1: \documentclass[floats,aps]{revtex4}
2: %\documentclass[aps,prb,%
3: %draft,%
4: %twocolumn
5: %,nobalancelastpage,%
6: %onecolumn
7: %,preprint,%
8: %tightenlines,%
9: %amssymb,amsmath,%
10: %nopreprintnumbers,%
11: %showpacs,%
12: %eqsecnum,%
13: %superscriptaddress,%
14: %endfloats,%
15: %byrevtex]{revtex4}
16: \usepackage{calc}
17: %\usepackage{graphics}
18: \usepackage{graphicx}
19: \begin{document}
20: \def \beq{\begin{equation}}
21: \def \eeq{\end{equation}}
22: \begin{abstract}
23: We argue that the naively expected singularities of the Fermi surface,
24: in the mixed composite boson-composite fermion states proposed [S.H. Simon {\it et al.},
25: PRL {\bf 91}, 046803 (2003)] for the evolution of $\nu =1$ bilayer quantum Hall system with distance, are obliterated. Our conclusion is based on a careful analysis of the momentum distribution in $\nu = \frac{1}{2}$ single-layer composite-
26: fermion state. We point out to a possibility of the phenomenon hitherto unknown outside
27: Kondo lattice systems when, in a translationally invariant system,
28: Fermi-liquid-like portion of electrons enlarges its volume.
29: \end{abstract}
30: \pacs{}
31: \title{Status of the Fermi surface in mixed composite boson - composite fermion quantum Hall states}
32: \author{Milica V. Milovanovi\'{c}}
33: \address{Institute of Physics, P.O.Box 68, 11080 Belgrade, Serbia and Montenegro}
34: \date{\today}
35: \maketitle
36: \vskip2pc]
37: \narrowtext
38: The nature and physics of the transition in the bilayer $\nu = 1$
39: quantum Hall (QH) system \cite{halp} between the well-established phases:
40: one characteristic for the distances between the layers of the order
41: of or smaller than magnetic length, sometimes described as ``111'' state,
42: and the other for larger distances, described by two separate
43: Fermi-liquid-like states of composite fermions (CFs) attracted recently
44: the attention of experimentalists \cite{kelog} and is the focus of several
45: theoretical papers \cite{sch,stha,srm,kimv}. Only
46: references \cite{stha} and \cite{srm} make a prediction
47: for a coexistence region between two phases, with a unique
48: property, semicircle law for the longitudinal and Hall drag
49: resistance that was revealed in the experiments \cite{kelog}.
50: The reference \cite{srm}
51: introduces a form of the ground state of the system that may
52: continously interpolate between the 111 state, usually described
53: by composite bosons (CBs), and the two seperate Fermi-liquid-like states of
54: CFs. The ground state proved to be a good variational ansatz when compared
55: with the exact solution in numerical studies \cite{srm}. The form of
56: the variational state for certain distance between the layers may be
57: described as one in which classically speaking
58: some of the electrons are in the 111 state (they make CBs) and the
59: others participate in two Fermi seas of CFs. Gradually the
60: number of CFs increases as the distance becomes larger. Therefore
61: the description easily accounts for the continous nature of the transition
62: as observed in the experiments \cite{kelog}. On the other hand the
63: proposal that came first, based on a phase separated picture, \cite{stha},
64: in which percolating puddles of one phase are in the other, well
65: enough exibits the transport properties measured in the experiments.
66: The advantage of the homogenous model (\cite{srm}), which accounts
67: for the same transport properties, is that it also accounts
68: for the strong 111 (interlayer) correlations that occur even
69: deep in the CF region \cite{kelog}.
