one-stage scheme as compared with
97.7% for a two-step scheme.
Although cumene hydroperoxide has been shown to be
undesirable as the hydroperoxide in the inventive process,
hydroperoxides haviny a structure closer to that of t-bu-tyl
hydroperoxide, such as t-amyl hydroperoxide, is expected to
be useful in this method.
Many modifications could be made by one skilled in
-23-
~7~3~
the art in the invention without changing its spirit or scope
which are defined only by the appended claims. For example,
;` within the parameters of the claims, a particular combination
of reactants, catalysts, mode of addition or sequence procedure
may prove to be particularly advantageous.
-2~-
r-l ~2 ~ J ~ O CO ~0 ~J
aJ ~ ~~ O ~
O ~ r~ I~ 1~
o
m
~h
O c~ ~~ ~ O c~ ~ ~ O
o c, o u~ ou~ r~ ~ u~ ~ ~ r~ 1-- 0
X 1~; ~ 1~ r~ 1~ ~D ~ I` ~ r~ ~o ~ rl
O
~ U:l
~ I I ~ ~ O r~ O 00 r1 0 ~
~ ~ O ~ I~ CO C~ ~ O
C ~ n cn co i~ CCI o
rn .
~ ~ o
X C~ ~ ~ ~ ~
o ~ ~ ~ . o
O~ o~ O
O Q~ ~ ~ aJ ~ a~
~n ~ ~ ~ ¢
.~1 ~ ~ D ~ ~ ~ ~ ~ ~ ~ 0~ ~
~ _ ~1 `~ o~ Io 1~ E
C _ ~ ~ 1 ~ ~ ~/ rt ~ ~ ~ O
X ~C~ o~ CO O C~ 00 0~ It') ~
~3 ~ 3 ~ r; ~ rl S l
O O~ O ~ o
P~ ~ X ~ ~
O ~I Ulr~
O ~ ~ ul ~ ~ ~ o ~ ~ ~
r~ O
O O r-l
uj ~ o
o ~ ~ ~` ~ ~ c~ o r~r. ~ r~ ~ 3 ~ o~
G o ~ ~ ~ o 1--In u) ~D Ih u~ ul ~ ~ q ~ ~
C~ C~ ~ ~ h
~ ~ ~3 40~ ' o o
~ - ~ 3 3 d
s~ 1 u d d
~ 1 ~ ~I ~ l r1 r~l r-l r~ l r~ rl rl
U ~r~ ~ ~ O ~ C~iC~ r~ ~ ~ oo ~ ~ u u
O ~ S l O O O O ~ r~ 1 C~ 0 3
~J r 00 00
r~l r-l O U~ ~
E r~ v~ ~ o r~ Dr~ r~ ~~ o
