1: % pacs number
2: % consistent use of acronyms, like QSH, QH, TR etc.
3:
4: \documentclass[prl,amssymb,twocolumn,showpacs]{revtex4}
5:
6: \usepackage{graphicx}% Include figure files
7:
8: \begin{document}
9: \title{Topological Superconductivity and Superfluidity}
10: \author{Xiao-Liang Qi, Taylor L. Hughes, Srinivas Raghu and Shou-Cheng Zhang}
11:
12: \affiliation{Department of Physics, McCullough Building, Stanford
13: University, Stanford, CA 94305-4045}
14:
15: \pacs{74.20.Rp, 73.43.-f, 67.30.he, 74.45.+c}
16: \begin{abstract}
17: We construct time reversal invariant topological superconductors and
18: superfluids in two and three dimensions which are analogous to the
19: recently discovered quantum spin Hall and three-d $Z_2$ topological
20: insulators respectively. These states have a full pairing gap in the
21: bulk, gapless counter-propagating Majorana states at the boundary,
22: and a pair of Majorana zero modes associated with each vortex. We
23: show that the time reversal symmetry naturally emerges as a
24: supersymmetry, which changes the parity of the fermion number
25: associated with each time-reversal invariant vortex. In the presence
26: of external T-breaking fields, non-local topological correlation is
27: established among these fields, which is an experimentally
28: observable manifestation of the emergent supersymmetry.
29: \end{abstract}
30:
31: \maketitle
32:
33: The search for topological states of quantum matter has become an
34: active and exciting pursuit in condensed matter physics. The quantum
35: Hall (QH) effect \cite{prange1990} provides the first example of a
36: topologically non-trivial state of matter, where the quantized Hall
37: conductance is a topological invariant\cite{thouless1982}. Recently,
38: the quantum spin Hall (QSH) state\cite{kane2005a,bernevig2006a} has
39: been theoretically predicted \cite{bernevig2006b} and experimentally
40: observed in HgTe quantum well systems\cite{koenig2007}. The time
41: reversal invariant (TRI) QSH state is characterized by a bulk gap, a
42: $Z_2$ topological number \cite{kane2005b}, and gapless helical edge
43: states, where time-reversed partners
44: counter-propagate\cite{wu2006,xu2006}.
45:
46: Chiral superconductors in a time reversal symmetry breaking (TRB)
47: $(p_x+ip_y)$ pairing state in 2d have a sharp topological
48: distinction between the strong and weak pairing regimes
49: \cite{read2000}. In the weak pairing regime, the system has a full
50: bulk gap and gapless chiral Majorana states at the edge, which are
51: topologically protected. Moreover, a Majorana zero mode is trapped
52: in each vortex core\cite{read2000}, which leads to a ground state
53: degeneracy of $2^{n-1}$ in the presence of $2n$ vortices. When the
54: vortices wind around each other a non-Abelian Berry phase is
55: generated in the $2^{n-1}$ dimensional ground state manifold, which
56: implies non-Abelian statistics for the vortices\cite{ivanov2001}.
57: Chiral superconductors are analogous to the QH state--- they both
58: break time reversal (TR) and have chiral edge states with linear
59: dispersion. However, the edge states of a chiral superconductor have
60: only half the degrees of freedom compared to the QH state, since the
61: negative energy quasi-particle operators on the edge of a chiral
62: superconductor describe the same excitations as the positive energy
63: ones.
64:
65: Given the analogy between the chiral superconducting state and the
66: QH state, and with the recent discovery of the TRI QSH state, it
67: is natural to generalize the chiral pairing state to the helical
68: pairing state, where fermions with up spins are paired in the
69: $(p_x+ip_y)$ state, while fermions with down spins are paired in
70: the $(p_x-ip_y)$ state. Such a TRI state have a full gap in the
71: bulk, and counter-propagating helical Majorana states at the edge
72: (in contrast, the edge states of the TRI topological insulator are
73: helical Dirac fermions). Just as in the case of the QSH state, a
74: mass term for an odd number of pairs of helical Majorana states is
75: forbidden by TR symmetry, and therefore, a topologically protected
76: superconducting or superfluid state can exist in the presence of
77: time-reversal symmetry. Recently, a $Z_2$ classification of the
78: topological superconductor has been discussed in Refs
79: \cite{roy2006c,roy2008,schnyder2008}, by noting the similarity
80: between the Bogoliubov-de Gennes (BdG) superconductor Hamiltonian
81: and the QSH insulator Hamiltonian. The four types of topological
82: states of matter discussed here are summarized in Fig.
83: \ref{edgedispersion}. In this work, we give a $Z_2$ classification
84: of both the 2D and 3D cases which has a profound physical
85: implication. In two dimensions, we show that a time-reversal
86: invariant topological defect of a $Z_2$ non-trivial superconductor
87: carries a Kramers' pair of Majorana fermions. Let $N_F$ be the
88: operator which measures the number of fermions of a general
89: system, then the fermion-number parity operator is given by
90: $(-1)^{N_F}$. This operator is also referred to as the Witten
91: index\cite{witten1982}, which plays a crucial role in
92: supersymmetric theories. We prove the remarkable fact that in the
93: presence of a topological defect, the TR operator $\cal T$ changes
94: the fermion number parity, ${\cal T}^{-1} (-1)^{N_F} {\cal T} = -
95: (-1)^{N_F}$ locally around the defect in the $Z_2$ non-trivial
96: state, while it preserves the fermion number parity, ${\cal
97: T}^{-1} (-1)^{N_F} {\cal T} = (-1)^{N_F}$, in the $Z_2$ trivial
98: state. This fact gives a precise definition of the $Z_2$
99: topological classification of any TRI superconductor state and is
100: generally valid in the presence of interactions and disorder. A
101: supersymmetric operation can be defined as an operation which
102: changes the fermion number parity; therefore, in this precise
103: sense, we show that the TR symmetry emerges as a supersymmetry in
104: topological superconductors. Though supersymmetry has been studied
105: extensively in high energy physics, it has not yet been observed
106: in Nature. Our proposal offers the opportunity to experimentally
107: observe supersymmetry in condensed matter systems without any fine
108: tuning of microscopic parameters. The physical consequences of
109: such a supersymmetry is also studied.
110:
111: \begin{figure}[t!]
112: \begin{center}
113: \includegraphics[width=3in] {topdisp.eps}
114: \end{center}
115: \caption{(Top row) Schematic comparison of $2d$ chiral
116: superconductor and the QH state. In both systems, TR symmetry is
117: broken and the edge states carry a definite chirality. (Bottom row)
118: Schematic comparison of $2d$ TRI topological superconductor and the
119: QSH insulator. Both systems preserve TR symmetry and have a helical
120: pair of edge states, where opposite spin states counter-propagate.
121: The dashed lines show that the edge states of the superconductors
122: are Majorana fermions so that the $E<0$ part of the quasi-particle
123: spectra are redundant. In terms of the edge state degrees of
124: freedom, we have ${\rm (QSH)}={\rm (QH)}^2={\rm (Helical~SC)}^2={\rm
125: (Chiral~SC)}^4$.
126: } \label{edgedispersion}
127: \end{figure}
128:
129: As the starting point, we consider a TRI $p$-wave superconductor
130: with spin triplet pairing, which has the following $4\times 4$ BdG
131: Hamiltonian:
132: \begin{eqnarray}
133: H=\frac12\int d^2x\Psi^\dagger(x)
134: \left(\begin{array}{cc}\epsilon_{\bf
135: p}\mathbb{I}&i\sigma_2\sigma_\alpha\Delta^{\alpha
136: j}p_j\\h.c.&-\epsilon_{\bf
137: p}\mathbb{I}\end{array}\right)\Psi(x)\label{Hdouble}
138: \end{eqnarray}
139: with $\Psi(x)= \left(c_\uparrow(x),c_\downarrow(x),
140: c_\uparrow^\dagger(x),c_\downarrow^\dagger(x)\right)^T$,
141: $\epsilon_{\bf p}={\bf p}^2/2m-\mu$ the kinetic energy and
142: chemical potential terms and $h.c.\equiv
143: (i\sigma_2\sigma_\alpha\Delta^{\alpha j}p_j)^\dagger$. The TR
144: transformation is defined as $c_\uparrow\rightarrow
145: c_\downarrow,~c_\downarrow\rightarrow -c_\uparrow$. It can be
146: shown that the Hamiltonian (\ref{Hdouble}) is time-reversal
147: invariant if $\Delta_{\alpha j}$ is a real matrix. To show the
148: existence of a topological state, consider the TRI mean-field
149: ansatz $\Delta^{\alpha 1}=\Delta(1,0,0),~\Delta^{\alpha
150: 2}=\Delta(0,1,0)$. For such an ansatz the Hamiltonian
151: (\ref{Hdouble}) is block diagonal with only equal spin pairing:
152: \begin{eqnarray}
153: H=\frac12\int
154: d^2x\tilde{\Psi}^\dagger\left(\begin{array}{cccc}\epsilon_{\bf
155: p}&\Delta p_+&&\\
156: \Delta p_-&-\epsilon_{\bf p}&&\\
157: &&\epsilon_{\bf p}&-\Delta p_-\\
158: &&-\Delta p_+&-\epsilon_{\bf
159: p}\end{array}\right)\tilde{\Psi}\label{Hdoublepip}
160: \end{eqnarray}
161: with $\tilde{\Psi}(x)\equiv
162: \left(c_\uparrow(x),c_\uparrow^\dagger(x),
163: c_\downarrow(x),c_\downarrow^\dagger(x)\right)^T$, and
164: $p_\pm=p_x\pm ip_y$. From this Hamiltonian we see that the spin up
165: (down) electrons form $p_x+ip_y$ ($p_x-ip_y$) Cooper pairs,
166: respectively. In the weak pairing phase with $\mu>0$, the
167: $(p_x+ip_y)$ chiral superconductor is known to have chiral
168: Majorana edge states propagating on each boundary, described by
169: the Hamiltonian $H_{\rm edge}=\sum_{k_y\geq
170: 0}v_Fk_y\psi_{-k_y}\psi_{k_y}$,%\label{Hpipedge}
171: where $\psi_{-k_y}=\psi_{k_y}^\dagger$ is the quasiparticle creation
172: operator \cite{read2000} and the boundary is taken to be parallel to
173: the $y$ direction. Thus we know that the edge states of the TRI
174: system described by Hamiltonian (\ref{Hdoublepip}) consist of spin
175: up and spin down quasi-particles with opposite chirality:
176: \begin{eqnarray}
177: H_{\rm edge}=\sum_{k_y\geq
178: 0}v_Fk_y\left(\psi_{-k_y\uparrow}\psi_{k_y\uparrow}-\psi_{-k_y\downarrow}\psi_{k_y\downarrow}\right).\label{Hdoubleedge}
179: \end{eqnarray}
180: The quasi-particle operators
181: $\psi_{k_y\uparrow},~\psi_{k_y\downarrow}$ can be expressed in terms
182: of the eigenstates of the BdG Hamiltonian as
183: \begin{eqnarray}
184: \psi_{k_y\uparrow}&=&\int
185: d^2x\left(u_{k_y}(x)c_{\uparrow}(x)+v_{k_y}(x)c_{\uparrow}^\dagger(x)\right)\nonumber\\
186: \psi_{k_y\downarrow}&=&\int
187: d^2x\left(u_{-k_y}^*(x)c_{\downarrow}(x)+v_{-k_y}^*(x)c_{\downarrow}^\dagger(x)\right)
188: \end{eqnarray}
189: from which the time-reversal transformation of the quasiparticle
190: operators can be determined to be
191: ${\cal{T}}^{-1}\psi_{k_y\uparrow}{\cal{T}}=\psi_{-k_y\downarrow},~{\cal{T}}^{-1}\psi_{k_y\downarrow}{\cal{T}}=-\psi_{-k_y\uparrow}$.
192: In other words, $(\psi_{k_y\uparrow},~\psi_{-k_y\downarrow})$
193: transforms as a Kramers' doublet, which forbids a gap in the edge
194: states due to mixing of the spin-up and spin-down modes when TR is
195: preserved. To see this explicitly, notice that the only
196: $k_y$-independent term that can be added to the edge Hamiltonian
197: (\ref{Hdoubleedge}) is
198: $im\sum_{k_y}\psi_{-k_y\uparrow}\psi_{k_y\downarrow}$ with $m\in
199: \mathbb{R}$. However, such a term is odd under TR, which implies
200: that any back scattering between the quasi-particles is forbidden by
201: TR symmetry. The discussion above is exactly parallel to the $Z_2$
202: topological characterization of the quantum spin Hall system. In
203: fact, the Hamiltonian (\ref{Hdoublepip}) has exactly the same form
204: as the four band effective Hamiltonian proposed in
205: Ref.\cite{bernevig2006b} to describe HgTe quantum wells with the QSH
206: effect. The edge states of the QSH insulators consist of an odd
207: number of Kramers' pairs, which remain gapless under any small
208: TR-invariant perturbation\cite{wu2006,xu2006}. A no-go theorem
209: states that such a ``helical liquid" with an odd number of Kramers'
210: pairs at the Fermi energy can not be realized in any bulk 1d system,
211: but can only appear as an edge theory of a 2d QSH
212: insulator\cite{wu2006}. Similarly, the edge state theory
213: (\ref{Hdoubleedge}) can be called a ``helical Majorana liquid",
214: which can only exist on the boundary of a $Z_2$ topological
215: superconductor. Once such a topological phase is established, it is
216: robust under any TRI perturbations.
217:
218: The Hamiltonian (\ref{Hdouble}) can be easily generalized to three
219: dimensions, in which case $\Delta^{\alpha j}$ becomes a $3\times
220: 3$ matrix with $\alpha=1,2,3$ and $j=x,y,z$. An example of such a
221: Hamiltonian is given by the well-known $^3$He BW phase, for which
222: the order parameter $\Delta^{\alpha j}$ is determined by an
223: orthogonal matrix $\Delta^{\alpha j}=\Delta u^{\alpha j}$,
224: $u\in{\rm SO(3)}$\cite{vollhardt1990}. Here and below we ignore
225: the dipole-dipole interaction term \cite{leggett1975} since it
226: does not affect any essential topological properties. By applying
227: a spin rotation, $\Delta^{\alpha j}$ can be diagonalized to
228: $\Delta^{\alpha j}=\Delta \delta^{\alpha j}$, in which case the
229: Hamiltonian (\ref{Hdouble}) has the same form as a 3d Dirac
230: Hamiltonian with momentum dependent mass $\epsilon({\bf p})={\bf
231: p}^2/2m-\mu$. We know that a band insulator described by the Dirac
232: Hamiltonian is a 3d $Z_2$ topological insulator for
233: $\mu>0$\cite{fu2007b,moore2007,roy2006b}, and has nontrivial
234: surface states. The corresponding superconductor Hamiltonian
235: describes a topological superconductor with 2d gapless Majorana
236: surface states. The surface theory can be written as
237: \begin{eqnarray}
238: H_{\rm surf}=\frac12\sum_{\bf k}v_F\psi_{-\bf
239: k}^T\left(\sigma_zk_x+\sigma_xk_y\right)\psi_{\bf k}
240: \end{eqnarray}
241: which remains gapless under any small TRI perturbation since the
242: only available mass term $m\sum_{\bf k}\psi_{-\bf
243: k}^T\sigma_y\psi_{\bf k}$ is time-reversal odd. We would like to
244: mention that the surface Andreev bound states in $^3$He-B phase
245: have been observed experimentally\cite{aoki2005}.
246:
247: To understand the physical consequences of the
248: nontrivial topology we study the TRI topological defects of the
249: topological superconductors. We start by considering the equal-spin
250: pairing system with BdG Hamiltonian (\ref{Hdoublepip}) in which spin
251: up and down electrons form $p_x+ip_y$ and $p_x-ip_y$ Cooper pairs,
252: respectively. A TRI topological defect can be defined as a vortex of
253: spin-up superfluid coexisting with an anti-vortex of spin-down
254: superfluid at the same position. In the generic Hamiltonian
255: (\ref{Hdouble}), such a vortex configuration is written as
256: $\Delta^{\alpha j}=[\exp\left(i\sigma_2\theta({\bf
257: r-r_0})\right)]^{\alpha j},~\alpha=1,2$ and $\Delta^{3j}=0$, where
258: $\theta({\bf r-r_0})$ is the angle of ${\bf r}$ with respect to the
259: vortex position ${\bf r}_0$. Since in the vortex core of a weak
260: pairing $p_x+ip_y$ superconductor there is a single Majorana zero
261: mode\cite{read2000,stone2006}, one immediately knows that a pair a
262: Majorana zero modes exist in the vortex core we study here. In terms
263: of the electron operators, the two Majorana fermion operators can be
264: written as
265: \begin{eqnarray}
266: \gamma_\uparrow&=&\int
267: d^2x\left(u_0(x)c_\uparrow(x)+u_0^*(x)c_\uparrow^\dagger(x)\right)\nonumber\\
268: \gamma_\downarrow&=&\int
269: d^2x\left(u_0^*(x)c_\uparrow(x)+u_0(x)c_\uparrow^\dagger(x)\right)
270: \end{eqnarray}
271: where we have used the fact that the spin-down zero mode wave
272: function can be obtained from the time-reversal transformation of
273: the spin-up one. The Majorana operators satisfy the
274: anti-commutation relation
275: $\left\{\gamma_\alpha,\gamma_\beta\right\}=2\delta_{\alpha\beta}$.
276: The TR transformation of the Majorana fermions is
277: \begin{eqnarray}
278: {\cal{T}}^{-1}\gamma_\uparrow
279: {\cal{T}}=\gamma_\downarrow,~{\cal{T}}^{-1}\gamma_\downarrow
280: {\cal{T}}=-\gamma_\uparrow .\label{TofMajorana}
281: \end{eqnarray}
282: Similar to the case of the edge states studied earlier, the
283: Majorana zero modes are robust under any small TRI perturbation,
284: since the only possible term $im\gamma_\uparrow\gamma_\downarrow$
285: which can lift the zero modes to finite energy is TR odd, {\it
286: i.e.},
287: ${\cal{T}}^{-1}i\gamma_\uparrow\gamma_{\downarrow}{\cal{T}}=-i\gamma_\uparrow\gamma_{\downarrow}$.
288:
289: The properties of such a topological defect appear identical to
290: that of a $\pi$-flux tube threading into a TRI topological
291: insulator\cite{qi2008,ran2008}, where a Kramers' pair of complex
292: fermions are trapped by the flux tube. However, there is an
293: essential difference. From the two Majorana zero modes
294: $\gamma_{\uparrow},\gamma_\downarrow$ a complex fermion operator
295: can be defined as
296: $a=\left(\gamma_\uparrow+i\gamma_\downarrow\right)/2$, which
297: satisfies the fermion anticommutation relation
298: $\left\{a,a^\dagger\right\}=1$. Since
299: $\gamma_\uparrow,\gamma_\downarrow$ are zero modes, we obtain
300: $\left[a,H\right]=0$ which implies that $a$ is the annihilation
301: operator of a zero-energy quasiparticle. Consequently, the ground
302: state of the system is at least two-fold degenerate, with two
303: states $|G_0\rangle$ and $|G_1\rangle=a^\dagger|G_0\rangle$
304: containing $0$ and $1$ $a$-fermions. Since $a^\dagger
305: a=\left(1+i\gamma_\uparrow \gamma_\downarrow\right)/2$, the states
306: $\left|G_{0(1)}\right\rangle$ are eigenstates of
307: $i\gamma_\uparrow\gamma_\downarrow$ with eigenvalues $-1(+1)$,
308: respectively. Thus from the oddness of
309: $i\gamma_\uparrow\gamma_\downarrow$ under TR we know that
310: $\left|G_0\right\rangle$ and $\left|G_1\right\rangle$ are
311: time-reversed partners. Note that superconductivity breaks the
312: charge $U(1)$ symmetry to $Z_2$, meaning that the fermion number
313: parity operator $(-1)^{N_F}$ is conserved. Thus, all the
314: eigenstates of the Hamiltonian can be classified by the value of
315: $(-1)^{N_F}$. If, say, $\left|G_0\right\rangle$ is a state with
316: $(-1)^{N_F}=1$, then
317: $\left|G_1\right\rangle=a^\dagger\left|G_0\right\rangle$ must
318: satisfy $(-1)^{N_F}=-1$. Since $\left|G_0\right\rangle$ and
319: $\left|G_1\right\rangle$ are time-reversal partners, we know that
320: in the Hilbert space of the zero-energy states the TR
321: transformation changes the fermion number parity:
322: \begin{eqnarray}
323: {\cal{T}}^{-1}(-1)^{N_F}{\cal{T}}=-(-1)^{N_F}.\label{TofNF}
324: \end{eqnarray}
325:
326: Eq. (\ref{TofNF}) is the central result of this paper. At a first
327: glance it seems contradict the fundamental fact that the electron
328: number of the whole system is invariant under TR. Such a paradox
329: is resolved by noticing that there are always an even number of
330: topological defects in a closed system without boundary. Under the
331: TR transformation, the fermion number parity around each vortex
332: core is odd, but the total fermion number parity remains even as
333: expected. Once the anomalous transformation property (\ref{TofNF})
334: is established for a topological defect in a TRI superconductor,
335: it is robust under any TRI perturbation as long as the bulk
336: quasiparticle gap remains finite and other topological defects are
337: far away. Thus Eq. (\ref{TofNF}) is a generic definition of TRI
338: topological superconductors:
339:
340: \begin{itemize}\item{\textsc{Definition I.} A two-dimensional
341: TRI superconductor is $Z_2$ nontrivial if and only if fermion
342: number parity around a TRI topological defect is odd under
343: TR.}\end{itemize}
344:
345: A transformation changing fermion number by an odd number is a
346: ``supersymmetry"; thus, the TR symmetry emerges as a discrete
347: supersymmetry for each TRI topological defect. The same analysis
348: applies to the edge theory (\ref{Hdoubleedge}), which shows that
349: in the 1d helical Majorana liquid is a theory with TR symmetry as
350: a discrete supersymmetry.
351:
352: All the conclusions above can be generalized to 3d topological
353: superconductors. In the $^3$He BW phase the Goldstone manifold of
354: the order parameter is $\Delta^{\alpha j}=\Delta u^{\alpha j}\in
355: SO(3)\times U(1)$\cite{vollhardt1990,salomaa1987}. A time-reversal
356: invariant configuration satisfies $\Delta^{\alpha j}\in \mathbb{R}$,
357: which restricts the order parameter to $SO(3)$. Since
358: $\Pi_1(SO(3))=Z_2$, the TRI topological defects are 1d ``vortex"
359: rings. By solving the BdG equations in the presence of such vortex
360: rings, it can be shown that there are linearly dispersing
361: quasiparticles propagating on each vortex ring, similar to the edge
362: states of the 2d topological superconductor. However, for a ring
363: with finite length the quasi-particle spectrum is discrete.
364: Specifically, there may or may not be a pair of Majorana modes at
365: exactly zero energy. The existence of the Majorana zero modes on the
366: vortex rings turns out to be a topological property determined by
367: the linking number between different vortex rings. Due to the length
368: constraints of the present paper, we will write
369: our conclusion and leave the details for a separate work: {\em
370: There are a pair of Majorana fermion zero modes confined on a
371: vortex ring if and only if the ring is linked to an odd number of
372: other vortex rings.} Such a condition is shown in Fig.
373: \ref{fig:vortex3d}. Consequently, the generalization of Definition
374: I to 3d is:
375:
376: \begin{itemize}\item{\textsc{Definition II.} A 3d TRI superconductor is $Z_2$ nontrivial if and only if
377: the fermion number parity around one of the two mutually-linked
378: TRI vortex rings is odd under TR.}\end{itemize}
379:
380:
381: \begin{figure}[htpb]
382: \begin{center}
383: \includegraphics[width=2.8in]{vortex3d.eps}
384: \end{center}
385: \caption{Illustration of a 3d TRI topological superconductor with
386: two TRI vortex rings which are (a) linked or (b) unlinked. The $E-k_\parallel$ dispersion relations show schematically the
387: quasiparticle levels confined on the red vortex
388: ring in both cases. ``$\circ$" and ``$\times$" stand for the quasiparticle levels that are Kramers' partners of each other.
389: Only case (a) has a pair of Majorana zero modes located on each vortex ring.}
390: \label{fig:vortex3d}
391: \end{figure}
392:
393: \begin{figure}[htpb]
394: \begin{center}
395: \includegraphics[width=3in]{vortex2d.eps}
396: \end{center}
397: \caption{(a) Illustration of a 2d TRI topological superconductor with four TRI topological
398: defects coupled to a TR-breaking field. The arrows show the sign of the TR-breaking field ${\rm sgn}(M({\bf r}_s))$
399: at each topological defect. In the two configurations shown, only the field around vortex 1 is flipped,
400: leading to an opposite fermion number parity in the corresponding ground state $|G\rangle$ and $|G'\rangle$ (see text).
401: (b) Illustration showing the flow of the energy levels when the upper configuration in figure (a) is deformed to
402: the lower one. The flip of the TR-breaking field $M({\bf r}_1)$ leads to a level crossing at $M({\bf r}_1)=0$, where
403: the fermion number
404: parity in the ground state changes sign. }
405: \label{fig:vortex2d}
406: \end{figure}
407:
408: Besides providing a
409: generic definition of the $Z_2$ topological superconductors, such
410: an emergent supersymmetry also leads to physical predictions.
411: Consider the 2d topological superconductor coupled to a weak
412: TR-breaking field $M({\bf r})$, which is classical but can have
413: thermal fluctuations. This situation can be realized in an
414: isolated superconductor with vortices pinned to quenched weak
415: magnetic impurities. The $n$-point correlation function of $M({\bf
416: r})$ can be obtained by
417: \begin{eqnarray}
418: \left\langle{\prod_{s=1}^nM({\bf r}_s)}\right\rangle\equiv\int
419: \frac{D[M({\bf r})]}Z\prod_{s=1}^nM({\bf r}_s){\rm
420: Tr}\left(e^{-\beta H[M]}\right)_{\rm even}\label{Mcorrelation}
421: \end{eqnarray}
422: in which the trace is restricted to states with an even number of
423: fermions. For a closed system with $N$ vortices, the leading order
424: effect of the TR-breaking field is to lift the degeneracy between
425: the two Majorana fermions in each vortex core. Consequently, the
426: perturbed Hamiltonian $H[M({\bf r})]$ to first order can be
427: written as
428: \begin{eqnarray}
429: H[M({\bf r})]=\sum_{s=1}^{N}iM({\bf
430: r}_s)a_s\gamma_{s\uparrow}\gamma_{s\downarrow}\label{HofM}
431: \end{eqnarray}
432: in which $\gamma_{s\uparrow(\downarrow)}$ are the Majorana fermion
433: operators, and $a_s\in \mathbb{R}$ depend on the details of the
434: perturbation. The important fact is that the mass term induced is
435: linear in $M({\bf r})$ at the defect position ${\bf r}_s$, since
436: $i\gamma_{s\uparrow}\gamma_{s\downarrow}$ is TR odd.
437:
438: Since the superconductor has a full gap, naively one would expect
439: all the correlations of $M({\bf r})$ field to be short ranged.
440: However, for a system with $N$ topological defects the $N$-point
441: correlation function has a long range order when ${\bf
442: r}_s,s=1,2,..N$ are chosen to be the coordinates of the
443: topological defects. In other words, the correlation function
444: \begin{eqnarray}
445: \lim_{|{\bf r}_i-{\bf r}_j|\rightarrow \infty,~\forall
446: i,j}\left\langle{\prod_{s=1}^NM({\bf r}_s)}\right\rangle\neq
447: 0,\label{Mtopcorrelation}
448: \end{eqnarray}
449: though all the $n$ point correlations in Eq. (\ref{Mcorrelation})
450: with $n<N$ remain short ranged. Physically, such a non-local
451: correlation can be understood by comparing two states $|G\rangle$
452: and $|G'\rangle$, which are the ground states of the systems with
453: the field configurations $\mathcal{M}\equiv(M({\bf r}_1),M({\bf
454: r}_2),...,M({\bf r}_N))$ and $\mathcal{M}'\equiv(-M({\bf
455: r}_1),M({\bf r}_2),...,M({\bf r}_N))$, respectively. From
456: Hamiltonian (\ref{HofM}) it can be seen that $|G\rangle$ and
457: $|G'\rangle$ have opposite fermion number parity, since the fermion
458: number parity around the first topological defect
459: $i\gamma_{1\uparrow}\gamma_{1\downarrow}$ is reversed while that of
460: all the other topological defects remains invariant, as shown in
461: Fig. \ref{fig:vortex2d}. Without loss of generality, we can assume
462: $(-1)^{N_F}$ is even for $|G\rangle$ and odd for $|G'\rangle$. Since
463: the whole system is required to have an even number of fermions, the
464: lowest energy state in the Hilbert space for the field configuration
465: $\mathcal{M}'$ is not $|G'\rangle$, but the lowest quasiparticle
466: excitation $a_{\rm min}^\dagger|G'\rangle$. Thus, the two field
467: configurations $\mathcal{M}$ and $\mathcal{M}'$ have different free
468: energies, which leads to the non-vanishing correlation function in
469: Eq. (\ref{Mtopcorrelation}). Even when the topological defects are
470: arbitrarily far away, the energy difference between the two
471: configurations remains finite, which shows the non-local topological
472: correlation. Similar non-local correlations can also be obtained for
473: a 3d TRI topological superconductor with linked vortex rings.
474:
475: Acknowledgement.---We acknowledge helpful discussions with S. B.
476: Chung, A. L. Fetter, L. Fu, R. Roy and S. Ryu. This work is
477: supported by the NSF under grant numbers DMR-0342832, the US
478: Department of Energy, Office of Basic Energy Sciences under contract
479: DE-AC03-76SF00515, and the Stanford Institute for Theretical Physics (S.R.).
480:
481: \bibliography{TSC}
482: \end{document}
483: