betaFeedback

WorksPhysics and Astronomy

Adiabatic confinement and number-theoretic spectra of massless Dirac fermions in twisted trilayer graphene

Abstract

Electrons in graphene move like massless relativistic particles, making them exceptionally mobile but difficult to trap because Klein tunnelling permits unimpeded passage through electrostatic barriers. This challenge extends to atomic collapse and hadron confinement. Twisted trilayers form Bernal and rhombohedral domains joined by smooth boundaries, bringing their carriers into contact. Here we show that a new mechanism, distinct from established approaches, confines massless fermions in equilateral triangular Bernal domains without mass barriers, bulk gaps or physical edges. Simulations and analytical modeling reveal that massless carriers have no adiabatic continuation into rhombohedral domains, while escape through nonadiabatic coupling to massive carriers is exponentially suppressed by smooth stacking variation. Perfect reflection in the adiabatic limit turns Bernal domains into quantum cavities in a gapless, edgeless metal. Striking density-of-states oscillations encode Eisenstein-integer arithmetic, linking energies and degeneracies to norms and prime factorizations. These findings bring relativistic confinement and number theory into experimentally accessible electronic spectra.

Moiré materials formed from twisted bilayers of two-dimensional (2D) crystals provide versatile platforms for exploring quantum phenomena beyond those accessible in conventional solid-state systems [1]. Extending this approach to multilayers expands the design space, enabling control over moiré length scales, layer number, interactions, and symmetries [2, 3]. These degrees of freedom support emergent phenomena governed by physics distinct from that of the constituent layers [1, 2, 3].

Spatial variations in electronic structure provide another degree of freedom. Lateral interfaces between different 2D crystals enable control over band alignment, carrier injection and transport [4, 5, 6, 7], while coupling between states with distinct dispersions and symmetries governs transmission and reflection [8]. In reconstructed moiré systems, smooth stacking variations create such junctions within a chemically uniform lattice.

Confining massless Dirac fermions at such interfaces is challenging because Klein tunnelling permits perfect transmission through smooth electrostatic barriers, obstructing conventional confinement [9, 10]. Previous efforts to circumvent this limitation have explored etched structures [11], confining mass terms [12, 13], and inhomogeneous magnetic fields [14] to name a few. Whether an intrinsic mechanism can confine these massless carriers within a metal remains an appealing question.

Twisted trilayer graphene (TTG) [15, 16, 17] provides a natural setting for such confinement. Its moiré superstructure brings Bernal- and rhombohedral-stacked trilayer graphene (BTG and RTG) into close proximity [17]. At low energies, BTG hosts massless and massive Dirac fermions decoupled by mirror symmetry, whereas RTG exhibits a qualitatively different energy–momentum dispersion [18, 19]. These domains invite exploration of scattering, confinement and interactions between distinct quasiparticles across atomically smooth interfaces.

We combine large-scale simulations [20, 21] with analytical modeling to demonstrate a new mechanism for confining massless fermions in metallic TTG, requiring neither physical edges [11] nor mass-gap engineering [12, 13]. Smooth stacking boundaries break the mirror symmetry that decouples massless and massive fermions in BTG, hybridizing the two branches. We show that the massless branch has no propagating continuation into RTG domains and is therefore reflected adiabatically, while smooth stacking variation exponentially suppresses its nonadiabatic escape into the massive branch. Near-perfect backscattering thus confines massless Dirac fermions in a gapless metal. The computed density of states (DOS) exhibits intriguing oscillations that reveal quantization of the linear Dirac dispersion. The resulting spectrum encodes the arithmetic of Eisenstein integers [22, 23, 24], whose norms select the confined energies and whose prime factorizations determine their degeneracies. We believe that TTG establishes a new route to massless-particle confinement and brings number theory into the experimentally accessible spectrum of a solid.

Electronic structures of twisted trilayer graphene

Let us first discuss the stacking-dependent electronic structures of pristine systems. As shown in Figure 1a and Figure 1b, RTG has a pair of bands with the cubic dispersion touching at the Fermi level (EFE_{\mathrm{F}}), whereas BTG has the decoupled linear and quadratic dispersions [18, 19]. Besides these low energy bands, two quadratic conduction and valence bands appear beyond the van Hove singularities (vHSs) for RTG, and additional single quadratic band also appears beyond vHSs for the BTG. Those energy levels are indicated by dashed lines in Figure 1a and Figure 1b. The DOS reflects the band characteristics, such as the diverging DOS at vHSs and EFE_F in RTG and linear DOS with abrupt jumps at vHSs in BTG.

Electronic structures of pristine and twisted trilayer graphene

Figure 1. Electronic structures of pristine and twisted trilayer graphene. a, Band structures (left panel) and DOS (right panel) of RTG. The EFE_F is set to zero. The energy ranges 1, 2, and 3 are marked between van Hove singularities, whose energy levels are indicated by dashed lines at 0.440, 0.288, −0.281-0.281, −0.440-0.440 eV. b, same for BTG. Insets in a and b show the schematic interlayer configuration. Δk\Delta k denotes the momentum displacement from KK, with positive (negative) values toward MM (Γ\Gamma). c, Local stacking order parameter of TMB graphene. Red (blue) denotes perfect rhombohedral (Bernal) stacking (see Methods). The dashed circle, rectangles, and triangles indicate the regions used for the projected local DOSs. d, Same as c for ATT graphene. Insets in c and d are schematic stacking configurations. e and f, Total DOSs of TMB and ATT shown in c and d, respectively, where the blue solid lines represent our simulations and the gray dashed or dotted lines are the reference inferred from the pristine systems.

Comparing the two band structures, three distinct energy ranges separated by vHSs can be identified, which we denote as ranges 1, 2 and 3, respectively and mark them in Figure 1. It is noteworthy that, in ranges 1 and 2, BTG and RTG differ in the number of energy bands, with one stacking hosting an additional band relative to the other, while in range 3, the both have the same number of bands. Owing to this mismatch, at least one scattering channel must be completely reflective for electrons to propagate from one stacking to the other when they are joined side by side (See Method for details) [25]. Such a situation arises naturally in twisted trilayer systems.

To investigate the scattering behavior, we consider two representative TTGs characterized by twist angles θ12\theta_{12} and θ23\theta_{23} between adjacent layers.

The first is twisted mono-bilayer (TMB) graphene, formed by twisting a monolayer relative to a bilayer by θ12\theta_{12}. Its fully relaxed structure consists of alternating RTG and BTG domains separated exclusively by RTG–BTG boundaries (Figure 1c). For θ12<0.05∘\theta_{12}<0.05^\circ, these triangular domains undergo spontaneous symmetry breaking into a kagome-like pattern [17]. The second is alternately twisted trilayer (ATT) graphene with θ23=−θ12\theta_{23}=-\theta_{12}. After relaxation, it forms alternating triangular Bernal domains (ABA and ACA), as shown in Figure 1d. We note that, for angles around 1–5 degrees, renormalization of band structures can occur owing to moiré potentials [15, 16]. So, we adopt a marginal twist angle of θ12=0.11∘\theta_{12}=0.11^\circ for both systems (corresponding to a moiré period of LM=1280.13L_M=1280.13 Å), thereby minimizing the strong band renormalization [17] (for detailed methods, See Methods).

Figure 1e shows the DOS of the TMB system obtained from large-scale atomistic calculations, together with a reference constructed from the averaged DOSs of pristine RTG and BTG. Overall, the calculated DOS closely follows the reference, reproducing the singularity at EFE_F and other vHSs. Notably, however, ranges 1 and 2 develop oscillatory features, whose envelope approximately follows the reference DOS. By contrast, range 3 exhibits excellent agreement with the reference and no discernible oscillations. The DOS of ATT graphene is qualitatively different. As shown in Figure 1f, the DOS agrees well with that of pristine Bernal stacked trilayer over a broad energy range, yet additional peaks emerge at a few energies.

We have verified that the oscillations in Figure 1e are not artifacts of incomplete convergence (see Methods). They therefore indicate that the electronic structure of TMB differs qualitatively from those of pristine trilayer and ATT systems. This is unexpected because the twist angle is small enough to avoid moiré-driven band renormalization, suggesting another origin.

Oscillatory density of states in triangular domains

To elucidate origin of the intriguing oscillations, we examine the local density of states (LDOS) of ρA(E)\rho_A(E), projected onto specific regions A, indicated by dashed lines in Figure 1c. In Figure 2, we present the enlarged views of the ρA(E)\rho_A(E), together with its normalized value defined as ΔρA(E)=(ρA(E)−ρR(E))/ρR(E)\Delta\rho_A(E) = (\rho_A(E) - \rho_R(E))/\rho_R(E) for various local domains of the TMB graphene across energy ranges 1 and 2. Here, ρR(E)\rho_R(E) denotes the reference DOS of pristine BTG (RTG) for the blue (red) triangular domain in Figure 1c.

Dichotomy of local density of states in triangular domains

Figure 2. Dichotomy of local density of states in triangular domains. a, The LDOS (ρA(E)\rho_A(E)) on the triangular region (Figure 1c) of the rhombohedral domain in energy range 1 is shown as a solid line. For comparison, the reference DOS of the RTG is plotted by a gray dash-dot line. b, The normalized deviation of LDOS in a from the reference DOS, ΔρA(E)\Delta\rho_A(E) (defined in the text). c and d, The same quantities as in a and b, but for the Bernal domain in Figure 1c. e and f, The same data as in a and b, but over energy range 2 for the rhombohedral domain. g and h, Analogously, these panels show the Bernal domain LDOS over the range 2.

The oscillatory behavior of ΔρA(E)\Delta\rho_A(E) reveals a clear dichotomy between the two domains. In range 1, the Bernal domain shows pronounced oscillations that persist throughout the energy range (Figure 2c and Figure 2d) while rhombohedral domain does not (Figure 2a and Figure 2b). Beyond the first vHS, the behavior reverses. In range 2, the Bernal domain becomes featureless (Figure 2g and Figure 2h), while the rhombohedral domain develops similarly strong oscillations (Figure 2e and Figure 2f). We therefore identify the DOS oscillations in Figure 1e as originating from the Bernal domain in range 1 and from the rhombohedral domain in range 2. By contrast, no discernible oscillatory features appear in range 3 (Figure 1e). As discussed, these contrasting behaviors are closely linked to the different numbers of available bands in each domain at a given energy.

As mentioned already, two connected conductors with different numbers of bands have perfect reflection channels at their interface [25]. In range 1, a Bernal domain has one more conducting channel than a rhombohedral domain, so one BTG channel should undergo perfect reflection (See Methods). In the layer-resolved LDOS shown in Extended Data Fig. 1, the middle layer shows no oscillations, whereas the top and bottom layers do. This strongly suggests that the confined state is the massless odd mode in ABA stacking, since this mode resides only on the top and bottom layers [19]. As will be demonstrated below, this is indeed the case. In range 2, confinement shifts to the RTG. Here, the perfectly reflected channel arises from hybridized bands rather than a single band (not shown here). In the following, we focus on range 1, where massless-particle confinement occurs.

The LDOS in Figure 2d exhibits four notable features. First, the peak spacing in range 1 is remarkably uniform, with an average value of 13.60 ± 0.91 meV. Second, the noticeable oscillations in quantized spectra emerge only away, about ±60 meV, from EFE_F as shown in Figure 2c, not starting from EFE_F. Third, beyond the primary oscillation, the peak amplitudes exhibit a slower beating-like envelope modulation. Fourth, the DOS between peaks remains only weakly suppressed, in sharp contrast to conventional quantum dots [26]. As will be shown, these features are hallmarks of quantization of massless fermions confined in a triangular domain of TTG.

In the ATT system, the LDOS within the Bernal domains exhibits no discernible oscillations (Extended Data Fig. 2a and b). By contrast, ΔρA(E)\Delta\rho_A(E) projected onto the domain boundaries displays several pronounced peaks (Extended Data Fig. 2c and d), indicating that the deviations observed in Figure 1f originate primarily from the domain boundaries rather than the domain interiors. For both TMB and ATT systems, ΔρA(E)\Delta\rho_A(E) projected onto the vertex regions exhibits several strong peaks near EFE_F, accounting for features that deviate from the otherwise boundary-state signatures and regular oscillatory behavior (Extended Data Fig. 2e and Extended Data Fig. 3c).

Charge densities of confined triangular states

To visualize the confined states, we calculate the energy- and layer-resolved charge density ρE(r)\rho_E(\mathbf{r}) for several peaks of the oscillatory DOSs (see Methods). Figure 3 show ρE(r)\rho_E(\mathbf{r}) for the first nine consecutive peaks in Figure 2d. We display a sum of top and bottom layers for each peak since the massless states are only on these layers not on the middle one. The detailed ρE(r)\rho_E(\mathbf{r}) for each layer and corresponding scanning tunneling microscopy image simulations can be found in Extended Data Figure 4.

Energy- and layer-resolved electron density distributions of the TMB system

Figure 3. Energy- and layer-resolved electron density distributions of the TMB system. The energy level EE above each panel is where the density ρE(r)\rho_E(\mathbf{r}) is evaluated for peaks of oscillatory DOS in Figure 2d, where the first three peaks corresponding to a-c are indicated by markers. ρE(r)\rho_E(\mathbf{r}) represents the sum of the layer resolved densities of layer 1 and layer 3. The dotted line triangle indicates the Bernal domain. The colormap range is individually scaled to the minimum and maximum density values of each panel. Note that the minimum is a finite value as the system is semimetallic. Each panel’s a ratio of MIN/MAX value is: a 0.505, b 0.521, c 0.645, d 0.645, e 0.593, f 0.570, g 0.555, h 0.559, and i 0.601.

Layer- and energy-resolved electron density distributions of the TMB system

Extended Data Figure 4. Layer- and energy-resolved electron density distributions of the TMB system. a–f Energy-resolved total electron density distributions including contributions from all three layers, ρET(r)=ρE1(r)+ρE2(r)+ρE3(r)\rho_{E}^{T}(r)=\rho_{E}^{1}(r)+\rho_{E}^{2}(r)+\rho_{E}^{3}(r) where ρEi(r)\rho_{E}^{i}(r) denotes the energy-resolved density distribution of layer ii (i=1,2,3i=1,2,3). These energies correspond to the first six panels for ρE(r)=ρE1(r)+ρE3(r)\rho_{E}(r)=\rho_{E}^{1}(r)+\rho_{E}^{3}(r) in Figure 3. All panels within a given row are evaluated at the same energy EE, indicated on the leftmost side of the row. The colormap in each panel is independently normalized to its own minimum and maximum density values. Each panel’s MIN/MAX value is: a 0.629, b 0.606, c 0.694, d 0.676, e 0.673, f 0.666. g–l Layer-resolved decomposition of the total energy-resolved density distributions shown in a–f. Each panel contains the contributions from the three individual layers. Within a given panel, all three subpanels share a common colormap range determined by the global minimum and maximum among them, whereas different panels are normalized independently. Each panel’s MIN/MAX value is: g 0.475, h 0.489, i 0.555, j 0.565, k 0.594, and l 0.586. m-r Simulated scanning tunneling microscopy (STM) topographies obtained from the layer-resolved densities following [48]. We employ the constant-current mode with decay length λ=5\lambda=5 Å and interlayer spacing d=3.3724d=3.3724 Å, assuming an STM current at height zz of I(r,z)∝∑i=13ρEi(r)e−(z+(i−1)d)/λI(r,z)\propto\sum_{i=1}^{3}\rho_{E}^{i}(r)e^{-(z+(i-1)d)/\lambda}. We consider the case where the twisted monolayer (layer 1) is the topmost layer.

As shown in Figure 3 and Extended Data Figure 4, the Bernal domain exhibits pronounced nodal patterns, with the number of nodes increasing with energy, providing direct evidence of quantum confinement. Notably, the effective area occupied by the confined charge density is smaller than the geometric boundary between the two stacking domains (indicated by the white dotted lines) and expands as the state energy moves away from EFE_F. At the same time, boundary charge accumulation is observed along the geometric boundaries within the middle layer of the Bernal domain, as shown in Extended Data Figure 4–Extended Data Figure 4, of which origin will be discussed later.

Unlike the well-defined confined states in Bernal domain, the rhombohedral domain displays more uniform wave patterns, consistent with hybridized states with neighboring domains rather than confined states. We further find that the largest amplitudes for first few low-energy states are concentrated near the vertices, and gradually diminish as the energy moves away from the EFE_F.

Similarly, Extended Data Figure 5 shows ρE(r)\rho_E(\mathbf{r}) corresponding to DOS peaks in range 2 of Figure 2f. These states also clearly exhibit a strong confinement effect, but now confined only within the rhombohedral domain. We find that all the other DOS peaks also closely follow the nodal structures expected from confinement inside triangular domain as well as the confinement rules dictated by excess number of bands discussed above. In addition, we identify the boundary and vertex states contributing a couple of LDOS peaks of TMB and ATT systems as shown in Extended Data Figure 6 and Extended Data Figure 7, respectively.

Energy-resolved electron density distributions of the TMB system in range 2

Extended Data Figure 5. Energy-resolved electron density distributions ρET(r)\rho_E^T(\mathbf{r}) of the TMB system in range 2. The energy levels EE above panels a-h are where the density ρET(r)\rho_E^T(\mathbf{r}) is evaluated for and E±3E \pm3 meV ranges are shown by shades in i, which is identical data as Figure 2f, the rhombohedral domain LDOS deviation. For the ρET(r)\rho_E^T(\mathbf{r}) in range 2, the contributions from all three layers are included. The dotted line triangle indicates the rhombohedral domain. The maximum and minimum of the colormap correspond to each panel’s maximum and minimum value of density. Each panel’s MIN/MAX value is: a 0.869 , b 0.853 , c 0.833 , d 0.817 , e 0.823 , f 0.821 , g 0.816 , and h 0.815.

Energy-resolved electron density distributions of the TMB system

Extended Data Figure 6. Energy-resolved electron density distributions ρE(r)\rho_E(\mathbf{r}) of the TMB system. The energy level EE above each panel is where the density ρE(r)\rho_E(\mathbf{r}) is evaluated. Those EE values are also indicated in Extended Data Fig. 3 a-c. In this figure, the contributions from every layer are included. The sparse dotted line indicates the domain boundaries. Each panel’s MIN/MAX value is: a 0.256, b 0.213, and c 0.161.

Energy-resolved electron density distributions of the ATT system

Extended Data Figure 7. Energy-resolved electron density distributions ρE(r)\rho_E(\mathbf{r}) of the ATT system. The energy level EE above each panel is where the density ρE(r)\rho_E(\mathbf{r}) is evaluated, and the following σ\sigma determines the width of the Gaussian weight for the integral. Those EE values are also indicated in Extended Data Fig. 2 c,d for the boundary-localized densities c-f of this figure, and in Extended Data Fig. 2 e for the vertex-localized densities a,b of this figure. Note that among the boundary localized densities, c shows a single-wall shape, d,f a double-wall shape, which resembles a boundary state reported for bilayer graphene [49], and e a triple-wall shape. In this figure, the contributions from every layer are included. The sparse dotted line indicates the domain boundaries. Each panel’s MIN/MAX value is: a 0.062, b 0.475, c 0.561, d 0.278, e 0.388, and f 0.245.

Boundary as adiabatic confinement barrier

To analyze the quantized states in Bernal domains, we consider massless Dirac fermions confined to an equilateral triangle. The nonrelativistic counterpart has a long history, dating back to Lamé [27] and later approaches [22, 23, 28, 29]. The confined states found here are massless, however, so Schrödinger solutions with Dirichlet boundaries do not apply. Our simulations reveal out-of-plane boundary corrugations below 0.02 Å and in-plane profiles extending about 124 Å (Extended Data Figure 8), ruling out valley mixing [30, 31]. Moreover, no infinite mass exists outside the domain, unlike the Berry–Mondragon (BM) condition [12, 32] for neutrino billiards or the MIT bag model [13] for hadron confinement.

Height corrugations across the boundaries

Extended Data Figure 8. Height corrugations across the boundaries. a, The spatial variation of total thickness of the TMB for the same area as Figure 1c. A solid (dotted) arrow indicate the line along which height variations of each layer are plotted. b and c, From the bottom to top panels, the height variations (did_i, i=1,2,3i = 1, 2, 3) across the boundaries along the solid and dotted arrows in a. d, Along the solid arrow in a, the variation of interlayer distance between the bottom and middle (Δd12=d1−d2−dint0\Delta d_{12} = d_1 - d_2 - d_{\mathrm{int}}^0) and one between the middle and top layer (Δd23=d2−d3−dint0\Delta d_{23} = d_2 - d_3 - d_{\mathrm{int}}^0) with respect to those (dintd_{\mathrm{int}}) at the center of domain. e, The same as d, along the dotted arrow in a.

The confinement found in our simulations follows a different route from established ways of trapping a massless Dirac particle [11, 12, 30, 31]. We find that the boundary here does not add any mass to the odd ABA state (See Methods for proof). Instead, it turns out that the absence of an adiabatic continuation is origin of massless particle confinement. On the ABA side there are two propagating channels, but on the ABC side only one. We explicitly show that this single exit carries the massive even ABA mode adiabatically across the smooth wall (See details in Methods). By unitarity, the massless mode has nowhere to go and turns back. Reflection therefore follows from channel counting and smoothness, becoming exact at every incidence angle in the adiabatic limit.

Two unusual features follow. First, confinement improves with wall smoothness, while a sharp interface would leak appreciably (Methods). The lattice relaxation that creates the domains is therefore also what makes them quantum cavities. Second, the trapped fermions retain their linear dispersion, so the cavity is a true massless billiard with a current-free boundary (Methods). Although the microscopic mechanism is entirely different, the BM condition [12] emerges here as a phenomenological description of our adiabatic confinement.

Ideal Dirac Billiard and Eisenstein Spectrum

To reveal the spectral consequences of this confinement mechanism, we first consider massless Dirac fermions in an equilateral triangle of side LML_M subject to the BM boundary condition. The resulting ideal spectrum provides a natural reference for understanding how confinement is reshaped within the actual moiré domains.

The eigenproblem is H△Ψ=εΨH_{\triangle}\Psi= \varepsilon\Psi, with H△=−iℏvF(σx∂x+σy∂y)H_{\triangle} = -i\hbar v_F(\sigma_x\partial_x + \sigma_y\partial_y), where Ψ\Psi is a two-component sublattice spinor, vFv_F is the Fermi velocity, and σx,y\sigma_{x,y} are Pauli matrices acting in sublattice space. Smooth boundaries allow us to consider a single valley. The positive-energy eigenstates Ψmn\Psi_{mn} have energies εmn=ε0∣zmn∣\varepsilon_{mn} = \varepsilon_0|z_{mn}| where

ε0=4πℏvF3LM≃18.82 meV,zmn=m+nω\varepsilon_0 = \frac{4\pi\hbar v_F}{3L_M} \simeq18.82\ \mathrm{meV},\quad z_{mn} = m+n\omega

is an Eisenstein integer, ω=e2πi/3\omega= e^{2\pi i/3}, and ∣zmn∣2=m2−mn+n2|z_{mn}|^2 = m^2-mn+n^2 [32, 33]. The BM boundary condition selects the integer pairs D={(m,n)∣m>n≥0}D = \{(m,n)\mid m>n\ge0\}, which occupy one-sixth of the triangular lattice of Eisenstein integers, as shown in Figure 4a (see Methods).

Bound massless states in a triangular cavity and associated density of states and local density of states

Figure 4. Bound massless states in a triangular cavity. a, Eisenstein numbers, zmn=m+nωz_{mn}=m+n\omega, satisfying m>n≥0m>n\geq0, are marked by filled circles in shaded region of the complex plane. The zmnz_{mn} on the excluded (m=nm=n) and included (n=0n=0) boundaries are denoted by the empty circles and the blue squares, respectively. b, d(ϵ)d(\epsilon) is a number of degeneracy for a given energy ϵ\epsilon (upper panel) and DOS (low panel) as a function of ϵ\epsilon, D△(ϵ)D_{\triangle}(\epsilon). The degeneracy d(ϵ)d(\epsilon) originating from the filled circles (blue squares) in a are shown by black solid (blue dotted) lines. The solid line is computed by a Gaussian convolution (σ=0.2\sigma=0.2) of d(ϵ)d(\epsilon) and the dotted line is from Eq. (3) with the same σ\sigma. In a and b, the red solid lines represent the quasiperiodic levels ϵ=M3/2+3/16\epsilon=M\sqrt{3}/2+\sqrt{3}/16 (M=1,2,3,⋯M=1,2,3,\cdots). c, The normalized deviation of LDOS of BTG domain from simulations (solid line) in Figure 2d and from our model (dotted line). d, ρ△(r;ϵ)\rho_{\triangle}(\mathbf{r};\epsilon) of the first nine DOS peaks shown in b arranged in ascending energy order from the left to right. The densities are integrated using a Gaussian weight (σ=0.2\sigma=0.2) centered at the energy of the peaks.

This correspondence places arithmetic structure of Eisenstein number directly within the confinement spectrum [22, 23, 24]. Using the dimensionless energy ε=E/ε0\varepsilon= E/\varepsilon_0, the discrete DOS is d(ε)=∑(m,n)∈Dδ(ε−∣zmn∣)d(\varepsilon) = \sum_{(m,n)\in D}\delta(\varepsilon-|z_{mn}|). Although its level multiplicities appear irregular in the upper panel of Figure 4b, Gaussian broadening reveals a clear oscillatory pattern. To understand its origin, we introduce the counting function n△(ε)n_{\triangle}(\varepsilon), the number of allowed Eisenstein integers satisfying ∣zmn∣≤ε|z_{mn}|\le\varepsilon. The relation d(ε)=∂εn△(ε)d(\varepsilon)=\partial_{\varepsilon}n_{\triangle}(\varepsilon) then connects the DOS to the distribution of lattice points within a circle. Using the prime-factorization properties of zmnz_{mn} [22, 24], we obtain the following approximation for the Gaussian-broadened DOS,

d△(ε)≃4π3ε3[16+e−8π2σ2/3J0(4π3ε)],(1)d_{\triangle}(\varepsilon)\simeq\frac{4\pi}{\sqrt{3}\varepsilon^3}\left[\frac{1}{6}+e^{-8\pi^2\sigma^2/3}J_0\left(\frac{4\pi}{\sqrt{3}\varepsilon}\right)\right], \tag*{(1)}

where σ\sigma is the dimensionless Gaussian width and J0J_0 is the zeroth-order Bessel function of the first kind (see Methods). The linear background reflects the massless dispersion, while the Bessel term produces confinement oscillations with an asymptotic energy spacing of 3ε0/2\sqrt{3}\varepsilon_0/2. These features capture the principal structure of the simulated DOS in Figure 2c, linking its oscillations to the arithmetic of the triangular confinement spectrum.

Self-energy corrections from leakages near EFE_F

The actual moiré domains, however, depart from isolated triangular cavities [32, 33] owing to two energy-dependent leakage channels. First, one arises from ABB-stacked regions of area ∼l2\sim l^2 at the corners of the domains, that depends on the overlap between ABB states and the massless mode of ABA stacking. Below Ec=44.7E_c=44.7 meV, the relevant ABB bands have strong odd character, with ⟨Po⟩=0.8\langle P_o\rangle=0.8–0.90.9, where PoP_o projects onto the ABA mirror-odd subspace (Extended Data Figs. 9a and b; see Methods). Their momenta are also comparable to those of the incident massless mode, providing a matching channel through which carriers can escape. We therefore expect corner leakage to obstruct confinement below EcE_c. At this energy, the matching ABB band reaches its maximum and the channel closes. Above EcE_c, the remaining states lie on an outer branch with momenta several times larger than those of the incident mode and max⁡⟨Po⟩≃0.05\max\langle P_o\rangle\simeq0.05. This mismatching suggests a reduction in corner transmission by more than an order of magnitude, allowing well-defined cavity resonances to emerge.

The other involves modes propagating parallel to the triangle sides, which can couple to massive states without undergoing reflection from the boundaries. This residual leakage channel remains for the n=0n=0 modes, marked by blue rectangles in Figure 4a. These modes propagate parallel to the triangle sides without encountering the wall. Nevertheless, the varied interlayer coupling within the boundary mixes them with middle-layer states and thereby connects them to the massive branch. Only about 26% of their weight remains within the domain (see Methods). Their coupling to massive states is also visible in the enhanced line-shaped intensity along the boundaries in layer 2 of ρE(r)\rho_E(\mathbf{r}) (Extended Data Figure 4g–Figure 4l).

To incorporate these departures from ideal confinement, we adopt a minimal affine self-energy that shifts and rescales the bare spectrum,

Emn≃Ec+4π3ℏveffLMm2−mn+n2,(2)E_{mn}\simeq E_c+\frac{4\pi}{3}\frac{\hbar v_{\mathrm{eff}}}{L_M}\sqrt{m^2-mn+n^2}, \tag*{(2)}

where veff=ZvFv_{\mathrm{eff}}=Zv_F and Z(<1)Z(<1) is the renormalized factor due to low energy leakage sources. Here, EcE_c serves as the spectral offset associated with the suppression of corner leakage (see Methods). With the renormalized dimensionless energy ε~=E−EcZε0\tilde{\varepsilon}=\frac{E-E_c}{Z\varepsilon_0} and the retained weight Aw=0.26A_w=0.26 of the n=0n=0 modes, we obtain the renormalized DOS,

D△(ε~)≃Zd△(ε~)+Z(Aw−1)bwall(ε~),(3)D_{\triangle}(\tilde{\varepsilon})\simeq Zd_{\triangle}(\tilde{\varepsilon})+Z(A_w-1)b_{\mathrm{wall}}(\tilde{\varepsilon}), \tag*{(3)}

where bwall(ε~)≃1+2e−2π2σ2cos⁡(2πε~)b_{\mathrm{wall}}(\tilde{\varepsilon})\simeq1+2e^{-2\pi^2\sigma^2}\cos(2\pi\tilde{\varepsilon}) is the DOS of the boundary-parallel modes. The second term reduces their contribution to account for leakage (See Methods for detailed derivations), such that Eq. (3) closely reproduces the simulated DOS in Figure 4b and yields the normalized oscillatory contribution ΔρA(E)\Delta\rho_A(E). The fast oscillation has a periodicity of 32Zϵ0≃13.6\frac{\sqrt{3}}{2}Z_{\epsilon_0}\simeq13.6 meV, while its superposition with the n=0n=0 modes produces a slower beating pattern. Both features agree quite well with our simulations, as shown in Figure 4b and Figure 4c. Leakage thus modifies the weights and energy scale of the resonances while preserving the arithmetic structure underlying the DOS.

The same description also accounts for the spatial signatures of confinement. Each eigenstate gives a probability density ρmn(x,y)=∣Ψmn(x,y)∣2\rho_{mn}(x,y)=|\Psi_{mn}(x,y)|^2 (see Methods for explicit forms). Summing these densities over the allowed integer pairs, with Gaussian broadening centered on the corresponding energies EmnE_{mn}, yields the energy-resolved distribution ρ△(r;E)\rho_{\triangle}(\mathbf{r};E) (see Methods). The resulting patterns in Figure 4c closely reproduce those obtained from the atomistic simulations in Figure 3, connecting the arithmetic organization of the spectrum to the real-space structure of the confined Dirac states.

Discussions

Taken together, our million-atom simulations and analytical modelling establish that massless Dirac fermions can be confined within metallic domains without surrounding semiconducting barriers. Smooth stacking variations act as internal reflectors to massless carriers when a mirror-protected mode loses its propagating continuation and conversion into other channels is suppressed. The diverse energy–momentum dispersions of multilayer graphene [19, 34] suggest that this mechanism may extend to other multilayer moiré systems with suitable channel structures. Twisted trilayers already host a rich variety of stacking domains [2, 17], offering a natural setting in which geometry and quasiparticle character can be brought together to design confinement spectra.

Considering that manipulations of local and global boundary shapes [35, 36] are already demonstrated, this engineering possibility raises the prospect of reshaping quantum cavities within a continuous crystal. Deforming triangular domains could reveal how the degeneracies organized by Eisenstein integers evolve with geometry, while suitable boundary shapes could open a window onto relativistic quantum chaos [12, 37]. Spatially resolved tunnelling spectroscopy could follow this evolution through discrete resonances and their local DOS patterns. Tracking the level spacings as domain size varies would further test the geometric origin of quantization, bringing the relation between shape, arithmetic and quantum dynamics within experimental reach.

Methods

Computational methods

The fully relaxed atomic structures were calculated using the interatomic potential method developed for efficient multi-million atom simulations [20]. For intralayer interactions, the harmonic potential was fitted against the phonon dispersion obtained from the first principles calculations [17]. For interlayer interactions, we adopted the Kolmogorov-Crespi potential [38] combined with long-range inter-atomic forces across all three layers, which is necessary to reproduce the energetic relation between the different stacking types [17].

Electronic structures were calculated using the large-scale tight-binding (TB) method used in our previous work [17, 21]. For the TB Hamiltonian model, we followed the previous study [17], where the Slater-Koster hopping parameter between the pzp_z orbitals of two carbon atoms displaced by r\mathbf{r} is given as t(r)=n2Vσ(r)+(1−n2)Vπ(r)t(\mathbf{r}) = n^2 V_{\sigma}(\mathbf{r}) + (1-n^2)V_{\pi}(\mathbf{r}), where n=z/rn=z/r (r=∣r∣r=|\mathbf{r}|). The decaying hopping strengths are Vσ(π)(r)=γσ(π)exp⁡[qσ(π)(1−r/aσ(π))]V_{\sigma(\pi)}(r) = \gamma_{\sigma(\pi)} \exp[q_{\sigma(\pi)}(1-r/a_{\sigma(\pi)})], where γσ=0.48\gamma_{\sigma}=0.48 eV, γπ=−2.7\gamma_{\pi}=-2.7 eV, aσ=3.349a_{\sigma}=3.349 Å, aπ=1.418a_{\pi}=1.418 Å, and qπ/aπ=qσ/aσ=2.218q_{\pi}/a_{\pi}=q_{\sigma}/a_{\sigma}=2.218 Å−1^{-1} [17]. The nearest neighbors are considered for the intralayer hoppings, whereas the hoppings within the cut-off distance 5 Å are considered for the interlayer interactions (the nearest interlayer hopping is given by γ1=390\gamma_1=390 meV). We find that increasing cut-offs only changes the very low energy bands near the Fermi level (EFE_F), not altering any signature above. The ARPACK iterative solver [39, 40] is adopted to obtain the eigen-solutions of the multi-million-dimensional TB Hamiltonian matrix. We examined only certain energy ranges near EFE_F instead of the full energy spectrum. This is possible by performing the shift-invert mode calculation of the Lanczos scheme.

To obtain a high-resolution density of states (DOS), we calculated eigenstates on a 6×66 \times6 Γ\Gamma-centered kk-space grid for twisted trilayer graphene (TTG) with a moiré unit-cell length of 1280.13 Å. The DOS was obtained by applying Gaussian broadening to the eigenenergies with a broadening factor of σ=2\sigma=2 meV. By varying both the kk grid and σ\sigma, we carefully tested the convergence of the DOS to exclude spurious wiggles arising from insufficient eigenenergy sampling, while also avoiding artificial smoothing of genuine DOS variations. We find that the domain-resolved local DOS is well converged to the untwisted reference at σ=2\sigma=2 meV, while keeping its physical fine structures. In contrast, the larger broadening of 6 meV can suppress the physical oscillations [17].

The Gaussian convoluted energy-resolved spatial density around EE is defined as ρE(r)=∑n∫BZwσ(E−Enk)∣ψEnk(r)∣2d2k\rho_E(\mathbf{r}) = \sum_n \int_{\mathrm{BZ}} w_{\sigma}(E-E_{n\mathbf{k}})|\psi_{E_{n\mathbf{k}}}(\mathbf{r})|^2d^2k where wσ(E)=12πσe−E2/2σ2w_{\sigma}(E) = \frac{1}{\sqrt{2\pi}\sigma}e^{-E^2/2\sigma^2} and σ=3\sigma=3 meV is used unless specified elsewhere.

Stacking order parameter for trilayer graphene

We quantify the stacking orders for arbitrary trilayer stackings by approximating the local structure as a relative sliding of untwisted layers. Following a previous work for twisted bilayer graphene [41], we first introduce four parameters: u1B2Au_{1B}^{2A}, u1A2Bu_{1A}^{2B}, u2B3Au_{2B}^{3A}, and u2A3Bu_{2A}^{3B}, where numbers denote layers (1: bottom, 2: middle, 3: top), and A/B are sublattices. Each represents the shortest in-plane distance from the atom labeled by the superscript to the atom by the subscript (e.g., u1B2Au_{1B}^{2A} is the shortest in-plane distance from middle layer atom on A sublattice to bottom layer atom on B sublattice). They are evaluated at each atom and interpolated on a regular grid. For unstrained unit-cell parameters, the maximum value should be the bond length of the hexagonal lattice, umax⁡=a0u_{\max}=a_0. The local trilayer stacking order are then,

uABC=(1−u3B3A/umax⁡)(1−u1B2A/umax⁡)u_{\mathrm{ABC}}=(1-u_{3B}^{3A}/u_{\max})(1-u_{1B}^{2A}/u_{\max})
uACB=(1−u3B2A/umax⁡)(1−u1A2B/umax⁡)u_{\mathrm{ACB}}=(1-u_{3B}^{2A}/u_{\max})(1-u_{1A}^{2B}/u_{\max})
uABA=(1−u3A2B/umax⁡)(1−u1A2B/umax⁡)u_{\mathrm{ABA}}=(1-u_{3A}^{2B}/u_{\max})(1-u_{1A}^{2B}/u_{\max})
uACA=(1−u3A2B/umax⁡)(1−u1B2A/umax⁡)u_{\mathrm{ACA}}=(1-u_{3A}^{2B}/u_{\max})(1-u_{1B}^{2A}/u_{\max})

Each equals 1 in its corresponding stacking and 0 when another parameter reaches 1. To distinguish rhombohedral-stacked trilayer graphene (RTG) from Bernal trilayer graphene (BTG), we define

uRB=uABC+uACB−uABA−uACA.u_{\mathrm{RB}}=u_{\mathrm{ABC}}+u_{\mathrm{ACB}}-u_{\mathrm{ABA}}-u_{\mathrm{ACA}}.

which equals +1(-1) for RTG (BTG). Here each subscript denotes a specific stacking sequence of trilayer.

