Extrinsic and intrinsic fracture behavior of high pressure torsion deformed nickel.

Nickel discs (>99.5 wt.%) were deformed by high pressure torsion (HPT) at different temperatures (-196 °C, 25 °C, 200 °C, and 400 °C) until saturation was reached. The strength and fracture behavior of microdefect-free samples and samples with inclusions were investigated using micro and macro tensile tests, respectively. The fracture behavior is not sensitive to the HPT deformation temperature but differs significantly in the two types of sample. The ultimate tensile strength is not affected by inclusions or grain texture.

Severe plastic deformation (SPD) results in beneficial mechanical and physical properties [1-6]. An increase in strength due to SPD has been documented in a vast number of papers, however, the effect on the ductility of SPD materials is not so clear. Hence the fracture behavior of the often used model material pure nickel is investigated here. Micro tensile tests with a gauge length of 15 μm and macro tensile tests with a gauge length of 2.5 mm were carried out. The volume of the micro samples was kept small to reduce the probability of inclusions to almost zero. Therefore these tests should reveal the intrinsic strength and fracture behavior of HPT deformed nickel. Comparing the micro and macro results will help to understand the strength and fracture behavior from the millimeter regime down to the sub micrometer regime, which will be used to better understand the fracture behavior of SPD materials and can help to define critical design parameters or limits to the technical use of these materials.

Nickel (>99.5 wt.%) was HPT deformed at liquid nitrogen temperature (−196 °C), room temperature (RT) (25 °C), 200 °C and 400 °C. The discs (diameter 8 mm, final thickness 0.6 mm) were sheared for three revolutions (speed 0.2 rotations min−1, pressure 4 GPa). The large samples (diameter 30 mm, final thickness 7.5 mm) were deformed at RT for 13 revolutions (speed 0.067 rotations min−1, pressure 3.2 GPa). Further details of the HPT process and the heating and cooling set-up are given in Pippan [7] and Vorhauer and Pippan [8,9]. Vickers microhardness measurements were made with a load of 500 g on the top surface of the discs using a Buehler® Micromet 5104 with an indent spacing of 250 μm. If a saturation microstructure is achieved further straining leads to no change in the hardness and microstructure. The saturation radius marks the smallest radius where the hardness remains constant over the radius [8].

For the convenience of the reader in the following the term “parallel” denotes that the sample tension axis is parallel to the shear plane, whereas “perpendicular” denotes the orientation is perpendicular to the shear plane, i.e. parallel to the torsion axis. Two macro dog bone-shaped tensile samples were machined out of one 8 mm disc, which was first ground and polished to a final thickness of about 0.5 mm. The sample tension axis was parallel to the shear direction with a gauge length of 2.5 mm and a cross-section of 0.5 × 0.5 mm in the middle of the sample, which was slightly dog bone shaped. For the applied turns a saturation radius of 2 mm is required to guarantee a saturation microstructure in the test volume. Details of the tensile geometry are given in Scheriau and Pippan [10]. Tensile tests were carried out at RT using a Kammrath & Weiss® tensile stage. Focused ion beam (FIB) milled micro tensile samples with a gauge length of 15 μm and a cross section of 3 × 3 μm were used to investigate a “microdefect-free” sample, to determine the intrinsic fracture behavior of HPT deformed nickel. The small test volume guarantees an ideal HPT microstructure in the tensile sample without any fracture-relevant microdefects or inclusions. Bars with the dimensions 0.4 × 0.4 × 5–7 mm, parallel and perpendicular to the HPT shear direction at radii of 3 and 14 mm, respectively, were cut, fixed in a sample holder and finally electrochemically etched to a needle. The tensile samples on top of the needles were milled with final currents below 200 pA to reduce the FIB damage. The micro tensile tests were carried out in a scanning electron microscope at RT. Details of the milling process and the tensile test are given in Kiener et al. [11]. To exclude strain rate effects similar strain rates (macro samples 3.0 μm s−1, micro samples 0.018 μm s−1) were used. Fracture surfaces were investigated by scanning electron microscopy (SEM) to determine the reduction in area.

Table 1 gives the saturation hardness according to HPT deformation temperature along with the results of the micro + macro tensile tests. All samples showed a saturation microstructure at a radius >1.8 mm (8 mm disc) and at a radius >3 mm (30 mm disc). The average saturation hardness decreased with increasing HPT deformation temperature, as already shown [9,12]. The same HPT–temperature dependency was observed for the tensile strength. The differences in hardness and tensile strength reflect the differences in the grain sizes generated at the different HPT temperatures. The evaluated grain size depends strongly on the technique used to determine the grain size and due to anisotropy of the grain shape it also depends on the observation direction. In the following, the microstructures are characterized by the HPT temperature and loading direction. For further details regarding grain size evolution and microstructure characteristics see Pippan et al. [7], Hafok and Pippan [13], Hafok et al. [14], and Zhang et al. [15,16].

Table 1

Micro and macro tensile test results and their corresponding standard deviations.

HPT temperature [°C]	Saturation hardness [GPa]	Loading direction [−]	Macro samples				Micro samples
			Ultimate strength [MPa]	Uniform elongation [%]	Fracture stress [MPa]	Fracture elongation [%]	Ultimate strength [MPa]	Uniform elongation [%]
−196	3.90 ± 0.027	Parallel	1455 ± 33	1.8 ± 0.39	1025 ± 41	9.1 ± 0.71	1294 ± 8	2.9 ± 0.11
25	3.13 ± 0.041	Parallelperpend.	1040 ± 15	2.1 ± 0.44	684 ± 19	11.6 ± 0.07	1155 ± 65952 ± 145	2.3 ± 0.651.1 ± 0.58
200	2.29 ± 0.030	Parallel	764 ± 39	0.9 ± 0.19	487 ± 29	11.6 ± 0.91	790
400	2.13 ± 0.233	Parallel	610 ± 127	1.3 ± 0.44	398 ± 85	15.3 ± 1.21	419

As shown in Figure 1, the ultimate tensile strength (UTS) increased with decreasing HPT deformation temperature independent of the test type. The UTS in the micro and macro tests are similar. It seems that the micro samples are slightly stronger at lower HPT deformation temperatures and weaker at higher HPT deformation temperatures compared with the macro samples. In the case of the micro sample deformed at 400 °C this is a consequence of the relatively large grain size compared with the sample dimensions. The average grain size in this case was somewhat below 1 μm and therefore the cross-section contained only about 10 grains. Most of them are surface grains and not as constrained in plastic deformation. A systematic investigation of copper wires revealed that there is a critical ratio of wire thickness to grain size at which the flow stress and UTS significantly decrease [17]. Furthermore, a difference in the range of 10% must always be expected due to the completely different equipment used for the micro and macro tests.

Figure 1

(a) Macro tensile curves for the HPT deformed disc at different deformation temperatures (testing direction parallel to the HPT shear direction). (b) Micro tensile curves for the HPT deformed disc at different deformation temperatures. The tensile test direction for the filled symbols is parallel to the HPT shear direction whereas the tensile test direction is perpendicular to the HPT shear direction for the open symbols.

A strong dependency of the tensile properties on the evolved grain texture during deformation was not observed, as shown in Figure 1b for the micro samples from the HPT samples deformed at RT tested in different loading directions. Complete necking occurred for all micro samples, whereas the macro samples finally fractured after necking. The technical fracture stresses decrease with increasing HPT deformation temperature (Table. 1). The fracture strain of the macro samples was about 10% (plastic strain) and independent of the HPT deformation temperature. Only the macro sample HPT deformed at 400 °C reached a higher value of 15%. The plastic elongation at fracture of the micro samples was about 20% and the technical fracture stress was close to 0.

Figure 2 shows the fracture surfaces of the macro samples (Fig. 2a–d) and of the micro samples (Fig. 2e–h). All macro samples showed cup and cone like fracture and the reduction in area was about 60%, increasing to more than 70% at HPT deformation temperatures above 200 °C, which is an extraordinarily high value for nickel with 1000 MPa strength. Figure 3 shows a detailed SEM investigation of a macro sample which reveals that the large pores with inclusions are mainly situated in the middle of the sample. The large pores are surrounded by a small sized dimpled surface structure without inclusions. In the large pores inclusions between 200 nm and 500 nm (Fig. 3d) are seen. No brittle fracture features were found anywhere on the surface. The micro samples show a complete ductile fracture behavior independent of the HPT deformation temperature and sample orientation (Fig. 2e–h). The square pyramid is a consequence of the square cross-section of the tensile sample. The soft structures visible on the fracture surfaces on these micro samples are carbon contamination caused by SEM observation.

Figure 2

Fracture surface of the micro and macro samples in the parallel and perpendicular orientations: (a) parallel macro sample, HPT at −196 °C; (b) parallel macro sample, HPT at 25 °C; (c) parallel macro sample, HPT at 200 °C; (d) parallel macro sample, HPT at 400 °C; (e) parallel micro sample, HPT at 25 °C; (f) perpendicular micro sample, HPT at 25 °C; (g) side view, parallel micro sample, HPT at 400 °C; (h) parallel micro sample, HPT at 400 °C.

Figure 3

Details of the fracture surface in the middle of a parallel macro sample HPT deformed at 200 °C. Inclusions are visible in some of the pores.

The ratio between saturation hardness and UTS was in the range 2.8–3.5 independent of the test size, which fits well with the Tabor value [18]. The UTS value 5 for the micro tensile sample HPT deformed at 400 °C is a consequence of it containing less than 10 grains in the test cross-section, as explained above. From the fracture surfaces one can estimate that the mean distance between the inclusions is about 20 μm, hence the volume content is on the order of 10–100 p.p.m. This indicates that a very small content of inclusions has a significant affect on the fracture behavior of ultrafine grained and nanocrystalline materials. The UTS is not limited by the presence of inclusions. Furthermore, no significant influence of sample orientation, HPT shear texture [13] or grain shape, taken as the ratio between the longest grain axis and the shortest grain axis, was observed. It is well known that increasing the impurity content or decreasing the HPT deformation temperature increases the aspect ratio [15,16]. However, it seems that the UTS is mainly governed by the average grain size and is independent of the sample tension axis orientation to the HPT shear plane (parallel or perpendicular).

Due to the small dimensions an accurately mounted sample strain measurement system was unavailable for the macro and micro samples. However, such a system would be required to accurately determine, for example, the stress at 0.1% or 0.05% plastic strain. From the shape of the tensile curves it is clear that hardening occurs in all parallel samples in the micro and macro regime. No significant hardening is observed in case of the perpendicular micro sample, which may be an experimental artifact due to pausing to capture the images during the test. As already reported [9,12], an increase in the HPT deformation temperature leads to a decrease in strength due to the larger grains in the saturation microstructure. This relation was observed for both test set-ups, which further indicates that the Hall–Petch effect is the main factor controlling strengthening. The influence of the FIB milling process on the tensile properties of the micro samples is negligible. Only small milling currents and milling parallel to the final surface were used. Small amounts of impurity elements can have a much stronger influence on the final strength as, for example, carbon and silicon [12,19] lead to a change in the saturation grain size. Independent of this impurity influence, HPT nickel has no strong orientation dependency and inclusions do not decrease the UTS, which are very beneficial features in terms of technical uses of SPD processed nickel.

Independent of the sample size, both the micro and macro tests exhibited necking and ductile fracture behavior. Although inclusions have no influence on strength, they play a dominant role in the final fracture behavior. In the absence of inclusions (micro samples) complete necking occurs. Samples with inclusions (macro samples) show pore formation and coalescence, resulting in micro ductile fracture surfaces. A detailed investigation of the fracture surfaces (Fig. 3) reveals two classes of pores, large ones with an approximate diameter of 10 μm and smaller ones with diameters between 300 nm and 1 μm. The large pores are independent of the HPT deformation temperature and the smaller ones are smaller in the fine grained HPT nickel generated at lower HPT deformation temperatures than in the coarser HPT nickel generated at higher HPT deformation temperatures. SEM investigations show that in most of the large pores inclusions with a typical size below 0.5 μm are visible, which seem to be the origin of pore formation. According to these fracteograhic observations the following fracture mechanism in the macro sample is proposed.

During plastic deformation decohesion occurs from the few small inclusions. Pronounced growth of these pores in the necking area occurs due to the strain concentration in the necking region and the change in stress state from uniaxial tension to multiaxial tension. Few of them coalesce. In the rest of the ligament between the pores, small pores are formed at some of the triple junctions due to the high triaxiality and shear localization, which was also proposed by Kumar et al. [20] and Hohenwarter and Pippan [21]. The final fracture is then determined by coalescence of these small pores. No inclusions can be found in these small dimples. The smaller dimples can again be categorized into smaller and larger ones, which are often visible between two large pores. It seems that these somewhat larger small pores grow stably during the final part of the necking process, whereas the smaller ones are generated and grow during final unstable fracture. Figure 4 shows a cross-section of the fracture surface, in which defects beneath the pores are visible. These defects are a consequence of the large plastic deformation and are mainly formed during the final fracture event. Despite the large deformation in the necking region the frequency of such pores is very low. This supports the assumption that additional shear localization and or higher triaxiality are necessary for the formation of such pores at the triple junctions of grains.

Figure 4

Cross-section of a parallel macro fracture surface HPT deformed at 25 °C. The contrast has been increased to enhance the microdefects below the surface. An inclusion on the fracture surface is marked.

Further investigations are necessary to clarify these phenomena in detail.We can summarize as follows.

No sample size dependency on strength was observed.

The UTS is not influenced by the presence of inclusions, as shown on inclusion-free micrometer sized samples and standard macro samples.

The fracture mechanism after necking mainly depends on the presence of inclusions. In the macro sample the fracture mechanism comprises pore formation at inclusions, pore growth and final sub-micro ductile fracture due to the growth of small pores at triple junctions. Ideal ductile fracture was observed for the inclusion-free micro tensile specimens.

Despite the strong effect of HPT deformation temperature on UTS the ductility was not substantially affected.