$\mathbb{P}^n$-functors

TL;DR

This paper introduces a new framework for P^n-functors, unifying and extending previous concepts, and provides criteria and examples demonstrating their properties and applications in derived categories.

Contribution

It develops a comprehensive theory of non-split P^n-functors, including criteria for their identification and connections to autoequivalences and geometric transformations.

Findings

Construction of P-twists as autoequivalences

A practical criterion for identifying P^n-functors

Examples including spherical functors and geometric monodromies

Abstract

We propose a new theory of (non-split) P^n-functors. These are F: A -> B for which the adjunction monad RF is a repeated extension of Id_A by powers of an autoequivalence H and three conditions are satisfied: the monad condition, the adjoints condition, and the highest degree term condition. This unifies and extends the two earlier notions of spherical functors and split P^n-functors. We construct the P-twist of such F and prove it to be an autoequivalence. We then give a criterion for F to be a P^n-functor which is stronger than the definition but much easier to check in practice. It involves only two conditions: the strong monad condition and the weak adjoints condition. For split P^n-functors, we prove Segal's conjecture on their relation to spherical functors. Finally, we give four examples of non-split P^n-functors: spherical functors, extensions by zero, cyclic covers, and family P-twists. For the latter, we show the P-twist to be the derived monodromy of associated Mukai flop, the so-called `flop-flop = twist' formula.