Ultrahigh-strength CoCrFeMnNi high-entropy alloy wire rod with excellent resistance to hydrogen embrittlement

Author’s Accepted Manuscript Ultrahigh-strength CoCrFeMnNi high-entropy alloy wire rod with excellent resistance to hydrogen embrittlement Young Jin Kwon, Jong Woo Won, Sung Hyuk Park, Jeong Hun Lee, Ka Ram Lim, Young Sang Na, Chong Soo Lee www.elsevier.com/locate/msea

PII: DOI: Reference:

S0921-5093(18)30890-6 https://doi.org/10.1016/j.msea.2018.06.086 MSA36641

To appear in: Materials Science & Engineering A Received date: 17 April 2018 Revised date: 21 June 2018 Accepted date: 22 June 2018 Cite this article as: Young Jin Kwon, Jong Woo Won, Sung Hyuk Park, Jeong Hun Lee, Ka Ram Lim, Young Sang Na and Chong Soo Lee, Ultrahigh-strength CoCrFeMnNi high-entropy alloy wire rod with excellent resistance to hydrogen e m b r i t t l e m e n t , Materials Science & Engineering A, https://doi.org/10.1016/j.msea.2018.06.086 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Ultrahigh-strength CoCrFeMnNi high-entropy alloy wire rod with excellent resistance to hydrogen embrittlement

Young Jin Kwon a, Jong Woo Won b,*, Sung Hyuk Park c, Jeong Hun Lee d, Ka Ram Lim b, Young Sang Na b, Chong Soo Lee a,*

a

Graduate Institute of Ferrous Technology, Pohang University of Science and Technology,

Pohang 37673, Republic of Korea. b

Metal Materials Division, Korea Institute of Materials Science, Changwon 51508, Republic

of Korea c

School of Materials Science and Engineering, Kyungpook National University, Daegu

41566, Republic of Korea. d

Advanced Forming Process R&D Group, Korea Institute of Industrial Technology, Ulsan

44413, Republic of Korea [email protected] [email protected] *Corresponding Author: Jong Woo Won Tel.:82-55-280-3355, Fax.:82-55-280-3255 *Corresponding Author. Tel.:82-54-279-9009, Fax.:82-54-279-9099

Abstract An ultrahigh-strength CoCrFeMnNi high-entropy alloy wire rod was produced using cryogenic temperature caliber rolling. Because of the highly increased twinning activity caused by lowering the temperature to 77 K, significant twinning-induced grain refinement occurred; thus, an ultrafine (< 100 nm) grain structure could be achieved in the processed material. The processed material showed a remarkably high tensile strength of ~1.7 GPa, and also had excellent resistance to hydrogen embrittlement (HE), in contrast to the typical tradeoff relationship between these two properties. The exceptionally high resistance to HE was attributed to the combined effects of (1) difficulties in accumulating hydrogen owing to the sluggish hydrogen diffusion caused by the face-centered cubic crystal structure and the severe 1

lattice distortion, (2) the high hydrogen threshold required for HE at the dominant cracking sites of twin boundaries, and (3) absence of martensite transformation.

Keywords: High-entropy alloy; Hydrogen embrittlement; Ultrafine grained microstructure; Twinning

2

1. Introduction High-strength metallic wire rods are used as heavy-duty fasteners in many industries. The material strength required for this application is usually > 1 GPa [1]. High strength is important because it allows for a reduction in the diameter of the bolt while maintaining its load-bearing ability, which in turn reduces the amount of raw materials and the weight of the final products. However, strengthening increases the susceptibility of the material to hydrogen embrittlement (HE), which is attributed to the formation of microstructural defects such as dislocations, grain boundaries (GBs), twin boundaries (TBs), and interfaces [2–4]. These defects provide hydrogen-trapping sites, and thereby, increase the ability of the material to absorb hydrogen; hence, sudden brittle fracture occurs [2–5]. GBs, TBs, and interphases further contribute to HE by providing preferential cracking sites [4]. HE fracture is a serious problem in bolt fasteners because they are exposed to hydrogen atmospheres, and concentration of local stress and hydrogen atoms occurs in the threads as a result of the notch effect [6], thereby degrading the HE resistance [4,6]. This trade-off between strength and HE resistance has hindered the development of high-strength materials for fasteners. Hence, new metallic materials that do not show this trade-off should be developed. High-entropic alloys (HEAs) are receiving much attention from both academic and industrial perspectives because they have highly desirable mechanical properties [7–9]. A recent work [10] reported that the equiatomic CoCrFeMnNi HEA, which has a face-centered cubic (FCC) structure, requires a substantially large number of hydrogen atoms to trigger HE, and therefore, shows better resistance to HE-induced fracture as compared to FCC steels. Although strengthening promotes HE owing to the formation of microstructure defects [1,2,4,5], less attention has been paid to the HE behavior of strengthened FCC HEAs. Previous studies have investigated only the innate resistance to HE by testing a fully recrystallized HEA having very few dislocations and almost no twins [10–12]. In this work, cryogenic temperature caliber rolling (CTCR) was used to produce an ultrahigh-strength equiatomic CoCrFeMnNi HEA wire rod. We conducted the first investigation of HE behavior in a highly strengthened FCC HEA having numerous dislocations and deformation twins and identified the factors governing the HE behavior. We also examined the feasibility of using the CoCrFeMnNi HEA in bolt fasteners by comparing its tensile properties and HE resistance to those of conventional high-strength steels. 3

2. Materials and methods Equiatomic CoCrFeMnNi HEA ingots were prepared via vacuum induction melting using high-purity (> 99.9%) alloying elements. The equiatomic chemical composition was confirmed by energy-dispersive X-ray spectroscopy performed on a scanning electron microscope

(SEM,

model:

7001F,

JEOL);

the

representative

composition

was

Co20.09Cr20.16Fe20.09Mn19.77Ni19.89. The ingots were homogenized at 1373 K for 12 h, hotforged at 1273 K to break up their large columnar casting structure, and annealed at 1273 K for 1 h. The resulting specimens were used as the initial HEAs. For use in CTCR, the initial HEA was machined into rods with a diameter of 12.5 mm. The CTCR process was conducted at 77 K, using eleven circular holes with progressively smaller diameters. This process reduced the diameter of the rod from 12.5 mm to 7.5 mm by imposing a total area reduction of 64%. To retain cryogenic temperature during the CTCR, the sample was immersed in liquid nitrogen just before every rolling pass; hereafter, the processed HEA is referred to as CTCRed HEA. The microstructures were examined by electron backscatter diffraction (EBSD, model: NordlysNano, Oxford) at an acceleration voltage of 15 kV and a step size of 50 nm, and by transmission electron microscopy (TEM, model: JEM 2100F, JEOL) at an acceleration voltage of 200 kV. The crystal structures of the samples were determined by X-ray diffraction (XRD) sing a MXP21VAHF diffractometer with CuK radiation (model: D/Max-2500VL/PC, RIGAKU). The tensile properties were evaluated at room temperature and a strain rate of 10−3 s−1. The tensile specimens had a gauge length of 10 mm and diameter of 2.5 mm (ASTM-E8), and were machined from the core of the produced rods. A slow strain rate test (SSRT) was performed at a constant stroke speed of 0.005 mmmin−1 and room temperature, and a notched sample was used to simulate the thread, which is vulnerable to HE in real bolts [6]. A 60° notch with a root radius of 0.1 mm was machined at the center of the cylindrical sample. The notched samples had a diameter of 6 mm, but the notch depth was 1 mm; hence, their effective cross-sectional diameter was 4 mm (Fig. 1). The notched samples were hydrogen-charged by electrochemical charging before the SSRT. Various hydrogen-charging conditions (Table 1) were used to introduce different 4

amounts of hydrogen into the notched samples. After the SSRT, the hydrogen content ([H]) of the sample heated to 773 K at 100 Kh−1 was measured by thermal desorption spectroscopy (TDS) using a quadrupole mass spectrometer (model: HTDS-002N, R-DEC). The SSRT fracture surface was also examined by using an SEM at an acceleration voltage of 15 kV.

Fig. 1. Schematic of the notched sample used in the SSRT.

Table 1. Hydrogen-charging conditions for the CTCRed HEA and the corresponding hydrogen contents measured using TDS analysis.

Solution

Temperature (K)

Current density (A∙m-2)

Time (h)

Hydrogen contents (ppm)

-

-

-

-

0.102*

0.1 M NaOH

296

10

48

2.79

0.1 M NaOH

296

70

48

3.69

3% NaCl + 0.3% NH4SCN

340

10

21

4.87

3% NaCl + 0.3% NH4SCN

340

10

24

5.41

3% NaCl + 0.3% NH4SCN

360

10

24

7.83

*Value for the non-charged sample 5

For comparison, a conventional tempered-martensitic steel for fasteners was prepared, and its tensile properties and HE resistance were evaluated in the same manner as those of the CTCRed HEA. The tempered-martensitic steel had a typical martensitic microstructure with an ´ single phase (Fig. S1 in supplementary material). The evaluation results for the CTCRed HEA were also compared with those for a pearlitic steel, which is another conventional high-strength steel for fasteners. The data for the pearlitic steel were obtained from [2].

3. Results and discussion 3.1. Microstructure The EBSD band contrast map showed that the initial HEA had a fully recrystallized equiaxed grain structure with coarse annealing twins (Fig. 2a), and an average grain size of 50 m. The XRD patterns showed only FCC peaks (Fig. 2b), indicating that the initial HEA had an entirely FCC crystal structure. Upon CTCR, the initial microstructure was changed because of the introduction of numerous deformation twins in the grain matrix. High-density black lines were visible in the band contrast map (Fig. 3a), indicating that abundant deformation twins were formed in the CTCRed HEA. These twins intersected with each other (Fig. 3a). EBSD kernel average misorientation (KAM) analysis was performed on the same region (Fig. 3b). In materials with very few dislocations, such as the initial HEA [13], the KAM values were close to zero; however, in the CTCRed HEA, the KAM values increased (average KAM value = 1.1) (Fig. 3b), implying that many dislocations were present [14] in addition to deformation twins. The microstructure was further analyzed by TEM, and the twins were found to be very thin (520 nm) (Fig. 3c). Diffraction patterns (Fig. 3d) revealed that the intersection of twins was caused by the operation of different twin variants. As caliber rolling imposes stress in all directions along the circumference of the sample, the twelve possible twinning variants in the FCC crystal structure can operate [15,16].

6

Fig. 2. (a) EBSD band contrast map of the initial HEA and (b) its XRD patterns. In (b), XRD patterns of the CTCRed HEA and the CTCRed HEA subsequently subjected to SSRT after hydrogen charging were also presented.

7

Fig. 3. (a) EBSD band contrast map of the CTCRed HEA and (b) KAM map of the same area. (c) TEM bright-field image of the CTCRed HEA and (d) corresponding [011] diffraction patterns.

3.2. Grain refinement mechanism Deformation twinning divides the grain matrix by introducing TBs. Because of their high boundary angle of 60°, TBs behave similar to GBs and thus refine the microstructure [15,16]. However, twinning was limited in the present HEA because of its intrinsically low activity at room temperature [7–9]. In this work, the processing temperature of CTCR was ~77 K, and therefore, the twinning activity in the present HEA was increased. Thus, numerous twins were introduced in the CTCRed HEA (Fig. 3a), which is not possible when employing room-temperature processes [7-9,16]. Notably, secondary twins were formed within a very narrow matrix bounded by primary twins even though twinning activity significantly decreases as the matrix becomes narrow (Fig. 3c). These twins further divided the matrix into much smaller segments. Moreover, the intersected twin morphology, which was caused by multi-variant twinning, led to more effective refinement of the grain matrix as compared to that observed in the case of parallel twins [15,16]. The average size of the substructure bounded by twins was found to be ~94 nm by using the intersected line method, indicating that the microstructure was refined to an ultrafine grain (UFG) level. Accordingly, the increased twinning activity and its multi-variant operation via CTCR yielded a UFG microstructure in the CTCRed HEA.

3.3. Tensile properties The tensile stress–strain curves (Fig. 4a) demonstrate that the tensile strength of the CTCRed HEA was remarkably higher than that of the initial HEA. The yield stress (YS) of CTCRed HEA was ~1.54 GPa, which was ~4.7 times higher than that of the initial HEA (0.328 GPa). The ultimate tensile stress (UTS) was ~1.71 GPa, which was ~2.3 times higher than that of the initial HEA (0.745 GPa). Considering that the CTCRed HEA had numerous dislocations (Fig. 3b) and a UFG structure (Fig. 3c), this noticeable improvement in strength is attributed to the combined effect of dislocation accumulation strengthening and Hall–Petch 8

strengthening. The CTCRed HEA showed a reasonable fracture elongation (EL) of ~10%, although it was significantly strengthened. The tensile properties of the CTCRed HEA were compared with those of the tempered-martensitic steel (evaluated in this work) and the pearlitic steel [2]. Both YS and UTS were higher in the CTCRed HEA than in the temperedmartensitic steel (YS = 1.37 GPa and UTS = 1.47 GPa) (Fig. 4a), although these two materials had a similar EL of 10%. The YS, UTS, and EL were higher in the CTCRed HEA than in the pearlitic steel (Fig. 4a, YS = 1.45 GPa, UTS = 1.63 GPa, and EL = 8% [2]), although the differences between these values were not significant. These comparisons show that the CoCrFeMnNi HEA has better tensile properties than do the high-strength steels used for fasteners.

Fig. 4. (a) Stress vs. strain curves of the initial HEA, the CTCRed HEA, and the temperedmartensitic steel. (b) SSRT notch fracture stress with diffusible-hydrogen content for the CTCRed HEA and the tempered-martensitic steel. Curves for a 1.6 GP-grade pearlitic steel [2] are also presented for comparison.

3.4. HE resistance The HE behavior of the CTCRed HEA was examined by SSRT using notched samples after charging with various amounts of hydrogen; the measured fracture stresses were plotted as a function of [H] (Fig. 4b). For comparison, the tempered-martensitic steel was evaluated in the same manner; its fracture stress rapidly decreased when [H] was increased to only 1.8 ppm (85% reduction as compared with that in the non-charged sample). In contrast, the CTCRed HEA showed almost no reduction in the fracture stress with the increasing [H]. The 9

fracture stress decreased by only 6.6% even though [H] was increased significantly from ~0 ppm to ~7.8 ppm. We also compared the results of HE evaluation with the corresponding results for a pearlitic steel [2] with a similar tensile strength level (1.63 GPa, shown in Fig. 4a). Pearlitic steels are known to have high resistance to HE because their ferrite–cementite interphases act as non-diffusible hydrogen-trapping sites [2]. The CTCRed HEA showed much better resistance to HE than did the pearlitic steel.

3.5. Excellent HE resistance of CTCRed HEA 3.5.1. Fracture surface The SSRT fracture surfaces were examined to determine reasons for the extraordinarily high resistance to HE in the highly strengthened HEA (Fig. 5). Under the most severe hydrogen-charging condition ([H] = 8 ppm, Table 1), a quasi-cleavage layer appeared near the sample surface, but it was only ~150 µm thick (Fig. 5a-c). Some dimples were also formed in the quasi-cleavage region (Fig. 5c, yellow arrow). The rest of the interior showed perfectly ductile dimples (Fig. 5a, b, d). The thickness of the quasi-cleavage layer decreased as [H] was decreased from 8 ppm, and was only ~30 µm thick in the sample with [H] = 4 ppm (Fig. 5e, f). These observations indicated that HE occurred only at the sample surface, and hence, the fracture stress did not decrease notably. Thus, HE-accelerated fracture was hindered in the CTCRed HEA.

10

Fig. 5. (a-d) SSRT fracture surfaces of a CTCRed HEA sample severely hydrogen-charged at -2

10 A·m and 360 K for 24 h in a 3% NaCl + 0.3% NH4SCN aqueous solution: (a) whole fracture surface; (b) region A – near surface; (c) quasi-cleavage in region A; (d) region B – inner area. (e, f) SSRT fracture surfaces of a CTCRed HEA sample moderately hydrogencharged at 70 A·m-2 and room temperature for 48 h in a 0.1 M NaOH aqueous solution: (e) whole fracture surface; (f) region C – near surface.

The limited HE is related to the hydrogen diffusivity in the material [4,5,14,17]. A low diffusivity of hydrogen hinders its migration; hence, hydrogen cannot penetrate the material deeply during charging. The present HEA has an FCC crystal structure, and hence, the hydrogen diffusivity is very low; the diffusivity of hydrogen in FCC metals is only ~10-6 of that in body-centered cubic (BCC) metals [18,19]. Moreover, severe lattice distortion of the HEAs further obstructs hydrogen diffusion by constricting the hydrogen diffusion pathways [20]. Dislocation tangles were seen in the CTCRed HEA (Fig. 3c), which are known to impede hydrogen diffusion [21,22]. Therefore, hydrogen accumulated intensively near the surface, and thus, HE occurred only at this location. We confirmed this limited hydrogen penetration by comparing the TDS analysis results (Fig. 6a) of two samples: one immediately after charging with hydrogen, and the other after charging with hydrogen and subsequent removal of material up to a depth of 1 mm from the surface. The hydrogen peak was almost absent in the sample from which the surface material had been removed (Fig. 6a). This observation demonstrated that hydrogen was accumulated primarily at the sample surface.

11

Fig. 6. Hydrogen desorption rate curves of the hydrogen-charged sample with a diameter of 6 mm (black line): (a) the CTCRed HEA and (b) the tempered-martensitic steel. The curves of the sample after hydrogen charging and removal of surface material up to a depth of 1 mm are also shown (red line). Here, all the samples were hydrogen-charged at 10 A·m-2 and room temperature for 24 h in a 3% NaCl + 0.3% NH4SCN aqueous solution. Moreover, the quasi-cleavage region was slightly resistant to HE. The thickness of the quasi-cleavage layer in the most severely hydrogen-charged sample was ~150 µm (Fig. 5b), which accounted for ~14.4% of the total cross-sectional area of a notched sample with a diameter of 4 mm. If the quasi-cleavage surface layer were completely brittle owing to HE, it would have fractured prematurely during the SSRT. In this case, as the effective crosssectional area was reduced by 14.4%, the fracture stress (= [external load]/[effective crosssectional area]) was expected to decrease by 14.4% relative to that in the quasi-cleavage region, but fracture stress decreased by only 6.6% (Fig. 4b). The presence of some ductile dimples in the quasi-cleavage region (Fig. 5c) supported the hypothesis that HE is suppressed in the quasi-cleavage region. This result is remarkable, considering that the amount of hydrogen per unit area would be extremely high owing to the restriction of hydrogen absorption to the sample surface.

3.5.2. Hydrogen-detrapping behavior HE occurs according to the following processes. Hydrogen-trapping sites such as the lattice, dislocations, and GBs absorb hydrogen atoms during hydrogen charging. When the material is subjected to a load, these hydrogen atoms detach from the trapping sites to move and concentrate locally. HE is initiated when the local [H] exceeds a critical level at which the fracture mode changes from ductile to brittle. Therefore, an increase in the number of trapping sites causes an increase in the hydrogen absorption capacity, which favors HE. However, when hydrogen binds strongly to the trapping sites, it cannot be detrapped easily and is unlikely to concentrate (i.e., non-diffusible hydrogen-detrapping behavior is exhibited) [2,23]. In this case, HE can be suppressed even though [H] is high in the material. HE is accelerated due to the increased [H] when the hydrogen atoms are easily released from the trapping sites (i.e., diffusible hydrogen-detrapping behavior is exhibited). 12

The hydrogen-detrapping pattern in the CTCRed HEA was examined by observing the TDS results (Fig. 6a). A peak was observed around 160 °C, presumably owing to the release of hydrogen from the diffusible-trapping sites [23,24], indicating that the CTCRed HEA trapped hydrogen atoms mostly in a diffusible manner. The large number of dislocations in the CTCRed HEA, which are diffusible hydrogen-trapping sites [4,23], might be responsible for the diffusible hydrogen-detrapping behavior. Although TBs can act as non-diffusible trapping sites due to their high binding energy with hydrogen [23], we did not observe any TDS peak for non-diffusible hydrogen detrapping in the range 200–300 °C [23,24]. In the CTCRed HEA, the KAM angles were high along the TBs; i.e., the TBs accommodated a significant number of dislocations. This result suggested that the hydrogen trapping associated with the TBs might be caused not by the TBs themselves but by the dislocations at the TBs [25]. Based on this analysis, we conclude that hydrogen atoms were easily detrapped from the trapping sites during SSRT, and that the mode of hydrogen detrapping does not contribute to the suppression of HE in the quasi-cleavage region.

3.5.3. Hydrogen mobility The mobility of detrapped hydrogen atoms is important for their local concentration. If detrapped hydrogen atoms can easily move inside the material under stress, local [H] increases rapidly, thus promoting HE. The tempered-martensitic steel was vulnerable to HE (Fig. 4b) because the hydrogen atoms could easily migrate due to their fast diffusion in the BCC lattice [2,15]. In contrast, the CTCRed HEA had sluggish hydrogen diffusion, and hence, hydrogen movement by lattice diffusion would be difficult. Hydrogen transportation also occurs by the mobile dislocations of slip and twinning under loading [4,26]. These movements significantly contribute to local [H] by carrying hydrogen atoms to the GBs or TBs [4]. In low-stacking-fault-energy metals such as the present HEA, the strong planarity of the mobile dislocations causes significant accumulation of hydrogen atoms at the boundaries by hindering cross slip [4]. However, the slip planarity decreases with decreasing grain size [27], and hence, this method of hydrogen accumulation is likely to be ineffective in a CTCRed HEA with a UFG structure (Fig. 3c). Moreover, twinning probably did not occur in the CTCRed HEA during the SSRT, because its activation is expected to be inhibited by the significantly reduced grain size [27,28] and the nature of the HEA [7-9]. 13

Accordingly, the difficulty in accumulating hydrogen atoms is most likely to be the main cause of the suppressed HE in the CTCRed HEA.

3.5.4. HE-induced cracking site Generally, grain refinement occurs as the number of GBs increases. Such a grain-refined material tends to fracture in intergranular mode because the GBs serve as preferential sites for cracking [4,29]. However, in the CTCRed HEA, grain refinement was caused by TBs, and this process differs from general grain refinement based on an increase in GB. Thus, due to their relatively large number, TBs appear to have been more important as HE-induced cracking sites than GBs. Step-like ridges observed in the quasi-cleavage region (Fig. 5c) are typical evidence of TB cracking [4]. Unlike GBs, TBs are resistant to HE-induced cracking because of their coincident bonding (i.e., Σ3 coherent TBs) [30,31], suggesting that the CTCRed HEA has a high [H] threshold for HE. In particular, this factor explains why HE was suppressed in the quasi-cleavage region even though hydrogen charging had deposited a large amount of hydrogen atoms at the TBs.

3.6. Comparison with conventional steels Unlike the CTCRed HEA, tempered-martensitic steel induces ductile–brittle transition in the fracture mode with only a small [H] [2]. Moreover, the TDS curve did not differ substantially before and after the surface of the hydrogen-charged sample was removed (Fig. 6b), indicating that hydrogen atoms were distributed throughout the hydrogen-charged sample by easily diffusing into the interior of the sample and further confirming that the tempered-martensitic steel showed fast diffusion of hydrogen atoms. Thus, severe HE occurred on the entire fracture surface and therefore, significantly reduced the SSRT fracture stress in the hydrogen-charged tempered-martensitic steel (Fig. 4b). This work revealed that an HEA wire rod produced via CTCR had both ultrahigh strength and excellent HE resistance, in contrast to the usual trade-off relationship between these two properties. This finding is particularly remarkable because conventional (temperedmartensitic and pearlitic) steels for wire rods cannot achieve high strength without a reduction in HE resistance [2]. The HEA wire rod has significant potential for load-intensive 14

applications, such as fasteners, which require both high strength and excellent HE resistance. Notably, CTCR is suitable for FCC HEAs rather than for FCC steels because of two reasons: (1) as the HEA has exceptionally high FCC stability [7], martensite transformation does not occur during cryogenic deformation, which significantly degrades the ductility and HE resistance [4]; FCC steels undergo martensite transformation during low-temperature deformation [32]. XRD patterns revealed no peaks related to martensite phases in the case of the CTCRed HEA and the CTCRed HEA subjected to SSRT after hydrogen charging (Fig. 1b). (2) Because of its excellent ductility at cryogenic temperature, the HEA can withstand extreme plastic deformation without cracking.

4. Conclusions An ultrahigh-strength CoCrFeMnNi HEA wire rod was produced using CTCR. The highly increased twinning activity and its multi-variant operation caused by CTCR led to significant twinning-induced grain refinement, and hence, the CTCRed HEA had a UFG structure. The CTCRed HEA had a remarkably high strength of ~1.7 GPa and excellent resistance to HE, and therefore, it overcame the typical trade-off between these two properties. The exceptionally high resistance to HE was attributed to the combined effects of (1) the difficulty in accumulating hydrogen atoms owing to the sluggish hydrogen diffusion caused by the FCC crystal structure and severe lattice distortion; (2) high [H] threshold required for HE at the dominant cracking sites of the TBs, and (3) absence of martensite transformation. The present results demonstrate that this CoCrFeMnNi HEA is appropriate for use in bolt fasteners that require both high strength and excellent HE resistance.

Acknowledgements This work received financial support from Fundamental Research Program(PNK5610) of the Korean Institute of Materials Science (KIMS), and from the Future Material Discovery Project of the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT and Future Planning (MSIP) of Korea (NRF-2016M3D1A1023534).

15

References [1] Y. Namimura, M. Fujita, N. Ibaraki, Y. Oki, Improved delayed fracture resistance in 1600MPa class ultra-high-strength bolt steel, KOBE Steel Eng. Rep. 54 (2004) 16–20. [2] J.S. Kim, Y.H. Lee, D.L. Lee, K.T. Park, C.S. Lee, Microstructural influences on hydrogen delayed fracture of high strength steels, Mater. Sci. Eng. A. 505 (2009) 105– 110. [3] M. Nagumo, K. Takai, N. Okuda, Nature of hydrogen trapping sites in steels induced by plastic deformation, J. Alloys Compd. 293–295 (1999) 310–316. [4] M. Koyama, E. Akiyama, Y. Lee, D. Raabe, K. Tsuzaki, Overview of hydrogen embrittlement in high-Mn steels, Int. J. Hydrogen Energy. 42 (2017) 12706–12723. [5] H. Iwaoka, M. Arita, Z. Horita, Hydrogen diffusion in ultrafine-grained iron with the body-centered cubic crystal structure, Philos. Mag. Lett. 839 (2017) 1–11. [6] M. Wang, E. Akiyama, K. Tsuzaki, Effect of hydrogen and stress concentration on the notch tensile strength of AISI 4135 steel, Mater. Sci. Eng. A. 398 (2005) 37–46. [7] B. Gludovatz, A. Hohenwarter, D. Catoor, E.H. Chang, E.P. George, R.O. Ritchie, A fracture-resistant high-entropy alloy for cryogenic applications, Science 345 (2014) 1153. [8] G. Laplanche, A. Kostka, O.M. Horst, G. Eggeler, E.P. George, Microstructure evolution and critical stress for twinning in the CrMnFeCoNi high-entropy alloy, Acta Mater. 118 (2016) 152–163. [9] F. Otto, A. Dlouhý, C. Somsen, H. Bei, G. Eggeler, E.P. George, The influences of temperature and microstructure on the tensile properties of a CoCrFeMnNi high-entropy alloy, Acta Mater. 61 (2013) 5743–5755. [10] Y. Zhao, D.H. Lee, M.Y. Seok, J.A. Lee, M.P. Phaniraj, J.Y. Suh, H.Y. Ha, J.Y. Kim, U. Ramamurty, J.I. Jang, Resistance of CoCrFeMnNi high-entropy alloy to gaseous hydrogen embrittlement, Scr. Mater. 135 (2017) 54–58. [11] Y. Zhao, D.H. Lee, J.A. Lee, W.J. Kim, H.N. Nam, U. Ramamurty, J.Y. Suh, J.I. Jang, Hydrogen-induced nanohardness variations in a CoCrFeMnNi high-entropy alloy, Int. J. Hydrogen Energy 42 (2017) 12015–12021. [12] K.E. Nygren, K.M. Bertsch, S. Wang, H. Bei, A. Nagao, I.M. Robertson, Hydrogen 16

embrittlement in compositionally complex FeNiCoCrMn FCC solid solution alloy, Curr. Opin. Solid State Mater. Sci. 22 (2018) 1-7. [13] M. Calcagnotto, D. Ponge, E. Demir, D. Raabe, Orientation gradients and geometrically necessary dislocations in ultrafine grained dual-phase steels studied by 2D and 3D EBSD, Mater. Sci. Eng. A. 527 (2010) 2738–2746. [14] Y.J. Kim, S.H. Kim, J.U. Lee, B.S. You, S.H. Park, Evolution of high-cycle fatigue behavior of extruded AZ91 alloy by artificial cooling during extrusion, Mater. Sci. Eng. A. 707 (2017) 620-628. [15] Y.S. Chun, J. Lee, C.M. Bae, K.T. Park, C.S. Lee, Caliber-rolled TWIP steel for highstrength wire rods with enhanced hydrogen-delayed fracture resistance, Scr. Mater. 67 (2012) 681–684. [16] J.W. Won, S. Lee, S.H. Park, M.J. Kang, K.R. Lim, C.H. Park, Y.S. Na, Ultrafinegrained CoCrFeMnNi high-entropy alloy produced by cryogenic multi-pass caliber rolling, J. Alloys Compd. 742 (2018) 290-295. [17] P. Kumar, R. Balasubramaniarn, Determination of hydrogen diffusivity in austenitic stainless steels by subscale microhardness profiling, J. Alloys Compd. 255 (1997) 130– 134. [18] J.A. Ronevich, S.K. Kim, J.G. Speer, D.K. Matlock, Hydrogen effects on cathodically charged twinning-induced plasticity steel, Scr. Mater. 66 (2012) 956-959. [19] D.K. Han, Y.M. Kim, H.N. Han, H.K.D.H. Bhadeshia, D.-W. Suh, Hydrogen and aluminium in high-manganese twinning-induced plasticity steel, Scr. Mater. 80 (2014) 912. [20] N. Stepanov, M. Tikhonovsky, N. Yurchenko, D. Zyabkin, M. Klimova, S. Zherebtsov, A. Efimov, G. Salishchev, Effect of cryo-deformation on structure and properties of CoCrFeNiMn high-entropy alloy, Intermetallics 59 (2015) 8–17. [21] P. Tsong-Pyng, C.J. Altstetter, Effects of deformation on hydrogen permeation in austenitic stainless steels, Acta Metallurgica 34 (1986) 1771-1781. [22] R.K. Dayal, H.J. Grabke, Dependence of the hydrogen permeation in stainless steel on carbon content, heat treatment and cold work, Steel Research 71 (2000) 255-260. 17

[23] K.H. So, J.S. Kim, Y.S. Chun, K.-T. Park, Y.-K. Lee, C.S. Lee, Hydrogen Delayed Fracture Properties and Internal Hydrogen Behavior of a Fe–18Mn–1.5Al–0.6C TWIP Steel, ISIJ International 49 (2009) 1952-1959. [24] I.J. Park, J.G. Jung, S.Y. Jo, S.M. Lee, Y.K. Lee, The Effect of Pre-Strain on the Resistance to Hydrogen Embrittlement in 316L Austenitic Stainless Steel, Mater. Trans. 55 (2014) 964–970. [25] J.H. Ryu, S.K. Kim, C.S. Lee, D.W. Suh, H.K.D.H. Bhadeshia, Effect of aluminium on hydrogen-induced fracture behaviour in austenitic Fe–Mn–C steel, Proc. Royal Soc. A 469 (2013) article number 20120458. [26] J. Tien, A.W. Thompson, I.M. Bernstein, R.J. Richards, Hydrogen transport by dislocations, Metall. Mater. Trans. A 7 (1976) 821-829. [27] S. Kang, J.-G. Jung, M. Kang, W. Woo, Y.-K. Lee, The effects of grain size on yielding, strain hardening, and mechanical twinning in Fe–18Mn–0.6C–1.5Al twinning-induced plasticity steel, Mater. Sci. Eng. A 652 (2016) 212-220. [28] K.M. Rahman, V.A. Vorontsov, D. Dye, The effect of grain size on the twin initiation stress in a TWIP steel, Acta Mater. 89 (2015) 247–257. [29] Y.J. Kwon, T. Lee, J. Lee, Y.S. Chun, C.S. Lee, Role of Cu on hydrogen embrittlement behavior in Fe–Mn–C–Cu TWIP steel, Int. J. Hydrogen Energy 40 (2015) 7409-7419. [30] F. Otto, E. J. Payton, J. Frenzel, G. Eggeler, The effectiveness of coincidence site lattice criteria in predicting creep cavitation resistance, J. Mater. Sci. 47 (2012) 2915-1927. [31] S. Bechtle, M. Kumar, B.P. Somerday, M.E. Launey, R.O. Ritchie, Grain-boundary engineering markedly reduces susceptibility to intergranular hydrogen embrittlement in metallic materials, Acta Mater. 57 (2009) 4148-4157 [32] S.W. Song, J.H. Lee, T. Lee, C.S. Lee, Effect of the amount and temperature of prestrain on tensile and low-cycle fatigue properties of Fe-17Mn-0.5C TRIP/TWIP steel, Mater. Sci. Eng. A 696 (2017) 493-502.

18

Table and figure captions Table 1. Hydrogen-charging conditions for the CTCRed HEA and the corresponding hydrogen contents measured using TDS analysis.

Fig. 1. Schematic of the notched sample used in the SSRT. Fig. 2. (a) EBSD band contrast map of the initial HEA and (b) its XRD patterns. In (b), XRD patterns of the CTCRed HEA and the CTCRed HEA subsequently subjected to SSRT after hydrogen charging were also presented. Fig. 3. (a) EBSD band contrast map of the CTCRed HEA and (b) KAM map of the same area. (c) TEM bright-field image of the CTCRed HEA and (d) corresponding [011] diffraction patterns. Fig. 4. (a) Stress vs. strain curves of the initial HEA, the CTCRed HEA, and the temperedmartensitic steel. (b) SSRT notch fracture stress with diffusible-hydrogen content for the CTCRed HEA and the tempered-martensitic steel. Curves for a 1.6 GP-grade pearlitic steel [2] are also presented for comparison. Fig. 5. (a-d) SSRT fracture surfaces of a CTCRed HEA sample severely hydrogen-charged at -2

10 A·m and 360 K for 24 h in a 3% NaCl + 0.3% NH4SCN aqueous solution: (a) whole fracture surface; (b) region A – near surface; (c) quasi-cleavage in region A; (d) region B – inner area. (e, f) SSRT fracture surfaces of a CTCRed HEA sample moderately hydrogen- charged at 70 A·m-2 and room temperature for 48 h in a 0.1 M NaOH aqueous solution: (e) whole fracture surface; (f) region C – near surface. Fig. 6. Hydrogen desorption rate curves of the hydrogen-charged sample with a diameter of 6 mm (black line): (a) the CTCRed HEA and (b) the tempered-martensitic steel. The curves of the sample after hydrogen charging and removal of surface material up to a depth of 1 mm are also shown (red line). Here, all the samples were hydrogen-charged at 10 A·m-2 and room temperature for 24 h in a 3% NaCl + 0.3% NH4SCN aqueous solution.

19