Ionization potential metal carbonyls

There have been fewer studies of the reactions of M ions with potential ligand molecules. Laser ablation, which has been the major ionization source for the production of bare metal ions, produces very few negative ions. Electron impact with low-energy electrons (12 eV) of metal carbonyls has been used to produce [Co(CO)4]- and Fc( CO)4 from Co2(CO)8 and Fe(CO)5. Collision-induced dissociation of these two anions produced Co- and Fc, which could be isolated. Both Co- and Fe were reacted with H2S, aliphatic thiols, aromatic thiols, CS2, and disulfides (153). Reactions with H2S gave the metal monosulfide anion [MS]-, which reacted with H2S by two pathways. 

It is commonly accepted that chemisorption of CO on transition metals takes place in a way that is quite similar to bond formation in metal carbonyls (4). First experimental evidence for this assumption was obtained from a comparison of the C—O stretching frequencies (5) and was later confirmed by data on the bond strength (6) as well as by valence and core level ionization potentials obtained by photoelectron spectroscopy (7). Recent investigations have in fact shown that polynuclear carbonyl compounds with more than about 3-4 metal atoms exhibit electronic properties that are practically identical to those of corresponding CO chemisorption systems (8, 9), thus supporting the idea that the bond is relatively strongly localized to a small number of metal atoms forming the chemisorption site. 

Replacement of CO in a metal carbonyl by a ligand, L, having a lower ionization potential (I.P-) than CO, results in a decrease in the I.P. of the complex. Since the donor ability of L increases as the I.P. decreases, it is not surprising that fragmentation of these complexes occurs by loss of CO rather than of L, and that the relative abundances of M+ are lower (184). 

After our discovery of the metal carbonyl hydrides, other authors (32) pointed out their acidic character in aqueous solution. Potentiometric titrations by Reppe and later by us, showed that in water HCo(CO)4 possesses an acidity (pWa l) comparable to that of nitric acid. The first ionization stage for H2Fe(CO)4 corresponds approximately to that of acetic acid (33), whereas the pentacarbonyl hydrides HM(CO)5 (M = Mn or Re) (VII, 11, 26) are hardly acidic at all. The redox potentials of the cobalt and iron carbonyl hydrides were also measured (33). 

The growing demand for efficient chemical transformations and catalysts has inspired a few research groups in recent years to develop rare earth metal catalysts for organic synthesis [1, 2]. Triflates of rare earth metals are strong Lewis acids, which are stable in aqueous solution. Rare earth metal alkoxides on the other hand are of interest as Lewis bases, e.g. in the catalysis of carbonyl reactions, because of the low ionization potentials (5.4-6.4 eV) and electronegativities (1.1-1.3) of the 17 rare earth elements. Rare earth metal-alkali metal complexes in contrast show both Brpnsted-basic and Lewis-acidic properties. Impressive applications of such catalysts are presented and discussed here. 

Table 3. Vertical ionization potentials (IP) and oxidation (peak) potentials ( ox) of metal carbonyls.

The ionization and oxidation (peak) potentials [53] of metal carbonyls listed in Table 3 establish their mild donicity despite their (formally) neutral oxidation state. However, partial replacement of carbonyls by stronger donor ligands (such as phosphine, sulfide, etc.) effects an incremental increase in their reducing properties [54], Such a fine-tuning of the oxidation potentials of structurally similar organometallic donors is ideal for studies of the correlation between rate constants and the electron-transfer driving force (see Section 2.5). 

This explains why Pt(PF3)4 and Pd(PF3)4 are stable, while the corresponding carbonyls have not yet been prepared. The negative charge on the central atoms of these compounds should, in fact, be quite negligible, so that a back-donation mechanism should not be necessary to stabilize the zerovalent state. We pointed out before that with palladium and platinum the back donation would be difficult because of the high ionization potential of the metals. 

A major improvement was realized with the use of indium, a metal with a very low first ionization potential (5.8 eV) which works without ultrasonic radiation even at room temperature [87]. As the zero-valent indium species is regenerated by either zinc, aluminum, or tin, a catalytic amount of indium trichloride together with zinc, aluminum [88], or tin [89] could be utilized in the allylation of carbonyl compounds in aqueous medium. The regeneration of indium after its use in an allylation process could be readily carried out by electrodeposition of the metal on an aluminum cathode [90], Compared with tin-mediated allylation in ethanol-water mixtures, the indium procedure is superior in terms of reactivity and selectivity. Indium-mediated allylation of pentoses and hexoses, which were however facilitated in dilute hydrochloric acid, produced fewer by-products and were more dia-stereoselective. The reactivity and the diastereoselectivity are compatible with a chelation-controlled reaction [84, 91]. Indeed, the methodology was used to prepare 3-deoxy-D-galacto-nonulosonic acid (KDN) [92, 93], N-acetylneuraminic acid [93, 94], and analogs [95],... 

The correlation between AN and the first M-CO bond strength is seen to be very good. Large values of AN correspond to strong bonds. The smallest values of AN are calculated for metals where no stable carbonyls are known. The results for Cu, Ag and Au would be even less favorable for M-CO bonding if the ionization potential for removing an (n-1) d electron, rather than an ns electron, had been used. 

Bursten and co-workers correlated a wide range of experimental data including not only electrochemical potentials but molecular orbital energies, photoelectron spectroscopy, and ionization energies however, these were focused specifically on metal carbonyls. 

Despite the formally zerovalent state, the homoleptic metal cabonyls are relatively mild donors, consistent with their ionization potentials of 8.5 0.5 V (Table V). [In solution, these complexes are oxidized irreversibly.] Successive introduction of stronger donor ligands (e.g. phosphines, sulfides, and/or isonitriles) results in a decrease in IP and Eox, consistent with increasing electron density on the metal center. [Compare the IP of Mo(CO)6 (8.50 V) [53] with Mo(CO)2 (Tl2-dppe)2 (6.00 V)] [54]. These metal carbonyl derivatives, especially those derived from Cr(CO)6, Mo(CO)6 and Fe(CO)s, form sets of graded electron donors with potentials that can be varied by selection of the number and donor strength of the noncarbonyl ligands [55]. 

In summary, the initial formation of an allylic radical anion on the metal surface is the most likely event, which would explain the success of indium, as its first ionization potential is particularly low E - 5.79 eV). In tin- and indium-mediated reactions the second step should be the insertion of the metal cation into the carbon-bromine (chlorine) bond to afford organometallic intermediates, which are stable enough to be produced, but also highly reactive toward carbonyl compounds in aqueous media. 

---