Square function estimates and the Kato problem for second order parabolic operators in \(\mathbb{R}^{n + 1}\) (Q261160): Difference between revisions

A major result in the harmonic analysis of partial differential operators is the solution of Kato's square root problem given in [\textit{P. Auscher} et al., Ann. Math. (2) 156, No. 2, 633--654 (2002; Zbl 1128.35316)]: \[ \|\nabla u\|_{L^{2}} \sim \|\sqrt{-\operatorname{div} A \nabla}u\|_{L^{2}} \quad \forall u \in W^{1,2}, \] where \(L=-\operatorname{div} A \nabla\) is a uniformly elliptic divergence form operator with bounded measurable complex coefficients. In recent years, this result, and the techniques invented to prove it, have been at the heart of a range of exciting developments in harmonic analysis and PDE. In particular, they have been successfully used to solve non-smooth elliptic boundary value problems; see e.g. [\textit{P. Auscher} et al., J. Funct. Anal. 255, No. 2, 374--448 (2008; Zbl 1157.35028)], and [\textit{P. Auscher} and \textit{A. Axelsson}, Invent. Math. 184, No. 1, 47--115 (2011; Zbl 1231.35059)]. The paper under review is an important first step in a program that aims to extend the theory emanating from the solution of Kato's square root problem to parabolic differential operators of the form \(\partial_{t}+L\). It establishes fundamental square function estimates of the form: \[ \int \limits _{0} ^{\infty} \int \limits _{0} ^{\infty} \int \limits _{\mathbb{R}^{n}} \|\lambda (\partial_{t}+L)(I+\lambda^{2}(\partial_{t}+L))^{-1}f\|_{2} ^{2} \frac{d\lambda dt dx}{\lambda} \lesssim \|\nabla f\|_{2} ^{2} + \| D_{t}^{\frac{1}{2}} f\|_{2} ^{2}, \] where \(\widehat{D_{t} ^{\frac{1}{2}} f}(\tau,\xi)= |\tau|^{\frac{1}{2}} \widehat{f}(\tau,\xi)\). In the elliptic situation, the Kato square root estimates follow directly from such square function estimates. In the parabolic case, one cannot derive results regarding the domain of \(\sqrt{\partial_{t}+L}\) so easily, and the author leaves that question for future research. Nevertheless, by proving the square function estimates, the author has successfully extended the most difficult part of the solution of the Kato square root problem to parabolic operators. Moreover, in applications to boundary value problems, the square function estimates themselves, along with the techniques developed to prove them, are often more useful than the square root estimates per se. It is therefore highly likely that the paper under review will provide a foundation for a new theory of non-smooth parabolic boundary value problems. The proof of the main result has the same structure as its elliptic counterpart. Some basic operator theoretic and harmonic analytic properties of the operators \((I+\lambda^{2}(\partial_{t}+L))^{-1}\) have to be established first (Section 2), but the heart of the proof is still the Carleson measure estimate: \[ \int \limits _{0} ^{\ell(Q)} \int \limits _{Q} |\lambda (I+\lambda^{2}(\partial_{t}+L))^{-1} \operatorname{div} A|^{2}\frac{d\lambda dt dx}{\lambda} \lesssim |Q|, \] for all parabolic cubes \(Q\). It is proven in Section 3, and requires a careful parabolic variation on the stopping time arguments developed for the elliptic problem. The paper is concluded by an interesting discussion of the next steps that should be taken in the development of this theory: connecting the square function estimates to an appropriate functional calculus, and considering time dependent coefficients.