Programmable photoacoustic patterning of microparticles in air

Precise non-contact manipulation of micro/nano objects is vital for applications in biology, chemistry, and micro/nanotechnology. Existing methods—optical, magnetic, opto-electric, and acoustic tweezers—offer complementary strengths: optical tweezers enable nanoscale precision but are limited by particle optical properties and small trapping sizes, while acoustic tweezers provide larger forces and handle larger objects but struggle with high-throughput arbitrary-shape patterning. Holographic acoustic manipulation with phased arrays requires many transducers and complex control, and spatial ultrasound modulation using microbubble arrays is limited by bubble size. Hybrid optical–acoustic approaches improve capability but often have fixed trapping positions that limit flexibility. Photoacoustic tweezers leverage optically generated acoustic radiation, yet in liquids random bubble formation hampers precision; optically excited Lamb waves have been used for single-object rotation in nonliquid settings but not for versatile, large-scale patterning. This work addresses the open challenge of flexibly and simultaneously manipulating large numbers of particles with high throughput and arbitrary shapes by introducing a programmable photoacoustic patterning (PPAP) method that maps laser-defined patterns into localized Lamb-wave fields to rearrange microparticles in air.

The paper surveys prior manipulation techniques: optical tweezers provide high precision and have been widely applied but are constrained by optical properties and trap size; acoustic tweezers deliver stronger forces and larger-object handling but face challenges in generating arbitrary patterns at high throughput. Holographic acoustic methods using phased arrays can create complex fields yet demand many transducers and intricate control, limiting practicality. Spatial ultrasound modulation that controls microbubble arrays offers a promising route but its spatial precision is bounded by bubble dimensions. Hybrid optical–acoustic systems alleviate some limitations but typically use fixed traps. Early photoacoustic tweezers demonstrated optically induced acoustic forces, though liquid environments introduce random bubble formation that degrades precision. Nanoscale Lamb-wave-driven actuation has been shown for single-object rotation on fibers in nonliquid environments, indicating the potential of Lamb waves for solid-surface manipulation. These works motivate a method that combines the programmability of optics with the force advantages of acoustics without the hardware complexity and precision limits observed previously.

The authors develop a programmable photoacoustic patterning (PPAP) platform that converts laser-defined patterns into localized antisymmetric Lamb waves on a multilayer membrane, which then rearrange microparticles in air.

• System and operation: Target patterns are uploaded to a digital micromirror device (DMD), which modulates a 6 ns, 532 nm pulsed laser (single-pulse energy ~480 mJ, 10 Hz). The patterned beam passes a convex lens (f = 40 cm) and a 1 mm water layer and is projected onto a TiN–steel–graphite multilayer membrane. The TiN layer absorbs light and rapidly thermally expands, launching elastic waves in the membrane. Water serves to damp and localize the Lamb waves near the illuminated region. The vibration is localized (amplitude decays by ~3 orders within 10 µs), and pulses are effectively independent given the 100 ms inter-pulse interval.
• Membrane structure and materials: The multilayer consists of a 30 nm TiN absorption layer (on the backside of a 304 stainless-steel substrate), a 5 µm-thick stainless-steel membrane, and a 10 nm graphite top layer to reduce adhesion. The substrate is 50 mm × 50 mm stainless steel; the front side is coated with graphite, and the backside with TiN via e-beam evaporation. Particles used include silica (density 2500 kg/m³), zirconia, and pollen, with sizes from ~25–100 µm. The method is tolerant to particle size and material variations.
• Wave excitation and characterization: DMD-defined stripes (typically 0.4 mm wide and 4 mm long) are illuminated to excite antisymmetric Lamb waves that propagate outward along the membrane at ~60 m/s. Simulations and laser vibrometer measurements provide space–time deformation maps; 2D Fourier analysis of measured fields yields dispersion relations matching analytical Lamb-wave dispersion with wavelengths ~0.2–0.8 mm and frequencies ~0.1–1.0 MHz.
• Pattern resolution and stripe width: Simulations show increasing stripe width raises vibration strength but reduces pattern resolution; widths >430 µm cause overlap, <280 µm insufficient excitation. An optimal width of 400 µm achieves ~270 µm line resolution.
• Particle dynamics: Particles on the vibrating surface experience energy transfer from membrane deformation. If kinetic energy exceeds adhesion (dominated by van der Waals forces), particles detach, jump, and resettle. Numerical integration of the motion using simulated deformation fields yields spatial variations in jumping height and detachment conditions, validated by high-speed imaging. The minimum laser power density to induce particle jumping is ~1.2 MW/cm²; jumping height increases linearly with power density due to linear temperature–displacement scaling.
• Instrumentation: Deformations are measured with a laser vibrometer (Polytech VFX-F-110). A high-speed camera (Flare 2M360MCL) records particle jumps; another camera (MV-CE200-10UC) records pattern formation. The DMD chip (TI S1410-9032) is 0.95 inch, 1400×1050 pixels.
• Theory and simulation: The opto-thermal response is derived by computing surface temperature rise, heat diffusion, thermoelastic strain, and initial displacement, then applying it as input to the vibration equation of an effectively infinite beam to obtain space–time deformation, showing linear dependence on laser power density. COMSOL simulations couple radiative heat absorption, heat transfer, solid mechanics, and pressure acoustics with thermal expansion and acoustic–structure interaction. For efficiency, the multilayer is homogenized as stainless steel with TiN optical absorption. Simulated domain widths are 6 mm; thicknesses: water 100 µm, membrane 5 µm. Material parameters include density 7930 kg/m³, Young’s modulus 210 GPa, Poisson’s ratio 0.3, specific heat 0.5 kJ kg⁻¹ K⁻¹, thermal conductivity 45 W m⁻¹ K⁻¹, and optical absorption 7×10⁻⁷ m⁻¹. An analytical expression for antisymmetric Lamb-wave frequency in the presence of water provides dispersion curves consistent with measurements.
• Patterning modes: Intaglio-style patterning removes particles from illuminated curves to reveal the figure; letterpress-style patterning removes background while retaining particles along contours. Letterpress is achieved by dynamic sequences of DMD frames that progressively drive particles outward from designated regions and then transport inner-region particles to the contours.

• PPAP demonstrates programmable mapping of laser-defined patterns into localized antisymmetric Lamb waves that rearrange tens of thousands of microparticles simultaneously in air.
• Wave characteristics: Propagation speed ~60 m/s; wavelengths ~0.2–0.8 mm; frequencies ~0.1–1.0 MHz; strong localization (amplitude decays by ~1000× within 10 µs). Simulations agree with vibrometer measurements and analytical dispersion.
• Resolution and optimal excitation: A 0.4 mm-wide laser stripe yields clear dark–white–dark bands and an effective line resolution of ~270 µm. Stripe widths >430 µm cause overlap; <280 µm yield insufficient excitation.
• Thresholds and scaling: Minimum laser power density to induce particle jumping is ~1.2 MW/cm². Deformation and particle jumping height scale approximately linearly with laser power density. For a particle mass of ~2×10⁻¹⁰ g and maximum acceleration ~4×10⁴ m/s², the estimated pushing force is ~80 µN, comparable to traditional acoustic tweezers, while operating on much smaller particles (~25 µm) in air.
• Patterning density and area: From an initially uniform coating (~4500 particles/mm²), PPAP forms patterns over large areas (e.g., 22 × 16 mm²). Static Pigsy figures emerge within ~3.5 s, with detailed features faithfully reproduced (e.g., eye details at ~270 µm resolution). PPAP also works with different particle types (silica 25–55 µm, zirconia 30–100 µm, pollen ~50 µm).
• Dynamic letterpress patterning: A kapok flower is produced using 1152 DMD frames at 10 fps. The first 576 frames remove outer-background particles (outer region nearly cleared by 57.6 s). The next 576 frames transport inner-region particles to the contours, with visible ripple-driven transport (~64.8 s snapshot).
• Controlled transport of few particles: PPAP moves 30 silica particles along a prescribed path over 10.0 mm at ~0.35 mm/s with ±120 µm accuracy, demonstrating precise, small-number manipulation comparable to holographic optical tweezers but independent of particle optical properties.

PPAP addresses the longstanding challenge of high-throughput, flexible, and precise manipulation of many particles by leveraging photoacoustically excited, localized Lamb waves that inherit the programmability of optical holography while providing acoustic-strength forces. The approach overcomes limitations of optical tweezers (small trap sizes, material dependence) and hardware complexity of phased acoustic arrays, and it avoids the precision loss associated with bubble formation in liquid-based photoacoustic methods by operating in air on a solid membrane. The close agreement between simulation, analytical dispersion, and measurements validates the physical model and supports predictive control of particle trajectories via laser-defined wavefields. The demonstrated static and dynamic patterning—intaglio and letterpress modes—over centimeter-scale areas with sub-millimeter line resolution and the ability to translate small particle sets along paths show the breadth of applications, from rapid micro-assembly and reconfigurable layouts to animated displays of particulate matter. The method’s tolerance to particle size and composition broadens applicability. Anticipated improvements in optical aberration correction, laser uniformity, and feedback via image recognition can further refine resolution, fidelity, and robustness. The potential to transfer patterned layers enables micro-3D printing, and adapting PPAP to liquids could extend precise control to bio/chemical systems.

The work introduces a chip-scale, programmable photoacoustic patterning technique that maps DMD-defined laser patterns into localized Lamb waves on a multilayer membrane to simultaneously manipulate and assemble large numbers of microparticles in air. It unifies the adaptability of optical holography with acoustic force capability, achieving centimeter-scale, high-fidelity intaglio and letterpress patterns, dynamic animations, and precise transport of small particle groups. Future directions include enhancing resolution and uniformity through improved optics and lasers, integrating closed-loop image-based feedback for error correction, transferring patterned layers for micro-3D printing, and extending the approach to liquid environments for broader bio/chemical applications.

• Resolution constraints: The effective line resolution is ~270 µm under optimal 0.4 mm stripe excitation; wider stripes (>430 µm) cause wave overlap and blur, while narrower (<280 µm) provide insufficient excitation, limiting pattern clarity.
• Power threshold and adhesion: Overcoming strong van der Waals adhesion requires sufficient kinetic energy, imposing a minimum laser power density (~1.2 MW/cm²) and potentially constraining very small or strongly adhering particles/surfaces.
• Measurement and modeling discrepancies: Minor differences between calculated and measured jumping heights are attributed to unmodeled camera measurement errors and particle rotation during jumps.
• Hardware and process considerations: Achieving higher fidelity depends on optical system aberrations and laser uniformity; the process requires high-energy pulsed lasers, a water layer for damping/localization, and a specialized multilayer membrane with low-adhesion coating. Letterpress-style patterning involves long frame sequences (up to ~115 s total) for full background clearing and contour formation.
• Uniform particle coating and collisions: Initial uniform spreading is needed; at high local densities, particle–particle collisions during jumps can cause scattering and transient disorder before settling.