70:
71: Here we study the Fermi surface singularities in the proposed wave
72: functions \cite{srm}. Naively they are expected at the Fermi momenta
73: directly related to the number of CFs in the particular partition
74: of the overall number of electrons into CFs and CBs. The analysis
75: begins with a careful study of the $\nu = \frac{1}{2}$ CF problem,
76: so that the relationships found can be readily applied to the
77: mixed state case. We found that the CF momentum distribution near
78: the naively expected Fermi momenta depend analytically on the
79: distance to the Fermi momenta, therefore showing no signature of the
80: Fermi surfaces.
81:
82: Soon after Halperin, Lee, and Read \cite{hlr} proposed their theory for $\nu = \frac{1}{2}$ fractional QH effect Bares and Wen \cite{bw} considered fermions in low dimensions interacting via a long range $ - \frac{2 \pi}{|\vec{q}|^{2}}$ interaction. They used as a good ansatz for the ground state, a wave function of the Feenberg-Jastrow type,
83: \beq
84: \Psi_{o}(\{x\}) = \prod_{i<j} |x_{i} - x_{j}|^{m} \Psi_{FS}(\{x\}),
85: \label{fjs}
86: \eeq
87: where $\Psi_{FS}$ denotes a Slater determinant of filled Fermi sea of free single-particle states. If $m = 2$ this construction is the Rezayi-Read \cite{rr} ground state, in the representation of CFs and when the projection to the lowest Landau level (LLL) is neglected, found to correctly captures the physics at $\nu = \frac{1}{2}$. By doing a calculation of a random phase approximation (RPA) type on (\ref{fjs}) Bares and Wen found that the leading singularity of the momentum distribution near $ k_{F}$, in two dimensions, is
88: \beq
89: \label{ls}
90: \delta n_{k} \approx \frac{m}{2} \{n_{k}^{o} \ln|\delta k| -
91: (1 - n_{k}^{o}) \ln|\delta k|\},
92: \eeq
93: where $\delta k = |\vec{k}| - k_{F}$ and $n_{k}^{o}$ denotes the free-Fermi-gas momentum distribution. They also remarked that if we interpret the rhs of (\ref{ls}) as the first term in an expansion in powers of $m$ we can write (near $k_{F}$)
94: \beq
95: n_{k} = \frac{1}{2} + \frac{1}{2} \{n_{k}^{o} |k - k_{F}|^{\frac{m}{2}} - (1 - n_{k}^{o}) |k - k_{F}|^{\frac{m}{2}} \}.
96: \label{extrap}
97: \eeq
98: What they did not emphasize is that if $m = 2$ and although we have a
99: Luttinger-liquid type of expansion near $k_{F}$ \cite{ml} there is no nonanalytic behavior due to the odd power of $ |k - k_{F}|$ and all trace of the Fermi surface has been eliminated.
100:
101: We can come to the same expressions employing the weakly-screening plasma analogy \cite{meth}, which in considering quantum-mechanical expectations in the state, Eq.(\ref{fjs}), mimics Laughlin's plasma approach \cite{la}. In the Laughlin approach there is the perfect screening of the classical Coulomb plasma, when interaction
102: $- \frac{2 \pi}{|\vec{k}|^{2}}$ becomes screened as
103: \beq
104: \label{screenint}
105: \frac{-\frac{2 \pi}{|\vec{k}|^{2}}}{1 + \frac{2 \pi \beta m^{2}}{|\vec{k}|^{2}} s_{o}(k)},
106: \eeq
107: where $m$ is from $\nu = \frac{1}{m}$, filling factor; $\beta = \frac{2}{m}$ is the plasma inverse temperature, and $s_{o}(k)$ is the static structure factor of the noninteracting particles, in this case bosons, so that $s_{o}(k) = \rho$ - particle density, and hence a perfect screening. More precisely it can be found \cite{thesis} that the expansion in small $m$ of classical statistic averages (to which quantum expectations correspond) is well defined, gives the results that can be found by other methods, and allows continuation to larger than $m = 1$ values. In this context the screening is captured by the accustomed infinite summation of a geometric series described by Eq.(\ref{screenint}) and symbolically can be represented by the sum of diagrams as in Fig. \ref{one}
108: \begin{figure}
109: \includegraphics{one1.eps}
110: %\includegraphics[width=\columnwidth, height=0.25\columnwidth]{one1.eps}
111: \caption{Effective screened interaction}
112: \label{one}
113: \end{figure}
114:
115: In the case of the weakly-screening plasma analogy due to the presence of the free-fermion Slater determinant in (\ref{fjs}), the first summation, (\ref{screenint}), that is done while organizing diagrams, gets modified, having for $s_{o}(k)$ the static structure factor of free fermion gas, which in two dimensions for small $k$ can be found to be $s_{o}^{f}(k) = \frac{3}{4} \frac{k_{F}}{\pi^{2}} k $.
