Abstract
Let \({\mathcal {L}}=-\Delta _{{\mathbb {H}}^{n}}+V\) be the Schrödinger operator on the Heisenberg group \({\mathbb {H}}^{n}\), where \(\Delta _{{\mathbb {H}}^{n}}\) is the sub-Laplacian and the nonnegative potential V belongs to the reverse Hölder class \(B_{q}\) for \(q\ge Q/2\). Suppose that \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), which is larger than \(BMO({\mathbb {H}}^{n})\). We prove that the operator \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) is bounded from the Herz space \({\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})\) into \({\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})\). By a maximal estimate, we obtain the boundedness of the commutators \([b,T_{\beta _{1},\beta _{2}}]\) and \([b,T_{\beta }]\) from \({\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})\) into \({\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})\), where \(T_{\beta }=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta }\).
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Let \({\mathbb {H}}^{n}\) be the Heisenberg group with the homogeneous dimension \(Q=2n+2\). Let \({\mathcal {L}}=-\Delta _{{\mathbb {H}}^{n}}+V\) be the Schrödinger operator on \({\mathbb {H}}^{n}\), where \(\Delta _{{\mathbb {H}}^{n}}\) is the sub-Laplacian and the nonnegative potential V belongs to the reverse Hölder class \(B_{q}\), \(q\ge Q/2\). The Riesz transforms and the fractional integrals associated with \({\mathcal {L}}\) have been studied extensively. Shen [15] proved the \(L^{p}\)-boundedness of the operators \((-\Delta +V)^{i\gamma }\), \(\nabla ^{2}(-\Delta +V)^{-1}\), \(\nabla (-\Delta +V)^{-1/2}\), and \(\nabla (-\Delta +V)^{-1}\nabla \), where \(\Delta \) denotes the Laplace operator on \({\mathbb {R}}^{n}\), i.e., \(\Delta =\sum ^{n}_{i=1}\partial ^{2}/\partial x^{2}_{i}\). Yang-Yang-Zhou [19] obtained the endpoint properties of the Riesz transforms and the fractional integrals associated to \({\mathcal {L}}\). For further information, we refer the reader to Guo-Li-Peng [3], Liu [10], Liu-Huang-Dong [11], Tang-Dong [17] and the references therein. For \(0\le \beta _{2}\le \beta _{1}<Q/2\), let
For \(0<\beta _{2}\le \beta _{1}<1\), Sugano [16] obtained the estimates for \(T_{\beta _{1},\beta _{2}}\) and \(T^{*}_{\beta _{1},\beta _{2}}\) on \({\mathbb {R}}^{n}\). If \(\beta _{2}=0\), we know that \(T_{\beta _{1},0}=T_{\beta _{1}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}\). In recent years, the fractional integral operator \(I_{\alpha }=(-\Delta +V)^{-\alpha }\) has been studied extensively. We refer to Duong-Yan [2], Jiang [4], Tang-Dong [17] and Yang-Yang-Zhou [19] for details. For \(\beta _{1}=\beta _{2}=1\) and \(\beta _{1}=\beta _{2}=1/2\), \(T^{*}_{1,1}=V(-\Delta +V)^{-1}\) and \(T^{*}_{1/2,1/2}=V^{1/2}(-\Delta +V)^{-1/2}\) are studied by Shen [15]. We also refer the reader to [14] for the Fefferman-Phong type inequality for \(\mathcal L\) on \(\mathbb H^{n}\).
As a useful tool in analysis, the commutators related to singular integral operators play an important role in the study of harmonic analysis and partial differential equations. Let \(R_{{\mathcal {L}}}=\nabla _{{\mathbb {H}}^{n}}(-\Delta _{{\mathbb {H}}^{n}}+V)^{-1/2}\) be the Riesz transform associated with \({\mathcal {L}}\), \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) and \(b\in BMO({\mathbb {H}}^{n})\). Li-Peng [6] extended the results of [3] to the Heisenberg group \({\mathbb {H}}^{n}\). Li-Wan [7] studied the boundedness on generalized Morrey spaces of commutators \([b,R_{{\mathcal {L}}}]\), \([b,T_{\beta _{1},\beta _{2}}]\) and \([b,T^{*}_{\beta _{1},\beta _{2}}]\), see also Liu-Tang [12] and [13]. Bongioanni–Harboure–Salinas [1] introduced a new class of BMO-type spaces denoted by \(BMO_{\rho }^{\theta }({\mathbb {R}}^{n})\) as generalizations of \(BMO({\mathbb {R}}^{n})\). By use of a maximal estimate, the authors proved that the commutators \([b,R_{{\mathcal {L}}}]\) are bounded on \(L^{p}({\mathbb {R}}^{n})\) with \(b\in BMO^{\theta }_{\rho }({\mathbb {R}}^{n})\).
In this paper, we investigate the boundedness of commutators related to Schrodinger operators on Herz spaces in the setting of Heisenberg groups. In Sect. 3.1, suppose that \(V \in B_{q}\), \(q\ge Q/2\). Let \(q_{0}=\sup \{q:V\in B_{q}\}\). Therefore, \(q_{0}>Q/2\) under the assumption of \(V\in B_{q}\). Given \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\). Let \(1/p_{0}=(1/q_{0}-1/Q)^{+}\). We investigate the boundedness of the commutators \([b,R_{{\mathcal {L}}}]\) and \([b,R^{*}_{{\mathcal {L}}}]\) on \(L^{p}({\mathbb {H}}^{n})\). Let \(\beta _{1}=\beta _{2}=\beta \). Li-Wan [8] further extended the \(L^{p}\)-boundedness of \([b,R_{{\mathcal {L}}}]\) and \([b,T_{\beta ,\beta }]\) on Herz spaces \({\dot{K}}_{q}^{\alpha ,p}({\mathbb {R}}^{n})\) and \(K_{q}^{\alpha ,p}({\mathbb {R}}^{n})\), respectively. In Sect. 4, we prove that if the index \((\beta _{1}, \beta _{2}, p_{1}, p_{2}, \alpha )\) satisfies
then \(T_{\beta _{1},\beta _{2}}\) and \([b,T_{\beta _{1},\beta _{2}}]\) are bounded from homogeneous Herz spaces \({\dot{K}}^{\alpha ,p}_{p_{1}}({\mathbb {H}}^{n})\) to \({\dot{K}}^{\alpha ,p}_{p_{2}}({\mathbb {H}}^{n})\), respectively. The rest of this paper is organized as follows. In Sect. 2, we state some notations and known results which will be used throughout rest of this paper. In Sect. 3, we extend the results of [1] on Euclidean space to the Heisenberg group \({\mathbb {H}}^{n}\), see Theorem 1 and Theorem 2 for details. Moreover, we prove that the operator \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) and the commutator \([b,T_{\beta _{1},\beta _{2}}]\) are bounded from \(L^{p_{1}}({\mathbb {H}}^{n})\) to \(L^{p_{2}}({\mathbb {H}}^{n})\), where \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), \(0\le \beta _{2}\le \beta _{1}<Q/2\). In Sect. 4, with the help of the \(L^{p}\)-boundedness of the operator \(T_{\beta _{1},\beta _{2}}\) and the commutator \([b,T_{\beta _{1},\beta _{2}}]\), we prove that the commutators \([b,T_{\beta _{1}\beta _{2}}]\) and \([b,T_{\beta }]\) are bounded from Herz spaces \({\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})\) to \({\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})\) with \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), respectively.
In what follows we recall some basic facts for the Heisenberg group \({\mathbb {H}}^{n}\). The Heisenberg group \({\mathbb {H}}^{n}\) is a Lie group with the underlying manifold \({\mathbb {R}}^{n}\times {\mathbb {R}}^{n}\times {\mathbb {R}}\), and the multiplication
A basis for the Lie algebra of left-invariant vector fields on \({\mathbb {H}}^{n}\) is given by
All non-trivial commutation relations are given by \([\mathrm {X}_{j},\mathrm {Y}_{j}]=-4\mathrm {T}\), \(j=1,2,...,n\). Then the sub-Laplacian \(\Delta _{{\mathbb {H}}^{n}}\) is defined by \(\Delta _{{\mathbb {H}}^{n}}=\sum _{j=1}^{n}(\mathrm {X}_{j}^{2}+\mathrm {Y}_{j}^{2})\) and the gradient operator \(\nabla _{{\mathbb {H}}^{n}}\) is defined by
The dilation on \({\mathbb {H}}^{n}\) is of the form \(\delta _{\lambda }(x,y,t)=(\lambda x,\lambda y,\lambda ^{2} t)\), \(\lambda >0\). The Haar measure on \({\mathbb {H}}^{n}\) coincides with the Lebesgue measure on \({\mathbb {R}}^{n}\times {\mathbb {R}}^{n} \times {\mathbb {R}}\). We denote the measure of any measurable set E by |E|. Then \(|\delta _{\lambda }E|=\lambda ^{Q}|E|\), where \(Q=2n+2\) is called the homogeneous dimension of \({\mathbb {H}}^{n}\).
We define a homogeneous norm function on \({\mathbb {H}}^{n}\) by \(|g|=\Big ((|x|^{2}+|y|^{2})+|t|^{2}\Big )^{1/4}\), \(g=(x,y,t)\in {\mathbb {H}}^{n}\). This norm satisfies the triangular inequality and leads to a left-invariant distant function \(d(g,h)=|g^{-1}h|\). Then the ball of radius r centered at g is given by \(B(g,r)=\{h\in {\mathbb {H}}^{n}:|g^{-1}h|<r\}\). The ball B(g, r) is the left translation by g of B(0, r) and we have \(|B(g,r)|=\alpha ^{1}r^{Q}\), where \(\alpha ^{1}=|B(0,1)|\).
Throughout this paper, the symbol \(A\lesssim B\) means \(A\le CB\), where C is a universal positive constant and independent of all important parameters. If \(A\lesssim B\) and \(B\lesssim A\), we denote \(A\thickapprox B\).
2 Preliminaries
2.1 Schrödinger Operators and the Auxiliary Function
Definition 1
A nonnegative locally \(L^{q}\) integrable function V on \({\mathbb {H}}^{n}\) is said to belong to \(B_{q}\), \(1<q<\infty \), if there exists a constant \(C>0\) such that the reverse Hölder inequality
holds for every ball in \({\mathbb {H}}^{n}\).
In the analysis of Schrödinger operators, the following auxiliary function plays an important role.
Definition 2
For \(g\in {\mathbb {H}}^{n}\), the auxiliary function \(\rho (\cdot )\) is defined by
Now, we give some related lemmas about \(\rho (\cdot )\). Lemmas 1–4 have been proved in [10].
Lemma 1
Suppose that \(V\in B_{q}\) with \(q>Q/2\). The measure V(h)dh is a doubling measure, that is, there exist constants \(\lambda \ge 1\) and C such that
holds for every ball B(g, r) in \({\mathbb {H}}^{n}\) and \(t>1\).
Lemma 2
Suppose that \(V\in B_{q}\) with \(q>Q/2\). There exists a constant \(C>0\) such that for \(0<r<R<\infty \),
Lemma 3
There exist \(C>0\) and \(N_{0}>0\) such that
Lemma 4
Suppose that \(V\in B_{q}\) with \(q>Q/2\). There exist constants \(C>0\) and \(k_{0}>0\) such that
holds for all \(g,h\in {\mathbb {H}}^{n}\).
A ball centered at g and with radius \(\rho (g)\) is called a critical ball and is denoted by \({\mathcal {Q}}=B(g,\rho (g))\). (1) implies that if \(g,h\in {\mathcal {Q}}=B(g,\rho (g))\), then \(\rho (g)\le C_{0}\rho (h)\), where the constant \(C_{0}\) depends on the constants C and \(k_{0}\) in (1).
Lemma 5
Suppose that \(V\in B_{q}\) with \(q>Q/2\). Let \(l_{0}\in {\mathbb {N}}\) and \(g\in 2^{k+1}B(g_{0},r)\setminus 2^{k}B(g_{0},r)\). Then
Proof
The proof is similar to that of [8, Lemma 2.17]. \(\square \)
Proposition 1
There exists a sequence of points \(\left\{ g_{k}\right\} _{k=1}^{\infty }\subset {\mathbb {H}}^{n}\) such that the family of critical balls \(\{{\mathcal {Q}}_{k}=B(g_{k},\rho (g_{k}))\}_{k=1}^{\infty }\) satisfies:
-
(i)
\({\mathbb {H}}^{n}=\cup _{k}{\mathcal {Q}}_{k}\).
-
(ii)
There exists N such that for every \(k\in {\mathbb {N}}\), card\(\{j:4{\mathcal {Q}}_{j}\cap 4{\mathcal {Q}}_{k}\ne \varnothing \}\le N\).
Proof
We refer to [9, page 23 and page 24 ] \(\square \)
2.2 New BMO-Type Spaces \(BMO_{\mathcal {\rho }}^{\theta }({\mathbb {H}}^{n})\)
As in [1], we define the following new BMO-type space \(BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\).
Definition 3
The new BMO-type space \(BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\) with \(0<\theta <\infty \) is defined as the set of all locally integrable functions b on \({\mathbb {H}}^{n}\) such that
holds for all \(g\in {\mathbb {H}}^{n}\) and \(r>0\), where \(b_{B}=\frac{1}{|B|}\int _{B}b(h)dh\). The norm for \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), denoted by \([b]_{\theta }\), is given by the infimum of the constants satisfying (2). We define \(BMO_{\rho }^{\infty }({\mathbb {H}}^{n})=\cup _{\theta >0}BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\). Clearly, \(BMO({\mathbb {H}}^{n})\subset BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\subset BMO_{\rho }^{\theta '}({\mathbb {H}}^{n})\) for \(0<\theta <\theta '\), and hence \(BMO({\mathbb {H}}^{n})\subset BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\).
