The Einstein problem is a fundamental question in discrete geometry concerning the existence of a single tile, termed an "einstein" (from the Germanein Stein, meaning "one stone"), that can cover the Euclidean plane through tilings but only in an aperiodic manner, without allowing any periodic arrangements.[1] This longstanding challenge, rooted in the study of aperiodic tilings, was first explicitly framed as the search for a monotile following earlier breakthroughs in multi-tile aperiodic sets during the 1960s and 1970s.[2]The problem traces its origins to Hao Wang's 1961 conjecture on the decidability of tiling problems using square tiles with colored edges, which Robert Berger refuted in 1966 by constructing an aperiodic set of 20,426 such tiles, thereby proving the existence of aperiodic tilings.[2] Subsequent efforts reduced the size of aperiodic tile sets dramatically: Raphael Robinson achieved six tiles in 1971, while Roger Penrose developed a two-tile pair (the kite and dart) in 1973, establishing aperiodic tiling with minimal components.[2] The einstein problem specifically sought to eliminate the need for multiple tiles or reflections, posing whether a single, unreflected prototile could enforce aperiodicity through its geometry alone—a question that remained open for over 50 years despite extensive computational and theoretical searches.[1]In March 2023, the problem was resolved with the publication of a paper by amateur mathematician David Smith, software engineer Joseph Samuel Myers, computer scientist Craig S. Kaplan, and mathematician Chaim Goodman-Strauss, introducing the "hat" tile—a polykite composed of 8 unit kites that tiles the plane hierarchically via substitution rules but admits no periodic tilings, as proven through geometric incommensurability and computer-assisted combinatorial analysis.[1] The hat's tilings incorporate both reflected and unreflected copies. A follow-up paper in May 2023 introduced the strictly chiral "spectre", a 14-sided polygon that tiles aperiodically without reflections.[3] This breakthrough not only answered the einstein problem but spurred further research, including the 2024 journal publication of the chiral monotile and a 2025 experimental realization using molecular self-assembly to form aperiodic structures inspired by these tiles.[4] The solution has implications for quasicrystals, materials science, and computational geometry, highlighting the interplay between mathematical abstraction and physical applications.[2]
Background concepts
Aperiodic tilings
An aperiodic tiling is a covering of the Euclidean plane using a finite set of prototiles, where the tiles fit together without gaps or overlaps, but the resulting pattern lacks any translational periodicity across the entire plane. Specifically, no non-trivial translation can map the tiling onto itself, ensuring that arbitrarily large periodic patches do not occur.[5] This contrasts with periodic tilings, which repeat indefinitely via translations generated by two linearly independent vectors, forming a lattice structure that allows the pattern to be superimposed on itself after finite shifts in multiple directions.[6] Such periodic arrangements, like square or hexagonal grids, exhibit global symmetry and are common in crystalline structures, whereas aperiodic tilings introduce complexity through enforced non-repetition.The concept of aperiodic tilings emerged in the context of decidability problems in logic and geometry. In 1961, Hao Wang introduced Wang tiles—unit squares with colored edges that must match on adjacent sides—and conjectured that any set of such tiles capable of tiling the plane must admit a periodic tiling. This conjecture implied an algorithm to determine tilability, but it was disproved in 1966 by Robert Berger, who constructed the first known aperiodic set of 20,426 Wang tiles as part of his proof that the domino problem (determining whether a given set tiles the plane) is undecidable. Berger's construction demonstrated that aperiodicity could be forced through local matching rules, highlighting the intricate relationship between finite local constraints and global non-periodic behavior.Subsequent work reduced the size of aperiodic tile sets significantly. In 1971, Raphael M. Robinson developed an aperiodic set of 52 tiles, later refining it to six tiles that enforce non-periodicity via hierarchical subdivision rules. Around 1977, Robert Ammann independently discovered several compact aperiodic sets, including ones with as few as 16 prototiles, often using linear markings (Ammann bars) to impose matching conditions that prevent periodicity.[7] These reductions underscored the possibility of minimal aperiodic systems.A key property of aperiodic tile sets is their tendency to produce tilings with hierarchical or self-similar structures, where larger patterns emerge from repeated local motifs without ever achieving full periodicity. Roger Penrose's tilings from 1974 to 1978 exemplify this, using just two prototiles (such as kites and darts or thin and thick rhombs) with matching rules that generate infinite non-repeating patterns exhibiting fivefold symmetry. Monotiles, where a single prototile suffices, represent an extreme case of such prototile minimization.
Prototiles and monotiles
In the context of plane tilings, a prototile is a basic geometric shape, typically a closed set homeomorphic to a disk, that serves as the fundamental unit from which tilings are constructed, allowing for translations, rotations, and possibly reflections unless specified otherwise.[8] Prototiles form the building blocks of tessellations, where multiple copies are arranged to cover the Euclidean plane without gaps or overlaps.A monotile refers to a single prototile capable of tiling the plane on its own. An aperiodic monotile, often called an "einstein," is a monotile that admits tilings of the plane but only in an aperiodic manner, meaning no periodic tiling is possible with that shape alone.[1] The term "einstein" derives from the German phrase ein Stein, meaning "one stone" or "one tile," and was coined by mathematician Ludwig Danzer to playfully denote a singular tile solving the problem.[9]Aperiodicity itself has nuanced definitions: weak aperiodicity for a prototile set means it allows tilings of the plane but forbids any periodic ones, while strong aperiodicity demands that every valid tiling is aperiodic, excluding even those with infinite cyclic symmetries.[10] In the Euclidean plane, for standard (or "normal") prototile sets without pathological features, weak and strong aperiodicity are equivalent, as established by Goodman-Strauss.[10] The Einstein problem specifically seeks a strongly aperiodic monotile, ensuring all tilings are non-periodic.Formally, the Einstein problem asks whether there exists a single prototile in the Euclidean plane that constitutes an aperiodic set of prototiles by itself—that is, a shape that tiles the plane using only translations, rotations, and reflections, but admits no periodic tiling.[1] This quest narrows the broader study of aperiodic tilings, which typically involve multiple prototiles, to the challenge of achieving aperiodicity with just one shape and no additional matching rules.Related variants in higher dimensions, such as three-dimensional space, remain open for a strongly aperiodic monotile; however, a weakly aperiodic example exists in the form of the Schmitt–Conway–Danzer tile, which tiles \mathbb{R}^3 without translational periodicity but permits certain screw symmetries.[8]
The Einstein problem originates from early 20th-century inquiries into the nature of plane tessellations, particularly those posed in David Hilbert's 18th problem at the 1900 International Congress of Mathematicians. This problem encompassed several questions about constructing space from polyhedra, including whether the Euclidean plane admits a tessellation by mutually incongruent convex polygons and the existence of shapes that tile monohedrally (using congruent copies) but not isohedrally (with transitive symmetry action on tiles). While Ludwig Bieberbach partially addressed related aspects by proving in 1911–1912 that there are only finitely many crystallographic space groups in n dimensions up to isomorphism, limiting periodic tilings to 230 types in three dimensions, the broader questions about non-periodic and anisohedral configurations left significant gaps in tiling theory.[11]Pioneering work on anisohedral tiles—shapes that permit monohedral tilings of the plane but no isohedral ones—built directly on Hilbert's framework. In 1928, Karl Reinhardt provided the first examples of such tiles in three dimensions, resolving a key part of Hilbert's 18th problem by demonstrating polyhedra that tile space without the tiling symmetries acting transitively on the tiles. Extending this to the plane, Heinrich Heesch identified the initial two-dimensional anisohedral prototile in 1935, a convex polygon that covers the plane with congruent copies but lacks full symmetry transitivity in any such covering. These early constructions, however, all supported periodic tilings, highlighting a distinction from the aperiodic focus of the later Einstein problem.[12][13]Heesch's investigations in the 1930s further advanced tiling theory by examining the structural constraints of plane mosaics composed of congruent figures, laying groundwork for analyzing how local arrangements influence global coverings. His later formulation of the corona problem, involving successive layers (coronas) around a central tile, quantified the extent to which a shape can extend without gaps or overlaps; the Heesch number denotes the maximum such layers before obstruction. This approach revealed that finite Heesch numbers preclude complete planetilings and inspired ideas of unavoidable configurations—local patterns that inevitably lead to inconsistencies in periodic extensions—paving the way for aperiodic constraints in tiling design.