A canonical connection on bundles on Riemann surfaces and Quillen connection on the theta bundle (Q2048626)

Let \(X\) be a compact Riemann surface of genus \(g\geq 2\) with canonical line bundle \(K_X\), \(r\) a natural number, and \(\mathcal{M}\) the moduli space of stable vector bundles on \(X\) of rank \(r\) and degree \(0\). Given a theta characteristic \(K_{X}^{1/2}\), \(D_{\Theta}:=\{E\in\mathcal{M}: H^{0}(X,E\otimes K_{X}^{1/2})\neq 0\}\) denotes the theta divisor, and \(\Theta\) denotes the associated theta line bundle over \(\mathcal{M}\). The authors consider two fibered spaces over \(\mathcal{M}\). On the one hand, the moduli space \(\mathcal{C}\) of holomorphic connections, \(E\rightarrow E\otimes K_X\), with \(E\) a stable vector bundle of rank \(r\). The forgetful map induces a holomorphic map, \(\mathcal{C}\rightarrow \mathcal{M}\). On the other hand, the fiber bundle Conn\((\Theta)\rightarrow \mathcal{M}\) given by the sheaf of holomorphic connections on \(\Theta\). Sections of Conn\((\Theta)\rightarrow \mathcal{M}\) over an open subset \(U\subset \mathcal{M}\) are in one-to-one correspondence with holomorphic connections on \(\Theta|_{U}\). Both \(\mathcal{C}\) and Conn\((\Theta)\) are holomorphic \(T^{*}\mathcal{M}\)-torsors and carry holomorphic symplectic forms, \(\Phi_1\) and \(\Phi_2\), respectively. Recently, the authors have constructed a \(\mathcal{C}^{\infty}\) isomorphism of torsors \(F:\mathcal{C}\simeq \mathrm{Conn}(\Theta)\) with the property that \(F^{*}\Phi_2 \simeq \Phi_1\) up to multiplication by \(2r\); that is, it preserves the symplectic structures. Further, they have proved that \(F\) is, in fact, holomorphic. The main result of this article is that the restriction of \(F\) to \(\mathcal{M}_0 :=\mathcal{M}\setminus D_{\Theta}\) can be constructed by purely algebraic methods. More precisely, the authors construct holomorphic sections, \(\phi\) and \(\tau\), of the torsors \(\mathcal{C}|_{\mathcal{M}_0}\rightarrow \mathcal{M}_0\) and \(\mathrm{Conn}(\Theta)|_{\mathcal{M}_0}\rightarrow \mathcal{M}_0\) respectively, and prove that the holomorphic map \(G\) defined by \(G(\delta^0 (\phi(E),\nu))=\eta^0(\tau(E),2r\cdot \nu)\) is a holomorphic isomorphism between \(\mathcal{C}|_{\mathcal{M}_0}\) and \(\mathrm{Conn}(\Theta)|_{\mathcal{M}_0}\), and it coincides with \(F|_{\mathcal{M}_0}\). Here, \(\delta^0\) and \(\eta^0\) are the actions of \(T^{*}\mathcal{M}_{0}\) on \(\mathcal{C}|_{\mathcal{M}_0}\) and \(\mathrm{Conn}(\Theta)|_{\mathcal{M}_0}\) respectively. As a consequence of the main result is that the holomorphic isomorphism \(F\), which preserves the symplectic structures, depends holomorphically on the base Riemann surface. This fact lets the authors to extended the holomorphic isomorphism \(F\) to a relative context.