Existence of perfectly reflected channels

Let us suppose that BTG and RTG occupy the regions x<−lx<-l and x>lx>l, respectively, with a smooth boundary in ∣x∣≤l|x|\leq l that is uniform along yy. Then kyk_y is conserved, reducing scattering at fixed EE and kyk_y to a one-dimensional problem. From Extended Data Figs. 8b and d, the full width at half maximum of the stacking gradient is about 124 Å, so l≃62l\simeq62 Å. In the range 1 of Figure 1a, the BTG on the left has two propagating channels (nL=2n_L=2) while the RTG on the right has one (nR=1n_R=1). One can construct a scattering matrix SS, defined by ψout=Sψin\psi_{\mathrm{out}}=S\psi_{\mathrm{in}}, that connects the states incoming to the boundary, ψin\psi_{\mathrm{in}}, to the states outgoing from it, ψout\psi_{\mathrm{out}} [42]. In matrix form,

S=(rt′tr′).S=\begin{pmatrix} r & t' \\ t & r' \end{pmatrix}.

Here, ψin=(ϕ1L,…,ϕnLL,ϕ1R,…,ϕnRR)T\psi_{\mathrm{in}}=(\phi_1^L,\ldots,\phi_{n_L}^L,\phi_1^R,\ldots,\phi_{n_R}^R)^{\mathrm{T}} and ψout=(φ1L,…,φnLL,φ1R,…,φnRR)T\psi_{\mathrm{out}}=(\varphi_1^L,\ldots,\varphi_{n_L}^L,\varphi_1^R,\ldots,\varphi_{n_R}^R)^{\mathrm{T}}, where ϕcα\phi_c^{\alpha} (φcα\varphi_c^{\alpha}) is the amplitude of the cc-th incoming (outgoing) channel of region α=L,R\alpha=L,R (c=1,…,nαc=1,\ldots,n_{\alpha}). We also note that S†S=SS†=1nL+nRS^{\dagger}S=SS^{\dagger}=\mathbb{1}_{n_L+n_R}. The channels are flux-normalised so that ∣ϕcα∣2|\phi_c^{\alpha}|^2 is the probability flux carried by that channel [42]. Accordingly, rr (r′r') is an nL×nLn_L\times n_L (nR×nRn_R\times n_R) reflection matrix, and tt (t′t') is the nR×nLn_R\times n_L (nL×nRn_L\times n_R) transmission matrix from the left to the right (from the right to the left) side. Then, the matrix tt has rank at most min⁡(nL,nR)\min(n_L,n_R), which equals nRn_R in range 1. By the rank-nullity theorem applied to t†t^{\dagger}, dim⁡ker⁡(t†)=dim⁡ker⁡(t)+rank⁡(t)=nL\dim\ker(t^{\dagger})=\dim\ker(t)+\operatorname{rank}(t)=n_L, so dim⁡ker⁡(t)≥nL−nR\dim\ker(t)\geq n_L-n_R. Owing to this, at least nL−nRn_L-n_R linearly independent incoming states ϕ~L\tilde{\phi}^L, each a linear combination of the ϕcL\phi_c^L, should satisfy tϕ~L=0t\tilde{\phi}^L=0. For such a state, unitarity gives ∣rϕ~L∣2+∣tϕ~L∣2=∣ϕ~L∣2|r\tilde{\phi}^L|^2+|t\tilde{\phi}^L|^2=|\tilde{\phi}^L|^2, hence ∣rϕ~L∣2=∣ϕ~L∣2|r\tilde{\phi}^L|^2=|\tilde{\phi}^L|^2, i.e., a perfectly reflected state. Therefore, such a reflected state always exists on the side with more channels, whenever nL≠nRn_L\ne n_R. From Figure 1a, one finds that in range 1 (range 2) the BTG (RTG) side has one perfectly reflected state. We note that this conclusion does not rest on any assumption about the Hamiltonians of the domains or the profile of the boundary.

Adiabatic confinement of massless fermions

Here, we show that the massless fermion in ABA stacking domain is the perfectly reflected state shown above. In one valley and in the sublattice basis (∣A1⟩,∣B1⟩,∣A2⟩,∣B2⟩,∣A3⟩,∣B3⟩)(|A_1\rangle, |B_1\rangle, |A_2\rangle, |B_2\rangle, |A_3\rangle, |B_3\rangle) where ∣A(B)α⟩|A(B)_\alpha\rangle (α=1,2,3\alpha= 1, 2, 3) is an orthonormal pzp_z-orbital basis state at the sublattice A(B)A(B) of α\alpha-th layer, the Schrödinger equation of the junction geometry discussed above is given by

H[λ(x)]Ψ=[HD+H⊥ABA+λ(x)ΔH]Ψ=EΨ,(4)H[\lambda(x)]\Psi= \left[H_D + H_{\perp}^{\mathrm{ABA}} + \lambda(x)\Delta H\right]\Psi= E\Psi, \tag*{(4)}

where λ(x)=12[1+tanh⁡(x/l)]\lambda(x) = \frac{1}{2}[1+\tanh(x/l)], HD=ℏv(kxσx+kyσy)⊗13H_D = \hbar v(k_x\sigma_x+k_y\sigma_y)\otimes\mathbf{1}_3, H⊥ABA=γ1(∣B1⟩⟨A2∣+∣A2⟩⟨B3∣+h.c.)H_{\perp}^{\mathrm{ABA}} = \gamma_1(|B_1\rangle\langle A_2|+|A_2\rangle\langle B_3|+\mathrm{h.c.}), HABC=γ1(∣B1⟩⟨A2∣+∣B2⟩⟨A3∣+h.c.)H^{\mathrm{ABC}} = \gamma_1(|B_1\rangle\langle A_2|+|B_2\rangle\langle A_3|+\mathrm{h.c.}), and ΔH=HABC−H⊥ABA=γ1(∣B2⟩⟨A3∣−∣A2⟩⟨B3∣+h.c.)\Delta H = H^{\mathrm{ABC}}-H_{\perp}^{\mathrm{ABA}} = \gamma_1(|B_2\rangle\langle A_3|-|A_2\rangle\langle B_3|+\mathrm{h.c.}). Then, we confirm that, for λ(x=+∞)=1\lambda(x=+\infty)=1, H→HABC=HD+HABCH\to H^{\mathrm{ABC}}=H_D+H^{\mathrm{ABC}} and, for λ(x=−∞)=0\lambda(x=-\infty)=0, H→HABA=HD+H⊥ABAH\to H^{\mathrm{ABA}}=H_D+H_{\perp}^{\mathrm{ABA}}. We note that ΔH\Delta H contains no momentum and has the coupling between the states only in layer 2 and 3.

In the basis ∣L±⟩=(∣L1⟩±∣L3⟩)/2|L_\pm\rangle=(|L_1\rangle\pm|L_3\rangle)/\sqrt{2} for a mirror operator MzM_z satisfying Mz∣L±⟩=±∣L±⟩M_z|L_\pm\rangle=\pm|L_\pm\rangle (L=A,BL=A,B), the spatially varying part in Eq. (4) is rewritten as

λ(x)ΔH=−gλ(x)(∣B2⟩⟨A−∣−∣A2⟩⟨B−∣)+gλ(x)(∣B2⟩⟨A+∣−∣A2⟩⟨B+∣)+h.c.,(5)\lambda(x)\Delta H=-g\lambda(x)(|B_2\rangle\langle A_-|-|A_2\rangle\langle B_-|)+g\lambda(x)(|B_2\rangle\langle A_+|-|A_2\rangle\langle B_+|)+\mathrm{h.c.}, \tag*{(5)}

where g=γ1/2≃276 meVg=\gamma_1/\sqrt{2}\simeq276\ \mathrm{meV}. Let P±P_\pm be a project operator on its even (odd) or +1(−1)+1(-1) eigenspaces. HABAH^{\mathrm{ABA}} commutes with MzM_z, so in the ABA domain the odd block is exactly PoHABAPo=ℏv(kxσx+kyσy)P_oH^{\mathrm{ABA}}P_o=\hbar v(k_x\sigma_x+k_y\sigma_y) [19]. Inside the wall, ΔH\Delta H breaks MzM_z, but Eq. (5) carries at most one odd index, so PoΔHPo=0P_o\Delta H P_o=0. Then, H[λ]H[\lambda] in Eq. (4) satisfies PoH[λ]Po=ℏv(kxσx+kyσy)P_oH[\lambda]P_o=\hbar v(k_x\sigma_x+k_y\sigma_y), for every λ(x)\lambda(x) in the junction region of ∣x∣<l|x|<l, implying that the junction adds no mass on the propagating massless state from BTG region. Instead, what the junction wall does to the odd mode modifies its kinematics as discussed below.

We note that ΔH\Delta H in Eq. (5) couples the odd sector only through the layer 2 with gλ(x)g\lambda(x). Since kyk_y is conserved, we can rewrite Eq. (4) with fixed EE and kyk_y as a stationary problem for evolution of states Ψ\Psi along xx,

∂xΨ=iK(x;E,ky)Ψ,(6)\partial_x\Psi=iK(x;E,k_y)\Psi, \tag*{(6)}

where K(x;E,ky)=1ℏvB[E−ℏvky(13⊗σy)−H⊥ABA−λ(x)ΔH]K(x;E,k_y)=\frac{1}{\hbar v}B[E-\hbar vk_y(\mathbf{1}_3\otimes\sigma_y)-H_{\perp}^{\mathrm{ABA}}-\lambda(x)\Delta H] and B=13⊗σxB=\mathbf{1}_3\otimes\sigma_x. Owing to pseudo-unitarity BK=K†BBK=K^\dagger B, the current J≡Ψ†BΨJ\equiv\Psi^\dagger B\Psi is conserved, and a local mode of unu_n satisfying Kun=knunKu_n=k_nu_n carries the flux jn=un†Bunj_n=u_n^\dagger Bu_n. Here, jnj_n is zero for complex knk_n and nonzero for a real knk_n, that is positive (negative) for a right (left) mover.

By expanding the state Ψ\Psi on the local modes un(x)u_n(x) of eigenstates of KK, Ψ=∑ncn(x)un(x)exp⁡(i∫xkn dx)\Psi=\sum_n c_n(x)u_n(x)\exp(i\int^x k_n\,dx), each propagating mode can be normalised to unit flux at every xx. Inserting this into Eq. (6), the knk_n terms cancel and projecting with the left eigenvector um†B/jmu_m^\dagger B/j_m gives the following coupled-mode equations,

∂xcm=−∑nDmnei∫x(kn−km) dxcn,(7)\partial_xc_m=-\sum_nD_{mn}e^{i\int^x(k_n-k_m)\,dx}c_n, \tag*{(7)}

where Dmn≡um†B∂xun/jmD_{mn}\equiv u_m^\dagger B\partial_xu_n/j_m. The reformulations of Eq. (4) to Eqs. (6) and (7) can be regarded as a real space generalization [45, 46] of the typical time-dependent non-adiabatic coupled-mode equation [47].

From the eigenvalue equation, Kun=knunKu_n=k_nu_n, the allowed propagating modes in the range 1 in Figure 1a can be determined through det⁡[K(x)−kn]=0\det[K(x)-k_n]=0. The characteristic polynomial of KK can be compactly written as

R(u)≡u[(u−c)2−2γ12E~2]−γ14E2λ2=0,(8)R(u)\equiv u\left[(u-c)^2-2\gamma_1^2\widetilde{E}^2\right]-\gamma_1^4E^2\lambda^2=0, \tag*{(8)}

where u=ℏ2v2∣k∣2−E2u=\hbar^2v^2|k|^2-E^2, c=γ12λ(1−λ)c=\gamma_1^2\lambda(1-\lambda) and E~2=E2−2λ(1−λ)ℏ2v2ky2\widetilde{E}^2=E^2-2\lambda(1-\lambda)\hbar^2v^2k_y^2. We note that λ\lambda, EE and kyk_y are varying parameters, satisfying ℏv∣ky∣<E\hbar v|k_y|<E and 0≤λ≤10\leq\lambda\leq1. When λ=0\lambda=0 for ABA stacking, the roots of Eq. (8) are uo=0u_o=0 and ue±=±2γ1Eu_e^\pm=\pm\sqrt{2}\gamma_1E, which corresponds to massless and massive evanescent states, respectively, indicating two right-movers in x<−lx<-l, ko=∣E∣/ℏvk_o=|E|/\hbar v from uo=0u_o=0 and ke≃21/4γ1∣E∣/ℏvk_e\simeq2^{1/4}\sqrt{\gamma_1|E|}/\hbar v from ue+=2γ1Eu_e^+=\sqrt{2}\gamma_1E. When λ=1\lambda=1, it has only one positive root uABC≃γ14/3E2/3u_{\mathrm{ABC}}\simeq\gamma_1^{4/3}E^{2/3} that corresponds to the right-mover kABC≃γ12/3∣E∣1/3/ℏvk_{\mathrm{ABC}}\simeq\gamma_1^{2/3}|E|^{1/3}/\hbar v in ABC stacking. For 0<λ<10<\lambda<1, it is straightforward to show that Eq. (8) has only one propagating root of ubu_b larger than c+2γ1E~c+\sqrt{2}\gamma_1\widetilde{E} and that ubu_b is separated from all the other roots by at least (2−6/3)γ1E~(\sqrt{2}-\sqrt{6}/3)\gamma_1\widetilde{E}. Furthermore, one can confirm that ubu_b adiabatically connects ue+u_e^+ to uABCu_{\mathrm{ABC}} as λ\lambda varies from 0 to 1. In sharp contrast to this, the propagating mode for odd states ceases to exist when λ>(E/γ1)(1−ky2/(4∣k∣2))\lambda>(E/\gamma_1)(1-k_y^2/(4|k|^2)). From these, we can assign a unique right moving state ue+u_e^+ and its corresponding kek_e for adiabatically continued states from the left to side whereas the odd mode does not have such states.

Along the right-moving even branch ue+u_e^+, we find that the current je=(ue+)†Bue+j_e=(u_e^+)^\dagger Bu_e^+ obeys ∂xje=2jeRe⁡Dee\partial_xj_e=2j_e\operatorname{Re}D_{ee}. Then, the unit flux condition at every xx therefore forces Re⁡Dee=0\operatorname{Re}D_{ee}=0 or DeeD_{ee} is a pure imaginary number. The resulting diagonal equation integrates to a pure phase so that ce(+∞)=exp⁡(−∫Deedx)=eiγec_e(+\infty)=\exp(-\int D_{ee}dx)=e^{i\gamma_e} from Eq. (7), directly indicating that, to the first order, the transmission probability of the even massive fermion across ABA and ABC stacking is ∣ce∣2=1|c_e|^2=1 at every kyk_y. Since there is only one mode in ABC stacking in the energy range 1 of Figure 1a, there is no chance for the odd sector to be connected adiabatically. So, the odd mode is neither transmitted nor converted so that it is reflected as itself with unit probability, at every kyk_y in the range.

The only way out for the odd mode is a non-adiabatic coupling to other modes within the junction region of ∣x∣<l|x| < l for the higher-order process, and its size can be estimated from Eq. (7). By differentiating Kun=knun\mathcal{K}u_n = k_nu_n and projecting the differentiated equation with um†B/Jmu_m^\dagger B/J_m, then one can obtain analytic expression for DmnD_{mn} in case of m≠nm \ne n using ∂xK=−dλ(x)dxBΔHℏv\partial_x\mathcal{K} = -\frac{d\lambda(x)}{dx}\frac{B\Delta H}{\hbar v}, as following,

Dmn=−λ′(x)ℏvJm⟨um∣ΔH∣un⟩kn−km.(9)D_{mn}=-\frac{\lambda'(x)}{\hbar vJ_m}\frac{\langle u_m|\Delta H|u_n\rangle}{k_n-k_m}. \tag*{(9)}

The amplitude leaked onto branch mm from nn can be approximated as ∂xcm≃−Dmnei∫x(kn−km)dx\partial_xc_m \simeq-D_{mn}e^{i\int^x(k_n-k_m)dx}. The spinors of umu_m and unu_n and the mismatch Δk=kn−km\Delta k=k_n-k_m would be quite smoothly varied with xx. So, using Eq. (9), cm∝∫dxλ′(x)eixΔkc_m\propto\int dx\lambda'(x)e^{ix\Delta k}. The leaked amplitude is then proportional to the Fourier transform of the stacking gradient λ′(x)=12lsech⁡2(x/l)\lambda'(x)=\frac{1}{2l}\operatorname{sech}^2(x/l) at the boundary region,

I(Δk)=∫λ′(x)eiΔkxdx=πlΔk/2sinh⁡(πlΔk/2).(10)I(\Delta k)=\int\lambda'(x)e^{i\Delta kx}dx=\frac{\pi l\Delta k/2}{\sinh(\pi l\Delta k/2)}. \tag*{(10)}

For two limiting cases, I=1I=1 for lΔk≪1l\Delta k\ll1 and I≃πlΔke−πlΔk/2I\simeq\pi l\Delta k e^{-\pi l\Delta k/2} for lΔk≫1l\Delta k\gg1. Then, we expect a significant leakage of massless fermion to ABC region for the abrupt boundary, while no transmission through the higher order process regardless of kyk_y for the smooth one.

Quantized energy and wave functions for ideal triangular confinement

Here, we derive the quantization rule for a massless Dirac fermion in an equilateral triangle of side length LML_M satisfying the BM boundary condition [12]. Inside the triangle the particle is described by H△=−iℏv(σx∂x+σy∂y)H_\triangle=-i\hbar v(\sigma_x\partial_x+\sigma_y\partial_y) acting on the spinor ψ=(ψ1,ψ2)⊤\psi=(\psi_1,\psi_2)^\top. We note that the valley index is suppressed, since the smooth boundary does not mix valleys. As shown in Extended Data Figure 10, we take the vertices A=(0,0)A=(0,0), B=LM(1,0)B=L_M(1,0), C=LM(12,32)C=L_M(\frac{1}{2},\frac{\sqrt{3}}{2}) and label the sides ABAB, BCBC, CACA by s=0,1,2s=0,1,2. Side ss has unit tangent t^s=(cos⁡αs,sin⁡αs)\hat{t}_s=(\cos\alpha_s,\sin\alpha_s) and outward normal n^s=(sin⁡αs,−cos⁡αs)\hat{n}_s=(\sin\alpha_s,-\cos\alpha_s) with αs=2πs/3\alpha_s=2\pi s/3.

Schematic geometries of triangular quantizations

Extended Data Figure 10. Schematic geometries of triangular quantizations. a, An equilateral triangle ABCABC of which vertices are located on A=(0,0)A = (0, 0), B=(LM,0)B = (L_M, 0) and C=(LM/2,LM3/2)C = (L_M/2, L_M\sqrt{3}/2). For each side, ∣tAB∣=∣tBC∣=∣tCA∣=LM|\mathbf{t}_{AB}| = |\mathbf{t}_{BC}| = |\mathbf{t}_{CA}| = L_M and LML_M is a moiré periodicity. t^S\hat{\mathbf{t}}_S where S=AB,BC,CAS = AB, BC, CA is a unit vector for each side and n^S\hat{\mathbf{n}}_S is its normal vector. b, A triangular lattice constructed by mirror reflections of △ABC\triangle ABC with respect to its sides, e.g., △ABC\triangle ABC (black color) transforms to △AB′C\triangle AB'C (blue) through a mirror reflection with respect to side CACA. Each color denote the same triangle. a1\mathbf{a}_1 and a2\mathbf{a}_2 are unit vectors for the lattice. c, A reciprocal lattice of the triangular tiling in c and its unit vectors, b1\mathbf{b}_1 and b2\mathbf{b}_2.

A confining boundary condition must be local, linear and carry no current through the side [12, 32, 33]. The normal current is jn=vψ†(σ⋅n^s)ψ=−2vIm⁡(e−iαsψ1∗ψ2)j_n=v\psi^\dagger(\boldsymbol{\sigma}\cdot\hat{n}_s)\psi=-2v\operatorname{Im}(e^{-i\alpha_s}\psi_1^*\psi_2), with σ=x^σx+y^σy\boldsymbol{\sigma}=\hat{x}\sigma_x+\hat{y}\sigma_y, so it vanishes when ψ2=eiαsψ1\psi_2=e^{i\alpha_s}\psi_1. On side ss the BM condition is therefore ψ=Msψ\psi=M_s\psi with Ms=σ⋅t^sM_s=\boldsymbol{\sigma}\cdot\hat{t}_s.

Since H△2ψ=−ℏ2v2∇2ψ=E2ψH_\triangle^2\psi=-\hbar^2v^2\nabla^2\psi=E^2\psi, each spinor component ψa\psi_a (a=1,2a=1,2) obeys the Helmholtz equation (∇2+k2)ψa=0(\nabla^2+k^2)\psi_a=0 with k=E/ℏvk=E/\hbar v, so an eigenstate is a superposition of plane waves of the single modulus kk. For kθ=k(cos⁡θ,sin⁡θ)\mathbf{k}_\theta=k(\cos\theta,\sin\theta) the positive-energy spinor is ψ(r;θ)=(1,eiθ)⊤eikθ⋅r\psi(\mathbf{r};\theta)=(1,e^{i\theta})^\top e^{i\mathbf{k}_\theta\cdot\mathbf{r}}, and the eigenfunction is Ψ=∑jc(θj)ψ(r;θj)\Psi=\sum_j c(\theta_j)\psi(\mathbf{r};\theta_j). An arbitrary position on side ss can be written as r=rs+τt^s\mathbf{r}=\mathbf{r}_s+\tau\hat{t}_s with r0=A\mathbf{r}_0=A, r1=B\mathbf{r}_1=B, r2=C\mathbf{r}_2=C and 0≤τ≤LM0\leq\tau\leq L_M. The BM condition on side ss then reads ∑jc(θj)(eiθj−eiαs)eikj⋅r=0\sum_j c(\theta_j)(e^{i\theta_j}-e^{i\alpha_s})e^{i\mathbf{k}_j\cdot\mathbf{r}}=0 for every τ\tau. From this, one can check that waves with different tangential momenta along the sides are linearly independent in τ\tau and cannot cancel one another. So, only a direction θ\theta and its mirror image θˉ=2αs−θ\bar{\theta}=2\alpha_s-\theta share a tangential momentum, and the BM condition can be applied to those pairs.

As shown in Extended Data Figure 10, the triangle has a mirror image with respect to each of its three sides, and the set of directions must be closed under the three mirrors θ↦2αs−θ\theta\mapsto2\alpha_s-\theta. Two successive mirrors are rotations by ±2π/3\pm2\pi/3, so for a given kθ\mathbf{k}_\theta the triangular geometry generates three equivalent momenta k+,s\mathbf{k}_{+,s} with angles θ+,s=θ+αs\theta_{+,s}=\theta+\alpha_s and their mirror partners k−,s\mathbf{k}_{-,s} with angles θ−,s=−θ+αs\theta_{-,s}=-\theta+\alpha_s, six plane waves in total (on side s′s', θ+,s\theta_{+,s} is paired with θ−,2s′−s\theta_{-,2s'-s}, indices mod 3). For each mirror pair (θ,θˉ)(\theta,\bar{\theta}) on side ss the boundary condition yields

c(θˉ)c(θ)=eirs⋅(kθ−kθˉ)ei(θ−αs).(11)\frac{c(\bar{\theta})}{c(\theta)}=e^{i\mathbf{r}_s\cdot(\mathbf{k}_\theta-\mathbf{k}_{\bar{\theta}})}e^{i(\theta-\alpha_s)}. \tag*{(11)}

From this one notes that a wave running along the side (θ=αs\theta=\alpha_s) is self-paired and unconstrained, whereas a wave running on the opposite way (θ=αs+π\theta=\alpha_s+\pi) is not allowed.

Now, let’s apply the BM condition in Eq. (11) on each side ss directly. For sides ABAB and CACA, rs\mathbf{r}_s is orthogonal to kθ−kθˉ\mathbf{k}_\theta-\mathbf{k}_{\bar{\theta}} and eirs⋅(kθ−kθˉ)=1e^{i\mathbf{r}_s\cdot(\mathbf{k}_\theta-\mathbf{k}_{\bar{\theta}})}=1. On side ABAB, the mirror image of θ+,s\theta_{+,s} is θˉ=−θ−αs=θ−,−s\bar{\theta}=-\theta-\alpha_s=\theta_{-, -s}, giving c(θ−,−s)c(θ+,s)=ei(θ+αs)\frac{c(\theta_{-,-s})}{c(\theta_{+,s})}=e^{i(\theta+\alpha_s)}. Here, “−s-s” follows mod-3 convention such that s=−1→(−2)s=-1\to(-2) converts to 2(1)2(1). On side CACA, θˉ=2π3−θ−αs=θ−,1−s\bar{\theta}=\frac{2\pi}{3}-\theta-\alpha_s=\theta_{-,1-s}, so c(θ−,1−s)c(θ+,s)=ei(θ+αs−4π/3)\frac{c(\theta_{-,1-s})}{c(\theta_{+,s})}=e^{i(\theta+\alpha_s-4\pi/3)}. Taking c(θ)=1c(\theta)=1, then, we can set coefficients, c(θ+,s)=ωsc(\theta_{+,s})=\omega^s and c(θ−,s)=eiθωsc(\theta_{-,s})=e^{i\theta}\omega^s where ω=e2πi/3\omega=e^{2\pi i/3}.

On side BCBC, θˉ=4π3−θ=θ−,−2−s\bar{\theta}=\frac{4\pi}{3}-\theta=\theta_{-,-2-s}. Then, from Eq. (11) and solutions c(θ±,s)c(\theta_{\pm,s}) obtained above, one can find eir1⋅(k+,s−k−,−2−s)=e−i(θ+αs)c(θ−,−2−s)c(θ+,s)=ω3(1−s)=1e^{i\mathbf{r}_1\cdot(\mathbf{k}_{+,s}-\mathbf{k}_{-,-2-s})}=e^{-i(\theta+\alpha_s)}\frac{c(\theta_{-,-2-s})}{c(\theta_{+,s})}=\omega^{3(1-s)}=1. Now, plugging r1=LMx^\mathbf{r}_1=L_M\hat{x}, we have a quantization condition, eiksLM=1e^{ik_sL_M}=1 where ks=x^⋅(k+,s−k−,−2−s)=3ksin⁡(θ+2π(s−1)/3)k_s=\hat{x}\cdot(\mathbf{k}_{+,s}-\mathbf{k}_{-,-2-s})=\sqrt{3}k\sin(\theta+2\pi(s-1)/3). With (kx,ky)=k(cos⁡θ,sin⁡θ)(k_x,k_y)=k(\cos\theta,\sin\theta), we first set k1LM=3kyLM=2πnk_1L_M=\sqrt{3}k_yL_M=2\pi n, and then remaining two conditions k0LM=−2πmk_0L_M=-2\pi m and k2LM=2π(m−n)k_2L_M=2\pi(m-n) with integers mm and nn, satisfying ∑sks=0\sum_s k_s=0. Hence, the allowed momentum with the BM condition on the equilateral triangle is given by

kmn=2π3LM(2m−n,3n).k_{mn}=\frac{2\pi}{3L_M}(2m-n,\sqrt{3}n).

Then, ∣kmn∣=kmn=4π3LMN|\mathbf{k}_{mn}|=k_{mn}=\frac{4\pi}{3L_M}\sqrt{N} with N=m2−mn+n2N=m^2-mn+n^2. With all these ingredients, one can obtain the quantized energy level confined in the equilateral triangle of side LML_M,

Emn=ℏvkmn=4πℏv3LMm2−mn+n2.(12)E_{mn}=\hbar vk_{mn}=\frac{4\pi\hbar v}{3L_M}\sqrt{m^2-mn+n^2}. \tag*{(12)}

that corresponds to the bare energy ϵmn\epsilon_{mn} in Eq. (2) of the main text.

As shown in Extended Data Figure 10, kmnk_{mn} runs over the triangular lattice spanned by the reciprocal lattice vectors b1\mathbf{b}_1 and b2\mathbf{b}_2 of the periodic lattice generated by the mirror operations on the triangle ABCABC. Scaled by 3LM4π\frac{3L_M}{4\pi}, the lattice points mb1+nb2m\mathbf{b}_1+n\mathbf{b}_2 coincide exactly with the Eisenstein integers zmn=m+nωz_{mn}=m+n\omega in the complex plane. The quantized energy also can be expressed as a norm of zmnz_{mn}, N[zmn]=NN[z_{mn}]=N, where N[x+iy]=x2+y2N[x+iy]=x^2+y^2 for x,y∈Rx,y\in\mathbb{R}. Among these points, distinct states correspond to one lattice vector per 60∘60^\circ wedge, m>n≥0m>n\geq0, as shown in Figure 4a, because each of six waves occupies a corresponding 1/6 of entire lattices.

We can obtain the eigenstates as a linear combination of six plane waves for given mm and nn. Using computed coefficients above and eiθ=zmn/Ne^{i\theta}=z_{mn}/\sqrt{N}, the compact expression for each spinor component can be written as

ψ1mn=Φ+(kmn,r)+zmnNΦ+(Mkmn,r),\psi_1^{mn}=\Phi_+(\mathbf{k}_{mn},\mathbf{r})+\frac{z_{mn}}{\sqrt{N}}\Phi_+(\mathcal{M}\mathbf{k}_{mn},\mathbf{r}),
ψ2mn=zmnNΦ−(kmn,r)+Φ−(Mkmn,r).\psi_2^{mn}=\frac{z_{mn}}{\sqrt{N}}\Phi_-(\mathbf{k}_{mn},\mathbf{r})+\Phi_-(\mathcal{M}\mathbf{k}_{mn},\mathbf{r}).

where Φ±(k,r)=∑s=02ω±sexp⁡[i(C3sk)⋅r]\Phi_\pm(\mathbf{k},\mathbf{r})=\sum_{s=0}^{2}\omega^{\pm s}\exp[i(\mathcal{C}_3^s\mathbf{k})\cdot\mathbf{r}]. Here, C3\mathcal{C}_3 is the rotation by 2π/32\pi/3 and M\mathcal{M} the reflection in the xx axis (side ABAB). For a given kmn\mathbf{k}_{mn}, C3kmn=k−n,m−n\mathcal{C}_3\mathbf{k}_{mn}=\mathbf{k}_{-n,m-n}, C32kmn=kn−m,−m\mathcal{C}_3^2\mathbf{k}_{mn}=\mathbf{k}_{n-m,-m} and Mkmn=km−n,−n\mathcal{M}\mathbf{k}_{mn}=\mathbf{k}_{m-n,-n}. Then, with a eigenspinor Ψmn=(ψ1mn,ψ2mn)T\Psi_{mn}=(\psi_1^{mn},\psi_2^{mn})^{\mathsf{T}}, ρmn=∣Ψmn∣2\rho_{mn}=|\Psi_{mn}|^2 and the local density for an energy EN=4πℏv3LMNE_N=\frac{4\pi\hbar v}{3L_M}\sqrt{N} is

ρN(r)=∑{m,n}ρmn=16S∑{m,n}[9−∣Smn(r)∣2].(13)\rho_N(\mathbf{r})=\sum_{\{m,n\}}\rho_{mn}=\frac{1}{6S}\sum_{\{m,n\}}\left[9-|S_{mn}(\mathbf{r})|^2\right]. \tag*{(13)}

where S=3LM2/4S=\sqrt{3}L_M^2/4, Smn(r)=∑s=02exp⁡[i(C3sk)⋅r]S_{mn}(\mathbf{r})=\sum_{s=0}^{2}\exp[i(\mathcal{C}_3^s\mathbf{k})\cdot\mathbf{r}] and {m,n}\{m,n\} indicates integers mm and nn satisfying ∣zmn∣=N|z_{mn}|=\sqrt{N}.

Density of states of ideal triangular confinement of massless Dirac fermions

We first define r(N)≡#{(m,n)∈Z2∣∣zmn∣2=N}r(N)\equiv\#\{(m,n)\in\mathbb{Z}^2\mid|z_{mn}|^2=N\}, the number of zmn=m+ne2πi/3z_{mn}=m+ne^{2\pi i/3} whose norm is equal to NN. Then, the total number N(ε)\mathcal{N}(\varepsilon) of zmnz_{mn} inside the circle of radius ε\varepsilon is given by N(ε)=∑0≤N≤ε2r(N)=∑m,nΘ(ε−∣zmn∣)\mathcal{N}(\varepsilon)=\sum_{0\leq N\leq\varepsilon^2}r(N)=\sum_{m,n}\Theta(\varepsilon-|z_{mn}|) where Θ(x)=1\Theta(x)=1 (00) for x≥0x\geq0 (x<0x<0). Then, the total number n△(ε)n_\triangle(\varepsilon) of zmnz_{mn} in the domain D\mathcal{D} shown in Figure 4a can be written as n△(ε)=(N(ε)−1)/6n_\triangle(\varepsilon)=(\mathcal{N}(\varepsilon)-1)/6.

Using Proposition 4.7 in [24] for the prime factorization of zmnz_{mn}, the norm can be expressed as a product of primes, N=3α0∏p≡1(3)pαp∏q≡2(3)q2βqN=3^{\alpha_0}\prod_{p\equiv1(3)}p^{\alpha_p}\prod_{q\equiv2(3)}q^{2\beta_q}, where α0\alpha_0, αp\alpha_p and βq\beta_q are non-negative integers and p≡1(3)p\equiv1(3) and q≡2(3)q\equiv2(3) denotes primes with residues 1 and 2 modulo 3, respectively [22]. Applying Proposition 4 in [22] then gives r(N)=6∏p≡1(3)(αp+1)r(N)=6\prod_{p\equiv1(3)}(\alpha_p+1). As an illustration, for N=1,3,4,7N=1,3,4,7 and 99, the exponents (α0,αp,βq)(\alpha_0,\alpha_p,\beta_q) take values (0,0,0)(0,0,0), (1,0,0)(1,0,0), (0,0,1)(0,0,1), (0,1,0)(0,1,0) and (2,0,0)(2,0,0) yielding r(1)=r(3)=r(4)=6r(1)=r(3)=r(4)=6, r(7)=12r(7)=12 and r(9)=6r(9)=6.

To compute N(ε)\mathcal{N}(\varepsilon), we apply the Poisson summation formula to the dual space Λ∗\Lambda^* spanned by ξ=mp1∗+np2∗\boldsymbol{\xi}=m\mathbf{p}_1^*+n\mathbf{p}_2^*, where p1∗=(1,−1/3)\mathbf{p}_1^*=(1,-1/\sqrt{3}) and p2∗=(0,2/3)\mathbf{p}_2^*=(0,2/\sqrt{3}) with ∣ξ∣2=43N[zmn]=43N|\boldsymbol{\xi}|^2=\frac{4}{3}N[z_{mn}]=\frac{4}{3}N. Here Λ∗\Lambda^* is the dual lattice of the triangular lattice Λ\Lambda span by p1=(1,0)\mathbf{p}_1=(1,0) and p2=(1/2,3/2)\mathbf{p}_2=(1/2,\sqrt{3}/2) satisfying pi⋅pj∗=δij\mathbf{p}_i\cdot\mathbf{p}_j^*=\delta_{ij} (i,j=1,2)(i,j=1,2), which is equivalent to that generated by zmnz_{mn}. Using N(ε)=13/2∑ξ∈Λ∗∫0ερ dρ∫02πdθ e−2πiρ∣ξ∣cos⁡θ\mathcal{N}(\varepsilon)=\frac{1}{\sqrt{3}/2}\sum_{\boldsymbol{\xi}\in\Lambda^*}\int_0^\varepsilon\rho\,d\rho\int_0^{2\pi}d\theta\,e^{-2\pi i\rho|\boldsymbol{\xi}|\cos\theta}, one obtain

N(ε)=23πε2+ε∑N≥1r(N)NJ1(2πεξN),(14)\mathcal{N}(\varepsilon)=\frac{2}{\sqrt{3}}\pi\varepsilon^2+\varepsilon\sum_{N\geq1}\frac{r(N)}{\sqrt{N}}J_1(2\pi\varepsilon\xi_N), \tag*{(14)}

where ξN=∣ξ∣=23N\xi_N=|\boldsymbol{\xi}|=\frac{2}{\sqrt{3}}\sqrt{N} and J1(x)J_1(x) is the first-order Bessel function of the first kind. Here we note that a convolution with the Gaussian broadening kernel 12πσe−ε2/2σ2\frac{1}{\sqrt{2\pi}\sigma}e^{-\varepsilon^2/2\sigma^2} introduces a damping factor e−2π2σ2ξN2e^{-2\pi^2\sigma^2\xi_N^2} for each Bessel term. Owing to its rapid decay with increasing NN, the contribution from N=1N=1 is enough. Then, with d△(ε)=∂εn△(ε)=∂εN(ε)/6d_\triangle(\varepsilon)=\partial_\varepsilon n_\triangle(\varepsilon)=\partial_\varepsilon\mathcal{N}(\varepsilon)/6, one obtains Eq. (1).

Leakage of lower energy confined states near EFE_F

As discussed in the main text, the ABB stacking on the corner has different interlayer interaction compared with BTG and RTG. From Slater-Koster parameters adopted in our work (see Methods), the vertical interlayer hopping for direct overlap between atoms γBB\gamma_{BB} is given as γBB/γ1=exp⁡[−(dσ/aσ)Δd]=e−2.218Δd\gamma_{BB}/\gamma_1=\exp[-(d_\sigma/a_\sigma)\Delta d]=e^{-2.218\Delta d}, Δd\Delta d in Å. From the relaxed height corrugation across the corner shown in Extended Data Figure 9 and Extended Data Figure 9, one can estimate γBB≃244\gamma_{BB}\simeq244 meV. With this, we compute the energy bands of pristine ABB stacked trilayer and associated weight expectation ⟨Po⟩\langle P_o\rangle of odd state projector PoP_o where ⟨Po⟩=⟨ψnkABB∣Po∣ψnkABB⟩\langle P_o\rangle=\langle\psi_{nk}^{ABB}|P_o|\psi_{nk}^{ABB}\rangle, Po=∣A−⟩⟨A−∣+∣B−⟩⟨B−∣P_o=|A_-\rangle\langle A_-|+|B_-\rangle\langle B_-| and ψnkABB\psi_{nk}^{ABB} is a Bloch state of pristine ABB stacked trilayer at a given band nn at momentum k\mathbf{k}. From Extended Data Figure 9, one can estimate energy-dependent leakage from the triangular domain to the other as discussed in the main text.

Energy bands of ABB stacking and their overlap with massless states in ABA stacking

Extended Data Figure 9. Energy bands of ABB stacking and their overlap with massless states in ABA stacking. a, Energy bands along kxk_x direction for odd sector (massless bands) in ABA stacking (dash-dot line) and ABB stacking (solid lines), respectively. For ABB bands, the grey scale denotes the weight of projection operator ⟨Po⟩≡⟨ψnkABB∣Po∣ψnkABB⟩\langle P_o\rangle\equiv\langle\psi_{nk}^{ABB}|P_o|\psi_{nk}^{ABB}\rangle of ABA stacking, Po=∣A−⟩⟨A−∣+∣B−⟩⟨B−∣P_o = |A_-\rangle\langle A_-| + |B_-\rangle\langle B_-| where ∣A−⟩=∣A1⟩−∣A3⟩2|A_-\rangle= \frac{|A_1\rangle-|A_3\rangle}{\sqrt{2}} and ∣B−⟩=∣B1⟩−∣B3⟩2|B_-\rangle= \frac{|B_1\rangle-|B_3\rangle}{\sqrt{2}}. Here ∣A1(3)⟩|A_{1(3)}\rangle and ∣B1(3)⟩|B_{1(3)}\rangle are eigenstates of A and B sublattices of first (third) layer. The horizontal dotted line denotes the local maximum energy of Ec=44.8E_c = 44.8 meV at k=0.022k = 0.022 Å−1^{-1}. b, For a given energy of EE, the maximum value of projection among ⟨Po⟩\langle P_o\rangle is plotted. The vertical dotted line is EcE_c in a.

For a leakage from n=0n=0 modes in Figure 4a, we consider the mirror-odd propagating state in x<−lx<-l, that is tangent to the wall in ∣x∣<l|x|<l. The boundary geometry is the same as discussed above. Then, the state has kx=0k_x=0 and E=ℏvkyE=\hbar v k_y, and its spinor is given by ∣o⟩=12(∣A−⟩+i∣B−⟩)|o\rangle=\frac{1}{\sqrt{2}}(|A_-\rangle+i|B_-\rangle). The coupling λΔH\lambda\Delta H in Eq. (5) maps it onto a single normalized, mirror-even state on the layer 2, ∣p⟩=12(−∣B2⟩+i∣A2⟩)|p\rangle=\frac{1}{\sqrt{2}}(-|B_2\rangle+i|A_2\rangle), with element ⟨p∣λΔH∣o⟩=gλ\langle p|\lambda\Delta H|o\rangle=g\lambda. So, they can be regarded as two states with an effective Hamiltonian, Heff=E+gλσxH_{\mathrm{eff}}=E+g\lambda\sigma_x for (∣o⟩,∣p⟩)T(|o\rangle,|p\rangle)^{\mathsf{T}}. Under HeffH_{\mathrm{eff}}, the odd state will evolves as (cos⁡ϑ,−isin⁡ϑ)T(\cos\vartheta,-i\sin\vartheta)^{\mathsf{T}}. So, assuming that massless particle moves at speed vv with arbitrary momentum along a path ss, then ϑ=gλ(x)s/ℏv\vartheta=g\lambda(x)s/\hbar v. Near a side the n=0n=0 odd state is spread over a transverse width Lρ∼LML_\rho\sim L_M, each slice dx/Lρdx/L_\rho runs a path LρL_\rho. The rotation of the state can be the density-weighted average, δ=∫−∞0dxLρgλ(x)ℏvLρ=γ1lln⁡222ℏv≃0.330π\delta=\int_{-\infty}^{0}\frac{dx}{L_\rho}\frac{g\lambda(x)}{\hbar v}L_\rho=\frac{\gamma_1 l\ln2}{2\sqrt{2}\hbar v}\simeq0.330\pi. Then, this evolution estimation gives Aw=cos⁡2δ=0.259A_w=\cos^2\delta=0.259.

Self-energy corrections to confined massless states

Let’s split the Hilbert space into PP for the odd modes ψP=Ψmn\psi_P = \Psi_{mn} for the closed triangle with H△Ψmn=εnmΨmnH_{\triangle}\Psi_{mn} = \varepsilon_{nm}\Psi_{mn} and QQ for leaked states (HQψQ=εQψQ)(H_Q\psi_Q = \varepsilon_Q\psi_Q) that Ψmn\Psi_{mn} couples to with a coupling VV. Here, εnm=ε0∣zmn∣\varepsilon_{nm} = \varepsilon_0|z_{mn}| and ε0=4πℏv3LM\varepsilon_0 = \frac{4\pi\hbar v}{3L_M}. Writing H=(H△VV†HQ)H = \begin{pmatrix} H_{\triangle} & V \\ V^\dagger& H_Q \end{pmatrix} for (ψP,ψQ)T(\psi_P,\psi_Q)^T, and eliminating QQ based on a standard procedure [19],

[H△+Σ(E)]ψP=EψP,[H_{\triangle}+\Sigma(E)]\psi_P=E\psi_P,

where a self energy is given by Σ(E)=V(E−HQ+i0)−1V†\Sigma(E) = V(E-H_Q+i0)^{-1}V^\dagger. Here for Σ(E)≡Δ(E)+iΓ(E)\Sigma(E)\equiv\Delta(E)+i\Gamma(E), we assume the affine form without mode-coupling such that Δ(E)≃Δ(E0)+(E−E0)∂EΔ(E)∣E=E0\Delta(E)\simeq\Delta(E_0)+(E-E_0)\left.\partial_E\Delta(E)\right|_{E=E_0} for an energy E0>EcE_0>E_c. With α=−∂EΔ(E)∣E=E0\alpha=-\left.\partial_E\Delta(E)\right|_{E=E_0} and εc=Δ(E0)−E0∂EΔ(E)∣E=E0\varepsilon_c=\Delta(E_0)-E_0\left.\partial_E\Delta(E)\right|_{E=E_0}, we can obtain Emn≃Zε0∣zmn∣+EcE_{mn}\simeq Z\varepsilon_0|z_{mn}|+E_c where Z=11+αZ=\frac{1}{1+\alpha} and Ec=ZεcE_c=Z\varepsilon_c as shown in Eq. (2) and to match the simulated data, we set Z≃0.84Z\simeq0.84.

To account the boundary leakage for n=0n=0 modes in DOS, we define B(ε)=#{(m,0)∣0<m≤ε}=∑m≥1Θ(ε−m)\mathcal{B}(\varepsilon)=\#\{(m,0)\mid0<m\le\varepsilon\}=\sum_{m\ge1}\Theta(\varepsilon-m). For B(ε)\mathcal{B}(\varepsilon), an analogous calculation to N(ε)\mathcal{N}(\varepsilon) in Eq. (14) gives B(ε)=ε+1π∑n≥1sin⁡(2πnε)n\mathcal{B}(\varepsilon)=\varepsilon+\frac{1}{\pi}\sum_{n\ge1}\frac{\sin(2\pi n\varepsilon)}{n}. We note that a Gaussian convolution 12πσe−ε2/2σ2\frac{1}{\sqrt{2\pi}\sigma}e^{-\varepsilon^2/2\sigma^2} introduces a damping e−2π2σ2n2e^{-2\pi^2\sigma^2n^2} for sin⁡(2πnε)\sin(2\pi n\varepsilon) when ε≫σ\varepsilon\gg\sigma. Owing to their rapid decays with increasing nn, the contributions from n=1n=1 are enough. So, the DOS contribution from the modes along the walls is bwall(ε)=∂εB(ε)≃1+2e−2π2σ2cos⁡(2πε)b_{\mathrm{wall}}(\varepsilon)=\partial_\varepsilon\mathcal{B}(\varepsilon)\simeq1+2e^{-2\pi^2\sigma^2}\cos(2\pi\varepsilon). Then, considering a renormalization factor ZZ, the total DOS D△D_{\triangle} can be written as

D△=Z(d△−dwall)+ZAwdwallD_{\triangle}=Z(d_{\triangle}-d_{\mathrm{wall}})+ZA_{\mathrm{w}}d_{\mathrm{wall}}

as can be seen in Eq. (3).

By extracting the oscillating parts in D△D_{\triangle} and compare them with pristine DOS of ABA stacking, one can obtain the normalized deviation shown in Figure 4c as follows,

Δρ(E)=100×Dosc(E)De(E)+Do(E)\Delta\rho(E)=100\times\frac{D_{\mathrm{osc}}(E)}{D_{\mathrm{e}}(E)+D_{\mathrm{o}}(E)}

in %, (15)

where Dosc(E)=1Sε0[dbulk(ε)+(Aw−1)(bwall(ε)−1)]D_{\mathrm{osc}}(E)=\frac{1}{S\varepsilon_0}[d_{\mathrm{bulk}}(\varepsilon)+(A_{\mathrm{w}}-1)(b_{\mathrm{wall}}(\varepsilon)-1)] is the oscillatory part in D△D_{\triangle}, ε=E−EcZε0\varepsilon=\frac{E-E_c}{Z\varepsilon_0}, S=3LM24S=\frac{\sqrt{3}L_M^2}{4}, and dbulk(ε)=2πε33e−2π2σ2/3J0(4πε/3)d_{\mathrm{bulk}}(\varepsilon)=\frac{2\pi\varepsilon}{3\sqrt{3}}e^{-2\pi^2\sigma^2/3}J_0(4\pi\varepsilon/\sqrt{3}). Here, De(E)=2γ14πℏ2v2D_{\mathrm{e}}(E)=\frac{\sqrt{2}\gamma_1}{4\pi\hbar^2v^2} and Do(E)=E2πℏ2v2D_{\mathrm{o}}(E)=\frac{E}{2\pi\hbar^2v^2} are the pristine DOSs of even and odd modes in BTG, respectively.

Acknowledgments

We thank fruitful discussions with Mikito Koshino. K.Y. was supported by Korea Institute for Advanced Study (KIAS) individual grant (CG092502). C.P. was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT and MOE) (No. RS-2025-16063688), and the Basic Science Research Program (No. RS-2024-00349573) through NRF, funded by the MSIT of the Republic of Korea. Y-W.S. was supported by the KIAS individual Grant No. (CG031509) and the NRF grant funded by the Korea government (MSIT) (No. RS-2026-25508298). Computations were supported by CAC of KIAS.

References

Layer-resolved LDOS of the TMB system projected on each domain

Figure Extended Data Fig 1. Layer-resolved LDOS of the TMB system projected on each domain. a The layer-resolved LDOS projected on the rhombohedral domain of the TMB system (black triangle in Figure 1c) are shown, where ‘R1’, ‘R2’, and ‘R3’ refer to the LDOS projected on the layer 1, 2, and 3 of the rhombohedral domain, respectively. Here, layer 1 is the twisted monolayer section of the TMB. Reference layer-resolved DOS of the pristine RTG are shown together. Note that in the pristine system, the layer-resolved DOSs of the layer 1 and layer 3 are the same due to a symmetry (labeled as ‘1 & 3 ref.’), whereas that of the layer 2 is distinct (‘2 ref.’) b The layer-resolved LDOS projected on the Bernal domain of the TMB system (white triangle in Figure 1c) are shown, where ‘B1’, ‘B2’, and ‘B3’ refer to the LDOS projected on the layer 1, 2, and 3 of the Bernal domain, respectively.

LDOS of the ATT system projected on each regions

Figure Extended Data Fig 2. LDOS of the ATT system projected on each regions. a The LDOS (ρA(E)\rho_A(E)) projected on the Bernal (B) domains of the ATT system (white triangles in Figure 1d) is shown by a solid line, and the B reference DOS is shown by a gray dotted line. Similarly to the Figure 2, the corresponding normalized deviation is shown in b. c and d The LDOS of the boundary regions of the ATT system (rectangles in Figure 1d) and its normalized deviation ΔρA\Delta\rho_A. The gray shades are ±σ\pm\sigma range around the energy levels corresponding to Extended Data Figure 7-Extended Data Figure 7. e The LDOS of the vertex and the B reference DOS. The gray shades are ±σ\pm\sigma range for Extended Data Figure 7,Extended Data Figure 7.

LDOS of the TMB system projected on boundaries and vertex

Figure Extended Data Fig 3. LDOS of the TMB system projected on boundaries and vertex. a The LDOS (ρA(E)\rho_A(E)) projected on the boundary regions of the TMB system (rectangles in Figure 1c) is shown by a solid line, and the R+B reference DOS is shown by a gray dashed line. Similarly to the Figure 2, the corresponding normalized deviation ΔρA\Delta\rho_A is shown in b. c The LDOS of the vertex shown with the R+B reference DOS. d and e The same data as a and b, but for different energy range. The gray shades are ±σ\pm\sigma range around the energy levels corresponding to Extended Data Figure 6.

References

  1. [1]E. Y. Andrei, D. K. Efetov, P. Jarillo-Herrero, A. H. MacDonald, K. F. Mak, T. Senthil, E. Tutuc, A. Yazdani, and A. F. Young, The marvels of moiré materials, Nat. Rev. Mater. 6, 201 (2021).DOI
  2. [2]W. Ren, Z. Fu, Z. Xu, Z. Zhang, M. Luskin, and K. Wang, Review: moiré-of-moiré superlattice in twisted trilayer graphene, J. Phys.: Condensed Matter 37, 353001 (2025).DOI
  3. [3]D. M. Kennes, M. Claassen, L. Xian, A. Georges, A. J. Millis, J. Hone, C. R. Dean, D. N. Basov, A. Pasupathy, and A. Rubio, Moiré heterostructures as a condensed-matter quantum simulator, Nat. Phys. 17, 155 (2021).
  4. [4]M. P. Levenford, C.-J. Kim, L. Brown, P. Y. Huang, R. W. Havener, D. A. Muller, and J. Park, Graphene and boron nitride lateral heterostructures for atomically thin circuitry, Nature 488, 627 (2012).
  5. [5]C. Huang, S. Wu, A. M. Sanchez, J. J. P. Peters, R. Beanland, J. S. Ross, P. Rivera, W. Yao, D. H. Cobden, and X. Xu, Lateral heterojunctions within monolayer MoSe2–WSe2 semiconductors, Nature Materials 13, 1096 (2014).DOI
  6. [6]M.-Y. Li, Y. Shi, C.-C. Cheng, L.-S. Lu, Y.-C. Lin, H.-L. Tang, M.-L. Tsai, C.-W. Chu, K.-H. Wei, J.-H. He, W.-H. Chang, K. Suenaga, and L.-J. Li, Epitaxial growth of a monolayer WSe2-MoS2 lateral p-n junction with an atomically sharp interface, Science 349, 524 (2015).DOI
  7. [7]W. S. Leong, Q. Ji, N. Mao, Y. Han, H. Wang, A. J. Goodman, A. Vignon, C. Su, Y. Guo, P.-C. Shen, Z. Gao, D. A. Muller, W. A. Tisdale, and J. Kong, Synthetic Lateral Metal-Metal-Semiconductor Heterostructures of Transition Metal Disulfides, J. Am. Chem. Soc. 140, 12354 (2018).
  8. [8]T. Nakanishi, M. Koshino, and T. Ando, Transmission through a boundary between monolayer and bilayer graphene, Physical Review B 82, 125428 (2010).DOI
  9. [9]T. Ando, T. Nakanishi, and R. Saito, Berry’s phase and absence of backscattering in carbon nanotubes, J. Phys. Soc. Jpn. 67, 2857 (1998).DOI
  10. [10]M. I. Katsnelson, K. S. Novoselov, and A. K. Geim, Chiral tunnelling and the Klein paradox in graphene, Nat. Phys. 2, 620 (2006).DOI
  11. [11]L. A. Ponomarenko, F. Schedin, M. I. Katsnelson, R. Yang, E. W. Hill, K. S. Novoselov, and A. K. Geim, Chaotic Dirac billiard in graphene quantum dots, Science 320, 356 (2008).DOI
  12. [12]M. V. Berry and R. J. Mondragon, Neutrino billiards: time-reversal symmetry-breaking without magnetic fields, Proc. R. Soc. Lond. A 412, 53 (1987).DOI
  13. [13]A. Chodos, R. L. Jaffe, K. Johnson, C. B. Thorn, and V. F. Weisskopf, New extended model of hadrons, Phys. Rev. D 9, 3471 (1974).
  14. [14]A. De Martino, L. Dell’Anna, and R. Egger, Magnetic confinement of massless Dirac fermions in graphene, Phys. Rev. Lett. 98, 066802 (2007).
  15. [15]Z. Zhu, S. Carr, D. Massatt, M. Luskin, and E. Kaxiras, Twisted trilayer graphene: A precisely tunable platform for correlated electrons, Phys. Rev. Lett. 125, 116404 (2020).DOI
  16. [16]N. Nakatsuji, T. Kawakami, and M. Koshino, Multiscale lattice relaxation in general twisted trilayer graphenes, Phys. Rev. X 13, 041007 (2023).DOI
  17. [17]D. Park, C. Park, Y. Yananose, E. Ko, B. Kim, R. Engelke, X. Zhang, K. Davydov, M. Green, H.-M. Kim, S. H. Park, J. H. Lee, S.-G. Kim, H. Kim, K. Watanabe, T. Taniguchi, S. M. Yang, K. Wang, P. Kim, Y.-W. Son, and H. Yoo, Unconventional domain tessellations in moiré-of-moiré lattices, Nature 641, 896 (2025).DOI
  18. [18]S. Latil and L. Henrard, Charge carriers in few-layer graphene films, Phys. Rev. Lett. 97, 036803 (2006).pubmed.ncbi.nlm.nih.gov/16907528
  19. [19]H. Min and A. H. MacDonald, Electronic structure of multilayer graphene, Prog. Theor. Phys. Supple. 176, 227 (2008).DOI
  20. [20]C. Park and Y.-W. Son, Condensation of preformed charge density waves in kagome metals, Nat. Commun. 14, 7309 (2023).DOI
  21. [21]K. Yananose, C. Park, and Y.-W. Son, Super moiré domain tessellations, sliding ferroelectricity, and reconfigurable quantum dot arrays in twisted trilayer hexagonal boron nitride, Phys. Rev. B 113, 205407 (2026).DOI
  22. [22]M. A. A. Pinsky, The eigenvalues of an equilateral triangle, SIAM J. Math. Anal. 11, 819 (1980).DOI
  23. [23]C. Itzykson and J.-M. Luck, Arithmetical degeneracies in simple quantum systems, J. Phys. A: Math. Gen. 19, 211 (1986).
  24. [24]D. A. Cox, Primes of the Form x² + ny²: Fermat, Class Field Theory, and Complex Multiplication, Pure and Applied Mathematics: A Wiley Series of Texts, Monographs and Tracts (John Wiley & Sons, New York, 1989) p. 77.
  25. [25]L. I. Glazman, G. B. Lesovik, D. E. Khmel’nitskii, and R. I. Shekhter, Reflectionless quantum transport and fundamental ballistic-resistance steps in microscopic constrictions, JETP Lett. 48, 238 (1988).
  26. [26]U. Banin, Y. Cao, D. Katz, and O. Millo, Identification of atomic-like electronic states in indium arsenide nanocrystal quantum dots, Nature 400, 542 (1999).
  27. [27]G. Lamé, Leçons sur la théorie mathématique de l’élasticité des corps solides (Bachelier, Paris, 1852) pp. 131–136.
  28. [28]J. Mathews and R. L. Walker, Mathematical Methods of Physics, 2nd ed. (W. A. Benjamin, New York, 1970) pp. 237–239.
  29. [29]H. R. Krishnamurthy, H. S. Mani, and H. C. Verma, Exact solution of the Schrödinger equation for a particle in a tetrahedral box, J. Phys. A 15, 2131 (1982).DOI
  30. [30]E. McCann and V. I. Fal’ko, Symmetry of boundary conditions of the dirac equation for electrons in carbon nanotubes, J. Phys.: Condens. Matter 16, 2371 (2004).DOI
  31. [31]L. Brey and H. A. Fertig, Electronic states of graphene nanoribbons studied with the Dirac equation, Phys. Rev. B 73, 235411 (2006).DOI
  32. [32]B. Dietz, Relativistic quantum billiards with threefold rotational symmetry: Exact, symmetry-projected solutions for the equilateral neutrino billiard, Acta Phys. Pol. A 140, 473 (2021).DOI
  33. [33]W. A. Gaddah, Exact solutions to the dirac equation for equilateral triangular billiard systems, J. Phys. A: Math. Theor. 51, 385304 (2018).DOI
  34. [34]M. Koshino and E. McCann, Multilayer graphenes with mixed stacking structure: Interplay of bernal and rhombohedral stacking, Phys. Rev. B 87, 045420 (2013).DOI
  35. [35]M. Yankowitz, J. I.-J. Wang, A. G. Birdwell, Y.-A. Chen, K. Watanabe, T. Taniguchi, P. Jacquot, P. San-Jose, P. Jarillo-Herrero, and B. J. LeRoy, Electric field control of soliton motion and stacking in trilayer graphene, Nat. Mater. 13, 786 (2014).
  36. [36]L. Jiang, S. Wang, Z. Shi, C. Jin, M. I. B. Utama, S. Zhao, Y.-R. Shen, H.-J. Gao, G. Zhang, and F. Wang, Manipulation of domain-wall solitons in bi- and trilayer graphene, Nat. Nanotechnol. 13, 204 (2018).DOI
  37. [37]L. Huang, H.-Y. Xu, C. Grebogi, and Y.-C. Lai, Relativistic quantum chaos, Phys. Rep. 753, 1 (2018).DOI
  38. [38]A. N. Kolmogorov and V. H. Crespi, Registry-dependent interlayer potential for graphitic systems, Phys. Rev. B 71, 235411 (2005).DOI
  39. [39]R. B. Lehoucq, D. C. Sorensen, and C. Yang, ARPACK users’ guide: solution of large-scale eigenvalue problems with implicitly restarted Arnoldi methods (Society for Industrial and Applied Mathematics, Philadelphia, 1998).
  40. [40]Opencollab, Arpack-ng, https://github.com/opencollab/arpack-ng, [Accessed: 24-July-2025].
  41. [41]J. S. Alden, A. W. Tsen, P. Y. Huang, R. Hovden, L. Brown, J. Park, D. A. Muller, and P. L. McEuen, Strain solitons and topological defects in bilayer graphene, Proc. Natl. Acad. Sci. U.S.A. 110, 11256 (2013).
  42. [42]D. S. Fisher and P. A. Lee, Relation between conductivity and transmission matrix, Phys. Rev. B 23, 6851(R) (1981).DOI
  43. [45]S. G. Johnson, P. Bienstman, M. A. Skorobogatiy, M. Ibanescu, E. Lidorikis, and J. D. Joannopoulos, Adiabatic theorem and continuous coupled-mode theory for efficient taper transitions in photonic crystals, Phys. Rev. E 66, 066608 (2002).pubmed.ncbi.nlm.nih.gov/12513430
  44. [46]S. J. Jang, B. K. Min, and Y. M. Rhee, Fermi’s golden rule rate expression for transitions due to nonadiabatic derivative couplings in the adiabatic basis, J. Chem. Theory Comput. 21, 1850 (2025).DOI
  45. [47]J. C. Tully, Molecular dynamics with electronic transitions, J. Chem. Phys. 93, 1061 (1990).DOI
  46. [48]F. Joucken, F. R. Issing, and R. Sporken, Fourier transform analysis of STM images of multilayer graphene moiré patterns, Carbon 83, 48 (2015).arxiv.org/abs/1409.3105
  47. [49]L.-J. Yin, H. Jiang, J.-B. Qiao, and L. He, Direct imaging of topological edge states at a bilayer graphene domain wall, Nat Commun. 7, 11760 (2016).DOI

Paper details

vlnc:W_w3q3nkbz
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 1

downloads · 4

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents