www.sciencemag.org/cgi/content/full/333/6038/68/DC1

Supporting Online Material for Superelastic Effect in Polycrystalline Ferrous Alloys T. Omori,* K. Ando, M. Okano, X. Xu, Y. Tanaka, I. Ohnuma, R. Kainuma, K. Ishida *To whom correspondence should be addressed. E-mail: [email protected] Published 1 July 2011, Science 333, 68 (2011) DOI: 10.1126/science.1202232

This PDF file includes: Materials and Methods SOM Text Figs. S1 to S6 Table S1 References

Supporting Online Material for

Superelastic Effect in Polycrystalline Ferrous Alloys T. Omori*, K. Ando, M. Okano, X. Xu, Y. Tanaka, I. Ohnuma, R. Kainuma, K. Ishida Department of Materials Science, Graduate School of Engineering, Tohoku University, 6-6-02 Aoba-yama, Sendai 980-8579, Japan. *

To whom correspondence should be addressed. E-mail: [email protected]

1. Materials and methods Ingots of Fe43.5Mn34Al15Ni7.5 were prepared by induction melting in an argon atmosphere. The ingots were then hot-rolled at 1200 ºC and finally cold-rolled.

Thin sheet specimens

were sandwiched between Mo sheets and solution-treated at 1200 ºC followed by air cooling, the process being repeated to obtain coarse grains with a mean size of about 3.8 mm and a bamboo structure to provide good superelastic properties (Fig. S5).

Here, the

specimens have no strong texture. They were finally quenched from 1200 °C into water and were subsequently aged at 200 ºC for 15min, 3 hours, 6 hours or 24 hours to obtain fine CsCl-type ordered bcc (B2) precipitates where the aging temperature was selected on the basis of the experimental result that the aging treatment at higher temperature than 300 °C results in the decrease in ductility.

Magnetic properties were examined using a

superconducting quantum interference device (SQUID) for lower temperature range at a cooling and heating rate of 2 K min-1 and by a vibrating sample magnetometer (VSM) for temperatures higher than about 50 °C.

Superelasticity was evaluated by a tensile testing at

1

various temperatures using a sheet specimen of 0.25×1×50 mm with a gauge length of 20mm.

The temperature dependence of the superelastic stress was determined using one

specimen for each aging condition, but for 24 hours shown in Fig. 2B another piece with the same composition was used for the tests at -196 °C and 20 °C. Here, the strain was limited to 1% at each temperature so as to avoid the influence of the cycle on the stress in Fig. 2B.

Thermodynamic analysis was carried out by the CALPHAD approach using the

software package Thermo-Calc.

The input parameters were partially modified from those

in ref. 18 and the fcc (A1) and disordered bcc (A2) phases were considered in the calculation.

2. Supporting online text 2.1. Transformation entropy and the temperature dependence of superelasticity Tensile tests were conducted at temperatures between -60 °C and 210 °C using a Fe43.5Mn34Al15Ni7.5 single crystal, which was obtained by grain growth, aged at 200 °C for 3 hours.

Figure S1A shows the tensile cyclic curve at 30°C.

The transformation strain ε

evaluated from the plateau region during loading is 9.7%. The superelastic stress σc is plotted as a function of temperature T in Fig. S1B, where the slope (dσc/dT) is determined to be 0.60 MPa/°C.

Using these experimental values and the molar volume Vm = 7.366 ×

10-6 (m3 mol-1), the transformation entropy change ΔS is estimated by Clausius-Clapeyron relation (Eq. 1) as being -0.43 (J mol-1 K-1). The ΔS, superelastic strain εSE (= transformation strain ε – residual strain after removal of strain εr) and dσc/dT for various superelastic alloys including single crystals are

2

summarized in Table S1 (15,16,23,28-36).

The ΔS and dσc/dT of Fe-Mn-Al-Ni are very

low compared with other alloy systems.

2.2 Microstructure and crystal structure of Fe-Mn-Al-Ni superelastic alloy Superelasticity is obtained due to a thermoelastic martensitic transformation. Fe-Mn-Al alloy shows a non-thermoelastic martensitic transformation, but the addition of Ni changes the transformation feature to the thermoelastic type.

Figure S2A shows the

dark field image of transmission electron microscopy (TEM) taken from the (100)B2 super-lattice reflection in Fe43.5Mn34Al15Ni7.5 that had been aged at 200 ºC for 6 hours. Nano-sized particles with the B2 phase (ordered bcc structure) are observed in the A2 (α phase: disordered bcc) matrix phase.

This kind of two-phase microstructure of disordered

matrix and coherent ordered precipitate has been reported in Fe-Ni-Co-Ti and Fe-Ni-Co-Al alloys (8,9), which show a thermoelastic martensitic transformation, although the crystal structure is different. Figure S2B shows the TEM bright field image and the corresponding selected area diffraction pattern (SADP) taken from the as-solution-treated Fe43.5Mn34Al15Ni7.5 alloy. Martensite plates were partially observed in this specimen.

While plane defects were

hardly observed in this sample, the SADP indicates that the martensite phase has the 8M stacking layered structure, showing n/8{110}A2 extra-spots.

However, the extra-spots

were often unclear and observed with streaks, and the fundamental reflections show the fcc structure. Since only a small region is observed in the TEM, the XRD experiment was also carried out.

Figure S3 shows the XRD profile of Fe43.5Mn34Al15Ni7.5 alloy aged at

3

200 °C for 3h, subsequently deformed up to about 4% in tension at room temperature and unloaded. The XRD pattern includes reflections from a fcc phase, besides ones from the bcc parent phase, and the lattice parameters of the bcc and fcc phases are abcc = 0.2903 nm and afcc = 0.3672 nm, respectively. as-solution-treated specimen.

The similar result has been obtained in the

From these results, it can be concluded that the martensite

phase basically has the fcc structure and the transformation strains are estimated from lattice correspondence as being 26.5% in the <001> direction, which is much larger than that in conventional superelastic alloys (Table S1), 9.5% in the <011> direction and 3.3% in the <111> direction.

This suggests that as well as the small ΔS, the large

transformation strain also contributes to the small temperature dependence of the superelastic stress according to the Clausius-Clapeyron relation (Eq. 1). Psudoelasticity (i.e. nonlinear elasticity) is known to be originated from several mechanisms including the superelasticity and the rubber-like behavior. In the Fe-based bcc ordered alloys, such as Fe3Al and Fe3Ga, pseudoelastic behavior (37-40) due to superpartial dislocations, with a similar stress-strain response to superelasticity arising from stress-induced martensitic (SIM) transformation, has been reported.

Therefore, we

conducted in-situ observation of the microstructure during tensile testing at room temperature.

Figure S4 shows optical micrographs during loading and unloading in

Fe43.5Mn34Al15Ni7.5 aged at 200 °C for 3 hours.

It is confirmed by magnetization

measurement that although this alloy does not thermally transform to the martensite phase by cooling to 4.2 K, a martensite plate forms in the parent phase by 0.5 % strain and grows almost continuously with increasing applied strain.

During unloading, many parent plates

appear in the martensite domain and the martensite phase finally disappears.

4

Thus, the

martensitic transformation in the Fe43.5Mn34Al15Ni7.5 is thermoelastic and the reversible stress-strain response shown in Fig. 2A is the superelasticity due to the SIM transformation.

2.3 Effect of microstructure on superelasticity Superelasticity is strongly influenced by microstructure. Figure S5 shows the shape recovery ratio R of superelasticity defined as R = (εt - εe - εr)/(εt - εe) × 100 for an Fe43.5Mn34Al15Ni7.5 alloy aged at 200 °C for 6 hours with different grain diameters. εt (= 4% in this case) is the applied strain and εe and εr are the elastic strain and the residual strain after removal of stress in the tensile superelastic test, respectively.

It maybe seen

that the superelastic shape recovery is drastically enhanced in the specimens with relative grain sizes d/w larger than 1, where d and w (= 1.0 mm in this case) are the mean grain diameter and the width of the sheet specimen, respectively. A similar dependence of the relative grain size on superelasticity has been reported in Cu-Al-Mn-based shape memory alloys with large anisotropy of transformation (41), where this behavior is explained by grain constraint from neighbouring grains.

In the case of d/w < 1, the grains are

surrounded by many neighbouring grains, and the transformation strain yielded from each grain by external stress is constrained by the neighbouring grains.

This condition makes

the superelastic stress increase and slip deformation is introduced, leading to a deterioration in superelasticity.

The sheet specimen with d/w > 1 has a bamboo structure, where the

grain constraint from neighbouring grains decreases, and the superelasticity is excellent.

5

Thus, the microstructural control is also important to bring about an excellent superelasticity in Fe-Mn-Al-Ni.

2.4 Magnetic property Figure S6 shows the magnetization curves of Fe43.5Mn34Al15Ni7.5 aged at 200 °C for 6 hours for various strains at room temperature. Ferromagnetism is observed at zero strain state.

With increasing strain, the spontaneous magnetization decreases, because the

ferromagnetic parent phase gradually transforms to the antiferromagnetic martensite phase. On the other hand, the spontaneous magnetization reversibly increases with decreasing strain due to reverse transformation. This change of magnetization induced by loading is opposite to that found in conventional Fe-based alloys that exhibit an irreversible martensitic transformation from a paramagnetic parent phase to ferromagnetic martensite. This unique property can be applied to a noncontact strain sensor, such as those used for monitoring deformation of buildings by earthquake or of vehicles damaged by accidents.

6

800

A

Stress (MPa)

700

111

600 500

001

400

101

300 200 100 0

0

Superelastic stress (MPa)

500

2

4

6

8 10 Strain (%)

12

14

16

0 50 100 150 Temperature (°C)

200

250

B

400 300 200 100 0 -100 -50

Fig. S1 Superelasticity of Fe43.5Mn34Al15Ni7.5 single crystal aged at 200 °C for 3 hours. (A) Cyclic stress strain curve at 30 °C. The specimen was first loaded in tension up to a strain of 2.5% and unloaded and then loaded up to 4.5% and unloaded in the second cycle and so forth, and was finally fractured. The tensile direction of the single crystal is shown in the inset. (B) Temperature dependence of superelatic stress of the single crystal.

A 110 000

100

200

110

200nm a.

B 114 000

114 008

500nm

Fig. S2 TEM images of the parent and martensite phases in Fe43.5Mn34Al15Ni7.5. (A) Selected area diffraction (SAD) pattern and dark-field (DF) image around the (100)B2 super-lattice reflection taken from a Fe43.5Mn34Al15Ni7.5 alloy aged at 200 °C for 6 hours. Fine particles with the B2 structure precipitate in the A2 parent phase. (B) SAD pattern and bright-field image of the martensite phase taken for the Fe43.5Mn34Al15Ni7.5 alloy without aging.

(211)α

(110)α

30

40

50 60 70 2θ (degree)

80

(222)γ’ (220)α

(311)γ’

(220)γ’

(200)α

(200)γ’

(111)γ’

Intensity 20

90 100

Fig. S3 XRD pattern of Fe43.5Mn34Al15Ni7.5 alloy after deformation in tension. Parent and martensite phases are identified as being bcc (α) and fcc (γ’) structures.

Loading (1) ε=0.5%

(3) ε=4%

(2) ε=2%

P

P

P

M

M M

(6) ε=0.15%

(5) ε=1%

P

P

(4) ε=1.75%

M

Unloading

200μm

Fig. S4 In-situ observation, using an optical microscope, of the microstructure during loading and unloading at room temperature in Fe43.5Mn34Al15Ni7.5 after ageing at 200 °C for 3 hours. Stressinduced martensite plates are observed during loading and disappear during unloading. This reversible microstructural change means that this transformation is thermoelastic.

Shape recovery ratio (%)

100

Relative grain size for thickness, d/t 0 5 10 15 20

90 80

d

70

t

60

w

50 40 30 20

0

1 2 3 4 Relative grain size for width, d/w

5

Fig. S5 The shape recovery ratio of superelasticity vs. relative grain size for specimen width and thickness, d/w and d/t, in Fe43.5Mn34Al15Ni7.5 aged at 200 °C for 6 hours, where d is the mean grain diameter, w (= 1 mm) and t ( = 0.25 mm ) are the width and thickness of sheet specimen, respectively. The strain of 4% was applied in tension at room temperature. The superelasticity is apparently enhanced in the region of the relative grain size for widths of over d/w = 1, owing to the decrease in grain constraint from neighbouring grains during deformation.

