SAW-Bench: Programmable SAW Field Synthesis

• SAW-Bench is a multifunctional platform that synthesizes tailored surface acoustic wave-fields by combining an array of IDTs with inverse filtering algorithms.
• It achieves high spatial and temporal control of acoustic fields for advanced microfluidic operations such as droplet manipulation, division, fusion, and atomization.
• Its modular design, rigorous calibration workflow, and versatile field configurations enable rapid prototyping and extension to 3D control and closed-loop sensing.

Surface Acoustic Wave (SAW)-Bench refers to a multifunctional platform for the programmable synthesis of tailored surface acoustic wave-fields by combining an array of interdigitated transducers (IDTs) and inverse filtering algorithms. This system enables the generation of classical and complex SAW fields—plane progressive, focused, standing, and swirling (vortex)—within a unified hardware framework, primarily for microfluidic and biological applications. The SAW-Bench achieves high spatial and temporal control of acoustic fields, facilitating versatile manipulation of sessile droplets, fluids in microchannels, and suspended particles, and expanding the repertoire of microfluidic actuation and particle tweezing operations previously restricted by fixed IDT geometries (Riaud et al., 2016).

1. Hardware Architecture

The SAW-Bench is built on a 1.05 mm thick X-cut lithium niobate (LiNbO₃) substrate featuring weak anisotropy compared to standard 128° Y-cut, with electromechanical coupling coefficients $K^2_Z = 5.9\%$ and $K^2_Y = 3.1\%$. The IDT array (IDTA) comprises 32 unidirectional single-phase unidirectional transducer (SPUDT) elements. Each IDT is fabricated with a Ti adhesion layer (20 nm) and Au contacts (200 nm) via lift-off, and is geometrically curved and distributed along the substrate slowness curve to maximally illuminate a 5 mm-radius acoustical scene.

The designed operating frequency is $f_0 = 11.9$ MHz, and due to substrate anisotropy, the local SAW wavelength $\lambda(\varphi)$ varies as $c_R(\varphi)/f_0$. Each IDT is independently driven via 32 electronically programmable channels, providing arbitrary waveform generation up to 12 MHz and precise amplitude-phase control per transducer. Impedance matching is achieved for each IDT with tuned external inductances.

2. Inverse-Filter Algorithm

The inverse filter algorithm is central for synthesizing arbitrary target SAW fields. Let $E_j(\omega)$ be the frequency-domain drive voltage for IDT $j$ and $S_i(\omega)$ the complex surface displacement at control point $i$. The propagation matrix $H_{ij}(\omega)$ relates input and output via $K^2_Y = 3.1\%$0 (Einstein summation).

Measurement stage: The impulse response $K^2_Y = 3.1\%$1 from channel $K^2_Y = 3.1\%$2 to control point $K^2_Y = 3.1\%$3 is acquired with a Michelson interferometer at $K^2_Y = 3.1\%$4-spaced grid points. $K^2_Y = 3.1\%$5 is computed as the Fourier transform of $K^2_Y = 3.1\%$6.

Field synthesis stage: The target field $K^2_Y = 3.1\%$7 is prescribed. The optimization problem minimizes

$K^2_Y = 3.1\%$8

where $K^2_Y = 3.1\%$9 is a Tikhonov regularization parameter. The closed-form solution, typically computed via SVD, is

$f_0 = 11.9$0

The time-domain drive signals $f_0 = 11.9$1 are then uploaded to the waveform generator. The reconstructed field $f_0 = 11.9$2 is formed by a sum of convolutions: $f_0 = 11.9$3.

3. SAW Field Engineering

Wave-field synthesis is performed within the acoustical scene (radius ~5 mm, diameter ~10 mm) using polar coordinates $f_0 = 11.9$4 and phase velocity $f_0 = 11.9$5. The key synthesized field types include:

• Plane progressive SAWs: $f_0 = 11.9$6, where $f_0 = 11.9$7.
• Focused anisotropic beams: $f_0 = 11.9$8 with the apodization $f_0 = 11.9$9.
• Standing waves: Constructed as the superposition of two plane waves in opposite directions, $\lambda(\varphi)$0.
• Swirling (vortex) SAWs of topological charge $\lambda(\varphi)$1: $\lambda(\varphi)$2.

Spatial sampling at $\lambda(\varphi)$3 (typically $\lambda(\varphi)$4 μm grid for $\lambda(\varphi)$5 μm at 11.9 MHz) ensures full field reconstruction up to $\lambda(\varphi)$6.

4. Performance and Demonstrated Capabilities

The SAW-Bench attains peak-to-peak surface displacements of 6–10 nm for plane waves, up to 100 nm for focused beams, and 80–108 nm for swirling SAWs (depending on topological charge). Spatial focusing is typically within ±100 μm, with sidelobe levels generally under 20% of the main lobe, limited by the 32-element aperture.

Demonstrated microfluidic operations on 2 μL sessile water droplets include:

• Displacement: Beam-forming sequences displaced droplets at ~50 ms per motion step through cumulative radiation pressure and acoustic streaming.
• Division: Alternating focus points induced droplet division within ~27 ms.
• Fusion: Swirling SAW ($\lambda(\varphi)$7) merged two droplets in ~27 ms.
• Atomization: High-intensity annular SAW ($\lambda(\varphi)$8) nebulized droplets in ~13 ms.

Radiation-pressure–induced contact-line forces are on the order of μN; acoustic streaming velocities are on the mm/s scale.

5. Calibration, Workflow, and Extensions

For a new target field, the recommended procedure is:

1. Define a $\lambda(\varphi)$9 grid of control points.
2. Measure $c_R(\varphi)/f_0$0 for all $c_R(\varphi)/f_0$1 combinations.
3. Compute $c_R(\varphi)/f_0$2 over the frequency range.
4. Specify $c_R(\varphi)/f_0$3, embedding anisotropy.
5. Solve for $c_R(\varphi)/f_0$4, selecting $c_R(\varphi)/f_0$5 to balance field fidelity and regularization.
6. Perform inverse Fourier transform to obtain $c_R(\varphi)/f_0$6.
7. Upload $c_R(\varphi)/f_0$7, drive the IDT array, verify the field by interferometry, and iterate parameters as needed.

Among limitations, finite transducer count restricts maximum steering angles and sidelobe suppression; higher frequencies necessitate faster AWGs and denser spatial sampling. The methodology presumes linear SAW propagation; high-power nonlinear effects (e.g., droplet absorption) may impair fidelity. For 3D spatial targeting, extension to volumetric control points and incorporation of coupled-mode or multiphysics models is suggested. The platform can also be adapted for real-time closed-loop control by operating IDTs in a dual detect-actuate mode for echo imaging and dynamic reconfiguration.

6. Relevance and Future Directions

The SAW-Bench paradigm enables rapid prototyping and execution of diverse acoustic operations on a single chip, contrasting sharply with classical single-purpose IDT arrangements. Its synthesis architecture not only consolidates the implementation of existing microfluidic tasks but also offers a controlled testbed for the exploration of advanced SAW field topologies, including tailored vortex fields for complex particle manipulations. Prospective developments include scaling the IDT array for higher directivity, expanding operational bandwidth, integration with closed-loop sensing, and extension to 3D field control for in-channel and volumetric actuation in complex microfluidic devices (Riaud et al., 2016).