Plancherel Identities for unbounded subsets of ℝ d \mathbb{R}^{d}
Piyali Chakraborty, Dorin Ervin Dutkay · 2026
Read Plancherel Identities for unbounded subsets of ℝ d \mathbb{R}^{d} by Piyali Chakraborty, Dorin Ervin Dutkay (2026) 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-2606-27008.txt.
Introduction
Chakraborty and Dutkay study when the Fourier transform keeps lengths exactly on regions that stretch without bound. The Fourier transform rewrites a signal as a mix of wave patterns. Plancherel's identity says total energy is the same before and after that rewrite. For bounded regions this balance can fail, so pairs of special regions are needed. The authors describe a class of such pairs in flat space of any dimension. Their regions repeat under shifts from a full-rank lattice, a regular grid of points.
Restricted to those paired regions, the transform becomes an isometric isomorphism, meaning a perfect length-keeping match between two function spaces. That gives the Plancherel identity directly on unbounded sets, where naive energy sums would diverge. They also give a clean test for when energy balance survives on infinite regions.
Catalog record
| CREATOR | Piyali Chakraborty, Dorin Ervin Dutkay |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 57 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.27008
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Plancherel Identities for unbounded subsets of ℝ d \mathbb{R}^{d} Piyali Chakraborty, Dorin Ervin Dutkay Abstract . We present a class of pairs of subsets of ℝ d \mathbb{R}^{d} for which the Fourier transform, when restricted to these subsets, is an isometric isomorphism, and thus the Plancherel identity is satisfied. The sets are invariant under translations by dual full-rank lattices. 2010 Mathematics Subject Classification 47F05,47A10,42A38,52C22 1. Introduction Fuglede’s conjecture originated in a question posed by Irving Segal concerning the partial differential operators \frac{1}{2\pi i}\frac{\partial}{\partial x_{j}} defined on an open set \Omega\subset\mathbb{R}^{d} . The question asks: under what conditions do these operators admit commuting self-adjoint extensions on L^{2}(\Omega) ? In [ 4 ] , Fuglede proved that for finite-measure domains \Omega with some regularity restrictions, there are such extensions if and only if there are orthogonal bases of exponential functions \{e_{\lambda}:\lambda\in\Lambda\} for L^{2}(\Omega) , where e_{\lambda}(x)=e^{2\pi i\lambda\cdot x}. He also proved that even for some nice domains like the disk or a triangle, this condition is not satisfied. It is satisfied, of course, by the unit cube, with the classical Fourier series. Definition 1.1 . A measurable set \Omega\subset\mathbb{R}^{d} of finite Lebesgue measure is called spectral if there exists a set \Lambda\subset\mathbb{R}^{d} such that the family of exponential functions \{e_{\lambda}:\lambda\in\Lambda\} forms an orthogonal basis for L^{2}(\Omega) , e_{\lambda}(x)=e^{2\pi i\lambda\cdot x} . In this case, \Lambda is called a spectrum for \Omega . We say that \Omega tiles \mathbb{R}^{d} by translations if there exists a set \mathcal{T}\subset\mathbb{R}^{d} such that the translations \{\Omega+t:t\in\mathcal{T}\} form a partition of \mathbb{R}^{d} up to measure zero. The set \mathcal{T} is then called a tiling set for \Omega , and we say that \Omega tiles \mathbb{R}^{d} with the set \mathcal{T} . If only the condition that the sets \Omega+t , t\in\mathcal{T} are mutually disjoint (up to measure zero) is satisfied, (so the union \cup_{t}(\Omega+t) is not necessarily the whole space \mathbb{R}^{d} ), then we say that \Omega packs with \mathcal{T} . Fuglede then formulated his conjecture: Conjecture 1.2 . [Fuglede’s Conjecture] A measurable subset \Omega\subset\mathbb{R}^{d} of finite Lebesgue measure is spectral if and only if it tiles \mathbb{R}^{d} by translations. In the same paper, Fuglede proved that his conjecture is true when the spectrum or the tiling set is a lattice in \mathbb{R}^{d} . Theorem 1.3 . Let \Omega be a measurable subset of \mathbb{R}^{d} of finite measure and let A be some invertible d\times d matrix. Then \Omega tiles \mathbb{R}^{d} with the lattice A\mathbb{Z}^{d} if and only if \Omega has spectrum the dual lattice (A^{T})^{-1}\mathbb{Z}^{d} . The conjecture was later shown to be false in dimensions d\geq 3 [ 13 , 3 ] , but it is true for convex domains [ 9 ] . In [ 11 ] , Steen Pedersen extended Fuglede’s result by removing both the regularity condition and the finite-measure assumption. Naturally, the definition of a spectral set had to be adapted for infinite-measure domains (since e_{\lambda} is not square-integrable in that case). Definition 1.4 . For a function f\in L^{1}(\mathbb{R}^{d}) , define its classical Fourier transform as \hat{f}(t)=\int_{\mathbb{R}^{d}}f(x)e^{-2\pi it\cdot x}\,dx,\quad(t\in\mathbb{R}^{d}). Let \Omega\subset\mathbb{R}^{d} be measurable and let \mu be a positive Radon measure on \mathbb{R}^{d} . We say that (\Omega,\mu) is a spectral pair if: (1) for each f\in L^{1}(\Omega)\cap L^{2}(\Omega) , the continuous function t\mapsto\hat{f}(t) satisfies \int|\hat{f}|^{2}\,d\mu<\infty ; and (2) the mapping f\mapsto\hat{f} from L^{1}(\Omega)\cap L^{2}(\Omega)\subset L^{2}(\Omega) into L^{2}(\mu) is isometric and has dense range. This map then extends by continuity to an isometric isomorphism \mathscr{F}:L^{2}(\Omega)\to L^{2}(\mu). The isometry property yields a Plancherel identity obtained by taking the Fourier transform of functions in L^{2}(\Omega) , restricting it to the support of the measure \mu and integrating the absolute value squared of this restriction of the Fourier transform against the measure \mu . The set \Omega is called spectral if there exists a measure \mu such that (\Omega,\mu) forms a spectral pair; \mu is then called a pair measure or a dual measure for \Omega . When \Omega has finite measure, this definition coincides with the earlier one [ 11 , Corollary 1.11] , and the pair measure \mu is the counting measure on the spectrum \Lambda . Denote by \textup{Leb}_{\Omega} the Lebesgue measure restricted to a subset \Omega of \mathbb{R}^{d} . In [ 1 ] , we presented a class of unbounded sets in \mathbb{R} that are spectral: Theorem 1.5 . [ 1 , Theorem 1.8] Let \Omega be an open subset of \mathbb{R} and let a\in\mathbb{N} , a\geq 2 . Then \Omega tiles \mathbb{R} by \{0,1,\dots,a-1\} if and only if \Omega is a spectral set with pair measure \mu equal to a renormalized Lebesgue measure on \left[-\frac{1}{2a},\frac{1}{2a}\right]+\mathbb{Z} , \mu=a\textup{Leb}_{\left[-\frac{1}{2a},\frac{1}{2a}\right]+\mathbb{Z}} . In this paper we extend this theorem to higher dimensions. Our main results are Theorem 1.6 . Let \Omega_{0} be a bounded open set in \mathbb{R}^{d} of finite measure, let \Lambda_{0} be a discrete subset of \mathbb{R}^{d} , and let A be some invertible d\times d matrix. Assume in addition that (1.1) The set \Omega_{0} packs with the lattice A\mathbb{Z}^{d} , and (1.2) \lambda_{0}\cdot Ak\in\mathbb{Z} for all \lambda_{0}\in\Lambda_{0} and k\in\mathbb{Z}^{d} , equivalently \Lambda_{0}\subseteq(A^{T})^{-1}\mathbb{Z}^{d} . The following statements are equivalent: (i) \Omega_{0} has spectrum \Lambda_{0} . (ii) The set \Omega=\Omega_{0}+A\mathbb{Z}^{d} is spectral with pair measure \mu:=\frac{|\det A|}{|\Omega_{0}|}\textup{Leb}_{(A^{T})^{-1}[-\frac{1}{2},\frac{1}{2}]^{d}+\Lambda_{0}}. (iii) For the set \Omega=\Omega_{0}+A\mathbb{Z}^{d} , the restriction of the Fourier transform L^{2}(\Omega)\ni f\to\hat{f}|_{(A^{T})^{-1}[-\frac{1}{2},\frac{1}{2}]^{d}+\Lambda_{0}}\in L^{2}(\mu) is an isometry. As a fairly easy corollary, we obtain a generalization of Theorem 1.5 : Theorem 1.7 . Let \Omega_{0} be an open subset of \mathbb{R}^{d} , let A be an invertible d\times d integer matrix, and let \mathcal{R} be a complete set of representatives \operatorname{mod}A\mathbb{Z}^{d} . Assume in addition that the sets \Omega_{0} packs with the lattice A\mathbb{Z}^{d} , and define the set \Omega:=\Omega_{0}+A\mathbb{Z}^{d}. The following statements are equivalent: (i) \Omega tiles \mathbb{R}^{d} with \mathcal{R} . (ii) \Omega_{0} tiles \mathbb{R}^{d} with the lattice \mathbb{Z}^{d} and hence |\Omega_{0}|=1 . (iii) \Omega is spectral with pair measure \mu:=\frac{|\det A|}{|\Omega_{0}|}\textup{Leb}_{(A^{T})^{-1}[-\frac{1}{2},\frac{1}{2}]^{d}+\mathbb{Z}^{d}}. (iv) The Fourier transform f\to\hat{f} , from L^{1}(\Omega)\cap L^{2}(\Omega) into L^{2}(\mu) (with \mu as above) extends to an isometry from L^{2}(\Omega) into L^{2}(\mu) . Note that our sets \Omega are invariant under translations by elements from the lattice A\mathbb{Z}^{d} ; in other words, any element of the lattice A\mathbb{Z}^{d} is a period for \Omega (see Definition 2.2 ). In section 4 , we show that, in dimension 1, if a set \Omega tiles with a finite set, then \Omega has to be periodic (Theorem 4.1 ). If \Omega tiles \mathbb{R} by \{0,1,\dots,a-1\} for some integer a\geq 2 , then a is a period for \Omega (see the Proof of Theorem 1.5 ). Also, if the cardinality of the tiling set \mathcal{T} is a prime number p , and the greatest common divisor of the elements of \mathcal{T} is 1, then \Omega has period p (Theorem 4.3 ). The paper is structured as follows: in Section 2 , we introduce some definitions, notations, and some lemmas needed for the proofs of our main results; see in particular Lemma 2.5 and Lemma 2.8 , which might be of independent interest. In Section 3 , we present the proofs of our main results. In Section 4, we show that, in dimension 1, the subsets of \mathbb{R} that tile with a finite set have periodicity properties.