betaFeedback

WorksPhysics and Astronomy

Quarter-Metal Phases in Multilayer Graphene: Ising-XY and Annular Lifshitz Transitions

Abstract

Recent experiments have uncovered a distinctive magnetic metal in lightly-doped multilayer graphene, coined the quarter metal. This quarter metal consolidates all the doped carriers, originally distributed evenly across the four (or twelve) Fermi surfaces of the paramagnetic state, into one expansive Fermi surface by breaking time-reversal and/or inversion symmetry. In this work, we map out a comprehensive mean-field phase diagram of the quarter-metal in rhombohedral trilayer graphene within the four dimensional parameter space spanned by the density nen_e, interlayer electric potential UU, external magnetic field parallel to the two-dimensional material plane B∥B_{\parallel} and Kane-Mele spin-orbit coupling λλ. We found an annular Lifshitz phase transition and a Ising-XY phase transition and locate these phase boundaries on the experimental phase diagram. The movement of the Ising-XY phase boundary offers insights into λλ. Our analysis reveals that it moves along the line ∂ne/∂B∥∼−0.5×1011cm−2T−1\partial n_e/\partial B_{\parallel} \sim -0.5\times 10^{11} \text{cm}^{-2}\text{T}^{-1} within the nen_e-B∥B_{\parallel} parameter space when λ=30μλ=30μeV. Additionally, we estimated the in-plane spin susceptibility of the valley-Ising quarter-metal χ∥∼8 μeV T−2χ_{_\parallel}\sim 8~μ\text{eV} ~\text{T}^{-2}. Beyond these quantitative findings, two general principles emerge from our study: 1) The valley-XY quarter metal's dominance in the ne−Un_e-U parameter space grows with an increasing number of layers due to the reduce valley polarization variations within the Fermi sea. 2) Layer polarization near the band edge plays an important role in aiding the re-entrance of the paramagnetic state at low density. The insights derived from the quarter metal physics may shed light on the complex behaviors observed in other regions of the phase diagram.

Introduction

The recent discoveries of symmetry-broken electronic phases [1, 2, 3, 4, 5] in the density-interlayer potential (ne−Un_e-U) parameter space of multilayer graphene have opened new avenues to explore strongly correlated physics [6, 7, 8, 9, 10, 11, 12, 13, 14, 15] and energy-efficient electronic devices [16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. The observed phases include spin-singlet superconductor, spin-polarized superconductors and generalized metallic ferromagnets, where the order parameter is characterized by the 15 generators of the SU(4)SU(4) space spanned by spin (ss) and valley (τ\tau) degrees of freedom. Among the various magnetic phases, the so-called quarter metal is particular noteworthy because of its simplicity and its analogies to the relatively well-understood ν=±1\nu= \pm1 graphene quantum Hall magnet. In the large hole-density regime of the quarter metal, the area enclosed by the Fermi surface extracted from magnetic oscillation data precisely corresponds to 4π24\pi^2 times the carrier density. This implies that, despite the 4-fold degenerate density-of-states implied by time-reversal and inversion symmetry, all holes are enclosed by a single Fermi surface. Progressing to lower hole density in this phase, Zhou et al. [1] found a Lifshitz transition where the simple Fermi sea becomes an annular Fermi sea. Despite its extensive capability, magnetic oscillations on its own does not provide insight into the nature of the order-parameter. Indeed, the order parameter can rotate in 15 the dimensional space without altering the area enclosed by the Fermi surface(s). Such variations in the order-parameter leads to variations in groundstate energy, which we found to be much smaller than the Coulomb energy scale. We termed this energy landscape as magnetic anisotropy energy landscape, and it holds pivotal importance in the study of magnetism. The intrinsic spin-orbit coupling (SOC) of graphene profoundly influence this magnetic anisotropy energy landscape. Since Kane and Mele first introduced it [16], numerous experiments have employed varied techniques to measure the SOC [31, 32, 33, 34, 35, 36, 37].

In this article, we present a comprehensive study of quarter metal phase diagram in the parameter space spanned by nen_e, UU, in-plane magnetic field B∥B_{\parallel} and SOC in rhombohedral trilayer graphene (RTG). Our findings indicate two distinct magnetic ground states for the quarter metal, termed the valley-XY phase and valley-Ising phase, separated by a first order phase transition. The momentum distribution of their holes either forms a conventional Fermi sea or adopts an annular shape, with the two types being separated by an annular Lifshitz transition (ALT). This ALT corresponds to the enhanced resistance (depicted in yellow) line running from (ne,D)∼(−4.5×1011 cm−2,0.35V/nm)(n_e,D) \sim(-4.5 \times10^{11}\,\mathrm{cm}^{-2},0.35\mathrm{V/nm}) to (ne,D)∼(−4.2×1011 cm−2,0.3V/nm)(n_e,D) \sim(-4.2 \times10^{11}\,\mathrm{cm}^{-2},0.3\mathrm{V/nm}) in Figure 1 of Ref. [3]. The microscopic origin of magnetic anisotropic energy is attributed to the non-identical nature of wavefunctions across opposite valleys, SOC and UU. Notably, the Ising-XY phase-boundary shift generated by an in-plane field B∥B_{\parallel} is strongly influenced by the strength of Kane-Mele spin-orbit coupling λ\lambda. This correlation enables us to relate the slope dne/dB∥dn_e/dB_{\parallel} to the strength of the spin-orbit coupling, as shown in Figure 7. Additionally, we have calculated the in-plane magnetic susceptibility within the valley-Ising phase. These theoretical outcomes can be readily corroborated with experimental data to provide deeper insights into graphene magnetism.

Displacement of the valley-Ising and valley-XY phase boundary

Figure 1. The displacement of the valley-Ising and valley-XY phase-boundary induced by in-plane magnetic field depends on the strength of spin-orbit coupling λ\lambda. Experimental observation of these shifts can be used to estimate λ\lambda.

Phase boundary and magnetization ratio

Figure 7. a) The slope of the phase boundary for nen_e, U=−2.9×1011 cm−2U=-2.9\times10^{11}\ \mathrm{cm}^{-2}, 18 meV indicates that the rate at which B∥B_{\parallel} lowers the energy of valley XY state is less than the rate at which λ\lambda lowers the energy of the valley Ising state. b) M∥XYμXY−μI\frac{M_{\parallel}^{XY}}{\mu_{XY}-\mu_I} is calculated along the phase boundary which is equal to the slope ∂ne∂B∥\frac{\partial n_e}{\partial B_{\parallel}} indicated in Figure 1 for the specific nen_e and UU.

The remainder of this article is structured as follows: In Sec. II, we study the quarter metal phase diagram without SOC. Sec. III discuss the role of layer-polarization in assisting the reentrance of paramagnet state at low densities. Sec. IV examines the impact of SOC on the magnetic states within the quarter metal. In Sec. V, we use the displacement of the Ising-XY phase boundary to estimate the magnitude of SOC. In Sec. VI, we compare the quarter metal phase diagram between Bernal bilayer graphene and rhombohedral trilayer graphene. Finally, we summarize our work in Sec. VII.

Quarter-Metal Phase Diagram without Spin-Orbit Coupling

In this section, we study the evolution of the quarter-metal groundstate in the ne−Un_e-U parameter space when spin-orbit coupling is neglected. We first discuss the bandstructure of the valley-Ising phase ∣θv=0⟩\lvert\theta_v=0\rangle and the valley-XY phase ∣θv=π/2⟩\lvert\theta_v=\pi/2\rangle obtained from self-consistent mean-field calculation. Then we compare their band-energy and Fock exchange-energy evolution as a function of nen_e. These analysis reveal that the first-order phase transition between ∣θv=π/2⟩\lvert\theta_v=\pi/2\rangle and ∣θv=0⟩\lvert\theta_v=0\rangle is driven by the decrease of exchange energy associated with the winding of pseudospin enforced by the band topology of multilayer graphene.

In the absence of spin-orbit coupling, the spin-quantization axis of a spin-polarized groundstate can be rotated without any energy cost. In contrast, there is a preferred direction to polarize valley degree of freedom because the wavefunctions and energy dispersion are different in the two valleys.

We use self-consistent Fock approximation to estimate the groundstate energy of the quarter metal state. The itinerant exchange effect, as described by Herring [38], is arguably the most important factor in shaping the total energy of a quarter metal. This assertion is supported by the oscillation data, which strongly suggest that time-reversal and/or inversion symmetry of the quarter metal is broken to maximize its pseudospin polarization. Given these observations, it seems likely that the self-consistent mean-field approximation would work better for the quarter metal than for other identified magnetic states, such as the half-metal. The mean field eigenvalue equation is H^k∣ψnk⟩=ϵnk∣ψnk⟩\hat{H}_{\mathbf{k}}\lvert\psi_{n\mathbf{k}}\rangle=\epsilon_{n\mathbf{k}}\lvert\psi_{n\mathbf{k}}\rangle where

H^k=T^k+Σ^kF.(1)\hat{H}_{\mathbf{k}}=\hat{T}_{\mathbf{k}}+\hat{\Sigma}^{F}_{\mathbf{k}}. \tag*{(1)}

The band-Hamiltonian T^k\hat{T}_{\mathbf{k}} is parameterized by the experimentally informed [1] Slonczewski-Weiss-McClure parameters and the Fock self-energy is given by Σ^kF=−∑nk′Vk−k′ρ^k′\hat{\Sigma}^{F}_{\mathbf{k}}=-\sum_{n\mathbf{k}'}V_{\mathbf{k}-\mathbf{k}'}\hat{\rho}_{\mathbf{k}'}. Here VqV_{\mathbf{q}} is the Fourier components of the gate-screened

Coulomb potential Vq=2πkeAtanh⁡(∣q∣d)/(ϵr∣q∣)V_{\mathbf{q}}=\frac{2\pi k_e}{\mathcal{A}}\tanh(\lvert\mathbf{q}\rvert d)/(\epsilon_r\lvert\mathbf{q}\rvert) where A\mathcal{A} is area of the sample, Coulomb constant ke=1.44 eVnmk_e=1.44\,\mathrm{eVnm}, screening constant ϵr=15\epsilon_r=15 and d=5d=5 nm is the distance from the gate to the material. The density-matrix is constructed from the eigenvector and eigenvalue of the self-consistent Hamiltonian:

ρ^k=∑nnF(ϵnk)∣ψnk⟩⟨ψnk∣.(2)\hat{\rho}_{\mathbf{k}}=\sum_n n_F(\epsilon_{n\mathbf{k}})\lvert\psi_{n\mathbf{k}}\rangle\langle\psi_{n\mathbf{k}}\rvert. \tag*{(2)}

The total energy of the self-consistent mean-field approximation is a sum of the band-energy EbθvE_b^{\theta_v} and the Fock exchange energy EexθvE_{\mathrm{ex}}^{\theta_v},

Ebθv=∫d2k4π2Tr⁡[T^kρ^kθv],(3)E_b^{\theta_v}=\int\frac{d^2k}{4\pi^2}\operatorname{Tr}[\hat{T}_{\mathbf{k}}\hat{\rho}_{\mathbf{k}}^{\theta_v}], \tag*{(3)}
Eexθv=12∫d2k4π2Tr⁡[Σ^kFρ^kθv],(4)E_{\mathrm{ex}}^{\theta_v}=\frac{1}{2}\int\frac{d^2k}{4\pi^2}\operatorname{Tr}[\hat{\Sigma}_{\mathbf{k}}^{F}\hat{\rho}_{\mathbf{k}}^{\theta_v}], \tag*{(4)}

where the superscript θv\theta_v stands for the valley-polarization of the converged density matrix:

∫d2k4π2Tr⁡[ρ^kθv(τxx^+τzz^)]∼sin⁡(θv)x^+cos⁡(θv)z^.(5)\int\frac{d^2k}{4\pi^2}\operatorname{Tr}[\hat{\rho}_{\mathbf{k}}^{\theta_v}(\tau_x\hat{x}+\tau_z\hat{z})]\sim\sin(\theta_v)\hat{x}+\cos(\theta_v)\hat{z}. \tag*{(5)}

Here the trace Tr⁡\operatorname{Tr} operates on the 24 dimensional space generated by the 6pz6p_z orbitals, the 2 spin and 2 valley degree of freedoms. More discussion of the Hamiltonian can be found in Ref. [20] and Ref. [39]. We used a Lagrange multipliers to generate the energy landscape associated with order parameter rotations.

As shown in Figure 2, we identified four unique phases of the quarter metal. These phases are characterized by differences in the number of Fermi surfaces and their valley-polarization. They are separated by first-order Ising-XY transition boundaries and first-order annular-Lifshitz transition phase boundaries. The simplest phase among these four appears at large ∣ne∣|n_e| and UU. It has an Ising-like valley-polarization (∣θv=0⟩|\theta_v = 0\rangle or ∣θv=π⟩|\theta_v = \pi\rangle) and all holes are enclosed in a simply-connected Fermi surface (SFS). Designated as the Ising-SFS, this phase is colored blue in Figure 2. Its unconventional relationship between orbital magnetization and valley-polarization is discussed in Ref. [39]. As ∣ne∣|n_e| decreases in the interlayer potential range 30<U<3530 < U < 35 meV, the Ising-SFS undergoes an annular Lifshitz transition (ALT) where an electron-like Fermi surface appears at the centre of the Fermi sea. This leads to an expansion of the hole-like Fermi surface without changing the valley/spin polarization. We refer to this state as the Ising-AFS which is represented as green in Figure 2. In contrast, when ∣ne∣|n_e| decreases at smaller interlayer potential 17<U<3017 < U < 30 meV, we found that the Ising SFS goes through a magnetic phase transition and changes its valley order parameter from θv=0\theta_v = 0 to θv=π/2\theta_v = \pi/2. We dubbed this state as XY-SFS because it enclosed all the holes in a single Fermi surface. Note that the Fermi surface is intrinsically tied to the valley-order parameter: it adopts a C6C_6 symmetry for θv=π/2\theta_v = \pi/2 and C3C_3 for θv=0\theta_v = 0. As ∣ne∣|n_e| reduces to smaller valley-XY-SFS also encounters an ALT. We’ve labelled this phase as XY-AFS, and it is distinguished by purple color in Figure 2.

Mean field phase diagram showing four unique phases of quarter metal

Figure 2. Mean field phase diagram shows 4 unique phases of quarter metal distinguished by their valley-polarization (XY/Ising) and Fermi surface topology (Simple Fermi Sea/Annular Fermi Sea). The evolution of band and Fock energy along the white dashed line, as well as the bandstructure of the black and red points are further expanded in Figure 3.

Figure 3a and b) shows the bandstructure ϵn,kx,ky=0\epsilon_{n,k_x,k_y=0} v.s.v.s. kxk_x for the valley-Ising and valley-XY states respectively. While the low-energy electronic states in multilayer graphene are centered at the zone-corner, we fold the bands into the Γ\Gamma point to visualize valley-mixing [40]. The annular Lifshitz transition appears because the bandstructure has a local minimum at k=0k = 0, as shown in Figure 3b. The valence band-projected valley-polarization shows that the quasiparticles in valley-XY states is a linear superposition of states from opposite valleys. Such valley-mixing is enabled by tripling of the area of the real-space unit cell. In contrast, the valley Ising state merely redistribute the holes all into one valley (say τz=−1\tau_z = -1 in Figure 3a) without changing the valley content of the spinor.

Figure 3. a) Band-structure (along kya=0k_y a = 0) of valley-Ising phase at ne=−2.9×1011 cm−2n_e = -2.9 \times10^{11}\,\mathrm{cm}^{-2} and b) valley-XY phase at ne=−2.5×1011 cm−2n_e = -2.5 \times10^{11}\,\mathrm{cm}^{-2} respectively at U=25U = 25 meV. The chemical potential (μ\mu) is marked by gray horizontal dashed line in both cases. c,d) Band and Fock exchange energy difference of valley-Ising and valley-XY phase per charge as a function of density (marked by white dashed line in Figure 2) shows the energy difference is increasing while approaching charge neutrality. The band energy in Ising phase is higher than valley-XY phase, whereas exchange energy of valley-XY phase is higher than valley-Ising phase. e) Difference in layer pseudospin average LxL_x per charge in valley-Ising and valley-XY state decreases while approaching charge neutrality. This indicates why exchange energy difference between these two states are increasing with decreasing hole density. f) The total energy difference between valley-Ising and valley-XY phase v.s.v.s. nen_e identifies the phase-transition at ne≈−2.78×1011 cm−2n_e \approx-2.78 \times10^{11}\,\mathrm{cm}^{-2}. The positive slope indicates the chemical potential of the Ising state is higher than the XY state.

In order to understand the Ising-XY phase transition, we break down the energy difference between ∣θv=0⟩|\theta_v = 0\rangle and ∣θv=π/2⟩|\theta_v = \pi/2\rangle into the band contribution and exchange contribution across various nen_e. As illustrated by the purple curve in Figure 3c, the band-energy of the Ising state is higher than that of the XY state. This is because the XY state requires a lesser degree of quasiparticle redistribution from the paramagnetic state. Consequently, from a band energy perspective, the XY state is favored over the Ising state as the quarter metal’s magnetic ground state. Importantly, this preference is reinforced at lower ∣ne∣|n_e| thus the band energy contributes to driving the density-induced first-order Ising-XY phase transition. The chemical potential difference between Ising and XY state from the band-energy can be inferred from the slope in Figure 3c,

∂(Ebθv=0−Ebθv=π/2)∂ne≡μbθv=0−μbθv=π/2>0.(6)\frac{\partial\left(E_b^{\theta_v=0}-E_b^{\theta_v=\pi/2}\right)}{\partial n_e} \equiv\mu_b^{\theta_v=0}-\mu_b^{\theta_v=\pi/2}>0. \tag*{(6)}

Figure 3d shows that the XY state’s exchange-energy is elevated compared to the Ising state’s exchange-energy. This energy difference can be traced back to the non-uniform alignment of quasiparticle valley-polarization in the XY state, as shown by the color variations in Figure 3a. However, as ∣ne∣|n_e| decreases, the difference between these exchange energies Eexθv=π/2−Eexθv=0E_{\mathrm{ex}}^{\theta_v=\pi/2}-E_{\mathrm{ex}}^{\theta_v=0} narrows, leading to

∂(Eexθv=0−Eexθv=π/2)∂ne≡μexθv=0−μexθv=π/2>0.(7)\frac{\partial\left(E_{\mathrm{ex}}^{\theta_v=0}-E_{\mathrm{ex}}^{\theta_v=\pi/2}\right)}{\partial n_e}\equiv\mu_{\mathrm{ex}}^{\theta_v=0}-\mu_{\mathrm{ex}}^{\theta_v=\pi/2}>0. \tag*{(7)}

This trend is attributed to the reduction of interlayer coherence Lxθv=π/2−Lxθv=0L_x^{\theta_v=\pi/2}-L_x^{\theta_v=0} with decreasing ∣ne∣|n_e|, as shown in Figure 3e). The interlayer coherence and layer-polarization (per area) is defined as Ljθv=∫d2k4π2Ljθv(k)L_j^{\theta_v}=\int\frac{d^2k}{4\pi^2}L_j^{\theta_v}(k) where j=x,y,zj=x,y,z:

Lxθv(k)=Tr⁡[ρ^kθv(∣A1⟩⟨B3∣+∣B3⟩⟨A1∣)],(8)L_x^{\theta_v}(k)=\operatorname{Tr}\left[\hat{\rho}_k^{\theta_v}\left(|A1\rangle\langle B3|+|B3\rangle\langle A1|\right)\right], \tag*{(8)}
Lyθv(k)=−iTr⁡[ρ^kθv(∣A1⟩⟨B3∣−∣B3⟩⟨A1∣)],(9)L_y^{\theta_v}(k)=-i\operatorname{Tr}\left[\hat{\rho}_k^{\theta_v}\left(|A1\rangle\langle B3|-|B3\rangle\langle A1|\right)\right], \tag*{(9)}
Lzθv(k)=Tr⁡[ρ^kθv(∣A1⟩⟨A1∣−∣B3⟩⟨B3∣)].(10)L_z^{\theta_v}(k)=\operatorname{Tr}\left[\hat{\rho}_k^{\theta_v}\left(|A1\rangle\langle A1|-|B3\rangle\langle B3|\right)\right]. \tag*{(10)}

Here ∣A1⟩|A1\rangle and ∣B3⟩|B3\rangle represent the outermost layer sublattices. Note the antisymmetric layer coherence remains zero for both type of valley-polarization Lyθv=0=Lyθv=π/2=0L_y^{\theta_v=0}=L_y^{\theta_v=\pi/2}=0. For all nen_e and UU, although the layer-polarization is greater in magnitude than layer coherence, for example: Lzθv=0/∣ne∣∼−7.584L_z^{\theta_v=0}/|n_e|\sim-7.584, Lzθv=π/2/∣ne∣∼−7.583L_z^{\theta_v=\pi/2}/|n_e|\sim-7.583 and Lxθv=0/∣ne∣∼−1.283L_x^{\theta_v=0}/|n_e|\sim-1.283, Lxθv=π/2/∣ne∣∼−1.228L_x^{\theta_v=\pi/2}/|n_e|\sim-1.228 at ne,U=−2.9×1011 cm−2,25n_e,U=-2.9\times10^{11}\,\mathrm{cm}^{-2},25 meV, their difference (Lzθv=π/2−Lzθv=0)/∣ne∣∼10−3(L_z^{\theta_v=\pi/2}-L_z^{\theta_v=0})/|n_e|\sim10^{-3} is an order of magnitude smaller than those shown in Figure 3e). This is because the LzL_z is mainly determined by the interlayer potential and the position of the Fermi level.

As a result, while the exchange energy leans towards the Ising state as the magnetic ground state of quarter metal, this inclination becomes less pronounced at smaller ∣ne∣|n_e|. In this sense, the exchange energy also contributes to driving the Ising-XY phase transition as ∣ne∣|n_e| reduces. Figure 3f) plots the total energy difference between Ising and XY and indicated the phase-boundary at ne∼−2.78×1011 cm−2n_e\sim-2.78\times10^{11}\,\mathrm{cm}^{-2}.

Re-entrance of Paramagnetic Metal at Low Density and the Layer Vorticity

In the large interlayer potential (UU) and lowest-density (nen_e) region of the ne−Un_e-U phase-diagram for multilayer graphene, an interesting question arises: Why does the spin-and-valley polarized magnetic quarter-metal transition into a paramagnetic metal as density (∣ne∣|n_e|) decreases, in sharp contrast to the two-dimensional electron gas (2DEG)? In this section, we discuss how the distinctive topological band properties of multilayer graphene leads to a narrowing energy gap between the magnetic states and paramagnetic state as ∣ne∣|n_e| decreases. This trend provides a crucial ingredient in understanding the re-entrance of the paramagnetic phase observed at low densities [1, 41].

Figure 4a) and b) plot the band-energy and exchange-energy per particle of the Ising and XY states relative to the paramagnetic state. The first feature to notice is that the energy difference of the Ising and XY magnetic states relative to the paramagnetic state significantly surpasses the energy gap between the Ising and XY states themselves, i.e., the magnetic anisotropic energy (c.f. Figure 3.c,d and Figure 4.a,b). A second important feature is the swift decline in exchange energy per particle of the magnetic state relative to the paramagnetic state as ∣ne∣|n_e| decreases. Meanwhile, the band-energy difference per particle stays relatively constant. This disparity, which is on the order of a magnitude, contrasts significantly with behaviors seen in 2DEG. In 2DEG, as nen_e decreases, the kinetic energy difference between the fully spin-polarized state and the paramagnetic state,

Band energies, layer polarization, and layer pseudospin orientations

Figure 4. a) The band energy of the Ising state with respect to paramagnetic state decreases while approaching charge neutrality. b) The Ising Fock energy increases linearly with respect to paramagnetic Fock energy as ∣ne∣|n_e| deceases. c) Layer pseudospin orientation of the holes enclosed in a C3C_3 Fermi surface in the valley-Ising quarter metal. As we circle the Fermi surface anticlockwise, the pseudospin rotates by 6π6\pi. Result is obtained at ne,U=−2.9×1011 cm−2,25n_e,U=-2.9\times10^{11}\,\mathrm{cm}^{-2},25 meV. d) Layer pseudospin orientation of the hole-like Fermi pockets shown in one valley in paramagnetic phase at same ne,Un_e,U. e) The layer pseudospin LzL_z of paramagnetic phase changes faster than valley-Ising phase with respect to density, and consequently the exchange energy gap between valley-Ising and paramagnetic phase shrinks towards charge neutrality.

χ(EbFM−EbPM)/ne=1rs2Ry(11)\chi(E_{b}^{FM}-E_{b}^{PM})/n_e=\frac{1}{r_s^2}\mathrm{Ry} \tag*{(11)}

changes more quickly than their Fock energy difference

(EexFM−EexPM)/ne=−82(2−1)3πrsRy,(12)(E_{ex}^{FM}-E_{ex}^{PM})/n_e=-\frac{8\sqrt{2}(\sqrt{2}-1)}{3\pi r_s}\mathrm{Ry}, \tag*{(12)}

where π(rsaB)2=1ne\pi(r_sa_B)^2=\frac{1}{n_e} and Ry=13.6eV\mathrm{Ry}=13.6\mathrm{eV}. We found that such different density-dependence of the band and Fock energies between 2DEG and multilayer graphene arises from the topological band properties of multilayer graphene. In a NN-layer rhombohedral graphene electronic system, the first valence and conduction band can be approximately described by an effective two-orbital Hamiltonian comprised of the outermost layer pzp_z orbitals, e.g. A1−BNA_1-B_N. These two degrees of freedom A1A_1 and BNB_N form a layer pseudospin and it winds 2πN2\pi N around the Brillouin zone corner, leading to a vortex structure in the momentum space. The exchange energy, which favors a uniform distribution of layer pseudospin in momentum space challenges this vortex configuration. See [42] for more details.

Figure 4.d and e) depict the valence-band projected layer pseudospin orientation inside the hole-like Fermi surface for both the valley-Ising and paramagnetic phases. Arrows indicate the in-plane layer-polarization Lx(k)L_x(\boldsymbol{k}), Ly(k)L_y(\boldsymbol{k}) and (c.f Eq. (8)) the color indicates the values of Lz(k)L_z(\boldsymbol{k}). These vortex cores are highlighted by the black dots. The vortex cores are situated close to the valence band maximum and their corresponding layer polarization (Lx,Ly,Lz)∼(0,0,−1)(L_x,L_y,L_z)\sim(0,0,-1). As the pseudospins approach the vortex cores (or the band-edge), they changes rapidly to align their pseudospin towards the (0,0,−1)(0,0,-1) directions.

In the paramagnetic state, the Fermi level is situated very close to the valence band edge with a Fermi energy of ϵF∼2.5\epsilon_F\sim2.5 meV at ne,U=−2.9×1011 cm−2,25n_e,U=-2.9\times10^{11}\ \mathrm{cm}^{-2},25 meV. Thus, LzPML_z^{PM} changes rapidly with decreasing ∣ne∣|n_e|. On the other hand, in the quarter metal (be it Ising or XY), the area of the single hole-like Fermi surface expands by a factor of 12 compared to the area enclosed by the small hole-like Fermi pocket in the paramagnetic state. This causes the Fermi energy to be situated further from the valence band edge even when both states share the same values of nen_e and UU, with ϵF∼4.8\epsilon_F\sim4.8 meV. Therefore, Lzθv=0,π/2L_z^{\theta_v=0,\pi/2} changes slowly with decreasing ∣ne∣|n_e|. In Figure 4.e, we show (Lzθv=0−LzPM)Δex/∣ne∣(L_z^{\theta_v=0}-L_z^{PM})\Delta_{\mathrm{ex}}/|n_e| v.s. nen_e and note that this slope closely resembles the slope in (Eexθv=0−EexPM)/∣ne∣(E_{\mathrm{ex}}^{\theta_v=0}-E_{\mathrm{ex}}^{PM})/|n_e| v.s. nen_e shown in Figure 4.b. Thus, this density-dependence of the layer-polarization can reduce the exchange energy difference between the quarter metal and paramagnetic state thereby significantly influencing the re-emergence of the paramagnetic state.

Effect of Spin-Orbit Coupling On Quarter-Metal

It is not unreasonable to assume that SOC plays a more important role in shaping the magnetic anisotropy landscape compared to the lattice-scale Coulomb interaction [43][44]. When we approximate lattice-scale Coulomb interaction by a delta-function interaction, it is clear that it won’t alter the distribution of particles with identical spin-valley quantization. Notably, unlike the zeroth-Landau level wavefunctions in graphene where sublattice and valley degrees of freedom are identical, the quasiparticle wavefunctions in RTG exhibit rather complex orbital dependence. Nevertheless, a direct mean-field calculation suggests that inter-valley scattering with a magnitude of u⊥=0.5u_\perp=0.5 meV has very small impact on the quarter-metal phase diagram [see Appendix A]. Therefore, our analysis of magnetic anisotropy energy of quarter metal will center on the SOC.

Since spin-orbit coupling that is off-diagonal in layer degree of freedom is suppressed by the application of large electric displacement field, we focus on the simple momentum independent Kane-Mele-like SOC,

H^KM=λσ^zτ^zs^z.(13)\hat{H}_{KM}=\lambda\hat{\sigma}_z\hat{\tau}_z\hat{s}_z. \tag*{(13)}

This Hamiltonian reduces the full SU(2)SU(2) spin-rotation symmetry down to U(1)U(1) symmetry as it only commutes with the generator of szs_z rotation: [H^KM,eiϕs^z]=0[\hat{H}_{KM},e^{i\phi\hat{s}_z}]=0. Note σ^z=∣A1⟩⟨A1∣−∣B3⟩⟨B3∣\hat{\sigma}_z=|A1\rangle\langle A1|-|B3\rangle\langle B3| here is taken as the layer polarization between the A1 - B3 sublattice [42, 45]. This form of SOC can be derived by considering the point group symmetry of rhombohedral trilayer graphene [46, 47].

When adding Eq. (13) (with λ=30 μeV\lambda=30\ \mu\mathrm{eV}) to Eq. 1, and carry out the self-consistent calculations, we obtained the phase diagram shown in Figure 5.a. A quick comparison between Figure 5.a and Figure 2 shows that even a modest value of SOC (λ=30 μeV\lambda=30\ \mu\mathrm{eV}) has a pronounced effect on the ne−Un_e-U phase diagram for the quarter metal. Note λ\lambda has almost no effect on the annular Lifshitz transition (ALT) phase boundaries. This is because the two phases on both side of the ALT phase boundary have identical spin-valley order parameters. Thus, SOC lowers their energy in a similar way. In contrast, the Ising-XY phase boundary is shifted dramatically in the direction that expands the Ising phase, strongly influencing the magnetic anisotropy energy landscape.

Figure 5. a) Modification of the phase diagram in Figure 2 due to a Ising spin-orbit coupling of λ=30 μeV\lambda= 30\ \mu\mathrm{eV}. b) valley and spin texture of valley-Ising state with simply connected Fermi surface (Ising-SFS) for ne,U=−2.9×1011 cm−2n_e,U=-2.9 \times10^{11}\ \mathrm{cm}^{-2}, 18 meV. c) valley and spin texture of valley-XY state with simply connected Fermi surface (XY-SFS) for ne,U=−2.4×1011 cm−2n_e,U=-2.4 \times10^{11}\ \mathrm{cm}^{-2}, 18 meV.

The principal effect of H^KM\hat{H}_{KM} is to introduce correlation between the quasiparticle valley-polarization and its spin-quantization. In order to understand this correlation, we plot the momentum space distribution of spin and valley-polarization for the holes in the Ising and XY state. For the Ising ground state, the spin polarization of all the holes are aligned in the +z+z direction while their valley-polarization are all aligned in the −z-z direction. This state is represented by the notation ∣θv,θs⟩|\theta_v,\theta_s\rangle, specifying the valley-polarization angle θv\theta_v and the spin polarization angle θs\theta_s. This state is called the Ising state because the energies of ∣θv=π,θs=0⟩|\theta_v=\pi,\theta_s=0\rangle and its time-reversed counterpart ∣θv=0,θs=π⟩|\theta_v=0,\theta_s=\pi\rangle are the same. The effect of SOC in the Ising phase is to introduce an uniaxial easy-axis normal to two-dimensional material plane. Note while the spin-quantization axis contains spatial information (in this case, it’s perpendicular to the two-dimensional material plane), the valley-polarization, lacks any spatial interpretation. In addition to the the Ising-state, we identified another stationary solution characterize an order parameter defined by non-zero values of τx\tau_x, sxs_x and szτzs_z\tau_z. We named this stationary solution the vallley-XY, denoted as ∣θs=π/2,θv=π/2⟩|\theta_s = \pi/2,\theta_v = \pi/2\rangle. Note the total energy of this state remains unchanged under rotations generated by eiϕssze^{i\phi_s s_z} and eiϕvτze^{i\phi_v \tau_z} rotations, justifying the valley-XY nomenclature. Figure 5.c shows the valley-polarization and spin-polarization of ∣θs=π/2,θv=π/2⟩|\theta_s = \pi/2,\theta_v = \pi/2\rangle for the holes enclosed inside the C6C_6 Fermi surface. Navigating around the Fermi surface clockwise, the valley-polarization exhibits an oscillatory pattern about the τx\tau_x axis. This specific valley-polarization configuration aims to minimize both the exchange and band energy: it minimizes the exchange energy by predominantly aligning the valley-polarization along the x-direction and introduces a valley-population imbalance across six unique momentum regions to further reduce band energy. The presence of a finite SOC generates a similar oscillatory pattern in spin so that the valley-polarization (⟨ψnk∣τ^∣ψnk⟩\langle\psi_{nk}|\hat{\boldsymbol{\tau}}|\psi_{nk}\rangle) is almost locally anti-parallel to spin-polarization (⟨ψnk∣s^∣ψnk⟩\langle\psi_{nk}|\hat{\boldsymbol{s}}|\psi_{nk}\rangle) in momentum space. Contrary to the Ising state, which reduces its energy linearly with SOC, the XY state decrease its energy quadratically in λ\lambda.

The valley-XY and valley-Ising order parameters are visualized on two Bloch spheres in Figure 6. Arrows touching the Bloch sphere represent maximium polarization, as seen in the valley-Ising state. Note, the valley-XY phase lacks maximum spin and valley polarization because of its non-uniform momentum distribution of spin-polarization and valley-polarization as shown in Figure 5.c.

Bloch spheres showing valley-Ising and valley-XY order parameters

Figure 6. The effect of Kane-Mele spin-orbit coupling (SOC) on the order parameters of valley-Ising and valley-XY quarter metals is illustrated on the two Bloch sphere. The ±z\pm z directions on these spheres correspond to the ∣↑,↓⟩|\uparrow,\downarrow\rangle and ∣K,K′⟩|K,K'\rangle degrees of freedom. In the valley-Ising state, SOC pins the spin and valley polarization in the ±z\pm z and ∓z\mp z directions. Conversely, in the valley-XY phase, it aligns them in directions orthogonal to zz. The XY-phase features two soft-modes, arising from independent rotations of spin and valley order parameter within the magnetic easy-plane. Refer to the maintext (Sec. IV) for discussion on the magnitude of the arrows.

In the Appendix C, we use the Lagrange multiplier H^L(n)=λL[n⋅s^+n⋅(eiπ2τ^xτ^e−iπ2τ^x)]\hat{H}_L(\boldsymbol{n}) = \lambda_L[\boldsymbol{n}\cdot\hat{\boldsymbol{s}}+\boldsymbol{n}\cdot(e^{i\frac{\pi}{2}\hat{\tau}_x}\hat{\boldsymbol{\tau}}e^{-i\frac{\pi}{2}\hat{\tau}_x})] to further expand on the energy landscape of other possible quarter metal phases in the presence of spin-orbit coupling.

Displacement of the Ising-XY phase boundary

In this section, we extend the electronic phase diagram of the quarter metal from the two-dimensional ne−Un_e-U parameter space to the more comprehensive four dimensional ne−U−λ−B∥n_e-U-\lambda-B_{\parallel} space. Here, λ\lambda denotes the Kane-Mele SOC (introduced in the last section) and B∥B_{\parallel} represents the in-plane Zeeman field. In this expanded parameter landscape, we first examine the movement of Ising-XY phase-boundary as a function of λ\lambda and B∥B_{\parallel}, keeping nen_e, UU fixed. Then, we study the movement of this phase-boundary as a function of nen_e and B∥B_{\parallel}, keeping λ\lambda, UU fixed. The latter analysis enables us to correlate λ\lambda to the experimentally observable metric dne/dB∥dn_e/dB_{\parallel}.

We compute the ne−U−λ−B∥n_e-U-\lambda-B_{\parallel} phase diagram by solving the following self-consistent mean-field equation at various density nen_e and interlayer potential UU:

Hk=Tk+ΣkF,Tk=T^k+H^KM+H^B∥,(14)\mathcal{H}_{\mathbf{k}}=\mathcal{T}_{\mathbf{k}}+\Sigma_{\mathbf{k}}^{F},\qquad\mathcal{T}_{\mathbf{k}}=\hat{T}_{\mathbf{k}}+\hat{H}_{KM}+\hat{H}_{B_{\parallel}}, \tag*{(14)}

where T^k\hat{T}_{\mathbf{k}} and ΣkF\Sigma_{\mathbf{k}}^{F} are described in Eq. (1). The Kane-Mele SOC, H^KM\hat{H}_{KM} is defined in Eq. (13) and the Zeeman energy is given by the following

H^B∥=−12gμBs^xB∥.(15)\hat{H}_{B_{\parallel}}=-\frac{1}{2}g\mu_B\hat{s}_xB_{\parallel}. \tag*{(15)}

Here g=2g=2 is the gyromagnetic ratio, μB\mu_B is the Bohr magneton, and s^x\hat{s}_x is the 1st Pauli matrix. The valley Ising-XY phase boundary is defined by the following equation:

Eθv=0(ne,U,λ,B∥)=Eθv=π/2(ne,U,λ,B∥).(16)E^{\theta_v=0}(n_e,U,\lambda,B_{\parallel})=E^{\theta_v=\pi/2}(n_e,U,\lambda,B_{\parallel}). \tag*{(16)}

where the total energy is

Eθv(ne,U,λ,B∥)≡12∫d2k4π2Tr⁡[(T^k+Hk)ρ^kθv].(17)E^{\theta_v}(n_e,U,\lambda,B_{\parallel})\equiv\frac{1}{2}\int\frac{d^2k}{4\pi^2}\operatorname{Tr}\left[(\hat{T}_{\mathbf{k}}+\mathcal{H}_{\mathbf{k}})\hat{\rho}_{\mathbf{k}}^{\theta_v}\right]. \tag*{(17)}

Figure 7a shows the movement of the Ising-XY phase-boundary as a function of B∥B_{\parallel} and λ\lambda for fixed nen_e and UU. To leading order in B∥B_{\parallel}, the total energy of the state ∣θv=π/2,θs=π/2⟩\lvert\theta_v=\pi/2,\theta_s=\pi/2\rangle is lowered while ∣θv=0,θs=π⟩\lvert\theta_v=0,\theta_s=\pi\rangle is unaffected. In contrast, to leading order in λ\lambda, the total energy of the state ∣θv=0,θs=π⟩\lvert\theta_v=0,\theta_s=\pi\rangle is lowered while ∣θv=π/2,θs=π/2⟩\lvert\theta_v=\pi/2,\theta_s=\pi/2\rangle is unaffected. The slope d(gμBB∥)/dλ>1d(g\mu_BB_{\parallel})/d\lambda>1 means SOC is more effective in reducing the energy of the Ising state compared to the energy reduction achieved by the Zeeman effect on the XY-state ∣θv=π/2,θs=π/2⟩\lvert\theta_v=\pi/2,\theta_s=\pi/2\rangle. This is simply because the out-of-plane spin-polarization of the Ising state is greater than the net in-plane spin-polarization of the XY-state, as shown in Figure 5b and c). Note that when the Zeeman field reaches a sufficiently high value (i.e. gμBB∥≳λg\mu_BB_{\parallel}\gtrsim\lambda), the Ising state manifests field-induced in-plane spin-polarization, such that θs=π−δ\theta_s=\pi-\delta and θv=δ\theta_v=\delta, with δ≪π\delta\ll\pi.

Next, we consider the Ising-XY phase-boundary as a function of nen_e and B∥B_{\parallel} for fixed λ\lambda and UU. Since nen_e and B∥B_{\parallel} are thermodynamical conjugate to chemical potential and magnetic-moments, a small variation of density nen_e around Ising-XY boundary at ne∗n_e^* and B∥B_{\parallel} around 0, results in the following energy changes:

Eθv=0(ne,U,λ,B∥)=E∗+μθv=0dne−M∥θv=0dB∥(18)E^{\theta_v=0}(n_e,U,\lambda,B_{\parallel})=E^*+\mu^{\theta_v=0}dn_e-M_{\parallel}^{\theta_v=0}dB_{\parallel} \tag*{(18)}
Eθv=π/2(ne,U,λ,B∥)=E∗+μθv=π/2dne−M∥θv=π/2dB∥(19)E^{\theta_v=\pi/2}(n_e,U,\lambda,B_{\parallel})=E^*+\mu^{\theta_v=\pi/2}dn_e-M_{\parallel}^{\theta_v=\pi/2}dB_{\parallel} \tag*{(19)}

where E∗=Eθv=0(ne∗,U,λ,B∥=0)=Eθv=π/2(ne∗,U,λ,B∥=0)E^*=E^{\theta_v=0}(n_e^*,U,\lambda,B_{\parallel}=0)=E^{\theta_v=\pi/2}(n_e^*,U,\lambda,B_{\parallel}=0) is the equilibrium free energy energy when B∥=0B_{\parallel}=0 is zero. As the spins of the Ising phase are pinned out of the two-dimensional material plane by the Kane-Mele SOC, the in-plane spin-moment vanishes at leading order in B∥B_{\parallel}, M∥θv=0=0M_{\parallel}^{\theta_v=0}=0. Thus, the Ising-XY phase boundary evolves in ne−B∥n_e-B_{\parallel} space according to

∂ne∂B∥=M∥θv=π/2μθv=π/2−μθv=0.(20)\frac{\partial n_e}{\partial B_{\parallel}}=\frac{M_{\parallel}^{\theta_v=\pi/2}}{\mu^{\theta_v=\pi/2}-\mu^{\theta_v=0}}. \tag*{(20)}

Note M∥θv=π/2M_{\parallel}^{\theta_v=\pi/2} is the magnitude of the spin magnetization so the sign of ∂ne∂B∥\frac{\partial n_e}{\partial B_{\parallel}} is determined by the denominator:

μθv=π/2−μθv=0≡μbθv=π/2+μexθv=π/2−μbθv=0−μexθv=0<0.(21)\mu^{\theta_v=\pi/2}-\mu^{\theta_v=0}\equiv\mu_b^{\theta_v=\pi/2}+\mu_{ex}^{\theta_v=\pi/2}-\mu_b^{\theta_v=0}-\mu_{ex}^{\theta_v=0}<0. \tag*{(21)}

We used Eq. (6) and Eq. (7) to establish the above relationship. Figure 7b shows that the ratio M∥θv=π/2/(μθv=π/2−μθv=0)M_{\parallel}^{\theta_v=\pi/2}/(\mu^{\theta_v=\pi/2}-\mu^{\theta_v=0}) has very weak dependence on the strength of SOC λ\lambda. Given that M∥θv=π/2=gμBSxθv=π/2M_{\parallel}^{\theta_v=\pi/2}=g\mu_B S_x^{\theta_v=\pi/2} and considering Sxθv=π/2≈1S_x^{\theta_v=\pi/2}\approx1 per hole at the densities where XYXY state are the groundstate, it is clear that the weak λ\lambda dependence in ∂ne∂B∥\frac{\partial n_e}{\partial B_{\parallel}} is attributed to the small increment of ∣μθv=π/2−μθv=0∣|\mu^{\theta_v=\pi/2}-\mu^{\theta_v=0}| as λ\lambda increases.

Our analysis enables to compute the in-plane spin susceptibility χ∥\chi_{\parallel} of the Ising-state ∣θv=0,θs=π⟩|\theta_v=0,\theta_s=\pi\rangle:

χ∥=∂2Eθv=0∂B∥2∼8 μeV T−2.(22)\chi_{\parallel}=\frac{\partial^2 E^{\theta_v=0}}{\partial B_{\parallel}^2}\sim8\ \mu\mathrm{eV}\ \mathrm{T}^{-2}. \tag*{(22)}

See Appendix B for more about in-plane spin susceptibility. Since 1/χ∥1/\chi_{\parallel} represents the resistance to collectively rotating the spin away from the uniaxial magnetic easy-axis, 1/χ∥1/\chi_{\parallel} increases with the strength of spin-orbit coupling.

Quarter Metals in Bernal Bilayer Graphene and Beyond

So far, our study has primarily focused on the quarter metal phase diagram of rhombohedral trilayer graphene (RTG). In this section, we contrasts the phase diagram of RTG with Bernal bilayer graphene (BBG). BBG is another system where experimental studies [3, 8, 9] have confirmed the presence of the quarter metal. Our objective here is to elucidate the core principles of magnetism that could be universally applicable to all multilayer graphene electron gas systems, while also highlighting specific nuances that are harder to predict. The mean-field phase diagram of BBG is shown in Figure 8, which agrees qualitatively with Ref. [36]. Both RTG and BBG’s quarter metal phase diagrams feature the valley-XY and valley-Ising states, yet they differ significantly in certain aspects. First, quarter metal in BBG occurs at density ∣ne∣|n_e| that is 3 to 5 times smaller than that in RTG. This is consistent with the experimental observations in Ref. [1, 41]. Second, the

Mean-field phase diagram of quarter-metal phases in BBG

Figure 8. Mean-field phase diagram of quarter-metal phases in BBG shows valley-XY phase is the groundstate close to the half-metal. As the hole density decreases, the valley-XY state transitions into the valley-Ising state.

valley-XY phase in BBG occupies a very small region in the ne−Un_e-U phase diagram. Finally, there is a notable distinction in the magnetic anisotropic energy landscape: in RTG, the valley-Ising transitions into the valley-XY as ∣ne∣|n_e| decreases, whereas in BBG, the transition occurs from the valley-XY to the valley-Ising with decreasing ∣ne∣|n_e|. We’ll first address the initial two points using general principles inherent to multilayer graphene systems. Subsequently, we’ll discuss the difference in magnetic anisotropic energy landscape, which exemplifies the nuanced specifics that challenge predictability.

Given that BBG’s quarter metal is observed at a density roughly five times less than RTG, the area of the Fermi sea in BBG contracts by the same factor when compared to RTG. While this reduced Fermi sea area doesn’t alter the spin and valley order parameters of the Ising phase, it does enhance the variation of valley-polarization within the Fermi sea for the valley-XY phase. This variation raises valley-XY’s exchange energy rendering it less stable as the ground state. This underlines the reason why the valley-XY phase occupies a smaller ne−Un_e-U phase space in BBG compared to RTG. To further elucidate this point, we plot the valley-polarization τz\tau_z of the valley-XY phase in BBG in Table I, obtained from self-consistent mean-field calculation. Similar to RTG, as the Bloch states traverse the Fermi surface, the vector τk=(τx,k,τz,k)\boldsymbol{\tau}_{\mathbf{k}}=(\tau_{x,\mathbf{k}},\tau_{z,\mathbf{k}}) exhibits oscillatory behavior reminiscent of a domain wall. Yet, a closer look at the color bar in Table I shows that the variations of τz,k\tau_{z,\mathbf{k}} in BBG are more pronounced than in RTG.

RTGBBG
nen_e−2.7×1011 cm−2-2.7 \times10^{11}\ \mathrm{cm}^{-2}−1.0×1011 cm−2-1.0 \times10^{11}\ \mathrm{cm}^{-2}
Eex/∣ne∣E_{\mathrm{ex}}/|n_e|−4.645 meV-4.645\ \mathrm{meV}−2.349 meV-2.349\ \mathrm{meV}
⟨τx⟩/∣ne∣\langle\tau_x\rangle/|n_e|10.86
⟨τz⟩/∣ne∣\langle\tau_z\rangle/|n_e|00
τz,k\tau_{z,\mathbf{k}}

Table I. Comparison of the valley-XY quarter metal in rhombohedral trilayer graphene (RTG) and Bernal bilayer graphene (BBG). Here nen_e, Eex/∣ne∣E_{\mathrm{ex}}/|n_e| and ⟨τx⟩\langle\tau_x\rangle represent density, the exchange energy per particle (measured respect to the paramagnetic state), and the order parameter, respectively. τz,k=⟨ψnk∣τz,k∣ψnk⟩\tau_{z,\mathbf{k}}=\langle\psi_{n\mathbf{k}}|\tau_{z,\mathbf{k}}|\psi_{n\mathbf{k}}\rangle denotes the valley polarization of the hole-occupied Bloch states and ⟨τz⟩=∑kτz,k=0\langle\tau_z\rangle=\sum_{\mathbf{k}}\tau_{z,\mathbf{k}}=0. Note the stronger variation of τz,k\tau_{z,\mathbf{k}} in BBG raises the exchange energy per hole in comparison to RTG. For these calculations, we set U=28 meVU=28\ \mathrm{meV}.

We now turn our attention to the magnetic anisotropic energy landscapes of quarter metal in RTG and BBG. In both systems, the exchange energy invariably promotes the Ising phase as the ground state, while the band energy gives always preference to the XY phase. However, their mean-field energy differences are minimal and are swayed by detailed factors such as layer polarization. For BBG, our observations indicate that, at large ∣ne∣|n_e|, the band-energy difference is greater than the exchange-energy difference. This means the valley-XY phase is the dominant ground state. In contrast, RTG behaves oppositely.

Summary

In conclusion, we have addressed four pertinent questions concerning the low-density portion of the magnetic phase diagram in rhombohedral trilayer graphene:

  1. Magnetic anisotropy of the quarter-metal groundstate: We identified two competing groundstate for the quarter metal: the valley-Ising state and the valley-XY state, detailed extensively in Sec. II. At large values of ∣ne∣|n_e| and UU, the valley-Ising state is the quarter metal groundstate. As we decrease ∣ne∣|n_e| and UU, it undergoes a first order phase transition to become the valley-XY state. This transition is driven by both the band and exchange energy. Although the exchange effect unambiguously favors the valley-Ising state as the ground state, the exchange energy of the valley-XY state becomes similar to the valley-Ising state with reducing ∣ne∣|n_e|. This narrowing of exchange energy gap between them is due to the subtle differences in their layer polarization.

  1. Sublattice pseudospin polarization and the re-entrance of paramagnetic state: We found that the exchange energy difference between the quarter metal and the paramagnetic state, denoted as EexQ−EexPME_{\mathrm{ex}}^{\mathrm{Q}}-E_{\mathrm{ex}}^{\mathrm{PM}}, can be strongly influenced by the layer-pseudospin. Given that the area of the Fermi sea in the quarter metal is 12 times larger than the Fermi pockets in the paramagnetic state, the Fermi level of the quarter metal is situated further from neutrality compared to that of the paramagnetic state. As a result, as ∣ne∣|n_e| reduces, the layer polarization in the paramagnetic phase increases more rapidly than in the quarter metal phase, resulting in ∂(EexQ−EexPM)/∂ne>0\partial(E_{\mathrm{ex}}^{\mathrm{Q}}-E_{\mathrm{ex}}^{\mathrm{PM}})/\partial n_e>0.

  1. Magnetic anisotropy and Kane-Mele spin-orbit coupling: We examined the influence of Kane-Mele spin-orbit coupling (SOC) on the quarter-metal phase diagram. This SOC term lowers the energy of a many-body state where the z-component of the spin-polarization (szs_z) is anti-parallel to the z-component of the valley-polarization (τz\tau_z). Here z^\hat{z} is the unit vector normal to the two-dimensional material plane. In the case of valley Ising quarter metal, this SOC establishes a magnetic easy-axis. Here, the spin-quantization axis of all Bloch states are uniformly aligned along ±z^\pm\hat{z} and that the valley-polarization is everywhere antiparallel to it, i.e. ∓z^\mp\hat{z}. For the valley-XY quarter metal, this SOC introduces a magnetic easy-plane. In this scenario, both the spin-polarization and valley-polarization of the Bloch states predominantly orient themselves perpendicular to z^\hat{z}. Minor oscillations in τz\tau_z and szs_z are observed within the Fermi sea. This anisotropy indicates that while the Kane-Mele SOC gives preference to the valley-Ising as the ground state, an in-plane magnetic field favors the valley-XY as the ground state.

  1. Displacement of Ising-XY boundary: In our study of the Ising-XY boundary in the four dimensional parameter space — defined by density nen_e, interlayer potential UU, in-plane magnetic field B∥B_{\parallel} and SOC strength λ\lambda — we found that the Ising-XY phase boundary follows the trajectory ∂ne/∂B∥∼−0.5×1011 cm−2T−1\partial n_e/\partial B_{\parallel}\sim-0.5\times10^{11}\,\mathrm{cm}^{-2}\mathrm{T}^{-1} within the nen_e and B∥B_{\parallel} parameter space. Furthermore, we have computed the in-plane spin susceptibility of the valley-Ising quarter-metal to be χ∥∼8 μeV T−2\chi_{\parallel}\sim8\,\mu\mathrm{eV}\,\mathrm{T}^{-2}.

We centered our attention on RTG due to its experimental significance and its representation as the most rudimentary form of rhombohedral-stacked multilayer graphene. It is noteworthy that the density-of-states (DOS) in RTG significantly exceeds that of its ABA Bernal-stacked trilayer graphene counterpart. Levering on this pronounced DOS, [5, 6] has recently identified correlated metals and insulators in rhombohedral-stacked pentalayer graphene even without the application of an external electric displacement field.

Note added: During finalizing our work, we became aware of parallel experimental work focusing on determining the strength of Kane-Mele Spin-Orbit Coupling (SOC) using the effect of B∥B_{\parallel} on phase transition boundaries in RTG. In the quarter-metal region, two competing ground states have been identified: the valley-imbalanced (VI) orbital ferromagnet and the intervalley coherent (IVC) phase, which is equivalent to the valley Ising and valley XY phase, respectively, as discussed in this work. Ref. [31] provides dne/dB∥∼−0.43×1011 cm−2T−1dn_e/dB_{\parallel}\sim-0.43\times10^{11}\,\mathrm{cm}^{-2}\mathrm{T}^{-1}, which is in rough agreement with our finding. This experiment predicts the Kane-Mele SOC strength ∼50 μeV\sim50\,\mu\mathrm{eV}.

Acknowledgement: We acknowledgement insightful conversation with Nemin Wei, Andrea Young and Haoxin Zhou. We are grateful to the University of Kentucky Center for Computational Sciences and Information Technology Services Research Computing for their support and use of the Morgan Compute Cluster and associated research computing resources.

θv\theta_v0π/2\pi/2
ΔEb/∣ne∣\Delta E_b/|n_e| (meV)0.7610.753
ΔEex/∣ne∣\Delta E_{\mathrm{ex}}/|n_e| (meV)−0.713-0.713−0.703-0.703

Table 2. Change in band and Fock exchange energy for valley-Ising and valley-XY phase due to change in u⊥u_\perp from 0 meV to 0.5 meV at (ne,U)=−2.9×1011 cm−2,18 meV(n_e,U) = -2.9 \times10^{11}\ \mathrm{cm}^{-2}, 18\ \mathrm{meV}.

θv\theta_v0π/2\pi/2
E⊥H/∣ne∣E_{\perp H}/|n_e| (meV)00.41
E⊥ex/∣ne∣E_{\perp\mathrm{ex}}/|n_e| (meV)−55.814-55.814−56.228-56.228

Table 3. Hartree and exchange energy component of intervalley scattering for valley-Ising and valley-XY phase using u⊥=0.5u_\perp= 0.5 meV at (ne,U)=−2.9×1011 cm−2,18 meV(n_e,U) = -2.9 \times10^{11}\ \mathrm{cm}^{-2}, 18\ \mathrm{meV}.

Effect of inter-valley scattering on quarter metal groundstates

The landscape of magnetic anisotropy energy can be affected by several independent parameters. In the main text, we have examined four such parameters: nen_e, UU, λ\lambda, and B∥B_\parallel. In this section, we study the impact of inter-valley scattering on quarter-metal ground states and how it alters the nen_e-UU phase diagram outlined in Figure 2. We introduce the self-energy Σ^⊥k\hat{\Sigma}_{\perp\mathbf{k}} into Eq. (1) of the maintext, where

Σ^⊥k=u⊥∑a=x,y[Tr⁡(τ^aρ^kθv)τ^a−τ^aρ^kθvτ^a].(23)\hat{\Sigma}_{\perp\mathbf{k}}=u_\perp\sum_{a=x,y}\left[\operatorname{Tr}(\hat{\tau}_a\hat{\rho}_{\mathbf{k}}^{\theta_v})\hat{\tau}_a-\hat{\tau}_a\hat{\rho}_{\mathbf{k}}^{\theta_v}\hat{\tau}_a\right]. \tag*{(23)}

Here u⊥u_\perp represents the strength of the inter-valley Coulomb scattering. In the above equation, first term denotes the Hartree contribution, while the second one corresponds to the Fock contribution. Their contribution to the total energy is given by the following:

E⊥Hθv=u⊥2∑a=x,y∫d2k(2π)2[Tr⁡(τ^aρ^kθv)]2,(24)E_{\perp H}^{\theta_v}=\frac{u_\perp}{2}\sum_{a=x,y}\int\frac{d^2k}{(2\pi)^2}\left[\operatorname{Tr}(\hat{\tau}_a\hat{\rho}_{\mathbf{k}}^{\theta_v})\right]^2, \tag*{(24)}
E⊥exθv=−u⊥2∑a=x,y∫d2k(2π)2Tr⁡[τ^aρ^kθvτ^aρ^kθv],(25)E_{\perp\mathrm{ex}}^{\theta_v}=-\frac{u_\perp}{2}\sum_{a=x,y}\int\frac{d^2k}{(2\pi)^2}\operatorname{Tr}\left[\hat{\tau}_a\hat{\rho}_{\mathbf{k}}^{\theta_v}\hat{\tau}_a\hat{\rho}_{\mathbf{k}}^{\theta_v}\right], \tag*{(25)}

where ρ^kθv\hat{\rho}_{\mathbf{k}}^{\theta_v} is the density-matrix determined self-consistently. As shown in Table III, despite a notable exchange energy shift related to the Dirac sea, the energy difference between them is very small: E⊥,Hθv=0+E⊥,exθv=0−E⊥,Hθv=π/2−E⊥,exθv=π/2∼4 μeVE_{\perp,H}^{\theta_v=0}+E_{\perp,\mathrm{ex}}^{\theta_v=0}-E_{\perp,H}^{\theta_v=\pi/2}-E_{\perp,\mathrm{ex}}^{\theta_v=\pi/2}\sim4\ \mu\mathrm{eV} per-hole. Furthermore, Table II shows that the variations in both band energy and (long-range Coulomb) exchange energy, due to u⊥u_\perp, are similar for both the Ising and XY states. Consequently, the introduction of inter-valley scattering doesn’t bring about significant alterations to the quarter-metal phase diagram.

Spin susceptibility of valley-Ising quarter metal

We compute the in-plane spin susceptibility (χ∥\chi_\parallel) of the valley-Ising quarter metal for different values of spin-orbit coupling (λ\lambda). χ∥\chi_\parallel is defined as:

χ∥=∂2EB∥=0θv=0∂B∥2∣B∥=0=∂MB∥=0θv=0∂B∥∣B∥=0(26)\chi_\parallel=\left.\frac{\partial^2E_{B_\parallel=0}^{\theta_v=0}}{\partial B_\parallel^2}\right|_{B_\parallel=0}=\left.\frac{\partial M_{B_\parallel=0}^{\theta_v=0}}{\partial B_\parallel}\right|_{B_\parallel=0} \tag*{(26)}

Figure 9 shows χ∥\chi_\parallel v.s. λ\lambda at ne,U=−2.9×1011 cm−2,18 meVn_e,U=-2.9\times10^{11}\ \mathrm{cm}^{-2},18\ \mathrm{meV}. Since the SOC pinned the spin-quantization axis of the valley-Ising state in the direction normal to the applied in-plane field B∥B_\parallel, χ∥\chi_\parallel decreases with increasing λ\lambda.

In-plane spin susceptibility of valley-Ising quarter metal plotted against increasing spin-orbit coupling strength

Figure 9. In-plane spin susceptibility of valley-Ising quarter metal is plotted against increasing SOC strength λ\lambda for ne,U=−2.9×1011 cm−2,18 meVn_e,U=-2.9\times10^{11}\ \mathrm{cm}^{-2},18\ \mathrm{meV}.

Energy landscape of quarter-metal in the presence of spin-orbit coupling

So far, we have identified only two quarter metal ground-state: valley-Ising and valley-XY phases. They are separated by a first order phase transition boundaries in the ne−Un_e-U space. In this section, we used the following Lagrange multipliers to further examine their energy landscape:

H^L(n)=λL[n⋅s^+(n×z^)⋅t^]=λL(sin⁡θ s^x+cos⁡θ s^z+sin⁡θ τ^x−cos⁡θ τ^z).(27)\begin{aligned} \hat{H}_{L}(\mathbf{n})=\lambda_L[\mathbf{n}\cdot\hat{\mathbf{s}}+(\mathbf{n}\times\hat{\mathbf{z}})\cdot\hat{\mathbf{t}}] \\ &=\lambda_L(\sin\theta\,\hat{s}_x+\cos\theta\,\hat{s}_z+\sin\theta\,\hat{\tau}_x-\cos\theta\,\hat{\tau}_z). \tag*{(27)} \end{aligned}

Here n=(sin⁡θ,cos⁡θ)\mathbf{n}=(\sin\theta,\cos\theta) represents the angle of spin-polarization. We opted not to delve into the two-dimensional energy landscape generated by independent valley and spin rotations as HKM=λσzτzτzH_{K M}=\lambda_{\sigma_z\tau_z}\tau_z favors the anti-alignment of the szs_z with τz\tau_z. Note ⟨σz⟩>0\langle\sigma_z\rangle>0.

In Figure 10, we plot the evolution of both the energy and the associated spin and valley order parameters for the quarter-metal phases as functions of hole density. At a high density of ne=−3×1011 cm−2n_e=-3\times10^{11}\ \mathrm{cm}^{-2}, the valley-Ising state, denoted by ∣θv=π,θs=0⟩|\theta_v=\pi,\theta_s=0\rangle, establishes itself as the global minimum, whereas the valley-XY state ∣θv=π/2,θs=π/2⟩|\theta_v=\pi/2,\theta_s=\pi/2\rangle appears unstable. However, when the density is lowered to ne=−2.8×1011 cm−2n_e=-2.8\times10^{11}\ \mathrm{cm}^{-2}, the valley-XY state stabilizes and becomes a local minimum. Reducing the density further to ne=−2.7×1011 cm−2n_e=-2.7\times10^{11}\ \mathrm{cm}^{-2} prompts a shift in the global minimum from ∣θv=π,θs=0⟩|\theta_v=\pi,\theta_s=0\rangle to ∣θv=π/2,θs=π/2⟩|\theta_v=\pi/2,\theta_s=\pi/2\rangle, signifying a first-order phase transition. At ne=−2.1×1011 cm−2n_e=-2.1\times10^{11}\ \mathrm{cm}^{-2}, ∣θv=π,θs=0⟩|\theta_v=\pi,\theta_s=0\rangle becomes unstable. Note that the Lagrange multiplier λL\lambda_L has a minimal impact on shifting the Ising-XY phase boundary.

Figure 10. The first column shows the total energy v.s. θ\theta while the second and third columns show the order parameters of the converged solution v.s. θ\theta. θ\theta is defined in Eq. (27) and and we set the λL=1 μeV\lambda_L=1\ \mu\mathrm{eV}, λ=30 μeV\lambda=30\ \mu\mathrm{eV}.

References

References

  1. [1]H. Zhou, T. Xie, A. Ghazaryan, T. Holder, J. R. Ehrets, E. M. Spanton, T. Taniguchi, K. Watanabe, E. Berg, M. Serbyn, and A. F. Young, Half and quarter metals in rhombohedral trilayer graphene, Nature 598, 429 (2021). arXiv:2104.00653.DOI
  2. [2]H. Zhou, L. Holleis, Y. Saito, L. Cohen, W. Huynh, C. L. Patterson, F. Yang, T. Taniguchi, K. Watanabe, and A. F. Young, Isospin magnetism and spin-polarized superconductivity in bernal bilayer graphene, Science 375, 774 (2022).DOI
  3. [3]H. Zhou, T. Xie, T. Taniguchi, K. Watanabe, and A. F. Young, Superconductivity in rhombohedral trilayer graphene, Nature 598, 434 (2021), arXiv:2106.07640.
  4. [4]A. M. Seiler, F. R. Geisenhof, F. Winterer, K. Watanabe, T. Taniguchi, T. Xu, F. Zhang, and R. T. Weitz, Quantum cascade of correlated phases in trigonally warped bilayer graphene, Nature 608, 298 (2022).
  5. [5]T. Han, Z. Lu, G. Scuri, J. Sung, J. Wang, T. Han, K. Watanabe, T. Taniguchi, L. Fu, H. Park, et al., Orbital multiferroicity in pentalayer rhombohedral graphene, arXiv preprint arXiv:2308.08837 (2023).DOI
  6. [6]T. Han, Z. Lu, G. Scuri, J. Sung, J. Wang, T. Han, K. Watanabe, T. Taniguchi, H. Park, and L. Ju, Correlated insulator and chern insulators in pentalayer rhombohedral stacked graphene (2023), arXiv:2305.03151 [cond-mat.mes-hall].DOI
  7. [7]L. Du, T. Han, A. Castellanos-Gomez, G.-B. Liu, Y. Yao, C. N. Lau, and Z. Sun, Engineering symmetry breaking in 2d layered materials, Nature Reviews Physics 3, 193 (2021).
  8. [8]S. C. da Barra, S. Aronson, Z. Zheng, K. Watanabe, T. Taniguchi, Q. Ma, P. J. Marillo-Herrero, and R. Ashoori, Cascade of isospin phase transitions in bernal-stacked bilayer graphene at zero magnetic field, Nature Physics 18, 771 (2022).
  9. [9]Y. Zhang, R. Polski, A. Thomson, É. Latagne-Hurtubise, C. Lewandowski, H. Zhou, K. Watanabe, T. Taniguchi, J. Alicea, and S. Nadj-Perge, Enhanced superconductivity in spin–orbit proximitized bilayer graphene, Nature 613, 268 (2023).DOI
  10. [10]X. Han, Q. Liu, Y. Wang, R. Niu, Z. Qu, Z. Wang, Z. Li, C. Han, K. Watanabe, T. Taniguchi, Z. Song, J. Mao, Z. V. Han, Z. Gan, and J. Lu, Chemical potential characterization of symmetry-breaking phases in a rhombohedral trilayer graphene, Nano Letters 23, 6875 (2023), pMID: 37466217, https://doi.org/10.1021/acs.nanolett.3c01262.
  11. [11]X. Han, Q. Liu, R. Niu, Z. Qu, Z. Wang, Z. Li, C. Han, K. Watanabe, T. Taniguchi, Z. Gan, and J. Lu, Spontaneous isospin polarization and quantum hall ferromagnetism in a rhombohedral trilayer graphene superlattice, Chinese Physics B (2023).DOI
  12. [12]Y. Lee, D. Tran, K. Myhro, J. Velasco, N. Gillgren, C. Lau, Y. Barlas, J. Poumirol, D. Smirnov, and F. Guinea, Competition between spontaneous symmetry breaking and single-particle gaps in trilayer graphene, Nature communications 5, 5656 (2014).DOI
  13. [13]P. A. Pantaleón, A. Jimeno-Pozo, H. Sainz-Cruz, T. Cea, V. T. Phong, and F. Guinea, Superconductivity and correlated phases in bilayer, trilayer graphene and related structures, arXiv preprint arXiv:2211.02880 (2022).DOI
  14. [14]L. Holleis, C. L. Patterson, Y. Zhang, H. M. Yoo, H. Zhou, T. Taniguchi, K. Watanabe, S. Nadj-Perge, and A. F. Young, Ising superconductivity and nematicity in bernal bilayer graphene with strong spin orbit coupling, arXiv preprint arXiv:2303.00742 (2023).DOI
  15. [15]Z. Lu, T. Han, Y. Yao, A. P. Reddy, J. Yang, J. Seo, K. Watanabe, T. Taniguchi, L. Fu, and L. Ju, Fractional quantum anomalous hall effect in a graphene moire superlattice, arXiv preprint arXiv:2309.17436 (2023).arxiv.org/abs/2309.17436
  16. [16]C. L. Kane and E. J. Mele, Quantum spin hall effect in graphene, Phys. Rev. Lett. 95, 226801 (2005).
  17. [17]A. Avsar, H. Ochoa, F. Guinea, B. Özilmaz, B. J. van Wees, and I. J. Vera-Marun, Colloquium: Spintronics in graphene and other two-dimensional materials, Rev. Mod. Phys. 92, 021003 (2020).
  18. [18]W. Han, R. K. Kawakami, M. Gmitra, and J. Fabian, Graphene spintronics, Nature nanotechnology 9, 794 (2014).DOI
  19. [19]Y. Zhumagulov, D. Kochan, and J. Fabian, Emergent correlated phases in rhombohedral trilayer graphene induced by proximity spin-orbit and exchange coupling (2023), arXiv:2305.14277 [cond-mat.str-el].DOI
  20. [20]C. Huang, T. M. Wolf, W. Qin, N. Wei, I. V. Blinov, and A. H. MacDonald, Spin and orbital metallic magnetism in rhombohedral trilayer graphene, Physical Review B 107, L121405 (2023).DOI
  21. [21]A. Jimeno-Pozo, H. Sainz-Cruz, T. Cea, P. A. Pantaleón, and F. Guinea, Superconductivity from electronic interactions and spin-orbit enhancement in bilayer and trilayer graphene, Physical Review B 107, L161106 (2023).
  22. [22]R. Katti, H. S. Arora, O.-P. Saira, K. Watanabe, T. Taniguchi, K. C. Schwab, M. L. Roukes, and S. Nadj-Perge, Hot carrier thermalization and josephson inductance thermometry in a graphene-based microwave circuit, Nano Letters 23, 4136 (2023).DOI
  23. [23]S. Konschuh, M. Gmitra, D. Kochan, and J. Fabian, Theory of spin-orbit coupling in bilayer graphene, Phys. Rev. B 85, 115423 (2012).DOI
  24. [24]M. Gmitra, S. Konschuh, C. Ertler, C. Ambrosch-Draxl, and J. Fabian, Band-structure topologies of graphene: Spin-orbit coupling effects from first principles, Physical Review B 80, 235431 (2009).DOI
  25. [25]H. Ochoa, A. C. Neto, V. Fal’Ko, and F. Guinea, Spin-orbit coupling assisted by flexural phonons in graphene, Physical Review B 86, 245411 (2012).DOI
  26. [26]J. Boettger and S. Trickey, First-principles calculation of the spin-orbit splitting in graphene, Physical Review B 75, 121402 (2007).DOI
  27. [27]K.-I. Imura, Y. Kuromoto, and K. Nomura, Anti-localization of graphene under the substrate electric field, Europhysics Letters 89, 17009 (2010).DOI
  28. [28]J. Sichau, M. Prada, T. J. Anlauf, B. Bosnjak, L. Tiemann, and R. H. Blick, Resonance microwave measurements of an intrinsic spin-orbit coupling gap in graphene: A possible indication of a topological state, Phys. Rev. Lett. 122, 046403 (2019).DOI
  29. [29]L. Banszerus, B. Frohn, T. Fabian, S. Somanchi, A. Epping, M. Müller, D. Neumaier, K. Watanabe, T. Taniguchi, F. Libisch, B. Beschoten, F. Hassler, and C. Stampfer, Observation of the spin-orbit gap in bilayer graphene by one-dimensional ballistic transport, Phys. Rev. Lett. 124, 177701 (2020).DOI
  30. [30]E. Morissette, J.-X. Lin, D. Sun, L. Zhang, S. Liu, D. Rhodes, K. Watanabe, T. Taniguchi, J. Hone, J. Hollanen, M. S. Scheurer, M. Lilly, A. Mounce, and J. I. A. Li, Electron spin resonance and collective excitations in magic-angle twisted bilayer graphene (2022), arXiv:2206.08354 [cond-mat.mes-hall].DOI
  31. [31]T. Arp, O. Sheekey, H. Zhou, C. Tschirhart, C. L. Patterson, H. Yoo, L. Holleis, E. Redekop, G. Babikyan, T. Xie, et al., Intervalley coherence and intrinsic spin-orbit coupling in rhombohedral trilayer graphene, arXiv preprint arXiv:2310.03781 (2023).DOI
  32. [32]L. Banszerus, S. Möller, C. Steiner, E. Icking, S. Trellenkamp, F. Lentz, K. Watanabe, T. Taniguchi, C. Volk, and C. Stampfer, Spin-valley coupling in single-electron bilayer graphene quantum dots, Nature Communications 12, 5250 (2021).
  33. [33]J. Island, X. Cui, C. Lewandowski, J. Y. Khoo, E. Spanton, H. Zhou, D. Rhodes, J. Hone, T. Taniguchi, K. Watanabe, et al., Spin–orbit-driven band inversion in bilayer graphene by the van der waals proximity effect, Nature 571, 85 (2019).DOI
  34. [34]T. Wang, M. Vila, M. P. Zaletel, and S. Chatterjee, Electrical control of magnetism in spin-orbit coupled graphene multilayers, arXiv preprint arXiv:2303.04855 (2023).DOI
  35. [35]J. M. Koh, J. Alicea, and É. Lantagne-Hurtubise, Correlated phases in spin-orbit-coupled rhombohedral trilayer graphene, arXiv preprint arXiv:2306.12486 (2023).DOI
  36. [36]M. Xie and S. Das Sarma, Flavor symmetry breaking in spin-orbit coupled bilayer graphene, Phys. Rev. B 107, L201119 (2023).DOI
  37. [37]E. McCann and M. Koshino, Spin-orbit coupling and broken spin degeneracy in multilayer graphene, Phys. Rev. B 81, 241409 (2010).DOI
  38. [38]C. Herring, Magnetism: Exchange interactions among itinerant electrons (Academic Press, 1966).DOI
  39. [39]M. Das and C. Huang, Unconventional metallic magnetism: Non-analyticity and sign-changing behavior of orbital magnetization in abc trilayer graphene (2023), arXiv:2308.01996 [cond-mat.mes-hall].DOI
  40. [40]C. Huang, N. Wei, and A. H. MacDonald, Current-driven magnetization reversal in orbital chern insulators, Phys. Rev. Lett. 126, 056801 (2021).DOI
  41. [41]H. Zhou, Y. Saito, L. Cohen, W. Huynh, C. L. Patterson, F. Yang, T. Taniguchi, K. Watanabe, and A. F. Young, Isospin magnetism and spin-triplet superconductivity in Bernal bilayer graphene, (2021), arXiv:2110.11317 [cond-mat].DOI
  42. [42]H. Min, G. Borghi, M. Polini, and A. H. MacDonald, Pseudospin magnetism in graphene, Phys. Rev. B 77, 041407 (2008).DOI
  43. [44]In preparation, Guopeng Chunli.
  44. [45]E. McCann and M. Koshino, The electronic properties of bilayer graphene, Reports on Progress in physics 76, 056503 (2013).
  45. [46]D. Kochan, S. Irmer, and J. Fabian, Model spin-orbit coupling hamiltonians for graphene systems, Phys. Rev. B 95, 165415 (2017).DOI
  46. [47]S. Konschuh, M. Gmitra, and J. Fabian, Tight-binding theory of the spin-orbit coupling in graphene, Phys. Rev. B 82, 245412 (2010).

Paper details

vlnc:W_9gark8y4
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

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