The Cativanr Process for the Manufacture
`of Acetic Acid
`IRIDIUM CATALYST IMPROVES PRODUCTIVITY IN AN ESTABLISHED INDUSTRIAL PROCESS
`
`By Jane H. Jones
`BP Chemicals Ltd., Hull Research & Technology Centre, Salt End, Hull HU12 BDS, U.K.
`
`Acetic acid is an important industrial commodity chemical, '<Vith a world demand of about 6
`million tonnes per year and many industrial uses. The preferred industrial method for its
`manufacture is by the carbonylation of methanol and this accounts for approximately 60 per
`cent of the total world acetic acid manufacturing capacity. The carbonylation of methanol,
`catalysed by rhodium, was invented by Monsanto in the 1960s and for 25 years was the leading
`technology. In 1996 a new, more efficient, process for the carbonyl arion of methanol was
`announced by BP Chemicals, this time using an iridium catalyst. This article describes the
`new process and looks at the ways in which it improves upon the prior technology.
`
`In 1996 a new process for the carbonylation of
`methanol to acetic acid was announced by BP
`Chemicals, based on a promoted iridium catalyst
`package, named Cativa™. The new process offers
`both significant improvements over the conven(cid:173)
`tional rhodium-based Monsanto technology and
`significant savings on the capital required to build
`new plants or to expand existing methanol car(cid:173)
`bonylation units. Small-scale batch testing of the
`new Cativa ™ process began in 1990, and in
`November 1995 the process was first used com(cid:173)
`mercially, in Texas City, U.S.A., see Table I.
`The new technology was able to increase plant
`throughput significantly by removing previous
`process restrictions (debottlenecking), for instance
`at Hull, see Figure 1. The final throughput
`achieved has so far been determined by local avail-
`
`ability of carbon monoxide, CO, feedstock rather
`than any limitation imposed by the Cativa ™ sys(cid:173)
`tem. In 2000 the first plant to use this new
`technology will be brought on-stream in Malaysia.
`The rapid deployment of this new iridium-based
`technology is due to these successes and its many
`advantages over rhodium-based technology. The
`background to this industrial method of producing
`acetic acid is explained below.
`
`The Rhodium-Based
`Monsanto Process
`The production of acetic acid by the Monsanto
`process utilises a rhodium catalyst and operates at
`a pressure of 30 to 60 atmospheres and at temper(cid:173)
`atures of 150 to 200°C. The process gives
`selectivity of over 99 per cent for the major feed-
`
`Table I
`Plants Producing Acetic Acid Using the New Cativa™ Promoted Iridium Catalyst Package
`
`Plant
`
`Location
`
`Sterling Chemicals
`Samsung-BP
`BP Chemicals
`Sterling Chemicals
`BP Petronas
`
`Texas City, U.S.A.
`Ulsan, South Korea
`Hull, U.K.
`Texas City, U.S.A.
`Kertih, Malaysia
`
`Year
`
`1995
`1997
`1998
`1999
`2000
`
`Debottlenecking or
`increased throughput achieved,%
`
`20
`75
`25
`25
`Output 500,000 tonnes per annum
`
`Platinum MetaLs Rev., 2000, 44, (3), 94-105
`
`94
`
`CE Ex. 2033
`Daicel v. Celanese
`IPR2015-00171
`
`001
`
`

`
`stock, methanol (1). This reaction has
`been investigated in great detail by
`Forster
`and
`his
`co-workers
`at
`Monsanto and the accepted mecha(cid:173)
`nism is shown in Scheme I (2). The
`cycle is a classic example of a homoge(cid:173)
`neous catalytic process and is made up
`of six discrete but interlinked reactions.
`During the methanol carbonylation,
`methyl iodide is generated by the reac(cid:173)
`tion of added methanol with hydrogen
`iodide. Infrared spectroscopic studies
`have shown that the major rhodium
`catalyst species present is [Rh(CO),I,]-,
`A. The methyl iodide adds oxidatively
`to this rhodium species to give a rhodi(cid:173)
`um-methyl complex, B. The key to the
`process is that this rhodium-methyl
`complex undergoes a rapid change in
`which the methyl is shifted to a neigh(cid:173)
`bouring carbonyl group, C. After the
`subsequent addition of CO, the rhodi(cid:173)
`um complex becomes locked into this
`acyl fonn, D. Reductive elimination of
`the acyl species and attack by water can
`then occur to liberate the original
`rhodium dicarbonyl diiodide complex
`
`MeCOI
`
`Fig. I The Cativa"' acetic acid plant which is now operating at Hull.
`The plant uses a promoted iridium catalyst package for the
`carbonylation of methanol. The new combined light ends and drying
`column can be seen
`
`............
`
`Rh
`
`/co l- A
`
`1
`'-...co
`
`Me
`
`~ 1 /
`
`l-
`
`Me
`
`I
`co
`
`1
`
`/
`
`oc
`'-.......
`Rha 1// '-....co
`
`D
`
`•-.......___ //co
`Rha
`1// '-....co
`
`MeC01H
`
`HI
`
`MeOH
`
`l-
`
`B
`
`Scheme I
`The reaction cycle for the Monsamo
`rhodium-catalysed carbonylation of
`methanol to acetic acid
`
`co
`
`PlalinNm Metal.s &v., 2000, 44, (3)
`
`95
`
`002
`
`

`
`and to form acetic acid and hydrogen iodide, HI.
`When the water content is high(> 8 wt.%), the
`rate determining step in the process is the oxida(cid:173)
`tive addition of methyl iodide to the rhodium
`centre. The reaction rate is then essentially first
`order in both catalyst and methyl iodide concen(cid:173)
`trations, and under commercial reaction conditions
`it is largely independent of any other parameters:
`
`Rate oc [catalyst] X [CH,I]
`
`(i)
`
`However, if the water content is less than 8
`wt. %, the rate determining step becomes the
`reductive elimination of the acyl species, from cat(cid:173)
`alyst species D.
`Although rhodium-catalysed carbonylation of
`methanol is highly selective and efficient, it suffers
`from some disadvantageous side reactions. For
`example, rhodium will also catalyse the water gas
`shift reaction. This reaction occurs via the compet(cid:173)
`ing oxidative addition of HI to [Rh(CO)J,r and
`generates low levels of carbon dioxide, CO,, and
`hydrogen, H 2, from CO and water feed.
`
`[Rh(CO),Izr + 2HI ~ [Rh(CO),:q- + H,
`[Rh(CO),I.r + H,O + CO ~
`[Rh(CO),I,r + CO, + 2 HI
`
`(li)
`
`(Iii)
`
`system to give ethanol which subsequently yields
`propionic acid.
`One possible precursor for the generation of
`acetaldehyde is the rhodium-acyl species, D,
`shown in Scheme I. Reaction of this species with
`hydrogen iodide would yield acetaldehyde and
`[Rhl.cor, the latter being well known in this sys(cid:173)
`tem and proposed to be the principal cause of
`catalyst loss by precipitation of inactive rhodium
`triiodide. The precipitation is observed in CO(cid:173)
`deficient areas of the plant.
`
`[Rhi,(CO)(COCH,)] + HI ~
`[RhL(CO)r + CH,CHO
`[RhL(CO)] ~ Rhl, + 1- + CO
`
`(v)
`
`(vi)
`
`In addition to propionic acid, very small amounts
`of acetaldehyde condensation products, their
`derivatives and
`iodide derivatives are also
`observed. However, under the commercial operat(cid:173)
`ing conditions of the original Monsanto process,
`these trace compounds do not present a problem
`to either product yield or product purity. The
`major units comprising a commercial-scale
`Monsanto methanol carbonylation plant are
`shown in Figure 2.
`
`Overall: CO + H,O ~ CO, + H,
`
`This side reaction represents a loss of selectivi(cid:173)
`ty with respect to the CO raw material. Also, the
`gaseous byproducts dilute the CO present in the
`reactor, lowering its partial pressure -which would
`eventually starve the system of CO. Significant vol(cid:173)
`umes of gas are thus vented - with further loss of
`yield as the reaction is dependent upon a minimum
`CO partial pressure. However, the yield on CO is
`good (> 85 per cent), but there is room for
`improvement (3, 4).
`Propionic acid is the major liquid byproduct
`from this process and may be produced by the car(cid:173)
`bonylation of ethanol, present as an impurity in the
`methanol feed. However, much more propionic
`acid is observed than is accounted for by this
`route. As this rhodium catalysed system can gener(cid:173)
`ate acetaldehyde,
`it is proposed
`that
`this
`acetaldehyde, or its rhodium-bound precursor,
`undergoes reduction by hydrogen present in the
`
`(iv) The Monsanto Industrial Configuration
`The carbonylation reaction is carried out in a
`stirred tank reactor on a continuous basis. Liquid is
`removed from the reactor through a pressure
`reduction valve. This then enters an adiabatic flash
`tank, where the light components of methyl
`acetate, methyl iodide, some water and the product
`acetic acid are removed as a vapour from the top
`of the vessel. These are fed forward to the distilla-
`cion train for further purification. The remaining
`liquid in the flash tank, which contains the dis(cid:173)
`solved catalyst, is recycled to the reactor. A major
`limitation of the standard rhodium-catalysed
`methanol carbonylation technology is the instabili(cid:173)
`ty of the catalyst in the CO-deficient areas of the
`plant, especially in the flash tank. Here, loss of CO
`from the rhodium complexes formed can lead to
`the formation of inactive species, such as
`[Rh(CO),I.r. and eventually loss of rhodium as the
`insoluble Rhi,, see Equations (v) and (vi).
`Conditions in the reactor have to be maintained
`
`Platin11111 Metals &v., 2000, 44, (3)
`
`96
`
`003
`
`

`
`Electric
`motor
`providing
`agitation
`
`Acetic
`acid
`
`Propionic
`acid
`(byproduct)
`
`Reactor
`
`Flash tank
`(Catalyst rich
`stream recycled)
`
`Drying
`column
`
`•Heavies•
`·Lights•
`removal
`removal
`column
`column
`'------Distillation train___j
`
`Fig. 2 The major units comprising a commercial-scale Monsanto methanol operating plam. which uses a rhodium·
`based catalyst. The technology uses three distillation columns to sequemially remove low boilers (methyl iodide and
`methyl acetate), water, and high boilers (propionic acid) and deliver high purity acetic acid product
`
`within certain limits to prevent precipitation of the
`catalyst. This imposes limits on the water, methyl
`acetate, methyl iodide and rhodium concentra(cid:173)
`tions. A minimum CO partial pressure is also
`required. To prevent catalyst precipitation and
`achieve high reaction rates, high water concentra(cid:173)
`tions in excess of 10 wt.% are desirable. These
`restrictions place a limit on plant productivity and
`increase operating costs since the distillation sec(cid:173)
`tion of the plant has to remove all the water from
`the acetic acid product for recycling to the reactor.
`(The water is recycled to maintain the correct
`standing concentration.)
`Significant capital and operational costs are also
`incurred by the necessity of operating a large dis(cid:173)
`tillation column (the "Heavies" column) to
`remove low levels of high boiling point impurities,
`with propionic acid being the major component.
`
`The Cativa ™ Iridium Catalyst for
`Methanol Carbonylation
`Due to the limitations described above and also
`because of the very attractive price difference
`between rhodium ($5200 per troy oz) and iridium
`($300 per troy oz) which existed in 1990, research
`into the use of iridium as a catalyst was resumed by
`
`BP in 1990, after earlier work by Monsanto. The
`initial batch autoclave experiments showed signif(cid:173)
`icant promise, and the development rapidly
`required the coordinated effort of several diverse
`teams.
`One early finding from the investigations was
`of the extreme robustness of the iridium catalyst
`species (5). Its robustness at extremely low water
`concentrations (0.5 wt.%) is particularly significant
`and ideal for optimisation of the methanol car(cid:173)
`bonylation process. The iridium catalyst was also
`found to remain stable under a wide range of con(cid:173)
`ditions that would cause the rhodium analogues to
`decompose completely to inactive and largely
`irrecoverable rhodium salts. Besides this stability,
`iridium is also much more soluble than rhodium in
`the reaction medium and thus higher catalyst con(cid:173)
`centrations can be obtained, making much higher
`reaction rates achievable.
`The unique differences between the rhodium
`and iridium catalytic cycles for methanol carbony(cid:173)
`lation have been investigated in a close partnership
`between researchers from BP Chemicals in Hull
`and a research group at the University of Sheffield
`(6). The anionic iridium cycle, shown in Scheme II,
`is similar to the rhodium cycle, but contains
`
`Platinum Metals Rev., 2000, 44, (3)
`
`97
`
`004
`
`

`
`COMe
`co
`I
`I....__
`......._ m /
`/lr'-.....
`CO
`
`I
`
`Scheme ll
`Catalytic cycle for the
`arbonylation of methanol
`usmg iridium
`
`E
`
`CO
`1....__
`....__ 1 /
`lr
`1 / ........._co
`
`Me
`
`Me
`
`1~ I _...,..co
`lrm
`1 / 1-.........co
`co
`
`F
`
`co
`
`sufficient key differences to produce the major
`advantages seen with the iridium process.
`Model studies have shown that the oxidative
`addition of methyl iodide to the iridium centre is
`about 150 times faster than the equivalent reaction
`with rhodium (6). This represents a dramatic
`improvement in the available reaction rates, as this
`step is now no longer rate determining (as in the
`case of rhodium). The slowest step in the cycle is
`the subsequent migratory insertion of CO to form
`the iridium-acyl species, F, which involves the
`elimination of ionic iodide and the coordination of
`an additional CO ligand. This would suggest a
`totally different form of rate law:
`
`Rate oc [catalyst] X [CO]
`W1
`or, taking the organic equilibria into account:
`
`(vii)
`
`Rate oc [catalyst] X [CO] X [MeOAc]
`
`(viii)
`
`The implied inverse dependence on ionic iodide
`concentration suggests that very high reaction rates
`should be achievable by operating at low iodide
`concentrations. It also suggests that the inclusion
`of species capable of assisting in removing iodide
`should promote this new rate limiting step.
`Promoters for this system fall within two distinct
`groups:
`
`o simple iodide complexes of zinc, cadmium,
`mercury, gallium and indium (T), and
`o carbonyl-iodide complexes of tungsten, rhenium,
`ruthenium and osmium (8, 9).
`
`Batch Autoclave Studies
`The effect on the reaction rate of adding five
`molar equivalents of promoter to one of the iridi(cid:173)
`um catalyst is shown in Table II. A combination of
`promoters may also be used, see runs 13 and 14.
`None of these metals are effective as carbonylation
`catalysts in their own right, but all are effective
`when used in conjunction with iridium.
`The presence of a promoter leads to a substan(cid:173)
`tial increase in the proportion of "active anionic"
`species [Ir(CO),I,Mer, E, and a substantial
`decrease in the "inactive" [Ir(C0)2I.]-. A suggested
`mechanism for the promotion of iridium catalysis
`by a metal promoter [M(CO),ly], is given in
`Scheme III. The promotion is thought to occur via
`direct interaction of promoter and iridium species
`as shown. The rate of reaction is dependent upon
`the loss of iodide from [Ir(C0)2I,Mer. These metal
`promoters are believed to reduce the standing con(cid:173)
`centration of I- thus facilitating the loss of iodide
`from the catalytic species. It is also postulated that
`carbonyl-based promoters may then go on to
`donate CO in further steps of the catalytic cycle.
`
`Platit111m Metals Rev., 2000, 44, (3)
`
`98
`
`005
`
`

`
`Table II
`Effect of Various Additives on the Rate for the lridium-Catalysed Carbonylation of Methanol' from
`Batch Autoclave Data
`
`Experimental
`run
`
`1
`2
`3
`4
`5
`6
`7
`8
`9
`10
`11
`12
`13
`14
`15
`
`Additive
`
`None
`Lil
`Bu,NI
`Ru(CO),I,
`Os(CO),I,
`Re(CO),CI
`W(CO),
`Znl,
`Cdl,
`Hgl,
`Gal,
`In I,
`lni,/Ru(CO),I,
`Zni,/Ru(CO),I,
`Ru(CO),I,
`
`Additive:iridium,
`molar ratio
`
`Carbonylation rate,
`mol dm-' h-'
`
`-
`1:1
`1:1
`5:1
`5:1
`5:1
`5:1
`5:1
`5:1
`5:1
`5:1
`5:1
`5:1:1
`5:1:1
`Control: no iridiumb
`
`8.2
`4.3
`2.7
`21.6
`18.6
`9.7
`9.0
`11.5
`14.7
`11.8
`12.7
`14.8
`19.4
`13.1
`0'
`
`a Reaction conditions: 190"C, 22 barg, and 1500 rpm. Autoclave charge: methyl acetate (648 mmol), water (943 mmol), acetic acid
`( 1258 mmol), methyl iodide (62 mmol), and H,IrCl, ( 1.56 mmol) plus additive as required. Carbonylation rate, in mol dm-' h-',
`measured at 50 per cent conversion of methyl acetate.
`b Control experiment conducted in the absence of iridium. Amount of the ruthenium complex used is the same as in run 4.
`c No CO uptake observed
`
`Another key role of the promoter appears to be
`in the prevention of the build up of "inactive"
`forms of the catalyst, such as [Ir(C0)2I.r and
`[Ir(CO)J,]. These species are formed as intermedi(cid:173)
`ates in the water gas shift reaction.
`For the rhodium system the rate of the carl)ony(cid:173)
`lation reaction is dependent only upon the
`concentrations of rhodium and methyl iodide.
`However, the situation is more complex for the pro(cid:173)
`moted iridium system. Table ill illustrates the effect
`
`of the system parameters on the rate of reaction.
`The effect of water concentration on the car(cid:173)
`bonylation rates of a rhodium system and an
`iridium/ ruthenium system is illustrated in Figure
`3. For rhodium, a decline in carbonylation rate is
`observed as the water content is reduced below
`about 8 wt. %. There are a number of possible the(cid:173)
`ories for this, including a possible build up of the
`"inactive" [Rh(C0)2I.r species formed in the
`water gas shift cycle at lower water concentrations,
`
`)(MG·~~,
`
`Mel + AcOH
`(H20)
`
`Scheme III
`A proposed mechanism for the promotion of iridium catalysis by a metal promoter, [M( CO Jl..( solv) ].
`The solvent could be water or methanol
`
`Platin11m Metals Rev., 2000, 44, (3)
`
`99
`
`006
`
`

`
`Table Ill
`The Rate Dependence Differences between the Rhodium and Iridium Systems
`
`Water
`
`Methyl acetate
`
`Methyl iodide
`
`Rhodium
`
`I rid iu m/promoter
`
`1 st order below 8 wt.%
`Independent above 8 wt.%
`
`Increases with increasing water
`up to- 5 wt.%, then decreases
`with increasing water
`
`Independent above - 1 wt.%
`
`Increases with increasing methyl acetate
`
`1st order
`
`Increases with increasing methyl iodide
`up to- 6 wt.%, then independent
`
`Increases with increasing CO partial pressure.
`As the CO partial pressure falls below - 8 bara
`the rate decreases more rapidly
`
`As the corrosion metals increase in
`concentration, the rate decreases
`
`Non applicable
`
`1st order, effect tails off at high
`catalyst concentrations
`
`Increases with increasing promoter,
`effect tails off at higher concentrations
`
`CO partial pressure
`
`A minimum CO partial pressure is
`required; above this, independent
`
`Corrosion metals
`
`Independent
`
`Rhodium
`
`Iridium
`
`1st order
`
`Non applicable
`
`Promoter
`
`Non applicable
`
`bara u bar ab.wlute; atmosphenc pressure= I bar absolute(= 0 bar gauge, burg)
`
`which is a precursor for the formation of insoluble
`Rhl,.
`Another theory for the decline in the carbony(cid:173)
`lation rate is that the rate determining step in the
`catalytic cycle changes to the reductive elimination
`(attack by water) instead of oxidative addition. This
`is consistent with the increased amount of
`acetaldehyde-derived byproducts in a low water
`concentration rhodium system, as the rhodium-
`
`acyl species, D, is longer lived.
`At lower water concentrations, the addition of
`ionic iodides, especially Group I metal iodides, to
`the process has been found to stabilise the rhodi(cid:173)
`um catalysts and sustain the reaction rate by
`inhibiting the water gas shift cycle, inhibiting the
`formation of [Rh(C0)2I.]- and its degradation to
`Rhl, and promoting the oxidative addition step of
`the catalytic cycle (10-13).
`
`5
`10
`WATER CONCENTRATION, "/oW/w
`
`15
`
`20
`
`Fig. 3 A comparison of carbonylation rates
`for iridium/ruthenium and rhodium processes
`depending on water concentration. These
`batch autoclave data were taken under
`conditions of- 30 % w/w methyl acetate,
`8.4 % wlw methyl iodide, 28 burg total
`pressure and 190°C; (barg is a bar gauge,
`referenced to atmospheric pressure, with
`atmospheric pre.1·sure = 0 bar gauge)
`
`Platinum Metals Rev., 2000, 44, (3)
`
`100
`
`007
`
`

`
`Fig. 4 The effect of
`catalyst concentration on
`the carbonylation rate
`for an unpromoted and a
`ruthenium-promoted
`iridium catalyst. The
`ruthenium promoter is
`effective over a wide
`range of catalyst
`concentrations. Batch
`autoclave data were
`taken at - 20 % wlw
`methyl acetate, 8 % wlw
`methyl iodide, 5.7% wlw
`water, 28 barg total
`pressure and /90°C
`
`·~
`':"E so
`" 15
`
`E 40
`
`w
`~ 30
`a::
`z
`Q 20
`....
`
`:3 >- 10 g
`
`a:
`<(
`u
`
`___ _..
`
`---r--
`
`-t __ -.::::._lr only
`
`-·--
`
`500
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`3500
`
`4000
`
`IRIDIUM CONCENTRATION, ppm
`
`However, there is also a downside, in the
`lithium-promoted rhodium system, the acetalde(cid:173)
`hyde is not scavenged sufficiently by the catalyst
`system to form propionic acid and therefore the
`concentration of acetaldehyde increases, conden(cid:173)
`sation reactions occur and higher non-acidic
`compounds and iodide derivatives are formed, for
`example hexyl iodide. Further purification steps
`are then required (14).
`For a Cativa ™ system, in contrast to rhodium,
`the reaction rate increases with decreasing water
`content, see Figure 3. A maximum value is reached
`at around 5% w/w (under the conditions shown).
`Throughout this region of the curve the iridium
`species observed are [Ir(CO),I.r (the "inactive"
`species which is formed in the water gas shift
`cycle) and [Ir(CO)zi.Mer (the "active" species in
`the anionic cycle). When the water concentration
`falls bdow 5 % w /w the carbonylation rate declines
`and the neutral "active" species [Ir(CO),I) and the
`corresponding "inactive" water gas shift species
`[Ir(CO),I3] are observed.
`
`Other Factors Mfecting the Reaction Rate
`(i) Methyl acetate concentration
`In the rhodium system, the rate is independent
`of the methyl acetate concentration across a range
`of reactor compositions and process conditions
`(1). In contrast, the Cativa™ system displays a
`strong rate dependence on methyl acetate concen(cid:173)
`tration, and methyl acetate concentrations can be
`increased to far higher levds than in the rhodium
`system, leading to high reaction rates. High methyl
`acetate concentrations may not be used in the
`
`rhodium process because of catalyst precipitation
`in downstream areas of the plant.
`(ii) Methyl iodide concentration
`The reaction rate for Cativa ™ has a reduced
`dependency on the methyl iodide concentration
`compared with the rhodium system. This is con(cid:173)
`sistent with the fast rate of oxidative addition of
`methyl iodide to [Ir(CO)zlzr giving [Ir(CO)zi.Mer.
`(iii) CO partial pressure
`The effect of CO partial pressure in the
`Cativa ™ process is more significant than for the
`rhodium process with the rate being suppressed
`bdow 8 bara when operating in the ionic cycle.
`(iii) Poisoning the Cativa TM system
`Corrosion metals, primarily iron and nickd,
`poison the Cativa ™ process. However, it is not the
`corrosion metals themsdves that poison the
`process, but rather the ionic iodide which they
`support that inhibits the iodide loss step in the
`carbonylation cycle, see Scheme II.
`(iv) Catalyst concentration
`The effects of catalyst concentrations on the
`carbonylation rate for an unpromoted and for a
`ruthenium-promoted iridium catalyst are shown in
`Figure 4. The ruthenium promoter is effective
`over a wide range of catalyst concentrations. As
`high catalyst concentrations and high reaction
`rates are approached a deviation from first order
`behaviour is noted, and a small but significant loss
`in reaction sdectivity is observed.
`(v) Promoters
`The addition of further promoters, to the ones
`already present, for example iridium/ ruthenium,
`can have positive effects. For instance, a synergy is
`
`Piati1111m Metals Rev., 2000, 44, (3)
`
`101
`
`008
`
`

`
`Table IV
`
`Effect of Lithium Iodide Additions on the Carbonylation Rate for Iridium and Iridium/Ruthenium
`Catalysed Methanol Carbonylation' from Batch Autoclave Data
`
`Experimental
`run
`
`Catalyst system
`
`Water,
`%w/w
`
`Carbonylation rate,
`mol dm-' h-'
`
`1
`2
`3
`4
`
`Iridium only
`Iridium/lithium 1:1 molar ratio
`Iridium/ruthenium 1:2 molar ratio
`Iridium/ruthenium/lithium 1:2:1 molar ratio
`
`2.1
`2.0
`2.0
`2.0
`
`12.1
`6.3
`15.1
`30.8
`
`" Reaction conditions: 190"C, 28 barg total pressure, and 30 % wlw methyl acetate, 8.4 % w!w methyl iodide and 1950 ppm iridium
`
`observed between the promoters and iodide salts,
`such as lithium iodide (15). Iodides usually poison
`the iridium catalyst, for example, if lithium iodide
`is added to an iridium-only catalyst at low water
`(- 2% w/w) and high methyl acetate (30% w/w),
`there is a markedly reduced carbonylation rate. A
`ratio of one molar equivalent of lithium iodide:
`iridium reduces the reaction rate by 50 per cent,
`see run 2 in Table IV but, under the same reaction
`conditions two molar equivalents of ruthenium:
`iridium increases the carbonylation rate by 25 per
`cent. Remarkably, adding lithium iodide to the
`ruthenium-promoted catalyst under these condi(cid:173)
`tions further doubles the carbonylation rate (run
`4). The net effect is that ruthenium and lithium
`iodide in combination under certain conditions
`increase the reaction rate by 250 per cent with
`respect to an unpromoted iridium catalyst. Thus,
`adding low levels of iodide salts to a promoted irid(cid:173)
`ium catalyst allows the position of the rate
`maximum, with respect to the water concentration,
`
`to be moved to even lower water.
`The effect of the lithium iodide:iridium molar
`ratio on the carbonylation rate is shown in Figure
`5 for a ruthenium-promoted iridium catalyst, hav(cid:173)
`ing iridium:ruthenium molar ratios of 1:2 and 1:5.
`Under these conditions an exceptionally high rate
`of 47 mol dm-' h-' can be achieved with a molar
`ratio for iridium:ruthenium:lithium of 1:5:1.
`
`Interdependence of Process Variables
`The Cativa ™ process thus displays a complex
`interdependence "between all the major process
`variables, notably between [methyl acetate],
`[water], [methyl iodide], [Iridium], CO partial pres(cid:173)
`sure, temperature and the promoter package used.
`For example, the methyl iodide concentration,
`above a low threshold value, has only a small influ(cid:173)
`ence on the reaction rate under certain conditions.
`However, when the reaction rate is declining with
`reducing water concentration, as shown for a
`ruthenium-promoted iridium catalyst in Figure 3,
`
`'"'
`7E 50r-------------------------------------------------~
`"' 45
`~40
`35
`.....
`~ 30
`a:: 25
`~ 20
`~ 15
`..J
`~ 10
`16
`5
`~ ~------~--------~--------~------~--------~~
`u
`1.5
`2.5
`2
`0.5
`ADDED Lil, MOLAR EQUIVALENTS TO IRIDIUM
`
`Fig. 5 The effect of adding
`a second promoter of
`lithium iodide to ruthenium(cid:173)
`promoted iridium catalysts
`on the methanol
`carbonylation rates. Batch
`autoclave data taken at
`2 % w/w water and
`30 % w/w methyl acetate
`
`Plalin11m Metals Rev., 2000, 44, (3)
`
`102
`
`009
`
`

`
`increasing the methyl iodide concentration from
`8.4 to 12.6 % w/w doubles the reaction rate.
`Increasing the methyl iodide concentration under
`these conditions also increases the effectiveness of
`the ruthenium promoter (16). In the CativaTM
`process these interactions are optimised to max(cid:173)
`imise reactor productivity and reaction selectivity
`and minimise processing costs.
`In addition to the batch autoclave studies, a
`pilot plant unit operating under steady state condi(cid:173)
`tions was used to optimise the CativaTM process.
`The unit provided data on the carbonylation rate,
`the byproducts, catalyst stability, corrosion rates
`and product quality under continuous steady state
`operation.
`
`Purification
`The quality of the acetic acid produced in the
`Cativa TM process is exceptional. It is inherently low
`in organic iodide impurities, which trouble other
`low water,
`rhodium-based, processes
`(14).
`Acetaldehyde is responsible for the formation of
`the higher organic iodide compounds via a series
`of condensation steps and other reactions. These
`higher iodides are difficult to remove by conven(cid:173)
`tional distillation techniques and further treatment
`steps are sometimes necessary to ensure that the
`acetic acid is pure enough for all end uses.
`In particular ethylene-based vinyl acetate man(cid:173)
`ufacturers or those using palladium catalysts
`require the iodide concentration in the acetic acid
`to be at a low ppb level (14). In the CativaTM
`process the levels of acetaldehyde in the reactor
`are very low, typically less than 30 ppm, compared
`to a few hundred ppm in the conventional
`Monsanto process and several hundred ppm in the
`lithium-promoted rhodium process. Further treat(cid:173)
`ment steps are not therefore necessary to give a
`product that can be used directly in the manufac(cid:173)
`ture of vinyl acetate.
`The levels of propionic acid in the acetic acid
`from the Cativa TM process are substantially less
`than those from the rhodium process. In the con(cid:173)
`ventional high water content rhodium process, the
`propionic acid present in the acetic acid product
`prior to the "Heavies" removal column is between
`
`1200 and 2000 ppm. In the Cativa™ process these
`concentrations are reduced to about one third of
`these levels.
`
`The Environmental Impact of Cativa ™
`As the Cativa™ process produces substantially
`lower amounts of propionic acid compared to the
`rhodium process, much less energy is required to
`purify the product. As mentioned previously, the
`CativaTM system can be operated at much lower
`water concentrations, thus reducing the amount of
`energy require~ to dry the product in the distilla(cid:173)
`tion train. Steam and cooling water requirements
`are reduced by 30 per cent compared to the rhodi(cid:173)
`um system. The water gas shift reaction does
`occur with CativaTM, as with rhodium, but at a
`lower rate, resulting in - 70 per cent lower direct
`C02 emissions. Overall, including indirect C02
`emissions, the Cativa TM process releases about 30
`per cent less C02 per tonne of product than does
`the rhodium process. The comparative insensitivi(cid:173)
`ty of the system to the partial pressure of CO
`allows operation with lower reactor vent rates than
`in the rhodium system. This results in the com(cid:173)
`bined benefits of less purge gas released to the
`atmosphere via the flare system and also greater
`CO utilisation, leading to decreased variable costs.
`In practice, total direct gaseous emissions can be
`reduced by much more than 50 per cent.
`
`Cost Reductions
`As discussed before there are a number of fac(cid:173)
`tors which have lead to substantial variable cost
`reductions for the Cativa ™ process compared to
`the rhodium process. In particular, steam usage is
`reduced by 30 per cent, while CO utilisation is
`increased from - 85 per cent to > 94 per cent.
`The Cativa ™ process also allows simplification
`of the production plant, which reduces the cost of
`a new core acetic acid plant by - 30 per cent. As
`the Cativa TM catalyst system remains stable down
`to very low water concentrations, the purification
`system can be reconfigured to remove one of the
`distillation columns completely and to combine
`the light ends and drying columns into a single col(cid:173)
`umn. The lower production rates of higher acids,
`
`Platinum Metals Rev., 2000, 44, (3)
`
`103
`
`010
`
`

`
`Off gas to scrubber and flare
`
`Ace-tic
`acid
`
`Re.actor
`
`Flash tank
`(Catalyst rich
`stream recycled)
`
`Drying
`column
`
`Propionic
`acid
`(byproduct)
`
`"Heavies"
`removal
`column
`
`Fig. 6 Simplified process flow sheet for a commercial scale Cativa'M methanol carbonylation plant. The low boiler and
`water removal duties are combined into one, smaller, distillation column. The size of the high boiler removal column
`has also been reduced
`
`compared to the Monsanto process, allows the size
`and operating cost of the final distillation column
`to be reduced. The major units of a commercial
`scale Cativa TM methanol carbonylation plant are
`shown in Figure 6.
`The reactor in the CativaTM system does not
`require a traditional agitator to stir the reactor con(cid:173)
`tents. Eliminating this leads to further operational
`and maintenance cost savings. The reactor con(cid:173)
`tents are mixed by the jet mixing effect provided
`by the reactor cooling loop, in which material
`leaves the base of the reactor and passes through a
`cooler before being returned to the top of the reac(cid:173)
`tor. A secondary reactor after the main reactor and
`before the flash tank further increases CO utilisa(cid:173)
`tion by providing extra residence time under plug
`flow conditions for residual CO to react and form
`acetic acid.
`
`Conclusions
`The new Cativa TM iridium-based system delivers
`many benefits over the conventional Monsanto
`rhodium-based methanol carbonylation process.
`The technology has been successfully proven on a
`commercial scale at three acetic acid plants world(cid:173)
`wide having a combined annual production of 1.2
`million tonnes. These benefits include:
`• an inherendy stable catalyst system
`
`• less dependence on CO partial pressure
`• the