Research
I currently study the physics of intense beams generated in high-power hadron accelerators. Such beams are used to create secondary beams of neutrons, neutrinos, and muons for experiments in condensed matter physics, material science, nuclear physics, and elementary particle physics. In the future, intense beams could be used to drive subcritical nuclear reactors.
Several accelerators now generate more than 1 megawatt (MW) of beam power, with future machines moving toward 10 MW. As the beam power/intensity increases, particle-particle interactions become more significant. Particles interact directly via Coloumb forces—also called space charge forces—and indirectly via electromagnetic wakefields. When combined with the electromagnetic fields in the accelerator, these collective effects generate chaotic, difficult-to-predict particle trajectories. Unpredicted particle losses, which occur when particles wander outside the accelerator aperture, are especially important because lost particles generate radiation when they interact with the environment. The lost beam power must remain below one watt per meter (1 W/m)—around one out of every million particles—to keep radiation levels safe. Modern accelerators already operate near this limit, so it is not possible to simply scale the beam intensity by a factor of 10.
A complete understanding of beam loss will require detailed theoretical and computational models of halo formation. Such models are difficult to validate due to uncertainties in the beam and accelerator state. Thus, it is important to measure the particle distribution in the accelerator, characterize the uncertainties in the model parameters, and validate models against direct measurements. A separate but related problem is to control the beam in the accelerator. For example, can we avoid instabilities by shaping the particle distribution in phase space? Can we minimize particle losses by tweaking the force fields applied to the beam?
Below are short summaries of a few projects I’ve worked on in this area.
Halo formation in linear accelerators
Linear accelerators (linacs) generate tight bunches of particles. The combined effects of applied and self-generated fields drive some particles to large amplitudes, creating a low-density “halo” surrounding a dense core. Halo is nearly invisible and barely influences the core, but it places hard constraints on the beam power since it leads to beam loss. Halo formation is challenging to predict because of its sensitivity to model parameters. In particular, we do not know the initial particle distribution in the full six-dimensional (6D) position-momentum space, also called phase space. It is also challenging to measure such a faint signal on top of the bright core.
We are using the Beam Test Facility (BTF) at the Spallation Neutron Source (SNS) to address this problem on a small scale. The BTF is a replica of the first few meters of the 400-meter SNS linac. It is equipped to measure the initial 6D distribution and the output 2D halo distribution. We are leveraging these unique diagnostics to benchmark particle-in-cell (PIC) simulations at a new level of detail and better understand halo formation.
Novel beam injection techniques
Our primary method to generate intense beams is to repeatedly inject small pulses from a linear accelerator (linac) into a circular accelerator (ring). Space charge forces drive resonances and collective instabilities in the circulating beam, leading to particle loss. One approach to mitigate space charge is to vary the position and momentum of the injected beam relative to the circulating beam during accumulation. We call this phase space painting. We’re studying a new painting method to generate a vortex beam, where particles swirl inside an elliptical boundary. Such a beam would minimize the nonlinear component of space charge forces, potentially leading to higher intensities. At the same time, the beam would occupy a tiny volume of phase space, making it potentially useful in the first stage of a high-energy collider.
Proton bunch compression
A future muon collider (MC) would require extremely short, intense proton pulses to generate the initial muon beam. The proposed scheme uses an SNS-style linac and accumulator ring followed by a separate ring to compress the bunch to a few nanoseconds in length. This extreme bunch compression has not been tested for hadron beams; space charge effects, impedances, and electron cloud instabilities could be an issue. We are examining whether a similar bunch compression scheme could be tested at the SNS.
Phase space tomography
Classical particles are described by their position and momentum in three-dimensional space. Together, these form a six-dimensional phase space. Direct phase space measurements are not always possible; sometimes we have only partial information, such as the particle density along a single dimension. Still, partial information constrains the unknown phase space distribution, and one may attempt to find a distribution consistent with these constraints. I am particularly interested in applying maximum-entropy and Bayesian methods to this problem to incorporate prior knowledge and quantify uncertainty in the reconstruction.
