y3-Allyl Metal Hydride Complexes. Oxidative Addition of Cyclopropane

Tulip, Ibers / q3-AIlyl Metal Hydride Complexes. 420 1 y3-Allyl Metal Hydride Complexes. Oxidative Addition of Cyclopropane and Olefin Substrates to I...
1 downloads 0 Views 1MB Size
Tulip, Ibers

/

420 1

q3-AIlyl Metal Hydride Complexes

y3-Allyl Metal Hydride Complexes. Oxidative Addition of Cyclopropane and Olefin Substrates to Iridium( I ) Complexes. Structure of IrC1H[v3-C3H4( 1-C6Hd][P(C6HS)3]2 Thomas H. Tulip and James A. tbers* Contribution f r o m the Department of Chemistry, Northwestern University, Ecanston, Illinois 60201. Received Nocember 29, I978

Abstract: Phenylcyclopropane reacts w i t h trans- IrCI(N2)(PPh3)2 or “lrCI(PPh,)z” to yield IrClH [q3-C3HJ(I-Ph)](PPh3)?. This complex is also obtained from similar reactions with allylbenzene or tmns-P-methylstyrene. A series of analogous complexes is formed from the reaction of allylbenzene, [IrCI(COT)l]z (COT = cyclooctene), and the appropriate donor ligand i n which PPh3 (Ph = ChH5) is replaced by P(p-Tol), (p-To1= 4-tolyl), AsPh,, As(p-Tol)3, or SbPh3. These q3-allyl hqdride coniplexes do not exhibit dynamic behavior in solution a5 judged by their variable-temperature N M R spectra. They possess unusual thermal stability and are not highly air sensitive. Chloroform solutions of the q3-allyI hydride complexes do not effect thc conversions of cyclopropane to olefin or primary to internal olefin. The complex IrClH [q3-C3H4(1 -Ph)]( PPhj), reacts w i t h CO or PF3 to liberate omethylstyrene, which also results from the extended exposure of the complex to 0 2 . Reaction w i t h HCI yields a mixture of products from which a new isomer of IrC12li(CO)(PPh3)2 is obtained upon treatment with CO. Similar reaction w i t h H B r causes total substitution of chlorine. Spectroscopic data indicate that IrClH [q3-C3H4( I -Ph)](PPh3)? adopts a geometry in which the phosphine ligands are mutually cis, hydrido and chloro ligands are trans. a n d t h e a l l q l group occupies two coordination sites. This structure is substantiated by a single-crystal X-ray difrraction stud). The complex crystallizes in space group C2i.”-Pn21awith four formula units in a cell of dimensions a = 14.902 (2), b = 11.016 (2). and c = 22.456 (4) A. Based on 3738 unique reflections having F,’ > 3 a(F,?), the structure was refined by full-matrix least-squares techniques to conventional agreement indices (on F ) of R = 0.029 and H, = 0.041. The hydride and allyl hydrogen atoms were located and refined. The Ir-CI bond length is long, 2.549 (2) A, as a result of the trans influence of the hydrido ligand. Thc geometry of the allyl group, including hydrogen atoms, is similar to that found in other q3-a11y1complexes. The implications of t h e isolation of these q3-allyl hydride complexes to catalytic transformations of cyclopropanes and olefins are discussed.

The ability of transition-metal complexes to effect structural transformations in organic substrates constitutes a basic facet of organometallic chemistry. Two such transformations which have received much attention are the metal-assisted rearrangements of strained-ring molecules, particularly the conversions of cyclopropanes to olefins (eq I ) , ] and the metal-promoted isomerizations of olefins (eq 2).2 Allyl metal

Scheme I

B Scheme II

hydride (A-M-H) complexes have been implicated as key intermediates in both of these transformations. The former reaction is proposed to proceed via the mechanism presented in Scheme I . The intermediacy of an A-M-H complex, B, accounts for the variety, distribution, and stereochemistry of the observed products.’~’-8 The initial step in the sequence, insertion of a metal center into a three-membered ring to form a metallocyclobutane complex, A, has ample precedent9 I h and the conversion of A to C has recently also been reported.” However, the crucial central step, P-hydrogen abstractionlX with concomitant formation of the A-M-H complex, B, has not been demonstrated. Two major mechanisms have been envisioned for the metal-promoted isomerizations of olefins. Although much less

L,M

“16 H

H 1

/

H

L,M----

B

well documented than the metal hydride addition-elimination mechanism (eq 3),2c‘.’9 an allyl metal hydride mechanism’? 1q (Scheme I I ) has received support from detailed stereochemical studies.’(’ ?‘) However, the intermediacy of complex B has only been confirmed by direct observation in one system,’” although dynamic equilibria between species of forms D and B have been observed i n two cases.317 3

0 1979 American Chemical Society

1 101.15 /

Journal of the American Chemical Society

4202

July 18, 1979

Table I. Spectral Data for IrCIH[$-C3H4( I-Ph)]L*

L

PPh3

AsPh3

SbPh3

P(P-T~I)~

As@-Tol)3

IR" vir

tii.

cm-I 2208

2179

2152

2200

3.87 (d)

NMRh 3.82 (d)

4.02 (dd)

2213

It4

ppm

;I{,,(

3.90 (dd) = 8.4 Hz)

(JHbP,,,,,

hl?

(Jti?P,rrn,

2.72 (d) = 10.3 HZ)

2.88 (dd)

3.65 (d)

2.98 ( d )

3.75 (d) = 9.1 HZ)

2.73 (dd)

2.61 (d)

= I I .3 H7.1

2.87 (dd) (JIi,P,,,,,

'h14 hl,

2.79 (d)

2.74 (dd) (JH*P!,,,,

h13

(JH~P,,,,,

= 3.5 H7)

(Jti3P,,,,,

2.95 ( d ) = 4.4 H 4

5.59 (ddd)

5.70 (ddd)

5.64 (ddd)

-24.4 (dd) -26.0 ( s ) ( J t 1 5 p = 13.8, 16.2 H z )

-27.0 (s)

-24.5 (dd) ( J t , s p = 13.2. 17.0 Hz)

-26.1 (s)

other

6.8-7.4 ( m ) , Ph

6.7-7.3 (m). Ph

6.7-7.4 ( m ) , Ph

J ) A ,HI

10.3 10.7 7.2

10.3 10.9 7 .O

10.3 11.0 6.8

6.8-7.4 (111). Ph 2.30 (s), Me 10.0 10.9 7.3

6.7-7.2 ( m ) , Ph 2.29 (s), M e 10.1 10.8 6.8

./?A, J\4,

5.70 (ddd)

5.72 (ddd)

til H7

3 1 PN M R h

15 (ppin)" Ji,r,.H7

1 . l 6 . 0.87 5.0

0.32. - I .30 3.8

Measured as Nujol mulls. I 13 1'04.

Measured as CDCll solution at 30 "C.

The A-M-H complexes have also been suggested as precursors to the metallocyclobutane e carbene-olefin complex manifold implicated in olefin metathesis (eq 4).33Despite the

CH,

-

L,M

P\CH, \ /

CH,

H

d- (4) YCH2 CH,

+ L,M,

interest in A-M-H complexes generated by these important catalyses, few such complexes have been reported. The compounds N ~ H ( v ~ - C ~ H S )LL=, ~PPh3 ' (Ph = C ~ H Sor ) PF3, prepared at low temperature, decompose irreversibly above -30 " C . Between -40 and -SO "C the PF3 complex was shown to be in dynamic equilibrium with the corresponding nickel(O)-q'-propene complex. The complex RhCIH(q3CjH!)(PF3)234 was observed at -75 OC as an intermediate in thc reaction of Rh(q3-C3H5)(PF3)3 with HCI which yields [RhCI(PF3)2]2 and propene. Byrne et a1.32have reported the I H N M R characterization of MoH(q3-C3H!)(dppe)2, dppe = Ph?P(CH2)2PPh*,and have shown that, although stable to 1 I O "C, this complex exhibits dynamic behavior similar to that observed in the nickel system. Recently the complex RuH(q3-C3Hj)(NCCH3)(PPh3)2 has also been reported.j5 In an earlier ~ o m m u n i c a t i o nwe ~ ~described our preliminary rcsults concerning an iridium( I I I ) allyl hydride complex, IrCIH(q3-ClHJ( I-Ph)](PPh3)?, which is obtained from both cyclopropane and olefin starting materials. This complex exhibits exceptional stability and provides substantiation for the incchanisms shown i n Schemes I and I I . The isolation of this

Downfield from external Me&.

d

Downfield from external 85%

complex is also of interest as it provides an example of oxidative addition, and thus activation, of C-H and C-C bonds by metal complexes.37 W e present here the results of a more intensive study of a series of Ir(l1l) v3-allyl hydride complexes, including the determination of the structure of IrCIH[q3-C3H4(1Ph)](PPh3)2.

