2024 Volume 65 Issue 9 Pages 1049-1054
Formation of stacking fault tetrahedra and diffuse streaks along ⟨111⟩ in the equiatomic Cr-Co-Ni medium-entropy alloy subjected to different heat-treatments and different specimen preparation methods are examined by transmission electron microscopy. Neither stacking faults nor stacking fault tetrahedra are formed simply by quenching from a high temperature due to the high energy barrier for the formation and migration of vacancies. These defects, however, are found to form by Ar+-ion irradiation, as abundant vacancies are continuously introduced during ion-irradiation so as to bypass the high energy barrier for the formation. Diffuse streaks along ⟨111⟩ are usually observed in the SAED pattern with the ⟨110⟩ incidence regardless of heat-treatments and specimen preparation methods, indicating the occurrence of diffuse streaks is nothing to do with the formation of stacking faults and stacking fault tetrahedra.

Since the first observation by Silcox and Hirsch [1], the formation of stacking fault tetrahedra (SFTs) has been repeatedly reported in some face-centered cubic (FCC) metals and alloys subjected to rapid quenching, heavy plastic deformation and irradiation [1–4]. Silcox and Hirsch proposed that SFTs are formed in quenched Au by the condensation of vacancies and the subsequent dislocation glide [1]. Subsequently, de Jong and Koehler [5] proposed, in contrast, that SFTs could form directly from the collapse of vacancy clusters without dislocation glide. Although the exact formation mechanism of SFTs is still a subject of debate, the formation of SFTs has been known to sometimes alter the materials properties quite drastically through hardening [6, 7].
High- and medium-entropy alloys (HEA/MEAs), on the other hand, have attracted considerable interest over the past decades as some of them exhibit impressive mechanical properties [8–13]. Many studies have proposed a way to predict the characteristically high strength of HEA/MEAs with some materials parameters, including mean-square atomic displacement (MSAD), atomic misfit volume and size misfit [14–21]. Since the discovery by Zhang et al. [22] for the surprisingly large increase in room-temperature yield strength (∼24%) with the formation of short-range ordering (SRO) in the equiatomic Cr-Co-Ni MEA by simple heat-treatment (furnace cooling from 1273 K), SRO has been one of the main focuses in the research of HEA/MEAs. Many theoretical calculation studies indicate that when SRO is formed, the atomic arrangement may deviate from that for the ideally disordered state due to the enthalpic contribution [23–28]. Then, the existence and development of SRO in the Cr-Co-Ni MEA is indeed verified by experiment with resistometry [29, 30], dilatometry and calorimetry [31]. However, the small difference in atomic scattering factor for the constituent elements makes it difficult to gain x-ray/electron diffraction evidence for SRO in the Cr-Co-Ni MEA, although some reports claim diffuse intensity at 1/3{422} and 1/2{113} positions in the electron diffraction patterns with the ⟨111⟩ and ⟨112⟩ incidences respectively as the evidence of SRO [32–34]. Of interest to note is that Zhang et al. [22] insist that diffuse streaks along ⟨111⟩ directions is the evidence of SRO in the equiatomic Cr-Co-Ni MEA as it is observed in specimens slowly cooled from 1473 K but not in specimens quenched from 1473 K. Interestingly, however, the research group of Zhang et al. recently changed their original claim into that these diffuse intensities do not necessarily come from SRO [35] and that diffuse streaks along ⟨111⟩ directions come from a large density of stacking faults (SFs) and SFTs formed in the equiatomic Cr-Co-Ni MEA by quenching from 1273 K but not from SRO [36]. If this is true, their finding is remarkable in view of the fact that pure Cu, Ag and Au are only the FCC metals in which SFTs are reported to form only by quenching [37, 38].
In the present study, we investigate the formation of SFTs in the equiatomic Cr-Co-Ni MEA by transmission electron microscopy (TEM), to see if SFTs are indeed formed only by quenching. We prepare bulk specimens as well as thin TEM foil specimens after various heat treatments (quenching from high temperatures and subsequent annealing at intermediate temperatures) and with various preparation methods (electropolishing with and without subsequent Ar+ ion milling).
Single crystals of the Cr-Co-Ni MEA were grown by the optical floating-zone method with a growth rate of 10 mm/h under an Ar atmosphere from polycrystalline ingots prepared by arc-melting. The bulk single crystals were then homogenized at 1473 K for 168 h followed by water quenching and then cut into slabs of ∼2.5 mm in thickness. Some of the slabs were further annealed at 573 K and 623 K for 4 h followed by quenching into water. Thin foils parallel to (110) were then prepared by twin-jet electropolishing in a solution of perchloric acid, n-butanol and methanol (1:2:7 by volume). Some of the electropolished thin foils prepared from specimens quenched as well as furnace-cooled from 1473 K were further subjected to Ar+ ion milling with the Precision Ion Polishing System (PIPS, Gatan Model 691). Microstructures of specimens subjected to different heat-treatments and foil preparations were examined by transmission electron microscopy (TEM) with a JEOL JEM-2100FX electron microscope operated at 200 kV and scanning transmission electron microscopy (STEM) with a JEOL JEM-ARM200F electron microscope operated at 200 kV.
A bright-field (BF) TEM image and the corresponding selected-area electron diffraction (SAED) pattern with the [110] incident beam direction of a specimen quenched from 1473 K are shown in Fig. 1(a). The thin foil was prepared only by electro-polishing. No lattice defects such as dislocation loops and stacking fault tetrahedra are visible in the BF TEM image of Fig. 1(a). The absence of these lattice defects is further confirmed by atomic-resolution STEM imaging, as shown in Fig. 1(b). This is in sharp contrast to the report by Zhang et al. [36] that a large density of SFs and SFTs are formed in the equiatomic Cr-Co-Ni MEA only by quenching from 1473 K.

BF TEM image and the corresponding SAED pattern with the [110] incident beam direction of specimens (a) quenched from 1473 K, subsequently annealed at (c) 573 K and (d) 623 K. (b) atomic-resolution STEM image with the [001] incidence from a specimen quenched from 1473 K.
As it is well known that the SFT formation becomes very evident through coalescence of dislocation loops by annealing a high-temperature quenched speciemen at intermediate temperatures (at, for example, around 0.3–0.35 Tm: Tm is the melting point [3–5]), some specimens quenched from 1473 K were further annealed at 573 and 623 K for two hours, as their BF TEM image and the corresponding SAED pattern with the [110] incidence are shown in Figs. 1(c) and 1(d), respectively. The thin foils of Figs. 1(c) and 1(d) were also prepared only by electro-polishing. Unlike what is expected by the intermediate-temperature annealing, no defect evolution is noted even after annealing at 573 and 623 K. Of interest to note in the all SAED patterns of Figs. 1(a), 1(c) and 1(d) is that there are diffuse streaks extending along ⟨111⟩ directions, regardless of the annealing conditions even without any SFs and SFTs. This is again in contrast to the report by Zhang et al. [36] that diffuse streaks along ⟨111⟩ directions come from a large density of SFs in the Cr-Co-Ni MEA quenched from 1473 K. The nature of diffuse streaks along ⟨111⟩ directions in the Cr-Co-Ni MEA will be briefly discussed in the 4.2 section later.
In order to see the possibility of the formation of SFs and SFTs by ion-irradition, some TEM foils prepared from a specimen quenched and also furnace-cooled from 1473 K and preparated only by electro-polishing were subjected to 3 keV Ar+-ion milling at a shallow glancing angle (∼8 degrees) for 3 hours. In addition to many SFs (dots), some SFTs (triangles, indicated by arrows) are observed to form after ion-milling the foil specimen quenched from 1473 K, as shown in Fig. 2(a). The triangle defects can be identified as SFTs because their edges are parallel to [$\overline{1}$12], [1$\overline{1}$2] and [$\overline{1}$10] (intersections of four {111} planes in the [110] projection). The size (edge length) of SFTs ranges from as large as 6∼10 nm. The formation of many SFs and some SFTs are similarly observed even ion-milling was made for a thin foil specimen furnace-cooled from 1473 K, as shown in Fig. 2(b). The size (edge length) of SFTs in Fig. 2(b) is similar to that in Fig. 2(a) and ranges from as large as 6∼10 nm. This clearly indicates that SFTs can be formed by ion-irradiation regardless of the amount of quenched-in vacancies (i.e., whether the specimen was quenched or furnace-cooled from a high temperature) prior to ion-irradiation and that ion-irradition is an effective way to introduce SFs and SFTs though continuous formation of vacancies during irradiation. To note once again is that there are diffuse streaks extending along ⟨111⟩ directions in both the SAED patterns of the [110] incidence of Figs. 2(a) and 2(b) with their intensity being not so much different from those in specimens without any visible lattice defects such as SFs and SFTs (Figs. 1(a), 1(c) and 1(d)). This clearly indicates that diffuse streaks extending along ⟨111⟩ directions have an origin that is different from SFs and SFTs.

BF TEM image and the corresponding SAED pattern with the [110] incident beam direction after 3 keV Ar+-ion irradition for 3 hours for specimes (a) quenched and (b) furnace-cooled from 1473 K.
Clarebrough et al. [37, 38] reported that dislocation loops, SFTs and even voids are formed in zone-purified Cu, Ag and Au quenched from high temperature and subsequently annealed at relatively low temperatures (0.3–0.35 Tm: Tm is the melting temperature). In contrast, the present study clearly indicates that SFTs are not formed only by quenching the equiatomic Cr-Co-Ni MEA and subsequently annealed at intermediate temperature (in the same temperature range of ∼0.3–0.35 Tm). Although the exact formation mechanism has yet to be clarified [1, 5], it is explicitly described that vacancies must be involved in the formation of SFTs by quenching (and subsequent annealing). Therefore, it is not unreasonable to assume that the possibility of SFT formation by quenching may be correlated with the ease of vacancy formation and migration. To make a rough estimation for the ease of SFT formation, the enthalpies for the formation (Hf) and migration (Hm) of monovacancy for some FCC metals and the equiatomic Cr-Co-Ni MEA are tabulated in Table 1. The intrinsic stacking fault energies (SFEs) for these metals and alloy are also tabulated, since the stable edge length of SFT is believed to be governed by the SFE [1]. The Hf and Hm values of pure FCC metals are experimental values taken from Ref. [39], while those for the Cr-Co-Ni MEA are theoretically calculated in [27, 40]. The calculated Hf (1.62–2.03 eV) and Hm (0.31–1.38 eV) values of the Cr-Co-Ni MEA exhibit large variations but we take their averaged values (Hf∼1.83 eV, Hm∼0.85 eV) as they are comparable to those (Hf∼1.70 eV, Hm∼0.79–0.93 eV) obtained by positron lifetime spectroscopy for some other FCC HEA/MEAs in the Cr-Mn-Fe-Co-Ni system [41, 42]. The migration enthalpies of monovacancy (Hm) for Al, Au, Ag, Cu are similar to each other but they are all somehow slightly lower than the average value of the Cr-Co-Ni MEA. On the other hand, the formation enthalpy of monovacancy (Hf) for the Cr-Co-Ni MEA is noticeably large and is almost 2–3 times those for these pure FCC metals. In view of the fact that SFTs are formed in Au, Ag and Cu but not in the Cr-Co-Ni MEA, the formation of SFTs by quenching is concluded to be facilitated by the low values for formation and migration enthalpy of monovacancy, as they ensure the abundance of vacancies when quenched from high temperature and the ease of their motion during quenching (and subsequent annealing). Obviously, the very high Hf value and the relatively high Hm value of the Cr-Co-Ni MEA prevent from the SFT formation by quenching.
On the other hand, although the vacancy formation and migration enthalpies for Al are the lowest among these pure FCC metals, SFTs have never been reported to form only by quenching in Al [4]. This is usually attributed to the remarkably high SFE of Al such that the SFTs are less stable when compared to Au, Ag and Cu [1, 43, 44]. The energy of a SFT in a FCC metal is considered to be positively related to SFE (γ) as [43]
| \begin{equation} E_{\text{SFT}} = \frac{\mu a^{2}L}{12\pi (1 - \nu)}\left[\ln\frac{4L}{b} + 1.017 + 0.97\nu \right] + \sqrt{3} L^{2}\gamma \end{equation} | (1) |
where μ is the shear modulus, ν is Poisson ratio, b is the length of Burges vector, L is the edge length of a SFT given by
| \begin{equation} L^{2} = Na^{2}\sqrt{3} /2 \end{equation} | (2) |
where N is the number of vacancies involved in a SFT, a is the lattice parameter. Using the elastic constants and lattice parameters from Ref. [43, 45], the energies per vacancy of a SFT in Al, Au, Ag, Cu and the equiatomic Cr-Co-Ni MEA are plotted in Fig. 3 as a function of number of vacancies in a SFT. Figure 3 indicates that the stability of SFTs in Al is slightly higher than those for Au and Ag only when the number of vacancies in a SFT is less than ∼20 that corresponds to a very limited size (∼1.6 nm) of SFT in the equilibrium state in Al, indicating that the formation of SFTs is energetically less favored in Al. In contrast, when the size of SFTs increases so as to contain 20∼1000 vacancies (edge length 1.6∼12 nm), SFTs are supposed to be much more stable in Cu, Au and Ag than in Al, being consistent with the experimental observation of SFTs in these pure FCC metals. Figure 3 also indicates that while the SFT is the least stable in the Cr-Co-Ni MEA when the SFT is small containing less than ∼100 vacancies (edge length 3.2 nm), it becomes as stable as in Cu and Au when the size increases to contain ∼600 vacancies (edge length 7.2 nm). This is obviously because of the low SFE and is consistent with the size range of SFTs experimentally observed in the present study (Fig. 2).

Energy of SFT per vacancy plotted as a function of number of vacancies involved in a SFT for the Cr-Co-Ni MEA, Al, Au, Ag and Cu.
The formation of SFTs in the Cr-Co-Ni MEA is not possible only by quenching (and subsequent low-temperature heat-treatment) due to the high energy barrier for the formation and migration of vacancies but is possible by ion-irradiation even no quenched-in vacancies are contained prior to irradiation. This is because while the formation of vacancies is thermally difficult due to the high formation enthalpy, abundant vacancies are continuously introduced during ion-irradiation. Then, the formation of large SFTs as observed in Fig. 2 is stabilized by the low SFE of the Cr-Co-Ni MEA. This is consistent with the report by Lu et al. [46] that a high density of SFTs with an average size 8.0 nm is observed in the Cr-Co-Ni MEA after high-load nanoindentation followed by a large dose of ion irradiation (3 MeV Ni2+, fluence 5 × 1016/cm2) at 693 K. Then, we suspect that the formation of SFTs and SFs in the Cr-Co-Ni MEA quenched from 1473 K reported by Zhang et al. [36] is an artifact that occurred during TEM thin foil preparation, although they did not describe how their TEM foils were prepared. Of significance to notice is that the research group of Zhang et al. reported in their original paper [22] that the yield strength at room temperature increases remarkably by 24% in the specimen furnace-cooled from 1273 K when compared with the specimen water-quenched from 1473 K. In view of their new claim [36] that the formation of SFTs occurs when the specimen is water-quenched from 1473 K, these two results [22, 36] of the same research group obviously contradict with each other.
4.2 Diffuse Streaks along ⟨111⟩Zhang et al. [22] originally reported that the occurrence of diffuse streaks along ⟨111⟩ is the evidence for the formation of SRO in the Cr-Co-Ni MEA furnace-cooled from 1273 K, although they later changed this conclusion to that these diffuse streaks along ⟨111⟩ do not necessarily come from SRO [35] and further to that diffuse streaks along ⟨111⟩ directions come from a large density of SFs and SFTs formed in the equiatomic Cr-Co-Ni MEA by quenching from 1473 K [36]. In the present study, however, diffuse streaks along ⟨111⟩ is usually observed in the Cr-Co-Ni MEA regardless of heat-treatments and foil preparation methods (i.e., regardless of whether lattice defects such as SFs and SFTs are contained in the specimen or not). SFs and SFTs formed with the habit plane parallel to {111} are known also to produce diffuse streaks along respective ⟨111⟩ with the intensity depending on their density [47, 48]. However, diffuse streaks originated from these lattice defects are generally much sharper than diffuse streaks observed in the present study. Diffuse streaks observed in the present study are thus considered not to be related with SFs and SFTs but seem similar to those discussed by Honjo et al. [49] a long time ago. They observed similar diffuse streaks along ⟨111⟩ in the SAED patterns with the ⟨110⟩ incidence not only for Al and Au but also for Si and Ge. They concluded that the observed diffuse streaks occur as an interception of the Ewald sphere with {110} diffuse intensity walls formed by the displacements of atoms in ⟨110⟩ nearest-neighbor chains in the cluster motion (in a several atomic length) of thermal origin. The observed characteristics of diffuse streaks along ⟨111⟩ coincide well with what is expected from the predictions by Honjo et al. [49] when examined along ⟨110⟩. But, more work (such as observations along different crystal orientations and at low temperatures) is definitely needed to draw the final conclusion on the origin of diffuse streaks along ⟨111⟩ in the Cr-Co-Ni MEA.
This work was supported by Grant-in-Aids for Scientific Research on Innovative Areas on High Entropy Alloys through the Grant number JP18H05450 and JP18H05451, in part by JSPS KAKENHI (Grant numbers JP22H00262 and JP23K17338) and by JST Spring (Grant number JPMJSP2110).