- Email: [email protected]
Crack path identification in a nanostructured pearlitic steel using atom probe tomography
Scripta Materialia 142 (2018) 66–69
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular article
Crack path identification in a nanostructured pearlitic steel using atom probe tomography T. Leitner a, S. Sackl b, B. Völker a, H. Riedl c, H. Clemens b, R. Pippan d,⁎, A. Hohenwarter a a
Department Materials Physics, Montanuniversität Leoben, Jahnstr. 12, A-8700 Leoben, Austria Department Physical Metallurgy and Materials Testing, Montanuniversität Leoben, Franz-Josef Str. 18, A-8700 Leoben, Austria Institute of Materials Science and Technology, TU Wien, Getreidemarkt 9, A-1060 Wien, Austria d Erich Schmid Institute of Materials Science, Austrian Academy of Sciences, Jahnstr. 12, A-8700 Leoben, Austria b c
a r t i c l e
i n f o
Article history: Received 31 July 2017 Received in revised form 10 August 2017 Accepted 15 August 2017 Available online xxxx Keywords: Severe plastic deformation Pearlitic steels Nanocrystalline Fracture toughness 3-dimensional atom probe tomography
a b s t r a c t Severely plastically deformed pearlitic steels often possess poor crack-growth resistance along the deformationinduced elongated nanolamellar microstructure. However, it is unknown if the crack propagates in the nanocrystalline ferrite or along the ferrite-cementite interface. Here, a pearlitic steel subjected to high pressure torsion exhibiting a fracture toughness of only ~4 MPa·m1/2 along the elongated structure was selected to address this fundamental question. For the first time 3-dimensional atom probe tomography was employed to unravel the local atomistic fractography. We present clear evidence that the low fracture toughness is controlled by crack propagation along the interface between the nanocrystalline carbon-rich and ferritic phase. © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Nanostructured pearlitic steels are widely used in applications, such as the steel cording for tire reinforcement, requiring their unprecedented mechanical properties [1,2]. Their strength is a result of a nanoscaled cementite-ferrite lamellar microstructure, which can be very effectively refined for example through wire drawing. To date, the obtainable strength levels reach almost 7 GPa which is about 30% of the theoretical strength [3]. Besides the unusual high strength, the fracture properties and damage tolerance are remarkable [4]. One main factor leading to these exceptional fracture properties is the anisotropy in the fracture toughness, with relatively “weak” and “strong” crack paths evolving during deformation [4,5]. A weak direction with relatively poor fracture toughness, for example in wire products, is found parallel to the drawing direction and promotes markedly increased fracture toughness values in the strong testing direction. The strong direction is oriented perpendicular to the weak one and triggered by a delamination process dividing the thick specimen during the fracture process into thinner ligaments. The thin ligaments enable even in thick specimens critical stress intensities and, thus, fracture toughness as normally measured for thin specimens. This mechanism has been reported in metallic specimens subjected to traditional deformation processes such as rolling [6,7] and seems to be quite beneficial in high strength materials subjected to various severe plastic deformation processes as well. The crack plane of the delaminations corresponds to the weak crack path ⁎ Corresponding author. E-mail address: [email protected] (R. Pippan).
http://dx.doi.org/10.1016/j.scriptamat.2017.08.025 1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
and, therefore, this fracture toughness enhancing mechanism is primarily controlled by the low toughness crack plane. Focusing on nanostructured pearlitic steels deformed by novel severe plastic deformation processes, this behavior was first studied in a pearlitic steel processed by high pressure torsion (HPT) [8]. Here, the intense shear deformation leads to a strong reduction in the lamella spacing and an alignment of the lamellae parallel to the shearing direction, combined with increased strength levels close to 4 GPa performing compression tests [9]. The fracture toughness in this material state varies between 4 and 40 MPa·m1/2 in the weakest and strongest testing direction, respectively. This large variation is comparable to the difference between structural ceramics and modern high performance steels [10]. Even though the strengthening mechanisms of nanostructured pearlitic steels have been thoroughly scrutinized [3], their fracture behavior and especially the failure mechanism in the weak testing orientation have not been investigated in detail. On the microscopic scale using scanning electron microscopy it was observed that the crack propagates parallel and along the well aligned lamellar structure, irrespective of the used production method. The exact location of the material separation, i.e. the involved interface for the crack propagation process, is unknown. In general, different scenarios are conceivable. The crack may run along the interface between the ferrite and cementite (interlamellar), in the cementite (intralamellar) or in the nanocrystalline ferrite (intralamellar). A clarification of this issue would be fruitful for future work focusing on grain boundary and interface engineering. Strategies could be developed to strengthen the involved crack path and to increase the
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
fracture toughness or even lowering it further to tailor the degree of the fracture toughness anisotropy in a controlled fashion. Traditional fractography and metallographic cross-sections investigated with scanning electron microscopy (SEM) can provide information about the crack path, for example in wire-drawn pearlitic steels, as long as the lamella distance is still large enough to be resolved [11]. For lamella spacings in the nanometer range a clear distinction of the crack path would become impossible with SEM. Energy dispersive spectroscopy is also often used for fractographic investigations, however, with respect to the length scales in nanolamellar steels the electron interaction volume is too large to analyze the top surface of the fracture surface satisfactorily [12]. Advanced in-situ transmission electron microscopy (TEM) studies combined with chemical analyses would grant sufficient resolution, however, the necessity of using ultrathin specimens cannot guarantee the required constraint ahead of the crack tip. Auger-electron spectroscopy possesses exceptional chemical resolution but needs to be supported by complimentary methods or modelling to yield satisfactory quantitative depth-information. In our view, a feasible method to address this issue is accomplished by advanced 3dimensional atom probe tomography (3D APT). This technique is capable of resolving the microstructure and the involved interface at the atomistic scale directly on the fracture surface. In this study 3D APT will be employed for the first time to unravel the local atomistic fractography of a nanocrystalline material. To evaluate this approach and to identify the exact locally occurring crack path during fracture of nanolamellar pearlitic steels, a commercially available fully pearlitic steel was subjected to high pressure torsion using disks of 30 mm diameter and a thickness of approximately 8 mm. The composition of the steel is 0.76 wt% C, 0.35 wt% Si, 1 wt% Mn, 0.014 wt% S, 0.017 wt% P, and Fe in balance. In a former study [8] compact tension C(T) fracture toughness specimens were extracted from different disk-radii of the HPT-disk and orientations, see Fig. 1a, and related to the local imposed pre-strain by calculating the v. Mises equivalent strain. By using this approach it was possible to measure the fracture toughness as a function of pre-strain since the variation of the strain within a single sample is small compared to the investigated deformation regime. For more experimental details regarding the used deformation process and fracture toughness evaluation procedure the reader is referred to the previous work [8]. After imposing an equivalent v. Mises strain of approximately 16, which is realized at a radius of about 13 mm after 3 rotations a strength of nearly 4 GPa was measured [9]. However, a very low fracture toughness of only 4 MPa·m1/2 was found for cracks propagating parallel to the HPT deformation induced elongated and aligned microstructure. This “weak” orientation, see Fig. 1a, shows the lowest fracture toughness in the entire investigated deformation range and the crack propagates parallel to the structural alignment which is at the same time parallel to the tangential direction
67
of the HPT-disk, see Fig. 1a. Due to delamination formation the same material state exhibits a fracture toughness which is 10 times higher in another crack propagation orientation, where the crack propagates perpendicular to the tangential direction and is considered the “strong” orientation (Fig. 1a). The failure mechanism of the weak specimen orientation allowing for the exceptionally high damage tolerance in the strong orientation was the focus of the following APT-study. Fig. 1b shows the corresponding microstructure of the investigated material inspected parallel to the radial direction. The SPD-processing leads to a strongly aligned nanolamellar microstructure parallel to the tangential direction and the lamella spacing is reduced during deformation from about 200 nm of the initial material down to a value in the range of 10 to 20 nm. The thickness of the cementite phase is diminished to only a few nanometers. The fracture surface at this microstructural state, representative for a samples with a fracture toughness of 4 MPa·m1/2 in the weak orientation, is presented in Fig. 1c. The fracture surface is fairly smooth implying that the crack follows the strongly aligned structure (Fig. 1b) and shows no further distinctive fracture characteristics except for some step like features, where the crack locally changes its crack plane along the lamellar structure. To ensure that the pristine fracture surface (Fig. 1c) is accessible in the atom probe, a protective layer was applied to protect the surface during consecutive preparation steps. Furthermore, a site specific specimen preparation is necessary to ensure that the specimen originates from the very top of the fracture surface. In this work different approaches to apply protective layers on the fracture surface before specimen extraction were utilized. In the first approach a platinum layer was deposited on the fracture surface via the gas injection system in the vacuum chamber of a focused ion beam (FIB) Versa 3D from FEI. It was found that an unsatisfactory bonding of this layer to the nanocrystalline pearlite may lead to fracture events during the atom probe measurement. Therefore, a second approach was selected and a chromium layer was deposited by physical vapor deposition (PVD). The thin chromium layer was sputterdeposited in a pure argon atmosphere at room temperature without previous ion etching of the sample to ensure that the fracture surface is not altered in the lab scale deposition system (AJA Orion 5). After depositing the protective layer on the fracture surface, samples for 3D APT were prepared in a dual beam microscope FIB Versa 3D by using a lift-out technique [13], whereby final tip preparation was achieved as described by Felfer et al. [14]. On the final apex of the tips a protective layer of about 100 nm was left to protect the fracture surface before the final measurement. Afterwards the specimens were measured in a Local Electrode Atom Probe (LEAP) 3000 X HR from Cameca Instruments in the laser mode at 60 K. The pulse rate was set to 250 kHz and the laser energy was 0.3 nJ. In Fig. 2 3D APT reconstructions of a specimen coated with Pt (Fig. 2a) and Cr (Fig. 2b) is presented. For the color-code of the atoms
Fig. 1. Features of the investigated material. a) Schematic of the HPT-disk, the C(T) specimens with the weak and strong specimen orientation and the coordinate system with the three principal directions. b) TEM-brightfield micrograph of the nanolamellar steel inspected into the radial direction consisting of a nanocrystalline ferritic and carbon-rich (cementite) phase aligned parallel to the tangential direction. c) Typical fractograph of this microstructure tested in the weak orientation exhibiting a fracture toughness of approximately 4 MPa∙m1/2 with relatively flat fracture surface. The dotted line separates the fatigue pre-crack surface from the final fracture surface.
