Making and using finely focused X ray beams
June 17, 2020
We all know that X rays can penetrate thick objects like people, but two recent advances in Applied Physics at Northwestern have advanced the state of the art in focusing them and in interpreting x-ray images. Applied Physics graduate Kenan Li (PhD 2017) and present PhD student Sajid Ali teamed up to study the focusing properties of Fresnel zone plates. These diffractive optics have an old history, but Li’s work advanced the state of the art on making them have narrow structures (required for a tight focus) that are very thick (required for good focusing efficiency). In fact, the aspect ratio (the ratio of thickness over width) of these nanostructure is unprecedented in x-ray optics: over 100 to 1, so that Ali had to undertake calculations to understand the required accuracy of angular alignment. Their teamwork led to tests at Brookhaven Lab that showed these energy-tunable optics deliver an unparalleled combination of 14 nanometer resolution (a single atom is only about 0.2 nanometers across) and 6% efficiency - which may not sound very good until you learn that previous optics of this type were less than 1% efficient! This work was published in May in Optica, the premier journal of the Optical Society of America.
That’s how you can focus coherent x-ray beams. But what about using these beams for imaging? That’s where the work of Ming Du (Applied Physics PhD 2019) and present PhD student Saugat Kandel comes in. Just like a photographer may change the aperture on their camera to trade off a sharper image with a loss of the depth of focus (so that you don’t have a good view of features at different distances away), when improving the resolution of x-ray microscopes you decrease the thickness of sample which is in focus. But X rays are unique in being able to image things that are thick! That’s where the work of Du and Kandel comes in. We know the physics of how a coherent x-ray wave propagates through a thick object, so if we guess at the object we can calculate what we would see in our measurement. Yet we know what we measured, so we need to adjust our guess to get the calculations and the measurements closer to each other. This is done using a method called “automatic differentiation,” where a supercomputer does something akin to the chain rule in calculus to figure out the best adjustment to the guess of the object until we converge. This work was published in the March 27 issue of Science Advances.
These developments are timely. Argonne National Laboratory is gearing up for an $800 million upgrade of the Advanced Photon Source to make it the brightest hard x-ray synchrotron light source on the planet, and Argonne is also preparing to receive its first Exascale supercomputer, a $500 million machine called Aurora. With our advances in making x-ray beams and using supercomputers to improve x-ray images, Prof. Chris Jacobsen’s team in Applied Physics at Northwestern is helping blaze the trail to get the most from these national investments by the US Department of Energy.
The papers:
Kenan Li, Sajid Ali, Michael Wojcik, Vincent De Andrade, Xiaojing Huang, Hanfei Yan, Yong S. Chu, Evgeny Nazaretski, Ajith Pattammattel, and Chris Jacobsen, "Tunable hard x-ray nanofocusing with Fresnel zone plates fabricated using deep etching," Optica 7, 410-416 (2020).
https://www.osapublishing.org/optica/abstract.cfm?uri=optica-7-5-410
Ming Du, Youssef S. G. Nashed, Saugat Kandel, Doğa Gürsoy, and Chris Jacobsen, "Three dimensions, two microscopes, one code: automatic differentiation for x-ray nanotomography beyond the depth of focus limit," Science Advances 6, eaay3700 (2020)
https://advances.sciencemag.org/content/6/13/eaay3700
This research used resources of the Advanced Photon Source and the Argonne Leadership Computing Facility, which are U.S. Department of Energy (DOE) Office of Science User Facilities operated for the DOE Office of Science by Argonne National Laboratory under contract no. DE-AC02-06CH11357. We also acknowledge the National Institute of Mental Health, NIH, for support under grant R01 MH115265.