Hydrogen enhances strength and ductility of an equiatomic high-entropy alloy

Metals are key materials for modern manufacturing and infrastructures as well as transpot and energy solutions owing to their strength and formability. These properties can severely deteriorate when they contain hydrogen, leading to unpredictable failure, an effect called hydrogen embrittlement. Here we report that hydrogen in an equiatomic CoCrFeMnNi high-entropy alloy (HEA) leads not to catastrophic weakening, but instead increases both, its strength and ductility. While HEAs originally aimed at entropy-driven phase stabilization, hydrogen blending acts opposite as it reduces phase stability. This effect, quantified by the alloy’s stacking fault energy, enables nanotwinning which increases the material’s work-hardening. These results turn a bane into a boon: hydrogen does not generally act as a harmful impurity, but can be utilized for tuning beneficial hardening mechanisms. This opens new pathways for the design of strong, ductile, and hydrogen tolerant materials.


Results and Discussion
Starting microstructure. The equiatomic CoCrFeMnNi HEA was produced by melting and casting in a vacuum induction furnace with high-purity elemental starting materials. The cast alloy was hot-rolled and homogenized followed by water-quenching. Cold-rolling and recrystallization annealing were performed to refine the grain size. Figure 1 shows the typical microstructure and elemental distribution in the CoCrFeMnNi HEA after recrystallization annealing. The material has a grain size in the range of 1-25 μm with random crystallographic texture (Fig. 1a,b). The back-scattered electron (BSE) image (Fig. 1b) and the corresponding energy-dispersive X-ray spectroscopy (EDS) maps ( Fig. 1c) reveal that all elements (Co, Cr, Fe, Mn and Ni) are uniformly distributed. The X-ray diffraction (XRD) pattern in Fig. 1d confirms the f.c.c. phase structure.
Strength and ductility upon hydrogen charging. Figure 2 shows the tensile deformation behavior   Table 1.

Fracture behavior upon hydrogen charging.
To understand the fracture behavior of the hydrogen alloyed CoCrFeMnNi HEA, we studied the morphology of the fracture surfaces of all samples after in-situ tensile testing. Figure 3a shows a fully ductile fracture, characterized by the growth and coalescence of microvoids in both, edge and center regions of the sample without hydrogen alloying. Some particles, which had acted as typical initiation sites of microvoids, were found inside the voids on the fracture surface. The EDS results ( Supplementary  Fig. 4) show that they are enriched in Mn, Cr, S and Al. The hydrogen alloyed HEA samples also display the primary fracture mode of microvoid coalescence (Fig. 3b,c and d). However, also some intergranular fracture was observed on the edges of the hydrogen alloyed HEA samples.
Since hydrogenation in the present study was conducted by cathodic charging in alkaline solution and the tensile tests were performed under in-situ hydrogen charging conditions, a gradient in hydrogen concentration from the edge to the inner regions of the samples could have been induced according to previous studies on hydrogen embrittlement of other materials 4 . For this reason associated with the charging process, the intergranular fracture feature observed in the edges is attributed to the high hydrogen concentration near the edges of the samples during the tensile testing in alkaline solution. The fracture mode of microvoid coalescence in the inner regions suggests lower hydrogen concentration in these zones. We also observe that the very fine (<1 μm) ductile dimples in the inner regions of the fracture surface in the hydrogen alloyed samples (Figs. 3b,c and d) occur much more frequently compared to the reference alloy without hydrogen (Fig. 3a). This might be related to the more extensively occurring nanotwinning in the hydrogen alloyed HEAs discussed below.
Deformation mechanism upon hydrogen charging. To reveal the micro-mechanisms behind the bulk samples to avoid any thin slice size and related effects which alter the stress and microstructure state of the material.
The equiatomic CoCrFeMnNi HEA has a relatively low intrinsic stacking fault energy (~19 mJ·m 2 ) 38 , which explains the occurrence of mechanical twinning at room temperature without hydrogen charging. The strong increase in twinning density in the hydrogen alloyed HEA samples reveals that hydrogen reduces the stacking fault energy of the equiatomic CoCrFeMnNi HEA. This is consistent with the previous studies on the effect of hydrogen on the stacking fault energy in other f.c.c. structured alloys 39,40 . Moreover, no other phase formed during the in-situ tensile testing with hydrogen charging according to the XRD results shown in Supplementary  Fig. 6. These results also show that the increase in the density of nanotwins that we observe in the CoCrFeMnNi HEA leads to an associated increase in work hardening ( Supplementary Fig. 3), which postpones the formation of geometric instabilities to higher strain values 41 , thereby increasing strength and ductility simultaneously.
The beneficial influence of nanotwins on strain hardening is commonly attributed to the "dynamic Hall-Petch" effect 29 , i.e. the formation of nanotwins leads to continuous grain fragmentation by the introduction of new interfaces which effectively reduces the dislocation mean free path, hence, causing further strengthening 42,43 . Consistent with this, the strength and ductility of the CoCrFeMnNi HEA increased with the increase in the twinning density caused by the enhanced hydrogen concentration. However, during in-situ tensile testing, unfavorable mechanism caused by hydrogen may also become more significant with higher hydrogen concentration. For instance, hydrogen enhanced localized plasticity 16 can occur in materials that are prone to planar slip, such as the here studied CoCrFeMnNi HEA 29 . This unfavorable mechanism was also observed in a gaseous hydrogen charged CoCrFeMnNi HEA 35 . Thus, a competition may exist between the beneficial nanotwining on the one hand and the unfavorable hydrogen embrittlement mechanisms on the other hand during deformation of the hydrogen charged HEA samples. Furthermore, considering the gradient in hydrogen concentration through the sample thickness, the very high concentration of hydrogen at the sample surface may result in initial surface cracks, thus reducing the overall strength and ductility of the bulk sample. The above mechanisms might be responsible for the fact that only small increments in ultimate strength and uniform elongation were observed in the HEA sample when apoproaching the highest diffusible hydrogen concentration (Fig. 2c).

Conclusions
Our findings demonstrate that hydrogen can be utilized for the case of an equiatomic CoCrFeMnNi HEA to tune beneficial strengthening and toughening mechanisms rather than undergoing catastrophic failure due to hydrogen embrittlement. We show that hydrogen alloying with a proper concentration actually jointly increases both strength and ductility. By reducing the stacking fault energy and hence the phase stability of the CoCrFeMnNi HEA, hydrogen can cause a significant increase in the nano-twin density, thereby increasing the alloy's work-hardening capability, and thus both strength and ductility. This breaks the previous preconception on the deleterious hydrogen effects which lasted for more than a century and provides new strategies for the future design of hydrogen tolerant materials with superior mechanical properties.

Methods
Materials preparation. The equiatomic CoCrFeMnNi HEA ingot with dimensions of 25 × 60 × 65 mm 3 was cast in a vacuum induction furnace using pure metals with predetermined compositions (Co 20 Cr 20 Fe 20 Mn 20 Ni 20 , at.%). All the pure metals were cleaned properly and the carbon content was then controlled to be as low as possible to avoid the influence of interstitial carbon on the mechanical properties 44,45 . Samples with dimensions of 10 × 25 × 60 mm 3 machined from the original cast were subsequently hot-rolled at 900 °C to a thickness reduction of 50% (thickness changed from 10 to 5 mm). Homogenization was conducted at 1200 °C for 2 h in Ar atmosphere followed by water-quenching. To refine the grain size, samples were further cold-rolled to a thickness reduction of 60%, and subsequently annealed at a furnace temperature of 900 °C for 3 min in Ar atmosphere followed by water-quenching. Note that the true temperature that the samples actually reached during annealing might be lower than the furnace temperature (900 °C) due to the short annealing time.
Microstructural and elemental characterization. The microstructure of the recrystallized alloy (grain-refined) was analyzed using various methods. X-ray diffraction (XRD) measurements were performed using an X-Ray equipment ISO-DEBYEFLEX 3003 equipped with Co Kα (λ = 1.788965 Å) radiation operated at 40 kV and 30 mA. Electron backscatter diffraction (EBSD) measurements were carried out by a Zeiss-Crossbeam XB 1540 FIB scanning electron microscope (SEM) with a Hikari camera and the TSL OIM data collection software. Electron channeling contrast imaging (ECCI) analyses of the deformation microstructures were performed on a Zeiss-Merlin instrument. The elemental distributions in the recrystallized alloy were investigated using energy-dispersive X-ray spectroscopy (EDS). The fracture morphology was observed by a Zeiss-Merlin instrument.
Hydrogen charging and mechanical characterization. Hydrogen was introduced into the specimens by electrochemical charging with various current densities at ambient temperature (25 °C) in 0.1 M NaOH solution plus 0.05 g·L −1 CH 4 N 2 S at pH = 13. A platinum wire was used as the counter electrode. The samples were pre-charged for 12 h at 15 mA·cm −2 , 72 h at 25 mA·cm −2 , and 240 h at 100 mA·cm −2 , respectively. Then the samples were continuously in-situ charged during the entire tensile testing to avoid the release of hydrogen. The tensile tests were conducted in an Instron tensile machine at the tensile rate of 1 × 10 −4 S −1 . Uniaxial tensile tests were conducted using specimens with thickness of 1.5 mm and gauge length of 10 mm. Three samples for each condition were tested to confirm reproducibility. The hydrogen desorption rates were measured by using a custom-designed UHV-based Thermal Desorption Analysis instrument in conjunction with a Mass Spectrometer detector set up (TDA-MS) from 25 °C to 800 °C, and the corresponding heating rate was 26 °C min −1 . The diffusible hydrogen concentration was determined by measuring cumulative desorbed hydrogen from 25 °C to 600 °C.