betaFeedback

WorksPhysics and Astronomy

Hybrid Disclination Skin-topological Effects in Non-Hermitian Circuits

Abstract

The bulk-disclination correspondence (BDC) is a fundamental concept in Hermitian systems that has been widely applied to predict disclination states. Recently, disclination states have also been observed and experimentally verified in non-Hermitian systems with C6 lattice symmetry, where gain and loss are introduced to induce non-Hermiticity. In this Letter, we propose a non-Hermitian two-dimensional (2D) Su-Schrieffer-Heeger (SSH) disclination model with skin-topological (ST) disclination states, and calculate its biorthogonal Zak phase. Together with the real-space disclination index, we predict the emergence of disclination states in a C4-symmetric non-Hermitian lattice and the corresponding fractional charge. We also generalize the symmetry indicator within the biorthogonal framework to predict the anomalous filling near the disclination core. Experimentally, the model is implemented on a nonreciprocal circuit platform, where we analyze the impedance matrix characterized by complex eigenfrequencies and directly observe the ST disclination states. Our work further extends the bulk-disclination correspondence to the non-Hermitian realm.

Introduction

In classical conditions of topological phases, hermiticity guarantees a real energy spectrum and ensures the robustness of the defined topological invariants. However, in photonic systems [1, 2, 3, 4, 5, 6, 7, 8] and open quantum systems [9], the ubiquitous presence of gain and loss makes these conditions difficult to satisfy, thus necessitating the study of non-Hermitian topological phases [10, 11, 12, 13, 14, 15, 16, 17]. Compared with Hermitian systems, non-Hermitian lattices support richer spectra and unconventional boundary responses, including the non-Hermitian skin effect (NHSE) [18, 19, 20, 21, 22, 23, 24, 25, 26, 27] and hybrid skin-topological (ST) [28] boundary modes in which nonreciprocity and topology jointly dictate boundary localization. These features require the non-Bloch band description, formulated in the generalized Brillouin zone (GBZ) [18, 19, 20].

Conventional polarization theory treats polarization as a local dipole density; however, it becomes inadequate for crystalline insulators where Bloch states are extended, and periodic boundaries make the position operator ill-defined. In contrast, modern polarization theory [29, 30, 31] adopts a global, geometric viewpoint, defining macroscopic polarization by the integral of the Berry connection—equivalently, the U(1)U(1) Berry phase—accumulated over the occupied Bloch states across the entire Brillouin zone (BZ). Symmetry can quantize this geometric phase, so that a nontrivial polarization constitutes a bulk topological invariant. Equivalently, the nontrivial polarization indicates a quantized shift of the Wannier centers relative to the ionic background. Under bulk–boundary correspondence (BBC) [32, 33, 34, 35], this mismatch gives rise to anomalous boundary charges, which can be equivalently characterized as a filling anomaly [36, 37, 38, 39, 40, 41].

Geometric defects can couple to bulk topology and generate localized responses. Dislocations [42, 43, 44, 45, 46, 47, 48] and

disclinations [49, 50, 51, 52] break local translational symmetry and thus induce localized electronic states and topological effects of abnormal quantum transport. Moreover, the bulk–disclination correspondence (BDC) establishes a direct link between bulk topology and disclination states and fractional charge near the disclination core.[37, 42, 49, 51, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66] In non-Hermitian settings, gain–loss has been demonstrated to induce topological disclination states [67], while a complementary field-theoretic treatment connects disclinations to the NHSE [52]. However, there is a lack of a theoretical framework for predicting disclination states in non-Hermitian nonreciprocal systems.

In this paper, we theoretically establish and experimentally verify the non-Hermitian bulk–disclination correspondence in a non-Hermitian two-dimensional (2D) Su–Schrieffer–Heeger (SSH) disclination model. We find a topological invariant that incorporates both real-space and momentum-space information under non-Hermitian biorthogonal frameworks, which predicts the emergence of ST disclination states. We further explain these results by generalizing symmetry indicators to non-Hermitian biorthogonal settings. Finally, we verify the above theory on an electrical circuit platform [24, 68, 69, 70, 71, 72, 73, 74, 75] and propose an experimental method for measuring complex eigenfrequencies that provide a direct mapping to the eigenstates of a non-Hermitian Hamiltonian.

Models

We start from the non-Hermitian 2D SSH model, in which each unit cell consists of four sites with nonreciprocal couplings along both the xx and yy directions, and reciprocal intercell couplings of strength ww. In particular, we introduce unidirectional intracell couplings forming clockwise and counterclockwise loops denoted by vpv_{p} and vnv_{n}, respectively. Next, we introduce a disclination with a frank angle Ω=−π/2\Omega= -\pi/2 into the system (see the End Matter for details). There exists a pair of complex conjugate ST disclination modes (orange) together with a single purely real mode (red) for even ∣B∣|B|, Figure 1(b), while these disclination states are absent when ∣B∣|B| is odd Figure 1(c)}. Both types of disclination trap quantized fractional charge near disclination cores Fig. 1(d-e)`, despite the presence of nonreciprocity. These ST disclination states and their fractional charge are intimately linked to the bulk topological characteristics of the corresponding model without disclinations. We first study the case without disclination states.

Figure 1. (a) Complex-energy spectrum for lattice size l=12l = 12 and the mode distribution, with corner modes (red), edge modes (blue), disclination modes (orange), and bulk modes (gray). (b) Complex-energy spectrum for l=10l = 10. (c-d) Density of states (DOS) of the occupied bands computed in the biorthogonal framework for (c) s=(0,0)s = (0, 0), (d) s=(1/2,1/2)s = (1/2, 1/2). Red markers indicate the presence of bound states at the corresponding site positions. Parameters for all cases are vp=0.5v_p = 0.5, vn=0v_n = 0, w=1w = 1.

