betaFeedback

WorksPhysics and Astronomy

Hopf Semimetals

Abstract

We construct two-band topological semimetals in four dimensions using the unstable homotopy of maps from the three-torus T3T^3 (Brillouin zone of a 3D crystal) to the two-sphere S2S^2. Dubbed ``Hopf semimetals'', these gapless phases generically host nodal lines, with a surface enclosing such a nodal line in the four-dimensional Brillouin zone carrying a Hopf flux. These semimetals show a unique class of surface states: while some three-dimensional surfaces host gapless Fermi-arc states and drumhead states, other surfaces have gapless Fermi surfaces. Gapless two-dimensional corner states are also present at the intersection of three-dimensional surfaces.

Introduction

The understanding and classification of gapped topological phases [1, 2, 3, 4, 5, 6, 7, 8] of non-interacting fermions has not only provided deeper insights, but also, stimulated wider generalizations [9, 10] and the search for topological materials [11]. The current understanding of these gapped phases is built on the symmetry classification of the fermionic systems that arise from the presence or absence of intrinsic symmetries such as time reversal, charge conjugation and sublattice symmetries [8, 12, 13, 14]. In a crystalline system in dd-dimensions, the ground state of a gapped fermionic system is obtained by the state of occupied valance bands in the first Brillouin zone (BZ), the dd-torus TdT^d. Interestingly, the occupied states at any point in the BZ can be viewed as a point in one of the ten symmetric spaces SS, the specific one being determined by the intrinsic symmetry. Topologically distinct gapped ground states are identified with the homotopy classes of maps from TdT^d to SS, resulting in the periodic table of strong topological phases [5].

Apart from these symmetry protected topological phases, a class of gapless phases have elicited attention, beginning with graphene [15], and more recently, Weyl and Dirac semimetals [16, 17, 18, 19, 20, 21, 22]. Weyl semimetals arise in three dimensions, exploiting the topology in a lower dimensional slice of the Brillouin zone (say the k1−k2k_1-k_2 plane) that undergoes a “phase transition” as the k3k_3 of the slice is varied. Thus, these semimetals are protected by the topology of the two adjacent two-dimensional phases, the gapless points being those k3k_3 at which the quantum phase transition between the two-dimensional phases is affected. They have received considerable attention owing to the exotic properties such as Fermi-arc surface states, interesting nonlinear responses related to the chiral anomaly etc.

The classification of gapped phases hinges on the number of bands being large. In more mathematical terms, these are determined by the stable homotopies of maps from TdT^d to SS which are realized when SS is large dimensional. In the absence of a large number of bands, one can still obtain topological phases that arise from unstable homotopies of maps from TdT^d to SS i. e., when the space SS is “small dimensional”. An example in a three-dimensional lattice that hosts a two-band gapped system is dubbed as a “Hopf insulator” [23, 24, 25] whose topology can be traced to the homotopies of maps from the three-sphere S3S^3 to the two sphere S2S^2.

In this paper, we show how a Hopf insulator in 3 dimensions can be used to construct interesting gapless phases in 4 dimensions. These gapless phases have several new features. Unlike the Weyl semimetal in three dimensions, these four-dimensional semimetals host nodal lines of gapless points (a one-dimensional submanifold) in the four-dimensional Brillouin zone. Remarkably, any three-dimensional surface that encloses one of these rings carries an integer Hopf number that characterizes the phases on either side of these rings. These features manifest spectacularly in the nature of gapless surface states. There are three-dimensional surfaces, which host Fermi-arc states and, in addition, gapless drumhead states [26]. Further, we also find evidence of two-dimensional corner states that arise at the intersection of two three-dimensional surfaces of the four-dimensional insulator. This work presents a new class of interesting topological phases in higher dimensions [27, 28, 29].

Hopf and Hopf-Chern Insulators

We begin with a two-band system that realizes an insulating phase on a 3D cubic lattice with a unit lattice spacing. The Brillouin zone (BZ) of this system is the three torus T3T^3 corresponding to [−π,π]3[-\pi,\pi]^3. A generic point in the BZ is denoted by k=(k1,k2,k3)\mathbf{k}=(k_1,k_2,k_3). A two-band Hamiltonian is defined by

H(k)=d(k)⋅σ(1)H(\mathbf{k})=\mathbf{d}(\mathbf{k})\cdot\boldsymbol{\sigma} \tag*{(1)}

where d(k)=(d1(k),d2(k),d3(k))\mathbf{d}(\mathbf{k})=(d_1(\mathbf{k}),d_2(\mathbf{k}),d_3(\mathbf{k})), and σ=(σ1,σ2,σ3)\boldsymbol{\sigma}=(\sigma_1,\sigma_2,\sigma_3) where σi\sigma_i are the 2×22\times2 Pauli matrices. The chemical potential here and henceforth in this paper is set to zero so that the fermionic many-body system is half-filled (1 particle per site). Existence of a gap necessitates that ∣d(k)∣>0\lvert\mathbf{d}(\mathbf{k})\rvert>0 for k∈T3\mathbf{k}\in T^3, and thus the unit vector d^(k)=d(k)/∣d(k)∣\hat{\mathbf{d}}(\mathbf{k})=\mathbf{d}(\mathbf{k})/\lvert\mathbf{d}(\mathbf{k})\rvert can be identified with a point on the two-sphere S2S^2. Consequently, the Hamiltonian Eq. (1) can be viewed as a map from T3T^3 to S2S^2.

Distinct insulating topological phases are obtained depending on the homotopy class of the map from T3T^3 to S2S^2, with two insulators being identical if they can be smoothly deformed to each other (i.e., homotopic) without closing the gap. Such maps have been extensively studied both from the mathematical and physical perspectives [30, 31, 32, 33, 34]. The homotopy classes of the maps are characterized by four (integer) numbers (χ,(C1,C2,C3))(\chi,(C_1,C_2,C_3)). The numbers CαC_\alpha are the Chern numbers associated with two-dimensional T2T^2 submanifolds of T3T^3, where α\alpha indicates the normal direction to the T2T^2-submanifold. The number χ\chi is in Z2Q\mathbb{Z}_{2Q} where Q=GCD⁡(C1,C2,C3)Q=\operatorname{GCD}(C_1,C_2,C_3). Thus, if all CαC_\alpha are zero, χ\chi can take any integer value, and such insulators are termed as Hopf insulators [23, 24, 25]. On the other hand, if any of the CαC_\alpha is nonzero, then χ\chi takes on only a finite set of values, and such insulators are dubbed as Hopf-Chern insulators [25].

Hopf insulators can be constructed [23, 24] using an intermediate map from T3T^3 to S3S^3 (the three-sphere). Since S3S^3 is described by two complex number z1,z2z_1,z_2 such that ∣z1∣2+∣z2∣2>0\lvert z_1\rvert^2+\lvert z_2\rvert^2>0, the prescription

z1(k,h)=sin⁡k1+isin⁡k2z2(k,h)=sin⁡k3+i(cos⁡k1+cos⁡k2+cos⁡k3+h)(2)\begin{aligned} z_1(\mathbf{k},h)&=\sin k_1+\mathrm{i}\sin k_2\\ z_2(\mathbf{k},h)&=\sin k_3+\mathrm{i}(\cos k_1+\cos k_2+\cos k_3+h) \tag*{(2)} \end{aligned}

(hh is a parameter, i=−1\mathrm{i}=\sqrt{-1}), is a map from T3T^3 to S3S^3. The topological index Γ\Gamma ([35], section S1) of this map, vanishes when ∣h∣>3\lvert h\rvert>3, is 1 for 1<∣h∣<31<\lvert h\rvert<3, and −2-2 for ∣h∣<1\lvert h\rvert<1. The map is thus topologically nontrivial for ∣h∣<3\lvert h\rvert<3. Finally, to obtain a two-band model, the point on S3S^3 is mapped to S2S^2 via the Hopf map [36]

d(p,q)(k,h)=(2ℜ(z1p(k,h)z2∗q(k,h)),2ℑ(z1p(k,h)z2∗q(k,h)),∣z1(k,h)∣2p−∣z2(k,h)∣2q)(3)\mathbf{d}^{(p,q)}(\mathbf{k},h)=\left(2\Re(z_1^p(\mathbf{k},h)z_2^{*q}(\mathbf{k},h)),2\Im(z_1^p(\mathbf{k},h)z_2^{*q}(\mathbf{k},h)),\lvert z_1(\mathbf{k},h)\rvert^{2p}-\lvert z_2(\mathbf{k},h)\rvert^{2q}\right) \tag*{(3)}

where p,qp,q are co-prime integers, ∗* denotes complex conjugation. Such a map has a Hopf index [36] H=±pq\mathcal{H}=\pm pq. The Hopf insulator defined using Eq. (3) has vanishing Chern number CαC_\alpha, and is thus characterized by (χ,(0,0,0))(\chi,(0,0,0)) where χ=ΓH\chi=\Gamma\mathcal{H} [24].

Hopf-Chern insulators are those which have a non-zero Chern numbers CαC_\alpha. These are obtained [25, 37] using

(d1(m)(k)d2(m)(k))=(cos⁡mk1−sin⁡mk1sin⁡mk1cos⁡mk1)(sin⁡k2sin⁡k3)d3(m)(k)=1+Δ1(cos⁡k2+cos⁡k3)+Δ2cos⁡k2cos⁡k3(4)\begin{aligned} \begin{pmatrix}d_1^{(m)}(\mathbf{k})\\d_2^{(m)}(\mathbf{k})\end{pmatrix} &= \begin{pmatrix}\cos mk_1&-\sin mk_1\\\sin mk_1&\cos mk_1\end{pmatrix} \begin{pmatrix}\sin k_2\\\sin k_3\end{pmatrix}\\ d_3^{(m)}(\mathbf{k})&=1+\Delta_1(\cos k_2+\cos k_3)+\Delta_2\cos k_2\cos k_3 \tag*{(4)} \end{aligned}

where mm is an integer and Δ1,Δ2\Delta_1,\Delta_2 are real parameters. For this model, C2=C3=0C_2=C_3=0 always, and C1C_1 is determined by the values of Δ1\Delta_1 and Δ2\Delta_2 (and is not affected by the value of mm). The quantity χ\chi is determined by mm as χ=m∣C1∣ mod 2∣C1∣\chi=m\lvert C_1\rvert\mathop{\bmod}2\lvert C_1\rvert.

Hopf semimetals: Realization of such topological phases in dd dimensions allows us to construct interesting semimetallic phases in d+1d+1 dimensions. In fact, the well-studied Weyl semimetallic phases are examples of such physics. These phases in three dimensions arise from the topological Chern insulators in two dimensions and enjoy a degree of protection owing to the stability of Weyl points where di(k)=0d_i(\mathbf{k})=0. The Weyl points are

those at which all components of dd vanish, each such equation describing a surface embedded in three dimensions. Three such surfaces generically intersect at isolated points, leading to the stability of Weyl points to small perturbations of the Hamiltonian. Taking this idea to a four-dimensional two-band system, semimetallicity will require that di(K)=0d_i(\mathbf{K})=0 for i=1,2,3i=1,2,3 where the gap closes; K\mathbf{K} is a point in the BZ T4T^4 (four-torus, [−π,π]4[-\pi,\pi]^4) of the 4D cubic lattice with K=(k1,k2,k3,k4)=(k,k4)\mathbf{K}=(k_1,k_2,k_3,k_4)=(\mathbf{k},k_4). This condition, if satisfied, will be generically met on a one-dimensional submanifold of T4T^4. The conclusion is that the semimetals arising in these four-dimensional systems will generically possess nodal lines. The exciting aspect here is that these line nodes enjoy a degree of protection in that small perturbations cannot remove them but, at best, change their shape.

To construct such a semimetal arising from the topology of the Hopf insulator, we first study a quantum phase transition that occurs in the 3D Hopf insulator. Consider a three-dimensional system with a tuning parameter λ\lambda

d(k,h,λ)=(1−λ)d(1,1)(k,h)+λdf(k)(5)\mathbf{d}(\mathbf{k},h,\lambda)=(1-\lambda)\mathbf{d}^{(1,1)}(\mathbf{k},h)+\lambda\mathbf{d}^{f}(\mathbf{k}) \tag*{(5)}

where d(1,1)(k)\mathbf{d}^{(1,1)}(\mathbf{k}) is the dispersion in Eq. (3), and df(k)=(0,0,1)\mathbf{d}^{f}(\mathbf{k})=(0,0,1) is the dispersion of a gapped flat band system which is topologically trivial. When λ=0\lambda=0, the system hosts Hopf insulating phases in a range of the parameter hh, and a topologically trivial phase for λ=1\lambda=1. For intermediate values of λ\lambda we obtain a variety of semimetallic phases (see Figure 1). These semimetallic phases are characterized by a change in Chern numbers of the T2T^2 submanifolds of T3T^3 [35], section S2), as illustrated for λ=1/4\lambda=1/4 in Figure 1 (inset).

Image not extracted

Figure 1. Phase diagram of system in Eq. (5). Semimetallic regions with horizontal lines have band touchings at (0,0,±kt(1))(0,0,\pm k_t^{(1)}), those with vertical lines have touchings at (π,π,±kt(−1))(\pi,\pi,\pm k_t^{(-1)}), and those with slanted lines have touching at (π,0,±kt(0)),(0,π,±kt(0))(\pi,0,\pm k_t^{(0)}),(0,\pi,\pm k_t^{(0)}), where ktl=arccos⁡(λ−1−(h+2l)22(h+2l))k_t^{l}=\arccos\left(\frac{\lambda-1-(h+2l)^2}{2(h+2l)}\right). Inset: Chern number of T2T^2 submanifold of T3T^3 BZ labelled by k3k_3, for h=−3,λ=1/4h=-3,\lambda=1/4.

We construct a semimetallic phase on a four-dimensional cubic lattice by defining for each K∈T4K \in T^{4}

d(K,λ)=(1−λ)d(1,1)(k,−3+cos⁡k4)+λdf(k)(6)d(K,\lambda)=(1-\lambda)d^{(1,1)}(k,-3+\cos k_{4})+\lambda d^{f}(k) \tag*{(6)}

The Hamiltonian Eq. (1) obtained using this produces a semimetallic phase for a range 0≤λ<1/20 \leq\lambda< 1/2. Focusing first on λ=0\lambda=0, we find that the bulk gap closes at two points in T4T^{4}, namely KH±(0,0,0,±π/2)K^{\pm}_{H}(0,0,0,\pm\pi/2), where the bands touch quadratically [35], section S3). Most interestingly, these points are a source of “Hopf flux” in T4T^{4}; this is most easily seen by enclosing, for example, the point K+K^{+} by a ball B+=∣K−K+∣≤ϵB^{+}=\lvert K-K^{+}\rvert\leq\epsilon where ϵ\epsilon is a small number, the boundary of this ball ∂B+\partial B^{+} is homeomorphic to S3S^{3}, and d(K,0)d(K,0), K∈∂B+K\in\partial B^{+} defines a map from S3S^{3} to S2S^{2}. Interestingly, the map carries a non-vanishing Hopf index H=−1\mathcal{H}=-1! ([35], section S3), pointing to the topological nature of this “Hopf semimetal” similar to what is found in a three-dimensional Weyl semimetal.

However, as noted in the discussion above, point touching of two bands in 4 dimensions is not stable (in contrast to the 3D Weyl semimetal), and this is indeed seen in our construction. For a small λ>0\lambda>0, we find that the two Hopf points evolve to nodal lines where the bands touch linearly except at two points on the nodal line ([35], section S3). With increasing λ\lambda, the size of the nodal lines centered around K±K_{\pm} increases. Figure 2 (thick blue lines) shows the nodal lines for λ=1/4\lambda=1/4 in the k1=0k_{1}=0 T3T^{3} sub-manifold of the T4T^{4} BZ. The nodal lines appear in the k3−k4k_{3}-k_{4} plane (k1=k2=0k_{1}=k_{2}=0), and encircle the Hopf points KH±K^{\pm}_{H} extending from k4min⁡≤∣k4∣≤k4max⁡k_{4}^{\min}\leq\lvert k_{4}\rvert\leq k_{4}^{\max}, k4min⁡=arccos⁡λ1−λk_{4}^{\min}=\arccos\sqrt{\frac{\lambda}{1-\lambda}}, k4max⁡=π−arccos⁡λ1−λk_{4}^{\max}=\pi-\arccos\sqrt{\frac{\lambda}{1-\lambda}} and described by the equation

Image not extracted

Figure 2. Hopf nodal line semimetal for λ=1/4\lambda= 1/4 in Eq. (6). Blue: nodal lines (nodal lines lie in the k1=0k_{1}=0 submanifold of T4T^{4}). The figure also depicts T3T^{3} the surface BZ of the (1,0,0,0)(1,0,0,0) surface. Green: Fermi arc surface states that correspond to the (1,0,0)(1,0,0) surface states of the χ=1\chi=1 Hopf insulator. Red: Drumhead surface states that correspond to the edge states of the C3=1C_{3}=1 Chern insulator.

2(1−cos⁡k3)(1−cos⁡k4)+cos⁡2k4=λ1−λ.(7)2(1-\cos k_{3})(1-\cos k_{4})+\cos^{2}k_{4}=\frac{\lambda}{1-\lambda}. \tag*{(7)}

The nodal lines L±L^{\pm} respectively encircle KH±K^{\pm}_{H}. The intriguing aspect is that the nodal lines also carry the same Hopf number, i.e. if we place balls B±B^{\pm} centered around KH±K^{\pm}_{H}, and enclosing the nodal lines L±L^{\pm}, then the Hamiltonian on the surface of the ball ∂B±\partial B^{\pm} defines a Hopf map such that the Hopf invariant associated with the L±L^{\pm} nodal lines are opposite of each other. This demonstrates the topological origin of the nodal lines and their stability. The nodal lines which appear between ±k4min⁡\pm k_{4}^{\min} and ±k4max⁡\pm k_{4}^{\max} separate three dimensional T3T^{3} sub-manifolds of T4T^{4} that carry distinct invariants. Indeed, for all the T3T^{3} submanifolds with ∣k4∣<k4min⁡\lvert k_{4}\rvert<k_{4}^{\min}, the invariant χ=1\chi=1. This change of topology of the bands along k4k_{4} is encoded in the Hopf number on the surface of ∂B±\partial B^{\pm}.

We next investigate the nature of the surface states of the four-dimensional Hopf semimetal. The surface of this system is characterized by a normal direction, and is a “three-dimensional crystal” with a T3T^{3} surface Brillouin zone. Depicted in Figure 2 for the surface with the (1,0,0,0)(1,0,0,0) normal, are a remarkably rich set of surface states. First, there is a set of gapless “Fermi-arc” states that exist between ±k4min⁡\pm k_{4}^{\min}, depicted by the solid green line in Figure 2. These arise from the (1,0,0)(1,0,0) surface states of the χ=1\chi=1 Hopf insulator realized in the T3T^{3} submanifolds in this regime of k4k_{4}. There are additional surface states that arise in the regime k4min⁡<∣k4∣<k4max⁡k_{4}^{\min}<\lvert k_{4}\rvert<k_{4}^{\max}. In fact, all the points in the T3T^{3} surface BZ that are inside the nodal line projected onto the surface BZ host gapless states that are higher-dimensional analogs of drumhead states ([26] and references therein). Details of all of these states may be found in [35], section S4.

The Hopf semimetal holds further interesting aspects when we study the surface states on the (0,0,1,0)(0,0,1,0) surface. This surface hosts two types of gapless states (see

Figure 3. We find, first, a “Fermi surface” of gapless states between ∣k4∣<k4min⁡|k_4|<k_4^{\min}; these are the surface states of the χ=1\chi=1 Hopf insulator that is realized in the T3T^3 submanifolds. In addition, there are other gapless states shown by the blue lines of the same figure; these are gapless states corresponding to the projection of the gapless nodal line on to the surface BZ.

Image not extracted

Figure 3. Surface states of the Hopf nodal line semimetal in the (0,0,1,0)(0,0,1,0) surface BZ. Red: “Fermi surface” states that correspond to the (0,0,1)(0,0,1) surface states of the χ=1\chi=1 Hopf insulator. Blue: Fermi arc states that arise from the projection of the nodal lines onto the surface BZ.

Finally, we also point out the possibility of interesting “corner states” in this Hopf semimetal that arise in the two-dimensional intersection of two three-dimensional surfaces. As an example, The corner formed by the intersection of two surfaces (1,0,0,0)(1,0,0,0) and (0,1,0,0)(0,1,0,0), will have a two-dimensional T2T^2 Brillouin zone labeled by (k3,k4)(k_3,k_4). The corner states arise because the corner terminates 1−21-2 planes of the four-dimensional crystal. In the instance of λ=1/4\lambda=1/4, some of the 1−21-2 submanifolds (which are T2T^2) host non-zero Chern numbers in the regime k4min⁡<∣k4∣<k4max⁡k_4^{\min}<|k_4|<k_4^{\max}, and should result in the “corner drumhead states” in the T2T^2 Brillouin zone. Other corners (intersections of different three-dimensional surfaces) will host Fermi arc states. While our calculations are consistent with this possibility, a full demonstration of this requires very large system sizes.

In the last part of this paper, we demonstrate that such semimetals in four dimensions can also be constructed out of the Hopf-Chern insulators. To achieve this we first construct as 3d system with a parameter λ\lambda,

d(k,λ)=(1−λ)d(1)(k)+λd(0)(k)(8)d(k,\lambda)=(1-\lambda)d^{(1)}(k)+\lambda d^{(0)}(k) \tag*{(8)}

where d(1)d^{(1)} and d(0)d^{(0)} are obtained from Eq. (4) with 3Δ1=Δ2=323\Delta_1=\Delta_2=\frac{3}{2}.

This system undergoes a quantum phase transition from a Hopf-Chern insulator with (χ=2,(C1=2,C2=0,C3=0))(\chi=2,(C_1=2,C_2=0,C_3=0)) to another insulator with (χ=0,(C1=2,C2=0,C3=0))(\chi=0,(C_1=2,C_2=0,C_3=0)) which can be viewed as a stack of Chern insulators. The phase transition[38] occurs at λ=1/2\lambda=1/2 where the gap closes along two nodal-line rings, and χ\chi changes from 2 to 0 for λ>1/2\lambda>1/2, with no change in the Chern numbers, i.e., C1=2C_1=2 for λ≠1/2\lambda\ne1/2 (C2,3=0C_{2,3}=0).

We use the above to construct a Hopf-Chern semimetal in a four dimensional cubic lattice, via

d(K)=(1−cos⁡k42)d(1)(k)+(1+cos⁡k42)d(0)(k)(9)d(\mathbf{K})=\left(\frac{1-\cos k_4}{2}\right)d^{(1)}(k)+\left(\frac{1+\cos k_4}{2}\right)d^{(0)}(k) \tag*{(9)}

where K=(k,k4)\mathbf{K}=(k,k_4) is a point in BZ T4T^4. This system hosts nodal lines in the k2−k3k_2-k_3 planes with k4=±π2k_4=\pm\frac{\pi}{2} as shown by blue lines in Figure 4(a). The T3T^3 submanifolds of T4T^4 with ∣k4∣<π/2|k_4|<\pi/2 have Hopf-Chern character with χ=2\chi=2. The topological aspects of this semimetal are again evident in the nature of the surface states that it hosts. On a surface with (1,0,0,0)(1,0,0,0) as the normal, we find a “Fermi-cylinder” of gapless states (see Figure 4(a).) The end-arcs of these cylinders are the projection of the nodal lines onto the surface BZ. These arise from the surface states of the Hopf-Chern insulators residing in ∣k4∣<π/2|k_4|<\pi/2 region of T4T^4 BZ. Turning now to the (0,1,0,0)(0,1,0,0) surface (see Figure 4(b)), we find that there are two types of gapless states. Since this also terminates at 2-direction of the crystal, it terminates 2−3 planes of the crystal, which carry a Chern number (independent of the value of k4k_4). Thus, there is a set of gapless states on the planes with k3=0k_3=0 and k3=πk_3=\pi. Finally there is a second set of gapless states that arise as disjoint nodal lines on the k1=πk_1=\pi plane of the surface BZ; these states are the projections of the nodal lines on to the surface BZ. It is also clear that this system can host a variety of corner states, states residing on the two-dimensional intersection of two three-dimensional surfaces.

Image not extracted

Figure 4. (a) Hopf-Chern nodal line semimetal. Blue: Nodal lines that lie in the T3T^3 submanifold with k1=πk_1=\pi of T4T^4. The figure also depicts the (1,0,0,0)(1,0,0,0) surface BZ. Red: “Fermi-cylinder” states that correspond to surface states of (1,0,0)(1,0,0) surface of the (2mod4,(2,0,0))(2\mathrm{mod}4,(2,0,0)) Hopf-Chern insulator. (b) (0,1,0,0)(0,1,0,0) surface states of the Hopf-Chern nodal line semimetal. They correspond to the edge states of C=2C=2 Chern insulator. Here, the blue lines depict the projection of the nodal lines on to the surface Brillouin zone.

It is interesting to explore the possibilities of experimental realization of these four-dimensional Hopf-semimetals and their surface states exploiting the ideas of synthetic dimensions[39] in cold atoms[40] and photonic systems[41]. Further theoretical investigations should also be fruitful. It will be interesting to find generalizations of such semimetals using other recently proposed topological phases[42] in three dimensions. Understanding the responses[37, 43] of the Hopf semimetals also provides an exciting direction.

Acknowledgement: BPP thanks the PMRF program for support. VBS acknowledges DST-SERB, India, for support through a MATRICS grant.

References

Supplemental Material for Hopf Semimetals

Bhandaru Phani Parasar; Vijay B. Shenoy

Topological Index of Map from T3T^3 to S3S^3

A gapped two-band hamiltonian in three dimensions can be thought of as a map from the three-torus T3T^3 to the two-sphere S2S^2. A tight binding model for Hopf insulator on a cubic lattice in three dimensions can thus be constructed [23, 24] by first mapping a point on T3T^3 to a point on S3S^3 (Eq. (2), main text). This point is then mapped to a point on S2S^2 via the Hopf map (Eq. (3), main text). The T3T^3 to S3S^3 map is given by

z1(k,h)=sin⁡k1+isin⁡k2z2(k,h)=sin⁡k3+i(cos⁡k1+cos⁡k2+cos⁡k3+h)(10)\begin{aligned} z_1(k,h) &= \sin k_1+\mathrm{i}\sin k_2 \\ z_2(k,h) &= \sin k_3+\mathrm{i}(\cos k_1+\cos k_2+\cos k_3+h) \tag*{(10)} \end{aligned}

where hh is a real parameter such that ∣z1∣2+∣z2∣2>0 ∀k∈T3|z_1|^2+|z_2|^2>0\ \forall k\in T^3. The homotopic classification of maps from T3T^3 to S3S^3 is Z\mathbb{Z} and the homotopic invariant Γ\Gamma encodes the winding of T3T^3 on to S3S^3. For the map given above, it is

Γ=112π2∫BZdk ϵαβγρϵμντ1∣z∣4zα∂μzβ∂νzγ∂τzρ(11)\Gamma=\frac{1}{12\pi^2}\int_{\mathrm{BZ}} \mathrm{d}k\,\epsilon_{\alpha\beta\gamma\rho}\epsilon_{\mu\nu\tau}\frac{1}{|z|^4}z_\alpha\partial_\mu z_\beta\partial_\nu z_\gamma\partial_\tau z_\rho \tag*{(11)}

where z=(Re(z1),Im(z1),Re(z2),Im(z2))z=(\mathop{\mathrm{Re}}(z_1),\mathop{\mathrm{Im}}(z_1),\mathop{\mathrm{Re}}(z_2),\mathop{\mathrm{Im}}(z_2)). The homotopic invariant depends on the parameter hh according to

Γ={0∣h∣>311<∣h∣<3−2∣h∣<1(12)\Gamma= \begin{cases} 0 & |h|>3 \\ 1 & 1<|h|<3 \\ -2 & |h|<1 \end{cases} \tag*{(12)}

Finally, the T3T^3 to S2S^2 map is given by composing the above with the Hopf map to give:

d(p,q)(k,h)=(2Re(z1p(k,h)z2∗q(k,h)),2Im(z1p(k,h)z2∗q(k,h)),∣z1(k,h)∣2p−∣z2(k,h)∣2q)(13)\mathbf{d}^{(p,q)}(k,h)=\left(2\mathop{\mathrm{Re}}(z_1^p(k,h)z_2^{*q}(k,h)),2\mathop{\mathrm{Im}}(z_1^p(k,h)z_2^{*q}(k,h)),|z_1(k,h)|^{2p}-|z_2(k,h)|^{2q}\right) \tag*{(13)}

Here p,qp,q are co-prime integers. For a map from T3T^3 to S2S^2 constructed this way, it can be shown [24] that the Chern numbers C1,C2,C3C_1,C_2,C_3 associated with the T2T^2 sub-manifolds of T3T^3 vanish. Thus the homotopy of this map is characterized by χ\chi, which is equal to the product of the homotopic invariant from T3T^3 to S3S^3 (Γ\Gamma) and the Hopf index from S3S^3 to S2S^2 (H=±pq\mathcal{H}=\pm pq): χ=ΓH\chi=\Gamma\mathcal{H}

Phase transition in 3D Hopf Insulator

To illustrate the nature of phase transition between distinct Hopf insulators, and its implications for Hopf semimetals, we consider the following two-band system in three dimensions, with real parameters λ,h\lambda,h (Eq. (5), main text):

d(k,h,λ)=(1−λ)d(1,1)(k,h)+λdf(k)(14)\mathbf{d}(k,h,\lambda)=(1-\lambda)\mathbf{d}^{(1,1)}(k,h)+\lambda\mathbf{d}^{f}(k) \tag*{(14)}

Here, d(1,1)(k,h)\mathbf{d}^{(1,1)}(k,h) is as given in Eq. (13), and df(k)=(0,0,1)\mathbf{d}^{f}(k)=(0,0,1) corresponds to a flat band dispersion. The phase diagram of this system in λ−h\lambda-h plane is given in Figure 1, main text. When λ=0\lambda=0, the system is a trivial insulator for h<−3h<-3, Hopf insulator with χ=1\chi=1 for −3<h<−1-3<h<-1, and a Hopf insulator with χ=−2\chi=-2 for h>−1h>-1. i.e., if λ=0\lambda=0 is fixed and hh is varied, there is a quantum phase transition occurring at h=−3h=-3 between χ=0\chi=0 and χ=1\chi=1 insulators. The semimetal at the critical point has quadratic band touching at k0=(0,0,0)k^0=(0,0,0). To see this, we expand the Bloch hamiltonian to the lowest non-trivial order in q=k−k0q=k-k^0:

H(k0+q)≈(q12+q22−q32)σ3+2q3(q1σ1+q2σ2)(15)H(k^0+q)\approx(q_1^2+q_2^2-q_3^2)\sigma_3+2q_3(q_1\sigma_1+q_2\sigma_2) \tag*{(15)}

Note that as hh is varied, the transition between distinct Hopf insulators occurs at a sharply defined value of hh only for the finely tuned value of λ=0\lambda=0. Generically, the transition between distinct Hopf insulators involves a region of semimetal.

Now, we identify various semimetallic regions shown in Figure 1, main text. It is immediately seen that the spectrum of S2.1 has a vanishing gap at kk if

z1(k)=0(16)z_1(k)=0 \tag*{(16)}
∣z2(k)∣2=λ1−λ(17)|z_2(k)|^2=\frac{\lambda}{1-\lambda} \tag*{(17)}

The possible solutions are

k=(0,0,±kt(1));kt(1)=arccos⁡(λ1−λ−1−(h+2)22(h+2))k=(0,0,\pm k_t^{(1)}); \quad k_t^{(1)}=\arccos\left(\frac{\lambda}{1-\lambda}-1-(h+2)^2\over2(h+2)\right)
k=(π,π,±kt(−1));kt(−1)=arccos⁡(λ1−λ−1−(h−2)22(h−2))k=(\pi,\pi,\pm k_t^{(-1)}); \quad k_t^{(-1)}=\arccos\left(\frac{\lambda}{1-\lambda}-1-(h-2)^2\over2(h-2)\right)
k=(0,π,±kt(0)),  (π,0,±kt(0));kt(0)=arccos⁡(λ1−λ−1−h22h)(18)k=(0,\pi,\pm k_t^{(0)}),\;(\pi,0,\pm k_t^{(0)});\quad k_t^{(0)}=\arccos\left(\frac{\lambda}{1-\lambda}-1-h^2\over2h\right) \tag*{(18)}

In each of the above, a solution exists only in the region of the λ−h\lambda-h plane for which the argument of arccos lies between −1-1 and +1+1. The regions where these solutions exist are shown by horizontal, vertical and slanted lines respectively in Figure 1, main text.

In the following, we focus on the semimetal labelled SM1, (Figure 1, main text) which separates the χ=0\chi=0 and χ=1\chi= 1 insulators. The band gap of this semimetal vanishes at k±=(0,0,±kt(1))\mathbf{k}^{\pm} = (0, 0, \pm k_t^{(1)}). A direct computation shows that this semimetal is characterized by a change in the Chern number C3C_3 across the BZ. The corresponding plot for λ=14\lambda= \frac{1}{4}, h=−3h = -3 is shown in Figure 1 (inset), main text.

C3(k3)={0kt(1)<∣k3∣<π1∣k3∣<kt(1)(19)C_3(k_3) = \begin{cases} 0 & k_t^{(1)} < |k_3| < \pi\\ 1 & |k_3| < k_t^{(1)} \end{cases} \tag*{(19)}

Expanding the Bloch hamiltonian about k+\mathbf{k}^{+}, we find that the bands cross linearly: H(k++q)≈2(1−λ) a(q)⋅σH(\mathbf{k}^{+}+\mathbf{q}) \approx2(1-\lambda)\,\mathbf{a}(\mathbf{q})\cdot\boldsymbol{\sigma} where

a1(q)=q1sin⁡kt(1)+q2(h+2+cos⁡kt(1))a2(q)=q2sin⁡kt(1)−q1(h+2+cos⁡kt(1))a3(q)=q3(h+2)sin⁡kt(1)(20)\begin{aligned} a_1(\mathbf{q}) &= q_1 \sin k_t^{(1)} + q_2\left(h+2+\cos k_t^{(1)}\right) \\ a_2(\mathbf{q}) &= q_2 \sin k_t^{(1)} - q_1\left(h+2+\cos k_t^{(1)}\right) \\ a_3(\mathbf{q}) &= q_3(h+2)\sin k_t^{(1)} \tag*{(20)} \end{aligned}

Note that, for the semimetal corresponding to a point on the boundary between the χ=1\chi= 1 insulator and SM1, the coefficient of q3q_3 in the Taylor expansion above vanishes, and the bands touch quadratically (along the k3k_3 direction) at k±=(0,0,0)\mathbf{k}^{\pm} = (0,0,0)

Hopf semimetals

We construct a two-band semimetallic system on a cubic lattice in four dimensions with 0<λ<120 < \lambda< \frac{1}{2} as a tuning parameter. Let us denote K=(k,k4)∈T4\mathbf{K} = (\mathbf{k},k_4) \in T^4 and k=(k1,k2,k3)\mathbf{k} = (k_1,k_2,k_3)

d(K,λ)=(1−λ)d(1,1)(k,−3+cos⁡k4)+λdf(k)(21)\mathbf{d}(\mathbf{K},\lambda) = (1-\lambda)\mathbf{d}^{(1,1)}(\mathbf{k},-3+\cos k_4)+\lambda\mathbf{d}^{f}(\mathbf{k}) \tag*{(21)}

When λ=0\lambda= 0, we obtain a semimetal with bands touching at exactly two points K±=(0,0,0,±π2)\mathbf{K}^{\pm} = (0,0,0,\pm\frac{\pi}{2}). Expanding the hamiltonian to lowest order near the band touching points reveals the interesting topological aspects of this semimetal. Writing H(K±+Q)≈a±(Q)⋅σH(\mathbf{K}^{\pm}+\mathbf{Q}) \approx\mathbf{a}^{\pm}(\mathbf{Q})\cdot\boldsymbol{\sigma}, where Q=(q1,q2,q3,q4)\mathbf{Q}=(q_1,q_2,q_3,q_4)

a1±(Q)=2(q1q3∓q2q4)(22)a_1^{\pm}(\mathbf{Q}) = 2(q_1q_3 \mp q_2q_4) \tag*{(22)}
a2±(Q)=2(q2q3±q1q4)(23)a_2^{\pm}(\mathbf{Q}) = 2(q_2q_3 \pm q_1q_4) \tag*{(23)}
a3±(Q)=q12+q22−q32−q42(24)a_3^{\pm}(\mathbf{Q}) = q_1^2+q_2^2-q_3^2-q_4^2 \tag*{(24)}

Note that these are the well-known Hopf maps with Hopf indices −1,+1-1,+1 respectively. This beautifully demonstrates that the points K±\mathbf{K}^{\pm} are sources of “Hopf flux”. Consider a ball B+B^{+} centered at the point K+\mathbf{K}^{+}. Its boundary ∂B+\partial B^{+}, homeomorphic to S3S^3 encloses the point. Now, the restriction of the hamiltonian S3.1 to ∂B+\partial B^{+} defines a map from S3S^3 to S2S^2. From the above, we see that it belongs to the homotopic class labelled by the Hopf index H=−1\mathcal{H}=-1. Similarly, the restriction of the hamiltonian to a surface homeomorphic to S3S^3, and one which encloses the point K−\mathbf{K}^{-} defines a map with Hopf index H=+1\mathcal{H}=+1

As discussed in the main text, point touching in a two-band system in four dimensions is unstable, and the band touching points evolve to nodal lines L±L^{\pm} for an infinitesimal λ>0\lambda>0. Importantly, any surface homeomorphic to S3S^3 and enclosing the nodal lines L±L^{\pm} still carries the Hopf index H=∓1\mathcal{H}=\mp1. It is easily seen that the nodal lines lie in the k3−k4k_3-k_4 plane (i.e., k1=k2=0k_1=k_2=0) of the BZ. The equation of the nodal lines can be obtained by the replacement kt(1)→k3k_t^{(1)}\to k_3, h→−3+cos⁡k4h\to-3+\cos k_4 in Eq. (18). We find that the nodal lines are given by the curves

2(1−cos⁡k3)(1−cos⁡k4)+cos⁡2k4=λ1−λ(25)2(1-\cos k_3)(1-\cos k_4)+\cos^2 k_4=\frac{\lambda}{1-\lambda} \tag*{(25)}

The nodal line lying in the region k4>0k_4>0 extends from k4min⁡≤k4≤k4max⁡k_4^{\min}\le k_4\le k_4^{\max}. Observing that k3=0k_3=0 when k4=k4min⁡k_4=k_4^{\min} or k4=k4max⁡k_4=k_4^{\max}, we obtain

k4min⁡=arccos⁡λ1−λ(26)k_4^{\min}=\arccos\sqrt{\frac{\lambda}{1-\lambda}} \tag*{(26)}
k4max⁡=π−arccos⁡λ1−λ(27)k_4^{\max}=\pi-\arccos\sqrt{\frac{\lambda}{1-\lambda}} \tag*{(27)}

Finally, for the semimetal λ>0\lambda>0, a Taylor expansion of the Bloch hamiltonian about a point on the nodal line shows that the bands cross linearly expect at the points given by k3=0k_3=0. Let K0=(0,0,k30,k40)\mathbf{K}^{0}=(0,0,k_3^0,k_4^0) be a point on the nodal line. i.e., (k30,k40)(k_3^0,k_4^0) lies on the curve Eq. (25). Now, taylor-expanding the hamiltonian about this point gives H(K0+Q)≈2(1−λ)a0(Q)⋅σH(\mathbf{K}^{0}+\mathbf{Q})\approx2(1-\lambda)\mathbf{a}^{0}(\mathbf{Q})\cdot\boldsymbol{\sigma} with Q=(q1,q2,q3,q4)\mathbf{Q}=(q_1,q_2,q_3,q_4) and

a10(Q)=q1sin⁡k30+q2(cos⁡k30+cos⁡k40−1)a20(Q)=q2sin⁡k30−q1(cos⁡k30+cos⁡k40−1)a30(Q)=q3sin⁡k30(cos⁡k40−1)+q4sin⁡k40(cos⁡k30+cos⁡k40−1)(28)\begin{aligned} a_1^0(\mathbf{Q}) &= q_1\sin k_3^0+q_2(\cos k_3^0+\cos k_4^0-1) \\ a_2^0(\mathbf{Q}) &= q_2\sin k_3^0-q_1(\cos k_3^0+\cos k_4^0-1) \\ a_3^0(\mathbf{Q}) &= q_3\sin k_3^0(\cos k_4^0-1)+q_4\sin k_4^0(\cos k_3^0+\cos k_4^0-1) \tag*{(28)} \end{aligned}

At a given point on the nodal line, a(Q)\mathbf{a}(\mathbf{Q}) must vanish along the tangent to the curve in the k3−k4k_3-k_4 plane, and the same is seen from above.

Surface states in Hopf semimetals

We have seen that the T3T^3 submanifolds of the BZ of the Hopf semimetal carry a Hopf invariant χ=1\chi=1 for ∣k4∣<k4min⁡|k_4|<k_4^{\min}. Hence, we expect Fermi-arc like surface states on the surfaces normal to (1,0,0,0)(1,0,0,0), (0,1,0,0)(0,1,0,0) and (0,0,1,0)(0,0,1,0), from the fact that the three-dimensional Hopf insulators themselves host surface states. To investigate the nature of surface states on the surface normal to (1,0,0,0)(1,0,0,0) direction, we consider a system of finite size with open boundary conditions (slab) in the (1,0,0,0)(1,0,0,0) direction. This explicitly breaks the lattice translation symmetry along the (1,0,0,0)(1,0,0,0) direction, but crystal momenta k2,k3,k4k_2,k_3,k_4 are still good quantum numbers. Thus, the Brillouin zone of the surface is T3T^3. The real space hoppings along (1,0,0,0)(1,0,0,0) can be found by inverse-Fourier transforming the Bloch hamiltonian, and the spectrum of single-particle energies is thus obtained. Assuming that the slab is LL-site thick, with two orbitals AA, BB per site, we have 2L2L bands in the BZ T3\mathbb{T}^3.

Image not extracted

Figure S.1. Surface states of the surface normal to (1,0,0,0)(1,0,0,0) on an L=256L=256-site thick slab. Energies of the two bands above and the two below the Fermi energy are shown. (a) Band dispersion for k2=k3=0k_2=k_3=0, k4∈[0,π)k_4\in[0,\pi). (b) Band dispersion for k2=0k_2=0, k4=π2k_4=\frac{\pi}{2}, k3∈[0,π2]k_3\in[0,\frac{\pi}{2}]. (c) The magnitudes of wavefunctions on AA orbital for the two fermi-arc states at k2=k3=k4=0k_2=k_3=k_4=0, localized at the same end. Even though L=256L=256, wavefunctions are shown only on the first thirty sites for clarity.

The (1,0,0,0)(1,0,0,0) surface has “Fermi-arc” surface states for k2=0k_2=0, k3=0k_3=0, ∣k4∣<k4min|k_4|<k_{4\mathrm{min}}. On this fermi-arc, there are four surface states (two orthogonal states localized on each end of the slab), which correspond to the (1,0,0)(1,0,0) surface states of the three-dimensional Hopf insulator (S1.4, p=q=1p=q=1). The wavefunctions of fermi-arc states at the point k2=k3=k4=0k_2=k_3=k_4=0 are shown in Fig. S.1(c). In addition to these, there are two surface states at every point interior to the curve bounded by the projected nodal lines in the k3−k4k_3-k_4 plane. They correspond to the edge states of the 1−21-2 submanifolds of T4\mathbb{T}^4 which have

non-zero Chern numbers for k4min<∣k4∣<k4maxk_{4\mathrm{min}}<|k_4|<k_{4\mathrm{max}}. The energies of the bands just above and below the Fermi energy as k4k_4 is varied, at k2=k3=0k_2=k_3=0, are shown in Fig. S.1(a). In Fig. S.1(b), the band dispersion is shown as k3k_3 is varied at k2=0k_2=0, k4=π2k_4=\frac{\pi}{2}. The nature of surface states on the surface normal to (0,1,0,0)(0,1,0,0) is similar to those on the (1,0,0,0)(1,0,0,0) surface discussed above. Analogous computation for the (0,0,1,0)(0,0,1,0) surface reveals that there is “Fermi-surface” of surface states which joins with the nodal lines projected on to the surface BZ, as shown in Figure 3, main text. There are two surface states on each point of the fermi-surface, with one state localized on each end of the slab. The band energies for k2=k4=0k_2=k_4=0, as k1k_1 is varied, are shown in Fig. S.2(a). The surface state wavefunction for (k1,k2,k4)=(0.841,0,0)(k_1,k_2,k_4)=(0.841,0,0) (a point on the Fermi-surface) is also plotted in Fig. S.2(b).

Image not extracted

Figure S.2. Surface states of the surface normal to (0,0,1,0)(0,0,1,0) surface on a L=256L = 256-site thick slab. (a) Dispersion of the band just above and the band just below the Fermi energy for k2=0k_2 = 0, k4=0k_4 = 0, k1∈[0,π)k_1 \in[0,\pi). (b) The magnitude of the surface state wavefunction on AA orbital at the point (k1,k2,k4)=(0.841,0,0)(k_1,k_2,k_4) = (0.841,0,0) on the Fermi-surface. L=256L = 256, but the wavefunction is shown only on the first thirty sites.

References

  1. [1]L. Fu, C. L. Kane, and E. J. Mele, Phys. Rev. Lett. 98, 106803 (2007).
  2. [2]C. L. Kane and E. J. Mele, Phys. Rev. Lett. 95, 146802 (2005).
  3. [3]J. E. Moore and L. Balents, Phys. Rev. B 75, 121306 (2007).
  4. [4]R. Roy, Phys. Rev. B 79, 195322 (2009).
  5. [5]A. Kitaev, AIP Conference Proceedings 1134, 22 (2009).
  6. [6]S. Ryu, A. P. Schnyder, A. Furusaki, and A. W. Ludwig, New Journal of Physics 12, 065010 (2010).
  7. [7]X.-L. Qi and S.-C. Zhang, Rev. Mod. Phys. 83, 1057 (2011).
  8. [8]C.-K. Chiu, J. C. Y. Teo, A. P. Schnyder, and S. Ryu, Rev. Mod. Phys. 88, 035005 (2016).
  9. [9]R.-J. Slager, A. Mesaros, V. Juričić, and J. Zaanen, Nature Physics 9, 98 (2013).
  10. [10]W. A. Benalcazar, T. Li, and T. L. Hughes, Phys. Rev. B 99, 245151 (2019).
  11. [11]B. Bradlyn, L. Elcoro, J. Cano, M. G. Vergniory, Z. Wang, C. Felser, M. I. Aroyo, and B. A. Bernevig, Nature 547, 298 (2017).
  12. [12]A. Altland and M. R. Zirnbauer, Phys. Rev. B 55, 1142 (1997).
  13. [13]M. R. Zirnbauer, ArXiv e-prints (2010), arXiv:1001.0722 [math-ph].arxiv.org/abs/1001.0722
  14. [14]A. Agarwala, A. Haldar, and V. B. Shenoy, Annals of Physics 385, 469 (2017).
  15. [15]A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and A. K. Geim, Rev. Mod. Phys. 81, 109 (2009).
  16. [16]A. A. Burkov, M. D. Hook, and L. Balents, Phys. Rev. B 84, 235126 (2011).
  17. [17]P. Hosur and X. Qi, Comptes Rendus Physique 14, 857 (2013), topological insulators / Isolants topologiques.
  18. [18]O. Vafek and A. Vishwanath, Annual Review of Condensed Matter Physics 5, 83 (2014).
  19. [19]C. Fang, Y. Chen, H.-Y. Kee, and L. Fu, Phys. Rev. B 92, 081201 (2015).
  20. [20]B. Yan and C. Felser, Annual Review of Condensed Matter Physics 8, 337 (2017).
  21. [21]N. P. Armitage, E. J. Mele, and A. Vishwanath, Rev. Mod. Phys. 90, 015001 (2018).
  22. [22]A. A. Burkov, Annual Review of Condensed Matter Physics 9, 359 (2018).
  23. [23]J. E. Moore, Y. Ran, and X.-G. Wen, Phys. Rev. Lett. 101, 186805 (2008).
  24. [24]D.-L. Deng, S.-T. Wang, C. Shen, and L.-M. Duan, Phys. Rev. B 88, 201105 (2013).
  25. [25]R. Kennedy, Phys. Rev. B 94, 035137 (2016).
  26. [26]G. Bian, T.-R. Chang, H. Zheng, S. Velury, S.-Y. Xu, T. Neupert, C.-K. Chiu, S.-M. Huang, D. S. Sanchez, I. Belopolski, N. Alidoust, P.-J. Chen, G. Chang, A. Bansil, H.-T. Jeng, H. Lin, and M. Z. Hasan, Phys. Rev. B 93, 121113 (2016).
  27. [27]H. M. Price, O. Zilberberg, T. Ozawa, I. Carusotto, and N. Goldman, Phys. Rev. Lett. 115, 195303 (2015).
  28. [28]B. Lian and S.-C. Zhang, Phys. Rev. B 94, 041105 (2016).
  29. [29]I. Petrides, H. M. Price, and O. Zilberberg, Phys. Rev. B 98, 125431 (2018).
  30. [30]L. Pontrjagin, Rec. Math. [Mat. Sbornik] N.S. 9, 331 (1941).
  31. [31]D. Auckly and L. Kapitanski, Communications in Mathematical Physics 256, 611 (2005).
  32. [32]D. DeTurck, H. Gluck, R. Komendarczyk, P. Melvin, C. Shonkwiler, and D. S. Vela-Vick, Journal of Mathematical Physics 54, 013515 (2013).
  33. [33]M. Kobayashi and M. Nitta, Nuclear Physics B 876, 605
  34. [34]F. N. Ünal, A. Eckardt, and R.-J. Slager, Phys. Rev. Res. 1, 022003 (2019).
  35. [35]See Supplemental Material.
  36. [36]L. W. T. Raoult Bott, Differential forms in algebraic topology (Springer, 1982).
  37. [37]Y. Wang, A. C. Tyner, and P. Goswami, Fundamentals of crystalline hopf insulators (2023), arXiv:2301.08244 [cond-mat.mes-hall].arxiv.org/abs/2301.08244
  38. [38]We note that a more generic version of this problem will have a gapless regime of separating the two phases rather than a single quantum critical point.
  39. [39]T. Ozawa and H. M. Price, Nature Reviews Physics 1, 349 (2019).
  40. [40]M. Lohse, C. Schweizer, H. M. Price, O. Zilberberg, and I. Bloch, Nature 553, 55 (2018).
  41. [41]O. Zilberberg, S. Huang, J. Guglielmon, M. Wang, K. P. Chen, Y. E. Kraus, and M. C. Rechtsman, Nature 553, 59 (2018).
  42. [42]Z. Davoyan, W. J. Jankowski, A. Bouhon, and R.-J. Slager, arXiv e-prints, arXiv:2308.15555 (2023), arXiv:2308.15555 [cond-mat.mes-hall].DOI
  43. [43]A. Sekine and K. Nomura, Journal of Applied Physics 129, 141101 (2021).

Paper details

vlnc:W_f74my7au
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