68
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
Fig. 2. 3D APT reconstructions of the volume underneath the fracture surface of the nanocrystalline pearlite. For clarity only O, Pt, Cr, and C atoms are indicated. The lamellar structure is aligned parallel to the tangential direction similar to the case in Fig. 1b. a) The specimen exhibits a carbon-rich phase on the fracture surface (coated with a platinum protective layer). b) The specimen displays a carbon-free phase at the apex (coated with a chromium protective layer). The concentration profile in Fig. 3b, is taken from the green cylinder and is aligned perpendicular to the cementite lamellae.
the reader is referred to the legend of the figure and to the figure caption. The example of Fig. 2a shows a carbon-rich phase on the fracture surface that was coated with Pt. The presence of the carbon-rich phase at the top of the fracture surface is particularly reflected by the related 1-dimensional concentration profile in Fig. 3a calculated from the reconstruction parallel to the long axis of the specimen within a cylinder having a diameter of ~20 nm. Within the presented region of interest (ROI) Pt from the coating is detectable that rapidly decreases to zero (see inset in Fig. 3a). With decreasing Pt-content the C-content
Fig. 3. 1-dimensional concentration profiles of the 3D APT reconstruction based on Fig. 2a and b.
increases and has a first peak at about 3 nm distance from the top with around 7 at.%. With increasing distance from the top surface concentration peaks of C repeatedly occur with an approximate distance of 10 nm, which corresponds to the typical lamella spacing of the carbon-rich phase varying between 10 and 20 nm according to TEM investigations. The fluctuation in the spacing is caused by the various initial orientations of the pearlitic colonies in the undeformed prematerial. Depending on their initial orientation they were refined to different amounts during HPT deformation. It is important to mention that the measurements also show that the carbon-rich phase does not correspond to the classical cementite phase (Fe3C) that should contain 25 at.% C. This dissolution or decomposition has been described in literature as a supersaturation process transferring C into the ferritic phase promoted by intense deformation. It was found for various SPDprocesses including HPT [15,16] and cold drawing [17,18]. Oxygen is measured at the very top of the sample as well, which is unavoidable as the fracture experiment was performed at air. Only under ultra-high vacuum conditions during the fracture process the deposition of impurities could be avoided, however, only for a limited time as any technical vacuum contains residual gas. With decreasing Ocontent the C-content increases, as mentioned, with the first peak at around 3 nm distance from the top. The presence of oxygen and its distribution is a clear proof that the genuine fracture surface is being analyzed within the tip and not a volume element underneath it. O will cover the fracture surface instantaneously after fracture and serves in this way as a natural marker. The width of the first carbon-rich lamella at the interface is comparable with the thickness of the consecutive ones in the bulk. This is a strong indication for interlamellar fracture. In case of intralamellar fracture (fracture within the carbon-rich lamella) a distinctively thinner carbon-rich lamella is to be expected. This theory is further supported by the second analysis presented in Figs. 2b and 3b. Already the visual inspection of Fig. 2b demonstrates that C was not observed adjacent to the fracture surface, which is again marked by the presence of O. The concentration profile in Fig. 3b, represented by the green cylinder (diameter 15 nm) in Fig. 2b, is aligned perpendicular to the cementite lamellae. The concentration profile shows that with decreasing Cr no enrichment with C is visible. Instead, the Feconcentration strongly rises and the first C-peak occurs approximately 20 nm behind the ferritic phase. The slow rise of the Fe-content can be linked to the 3-dimensional curvature of the fracture surface, see Fig. 2b, so that the region of interest for the calculation of the 1dimensional concentration profile cannot be aligned perfectly perpendicular to the interface. This might be also a contributing factor for the somewhat larger thickness of the first ferrite lamella in comparison with the consecutive ones. From these measurements displaying the absence of carbon adjacent to the Cr and O layer it can be concluded that the ferritic phase forms the fracture surface in this example. A small shortcoming of using Cr for such investigations is the isotopes of Cr and Fe, which share one with the same mass/charge ratio (54Fe and 54 Cr). This is reflected in the raw data by one overlap-peak of Cr and Fe. Since the corresponding concentration of each element forming this joint peak cannot be separated it was excluded from the evaluation process. An inclusion of this peak would suggest a distinctive Cr-content within this steel, which is inconsistent with its actual chemical composition. As a consequence the Cr and Fe-content become slightly underestimated. Nevertheless, this restriction has no impact on the interpretation of the results with the focus on the C-content at the interface. By combining the results of the 3D APT measurements which exemplified carbon-rich and ferritic interfaces it becomes clear that the weak link during the fracture process of the nanolamellar pearlitic steel is the interface between the ferritic and the carbon-rich (formerly cementite) phase, which is called interlamellar fracture. As the fractograph in Fig. 1c suggests, the crack propagates along these interfaces and is occasionally deflected onto other crack planes, possibly due to imperfections in the alignment, which results in step-like features as presented in Fig. 1c. Therefore, on one specimen half both phases, namely the ferritic and
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
carbon-rich phase form the fracture surface and are individually detectable by 3D APT measurements. The occurrence of intralamellar fracture within the ferritic or carbon-rich phase can be principally excluded based on this study. Due to the technological relevance, delamination cracking during twisting of drawn wires is frequently investigated [11,19] and the crack plane of this crack type is comparable to the present case. For such delamination cracks an additional severe deformation near the fracture surface leading to a further local grain refinement was found [19]. In this study the lamellar structure was preserved during crack propagation. This can be explained by the fact that the fracture surface in the present study was subjected to pure Mode-I loading instead of cracks initiated during twisting that are additionally loaded under Mode-II. Secondly, by comparing the lamella spacing of this study with the work of Tanaka et al. [19] a distinctively smaller lamella spacing applies in the present case. This results in a higher strength and markedly smaller plastic zone size during crack propagation. In summary, the present investigations illustrate a strong departure from the classical model characterizing the failure process in coarsestructured pearlitic steels. In the classically coarse-lamellar regime, it is described as a quasi-cleavage fracture process across the pearlitic colonies as proposed by Miller and Smith [20]. In the nanolamellar regime it changes to a pure interface mediated failure type. This conclusion may provide helpful insights into future interface engineering and modelling activities of ultrastrong nanostructured steels. Acknowledgements Funding of this work has been provided by the European Research Council under ERC Grant Agreement No. 340185 USMS and by the
69
Austrian Science Fund (FWF) in the framework of research project P26729-N19. References [1] D. Raabe, P.P. Choi, Y. Li, A. Kostka, X. Sauvage, F. Lecouturier, K. Hono, R. Kirchheim, R. Pippan, D. Embury, MRS Bull. 35 (2010) 982. [2] C. Borchers, R. Kirchheim, Prog. Mater. Sci. 82 (2016) 405. [3] Y. Li, D. Raabe, M. Herbig, P.-P. Choi, S. Goto, A. Kostka, H. Yarita, C. Borchers, R. Kirchheim, Phys. Rev. Lett. 113 (2014) 106104. [4] A. Hohenwarter, B. Völker, M.W. Kapp, Y. Li, S. Goto, D. Raabe, R. Pippan, Sci Rep 6 (2016) 33228. [5] J. Toribio, A. Valiente, Mater. Lett. 58 (2004) 3514. [6] K.T. Venkateswara Rao, W. Yu, R.O. Ritchie, Metall. Trans. A. 20 (1989) 485. [7] R.W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, 4th ed. Wiley, New York, 1996. [8] A. Hohenwarter, A. Taylor, R. Stock, R. Pippan, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 42 (2011) 1609. [9] M.W. Kapp, A. Hohenwarter, S. Wurster, B. Yang, R. Pippan, Acta Mater. 106 (2016) 239. [10] M.F. Ashby, Materials Selection in Mechanical Design, 4th ed. ButterworthHeinemann/Elsevier, Amsterdam, 2010. [11] X. Hu, L. Wang, F. Fang, Z. Ma, Z.H. Xie, J. Jiang, J. Mater. Sci. 48 (2013) 5528. [12] J. Goldstein, D.E. Newbury, D.C. Joy, C.E. Lyman, P. Echlin, E. Lifshin, L. Sawyer, J.R. Michael, Scanning Electron Microscopy and X-ray Microanalysis: Third Edition, 3rd ed. Springer US, New York, 2007. [13] M.K. Miller, K.F. Russell, Ultramicroscopy 107 (2007) 761. [14] P.J. Felfer, T. Alam, S.P. Ringer, J.M. Cairney, Microsc. Res. Tech. 75 (2012) 484. [15] Y. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.-J. Fecht, Acta Mater. 51 (2003) 5555. [16] F. Wetscher, R. Pippan, S. Sturm, F. Kauffmann, C. Scheu, G. Dehm, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 37 (2006) 1963. [17] X. Sauvage, N. Guelton, D. Blavette, Scr. Mater. 46 (2002) 459. [18] Y.J. Lia, P. Choi, C. Borchers, Y.Z. Chen, S. Goto, D. Raabe, R. Kirchheim, Ultramicroscopy 111 (2011) 628. [19] M. Tanaka, H. Saito, M. Yasumaru, K. Higashida, Scr. Mater. 112 (2016) 32. [20] L.E. Miller, G.C. Smith, J. Iron Steel Inst. (London) 208 (1970) 998.
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular article
Crack path identification in a nanostructured pearlitic steel using atom probe tomography T. Leitner a, S. Sackl b, B. Völker a, H. Riedl c, H. Clemens b, R. Pippan d,⁎, A. Hohenwarter a a
Department Materials Physics, Montanuniversität Leoben, Jahnstr. 12, A-8700 Leoben, Austria Department Physical Metallurgy and Materials Testing, Montanuniversität Leoben, Franz-Josef Str. 18, A-8700 Leoben, Austria Institute of Materials Science and Technology, TU Wien, Getreidemarkt 9, A-1060 Wien, Austria d Erich Schmid Institute of Materials Science, Austrian Academy of Sciences, Jahnstr. 12, A-8700 Leoben, Austria b c
a r t i c l e
i n f o
Article history: Received 31 July 2017 Received in revised form 10 August 2017 Accepted 15 August 2017 Available online xxxx Keywords: Severe plastic deformation Pearlitic steels Nanocrystalline Fracture toughness 3-dimensional atom probe tomography
a b s t r a c t Severely plastically deformed pearlitic steels often possess poor crack-growth resistance along the deformationinduced elongated nanolamellar microstructure. However, it is unknown if the crack propagates in the nanocrystalline ferrite or along the ferrite-cementite interface. Here, a pearlitic steel subjected to high pressure torsion exhibiting a fracture toughness of only ~4 MPa·m1/2 along the elongated structure was selected to address this fundamental question. For the first time 3-dimensional atom probe tomography was employed to unravel the local atomistic fractography. We present clear evidence that the low fracture toughness is controlled by crack propagation along the interface between the nanocrystalline carbon-rich and ferritic phase. © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Nanostructured pearlitic steels are widely used in applications, such as the steel cording for tire reinforcement, requiring their unprecedented mechanical properties [1,2]. Their strength is a result of a nanoscaled cementite-ferrite lamellar microstructure, which can be very effectively refined for example through wire drawing. To date, the obtainable strength levels reach almost 7 GPa which is about 30% of the theoretical strength [3]. Besides the unusual high strength, the fracture properties and damage tolerance are remarkable [4]. One main factor leading to these exceptional fracture properties is the anisotropy in the fracture toughness, with relatively “weak” and “strong” crack paths evolving during deformation [4,5]. A weak direction with relatively poor fracture toughness, for example in wire products, is found parallel to the drawing direction and promotes markedly increased fracture toughness values in the strong testing direction. The strong direction is oriented perpendicular to the weak one and triggered by a delamination process dividing the thick specimen during the fracture process into thinner ligaments. The thin ligaments enable even in thick specimens critical stress intensities and, thus, fracture toughness as normally measured for thin specimens. This mechanism has been reported in metallic specimens subjected to traditional deformation processes such as rolling [6,7] and seems to be quite beneficial in high strength materials subjected to various severe plastic deformation processes as well. The crack plane of the delaminations corresponds to the weak crack path ⁎ Corresponding author. E-mail address: [email protected] (R. Pippan).
http://dx.doi.org/10.1016/j.scriptamat.2017.08.025 1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
and, therefore, this fracture toughness enhancing mechanism is primarily controlled by the low toughness crack plane. Focusing on nanostructured pearlitic steels deformed by novel severe plastic deformation processes, this behavior was first studied in a pearlitic steel processed by high pressure torsion (HPT) [8]. Here, the intense shear deformation leads to a strong reduction in the lamella spacing and an alignment of the lamellae parallel to the shearing direction, combined with increased strength levels close to 4 GPa performing compression tests [9]. The fracture toughness in this material state varies between 4 and 40 MPa·m1/2 in the weakest and strongest testing direction, respectively. This large variation is comparable to the difference between structural ceramics and modern high performance steels [10]. Even though the strengthening mechanisms of nanostructured pearlitic steels have been thoroughly scrutinized [3], their fracture behavior and especially the failure mechanism in the weak testing orientation have not been investigated in detail. On the microscopic scale using scanning electron microscopy it was observed that the crack propagates parallel and along the well aligned lamellar structure, irrespective of the used production method. The exact location of the material separation, i.e. the involved interface for the crack propagation process, is unknown. In general, different scenarios are conceivable. The crack may run along the interface between the ferrite and cementite (interlamellar), in the cementite (intralamellar) or in the nanocrystalline ferrite (intralamellar). A clarification of this issue would be fruitful for future work focusing on grain boundary and interface engineering. Strategies could be developed to strengthen the involved crack path and to increase the
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
fracture toughness or even lowering it further to tailor the degree of the fracture toughness anisotropy in a controlled fashion. Traditional fractography and metallographic cross-sections investigated with scanning electron microscopy (SEM) can provide information about the crack path, for example in wire-drawn pearlitic steels, as long as the lamella distance is still large enough to be resolved [11]. For lamella spacings in the nanometer range a clear distinction of the crack path would become impossible with SEM. Energy dispersive spectroscopy is also often used for fractographic investigations, however, with respect to the length scales in nanolamellar steels the electron interaction volume is too large to analyze the top surface of the fracture surface satisfactorily [12]. Advanced in-situ transmission electron microscopy (TEM) studies combined with chemical analyses would grant sufficient resolution, however, the necessity of using ultrathin specimens cannot guarantee the required constraint ahead of the crack tip. Auger-electron spectroscopy possesses exceptional chemical resolution but needs to be supported by complimentary methods or modelling to yield satisfactory quantitative depth-information. In our view, a feasible method to address this issue is accomplished by advanced 3dimensional atom probe tomography (3D APT). This technique is capable of resolving the microstructure and the involved interface at the atomistic scale directly on the fracture surface. In this study 3D APT will be employed for the first time to unravel the local atomistic fractography of a nanocrystalline material. To evaluate this approach and to identify the exact locally occurring crack path during fracture of nanolamellar pearlitic steels, a commercially available fully pearlitic steel was subjected to high pressure torsion using disks of 30 mm diameter and a thickness of approximately 8 mm. The composition of the steel is 0.76 wt% C, 0.35 wt% Si, 1 wt% Mn, 0.014 wt% S, 0.017 wt% P, and Fe in balance. In a former study [8] compact tension C(T) fracture toughness specimens were extracted from different disk-radii of the HPT-disk and orientations, see Fig. 1a, and related to the local imposed pre-strain by calculating the v. Mises equivalent strain. By using this approach it was possible to measure the fracture toughness as a function of pre-strain since the variation of the strain within a single sample is small compared to the investigated deformation regime. For more experimental details regarding the used deformation process and fracture toughness evaluation procedure the reader is referred to the previous work [8]. After imposing an equivalent v. Mises strain of approximately 16, which is realized at a radius of about 13 mm after 3 rotations a strength of nearly 4 GPa was measured [9]. However, a very low fracture toughness of only 4 MPa·m1/2 was found for cracks propagating parallel to the HPT deformation induced elongated and aligned microstructure. This “weak” orientation, see Fig. 1a, shows the lowest fracture toughness in the entire investigated deformation range and the crack propagates parallel to the structural alignment which is at the same time parallel to the tangential direction
67
of the HPT-disk, see Fig. 1a. Due to delamination formation the same material state exhibits a fracture toughness which is 10 times higher in another crack propagation orientation, where the crack propagates perpendicular to the tangential direction and is considered the “strong” orientation (Fig. 1a). The failure mechanism of the weak specimen orientation allowing for the exceptionally high damage tolerance in the strong orientation was the focus of the following APT-study. Fig. 1b shows the corresponding microstructure of the investigated material inspected parallel to the radial direction. The SPD-processing leads to a strongly aligned nanolamellar microstructure parallel to the tangential direction and the lamella spacing is reduced during deformation from about 200 nm of the initial material down to a value in the range of 10 to 20 nm. The thickness of the cementite phase is diminished to only a few nanometers. The fracture surface at this microstructural state, representative for a samples with a fracture toughness of 4 MPa·m1/2 in the weak orientation, is presented in Fig. 1c. The fracture surface is fairly smooth implying that the crack follows the strongly aligned structure (Fig. 1b) and shows no further distinctive fracture characteristics except for some step like features, where the crack locally changes its crack plane along the lamellar structure. To ensure that the pristine fracture surface (Fig. 1c) is accessible in the atom probe, a protective layer was applied to protect the surface during consecutive preparation steps. Furthermore, a site specific specimen preparation is necessary to ensure that the specimen originates from the very top of the fracture surface. In this work different approaches to apply protective layers on the fracture surface before specimen extraction were utilized. In the first approach a platinum layer was deposited on the fracture surface via the gas injection system in the vacuum chamber of a focused ion beam (FIB) Versa 3D from FEI. It was found that an unsatisfactory bonding of this layer to the nanocrystalline pearlite may lead to fracture events during the atom probe measurement. Therefore, a second approach was selected and a chromium layer was deposited by physical vapor deposition (PVD). The thin chromium layer was sputterdeposited in a pure argon atmosphere at room temperature without previous ion etching of the sample to ensure that the fracture surface is not altered in the lab scale deposition system (AJA Orion 5). After depositing the protective layer on the fracture surface, samples for 3D APT were prepared in a dual beam microscope FIB Versa 3D by using a lift-out technique [13], whereby final tip preparation was achieved as described by Felfer et al. [14]. On the final apex of the tips a protective layer of about 100 nm was left to protect the fracture surface before the final measurement. Afterwards the specimens were measured in a Local Electrode Atom Probe (LEAP) 3000 X HR from Cameca Instruments in the laser mode at 60 K. The pulse rate was set to 250 kHz and the laser energy was 0.3 nJ. In Fig. 2 3D APT reconstructions of a specimen coated with Pt (Fig. 2a) and Cr (Fig. 2b) is presented. For the color-code of the atoms
Fig. 1. Features of the investigated material. a) Schematic of the HPT-disk, the C(T) specimens with the weak and strong specimen orientation and the coordinate system with the three principal directions. b) TEM-brightfield micrograph of the nanolamellar steel inspected into the radial direction consisting of a nanocrystalline ferritic and carbon-rich (cementite) phase aligned parallel to the tangential direction. c) Typical fractograph of this microstructure tested in the weak orientation exhibiting a fracture toughness of approximately 4 MPa∙m1/2 with relatively flat fracture surface. The dotted line separates the fatigue pre-crack surface from the final fracture surface.
68
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
Fig. 2. 3D APT reconstructions of the volume underneath the fracture surface of the nanocrystalline pearlite. For clarity only O, Pt, Cr, and C atoms are indicated. The lamellar structure is aligned parallel to the tangential direction similar to the case in Fig. 1b. a) The specimen exhibits a carbon-rich phase on the fracture surface (coated with a platinum protective layer). b) The specimen displays a carbon-free phase at the apex (coated with a chromium protective layer). The concentration profile in Fig. 3b, is taken from the green cylinder and is aligned perpendicular to the cementite lamellae.
the reader is referred to the legend of the figure and to the figure caption. The example of Fig. 2a shows a carbon-rich phase on the fracture surface that was coated with Pt. The presence of the carbon-rich phase at the top of the fracture surface is particularly reflected by the related 1-dimensional concentration profile in Fig. 3a calculated from the reconstruction parallel to the long axis of the specimen within a cylinder having a diameter of ~20 nm. Within the presented region of interest (ROI) Pt from the coating is detectable that rapidly decreases to zero (see inset in Fig. 3a). With decreasing Pt-content the C-content
Fig. 3. 1-dimensional concentration profiles of the 3D APT reconstruction based on Fig. 2a and b.
increases and has a first peak at about 3 nm distance from the top with around 7 at.%. With increasing distance from the top surface concentration peaks of C repeatedly occur with an approximate distance of 10 nm, which corresponds to the typical lamella spacing of the carbon-rich phase varying between 10 and 20 nm according to TEM investigations. The fluctuation in the spacing is caused by the various initial orientations of the pearlitic colonies in the undeformed prematerial. Depending on their initial orientation they were refined to different amounts during HPT deformation. It is important to mention that the measurements also show that the carbon-rich phase does not correspond to the classical cementite phase (Fe3C) that should contain 25 at.% C. This dissolution or decomposition has been described in literature as a supersaturation process transferring C into the ferritic phase promoted by intense deformation. It was found for various SPDprocesses including HPT [15,16] and cold drawing [17,18]. Oxygen is measured at the very top of the sample as well, which is unavoidable as the fracture experiment was performed at air. Only under ultra-high vacuum conditions during the fracture process the deposition of impurities could be avoided, however, only for a limited time as any technical vacuum contains residual gas. With decreasing Ocontent the C-content increases, as mentioned, with the first peak at around 3 nm distance from the top. The presence of oxygen and its distribution is a clear proof that the genuine fracture surface is being analyzed within the tip and not a volume element underneath it. O will cover the fracture surface instantaneously after fracture and serves in this way as a natural marker. The width of the first carbon-rich lamella at the interface is comparable with the thickness of the consecutive ones in the bulk. This is a strong indication for interlamellar fracture. In case of intralamellar fracture (fracture within the carbon-rich lamella) a distinctively thinner carbon-rich lamella is to be expected. This theory is further supported by the second analysis presented in Figs. 2b and 3b. Already the visual inspection of Fig. 2b demonstrates that C was not observed adjacent to the fracture surface, which is again marked by the presence of O. The concentration profile in Fig. 3b, represented by the green cylinder (diameter 15 nm) in Fig. 2b, is aligned perpendicular to the cementite lamellae. The concentration profile shows that with decreasing Cr no enrichment with C is visible. Instead, the Feconcentration strongly rises and the first C-peak occurs approximately 20 nm behind the ferritic phase. The slow rise of the Fe-content can be linked to the 3-dimensional curvature of the fracture surface, see Fig. 2b, so that the region of interest for the calculation of the 1dimensional concentration profile cannot be aligned perfectly perpendicular to the interface. This might be also a contributing factor for the somewhat larger thickness of the first ferrite lamella in comparison with the consecutive ones. From these measurements displaying the absence of carbon adjacent to the Cr and O layer it can be concluded that the ferritic phase forms the fracture surface in this example. A small shortcoming of using Cr for such investigations is the isotopes of Cr and Fe, which share one with the same mass/charge ratio (54Fe and 54 Cr). This is reflected in the raw data by one overlap-peak of Cr and Fe. Since the corresponding concentration of each element forming this joint peak cannot be separated it was excluded from the evaluation process. An inclusion of this peak would suggest a distinctive Cr-content within this steel, which is inconsistent with its actual chemical composition. As a consequence the Cr and Fe-content become slightly underestimated. Nevertheless, this restriction has no impact on the interpretation of the results with the focus on the C-content at the interface. By combining the results of the 3D APT measurements which exemplified carbon-rich and ferritic interfaces it becomes clear that the weak link during the fracture process of the nanolamellar pearlitic steel is the interface between the ferritic and the carbon-rich (formerly cementite) phase, which is called interlamellar fracture. As the fractograph in Fig. 1c suggests, the crack propagates along these interfaces and is occasionally deflected onto other crack planes, possibly due to imperfections in the alignment, which results in step-like features as presented in Fig. 1c. Therefore, on one specimen half both phases, namely the ferritic and
T. Leitner et al. / Scripta Materialia 142 (2018) 66–69
carbon-rich phase form the fracture surface and are individually detectable by 3D APT measurements. The occurrence of intralamellar fracture within the ferritic or carbon-rich phase can be principally excluded based on this study. Due to the technological relevance, delamination cracking during twisting of drawn wires is frequently investigated [11,19] and the crack plane of this crack type is comparable to the present case. For such delamination cracks an additional severe deformation near the fracture surface leading to a further local grain refinement was found [19]. In this study the lamellar structure was preserved during crack propagation. This can be explained by the fact that the fracture surface in the present study was subjected to pure Mode-I loading instead of cracks initiated during twisting that are additionally loaded under Mode-II. Secondly, by comparing the lamella spacing of this study with the work of Tanaka et al. [19] a distinctively smaller lamella spacing applies in the present case. This results in a higher strength and markedly smaller plastic zone size during crack propagation. In summary, the present investigations illustrate a strong departure from the classical model characterizing the failure process in coarsestructured pearlitic steels. In the classically coarse-lamellar regime, it is described as a quasi-cleavage fracture process across the pearlitic colonies as proposed by Miller and Smith [20]. In the nanolamellar regime it changes to a pure interface mediated failure type. This conclusion may provide helpful insights into future interface engineering and modelling activities of ultrastrong nanostructured steels. Acknowledgements Funding of this work has been provided by the European Research Council under ERC Grant Agreement No. 340185 USMS and by the
69
Austrian Science Fund (FWF) in the framework of research project P26729-N19. References [1] D. Raabe, P.P. Choi, Y. Li, A. Kostka, X. Sauvage, F. Lecouturier, K. Hono, R. Kirchheim, R. Pippan, D. Embury, MRS Bull. 35 (2010) 982. [2] C. Borchers, R. Kirchheim, Prog. Mater. Sci. 82 (2016) 405. [3] Y. Li, D. Raabe, M. Herbig, P.-P. Choi, S. Goto, A. Kostka, H. Yarita, C. Borchers, R. Kirchheim, Phys. Rev. Lett. 113 (2014) 106104. [4] A. Hohenwarter, B. Völker, M.W. Kapp, Y. Li, S. Goto, D. Raabe, R. Pippan, Sci Rep 6 (2016) 33228. [5] J. Toribio, A. Valiente, Mater. Lett. 58 (2004) 3514. [6] K.T. Venkateswara Rao, W. Yu, R.O. Ritchie, Metall. Trans. A. 20 (1989) 485. [7] R.W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials, 4th ed. Wiley, New York, 1996. [8] A. Hohenwarter, A. Taylor, R. Stock, R. Pippan, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 42 (2011) 1609. [9] M.W. Kapp, A. Hohenwarter, S. Wurster, B. Yang, R. Pippan, Acta Mater. 106 (2016) 239. [10] M.F. Ashby, Materials Selection in Mechanical Design, 4th ed. ButterworthHeinemann/Elsevier, Amsterdam, 2010. [11] X. Hu, L. Wang, F. Fang, Z. Ma, Z.H. Xie, J. Jiang, J. Mater. Sci. 48 (2013) 5528. [12] J. Goldstein, D.E. Newbury, D.C. Joy, C.E. Lyman, P. Echlin, E. Lifshin, L. Sawyer, J.R. Michael, Scanning Electron Microscopy and X-ray Microanalysis: Third Edition, 3rd ed. Springer US, New York, 2007. [13] M.K. Miller, K.F. Russell, Ultramicroscopy 107 (2007) 761. [14] P.J. Felfer, T. Alam, S.P. Ringer, J.M. Cairney, Microsc. Res. Tech. 75 (2012) 484. [15] Y. Ivanisenko, W. Lojkowski, R.Z. Valiev, H.-J. Fecht, Acta Mater. 51 (2003) 5555. [16] F. Wetscher, R. Pippan, S. Sturm, F. Kauffmann, C. Scheu, G. Dehm, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 37 (2006) 1963. [17] X. Sauvage, N. Guelton, D. Blavette, Scr. Mater. 46 (2002) 459. [18] Y.J. Lia, P. Choi, C. Borchers, Y.Z. Chen, S. Goto, D. Raabe, R. Kirchheim, Ultramicroscopy 111 (2011) 628. [19] M. Tanaka, H. Saito, M. Yasumaru, K. Higashida, Scr. Mater. 112 (2016) 32. [20] L.E. Miller, G.C. Smith, J. Iron Steel Inst. (London) 208 (1970) 998.