An improved bound for the ground state energy of a Schrödinger operator on a loop

Helmut Linde · 2025

Read An improved bound for the ground state energy of a Schrödinger operator on a loop by Helmut Linde (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-2504-20229.txt.

Introduction

Helmut Linde sharpens a bound in quantum mechanics on a loop in a 2025 paper. The setup is a closed curve of fixed length with a Schrödinger operator whose potential is the square of the curve's bending. The lowest energy of that operator is the key value, and a conjecture by Benguria and Loss says it never drops below one.

The conjecture is still open. Linde proves a new floor of zero point eight one for the lowest energy. That mark beats the best past bound of about zero point six one by a wide step. Progress of this kind narrows the gap toward the conjectured value and tests the methods used to bound quantum energies on curved paths.

Catalog record

CREATORHelmut Linde
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-20229

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

An improved bound for the ground state energy of a Schrödinger operator on a loop Helmut Linde Abstract Consider a closed curve of length 2 ​ π 2\pi with curvature κ ⁡ ( s ) \kappa(s) and the Schrödinger operator H H with κ 2 \kappa^{2} as the potential term. Let λ Γ \lambda_{\Gamma} be the lowest eigenvalue of H H . The Ovals Conjecture proposed by Benguria and Loss states that λ Γ ≥ 1 \lambda_{\Gamma}\geq 1 . While the conjecture remains open, the present work establishes a new lower bound of 0.81 0.81 on λ Γ \lambda_{\Gamma} , improving on the best previous estimate of approximately 0.6085 0.6085 . 1 Introduction Let \Gamma be a closed curve of length 2\pi in \mathbb{R}^{n} with the curvature \kappa(s) as a function of the arc length. Let H_{\Gamma} be the Schrödinger operator H_{\Gamma}=-\Delta+\kappa^{2}(s)\quad\mathrm{in}\quad L^{2}([0,2\pi)) (1) with periodic boundary conditions and \lambda_{\Gamma} its lowest eigenvalue. The topic of the present article is the ‘Ovals Conjecture’ which states that this eigenvalue cannot be smaller than one. The question about the minimal possible value of \lambda_{\Gamma} has attracted attention since the year 2004, when Benguria and Loss established a connection between this problem and the Lieb-Thirring conjecture in one dimension [ 1 ] . They also conjectured that \lambda_{\Gamma}\geq 1 for any \Gamma . The class {\cal F} of putative minimizers of \lambda_{\Gamma} contains the circle and certain point-symmetric oval loops. It can be extended to the edge case of a straight segment of length \pi , traversed back and forth, that becomes admissible in a more generalized setting. For all curves in {\cal F} the equality \lambda_{\Gamma}=1 holds, but it has not yet been proved that this is actually the smallest possible value of \lambda_{\Gamma} . Several other authors have made contributions towards proving the conjecture of Benguria and Loss: Burchard and Thomas have demonstrated [ 2 ] that the curves in {\cal F} minimize \lambda_{\Gamma} locally, i. e., there is no small variation around these curves that reduces \lambda_{\Gamma} . Denzler [ 3 ] has shown that closed curves minimizing \lambda_{\Gamma} exist and that these minimizers are planar, smooth, strictly convex curves. Throughout this paper we can therefore assume that \Gamma has these properties without losing generality. The highest global bound on the eigenvalue so far was established in the year 2006, showing that \lambda_{\Gamma}>0.6085 for all possible curves that meet the specifications of the conjecture [ 4 ] . In the present article, that bound on \lambda_{\Gamma} is improved to 0.81 . 2 Result statement The main result of this article is: Theorem 2.1 . Let \Gamma be a smooth, strictly convex, closed curve of length 2\pi in the plane with curvature \kappa(s) and H_{\Gamma} the Schrödinger operator ( 1 ). Then its lowest eigenvalue \lambda_{\Gamma} satisfies the bound \lambda_{\Gamma}>\inf_{\begin{subarray}{c}\tilde{\nu}\in[0,1]\\ \Delta\in(0,\frac{\pi}{2})\end{subarray}}\max(B_{1}(\tilde{\nu},\Delta),B_{2}(\tilde{\nu},\Delta)), (2) where \displaystyle B_{1}(\tilde{\nu},\Delta) \displaystyle=\left(1+2\tilde{\nu}G(\Delta)\right)^{-2}, (3) \displaystyle B_{2}(\tilde{\nu},\Delta) \displaystyle=1-\left(2-\sec^{2}\frac{\Delta}{2}\right)\left(1-\left(2+2\tilde{\nu}G(\Delta)-\tilde{\nu}\right)^{-2}\right) (4) with G(\Delta):=\left(1+\cot\frac{\Delta}{4}\right)^{-2}. (5) Despite its involved form, the argument on the r.h.s. of ( 2 ) ultimately represents a combination of elementary functions on the rectangular domain [0,1]\times(0,\frac{\pi}{2}) . It is straightforward to evaluate the expression ( 2 ) with a computer and the result is \lambda_{\Gamma}>0.8246\text{ (numerically).} (6) Also see Figure 1 for a visual representation of B_{1} and B_{2} . Figure 1: Superposition of contour plots of B_{1} (solid lines) and B_{2} (dashed lines). The X marks the point where \max(B_{1},B_{2}) is minimal. Alternatively, one can derive analytical bounds on \lambda_{\Gamma} from Theorem 2.1 to avoid the use of a computer. The function B_{1}(\tilde{\nu},\Delta) decreases strictly monotonously with \tilde{\nu} and with \Delta . Therefore, a crude but simple bound on \lambda_{\Gamma} can be given based on B_{1} alone: \displaystyle\lambda_{\Gamma} \displaystyle>\inf_{\begin{subarray}{c}\tilde{\nu}\in[0,1]\\ \Delta\in(0,\frac{\pi}{2})\end{subarray}}B_{1}(\tilde{\nu},\Delta)=\lim_{\Delta\rightarrow\frac{\pi}{2}}B_{1}(1,\Delta) \displaystyle=\frac{3+2\sqrt{2}}{8}\approx 0.7285 To obtain better estimates leveraging B_{1} and B_{2} jointly, there is a trade-off to be made between accuracy and simplicity of the analytical treatment. The following theorem and its proof are one example for such a trade-off: Theorem 2.2 . The inequality ( 2 ) implies the weaker bound \lambda_{\Gamma}>0.81 (7) The rest of the article is devoted to proving Theorem 2.1 and Theorem 2.2 : In Section 3 , some groundwork is laid by summarizing results from previous work, introducing the notation, and proving several lemmata. The main part of the proof of Theorem 2.1 is given in Section 4 and this is where new methods of treating the Ovals Conjecture are developed. Section 5 contains the purely technical proof of Theorem 2.2 , which is given here for completeness and to obtain the clean result ( 7 ), but which adds very little to the understanding of the Ovals Conjecture as such. Finally, Section 6 fills a gap in the proof of Theorem 2.1 by establishing an important lemma where lengthy and, again, rather technical work is required. 3 Preliminaries 3.1 Notation and closure condition To prepare the proof of Theorem 2.1 , this section introduces additional notation and collects some basic facts which have already been established in [ 4 ] . For a given smooth curve \Gamma as in Theorem 2.1 with an arc length parameter s let \phi(s) be the angle between the tangent on \Gamma in s and some fixed axis, which implies \phi^{\prime}(s)=\kappa(s) . To keep the notation compact we write \phi:\Omega\rightarrow\Omega\quad\mathrm{with}\quad\Omega:=\mathbb{R}/\penalty 2\pi\mathbb{Z}, considering numbers that differ by an integer multiple of 2\pi as identical. Since the problem is invariant under a reflection of \Gamma in the plane one can assume that \phi^{\prime}>0 without loss of generality. For convenience, this and other assumptions made on \Gamma throughout the article are summarized in Table 1 . The assumption that \Gamma is closed implies \displaystyle\int_{\Omega}\cos\phi(s)\,\mathrm{d}s=\int_{\Omega}\sin\phi(s)\,\mathrm{d}s=0, (8) \displaystyle\int_{\Omega}\phi^{\prime}(s)\,\mathrm{d}s=2\pi. The eigenvalue \lambda_{\Gamma} is the minimal value - for given \phi - of the energy functional E(\phi,\psi):=\frac{\int_{\Omega}(\psi^{\prime}(s)^{2}+\kappa(s)^{2}\psi(s)^{2})\,\mathrm{d}s}{\int_{\Omega}\psi(s)^{2}\,\mathrm{d}s} (9) It turns out that the curve \Gamma and the lowest eigenfunction of the Schrödinger operator H_{\Gamma} can be characterized in two more ways, which will be called the \phi^{-1} view and the x / y coordinates view . The views are equivalent but in each of them some results are more accessible than in the other. One key idea for proving Theorem 2.1 is to gather information from the different views and put them into contrast. Each view requires its own notation, which is introduced in the following two subsections. 3.2 The \phi^{-1} view The closure condition on \Gamma can be expressed in a more convenient way in terms of the inverse function \phi^{-1}:\Omega\rightarrow\Omega , where the unpleasant integrals over \cos\phi(s) and \sin\phi(s) in ( 8 ) turn into the simple demand that the first non-trivial Fourier component of its derivative be zero: \displaystyle\int_{\Omega}(\phi^{-1})^{\prime}(t)\sin t\,\mathrm{d}t=\int_{\Omega}(\phi^{-1})^{\prime}(t)\cos t\,\mathrm{d}t=0, (10) \displaystyle\int_{\Omega}(\phi^{-1})^{\prime}(t)\,\mathrm{d}t=2\pi. The function (\phi^{-1})^{\prime} can therefore be written as a Fourier series (\phi^{-1})^{\prime}(t)=1+\sum\limits_{n=2}^{\infty}na_{n}\cos nt-nb_{n}\sin nt, (11) such that \phi^{-1}(t)=C+t+g(t)+f(t), (12) where C\in\mathbb{R} is some integration constant and \displaystyle g(t) \displaystyle:= \displaystyle\sum\limits_{n=2,4,6,...}^{\infty}a_{n}\sin nt+b_{n}\cos nt, (13) \displaystyle f(t) \displaystyle:= \displaystyle\sum\limits_{n=3,5,7,...}^{\infty}a_{n}\sin nt+b_{n}\cos nt. (14) Note the periodicity properties f(t+\pi)=-f(t),\quad g(t+\pi)=g(t)\quad\textmd{for all}\quad t\in\Omega. (15) The functions f and g are continuous, since a discontinuity in either of them at some point t would lead to at least one discontinuity of \phi^{-1} in t or in t+\pi due to the periodicity properties ( 15 ). 3.3 Estimates on the total variation of f

Read the complete text