Direct Imaging of Hydrogen-Driven Dislocation and Strain Field Evolution in a Stainless Steel Grain

TL;DR

This work uses in situ Bragg coherent X-ray diffraction imaging (BCDI) to watch, in a bulk grain of austenitic 316 stainless steel, how hydrogen charging alters dislocation behavior and strain fields. By tracking a single dislocation, the study reveals hydrogen-enhanced mobility and subsequent climb driven by osmotic forces, alongside direct measurements of hydrogen-induced elastic shielding of the dislocation strain field. The results validate theoretical predictions of hydrogen-dislocation interactions, quantify nanoscale phenomena, and provide inputs for multiscale models aimed at predicting bulk material response and guiding the design of hydrogen embrittlement-resistant alloys. The approach demonstrates the power of combining BCDI with dislocation dynamics modeling to unravel complex HE mechanisms under realistic, bulk conditions, with implications for future high-coherence, Bragg ptychography studies at next-generation light sources.

Abstract

Hydrogen embrittlement (HE) poses a significant challenge to the durability of materials used in hydrogen production and utilization. Disentangling the competing nanoscale mechanisms driving HE often relies on simulations and electron-transparent sample techniques, limiting experimental insights into hydrogen-induced dislocation behavior in bulk materials. This study employs in situ Bragg coherent X-ray diffraction imaging to track three-dimensional dislocation and strain field evolution during hydrogen charging in a bulk grain of austenitic 316 stainless steel. Tracking a single dislocation reveals hydrogen-enhanced mobility and relaxation, consistent with dislocation dynamics simulations. Subsequent observations reveal dislocation unpinning and climb processes, likely driven by osmotic forces. Additionally, nanoscale strain analysis around the dislocation core directly measures hydrogen-induced elastic shielding. These findings experimentally validate theoretical predictions and offer mechanistic insights into hydrogen-driven dislocation behavior. The quantified nanoscale phenomena serve as critical inputs for multiscale modeling frameworks to predict bulk material responses and accelerate the development of HE-resistant alloys.

Figures (5)

• Figure 1: In situ hydrogen charging BCDI setup and evolution of the Bragg peak. a) The microcrystalline 316 SS disk is mounted in the electrochemical flow cell (see Supporting Information and Figure \ref{['supp-fig:flowcell']}, Supporting Information). An incoming coherent X-ray beam (red) illuminates a grain within the SS disk, and a slice through the reflected $111$ Bragg peak is captured on the detector. During in situ charging, a bias is applied to the SS (working electrode). A Pt wire coil (counter electrode) sits inside the flow cell moat. The blue arrows indicate the flow of the hydrogen charging solution. b) The lattice parameter and inferred hydrogen concentration before and during the experiment, with insets showing slices through the center of the Bragg peak corresponding to different times. The shaded areas correspond to the uncertainty associated with the Bragg peak position. c) 3D Pearson correlation of the Bragg peaks. Hydrogen charging starts at 0 h.
• Figure 2: Evolution of the heterogeneous strain field on the grain surface before and during hydrogen charging. a) Different views of the grain surface, colored by $\varepsilon_{111,\ \mathrm{surf.}}$, the strain relative to the average lattice parameter indicated at the bottom. See Figure \ref{['supp-fig:all isosurfaces']}, Supporting Information and Video S1 for reconstructions from all time points. b) Histogram distribution of the surface strain for different time points during the hydrogen charging history. c) Evolution of average surface strain. The shaded region corresponds to one standard deviation.
• Figure 3: Evolution of the dislocations before and during hydrogen charging. a,b,c) Columns represent a different orthogonal view of the grain and the dislocations. The top row shows the initial morphology of the grain rendered as a grey isosurface. The subsequent rows have translucent renderings of the same grain morphology (largely unchanged, see Figure \ref{['main3:fig:Fig2']}) along with the dislocations at different states, with glide and climb events indicated. The dislocation dynamics simulation of glide is indicated in green. Dislocations are colored according to the time of observation. The black arrow indicates the Burgers vector. See Video S2 (Supporting Information) for the dislocations at each measurement time.
• Figure 4: Dislocation unpinning and climb. A translucent isosurface of the grain showing only the large dislocation before (1, at 3.2 h) and after the climb event (2, at 3.9 h). The dislocation is colored between two nodes based on the $b_e$ (Equation \ref{['main3:eq:b_e']}), where $b_e/|\mathbf{b}| = 1$ represents a pure edge dislocation and $b_e/|\mathbf{b}| = 0$ represents a pure screw dislocation. The black arrow indicates the Burgers vector direction. The magenta arrow, which is nearly perpendicular to the Burgers vector direction, indicates the direction of climb. a) Orthogonal views based on sample coordinates. b) Crystallographic views based on the dislocation loop plane, with the dislocation initially lying on the $\left(\bar{1}11\right)$ plane, and later one end unpins and climbs onto the $\left(11\bar{1}\right)$ plane.
• Figure 5: Evolution of the internal strain field surrounding the large dislocation. Three time points are presented as rows in (a--d). a) Translucent morphologies of the grain with dislocations. Slices through $\varepsilon_{111}$ are in-plane to the Burgers vector and the normal to the $\left(\bar{1}11\right)$ plane. The average lattice parameter is listed for each time. b) A $\varepsilon_{111}$ slice capturing a section of the large dislocation with pure edge character. See Video S3 (Supporting Information) for $\varepsilon_{111}$ slices through the entire grain. c) Theoretical elastic model, $\varepsilon_{111,\ \mathrm{model}}$, of the large dislocation devoid of hydrogen (Experimental Section). d) Circular line profiles, drawn at a 30 nm radius from each dislocation core, compared to the model. The shaded region is the standard deviation of the experimental values in the neighboring 26 pixels. Line profiles start at the Burgers vector (horizontal) and run anticlockwise. e) Circular line profiles drawn for each time point of the experiment, plotted as a surface indicating $\varepsilon_{111}$ values. f) Top-down view of (e). g) Maximum and minimum values of the line profiles, averaged over a range of $\pi/4$. The shaded region corresponds to one standard deviation of each value. To better highlight the differences between the model and experimental data at positive charging times, the model's maximum and minimum values were slightly offset to align with the experimental data at negative times.