HKUST Develops Ultra-Robust Acoustic Chip Technology To Advance 6G And Satellite Communications

Researchers from The Hong Kong University of Science and Technology (HKUST) have developed a novel chip architecture that enables the miniature acoustic devices inside smartphones to withstand over 12 times higher power loads while maintaining cooler and more stable operation. Tested on transducers vibrating more than two billion times per second, the new platform – named "Layered Acoustic Wave" (LAW) – reduced the operating temperature rise by 70% and set a record threshold power density of 36.4 W/mm². The breakthrough lifts a decades-old restriction that confined acoustic chips to small-signal processing, paving the way for a new generation of applications from direct-to-cell satellite communications and 6G networks to compact power conversion.

The research was carried out by a research team led by Prof. YANG Yansong, Assistant Professor of the Department of Electronic and Computer Engineering (ECE), with his PhD student QIAN Fangsheng as first author. The research result was published in the international journal Nature Communications titled "Suppressing Acoustomigration and Temperature Rise for High-Power Robust Acoustics". All authors of the paper are members of the Radio Frequency Microsystem (RFMY) lab in the Department of Electronic and Computer Engineering at HKUST.

Acoustic wave devices are among the most ubiquitous yet least-known chips in modern life. They convert electromagnetic signals into sound waves vibrating at gigahertz (GHz) frequencies, filtering the radio-frequency signals for every smartphone and mobile base station. Because sound travels about 100,000 times slower than light, acoustic devices can compress GHz-frequency energy down to chip scale, making them irreplaceably compact. Today, acoustic devices are also expanding into interdisciplinary frontiers – coupling with qubits to process quantum information, driving microfluidic and acousto-optic systems, and serving as non-magnetic energy-storage units for sustainable power conversion.

The strategic value of high-power acoustics is increasingly evident: in 2025, SpaceX acquired acoustic filter company Akoustis Technologies to develop high-performance acoustic filters for direct-to-cell satellite networks – applications in which power-handling capability directly determines the success or failure of the link between the satellite and the user's phone.

Yet a fundamental bottleneck has remained unsolved for decades: acoustic devices struggle to operate reliably at high power. Just as violent shaking can rattle objects apart and generate heat, high-power GHz vibration concentrates enormous energy density in a tiny volume, triggering three intertwined failure mechanisms: mechanical stress and self-heating drive "acoustomigration", in which metal atoms in the electrodes migrate to form protrusions and voids, causing circuit failure; accumulated heat changes the acoustic velocity, causing the operating frequency to drift off target; and concentrated stress cracks or delaminates the piezoelectric thin film.

Existing countermeasures focus almost entirely on the bottom of the device – switching to expensive high-thermal-conductivity substrates such as silicon carbide or diamond. But the piezoelectric layer itself conducts heat poorly, so heat and stress remain trapped near the top surface where the device actually vibrates. Meanwhile, a long-standing consensus in the field has held that the top surface of an acoustic wave device must stay in contact with air in order to confine the acoustic energy. This view has seemingly ruled out the possibility of engineering the very boundary where failure originates.

The research team led by Prof. Yang has overturned that consensus. Their LAW architecture "buries" the vibrating surface of a lithium niobate thin-film device beneath a composite top boundary: first a silicon dioxide isolation layer, then a "quasi-infinite" amorphous silicon overlayer far thicker than the acoustic wavelength. Through theoretical analysis and simulations, the team found that the acoustic wave, contrary to expectation, remains firmly confined near the surface – while the added stack takes on three roles at once.

The thick amorphous silicon overlayer serves first as a mechanical confinement layer, reshaping the stress-field distribution to cut the peak stress at the electrode interface to about one quarter – directly eliminating the driving force behind acoustomigration. Next, it acts as an integrated heat spreader: with a thermal conductivity roughly two orders of magnitude higher than that of air, it opens vertical channels that rapidly carry heat away from the hot spots. At the same time, it works as a thermal-expansion compensation layer, keeping the frequency stable under large temperature swings. Far from sacrificing performance, the redesigned boundary raises the operating frequency at the same wavelength by about 25% and lowers damping loss by about 30%, with the capability to scale across the sub-6 GHz bands – all achieved with standard, low-cost fabrication processes.

In standardized component-level high-power tests, LAW's advantages were striking. Under identical power loads, the steady-state temperature rise of the LAW transducer was only 5.2°C, versus 17.4°C for a state-of-the-art thin-film surface acoustic wave (TF-SAW) control device – equivalent to a 70% reduction. The LAW device withstood an injected power density of 45.61 dBm/mm² (36.4 W/mm²), 12.73 times the threshold of the TF-SAW control, while maintaining a first-order temperature coefficient of frequency of −13 ppm/°C across an ultra-wide 475°C temperature range (−150°C to 325°C). At −85°C, the threshold climbed further to 49.45 dBm/mm² (88.11 W/mm²), representing a 13.85-fold increase over the TF-SAW control threshold, of particular significance for cryogenic quantum acoustic systems. Failure analysis corroborated the underlying mechanism: TF-SAW devices exhibited severe electrode migration and cracking of the piezoelectric film, whereas LAW devices operating at far higher power showed no such severe migration.

Prof. Yang said, "For decades, the community accepted two 'givens': that the surface of an acoustic wave device must stay open to the air, and that high-power operation was far out of reach. We have proven that both assumptions can be broken at the same time. Rather than chasing increasingly expensive substrate materials at the bottom of the device, which can only improve thermal dissipation slightly, we chose to redefine the boundary above it – turning the most fragile region (worst heat dissipation) of the device into its strongest asset. This takes acoustic technology from small-signal components to a true high-power platform, meeting the demands of emerging applications such as direct-to-cell satellite connectivity, 6G networks and energy conversion."

The paper's first author Qian Fangsheng said, "The beauty of this architecture is that a single simple, low-cost layer of material accomplishes three tasks at once: it carries heat away from the hot spots, compensates for frequency drift and – most importantly – redistributes the mechanical stress, bringing the peak stress that drives metal migration down to about one quarter. We are not merely delaying failure by cooling the device; we are eliminating failure itself at its physical root."

Because the approach redefines boundary conditions rather than relying on any particular material, its design principles are broadly applicable to the wide family of acoustic systems based on interdigital transducers. The team envisions the LAW platform underpinning the development of next-generation radio-frequency filters, cryogenic quantum acoustic circuits, on-chip acousto-optic and microfluidic systems, and miniaturized non-magnetic power conversion modules.