Abstract
We extend the criterion of Kawatani and Okawa for indecomposability of the derived category of a smooth projective variety to arbitrary schemes. For relative schemes, we also give a criterion for the nonexistence of semiorthogonal decompositions that are linear over the base. These criteria are based on the base loci of the global or relative dualizing complexes.
Similar content being viewed by others
References
Bondal, A. I., Kapranov, M. M.: Representable functors, Serre functors, and reconstructions. Izv. Akad. Nauk SSSR Ser. Mat. 53, 1183–1205 (1989). English transl. in Math. USSR-Izv. 35, 519–541 (1990)
Catanese, F.: Pluricanonical–Gorenstein-curves. In: Enumerative Geometry and Classical Algebraic Geometry (1982) (Nice,: vol. 24 of Progr. Math. Birkhäuser Boston, Boston, MA, pp. 51–95, 1981)
Catanese, F., Franciosi, M., Hulek, K., Reid, M.: Embeddings of curves and surfaces. Nagoya Math. J. 154, 185–220 (1999)
Caucci, F.: Paracanonical base locus, Albanese morphism and semiorthogonal indecomposability of derived categories (2022). arXiv:2110.06795v2
Cossec, F. R., Dolgachev, I.: Enriques surfaces. I. In: Progress in Mathematics, Vol. 76, pp. x+397. Birkhäuser Boston Inc., Boston, MA (1989)
Elagin, A., Lunts, V.: Regular subcategories in bounded derived categories of affine schemes. Sbornik Math. 209(12), 1756–1782 (2018)
Hernández Ruipérez, D., López Martín, A.C., de Salas, F.S.: Relative integral functors for singular fibrations and singular partners. J. Eur. Math. Soc. JEMS 11, 597–625 (2009)
Kawatani, K., Okawa, S.: Nonexistence of semiorthogonal decompositions and sections of the canonical bundle (2018). arXiv:1508.00682v2
Kodaira, K.: On compact analytic surfaces. II, III. Ann. Math. 77(2), 563–626 (1963)
Kuznetsov, A.: Hyperplane sections and derived categories. Izv. Ross. Akad. Nauk Ser. Mat. 70(3), 23–128 (2006); translation in Izv. Math. 70(3), 447–547 (2006) (Russian)
Kuznetsov, A.: Semiorthogonal decompositions in algebraic geometry. In: Proceedings of the International Congress of Mathematicians—Seoul 2014, vol. II pp. 635–660. Kyung Moon Sa, Seoul (2014)
Kuznetsov, A.: Semiorthogonal decompositions in families (2021). arXiv:2111.00527
Kuznetsov, A., Shinder, E.: Homologically finite-dimensional objects in triangulated categories (2022). arXiv:2211.09411
Kuznetsov, A., Shinder, E.: Categorical absorptions of singularities and degenerations (2022). arXiv:2207.06477
Lin, X.: On existence of semiorthogonal decompositions in algebraic geometry (2021). arXiv:2107.09564
Lipman, J.: Notes on derived functors and Grothendieck duality. In: Foundations of Grothendieck duality for Diagrams of Schemes, Lecture Notes in Mathematics, vol. 1960, pp. 1–259. Springer, Berlin (2009)
Neeman, A.: The Grothendieck duality theorem via Bousfield’s techniques and Brown representability. J. Am. Math. Soc. 9(1), 205–236 (1996)
Okawa, S.: Semi-orthogonal decomposability of the derived category of a curve. Adv. Math. 228, 2869–2873 (2011)
Okawa, S.: Semiorthogonal indecomposability of minimal irregular surfaces (2023). arXiv:2304.14048v1
Orlov, D.O.: Triangulated categories of singularities, and equivalences between Landau–Ginzburg models. Mat. Sb. 197, 117–132 (2006)
Pirozhkov, D.: Stably semiorthogonally indecomposable varieties (2023). arXiv:2011.12743v3
Ran, Z.: Canonical systems and their limits on stable curves. J. Algebra 399, 634–656 (2014)
Sancho de Salas, F.: Koszul complexes and fully faithful integral functors. Bull. Lond. Math. Soc. 41, 1085–1094 (2009)
Spence, D.: A note on semiorthogonal indecomposability of some Cohen–Macaulay varieties. J. Pure Appl. Algebra 226, 107076 (2022)
Acknowledgements
We would like to thank Evgeny Shinder for his comments and suggestions and for pointing out a mistake in the proof of Theorem 3.3 in an earlier version of the paper. We are grateful to Alexander Kuznetsov for his comments and for reading the manuscript. We also thank Shinnousuke Okawa for his interest in this work and for sharing with us the draft of his article.
