Micropillar compression of single-crystal single-phase (Co, Cu, Mg, Ni, Zn)O

(Co, Cu, Mg, Ni, Zn)O was first developed by Rost et al. in 2015 focusing on the formation of this novel material from both a characterization and thermodynamic standpoint.¹ Subsequent studies on this high entropy oxide (HEO) have characterized its functional and mechanical behavior.^2–10 Studies on mechanical behavior have included the application of microscopic techniques, such as nanoindentation and nanoscratch, and examined either the elastic or plastic response of the material, respectively.^9,10 Pitike et al. used nanoindentation and computational methods to evaluate the effects of crystallographic orientation and magnetic ordering on the elastic properties of (Co, Cu, Mg, Ni, Zn)O, while Wang et al. used nanoscratch techniques to examine the effect of grain size on room-temperature mechanical behavior and deformation mechanisms and determined that dislocation activity played a key role in the observed behavior. Further examination of these two studies highlights the lack of understanding of the interplay between grain orientation, plastic response, and deformation mechanisms in (Co, Cu, Mg, Ni, Zn)O.

Micropillar compression has been used to investigate the plastic response in ceramic materials including high entropy carbides, sapphire crystals, and MgO.^11–14 Quantitative analysis of slip systems, such as the calculation of the Schmid factor (SF), can facilitate interpretation of dislocation activity in micropillar compression experiments.¹⁵ For instance, under micropillar compression, the grain orientation of MgO, one of the components in (Co, Cu, Mg, Ni, Zn)O and with the same crystal structure (rock salt), is known to affect its mechanical behavior, with {100} grains having a lower yield stress than {111} grains.¹⁶ Wang et al. concluded that the nanoscratch deformation behavior of (Co, Cu, Mg, Ni, Zn)O was similar to that of MgO. MgO is known to form kink bands, dislocation walls, and dislocation entanglements as a consequence of deformation, as observed during high temperature deformation on bulk samples, room-temperature nanoindentation on bulk samples, and room-temperature microcompression on micropillars.^14,16–19 With enough misorientation created at or near the kink bands or dislocation walls, subgrains can form within a single grain of MgO, as well as during recrystallization experiments, changing the deformation response.^19–21 Additionally, the formation of dislocation entanglements in MgO can lead to dislocation entanglement-mediated cracking, particularly due to the Zener–Stroh mechanism, where the induced stress state at the head of the dislocation entanglement creates an energetically favorable cracking mechanism compared to continued dislocation climb or glide.^22–24 With inspiration taken from these previous studies on (Co, Cu, Mg, Ni, Zn)O and MgO, we qualitatively explore the role of crystallographic orientation on room-temperature mechanical deformation behavior in (Co, Cu, Mg, Ni, Zn)O using post-deformation electron microscopy, with particular emphasis on the role of dislocation activity in the absence of elevated temperatures, since elevated temperatures are known to affect slip activation and plastic response.^11,12

Bulk, polycrystalline (Co, Cu, Mg, Ni, Zn)O samples were prepared from equimolar amounts of each constituent binary oxide with a purity greater than 99.7% and blended together a Fritsch Premium 7 planetary ball mill (PBM). The bulk, polycrystalline samples, with an average grain size of approximately 25 μm, were consolidated using conventional sintering at 1100 °C for 12 h in a CM Furnaces 1210BL elevator furnace.³ The phase state and microstructure of the samples have been characterized in detail using x-ray diffraction, scanning electron microscopy, and energy dispersive x-ray spectroscopy in a previously published study by Dupuy et al.³ The consolidated samples were polished to a 1 μm finish before using electron backscatter diffraction (EBSD) in a Tescan GAIA3 scanning electron microscope (SEM) to map grain orientation relative to the z-axis (perpendicular to the surface), as seen in the micrographs and inverse pole figure (IPF) color maps provided in the supplementary material (Fig. S1). Three grain orientations were selected for micropillar compression: (001), (101), and (111). The micropillars, hereafter labeled as Pillar (001), Pillar (101), and Pillar (111), respectively, were made using a FEI Quanta-3D dual beam SEM with focused ion beam (FIB) capabilities to dimensions of approximately 2 μm in diameter and 5 μm in height, following the methods described in the literature.^17,25 To minimize gallium contamination and tapering, a low ion beam current of 10 pA was used for the final cutting step. Because the micropillars are FIB'ed from the center of grains with select orientations, and the micropillar size is about five times smaller than the average grain size, the micropillars are considered to be single-crystal rather than polycrystalline.

Each micropillar was compressed using a Hysitron PI-85 nanoindentation system equipped with a picoindentor with a flat top conical tip placed inside the Quanta SEM for in situ imaging. Each micropillar was compressed under displacement control mode using an initial strain rate of 10⁻³ s⁻¹ to a displacement corresponding to ∼10% strain, at which point the compression was stopped. Engineering stress vs engineering strain curves were calculated using the methods of Greer et al., while using a previously reported elastic modulus from Cortez et al., experimentally measured on the same bulk sample.^26,27

Post-compression characterization was performed on the micropillars using both SEM and a JEOL JEM-2800 transmission electron microscope (TEM). Micrographs of the deformed micropillars from 0° to 270° angles were taken in the SEM before preparing TEM lift-outs from each micropillar. The post-compression TEM lift-outs were made by cross-sectioning the deformed micropillars using FIB.

The micropillars, before compression, are shown in Figs. 1(a), 1(f), and 1(k), respectively, with the picoindentor tip hovering above, while the views of the micropillars post-compression are shown in Figs. 1(b)–1(e), 1(g)–1(j), and 1(l)–1(o), respectively. The micropillars were compressed, producing the engineering stress (MPa) vs engineering strain curves shown in Fig. 1(p). These figures provide qualitative evidence to support the interpretation of dislocation activity as a function of crystallographic orientation in (Co, Cu, Mg, Ni, Zn)O, as further described below.

As shown in Fig. 1(p), Pillar (001) exhibits a large stress drop at an engineering strain value of approximately ∼7% indicating the potential formation of a slip band, with additional small stress drops occurring as compression continued. Stress drops in micropillars can also indicate dislocation formation or crack formation, with crack formation being less likely due to the size of the pillars.^28–32 It is hypothesized that the stress drop observed in Fig. 1(p) is due to slip activation and dislocation formation, not cracking. Post-compression imaging of Pillar (001) did not reveal any visible slip band traces on the micropillar surface [Figs. 1(b)–1(e)], potentially obscured by the cracks that formed (highlighted with white arrows). The observed cracks on both the top and sides of the micropillar indicated fracture occurred, with the cracks on the top of the micropillar being similar to those reported for MgO micropillars.²⁵ Pillar (101) exhibited small stress drops throughout the compression experiment and did not exhibit any visible cracks [Figs. 1(g)–1(j)] but was slightly bent due to the formation of slip bands (highlighted with blue arrows) across the micropillar. Pillar (111) had small stress drops during compression and displayed brittle behavior typical of transgranular fracture.³³ As observed in the 360° view in Figs. 1(l)–1(o), a portion of the micropillar fractured off of the sample (highlighted with green arrows) and another large crack formed on the opposite side (highlighted with white arrows).

In the rock salt crystal structure into which (Co, Cu, Mg, Ni, Zn)O forms, the most relevant family of slip systems is $1 / 2⟨110⟩ 1 1 ¯ 0$ type with six individual slip systems, as previously identified by Wang et al.¹⁰ The other two families of slip common in this crystal structure (with planes {111} and {001}) were not considered in the analysis based on what is known about their deformation behavior in MgO: the {111} system is only activated at elevated temperatures, and the {001} system, while active at room temperature, exhibits a higher critical resolved sheer stress (CRSS) than the {101} system, indicating that the {101} type slip system is most likely to activate.^22,34–36 These six slip systems in $1 / 2⟨110⟩ 1 1 ¯ 0$ can be evaluated relative to the crystallographic orientation of the micropillars and the applied stress to calculate the corresponding SF values and to identify the number of inactive (SF = 0) and potentially active (0 < SF ≤ 0.5) slip systems for each micropillar orientation.^37–41 The results of this analysis are shown in Fig. 2.

This slip system analysis revealed SF values of varying magnitude. The (001) orientation corresponds to SF values of 0 and 0.5, the (101) orientation corresponds to SF values of 0 and 0.25, and the (111) orientation corresponds to a SF value of 0. The (001) orientation can activate four of the six systems, the (101) orientation can activate four, and the (111) orientation cannot activate any. A higher number of potentially active slip systems indicates that there are more paths for deformation to occur, which is advantageous since slip bands, created from activated slip systems, help mitigate the formation of cracks.

The (001) orientation is the only orientation with a SF = 0.5 (the maximum SF value possible), for which activation of the slip system is expected. Moreover, all four of these (001) slip systems have the same SF value, indicating they should have an equal probability of activation. Notably, the results shown in Fig. 1(p), particularly the large stress drop, reveal the possible activation of one slip system. The smaller stress drops seen throughout the compression experiment are possible indicators that additional slip systems could have been activated as well, agreeing with the SF analysis that more than just one slip system is expected to be activated for this crystallographic orientation. The (101) orientation also had four potentially active slip systems, all four having a SF = 0.25. The post-compression micrographs of Pillar (101) exhibit two visible slip bands [Figs. 1(f)–1(j)], indicating that the two potentially active slip systems were most likely activated. Unlike the other two orientations, the (111) orientation had no potentially active slip systems since the SF = 0 for all six slip systems. With all six slip systems being inactive, dislocation activity is not expected, and cracking should occur instead, which agrees with the cracking behavior observed in Figs. 1(l)–1(o).

The micropillars were evaluated post-compression by TEM to characterize the morphology, size, and distribution of defects. The TEM micrographs for all three micropillars, post-compression, are shown in Fig. 3, where the axis of compression is indicated by the black arrow; the orientation of the pillar is indicated with a schematic of a unit cell and the crystallographic plane.

An entanglement of dislocations in Pillar (001) is observed in Fig. 3(a) (outlined with a white dashed rectangle). The entanglement of dislocations is roughly parallel to the axis of compression. Figure 3(b) is a higher magnification image of the inset from Fig. 3(a) and reveals a nanocrack, based on the distinct contrast compared to the bulk of the material. The crack is more than 250 nm in length and transgranular in type, which is expected in a single-crystal micropillar. Based on similar reports and the observed behavior in the current study, the formation of the nanocrack (roughly perpendicular to the dislocation entanglement) is attributed to the Zener–Stroh mechanism where it was more energetically favorable to crack transgranularly instead of continuing with dislocation glide or climb.^22–24 Additional dislocation entanglements with similar orientations, i.e., the angle between the dislocation entanglement and the axis of compression, to the one shown in Fig. 3(a) were observed throughout the micropillar cross section. We propose that these dislocation entanglements propagated throughout the material via dislocation glide, similar to what occurs in MgO at room temperature in the same slip system: $1 / 2⟨110⟩ 1 1 ¯ 0$⁠.^22,41–43

The post-compression TEM micrographs for Pillar (101) reveal dislocations (indicated by the white dashed ovals) that formed during deformation [Fig. 3(c)]. Compared to the entanglement of dislocations observed in Pillar (001), the dislocations here were individual. These dislocations propagated by dislocation cross-slip and/or glide, similar to the behavior observed in Pillar (001). Fewer dislocations propagated in the micropillar preventing entanglement or generalized clustering. Less entanglement, as observed in Pillar (101), can be attributed to a potentially smaller Peierls stress compared to that for Pillar (001), which limits the number of dislocations that can glide.^44,45 Also of note, the dislocations were not oriented parallel to the axis of compression, with some forming ∼40° relative to the compression axis. Their contrasting directions inhibit further dislocation formation and propagation as they collide with one another, essentially pinning each other until more stress is added to overcome this interaction.⁴⁶ 

Pillar (111) did not form any dislocations across the cross section, instead having generalized regions of amorphization created due to the compression experiment, as observed in Fig. 3(d). (S)TEM imaging from previous studies revealed no inherent dislocations or amorphization prior to deformation.^10,47,48 With none of the six slip systems being potentially active, no slip bands could form, and no dislocations were activated, leading to the brittle fracture observed in Figs. 1(k)–1(o). The three deformed micropillars had varying degrees of dislocation activity, with Pillar (001) forming large entanglements, Pillar (101) forming singular dislocations, and Pillar (111) forming no visible dislocations.

With no visible dislocations seen in Pillar (111), HRTEM imaging was done to look for other signs of deformation besides the regions of amorphization [Figs. 4(a) and 4(b)]. Notably, no additional signs of deformation or dislocations were observed [Fig. 4(a)]. In Fig. 4(b), the variations in contrast revealed a coherent precursor phase, outlined in blue, embedded within the rock salt primary phase, outlined with white, to show the contrast differences. The coherency and similar orientation to the rock salt phase indicate that this phase has a cubic crystal structure. Based on our previous study, this phase is most likely a precursor phase to a Co-rich spinel secondary phase.⁶ Since all three micropillars were made from the same bulk sample, these precursor phases are most likely distributed across all three micropillars. These precursor phases are inherently a defect site and contribute to the overall deformation behavior. They present a potential barrier to dislocation motion, leading to dislocation entanglement and/or slip bands, similar to how GP zones and intermetallic phases act as barriers to dislocation motion.^49–52 The coherent interfaces between the precursor phase and the primary rock salt phase also act as potential sources for dislocation nucleation.⁵³ Depending on the slip system activation for each micropillar, these precursor phases have the potential to impact the deformation mechanism. With their presence in all three micropillars, they are not expected to contribute to the crystallographic comparison this study is focused on, but acknowledging their existence and potential contribution is important. Their exact role requires further investigation.

As the micropillars deform and form dislocations, regions of high strain and misorientation are created within the single-crystal micropillars. If enough misorientation is created during micropillar compression, subgrains are likely to form, similar to the behavior observed in MgO or ZnO.^18,33 With a high concentration of dislocation entanglements in Pillar (001), there was sufficient misorientation to form subgrains, where the dislocation entanglements could become the subgrain boundaries, consistent with descriptions in the literature.⁵⁴ Although subgrains were not directly observed, Pillar (001) had high concentrations of dislocations across the entirety of the pillar, creating misorientation globally, which could lead to subgrain formation. In contrast, the low concentration of dislocations in Pillar (101) fails to create the misorientation necessary to form subgrains. Pillar (111) lacked the dislocation formation necessary to form subgrains and proceeded to undergo brittle fracture with little-to-no plastic response. Figure 5 is a schematic summarizing the deformation mechanisms in the micropillars for each crystallographic orientation during compression, particularly highlighting the differences in dislocation formation, as described above.

Once the dislocations form and/or enough stress is added to the micropillar, cracking is inevitable. Single-crystal materials tend to form cracks along crystallographic planes. This behavior is equivalent to when transgranular-type cracks form in polycrystalline materials. Because subgrain boundaries, which could form from the dislocation entanglements observed in Pillar (001), are weaker than the single-crystal region, they would create additional routes for crack propagation, specifically intergranular-type cracking. The potential for both transgranular-type and intergranular-type cracks to form in single-crystal, single-phase materials is unexpected.⁵⁵ Yet, this atypical cracking behavior provides an explanation for the variations in fracture behavior observed in Pillar (001) and Pillar (111), while Pillar (101) had no crack formation. Pillar (001) displayed both intergranular-type and transgranular-type cracks, specifically, the large cracks along the side of the micropillar observed in Figs. 1(a)–1(e) and the nanocrack (attributed to Zener–Stroh dislocation entanglement-mediated cracking) observed in Fig. 3(b), respectively, while Pillar (111) appeared to form only brittle, transgranular-type cracks, as observed in Figs. 1(k)–1(o).

Variations in crystallographic orientation created unique mechanical responses, which were observed during micropillar compression of (Co, Cu, Mg, Ni, Zn)O. Schmid factor analysis revealed consistent trends with the post-compression results, deepening the understanding of slip system activation in high entropy oxides with the rock salt crystal structure. Each micropillar had a distinct dislocation structure after micropillar compression. Pillar (001) had extensive dislocation entanglement across the TEM cross section. Pillar (101) had no dislocation entanglement, exhibiting isolated dislocations across the TEM lamella. Pillar (111) had no evidence of dislocation formation, instead forming amorphous regions during the micropillar compression experiment. Misorientation due to the dislocation entanglement occurring in Pillar (001) led to the possible creation of subgrains. Pillar (001) presented both intergranular (made possible based on the weakened subgrain boundaries compared to the bulk material) and transgranular style cracks (attributed to the Zener–Stroh mechanism for dislocation entanglement-mediated cracking), atypical in single-phase, single-grain ceramics. Pillar (101) remained crack free. Pillar (111) only formed brittle, transgranular cracks, resulting in brittle fracture typical of ceramic micropillars. This micro-mechanical experimentation coupled with in-depth microstructural characterization revealed unusual room-temperature, deformation behavior in these ceramic materials: dislocation formation, slip band activation, and subgrain formation. Deeper understanding gained from this study highlights the unusual compression-based deformation of (Co, Cu, Mg, Ni, Zn)O and how it can be used to further explore the mechanical behavior of this and other high entropy oxides.