Proposition 9
Tag 0ADB, permanent. Cite it as [Bourbaki, Tag 0ADB].
Lie Groups and Lie Algebras, chapter III, § 4. Passage from Lie algebras to Lie groups
Let be a finite-dimensional Lie group and a subgroup of .
(i) There exists on one and only one analytic manifold structure with the following property: for all between 1 and , for every manifold of class and for every mapping of into , is of class as a mapping of into if and only if is of class as a mapping of into .
(ii) With this structure, is a Lie group, the canonical injection of into is an immersion and is the Lie subalgebra tangent at to .
In (i) the uniqueness is obvious. We prove the existence. Let be the Lie algebra tangent at to . Let be a Lie subgroup germ of with Lie algebra . By replacing by an open subgroup germ of , it can be assumed that (Lemma 4 (iv)). For all , is a Lie subgroup germ of with Lie algebra . Hence is open in (no. 2, Theorem 3) and the mapping is an isomorphism of onto . Using Proposition 18 of § 1, no. 9, there exist an open Lie subgroup germ of and a Lie group structure on with the following properties: is open in and the manifold structures on and induce the same structure on . It then follows that the canonical injection of into is an immersion and that . Moreover, let and be as in (i). If is of class , is of class . Suppose that is of class ; then we prove that is of class . By translation, it suffices to consider the case where there exists such that and to prove that is of class in an open neighbourhood of . Now, by Lemma 4 (iii), for sufficiently close to , whence our assertion. Thus we have proved (i) and (ii) has been obtained on the way.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.