Experimental Section IR spectra in the range 4000-400 cni-' were recorded from Nujol iiiulls using Perkin-Elmer 283 or Nicolet 7199 FT spectrometers. In the range 400-230 c m - ' samples were r u n as Nujol mulls between high-density polythcnc windows using thc formcr spcctrometcr. ' H Y M R spectra were recorded on Varian C F T - 2 0 or Hitachi PerkinElmer R20-B spectrometers. a t 80 or 60 MH7. respectively. "PI'HJ N M R spectra were measured on a Varian C F T - 2 0 instrument operating at 32.199 MHz. Melting points were measured for encapsulated samples w i t h a Mel-Temp hot-stage apparatus and arc uncorrected. Microanalyses were carried out by Micro-Tech Laboratories, Inc.. Skokie. I l l . , or Ms. Hilda Beck. Northwestern University. All reactions and manipulations involving air-sensitive materials \\ere carried out under inert atmosphere. either dinitrogen or argon. Phcnylcyclopropane, allylbenzenc. and rrrrns-a-methylstyrene were obtained from Aldrich Chemical Co. and used without additional purification. These organic substrates wcrc deoxygenated by vigorously passing a stream of dry dinitrogen through them for a t lciist I 11 imincdintely prior to use. The complex t m n s - l r C I ( N r ) ( P P h 3 ) ~ 3 ~ \ v a s prepared by the literature procedure as was [ lrCI(COT)r]234 (COT = cyclooctenc) cxccpt that this latter complex was washed with chillcd. degassed 2-propanol rather thun with methanol. The usc 01' lhc primary alcohol often resulted i n formation of / r m ~ - l r C l ( C O ) (f'Ph>)?as a minor side product in subsequent reactions. This presumably ariscs from decarbonylation of the traces of residual alcohol b) ii highly reactive Ir( I ) phosphine complex.3x.40Phosphine. arsinc, and stibine ligands wcrc obtained frum comnicrciul sources and wcrc uhed as received. Solvent\ wcrc dried and distillcd under dinitrogen prior to use. The spectral data of the new complexes arc summarized i n Table I .

Tulip, Ibers

/ q3-Allyl Metal Hydride Complexes

Preparation of IrCIH[q3-C3H4(1-Ph)](PPh3)~ (1). A. A sample of was suspended in 8.0 mL I T ( ~ ~ z s - I ~ C I ( N ~( )I (.00 P Pg,~ 1.3 ~ ) mniol) ~ of neat, degassed phenylcyclopropane (64 mmol). With stirring at room temperature the yellow solid slowly dissolved to form an orange colution from which an off-white solid was deposited. After I5 days thc solid was collected, washed with ether and acetone, and dried in vacuo. Recrystalli7ation [rom chloroforni/methanol produced colorless crystals (0.47 g, 42%), nip 19 I " C dec. Anal. (C45H40CllrP2) C, H. CI. B. A flask containing trnns-lrCI(Nz)(PPh,)2 (0.50 g, 0.64 mmol) w a s charged with 8.5 mL (64 mmol) of freshly degassed allylbenzene and the mixture was stirred at room temperature. The qellow solid rapidly dissolved to produce a deep red solution. After 8 h the solution had become pale orange and a n off-white precipitate had formed. t t h e r (20 mL) was added and workup proceeded as in A to yield 0.42 f (78%) of 1. C. The olefin Irans-p-methylstyrene was substituted for allylbcn7enc in B a n d the reaction mixture was stirred for 12 h before addition of ether. Recrystalliiiition gave 1 in 7 I %yield. D.Thecompounds [IrCI(COT)2]? ( 0 . 1 6 g . 0 . 1 8 mmol) and P P h j (0.19 g. 0.72 mmol) were vigorously stirred together as solids. Phenylcyclopropane (4.6 m L . 37 rninol) \tiis added and a deep rcd solution quick11 ensued. The iiiixturc \\;is stirred at room temperature for I8 h. at which time it had become ;i yellob\-brown suspension. The usuiil workup gave 1 in 499 lield (0.15 g ) . E. I n a similar manner [Ir(CI(COT)l]2(0.19 g. 0.21 mmol), PPhj (0.23 g. 0.86 mniol). and allylben7ene (5.7 niL, 43 mmol) were combined to form a deep red solution. The mixture rapidly became a yellow-brown slurry, After 3 h t h e reaction mixture was worked up i n the usual manner to yicld 0.32 g of 1 (85%). F. The compound rruns-i3-niethylstyrene was substituted for allylbenzene in E and the reaction mixture stirred for I2 h. Workup gave 1 in 79% yield. Preparation of Other q"-AIIyI Hydrides, IrCIH[q3-C3H4(1-PhllLZ. L = AsPhJ (2).Samples of [ IrCI(COT)2]: (0.20 g, 0.22 mmol). AsPh, (0.27 g. 0.88 nimol), and allylbenzene (5.8 mL, 44 mmol) were combined as in E. A white solid rapidly formed and the mixture was worked up after 2 h to yield 2 (0.29 g. 69%). nip 221 " C dec. Anal. (C4iH40AslCllr) C, H, CI. Complexes 3,4, and 5 ( L = SbPh3, P(p-Tob3, and 4s(p-Tol)3(p-To1 = 4-ToIyl).Respectirelyi. These complexes were prepared in a similar inanncr using the appropriate donor ligand, a reaction time o f I8 h , and a 50-fold. rather thiin I00-fold. excess of allylbenzene. Complex 3 was recrystallized from chloroform/methanol (30% yield), while 4 and 5 were recrystallized from dichloromethane/methanol in 63 and 52% lields, respectively. Complex 3: mp 198 " C dec. Anal. (CJ5H4oCIIrSb?) C . H. CI. Complex 4: mp 146 OC dec. Anal. (CilH5:CllrP?) C , H . Complex 5 : mp 166 " C dec. A n a l . (C51H+s2Cllr) C . H. Similar procedures using othcr donor ligands (PEt3, PMezPh, PUePh:. P(o-Tol)?(o-To1 = ?-tolyl). dppe. P(OPh)3) failed to yield the desired q3-a11yI metal hydride complexes. I n all reactions the residual substrate mixturc was isolated from the supernatant mixture bq viicuum distillation at room temperature Lind its ' H N M R spectrum recorded in order to ascertain the extent of any isomerization. In the absence of metal complexes no isomeri7ation wab detected for a n y of the substratea. Reaction of 1 w i t h CO. Into ii dichloroinethanc ( 5 ml.) solution of 1 (36 ing. 0.041 m i n d ) a t room temperature and pressure CO \+;is bubbled for I 5 iiiin during which time the solution became yellow. The solution w s allo\icd to stand under a C O atmosphere for an additional I 5 m i n . L p o n addition of methanol (20 mL) ;I yellow solid precipitated. Recrystallization from chloroforni/methanol ( I : I ) gave )ello\+ crystals of rrcitis-lrCI(CO)(PPh3)r (28 mg, X4Yo). The identity of this complex was verified by comparison with a n authentic sample [ V C O 1960 c1n-I. v l r CI 313. 31 1 ( s h ) cni-I]. A n a l . (C37H30CllrOP?)C. H. CI. I n a n alternntive experiment. C O was bubbled into a deuteriochloroform solution of 1 i n a n N M R tube. I n the ' H N M R spectrum of the resulting solution the resonances of the ullyl and hydrido ligands were completely replaced by those of/3-methylstyrene. as verified by comparison with an authentic sample. Reaction of 1 with PF3. PF? \+:is substituted for C O in the above proccdureh dnd irnns-lrCI(PF?)(PPh3)2 was isolated in 75% yield as a yello\+ solid. Its identity \\;is verified by comparison with a n au~ c mI - I ] . Anal. (C~61i30C1F31rP~) C. H. thentic s;iniple4"." [ v ~ ~300

4203 0-Methylstyrene was the only hydrocarbon product observed by N M R spectroscopy . Reaction of 1 with 0 2 . A saturated solution of 1 in CDC13 in an N M R tube was flushed with d r j 0 2 for IO min. The system was then sealed and stored at room temperature. The solution gradually became green and IH N M R spectroscopy indicated the slow production of 0-methylstyrene. Aiter 10 weeks the solution was concentrated to yield a mixture of 1 and the green product, which could be separated by extraction with acetone. The IR spectrum o f the green solid shows bands at I I20 and 720 cm-' characteristic of OPPh3 and is identical with that of authentic samples of the oligomeric dioxygen-decomposition products of trnns-lrCI(N2)(PPh,)2:'X and "lrCI( PPh3)2".42 Reaction of 1 with HCI. Anhydrous HCI was bubbled into a saturated CDCI, solution of 1 in an Y M R tube. The solution immediately turned qellow and in its IH N M R spectrum there were no resonances arising from I . These were replaced by those of both /3-niethylstyrene and allylbenzene. in approximatelk, equal intensity. The upfield hydride signal of I had been replaced b) two sets of triplets ofapproximatcly equal intensity at d -23.2 and -24.0 ( ? J P H= 19.5, 19.8 Hz, respectively). In a separate experiment. 0.23 g (0.2b mmol) of 1 was dissolved in 20 mL of dichloromethane. Anhydrous HCI w a s bubbled into the solution for 10 m i n and the solution \+;is concentrated. RccrSstalli7ation from bcnrcnc-ti-hexane yielded 0. I9 g of pale yellow solid (92% based on IrC12H(PPh3)?). Anal. (C?hH31CI?IrP2)C, H. CI: calcd. 8.0: found. 8.7. The I H lCMR spectrum of this solid shows hydride resonances equicalcnt to those observed above. The IR spectrum contains ii weak. broad band centered a t ca. 2260 cm-' which w'e ascribe to VI, ti.

Reaction of IrC12H(PPh3)2 with CO. Into a CH2Clz solution of IrCI?H(PPh3)2, prepared in situ from 1. w;is bubbled CO. The yellow solution rapidly decolorixd and a n ol'f-white solid began to precipitate almost immediately. .After 5 min the addition of CO was discontinued and the volume of the suspension reduced. Ether was added and the resulting suspension was filtered to Sield a n off-\+hitc solid i n quantitative yield. Anal. (C37H31Cl2lrOP?) C , H. CI. The IR spectrum shows bands at 2245 ( q r 1,). 2078 ( V C O ) . 300, and 260 ( ~ 1 ~

~ ~ 1 )

cm-1.

Reaction of 1 with HBr and CO. A saturated CH2CI2 solution of 1 w a s treated sequentially with anhydrous HBr and C O in the manner employed above for HCI and C O . The off-ahite solid obtained contains no chlorine. Anal. (C37H31Br21rOP?)C , H . Bands at 2240 ( V I r H) and 2070 (vco)cm-I are present in the IR spectrum. Collection and Reduction of the X-ray Data. Vapor diffusion of ;icetone into a dichloromethane solution of IrClH [$-CJHJ( I-Ph)](PPh3)l (1) yielded a crop of colorless. rectangular needles suitable for diffraction. Preliminary photogrLiphic data from a crystal mounted i n air indicated that the c'r)stal belongs to the orthorhombic s1steni. Sqstemntic extinction\ (Oh/. h I = 211 I : hXO, h = 212 I ) characteristic of t h e space groups D:,,lh-Ptitiin and c ' 2 ! y - P ~ ~ 2were 1n observed. Crystallographic data arc tnbulated in Table II. The obscr\,cd density is in accord v.ith four molecules per u n i t cell. Consequently in t h e centrosymmetric group D ? l j 1 6 - f t ~ u in ~ athe , absence of disorder. the molecule must contain eitlicr ;I center or ;I plane of . .4a spectroscopic data do not indicate cither t l p e of nioninictry, the polar noncentrosl mmctric group ('2, '-PI12 1 0 \\;is chosen initially. This choice u a s wbsequentlk shown to be correct ;is reasonable positional a n d thermal par:inicters, including those of tlic allylic and hydride hydrogen atoms. \+ere rel'ined successfully. Cell constants (Table I I) \\ere obtained ;is prcviouslq describedj7 b) ;I least-squares rcrincment of 1 5 reflections manually ccntercd on ;I Picker FACS-I diffr,ictonictcr. Thehe rel'lections \+crc generated using a n;irrow source and chosen l'rom diverse regions ol'reciprocal 124.5' I O s a t each end of scan with rescan option (see t e x t ) 5.0- I 60.0' 28 5 40' +h, k k , +I 28 > 40' +h. + k . +I 4346 3738 I68 0.029 0.041 I .58 electrons

rcllcctions were measured. N o significant deviations in the intensities of these standards were observed. T h e data were processed A S dcscribed p r e v i o u ~ l yusing . ~ ~ a p value of 0.04. An absorption correction was applied to the data.45The total number of independent data with F,,? > 3a(F,') was 3738. Only +k data were used in the solution and preliminark refinement of the structure. Solution and Refinement of the Structure. T h e iridium atom \+;is located froni an origin-removed Patterson synthesis. Structure factor kind Fourier synthesis calculations revealed the positions of the phosphorus and chlorine atoms. Subsequentlq. the positions of the remaining nonhydrogen atoms were obtnined by the usu:il conibination of least-squares refinements a n d difference Fourier syntheses. The function niinimi7ed was Zw((F,I - l F c l ) ' . where lFc)land IF,\ arc respectively the observed and calculated structure amplitudes and \\here M' = 4F,12/a2(Fc12). Atomic scattering factors for the nonhhdrogen atonis were taken from the usual tabulation.jh Anomalous dispersion terms \\ere not included at this stage of refinement. Thc phenyl rings were treated as rigid groups2' and restricted to ;I g o m e t r y having uniform C - C distances of 1.395 A and ideal D,,/,sqnimctr) 1:ach of the group atoiiis was refined \ 4 l t h a n individual isotropic tIIcrn1;il paranictcr. Refinement of the structure with ;ill nonhldrogcn toni is included a s isotropic bodies resulted in agreement indicch of K = 0.073 and K , = 0.104,whcre K = Z~IF,J - lFc~l/ZIF~ and , l ti,, = (XW((F(,\ (F,()'/~\~,F,')I~'. ,A t e s t \+;is next made to detcrminc the correct enantiomer for this crkhtal. Separate least-squares caIcuI3tions for the two enantiomers \\crc carried out i n u hich all of the nongroup iitoins were nlloucd to vibratc anisotropically and the iinoniiilous dispersion term> for Ir. CI. ;ind P atoms u e r c I n thcsc calculations - k data were included. I n t h e first calculation the original enantiomer refined to K = 0.054 a n d K,, = 0.087. The same least-squares cilculation for t h e altcrnatibc enantiomer yielded agrcemcnt indices of K = 0.057 and K,, = 0.090. I n view of the abscncc o f a drnmntic difference in these

1 101:15 1 July

18, 1979

agreenient indices \+K pcrfornied a subsequent comparison of Friedel pairs. Of the 138 Frieiel pairs collected. 35 (for which Fc L 2Oc and llF,(khl)l - I F , ( h X / ) ' l / I F , ( h X I ) I > 0.05) \rereconipared. O f t h i s number the trend among 34 confirmed the original enationicr ;is the correct choice. The positions of the hldrogen atonis ofthc allyl and phenyl groups were determined from ideal gcometrics Lind a C - tI distance of 0.95 A. Each 01' these 39 hgdrogen atoms \ + a s then assigned a n isotropic thermal par;imcter I .&? greater than t h a t o f the carbon atom to uhich it is attached. These atoms \vert: included in subsequent calculations as fixed contributions. The hldrogen atoni scattering factors used wcrc those of S t e n a r t e t al.iy ,After a le of least-squares refinement a difference Fourier sjnthesis revealed the position of the hjdride hydrogen atom. This atom \+:I> rel'incd as an isotropic bod) in subsequent calculations. The fixcd contribution of the phenyl hbdrogen a t o m \+as reset. based on the ne\\ positions o f t h c group carbon atonis. R a t h e r than being included a i t h the fixed contributions. the hydrogen atoms of the allbl group \\ere alloucd to var) isotropiciilll in the final cqclcs of refinement. Thcsc calcuhtioiis converged to final agrcciiicnl indiccc of K = 0.029 and K,, = 0.032. The error in an obscrLation of unit weight is 1.58 electrons. The largest peaks in the fin;.il difference Fourier s> nthesis are ~ipprouimatcl)0 . 5 7 ~ -A-3 and a r e associated uith the iridium atoni. ,\n analgsis o f Z w ( l F,,I - I F , ( ) ? a s a function of l F L ) l ,setting angles. and Miller indices reveal5 no unexpected trends. The final positional and thermal parameters of t h e rionh)drogen atoms appear in Table I I I . Table IV lists the derived p;irameters of t h e 42 group atonis. Thc root-mean-square Limplitudcsof vibration and the idenliycd positions for the phen! I hbdrogen ;itoms ;ire given in Tables V and V I . rcspcctively.'" )\ listing of the observed and calcu l a t ed structure a iii pl i t udcs is ;I Iso a v a i l a bl e .50

Results Syntheses and Reactions. When a suspension of transIrCI(N2)(PPh3)2 in neat phenylcyclopropane is stirred at room temperature the yellon solid dissolves slowly to yield an orange solution from which a n off-white solid, IrCIH[$-C3HA( 1 Ph)](PPh,), ( l ) ,is deposited. N o Ir(1ll) complex is obtained when benzene or toluene is used as solvent. Complex 1 is sparingly soluble i n TH F and chlorinated hydrocarbons but insoluble in benzene. toluene, ether, or hydrocarbon solvents. It is stable indefinitely in air in the solid state and decomposes only very slowly in solution on exposure to the atmosphere (vide infra). The most prominent feature of the IR spectrum of 1 is a strong. sharp band at 2208 cm-l, which we attribute to the iridiuin-hydrogen stretching mode. The high value of vlr t i suggests that the hydrido ligand is trans to the chloro ligand." We had initial11 reported3" that the spectrum also exhibits a band at 245 cm-I, which we ascribed to vlr (I: and took to be support for trans H-CI stereochemistry.52W e now find that the 245-cm-' band is actually an instrumental artifact and we have been unable to locate unequivocally a i'lr [ I band. Either this band is of low intensity or the Ir-CI bond is particularly weak. causing I J C~I to ~ fall below 230 cm-I. The low-field region of the H Y M R spectrum of 1 consists of a broad area of aromatic resonances and a series of multiplets consistent with a n v3-alIyl moiety.', T h e details of this spectrum are compiled i n Table I . The assignments therein have been confirmed by homonuclear decoupling and comparison ~ i t the h literature.54The observation of a set of resonances in the high-field region confirms the presence of a hydrido ligand and the doublet of doublets pattern with l J ~ values of 13.8 and 16.2 H 7 indicates that the two P nuclei are Ph

L1

I

-

H,

p

Tulip, Ibers

/ v3-AIlyl Metal Hydride Complexes

4205

Table III. Positional and Thermal Parameters for the Nongroup Atoms of IrCIH($-C3H4( I-Ph)](PPh&

I R

0.127301(131

114

o.

31.70111 I

45.3oi26)

19.3815

-2.071 1 9 )

0.6ii41

CL

0.17615111)

0.02951(17)

0.519?417)

52.718)

42.9113)

1k.64125)

-0en18)

0.16118)

Pl1)

0.2542RllOl

0.5003317l

33.616)

53.0113)

12.92126)

2.11.)

50621n 1 8 )

)

PI21

0.171291101

0.30228l19) 0.296871181

0.4105?16)

37.216)

0.00011b)

0.1667 19)

0.k7R50115)

45.31321

46.5113) 65.181

10.98124)

C I I I

CIZ)

-0.516921381

0.23P318)

0.527Q2126)

31.3122)

61.181 57.171

Cl31

0.03033144)

0.214717)

0 . 5 ~ ~ 1 ~ 1 z e i 38.2127)

H

0.10515)

o.380191

0.501ql??)

2,7116)

Hicii)

o.ooain)

0.096(12)

0.41171146)

3.31311

HOC 1 1 )

-0.02716)

0.1P2l81

0.k41ql 41)

5.2121)

HCIZI

-0.045a138)

0.315161

0. 5 F 5 ? ( ? 7 1

1.61111

0.06616)

0.12919)

0.5119’151)

O.RIZ4)

HCO)

13.8112)

-0.617)

-1.67021

2.33 1 3 1 )

-1k.71 k 7 )

-3.8117)

15.3111 1

-4.7 1 3 7 )

2.9112)

13.1l11l

-0.01311

3 - 9 11 4 I

0.46119) k.Ollk0l o.ai51

0.751 631 O.kl25)

5.61251 -3.0 1 1 9 )

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A E S T 1 7 A T E 0 S T A Y O A R O DEVIATIONS I N TL(F LFAqT T l G N I F I C A N T F I G U P E I S I A95 GIVEN I N PAREYTHESES I q T Y I S b Y D 4 L L 9UESEOUENT TABLES. F O R M OF THE A M I S O T P O P I C THERNAL E L L I P S ‘ l l Q

A 9 E THE THERRAL COEFFICIENTS X 10

.

TSI

EWPC-18114

2

t8ZZK

2

*833L

8

THE

t Z B 1 2 ~ ~ r 2 3 1 3 H L + 2 8 2 3 ~ LTHE ) l ~ OULNTITICS G I V E d I N THE TABLE

Table IV. Derived Parameters for the Rigid Group Atoms of IrCl H[$-C;H4( I-Ph)](PPh,)2 2

2

:!% * * * * * * 8 : * 8 * * *I* ** * *: * * 8 * $ 8 * * * * * * * *: * * ** **** *I* R* t A** *** * * * AT011 **** * *** *I * **:* * * * * * * * * ** * * * *: * ** * * * ** * * * * * f * * * * * * * ** * * * 3,A * **** $ 1 I

I

I

nonequivalent and both cis to the hydrido ligand.55The nonequivalence of the phosphorus nuclei in 1 is substantiated by its 3 1 P ( l H N } M R spectrum, which consists of an AB quartet (Table I ) . The small value of 2 J p p , 5.0 Hz, indicates that the phosphine ligands occupy mutually cis positions. The combined results of these stereochemical indicators suggest that the structure of 1 corresponds to isomer I , which is corroborated by the solid-state structure discussed below. Complex 1 is also obtained in improved yields from the reactions of allylbenzene or trans-0-methylstyrene with transIrCI(N?)(PPh3)?. In each case the product arises from oxidative addition of an Ir(l) center to an allylic C-H bond. Presumably this insertion is preceded by the formation of an

Ir( [)-olefin complex. In these intermediate complexes the allylic C-H bonds are held in proximity to the metal center by olefin coordination. The need for such a juxtaposition has been well d o c ~ m e n t e d . ~The ’ formation of complex 1 from these olefins lends credence to the proposed allyl metal hydride mechanism for metal-catalyzed olefin isomerizations, as these coordination and oxidative addition steps are entirely analogous to the initial portion of this mechanism as depicted in Scheme I I . At the time of our initial observation of these reactions the formation of MoH($-CjH!)(dppe)? by addition of a Mo center to propene was the only analogous reaction reported.j? I n the interim similar reactions of olefins with a Ru(0) complex have also

4206

Journal of the American Chetnical Society

A more convenient iridium-containing starting material for the preparation of 1 is [IrCI(COT)r]l. Mixtures of PPh3 (4 mol) and [IrCI(COT)2]2 ( 1 mol) are thought to contain monomeric “lrCI( PPh3)2”,4’ possibly solvated, which is analogous to the reactive species generated by transIrCl(N2)(PPh3)2.38 The [IrCI(COT)2]2-PPh3 mixture affords shorter reaction times as a result of increased solubility. Cleaner products in enhanced yields are also obtained as trans-lrCI(N2)(PPh3)2 invariably contains trans- IrCI(C0)(PPh3)2 as a contaminant not found i n carefully prepared samples derived from [ IrCI(COT)>]1. Thus phenylcyclopropane, allylbenzene, and trans-,h’-inethylstyrene each react with “ l r C l ( P P h 3 ) ~ ”generated in situ to form 1 in high yield. Use of [IrCI(COT)1]2 also allows the introduction of donor ligands other than PPh3, whereas the dinitrogen complex can be satisfactorily obtained only with this phosphine. Thus we have also prepared complexes 2-5, i n which AsPh3, SbPh3, P(p-Tol)i, or As(p-Tol)3 are substituted for PPh3 in 1 by the reaction of [ IrCI(COT)2]2 with allylbenzene and the appropriate ligand. The similarity of the I R and N M R spectra of these complexes with those of 1 suggests that they also adopt geometry 1. Attempts to prepare additional A-M-H complexes using PEt3, PMelPh, PMePhz, P(o-Tol)3, dppe, or P(OPh)3 as ligands were unsuccessful. I n the course of these synthetic reactions we routinely monitored the residual substrate to ascertain the extent of any isomerization. Although no attempt has been made to quantify these isomerizations, the general trends are interesting. The reactions of allylbenzene with trans- IrCI(N?)(PPh3). or “lrCI(PPh3).” both yield substrate which has been isomerized to P-methylstyrene to the extent of approximately 20 f 5%. T h e analogous reactions with P-methylstyrene produce no observable isomerization. This is not surprising as the equilibrium concentation of the terminal olefin, allylbenzene. is only 0.05% a t 25 oC.5hThe extent of isomerization of allylbenzene is very dependent on L in the “IrCIL,“ systems, increasing as the basicity of L decreases. When L is the highly basic PEt3 no isomerization is observed even after 1 week. The activity of the system increases as the ligands become less basic” in the order PMe1Ph < PMePh. < PPh3. With the weakly basic ligand P ( 0 P h ) j complete conversion of allylbenzene to p-methylstyrene is rapidly achieved. Ligands such as P(OPh)3 lower the electron density at the metal center thereby stabilizing the lower formal oxidation state. This should promote the reductive elimination of hydrido and allyl ligands from an intermediate A-M-H complex and facilitate the isomerization if this reductive elimination is the rate-determining step. Highly basic ligands such as PEtj decrease the effective oxidative state of the iridium center in “lrCIL2“ and thus favor both olefin coordination and oxidative addition while prevent i ng reductive e I i m i n a ti on. T ha t disc ret e A - M - H complexes are not obtained with these alkyl phosphines ma) result from interactions between the metal and C-H bonds of the ligands, as has been previously demonstrated.3xW e have noted3” that no P-methylstyrene or allylbenzene is observed i n the preparation of complex 1 from trans- IrC1(%2)(PPh3)2 and phenylcyclopropane. This is also the case &hen “lrCI(PPh3)2” is used. The facile formation of 1 from either olefin suggests that the complex possesses high stability and that any small concentration of olefin would be quickly scavenged by “lrCI(PPh3),”. The thermodynamic stability of 1 is consistent with its lack of catalytic activity. Mixtures of complex 1 ( I mol) and allylbenzene ( I 00 mol) in deuteriochloroform showed no production of p-methylstyrene over the course of I O weeks a t 40 “ C , as judged by H N M R spectroscopy. Reductive elimination of the hydrido and allyl ligand appears to be unfavorable. This is substantiated by the variable temperature (-60 to 60 “ C ) H and ? ‘ PN M R spectra of 1-5, which indicate that no

/

101:15

1 J u l y 18, 1979

dynamic equilibria with corresponding Ir( I)--olefin complexes take place, in contrast to the results observed with NiH(v3C3 H s)( PF3) and MoH (v3-C3H 5 ) ( d ~ p e ) ? . ~ ’ When heated at 60 “C for prolonged periods complex 1 does slowly decompose to yield P-methylstyrene. This olefin is also produced in the reactions of CO or PF3 with 1. N o such reaction takes place between 1 and propene ( 1 atm) or triphenylphosphine. Here the stronger r-bonding ligands stabilize the metal in the 1 oxidation state, as trans- IrCI(CO)(PPh3)2 and trans- IrCI(PF3)(PPh3)2, respectively, and thereby facilitate reductive elimination of olefin from I . Solutions of 1 in chloroform are very slowly decomposed by dioxygen, again yielding i3-methylstyrene. In none of the reactions is the production of allylbenzene observed. The direction of these processes suggests that if the v3-allyliridium(111) hydride complex were less stable it could serve as a good catalyst for the isomerization of primary to internal olefins, as depicted in Scheme l l . Treatment of complex I with HCI ( 1 atm) at 25 “ C produces both P-methylstyrene and allylbenzene in approximately equivalent amounts. This result, in contrast to the formation of only the former, thermodynamically favored olefin in the reaction of 1 with C O or PF3, indicates that the HCI reaction is kinetically controlled. A similar result has been reported by Nixon and Wilkins’l for the reactions of Rh($-allyl)( PF3)3 complexes with HCI. At -75 “ C the intermediate complex RhCIH(q3-C3Hj)(PF3)~is observed but it decomposes to yield propene on warming. I n those cases involving asymmetric $-allyl ligands, roughly 1 / 1 mixtures of the two possible olefins are observed, regardless of the nature of the allyl substituents. These results clearly indicate that the formation of olefin does not proceed i n the expected manner, ix., by reductive coupling of hydrido and qi-allyl ligands in a complex analogous to I I . The most stable vl-allyl group corresponds to the ther-

+

Ph3P\

H

Ph HC

Ir-

C’

I

%-I

H,

PhP ,’I C1

I1 modynamically favored olefin and it is initial rearrangement ( q 3 - + VI-allyl) which directs the course of the reaction. A number of mechanistic pathways, including outer-sphere protonation of the $-allyl group, are conceivable but labeling studies are required to discriminate among these. Two hydride complexes are also formed in the reaction of HCI with 1, as judged by the high-field ‘ H N M R spectrum of the iridium-containing product. Two hydride signals are present, again of approximately equivalent intensity, as binomial triplets with H - P coupling constants of about 20 Hz. Thus in both species the hydrido ligand is cis to two equivalent phosphine ligands. The absence of H-H coupling indicates that the signals arise from different species rather than from a single dihydride complex. The IR spectrum of this mixture contains ;I broad band centered at 2260 c n - ’ which we attribute to u l r 1 1 . These results suggest that this reaction proceeds as in cq 5 . Complexes of stoichiometry IrC12H(PPh3)2 have also

-

IrCIH[$-C3HJ( I-Ph)](PPhi)?

+ HCI

+ tu-lrCI?H(PPh3)2] + H + 8-IrCI2H (PPh3),]

[PhCH=CHCH?

[ PhC H :C H=C

(5)

2

been reported as the products of the reactions of IrH3( PPh3)2’” or trans-lrCI(N2)(PPhj)~””.”’ with HCI. These previously reported complexes react with CO to produce compounds of stoichiometry IrC12H(CO)( PPh3)2 whose stereochemistries

Tulip, lbers

/ q3-AIlyl Metal Hydride Complexes

have been assigned as 11161 and IV,62where P represents triphenylphosphine.

VI+H

Vco Vk-ci

H

H

h

c

I11

Iv

0

2240 cm-' 2024 313,265

2154 cm-I 2003 320

Vaskah2reported that the HCI adduct of trans- IrCI(C0)(PPh3)2 also adopts structure I I I. These structural assignments are based upon the empirical rules for v l r . . ~and vir-c1 outlined by Chatt et a1.5' and Jenkins and Shaw.52Thus in 111, the high value of vlr-k+ and the low value of vlr.cl indicate that the CI and H ligands are mutually trans while the remaining value of vlr.c1, 3 I3 cm-I, is in the region for CI trans to C O . For IV the decreased value of vir H suggests that the hydrido ligand is not trans to a halide while the single high value of v l r - ~ at l 320 cm-' is consistent with mutually trans chloro ligands. Structure IV is further supported by the observation of H-CO vibrational i n t e r a ~ t i o n The . ~ ~ IrCI2H(PPh3)2 mixture obtained from 1 and HCI reacts with C O to yield a single complex of stoichiometry IrCIlH(CO)(PPh3)2 unlike either 111 or I V . Thus neither CY- nor P-lrC12H(PPh3)2 is equivalent to those complexes previously reported. The IR spectrum of the new complex exhibits bands a t 2245, 2075, 300, and 260 cm-l which we attribute to vlr-tj, U C O , and vlr-ci(2), respectively. The high value for v l r - ~in conjunction with the low value of vlr-c1at 260 cm-I again suggest a trans H-Ir-CI array. Of the six possible octahedral isomers of IrCI?H(CO)(PPh3)2 four are thus eliminated, leaving only structures 111 and V. The I'CO

H

C1

V and remaining vlr-c1 values of the new complex are clearly dissimilar to those reported by Vaskah2for 111. Thus the new complex probably adopts structure V, although zqr CI at 300 cm-I is at slightly higher energy than is usually observed when CI and PR3 ligands are trans.52 In the new complex the increased value of vco with respect to that of I I I is consistent with a trans P-CO stereochemistry, as the T acidity of the phosphine ligand allows it to complete more successfully for electron density than the CI ligand thereby decreasing the back-donation to C O . The low solubility of these complexes prohibits further comparisons based on ' H and 3 ' P N M R spectra. The novel feature of structure V is the cis coordination of the phosphine ligands. A trans arrangement for these bulky groups, as in I11, should be favored on steric grounds. This is consistent with the observation that upon standing in benzene or acetone at room temperature V slowly isomerizes to I 1 I , as verified by comparison with an authentic sample.h2 In order to determine the stereochemical course of the reaction of HCI with 1, we substituted HBr to ascertain which CI ligand in V is retained from 1. Subsequent treatment with C O produced a complex which contained no chlorine as determined by the absence of vlr CI bands in its IR spectrum. This was verified by elemental analysis. Other bands in the IR spectrum at 2240 and 2070 cm-I, ulr t i and V C O . respectively, suggest that the complex is IrBr?H(CO)(PPh3)2 with a structure analogous to V. The facile replacement of the CI l i -

4207 gand from 1 is presumably a result of the trans labilization of the hydrido ligand. The formation of two products of apparent stoichiometry IrC12H(PPh3)2 from the reaction of 1 with HCI is unusual. Recently a number of five-coordinate Ir( 111) complexes have been prepared.61sh4-66The stability of the complexes requires the presence of a ligand of high trans influence, such as the hydride here. Although no structural reports have appeared, these complexes are proposed, by analogy with two similar rhodium complexes, R h l l ( C H 3 ) ( PPh3)lhi and RhClzH [ P ( n - P r ) z ( t - B ~ ) ] l , to ~ ~ adopt square-pyramidal geometries with the ligand of highest trans influence occupying the apical site. This is substantiated by the observation0i.64J'6 that addition of a nucleophile occurs trans to the ligand of highest trans influence, which would correspond to attack at the open coordination site below the basal plane. Our results are not readily accommodated within this scheme. The IH N M R spectrum of the IrCI>H(PPh3)2mixture indicates that the hydride complexes each contain equivalent phosphines in positions cis to the hydrido ligand. The reaction of C O clearly does not proceed by addition trans to the hydride, the ligand of highest trans influence, and the retention of cis phosphine ligands in the C O adduct strongly suggests that these ligands are also cis in the IrC12H(PPh3)2 complexes. These results are incompatible with square-pyraniidal structures for the IrCI?H( PPh3)1 complexes. One possible explanation is that the IrCI?H(PPhj), mixture contains dimeric rather than monomeric complexes. The two complexes would then be V I and V I I . I n each complex the phosphine ligands are in the H H H ci

Cl

I

H

VI VI1 correct environment and addition of C O will cleave one of the bridge bonds affording the carbonyl adduct observed. This explanation is supported by the pale yellow color of the IrCIlH(PPh3)2 mixture. Five-coordinate Ir( 111) complexes generally are intensely colored.64 I n contrast to these results the HCI adduct of trans-IrCI(N2)(PPh3)r conforms to the criteria for a true five-coordinated species. The trans phosphine ligands in this complex require that in a dimer such as V I or VI1 the hydrido ligand be trans to a chloro ligand and thus the trans bond weakening effect of the hydrido ligand would retard dimer formation. A more complete elucidation of these results must await further study. Description of the Crystal and Molecular Structure of I ~ C I H [ O " - C ~ H ~ ( ~ - P ~(1). ) ] ( The P P ~crystal ~ ) ~ structure of 1 consists of four well-separated niolecules with a closest intermolecular Ha-H approach of 2.24 A. A stereoscopic packing diagram is shown in Figure 1 uhile Figure 2 presents a stereoview of the molecule. Intrnmolecular distances and angles are compiled i n Table V11. The inner coordination sphere is depicted in Figure 3, which also contains atom labels and selected bond distances. The observed structure is in accord with that predicted from spectroscopic data. The coordination geometry is pseudooctahedral with cis phosphine ligands and an q3-a11y1 moiety which occupies two coordination sites. The Ir-P bond lengths are normal. The small difference between these values i h presumably a result of the differing trans influences of the C( 1 ) and C(3) extremes of the allyl group. The Ir-C( I ) bond is stronger than that of lr-C(3), as judged by thc Ir-C distances, and thus exerts a stronger trans influence. thereby causing Ir-P( 1 ) to be longer than lr-P(2). The hydride hydrogen atom was easilq located and readily refined to a chemically reasonable position. The H-lr-CI angle

4208

Journal of the American Chemical Society

/

101:15

1 July

18, 1979

Figure I . A stereoview of the unit cell of 1, IrCltl[q3-C3H4(l-Ph)](PPh3)*.Phenyl and allylic hydrogen atoms have been omitted for clarity. The x axis is vertical from bottom to top, the). axis is horizontal to the right. and the z axis is perpendicular to the paper going away from the reader. The vibrational cllipsoids are drawn at the 20% level here and i n the following figure.

Figure 2. A stereoview of the overall molecule of IrCIH[q3-C3H4(I-Ph)](PPh3)*. Phenyl hydrogen atoms have been omitted for clarity. A portion of the labeling scheme is included.

A have been found."

Figure 3. /Z perspective view of tlie inner coordination sphere o f IrCIH[q3-C3H4( l-Ph)](PPh3)*. Thc numbering scheme and selected distances are shown. The vibrational ellipsoids arc drawn a( the 50% IcveI.

of 176 (3)O is consistent with the spectroscopic prediction of trans H and CI ligands. The Ir-H distance,'l.5 ( I ) A, may be somewhat shorter than that observed in other Ir( I l l ) hydride c o m p l e x e ~ " ~and ~ ~ ) transition-metal hydrides i n general, approximately 1.7 A . 7 1This short bond length may reflect the weak trans influence of the chloro ligand as does the high value of U l r 11.5' The Ir-CI bond is extremely long, 2.549 (2) A, presumably as a result of the trans influencc of the strongly bound hydrido ligand. Iridium( Ill)-chlorine bond distances from 2.33 to 2.S1

X h W e believe that this bond in 1 is the longest of its kind reported. A longer terminal Ir-CI distance, 2.599 A, has been determined only for the five-coordinate Ir( I ) complex IrCI(C~H,~)(Ph~PCHl(C~H80~)CH~PPhl).X7 The lability of the CI ligand in 1, as shown by its facile substitution by the Br- ion, is consistent with this long bond. Considerable use has been made of iridium-chlorine stretching frequencies to diagnose the trans bond weakening effect of ligands in octahedral d h Ir(ll1) complexes. Values of u l r CI decrease for trans ligands in the order CI-, CO, PR3, alkyl- or aryl-, H-.52,62,65 A similar correlation with VR,, CI has been demonstrated for Ru(l I)-CI systemsXX and we have recently shownxythat the Ru-CI bond length can also be used as an indicator of the trans influence of a ligand. The Ir(ll1)-CI bond lengths may be used similarly. Table V I l l contains a comparison of iridium-chlorine stretching frequencies and bond lengths and shows that, as expected, the bond length increases as the value of vir ~1 decreases. T h e geometry and dimensions of the allyl group in 1 are normal. The C( I)-C(2)-C(3) angle, 120.8 (8)", is as expected for an sp'-hybridized carbon center and the dihedral angle between the planes defined by the three carbon atoms of the allyl group and C( I ) , Ir, and C(3), 73.1 ( S ) O , is consistent with that found i n other V-?-allyl complexes, 70-80°.')0 Useful comparisons can be made between the structures of 1 and another octahedral Ir( I l l ) q3-allyl complex, [lrCl($C3H5)(CO)(PMe?Ph)?][PF6I7" (6). As is characteristic of $-allyl complexes the metal-carbon (central) distances, 2 . I78 (6) and 2.24 ( 1 ) A, in complexes 1 and 6, respectively, are shorter than the respective metal-carbon (terminal) distances.

Tulip, Ibers

/ q3-AllylMetal Hydride Complexes

4209

Table VII, Selected Distances (A) and Angles (deg) in IrCIH[q3-C3H4( I-Ph)](PPh3)2 Ir-CI I r--P(1 ) lr-P(2) Ir-C(1 j lr-C(2) lr-C(3) Ir-H C ( I )-C(2) C(2)-C(3) C ( 3 ) - C ( I)Ph

2.549 (2) 2.321 ( I ) 2.305 ( I ) 2. I96 (7) 2.178 (6) 2.276 ( 6 ) 1.5 ( I ) 1.378 (12) I .427 ( 8 ) 1.483 (8)

0.81 (13) 0.90 (9) 0.96 (7) 0.98 ( I O ) 1.835 (5) 1.843 ( 4 ) 1.842 (5) 1.870 (4) 1.825 (5) 1.848 (5)

Ir-HIC( I ) Ir-H2C( I ) H -C(2)

2.49(12) 2.80 ( I O ) 2.44 (9)

2.71 (8) 2.59 (9)

Cl-lr-P( I ) CI-lr-P(2) Cl-lr-C( I ) CI-lr-C(Z) CI-lr-C(3) CI-lr-H P( I )- Ir-P( 2) P( I)-lr-C( I ) P( I )-I r-C( 2) P( I ) -1r-C( 3) P( I )-I r-H P(2)-lr-C( I ) P(2)- lr-C(2) P( 2)-- lr-C(3) P( 2)- Ir- H C ( I)-lr-H C(Z)-lr-H C( 3)-lr-H

86.7 ( I ) 103.8 ( I ) 83.2 (3) 101.3 ( 2 ) 86.1 ( 2 ) 176 (3) 101.6 ( I ) 162.9 ( 2 ) 133.7 ( 2 ) 99.5 ( 2 ) 89 (3) 94.3 (2) 119.8 (2) 157.1 ( 2 ) 78 (2) 101 (3) 81 (3) 94 ( 3 )

Bond Anglca C ( I ) - l r - C ( 2) C ( I)-lr-C(3) C(Z)-lr-C(3) C(I)-C(2) C(3) C ( I)-C(2)-HC(Z) C ( 2 ) - C ( I )-H IC( I ) C(2)-C( I)-H2C( I ) HIC( I)-C(I)-HZC( I ) C(3)-C(2)-HC(Z) c ( 2)-C( 3 ) -C( I ) PI1 C(2)-C(3)-HC(3) C ( I)Ph-C(3)-HC(3) C ( l ) R ( I)P( I)-P( I ) - C ( I ) R ( Z ) P ( I ) C(I)R(I)P(I)-P(I)-C(I)R(3)P(I) C ( I ) R ( 2 ) P ( I)-P( I ) -C( I ) R ( 3 ) P ( I ) C ( l ) R ( I)P(Z)-P(2)-C( I)R(Z)P(Z) C ( I ) R ( I )P( 2)-P( 2)-C( I ) R ( 3)P( 2) C ( 1 )R(Z)P(2)-P(2) C ( I ) R ( 3 ) P ( 2 )

36.7 (3) 66.1 ( 3 ) 37.3 (2) 120.8 (8) I22 (4) I 1 2 (8) I22 (6) I I7 (9) 116 (4) 120.6 (6) I15 (5) 118 (S)

103.3 (2) 101.4 ( 2 ) 103.0 (2) 100.1 ( 2 ) 99.5 (2) 107.1 ( 2 )

lnterplandr Angle\ 74.7 (7) 73.1 (8) 52.6 ( 7 ) I44 ( I O )

However, i n complex 1 there is an additional asymmetry, as the lr-C(3) distance, 2.276 (8) A, is significantly longer than that of lr-C(l), 2.196 (7) Complex 1 also exhibits an asymmetry in the C - C bonds of the allyl group (C(I)-C(2) = 1.378 (12), C(2)-C(3) = I .427 (8) 8,);these have equivalent values of 1.38 (3) and 1.40 (3) 8, in complex 6. The long bonds in 1 involve C(3) and presumably result from phenyl substitution. However, there is little conjugation between allyl and phenyl groups, as indicated by a dihedral angle of 52.6 (7)' and C(3)-C( I)PhY1distance of 1.483 (8) A. These structural variations do not reflect the incipient formation of an ql-(a)allyl complex such as 11. Unlike the observed results, the C(2)-C(3) bond length i n such a complex would be shorter than that of C ( l)-C(2). This is also consistent with the lack of dynamic equilibria in 1. I n contrast, note that the bondlength variations i n 6, although only possibly significant, are consistent with the incipient formation of an ql-allyl group and that this process readily The bonding parameters are generally consistent with the q'-allyl group of 1 being bound more strongly than that of 6. In conjunction with the structural determination of complex 6 we have described a set of parameters which facilitate comparisons between the geometries of q3-allyl complexes.7(>Although these parameters were designed to deal only with unsubstituted q3-allyl systems, their values for complex 1 are consistent with the observed trends. The value of D, the distance from Ir to the center of mass 0 of the C3 allyl group, for

Table VIII. Iridium-Chlorine Stretching Frequencies and Bond Lengths in Octahedral I r ( I I I ) Complexes

CICO I'R3 Lilkyl-

H-

or aryl-

310-330 300-3 15 265-280 250-260 245-250

2.33-2.36h 2.3 7 - 2.4 1 2.42" 2.44-2.5 I e 7 C49J ('

Rcferenccs 52. 62, and 65. Rcfcrences 72 and 73. References 74-79. d Reference 80. '' References 74 and 8 1-86. J This work. i]

i

1, 1.95 A, is in accord with those found in other dh complexes; e.g., the complex Ru(q3-C3H5)I(PPh3)2-C7HXy2has a value of I .94 A for each allyl group. Similarly, a comparison of the values of D and the C ( I)-C(2)-C(3) angle for 1 shows that 1 falls exactly on the least-squares line generated for other dh complexes. Consistent with the /5' angles of 90-93' found i n other q3-allyl complexes, the value of /5', the angle between the I r - 0 and the C( I)-C(3) vectors, in 1 is 91.9'. The orientation of the allyl group is such that atom C(2) is closer to the hydrido than to the chloro ligand. E:arlier36 we pointed out a close contact between atoms C ( 2 ) and H (Table VI I) and suggested a potential decomposition pathway which could directly produce an iridium metallocyclobutane complex by hydride migration to atom C(2). Subsequent reductive C-C

4210

Journal of the American Chemical Society

coupling would then yield phenylcyclopropane. As discussed above, only 0-methylstyrene or allylbenzene, produced by hydride transfer to a terminal carbon atom, have been observed. This lack of cyclopropane formation is consistent with symmetry restrictions recently outlined by M i n g ~ s . ~ ~ T h e positions of the hydrogen atoms of the allyl group are of interest as we know of only one other complex, V(q3C3Hj)(C0)3(dppe)'0 (7), for which these atoms have been refined. The bond angles of the allyl group, ranging from 1 I 2 to 122" (av I19"), are as expected for sp'-hybridized carbon atoms. However, there are deviations from ideal geometry as the planes H I C ( l ) - C ( I ) - H 2 C ( I ) and HC(3)-C(3)-C(l) Ph form angles of 36 ( I O ) and 30 (6)" with the C ( I)-C(2)-C(3) planes. Similar values of 26 and 27" were found in complex 7.9" The C ( 1)Ph and four hydrogen atoms deviate (A) from the allyl plane as follows. HC(2) (-0.18)

I

H2C(1) (-0.02)

C(1)Ph (0.01) h ( 2 ) ( 0 0 )

\A \C(1) / C(3) (0.0) (0.0)

I

HC(3) (0.43)

I

HlC(1) (OA4)

Errors in these displacements are approximately 0.1 A. T h e Ir atom is displaced - 1.79 A in this scheme. The most striking feature is the bending of the anti protons, H lC(1) and HC(3), away from the metal. For complex 7 analogous displacements of 0.43 and 0.45 A were r e p ~ r t e d . T ~ h" e other displacements for complexes 1 and 7 are also similar. The syn protons of 7, corresponding to H 2 C ( I ) and C ( I ) P h , vary only 0.02 and -0.01 A from the C ( I)-C(2)-C(3) plane. Interestingly, in both 1 and 7 the hydrogen of the methine carbon atom, HC(2), is displaced toward the metal center, by -0.1 1 A in 7. Despite the bending back of the anti protons, close contacts exist between these atoms and the metal centers in both complexes l (Table VII) and 7. Franke and Weiss'O have suggested that this feature is responsible for the shielding experienced by these protons, a characteristic of the I H N M R spectra of q3-allyl complexes.

Discussion Recently the reactions of transiton-metal complexes with small-ring organic molecules have received intensive examination.) A common feature in many of these reactions is metal insertion into a three-membered ring to form a metallocycle, ;I number of which have been isolated.' A prerequisite for ring opening is the availability of two adjacent, vacant metal coordination sites. The complexes trans- IrCI(NZ)(PPh3)2, because of the lability of its N ? ligand, and ' ' l r C l ( P P h ~ ) ~are " such precursors. Phenylcyclopropane does react with these complexes, but iridium metallocyclobutane complexes have not been forthcoming. The isolation instead of an q3-allyl metal hydride complex, 1, is readily explained by Scheme I and lends credence to this mechanism for cyclopropane isomerization. Although we have been unable to detect a transient nietallocyclobutane complex, we believe that such a complex is initially formed by insertion of the iridium atom into the 1,2 bond of the cyclopropane. The complex then would rapidly undergo $-hydrogen abstraction to yield 1, presumably by intermediate formation of an ql-allyl complex such as I I . Insertion into a bond adjacent to the phenyl ring is consistent with the preferred rcgioselectivity observed in all other metal-promoted phenylcyclopropane ring opening^.^.'^^."^ yh An earlier suggestionI5 that metal atom insertion occurs at the least substituted edge has recently been refuted by Puddephatt et al.9hThey have shown that platinum atom insertion into the 1,2 bond is kinetically controlled. The observed product. which corresponds to insertion into the 2.3 bond, is thermodynamically favored

/

101.15

/

July 18. 1979

and results from rearrangement of the initially formed metallocycle. The preferential 1,f-bond cleavage of phenylcyclopropane probably arises from initial metal-aryl interaction which positions the cyclopropane near the metal center and thereby facilitates the opening of the adjacent bond. Here, as in a number of other case^,^^^^ the ring-opening reaction is promoted by the presence of phenyl substitution. This reactivity enhancement is presumably also a result of the initial metalaryl interaction responsible for the direction of ring opening. Note, however, that, although unsubstituted cyclopropane does not react with trans-IrCI(Nl)(PPh3)2 or "lrCI(PPh3)2" under the conditions described here, an q3-(C3Hs) hydride complex analogous to 1 is formed using an alternative p r ~ c e d u r e . ~ ' As described above, no b-methylstyrene, the olefin expected from reductive elimination of the hydrido and allyl ligands from 1, is observed in the phenylcyclopropane reaction. This result is explained by the preferential interactions of the Ir( I ) complexes with olefinic substrates. Thus any 13-methylstyrene produced as in Scheme 1 would quickly be coordinated and 1 would be re-formed. Small amounts of a-methylstyrene have been observed. Unlike allylbenzene or p-methylstyrene this olefin does not react with trans-lrCI(Nz)(PPh3)? or "lrCI( PPh3)2" under the conditions of the phenylcyclopropane reaction. Thus once pr0duce.d a-methylstyrene will not be scavenged. This olefin presumably arises as in Scheme I from the 2-phenylallyl isomer of 1. In fact this complex has been detected in small amounts. These observations are the subject of ongoing investigations. The formation of 1 by apparent 0-hydrogen abstraction from an initially formed iridium metallocyclobutane complex contrasts to the isolation of a number of these complexes. Although a similar facile abstraction of a cyano group from the 8 position in a series of oxygen-containing metallocyclobutanes has been observed,98 hydrogen abstractions have not been observed in metallocycles formed in reactions of Pt,'".' 1 . 1 5 . 1 h Rh,7."."".'00 and Fe14 complexes with cyclopropanes. In a number of these cases the formation of oligomeric products, involving p-chloro ligands, blocks the vacant coordination site required for P-eliminationIx while in others the reaction is subject to stereochemical restraints, e.g., the production of anti-Bredt olefins.5 The bulk of the ligands in the metallocyclobutane precursor to 1 may prevent dimerization. The iridium and P-hydrogen atoms in such a metallocycle would be in a syn-periplanar configuration, exactly that required for /3-elimination. The lack of catalytic olefin isomerization using complexes 1-5 presumably arises from their exceptional stability as octahedral Ir( I I I ) complexes.4oThe strong bonding of the iridium atom to both hydrido and allyl ligands has already been noted. However, additional factors must also be important as we have rccently synthesized other Ir( I I I ) $-allyl hydride complexes from propene or cyclopropane, isobutene, and I-butene or rran.s-2-buteney7and find that the stability of these complexes is much less than that of 1-5. The phenyl group o f t h e allyl ligand i n 1-5 may be the cause of their stability. The formation of complexes 1-5 from allylbenzene and P-methylstyrene strongly supports the allyl metal hydride mechanism for olefin isomerizations (Scheme I I ) . Although the complexes themselves do not act as catalysts i n this reaction. their production of p-methylstyrene upon treatment with C O or PFj'is in accord with the expected results of Scheme I I , in that the thermodynamically favored internal olefin results. I n complexes 1-5 strong ligands, such as CO or PF3, are required to cause reductive elimination. On the other hand, in allyl metal hydride complexes in which the metal-hydride and metal-allyl bonding is not as strong as that in complexes 1-5, this reductive elimination will more readily occur. Thus, a less efficient incoming ligand, for example, a terminal olefin, could be effective with less stable allyl metal hydride complexes. I n

Tulip, Ibers

/ q3-AllylMetal Hydride Complexes

this manner the catalytic cycle for converting terminal to internal olefins shown in Scheme I1 would result.

Acknowledgments. This work was kindly supported by the National Science Foundation (Grant CHE 76-10335). W e are indebted to Matthey-Bishop, Inc. for the generous loan of precious metals used in our studies. The authors thank Professor E. 0.Sherman for a preprint of his recent results. of

Supplementary Material Available: Root-mean-square amplitudes v i b r a t i o n ( T a b l e V ) , idealized positions f o r the phenyl hydrogen

atoms ( T a b l e V I ) . a n d a l i s t i n g o f observed a n d calculated s t r u c t u r e amplitudes (28 pages). O r d e r i n g i n f o r m a t i o n is given on any current masthead page.

References and Notes (1) For a recent review see Bishop, K. C. Ill. Chem. Rev. 1976, 76, 461486. (2) For reviews of transition metal catalyzed olefin isomerizations see (a) Tolman, C. A. In "Transition Metal Hydrides", Muetterties, E. L., Ed.; Marcel Dekker: New York, 1971; pp 271-312. (b) Hubert, A. J.; Reimlinger, H. Synthesis 1970, 2, 405-430. (c) Davies, N. R. Rev. Pure Appl. Chem. 1967, 77, 83-93. (d) Orchin. M. Adv. Catal. Relat. Sub]. 1966, 76, 1-47. (3) Paquette. L. A.; Gree, R. J. Organomet. Chem. 1978, 746, 319-329. (4) Beach, D. L.; Barnett, K. W. J. Organomet. Chem. 1977, 142, 225232. (5) Salomon, R. G.; Salomon, M. F.; Kachinski, J. L. C. J. Am. Chem. SOC. 1977, 99, 1043-1054. (6) Wilberg, K. B.; Bishop, K. C. Ill Tetrahedron Lett. 1973, 2727-2730. (7) McQuillin, F. J.; Powell, K. G. J. Chem. SOC..Dalton Trans. 1972, 2129-2133. ( 8 ) Katz, T. J.: Cerefice, S.A. J. Am. Chem. SOC.1971, 93, 1049-1050. (9) Rajaram, J.; Ibers, J. A. J. Am. Chem. SOC. 1978, 700, 829-838. (10) AI-Essa, R. J.; Puddephatt, R. J.; Quyser, M. A.; Tipper, C. F. H. J. Organomet. Chem. 1978, 750, 295-307. (11) Brown, D. B.; Viens, V. A. J. Organomet. Chem. 1977, 742, 117-121. (12) Yarrow, D. J.; Ibers, J. A.; Lenarda, J.; Graziani, M. J, Organomet. Chem. 1974, 70, 133-145. Lenarda, M.; Ros. R.; Graziani, M.; Belluco, U. ibid. 1974, 65, 407-416. (13) Johnson, B. F. G.; Lewis, J.; Tam, S. W. J. Organomet. Chem. 1976, 705, 271-279. (14) Moriarty, R. M.; Chen, K.-N.; Yeh, C.-L.; Flippen, J. L.; Karle, J. J. Am. Chem. SOC.1972, 94,8944-8946. (15) McQuillin, F. J.; Powell, K. G. J. Chem. Soc., Dalton Trans. 1972, 2123-2129. (16) Tipper, C. F. H. J. Chem. SOC. 1955, 2045-2046. Adams, D. M.; Chatt, J.; Guy, R. G.; Sheppard, N. ibid., 1961, 738-742. (17) Cushman, B. M.; Brown, D. B. J. Organomet. Chem. 1978, 752, C42c44. (18) Schrock, R. R.; Parshall, G. W. Chem. Rev. 1976, 76, 243-268. (19) Hendrix, W. T.; von Rosenberg. J. L. J, Am. Chem. SOC. 1976, 98, 4850-4852. (20) Casey, C. P.; Cyr, C. R. J. Am. Chem. SOC.1973, 95, 2240-2247. (21) Taylor, P.; Orchin, M. J. Am. Chem. SOC.1971, 93, 6504-6506. (22) Cramer, R. Acc. Chem. Res. 1968, 7, 186-191. (23) Manuel, T. A. J. Org. Chem. 1962, 27, 3941-3945. (24) Emerson, G. F.: Pettit, R. J. Am. Chem. SOC.1962, 84, 4591-4592. 125) Harrod, J. F.; Chalk, A. J. J. Am. Chem. SOC.1966, 88, 3491-3497. Green, M.: Hughes, R. P. J. Chem. Soc.. Dalton Trans. 1976, 1907-1914. Green. M. Ann N.Y Acad. Sci. 1977. 295. 160-173. Whitesides, T. H.: Neilan, J. P. J. A m . Chem. SOC.1976, 98, 63-73. Casey, C. P.; Cyr, C. R. J. Am. Chem. SOC.1973, 95, 2248-2253. Cowherd, F. G.; von Rosenberg, J. L. J. Am. Chem. SOC. 1969, 97, 2157-2158. Sherman, E. 0..Jr.; Olsen, M. E. Unpublished results. Bonnemann. H. Angew. Chem.. lnt. Ed. Engl. 1970, 9, 736-737. Byrne, J. W.; Blaser, H. U.; Osborn, J. A. J. Am. Chem. SOC.1975, 97, 387 1-8373. Ephritikhine, M.; Green, M. L. H.; MacKenzie, R. E . J. Chem. Soc., Chem. Commun. 1976,619-620. Nixon, J. F.; Wilkins, 6 .J. Organomet. Chem. 1974, 80, 129-137. Chem. Commun. 1978, Sherman. E. 0.;Schreiner, P. R. J. Chem. SOC., 223-224. Tulip, T. H.; Ibers, J. A. J. Am. Chem. SOC.1978, 700, 3252-3254. Parshall, G. W. Catalysis (London) 1977, 7. Collman, J. P.; Kubota, M.; Vastine. F. D.; Sun, J. Y.; Kang, J. W. J. Am. Chem. SOC.1968, 90, 5430-5437. Herde. J. L.; Senoff, C. V. lnorg. Nucl. Chem. Lett. 1971, 7, 10291031. Bennett, M. A,; Milner, D. L. J. Am. Chem. SOC. 1969, 97, 69836994. Schramm. K. D.; Tulip, T. H.; Ibers, J. A. Unpublished results. Van der Ent. A.: Onderdelinden. A. L. lnorg. Chim. Acta 1973, 7, 203208. Corfield, P. W. R.; Doedens. R. J.; Ibers. J. A. lnorg. Chem. 1967, 6, 197-204. Doedens. R. J.; Ibers, J. A. lnorg. Chem. 1967, 6, 204-210. The Northwestern absorption program. AGNoST, includes both the Coppens-Leiserowitz-Rabinovitch logic for Gaussian integration and the Tompa-De Meulenaur analytical method. In addition to various local

421 1 programs for the CDC 6600 computer, modified versions of the following programs were employed: Zalkin's FORDAP Fourier summation program, Johnsons' oRTEP thermal ellipsoid plotting program, and Busing's and Levy's ORFFE error function program. Our full-matrix, least-squares program, NUCLS, in its nongroup form. closely resembles the Busing-Levy ORFLS program. The diffractometer was run under the disk-oriented Vanderbilt system (Lenhert. P. G. J. Appl. Crystallogr. 1975, 8, 568570). (46) Cromer, D. T.; Waber, J. T. "International Tables for X-ray Crystallography", Vol. IV; Kynoch Press: Birmingham, England, 1974; Table 2.2A. (47) La Placa. S. J.; Ibers, J. A. Acta. Crystallogr. 1965, 78, 511-519. (48) Cromer, D. T.; Liberman, D. J. Chem. Phys. 1970, 53, 1891-1899. (49) Stewart. R. F.; Davidson, E. R.; Simpson, W. T. J. Chem. Phys. 1965, 42, 3175-3187. (50) See paragraph at end of paper regarding supplementary material. (51) Chatt, J.; Coffey, R. S.; Shaw. B. L. J. Chem. SOC.1965, 7391-7405. (52) Jenkins, J. M.; Shaw, B. L. J. Chem. SOC.1965, 6789-6796. (53) Vrieze, K. In "Dynamic NMR Spectroscopy", Cotton, F. A ; Jackman, L. M.. Eds.; Academic Press: New York, 1975; pp 441-487. (54) Deeming, A. J.; Shaw, 8. L. J. Chem. SOC.A 1969, 1562-1567. (55) Malatesta, L.; Angoletla, M.; Araneo, A.; Canziani. F. Angew. Chem. 1961, 73, 273. (56) Ela, S. W.; Cram, D. J. J. Am. Chem. SOC.1966, 88, 5791-5802. (57) Tolman, C. A. Chem. Rev. 1977, 77, 313-348. (58) Hietkamp, S.; Stufkens, D. J.: Vrieze, K. J. Organomet. Chem. 1977, 739, 189-198 _.._ (59) Araneo, A.; Martinengo, S. Gazz. Chim. /tal. 1965, 95, 61-66. (60) Chatt, J.; Richards, R L.; Sanders, J. R.; Fergusson. J. E. Nature(London) 1969. 227. 551-552. (61) Blake, D. M.; Kubota, M. J. Am. Chem. SOC.1970, 92, 2578-2579. (62) Vaska, L. J. Am. Chem. SOC.1966, 88. 5325-5327. (63) Vaska, L. J. Am. Chem. SOC.1966, 88, 4100-4101. (64) Blake, D. M.; Smith, L. R. J. Am. Chem. SOC.1977, 99, 3302-3309. (65) Strope, D.; Shriver, D.F. lnorg. Chem. 1974, 73, 2652-2655. (66) Masters, C.; Shaw, B. L.; Stainbank, R. E. Chem. Commun. 1971, 209. (67) Troughton, P. G. H.; Skapski, A. C. Chem. Commun. 1968, 575-576. (68) Masters. C.; McDonald. W. S.; Raper, G.: Shaw, B. L. Chem. Commun. 1971, 210-211. (69) Del Piero, G.; Perego, G.; Zazzetla. A,: Cesari, M. Cryst. Struct. Commun. 1974, 3, 725-729. (70) Clark, G. R.; Skelton, B. W.; Waters. T. N. lnorg. Chim. Acta 1975, 72, 235-245. (71) Ibers, J. A. Adv. Chem. Ser. 1978. 767, 26-35. (72) McPartlin, M.; Mason, R. Unpublished results. Quoted in McPartlin, M.; Mason, R. J. Chem. SOC.A 1970, 2206-2212. (73) Bonnet, J.-J.; Jeannin, Y. J. lnorg. Nucl. Chem. 1973, 35. 4103-4111. (74) Schultz, A. J.; McArdle, J. V.: Khare, G. P.; Eisenberg, R. J. Organornet. Chem. 1974, 72, 415-423. (75) Schultz, A. J.; Khare, G. P.; Meyer. C. D.; Eisenberg, R. lnorg. Chem. 1974, 73, 1019-1024. (76) Kaduk, J. A.; Poulos, A. T.; Ibers. J. A. J. Organomet. Chem. 1977, 727, 245-260. (77) Cobbledick, R. E.; Einstein, F. W. B.; Farrell, N.; Gilchrist. A. B.; Sutton, D. J. Chem. Soc.. Dalton Trans. 1977, 373-379. (78) Ibers, J. A.; Hamilton, D. S.; Baddley. W. H. lnorg. Chem. 1973, 72, 229-232. (79) Brotherton. P. D.; Raston, C. L.; White, A. H.: Wild, S. B. J. Chem. Soc., Dalton Trans. 1976, 1799-1802. (80) Guss, J. M.; Mason. R. J. Chem. SOC., Dalton Trans. 1972, 21932196. (81) Nolte, M.; Singleton, E.: van der Stok. E. J. Organomet. Chem. 1977, 742, 387-395. (82) Herskovitz, T.; Guggenberger. L. J. J. Am. Chem. SOC. 1976, 98, 16 15-1 616. (83) Van Baar. J. F.; Meij, R.; Olie. K . Cryst. Struct. Commun. 1974, 3, 587-593. (84) Bellon, P. L.: Caglio, G.; Manassero, M.; Sansoni. M. J. Chem. SOC.Dalton Trans. 1974, 897-901. (85) Callahan, K. P.; Strouse. C. E.; Layten, S. W.; Hawthorne, M. F. J. Chem. Soc.. Chem. Commun. 1973. 465--466. (86) Einstein, F. W. B.; Sutton. D.'J. Chem. SOC.Dalton Trans. 1973, 434438. (87) Brunie, S.; Mazan. J.; Langlois, N.; Kagan, H. B. J. Organomet. Chem. 1976, 7 74, 225-232. (88) Lupin, M. S.; Shaw, 8. L. J. Chem. SOC. A 1968, 741-749. (89) Brown, L. D.; Barnard, C. F. J.: Daniels, J. A.; Mawby, R. J.; Ibers, J. A. lnorg. Chem. 1970, 77, 2932-2935. (90) Franke, U.; Weiss, E. J. Organomet. Chem. 1977, 739, 305-313. (91) In the text and Tables VI and VI1 the phenyl group of the allyl ligand is abbreviated Ph whereas in Table IV and Figure 2 RALL I S the notation used. (92) Smith, A. E. lnorg. Chem. 1972, 7 7, 2306-231 1. (93) Mingos. D. M. P. J. Chem. Soc., Dalton Trans. 1977, 26-30. (94) (a) Voigt, H. W.: Roth, J. A. J. Catal. 1974, 33, 91-97. (b)Chum. P.-W.; Roth, J. A. ibid. 1975, 39, 198-204. (95) Ouellette. R . J.; Levin, C. J. Am. Chem. SOC.1968, 90, 6889-6891. Chem. (96) Puddephatt, R. J.; Quyser. M. A.; Tipper, C. F. H. J. Chem. SOC., Commun. 1976, 626-627. (97) Full details to be published elsewhere. (98) (a) Lenarda. M.; Ros, R.; Traverso, 0.;Pitts. W. D.; Baddley, W. H.; Graziani, M. Inorg. Chem. 1977, 76,3178-3182. (b)Tulip, T. H.; Ibers, J. A. Unpublished results. (99) Roundhill, D. M.; Lawson, D. N.; Wilkinson, G. J. Chem. SOC.A 1968, 845-849. (100) (a) Gassman. P. G.; Nikora, J. A. J. Organomet. Chem. 1975, 92, 81-87. (b) Cassar, L.; Halpern, J. Chem. Commun. 1970, 1082-1083.