Complexes of the form RhX(CO)(PR 3 ) 2 [X = Cl, Br or I; R = Me or Ph] reacted with H 2 to form a series of binuclear complexes of the type (PR 3 ) 2 H 2 Rh(µ-X) 2 Rh(CO)(PR 3 ) [X = Cl, Br or I, R = Ph; X = I, R = Me] and (PMe 3 ) 2 (X)-HRh(µ-H)(µ-X)Rh(CO)(PMe 3 ) [X = Cl, Br or I] according to parahydrogen sensitised 1 H, 13 C, 31 P and 103 Rh NMR spectroscopy. Analogous complexes containing mixed halide bridges (PPhare detected when RhX(CO)(PPh 3 ) 2 and RhY(CO)(PPh 3 ) 2 are warmed together with p-H 2 . In these reactions only one isomer of the products (PPh 3 ) 2 H 2 Rh(µ-I)(µ-Cl)Rh(CO)(PPh 3 ) and (PPh 3 ) 2 H 2 Rh(µ-I)-(µ-Br)Rh(CO)(PPh 3 ) is formed in which the µ-iodide is trans to the CO ligand of the rhodium() centre. When (PPh 3 ) 2 H 2 Rh(µ-Cl)(µ-Br)Rh(CO)(PPh 3 ) is produced in the same way two isomers are observed. The mechanism of the hydrogen addition reaction is complex and involves initial formation of RhH 2 X(CO)(PR 3 ) 2 [R = Ph or Me], followed by CO loss to yield RhH 2 X(PR 3 ) 2 . This intermediate is then attacked by the halide of a precursor complex to form a binuclear species which yields the final product after PR 3 loss. The (PPh 3 ) 2 H 2 Rh(µ-X) 2 Rh(CO)(PPh 3 ) systems are shown to undergo hydride self exchange by exchange spectroscopy with rates of 13.7 s Ϫ1 for the (µ-Cl) 2 complex and 2.5 s Ϫ1 for the (µ-I) 2 complex at 313 K. Activation parameters indicate that ordering dominates up to the rate determining step; for the (µ-Cl) 2 system ∆H ‡ = 52 ± 9 kJ mol Ϫ1 and ∆S ‡ = Ϫ61 ± 27 J K Ϫ1 mol Ϫ1 . This process most likely proceeds via halide bridge opening at the rhodium() centre, rotation of the rhodium() fragment around the remaining halide bond and bridge re-establishment. If the triphenylphosphine ligands are replaced by trimethylphosphine distinctly different reactivity is observed. When RhX(CO)(PMe 3 ) 2 [X = Cl or Br] is warmed with p-H 2 the complex (PMe 3 ) 2 (X)HRh(µ-H)(µ-X)Rh(CO)(PMe 3 ) [X = Cl or Br] is detected which contains a bridging hydride trans to the rhodium() PMe 3 ligand. However, when X = I, the situation is far more complex, with (PMe 3 ) 2 H 2 Rh(µ-I) 2 Rh(CO)(PMe 3 ) observed preferentially at low temperatures and (PMe 3 ) 2 (I)HRh(µ-H)-(µ-I)Rh(CO)(PMe 3 ) at higher temperatures. Additional binuclear products corresponding to a second isomer of (PMe 3 ) 2 (I)HRh(µ-H)(µ-I)Rh(CO)(PMe 3 ), in which the bridging hydride is trans to the rhodium() CO ligand, and (PMe 3 ) 2 HRh(µ-H)(µ-I) 2 Rh(CO)(PMe 3 ) are also observed in this reaction. The relative stabilities of related systems containing the phosphine PH 3 have been calculated using approximate density functional theory. In each case, the (µ-X) 2 complex is found to be the most stable, followed by the (µ-H)(µ-X) species with hydride trans to PH 3 .

Density functional theory is used to probe the reaction of O(2) with the unsaturated transition-metal fragments [M(bipy)(2)](2+), M = Cr, Fe. In both cases, calculations indicate that the O(2) molecule is initially trapped as an eta(2)-bound superoxide ion, where the unpaired electron in the out-of-plane pi orbital of O(2) is weakly coupled to those on the trivalent metal ion. In the chromium case, a cis-dioxo Cr(VI) complex is found to be significantly more stable than the superoxo species. The two minima are, however, separated by a large barrier, along with a change in spin state. For the iron analogue, the relative energies of the two minima are reversed, the superoxo complex being the global minimum. The energetics of the O(2) activation processes are consistent with previously reported mass spectrometric experiments, where an adduct, [M(bipy)(2)(O(2))](2+), was detected only for chromium.