A Gaussian estimate for the heat kernel on differential forms and application to the Riesz transform (Q2439284)

Consider a complete Riemannian \(n\)-manifold \((M,g)\) which satisfies a Sobolev inequality of dimension \(n\) and on which the volume growth is comparable to the one of \(\mathbb{R}^n\) for big balls. The author proves a Gaussian estimate on the heat kernel of the Hodge Laplacian acting on \(1\)-forms in the hypotheses that there is no non-zero \(L^2\) harmonic \(1\)-form and the Ricci tensor is in \(L^{{{n}\over{2}}-\varepsilon }\cap L^\infty ; \;\varepsilon >0\). After presenting some concepts such as the volume doubling property, Gaussian upper estimate for the heat kernel, the author presents the statement of the theorem of Coulhon-Zhang. A manifold \(M\) satisfies a Sobolev inequality in dimension \(n\) if there is a constant \(C\) such that \( \|f\|_{{2n}\over{n-2}}\leq C\|\nabla f\|_2\), \(\;\forall f\in C^\infty_0(M)\). Some results are proved. Theorem 3. Let \(M\) be a complete, non-compact Riemannian manifold that satisfies a Sobolev inequality of dimension \(n\) and assume that the volume of big balls is Euclidean of dimension \(n\): \( V(x,R)\simeq R^n, \forall x\in M, \;\forall R\geq 1\). Let \(E\to M\) be a Riemannian vector bundle, endowed with a compatible connection \(\nabla\), and let \(L\) be a generalized Schrödinger operator acting on sections of \(E\): \(\;L=\nabla^*\nabla + \mathcal{R}\), where \(\mathcal{R}\) is a field of symmetric endomorphisms. Assume that \(L\) is non-negative, that \(\mathcal{R}_{-}\) lies in \(L^{{n}\over{2}}-\varepsilon \cap L^\infty\) for some \(\varepsilon >0\) and that \(Ker_{L^2}(L)=\{0\}\). Then the Gaussian estimate holds for \(e^{-tL}\): if \(K_{\exp (-tL)}(x,y)\) denotes its kernel , for every \(\delta >0\) there is a constant \(C\) such that for every \((x,y)\in M\times M\) and \(t>0\), \[ \|K_{\exp (-tL)}(x,y)\|\leq (C/V(x,\sqrt{t}))\exp(-d^2(x,y)/(4+\delta)t). \] Next the author finds a boundedness result for the Riesz transform in the case where \(L\) is the Hodge-Laplacian acting on \(1\)-forms. In the second part of the paper the author studies the behavior of the Riesz transform when one removes the hypothesis \( \mathcal{H}^1(M)=0\). In fact he shows that the Riesz transform is bounded on \(L^p\) for all \(1<p<n\). At the end, the author studies the reduced cohomology \(H^1_p(M)\) of \(M\).