Heat content with singular initial temperature and singular specific heat (Q2512906)

The heat content of a compact submanifold \(\Omega\) in a closed manifold \(M\) is provided by \(\beta_\Omega(\phi, \rho,D_M)(t)= \int_{\Omega\times\Omega}K_M(x,y,t)\phi(x)\rho(y)dxdy\) for \(t>0\) such that \(K_M\) is the kernel of \(\exp(-tD_M)\), the fundamental solution of the heat equation for \(D_M\), the Laplace-Beltrami operator on \(M\), \(\phi\) and \(\rho\) are the initial and the specific temperatures, respectively. The first result states that for \(r^a\phi\) and \(r^b\rho\) smooth functions in a neighborhood of the boundary \(\partial\Omega\), such that \(r\) stands for the geodesic distance and \((a,b)\in\mathcal{O}:=\{(z_1,z_2)\in \mathbb C^2\text{ such that } (\Re z_1,\Re z_2)\in (-\infty,1)^2,\;z_1+z_2\neq1-2k,\text{ for } k\in \mathbb N\}\), then (up to an error) the heat content has a finite expansion whenever \(t\) is closer to zero, besides the coefficients satisfy meaningful conditions (Theorem 1.1). The proof is essentially based on the following result that for \(N\geq-\Re(a+b)+3\), the remainder term \(R_N(a,b)\) -- associated to the expansion of the heat content \(h_{a,b}^{[0,1]}(t):=h_{[0,1]}(x^{-a},x^{-b},-\frac{\partial^2}{\partial x^2})(t)\) -- is uniformly bounded whenever the minus real part of \(a+b\) is either greater or less than seven (Theorem 3.1), furthermore the explicit coefficients are holomorphic functions in \(\mathcal{O}\) (Lemma 3.6). For the precise control of \(R_N(a,b)\), see Lemma 3.7. Whence, by using the regularization process of \(\beta_\Omega\) on the real line (Section 3.4), on the half line (Section 3.5) and on an interval (Section 3.6), respectively, and calculus of pseudo-differential operators, they end the proof of Theorem 1.1. Concerning the second result, the authors consider two cut-off functions \(\psi_1,\psi_2\) and \(a<1\), then for \(t\) closer to zero, they unfold \(\beta_{[0,1]}(x^{-a}\psi_1,x^{a-1}\psi_2, D_{\mathbb R})(t)\) in three terms: The first one is given explicitly, the second one is minus \(\frac{\log(t)}{2}\), and the last one corresponds to the \(O(t^{1/2}\log(t))\), see Theorem 1.2, and its general version is given in Theorem 3.2 when \(a\) and \(b\) are real numbers such that \(a+b=1-2k\) for \(k\in \mathbb N\). The proof is based on using several lemmas, essentially on the properties of the heat content when \(\phi\) and \(\rho\) satisfy smoothness conditions and where the ambient space is the unit circle (Lemma 3.4).