Figure 2(a) shows the band structure under PBC along both directions, where there is a finite real band gap. The model exhibits only line gaps [76] and therefore does not host any nontrivial point gap topology [76]. The complex energy spectra under xx-open boundary condition (OBC) and yy-OBC, and the corresponding mode distributions are shown in Figure 2(b-Figure 2c). We identify four ST corner modes, as well as twenty ST edge modes residing in the non-Hermitian line gap, whose number is determined by the finite lattice size. Their emergence is jointly governed by the intrinsic topology and the boundary localization direction set by NHSE [28] [77].

Figure 2. Properties of the 2D non-Hermitian SSH model without disclinations when vp=0.5v_p = 0.5, vn=0v_n = 0, w=1w = 1. (a) The real band structures under x-PBC, y-PBC. (b) Complex spectrum of the 12×1212 \times12 system under double OBC conditions. Four ST corner modes appear at the center, and twenty ST edge modes are located in the line gap (blue shading). (c) Spectrum of real part and the corresponding mode distributions.

In this regime, the line gap structure allows the Hamiltonian to be continuously deformed into an effectively Hermitian one without closing the gap. As a result, we can define a biorthogonal polarization vector P=(Px,Py)\boldsymbol{P} = (P_x, P_y) as a topological invariant associated with the geometric phase. In the presence of crystalline symmetries, the biorthogonal polarization is quantized. Mirror symmetries Mx,yM_{x,y} quantize PiP_i (i=x,yi = x,y) to Pi=−PiP_i = -P_i (mod ee), leading to a Z2⊕Z2\mathbb{Z}_2 \oplus\mathbb{Z}_2 classification. If only C4C_4 symmetry is preserved, as in our model, it enforces Px=PyP_x = P_y and reduces the invariant to a single Z2\mathbb{Z}_2. The biorthogonal Berry connection for the occupied bands is defined in terms of the left and right eigenstates as i⟨unL∣∂kiunR⟩i\langle u_n^L|\partial_{k_i}u_n^R\rangle. Accordingly, the biorthogonal polarization PxP_x can be formulated as

Px=e(2π)2∫−ππdky arg⁡∏n=1N⟨uL(kx,n,ky)∣uR(kx,n−1,ky)⟩⟨unL∣unR⟩⟨un−1L∣un−1R⟩.(1)P_x = \frac{e}{(2\pi)^2}\int_{-\pi}^{\pi} \mathrm{d}k_y\ \arg\prod_{n=1}^{N} \frac{\langle u^L(k_{x,n},k_y)|u^R(k_{x,n-1},k_y)\rangle}{\sqrt{\langle u_n^L|u_n^R\rangle\langle u_{n-1}^L|u_{n-1}^R\rangle}}. \tag*{(1)}

Here, NN denotes the number of discrete sampling points used to partition the Brillouin zone along the kxk_x directions, respectively. Because the system respects C4C_4 symmetry, we have Px=PyP_x = P_y. The results of biorthogonal polarization PxP_x are shown in Figure 3(a). In the regime 0<vp<10 < v_p < 1, the polarization is P=(1/2,1/2)\boldsymbol{P} = (1/2, 1/2), whereas for vp>2v_p > 2, it becomes P=(0,0)\boldsymbol{P} = (0, 0). As shown by the upper-left inset of Figure 3(a), when vp=0.5v_p = 0.5, a finite band gap opens at the high-symmetry points (HSPs). When vp=1v_p = 1 (gray dashed line), the two upper bands intersect, and the global real and imaginary [76] line gap no longer exists. The system remains gapless for 1<vn<21 < v_n < 2 (lower-right inset of Figure 3(a)). As vpv_p further increases, the system reaches the critical point for reopening the energy gap at vp=2v_p = 2 (lower-right inset of Figure 3(a)). At vp=2.5v_p = 2.5, the gap is fully opened. We further consider the more general case with vn≠0v_n \ne0, and obtain the phase diagram shown in Figure 3(b). When 0<vn<10 < v_n < 1 and 0<vp<10 < v_p < 1 (purple region), P=(1/2,1/2)\boldsymbol{P} = (1/2, 1/2). The yellow region corresponds to the gapless phase in which the polarization vector no longer exists. In the blue region, the energy gap reopens, P=(0,0)\boldsymbol{P} = (0, 0).

Figure 3. (a) Polarization along the xx-direction, PxP_x (vn=0v_n = 0). The system becomes gapless in the interval 1≤vp≤21 \leq v_p \leq2. The band structures are shown for vp=0.5v_p = 0.5 (upper left, with the gray dashed curve indicating vp=1v_p = 1), vp=1.5v_p = 1.5 (lower left), and vp=2.5v_p = 2.5 (right, with the gray dashed curve indicating vp=2v_p = 2). (b) Phase diagram in the (vp,vn)(v_p,v_n) plane when w=1w = 1.

Next, we investigate the case after introducing a disclination. According to homotopy theory, the equivalence class of the Burgers vector B\boldsymbol{B} forms a Z2\mathbb{Z}_2 group when Ω=±π/2\Omega= \pm\pi/2, and a Z2⊕Z2\mathbb{Z}_2 \oplus\mathbb{Z}_2 group when Ω=±π\Omega= \pm\pi. Since the parity of 2B2\boldsymbol{B} carries a geometric meaning, the disclination index ss is introduced as a real-space topological invariant, defined by 2s=(2B) mod 22s = (2\boldsymbol{B}) \bmod2 [66]. The conventional bulk-disclination correspondence theory applies only to Hermitian systems. Here, combined with the biorthogonal polarization vector P\boldsymbol{P}, we extend this correspondence to non-Hermitian systems and define a new topological invariant K=(s+P) mod 1\boldsymbol{K} = (s + \boldsymbol{P}) \bmod1, which can indicate the presence of ST disclination states.

In our model with C4C_4 rotational symmetry, the relation between the lattice size ll and ss is given by sx=sy=s_x = s_y = 14(l mod 4)\frac{1}{4}(l \bmod4). Hence, for finite lattices whose l mod 4=0l \bmod4 = 0, s=(0,0)s = (0,0), and for l mod 4=2l \bmod4 = 2, s=(1/2,1/2)s = (1/2,1/2). Figure 1(b) shows the complex energy spectrum of the disclination lattice with l=12l = 12, corresponding to K=(1/2,1/2)\mathbf{K} = (1/2,1/2). A pair of conjugate disclination states emerges within the loop of the ST edge state on the negative real axis, and an additional pure disclination state is embedded with the bulk states on the positive real axis. In contrast, Figure 1(c) shows the case of l=10l = 10, corresponding to K=(0,0)\mathbf{K} = (0,0), where no ST disclination state is observed.

The fractional charge is also essential, as it is equivalent to the anomalous filling of the Wannier center. From the biorthogonal polarization vector, we infer that the Wannier center resides at the Wyckoff position 1a, implying anomalous filling at the corners and at the disclination core. We confirm this by the biorthogonal density of states (DOS) within the occupied bands of the disclination model,

Qd=∑m∑n∣⟨ψnL∣m⟩⟨m∣ψnR⟩∣(mod1).(2)Q_d = \sum_m \sum_n \left|\langle\psi_n^L|m\rangle\langle m|\psi_n^R\rangle\right| \pmod1. \tag*{(2)}

Here, the summation over mm runs over the sites surrounding the disclination core, while nn traverses all occupied bands. The biorthogonal density of states (DOS) is shown in Figure 1(d-Figure 1e). For s=(0,0)s = (0,0), the result is Qd=0.75Q_d = 0.75. For the case with s=(1/2,1/2)s = (1/2,1/2), the fractional charge per unit cell near the disclination core, which is Qu=0.81Q_u = 0.81, slightly deviates from the theoretical value due to the limited lattice size [76]. In the thermodynamic limit, the total fractional charge near the disclination core is Qd=(3Qu) mod 1≈0.25Q_d = (3Q_u) \bmod1 \approx0.25.

Biorthogonal symmetry indicators

Here, we generalize the Hermitian rotation invariants to a biorthogonal non-Hermitian platform to explain the results. For a 2D insulator with CnC_n rotation symmetry [37], the Bloch Hamiltonian satisfies r^nH(k)r^n†=H(Rnk)\hat r_n H(\mathbf{k})\hat r_n^\dagger= H(R_n\mathbf{k}), where r^n\hat r_n is the rotation operator corresponding to a rotation by 2π/n2\pi/n. In a Hermitian system, at rotation-invariant HSP Π\Pi with RnΠ=ΠR_n\Pi= \Pi, one has [r^n,H(Π)]=0[\hat r_n,H(\Pi)] = 0. Whereas in the non-Hermitian case, the occupied subspace is spanned by biorthogonal left and right Bloch eigenstates satisfying ⟨ulL(Π)∣umR(Π)⟩=δlm\langle u_l^L(\Pi)|u_m^R(\Pi)\rangle= \delta_{lm}. Accordingly, the CnC_n-rotation eigenvalues at Π\Pi are obtained by diagonalizing the biorthogonal rotation matrix

RlmΠ=⟨ulL(Π(n))∣r^n∣umR(Π(n))⟩,l,m∈occ.,(3)\mathcal{R}_{lm}^{\Pi} = \langle u_l^L(\Pi^{(n)})|\hat r_n|u_m^R(\Pi^{(n)})\rangle,\qquad l,m \in\mathrm{occ.}, \tag*{(3)}

whose eigenvalues define Πp,nNH(n)\Pi_{p,nNH}^{(n)} (p=1,…,np = 1,\ldots,n). This enables the definition of symmetry indicators purely from band representations,

[Πp,nNH(n)]≡#Πp,nNH(n)−#Γp,nNH(n),(4)[\Pi_{p,nNH}^{(n)}] \equiv\#\Pi_{p,nNH}^{(n)} - \#\Gamma_{p,nNH}^{(n)}, \tag*{(4)}

where #Πp,nNH(n)\#\Pi_{p,nNH}^{(n)} counts the number of occupied bands at HSP Π\Pi with the eigenvalue Πp,nNH(n)\Pi_{p,nNH}^{(n)}, and Γ\Gamma is the center of the BZ. Symmetry-indicator relates rotation invariants in momentum space to the generalized Wannier-orbital content in real space, so that the fractional charge bound to a disclination can be written as an index depending solely on {[Πp,nNH(n)]}\{[\Pi_{p,nNH}^{(n)}]\}. For C4C_4 symmetry, the fractional charge bound to a disclination is given by [37]

Qd=Ω2π([X1,NH(2)]+32[M3,NH(4)]−12[M1,NH(4)])+T⋅P mod 1.(5)Q_d = \frac{\Omega}{2\pi}\left([X_{1,NH}^{(2)}] + \frac{3}{2}[M_{3,NH}^{(4)}] - \frac{1}{2}[M_{1,NH}^{(4)}]\right) + \mathbf{T}\cdot\mathbf{P} \bmod1. \tag*{(5)}

where Ω\Omega is the Frank angle, [X1,NH(2)][X^{(2)}_{1,NH}] and [M1,3,NH(4)][M^{(4)}_{1,3,NH}] are the biorthogonal rotation invariants. Here T⋅PT \cdot P captures the translational part of the holonomy, with TT the normal direction associated with this translation [37]. In our model, we obtain [X1,NH(2)]=[M3,NH(4)]=−1[X^{(2)}_{1,NH}] = [M^{(4)}_{3,NH}] = -1, and [M1,NH(4)]=1[M^{(4)}_{1,NH}] = 1. Specifically, for lmod4=2l \mathop{\mathrm{mod}} 4 = 2, (T⋅P)mod1=0(T \cdot P) \mathop{\mathrm{mod}} 1 = 0, leading to Qd=3/4Q_d = 3/4. For lmod4=0l \mathop{\mathrm{mod}} 4 = 0, (T⋅P)mod1=1/2(T \cdot P) \mathop{\mathrm{mod}} 1 = 1/2, and hence Qd=1/4Q_d = 1/4, which exactly coincide with the numerical results.

Experiments.— Hamiltonian can be mapped to the circuit Laplacian. In experiments, we fabricate an 8×88 \times8 circuit lattice with a Frank angle of −π/2-\pi/2, shown in Figure 4(a-Figure 4b). The ideal circuit Lagrangian matrix is given by J(ω)=iωC−(i/ω)WJ(\omega) = i\omega C - (i/\omega)W, where CC and WW are the matrices of capacitance and inverse inductance. When driven at the resonant frequency, the circuit Laplacian J(ω0)=iω0CtHJ(\omega_0) = i\omega_0 C_t H is linearly related to the Hamiltonian of the system, where Ct=C1/vpC_t = C_1/v_p. The circuit parameters are chosen as C1=1nFC_1 = 1\mathrm{nF}, C2=2nFC_2 = 2\mathrm{nF} and Lg=18μHL_g = 18\mathrm{\mu H}, in both simulations and experiments, leading to the on-site resonant frequency ω0=531kHz\omega_0 = 531\mathrm{kHz}. The nonreciprocal coupling is realized via voltage followers [Figure 4(c)] employing operational amplifiers (AD817) capable of driving sufficiently large capacitive loads.

Figure 4. Experimental results of an 8×88 \times8 disclination circuit with Frank angle −π/2-\pi/2. (a) Schematic of the nonreciprocal disclination circuit. Red components indicate voltage followers connected to a terminal capacitor C1C_1, and blue components denote coupling capacitors C2C_2. (b) The picture of the disclination circuit board. (c) Internal structure of the voltage follower. (d) Simulated, (e) experimental results of the spectrum of iJiJ at the resonance frequency 531kHz. (f) Case of σg=0\sigma_g = 0. A resonance peak appears at 566kHz, corresponding to a ST corner state located on the real energy axis. The right inset shows the distribution of the right eigenstate amplitude. (g) Case of σg=0.0002697S\sigma_g = 0.0002697\mathrm{S}. A conjugate ST corner mode with negative imaginary part appears at 578kHz. (h) Case of σg=0.0006255S\sigma_g = 0.0006255\mathrm{S}. A conjugate ST disclination mode with negative imaginary part appears at 493kHz. (i) Case of σg=0\sigma_g = 0. An ST disclination state with purely real eigenfrequency is observed at 1237kHz.

Next, we express the Hamiltonian in the biorthogonal basis as H=VRΛVL†H = V_R \Lambda V_L^\dagger, where Λ\Lambda denotes the diagonal matrix of eigenvalues, and VRV_R and VLV_L collect the right and left eigenvectors, respectively. The impedance matrix then takes the following form:

Zab=∑n1iωCt(VR)an1λn−λ(ω)(VL†)nb,(6)Z_{ab} = \sum_n \frac{1}{i\omega C_t}(V_R)_{an}\frac{1}{\lambda_n-\lambda(\omega)}(V_L^\dagger)_{nb}, \tag*{(6)}

where λ(ω)=−Cg/Ct+1/(ω2CtLg)\lambda(\omega) = -C_g/C_t + 1/(\omega^2 C_t L_g) and λn\lambda_n is the nn-th eigenvalue of the Hamiltonian. When the circuit is driven at an eigenfrequency ωn\omega_n corresponding to the nn-th mode of Hamiltonian, the eigenstates of the electrical circuit can be mapped exactly onto those of Hamiltonian via the following relations: ∣Za1(ωn)∣∝∣ψRna∣|Z_{a1}(\omega_n)| \propto|\psi_R^{na}|, ∣Z1a(ωn)∣∝∣ψLna∣|Z_{1a}(\omega_n)| \propto|\psi_L^{na}|, ∣Zaa(ωn)∣∝∣ψRnaψLna∣|Z_{aa}(\omega_n)| \propto|\psi_R^{na}\psi_L^{na}|. Here, we choose the sum of the magnitudes of each row of the impedance matrix to characterize the corresponding eigenstate ∣ψRna∣|\psi_R^{na}|.

However, in non-Hermitian circuits, it should be noted that the eigenfrequencies are generally complex. Instead of exciting the circuit at a complex frequency, we introduce a compensating conductance term σg(ωn)\sigma_g(\omega_n) grounded at each circuit node (see the End Matter for details),

σg(ωn)=−2ωnrωni(ωnr2+ωni2)ωnr2−ωni2Lg.(7)\sigma_g(\omega_n) = -\frac{2\omega_{nr}\omega_{ni}}{(\omega_{nr}^2+\omega_{ni}^2)\sqrt{\omega_{nr}^2-\omega_{ni}^2}L_g}. \tag*{(7)}

ωnr\omega_{nr} and ωni\omega_{ni} are real and imaginary parts of ωn\omega_n. By exciting the circuit at a real frequency ωeff=∣ωn2∣/Re⁡(ωn2)\omega_{\mathrm{eff}} = |\omega_n^2|/\sqrt{\operatorname{Re}(\omega_n^2)}, we can effectively access the eigenstate of the nn-th complex eigenmode ωn\omega_n. The detailed derivation and verification by ADS simulations are provided in [76].

Figure 4(d-Figure 4e) shows, respectively, the simulated and experimental spectrum of the circuit Laplacian at the resonant frequency 531kHz. By introducing the appropriate compensating conductance term σg(ωn)\sigma_g(\omega_n) into the diagonal of the circuit Laplacian, we obtain the corresponding impedance spectra and the mode distributions at each resonance frequency ωn\omega_n, as shown in Figure 4(f-Figure 4i). With σg=0\sigma_g = 0, we observe two real modes at 566kHz and 1237kHz that correspond, respectively, to the ST corner mode [Figure 4(i)] and the ST disclination mode [Figure 4(f)]. To access the complex eigenmode with a negative imaginary part, an additional conductance σg=0.0006255S\sigma_g = 0.0006255\mathrm{S} and σg=0.0002697S\sigma_g = 0.0002697\mathrm{S} is manually added to the diagonal of iJ(ω0)iJ(\omega_0). This procedure shifts the imaginary parts to zero, resulting in pronounced impedance resonance peaks at 578kHz and 493kHz, which correspond, respectively, to the ST corner mode [Figure 4(g)] and ST disclination mode [Figure 4(h)].

Conclusion.— We establish a non-Hermitian 2D SSH disclination model that simultaneously hosts ST corner, ST edge, and ST disclination modes, and extend the BDC to the non-Hermitian framework. In such cases, the reciprocal-space topological invariant should be replaced by the biorthogonal Zak phase. Meanwhile, we formulate the symmetry indicators within the non-Hermitian biorthogonal framework, enabling an exact prediction of the fractional charge near the disclination core. In the experiment, we establish a direct correspondence between the impedance matrix and the non-Hermitian biorthogonal eigenvector, and propose a method for resolving the complex eigenfrequencies in non-Hermitian electrical circuits. This allows us to experimentally observe the ST corner and ST disclination modes in the complex frequency plane.

Acknowledgements

This work acknowledges funding from the National Key Research and Development Program of China under Grant No. 2022YFA1404903, the National Natural Science Foundation of China under Grant No. 62288101, and Project Nos. E4BA270100, E4Z127010F, E4Z6270100, E53327020D of the Chinese Academy of Sciences.

The non-Hermitian 2D SSH model— We can write the Hamiltonian as follows:

H(kx,ky)=σ0⊗[(v+wcos⁡kx)τ1+(wsin⁡kx)τ2]+σ3⊗[(−iδ)τ2]+(v+wcos⁡ky)(σ1⊗τ0)+σ2⊗[(wsin⁡ky)τ0+iδτ3].(8)H(k_x,k_y)=\sigma_0\otimes\left[(v+w\cos k_x)\tau_1+(w\sin k_x)\tau_2\right] +\sigma_3\otimes\left[(-i\delta)\tau_2\right]+(v+w\cos k_y)(\sigma_1\otimes\tau_0) +\sigma_2\otimes\left[(w\sin k_y)\tau_0+i\delta\tau_3\right]. \tag*{(8)}

where v=(vp+vn)/2v=(v_p+v_n)/2 and δ=(vp−vn)/2\delta=(v_p-v_n)/2, σi\sigma_i and τi\tau_i are Pauli matrices with i=0,1,2,3i=0,1,2,3. By cutting and gluing the lower-right quadrant of a square lattice with 4×44\times4 unit cells, as illustrated in Figure 5, we introduce a disclination with a frank angle Ω=−π/2\Omega=-\pi/2 into the system.

Figure 5. (a) Schematic of an 8×88\times8 non-Hermitian 2D SSH lattice with a disclination of Frank angle −π/2-\pi/2. Clockwise intracell couplings (pink) are vpv_p and counterclockwise couplings (blue) are vnv_n. The lower-right quarter is cut out, and the exposed edges are glued by the intercell coupling ww.

Effective Protocol for Probing Complex eigenmodes— We establish the correspondence between non-Hermitian systems and nonreciprocal circuit parameters. We also present the treatment of imaginary frequencies, deriving an effective real frequency ωeff\omega_{\mathrm{eff}} and compensating conductance σg\sigma_g, mentioned in the main text. Firstly, we begin with the basic formalism of the circuit Laplacian,

J(ω)=iωC+Wiω=iω(−CgI+CtH)−IiωLg,(9)J(\omega)=i\omega C+\frac{W}{i\omega}=i\omega(-C_gI+C_tH)-\frac{I}{i\omega L_g}, \tag*{(9)}

which is equal to,

J(ω)=iωCt(H−λ(ω)I),λ(ω)=CgCt−1ω2LgCt.(10)J(\omega)=i\omega C_t(H-\lambda(\omega)I),\qquad\lambda(\omega)=\frac{C_g}{C_t}-\frac{1}{\omega^2L_gC_t}. \tag*{(10)}

Since the Hamiltonian is non-Hermitian, we can expand it in terms of left and right eigenvectors as H=VRΛVL†H=V_R\Lambda V_L^\dagger, then we obtain:

J(ω)=iωCtVR(Λ−λ(ω)I)VL†.(11)J(\omega)=i\omega C_tV_R(\Lambda-\lambda(\omega)I)V_L^\dagger. \tag*{(11)}

By inverting the circuit Laplacian above, the impedance matrix is obtained as follows,

Zab=∑n1iωCt(VR)an1λn−λ(ω)(VL†)nb.(12)Z_{ab}=\sum_n\frac{1}{i\omega C_t}(V_R)_{an}\frac{1}{\lambda_n-\lambda(\omega)}(V_L^\dagger)_{nb}. \tag*{(12)}

where Λmn=λnδmn\Lambda_{mn}=\lambda_n\delta_{mn}, and we also use the biorthogonal relation VL†VR=IV_L^\dagger V_R=I. Here VR=[∣unR⟩]V_R=[|u_n^R\rangle] and VL=[∣unL⟩]V_L=[|u_n^L\rangle] collect the left and right eigenvectors with H∣unR⟩=λn∣unR⟩H|u_n^R\rangle=\lambda_n|u_n^R\rangle and ⟨unL∣H=⟨unL∣λn\langle u_n^L|H=\langle u_n^L|\lambda_n. When λ(ωn)=λn\lambda(\omega_n)=\lambda_n, the denominator approaches zero, resulting in a pronounced impedance peak. Through this formulation, the eigenstates of the electrical circuit can be mapped exactly onto those of the Hamiltonian via the relations: ∣Za1(ωn)∣∝∣ψRna∣|Z_{a1}(\omega_n)|\propto|\psi_R^{na}|, ∣Z1a(ωn)∣∝∣ψLna∣|Z_{1a}(\omega_n)|\propto|\psi_L^{na}|, ∣Zaa(ωn)∣∝∣ψRnaψLna∣|Z_{aa}(\omega_n)|\propto|\psi_R^{na}\psi_L^{na}|.

Considering that the eigenfrequency generally takes complex values ωn=ωnr+iωni\omega_n=\omega_{nr}+i\omega_{ni}, it is impossible to excite the system at purely real frequencies to get the resonance peak associated with the corresponding mode. To overcome this limitation, we introduce a compensating conductance σg(ωn)\sigma_g(\omega_n). By substituting ωn=ωnr+iωni\omega_n=\omega_{nr}+i\omega_{ni} into Eq. (10), we can rewrite λ(ωn)\lambda(\omega_n) as:

λ(ωn)=CgCt−ωnr2−ωni2(ωnr2+ωni2)2LgCt+2iωnrωni(ωnr2+ωni2)2LgCt.(13)\lambda(\omega_n)=\frac{C_g}{C_t}-\frac{\omega_{nr}^2-\omega_{ni}^2}{(\omega_{nr}^2+\omega_{ni}^2)^2L_gC_t}+\frac{2i\omega_{nr}\omega_{ni}}{(\omega_{nr}^2+\omega_{ni}^2)^2L_gC_t}. \tag*{(13)}

We then consider the circuit Laplacian with the compensating conductance σ=−σgI\sigma=-\sigma_gI:

J(ωn′)=iωn′Ct(H−λ′(ωn′)I),(14)J(\omega_n')=i\omega_n'C_t\left(H-\lambda'(\omega_n')I\right), \tag*{(14)}

where λ′(ωn′)=Cg/Ct−1/(ωn′2LgCt)−iσg/(ωn′Ct)\lambda'(\omega_n')=C_g/C_t-1/(\omega_n'^2L_gC_t)-i\sigma_g/(\omega_n'C_t). Comparing the real and imaginary parts of Eq. (13) with (14), we obtain:

ωeff=ωn′=∣ωn2∣Re⁡(ωn2),(15)\omega_{\mathrm{eff}}=\omega_n'=\frac{|\omega_n^2|}{\sqrt{\operatorname{Re}(\omega_n^2)}}, \tag*{(15)}
σg(ωn)=−2ωnrωni(ωnr2+ωni2)ωnr2−ωni2Lg.(16)\sigma_g(\omega_n)=-\frac{2\omega_{nr}\omega_{ni}}{(\omega_{nr}^2+\omega_{ni}^2)\sqrt{\omega_{nr}^2-\omega_{ni}^2}L_g}. \tag*{(16)}

These two equations show that, after introducing compensating conductance, the required excitation frequency ωn′\omega_n' is no longer simply given by the real part of the original eigenfrequency. Therefore, to observe the eigenmode associated with the nn-th eigenfrequency, we could first compute the required compensating conductance theoretically, and then use Z′(ωn)=(Z−1(ωn)+σg(ωn))−1Z'(\omega_n)=(Z^{-1}(\omega_n)+\sigma_g(\omega_n))^{-1} to process the measured data at the corresponding frequency. Finally, we can observe the effective impedance peak corresponding to the target eigenstate. The corresponding ADS simulation results are provided in [76].

References

  1. [1]K. G. Makris, R. El-Ganainy, D. N. Christodoulides, and Z. H. Musslimani, Phys. Rev. Lett. 100, 103904 (2008).
  2. [2]R. El-Ganainy, K. G. Makris, M. Khajavikhan, Z. H. Musslimani, S. Rotter, and D. N. Christodoulides, Nat. Phys. 14, 11 (2018).
  3. [3]A. Regensburger, C. Bersch, M.-A. Miri, G. Onishchukov, D. N. Christodoulides, and U. Peschel, Nature 488, 167 (2012).
  4. [4]H. Jing, S. K. Özdemir, X.-Y. Liu, J. Zhang, L. Yang, and F. Nori, Phys. Rev. Lett. 113, 053604 (2014).
  5. [5]M. Hodaei, M.-A. Miri, M. Heinrich, D. N. Christodoulides, and M. Khajavikhan, Science 346, 975 (2014).
  6. [6]B. Peng, S. K. Özdemir, F. Lei, F. Monifi, M. Gianfreda, G. L. Long, S. Fan, F. Nori, C. M. Bender, and L. Yang, Nat. Phys. 10, 394 (2014).
  7. [7]Y. Wu, W. Liu, J. Geng, X. Song, X. Ye, A. Ji, X. Rong, and J. Du, Science 364, 878 (2019).
  8. [8]R. El-Ganainy, K. G. Makris, M. Khajavikhan, Z. H. Musslimani, S. Rotter, and D. N. Christodoulides, Nat. Phys. 14, 11 (2018).
  9. [9]F. Minganti, A. Miranowicz, R. W. Chhajlany, and F. Nori, Phys. Rev. A 100, 062131 (2019).
  10. [10]S. Yao, F. Song, and Z. Wang, Phys. Rev. Lett. 121, 136802 (2018).
  11. [11]T. Liu, Y.-R. Zhang, Q. Ai, Z. Gong, K. Kawabata, M. Ueda, and F. Nori, Phys. Rev. Lett. 122, 076801 (2019).
  12. [12]N. Okuma, K. Kawabata, K. Shiozaki, and M. Sato, Phys. Rev. Lett. 124, 086801 (2020).
  13. [13]K. Zhang, Z. Yang, and C. Fang, Phys. Rev. Lett. 125, 126402 (2020).
  14. [14]D. S. Borgnia, A. J. Kruchkov, and R.-J. Slager, Phys. Rev. Lett. 124, 056802 (2020).
  15. [15]K. Kawabata, M. Sato, and K. Shiozaki, Phys. Rev. B 102, 205118 (2020).
  16. [16]R. Okugawa, R. Takahashi, and K. Yokomizo, Phys. Rev. B 102, 241202 (2020).
  17. [17]C. Yuce and H. Ramezani, Phys. Rev. A 100, 032102 (2019).
  18. [18]S. Yao and Z. Wang, Phys. Rev. Lett. 121, 086803 (2018).
  19. [19]F. K. Kunst, E. Edvarsson, J. C. Budich, and E. J. Bergholtz, Phys. Rev. Lett. 121, 026808 (2018).
  20. [20]K. Yokomizo and S. Murakami, Phys. Rev. Lett. 123, 066404 (2019).
  21. [21]F. Song, S. Yao, and Z. Wang, Phys. Rev. Lett. 123, 170401 (2019).
  22. [22]N. Okuma, K. Kawabata, K. Shiozaki, and M. Sato, Phys. Rev. Lett. 124, 086801 (2020).
  23. [23]D. S. Borgnia, A. J. Kruchkov, and R.-J. Slager, Phys. Rev. Lett. 124, 056802 (2020).
  24. [24]T. Helbig, T. Hofmann, M. Imhof, M. Abdelghany, T. Kiessling, L. W. Molenkamp, C. H. Lee, A. Szameit, M. Greiter, and R. Thomale, Nat. Phys. 16, 747 (2020).
  25. [25]L. Xiao, T. Deng, K. Wang, G. Zhu, Z. Wang, W. Yi, and P. Xue, Nat. Phys. 16, 761 (2020).
  26. [26]S. Weidemann, M. Kremer, T. Helbig, T. Hofmann, A. Stegmaier, M. Greiter, R. Thomale, and A. Szameit, Science 368, 311 (2020).
  27. [27]T. Hofmann, T. Helbig, F. Schindler, N. Salgo, M. Brzezinska, M. Greiter, T. Kiessling, D. Wolf, A. Vollhardt, A. Kabaši, C. H. Lee, A. Bilušič, R. Thomale, and T. Neupert, Phys. Rev. Research 2, 023265 (2020).
  28. [28]C. H. Lee, L. Li, and J. Gong, Phys. Rev. Lett. 123, 016805 (2019).
  29. [29]D. Vanderbilt and R. D. King-Smith, Phys. Rev. B 48, 4442 (1993).
  30. [30]X. Gonze, P. Ghosez, and R. W. Godby, Phys. Rev. Lett. 74, 4035 (1995).
  31. [31]R. D. King-Smith and D. Vanderbilt, Phys. Rev. B 47, 1651 (1993).
  32. [32]M. Z. Hasan and C. L. Kane, Rev. Mod. Phys. 82, 3045 (2010).
  33. [33]D. Xiao, M.-C. Chang, and Q. Niu, Rev. Mod. Phys. 82, 1959 (2010).
  34. [34]X.-L. Qi and S.-C. Zhang, Rev. Mod. Phys. 83, 1057 (2011).
  35. [35]A. Bansil, H. Lin, and T. Das, Rev. Mod. Phys. 88, 021004 (2016).
  36. [36]W. A. Benalcazar, T. Li, and T. L. Hughes, Phys. Rev. B 99, 245151 (2019).
  37. [37]T. Li, P. Zhu, W. A. Benalcazar, and T. L. Hughes, Phys. Rev. B 101, 115115 (2020).
  38. [38]C. W. Peterson, T. Li, W. A. Benalcazar, T. L. Hughes, and G. Bahl, Science 368, 1114 (2020).
  39. [39]M. Geier, L. Trifunovic, M. Hoskam, and P. W. Brouwer, Phys. Rev. B 97, 205135 (2018).
  40. [40]Y. Xu, L. Elcoro, Z.-D. Song, M. G. Vergniory, C. Felser, S. P. Parkin, J. L. Regnault, Nicolas Regnault, and B. A. Bernevig, Phys. Rev. B 109, 165139 (2024).
  41. [41]H. Wada, T. Zhang, and S. Murakami, Phys. Rev. B 111, 155305 (2025).
  42. [42]Y. Ran, Y. Zhang, and A. Vishwanath, Nat. Phys. 5, 298 (2009).
  43. [43]G. van Miert and C. Ortix, Phys. Rev. B 97, 201111 (2018).
  44. [44]F.-F. Li, Y.-X. Wang, Z. Xiong, Q. Lou, P. Chen, R.-X. Wu, Y. Poo, J.-H. Jiang, and S. John, Nat. Commun. 9, 2462 (2018).
  45. [45]A. K. Nayak, J. Reiner, R. Queiroz, R. Hu, C. Shekhar, B. Yan, C. Felser, N. Avraham, and H. Beidenkopf, Sci. Adv. 5, eaax6996 (2019).
  46. [46]B. Roy and V. Juricić, Phys. Rev. Research 3, 033107 (2021).
  47. [47]F. Schindler and A. Prem, Phys. Rev. B 104, L161106 (2021).
  48. [48]B. A. Bhargava, I. C. Fulga, J. van den Brink, and A. G. Mogadhdam, Phys. Rev. B 104, L241402 (2021).
  49. [49]C. W. Peterson, T. Li, W. Jiang, T. L. Hughes, and G. Bahl, Nature 589, 376 (2021).
  50. [50]Q. Wang, H. Xue, B. Zhang, and Y. D. Chong, Phys. Rev. Lett. 124, 243602 (2020).
  51. [51]Y. Liu, S. Leung, F.-F. Li, Z.-K. Lin, X. Tao, Y. Poo, and J.-H. Jiang, Nature 589, 381 (2021).
  52. [52]X.-Q. Sun, P. Zhu, and T. L. Hughes, Phys. Rev. Lett. 127, 066401 (2021).
  53. [53]J. C. Y. Teo and T. L. Hughes, Phys. Rev. Lett. 111, 047006 (2013).
  54. [54]A. Rüegg and C. Lin, Phys. Rev. Lett. 110, 046401 (2013).
  55. [55]W. A. Benalcazar, J. C. Y. Teo, and T. L. Hughes, Phys. Rev. B 89, 224503 (2014).
  56. [56]F.-F. Li, H.-X. Wang, Z. Xiong, Q. Lou, P. Chen, R.-X. Wu, Y. Poo, J.-H. Jiang, and S. John, Nat. Commun. 9, 2462 (2018).
  57. [57]J. C. Y. Teo and C. L. Kane, Phys. Rev. B 82, 115120 (2010).
  58. [58]Q. Wang, H. Xue, B. Zhang, and Y. D. Chong, Phys. Rev. Lett. 124, 243602 (2020).
  59. [59]Q. Wang, Y. Ge, H. X. Sun, H. Xue, D. Jia, Y. J. Guan, S. Q. Yuan, B. Zhang, and Y. D. Chong, Nat. Commun. 12, 3654 (2021).
  60. [60]H. Xue, D. Jia, Y. Ge, Y. J. Guan, Q. Wang, S. Q. Yuan, H. X. Sun, Y. D. Chong, and B. Zhang, Phys. Rev. Lett. 127, 214301 (2021).
  61. [61]M. Geier, I. C. Fulga, and A. Lau, SciPost Phys. 10, 092 (2021).
  62. [62]Y. Deng, W. A. Benalcazar, Z.-G. Chen, M. Oudich, G. Ma, and Y. Jing, Phys. Rev. Lett. 128, 174301 (2022).
  63. [63]B.-Y. Xie, O. You, and S. Zhang, Phys. Rev. A 106, L021502 (2022).
  64. [64]Z.-K. Lin, Q. Wang, Y. Liu, H. Xue, B. Zhang, Y. D. Chong, and J.-H. Jiang, Nat. Rev. Phys. 5, 483 (2023).
  65. [65]M.-S. Hwang, H.-R. Kim, J. Kim, B.-J. Yang, Y. Kivshar, and H.-G. Park, Nat. Photonics 18, 286 (2024).
  66. [66]Q. He, J. Sun, H.-Y. Deng, K. Wakabayashi, and F. Liu, Front. Phys. 11, 1213158 (2023).
  67. [67]R. Banerjee, S. Mandal, Y. Y. Terh, S. Lin, G.-G. Liu, B. Zhang, and Y. D. Chong, Phys. Rev. Lett. 133, 233804 (2024).
  68. [68]S. Imhof, C. Berger, F. Bayer, J. Brehm, L. W. Molenkamp, T. Kiessling, F. Schindler, C. H. Lee, M. Greiter, T. Neupert, and R. Thomale, Nat. Phys. 14, 925 (2018).
  69. [69]C. H. Lee, S. Imhof, C. Berger, F. Bayer, J. Brehm, L. W. Molenkamp, T. Kiessling, and R. Thomale, Commun. Phys. 1, 39 (2018).
  70. [70]S. Liu, S. Ma, Q. Zhang, L. Zhang, C. Yang, O. You, W. Gao, Y. Xiang, T. J. Cui, and S. Zhang, Light Sci. Appl. 9 (2020).
  71. [71]Y. Wang, H. M. Price, B. Zhang, and Y. D. Chong, Nat. Commun. 11, 2356 (2020).
  72. [72]S. Liu, R. Shao, S. Ma, L. Zhang, O. You, H. Wu, Y. J. Xiang, T. J. Cui, and S. Zhang, Research 2021, 5608038 (2021).
  73. [73]C. Shang, S. Liu, R. Shao, P. Han, X. Zang, X. Zhang, K. N. Salama, W. Gao, C. H. Lee, R. Thomale, A. Manchon, S. Zhang, T. J. Cui, and U. Schwingenschlögl, Adv. Sci. 9, 2202922 (2022).
  74. [74]S. Qiu, J. Hu, Y. Yang, C. Shang, S. Liu, and T. J. Cui, Natl. Sci. Rev. 12, nwaf137 (2025).
  75. [75]Supplementary Material contains information on the analytical determination of the critical point for Re(E), the complex band spectrum and topological characterization, the fractional-charge in the enlarged system, the detailed procedure for processing the experimental data, a comparison between theory and experiment, and details of the circuit board and experimental procedure.
  76. [76]L. Li, Y. Wei, G. Wu, Y. Ruan, S. Chen, C. H. Lee, and Z. Ni, Phys. Rev. B 111, 075132 (2025).

Paper details

vlnc:W_t7dvcbnn
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