Chemistry. A tamed reactive intermediate.

Abstract

F or more than 100 years (1), chemists have hypothesized unobserved intermediates in reaction schemes—seemingly invisible chemical spirits that explain the experimental facts. Often, such species are the key to understanding entire areas of chemistry. Indirect evidence may point to their existence, but their high reactivity shortens their lifetime and prevents isolation. Under particular circumstances, reactive intermediates have been isolated and characterized, but establishing the existence of an entirely new class of intermediates is rare indeed. The isolation of a novel reactive intermediate by Rupar et al. on page 1360 of this issue (2) is thus a remarkable and exciting achievement. The organic chemistry of reactive intermediates has focused largely on carbenium ions (R3C +), carbanions (R3C: –), and carbenes (R2C:). Open-shell intermediates (with unpaired electrons) include neutral (R3C •) and charged radicals. Much of the reactivity of these organic intermediates comes from their charges and their deviations from a closed shell of eight electrons. Thus, higher charge and further loss of electrons from the closed shell might imply even greater reactivity and more difficulty in isolation. A series of sixelectron carbon intermediates can be imagined by successively removing a ligand without its electron pair (R+) from the carbenium ion, resulting in R3C +, R2C:, R : , and : :2–. A series of increasingly electron-deficient carbon species is generated by successive removal of ligands with their electron pair (R:–) from the carbanion, resulting in R3C: –, R2C:, RC: +, and C:2+, with eight, six, four, and two valence electrons, respectively. At the end of each series is an unknown, doubly charged atomic ion that lacks all coordination with other ligands or molecules. Of the seven distinct species, only three to date have been isolated and characterized in stable forms. Possibly the most reactive of the unknown species is carbon with two positive charges and only two valence electrons, C:2+. Rupar et al. now report the isolation of the germanium(II) dication, Ge:2+, which is the germanium analog of C:2+ (both germanium and carbon are in group 14 of the periodic table). No previous doubly charged ion of a nonmetal has been reported, except in highly coordinated forms such as R3Ge: 2+ (3). Singly charged, monocoordinated silicon, germanium, and tin cations exist, which are analogs of RC:+ (4, 5). Nonmetallic cations such as Ge:2+ differ from the many metallic cations of general chemistry, such as Na+ and Ca2+, in many important ways. Whereas the metallic cations have four valence electron lone pairs, Ge:2+ possesses just one lone pair (explicitly drawn) and three empty orbitals in the valence shell. The metallic cations are loosely solvated in aqueous solution, whereas the empty orbitals of Ge:2+ would react instantly with any nucleophile such as water. The legendary reactivity of the carbenium and silylium cations (R3C + and R3Si +) derives from their single empty orbital. For decades, chemists have aimed to tame high reactivity by finding friendly environments with low reactivity and by arranging steric protection around the reactive species. The preparation of the free silylium cation relied on these techniques (6). In 1991, Cram et al., in a paper famously entitled “The Taming of Cyclobutadiene,” reported isolation of the highly reactive four-membered ring containing two double bonds by protecting it within a molecular cage or prison (7). The cage lacked molecular components that could react with cyclobutadiene and barred external molecules from reacting with it. Many molecular cages have been used to encapsulate ions or molecules (8), but only a few have served to accommodate reactive intermediates. Rupar et al. now use a cryptand cage, composed of carbon chains interspersed with nitrogen and oxygen atoms, to isolate Ge:2+ from reactive neighbors. But is their germanium cation free, or “naked” as many authors like to call such species? A truly free cation must resemble its analog in vacuum, free of all surrounding species. No atoms can be within bonding distance. This extreme criterion was, for example, fulfilled in the case of the silyl cation (6). In the present case, there is no doubt from the mass spectrum that the imprisoned atom is germanium and from the x-ray data that the triflate counterions are well removed from bonding distances. The remaining suspects are the atoms of the cryptand (see the figure). The carbon and hydrogen atoms are nonreactive and well removed from the germanium, but what about the nitrogen and oxygen atoms? It seems likely that there is some interaction between germanium and these other atoms, as germanium falls midway on the line connecting the two nitrogens and is equidistant from the six oxygens. The experimentally determined distances (Ge–N, 2.524 Å; Ge–O, 2.4856 Å) are well outside typical Ge–N (1.85 Å) and Ge–O (1.80 Å) single-bond lengths. The Ge–O distance, however, is similar to that (2.4 Å) between the solvent O and Ge in the GeCl2-dioxane complex, in which germanium is not considered entirely free (9). The fractional bond orders calculated from the crystallographic parameters are 0.11 for Ge–N and 0.10 for Ge–O, indicating that there is little bonding between Ge and any single atom. Six oxygens and two nitrogens, however, each provide this small amount of bonding, so that the total bonding accumulates. It is noteworthy that germanium is midway between the two nitrogens, indicating a single-minimum rather than a double-minimum potential well. This situation is reminiscent of hydrogen bonds, in which hydrogen can reside in either type of well. Shorter distances favor the single minimum. At longer distances, the hydrogen bond moves to a double minimum (10). It is possible, then, that a larger cryptand would allow germanium to move off center and have stronger bonding interactions with one nitrogen or with one or more oxygen atoms. C : C : The highly reactive germanium dication is isolated by trapping it in a molecular cage. A Tamed Reactive Intermediate

DOI: 10.1126/science.1167321

Cite this paper

@article{Lambert2008ChemistryAT, title={Chemistry. A tamed reactive intermediate.}, author={Joseph B. Lambert}, journal={Science}, year={2008}, volume={322 5906}, pages={1333-4} }