Characterisation of Meyer sets via the Freiman-Ruzsa theorem (Q6093380)

The discovery of quasicrystals [\textit{D. Shechtman} et al., Phys. Rev. Lett. 53, No. 20, 1951--1953 (1984; \url{doi:10.1103/PhysRevLett.53.1951})] led to considerable interest in non-periodic, discrete subsets of \(\mathbb R^d\) which have approximate symmetry. Meyer sets, originally introduced in [\textit{Y. Meyer}, Nombres de Pisot, nombres de Salem et analyse harmonique. Berlin-Heidelberg-New York: Springer-Verlag (1970; Zbl 0189.14301); [\textit{Y. Meyer}, Algebraic numbers and harmonic analysis. Amsterdam-London: North- Holland Publishing Company (1972; Zbl 0267.43001)], are one of the fundamental concepts. For the definition, it is required the following definitions. Let \(A \subset \mathbb R^d\). The upper uniform asymptotic density and the lower uniform asymptotic density of \(A\) are respectively given by \begin{align*} D^+(A) &= \limsup_{R \to \infty} \sup_{x \in \mathbb R^d} \frac{|\{a \in A; \lVert a-x \rVert_\infty < R\}|}{(2R)^d}, \quad \text{and} \\ D^-(A) &= \liminf_{R \to \infty} \inf_{x \in \mathbb R^d} \frac{|\{a \in A; \lVert a-x \rVert_\infty < R\}|}{(2R)^d}. \end{align*} The difference set of \(A\) is \(A - A = \{a - b; a,b \in A\}\). \(A\) is relatively dense if it intersects each ball of radius \(R\) for some \(R>0\), and it is uniformly discrete if for some \(r>0\) each ball with radius \(r\) contains at most one point in \(A\). A cut-and-project set takes the form \(\pi_1(\Gamma \cap (\mathbb R^d \times \Omega))\) for some lattice \(\Gamma \subset \mathbb R^{d+e}\), some bounded open set \(\Omega \subset \mathbb R^e\) and the standard projection \(\pi_1: \mathbb R^{d+e} \to \mathbb R^d\). Meyer sets are sets satisfying the equivalent conditions of the following theorem. Theorem 1.1. Let \(d \ge 1\) and let \(A \subset \mathbb R^d\) be a relatively dense set. Then the following are equivalent: (1) \(D^+(A-A) < \infty\); (2) \(A-A\) is uniformly discrete; (3) \(A\) is uniformly discrete and \(A-A \subset A+F\) for a finite set \(F\); (4) \(A \subset M+F\) for a cut-and-project set \(M\) and a finite set \(F\). It is easy to verify the chain of implications (4) \(\Rightarrow\) (3) \(\Rightarrow\) (2) \(\Rightarrow\) (1), and the hardest part is that (1) \(\Rightarrow\) (4). Theorem 1.1 has a great similarity with the Freiman-Ruzsa theorem [\textit{G. A. Freiman}, Foundations of a structural theory of set addition. Translation from Russian. Providence, RI: American Mathematical Society (AMS) (1973; Zbl 0271.10044); \textit{I. Z. Ruzsa}, Sci., Ser. A 3, 97--109 (1989; Zbl 0743.05052)], which is very important in additive combinatorics. Before stating such theorem, we need the following definition. A generalised arithmetic progression of rank \(d\) is a set of the form \(P_1+\dots+P_d\), where \(P_i\) are arithmetic progressions. Theorem 1.2 (Freimann-Ruzsa). Let \(A\) be a finite, non-empty subset of an abelian torsion-free group. Consider the conditions: (i) \(_K\) \(|A+A| < K|A|\) for some \(K \ge 2\); (ii) \(_{d,C}\) \(A\) is contained in a generalised arithmetic progression \(P\) of rank \(d\) and size \(\le C|A|\), for some constants \(d \in \mathbb N\) and \(C\). Then (i) and (ii) are equivalent in the following sense: For each \(K\) there exist \(d = d(K)\) and \(C = C(K)\) such that (i)\(_K\) implies (ii)\(_{d,C}\). Conversely, for each \(C,d\) there exists \(K = K(C,d)\) such that (ii)\(_{d,C}\) implies (i)\(_K\). Despite the similarity between Theorems 1.1 and 1.2, their proofs use rather different techniques. Hence, it is natural to inquire if such similarity is only superficial or if there is a more tangible connection. The purpose of the paper is to exhibit such a connection by showing that Theorem 1.2 together with some basic tools implies Theorem 1.1 in a relatively elementary way. Additionally, the authors point out that quantitative variants of Theorem 1.2 have been extensively investigated by many authors. On the other hand, quantitative bounds in Theorem 1.1 are rarely addressed. The argument of the authors preserves quantitative bounds, and as a consequence they obtain the following new result. Theorem 1.3. Suppose that \(A \subset \mathbb R^d\) is relatively dense and \(D^+(A-A) < KD^-(A)\) for some \(K \ge 2\). Then \(A \subset M+F\), where \(F\) is a finite set and \(M\) is a cut-and-project set with at most \(O\left( (d+\log K)^{6+o(1)} \right)\) internal dimensions. Further improvements in the exponent \(6+o(1)\) would follow from stronger quantitative versions of the Freiman-Ruzsa theorem. In particular, it is plausible that \(O(d+\log K)\) internal dimensions suffice.