1: \documentclass[aps,twocolumn,prl,showpacs]{revtex4}
2: \usepackage{epsfig}
3: \usepackage{graphicx}
4: \usepackage{amsfonts}
5: \usepackage[figuresright]{rotating}
6: \usepackage{amssymb}
7: \usepackage{amsmath}
8: %\renewcommand{\baselinestretch}{2.0}
9:
10: \def\avg#1{\langle#1\rangle}
11: \def\Re{\rm{Re}}
12: \def\Im{\rm{Im}}
13: \def\be{\begin{equation}} \def\ee{\end{equation}}
14: \def\bea{\begin{eqnarray}} \def\eea{\end{eqnarray}}
15: \def\PRB{Phys. Rev. B}
16: \def\PRA{Phys. Rev. A}
17: \def\PRL{Phys. Rev. Lett.}
18: \def\nn{\nonumber}
19: \def\pp{\parallel}
20:
21: \newcommand{\ket}[1]{|#1 \rangle}
22: \newcommand{\bra}[1]{\langle #1 |}
23:
24: %\documentstyle[aps,epsfig,preprint]{revtex}
25: \begin{document}
26:
27:
28: \title{Orbital analogue of quantum anomalous Hall effect in $p$-band
29: systems}
30: \author{Congjun Wu}
31: \affiliation{Department of Physics, University of California, San Diego,
32: CA 92093}
33:
34: \begin{abstract}
35: We investigate the topological insulating states of the $p$-band
36: systems in optical lattices induced by the onsite orbital angular
37: momentum polarization, which exhibit gapless edge modes in
38: the absence of Landau levels.
39: This effect arises from the energy level splitting between the
40: onsite $p_x+ip_y$ and $p_x-ip_y$ orbitals by rotating each optical
41: lattice site around its own center.
42: At large rotation angular velocities, this model naturally reduces
43: to two copies of Haldane's quantum Hall model without Landau levels.
44: The distribution of Berry curvature in the momentum space and the
45: quantized Chern numbers are calculated.
46: The experimental realization is also discussed.
47: \end{abstract}
48: \pacs{03.75.Ss, 05.50.+q, 73.43.-f, 73.43.Nq}
49: \maketitle
50:
51: The integer quantum Hall (QH) effect has generate tremendous research
52: interests for several decades.
53: The precise quantization of the Hall conductance is due to the
54: topologically non-trivial band structure characterized by the
55: Thouless-Kohmoto-Nightingale-den Nijs (TKNN) number, or the Chern
56: number \cite{thouless1982, kohmoto1985}.
57: The origin of the QH effect has also deep connections to the parity anomaly of
58: 2D Dirac fermions \cite{jackiw1984, fradkin1986, haldane1988}.
59: Although breaking time-reversal (TR) symmetry is required, Landau levels
60: (LL) are not necessary for the QH effect.
61: For example, Haldane \cite{haldane1988} constructed a QH model with
62: average zero flux per unit cell but with complex-valued hopping
63: integrals. %which has non-zero Chern number and exhibits gapless edge states.
64: Recently, QH insulators have been generalized to the topological quantum
65: spin Hall (QSH) insulators which keep time-reversal (TR) symmetry
66: and are characterized by a Z$_2$ topological number \cite{
67: bernevig2006a, qi2008a, kane2005, hirsch1989, sheng2006,
68: moore2007, roy2006}.
69: %Interaction effects in QSH insulators are investigated in
70: %\cite{wu2006, xu2006, lee2008, raghu2008}.
71: Excitingly, experimental evidence for the QSH insulating
72: states has been found \cite{konig2007, hsieh2008}.
73:
74: Anomalous Hall (AH) effect describes the dependence of the Hall
75: current on the spin magnetization not the external magnetic field,
76: whose mechanism has been debated for a long time,
77: including the anomalous velocity from the interband matrix element
78: \cite{karplus1954}, screw scattering\cite{smit1958}, and side jump
79: \cite{berger1970}.
80: Recently, a new perspective on the AH effect has been developed from the
81: topological Berry curvature of the band structure which is a combined
82: effect from spin-orbit coupling and spin polarization \cite {jungwirth2002,
83: nagaosa2006}.
84: %This mechanism describes the AH effect well in several materials
85: %such as SrRuO$_3$, (In,Mn)As and (Ga,Mn)As systems.
86: Its quantum version, topological insulators arising from spin
87: magnetization has been proposed and investigated in semiconductor
88: systems \cite{qi2006a, liu2008, onoda2003a}.
