Randomly twisted bilayer graphene -- the cascade transitions

Baruch Horovitz; Pierre Le Doussal

REVIEW 4 major objections 4 minor 42 references

Randomly twisted bilayer graphene -- the cascade transitions

T0 review · 4 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that twist-angle disorder, amplified into a huge random gauge field near the magic angle, forces every electron to zero energy with equal Coulomb energy per electron, and Hartree-Fock then jumps the chemical potential at ea

desk verdict Twist-disorder-as-random-gauge mechanism for TBG cascades is genuinely new and worth refereeing, but the step from a total sum rule to equal pair overlaps is asserted, not proven, and the HF jumps rest on it. read the letter →

arxiv 2508.07024 v2 pith:YY3YEAEP submitted 2025-08-09 cond-mat.mes-hall cond-mat.dis-nn

Baruch Horovitz , Pierre Le Doussal This is my paper

classification cond-mat.mes-hallcond-mat.dis-nn PACS 73.22.Pr

keywords twistedbilayergraphenetwist-angledisorderrandomgaugepotentialmagicangledensityofstatescascadetransitionsHartree-FockDiracfermions

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Twisted bilayer graphene shows "cascade" transitions: sharp jumps in the inverse compressibility when the Moiré-band filling passes an integer. The paper claims these are a disorder effect. Local fluctuations of the twist angle act as a random gauge potential in the Dirac description of the bands, and near the magic angle, where the Dirac velocity vanishes, that potential is amplified by roughly a factor of 40, making the disorder enormous ($\sigma \sim 10^3$ to $10^4$). In this regime the disorder-averaged density of states diverges at zero energy as $\rho(E)\sim E^{(2/z)-1}$ with $z\gg1$, so every added electron sits at essentially the same energy. A sum rule on the averaged wavefunction overlaps then forces each electron to carry the same intraband Coulomb energy, and with the flavor symmetries of the eight bands, Hartree-Fock produces chemical-potential jumps at every integer filling. If the paper is right, the cascade phenomenon is a large-area averaged effect: probes averaging over regions much larger than the disorder correlation length should see sharp, integer-locked peaks, matching the cleaner experimental data.

What carries the argument

The central object is the random transverse gauge potential $A(\mathbf{r})$ of the effective 2D Dirac equation for each Moiré-band flavor, with $A_x=-\partial_y V$, $A_y=\partial_x V$, and fictitious field $B=\nabla^2 V$. It does two jobs: its disorder variance is multiplied by the inverse-velocity factor $(1+3\bar u_1^2)/(1-3\bar u_1^2)$, making the disorder enormous ($z\sim\sqrt{\sigma}$) near the magic angle and producing the DOS divergence; and its eigenfunction overlaps obey the sum rule $\sum_{|E_i|,|E_j|<\mu}\int d^2r\,|\varphi_i|^2|\varphi_j|^2\approx N/(\pi\ell_0^2)$, which, combined with the symmetry relations $A(\mathbf{r})\to -A(\mathbf{r})$ (layer exchange / Moiré valley) and $A

What would settle it

Numerically average $|\varphi_i(\mathbf{r})|^2|\varphi_j(\mathbf{r})|^2$ over disorder for $E\neq0$ eigenstates of the Dirac equation with a random gauge potential at $\sigma \approx 10^3$: if the per-pair overlaps spread substantially beyond $1/(\pi\ell_0^2)$, the equal-$U$ step fails. Experimentally, a planar-tunneling or large-area compressibility measurement with averaging over $L \gg \ell_0$ should show the zero-bias DOS divergence $\rho(E) \sim E^{(2/z)-1}$, and the cascade peaks should sharpen and lock onto integer $\nu$ as the probed area grows.

Watch

Extended reading notes

Core claim

Near the magic angle the renormalized Dirac velocity nearly vanishes, so the random gauge potential $A(\mathbf{r})$ from twist-angle fluctuations is amplified to disorder strength $\sigma\approx 10^3$–$10^4$. At such strong gauge disorder the averaged density of states diverges as $\rho(E)\sim E^{(2/z)-1}$ with $z\gg1$, so all electrons sit at $E\approx0$. A sum rule on averaged eigenfunction overlaps (Eq. 14) forces each added electron to carry the same intraband Coulomb energy $U$, and the flavor symmetries $A(\mathbf{r})\to -A(\mathbf{r})$, $A(\mathbf{r})\to A(-\mathbf{r})$ fix the interband overlaps: vanishing between opposite-Moiré-valley states, suppressed by $\ell_0^2/L^2$ between dis

Load-bearing premise

The step from the sum rule (which fixes only the total eigenfunction-overlap sum over all occupied pairs) to the conclusion that every individual pair overlap equals $\delta_{ij}/(\pi\ell_0^2)$, so that each electron carries the same Coulomb energy $U$; if overlaps vary from pair to pair, the Hartree-Fock ladder in Table II and the predicted cascade jumps lose their footing.

Editorial extensions

If this is right

Near the magic angle kinetic energy is negligible: every electron in a partially filled band sits at $E \approx 0$, so the filling sequence is set by interactions, not dispersion.
Hartree-Fock gives chemical-potential jumps at every integer filling $\nu = -3, \dots, 3$, with larger jumps at even fillings (order $U$) and weaker, size-dependent jumps at odd fillings (order $U\ell_0^2/L^2$).
Cascade peaks sharpen and lock to integer $\nu$ when the measurement averages over areas $L \gg \ell_0$, which distinguishes the broad small-area data from the sharp large-area data.
Equal-energy states of opposite Moiré valley have vanishing interaction, so the eight non-interacting flavors effectively form four bands with fillability $-1 < \nu < 1$, reconciling an apparent band-count contradiction.
Averaged planar tunneling spectroscopy over $L \gg \ell_0$ should reveal the DOS divergence $\rho(E) \sim E^{(2/z)-1}$ at the Dirac point.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Testable without any experiment: diagonalize the single-particle Dirac equation with a random transverse gauge potential at $\sigma \approx 10^3$ and average $|\varphi_i(\mathbf{r})|^2|\varphi_j(\mathbf{r})|^2$ over disorder; if the $E\neq0$ overlaps are not flat across pairs, the equal-$U$ ladder loses its basis independently of all physics beyond the model.
The disorder mechanism predicts a monotonic sharpening of cascade peaks with probed area on a single device, which imaging compressibility at several patch sizes could verify.
The same inverse-velocity amplification should apply to any near-magic-angle Moiré platform whose Dirac velocity passes through zero, so a similar DOS divergence would be expected in twisted double-bilayer or trilayer graphene.
The argument relies on particle-hole symmetry surviving disorder averaging; any measurable particle-hole asymmetry between positive and negative fillings would demand a revisit of the equal-overlap step for $E>0$ versus $E<0$ states.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know. First, the core idea is new and interesting: near the magic angle the effective Dirac velocity vanishes but the random gauge field from twist disorder does not, so the disorder parameter is boosted by roughly v_F/v to 10^3–10^4, the DOS diverges at zero energy, and a Hartree-Fock treatment with equal intraband interactions produces chemical potential jumps at integer fillings. That is a real alternative to purely interaction-driven cascade explanations, and the probe-size contrast between Refs. [3] and [4,6] is a clever observation. Second, the central derivation has a load-bearing gap that the paper does not close, and the authors half-admit it.

What is done well: the velocity renormalization argument is clean and the mapping from twist-angle fluctuations to a transverse gauge potential is worked through carefully. The sum rule in Eq. (14) is a nontrivial result, derived with the two-layer bosonization. The symmetry table for the eight flavors and the exact E=0 overlaps is useful and compact. The paper is also honest about the variational method giving exponents only and about the missing sum rule for U2 overlaps.

