A Stanford University team has reported the first observation of quantum jumps in phonons, after boosting phonon coherence times by four to five orders of magnitude on thin-film lithium niobate devices. The result, published in Science, places mechanical vibrations alongside the ions measured in the 1980s and the photons demonstrated in the 2000s as systems where discrete quantum transitions can be caught in real time.

"Until now we had quantum jumps in ions and in photons; phonons were the missing piece," team leader Amir Safavi-Naeini tells Physics World. "We could see it by eye in the data, which was very satisfying."

What changed in the fabrication?

The group's recipe starts with thin-film lithium niobate, a material commercialized about a decade ago and now a workhorse for acoustic and photonic integration. The crystal is piezoelectric: stress from a mechanical vibration produces a voltage. The Stanford team uses that voltage to couple vibrational quanta to a superconducting qubit, which reads out a binary "phonon present or absent" signal.

Early devices were too lossy. Phonons lived for only 800 nanoseconds, well short of the millisecond window needed to watch a transition from the excited state to the ground state. "Extending these phonon coherence times took a lot of work by a number of graduate students in my group," Safavi-Naeini says.

The team traced dissipation to surface defects, including residue left by polymer processing. They replaced polymer lithography with shadow masks that align one chip above another and evaporate metal through the pattern. The switch yielded the four to five orders of magnitude gain in device performance.

Students Takuma Makihara and Erik Szakiel then integrated the long-lived mechanical oscillators with the superconducting qubit system without degrading either. "That was the last piece," Safavi-Naeini recalls. "Once we had that, we could do these experiments."

How does the qubit read the phonon state?

The qubit discriminates between an even and odd number of phonons in the resonator. At very low energies, an odd count effectively means a single phonon exists. That makes the read-out a parity measurement, a useful primitive in quantum error correction where the goal is to flag errors without collapsing the underlying quantum state.

The team points to the so-called binomial code, already implemented for microwave photons, as a natural fit. Phonon pairs encode the logical qubit; if the system detects an odd number, an error has occurred and the lost excitation becomes apparent without measuring every state. Safavi-Naeini notes that binomial coding works best for long-lived particles, a built-in advantage of the Stanford device.

Which applications are within reach?

The group has opened two collaboration tracks. With Michael Roukes at the California Institute of Technology, they are exploring single-protein mass measurements with a mechanical sensor. They are also weighing whether mechanical resonators could be entangled through gravitational fields, a test so far confined to heavier objects.

Both targets demand further lifetime gains. Safavi-Naeini estimates a factor of 10 more coherence for error correction and mass spectrometry, and up to six orders of magnitude for the gravity experiment, with new modifications to the experimental set-up also required.

"The hope is that by figuring out the fundamental sources of loss in these materials we can build our way up to those experiments in the coming decade," Safavi-Naeini says. "Every time you improve, you hit some new source of noise, you understand it, then you can deal with it, and then you go to the next one. It's a fun physics problem."