A Lieb-Thirring inequality for extended anyons (Q2112249)

The authors prove a kinetic Lieb-Thirring inequality for \(N\) extended anyons. The fractional statistics is taken into account by considering the \(N\)-anyon wave function in the magnetic gauge picture, i.e., the anyons are treated as fermions coupled to magnetic flux tubes of finite radius \(R>0\). In this picture, the total kinetic energy is \[ \mathcal{E}_{\alpha}^{R}[\Psi_N] := \sum_{j=1}^N \int_{\mathbb{R}^{2N}}\left|\left(-i\nabla_{x_j} + \alpha \mathbf{A}^R(x_j)\right)\Psi_N\right|^2\,dx_1\dots\,dx_N, \quad \Psi_N \in \bigwedge_{j=1}^N L^2(\mathbb{R}^2) \] with \[ \mathbf{A}^R(x_j) := \sum_{k\neq j}^N \frac{(x_j-x_k)^\perp}{\max\{|x_j-x_k|,R\}^2} \] and \(\alpha\in[0,2]\) determining the statistics of the anyons. If \(\alpha=0\), the anyons are fermions, while if \(\alpha=1\), they are bosons. The main result of the paper under review (Theorem~1.2) states that for all \(\Psi_N\in\bigwedge_{j=1}^N L^2(\mathbb{R}^2)\) with associated one-particle density \[ \rho_{\Psi_N}(x) = \sum_{j=1}^N\int_{\mathbb{R}^2(N-1)}|\Psi_N(x_1,\dots,x_j=x,\dots,x_N)|^2\prod_{k\neq j}dx_k, \] there is \(C>0\) such that the ``anyonic but non-magnetic'' Lieb-Thirring inequality \[ \mathcal{E}_\alpha^R[\Psi_N] \geq C |\alpha-1| \int_{\mathbb{R}^2}\rho_{\Psi_N}(x)^2\,dx \] holds for all \(R,\alpha,N\). The proof follows the lines of [\textit{D. Lundholm} and \textit{J. P. Solovej}, Commun. Math. Phys. 322, No. 3, 883--908 (2013; Zbl 1270.81248); \textit{D. Lundholm} and \textit{J. P. Solovej}, Phys. Rev. A 88 (2013); \textit{D. Lundholm} and \textit{J. P. Solovej}, Ann. Henri Poincaré 15, No. 6, 1061--1107 (2014; Zbl 1294.81385)] -- that treated ideal anyons, i.e., \(R=0\) -- by covering \(\mathbb{R}^2\) -- using the Besicovitch covering lemma as in [\textit{P. T. Nam}, Acta Math Vietnam (2023), doi: 10.1007/s40306-022-00490-1] -- with boxes, proving a local Lieb-Thirring inequality on each box separately, and adding all contributions up. As in these works, the proof of the local Lieb-Thirring inequality (Theorem~2.1) follows by combining a local uncertainty principle, i.e., a Sobolev-type inequality (Lemma~2.4), and a local exclusion principle (Theorem~2.3 and Proposition~3.1). The latter is the main technical contribution of the paper and states that on a box with side length \(L\), the kinetic energy of a fixed \(N\)-independent number of anyons is bounded from below by a constant times \(L^{-2}\). (When \(R=0\), two anyons suffice to prove such a bound.) To prove the local exclusion principle, the authors distinguish between \(R\ll L\), \(R\sim L\), and \(R\gg L\) and proceed as follows. \begin{itemize} \item[(a)] \(R\ll L\). The anyons are close to being ideal and the proofs of [\textit{D. Lundholm} and \textit{R. Seiringer}, Lett. Math. Phys. 108, No. 11, 2523--2541 (2018; Zbl 1402.81269); \textit{S. Larson} and \textit{D. Lundholm}, Arch. Ration. Mech. Anal. 227, No. 1, 309--365 (2018; Zbl 1391.81217)] can be adapted. \item[(b)] \(R\sim L\). The cases where the anyons are not close to be fermionic (\(\alpha\geq c>0\)) and where they are close to be fermionic (\(\alpha\ll1\)) are treated separately. \begin{itemize} \item[(1)] If \(\alpha\geq c>0\), the methods of [\textit{S. Larson} and \textit{D. Lundholm}, Arch. Ration. Mech. Anal. 227, No. 1, 309--365 (2018; Zbl 1391.81217)] can be used. \item[(2)] If \(\alpha\ll1\), the magnetic field is treated perturbatively and the local exclusion principle for free fermions is used. \end{itemize} \item[(c)] \(R\gg L\). The magnetic flux covers the whole box, and the analysis reduces to a fermionic one-body problem on a box with an external magnetic field. As for magnetic Lieb-Thirring inequalities on \(\mathbb{R}^d\), the external magnetic field is not expected to raise the eigenvalues significantly. By diamagnetic considerations, the authors indeed prove estimates for the eigenvalue sums, which are independent of the external magnetic field. (The external magnetic field can be rather general because of the influence of particles outside the box). The authors remark that their estimates can alternatively be obtained by combining arguments in [\textit{R. L. Frank}, Contemp. Math. 500, 63--86 (2009; Zbl 1183.81068)] and [\textit{D. Hundertmark} and \textit{B. Simon}, J. Reine Angew. Math. 571, 107--130 (2004; Zbl 1068.47052)], as pointed out by the referees of their paper. \end{itemize}