116: This leads to not so perfect screening of the long-range interaction which becomes as $\frac{1}{r}$ instead of $\ln r $ in real space. The approach introduced parallels the RPA calculation in Ref. \cite{bw} in getting (\ref{ls}) when Fig. 1 corresponds to an RPA summation with the value of the bubble equal to $ s_{o}^{f}(k) \beta m^{2}$.
117:
118: We want to see in more detail how the equal-time CF propagator can be found, and, possibly, which additional diagrams in its calculation would lead to the conjectured expression for the CF occupation number. It is instructive to first consider how we can get the equal-time CB correlator i.e. Girvin - MacDonald correlations \cite{gmd} in the Laughlin case using the diagramatic expansion \cite{thesis}. As introduced by Girvin and MacDonald we in fact in the plasma language have to deal with two impurities of charge $\frac{m}{2}$ each, which do not interact directly. Therefore we have
119: \beq
120: \label{bplasma}
121: G_{B}(z,z^{'}) \sim |z - z^{'}|^{-\frac{m}{2}} \frac{Z(z,z^{'})}{Z(z,z)},
122: \eeq
123: where $Z(z,z^{'})$ is the partition function of the classical 2D plasma with inverse temperature $\beta = \frac{2}{m}$, each particle with charge $m$, as before, and two impurities with charge $\frac{m}{2}$ each at the locations $z$ and $z^{'}$. ($Z(z,z)$ is the partition function with one impurity of charge $m$ at an arbitrary location because the value of the partition function does not depend on $z$.) What we expect is that the ratio will have the following form,
124: \beq
125: \label{form}
126: \frac{Z(z,z^{'})}{Z(z,z)} = \exp\{- \beta \Delta f(z,z^{'})\},
127: \eeq
128: where $\Delta f(z,z^{'})$ represents the difference in the free energy between the two configurations. Indeed we can find doing the simple expansion in $m$ that the term right after the first term (of value one) is
129: \beq
130: \label{veff}
131: V_{eff}(|z - z^{'}|) = (\frac{m}{2})^{2} \; \int \frac{d^{2}k}{(2 \pi)^{2}} \exp\{i \vec{k}(\vec{r} - \vec{r}^{'})\}
132: \frac{-\frac{2 \pi}{|\vec{k}|^{2}}}{1 + \frac{2 \pi \beta m^{2}}{|\vec{k}|^{2}} \rho},
133: \eeq
134: $(\rho = \frac{1}{2 \pi m})$, which represents an effective screened interaction between two impurities and extract to mimic (\ref{form}), contributions of disconnected $V_{eff}(|z - z^{'}|)$ parts that follow so that for the final expression we get
135: \beq
136: \label{expr}
137: \frac{Z(z,z^{'})}{Z(z,z)} = \exp\{V_{eff}(|z - z^{'}|) \}.
138: \eeq
139: Therefore we can conclude that in calculating $G_{B}(z,z^{'})$ we have to exponentiate the value of the diagram shown in Fig. \ref{two},
140: \begin{figure}
141: \includegraphics{two1.eps}
142: %\includegraphics[width=\columnwidth, height=0.50\columnwidth]{two1.eps}
143: \caption{Effective diagram contribution}
144: \label{two}
145: \end{figure}
146: and get, due to the screening, the famous algebraic decay.
