We design filters, resonators, oscillators, delay lines and sensors, from specification through simulation, prototype and characterisation to manufacturing transfer. Our services also cover process design kits (PDKs) and compact models. The requirement, not a fixed product line, selects the wave family.
Surface, bulk and plate waves differ in where the energy sits and in what sets the frequency. Those two facts drive almost every trade-off that follows, which is why choosing the family is the first real decision in a design.
Energy is guided along the polished surface of a piezoelectric substrate. Frequency is set lithographically by the transducer pitch, so many frequencies can be patterned on one wafer. Modern thin-film and bonded substrates reach high Q with low temperature drift.
Energy resonates through the thickness of a thin piezoelectric film between two electrodes, so frequency is set by film thickness. Solidly mounted resonators conduct heat into the substrate and carry power well; membrane resonators trade that path for isolation.
Laterally excited modes in a thin suspended plate take their frequency from electrode pitch and plate thickness together. The coupling available in lithium niobate plates supports bandwidths the other two families reach only with difficulty.
| Class | Depends on | Notes |
|---|---|---|
| Filters and duplexers | coupling, Q, mode purity | Band-pass, band-reject, duplexers and multiplexers for RF front-ends. |
| Resonators and oscillators | Q, temperature drift | One- and two-port resonators; references for timing and frequency control. |
| Delay lines and correlators | propagation velocity | Signal processing carried out in the acoustic domain rather than electrically. |
| Physical sensors | frequency shift under load | Temperature, pressure, strain, torque and mass sensing. |
| Chemical and bio sensors | surface mass loading | Gas, humidity and biosensing on functionalised surfaces. |
| Front-end modules | integration and matching | Packaged assemblies specified to a system interface. |
| Acousto-electric devices | carrier–phonon coupling | Amplification and non-reciprocal behaviour from acoustic waves interacting with carriers in an adjacent semiconductor layer. |
| Magnetoacoustic devices | magnon–phonon coupling | Magnetostrictive layers driven acoustically; tunable and non-reciprocal components. |
| Acousto-optic devices | photoelastic interaction | Modulators, deflectors and tunable optical filters driven by an acoustic wave. |
| Phononic structures | engineered dispersion | Waveguides, mirrors and band-gap structures that route or confine acoustic energy on chip. |
The list is indicative rather than exhaustive. Our field is microwave acoustics as a whole: if a device works by moving an acoustic wave through a solid at radio or microwave frequencies, it is worth asking us about.
Every stage ends with something you can check, so you always know where a design has got to. Here is what happens in each one, and what you have in hand at the end.
Electrical and environmental specification: frequency, bandwidth, insertion and return loss, power, temperature range, shock and vibration, and the package the part has to live in.
Exit An agreed specification and the acoustic approach that can meet it.
Finite-element analysis of the resonator structure, coupled to circuit, electromagnetic and electrothermal simulation so that the whole device is predicted rather than the acoustic cell alone. PDK and compact-model development connects device physics and process parameters to reusable circuit-level models.
Exit A predicted response, a mode inventory and an identified spurious-mode risk.
Wafer-scale builds through established process partners across surface, bulk and plate processes, with the mask set laid out for measurement as well as for function.
Exit Measurable parts and a test structure set on the same wafer.
On-wafer and packaged measurement, thermal cycling, power and environmental stress, then correlation of the measured response against the model so the next iteration is informed rather than guessed.
Exit Measured performance against specification, with the model corrected to match.
Design-for-manufacture and design-for-test review, process window and yield analysis, and the documentation a volume line needs to build the part without the design team present.
Exit A released design package and a qualified process at the manufacturing partner.
Four interactive figures: the filter response, the wave itself, where the bands sit, and how a front-end module fits together. These illustrations show qualitative behaviour; their values are representative, not solver output or measured data. They are not a substitute for full-wave analysis.
Confining acoustic energy in the intended mode is most of the work. Finding where the rest of it went is the other half.
Confine it cleanly and you get the high Q that produces steep skirts and low insertion loss. Let it escape through the anchors, the edges or the substrate and the part becomes lossy and hard to match.
Energy that escapes tends to come back as trouble: transverse and lateral spurious modes show up as ripple inside the passband. We map both from first principles in our own solvers, confirm them on real material by laser Doppler vibrometry, and engineer them out with apodised geometries, piston-mode borders and tuned boundary conditions.
Simulation, radio-frequency measurement and materials analysis sit side by side. When model and measurement disagree, we settle it against the same hardware that produced the result instead of passing the question on.
| Capability | Applied to |
|---|---|
| Compute cluster | Large finite-element models of full resonator structures. |
| In-house acoustic solvers | Mode inventories and spurious-mode prediction, alongside commercial multiphysics and electromagnetic tools. |
| Coupled co-simulation | Acoustic, piezoelectric, electromagnetic and electrothermal behaviour solved together rather than in sequence. |
| PDK and model development | Process design kits (PDKs), compact device models and parameter extraction for circuit-level simulation, with documented process assumptions and model limits. |
| Design of experiments | Automated parameter sweeps and optimisation across the process window. |
| Capability | Applied to |
|---|---|
| Vector network analysis | Response measurement into the millimetre-wave range. |
| On-wafer probing | Device and test-structure measurement before dicing. |
| Power and environmental test | Power handling, thermal cycling and environmental stress. |
| Calibration standards | Traceable reference for every measurement above. |
| Method | What it resolves |
|---|---|
| SEM, EDX, XPS, AFM | Surface condition, composition and film topography. |
| X-ray diffraction | Crystallographic texture of the piezoelectric film, which sets coupling. |
| Laser Doppler vibrometry | Direct visualisation of the acoustic mode on a real device. |
| White-light interferometry | Surface profile, film thickness and flatness. |
| Nanoindentation | Elastic constants and stress state of deposited films. |
| TCF characterisation | Frequency drift against temperature across the operating range. |
ResonixRF was founded in 2025 by microacoustic specialists who previously ran research and development at established device makers.
We keep the field deliberately narrow: acoustic device physics, the modelling that predicts it, and the process knowledge that turns a prediction into a part that yields. That is where we are genuinely useful, so that is the work we take on.
Send a briefCombined across the founding team. We manufacture with established foundry and fab partners, and draw on academic advisors in piezoelectric materials and microwave acoustics.
The more specific you can be, the more useful our reply will be — target band, bandwidth, loss budget, power, environment and volume are usually what decide whether an approach is viable. For PDK or model development, include the process, available device data and intended simulation workflow. If you are still scoping, send what you have and we will help you narrow it down.
