The geometry of arithmetic noncommutative projective lines (Q740250): Difference between revisions

Let \(k\) be a perfect field, and let \(K/k\) be a finite extension. This article studies the geometry of noncommutative spaces (Grothendieck categories) of the form \(\text{Proj }\mathbb S_K(V)\) where \(V\) is a \(k\)-central \(K - K\)- bimodule which is two-dimensional as a vector space via the right and left actions of \(K\) on \(V\). \(\text{Proj }\mathbb S_K(V)=\mathbb P_K(V)\) denotes the quotient of the category of graded right \(\mathbb S_K(V)\)-modules modulo the full subcategory of direct limits of right bounded modules. The space \(\mathbb P_K(V)\) is called the \textit{arithmetic} projective line, and one of the main issues of this article is to show that its geometry has close connection to data associated with \(K/k\). The noncommutative projective lines serves as basic examples in the field of noncommutative arithmetic geometry. Artin's conjecture states that the division ring of fractions of a noncommutative surface, not finite over its center, is the function field of a noncommutative \(\mathbb P^1\)-bundle over a smooth commutative curve. As a noncommutative projective line is a noncommutative \(\mathbb P^1\)-bundle over the point \(\text{Spec }K\), this article provides a model for the geometry of noncommutative \(\mathbb P^1\)-bundles over smooth commutative curves. The article gives isomorphism invariants of noncommutative projective lines, and questions if these are also birational. If Artins conjecture is true, then noncommutative surfaces, infinite over their center, will have the same birational invariants as noncommutative \(\mathbb P^1\)-bundles over smooth commutative curves, and this indicates which invariants that are the interesting ones. The author justifies the name ``noncommutative projective line''. It is proved that these spaces are noetherian, Ext-finite, homological dimension one categories, also having a Serre functor. The first main result states that noncommutative projective lines are integral (in a defined sense), that the line bundles over a noncommutative projective line are indexed by \(\mathbb Z\), and that every vector bundle over a noncommutative projective line is a direct sum of line bundles. Using the classification of vector bundles, the author classify noncommutative projective lines up to \(k\)-linear equivalence. The statement is that there is a \(k\)-linear equivalence \(\mathbb P_K(V)\rightarrow\mathbb P_K(W)\) if and only if there exists a \(\sigma\in\mathrm{Gal}(K/k)\) such that either \(V\simeq K_\delta\otimes_K K_{\epsilon}\) or \(V\simeq K_\delta\otimes_K W^\ast\otimes_K K_{\epsilon}\). Here \(K_\sigma\) for \(\sigma\in\mathrm{Gal}(K/k)\) denotes the two-sided vector space whose left \(K\)-action is ordinary multiplication in \(K\), and whose right action is defined as \(x\cdot a=x\sigma(a)\). It is known that there is a one-to-one correspondence between the finite orbits \(\text{Emb}(K)\) under the action of \(G=\mathrm{Gal}(\overline K/K)\) and the isomorphism classes of simple left finite dimensional two-sided vector spaces. This allows the author to sharpen the classification of noncommutative projective lines: For \(\mathrm{char} k\neq 2\) and \(V_i\) a rank 2 two-sided vector space for \(i=1,2\), there is a \(k\)-linear equivalence \(\mathbb P_K(V_1)\rightarrow\mathbb P_K(V_2)\) if and only if (1) there exists \(\sigma_i\in\mathrm{Gal}(K/k)\) such that \(V_i\simeq K_{\sigma_i}\oplus K_{\sigma_i}\). In this case, \(\mathbb P(V_i)\) is equivalent to the commutative projective line over \(K\), (2) there exists \(\sigma_i, \tau_i\in\mathrm{Gal}(K/k),\) with \(\sigma_i\neq\tau_i\), \(V_i\simeq K_{\sigma_i}\oplus K_{\tau_i}\) and under the action of \(\mathrm{Gal}(K/k)^2\) on itself defined by \((\sigma,\tau)\cdot(\delta,\epsilon)=(\delta^{-1}\sigma\epsilon,\delta^{-1}\tau\epsilon)\) the orbit of \((\sigma_1,\tau_1)\) contains an element of the set \(\{(\sigma_2,\tau_2),(\sigma_2^{-1},\tau_2^{-1}),(\tau_2,\sigma_2),(\tau_2^{-1},\sigma_2^{-1})\}\), (3) \(V_i\simeq V(\lambda_i)\), and under the action of \(\mathrm{Gal}(K/k)^2\) on \(\Lambda(K)\) defined by \(\lambda^G\cdot(\delta,\epsilon)=(\overline{\delta}^{-1}\lambda\epsilon)^G\), the orbit of \(\lambda_1^G\) contains either \(\lambda_2^G\) or \(\mu_2^G\). \(\overline{\delta}\) denotes an extension of \(\delta\) to \(\overline{K}\). The author gives the definition of the three types of canonical equivalences between noncommutative projective lines. Then the classification of \(k\)-linear equivalences \(\mathbb P_K(V)\rightarrow\mathbb P_K(W)\) upto isomorphism can be given for \(\mathrm{char} k \neq 2\): Let \(F:\mathbb P_K(V)\rightarrow\mathbb P_K(W)\) be a \(k\)-linear equivalence. Then there exists \(\delta_i\in\mathrm{Gal}(K/k)\), an isomorphism \(\phi:K_{\delta}^{-1}\otimes_K V\otimes_K K_{\epsilon}\rightarrow W^{i^\ast}\) inducing an equivalence \(\Phi:\mathbb P_K(K_{\delta^{-1}}\otimes_K V\otimes_K K_{\epsilon})\rightarrow\mathbb P_K(W^{i\ast})\) and an integer \(i\) such that \(F[-i]\circ\Phi\circ T_{\delta,\epsilon}\). Also, \(\delta,\epsilon, i\) are unique up to natural equivalence, while \(\Phi\) is naturally equivalent to \(\Phi^\prime\) if and only if there exist nonzero \(a,b,\in K\) such that \(\phi^\prime\phi^{-1}(w)=a\cdot w\cdot b\) for all \(w\in W^{i\ast}\) The last result is a computation of the automorphism group of a noncommutative projective line. The automorphism group of \(\mathbb P_K(V)\) is the set of \(k\)-linear shift-free equivalences \(\mathbb P_K(V)\rightarrow\mathbb P_K(V)\) modulo natural equivalence. Also, the author discuss some of the other notions of noncommutative projective line given in the literature. For some of these notions, the results in this article proves that a noncommutative projective line is a commutative curve of genus \(0\), and necessary and sufficient conditions for noncommutative curves of genus zero to be noncommutative projective lines are given. The theory presented in this article, just don't fit into other notions. All necessary definitions and explicit computations are given in the text. The article is a long and thorough explanation of the geometry of noncommutative projective lines, and the arithmetic of these lines are explicit.