Abstract
To create comprehensive models of mechanical deformation in Zirconium (Zr) it is important to observe the effect of high strain on the material. The mechanical behavior and damage evolution in textured, high-purity zirconium (Zr) is influenced by strain rate, temperature, stress state, grain size, and texture. In particular, texture is known to influence the slip-twinning response of Zr, which directly affects the work hardening behavior at both quasi-static and dynamic strain rates. However, while microstructural and textural evolution of Zr in compression and to relatively low strains in tension has been studied, little is understood about the dynamic, high strain, tensile response of Zr. Here, the influence of texture on the dynamic, tensile, mechanical response of high-purity Zr is correlated with the evolution of the substructure. Experiments will be conducted using dynamic-tensile-extrusion process. A bullet-shaped sample has been impacted into a high-strength steel extrusion die and soft recovered in the Taylor Anvil Facility at Los Alamos National Laboratory. Finite element modeling that employs a continuum level constitutive description of Zr will be performed to provide insight into the dynamic extrusion process. Current experimental findings will be presented.
Introduction
The hexagonal close packed (HCP) metal, zirconium (Zr) lacks the symmetry and isotropy commonly exhibited in cubic materials. This anisotropy creates a unique challenge in modeling the deformation in the material. To develop accurate predictive models, extensive characterization of the mechanical response must be performed.
Statistically understanding the deformation behavior over a wide range of strain rates, temperatures and stress states is key.
Zirconium has been the subject of significant research. A great deal of this work has focused on microstructural evolution due to the deformation dominated by a combination of twinning and slip. Additionally many of these same studies have examined the effect of texture on deformation [1]. However much of this work has been conducted under uni-axial stress conditions. While a few studies such as the one performed by Vogel et al. [2] have macroscopically observed the effects of high pressure on Zr, systematic studies on the effects of high strain and high strain rates on zirconium have not been performed.
In this study, high strain and high strain rates will be established using the dynamic extrusion process developed at Los Alamos National Laboratory. Dynamic extrusion is a method for imparting relatively high strains to specimens that have been accelerated into a steel die at velocities of up to 640m/s using a modified Taylor Gun Facility built at Los Alamos National Laboratory[3].
The study of dynamic extrusion on other materials such as Cu and Ta, has found that considerable grain elongation as well as shear localization leading to void formation accommodates deformation at these strains (Cao et al. [6]). Instability develops and the fragments are extruded from the die during testing.
The first portion of the study will focus on the effects of varying velocity and therefore varying strain and strain rate on ductility of Zr. Given prior understanding of zirconium, both twinning and slip are expected to dominate deformation. Twinning has been found to be vital in deformation of HCP metals as they have fewer slip systems than cubic materials (Song and Gray [4]). Zirconium has three prominent twin systems: (1-102), (11-21), and (11-22), and extensive twinning is expected to be observed in each at high strain rates.
The next portion of the study will focus on the effects of temperature on the dynamic ductility of Zr. Relative differences as a function of temperature and texture will be examined.
Experimental Method
This study examines high-purity, alpha zirconium which has been clock rolled. The average grain size is 35 ^m, as measured by the Heyn method. Using Electron Back Scatter Diffraction (EBSD) a strong 0001 basal texture was observed as shown in Figure 1, which is a characteristic of the plate having been clock rolled during processing. The plate had been annealed at 550°C for one hour and as such the initial dislocation density is very low. Bullet shaped specimens of approximately 0.299" in diameter and 0.310" in length were obtained from the as-received plate.
Figure 1. Pole figure from EBSD scans of as-annealed Zr .
High-speed photography was used to view the in-siu extrusion process on a macroscopic level (Figure 3). The high-speed photographs were utilized to observe the elongation of the material as it exited the die and to reassemble fully extruded and soft recovered segments for post mortem characterization. Additionally, these images can be utilized to determine the exit velocity of the specimens.
Figure 3. High speed photography of Zr specimens extruded in the following test conditions: A) 534 m/s, in-plane direction, 25°C; B) 535 m/s, through thickness direction, 25°C; and C) 602 m/s, through thickness direction, 250°C.
For all tests, the extruded pieces were soft captured and collected. The portion of the bullet that remained caught in the die was removed by cutting the die near the sample to subsequently weaken the die thus allowing the specimen ejection. All the soft recovered pieces as well as to verify that all pieces were captured. Measuring the length of each soft recovered piece and summing these lengths established the total elongation. Three exampled of reassembled, fully extruded segments along with the portion left in the die are shown in Fig. 4.
For many of the tests, each extruded segment along with the die left in the die was examined using scanning electron microscopy (SEM). These pieces were then mounted in epoxy, ground to the mid-section, prepared using standard metallographic preparation techniques, and chemically etched with a solution of 45 ml H2O, 45 ml HNO3, and 10 ml HF for 20-25 seconds. Optical microscopy was then performed on the samples using a microscope equipped with polarized light.
Figure 4. Reassembled soft captured fragments of extruded Zr in the following test conditions: A) 534 m/s, in-plane direction, 25°C; B) 535 m/s, through thickness direction, 25°C; and C) 602 m/s, through thickness direction, 250°C.
Results and Discussion
Dynamic extrusion tests have been performed on both IP and TT specimens at velocities of 400-640 m/s and at room temperature and 250°C. Total elongations for each test are given in Fig. 5.
Figure 5. Elongation (mm) as a function of velocity
Figure 6. SEM image of the fracture surface of an extruded specimen
Although in all tests the metal was not fully extruded, velocity is seen to strongly affect the fragmentation and elongation of the specimen. Total elongation was measured after reassembling the fragments in the extrusion order according to the high-speed images. From Fig. 5 it is evident that IP samples consistently display more elongation than the TT specimens. Additionally, elongation increases with increasing impact velocity regardless of specimen texture and test temperature. Finally, the role of temperature on dynamic extrusion is less clear. However, temperature slightly increased the elongation in the TT specimens. Additionally, more instability is developed with higher velocity. The number of fragments varies depending on the velocity (ex. five segments for IP specimen tested at 479.8 m/s whereas nine segments were recovered for an IP tested at 654.6 m/s).
Scanning electron microscopy (SEM) was used to observe the exterior appearance of the soft recovered segments as well as the tips of the first and last extruded segments to examine the fracture surfaces. This analysis revealed that rather than having failed due to shear as seen in previous studies of cubic materials, the end of the fragments displayed a ductile, fracture surface similar to that seen in tensile tests (Figure 6). !
Optical microscopy was performed on many of the tested specimens. Rather than the expected grain elongation, all samples experienced significant recrystallization in most of the fully extruded segments (Figure 7). However, the initial fully extruded segment and segment remaining the die displayed a range of microstructures as shown in Fig. 7.
Figure 7. Optical images of the IP, 421m/s, 25°C extrusion. The microstructure of the (a) segment left in the die and (b) the first fully extruded piece display deformed grains in region 1, elongated grains in region 2, and recrystallization in region 3.
Electron back scattered diffraction was utilized to examine differences in twinning and evolving texture in IP and TT specimens as a function of test velocity. Twinning was more significant in the TT specimens than in the IP specimens and more twinning was observed with increasing test velocity. The significant differences in twinning result if difference texture evolutions between the TT and IP specimens, as is shown in Figs. 8 and 9. This difference in active deformation mechanism as a function of specimen texture is also likely the reason for the differences in the development of instability as a function of texture and velocity and this likely directly influences elongation of the specimens.
Figure 8. IP case: (a) the undeformed microstructure and texture, (b) the microstructure and local texture in the segment left in the die tested at 25°C and 460m/s, and (c) the microstructure and local texture in the segment left in the die tested at 25°C and 600m/s.
Figure 9. TT case: (a) the undeformed microstructure and texture, (b) the microstructure and local texture in the segment left in the die tested at 25°C and 468m/s, and (c) the microstructure and local texture in the segment left in the die tested at 25°C and 603m/s velocity.
Conclusions
Thus far, we can conclude from this study the following about the dynamic tensile extrusion response of high purity Zr:
1. Impact velocity strongly influences the large-strain tensile ductility of Zr and thus elongation.
2. Impact velocity influences the development of instability and therefore the number of extruded segments during the extrusion process; higher velocities produce more fragments.
3. Differences in the relative activation of twinning as a function of velocity and texture correlated with the observed differences in the development of instability and the total elongation of specimens.
