On weighted compactness of commutators related with Schrödinger operators (Q2150680)

In this paper, one studies the weighted compactness of commutators of some Schrödinger operators, which include Riesz transforms, Calderón-Zygmund operators, and Littlewood-Paley functions. The operators considered will be related to the Schrödinger operator \(\mathcal{L}=-\Delta+V\) on \(\mathbb{R}^d\), \(d\ge 3\), where \(\Delta\) is the Laplace operator and \(V\ge 0\) a potential in the reverse Hölder class \(B_q\), for some \(q>d/2\). Let us consider the function \(\rho(x)\) which is the radius of the maximum ball \(B(x,r)\) where the mean value of the potential \(V\) is less or equal to \(1\). One introduces then a new BMO space, BMO\((\rho)\), defined as the usual BMO but modifying the measure of a ball \(B=B(x,r)\) by the factor \(\psi(B)=\psi_\theta(B)=(1 + r/\rho(x))^\theta\), for \(\theta>0\). Then CMO\((\rho)\) denotes the closure of \(C^\infty_c\), the set of all smooth functions with compact support, in the topology of BMO\((\rho)\). As well, the class of weights \(A^\rho_p= A^{\rho,\theta}_p\) is the set of functions satisfying the ordinary weight condition but with the measure of the balls modified by the function \(\psi(B)\). If \(T\) is a Calderón-Zygmund operator related to \(-\Delta\), as for instance \[ (-\Delta+V)^{(-1)}\nabla,\ (-\Delta+V)^{(-1/2)},\ (-\Delta+V)^{(-1/2)}\nabla,\ (-\Delta+V)^{i\gamma},\ \gamma\in\mathbb{R}, \] then one proves Theorem: Let \(1<p<\infty\), \(b\in\mathrm{CMO}(\rho)\) and \(w\in A^\rho_p\). Then the commutator \([b,T]\) is a compact operator from the weighted space \(\mathcal{L}^p(w)\) to itself. It is also considered the weighted compactness of square functions. More concretely, if \(g_Q\) is the Littlewood-Paley \(g\)-function and \(S_Q\) the area function related to Schrödinger operators, let \(g_Q,b\) and \(S_Q,b\) be the commutators of \(g_Q\) and \(S_Q\) with \(b\in\mathrm{BMO}(\rho)\), respectively. Then the following result is proved Theorem: Let \(1< p<\infty\), \(b\in\mathrm{CMO}(\rho)\) and \(w\in A^\rho_p\). Then \(g_{Q,b}\) and \(S_{Q,b}\) are compact operators from \(\mathcal{L}^p(w)\) to itself. Finally, for \(V\in B_q\), \(q> d/2\), one deals with the Riesz transforms \(T_i\), \(i=1,2,3\) associated with Schrödinger operator, defined as \[ T_1= (-\Delta+ V)^{(-1)}V,\ T_2= (-\Delta+ V)^{(-1/2)}V^{(1/2)},\ T_3= (-\Delta+ V)^{(-1/2)}\nabla. \] One obtains Theorem: Suppose that \(V\in B_q\), \(q > d/2\). Let \(b\in\mathrm{CMO}(\rho)\). Then the following statements hold: \begin{itemize} \item [(i)] If \(q^\prime\le p<\infty\) and \(w\in A^\rho_{(p/q^\prime)}\), \([b, T_1]\) is a compact operator from \(\mathcal{L}^p(w)\) to itself; \item [(ii)] If \((2q)^\prime<p<\infty\) and \(w\in A^\rho_{(p/(2q)^\prime)}\), \([b, T_2]\) is a compact operator from \(\mathcal{L}^p(w)\) to itself; \item [(iii)] If \(p^\prime_0< p<\infty\) and \(w\in A^\rho_{(p/p_0^\prime)}\), where \(1/p_0= 1/q-1/d\) and \(d/2<q< d\), \([b, T_3]\) is a compact operator from \(\mathcal{L}^p(w)\) to itself. \end{itemize}