Some Bonding Questions Prompted by Studies of the FluxionalMolecule TriscyclopentadienylnitrosylmolybdenumBY F. A. COTTONDept. of Chemistry, Massachusetts Institute of Technology, Cambridge,Massachusetts 02 139, U.S.A.Received 20th January, 1969The compound (C5H5),MoN0 has been prepared. Its proton magnetic resonance spectrumin solution has been studied as a function of temperature down to - 1 10". The crystal structurehas been determined with relatively high precision at 22". The combined results of these studieslead to some new observations concerning the manner in which cyclopentadienyl rings may bebonded to metal atoms. Specific conclusions are (a) that a trihapto (i.e., x-allylic) form of bondingis not a real possibility, and (6) unsymmetrical relationships of ring to metal, which cannot be readilydescribed in terms of any simple limiting structure, are possible under certain conditions.The molecule (C,H,),(NO)Mo was prepared in these laboratories in early 1968and has been subject to two physical studies.(i) The p.m.r. spectrum in the tempera-ture range - 110 to + 14" has been recorded ; the molecule is shown thereby to bea fluxional one.2* (ii) A single-crystal X-ray structure determination at ca. 22°Chas been carried The results of these studies raise several significant questionsabout bonding in this and related molecules. As a basis for discussion, the back-ground and experimental results will be briefly summarized.Synthesis of (C5H5)3(NO)M~ was prompted by the idea that it might have astructure of the type shown in fig.1, in which there are three different (but eachFIG. 1.-A hypothetical structure for the (C5H,)3MoN0 molecule in which there is a pentahapto(IT) ring, a trihapto (x-allylic) ring and a monohapto (0) ring.well-defined) forms of C,H,-to-metal bonding, viz., monohapto (or a-), trihapto(or " n-ally1 ") and pentahapto (n- or '' sandwich "). This was considered plausibleon the basis that the Mo atom has six electrons, a h5-CSH5 ring contributes fiveand the NO ligand formally contributes three. Thus, in order to achieve the comple-ment of 18 electrons which usually affords maximum stability, the remaining twoC5HS rings together would have to supply 18-6-5-3 = 4 electrons to bondingorbitals. It seemed plausible in the beginning that this would be achieved by having780 TRISCYCLOPENTADIENYLNITROSYLMOLYBDENUMone ring act as a 1-electron donor and the other as a 3-electron donor, with thegeometric relationship of these to the metal atom and its other ligands being asdepicted in fig.1. The only part of this overall structure for which there was nodistinct precedent was the trihapto- (or n-allyl) type ring. However, the idea ofsuch a metal-ring interaction had been mentioned previously in the literature, specifi-cally for the closely related compounds (CSH,),Mo(NO)I and (C,H&NO(NO)CH~,~and it did not seem to be an unreasonable idea at the outset.SUMMARY OF EXPERIMENTAL DATAThe main results of the n.m.r. study are the following. (i) At room temperatureall 15 protons together give a single sharp line, but even before the temperaturereaches 0" this line broadens appreciably.(ii) By about -30" the spectrum of aslowly moving h1-C5H5 ring appears and becomes a well-resolved AA'BB'X patternby about - 50°, while the line due to the remaining 10 protons simultaneously becomes3 4 5 67FIG. 2.-The proton n.m.r.spectrum (60 mHz ; 4 . 2 M in CS2) of (CSH5)3MoN0 at various tempera-tures (reproduced by permission from ref. (1 >).sharp again. (iii) Between - 50" and - 110" two things happen simultaneously.The line of intensity 10 broadens, separates into two lines of equal intensity and theselines then become sharp. At the same time the AA'BB'X pattern is transformedinto an ABCDX pattern. These changes are shown in fig.2.Clearly, observation (i) means that intramolecular rearrangements occur rapidlyenough at 25" to give all three rings time-average equivalence. Each ring passeF . A. COTTON 81from one to another of the available environments rapidly on the n.m.r. time scale.Observation (ii) clearly shows that the slowest processes are entering and leaving theh1-C5H5 environment and the eccentric rotation (" ring-whizzing ") of a ring whenit is in this environment. Beyond this, however, the interpretation is not entirelyunequivocal, although it might appear to be so.It is true that observations cited under (iii) are consistent with the notion that theinstantaneous structure is that shown in fig. 1. Thus, we might assume that at-50" the h3-C5H5 ring is still exchanging roles rapidly with the h5-C5H5 ring (andpresumably also executing its own form of eccentric rotation).We might then assumefurther that this exchange is gradually slowed between - 50" and - 110" so that theh3 and h5 rings become distinguishable, although the five protons within each oneremain indistinguishable due to rotations. Because of the dissymmetry of thepostulated (h5-C,H5)(h3-CSH5)(NO)Mo moiety, the A and A' and also the B andB' protons of the h1-CgH5 ring would not actually be equivalent and as the rate ofinterchange of the (hS-CSH5) and the (h3-C5H5) rings slowed, this non-equivalenceshould be revealed, thus accounting for the changes in the AA'BB'X spectrum.However, this is not the only possible interpretation of the observations and,conversely, the observations do not prove that the structure shown in fig.1 is correct.All of the spectral changes which are seen below - 50" could be accounted for simplyby slowing the rotation about the bond from Mo to C1 of the h1-C5H5 ring; noparticular kind of difference-or any dzflereizce at all-in the way the other tworings are bound to Mo is necessarily implied.In order to establish more definitely the structure of the (C5Hs),MoN0 molecule,an X-ray crystallographic study of the crystalline compound was carried out. Fig. 3, I ! ,'- - - - - A+FIG. 3.-A projection of the structure of the (C5H5)3MoN0 molecule as it occurs in the crystal.and 4 show the results. It is dear that the ring consisting of C(11) to C(15) is agenuine h1-C5H5 ring.Also, the NO group is boundin the expected manner (Mo-N,1.75 l(3) A ; N-0, 1.207(4) A ; angle Mo-N-0, 179.2(2)") [or a three-electrondonor. The relationship of the other two rings to the Mo atom is novel. We believethat it is, in fact, unprecedented. First, each of these two rings is related in essentiallythe same way to the metal atom. There is no possibility of regarding one as a82 TRISCYCLOPENTADIENYLNITROSYLMOLYBDENUMhS-C5H5 and the other as an h3-C5H5 ring. Secondly, each one has a curiousorientation relative to the metal, in which there are two short (2.32-2.35a), onemedium (2.42-2.44 A) and two long (2.58-2.68 A) Mo-C distances.* J 1-00 70.93FIG. 4.-Some bond distances in the (CSH&MoNO molecule.Figures beside bond lines giveC-C and C-H bond lengths in A. Figures beside carbon atoms give Mo-C distances in A.The e.s.d. of C - 4 distances are 0.005-0-008 A ; the e.s.d. of Mo-C distances are 0.003-0-005 A.If this structure (or something essentially similar) is the instantaneous structurein solution, then it is clear that the spectral changes below -50" must indeed beattributed to slowing rotation about the Mo-C bond as mentioned above.DISCUSSION OF CSHS-MO BONDINGThe absence of a trihapto or Ir-allylic C5H5 ring, especially when this apparentlyrequires that the presumably favourable pentahapto arrangement for the other ringalso has to be sacrificed, prompts a closer examination of the concept of a (trihapto-cyclopentadienyl) metal bond, which in turn leads to the conclusion that this is amythical concept.Tt is first to be noted that a large number of (n-allyl) metal bonds have now beenstructurally characterized.' Unless there is some great disparity among the otherligands also bound to the metal, which could be expected to cause a skewing of theallyl plane, the three metal-carbon distances are equal, or nearly equal.Mostimportant, the distance from the metal atom to the centre carbon atom of the allylgroup is usually equal to, and never more than 0.1 A less than the distances to theouter carbon atoms.If we assume that in the postulated (h3-C5H5)M group the ring would remain aregular plane pentagon then by simple geometry it follows that a metal atom whichis equidistant from three carbon atoms is, in fact, equidistant from all five.Inshort, (h3-C5H5)M is geometrically indistinguishable from (h5-C5H5)M. It is notlikely that the ring could deviate very much from planarity, nor are the angles likelyto become greatly distorted in real cases. Hence this conclusion concerning theidealized system should apply in a practical sense.Implicit in the above argument is the fundamental assumption that if an atom Alies a distance d from two other identical atoms B and B' and A is considered to bebonded to B, it must also be considered to be equally bonded to B'. This notioncan be generalized to the idea that, under proper conditions, interatomic distances(especially those which are near but not quite as short as normal bond distances)give a measure of bond energy.While this is not the place to conduct a detaileddiscussion of this point, it does seem applicable to the present case in that thereappears to be appreciable bonding between the Mo atom and all ten carbon atomsof the two comparable rings in (C5H,),MoN0. Thus, according to a previouslyproposed * correlation of bond lengths and bond orders for Mo-C bonds, theshort Mo-C distances indicate approximately single bonds, the medium ones slightlyweaker ones, while the distances 2-58-2.68 A suggest bond orders of ca. 0.6F . A . COTTON 83While none of the foregoing estimates is intended to have literal significance, thepoint is that we have here a very unsymmetrical interaction in which all five ringatoms participate in varying degrees. It does not appear possible to give any simpleelectronic description of this bonding however. Evidently, the flexibility of C5H5rings in respect to bonding with metals is very great and less reliance should beplaced on the simple prototype modes than might previously have seemed safe.F. A. Cotton and P. Legzdins, J . Amer. Chem. SOC., 1968, 90,6232.F. A. Cotton, Chemistry in Britain, 1968, 4, 345.idem, Acct. Chem. Res., 1968, 1,257.J. L. Calderon, F. A. Cotton and P. Legzdins, J. Amer. Chem. SOC. in press.This relatively new notation is explained in F. A. Cotton, J . Amer. Chem. SOC., 1968,90,6230.R. B. King, inorg. Chem., 1968, 7,90. ’ Recently reported structures are found in (a) R. Mason and A. G. Wheeler, J. Chem. SOC.,1968,2543,2549 ; T. G. Hewitt and J. J. deBoer, Chem. Comm., 1968, 1413 ; B. T. Kilbourne,R. H. B. Mais and P. G. Owston, Chem. Comm., 1968, 1438. * F. A. Cotton and R. M. Wing, Inorg. Chem., 1965,4,314; cf., fig. 2