Funding
Work supported by Grant PID2021-128665NB-I00 funded by MCIN/AEI/10.13039/501100011033 and, as appropriate, by “ERDF A way of making Europe”.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Indecomposability of semistable curves (by Shinnosuke Okawa)
Appendix A: Indecomposability of semistable curves (by Shinnosuke Okawa)
Let X be a semistable curve over an algebraically closed field \({\textbf {k}}\); namely, a connected projective nodal curve such that the relative dualizing sheaf \(\omega _{ X / {\textbf {k}}}\) of X over \({\textbf {k}}\) is nef, in the sense that \(\deg \omega _{ X / {\textbf {k}}} \vert _{ C } \ge 0\) for any irreducible component \(C \hookrightarrow X\). It is shown in Corollary 4.4 that \({{\text{Perf }}}(X)\) admits no semiorthogonal decomposition provided that \(\omega _{ X /{\textbf {k}}}\) is ample and there is no LCRT. The aim of this appendix is to answer the question by Alexander Kuznetsov whether Corollary 4.4 generalizes to all semistable curves. The author is indebted to him for asking the question and for his useful comments. The author was partially supported by JSPS Grants-in-Aid for Scientific Research (18H01120, 19KK0348, 20H01797, 20H01794, 21H04994, 23H01074).
Proposition 5.1
Let X be a nodal connected projective curve over an algebraically closed field \({\textbf {k}}\). Then \({{\text{Perf }}}(X)\) is semiorthogonally indecomposable if and only if X is semistable.
For the sake of completeness, we first show the following easy lemma. Below \({{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\) denotes the identity component of the Picard scheme.
Lemma 5.2
Let X be a reduced and connected projective curve over \({\textbf {k}}\) and \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\) such that \(L \not \simeq \mathcal{O}_{ X }\). Then \(H ^{ 0 } \left( X, L \right) = 0\).
Proof
Suppose \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X \right)\) admits a global section \(0 \ne s \in H ^{ 0 } \left( X, L \right)\).
If \(s \vert _{ C } \ne 0\) for any irreducible component \(C \hookrightarrow X\), then obviously s must be nowhere vanishing. This would imply \(L \simeq \mathcal{O}_{ X }\).
Suppose that \(s \vert _{ C } = 0\) for some irreducible component C. We can assume without loss of generality that there is another component \(D \hookrightarrow X\) which intersects C and \(s \vert _{ D } \ne 0\). Since \(L \vert _{ D } \in {{\,\mathrm{{Pic}}\,}}^{ 0 } \left( D \right)\), it follows that \(s \vert _{ D }\) is a nowhere vanishing global section of \(L \vert _{ D }\). This contradicts \(C \cap D \ne \emptyset\) and \(s \vert _{ C } = 0\). \(\square\)
Lemma 5.3
Let X be a nodal projective curve over \({\textbf {k}}\) such that \(\dim _{ {\textbf {k}}} H ^{ 1 } \left( X, \mathcal{O}_{ X } \right) > 0\). Then there exists an invertible sheaf \(L \not \simeq \mathcal{O}_{ X }\) such that \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\).
Proof
Since \(\dim X = 1\), it follows that \(H ^{ 2 } \left( X, \mathcal{O}_{ X } \right) = 0\). The deformation theory of coherent sheaves implies that \({{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X \right)\) is a smooth connected group scheme of dimension \(\dim _{ {\textbf {k}}} H ^{ 1 } \left( X, \mathcal{O}_{ X } \right) > 0\). Hence we can take L to be the line bundle corresponding to any point of \({{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X \right) {\setminus } \left\{ \mathcal{O}_{ X } \right\}\). \(\square\)
Proof of Proposition 5.1
Let us first assume X is semistable, and show that \({{\text{Perf }}}(X)\) is indecomposable. As in Theorem 1.4 in [15], let us consider the paracanonical base locus:
As in the smooth case, if one can show that \(| \textsf {PBs}( X ) | < \infty\), then the semiorthogonal indecomposability of \({{\text{Perf }}}(X)\) follows. This follows from that any semiorthogonal decomposition of \({{\text{Perf }}}(X)\) is stable under the action of \({{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X \right)\) by Theorem 3.8 in [8]. Indeed, take a closed point \(x \in X _{ \text{sm} } {\setminus } \textsf {PBs}( X )\). Let \({{\text{Perf }}}(X) = \langle \mathcal{A}, \mathcal{B}\rangle\) be a semiorthogonal decomposition and
be the corresponding distinguished triangle. Then for any \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\) it follows that \(b \otimes L \in \mathcal{B}\) and \(a \otimes L \in \mathcal{A}\) by the stability. Hence from (5.2) \(\otimes L\) we get \(b \otimes L \simeq b\) and \(a \otimes L \simeq a\). By assumption there are \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\) and \(s \in H ^{ 0 } \left( X, L \otimes \omega _{ X / {\textbf {k}}} \right)\) such that \(s ( x ) \ne 0\). Thus we get a map
which is an isomorphism on an open neighborhood of x. Then by standard arguments (see, say, the proof of Theorem 3.3) we can conclude that \(\mathcal{O}_{ x }\) is contained in either \(\mathcal{A}\) or \(\mathcal{B}\). Then, finally, one can show that all points of \(X _{ \text{sm} } {\setminus } \textsf {PBs}( X )\) are simultaneously contained in either \(\mathcal{A}\) or \(\mathcal{B}\) by the arguments of Corollary 3.9.
In the rest we show that \(\textsf {PBs}( X )\) is a finite set if X is a semistable curve. By Corollary 1.2 in [22],Footnote 1 it is enough to show that a general closed point of any LCRT component (or a separating line, after [22]) is not contained in \(\textsf {PBs}( X )\).
Let \({\mathbb P} ^{ 1 } \simeq E \hookrightarrow X\) be an LCRT component. By definition, there are mutually disjoint subcurves \(X _{ 1 }, \ldots , X _{ r } \hookrightarrow X\)
for some \(r \ge 2\) such that \(X = \left( \bigcup _{ i = 1 } ^{ r } X _{ i } \right) \cup E\) and \(| X _{ i } \cap E | = 1\) for \(i = 1, \ldots , r\) (hence the name comb). Then it follows for each i that \(\dim _{ {\textbf {k}}} H ^{ 1 } \left( X _{ i }, \mathcal{O}_{ X _{ i } } \right) > 0\). Indeed, otherwise \(X _{ i}\) is a tree of \({\mathbb P} ^{ 1 }\)s and hence the component of \(X _{ i }\) which corresponds to a leaf of the dual graph of \(X _{ i }\) would destabilize X. Hence we can apply Lemma 5.3 to each \(X _{ i }\) to obtain a non-trivial invertible sheaf \(L _{ i } \in {{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X _{ i } \right)\). Now let \(L \in {{\,\mathrm{{Pic}}\,}}^{ 0 } ( X )\) be such that \(L \vert _{ X _{ i } } \simeq L _{ i }\) for all \(i = 1, \ldots , r\) (see, say, the 1st paragraph of Example 0.7.2 in [5] for the existence of such L).
For a general closed point \(x \in E\) consider the following short exact sequence
from which we obtain the following exact sequence.
The last term is isomorphic to the dual of the following vector space by the Serre duality.
By Lemma 5.2, for each \(i = 1, \ldots , r\) we have \(H ^{ 0 } \left( X _{ i }, L ^{ - 1 } ( x ) \vert _{ X _{ i } } \right) = 0\). Finally, as \(L ^{ - 1 } ( x ) \vert _{ E } = \mathcal{O}_{ E } ( 1 )\) and \(r \ge 2\), we conclude that (5.6) vanishes.
Conversely, let X be a curve as in the assertion which is not semistable. Then there must be an irreducible component \(C \hookrightarrow X\) such that \(C \simeq {\mathbb P} ^{ 1 }\) and C meets the other irreducible components of X in exactly 1 point. Then Proposition 6.15 in [14] implies that there is a nontrivial semiorthogonal decomposition of \({D}^b_c( X)\), which implies that \({{\text{Perf }}}(X)\) admits a nontrivial semiorthogonal decomposition by Corollary 6.6 in [13]. \(\square\)
Remark 5.4
When the components \(X _{ 1 },\ldots , X _{ r }\) in the proof of Proposition 5.1 are all smooth, then we can prove Proposition 5.1 by means of the relative canonical base locus. This can be regarded as a sophisticated version of Lin’s argument.
Indeed, in this case we have a morphism as follows, which contracts the LCRT E in the proof of Proposition 5.1 to a point.
One can easily show that \(\text{Bs} |f| = \emptyset\) (recall that \(r \ge 2\)). Hence there is no f-linear semiorthogonal decomposition of \({{\text{Perf }}}(X)\) by Theorem 3.6. Now since \(\prod _{ i = 1 } ^{ r } {{\,\mathrm{{Pic}}\,}}^{ 0 } \left( X _{ i } \right)\) is an abelian variety, any semiorthogonal decomposition of \({{\text{Perf }}}(X)\) is f-linear by Theorem 1.4 in [21]. Hence the conclusion. It is interesting to ask if one can push the same strategy without assuming \(X _{ i }\) are smooth.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
López Martín, A.C., Sancho de Salas, F. Indecomposability of derived categories for arbitrary schemes. Collect. Math. (2023). https://doi.org/10.1007/s13348-023-00404-w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s13348-023-00404-w
Keywords
- Semiorthogonal decompositions
- Derived categories
- Indecomposable
- Perfect complexes
- Minimal model program
- Stable curves
- Kodaira degenerations