materials science

In the world of magnetism, a simple dichotomy long reigned: ferromagnets, where spins point in one direction, and antiferromagnets, where they line up in a checkerboard pattern, canceling each other out. But a recent 'third player' — altermagnets — forced a rewrite of the rules.

Ordinary neural networks learn from examples, like a person memorizing tunes by ear. But networks with physical laws embedded from the start resemble a musician who knows notes and harmony: they can play a new piece from just a few chords. Instead of mindless memorization, they use equations like energy conservation or motion.

In the orchestra of atoms, each instrument has its own score. But sometimes two musicians start playing in unison, and the sound amplifies many times over. In the quantum world, this 'harmony' is called a Feshbach resonance. It allows you to control atomic interactions with jeweler-like precision.

When you pull on a piece of metal, it changes shape and simultaneously becomes stronger. The secret lies in tiny defects in the crystal lattice called dislocations. They are like wrinkles in a rug: moving a wrinkle shifts the rug more easily than dragging the whole rug.

Deep inside every hard drive or flash drive lies a material that can 'remember' an electric field. Such a material is called a ferroelectric. When you apply a voltage, tiny regions inside the crystal (domains, like compass needles) flip. Flip the needle one way, you've written a one; flip it back, a zero. But what happens to the atoms at the moment of that flip?

Steel is not just a piece of iron. Inside it, like in a layered pie, different structures, phases, coexist. In ultra-strong steel for cars, there are three such phases: soft ferrite, strong martensite, and ductile austenite. Medium-manganese steel owes its strength precisely to this trio.

Pour a handful of dry sand: it will spread into a neat cone. Now grab a box of star-shaped pasta: the pile will stand almost vertical. Why? It's all about particle shape. Physicists built cylindrical piles from particles of different geometries — from smooth spheres to 'polypods' with several arms, like a sea star.

A Swiss Army knife is simple: one handle, many blades, each for its own task. Grab what you need and get to work. In materials science, graph neural networks have long resembled a scattered set of tools: MEGNet sat in the TensorFlow drawer, M3GNet in the PyTorch drawer, and CHGNet in a separate box with its own scripts.