cond-mat0412701/Crt.tex
1: \documentclass[12pt]{article}
2: \usepackage{epsf}
3: \usepackage{epsfig}
4: \usepackage{wrapfig}
5: \usepackage{amsmath}
6: \usepackage{amssymb}
7: \usepackage{eufrak}
8: \usepackage{graphicx}
9: \textwidth=160mm
10: \textheight=238mm
11: \setlength{\voffset}{-20mm}
12: \setlength{\hoffset}{-5mm}
13: \setcounter{section}{0}
14: \setcounter{subsection}{0}
15: \setcounter{equation}{0}
16: \setcounter{figure}{0}
17: \setcounter{footnote}{0}
18: \setcounter{table}{0}
19: %  \usepackage[CP866]{inputenc}
20: %  \usepackage[russiane]{babel}
21: %  \usepackage{pscyr}
22: %  \usepackage[english,russian]{babel}
23:   \renewcommand{\rmdefault}{ftm} % Times
24: \begin{document}
25: 
26: 
27: 
28: 
29: 
30: 
31: 
32: \begin{center}
33: \section*{Collapse of Bose component in Bose-Fermi mixture with attraction between components}
34: 
35: 
36: \vskip 5mm S.-T. Chui$^{1}$, V. N. Ryzhov$^{2}$ and E. E.
37: Tareyeva$^{2}$ \vskip 5mm
38: 
39: {\small (1) {\it Bartol Research Institute, University of
40: Delaware, Newark, DE 19716}
41: \\
42: (2) {\it Institute for High Pressure Physics, Russian Academy of
43: Sciences, 142 190 Troitsk, Moscow region, Russia}
44: 
45: 
46: }
47: \end{center}
48: \vskip 5mm
49: 
50: \begin{abstract}
51: An effective Hamiltonian for the Bose system in the mixture of
52: ultracold atomic clouds of bosons and fermions is obtained by
53: integrating out the Fermi degrees of freedom. An instability of
54: the Bose system is found in the case of attractive interaction
55: between components in good agreement with the experiment on the
56: bosonic $^{87}$Rb and fermionic $^{40}$K mixture.
57: \end{abstract}
58: \vskip 8mm
59: 
60: The seminal paper by N.N. Bogoliubov "To the theory of
61: superffluidity", published in 1947, had a great influence on the
62: development of the modern theory of quantum many-body systems. It
63: is cited in all textbooks devoted to this topic. However, for a
64: long time the real experimental systems which could be
65: quantitatively described by the model of dilute Bose gas developed
66: in this article were absent. Only about 50 years later, the
67: Bose-Einstein condensation was achieved in the dilute ultracold
68: gas of magnetically trapped alkali atoms. Since the first
69: realization of Bose-Einstein condensation in ultracold atomic gas
70: clouds \cite{[1],[3]}, studies in this direction have yielded
71: unprecedented insight into the quantum statistical properties of
72: matter. Besides the studies using the bosonic atoms, growing
73: interest is focused on the cooling of fermionic atoms to a
74: temperature regime where quantum effects dominate the properties
75: of the gas \cite{sci99}. This interest is mainly motivated by the
76: quest for the Bardeen-Cooper-Schrieffer (BCS) transition in
77: ultarcold atomic Fermi gasses \cite{nat03_1}.
78: 
79: Strong {\it s}-wave interactions that facilitate evaporative
80: cooling of bosons are absent among spin-polarized fermions due to
81: the exclusion Pauli principle. So the fermions are cooled to
82: degeneracy through the mediation of fermions in another spin state
83: \cite{sci99,nat03_1} or via a buffer gas of bosons \cite{sci02bf}
84: (sympathetic cooling). The Bose gas, which can be cooled
85: evaporatively, is used as a coolant, the fermionic system being in
86: thermal equilibrium with the cold Bose gas through boson-fermion
87: interaction in the region of overlapping of the systems.
88: 
89: However, the physical properties of Bose-Fermi mixtures are
90: interesting in their own rights and are the subject of intensive
91: investigations including the analysis of ground state properties,
92: stability, effective Fermi-Fermi interaction mediated by the
93: bosons, and new quantum phases in an optical lattices. Several
94: successful attempts to trap and cool mixtures of bosons and
95: fermions were reported. Quantum degeneracy was first reached with
96: mixtures of bosonic $^7$Li and fermionic $^6$Li atoms. Later,
97: experiments to cool mixtures of $^{23}$Na and $^6$Li, as well as
98: $^{87}$Rb and $^{40}$K \cite{sci02bf}, to ultralow temperatures
99: succeeded.
100: 
101: In this article we study the instability and collapses of the
102: trapped boson-fermion mixture due to the boson-fermion attractive
103: interaction, using the effective Hamiltonian for the Bose system
104: \cite{ChRyz04,CTR04}. We analyze quantitatively properties of the
105: $^{87}$Rb and $^{40}$K mixture  with an attractive interaction
106: between bosons and fermions recently studied by Modugno and
107: co-workers \cite{sci02bf}. They found that as the number of bosons
108: is increased there is an instability value $N_{Bc}$ at which a
109: discontinues leakage of of the bosons and fermions occurs, and
110: collapse of boson and fermion clouds is observed. Using the
111: experimental parameters we estimated the instability boson number
112: $N_{Bc}$ for the collapse transition as a function of the fermion
113: number and temperature and found a good agreement with
114: experimental results.
115: 
116: First of all we briefly discuss the effective boson Hamiltonian
117: \cite{ChRyz04,CTR04}.  Our starting point is the
118: functional-integral representation of the grand-canonical
119: partition function of the Bose-Fermi mixture. It has the form
120: \cite{popov1}:
121: \begin{eqnarray}
122: Z&=&\int
123: D[\phi^*]D[\phi]D[\psi^*]D[\psi]\exp\left\{-\frac{1}{\hbar}\left(
124: S_B(\phi^*,\phi)+\right.\right.\\
125: \nonumber
126: &+&\left.\left.S_F(\psi^*,\psi)+S_{int}(\phi^*,\phi,\psi^*,\psi)\right)\right\}.
127: \label{1}
128: \end{eqnarray}
129: and consists of an integration over a complex field
130: $\phi(\tau,{\bf r})$, which is periodic on the imaginary-time
131: interval $[0,\hbar\beta]$, and over the Grassmann field
132: $\psi(\tau,{\bf r})$, which is antiperiodic on this interval.
133: Therefore, $\phi(\tau,{\bf r})$ describes the Bose component of
134: the mixture, whereas $\psi(\tau,{\bf r})$ corresponds to the Fermi
135: component. The term describing the Bose gas has the form:
136: \begin{eqnarray}
137: S_B(\phi^*,\phi)&=&\int_0^{\hbar\beta}d\tau\int d{\bf
138: r}\left\{\phi^*(\tau,{\bf r})\left(\hbar \frac{\partial}{\partial
139: \tau}-\frac{\hbar^2\nabla^2}{2 m_B} + \right.\right.\nonumber\\
140:  &+&V_B({\bf r})-\left.\left.\mu_B\right)\phi(\tau,{\bf
141: r})+ \frac{g_B}{2}|\phi(\tau,{\bf r})|^4\right\}. \label{2}
142: \end{eqnarray}
143: Because the Pauli principle forbids $s$-wave scattering between
144: fermionic atoms in the same hyperfine state, the Fermi-gas term
145: can be written in the form:
146: \begin{eqnarray}
147: S_F(\psi^*,\psi)&=&\int_0^{\hbar\beta}d\tau\int d{\bf
148: r}\left\{\psi^*(\tau,{\bf r})\left(\hbar \frac{\partial}{\partial
149: \tau}-\frac{\hbar^2\nabla^2}{2 m_F}+ \right.\right.\nonumber\\
150: &+&\left.\left.V_F({\bf r}) -\mu_F\right)\psi(\tau,{\bf r})
151: \right\}. \label{3}
152: \end{eqnarray}
153: The term describing the interaction between the two components of
154: the Fermi-Bose mixture is:
155: \begin{equation}
156: S_{int}(\phi^*,\phi,\psi^*,\psi)=g_{BF}\int_0^{\hbar\beta}d\tau\int
157: d{\bf r} |\psi(\tau,{\bf r})|^2|\phi(\tau,{\bf r})|^2, \label{4}
158: \end{equation}
159: where $g_B=4\pi \hbar^2a_B/m_B$ and $g_{BF}=2\pi
160: \hbar^2a_{BF}/m_I$, $m_I=m_B m_F/(m_B+m_F)$, $m_B$ and $m_F$ are
161: the masses of bosonic and fermionic atoms respectively, $a_B$ and
162: $a_{BF}$ are the $s$ wave scattering lengths of boson-boson and
163: boson-fermion interactions.
164: 
165: Integral over Fermi fields is Gassian, we can calculate this
166: integral and obtain the partition function of the Fermi system as
167: a functional of Bose field $\phi(\tau, {\bf r})$. In the
168: semiclassical Thomas-Fermi approximation one has
169: \cite{ChRyz04,CTR04}:
170: \begin{eqnarray}
171: S_{eff}&=&\int_0^{\hbar\beta}d\tau d{\bf r}f_{eff}(|\phi(\tau,{\bf
172: r})|),\label{18}\\
173: f_{eff}&=&-\frac{3}{2}\kappa\beta^{-1}\int_0^\infty\sqrt{\epsilon}
174: d\epsilon\ln\left(1+e^{\beta
175: (\tilde{\mu}-\epsilon)}\right)=\nonumber\\
176: &=&-\kappa\int_0^\infty\frac{\epsilon^{3/2}d\epsilon}{1+e^{\beta(\epsilon-\tilde{\mu})}},
177: \label{20}
178: \end{eqnarray}
179: where $\epsilon=p^2/2m_F$, $\tilde{\mu}=\mu_F-V_F({\bf
180: r})-g_{BF}|\phi(\tau,{\bf r})|^2$ and
181: $\kappa=2^{1/2}m_F^{3/2}/(3\pi^2\hbar^3)$.
182: 
183: So we can write the effective bosonic Hamiltonian in the form:
184: \begin{eqnarray}
185: H_{eff}&=&\int d{\bf
186: r}\left\{\frac{\hbar^2}{2m_B}|\nabla\phi|^2+(V_B({\bf
187: r})-\mu_B)|\phi|^2 +\right.\nonumber
188: \\ &+&\left.\frac{g_B}{2}|\phi|^4+f_{eff}(|\phi|)\right\}. \label{21}
189: \end{eqnarray}
190: The first three terms in (\ref{21}) have the conventional
191: Gross-Pitaevskii form, and the last term is a result of
192: boson-fermion interaction. In low temperature limit
193: $\tilde{\mu}/(k_BT)\gg 1$ one can write $f_{eff}(|\phi|)$ in the
194: form:
195: \begin{equation}
196: f_{eff}(|\phi|)=-\frac{2}{5}\kappa\tilde{\mu}^{5/2}-\frac{\pi^2}{4}\kappa(k_BT)^2
197: \tilde{\mu}^{1/2}. \label{25}
198: \end{equation}
199: 
200: As usual, $\mu_F$ can be determined from the equation
201: \begin{equation}
202: N_F=\int d{\bf r}n_F({\bf r}), \label{33}
203: \end{equation}
204: where
205: \begin{equation}
206: n_F({\bf r})=
207: \frac{3}{2}\kappa\int_0^\infty\frac{\sqrt{\epsilon}d\epsilon}{1+e^{\beta
208: (\epsilon-\tilde{\mu})}}. \label{31}
209: \end{equation}
210: 
211: 
212: Let us consider now the $^{87}$Rb and $^{40}$K mixture  with an
213: attractive interaction between bosons and fermions \cite{sci02bf}.
214: The parameters of the system are the following: $a_B=5.25\,\, nm$,
215: $a_{BF}=-21.7^{+4.3}_{-4.8}\,\, nm$. K and Rb atoms were prepared
216: in the doubly polarized states $|F=9/2, m_F=9/2>$ and $|2,2>$,
217: respectively. The magnetic potential had an elongated symmetry,
218: with harmonic oscillation frequencies for Rb atoms
219: $\omega_{B,r}=\omega_B=2\pi\times 215 Hz$ and $
220: \omega_{B,z}=\lambda\omega_B=2\pi\times 16.3 Hz,
221: \omega_F=\sqrt{m_B/m_F}\omega_B\approx 1.47 \omega_B$, so that
222: $m_B\omega_B^2/2=m_F\omega_F^2/2=V_0$.  The collapse was found for
223: the following critical numbers of bosons and fermions:
224: $N_{Bc}\approx 10^{5}; N_K\approx 2\times 10^{4}$.
225: 
226: At the zero temperature limit, expanding $f_{eff}(|\phi|)$ up to
227: the third order in $g_{BF}$ we obtain the effective Hamiltonian in
228: the form:
229: \begin{eqnarray}
230: H_{eff}&=&\int d{\bf
231: r}\left\{\frac{\hbar^2}{2m_B}|\nabla\phi|^2+(V_{eff}({\bf
232: r})-\mu_B)|\phi|^2 +\right. \nonumber\\
233: &+&\left.\frac{g_{eff}}{2}|\phi|^4+\frac{\kappa}{8\mu_F^{1/2}}g_{BF}^3|\phi|^6\right\}.
234: \label{he}
235: \end{eqnarray}
236: where
237: \begin{eqnarray}
238: V_{eff}({\bf r})&=&\left(1-\frac{3}{2}\kappa\mu_F^{1/2}
239: g_{BF}\right)\frac{1}{2}m_B\omega^2_B\left(\rho^2+\lambda^2
240: z^2\right),\label{p1}\\
241: g_{eff}&=&g_B-\frac{3}{2}\kappa\mu_F^{1/2}g_{BF}^2,\label{p2}
242: \end{eqnarray}
243: and  $\rho^2=x^2+y^2$.
244: 
245: In principle, one can study the properties of a Bose-Fermi mixture
246: with the help of $f_{eff}$ (\ref{25}) without any expansion.
247: However, the form of the Hamiltonian (\ref{he}) gives the
248: possibility to get a clear insight into the physics of the
249: influence of the Fermi system on the Bose one (see discussion
250: below). It may be easily verified that the expansion of the
251: function $f(x)=(1+x)^{5/2}$ (see Eq. (\ref{25})) up to the third
252: order in $x$ gives a reasonably good approximation for $f(x)$ even
253: for rather large values of $x$, in contrast with the higher order
254: expansions, so one can safely use Eq. (\ref{he}) as a starting
255: point for the investigation of the properties of the Bose
256: subsystem.
257: 
258: In derivation of Eqs. (\ref{he}-\ref{p2}) we also use the fact
259: that due to the Pauli principle (quantum pressure) the radius of
260: the Bose condensate is much less than the radius of the Fermi
261: cloud $R_F\approx \sqrt{\mu_F/V_0}$, so one can use an expansions
262: in powers of $V_F({\bf r})/\mu_F$.
263: 
264: From Eq. (\ref{p1}) one can see that the interaction with Fermi
265: gas leads to modification of the trapping potential. For the
266: attractive fermion-boson interaction the system should behave as
267: if it was confined in a magnetic trapping potential with larger
268: frequencies than the actual ones, in agreement with experiment
269: \cite{sci02bf}. Boson-fermion interaction also induces the
270: additional attraction between Bose atoms which does not depend on
271: the sign of $g_{BF}$.
272: 
273: The last term in $H_{eff}$ (\ref{he}) corresponds to the
274: three-particle \emph{elastic} collisions induced by the
275: boson-fermion interaction. In contrast with \emph{inelastic}
276: 3-body collisions which result in the recombination and removing
277: particles from the system \cite{kagan2}, this term for $g_{BF}<0$
278: leads to increase of the gas density in the center of the trap in
279: order to lower the total energy. The positive zero point energy
280: and boson-boson repulsion energy (the first two terms in Eq.
281: (\ref{he})) stabilize the system. However, if the central density
282: grows too much, the kinetic energy and boson-boson repulsion are
283: no longer able to prevent the collapse of the gas. Likewise the
284: case of Bose condensate with attraction (see, for example,
285: \cite{kagan2}), the collapse is expected to occur when the number
286: of particles in the condensate exceeds the critical value
287: $N_{Bc}$.
288: 
289: The critical number $N_{Bc}$ can be calculated using the
290: well-known ansatz for the Bosonic wave function:
291: \begin{equation}
292: \phi({\bf
293: r})=\left(\frac{N_B\lambda}{w^3a^3\pi^{3/2}}\right)^{1/2}
294: \exp\left(-\frac{\left(\rho^2+\lambda^2
295: z^2\right)}{2w^2a^2}\right), \label{38}
296: \end{equation}
297: where $w$ is a dimensionless variational parameter which fixes the
298: width of the condensate and $a=\sqrt{\hbar/m_B\omega_B}$.
299: 
300: In this case the variational energy $E_B$ has the form:
301: \begin{equation}
302: \frac{E_B}{N_B\hbar\omega_B}=\frac{2+\lambda}{4}\frac{1}{w^2}+b
303: w^2+\frac{c_1N_B}{w^3}+\frac{c_2N_B^2}{w^6},\label{var}
304: \end{equation}
305: \begin{eqnarray*}
306: b&=&\frac{3}{4}\left(1-\frac{3}{2}\kappa\mu_F^{1/2}g_{BF}\right),\\
307: c_1&=&\frac{1}{2}\left(g_B-\frac{3}{2}\kappa\mu_F^{1/2}g_{BF}^2\right)
308: \frac{\lambda}{(2\pi)^{3/2}\hbar\omega_Ba^3},\\
309: c_2&=&\frac{\kappa}{8\mu_F^{1/2}}g_{BF}^3\frac{\lambda^2}{3^{3/2}
310: \pi^3\hbar\omega_Ba^6}.
311: \end{eqnarray*}
312: This energy is plotted in Fig.1 as a function of $w$ for several
313: values of $N_B$. It is seen that when $N_B<N_{Bc}$ there is a
314: local minimum of $E_B$ which correspond to a metastable state of
315: the system. This minimum arises due the competition between the
316: positive first three terms in Eq. (\ref{var}) and negative fourth
317: term. The local minimum disappears when the number of bosons $N_B$
318: exceeds the critical value which can be calculated by requiring
319: that the first and second derivatives of $E_B$ vanish at the
320: critical point. In this case the behavior of $E_B$ is mainly
321: determined by the second and fourth terms in Eq. (\ref{var}). For
322: $N_K= 2\times 10^{4}$ and $a_{BF}=-19.44\, nm$ we obtain
323: $N_{Bc}\approx 9\times10^4$ in a good agreement with the
324: experiment \cite{sci02bf}. It is interesting to note that the
325: critical number of Bose atoms in Bose-Fermi mixture is about two
326: orders larger than the critical number for the condensate with a
327: purely attractive interaction. For example, in the experiments
328: with trapped $^7$Li \cite{[3]} it was found that the critical
329: number of bosons is about $1000$.
330: 
331: \begin{figure}
332: \begin{center}
333: \includegraphics[width=8cm]{RyzhovFig1.eps}
334: \caption{Variational energy $\frac{E_B}{N_B\hbar\omega_B}$ as a
335: function of $w$ for various numbers of bosons.}
336: \end{center}
337: \end{figure}
338: 
339: 
340: Upon increasing the number of fermions, the repulsion between
341: bosons decreases leading to the collapse for the smaller numbers
342: of the bosonic atoms. It should be also noted that an increase of
343: the temperature results in a decrease of the local density of
344: fermions and reduces the interaction energy between Bose and Fermi
345: systems increasing the critical number of bosons.
346: 
347: Finally, we make a short remark on the nature of the collapse
348: transition. In this article we found the instability point of the
349: Bose-Fermi mixture with attractive interaction between components.
350: A strong rise of density of bosons and fermions (in the collapsing
351: condensate enhances intrinsic inelastic processes, in particular,
352: the recombination in 3-body interatomic collisions, as is the case
353: for the well-known $^7$Li condensates \cite{kagan2}. However,
354: recently M. Yu. Kagan and coworkers suggested the new microscopic
355: mechanism of removing atoms from the system which is specific for
356: the Bose-Fermi mixtures with attraction between components and is
357: based on the formation of the boson-fermion bound states
358: \cite{MKag}. It seems that the description of the evolution of the
359: collapsing condensate should include both these mechanisms.
360: 
361: VNR and EET are grateful to N.M. Plakida for the interest to this
362: work. VNR acknowledges valuable discussions with M. Yu. Kagan. The
363: work was supported in part by the Russian Foundation for Basic
364: Research (Grants No 02-02-16622 and No 02-02-16621). STC is partly
365: supported by a grant from NASA.
366: 
367: 
368: 
369: \begin{thebibliography}{99}
370: \itemsep=-2pt
371: \bibitem{[1]}  M.\,N. Anderson, J.\,R. Ensher, M.\,R. Matthews
372: {\it et al.}, Science {\bf 269}, 198 (1995).
373: \bibitem{[3]} C.\,C. Bradley, C.\,A. Sackett, and R.\,G. Hulet, Phys.
374: Rev. Lett. {\bf 78}, 985 (1997).
375: \bibitem{sci99} B. DeMarco and D.\,S. Jin, Science {\bf 285}, 1703
376: (1999).
377: \bibitem{nat03_1} C.\,A. Regal, C. Ticknor, J.\,L. Bohn, and D.\,S. Jin,
378: Nature {\bf 424}, 47 (2003).
379: \bibitem{sci02bf} G. Modungo, G. Roati, F. Riboli {\it et al},
380: Science {\bf 297}, 2240 (2002).
381: \bibitem{ChRyz04} S.\,T. Chui and V.\,N. Ryzhov, Phys. Rev. A {\bf
382: 69}, 043607 (2004).
383: \bibitem{CTR04} S.\,T.\, Chui, V.\,N.\,Ryzhov, E.\,E.\,Tareyeva, JETP
384: Lettes {\bf 80}, 274 (2004).
385: \bibitem{popov1} V.\,N. Popov, {\it Functional Integrals in Quantum Field Theory
386: and Statistical Physics} (Reidel, Dordrecht, 1983).
387: \bibitem{kagan2} Yu. Kagan, A.\,E. Muryshev, and G.\,V.
388: Shlyapnikov, Phys. Rev. Lett. {\bf 81}, 933 (1998).
389: \bibitem{MKag} M.\,Yu. Kagan, I.\,V. Brodsky, D.\,V. Efremov, A.\,V. Klaptsov,
390: cond-mat/0209481.
391: 
392: 
393: \end{thebibliography}
394: 
395: \end{document}
396: