The strong ultrafilter topology on spaces of ideals (Q298036)

Results concerning the use of topology to study the prime spectrum of a commutative ring are long-standing and extensive, starting from the pioneering work by M. H. Stone, O. Zariski and N. Jacobson among others. The Zariski topology and the constructible (or patch) topology are especially well-known after the work by H. Cartan and C. Chevalley, A. Grothendieck and M. Hochster. Recently, \textit{M. Fontana} and \textit{K. A. Loper} [Commun. Algebra 36, No. 8, 2917--2922 (2008; Zbl 1152.13003)] introduced a new topology (called the ultrafilter topology) on the prime spectrum of a ring using ultrafilters and demonstrated that it is equivalent to the constructible (or, patch topology). In the present paper, the authors investigate more deeply at the way that ultrafilters are used to define closure in the ultrafilter topology (mentioned above) and define two refinements called the strong ultrafilter topology and the Goldman topology. Note that the Goldman topology was essentially introduced by \textit{G. Picavet} [Ann. Sci. Univ. Clermont-Ferrand II, Math. 57(11), 73--90 (1975; Zbl 0317.13002)]. The authors declare that the inspiration for their definitions comes from the fact that it often happens in Noetherian rings that a prime ideal can be expressed as the intersection of properly larger primes and the intersection prime can be viewed as a limit of the primes that were intersected. On the other hand, in rings that are not Noetherian a prime can be a limit point in the ultrafilter topology without being representable as an intersection of primes. These two notions give rise to the two topologies investigated in the present paper, Goldman (intersection) and strong (non-intersection) and particular attention is devoted to the case of the strong ultrafilter topology. With this inspiration, the authors study the connections with the Noetherian motivating ideas and demonstrate that the strong ultrafilter topology on the prime spectrum of a ring \(R\) gives rise to a discrete space if and only if \(R\) has a Noetherian spectrum with respect to the Zariski topology. They also show that the strong topology is identical to the ultrafilter topology if and only if all prime ideals of the ring are G-ideals, where a G(oldman)-ideal is a prime ideal which is properly contained in the intersection of all properly larger primes. In this case, the ``substance'' of the ultrafilter topology comes entirely from the strong ultrafilter topology. As an application of their theory, the authors obtain a characterization, using ultrafilters, of which subsets of Spec(\(R\)) are compact in the Zariski topology. Another application concerns Grothendieck's notion of a retrocompact set: using the two new topologies, they characterize when a subset of Spec(\(R\)) is retrocompact.