A 108 MHz four-vane radio frequency quadrupole (RFQ) with a piecewise linear intervane voltage profile has been manufactured. The intervane voltage is designed to increase linearly within the shaper and bunching section of the RFQ to enhance the acceleration gradient, while remaining constant in the other sections to control the peak surface electric field. The feasibility of this configuration was achieved through a specialized design of the vane base featuring a localized dip and was demonstrated by RF field tuning, which involved measuring the electric field near the axis and adjusting the insertion depths of 48 tuners.

This work presents a dynamic phase-driven synchronization scheme for inverse Smith-Purcell dielectric laser acceleration, leveraging controlled phase slippage to enable simultaneous transverse alternating focusing and continuous acceleration. Unlike conventional alternating phase focusing approaches, the proposed scheme supports particle capture over a continuous phase range, greatly relaxing alignment tolerances and simplifying device fabrication. Full-wave simulations demonstrate outstanding performance: a peak acceleration gradient of 39 MV/m with 95% capture efficiency for a 29.5 keV electron bunch, all within a compact acceleration channel of only 120 μm in length.

Few-cycle laser pulses offer new routes to drive ion acceleration via transient electrostatic shocks, yet how the carrier-envelope phase (CEP) governs these processes remains poorly understood. Using particle-in-cell simulations and theoretical modeling, we show that the CEP critically selects which laser field peak initiates electron displacement, thereby controlling shock strength and proton energy. We find a cosine-like dependence of proton cutoff energy on the CEP. A predictive framework for CEP-controlled plasma dynamics is established, advancing the optimization of compact high-energy proton sources and offering insights into astrophysical particle acceleration.

We propose a new acceleration scheme for fixed-field accelerators in which beams corresponding to different RF harmonic numbers are accelerated simultaneously and connected into a single, continuous acceleration path. By completing acceleration over two successive RF frequency sweeps, the method limits the overall sweep range while shortening the repetition cycle.Longitudinal simulations based on realistic machine parameters demonstrate stable operation of the second-harmonic bucket, establishing the practical feasibility of wide-range acceleration within a unified framework.

The engineering reliability of high-power accelerator depends critically on understanding complex error couplings. We perform a comprehensive global Sobol sensitivity analysis of the Medium Energy Beam Transport (MEBT) section in an accelerator-driven system test facility. The results identify dominant tolerance drivers and quantify nonlinear interactions that govern beam loss and emittance growth. This approach establishes a practical methodology for physics-informed tolerance design in high-intensity accelerators.

Muonium-to-antimuonium conversion provides a sensitive probe of charged lepton flavor violation. We present the design and performance of the positron transport system for the Muonium-to-Antimuonium Conversion Experiment (MACE), combining an electrostatic accelerator with a solenoid system. Simulations show high geometric acceptance and precise position resolution, enabling efficient signal transport and background rejection. Dedicated studies demonstrate that the system, together with optimized collimation and time-of-flight selection, can effectively suppress internal conversion decay backgrounds, providing a new approach to positron transport for experiments at the high-intensity frontier.

The accelerating structure is a key component of particle linear accelerators. The innovative S-band RF cavity (operating at 3 GHz) that we have developed and tested at high-power pushes the frontiers of high gradient RF cavities, achieving an accelerating gradient of 38 MV/m with duty factor of up to 0.035% and accelerating gradient of 24 MV/m with duty factor of 0.1%. This research contributes to the advancement of compact particle linear accelerators.

In 2025, the 166.6 MHz high-power, high-order-modes-damped β = 1 quarter-wave superconducting cavity modules—developed for the High Energy Photon Source—successfully accelerated a 100 mA electron beam. This milestone marks the world’s first use of a superconducting quarter-wave cavity as the main accelerating system in a high-current storage ring. By extending the application limits of quarter-wave cavities from their conventional use in low-β accelerators to high-current electron storage rings, this achievement establishes a proven pathway for future high-performance light sources.

Screen-based transverse beam imaging in radiation environments is limited by the survivability of cameras and electronics. We demonstrate a proof-of-concept chain that relays the optical signal through a multimode fiber and uses a synthetic-data-trained autoencoder to reconstruct the screen intensity distribution from the fiber output pattern. Tests using beam images recorded at CERN’s CLEAR facility and replayed in a lab setup show good agreement for beam centroid and transverse sizes, suggesting a route to remote imaging with reduced reliance on beam data.

We demonstrate for the first time how dynamic models of global rf control loops, such as beam-phase loops, and local rf loops, such as fast rf feedbacks, can be coupled in a macroparticle tracking code. Extensive benchmarks against measurements from the CERN accelerators show excellent agreement and predictive capability. This advance enables highly realistic beam dynamics simulations capturing the interplay between rf control systems and the circulating beam in critical transient regimes, such as at bunch-to-bucket transfers.

We employ hydroforming technology to develop a 1.3 GHz one-cell Nb/Cu full-seamless cavity to reduce the cost and improve performance. Following hydroforming, the cavities were coated with niobium via high-power impulse magnetron sputtering (HiPIMS). We conducted RF measurements of the cavities, achieving an accelerating gradient of 12 MV/m at 4.2 K. The quality factor was slightly below requirements, attributed to surface roughness from the hydroforming process. The maximum accelerating gradient reached 15.7 MV/m at 1.85 K with no field emission. We demonstrate the advantages of full-seamless cavities. including reduced manufacturing costs, high reproducibility, and the elimination of welding seams.

Superconducting radio-frequency (SRF) cavities were first used to accelerate high-velocity electron/positron beams and low-velocity ion beams. High-velocity SRF cavities in particular have benefited from extensive studies and significant performance improvements in the past 6 decades. More recently, SRF technology has been applied to intermediate-velocity beams using half-wave resonators (HWRs) and spoke cavities. Methods to improve HWR performance for the Facility for Rare Isotope Beams (FRIB) are being explored. With electro-polishing (EP), low-temperature baking (LTB), and mitigation of thermo-electric currents, the quality factor at high accelerating gradient improves significantly.

Nonlinear lattice design traditionally relies on minimizing resonance driving terms (RDT) and amplitude-dependent detuning (ADD) to enlarge dynamic aperture. We show at NSLS-II that directly suppressing global chaos provides a stronger correlation with dynamic aperture and enables comparable performance using fewer sextupoles. This challenges the conventional emphasis on RDT and ADD minimization and suggests chaos-based optimization as a more robust strategy.

While complex, caustic-like charge patterns driven by dynamic Coulomb expansion have been observed in ultracold ion beams, their manifestation in electrons remains unexplored due to much faster timescales involved. We employ high-fidelity N-body simulations to show how dense, nonuniform electron beamlets emitted from nanostructured cathodes form coherent patterns or merge into a single beam. By identifying the temperature and density thresholds necessary to sustain these patterns, and the effect of geometry and asymmetry, we establish practical design criteria for optimizing next-generation cathodes and introduce a novel virtual diagnostic framework for emittance measurement.

The shape design of radio frequency (RF) cavities in particle accelerators is a computationally expensive multiobjective problem with extremely strict equality constraints. We report the first application of multiobjective Bayesian optimization (MOBO) to the shape design of RF cavities and introduce a two-stage acquisition strategy that effectively handles such constraints within the MOBO framework. The method rapidly identifies high-quality Pareto-optimal solutions, demonstrating substantial advantages over conventional approaches in both convergence speed and solution quality.

This work combines convolutional layers and attention mechanisms to enable a model to effectively extract both global and local features. This approach optimizes the use of the most effective correctors during closed orbit correction, enhancing process efficiency while better mitigating corrector saturation, thereby contributing to long term stability. By integrating Gaussian process models, the neural network addresses input length mismatches caused by Beam Position Monitor failures, ensuring its applicability to real world devices.