A classification of K-local spectra (Q749918)

In earlier work the author showed that it is possible to define localization with respect to any generalized homology theory. When either the theory or the spectrum being localized is not connective, then the result of this construction cannot be described in terms of arithmetic localization or completion. The case of K-theory is particularly interesting because it is possible to give a complete classification of the spectra that can be obtained in this way. This paper is a continuation of the study of this problem begun by the author in Am. J. Math. 107, 895-932 (1985; Zbl 0585.55004). The results of that paper were limited to odd primes. The generalization to the prime 2 is accomplished in the paper under review in the following way. One considers complex K-theory \(K_*\), real K-theory \(KO_*\) and self- conjugate K-theory \(KT_*\) simultaneously, along with certain natural transformations between them. If one localizes at an odd prime, then these other theories are functorially determined by \(K_*.\) These collectively form a functor the author calls `united K-homology theory' \(K_*^{CRT}\). The homological properties of the abelian category where this theory takes its values are developed and it is shown to be more convenient than the target categories of any of the constituent theories. Roughly speaking, a K-theoretically local spectrum X is determined by the structure of \(K_*^{CRT}(X)\) enriched with Adams operations, and each object in the abelian category can be realized by such a spectrum. In particular \(\pi_{*}(X)\) is determined by certain Ext groups.