Molecules and Calderón-Zygmund operators with noncommuting kernels on H 1 c \mathrm{H}_{1}^{c}
Antonio Ismael Cano-Mármol · 2025
Read Molecules and Calderón-Zygmund operators with noncommuting kernels on H 1 c \mathrm{H}_{1}^{c} by Antonio Ismael Cano-Mármol (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-01798.txt.
Introduction
Antonio Ismael Cano-Marmol studies Hardy spaces in a setting where values do not commute, meaning order matters when they combine. These semicommutative Hardy spaces extend classic spaces used to study functions on a disk. He describes them through molecules, which are small building blocks that can rebuild any member of the space. That description is the core tool of the work.
With the molecule picture in hand, he bounds a class of singular integral operators named for Calderon and Zygmund. The operators take values in a semifinite von Neumann algebra, a broad noncommuting number system. The bounds hold from the space back into itself. Such bounds matter for harmonic analysis, the study of waves and signals, in quantum settings.
Catalog record
| CREATOR | Antonio Ismael Cano-Mármol |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-01798
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Molecules and Calderón-Zygmund operators with noncommuting kernels on H 1 c \mathrm{H}_{1}^{c} Antonio Ismael Cano-Mármol Abstract . We study the description of semicommutative Hardy spaces in terms of molecules. We use this characterization to obtain H 1 c − H 1 c \mathrm{H}_{1}^{c}-\mathrm{H}_{1}^{c} estimates for Calderón-Zygmund operators with kernels with values in a semifinite von Neumann algebra ℳ \mathcal{M} . 2020 Mathematics Subject Classification 42B20, 42B35, 46L51, 46L52 Introduction In this paper, we introduce sufficient conditions for the boundedness of Calderón-Zygmund operators with noncommuting kernels from the operator-valued version of the Hardy space \mathrm{H}_{1} to itself. This complements the results which were obtained in the work by the author and Ricard [ 1 ] , and can be framed within the theory of semicommutative Calderón-Zygmund operators. Let (\mathcal{M},\tau) be a semifinite von Neumann algebra of operators on a separable Hilbert space, equipped with a normal semifinite faithful trace \tau . Denote by \mathcal{A} the weak operator closure of the space of essentially bounded (strongly measurable) functions f:\mathbb{R}\longrightarrow\mathcal{M} acting on L_{2}(\mathbb{R};L_{2}(\mathcal{M})) . The von Neumann algebra \mathcal{A} can be identified with the tensor product L_{\infty}(\mathbb{R})\overline{\otimes}\mathcal{M} equipped with the trace \displaystyle\varphi(f)=\int_{\mathbb{R}}\tau(f(x))\ dx. We will restrict ourselves to dimension 1, even though our arguments extend trivially to any finite dimension namely for L_{\infty}(\mathbb{R}^{n})\overline{\otimes}\mathcal{M} . The noncommutative L_{p} -spaces associated with \mathcal{A} are indeed vector-valued L_{p} -spaces: more clearly [ 10 , Chapter 3] L_{p}(\mathcal{A})=L_{p}(\mathbb{R};L_{p}(\mathcal{M})), for 1\leq p<\infty . However, in this note we will discuss the boundedness of operators on the Hardy space associated to \mathcal{A} . More clearly, boundedness results of the type \mathrm{H}_{1}\longrightarrow\mathrm{H}_{1} . This question was studied for scalar-valued functions [ 2 , 9 , 8 ] as well as for the vector-valued setting [ 4 , 6 ] , where the existence of the atomic decomposition plays an essential role. This technique does not seem to have been exploited as often in the noncommutative setting except perhaps in [ 5 ] and more recently in [ 1 ] . Mei [ 7 ] was the first to introduce the so-called operator-valued Hardy space \mathrm{H}_{1}(\mathbb{R},\mathcal{M}) in this context via noncommutative equivalents of the Poisson integral, the Lusin area integral and the Littlewood-Paley g function. These techniques allowed Mei to identify the dual space of \mathrm{H}_{1}(\mathbb{R},\mathcal{M}) , which is denoted by \mathsf{BMO}(\mathbb{R},\mathcal{M}) , in the spirit of the classical argument by Fefferman and Stein [ 3 ] . More recently, the author and Ricard [ 1 ] introduced an alternative definition of the operator-valued Hardy space via a “new” atomic decomposition of the Hardy space. A c -atom is a function a\in L_{1}(\mathcal{A}) which admits a factorization of the form a=bh for some function b:\mathbb{R}\rightarrow L_{2}(\mathcal{M}) and an norm-one operator h\in L_{2}(\mathcal{M}) , satisfying (1) \mathrm{supp}_{\mathbb{R}}(b)\subseteq I for some interval I , (2) \displaystyle\int_{I}b=0 , (3) \displaystyle\|b\|_{L_{2}(\mathbb{R};L_{2}(\mathcal{M}))}\leq\frac{1}{\sqrt{|I|}} . Then, the column Hardy space \mathrm{H}_{1}^{c}(\mathcal{A}) is defined to be the subspace of elements in L_{1}(\mathcal{A}) of the form \displaystyle\sum\limits_{i=0}^{\infty}\lambda_{i}a_{i}\ \mathrm{where\ }(\lambda_{i})_{i}\in\ell_{1}\ \mathrm{and}\ (a_{i})_{i}\ c\mbox{-}\mathrm{atoms} with respect to the norm \displaystyle\|f\|_{\mathrm{H}_{1}^{c}(\mathcal{A})}=\inf\Big\{\sum_{i=0}^{\infty}|\lambda_{i}|\ :\ f=\sum_{i=0}^{\infty}\lambda_{i}a_{i}\Big\}. The row space \mathrm{H}_{1}^{r}(\mathcal{A}) is defined analogously via r -atoms of the form a=hb , and \mathrm{H}_{1}(\mathcal{A})=\mathrm{H}_{1}^{c}(\mathcal{A})+\mathrm{H}_{1}^{r}(\mathcal{A}) . Let \mathcal{S} denote the set of compactly supported essentially bounded functions \mathbb{R}\to L_{\infty}\cap L_{1}(\mathcal{M}) (measurable with values in L_{1} ). Let T be a bounded operator on L_{2}(\mathcal{A}) for which there exists a kernel K:\mathbb{R}\times\mathbb{R}\ \setminus\ \{x=y\}\longrightarrow\mathcal{M} such that for every pair of intervals I , J satisfying d(I,J)>0 , there exists K_{I,J}\in L_{\infty}(I\times J;\mathcal{M}) such that K(t)=K_{I,J}(t)\ \mbox{for any }t\in I\times J and \displaystyle\int T(f)(x)g(x)\ dx=\int\int K_{I,J}(x,y)f(y)g(x)\ dx\ dy holds for any f,g\in\mathcal{S} satisfying \mathrm{supp}\|f\|_{L_{2}(\mathcal{M})}\subset J and \mathrm{supp}\|g\|_{L_{2}(\mathcal{M})}\subset I . More technical details on this definition can be found in [ 1 ] . Under these assumptions, we say that T is a Calderón-Zygmund operator with kernel K . Moreover, if T fulfills a right-modularity condition, that is, \displaystyle T(fh)=T(f)h for any f\in L_{2}(\mathcal{A}) with compact support and h\in\mathcal{M} , we say that T is a left Calderón-Zygmund operator . The main result in [ 1 ] states that whenever the kernel K satisfies the Hörmander condition , that is, that for some \lambda>0 there holds (1) \displaystyle\int_{|x-y|\geq\lambda|y^{\prime}-y|}\|K(x,y)-K(x,y^{\prime})\|_{\mathcal{M}}\ dx<\infty, then T is bounded from \mathrm{H}_{1}^{c}(\mathcal{A}) to L_{1}(\mathcal{A}) . As a straightforward consequence, there follows that the Hardy space \mathrm{H}_{1}(\mathcal{A}) coincides with the one introduced by Mei [ 7 ] . Nonetheless, the condition ( 1 ) is not sufficient to prove the boundedness of Calderón-Zygmund operators from \mathrm{H}_{1}^{c}(\mathcal{A}) to itself. Instead, a stronger assumption is required, namely the Lipschitz condition : there exists some \lambda>0 and \gamma\in(1/2,1] such that (2) \displaystyle\|K(x,y)-K(x,y^{\prime})\|_{\mathcal{M}}\lesssim\frac{|y^{\prime}-y|^{\gamma}}{|x-y|^{1+\gamma}}\mbox{ whenever }|y^{\prime}-y|\leq\frac{|x-y|}{\lambda}. Theorem 1 . Let \mathcal{M} be a von Neumann algebra. Let T be a left Calderón-Zygmund operator with associated kernel K:\mathbb{R}\times\mathbb{R}\ \setminus\ \{x=y\}\longrightarrow\mathcal{M} . If K satisfies the Lipschitz condition ( 2 ) and \int_{\mathbb{R}}T(b)=0 for every c -atom a=bh , then T extends to a bounded operator from \mathrm{H}_{1}^{c}(\mathcal{A}) to \mathrm{H}_{1}^{c}(\mathcal{A}) . This result heavily relies on the connection between the theory of vector-valued Hardy spaces and the semicommutative Hardy space \mathrm{H}_{1}^{c}(\mathcal{A}) . More clearly, it is based on the decomposition of \mathrm{H}_{1}^{c}(\mathcal{A}) into column-valued versions of molecules , which have been widely studied in the classical setting [ 2 , 9 , 8 ] . Analogous statements follow for right-Calderón-Zygmund operators on \mathrm{H}_{1}^{r}(\mathcal{A}) and the vector-valued setting. Therefore, operators with scalar-valued kernels satisfying both modularity conditions happen to be bounded on the full Hardy space \mathrm{H}_{1}(\mathcal{A}) . 1. Preliminaries Noncommutative spaces L_{p}(\mathcal{M};L_{2}^{c}(\Omega)) Let H be a separable Hilbert space. Let \mathds{1} be a norm-one element in H , and let p_{\mathds{1}}=\mathds{1}\otimes\overline{\mathds{1}} denote the rank-one projection onto \mathrm{span}\{\mathds{1}\} . Given 0<p\leq\infty , we define the column H -valued L_{p} space as \displaystyle L_{p}(\mathcal{M};H^{c})=L_{p}(\mathcal{M}\overline{\otimes}B(H))(\mathbf{1}_{\mathcal{M}}\otimes p_{\mathds{1}}) and the row H -valued L_{p} space as \displaystyle L_{p}(\mathcal{M};H^{*r})=(\mathbf{1}_{\mathcal{M}}\otimes p_{\mathds{1}})L_{p}(\mathcal{M}\overline{\otimes}B(H)). Identify L_{p}(\mathcal{M}) as a subspace of L_{p}(\mathcal{M}\overline{\otimes}B(H)) via the map m\mapsto m\otimes p_{\mathds{1}} . This is equivalent to the identity \displaystyle L_{p}(\mathcal{M})=(\mathbf{1}_{\mathcal{M}}\otimes p_{\mathds{1}})L_{p}(\mathcal{M}\overline{\otimes}B(H))(\mathbf{1}_{\mathcal{M}}\otimes p_{\mathds{1}}). Thus, given an element f in L_{p}(\mathcal{M};H^{c}) , then f^{*}f\in L_{p/2}(\mathcal{M}) , which justifies defining \displaystyle\|f\|_{L_{p}(\mathcal{M};\mathcal{H}^{c})}=\|f\|_{L_{p}(\mathcal{M}\overline{\otimes}B(H))}=\|(f^{*}f)^{1/2}\|_{L_{p}(\mathcal{M})}. Analogously, if f\in L_{p}(\mathcal{M};H^{*r}) , then ff^{*}\in L_{p/2}(\mathcal{M}) , which enables us to set \|f\|_{L_{p}(\mathcal{M};H^{*r})}=\|f^{*}\|_{L_{p}(\mathcal{M};H^{c})} . We will use without reference that the algebraic tensor L_{p}(\mathcal{M})\otimes H is dense (resp. weak ∗ dense) in L_{p}(\mathcal{M};H^{c}) for 1\leq p<\infty (resp. p=\infty ) and similarly for rows. Column and row Hilbert-valued L_{p} -spaces satisfy the expected duality relations expressed via the natural duality bracket (3) \displaystyle\langle f,g\rangle_{r,c}=\mathrm{Tr\otimes\tau}(fg) where \mathrm{Tr} denotes the trace of B(H) . More clearly, there holds linearly isometrically \displaystyle L_{p}(\mathcal{M};\mathcal{H}^{c})^{*}=L_{p^{\prime}}(\mathcal{M};H^{*r})\ \mathrm{and}\ L_{p}(\mathcal{M};H^{*r})^{*}=L_{p^{\prime}}(\mathcal{M};\mathcal{H}^{c}). for any 1\leq p<\infty whenever 1/p+1/p^{\prime}=1 . Let (\Omega,\mu) be a \sigma -finite measure space. A remarkable setting for noncommutative Hilbert-valued column/row L_{p} -spaces is the case H=L_{2}(\Omega):=L_{2}(\Omega,\mu) . Identifying L_{2}(\Omega,\mu)^{*} and L_{2}(\Omega,\mu) and using the bilinear pairing (f,g)\mapsto\int_{\Omega}fg\ d\mu yields the following duality identity \displaystyle L_{p^{\prime}}(\mathcal{M};L_{2}^{r}(\Omega,\mu))=L_{p}(\mathcal{M};L_{2}^{c}(\Omega,\mu))^{*}\ \mathrm{for\ }1\leq p<\infty. Moreover, for F=\sum_{i=1}^{n}m_{i}\otimes f_{i}\in L_{p}(\mathcal{M})\otimes L_{2}(\Omega) with p<\infty : (4) \displaystyle\|F\|^{p}_{L_{p}(\mathcal{M};L_{2}^{c}(\Omega,\mu))}=\tau\Big(\int_{\Omega}\Big|\sum_{i=1}^{n}f_{i}(t)m_{i}\Big|^{2}d\mu\Big)^{p/2} In this work, only the case p=1 will be relevant. It turns out that L_{1}(\mathcal{M};L_{2}^{c}(\Omega,\mu))\subseteq L_{2}(\Omega;L_{1}(\mathcal{M})) so that the former space can be identified with a.e. Bochner measurable functions from \Omega to L_{1}(\mathcal{M}) . Moreover, according to the discussion above, L_{1}(\mathcal{M})\otimes L_{2}(\Omega) is dense in L_{1}(\mathcal{M};L_{2}^{c}(\Omega,\mu)) with respect to the topology given by ( 4 ) for p=1