147:
148: Similarly, applying the same type of approximation we can get in the CF case
149: \beq
150: G_{F}(z,z^{'}) \sim G_{F}^{o}(z,z^{'}) \; \exp \{(\frac{m}{2})^{2}
151: \; \int \frac{d^{2}k}{(2 \pi)^{2}}
152: ( \frac{2 \pi \beta}{|\vec{k}|^{2}}-
153: \frac{2 \pi \beta}{|\vec{k}|^{2}+ 2 \pi \beta m^{2} s_{o}^{f}(k)})
154: \exp\{i \vec{k} (\vec{r}-\vec{r}^{'})\},
155: \label{thgreen}
156: \eeq
157: where (the screening bubble is proportional to the static structure factor of free fermions and) $G_{F}^{o}(z,z^{'}) = \sum_{|\vec{k}|< k_{F}} e^{i \vec{k}(\vec{r}-\vec{r}^{'})}$ is the equal-time correlator of free Fermi gas. To fix the normalization we demand that the total number of CFs is the same as of noninteracting particles so that
158: \beq
159: \sum_{\vec{k}}\; \int d^{2}r \; G_{F}^{o}(\vec{r}) \; e^{- i \vec{k} \vec{r}} = N =
160: \sum_{\vec{k}}\; \int d^{2}r \; G_{F}(\vec{r}) \; e^{- i \vec{k} \vec{r}},
161: \label{equal}
162: \eeq
163: and $G_{F}^{o}(0) = G_{F}(0)$ follows. Therefore
164: \beq
165: G_{F}(z,z^{'}) = G_{F}^{o}(z,z^{'}) \; \exp \{(\frac{m}{2})^{2}
166: \; \int \frac{d^{2}k}{(2 \pi)^{2}}
167: ( \frac{2 \pi \beta}{|\vec{k}|^{2}}-
168: \frac{2 \pi \beta}{|\vec{k}|^{2}+ 2 \pi \beta m^{2} s_{o}^{f}(k)})
169: [\exp\{i \vec{k} (\vec{r}-\vec{r}^{'})\}-1]\}.
170: \label{finalex}
171: \eeq
172:
173: And indeed by taking $n_{k} = \int e^{-i \vec{k} \vec{r}} G_{F}(\vec{r})$ and considering the first nontrivial contribution in the expansion of the exponential in (\ref{finalex}) we get
174: \beq
175: \delta n_{k} = (\frac{m}{2})^{2}\; \int \frac{d^{2}q}{(2 \pi)^{2}}
176: ( \frac{2 \pi \beta}{|\vec{q}|^{2}}-
177: \frac{2 \pi \beta}{|\vec{q}|^{2}+ 2 \pi \beta m^{2} s_{o}^{f}(q)})
178: [n_{\vec{k}-\vec{q}}^{o} (1 - n_{k}^{o})-
179: n_{k}^{o}(1 - n_{\vec{k}-\vec{q}}^{o})],
180: \label{occbw}
181: \eeq
182: exactly the same expression as Eq.(86) of Ref. \cite{bw}. Once we specify that $k$ is near $k_{F}$, assume a flat Fermi surface and neglect the contribution of the (weakly) screened interaction in (\ref{occbw}) we can get, as in \cite{bw}, the leading singularity in $n_{k}$ given by Eq.(\ref{ls}).
183:
184: But unfortunately after a suitable regrouping of free-Fermi-gas occupation numbers we can prove that the second nontrivial contribution in the expansion of the exponential (with the neglect of the screened interaction) is equal to zero. That does not mean that the conjectured contribution (Eq.(\ref{extrap})) is absent. Here we have very likely the situation that due to the nonanlytic nature of the attempted expansion in the CF case we can not generate corrections to the terms linear in $m$. That conclusion supports also the finding \cite{thesis} that when the same expansion was applied in the calculation of the static structure factor of the CF state a first correction to the RPA result could not be generated although it was expected on the grounds that the correction would have made the infered LLL-projected static structure factor positive definite what by its definition it should be.