89:
90: The current development of cold atom physics has provided another
91: new opportunity to investigate the QH effect.
92: Several methods have been proposed including globally rotating the
93: trap or optical lattice, or introducing effective gauge potential generated by
94: laser beams \cite{ho2000, scarola2007, shao2008, umucalilar2008, zhu2006}.
95: However, the rotating angular velocity cannot be large enough otherwise
96: the centrifugal potential will throw atoms away.
97: It is also difficult to make the light induced gauge potential strong
98: in a large region.
99: %Recently, a proposal to simulate Haldane's model
100: %has been suggested \cite{shao2008} by superpose a periodic light
101: %-induced gauge potential with a honeycomb optical lattice.
102: %However, it needs a high precision of alignment which might be difficult.
103:
104: In this article, we propose an orbital analogue of the quantum anomalous
105: Hall (QAH) effect in solid state systems, i.e., the QAH effect
106: arising from orbital angular momentum polarization without LLs.
107: This can be achieved by rotating each optical site around its own center
108: which is an experimentally feasible technique \cite{gemelke2007}.
109: The lift of the degeneracy between $p_x\pm i p_y$ orbitals
110: gives rise to non-trivial topological band structures,
111: and provides a natural way to realize Haldane's model.
112: Increasing rotation angular velocity induces topological phase
113: transition by changing the Chern number of the band structure.
114: We also consider the QH effect arising with LLs in
115: such systems.
116:
117: % an overall rotation in the
118: %$p$-orbital system in the honeycomb lattices.
119: %which has a combined effect of orbitals and Landau levels.
120:
121:
122:
123: The experiment to rotate each site in the lattice around its own site
124: center has been performed by Gemelke {\it et al.} \cite{gemelke2007}.
125: Electro-optic phase modulators are applied to the laser beams forming
126: the lattice, which results in a periodical overall translation of
127: the lattice at a radio-frequency $\omega_{RF}$ but without the internal
128: lattice distortions.
129: $\omega_{RF}$ is hundreds of times larger than the harmonic frequency
130: $\omega_L$ of each site, thus atoms only feel an averaged potential with a
131: small distortion along the oscillation axis.
132: This axis can be controlled to rotate at an audio frequency $\Omega_z$,
133: which induces the rotation of each site around its own center at
134: such a frequency.
135: $\Omega_z$ can be much larger than the overall parabolic trapping
136: frequency and reach a few kilo-Hertz.
137: This technique has been applied in the triangular lattice described in
138: Ref. \cite{gemelke2007}.
139:
140: Let us consider to apply this technique to the honeycomb lattice
141: which has been constructed quite some ago \cite{grynberg1993}.
142: We study the $p_{x,y}$-orbital band filled with spinless fermions
143: described in Ref. \cite{wu2007,wu2007a,wu2008} with the new
144: ingredient of rotation as
145: \bea
146: \label{eq:ham0}
147: H_0&=&t_{\pp} \sum_{ \vec{r} \in A} \big\{ p^\dagger_{\vec{r},i}
148: p_{\vec{r}+a\hat{e}_i,i}+h.c. \big\} -\mu\sum_{\vec{r}} n_{\vec{r}},
149: \nonumber \\
150: H_L&=& i\Omega_z \sum_{\vec{r}} \big\{ p^\dagger_{\vec{r},x} p_{\vec{r},y}
151: - p^\dagger_{\vec{r},y} p_{\vec{r},x} \big\},
152: \eea
153: where $\hat{e}_{1,2}=\pm\frac{\sqrt{3}}{2}\hat{e}_x+\frac{1}{2}\hat{e}_y$
154: and $\hat{e}_3=-\hat{e}_y$ are the unit vectors pointing from a
155: site in the $A$-sublattice to its three neighbors in the $B$-sublattice;
156: $p_i\equiv (p_x\hat{e}_x+p_y\hat{e}_y)\cdot \hat{e}_i~(i=1\sim 3)$
157: are the projections of the $p$-orbitals along the $\hat e_i$ direction;
158: $\mu$ is the chemical potential;
159: $a$ is the nearest neighbor bond length.
160: Since there is no the overall lattice rotation, the vector potential
161: due to the Coriolis force and the centrifugal potential across the
162: entire lattices do not appear.
163: The effect is to break the degeneracy between $p_x\pm i p_y$
164: as described by $H_L$.
165: $\Omega_z$ can easily reach the order of the recoil energy $E_R$,
166: and $t_\pp$ can be tuned one order smaller than E$_R$ \cite{wu2007a},
167: thus we have a large flexibility of tuning $\Omega_z/t_\pp$.
168:
169:
170: The band structure of Eq. \ref{eq:ham0} is presented as follows.
171: Under the chiral transformation $P$, i.e., $p_{r_A, x,y} \rightarrow
172: p_{r_A, x,y}$, $p_{r_B, x,y}\rightarrow -p_{r_B, x,y}$, combined by
173: the time-reversal transformation $T$, Eq. \ref{eq:ham0}
174: transforms as $ (TP)^{-1} (H_0+H_L) (TP)= -(H_0+H_L)$, thus its spectra are
175: symmetric respect to the zero energy.
176: At $\Omega_z=0$ it exhibits two dispersive bands touching at
177: Dirac cones located at $K_{1,2}=(\pm \frac{4\pi}{3\sqrt 3 a}, 0)$
178: and other two flat bands \cite{wu2007,wu2007a}.
179: The dispersive bands touch the flat bands at the
180: Brillouin zone (BZ) center $K_0=(0,0)$.
181: We define the 4-component spinor as
182: $\psi(\vec k)=(p_{Ax}(\vec k),p_{Ay}(\vec k),
183: p_{Bx}(\vec k),p_{By}(\vec k))^T$, and
184: the two bases for the dispersive bands as
185: $\phi_1(\vec k)=\sqrt{\frac{2}{N_0}}\Big\{f_{12}(\vec k),\ \ \
186: \frac{1}{\sqrt 3}
187: (f_{23}(\vec k)-f_{31}(\vec k)),~ 0,~ 0\Big\}$, and
188: $\phi_2(\vec k)=\sqrt{\frac{2}{N_0}}
189: \Big\{0,~ 0,~ f^*_{12}(\vec k),\ \ \ \frac{1}{\sqrt 3}
190: (f^*_{23}(\vec k)-f^*_{31}(\vec k))\Big\},
191: $
192: where $f_{ij}=e^{i \vec k \cdot \hat e_i}
193: -e^{i \vec k \cdot \hat e_j}$ and $N_0(\vec k)$ is the
194: normalization factor.
195: At nonzero $\Omega_z$, gaps open between different bands as
196: depicted in Fig. \ref{fig:spectra} A, B, and C.
197: If $\Omega_z$ is small, the effective Hamiltonian close to
198: the Dirac points of $K_{1,2}$ can be written in the bases of
199: $\phi_1, \phi_2$ as
200: \bea
201: H_1(\vec k)=\left( \begin{array}{cc}
202: \frac{8\Omega_z}{\sqrt 3 N_0(\vec k)} \sum_i \sin \vec k \cdot \vec b_i
203: & -\frac{t_\pp}{2} \sum_i e^{-i \vec k \cdot \hat e_i}\\
204: -\frac{t_\pp}{2} \sum_i e^{i \vec k \cdot \hat e_i} &
205: \frac{-8\Omega_z}{\sqrt 3 N_0(\vec k)} \sum_i \sin \vec k \cdot \vec b_i
206: \end{array} \nonumber
207: \right ),
208: \label{eq:dirac1}
209: \eea
210: where $\vec b_i=\frac{1}{2}\epsilon_{ijk} (\hat e_j -\hat e_k)$ are the
211: vectors connecting the next nearest neighbors.
212: The Dirac cones become gapped with $\Delta= \Omega_z$
213: and the masses are of the opposite sign at $K_{1,2}$.
214: %This is similar to Haldane's construction of the topological
215: %insulator in the hexagonal lattice with the complex-valued next
216: %nearest neighbor hopping term.
217: %In comparison, only the nearest neighbor hopping is involved
218: %in Eq. \ref{eq:ham0}.
219: The bottom band is no longer flat at nonzero $\Omega_z$.
220: Its minimum at $K_0$ is pushed down by a value of $\frac{3}{2}
221: \Omega_z$ and that of the second band is pushed up by
222: $\frac{3}{2} \Omega_z$.
223: This opens a gap of $3\Omega_z$.
224: A similar analysis applies to the top and the third bands.
225:
226: \begin{figure}
227: \centering\epsfig{file=Hnycmb_rot1.eps,clip=1,width=0.45\linewidth,angle=0}
228: \centering\epsfig{file=Hnycmb_rot2.eps,clip=1,width=0.45\linewidth,angle=0}
229: \centering\epsfig{file=Hnycmb_rot3.eps,clip=1,width=0.45\linewidth,angle=0}
230: \centering\epsfig{file=haldane.eps,clip=1,width=0.52\linewidth,angle=0}
231: \caption{The band structure of Eq. \ref{eq:ham0} at $\Omega_z>0$ as
232: shown in A, B and C. Only the lower two bands are presented,
233: and the upper two are symmetric respect to the zero energy.
234: A) $\Omega_z/t_\pp=0.3$; B) $\Omega_z/t_\pp=1.5$ where a single
235: gapless Dirac cone appears; C) $\Omega_z/t_\pp=3$
236: where two massive Dirac cones appear at $K_{1,2}$ between the lower
237: two bands and also between the upper two bands;
238: D) The pattern of the induced next nearest neighbor hopping (complex-valued)
239: at $\Omega_z\gg t_\pp$, which is generated by the virtual hopping
240: between orbitals with opposite chirality.
241: }
242: \label{fig:spectra}
243: \end{figure}
244:
245:
246: As $\Omega_z$ approaches $\frac{3}{2}t_\pp$, the middle two bands at $K_0$
247: are pushed to zero from both up and below respectively, and form a single
248: gapless Dirac cone in the BZ as depicted in Fig. \ref{fig:spectra} B.
249: Let us define another two bases as
250: $\phi_1^\prime=\frac{1}{2}\{1, i, -1, -i\}$,
251: and $\phi_2^\prime=\frac{1}{2}\{1,-i, 1, -i\}$ for the Dirac cone
252: at $K_0$, and the Hamiltonian matrix becomes
253: \bea
254: H_2(\vec k)=\left( \begin{array}{cc}
255: -(\Omega_z-\frac{3}{2}t_\pp) & -\frac{3}{2} t_\pp (k_x+i k_y) \\
256: -\frac{3}{2} t_\pp (k_x-i k_y)& \Omega_z-\frac{3}{2}t_\pp
257: \end{array}
258: \right ).
259: \label{eq:dirac2}
260: \eea
261: We notice that a single Dirac cone of the chiral fermion is allowed in
262: the 2D bulk lattice systems, which
263: actually does not contradict to the fermion doubling theory
264: proved for 3D lattices \cite{nielsen1981}.
265:
266:
267: As $\Omega_z$ goes even larger, the lower and upper two pairs of
268: bands are projected into the single orbital bands of $p_x \pm i p_y$
269: respectively.
270: The lower two are described by the $p_x + i p_y$ orbital
271: with a nearest neighbor hopping of $\frac{t_\pp}{2}$.
272: Furthermore, a Haldane type next nearest neighbor hopping is generated
273: as depicted in Fig. \ref{fig:spectra} D:
274: one particle at site $A$ in the $p_x + i p_y$ orbital hops to the
275: high energy orbital of $p_x -i p_y$ at its nearest neighbor $B$,
276: and hops back into the $p_x +i p_y$ state at the next nearest
277: neighbor site $A^\prime$.
278: Along the directions indicted by arrows, this
279: hopping amplitude can be calculated from the second order
280: perturbation theory as $t_{nn}= t^2_\pp/(2\Omega_z) e^{i\frac{2}{3} \pi}.$
281: As pointed out in Ref. \cite{haldane1988}, this generates two
282: massive Dirac cones with gap $\Delta=\frac{9}{2} t_{nn}$
283: at $\vec K_{1,2}$ of masses with opposite signs.
284:
285: \begin{figure}
286: \centering\epsfig{file=curvature_1a.eps,clip=1,width=0.49\linewidth,angle=0}
287: \centering\epsfig{file=curvature_1b.eps,clip=1,width=0.49\linewidth,angle=0}
288: \centering\epsfig{file=curvature_2a.eps,clip=1,width=0.49\linewidth,angle=0}
289: \centering\epsfig{file=curvature_2b.eps,clip=1,width=0.49\linewidth,angle=0}
290: \centering\epsfig{file=curvature_3a.eps,clip=1,width=0.49\linewidth,angle=0}
291: \centering\epsfig{file=curvature_3b.eps,clip=1,width=0.49\linewidth,angle=0}
292: \caption{The distribution of Berry curvature $F_{xy} (\vec k)$ in the
293: Brillouin zone for the lower two bands at different $\Omega_z$.
294: A, C and E (B, D, and F) are $F_{xy} (\vec k)$ of the first (second)
295: band at $\Omega_z/t_\pp=0.3$ and $1.7$, respectively.
296: The Chern number of the first band is 1, and that of the second band
297: changes from 0 to -1 at $\Omega_z/t_\pp=\frac{3}{2}$.
298: }
299: \label{fig:chern}
300: \end{figure}
301:
302:
303: The above band structures exhibit non-trivial topological properties.
304: The Berry curvature $F_{xy}(\vec k)$, or the gauge field strength,
305: in the momentum space for the $n$-th band $(n=1\sim 4)$ is defined as
306: $
307: F_{n,xy}(\vec k)=\partial_{k_x} A_{n,y} (\vec k) - \partial_{k_y} A_{n,x}
308: (\vec k),
309: $
310: where $A_{n,i}(i=x,y)$ is gauge potential defined as
311: $A_{n,i}=i\avg{\psi_n(\vec k)| \partial_{k_i}| \psi_n (\vec k)}$
312: \cite{thouless1982,kohmoto1985}.
313: The eigenstates of the lower two bands are related to those of the
314: upper two by the transformation
315: $|\psi_{4-n}(-\vec k)\rangle = (T P)|\psi_n(\vec k)\rangle (n=1,2)$,
316: thus the Berry curvature satisfies $F_{4-n,xy}(-\vec k)=-F_{n,xy}(\vec k)$.
317: The field strength $F_{xy}$ of the lower two band is depicted
318: in Fig. \ref{fig:chern} at different angular velocities.
319: $F_{n,xy}$ mainly distributes at wavectors $\vec k$ with small gap values
320: of $|E_{n\pm 1}(\vec k)-E_{n} (\vec k)|$.
321: The total flux in the BZ for each band is quantized
322: known as the Chern number
323: $C_n=\frac{1}{2\pi} \int d^2 k ~ F_{n,xy} (\vec k)$
324: \cite{thouless1982,kohmoto1985}.
325: At all values of $\Omega_z>0$, the Chern number of band 1 is
326: quantized to $1$, in spite of a significant change of distribution
327: of $F_{xy}$ as increasing $\Omega_z$ as depicted in Fig.
328: \ref{fig:chern} A, C and E.
329: The maximal of $F_{xy}$ are distributed among a ring around the
330: BZ center at small values of $\Omega_z$, and are
331: pushed to the two vertexes $K_{1,2}$ of BZ as $\Omega_z$
332: increases.
333: The Chern number of band 2 is more subtle.
334: At small $\Omega_z$, each of two massive Dirac points at $K_{1,2}$
335: approximately contribute a flux of $\frac{1}{2}$.
336: As $\Omega_z/t_\pp\rightarrow \frac{3}{2} $ from below, the maximum
337: of $F_{xy}$ is shifted to the new Dirac point at the BZ center,
338: which approximately contributes the flux of $\frac{1}{2}$.
339: However, these contributions are canceled by the background negative
340: flux at $\Omega_z/t_\pp<\frac{3}{2}$, and thus the Chern number
341: is $0$.
342: A topological quantum phase transition occurs at
343: $\Omega_z/t_\pp>\frac{3}{2}$ beyond which the flux from the
344: Dirac point $K_0$ flips the sign to $-\frac{1}{2}$.
345: Combined with the background contribution, the Chern number of
346: $C_2$ changes to $-1$.
347: In analogy to electron systems, the transverse conductivity
348: can be defined as the ration between the mass flow and
349: the potential gradient as $\sigma_{xy}= -J_x/\partial_y V$.
350: % which can be accordingly expressed as
351: %\bea
352: %\sigma_{xy}=\frac{m}{\hbar}\frac{1}{2\pi}
353: %\sum_n \int d^2 k F_{n,xy} (\vec k) n_f(E_n (\vec k)-\mu),
354: %\eea
355: %where $n_f$ is the Fermi distribution \cite{nagaosa2006}.
356: When the Fermi level lies in the band gap, $\sigma_{xy}$
357: is quantized as the sum of
358: the Chern numbers of the occupied bands.
359:
360:
361: \begin{figure}
362: \centering\epsfig{file=edge_L03.ps,clip=1,width=0.5\linewidth,angle=-90}
363: \centering\epsfig{file=edge_L17.ps,clip=1,width=0.5\linewidth,angle=-90}
364: \caption{The gapless edge excitations with the open boundary condition
365: along the zig-zag edge of the hexagon lattice.
366: A) $\Omega_z/t_\pp=0.3$; B) $\Omega_z/t_\pp=1.7$.
367: A topological phase transition occurs at $\Omega_z/t_\pp=\frac{3}{2}$
368: above which the edge modes between the middle two bands disappear.
369: }
370: \label{fig:edge}
371: \end{figure}
372:
373: The above band structure with non-vanishing Chern numbers gives rise to
374: topological stable gapless edge modes lying inside the band gap.
375: Fig. \ref{fig:edge} depicts the spectra with the open boundary condition
376: on the zig-zag edges.
377: The number of chiral edge modes inside the gap between $n$ and $n+1$
378: band is the sum of Chern numbers from band 1 to $n$, i.e., $\sum_{i=1,n} C_n$.
379: At $\Omega_z<\frac{3}{2} t_\pp$, the Chern numbers reads
380: $C_1=-C_4=-1$ and $C_2=-C_3=0$, thus edge modes exist in all
381: of the three band gaps with the same chirality.
382: At $\Omega_z>\frac{3}{2} t_\pp$, $C_2$ and $C_3$
383: change to $C_2=-C_3=-1$.
384: Thus the edge modes between band 1 and 2, and that between band 3
385: and 4 are of the opposite chiralities.
386: No edge mode appears between band 2 and 3.
387: This agrees with the picture that Eq. \ref{eq:ham0} reduces
388: to two copies of Haldane's model at $\Omega_z\gg t_\pp$.
389:
390:
391: \begin{figure}
392: \centering\epsfig{file=pQHEB05L20.ps,clip=1,width=0.6\linewidth,angle=-90}
393: \centering\epsfig{file=pQHEB05L150.ps,clip=1,width=0.6\linewidth,angle=-90}
394: \caption{Edge and bulk states spectra of Eq. \ref{eq:QHE} with the open
395: boundary condition along the zig-zag edge. The flux per plaquette
396: $\phi/(2\pi)=1/20$. A) $\Omega_z/t_\pp=0.2$; B)$\Omega_z/t_\pp=1.5$.
397: }
398: \label{fig:QHEedge}
399: \end{figure}
400:
401:
402:
403: We also study the QH effect on Eq. \ref{eq:ham0} arising from Landau levels
404: (LL) by replacing the hopping part to
405: \bea
406: H_{hop}=t_{\pp} \sum_{ \vec{r} \in A}
407: \big\{ p^\dagger_{\vec{r},i} p_{\vec{r}+a\hat{e}_i,i}
408: e^{i\int_{r+ a e_i}^r \vec A \cdot d \vec r} +h.c. \big\},
409: \label{eq:QHE}
410: \eea
411: where the vector potential-$\vec A$ can be generated through another
412: overall lattice rotation or by light induced gauge potential.
413: We will take the flux per plaquette $\Phi$ and $\Omega_z$
414: as two independent variables.
415: The spectra of the above Hamiltonian does not depend on the gauge choice
416: but the physical wavefunctions differ by a gauge transformation.
417: For the calculation convenience, we use the Landau gauge for an open
418: boundary system along the zig-zag edge and take $\Phi/(2\pi)=0.05$.
419:
420: Due to the vector potential $\vec A$, $(TP) (H_{hop}+H_L)
421: (TP)^{-1} \neq -(H_{hop}+H_L) $, thus the spectra are no longer symmetric
422: respect to the zero energy.
423: Generally speaking, all of the four bands split into a number of
424: flat LLs with dispersive edge modes lie in between.
425: The pattern of edge modes does not change much as varying the value
426: of $\Phi$, but significantly changes as increasing $\Omega_z$.
427: At small values of $\Omega_z$ (e.g., $\Omega_z =0.2 t_\pp$ as
428: shown in Fig. \ref{fig:QHEedge}. A ), gapless edge modes go through
429: the entire spectra from the very band bottom to top.
430: Landau levels close to the zero energy arise from Dirac cones at $K_{1,2}$
431: with opposite masses as shown in Eq. \ref{eq:dirac1}.
432: The $0$th LL is pushed to negative energy at the gap value
433: around $-0.26t_\pp$.
434: The number of chiral edge modes between levels of $n=0$ to $\pm1$
435: is 1 with opposite chirality and that between $n=\pm 1$ and $\pm 2$ is 3.
436: The energies of $n=\pm1$ and $n=\pm 2$ appear roughly symmetric to
437: zero energy.
438: Next let us look at $\Omega_z=\frac{3}{2} t_\pp$ where a single gapless
439: Dirac cone appears as shown in Eq. \ref{eq:dirac2}.
440: Indeed the $0$th LL appear close to the zero energy but with a small
441: deviation, which is understandable as no exact symmetry to protect it
442: right at the zero energy.
443: It is tempting to think the appearance of the half-integer QH effect,
444: but this is impossible in free lattice fermion systems \cite{haldane1988}.
445: Another half has to be contributed from the high energy part of the
446: band structure.
447: As a result, the number of chiral edge modes between LLs $n=0$
448: and $1$ is $\frac{1}{2}+\frac{1}{2}=1$, while that between LLs
449: $n=0$ and $-1$ is $-\frac{1}{2}+\frac{1}{2}=0$.
450: Thus the spectra from bottom to top become disconnected without
451: edge modes connecting them.
452: This disconnection actually begins to appear even earlier at
453: $\Omega_z=1.2 t_\pp$, and is enhanced as $\Omega_z$ goes larger.
454: At large values of $\Omega_z$, the model reduces to two copies
455: ($p_x\pm ip_y$) of Haldane's model.
456: The patterns of LLs between band 1 and 2, and between band 3 and 4
457: become the those of the two massive Dirac cones with opposite
458: mass signs.
459: When Fermi level lies in between LLs, the transverse conductance
460: $\sigma_{xy}$ is quantized at the number of chiral
461: edge modes.
462:
463:
464: In summary, we propose to investigate the topological insulating
465: states in the $p$-orbital systems in the honeycomb lattice, which
466: can be realized by the current available experimental techniques.
467: The orbital angular momentum polarization generates non-trivial
468: Chern numbers in the band structure, which gives rise to
469: the orbital counterpart of QAH effect without LLs.
470: QH effect arising for LLs are also investigated, which shows
471: quantitative different features from those in graphene.
472:
473:
474: C. W. thanks D. Arovas, M. Fogler and J. Hirsch for helpful discussions,
475: and N. Gemekel for the introduction the method of rotating optical lattices.
476: C. W. is supported by the start-up funding at UCSD, and the
477: Sloan Research Foundation.
478:
479: %\bibliographystyle{prsty}
480: %\bibliography{orbital,spin32,extra,topo}
481:
482: \begin{thebibliography}{10}
483:
484: \bibitem{thouless1982}
485: D.~J. Thouless, M. Kohmoto, M.~P. Nightingale, and M. den Nijs, Phys. Rev.
486: Lett. {\bf 49}, 405 (1982).
487:
488: \bibitem{kohmoto1985}
489: M. Kohmoto, Annals of Physics {\bf 160}, 296 (1985).
490:
491: \bibitem{jackiw1984}
492: R. Jackiw, Phys. Rev. D {\bf 29}, 2375 (1984).
493:
494: \bibitem{fradkin1986}
495: E. Fradkin, E. Dagotto, and D. Boyanovsky, Phys. Rev. Lett. {\bf 57}, 2967
496: (1986).
497:
498: \bibitem{haldane1988}
499: F.~D.~M. Haldane, Phys. Rev. Lett. {\bf 61}, 2015 (1988).
500:
501: \bibitem{bernevig2006a}
502: B.~A. Bernevig, T.~L. Hughes, and S.-C. Zhang, Science {\bf 314}, 1757
503: (2006).
504:
505: \bibitem{qi2008a}
506: X.-L. Qi, T. Hughes, and S.-C. Zhang, Topological Field Theory of Time-Reversal
507: Invariant Insulators, arXiv.org:0802.3537, 2008.
508:
509: \bibitem{kane2005}
510: C.~L. Kane and E.~J. Mele, Physical Review Letters {\bf 95}, 146802 (2005).
511:
512: \bibitem{hirsch1989}
513: J.~E. Hirsch, Phys. Rev. B {\bf 40}, 2354 (1989).
514:
515: \bibitem{sheng2006}
516: D.~N. Sheng, Z.~Y. Weng, L. Sheng, and F.~D.~M. Haldane, Physical Review
517: Letters {\bf 97}, 036808 (2006).
518:
519: \bibitem{moore2007}
520: J.~E. Moore and L. Balents, Physical Review B {\bf 75}, 121306 (2007).
521:
522: \bibitem{roy2006}
523: R. Roy, On the \$Z\_2\$ classification of Quantum Spin Hall Models, 2006.
524:
525: \bibitem{konig2007}
526: M. K\"onig {\it et~al.}, Science {\bf 318}, 766 (2007).
527:
528: \bibitem{hsieh2008}
529: D. Hsieh {\it et~al.}, Nature {\bf 452}, 970 (2008).
530:
531: \bibitem{karplus1954}
532: R. Karplus and J.~M. Luttinger, Phys. Rev. {\bf 95}, 1154 (1954).
533:
534: \bibitem{smit1958}
535: J. Smit, Physica {\bf 24}, 39 (1954).
536:
537: \bibitem{berger1970}
538: L. Berger, Phys. Rev. B {\bf 2}, 4559 (1970).
539:
540: \bibitem{jungwirth2002}
541: T. Jungwirth, Q. Niu, and A. MacDonald, Phys. Rev. Lett. {\bf 88}, 207208
542: (2002).
543:
544: \bibitem{nagaosa2006}
545: N. Nagaosa, J. Phys. Soc. Jpn {\bf 75}, 42001 (2006).
546:
547: \bibitem{qi2006a}
548: X.-L. Qi, Y.-S. Wu, and S.-C. Zhang, Phys. Rev. B {\bf 74}, 085308 (2006).
549:
550: \bibitem{liu2008}
551: C.-X. Liu {\it et~al.}, Quantum Anomalous Hall Effect in
552: Hg\$\_\{1-y\}\$Mn\$\_\{y\}\$Te Quantum Wells, arXiv.org:0802.2711, 2008.
553:
554: \bibitem{onoda2003a}
555: M. Onoda and N. Nagaosa, Physical Review Letters {\bf 90}, 206601 (2003).
556:
557: \bibitem{ho2000}
558: T.~L. Ho and S.~K. Yip, Phys. Rev. Lett. {\bf 84}, 4031 (2000).
559:
560: \bibitem{scarola2007}
561: V.~W. Scarola and S.~D. Sarma, Physical Review Letters {\bf 98}, 210403
562: (2007).
563:
564: \bibitem{shao2008}
565: L.~B. Shao {\it et~al.}, Realizing and Detecting the Haldane's Quantum Hall
566: effect with Ultracold Atoms, arXiv.org:0804.1850, 2008.
567:
568: \bibitem{umucalilar2008}
569: R.~O. Umucalilar, H. Zhai, and M.~O. Oktel, Physical Review Letters {\bf 100},
570: 070402 (2008).
571:
572: \bibitem{zhu2006}
573: S.-L. Zhu {\it et~al.}, Physical Review Letters {\bf 97}, 240401 (2006).
574:
575: \bibitem{gemelke2007}
576: N. Gemelke, Quantum Degenerate Atomic Gases in Controlled Optical Lattice
577: Potentials, Ph. D. Thesis, 2007.
578:
579: \bibitem{grynberg1993}
580: G. Grynberg {\it et~al.}, Phys. Rev. Lett. {\bf 70}, 2249 (1993).
581:
582: \bibitem{wu2007}
583: C. Wu, D. Bergman, L. Balents, and S.~D. Sarma, Phys. Rev. Lett. {\bf 99},
584: 70401 (2007).
585:
586: \bibitem{wu2007a}
587: C. Wu and S. Das~Sarma, The \$p\_\{x,y\}\$-orbital counterpart of graphene:
588: cold atoms in the honeycomb optical lattice, arXiv.org:0712.4284, 2007.
589:
590: \bibitem{wu2008}
591: C. Wu, Orbital ordering and frustration of \$p\$-band Mott-insulators,
592: arXiv.org:0801.0888.
593:
594: \bibitem{nielsen1981}
595: H.~B. Nielsen and M. Ninomiya, Nucl. Phys. B {\bf 185}, 20 (1981).
596:
597: \end{thebibliography}
598:
599:
600:
601: \end{document}
602:
603: %where $\vec A=\frac{m}{h} \vec \Omega_z \times \vec r$.
604: %For simplicity, we neglect the centrifugal potential part which
605: %just softens the trapping potential.
606: %The flux per plaquette is $\phi/(2\pi)=\frac{3\sqrt 3}{h} m\Omega_z a^2$.
607: %If we use $^{40}$K and take $a=0.5 \mu m$, a flux of $\phi/(2\pi)=0.05$
608: %corresponds to $\Omega_z\approx 384 Hz$, which is still larger than
609: %the trapping frequency.
610: