A mathematical preprint presents an MU-module version of a Viterbo-type relation linking the cotangent bundle of a closed smooth manifold with its free-loop space — the space of loops that start and finish at the same point. It also gives a framed sphere-spectrum version for the same setting. In each case, the central claim is a homotopy equivalence: an identification of the relevant mathematical objects up to the kind of continuous deformation used in homotopy theory.
The paper’s warning is as important as its positive results. It states that the sphere-level construction does not generally recover the canonical MU-level statement after base-change, even when the base is spin. The two formulations therefore capture related structures without making one a universal replacement for the other.
What the paper identifies
The mathematical input is a closed smooth manifold Q and its cotangent bundle, written T*Q. The construction does not assume that Q is oriented or spin, putting the results in a broad non-oriented setting rather than restricting them to manifolds with those additional structures.
At the MU level, the canonical Floer homotopy type of T*Q is identified, in the category of MU-modules, with an MU-module built from the free-loop space LQ. The local system used on the loop-space side includes a suspension by the dimension of Q and a twist arising from the index data. The paper constructs that MU local system through index theory for Cauchy–Riemann operators.
The corresponding homological statement recovers symplectic cohomology as the homology of LQ with a local system called η. A local system serves here as a varying system of coefficients attached to the loop-space construction, rather than a single untwisted coefficient system used throughout.
The framed result is stated for the standard stable R-polarization, a chosen stable geometric background. In spectra, it identifies the framed Floer homotopy type with the suspension spectrum of LQ. Its homological counterpart gives the corresponding integral-homology statement for symplectic cohomology, twisted by the inverse local system.
A proof assembled from finite approximations
This is a theoretical argument, not a statistical study. It works with Floer and Morse moduli spaces, index bundles, flow categories, local systems and directed colimits — constructions used to organize compatible mathematical data. On the free-loop-space side, the topology is approached through finite-dimensional approximations and Morse theory.
Compatibility is a central part of the construction. The Viterbo map is made compatible with the inclusions used in the finite approximations and with Floer continuation maps, allowing the separate stages to fit together into the final statement.
The proof passes through an intermediate spectrum with a spherical complex orientation. The paper states that base-change from MU to this spectrum is conservative, meaning that the relevant equivalence can be detected after moving to that intermediate setting. A spectral Whitehead theorem for bounded-below HZ-module spectra is then used to promote the constructed map to a homotopy equivalence.
The paper presents the MU local system as an index-theoretic construction. It leaves open the possibility that the local system may have a purely homotopy-theoretic origin, so the preprint does not settle that deeper explanation.
Why sphere-level data fall short
In the paper’s account, the difference between the two formulations is structural rather than cosmetic. Sphere-level base-change cannot generally be treated as a shortcut to the canonical MU-level object. The stated failure persists even for spin bases.
The preprint reports a cited prior obstruction for complex projective space CP^n when n is an odd integer at least 3. In that case, the canonical MU-Floer type cannot be equivalent to any sphere-spectrum object after tensoring with MU. The associated MU local system also cannot be lifted to BGL1(S), the classifying space appearing in the sphere-spectrum formulation. The paper attributes this statement to the cited result BP26.
That example gives the limitation a precise form. The obstruction applies both to the full MU-Floer type and to the local system used to build the loop-space object, showing why the MU construction cannot generally be reconstructed from sphere-spectrum data by base-change alone.
The edges of the result
The framed theorem is tied to the standard stable R-polarization. The paper emphasizes that framed Floer homotopy type depends on the chosen stable polarization, so the result does not establish the same identification for every possible choice.
The non-oriented construction also comes with a symmetry cost. It fixes a straight line on the cylinder, explicitly breaking circle symmetry. The paper identifies equivariant refinements as an open direction and does not establish a generally circle-equivariant or framed-E2-compatible Viterbo isomorphism.
The supplied document is an arXiv version-one preprint dated 20 Aug 2026. Journal publication and peer-review status are not reported in the supplied material. There are no participants, empirical dataset or statistical validation; the conclusions are mathematical homotopy-equivalence statements within Floer homotopy theory, not experimental evidence.
The paper leaves open whether the MU local system can be constructed purely homotopy-theoretically, whether the suggested equivariant structures can be established in the non-oriented setting, and whether an MU-level Viterbo isomorphism can incorporate framed E2 structures without breaking circle symmetry.
The work was partially supported by an NSF Graduate Research Fellowship award.
Paper data and sources
Original title: Spectral Viterbo isomorphism: complex-oriented versus framed
Authors: Kenneth Blakey
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text