Frequency-Dependent Dynamic Apertures

• Frequency-dependent dynamic apertures are systems whose effective area varies with signal frequency, influencing stability in accelerators, resolution in imaging, and beamforming performance.
• They employ techniques such as parametric resonance analysis, adaptive F-number scheduling, and discrete state-space tuning to balance competing design criteria.
• Experimental outcomes demonstrate improvements like <20% simulation error in stability predictions, 12.8–24% enhanced imaging resolution, and beamforming gains within 3 dB of ideal performance.

A frequency-dependent dynamic aperture is a physical or algorithmic aperture whose transmission, reception, or effective area varies as a function of signal frequency. These systems arise naturally or are engineered in domains such as synchrotron light sources, ultrasound imaging, and programmable antenna arrays, where both the spatial filtering effect of the aperture and its response or acceptance vary non-trivially with frequency. Frequency-dependence in the dynamic aperture directly affects operational metrics such as stability regions in storage rings, image resolution and artifact suppression in medical imaging, and beamforming gain/bandwidth tradeoffs in reconfigurable antennas.

1. Physical and Mathematical Mechanisms in Frequency-Dependent Dynamic Apertures

In high-energy storage rings (e.g., $e^+e^-$ Higgs factories), the physical dynamic aperture is determined by the set of initial phase-space coordinates for which particle trajectories remain stable over long times. A critical, nontrivial mechanism leading to a frequency-dependent dynamic aperture is the interplay between synchrotron radiation losses—particularly from quadrupole magnets—and the coherent response of the betatron motion. Specifically, betatron oscillations in a quadrupole create a radiative energy modulation at twice the betatron frequency, $2\omega_\beta$. The resulting energy loss feeds back into the focusing strength, such that the restoring force in the transverse direction oscillates at this double frequency.

This leads to the following parametric-oscillator equation for the transverse displacement $y$:

$$\ddot y + 2\alpha_y\,\dot y + \Bigl[\omega_\beta^2 + \Delta K\,J_y\cos(2\omega_\beta t)\Bigr]\,y = 0,$$

where $\alpha_y$ is the radiation damping rate, $\omega_\beta$ is the betatron frequency, $\Delta K$ parameterizes the resonance drive strength, and $J_y$ is the vertical action. This equation describes a parametric resonance with several distinctive properties:

• The resonance is self-induced with zero detuning; the amplitude-dependent drive term tunes the system onto resonance regardless of initial settings.
• The strength of the resonance is proportional to the square of oscillation amplitude ($J_y$), leading to a nonlinear instability threshold.

This parametric excitation and the resulting boundary for stable motion (dynamic aperture) depend sensitively on the frequency content and harmonics of the lattice optics—specifically, the Fourier spectrum of the $K_1^2(s)\beta_y(s)$ pattern, where $K_1$ is the quadrupole gradient and $\beta_y$ is the vertical beta function (Bogomyagkov et al., 2018).

2. Frequency-Dependent Dynamic Apertures in Ultrasound Imaging

A digital array in ultrasound imaging utilizes dynamic receive and/or transmit apertures—subsets of array elements that are adaptively engaged as a function of focal depth. The classical "F-number" $F(z) = z/A(z)$ (with $A(z)$ the subaperture width at focal depth $z$) governs lateral beamwidth and artifact suppression. A frequency-independent (fixed) $F$-number imposes a trade-off between lateral resolution (improved with larger aperture/small $F$) and grating-lobe artifact suppression (improved with smaller aperture/large $F$).

The frequency-dependent F-number, $F(f)$, is introduced to dynamically widen the subaperture at low frequencies (where grating lobes are not a risk and the main lobe can be sharpened) and restrict it at higher frequencies to maintain appropriate spatial sampling and suppress grating lobes. The optimal $F(f)$ at each frequency is given by (Schiffner et al., 2021, Schiffner, 2024):

• At low $f$, $F(f)=1$ (full aperture).
• At high $f$, $F(f)$ increases so that the first grating lobe is kept beyond a design angular bound $\theta_{1b}$:

$$F(f) \geq 1 / (p/\lambda - \sin\theta_{1b}), \quad \text{for } p/\lambda > 1/(1+\sin\theta_{1b}).$$

• The effective subaperture width is $D(f,z) = z / F(f)$.
• Additional constraints (e.g., minimum depth of field for focal zone coverage, maximum allowable grating-lobe level) are included by forming

$$F(f) = \max\left\{ F_{\mathrm{DOF}}(f), F_{\mathrm{GL}}(f) \right\}$$

where $F_{\mathrm{DOF}}$ and $F_{\mathrm{GL}}$ are lower bounds for depth-of-field and grating-lobe suppression, respectively (Schiffner, 2024).

This adaptivity leads to measurable improvements in image contrast, lateral and axial resolution, and the uniformity of resolution across the imaging field.

3. Engineering and Optimization of Frequency-Selectable Dynamic Apertures in Metasurface Antennas

Dynamic metasurface antennas (DMAs) implement beamforming by adjusting the resonance frequencies of constituent elements (slots), each modeled as a Lorentzian magnetic dipole with tunable polarizability:

$$\alpha_{M,n}(f) = \frac{F \cdot (2\pi f)^2}{2\pi f_{r,n}^2 - 2\pi f^2 + \Gamma f}$$

where $f_{r,n}$ is the programmable resonance of element $n$ and $\Gamma$ denotes damping. The beamforming gain for a desired direction and operating frequency $f_{\mathrm{op}}$ is a product of the array factor and the frequency-dependent element gain.

A two-stage optimization is executed:

1. For a given operating frequency and target direction, assign $f_{r,n}$ for each element to maximize gain.
2. Optimize $f_{\mathrm{op}}$ itself to compensate for Lorentzian roll-off and maximally align the beam with available element responses.

In hardware, element resonance states are discretized ($M$-state codebook), but even with a small $M$ (e.g., $M=4$), the system achieves near-ideal performance. Frequency-selective dynamic apertures produced by this process yield robust wideband beamforming across a controlled angular sector and allow for rapid single-shot beam training using frequency-division pilots (Deshpande et al., 2024).

4. Sensitivity to Frequency and System Parameters

The effectiveness and limitations of a frequency-dependent dynamic aperture are acutely sensitive to the underlying frequency structure of the system:

• In storage rings, the vertical dynamic aperture $y_{\max}(\nu_y)$ exhibits deep minima when $2\nu_y$ (twice the betatron tune) aligns with strong Fourier harmonics of $K_1^2(s)\beta_y(s)$. Small changes in $\nu_y$ can markedly reduce stability through enhanced parametric driving at these "resonant" harmonics (Bogomyagkov et al., 2018).
• For imaging arrays, the resolution/grating lobe balance transitions sharply as $p/\lambda$ (array pitch to wavelength) crosses thresholds set by the angular bound, with $F(f)$ diverging at singular points requiring practical apodization or saturation (Schiffner et al., 2021).
• In DMA-based antennas, the Lorentzian element response bandwidth (parametrized by $\Gamma$) controls the operational frequency window, with beamforming gain degrading outside the 3 dB bandwidth. The quantization of element resonance settings minimally impact achievable performance until the codebook becomes overly coarse (Deshpande et al., 2024).

5. Experimental Outcomes and Quantitative Metrics

Experimental validation across domains demonstrates the efficacy of frequency-dependent dynamic apertures:

• In synchrotron storage rings (FCC-ee, 45 GeV), the analytical expressions for dynamic aperture accurately predict ($<$20% error) the simulation/tracking limit for vertical stability, and reveal the tune-dependence and sensitivity to optics harmonics (Bogomyagkov et al., 2018).
• In ultrasound imaging (SonixTouch L14-5/38, CIRS 040 phantom), frequency-dependent subapertures improve contrast-to-noise ratio (gCNR) by 2.0–3.2%, lateral resolution by 12.8–24%, and uniformity by up to 14.1%, compared to fixed F-number or full-aperture imaging, all at real-time computational cost (Schiffner et al., 2021, Schiffner, 2024).
• For DMA antenna arrays ($N=8$, $f\in[12,18]\,\mathrm{GHz}$), beamforming gain remains within 3 dB of the ideal array value over a 300 MHz bandwidth and an angular sector of $\pm30^\circ$, using only four quantized resonance states. Single-shot beam training incurs less than 0.1 dB penalty versus perfect angle estimation (Deshpande et al., 2024).

6. Design Strategies and Outlook

Design guidance for frequency-dependent dynamic apertures includes:

• In accelerators, suppress or redistribute the spectral power of $K_1^2(s)\beta_y(s)$ harmonics near $n\approx-2\nu_y$ via optics flattening or compensating magnets to mitigate parametric resonance (Bogomyagkov et al., 2018).
• In imaging arrays, select imaging bands and F-number schedules that match clinical depth-of-field requirements and minimize resolvable grating lobes, possibly via precomputed $F(f)$ maps and filter-bank implementations for real-time operation (Schiffner et al., 2021, Schiffner, 2024).
• For programmable apertures, treat frequency as a co-optimized parameter alongside spatial phase; use small discrete state-space tunable elements rather than ideal analog settings for efficient hardware (Deshpande et al., 2024).

A plausible implication is that frequency-dependent dynamic apertures unify the physical and algorithmic tuning of system performance in a wide range of fields, offering principled, analytically tractable, and experimentally validated strategies for balancing competing operational criteria such as resolution, artifact suppression, stability, and power consumption.