Anisotropic strengthening of nanotwinned austenitic grains in a nanotwinned stainless steel

Anisotropic strengthening of nanotwinned austenitic grains in a nanotwinned stainless steel

Scripta Materialia 142 (2018) 15–19 Contents lists available at ScienceDirect Scripta Materialia journal homepage: www.elsevier.com/locate/scriptama...

3MB Sizes 1 Downloads 14 Views

Scripta Materialia 142 (2018) 15–19

Contents lists available at ScienceDirect

Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat

Regular article

Anisotropic strengthening of nanotwinned austenitic grains in a nanotwinned stainless steel F.K. Yan ⁎, Q. Li, N.R. Tao Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese Academy of Sciences, Wenhua Road 72, Shenyang 110016, China

a r t i c l e

i n f o

Article history: Received 14 February 2017 Received in revised form 2 August 2017 Accepted 9 August 2017 Available online xxxx Keywords: Nanotwins Orientation Strengthening Austenitic stainless steel

a b s t r a c t Nanotwin lamellae orientation effect on tensile properties of nanotwinned grains strengthening stainless steels was investigated. It is found that the nanotwinned grains with twin lamellae oriented roughly parallel to loading direction exhibit the much higher strengthening effect (at least a GPa higher) than the nanotwinned grains containing twin lamellae inclined to loading direction, which is related to the anisotropic strengthening mechanisms of nanotwins. © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Nanotwinned austenitic steels have attracted great interests in the past few years as a novel type of advanced high performance steels [1–8]. They are commonly based on a single-phase duplex-microstructure consisting of nanotwinned grains embedded in the statically recrystallized (SRX) grained matrix. These nanotwinned austenitic (referred to as nt-γ) grains act as “hard inclusions” to strengthen austenitic matrix due to containing a high density of nanoscale deformation twins [1–4]. Attractively, nt-γ grains not only possess the same elastic modulus as the matrix, but also co-deform homogeneously together with the surrounding soft matrix without generating notable strain localizations around the interfaces at low strains [9,10]. Hence, in comparison with conventional dual-phase (DP) steels, the nanotwinned austenitic steels exhibit an enhanced combination of strength and ductility [1,2,4]. Apparently, the mechanical properties of nanotwinned austenitic steels can be optimized by tailoring the structural parameters of nt-γ grains through various thermomechanical treatments. It has been reported that the higher volume fraction of nt-γ grains and the thinner of twin/matrix (T/M) lamellae in the samples are introduced, the higher strength can be achieved [2,3], like that in traditional DP steels [11,12]. But recent investigations [2,3,10] indicated that such relationship is not monotonously changing. For example, the strength did not rise but slightly reduced when the volume fraction of nt-γ grains increased from ~ 25 vol% to ~ 50 vol% in nanotwinned 316L samples with the fixed twin thickness [2,10]. So the strengthening of nt-γ grains is also influenced by other factors.

⁎ Corresponding author. E-mail address: [email protected] (F.K. Yan).

http://dx.doi.org/10.1016/j.scriptamat.2017.08.015 1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Investigations of electro-deposited nanotwinned Cu [13] indicated that the twin boundary (TB) orientation with respect to loading direction leads to an anisotropic nanotwin strengthening response. It implies twin lamellae orientation-dependent strengthening of nt-γ grains. Accordingly, in the present work, two types of nanotwinned 316L stainless steels were produced by means of dynamic plastic deformation (DPD) followed thermal annealing. One is strengthened by large nt-γ grains containing T/M lamellae which are roughly inclined to the tensile axis (TA) direction (referred to as I-nt 316L) and the other is strengthened by fine nt-γ grains containing T/M lamellae which are parallel to TA (referred to as P-nt 316L). The objective of this work is to investigate the relationship between the T/M lamellae orientation and the nt-γ grain strengthening effect. The material studied in this work is a commercial AISI 316L stainless steel with a composition of Fe-16.42Cr-11.24Ni-2.12Mo-0.02C-0.37Si1.42Mn-0.011S-0.040P (wt%). The original samples were solutionheat-treated at 1200 °C for 1 h and followed by water quenching to obtain a uniform austenitic structure with an average grain size of ~ 100 μm. The cylindrical samples were processed on a dynamic plastic deformation (DPD) facility with a strain rate of ~102–103 s−1 at room temperature to an accumulative strain of ε = 0.8. The DPD set up and processing parameters were described elsewhere in details [2]. Microstructures were characterized by using field emission gun scanning electron microscope (SEM) FEI Nova NanoSEM 430 with electron channeling contrast (ECC) imaging and a transmission electron microscope (TEM) JEOL 2010 operated at 200 kV. The I-nt 316L stainless steels were produced by means of DPD to the coarse grained samples with random grain orientation followed by annealing at 770 °C for 90 min. As shown in Fig. 1a, the microstructure is mainly composed of nt-γ grains (outlined) embedded in SRX grains.

16

F.K. Yan et al. / Scripta Materialia 142 (2018) 15–19

Fig. 1. (a) A typical SEM-ECC image showing the microstructure of the I-nt 316L samples mainly consisting of large nt-γ grains (outlined) embedded in the recrystallized (SRX) grain matrix. Statistical distribution of nt-γ grain size (b) and the twin/matrix (T/M) lamellar thickness (c) as well as angle distribution of twin boundary (TB) traces with respect to the tensile direction (d).

These nt-γ grains exhibit an average size of ~23 μm, much larger than SRX grains (averagely 2.7 μm) (Fig. 1b). They constitute ~20% in volume and contain a high density of T/M lamellae with the average thickness of ~31 nm by TEM statistics (Fig. 1c). The nt-γ grains distribute randomly in the matrix, but the orientation distribution of T/M lamellae inside ntγ grains is not homogeneous. Most TBs are inclined to the tensile axis (TA) direction, as shown in Fig. 1d, ranging from 0° to ~45°, averagely ~19°. Rough estimation shows that the ~75% traces of TBs are inclined (N 10°) to the TA direction while ~ 25% parallel to the TA direction (b 10°). It is noted that the statistical measurement of the angles between the TA and TB traces is more operable and efficient than that between the TB and the DPD direction (i.e. Compression axis, CA) by EBSD. In fact, the two measurements are self-consistent since the normal directions of most twin planes distribute randomly in the 3D-space [2,3, 14]. That is to say, most twin planes in I-nt 316L are also inclined to the CA direction. The P-nt 316L samples were produced by DPD to the fully recrystallized samples (with average size of 4.6 μm and random orientation distribution) and then subsequently annealing at 730 °C for 30 min. As shown in Fig. 2a-b, the typical microstructure is also mainly consisted of the SRX grains matrix “dispersed” with nt-γ grains. Attractively, the nt-γ grain size is very small, ranging from several hundred nanometers to several micrometers (usually b 4.5 μm) with the average transverse size of ~ 1.7 μm (Fig. 2c), which is nearly identical to the size of SRX grains (averagely 1.6 μm). Statistics show that the volume fraction of nt-γ grains and the T/M lamellar thickness is ~ 16% and ~ 29 nm (Fig.2d), respectively. These nt-γ grains “disperse” very homogeneously in the samples. But the angle distribution of TB traces with respect to TA is narrow, as shown in Fig. 2e, roughly ~70% angles below 10°, averagely

9°. Clearly, most T/M lamellae are roughly parallel to the TA direction, which means that they are almost perpendicular to the CA direction, analogous to the previous results [3,14]. Table 1 summarizes the microstructure characteristics of the two types of samples. Clearly, there are no significant differences in constitution and characteristic size except the size of nt-γ grains and T/M lamellar orientation distribution. The average size is ~ 23 μm for large nt-γ grains and ~1.7 μm for small ones. Meanwhile, most T/M lamellae are inclined to TA direction in large nt-γ grains of the I-nt samples but parallel to TA direction for small nt-γ grains of the P-nt samples. Such different twin lamellae orientation distribution may be due to the different degree of grain rotation at the same DPD strain of 0.8 [14,15]. The large grains containing multiple nanotwins may rotate slowly while the small nanotwinned grains rotate easily with increasing the strains. This microstructural distinction significantly influences the mechanical properties of nanotwinned 316L samples. Uniaxial tensile tests with an initial strain rate of 5 × 10−3 s−1 at room temperature were performed in an Instron 5848 Micro-Tester system equipped with a contactless MTS LX300 laser extensometer to measure the tensile strain upon loading. As shown in Fig. 3, the P-nt 316L samples exhibit a high yield strength of ~760 MPa, which is 220 MPa higher than that of I-nt ones. Interestingly, the uniform elongation of the P-nt 316L is still as high as ~20%, ~6% lower than of the I-nt counterparts. It means that the yield strength is elevated by ~40% at the expense of a 20% loss in uniform elongation for the P-nt 316L samples, exhibiting an enhanced combination of strength and ductility. To investigate and compare the strengthening mechanisms between the two samples, tensile deformation behaviors of SRX grains surrounding nt-γ grains in both samples are characterized at uniform strains of ε

F.K. Yan et al. / Scripta Materialia 142 (2018) 15–19

17

Fig. 2. The typical SEM-ECC (a) and TEM (b) images showing the microstructure of P-nt 316L samples mainly composing of small nt-γ grains (outlined) “dispersed” in the SRX grain matrix. Statistical distribution of nt-γ grain size (c) and the T/M lamellar thickness (d) as well as angle distribution between the TB traces and the tensile direction (e).

= 2% and ε = 15%, respectively. The analysis methods and systematical study were described elsewhere [9]. As shown in Fig. 4a–b, in the initial stage of tensile deformation (ε = 2%), the deformation characteristic between the two samples shows no notable differences. Dislocations distribute randomly in the SRX grains adjacent to nt-γ grains in both samples (Fig. 4a–b) and their density is roughly comparable. Specifically, there are no high density dislocations pile-uping at the nt-γ/SRX interfaces. Such phenomenon is consistent with our previous EBSD results

that the strain gradient in SRX grains surrounding nt-γ grains is not pronounced [9]. With straining to 15% (Fig. 4c-d), dislocation density increases in SRX grains, forming some dislocation substructures, such as dislocation walls, cells, and tangles, like those observed in conventional deformed metals [16,17]. But the characteristic sizes of dislocation substructures surrounding the P-nt-γ grains (Fig. 4c) are slightly smaller than those close to I-nt counterparts (Fig. 4d), roughly 200–500 nm for the former and 300–900 nm for the latter.

Table 1 Microstructure characteristics of the I-nt and P-nt 316L samples. T/M: twin/matrix; TA: tensile axis; DS: dislocation substructures; SRX: recrystallized grains. Samples

Microstructures nt-γ Grains

P-nt 316L I-nt 316L

SRX grains

DS

Volume (%)

T/M lamellae thickness (nm)

Grain size (μm)

Angle of T/M lamellae with TA

Volume (%)

Grain size (μm)

Volume (%)

16 ± 2 20 ± 4

29 ± 4 31 ± 4

1.7 ± 0.1 23 ± 2

~9° ~19°

74 ± 2 68 ± 2

1.6 ± 0.1 2.7 ± 0.1

~10 ~12

18

F.K. Yan et al. / Scripta Materialia 142 (2018) 15–19

It is well accepted that the total yield strength (σy) of nanotwinned samples can be expressed roughly according to rule-of-mixture [2,3]: σy ¼ f

nt−γ

∙σ nt−γ þf y

SRX

∙σ SRX þf y

DS

∙σ DS y

ð1Þ

where fnt−γ, fSRXand fDSis the volume fraction of nt-γ grains, SRX grains γ , σSRX and σDS and dislocation substructures, respectively. And σnt− y y y is their corresponding yield strength. Clearly, the contribution of SRX grains and dislocation substructures to the total strength can be described in terms of the “Hall-Petch relationship” [20] and “similitude principle” [5,21,22], respectively: −1=2

¼ σ 0 þ K P−H ∙d σ SRX y

ð2Þ

−1 σ DS y ¼ σ 0 þ KμbM∙D

ð3Þ

where in Eq. (2): σ0 (=181 MPa [16]) is the friction stress, KP−H (¼ 26 Fig. 3. Tensile true stress-strain curves for I-nt and P-nt 316L samples, respectively.

The above results indicated that both P-nt-γ and I-nt-γ grains can sustain a certain of tensile uniform strains and co-deform together with their neighboring SRX grains without generate notable strain localization (i.e. high density dislocations) around the nt-γ/SRX interfaces at small strains. But the deformation capability between the two nt-γ grains is different. I-nt-γ grains seem to possess a higher ductility and undertake a slightly higher tensile strain than that for P-nt-γ ones, which results in the less deformation of SRX grains adjacent to I-nt-γ grains in comparison with that close to P-nt-γ grains. It implies that the deformation mechanism between the two nt-γ grains is different. Notably, no extra local strengthening resulted from high density of dislocations around P-nt-γ and I-nt-γ grains contributes to the total strength of 316L samples at the beginning of yielding. Such phenomenon is distinct from that in conventional second-phase strengthening composites in which local strengthening around hard particles is pronounced [18,19]. Hence, the large increment of yield strength of P-nt samples is mainly due to the higher strengthening effect of P-nt-γ grains themselves.

1

7 MPa∙μm2 [16]) is the Hall-Petch constant and d is the average SRX grain size. And in Eq. (3): K (= 2.8 [23]) is the similitude constant, μ (=78 GPa [23]) is the shear modulus, b (=0.258 nm [24]) is the Burgers vector, M (=3.06 [23]) is the Taylor factor and D (=~200 nm in both samples) is the average dislocation substructures size. Other parameters, such as the volume fraction, the average SRX grain size are , σSRX and σDS shown in Table 1. So the σnt−γ y y y can be roughly calculated by Eq. (1–3). Surprisingly, the effective yield strength of P-nt-γ grains is 2.29 GPa, which roughly conforms to the Hall-Petch relationship. But the I-nt-γ grains exhibit an effective yield strength of ~0.91 GPa, deviating largely from the Hall-Petch value. Apparently, the P-nt-γ grains are much stronger than the I-nt-γ ones. The nt-γ grain size effect on its yield strength is slight since there is only ~150 MPa increment in the yield strength for fine nt-γ grains by calculations. Thus, such significant distinction is primarily determined by the different orientation of T/M lamellae inside nt-γ grains. Previous investigations [13,25,26] indicated that the effect of nanotwin strengthening exhibited a strong dependence on loading direction with respect to TBs. When the loading direction is perpendicular or parallel to twin planes, TB strengthening dominates. While TB softening prevails when twin

Fig. 4. Typical TEM images of dislocation distribution and substructure characteristic in recrystallized grains surrounding the P-nt and I-nt austenitic grains at the various tensile strains of 2% (a, b) and 15% (c, d), respectively.

F.K. Yan et al. / Scripta Materialia 142 (2018) 15–19

planes are inclined to loading direction. For example, when the electrodeposited nanotwinned Cu samples are tested with the compression axis oriented at 0° with respect to TBs, the measured yield strength was 463 MPa. But a much lower yield stress of ~230 MPa was achieved in the case of CA oriented at 45° to TBs. Such strengthening anisotropy of nanotwins also significantly changes the strengthening mechanism of nt-γ grains with different twin lamellae orientation in nanotwinnd 316L samples. For P-nt-γ grains, most T/M lamellae traces are parallel to TA direction (Fig. 2e). At the beginning of yielding, the T/M lamellae twisted and deformed severely by many TEM observations (Fig. 4a). It mainly originates from a large amount of TB strengthening activation in terms of the threading dislocation motion in the T/M lamellae channels [25]. The neighboring nanoscale TBs act as strong barriers to block these dislocations trailing motion. Specifically, prior dislocations of high density existing in T/M lamellae further restrain these dislocations motion. Such superposed effects lead to more pronounced TBs strengthening to P-nt-γ grains, nearly identical to Hall-Petch strengthening of the TBs. So they exhibit an effective yield strength as high as 2.29 GPa and contribute largely to the strength of the 316L samples. In contrast, most T/M lamellae are inclined to TA direction in the I-nt-γ grains, (Fig. 1d). So TB softening activated and dominated at the beginning of tensile tests [25]. Multiple partial dislocations slip along the TBs prevails with little TB blocking, leading to a very limited strengthening effect of Int-γ grains to the SRX matrix. In summary, we produced two types of, i.e. I-nt and P-nt 316L stainless steels by means of DPD with subsequent annealing. It is found that the nanotwinned grain strengthening is influenced slightly by their size but significantly by twin lamellae orientation. The nanotwinned grains containing twin lamellae oriented roughly parallel to TA can enhance the strength of nanotwinned 316L samples more significantly than nanotwinned grains containing twin lamellae inclined to TA. It is mainly ascribed to the unique anisotropic strengthening mechanisms of nanotwins. The results provide important implications by tuning the twin lamellae orientation within nt-γ grains to optimize the mechanical properties of nanotwinned austenitic steels.

19

Acknowledgements Financial support from the National Natural Science Foundation (Grant No. 51371172 and 51501191), the Doctoral Initiation Foundation of Liaoning Province (No. 201501042) and Shenyang National Laboratory for Materials Science, IMR, CAS (Grant No. Y4N16F3161) is acknowledged. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

K. Lu, F.K. Yan, H.T. Wang, N.R. Tao, Scr. Mater. 66 (2012) 878. F.K. Yan, G.Z. Liu, N.R. Tao, K. Lu, Acta Mater. 60 (2012) 1059. H.T. Wang, N.R. Tao, K. Lu, Acta Mater. 60 (2012) 4027. O. Bouaziz, S. Allain, C.P. Scott, P. Cugy, D. Barbier, Curr. Opin. Solid State Mater. Sci. 15 (2011) 141. I. Gutierrez-Urrutia, D. Raabe, Acta Mater. 59 (2011) 6449. H. Ueno, K. Kakihata, Y. Kaneko, S. Hashimoto, A. Vinogradov, Acta Mater. 59 (2011) 7060. P. Zhou, Z.Y. Liang, R.D. Liu, M.X. Huang, Acta Mater. 111 (2016) 96. A.Y. Chen, H.H. Ruan, J. Wang, H.L. Chan, Q. Wang, Q. Li, J. Lu, Acta Mater. 59 (2011) 3697. F.K. Yan, N.R. Tao, F. Archie, I. Gutierrez-Urrutia, D. Raabe, K. Lu, Acta Mater. 81 (2014) 487. F.K. Yan, N.R. Tao, K. Lu, Scr. Mater. 84-85 (2014) 31. N.K. Balliger, T. Gladman, Metal Science. 15 (1981) 95. A. Bag, K.K. Ray, E.S. Dwarakadasa, Metall. Mater. Trans. A 30A (1999) 1193. Z.S. You, X.Y. Li, L.J. Gui, Q.H. Lu, T. Zhu, H.J. Gao, L. Lu, Acta Mater. 61 (2013) 217. F.X. Lin, Y.B. Zhang, N.R. Tao, W. Pantleon, D.J. Jensen, Acta Mater. 72 (2014) 252. L. Meng, P. Yang, Q. Xie, H. Ding, Z. Tang, Scr. Mater. 56 (2007) 931. B.P. Kashyap, K. Tangri, Acta Metall. Mater. 43 (1995) 3971. N.R. Tao, Z.B. Wang, W.P. Tong, M.L. Sui, J. Lu, K. Lu, Acta Mater. 50 (2002) 4603. M.F. Ashby, Philos. Mag. 21 (1970) 399. D.A. Korzekwa, D.K. Matlock, G. Krauss, Metall. Trans. A. 15 (1984) 1221. J. Gil Sevillano, P. van Houtte, E. Aernoudt, Prog. Mater. Sci. 25 (1981) 69. I. Gutierrez-Urrutia, D. Raabe, Acta Mater. 60 (2012) 5791. M. Sauzay, L.P. Kubin, Prog. Mater. Sci. 56 (2011) 725. X. Feaugas, Acta Mater. 47 (1999) 3617. G. Angella, B.P. Wynne, W.M. Rainforth, J.H. Beynon, Acta Mater. 53 (2005) 1263. T. Zhu, H.J. Gao, Scr. Mater. 66 (2012) 843. Z.S. You, L. Lu, K. Lu, Acta Mater. 59 (2011) 6927.