Deceptions in quasicrystal growth (Q1894835)

The authors consider non-periodic tilings in \(\mathbb{R}^m\) (i.e. which admit no nontrivial translational symmetry) by a given (aperiodic) protoset of tiles \(S\). In such a tiling a regular patch of order \(r\) is defined as a sufficiently large (covering a ball of radius at least \(r\) if the tiles are small), finite set of non-overlapping tiles homeomorphic to the closed \(m\)-ball. A deception of order \(r\) is a regular patch of order \(r\) which is not a subset of any tiling of the space and such that every connected subpatch of cardinality less than \(r\) is a subset of some tiling of the space. The following two results are proved: \(\bullet\) If there exists a tiling in \(\mathbb{R}^3\) based on an aperiodic prototile set \(S\) which admits only finitely many orientations of the prototiles of \(S\), then there are tilings based on \(S\) with deceptions of all orders. \(\bullet\) For a given aperiodic set \(S\) of prototiles in the plane \(\mathbb{R}^2\) there exist tilings with deceptions of all orders.