[14][12]The nomenclature "Einstein problem" emerged much later as a playful reference to an aperiodic monotile, derived from the German phrase ein Stein ("one stone" or "one tile"), bearing no connection to the physicist Albert Einstein.[2]A pivotal shift toward aperiodicity and computability occurred with Hao Wang's 1961 conjecture that any finite set of square tiles (Wang tiles) capable of tiling the plane must admit a periodic tiling, suggesting the decidability of the tiling problem. RobertBerger refuted this in 1966 by constructing an aperiodic set of 20,426 Wang tiles—proving that tilings exist but none are periodic—and demonstrating the undecidability of the domino problem via reduction to the halting problem. This resolution transitioned geometric tiling inquiries into computability theory, establishing aperiodic tilings as a rigorous concept emerging from these foundational origins.[15]
Early aperiodic constructions
The first explicit construction of an aperiodic tile set was provided by Robert Berger in 1966, who demonstrated the existence of a finite set of 20,426 Wang tiles capable of tiling the plane only in non-periodic ways.[16] This set, consisting of unit squares with colored edges that must match adjacently, encoded simulations of Turing machine computations to enforce aperiodicity, simultaneously proving the undecidability of the domino problem—the question of whether a given finite set of tiles can tile the plane at all.[16]Subsequent efforts focused on reducing the number of tiles while preserving aperiodicity. In 1971, Raphael M. Robinson constructed a smaller aperiodic set of six prototiles, introducing a hierarchical structure based on larger "supertiles" that recursively subdivide into smaller versions of themselves. Further reductions followed in the 1970s and 1980s; for instance, Robert Ammann developed an aperiodic set of 16 Wang tiles around 1977, and work by Branko Grünbaum and Geoffrey C. Shephard in their 1987 book Tilings and Patterns explored additional families of aperiodic prototiles, contributing to progressive minimizations that eventually reached sets as small as 13 Wang tiles by the mid-1990s.[17] These constructions typically relied on edge-matching rules or decorations to prevent periodic arrangements. These early sets were primarily Wang tiles, but later constructions shifted to more geometrically diverse prototiles.Parallel developments shifted away from square Wang tiles toward more geometrically diverse prototiles. In 1974, Roger Penrose introduced an aperiodic set comprising a "dart" and a "kite"—two rhombi derived from a 72°-108° rhombus, with ratios governed by the golden ratio \phi = \frac{1 + \sqrt{5}}{2} to ensure matching constraints. By 1976, Penrose refined this to a pair of thin and thick rhombi, also scaled by \phi, which tile the plane aperiodically through forced hierarchical subdivisions. These sets exemplified how irrational ratios in tile dimensions or angles could inherently prohibit translational periodicity.Aperiodicity in these early constructions was often enforced through substitution rules, where basic prototiles are replaced by larger supertiles composed of multiple copies, scaled by an inflation factor (such as \phi in Penrose tilings). This process generates a self-similar hierarchy: each supertile level mimics the overall structure at a larger scale, but the irrational scaling prevents the pattern from repeating periodically across the infinite plane. In Robinson's framework, for example, macro-tiles emerge that must align in ways that propagate defects or shifts, ensuring no global periodicity is possible despite local tiling validity. Grünbaum and Shephard analyzed such mechanisms in their comprehensive study, highlighting how local matching rules cascade into global non-repetitive order.Despite these advances, all known aperiodic tile sets from this era required multiple prototiles—at minimum two, as in Penrose's constructions—and no single tile (monotile) capable of aperiodic tilings of the plane had been identified, setting the stage for the longstanding Einstein problem.
Proposed solutions
Pre-2023 attempts
Early efforts to find an aperiodic monotile in the plane before 2023 focused on shapes that imposed non-periodicity through additional constraints, higher dimensions, or non-standard tiling rules, but none satisfied the strict criteria of a single connected prototile forming only aperiodic edge-to-edge tilings without markings, overlaps, or reflections. These attempts highlighted persistent challenges, such as the need for supplementary rules or deviations from planar geometry, underscoring the difficulty of achieving strong aperiodicity through shape alone.[8]In 1988, Peter Schmitt constructed a convex polyhedron prototile for three-dimensional Euclidean space that tiles aperiodically via screw symmetries, producing tilings invariant under translations but not the full group of Euclidean motions; however, this solution operates outside the plane and thus does not address the two-dimensional Einstein problem.[18] The tile's design inspired later work but failed to yield a planar monotile, as its aperiodicity relies on 3D structure.[8]Petra Gummelt proposed a star-shaped decagon in 1996 that covers the plane aperiodically through substitutions and overlaps in five specific ways, corresponding to Penrose-like patterns; this approach deviates from standard edge-to-edge tilings by permitting overlaps, disqualifying it as a true monotile for the Einstein problem.[19] The construction demonstrates how overlap rules can enforce aperiodicity but highlights the need for non-overlapping alternatives.[8]In 2010, Joshua Socolar and Joan Taylor introduced a hexagonal prototile with arrow and color markings to enforce local matching rules that force aperiodicity, drawing from Penrose tilings; when realized without markings via shape alone, the effective tile consists of disconnected components and necessitates reflections (mirror images) to complete tilings, violating connectivity and chirality requirements for a strong planar monotile.[20] This near-miss advanced understanding of rule-based enforcement but underscored limitations in geometric realization.[21]The Voderberg tile, a non-convex enneagon discovered by Heinz Voderberg in 1936 and revisited in later analyses, admits both periodic and spiral (non-periodic) tilings of the plane in an edge-to-edge manner without inherent overlaps or rules to exclude periodicity; its ability to form repeating patterns prevents it from qualifying as aperiodic, serving instead as an early example of a monotile with flexible tiling behaviors.Weak aperiodic candidates, such as certain chiral shapes or those relying on orientation restrictions, can produce only non-periodic tilings when reflections are forbidden by fiat, but they admit periodic arrangements if mirrors are allowed, failing the strong aperiodicity criterion that prohibits all periodic tilings under the full Euclidean group.[8] Examples include modifications of earlier prototiles where chirality enforces non-repetition, yet these depend on imposed constraints rather than intrinsic geometry.[22]Computational searches prior to 2023 exhaustively enumerated tiling polygons up to pentagons, confirming that no convextriangle, quadrilateral, or pentagon serves as an aperiodic monotile, as all known convex tilers admit periodic arrangements; for instance, the complete classification of convexpentagons that tile the plane revealed 15 families, none inherently aperiodic. These efforts established that simple convex shapes cannot solve the problem, shifting focus to non-convex forms.[23]
The hat tile
The hat tile, also known as an "einstein" in reference to the Einstein problem, was discovered in November 2022 by amateur mathematician David Smith during an exploratory search for polyforms capable of tiling the plane, using software such as the PolyForm Puzzle Solver developed by Jaap Scherphuis.[1][24] Smith, a retired printing technician from East Yorkshire, England, identified the shape while experimenting with irregular polygons, initially suspecting its aperiodic properties after generating numerous tilings that appeared non-periodic.[25] He shared his findings with mathematician Craig S. Kaplan at the University of Waterloo, leading to collaborative verification.[1]Geometrically, the hat tile is a 13-sided polygon composed of eight congruent kites, each with angles of 60°, 90°, 120°, and 90°.[1] Its side lengths incorporate rational multiples of 1 and \sqrt{3}, along with irrational values derived from these to ensure incommensurability, which contributes to the enforced aperiodicity; specifically, it is denoted as Tile(1, \sqrt{3}) in the foundational construction.[1] The overall shape vaguely resembles a fedora hat, hence its name, and spans a total area that allows for dense packing without gaps.[24]Tilings using the hat tile are edge-to-edge, covering the plane without gaps or overlaps, and require both the tile and its mirror reflection to achieve full aperiodicity, resulting in an achiral tiling overall.[1] The matching rules enforce a hierarchical structure where edges align precisely based on the kite subdivisions, preventing periodic arrangements through forced irregularities in orientation and positioning.[1]The proof of aperiodicity, established in 2023 by Smith, Joseph Samuel Myers, Kaplan, and Chaim Goodman-Strauss, demonstrates that every valid tiling of the plane by the hat tile must contain arbitrarily large finite supertiles exhibiting 12-fold rotational symmetry, which inherently disrupts translational periodicity across the infinite plane.[1] This geometric forcing arises from substitution rules that build larger metatiles iteratively, ensuring no global lattice structure can emerge.[1]The hattile serves as the prototype for an infinite family of aperiodic monotiles, generalized through hierarchical meta-tile constructions and substitution systems; notable variants include the "turtle" tile, denoted Tile(\sqrt{3}, 1), and the "spectre" tile, all derived by parameterizing the side length ratios and angles while preserving the core aperiodic mechanism.[1]The discovery and proof were first detailed in an arXiv preprint released in March 2023, with the formal publication appearing in Combinatorial Theory in July 2024.[1]
The spectre tile
The spectre tile, introduced in May 2023 by David Smith, Joseph Samuel Myers, Craig S. Kaplan, and Chaim Goodman-Strauss, represents a strictly chiral aperiodic monotile that tiles the plane only through non-periodic arrangements without requiring reflections.[26] This development builds directly on their earlier "hat" tile by enforcing homochirality, using either left-handed or right-handed enantiomers exclusively to address whether a single chiral shape can force aperiodicity independently of its mirror image.[26]Geometrically, the spectre is a 14-sided polygon derived from an equilateral variant in the hat-turtle continuum, specifically Tile(1,1), but with adjusted kite-like proportions and edges modified by smooth, non-straight S-curves to prevent mirror-image matches.[26] These modifications ensure that tilings rely solely on translations and rotations, maintaining consistent handedness across the plane while allowing mixtures of orientations within that chirality.[26]The proof of aperiodicity demonstrates that all valid tilings of the spectre generate hierarchical clusters through substitution rules, inevitably producing supertiles with irrational scaling ratios, such as $4 + \sqrt{15}, that preclude any periodic lattice structure.[26] Central to this are nine marked clusters—labeled with the Greek letters \Gamma, \Delta, \Theta, \Lambda, \Xi, \Pi, \Sigma, \Phi, and \Psi—which form supertiles combinatorially equivalent to marked hexagons; two substitution rules apply: one replaces a single spectre with a cluster containing one "mystic" (a reflected auxiliary shape) and seven spectres, while the other replaces a mystic with one mystic and six spectres, forcing an infinite hierarchy of increasingly complex, non-repeating supertiles.[26]As a meta-tile within the hat family, the spectre shares the foundational continuum of the hat but refines it to eliminate reflections entirely, enabling purely chiral tilings that the hat's achiral design cannot achieve without mirrors.[26] This relation highlights the spectre's role in extending the hat's weakly chiral properties to strict chirality.A formal proof confirming strong aperiodicity for a single enantiomer of the spectre was published in 2024 under the title "A chiral aperiodic monotile" in Combinatorial Theory.[26] The key innovation lies in resolving the longstanding question of whether a lone chiral prototile can enforce aperiodicity without its mirror counterpart, advancing the understanding of chiral constraints in aperiodic tilings.[26]
Applications and implications
In mathematics
The resolution of the Einstein problem in 2023 with the discovery of the hat tile, a 13-sided polykite, confirmed the existence of a strong aperiodic monotile capable of tiling the Euclidean plane solely in non-periodic ways, though requiring reflected copies and hierarchical substitution rules. This breakthrough directly addresses a conjecture posed by mathematicians like Roger Penrose, whose 1970s work established aperiodic tilings using two prototiles, by demonstrating that a single tile suffices to enforce aperiodicity through geometric incommensurability and hierarchical substitution rules. The proof combines geometric arguments showing the absence of translational symmetries with exhaustive case analysis of local patch configurations, marking a significant advancement in discrete geometry and tiling theory.[1]Subsequent work introduced the spectre tile, a chiral variant that tiles the plane aperiodically without needing mirror images, further solidifying the existence of such monotiles and expanding the theoretical framework for non-periodic structures. These tiles serve as discrete mathematical analogs to quasicrystals, the aperiodic atomic arrangements first observed by Dan Shechtman in 1982 via electron diffraction patterns in aluminum-manganese alloys, which challenged crystallographic periodicity and earned Shechtman the 2011 Nobel Prize in Chemistry. The hat and spectre tilings, with their substitution-based hierarchies and diffraction-like properties, provide idealized models for analyzing the long-range order in quasicrystals, bridging pure geometry with the mathematical description of physical aperiodic order. In January 2025, researchers demonstrated an experimental realization of aperiodic chiral tilings through molecular self-assembly of tris(tetrahelicenebenzene) on silver surfaces, forming quasicrystal-like structures.[4][3][27][1]Computationally, the proofs for these monotiles relied on software for enumerating thousands of local configurations and verifying substitution rules, positioning the Einstein problem as a benchmark for constraint satisfaction problems in computer science, where tiles must satisfy adjacency rules without global periodicity. This approach highlights connections to automated theorem proving, as tools like those used by Craig Kaplan systematically checked finite cases to establish global aperiodicity. The work also bears on decidability in geometry: while the general domino problem of whether a tile set admits a tiling is undecidable (as shown by Berger in 1966 using aperiodic sets), the monotile solutions raise refined questions about algorithmic verifiability for single-tile aperiodicity, with implications for higher-dimensional or non-Euclidean settings where undecidability persists.[28][29][30]The hat tile generalizes to an infinite continuum of aperiodic monotiles, parameterized by side-length ratios that preserve combinatorial equivalence and tunable symmetries while maintaining strict aperiodicity, allowing for families with varying geometric properties. Extensions to other spaces, such as aperiodic monotilings of the hyperbolic plane or three-dimensional Euclidean space, remain open challenges, with current results limited to the planar case and higher-dimensional undecidability results suggesting significant obstacles. To foster broader mathematical engagement, the National Museum of Mathematics hosted the 2023 Hat Tile Art Contest, which drew over 245 entries from 30 countries and awarded three winners—Evan Brock and Shiying Dong in the global category, and Devi Kuscer in the scholastic category—for creative artworks inspired by the tiles, thereby promoting public understanding of aperiodic structures.[1][29][31][32]
In engineering and materials science
Aperiodic monotiles, such as the hat and spectre tiles, have inspired designs for quasicrystal analogs in materials science, where molecular self-assembly mimics hat-like patterns to create structures with forbidden rotational symmetries. These assemblies enable photonic and electronic materials exhibiting unique optical and conductive properties due to their aperiodic order, as demonstrated by the formation of chiral tilings using tris(tetrahelicenebenzene) molecules on silver surfaces, achieving dense packing with topological defects that promote entropy-driven aperiodicity.[4] Such self-assembled quasicrystals offer potential for advanced devices like chiral metamaterials for light manipulation, where the lack of translational symmetry prevents diffraction in conventional crystals.[4]In metamaterials engineering, aperiodic honeycombs derived from hat tilings have been developed to achieve zero or negative Poisson's ratios, resulting in auxetic behavior that enhances impact resistance by expanding laterally under compression. For instance, hat-based lattice metamaterials exhibit anisotropic Poisson's ratios ranging from -0.064 to 0.215, with directional auxetic effects that improve energy absorption compared to periodic counterparts.[33] Similarly, isotropic zero-Poisson's-ratio honeycombs constructed from the hat monotile maintain consistent deformation across orientations at relative densities above 0.225, making them suitable for applications requiring uniform shock absorption without lateral strain transfer, such as protective coatings.[34]Aperiodic monotiles facilitate the design of composites with tunable mechanicalproperties, leveraging infinite families of tile variants to optimize stiffness, toughness, and fractureresistance. These structures promote zigzag crack propagation, increasing energy dissipation and defect tolerance over periodic honeycombs; for example, hat-based composites with 80% VeroClear volume fraction show 103% higher stiffness, 34.5% greater strength, and 15.9% improved toughness.[35] The chiral spectretile, in particular, enables directional properties in multi-phase composites, where curved edges enhance fractureresistance through curvature-induced stress distribution, as explored in recent designs that integrate soft and hard phases for superior load-bearing capacity.[36]Prototypes of aperiodic structures have been fabricated via 3D printing for engineering applications, including lightweight aerospace components and vibration-damping surfaces. Polyjet-printed hat-derived honeycombs demonstrate enhanced compressive strength under varying temperatures, reducing weight while maintaining structural integrity in high-stress environments like aircraft panels.[37] These additively manufactured metamaterials also exhibit superior damping due to their irregular geometries, which disrupt wave propagation more effectively than periodic lattices, as validated in prototypes for noise and vibration control in aerospace assemblies.[37]Despite these advances, challenges persist in scaling aperiodic 2D tiles to 3D volumes, including geometric complexity in boundary triangulation and assemblability issues that lead to collisions during topological interlocking. Ongoing research addresses these through multi-phase chiral composites, with 2024 studies on aperiodic monotile-patterned structures exploring deformation-tolerant assembly methods to enable practical 3D fabrication.[38][36]