Moiré straintronics: a universal platform for reconfigurable quantum materials

Strong electron correlations emerge when Coulomb interactions are comparable to or exceed kinetic energy. Long-period moiré patterns in stacked 2D crystals (with lattice mismatch δ and/or twist θ) enable broad tunability of Coulomb and kinetic energy scales and carrier density, providing a versatile platform to study correlated phases (e.g., twisted bilayer graphene near the magic angle and TMD moiré heterostructures). The Hubbard model captures key physics via hopping t and on-site repulsion U; when U/t > 1, strong correlations dominate, but higher-dimensional solutions are sensitive to parameters and lattice geometry, motivating quantum simulators such as ultracold atoms with tunable U/t and geometry. In 2D moiré materials, in situ tunability of lattice symmetry, U, and t comparable to optical lattices has been limited: displacement fields tune interlayer coupling and band alignment, while U and t have been controlled mainly by material choice, twist θ, or dielectric environment—parameters typically fixed during fabrication. This hampers fine tuning near critical points and exploration of phase diagrams. Heterostrain (differential strain between layers) has been proposed/theorized and observed as a route to tailor electronic properties and design flat bands, but experimental studies largely involved unintentional or fixed strain; few works realized in situ heterostrain devices. Its use to directly control the moiré superlattice—and thereby tune U/t—has not been fully explored. This work develops a general, exact geometrical framework for arbitrary in-plane strain in bilayers to analyze how biaxial, uniaxial, and shear heterostrain tune the size (wavelength) and symmetry (geometry) of moiré lattices, enabling reconfigurable quantum materials and precise tuning around critical regimes (e.g., magic-angle graphene).

The paper situates its contribution within extensive prior work on correlated phases in moiré materials, including magic-angle graphene (Cao et al., 2018; Balents et al., 2020) and TMD moiré systems exhibiting Hubbard physics and Wigner crystal states (Tang et al., 2020; Regan et al., 2020). Synthetic moiré platforms have revealed moiré excitons, twisted magnets, solitons, polaritons, and ferroelectricity. Prior tunability relied mainly on twist angle and dielectric environment, or displacement fields for interlayer coupling, but real-time tuning of U/t remained limited. Theoretical studies on heterostrain predicted designed flat bands in graphene and TMDs (Bi et al., 2019; Mannai & Haddad, 2021; Tong et al., 2017), with experimental observations of heterostrain effects and 1D moiré patterns in select systems (Huder et al., 2018; Bai et al., 2020). Moiré metrology and reconstruction phenomena have been reported (Weston et al., 2020; Halbertal et al., 2021). The present work advances by providing an analytically exact, universal geometrical description of moiré lattices under arbitrary in-plane heterostrain for general bilayers, beyond hexagonal approximations and specific strain limits.

The authors develop a general, analytically exact geometrical framework for moiré superlattices formed by two stacked 2D layers under arbitrary global in-plane strain applied to one layer. Key elements:

• Lattice description: Two monoclinic lattices with primitive vectors a_i (lower layer) and b_i (upper layer), with lattice constant mismatch δ and relative twist θ. The formulation assumes similarly shaped unit cells with scaling by (1+δ).
• Strain model: A uniform 2D strain tensor ε is applied to the lower layer; strain transfer to the upper layer is parameterized by μ (0 for pure heterostrain; 1 for homostrain; 0<μ<1 partial transfer). The deformed lattice vectors incorporate ε and μ. High interlayer friction in commensurate cases (small θ or δ) can lead to μ>0, while incommensurate cases allow μ≈0.
• Strain decomposition: ε is parameterized into biaxial strain ε_c and complex shear ε_s with shear orientation φ_s via ε = ε_c I + ε_s S_φs, where ε_c=(ε_xx+ε_yy)/2, ε_s=ε_xx−ε_yy + i2ε_xy, S_φs=cos(φ_s)σ_x+sin(φ_s)σ_y, and φ_s determined by ε components. Uniaxial strain is treated as a mixture of biaxial and shear with φ_s=90°, accounting for Poisson ratio ν (ε_yy=−ν ε_u for strain along x).
• Exact moiré vectors and area: Closed-form expressions for the strained moiré lattice vectors A′_i (Eq. 5) are derived with denominator Δ dependent on δ, θ, ε_c, ε_s, and μ. The moiré unit cell area M=||A′_1×A′_2|| is obtained (Eq. 7). No approximations are made regarding lattice symmetry; results apply to general homo- and heterobilayers with similarly shaped unit cells.
• Special cases and assumptions: For homostrain (μ=1), A′_i=(1+ε_c)A_i, implying small effects for realistic ε. For μ<1 (heterostrain), Δ dominates, enabling substantial tuning of moiré size/shape. Atomic reconstruction is neglected in the geometric derivation (rigid layers), but discussed regarding its suppression at sufficiently large θ (e.g., >2.5° for MoSe2–WSe2 3R, ~1.0° for 2H; ~3.0° for 2H homobilayers) or via sufficient heterostrain (ε_c≈3% can reduce reconstruction at θ=0°). Supplementary Notes provide derivations and experimental implementation strategies.
• Case studies: Calculations focus on hexagonal homo- (WSe2) and heterobilayers (MoSe2/WSe2) with realistic lattice constants and Poisson ratios, exploring pure heterostrain (μ=0) under biaxial, uniaxial, and shear strain, mapping moiré parameters (A′_1, A′_2), angle α between A′ vectors, and area M versus ε and θ.

• Universal exact framework: Provides a general, analytically exact description of moiré lattice vectors and area under arbitrary in-plane strain for any bilayer with similarly shaped unit cells, extending beyond prior hexagonal/approximate treatments.
• Control of interaction scales: In heterostrain (μ<1), moiré period A can be tuned strongly via ε; since t decreases exponentially with A and U ∝ 1/A, the ratio U/t is directly tunable in situ by strain (particularly ε_c), enabling exploration of correlated regimes and fine-tuning near critical points.
• Biaxial heterostrain (hexagonal bilayers): • Decreases moiré lattice parameters with approximate A′ ∝ 1/|ε_c| (homobilayer case). • In homobilayers, A′ diverges at θ=0°, ε_c=0 (no moiré in natural 2H/3R stacking), while any θ≠0 or ε_c≠0 yields finite moiré periods. • In MoSe2/WSe2 heterobilayers (lattice mismatch ≈0.4%), stretching WSe2 by ~0.4% (ε_c≈+0.4%) compensates mismatch and recovers divergence at θ=0°, enabling arbitrarily large moiré lattices analogous to homobilayers. • For hexagonal layers, ε_c (and ε_s=0) does not change moiré geometry: A′_1=A′_2 and α=60° persist.
• Uniaxial heterostrain: • Produces stronger effects than biaxial: divergences of A′_1 and A′_2 occur along curves in (ε_u, θ) space (not just a single point), and degeneracy between A′_1 and A′_2 is lifted. • Enables continuous reconfiguration of moiré geometry; by tuning ε_u and θ one can realize rectangular lattices (α=90°) or approach 1D moiré (α→0° or 180°) while M diverges along red curves. • In homobilayers, divergence curves intersect at (ε_u=0, θ=0). In heterobilayers, curves show avoided crossing with vertices: v2 at ε_u≈+0.4% (lattice mismatch), v1 at ≈−2.1% set by Poisson deformation (−0.4/0.19), illustrating sensitivity to ν; ν→0 can eliminate divergence in negative ε_u.
• Shear heterostrain: • The denominator Δ can vanish, leading to divergence of A′ (1D moiré patterns), independent of lattice orientation θ0, shear angle φ_s, or lattice shape. • As with uniaxial strain, divergences appear along curves in (ε_s, θ) space, with α→0° or 180°. In heterobilayers, divergence curves have vertices at ±0.4% shear to compensate mismatch. • The framework generalizes to non-hexagonal lattices (e.g., rectangular WTe2 homobilayer; similar behaviors shown in Supplementary Note 8).
• Combined strain control (ε_c and ε_s): Demonstrated roadmap (θ=1° homobilayer) for: • Tuning moiré size (thus U/t) via ε_c at fixed geometry (points III/IV). • Adjusting anisotropy A1/A2 via ε_s along A1=A2 contours to realize triangular Hubbard models with two tunable hoppings t1 and t2. • Accessing square-lattice-like geometries (point V) under appropriate ε_s and ε_c.
• Practical considerations and benchmarks: • Avoid atomic reconstruction by using θ above critical angles (e.g., ~2.5° for MoSe2–WSe2 3R; ~1.0° for 2H; ~3.0° for 2H homobilayers) or applying sufficient heterostrain (ε_c≈3% at θ=0°). • Strain tuning accuracy can correspond to ~0.0001° equivalent twist precision; shear ~0.36% at θ≈1.25° can yield flat bands in twisted bilayer graphene, enabling fine tuning post-fabrication. • Heterostrain can lift particle–hole symmetry and valley degeneracy and tune miniband bandwidths in graphene, offering an extra knob to engineer flat bands and correlated phases.

The work addresses the challenge of in situ tunability in moiré quantum materials by establishing strain—particularly heterostrain—as a universal, precise, and flexible control parameter for both the strength of electronic interactions (via moiré length scale) and lattice symmetry (via moiré geometry). The exact geometrical framework links arbitrary in-plane strain components to moiré vectors and unit cell area, enabling predictive design across material systems and lattice types. By mapping how biaxial, uniaxial, and shear heterostrain reshape moiré superlattices, the authors demonstrate routes to: (i) continuously tune U/t by resizing the moiré period; (ii) reconfigure lattice symmetry from hexagonal to rectangular to effectively 1D, opening access to models beyond the standard Hubbard model (e.g., coupled Luttinger liquids) and anisotropic Hubbard models with independently tunable hoppings; and (iii) compensate fabrication variabilities (e.g., slight twist-angle deviations) with high precision using strain. The analysis highlights experimental feasibility, with established strain-tuning platforms and clear metrology (e.g., piezo response force microscopy) to locally validate moiré size and shape under strain. This strain-based approach complements and surpasses twist-only control by enabling real-time, reversible, and fine-grained tuning around critical points such as magic-angle conditions in graphene and correlated phase boundaries in TMD moirés.

The paper introduces a universal, analytically exact framework to predict and engineer moiré superlattices under arbitrary in-plane heterostrain, applicable to general bilayer lattices. It shows that heterostrain offers unprecedented in situ control over moiré size and symmetry, enabling direct tuning of the interaction-to-kinetic energy ratio U/t and reconfiguration of lattice geometry (hexagonal, rectangular, 1D). Case studies on hexagonal homo- and heterobilayers demonstrate divergence conditions, compensation of lattice mismatch (~0.4%), and pathways to design anisotropic Hubbard models with multiple hopping parameters. The approach provides a practical solution to fabrication variability (e.g., near magic angles) and points to strain as a precise knob for reconfigurable quantum materials and quantum simulation. Future directions include integrating the framework with atomic reconstruction effects, extending to multilayer and non-monoclinic systems, and experimental realization of combined strain fields to dynamically traverse correlated phase diagrams (e.g., flat-band engineering, Luttinger liquids, and beyond-Hubbard physics).

• The core derivation assumes rigid layers and neglects atomic reconstruction; while the authors discuss regimes to suppress reconstruction (θ above material-dependent critical angles or sufficient heterostrain), reconstruction can influence real systems near small θ or δ.
• The practical strain range is limited (|ε_c|, |ε_s| ≪ 1), and homostrain effects on moiré geometry are small; significant tuning relies on imperfect strain transfer (μ<1), which depends on interlayer friction and commensurability.
• Real devices may exhibit partial strain transfer (0<μ<1), spatial nonuniformities, and slippage, which can complicate precise control compared to the idealized global uniform strain model.
• The study focuses on geometric control; translating geometry changes to quantitative electronic properties (band structures, interactions) requires additional material-specific modeling beyond the geometric framework.