1: \documentstyle[floats,twocolumn,aps,prl]{revtex}
2: \begin{document}
3: \ifpreprintsty\else
4: \twocolumn[\hsize\textwidth%
5: \columnwidth\hsize\csname@twocolumnfalse\endcsname
6: \fi
7: \title{Theory of the Half-Polarized Quantum Hall States}
8: \author{V.M. Apalkov$^\ast$, T. Chakraborty$^\ast$,
9: P. Pietil\"ainen$^\dagger$, and K. Niemel\"a$^\dagger$}
10: \address{$^\ast$Max-Planck-Institut f\"ur Physik Komplexer Systeme,
11: Dresden, Germany \\
12: $^\dagger$Theoretical Physics, University of Oulu, Finland}
13: \maketitle
14: \begin{abstract}
15: We report a theoretical analysis of the half-polarized quantum
16: Hall states observed in a recent experiment. Our numerical results
17: indicate that the ground state energy of the quantum Hall $\nu=\frac23$
18: and $\nu=\frac25$ states versus spin polarization has a downward
19: cusp at half the maximal spin polarization. We map the two-component
20: fermion system onto a system of excitons and describe the ground state
21: as a liquid state of excitons with non-zero values of exciton angular
22: momentum.
23: \end{abstract}
24: \ifpreprintsty\clearpage\else\vskip1pc]\fi
25: %\pacs{73.40.Hm,71.45.Gm,73.20,Dx}
26: \narrowtext
27:
28: In recent years it has become increasingly clear that spin
29: degree of freedom plays an important role in the fractional quantum
30: Hall effect (FQHE) where many novel and interesting spin-related
31: phenomenon have been observed both theoretically and experimentally
32: \cite{book}. One of the important problems in this field is the
33: influence of Zeeman splitting on the properties of FQHE systems,
34: in particular, on the ground state spin-polarizations. It is now well
35: established that for some filling factors ($\nu=\frac1m$) the ground
36: state is fully spin-polarized for all values of Zeeman splitting,
37: while for other filling factors (for example, $\nu =2/3$, $2/5$) the
38: ground state is fully polarized only for large values of Zeeman
39: splitting but unpolarized or partially polarized for small (or zero)
40: values of Zeeman energy\cite{book,rev}. One interesting problem then
41: is to find the state for intermediate values of Zeeman energy. That
42: problem was highlighted in a recent experimental work \cite{kukushkin},
43: where the magnetic field driven spin transitions at various FQHE
44: states were reported and in particular, at $\nu=\frac23$ and
45: $\nu=\frac25$, weak features (plateau-like singularities) were observed
46: at half the maximal spin polarization of the system. Observed stability
47: of the half-polarized states means that the ground state energy of the
48: system as a function of spin polarization should have non-monotonic
49: behavior at half polarization. Our earlier work \cite{karri} did not
50: provide much information about the nature of states at
51: half-polarization. In this paper, we have explored possible
52: ground states at half-polarization for filling factors $\nu=\frac25$
53: and $\nu=\frac23$. We find that the ground state energy versus spin
54: polarization has a downward cusp at half polarization that might
55: describe stability of the observed state.
56:
57: The FQHE system at filling factor $\nu=\frac25$ can be
58: described as a composite fermion (CF) system with total filling
59: factor $\nu=2$ \cite{cfermion}. At high values of the Zeeman energy
60: the composite fermions will occupy $n=0$ $\uparrow$-spin and
61: $n=1$ $\uparrow$-spin Landau levels of composite fermions.
62: As a result we have a fully spin-polarized state.
63: At low Zeeman energies they will occupy $n=0$
64: $\uparrow$-spin and $n=0$ $\downarrow$-spin Landau levels
65: which will result in an unpolarized state. At intermediate
66: values of Zeeman energy the composite fermions will
67: fully occupy $n=0$ $\uparrow$-spin Landau level and partially
68: occupy $n=0$ $\downarrow$-spin and $n=1$ $\uparrow$-spin
69: levels with filling factors $\nu_1$ and $\nu_2$ respectively,
70: with $\nu_1+\nu_2=1$. The half polarized state corresponds to
71: $\nu_1=\nu_2=\frac12$. As the fully occupied state can
72: be considered as a non-dynamical background, the composite
73: fermions occupying the partially filled levels can be thought of
74: as a system consisting of two types of fermions with $\nu=1$.
75: The Hamiltonian of the two-component system has the form
76: \begin{eqnarray}
77: \nonumber
78: {\cal H}= \frac12 \sum_{\alpha \beta}
79: && \int \int d\vec{r}_1 d\vec{r}_2
80: V_{\alpha\beta}\left(\left|\vec{r}_1-\vec{r}_2\right|\right)\\
81: &&\times \psi^{+}_{\alpha} (\vec{r}_1) \psi^{+}_{\beta} (\vec{r}_2)
82: \psi_{\beta} (\vec{r}_2) \psi_{\alpha}(\vec{r}_1)
83: % \nonumber
84: \end{eqnarray}
85: where $\alpha, \beta=1$ and 2, $V_{11}$ is the interaction potential
86: between fermions of type 1, $V_{22}$ is the interaction potential
87: between fermions of type 2, and $V_{12}$ is the potential
88: between fermions of types 1 and 2.
89: The specific feature of this system is that the Hamiltonian is
90: completly ``non-symmetric'', i.e., all interaction potentials $V_{11}$,
91: $V_{22}$ and $V_{12}$ are different.
92:
93: Alternatively, we can also describe the $\nu=\frac25$ state as the
94: daughter state of the $\nu=\frac13$ system \cite{book}. Then the
95: spin-polarized state of the $\nu=\frac25$ system is due to
96: condensation of spin-polarized quasiparticles with filling factor
97: $\nu=\frac12$ and the unpolarized state as the condensation of
98: spin-reversed quasiparticles \cite{kane}. For intermediate polarization
99: we have the system of spin-polarized and spin-reversed quasiparticles
100: with $\nu=\frac12$. Because they are Bose particles we can
101: map this system into the system of fermions with $\nu=1$. Here again,
102: as for the composite fermion picture, we have two types
103: of fermions with Hamiltonian (1). We also have the similar picture
104: for $\nu=\frac23$ state, which can be described as the daughter state
105: of $\nu=1$ system with condensation of spin-polarized and spin-reversed
106: holes of $\nu=1$.
107:
108: For strong enough repulsion between the fermions at the same point
109: we expect that the ground state of the
110: system at $\nu_1=\nu_2 =\frac12$ and other values of $\nu_1$
111: to be the Halperin-(1,1,1) state \cite{book,bert}
112: \begin{eqnarray*}
113: \psi=&&\prod_{i<j=1}^{N_1}(z_i-z_j)\prod_{k<m=1}^{N_2}
114: (\tilde z_k-\tilde z_m)\prod_{i=1}^{N_1}\prod_{k=1}^{N_2}
115: (z_i-\tilde z_k)\\
116: &&\times\prod_{i=1}^{N_1}{\rm e}^{-\vert z_i\vert^2/4\ell_0^2}
117: \prod_{k=1}^{N_2}{\rm e}^{-\vert\tilde z_k\vert^2/4\ell_0^2}
118: \end{eqnarray*}
119: where $N_1$ is the number of fermions of type 1, $N_2$ is the
120: number of fermions of type 2; $N_1+N_2=N$. Here $z_{j}$ are the
121: coordinates (complex) of fermions of type 1, $\tilde z_{j}$ are
122: those of fermions of type 2. If we consider the system of two
123: types of fermions as a two-level system and introduce a pseudo-spin
124: $\tau$ for the states at different levels then the Halperin-(1,1,1)
125: liquid state has $\tau=N/2$ and $N/2 \ge \tau_z \ge -N/2$. If this
126: state is the correct ground state then the transition from a polarized
127: state to an unpolarized state of the system is just the rotation
128: of pseudo-spin vector from $\tau_z =-N/2$ to $\tau_z =N/2$ with
129: fixed value of the total pseudo-spin $\tau=N/2$. But then
130: the ground state energy of the system is monotonic with
131: polarization of the system (quadratic function) without
132: any singularity at half polarization. This means that the Halperin
133: state is not the true ground state of the half-polarized state.
134:
135: \begin{figure}
136: \begin{center}
137: \begin{picture}(120,130)
138: \put(0,0){\special{psfile=Fig1.ps angle=0
139: voffset=-420 hoffset=-250 hscale=100 vscale=100}}
140: \end{picture}
141: \vspace*{4.8cm}
142: \caption{Ground state energy as a function of filling factor of
143: fermions of type 1, $\nu_1=N_1/12$, for (a) ``symmetric'',
144: (b) ``quasiparticle'' and (c) ``quasihole'' systems. The energy is in
145: units of $e^2/\epsilon\ell_0$.
146: }
147: \end{center}
148: \end{figure}
149:
150: In Ref.\cite{murthy} it was proposed that the half-polarized ground
151: state is a charge-density-wave (CDW) of composite fermions. To check
152: this claim we have compared the energies of proposed CDW state and
153: the Halperin-(1,1,1) state \cite{vadim}. The CDW of Ref.\cite{murthy}
154: in our notations is formed by type-1 and type-2 fermions on a square
155: lattice. We calculate the cohesive energy of this state from
156: \cite{fukuyama}
157: \begin{eqnarray}
158: \nonumber
159: &&E_{\rm CDW}=-\frac12 \sum_{\vec{Q}} \left(
160: V^{\rm HF}_{11}(\vec{Q})\Delta_{1}(\vec{Q}) \Delta_{1}(-\vec{Q})
161: \right. \\
162: \nonumber
163: &&+ V^{\rm HF}_{22}(\vec{Q})\Delta_{2}(\vec{Q})\Delta_{2}(-\vec{Q})
164: + \left. 2 V^{\rm H}_{12}(\vec{Q})\Delta_{1}(\vec{Q})\Delta_{2}(-\vec{Q})
165: \right)\\
166: \end{eqnarray}
167: where $V^{\rm HF}_{11}(\vec{Q})$ and $V^{\rm HF}_{22}(\vec{Q})$ are
168: the Hartree-Fock potentials for fermions of type 1 and 2, respectively,
169: $V^{\rm H}_{12}(\vec{Q})$ is the Hartree interaction potential between
170: fermions of types 1 and 2. The order parameter $\Delta_{\alpha}(\vec{Q})$
171: of the CDW corresponding to wave vector $\vec{Q}$ for fermions of
172: type $\alpha $ is taken to be non-zero only for reciprocal vectors:
173: $\vec{Q}=(\pm Q_0,0),(0,\pm Q_0)$ and $(\pm Q_0,\pm Q_0)$,
174: $(\pm Q_0,\mp Q_0)$, where $Q^2_0\ell_0^2=\pi$ ($\ell_0$ is the
175: magnetic length for composite fermions). The energy of the
176: Halperin-(1,1,1) state is calculated from
177: \[
178: E_{(1,1,1)}=-\frac1{4\pi}\int d^2 r V_{\rm eff}(r)[g(r)-1],
179: \]
180: where $g(r)=1-\exp(-r^2/2\ell_0^2)$ is the correlation function
181: of a fully occupied Landau level and
182: \[
183: V_{\rm eff}(r)=\frac14\left[V_{11}(r)+V_{22}(r)+2V_{12}(r)\right]
184: \]
185: is the effective interaction between composite fermions.
186:
187: We have considered Coulomb interaction between composite fermions
188: of types 1 and 2. The interaction asymmetry here results from their
189: different form factors because fermions belong to different Landau
190: levels. The energy of the Halperin liquid state at half polarization
191: is $-0.196 e^2/\varepsilon \ell_0$ and is {\it lower} than that of
192: the proposed CDW state, $-0.123 e^2/\varepsilon \ell_0$.
193: This means that the proposed CDW state is not the lowest energy state
194: and therefore not the ground state at half-polarization.
195:
196: To find the true ground state of the system that has lower energy
197: than the Halperin state and that can also explain the half-polarized
198: singularity, we considered the system described by the Hamiltonian
199: (1) numerically in a spherical geometry \cite{haldane,fano}. All
200: computations were done for a 12 fermion system on a sphere with sphere
201: parameter $q=5.5$, where the radius of the sphere
202: $R=\sqrt{q}=2.34$ in units of magnetic length $\ell_0$. Because the
203: Hamiltonian (1) does not change the number of fermions of a given
204: type the eigenstates of the system can be classified by the number
205: of fermions of type 1 ($12\ge N_1\ge 0$) and by angular momentum $L$
206: (due to spherical geometry). In what follows, we have investigated
207: three systems:
208: (i) a ``symmetric'' system $(V_{11}=V_{22}=V_{12})$, where the
209: interaction potentials are Coulombic except at the origin where it
210: is taken to be less repulsive,
211: (ii) the nonsymmetric system where the interactions between fermions
212: were taken as the interactions between quasiparticles of $\nu=\frac13$
213: state -- the ``quasiparticle'' system. These interaction potentials
214: were found from finite size computations following method of
215: Ref. \cite{morf}, and, (iii) the nonsymmetric system where the
216: interactions between fermions were taken as the interactions between
217: quasiholes of $\nu =1$ state -- the ``quasihole'' system.
218: For all these systems we found a similar behavior: The ground
219: state energy as a function of filling factor $\nu_1$ ($N_1$) has
220: a {\it downward cusp} at $\nu_1=\frac12$ ($N_1=6$), as seen in
221: Fig. 1. This point corresponds to a half-polarized state of the
222: original system and the cusp indicates stability of the observed
223: \cite{kukushkin} half-polarized state.
224:
225: It should be mentioned that for ``quasiparticle'' and ``quasihole''
226: systems we do not include in the Hamiltonian (1) the different
227: creation energies for polarized and spin-reversed quasiparticles
228: and quasiholes, which acts as an effective internal Zeeman splitting
229: and makes the unpolarized state ($\nu_1=1$) the ground state at
230: zero value of real Zeeman energy. These terms are monotonic functions
231: of polarization and do not change the singular behaviour at half
232: polarization.
233:
234: In order to analyze the symmetric system more carefully,
235: we consider our two-component fermion system with $\nu=1$
236: as a system of excitons, i.e., the electron-hole pairs. We choose
237: type-1 fermions as electrons and the absence of type-2 fermions as holes.
238: In this case the filling factors for electrons and holes are the same and
239: is equal to $\nu_1$, which is also the exciton filling factor.
240:
241: \begin{figure}
242: \begin{center}
243: \begin{picture}(100,130)
244: \put(0,0){\special{psfile=Fig2.ps angle=0
245: voffset=-380 hoffset=-270 hscale=100 vscale=100}}
246: \end{picture}
247: \vspace*{5.3cm}
248: \caption{Energy spectrum of the symmetric system for (a) $N_1=5$
249: and (b) $N_1=6$. Energy is in the units of $e^2/\epsilon\ell_0$.
250: }
251: \end{center}
252: \end{figure}
253:
254: It is well known that for a symmetric (Coulomb) multi-exciton system
255: the ground state is a Bose-condensed state of excitons with zero
256: momentum \cite{lerner,rashba,lozovik,rice}. In original fermion
257: language these
258: states are the Halperin states. For Coulomb interaction, these
259: Bose-condensed states are the ground states
260: for all values of filling factor $\nu_1$ ($1>\nu_1>0$). Let us now
261: decrease the repulsion between original fermions at the origin.
262: We do it numerically by decreasing the value of the pseudopotential
263: with zero angular momentum, $V_0$. A decrease of $V_0$ by about 40\%
264: results in ground states of the multi-exciton system that are not
265: the Bose-condensed states of excitons with $L=0$ (Fig. 1). These transitions
266: are also accompanied by transition of the ground state of the one-exciton
267: system to the state with non-zero angular momentum, $L=1$.
268:
269: Interestingly, the transition from a Bose-condensed state of
270: zero-momentum excitons to a new state was also observed in a double
271: layer system \cite{tapash} with $\nu=1$ when the
272: separation between the layers is increased beyond some critical value
273: \cite{fertig,MacDonald}. At a critical layer separation of this system,
274: the dispersion relation of the collective mode becomes negative at
275: momentum $q\sim 1.3/\ell_0$ and it was proposed that this transition
276: is the transition to a charge density wave state \cite{quinn}.
277: In our case the dispersion relation of the collective mode of the
278: Bose-condensed state also becomes negative but for much smaller momentum
279: $q\sim L/R\sim 0.4/\ell_0$, which results from the fact that the ground
280: state of one-exciton system has angular momentum $L=1$.
281:
282: As stated above,
283: for the Coulomb interaction the ground state of the one-exciton system has
284: zero momentum and the multi-exciton ground state is the Bose-condensed
285: liquid state of excitons with zero momentum \cite{rice}. In analogy
286: with the Coulomb system, we describe our multi-exciton ground state
287: as a liquid state of excitons with non-zero angular momentum $L=1$,
288: a {\it non-symmetric exciton liquid}.
289: Our multi-exciton system does not contain $L=0$ excitons because they
290: are non-interacting and removing one such exciton would not change the
291: energy of the multi-exciton system. However, in Fig. 1, the energy of
292: our 5-exciton system is higher then the energy of the 6-exciton system.
293: That means the 6-exciton system can have only excitons with non-zero
294: angular momemtum.
295:
296: In Fig. 2, the energy spectrum of multi-exciton symmetric system is
297: shown for 5- and 6-exciton systems. We can see that the ground state
298: of the 6-exciton system has zero angular momentum, while the ground state of
299: the 5-exciton system has angular momentum $L=1$. For the Halperin liquid each
300: exciton effectively occupies only one state (electron and hole are in
301: the same place) and the filling factor of the exciton is equal to the
302: filling factor of electrons, $\nu_1$. In the non-symmetric exciton
303: liquid, each exciton occupies effectively two states. As a result, the
304: effective filling factor of excitons is $2\nu_1$. Then for
305: $\nu_1=\frac12$ the filling factor of excitons is 1 which means that
306: we have completely occupied Landau level. Results of Fig. 2 do support
307: this contention because removing one exciton with angular momentum
308: $L=1$ from 6-exciton system left the exciton hole with the same angular
309: momentum, which means that the system behaves as though the levels
310: are completely filled. The energy spectrum of the 6-exciton system,
311: which corresponds to a half-polarized state of the original system,
312: also has a gap. However, from the finite size results we can not
313: ay with certainty if this gap will survive in the thermodynamic limit.
314:
315: \begin{figure}
316: \begin{center}
317: \begin{picture}(100,130)
318: \put(0,0){\special{psfile=Fig3.ps angle=0
319: voffset=-420 hoffset=-250 hscale=95 vscale=95}}
320: \end{picture}
321: \vspace*{7.0cm}
322: \caption{Ground state pair correlation functions $g_{11}(r)$
323: (curves labelled ``1'') and $g_{12}(r)$ (curves labelled ``2'')
324: for (a) symmetric, (b) ``quasiparticle'' and (c) ``quasihole''
325: systems. The solid lines are for $N_1=4$ systems, and dashed lines
326: are for $N_1=6$ systems. Correlation functions are shown in
327: units of maximum electron density, and $r$ is in units of the
328: magnetic length.
329: }
330: \end{center}
331: \end{figure}
332:
333: In Fig. 3, the pair correlation functions $g_{11}(r)$ and $g_{12}(r)$
334: are shown for $N_1=4$ and 6, for symmetric (a), ``quasiparticle'' (b)
335: and ``quasihole'' (c) systems. The general feature of all these
336: correlation functions is the non-zero value of $g_{12}(0)$.
337: For the Bose-condensed state of excitons with zero momentum
338: (the Halperin state) we would expect $g_{12}(0)=0$, i.e.,
339: the holes in the multi-exciton picture are sitting exactly at the
340: position of the electrons. The non-zero values of $g_{12}(0)$ can
341: therefore be directly associated with the cusp at $N_1=6$.
342:
343: In closing, we have investigated the possible ground states at half
344: the maximal spin polarization for $\nu=\frac25, \frac23$ FQHE
345: states. Our results indicate that for the systems studied here,
346: there is a downward cusp at half polarization that reflects the
347: observed structure in a recent experiment \cite{kukushkin}. We
348: interpret this result as due to condensation of a non-symmeric
349: exciton liquid.
350:
351: Two of us (V.M.A. and T.C.) would like to thank Peter Fulde for
352: his support and kind hospitality in Dresden.
353:
354: \begin{references}
355: \vspace*{-0.8cm}
356: \bibitem{book} T. Chakraborty and P. Pietil\"ainen, {\it The
357: Quantum Hall Effects}, (Springer, New York, 1995), 2nd Edition.
358: \bibitem{rev} T. Chakraborty, Adv. Phys. (to be published).
359: \bibitem{kukushkin} I.V. Kukushkin, K. von Klitzing, and K. Eberl,
360: Phys. Rev. Lett. {\bf 82}, 3665 (1999).
361: \bibitem{karri}
362: K. Niemel\"a, P. Pietil\"ainen, and T. Chakraborty, Physica B
363: {\bf 284}, 1716 (2000).
364: \bibitem{cfermion} S. Das Sarma and A. Pinczuk, (Eds.), {\it
365: Perspectives in Quantum Hall Effects}, (Wiley, New York, 1997).
366: \bibitem{kane} D.-H. Lee and C.L. Kane, Phys. Rev. Lett. {\bf 64},
367: 1313 (1990).
368: \bibitem{bert} B.I. Halperin, Helv. Phys. Acta {\bf 56}, 75 (1983).
369: \bibitem{murthy} G. Murthy, Phys. Rev. Lett. {\bf 84}, 350 (2000).
370: \bibitem{vadim} V.M. Apalkov, T. Chakraborty, P. Pietil\"ainen, and
371: K. Niemel\"a, Cond-mat/0007043.
372: \bibitem{fukuyama} D. Yoshioka and H. Fukuyama, J. Phys. Soc. Jpn. {\bf 47},
373: 394 (1979).
374: \bibitem{haldane} F.D.M. Haldane, Phys. Rev. Lett. {\bf 51}, 605 (1983).
375: \bibitem{fano} G. Fano, F. Ortolani, and E. Colombo, Phys. Rev. B {\bf 34},
376: 2670 (1989).
377: \bibitem{morf} P. Beran, and R. Morf, Phys. Rev. B {\bf 43}, 12654 (1991).
378: \bibitem{lerner} I.V. Lerner and Yu.E. Lozovik, Zh.Eksp.Teor.Fiz. {\bf 80},
379: 1488 (1981) [Sov. Phys.-JETP {\bf 53}, 763 (1981)].
380: \bibitem{rashba} Y.A. Bychkov and E.I. Rashba, Solid State Commun. {\bf 48},
381: 399 (1983).
382: \bibitem{lozovik} A.B. Dzyubenko and Yu.E. Lozovik, Fiz. Tverd. Tela {\bf 25},
383: 1519 (1983) [Sov. Phys. Solid State {\bf 25}, 874 (1983)].
384: \bibitem{rice} D. Paquet, T.M. Rice, and K. Ueda, Phys. Rev. B {\bf 32},
385: 5208 (1985).
386: \bibitem{tapash} T. Chakraborty and P. Pietil\"ainen, Phys.
387: Rev. Lett. {\bf 59}, 2784 (1987).
388: \bibitem{fertig} H.A. Fertig, Phys. Rev. B {\bf 40}, 1087 (1989).
389: \bibitem{MacDonald} A.H. MacDonald, P.M. Platzman, and G.S. Boebinger,
390: Phys. Rev. Lett. {\bf 65}, 775 (1990).
391: \bibitem{quinn} X.M. Chen and J.J. Quinn, Phys. Rev. Lett. {\bf 67},
392: 895 (1991).
393: \end{references}
394:
395: \end{document}
396:
397: