Millisecond phonon lifetimes in piezoelectric nanomechanical resonators

The pursuit of mechanical resonators capable of storing quantum information for extended periods has long been shaped by a tension between two material properties: the intrinsic quality of the mechanical mode and the strength with which that mode can be driven and read out electrically. Crystalline quartz and sapphire have historically set the benchmark for mechanical coherence, with quality factors reaching into the hundreds of millions or beyond at cryogenic temperatures, but these materials couple only weakly to electric fields, complicating their integration with superconducting circuits. Piezoelectric materials such as lithium niobate invert this trade-off, offering strong electromechanical coupling that enables efficient transduction between microwave photons and phonons, yet their measured quality factors have stubbornly remained orders of magnitude below those of the best non-piezoelectric resonators. The new results from the Stanford group demonstrate that this gap was not a property of lithium niobate itself but of the way lithium niobate devices were conventionally made.

The central insight of the work is that the dominant loss channel in gigahertz-frequency lithium niobate resonators has been fabrication-induced rather than fundamental. When such devices are patterned, the standard approach involves ion milling, a physical etching process in which energetic ions bombard the surface to remove material. This process leaves behind a damaged layer in which the crystalline lattice is disrupted, and it is precisely this damaged layer that hosts the defects responsible for phonon loss. Recent studies of defect-mediated loss in lithium niobate had begun to clarify this picture, identifying two-level-system defects at surfaces and interfaces as significant contributors to dissipation at cryogenic temperatures. Building on that understanding, the researchers developed a process that chemically restores the ion-milled surfaces, effectively healing the damage that the etching step had introduced. The chemical treatment dissolves or reorganizes the defective near-surface region, returning the material closer to its pristine crystalline state.

A second challenge addressed by the work concerns metallization. Piezoelectric resonators require metal electrodes to convert electrical signals into mechanical motion and vice versa, but depositing metal directly onto a resonator surface introduces additional loss channels. Metal films can host their own two-level-system defects, and the metal-semiconductor or metal-insulator interface can further degrade coherence. The team circumvented this problem with a shadow-mask metallization process, in which a physical mask is held in close proximity to the device during metal evaporation, defining the electrode pattern without requiring a resist layer or direct-contact lithography on the resonator itself. The electrodes are then separated from the resonator body by a small gap, allowing the electric field to couple to the piezoelectric mode while keeping lossy metal from touching the regions of highest acoustic strain. The mechanical drawings for the flip-chip shadow-mask evaporation apparatus used in this process have been released publicly, which should allow other groups to reproduce the technique without needing to redesign the tooling from scratch.

The headline result is an inferred internal phonon lifetime of up to 21 milliseconds at a frequency of 3.6 gigahertz, corresponding to a quality factor exceeding four hundred million. Expressed in the conventional figure of merit for mechanical resonators, the product of frequency and internal quality factor exceeds ten to the eighteenth power, placing these devices in the same regime as the highest-performing mechanical resonators made from any material. Achieved at millikelvin temperatures, this lifetime means that a single phonon at gigahertz frequency, once created, persists long enough — roughly twenty milliseconds — for a sequence of quantum operations to be performed on it, or for it to be exchanged with a microwave photon in a superconducting circuit and then recovered. For context, the inverse of the lifetime sets the linewidth of the mechanical resonance, and a linewidth of tens of hertz at gigahertz frequencies represents a spectral purity that was previously thought unattainable in piezoelectric materials.

Equally important is the way the authors established that these lifetimes reflect intrinsic material quality rather than measurement artifacts. By systematically varying the spacing between the electrode and the resonator, they changed the effective electromechanical coupling between the device and the external measurement circuit. When a resonator is strongly coupled to its measurement apparatus, energy can leak out through that coupling channel, shortening the apparent lifetime even if the material itself is lossless. The measured decay rate in such experiments is a combination of internal loss and external loading, and disentangling the two requires careful variation of the coupling strength. The systematic study showed that the measured lifetime was set by this external coupling rather than by material dissipation, meaning that the internal quality factor — the figure that matters for how long a phonon truly survives in the device — is even higher than the raw measured values suggest. This distinction between loaded and internal quality factor is a recurring subtlety in resonator physics, and the demonstration that the internal quality factor exceeds the reported values strengthens the claim that lithium niobate itself, once surface damage is removed, is comparable in intrinsic quality to the best mechanical materials known.

The implications for quantum information processing are substantial. Mechanical resonators have emerged as compelling candidates for quantum memories in superconducting quantum computing architectures, where microwave photons decohere on microsecond timescales but phonons in high-quality mechanical modes can persist for far longer. Previous experiments with lithium niobate phononic crystal resonators demonstrated quantum state transfer between microwave photons and phonons, phonon-mediated entanglement between distant qubits, and the resolution of individual phonon energy levels, but the utility of these protocols was constrained by phonon lifetimes in the microsecond range. Extending those lifetimes by three to four orders of magnitude transforms the calculus: a mechanical memory that holds a quantum state for twenty milliseconds can outlast many microwave-coherence budgets and can serve as a genuine storage element rather than a transient buffer. The strong piezoelectric coupling of lithium niobate, which was always its advantage, now comes without the coherence penalty that previously accompanied it.

The work also connects to a broader body of research on two-level-system defects in quantum materials. Defects with quantum character — atomic-scale tunneling systems that can absorb and re-emit energy — are now understood to plague not only mechanical resonators but also superconducting qubits and microwave cavities. Studies of spectral diffusion in nanomechanical resonators have shown that single quantum defects can modulate resonator frequency and dephase coherent motion, and correlated dephasing phenomena in piezoelectrically transduced waveguides have provided further evidence that defect dynamics are central to coherence in these systems. The chemical restoration process developed here represents a materials-science approach to mitigating such defects, complementing the device-engineering approaches of phononic bandgap design and mode isolation that have dominated the field. As understanding of defect physics matures, the combination of surface chemistry, careful metallization, and phononic engineering may become a standard recipe for high-coherence acoustic devices.

It is worth noting the broader landscape of piezoelectric quantum acoustodynamics. Surface acoustic wave resonators, bulk acoustic resonators, and phononic crystal modes each occupy different niches, with varying trade-offs between mode volume, frequency, coupling strength, and coherence. Quartz phononic crystal resonators have recently been explored as hybrid acoustic memories, and thin-film quartz has been proposed as a platform for high-coherence piezoelectric phononic devices, motivated by quartz’s favorable defect properties. The new lithium niobate results suggest that the choice between piezoelectric materials may be less constrained by intrinsic loss than previously assumed, allowing device designers to prioritize coupling strength, fabrication compatibility, or nonlinear properties without sacrificing coherence. The availability of the underlying datasets through a public repository further supports the reproducibility of these findings and should accelerate adoption across the community.

Looking forward, several directions follow naturally from this result. Longer phonon lifetimes enable stronger cooperative coupling in hybrid quantum systems, where the coherent exchange rate between a phonon and a qubit must exceed the decoherence rates of both for useful quantum state transfer. They also open the door to experiments that store and manipulate single phonons for extended periods, potentially including phonon blockade, squeezed mechanical states with long lifetimes, and entanglement distribution across acoustic channels. In precision sensing, long-lived gigahertz phonons could serve as frequency references or as probes of faint forces, building on earlier demonstrations of nanomechanical mass spectrometry with single-molecule sensitivity. The demonstration that fabrication-induced damage, rather than material physics, set the previous limits is a reminder that in nanoscale quantum devices, the path to fundamental performance limits often runs through process control as much as through material selection.

Subject of Research: Millisecond phonon lifetimes in piezoelectric nanomechanical resonators

Article Title: Millisecond phonon lifetimes in piezoelectric nanomechanical resonators

Article References: Hitchcock, O. A., Maksymowych, M. P., Pezeshki, K., Makihara, T., Szakiel, E., Kiang, C.-W., Rugar, D., & Safavi-Naeini, A. H. (2026). Millisecond phonon lifetimes in piezoelectric nanomechanical resonators. Nature Nanotechnology. https://doi.org/10.1038/s41565-026-02289-9

Image Credits: AI Generated

DOI: 10.1038/s41565-026-02289-9

Keywords: nanotechnology, Millisecond, phonon, lifetimes, piezoelectric, nanomechanical, resonators, scientific research