185:
186: Therefore, very likely the exponential prescription (used to get Eq.(\ref{extrap})) is a
187: valid one although there is no straightforward expansion to prove it. Once we accept the prescription
188: we are left to wonder where is the expected nonanalyticity at $ \nu = \frac{1}{m}, \frac{m}{2} =
189: odd \; integer$ (see Eq.(\ref{extrap}) for that case).
190: A trace of the real Fermi-surface nonanalyticity at $\frac{m}{2}= odd \; integer$ may be seen in the expansion only if we take into account the screened interaction (second part) in Eq.(\ref{occbw}). As an effective contribution from this part we have
191: \beq
192: \delta n_{k} = \frac{1}{2 \pi m s_{o}} [ |\delta k|
193: \ln|\delta k|\; n^{o}(k) - (1 - n^{o}(k))\;
194: |\delta k|
195: \ln|\delta k|],
196: \label{nona}
197: \eeq
198: where $s_{o} = \frac{3}{4} \frac{k_{F}}{\pi^{2}}$.
199: %, although certainly due to the impossibility of finding out further contributions this can not be a final expression with fixed, final coefficients for the nonanalytic behavior of the occupation number near the Fermi surface.
200: If we apply the exponential prescription again, taking also into account this second
201: contribution, we have, with $ \frac{1}{2 \pi m s_{o}} \equiv c$ and, for $\frac{m}{2} = 1$,
202: for the contribution in the vicinity of $k_{F}$,
203: \begin{equation}
204: n_{k} = \frac{1}{2} + \frac{1}{2} [ |\delta k| \exp\{c |\delta k| \ln|\delta k|\} \; n^{o}(k)
205: - ( 1 - n^{o}(k)) |\delta k| \exp\{c |\delta k| \ln|\delta k|\} ].
206: \label{fin}
207: \end{equation}
208: Here a (weak) nonanalyticity is retained. Namely, in Eq.(\ref{fin}) we have singular (at $k_{F}$) the
209: second derivative of $ |\delta k| \exp\{c |\delta k| \ln|\delta k|\} $ with respect to $|\delta k|$.
210: Therefore, a trace of the Fermi surface at $ \nu = \frac{1}{m} , \frac{m}{2}=1 $, is present
211: because of the found nonanalytic behavior. (Such a behavior exists also for $ \frac{m}{2} =
212: odd > 1$ but is weaker having singular higher derivatives.)
213:
214: In the following we will give an example where aforementioned mechanism for getting the Fermi surface (nonanalyticity) does not work due to strong correlations of the CFs with other particles of the system. This is the case of the mixed CB - CF quantum Hall states proposed in \cite{srm}
215: to describe the evolution of the bilayer $\nu = 1$ QH system with distance between layers.
216:
217: If we neglect the LLL projection again and assume that for our purposes we can also neglect the overall antisymmetrization between CB and CF parts that makes the mixed state completely antisymmetric and an electronic wave function, we can write it in the quasiparticle representation as
218: \begin{eqnarray}
219: &\Psi_{o}(z_{\uparrow},z_{\downarrow},w_{\uparrow},w_{\downarrow})= \;\; \; \; \; \; \; \; \; & \nonumber \\
220: & \prod_{i<j} |z_{i \uparrow} - z_{j \uparrow}|^{n}
221: \prod_{k<l} |z_{k \downarrow} - z_{l \downarrow}|^{n}
222: \prod_{p<q} |z_{p \uparrow} - z_{q \downarrow}|^{n} & \nonumber \\
223: & \prod_{i,j} |z_{i \uparrow} - w_{j \uparrow}|^{n}
224: \prod_{k,l} |z_{k \uparrow} - w_{l \downarrow}|^{n}
225: \prod_{p,q} |z_{p \downarrow} - w_{q \uparrow}|^{n}
226: \prod_{r,s} |z_{r \downarrow} - w_{s \downarrow}|^{n} & \nonumber \\
227: & \prod_{i<j} |w_{i \uparrow} - w_{j \uparrow}|^{m}
228: \Psi_{FS}(\{w_{\uparrow}\})
229: \prod_{k<l} |w_{k \downarrow} - w_{l \downarrow}|^{m}
230: \Psi_{FS}(\{w_{\downarrow}\}), &
231: \label{bofe}
232: \end{eqnarray}
233: where $z_{\uparrow}$ and $z_{\downarrow}$ denote CB coordinates and $ w_{\uparrow}$ and $w_{\downarrow}$ denote CF coordinates with arrows specifying to which layer quasiparticles belong. The total numbers of bosons are equal as well the total numbers of fermions, and $n = \frac{m}{2} = $ odd integer.
234:
235: We want to find out (the asymptotic behavior of ) the equal-time
236: correlator of a CF (belonging to one of the layers). It is not hard to conclude that in this case with assumptions similar to the ones done in the single-layer case, we get $G_{F}(w,w^{'})$ by simply taking for the value of the ``polarization'' bubble -
237: \beq
238: \beta m^{2} s_{o}^{f}(k) + \beta n^{2} \rho_{b},
239: \label{sub}
240: \eeq
241: instead of $\beta m^{2} s_{o}^{f}(k)$ only in Eq.(\ref{finalex}), where $\rho_{b}$ denotes the total (up plus down) density of bosons and in $s_{o}^{f}(k)$ we have to take $k_{F} = \sqrt{4 \pi \rho_{f}}$ where $\rho_{f}$ is the density of fermions of one layer only. In this case we work with a completely screened interaction between the two impurities which does not produce nonanalytic contributions.
242:
243: Therefore we find that at the total fillings of bilayer at which we can expect bose-fermi mixed states, $\nu = \frac{2}{m}; \; m = 2,6,\ldots$ the naively expected Fermi surface(s)
244: can not exist due to our analysis.
245: This outcome remind us of the similar disappearance of the small
246: (naively) expected Fermi momentum in the Kondo lattice systems
247: \cite{yoa,osh} due to the Luttinger theorem \cite{lutt}.
248: In the case considered in the paper we do not know for sure if we
249: deal with (overall) Fermi-liquid-like states and a complete analogy
250: (in which CBs and CFs play the roles of
251: of localized spin $1/2$ local moments and conduction electrons respectively)
252: is still missing. Further insights into the physics of the mixed states are
253: necessary.
254: %In a broader sense this conclusion
255: %also complies with the Luttinger theorem for interacting fermions \cite{lutt}
256: %similarly to the generalizations in \cite{yoa} and \cite{osh} in the
257: % Kondo lattice case. In this case of CFs and CBs (instead of
258: %conduction electrons and localized spin $1/2$ local moments)
259: %we generalize, and view the mixed states as Fermi-liquid-like states with
260: %necessarilly large Fermi volume, evolving from two decoupled
261: %Fermi-liquid-like states with the gradual inclusion of
262: %interlayer interaction that preserves the Fermi volume, so that
263: %at the transition from pure (111) state to the mixed region there is a jump
264: %in the Fermi volume. The transition is still continuous as we characterize
265: %the ordering in the CF region as CB $p$-wave pairing \cite{mvm},
266: %so that the number of CFs is still meaningful number continously
267: %changing from zero at the transition point to the number of electrons
268: %when the layers are far apart \cite{pivo}.
269:
270: The author thanks Ashvin Vishwanath for asking the question discussed in the paper,
271: and Aspen Center for Physics for its hospitality. This work was supported by
272: Grant No. 1899 of the Serbian Ministry of Science, Technology, and Development.
273: \begin{references}
274: \bibitem{halp} For an introduction see {\it Perspectives in
275: Quantum Hall Effects}, edited by Sankar Das Sarma and Aron Pinczuk (Wiely, New York, 1997).
276: \bibitem{kelog} M. Kellogg, J.P. Eisenstein, L.N. Pfeiffer, and K.W. West,
277: Phys. Rev. Lett. {\bf 90}, 246801 (2003).
278: %\bibitem{luin} S. Luin, V. Pellegrini, A. Pinczuk, B.S. Dennis, L.N.
279: %Pfeiffer, and K.W. West, Phys. Rev. Lett. {\bf 90}, 236802 (2003).
280: \bibitem{sch} J. Schliemann, S.M. Girvin, and A. H. MacDonald,
281: Phys. Rev. Lett. {\bf 86}, 1849 (2001); D.N. Sheng, L. Balents,
282: and Z. Wang, Phys. Rev. Lett. {\bf 91}, 116802 (2003).
283: \bibitem{stha} A. Stern and B.I. Halperin, Phys. Rev. Lett. {\bf 88}, 106801 (2002).
284: \bibitem{srm} S.H. Simon, E.H. Rezayi, and M.V. Milovanovic, Phys. Rev. Lett. {\bf 91}, 046803 (2003).
285: \bibitem{kimv} Y.B. Kim, C. Nayak, E. Demler, N. Read, and S. Das Sarma,
286: Phys. Rev. B {\bf 63}, 205315 (2001); M. Viellette, L. Balents,
287: and M.P.A. Fisher, Phys. Rev. B {\bf 66} 155401 (2002).
288: \bibitem{hlr} B.I. Halperin, P.A. Lee, and N. Read, Phys. Rev. B {\bf 47}, 7312 (1993).
289: \bibitem{bw} P.-A. Bares and X.-G. Wen, Phys. Rev. B {\bf 48}, 8636 (1993).
290: \bibitem{rr} E. Rezayi and N. Read, Phys. Rev. Lett. {\bf 72}, 900 (1994); {\bf 73}, 1052 (1994).
291: \bibitem{ml} D.C. Mattis and E.H. Lieb, J. Math. Phys. {\bf 6} 375 (1965).
292: \bibitem{meth} M.V. Milovanovi\'{c} and E. Shimshoni, Phys. Rev. B {\bf 59}, 10757 (1999) and references therein.
293: \bibitem{la} R.B. Laughlin, Phys. Rev. Lett. {\bf 50}, 1395 (1983);
294: R.B. Laughlin in {\it The Quantum Hall Effect}, 2nd ed., edited by R.E. Prange and S.M. Girvin (Springer-Verlag, New York, 1990).
295: \bibitem{thesis} M. Milovanovi\'{c}, Ph. D. thesis, Yale University, 1996.
296: \bibitem{gmd} S.M. Girvin and A.H. MacDonald, Phys. Rev. Lett. {\bf 58}, 1252 (1987).
297: \bibitem{yoa} M. Yamanaka, M. Oshikawa, and I. Affleck, Phys. Rev. Lett. {\bf 79}, 1110 (1997).
298: \bibitem{osh} M. Oshikawa, Phys. Rev. Lett. {\bf 84}, 3370 (2000).
299: \bibitem{lutt} J.M. Luttinger, Phys. Rev. {\bf 119}, 1153 (1960).
300: %\bibitem{mvm} M.V. Milovanovi\'{c}, unpublished.
301: %\bibitem{pivo} An interesting example for such a transition
302: %(continous but with a Fermi momentum jump) in a Kondo lattice, onedimensional
303: %system can be found in E. Pivovarov and Q. Si, cond-mat/0304129
304: \end{references}
305: \end{document}