1: \documentclass[aps,prl,preprint,a4paper]{revtex4}
2: %\documentclass[aps,prl,preprint,superscriptaddress]{revtex4}
3: %\documentclass[aps,prl,twocolumn,a4paper]{revtex4}
4: \usepackage{amsmath}
5: \usepackage{amsfonts}
6: \usepackage{graphicx}
7:
8: \begin{document}
9:
10: \preprint{ITP-UH-11/05}
11:
12: \title{Correlation functions of one-dimensional Bose-Fermi mixtures}
13:
14: \author{Holger Frahm}
15: \author{Guillaume Palacios}
16: %\email[]{Your e-mail address}
17: %\homepage[]{Your web page}
18: %\thanks{}
19: %\altaffiliation{}
20: \affiliation{Institut f\"ur Theoretische Physik, Universit\"at Hannover,
21: Appelstr.\ 2, 30167 Hannover, Germany}
22:
23:
24: \date{July 14, 2005}
25:
26: \begin{abstract}
27: We calculate the asymptotic behaviour of correlation functions as a function
28: of the microscopic parameters for a Bose-Fermi mixture with repulsive
29: interaction in one dimension. For two cases, namely polarized and unpolarized
30: fermions the singularities of the momentum distribution functions are
31: characterized as a function of the coupling constant and the relative density
32: of bosons.
33: \end{abstract}
34: \maketitle
35:
36: In the past years, the advances in cooling and trapping of atomic gases have
37: opened the possibility to realize quasi one-dimensional (1D) systems with
38: tunable strength of the interactions in optical lattices. This gives rise to
39: new opportunities for the investigation of the striking phenomena appearing in
40: correlated systems as a consequence of the enhanced quantum fluctuations in
41: reduced spatial dimensions. The observable signatures of these phenomena are
42: encoded in the correlation functions of the system such as the momentum
43: distribution function which can be measured directly in time-of flight
44: experiments or using Bragg spectroscopy \cite{EXPnk}. A setup for the
45: measurement of various density correlation functions has been proposed for the
46: identification of dominant correlations in the atomic gas
47: \cite{CORR}. Theoretical studies of correlation functions in cold
48: atomic gases have been performed both using analytical methods, e.g.\
49: bosonization \cite{GoNeTs:boson} combined with exact results from integrable
50: models such as the Bose-gas with repulsive $\delta$-function interaction
51: \cite{BOSE1d}, and numerically.
52: %
53: New correlation effects appear when the particles considered have internal
54: degrees of freedom. In cold gases containing different constituent atoms
55: Bose-Fermi mixtures can be realized \cite{MIX}. Extensive theoretical results
56: exist for 1D Fermi gases due to their equivalence with the Tomonaga-Luttinger
57: (TL) liquids realized by correlated electrons in 1D lattices
58: \cite{GoNeTs:boson,TheBook}. Only recently, theoretical investigations have
59: been extended to Bose-Fermi mixtures: some correlation functions have been
60: calculated numerically in the strong coupling limit \cite{ChSZ05,ImDe05} where
61: the problem simplifies due to the factorization of the many-particle wave
62: function (see e.g.\ \cite{OgSh90}).
63: %
64: For analytical results on these systems one has to go beyond mean-field
65: approximations and use methods which can capture the strong quantum
66: fluctuations in 1D systems. The phase diagram and certain correlation
67: functions of atomic mixtures have been studied in the Luttinger liquid picture
68: \cite{CaHo03,MatX04}. Without further input, however, these results are
69: limited to the weakly interacting regime since the TL parameters which
70: determine the low-energy theory cannot easily be related to the microscopic
71: parameters describing the underlying gas. Therefore, instabilities predicted
72: within this approach may not appear in a specific realizations
73: \cite{LaYa71,ImDe05,BatX05b}.
74:
75: In this paper we establish the relation between the TL and the microscopic
76: parameters for an integrable Bose-Fermi mixture \cite{LaYa71}. We employ
77: methods from conformal quantum field theory (CFT) to determine the asymptotic
78: (long distance, low-energy) behaviour of correlation functions in the model
79: from a finite size scaling analysis of the exact spectrum obtained by means of
80: the Bethe ansatz. This approach gives the complete set of critical exponents
81: of the model as a function of the parameters in the microscopic Hamiltonian
82: (see e.g. \cite{FrKo,KaYa91,TheBook} for applications to 1D
83: correlated electrons). As an application we compute the momentum distribution
84: function of bosons and fermions in the atomic mixture as a function of their
85: respective densities and the effective coupling constant.
86: %
87: It should be emphasized that our results can be expected to describe the
88: generic (universal) low-energy behaviour of atomic mixtures. Additional
89: interactions -- as long as they do not lead to a phase transition -- will
90: merely change the anomalous exponents but not the qualitative behaviour of the
91: correlation functions.
92:
93: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
94:
95: The 1D Bose-Fermi mixture of $N=M_f+M_b$ particles with repulsive interaction
96: ($c>0$) on a line of length of length $L$ subject to periodic boundary
97: conditions is described by the integrable Hamiltonian \cite{LaYa71}
98: \begin{equation}
99: {\cal H} = -\sum_{i=1}^N \frac{\partial^2}{\partial x_i^2}
100: + 2c\sum_{i<j} \delta(x_i-x_k)\ .
101: \label{hamil}
102: \end{equation}
103: Here $M_f=M_\uparrow+M_\downarrow$ of the particles are fermions carrying spin
104: $\sigma=\uparrow,\downarrow$ and $M_b$ of them are bosons. The many-particle
105: eigenstates of (\ref{hamil}) are parametrized by the solutions of the Bethe
106: ansatz equations (BAE) \cite{LaYa71}
107: \begin{align}
108: & \exp({iq^{(0)}_j L}) = \prod_{k=1}^{M_1} e_c(q^{(0)}_j-q^{(1)}_k)
109: \notag\\
110: & \prod_{j=1}^{M_0} e_c(q^{(1)}_k-q^{(0)}_j)
111: \prod_{\ell=1}^{M_2} e_c(q^{(1)}_k-q^{(2)}_\ell) =
112: \prod_{k'\ne k}^{M_1} e_{2c}(q^{(1)}_{k}-q^{(1)}_{k'})
113: \notag\\
114: & \prod_{k=1}^{M_1} e_c(q^{(2)}_\ell-q^{(1)}_k) = 1
115: \label{bae}
116: \end{align}
117: where $e_a(x) = (x+ia/2)/(x-ia/2)$ and $M_0=N$, $M_1=N-M_\uparrow$,
118: $M_2=M_b$. The corresponding eigenvalue of (\ref{hamil}) is $E=\sum_{j=1}^N
119: (q_j^{(0)})^2$.
120: %
121: In the thermodynamic limit $L\to\infty$ with $M_i/L$ kept fixed, the root
122: configurations $\{q^{(i)}\}$ of Eqs.~(\ref{bae}) can be described by
123: distribution functions $\rho_i$ which, as a consequence of (\ref{bae}) are
124: solutions to \cite{LaYa71}
125: \begin{equation}
126: \rho_i(x) = \rho^{(0)}_i + \sum_j (\hat{K}_{ij}\otimes\rho_j)(x)
127: \label{irho}
128: \end{equation}
129: Here $\rho^{(0)}_i = (c/2\pi) \delta_{i0}$ and $\hat{K}_{ij}$ are linear
130: integral operators acting as ($\int_j \equiv \int_{-Q_j}^{Q_j} {\mathrm{d}}y$)
131: \begin{equation}
132: (\hat{K}_{ij} \otimes f)(x) =
133: \int_j k_{ij}(x-y) f(y)\ .
134: \end{equation}
135: The kernels of these integral operators are $k_{ij}(x) = a_1(x)
136: \delta_{|i-j|,1} - a_2(x) \delta_{i,1}\delta_{j,1}$
137: %%\begin{align}
138: %% k_{ij}(x) = \left(\begin{array}{crc} 0 & a_1(x) & 0 \\
139: %% a_1(x) & -a_2(x) & a_1(x)\\ 0 & a_1(x) & 0
140: %% \end{array} \right)
141: %%\end{align}
142: where $2\pi a_n(x) = 4n/(4x^2+n^2)$. The properties of the system are
143: completely characterized by the densities $m_i\equiv M_i/L = \int_i
144: \rho^{(i)}(x)$ of the components of the mixture (these relations determine the
145: boundaries $Q_i$ of the above integral equations) and the dimensionless
146: coupling strength $\gamma = Lc/N$. For later use we also introduce the
147: fraction of bosons in the system $\alpha=M_b/N$.
148:
149: %%It should be stressed that although the $\{q^{(i)}\}$ parametrize the
150: %%many-particle eigenstates of the system (\ref{hamil}) they do not correspond
151: %%to physical observables. In particular, the distribution functions $\rho_i$
152: %%are not correlation functions of local operators.
153:
154: Generically, i.e.\ for $m_i>0$, there are three modes of collective elementary
155: excitations above the many-particle ground state of (\ref{hamil}).
156: %
157: Their dispersion $\epsilon_i(k)$
158: %are obtained from the Bethe ansatz within the
159: %dressed energy formalism. For low energies the dispersion
160: is linear at low energies with different sound velocities $v_i=\partial
161: \epsilon_i/\partial k$, $i=0,1,2$ \cite{BatX05b}. These quantities determine
162: the finite size scaling behaviour of the ground state energy
163: \begin{equation}
164: E_0 - L\epsilon_\infty = -\frac{\pi}{6L} \sum_i v_i +
165: \mathit{o}\left(\frac{1}{L}\right)\ .
166: \label{fs0}
167: \end{equation}
168: Physical excitations of the system are combinations of the elementary ones.
169: Due to the interacting nature of the system the different modes are coupled
170: and excitations in one of the modes shift the energies in the other ones. In
171: general, this effect can be described in terms of generalized susceptibilities
172: which may be determined in an experiment or numerically from studies of small
173: systems \cite{TheBook}. For the Bethe ansatz solvable models it is possible
174: to describe the coupling of the modes in terms of the dressed charge matrix
175: \cite{DCREF} which in this case reads
176: \begin{equation}
177: Z_{ij} = \xi_{ij}(Q_i)\ .
178: \label{zmatrix}
179: \end{equation}
180: The functions $\xi_{ij}$ are given in terms of integral equations
181: %very similar to the ones (\ref{irho}) for the distribution functions
182: %introduced above
183: %\begin{equation}
184: $\xi_{ij}(x) = \delta_{ij} + \sum_k (\hat{K}_{ik}\otimes\xi_{kj})(x)$.
185: %\label{idc}
186: %\end{equation}
187:
188: $Z$ determines the general form of the finite size corrections to the energies
189: of low-lying excitations
190: \begin{align}
191: &\Delta E(\Delta\mathbf{M},\mathbf{D}) = {2\pi\over L}\Bigl(
192: {1\over4} \Delta \mathbf{M}^\top
193: \left(Z^\top\right)^{-1} V Z^{-1} \Delta \mathbf{M}
194: \notag\\
195: & + \mathbf{D}^\top Z V Z^{\top} \mathbf{D}
196: + \sum_{k} v_k \left(N_k^+ + N_k^-\right)\Bigr)
197: +\mathit{o}\left(1\over L\right)\ .
198: \label{fse}
199: \end{align}
200: Here, $V=\mathrm{diag}(v_0,v_1,v_2)$ is a $3\times3$ matrix of the sound
201: velocities, $N_k^\pm$ are non-negative integers, $\Delta M$ is a vector of
202: integers denoting the change of $M_i$ with respect to the ground state for
203: charged excitations. The $D_i$ are integers or half-odd integers according to
204: \begin{align}
205: D_0 \sim \left( \Delta M_0+\Delta M_1\right)/2
206: = \Delta M_\uparrow/2 \mod 1
207: \notag\\
208: D_1 \sim \left( \Delta M_0+\Delta M_2\right)/2
209: = \Delta M_f/2 \mod 1
210: \label{selrul}\\
211: D_2 \sim \left( \Delta M_1+\Delta M_2\right)/2
212: = \Delta M_\downarrow/2 \mod 1
213: \notag
214: \end{align}
215: and enumerate finite momentum transfer processes:
216: \begin{multline}
217: \Delta P(\Delta\mathbf{M},\mathbf{D}) = {2\pi\over L} \left(
218: \Delta \mathbf{M}^\top \cdot {\mathbf D} +
219: \sum_{k}\left(N_k^+ - N_k^-\right)\right) \\
220: \qquad + 2 {k}_{F,\uparrow} D_0 + 2
221: {k}_{F,\downarrow}\left( D_0 + D_1\right) + 2 k_B \sum_j D_j\ .
222: \label{fsp}
223: \end{multline}
224: Here $k_{F,\sigma} = \pi M_\sigma/L$ are the Fermi momenta of the fermion
225: components, $k_B = \pi M_b/L$ is the corresponding quantity for the
226: interacting bosons.
227:
228: In the framework of CFT \cite{BPZ} the finite size spectrum (\ref{fs0}),
229: (\ref{fse}) can be understood as that of a critical theory based on the
230: product of three Virasoro algebras each having central charge $C=1$
231: \cite{FrKo,TheBook}. Correlation functions of a general operator in the
232: theory -- characterized by the quantum numbers $\Delta M_i$ and $D_i$ -- will
233: contain contributions from these three sectors. The simplest ones, analogues
234: of primary fields in the CFT, have correlation functions (in Euclidean time
235: $\tau$)
236: \begin{widetext}
237: \begin{equation}
238: \left\langle \phi_{\Delta}(x,\tau) \phi_{\Delta}(0,0) \right\rangle =
239: \frac{\exp\left(2i D_0 {k}_{F,\uparrow}x
240: +2i \left(D_0+D_1\right) {k}_{F,\downarrow}x
241: +2i \left(\sum_j D_j\right) k_B x\right)
242: }{\prod_{k} (v_k\tau +ix)^{2\Delta^+_k}(v_k\tau -ix)^{2\Delta^-_k}}\ .
243: \label{corr0}
244: \end{equation}
245: \end{widetext}
246: The operators $\phi_{\Delta}$ are characterized by their scaling dimensions
247: $\Delta_k^\pm$ in the chiral (left- and right moving) components of all three
248: constituent theories. The latter are uniquely determined from the finite size
249: energies (\ref{fse}) and momenta (\ref{fsp}) and form towers starting at
250: \begin{align}
251: 2\Delta_k^\pm = \left( \sum_j Z_{kj} D_j \pm
252: \frac{1}{2}\sum_j \Delta M_j (Z^{-1})_{jk} \right)^2 .
253: \label{cdim}
254: \end{align}
255: The asymptotic exponential decay of correlation functions in a large but
256: finite system or at finite temperature $T$ can be obtained from (\ref{corr0})
257: by conformal invariance. For example, at $T>0$ the denominators in
258: (\ref{corr0}) have to be replaced by by
259: %\begin{equation}
260: $
261: (v_k\tau \pm ix)^{-2\Delta^\pm_k} \to
262: \left({\pi T}/v_k{\sin \pi T (\tau \pm ix/v_k)}\right)^{2\Delta^\pm_k}
263: $.
264: %\end{equation}
265:
266:
267: With (\ref{cdim}) the critical exponents which determine the long-distance
268: asymptotics of \emph{any} correlation function are known as soon as we have
269: computed the dressed charge matrix (\ref{zmatrix}).
270: %
271: To calculate the correlation functions of a given local operator ${\cal O}$ in
272: the microscopic theory (\ref{hamil}) one needs to know its expansion in terms
273: of the fields $\phi_{\Delta}$ of the CFT. Usually, this expansion is not
274: known but ${\cal O}$ and $\phi_{\Delta}$ have to generate the same set of
275: selection rules in calculating the correlation function. This drastically
276: reduces the number of possible terms in the expansion: As an example consider
277: the bosonic Green's function $G_b(x,\tau)= \langle \Psi_b(x,\tau)
278: \Psi_b^\dagger(0,0)\rangle$: clearly $\Psi_b^\dagger$ generates a state with
279: $\Delta M_b=1$ which implies $\Delta M_j\equiv1$ in (\ref{cdim}). By
280: (\ref{selrul}) the quantum numbers $D_j$ are further restricted to integers:
281: the uniform part of the $G_b(x) \sim |x|^{-1/2K_b}$ is described by the
282: operator with $D_j\equiv0$ which allows to identify the TL parameter $K_b$
283: \cite{HALD} from (\ref{cdim}). The interactions lead to additional
284: contributions to $G_b$ oscillating with wave numbers $k_0=2k_{F\sigma}$,
285: $2k_B$, \ldots
286:
287: For a comparison with experimental data one is often interested in Fourier
288: transforms of the two-point correlation functions given above. The large
289: distance behaviour of (\ref{corr0}) determines the singularities of spectral
290: functions near $\omega\approx \pm v_k(k-k_0)$ (see e.g.\ \cite{TheBook}).
291: %
292: Quantities accessible in experiments with cold gases \cite{EXPnk} are
293: the momentum distribution functions of the constituent particles. For the
294: bosons this is the Fourier transform of the equal time Green's function
295: $G_b(x)$. From (\ref{corr0}) its singularities at wave numbers $k_0$ are then
296: $n_b(k) \sim |k-k_0|^{\nu_b}$ near $k\approx k_0$. The exponent $\nu_b$ is
297: the minimal value of $2\left(\sum_k \Delta_k^++\Delta_k^- \right) -1$
298: compatible with the quantum numbers $\Delta M$ and the selection rules for the
299: $D$ for the given $k_0$, e.g.\ $1/2K_b=\nu_b+1=\frac{1}{4}
300: \sum_k(\sum_j (Z^{-1}_{jk}))^2$ for $k_0=0$.
301:
302: Using the same procedure for the fermionic Green's function $G_\sigma(x,\tau)
303: = \langle\Psi_\sigma(x,\tau)\Psi_\sigma^\dagger(0,0)\rangle$ we find that
304: their asymptotic behaviour is determined by the conformal fields with $\Delta
305: M_0=1$, $\Delta M_1 = \Delta M_2 = 0$, half-odd integers $D_0$, $D_1$ and
306: integer $D_2$ for $G_\uparrow$ and $\Delta M_0=1=\Delta M_1$, $\Delta M_2 =
307: 0$, half-odd integers $D_1$, $D_2$ and integer $D_0$ for $G_\downarrow$.
308: Again, the singularities of the fermions' distribution functions $n_\sigma(k)$
309: follow from (\ref{corr0}).
310: % Due to the half-integer conformal spin of the Fermi fields the behaviour
311: % near
312: Near $k_0-k_{F\sigma}=0$, $\pm 2k_B$, \ldots they are given by
313: \begin{equation}
314: n_\sigma(k) \sim \mathrm{sign}(k-k_0)|k-k_0|^{\nu_f},\quad
315: \mathrm{for~}k\approx k_0\ .
316: \end{equation}
317: $\nu_f$ is related to the dimensions (\ref{cdim}) for the quantum
318: numbers $\Delta M$ and $D$ as $\nu_b$ above. The Fermi distribution of
319: non-interacting particles corresponds to $\nu_f(k_{F\sigma})=0$.
320:
321: %The integral equations for the densities and the dressed charge introduced
322: %above can be reduced to pairs of coupled ones in two cases of particular
323: %relevance \cite{LaYa71,BatX05b}: (i) the unpolarized case where
324: %$M_\uparrow=M_\downarrow=M_f/2$ and the ground state of the system is
325: %invariant under rotations in the spin index of the fermions and (ii) the fully
326: %polarized case where there is only one spin component of the fermions.
327:
328: In the following we consider two cases of particular relevance
329: \cite{LaYa71,MatX04,BatX05b}, namely (i) the unpolarized case where
330: $M_\uparrow=M_\downarrow=M_f/2$ and the ground state of the system is
331: invariant under rotations in the spin index of the fermions and (ii) the fully
332: polarized case where there is only one spin component of the fermions.
333:
334:
335: \paragraph{The unpolarized gas.}
336: %
337: For $Q_1=\infty$ one obtains $M_\uparrow=M_\downarrow$ from (\ref{irho}),
338: %corresponding to a gas
339: i.e.\ with vanishing net magnetization. In this case the dressed charge
340: matrix (\ref{zmatrix}) takes the form
341: \begin{equation}
342: Z = \frac{1}{2}\left(\begin{array}{ccc}
343: 2\zeta_{00} & (\zeta_{00}+\zeta_{01}) & 2\zeta_{01} \\
344: 0 & {\sqrt{2}} & 0 \\
345: 2\zeta_{10} & (\zeta_{10}+\zeta_{11}) & 2\zeta_{11}
346: \end{array}\right)\ .
347: \label{zmat00}
348: \end{equation}
349: Here the Wiener-Hopf method has been used to determine $Z_{11}=1/\sqrt{2}$ and
350: $\zeta_{0j} = \zeta_{0j}(Q_0)$, $\zeta_{1j} = \zeta_{1j}(Q_2)$. The functions
351: $\zeta_{ij}(x)$ are given by
352: \begin{equation}
353: \zeta_{ij}(x) = \delta_{ij} + \int_0 R(x-y) \zeta_{0j}(y)
354: + \int_2 R(x-y) \zeta_{1j}(y)\
355: \label{dc_up}
356: \end{equation}
357: %
358: with $R(x) = (1/\pi) \int_0^\infty \mathrm{d}\omega\ \mathrm{e}^{-|\omega|/2}
359: \cos(\omega x)/\cosh(\omega/2)$. Using (\ref{zmat00}) the scaling dimensions
360: $\Delta_1^\pm$ in (\ref{cdim})
361: %%\begin{align}
362: %% \Delta_1^\pm = \left(\frac{1}{2}D_1
363: %% \mp \frac{1}{4}(\Delta M_\uparrow
364: %% -\Delta M_\downarrow)\right)^2 + N_1^\pm\
365: %%\end{align}
366: are independent on the remaining system parameters, i.e.\ the effective
367: coupling $\gamma$ and the bosonic fraction $\alpha$. This a consequence of
368: the $SU(2)$ invariance of the system in this case. The mode $\epsilon_1(k)$
369: is the spinon mode of the unpolarized system, the CFT describing its
370: low-energy properties is an $[SU(2)]_1$ Wess-Zumino-Witten model.
371: %
372:
373: Additional simplifications arise in the strong coupling limit
374: $\gamma\to\infty$ (i.e.\ $Q_0\to0$) where $\zeta_{00}=1$, $\zeta_{10}=0$,
375: $\zeta_{10}=\zeta_{11}(0)-1=\alpha$ and $\zeta_{11}(x)$ is given by a scalar
376: integral equation resulting from (\ref{dc_up}).
377: %
378: In Fig.~\ref{fig-up} we present results obtained from the numerical solution
379: of these integral equations for the exponents which determine the
380: singularities of the momentum distribution functions for bosons at $k=0$ and
381: fermions at $k=k_F$ as a function of the bosonic fraction $\alpha$ for various
382: values of $\gamma$. The exponents $\nu_{b,f}$ at the other wave numbers are
383: always larger than 1.
384: %
385: \begin{figure}
386: \includegraphics[width=0.5\textwidth]{unpol.eps}
387: \caption{
388: \label{fig-up}
389: Exponents characterizing the singularities in the bosonic (upper panel) and
390: fermionic (lower panel) momentum distribution function for the unpolarized
391: gas at $k=0$ and $k_F$, respectively, as a function of the bosonic fraction
392: $\alpha$ in the mixture for $\gamma=0.2$, $1.0$, $5.0$, $25.0$, $\infty$
393: (bottom to top).}
394: \end{figure}
395: %
396: Note that the system is in a different universality class for $\alpha=0$ or
397: $1$. At $\alpha=0$, all particles are fermionic and the critical exponents
398: are those of the 1D Fermi gas \cite{TheBook}. Here the exponent $\nu_f$ for
399: the singularity at the Fermi point varies between $0$ and $1/8$ as a function
400: of $\gamma$. On the other hand, the limit of $\nu_b$ as $\alpha\to1$ gives
401: exactly the exponent of the 1D Bose gas with $\delta$-interaction \cite{HALD}.
402:
403:
404: \paragraph{The spin-polarized gas.}
405: %
406: Setting $Q_2=\infty$ in (\ref{irho}) corresponds to $M_\downarrow=0$. This
407: case has been discussed recently in Ref.~\onlinecite{ImDe05} where some
408: correlation functions have been computed numerically in the strong coupling
409: limit.
410: %Integrating out the $2$-components the integral equations for the
411: %distribution function and the $2\times2$ dressed charge reduce to coupled
412: %pairs with
413: %\begin{equation}
414: % \hat{K} = \left(\begin{array}{cc}
415: % 0 & \int_1 a_1(x-y)\,\cdot \\
416: % \int_0 a_1(x-y)\,\cdot & 0
417: % \end{array} \right)
418: %\end{equation}
419: In this case, the finite size spectrum and the scaling dimensions are
420: determined by two gapless modes. Again, the equations simplify in the strong
421: coupling limit where all exponents can be given as a function of $\alpha$
422: directly, e.g.\ $\nu_b(0)=\alpha^2/2-\alpha$ and
423: $\nu_f(k_F)=\alpha^2-\alpha+1/2$ for the dominant singularities of the bosonic
424: and fermionic momentum distribution functions, respectively (see
425: Fig.~\ref{fig-pp} for the $\gamma$ dependence).
426: %
427: \begin{figure}
428: \includegraphics[width=0.5\textwidth]{fmpol.eps}
429: \caption{
430: \label{fig-pp}
431: Same as Fig.~\ref{fig-up} for a mixture with polarized fermions.}
432: \end{figure}
433: %
434: While the dependence of $\nu_b$ on $\alpha$ is similar to the one found in the
435: unpolarized case, the strong coupling behaviour of $\nu_f$ at small bosonic
436: fraction is seen to be very different.
437: %
438: Note that the singularity at $k_F+2k_B$ becomes very pronounced for
439: sufficiently small $\alpha$ ($\nu_f=\alpha^2+\alpha+1/2$ at strong coupling).
440: This feature of the fermionic distribution function is a direct signature of
441: the interaction and should be observable in experiments.
442:
443: In summary we have used predictions from CFT on the finite size scaling of the
444: low-energy spectrum to compute the critical properties of a 1D Bose-Fermi
445: mixture for the integrable model (\ref{hamil}). In the generic case there are
446: three linearly dispersing modes which determine the low-energy effective
447: theory and the asymptotic behaviour of the correlation functions. Within this
448: formalism we have related the critical exponents directly to the parameters
449: describing the microscopic Hamiltonian, i.e.\ the coupling strength, the
450: fraction of bosons and polarization of the fermions.
451: %
452: The approach can be used to investigate the phase diagram of the 1D mixture by
453: identifying the order parameter with the slowest long-distance decay of its
454: correlation functions (smallest exponent) \cite{CaHo03,MatX04}. Within the
455: integrable model there is no instability leading to a phase transition
456: \cite{LaYa71,BatX05b}. The general expression (\ref{cdim}), however, allows
457: to obtain estimates on the exponents in a more general system based on
458: numerical data on the spectrum of finite systems.
459:
460: \begin{acknowledgments}
461: This work has been supported by the Deutsche Forschungsgemeinschaft.
462: \end{acknowledgments}
463:
464: % Create the reference section using BibTeX:
465: %\bibliography{base,frahm,books,coldgas}
466: \begin{thebibliography}{26}
467: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
468: \expandafter\ifx\csname bibnamefont\endcsname\relax
469: \def\bibnamefont#1{#1}\fi
470: \expandafter\ifx\csname bibfnamefont\endcsname\relax
471: \def\bibfnamefont#1{#1}\fi
472: \expandafter\ifx\csname citenamefont\endcsname\relax
473: \def\citenamefont#1{#1}\fi
474: \expandafter\ifx\csname url\endcsname\relax
475: \def\url#1{\texttt{#1}}\fi
476: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
477: \providecommand{\bibinfo}[2]{#2}
478: \providecommand{\eprint}[2][]{\url{#2}}
479:
480: \bibitem{EXPnk}
481: %\bibitem[{\citenamefont{Paredes et~al.}(2004)\citenamefont{Paredes, Widera,
482: % Murg, Mandel, F{\"o}lling, Cirac, Shlyapnikov, H\"ansch, and
483: % Bloch}}]{ParX04}
484: \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Paredes}} et~al.,
485: \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{429}}, \bibinfo{pages}{277
486: } (\bibinfo{year}{2004});
487: %
488: %\bibitem[{\citenamefont{Richard et~al.}(2003)\citenamefont{Richard, Gerbier,
489: % Thywissen, Hugbart, Bouyer, and Aspect}}]{RicX03}
490: \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Richard}} et~al.,
491: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{91}},
492: \bibinfo{pages}{010405} (\bibinfo{year}{2003}).
493:
494: \bibitem{CORR}
495: %\bibitem[{\citenamefont{Altman et~al.}(2004)\citenamefont{Altman, Demler, and
496: % Lukin}}]{AlDL04}
497: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Altman}},
498: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Demler}}, \bibnamefont{and}
499: \bibinfo{author}{\bibfnamefont{M.~D.} \bibnamefont{Lukin}},
500: \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{70}},
501: \bibinfo{pages}{013603} (\bibinfo{year}{2004});
502: %
503: %\bibitem[{\citenamefont{Mathey et~al.}()\citenamefont{Mathey, Altman, and
504: % Vishwanath}}]{MaAV05}
505: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Mathey}},
506: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Altman}}, \bibnamefont{and}
507: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Vishwanath}},
508: \bibinfo{note}{preprint}, \eprint{cond-mat/0507108}.
509:
510: \bibitem[{\citenamefont{Gogolin et~al.}(1998)\citenamefont{Gogolin, Nersesyan,
511: and Tsvelik}}]{GoNeTs:boson}
512: \bibinfo{author}{\bibfnamefont{A.~O.} \bibnamefont{Gogolin}},
513: \bibinfo{author}{\bibfnamefont{A.~A.} \bibnamefont{Nersesyan}},
514: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Tsvelik}},
515: \emph{\bibinfo{title}{Bosonization and Strongly Correlated Systems}}
516: (\bibinfo{publisher}{Cambridge University Press},
517: \bibinfo{address}{Cambridge}, \bibinfo{year}{1998}).
518:
519: \bibitem{BOSE1d}
520: %\bibitem[{\citenamefont{Lieb and Liniger}(1963)}]{LiLi63}
521: \bibinfo{author}{\bibfnamefont{E.~H.} \bibnamefont{Lieb}} \bibnamefont{and}
522: \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Liniger}},
523: \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{130}},
524: \bibinfo{pages}{1605} (\bibinfo{year}{1963});
525: %
526: %\bibitem[{\citenamefont{Lieb}(1963)}]{Lieb63}
527: \bibinfo{author}{\bibfnamefont{E.~H.} \bibnamefont{Lieb}},
528: \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{130}},
529: \bibinfo{pages}{1616} (\bibinfo{year}{1963}).
530:
531: \bibitem{MIX}
532: %\bibitem[{\citenamefont{M{\o}lmer}(1998)}]{Molm98}
533: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{M{\o}lmer}},
534: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{80}},
535: \bibinfo{pages}{1804} (\bibinfo{year}{1998});
536: %
537: %\bibitem[{\citenamefont{Lewenstein et~al.}(2004)\citenamefont{Lewenstein,
538: % Santos, Baranov, and Fehrmann}}]{LewX04}
539: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Lewenstein}} et~al.,
540: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{92}},
541: \bibinfo{pages}{050401} (\bibinfo{year}{2004}).
542:
543: \bibitem[{\citenamefont{Essler et~al.}(2005)\citenamefont{Essler, Frahm,
544: G\"ohmann, Kl\"umper, and Korepin}}]{TheBook}
545: \bibinfo{author}{\bibfnamefont{F.~H.~L.} \bibnamefont{Essler}} et~al.,
546: \emph{\bibinfo{title}{The One-Dimensional {H}ubbard
547: Model}} (\bibinfo{publisher}{Cambridge University Press},
548: \bibinfo{address}{Cambridge}, \bibinfo{year}{2005}).
549:
550: \bibitem[{\citenamefont{Imambekov and Demler}()}]{ImDe05}
551: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Imambekov}} \bibnamefont{and}
552: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Demler}},
553: \bibinfo{note}{preprint}, \eprint{cond-mat/0505632}.
554:
555: \bibitem[{\citenamefont{Cheianov et~al.}(2005)\citenamefont{Cheianov, Smith,
556: and Zvonarev}}]{ChSZ05}
557: \bibinfo{author}{\bibfnamefont{V.~V.} \bibnamefont{Cheianov}},
558: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Smith}}, \bibnamefont{and}
559: \bibinfo{author}{\bibfnamefont{M.~B.} \bibnamefont{Zvonarev}},
560: \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{71}},
561: \bibinfo{pages}{033610} (\bibinfo{year}{2005}).
562:
563: \bibitem[{\citenamefont{Ogata and Shiba}(1990)}]{OgSh90}
564: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Ogata}} \bibnamefont{and}
565: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Shiba}},
566: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{41}},
567: \bibinfo{pages}{2326} (\bibinfo{year}{1990}).
568:
569: \bibitem[{\citenamefont{Cazalilla and Ho}(2003)}]{CaHo03}
570: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Cazalilla}}
571: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.~F.} \bibnamefont{Ho}},
572: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{91}},
573: \bibinfo{pages}{150403} (\bibinfo{year}{2003});
574:
575: \bibitem[{\citenamefont{Mathey et~al.}(2004)\citenamefont{Mathey, Wang,
576: Hofstetter, Lukin, and Demler}}]{MatX04}
577: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Mathey}} et~al.,
578: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{93}},
579: \bibinfo{pages}{120404} (\bibinfo{year}{2004}).
580:
581: \bibitem[{\citenamefont{Lai and Yang}(1971)}]{LaYa71}
582: \bibinfo{author}{\bibfnamefont{C.~K.} \bibnamefont{Lai}} \bibnamefont{and}
583: \bibinfo{author}{\bibfnamefont{C.~N.} \bibnamefont{Yang}},
584: \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{3}},
585: \bibinfo{pages}{393} (\bibinfo{year}{1971}).
586:
587: \bibitem[{\citenamefont{Batchelor et~al.}()\citenamefont{Batchelor, Bortz,
588: Guan, and Oelkers}}]{BatX05b}
589: \bibinfo{author}{\bibfnamefont{M.~T.} \bibnamefont{Batchelor}} et~al.,
590: \bibinfo{note}{preprint}, \eprint{cond-mat/0506478}.
591:
592: %\bibitem[{\citenamefont{Frahm and Korepin}(1990)}]{FrKo90}
593:
594: \bibitem[{\citenamefont{Frahm and Korepin}(1990)}]{FrKo}
595: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Frahm}} \bibnamefont{and}
596: \bibinfo{author}{\bibfnamefont{V.~E.} \bibnamefont{Korepin}},
597: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{42}},
598: \bibinfo{pages}{10553} (\bibinfo{year}{1990});
599: %\bibitem[{\citenamefont{Frahm and Korepin}(1991)}]{FrKo91}
600: %\bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Frahm}} \bibnamefont{and}
601: % \bibinfo{author}{\bibfnamefont{V.~E.} \bibnamefont{Korepin}},
602: \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{43}},
603: \bibinfo{pages}{5653} (\bibinfo{year}{1991}).
604:
605: \bibitem[{\citenamefont{Kawakami and Yang}(1991)}]{KaYa91}
606: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Kawakami}} \bibnamefont{and}
607: \bibinfo{author}{\bibfnamefont{S.-K.} \bibnamefont{Yang}},
608: \bibinfo{journal}{J. Phys. Condens. Matter} \textbf{\bibinfo{volume}{3}},
609: \bibinfo{pages}{5983} (\bibinfo{year}{1991}).
610:
611:
612: \bibitem{DCREF}
613: %\bibitem[{\citenamefont{Izergin et~al.}(1989)\citenamefont{Izergin, Korepin,
614: % and Reshetikhin}}]{izkr:89}
615: \bibinfo{author}{\bibfnamefont{A.~G.} \bibnamefont{Izergin}},
616: \bibinfo{author}{\bibfnamefont{V.~E.} \bibnamefont{Korepin}},
617: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{N.~{\ Yu}.}
618: \bibnamefont{Reshetikhin}}, \bibinfo{journal}{J. Phys. A}
619: \textbf{\bibinfo{volume}{22}}, \bibinfo{pages}{2615} (\bibinfo{year}{1989});
620: %\bibitem[{\citenamefont{Woynarovich}(1989)}]{Woyn89}
621: \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Woynarovich}},
622: \bibinfo{journal}{J. Phys. A} \textbf{\bibinfo{volume}{22}},
623: \bibinfo{pages}{4243} (\bibinfo{year}{1989});
624: %\bibitem[{\citenamefont{Frahm and Yu}(1990)}]{FrYu90}
625: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Frahm}} \bibnamefont{and}
626: \bibinfo{author}{\bibfnamefont{N.-C.} \bibnamefont{Yu}}, \bibinfo{journal}{J.
627: Phys. A} \textbf{\bibinfo{volume}{23}}, \bibinfo{pages}{2115}
628: (\bibinfo{year}{1990}).
629:
630: \bibitem[{\citenamefont{Belavin et~al.}(1984)\citenamefont{Belavin, Polyakov,
631: and Zamolodchikov}}]{BPZ}
632: \bibinfo{author}{\bibfnamefont{A.~A.} \bibnamefont{Belavin}},
633: \bibinfo{author}{\bibfnamefont{A.~M.} \bibnamefont{Polyakov}},
634: \bibnamefont{and} \bibinfo{author}{\bibfnamefont{A.~B.}
635: \bibnamefont{Zamolodchikov}}, \bibinfo{journal}{Nucl. Phys. B}
636: \textbf{\bibinfo{volume}{241}}, \bibinfo{pages}{333} (\bibinfo{year}{1984}).
637:
638: \bibitem{HALD}
639: %\bibitem[{\citenamefont{Haldane}(1981)}]{Hald81d}
640: \bibinfo{author}{\bibfnamefont{F.~D.~M.} \bibnamefont{Haldane}},
641: \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{47}},
642: \bibinfo{pages}{1840} (\bibinfo{year}{1981});
643: %\bibitem[{\citenamefont{Cazalilla}(2004)}]{Caza04}
644: \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Cazalilla}},
645: \bibinfo{journal}{J. Phys. B} \textbf{\bibinfo{volume}{37}},
646: \bibinfo{pages}{S1} (\bibinfo{year}{2004}).
647:
648:
649: \end{thebibliography}
650:
651: \end{document}
652:
653: