1: \documentclass[aps,twocolumn,showpacs]{revtex4}
2: \usepackage{graphicx,epsfig,amsmath,amsfonts,epic,eepic,psfrag,latexsym,amssymb}
3:
4: \renewcommand{\vec}[1]{{\mathbf #1}}
5: \newcommand{\calM}{{\mathcal M}}
6: \newcommand{\epstil}{\tilde{\epsilon}}
7: \newcommand{\omtil}{\tilde{\omega}}
8:
9: \begin{document}
10:
11: \title{Dynamics of the BCS-BEC crossover in a degenerate Fermi gas}
12:
13: \author{M. H. Szyma{\'n}ska${}^{1}$}
14: \author{B. D. Simons${}^{1}$}
15: \author{K. Burnett${}^{2}$}
16:
17: \affiliation{${}^{1}$ Cavendish Laboratory, University of Cambridge,
18: Madingley Road, Cambridge CB3 OHE, UK \\
19: ${}^{2}$ Clarendon Laboratory, University of
20: Oxford, Parks Road, Oxford, OX1 3PU, UK}
21: \pacs{03.75.Kk, 03.75.Ss, 05.30.Fk}
22:
23: \begin{abstract}
24: We study the short-time dynamics of a degenerate Fermi gas positioned
25: near a Feshbach resonance following an abrupt jump in the atomic
26: interaction resulting from a change of external magnetic field. We
27: investigate the dynamics of the condensate order parameter and pair
28: wavefunction for a range of field strengths. When the abrupt jump is
29: sufficient to span the BCS to BEC crossover, we show that the rigidity
30: of the momentum distribution precludes any atom-molecule oscillations
31: in the entrance channel dominated resonances observed in the
32: ${}^{40}$K and ${}^{6}$Li. Focusing on material parameters tailored to
33: the ${}^{40}$K Feshbach resonance system at $202.1$\, gauss, we
34: comment on the integrity of the fast sweet projection technique as a
35: vehicle to explore the condensed phase in the crossover region.
36: \end{abstract}
37:
38: \maketitle
39:
40: Ultracold alkali atomic gases provide a valuable arena in which to
41: explore molecular Bose-Einstein condensation (BEC)~\cite{Greiner03}
42: and fermionic pair condensation~\cite{Regal04-2,Zwierlein04}. Further,
43: the facility to control interparticle interactions via a
44: magnetically-tuned Feshbach resonance (FR) provide a unique
45: opportunity to investigate the BCS-BEC crossover and the dynamics of
46: condensate formation. As well as the adiabatic association of
47: molecules~\cite{Greiner03}, both fast sweep `projections' of fermionic
48: pair condensates onto the molecular BEC~\cite{Regal04-2,Zwierlein04},
49: and atom-molecule Ramsey fringes~\cite{Donley} have been reported in
50: the recent literature. Lately, motivated by earlier work on the
51: mean-field BCS system~\cite{Levitov1}, it was shown in separate
52: works~\cite{Levitov2,Anton} that the mean-field equations of motion of
53: a Bose-Fermi (BF) model, commonly used to describe the FR system, are
54: characterised by an integrable nonlinear dynamics. From these works,
55: three striking predictions emerged: Firstly, when perturbed by an
56: abrupt change in the strength of the pair interaction, the condensate
57: order parameter exhibits substantial oscillations which range in
58: magnitude between some initial state value, $\Delta_\mathrm{I}$, and
59: that expected for the equilibrium final state configuration
60: $\Delta_{\mathrm{eq}}$. Secondly, in the absence of energy relaxation
61: processes, these oscillations remain undamped suggesting the potential
62: to observe coherent atom-molecule oscillations in the FR system.
63: Thirdly, it is proposed that a spectral `hole-burning' phenomena in
64: the atomic momentum distribution, at a frequency associated with one
65: half of the molecular binding energy, provides a signature of such
66: atom-molecule oscillations~\cite{Anton}.
67:
68: In the following, we will argue that this behaviour rests on an
69: auxiliary constraint that, in the present system, seems hard to
70: justify. Drawing on the results of a numerical analysis of the
71: unconstrained dynamics, we will show that, in the absence of
72: relaxational processes, the oscillations of the order parameter are
73: damped substantially, even at the level of the mean-field. When the
74: abrupt change of the interaction is sufficient to span the BCS-BEC
75: crossover, for both the single channel and BF systems, the magnitude
76: of oscillations remain small attenuating to some value $\Delta_{\rm
77: F}$ which lies close to the initial state, and much smaller than
78: the expected equilibrium state value, $\Delta_{\rm eq}$, while the
79: distribution remains essentially ``frozen'' to the initial state. The
80: hole-burning oscillations of the form predicted in Ref.~\cite{Anton}
81: do not appear in either regime. From these results, we are able to
82: conclude that, although the observation of BCS-BEC like atom-molecule
83: oscillations in alkali Fermi gases seems infeasible, the rigidity of
84: the initial state distributions validate fast sweep
85: techniques~\cite{Regal04-2} as a reliable experimental tool to explore
86: the crossover region.
87:
88:
89: Formally, a detailed microscopic theory of FR phenomena demands
90: consideration of all matrix elements connecting different spin states
91: participating in condensate formation. In practice, an accurate
92: description of the resonance can be obtained either by using a
93: single-channel magnetic field dependent effective interaction between
94: atoms in the entrance channel or, more widely, in the two-most
95: relevant channels. For low relative momenta, relevant to cold atom
96: physics, the full form of complex atomic potentials are not
97: resolved. Indeed, separable potentials with microscopic parameters
98: drawn from experiment and exact multichannel calculations can be used
99: to recover all low-energy binary scattering
100: observables~\cite{Oxford1,Oxford2}. Since, very often, only one bound
101: state of the closed channel potential is relevant, it has been
102: traditionally replaced by a fictitious Bose-particle and FR phenomena
103: captured by an effective two-channel BF
104: Hamiltonian~\cite{Holland,meera}.
105:
106: Now, applied to the entrance channel dominated resonances observed in
107: the ${}^{40}$K and ${}^{6}$Li system, formal
108: calculations~\cite{Oxford1,Oxford2} support a picture in which the
109: BCS-BEC crossover \emph{is mediated by only a very small admixture of
110: closed channel states} --- e.g., in the ${}^{40}$K FR at 202.1 gauss,
111: the admixture of the closed channel is ca.~$8\%$ or less of the total
112: \cite{Oxford2}. Rather, the weakly bound molecular state appears at a
113: detuning $E_0$ which lies far from the value where the resonance state
114: of the closed channel crosses the dissociation threshold. Since the
115: two-body observables drawn from the exact numerical solution of the
116: Shr\"odinger equation within finite-range single and two-channel
117: models do not differ over a wide range of fields~\cite{Oxford2}, FR
118: phenomena in ${}^{40}$K can be equally-well described by a
119: single-channel theory,
120: %
121: \begin{equation}
122: \hat{H}=\sum_{\mathbf{k}s} \epsilon_{\mathbf{k}}
123: a^{\dagger}_{\mathbf{k}s} a_{\mathbf{k}s}
124: +\sum_{\bf{k}\bf{k'}\bf{q}} V_{{\bf{k}}{\bf{k}}'}
125: a^{\dagger}_{{\bf{k}}+{\bf{q}}\uparrow}
126: a^{\dagger}_{-\bf{k}\downarrow} a_{-{\bf{k}}'\downarrow}
127: a_{{\bf{k}}'+{\bf{q}}\uparrow},
128: \label{HMB1ch}
129: \end{equation}
130: %
131: involving Fermi operators $a^{\dagger}_{\mathbf{k}s}$ and
132: $a_{\mathbf{k}s}$. In the following, to account for the entire region
133: of crossover (and not only the universal regime) we take as matrix
134: elements $V_{{\bf{k}}{\bf{k}}'}=V_0(B)\chi_{\bf
135: k}(\sigma_{\mathrm{bg}}) \chi_{{\bf k}'}(\sigma_{\mathrm{bg}})$ with
136: $\chi_{\bf k}( \sigma_{\mathrm{bg}})=\exp[-({\bf
137: k}\sigma_{\mathrm{bg}})^2/2]$ and the parameters $V_0(B)$ and
138: $\sigma_{\mathrm{bg}}$ chosen to recover the correct magnetic field
139: dependence of the scattering length and the highest vibrational bound
140: state~\cite{Oxford2}.
141:
142: Applied to the fields $\Phi_{\bf k}=\sum_s\langle a^{\dagger}_{\bf{k}
143: s}a_{\bf{k}s}\rangle$/2 and $\kappa_{\bf{k}}=\langle a_{-\bf{k}
144: \downarrow}a_{\bf{k},\uparrow}\rangle$, the Heisenberg equations of motion
145: $i\langle\dot{A}\rangle=\langle[A,\hat{H}]\rangle$ translate to the
146: relations
147: %
148: \begin{eqnarray}
149: i\dot{\kappa}_{\bf{k}} &=& 2\epsilon_{\bf{k}}\kappa_{\bf{k}} -
150: \Delta_{\bf{k}}(2\Phi_{\bf{k}}-1)\nonumber \\ i\dot{\Phi}_{\bf{k}} &=&
151: \Delta_{\bf{k}} \kappa_{\bf{k}}^{\star} -
152: \Delta^{\star}_{\bf{k}}\kappa_{\bf{k}},
153: \label{eqofmotion}
154: \end{eqnarray}
155: %
156: where $\Delta_{\bf k}=V_0(B)\chi_{\bf
157: k}(\sigma_{\mathrm{bg}})\sum_{{\bf k}'} \chi_{{\bf
158: k}'}(\sigma_{\mathrm{bg}}) \kappa_{{\bf k}'}$ denotes the complex
159: order parameter. Similarly for a BF theory, defining $g_{\bf
160: k}=g_0\chi_k(\sigma)$ as the coupling of the entrance channel states
161: to the bosonic field $b_{\bf k}$ associated with the Feshbach
162: resonance level configuration of the closed channel, the equations of
163: motion acquire the same form as~(\ref{eqofmotion}) with $\Delta_{\bf
164: k} = g_0 \chi_{\bf k}(\sigma) b_0 +V_{\mathrm{bg}}\chi_{\bf
165: k}(\sigma_{\mathrm{bg}})\sum_{{\bf k}'} \chi_{{\bf
166: k}'}(\sigma_{\mathrm{bg}}) \kappa_{{\bf k}'}$ and $b_0=\langle b_{{\bf
167: k}=0}\rangle$ obeying the supplementary equation $i\dot{b}_0=E_0(B)
168: b_0 +g_0\sum_{\bf{k}} \chi_{\bf k}(\sigma) \kappa_{\bf k}$. As with
169: the single-channel theory, the five parameters which characterise the
170: resonance, the background potential strength $V_{\mathrm{bg}}$ and
171: range $\sigma_{\mathrm{bg}}$ (which define the entrance channel
172: scattering length and its highest vibrational bound state), the
173: interchannel coupling $g_0$ and its range $\sigma$, and the detuning
174: $E_0(B)$ (which specifies the position and width of the resonance),
175: are determined from experiment and exact multichannel
176: calculations~\cite{Oxford2}.
177:
178: Before turning to the numerical analysis, it is instructive to
179: contrast the present approach to that adopted in
180: Refs.~\cite{Levitov1,Levitov2,Anton}. Given an initial condition, the
181: Heisenberg equations of motion~(\ref{eqofmotion}) present a
182: deterministic time evolution of the density distributions. Indeed, the
183: conservation of total atomic density $n$, implicit in the
184: dynamics~(\ref{eqofmotion}), provides a check on the integrity of the
185: numerical integration described below. By contrast, the integrability
186: of the equations of motion as described by
187: Refs.~\cite{Levitov1,Levitov2,Anton} relies on an additional
188: constraint involving the density and an auxiliary parameter playing
189: the role of a ``chemical potential''. The need for this constraint
190: originates from the fact that, as with the equilibrium steady state
191: solution, the analytical solution of the dynamical equations is
192: determined only up to an arbitrary momentum dependent sign factor
193: which is fixed by the auxiliary constraint \cite{Anton}, instead of
194: naturally following from the initial conditions. A similar
195: phenomenology describes the effect of a classical laser source on a
196: semiconductor electron/hole system where the laser frequency
197: ``imprints'' a chemical potential onto the system~\cite{SchR}. While
198: equilibration processes are sufficiently small, the electrons and
199: holes assume a non-equilibrium distribution around the externally
200: imposed chemical potential resulting in a phenomenon of
201: ``spectral-hole burning'' in which the density distribution is
202: depleted at the laser frequency. In the atomic gas system, where the
203: degrees of freedom remain internal, it is difficult to see how such a
204: choice is motivated or justified. Crucially, we will see that the
205: unconstrained dynamics associated with the equations of
206: motion~(\ref{eqofmotion}) lead to physical behaviour very different
207: from that obtained from the constrained
208: dynamics~\cite{Levitov1,Levitov2,Anton}.
209:
210: With this background, let us now turn to the results of the numerical
211: investigation of the dynamics~(\ref{eqofmotion}) performed using an
212: adaptive step Runge-Kutta algorithm applied on a grid in momentum
213: space chosen fine enough to ensure convergence at each time step.
214: Although we find that the qualitative behaviour of BCS and BF dynamics
215: is generic, we focus specifically on potentials characteristic for the
216: ${}^{40}$K resonance at $B_0=202.1\,$gauss with a density of
217: $n=1.5\times 10^{13}\,{\rm cm}^{-3}$ (i.e. $T_{\mathrm{F}}=0.35\,\mu
218: \mathrm{K}$) comparable to that used in
219: experiment~\cite{Regal04-2}. With these parameters, the equilibrium
220: properties of the effective single-channel and BF model essentially
221: coincide (for further details and values of parameters, we refer to
222: Ref.~\cite{Oxford2}). Therefore, to keep our discussion concise, we
223: will focus on the single-channel theory noting that the parallel
224: application to the BF model with appropriate physical parameters
225: generates \emph{quantitatively} similar results.
226: %
227: \begin{figure}[htbp]
228: \centering
229: \includegraphics[width=3in]{chem2.eps}
230: \caption{Variation of the chemical potential $\mu$ with field $B$ at
231: $T=0$ for a mixture of fermionic $^{40}$K atoms prepared in the
232: $(f=9/2,m_f=-9/2)$ and $(f=9/2,m_f=-7/2)$ Zeeman states at a density
233: of $n=1.5\times 10^{13}\,{\rm cm}^{-3}$
234: (i.e. $T_{\mathrm{F}}=0.35\,\mu \mathrm{K}$). The FR takes place at
235: a field $B_0=202.1\,$gauss while the BEC-BCS crossover ($\mu=0$)
236: occurs when $B-B_0\simeq 0.3\,$gauss. The inset indicates the
237: scattering length $a(B)=a_{\mathrm{bg}}(1-\frac{\Delta B}{B-B_0})$
238: in the universal regime, where a$_\mathrm{bg}$ is the background
239: potential scattering length and $\Delta$B the width of the
240: resonance. Note that, in the present theory, we use finite range
241: potentials with parameters which describe the FR also far outside
242: the universal region \cite{Oxford2}.}
243: \label{fig:chem}
244: \end{figure}
245: %
246: At a field of ca.~$1.0\,$gauss above the FR, the condensate has an essentially
247: BCS-like character while the experiment using the fast sweep technique
248: observed the condensate starting from $0.5\,$gauss above $B_0$. To explore
249: the entire region of interest, we choose as initial conditions field values
250: $B_\mathrm{I}$ which span the entire crossover region (marked by stars
251: in Fig.~\ref{fig:chem}). Starting from the ground state $T=0$ distribution,
252: we follow the dynamics of the condensate after an abrupt switch to some
253: different value of magnetic field $B_\mathrm{F}$.
254:
255: \begin{figure}[htbp]
256: \centering
257: \includegraphics[width=3.2in]{delfracto-10.eps}
258: \caption{Time dependence of the order parameter amplitude
259: $|\Delta_{{\bf k}=0}|/\Delta_{\rm eq}$ following an abrupt switch
260: from a field of $(B_\mathrm{I}-B_0)/{\rm gauss}=-0.5$ (top),
261: $0.0$, $0.5$, $1.0$ (bottom) to $(B_\mathrm{F}-B_0)/{\rm
262: gauss}=-10$, and from $(B_\mathrm{I}-B_0)/ {\rm gauss}=-0.5$ (top
263: inset) and $(B_\mathrm{I}-B_0)/{\rm gauss}=0.5$ (bottom inset) to
264: $(B_\mathrm{F}-B_0)/{\rm gauss}=-1.0$. In addition to the constant
265: phase velocity of $\Delta_{{\bf k}=0}$ (found numerically to be
266: set by the final state equilibrium chemical potential $\mu$),
267: there is an additional time-dependent phase modulation whose
268: characteristics mirror closely that of the amplitude oscillations. }
269: \label{fig:delta}
270: \end{figure}
271:
272: Figure~\ref{fig:delta} shows the time-evolution of the order parameter
273: $|\Delta_{{\bf k}=0}|$, normalised by the value,
274: $\Delta_{\mathrm{eq}}$ (i.e. the value that it would acquire were the
275: system to reach the $T=0$ ground state at the final field
276: $B_\mathrm{F}$). Here we have chosen a field
277: $B_\mathrm{F}-B_0=-10\,$gauss deep within the BEC phase where the
278: large binding energy of molecules allows their momentum distribution
279: to be inferred from time of flight measurements~\cite{Regal04-2}. In
280: contrast to the predictions of the constrained
281: dynamics~\cite{Levitov1,Levitov2,Anton}, these results show that (a)
282: the coherent oscillations are substantially damped even at the level
283: of mean-field, (b) the amplitude of the oscillations is small, and (c)
284: the order parameter $|\Delta_{{\bf k}=0}|$ asymptotes to a value much
285: less than the expected final state equilibrium value
286: $\Delta_{\mathrm{eq}}$. Referring to the insets of
287: Fig.~\ref{fig:delta}, one may note that, when the initial and final
288: conditions are drawn closer, the period of the oscillations becomes
289: longer and the effects of the damping more pronounced. Moreover,
290: although the period of the oscillations increases monotonically with
291: $\Delta_{\rm eq}$, the dependence is nonlinear and, referring to the
292: bottom inset of Fig.~\ref{fig:ukvk}, one may note that
293: $\Delta_{\mathrm{eq}}$ does not provide a ceiling for the magnitude of
294: the oscillation.
295:
296: To interpret the generic behaviour, it is instructive to access the
297: time-dependence of the pair wavefunction $\kappa_{\bf k}$ and the
298: distribution function $\Phi_{\bf k}$.
299: %
300: \begin{figure}[htbp]
301: \centering
302: \includegraphics[width=3.2in]{ukvk-from0.5to-10.eps}
303: \caption{The pair wave-function $|\kappa_{\bf{k}}|$ (upper panel)
304: and the density distribution function $\Phi_{\bf k}$ (lower panel)
305: shown as function of $k=|{\bf k}|$ with $B_\mathrm{I}-B_0=0.5$\,
306: gauss (dashed lines) and $B_\mathrm{F}-B_0=-10\,$ gauss after
307: $50\,\mu$s (dashed-dotted line) and $B_\mathrm{F}-B_0=-1.0$\, gauss
308: after $800\,\mu$s (solid line) following the abrupt switch as in
309: Fig. \ref{fig:delta}. The dotted lines signify the ground state
310: equilibrium distributions at $B_\mathrm{F}-B_0=-1.0$\, gauss
311: included for comparison. The upper inset shows oscillations of
312: $|\kappa_{{\bf k}=0}|$ for $B_\mathrm{F}-B_0=-10$\, gauss. The lower
313: two insets refer to a rather weak perturbation on the BCS side from
314: $B_\mathrm{I}-B_0=0.3$\, gauss (dashed line) to
315: $B_\mathrm{F}-B_0=0.2$\, gauss. The dotted lines shows the
316: equilibrium distribution while the solid line provides a snap-shot
317: of the distribution at $3$\,ms after the switch. Note that the
318: harmonic modulations visible in the distribution functions translate
319: to a single energy scale of the same order of magnitude as the
320: period of oscillations seen in $|\Delta_{{\bf k}=0}|$. }
321: \label{fig:ukvk}
322: \end{figure}
323: %
324: Figure~\ref{fig:ukvk} shows that, when the abrupt switch takes place
325: from $B_{\mathrm{I}}-B_0=0.5\,$gauss to
326: $B_{\mathrm{F}}-B_0=-10\,$gauss, although there is a slight tendency
327: to shift towards the final state equilibrium distribution, the pair
328: wavefunction and the density distribution remain essentially frozen
329: close to the initial BCS-like distribution, exhibiting only small
330: oscillations in time. In the absence of energy relaxational processes,
331: the system is unable to significantly redistribute weight. By
332: contrast, when the abrupt switch takes place from
333: $B_{\mathrm{I}}-B_0=0.5\,$gauss to $B_{\mathrm{F}}-B_0=-1.0\,$gauss,
334: the proximity of the two phases allows the system to converge to a
335: (non-stationary) modulated distribution whose (stationary) envelope
336: reflects more closely the final state equilibrium distribution.
337: Referring to insets in the bottom panel one may note that a weakly
338: perturbed condensate on the BCS side attenuates to its equilibrium
339: value with the pair distribution showing only small modulations around
340: the equilibrium in accord with the linear stability analysis of the
341: weakly perturbed BCS system~(\ref{eqofmotion}) discussed in
342: Ref.~\cite{Volkov}. In particular, one may note that here (and,
343: indeed, for other values of the initial and final conditions) the
344: hole-burning phenomenon predicted by the constrained
345: dynamics~\cite{Anton} does not appear.
346:
347: To assess the potential to observe coherent atom-molecule oscillations, one
348: can monitor the time evolution of the number of condensed molecules,
349: %
350: \begin{equation}
351: n_{\mathrm{mc}}(t)=\left|\int d^3 k \ \kappa_k(t)\,
352: \phi_\mathrm{B}(k,B_\mathrm{F})\right|^2\,.
353: \label{nmc}
354: \end{equation}
355: %
356: Here $\phi_\mathrm{B}(k,B_\mathrm{F})$ denotes the wavefunction of
357: the highest vibrational bound state being an exact eigenstate of the
358: two-body problem~\cite{Oxford2}.
359: %
360: \begin{figure}[htbp]
361: \centering
362: \includegraphics[width=3.2in]{mol.eps}
363: \caption{Time-dependence of the relative number of condensed
364: molecules
365: $(n_\mathrm{mc}(t)-n_\mathrm{mc}(0))/n_\mathrm{mc}(0)$. Here we have
366: used the same field values as that used in Fig.~\ref{fig:delta}
367: (main) with $(B_{\rm I}-B_0)/{\rm gauss}=-0.5\,$ (top), $0.0$, $0.5$,
368: and $1.0$ (bottom) and $(B_{\rm F}-B_0)/{\rm gauss}=-10$. For these
369: field values, when normalised to one half of the total atomic
370: density, $n_\mathrm{mc}(0)=0.0005$, $0.012$, $0.1$, and $0.3$
371: respectively.}
372: \label{fig:mol}
373: \end{figure}
374: %
375: Referring to Fig.~\ref{fig:mol}, one may note that the amplitude of
376: oscillations is negligible (less than one percent). Although one may
377: adjust the final field $B_{\rm F}$ to lie closer to the FR, the
378: amplitude of the oscillations increases only slightly while the
379: damping rate is also enhanced. Moreover, the inclusion of processes
380: beyond the mean-field approximation considered here would simply
381: increase the damping rate and not enhance the oscillations. One may
382: therefore conclude that the observation of atom-molecule oscillations,
383: following an abrupt change in the interaction strength, is infeasible
384: for the entrance channel dominated resonances currently discussed in
385: the context of the ${}^{40}$K and ${}^{6}$Li system.
386:
387: To conclude, we have presented a numerical analysis of the dynamical
388: mean-field equations for the single-channel theory of the FR following
389: an abrupt field change. In the range of physical parameters
390: appropriate to the ${}^{40}$K system, the consideration of the
391: two-channel BF (Bose-Fermi) theory does not change the results
392: \emph{quant}itatively. When applied to a theoretical regime where the
393: population of the closed channel states below resonance is high, the
394: numerical findings do not change \emph{qual}itatively. Relying on the
395: deterministic time-evolution of the initial state according to the
396: Heisenberg mean-field equations of motion, our results differ
397: substantially from the findings of the constrained dynamics considered
398: previously~\cite{Levitov1,Levitov2,Anton}. Over a wide range of
399: initial conditions, we observe substantially damped oscillations with
400: an amplitude strongly dependent on initial conditions and a frequency
401: set by the interaction strength after the switch. To assess the
402: capacity for BCS-BEC like atom-molecule oscillations following an
403: abrupt change in the interaction, we have chosen initial and final
404: conditions to span the crossover from the BCS to the BEC limits. We
405: have found that the amplitude of atom-molecule oscillations is
406: negligible and the distribution is essentially frozen to the initial,
407: a behaviour reminiscent of an orthogonality catastrophe. We conclude
408: that, in the entrance channel dominated resonances observed in
409: ${}^{40}$K and ${}^{6}$Li, the observation of atom-molecule
410: oscillations is infeasible. The rigidity of the condensate
411: wavefunction distribution at short-time scales, where the time
412: evolution is mean-field in character, supports the method of the fast
413: sweep as a reliable technique to probe the fermionic pair condensate.
414:
415: We are grateful to Krzysztof G{\'o}ral and Thorsten K\"ohler for
416: numerous discussions concerning the FR physics, to Peter Littlewood
417: for stimulating discussions and to S{\l}awomir Matyja{\'s}kiewicz for
418: advice concerning the numerical techniques. This research has been
419: supported by Gonville and Caius College Cambridge (M.H.S.)
420:
421: \begin{thebibliography}{20}
422:
423: \bibitem{Greiner03}
424: %3 November 2003 K2 BEC
425: M. Greiner, \emph{et al.},
426: %C. A. Regal, and D. S. Jin,
427: Nature \textbf{426}, 537 (2003);
428: %\bibitem{Grimm-mol}
429: S. Jochim, \emph{et al.},
430: %M. Bartenstein, A. Altmeyer, G. Hendl, C. Chin, J. Hecker Denschlag,
431: %R. Grimm,
432: %\emph{Bose-Einstein Condensation of Molecules},
433: Science, {\bf 302}, 2102 (2003);
434: %\bibitem{Ket-mol}
435: M. W. Zwierlein, \emph{et al.},
436: %C. A. Stan, C. H. Schunck, S. M. F. Raupach, S. Gupta, Z. Hadzibabic,
437: %W. Ketterle,
438: %\emph{Observation of Bose-Einstein condensation of molecules},
439: Phys. Rev. Lett. {\bf 91}, 250401 (2003).
440:
441: \bibitem{Regal04-2}
442: %13 January K fermionic condensate
443: C. A. Regal, \emph{et al.},
444: %M. Greiner, and D. S. Jin,
445: Phys. Rev. Lett. \textbf{92}, 040403 (2004).
446:
447: \bibitem{Zwierlein04}
448: %1 March 2004 Li fermionic condensate
449: M. W. Zwierlein, \emph{et al.},
450: %C. A. Stan, C. H. Schunck, S. M. F. Raupach, A. J. Kerman, and W. Ketterle,
451: Phys. Rev. Lett. \textbf{92}, 120403 (2004).
452:
453: \bibitem{Donley} E. A. Donley, \emph{et al.},
454: %N. R. Claussen, S. T. Thompson, C. E. Wieman
455: Nature \textbf{417}, 529 (2002).
456:
457: \bibitem{Levitov1}R. A. Barankov, \emph{et al.},
458: %L. S. Levitov, and B. Z. Spivak
459: Phys. Rev. Lett. \textbf{93}, 160401 (2004).
460:
461: \bibitem{Levitov2} R. A. Barankov, \emph{et al.},
462: %and L. S. Levitov,
463: Phys. Rev. Lett. \textbf{93}, 130403 (2004).
464:
465: \bibitem{Anton} A. V. Andreev, \emph{et al.},
466: %V. Gurarie, and L. Radzihovsky,
467: Phys. Rev. Lett. \textbf{93}, 130402 (2004).
468:
469: \bibitem{Boris} E. A. Yuzbashyan, \emph{et al.},
470: %B. L. Altshuler, V. B. Kuznetsov, V. Z. Enolskii,
471: cond-mat/0407501.
472:
473: \bibitem{SchR} S. Schmitt-Rink, \emph{et al.}
474: %D. S. Chemla, and H. Haug
475: Phys. Rev. B \textbf{37}, 941-955 (1988).
476:
477: \bibitem{Oxford1} K.~G\'oral, \emph{et al.},
478: %T.~K\"ohler, S.A.~Gardiner, E.~Tiesinga, and P.S.~Julienne,
479: J. Phys. B \textbf{37}, 3457 (2004).
480:
481: \bibitem{Oxford2} M. H. Szyma{\'n}ska, K.~G\'oral, T.~K\"ohler,
482: K. Burnett, unpublished
483: %cond-mat/ (2004).
484:
485: \bibitem{Holland} M. Holland, \emph{et al.},
486: %S. J. J. M. F. Kokkelmans, M. L. Chiofalo and R. Walser,
487: %\emph{Resonance Superfluidity in a Quantum Degenerate Fermi Gas},
488: Phys. Rev. Lett. {\bf 87}, 120406 (2001);
489: %\bibitem{Timm}
490: E. Timmermans, \emph{et al.},
491: %K. Furuya, P. W. Milonni and A. K. Kerman,
492: %\emph{Prospect of creating a composite Fermi-Bose superfluid},
493: Phys.\ Lett.\ A {\bf 285}, 228 (2001);
494: %\bibitem{Griff}
495: Y. Ohashi, \emph{et al.},
496: %and A. Griffin,
497: %\emph{BCS-BEC Crossover in a Gas of Fermi Atoms with a Feshbach Resonance},
498: Phys.\ Rev.\ Lett. {\bf 89}, 130402 (2002).
499:
500: \bibitem{meera} Note that the particular architecture of the hyperfine
501: states in the $^{40}$K system calls into question the validity of a
502: BF model as a microscopic theory~\cite{meera2}.
503:
504: \bibitem{meera2} M. M. Parish, \emph{et al.},
505: %B. Mihaila, B. D. Simons, P. B. Littlewood,
506: cond-mat/0409756 (2004).
507:
508: \bibitem{Volkov} A. F. Volkov and Sh. M. Kogan,
509: %\emph{Collisionless relaxation of the energy gap in superconductors},
510: Sov. Phys. JETP, {\bf 38}, 1018 (1974).
511:
512: \end{thebibliography}
513:
514:
515: \end{document}
516:
517:
518:
519:
520:
521: