Applied Industrial Catalysis
`
`Volume 1
`
`Edited by
`BRUCE E. LEACH
`Conoco Inc .
`Research and Development
`Ponca City, Oklahoma
`
`1983
`
`ACADEMIC PRESS
`A Subsidiary of Harcourt Brace Jovanovich, Publishers
`New York London
`Paris San Diego San Francisco Sao Paulo Sydney Tokyo Toronto
`
`IPR 2015-00171
`Exhibit 1066
`
`000001
`
`

`
`COPYRIGHT@ 1983, BY ACADEMIC PRESS, ]NC.
`ALL RIGHTS RESERVED.
`NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR
`TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC
`OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY
`INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT
`PERMISSION IN WRITING FROM THE -PUBLISHER.
`
`ACADEMIC PRESS, INC.
`111 Fifth Avenue, New York, New York 10003
`
`United Kingdom Edition published by
`ACADEMIC PRESS, INC. (LONDON) LTD.
`24/28 Oval Road , London NW! 7DX
`
`Library of Congress Cataloging in Publication Data
`
`Main entry under title :
`
`Applied industrial catalysis.
`
`Includes inde x.
`I . Leach, Bru~e E.
`I. Catalysis.
`660.2'995
`TPl56.C35A66
`1983
`ISBN 0-12-440201-1 (v. l)
`
`82-22751
`
`PRINTED IN THE UNITED STATES OF AMERICA
`
`83 84 85 86
`
`987654321
`
`000002
`
`

`
`CHAPTER 10
`
`Methanol Carbonylation to
`Acetic Acid
`
`R. T. EBY T. c. SINGLETON
`
`Process Technology Department
`Monsanto Company
`Texas City, Texas
`
`.
`
`.
`
`I. History of Acetic Acid Processes
`II. Process Description.
`.
`A. Reaction Area .
`.
`B. Purification Area .
`C. Control . . . .
`Ill. Licensing Activities.
`.
`JV. Process Chemistry
`.
`.
`A. Methanol Carbo nylation Reaction
`B. Competing Reactions .
`References
`.
`.
`.
`.
`.
`.
`.
`
`.
`
`.
`
`.
`
`275
`276
`276
`277
`278
`278
`279
`279
`284
`295
`
`I. History of Acetic Acid Processes
`
`Throughout civilization acetic acid (CH 3COOH) has played an important
`role. For many years the only source of this acid was the oxidation of ethanol
`in fermented liquids. It was not until 1920 in the United States that catalytic
`production of acetic acid was begun by the oxidation of acetaldehyde in the
`presence of manganous acetate. From these beginnings, both the need for
`improved production methods and the development of these methods grew
`rapidly. In 1960, Badische Anilin- & Soda-Fabrik AG (BASF) introduced
`the carbonylation of methanol to acetic acid to the list of several oxidation
`processes currently used. This method employs the bubbling of carbon
`monoxide through a reacting mixture containing cobalt as a catalyst at
`l 80°C and 3000-10,000 psig. Acetic acid yields of 90% based on methanol
`and 70% based on CO have been reported by Von Kutepow et al.[!]. This
`was one of the first examples of commercially reacting simple reactants to
`form the more complex acetic acid molecule. All previous chemistry had
`involved various levels of oxidation of ethanol, acetaldehyde, or hydrocar(cid:173)
`bons to produce the acetic acid. One of these plants, licensed by BASF, was
`
`Applied Industrial Catal ysis, Volume I
`
`275
`
`Copyright© 1983 Academic Press, Inc.
`All rights or reproduction in any form reserved.
`ISBN: 0-1 2-440201-1
`
`000003
`
`

`
`276
`
`R. T. Eby and T. C. Singleton
`
`the Borden Company plant in Geismar, Louisiana. It was the only BASF
`process built in the United States before a new breakthrough occurred.
`In the late 1960s the Monsanto Company developed a process for car(cid:173)
`bonylating methanol in the presence of a rhodium catalyst to produce acetic
`acid in high yields at low pressures and reasonable temperatures. High
`selectivity with a minimum number of by-products has made this process
`the current choice of companies throughout the world. After successful
`initial batch studies, development of a continuous process was initiated.
`Close coordination among the development, engineering, and manufactur(cid:173)
`ing groups, resulted in the start-up of a commercial unit just 4 yr after
`generation of the idea in 1966. Major hurdles overcome during its develop(cid:173)
`ment involved materials of construction, catalyst losses, catalyst inventory,
`product purity, and process control. The resulting process proved to be a
`simple, efficient, computer-controlled unit producing acetic acid that ex(cid:173)
`ceeded the quality of USP, food grade, and reagent grade acid and was
`suitable for all commercial uses. Monsanto was awarded the Kirkpatrick
`Merit Award in 1971 for this development.
`
`II. Process Description
`
`Acetic acid is produced by continuously reacting methanol and carbon
`monoxide in a homogeneous catalytic reactor at < 200°C and < 500 psi
`pressure, as shown schematically in Fig. 1. The product is glacial acetic acid
`of 99.9 + %purity. Small amounts of propionic acid, hydrogen, methane,
`and carbon dioxide are produced in the process. Propionic acid is inciner(cid:173)
`ated. Hydrogen, methane, and carbon dioxide are flared.
`
`A. REACTION AREA
`
`Acetic acid is produced in a liquid phase system in a continuously agitated
`reactor. The reaction is catalyzed by soluble rhodium compounds with an
`iodine promoter. There are two primary reactions that occur in the acetic
`acid process that consume reactants. These are carbonylation of methanol
`with carbon monoxide to form acetic acid, and a water gas shift reaction that
`forms carbon dioxide and hydrogen from carbon monoxide and water. As in
`any catalyst and promoter system, the rhodium and iodine compounds
`participate in exchange reactions but are not consumed. The theoretical
`reaction rate of the carbonylation reaction has been determined to be a
`function of at least temperature, promoter, and rhodium concentration. It
`has been found that the theoretical reaction rate of methanol carbonylation
`
`000004
`
`

`
`I 0 Methanol Carbonylation to Acetic Acid
`
`277
`
`are
`
`__ _,.,.,[ Recycle
`
`-lr'>-......-i
`
`Product
`to
`Storage
`
`Carbon
`Monox ide
`Methanol
`
`REACTOR
`SYSTEM
`
`LIGHT ENDS
`COLUMN
`
`DRYING
`HEAVY ENDS
`COLUMN
`COLUMN
`Fig. I. Monsanto acetic acid process.
`
`increases with (a) increases in temperature, (b) increases in iodine concen(cid:173)
`tration, and (c) increases in rhodium concentration. It has also been found
`that the reaction rate is independent of the concentration of methanol and
`carbon monoxide as long as both methanol and carbon monoxide are
`available for reaction. This condition requires the carbon monoxide to be
`dissolved in the reaction liquid before it can participate in the reaction. The
`two conditions that determine the speed with which carbon monoxide goes
`into solution are the rate of agitation and the partial pressure of carbon
`monoxide in the gas above the liquid.
`An unusual aspect of the acetic acid process is that the reactor is not
`operated at the maximum theoretical rate because at the maximum theoret(cid:173)
`ical methanol feed rate any slight variation in operating conditions could
`mean that more methanol would be fed than could be reacted. Such a
`situation, if continued, would result in upset conditions, making the reac(cid:173)
`tion and purification sections difficult to operate.
`
`B. PURIFICATION AREA
`
`The resulting crude acetic acid is sent from the reaction system to the first
`column which separates the light ends from a heavy recycle stream. Wet
`
`000005
`
`

`
`278
`
`R. T. Eby and T. C. Singleton
`
`acetic acid is sent as a side stream into the drying column. A water-acetic
`acid mixture is removed overhead in the drying column and recycled to the
`reactor to provide part of the cooling for the reaction. Dry acetic acid is sent
`to the final purification column in which propionic acid is separated as a
`waste stream from the bottom. Product acetic acid is withdrawn as a side
`stream, cooled, and sent to storage. The waste stream is concentrated in a
`small waste acid stripper. The inert gases resulting from the water gas shift
`reaction are sent to a scrubber system that recovers light ends before the gas
`reaches the flare.
`
`C. CONTROL
`
`It became obvious early in the development of this process that efficient
`control of the variables was necessary to ensure high yields and control the
`effect several recycles have on the process. An efficient supervisory control
`was developed that ensures methanol yields above 98% and greater than
`90% yields for carbon monoxide. The scheme effectively controls the sys(cid:173)
`tem's heat balance.
`In addition to the continuous-process equipment, an auxiliary system
`operates under batch conditions to prepare the catalyst and promoter and to
`regenerate the catalyst.
`
`III. Licensing Activities
`
`In 1973, a licensing program began. The first license was granted to the
`USSR. Since then, a total of nine active licenses have been granted through(cid:173)
`out the world, as shown in Table I. Of these, five plants are on stream:- USSR,
`
`T ABLE I
`
`Acetic Acid Licensees (Monsanto Process)
`
`Company
`
`Country
`
`Capacity (MTA)
`
`Celanese
`USI
`Dainihon
`Kyodo Sakusan
`British Petroleum
`Techmashimport
`Rhone-Poulenc
`MSK
`China Petroleum Development Corporation
`
`United States
`United States
`Ja pan
`Japan
`England
`USSR
`France
`Yugoslavia
`T aiwan
`
`272,000
`272,000
`200,000
`200,000
`170,000
`150,000
`225 ,000
`100,000
`80,000
`
`000006
`
`

`
`IO Methanol Carbonylation to Acetic Acid
`
`279
`
`Celanese, USI, Kyodo Sakusan, and Rhone-Poulenc. These plants represent
`over 90% of the world's new production since 1973. Together with Mon(cid:173)
`santo's plant, carbonylation of methanol by the Monsanto process repre(cid:173)
`sents about 40% of the world's current production. The oxidation processes
`suffer from high raw material costs and low yields. Replacement of these
`processes is forseen in the future, except where the by-products play an
`important economic role in a specific situation.
`
`IV. Process Chemistry
`
`A. METHANOL CARBO NYLA TION REACTION
`
`1. Effects of Reaction Parameters
`
`These studies have been carried out with both batch and continuous
`reactor equipment. Any rhodium compound in a solution under reaction
`conditions forms an active catalyst. The use of many rhodium compounds
`for this reaction has been reported [2]. Iodine may be charged as hydrogen
`iodide, methyl iodide, or iodine. All these iodine components are of equiva(cid:173)
`lent activity. Iodine is reduced to hydrogen iodide by this catalyst system.
`The reaction proceeds at measurable rates at moderate temperature and
`carbon monoxide partial pressure (> l 50°C and > 100 psi, respectively).
`Batch reactions are normally carried out in rapidly stirred autoclaves.
`Carbon monoxide is fed to maintain constant pressure in the reactor as it is
`consumed in the reaction. Reaction rates in batch reactions may be deter(cid:173)
`mined by either the rate of carbon monoxide uptake or by the rate of
`decrease in active methyl groups in the catalyst solution. An active methyl
`group is defined as a methyl (CH3) group that can be converted to an acetyl
`
`[CHaj >[I]
`
`z
`0
`I-
`C
`Ill:
`
`1-z ... <..> z
`
`0
`<..>
`
`TIME
`Fig. 2. Rhodium-catalyzed carbonylation of methanol in a batch reaction.
`
`000007
`
`

`
`280
`
`R. T . Eby and T. C. Singleton
`
`(CH 3CO) group under the conditions of this reaction, e.g. , methanol, methyl
`acetate, methyl iodide, and dimethyl ether. Figure 2 shows a typical rate
`order for this reaction.
`In a batch reaction of this type in which the initial methanol is present in a
`significant molar excess over the iodide concentration, the carbonylation
`rate is apparently zero order over a major portion of the reaction. The
`deviation from zero order occurs at the reaction stage when the molar
`concentrations of iodide and active methyl groups are approximately equiv(cid:173)
`alent. Under reaction conditions all components ofEqs. (l)-(5) are present
`in equilibrium concentrations. At normal reaction temperatures(> 150°C)
`these equilibria are established instantaneously.
`
`CH 3COOH + CH 30H ;:::: CH 3COOCH 3 + Hp
`2CH 30H ;:::: CH 30CH 3 + H 20
`CH 30H + HI ..,,. CH 3l + H 20
`CH 3COOCH 3 + HI ~ CH 31 + CH 3COOH
`CH 30CH 3 + HI ..,,. CH 31 + CH 30H
`
`(l)
`(2)
`(3)
`(4)
`(5)
`
`Dimethyl ether is normally present in the catalyst solution in only a trace
`amount. Methyl iodide is the equilibrium-favored component in Eqs.
`(3)-(5). Therefore, when active methyl groups are present in a molar excess
`over iodine, a major fraction of the iodine exists in equilibrium as methyl
`iodide. Concentrations of methanol and methyl acetate decrease as these
`components are converted to acetic acid during the course of the reaction.
`The methyl iodide concentration remains nearly constant during the zero(cid:173)
`order portion of a batch reaction. The rate during this zero-order portion of
`the reaction is directly proportional to both the rhodium and iodide con(cid:173)
`centrations over very wide ranges of these two variables.
`The carbonylation rate is essentially independent of the initial methanol
`concentration. The only effect of the initial methanol concentration is a
`catalyst dilution factor due to an increasing liquid volume during the course
`of the reaction. This volume change is greater at higher initial methanol
`concentrations. Water, produced initially by the reactions in Eqs. (I )-(3), is
`reduced in equilibrium concentration as the reaction proceeds. An initial
`water concentration over a moderate range has no significant effect on the
`methanol carbonylation rate. It has been reported that the reaction rate is
`independent of carbon monoxide partial pressure in the range 200-800 psi
`[3 , 4). The rhodium-catalyst complex is unstable at low carbon monoxide
`partial pressure, and therefore the reaction subsides and eventually ceases
`with decreasing carbon monoxide partial pressure. However, if the mini(cid:173)
`mum carbon monoxide pressure sufficient to sustain an active catalyst is
`attained, further increases in carbon monoxide partial pressure fail to
`accelerate the reaction. The minimum carbon monoxide partial pressure is
`
`000008
`
`

`
`I 0 Methanol Carbonylation to Acetic Acid
`
`281
`
`dependent on other reaction parameters, e.g. , temperature, concentrations
`of rhodium and iodine, and vapor pressure of the catalyst solution. This
`minimum threshold carbon monoxide partial pressure differs with each
`combination of these reaction parameters.
`The anionic iodocarbonyl-rhodate complexes that can exist in acetic
`acid-water solutions under various conditions under carbon monoxide
`pressure have been investigated by Denis Forster [5]. The compositions of
`these anions were assigned on the basis of their infrared spectra and elemen(cid:173)
`tal analyses of their quaternary ammonium salts. Samples of catalyst solu(cid:173)
`tions obtained during the zero-order portion of a batch methanol carbony(cid:173)
`lation reaction were light yellow in color. This color is typical of the d8
`square planar [Rh(CO)J2]- anion. An in situ infrared spectrum of this
`catalyst solution was obtained by the use of a high-pressure spectrophoto(cid:173)
`metric cell [6]. The spectrum showed bands at 1994 and 2064 cm- I, typical
`of the [Rh(C0)2I2]- anion. No infrared bands characteristic of other rho(cid:173)
`dium complexes were observed in the catalyst solution during the zero-order
`portion of a batch reaction. Near the end of the reaction, when the excess
`methanol and methyl acetate have been converted, further carbonylation of
`the remaining methyl iodide increases the hydrogen iodide concentration,
`as in Eq. (6).
`
`(6)
`
`As the hydrogen iodide concentration increases, the reaction subsides and,
`under most conditions, finally stops completely without the total conversion
`of methyl iodide. An infrared spectrum of the catalyst solution at the end of
`the reaction showed only a band at 2090 cm-I, typical of the trans(cid:173)
`[Rh(C0)214]- anion. This effect of acidity on the transition of the rhodium
`complexes is discussed in Section IV.B.1 on the water gas shift reaction.
`Iodide salts of alkali metals are inactive as cocatalysts in the rhodium(cid:173)
`catalyzed carbonylation of methanol, even though the [Rh(CO)il2]- com(cid:173)
`plex is formed in the presence of alkali metal iodides. Iodine remains
`combined as salts of these basic metals under reaction conditions. No
`methyl iodide was detected in the catalyst solution of an attempted batch
`reaction with potassium iodide as the iodine source. Iodide salts of more
`acidic metals, e.g., iron and nickel, are partially active as cocatalysts. These
`metal iodides are partially solvolyzed and converted to methyl iodide, as in
`Eq. (7).
`
`(7)
`
`Analyses of catalyst solution samples from batch reactions with ferrous or
`nickelous iodide as the iodine source show the presence of low concentra-
`
`000009
`
`

`
`282
`
`R. T. Eby and T. C. Singleton
`
`tions of methyl iodide. Carbonylation rates in these cases are lower than in
`reactions in which the iodine source is hydrogen iodide or methyl iodide.
`
`2. Mechanistic Interpretation
`
`The study of reaction parameters demonstrated that the methanol car(cid:173)
`bonylation rate was first-order-dependent on rhodium and iodine concen(cid:173)
`trations and independent of methanol concentration and carbon monoxide
`partial pressure. Under conditions of maximum rate the [Rh(C0)2I2]- and
`methyl iodide concentrations were in direct proportion to the total rhodium
`and iodine concentrations, respectively. On this basis it was suggested by
`Roth et al. [4] that the rate-controlling step involves the oxidative addition
`of methyl iodide to the monovalent [Rh(C0)2I2]- complex. They suggested
`that this rate-limiting step is followed by a sequence of more rapid reactions
`involving rearrangements, addition of carbon monoxide, and finally hydro(cid:173)
`lysis to produce acetic acid. Additions of covalent compounds to d8 and d 10
`complexes are well known [7, 8]. Additions of carbon monoxide to metal(cid:173)
`carbon or metal-hydride complexes have also been reported [9].
`A later investigation by Denis Forster [ 10-12] involved spectrophotome(cid:173)
`tric studies on the rhodium complexes in this reaction cycle. These studies
`provided evidence for the structures of these intermediates. On the basis of
`this study a mechanism consistent with the reaction kinetics was proposed
`for the rhodium-iodine-catalyzed carbonylation of methanol. Forster's
`mechanism is presented in Fig. 3.
`It was observed that, when a solution of [Rh(C0)2I2]- ions was reacted
`with methyl iodide at ambient temperature, the infrared bands of the
`original diiododicarbonylrhodate anion were replaced by bands at 2062 and
`1711 cm- 1 in the product. The 1711-cm-1 band is typical of an acetyl
`
`CH3
`CHJ -
`[
`r+co-
`r-'o
`I Rh I -
`I
`I
`CH3I I--tco
`-
`1 or 2
`Rh
`I-t-co
`I- CO~
`I Rh I
`1<II>
`I (Ill)
`~It::
`~co
`CO CH3-
`CO
`r+1co- Cff3COOCH3
`/1+/do
`+
`/
`or
`Rh
`Rh
`I -t H
`CH3COOH
`r---f- co
`I
`(V)
`I
`(IV)
`Fig. 3. Proposed mechanism of the rhodium-catalyzed carbonylation of methanol to acetic
`acid.
`
`CH30ff
`H,O
`
`-CH3COI
`
`000010
`
`

`
`IO Methanol Carbonylation to Acetic Acid
`
`283
`
`frequency. These infrared bands were assigned to intermediate III in the
`proposed reaction cycle. It is believed that the methyl iodide adduct (inter(cid:173)
`mediate II) exists only in transient form and rapidly rearranges to interme(cid:173)
`diate III. The elemental analysis of a quaternary amine salt of intermediate
`III is consistent with an atomic ratio calculated for the salt of intermediate
`III. The structure of this compound was also confirmed by x-ray diffraction
`[ 13]. This x-ray diffraction study indicated that intermediate III existed as a
`dimer. Vacuum distillation ofa tetraphenylarsonium salt of intermediate III
`produced the [Rh(CO)J2]- ion, demonstrating the reversibility of these
`reactions. The treatment of a solution of intermediate III with carbon
`monoxide at atmospheric pressure and ambient temperature rapidly con(cid:173)
`verted it to a component with CO stretching frequencies at 2 141 and
`2084 cm- 1 and an acetyl frequency at 1708 cm- 1• This species, which was
`assigned the structure of intermediate IV, slowly decomposes at room
`temperature to produce the original complex I of this reaction cycle.
`The formation of an acetyl complex (III) in the absence of carbon
`monoxide and the rapid reaction of intermediate III with carbon monoxide
`at low pressure are consistent with the lack of rate dependence of the
`carbonylation reaction on carbon monoxide partial pressure. The conver(cid:173)
`sion of methyl iodide to these intermediates explains the lack of rate
`dependence on methanol concentration. The first-order rate dependence on
`both rhodium and iodine suggests that the addition of methyl iodide to the
`diiododicarbonylrhodate anion (I) is the rate-controlling step in the metha(cid:173)
`nol carbonylation reaction. As mentioned previously, an infrared spectrum
`of the catalyst solution obtained under actual reaction conditions at elevated
`temperature showed the presence of only the [Rh(CO)zl2] - anion. The other
`steps in the reaction cycle after oxidative addition of methyl iodide are
`apparently very rapid. The other intermediates (II-IV) in the reaction cycle
`in Fig. 3 are not present in detectable concentrations under steady state
`conditions. This observation supports the conclusion that the oxidative
`addition of methyl iodide to the diiododicarbonylrhodate ion is rate-deter(cid:173)
`mining in this reaction.
`The mechanism proposed by Forster has alternative elimination steps.
`One of these involves solvolysis of the acetyl complex (IV) to form a hydride
`intermediate (V), followed by reductive elimination of hydrogen iodide.
`This is similar to a mechanism suggested for the carboxylation of alkyl
`chlorides by cobalt catalysts [ 14 ]. It was observed by Forster that carbonyla(cid:173)
`tion of anhydrous methyl iodide with a tetraphenylarsonium salt of a
`dihalodicarbonylrhodate catalyst at low temperature and pressure produced
`detectable amounts of acetyl iodide. Reductive elimination of an acyl halide
`by carbonylation of a trivalent rhodium phosphine acyl complex has been
`reported [ 15]. No oxidative addition of an acetyl halide to a [Rh(CO)zX2]-
`
`000011
`
`

`
`284
`
`R. T. Eby and T. C. Singleton
`
`complex occurred in 24 hr at 50°C. On the basis of these observations, the
`reductive elimination of acetyl iodide (IV --+ I) is the favored alternative for
`the final step of this reaction cycle.
`
`B. COMPETING REACTIONS
`
`The yields of acetic acid, based on both methanol and carbon monoxide,
`in this process are very high. Minor competing reactions have been ob(cid:173)
`served. These side reactions lead to a very small yield loss. They are
`discussed in the following sections.
`
`I. Water Gas Shift Reaction
`
`The water gas shift reaction was observed to occur as a side reaction in the
`rhodium-catalyzed methanol carbonylation reaction in early exploratory
`and process development studies [ 4, 16]. In this reaction, carbon monoxide
`and water are converted to hydrogen and carbon dioxide. This reaction also
`proceeds at moderate rates in acetic acid solutions in the absence of active
`methyl groups in this catalyst system under conditions similar to the
`methanol carbonylation reaction. Workers at the University of Rochester
`have observed this reaction to proceed at measurable rates in this catalyst
`system at a low temperature and subatmospheric pressure [ 17, 18]. The
`water gas shift reaction has been investigated the most extensively of any of
`the competing reactions in the rhodium-catalyzed methanol carbonylation
`process [19, 20].
`
`A. EFFECTS OF REACTION PARAMETERS. Experiments for determining
`reaction parameter effects were carried out in semibatch reactions by purg(cid:173)
`ing carbon monoxide through a catalyst solution of rhodium, hydrogen
`iodide, water, and acetic acid in a rapidly agitated autoclave at constant
`pressure and temperature. The water gas rates were determined by the total
`rate and carbon dioxide content of the off-gas. Mass spectrometric analyses
`of the off-gas showed the hydrogen and carbon dioxide contents to be
`equivalent in all cases. These reactions were not carried out in the presence
`of active methyl groups because of the high volatility of methyl iodide and
`because of the increase in hydrogen iodide concentration via carbonylation
`of methyl iodide during the reaction. Steady state conditions could not be
`attained in a semibatch process in the presence of methyl iodide.
`A variable study was carried out to determine the effects of temperature
`and the concentrations ofrhodium, hydrogen iodide, and water on the water
`gas rate. Four series of runs were carried out. Temperature, rhodium, and
`hydrogen iodide were maintained at constant levels in each separate group
`
`000012
`
`

`
`I 0 Methanol Carbonylation to Acetic Acid
`
`285
`
`of runs, whereas the water concentration was varied within each group.
`These experiments demonstrated a complex interaction of hydrogen iodide
`and water. In each of these series, a peak rate was observed with varying
`water content when the other variables were constant. The peak rates
`occurred at higher water levels at the higher hydrogen iodide concentrations.
`The presence of sodium iodide also affects the water gas rate. At low water
`levels the addition of sodium iodide reduces the rate, whereas sodium iodide
`addition enhances the rate at higher water levels. Water gas rates are
`enhanced by higher temperature and higher rhodium concentrations. Inter(cid:173)
`actions of carbon monoxide partial pressure with other variables were
`observed. This effect of carbon monoxide partial pressure is discussed in
`Section IV.B. l.b, on the mechanism of this reaction.
`
`It has been reported that the
`B. MECHANISTIC INTERPRET A TIO NS.
`diiododicarbonylrhodate anion reacts with aqueous hydriodic acid in the
`absence of carbon monoxide pressure to produce hydrogen, carbon monox(cid:173)
`ide, and the tetraiodocarbonylrhodate anion, as in Eq. (8) [5 , 21].
`
`(8)
`
`[Rh(C0),! 2] - + 2HI ~ [Rh(C0)14]- + CO + H2
`As mentioned previously in Section IV.A. I, an anionic monovalent and
`four different anionic trivalent rhodium carbonyl iodide complexes have
`been formed under different conditions (Table II). The structures of these
`rhodium complexes were assigned on the basis of their infrared spectra and
`elemental analyses of their tetraalkylammonium salts and, in the case of the
`trans-[Rh(C0)2I4]- anion, by x-ray crystallography [22].
`It is also known that most soluble trivalent rhodium compounds are
`converted to the monovalent complex [Rh(C0)2X2]- by heating under
`carbon monoxide pressure in the presence of water and hydrohalic acids
`with the liberation of carbon dioxide, as in Eq. (9) [23, 24 ]. It was therefore
`proposed that the water gas shift reaction in this homogeneous catalyst
`
`TABLE II
`
`Iodocarbonylrhodate Anions and Infrared Bands
`
`Anion
`
`Infrared wave number (cm- 1)
`
`[Rh(CO),l 2] (cid:173)
`[Rh(CO)l4](cid:173)
`cis-[Rh(C0),14](cid:173)
`trans-[Rh(C0),14](cid:173)
`[Rh(C0)1 5]2-
`
`1994, 2064
`2075
`2092, 2122
`2090
`2047
`
`Oxidation state
`of rhodium
`
`I+
`3+
`3+
`3+
`3+
`
`000013
`
`

`
`286
`
`R. T. Eby and T. C. Singleton
`
`system occurs by a combination of these oxidation and reduction steps
`under steady state conditions.
`
`Rh'•+ 3CO + H,O + 2x- -
`
`[Rh(CO),X,J- +CO,+ 2H+
`
`(9)
`
`An in situ infrared spectral study of water gas catalysts at< 200°C and < 500
`psig total pressure was performed by obtaining catalyst samples in acetic
`acid solutions in the high-pressure infrared cell mentioned in Section
`IV.A. l. This showed that the monovalent diiododicarbonylrhodate anion,
`[Rh(CO)zl2] - , was the predominant rhodium component at low hydrogen
`iodide and high water concentrations. At high hydrogen iodide and low
`water levels the rhodium is present primarily as the trivalent trans-tetraio(cid:173)
`dodicarbonylrhodate anion, [Rh(C0)214]-. Similar results were observed in
`another study with nonanoic acid as solvent.
`The tetraiodocarbonylrhodate anion, [Rh(CO)l4]-, is the favored triva(cid:173)
`lent rhodium component in coordinating solvents at low carbon monoxide
`pressure and at high hydrogen iodide and low water levels. In a separate
`study 0. 1 M rhodium solutions were treated with carbon monoxide pressure
`at IOO°C and 45 psig in water-hydrogen iodide-acetic acid mixtures until
`steady state conditions were obtained. The monocarbonyl tetraiodide anion
`was the only trivalent rhodium component detected under these conditions.
`The relative amounts of monovalent and trivalent rhodium carbonyl iodide
`anions produced under steady state conditions were determined by infrared
`spectra of the solutions. It was observed that higher concentrations of iodide
`and lower water levels produced more of the trivalent rhodium complex in
`the steady state. When the total iodide concentration remains constant, a
`higher hydrogen iodide fraction leads to a greater proportion of the trivalent
`complex, whereas a larger fraction of the iodide in the sodium salt form
`produces more monovalent rhodium in the steady state, as seen in Table III.
`
`TABLE Ill
`
`Effect of Water and Iodide on the Rhodium Oxidation State
`(IOO°C, 0.1 M Rh, 45 psig total pressure)
`
`Total rhodium in Rh(lll) state(%)
`
`Iodide concentration
`
`1.65 MH,O
`
`5.7 MH,O
`
`7.5 MH,O
`
`0.3MHI
`0.5MHI
`0.7MHI
`IMHI
`IMNal
`0.3 MH I +0.7 MNal
`0.5 M HI+ 0.5 MNal
`
`II
`
`100
`
`0
`0
`36
`100
`0
`IO
`50
`
`0
`0
`14
`40
`
`000014
`
`

`
`10 Methanol Carbonylation to Acetic Acid
`
`287
`
`The distribution between the mono- and trivalent rhodium complexes
`under steady state conditions is dependent on the relative rates of oxidation
`and reduction steps. As the rate of oxidation ofrhodium(l) increases relative
`to the rate of reduction of rhodium(III), the ratio of rhodium(!) to rho(cid:173)
`dium(III) should decrease under steady state conditions.
`James and Rempel [23, 24] described their studies on the reduction of
`trivalent rhodium with carbon monoxide and water in aqueous and non(cid:173)
`aqueous solutions to produce a monovalent dichlorodicarbonylrhodate
`anion. The reaction rate is retarded by higher acid concentrations. They
`suggested that the inverse acid dependence indicates that the rate-determin(cid:173)
`ing step is the reaction of carbon monoxide with a hydroxy complex in
`equilibrium with hydrated rhodium chloride. Hydroxide migration could
`form a transient carboxylate ligand which decarboxylates to produce a
`monovalent rhodium ion. Their data did not allow them to determine
`whether prior coordination of hydroxide ion or water to a trivalent rho(cid:173)
`dium(III) carbonyl species was necessary for the reaction. A similar mecha(cid:173)
`nism for the water gas reaction catalyzed by rhodium carbonyl iodide
`complexes has been suggested by Baker et al. [ 17] and by Cheng et al. [ 18] at
`the University of Rochester. Based on several studies (and references there(cid:173)
`in), the reactions of metal carbonyls with water to generate hydroxycarbonyl
`species can be interpreted as involving the attack of hydroxide ion on
`coordinated carbon monoxide without prior coordination of the nucleo(cid:173)
`phile [25].
`A similar mechanism consistent with the kinetic studies at Monsanto's
`laboratories with the acetic acid-iodide-water system is suggested for the
`step involving reduction of the rhodium(III) anion to the rhodium(!) anion.
`This reaction should have an inverse rate dependence on acidity. This
`mechanism is presented in Fig. 4.
`
`co
`I+-x-
`I Rh I
`co
`I+I~I-j-I
`co ~H2o cooH
`I Rh I ~co
`H20\: I+-I•-
`I -
`I
`/ Rh/ +H +
`
`I+I
`co-
`I -
`/
`I
`co
`I
`I~CO~ I -I 2
`I Rh I +H++ I - +CO ,
`+ _
`I
`I-CO
`Fig. 4. Proposed mechanism of the reduction step of the rhodium-catalyzed water gas shift
`reaction .
`
`-
`
`000015
`
`

`
`288
`
`R. T. Eby and T. C. Singleton
`
`I
`I--co- HI
`I+co(cid:173)
`I Rh I~
`/ Rh I
`I -- co -HI
`I-+H
`co
`H
`CO
`I
`I+I-
`I+co-
`HI
`/ Rh I+ H2 - - - I Rh I + co
`
`I~-I
`I~-H
`Fig. S. Proposed mechanism of the oxidation step of the rhodium-catalyzed water gas shift
`reaction.
`
`A suggested mechanism for the oxidation of rhodium(!) to rhodium(III)
`involves the attack of an acidic species on the [Rh(C0)2IX anion, followed
`by elimination of carbon monoxide and hydrogen. This oxidation of rho(cid:173)
`dium(I) should be enhanced by higher acidity and lower carbon monoxide
`partial pressure. This mechanism is given in Fig. 5.
`The structures of the transient complexes in the oxidation and reduction
`steps are speculative. The experimental data obtained do not definitely
`establish the nature of the transient complexes.
`Under steady state conditions the rates of the two reactions are equiva(cid:173)
`lent. If this proposed mechanism is valid, the overall rate of the water gas
`reaction ca