Fracture Properties of a Refractory High- Entropy Alloy: in Situ Micro- Cantilever and Atom Probe Tomography Studies

Home , Fracture toughness

TSpace Research Repository tspace.library.utoronto.ca Fracture properties of a refractory high- entropy alloy: In situ micro- cantilever and atom probe tomography studies

Y. Zou, P. Okle, H. Yu, T. Sumigawa, T. Kitamura, S. Maiti, W. Steurerd and R. Spolenaka

Version Post-print/accepted Manuscript

Citation Zou, Y., et al. "Fracture properties of a refractory high-entropy alloy: In (published version) situ micro-cantilever and atom probe tomography studies." Scripta Materialia 128 (2017): 95-99.

Publisher’s Statement The article has been published in final form at [10.1016/j.scriptamat.2016.09.036]

Copyright / License This work is licensed under the Creative Commons Attribution-NonCommercial- NoDerivatives 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by- nc-nd/4.0/.

How to cite TSpace items

Always cite the published version, so the author(s) will receive recognition through services that track citation counts, e.g. Scopus. If you need to cite the page number of the author manuscript from TSpace because you cannot access the published version, then cite the TSpace version in addition to the published version using the permanent URI (handle) found on the record page.

This article was made openly accessible by U of T Faculty. Please tell us how this access benefits you. Your story matters.

Fracture properties of a refractory high-entropy alloy: In situ micro-cantilever and atom probe

tomography studies

Y. Zou,a,b,* P. Okle,a H. Yu,b,c T. Sumigawa,b T. Kitamura,b S. Maiti,d W. Steurerd and R. Spolenaka,*

aLaboratory for Nanometallurgy, Department of Materials, ETH Zürich, Vladimir-Prelog-Weg 5, CH-

8093 Zurich, Switzerland

bDepartment of Mechanical Engineering and Science, Graduate School of Engineering, Kyoto University,

Nishikyo-ku, Kyoto, 615-8540, Japan

cInstitute of Applied Mathematics, Harbin Institute of Technology, Harbin 150001, China

dDepartment of Materials, ETH Zürich, Leopold-Ruzicka-Weg 4, CH-8093 Zurich, Switzerland

Abstract

Most refractory high-entropy alloys (HEAs) are brittle and suffer from limited formability at ambient temperature. Previous studies imply that grain boundaries affect their fracture behavior, but quantitative studies on the fracture properties of body-centered-cubic HEAs are scarce. Here, using in situ micro- cantilever tests, we show that the fracture toughness of a bi-crystal HEA, Nb25Mo25Ta25W25, is one order of magnitude lower than that of single crystalline ones. Atom probe tomography of the bi-crystal HEA reveals element segregation and formation of oxides and nitrides at grain boundaries, suggesting that minimizing grain boundary segregation is critical to improving fracture properties in refractory HEAs.

Keywords: high-entropy alloys; refractory metals; micromechanics; fracture toughness; atom probe

tomography

*Correspondence should be addressed to [email protected] (R. S.) or [email protected] (Y.Z., current address: Department of Mechanical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, USA)

Refractory high-entropy alloys (HEAs) are emerging metallic systems that consist of four or more equiatomic refractory elements (e.g., Nb, Mo, Ta, W, and V) and tend to form single solid-solution-like body-centered cubic (bcc) phases with a strong tendency to solid solution strengthening [1-7]. For the last five years, such alloys have attracted significant attention, because of their superior and useful properties at elevated temperatures (above ~1100 °C): high strength and hardness [2, 8], enhanced thermal and microstructural stability [9, 10], good oxidation resistance [7, 11], etc. A vast majority of them are, however, extremely brittle at room temperature and suffer from poor ductility, rendering them difficult to further

process and use [2, 6, 12, 13]. For example, two typical refractory HEAs (i.e., Nb25Mo25Ta25W25 and

