Bilinear decompositions and commutators of singular integral operators (1105.0486v5)
Abstract: Let $b$ be a $BMO$-function. It is well-known that the linear commutator $[b, T]$ of a Calder\'on-Zygmund operator $T$ does not, in general, map continuously $H1(\mathbb Rn)$ into $L1(\mathbb Rn)$. However, P\'erez showed that if $H1(\mathbb Rn)$ is replaced by a suitable atomic subspace $\mathcal H1_b(\mathbb Rn)$ then the commutator is continuous from $\mathcal H1_b(\mathbb Rn)$ into $L1(\mathbb Rn)$. In this paper, we find the largest subspace $H1_b(\mathbb Rn)$ such that all commutators of Calder\'on-Zygmund operators are continuous from $H1_b(\mathbb Rn)$ into $L1(\mathbb Rn)$. Some equivalent characterizations of $H1_b(\mathbb Rn)$ are also given. We also study the commutators $[b,T]$ for $T$ in a class $\mathcal K$ of sublinear operators containing almost all important operators in harmonic analysis. When $T$ is linear, we prove that there exists a bilinear operators $\mathfrak R= \mathfrak R_T$ mapping continuously $H1(\mathbb Rn)\times BMO(\mathbb Rn)$ into $L1(\mathbb Rn)$ such that for all $(f,b)\in H1(\mathbb Rn)\times BMO(\mathbb Rn)$, we have \label{abstract 1} b,T= \mathfrak R(f,b) + T(\mathfrak S(f,b)), where $\mathfrak S$ is a bounded bilinear operator from $H1(\mathbb Rn)\times BMO(\mathbb Rn)$ into $L1(\mathbb Rn)$ which does not depend on $T$. In the particular case of $T$ a Calder\'on-Zygmund operator satisfying $T1=T*1=0$ and $b$ in $BMO{\rm log}(\mathbb Rn)$-- the generalized $\BMO$ type space that has been introduced by Nakai and Yabuta to characterize multipliers of $\BMO(\bRn)$ --we prove that the commutator $[b,T]$ maps continuously $H1_b(\mathbb Rn)$ into $h1(\mathbb Rn)$. Also, if $b$ is in $BMO(\mathbb Rn)$ and $T*1 = T*b = 0$, then the commutator $[b, T]$ maps continuously $H1_b (\mathbb Rn)$ into $H1(\mathbb Rn)$. When $T$ is sublinear, we prove that there exists a bounded subbilinear operator $\mathfrak R= \mathfrak R_T: H1(\mathbb Rn)\times BMO(\mathbb Rn)\to L1(\mathbb Rn)$ such that for all $(f,b)\in H1(\mathbb Rn)\times BMO(\mathbb Rn)$, we have \label{abstract 2} |T(\mathfrak S(f,b))|- \mathfrak R(f,b)\leq |b,T|\leq \mathfrak R(f,b) + |T(\mathfrak S(f,b))|. The bilinear decomposition (\ref{abstract 1}) and the subbilinear decomposition (\ref{abstract 2}) allow us to give a general overview of all known weak and strong $L1$-estimates.