Where it is soft: Eq. (14) fixes only the disorder-averaged total of all occupied pair overlaps, N/(πℓ0²). The deduction that every individual pair equals δ_ij/(πℓ0²) does not follow from that total, the N→N+1 increment, and the known E=0 behavior. Those pin down the total and a measure-zero set, not the finite-energy spectrum. Since Table II's chemical potentials are differences of U1 and 2U2 terms, any i,j-dependence in the overlaps changes the predicted jumps. The U2 overlaps are explicitly without a sum rule, so the odd-ν structure is a motivated estimate. The prefactor in Eq. (13) is also known only up to a numerical factor, so the normalization of the sum rule itself is approximate. None of this kills the qualitative mechanism, but it means the quantitative cascade structure is not derived.

For whom: theorists interested in random Dirac materials or TBG correlated states. A serious referee should take it; the idea is testable (large-scale-averaged tunneling DOS should show the divergent DOS, and the dependence on averaging length is a falsifiable trend). My recommendation: send it to peer review with a request for either a proof of the individual-overlap equality or a numerical check of Eq. (16) in a simplified model. It is a conditional accept at best.

Referee Report

4 major / 4 minor

Summary. The manuscript proposes that twist-angle disorder in twisted bilayer graphene, represented as a random gauge potential in a low-energy Dirac Hamiltonian, is strongly amplified near the magic angle. The authors argue that the resulting disorder-averaged density of states diverges as rho(E) ~ E^{(2/z)-1} with z >> 1, so occupied states cluster near zero energy. They derive a disorder-averaged sum rule for two-point density overlaps and use it to assert that each added electron contributes an equal intraband Coulomb interaction. Flavor symmetries then yield interband interactions, and a Hartree-Fock calculation produces chemical-potential jumps at integer fillings, which they identify with the observed cascade transitions. The paper also connects the sharpness and position of the cascade peaks to the measurement length scale relative to the disorder correlation length.

Significance. If established, the paper would provide a disorder-driven mechanism for the cascade transitions, with a falsifiable prediction that the transitions sharpen and move to integer fillings when averaged over scales L >> l0. The technical content is substantial: the symmetry classification in Table I, explicit E=0 eigenstates, the replica calculation of the DOS and of rho^2_epsilon, and the sum-rule apparatus are nontrivial and go beyond a phenomenological fit. However, the central step from the total sum rule to individual overlap equality is not a proof, and the U2 overlap is explicitly asserted without an exact sum rule. Because the Hartree-Fock jump magnitudes and the very existence of jumps at odd fillings depend on these individual overlaps, the central claim is not yet established. The paper is nonetheless a serious candidate explanation, and the gaps are local enough to be addressed by further derivation or numerical checks.

major comments (4)

[§III, Eqs. (14)–(16)] The deduction that every individual disorder-averaged overlap equals delta_{ij}/(pi l0^2) does not follow from Eq. (14). The sum rule controls only the total sum over |E_i|,|E_j|<mu of the overlaps; it is satisfied by many non-uniform distributions. The N -> N+1 increment fixes the change in the total, not the identity of the new terms, and the E=0 anchor constrains only a measure-zero subset of states. Since Table II is built from U1 and 2U2 obtained from these individual overlaps, the predicted cascade jumps depend on this equality. Please provide a derivation of the individual overlaps, or a concentration/self-averaging argument, or test Eq. (16) numerically on the random Dirac equation.
[§III, U2 after Eq. (16); Table II] The interband overlap U2 is explicitly said to have "no exact sum rule"; it is motivated by a displaced-overlap calculation Q(r0) in the supplementary that concerns E=0 states with r0 >> l0. Extending this to all E != 0 states and to all pairs between H++ and H+-/H-- is an ansatz. The odd-filling jumps (2U2) in Table II therefore rest on a weaker foundation than the intraband jumps. This should be stated as an assumption, or supported by a sum rule / numerical check.
[§II, Eqs. (8)–(14)] The normalization of the sum rule (14) inherits the prefactor uncertainty acknowledged above Eq. (8): the variational method gives the exponent but not the prefactor. Eq. (13) uses sigma_c(2K, eps_bar^2) ~ alpha^{-2} eps_bar^{2/z}, which is stated only up to a numerical factor, so the value of U1 is not fixed quantitatively. In addition, the inference from Eq. (10) that only low energies are occupied uses alpha^2/A_m < 1, justified only by the absence of an unobserved jump. This does not invalidate the qualitative cascade picture, but the word "prove" in "we prove a sum rule" overstates the status of Eq. (14).
[§I, Eq. (5); Supplementary Eq. (27)] The effective Hamiltonian is obtained by projecting to first order in k and A~(r) (Supplementary Eq. (27)), but the central regime sigma ~ 10^3–10^4 corresponds to A(r) amplified to values comparable to or larger than the moiré momenta. The first-order projection may break down in exactly the regime where the paper claims the disorder becomes huge. A concrete check would be a comparison with a tight-binding or continuum model with the same delta-theta(r) statistics to verify Eq. (5) quantitatively before the DOS/HF analysis is built on it.

minor comments (4)

[§III, Eq. (16)] The same symbol U1 is used for both delta_{ij} and delta_{i,-j}; use U1^+ and U1^- to avoid ambiguity.
[Table II] The table caption should define the flavor notation (↑/↓, ±±, sign(E)); as printed the entries are difficult to parse.
[§III, Eq. (14)] The counting behind Eq. (14) is not fully explained: Eq. (12) sums over all pairs with |E_i|,|E_j|<mu, while N is introduced as the number of electrons in the partially filled E>=0 band. Clarify whether the sum is over N terms or N^2 terms.
[General] Typos and minor presentation issues: "scaler V(r)" in §III; "density verlaps" in the supplementary heading; the inline equation for H_eff in Eq. (5) is run together.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the cascade prediction is not fitted to the target data, and the self-cited disorder/DOS results are independent prior derivations.

full rationale

The paper's central chain—random twist-angle disorder generating a random gauge potential, a huge disorder parameter σ near the magic angle, a divergent DOS, sum rules on disorder-averaged eigenfunctions, and Hartree-Fock chemical-potential jumps—is not circular in the evidentiary sense required here. The disorder parameter σ is estimated from independent experimental twist-angle disorder data (Ref. 2) and the deformation-potential constant β≈3 from prior published work; the cascade data are not used to set σ or U. The DOS exponent z(K, σ) and the freezing behavior are imported from the authors' earlier random-gauge Dirac derivations [12], which are parameter-free for that model and do not assume the TBG cascade result, so these self-citations are real independent support rather than a self-justifying loop. The sum rule (14) is derived from consistency between fermionic and bosonic forms of the disorder-averaged DOS, not from the target interaction phenomenon; it is a rearrangement of the DOS calculation. The main weak point—deducing that every individual overlap equals δ_{ij}/πℓ0² from the aggregate sum rule (14), leading to Eq. (16) and the Table II jumps—is an unsupported inference (a logical gap), not an equivalence by construction or a fitted parameter renamed as a prediction. The paper even flags the weaker U2 estimate with “we do not have an exact sum rule; yet this can be motivated.” Thus no circular step meets the standard of being exhibited as Eq. X = Eq. Y by construction or a fitted input called a prediction.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the random-gauge disorder framework and several unproven inductive steps; see free parameters and axioms above.

free parameters (6)

disorder correlation length l0 = 0.13 to 0.5 um (estimated from Ref. 2 gradients)
Sets the disorder parameter and the U2 overlap scale; chosen from experimental data, not derived.
twist disorder amplitude gamma = about 5e-4
Inferred from standard deviations 0.025 and 0.022 degrees in Ref. 2; controls the disorder parameter.
local Coulomb interaction U = not fixed
Defined as Integral V_c d^2r / l0^2; sets the overall scale of chemical potential jumps but is not predicted.
proximity to magic angle, v/v_F = about 1/40, assuming delta_theta = 0.025 deg as closest approach
Enhances the disorder parameter to 10^3 to 10^4; depends on the sample-specific mean angle.
ratio alpha^2 / A_m = <1, assumed
Assumed so that the chemical potential remains small; not measured.
prefactor of sigma_c(K, epsilon_bar) = unknown, 'up to a numerical prefactor'
DOS prefactor inferred from the classical K=0 limit; affects absolute values.

assumptions (6)

domain assumption Effective single Dirac cone with random gauge potential Eq. (5) is valid; interlayer tunnelling renormalizes velocity and A(r) is locally constant over the moire scale.
Requires l0 >> L_m and perturbative expansion in u0, u1; near the magic angle higher-order terms may matter.
domain assumption Random gauge potential is Gaussian, transverse, and has power-law correlations; replica bosonization applies.
Framework taken from prior random Dirac literature; not re-derived for twist disorder.
standard math The variational/freezing result z = K(sqrt(8 sigma) - 1) and sigma_c ~ epsilon_bar^{2/z} applies with K=1 and with prefactors from the classical limit.
Takes prior same-author results as benchmarks; the prefactor is coarse.
domain assumption Particle-hole symmetry C = eta_x sigma_x K_bar is exact for the disordered Hamiltonian.
Used to relate E>0 and E<0 states; small theta rotations are neglected.
domain assumption Each of the 8 flavor bands has a single Dirac cone and the same disorder-averaged DOS divergence.
Justified by the symmetry table, but no explicit proof that disorder does not mix flavors.
ad hoc to paper All individual overlaps in Eq. (16) equal delta_{ij}/(pi l0^2) or delta_{i,-j}/(pi l0^2).
This is the load-bearing unproven deduction from the sum rule; not established.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Randomly twisted bilayer graphene -- the cascade transitions." pith.science (2026). https://pith.science/paper/YY3YEAEP

@misc{pith2026250807024,
  author       = {Pith},
  title        = {Pith review of: Randomly twisted bilayer graphene -- the cascade transitions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YY3YEAEP}},
  note         = {Machine review of arXiv:2508.07024}
}

abstract

Twisted bilayer graphene (TBG) is known to have disorder in its twist angle. We show that in terms of a Dirac equation with a random gauge potential ${\bf A}({\bf r})$ this disorder becomes huge when the average twist angle is near the magic angle where the Dirac velocity vanishes. The density of states (DOS) then diverges at the Dirac point as $\rho(E)\sim E^{(2/z)-1}$ with $z\gg 1$ and we deduce that all electrons occupy energies very near $E=0$. We prove a sum rule on the disorder averaged eigenfunctions from which we deduce that each added electron contributes equal intraband Coulomb interaction energy. The various bands in TBG are related by either ${\bf A}({\bf r})\rightarrow {\bf A}({-\bf r})$ or ${\bf A}({\bf r})\rightarrow -{\bf A}({\bf r})$ which affects the interband interaction energy. We find, within Hartree-Fock, jumps in the chemical potential at each integer filling, accounting for the cascade transitions.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

42 extracted references · 40 canonical work pages

[3]

Since the range of the eigenstates is ℓ0 is much longer we can use a local interaction defined byU ≡ R Vc(r)d2r/ℓ2

of d/ϵB ≈ 10nm, where d is the distance to the gate the ϵB is the dielectric constant of the spacer h-BN. Since the range of the eigenstates is ℓ0 is much longer we can use a local interaction defined byU ≡ R Vc(r)d2r/ℓ2

work page
[1]

The velocity v coincides with the original [1] result when u0 = u1

, A(r) = 1 + 3¯u2 1 1 − 3¯u2 1 ˜A(r)(5) where ¯ui = ui 3a 4πvF θ , i= 1 , 2. The velocity v coincides with the original [1] result when u0 = u1. It indicates the presence of a magic angle, and is in fact very close to the exact (numerical) result [1–3] for A = 0. It is remarkable that at the magic angle 3¯ u2 1 → 1 the coefficient of A(r) remains finite, ...

work page
[2]

A. Uri, S. Grover, Y. Cao, J. A. Crosse, K. Bagani, D. Rodan-Legrain, Y. Myasoedov, K. Watanabe, T. Taniguchi, P. Moon, M. Koshino, P. Jarillo-Herrero and E. Zeldov, Nature 581, 47 (2020)

work page 2020
[4]

the eigenstates are now 16 long spinors φiα(r)

The average interaction is for occupied states in the flavors α (including spin) with i are states within each band, i.e. the eigenstates are now 16 long spinors φiα(r). ⟨H⟩int = 1 2 U ℓ2 0 Z d2r X ij,αβ [φ† iα(r)φiα(r)φ† jβ (r)φjβ (r) −φ† jβ (r)φiα(r)φ† iα(r)φjβ (r)] (17) The first term is the direct (Hartree) interaction and the second one is the exchan...

work page
[6]

Zondiner, A

U. Zondiner, A. Rozen, D. Rodan-Legrain, Y. Cao, R. Queirpz, T. Taniguchi, K. Watanabe, Y. Oreg, F. von- Oppen, A. Stern, E. Berg, P. Jarillo-Herrero, and S. Ilani, Nature 582, 203 (2020)

work page 2020
[7]

A. T. Pierce, Y. Xie, J. M. Park, E. Khalaf, S. H. Lee, Y. Cao, D. E. Parker, P. R. Forrester, S. Chen, K. Watan- abe, T. Taniguchi, A. Vishwanath, P. Jarillo-Herrero and A. Yacoby, Nat. Phys. 17, 1210–1215 (2021)

work page 2021
[8]

J. M. Park, Y. Cao, K. Watanabe, T. Taniguchi, and P. Jarillo-Herrero, Nature 592, 43 (2021)

work page 2021
[9]

Saito, F

Y. Saito, F. Yang, J. Ge, X. Liu, T. Taniguchi, K. Watan- abe, J. I. A. Li, E. Berg and A. F. Young, Nature 593, 220 (2021)

work page 2021

Show all 42 references

[10]

Its longitudinal part can be gauged away as a random phase, immaterial in the following
[11]

A. W. W. Ludwig, M. P. A. Fisher and G. Grinstein, Phys. Rev. B50, 7526 (1994)

1994
[13]

D. R. Nelson and B. I. Halperin, Physical Review B 19, 2457 (1979)

1979
[14]

Guinea, B

F. Guinea, B. Horovitz and P. Le Doussal, Phys. Rev. B77, 205421 (2008)

2008
[16]

N. N. T. Nam and M. Koshino, Phys. Rev. Bbf 96, 075311 (2017)

2017
[18]

Supplementary material
[19]

Andrei Bernevig, Zhi-Da Song, Nicolas Regnault, and Biao Lian, Phys

B. Andrei Bernevig, Zhi-Da Song, Nicolas Regnault, and Biao Lian, Phys. Rev. B 103, 205413 (2021)

2021
[20]

Sheffer, R

Y. Sheffer, R. Queiroz and A. Stern, Phys. Rev. X 13, 021012 (2023)

2023
[21]

Christos, S

M. Christos, S. Sachdev and S. Scehurer, PNAS 117, 29543 (2020)

2020
[22]

Giamarchi, Quantum Physics in One Dimension (Clarendon Press, Oxford 2003)

T. Giamarchi, Quantum Physics in One Dimension (Clarendon Press, Oxford 2003)

2003
[23]

Scheidl, Phys

S. Scheidl, Phys. Rev. B 55, 457 (1997)

1997
[24]

Carpentier and P

D. Carpentier and P. Le Doussal, Phys. Rev. Lett. 81, 2558 (1998); Nucl. Phys. B 588, 565 (2000)

1998
[25]

Carpentier and P

D. Carpentier and P. Le Doussal, Phys. Rev. E63, 026110 (2001)

2001
[26]

Horovitz and P

B. Horovitz and P. Le Doussal, Phys. Rev. Lett.,84, 5395 (2000); Phys. Rev. B, 71, 134202 (2005)

2000
[27]

H. E. Castillo, C. de C. Chamon, E. Fradkin, P. M. Gold- bart and C. Mudry, Phys. Rev. B 56, 10668 (1997)

1997
[28]

Rozen, J

A. Rozen, J. M. Park, U. Zondiner, Y. Cao, D. Rodan- Legrain, T. Taniguchi, K. Watanabe, Y. Oreg, A Stern, E. Berg, P. Jarillo-Herrero and S. Ilani, Nature 592, 214 (2021)

2021
[29]

Zhang, S

Z. Zhang, S. Wu, D. C˘ alug˘ aru, H. Hu, T. Taniguchi, K. Wanatabe, A. B. Bernevig and E. Y. Andrei, arXiv:2503.17875 Randomly twisted bilayer graphene – the cascade transitions Supplementary material Baruch Horovitz 1 and Pierre Le Doussal 2 1Department of Physics, Ben Gurion...

arXiv
[30]

using ( h−1 j )2 = 1 /(vF kθ)2 and defining ¯ui = ui/vF kθ, i= 1, 2. Consider next the effective Hamiltonian projected to the ψ0 space, to 1st order in −i∇, ˜A(r), defined by ⟨ψ0|Hef f|ψ0⟩ ∼ ⟨Ψ0 + O(k, ˜A)|vF σ · [−i∇1 − ˜A(r)D]|Ψ0 + O(k, ˜A)⟩ (27) where D is a diagonal matrix...
[31]

⟨Ψ0|vF σ · [k1 − ˜A(r)D]|Ψ0⟩ = vF 1 + 3(¯u2 0 + ¯u2
[32]

⟨ψ0|σ · [k − ˜A(r)] + X j Tj(h−1 j )†σ · [k + ˜A(r)]h−1 j Tj|ψ0⟩ = vF 1 + 3(¯u2 0 + ¯u2
[33]

(5) in the main text

⟨ψ0|σ · [k − ˜A(r)] − 3¯u2 1σ · [k + ˜A(r)]|ψ0⟩ (28) This then identifies the reduced Hamiltonian, Eq. (5) in the main text. We next interchange the two layers with an operator ηx, i.e. ηxH(1)ηx = σ · [−i∇r + ˜A(r)] e −i(θ/4)σz ˆT †(r)e−i(θ/4)σz ei(θ/4)σz ˆT (r)ei(θ/4)σz σ · [...
[34]

− 2 logZ(1) which leads to typical exponent −2 for σ < 1, 4( √σ − 1 − σ/2) for 1 < σ <2 and 4 p σ/2( √ 2 − 2) for σ >2. Consider next Q1,1 = R d2r e2V (r)+2V (−r) R d2r e2V (r) R d2r e2V (−r) = Z d2r 2n−2 Π i=1 d2ri e2V (r)+2V (−r)+2 P2n−2 i=1 V (ri) (45) where in the denomina...
[35]

Bistrizer and A

R. Bistrizer and A. H. MacDonald, Proc. Natl. Acad. Sci. USA 108, 12233 (2011)

2011
[36]

Ceferino and F

A. Ceferino and F. Guinea, 2D Mater. 11, 035015 (2024)

2024
[37]

M. M. Al Ezzi, G. N. Pallewela, C. De Beule, E. J. Mele and S. Adam, Phys. Rev. Lett. 133, 266201 (2024)

2024
[38]

Sheffer, R

Y. Sheffer, R. Queiroz and A. Stern, Phys. Rev. X 13, 021012 (2023), Appendix B

2023
[39]

Andrei Bernevig, Zhi-Da Song, Nicolas Regnault, and Biao Lian, Phys

B. Andrei Bernevig, Zhi-Da Song, Nicolas Regnault, and Biao Lian, Phys. Rev. B 103, 205413 (2021), Eqs. A11-A15

2021
[40]

A. W. W. Ludwig, M. P. A. Fisher and G. Grinstein, Phys. Rev. B 50, 7526 (1994)

1994
[41]

Scheidl, Phys

S. Scheidl, Phys. Rev. B 55, 457 (1977)

1977
[42]

Carpentier and P

D. Carpentier and P. Le Doussal, Phys. Rev. E 63, 026110 (2001)

2001
[43]

H. E. Castillo, C. de C. Chamon, E. Fradkin, P. M. Goldbart and C. Mudry, Phys. Rev. B 56, 10668 (1997)

1997
[44]

Y. V. Fyodorov, J. Stat. Mech. P07022 (2009)

2009
[45]

Horovitz and P

B. Horovitz and P. Le Doussal, Phys. Rev. B 71, 134202 (2005)

2005
[46]

Horovitz and P

B. Horovitz and P. Le Doussal, Phys. Rev. B 65, 12532 (2002)

2002

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [3]

Since the range of the eigenstates is ℓ0 is much longer we can use a local interaction defined byU ≡ R Vc(r)d2r/ℓ2

of d/ϵB ≈ 10nm, where d is the distance to the gate the ϵB is the dielectric constant of the spacer h-BN. Since the range of the eigenstates is ℓ0 is much longer we can use a local interaction defined byU ≡ R Vc(r)d2r/ℓ2

work page

[2] [1]

The velocity v coincides with the original [1] result when u0 = u1

, A(r) = 1 + 3¯u2 1 1 − 3¯u2 1 ˜A(r)(5) where ¯ui = ui 3a 4πvF θ , i= 1 , 2. The velocity v coincides with the original [1] result when u0 = u1. It indicates the presence of a magic angle, and is in fact very close to the exact (numerical) result [1–3] for A = 0. It is remarkable that at the magic angle 3¯ u2 1 → 1 the coefficient of A(r) remains finite, ...

work page

[3] [2]

A. Uri, S. Grover, Y. Cao, J. A. Crosse, K. Bagani, D. Rodan-Legrain, Y. Myasoedov, K. Watanabe, T. Taniguchi, P. Moon, M. Koshino, P. Jarillo-Herrero and E. Zeldov, Nature 581, 47 (2020)

work page 2020

[4] [4]

the eigenstates are now 16 long spinors φiα(r)

The average interaction is for occupied states in the flavors α (including spin) with i are states within each band, i.e. the eigenstates are now 16 long spinors φiα(r). ⟨H⟩int = 1 2 U ℓ2 0 Z d2r X ij,αβ [φ† iα(r)φiα(r)φ† jβ (r)φjβ (r) −φ† jβ (r)φiα(r)φ† iα(r)φjβ (r)] (17) The first term is the direct (Hartree) interaction and the second one is the exchan...

work page

[5] [6]

Zondiner, A

U. Zondiner, A. Rozen, D. Rodan-Legrain, Y. Cao, R. Queirpz, T. Taniguchi, K. Watanabe, Y. Oreg, F. von- Oppen, A. Stern, E. Berg, P. Jarillo-Herrero, and S. Ilani, Nature 582, 203 (2020)

work page 2020

[6] [7]

A. T. Pierce, Y. Xie, J. M. Park, E. Khalaf, S. H. Lee, Y. Cao, D. E. Parker, P. R. Forrester, S. Chen, K. Watan- abe, T. Taniguchi, A. Vishwanath, P. Jarillo-Herrero and A. Yacoby, Nat. Phys. 17, 1210–1215 (2021)

work page 2021

[7] [8]

J. M. Park, Y. Cao, K. Watanabe, T. Taniguchi, and P. Jarillo-Herrero, Nature 592, 43 (2021)

work page 2021

[8] [9]

Saito, F

Y. Saito, F. Yang, J. Ge, X. Liu, T. Taniguchi, K. Watan- abe, J. I. A. Li, E. Berg and A. F. Young, Nature 593, 220 (2021)

work page 2021