The space \(BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\) is a generalization of the classical BMO-type space associated with Schrödinger operators which is defined as follows.
Definition 4
The space \(BMO_{{\mathcal {L}}}({\mathbb {H}}^{n})\) is defined as the set of all locally \(L^{1}\) integrable functions f on \({\mathbb {H}}^{n}\) such that
where C is a constant and \(B=B(g,r)\).
Definition 5
The space \(BMO_{\rho }^{\theta ,log}({\mathbb {H}}^{n})\) is defined as the set of all functions b such that
holds for all \(g\in {\mathbb {H}}^{n}\) and \(r>0\). We define \(BMO_{\rho }^{\infty ,log}({\mathbb {H}}^{n})=\cup _{\theta >0}BMO_{\rho }^{\theta ,log}({\mathbb {H}}^{n})\).
We state the following lemma concerning the function \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\) which will be used in the proofs of the main results.
Lemma 6
Let \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), \(B=B(g,r)\), \(\theta >0\). If \(1\le s<\infty \), \(\theta '=(k_{0}+1)\theta \), and \(k_{0}\) is the constant appearing in (1). Then
-
(i)
for all \(g\in {\mathbb {H}}^{n}\) and \(r>0\),
$$\begin{aligned} \left( \frac{1}{|B|}\int _{B}|b(h)-b_{B}|^{s}dh\right) ^{1/s}\lesssim [b]_{\theta }\Big (1+\frac{r}{\rho (g)}\Big )^{\theta '}. \end{aligned}$$ -
(ii)
For all \(g\in {\mathbb {H}}^{n}\) and \(k\in {\mathbb {N}}\) with \(r>0\),
$$\begin{aligned} \Big (\frac{1}{|2^{k}B|}\int _{2^{k}B}|b(h)-b_{B}|^{s}dh\Big )^{1/s}\lesssim k[b]_{\theta }\Big (1+\frac{2^{k}r}{\rho (g)}\Big )^{\theta '}. \end{aligned}$$
Proof
The proof of (i) is similar to that of [1, Proposition 1] and the proof of (ii) is similar to that of [1, Lemma 1]. \(\square \)
For \(k\in {\mathbb {Z}}\), let \(B_{k}=\{g\in {\mathbb {H}}^{n}:|g|\le 2^{k}\}\) and \(E_{k}=B_{k}\setminus B_{k-1}\). Denote by \(\chi _{k}\) the characteristic function of \(E_{k}\). The Herz spaces are defined as follows.
Definition 6
Let \(\alpha \in {\mathbb {H}}^{n}\), \(0<p,q\le \infty \). The homogeneous Herz space \({\dot{K}}_{q}^{\alpha ,p}({\mathbb {H}}^{n})\) is defined by
where
2.3 Estimates of the Kernels
We recall the estimates of kernels. Let \(\Gamma (g,h,\tau )\) denote the fundamental solution for the operator \(-\Delta _{{\mathbb {H}}^{n}}+i\tau \), where \(\tau \in {\mathbb {R}}\). For any \(N>0\), there exists \(C_{N}>0\) such that
where \(\nabla _{{\mathbb {H}}^{n},g}\) denote the gradient for variable g. The estimate (3) still holds for \(\nabla _{{\mathbb {H}}^{n},h}\). We remark that the explicit expression of \(|\Gamma (g,h)|=|\Gamma (g,h,0)|\) is obtained by Folland: \(|\Gamma (g,h)|=\frac{2^{n-2}\Gamma (n/2)^{2}}{\pi ^{n+1}}\frac{1}{|h^{-1}g|^{Q-2}}\).
Let \(\Gamma ^{{\mathcal {L}}}(g,h,\tau )\) denote the fundamental solution for the operator \(-\Delta _{{\mathbb {H}}^{n}}+V+i\tau \), where \(\tau \in {\mathbb {R}}\). For any \(N>0\), there exists \(C_{N}>0\) such that
Let \(R=\nabla _{{\mathbb {H}}^{n}}(-\Delta _{{\mathbb {H}}^{n}})^{-1/2}\) be the classical Riesz transforms and \(R^{*}\) be its adjoint operator. The kernels of R and \(R^{*}\) are denoted by \(R(\cdot ,\cdot )\) and \(R^{*}(\cdot ,\cdot )\), respectively. Let \(R_{{\mathcal {L}}}=\nabla _{{\mathbb {H}}^{n}}(-\Delta _{{\mathbb {H}}^{n}}+V)^{-1/2}\) be the vector-valued Riesz transform related to \({\mathcal {L}}\) and \(R^{*}_{{\mathcal {L}}}\) be its adjoint operator. The kernels of \(R_{{\mathcal {L}}}\) and \(R^{*}_{{\mathcal {L}}}\) are denoted by \(R_{{\mathcal {L}}}(\cdot ,\cdot )\) and \(R^{*}_{{\mathcal {L}}}(\cdot ,\cdot )\), respectively. We have
Definition 7
The operator \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) is defined by \(T_{\beta _{1},\beta _{2}}f(g)=\int _{{\mathbb {H}}^{n}}K_{\beta _{1},\beta _{2}}(g,h)f(h)dh\), where \(K_{\beta _{1},\beta _{2}}(g,h)=\Gamma (g,h)V^{\beta _{2}}(h)\) and \(\Gamma (g,h)=\Gamma (g,h,0)\).
Definition 8
The operator \(R_{{\mathcal {L}}}=\nabla _{{\mathbb {H}}^{n}}(-\Delta _{{\mathbb {H}}^{n}}+V)^{-1/2}\) is defined by \(R_{{\mathcal {L}}}f(g)=\int _{{\mathbb {H}}^{n}}R_{{\mathcal {L}}}(g,h)f(h)dh\).
Next, we give some lemmas about the estimates for the kernels of \(T_{\beta _{1},\beta _{2}}\), \(R_{{\mathcal {L}}}\) and \(R_{{\mathcal {L}}}^{*}\).
Lemma 7
( [18, Lemma 3.2]) Suppose \(V\in B_{q}, q>Q/2\). For any integer \(N>0\), there exists \(C_{N}>0\) such that
Lemma 8
If \(V\in B_{Q}\), then
-
(i)
for every N, there exists a constant C such that
$$\begin{aligned} |R_{{\mathcal {L}}}(g,h)|\le \frac{C}{(1+|gz^{-1}|/\rho (g))^{N}}\frac{1}{|gz^{-1}|^{Q}}. \end{aligned}$$ -
(ii)
For every N and \(0<\delta <\min \left\{ 1,1-Q/q_{0}\right\} \), there exists a constant C such that for \(|gh^{-1}|<\frac{2}{3}|gz^{-1}|\),
$$\begin{aligned} |R_{{\mathcal {L}}}(g,z)-R_{{\mathcal {L}}}(h,z)|\le \frac{C}{(1+|gz^{-1}|/\rho (g))^{N}}\frac{|gh^{-1}|^{\delta }}{|gz^{-1}|^{Q+\delta }}. \end{aligned}$$ -
(iii)
For every \(0<\sigma <2-Q/q_{0}\), then
$$\begin{aligned} |R(g,z)-R_{{\mathcal {L}}}(g,z)|\le \frac{C}{|gz^{-1}|^{Q}}\Big (\frac{|gz^{-1}|}{\rho (z)}\Big )^{\sigma }. \end{aligned}$$
Proof
The proof is similar to that of [15, Theorem 0.4] \(\square \)
Lemma 9
( [6, Lemma 5]) If \(V\in B_{Q/2}\), then
-
(i)
for every N, there exists a constant C such that
$$\begin{aligned}&|R_{{\mathcal {L}}}^{*}(g,z)|\le \frac{C(1+|gz^{-1}|/\rho (g))^{-N}}{|gz^{-1}|^{Q-1}}\nonumber \\&\quad \Big (\int _{B(z,|gz^{-1}|/4)}\frac{V(u)}{|uz^{-1}|^{Q-1}}du+\frac{1}{|gz^{-1}|}\Big ). \end{aligned}$$(4) -
(ii)
For every N and \(0<\delta <\min \left\{ 1, 2-Q/q_{0}\right\} \), there exists a constant C such that for \(|gh^{-1}|<\frac{2}{3} |gz^{-1}|\),
$$\begin{aligned} |R_{{\mathcal {L}}}^{*}(g,z)-R_{{\mathcal {L}}}^{*}(h,z)|\le & {} \frac{C(1+|gz^{-1}|/\rho (g))^{-N}|gh^{-1}|^{\delta }}{|gz^{-1}|^{Q-1+\delta }}\nonumber \\&\times \Big (\int _{B(z,|gz^{-1}|/4)}\frac{V(u)}{|uz^{-1}|^{Q-1}}du+\frac{1}{|gz^{-1}|}\Big ). \end{aligned}$$(5)
Lemma 10
( [5, Lemma 6.3]) Let \(V\in B_{Q/2}\). For every \(0<\sigma <2-Q/q_{0}\),
Lemma 11
If \(V\in B_{q_{0}}\), when \(q_{0}>Q\), then \(R_{{\mathcal {L}}}\) and \(R_{{\mathcal {L}}}^{*}\) are Calderón-Zygmund operators. The term involving V can be dropped from inequalities (4), (5), and (6).
Proof
We refer to [9, page 2]. \(\square \)
3 Boundedness of the Commutators on \(L^{p}({\mathbb {H}}^{n})\)
In this section, we consider the boundedness of commutators \([b,R_{{\mathcal {L}}}]\) and \([b,R^{*}_{{\mathcal {L}}}]\) on \(L^{p}({\mathbb {H}}^{n})\), where \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\). Furthermore, let \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\). Suppose \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\). We prove the commutator \([b,T_{\beta _{1},\beta _{2}}]\) is bounded from \(L^{p_{1}}({\mathbb {H}}^{n})\) to \(L^{p_{2}}({\mathbb {H}}^{n})\).
3.1 Technical Lemmas
We give some technical lemmas which will be used in the sequel.
Definition 9
Let \(f\in L_{loc}^{q}({\mathbb {H}}^{n})\). Denote by |B| the Lebesgue measure of the ball \(B\subset {\mathbb {H}}^{n}\). The Hardy–Littlewood maximal function Mf and its variant \(M_{\sigma ,s}f\) are defined by
If \(\sigma =0\), then \(M_{0,s}f(g)\) is denoted by \(M_{s}f(g)\).
Definition 10
Let \(\gamma >0\) and \({\mathcal {B}}_{\rho ,\gamma }=\{B(h,r):h\in {\mathbb {H}}^{n},r\le \gamma \rho (h)\}\). For \(f\in L_{loc}^{1}({\mathbb {H}}^{n})\) and \(g\in {\mathbb {H}}^{n}\), we define the following two maximal functions:
Definition 11
Let \({\mathcal {S}}_{\mathrm {Q}}=\{B(h,r):h\in \mathrm {Q}, r>0\}\) and let \(\mathrm {Q}\) be a ball in \({\mathbb {H}}^{n}\). For \(f\in L_{loc}^{1}({\mathbb {H}}^{n})\) and \(h\in \mathrm {Q}\), we define
Lemma 12
(Fefferman–Stein-type inequality) For \(1<p<\infty \), there exist \(\xi \) and \(\gamma \) such that if \(\{{\mathcal {Q}}_{k}\}_{k=1}^{\infty }\) is a sequence of the balls as those in Proposition 1, then for all \(f\in L_{loc}^{1}({\mathbb {H}}^{n})\),
Proof
The proof is similar to that of [1, Lemma 2]. \(\square \)
Lemma 13
Let \(V\in B_{Q/2}\), \(1/p_{0}=(1/q_{0}-1/Q)^{+}\), and \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\). Then, for any \(s>p_{0}'\), there exists a constant C such that
holds for all \(f\in L_{loc}^{s}({\mathbb {H}}^{n})\) and every ball \({\mathcal {Q}}=B(g_{0},\rho (g_{0}))\subset {\mathbb {H}}^{n}\). Moreover, if \(q_{0}>Q\), the above estimate also holds for \(R_{{\mathcal {L}}}\) instead of \(R_{{\mathcal {L}}}^{*}\).
Proof
Let \(f\in L^{p}({\mathbb {H}}^{n})\) and \({\mathcal {Q}}=B(g_{0},\rho (g_{0}))\). We write \([b,R_{{\mathcal {L}}}^{*}]f(g):=I-II\), where
Next we deal with the average on \({\mathcal {Q}}\) of each term.
For I, it follows from Hölder’s inequality with \(s>p_{0}'\) and Lemma 6 that
We set \(f(g)=f_{1}(g)+f_{2}(g)\), where \(f_{1}(g)=f\chi _{2{\mathcal {Q}}}(g)\) and \(f_{2}(g)=f\chi _{(2{\mathcal {Q}})^{c}}(g)\). Then, by use of the \(L^{s}({\mathbb {H}}^{n})\)-boundedness of \(R^{*}_{{\mathcal {L}}}\), we obtain
Now, for \(g\in {\mathcal {Q}}\), by (4), we have \(|R_{{\mathcal {L}}}^{*}f_{2}(g)|\lesssim I_{1}+I_{2}\), where
Via a simple computation, for \(z\in 2^{k+1}{\mathcal {Q}} \setminus 2^{k}{\mathcal {Q}}\) and \(g\in {\mathcal {Q}}\), we can deduce that \(|gz^{-1}|\thickapprox |g_{0}z^{-1}|\thickapprox 2^{k}\rho (g_{0})\). With the help of the condition \(\rho (g)\thickapprox \rho (g_{0})\), we obtain
For \(I_{2}\), we assume \(Q/2<q_{0}<Q\). Let \(p_{0}'<s<Q\). By Hölder’s inequality and the boundedness of the fractional integral \({\mathcal {I}}_{1}:L^{s'}\mapsto L^{q}\) with \(1/s'=1/q-1/Q\), we can get
where in the last inequality we have taken N large enough.
For II, we split \(f(g)=f_{1}(g)+f_{2}(g)\), where \(f_{1}(g)=f\chi _{2{\mathcal {Q}}}(g)\) and \(f_{2}(g)=f\chi _{(2{\mathcal {Q}})^{c}}(g)\). We select \(p_{0}'<{\widetilde{s}}<s\) and denote \(m=\frac{{\widetilde{s}}s}{s-{\widetilde{s}}}\). With the help of Hölder’s inequality and Lemma 6, we can deduce that
Now, we consider the term \(R_{{\mathcal {L}}}^{*}((b-b_{{\mathcal {Q}}})f_{2})(g)\). Applying (4), we have \(|R_{{\mathcal {L}}}^{*}((b-b_{{\mathcal {Q}}})f_{2})(g)|\lesssim II_{1}+II_{2}\), where
For \(1\le {\widetilde{s}}<s\) and \(m=\frac{{\widetilde{s}}s}{s-{\widetilde{s}}}\), with the help of Lemma 6, we obtain
Let \({\widetilde{s}}=1\). We have
Next, we deal with \(II_{2}\). Similar to \(I_{2}\), with the help of (3.1), we have
where in the last inequality we choose N large enough. For the case \(q_{0}>Q\), we have the result for \(R_{{\mathcal {L}}}\) in view of Lemma 8. \(\square \)
Remark 1
It is easy to check that if the critical ball \({\mathcal {Q}}\) is replaced by \(2{\mathcal {Q}}\), Lemma 13 also holds.
Lemma 14
Let \(V\in B_{Q/2}\) and \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\). Then for any \(s>p_{0}'\) and \(\gamma \ge 1\), there exists a constant C such that
holds for all f and \(g,h\in B=B(g_{0},r)\) with \(r<\gamma \rho (g_{0})\). Moreover, if \(q_{0}>Q\), the above estimate also holds for \(R_{{\mathcal {L}}}(\cdot ,\cdot )\) instead of \(R_{{\mathcal {L}}}^{*}(\cdot ,\cdot )\).
Proof
Let \({\mathcal {Q}}=B(g_{0},\gamma \rho (g_{0}))\). For \(g\in B(g_{0},r)\) with \(r<\gamma \rho (g_{0})\), \(z\in (2B)^{c}\), we can deduce that \(|gz^{-1}|\thickapprox |g_{0}z^{-1}|\thickapprox 2^{j}r\). Applying (5) and \(\rho (g)\thickapprox \rho (g_{0})\), we consider the following four parts:
To deal with \(I_{1}\), for all \(j\le j_{0}\) such that \(2^{j_{0}}\ge \gamma \rho (g_{0})/r\), by Lemma 6, we use Hölder’s inequality to obtain
For \(I_{2}\), using Lemma 6 and choosing \(N>\theta '\), we have
Next, for \(I_{3}\), we assume \(Q/2<q_{0}\le Q\). Then
Take \(p_{0}'<{\widetilde{s}}<s\), \(m=\frac{{\widetilde{s}}s}{s-{\widetilde{s}}}\) and q such that \(1/q=1/{\widetilde{s}}'+1/Q\). Using Lemma 6 and \(j\le j_{0}\), we have
For all \(j\le j_{0}\) such that \(2^{j_{0}}\ge \gamma \rho (g_{0})/r\), since \(Q/{\widetilde{s}}=Q+1-Q/q\) and \(2-Q/q>0\), then
Finally, to deal with \(I_{4}\), we get
For \(j>j_{0}\), we have
Noticing
we obtain
where we take N large enough such that \(N+Q+\delta -2-\theta '-Q\lambda >0\). Now, suppose \(q_{0}>Q\). In order to obtain the estimate of \(R_{{\mathcal {L}}}(\cdot ,\cdot )\), we apply Lemma 8 (ii) to get
This completes the proof of Lemma 14. \(\square \)
3.2 The \(L^{p}({\mathbb {H}}^{n})\) Boundedness of \([b,R_{{\mathcal {L}}}]\) and \([b,R_{{\mathcal {L}}}^{*}]\)
Theorem 1
Suppose that \(V\in B_{Q/2}\), \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\) and \(p_{0}\) such that \(1/p_{0}=(1/q_{0}-1/Q)^{+}\).
-
(i)
If \(1<p<p_{0}\), then \(\Vert [b,R_{{\mathcal {L}}}]f\Vert _{L^{p}({\mathbb {H}}^{n})}\lesssim [b]_{\theta }\Vert f\Vert _{L^{p}({\mathbb {H}}^{n})}\) for all \(f\in L^{p}({\mathbb {H}}^{n})\).
-
(ii)
If \(p_{0}'<p<\infty \), then \(\Vert [b,R_{{\mathcal {L}}}^{*}]f\Vert _{L^{p}({\mathbb {H}}^{n})}\lesssim [b]_{\theta }\Vert f\Vert _{L^{p}({\mathbb {H}}^{n})}\) for all \(f\in L^{p}({\mathbb {H}}^{n})\).
Proof
We only prove (ii), and (i) holds by duality. Because \(f\in L^{p}({\mathbb {H}}^{n})\) with \(p_{0}'<p<\infty \), according to Lemma 13, we have \([b,R_{{\mathcal {L}}}^{*}]f\in L_{loc}^{1}({\mathbb {H}}^{n})\). Applying Lemmas 12 and 13 with \(p_{0}'<p<\infty \) and Remark 1, we get
For the term \(\int _{{\mathbb {H}}^{n}}|M_{\rho ,\gamma }^{\sharp }([b,R_{{\mathcal {L}}}^{*}]f)(g)|^{p}dg\), we write \([b,R_{{\mathcal {L}}}^{*}]f(g):=B_{1}(g)-B_{2}(g)\), where
This implies that \(\frac{1}{|B|}\int _{B}|[b,R_{{\mathcal {L}}}^{*}]f(g)-([b,R_{{\mathcal {L}}}^{*}]f)_{B}|dg\lesssim I+II\), where
For I, let \(s>p_{0}'\). An application of Lemma 6 and Hölder’s inequality gives
For II, let \(g\in {\mathbb {H}}^{n}\) and \(B=B(g_{0},r)\) with \(r<\gamma \rho (g_{0})\) such that \(g\in B\). We split \(f=f_{1}+f_{2}\) with \(f_{1}=f\chi _{2B}\). We further divide II into two parts as \(II:=II_{1}+II_{2}\), where
For \(II_{1}\), let \(p_{0}'<{\widetilde{s}}<s\) and \(m=\frac{{\widetilde{s}}s}{s-{\widetilde{s}}}\). Then
For \(II_{2}\), it follows from Lemma 14 that
Finally, the \(L^{p}\)-boundedness of \(M_{s}\) implies that
This completes the proof of Theorem 1. \(\square \)
Theorem 2
Let \(V\in B_{Q/2}\) and \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\). Then
-
(i)
\([b,R_{{\mathcal {L}}}^{*}]: L^{\infty }({\mathbb {H}}^{n}) \mapsto BMO_{{\mathcal {L}}}({\mathbb {H}}^{n})\) if and only if \(b\in BMO_{\rho }^{\infty ,log}({\mathbb {H}}^{n})\).
-
(ii)
If \(V\in B_{Q}\), the above result is also true for \([b,R_{{\mathcal {L}}}]\).
Proof
We only prove (i), and the proof of (ii) is similar to that of (i). Let \(f\in L^{\infty }({\mathbb {H}})\) and \({\mathcal {Q}}=B(g_{0},\rho (g_{0}))\). Due to Lemma 13, we have
In order to deal with the oscillations, let \(B=B(g_{0},r)\) with \(r<\rho (g_{0})\). From Lemma 13, we know the function \([b,R^{*}_{{\mathcal {L}}}]f\) belongs to \(L_{loc}^{1}({\mathbb {H}}^{n})\).
We write \([b,R^{*}_{{\mathcal {L}}}]f:=I-II-III\), where
We have done the estimates of II and III in (3.2) and (3.3). Therefore, both terms are bounded by \([b]_{\theta }\left\| f\right\| _{\infty }\).
To deal with I, we fix \(h\in B\) and write
where \(f_{2}=f_{2,1}(u)+f_{2,2}(u)\) with \(f_{2,2}(u)=f\chi _{4{\mathcal {Q}}\setminus 2B}\) and \({\mathcal {Q}}=B(g_{0},\rho (g_{0}))\). Denote by \(I_{1}\), \(I_{2}\), \(I_{3}\) and \(I_{4}\) the each term oscillation.
Notice that \(R^{*}_{{\mathcal {L}}}f_{2,1}(u)\) and \(R^{*}_{{\mathcal {L}}}f_{2,2}(u)\) are finite for any \(u\in B\) since \(f\in L^{\infty }({\mathbb {H}}^{n})\) and
Next, we consider the boundedness of \(I_{1}\), \(I_{2}\) and \(I_{3}\) under the condition that \(b\in BMO^{\infty }_{\rho }({\mathbb {H}}^{n})\). For \(I_{1}\), selecting s such that \(R^{*}_{{\mathcal {L}}}\) is bounded on \(L^{s}({\mathbb {H}}^{n})\), we have
For the estimate of \(I_{2}\), we consider the term \(R^{*}_{{\mathcal {L}}}f_{2}(g)-R^{*}_{{\mathcal {L}}}f_{2}(u)\). Note that when \(V\in B_{Q}\), then \(|R^{*}_{{\mathcal {L}}}f_{2}(g)-R^{*}_{{\mathcal {L}}}f_{2}(u)|<\infty \). For \(V\in B_{q}\) with \(Q/2\le q<Q\), according to (5), we only need to estimate
Similar to \(I_{3}\) and \(I_{4}\) in Lemma 14, we have
and
which gives \(|R^{*}_{{\mathcal {L}}}f_{2}(g)-R^{*}_{{\mathcal {L}}}f_{2}(u)|\lesssim \left\| f\right\| _{\infty }\).
Applying Lemma 6 and Hölder’s inequality, we get
According to (3.4), we can get \(I_{3}\lesssim [b]_{\theta }\left\| f\right\| _{\infty }\) .
Finally, the theorem will follow if and only if there exists a constant \(C_{b}\) such that for any \(B\in {\mathcal {B}}_{\rho ,1}\) and \(u\in B\),
However, the kernel \(R^{*}(u,z)\) is added or subtracted, (3.5) is true, and only if the estimate (3.6) is true.
In fact, with the help of (6), we can deduce that \(\int _{4{\mathcal {Q}}}|R^{*}_{{\mathcal {L}}}(u,z)-R^{*}(u,z)|dz\) is bounded independently of the critical ball \({\mathcal {Q}}\) and
Assume \(V\in B_{q}\) for \(Q/2<q<Q\). Let \(1/q=1/s-1/Q\). We have
Notice, so far, we have only used \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\). Now, we assume that b satisfies the stronger condition \(b\in BMO_{\rho }^{\infty ,log}({\mathbb {H}}^{n})\), since \(|\int _{4{\mathcal {Q}}\setminus 2B}R^{*}(u,z)f(z)dz|\lesssim \left\| f\right\| _{\infty }\) and
we obtain the conclusion (3.6), thus proving the boundedness of the commutator \([b,R_{{\mathcal {L}}}^{*}]\).
On the other hand, we suppose that the commutator \([b,R_{{\mathcal {L}}}^{*}]\) is bounded with \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\), then (3.6) holds for each component \(R^{*}_{i}\), \(i=1,...,Q\) of \(R^{*}\) and for any \(f\in L^{\infty }({\mathbb {H}}^{n})\). Choosing \(f=sg(uz^{-1})\), and adding over i, inequality (3.6) implies
Since \(|zu^{-1}|\thickapprox |zg_{0}^{-1}|\), performing the integration, the inequality \(\frac{1}{|B|}\int _{B}|b(z)-b_{B}|dz\le \frac{C_{b}}{1+log(\rho (g_{0})/r)}\) holds for any \(B\in {\mathcal {B}}_{\rho ,1}\). Since \(b\in BMO_{\rho }^{\infty }({\mathbb {H}}^{n})\), we conclude that \(b\in BMO_{\rho }^{\infty ,log}({\mathbb {H}}^{n})\). This completes the proof of Theorem 2. \(\square \)
3.3 The \(L^{p}({\mathbb {H}}^{n})\) Boundedness of the Commutator \([b,T_{\beta _{1},\beta _{2}}]\)
Theorem 3
Suppose that \(V\in B_{q}\) for \(q>Q/2\). Let \(0\le \beta _{2}\le \beta _{1}<Q/2\). If \(1<(q/\beta _{2})'<p_{1}<Q/(2\beta _{1}-2\beta _{2})\) and \(1/p_{2}=1/p_{1}-(2\beta _{1}-2\beta _{2})/Q\), then
where \((q/\beta _{2})'\) is the conjugate of \(q/\beta _{2}\).
Proof
We only prove the following inequality:
Let \(r=\rho (g)\), \(B=B(g_{0},r)\). With the help of Lemma 7, we use Hölder’s inequality to deduce that
For \(k\ge 1\), because V(h)dh is a doubling measure, we have
Take N large enough. We can get
For \(k\le 0\), Lemma 2 implies that
Taking N large enough, we obtain
Finally, it holds
By (3.7) and the boundedness of the fractional maximal function, we obtain Theorem 3. \(\square \)
Theorem 4
Suppose \(V\in B_{q}\), \(q\ge Q/2\) and \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), \(0<\theta <\infty \). If \(0\le \beta _{2}\le \beta _{1}<Q/2\), \(q/(q-\beta _{2})<p_{1}<Q/(2\beta _{1}-2\beta _{2})\), \(1/p_{2}=1/p_{1}-(2\beta _{1}-2\beta _{2})/Q\), then
Proof
Because \(\beta _{2}\le \beta _{1}\), we can decompose the operator \((-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) as
Denote by \(L^{\beta _{2}-\beta _{1}}\) and \(T_{\beta _{2}}\) the operators \((-\Delta _{{\mathbb {H}}^{n}}+V)^{\beta _{2}-\beta _{1}}\) and \((-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{2}}V^{\beta _{2}}\), respectively. Then we can get
By Theorem 3, we obtain
This completes the proof. \(\square \)
4 Boundedness of the Commutators on Herz Spaces \({\dot{K}}_{q}^{\alpha ,p}({\mathbb {H}}^{n})\)
In this section, we consider the boundedness of the operator \(T_{\beta _{1},\beta _{2}}=(-\Delta _{{\mathbb {H}}^{n}}+V)^{-\beta _{1}}V^{\beta _{2}}\) and the commutator \([b,T_{\beta _{1},\beta _{2}}]\) on Herz spaces \({\dot{K}}_{q}^{\alpha ,p}({\mathbb {H}}^{n})\), where \(0\le \beta _{2}\le \beta _{1}<Q/2\).
Theorem 5
Suppose that \(V\in B_{q}\), \(q\ge Q/2\). Let \(0<p<\infty \), \(0\le \beta _{2}\le \beta _{1}<Q/2\). If \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta _{1}/Q)\) and \(q/(q-\beta _{2})<p_{1}<Q/(2\beta _{1}-2\beta _{2})\) with \(1/p_{2}=1/p_{1}-(2\beta _{1}-2\beta _{2})/Q\), then
Proof
We write \(f(h){=}\!\sum \limits _{j=-\infty }^{\infty }\! f(h)\chi _{E_{j}}(h){=}\!\sum \limits _{j=-\infty }^{\infty }\! f_{j}(h)\), which gives \(\Vert T_{\beta _{1},\beta _{2}}f\Vert _{{\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})}^{p} \lesssim M_{1}+M_{2}+M_{3}\), where
For \(M_{2}\), by Theorem 3, we obtain \(M_{2}\lesssim \sum \limits _{k=-\infty }^{\infty }2^{k\alpha p}\Big (\sum \limits _{j=k-1}^{k+1}\Vert f_{j}\Vert _{L^{p_{1}}({\mathbb {H}}^{n})}\Big )^{p} \lesssim \Vert f\Vert _{{\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})}^{p}.\)
Now we estimate \(M_{1}\). Via a simple computation, for \(g\in E_{k}\) and \(h\in E_{j}\) with \(j\le k-2\), we can deduce that \(|gh^{-1}|\thickapprox 2^{k}\). Applying Lemmas 7, 5, and 3, we use Hölder’s inequality to get
Take N large enough. For \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta _{1}/Q)\), we have
Below two cases are considered.
Case 1:\(p\le 1\). By the p-triangle inequality,
we have
Case 2:\(p> 1\). Using the Hölder inequality, we can get
For \(M_{3}\), because \(g\in E_{k}\), \(h\in E_{j}\), and \(j\ge k+2\), then \(|gh^{-1}|\thickapprox 2^{j}\). Similar to \(M_{1}\), we have
Because \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta _{2}/Q)\), we use (4.1) and Hölder’s inequality to deduce that
Finally, we get \(\Vert T_{\beta _{1},\beta _{2}}f\Vert _{{\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})}\lesssim \Vert f\Vert _{{\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})}\). \(\square \)
Theorem 6
Suppose that \(V\in B_{q}\), \(q\ge Q/2\) and \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), \(0<\theta <\infty \). Let \(0<p<\infty \), \(0\le \beta _{2}\le \beta _{1}<Q/2\). If \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta _{1}/Q)\) and \(q/(q-\beta _{2})<p_{1}<Q/(2\beta _{1}-2\beta _{2})\) with \(1/p_{2}=1/p_{1}-(2\beta _{1}-2\beta _{2})/Q\), then
Proof
We decompose f as follows.
which implies that \(\left\| [b,T_{\beta _{1},\beta _{2}}]f \right\| _{{\dot{K}}_{p_{2}}^{\alpha ,p}({\mathbb {H}}^{n})}^{p} \lesssim I+II+III\), where
Applying Theorem 4, we can get \(II\lesssim [b]_{\theta }^{p}\sum \limits _{k=-\infty }^{\infty }2^{k\alpha p} \Big (\sum \limits _{j=k-1}^{k+1}\Vert f_{j}\Vert _{L^{p_{1}}({\mathbb {H}}^{n})}\Big )^{p} \lesssim [b]_{\theta }^{p}\Vert f\Vert _{{\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})}^{p}\).
For I, we write \(b(g)-b(h):=(b(g)-b_{B})-(b(h)-b_{B})\). By Hölder’s inequality and the fact that \(V\in B_{q}\), we apply Lemmas 7 and 5 to deduce that
Now we consider the term \(\Big (\int _{E_{k}}\Big |\int _{E_{j}}(b(g)-b(h))V^{\beta _{2}}(h)f(h)dh\Big |^{p_{2}}dg\Big )^{1/p_{2}}\). We have
which implies that \(\Vert \chi _{k}[b,T_{\beta _{1},\beta _{2}}]f_{j}\Vert _{L^{p_{2}}({\mathbb {H}}^{n})}\lesssim I_{1}+I_{2}\), where
For \(I_{1}\), we use the Hölder inequality to deduce that
Note that \(\Big (\int _{E_{k}}|b(g)-b_{B}|^{p_{2}}dg\Big )^{1/p_{2}}\lesssim (k-j)[b]_{\theta }(1+2^{k}/\rho (0))^{\theta '}|E_{k}|^{1/p_{2}}\). Then, we have
Therefore, we can get
For the term \(I_{2}\), denote \(m=sp_{1}/(p_{1}(s-\beta _{2})-s)\). We use Hölder’s inequality to obtain
which gives
Take N large enough. Because \(-Q/p_{2}<\alpha <Q(-1/p_{2}+1-2\beta _{1}/Q)\), we obtain
We still divide the proof into two cases.
Case 1: \(p\le 1\). By (4.1), we have
Case 2: \(p>1\). The Hölder inequality implies that
For III, note that when \(g\in E_{k}\), \(y\in E_{j}\), and \(j\ge k+2\), then \(|gh^{-1}|\thickapprox 2^{j}r\). Using the following decomposition: \(b(g)-b(h):=(b(g)-b_{B})-(b(h)-b_{B})\) and Lemmas 7 and 5, we can get
Now we consider the term \(\Big (\int _{E_{k}}\Big |\int _{E_{j}}(b(g)-b(h))V^{\beta _{2}}(h)f(h)dh\Big |^{p_{2}}dg\Big )^{1/p_{2}}\). We have
which implies that \(\Vert \chi _{k}[b,T_{\beta _{1},\beta _{2}}]f_{j}\Vert _{L^{p_{2}}({\mathbb {H}}^{n})}\lesssim III_{1}+III_{2}\), where
For \(III_{1}\), we use the Hölder inequality to deduce that
Note that \(\Big (\int _{E_{k}}|b(g)-b_{B}|^{p_{2}}dg\Big )^{1/p_{2}}\lesssim (j-k)[b]_{\theta }(1+2^{k}/\rho (0))^{\theta '}|E_{k}|^{1/p_{2}}\). Then, we have
Therefore, we can get
For \(III_{2}\), denote \(m=sp_{1}/(p_{1}(s-\beta _{2})-s)\). The Hölder inequality implies that
which gives
Since N is large enough and \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta _{1}/Q)\), we obtain
Similar to I, it holds \(III\lesssim [b]_{\theta }^{p}\Vert f\Vert _{{\dot{K}}_{p_{1}}^{\alpha ,p}({\mathbb {H}}^{n})}^{p}\). Finally, we get
\(\square \)
Let \(\beta _{2}=0\). The following result is an immediate consequence of Theorem 6.
Corollary 1
Suppose that \(V\in B_{q}\), \(q\ge Q/2\) and \(b\in BMO_{\rho }^{\theta }({\mathbb {H}}^{n})\), \(0<\theta <\infty \). Let \(0<p<\infty \), \(0< \beta <Q/2\). If \(-Q/p_{2}<\alpha <Q(1-1/p_{2}-2\beta /Q)\) and \(1<p_{1}<Q/2\beta \) with \(1/p_{2}=1/p_{1}-2\beta /Q\), then
References
Bongioanni, B., Harboure, E., Salinas, O.: Commutators of Riesz transforms related to Schrödinger operators. J. Fourier Anal. Appl. 17, 115–134 (2011)
Duong, X., Yan, L.: On Commutators of fractional integrals. Proc. Amer. Math. Soc. 132, 3549–3557 (2004)
Guo, Z., Li, P., Peng, L.: \( L^{p}\) Boundedness of commutators of Riesz Transforms Associated to Schrödinger Operators. J. Math. Anal. Appl. 341, 421–432 (2008)
Jiang, Y.S.: Endpoint estimates for fractional integral associated to Schrödinger operators on the Heisenberg groups. Acta Math. Sci. Ser. B. 31, 993–1000 (2011)
Li, H.: Estimations Lp des opérateurs de Schrödinger sur les groupes nilpotents. J. Funct. Anal. 161, 152–218 (1999)
Li, P., Peng, L.: \(L^{p}\) boundedness of commutator operator associated with schrödinger operators on heisenberg group. Acta Math. Sci. 32, 568–578 (2012)
Li, P., Wan, X.: Schrödinger Operators on generalized Morrey spaces. J. Inequal. Appl. 1, 229 (2015)
Li, P., Wan, X.: The boundedness of commutators associtated with Schrödinger Operators on Herz spaces. J. Inequal. Appl. 2016, 172 (2016)
Lin, C., Liu, H., Liu, Y.: Hardy spaces associated with Schrödinger operators on the Heisenberg group. Preprint Available at arXiv:1106.4960. (2011)
Liu, Y.: The weighted estimates for the operators \(V^{\alpha }(\Delta _{G}+V)^{-\beta }\) and \(V^{\alpha }\nabla _{G}(-\Delta _{G}+V)^{-\beta }\) on the stratified Lie group G. J. Math. Anal. Appl. 349, 235–244 (2009)
Liu, Y., Huang, J., Dong, J.: Commutators of Calderón-Zygmund operators related to admissible functions on spaces of homogeneous type and applications to Schrödinger operators. Sci. China Math. 56, 1895–1913 (2013)
Liu, Y., Tang, G.: A note for Riesz transforms associated with Schrodinger operators on the Heisenberg group. Anal. Math. Phys. 7(1), 31–45 (2017)
Liu, Y., Tang, G.: Hardy type estimates for Riesz transforms associated with Schrodinger operators on the Heisenberg group. Anal. Theory Appl. 32(1), 78–89 (2016)
Lu, G.: A Fefferman-Phong type inequality for degenerate vector fields and applications. Panamer. Math. J. 6, 37–57 (1996)
Shen, Z.: \(L^{p}\) estimate for Schrödinger operator with certain potentials. Ann. Inst. Fourier. 45, 513–546 (1995)
Sugano, S.: Estimates for the operators \(V^{\alpha }(-\Delta +V)^{-\beta }\) and \(V^{\alpha }\nabla (-\Delta +V)^{-\beta }\) with certain nonnegative potentials V. Tokyo J. Math. 21, 441–452 (1998)
Tang, L., Dong, J.: Boundedness for some Schrödinger type operators on Morrey spaces related to certain nonnegative potentials. J. Math. Anal. Appl. 355, 101–109 (2009)
Tang, G., Liu, Y.: Hardy type estimates for Riesz transform associated with Schrödinger Operators on the Heisenberg group. Pure Math. 5, 291–297 (2015)
Yang, Da., Yang, Do., Zhou, Y.: Endpoint properties of localized Riesz transforms and fractional integrals associated to Schrödinger operators. Potential Anal. 30, 271–300 (2009)
Acknowledgements
The research was supported by Shandong Natural Science Foundation of China (No. ZR2020MA004) and National Natural Science Foundation of China (No. 11871293, No. 12071272).
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by V. Ravichandran.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Gao, C., Li, P. The Herz-Boundedness of Commutators Related to Schrödinger Operators in the Setting of Heisenberg Group. Bull. Malays. Math. Sci. Soc. 45, 1213–1240 (2022). https://doi.org/10.1007/s40840-022-01261-9
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40840-022-01261-9