Conformal structures in noncommutative geometry (Q2471146): Difference between revisions

Let \((M, g)\) be a compact Riemannian spin manifold. Then it is reconstructed from the spectral triple \(({\mathcal A},{\mathcal H},D)\), where \({\mathcal A}= C^\infty(M)\), \({\mathcal H}= L^2(M,\Sigma M)\), \(\Sigma M\) is the spinor bundle, and \(D\) is the Dirac operator. \(M\) is reconstructed from \({\mathcal A}\), while \(g\) is reconstructed from the pair \(({\mathcal H}, D)\), together with the action of \({\mathcal A}\) on \({\mathcal H}\) (this is reviewed in \S2 of this paper). In this paper, the conformal structure of \(M\) is reconstructed from the pair \(({\mathcal H},\text{sign}(D))\). Here \[ \text{sign}(D)(e_n)= \begin{cases} e_n,& \lambda_n> 0,\\ 0, &\lambda_n= 0,\\ -e_n, &\lambda_n< 0,\end{cases} \] where \(De_n= \lambda_n e_n\). The pair \(({\mathcal H},\text{sign}(D))\) together with the action of \({\mathcal A}\) is called canonical Fredholm module of \((M, g)\). If \(g'= e^{2v}g\), then Riemannian metrics \(g\) and \(g'\) on \(M\) are called conformally equivalent. Then, the result is stated that \(g\) and \(g'\) are conformally equivalent if and only if canonical Fredholm modules \(({\mathcal H},\text{sign}(D))\) and \(({\mathcal H},\text{sign}(D'))\) are weakly unitary equivalent (Th. 3.1). To show the Theorem, by using precise symbol calculus, \[ \sigma_Q(\xi)= \Psi(x)\cdot \sigma_P(\xi)\cdot\Phi(x),\quad x\in M,\;\xi\in S^*_x M, \] where \(P\) is a classical pseudo-differential operator of order \(0\) acting on sections in \(E\), \(\Phi\in L^\infty(M,\Hom(E',E))\), \(\Psi\in L^\infty(M, \Hom(E, E'))\) and \(Q=\Psi\circ P\circ \Phi\) is shown (Lemma 4.3). Since the principal symbol of the Dirac operator is given by Clifford multiplication with respect to the metric \(g\); \(\sigma_D(\xi)= ic_g(\xi)\), the principal symbol of \(\text{sign}(D)\) is given by \(\sigma_{\text{sign}(D)}(\xi)= {ic_g(\xi)\over \|\xi\|_g}\). By Lemma 4.3, we have \[ \sigma_{U\circ\operatorname{sign}(D)\circ U^{-1}}(\xi)= \Psi(x)\cdot \sigma_{\text{sign}(D)}(\xi)\circ \Psi^{-1}(x). \] Therefore we obtain \[ {-2g'(\xi,\eta)\over \|\xi\|_{g'}\|\eta\|_{g'}}= {-2g(\xi,\eta)\over \|\xi\|_g\|\eta\|_g}. \] Hence, we have the only if part of Theorem 3.1. The if part follows from the formula \[ \sigma_{\text{sign}(D)}(\xi)= {\sigma_D(\xi)\over |\sigma_D(\xi)|}. \]