Organometallics 1986, 5, 561-566
is comparable to that of relatively “weak” actinide-tocarbon u bonds. Much of the chemistry of actinide and by implication, group 4,diene complexes can be understood on the basis of this characteristic as well as the relatively weak C-H bonds which would be formed on transferring hydrogen atoms (or ions) to the termini of the butadiene fragment.
561
Division of Chemial Sciences, U S . Department of Energy, under Contract W-31-109-ENG-38a t Argonne National Laboratory. We thank Dr. Lester Morss for helpful comments. Registry No. 1, 67506-91-6; 2, 99829-51-3;3, 99668-63-0;4, 99829-52-4;5, 99837-57-7;6,99837-58-8;Cp’zThClZ, 67506-88-1; Cp’,UCl,, 67506-89-2;Cp’,Th(Me)Cl, 79301-20-5;(THF),Mg(CH,CH=CHCH2), 83995-88-4;(THF),Mg(CH,CMe=CMeCH,), 99829-53-5;LiCH=CHz, 917-57-7;3-butenylmagnesiumbromide, 7103-09-5; tetravinyltin, 1112-56-7.
~~
AcknowledWent- This research W a s SuPPoad by the National Science Foundation under Grant CHE8306255 to T.J.M. and by the Office of Basic Energy Sciences,
Supplementary Material Available: A table of anisotropic thermal parameters for non-hydrogen atoms (Table 11) and a listing of observed and calculated structure factors from the final cycle of least-squares refinement (Table 111) (12 pages). Ordering information is given on any current masthead page.
Kinetics of Intramolecular Interconversion of Two Isomers of (P-W M, (P- CNMe, 1(CO) ,L (M = Ru, L = PR,, AsPh,, or SbPh,; M = Os, L = AsPh,) Mark R. Shaffer and Jerome B. Keister* Department of Chemistv, University at Buffalo, State University of New York, Buffalo, New York 14214 Received July 3 1, 1985
The clusters (fi-H)M,(p-CNMe,)(CO),L (M = Ru, L = PPh3, PMePhz,PMe,Ph, PBu3, PCyc,, P(OMe)3, P(O-i-Pr),, P(OPh)3, AsPh,, SbPh3; M = Os, L = AsPh,) exist in solution as equilibrium mixtures of two isomers, L being coordinated to the nonbridged metal atom in one (n-e isomer) and to the bridged metal atom in the other (b-e isomer). The n-e isomer is the kinetic product formed by addition of L to ( p H)M3(p-CNMe2)(C0),L’(M = Ru, L’ = py; M = Os, L’ = NCMe). The kinetics of the intramolecular rearrangements of the n-e isomers to the equilibrium mixtures have been determined. The equilibrium constant depends upon the identity of L, increasing in the order SbPh, (0.37) < PMe2Ph (0.59) < PBu, (0.67) < PMePh2 (1.75) < AsPh3 (2.5) < P(OMe)3(3.2) < P(OPh), (4.5) < P(O-i-Pr)3(6.9) < PCyc, (7.2) < PPh3 (7.51, and values of the forward rate constant 104kf(s-’) vary slightly (at 5 “C): SbPh3 (0.27) < AsPh3 (0.87) < PMezPh (1.1)< PBu, (1.4) < P(O-i-Pr), (1.9) < PMePhz (2.7) = PPh, (2.7) ,< P(OPh), (2.8) < P(OMe)3(3.6) < PCyc,. These trends indicate that both steric and electronic factors are important, K and kf increasing as the size and hardness of L increase. Activation parameters for kf for rearrangement O~?P-H)M~(~-CNM~~)(CO)~(A~P~~) are AH* = 20.7 (M = Ru) and 15.5 kcal/mol (M = Os) and AS* = -3.2 (M = Ru) and -27 eu (M = Os). The deuterium isotope effect Hkf/Dkf for the Ru complex is 1.24. The proposed mechanism involves migration of hydride, methylidyne, and carbonyl ligands through pairwise bridge opening to form intermediates having only terminally bound ligands. One of the most active areas of research concerning metal cluster chemistry has been ligand migrations, most commonly involving carbonyls, hydrides, or hydrocarbon fragments, the rates of which are frequently rapid on the NMR time scale.’ It has been recognized for some time that ligand migration may play an important role in ligand substitution or other reactions of metal clusters,2but because ligand migration processes are frequently orders of magnitude faster than substitutions of strongly bound ligands such as CO, there has been little definitive evidence regarding this. We have shown that (fi-H)Ru3(fi-CNMe2)(C0),L(L = py, PPh,, and other group 15 donor ligands) exist in solution as three isomeric forms in which the ligand L is (1) (a) Band, E.;Muetterties, E. L. Chem. Rev. 1978, 78,639. (b) Evans, J. Ado. Organomet. Chem. 1977, 16, 319. (2) (a) Sonnenberger, D.; Atwood, J. D. J. Am. Chem. SOC.1980,102, 3484. (b) Sonnenberger, D. C.; Atwood, J. D. Ibid. 1982,104, 2113. (c) Sonnenberger, D. C.; Atwood, J. D. OrganometaZlics 1982, I, 694. (d) Atwood, J. D.; Brown, T. L. J. Am. Chem. SOC. 1976,98, 3160.
0276-7333/86/2305-0561$01.50/0
coordinated (1) on a bridged metal atom in the axial position trans to the CNMez ligand (Figure 1,bridged-axial (b-a) isomer), (2) on a bridged metal atom in the equatorial position trans to the Ru(CO), unit (Figure 1, bridgedequatorial (b-e) isomer), or (3) on the nonbridged metal atom in the equatorial position (Figure 1, nonbridgedequatorial (n-e) i ~ o m e r ) . ~The ? ~ only cluster for which the b-a isomer has been found to be the most stable is ( p H)Ru,(p-CNMe,)(CO),(py). In general, when L is a phosphine, phosphite, arsine, or stibine, the b-e and n-e isomers are both present in solution, and thermally induced substitution on ( ~ - H ) R U ~ ( ~ - C N M ~ ~ by) (LC O ) ~ ~ produces the equilibrium mixture. However, the kinetic product from replacement of py from (p-H)Ru3(pCNMe,)(CO),(py) by L a t 25 “C is exclusively the n-e (3) Dalton, D. M.; Barnett, D. J.; Duggan, T.P.;Keister, J. B.; Malik P.T.;Modi, S. P.; Shaffer, M. R.; Smesko, S.A. Organometallics 1985, 4., 1854. (4) Churchill, M. R.;Fettinger, J. C.; Keister, J. B. Organometallics 1985, 4 , 1867. ~~~
0 1986 American Chemical Society
562 Organometallics, Vol. 5, No. 3, 1986
Ru Ru Ru Ru Ru Ru Ru Ru Ru Ru
os
os
Shaffer and Keister
Table I. 'H NMR Spectral D a t a of (rr-H)M?(rr-CNMer)(CO)oL" resonances, ppm L isomer hydride N-methyl PPh, n-e -14.15 (d, J = 3.0 Hz) 3.72 (s, 6 H) b-e -14.25 (d, J = 8.2 Hz) 3.58 (d, 3 H, J p H = 1.3 Hz) 2.80 (9. 3 H) PMePhs n-e -14.40 (d, J = 2.9 Hz) 3.72 (s', 6 H) b-e -14.70 (d, J = 9.5 Hz) 3.56 (d, 3 H, J p H = 1.5 Hz) 2.99 (s, 3 H) PMezPh n-e -14.40 (d, J = 2.9 Hz) 3.75 (s, 6 H) b-e -14.70 (d, J = 9.5 Hz) 3.63 (d, 3 H, J p H = 1.5 Hz) 3.13 (s, 3 H) -14.26 (d, J = 2.9 Hz) PBu~ n-e -14.81 (d, J = 9.5 Hz) b-e PCyc, n-e -14.83 (s) 3.81 (s, 6 H) b-e -14.66 (d, J = 7.7 Hz) 3.76 (s, 3 H) 3.73 (s, 3 H) P(OMeh n-e -14.60 (d, J = 5.1 Hz) b-e -15.09 (d, J = 10.3 Hz) P(0-i-Pr), n-e -14.49 (d, J = 4.4 Hz) b-e -14.98 (d, J = 10.3 Hz) -14.61 (d, J = 6.6 Hz) P(OPh), n-e b-e -15.17 (d, J = 11.0 Hz) AsPh3 n-e -14.28 (s) 3.80 (9, 6 H) b-e 3.60 (s, 3 H) -14.28 (s) 2.90 (s, 3 H) -14.47 (s) SbPh, n-e 3.72 (s, 3 H) 3.67 (s, 3 H) 3.44 (s, 3 H) -14.62 (s) b-e 3.01 (s, 3 H) 3.78 (9, 6 H) AsPh, n-e -16.04 (s) -16.36 (s) b-e 3.59 (9, 3 H) 2.89 (9, 3 H) 3.79 (s, 6 H) NCMeb -16.08 (s)
"In deuteriochloroform at 21 "C unless otherwise noted.
resonances: 6 2.58 (s, 3 H, NCMe).
Experimental Section
F i g u r e 1. Structures of bridged-axial, bridged-equatorial, and nonbridged-equatorial isomers (pH)M3(p-CNMe2)(CO),L. isomer. R e a r r a n g e m e n t of the n-e i s o m e r to the n - e / b - e equilibrium m i x t u r e occurs at a rate conveniently followed b y NMR spectroscopy. S i n c e relatively few s t u d i e s of nondegenerate r e a r r a n g e m e n t s of ligands on clusters have been c o n d u c t e d and few i n w h i c h the effects of c h a n g i n g the steric and electronic properties of the ligands have been e x a m i n e d , w e h a v e undertaken a s t u d y of the k i n e t i c s of interconversion of these isomeric clusters. These r e s u l t s h a v e i m p l i c a t i o n s f o r other reactions, such as alkyne-alk y l i d y n e coupling, w h i c h p r o c e e d t h r o u g h H R u , ( p - C X ) (CO),L intermediate^.^ (5) (a) Beanan, L. R.; Abdul Rahman, 2.; Keister, J. B. Organometallics 1983,2, 1062. (b) Beanan, L. R.; Keister, J. B. Organometallics 1985. 4, 1713.
Chemicals. ( ~ - H ) R U ~ ( ~ - C N M ~ ~ (p-H)Ru3(p)(CO)~~,~ ( ~ - H ) O S ~ ( ~ - C N M ~ ~PMe2Ph,7 ) ( C O ) ~ and ~,~ CNMe2){CO)9(py),3 PMePh2' were prepared according to previously reported procedures. Solvents and all other ligands were obtained from commercial sources and were used as received. C h a r a c t e r i z a t i o n of Compounds. Characterizations of (pH)Ru~(~-CNM~~)(C (LO=) PPh3, ~ L AsPh3, SbPh3, PBu,, PCyc,, and py)3'4and ( ~ - H ) O S ~ ( ~ - C O M ~ ) ( Cwere O ) ~reported (PP~~)~ previously. All new compounds reported here were characterized by comparison of the infrared (Perkin-Elmer 457 spectrophotometer) and 'H NMR (JEOL FX-9OQ instrument) spectra with those of the other derivatives. Infrared spectral data are available as supplementary material (Table A), and 'H NMR data are given in Table I. T h e phosphite complexes decomposed during chromatography and characterization is of the crude products. (p-D)Ru3(pL-CNMe2)(CO),,.A solution of R u ~ ( C O (84 ) ~ ~mg, 0.13 mmol), triethylamine (2 mL, dried over potassium hydroxide and freshly distilled), and D,O (2 mL) in dry T H F was placed in a 50-mL Schlenk flask equipped with a reflux condenser, nitrogen gas inlet, and stir bar. The solution was heated a t 60-70 "C for 1 h. After the solution was cooled, a solution of tetraethylammonium bromide (250 mg) in D,O was added and then the volatile components were removed by vacuum transfer until only the water and the purple precipitate [NEt4][ D R U ~ ( C O ) ~ J remained. The water was removed by pipet, and the precipitate was washed three times with water (10 mL) and dried with vacuum. Next a solution of methyl iodide (80 pL) in dichloromethane (5 mL) was added, and to this purple solution was added dropwise with stirring a solution of methyl isocyanide (10 pL) in dichloromethane (10 mL) until the solution turned yellow. Next the solution was evaporated to dryness on a rotary evaporator, and the residue was purified by thin-layer chromatography on (6) Keister, J. B. Ph.D. Dissertation,University of Illinois, Urbana, IL, 1978. (7) Mathur, M. A,; Myers, W. H.; Sisler, H. H.; Ryschkewitsch,G. E. Inorg. S y n t h . 1974, 15, 128. (8) Bavaro, L. M.; Keister, J. B. J . Organornet. Chem. 1985,287, 357.
Organometallics, Vol. 5, No. 3, 1986 563
Two Isomers of (p-H)M3(p-CNMez)(CO)& silica eluting with cyclohexane. The second yellow band was extracted with dichloromethane;evaportion yielded the product as a yellow solid (31 mg, 37%). The 'H NMR spectrum indicated that the product contained 98% deuterium in the hydride position. (~-H)OS,(~-CNM~~)(CO)~(NCM~). To a stirred solution of (p-H)Os3(pCNMez)(C0)10 (31mg, 0.034 mmol) in acetonitrile (10 mL) under a nitrogen atmosphere was added dropwise with stirring from a pressure-equalizing dropping funnel a solution of trimethylamine N-oxide dihydrate (4 mg, 0.036 mmol) in acetonitrile (10 mL). After addition was complete, the solvent was removed by vacuum transfer. The 'H NMR spectrum of the residue indicated that the only cluster products were (p-H)Os3(pCNMe,) (CO)sLand unreacted starting material. The compound is unstable in solution. Recrystallization from acetonitrile gave yellow crystals. Anal. Calcd for Cl4Hl0N2OS0s3; C, 18.26; H, 1.09; N, 3.04. Found: C, 18.46; H, 1.33; N, 2.84. (~-H)OS~(~-CNM~~)(CO)~(A~P~,). This product was prepared by addition of a slight excess of AsPh, to a solution of (~-H)OS,(~-CNM~~)(CO)~(NCM~) in deuteriochloroform. The compound was purified by thin-layer chromatography on silica eluting with 10% dichloromethanein cyclohexane. The product was characterized by comparison of its 'H NMR and IR spectra previously rewith those of (~-H)OS~(~-COM~)(CO)~(ASP~~) ported.* The compound was recrystallized from methanol/2propanol for analysis. Anal. Calcd for C,H2,NO&Os3: C, 30.38 H, 1.87. Found: C, 30.35; H, 1.92. Kinetic Measurements. The progress of each isomerization was monitored by 'H NMR spectroscopy using a JEOL FX-9OQ spectrometer. For each run a sample of (p-H)Ru3(p-CNMez)N CdeuterioM~) (CO),(py) or ( ~ - H ) O S , ( ~ - C N M ~ ~ ) ( C O ) ~ (in chloroform (ca. 0.6 mL) was prepared. Then the ligand L was added and the solution transferred immediately to a 5-mm NMR tube which was placed into the probe already thermostated at the desired temperature (It1 "C, temperature measured using the methanol chemical shift methods). Spectra were recorded automatically at regular intervals using the STACK routine. Equilibrium constants Kq = k f / k ,were determined by integration of the N-methyl or hydride resonances due to each isomer at 2 1 O . Error limits are given as the standard deviation from the mean for three determinations. The rate constant koM for the isomerization to equilibrium was determined from a plot of In (mt- meq}vs. time t. The quantity m represents the mole fraction due to the n-e isomer, which was measured by the function m = h,/{h, + 2hb),where h, and hb are the peak heights of the methyl resonance due to the n-e isomer and the upfield methyl resonance due to the b-e isomer, respectively, or which was measured by the function m = h,/(h, + hb),where h, and hb are the peak heights of the hydride resonances of the n-e and b-e isomers. Although peak areas should be more properly used, satisfactory measurements could be made by using peak heights much more quickly. Equilibrium constants calculated from both peak height and area measurements were in good agreement. Good linear plots were obtained over 2-3 half-lives. A computer-calculatedleast-squares procedure was used to determine the slope of the plot for each run. Error limits of one standard deviation were also computer calculated. Two to three runs were made at each temperature and the average value used as the best number for the rate constant k0bsd. Forward and reverse rate constants were calculated from k, = k o b s ~ / { liKeg)and kf = kobsd - k,. There was no measureable variation in Keqwith temperature over the range examined. Activation parameters for rearrangements of (p-H)M,(p CNMe2)(CO)9(AsPh3) (M = Ru and Os) were obtained from plots of In { k f /.r? vs. 1/T using a computer-calculatedleast-squares procedure. Error limits are given as the computer-calculated standard deviation.
Results The structure of (p-H)Ru,(p-CNMez)(CO),(py) in solution and in the solid state contains py coordinated in the bridged-axial position (Figure 1). The structure of ( p H)0s,(p-CNMez)(CO),(NCMe) is unknown. The lH NMR
spectrum of the latter (Table I) is consistent with NCMe coordination on the nonbridged metal atom and in an axial position, as was shown for ( ~ . - H ) O S , ( ~ - C O M ~ ) ( C O ) ~ (CNCMe3),I0 since this is the only isomer containing equivalent N-methyl groups. However, the N-methyl resonances may be coincidently isochronous; the chemical shifts of the two N-methyl resonances of (p-H)Ru3(pCNMe,)(CO),(py) are not very different from one another. As we have previously reported, the kinetic product formed upon mixing (p-H)Ru3(p-CNMez) (CO),(py) and L = PR3, AsPh,, or SbPh, is the n-e isomer (p-H)Ru3(pCNMez)(C0)9L.3 The n-e isomer is also the kinetic product from replacement of NCMe on (p-H)Os,(pCNMe,)(CO),(NCMe) by AsPh,. Rearrangement to the equilibrium mixture of n-e and b-e isomers occurs with a rate conveniently measured by using 'H NMR spectroscopy, Unless noted otherwise, all experiments concern deuteriochloroform solutions. The n-e and b-e isomers are easily differentiated in the NMR spectrum (Table I). When L is a phosphorus donor ligand, the ,lP-hydride coupling constant is 0 4 Hz for the n-e isomer and always a larger value, 7-11 Hz, for the b-e form. For complexes of ligands having phenyl substituents, the N-methyl resonances have very similar chemical shifts for the n-e form but are separated by ca. 0.6 ppm for the b-e isomer. In previous papers we have fully characterized these isomers by lH and 13C NMR spectroscopy, mass spectrometry, elemental analysis, and X-ray crystallography for representative example^.^^^ The value of the equilibrium constant for the n-e to b-e rearrangement (Table 11) depends upon the steric and electronic properties of L. Comparison of the values of Kq for L = PCyc, (7.2) and PBu, (0.67) suggests that large ligands stabilize the b-e isomer relative to the n-e isomer. We have previously proposed that the b-e form is the most favorable on steric ground^.^ However, the values of K,, for L = EPh, (E = P (7.51, As (2.5), and Sb (0.4)), ligands which have very similar cone angles,27suggest that electronic effects are also important. The kinetics of rearrangement from the n-e isomer of (p-H)Ru3(p-CNMez)(C0),Lto the equilibrium mixture were measured from the relative peak heights of the resonances due to each isomer, either the methyl or hydride resonances. For relaxation to equilibrium first order kinetics are observed and the first-order rate constant, kobsd, is equal to the sum of the forward and reverse rate constants, kf k,. The equilibrium constant, Keq,is given by kf/k, and was determined from the relative amounts of the n-e and b-e isomers a t equilibrium. Therefore, we were able to determine values for kf and k , (Table 11). The isomerization does not proceed by dissociation of L since there is no formation of (p-H)Ru,(p-CNMe,)(CO),(PPh,) during the isomerization of (p-H)Ru,(pCNMe2)(CO),(AsPh,) in the presence of added PPh,, even though exchange is noted over longer time periods. Furthermore, the forward rate constant kf of 3.6 X s-l a t 17 "C for isomerization of (p-H)Ru3(p-CNMe2)(C0),(AsPh,) is 240 times larger than the rate constant for replacement of AsPh, by CO (1.6 X s-I, extrapolated from 40 O C ) . , We can also rule out CO dissociation as the mechanism since the rate constant for the second ligand substitution on (~-H)Ru~(~-CNM~~)(CO)~(PP~~) is much smaller than its rate constant for isomerization. Therefore, the isomerization is an intramolecular process. There is only a small dependence of the values for kobd, kf, and k, upon the nature of L (Table 11). Both steric and
+
~
( 9 ) Van Geet, A. L. Anal. Chem. 1968,40, 2227; 1970,42, 679.
(10)Gavens, P. D.; Mays, M. J. J. Organomet. Chem. 1978,162,389.
564 Organometallics, Vol. 5, No. 3, 1986
S h a f f e r a n d Keister
Table 11. Equilibrium and Rate Constants for Isomerization of (w-H)M3(w-CNMe2)(CO),L" M Ke,* 104kObs,cS-1 104kf,d s-1 104k,,d s-1 Ru 7.5 (0.2) 3.1 (0.1) 2.7 (0.2) 0.36 (0.02) 2.9 (0.1) 1.4 (0.2) 1.5 (0.1) Ru 0.67 (0.08) PBu,3 1.19 (0.04) 0.6 (0.2) 0.7 (0.1) PMePh, Ru 1.75 (0.04) 4.2 (0.1) 2.7 (0.2) 1.5 (0.2) PMe,Ph Ru 0.59 (0.04) 2.8 (0.3) 1.1 (0.2) 1.7 (0.1) P(OMe), Ru 3.2 (0.2) 4.8 (0.2) 3.6 (0.5) 1.2 (0.1) P(0-i-Pr,) Ru 6.9 (0.7) 2.2 (0.1) 1.9 (0.2) 0.28 (0.04) 2.8 (0.1) 0.62 (0.02) P(OPh), Ru 4.5 (0.2) 3.4 (0.1) PCyc, Ru 7.2 (0.2) 5.6 (0.2) 4.9 (1.0) 0.7 (0.1) AsPh, Ru 2.5 (0.1) 0.51 (0.02) 0.36 (0.04) 0.15 (0.01) 1.22 (0.02) 0.87 (0.09) 0.35 (0.02) 3.1 (0.1) 2.2 (0.2) 0.90 (0.05) 5.0 (0.1) 3.6 (0.3) 1.43 (0.07) 6.8 (0.3) 4.9 (0.4) 1.9 (0.1) SbPh, Ru 0.37 (0.01) 0.76 (0.02) 0.27 (0.03) 0.73 (0.02) AsPh, os 4.4 (0.2) 0.51 (0.04) 0.42 (0.07) 0.09 (0.01) 0.90 (0.04) 0.74 (0.08) 0.16 (0.01) 1.22 (0.02) 1.0 (0.1) 0.21 (0.01) 0.37 (0.03) 2.0 (0.1) 1.63 (0.02) 3.14 (0.05) 2.60 (0.03) 0.54 (0.03)
L PPh,
T , "C +5 +5 -5 +5 +5 +5 +5 +5 -5 0 +5 +12 +17 +19 +5 +27 +34 +39 +44 +49
In deuteriochloroform solution unless otherwise indicated. Measured at 21 "C. Error limits are the standard deviation for three measurements. Error limits are the computer-calculated standard deviation. Error limits are calculated from those for Kegand kobsd.
electronic properties of L seem important. The trend in the values of kobsdparallels the trend in the values of Keg. It appears that hf is somewhat more sensitive to the identity of L than k,. Comparison of the values of kf for L = PCyc, (4.9 X s-l) and PBu3 (0.6 X s-l) a t -5 "C suggests that the rate is increased by increasing the size of L. As for the values of Keg,comparison of the values of kffor L = EPh, (E = P (2.7 X s-l), As (0.9 X s-l), and Sb (0.3 X lo-* s-l)) indicate that electronic effects are also important. A small but normal deuterium kinetic isotope effect is observed. The values of Hkobsd/Dkobsdand Hkf/Dkffor ( p H)RU,(~-CNM~~)(CO)~(A~P~J a t 13 O C are both 1.24 f 0.10. Activation parameters were determined for rearrangements of (p-H)M,(p-CNMe2)(CO)g(AsPh3) (M = Ru and Os). For the Ru cluster values of AH* = 20.7 f 1.0 kcal and A S = -3.2 f 0.2 eu were found, while values for the Os analogue were 15.5 f 1.3 kcal and -27 f 5 eu, respectively. Only a small influence of the solvent upon the rate constants or equilibrium constants was noted. The equilibrium constants for (p-H)Ru3(p-CNMez)(CO),(AsPh,) in benzene, chloroform, and acetone were 2.2, 2.5, and 2.8, respectively. Rate constants kobsdfor n-e to b-e rearrangement a t 19 "C in chloroform and in acetone were 6.8 X low4and 4.9 X s-', respectively, not significantly different.
F i g u r e 2. Proposed mechanism of carbonyl fluxionality on (wH)M3(~-CO)(CO)Io1-(M = Fe, Ru, Os). \ '
t
i
F i g u r e 3. Proposed mechanism for isomerization of (w-H)M,(pCX)(CO)9L ( X = CNMeJ.
Discussion The isomerization of (p-H)M3(p-CNMe2)(CO),Lfrom the n-e form to the equilibrium mixture is an intramolecular process. Evidence for this includes the lack of exchange with PPh3 during the isomerization of (p-H)RU,(~-CNM~~)(CO)~(A the S Pmuch ~ ~ ) , greater rate for isomerization of this cluster than for dissociation of AsPh,, the negative values of the entropies of activation (inconsistent with a dissociative process), and the very small dependence of the rate constant upon the identity of L. This last point and the fact that there are no examples of intramolecular migrations of phosphine ligands suggest that the ligand L remains attached to the same metal atom during the rearrangement. Then the isomerization requires that both the hydride ligand and the methylidyne ligand move. Although there is little information concerning the migratory aptitudes of methylidyne ligands, carbonyl and
hydride migrations for the isoelectronic and isostructural clusters (p-H)M3(p-CO)(CO)lo1(M = Fe, Ru, or Os) have been e~tab1ished.ll-l~ The proposed mechanism for carbonyl migration on (p-H)M3(p-CO)(CO)lo1-is shown in Figure L1',14 Simul(11) Wilkinson, J. R.; Todd, L. J. J. Organomet. Chem. 1976,118, 199.
(12) (a) Ungermann, C.; Landis, V.; Moya, S. A.; Cohen, H.; Walker, H.; Pearson, R. G.; Ford, P. C. J . Am. Chem. SOC.1979, 101, 5922. (b) Johnson, B. F. G.; Lewis, J.; Raithby, P. R.; Suss, G. J . Chem. SOC., Dalton Trans. 1979, 1356. (13) Eady, C. R.; Johnson, B. F. G.; Lewis, J.; Malatesta, M. C. J . Chem. SOC.,Dalton Trans. 1978, 1358. (14) Although a different mechanism for carbonyl fluxionality for (cr-H)Ru,(~-CO)(CO)~~,lwas proposed by Johnson, Lewis, Raithby, and Suss,7bthe correct assignment of the I3C NMR spectrum by Ford and ~o-workers'~ leads us to the conclusion that the lowest energy process is the same as that observed for the Fe analogue."
Two Isomers of (p-H)M,(p-CNMe,)(CO)SL taneous opening of the hydride and carbonyl bridges leads to a structure having only terminal ligands. Then closing of the hydride bridge to a different metal atom, in conjunction with formation of a carbonyl bridge, exchanges the carbonyl ligands. Similar mechanisms have been proposed for carbonyl exchange on Fe,(p-C0),(CO),Ll5 and Ru,(p-CH,) h-co)(CO)IO.’~ We propose the mechanism in Figure 3 for isomerization of (p-H)M,(p-CNMe,)(CO)& Opening of the methylidyne and hydride bridges of the n-e isomer produces an intermediate containing only terminally coordinated ligands. Closing of carbonyl and hydride bridges across a second metal-metal vector while retaining the terminally bound methylidyne generates a new dibridged cluster isostructural with the ground-state structures but having exchanged positions of carbonyl and methylidyne ligands. Opening of the hydride and carbonyl bridges positions the hydride on the L-substituted metal atom. Finally, closing hydride and methylidyne bridges across the third metalmetal vector generates the b-e isomer. Although there is no direct evidence for this mechanism, the small normal deuterium kinetic isotope effect is consistent with hydride bridge opening and the order of decreasing ease of carbonyl fluxionality for (p-H)Ru,(p-CX)(CO),,, (X = 0- > OMe > NMe2),3,’3,‘7which would be expected to be the order of decreasing migratory aptitude of the methylidyne ligand (vide infra), is consistent with this mechanism. Deuterium kinetic isotope effects of 1.5 and 1.8were found for hydride fluxionality on H20s,(CO)lo(PPh,) and H,RU,(CO)~(HC=CCMe#+, respectively.ls While there has been no direct comparison of the migratory abilities of CO, COMe+, and CNMe2+ ligands, analogy to related systems suggests the order CO > COMe+ > CNMe2+. Cis-trans isomerization and carbonyl fluxionality of C P ~ M ~ ( C O ) ~ ( ~(M -C= O Fe, ) ~ Ru)19have been shown to proceed via bridge-terminal carbonyl exchange through an intermediate having only terminal carbonyls. Comparison of Cp,Fe,(CO),(p-CO), and Cp,Fe,(CO),(pCO)(p-CNPh)20finds that the latter isomerizes a t a slower rate, presumably because of the larger activation energy required to convert the isocyanide ligand to the terminal coordination mode. Finally, 0-coordination of Lewis acids such as AlEt, to the bridging carbonyl has been shown to increase to barrier to isomerization of Cp,Fe,(CO),(pCO)2;21this is presumably because 0-coordination stabilizes the bridged coordination mode relative to the terminal mode.22 Since COMe+ and CNMe2+may be considered to be adducts of the Lewis acid Me+ with CO and CNMe, respectively, this suggests that the barrier to bridge-terminal ligand interconversion should increase in the order CO < COMe+ and COMe+ < CNMe2+. Wilkinson and Todd have attributed the decrease in rate of carbonyl (15) (a) Benfield, R. E.; Gavens, P. D.; Johnson, B. F. G.; Mays, M. J.; Aime, S.; Milone, L.; Osella, D. J. Chem. SOC., Dalton Trans. 1981, 1535. (b) Cotton, F. A.; Troup, J. M. J . Am. Chem. SOC.1974, 96, 4155. (16) Holmgren, J. S.; Shapley, J. R. Organometallics 1985, 4, 793. (17) Johnson, B. F. G.; Lewis, J.; Orpen, A. G.; Raithby, P. R.; Suss, G. J . Organomet. Chem. 1983, 173, 187. (18) Rosenberg, E.; Anslyn, E. V.; Barner-Thorsen, C.; Aime, S.; Osella, D.; Gobetto, R.; Milone, L. Organometallics 1984, 3, 1790. (19) (a) Bullitt, J. G.; Cotton, F. A.; Marks, T. J. J . Am. Chem. SOC. 1970,92, 2155. (b) Gansow, 0. A.; Burke, A. R.; Vernon, W. D. J . Am. 1972,94,2550. (c) Adams, R. D.; Cotton, F. A. J. Am. Chem. Chem. SOC. SOC.1973,95,6589. (d) Cotton, F. A.; Hunter, D. L.; Lahuerta, P.; White, A. J. Inorg. Chem. 1976,15,557. (e)Cotton, F. A.; Kruczynski, L.; White, A. J. Inorg. Chem. 1974,13, 1402. (20) Howell, J. A. S.;Matheson, T. W.; Mays, M. J. J . Chem. SOC., Chem. Commun. 1975, 865. (21) Alich, A.; Nelson, N. J.; Strope, D.; Shriver, D. F. Inorg. Chem. 1972, 11, 2976. (22) Horwitz, C. P.; Shriver, D. F. Adu. Organomet. Chem. 1984,23, 219.
Organometallics, Vol. 5, No. 3, 1986 565 fluxionality for (p-H)Fe,(p-CO)(CO)lo’- in the presence of BF, t o 0-coordination.” Other mechanisms, such as ones involving p3-CNMe,, p3-H, and/or p&O ligands, may be postulated which are consistent with our data, but these are intuitively less attractive on the basis of electron counting and are without precedent in analogous HM,(P-X)(CO)~~ systems. These mechanisms also require several intermediates to convert the n-e isomer to the b-e isomer. Few studies of ligand effects upon fluxional processes on clusters have appeared,’ but a number of studies of fluxional processes a t single metal centers have found that the energy barrier to ligand rearrangement increases as the size of the ligand increases. For example, the free energy of activation for intramolecular CO rearrangement on Co(CO),(EX,) increases in the order EX, = SiF, < SiC1, < SiPh,,,, and for intramolecular cis-trans isomerization of M(C0)4(13CO)(PR3)(M = Cr, Mo, W) the barrier increases in the order PR, = PMe, < PEt, < P(i-Pr),.24 Activation entropies for these isomerizations are generally negative, as found here for n-e to b-e isomerization, and it is interesting that the slower rate for isomerization of W(CO),(l3C0)(PR,) compared to the Cr analogue is due to a more negative entropy of activation (-54.9 vs. +1.8 eu), as is the case for the Os and Ru clusters. However, these studies of isomerizations a t a single metal center are not directly related to ligand exchange between different metal centers reported here because of the different mechanisms, trigonal twist for the former and bridgeterminal ligand interconversion for the latter. Furthermore, the electronic effects upon isomerizations of mononuclear complexes and clusters are different, the free energies of activation for fluxionality for Co(CO),(EPh,) decreasing as the atomic numbers of the donor atoms increase ( E = Si (9.5 kcal) > Ge (8.8) > Sn (7.2) > P b (6.6)),14 in contradistinction to the trend seen for the isomerization of (p-H)M,(p-CNMe,)(CO),L. The variation of K , with the steric properties of L is explained by the fact tkat the b-e coordination site is less sterically hindered than the n-e site., Thus, the b-e isomer becomes more stable relative to the n-e form as the size of L increases. The dependence of K,, upon the electronic properties of L is more difficult to rationalize. The trend in K,, for ligands of similar cone L = PPh, > AsPh, > SbPh3 and L = P(OPhI3 > PBu,, can be analyzed in terms of a number of ligand properties-r-acceptor ability, u-donor ability, basicity, polarizability, and others. We have chosen to rationalize the variation in Keqin terms of the “softness” of L.2s The problem can be analyzed in terms of the formation of t h e most favorable acid-base combinations-(Ru(n-e)-L, Ru(b-e)-CO) or (Ru(n-e)-CO, Ru(b-e)-L). According to the HSAB theory25the most favorable combination will pair the softer metal site with the softer ligand and the harder metal site with the harder ligand. Of the two metal sites we propose that the nonbridged metal atom is in the lower oxidation state and is thus the softer; in support of this, a study of HOs,(pX)(CO),, (X = C1, Br, and I) found that in each case the (23) Lichtenberger, D. L.; Brown, T. L. J. Am. Chem. SOC.1977, 99, 8187. (24) Darensbourg, D. J.; Gray, R. L. Inorg. Chem. 1984, 23, 2993. (25) Pearson, R. G. In “Survey of Progress in Chemistry”; Scott, A., Ed.; Academic Press: New York, 1969; Vol. 5, pp 1-52. (26) Chesky, P. T.; Hall, M. B. Inorg. Chem. 1983,22, 3327. (27) Tolman, C. A. Chem. Reu. 1977, 77, 313. (28) Few studies have compared the softness of various phosphine, arsine, and stibine ligands. Klopmanzghas proposed a softness parameter which is related to the electronegativity of the donor atom. (29) Klopman, G. J . Am. Chem. SOC.1968, 90, 223.
Organometallics 1986, 5 , 566-574
566
gross atomic charges on the bridged Os atoms are positive while the charge on the nonbridged metal atom is negative.26 By this argument the softer the ligand L, the more favorable should be the n-e isomer. This is also consistent with the observation that py, a harder ligand, coordinates to a bridged metal at0m~9~ but the softer isocyanide ligands coordinate to the nonbridged atom (e.g., (y-H)Os,(yCOMe)(CO)9(CNCMe3)10).a-Bonding arguments do not explain these observations.
Acknowledgment. This work has been supported by the National Science Foundation (Grant No. CHE8121059). Registry No. (p-H)Ru3(p-CNMe2) (CO),(PPh,) (n-e),98065(b-e),98065-41-9;(pH)42-0; (~H-)RU~(~-CNM~,)(CO)~(PP~,) R U ~ ( ~ - C N M ~ ~ ) ( C O ) (n-e), ~ ( P M99921-89-8; ~ P ~ ~ ) (p-H)Ru,(pCNMe2)(CO)9(PMePh2)(b-e), 99901-46-9; (p-H)Ru,(pCNMe2)(CO),(PMe,Ph) (n-e), 99921-90-1; (p-H)Ru3(pCNMe2)(CO),(PMe2Ph) (b-e), 99901-47-0; (p-H)Ru,(p-
CNMe2)(CO),(PBu3) (n-e), 98065-48-6; (p-H)Ru,(p-CNMe,)(CO),(PBu,) (b-e),98065-47-5;(~-H)RU~(~-CNM~~)(CO)~(PC~,) (n-e),98065-46-4;(p-H)Ru,(p-CNMez)(CO),(PCy,) (b-e),9806545-3; (p-H)Ru3(p-CNMe,)(C0),(P(OMe),) (n-e),99901-43-6;(pH)RU,(~-CNM~~)(C~)~(P(OM~)~) (b-e),99901-48-1;(p-H)Ru3(p-CNMe2)(CO)g(P(O-i-Pr)3)(n-e), 99901-44-7; (p-H)Ru3(pCNMe,)(CO),(P(O-i-Pr),) (b-e), 99901-49-2; (p-H)Ru,(pCNMeJ(CO),(P(OPh),) (n-e), 99901-45-8; (p-H)Ru3(pCNMe2)(CO),(P(OPh),) (b-e), 99901-50-5; (p-H)Ru3(pCNMe2)(CO)g(AsPh3) (n-e), 98065-40-8; (p-H)Ru,(p-CNMe,)(CO),(AsPh,) (b-e),98065-39-5;(~-H)RU,(~-CNM~,)(CO)~(S~P~,) (n-e), 98065-44-2; (p-H)Ru3(p-CNMe2)(CO),(SbPh,) (b-e), ~ ( A S99921-91-2; P~,) 98065-43-1;( ~ - H ) O S ~ ( ~ - C N M ~ ~ ) ( C O )(n-e), (p-H)0s3(p-CNMe2)(CO),(AsPh3)(b-e),99901-53-8;(p-H)Os,(pCNMe,) (CO),(NCMe), 99901-51-6;(p-D)Ru,(p-CNMe,)(CO) 99901-52-7;RU,(CO)~~, 15243-33-1;(p-H)Ru3(p-CNMe2) (CO),(py), 97415-87-7;D2, 7782-39-0;OS, 7440-04-2;Ru, 7440-18-8. Supplementary Material Available: Infrared spectral data for (p-H)M,(p-CNMe,)(CO),L (1 page). Ordering information is given on any current masthead page.
Synthesis and Structural Characterization of Binuclear Gold( I ) Complexes Derived from Two New A-Frame Ligands H. Schmidbaur,*+ Th. Pollok,t R. Herr,+ F. E. Wagner,$ R. Bau,*sN J. Riede,+ and G. Mullert Anorganisch-Chemisches Institut und Physik-Department der Technischen Universitat Munchen, 0-8046 Garching, West Germany, and Department of Chemistry, University of Southern California, Los Angeles, California Received July 18, 1985
The molecular structure of the potential A-frame ligand cyclopropylidenebis(diphenylphosphine),cdpp, has been determined by an X-ray diffraction analysis. The ground-state conformation in the crystal is a rotamer with the lone pairs of electrons at phosphorus in cis and trans (2 or E ) positions relative to the cyclopropane plane. This ligand forms 1:2 complexes with AuCl, AuI, and AuCH3. The 1:1 complex with AuCl(7b) has the expected eight-membered ring structure, as determined by an X-ray diffraction study of the trimethanol solvate, but one of the chloro ligands is found to be dissociated from the gold(1) centers. In the remaining dinuclear cation with an Au-Au contact of 2.907 (1) A, the residual chloro ligand is positioned above one gold atom with an Aul-Au2-C11 angle of only 75.4 (1)'. The structural inequivalence of the metal atoms in the solid is confirmed by lWAuMossbauer data. Solution NMR data in CDCl, indicate a virtually symmetrical structure due to exchange processes. A related autodissociation of only one chloro ligand from a dinuclear gold(1) complex was also found in the X-ray analysis of the 1:l complex of AuCl with 2-methoxyethylidenebis(diphenylphosphine)(3b). In this complex, obtained by methanol addition to the vinylidene precursor as the tetramethanol solvate, the residual chlorine is attached to one gold atom on the projection of the Au-Au axis (Aul-Au2 = 3.002 (1)A; Aul-C13 = 2.963 (3) A; Au2-Aul-Cl3 = 176.2 (1)O). The methoxy groups of the two side chains are oriented toward Au2. These peculiar interactions lead to very uncommon coordination spheres for both gold(1) centers in the cation of 3b. Au Mossbauer data of the solid and NMR spectra of the solution are again in agreement with an unsymmetrical and symmetrical structure, respectively. Crystal data for cdpp: C2/c, a = 27.064 (6) A, b = 10.146 (2) A, c = 17.562 (4)A, 13 = 114.30 ( 2 ) O , V = 4395.1 A3,2 = 8, R, = 0.041 for 262 refined parameters and 2437 reflections with F, 1 447,). Crystal data for 7b: P2*/c, a = 17.283 (4) A, b = 17.027 (4) 4, c = 19.284 (4) A, /3 = 108.75 (2)O, V = 5373.7 A3, 2 = 4, R, = 0.044 for 343 parameters and 6045 observed data. Crystal data for 3b: P212121,a = 11.388 (2) A, b = 22.588 ( 5 ) A, c = 22.737 ( 5 ) A, V = 5840.9 A3, R, = 0.042 for 378 parameters and 4549 observed reflections.
Introduction
Bis(diPhenYIPhosPhino)methane(dPPm) is the Prototype of a class of difunctional phosphines known to favor +
Anorganisch-Chemisches Institut.
* Physik-Department der Technischen Universitat Munchen.
§University of Southern California. .$A. v. Humboldt Awardee, 1985, Technical University of Munich.
0276-7333/86/2305-0566$01.50/0
the construction of binuclear complexes with short metal-metal contacts.' This close metal-metal proximity is responsible for a variety of novel addition reactions which situation designated as an "A-frame>* in a structure. Many of these reactions are at least partially reversible as other substrates are added to the system, and (1) Puddephatt, R. J. Chem. SOC.Reu. 1983, 12, 99.
0 1986 American Chemical Society