Anisotropy of Tensile and Fracture Behavior of Pure Titanium after Hydrostatic Extrusion

Abstract

Commercially pure titanium was subjected to hydrostatic extrusion resulting in formation of an ultra-fine grained microstructure with a strong α-fiber texture and significant improvement of mechanical strength. Anisotropy of the tensile and fracture behavior of the hydrostatically extruded material was studied. It will be demonstrated that the material has significantly higher yield strength along the extrusion direction, while in transversal direction it shows higher work hardening ability related to the α-fiber crystallographic texture. The anisotropy of the fracture behavior in these two directions is less pronounced. A slightly lower fracture initiation toughness and crack growth resistance along the extrusion axis can be related to a lower crack propagation resistance along the boundaries of the elongated grains.

Fig. 5 Engineering stress-engineering strain curves for longitudinal (L) and transversal (T) specimens of the as-received (AR) CP Ti and HE processed CP Ti.

1. Introduction

There is a significant body of research on mechanical properties of ultra-fine grained (UFG) and nanostructured (NS) metallic materials processed via severe plastic deformation (SPD).¹⁾ It has been shown that their yield strength increases with reducing grain size according to the Hall-Petch relation, whereas their tensile ductility shows an opposite trend.²⁾ Numerous strategies have been developed to improve low tensile ductility in these materials, and their detailed description can be found in several review articles.^3–5) The effect of grain refinement on fatigue properties is unambiguous: UFG and NS metallic materials can demonstrate improved or degraded fatigue properties depending on the material, microstructure evolved in the material during SPD processing, and applied fatigue testing conditions, as described in the review article.⁶⁾ Fatigue crack growth anisotropy was reported for single-phase UFG pure Fe⁷⁾ and two-phase pearlitic steel⁸⁾ processed by HPT. The fatigue crack growth behavior can be improved by annealing of the SPD processed nanocrystalline steels.⁹⁾ Fracture properties and behavior still remain as a ‘white spot’, since research activities in this area are limited, despite information about fracture toughness and damage tolerance of any material is of great concern for its structural applications.^10–12) One of the limitation of research activities on this topic is related to the relatively small size of SPD processed samples and the resulting fracture specimens with dimensions often being much smaller than required by standards, such as ASTM or ESIS. However, the compliance with standards is of great importance, especially when the validity of the measurements is based on fulfilling small scale yielding conditions. This would finally also allow a better fracture behavior comparison of SPD processed materials with their coarse-grained counterparts and other commercial alloys available on the market.

Up-to-date, there are only a few studies describing the effect of SPD processing on fracture properties, such as the plain strain fracture toughness K_IC (linear-elastic fracture mechanics), the J-integral and the crack growth resistance (elastic-plastic fracture mechanics). It has been demonstrated that despite UFG pure Ti¹²⁾ and Cu¹³⁾ show lower fracture initiation toughness, their damage tolerance is comparable to that of their coarse-grained counterparts. Stable crack growth resistance was observed in UFG Cu,¹³⁾ whereas the fracture behavior of the UFG pure Ti at room temperature was characterized by instability of crack propagation after relatively small crack extensions due to a significantly lower total crack growth resistance of the material.¹²⁾ Very strong anisotropy of fracture toughness was shown for pure iron processed by ECAP¹⁴⁾ and HPT.¹⁵⁾ Different crack plane orientations led to either crack deflection or delamination, resulting in increased fracture resistance in comparison to one remarkably weak specimen orientation, where intercrystalline failure along the elongated grains occurred. Similar observations were reported for a duplex steel.¹⁶⁾ The weakest testing direction was oriented parallel to the lamellar structure, while other directions showed enhanced values of fracture toughness due to crack deflection and delamination induced toughening effects. The effect of heat treatments on fracture toughness and its anisotropy in the fully pearlitic steel subjected to high pressure torsion (HPT) was studied in Ref. 17). HPT processing led to formation of a nanostructured lamellar aggregate of alternately arranged ferrite and carbon enriched areas, which were aligned with respect to the shear direction leading to a dramatic reduction of its fracture toughness. Tempering at 420°C for 2 h resulted in formation of a bimodal microstructure consisting of coarsened pearlite coexisting next to a necklace-like structure of the original deformation structure. This resulted in an improvement of fracture toughness of the crack propagation direction parallel to the initially aligned nanostructured pearlite up to 3–12 MPa·m^1/2, while the strength remained as high as 2 GPa. Microstructure consisting of spherical cementite particles in an equiaxed ferritic matrix (formed after tempering at 600°C for 2 h) showed higher isotropic fracture toughness with ∼45 MPa·m^1/2 and strength of 1.1 GPa that was attributed to a change of the failure mechanism from intercrystalline to microductile failure. A remarkable fracture toughness in combination with very high strength was also found in the UFG pure Ni processed via HPT, though anisotropy of fracture toughness was also present.¹⁸⁾ Slightly improved fracture toughness was found in the ECAP processed Al 7075 alloy compared to its coarse-grained counterpart in Ref. 19), while ECAP processed Al 6063 alloy showed somewhat decreased fracture toughness.²⁰⁾ Significant improvement of fracture toughness with increasing number of accumulative roll bonding cycles was reported in Ref. 21).

As one can see, the focus of research on the fracture behavior of UFG and NS metallic materials has been on materials mainly processed by ECAP or HPT so far.^10,11) The present work focuses on the effect of hydrostatic extrusion on the fracture properties and their anisotropy of metallic materials applying this process on pure Ti (grade 3). Hydrostatic extrusion is one of the most promising SPD methods for fabrication of UFG CP Ti since it has some advantages compared to other SPD processing techniques.²²⁾ First, very long rods can be processed via hydrostatic extrusion and, second, the method has a very high efficiency due to the very high processing strain rates, usually >10 s⁻¹, but can exceed 10⁴ s⁻¹.²³⁾ Long rods produced via hydrostatic extrusion can be used in various structural application, where their fracture toughness and its anisotropy is of great importance,²³⁾ but their fracture properties have not been investigated yet.

2. Material and Processing

Commercially pure (CP) Ti (grade 3) with specifications corresponding to the ASTM B348-09 standard²⁴⁾ was selected as a material for this investigation. The impurity level (in mass%) was: N-0.05, C-0.08, O-0.35, Fe-0.30, H-0.015. It was received in form of billets having a diameter of 50 mm. The microstructure of the as-received material was homogeneous with the average grain size of 42 µm. Hereafter, this material condition is referred to as AR Ti.

The billets were subjected to hydrostatic extrusion (HE) at room temperature using a 45° die in order to refine their microstructure. The HE processing parameters are listed in Table 1. The material was processed in 4 passes that induced a total strain of 3.24. Hydrostatic extrusion is always accompanied by significant adiabatic heating which might significantly affect the microstructure developed during processing.²⁵⁾ The temperature rise due to adiabatic heating was estimated by (1).^25,26)   

\begin{equation} \Delta T = \beta\frac{p}{c\rho} \end{equation} (1)

where p is the extrusion pressure, c the specific heat, ρ the density, and β denotes the fraction of the plastic work converted into heat during deformation. For HE with good lubrication, β ∼ 0.9.²⁵⁾ The calculated ΔT-value is also given in Table 1. It is demonstrated that the temperature of adiabatic heating in the hydrostatically extruded CP Ti, ΔT ≈ 360°C, is well below the recrystallization temperatures reported in the earlier works.^27,28) Moreover, it should be noted that the samples during HE processing are heated only in the local deformed area just less than 1 s followed by immediate cooling down to room temperature, that is not sufficient neither for recrystallization nor for significant recovery of the developed microstructure.^27,28)

Table 1 Parameters of hydrostatic extrusion of CP Ti.

3. Experimental Procedures

3.1 Microstructural characterization

Microstructure analysis was carried out using a JEOL-200 microscope operating at 200 kV. Observations were made in both bright and dark field imaging modes. The selected area electron diffraction (SAED) patterns were also recorded from the areas of interest. Specimens were cut from the longitudinal plane of the HE processed CP Ti and thinned down to the thickness of ∼100 µm. The Ti specimens were prepared by electropolishing in a TENUPOL 5 twin-jet polisher using a 1:4 solution of nitric acid in methanol at T ∼ −30°C. Texture measurements were performed at the CAI DRX using a Phillips Xpert’PRO diffractometer furnished with a PW3050/60 goniometer. Measurements were taken in a range of Psi angles from 0° to 75° at 3° steps. The pole figures for the planes (0001) and (10–10) were plotted.

3.2 Mechanical characterization

Tensile specimens with a gauge length of 3.2 mm, a gauge width of 0.8 mm and a gauge thickness of ∼0.8 mm were machined from the longitudinal (L) and transversal (T) sections of as-received and HE processed bars. The tensile axis of T-specimens was perpendicular to the bar axis, whereas in L-specimens the tensile axis was parallel to the bar axis. These tensile specimens were mechanically polished to mirror-like surface using colloidal silica solution at the final stage. Tensile tests were carried out at room temperature using a Kammrath & Weiss testing module. Tensile specimens were deformed to failure with constant cross-head speed corresponding to the initial strain rate of 10⁻³ s⁻¹. At least three tensile specimens were tested for each material’s condition and the results were found to be reproducible.

Compact and round compact tension specimens for fracture characterization were extracted with different crack propagation directions from the HE processed bars. The nomenclature of the specimens depends on the designated crack propagation direction and correspond to the letter code of the ASTM. C-R specimens are machined as round compact tension specimens with the circumferential direction, C, as the crack plane normal and the radial direction, R, as the crack propagation direction. R-L specimens have the radial direction as the crack plane normal and the longitudinal direction (L) as the crack propagation direction which is equivalent to the extrusion axis, see Fig. 1. The exact dimensions of both specimen types are presented in Fig. 2.

Fig. 1

Orientation of fracture specimens.

Fig. 2

Drawing of the fracture specimens: (a) R-L specimen; (b) C-R specimen. All dimensions are in mm. Specimen thickness is B = 3.7 mm.

A sharp fatigue pre-crack was introduced before the actual fracture experiment by compression-compression loading using stress-ration of 20 and a ΔK of approximately 12 MPa·m^1/2. Initial crack length was a_o ∼ 3.4 mm (a_o ∼ W/2) (Fig. 2). Fracture tests were performed at room temperature following the ASTM E1820 standard²⁹⁾ as far as possible. At least 3 specimens were tested for each material condition, and the results were found to be reproducible.

3.3 Quantitative characterization of fracture surfaces and experimental determination of the COD- and COA-values

Fracture surface of tested compact tensile specimens was analyzed in a scanning electron microscope (SEM) LEO EVO MA 15. Quantitative analysis of fracture surface was performed using an automatic fracture surface analysis system.^30–32) This system generates a digital elevation model (DEM) of the surface in a stereoscopic image from two SEM images taken by tilting the specimen by 5°. The software automatically finds homologue points in the SEM images taken at two different angles and computes the three-dimensional coordinates of the surface points, thus reconstructing the fracture surface. Fracture surfaces profiles can be extracted from the DEM in order to automatically evaluate the surface roughness parameters and the fractal dimension. The system has been widely used for experimental study of fracture surfaces of diverse metallic materials.^12,13,33) The crack profiles can be drawn over the stereoscopic images of both halves in such a way, that the crack propagation can be reconstructed. An example of crack profiles extracted for the studied material is presented in Section 4.3. From the crack profiles, crack tip opening displacement (CTOD) and crack tip opening angle (CTOA) can be determined, as escribed in Refs. 12, 13, 33). These parameters are measurements of local fracture initiation toughness and total crack growth resistance, respectively. CTOD is estimated by separating the crack profiles from each other until the coalescence point of the first pore with the fatigue pre-crack is reached. The CTOD-value is related to the J-integral (eq. (2)), which characterizes the intensity of elastic-plastic crack-tip field and is understood as the difference of potential energy between two identically loaded specimens with different crack lengths.   

\begin{equation} J = \frac{1}{d_{N}}\sigma_{0}\mathit{CTOD} \end{equation} (2)

In eq. (2), yield stress σ₀ was extracted from tensile testing, and d_N is a dimensionless parameter depending on the strain hardening exponent (n) and σ₀/E ratio,³⁴⁾ where E is the Young's modulus. The strain hardening exponent can be determined through fitting the experimental tensile stress-strain curves by power law $\frac{\sigma }{\sigma _{0}} = ( \frac{\varepsilon }{\varepsilon _{0}} )^{n}$, where ε₀ = σ₀/E.³⁴⁾

The crack opening angle can also be easily measured for the extracted profiles (see Section 4.3). The total crack growth resistance (R_tot) can be calculated as a function of CTOA according to³³⁾   

\begin{equation} R_{\textit{tot}} = \frac{\mathrm{m}\sigma_{\text{y}}\mathrm{b}}{\eta}\mathit{CTOA} \end{equation} (3)

η is a pre-factor to calculate J from the area below load versus load-line displacement curve.³³⁾

In eq. (3), m is estimated according to³³⁾ as   

\begin{equation} m = \frac{\sigma_{\textit{UTS}}}{\sigma_{y}}\frac{\exp(n)}{(1 + n)n^{n}} \end{equation} (4)

where b is the ligament length (b = W − a).

At least 10 profiles were analyzed for each specimen and the average values and their standard deviation were calculated.

4. Results and Discussion

4.1 Effect of HE processing on microstructure and crystallographic texture

CP Ti after HE processing shows a very complex lamellae-type microstructure with long lamellae of width varying in the range from 60 to 500 nm (Fig. 3(a)). Within the lamellar structure, grains and subgrains are again elongated along the extrusion direction and have a width of 100–240 nm and an aspect ratio of 1…2.6 as it can be seen on the TEM image taken at higher magnification (Fig. 3(b)). The grain boundaries are ill-defined and selected area electron diffraction (SAED) patterns are spread. The microstructure is homogeneous over the cross section of the HE processed rods.³⁵⁾ Similar microstructures were reported in previous works focusing on the microstructure evolution during hydrostatic extrusion of CP Ti.^36,37) Thorough microstructural characterization showed that its formation is related to twinning and dislocation glide operating in the material during HE processing.³⁷⁾ There is a body of research showing an α → ω phase transformation during high pressure torsion of pure Ti under pressure of ≥4 GPa.^38,39) It should be noted that no ω phase was found in the HE processed Ti, as much lower pressure (940 MPa) was applied during HE. An earlier study of the HE processed CP Ti also did not reveal the formation of omega phase.³⁷⁾

Fig. 3

Bright field TEM images and SAED pattern of the HE processed Ti. The extrusion direction is marked by white arrows.

Figure 4 illustrates pole figures for the transversal section of the as-received and HE processed Ti bars. The strong α-fiber texture is developed during hydrostatic extrusion with the ⟨10-10⟩ direction and the basal plane (0001) parallel to the rod axis (Fig. 4(b)) from the relatively weak hot rolling texture in the as-received material (Fig. 4(a)). Such fiber texture is typical for CP Ti subjected to extrusion, drawing and/or swaging.⁴⁰⁾

Fig. 4

Pole figures for the transversal section of the as-received (a) and HE processed (b) Ti.

From these microstructure analyses, a significant anisotropy of mechanical properties and deformation behavior in longitudinal and transverse directions can be expected in the material due to the pronounced grain elongation and crystallographic texture as presented in Figs. 3 and 4.

4.2 Effect of HE processing on mechanical properties of the studied materials

Figure 5 illustrates typical engineering stress - engineering strain curves from tensile testing at room temperature for longitudinal and transversal tensile samples of the hydrostatically extruded Ti bars in comparison with the as-received material. Mechanical properties obtained from tensile testing are listed below (Table 2). Despite the as-received CG Ti specimens show a slight anisotropy of mechanical properties, pronounced anisotropy of mechanical behavior is observed after hydrostatic extrusion. It is in a good agreement with the literature data where strong mechanical anisotropy was observed for Ti billets processed via different SPD routes.^41,42) The T-specimens of the CG material tend to show somewhat higher mechanical strength and tensile ductility compared to the L-specimens (Table 2) due to a weak crystallographic texture (Fig. 4(a)). No significant effect of the sample orientation on the work hardening behaviour is observed (Fig. 5). The mechanical strength of CP Ti dramatically increases after hydrostatic extrusion due to grain refinement, whereas both uniform elongation and elongation to failure show an opposite trend. After HE processing, the L-specimens display very high 0.2% proof strength of 915 MPa but low work hardening capacity, so the ultimate tensile strength of the material is 970 MPa. The T-specimens show much lower 0.2% proof strength of 562 MPa, which is higher compared to that of the as-received material though. However, very high work hardening capacity of the T-specimens from the HE processed material leads to a high ultimate tensile strength of 995 MPa, which slightly exceeds that of the L-specimens (Table 2).

Fig. 5

Engineering stress-engineering strain curves for longitudinal (L) and transversal (T) specimens of the as-received (AR) CP Ti and HE processed CP Ti.

Table 2 Mechanical properties of the as-received (AR) and HE processed bars of CP Ti from tensile testing on transversal (T) and longitudinal (L) planes.

This anisotropy in mechanical properties of the extruded CP Ti may be rationalized based on the crystallographic texture developed in the microstructure during processing. The orientation of the slip systems of the metal lattice with respect to the tensile axis determines its deformation behaviour.^40–44) The main dislocation slip mode in CP Ti at room temperature is the prismatic slip system, which is followed in importance by the basal slip system. Dislocation glide on pyramidal systems can also be activated. The operating systems are generally determined by the von Mises criterion, the Schmid factors, and the critical resolved shear stress. In the L-specimens, two prismatic planes and the basal planes are suppressed for dislocation slip, while remaining four prismatic planes remain active (Fig. 4(b)). Therefore, the high yield strength is related to the limited number of slip systems available. In the T-specimens, the c-axis of the HCP lattice of individual grains is randomly inclined with respect to the tensile axis (Fig. 4(b)). Thus, the grains that are most favourably orientated for prismatic and basal slip can be easily deformed at lower values of applied stress. Localization of plastic slip within these grains results in local strain hardening, and, in turn, in an increase of the flow stress, and in the spread of plastic slip to grains that are less favourably oriented for prismatic and basal slip. This scenario leads to overall high work hardening capacity of the T-specimens. It should also be noted that a higher work hardening ability of these T-specimens delays macro-localization of plastic flow resulting in the higher work uniform elongation (∼2.7%) and higher elongation to failure (∼6%) compared to those for the L-specimens (∼0.9% and ∼3.1%) according to the well-known Considère criterion.⁴⁵⁾

In Table 3, data on mechanical properties of CP Ti after SPD processing using ECAP,⁴⁶⁾ ECAP combined with other metalforming techniques,^41,46–48) and HPT⁴⁹⁾ are compared. It is seen that the HE processed CP Ti in the L direction shows much better tensile strength and somewhat lower tensile ductility compared to the ECAP processed counterpart. Combination of ECAP with other processing methods, such as rolling, swaging, drawing (resulting in further grain refinement and formation of strong crystallographic texture) can dramatically improve strength of CP Ti, so it can exceed the strength of the HE processed CP Ti. Nevertheless, from the industrial viewpoint, HE appears as the most preferable processing technique due to its much higher efficiency and low cost compared to the complex SPD processing (i.e. combination of several processing methods).

Table 3 Data on mechanical properties of CP Ti after SPD processing reported in the literature.

4.3 Effect of microstructure and crystallographic texture on fracture toughness

Figure 6 corresponds to SEM images of fracture surfaces of C-R and R-L HE processed CP Ti specimens. The fatigue pre-crack area and fracture surface can be easily identified on the images. The fatigue pre-crack area is characterized by aligned fractographic features parallel to the ED. These aligned fractographic features are perpendicular to the fatigue crack growth direction in the C-R specimen (Fig. 6(a), (b)), and parallel to the fatigue crack growth direction in the R-L specimen (Fig. 6(c), (d)). Such morphology of fatigue pre-crack surface is determined by the microstructure of the HE processed material (see Section 4.1). It is seen that both materials have undergone ductile fracture via formation of voids, their growth and coalescence resulting in formation of a dimple fracture surface. Dimples having a size in the range of 3–15 µm somewhat elongated along the extrusion direction are seen: i.e. they are elongated perpendicular to the crack front in the C-R specimens (Fig. 6(a), (b)) and parallel to the crack front in the R-L specimens (Fig. 6(c), (d)).

Fig. 6

SEM images of fracture surface of both halves for (a, b) C-R specimen, (c, d) R-L specimen from the HE processed CP Ti bars. The extrusion direction is marked by the black arrows.

In Table 4, the fracture parameters estimated using the automatic fracture surface analysis system and calculated by eqs. (2)–(4) are summarized. An example of a single measurement for evaluating CTOD is presented in Fig. 7. A slight anisotropy of the fracture behavior observed in the SEM images is supported by the measures in Table 4. The average CTOD- and CTOA-values for the C-R specimens are higher than those of the R-L specimens (Table). Thus, the average COA- and J_i-values are also somewhat higher for the C-R specimens. It should be noted that in eq. (3), the pre-factor η is independent of specimen orientation. Therefore, a ratio of R_tot values for C-R and R-L specimens can be determined. Simple calculations show that the C-R specimens have the crack growth resistance by a factor of 1.5 higher compared to the R-L specimens.

Table 4 Results of fracture testing of the C-R and R-L specimens for the HE processed CP Ti.

Fig. 7

(a) and (b) SEM images of corresponding regions on both R-L specimen halves of the HE processed CP Ti; (c) Profiles through a fracture surface element at the moment of local failure.

The anisotropic fracture behavior can be related to the microstructure developed in the material during HE processing: the elongated grains/subgrains aligned along the extrusion direction and the lamellae-type microstructure. Since the grains/subgrains are elongated along the extrusion direction, the crack growth resistance is lower in the extrusion direction as the crack can easily grow along the lamellae boundaries, whereas the crack growing in the perpendicular direction has to kink more often. Also, the length of dimples along the direction of crack propagation is larger in the R-L specimens compared to the C-R specimens. This is also supported by the estimated values of J_i and R_tot (Table).

5. Conclusions

Tensile and fracture behavior of commercially pure (CP) Ti after hydrostatic extrusion was studied with respect to its microstructure and texture. It is shown that HE processing results in a complex microstructure consisting of ultrafine lamellae and grains/subgrains slightly elongated along extrusion direction. In addition, a strong α-fiber texture is formed after HE. Grain refinement during HE processing results in significant improvement of mechanical strength and reduction of tensile ductility. Significant anisotropy of tensile mechanical behavior is observed between longitudinal and transversal directions. Longitudinal tensile specimens show significantly higher yield strength, while transversal tensile specimens demonstrate higher work hardening ability due to the α-fiber crystallographic texture. The anisotropy of fracture behavior (crack propagation) in these two directions is also observed though it is less pronounced. The difference in the fracture behavior can be attributed to a lower fracture resistance in the extrusion direction along the boundaries of the lamellar and elongated microstructure.

REFERENCES

• 1)   T.G.  Langdon: Acta Mater. 61 (2013) 7035–7059.
• 2)   N.  Balasubramanian and  T.G.  Langdon: Metall. Mater. Trans. A 47 (2016) 5827–5838.
• 3)   E.  Ma: JOM 58 (2006) 49–53.
• 4)   Y.  Zhao,  Y.T.  Zhu and  E.  Lavernia: Adv. Eng. Mater. 12 (2010) 769–778.
• 5)   I.A.  Ovid’ko and  T.G.  Langdon: Rev. Adv. Mater. Sci. 30 (2012) 103–111.
• 6)   Y.  Estrin and  A.  Vinogradov: Int. J. Fat. 32 (2010) 898–907.
• 7)   T.  Leitner,  A.  Hohenwarter,  W.  Ochensberger and  R.  Pippan: Acta Mater. 126 (2017) 154–165.
• 8)   T.  Leitner,  G.  Trummer,  R.  Pippan and  A.  Hohenwarter: Mater. Sci. Eng. A 710 (2018) 260–270.
• 9)   T.  Leitner,  S.  Pillmeier,  K.S.  Kormout,  R.  Pippan and  A.  Hohenwarter: Scr. Mater. 139 (2017) 39–43.
• 10)   R.  Pippan and  A.  Hohenwarter: Mater. Res. Lett 4 (2016) 127–136.
• 11)   A.  Hohenwarter and  R.  Pippan: Phil. Trans. Royal Soc. A 373 (2015) 20140366.
• 12)   I.  Sabirov,  R.Z.  Valiev,  I.P.  Semenova and  R.  Pippan: Metall. Mater. Trans. A 41 (2010) 727–733.
• 13)   A.  Hohenwarter and  R.  Pippan: Mater. Sci. Eng. A 540 (2012) 89–96.
• 14)   A.  Hohenwarter and  R.  Pippan: Acta Mater. 61 (2013) 2973–2983.
• 15)   A.  Hohenwarter and  R.  Pippan: Mater. Sci. Eng. A 527 (2010) 2649–2656.
• 16)   K.T.  Schwarz,  K.S.  Kormout,  R.  Pippan and  A.  Hohenwarter: Mater. Sci. Eng. A 703 (2017) 173–179.
• 17)   C.  Kammerhofer,  A.  Hohenwarter,  S.  Scheriau,  H.P.  Brantner and  R.  Pippan: Mater. Sci. Eng. A 585 (2013) 190–196.
• 18)   A.  Hohenwarter and  R.  Pippan: Scr. Mater. 64 (2011) 982–985.
• 19)   H.  Darban,  B.  Mohammadi and  F.  Djavanroodi: Eng. Failure Anal. 65 (2016) 1–10.
• 20)   M.A.  Kazemi and  R.  Seifi: Mater. Sci. Eng. A 733 (2018) 71–79.
• 21)   D.  Rahmatabadi,  R.  Hashemi,  B.  Mohammadi and  T.  Shojaee: Mater. Sci. Eng. A 708 (2017) 301–310.
• 22)   W.  Pachla,  M.  Kulczyk,  M.  Sus-Ruszkowska,  A.  Mazur and  K.J.  Kurzydlowski: J. Mater. Process. Technol. 205 (2008) 173–182.
• 23)   M.  Lewandowska and  K.J.  Kurzydlowski: J. Mater. Sci. 43 (2008) 7299–7306.
• 24)  ASTM: Standard specification for Titanium and Titanium alloy bars and billets, ASTM B 348-09, (2013).
• 25)  W. Pachla, M. Kulczyk, A. Świderska-Środa, M. Lewandowska, H. Garbacz, A. Mazur and K.J. Kurzydłowski: Proc. of 9th Int. Conf. on Mat. Forming ESAFORM-2006, (Glasgow, UK, 2006) pp. 535–538.
• 26)   A.I.  Ryazanov,  S.A.  Pavlov and  M.  Kiritani: Mater. Sci. Eng. A 350 (2003) 245–250.
• 27)   K.  Hajizadeh,  S.  Ghobadi Alamdari and  B.  Eghbali: Physica B 417 (2013) 33–38.
• 28)   R.J.  Contieri,  M.  Zanotello and  R.  Caram: Mater. Sci. Eng. A 527 (2010) 3994–4000.
• 29)  ASTM E1820-01, Standard Test Method for Measurement of Fracture Toughness, Annual Book of Standards, (ASTM, Philadelphia, PA, 2001).
• 30)   A.  Tatschl and  O.  Kolednik: Mater. Sci. Eng. A 339 (2003) 265–280.
• 31)   S.  Scherer and  O.  Kolednik: Microscopy Anal. 336 (2001) 15–17.
• 32)   J.  Stampfl,  S.  Scherer,  M.  Gruber and  O.  Kolednik: Appl. Phys. A 63 (1996) 341–346.
• 33)   J.  Stampfl and  O.  Kolednik: Int. J. Fract. 101 (2000) 321–345.
• 34)  C.F. Shin: Tables of HRR Singular Field Quantities, MRL E-147, (Providence, RI, Materials Research Laboratory, Brown University, USA, 1983).
• 35)   W.  Pachla,  M.  Kulczyk,  S.  Przybysz,  J.  Skiba,  K.  Wojciechowski,  M.  Przybysz,  K.  Topolski,  A.  Sobolewski and  M.  Charkiewicz: J. Mater. Process. Tech. 221 (2015) 255–268.
• 36)   J.  Kawałko,  M.  Wroński,  M.  Bieda,  K.  Sztwiertnia,  K.  Wierzbanowski,  D.  Wojtas,  M.  Łagoda,  P.  Ostachowski,  W.  Pachla and  M.  Kulczyk: Mater. Charact. 141 (2018) 19–31.
• 37)   S.  Zherebtsov,  W.  Lojkowski,  A.  Mazur and  G.  Salishchev: Mater. Sci. Eng. A 527 (2010) 5596–5603.
• 38)   K.  Edalati,  E.  Matsubara and  Z.  Horita: Metall. Mater. Trans. A 40 (2009) 2079–2086.
• 39)   M.  Shirooyeh,  J.  Xu and  T.G.  Langdon: Mater. Sci. Eng. A 614 (2014) 223–231.
• 40)   S.  Nemat-Nasser,  W.G.  Guo and  J.Y.  Cheng: Acta Mater. 47 (1999) 3705–3720.
• 41)   I.  Sabirov,  M.T.  Perez-Prado,  J.M.  Molina-Aldareguia,  I.P.  Semenova,  G.  Salimgareeva and  R.Z.  Valiev: Scr. Mater. 64 (2011) 69–72.
• 42)   G.G.  Yapici,  I.  Karaman and  H.J.  Maier: Mater. Sci. Eng. A 434 (2006) 294–302.
• 43)   E.C.  Moreno-Valle,  W.  Pachla,  M.  Kulczyk,  B.  Savoini,  M.A.  Monge,  C.  Ballesteros and  I.  Sabirov: Mater. Sci. Eng. A 588 (2013) 7–13.
• 44)   J.  Gong and  A.J.  Wilkinson: Acta Mater. 57 (2009) 5693–5705.
• 45)  G.E. Dieter: Mechanical Metallurgy, (McGraw-Hill, Boston, MA, 1986).
• 46)   V.V.  Stolyarov,  Y.T.  Zhu,  I.V.  Alexandrov,  T.C.  Lowe and  R.Z.  Valiev: Mater. Sci. Eng. A 299 (2001) 59–67.
• 47)   J.  Gubicza,  Zs.  Fogarassy,  Gy.  Krállics,  J.  Lábár and  T.  Törköly: Mater. Sci. Forum 589 (2008) 99–104.
• 48)   D.V.  Gunderov,  A.V.  Polyakov,  I.P.  Semenova,  G.I.  Raab,  A.A.  Churakova,  E.I.  Gimaltdinova,  J.  Segurado,  V.D.  Sitdikov,  I.V.  Alexandrov,  N.A.  Enikeev and  R.Z.  Valiev: Mater. Sci. Eng. A 562 (2013) 128–136.
• 49)   A.V.  Sergueeva,  V.V.  Stolyarov,  R.Z.  Valiev and  A.K.  Mukherjee: Scr. Mater. 45 (2001) 747–752.