Topological equivalences for differential graded algebras (Q881958): Difference between revisions

This intriguing paper studies the functor \(H(-)\) which assigns to each differential graded algebra (dga) \(C\) the associated Eilenberg-MacLane ring spectrum \(HC\). This functor sends quasi-isomorphic dgas to weakly equivalent ring spectra; however, it is possible for \(HC\) and \(HD\) to be weakly equivalent even though \(C\) and \(D\) are not quasi-isomorphic. The authors call such dgas \(C\) and \(D\) topologically equivalent. One explanation for this phenomenon (explored in sections 3 and 4) is that the \(k\)-invariants for the Postnikov tower of \(C\) are in Hochschild cohomology, while the \(k\)-invariants for the Postnikov tower of \(HC\) are in topological Hochschild cohomology. There is a map from Hochschild cohomology to topological Hochschild cohomology (Section 4.4), but it need not be an injection; thus a non-trivial \(k\)-invariant in dgas can become trivial in ring spectra. The authors go on to discuss the question of when the module categories for two dgas are Quillen equivalent (which implies they have isomorphic derived categories). They give a complete answer via a topological analog of tilting theory; it involves the notion of topological equivalence. As a sidebar, the authors take real care in constructing the functor \(H(-)\). While it is intuitively obvious that it should exist, the state of the theory of ring spectra is such that it requires a certain amount of care to define.