- Email: [email protected]
On the effect of gamma phase formation on the pseudoelastic performance of polycrystalline Fe–Mn–Al–Ni shape memory alloys
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 Scripta Materialia xxx (2015) xxx–xxx 1
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat 5 6 3 4 7 8 9 10 11 12 1 2 4 7 15 16 17 18 19 20 21 22 23 24 25 26
On the effect of gamma phase formation on the pseudoelastic performance in polycrystalline Fe–Mn–Al–Ni shape memory alloys M. Vollmer a,⇑, C. Segel a, P. Krooß a, J. Günther a, L.W. Tseng b, I. Karaman b, A. Weidner a, H. Biermann a, T. Niendorf a a b
Technische Universität Bergakademie Freiberg, Institut für Werkstofftechnik (Materials Engineering), 09599 Freiberg, Germany Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA
a r t i c l e
i n f o
Article history: Received 8 May 2015 Revised 4 June 2015 Accepted 9 June 2015 Available online xxxx Keywords: Shape memory alloys (SMAs) Microstructure Quenching Precipitation Mechanical properties
a b s t r a c t In recent years, iron-based shape-memory-alloys such as Fe–Mn–Al–Ni came into focus, due to superior pseudoelastic performance and concomitantly low costs for alloying elements. The formation of the ductile c-phase in polycrystalline Fe–Mn–Al–Ni and its impact on the pseudoelastic performance have not been addressed, yet. Based on modification of quenching conditions this study shows that controlled c-phase precipitation at grain boundaries can suppress intergranular cracking without affecting pseudoelasticity. In situ incremental strain tests revealed fairly good recoverability up to about 8% strain. Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.
28 29 30 31
32 33 34 35
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
58 59 60 61
Iron based shape memory alloys (SMAs) such as Fe–Ni–Cobased and Fe–Mn-based alloys have attracted considerable attention in recent years [1–12]. Following the development of Fe–Ni–Co–Ti SMAs, which are characterized by relatively low reversible transformation strains [7–10], Fe–Ni–Co–Al-based alloys were introduced showing almost perfect pseudoelastic strains up to 13% in the polycrystalline state, as shown for Fe–Ni–Co–Al–Ta by Tanaka, Kainuma and co-workers [11]. More recently, Omori et al. [13] introduced a new Fe–Mn-based SMA, namely Fe–Mn–A l–Ni, showing about 5% pseudoelastic strain in the polycrystalline state in single cycle tests and an extremely low slope for the Clausius–Clapeyron relationship (0.53 MPa °C1) over a wide temperature range (196 °C to 240 °C). In single crystalline samples with [0 0 1] orientation even more than 20% transformation strain was demonstrated by Tseng et al. [14] under tensile loading. Due to these excellent properties and the cheap alloying elements Fe–Mn–Al–Ni SMA became a candidate material for use in aerospace, automobile and seismic applications. The high amount of Mn in Fe–Mn–Al alloys introduces a change in a/c equilibria and, thus, an unusual transformation between an a (A2 – bcc) austenitic high-temperature parent phase and a c0 (A1/2M – fcc) martensitic product phase [13,15]. In order to promote a thermoelastic transformation, Omori et al. added Ni to Fe–Mn–Al resulting in the formation of nano-sized, coherent b ⇑ Corresponding author.
(B2 – bcc/NiAl) precipitates in the disordered a matrix induced by a low temperature heat treatment [13,16]. These precipitates strengthen the matrix and suppress plastic deformation. Besides the orientation dependency of martensitic transformation in Fe–Mn–Al–Ni [13,14,17], a superior pseudoelastic performance was shown for oligocrystalline, i.e. bamboo-structured samples, characterized by a mean grain diameter to wire diameter [17] or sheet width [13], respectively, ratio of larger than one. Ueland et al. [18–20] demonstrated that bamboo structures with grain sizes larger than the wire diameter and grain boundaries oriented only perpendicular to the loading direction, i.e. preventing grain boundary triple junctions and concomitant grain constraints, led to superior pseudoelastic performance in micro wires of Cubased SMAs. Similarly, damage evolution in bulk polycrystalline Co–Ni–Ga high-temperature alloys was investigated [21], revealing that bamboo-structured samples showed superior performance. It was shown that intergranular fracture is strongly affected not only by the grain boundary arrangement, but also by the formation of a ductile phase decorating the grain boundaries, i.e. the c-phase in Co–Ni–Ga alloys [21,22]. The Fe–Mn–Al–Ni SMA in the present study is prone to intergranular cracking upon rapid quenching. Omori et al. used Mo-sheets during the solution treatment [13]. However, to the best of authors’ knowledge, data reporting on crack formation with an emphasis on its underlying mechanisms are still lacking in open literature.
http://dx.doi.org/10.1016/j.scriptamat.2015.06.013 1359-6462/Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 2 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
The present study focusses on the role of segregations in Fe– Mn–Al–Ni evolving during quenching, their impact on the pseudoelastic behavior as well as the morphology and sites of the phases formed as a result of the different quenching conditions. Tension samples were first cyclically heat treated in order to obtain a coarse-grained microstructure, similar to the procedure reported for a Cu–Al–Mn SMA by Omori et al. [23]. They were able to show that by increasing the fraction of semi-coherent second phase particles during cooling into a two phase region, the development of subgrain structures is promoted. Omori et al. attributed this to the loss of coherency between the matrix and the particles and concurrent dislocation formation. Upon heating to temperatures above the solvus temperature, the subgrains remain and the associated internal stress triggers abnormal grain growth [23]. However, this kind of procedure results in intense crack formation upon final quenching to room temperature in Fe–Mn–Al–Ni. Commercially pure polycrystalline Fe–Mn–Al–Ni, produced by vacuum induction melting, with a nominal chemical composition of Fe–34% Mn–14% Al–7.5% Ni (at.%) was used in the current work. The alloy was cast by Stahlzentrum Freiberg e.V. (Freiberg, Germany) and subsequently hot-forged at 1150 °C. Dog-boneshaped tension samples were cut using electro discharge machining featuring a gage length of 12 mm and a cross section of 1.6 mm 1.5 mm. Subsequently, samples were sealed in quartz tubes under high vacuum. In order to obtain a high fraction of the second phase, samples were solution treated at 1200 °C for 12 h and then air cooled for 5 s before quenching into 15 °C cold water. The second set of samples was solution heat treated at 1200 °C for 30 min, cooled down into the a + c two phase region at 900 °C within 30 min, held for 15 min and then heated up to 1200 °C within 30 min again. This procedure was repeated for four times. Finally, the samples were homogenized at 1200 °C for 1 h to ensure that they were free of the c-phase. In order to generate two different conditions, some samples were quenched into water at 15 °C, referred to as cold water in the remainder of the text, and the others were quenched into water heated up to 80 °C, referred to as hot water. All samples were subsequently aged at 200 °C for 3 h in air, ground down to 5 lm grit size and finally vibration-polished using colloidal SiO2 suspension with 0.02 lm particle size. For optical microscopy the air cooled condition was additionally etched using a solution of 7% nitric acid and 93% ethanol. Microstructural analyses were conducted using a scanning electron microscope (SEM) operated at 20 kV equipped with an electron-backscatter diffraction (EBSD) system and an energy dispersive X-ray spectroscopy (EDS) unit. For in situ high-resolution measurements, a miniature load frame was used under a constant crosshead displacement rate of 2 lm s1 in combination with the SEM system. Incremental strain tests were conducted with a step size of 120 lm (e.g. about 1% strain) until the samples failed. Strain values for the stress–strain diagram were then calculated from displacement data. Surface images were taken using either secondary electron (SE) or backscattered electron (BSE) contrasts, with and without superimposed mechanical load. Fig. 1 shows SEM micrographs and results from the EBSD measurements for the cold water quenched state after cyclic heat treatment revealing no traces of evolution of second phases. It is obvious from Fig. 1a that this state is prone to crack formation and further investigations revealed that a high density of multi-variant martensite appeared in the vicinity of the cracks (not shown). Since temperature induced martensitic transformation can be excluded according to the magnetic measurements conducted by Omori et al. [13], high internal stresses resulting from quenching in cold water are expected to be responsible for the martensitic transformation. The EBSD scan and the associated inverse pole figure (IPF) in Fig. 1b reveal that crack formation
Fig. 1. Characteristic crack formation during cold water quenching after cyclic heat treatment: (a) SEM micrograph depicting crack formation, (b) EBSD inverse pole figure mapping of the highlighted area depicting orientations with respect to the loading direction, (c) SEM micrograph of a representative fracture surface.
occurred mainly along the grain boundaries. As shown in Fig. 1c, crack surfaces are very smooth revealing a brittle behavior in this condition. A bamboo structure as induced by the cyclic heat treatment seems to promote cracking most probably due to an increasing grain boundary area. SMAs are characterized by pronounced anisotropic material behavior, e.g. high elastic anisotropy [3]. Consequently, grain boundaries of high misorientation, separating large grains of significantly different crystallographic orientation, are prone to rapid crack formation upon deformation [21]. Similarly, the anisotropic nature seems to promote cracking upon quenching at those large grain boundaries. In samples quenched at significantly slower cooling rates, a second phase with a serrated interface formed mainly at the grain boundaries and subsequently grew into the grain interiors (Fig. 2a). The EBSD measurement in Fig. 2b clearly reveals the fcc structure of the second phase and further investigations of a representative, larger section by EBSD revealed a c-phase fraction of about 40% (not shown). Additional EDS measurements (Fig. 2c–f) show a decrease of Al content and a slight increase of Mn content, whereas changes in Fe and Ni can hardly be resolved. Al is known to be an a-phase stabilizer and Mn is a c-phase stabilizer. These local segregations form a Widmanstätten-like c-phase especially at the grain boundaries. The slightly different alloy composition with a lower content of Al and Ni in comparison with the alloy introduced by Omori et al. [13] additionally promotes c-phase formation. Fast precipitation of the c-phase along grain boundaries in Fe–Mn– Al alloys, evolving despite rapid water quenching, was also observed in other studies [15,24,25]. Hwang et al. [24] attributed the segregations to a massive transformation due to short-range diffusion of Mn and Al across the a/c boundaries. Another microstructural feature resulting from fast segregation in the Fe–Mn–Al–Ni alloy is acicular precipitates inside the grains indicated by the white arrow in Fig. 2g. These are characterized by a length of about 5–10 lm, which is three orders of magnitude larger than the size of the fine b precipitates [16]. Since their volume fraction is much lower in vicinity of c-phase covered grain boundaries, it can be assumed that these segregations are of similar composition. However, detailed analyses will be subject of future work. In contrast to the cold water quenched condition, the slow cooled condition was free of martensite as well as free of cracks along grain boundaries. However, this condition did not show satisfactory pseudoelastic behavior as the high fraction of the non-transforming c-phase is detrimental with respect to recoverability of martensitic transformation.
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188
189 190 191 192 193 194 195 196
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
Fig. 2. c-Phase formation following slow cooling: (a) optical micrograph depicting the etched surface, (b) EBSD phase mapping, (c–f) EDS element mappings, and (g) SEM micrograph of an area showing fine acicular precipitates (white arrow) in the grain interior.
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
It is well known that grain boundary decoration using suitable ductile second phases can reduce crack susceptibility in brittle SMAs without deterioration of the pseudoelastic performance, e.g. in Co–Ni–Ga [21,22]. This method for improving ductility of brittle SMAs was first shown by Nishiziwa and co-workers for B2-type Ni–Al-based alloys [26,27]. Already small volume fractions of ductile secondary phases are sufficient. Based on all aforementioned considerations, samples were quenched into hot water in order to slightly reduce the quenching rate and eventually promote grain boundary segregation of the c-phase at small volume fractions. The EBSD image in Fig. 3a shows a representative coarse grained, bamboo-like sample. Crystallographic orientations are illustrated with respect to the loading direction (LD). Since Tseng et al. [14] demonstrated formation of retained martensite in a [0 0 1] oriented single crystal in tension, even at relatively small strains, grain orientations shown in Fig. 3a are thought to be more beneficial for the evaluation of the influence of c-phase coverage of grain boundaries on the overall pseudoelastic behavior. It is obvious from the insets (Fig. 3b–e) that the slightly lower cooling rate and the slightly higher quenching temperature, are sufficient to introduce a small fraction of the c-phase (1–3 lm thickness) exclusively at the grain boundaries. Since all samples treated this way were free of cracks and martensite, grain boundary decoration reveals to be a suitable way to prevent crack formation. In order to evaluate the impact of c-phase grain boundary decoration on the pseudoelastic stress–strain response, incremental strain tests were carried out in situ using the SEM. The stress–strain response shown in Fig. 4a clearly reveals a good pseudoelastic response up to 8% of nominal strain before the sample failed. Using the in situ measurements, it was possible to assign the two essential stress drops, i.e. at 2.8% and 4.7%, respectively, to the martensitic transformation of a single variant in the [2 0 3] oriented grain and to the transformation of a single variant in the [1 2 3]
3
Fig. 3. Bamboo-like structures after cyclic heat treatment and subsequent quenching into hot water: (a) IPF mapping depicting orientations with respect to the loading direction, (b–e) SEM micrographs of the grain boundaries highlighted in (a) revealing strongly localized formation of the c-phase.
Fig. 4. (a) Stress–strain curves for the sample shown in Fig. 3 obtained by in situ measurements, (b) BSE image highlighting the area of final fracture (grain boundaries are marked by the red dashed lines), (c) SE micrograph depicting the fracture surface. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
oriented grain (not shown for the sake of brevity). Omori et al. [17] made similar observations in Fe–Mn–Al–Ni wires with a high grain diameter to wire diameter ratio and they assumed that such stress drops are related to nucleation barriers for the martensite plates. After the initial activation of these martensite variants no further peak stress level is seen concomitant to the repeated transformation events in the subsequent cycles. Almost the complete grains within the gage length were transformed before the sample failed and the residual strain after the last completed cycle was only about 0.9% of nominal strain. Considering the actual position of sample failure, shown in Fig. 4b, it is important to note that a transgranular crack occurred instead of an intergranular one. In comparison with the fracture surface of the cold water quenched
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
230 231 232 233 234 235 236 237 238 239 240 241 242
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 4 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
specimens (Fig. 1c), the fracture surface shown in Fig. 4c is more serrated revealing a more ductile failure. Based on the different quenching treatments and their impact on the pseudoelastic performance of Fe–Mn–Al–Ni SMA, the following conclusions can be drawn: Rapid quenching of Fe–Mn–Al–Ni into cold water results in crack formation mainly linked to the grain boundaries. Multi-variant martensite was found to form in the vicinity of those grain boundaries. This is attributed to high internal stresses due to high quenching rates, pronounced undercooling and the highly anisotropic nature of SMAs, respectively. Based on rapid segregation and phase decomposition, a significant fraction of the ductile c-phase formed mainly at the grain boundaries. Additionally, small precipitates were found in the grain interior after air cooling. Through suitable quenching conditions, i.e. slower quenching rate and increased quenching temperature, it is feasible to control the amount of the secondary phase at the grain boundaries. Grain boundary decoration following quenching into hot water suppressed crack formation and evolution of martensite. In situ incremental strain testing clearly revealed that the grain boundary precipitation does not affect the pseudoelastic behavior detrimentally. Final failure occurs through transgranular fracture as opposed to intergranular fracture in rapidly quenched, c-phase-free samples.
268 269
Acknowledgments
270
The authors acknowledge financial support by Deutsche Forschungsgemeinschaft (Contract No. NI1327/7-1). I. Karaman acknowledges financial support from the U.S. National Science Foundation, under Grant No. DMR 08-44082, funding the International Institute for Multifunctional Materials for Energy Conversion at Texas A&M University.
271 272 273 274 275
References
276
[1] A. Sato, E. Chishima, K. Soma, T. Mori, Acta Metall. 30 (1982) 1177–1183. [2] A. Sato, E. Chishima, Y. Yamaji, T. Mori, Acta Metall. 32 (1984) 539–547. [3] C.M. Wayman, K. Otsuka, Shape Memory Materials, first ed., Cambridge University Press, Cambridge, New York, 1999. [4] S. Kajiwara, Mater. Sci. Eng. A 273–275 (1999) 67–88. [5] A. Ariapour, D.D. Perovic, I. Yakubtsov, Metall. Mater. Trans. A 32 (2001) 1621– 1628. [6] A. Sato, H. Kubo, T. Maruyama, Mater. Trans. 47 (2006) 571–579. [7] T. Maki, K. Kobayashi, M. Minato, I. Tamura, Scr. Metall. 18 (1984) 1105–1109. [8] R. Hayashi, S.J. Murray, M. Marioni, S.M. Allen, R.C. O’Handley, Sens. Actuators A 81 (2000) 219–223. [9] Y. Tanaka, Y. Himuro, T. Omori, Y. Sutou, R. Kainuma, K. Ishida, Mater. Sci. Eng. A 438–440 (2006) 1030–1035. [10] H. Sehitoglu, I. Karaman, X. Zhang, Y. Chumlyakov, H. Maier, Scripta Mater. 44 (2001) 779–784. [11] Y. Tanaka, Y. Himuro, R. Kainuma, Y. Sutou, T. Omori, K. Ishida, Science (New York, N.Y.) 327 (2010) 1488–1490. [12] P. Krooß, C. Somsen, T. Niendorf, M. Schaper, I. Karaman, Y. Chumlyakov, G. Eggeler, H.J. Maier, Acta Mater. 79 (2014) 126–137. [13] T. Omori, K. Ando, M. Okano, X. Xu, Y. Tanaka, I. Ohnuma, R. Kainuma, K. Ishida, Science (New York, N.Y.) 333 (2011) 68–71. [14] L.W. Tseng, J. Ma, S.J. Wang, I. Karaman, M. Kaya, Z.P. Luo, Y.I. Chumlyakov, Acta Mater. 89 (2015) 374–383. [15] K. Ando, T. Omori, I. Ohnuma, R. Kainuma, K. Ishida, Appl. Phys. Lett. 95 (2009) 212504. [16] T. Omori, M. Nagasako, M. Okano, K. Endo, R. Kainuma, Appl. Phys. Lett. 101 (2012) 231907. [17] T. Omori, M. Okano, R. Kainuma, APL Mater. 1 (2013) 032103. [18] S.M. Ueland, Y. Chen, C.A. Schuh, Adv. Funct. Mater. 22 (2012) 2094–2099. [19] S.M. Ueland, C.A. Schuh, Acta Mater. 60 (2012) 282–292. [20] S.M. Ueland, C.A. Schuh, J. Appl. Phys. 114 (2013) 053503. [21] M. Vollmer, P. Krooß, C. Segel, A. Weidner, A. Paulsen, J. Frenzel, M. Schaper, G. Eggeler, H.J. Maier, T. Niendorf, J. Alloy. Compd. 633 (2015) 288–295. [22] E. Dogan, I. Karaman, Y.I. Chumlyakov, Z.P. Luo, Acta Mater. 59 (2011) 1168– 1183. [23] T. Omori, T. Kusama, S. Kawata, I. Ohnuma, Y. Sutou, Y. Araki, K. Ishida, R. Kainuma, Science (New York, N.Y.) 341 (2013) 1500–1502. [24] K.-H. Hwang, C.-M. Wan, J.G. Byrne, Scr. Metall. Mater. 24 (1990) 979–984. [25] W.-C. Cheng, H.-Y. Lin, C.-F. Liu, Mater. Sci. Eng. A 335 (2002) 82–88. [26] K. Ishida, R. Kainuma, N. Ueno, T. Nishizawa, MTA 22 (1991) 441–446. [27] R. Kainuma, K. Ishida, T. Nishizawa, MTA 23 (1992) 1147–1153.
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
318
No. of Pages 4, Model 5G
11 June 2015 Scripta Materialia xxx (2015) xxx–xxx 1
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat 5 6 3 4 7 8 9 10 11 12 1 2 4 7 15 16 17 18 19 20 21 22 23 24 25 26
On the effect of gamma phase formation on the pseudoelastic performance in polycrystalline Fe–Mn–Al–Ni shape memory alloys M. Vollmer a,⇑, C. Segel a, P. Krooß a, J. Günther a, L.W. Tseng b, I. Karaman b, A. Weidner a, H. Biermann a, T. Niendorf a a b
Technische Universität Bergakademie Freiberg, Institut für Werkstofftechnik (Materials Engineering), 09599 Freiberg, Germany Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA
a r t i c l e
i n f o
Article history: Received 8 May 2015 Revised 4 June 2015 Accepted 9 June 2015 Available online xxxx Keywords: Shape memory alloys (SMAs) Microstructure Quenching Precipitation Mechanical properties
a b s t r a c t In recent years, iron-based shape-memory-alloys such as Fe–Mn–Al–Ni came into focus, due to superior pseudoelastic performance and concomitantly low costs for alloying elements. The formation of the ductile c-phase in polycrystalline Fe–Mn–Al–Ni and its impact on the pseudoelastic performance have not been addressed, yet. Based on modification of quenching conditions this study shows that controlled c-phase precipitation at grain boundaries can suppress intergranular cracking without affecting pseudoelasticity. In situ incremental strain tests revealed fairly good recoverability up to about 8% strain. Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.
28 29 30 31
32 33 34 35
36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
58 59 60 61
Iron based shape memory alloys (SMAs) such as Fe–Ni–Cobased and Fe–Mn-based alloys have attracted considerable attention in recent years [1–12]. Following the development of Fe–Ni–Co–Ti SMAs, which are characterized by relatively low reversible transformation strains [7–10], Fe–Ni–Co–Al-based alloys were introduced showing almost perfect pseudoelastic strains up to 13% in the polycrystalline state, as shown for Fe–Ni–Co–Al–Ta by Tanaka, Kainuma and co-workers [11]. More recently, Omori et al. [13] introduced a new Fe–Mn-based SMA, namely Fe–Mn–A l–Ni, showing about 5% pseudoelastic strain in the polycrystalline state in single cycle tests and an extremely low slope for the Clausius–Clapeyron relationship (0.53 MPa °C1) over a wide temperature range (196 °C to 240 °C). In single crystalline samples with [0 0 1] orientation even more than 20% transformation strain was demonstrated by Tseng et al. [14] under tensile loading. Due to these excellent properties and the cheap alloying elements Fe–Mn–Al–Ni SMA became a candidate material for use in aerospace, automobile and seismic applications. The high amount of Mn in Fe–Mn–Al alloys introduces a change in a/c equilibria and, thus, an unusual transformation between an a (A2 – bcc) austenitic high-temperature parent phase and a c0 (A1/2M – fcc) martensitic product phase [13,15]. In order to promote a thermoelastic transformation, Omori et al. added Ni to Fe–Mn–Al resulting in the formation of nano-sized, coherent b ⇑ Corresponding author.
(B2 – bcc/NiAl) precipitates in the disordered a matrix induced by a low temperature heat treatment [13,16]. These precipitates strengthen the matrix and suppress plastic deformation. Besides the orientation dependency of martensitic transformation in Fe–Mn–Al–Ni [13,14,17], a superior pseudoelastic performance was shown for oligocrystalline, i.e. bamboo-structured samples, characterized by a mean grain diameter to wire diameter [17] or sheet width [13], respectively, ratio of larger than one. Ueland et al. [18–20] demonstrated that bamboo structures with grain sizes larger than the wire diameter and grain boundaries oriented only perpendicular to the loading direction, i.e. preventing grain boundary triple junctions and concomitant grain constraints, led to superior pseudoelastic performance in micro wires of Cubased SMAs. Similarly, damage evolution in bulk polycrystalline Co–Ni–Ga high-temperature alloys was investigated [21], revealing that bamboo-structured samples showed superior performance. It was shown that intergranular fracture is strongly affected not only by the grain boundary arrangement, but also by the formation of a ductile phase decorating the grain boundaries, i.e. the c-phase in Co–Ni–Ga alloys [21,22]. The Fe–Mn–Al–Ni SMA in the present study is prone to intergranular cracking upon rapid quenching. Omori et al. used Mo-sheets during the solution treatment [13]. However, to the best of authors’ knowledge, data reporting on crack formation with an emphasis on its underlying mechanisms are still lacking in open literature.
http://dx.doi.org/10.1016/j.scriptamat.2015.06.013 1359-6462/Ó 2015 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 2 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
The present study focusses on the role of segregations in Fe– Mn–Al–Ni evolving during quenching, their impact on the pseudoelastic behavior as well as the morphology and sites of the phases formed as a result of the different quenching conditions. Tension samples were first cyclically heat treated in order to obtain a coarse-grained microstructure, similar to the procedure reported for a Cu–Al–Mn SMA by Omori et al. [23]. They were able to show that by increasing the fraction of semi-coherent second phase particles during cooling into a two phase region, the development of subgrain structures is promoted. Omori et al. attributed this to the loss of coherency between the matrix and the particles and concurrent dislocation formation. Upon heating to temperatures above the solvus temperature, the subgrains remain and the associated internal stress triggers abnormal grain growth [23]. However, this kind of procedure results in intense crack formation upon final quenching to room temperature in Fe–Mn–Al–Ni. Commercially pure polycrystalline Fe–Mn–Al–Ni, produced by vacuum induction melting, with a nominal chemical composition of Fe–34% Mn–14% Al–7.5% Ni (at.%) was used in the current work. The alloy was cast by Stahlzentrum Freiberg e.V. (Freiberg, Germany) and subsequently hot-forged at 1150 °C. Dog-boneshaped tension samples were cut using electro discharge machining featuring a gage length of 12 mm and a cross section of 1.6 mm 1.5 mm. Subsequently, samples were sealed in quartz tubes under high vacuum. In order to obtain a high fraction of the second phase, samples were solution treated at 1200 °C for 12 h and then air cooled for 5 s before quenching into 15 °C cold water. The second set of samples was solution heat treated at 1200 °C for 30 min, cooled down into the a + c two phase region at 900 °C within 30 min, held for 15 min and then heated up to 1200 °C within 30 min again. This procedure was repeated for four times. Finally, the samples were homogenized at 1200 °C for 1 h to ensure that they were free of the c-phase. In order to generate two different conditions, some samples were quenched into water at 15 °C, referred to as cold water in the remainder of the text, and the others were quenched into water heated up to 80 °C, referred to as hot water. All samples were subsequently aged at 200 °C for 3 h in air, ground down to 5 lm grit size and finally vibration-polished using colloidal SiO2 suspension with 0.02 lm particle size. For optical microscopy the air cooled condition was additionally etched using a solution of 7% nitric acid and 93% ethanol. Microstructural analyses were conducted using a scanning electron microscope (SEM) operated at 20 kV equipped with an electron-backscatter diffraction (EBSD) system and an energy dispersive X-ray spectroscopy (EDS) unit. For in situ high-resolution measurements, a miniature load frame was used under a constant crosshead displacement rate of 2 lm s1 in combination with the SEM system. Incremental strain tests were conducted with a step size of 120 lm (e.g. about 1% strain) until the samples failed. Strain values for the stress–strain diagram were then calculated from displacement data. Surface images were taken using either secondary electron (SE) or backscattered electron (BSE) contrasts, with and without superimposed mechanical load. Fig. 1 shows SEM micrographs and results from the EBSD measurements for the cold water quenched state after cyclic heat treatment revealing no traces of evolution of second phases. It is obvious from Fig. 1a that this state is prone to crack formation and further investigations revealed that a high density of multi-variant martensite appeared in the vicinity of the cracks (not shown). Since temperature induced martensitic transformation can be excluded according to the magnetic measurements conducted by Omori et al. [13], high internal stresses resulting from quenching in cold water are expected to be responsible for the martensitic transformation. The EBSD scan and the associated inverse pole figure (IPF) in Fig. 1b reveal that crack formation
Fig. 1. Characteristic crack formation during cold water quenching after cyclic heat treatment: (a) SEM micrograph depicting crack formation, (b) EBSD inverse pole figure mapping of the highlighted area depicting orientations with respect to the loading direction, (c) SEM micrograph of a representative fracture surface.
occurred mainly along the grain boundaries. As shown in Fig. 1c, crack surfaces are very smooth revealing a brittle behavior in this condition. A bamboo structure as induced by the cyclic heat treatment seems to promote cracking most probably due to an increasing grain boundary area. SMAs are characterized by pronounced anisotropic material behavior, e.g. high elastic anisotropy [3]. Consequently, grain boundaries of high misorientation, separating large grains of significantly different crystallographic orientation, are prone to rapid crack formation upon deformation [21]. Similarly, the anisotropic nature seems to promote cracking upon quenching at those large grain boundaries. In samples quenched at significantly slower cooling rates, a second phase with a serrated interface formed mainly at the grain boundaries and subsequently grew into the grain interiors (Fig. 2a). The EBSD measurement in Fig. 2b clearly reveals the fcc structure of the second phase and further investigations of a representative, larger section by EBSD revealed a c-phase fraction of about 40% (not shown). Additional EDS measurements (Fig. 2c–f) show a decrease of Al content and a slight increase of Mn content, whereas changes in Fe and Ni can hardly be resolved. Al is known to be an a-phase stabilizer and Mn is a c-phase stabilizer. These local segregations form a Widmanstätten-like c-phase especially at the grain boundaries. The slightly different alloy composition with a lower content of Al and Ni in comparison with the alloy introduced by Omori et al. [13] additionally promotes c-phase formation. Fast precipitation of the c-phase along grain boundaries in Fe–Mn– Al alloys, evolving despite rapid water quenching, was also observed in other studies [15,24,25]. Hwang et al. [24] attributed the segregations to a massive transformation due to short-range diffusion of Mn and Al across the a/c boundaries. Another microstructural feature resulting from fast segregation in the Fe–Mn–Al–Ni alloy is acicular precipitates inside the grains indicated by the white arrow in Fig. 2g. These are characterized by a length of about 5–10 lm, which is three orders of magnitude larger than the size of the fine b precipitates [16]. Since their volume fraction is much lower in vicinity of c-phase covered grain boundaries, it can be assumed that these segregations are of similar composition. However, detailed analyses will be subject of future work. In contrast to the cold water quenched condition, the slow cooled condition was free of martensite as well as free of cracks along grain boundaries. However, this condition did not show satisfactory pseudoelastic behavior as the high fraction of the non-transforming c-phase is detrimental with respect to recoverability of martensitic transformation.
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188
189 190 191 192 193 194 195 196
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
Fig. 2. c-Phase formation following slow cooling: (a) optical micrograph depicting the etched surface, (b) EBSD phase mapping, (c–f) EDS element mappings, and (g) SEM micrograph of an area showing fine acicular precipitates (white arrow) in the grain interior.
197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
It is well known that grain boundary decoration using suitable ductile second phases can reduce crack susceptibility in brittle SMAs without deterioration of the pseudoelastic performance, e.g. in Co–Ni–Ga [21,22]. This method for improving ductility of brittle SMAs was first shown by Nishiziwa and co-workers for B2-type Ni–Al-based alloys [26,27]. Already small volume fractions of ductile secondary phases are sufficient. Based on all aforementioned considerations, samples were quenched into hot water in order to slightly reduce the quenching rate and eventually promote grain boundary segregation of the c-phase at small volume fractions. The EBSD image in Fig. 3a shows a representative coarse grained, bamboo-like sample. Crystallographic orientations are illustrated with respect to the loading direction (LD). Since Tseng et al. [14] demonstrated formation of retained martensite in a [0 0 1] oriented single crystal in tension, even at relatively small strains, grain orientations shown in Fig. 3a are thought to be more beneficial for the evaluation of the influence of c-phase coverage of grain boundaries on the overall pseudoelastic behavior. It is obvious from the insets (Fig. 3b–e) that the slightly lower cooling rate and the slightly higher quenching temperature, are sufficient to introduce a small fraction of the c-phase (1–3 lm thickness) exclusively at the grain boundaries. Since all samples treated this way were free of cracks and martensite, grain boundary decoration reveals to be a suitable way to prevent crack formation. In order to evaluate the impact of c-phase grain boundary decoration on the pseudoelastic stress–strain response, incremental strain tests were carried out in situ using the SEM. The stress–strain response shown in Fig. 4a clearly reveals a good pseudoelastic response up to 8% of nominal strain before the sample failed. Using the in situ measurements, it was possible to assign the two essential stress drops, i.e. at 2.8% and 4.7%, respectively, to the martensitic transformation of a single variant in the [2 0 3] oriented grain and to the transformation of a single variant in the [1 2 3]
3
Fig. 3. Bamboo-like structures after cyclic heat treatment and subsequent quenching into hot water: (a) IPF mapping depicting orientations with respect to the loading direction, (b–e) SEM micrographs of the grain boundaries highlighted in (a) revealing strongly localized formation of the c-phase.
Fig. 4. (a) Stress–strain curves for the sample shown in Fig. 3 obtained by in situ measurements, (b) BSE image highlighting the area of final fracture (grain boundaries are marked by the red dashed lines), (c) SE micrograph depicting the fracture surface. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
oriented grain (not shown for the sake of brevity). Omori et al. [17] made similar observations in Fe–Mn–Al–Ni wires with a high grain diameter to wire diameter ratio and they assumed that such stress drops are related to nucleation barriers for the martensite plates. After the initial activation of these martensite variants no further peak stress level is seen concomitant to the repeated transformation events in the subsequent cycles. Almost the complete grains within the gage length were transformed before the sample failed and the residual strain after the last completed cycle was only about 0.9% of nominal strain. Considering the actual position of sample failure, shown in Fig. 4b, it is important to note that a transgranular crack occurred instead of an intergranular one. In comparison with the fracture surface of the cold water quenched
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
230 231 232 233 234 235 236 237 238 239 240 241 242
SMM 10682
No. of Pages 4, Model 5G
11 June 2015 4 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
M. Vollmer et al. / Scripta Materialia xxx (2015) xxx–xxx
specimens (Fig. 1c), the fracture surface shown in Fig. 4c is more serrated revealing a more ductile failure. Based on the different quenching treatments and their impact on the pseudoelastic performance of Fe–Mn–Al–Ni SMA, the following conclusions can be drawn: Rapid quenching of Fe–Mn–Al–Ni into cold water results in crack formation mainly linked to the grain boundaries. Multi-variant martensite was found to form in the vicinity of those grain boundaries. This is attributed to high internal stresses due to high quenching rates, pronounced undercooling and the highly anisotropic nature of SMAs, respectively. Based on rapid segregation and phase decomposition, a significant fraction of the ductile c-phase formed mainly at the grain boundaries. Additionally, small precipitates were found in the grain interior after air cooling. Through suitable quenching conditions, i.e. slower quenching rate and increased quenching temperature, it is feasible to control the amount of the secondary phase at the grain boundaries. Grain boundary decoration following quenching into hot water suppressed crack formation and evolution of martensite. In situ incremental strain testing clearly revealed that the grain boundary precipitation does not affect the pseudoelastic behavior detrimentally. Final failure occurs through transgranular fracture as opposed to intergranular fracture in rapidly quenched, c-phase-free samples.
268 269
Acknowledgments
270
The authors acknowledge financial support by Deutsche Forschungsgemeinschaft (Contract No. NI1327/7-1). I. Karaman acknowledges financial support from the U.S. National Science Foundation, under Grant No. DMR 08-44082, funding the International Institute for Multifunctional Materials for Energy Conversion at Texas A&M University.
271 272 273 274 275
References
276
[1] A. Sato, E. Chishima, K. Soma, T. Mori, Acta Metall. 30 (1982) 1177–1183. [2] A. Sato, E. Chishima, Y. Yamaji, T. Mori, Acta Metall. 32 (1984) 539–547. [3] C.M. Wayman, K. Otsuka, Shape Memory Materials, first ed., Cambridge University Press, Cambridge, New York, 1999. [4] S. Kajiwara, Mater. Sci. Eng. A 273–275 (1999) 67–88. [5] A. Ariapour, D.D. Perovic, I. Yakubtsov, Metall. Mater. Trans. A 32 (2001) 1621– 1628. [6] A. Sato, H. Kubo, T. Maruyama, Mater. Trans. 47 (2006) 571–579. [7] T. Maki, K. Kobayashi, M. Minato, I. Tamura, Scr. Metall. 18 (1984) 1105–1109. [8] R. Hayashi, S.J. Murray, M. Marioni, S.M. Allen, R.C. O’Handley, Sens. Actuators A 81 (2000) 219–223. [9] Y. Tanaka, Y. Himuro, T. Omori, Y. Sutou, R. Kainuma, K. Ishida, Mater. Sci. Eng. A 438–440 (2006) 1030–1035. [10] H. Sehitoglu, I. Karaman, X. Zhang, Y. Chumlyakov, H. Maier, Scripta Mater. 44 (2001) 779–784. [11] Y. Tanaka, Y. Himuro, R. Kainuma, Y. Sutou, T. Omori, K. Ishida, Science (New York, N.Y.) 327 (2010) 1488–1490. [12] P. Krooß, C. Somsen, T. Niendorf, M. Schaper, I. Karaman, Y. Chumlyakov, G. Eggeler, H.J. Maier, Acta Mater. 79 (2014) 126–137. [13] T. Omori, K. Ando, M. Okano, X. Xu, Y. Tanaka, I. Ohnuma, R. Kainuma, K. Ishida, Science (New York, N.Y.) 333 (2011) 68–71. [14] L.W. Tseng, J. Ma, S.J. Wang, I. Karaman, M. Kaya, Z.P. Luo, Y.I. Chumlyakov, Acta Mater. 89 (2015) 374–383. [15] K. Ando, T. Omori, I. Ohnuma, R. Kainuma, K. Ishida, Appl. Phys. Lett. 95 (2009) 212504. [16] T. Omori, M. Nagasako, M. Okano, K. Endo, R. Kainuma, Appl. Phys. Lett. 101 (2012) 231907. [17] T. Omori, M. Okano, R. Kainuma, APL Mater. 1 (2013) 032103. [18] S.M. Ueland, Y. Chen, C.A. Schuh, Adv. Funct. Mater. 22 (2012) 2094–2099. [19] S.M. Ueland, C.A. Schuh, Acta Mater. 60 (2012) 282–292. [20] S.M. Ueland, C.A. Schuh, J. Appl. Phys. 114 (2013) 053503. [21] M. Vollmer, P. Krooß, C. Segel, A. Weidner, A. Paulsen, J. Frenzel, M. Schaper, G. Eggeler, H.J. Maier, T. Niendorf, J. Alloy. Compd. 633 (2015) 288–295. [22] E. Dogan, I. Karaman, Y.I. Chumlyakov, Z.P. Luo, Acta Mater. 59 (2011) 1168– 1183. [23] T. Omori, T. Kusama, S. Kawata, I. Ohnuma, Y. Sutou, Y. Araki, K. Ishida, R. Kainuma, Science (New York, N.Y.) 341 (2013) 1500–1502. [24] K.-H. Hwang, C.-M. Wan, J.G. Byrne, Scr. Metall. Mater. 24 (1990) 979–984. [25] W.-C. Cheng, H.-Y. Lin, C.-F. Liu, Mater. Sci. Eng. A 335 (2002) 82–88. [26] K. Ishida, R. Kainuma, N. Ueno, T. Nishizawa, MTA 22 (1991) 441–446. [27] R. Kainuma, K. Ishida, T. Nishizawa, MTA 23 (1992) 1147–1153.
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
Please cite this article in press as: M. Vollmer et al., Scripta Mater. (2015), http://dx.doi.org/10.1016/j.scriptamat.2015.06.013
318