$d$-abelian quotients of $(d+2)$-angulated categories

Abstract

Let ${\mathscr T}$ be a triangulated category. If $T$ is a cluster tilting object and $I = [ \operatorname{add} T ]$ is the ideal of morphisms factoring through an object of $\operatorname{add} T$, then the quotient category ${\mathscr T} / I$ is abelian. This is an important result of cluster theory, due to Keller-Reiten and König-Zhu. More general conditions which imply that ${\mathscr T} / I$ is abelian were determined by Grimeland and the first author. Now let ${\mathscr T}$ be a suitable $( d+2 )$-angulated category for an integer $d \geqslant 1$. If $T$ is a cluster tilting object in the sense of Oppermann-Thomas and $I = [ \operatorname{add} T ]$ is the ideal of morphisms factoring through an object of $\operatorname{add} T$, then we show that ${\mathscr T} / I$ is $d$-abelian. The notions of $( d+2 )$-angulated and $d$-abelian categories are due to Geiss-Keller-Oppermann and Jasso. They are higher homological generalisations of triangulated and abelian categories, which are recovered in the special case $d = 1$. We actually show that if $Γ= \operatorname{End}_{ \mathscr T }T$ is the endomorphism algebra of $T$, then ${\mathscr T} / I$ is equivalent to a $d$-cluster tilting subcategory of $\operatorname{mod} Γ$ in the sense of Iyama; this implies that ${\mathscr T} / I$ is $d$-abelian. Moreover, we show that $Γ$ is a $d$-Gorenstein algebra. More general conditions which imply that ${\mathscr T} / I$ is $d$-abelian will also be determined, generalising the triangulated results of Grimeland and the first author.