betaFeedback

WorksPhysics and Astronomy

Emergent fractals in dirty topological crystals

Abstract

Non-trivial geometry of electronic Bloch states gives rise to topological insulators which are robust against sufficiently weak randomness inevitably present in any quantum material. However, increasing disorder triggers a quantum phase transition into a featureless normal insulator. As the underlying quantum critical point is approached from the topological side, small scattered droplets of normal insulators start to develop in the system and their coherent nucleation causes ultimate condensation into a trivial insulator. Unless disorder is too strong, the normal insulator accommodates disjoint tiny topological puddles. Furthermore, in the close vicinity of such a transition the emergent islands of topological and trivial insulators display spatial fractal structures, a feature that is revealed only by local topological markers. Here we showcase this (possibly) generic phenomenon that should be apposite to dirty topological crystals of any symmetry class in any dimension from the Bott index and local Chern marker for a square-lattice-based disordered Chern insulator model.

Introduction. Thermal and quantum continuous phase transitions are fascinating phenomena in modern physics as in the vicinity of the associated critical points physical observables manifest power-law behavior in terms of universal scaling exponents [1, 2, 3, 4, 5]. Also, when such a critical point is approached from an ordered phase, it gradually loses spatial coherence and small incoherent islands of the disordered phase start to develop in the system. In the close proximity to the critical points, such flakes of ordered and disordered phases display fractal structures. While this phenomenon is well-appreciated in interacting systems, its jurisdiction across the disorder-controlled quantum phase transition (QPT) in noninteracting topological quantum materials remains unclear; raising the following question. How do dirty topological insulators (TIs) become trivial or normal insulator (NI) ?

To answer this question, we consider a specific lattice-regularized model for disordered TI and come to the following (possibly generic) conclusions that should be pertinent in dirty topological crystals from any symmetry class in any dimension. We compute a global topological invariant and the associated local topological marker in tandem. While the former one allows us to pin the critical disorder strength (WcW_{\rm c}) for a topological-normal insulator QPT, the latter one provides invaluable insights into the spatial profile of local topology. We show that as a dirty TI arrives at the shore of such a QPT, small isolated droplets of a NI start to form in the system. Near the TI-NI quantum critical point (QCP), the islands of TI and NI display spatial fractal structures. We identify a few isolated tiny pockets of TI inside a NI phase, unless the disorder strength W≫WcW \gg W_{\rm c}.

As a demonstrative example, we showcase these outcomes by numerically computing the disorder-averaged Bott index (BI) [6] and local Chern marker (LCM) [7, 8, 9] from a square lattice-based Chern insulator model. While the BI jumps from an integer value to zero across the TI-NI QPT, the LCM displays the rich structure mentioned above. The associated fractal and anomalous dimensions are reasonably close to the ones known for the two-dimensional (2D) uncorrelated Ising-like percolation theory [10, 11, 12, 13, 14]; slight deviations from which stem from the unavoidable spread of the LCM around the integer and trivial values in disordered systems. See Figure 1 and Figure 2. Therefore, our theoretical predictions unify emergent fractal phenomenon near critical points that so far has been studied in interacting, non-topological disordered, and statistical systems by introducing a new member in this family: dirty topological crystals.

Figure 1. (a) Disorder-averaged BI ⟨B⟩\langle B\rangle as a function of disorder strength (WW), obtained from exact diagonalization (ED) and kernel polynomial method (KPM), cross at W≈6.0W \approx6.0 for different LL (linear dimension of square lattice), where ⟨B⟩≈0.5\langle B\rangle\approx0.5, marking the critical disorder WcW_{\mathrm{c}} for the TI-NI insulator QPT. Inset: Data collapse for δL1/ν\delta L^{1/\nu} vs ⟨B⟩\langle B\rangle with ν=2.49\nu= 2.49, where δ=(W−Wc)/Wc\delta= (W-W_{\mathrm{c}})/W_{\mathrm{c}} and ν\nu is the correlation length exponent. (b) Spatial distribution of the LCM Cloc(r)C_{\mathrm{loc}}(r) on a L=40L=40 square lattice is shown after excluding four sites from each end of the system in both directions for a single disorder realization for each WW. The pair of numbers correspond to (W,⟨B⟩)(W,\langle B\rangle). (c) Fraction of such an interior area (ff) occupied by regions with Cloc(r)=1.0±0.2C_{\mathrm{loc}}(r)=1.0\pm0.2 (blue) and 0.0±0.20.0\pm0.2 (red). (d) Disorder-averaged LCM ⟨Cloc⟩\langle C_{\mathrm{loc}}\rangle in a square box of linear dimension ℓ=L/4\ell=L/4 at the center of the system as a function of WW, which for all choice of LL cross roughly at W=WcW=W_{\mathrm{c}}. (e) Data collapse for δL1/ν\delta L^{1/\nu} vs ⟨Cloc⟩\langle C_{\mathrm{loc}}\rangle with ν=2.51\nu=2.51 for ℓ=L/4\ell=L/4. (f) Variation of ν\nu with ℓ/L\ell/L. (g) Distributions of Cloc(r)C_{\mathrm{loc}}(r) for a single disorder realization with W=WcW=W_{\mathrm{c}} on a L=100L=100 square lattice is shown after excluding ten sites from each end of the system in both directions. Patches of (h) trivial [(i) topological] insulators with Cloc(r)=0.0±0.2C_{\mathrm{loc}}(r)=0.0\pm0.2 [1.0 ±\pm 0.2] are shown in different colors, exhibiting fractal structures (see Figure 2). LCM is always computed from ED. We average over 50 (200) disorder realizations far from (near) WcW_{\mathrm{c}}.

Figure 2. Computation of (a) hull fractal dimension dhd_h, (b) volume fractal dimension dvd_v, and (c) anomalous dimension η\eta and correlation length ξ\xi from the pair-connectivity G(r)G(r) at critical disorder W=Wc=6.0W = W_c = 6.0 from the islands of NIs in a L=120L = 120 system (in black). Panels (d)-(f) same as (a)-(c), respectively, for the islands of TIs. Dependance of (g) dhd_h, (h) dvd_v, (i) η\eta, and (j) ξ\xi on LL for a few W≲WcW \lesssim W_c (color coded) for the islands of NIs. Panels (k)-(n) are same as (g)-(j), respectively, for the islands of TIs. We average over 50 disorder realizations. For each each WW we quote (color coded) the values of dhd_h [(a) and (d)], dvd_v [(b) and (e)], and η\eta and ξ\xi [(c) and (f)]. Error bars capture bootstrap standard error [19].

Model. The minimal Hamiltonian for 2D Chern insulators of spinless or spin-polarized fermions is h=d(k)⋅τh = \mathbf{d}(\mathbf{k}) \cdot\boldsymbol{\tau}. The vector Pauli matrix τ=(τx,τy,τz)\boldsymbol{\tau} = (\tau_x,\tau_y,\tau_z) operates on orbitals with parity eigenvalues τ=±\tau= \pm. We choose d1(k)=tsin⁡(kxa)d_1(\mathbf{k}) = t\sin(k_xa), d2(k)=tsin⁡(kya)d_2(\mathbf{k}) = t\sin(k_ya), and d3(k)=m0−t0[cos⁡(kxa)+cos⁡(kya)]d_3(\mathbf{k}) = m_0 - t_0[\cos(k_xa) + \cos(k_ya)] [15]. Here, aa is the lattice constant, (t0)(t_0) is the hopping amplitude between the opposite (same) parity orbitals living on the nearest-neighbor sites of a square lattice, and m0m_0 denotes on-site staggered potential, yielding a uniform Dirac mass. This system supports TI (NI) for ∣m0/t0∣<2|m_0/t_0| < 2 (∣m0/t0∣>2|m_0/t_0| > 2). We set t=t0=1t = t_0 = 1, yielding a TI with the band inversion at the Γ\Gamma point, characterized by the first Chern number C=1C = 1 for the filled valence band, obtained by integrating its Berry curvature over the first Brillouin zone [16]. We disregard particle-hole asymmetry as it does not affect the topology of an insulating system.

Disorder. In the presence of disorder, the translational symmetry gets broken and the notion of a Bloch Hamiltonian becomes moot. We are then forced to compute the associated topological invariant by diagonalizing the corresponding real space tight-binding Hamiltonian, which on a square lattice for the chosen d(k)\mathbf{d}(\mathbf{k}) vector reads as

HTBSL=∑r[m0Ψr†τzΨr+∑j=x,y{t2iΨr†τjΨr+e^j+t02Ψr†τzΨr+e^j}]+∑rΨr†τ0V(r)Ψr.(1)H_{\rm TB}^{\rm SL} = \sum_r \left[m_0\Psi_r^\dagger\tau_z\Psi_r + \sum_{j=x,y}\left\{\frac{t}{2i}\Psi_r^\dagger\tau_j\Psi_{r+\hat{e}_j} + \frac{t_0}{2}\Psi_r^\dagger\tau_z\Psi_{r+\hat{e}_j}\right\}\right] + \sum_r \Psi_r^\dagger\tau_0 V(r)\Psi_r. \tag*{(1)}

Here, e^j=aj^\hat{e}_j=a\hat{j} with j^\hat{j} as the unit vector along j=x,yj=x,y, ΨrT=(cr,+,cr,−)\Psi_{\mathbf{r}}^{\mathrm{T}}=(c_{\mathbf{r},+},c_{\mathbf{r},-}), and cr,τc_{\mathbf{r},\tau} is the fermionic annihilation operator at position r\mathbf{r} with parity τ\tau. The last term encodes on-site potential disorder, the dominant source of elastic scattering in real materials. On each lattice site, we sample V(r)V(\mathbf{r}) uniformly and randomly from a box distribution [−W/2,W/2][-W/2,W/2], where WW is the disorder strength. We denote the corresponding matrix operator as W\mathbf{W} with its elements given by Wi,j=wiδijW_{i,j}=w_i\delta_{ij}, where ii and jj are site indices, wi∈[−W/2,W/2]w_i\in[-W/2,W/2], and δij\delta_{ij} is the Kronecker delta symbol. While W\mathbf{W} is traceless in sufficiently large systems for each disorder realization, in order to ensure this property that minimizes the shift in the Fermi level in moderate systems we replace W\mathbf{W} by W−δWIN\mathbf{W}-\delta_W I_N in numerical calculations. Here, δW=Tr⁡(W)/N\delta_W=\operatorname{Tr}(\mathbf{W})/N is a small constant, NN is the total number of sites in the system, and INI_N is an NN-dimensional identity matrix.

Topological invariants. In clean systems, the BI (BB), computed upon diagonalizing HTBSLH_{\mathrm{TB}}^{\mathrm{SL}} [Eq. (1)], enjoys a one-to-one correspondence with the first Chern number (CC); C≡BC\equiv B. In dirty systems, we compute the disorder-averaged Bott index ⟨B⟩\langle B\rangle to underpin the TI-NI QPT. We assume that the system is at half-filling and zero temperature, and define the projection operator P\mathcal{P} onto the valence band and its complementary operator Q\mathcal{Q} as

P=∑n=1N∣n⟩⟨n∣andQ=IN−P,(2)\mathcal{P}=\sum_{n=1}^{N}|n\rangle\langle n| \quad\text{and} \quad\mathcal{Q}=I_N-\mathcal{P}, \tag*{(2)}

respectively. Here ∣n⟩|n\rangle are the algebraically ordered eigenstates of HTBSLH_{\mathrm{TB}}^{\mathrm{SL}}. A unitary phase operator in the xx direction is defined as Ux=exp⁡(2πiX⊗τ0/Lx)U_x=\exp(2\pi i X\otimes\tau_0/L_x), where ⊗\otimes stands for the Kronecker product and the matrix elements of the position operator are Xi,j=xiδijX_{i,j}=x_i\delta_{ij} with xi∈[1,Lx]x_i\in[1,L_x]. The linear dimension of the system in the xx direction is LxL_x. Notice that Ux∣xiτ⟩=exp⁡(i2πxi/Lx)∣xiτ⟩U_x|x_i^\tau\rangle=\exp(i2\pi x_i/L_x)|x_i^\tau\rangle. Here, ∣xiτ⟩|x_i^\tau\rangle is the site-localized Wannier state at xix_i on an orbital with parity eigenvalue τ\tau. Similarly, we define a unitary phase operator UyU_y in the yy direction. Throughout we take Lx=Ly=LL_x=L_y=L (say). The projections of these phase operators onto the filled states are Vq=Q+PUqPV_q=\mathcal{Q}+\mathcal{P}U_q\mathcal{P} for q=xq=x and yy. Then the net phase around closed loop constructed with VxV_x and VyV_y yields [6]

B=12πIm⁡[Tr⁡[ln⁡(VxVyVx†Vy†)]].(3)B=\frac{1}{2\pi}\operatorname{Im}\left[\operatorname{Tr}\left[\ln\left(V_xV_yV_x^\dagger V_y^\dagger\right)\right]\right]. \tag*{(3)}

To respect the periodic nature of the phase operators, we always compute the BI on a square lattice with periodic boundary conditions.

Naturally, unveiling the spatial variation of topology is beyond the scope of BI, for which we compute the LCM at every site, encoded in the operator CLCM=4πIm⁡[PXQYP]C_{\mathrm{LCM}} = 4\pi\operatorname{Im}[PXQ_YP]. On a site at rr the LCM is [7, 8, 9]

Cloc(r)=∑τ=±⟨r,τ∣CL∣r,τ⟩.(4)C_{\mathrm{loc}}(r) = \sum_{\tau=\pm}\langle r,\tau|C_L|r,\tau\rangle. \tag*{(4)}

The LCM is insensitive to the boundary conditions, as XX and YY are not periodic. The LCM near the boundary of the system deviates considerably from the bulk quantized value even in clean systems [7]. Thus, we always exclude several sites near the boundaries of the square lattice while displaying and analyzing LCM.

Usually, we rely on exact diagonalization (ED) to compute P\mathcal{P}, which becomes time consuming in large systems when in addition we need to perform disorder averaging. This limitation can be circumvented by computing P\mathcal{P} using the kernel polynomial method (KPM) [17, 18], detailed in the Supplemental Material (SM) [19]. However, KPM does not count and order the eigenstates while constructing P\mathcal{P}. In KPM, the eigenvalue spectrum can only be truncated at a specified energy value. Therefore, we work under the assumption that the numbers of states below and above the Fermi energy (EF=0E_F = 0) remain equal even in the presence of disorder, which is typically the case when disorder configuration is set by the matrix W−(δW)INW - (\delta W)I_N. As the BI can only take integer values, a slight deviation from this assumption in KPM averages out and we obtain the same results for ⟨B⟩\langle B\rangle as from ED. However, no such constraint applies for the LCM, for which we always rely on ED.

Phase diagram. First, we compute ⟨B⟩\langle B\rangle using ED (for smaller LL) and KPM (for larger LL) as a function of WW, see Figure 1(a). In the weak (strong) disorder regime the system describes a TI (NI) with ⟨B⟩=1.0\langle B\rangle= 1.0 (0.0). The curves for WW vs ⟨B⟩\langle B\rangle for different LL cross around W≈6.0W \approx6.0, where ⟨B⟩≈0.5\langle B\rangle\approx0.5, defining the critical disorder WcW_c for the TI-NI QPT. With a single-parameter scaling ansatz ⟨B⟩=F(δL1/ν)\langle B\rangle= F(\delta L^{1/\nu}), where FF is an unknown universal function of its argument, δ=(W−Wc)/Wc\delta= (W - W_c)/W_c, and ν\nu is the correlation length exponent, determining the universality class of this QPT, we find best quality data collapse for ν=2.49\nu= 2.49, close to its current estimation for the quantum Hall plateau transition [20, 21, 22, 23, 24, 25], see inset of Figure 1(a).

Such a seemingly featureless QPT encodes fascinating rich structure, revealed by the LCM, see Figure 1(b). In the clean and weakly disordered systems Cloc(r)≈1C_{\mathrm{loc}}(r) \approx1 in the entire interior of the system, when ⟨B⟩=1\langle B\rangle= 1. With increasing disorder isolated droplets of NIs where Cloc(r)≈0C_{\mathrm{loc}}(r) \approx0 start to form in the system even inside the TI phase with ⟨B⟩=1\langle B\rangle= 1. As the system approaches the TI-NI QCP, the droplets of topological and normal insulators occupy almost equal area of the system when ⟨B⟩=0.5\langle B\rangle= 0.5, see Figure 1(c). By the same token, a NI with ⟨B⟩=0\langle B\rangle= 0 fosters a few isolated small flakes of TIs when W≳WcW \gtrsim W_c, which completely disappear only for W≫WcW \gg W_c.

The disorder-averaged LCM ⟨Cloc⟩\langle C_{\mathrm{loc}}\rangle, averaged over the sites of a square box of linear dimension ℓ=L/4\ell= L/4, placed near the center of the system, displays a similar behavior as ⟨B⟩\langle B\rangle, see Figure 1(d). With the scaling ansatz ⟨Cloc⟩=G(δL1/ν,ℓ/L)\langle C_{\mathrm{loc}}\rangle= G(\delta L^{1/\nu},\ell/L), where GG is another universal function of δL1/ν\delta L^{1/\nu} when ℓ/L≪1\ell/L \ll1, we obtain best quality data collapse for ν=2.51\nu= 2.51, see Figure 1(e). Figure 1(f) shows that only for 0.2≲ℓ/L≲0.40.2 \lesssim\ell/L \lesssim0.4, the values of ν\nu obtained from the data collapses of ⟨B⟩\langle B\rangle and ⟨Cloc⟩\langle C_{\mathrm{loc}}\rangle are sufficiently close.

Since the value of Cloc(r)C_{\mathrm{loc}}(r) is not strictly restricted to any integer value, especially in the presence of disorder, the islands of TI and NI in dirty topological crystals are identified with Cloc(r)=1.0±0.2C_{\mathrm{loc}}(r) = 1.0 \pm0.2 and Cloc(r)=0.0±0.2C_{\mathrm{loc}}(r) = 0.0 \pm0.2, respectively, in Figure 1(g). In the close proximity to the TI-NI QCP, individual droplets of NI and TI assume irregular fractal structure, as shown in Figure 1(h) and 1(i), respectively. Next, we quantitatively establish such emergent fractal structures near the TI-NI QPT.

Fractal analysis. Analysis of the fractal clusters begins with the identification and labeling of individual ones. For each fractal, we compute three geometric quantities; the radius of gyration (RR), perimeter (PP), and area (AA). Here we mention the key steps, while relegating details to the SM [19]. Since the fractals are embedded in two dimensions, the number of sites within it yields AA, the number of sites bordering the interior of the cluster determines PP, and RR is the root mean square distance of each site at rir_i from the geometric center of the fractal at rcr^c, given by R=[R12+R22⋯+RA2]1/2R = [R_1^2 + R_2^2 \cdots+ R_A^2]^{1/2}, where Ri=∣ri−rc∣/AR_i = |r_i - r^c|/\sqrt{A}. The universal and scale-independent (when L≫RL \gg R) hull (dhd_h) and volume (dvd_v) fractal dimensions are given by P∼RdhP \sim R^{d_h} and A∼RdvA \sim R^{d_v}, respectively. The pair connectivity G(r)G(r), measuring the probability of a given pair of sites separated by distance rr belonging to the same cluster, in general feature both power-law and exponential decays with the scaling form G(r)∼(r/ξ)−(d−2+η)exp⁡(−r/ξ)G(r) \sim(r/\xi)^{-(d-2+\eta)} \exp(-r/\xi). Here, d=2d = 2 is the embedding dimension, η\eta is the anomalous dimension, and ξ\xi is the correlation length, representing the average scale over which two points remain connected with a given

cluster. At the QCP, the distribution of the LCM ideally forms two large cohesive fractals spanning the entire system and thus ξ≈L\xi\approx L therein. At this point the exponential decay of G(r)G(r) gets suppressed, thereby featuring a power-law decay, determined by η\eta. These scaling forms capture how clusters fill the space in a self-similar manner, deviating from integer dimensions, near W=WcW = W_c.

Such an analysis is explicit shown on a L=120L = 120 system for W=6.0W = 6.0 and for the clusters of both TI and NI in Figure 2, where we also display the LL-dependence of dhd_h, dvd_v, η\eta, and ξ\xi for a few values of W≲WcW \lesssim W_c. When averaged over LL, the values of these quantities are reasonably insensitive to WW near WcW_c and are close, yet slightly far from the ones for the 2D uncorrelated Ising-like percolation theory for which dh=1.75d_h = 1.75, dv=1.9d_v = 1.9, and η=0.207\eta= 0.207 [10, 11, 12, 13, 14]. In particular for W=6.0W = 6.0, we obtain (after averaging the values over various LL) dh=1.67±0.12d_h = 1.67 \pm0.12 (1.67 ±\pm 0.09), dv=1.92±0.15d_v = 1.92 \pm0.15 (1.84 ±\pm 0.11), η=0.243±0.089\eta= 0.243 \pm0.089 (0.272 ±\pm 0.079), and ξ=46.0±17.3\xi= 46.0 \pm17.3 (35.8 ±\pm 11.6) for the fractal droplets of NI (TI). Such mismatches stem from the fact that the LCM in dirty systems spreads considerably around expected integer and trivial values, introducing an unavoidable ambiguity in properly identifying the sites belonging to specific fractal clusters of NI and TI, while Ising spins can only have two orientations. Otherwise, a correspondence between these two systems can be established as follows. Two possible distributions of the LCM, centered around 0 and 1, map onto two projections of spin. And with increasing disorder randomly-placed droplets of NI fluid with Cloc(r)≈0.0C_{\mathrm{loc}}(r) \approx0.0 percolates through the large cluster of TI with Cloc(r)≈1.0C_{\mathrm{loc}}(r) \approx1.0.

To establish the universal nature of emergent fractal droplets of local TIs and NIs, we repeat the entire analysis, detailed in the SM [19], for t=t0=1.0t = t_0 = 1.0 and m0=1.5m_0 = 1.5 for which the band gap is half of that with m0=1.0m_0 = 1.0. In this case Wc=6.70W_c = 6.70 around which we find dh=1.665±0.084d_h = 1.665 \pm0.084 (1.646 ±\pm 0.052), dv=1.882±0.068d_v = 1.882 \pm0.068 (1.850 ±\pm 0.069), η=0.277±0.028\eta= 0.277 \pm0.028 (0.241 ±\pm 0.086), and ξ=62.8±13.4\xi= 62.8 \pm13.4 (55.3 ±\pm 14.0) from the fractal islands of NIs (TIs), after averaging the values over various LL. Hence, emergent fractal islands of TIs and NIs share the same universal behavior irrespective of the band gap, and in two dimensions they are always similar to the uncorrelated 2D Ising-like percolation theory. Although, existing numerical results strongly suggest that these conclusions should hold sufficiently close to the band gap closing point at m0=2.0m_0 = 2.0, the corresponding numerical analysis must be performed in much larger systems to account for finite-size effects in the presence of a small band gap. The requisite computation time then grows very rapidly.

Discussions. We show that as a precursor of the disorder-induced TI-NI QPT, isolated small droplets of incipient NI start to nucleate inside the TI phase. In the close proximity to the associated QCP, droplets of TI and NI display fractal structures. The corresponding fractal and anomalous dimensions are reasonably close to the ones for the 2D Ising-like percolation theory. The predicted fractal structure in the local topology should be germane to dirty topological crystals of any Altland-Zirnbauer symmetry class and dimensionality [26, 27, 28, 29], as for all of them local topological markers can be computed [9], including topological superconductors [30]. The emergent fractal structures of the ground state, manifesting via local irregular droplets of TIs and NIs, near the disorder-driven QCP, separating two topologically distinct insulators, are distinct from the fractal-type structure of a few isolated critical wavefunctions residing near the Fermi energy, previously reported for Anderson [31, 32] and disordered topological [33, 34] models.

As the fractal and anomalous dimensions are sufficiently close for the clusters of TI and NI, the predicted fractal structure in real materials can be mapped from the local density of states (LDOS), measurable via scanning tunneling microscopy (STM) [14, 35], by solely focusing on the islands of TI. Such measurements should feature multi-gapped (gapless) LDOS spectrum near the Fermi energy inside (along boundaries of) the topological islands [see Figure 3], falling within the spatial resolution window of STM measurements (a few Å). The fact that a gapped (gapless) spectrum is observed in the interior (at the edges) of a cluster, as shown in Figure 3, justifies the working assumption that regions with LCM C=1.0±0.2C = 1.0 \pm0.2 and 0.0±0.20.0 \pm0.2 represent local topological and trivial insulators, respectively. Gapless edge modes residing along the boundaries between topologically distinct islands can also be responsible for the enhancement of electrical and thermal Hall and longitudinal conductivities near the topological QPTs, as reported in recent numerical works [36, 37]. The STM measurements can be performed on available quantum anomalous Hall or Chern insulators, realized thin films of Bi2_2Se3_3, Bi2_2Te3_3, and Sb2_2Te3_3 doped by Cr or V or Fe, for example [38, 39, 40], and quantum spin Hall insulators in CdTe-HgTe [41, 42] and InAs-SbTe [43] quantum wells by tuning the disorder strength therein. Recently the lattice model for the Chern insulator [15] has been engineered on optical lattices [44], where disorder strength can be tuned in a controlled fashion and LDOS can be measured using local radio frequency spectroscopy [45, 46, 47], to identify the proposed emergent fractal structures near the TI-NI QPT.

Figure 3. Distribution of the LCM inside a L=120L = 120 square lattice for a single disorder realization of strenght W=5.25W = 5.25 showing the topological (trivial) regions with Cloc(r)≈1.0C_{\mathrm{loc}}(r) \approx1.0 (0.0) in black (white). Local density of states ρloc(E)\rho_{\mathrm{loc}}(E) (b) at the edge and (c) in the interior of the topological region, computed over five green colored sites within the golden and red boxes, respectively, showing gapless and multi-gapped structures near the Fermi energy EF=0E_{\mathrm{F}} = 0. (d) Corresponding average density of state D(E)D(E) as a function of energy EE.

Acknowledgments. This work was supported by the NSF CAREER Grant No. DMR-2238679 of B.R.

Data availability. Numerical codes developed and data generated in this work are available via Zenodo [48].

References

References

  1. [1]L. P. Kadanoff, Statistical Physics: Statics, Dynamics and Renormalization (World Scientific Publishing Company, Singapore, 2000).
  2. [2]P. Chaikin and T. C. Lubensky, Principles of Condensed Matter Physics (Cambridge University Press, Cambridge, UK, 2000).
  3. [3]J. Zinn-Justin, Quantum Field Theory and Critical Phenomena (Oxford University Press, Oxford, UK, 2002).
  4. [4]S. Sachdev, Quantum Phase Transitions (Cambridge University Press, 2nd ed., Cambridge, UK, 2007).
  5. [5]I. F. Herbut, A modern approach to critical phenomena (Cambridge University Press, Cambridge, UK, 2007).DOI
  6. [6]T. A. Loring and M. B. Hastings, Disordered topological insulators via C*-algebras, Europhys. Lett. 92, 67004 (2010).DOI
  7. [7]R. Bianco and R. Resta, Mapping topological order in coordinate space, Phys. Rev. B 84, 241106(R) (2011).
  8. [8]A. Panigrahi, V. Juričić, and B. Roy, Projected topological branes, Commun. Phys. 5, 230 (2022).DOI
  9. [9]W. Chen, Universal topological marker, Phys. Rev. B 107, 045111 (2023).DOI
  10. [10]D. Stauffer and A. Aharony, Introduction To Percolation Theory, 2nd ed. (Taylor & Francis, London, 2018).
  11. [11]D. Stauffer, Scaling theory of percolation clusters, Phys. Rep. 54, 1 (1979).
  12. [12]T. Grossman and A. Aharony, Accessible external perimeters of percolation clusters, J. Phys. A: Math. Gen. 20, L1193 (1987).DOI
  13. [13]W. Janke and A. M. J. Schakel, Fractal structure of spin clusters and domain walls in the two-dimensional Ising model, Phys. Rev. E 71, 036703 (2005).pubmed.ncbi.nlm.nih.gov/15903629
  14. [14]C.-L. Song, E. J. M. F. Simmons, S. Liu, B. Phillabaum, K. A. Dahmen, E. W. Hudson, J. E. Hoffman, and E. W. Carlson, Critical nematic correlations throughout the superconducting doping range in Bi₂₋ₓPbₓSr₂₋ᵧLaᵧCuO₆₊ₓ, Nat. Commun. 14, 2622 (2023).DOI
  15. [15]X.-L. Qi, Y.-S. Wu, and S.-C. Zhang, Topological quantization of the spin Hall effect in two-dimensional paramagnetic semiconductors, Phys. Rev. B 74, 085308 (2006).DOI
  16. [16]D. J. Thouless, M. Kohmoto, M. P. Nightingale, and M. den Nijs, Quantized Hall Conductance in a Two-Dimensional Periodic Potential, Phys. Rev. Lett. 49, 405 (1982).DOI
  17. [17]A. Weiße, G. Wellein, A. Alvermann, and H. Fehske, The kernel polynomial method, Rev. Mod. Phys. 78, 275 (2006).
  18. [18]D. Varjas, M. Fruchart, A. R. Akhmerov, and P. M. Perez-Piskunow, Computation of topological phase diagram of disordered Pb₁₋ₓSnₓTe using the kernel polynomial method, Phys. Rev. Res. 2, 013229 (2020).DOI
  19. [19]See Supplemental Material at XXX-XXXX for the details of the kernel polynomial method, numerical analysis of the fractals, and complete numerical analysis for t = t₀ = 1 and m₀ = 1.5 in the presence of disorder.
  20. [20]J. T. Chalker and P. D. Coddington, Percolation, quantum tunnelling and the integer Hall effect, J. Phys. C 21, 2665 (1988).DOI
  21. [21]K. Slevin and T. Ohtsuki, Critical exponent for the quantum Hall transition, Phys. Rev. B 80, 041304(R) (2009).DOI
  22. [22]B. Sbierski, E. J. Dresselhaus, J. E. Moore, and I. A. Gruzberg, Criticality of Two-Dimensional Disordered Dirac Fermions in the Unitary Class and Universality of the Integer Quantum Hall Transition, Phys. Rev. Lett. 126, 076801 (2021).DOI
  23. [23]E. J. Dresselhaus, B. Sbierski, and I. A. Gruzberg, Scaling Collapses of Longitudinal Conductance near the Integer Quantum Hall Transition, Phys. Rev. Lett. 129, 026801 (2022).DOI
  24. [24]J. Mildner, M. D. Caio, G. Möller, N. R. Cooper, M. J. Bhaseen, Topological Phase Transitions in the Disordered Haldane Model, arXiv:2312.16689.arxiv.org/abs/2312.16689
  25. [25]C. Grindall, A. C. Tyner, A-K. Wu, T. L. Hughes, and J. H. Pixley, Separate surface and bulk topological Anderson localization transitions in disordered axion insulators, arXiv:2501.16413.arxiv.org/abs/2501.16413
  26. [26]A. Altland and M. R. Zirnbauer, Nonstandard symmetry classes in mesoscopic normal-superconducting hybrid structures, Phys. Rev. B 55, 1142 (1997).DOI
  27. [27]A. Kitaev, Periodic table for topological insulators and superconductors, AIP Conf. Proc. 1134, 22 (2009).
  28. [28]S. Ryu, A. P. Schnyder, A. Furusaki, and A. W. W. Ludwig, Topological insulators and superconductors: tenfold way and dimensional hierarchy, New J. Phys. 12, 065010 (2010).DOI
  29. [29]C.-K. Chiu, J. C. Y. Teo, A. P. Schnyder, and S. Ryu, Classification of topological quantum matter with symmetries, Rev. Mod. Phys. 88, 035005 (2016).DOI
  30. [30]S. Manna, S. K. Das, and B. Roy, Noncrystalline topological superconductors, Phys. Rev. B 109, 174512 (2024).DOI
  31. [31]D. A. Parshin and H. R. Schober, Distribution of Fractal Dimensions at the Anderson Transition, Phys. Rev. Lett. 83, 4590 (1999).DOI
  32. [32]For review, see, F. Evers and A. D. Mirlin, Anderson transitions, Rev. Mod. Phys. 80, 1355 (2008).
  33. [33]D. Xu, J. Qi, J. Liu, V. Sachdeva, T. X. C. Xie, and H. Jiang, Phase structure of the topological Anderson insulator, Phys. Rev. B 85, 195140 (2012).DOI
  34. [34]I. Sahlberg, A. Westström, K. Pöyhönen, and T. Ojanen, Topological phase transitions in glassy quantum matter, Phys. Rev. Res. 2, 013053 (2020).DOI
  35. [35]C. J. Chen, Introduction to Scanning Tunneling Microscopy (Oxford University Press, Oxford, UK, 2008).
  36. [36]S. K. Das and B. Roy, Quantized thermal and spin transports of dirty planar topological superconductors, Phys. Rev. B 109, 195403 (2024).DOI
  37. [37]S. K. Das and B. Roy, Quantized electrical, thermal, and spin transports of non-Hermitian clean and dirty two-dimensional topological insulators and superconductors, Phys. Rev. B 110, L241402 (2024).DOI
  38. [38]R. Yu, W. Zhang, H.-J. Zhang, S.-C. Zhang, X. Dai, and Z. Fang, Quantized anomalous Hall effect in magnetic topological insulators, Science 329, 61 (2010).
  39. [39]C.-Z. Chang, J. Zhang, X. Feng, J. Shen, Z. Zhang, M. Guo, K. Li, Y. Ou, P. Wei, L.-L. Wang, Z.-Q. Ji, Y. J. Feng, S. Ji, X. Chen, J. Jia, X. Dai, Z. Fang, S.-C. Zhang, K. He, Y. Wang, L. Lu, X-C. Ma, and Q-K. Xue, Experimental observation of the quantum anomalous Hall effect in a magnetic topological insulator, Science 340, 167 (2013).DOI
  40. [40]C.-Z. Chang, W. Zhao, D. Y. Kim, H. Zhang, B. A. Assaf, D. Heiman, S.-C. Zhang, C. Liu, M. H. W. Chan, and J. S. Moodera, High-precision realization of robust quantum anomalous Hall state in a hard ferromagnetic topological insulator, Nat. Mater. 14, 473 (2015).
  41. [41]B. A. Bernevig, T. L. Hughes, and S.-C. Zhang, Quantum spin Hall effect and topological phase transition in HgTe quantum wells, Science 314, 1757 (2006).
  42. [42]M. König, S. Wiedmann, C. Brüne, A. Roth, H. Buhmann, L. W. Molenkamp, X.-L. Qi, and S.-C. Zhang, Quantum spin Hall insulator state in HgTe quantum wells, Science 318, 766 (2007).pubmed.ncbi.nlm.nih.gov/17885096
  43. [43]I. Knez, R.-R. Du, and G. Sullivan, Evidence for helical edge modes in inverted InAs/GaSb quantum wells, Phys. Rev. Lett. 107, 136603 (2011).DOI
  44. [44]M-C. Liang, Y-D. Wei, L. Zhang, X-J. Wang, H. Zhang, W-W. Wang, W. Qi, X.-J. Liu, and X. Zhang, Realization of Qi-Wu-Zhang model in spin-orbit-coupled ultra-cold fermions, Phys. Rev. Res. 5, L012006 (2023).
  45. [45]R. Jördens, N. Strohmaier, K. Günter, H. Moritz, and T. Esslinger, A Mott insulator of fermionic atoms in an optical lattice, Nature (London) 455, 204 (2008).DOI
  46. [46]A. Schirotzek, C-H. Wu, A. Sommer, and M. W. Zwierlein, Observation of Fermi Polarons in a Tunable Fermi Liquid of Ultracold Atoms, Phys. Rev. Lett. 102, 230402 (2009).
  47. [47]D. Mitra, P. T. Brown, E. Guardado-Sanchez, S. S. Kondov, T. Devakul, D. A. Huse, P. Schauß, and W. S. Bakr, Quantum gas microscopy of an attractive Fermi-Hubbard system, Nat. Phys. 14, 173 (2018).DOI
  48. [48]D. J. Salib, Emergent fractals in dirty topological crystals, https://zenodo.org/records/17095623.

Paper details

vlnc:W_8cfafh29
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