A Bose–Einstein condensate of dipolar molecules

In a gas chilled to the nanokelvin regime, particles with integer spin can all settle into the same ground state. The resulting Bose–Einstein condensate, or BEC, has proved to be a rich laboratory for quantum phenomena such as superconductivity and superfluidity (see the article by Carlos Sá de Melo, Physics Today, October 2008, page 45 ).

Although BECs of polar atoms and of Rydberg atoms have given physicists glimpses of magnetism in ultracold systems, the former have weak polar interactions, and the latter have short lifetimes. Researchers have sought to create a BEC composed of dipolar molecules (see the article by Deborah Jin and Jun Ye, Physics Today, May 2011, page 27 ). Such a BEC would be an ideal quantum simulator. Whereas atoms in an optical lattice can be used to simulate the Hubbard model, for example, molecules with much stronger dipole interactions might exhibit novel physics. “There are many new properties of matter relevant to quantum magnetism and all sorts of other things that open up if you can make lattices of dipolar particles,” says Jeremy Hutson, a theorist at Durham University in the UK.

But in their quest to cool dipolar molecules, researchers ran into a problem: They lost molecules too quickly. A collision between molecules results in a chemical reaction that imparts enough kinetic energy to kick the molecules out of the optical trap. The losses prevented researchers from trapping and cooling enough dipolar molecules to form a BEC. “What’s really changed in the last couple years is people have found ways to basically shut the losses off,” says Kaden Hazzard, a theorist at Rice University.

Now a team led by Sebastian Will at Columbia University has created the first BEC from dipolar molecules. The BEC, reported in Nature on 3 June, is composed of sodium cesium molecules—bosonic particles with a strong dipole—cooled to 6 nK. If its long-range dipole interactions could be preserved, the system could be a window to view exotic phases of quantum matter, such as supersolids.

The first ultracold polar gas was constructed in 2008, when Deborah Jin and Jun Ye of JILA in Boulder, Colorado, and their colleagues managed to cool potassium rubidium to 350 nK. They used optical tweezers to perform evaporative cooling, which allows energetic particles to float away and leaves behind only those with the lowest energy.

But Jin and Ye were losing too many KRb molecules to create a BEC. It seemed the molecules were reacting to create K₂ and Rb₂ and getting ejected from the tweezers. “People thought, ‘Well, maybe we only have to switch to nonreactive molecules like sodium potassium, and it’s going to be good,’ ” says Immanuel Bloch, director of the quantum many-body systems department at the Max Planck Institute of Quantum Optics in Germany. “But that turned out to be wrong.” No matter what molecules the research groups tried, the loss rate remained about the same.

Scientists tried various approaches, including flattening the gas into a pancake, but persistent losses prevented the creation of a dipolar molecular BEC. In 2018 Tijs Karman at Radboud University in the Netherlands and Hutson proposed a new approach to prevent loss: Microwave fields could shield most of the two-body interactions and keep dipolar molecules in line. Initial experiments proved promising, and in 2022 Bloch and his colleagues cooled NaK, a fermionic molecule, to 21 nK—about a third of the Fermi temperature, the point at which molecules occupy the lowest energy levels.

To keep losses low in NaCs, the Columbia researchers needed an exceptionally uniform microwave field, so they built a custom antenna. The circularly polarized field it generated increased the short-range repulsion between molecules and lowered the rate of two-body losses by a factor of about 200 . But with two-body losses subdued, another problem reared its head: three-body losses. Although the microwave field repelled NaCs molecules in the short range, it introduced a long-range attraction between molecules, which led to three-body collisions.

Will and his colleagues realized that they could control three-body losses by adding another antenna with a vertically polarized field. With their new alignment, the molecules lost their long-range attraction. The Columbia team ended up with a dipolar molecular BEC with about 200 molecules and a lifetime of about 1.8 seconds; the researchers say they have since boosted the number of molecules to roughly 1000. In comparison, the first BEC from the mid 1990s contained about 2000 rubidium-87 atoms and lasted roughly 15 seconds.

Adjusting the strength of the microwave fields also “offers a knob to control the strengths of the dipole–dipole interactions,” Will says. Initially, the microwave fields were set to reduce the dipole interactions and maximally suppress losses. But Will and his colleagues have since made preliminary measurements with the dipolar interactions turned back on—to about 5% of their normal strength. Although allowing some interactions shortens the lifetime of the BEC, Will says the Columbia team has already seen some interesting effects. “Spontaneously, the BEC splits up into droplets,” he says. “These are new types of self-organization.” The researchers are watching for even odder phases of matter, like a supersolid, in which particles can flow with zero viscosity yet still end up in rigid structures.

Looking ahead, researchers are interested in expanding the types of molecules that can be cooled. Currently, dipolar molecules like NaCs must be assembled by laser cooling alkali or alkaline elements to the millikelvin regime and then combining them. Only afterward do the molecules undergo evaporative cooling to get to the nanokelvin regime. Developing ways to directly laser cool molecules instead of elements would allow experimenters to incorporate more of the periodic table.

For now, though, there is plenty more to learn from NaCs, Will says: “I think this experiment is going to keep us busy.”