This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Donovan–Karoubi | |
|---|---|
| Name | Donovan–Karoubi |
| Field | Algebraic topology; K-theory |
| Introduced | 1970s |
| Authors | Patrick Donovan; Max Karoubi |
Donovan–Karoubi is a construction in algebraic topology and operator algebras that links vector bundle classification, projective module theory, and cohomological twisting in K-theory. It arose in the study of topological obstruction classes and central simple algebras over spaces, providing a bridge between geometric bundles on CW complexes, gerbes on manifolds, and analytic families of Fredholm operators. The construction has influenced developments in twisted K-theory, Atiyah–Singer index theorem, and classifications related to Brauer group elements.
The Donovan–Karoubi framework formalizes how torsion classes in cohomology detect obstructions to lifting PU(n)-bundles to U(n)-bundles and how such obstructions correspond to classes in the Brauer group and to twists of K-theory. It connects work of Michael Atiyah, Israel Gelfand, Daniel Quillen, Jean-Louis Koszul, and Jean-Pierre Serre on projective modules, and it complements perspectives from Alain Connes and Gennadi Kasparov in noncommutative topology. The theory plays a role alongside the Chern character, the Bockstein homomorphism, and the Steenrod algebra in detecting secondary invariants.
Donovan and Karoubi developed their ideas in the 1970s influenced by earlier classification results of Hassler Whitney, John Milnor, and Raoul Bott. The motivating problems included lifting problems studied by I. M. Singer and obstruction theories formalized by Jean Leray and Samuel Eilenberg. Their papers responded to questions raised in the wake of the Atiyah–Hirzebruch spectral sequence and the emerging picture of twisted cohomology appearing in work by Daniel Freed, Greg Moore, and Edward Witten. Interactions with the literature on the Brauer group of schemes from Alexander Grothendieck and on central simple algebra classification from Richard Brauer also informed the formulation.
At its core, the Donovan–Karoubi construction associates to a principal PU(n), or more generally a principal PGL(n)-bundle over a paracompact CW complex or smooth manifold, a twisting class in integral cohomology that obstructs the existence of an associated vector bundle with structure group U(n) or GL(n,C). The obstruction takes values in the torsion subgroup of H^3(X;Z) connected to the Brauer group of topological space X and is computed via a Bockstein from exact sequences akin to those used by Henri Cartan and Samuel Eilenberg–MacLane. Key properties echo results of Max Karoubi on graded algebras and Bott periodicity as proven by Raoul Bott and Michel Atiyah: the Donovan–Karoubi class behaves functorially under pullback along maps between CW complexes, is stable under stabilization by trivial bundles studied by John Milnor, and interacts compatibly with the Atiyah–Hirzebruch spectral sequence and with operations studied by Jean-Pierre Serre.
Standard examples include the obstruction for lifting the Hopf fibration and for projective bundles over complex projective spaces such as CP^n. For a principal PU(1)=1 case the obstruction vanishes, while nontrivial examples arise on lens spaces analyzed by Poincaré-type constructions and by computations using the Serre spectral sequence as in work of Jean-Pierre Serre and Armand Borel. Calculations for low-dimensional manifolds like S^3, S^2 × S^1, and T^n (the n‑torus studied by Carl Friedrich Gauss-related lattice problems) illustrate how torsion in H^3 produces nontrivial Donovan–Karoubi classes. Explicit connections to Chern–Simons invariant computations and to Dixmier–Douady class evaluations demonstrate the role of the theory in concrete index problems considered by Atiyah and Dixmier.
The Donovan–Karoubi classes provide the prototypical twists used to define twisted versions of topological K-theory introduced by Michael Atiyah and Graeme Segal. When a space carries a nontrivial Donovan–Karoubi class, ordinary vector bundle-valued K-theory is replaced by twisted K-theory groups that account for projective module refinements analogous to the passage from C*-algebras to continuous trace C*-algebras classified by the Dixmier–Douady invariant studied by Jacques Dixmier and Adrien Douady. These twists relate to the formulation of the Atiyah–Singer index theorem for families on spaces with gerbe background studied by Daniel Freed and Gerald Moore, and they interface with the Kasparov product in KK-theory as developed by Gennadi Kasparov.
Applications span classification problems for continuous-trace C*-algebras in the tradition of Victor Nistor and Iain Raeburn, to the role of twists in string-theoretic constructions by Edward Witten and Alain Connes in noncommutative geometry. Subsequent developments link Donovan–Karoubi concepts to gerbe theory advanced by Jean Giraud, to higher categorical formulations in the work of Jacob Lurie and André Henriques, and to refined invariants in equivariant settings studied by Gunnar Wasserman and Gordon Segal. Modern research continues to explore computational frameworks from stable homotopy theory influenced by J. Peter May and connections to derived algebraic geometry promoted by Bertrand Toën and Gabriele Vezzosi.