Our previous work established two successful methods for producing ultracold diatomic molecules: association of ultracold atoms and direct laser cooling. This led to the development of two state-of-the-art molecular sources – ultracold RbCs in Durham and CaF at Imperial. We have also developed the techniques for producing single organic dye molecules in a host matrix, which behave as nearly ideal two-level systems when cooled to low temperature. Over the course of QSUM, we will investigate how to align the dipoles of the dye molecules with respect to nanophotonic structures on surfaces, we will enhance the sources of RbCs and CaF, and extend atom association to develop sources of CsYb and KCs molecules.


We produce ultracold RbCs molecules in the rovibrational and hyperfine ground state by  associating ultracold Rb and Cs atoms using a two-step process. Starting from an atomic mixture with high phase-space density, we use magnetoassociation on a Feshbach resonance to produce weakly bound molecules. These molecules are then coherently transferred to a single deeply bound rovibrational state of the ground electronic state with near unity efficiency, using  stimulated Raman adiabatic passage (STIRAP). The final temperature mirrors that of the original atomic gas, allowing us to exploit the established techniques of atomic cooling to bring molecules into the ultracold regime.

Transfer of RbCs molecules to the absolute ground state using STIRAP.

From the outset we will use our existing RbCs apparatus to (i) further characterise the ground state molecules, developing quantum control of the hyperfine and rotational states, (ii) load ‘tunable’ optical lattices and (iii) investigate quantum magnetism in 1D, 2D and 3D lattices. In parallel, we will build a new molecular microscope apparatus capable of single-site imaging in an optical lattice, and will use RbCs to develop the tweezers approach to realising small molecular arrays.


We use direct laser cooling to cool CaF molecules. An intense beam of CaF is produced by a cryogenic buffer-gas source and slowed to low velocity by a frequency-chirped counter-propagating laser beam containing all frequencies needed to produce a closed cycling transition. The slow molecules will now be captured in a MOT that uses these same laser cooling methods.

To lower the temperature, it is desirable to apply sub-Doppler cooling methods, but it is not obvious how to do that with molecules. Our detailed theoretical study suggests that molecules captured in a red-detuned MOT can be further cooled in a blue-detuned molasses. It shows us the configuration that will optimize the cooling, and suggests that this will be just as effective as it is for atoms in cooling towards the recoil limit. Our MOT setup will have the flexibility to test these ideas.

While the density and temperature of the MOT will be suitable for loading molecules into tweezer traps and chip traps, the phase-space density is too low for loading into an optical lattice. We aim to increase the phase-space density by sympathetic cooling of the molecules using a gas of ultracold atoms. The experimental work needed to develop sympathetic cooling as a useful tool is already underway. We will develop the theoretical tools needed to guide the experiments and interpret the results.


The ground states of all heteronuclear bialkali molecules have 1Σ symmetry and so have only tiny magnetic moments due to nuclear spins. In contrast, 2Σ molecules possess both an electric and a magnetic dipole moment, allowing a greater range of control and scientific applications. Our laser-cooled source of CaF is one route to such molecules. An alternative, which may be more suitable for loading lattices, is to work with molecules associated from ultracold atoms which can themselves be loaded into 3D lattices at close to unity filling. Yb is advantageous because it possesses 7 stable isotopes (5 bosons and 2 fermions), many of which can be cooled to quantum degeneracy, which allow mass-tuning of the interspecies scattering length. Theoretical predictions show that CsYb is a favourable combination for magnetoassociation.

We have constructed a dual-species apparatus for Cs and Yb. With this apparatus, large numbers (>10 8 ) of both species can be reliably collected from a versatile Zeeman slower into adjacent MOTs. The capability of the system has been demonstrated by the production of Yb Bose-Einstein condensates (BEC). This project requires further development before a source of CsYb is available for science applications. We will first find the near-threshold bound states by 2-photon spectroscopy and then model the binding energies from these spectra to estimate ground-state scattering lengths for all the Yb isotopes. With this information, we will determine the best approach to producing CsYb molecules. Whichever association method works, the molecules are likely to be in a rovibrationally excited state that we will then transfer to the lowest level using STIRAP.

Yb BEC: Absorption images (top) and horizontal cuts (bottom) of the transition from thermal cloud (left) to pure BEC (right).


The production of bialkali dimers in the rovibrational ground state has now been achieved by five groups, working with KRb, RbCs, NaK and NaRb. However, amongst the possible bialkali combinations, there are only two nonreactive 5 fermionic molecules: NaK and KCs. These hold great promise for quantum simulation of many-body problems relevant to condensed matter physics. Motivated by this, we have begun a project to produce an ultracold gas of fermionic 40KCs molecules.

We will extend this work to explore the bosonic isotopes 39KCs and 41KCs as a resource for quantum science. Our predictions of atomic scattering properties show that the interspecies atomic scattering lengths for 41KCs are highly favourable for sympathetic cooling and loading of 3D lattices. Moreover, there are several interspecies resonances for magnetoassociation at magnetic fields below 200 G, including three resonances close to the 21 G window where evaporative cooling of Cs proceeds well. Accordingly, we will incorporate K sources into the molecular microscope project. A key decision will be to select the best bialkali molecule for the lattice experiments.


One of our key scientific challenges is to learn how to control the interactions between molecules and optical fields at the single photon level. Programme grant support allows us to adopt a synergistic approach and investigate a completely different type of source to develop quantum interfaces between molecules and photons. Here we will focus on single polycyclic aromatic hydrocarbon molecules in a host material, because they are easy to handle and have a  fluorescence quantum yield close to unity. These were among the first solid-state systems to exhibit single-photon emission and are arguably the most favourable. Their suitability for quantum applications has already been demonstrated in bulk optics through two-photon interference and the extinction/amplification of a laser beam. Among the few suitable molecule-host systems, dibenzoterrylene molecules (DBT) in anthracene offer near-ideal optical properties, with a branching ratio into the zero-phonon line of ∼40%, a ∼4 ns excited state lifetime permitting photon emission rates over ∼200 MHz, and a lifetime-limited linewidth at cryogenic temperatures of ∼2π×40 MHz.

(a) Energy levels of a typical organic molecule, in this case DBT, showing the singlet ground and excited state, including vibrational manifolds. (b) The structure of DBT and anthracene (top), and the lowest energy configuration of DBT in free space (bottom).

We have recently developed two techniques for growing DBT-doped anthracene crystals, a co-sublimation method which produces crystals with mm2 faces that are ∼1 μm thick and a super-saturated vapour growth for depositing crystals ∼100 nm in height directly on nanophotonic devices. Both techniques allow control over the level of DBT-doping depending on the intended application or experiment. The DBT is also aligned to the b-axis of the anthracene crystals in both cases. We will investigate various strategies to align the dipole of DBT molecules with respect to nanophotonic structures on surfaces including nano-patterning to match the anthracene crystal angles, functionalising the waveguide material before growth using self-assembled monolayers, 39 using different materials with intrinsic surface charges, such as lithium niobate, and applying an electric field during deposition to align the crystals as they grow.