|
1. |
Kinetic and spectroscopic studies of the carbonylation of methanol with an iodide-promoted iridium catalyst |
|
Dalton Transactions,
Volume 1,
Issue 11,
1979,
Page 1639-1645
Denis Forster,
Preview
|
PDF (821KB)
|
|
摘要:
1979 1639Kinetic and Spectroscopic Studies of the Carbonylation of Methanol withan lodide-promoted Iridium CatalystBy Denis Forster, Corporate Research Laboratories, Monsanto Company, 800 N. Lindbergh Boulevard, St. Louis,Missouri 631 66, U.S.A.A mechanistic interpretation is proposed for the carbonylation of methanol in the presence of iridium halides as thecatalyst precursor and methyl iodide as promoter. The scheme is based on a combination of kinetic observations,in situ spectral studies of reactions, and studies of the chemistry of the observed intermediates. It is shown thatthere are two principal catalytic cycles for the methanol carbonylation, one involving neutral iridium carbonyl iodidecomplexes and the other involving anionic iridium species. In addition, a competitive water-gas shift reaction isobserved under many conditions. The catalytic cycle which predominates is influenced by iodide-ion level andhydrogen iodide concentration.These two factors are influenced by variables such as water, methanol, and methyliodide concentrations as well as salt additives and temperature. The kinetic dependencies for the reaction aredramatically dependent on the catalytic cycle which predominates.THE carbonylation of methanol to acetic acid using ahomogeneous iodide-promoted rhodium catalyst systemis now conducted on a large scale commercia1ly.l Agood understanding of the basic metal-complex chemistryinvolved in the process has been obtained.1,2 I t isknown that iridium compounds will also act as catalystsfor this reaction3 with comparable reactivity to thatdisplayed by the rhodium-containing ~ y s t e m .~ * ~ How-ever, kinetic observations 495a on the iriclium-catalysedreaction have given indications of considerably greatermechanistic complexity than displayed by the rhodiumcatalyst.I report here an attempt to gain some mechanisticinsight into the iridium-catalysed reaction by using acombination of techniques.EXPERIMENTALThe preparations of the following compounds have beendescribed elsewhere : [AsPh,] [Ir(CO)212] ,6 [AsPh,] [Ir-(CO) J,], [AsPh,] [IrMe(CO)21,],B [AsPh,] [Ir(OCMe) (CO),-C13] and [IrH(CO)213]-.8(Cyclo-octa- 1,5-diene)iridium(1) Iodide.-This compoundwas prepared in variable purity by the technique describedfor the analogous chloro-complex with the modification ofusing 5034 more cyclo-octadiene (Found: C, 22.85; H,2.85; I, 29.25.Calc. for C8H1411r: C, 22.5; H, 2.85; I,29.7%).Tricarbonylioctoiridiu~~ ( I) .-This compound was origin-ally described by Hieber et aZ.1° but no spectroscopiccharacterization has been presented. We find that car-bonylation of [Ir(cod) I] in a variety of non-co-ordinatingsolvents produces a soluble brown species with CO stretchingfrequencies a t 2 074 and 2 046 cm-l. This compoundreacts immediately with quaternary ammonium iodidesalts to give the [Ir(CO)212]- ion and also with iodine togive [ Ir (CO) 313]. l19 l2The solutions appear to be somewhat unstable in theabsence of an elevated pressure of CO and a pure solid ofcomposition Ir(CO),I was not isolated from solution.Methanol-carbonylation Reactions.-Experiments werecarried out a t constant pressure in a Hastelloy C.Magne-drive autoclave (300 cm3). The temperature of the reactionwas measured with an internal thermocouple and regulatedwith a West Gardsman controller. The progress of thet Throughout this paper: 1 atm = 101 325 Pa.reaction was followed by automatically recording thepressure drop in a high-pressure reservoir employing aFoxboro model 40 pressure recorder.Typical operating procedure. The iridium precursor(usually IrCl3*4H2O), iodide promoter (usually MeI), andsolvent (usually nonanoic acid or methyl acetate) werecharged to the autoclave and heated to the desired temper-ature under ca.15 atm t of carbon monoxide with stirring.After the reaction temperature had been reached, theautoclave was charged with the reactant (MeOH) from apressurized charging bomb. The reactor pressure wasbrought up to the desired value with additional carbonmonoxide.Spectroscopic Studies.-Infrared spectral studies of thereaction solutions under operating conditions were per-formed using the high-pressure high-temperature celldescribed previ0us1y.l~ Reaction solutions were fed to thecell under pressure and the cell was maintained a t therelevant reaction temperature and pressure while the spec-trum was recorded. Reaction conditions were chosen suchthat the rate of reaction was relatively slow so that COdiffusion would not become the rate-limiting process in thetime taken to record the spectrum (< 1 min) in the non-agitated reaction medium inside the spectrophotometriccell.Nonanoic acid was chosen as the reaction solventbecause of its low absorption in the metal carbonyl regionof the infrared.Low-pressure Carbonylation Reactions.-Reactions con-ducted under carbon monoxide pressures of < 8 atm werecarried out in Fischer-Porter aerosol compatibility tubeswith magnetic stirring.A nalyses.-Reaction solutions were examined by a gaschromatographic (g.c.) technique using a stainless-steelcolumn (10 f t x 0.125 in) packed with OV101. Gasanalyses were obtained by a g . c . technique using dualcolumns of Carbowax 400-modified Porasil A and 13Xmolecular sieves.RESULTS AND DISCUSSIONPreliminary investigations showed that a variety ofspecies could be detected in the reaction media and thepredominant form depended upon the reaction para-meters chosen. It also became apparent that quitedifferent kinetic dependencies were manifested as theconditions were varied.It is therefore convenient tobreak down the discussion of the process into thes1640 J.C.S. Daltonrecognizably different regimes, although in practicesharp break points are not observed.Regime I : The Neutral Iridium Complex CatalyticCycle.-Experiments performed with methyl iodideconcentrations of (0.15 mol dm-, at an iridium con-centration of 0.01 mol dm-,, in media in which low levelsof water and ionic iodide are expected (eg. methylacetate with water pumped into the reactor such thatthe water level was maintained at ca.3 wt. yo),* arecharacterized by kinetics in which an inhibiting effectof carbon monoxide pressure is observed (see Table 1).Infrared spectra of reacting solutions taken under theseconditions show strong metal-carbonyl stretching fre-quencies at 2 073 and 2 046 cm-l (see Figure 1). If theI I I I 1 I I2200 2100 2000 2200 2100 2000 1900T 1 cm-'FIGURE 1 Infrared spectra of reacting solution a t low methyliodide levels: (a) observed a t 160 "C and 33 atm pressure (COpartial pressure ca. 30 atm) for a reaction mixture consistingof 5 wt.% methanol and 2 wt. yo water in nonanoic acid with[Ir] = 1.0 x mol dm-3;(b) of solution (a) after cooling and releasing the carbonmonoxide pressure, followed by treatment with tetraheptyl-ammonium iodidemol dm-3 and [MeI] = 4.0 xsamples from the autoclave are rapidly cooled, thespecies is stable in solution for some time under ambientconditions and when treated with quaternary ammoniumiodides gives [Ir(CO),I.j- and when treated with iodinegives [Ir(CO),I,].The species observed under reactionconditions is therefore [Ir(CO),I].I have independently synthesized [Ir(CO),I] andstudied some of its reactions relevant to the methanol-carbonylation reaction. Thus a solution generated bythe action of carbon monoxide on [Ir(cod)I] in CH,Cl,reacts slowly (several hours