1: %\documentclass[preprint,aps]{revtex4}
2: \documentclass[twocolumn,aps,graphicx]{revtex4}
3: %\documentclass[preprint,aps,graphicx]{revtex4}
4:
5: \usepackage{graphicx}% Include figure files
6: \usepackage{dcolumn}% Align table columns on decimal point
7: \usepackage{bm}% bold math
8:
9: \begin{document}
10: \title{Ultracold Collisions of Fermionic OD Radicals}
11: \author{A. V. Avdeenkov and John L. Bohn}
12: \affiliation{JILA and Department of Physics,
13: University of Colorado, Boulder, CO 80309-0440}
14: \date{\today}
15:
16: \begin{abstract}
17: We discuss consequences of Fermi exchange symmetry on collisions
18: of polar molecules at low temperatures (below 1 K), considering
19: the OD radical as a prototype. At low fields and low temperatures,
20: Fermi statistics can stabilize a gas of OD molecules against
21: state-changing collisions. We find, however, that this stability
22: does not extend to temperatures high enough to assist with
23: evaporative cooling. In addition, we establish that a novel
24: ``field-linked'' resonance state of OD dimers exists, in analogy with
25: the similar states predicted for bosonic OH.
26: \end{abstract}
27:
28: \maketitle
29:
30: \section{Introduction}
31:
32: Cooling and trapping molecules in their ro-vibrational
33: ground states has proven to
34: be a daunting experimental task, yet now it has been
35: achieved \cite{Weinstein,Bethlem,Meerakker}. Currently, the
36: samples produced by Stark slowing are limited to temperatures around
37: 1 mK, which is cold enough to trap, but not yet cold enough
38: for interesting applications to novel dilute quantum gases of
39: fermions~\cite{You,Goral1,Baranov1,Baranov2,Baranov3}
40: or bosons~\cite{Yi1,Santos,Yi2,Goral2,Giovanazzi,Odell}.
41: To produce colder, denser samples for these applications,
42: an attractive approach may be to use sympathetic
43: cooling with an easily cooled species (such as Rb) \cite{Soldan},
44: or else evaporative cooling. Knowledge of collision cross sections
45: is therefore essential in understanding prospects for the
46: success of either approach.
47:
48: We have previously considered electrostatic trapping of polar
49: $\Pi$-state molecules
50: from the point of view of stability with respect to collisions
51: \cite{Avdeenkov1}.
52: The main bottleneck here is that electro{\it static} trapping requires
53: the molecules to be in a weak-field-seeking state, in which case the molecules
54: of necessity have an even lower-energy strong-field seeking state. Collisions
55: involving the strong and anisotropic dipole-dipole interaction between
56: molecules appear more than adequate to drive the molecules into these
57: unfavorable states, leading to unacceptably high trap loss and
58: heating.
59: For this reason, it may be necessary to seek alternative methods
60: that can confine the strong-field-seeking, absolute lowest-energy ground
61: state of the molecule using time-varying electric
62: fields~\cite{Cornell,Spreeuw,Junglen,DeMille}.
63: A recent proposal for an electro{\it dynamic} trap
64: is based on the microwave analogue of the familiar far-off-resonant
65: optical dipole trap -- only the microwave version doesn't have to be
66: far off-resonance, making the trap very deep \cite{DeMille}.
67:
68: Polar fermions may have an important advantage for
69: electrostatic trapping, namely, low inelastic rates at cold temperatures.
70: Kajita has discussed state-changing collisions
71: of dipolar fermionic molecules \cite{Kajita}, based
72: on the well-known Wigner threshold laws for dipolar interactions.
73: Namely, elastic scattering cross sections are essentially
74: independent of collision energy $E$ at low energies, but state-changing
75: cross sections scale as $E^{1/2}$. Therefore, at ``sufficiently low''
76: temperatures, elastic scattering always wins, and evaporative cooling
77: should be possible. Using the
78: %first
79: Born approximation, Kajita concludes
80: that this is the case for the molecules OCS and CH$_3$Cl,
81: at reasonable experimental temperatures \cite{Kajita}. This analysis
82: may yet prove too optimistic, since the results include regions where
83: the Born approximation may not be strictly applicable \cite{Avdeenkov2}.
84: Still, the idea is a sound one that deserves further exploration.
85:
86: A complete theoretical description of molecule-molecule scattering
87: is complicated by the complexity of the short-range interaction
88: between molecules. Indeed, for open shell molecules the
89: potential energy surface is difficult to compute by {\it ab intio}
90: methods, and remains inadequately known. It is therefore worthwhile
91: to seek situations in which the influence of short-range physics is minimal.
92: It appears that for weak- field seeking states
93: the influence of the short-range potential is weak, owing to
94: avoided crossings in the long-range interaction~\cite{Avdeenkov1}.
95: For collisions of identical fermionic molecules, the influence of
96: short-range physics may be even smaller, since only partial waves
97: with $l \ge 1$ are present, and there is centrifugal repulsion
98: in all scattering channels.
99:
100: A main aim of the present paper is thus to explore the suppression of inelastic
101: collisions in fermionic $\Pi$-state molecules, using the OD radical as an
102: example. This is an illustrative choice of molecule, since we have studied
103: its bosonic counterpart, OH, extensively in the past
104: \cite{Avdeenkov1,Avdeenkov3,Avdeenkov4}. It is also a species
105: at the center of current experimental interest \cite{Meerakker,Bochinski}.
106: To this end we employ
107: full close-coupling calculations to a model of the OD-OD interaction that
108: includes only the dipolar part. We find, as we must, that the fermionic
109: threshold laws ultimately favor elastic over inelastic scattering
110: at low temperatures. For OD, however, we find that the energy
111: scales for this to happen remain quite low, on the order of microKelvins
112: or below, so that the usefulness of this result
113: to evaporative cooling remain questionable.
114:
115: On the bright side, the suppression of inelastic collisions does mean that
116: a gas that is already cold will be stable under collisions, even in an
117: electrostatic trap. This is a similar conclusion to one we have drawn
118: in the past for magnetostatic trapping of spin-polarized paramagnetic
119: (nonpolar) species \cite{Avdeenkov5,Volpi}. This is useful
120: for cold collisions studies, since it is believed that collisions
121: of weak-electric-field seekers are dominated by, and can be
122: understood in terms of, purely long-range dipolar forces \cite{Avdeenkov3}.
123: In particular, such collisions are predicted for bosons
124: to have long-range resonant states, termed field-linked resonances, that may
125: be useful in understanding cold collisions. A second goal of this paper
126: is to verify that the fermionic OD molecules also possess these resonances.
127:
128: \section{Threshold laws in the Born approximation}
129:
130: Threshold laws for various power-law long-range potentials have
131: been written about extensively in the past
132: \cite{Wigner,Landau,Spruch,OMalley,Hinckelman,Shakeshaft,Holzwarth,Peach,Fabrikant,Rosenberg,Cavagnero,Gao,Sadeghpour,Deb}.
133: In this
134: section we summarize the main results relevant to the energy
135: dependence of cross sections, using the first Born approximation to
136: make the math transparent. Similar arguments are presented
137: in Refs.~\cite{Yi1,Kajita}.
138:
139: A first point to be considered is why the Born approximation should
140: be of any use at all, since it is ordinarily associated with collisions
141: of ``fast'' particles. Strictly speaking, however, the Born approximation
142: is valid when the potential responsible for scattering is suitably
143: ``weak,'' meaning that the true scattering wave function is
144: well-approximated by the unperturbed wave function. For dipolar
145: scattering, the argument is as follows. Consider elastic scattering
146: in a single-channel whose long-range potential varies as $1/R^s$.
147: Then, in partial wave $l$, the elastic scattering phase shift
148: will vary with wave number $k$ as \cite{Sadeghpour}
149: \begin{equation}
150: \label{delta}
151: \delta_l \sim \alpha k^{2l+1} + \beta k^{(s-1)},
152: \end{equation}
153: where $\alpha$ and $\beta$ are constants depending on details
154: of the potential.
155:
156: Thus for a dipolar potential with $s=3$, the second term in
157: Eqn.~(\ref{delta}) is the dominant contribution to the phase
158: shift for any partial wave $l \ge 1$, yielding $\delta_l \sim k$.
159: (Moreover, the $l=0$ contribution to a realistic dipole-dipole
160: interaction rigorously vanishes by symmetry.) It can be shown
161: (for example, using the JWKB approximation~\cite{Sadeghpour}) that the
162: second contribution in Eqn.~(\ref{delta}) arises from purely
163: long-range physics, i.e., for intermolecular separations
164: outside the centrifugal barrier imposed by the partial wave.
165: As the collision energy approaches threshold, this distance
166: gets ever larger, and the influence of the $1/R^3$ perturbing
167: potential gets ever weaker. Thus, near threshold, the wave function
168: is well-approximated by unperturbed spherical Bessel functions
169: in each partial wave, and the Born approximation can be used.
170:
171: We adopt this view in the multichannel case. Because the
172: dipole-dipole interaction is anisotropic, different partial
173: waves are coupled together.
174: Nevertheless, the diagonal pieces of the Hamiltonian matrix
175: have the general form
176: \begin{equation}
177: \label{potential}
178: {\hbar ^2 l(l+1) \over 2 \mu^2 R^2} + { C_{3}^{\rm eff} \over R^3},
179: \end{equation}
180: where $R$ is the distance between molecules, $\mu$ is their reduced mass,
181: and the effective $C_3$ coefficient depends, in general,
182: on the channel as well as on the degree of
183: electric field polarization (see Ref.\cite{Avdeenkov1}).
184: When $C_3^{\rm eff}$ is negative, the potential (\ref{potential})
185: presents a finite barrier of height
186: \begin{equation}
187: E_b = {4 \over 27} \left[ {\hbar^2 l(l+1) \over 2 \mu} \right]^3
188: {1 \over (C_3^{\rm eff})^2 }.
189: \end{equation}
190: For energies $E$ considerably less than $E_b$, scattering only
191: occurs from outside the barrier (barring resonances~\cite{Yi1}), thus
192: setting an energy scale for the utility of the Born approximation.
193: To make an estimate of this energy scale, consider the strong-field
194: limit, where polarized molecules have $C_3 \sim d^2$, the square
195: of the dipole moment. For OD, this sets the relevant p-wave
196: centrifugal barrier height at $\sim 10$ nK. At higher energies,
197: the incoming wave spills over the barrier, samples smaller-$R$
198: interactions, and is no longer well-described as a plane wave.
199:
200: Assuming that the Born approximation holds, we proceed as follows.
201: The partial scattering cross section for a collision process entering on
202: channel $i$ and exiting on channel $f$ is given in terms of
203: the transition matrix $T$ by
204: \begin{equation}
205: \label{cross}
206: \sigma_{if} = {\pi \over k_i^2} |\langle i | T | f \rangle|^2,
207: \end{equation}
208: where the channel indices $i$ and $f$ include partial wave
209: contributions $l_i$ and $l_f$, which need not be the same.
210: In the first Born approximation, the $T$ matrix elements are given
211: by the matrix elements of the potential (Chap. 7 of Ref.~\cite{Child},
212: where we have re-inserted the dimensionful factors)
213: \begin{eqnarray}
214: \label{Born}
215: \langle i | T | f \rangle = && 2 \left({2 \mu \over \hbar^2}\right)
216: \left( k_i k_f \right)^{1/2} \\
217: & & \times \int_0^{\infty} R^2 dR j_{l_i}(k_iR)
218: {C_3(l_il_f) \over R^3} j_{l_f}(k_fR) . \nonumber
219: \end{eqnarray}
220: Here $C_3(l_il_f)$ represents the appropriate off-diagonal
221: coupling matrix element, which, again, depends on field.
222:
223: For {\it elastic} scattering, where the initial and final wave numbers
224: are equal, $k_f=k_i$, we can rewrite Eqn.~(\ref{Born}) in terms of
225: the dimensionless variable $x=k_iR$,
226: \begin{eqnarray}
227: \label{Bornelastic}
228: \langle i | T | f \rangle = && {4 \mu C_3(l_il_f) \over \hbar^2} k_i \\
229: && \times \int_0^{\infty} dx {j_{l_i}(x) j_{l_f}(x) \over x}. \nonumber
230: \end{eqnarray}
231: The integral in Eqn.~(\ref{Bornelastic}) converges whenever
232: $l_i + l_f > 0$, and is moreover independent of $k_i$. Therefore,
233: for any elastic scattering process by dipolar forces that changes
234: $l$ by at most 2 units, $T \sim k_i$ at low energies, and by
235: Eqn.~(\ref{cross}), the associated cross section is independent
236: of collision energy. In particular, the elastic scattering cross
237: section of identical fermions does not vanish, if they interact
238: via dipolar forces.
239:
240: For completeness, we give the value of the integral. This is found by
241: substituting ordinary Bessel functions for the spherical Bessel
242: functions, $j_n(x) = \sqrt{ \pi /2x} J_{n+1/2}(x)$, and using
243: standard formulas for integrals \cite{Gradshteyn}:
244: \begin{eqnarray}
245: \int_0^{\infty} && dx {j_{l_i}(x) j_{l_f}(x) \over x} \\
246: =&& { \pi \Gamma ({l_i+l_f \over 2}) \over
247: 8 \Gamma ({-l_i+l_f+3 \over 2}) \Gamma ({l_i+l_f+4 \over 2})
248: \Gamma ({l_i-l_f+3 \over 2}) }. \nonumber
249: \end{eqnarray}
250:
251:
252: For an {\it exothermic} process, with $k_f > k_i$, a similar argument
253: yields for the transition amplitude
254: \begin{eqnarray}
255: \label{Borninelastic}
256: \langle i | T | f \rangle &&= \pi {2 \mu C_3(l_il_f) \over \hbar^2} \\
257: && \times \int_0^{\infty} dR
258: J_{l_i+1/2}(k_iR) J_{l_f+1/2}(k_fR) R^{-2}. \nonumber
259: \end{eqnarray}
260: This integral, too, can be done as long as $l_i+l_f>0$
261: \cite{Gradshteyn}:
262: \begin{eqnarray}
263: \label{inelastic}
264: && \int_0^{\infty} dR J_{l_i+1/2}(k_iR) J_{l_f+1/2}(k_fR) R^{-2} \\
265: \nonumber
266: && = {k_i^{l_i+1/2} \Gamma( {l_i+l_f \over 2}) \over
267: 4 k_f^{l_i-l_f+1} \Gamma ({ -l_i+l_f+3 \over 2} ) \Gamma( l_i+3/2)} \\
268: \nonumber
269: && \times F \left( {l_i+l_f \over 2}, {l_i-l_f-1 \over 2},
270: l_i+3/2; \left( {k_i \over k_f} \right)^2 \right), \nonumber
271: \end{eqnarray}
272: where $F$ stands for a hypergeometric function.
273:
274: Near threshold in an exothermic process, we have $k_f \gg k_i$.
275: In this case the leading order term of the hypergoemetric function
276: $F$ is a constant, and the only remaining dependence of
277: (\ref{inelastic}) on $k_i$ is in its prefactor. Thus
278: $T \sim k_i^{l_i+1}$, and $\sigma \sim k_i^{2l_i-1} \sim
279: E^{l_i-1/2}$.
280: When the incident partial wave is $l_i = 0$, as would be the case
281: for identical bosons, the inelastic scattering cross section
282: diverges at threshold. For any higher partial wave, say the
283: $l_i=1$ partial wave that dominates scattering of identical
284: fermions, the inelastic cross section instead vanishes in the
285: threshold limit.
286:
287:
288: \section{Collision cross sections for OD}
289:
290: The OD radical differs from OH in two significant ways, for our present
291: purposes: first, its lambda-doubling
292: constant is somewhat smaller \cite{Abrams} . Second, its hyperfine structure
293: depends on the nuclear spin of deuterium being 1 instead of 1/2 for hydrogen,
294: meaning that total spin states $f=1/2, 3/2$, and 5/2 are possible in
295: the $^2\Pi_{3/2}$ electronic ground state of OD (as in our OH work,
296: we consider exclusively in the electronic
297: ground state, and neglect excited vibrational and rotational levels).
298: Figure 1 presents the Stark effect for OD, which can be compared to
299: the similar figure for OH, Fig. (1) of Ref.\cite{Avdeenkov1}.
300: Note that, due to the smaller lambda-doublet in OD, this radical
301: enters the linear Stark regime at applied electric fields of
302: $\sim 200$ V/cm, as opposed to $\sim 1000$ V/cm in OH.
303:
304: \begin{figure}
305: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig1a.ps}}
306: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig1b.ps}}
307: \caption{Stark effect for the $^2\Pi_{3/2}|f,m_f,{\rm parity} \rangle$
308: ground states of OD. Shown are state of f parity (a) and e parity (b).}
309: \end{figure}
310:
311: We consider collisions of the highest-energy weak-field seeking state in
312: Fig.1, with quantum numbers $|f,m_f, {\rm parity} \rangle =$
313: $|5/2,5/2,f \rangle$.
314: The details of our scattering theory have been presented elsewhere,
315: for OH \cite{Avdeenkov1}. The
316: main difference in handling OD is to incorporate Fermi exchange symmetry,
317: which amounts to changing plus signs to minus signs in Eqn. (17)
318: of Ref. \cite{Avdeenkov1}.
319: Otherwise, we treat the scattering in the same
320: way, by including only the Stark and dipole-dipole interactions, along
321: with the hyperfine structure.
322:
323: \begin{figure}
324: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig2.ps}}
325: \caption{Elastic (solid lines) and total inelastic(dashed) cross sections
326: for several values of electric field: ${\cal E}=0,100,1000 V/cm$.}
327: \end{figure}
328:
329: Fig. 2 shows the main scattering results, as collision cross sections
330: versus collision energy at different electric fields. Three different
331: applied electric fields are indicated by color coding: ${\cal E}=0$ (black),
332: ${\cal E}=100$ V/cm (red), and ${\cal E}=1000$ V/cm (blue).
333: In each case, the solid line denotes elastic scattering, while
334: the dashed line represents total inelastic scattering to channels
335: where one or both molecules loses internal energy.
336: In zero field, the molecules are completely
337: unpolarized, and the dipole-dipole interaction vanishes. Thus the
338: cross sections obey the familiar Wigner threshold laws for short-ranged
339: interactions between fermions:
340: the elastic cross section $\sigma_{el} \propto E^2$, whereas the
341: (exothermic) inelastic cross section $\sigma_{inel} \propto E^{1/2}$
342: \cite{Sadeghpour}. Thus in zero field elastic scattering is
343: actually {\it less} likely than inelastic scattering at lower
344: energies (below about 10$\mu$K in this example). Above this energy
345: elastic scattering appears somewhat more favorable than inelastic
346: scattering, at least until
347: several mK, where both cross sections start to hit the unitarity limit.
348:
349: Turning on the electric field partially polarizes the molecules, so that the
350: dipole-dipole interaction is ``activated.'' Then the dipole-dipole
351: threshold laws take effect:
352: $\sigma_{el} \propto const.$, whereas we still have
353: $\sigma_{inel} \propto E^{1/2}$. Fig.2
354: illustrates where this threshold behavior kicks in for different electric field
355: values. Notice that the higher the electric field, the lower is the
356: energy where the threshold behavior is attained.
357: This is because the effective $C_3$ coefficient that determines
358: the barrier height $E_b$ is an increasing function of
359: electric field, at least until it saturates \cite{Avdeenkov1}.
360:
361:
362:
363: \begin{figure}
364: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig3.ps}}
365: \caption{Comparison of cross sections for OD (red) and OH
366: (black) molecules.
367: Solid and dashed lines refer to elastic and inelastic cross sections,
368: respectively. }
369: \end{figure}
370:
371: On the other hand, a Fermi gas of molecules that is {\it already
372: cold} will enjoy the benefits of Wigner-law suppression of
373: inelastic collisions. Suppose a quantum degenerate gas of
374: OD could be produced at nK temperatures, as is the case
375: for current experiments in $^{40}$K and $^6$Li. Then Fig. 2
376: suggests that a small bias field of $\sim 100$ V/cm reduces
377: inelastic cross sections to an acceptable level of
378: $\sim 2 \times 10^{-14}$ cm$^2$,
379: corresponding to a rate constant $\sim 10^{-16}$ cm$^3$/sec.
380:
381: To emphasize the difference between bosons and fermions,
382: we reproduce the ${\cal E}=100$ V/cm cross sections
383: in Fig. 3 (red) along with the corresponding
384: cross sections for OH in the same field (black). It is clear that in
385: both cases elastic scattering (solid lines) is quite similar,
386: whereas the behavior of inelastic scattering is utterly
387: different at low energies for the two species. Equally
388: importantly, at collision energies above about 1 mK, all the cross
389: sections have the same general behavior. This is a manifestation
390: of the strength of the dipolar interactions, and the fact that
391: in this energy range all processes are essentially unitarity-limited.
392:
393: \section{On the question of field-linked resonances}
394:
395: Finally, we comment on the occurrence of field-linked (FL) resonance
396: states in this system. Fig. 4a shows the elastic and inelastic
397: cross sections versus electric field, at a fixed collision energy of
398: 1 $\mu$K. This figure exhibits the characteristic peaks indicative
399: of field-linked resonances; compare Fig.(2) of Ref. \cite{Avdeenkov1}.
400: To converge these results at higher field demands an increasing number
401: of partial waves. Fig. 4b illustrates the convergence of the
402: resonant scattering cross section for various numbers of partial
403: waves included. For partial waves $L = 1, 3, 5, 7$, the cross
404: section is well-converged up to several hundred V/cm. This is
405: sufficient to compute the first two resonance states, which are the only
406: well-resolved ones anyway.
407:
408: \begin{figure}
409: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig4a.ps}}
410: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig4b.ps}}
411: \caption{a) Elastic (solid) and inelastic (dashed) cross sections
412: for OD scattering as a function of applied electric field. The
413: Collision energy is $E=1\mu$K. b) Convergence of elastic cross
414: section upon increasing the number of partial waves included in the
415: calculation.}
416: \end{figure}
417:
418: As discussed in Ref. \cite{Avdeenkov1} and elaborated on in
419: Ref. \cite{Avdeenkov4}, the FL resonances live
420: in adiabatic potential energy surfaces generated by avoided crossings in the
421: long-range dipole-dipole interaction. The surfaces have a somewhat different
422: character for fermions than for bosons, however. Fig. 5 shows a set
423: of adiabatic curves for a single partial wave $L=1$; this is the
424: analogue of Fig. (1) in Ref. \cite{Avdeenkov3}.
425:
426: \begin{figure}
427: \centerline{\includegraphics[width=0.9\linewidth,height=1.08\linewidth,angle=-90]{Fig5.ps}}
428: \caption{A set of avoided crossings that generate FL states for OD.}
429: \end{figure}
430:
431: For identical bosons, the long-range attractive part of the relevant curve is
432: predominantly s-wave in character, hence has a $1/R^4$ behavior in an
433: electric field. This reflects the fact that the direct dipolar
434: interaction vanishes for s-waves, and makes an effect only at
435: second order \cite{Avdeenkov1}. For fermions in identical spin states,
436: however, the attractive part involves the p-wave interaction, where
437: the dipole is already nonzero, so that the long-range interaction
438: scales as $1/R^3$. The net effect is that
439: the inner turning point of the
440: s-wave FL states approaches smaller $R$ as the field is increased for bosons,
441: but that this inner turning point is relatively fixed for fermions. (The outer
442: turning point is set by the energy of the resonant state relative to the
443: threshold, and is thus arbitrarily large.)
