1: %\documentstyle[preprint,aps]{revtex}
2: \documentstyle[twocolumn,aps]{revtex}
3:
4: \def\btt#1{{\tt$\backslash$#1}}
5: \input epsf
6:
7: \begin{document}
8: \title{Rotational Feshbach Resonances in Ultracold Molecular Collisions}
9: \author{J. L. Bohn$^1$, A. V. Avdeenkov$^1$, and M. P. Deskevich$^2$}
10: \address{$^1$JILA and Department of Physics,
11: University of Colorado, Boulder, CO 80309 \\
12: $^2$Department of Chemistry and Biochemistry, University of Colorado, Boulder, CO
13: 80309}
14: \date{\today}
15: \maketitle
16:
17: \begin{abstract}
18: In collisions at ultralow temperatures, molecules will possess Feshbach
19: resonances, foreign to ultracold atoms, whose virtual excited states
20: consist of rotations of the molecules. We estimate the mean
21: spacing and mean widths of these resonant states, exploiting the fact
22: the molecular collisions at low energy display chaotic motion.
23: As examples, we consider the experimentally relevant molecules O$_2$, OH,
24: and PbO. The density of s-wave resonant states for these species
25: is quite high, implying that a large number of narrow resonant states
26: will exist.
27: \end{abstract}
28:
29: \pacs{34.10.+x,34.50.Ez}
30:
31: \narrowtext
32:
33: Scattering resonances are of great importance in ultracold collisions.
34: Feshbach resonances occur when the energy of a pair of free atoms (or molecules)
35: is nearly degenerate with that of a quasi-bound state of the pair. The
36: quasi-bound state is characterized by the promotion of one or both atoms
37: to an excited internal state, for example an excited hyperfine state
38: \cite{Stwalley}. External magnetic fields can then shift these states
39: in or out of resonance, giving the experimenter direct control over
40: interparticle interactions \cite{Tiesinga}. In this way, Bose-Einstein
41: condensates of $^{85}$Rb have been made stable or unstable on command, leading to
42: novel many-body effects \cite{Weiman}. It is moreover predicted that this
43: kind of control will be useful in preparing degenerate Fermi gases in
44: ``resonant superfluid'' states with any desired interaction strength
45: \cite{Holland,Timmermans}.
46: Magnetic field Feshbach resonances have now been observed in the alkali
47: atoms $^{23}$Na \cite{Ketterle}, $^{85}$Rb \cite{Heinzen,Roberts}, $^{133}$Cs
48: \cite{Chu}, and $^{40}$K \cite{Jin}. While not uncommon, these resonances
49: are far from ubiquitous in ultracold alkali atoms. This is largely due to
50: the relatively small number of hyperfine states available to form resonances.
51:
52: Ultracold molecules, by contrast, offer a far greater number
53: of resonances than alkali atoms, because rotational excitations can also
54: contribute to resonant states. In contrast to the two hyperfine states
55: in an alkali atom, a molecule can possess
56: a large number of energetically available rotational states. Considering
57: that ``typical'' rotational energy splittings are of order 1-10 K, while
58: the well depths of intermolecular potential energy surfaces (PES's) can be
59: hundreds or thousands of K, it is clear that tens of rotational states may contribute,
60: including their degeneracies arising from magnetic quantum numbers.
61:
62: Identifying the number and properties of these resonances is
63: important for understanding the behavior of a molecular gas reduced to
64: extremely low temperatures. These resonances may be useful as tools
65: for pinning down the details of multidimensional PES's. They may also
66: imply resonant control over cold molecular collisions, including, possibly,
67: control over chemical reactions \cite{Bala}. In addition,
68: the scattering dynamics is chaotic, implying that ultracold resonances
69: will form a new laboratory for studying quantum chaos \cite{Qchaos}.
70: In this Letter we estimate the most basic properties of these
71: resonances, namely, how many we can expect, and what their widths might be.
72: We are interested here in molecules produced in their ro-vibrational
73: ground states by buffer-gas cooling \cite{Doyle} or Stark slowing
74: \cite{Meijer} techniques. However, vibrationally excited cold molecules
75: produced by photoassociation of cold atoms will
76: also exhibit rich resonant dynamics \cite{Forrey}.
77:
78: Classical chaos at low energies has been studied extensively in
79: molecular scattering problems \cite{Atkins}. Of greatest relevance to
80: our present purposes are multiple collision events: it is possible
81: that a collision deposits sufficient energy in internal molecular degrees
82: of freedom (e.g., rotations and vibrations) that there is not enough
83: translational energy left to allow the molecules to separate.
84: The molecules may therefore collide many times before finally
85: shedding enough energy to separate. Multiple collisions are
86: increasingly likely as the collision energy is lowered, and in fact
87: are useful in quantifying the onset of chaos in this regime
88: \cite{Pattard}.
89:
90: \epsfxsize = 3in
91: \begin{figure}
92: \epsfbox{fig1.eps}
93: \vskip .5in
94: \caption{A classical trajectory for O$_2$-O$_2$ scattering at a collision
95: energy 1 mK. See text for details.}
96: \end{figure}
97:
98: Our classical calculations verify that multiple-collision resonances
99: persist at ultracold temperatures. We consider scattering of $^{17}$O$_2$
100: molecules, using the singlet rigid rotor potential energy surface of Ref.
101: \cite{Avoird}. At the collision energies considered, vibrational
102: excitation is not energetically allowed. To simplify the calculation
103: we constrain all the coordinates to fixed values except the intermolecular
104: distance $R$ and the orientations of the molecules in the scattering plane, as
105: shown in the inset to Figure 1. The main part of the figure shows
106: a slice through a sample phase space trajectory for collision energy 1 mK.
107: The axes are $R$ and the angular momentum $P_{\theta1}$
108: (in units of $\hbar$) of one of the molecules. As a guide, the
109: heavy solid line indicates the allowed region of phase space.
110: For the molecules to separate at this energy, the greatest
111: angular momentum either molecule could have is $0.027 \hbar$.
112: Instead, the molecules can exchange several $\hbar$ of angular momentum
113: in each collision, and the trajectory eventually fills the allowed
114: phase space.
115:
116: \epsfxsize = 3in
117: \begin{figure}
118: \epsfbox{fig2.eps}
119: \vskip .5in
120: \caption{Evidence of chaos in quantum mechanical scattering of
121: oxygen molecules. (a) shows the complete elastic scattering
122: cross section, computed as in Ref. [19]. In (b) the Dyson
123: rigidity function (points) is computed for the energy spacings of
124: the 33 resonances in (a). Our results are much more closely
125: related to the results of a gaussian orthogonal ensemble (GOE)
126: approximation to a chaotic system than to a nonchaotic regular
127: system.}
128: \end{figure}
129:
130: Chaotic motion in the classical realm leaves its signature in
131: the quantum mechanical spectrum as well. We have computed the quantum-
132: mechanical elastic scattering cross section for the $|N=0,J=1,M_J=1
133: \rangle$ fine-structure state of $^{17}$O$_2$ [Fig. 2(a)], following the model
134: described in Ref. \cite{Avdeenkov}. As in Ref. \cite{Avdeenkov},
135: we have neglected the nuclear spin of the $^{17}$O atoms, to emphasize
136: the effect of from rotational motion. This
137: model includes rotational levels up to $N=2$, and the partial
138: waves $L=0,2,...14$, and finds 33 resonances in this energy range.
139: Treating these levels as bound states of the (O$_2$)$_2$ collision
140: complex, we expect to find evidence of quantum chaos in statistical measures of
141: the energy level distributions. A glance at the spectrum in Fig. 2a)
142: suggests that there is no clustering of levels, but rather that they
143: are roughly evenly spaced. This ``rigidity'' of level spacings
144: is characteristic of the quantum mechanical
145: spectrum of a system that displays classical chaos, and is usually
146: quantified in terms of the spectral rigidity function $\Delta (x)$
147: of Dyson and Mehta \cite{Dyson}. If the number of states with energy
148: less than or equal to $E$ is denoted $N(E)$, then $\Delta (x)$
149: represents the root-mean-squared deviation of $N(E)$ from a
150: straight line over an energy range $x$, where $x$ is measured in units of the
151: mean level spacing. Figure 2(b) shows this function, computed using the
152: resonances from Fig 2(a). A spectrum consisting of uncorrelated energy levels would
153: yield a rigidity $\Delta (x) = x/15$, as shown with the dashed line.
154: A spectrum generated in the Gaussian orthogonal ensemble (GOE) approximation,
155: representing a chaotic system, yields instead the solid line.
156: The $\Delta (x)$ computed from the spectrum in Fig. 2(a) more closely
157: resembles the GOE result, suggesting that this system is
158: chaotic. A similar analysis has been applied to the eigenphase shifts
159: in reactive scattering processes \cite{shifts}.
160:
161:
162: Given that the classical trajectories can fill phase space (Fig.1),
163: each energetically allowed state is equally likely to be populated
164: during a resonant collision. This observation has been exploited in the theory of
165: unimolecular chemical reactions, which considers the breakup of a
166: large molecule under the influence of random collisions with other molecules
167: in a solution \cite{unimol}. We use the same idea here to estimate
168: the total number of resonant states available in a given energy range.
169: To this end we separate the molecule-molecule Hamiltonian into three
170: independent terms:
171: \begin{equation}
172: {\hat H} = {\hat H}_1 + {\hat H}_2 + {\hat H}_{\rm int},
173: \end{equation}
174: where ${\hat H}_1$ and ${\hat H}_2$ represent the rotational fine structure
175: of each individual molecule, and ${\hat H}_{\rm int}$ is an effective
176: intermolecular potential that depends only on $R$.
177: Each Hamiltonian is represented by a set of
178: independent energy eigenvalues $E_1$, $E_2$, and $E_{\rm int}$.
179: During a resonant collision any partition of the
180: total energy $E = E_1 + E_2 + E_{\rm int}$ is equally likely, provided that
181: angular momentum is conserved. Estimating the number of states $N(E)$
182: at or below energy $E$ is then a simple counting exercise. The density of
183: states, $\rho(E) = dN/dE$, is also easily computed.
184:
185:
186: To illustrate that this is a reasonable estimate of the density of states,
187: we compare the results to those of a fully quantum mechanical calculation
188: for $^{17}$O$_2$ - $^{17}$O$_2$ scattering \cite{Avdeenkov}.
189: Table I tabulates, for the quantum mechanical calculation, the number of
190: resonances in each partial wave $L$ in an energy range 1K above the incident
191: threshold. Each row in the table represents a separate calculation,
192: characterized by the maximum number of molecular rotation states ($N_{\rm max}$)
193: and partial wave states ($L_{\rm max}$) included in the basis set.
194:
195:
196: Table II reports the same information, but
197: this time estimated from the statistical model. Statistical fluctuations in the
198: density of states function $\rho(E)$ in the relevant energy range
199: correspond to uncertainties of $\sim 50$ \% in the numbers given.
200: Within this uncertainty the agreement with the quantum mechanical
201: calculation in Table I is quite good, giving us further confidence that the
202: statistical estimate is reasonable for this purpose.
203:
204: A clear advantage of the statistical model over the full calculation is that
205: it can easily be extrapolated to include {\it all} allowed values of
206: $N$ and $L$ consistent with conservation of energy and angular
207: momentum, as well as Bose symmetry. The result
208: is given by the final row in Table II, for small values of the incident
209: partial wave. Considering all partial waves yields a total number of
210: resonances in this 1K interval of order $\sim 10^3$. However, in ultracold
211: collisions we are limited to low values of $L$, typically $L=0$ alone,
212: barring occasional shape resonances with higher values of $L$. We therefore
213: conclude from Table II that the mean density of s-wave rotational resonances
214: in O$_2$ is on the order of 10 K$^{-1}$. For comparison, alkali atoms
215: typically exhibit on the order of 1 s-wave Feshbach resonance per 1 K energy
216: interval. Thus rotational Feshbach resonances are slightly more numerous
217: than atomic hyperfine Feshbach resonances, but not strikingly so, at least for
218: nonpolar molecules.
219: (In practice the numerous hyperfine levels in $^{17}$O$_2$ will play a
220: role also, but we are concerned here strictly with rotation.)
221:
222: This simple statistical theory can also estimate the widths of these resonances.
223: Again borrowing from the theory of unimolecular dissociation, the
224: resonance width may be approximated in the
225: Rice-Ramsperger-Kassel-Marcus (RRKM) approximation as
226: \cite{unimol}
227: \begin{equation}
228: {\bar \Gamma } = {W(E^+) \over 2 \pi {\bar \rho} },
229: \end{equation}
230: where ${\bar \rho}$ is the mean density of states we have been considering,
231: and $W(E^+)$ is the number of states that permit the molecules to
232: actually escape to infinite separation. For the low collision energies
233: considered here, both molecules must return to their ground state to separate,
234: so that $W(E^+)=1$. In the quantum mechanical calculation in Fig. 2(a),
235: there are 33 resonances in a 1 Kelvin interval, implying a mean width
236: in this approximation of ${\bar \Gamma} \approx 5$ mK. For comparison,
237: the arithmetic mean of the quantum mechanical resonance widths is 8 mK.
238: Note that the quantum mechanical widths span several orders of magnitude;
239: their standard deviation is not given by the RRKM result \cite{note}.
240:
241: The situation is dramatically different for polar molecules, whose
242: long range dipole-dipole interaction potential, proportional to $1/R^3$,
243: holds many more intermolecular bound states near threshold.
244: To illustrate this point we consider two examples: OH,
245: which is suitable for Stark slowing \cite{Meijer}; and the $a(1)$ state
246: of PbO, which is a leading candidate in which to measure the electric dipole
247: moment of the electron \cite{DeMille,Egorov}. Both molecules possess,
248: in addition to a dipole moment, a pair of closely-spaced states of opposite
249: parity: a $\Lambda$ doublet of splitting $\Delta = 0.081$ K in OH,
250: and an $\Omega$-doubling of splitting $\Delta =260$ $\mu$K in PbO.
251: Intermolecular potential curves correlating to these thresholds will
252: hold a large number of dipole-bound resonant states. Details of
253: the fine structure of OH is given in \cite{Coxon}, and that of PbO
254: in \cite{Martin}.
255:
256: For the OH-OH interaction an approximate PES exists \cite{Kuhn}.
257: We include the long-range part of this interaction, which has
258: a minimum at $R \approx 6$ a.u. that arises from hydrogen bonding
259: forces. For the sake of our statistical model, we approximate
260: this surface with a single curve by fixing the two OH molecules in
261: their linear geometry. The intermolecular bound state energies
262: $E_{\rm int}$ are then simply the bound states of this potential,
263: for each of the partial waves considered. We ignore for our present
264: purposes an even deeper part of the PES corresponding to chemical binding
265: of hydrogen peroxide, H$_2$O$_2$.
266: For PbO-PbO the PES is entirely unknown; however, since the bound states of interest
267: depend primarily on the long-range potential, we estimate the PbO-PbO
268: PES by multiplying that of OH-OH by the appropriate ratio of squared
269: dipole moments, $[d_{\rm PbO}/d_{\rm OH}]^2 \approx 7.7$. Table III tabulates
270: the number of s-wave resonances to be expected for each species.
271: These numbers refer to the important energy range $2 \Delta$ above
272: threshold, where the number of resonances is greatest. Expressed as
273: a mean density-of-states over this range, the densities are 240
274: K$^{-1}$ for OH and $2 \times 10^5$ K$^{-1}$ for PbO,
275: significantly larger than for the nonpolar O$_2$.
276:
277: In general, the resonances shift when an external electric field
278: is applied, so that one after the other can be brought into resonance
279: with the essentially zero-energy molecules. Given a mean density of
280: states $\approx 10^2$ K$^{-1}$ for OH,
281: and given the OH dipole moment (=1.67 D), the mean spacing
282: of these resonances as a function of field would be of order $\sim 250$ V/cm.
283: This means that, in contrast to the magnetic field resonances in alkali
284: atoms, there will be many available resonances in ultracold polar
285: molecule collisions.
286:
287: We conclude by sketching the implications of this work for the theory
288: of ultracold collisions. As is well-known, ultracold collision theory
289: requires a certain level of experimental input to fine-tune the scattering
290: Hamiltonian. It is not yet clear how much or what kind of information
291: is required to perform this fine-tuning for molecules \cite{Hutson}.
292: Knowing roughly how many resonances to expect can greatly simplify theory, by
293: providing an estimate of how many parameters are strictly necessary
294: to reproduce the data. With this information theory could then turn around
295: and predict how these resonances evolve with applied external electric and
296: magnetic fields.
297:
298: A second, equally important implication concerns experiments at temperatures
299: where $kT$ is large compared to the mean spacing between resonances.
300: In this case individual resonances are not resolved, and only the overall
301: modulation of the spectrum matters. It is then likely that
302: semiclassical methods \cite{Delos,Granger} should prove insightful.
303:
304: This work was supported by the NSF and by the OSEP program at the
305: University of Colorado. We acknowledge an allocation of
306: parallel computer time by NPACI, and useful discussions with
307: D. Nesbitt.
308:
309: \begin{references}
310: \bibitem[*]{byline} email address: bohn@murphy.colorado.edu
311: \bibitem{Stwalley} W. C. Stwalley, Phys. Rev. Lett. {\bf 37}, 1628 (1976).
312: \bibitem{Tiesinga} E. Tiesinga, B. J. Verhaar, and H. T. C. Stoof,
313: Phys. Rev. A {\bf 47}, 4114 (1993).
314: \bibitem{Weiman} J. L. Roberts {\it et al.}, Phys. Rev. Lett. {\bf 86}, 4211 (2001).
315: \bibitem{Holland} M. Holland, S. J. J. M. Kokkelmans, M. L. Chiofalo,
316: and R. Walser, Phys. Rev. Lett. {\bf 87}, 120406 (2001).
317: \bibitem{Timmermans} E. Timmermans, K. Furuya, P. W. Milloni, and
318: A. K. Kerman, Phys. Lett, A {\bf 285}, 228 (2001).
319: \bibitem{Ketterle} S. Inouye {\it et al.}, Nature {\bf 392}, 151 (1998).
320: \bibitem{Heinzen} Ph. Courteille {\it et al.}, Phys. Rev. Lett. {\bf 81}, 69 (1998).
321: \bibitem{Roberts} J. L. Roberts {\it et al.}, Phys. Rev. Lett. {\bf 81}, 5109 (1998).
322: \bibitem{Chu} V. Vuleti\'{c}, A. J. Kerman, C. Chin, and S. Chu, Phys. Rev. Lett.
323: {\bf 82}, 1406 (1999).
324: \bibitem{Jin} T. Loftus {\it et al.}, Phys. Rev. Lett. {\bf 88}, 173201 (2002).
325: \bibitem{Bala} That chemical reactions may occur at ultralow temperatures
326: is suggested by N. Balakrishnan and A. Dalgarno, Chem. Phys. Lett. {\bf 341},
327: 652 (2001).
328: \bibitem{Qchaos} M. C. Gutzwiller, {\it Chaos in Classical and Quantum Mechanics}
329: (New York, Springer-Verlag, 1990).
330: \bibitem{Doyle} J. D. Weinstein {\it et al.}, Nature {\bf 395}, 148 (1998).
331: \bibitem{Meijer} H. L. Bethlem {\it et al.}, to appear in Phys. Rev. A.
332: \bibitem{Forrey} R. C. Forrey {\it et al.}, Phys. Rev. Lett. {\bf 82}, 2657 (1999).
333: \bibitem{Atkins} K. M. Atkins and J. M. Hutson, J. Chem. Phys.
334: {\bf 103}, 9218 (1995), and references therein.
335: \bibitem{Avoird} P. E. S. Wormer and A. van der Avoird, J. Chem. Phys.
336: {\bf 81}, 1929 (1984).
337: \bibitem{Pattard} T. Pattard and J. M. Rost, Chem. Phys. Lett. {\bf 291}, 360
338: (1998).
339: \bibitem{Avdeenkov} A. V. Avdeenkov and J. L. Bohn, Phys. Rev. A {\bf 64},
340: 052703 (2001).
341: \bibitem{Dyson} F. J. Dyson and M. L. Mehta, J. Math. Phys. {\bf 4},
342: 701 (1963).
343: \bibitem{shifts} P. Honvault and J.-M. Launay, Chem. Phys. Lett. {\bf 329},
344: 233 (2000).
345: \bibitem{unimol} R. G. Gilbert and S. C. Smith, {\it Theory of Unimolecular
346: and Recombination Reactions} (Blackwell Scientific, 1990).
347: \bibitem{note} In addition, resonance widths can display signautres of
348: quantum chaos, but the limited statistics of this sample prevent us from
349: drawing any conclusions in this regard.
350: \bibitem{Kuhn} B. Kuhn {\it et al.}, J. Chem. Phys. {\bf 111}, 2565 (1999).
351: \bibitem{DeMille} D. DeMille {\it et al.}, Phys. Rev. A {\bf 61}, 052507 (2000).
352: \bibitem{Egorov} D. Egorov {\it et al.}, Phys. Rev. A {\bf 63}, 030501 (2001).
353: \bibitem{Coxon} J. A. Coxon, Can. J. Phys. {\bf 58}, 933 (1980).
354: \bibitem{Martin} F. Martin {\it et al.}, Spectrochim. Acta {\bf 44A}, 889 (1988).
355: \bibitem{Hutson} M. Muewly and J. M. Hutson, J. Chem. Phys. {\bf 110},
356: 8338 (1999).
357: \bibitem{Delos} M. L. Du and J. B. Delos, Phys. Rev. A {\bf 38},
358: 1896 (1988); {\it ibid.}, {\bf 38}, 1913 (1998).
359: \bibitem{Granger} B. E. Granger and C. H. Greene, Phys. Rev. A {\bf 62},
360: 012511 (2000).
361:
362: \end{references}
363:
364:
365: \begin{table}
366: \caption{Number of resonances in a 1K energy interval above threshold for
367: cold collisions of $^{17}$O$_2$
368: molecules in their $J=M_J=+1$ state, neglecting nuclear spin. These results
369: are determined from a quantum mechanical calculation. Various levels
370: of completeness of this calculation are specified by the maximum value
371: of rotational quantum number $N_{\rm max}$ and maximum partial wave
372: $L_{\rm max}$. For each such calculation, the table gives the number
373: of resonances found separately in each incident partial wave $L$.}
374: \label{table1}
375: \begin{tabular}{cccccccc}
376: $N_{\rm max}$ & $L_{\rm max}$ &
377: $L=0$ & $L=2$ & $L=4$ & $L=6$ & $L=8$ & $L=10$ \\
378: \tableline
379: 2 & 2 & 1 & 4 & $-$ & $-$ & $-$ & $-$ \\
380: 2 & 4 & 2 & 7 & 8 & $-$ & $-$ & $-$ \\
381: 2 & 6 & 2 & 7 & 9 & 8 & $-$ & $-$ \\
382: 2 & 8 & 2 & 10 & 14 & 9 & 10 & $-$ \\
383: 2 & 10 & 3 & 10 & 14 & 11 & 11 & 6 \\
384: 4 & 4 & 3 & 11 & 15 & $-$ & $-$ & $-$ \\
385: \end{tabular}
386: \end{table}
387:
388: \begin{table}
389: \caption{Number of resonances for $^{17}$O$_2$ cold collisions,
390: as in Table I. Here, however, the
391: resonances are estimated using the statistical counting model described
392: in the text. In addition, the final row extrapolates to the limit where
393: $N_{\rm max}=8$ and $L_{\rm max}=36$, thus including all possible
394: energetically allowed states of the $(^{17}$O$_2)_2$ dimer.}
395: \label{table2}
396: \begin{tabular}{cccccccc}
397: $N_{\rm max}$ & $L_{\rm max}$ &
398: $L=0$ & $L=2$ & $L=4$ & $L=6$ & $L=8$ & $L=10$ \\
399: \tableline
400: 2 & 2 & 2 & 5 & $-$ & $-$ & $-$ & $-$ \\
401: 2 & 4 & 3 & 8 & 11 & $-$ & $-$ & $-$ \\
402: 2 & 6 & 3 & 10 & 15 & 11 & $-$ & $-$ \\
403: 2 & 8 & 3 & 10 & 16 & 14 & 9 & $-$ \\
404: 2 & 10 & 3 & 10 & 16 & 16 & 12 & 9 \\
405: 4 & 4 & 5 & 15 & 24 & $-$ & $-$ & $-$ \\
406: \tableline
407: All & All & 14 & 45 & 78 & 83 & 85 & 84 \\
408: \end{tabular}
409: \end{table}
410:
411:
412:
413: \begin{table}
414: \caption{Summary of s-wave resonances for two polar molecules.
415: $\rho_{\Delta}$ is the average density of states consistent with
416: s-wave initial conditions, over an
417: energy range $2 \Delta$ above threshold, where $\Delta$ refers to the
418: $\Lambda$ ($\Omega$)-doublet of OH (PbO). Using the RRKM result,
419: this density translates into the mean energy width ${\bar \Gamma}_{\Delta}$.}
420: \label{table3}
421: \begin{tabular}{cccc}
422: Molecule & $\Delta$ (K) & $\rho _{\Delta}$ (K$^{-1}$) & ${\bar \Gamma}_{\Delta}$
423: ($\mu$K)\\
424: \tableline
425: OH & 0.081 & 240 & 660\\
426: PbO & $2.6 \times 10^{-4}$ & $1.9 \times 10^5$ & $5 $\\
427: \end{tabular}
428: \end{table}
429:
430:
431:
432:
433:
434:
435: \end{document}
436: