Leibniz-type rules for bilinear Fourier multiplier operators with Besov regularity (Q2071836)

The main result of the paper is a Leibniz-type estimate for the bilinear Fourier multiplier operator \[ T_m(f,g)(x) = \int_{\mathbb{R}^{2n}} m(\xi,\eta)e^{i(\xi+\eta)\cdot x}\widehat{f}(\xi)\widehat{g}(\eta) d\xi d\eta \] under Besov regularity assumptions on the symbol \(m\). To formulate this more precisely let \(\Psi\) be a Schwartz function on \(\mathbb{R}^{2n}\), supported in an annulus centered at the origin of inner radius \(\frac{1}{2}\) and outer radius \(2\) such that \(\sum\limits_{k\in\mathbb{Z}}\Psi(\xi/2^k,\eta/2^k) = 1\) for all \((\xi,\eta)\in\mathbb{R}^{2n}\setminus\lbrace (0,0)\rbrace\). For all \(k\in\mathbb{Z}\) and \(0 \leq \nu < 2n\) define \[ m_k^{\nu}(\xi,\eta) 2^{k\nu}m(2^k \xi, 2^k \eta)\Psi(\xi,\eta) .\] Under the assumption that \(m\) satisfies \[ \sup\limits_{k\in\mathbb{Z}} \| m_{k}^{\nu}\|_{B_{2,1}^{(n/2,n/2)}} < \infty \] for some \(0\leq \nu < 2n\), where \(B_{2,1}^{(n/2,n/2)}\) is a product Besov space, the authors show for \(s>0\) and \(f\), \(g\) Schwartz functions on \(\mathbb{R}^n\) \[ \| D^{s}T_m(f,g) \|_{L^p} \lesssim \sup\limits_{k\in\mathbb{Z}}\| m_{k}^{\nu} \|_{B_{2,1}^{(n/2,n/2)}}\left( \| D^{s-\nu}f \|_{L^{p_1}}\| g \|_{L^{p_2}} + \| f \|_{L^{p_1}}\| D^{s\nu} g \|_{L^{p_2}} \right) \] whenever \(1/p = 1/p_1 + 1/p_2\) with \(1< p < \infty\) and \(2 < p_1\), \(p_2 < \infty\), where \(D^s\) encodes derivatives in an appropriate way. This extends and can be viewed as a limiting case of an analogous result of \textit{J. Hart} et al. [Trans. Am. Math. Soc. 370, No. 12, 8581--8612 (2018; Zbl 1409.42009)] that assumes Sobolev bounds on \(m\). Analogous mixed Lebesgue space results are also presented.