Quasi-In Situ EBSD Study of Anisotropic Mechanical Behavior and Associated Microstructure Evolution in Zircaloy-4

Abstract

The anisotropic mechanical behavior and associated microstructure evolution in annealed Zircaloy-4 were investigated at room temperature, using quasi-in situ tensile tests along the typical direction, rolling direction (RD), and transverse direction (TD). Herein, the in-grain misorientation axes (IGMA) and the nominal Schmid factors were evaluated to analyze the slip mode based on the electron backscatter diffraction. The IGMA result shows that there were anisotropic slip modes within grains, whose basal poles were parallel with the TD (TB) and placed within 40 to 50 degrees from the normal direction (ND) to the transverse direction (N (40°–50°) TB)), under different loading directions. When loading along the RD, the basal <a> slips were activated in the N (40°–50°) TB and TB orientation grains, while the second-order pyramidal slips were activated in the grains when loading along the TD. The relatively higher ultimate tensile strength and elongation in Zircaloy-4 when tensile along RD occurs due to its much higher frequency of soft grains (88.54%) than the TD sample (64.29%), and the synergy deformation among local grains. The present study demonstrated that the anisotropic mechanical behavior of Zircaloy-4 was attributed to the combined effects that exist between the anisotropic slip behavior and the different compatible deformation capabilities. Many shallow dimples and cleavage regions were observed on the fracture surface in the TD sample. Such fracture features are consistent with the lower ultimate tensile strength ~470 MPa and elongation ~14.5% in the deformed tensile Zircaloy-4 along the TD.

1. Introduction

Zirconium alloys are extensively applied in the nuclear power field, such in fuel cladding materials and position frameworks, due to their combination of good neutron transparency, excellent corrosion resistance, and outstanding mechanical properties. However, the α-Zr alloy exhibits pronounced anisotropic mechanical responses owing to its hexagonal close-packed (hcp) crystal structure with low crystallographic symmetry at room temperature [1]. These behaviors have also been extensively reported in other hcp metals e.g., titanium alloy [2,3] and magnesium alloy [4,5]. The same crystal structures are apparent between Zr (c/a = 1.593), Ti (c/a = 1.58), and Mg (c/a = 1.624); the different axis ratios of c/a result in discrepancies in the predominant slip mode [6].

To date, numerous simulations and experiments have been conducted on the plastic deformation mechanism of zirconium alloy. Whether there are polycrystals or single crystals, the prismatic <a> slip is most readily activated in plastic-deformed zirconium [7,8,9], which is consistent with the ab initio calculations by Clouet [10]. The first-order pyramidal <c + a> slip, {10$\stackrel{¯¯}{1}$2} <10$\stackrel{¯¯}{1}$1> tensile twinning, and basal slip are also crucial plastic deformation mechanisms [9,11,12,13,14,15]. Twinning always exhibits sensitivity to temperature [16], loading direction [17,18], and strain rate [9] as experimental factors. The inhomogeneous microstructure in zirconium alloy leads to inherently anisotropic mechanical behavior. For example, the yield strength in tensile Zircaloy-4 plates along the radial direction is higher than the transverse direction, in both the cold-worked stress-relieved or the various Nb-modified Zircaloy-4 alloys [19]. A similar research result has also been shown in a the zircaloy-2 cladding tube [20]. The heat treatment regime does not decrease the degree of anisotropy in zirconium alloys [21]. The effect of irradiation on the plastic anisotropy of zircaloy cladding tubes shows, through experiments and simulations, that the anisotropy of zircaloy is decreased after irradiation, because the prismatic <a> slip is more strongly suppressed, while the pyramidal slip mode becomes the dominate slip mode along with the prismatic <a> slip [22]. The studying of anisotropic mechanical responses based on the dislocation evolution in zirconium alloys has been reported [1,8,23]. However, research on the anisotropic mechanical behavior of hcp-Zr under external loading from the grain orientation viewpoint is lacking, especially related to the discrepancy of slip behaviors at the grain scale that through the quasi-in situ electron backscattered diffraction (EBSD) method.

Transmission electron microscopy techniques [1,7,8,24,25] and slip trace analysis [26,27,28] have been used to reveal the activation of slip and/or twinning modes, in both Zr and in other hcp metals. Despite this, both methods show obvious limitations in applied circumstances. The former requires substantial time to determine the slip mode of a large number of grains. In addition, this method has only been applied to materials with a small strain and it barely distinguishes the glide plane of severely deformed materials [7,8]. The latter is only suitable for coarse-grained metal materials exposed to a limited strain. In order to fulfill the reliable and efficient statistical requirements of slip modes in small-grain polycrystal samples, the in-grain misorientation axes (IGMA) method based on the orientation of individual grains was introduced in this study. The validity of the IGMA method for the determination of slip modes in hcp metals, such as Zr [29], Ti [30], and Mg [31], has been demonstrated in prior research.

Much less is known to date about the effects of grain orientation on anisotropic mechanical behavior, such as slip modes. In this paper, the combination of a quasi-in situ EBSD technique and the IGMA method was used to reveal the activity of slip modes in grains and the corresponding microstructure evolution, while the mechanical response and fracture behavior were also investigated.

2. Materials and Methods

Full recrystallized Zircaloy-4 was selected for investigation in this study. It contained 1.5, 0.2, 0.1, and 0.13 wt.% of Sn, Fe, Cr, and O, respectively, the remainder being Zr. The dimensions and structure of the dog-bone shape are shown in Figure 1, where RD, TD, and ND represent the rolling direction, transverse direction, and normal direction, respectively. The tensile samples were prepared by electrical discharge machining along the RD and TD. An electro-mechanical universal testing machine was used to conduct the tensile tests with a constant strain rate of 0.5 × 10⁻³/s at room temperature, and the quasi-in situ tests were interrupted at the presetting strain. The plastic strain ratio, ε, is the actual calculated strain value.

During tensile testing, the microstructure evolution was recorded by a NOVA400 field emission scanning electron microscope (SEM) coupled with EBSD and an electron channeling contrast (ECC) probe. The acceleration voltage and scanning step size were 20 KV and 0.5 μm, respectively. EBSD data were processed using professional software (Channel 5, HKL Technology-Oxford Instruments). At −30 °C, electropolishing of Zircaloy-4 samples was performed in a solution of 10% perchloric acid, 20% 2-butoxyethanol, and 70% methanol for 15~18 s after mechanically polishing to meet the surface quality requirements for the microstructure characterization.

The IGMA method was used to study the anisotropic mechanical behavior of Zircaloy-4. The active slip modes can be determined by matching the Taylor axis for the given slip mode and the IGMA experimental results. The activation of the slip mode induced a crystal lattice rotation that could be detected by EBSD. The Taylor axis corresponds to the common slip modes of α-Zr, as shown in

Table 1. Slip modes available in α-Zr and the corresponding Taylor axis.

Slip System	Slip Plan, Direction	Taylor Axis	Variants Number of Taylor Axis
Bas. <a> slip	(0002), <$\overline{2}$10>	<0$\overline{1}$10>	3
Prism. <a> slip	(1$\overline{1}$00), <$\overline{1}$110>	<0001>	1
Pyram. <a> slip	(01$\overline{1}$1), <$\overline{1}$110>	<0$\overline{1}$12>	6

. In order to meet the IGMA analysis requirements, the misorientation axes were only plotted with θ between 1.0 and 2.0 deg. Other detailed information on the IGMA method has been reported by Chun [30]. Meanwhile, the nominal Schmid factor (SF_N) was evaluated to verify the universality of IGMA in the deformed Zircaloy-4 alloy at room temperature.

3. Results and Discussion

3.1. Mechanical and Fracture Behavior Response to the Loading Direction

The obtained engineering stress-strain curves of Zircaloy-4 along the RD and TD are shown in Figure 2. The TD sample exhibited a higher yield strength (YS~433 MPa) compared with the RD sample (~322 MPa). This yielding anisotropy has also been reported in previous studies of Zr, Mg, etc. [1,32], which has been ascribed to the deficiency of mobile dislocations [33]. The ultimate tensile strength (UTS~486 MPa) and uniform elongation (UE~22.8%) of the RD sample are notably higher than the TD sample (UTS~470 MPa, UE~14.5%). Discrepancies among the YS, UTS, and UE of Zircaloy-4 indicate that the anisotropic mechanical properties change along the loading direction. Additionally, a remarkable yield drop occurs in Zircaloy-4 along the TD compared with the RD sample. Shi et al. reported a gradual disappearance and re-appearance of yield drop upon introducing different pre-strain histories in zirconium alloy. This phenomenon is commonly associated with dislocation mobility and dislocation density. A lack of dislocation density causes the increase in flow stress to trigger the movement of dislocation. Shi et al. attributed such features to the lack of glissile dislocations at the initial plastic deformation stage [34], which is consistent with the mechanism proposed by Johnstok et al. [34].

Figure 3 presents the microscopic fractography of Zircaloy-4 after tensile fractures along different loading directions at 1000× and 4000× magnification. Both specimens show many dimples and tearing edges (Figure 3a–h). A relatively higher number of cleavage surfaces were observed in the TD sample, as the yellow lines denote in Figure 3c,g. Such features in the fracture morphology in the outer side position are more significant (Figure 3d,h). The hardening ability and deformation mechanisms of metals are more strongly associated with ductile fracture behaviors [2]. Shi et al. indicate that zirconium alloys exhibit promoted strain hardening capability when tensile along the RD than the TD from the dislocation viewpoint [1]. Compared with the RD tensile sample (Figure 3a,b,e,f), the TD sample was conducive to void nucleation, growth, and coalescence due to its weak hardening ability. Therefore, many relatively shallow dimples were observed on the fractured surface, as shown in Figure 3c,d,g,h. The macroscopic fracture morphology showed a cup-cone fracture surface, as the inserted figures in Figure 3a–d show. Necking at the fracture position was more remarkable in the RD sample than the TD sample, which indicates that the RD sample undergoes severe plastic deformation. Both the microstructure and macroscopic morphologies observed in the TD sample were associated with features in a lower ductility sample, which is consistent with the UE result in Zircaloy-4 under different loading directions.

3.2. Microstructure Evolution during Deformation

To study the anisotropic mechanical behavior, the evolution of the microstructure during the RD/TD tensile test was observed. The inverse pole figure (IPF) and the associated grain boundary (GB) distribution map in the RD sample at 0.00% and 6.64% tensile strain are shown in Figure 4. High-angle grain boundaries (HAGBs, >10°) and low-angle grain boundaries (LAGBs, 2~10°) are denoted as the black and red lines, respectively. The initial undeformed microstructure consisted of equiaxed grain with a 4.78 μm average size, and the HAGBs were the predominate grain boundary, as depicted in Figure 4a,d. As the strain increased to 6.64%, the color distribution among grains exhibited evident evolution compared with the undeformed sample, which suggests that significant plastic strain occurs inside these grains, as Figure 4b shows. Combined with the LAGB distribution in Figure 4c, it can be seen that all of the grains suffered plastic deformation to some extent. Meanwhile, the proportions of HAGBs and LAGBs were 36.3% and 63.7%, respectively. It can also be observed that plastic deformation was in homogeneously distributed among grains, both in the IPF and GB distribution maps, which is attributed to the discrepancies in activated slip modes among grains with different crystal orientations under loading conditions, such as the activated likelihood of the identical slip mode or the transmission of slip mode between grains [28]. With an increase in the tensile strain, both LABGs formation and transformation to HAGBs were observed, as the white arrow denotes in Figure 4b. At the initial stage of plastic deformation, the local rearrangement of the high-density dislocations promoted the LAGBs to form and release energy. The GB misorientation increases further, which facilitated the transformation of LAGBs to HAGBs with progressive strain increases [1,35]. More importantly, an evident lattice rotation among grains with an increase in strain can be observed, as the yellow ellipse region denotes in Figure 4a,b.

The IPF and GB distribution map at 0.00% and 6.20% strain during the tensile process in Zircaloy-4 along TD are shown in Figure 5. The main features, i.e., the average size (4.69 μm) and LAGBs ratio (2.5%), of the as-received materials are in accordance with the observed information in Figure 4a,c. As the strain reached 6.20%, Figure 5b shows that most grains suffered severe plastic deformation, and their morphology was elongated significantly along the loading direction compared with the RD sample (Figure 4b). Additionally, the transformation from LAGBs to HAGBs was also observed in partial grains, as the white arrows denote in Figure 5b, which suggests that significant heterogeneous deformation exists within grains. At the same observed region, the density of LAGBs in the TD tensile sample was larger than in the RD sample, which indicates that the hindering effect of dislocation motion in the RD sample was less than in the TD sample. The feature where LAGBs were concentrated along GBs verified this inference further. The stronger hindering effect for dislocation motion in the TD sample may be the key factor that results in its higher yield strength than the RD sample. In order to evaluate the strain distribution during the tensile process under different loading directions, the local strain data quantified by kernel average misorientation (KAM) that based on the third nearest neighborhood in Zircaloy-4 with various strains under different loading directions are plotted in Figure 6. A higher KAM value corresponds to a higher local stress concentration or greater degree of plastic deformation [36,37]. It can be found that local strain concentration was positively proportional to the tensile strain, regardless of loading direction, and the local strain in the undeformed sample maintained a lower value. More importantly, the local grain misorientation in the TD sample was much larger than the RD sample, i.e., 1.76 at 6.64% strain for the RD sample versus 2.04 at 6.20% strain for the TD sample. Therefore, these observations, combined with the microstructure evolution, suggest that Zircaloy-4 possesses better synergistic deformation effects, namely good uniform plasticity, when tensile along the RD compared to the TD.

3.3. Slip Behaviors during Tensile along RD/TD

In order to further reveal the anisotropic mechanical and fracture behavior responses to the loading direction, slip modes of individual grains were statistically analyzed based on the IGMA and SF_N methods. The lattice rotation caused by plastic deformation within the grain can be detected by EBSD, as introduced by Luan et al. [29] and Chun et al. [30]. The rotation axes that correspond to different slip modes in Zircaloy-4 alloy are listed in Table 1. Figure 7 shows the IGMA distribution that corresponds to the nine selected grains, as denoted A–I in Figure 4a. The IGMA distribution that corresponds to the overall region (Figure 4c) presents high concentrations of IGMA at <0001>, namely activated prismatic <a> slip systems [29,30]. The selected grain in the present study was divided into three groups based on its crystallographic orientation in terms of the position of the basal pole [30], with the basal pole parallel with the TD and ND denoted as TB (A–C) and NB orientation grains (D–F), respectively. “N (40°–50°) TB” means the basal poles of a grain located 40° to 50° from the ND to the TD. It can be observed that the activated prismatic <a> slip exists in three groups of grains without discrepancies, while the basal <a> slip only exists in the NB and TB orientation grains for the IGMA distribution around <10$\stackrel{¯¯}{1}$0> axis with higher intensity.

Table 2. Calculated SF_N values for slip modes of selected grains in the RD sample.

No.	Euler Angle (φ1, Φ, φ2)			Bas. <a>	Prism. <a>	Pyram. <a>
A	66.65	168.71	59.21	0.109	0.465	0.212
B	6.61	169.17	44.32	0.014	0.484	0.192
C	140	26.12	4.60	0.089	0.48	0.206
D	91.35	36.09	55.96	0.317	0.295	0.169
E	170.46	41.68	36.33	0.073	0.435	0.182
F	132.2	38.91	34.51	0.251	0.373	0.216
G	4.29	83.95	22.87	0.049	0.476	0.196
H	2.2	96.9	51.25	0.024	0.488	0.197
I	0.42	81.45	1.83	0.004	0.449	0.176

shows the calculated SF_N value for the nine selected grains, and the bold font represents the maximum value of the slip mode which is the most readily activated, indicating that prismatic <a> slip is the predominant activated slip system. However, the basal <a> slip may also be activated, regardless of whether it is in the IGMA evaluation or the prediction from the calculated SF_N. The difference between IGMA and SF_N may be ascribed to the lack of neighboring grain influences, because the SF_N is only calculated based on the Euler angle inside the grain at the fixed loading direction.

The IGMA distribution for the selected grains in the TD sample is presented in Figure 8, and the corresponding SF_N calculation results are listed in

Table 3. Calculated SF_N values for slip modes of selected grains in TD sample.

No.	Euler Angle (φ1, Φ, φ2)			Bas. <a>	Prism. <a>	Pyram. <a>
A	101.3	174.02	23.88	0.067	0.493	0.207
B	164.89	5.89	50.51	0.020	0.472	0.188
C	165.99	170.3	45.05	0.027	0.442	0.180
D	93.93	138.7	48.02	0.240	0.281	0.187
E	83.84	136.15	38.39	0.327	0.263	0.173
F	42.97	135.45	52.6	0.273	0.380	0.204
G	79.52	98.61	52.25	0.147	0.030	0.049
H	84.32	102.02	40.85	0.060	0.032	0.049
I	89.67	97.2	9.86	0.080	0.008	0.027

. The selected typical grains in the TD sample obey the identical definition as the RD sample. The only remaining three grains belong to the TB orientation and are marked as G through I in Figure 5a. Unfortunately, the H and I grains cannot be indexed by EBSD for further IGMA analysis, because of the severe plastic deformation within grains. Both characteristic IGMA distributions of overall grains in the deformed RD and TD samples exhibit stronger intensity at the <0001> axis (Figure 7 and Figure 8), which means the activation of prismatic <a> slip, because <0001> is a unique Taylor axis for the prismatic <a> slip mode [29,30,31]. The IGMA distributions of the selected individual grains between the TD and RD samples exhibit significant discrepancies. Firstly, the ratio of activated prismatic <a> slip in the TD sample is lower than in the RD sample, regardless of the grain’s crystallographic orientation, both the IGMA and SF_N result supports this viewpoint. Secondly, only partially N (40°–50°) TB orientation grains (F grain in Figure 7 and Figure 8) can exhibit relatively uniform IGMA distribution with a maximum intensity of less than 2.0. A possible explanation for this is the coactivation of various slip modes around different Taylor axis within the deformed grains. Thirdly, the grains that belong to N (40°–50°) TB and TB orientations showed relatively higher intensities of IGMA around the <uvt0>, while the same type of grains in the RD sample showed IGMA distributions around <0001> with a higher intensity. The former distribution pattern was ascribed to coactivation of several slip variants of {11$\stackrel{¯¯}{2}$2} <$\stackrel{¯¯}{1}$$\stackrel{¯¯}{1}$23> mode, as proposed by Chun et al., because the {11$\overline{2}$2} plane is more favored than the {10$\overline{1}$1} plane in hcp metals [30]. The latter distribution pattern led to the activation of basal <a> slip. As for the TB orientation grains (G–I), the SF_N calculated result for the prismatic <a> slip in the RD sample is remarkably larger than the TD sample. More importantly, the calculated maximum SF_N value for the G–I grain in the TD sample was less than 0.1, which means that it was almost impossible to activate the listed slip modes in Table 3.

Based on the combined analyses of IGMA distributions and SF_N calculations in the tensile deformed Zircaloy-4 along the RD and TD, it may be reasonably concluded that a significant discrepancy exists in the N (40°–50°) TB and TB orientation grains. This is because the former tended to induce the basal <a> slip, while the latter tended to induce the coactivation of several slip variants of the second-order pyramidal slip.

By careful inspection of the IPF colored maps of RD and TD samples, evidence of coordination deformation can be observed, as the yellow ellipses show in Figure 4a. The color distribution of grains gradually became the same, which means that synergistic deformation effects exist in tensile Zircaloy-4 along the RD, while similar features did not exist in the tensile sample along the TD. Figure 9 shows the enlarged map of the region of interest region in Figure 4a and the corresponding misorientation angles distribution. It can be seen from the misorientation angles distribution in red and blue lines (Figure 9d–f) that the misorientation angles between 1#, 2#, and 3# grains (red dotted line) increased with an increase in strain. The misorientation angles between 1# and 2# increased slightly, while this value between 2# and 3# experienced a significant drop from 21.77° to 6.81°, indicating the disappearance of HAGB (>10°). However, similar features were not exhibited in the blue dotted line that consisted of 1#, 4#, and 5# grains, indicative of discrepancies compatible with deformation capability among grains. The strain could also be further accommodated by grain rotation, in addition to the activated slip modes [38]. Furthermore, we can infer that there were significant discrepancies in synergistic plastic deformation behavior in Zircaloy-4 alloys under different loading directions. Such anisotropic behavior has also been mentioned in other hcp metals [2,6].

To ensure statistical significance in the slip modes analyzed, SF distributions of various slip modes in the RD and TD samples are shown in Figure 10. Both show that the prismatic <a> slip is the preferable deformation mechanism. Further evidence can be found in Figure 11, which displays the cumulative distribution results of SF_N with different slip modes at 6.64% strain in the RD sample and 6.20% strain in the TD sample. It is readily observed that both samples possess higher fractions of SF_N for the prismatic <a> slip with higher SF_N values (>0.4). This means prismatic <a> slip is the predominant activated slip mode [12,29,37]. Relative frequencies (RD~88.54% and TD~64.29%) and values of SF_N for the prismatic <a> slip are remarkably higher in the RD sample than the TD (Figure 11), implying that most grains are favorably oriented for activating the prismatic slip, and then contribute to the plastic deformation. In addition to the prismatic <a> slip, the basal <a> slip possesses higher SF_N and lower critical resolved shear stress (CRSS) than the pyramidal <a> slip in the RD sample [1]. This further verified the possibility of the activated basal <a> slip evaluated through IGMA. As for the TD sample, it is reasonable to conclude that the plastic slip is partially restricted for its weak compatible deformation capability and higher frequency of pyramidal <a> slip tendency.

The activation of the slip mode in an individual grain is determined by the IGMA method, which, combined with SF_N calculation results, suggests there is the applicability of the IGMA method for slip behavior analysis in deformed Zircaloy-4 at the grain scale. This plays a significant guiding role in the analysis of the micromechanical behavior of local deformed circumstances in Zircaloy-4. These methods are applicable to investigate the plastic deformation at local areas, especially for the circumstances of hydrogen embrittlement, delayed hydrogen cracking, and oxidation corrosion, etc. [39]. In the present study, the quasi-in situ EBSD combined with mechanical tensile tests had limitations to some extent. For example, it was difficult to track the same characterized region when the Zircaloy-4 sample suffered severely from plastic deformation. Furthermore, a certain degree of manual error exists when being re-mounted to the specimen mount. Hence, more advanced and precise characterization methods, such as in situ EBSD, may be applied to obtain the observed microstructure evolution in real-time, and provide detailed experimental results revealing the potential plastic deformation mechanisms of this metal and its alloys.

4. Conclusions

In this study, the anisotropic mechanical behavior and the associated microstructure evolution in Zircaloy-4 along RD/TD tensile conditions were investigated via a quasi-in situ EBSD method. The following conclusions were drawn based on the experimental results and discussions.

(1)

Many shallow dimples and cleavage regions on the fracture surface and a cup-cone fracture morphology were observed. Such features are consistent with the lower ultimate tensile strength ~470 MPa and elongation ~14.5% in deformed tensile Zircaloy-4 along the TD;

(2)

Prismatic <a> slip is the predominant slip mode in Zircaloy-4 under tensile test, regardless of the loading direction;

(3)

The anisotropic slip behavior in individual grains of deformed Zircaloy-4 is attributed to the N (40°–50°) TB and TB orientation grains. The RD sample tends to induce basal <a> slip, while the TD sample tends to induce the coactivation of several slip variants of second-order pyramidal slip modes. Meanwhile, the coactivation of various slip modes only exists within grains belonging to the N (40°–50°) TB orientation;

(4)

The RD sample exhibits excellent compatible deformation capability, not only due to its much higher frequency (88.54%) of soft grains than the TD sample (64.29%), but also due to the synergy deformation among local grains.