We have developed a versatile empirical model to predict the correct mean charge states ($q_m^0$) and the charge state distributions (CSDs) outside the target, which compare well with the experimentally measured quantities for the entire range of projectile ions in the energy range of 1-4 MeV/u. Previous models such as SGM (Schiwietz and Grande), and ISGM (Schiwietz et al) could predict only the $q_m^0$. Comparison between previous models and the present one on $q_m^0$ with experimental data shows that the present model is closest to the measured data. We believe that such precise $q_m^0$ and CSDs will be highly useful in both ion-atom collisions and accelerator physics.

A critical problem in Laser Wakefield Acceleration (LWFA) is converting beams with 100% energy spread into narrow ones with minimal dark current, which are essential for many applications. We demonstrate for the first time how Bayesian optimization can rapidly achieve this without quadrupole assistance. Correct spectral phase is key to performance, but inter-parameter couplings obscure its role. Realistic simulations that match the experiment closely reveal how a positively skewed temporal profile is essential for improving spectral density. These results highlight the impact of algorithmic optimization as well as the underutilized role of temporal shaping in maximizing LWFA performance.

Accurate knowledge of beam-pipe impedance is essential for controlling wakefields and energy loss in modern accelerators. We derive a comprehensive theory that predicts the impedance of metallic vacuum pipes under a wide range of materials, temperatures, and bunch durations including regimes where classical models fail. This unified approach helps improve the reliability and performance of future high-brightness machines.

This work reports the design and experimental validation of a novel tunable-gradient permanent quadrupole magnet (PQM) that combines compactness, energy efficiency, and precise field control. The prototype delivers a continuous tuning range of 39 T/m with a peak gradient of 65 T/m, while maintaining excellent field quality with multipole errors below 5 × 10⁻⁴. Its fully permanent-magnet configuration eliminates the need for power supplies andmcooling systems, offering a compact and sustainable solution for next-generation light sources and free-electron lasers.

On-line monitoring of beam status is crucial for stable operation of accelerators, especially high-power machines. Here we introduce two novel methods for the on-line monitoring of beam phase and energy gain, based on the beam-cavity interaction. The methods are successfully validated in both single-cell and multicell cavities, demonstrating their broad applicability for accelerator control and diagnostics. The precision of the methods improves with the increase of the beam current, and remains satisfactory even when the current is only about 3 mA.

A reinforcement learning algorithm for real time luminosity optimization in colliders is designed and benchmarked. Prior to training, the input data undergoes multi-stage preprocessing before being fed into a neural network. The approach combines offline pretraining on historical accelerator data with online fine-tuning during operation and shows an average improvement of about 4% in luminosity. Detailed algorithmic descriptions as well as a high-level statistical summary of the new tuning algorithm are provided.

As modern accelerators reach higher and higher brightness beams, there is more potential for damage due to direct beam strikes with beamline components. Particle tracking and matter interaction simulations are commonly used in accelerator physics but there are no simulation codes which capture all the physics present in these interactions. We present a method for coupling several software programs to capture the dynamic physics present and demonstrate its use in modeling beam losses in the Advanced Photon Source.

Carbon is an unavoidable contaminant on metal surfaces, yet its precise impact on SRF cavity performance remains a critical question. This study systematically investigates the underlying mechanisms through which distinct forms of carbon contamination, including adsorbed carbon and deposited carbon, degrade SRF cavity performance, and explores corresponding recovery strategies. The findings advance our understanding of performance degradation and recovery mechanisms induced by carbon contamination, thereby establishing a foundation for selecting optimal online recovery methods.

The separation of photo-electron excitation and electron transport processes in the studied novel hybrid metallized diamond(001) photocathode highlights a potential path for the development of robust wide bandgap photocathodes emitting at visible wavelengths.

The brightness of an electron source is strongly influenced by the underlying electron-emission mechanisms. This work presents a unified theory for calculating the mean transverse energy of electrons produced through mixed thermionic, multiphoton, and field emission processes, enabling accurate estimates of source brightness when multiple mechanisms are involved. The theory successfully accounts for thermionically assisted multiphoton emission from a copper photocathode illuminated by an intense infrared laser.

Fluctuations in electron time-of-flight (ToF) during lattice transport is crucial for various applications, such as ultrafast electron diffraction, optical stochastic cooling in storage rings (SR), and sustaining steady-state microbunching in SR. This paper introduces a design concept aimed at reducing the electron beam’s ToF fluctuation in an isochronous lattice up to the sub-attosecond level, while nonlinearity is minimized based on high order achromat lattice. We also present an optimized design example achieving $\approx$0.3 attoseconds root-mean-square ToF fluctuation (equivalent to 0.1 nm) with specific beam emittance values of ${ϵ}_x$ ~1 nm, ${ϵ}_y$ ~1 nm, and ${\sigma }_{\delta }$ ~${10}^{-3}$

This work is dedicated to the study of X-ray transition radiation by 2–10 GeV electrons in a thin target placed in a strong magnetic field. Due to interference effects, the total emission can differ significantly from the simple sum of transition and synchrotron radiation and both constructive and destructive interference is possible. The magnitude of interference effects as a function of electron energy and detector acceptance angle is studied.

Unplanned anomalies in radio-frequency (RF) stations limit uptime and degrade beam quality in large accelerators such as SLAC’s Linac Coherent Light Source (LCLS). Machine learning is widely used for accelerator fault detection, but most methods require labels or assume entirely normal training data. We instead employ Coincident Anomaly Detection (CoAD), a multimodal, label-free framework that remains robust even when the training data contain many anomalies. When applied to high-rate RF phase and beam data at LCLS, phase-based CoAD detects nearly three times as many RF station faults as the amplitude-based baseline and improves diagnosis and root-cause analysis.