Elastic Lattices Inspired by Ulam-Warburton Cellular Automaton

Periodic lattices have been widely explored for decades, owing to their peculiar vibrational behavior. On the other hand, certain types of aperiodic lattices have enabled new phenomena that may not be otherwise attainable in periodic ones. In this paper, a new class of aperiodic lattices inspired by cellular automaton is introduced. Cellular automata were originally developed as a machine replication algorithm and it has been intensively explored in computer science. These algorithms yield structures that are not necessarily periodic, yet follow well-defined rules that lead to interesting patterns. The concept is utilized here to build elastic lattices following such rules, and Ulam-Warburton Cellular Automaton (UWCA) is demonstrated as an example. Starting from a square monatomic lattice, an UWCA lattice is constructed and its vibrational behavior is analyzed, showing unique dynamical properties, including symmetric eigenfrequency spectra, repeated natural frequencies of large multiplicity, and the emergence of strongly localized corner modes. It is envisioned that computer-algorithm-inspired lattices may unlock new wave phenomena that could outperform existing lattice designs.

💡 Research Summary

The manuscript introduces a novel class of aperiodic elastic lattices derived from the Ulam‑Warburton cellular automaton (UWCA). Starting from an infinite square monatomic lattice whose masses are initially pinned (OFF), the central mass is released (ON) to define generation zero. Subsequent generations follow the UWCA rule: an OFF cell becomes ON only if it shares exactly one edge with an ON neighbor, and once turned ON it remains ON forever. In the mechanical analogue, “ON” corresponds to an unpinned mass free to vibrate, while “OFF” corresponds to a pinned mass that is constrained. This mapping yields a deterministic, rule‑based growth of free degrees of freedom that is inherently non‑periodic.

The authors formulate the dynamics using standard mass‑spring matrices. The mass matrix is diagonal (M = m I) and the stiffness matrix K reflects the square‑grid connectivity; all diagonal entries of K equal 4k, a property that later explains the observed spectral symmetry. By normalizing with ω₀ = √(k/m) the eigenvalue problem reduces to D u = λ u with λ = (ω/ω₀)². Numerical eigenvalue analysis (MATLAB) is performed for the first eleven generations, and results are visualized in a three‑dimensional plot of normalized frequency versus log₁₀(number of degrees of freedom) and generation index.

Key findings include: (1) the natural frequencies are confined to the band ω/ω₀ ≈ 1.2–2.6. The lower bound originates from the effective elastic foundation created by the pinned masses, while the upper bound is the usual cutoff of a discrete lattice. (2) Repeated (degenerate) frequencies appear at several generations, most prominently at ω/ω₀ ≈ 2. The multiplicity of these flat points depends on the overall shape of the “ON” region for a given generation. (3) The eigenvalue spectrum is symmetric about λ = 4. Shifting the spectrum by 4 reveals that each eigenvalue λ has a counterpart –λ, an analogue of particle‑hole symmetry in condensed‑matter systems, directly linked to the constant diagonal of K.

A particularly striking phenomenon is the emergence of strongly localized corner modes. The authors derive that such modes exist only when the generation number satisfies n_g = 4r – 1 (r ≥ 2). At these generations, two distinct frequencies, ω/ω₀ = √3 (λ = 3) and ω/ω₀ = √5 (λ = 5), support modes whose deformation is confined to the lattice corners. The corners possess a “⊥” geometry: three arms of equal length extending from a corner mass. Each corner contributes two degenerate modes per frequency, leading to a total of 2 × (number of corners) localized modes. The number of such corners does not increase monotonically with r; it follows a pattern where r = 2z (z ≥ 2) yields a maximum of 4·3^{z‑1} corners, while r = 2z + 1 collapses the count back to four. This non‑monotonic behavior reflects the underlying self‑replicating rule of the UWCA.

The degree of freedom count grows rapidly with generation, as shown in Figure 3c, indicating that the aperiodic design can accommodate many more vibrational modes within a comparable spatial footprint than a periodic counterpart. The authors argue that this growth, together with the spectral degeneracies and corner localization, offers new avenues for wave manipulation: band‑gap engineering, mode‑multiplicity exploitation, and robust edge or corner confinement without requiring traditional topological design.

In conclusion, the paper demonstrates that cellular‑automaton‑inspired lattice synthesis yields elastic structures with (i) symmetric and highly degenerate eigenfrequency spectra, (ii) generation‑dependent, strongly localized corner modes, and (iii) a scalable increase in dynamic degrees of freedom. These attributes differentiate UWCA lattices from periodic, fractal, and topological metamaterials and suggest potential applications in vibration isolation, energy harvesting, and programmable waveguiding. The work also opens the door to exploring other cellular automata (hexagonal, triangular, or rule‑variations) and incorporating nonlinear or anisotropic material models to further expand the design space.