RbCs Project: Research

We have built an experiment that produces a two-species quantum degenerate gas of rubidium-87 and caesium in the same trapping potential [1]. This result acts as the starting point for two main avenues of research. The first is the direct study of the quantum degenerate mixture and will allow investigations into tunable interspecies interactions [2] and the myriad of quantum phases displayed in optical lattices [3]. The second involves the efficient production of RbCs molecules via magneto-association followed by optical transfer into the rovibrational ground state. Possessing long-range, anisotropic dipole-dipole interactions, such molecules offer a wealth of new research directions including quantum simulation [4] and computation [5].

Below you will find a brief summary of the experimental method and future steps along the two research avenues of studying a quantum degenerate mixture and producing ultracold RbCs molecules in the rovibrational ground state. Further details can be found in our recent publications and references therein.

If you are totally unfamiliar with Bose-Einstein condensation, we suggest you first take a look at the Bose-Einstein condensation section of this page, which contains a very basic non-mathematical introduction to the central concepts. More technical information may be found by following the links on our website.

Dual-Species BEC

Loading the Magnetic Trap

Initially ultracold mixtures of 87Rb and133Cs are collected in a two-species MOT in a UHV cell. Up to 3x108 Rb atoms are collected while the Cs atom number is servomechanically controlled between 5x105 and 3x108. To create a degenerate mixture of 87Rb and 133Cs 2x108 87Rb and 4x106 133Cs atoms are collected in the MOT. Following a compressed MOT and molasses phase the atoms are optically pumped into the 87Rb |F=1, mF=-1> and 133Cs |3, -3> states and loaded into the magnetic quadrupole trap. After loading and adiabatic compression to 187 G/cm the trap contains 1.4x108 87Rb atoms and 2.7x106 133Cs atoms at a temperature of 160 µK.

RF Evaporation and Loading the Dipole Trap

In the quadrupole trap the trap depth set by the RF frequency is three times deeper for 133Cs than for 87Rb. This allows the selective removal of 87Rb while interspecies elastic collisions sympathetically cool the 133Cs. Once the mixture is cooled below ~70 µK Majorana spin-flip losses begin to limit the efficiency of any further cooling. At this point the trap contains 3.1x107 87Rb atoms and 1.8x106 133Cs atoms at 68 µK.

At this point the crossed dipole trap is loaded by reducing the magnetic field gradient to 29 G/cm and the atomic spins are flipped into the |1, +1> and |3, +3> states for Rb and Cs respectively as a 22.4 G bias field is switched on. In the dipole trap the mixture consists of 2.8x106 87Rb and 5.1x105 133Cs atoms at 9.6 µK. These parameters correspond to phase-space densities of 2x10-3 for 87Rb and 9x10-4 for 133Cs.

Interspecies Three-Body Loss

Once loaded into the crossed dipole trap the peak atomic densities for both species are typically >1013 cm-3. When both species are present in the trap very strong interspecies inelastic losses are observed, see figure. The open symbols are for 87Rb and the closed symbols are for 133Cs. In the main plot the 133Cs minority species lifetime is 0.8 s and the two 87Rb lifetimes are 4 s and 70 s. In the inset the 87Rb minority species has a lifetime of 0.9 s and the two 133Cs lifetimes are 2 s and 10 s. A simple calculation assumes that the trap lifetime is due exclusively to three-body losses and does not account for any evaporation from the trap. This method estimates upper limits for the interspecies three-body loss rate coefficients of ~10-25 - 10-26 cm6/s. Data taken from reference [6].

Initial Evaporation in the Dipole Trap

To combat the strong interspecies inelastic losses the trap depth is reduced from 90 µK to 2 µK in just 1.0 s by reducing the beam powers from 6 W to 120 mW. Accounting for the adiabatic expansion of the cloud resulting from the commensurate relaxation of the trap frequencies, this corresponds to an effective reduction of the trap depth by a factor of 7. The result is a factor of ~3 reduction in the density and consequently an order of magnitude increase in the 133Cs trap lifetime. Here 4.7x105 87Rb atoms and 1.5x105 133Cs atoms have phase-space densities of 5.3x10-2 and 2.1x10-2 respectively.

Evaporation to BEC

Further evaporation is performed by reducing beam powers over 2.5 s to produce a trap 800 nK deep. Finally by tilting the trap by increasing the applied magnetic field gradient to 32.7 G/cm over 1.0 s dual-species condensates are produced in the same trapping potential containing up to ~2x104 atoms of each species. The trajectory to BEC for both species is shown in the figure and is divided into four sections: (i) RF evaporation in the magnetic quadrupole trap, (ii) dipole trap loading and internal state transfer, (iii) evaporation by reducing the dipole trap beam powers, and (iv) evaporation by trap tilting. Open (closed) symbols show data for 87Rb (133Cs).

Another experiment based in Innsbruck produces condensates of 87Rb and 133Cs in separate dipole traps to avoid the strong interspecies inelastic losses [7].

Immiscibility in a Degenerate Mixture

A dramatic spatial separation of the two condensates in the trap is observed, revealing the mixture to be immiscible at 22.4 G. As shown in the figure the immiscible behaviour is exclusive to the condensed atoms. The condensates always forms one of three structures, either one of two symmetric cases (triangles and circles) or an asymmetric case (squares). Experimentally, the structures are correlated with the number of atoms in each condensate. The strong interspecies losses make the formation of a dual-species condensate with >2x104 atoms of each species a challenge and lead to the asymptotic nature of the data in the figure.

The next step is to explore the miscibility of the degenerate mixture by Feshbach tuning both the 133Cs and interspecies scattering lengths [2].

Ultracold heteronuclear molecules

Interspecies Feshbach Resonance

Once a high phase-space density 87Rb 133Cs mixture has been made via the method (briefly!) outlined above the search for interspecies Feshbach resonances can begin. These features will be used to magneto-associate weakly bound Rb-Cs Feshbach molecules.

Feshbach resonances are most easily detected through an enhancement of trap loss. To increase the sensitivity of heteronuclear Feshbach spectroscopy a significant imbalance between the two-species atom numbers is useful. Here the majority species acts as a collisional bath for the minority species which is used as a probe. If the atom number imbalance is large the probe species will be significantly depleted when a resonance is encountered.

An RbCs interspecies Feshbach resonance is shown in the figure. Here the 133Cs atom number was measured after a 5 s hold at a specific bias field. A Lorentzian fit yields a resonance position of 181.7(5) G in excellent agreement with previous measurements of this resonance [8].

Locating an interspecies Feshbach resonance now opens the possibility of creating RbCs molecules via magneto-association.

Feshbach Association of Cs2 Dimers

To develop experimental protocols and detection methods the magneto-association of Cs2 dimers has been explored using a 133Cs Feshbach resonance at 19.8 G. This resonance is crossed from high to low field at a rate of 47 G/s to form the weakly bound Feshbach molecules. The magnetic field gradient is fixed at 31.1 G/cm to levitate the atomic cloud. During this variable hold time the molecules fall away from the atomic cloud due to their smaller magnetic moment to mass ratio. To dissociate the molecules for imaging the field is jumped back across the resonance to 21 G in 130 µs.

The figure shows the atomic and molecular clouds positions as a function of time and a typical absorption image. Magneto-association efficiencies are ~12% and the molecular acceleration at 31.1 G/cm is measured to be 3.86 m/s2 which corresponds to a magnetic moment of 0.92μB. This value concurs with another measurement performed by levitating the molecules at 51.4 G/cm.

Future Work and Applications

STIRAP Transfer

Once weakly bound RbCs Feshbach molecules have been produced optical transfer to the rovibrational ground state will be performed via stimulated Raman adiabatic passage (STIRAP) [9]. This coherent transfer uses two laser frequencies to couple the initial and final molecular states via a an intermediate excited state. The complete process for the production of ultracold polar RbCs molecules is presented in the figure.

Work on the cavity that will be used to lock the two laser frequencies is underway and can be seen in our news section. Recent molecular spectroscopy by the Innsbruck group [10] has identified the transition frequencies for these states in Rb Cs to be ~1570 nm and ~980 nm. This knowledge will greatly accelerate our progress towards polar RbCs molecules.

Applications of Polar RbCs Molecules

Ground state RbCs molecules possess permanent electric dipole moments which give rise to anisotropic, long range dipole-dipole interactions. To realise an electric dipole moment an external electric field must be applied in the laboratory frame to polarise the molecular sample. The direction of the external electric field controls the orientation of the molecular dipoles. This control, combined with the trapping geometry, enables interactions within this quantum system to be tuned with exquisite sensitivity.

When loaded into an optical lattice the long-range interaction of these dipoles leads to a rich spectrum of quantum phases [11] and offers potential applications for quantum information processing [5], simulation [4] and precision metrology [12]. Zoller et al have proposed a method for creating such molecules in which a Mott Insulator (MI) transition is applied to a two-species BEC confined in a lattice and then a photoassociation sequence is used to produce a molecular BEC [13,14].


  1. D. J. McCarron, H. W. Cho, D. L. Jenkin, M. P. Koeppinger, and S. L. Cornish, Phys. Rev. A 84, 011603 (2011)
  2. S. B. Papp, J. M. Pino, and C. E. Wieman, Phys. Rev. Lett. 101, 040402 (2008)
  3. E. Altman, W. Hofstetter, E. Demler, and M. D. Lukin, New J. Phys. 5, 113 (2003)
  4. A. Micheli, G. K. Brennen and P. Zoller, Nat. Phys. 2, 341 (2006)
  5. D. DeMille, Phys Rev Lett 88, 067901 (2002)
  6. H. W. Cho, D. J. McCarron, D. L. Jenkin, M. P. Koeppinger and S. L. Cornish, Eur. Phys. J. D (DOI: 10.1140/epjd/e2011-10716-1)
  7. A. D. Lercher, T. Takekoshi, M. Debatin, B. Schuster, R. Rameshan, F. Ferlaino, R. Grimm, and H.-C. Naegerl, Eur. Phys. J. D (DOI: 10.1140/epjd/e2011-20015-6)
  8. K. Pilch, A. D. Lange, A. Prantner, G. Kerner, F. Ferlaino, H.-C. Naegerl, R. Grimm, Phys. Rev. A 79, 042718 (2009)
  9. K. Bergmann, H. Theuer, and B. W. Shore, Rev. Mod. Phys. 70, 1003 (1998)
  10. M. Debatin, T. Takekoshi, R. Rameshan, L. Reichsoellner, F. Ferlaino, R. Grimm, R. Vexiau, N. Bouloufa, O. Dulieu, H.-C. Naegerl, arXiv:1106.0129 (2011)
  11. B. Capogrosso-Sansone, C. Trefzger, M. Lewenstein, P. Zoller, and G. Pupillo, Phys. Rev. Lett. 104, 125301 (2010)
  12. T. Zelevinsky, S. Kotochigova, and J. Ye, Phys. Rev. Le>tt. 100, 043201 (2008)
  13. D. Jaksch, V. Venturi, J. I. Cirac, C. J. Williams, P. Zoller, Phys Rev Lett 89, 040402 (2002)
  14. B. Damski, L. Santos, E. Tiemann, M. Lewenstein, S. Kotochigova, P. Julienne, P. Zoller, Phys Rev Lett 90, 110401 (2003)
Updated by Simon L. Cornish, May 2010