摘要:

J. CHEM. soc. DALTON TRANS. 1982 1069Catalysed and Non-catalysed Reaction Between [ Fe(CO)5] and IsonitrilesBy Michel 0. Albers and Neil J. Coville,' Department of Chemistry, University of the Witwatersrand, 1 JanEric Singleton, National Chemical Research Laboratory, CSIR, P.O. Box 395, Pretoria 0001, Republic ofSmuts Avenue, Johannesburg 2001, Republic of South AfricaSouth AfricaThe reaction between [Fe(CO),J and isonitrile, RNC, is catalysed by CoCI2*H20 and readily yields the complexes[Fe(CO),-,(CNR),] ( n = 1-3, R = Me, C&11, But, PhCH2, Ph, 2,6-Me2C6H3, or 2,4,6-Me,C,H2; n = 4,R = But; n = 4 or 5, R = Ph, 2,6-Me2C6H3, or 2,4,6-Me3C6H2). The high-yield synthesis of [Fe(CO),(CNR)]from [Fe(CO),] and RNC in the absence of catalyst is also reported. Trimethylamine N-oxide has been used tosynthesize [Fe(C0)3(CNR)2] from [Fe(CO),(CNR)] and RNC and results are compared with the CoCI2 catalysedreaction. All products were characterized by i.r.and n.m.r. spectroscopy. The higher substituted derivatives werefurther characterized by reaction with l2 and tetracyanoethylene (tcne) and gave [Fe(CO),(CNR),(tcne)]( n = 2, R = But, PhCH2, or 2,6-Me2C6H3; n = 3, R = But or 2,6-Me2C6H3; n = 4, R = 2,6-Me2CeH3) andcis- and trans- [Fe(CNC6H3Me2-2,6),I2] from appropriate starting materials. Mechanistic data suggest that thereaction occurs via attack of catalyst at a co-ordinated CO ligand. Subsequent attack by unco-ordinated RNCin an intermolecular, non-bridging mechanism leads to the required isonitrile derivatives.AN extensive chemistry of transit ion-met al-isonit rilecomplexes has been developed and the synthesis of thisclass of compounds has been well reviewed.14 Syn-thetic methods are varied and include the substitutionof metal carbonyl complexes by isonitriles (RNC)Although a viable and attractive route, direct multiplereplacement of CO groups by RNC has in general failedas high temperatures, i.e.those which also induce pro-duct or reactant decomposition, are required to bringabout the CO substitution reaction.This is exemplified by the reaction between [Fe(CO),]and RNC. Attempts to prepare [Fe(CO)Ln(CNR),] fn =1-5) by heating the reactants in the absence of solventhave resulted in the synthesis of only mono- (n = 1) ordi-substituted (n = 2) product^.^^^ Alternative pro-cedures starting from [Fe,(C0)Q],7 [Fe3(CO)1,],s or thephotochemical reaction between [Fe(CO),] and RNChave only led to complexes with n < 2.Similar resultswere obtained when [Fe(CO),] was reacted with iso-nitrile precursors such as isocyanates,l0 phosphine-imines,ll isocyanide d i ~ h l o r i d e s , ~ ~ . ~ ~ and metal bis(tri-methy1~ilyl)amides.~~ Indeed, the only iron(0) complexknown with n > 2 is the complex [Fe(CNR),],. preparedeither via metal-atom techniques or by reducing FeBr,in the presence of RNC.15In this publication we wish to report on our successfulattempts to synthesize [Fe(CO)ti,(CNR),] (n = 1-5)from [Fe(CO),] and RNC in the presence of transition-metal catalysts under mild reaction conditions. Wehave concentrated our efforts on the use of CoCI,*ZH,O ascatalyst because of its low cost, ready availability, andthe facility with which it catalyses multiple CO substi-tution reactions.A preliminary report of this work hasbeen published.16For comparison we have also included our data on thethermal reaction between [Fe(CO),] and RNC in benzeneand the ' reagent induced ' reaction between [Fe(CO),-(CNR)] and RNC in the presence of trimethylamine N -oxide. l 7EXPERIMENTALPentacarbonyliron was purchased from Strem Chemicals.The isonitriles were either purchased from Fluka A.G.(ButNC, C,H,,NC, and 2,6-Me,C,H3NC), Aldrich (PhCH,-NC), or prepared by the literature methods 18v1s (2,4,6-Me,C,H,NC, PhNC, and MeNC) .The catalyst CoC1,*2H,Owas obtained by heating CoCl,*GH,O (Riedel-de-Haen A.G.)in vacuo at ca. 50 "C for 5 h. All reactions were routinelycarried out under an argon atmosphere with dry, degassedsolvents. Merck Kieselgel 60 (60-200 pm) silica gel and150 x 2 cm columns were used throughout.Melting points were determined with a Kofler hot-stageapparatus and are uncorrected. Infrared spectra wererecorded on a Pye Unicam SP 300 spectrophotometer.Hydrogen-1 n.m.r. spectra were recorded on a Brukei WP 80FTNMR spectrometer, and mass spectra were recorded on aVarian CH5 spectrometer (operating at 70 eV).t Micro-analyses were performed by the Microanalytical Labora-tories, C.S.I.R. Magnetic susceptibility measurements weredetermined using a variable-temperature Gouy BalanceSystem (Newport Instruments) fitted with a 10 cm electro-magnet and a Stanton model S.M.12 precision balance.Preparation of [Fe(CO),(CNR)] (R = Me, C,H,,, But, Ph-CH,, Ph, 2,6-Me2C,H,, or 2,4,6-Me3C,H,) (la)-( lg).--Method (A) (thermal reaction). [Fe(CO),] (7.84 g, 40 mmol)and isonitrile (20 mmol) were refluxed for 0.5 h in benzene(60 cm3). The end of the reaction was evidenced by cess-ation of CO evolution and a rapid darkening of the reactionsolution. The cold solution was passed down a short silicacolumn (ca. 50 g) and eluted with benzene. Excess [Fe-(CO),] and solvent were removed on a rotary evaporator togive the products as indicated (Table 1).Method (B) (CoC1, catalysed reaction). [Fe(CO),] (3.92 g,20 mmol), isonitrile (20 mmol), and CoC1,*2H20 (O.OS0 g,0.3 mmol) were combined with benzene (100 cm3) in a250 cm3 round-bottomed flask, and refluxed for the appro-priate time (Table 1).The green reaction solution wasallowed to cool, then treated as in Method (A) above, to givethe products in the recorded yields (Tables 1 andPreparation of [Fe(CO),(CNR)J (R = Me, C6H11, But,t Throughout this paper: 1 eV x 1.602 x J.2)1070 J. CHEM. SOC. DALTON TRANS. 1982PhCH,, Ph, 2,6-Me,C,H3, or 2,4,6-Me3C,H,) (2a)-(Zg).-Method ( A ) (CoCl, catalysed reaction). [Fe(CO),(CNR)J (3mmol), CoC1,*2H20 (0.017 g, 0.1 mmol), and benzene (20cm3) were combined in a reaction flask. Isonitrile (3 mmol)was added to give a green solution, which was then refluxedand benzene (20 cm3) were combined in a reaction flask andthe solution brought to reflux.Addition of isonitrile (4mmol) gave a green solution which eventually turned orangeafter continued reaction in refluxing benzene. The solutionwas allowed to cool and then filtered under nitrogen. TheTABLE 1Experimental details for the thermal and CoCl, catalysed reactions a [Fe(CO),] + RNC - [Fe(CO),(CNR)] + COThermal reaction7 .A rReactionProduct time b (min) Yield (%)Fe(CO),(CNMe)] 30 79 ;:y{ :[E;f-f) 1 30 9130 96:Fe (CO), (CNCH,Ph)] 30 87'Fe(CO),(CNPh)] 30 95.Fe (CO) ,( CNC,H3Me,-2,6) J 30 95IFe(C0) ,(CNC,H,Me,-2.4, 6)] 30 95See Experimental section. As determined by i.r.CoC1, catalysed L reactiontime (min) Yield (yo)10 655 895 855 8810 805 885 88r 7 Reactionspectroscopy.for the appropriate time (Table 3).The reaction solutionwas allowed to cool, then passed down a silica column (ca.20 g) and eluted with benzene. The solvent was removed ona rotary evaporator, t o give the yellow products (Tables 3and 4). (Table 5 ) .Method ( B ) (trimethylamine N-oxide promoted reaction).residue was washed with 10 cm3 portions of dry, degassedbenzene. Solvent was then removed under vacuum and thepure complex recrystallised from toluene-pentane at - 78 "Cto yield an air-sensitive yellow powder (yield 50-60%)Preparation of [Fe(CO)(CNR),] (R = But, Ph, 2,6-Me2-TABLE 2Analytical and spectroscopic data for the complexes [Fe(CO),(CNR)] (la)-( lg)'H N.m.r.d (7) . , Analysis (%) 1.r.e (cm-1) r - vL M.p." > I - - 7 Aro- MassColour ("C) C H N v(NC) V(C0) CH3 CH, matic spectrum b +209 (208.94) (la) f Pale yellow 34-35 34.4 (34.5) 1.45 (1.45) 6.65 (6.7) 2 199 2 060, 1 993, 1 970 8.18(lb) Yellow oil 48.0 (47.65) 4.0 (3.95) 5.45 (5.05) 2 183 2 063, 1 994, 1 967 8.3 277 (277.06)(lc) Yellow 50-52 42.3 (43.05) 3.45 (3.6) 5.5 (5.6) 2 167 2 051, 1 992, 1 968 9.22 251 (251.02)(Id) Yellow 14-16 51.0 (50.55) 2.5 (2.45) 5.2 (4.95) 2 190 2 065, 1 998, 1 970 6.46 3.25 285 (285.04)(le) 6 Yellow 56-57 49.05 (48.7) 1.9 (1.85) 5.25 (5.15) 2 150 2 050, 1995, 1975 3.42 271 (271.01)(If) 5 Yellow 82-83 52.45 (52.2) 3.15 (3.05) 4.7 (4.7) 2 151 2 051, 1 999, 1 975 8.07 3.32 299 (279.09)(lg) Yellow 38-39 53.55 (53.65) 3.65 (3.5) 4.5 (4.45) 2 185 2 060, 2 000, 1 872 k 3.75 313 (313.09)Uncorrected.b Calculated values in parentheses. C Recorded in hexane. Recorded in C6D, relative to SiMe,. e MolecularY. Yamamoto and N. Yamazaki, J. Org. Chem., 1977, 42, 4138. ion. f Ref. 5. g Refs. 9 and 20. Ref. 25. i Refs. 5 and 10.8.13 and 8.11, ratio 2 : 1.[Fe(CO),(CNR)] (3 mmol), trimethylamine N-oxide (0.50 g ,4.5 mmol), and isonitrile (3 mmol) were added to 20 cm3 ofbenzene. The reaction solution was refluxed for theindicated time and purification, as in Method (A) above,gave the required products (Tables 3 and 4).Preparation of [Fe(CO),(CNR),] (R = Me, C,H,,, But,PhCH,, Ph, 2,6-Me,C,H3, or 2,4,6-Me3C,H,) (3a)-( 3g) .-[Fe(CO),(CNR)] (2 mmol), CoC12*2H20 (0.017 g, 0.1 mmol),C,H3, or 2,4,6-Me3C,H,) (4a)-(4d) .-[Fe(CO),(CNR)] (2mmol), CoC1,.2H2O (0.017 g, 0.1 mmol), and benzene (20cm3) were combined in the reaction flask.At reflux,isonitrile (6 mmol) was added and the green solution re-fluxed for the indicated time (Table 5). The cold reactionsolution was treated as above to yield the required air-sensitive products.Preparation of [Fe(CNR),] (R = Ph, 2,6-Me,C,H3, OYTABLE 3Experimental details for I he CoCl, catalysed and trimethylamine N-oxide promoted reactions a [Fe(CO) 4(CNR)] +RNC - [Fe(CO),(CNR),] + COCoC1, catalysed reactionTrimethylamine N-oxide promotedReaction time b (min)reactionAI 7 Yield (%)15 8820 9312 8712 8315 9515 9415 76spectroscopyJ . CHEM. SOC. DALTON TRANS. 1982 1071TABLE 4Analytical and spectroscopic data for the complcxes [Fe(CO),(CNR) ,]1H N.m.r.d (T)Analysis 6 (%) I .r.c (cm-l) r-h- 7 - A .-A- - Aro- MassColour M.p.5 ("C) C(2a) f Yellow 141-144 g 37.65 (37.85)(2b) Yellow oil 52.45 (53.4)(2c) Yellow 98-99 50.2 (51.0)(2d) Yellow 72-73 61.15 (60.95)(2e) Yellow 72-73 59.15 (58.95)(2f) Yellow 132-134 0 62.3 (62.7)(2g) Yellow 134-135 63.15 (63.9)a Uncorrected.Calculated valuesf M.p. 100-130 "C, from ref. 5. ion.1 . H N v(NC) v(C0) CH, CH, matic spectrum"@2.65 (2.7) 12.3 (12.6) 2 160 2 000, 1 922 7.97 222 (221.99)6.0 (6.0) 7.1 (7.3) 2 130 1 995, 1 922 8.52 358 (358.22)5.9 (5.95) 9.1 (9.15) 2 130 1998, 1922 9.08 306 (306.15)4.05 (3.75) 7.75 (7.5) 2 140 1 998, 1 928 6.27 3.22 374 (374.18)2.9 (2.9) 8.2 (8.1) 2 103 1 996, 1 943 3.28 346 (346.13)4.6 (4.5) 6.95 (6.95) 2 108 2 000, 1938 7.92 3.35 402 (402.23)5.05 (5.1) 6.3 (6.5) 2 105 1995, 1935 i 3.53 430 (430.29)in parentheses.Recorded in CHC1,. d Recorded in C,D, relative t o SiMe,. e MolecularWith decomposition, li Ref. 25. i 7.86 and 8.08, intensity ratio 2 : 12,4,6-Me3C,H,) (5a)-(5c).-[Fe(CO),(CNR)] (2 mmol),CoC1,*2H2O (0.017 g, 0.1 mmol), and benzene (20 cm3) wereadded to the reaction flask. Isonitrile (8 mmol) was added,and the solution refluxed for the appropriate time (Table 5 ) .The cold reaction solution was purified as above.Preparation of [Fe(CO),-,(CNR),(tcne)] (n = 2--4; R =But, PhCH,, or 2,6-Me2C,H,) (6a)-(6f) .-The appropriate[Fe(CO),-,(CNR),] (n = 2-4) complex (2 mmol) was pre-I).-CoX2~6H,O (X = C1, Br, or I ; 2.0 mmol) was dissolvedin acetone (40 cm3), with heating to boiling if necessary.2,6-Me2C,H,NC (8.5 mmol) was added, and the mixturestirred for 15 min.The green (or brown, X = I) pre-cipitate was filtered off and washed with several 10 cm3portions of cold diethyl ether. Recrystallisation wasachieved from hot methanol solutions to give the requiredproduct in > 70% yield (Table 8).TABLE 5Reaction times and spectroscopic data for the complexes [Fe(CO),-,(CNR),] (n = 3-5)1H N.m.r.C (T)r > r-+7I.r.6 (cm-1).A ReactionComplex time 5 (min) v(NC) d 4CO) CH, or CH,(3a) [Fe(CO),(CNMe),l 180 2 128, 2 098 (sh) 1923, 1885 7.6235 2 108, 2 090 (sh) 1926, 1884 8.7235 2 092, 2 061 (sh) 1922, 1 883 8.8325 2 134 (sh), 2 109 1934, 1 892 6.06( 3 C ) [Fe(CO) 2(CNBut) 315 S 083, 2 053 (sh) 1947, 1917(3d) [Fe(CO)2(CNCHzPh) 315 2 065, 2 045 (sh) 1940, 1906 7.83 (3f) [Fe(CO) ,(CNC,H3Me,-2, 6) 3](36) [Fe(CO) ,(CNC,H,Me3-2, 4,6) 31 5 2 075, 2 050 (sh) 1940, 1 905 7.73, 8.04(4a) [ Fe (CO) ( CNBut ) 4] 400 2 086, 2 041 1881 8.885 2 055 19105 2 045, 1 990 (sh) 1 903 7.69(4b) [Fe(CO) (CNPh)4I(4c) [Fe(CO) (CNC,H3Me,-2, 6) 4](4d) [ Fe (CO) ( CNC,H,Me3-2, 4,6) dl 5 2 040, 2 000 1899 7.61, 8.01 c(5b) [Fe(CNC,H,Me,-2,6),](3b) [Fe(Co) 2(CNC6H11) 31(3e) [Fe(CO),(CNPh)311010 2 028, 1960, 1920 (sh) 7.612 041, 1 990, 1 917 (sh) ( 5 4 [Fe(CNPh),l(5c) [Fe(CNC,H,Me3-2,4, 6) 5] 10 2 014, 1 989, 1 900 (sh) 7.53, 7.99 @Estimated b y i.r.spectroscopy.' Recorded in C&,. Recorded in C,D, relative t o SiMe,. sh = Shoulder. c Intensityratio 2 : 1.pared in situ as described above. To this warm reactionsolution (ca. 40 "C) was added tetracyanoethylene (tcne)(2.2 mmol) and stirring was continued until CO evolutionhad ceased. The reaction solution was filtered and thesolvent removed on a rotary evaporator. The crude pro-duct was purified by column chromatography (silica gel orneutral alumina, eluant CH,Cl, or CHC1,) or by crystallis-ation (n = 2, benzene; n = 3 or 4, CH,Cl,-hexane mix-tures) to give the required products in typically 40-60%yield (Tables 6 and 7).Preparation of cis- and trans-[Fe(CNC,H,Me,-2,6),I,](6g) and (6h).