Hydrogen embrittlement has been destroying steel structures for over a century: pipelines crack, bolts fracture, pressure vessels fail. The mechanism involves hydrogen atoms — small enough to diffuse between iron atoms — accumulating at dislocations, the line defects that allow metals to deform plastically. The conventional picture is simple: hydrogen makes dislocations move more easily (the HELP mechanism — hydrogen-enhanced localized plasticity) or it reduces the cohesive energy of grain boundaries (the HEDE mechanism — hydrogen-enhanced decohesion). Both mechanisms predict that hydrogen makes failure easier. The question is how.
Manda, Gupta, Kumar, Akhter, Guruprasad, Samajdar, and Panwar (arXiv 2602.22995, February 2026) show that hydrogen's effect on dislocation dynamics is not simply one-directional. Using molecular dynamics coupled with kinetic Monte Carlo across multiple time and length scales, they find that hydrogen simultaneously accelerates one process and retards another.
Hydrogen lowers the energy barrier for kink nucleation on dislocation lines in BCC iron. Kinks are the elementary excitations by which dislocations move in body-centered cubic metals — the dislocation advances by forming a small step (a kink pair) that then migrates along the line. By reducing the nucleation barrier, hydrogen makes it easier to initiate dislocation motion.
But hydrogen increases the resistance to dislocation migration itself. Once the kink forms, the dislocation must drag along its hydrogen atmosphere — the Cottrell cloud of hydrogen atoms that decorates the dislocation core. The cloud creates friction. Moving the dislocation means moving the cloud, and the cloud's mobility is limited by hydrogen diffusion.
The net effect depends on which process is rate-limiting. At low temperatures or high strain rates, kink nucleation limits dislocation motion, and hydrogen accelerates it. At high temperatures or low strain rates, migration limits motion, and hydrogen retards it. The same impurity can enhance or suppress plasticity depending on the conditions.
A linear correlation emerges between hydrogen concentration and internal friction loss — the energy dissipated during cyclic mechanical loading. This makes mechanical spectroscopy a quantitative hydrogen detector: the Snoek-Koster relaxation peak from the hydrogen Cottrell atmosphere provides a direct measurement of hydrogen content at dislocation cores.
Hydrogen helps and hinders. The same atom that starts the motion drags the motion down.