Aug. 4, 2023
As an example of explainable AI (XAI), a real scanning electron microscope image of TATB powder (left) was modified by a trained neural network to increase its peak stress. The hypothetical image (right) outputted by the neural network is composed of smaller grains, indicating to materials scientists which changes need to be made to increase the material’s strength. Alternatively, a neural network can be asked to modify the powder in a certain way—by changing its size or porosity, for example—and the hypothetical image fed into a predictive model to predict the modified material’s physical properties. Photo courtesy of Shusen Liu.
Getting under the hood of machine learning
As machine learning techniques become increasingly used in the sciences, a team of researchers in Lawrence Livermore National Laboratory’s Computing and Physical and Life Sciences directorates are trying to provide a reasonable starting place for scientists who want to apply machine learning (ML) but lack a strong ML background.
The goal is to help users understand what’s going on behind the machine learning curtain, increasing trust in ML models.
The team’s work grew out of a Laboratory Directed Research and Development project on feedstock materials optimization, which led to a pair of papers about the types of questions a materials scientist may encounter when using machine learning tools, and how these tools behave.
Trusting artificial intelligence is easy when its conclusion is a simple ground truth, like identifying an animal. But when it comes to abstract scientific concepts, ML can seem more ambiguous.
The molecular crystal structure of (left) barium and (right) radium. Image courtesy of Oak Ridge National Laboratory.
Making radium crystal-clear
A team at Oak Ridge National Laboratory has reported the first synthesis and characterisation of a molecular radium compound by single crystal x-ray diffraction. The findings reveal the element’s coordination chemistry and could help develop new radium-based treatments for cancer.
The study’s results provide an important insight into radium’s chemistry. Up to now, researchers have used barium to make assumptions about radium’s interactions, yet the crystal structure of the equivalent barium complex showed significant differences.
Gauthier Deblonde, a radiochemist in the nuclear and chemical sciences division of Lawrence Livermore National Laboratory who was not involved in the research, said he is “really excited” about this latest preprint. “The new study is a veritable tour de force because the ORNL team managed to crystallise a pure radium compound and determined its single crystal XRD structure.”
NIF beamlines entering the lower hemisphere of the target chamber.
Break on through
For a moment, imagine a world of limitless energy – one where energy is so abundant that everyone can power their homes and businesses for mere pennies.
But thanks to a potential world-changing scientific breakthrough at Lawrence Livermore when researchers achieved fusion ignition, the ostensibly utopian world of limitless energy could soon become a reality.
Investors who place the right bets on the right stocks in this industry could mint fortunes over the next few years.
An overview of Saturn, which is believed to rain helium. Image courtesy of NASA.
It’s a gas
Have you ever wondered what makes it rain on other planets? So far, Earth is the only planet that receives rain because it has liquid water on its surface. However, there are records of precipitation on other planets as well, but not of water.
For example, on Saturn and Jupiter there is helium rain. It was discovered in 1977 by a professor at Cornell University.
However, so far it has not been possible to demonstrate that helium rain exists. This happens because on these planets “pressure and temperature conditions are close to those needed to convert hydrogen into a metallic liquid, which triggers dissociation and subsequent helium precipitation,” said Lawrence Livermore National Laboratory scientist Marius Millot.
American Rare Earths core drilling campaign at the company’s 100% owned Halleck Creek Rare Earth project site in Wyoming. Image courtesy of American Rare Earths.
Making rare earths greener
The world's two biggest rare earths companies outside of China are facing challenges turning rock from their mines into the building blocks for magnets used across the global economy, from Apple's iPhone to Tesla's Model 3 to Lockheed Martin's F-35 fighter jet.
The West's push to develop independent supplies of critical minerals took on greater urgency after Beijing imposed export controls last month on the strategic metals gallium and germanium, raising global fears that China could block exports of rare earths or processing technology next.
Companies refining their own rare earths highlight the difficult task the rest of the world faces to break China's stranglehold on the key group of 17 metals needed for the clean energy transition.
“China made a strategic decision decades ago to develop its rare earth processing capability, despite the environmental consequences of the available technology," said Melissa Sanderson, president of American Rare Earths, which is developing several U.S. rare earths projects.
American Rare Earths is working with U.S. scientists at Lawrence Livermore National Laboratory to develop bacteria that could process rare earths. Privately held Locus Mining and Aether Bio also are studying ways to use biosurfactants and nanotechnology, respectively.