-Iodine (2.0 mmol), in benzene ( 5 cm3), wasadded dropwise to a warm reaction solution containing[Fe(CO) (CNC,H,Me,-2,6),] (2 mmol).On completion of I,addition the solvent was removed on a rotary evaporator.The solid residue was then chromatographed on silica gel(eluant benzene, followed by CH,CI,) and yielded greentrans-[Fe(CNC,H,Me2-2,6),I,] and finally brown cis-[Fe-(CNC,H3Me,-2,6),I,]. Both materials were recrystallisedfrom CH,Cl,-hexane (combined yield, 80%) (Tables 6 andPreparation of P-[CoX,(CNC,H3Me2-2,6),1 (X = C1, Br, or7).RESULTS AND DISCUSSION[Fe(CO),(CNR)] (la)-(lg) can be prepared either byrefluxing a two-fold excess of [Fe(CO),] with RNC inbmzene or via a transition-metal catalysed reactionbetween equimolar quantities of [Fe(CO),] and RNC, alsoin refluxing benzene.In both instances reaction timesare short, product purification trivial, and yields good toexcellent. Surprisingly, only one previous attempt hasbeen made to synthesize [Fe(CO),(CNR)] derivativesthermally in solution (heptane) .20 The derivatives arepale yellow to yellow crystalline solids [except (lb), ayellow oil] and all are mildly oxygen and light sensitive.When pure, the compounds can be stored in the dark andunder an inert atmosphere for several months with onlyminimal decomposition.The catalysed and non-catalysed reactions are bothtrivial; however, purification of the product from thenon-catalysed reaction is more simple (see below).Thus, for the studies involving the use of [Fe(CO),1072 J. CHEM. soc. DALTON TRANS. 1982(CNR)] reported herein, we have preferred to synthesize reaction solution.The reaction was monitored by i.r.the monosubstituted derivative in the absence of a spectroscopy and indicated at most, small amounts ofcatalyst. Care, however, must be exercised in the use of [Fe(CO),(CNR),] as impurity in the reactions. Nothe non-catalysed reactions. The end of the reaction is darkening of the reaction solution on completion of thevery often evidenced by the rapid darkening of the reaction was evidenced in these catalysed reactions.TABLE 6Analytical data for the complexes [Fe(CO),-,(CNR),(XY)] (n = 2-4; XY = tcne or I,)Analysis a (%)Complex[Fe( CO) ,( CNBut ),(tcne) J[Fe(CO),(CNCH,C,H,),(tcne)l[Fe(CO),(CNC,H,Me,-2, 6),(tcne)][Fe( CO) (CNBut) ,(tcne)][Fe (CO) (CNC,H,Me,-2,6) ,( tcne) ][Fe(CNC,H,Me,-2, 6),(tcne)]*0.5CHC13cis-[ Fe (CNC,H3Me,-2,6) al,]trans-[Fe(CNC,H,Me2-2,6),I,]Colour M.p.("C) CYellow >200 53.65 (53.2)Yellow b 61.55 (60.75)Yellow 197-198 66.2 (66.2)Yellow >200 56.9 (57.25)Yellow 198-200 66.25 (67.45)Yellow >200 66.6 (66.4)Brown 195-200 bGreen 198-200 b 1 51.25 (51.3)H4.65 (4.45)3.15 (2.95)4.7 (4.15)5.8 (5.85)4.6 (4.7)4.3 (4.45)4.35 (4.35)N halogen20.8 (20.7)17.7 (17.7)15.25 (14.5)21.35 (21.25)16.1 (16.2)14.65 (14.6) 6.35 (6.95)6.65 (6.7) 31.3 (30.45)0 Calculated values in parentheses. b Decomposes without melting.reaction solution ( i ~ . from yellow or orange to green orbrown). Although this effect does not appear adverselyto affect the product yield, an undesirable colour isimparted to the product and purification to remove thediscolouration has proved difficult. The reason for thisdarkening has not been thoroughly investigated, but isalmost certainly due to reaction of the excess [Fe(CO),]with the product, [Fe(CO),(CNR)].Thus, reaction ofPurification of the reaction mixture from the catalysedreaction requires elution through a silica gel column toremove catalyst (and possibly other impurities) followedby crystallisation. By contrast the products from thenon-catalysed reaction can be purified by crystallisationalone, although in general columns were used. The onlydifficulties that occur with the purification of the non-catalysed reaction product relate to the use of C,H,,NCTABLE 7Spectroscopic data for the complexes [~e(CO),_,,(C~uK),~(X'L')J ( n -- 2-4; XY = tcne or I,)I.r.".b (cm-I) IH N.m.r.C*d (T)hT I - - r \hv(CN) V ( W V(C0) CH, CH, Aromatic(6a) 2 230 (sh) 2 206, 2 186 (sh) 2 070, 2 026, 2 012 8.48(6b) 2 231 (sh) 2 216 2 081, 2 034 4 97 2.50(6c) 2 236w 2 201 (sh), 2 176, 2 156 (sh) 2 070, 2 026 7.48 2.75(6d) 2 225 (sh) 2 210 (sh), 2 185 1979 8.41, 8.51(6e) 2 230w 2 189m, 2 155 2 014 7.47, 7.53 2.80, 2.90 e(6f) 2 224w 2 176m, 2 131, 2 106 (sh) 7.45, 7.54 f 2.80, 2.92 f2.80, 2.82 f 2 18Ow, 2 142, 2 134 (sh), 2 1122 134 7.34 2.77KBr matrix.Recorded in CDC1,f Relative intensity 1 : 1.7.29, 7.49 f (6g)(6h)All bands strong unless otherwise stated; w = weak, m 2 medium, sh L shoulder.relative t o SiMe,.d Methyl and methylene resonances, singlets. e Relative intensity 1 : 2.J?e(CO),( CNC,H,Me,-2,6)] with [Fe( CO),] under thereaction conditions produces an identical darkening ofthe react ion solution.The catalysed synthesis of (la)-( lg) has been achievedusing a variety of transition-metal and other catalysts(see below) but only the use of CoC1,*2H20 as catalyst hasbeen systematically investigated in detail. The additionof catalytic amounts (ca. 1.5 mol yo) of CoCl, to [Fe(CO),]and RNC results in vigorous CO evolution from the greenand PhNC. Trace amounts of the unreacted RNC madepurification of the products difficult and necessitated theuse of columns to obtain [Fe(CO),(CNR)] free from RNC.The disubstituted derivatives, [Fe(CO),(CNR),] (2a)-(Zg), have been synthesized by two routes; one involvingthe reagent trimethylamine N-oxide and the other routeinvolving the use of CoC1, as catalyst.The reaction between [Fe(CO),(CNR)] and RNC (re-fluxing benzene) in the presence of CoC1, as catalystTABLE SAnalytical and spectroscopic data for the complexes P-[CoX,(CNC,H,Me,-2,6),] (X = C1, Br, or I)1.r.(cm-I) Analysis (%)c7-A l r - \ Complex v(NC) v(NC) C H N halogen pea.d/B .M.P-[CoC12(CNC,H,Me,-2,6),]*MeOH 2 180 2 184 63.65 (64.7) 5.7 (5.85) 8.35 (8.15) 10.5 (10.35) 2.06P-[CoBr,( CNC,H,Me,-2,6) ,] 2 180 2 182 56.7 (58.15) 5.05 (4.85) 7.3 (7.55) 22.15 (21.55) 2.00p-[Co12(CNC,H,Me,-2, 6) ,] 2 170 2 170 50.75 (51.6) 6.25 (6.7) 6.45 (6.7) 30.7 (30.35) 1.82Recorded in CHCI,.KBr matrix. Calculated values in parentheses. Gouy method (293 K)J. CHEM. SOC. DALTON TRANS. 1982 1073results in the rapid, facile synthesis of (2a)-(2g) in highyield (Table 3). In the absence of a catalyst the abovereaction proceeds with difficulty (R = Me, C6H11, But, orPhCH,) or not at all (R = Ph, 2,6-Me,C6H,, or 2,4,6-The products are generally free of contamination byeither starting material or higher substituted derivatives.Purification by silica-gel column chromatography (in theabsence of light) is necessary to remove the catalyst whichotherwise discolours the final product.The use of trimethylamine N-oxide in organometallicsynthesis has yielded a range of products not readily pre-pared by conventional means.Thus the CO substitutionof [OS,(CO),,],~~ [Mn,(CO)t0],22 [Fe(CO),],23 etc. by ligandssuch as PR, and CH,CN is readily achieved under mildconditions. We have thus carried out an investigationof the reaction [Fe(CO),(CNR)] + RNC ---+. [Fe(CO),-(CNR),] in the presence of trimethylamine N-oxide tocompare the results with the CoC1, catalysed reaction.Results are reported in Table 1 and it can be seen that theamine oxide route also gives the required product,rapidly, and in high yield. However, the reaction onlyproceeds to completion in the presence of excess of bothamine oxide and RNC. Consequently excess amineoxide, isonitrile, and [Fe(CO),(CNR),] are found mixedwith the required product. We have also observed thatthe final reaction solutions are often coloured a brightyellow and this colouration is difficult to remove fromthe product, even by column chromatography.Disubstituted derivatives containing two differentRNC groups can also be prepared by the CoCl, catalysedroute.Thus, the react ion of [Fe (CO),( CNC6H,Me,-2 ,6)]with ButNC readily produces [Fe(CO),(CNC,H,Me,-2,6)-(CNBut)]. In general, however, the preparation of themixed derivatives is complicated by RNC exchange (seebelow) and separation of the disubstituted products hasproven difficult.The catalytic synthesis of [Fe(CO),(CNR),] complexescan also be achieved by addition of two equivalents ofRNC (either stepwise or by simultaneous addition) to[Fe(CO),]. The products prepared in this manner con-tain a larger percentage of [Fe(CO),(CNR),] than thatobtained from the [ Fe (CO),( CNR)]-RNC reaction.The complexes (2a)-(2g) are yellow crystalline solids,oxygen and light sensitive, particularly in the presence ofimpurities.They are generally difficult to store forextended periods of time but, if necessary, can be storedin the dark, under nitrogen, below 0 "C.The [Fe(CO),(CNR),] derivatives (3a)-(3g) have beensynthesized by the CoCl, catalysed reaction between[Fe(CO),(CNR)] and RNC in refluxing benzene (Table 5).In the absence of a catalyst no reaction between [Fe-(CO),(CNR)J and RNC has been detected (or reported inthe literature). The only evidence we have obtained fora non-catalytic synthesis of [Fe(CO),(CNR),] is via thereaction between [Fe,(CO),,] and ButNC (benzene,80 "C). Infrared spectra recorded on the reactionmixture indicated the presence of [Fe( CO),( CNBut),](ca.30%).'Whereas the synthesis of [Fe(CO),(CNR)] (la)-(lg)and [Fe(CO),(CNR),j (2a)-(2g) all required reactiontimes of less than 10 min, a marked difference in thereactivity of alkyl and aryl isonitriles is apparent in thesynthesis of [Fe(CO),(CNR),] (3a)-(3g). Thus com-plete synthesis of (3a)-(3d) requires 25-35 minwhereas those for the aryl derivatives (3e)-(3g) are allcomplete in <5 min.The synthesis of (3a)-(3g) can also be achieved by thestepwise addition of two or three equivalents of RNC toeither [Fe(CO),(CNR)] or [Fe(CO),] and the final productis obtained in good yield and purity.Compounds (3a)-(3g) are all yellow solids which areoxygen and light sensitive.The complexes rapidlydarken and decompose even under an inert atmosphere.Stability has been found to decrease from aryl- to alkyl-isonitrile complexes. Due to their oxygen sensitivitythe [Fe(CO),(CNR),] complexes must be handled underan inert atmosphere and in dry, degassed solvents at alltimes. Purification was achieved by precipitation fromthe reaction mixture followed by recrystallisation. Thealkyl derivatives (e.g. RNC = MeNC or ButNC) reactwith solvents such as CHCl, and CH,Cl,, presumably viaan oxidat ion mechanism.Addition of three equivalents of RNC to [Fe(CO),-(CNR)] in the presence of CoCl, in refluxing benzeneyields the new complexes [Fe(CO) (CNR),] (4a)-(4d).Simultaneous or stepwise addition of the equivalents ofRNC leads to the same product.Attempts to prepare[Fe(CO)(CNR),] (R = Me, C6H11, or PhCH,), even withextended reaction times (ca. 24 h) were unsuccessful andthe prolonged reaction time only resulted in reactantdecomposition. [Fe(CO) (CNBut),] was obtained onlywith difficulty (ca. 50% yield after 7 h) and the use ofother solvents, e.g. tetrahydrofuran (thf) or toluene, didnot give this product in higher yield.Complex (4a) was found to precipitate from the hotreaction solution as a yellow crystalline material.Filtration under nitrogen followed by recrystallisationfrom toluene-pentane at -78 "C gave the pure product.Complexes (4b)-(4d) were completely soluble in thereaction solution and were purified by crystallisation fromtoluene-pentane at -78 "C; (4a)-(4d) are all yellow ororange, oxygen and light sensitive materials,The simultaneous or stepwise addition of four equi-valents of RNC to [Fe(CO),(CNR)] in refluxing benzeneand in the presence of CoCl, leads to the [Fe(CNR),]derivatives (5a)-(5c).Attempts to prepare [Fe(CN-But),] l59Z4 via this route have been unsuccessful. Thecomplexes (5a)-(5c) are red, oxygen and light sensitivematerials, susceptible to oxidation by 0, and solventssuch as CHCl, and CH,Cl,.Product Characterization .-All the complexes have beencharacterized by i.r. and n.m.r. spectroscopy (Tables2, 4, and 5). Where possible, the complexes have alsobeen characterized by mass spectrometry and elementalanalyses.Possible structures for the complexes [Fe-(CO)5-,(CNR)n] (n = 1-5), assuming idealized trigonal-bipyramidal structures, are shown below1074 J. CHEW SOC. DALTON TRANS. 1982RNOCJ coOC c0R RNcoNRR Poc. INRNRThe [Fe(CO),(CNR)] derivatives (la)-( lg) all containone v(NC) and three v(C0) stretching frequencies con-sistent with axial substitution of the trigonal bipyramidby RNC, as reported previously.25. The v(NC) stretchingfrequencies are all higher than v(NC) for the free ligand,consistent with the higher x-acceptor ability of COrelative to RNC.l-, The lH n.m.r. spectra all show theexpected singlet for the CH, or CH, resonance (whereapplicable) and in every case the resonance is moveddownfield relative to the free isonitrile.No long-rangelH-14N coupling, as detected for [Cr(CO),(CNR)],26 wasobserved. The [Fe(CO),(CNR)] complexes were furthercharacterized by mass spectrometry. In every case thespectra indicated the presence of the parent ion, M+, aswell as fragments corresponding to the stepwise loss ofCO groups. In addition, fragments corresponding tobreakdown of the RNC ligand were also observed onoccasion. The mass spectral data (fragments, intensitydata) for a typical example, [Fe(CO),(CNC6H,Me,-2,6)],are given in Table 9.TABLE 9Mass spectral data for the complexes [Fe(CO)s-n-(CNC,H,Me,-2,6),] (n = 1-5) an = l b n = 2 c n = 3 d n = 4 n = 5FeL,+ 5Fe(CO)L,+ (1 < 1 ( 5FeL,+ 1 < 1 12Fe(C0) ,L,+ 3Fe(CO)L,+ 5FeL,+ 23 5 22Fe( CO) 3L2+ 16 9FeL,+ 100 100 100 74Fe( CO) ,L+ 7Fe(CO),L+ 12Fe(C0) ,L,+ 24 16Fe( CO) L,f 20 13Fe(CO),L+ 14Fe( CO) L+ 30FeL+ 100 92 91 74 40L+ 49 60 91 78 100Fe+ 63 57 39 72 51only m/z values >lo0 reported (except Fe+); spectrarecorded at 25-200 "C.FeL+] ;doubly charged ions, Fe(CO)Q-nL2+ (n = 0-3) and L2+, ob-served at m/z = 135.5, 121.5, 107.5, 93.5, and 64.5 respectively(intensity <lye). c m* = 292 [Fe(CO)L,+ ----r FeL,+].d m* = 292 [Fe(CO)L,+ + FeL,+] : doubly charged ionsFeL,,+ (mlz = 224.5) Fe(CO)LZ2+ (173), FeLZ2+ (159), andFeL2+ (93.5) Were observed.b m* = 162.5 [Fe(CO)L+The i.r. data for the [Fe(CO),(CNR),] derivatives areconsistent with the data reported previously, i.e. diaxialsubstitution of the trigonal bi~yrarnid.~~ The un-expected observation of two v(C0) bands has alreadyreceived attention and is most probably a result of adeviation from the idealised D3h symmetry obtained bythe bending of the CNC unit of the isonitrile ligand.Acrystal structure determination is presently underway toconfirm this prediction. The frequency of v(NC) in(2a)-(2g) is lowered with respect to v(NC) in (la)-(lg)and this is interpreted in terms of an increase in electrondensity on the Fe atom due to the greater a-donor abilityof the RNC ligand, relative to CO.The lH n.m.r. spectra show the expected singlets forthe CH, and CH, resonances where applicable. Themass spectra of (2a)-(2g) all show fragments correspond-ing to M+, loss of CO, and breakdown of the RNC ligand.Mass spectral data for [Fe(CO),(CNC6H,Me2-2,6)~ areshown in Table 9.Trisubstitution of [Fe(CO),] leads to [Fe(CO),(CNR),]and the overall geometry is expected to consist of atrigonal-bipyramidal structure containing two CO ligandsin the equatorial plane.An analysis of this structure(C, symmetry) leads to the expectation of two v(C0)and three v(NC) absorptions in the i.r. spectrum. Thei.r. spectra are all found to contain two v(C0) and onebroad v(NC) band centred at 2 060-2 110 cm-l whichare reasonably consistent with the analysis. However,a structure containing one axial and one equatorial COgroup (C, symmetry) on analysis would also be con-sistent with the data, i.e. two v(C0) and one v(NC) band.The lH n.m.r.data all show singlets for the methyl ormethylene resonances, where appropriate. The n.m.r.data thus indicate that the structure is fluxional atambient temperature and hence do not differentiatebetween the C, and CzV structures. The mass spectrumof [Fe(CO),(CNC6H3Me2-2,6),] is in agreement with thestructure formulation and the data are recorded in Table9. A feature of these data is the appearance of fragmentscorresponding to Fe(CO),L,+ and Fe(CO)L,+. Thissuggests that a reaction such as 2[Fe(CO),L,] - [Fe(CO),L,] + [Fe(CO)L4] is occurring in the massspectrometer, as has been observed for the isonitrilederivatives of Group 6 metal hexacarbonyls.26 Twoindependently synthesized and purified samples of[Fe(CO),(CNC,H,Me,-2,6),] gave mass spectra withidentical intensity data ratios (180 "C) and this impliesthat the above results are not attributable to the pre-sence of [Fe(CO),L,] and [Fe(CO)L,] impurities.The [Fe(CO)(CNR),] complexes can have either of twopossible geometries, a structure with the CO axial (C3v) orequatorial (C2,) with respect to a trigonal-bipyramidalstructure. An analysis of both structures indicates thateither one v(C0) and three v(NC) (C3u) or one v ( C 0 ) andfour v(NC) (C,,) absorption bands are expected.The i.r.data indicate a single v(C0) band at 1880-1 910 cm-land one or two broad bands [v(NC)], between 1990 and2 100 cm-l. Band overlap thus makes it impossible todecide on the structure geometry on i.r. evidence aloneJ. CHEM. SOC. DALTON TRANS.1982although the C% structure is expected on electronicgrounds.,' The lH n.m.r. spectra all show singlemethyl resonances for (4a), (4c), and (4d), and againfluxional behaviour is observed. The mass spectrum ofone of the derivatives, [Fe(CO) (CNC,H3Me,-2,6),] (Table9), confirms the identity of the tetrasubstituted com-plexes.Characterization of the [Fe(CNR),] derivatives hasalso been achieved by a combination of i.r. and n.m.r.spectroscopy. The i.r. spectra all contain a broad com-plex absorption in the region 1 900-2 050 cm-l and thei.r. spectrum of (5b) is in complete agreement with thatreported previou~ly.~~ The 1H n.m.r. spectra of (5b) and(5c) both show singlets for the methyl resonances. I thas been reported that (5b) is fluxional down to lowtemperatures (-90 "C) and (5a) and (5c) most probablybehave similarly at low temperatures.The massspectrum of [Fe(CNC6H,Me,-2,6),] is shown in Table 9and further confirms the proposed formulation of thesubstituted derivatives.We have further characterized the new complexes[Fe(CO)5-,(CNR),] (n = 2 4 ) by chemical reaction withtcne and iodine. The oxidation reactions lead to stableproducts of the type [Fe(CO)e-,(CNR),L] (n = 2 4L = tcne or I,). The tcne derivatives are similar to the[Fe(CO)4-,(CNR),(olefin)] (n = 1-3) complexes whichwe have previously prepared from [Fe(CO),(olefin)] andRNC in the presence of [RhCl(CO)(PPh,),] as catalyst.,*Thus addition of tcne to [Fe(CO)s-,(CNR),] (n = 2,R = But, PhCH,, or 2,6-Me,C6H,; n = 3, R = But or2,6-Me,C6H,; n = 4, R = 2,6-Me,C6H,) gives the com-plexes [Tie( C0)4 -,(CNR),(tcne)] (6a)-(6f).Similarly,addition of I, to [Fe(CO)(CNC,H,Me,-2,6),] gives cis- andtrans-[Fe(CNC6H3Me-2,6),I,] (6g)-(6h).The reactions are facile and reflect the ease of oxid-ation of the parent zerovalent isocyanide complexes.Although the displacement of either CO or RNC is pos-sible, on the basis of the products detected, the dominant,if not exclusive reaction , is CO displacement.The new complexes (6a)-(6h) have been completelycharacterized by i.r. and n.m.r. spectroscopy (Table 7)and elemental analyses (Table 6). Complexes (6b) and(6c) each show two bands while (6a) shows three bandsattributable to vlC0) in their respective i.r.spectra(KBr). The latter is probably due to a solid-state effectsince in CH,Cl, only two v(C0) bands are observed(2063 and 2009 cm-l) for (6a). The lH n.m.r. spectraof (6a)-(6c) show only single peaks for the methyl andmethylene resonances. Neither i.r. nor n.m.r. spectro-scopy can thus be used to distinguish between the threepossible isomeric structures for (6a)-(6c) shown below.R RN N 00CN 0R1075Complexes (6d)-(6e) both show a single v(C0) band inthe i.r. spectrum. This together with the lH n.m.r.spectra (2: 1 ratio for the CH, resonance) is consistentwith either of the two structures shown below.R!RNC OC*NRThe similarity between the i.r.RN,RNCband n.m.r. data for (6f)and (6g) suggest that (6g) is the cis isomer of [Fe(CNC,H,-Me2-2,6),I,].This is also apparent from the n.m.r. andi.r. spectra of (6h) which are consistent with a trans con-figuration (see below). The isomers (6g) and (6h) areRNNRcis transreadily separated by column chromatography butthermally interconvert in benzene (80 "C).Mechanistic Studies.-Addition of ButNC to [Fe-(CO),(CNBut)] and CoX,*nH,O (X = C1, Br, or I ; n =2,3, or 4 respectively) under standard reaction conditions(see Experimental section) results in the rapid formationof [Fe(CO),(CNBut),]. During this reaction the catalystchanges colour from blue to green, the solution turnsgreen, and vigorous evolution of CO (as detected by1205) takes place. No CO, evolution, as detected byaqueous solutions of Ba(OH),, is observed.Infraredspectra and elemental analyses of the used catalyst are inaccordance with its formulation as [CoX,(CNBut),].This compound can readily be prepared from COX,*1zH,0 and ButNC in refluxing methanol, benzene, oracetone. Addition of catalytic amounts of [COX,-(CNBut),] , prepared as above, to [Fe(CO),(CNBut)] andButNC results in the same rapid catalysis and colourchanges as observed for the addition of the appropriateCoX,*nH,O catalyst (see below). Identical results werealso observed when [Fe(CO),(CNR)] was reacted withRNC (catalysed by CoC1,*2H2O) to give [Fe(CO),(CNR),](R = Me, C&11, PhCH,, Ph, 2,6-Me2C6H,, or 2,4,6-Me&,H,) and for the catalytic synthesis of [Fe(C0)5-,-(CNR),] (n = 1-5) derivatives.I t is thus apparent that the initial step in the reactionsunder investigation is the formation of [CoX,(CNR),]from COX, and RNC.Complexes of this type are knownto exist in two forms, one diamagnetic [a modification,(a)] and the other paramagnetic [p modification, (b)].29A magnetic moment determination (Gouy method) on[CoX,(CNC,H3Me2-2,6)4] (X = C1, Br, or I) has estab1076 J. CHEM. SOC. DALTON TRANS. 1982lished the paramagnetic nature of these complexes (Table8). Further, the data are in agreement with previouslyreported magnetic moment data on related complexes.mpmr 7 + YLFurther observations pertaining to the use of thecatalysts have been made and are listed below.(1) Water has no effect on the reaction. Although wehave routinely dried the cobalt salts in vacuo, use ofCoCl2*6H,O does not detectably influence the course ofthe reaction.In practice, addition of CoCl2*6H,O torefluxing benzene prior to reactant addition was found todehydrate the catalyst, as detected by a colour change ofthe catalyst from pink to blue.(2) The use of Co(PF,),, CoSO,, Co(NO,),, and co-(O,CMe), as catalysts for the substitution reaction rulesout the need for halide participation and the use ofCo(PF,), rules out the need for a bridging anion in thereaction mechanism (see below).(3) The catalysed and uncatalysed reaction [Fe-(CO),(CNR)] + RNC d [Fe(CO),(CNR),] is unaffectedby light, by radical inhibitors (galvinoxyl,* hydroquin-one), or radical initiators (azobisisobutyronitrile) .(4) The reaction is not limited to cobalt(I1) catalysts.For instance, NiX,*nH,O (X = C1, Br, or I) has beenfound to catalyse efficiently the reactions [Fe(C0)5-,-(CNR),] + RNC - [Fe(C0)4-,(CNR),+1] (n = 0 or 1).Typically, a yield of 80-90% of product has beenobtained for the reaction with n = 1 and NiC1,*4H2O ascatalyst (R = Me, C6H11, ButNC, PhCH,, 2,6-Me,C,H3,or 2,4,6-Me,C,H2).Reaction times are all (10 min (1mmol reactants, 20 mg catalyst, in benzene at 80 "C).In addition, NiC1,*4H20 has also been used to synthesize[ Fe (CO),( CNC,H,Me2-2 ,6),] but the at tempted synthesisof [Fe(C0)6-,(CNC,H3Me,-2,6),] (n = 4 or 5 ) was un-successful. The major disadvantage to the use ofnickel(I1) halides as catalysts for the above substitutionreactions is the competing catalytic polymerization ofisonitrile that occurs in the presence of the nickel salts(see below).(5) The reaction between [Fe(CO),] and RNC is notonly catalysed by metal@) salts.To date we haveobtained catalysis, with varying degrees of success, withover 50 transition-metal complexes. Our results usingsome of these catalysts have already been reported :[RhC1(PPh,),],28*31 [RuC~,(PP~,),],~~ and [{Fe(C0),(q5-C,H,)},].32 The use of other catalysts such as Pd/C,PtO,, and [Mn,(CO),,], [Os,(CO),,], and [Ir4(C0),,] will bereported in the near future.=(6) The effect of different solvents on the reaction[Fe(C0)4(CNBut)] + ButNC - [Fe(CO),(CNBut),] +CO has also been investigated (Table 10). No apparent* 2,6-Di-t-butyl-a-( 3,5-di-t-butyl-4-oxocyclohexa-2, B-dien-l-yl-idene) -p-tolyloxy .correlation between solvent properties and catalyticactivity is apparent from the data.The excellentresults for the solvents thf and benzene are, however, tobe noted. Since the CoC1, reaction takes place underboth homogeneous (acetone as solvent) and heterogeneousconditions (hexane as solvent) the reaction using CoCl,=2H,O as catalyst could occur at both the surface of theundissolved CoCl, as well as via interaction with thedissolved [CoCl,(CNR)4-,] (TZ = 1-4) species. Thedata could thus reflect the number of active sites avail-able for catalysis but, to date, we have made no effort toascertain the real number of active catalyst molecules inthe different solvents.TABLE 10Effect of solvent on the CoC1, catalysed reaction a [Fe(CO),-(CNBut)] + ButNC -+ [Fe(CO),(CNBut).j + COPentane (36.1) 15 (n.r.) ,IHeptane (98.4)Dichloromethane (40.0) 15Chloroform (61.2) 15Solvent b Reaction time c (min)Hexane (68.9) ;; (30%) *Cyclohexane (80.7) 101,S-Dichloroethane (83.4) 4Tetrahydro furan (65.4) 3Benzene (80.1) 2Acetonitrile (8 1.6) 4Acetone (56.2) 5Methyl ethyl ketone (79.6) 4Methanol (64.5) 10Ethanol (78.3) 8a [Fe(CO),(CNBut)] : ButNC : CoC1,*2H20, 1.0 : 1.4 : 0.1mmol; solvent (reflux), 10 cm3.Boiling point ("C) in paren-theses. c As determined by i.r. spectroscopy. n.r. = Noreaction. 30% reaction in 15 min.(7) Transition-metal salts are known to catalyse thepolymerization of isonitriles and this competing reactionhas been found to cause loss of catalyst activity.34 Thus,relatively long react ion times {particularly for thesynthesis of the higher substituted isonitrile derivativesof [Fe(CO),]} and reactions carried out in high-boilingsolvents have on occasion been found to result in catalystdeactivation.This may be observed by the slow form-ation of yellow reaction solutions as well as by theappearance of a yellow insoluble material. Thus, adecrease in catalytic activity obtained from qualitativekinetic data on the reaction between [Fe(CO),(CNC,H,,)]and C,H,,NC in the presence of CoC1,*2H20 as catalyst(1 : 5 : 0.1 mmol respectively; in benzene at 45 "C) can beassociated with a simultaneous gradual change in colourof the reaction solution to yellow.At increased isonitrileconcentrations (e.g. C,H,,NC : CoC1,*2H2O, 100 : 1) ayellow solid (>95% yield based on isonitrile) is rapidlyformed, even at room temperature. This material,insoluble in common organic solvents and water, wasshown by i.r. spectroscopy [v(NC) ca. 1630 cm-l, v(NH)ca. 3 400 cm-l; KBr] to have the same structure as thatof a polyisonitrile with a poly(Schiff base) structure 34 ofthe type shown bdow. Similar observations have beenmade for the CoCl, catalysed substitution reactions of theGroup 6 metal hexacarbonyls with isonitri1es.S Ingeneral our observations suggest that alkyl isonitrileJ. CHEM. soc. DALTON TRANS. 1982 1077are more susceptible to catalytic polymerization thanthe corresponding aryl derivatives.It is thus apparentfrom our results that the CO substitution reaction com-petes favourably with the isonitrile polymerisation re-action but that deviations from the reaction procedureslisted can result in poor substitution reactions.% Un-fortunately, the competing polymerization reaction hasthus far had adverse effects on our attempts to obtainquantitative kinetic data for the substitution reaction.(8) The isonitrile originally attached to the catalystcan readily be transferred to the iron-carbonyl reagent.For instance, reaction of [Fe(CO),] with 2,6-Me2C,H,NC,in the presence of [CoC1,(CNC,H3Me2-2,6),] as catalyst{[Fe(CO) J : RNC : catalyst, 10 : 2 : 0.3 mmol; in ben-zene a t 80 “C} yields 2.7 mmol of [Fe(CO),(CNC,H,Me,-2,6)], i.e.0.7 mmol more than the yield (2 mmol) ex-pected from free isonitrile alone. A similar effect wasobserved for the reaction between [Fe(CO),] and [CoCl,-(CNBut),] (ratio, 10 : 0.3 mmol) in which [Fe(CO),-(CNBut)] was readily formed. During this reactionthe solution changed from blue-green to yellow (ca. 5min). Addition of further amounts of ButNC to thereaction solution resulted in a regeneration of the blue-green colour and a renewal of catalytic activity. Similarobservations have been made on the reaction between[Mo(CO)J and ButNC.= The mechanism of thetransfer of the isonitrile to the metal carbonyl substratecould involve either an inter- or intra-molecular process.In an intermolecular process, [CoCl,( CNR),] could pro-duce free isonitrile by the dissociation mechanism shownbelow. Any one of the cobalt species could thus act as a[CoCl,(CNR),] [CoCl,(CNR),] + RNC[CoCl,(CNR),] [CoCl,(CNR),] + RNC[CoCl,(CNR),] [CoCl,(CNR)] + RNC[CoCl,(CNR)] @ CoCl, + RNCcatalyst in the reaction below (m < 4; R’NC = solutionisonitrile).Alternatively, a bridging mechanism (intra-[C@Js(CNR) 1 [M(CO),] + R’NC mL [M(CO),-l(CNR’)]molecular process) involving direct transfer of isonitrilefrom cobalt to iron with simultaneous tranfer of CO fromiron to colbalt could also explain the data. {Thismechanism has been observed, for instance, when[RhCl(CO)(PPh,),] is used as a ~atalyst.~’} The twopossible mechanisms are shown in the Scheme.On the basis of the results obtained above, attemptshave been made to establish the source of the iron-co-ordinated isonitrile.Thus, reactions were carried out inwhich [Fe(CO),(CNC,H,Me2-2,6)], [CoCl,(CNR),], andR’NC (1.0 : 1.0 : 1.0 mmol) were heated in benzene (10cm3, 80 “C) for short periods of time (3 min). The re-action solutions were then analysed by n.m.r. spectro-scopy. The exchange reaction, [CoCl,(CNR),] + R’NC - [CoCl,(CNR),(CNR’)] + RNC, was also studiedunder identical reaction conditions (1.0 mmol reagents, inbenzene at 80 “C). These results are summarized byreactions (i)-(viii).It is apparent from the above results that ligand ex-change is occurring at the catalyst centre and that theexchange reaction is in complete agreement with theisonitrile distribution eventually obtained on the ironsubstrate.Thus, the major pathway involves attack[ Fe(CO)51 + [ C O C I ~ ( C N R ) ~ ]1of solution isonitrile on the iron-carbonyl substrate.The results further indicate that cobalt has moreaffinity for alkyl than aryl isonitriles, i.e. for a-donatingligands rather than for x-accepting ligands. By ex-tension, the possibility of replacing an isonitrile by CO(a poor a-donor, good x-acceptor ligand) seems highlyunlikely and further rules out the possibility of an intra-molecular mechanism.It must be noted however that a pathway involvingcatalytic exchange between isonitrile bound to the ironsubstrate and free isonitrile may also be significant:[Fe(CO),(CNR)] + R’NC - [Fe(CO),(CNR’)] + RNC.We have found no evidence (by n.m.r.spectroscopy) forthe exchange reaction in the absence of catalyst, but suchexchange in the presence of Corl has been detected.Thus, the reaction between [Fe(CO),(CNC,H,Me,-2,6)]and ButNC (1.0 : 1.0 mmol, in benzene at 80 “C) carriedout in the presence of CoC1, as catalyst yields [Fe(CO),-(CNC,H,Me,-2,6) ( CNBut)] (60 yo), [Fe( CO),( CNC,H,Me,-2,6),] (20y0), and [Fe(CO),(CNBut),] (20Y0), as detectedby n.m.r. spectroscopy (similar results have been foundfor the [Mo(CO),]-RNC-CoC1, reaction 35}.(9) Further evidence against a bridging intramolecularexchange mechanism is provided by the following re-sults. (i) No spectroscopic evidence for the formation ofa Co-CO complex has been observed in our reactions.(ii) CoC1, catalyses the substitution of [Fe(CO),] by non-bridging ligands such as phosphines and ph0sphites.1078 J.CHEM. SOC. DALTON TRANS. 1982[Fe(CO),(CNC,H,Me,-2,6)] + [CoC1,(CNC6H,Me2-2,6),] + PhCH,NC - [Fe(CO),(CNC6H,Me,-2,6),] (100%)[CoCl,(CNC,H,Me,-2,6),] + PhCH,NC - [CoCl,(CNC6H,Me,-2,6),(CNCH2Pli)] + 2,6-hfe,C6H,NC ( 100~o)[Fe (co),( CNC,H,Me,-2,6)] + [CoCl,( CNCH,Ph),] + 2, 6-Me,C6H3NC - [Fe(CO),(CNC,H,Me2-2,6)(CNCH2Ph)](ca. SOY0) + [Fe(C0)3(CNC6H,Me,-2,6),] (ca. 40%)[CoCl,(CNCH,Ph),] + 2,6-Me,C6H,NC - partial exchange to leave PhCH,NC (ca. 55%) and 2,6-Me,C6H,NC (ca. 45%) in solution[Fe(CO),(CNC6H,Me,-2,6)] + [CoCl,(CNC,H,Me,-2,6),] + ButNC - [Fe(CO),(CNC6H,Me,-2,6),] (ca. 95%) + [Fe(CO),(CNC6H3Me,-2,6)(CNBut)] (ca. 5%)[CoC1,(CNC6H,Me,-2,6),] + ButNC - partial exchange to leave 2,6-Me,C6H3NC (ca.90%) and ButNC (ca. 10%) in solution[Fe (CO),( CNC,H,Me,-2,6)] + [ CoCl,( CNBu t)4] + 2 ,6-Me,C6H,NC - [ Fe (CO),( CNC6H,Me,-2 ,6) ( CNBut)j(i)(ii)(iii)(iv)(v)(vi)(ca. 60%) and [Fe(CO),(CNC,H,Me,-2,6),] (ca. 40%) N . B : in the early stages of the reaction (< 10% reaction)(vii) and also under milder reaction conditions (45 "C) the reaction product is [Fe(CO),(CNC6H,Me,-2,6)2][CoC1,(CNBut),] + 2,6-Me,C6H,NC - partial exchange to leave ButNC (caIn summary, our mechanistic data indicate that theinitial interaction between COX, and RNC leads to thecomplex [CoX,(CNR)J (n = 4). These complexes, withn = 1 4 , are believed to be the active form of thecatalyst. Further, a reaction mechanism involvingthese species does not require a bridging mechanismbetween catalyst and substrate which involves either thehalide (X) or RNC.Since the predominant effect of the catalyst must be todestabilize the M-CO bond, this destabilization mustoccur via catalyst interaction with the metal (e.g.Fe) orthe CO group attached to the metal. Although attackat five-co-ordinate iron by cobalt might be possible,attack at the sterically crowded [cr(co),] molecule 35seems unlikely. Thus, attack of the catalyst presumablyoccurs at the CO ligand, either at C or 0, as shown below.Carbonyl attack at 0 by Lewis acids such as AlBr, isknown3' but AlBr, has not been found to catalyse thesubstitution reaction under investigation. In the case ofCO attack at C, the carbonyl ligand attacked is notnecessarily the CO displaced and cis 38 or trans 39 effects,resulting from the novel ligand -C(O)CoCl,(CNR),, couldbe dominant.Further work will be required to deter-mine the exact nature of the CO-catalyst interaction.An alternative mechanism based on the expectedoxidizing ability of CoC1, (and presumably [CoCl,-(CNR),]) and involving an outer-sphere electron-transferprocess can also be envisaged: [Fe(CO),] + [COX,-(CNR),] - [Fe(CO),]+ + [CoX,(CNR),]-. The 17-electron [Fe(CO),]+ species formed could be expected tobe involved in a chain or non-chain 32 radical processin which the CO was replaced by isonitrile. Our resultson the use of radical inhibitors, initiators, etc. (see above)do not favour such a mechanism for the CoCl, reactions.However, our results on the use of [(Fe(CO),(q5-C,H,)),]as catalyst for the above substitution reaction,32 as well55%) and 2,6-Me2C,H,NC (ca.45%) in solution (viii)as the recent electrochemical studies on the substitutionof CO on [Fe(CO),] by PPh3,l1 suggest that in general apathway involving radical anions and cations is viable.We thank the C.S.I.R. and the University of the Wit-watersrand for financial assistance (to N. J. C . and M. 0. A.).[1/1782 RecPived, 17th November, 19811REFERENCESMetals,' Wiley, London, 1969.L. Malatesta and F. Bonati, ' Isocyanide Complexes ofF. Bonati and G. Minghetti, Inorg. Chim. Acta, 1974, 9, 95.Y. Yamamoto, Coord. Chem. Rev., 1980, 32, 193. * P. Treichel, Adv.Organomet. Chem., 1973, 11, 21.W. Hieber and D. von Pigenot, Chem. Ber., 1956, 89, 193.W. Hieber and D. von Pigenot, Chem. Ber., 1956, 89, 610. ' A. Reckziegel and M. Bigorgne, J . Organomet. Chem., 1965,S. Grant, J . Newman, and A. R. Manning, J . Organomet.W. Strohmeier and F. J . Muller, Chem. Ber., 1969, 102, 3613.lo T. A. Manuel, Inorg. Chem., 1964, 3, 1703.l1 H. Alper and R. A. Partis, J . Organomet. Chem., 1972, 35,C40.l2 W. P. Fehlhammer, A. Mayr, and B. Olgemoller, Angew.Chem., Int. Ed. Engl., 1975, 14, 369.l3 J . Furuwaka, A. Matsumura, Y . Matsuoka, and J. Kiji,Bull. Chem. SOC. Jpn., 1976, 49, 829.l4 U. Wannagat and H. Seyffert, Angew. Chem., 1965, 77, 457.l6 J . M. Basset, D. E. Berry, G. K. Barker, M. Green, J. A. K.Howard, and F. G. A. Stone, J . Chem. SOC., Dalton Trans., 1979,1003.l6 M. 0. Albers, N. J . Coville, T. V. Ashworth, E. Singleton,and H. E. Swanepoel, J . Chem. SOC., Chem. Commun., 1980, 489.17 U. Koelle, J . Organomet. Chem., 1977, 133, 53.18 J . Casanova, R. E. Schuster, and N. D. Werner, Org. Synth.,1966, 46, 75.19 W. P. Weber, G. W. Gokel, and I. K. Ugi, Angew. Chem.,Int. Ed. Engl., 1972, 11, 530.2o N. J. Bremer, A. B. Cutcliffe, M. F. Farona, and W. G.Kofron, J . Chem. SOC. A , 1971, 3264.21 B. F. G. Johnson, J. Lewis, and D. A. Pippard, J . Chem.Soc., Dalton Trans., 1981, 407.22 U. Koelle, J . Organomet. Chem., 1978, 155, 53.23 Y. Shvo and E. Hazum, J . Chem. SOC., Chem. Commun.,1975, 829.2* D. Gladowski and F. R. Scholer, Abstracts from CentennialMeeting of the American Chemical Society, New York, 1976,paper INOR 133.J. A. McCleverty, J. Chem. SOC., Dalton Trans., 1972, 1246.3, 341.Chem., 1975, 96, C11.26 F. A. Cotton and R. V. Parish, J . Chem. SOC., 1960, 1446.26 J. A. Connor, E. M. Jones, G. K. McEwen, M. K. Lloyd, anJ. CHEM. SOC. DALTON TRANS. 1982 1079J. R. Shapley and J. A. Osborn, Acc. Chem. Res., 1973, 6,28 M. 0. Albers, N. J. Coville, P ten Doeschate, and E. Single-29 D. Baumann, H. J. Keller. D. Nothe. H. H. Rupp, and G.30 L. Malatesta and A. Sacco, Gazz. Chim. Ital., 1953, 83, 499.31 M. 0. Albers, N. J. Coville, C. P. Nicolaides, R. A. Webber,T. V. Ashworth, and E. Singleton, J. Organomet. Chem., 1981,217, 247.3p M. 0. Albers, N. J. Coville, T. V. Ashworth, and E. Single-ton, J. Chem. SOC., Chem. Commun., 1981, 408.33 M.O. Albers, N. J. Coville, and E. Singleton, J. Chem. SOC.,Chem. Commun., 1982, 96.84 R. J. M. Drenth, R. W. Stephany, and W. Drenth, Red.305.ton, S. Afr. J. Chem., 1981, 34, 81.Uhlmann, 2. Naturforsch., Teil B, 1976, 31, 912.Trav. Chirn. Pays-Bas, 1973, 92, 83; W. Drenth and R. J. M .Nolte, Acc. Chem. Res., 1979, 12, 30.35 M. 0. Albers, N. J. Coville, T. V. Ashworth, E. Singleton,and H. E. Swanepoel, J . Organomet. Chem., 1980, 199, 55.M. 0. Albers, N. J. Coville, T. V. Ashworth, and E. Single-ton, J. Organomet. Chem., 1981, 217, 385.37 D. F. Shriver and A. Alich, Coord. Chem. Rev., 1972, 8, 15.38 T. L. Brown and P. A. Bellus, Inorg. Chem., 1978, 17, 3726and refs. therein.39 F. A. Cotton and G. Wilkinson, ' Advanced InorganicChemistry,' 4th edn., Wiley, New York, 1980.40 B. H. Byers and T. L. Brown, J . Am. Chem. SOC., 1977, 99,2527.41 G. J. Bezems, P. H. Rieger, and S. Visco, J. Chem. SOC.,Chem. Commun., 1981, 265

ISSN:1477-9226    

DOI:10.1039/DT9820001069

出版商:RSC    

年代:1982

数据来源: RSC