Cp2M2( C0)4--Quadruply Bridging, Doubly Bridging, Semibridging, or Nonbridging?

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Information Report
Category:

Sheet Music

Published:

Views: 2 | Pages: 10

Extension: PDF | Download: 0

Share
Related documents
Description
2576 Journal of the American Chemical Society / 102:8 / April 9, 1980 Cp2M2( C0)4--Quadruply Bridging, Doubly Bridging, Semibridging, or Nonbridging? Eluvathingal D. Jemmis, Allan R. Pinhas, and Roald
Transcript
2576 Journal of the American Chemical Society / 102:8 / April 9, 1980 Cp2M2( C0)4--Quadruply Bridging, Doubly Bridging, Semibridging, or Nonbridging? Eluvathingal D. Jemmis, Allan R. Pinhas, and Roald Hoffmann* Contribution from the Department of Chemistry, Cornell Unicersity. Ithaca, New York Received September 20, 1979 Abstract: Electronic structures of alternative geometries of complexes represented by the molecular formula CP~M~(CO)~ are analyzed using the fragment molecular orbital approach. The tendency for CO semibridging found in d5-d5 complexes is traced back to M2(CO)lo via a hypothetical nonbridged Cp2M2(C0)4 structure. In d7-d7 complexes the transformation from the doubly bridging structure to either a nonbridging or a tetrabridging structure is symmetry allowed. The rotational barrier around the M-M bond in the nonbridging structure is calculated to be low, supporting the mechanism of carbonyl scrambling via nonbridging intermediates. An explanation is also offered for the puckering of the M2(C0)2 rhomboid in cis- Cp2Fez(C0)4. The subject of this paper is the class of binuclear transition metal complexes which contain two metals, two $-cyclopentadienyls, and four carbonyls, Cp2M2(C0)4. Two distinct structural types are represented among these-cis and trans isomeric compounds with two bridging and two terminal carbonyls, 1,' and 2, with four semibridging carbonyls.2 A third 1 cis 1 trans 2 3 structural type, 3, four carbonyls in the bridge, has not yet been observed. The nature of the metal-metal bond, the tendency to bridge or not to bridge, what makes for a semibridging interaction, conformational mobility of carbonyls-these are obvious questions that come to mind about these lovely molecules. These queries also reflect some of the more intriguing currents in contemporary organometallic and inorganic chemistry. We will try to make a minor contribution here in unraveling some aspects of the electronic structure of these molecules. Electron Counting Preliminaries The thread of a metal-metal bond, of multiplicity to be defined, runs throughout literature discussions (and the structural formula representations) of binuclear complexes in general, and these molecules in particular. So perhaps it is worth repeating explicitly the imperatives of the 18-electron rule, spin state, and bond length which are behind the generally accepted bond multiplicity assignments in these complexes. Cp2Fe2(C0)4, iron formally in oxidation + 1 if the Cp ring is taken as anionic, 0 if Cp is a neutral five-electron donor, acquires a 17-electron count around the iron. With a single iron-iron bond in the complex both metals achieve an 18- electron configuration, consistent with the diamagnetic character of the compound. The iron-iron distance in the /80/ $01.OO/O multitude of complexes whose crystal structure is known ranges between 2.49 and 2.51 A.' Some case could be made for an Fe-Fe single bond on this basis but it is not an awfully strong one. Replacement of the bridging or terminal carbonyls by linear nitrosyls, cyanide, or isocyanide groups, of terminal carbonyls by phosphines, of a Cp ring by a trio of carbonyls-all of these are relatively trivial electronic perturbations, and the doubly bridged structure 1 is maintained, for iron complexes. Unbridged dimers are not found though they are strongly implicated as intermediates underlying carbonyl fl~xionality.~ For CpzRu2(C0)4 an equilibrium between the bridged form and an unbridged isomer of unspecified structure is observed in soi~tion,~~~~~-~ though the solid-state structure is bridged.4d The corresponding Os complex is claimed to exist in an unspecified unbridged structure in the solid and solution.4e On the other hand we have the smaller class of Cp2M2(C0)4 complexes with a d5 configuration, M = Cr, Mo, W, in oxidation state I. The individual metal has a 15-electron count, a triple bond is indicated by the 18-electron rule, and indeed these metal-metal bonds are short: Cr-Cr, ;2c Mo- Mo, 2.452b A. A pervasive structural feature of these compounds is the involvement of all the carbonyls in semibridging, Le., in short, bonding approaches of carbonyls of one metal to those of the other. The concept of metal-metal bonding, especially its extension to extremely strong and short multiple bonds, exemplified by the work of the Cotton group, forms a beautiful chapter of modern chemistry. The idea is heuristically useful and elegant. Yet in the case of bridged, supported metal centers the nagging doubt always remains as to the nature of the forces holding the two metals a certain distance apart and making for a low-spin ground-state configuration. This is especially so in the case of bridging carbonyls, where the bonding is highly delocalized, similar to that in diborane. We5 and others6 have argued that there is not much to be gained by imagining that there is a metal-metal bond in such carbonyl bridged complexes, and we would like to discuss the question again in the context of complexes 1 and 2. It may be indeed pleasing or even correct to assign a metal-metal bond order to a given compound of type 1 or 2. But we view this as a description of nature and not as its understanding. Why do the d7 Cp~M2(C0)4 complexes all assume structural type 1 and the d5 ones 2? In principle geometries 1 and 2 (with or without semibridging), and even the unlikely looking 3, are available to singly or triply bonded, d7 or d5, molecules. What makes a certain electron count opt for one structural type? That is the basic question, and if we can answer it we can say that we comprehend these molecules American Chemical Society Jemmis, Pinhas, Hoffmann / Cp2M2(C0) Setting Up We will build up the orbitals of the structural alternatives 1-3 in order of.increasing number of bridging carbonyls. This will be accomplished by a retrotheoretical analysis, in which CpM(C0)2, CpM(CO), and CpM building blocks will be combined with each other and an increasing number of bridging ligands. The extended Hiickel procedure, an approximate molecular orbital method, is used, with details provided in the Appendix. The required orbitals are those of 4,5,6, CpM(CO),, n = 0, 1,2. One privileged coordinate system would place the z axis 3 s P I I x along the metal to cyclopentadienyl normal, which would emphasize the natural descent of the fragments from a CpM(C0)3 parent. Another coordinate system, however, will be chosen here. This prepares the fragment for its eventual incorporation in a metal-metal bonded dimer. To explore the characteristics of the metal-metal bond u, ir, and 6 designations of pseudosymmetry around the M-M axis are useful. To prepare for this we will orient the z axis along the eventual direction of approach of the other metal atom. The orbitals of all of these fragments have been discussed in the literat~re.~ However we doubt that they are yet a household word, so let us review them. The CpM(CO), orbitals are related to those of M(CO),+3 by the isolobal replacement of a cyclopentadienyl by three carbonyls. The ML, orbitals (7-9) are simple.8 Above a nest of three levels, the -12 Figure 1. Frontier orbitals of CpzMoz(C0)4 and Mo2(CO)lo2+. At left are the orbitals corresponding to the experimental geometry of Cpz- M02(C0)4 and in the middle those corresponding to an idealized octahedral geometry. The same MM distance (2.448 %.) is used in all the three structures. 0 0 I io ti which would bring M and CO together. In practice this is difficult to do, but let us see what happens. The observed structures have an M -M-CO angle of Let us imagine that 0 = 90 were a good model for no semibridging, and bring together two CpM(C0)2 units in such a geometry, 12, for a metal-metal separation of A, that remnants of the octahedral t2g, there are disposed the delocalized equivalents of 6 - n hybrids pointing toward the missing octahedral sites. On going to CpM(CO),, n = 1 or 2, the symmetry falls precipitously, to a single mirror plane. The electronic pseudosymmetry is also unbalanced, for a Cp unit lacks the acceptor characteristics of the three carbonyls it replaces. Nevertheless the gross features of the ML, parent are discernible in the CpM(CO), frontier orbitals. Each has a group of higher lying hybrids and a set of three t2,-like orbitals below. The hybrids and t2g sets are less well separated in the Mo case. We are now ready for the construction of the various dimers. The Cp2Crz(C0)4 Type-Four Semibridging Carbonyls One would like to think that it is possible to unscramble the semibridging carbonyl interaction, Le., one metal bonding to a carbonyl terminally connected to another metal, 10, by first creating a situation where the M carbonyl and M are too far removed to interact, and then initiating a geometrical motion U 12 found in the Mo dimer. Here is the problem that arises. If 0 is 90 and the Cp indeed takes the place of three carbonyls in an octahedral fragment (so that the M -M-normal to Cp angle cp is ), then for most values of the torsion around the MM axis it is impossible to avoid unacceptable steric contacts. These occur either between the cyclopentadienyls or between a Cp and a carbonyl on the other metal. The most severe interactions are between Cp ring hydrogens near a = 0 (a is the torsion angle around the MM bond, a = 0 corresponding to eclipsed Cp s). Even if the Cp rings are rotated to create a cogwheel, H-H contacts as short as 1.8 remain at a = 0. The result is a destabilization of that torsional region relative to large a-the anti orientation, a = 180, actually depicted in 12 and related to the observed dimer structures, is some 32 kcal/mol more stable than the a = 0 rotamer. As we shall see, this number will be smaller for the iron case. We now focus on the orbital pattern of the idealized octahedral dimer 12 at a = 180. This is shown in the middle of Figure 1. This figure contains also at left the orbitals of the real 2578 Journal of the American Chemical Society 1 102:8 1 April 9, I980 Table 1. Composition of Frontier Orbitals of Cp2Moz(CO)4 in an Idealized Octahedral Fragment Geometry, 0 = 90' orbital type energy, ev % on metals 70 composition of metal part b8 K* yz 2xy ly 3% it* LUMO 52 43xz 6z2 2(x2 - y2) IS 2% U HOMO 64 36z (x2 - y2) + 6s 2bu 6* I(x2 -y2) 4z2 + 1~ 1 bu 7r xz + 4z Is 1% (x2 - y2) 19~' + 2x2 + 2x a, it yz + loxy + 3y - Table 11. Structural Parameters for Idealized Octahedral Fragment and Observed Cp~M2(C0)4 molecule MM,A 0,deg P,deg 4,deg ref octahedral fragment CPZMQ(CO)~ b CpZCrZ(C0) c (C~MeWrdC0) a geometry of Cp2Mo2(C0)4 and at right the orbitals of a hypothetical decacarbonyl Mo2(CO) lo2+. We will return to these cogs in an explanation in a moment, but note immediately the five below two level pattern of the idealized and real geometries. A nice closed-shell structure is obtained for a d5-d5 electron count. The filled levels in Cp2Mo2(CO)4 at any a are two of ag symmetry, two b,,, and one a,. Table I summarizes their composition in the idealized geometry. The actual symmetry, C2h, is so much lower than cylindrical that much mixing naturally occurs. The characterization of these orbitals as cr, x, or 6 is to some extent arbitrary (for instance, the lowest a, level is both x bonding and 6 antibonding), but it is nevertheless a useful distinction based on predominant orbital type. It allows one to see explicitly the triple bond assigned to these molecules on the basis of electron-counting considerations. The occupied orbitals are of u, x, x, 6, and 6* type. These five orbitals contribute of the overlap population between the metal atoms, and that partitions into 53% cr, 42% x, and 5% 6. The five below two pattern is the most interesting aspect of the electronic structure of these molecules, for as we will see it differs significantly from the CpzFez(C0)4 geometries. At the same time it assures a low-spin ground state for the Cr, Mo, and W dimers. Where does that pattern come from? To answer that question we retreat in an isolobal replacement scheme, from 12 to an isoelectronic Mo2(C0)102+ (13). The orbitals This is a consequence of replacing three good acceptors (carbonyls) by their electronic equivalent but a much poorer acceptor, a cyclopentadienyl. The symmetry is also greatly lowered by the substitution, which is responsible for the admixture of cr, x, and 6 character exhibited in Table I. To go from idealized octahedral fragment dimer 12 to the real semibridging geometry is a process that involves minimally a readjustment of three angles shown in 14: the M'-M-CO angle 6' defined earlier, the angle fi between the carbonyls, and the angle 4 between the MM axis and the normal to the Cp 14 plane. These angles in the idealized structure and the two known crystal structures are summarized in Table 11. The excursion in fi is small, in 6' also small but much more important because it gauges the semibridging, and the change in 4 is very great. We have explored several cuts through a potential-energy surface that connects the dimer made up of idealized octahedral fragments (6' = 90, fi = 90, 4 = '), with a real- istic structure for the Mo case. The angular degrees of freedom are linked-for instance, it is costly to decrease 6' unless the Cp rings on the other metal are moved out of the way through increasing 4. J%t the overall surface is in our calculations a soft one. Only 2 kcal/mol separate the idealized geometry from the real solid-state one-the two defined by the angles in the first two entries of Table I. Thus it is no surprise that the angle 4 in the three structures known varies over a wide range. The energy does not change much as all four carbonyls enter the semibridging region. And while there are changes in the individual d-block levels (center to left of Figure l), the most dramatic effects occur in a theoretical descriptor of bonding, the Mulliken overlap populations-structures 15 and 16 show./?? U le 13 of this molecule, calculated at the same Mo-Mo separation of A as for 12, are shown at right in Figure 2. The levels may be trivially constructed from those of two M(CO)s fragments, 7, much as we and others have done for Mn2(C0),~,.~,~ There is a metal-metal u bonding orbital and a total of six orbitals based on the t2g set xy, xz, yz of each metal center. The symmetry of these is x, 6,6*, x*. Their splitting, indicated in Figure 2, is quite substantial. This is a consequence of the good d orbital overlap at the relatively short Mo-Mo distance. The five below two pattern is thus established already in the M~L~o dimer and, as we shall shortly see, so is the incipient carbonyl bridging. The two orbitals that split away are of x* symmetry. Now we return to the idealized CpzMo2(CO)4 structure. Figure 2 shows the aforementioned splitting pattern and an expected movement to higher energy of all the orbitals. O b \ ' idealized real the MM and the M-CO overlap populations in the geometries of the first two entries of Table 11. The MM overlap population in the real structure is much smaller than in the idealized one, for the same metal-metal separation. The M to semibridging or distal CO carbon overlap population, on the other hand, is much larger. All other overlap populations change less drastically, though in a well-defined pattern: in the real structure the M-terminally or directly bound C bond is weaker, as is the CO bond itself. The overlap Jemmis, Pinhas, Hoffmann / Cp2M2(C0) population between the metal and the oxygen of the distal CO goes from to The change is not great, but is especially significant because the metal-carbon bond is growing in strength in the course of the same deformation. Along with these overlap population changes there is a shift of electron density, summarized in the charges of 17 and 18. ' bu idealized CP cp+cp ---=!-I idealized real On going to the experimental geometry the metal atoms lose electron density to the carbonyls. The electron redistribution attendant upon semibridging is entirely consistent with the replacement of metal-metal bonding by a metal-remote carbonyl interaction. More specifically that distal interaction involves occupied metal based orbitals acting as donors to carbonyl a* acceptor orbitals.1 The orbital details will be filled in momentarily, but the overall effect is indicated schematically in 19. Note that the buildup 19 of M-distal CO bonding, diminution of M-direct CO and C-0 bonding, the electron drift from metal to CO, all of these symptoms are consistent with the charge transfer being in the direction indicated. The M-0 antibonding coupled with M-C bonding is a particularly striking sign. Why is metal-metal bonding diminished as the semibridging interaction is turned on? The answer to this question is tied up with the details of the interaction schematically and incompletely indicated in 19. The careful reader will have noted that in Figure 2 the a, and 1 b, levels have changed designation on going from the idealized to the real structure, a, from a to 6*, 1 b, from a to cr*. This is based upon their calculated composition. For instance, a, in the idealized geometry was given in Table I as being made up of 71% on metal, 61% a(yz andy) and 10% 6*(xy). In the real geometry the composition changes to 52% on metal, 5% a and 46% 6*. A similar transformation occurs in 1 b,, which goes from 68% on the metal, mostly a, in the idealized structure to 46% on the metal, predominantly U*. These two orbitals are graphed in Figure 2. The greater part of what happens is a simple reorientation, shown in 20 and 21. Mc;- 2' a, CP cp*cp The molecular orbitals remember their CpM(C0)z fragment parentage and follow the fragment around as it reorients. The transformation of a character into 6* and cr* appears quite natural from this viewpoint. OU Figure 2. Contour diagrams of the occupied lb,, and a, orbitals of Cp2Mo2(C0)4 showing only Ma contributions. The idealized octahedral geometry is on the left and the experimental geometry on the right. The 1 b, orbital is in the xz plane and the a, orbital is in a plane parallel to xz and 0.5 8, in they direction. While this picture explains the loss of metal-metal bonding, it does not yet have the compensatory factor-metal carbonyl semibridging. To bring this into focus we must restore the carbonyl components omitted in 20 and 21 to these a, and 1 b, orbitals. This is done in 22 and and 23 are drawn for the idealized octahedral fragment structure. A diminution of 8, Le., a motion toward semibridging, the very same motion which P we claim demolishes metal-metal a bonding, that same distortion increases the metal to carbon (of semibridging carbonyl) interaction. This analysis is confirmed by an examination of the contributions to the metal to distal CO overlap populations orbital by orbital. That the secondary interaction responsible for semibridging is a bonding one is a consequence of the way the phases of the M' and the carbonyl on M in M'-M-CO are tied together. The COT* is tied to the d orbital on its own metal, M, through the primary back-bonding interaction. M is linked to M' by metal-metal a bonding. This forces a bonding phase relationship between M and CO, which is increased upon the carbonyls bending over. The phase relationship invoked and illustrated in 22 and 23 is present in the idealized dimer formed from octahedral fragments. It is enhanced in the real geometry, but the overlap population is a definite sign of its existence in the octahedral dimer. But it can be traced back even further. Its roots are in the binuclear decacarbonyl M02(CO) lo2+, where orbitals analogous to 22 and 23, namely, 24a,b, are to be found. The M-C(O) overlap population in Mo2(CO)lo2+ is very 24 Q 2Sb 2580 Journal of the American Chemical Society / / April 9, :J 7T* combination of the CpM(C0) hybrid levels, c* in symmetry, is too high to enter into further bonding interactions. It does, however, become the LUMO of the final composite molecule. Now two carbonyls come in, at the right side of Figu
Recommended
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x