Three scientists from the U.S. Department of Energy's (DOE's) SLAC National Accelerator Laboratory will receive grants from the Early Career Research Program for their cutting-edge work in developing tools to better understand the physics of X-ray beams and push the boundaries of free-electron laser performance and ultrafast science; advancing microelectronics and machine learning to help hunt for new physics; and developing machine learning-enabled searches for primordial gravitational waves that could reveal more about the origins of the universe.
David Cesar, Julia Gonski and W.L. Kimmy Wu are among 91 recipients selected from a large pool of applicants from U.S. universities and DOE national laboratories, the DOE Office of Science announced today. In total, the DOE issued $138 million to support their work. Since 2010, the DOE has awarded these grants to support outstanding early career scientists who are performing research that supports the DOE Office of Science mission.
"Investing in cutting edge research and science is a cornerstone of DOE's mission and essential to maintaining America's role as a global innovation leader," said U.S. Secretary of Energy Jennifer M. Granholm. "The Biden-Harris administration is funding scientists and researchers at our nation's national labs and universities, early in their careers, ensuring they have the resources to expand scientific discovery and pursue solutions to some of the most complex questions."
Pushing the frontiers of ultrafast science
As a physicist, David Cesar has enjoyed trying to unlock the mysteries of how ultrafast phenomena, reactions and behavior that happen at billionths of a billionth of a second, or attoseconds, coalesce into macroscopic changes in the world - and building the tools to make that happen. "I love pushing the boundaries of what we can measure. Making a tool that is faster, is more powerful and can resolve at smaller length and timescales means I have a chance to see something no one else has ever seen before."
After completing his PhD in physics at the University of California, Los Angeles, where he worked on probing ultrafast dynamics with electrons traveling at nearly the speed of light, Cesar joined SLAC in 2019 as a research associate. "Attosecond science allows us to access the fastest, most fundamental subprocesses relevant for chemistry, materials science and more, and so promises to give an atomistic picture of the ultrafast universe," he said.
At SLAC - home to the Linac Coherent Light Source (LCLS), the world's most powerful X-ray laser - his research focuses on developing tools to create, control and measure ultrafast electron and X-ray beams and study how these intense beams interact with matter.
Although the lab's powerful X-ray laser beams allow scientists to peer into the inner workings of materials behavior or chemical reactions, Cesar would like to sharpen that view further. One way to do that is to produce shorter and more powerful X-ray pulses, called single-spike X-ray pulses, through X-ray laser-enhanced attosecond pulse generation (XLEAP). The single-spike X-ray pulses are partially synchronized, or partially coherent. "The coherence manifests itself as a stable, repeatable pulse structure," he said, which can be used to generate, or "seed," similar pulses downstream.
The physics behind single-spike pulses is largely unexplored because scientists do not have any instruments capable of measuring the relationship between the electron beam and X-rays, but Cesar would like to remedy that.
"This award will help me build a small team within the Accelerator Directorate focused on attosecond metrology, or measuring of electron beams and free-electron lasers," Cesar said. "The primary goal of this research is to directly measure the physics of single-spike attosecond free-electron lasers. In particular, we will build a new instrument that can simultaneously measure both the electron beam and the X-rays involved in the free-electron laser interaction. By referencing the X-ray pulse structure to the electron beam current, we will gain insight into the unique partial coherence of the single-spike X-ray pulse."
Cesar hopes the measurements from this instrument can be used to understand and take advantage of the seed-like properties of partially coherent radiation in delivering advanced attosecond lasers. Doing so could improve peak power and stability, shorten pulse length, add additional wavelengths, or colors, and more.
"Making shorter pulses allows us to study faster and faster dynamics. We can use multiple colors to do 'pump-probe' measurements, where one color triggers an interesting process and the second measures the system after a fixed time delay," Cesar said. "We are always trying to find ways to explore new frontiers and find new science."
Machine learning and microelectronics in the search for new physics
Julia Gonski's interest in particle physics started when she was 16. She was reading pop science books and got curious about the composition and rules of the fundamental universe. "I decided I wanted to be a particle physicist and have never looked back," she said.
During her undergraduate studies at Rutgers University, she analyzed data from the CMS detector at CERN's Large Hadron Collider (LHC) to search for evidence of supersymmetry, which predicts every elementary particle in the Standard Model has a partner particle. She then continued with graduate studies in the search for new physics at LHC's ATLAS experiment, earning her PhD from Harvard University. As a postdoc at Columbia University, she continued her work on ATLAS, focusing on detector electronics and using machine learning to detect anomalies in the data that could be evidence of new physics.
Gonski joined SLAC as a Panofsky Fellow in 2023, and today her research involves the development of and data collection from high-energy particle detectors, including the one at ATLAS.
"The main goal is to understand how the universe works at its most fundamental level, which requires understanding what particles exist that can't be broken down any further into constituents, and how they interact with one another," Gonski said.
Searching for these elusive particles requires developing advanced technologies, including machine learning. Machine learning, Gonski said, helps identify anomalies in a dataset - outliers and other oddities - that could indicate something new. She also wants to adopt "fast machine learning," where she can run real-time machine learning algorithms at the data source, known as edge machine learning, and throughout the data processing chain, a promising but relatively unexplored area.
In parallel, Gonski works to develop microelectronics, which involve special integrated circuits called field programmable gate arrays (FPGAs) and application specific integrated circuits (ASICs), that turn signals from particle collision events into data for offline analysis. "The unique challenges of on-detector data processing in collider physics, namely the high radiation environment and ultrafast latencies, mean that there's a lot of advanced engineering in this area that we have to draw from," Gonski said. She has been working with SLAC engineers to develop the algorithm for the hardware and the hardware that implements the algorithm at the same time, a concept called co-design. "The scientist-engineer relationship is crucial and a unique benefit of being in the SLAC lab environment," she said.
The Early Career Research Program grant, Gonski said, will provide support across all levels of her research. "The grant allows me to bring early career scientists, who are the lifeblood of research, into this work, while also giving them a valuable training opportunity in these key technologies that are driving U.S. innovation right now," she said.
Looking farther down the road, she feels that colliders benefit both science and society. "This award will be an excellent opportunity to drive the future of the field," Gonski said. "With some hard work and a bit of luck, I hope the projects laid out in my Early Career Research Program will lead to some novel search results in existing collider datasets and lay the groundwork for advanced detectors in future ones. Maintaining large international experiments, with thousands of people across the globe training in advanced technology and working toward a single mission, is a real force for good in the world."
Machine learning and the origins of the universe
Much of Kimmy Wu's day is spent searching for clues from the distant past to shed light on the origins of the universe. "It is incredible that the universe we inhabit gives us a way to probe processes that happened during its earliest moments," Wu said.
Wu, an associate scientist at SLAC, obtained her PhD in physics from Stanford University, where she was making observations of the cosmic microwave background, the radiation in the universe dating back approximately 380,000 years after the Big Bang. Wishing to continue work on the cosmic microwave background, she did her postdoctoral research first at the University of California, Berkeley, and then at the University of Chicago, working with the South Pole Telescope collaboration. She returned to the Bay Area and joined SLAC as a Panofsky Fellow in 2020.
Her research focuses on searching for evidence of primordial gravitational waves, which are from the earliest moments of the universe, using the BICEP and Keck Array telescopes and South Pole Telescope. Detecting these signatures, however, is challenging. "The signature is very, very faint," Wu explained. In addition, signatures of primordial gravitational waves are obscured by two types of contamination. One is foregrounds from within our own Galaxy: thermal radiation from interstellar dust, or microscopic bits of matter, and radiation from charged particles that travel in a spiral. Another is from weak gravitational lensing, a particular way that the structure of our universe bends and distorts cosmic microwave background light before it reaches us.
To get a more accurate picture of both contaminants and primordial gravitational waves, Wu is developing machine learning models that will be part of a larger analysis framework to find the most likely images of each component, given the input observations. With a good training dataset that avoids biased data, Wu hopes to create and deploy a robust machine learning model. "To find the signature, we need to better model these sources," she said. "The award will allow me to deploy some of these tools to tackle these problems."
The techniques her team develops, she believes, will also be broadly applicable for the cosmology community in analyzing survey data, for example, from the Vera C. Rubin Observatory/Legacy Survey of Space and Time, Dark Energy Spectroscopic Instrument and Cosmic Microwave Background Stage 4 projects.
"With the resources from this award, I am excited about exploring some super cool techniques, like advancing the modeling of the cosmic microwave background foregrounds and incorporating this model in the analysis of data from the BICEP/Keck Array and South Pole Telescope," Wu said.
LCLS is a DOE Office of Science user facility.