V20Nb20Mo20Ta20W20) have been found to exhibit good ductility above 600 °C, but at room temperature,

they fail easily by cracking at low compressive strains (~2.1% and ~1.7%, respectively) [12]. In a recent study, we compressed Nb25Mo25Ta25W25 HEA micro-pillar samples and observed that the micro-pillars with

a grain boundary (GB) fractured along the GB at a much lower strain than single crystalline (SC) HEA

micro-pillars, which can bear large-strain bending (> 75°) without any fracture [6]. Our previous observation implies that the refractory HEAs might not be intrinsically brittle, but it is their GBs that significantly reduce their ductility. Thus, two questions arise: how much does a GB influence the fracture properties (i.e., fracture toughness and fracture strength) in a refractory HEA? And what characteristics of the GBs induce the brittleness of the HEAs?

Although face-centered cubic (fcc) HEAs, such as CrMnFeCoNi, have been reported to exhibit promising fracture resistance [14], the inadequate fracture-resistance property of bcc HEAs is, in fact, a bottleneck that limits their usage. So far, to the authors’ knowledge, study of the fracture properties of bcc HEAs is still lacking. The main obstacle to measuring their fracture toughness is that the materials are generally too brittle to be fabricated as standard macroscopic fracture toughness specimens (as illustrated in supplementary information Fig. S1). In this work, we have applied the methodology of in situ micro- cantilever fracture tests [15] to study the fracture behavior of SC Nb25Mo25Ta25W25 HEAs and bi-crystal

(BC) HEAs (i.e., those containing a GB) and evaluate the effect of a GB on the fracture properties of

2 refractory HEAs. Using this technique, combined with finite element method (FEM) simulations, we are able to calculate the fracture toughness and strengths of the HEA micro-cantilevers and also compare them with other reported micrometer-sized materials, such as ceramics [16, 17], intermetallics [18, 19], and metals [20, 21].

A bulk Nb25Mo25Ta25W25 HEA was produced using the arc melting technique in an argon atmosphere and

then homogenized at 1800 °C for seven days (as described in [6]). The crystal orientations of the as-prepared

HEA specimen were characterized using the electron back-scatter diffraction (EBSD) technique (Figure

1a). The specimen shows an equiaxed grain microstructure with grain sizes of a few hundred microns. SC- and BC-cantilevers were fabricated from two adjacent <110>-oriented grains (misorientation angle <5°) using the focused ion beam (FIB) technique (Hitachi, FB-2200), as shown in Figures 1b and 1c, respectively.

Both SC- and BC-cantilever beams have a length (L0) of ~6-8 µm, a width (W) and thickness (B) of ~1.5-2

µm, as schematically illustrated in Figure 1d. A notch with a depth of ~0.3-0.5 µm and a tip radius of ~10

nm was fabricated in each cantilever using a fine milling current (5 kV, 5 pA) and a milling time of ~5-10

seconds. In SC-cantilevers, the notches were close to the cantilever beam support and parallel to {100}

planes (primary cleavage planes for bcc metals). In BC-cantilevers, the notches were cut along the GBs.

The micro-cantilever specimens were mounted in an indenter holder (Nanofactory Instruments AB,

SA2000N), which was fitted to a transmission electron microscope (TEM, JEOL JEM-2100). A sharp diamond tip was used to load at the beam (close to the free end) in a displacement control mode (2 nm/s) by feedback mechanism (the details of the loading apparatus were described in [22]). Four specimens were

measured for both SC- and BC-cantilevers. A force-displacement curve during the fracture process was

recorded.

Figures 2a-c and 2e-g present snapshots from movies of typical SC- and BC-cantilevers upon loading,

respectively. Figures 2d and 2h show their corresponding load–displacement curves. The SC-cantilever

exhibits a linear elastic behavior at the initial loading stage (between a and b in Figure 2d), a slight yielding

before reaching the maximum load, and a subsequent gradual force drop. In contrary to the SC-cantilevers, all the BC-cantilevers experienced a catastrophic event at the maximum load. They did not show any plastic

yielding before fracture, and the crack tips suddenly opened and advanced along the GBs (as shown in Fig.

2g), indicating that the BC-specimens are more brittle than the SC ones. After in situ cantilever tests, we characterized the fracture surfaces using a scanning electron microscope (SEM). The two types of

cantilevers reveal distinct surface morphologies. The SC-specimen shows a quasi-cleavage feature with

river markings (Figure 2i), as being cleavage-like but not along a single sharp plane, suggesting that SC-

HEAs fracture in a mode between brittle fracture and ductile fracture. During the bending test, the secondary

fracture planes {110} might be also activated due to a few degrees of the misalignment between the notch

and the fracture plane. The BC-specimen exhibits an extremely flat and smooth surface along the grain

boundary (Figure 2j), showing a typical feature of brittle intergranular fracture.

Figure 2 suggests that both the SC- and BC-specimens show a limited amount of crack tip plasticity before fracture because a linear elastic behavior is present and no force-displacement plateau has been observed.

To calculate the plane strain fracture toughness, KIc, the following equation, according to linear elastic

fracture mechanics (LEFM), can be applied [23]:

= (1) 𝐹𝐹max𝐿𝐿f 𝑎𝑎 3 Ic 2 𝑊𝑊 𝐾𝐾 𝐵𝐵𝑊𝑊 𝑓𝑓 � � where Fmax is the maximum load before fracture (i.e., fracture force) and f (a/W) is a geometry factor, which can be calculated using FEM simulations. In this study, two-dimensional extended FEM modeling was used to calculate the values of KIc for both SC- and BC-cantilevers using the J-integral method. The details of the method are explained in the supplementary data and ref. [24]. Additionally, because the geometrical dimensions of the cantilevers in this study are comparable to the cantilevers in literature, the values of f(a/W) can also be estimated using the formula given by Iqbal et al. [19], as:

f(a/W)=77.608(a/W)3-48.422(a/W)2+24.184(a/W)+1.52 (2)

The KIc values of SC- and BC-cantilevers obtained using the two means are illustrated in Figure 3a. The KIc

of SC-cantilevers is in the range of 1.3-2.1 MPa·m1/2 with an average value of 1.6 MPa·m1/2, which is nearly

one order of magnitude higher than that of the BC-cantilevers (~0.2 MPa·m1/2). It should be noted that for

the LEFM method the size of the plastic zone, ω, has to be smaller than the specimen dimensions. It requires

the specimen dimension to be above a critical thickness, ωth, of 2.5 / , as elucidated in ASTM E399 2 2 Ic y [23], where σy is the yield strength (obtained from the micro-compression𝐾𝐾 𝜎𝜎tests [6]) and KIc can be estimated

using the values of single-crystalline tungsten [25]. Here, the obtained ωth is approximately 1 μm, which is still smaller than the specimen dimensions in this study. The plasticity of the crack tip might slightly affect

KIc, and the SC-specimens tested here may fracture in a mixed-mode condition, but mainly in a brittle

cleavage mode. Nevertheless, the KIc shown in Figure 3a at least gives a lower limit for critical fracture

toughness values. Furthermore, the fracture strength, σF, which is applied to a uniformly stressed tension

plate with a center-crack of 2a, can also be estimated as [26]:

1/2 σF = KIc/(πa) (3)

Figure 3b shows that σF of the SC-specimens is 950-1750 MPa, while the BC ones exhibit a much lower σF

of ~100 MPa. The BC-cantilevers are less fracture-resistant than the SC-cantilevers, suggesting that the

GBs in this HEA are weak and the brittle intergranular fracture is its major failure mode rather than a

transgranular fracture in the polycrystalline HEAs. Figure 3c illustrates the fracture toughness as a function

of yield strength for various materials tested using the micro-cantilever method. SC-Nb25Mo25Ta25W25

HEAs exhibit higher fracture toughness than do the ceramics, but slightly lower than that of pure SC tungsten and close to that of intermetallics. Compared to the SC-HEAs, the BC-HEAs show both lower yield strength and fracture toughness. The sharp fracture surfaces (Figure 2j and supplementary information

Fig. S1) also confirm the low fracture toughness of the BC- and bulk HEAs.

Early studies indicate that the segregation of impurities to grain boundaries of alloys results in significant reductions of fracture toughness in metals, by even one order of magnitude [27]. The decrease of fracture toughness is attributed to intergranular brittle fracture along grain boundaries, although the relative decrease

5 may also depend on grain boundary misorientations [20]. To characterize the GB chemical compositions of the HEAs, we applied the atom probe tomography (APT) technique to polycrystalline HEAs. Because of the considerable challenges of including a grain boundary in an ATP tip prepared from a bulk HEA specimen (such as including a GB in a ~ 10-nm tip, difficult alignment of the GB in the tip, and high field strength in the ATP measurement breaking the ATP samples), we used annealed nanocrystalline HEA samples, which are comparable to their bulk forms [10]. The ATP tip was measured using a LEAP 4000X

HR (Cameca) in a laser mode with a wavelength of 355 nm, the specimen temperature of 40 K, and pulse frequency of 200 kHz.

We reconstructed an atom map from the HEA tip containing a GB. The concentrations of Nb, Mo, Ta, and

W are 22.3 at.-%, 22.5 at.-%, 26.3 at.-% and 25.4 at.-%, respectively. The four elements are homogeneously

distributed within the HEA tip without clustering (Figure 4a). Moreover, a one-dimensional (1D) concentration profile perpendicular to the GB and across the whole dataset indicates that there is no obvious

segregation at the GB for Nb, Mo, Ta, and W, as shown in Figure 4b. Moreover, we can clearly identify a

band region with enriched N, C, and O from the top to the bottom of the tip (Figure 4c), which can be

correlated to the GB in the HEA tip as observed in a SEM. The corresponding concentration profile

indicates that N, C, and O are segregated at the GB, with the highest concentrations of approximately 0.5

at.-%, 0.2 at.-%, and 0.05 at.-%, respectively. N and O might be introduced during sample preparation and

annealing processes; the existence of C could be due to the impurity of raw materials, although they are

99.98%-99.99% pure. Due to FIB milling and sputtering, a small amount of Ga and Ar are also detected

but exhibit no segregation at GBs, suggesting that they do not result in the difference of the fracture behavior

in the SC- and BC-HEA samples. Oxides and nitrides of the refractory metals are segregated at the GB

region with concentrations of between 1.6 at.-% and 0.05 at.-% (Figure 4e and 4f). Interestingly, in the mass-to-charge-state ratios, TaN, TaO, NbN, NbO, and WN are identified, but other oxides and nitrides from the refractory metals are not observed. Although, to our knowledge, there is no direct comparison of the reaction enthalpies of the oxides and nitrides of these refractory metals in literature, the oxide and nitride

6 compounds of Mo and W show relatively high formation enthalpies [28]. Moreover, it is well known that

Ta and Nb exhibit considerably higher oxygen/nitrogen solubility than Mo and W [29, 30], and TaN and

TaO are slightly more stable than NbN and NbO [31]. This may explain why TaN and TaO exhibit higher concentrations than the other nitrides and oxides at the GB region. At GBs, the materials that exhibit significant oxygen/nitrogen solubility (e.g., Ta) can harden and embrittle samples even at low oxygen pressures. Oxygen dissolves into the tantalum at 350 °C rapidly enough to cause embrittlement after a few hours [32]. Thus, the results in Figures 2 and 4 suggest that the intergranular fracturing of the refractory

HEAs is attributed to the GB segregation of foreign elements (i.e., N, O, and C) and the formation of brittle intermetallic phases (e.g., TaO and TaN). Consequently, avoiding the formation of brittle oxides and nitrides (especially TaO and TaN) and segregation (e.g., N and C) at GBs during sample fabrication and post-processing processes is critical to improving the fracture resistance of refractory HEAs.

In summary, we have studied the fracture toughness and fracture strength of Nb25Mo25Ta25W25 HEAs using micro-cantilever fracture tests and FEM simulations. The single-crystal cantilevers fail by quasi-cleavage

1/2 fracture with KIc of ~1.3-2.1 MPa·m , while the bi-crystal cantilevers exhibit brittle intergranular fracture

1/2 with much lower KIc of ~0.2 MPa·m . The poor fracture resistance of the polycrystalline refractory HEAs

is attributed to the GB segregation and formation of brittle oxides and nitrides at GBs. For future studies, it would be interesting to study the oxidation and nitridation behavior of refractory HEAs in different temperature and oxygen conditions, and minimize the segregation and formation of brittle phases at GB regions.

Acknowledgements

The authors thank H. Ma (ETH Zurich), E. Kawai and S. Ashida (Kyoto University) for their experimental help and Dr. S. Gerstl (ScopeM ETH Zurich) for his help in Atom Probe analysis. Y.Z and S.M. acknowledge financial support through SNF Grants (200021_143633, P2EZP2_165278 and

200020_144430); Y.Z and H.Y. also acknowledge financial support through the JSPS program (GR14103 and P13055).

References

[1] O.N. Senkov, G.B. Wilks, D.B. Miracle, C.P. Chuang, P.K. Liaw, Intermetallics, 18 (2010) 1758-1765. [2] O.N. Senkov, G.B. Wilks, J.M. Scott, D.B. Miracle, Intermetallics, 19 (2011) 698-706. [3] M.F. del Grosso, G. Bozzolo, H.O. Mosca, J Alloy Compd, 534 (2012) 25-31. [4] L. Lilensten, J.P. Couzinié, L. Perrière, J. Bourgon, N. Emery, I. Guillot, Mater Lett, 132 (2014) 123-125. [5] M. Widom, W.P. Huhn, S. Maiti, W. Steurer, Metallurgical and Materials Transactions A, 45 (2014) 196-200. [6] Y. Zou, S. Maiti, W. Steurer, R. Spolenak, Acta Mater, 65 (2014) 85-97. [7] B. Gorr, M. Azim, H.J. Christ, T. Mueller, D. Schliephake, M. Heilmaier, J Alloy Compd, 624 (2015) 270-278. [8] O.N. Senkov, J.M. Scott, S.V. Senkova, F. Meisenkothen, D.B. Miracle, C.F. Woodward, J Mater Sci, 47 (2012) 4062-4074. [9] Y.D. Wu, Y.H. Cai, T. Wang, J.J. Si, J. Zhu, Y.D. Wang, X.D. Hui, Mater Lett, 130 (2014) 277-280. [10] Y. Zou, H. Ma, R. Spolenak, Nat Commun, 6 (2015). [11] C.M. Liu, H.M. Wang, S.Q. Zhang, H.B. Tang, A.L. Zhang, J Alloy Compd, 583 (2014) 162-169. [12] O.N. Senkov, C.F. Woodward, Materials Science and Engineering: A, 529 (2011) 311-320. [13] O.N. Senkov, S.V. Senkova, C. Woodward, Acta Mater, 68 (2014) 214-228. [14] B. Gludovatz, A. Hohenwarter, D. Catoor, E.H. Chang, E.P. George, R.O. Ritchie, Science, 345 (2014) 1153- 1158. [15] T.P. Halford, K. Takashima, Y. Higo, P. Bowen, Fatigue Fract Eng M, 28 (2005) 695-701. [16] K. Matoy, H. Schönherr, T. Detzel, T. Schöberl, R. Pippan, C. Motz, G. Dehm, Thin Solid Films, 518 (2009) 247-256. [17] W. Cao, A. Kundu, Z. Yu, M.P. Harmer, R.P. Vinci, Scripta Materialia, 69 (2013) 81-84. [18] S. Massl, W. Thomma, J. Keckes, R. Pippan, Acta Mater, 57 (2009) 1768-1776. [19] F. Iqbal, J. Ast, M. Göken, K. Durst, Acta Mater, 60 (2012) 1193-1200. [20] D.E.J. Armstrong, A.J. Wilkinson, S.G. Roberts, Philosophical Magazine Letters, 91 (2011) 394-400. [21] S. Wurster, C. Motz, R. Pippan, Philos Mag, 92 (2012) 1803-1825. [22] T. Sumigawa, T. Shishido, T. Murakami, T. Kitamura, Materials Science and Engineering: A, 527 (2010) 4796- 4803. [23] A. International: ASTM Standard E399-09, West Conshohocken, PA, 2003. [24] H. Yu, L. Wu, L. Guo, S. Du, Q. He, International Journal of Solids and Structures, 46 (2009) 3710-3724. [25] A. Giannattasio, S.G. Roberts, Philos Mag, 87 (2007) 2589-2598. [26] T.H. Courtney, Mechanical behavior of materials, Waveland Press, 2005. [27] C.J. McMahon, V. Vitek, Acta Metallurgica, 27 (1979) 507-513. [28] B. Landolt, R. Börnstein, Scientific Group Thermodata Europe (SGTE), Springer-Verlag, Berlin Heidelberg, (1999). [29] J.R. DiStefano, B.A. Pint, J.H. DeVan, International Journal of Refractory Metals and Hard Materials, 18 (2000) 237-243. [30] C.A. Steidel, D. Gerstenberg, J Appl Phys, 40 (1969) 3828-3835. [31] A.D. Mah, N.L. Gellert, Journal of the American Chemical Society, 78 (1956) 3261-3263. [32] D.A. Vermilyea, Acta Metallurgica, 6 (1958) 166-171. [33] F. Ostlund, K. Rzepiejewska-Malyska, K. Leifer, L.M. Hale, Y.Y. Tang, R. Ballarini, W.W. Gerberich, J. Michler, Adv Funct Mater, 19 (2009) 2439-2444.

Figure 1. a. An EBSD inverse pole figure map of the cross-section of the HEA bulk specimen. Two adjacent <110>-oriented grains were selected to fabricate micro-cantilevers, as indicated by boxes. b. and c. Typical single crystalline (SC) and bi-crystal (BC) cantilevers fabricated by FIB, respectively. The notch,

crystal orientation, and grain orientation are indicated in each figure. d. A schematic of the shape and

dimension of an FIB-notched cantilever with a beam length, L0, width, W, thickness, B, loading length, Lf,

and notch depth, a.

Figure 2. Representative in situ TEM images of deflected SC- (a, b, and c) and BC- (e, f, and g) cantilevers:

a and e. initial contacts; b. and f. crack tip opening at the maximum load; c. and g. fracture and load drops.

d and h. the corresponding indenter load-displacement curves for the SC- and BC-cantilevers, respectively.

The indenter displacement was evaluated using an image correction software. Typical post-mortem SEM

images of the fracture surfaces: i. the SC-cantilever specimen shows a quasi-cleavage feature with river

markings, suggesting SC-HEAs are not intrinsically brittle; j. the BC-cantilever specimen exhibits a typical feature of brittle intergranular fracture. More SEM images of fracture surfaces are shown in supplementary information Fig. S2.

Figure 3. a. Comparison of the fracture toughness between SC- and BC-cantilevers, using the extended

FEM modeling and the formula obtained in the literature [19]. b. Fracture stress of SC-and BC-HEAs

determined by the fracture toughness values applied to a uniformly stressed tension plate with a center-

crack of 2a. c. Ashby map showing fracture toughness as a function of yield strength for the micro-

cantilever HEAs and other materials [16, 18, 19, 21, 33] that were also tested using the micro-cantilever method. The yield strengths are obtained from micro-compression tests or estimated by nanoindentation hardness, and KIc is calculated using the LEFM method and the values for Si are obtained from micro-pillar compression.

Figure. 4. Reconstruction of the APT tip showing the HEA elemental distribution of (a) Nb, Mo, T and

W, (b) N, O and C, and (c) detected oxides and nitrides (i.e., TaN, TaO, NbN, NbO, and WN) of the refractory metals; d, e, and f are corresponding one-dimensional concentration profiles of the GB by measuring the concentration along the perpendicular axis of a box wrapped around the whole GB with a fixed bin width of 0.6 nm. The APT results indicate the segregation of N, C, O, and the refractory metal

oxides and nitrides at GBs.