at room temperature) with alarge excess of methyl iodide to give two new species.The principal product has two strong carbonyl bands at2 118 and 2 077 cm-l and is immediately transformed byadded iodide ion to [IrMe(CO),I,]- without gas evolu-tion and is therefore formulated as [IrMe(CO),I,].ThisSimilar kinetic behaviour and spectroscopic observationswere noted in media with ca. 10% methanol but no added waterwith Me1 : Ir ratios as high as 60 : 1.compound may be dimerized as is the correspondingchloro-complex ; l4 however, the bridge between eachcomponent of the dimer must be very labile in view ofthe very rapid reaction with iodide ion. The lesser com-ponent in the reaction mixture is a species with carbonyl-stretching frequencies a t 2 125, 2 090, and 1710 cm-land is tentatively identified as [(Ir(OCMe)(CO),I,},] (n =1 or 2) because of its immediate transformation into the[Ir(OCMe) (CO),I,]- ion by treatment with iodide ion.The ra.te of reaction of [Ir(CO),I] with Me1 is far slowerthan the corresponding reaction with [Ir(CO),I,]- and inview of the product observed may involve prior lossof carbon monoxide before reaction.This reactivitydifference between the neutral and anionic iridium(1)halocarbonyl species is to be expected if the reactionpathway follows the anticipated nucleophilic attack ofthe metal species on the alkyl halide.l5?l6The reaction between methyl iodide and [Ir(CO),I] (inCH,Cl, solvent) was also conducted under low pressures(3 atm) of carbon monoxide. At room temperature, avery slow reaction occurred and after ca. 19 h the solespecies present had i.r.stretching frequencies at 2 176w,2 116vs, and 1712ms cm-l and this compound wasimmediately transformed into [Ir(OCMe) (CO),I,]- byaddition of iodide ion. The species formed under pres-sure is tentatively identified as [Ir(OCMe) (C0)312j.When the reaction between Me1 and [Ir(CO),I] iscarried out at 100 "C and 3 atm of carbon monoxide, inaddition to the iridium species tentatively identified as[Ir(OCMe)(CO),I,] above, a strong band at 1 805 cm-l isobserved for the solutions. The species responsible forthis band was identified as acetyl iodide both by theposition of the carbonyl band and by its immediatereaction with methanol to give methyl acetate. It isthus apparent that a catalytic methanol-carbonylationcycle is possible when predominantly neutral iridiumspecies are present in the medium.The strong inhibitingeffect of carbon monoxide evident when [Ir(CO),I] is theprincipal iridium species present in the catalytic reactionis probably caused by the existence of an equilibriumbetween the very weakly nucleophilic [Ir(CO),I] and amore reactive dicarbonyl species.e.g. [Ir(CO),I] + [(Ir(CO),I}]n (n = 1 or 2) + CO[I~(CO),II I-_ [I~(co),I,I- + co[Ir(CO),I] + Me1 + H,O +[Ir(CO),I] + LH[Ir(CO),I,] + MeOH + CO[Ir(CO),I(L)] + CO2 [Ir(CO),I] + 2L -----L [I~I(CO)~L~][II-I(C~)~I,] + 2 COwhere L = oxy-donor such as methanol or waterAn extensive effort to distinguish among the variouspossibilities has not been made although it is noted thatthe [Ir(CO),I,]- ion in MeNO, was not measurablyconverted into [Ir(CO),I] under 35 atm of carbon mon-oxide when observed in the high-pressure spectrophoto-metric cell at room temperature1979The kinetic dependence on methyl iodide concen-tration for this regime was not determined since theobservation of [Ir(CO),I] as the principal iridium speciesis only found over a very narrow range of methyl iodideconcentration. In addition, as the water or methanollevel is raised at low I- : I r ratios, a substantial amountof the products of the water-gas shift reaction (i.e.CO + H,O CO, + H,) is noted in gas samples andan i.r.spectrum taken under reaction conditions (Figure2) shows that the principal form of iridium has changedto a species with a strong CO stretching frequency at2 099 cm-l.Samples containing this species are dark redand when treated with iodide ion,[IrH(CO),I,]- is gen-erated. A more extensive description of the identifi-la) Ibl1641-uu2200 2100 2000 1900 2200 2100 2000 2200 2100 2000v/crn-'FIGURE 3 Infrared spectra of a reaction solution containingadded salt: (a) observed at 160 "C and 30 atm partial pressureof carbon monoxide, run 6 from Table 1; (b) of run 7 underreaction condtions; (c) of run 8 under reaction conditionsa considerably faster rate than the water-gas shiftreaction while the hydridoiridium(II1) species was theprincipal iridium species in solution. A similar situationexists in the anionic catalytic cycle described in a sub-sequent section. Discussion of this phenomenon will bereserved for this later section where more observations onthe equilibria involving oxidative addition of HI toiridium(1) species will be presented.Regime I I : The Anionic Iridium Complex CatalyticCycle.-In reactions conducted with the addition of low -- 2200 2100 2000 1900 2200 2100 2000 190071crn-lFIGURE 2 Infrared spectra of reacting solution with lowmethyl iodide and water levels: (a) observed a t 150 "C and 33atm total pressure (CO partial pressure ca.30 atm) for thereaction mixture of run number 4 from Table 1; (b) of thesolution ( a ) after cooling and releasing the carbon monoxidepressure, followed immediately by treatment with tetraheptyl-ammonium iodidecation of this species will be given elsewhere l2 and itsuffices to note that the species is formulated as [IrH-(CO),I,(OH,)] and can be produced by oxidative additionof H I to [Ir(CO),I] in the presence of low levels of water.This species is a logical intermediate in a water-gas shiftcycle.i.e.IrI + HI + HIrIIII (1) 'p (2) HIrIIII + HI @ Ir11112 + H, -- 2200 2100 2000 2200 2100 2000 1900FIGURE 4 Infrared spectra of a reaction solution in regime 111:(b) ofIrIrrI2 + CO + H,O - IrI + CO, + 2HI (3)The surprising feature about the observation of thisspecies under reaction conditions (Figure 3) and thedetection of co, and H2 in the gas phase is that themet hanol-carbonylation reaction was proceeding atFJcrn-1(a) under reaction conditions of run 12 from Table 1;solution (a) after cooling and releasing the pressur1642 J.C.S.DaltonTABLE 1Carbonylation of methanol with iodide-promoted iridium catalystConcentration/ Rate of Water-gas- _ Concentration/ mol dm-3 MeOH shiftLO mol dm-3 r -A- carbonyl- rate e 0, pressure r---.--~-, Initial Charged Charged ation mol dm-3Experiment "c atm I r Me1 Solvent Methanol water I-Reactions where the predominant iridium species was [Ir(CO),I]1 175 23 0.01 0.01 Methyl 2.22 175 50 0.01 0.10 Methyl 2.23 175 100 0.01 0.10 Methyl 2.2acetateacetateacetateReactions where the predominant iridium species was [IrH(CO),I,(OH,)]4 150 30 0.02 0.085 160 30 0.02 0.08Reactions where the predominant iridium species6 160 30 0.02 1 .o7 160 65 0.02 1 .o8 160 100 0.02 1.09 160 65 0.02 1 .o10 175 20 0.02 0.211 175 50 0.02 0.2Nonanoic 2.47acidNonanoic 2.47acidwas [IrMe(CO),I,]-Nonanoic 2.47acidNonanoic 2.47acidNonanoic 2.47acidNonanoic 2.47acidMethyl 16acetateMethyl 16acetateReactions where the predominant iridium species was [IrH(CO) ,I3]-12 160 30 0.01 0.10 Nonanoic 1.213 160 30 0.01 0.20 Nonanoic 1.214 160 30 0.01 0.40 Nonanoic 1.215 160 50 0.01 0.40 Nonanoic 1.216 160 30 0.01 0.60 Nonanoic 1.217 160 30 0.01 0.80 Nonanoic 1.218 160 30 0.01 1.0 Nonanoic 1.219 160 30 0.01 1.20 Nonanoic 1.2acidacidacidacidacidacidacidacid1.71.70.100.100.100.040.040.041.11.11.11.11.11.11.11.12.71.7 ca.0.03(0.2 Very low0.6 ca. 0.52.2Very slowca. 0.060.100.162.28.0ca.0.15 ca. 0.040.8 ca. 0.121.1 ca. 0.121.05 ca. 0.121.3 ca. 0.171.7 ca. 0.31.3 ca. 0.31.2 ca. 0.3CommentsSmallselectivityloss to CH,4% of MeOHconvertedinto CH,10% ofMeOHconvertedinto CH,a Iodide added as [N(C,H,,),]I- or LiI. Rate for initial 20% methanol conversion. Determined by rate of gas uptakelevels of salts (quaternary ammonium iodide or lithium ative substitution mechanism commonly found foriodide) the spectra of reacting solutions contain pre- octahedral d6 metal species.*dominantly [IrMe(CO),I,]- (see Figure 4). When thisiridium species predominates the reaction rate increaseswith increasing carbon monoxide pressure (see Table 1,experiments 6-1 1) and decreases with increasing ioniciodide level.These results are compatible with the rate-determining step in a catalytic cycle being formation of atricarbonyl(methyl)iridium(Irr) species by the dissoci-* I favour an equilibrium limitation on the overall rate oftransformation to acyl species as indicated, not only because ofthe observed kinetic dependencies but because there is goodreason to believe that the [IrMe(CO),I,]- ion is highly labile a ttemperatures >lo0 "C. Thus the halide positions trans to thecarbonyls in [Ir(CO),X,]- are very labile even a t room temper-ature (D. Forster, Inorg. Chem., 1972, 11, 1686) and this togetherwith the high trans-labilizing effect of the methyl group shouldensure that all of the iodide positions are labile.i.e. [IrMe(CO),I,]- "- [IrMe(CO),I,] + I-I C0 .f[IrMe (CO),I,IIn agreement with this, [IrMe(CO),I,]- can be convertedinto [Ir(OCMe)(CO),I,]- by treatment with 5 atm ofCO at 80 "C (4 h) in chlorobenzene.This transformationcan be completely prevented by the addition of 5 mol of[NBu,]I per mol of iridium. In addition, the [IrMe-(CO),I,]- ion shows no tendency to isomerize to anacetyl species even on heating to 150 "C. This is inmarked contrast with the corresponding rhodium syste1979where an acetyl species is formed from the reaction of TABLE 2methyl iodide with the [Rh(CO),I,]- ion in the absenceof carbon monoxide and with the species [RhMe(CO)213]-being an unstable intermediate.2 Apparently, a tri-carbonyl(methyl)iridium(III) species must form in orderto induce alkyl migration in this system.In support of this concept, it should be noted that[{ IrMe(CO),Cl,),] is prepared by oxidative addition ofacetyl chloride to an iridium(1) monocarbonyl species,14where it appears that alkyl migration occurs on anunstable intermediate, an iridium(II1) acyl monocar-bony1 complex.In attempting to define the elimination step in thecatalytic cycle, I studied the decomposition of [Ir(OC-Me)(CO),Cl,] - (ref.8) in chlorobenzene and nitrobenzene,both with and without added methanol. In the absenceof methanol, acetyl chloride and [Ir(CO),CI,]- were thedecomposition products and with methanol presentmethyl acetate was formed. Methanol did not appearto affect the initial rate of formation of the [Ir(CO),-Cld- ion (ti ca.15 min in chlorobenzene at 35 "C). It isconcluded that solvolysis of the acetyl complex is notrequired to regenerate iridium(1).Formation of methane and CO, becomes a verysignificant competing reaction under these carbonyl-ation-reaction conditions, particularly at low Me1 : Irratios and low carbon monoxide pressure. In sever81reactions no detectable hydrogen was produced althoughlarge amounts of methane were formed, implying amechanism for cleavage of the methyliridium speciesinvolving either attack by a proton or a separate hydrido-iridium species.Regime 111: Competition between the Water-GasShift Reaction and the Methanol-carbonylation Reaction.-Carbonylation reactions conducted at methyl iodideconcentrations greater than 0.1 mol dmP3 at [Ir] = 0.01mol dm-, and with significant concentrations of waterand/or methanol exhibit a different type of kinetic andspectroscopic behaviour. This is the regime wheremost of the previously published work 4 9 5 a was conductedand is characterized by an apparent first-order depen-dence on methanol concentration and an independence ofcarbon monoxide pressure (see Table 1 and refs.4 and 5 ) .Here, an optimum rate of reaction is found with respectto the methyl iodide concentration (see Table 1, experi-ments 12-19). Spectroscopically, a species with strongCO stretching bonds at 2 102 and 2 050 cm-l is observed(Figure 4). This species dominates the spectra ofreacting solutions over the whole range of MeI: I rratios. It is unchanged on cooling and releasing the COpressure and it is not immediately changed on additionof ionic iodide, although it gradually transforms to amixture of [Ir(CO),I,]- with some [Ir(CO),I,]-.Fur-ther, by treating a reaction solution (Me1 : Ir = 40 : l )with PPh, a mixture of phosphine complexes precipi-* This salt may contain the [PMePh3]+ cation formed byquaternization of some of the triphenylphosphine by the methyliodide present. The finding of the [Ir(CO),IJ-- ion implies thatthis species is present in the reaction solution but is spectrallymasked by the much higher concentration of [IrH(CO)213]--.Infrared spectra of iridium carbonyl speciesrelevant to the methanol-carbonylation reaction[ Ir Me (CO) J 3] -Colour SolventYellow nonanoic acidnitromethaneOrange nonanoic acidnitromethanePale nonanoic acidorangenitrome thaneYellow nonanoic acidnitromethaneCarbonyl-stretchingvibrations(cm-1) b2 048, 19702 051, 19732 110, 2 0702 115, 2 071ca.2 160w,br2 101, 2 051ca. 2 160w,bre2 107, 2 0562 096, 2 0462 102. 2 050[Ir( OCMe) - Orange nonanoic acid 2 108; 2 062chlorobenzene 2 110, 2 062,(co) 2131-1 685Dark nonanoic acid 2 073, 2 046mbrownorange 2 170vw,[Ir(CO) 311[Ir(co)&l Pale benzene[{IrMe(CO) 21z}n] Orange benzene 2 116, 2 069methylene 2 118, 2 077[{ Ir(0CMe) - Orange benzene 2 125, 2 090,2 186w,2 132(n = 1 or 2)chloride(CO),T,)n1 1710m(n = 1 or 2)[Ir (OCMe) - Orange methylene 2 176w,KO) 3 1 2 1 chloride 2 116, 1712m2 155w, [IrH(CO),I,- Blood red nonanoic acid(OH,)] 2 098vscation.The anionic species were studied as salts of the [AsPh,]+All bands were very strong unless otherwise noted.Ir-H stretch.tates.The i.r. spectrum of this mixture shows it tocontain primarily [IrH(CO) (PPh,),12] l 7 with a character-istic Ir-H stretch at ca. 2 215 cm-l and [Ir(CO),(PPh,),I].A minor amount of a salt of the [Ir(CO),I,]- ion * is alsopresent. The predominant species in the reactionsolutions thus appears to be the [IrH(CO),I,]- ion. Thereaction of PPh, with this ion, synthesized in MeNO, byreaction of a salt of [1r(C0),I2]- with aqueous HI, wasstudied and it was found that the reaction gave exclu-sively [IrH(CO)(PPh,),I,] when the reaction was conduc-ted with a large excess of HI, but a mixture of [Ir(CO),-(PPh,),I] together with the hydride complex when thesolutions contained little or no excess of HI.? I t can beinferred from these experiments that the carbonylation-reaction solutions contain predominantly [IrH(C0)213]-and that little or no excess of HI is present.At highmethyl iodide concentrations where the rate begins toslow down, significant amounts of [Ir(CO),I,]- areobserved in the spectra of the reacting solutions.The water-gas shift reaction is also proceeding con-currently in these reactions although the rate is sub-stantially slower than the methanol carbonylation rate(typically about a fifth of the methanol carbonylationrate). This slower rate appears unusual in view of thet This is probably because of the rapidly established equilib-rium between [II-(CO)~I,J- and [IrH(CO),T,]- which allows thephosphine to attack the more labile iridium@) species1644 J.C.S.Daltonobservation of a hydridoiridium(Ir1) species which wouldbe a logical first step in a water-gas shift cycle.e.g. [Ir(CO),I,]- + HI [IrH(CO),I,]- (4)[IrH(CO),I,]- + H I T- [Ir(CO),IJ + H, (5)[Ir(CO),I,I- + co, + 2 HI (6)[Ir(CO),I,]- + H,O + CO -Thus in order to accommodate the observation ofdominance of the [IrH(CO),I,]- ion with the observedpredominance of the methanol-carbonylation reactionin the system we have to propose that: ( a ) reaction (4)is readily reversible and probably occurs faster thanmethyl iodide addition ; ( b ) reaction (5) is relatively slow ;and (c) when methyl iodide addition occurs it is effect-ively non-reversible.In order to test these hypothesesseveral experiments were performed at ambient condi-tions. First, [Ir(CO),I,]- was allowed to react with asmall amount of aqueous hydriodic acid in nitromethanewhereupon the [IrH(CO),I,]- ion forms. The equili-brium for this reaction is very favorable at room temper-ature ( K > 50 dm3 mol-l). A small amount of solidcalcium carbonate when added to this solution immedi-ately reverses the addition reaction and [Ir(CO),I,]-reforms.The oxidative-addition reactions of methyl iodide andhydrogen iodide with [Ir(CO),I,] - were performed undercompetitive conditions. We found that if [IrH(CO),I,]-,prepared by the addition of slightly more than 1 mol ofHI to [Ir(CO),I,]- in MeNO,, was treated with a largeexcess of MeI, the hydride was slowly transformed to[1rMe(CO),I3]- over 60 min.I n . another experiment[Ir(CO),I,]- was treated with a mixture of HI and MeI,each in a ten-fold excess. The [IrH(CO),I,]- ion wasthe only species detected initially and over a period of50 min it was transformed into [Ir(CO),I,]-. Thecomparative stability of the [IrMe(CO),I,]- ion wasconfirmed in a separate experiment in which a MeNO,solution of a salt of this ion was heated to 100 "C in anopen system (where Me1 could escape) and no decomposi-tion was evident after 10 min.Thus the oxidative addition of H I is apparently muchfaster than that of methyl iodide, but the reversibility ofthe H I addition combined with the relative non-reversi-bility of the Me1 addition means that [IrMe(CO),I,]-can form in situations where there is little or no excess ofHI.Obviously, if excess of HI is present, methyl iodidecannot compete, [Ir(CO),14]- forms, and under carbonyl-ation-reaction conditions the water-gas shift reactionwould predominate.I therefore propose that in this regime of the carbonyl-ation reaction the rate-determining step is methyl iodideaddition to the low level of [Ir(CO),I,]- which is inequilibrium with the [IrH(CO),I,]- ion. The reactionrate should therefore show a positive dependence onmethyl iodide level although the equilibrium Me1 +H,O + MeOH + HI means that varying the methyliodide independent of HI will be difficult in a practicalsituation, where ester formation also influences themethanol, water, and hence HI and methyl iodide levels.In my limited variation of the methyl iodide concen-tration (Table 1) it can be seen that there is an initialsharp drop in reaction rate which occurs as the catalystsystem changes over from primarily iridium(1) species tothe [IrH(CO),I,]- ion and then the rate of the reactionincreases with increasing Me1 concentration up to aconcentration of 0.80 mol dm-3 at [Ir] = 0.01 mol dm-3,and then starts to decrease again.The dependence onmethyl iodide concentration is somewhat less than firstorder over the range where the rate is showing a positivedependence on Me1 concentration.A detailed understanding of this kinetic phenomenonwill require a knowledge of the various equilibriumconstants under actual reaction conditions.However,it would appear that increasing methyl iodide wouldincrease the HI level proportionately if the organicequilibria alone were controlling and hence that nopositive dependence on methyl iodide should be observed.I tentatively propose that the amount of [IrH(CO),I,]-present in the reacting solutions is not primarily areflection of the amount of free HI involved in theorganic equilibria alone, but rather reflects the rapid rateand large equilibrium constant of the reaction HI +[Ir(CO),I,] - + [IrH(CO),I,] -.The experiments referred to earlier in which tri-phenylphosphine was added to the reaction solutionswere interpreted as indicating little or no excess of HIin the solutions and experiments in Monsanto's TexasCity laboratories on the equilibria involved in themethanol-acetic acid-water-hydriodic acid system sug-gest that the level of free H I in the experiments describedherein should be very small.Thus, if there is little or no excess of HI present in thereaction solutions, increasing the methyl iodide con-centration would be expected to increase the reactionrate provided that there is sufficient methanol present toact as a ' sink ' for the HI.Continued increases in themethyl iodide level will eventually lead to appreciablelevels of HI being produced if the methanol level is notraised simultaneously. The reaction rate will not thenincrease with increasing H I level but will start to de-crease as the sequential reaction of the hydride withH I to give [Ir(CO),I,]- starts to become appreciable.As noted earlier, in the carbonylation experiments at thehighest Me1 : I r ratio examined, the [Ir(CO),I,]- ion * wasclearly observed in the reacting solutions.A scheme which summarizes the observations on thecatalytic reaction is given below.I t is apparent thatboth hydriodic acid and ionic iodide concentrations canplay very significant roles in controlling not only thecatalytic cycle in which the iridium complexes operatebut the actual reaction which predominates, i.e. methanolcarbonylation versus water-gas shift versus methanolto methane. The hydriodic acid and ionic iodide levelsare affected by methyl iodide, water, methanol, andmethyl acetate concentrations all of which are not onlyinfluenced by the initial charge to the catalytic reactionbut also vary markedly during a batch reaction.Thus, 1979 1645\co 1-4co[IrMe( CO 131 21I'Organic reactionsHI+ MeOH = H20 + Me1MeCOzMe MeCOIorMeC02H H20 Or MeOHSCHEME Catalytic cycles for the iridium-catalysed carbonylation of methanoldetailed mechanistic interpretation of this system willrequire a considerably sharper definition of the reactionparameters than is required for the rhodium system.I acknowledge the outstanding experimental assistance ofMr. R. G. Smith. Many helpful discussions were held withDr. A. Hershman, Dr. D. E. Morris, and Professor J.Halpern.[8/1490 Received, 14th August, 19781REFERENCESChem. Tech., 1971, 600.J. F. Roth, J. H. Craddock, A. Hershman, and F. E. Paulik,F. E. Paulikand J. F. Roth, Chem. Comm., 1968, 1578.a D. Forster, J . Amer. Chem. SOC., 1976, 98, 846..A D. Brodzki, B. Denise, and G. Pannetier, J . Mol. Cut., 1977,2, 149.5 (a) T. Matsumoto, T. Mjzoroki, and A. Ozaki, J . Catalysis,1978, 51, 96; (b) A. Hershman and J. H. Craddock, personalcommunication.D. Forster, Inorg. Nuclear Chem. Letters, 1969, 5, 433.7 D. Forster, Synth. Inorg. Metal-Org. Chem., 1971, 1, 221.D. Forster, Inorg. Chem., 1972, 11, 473.G. Winkhaus and H. Singer, Ber., 1966, 99, 3610.lo W. Hieber, H. Logally, and A. Mayr, 2. anorg. Chem., 1941,11 L. Malatesta, L. Naldini, and F. Cariati, J . Chem. SOC., 1964,12 D. Forster, A. Hershman, and D. E. Morris, unpublishedl3 D. E. Morris and H. B. Tinker, Chem. Tech., 1972, 555;l4 N. A. Bailey, C. J . Jones, B. L. Shaw, and E. Singleton,l5 P. R. Chock and J. Halpern, J . Amer. Chem. Soc., 1966, 88,l6 D. Forster, J . Amer. Chem. SOC., 1975, 97, 951.l7 D. M. RlakeandM. Kubota, Inorg. Chem., 1970, 9. 989.246, 138.961.work.H. B. Tinker and D. E. Morris, Rev. Sci. Inst., 1972, 45, 1024.Chem. Comm., 1967, 1051.3511
ISSN:1477-9226
DOI:10.1039/DT9790001639
出版商:RSC
年代:1979
数据来源: RSC
|
|