Estimates for differential operators of vecor analysis involving \(L^{1}\)-norm (Q845291)

In this nice paper, new Hardy and Sobolev type inequalities involving \(L^1\)-norms of scalar and vector-valued functions in \(\mathbb{R}^n\) are obtained. The work is related to some open problems stated in the recent paper by \textit{J. Bourgain} and \textit{H. Brezis} [J. Eur. Math. Soc. 9, 227--315 (2007; Zbl 1176.35061)]. Let \(\mathbf{f}\) be an \(n\)-dimensional vector-valued function in \(L^1({\mathbb R}^n)\) subject to \(\text{div}\mathbf{f}=0.\) Also, let \(\mathbf{u}\) denote the solution of \(-\Delta\mathbf{u}=\mathbf{f}\) in \(\mathbb{R}^n\) represented in the form \[ \mathbf{u}(x)=\int_{{\mathbb R}^n}\Gamma(x-y)\mathbf{f}(y)\,dy \quad (x\in{\mathbb R}^n), \] where \(\Gamma\) is the fundamental solution of \(-\Delta.\) In this paper, the author proves that there exists a positive constant \(c\) depending on \(n\) and \(q\in[1,n/(n-1))\) such that \[ \left(\int_{{\mathbb R}^n}|D\mathbf{u}(x)|^q |x|^{n(q-1)-q}\,dx\right)^{1/q}\leq c\int_{{\mathbb R}^n}|\mathbf{f}(x)|\,dx, \] where \(D\mathbf{u}\) is the Jacobi matrix \((\partial u_i/\partial x_j)_{i,j=1}^n\). This inequality is included in \textit{Open Problem 1} of Bourgain and Brezis's paper mentioned as above. Furthermore, certain inequalities for vector-valued functions involving Hilbert-Sobolev spaces \(\mathcal{H}^{-s}(\mathbb{R}^n)\) of negative order are also obtained in this paper. For example, Theorem 3(iii) of the paper states that the estimate \[ \left|\|\mathbf{g}\|^2_{\mathcal{H}^{-n/2}({\mathbb R}^n)} -n\|\text{div}\mathbf{g}\| ^2_{\mathcal{H}^{-1-n/2}({\mathbb R}^n)}\right|\leq\frac{(2\sqrt{\pi})^{-n}} {\Gamma(n/2)}\|\mathbf{g}\|^2_{L^1({\mathbb R}^n)} \eqno(\ast) \] holds for all \(\mathbf{g}\in L^1({\mathbb R}^n)\) with \(\text{div} \mathbf{g}\in \mathcal{H}^{-1-n/2}({\mathbb R}^n)\). An assertion dual to (\(\ast\)) replies in affirmative to \textit{Open Problem 2} of Bourgain and Brezis's paper mentioned above for the particular case \(l=1,\,p=2,\,s=n/2.\)