An arXiv preprint in pure mathematics sets out a way to give a concrete algebraic home to a kernel created by a derived tensor operation. Its central result identifies that kernel with the derived category of an exact category, extending the framework to cases where an ordinary ring model is not guaranteed.
The work studies n-term big tilting complexes, algebraic objects used to generate a derived tensor functor between structures associated with rings A and B. It is entirely theoretical: the subject is a collection of rings, complexes, differential graded algebras and derived categories, not an empirical sample or a statistical dataset.
From a tensor kernel to a workable model
To reach its main result, the paper uses a differential graded double-centralizer description together with an exact-realisation theorem. These tools connect the derived tensor construction to a category E whose exact structure can be studied directly. The resulting equivalence concerns the full tensor kernel: the derived category D(E) is equivalent to that kernel, rather than describing only selected examples.
The paper then places that kernel in a wider structure. It proves a recollement, an arrangement in which one derived category is assembled from two others: D(B) is assembled from D(E) and D(A). The result links the kernel to the surrounding algebraic categories instead of leaving it as an isolated construction.
The exact category E is not merely a placeholder. The paper calls it definable and shows that it is closed under filtered colimits, pure submodules, direct summands, products, coproducts and extensions. In less technical language, the proposed model remains within the same class after each of those standard algebraic constructions.
The test for an ordinary-ring model
The key dividing line is whether E is abelian, the condition the paper uses for an ordinary-ring description. Three conditions are shown to be equivalent: E is abelian; the standard t-structure, a way of sorting derived objects by degree, restricts to the tensor kernel; and the kernel is the derived image of a ring map called a homological ring epimorphism.
That equivalence also sets a limit on the result. It does not say that every tensor kernel comes from an ordinary ring. In higher amplitude, when the tilting complex spans a wider range of degrees, E need not be abelian, so the exact category can remain the relevant algebraic model.
One clear sufficient case appears when the relevant perfect model on the B side has amplitude width at most one, meaning that its activity is confined to a narrow range of degrees. Then E is abelian and the equivalent ring-epimorphism conditions hold. In the stated semiperfect-ring setting, the paper also identifies the optimal right B-amplitude from reduced cohomological support, writing it as s minus r.
A related criterion says that concentration in one degree or in two consecutive degrees implies that the tensor kernel is induced by a homological ring epimorphism. The boundary is therefore tied to how tightly the tilting data are concentrated, rather than being guaranteed simply because a tensor kernel has been defined.
Examples put the theory to work
The authors turn next to explicit constructions. They produce genuinely n-term, non-compact big tilting complexes for every n at least two. That shows the higher-term setting addressed by the theory is a broad constructional family, not a single isolated case.
One example gives the abstract model a direct description. There, E consists of B-modules annihilated by an idempotent e, and restriction of scalars along B to C = B/BeB identifies C-Mod with E. In that construction, the exact-category kernel can therefore be described through modules over the quotient algebra C.
A separate three-term construction produces a nontrivial homological epimorphism onto an algebraic Calkin quotient. The example is significant within the paper because it turns the higher-term framework into a specific, nontrivial ring-theoretic map.
Another direct-product construction produces genuine n-term big tilting complexes without adding a further infinite indexing set beyond the two-term building block. The result shows that the higher-term objects can be constructed while keeping that part of the setup controlled.
A theorem, not an empirical claim
Because this is a theoretical paper, there is no empirical sample, comparison group or statistical analysis behind its conclusions. The evidence consists of proofs in derived and triangulated categories and the stated algebraic constructions. The results are conditional theorems, applying under the ring, tilting, projectivity, semiperfectness and amplitude assumptions specified for each result.
The examples do not establish behavior for all rings or all n-term big tilting complexes, and no independent replication is reported. The document is listed as arXiv version 1, dated 20 August 2026, with no journal publication listed in the supplied record. It should therefore be read as a preprint presenting a mathematical framework, not as evidence of an application outside mathematics.
The open questions are correspondingly mathematical: finding broader classes in which E is abelian, identifying further ring-theoretic conditions for ordinary-ring models beyond the stated amplitude criteria, and computing more explicit higher-term kernels and homological ring epimorphisms.
Paper data and sources
Original title: Recollements of derived categories from $n$-term big tilting complexes
Authors: Shengyong Pan, Huabo Xu
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text