To create a Bose-Einstein condensate, all of the constituent particles must be cooled to just above absolute zero so that their de Broglie wavelength is comparable to their separation distance. The first condensate of atoms was created in 1995, but it has been difficult to make the molecular analogue because many of the techniques used to cool atoms – such as laser cooling – do not necessarily work for molecules.

Rather than cooling molecules until they condensed, Donley and co-workers converted part of an atomic condensate into molecules. To do this they used a magnetic pulse to tune the energy of pairs of colliding atoms within the condensate. However, since only a fraction of the atoms could have been converted into molecules, the condensate existed in a quantum superposition of atoms and molecules.

To establish that the condensate was indeed in a superposition, the researchers applied a second magnetic pulse to break the superposition and then immediately measured the number of atoms in the condensate. They discovered that this number – a measure of the relative phase between the atomic and molecular states – varied sinusoidally as the time between the pulses was increased. The frequency of this variation matched that expected from quantum mechanical calculations describing the atomic collisions within the condensate. The JILA researchers say that this agreement between theory and experiment provides strong evidence for a coherent superposition of atomic and molecular states.

The work could lead to developments in cold molecular spectroscopy and to a better understanding of molecular collisions. It could also be used to explore the physics of zero-temperature chemical reactions, and to create entangled molecules for quantum computing.

However, other physicists are not convinced that the JILA group has observed a molecular Bose-Einstein condensate. Daniel Heinzen of the University of Texas says that the research is “tantalizingly close” to demonstrating a molecular condensate, but says that what is actually generated in the experiment is a gas of pairs of atoms, each of which is in a coherent superposition of a bound state molecule, a pair of atoms in the condensate, and pair of atoms that are not in the condensate. “I think this is somewhat different to a mixture or superposition of a molecular condensate and an atomic condensate,” he says. He adds that the short lifetime of the gas may in fact make it impossible to observe molecular coherence within the gas.

Carl Wieman, one of the JILA researchers and a recipient of the 2001 physics Nobel prize for his creation of atomic Bose-Einstein condensation, believes that further proof is not really necessary. "If it walks like a duck and quacks like a duck then it's a duck," he says of his group's work.