Some combinatorially defined subsets of \(\beta \mathbb N\) and their relation to the idempotents (Q618294)

The authors introduce and study methodically a couple of subsets and subsemigroups of \(\beta\mathbb{N}\) arising in attempts to solve or around ``a favorite problem of I.~Leader'': does there exist a set \(A\subseteq\mathbb{N}\) such that whenever \(r\in\mathbb{N}\) and \(A=\bigcup_{i=1}^r C_i\), there must exist \(i\in\{1,2,\dots,r\}\) and \(\langle x_t \rangle_{t=1}^\infty\) such that \(FS_{\leqslant 2}(\langle x_t \rangle_{t=1}^\infty)\subseteq C_i\) but there does not exist \(\langle y_t \rangle_{t=1}^\infty\) with \(FS(\langle y_t \rangle_{t=1}^\infty)\subseteq A\)? Here, \(FS(\langle y_t \rangle_{t=1}^\infty)\) is a set of all finite sums of distinct members of \(\langle y_t \rangle_{t=1}^\infty\), \(FS_{\leqslant 2}(\langle x_t \rangle_{t=1}^\infty)=\{x_i+x_j: 1\leqslant i < j < \infty\}\). This question remains open but has an affirmative answer (Theorem 1.11) if we take, instead of \(\mathbb{N}\), the direct sum of countably many copies of \(\mathbb{Z}_2\).