Vanishing of \(H^1\) for Dedekind rings and applications to loop algebras (Q556912)

Let \(k\) be an algebraically closed field of characteristic \(0\) and let \(\{\zeta_n\}_{n\geq 1}\) be compatible primitive \(n\)-roots of unity in \(k^\times\). If \(A\) is any \(k\)-algebra , and if \(\sigma\) is an automorphism of \(A\) of finite order \(m=\sigma(m)\), then \(A\) has an eigenspace decomposition \(A=\sum_{i=0}^{m-1}A_{\bar i}\) where \(\bar i\) is the class of \(i\) in \(\mathbb{Z}/m\) and where \(A_{\bar i}=\{a\in A\mid \zeta(a)=\zeta^i_m a\}.\) The {loop algebra} \(L(A,\sigma)\) of the pair \((L,\sigma)\) is defined as \[ L(A,\sigma)=\bigoplus_{i\in\mathbb{Z}}A_{\bar i}\otimes t^{i/m}\subset A\otimes_k S_m, \] with \(S_m=k[t^{1/m},t^{-1/m}]\). The author's (at least in this article) main interest in these loop algebras, stems from the fact that every affine Kac-Moody Lie algebra is isomorphic to the loop algebra of some finite dimensional simple \(k\) Lie algebra. The contents of this article is to study the {Loop Algebra Principle}: Let \(A\) be a \(k\)-algebra. There exists a normal subgroup \(\text{Int} A\) of \(\text{Aut} A\) with the property that for any two finite order automorphisms \(\sigma\) and \(\tau\) of \(A\), \(L(A,\sigma)\cong L(A,\tau)\Leftrightarrow\bar\sigma=\bar\tau^{\pm 1}\) where \(\bar{}\) denotes the class in \(\text{Aut} A/\text{Int} A\). The loop algebra principle is known in a series of cases listed in the article. It is remarked that the classification of loop algebras depends solely on the structure of \(\text{Aut} A\), and not on the nature of \(A\) itself. The first main theorem in the article proves that the loop algebra principle holds in the Galois cohomological case. The proof of this theorem shows that the vanishing of \(H^1\) for for certain reductive group schemes over \(k[t,t^{-1}]\) is essential. The rest of the article proves this vanishing for a Dedekind scheme whose generic fiber is the spectrum of a field \(K\) of dimension \(1\). This proof also includes the rest of the proof of the first main theorem.