Non-commutative analytic subgroup theorem (Q2235730)

Let \(G\) be a connected algebraic group defined over the field \({\overline{\mathbb{Q}}}\) of algebraic numbers. Let \({\mathrm{exp}}_G:{\mathrm{Lie}}(G)\to G\) be the exponential map and \(u\in{\mathrm{Lie}}(G)({\mathbb{C}})\) such that \({\mathrm{exp}}_G(u)\in G({\overline{\mathbb{Q}}})\). Let \(\Theta\) be the smallest \({\overline{\mathbb{Q}}}\)-vector subspace of \({\mathrm{Lie}}(G)_{\overline{\mathbb{Q}}}\) such that \(\Theta\otimes{\mathbb{C}}\) contains \(u\). Then \(\Theta\) is the Lie algebra of a connected algebraic subgroup \(H_u\) of \(G_{\overline{\mathbb{Q}}}\) which is commutative. This result was proved by \textit{G. Wüstholz} [Ann. Math. (2) 129, No. 3, 501--517 (1989; Zbl 0675.10025)] under the assumption that \(G\) is commutative. The general case follows from the commutative case: the one parameter subgroup \(\Phi_u={\mathrm{exp}}_G({\mathbb{C}}\cdot u)\) of \(G({\mathbb{C}})\) is commutative, its Zariski closure is commutative and is defined over \({\overline{\mathbb{Q}}}\). The author gives the following application of his result : if \(G\) is an affine algebraic group over \({\overline{\mathbb{Q}}}\) and \(z\) an element of \({\mathrm{Lie}}(G)_{\overline{\mathbb{Q}}}\) such that \( {\mathrm{exp}}_G(z)\in G({\overline{\mathbb{Q}}})\), then \({\mathrm{exp}}_G(z)\) is contained in a subgroup isomorphic to the additive group \({\mathbb G}_a\); further, if \(G\) is an affine algebraic group, then \(z\) is nilpotent in \({\mathrm{Lie}}(G)({\mathbb{C}})\).