Possible Boron Hydride Ions

It is easy to extend the foregoing topological arguments to boron hydride ions (4 )- Consider ions of the charges c = —2, —1, 1, and 2 in the form BpR p+ +c. The modified equations of balance are s + a = q +c, s +t = p + c, and t + y = p — c —q/2. The structures derived by the procedures described in Section VI can easily be drawn in topological form from the four numbers s, t, y, and x, which precede the formula. These possible ions are classified as the following types  

While the above classifications and lists probably do not exhaust all possibilities, it may be hoped that the omissions do not overlook any obviously promising stable ions. For example, we have not considered ions based on linked polyhedra or linked fragments of polyhedra. We have not tried to list completely the results of applying the processes of classes (c), (d), and (e) to all ions listed above. Also, we have not 

---

Lipscomb, W. N., Possible Boron Hydride Ions. J. Phys. Chem. 1958, 62, 381-382. 

For some mass numbers, it might be possible to resolve several peaks arising from different ions. In the spectra of boron hydrides, for example, it is often difficult to sort out what is going on, because boron has isotopes and B, and of course °BH and B are both of mass 11 Da. But not quite. The two masses are actually 11.0209 and 11.0093 Da, and so with high-resolution instruments the various ions contributing to one composite peak can be separately identified. 

It is necessary for the intermediate cation or complex to bear considerable car-bocationic character at the carbon center in order for effective hydride transfer to be possible. By carbocationic character it is meant that there must be a substantial deficiency of electron density at carbon or reduction will not occur. For example, the sesquixanthydryl cation l,26 dioxolenium ion 2,27 boron-complexed imines 3, and O-alkylated amide 4,28 are apparently all too stable to receive hydride from organosilicon hydrides and are reportedly not reduced (although the behavior of 1 is in dispute29). This lack of reactivity by very stable cations toward organosilicon hydrides can enhance selectivity in ionic reductions. 

The Lewis basic carbonyl group forms a complex with the empty p orbital of the Lewis acidic borane. Hydride transfer is then possible from anionic boron to electrophilic carbon. The resulting tetrahedral intermediate collapses to an iminium ion that is reduced again by the borane. 

The codeposition of elements, such as phosphorus or boron, by direct interaction of the reducing agent with the catalytic surface was examined by some authors however, the interpretations are not completely satisfying. An indirect mechanism for hypophosphite reduction was proposed by Zeller and Landau.23 In this approach formation of phosphine as an intermediate and a direct interaction with radicals such as hydrogen atom H, hydride H , or NiOH was suggested.5, 12 24 26 However, the possible direct reduction of phosphite to phosphorous must also be considered. In favor of this hypothesis, there is the electrodeposition of NiP alloys from solutions containing Ni2+ and phosphite. The interaction of Ni2+ with the phosphite ions produces, at high pH, nickel phosphate precipitate when nickel ions are not adequately complexed. 

Meerwein was the first to succeed in obtaining dioxolanylium ions of type 2, sufficiently stabilized as salts with non-polarizable anions that they could be isolated crystalline. The compounds can be prepared by splitting out of an anion from cyclic ortho esters or acetals wherein the required ring-system is already present. The ortho ester 1 reacts with antimony pentachloride or boron trifluoride, with splitting out of OR, to give 2. Acetals (3) from aldehydes can be converted, by hydride abstraction with triphenylmethyl or triethyl-oxonium fluorohorate, into salts (2) this reaction proceeds well only with acetals of the 1,3-dioxolane type (3) that have little steric hindrance. With acetals of the 1,3-dioxane type, formed from aldehydes, the reaction of hydride abstraction is not, as a rule, possible. In all such reactions, the anion involved is either SbClg or BF4 . 

---