Cyclobutane-1,2,3,4-tetrone has been both predicted and found to have a triplet ground state, in which a b(2g) σ MO and an a(2u) π MO are each singly occupied. The nearly identical energies of these two orbitals of (CO)(4) can be attributed to the fact that both of these MOs are formed from a bonding combination of C-O π* orbitals in four CO molecules. The intrinsically stronger bonding between neighboring carbons in the b(2g) σ MO compared to the a(2u) π MO is balanced by the fact that the non-nearest-neighbor, C-C interactions in (CO)(4) are antibonding in b(2g), but bonding in a(2u). Crossing between an antibonding, b(1g) combination of carbon lone-pair orbitals in four CO molecules and the b(2g) and a(2u) bonding combinations of π* MOs is responsible for the occupation of the b(2g) and a(2u) MOs in (CO)(4). A similar orbital crossing occurs on going from two CO molecules to (CO)(2), and this crossing is responsible for the triplet ground state that is predicted for (CO)(2). However, such an orbital crossing does not occur on formation of (CO)(2n+1) from 2n + 1 CO molecules, which is why (CO)(3) and (CO)(5) are both calculated to have singlet ground states. Orbital crossings, involving an antibonding, b(1), combination of lone-pair MOs, occur in forming all (CO)(2n) molecules from 2n CO molecules. Nevertheless, (CO)(6) is predicted to have a singlet ground state, in which the b(2u) σ MO is doubly occupied and the a(2u) π MO is left empty. The main reason for the difference between the ground states of (CO)(4) and (CO)(6) is that interactions between 2p AOs on non-nearest-neighbor carbons, which stabilize the a(2u) π MO in (CO)(4), are much weaker in (CO)(6), due to the much larger distances between non-nearest-neighbor carbons in (CO)(6) than in (CO)(4).

Density functional theory (DFT) is used to investigate the geometries and metal−ligand bonding in nickel complexes of bidentate phosphines, NiX 2 (R 2 P(CH 2 ) n PR 2 ), where X = H, CO, n = 1−3, and R = H, Me, CF 3 , Et, i-Pr, t-Bu, Ph, OMe, F. The net donor−acceptor properties of the phosphine ligands can be deduced from the computed frequency of the symmetric CO stretch of the Ni(CO) 2 (R 2 P(CH 2 ) n PR 2 ) carbonyl complexes. This frequency (in cm −1 ) can be estimated from the empirical expression ν(CO) = 1988 + ∑χ B − 4n, where the sum is over the four substituents on the bidentate phosphine, χ B is a substituentdependent parameter, and n is the number of carbon atoms in the backbone (1 ≤ n ≤ 3). The deduced values of χ B (in units of cm −1 )t-Bu (0.0), i-Pr (0.8), Et (3.0), Me (4.0), Ph (4.3), H (6.3), OMe (10.8), CF 3 (17.8), and F (18.3)are generally similar to Tolman's electronic parameter χ derived from nickel complexes of unidentate phosphines. For the NiH 2 (R 2 P(CH 2 ) n PR 2 ) hydride complexes, the global minimum is a nonclassical dihydrogen structure, irrespective of the nature of the phosphine. For bidentate phosphines that are strongly donating, a classical cis-dihydride structure lies higher in energy (in some cases, by only 0.4 kcal mol −1 above the global minimum). For phosphines that are less electron donating, the dihydride structure is no longer a local minimum but instead is an inflection point on the potential energy surface. Atoms in molecules (AIM) and natural bond order (NBO) analyses confirm that the nickel−dihydrogen interaction involves a threecenter−two-electron bond. The Kohn−Sham molecular orbital diagram and energy decomposition analysis of these complexes show that metal to H 2 π back-donation is the dominant orbital component for phosphines with electron-donating substituents, whereas H 2 to metal σ donation is dominant for phosphines with electron-withdrawing substituents. The EDA results clearly indicate that long H−H distances are seen when the metal to H 2 π back-donation dominates over H 2 to M σ donation.

(15)) have been computed, using both UB3LYP and (6/6)CASPT2 calculations. ∆E ST in 2c, in which a fourmembered ring is anti-bridged by two allylic radicals, is computed to be larger by a factor of 5 than ∆E ST in 15, in which the anti-bridged ring is five-membered, and by a factor of 10 than that in 12b, in which the anti-bridged ring is six-membered. The reasons for the much larger interaction between two allylic radicals through the bonds of the four-membered ring in 2c than through the bonds of the fivemembered ring in 15 or the six-membered ring in 12b are discussed, and the consequences of the large, through-bond stabilization of the singlet state of 2c are described.

The compound Ti 2 Cl 6 [N(t-Bu) 2 ] 2 (1) has been synthesized by treating TiCl 4 with di(tert-butyl)amine, HN(t-Bu) 2 . Compound 1 crystallizes in two different polymorphs from pentane, both conforming to the space group P2 1 /n. In both polymorphs, 1 exhibits a close Ti•••C contact of 2.634(3) Å between titanium and a γmethyl group in one of the two tert-butyl groups of the bound amido ligand. Interestingly, the γ-methyl group adopts a rotational conformation that maximizes the Ti•••H distances, the shortest of which are 2.36(2) and 2.62(2) Å. Even though the former distance is within the range characteristic of agostic interactions, the rotational orientation of the methyl group suggests that the Ti•••H interactions are repulsive rather than attractive. DFT and NBO analysis confirms this supposition: there is no evidence of weakening of the C−H bond closest to the titanium and no evidence of significant overlap of titanium orbitals with the C−H bonding orbitals of the γ-methyl group involved in the close contact. Further evidence that the close contact is repulsive was obtained from a DFT study of a series of related complexes in which the N(t-Bu) 2 ligand is replaced with a NR(t-Bu) ligand, where the substituent R not involved in the close contact is Et, Me, or SiMe 3 . All of these latter substituents, which are sterically smaller than a t-Bu group, enable the amide group to pivot in such a way as to move the tert-butyl group farther from the metal center. The results suggest that the short Ti•••C and Ti•••H distances seen crystallographically for 1 are actually the result of intraligand and interligand steric repulsions involving the amide substituent not involved in the close contact. The lack of an agostic interaction despite the close contact (and the low electron count of the Ti center) is ascribed to the strong σand π-donor properties of the amide and chloride ligands, which raise the energies of the empty orbitals on Ti.

No abstract