94
Organometallics 1982, 1, 94-96
W(CO)4(alkene),oand Mo(CO),P(C,H& (vacant site trans to pho~phine).~’ In the series of metal-carbene complexes studied here, examples of both rigid and fluxional intermediates have been seen. It is likely that the relative rates of rearrangement and of CO capture of the 5-coordinate intermediate are similar in all the cases studied. In some cases, rearrangement is faster than CO capture, and the intermediate is called “fluxional”. In other cases, CO capture is more rapid and the intermediate is called “rigid”. The rate of capture of 5-coordinate intermediates by ligands has been measured to be between 3 X 106 and 2 X l@M-’ s-1.42r43Under the conditions used here I1 atm 10.01 M (38)Darembourg, D. J.; Darensbourg, M. Y.; Dennenberg, R. J. J. Am. Chem. SOC.1971,93,2807. (39)Dobson, G. R.;Aaali, K. J. J. Am. Chem. Soc. 1979,101,5433. (40)Darembourg, D. J.; Nelson, H. H.; Murphy, M. A. J.Am. Chem. SOC.1977,99,896. (41)Darenebourg, D. J.; Graves, A. H. Inorg. Chem. 1979,18,1257. (42)Kelly, J. M.; H e m , H.; Kcemer von Gustorf, E. J. Chem. Soc., Chem. Commun. 1973,105.
CO), the lifetime of the 5-coordinate intermediate before CO capture can be estimated to be between 3 X lo-, and 5 X lO-”s. The lifetime before rearrangement is probably also within an order of magnitude of these values. Acknowledgment. Support from the National Science h ~ n d a t i o nis gratefully acknowledged. Registry No. 3Cr,27436-93-7; 3Mo,38797-47-6; 3 W ,37823-96-4;
4, 50276-12-5; 7,54040-18-5.
Supplementary Material Available: An appendix describing kinetic models and listing differential equations for the rigid mechanism and the fluxional mechanism (9 pages). Ordering information is given on any current masthead page. (43)Kelly, J. M.; Bent, D. V.; Hermann, H.; Schulte-Frohlinde, D.; Koemer von Gustorf, E. J. Organomet. Chem. 1974,69,259. (44) We wish to thank Professor F. A. Weinhold and Dr. A. J. Shusterman for writing the computer program. (45)Johnson, L. W.;Riess, R. D. “Numerical Analysis”; AddisonWesley: Reading, MA, 1977;pp 299-302.
An X-ray Photoelectron Spectroscopic Study of Groups
0-
and .;rr-Allyl
Antonio J. Ricco, Albert A. Bakke, and William L. Jolly’ Department of Chemistry, University of California, and the Materials and Molecular Research Division, Lawrence Berkeley Laboratoty, Berkeley, California 94720 Received June 29, 198 1
From a comparison of the atomic core binding energies of CH,Mn(CO),, (n-C,H,)Mn(CO),, (u-C3H5)Mn(CO),, and Mn2(CO)lo,we conclude that the methyl, propyl and a-allyl groups in the first three compounds are negatively charged. Both the carbon 1s binding energies and theoretical considerations indicate an unusually large relaxation energy associated with the ionization of the terminal carbon atom of the u-allyl group. The 7r-C3H5group in (?r-C,H,)Mn(CO), has a relatively low electron density, undoubtedly because it is a formal 3-electrondonor group. The carbon 1s data indicate that the C5H5group in (75-C5H5)Mn(CO), is probably positively charged. The allyl group is a commonly encountered ligand in organometallic reactions’ and catalytic processes.2 The ligand is of particular interest because it can engage in either 7’ or 7, bonding to a transition metal; that is, it can be either a bonded, CH2=CH-CH2-M, or ?r bonded.
In this gas-phase X-ray photoelectron spectroscopic (XPS) investigation we have studied the nature of the bonding of the a-allyl and ?r-allyl groups to manganese by comparing the core binding energies of (a-C,H,)Mn(CO), and (.rr-C,H,)Mn(CO), with those of CH,Mn(CO),, (n-C3H7)Mn(CO),, Mn2(CO)lo,and (q5-C5H5)Mn(C0),. In a previous study of a wide variety of LMn(CO), compounds, we observed that the manganese, carbonyl carbon, and oxygen (1)Chiusoli, G. P.; Caasar, L. In “Organic Synthesis Via Metal Carbonyls”,Wender, I.; Pino, P.; E&.; Wiley: 1977;Vol. 2,pp 297-417. (2)Keim, W. In “Transition Metals in Homogeneous Catalysis”; Schrauzer, G.N., Ed.;Marcel Dekker, New York, 1971;Chapter 3, pp 69-91. Muetterties, E. L.;Bleeke, J. R. Acc. Chem. Res. 1979,12,324.
core binding energies are linearly correlated with one another., Although the electronegativities and polarizabilities of the L group studied varied over a considerable range, the electronegativities of the L groups were not correlated in any obvious way with the corresponding polarizabilities. If the binding energy shifts were significantly affected by changes in both atomic charge and relaxation energy, one would expect the relative proportions of these changes to be different for the manganese, carbonyl carbon, and oxygen atoms, and therefore that the binding energy shifta would not be linearly related. Hence we believe that those results indicate that the binding energy shifta in such closely related compounds are mainly due to changes in atomic charges and that changes in electronic relaxation energy are relatively small. This conclusion is supported by the general observation that calculated relaxation energies for an atom are similar when the bonding environments of the atom are ~ i m i l a r . ~ (3)Avanzino, S.C.; Chen, H.-W.; Donahue, C. J.; Jolly, W. L. Inorg. Chem. 1980,19,2201. (4)Davis, D. W.; Shirley, D. A. J.Electron Spectrosc. Relat. Phemm. 1974,3, 137.
0276-7333/82/2301-0094$01.25/0 0 1982 American Chemical Society
Organometallics, Vol. 1, No. 1, 1982 95
Study of a- and *-Allyl Groups
Table I. Core Bindine Energies (eV) carbonyl groups compd
Mn 2 P m
CH,Mn( CO), (n-C, H,)Mn( CO 1, (0-C,H,)Mn(CO), a MnACO),, a (n-C,H,)Mn(CO), (T) 5-CsH5)Mn( CO), CH,CH=CH,
647.30 ( 5 ) 647.13 ( 2 ) 647.13 ( 3 ) 647.01 (3) 646.80 ( 4 ) 646.72 ( 3 )
c 1s 293.62 293.44 293.47 293.28 293.03 292.30
C2H6
(7) (7) (4) (3) (4) (7)
0 Is 539.91 ( 7 ) 539.89 (4) 539.82 (3) 539.57 (7) 539.41 ( 4 ) 538.90 ( 5 )
c 1s (organic ligand) 289.5 (1) 290.32 (11) 289.85 (3) 290.44 (6) 290.85 (3) 290.68 ( 4 ) 290.74
a Reference 3. b The uncertainty in the last digit, estimated as twice the standard deviation determined by the leastsquares fit, is indicated parenthetically. C Perry, W. B.; Jolly, W. L. Inorg. Chem. 1974,13, 1211. I
,OOt
c
I
,,
i
pounds), we conclude that the organic ligands have net negative charges. Indeed, if we interpret the manganese and carbonyl data of Table I with the potential equation = (l/r) AQA + A C (QB/RAEJ B#A
305
I
I
I
300
295
290
I 285
Binding E n e r g y , eV
Figure 1. Carbon 1s spectrum of (?r-C3Hs)Mn(C0)4, showing
shake-up,carbonyl carbon, and allyl carbon bands.
Therefore in this study we shall generally interpret core binding energy shifts in terms of changes in atomic charge. The only previous XPS study of the allyl group of which we are aware involved solid complexes containing the (TC3H5)M~(C0)2 m ~ i e t y . ~Peaks due to the C3H5group were not deconvoluted, and study of the bonding of the allyl group was not the main thrust of the study. Our work complements gas-phase UPS studies of (r-C3H5)Mn(c0),.el7 A typical carbon 1s spectrum, that of (?r-C,H,)Mn(CO), (Figure l),shows, from left to right, shake-up, carbonyl carbon, and allyl carbon bands. Although there are two stereochemically distinct types of carbon atom in the Tallyl ligand, resolution was inadequate to justify deconvolution of the allyl carbon 1s band into two lines. Similarly, neither the a-allyl nor n-propyl carbon 1s bands were deconvolutable. The binding energy data are given in Table I, where we have listed the manganese compounds in order of decreasing Mn 2pSl2 binding energy. Almost without exception, the ordering of the carbonyl C 1s and 0 1s binding energies is the same as that of the corresponding Mn 2 ~ 3 1 2 binding energies. That is, the charge on a carbonyl group is essentially a measure of the charge on the manganese atom to which it is attached. The binding energies of Mn2(CO)lomay be taken as characteristic of an MxI(CO)~group attached to ligand with no charge. Because the manganese and carbonyl groups of the methyl, n-propyl, and a-allyl compounds have higher binding energies than those of Mn2(CO)lo(corresponding to positively charged Mn(CO)s groups in the former com(5)Brisdon, B. J.; Mialki, W. S.;Walton, R. A. J. Orgarwmet. Chem. 1980,187,341. (6) Jackson, S. E. D. Phil. Thesis, Oxford, 1963,cited by: Green, J. C. Structure Bonding (Berlin) 1981,43,37. (7) Worley, S.D.; Gibson, D. H.; Hsu, W.-L. Znorg. Chem. 1981,20, 1327.
(assuming zero relaxation energies), we estimate net charges of -0.14,-0.07, and -0.07 for the methyl, propyl, and a-allyl groups, respectively.g10 The organic ligand C 1s binding energies of the n-propyl and a-allyl compounds are considerably higher than that of the methyl compound partly because of a more negative charge on the methyl group and partly because the measured binding energies of the former compounds are the weighted averages for three carbon atoms and, to some extent, the negative charges of those groups are spread over all the atoms of the groups. I t is remarkable that the manganese and carbonyl binding energies of the n-propyl and u-allyl compounds are essentially identical, whereas the C 1s binding energy for the a-allyl group is 0.47 eV lower than that for the n-propyl group. We believe this difference can be accounted for in terms of an unusually large relaxation energy associated with the ionization of the terminal carbon atom of the a-allyl group. Core ionization of the terminal carbon would yield a species which can be well represented by the following equivalent-core formula,11J2in which a full unit of formal negative charge has been withdrawn from the Mn(CO)5 group. H2N-CH=CH2 MII(CO)~+ No such extraordinary relaxation would be expected for the terminal carbon atom of the n-propyl group. In that (8) In calculating the potentials at the Mn, C, and 0 sites in these compounds, it was assumed that the potential due to the neighboring Mn(C0)6 group in Mn2(CO)lois zero and that the charge of the hydrocompounds is concentrated at a point on carbon ligand in the RMII(CO)~ the C,, axis of the Mn(C0)6 group, 1.95 A away from the Mn atom in CH,Mn(CO), and 2.925 A in CSH6Mn(C0)6and C3H7Mn(C0)5. The geometry of the Mn(C0)6 group was taken from the structure of HMn(CO)6.9Values of (l/r)were taken to be 23.0,32.5,and 15.4 for C, 0,and Mn, respectively. The values for carbon and oxygen were taken from ref 10. For carbon, the value for an sps hybrid orbital was used, and for oxygen, the value for an orbital with 20% 8 character was used. For manganese, the (l/r)value was taken to be that for carbon times the ratio of covalent radii of carbon and manganese, 23.0 (0.77/1.18)= 15.0. The covalent radius for manganese was obtained bv subtractine the covalent radius of double-bonded-carbon (0.67A) from the metal-&bon distance in HMn(C(& (1.85A). (9)La Placa, S.J.; Hamilton, W. C.; Ibers, J. A.; Davison, A. Znorg. Chem. 1969,8,1928. (10)Lu, C. C.; Carlson, T. A.; Malik, F. B.; Tucker, T. C.; Nestor, C. W., Jr. At. Data 1971,3, 1. (11)According to the equivalent-core approximation,'* core ionization of an atom affecta the valence electrons essentially as if the nuclear charge were increased by one unit. The no-bond resonance structure shown surely contributes importantly to the core-ionized state and is readily derived from the resonance structure (12)Jolly, W.L.In "Electron Spectroscopy: Theory, Techniques and Applications"; Brundle, C. R., Baker, A. D.; Eds.; Academic Press, New York, 1977;Vol. 1, Chapter 3,pp 119-149.
96 Organometallics, Vol. 1, No. 1, 1982
case, a reasonable equivalent-core formula of the coreionized species has the positive formal charge on the terminal atom. H3fiCH2-CH2-Mn
(CO),
This explanation is supported by calculations of the C 1s relaxation energies by using the transition-state concept and CNDOIB wave functi~ns.'~The terminal carbon of the a-allyl group is calculated to have a relaxation energy 0.75 eV greater than that of the terminal carbon of the n-propyl group, whereas the carbon atoms attached to the Mn atom are calculated to have relaxation energies differing by only 0.02 eV. The relatively high binding energies of the carbon atoms in the hydrocarbons propene and ethane (see Table I) are consistent with negative charges on the a-bonded organic ligands in the Mn(CO)6complexes, but in these cases the assumption regarding relaxation energy is less valid. In the case of (?r-C,H,)Mn(CO),, there are only four carbonyl groups withdrawing electron density from the manganese atom; thus the manganese atom and carbonyl groups have greater electron densities than they have in the LMn(CO), compounds. In spite of the more negative charge on the manganese atom in (?r-C,H,)Mn(CO),, the ?r-C3H6group has a considerably lower electron density than the a-C3H6group in ( u - C , H ~ ) M ~ ( C OThis ) ~ result is undoubtedly due to the fact that the 7r-C3H5group is formally a 3-electron donor, whereas the a-C3Hsgroup is formally a 1-electron donor. The actual net charge on the ?r-C3Hsgroup is probably close to zero; it is clear from the data that back-bonding to a ?r-C3H5group is much less than that to a CO group. The data for (q5-C5H,)Mn(CO), show that, when one removes another CO group and goes to an aromatic 5-electron donor group, the manganese atom and carbonyl groups become even more negatively charged and the organic ligand carbon atoms have lower electron density. It is not surprising (in view of the strong ?r-acceptor character of the carbonyl groups) that the C 1s binding energy of the C a s group in (q5-C5H5)Mn(C0), is higher than that in (qs-CsHS)2Mn(290.31 eV) or (q5(13) The electronic relaxation energy is taken to be half the difference in electrostatic potentials at the core-ionized site between the neutral molecule and the equivalent-core core-ionized structure, where the valence electron distributions me taken from CND0/2 wave functions. See ref 3 for previous use of this method.
Ricco, Bakke, and Jolly Cd-€5)2Fe(290.03 eV).14 Indeed, the C 1s binding energy of the CSH5group in (qs-C5H5)Mn(CO),is higher than that of benzene (290.42 eV).16 The relaxation energy associated with the C 1s ionization of (qs-CSHs)Mn(CO),is probably greater than that for benzene, and so it seems probable that the C6HSgroup bears a positive charge in this complex.le
Experimental Section Spectra were obtained in the gas phase by using a GCA/ McPherson ESCA 36 spectrometer and Mg Ka X-rays. The Ne Is,N 1s of N2, and Ne 2s lines served as calibration standards." The general procedureshave been described previo~sly.~~ Spectra were fit to Gaussian/Lorentzian line shapes by using the nonlinear least-squares program CURW.'~ The compounds CHSMn(C0)6,( U - C ~ H ~ ) M ~ ( C (a-C3H5)O)~, Mn(CO),, and (q5-C5Ha)Mn(C0), were synthesized according to literature procedures,*29 and the preparative procedure for (nC3H7)Mn(CO)5 was adapted from that of a similar compound.24 The purity of each compound was verified by infrared and/or NMR spectroscopies, as well as melting point determinations where appropriate.
Acknowledgment. This work was supported by the Diredor, Office of Energy Research, Office of Basic Energy Sciences, Chemical Sciences Division of the U.S.Department of Energy, under Contract No. W-7405-Eng-48. Registry No. CH3Mn(CO)s, 13601-24-6; (n-C,H,)Mn(CO),, 15628-53-2; (6-C3H,)Mn(CO)S, 14057-83-1; Mnz(CO)lo, 10170-69-1; (r-C,H&Mn(CO),,33307-28-7; (qS-C5Hs)Mn(CO),, 12079-65-1. (14) Bakke, A. A.; Jolly, W. L.; Schaaf, T. F. J. Electron Spectros. Relat. Phenom. 1977,II, 339. (15) Lindberg, B.; Svenseon, S.; Malmquist, P.-A.;Basilier, E.; Gelius, U.; Siegbahn, K. Uppsala Univ. Inst. Phys., Tech. Rep. 1975, UUIP-910. (16) We have no data with which to reliably estimate the potential at a C a H carbon , atom due to the Mn(CO), group, but, in view of the very low core binding energiea of the manganese atom and the carbonyl group in the complex, we do not believe the potential term can contribute much to the increase in binding energy from benzene. (17) Johaneeon, C.; Hedman, J.; Bemdtsson, A,; Klaeaon, M.; N h o n , R. J. Electon Spectros. Relat. Phenom. 1973,2, 295. (18) Chen, H.-W.; Jolly, W. L.; Kopf, J.; Lee, T. H. J.Am. Chem. SOC. 1979,101,2607. (19) The curve-fitting program CURVY is described by: Bakke, A. A. Ph.D. Thesis. Universitv of California. Berkelev. 1981. (20) King,'R. B. 'Or&nometallic Syktheses", Academic Press: New York, 1966; Vol. 1, p 148. (21) Closson, R. D.; Kozikowski, J.; Coffield, T. H. J.Org. Chem. 1957, 22. 598. (22) Kaesz, H. D.; King, R. B.; Stone, F. G. A. 2.Naturjorsch., B , Anorg. Chem., Org. Chem., Biochem., Biophys., Biol. 1960, 15B,682. (23) Reference 20, p 111. (24) Hieber, W.; Wagner, G. Jwtw Liebigs Ann. Chem. 1958,618,24.
