Multiplexed Tweezer Generation Methods

• Multiplexed tweezer generation is the creation of arrays of finely controlled trap potentials for parallel manipulation of atoms, molecules, or nanoparticles.
• It employs methods such as static beam splitting, acousto-optic deflection, and SLM/metasurface holography to achieve high configurability and uniformity.
• These techniques are pivotal for applications in quantum simulation, biophysics, and colloidal science, enabling scalable and precision-controlled experiments.

Multiplexed tweezer generation refers to the creation of arrays of tightly focused optical, acoustical, or other potential minima, with spatial and temporal control, such that multiple particles—including atoms, molecules, or micro/nanoparticles—can be trapped and manipulated in parallel. This capability enables scalable implementations for quantum simulation, quantum information processing, precision measurements, biophysics, and colloidal science. Multiplexing techniques span static beam splitting (e.g., using diffractive or refractive elements), spatial light modulation, acousto-optic or electro-optic beam steering, metasurface holography, and time-division “painting” schemes. Each method has distinct performance characteristics in terms of geometry flexibility, scalability, uniformity, and dynamical reconfigurability.

1. Principles of Multiplexed Tweezer Array Formation

Multiplexed tweezer generation exploits the superposition and control of optical or acoustic fields to produce multiple independent or correlated trapping sites. The essential theoretical basis is the design of a target intensity/potential landscape, typically in the focal plane of a high numerical aperture (NA) objective, achieved by modulation of the illuminating field’s amplitude, phase, polarization, or frequency.

For static optical configurations, the trap potential for a focused Gaussian beam is $$U(r, z) = U_0 \exp[-2r^2 / w_0^2] \cdot (1 + z^2 / z_R^2)^{-1}$$, where $w_0$ is the beam waist and $z_R$ is the Rayleigh range, while the oscillation frequencies are $\omega_r = \sqrt{4U_0/(mw_0^2)}$ (radial) and $\omega_z = \sqrt{2U_0/(mz_R^2)}$ (axial) (Anderegg et al., 2019). In time-multiplexed setups, the time average of sequential potentials yields an effective multi-well landscape provided the dwell and cycle rates satisfy parametric heating and trap-depth dilution constraints (Gosar et al., 2022, Chisholm et al., 2018). Inter-trap crosstalk, power allocation, and aberration correction are key design factors regardless of multiplexing modality.

2. Static and Dynamic Acousto-Optic Multiplexing

Acousto-optic deflectors (AODs) and modulators (AOMs) enable fast and programmable two- or three-dimensional beam steering and can generate multiplexed arrays via frequency multiplexing or time-division.

• Multi-tone static drive: By feeding the AOD a set of simultaneous radiofrequency (RF) tones, each frequency diffracts a proportion of the input laser to a distinct angle, forming a linear or two-dimensional array. The trap positions $x_i$ scale with the RF frequency $f_i$ ($x_i \propto f_i$), and independent amplitude control $A_i$ enables trap-depth balancing (e.g., $U_{0,i} \propto A_i$) (Anderegg et al., 2019). This technique is effective up to tens of traps but is practically limited by RF bandwidth and inter-tone crosstalk.
• Time-averaged (stroboscopic/painting) multiplexing: Rapid toggling among $N$ frequency settings generates an effective $N$-site array. Each site is “on” for $1/N$ of the cycle and dwell/cycle rates must exceed atomic oscillation frequencies to avoid heating and atom loss. Under these conditions, the effective potential is the arithmetic mean of the instantaneous trap potentials (Gosar et al., 2022, Chisholm et al., 2018).
• Hybrid schemes: Static multi-tone arrays can be overlaid with fast time-multiplexed “correction” fields to balance depths and extend the number of viable traps (Yan et al., 2022).

Table 1: Comparison of AOD-driven multiplexed tweezer schemes

Scheme	Number of Traps	Dynamic Reconfigurability	Uniformity Control
Static multi-tone	up to 10–50	Slow (RF update)	Per-tone amplitude
Time-multiplexing	10–1000*	Sub-ms by FPGA/DDS update	Via dwell/power
Hybrid static+strobe	Dozens (extendable)	Moderate, amplitude/frequency	Optimization required

*Upper limit is set by the update rate and available optical power.

3. Holographic, Metasurface, and SLM-Based Multiplexing

Spatial light modulators (SLMs)—including transmissive and reflective liquid-crystal devices, digital micromirror devices (DMDs), and recently, metasurfaces—enable high-dimensional and arbitrary pattern generation by controlling the phase and/or amplitude of the trapping field at thousands to millions of pixels.

• Gerchberg–Saxton (GS) and weighted GS (GSW) algorithms: These Fourier-domain iterative phase-retrieval methods are widely employed to encode arbitrary trap arrays into the SLM phase for static or quasi-static operation. Incorporation of per-spot weighting and successive feedback allows for non-uniformity reduction below 1.1% in arrays exceeding 1000 spots (Wang et al., 2024).
• Metasurface holography: Nanofabricated subwavelength structures enable direct phase and polarization engineering at the input aperture. Their high pixel density (e.g., $d\approx\lambda/2$) yields "effective NA" $>0.6$, enabling tight focusing and uniformity $\sim7.5\%$ with trap-counts exceeding $10^5$ (Holman et al., 2024). Reverse-projection algorithms operating on the relative-phase manifold further achieve $<5\%$ intensity variations with robust resistance to aberrations (Zhu et al., 1 Dec 2025).
• DMD-coupled microlens arrays: These combine the mechanical stability of solid lenslets with real-time per-site digital control. Arrays of up to 97 individually controllable traps have been demonstrated, with reconfigurable patterns up to 4 kHz and intensity uniformity adjustable to $±0.2\%$ (Schäffner et al., 2019).

4. Specialized Multiplexing: Vortex, Acoustic, and Advanced Trap Designs

• Multiplexed vortex tweezers: A spiral-phase mask with concentric radial zones carrying independent topological charges generates multiple concentric vortex traps, each imparting orbital angular momentum (OAM) of different magnitude and direction to trapped particles. Independent rotation and counter-rotating dynamics have been demonstrated for up to $N=4$–6 concentric rings, constrained by SLM pixel count and annular thickness (Muñoz-Pérez et al., 2023).
• Single-beam multiplexed acoustical tweezers: A focused ultrasonic transducer naturally generates an axial standing-wave with multiple prefocal intensity maxima. Each local minimum in the gradient potential provides a stable spatially separated trap, with the number and spacing tunable via geometric and acoustic parameters, e.g., $Δz\approx 2πzz_0/(kb^2)$ (Silva et al., 2014).
• Polarization- and phase-encoded tweezers: Metasurface holography can endow each trap with arbitrary polarization (e.g., $|H\rangle$, $|V\rangle$, $|L\rangle$, $|R\rangle$), or global phase structures (vortex, radial, azimuthal), using reverse-projection algorithms that optimize only the relative phase degrees of freedom across traps (Zhu et al., 1 Dec 2025).

5. Uniformity, Stability, and Feedback Correction

Achieving uniform trap depths and positions across large arrays is critical for quantum simulation and many-body experiments. Primary sources of non-uniformity are optical aberrations, SLM/DMD pixelation, and fabrication errors in metasurfaces.

• Feedback-intensity equalization: Real-time measurement of the array intensities (via CCD imaging) and iterative adjustment of target weights in the GSW loop reduces the root-mean-square (RMS) variation from $\sim11\%$ (raw GSW) to $<1.1\%$ in $>10^3$-site SLM arrays (Wang et al., 2024).
• Phase stability: Holographic approaches based on the relative-phase manifold ensure minimal intra-trap phase noise and enable robustness to aberrations; for example, intensity-pattern fidelity $F>0.98$ and crosstalk $<1\%$ have been obtained over $N=100$–400 sites (Zhu et al., 1 Dec 2025).

Table 2: Reported uniformity metrics for large-scale arrays

Approach	Trap Number	Uniformity (σ)	Reference
SLM GSW + feedback	>1000	<1.1%	(Wang et al., 2024)
Metasurface	256	~7.5%	(Holman et al., 2024)
DMD-MLA	97	±0.2% (adjustable)	(Schäffner et al., 2019)

6. Scalability, Dynamic Reconfiguration, and Limitations

The scalability of multiplexed tweezer platforms is set by the modulation or readout device bandwidth, available optical power per site, heating constraints, and physical size/pixelation.

• Optical devices: Metasurfaces can, in principle, realize $>10^5$ traps with uniformity $<10\%$ and sub-$\mu$m spacing (Holman et al., 2024); SLMs with feedback reach $>10^3$ traps and $<1.1\%$ non-uniformity (Wang et al., 2024). DMD-MLA approaches are currently limited to $\sim10^2$ spots, but rapid printing and demagnification re-scaling offer paths to higher densities (Schäffner et al., 2019).
• AOD/AOM systems: Multiplexing is bandwidth and update rate limited. Current AOD-based painting methods achieve $\sim100$--$1250$ sites in 1D/2D with negligible heating under $f_\text{trig}/N \gg 2\omega_r/2\pi$ (Chisholm et al., 2018). Power division and minimum dwell constraints set upper bounds on achievable $N$ for practical trap depths (Gosar et al., 2022).
• Limitations: Crosstalk and loss of uniformity arise with increasing N, especially in statically multiplexed AODs due to RF intermodulation; SLMs experience chromatic and polarization dependence, and metasurfaces are limited by nanofabrication tolerances and thermal handling.

Future prospects include integration of active feedback for ultra-large arrays, use of multi-wavelength or multi-plane metasurfaces for 3D trapping, and hybrid photonic-acoustic architectures for dynamic, high-stability trapping at scale.

7. Applications in Quantum, Biophysical, and Colloidal Systems

Multiplexed tweezer arrays underpin current neutral-atom quantum simulation architectures, especially for Rydberg-mediated entanglement and programmable Hubbard models (Yan et al., 2022). In ultracold molecules, multiplexed tweezers support collisional blockade single-molecule loading and control at high densities (e.g., single CaF arrays) (Anderegg et al., 2019). Vortex and polarization multiplexing open additional degrees of freedom for the study of OAM transfer, colloidal assembly, and rotational dynamics (Muñoz-Pérez et al., 2023). Multi-site acoustic trapping enables parallel microfluidic handling in optically opaque environments with forces $10^{-9}$–$10^{-8}$ N per trap (Silva et al., 2014). High-uniformity SLM and metasurface arrays are emerging as platforms for digital quantum computing and large-scale many-body state preparation (Wang et al., 2024, Holman et al., 2024).

Overall, multiplexed tweezer generation leverages advances in photonic, acoustic, and nanofabrication technologies to unlock new regimes of precise, parallel particle control across disciplines requiring high spatial, temporal, and configurational flexibility.