LITHIUM

PAUL V. R. SCHLEYER, UNIVERSITY OF GEORGIA

The third chemical element, lithium, was discovered in 1817 in a rocklike ("lithos") mineral, petalite, by J. August Arfvedson in J. J. Berzelius' laboratory in Stockholm. Uses, while currently ranging from batteries to antidepression treatments, developed very slowly. Unlike its close neighbors, hydrogen and helium, lithium has not been detected in interstellar space. Currently, adequate deposits exist on Earth in certain minerals and in brines. Lithium also is a trace element in the human body.

A century passed before Wilhelm Schlenk prepared the first alkyl lithium derivatives from organomercury precursors. Henry Gilman and Georg Wittig made important contributions to the development of the synthetic capabilities of lithium derivatives, which eventually replaced Grignard reagents as the primary anionic reactive intermediates. Preparative uses blossomed after both discovered independently that many compounds could be lithiated directly by hydrogen-metal exchange. This bypassed the often cumbersome preparations of Grignard reagents from magnesium metal and halogen derivatives. The latter had to be synthesized separately from the parent substrate.

With the readily available n-butyl lithium (used commercially as an anionic polymerization catalyst) typically serving as the reagent, hydrogen-lithium exchange often takes place specifically at positions in the vicinity of functional groups, where it is difficult to introduce a halogen. Such "directed metallations" are favored mechanistically as a result of the lowering of the metallation transition-state energy through lithium complexation with the functional group. Polar solvents and accelerators often, but not always, speed up metallation substantially and may even influence the site of substitution. Empirical trials and experience, rather than understanding, long guided the synthetic chemist.

One of my early mechanistic papers trying to unravel such mysteries computationally met the scorn of a referee: "Who cares what is in the pot, as long as it works." This comment equated lithium chemists with the witches in Shakespeare's "Macbeth." But it was my involvement with the highly peculiar structures of organolithium compounds that initiated my interest in lithium in the 1970s. At that time, alkyl lithium compounds were believed to be covalent. The evidence seemed clear. In contrast to insoluble ionic sodium and potassium alkyls with limited preparative utility, lithium alkyls are soluble in hydrocarbons and, when pure, either are liquids (n-butyl lithium) or have low melting points (ethyl lithium). The X-ray structures show tetramers with hexacoordinate carbons! Andrew Streitwieser pointed out later that a simple ionic cluster model with interpenetrating anionic and cationic tetrahedra (as in rock salt) could rationalize such structures. The hydrocarbon exterior of (RLi)₄ tetramers explains their alkane solubility.

John Pople and I joined forces in the early 1970s. Our research groups employed Pople's newly developed Gaussian 70 ab initio program to computionally discover, for example, doubly bridged dilithioacetylene and a dilithiocyclopropane with a planar tetracoordinate carbon. Octahedral CLi₆, with six equivalent bonds to carbon, is one example of "hypermetallation" out of many. These predictions were verified, at least in part, experimentally. Such rule-breaking structures illustrate the interplay of ionic and covalent bonding with some Li-Li interactions; the octet rule is not violated.

After my move to the University of Erlangen-Nuremberg in 1976, my coworkers added to the relatively small number of X-ray structures. Each new result revealed some new feature, which required a detailed computational study to understand. This complexity is still true. Lithium compounds are "self-assembling molecules" par excellence. They can aggregate in a variety of ways and bind not only to polar solvents (the "accelerators" mentioned above) and to benzene, but also to the substrates before reaction. The "witches brew" also can be clarified computationally, as can the reaction pathways and transition states. Attacking the planar tetracoordinate carbon problem again, we took advantage of computer modeling to design promising chelated (internally solvated) organolithium compounds. Their subsequent synthesis and X-ray analysis verified our prediction.

A chemical world based on electrostatic interactions, rather than just covalent bonds, is disclosed by the geometries, the bonding, and the course of reactions of lithium compounds. A goal of science is not just to understand what has happened, but to predict the outcome quantitatively.

Paul v. R. Schleyer is Graham Perdue Professor at the University of Georgia and professor emeritus of the University of Erlangen-Nuremberg, in Germany. He was the first recipient of the Arfvedson-Schlenk Award of the Society of German Chemists sponsored by Chemetall.

Top

Chemical & Engineering News
Copyright © 2003 American Chemical Society