Lie 2-algebras of vector fields (Q2226642)

This paper is concerned with the structure of the category of vector fields on a geometric stack. Ultimately, it is shown that this category is equivalent to the groupoid of multiplicative vector fields on any Lie groupoid presenting the stack, which in turn is proven to be a strict Lie 2-algebra. Recall that a stack \(\mathcal{A}\) is geometric, roughly, if it is presented by a groupoid \(G=G_1\rightrightarrows G_0\), i.e., if \(\mathcal{A}\) is isomorphic to \(\mathbb{B}G\), the stack of principal \(G\)-bundles. Two groupoids whose stacks of principal bundles are isomorphic are called Morita equivalent. In [\textit{R. Hepworth}, Theory Appl. Categ. 22, 542--587 (2009; Zbl 1206.37009)], Hepworth proved that a stack \(\mathcal{A}\) over the site of manifolds carries a category of vector fields \(\mathrm{Vect}(\mathcal{A})\) that is, in fact, a groupoid. In the same article, Hepworth proved that vector fields \(\mathbb{X}(G)\) on a Lie groupoid \(G\) also form a category. A natural question to ask is whether these categories of vector fields are related. The authors prove that this is indeed the case, that the groupoid \(\mathrm{Vect}(\mathcal{A})\) of vector fields on a geometric stack \(\mathcal{A}\) is equivalent to the groupoid \(\mathbb{X}(G)\) of vector fields on any groupoid presenting it (Theorem 6.1). The expected consequence of this result is that just in as much as the space of vector fields on a manifold is not only a set but a Lie algebra, the category of vector fields on a geometric stack is a (strict) Lie 2-algebra, i.e., a category internal to the category of Lie algebras (Theorem 3.4). To prove this result, the authors use the equivalence between strict Lie 2-algebras and crossed modules. In so, the Lie 2-algebra structure on \(\mathbb{X}(G)\) is built out of the Lie algebra of global sections of the Lie algebroid of \(G\) and Mackenzie and Xu's Lie algebra of multiplicative vector fields. The bulk of the article is concerned with how to use Theorem 3.4 to get a Lie 2-algebra structure on \(\mathrm{Vect}(\mathbb{B}G)\); in other words, given two Morita equivalent Lie groupoids \(G\) and \(H\), to establish that their Lie 2-algebras of vector fields \(\mathbb{X}(G)\) and \(\mathbb{X}(H)\) are too Morita equivalent (Theorem 4.1). Quoting Remark 1.2, ``At first pass this might seem surprising: neither multiplicative vector fields nor sections of Lie algebroids are well-behaved under Morita equivalence of Lie groupoids. Theorem 4.1 shows that combining this pair of Lie algebras into a Lie 2-algebra gives us an object that is preserved by Morita equivalence''. The authors prove Theorem 4.1 by first establishing that \(\mathbb{X}\) is a contravariant functor from the bicategory \(\mathcal{E}mb\) of Lie groupoids, essentially surjective open embeddings and natural transformations to the strict 2-category \(\mathrm{Lie2Alg}_{\mathrm{strict}}\) of Lie 2-algebras (Proposition 4.5), and then proving that this functor maps every morphism in \(\mathcal{E}mb\) to a fully faithful essentially surjective morphism of Lie 2-algebras (Lemmas 4.10 and 4.11). Thus, the authors extend the assignment \(G\mapsto\mathbb{X}(G)\) to a covariant functor \(\mathbb{X}:\mathcal{E}mb\to\mathrm{Lie2Alg}\) to the localization of \(\mathrm{Lie2Alg}_{\mathrm{strict}}\) at the class of essential equivalences (Definition 4.12). It turns out the localization of the bicategory \(\mathcal{E}mb\) at its full space of morphism coincides with \(\mathrm{Bi}_{\mathrm{iso}}\) hence implying the result. The proof of Lemma 4.10 as well as Theorem 6.1 hang on the technical Theorem 4.9, whose proof is the content of Section 7. This theorem establish an equivalence of categories between \(\mathbb{X}(G)\) and the category \(\mathbb{X}_{\mathrm{gen}}(G)\) of generalized vector fields (i.e. bibundles from the groupoid to its tangent that are sections in an appropriate sense), induced by the inclusion functor \(\iota_G:\mathbb{X}(G)\to \mathbb{X}_{\mathrm{gen}}(G)\). It should be noted that the statement Theorem 6.1 coincides with that of Theorem 4.15 in [\textit{R. Hepworth}, Theory Appl. Categ. 22, 542--587 (2009; Zbl 1206.37009)]; however, as the authors point out, during the proof of the latter the groupoid seems to be implicitly assumed to be proper. This signals the relevance of the technique involved in the proof of Theorem 4.9 as it ultimately implies that Theorem 6.1 truly holds for any Lie groupoid. The present article is, in a sense, complementary to [\textit{C. Ortiz} and \textit{J. Waldron}, J. Geom. Phys. 145, Article ID 103474, 34 p. (2019; Zbl 1469.58012)] which studies the categories of multiplicative sections of so-called \(\mathcal{LA}\)-groupoids (i.e. groupoid objects in the category of Lie algebroids), proving in particular that they also can be endowed with the structure of a strict Lie 2-algebra. In the authors own words ``to every stack Ortiz and Waldron assigned an ordinary Lie algebra and showed that in the case of proper geometric stacks the set underlying this Lie algebra is in bijective correspondence with isomorphism classes of vector fields in Hepworth's definition''. Indeed, this article offers categorified versions of Ortiz and Waldron's results and goes beyond proper Lie groupoids. An innocuous typo on the words following Theorem 4.9: ``[Hepworth 2009, Theorem 15]'' should read ``[Hepworth 2009, Theorem 4.15]''.