Bounded \(t\)-structures on the category of perfect complexes (Q7028451)

This paper is concerned with the relation between the regularity of a scheme and the existence of bounded \(t\)-structures on the category of perfect complexes of the scheme. The author proves the following main result labeled as Theorem 0.1:\N\NFor a noetherian, finite-dimensional scheme \(X\) and a closed subset \(Z \subseteq X\), the category \( \mathsf{D}^{\mathrm{perf}}_Z (X) \) of perfect complexes on \( X\) whose cohomology is supported on \( Z \) admits a bounded \(t\)-structure if and only if \(Z\) is contained in the regular locus of \(X\).\N\NNote that it is well known that, under the assumptions of the theorem, there exists a ``standard bounded \(t\)-structure'' on \( \mathsf{D}^{\mathrm{perf}}_Z (X) \) when \( Z \) is contained in the regular locus of \( X \). This structure is inherited from the canonical inclusion of triangulated categories \( \mathsf{D}^{\mathrm{perf}}_Z (X) \subseteq \mathsf{D}^b_{\mathrm{coh},Z} (X) \), where the latter category is the bounded derived category of coherent sheaves on \(X\) with cohomology supported on \(Z\). Thus, the other direction, i.e. the existence of a bounded \(t\)-structure implying the regularity, is the heart of the proof. In short, the author proves that the existence of a bounded \(t\)-structure on \( \mathsf{D}^{\mathrm{perf}}_Z (X) \) implies the other inclusion \( \mathsf{D}^b_{\mathrm{coh},Z} (X) \subseteq \mathsf{D}^{\mathrm{perf}}_Z (X) \).\N\NThis result is motived by the Conjecture 1.5 of [\textit{B. Antieau} et al., Invent. Math. 216, No. 1, 241--300 (2019; Zbl 1430.18009)] which posed this question for the case \( Z = X\). Hence, the result that the author provides is more general than just a positive answer to this conjecture.\N\NThe proof of the main theorem, which occupies section 3, is only the beginning; the author goes on to explore conditions under which all bounded \(t\)-structures on \( \mathsf{D}^b_{\mathrm{coh},Z} (X) \) are equivalent. The first result in this direction is related to the main theorem; the bounded \(t\)-structure of \( \mathsf{D}^{\mathrm{perf}}_Z (X) \), if it exists, is unique (up to equivalence). Hence, \(Z\) meets the singular locus of \( X \) if and only if there are no bounded \(t\)-structures on this category. Therefore, if \(Z\) sits in the regular locus there exists a unique bounded \(t\)-structure on \( \mathsf{D}^b_{\mathrm{coh},Z} (X) =\mathsf{D}^{\mathrm{perf}}_Z (X) \). Moreover, the author shows that all bounded \(t\)-structures on \(\mathsf{D}^b_{\mathrm{coh},Z} (X) \) are equivalent either when \(X\) admits a dualizing complex (Theorem 4.2) or when Z=X seperated and quasiexcellent (Theorem 5.4). All these results stem from the nice observation that is Lemma 2.5.\N\NThe last three sections of the article serve as (1) an appendix for a complete proof of Theorem 3.2, which is crucial in the proof of the main theorem, and (2) results and applications regarding the approximability of \( \mathsf{D}_{\mathsf{qc},Z} (X) \), the category of quasi-coherent sheaves on \(X\) with cohomology supported on \(Z\). The latter is not needed for the results explained above and it is addressed to experts and whomever is interested in this theory which, quoting the author, is a work-in-progress. It is worth pointing out that Theorem 3.2 is a major generalisation of old results and it is fully proved in its new form using the current ``state of the art'' tools.\N\NThe author provides a concise description of the theory needed regarding \(t\)-structures on triangulated categories, as well as the author's own terminology and work used for proofs, especially from Sections 1 and 2 of \textit{A. Neeman} [Ann. Math. (2) 193, No. 3, 689--732 (2021; Zbl 1478.18014); ``Triangulated categories with a single compact generator and a brown representability theorem'', Preprint, \url{arXiv:1804.02240}] regarding the approximability part. Nevertheless, the reader is advised to have some familiarity with these concepts.