444:
445: These resonant states, if sufficiently stable, may form a novel kind of pair of
446: fermions, which may ultimately lead to an exotic Fermi superfluid state.
447: Unfortunately, as seen in Fig. 4, these resonances are quite readily susceptible
448: to predissociation, indicated by the large inelastic cross sections near resonance.
449: In this they resemble their bosonic counterparts. However, stabilization of
450: cold dipolar gases using magnetic fields has been recently discussed
451: \cite{Ticknor}.
452: It is yet conceivable that these resonances could be tamed long enough to
453: put them to use.
454:
455: In summary, we have computed scattering cross sections for cold
456: collisions of the fermionic free radical OD, as functions of both
457: collision energy and electric field. We find that, similar to the
458: case of bosonic OH, these molecules are unlikely to be stable against
459: collisions in traps warmer than about 10 $\mu$K. Unlike OH, however,
460: they will be collisionally stable at lower temperatures, owing to
461: the unique Wigner threshold laws for fermionic polar particles.
462: in such a gas.
463:
464: This work has been supported by the NSF and by a grant from the
465: W. M. Keck Foundation. We acknowledge useful interactions with
466: M. Kajita and D. DeMille.
467:
468:
469: \begin{references}
470: \bibitem{Weinstein} J. D. Weinstein, R. deCarvalho, T. Guillet,
471: B. Friedrich, and J. M. Doyle, Nature (London) {\bf 395}, 148 (1998).
472: \bibitem{Bethlem} H. L. Bethlem {\it et al.}, Nature (London)
473: {\bf 406}, 491 (2000).
474: \bibitem{Meerakker} S. Y. T. van der Meerakker {\it et al.},
475: physics/0407116 (2004).
476: \bibitem{You} L.You and M.Marinescu, Phy. Rev. A {\bf 60}, 2324 (1999).
477: \bibitem{Goral1} K. Goral, B-G.Englert and K.Rzazewski, Phys.Rev.A {\bf 63}, 033606 (2001).
478: \bibitem{Baranov1} M.A.Baranov {\it et al.}, Phys.Rev.A {\bf 66},013606 (2002).
479: \bibitem{Baranov2} M.A.Baranov {\it et al.}, Physica Scripta, T.102,74-81, 2002.
480: \bibitem{Baranov3} M.A.Baranov, L.Dobrek, and M.Lewenstein, Phys.Rev.Lett. {\bf 92}, 250403 (2004).
481: \bibitem{Yi1} S.Yi and L.You, Phys.Rev.A, {\bf 61}, 041604(R) (2000).
482: \bibitem{Santos} L.Santos {\it et al.}, Phys.Rev.Lett. {\bf 85}, 1791 (2000).
483: \bibitem{Yi2} S.Yi and L.You, Phys.Rev.A, {\bf 63}, 053607 (2001).
484: \bibitem{Goral2} K.Goral and L.Santos, Phys.Rev.A. {\bf 66}, 023613 (2002).
485: \bibitem{Giovanazzi} S.Giovanazzi,A. Gorlitz and T.Pfau, Phys.Rev.Lett. {\bf 89}, 130401 (2002).
486: \bibitem{Odell} D.H.J. O'Dell, S.Giovanazzi, C.Eberlein, Phys.Rev.Lett. {\bf 92}, 250401 (2004).
487: \bibitem{Soldan} P. Sold\'{a}n and J. M. Hutson, Phys. Rev. Lett.
488: {\bf 92}, 163202 (2004).
489: \bibitem{Avdeenkov1} A. V. Avdeenkov and J. L. Bohn, Phys. Rev. A
490: {\bf 66} 052718 (2002).
491: \bibitem{Cornell} E. A. Cornell, C. Monroe, and C. E. Wieman,
492: Phys. Rev. Lett. {\bf 67}, 2439 (1991).
493: \bibitem{Spreeuw} R. J. C. Spreeuw {\it et al.}, Phys. Rev.
494: Lett. {\bf 72}, 3162 (1994).
495: \bibitem{Junglen} T.Junglen {\it et al.}, Phys.Rev.Lett. {\bf 92}, 223001 (2004).
496: \bibitem{DeMille} D. DeMille, D. R. Glenn, and J. Petricka,
497: physics/0407038.
498: \bibitem{Kajita} M. Kajita, Phys. Rev. A {\bf 69}, 012709 (2004).
499: \bibitem{Avdeenkov2} A. V. Avdeenkov, unpublished.
500: \bibitem{Avdeenkov3} A. V. Avdeenkov and J. L. Bohn, Phys. Rev.
501: Lett. {\bf 90}, 043006 (2003).
502: \bibitem{Avdeenkov4} A. V. Avdeenkov, D. C. E. Bortolotti, and J. L. Bohn,
503: Phys. Rev. A {\bf 69}, 012710 (2004).
504: \bibitem{Bochinski} J. R. Bochinski {\it et al.},
505: Phys. Rev. Lett. {\bf 91}, 243001 (2003).
506: \bibitem{Avdeenkov5} A. V. Avdeenkov and J. L. Bohn, Phys. Rev. A
507: {\bf 64}, 052703 (2001).
508: \bibitem{Volpi} A. Volpi and J. L. Bohn, Phys. Rev. A {\bf 65},
509: 052712 (2002).
510: \bibitem{Wigner} E. P. Wigner, Phys. Rev. {\bf 73}, 1002 (1948).
511: \bibitem{Landau} L. D. Landau and E. M. Lifshitz, {\it Quantum
512: Mechanics: Nonrelativistic Theory}, 4th ed. (Nauka, Moscow, 1989).
513: \bibitem{Spruch} L. Spruch, T. F. O'Malley, and L. Rosenberg,
514: Phys. Rev. Lett. {\bf 5}, 375 (1960).
515: \bibitem{OMalley} T. F. O'Malley, Phys. Rev. {\bf 134}, A1188 (1964).
516: \bibitem{Hinckelman} O. Hinckelman and L. Spruch, Phys. Rev. A
517: {\bf 3}, 642 (1971).
518: \bibitem{Shakeshaft} R. Shakeshaft, J. hys. B {\bf 5}, L115 (1972).
519: \bibitem{Holzwarth} N. A. Holzwarth, J. Math, Phys. {\bf 14},
520: 191 (1973).
521: \bibitem{Peach} G. Peach, J. Phys. B {\bf 12}, L13 (1979).
522: \bibitem{Fabrikant} I. I. Fabrikant, J. Phys. B {\bf 17}, 4223 (1984).
523: \bibitem{Rosenberg} L. Rosenberg, Phys. Rev. A {\bf 57}, 1862 (1988).
524: \bibitem{Cavagnero} M. Cavagnero, Phys. Rev. A {\bf 50}, 2841 (1994).
525: \bibitem{Gao} B. Gao, Phys. Rev. A {\bf 59}, 2778 (1999).
526: \bibitem{Sadeghpour} H. Sadeghpour {\it et al.}, J. Phys. B
527: {\bf 33}, R93 (2000).
528: \bibitem{Deb} B. Deb and L. You, Phys. Rev. A {\bf 64}, 022717 (2001).
529: \bibitem{Child} M. S. Child, {\it Molecular Collision Theory}
530: (Dover, New York, 1974).
531: \bibitem{Gradshteyn} I. S. Gradshteyn and I. M. Ryzhik, A. Jeffrey,
532: Ed, {\it Table of Integrals, Series, and Products} (Academic,
533: San Deigo, fifth edition), Sec. 6.574.
534: \bibitem{Abrams} M. C. Abrams, S. P. Davis, M.L.Rao, and R. Engelman, Jr.,
535: J. Mol. Spec. {\bf 165}, 57 (1994).
536: \bibitem{Ticknor} C. Ticknor and J. L. Bohn, unpublished.
537:
538: \end{references}
539:
540: \end{document}
541:
542:
543: