Formation of Ultracold NaLi Molecules
We have succeeded in forming NaLi molecules from an ultracold mixture of Na and Li atoms. This is done by carefully sweeping a magnetic field around a Feshbach resonance, where the energy for a state of two free atoms becomes degenerate with the energy for the bound molecular state of the two atoms, but with different internal electronic and nuclear spins. A magnetic field can be used to tune the energy difference between these two states to zero because the different spin orientations lead to different magnetic moments of the two-atom system.
Usually, when the magnetic field is swept across a Feshbach resonance, the atom pair is adiabatically transferred to the molecular bound state because the two are coupled by the hyperfine interactions in the system. However, in the Na + Li system, such hyperfine-induced Feshbach resonances are at very high magnetic fields that are out of experimental range. Instead, we worked around a Feshbach resonance at 745G that is produced by weak dipole-dipole coupling between the atoms. This coupling term is orders of magnitude weaker than the hyperfine interaction, meaning that the requirements for successful adiabatic conversion of atom pairs to molecules become extraordinarily demanding. However, by carefully controlling our magnetic field stability and sweep sequence, we were able to rapidly jump the field near resonance, do a slow, adiabatic sweep across the very narrow, ~mG wide range of the Feshbach resonance, and then immediately jump the field away from resonance again to isolate the molecules for imaging. This produced a fraction of NaLi molecules from our initial atomic mixture, and perhaps represents successful molecule formation around the most difficult Feshbach resonance ever used.
In the near future, we plan to take advantage of the fact that NaLi is the lightest heteronuclear alkali molecule. This means that it is expected to have a long-lived metastable spin-triplet ground state, with decay to lower lying spin-singlet states suppressed by the smallness of second-order spin-orbit coupling in systems with small constituent atomic charges. In the spin-triplet ground state, NaLi would be the first molecule with both a significant electric dipole moment and a nonzero magnetic dipole moment. The simultaneous presence of two independently tunable dipole moments opens the door to engineering exotic new Hamiltonians in optical lattices, as well as the possibility of exploring new ways to tune chemical reactions with applied fields. In particular, because of the small mass of the system, NaLi has an exceptionally low density of states, meaning that there can be the possibility of observing discrete, well-separated resonances in molecule-molecule collisions, which would enable resonant tuning of reaction rates for ultracold molecules.
The paper we wrote about our work can be found here
Forming Ultracold molecules from Ultracold Atoms
The power of Feshbach resonances
Feshbach resonances occur in a cold atomic gas when there is a possible bound molecular state with magnetic moment which differs from the vector sum of the magnetic moments of its constituent atoms. This allows the energy of the molecular state to be adjusted relative to the energy of the free atoms. Things become interesting when the energies are nearly equal at magnetic fields which can be practically generated in the presence of ultracold atoms. Under such circumstances, coupling between the free and bound states can raise or lower the energies of the free atoms (effectively causing repulsion or attraction as the atoms try to move to lower their energy), or can allow the free atoms to bind into a molecule while releasing very little energy. The repulsion and attraction have led to many interesting studies of strongly interacting many-body systems (see the ferromagnetism project on this page, or one of many projects on atomic BCS superfluids), but the ability to form molecules without releasing much energy allows ultracold chemistry to take place.
Feshbach association of Ultracold molecules
While molecules can be formed by allowing thermal atoms to mingle until they crash together and stick (as in traditional chemistry or 3-body collisions in clouds of atoms), or by blasting them with appropriately tuned photons (as in photoassociation), these processes tend to relase a lot of energy because the final molecular state is tightly bound and has a much lower energy than the inital atomic state. This energy must dissipate somehow, and it generally leads to heating of the sample of atoms. Such heating defeats the efforts taken to bring the atoms to ultracold temperatures in the first place.
By tuning the energy of the molecular state to be close to that of the free atomic states, very little if any energy is released into the sample. The atoms can be nudged into a molecules with a low energy radiofrequency photon, or the magnetic field can be changed gradually so that the atoms bind at the moment when there is no energy difference at all. Both of these methods have been used with some success in other atoms (for example, two lithium atoms can form very stable gasses of diatomic lithium molecules).
When the two pairing partners are different atoms, the resultant molecule can be quite complex. It can have a strong electric dipole moment, and the quantum dynamics resulting from the rotation and vibration of the molecules are much richer than the physics of atoms or diatomic molecules, which are simple by comparison. This has led to a push to discover and characterize heteronuclear molecules which can be formed at ultracold temperatures. There has already been great success in forming RbK molecules and studying their behavior, and work is well underway to study LiCs, LiK, and other combinations.
We are working to form stable molecules comprised of one lithium and one sodium atom. These two species have proven to cooperate well together, as their favorable collision properties allow them to collide many times without unwanted inelastic reactions. Furthermore, because we can generate large samples of bosonic Sodium-23 and fermionic Lithium-6 (several million atoms of each), we may be able to produce large clouds of fermionic NaLi molecules. Such a large fermionic dipolar gas has interesting applications in the study of complicated many-body systems, as well as potential applications to quantum information.
Practical challenges for forming ultracold NaLi molecules
One major concern when creating such strongly interacting clouds of atoms and encouraging them to form molecules is the potential for 3-body losses. Because the free and bound states are nearly resonant (very low energy difference), two atoms will spend a long time in close vicinity, greatly encouraging the chance that a third atom will approach and enable an inelastic 3-body collions. This is the uncontrolled formation of a deeply bound molecule which dumps its energy into the third particle, and all 3 atoms will be lost from the trap. Such 3-body losses effectively give a time limit to association molecules, isolate them from other atoms which could inelastically collide, and observe the presence of molecules.
Another challenge is the fact that properties of the NaLi system are not known more accurately than can be experimentally measured, as the 31 body system of 2 nuclei and 29 electrons is too complicated for accurate rigorous calculations. This means that times scales of 3-body loss, molecule association, and decay into more deeply bound molecular states are not accurately known, and binding energies of the molecule at specific magnetic fields cannot be determined without more accurate measurements of the exact resonance position. Unfortunately, more accurate measurements require the presence of stable, measurable molecules. Thus, forming the first repeatable, measurable clouds of NaLi molecules requires a nearly blind search over a variety of parameters and efforts to optimize 2-body molecule association rates, 3-body loss rates, removal of unpaired atoms, and lifetimes of weakly bound molecules. Although we have not yet observed stable, isolated molecules, all theoretical estimates and preliminary observations suggest that it should be possible, and nearly within reach.