Completely bounded \(\Lambda_p\) sets that are not Sidon (Q2802107)

Let \(G\) be a compact abelian group (\(\Gamma\) its dual group and \(\lambda\) the Haar measure) and \(p>2\). A set \(E\subset\Gamma\) is a \(\Lambda(p)\) set if the \(L^2\) norm and the \(L^p\) norm are equivalent on the subspace of trigonometric polynomials spanned by \(E\). A set \(E\subset\Gamma\) is a Sidon set if every continuous function with Fourier transform supported by a subset of \(E\) has a summable Fourier transform. It is known since the seminal paper of \textit{W. Rudin} [J. Math. Mech. 9, 203--227 (1960; Zbl 0091.05802)] that a Sidon set is \(\Lambda(p)\) for every \(p>2\). Conversely, there exist some sets \(E\) which are \(\Lambda(p)\) for every \(p>2\), but not Sidon ([\textit{A. Bonami}, Ann. Inst. Fourier 18, No. 2, 193--204 (1968; Zbl 0175.44801)], [\textit{R. E. Edwards} et al., Indiana Univ. Math. J. 21, 787--806 (1972; Zbl 0221.43007)]).NEWLINENEWLINEThe following notion was introduced in the 90's by \textit{A. Harcharras} [Stud. Math. 137, No. 3, 203--260 (1999; Zbl 0948.43002)]: \(E\subset\Gamma\) is a \(\Lambda^{cb}(p)\) set if, for every \({\mathcal S}^p\)-valued polynomials (with Fourier coefficients supported by a subset of \(E\)), the \(L^p({\mathcal S}^p)\)-norm of \(f\) is controlled by the maximum between the \({\mathcal S}^p\)-norm of \(\displaystyle\Big(\sum\hat f(\gamma)^\ast\hat f(\gamma)\Big)^{1/2}\) and the \({\mathcal S}^p\)-norm of \(\displaystyle\Big(\sum\hat f(\gamma)\hat f(\gamma)^\ast\Big)^{1/2}\) (be aware of a misprint in the definition in the article) , where \({\mathcal S}^p\) is the Schatten \(p\)-class of operators.NEWLINENEWLINEOne can prove that \(\Lambda^{cb}(p)\) sets are \(\Lambda(p)\) sets and the converse is false (see [Harcharras, loc. cit.]). It was shown in [\textit{W. D. Banks} and \textit{A. Harcharras}, Ill. J. Math. 47, No. 4, 1063--1078 (2003; Zbl 1035.43004)] that there exist sets \(E\) which are \(\Lambda^{cb}(p)\) sets for every \(p>2\), but not Sidon, when \(\Gamma={\mathbb Z}\).NEWLINENEWLINEIn the paper under review, this result is extended to any compact abelian (infinite) group \(G\) (see Theorem 4.2). The particular case of \(p=4\) is proved in Section 3 (Theorem 3.5.). The proofs rely exclusively on (nontrivial) combinatorial arguments, and make use of the condition of type \(Z(p)\) of Harcharras [loc. cit.].