5082
excess of the closed-shell electronic configuration of each metal atom, the paramagnetic Ni3(h5-CjH5)3(C0)2 molecule one electron over the “magic number” of each nickel atom, and the CoNi,(hg-C,Hg),(CO), and Co3(h5-CsH~)3(CO)(0) molecules a closed-shell electronic configuration for each metal atom. A qualitative LCAO-MO model (based on the nodal character of available metal orbitals and described elsewhere5 for Ni3(h5-C6H5)3(C0)~) predicts that the five excess electrons in Ni3(h5-CjH5)3S2and one unpaired electron in Ni3(h5-C5H5)3(C0)Z occupy primarily the strongly antibonding metal symmetry orbitals essentially localized in the trimetal plane. On the basis that the corresponding bonding and antibonding electron-filled energy levels effectively cancel each other with respect to bonding effects, it follows from this bonding model that the net stabilization of the trimetal system is a consequence of only one bonding electron in the Ni,(h5C5H5)3S2 molecule and five bonding electrons in the Ni3(hg-CjH5)3(CO)zmolecule compared to six bonding (12) H. Sakurai, A. Hosomi, and M. Kumada, Chem. Commun., 4 electrons (corresponding to electron-pair metal-metal (1969). ( l 3 j L. H . Sommer, L . A. Ulland, and A . Ritter,J. Amer. Chem. SOC., bonds) in the CoNi,(h5-C,Hj),(CO), and Co3(hj-C5Hj),90, 4486 (1968). (CO)(O) molecules. The number of antibonding elecHideki Sakurai,* Masashi Murakami trons in this localized symmetry-based bonding scheme Department of Chemistry, Faculty of Science is compatible with the observed average values of the Tohoku Uniaersity, Sendai 980, Japan metal-metal distances: (1) w)ich are unusually long ReceiDed March 2 , 1972 in Ni3(h5-CjH5)3S2(2.801 (5) 6) but much shorter in Ni3(hS-C5H&(CO),(2.389 (2) A); ( 2 ) yhich are shorter in C O N ~ ~ ( ~ ~ - C ~ H (2.358 ~ ) ~ ( C(2) O )A) ~ than in Ni,Organometallic Chalcogen Complexes. XXV. (h5-C5H5)3(C0)Zo; and (3) which in Co3(hWjHj),(CO)Structural and Magnetic Studies of (0)(2.365 (4)A) are analogous with those in CoNi2Co3(h5-C5H5),(C0)(S),Co3(h5-C5H&S2,and the (h”CsHb)3(CO),. Oxidized Monocation, [ C O ~ ( ~ ~ - C ~ H ~ ) ~ S ~ ] + . The recently reported syntheses of C O ~ ( ~ ~ - C ~ H ~ ) ~ S , The Sensitivity of the Geometry of a Triangular and C O ~ ( ~ ~ - C ~ H ~ ) ~ (by C OOtsuka, ) ( S ) Nakamura, and Metal Cluster System to Antibonding Electrons Yoshida6 appeared to provide a critical test case of whether this apparent correlation of metal-metal bond Sir: lengths with the number of “extra” presumably strongly We wish to present here the results of a stereochemantibonding electrons is a valid one in that (with the ical characterization of three compounds which not reasonable assumption of their configurations being only substantiate predictions concerning the dominant analogous to those of Nia(h5-C5H5)3S2 and Coa(h3influence of antibonding valence electrons on the metalC5H5)3(C0)(0)) the tricobalt disulfide molecule conmetal bond lengths of a metal cluster M3(h5-C5H&XY two electrons in excess of the magic number of tains system but also provide evidence that its architecture each metal atom and the latter molecule none. An imcan undergo a drastic Jahn-Teller deformation due to portant bonus to this problem was provided by our subthe effect of antibonding electrons. Previous X-ray analyses of Ni3(h5-C5Hg),S2, Ni3(h5-C5H5)3(CO)2,2s3 sequent discovery that the tricobalt disulfide molecule C O N ~ ~ ( ~ W ~ H ~ ) ~ (and C O ) C, ,O~ ~ ( ~ W ~ H ~ ) ~ ( C O could ) ( O ) ~be easily oxidized to give the [ C O ~ ( ~ ~ - C ~ H ~ ) ~ S ~ ] + monocation. The resulting structural and magnetic showed a basic averaged structure consisting of an investigations form the subject matter of this communiequilateral triangle of metal atoms capped above and cation. below by X and Y groups to give a trigonal-bipyC O ~ ( ~ W ~ H ~ ) ~ ( Cand O)C ( SO) ~ ( ~ ~ - C ~ Hwere , ) : ~ Spre, ramidal M3XY fragment. With the assumption of et u/.~-’O In pared and isolated according to Otsuka, cylindrical symmetry for each cyclopentadienyl ring pentahapto-coordinated to a metal atom, an idealized ( 5 ) C . E . Strouse and L. F. Dahl, ibid., 93, 6032 (1971); C. E. Strouse and L. F. Dahl, Discuss. Farnduy SOC.,No. 47,93 (1969). D3hgeometry was discerned when the triply bridging X (6) S. Otsuka, A. Nakamura, and T. Yoshida, Inorg. Chem., 7, 261 and Y ligands are identical and an idealized Csage(1968). (7) A characterization of C03(hWjHs)& and C O ~ ( ~ ~ - C ~ H S ) Z ( C O ) ( S ) ometry when the X and Y ligands are different. The by mass spectral, ir, 1H nmr, and temperature-dependent magnetic prime characteristic of this particular metal cluster susceptibility studies was reported elsewhere by Otsuka, Nakamura, system is that with the assumption of electron-pair and Yoshida.8 From heat capacity measurements of Co3(hj-CsHs)& over a 15-270°K range, Sorai, er aL,9 later discovered a crystalline bonds between each pair of metal atoms the paramagphase transition (192.5”K) whose behavior they interpreted in terms of a netic Ni3(h6-CjHj)3S2 molecule contains j i v e electrons in cooperative coupling between the librational motion of the cycloto the mono insertion product, higher homologs including di insertion (11, 0.103 mmol) and probably tri insertion products (12) were obtained.’, The configuration of these compounds could not be judged by means of chemical shifts of the methyl protons, but the stepwise mechanism of dimethylsilylene insertion’2 is enough to suggest the same configuration. The trans isomer 2b also gave 10b in a similar stereospecific way. It is concluded, therefore, that the insertion of the photochemically generated dimethylsilylene into the Si-H bond took place with retention of configuration, as is also true for the dichlorocarbene. l 3 Results of other stereochemical studies including conformational analyses, homolytic aromatic silylation, and desilylation reactions will be published in forthcoming papers. Acknowledgment. We thank Toshiba Silicone Co., Ltd., for a gift of chlorosilanes.
~
( I ) V. A. Uchtman, H . Vahrenkamp, and L. F. Dahl, J . Amer. Chem. SOC.,90, 3272 (1968). (2) A . A . Hock and 0. S . Mills, Proc. In?. Conf, Coord. Chem., 6th, 640 (1961). ( 3 ) V. A. Uchtman and L. F. Dahl, to be published. (4) V. A . Uchtman and L. F. Dahl, J . Amer. Chem. SOC.,91, 3763 (1969). Journal of the American Chemical S o c i e t y
94:14
pentadienyl rings in the crystalline lattice and the electronic states of the molecule, They also described their anomalous magnetic susceptibility data (which showed a maximum at 215°K) of Co3(h6-CsHs)3Sz on the basis of the introduction of temperature-dependent energy parameters and the inversion of spin states at higher temperatures to the phenomenological model derived by Chcsnut’o for magnetic exciton systems. Since our physical measurements of Co3(hWsH5)3S? and Cos(hb-CsHs)3(CO)(S)were performed independently, we have outlined
J u l y 12, 1972
5083
contrast to the diamagnetic C O ~ ( ~ ~ - C ~ H ~ ) ~ ( Cthe O)(S), tricobalt disulfide complex is paramagnetic at room temperature with p e f f = 2.83 BM1l (characteristic of two unpaired electrons in the absence of orbital contributions). The cyclopentadienyl 'H nmr resonance observed in carbon disulfide solution from 163 to 324°K indicated diamagnetism below 173°K where a sharp singlet at T 5.33 was observed (cf. Co3(hj-C5H&(CO)(S), T 5.35 at ambient room temperature), but at temperatures greater than 173"K a paramagnetic shift was observed which considerably broadened the signal and moved it monotonically to higher field with increasing temperature. 1 2 *l 3 Bulk susceptibility data Figure 1. The[Co3(h5-C,HJ3S,]+ monocation of the I- salt. T h e measured'' from 98 to 340°K exhibited Curie-Weiss monocation of idealized C ~ 2 m m geometry has crystallogralaw behavior at temperatures greater than 195"K, but phic site symmetry C,-m. at TN = 195°K an abrupt discontinuity was observed such that X M ~dropped to approximately one-half its the effective magnetic moment was found to decrease maximum value within a 30" temperature range. A linearlyfrom 1.95 BM at 297°K to 1.77 BM at 83"K.14 recent calorimetric study of Co3(h5-CjHj)aS2by Sorai Single-crystal X-ray diffraction investigations1j-IY reand coworkersgshowed that a phase transition occurs at vealed the following prominent structural features: 192.5"K. The complementary magnetic data in solid (1) with the assumption of cylindrical symmetry for state and in solution as well as the abnormally high each cyclopentadienyl ring, the C O ~ ( ~ ~ - C ~ H ~ ) ~ ( C O ) ( S ) transition entropyg provide conclusive evidence that molecule has an averaged C.$, geometry $th three this phase transition must be coupled with a change in equivalent Co-Co distances of 2.452 (2) A ; (2) the spin states as a function of temperature. The Chesnut C O ~ ( ~ ~ C ~ molecule H ~ ) ~ possesses S ~ an averaged D M exciton theory'O has provided a good model (modified geometry yith three equivalent Co-Co distances of by us only through inclusion of a TIP term) for our bulk 2.687 (3) A ; (3) the [ C O ~ ( ~ ~ - C ~ H ~monocation >~S,]+ susceptibility data of C O ~ ( ~ ~ - C ~ H ~ ) ~ S , . (Figure 1) is significantly deformed to a C?, geometry [ C O ~ ( ~ ~ - C ~ H ~ ) :was ~ S ~prepared ]+Iby the stoichiometric addition of Iz to the neutral complex and was (14) The magnetic susceptibility of the corresponding SbFa- salt, containing a similarly distorted [Co3(h6-C5Hs)3S?]' monocation (B. I(. purified by repeated recrystallizations from acetonitrile. Teo, P. D. Frisch, and L. F. Dahl, to be publlshcd), also obeyed the This ionic compound was analyzed by ir, solution conCurie-Weiss law and gave a magnetic moment of p e f f = 1.84 BM at ductivity, esr, and bulk susceptibility measurements. 272°K: this value remained unchanged within 0.02 BM over a temuerature range from 272 to 83 OK. Esr studies both in CHXl, solution and in polycrys(IS) Co3(h5-CrHs),(CO)(S): hexagonal with u = 6 = 9.184 (1) talline form showed a broad resonance signal centered US. Pcaled = 1.85-g Cm-3 (2= A, C = 10.639 (2) A; Pobsd = 1.88 g at g = 2.05 with no fine structure. Temperature2). oCo3(h5-C5Hs)3S~:hexagonal with a = 6 = 9.417 ( 1 ) A, c = 10.117 (2) A ; pobrd = 1.88 g cm-3 us. pcnlc,i = 1.86 cm-3 ( Z = 2). The strucdependent susceptibility measurements'' indicated that tures of C ~ ~ ( ~ ~ . C ~ H S ) ~ ( Cwhich O ) ( Sappears ), isomorphous with Cos(h6the cation is a simple paramagnet; for the iodide salt CsHs)r(CO)(0), and C O ~ ( ~ Y C S Hwhich ~ ) ~ Sappears ~, isomorphous with I
them in this communication. It is also noteworthy that our MO scheme used to rationalize the magnetic and X-ray data differs from that proposed by the Japanese workersg to account for their experimental results. (8) S. Otsuka, A. Nakamura, and T. Yoshida, Justus Liebigs A n n . Chem.. 719. 54 (1968). --,(9) M. Sorai, A. Kosaki, H . Suga, S. Seki, T. Yoshida, and S . Otsuka, Bull. Chem. SOC.Jup., 44, 2364 (1971). (10) D . B. Chesnut, J . Chem. Phys., 40,405 (1964). (1 1) The magnetic susceptibility measurements were kindly performed initially by D r . Michael Camp and later by Mr. James Kleppinger at the University of Wisconsin cia the Faraday method. ( 12) Three distinct regions of paramagnetic-shift dependence within the temperature range of 163-324°K can be differentiated: (1) a Curie range (S E 1) for T 2 260°K where the paramagnetic shift varies linearly with 1/T; (2) a region for T 5 173°K where with temperature change the shift is zero which is characteristif of a diamagnetic species (S=O); (3) an intermediate non-Curie region for 173 < T < 260°K where an equilibrium exists between a singlet and triplet species. On the basis of the inherent assumptions that the equilibrium interconversion between the two species occurs at a sufficient rate such that the observed shifts are averaged over both species and that dipolar contributions (which presumbly are small due to the relative isotropy of the g values, ciz., gl I = g l ) to the shifts are essentially invariant to the nondrastic change in geometry, the dipolar term can be combined with the isotropic hyperfine contact term to give the equation (AHilAHo) = K [SI (S 1)/3][exp (AGiRT) 11 -IT-1, where Kis a c o n ~ t a n t . ~ 3However, a least-squares analysis of this simple model did not fit thc observed shifts in the non-Curie region; a much better fit to the observed data was obtained for an analogous empirical model for which a temperature dependence was introduced into the energy separation between the singlet and triplet states (i.e., consistent with the premise that AG is a function of temperature). (13) For an excellent revie% of the application of nmr paramagnetic shifts to an investigation of structures and structural equilibria of metal complexes, see R. H. Holm, Accounfs Chem. Res., 2,307 (1969). I
+
~
\ - -
+
Ni~(h&-CsH5)3S2, were solved by the application of the Wei procedure16 based on an incoherent twinning mechanism involving a single-crystal component of above unit cell dimensions and of symmetryP63/m. This space group requires the C ~ ~ ( ~ ~ - C ~ H S ) ~ (and C O )Co3(h&-CsHs)3Sz (S) molecules to possess crystallographic site symmetry Cah-3im which necessitates for the C O ~ ( ~ ~ - C ~ H ~ ) ~ (molecule C O ) ( S )that the sulfur and carbonyl ligands are disordered into each other. Our crystal model additionally assumes that each cyclopentadienyl ring possesses a twofold orientational disorder of half-weighted atoms such that the crystallographic site symmetry for each moiecule of Co,(hs-CsHs)a(CO)(S) and C03(h5GHj)& is increased to Doh-62m. Anisotropic least-squares refinements o ~ C O ~ ( ~ - C ~ H ~ ) ~ ( C 197 O independent )(S)( observed diffractometry data) and Co3(hS-CeHs)& (218 independent observed diffractometry data) converged to values of 6.0 and 10.3 %, respectively, for R i ( F 2 ) and values of 9.3 and 10.4%, respectively, for Rr(F2). No significant changes in Co-Co distances (i.e., &shows no detectable through the University Research Committee. We also distortion of the c03s?fragment from Dahgeometry19 at express special thanks to Mr. Boon Keng Teo at the room temperature in contradistinction to the large C22r University of Wisconsin, to Dr. C. H . Wei at Oak deformation of this fragment in the [ C O ~ ( ~ ~ - C ~ H ~ )Ridge ~ S ~ ]National + Laboratory (Biology Division), and to monocation illuminates a key point with regard to the Dr. H. Vahrenkamp at the Institut fur Anorganische electronic structures of these complexes. For Co3Chemie der Universitat Munchen for helpful consulta(h5-CjH&S2 the nmr paramagnetic shift data, which tions. strongly indicate that the singlet-triplet energy separaP. Douglas Frisch, Lawrence F. Dahl* tion in the non-Curie region is temperature dependent, Department of Chemistry, Uniaersity of Wisconsin may be rationalized on the basis that at T > 260” the Madison, Wisconsin 53706 triplet state arising from double occupancy of a doublet Receiaed March I , 1972 degenerate level is more likely the ground state than one arising from occupancy of two close-lying energy levels (with one level nondegenerate and the other degenCarbon-13 Spectra of Allenes erate).20 On the other hand, at T < 260°K the observed nmr solution data can be explained in terms of Sir: two close-lying levels of which the lowest one is nonAs a result of technical advances carbon-13 nmr has degenerate. This situation may arise if the antifinally begun to achieve its obvious place as a convenient bonding l a t 2level5 is assumed to be lower in energy and and powerful tool for the elucidation of organic structherefore fully occupied at T < 173’K. With an intures and the probing of ground-state electronic districrease in temperature the la’* level is then presumed to butions of molecules. The unusual b o d i n g situation be destabilized relative to the antibonding 3e’ one5 such in the allene functionality has prompted us to underthat a crossover occurs to give the 3e’ level as the octake a systematic study of this class of compounds.3,4 cupied lower one at room temperature.*I The solidDetermination of the spectra5 of over 50 allenes has state magnetic data, whose interpretation is compliprovided data which show definite chemical-shift trends cated by the observed phase t r a n ~ i t i o n ,are ~ , ~consistent as a function of substitution. Moreover, these trends with these bonding hypotheses. We feel (in harmony are correlated with theoretical calculations of electron with Sorai, et ~ 1 . that ~ ) the change in electronic strucdensity. Representative carbon-1 3 nmr chemical shifts ture may be primarily a consequence of the metalfor allenes with simple alkyl substituents are given in cyclopentadienyl interactions due to a vibronic effect Table I. caused by more restricted ligand motion at lower temIn general, for allenes bonded to nonfunctionalized peratures. The Czageometry of the [Co3(h5-C5H,),carbon, the sp-hybridized carbons are found in the S2]+ monocation may be explained by a Jahn-Teller sparsely populated region at lower field than CS2 (-20 splitting of a doublet ?E ground state (which pret o - 5 ppm) whereas the sp2 carbons appear at slightly sumably would occur for a symmetrical triangular tricobalt cluster system. The unpaired electron would (19) For all P63/m twinned structures,1.4 a significant distortion from D3h (or CzC)symmetry may be undetected due to the twinning as well as the added possibility of a threefold crystal disorder of an unsymmetrical metal triangle. For this reason we stress that the averaged geometry is D3h (or Csv). However, the thermal ellipsoids of the metal atoms in C03(hK!5Hi)& did not give an indication of either any distortion or crystalline disorder. (20) A Jahn-Teller distortion of the molecule would be anticipated for such an electronic configuration (e.g., the antibonding la’n energy level being slightly lower than the antibonding 3e’ one). (21) Another possibility is that the 3e’ level is lower, but the degeneracy is broken due to some unknown and structurally nondetectable change in the Co3(h5-CjHs)& molecule from threefold symmetry (presumably involving the cyclopentadienyl rings and/or the c03s?cluster fragment).
Journal of the American Chemical S o c i e t y
1
94:14
1
(1) (a) E. F. Mooney and P. H. Winson, Annu. Reo. Nucl. Magn. Resonance Speczrosc., 2, 153 (1969); (b) J. B. Stothers, Quart. Rev., Chem. Soc., 19, 144(1965). (2) Examination of electronic environments in aromatic compounds by carbon-13 nmr has been quite successful. See, for instance, R. J. Pugmire, M. J. Robins, D. M. Grant, and R . K. Robins, J . Amer. Chem. Soc., 93, 1887 (1971), and references therein. (3) (a) R . A. Friedel and L. Retcofsky, ibid., 85, 1300 (1963); (b) R . Steur, J. P. C . M. van Dougen, M. J . A. de Bie, and W. Drenth, TetrahedronLert., 3307 (1971). (4) W. W. Conover, J. K. Crandall. and S. A . Sojka, paper presented at the International Symposium on Acetylenes, Allenes, and Cumulenes, Nottingham, England, July 5-8, 1971. (5) Spectra were obtained by using equipment previously described : A. Allerhand, D. Doddrell, V. Glushko, D. W. Cochran, E. Wenkert, P.’T. Lawson. and F. R . N. Gurd, J . Amer. Chem. Soc., 93,544 (1971).
July 12, 1972