Magnetization (emu g-1)

70 60

Loading 1.1%

50

2.2%

40 30

3.4% 4.1%

20

4.9%

1.1% 3.4% 4.1% 4.9%

10 0

ε=0% (before loading) 0.7%

5.2%

0

2

4

Unloading

6 8 10 12 14 16 18 20 Magnetic field (kOe)

Magnetization curves of Fe43.5Mn34Al15Ni7.5 alloy aged at Fig. S6 200 ° C for 6 hours at various fixed strains in the loading process (solid lines) and unloading process (dashed lines) at room temperature. Due to the superelasticity associated with reversible phase transformations between the ferromagnetic parent and antiferromagnetic martensite phases, spontaneous magnetization decreases on loading and increases on unloading, reversibly.

Table S1 The transformation entropy change ΔS, superelastic strain εSE, temperature dependence of the superelastic stress dσc/dT of various single crystal and polycrystalline superelastic alloys. The εSE and dσc/dT are values taken under tension except for Ni-Mn-Ga single crystal (compression).

Alloy system

Crystal structure

Chemical composition (at%)

ΔS (J mol-1K-1)

εSE (%)

dσc

Ref.

dT dσc (MPa °C-1) ΔS εSE dT

Parent

Martensite

Fe-Mn-Al-Ni

BCC (A2+B2)

FCC

Fe43.5Mn34Al15Ni7.5

-0.43

5.2

0.53

*

*

*

Fe-Ni-Co-Al

FCC

BCT

Fe40.95Ni28Co17Al11.5Ta2.5B0.05

-2.95

13.5

3.1

*

*

*

Ti-Ni

B2

B19'

Ti50Ni50

-4.37

7.3

5.7

23

*

*

Ti-Ni-Cu

B2

B19

Ti50Ni40Cu10

-2.92

3.2

8.4

28

28

28

D03

6M

-1.03

8.6

†

1.5

‡

29

29

D03

2O

-1.20

4.7

†

3.4

‡

29

29

Cu-Zn-Al

D03

6M

Cu67.9Zn16.1Al16

-1.45

8.5

†

2.1

‡

30

30

30

Cu-Al-Mn

L21

6M

Cu71.9Al16.6Mn9.3Ni2B0.2

-1.15

7.5

15

31

15

Ni-Mn-Ga

L21

10M

Ni52Mn24.4Ga23.6

-0.97

4.0

†

3.5

‡

32

33

32

Ni-Fe-Ga

L21

14M

Ni51Fe22Ga27

-0.91

6.2

†

1.9

‡

34

35

35

BCC

Orthorhombic

Ti74Nb26

-1.30

2.3

16

36

16

Cu-Al-Ni

Ti-Nb

Cu68.4Al27.8Ni3.8

* this work † maximum strain in single crystal ‡ calculated or experimental value of minimum temperature dependence in single crystal

2.4

4.4

References and Notes 1. A. Ölander, An electrochemical investigation of solid cadmium-gold alloys. J. Am. Chem. Soc. 56, 3819 (1932). doi:10.1021/ja01349a004 2. K. Otsuka, C. M. Wayman, in Shape Memory Materials, K. Otsuka, C. M. Wayman, Eds. (Cambridge Univ. Press, Cambridge, 1998), pp. 1–48. 3. T. Kakeshita, K. Ullakko, Giant magnetostriction in ferromagnetic shape-memory alloys. MRS Bull. 27, 105 (2002). doi:10.1557/mrs2002.45 4. R. Kainuma et al., Magnetic-field-induced shape recovery by reverse phase transformation. Nature 439, 957 (2006). doi:10.1038/nature04493 Medline 5. T. W. Duerig, K. N. Melton, D. Stöckel, C. M. Wayman, Engineering Aspects of Shape Memory Alloys (Butterworth-Heinemann, London, 1990). 6. P. Wollants, M. De Bonte, J. R. Roos, Thermodynamic analysis of the stress-induced martensitic-transformation in a single-crystal. Z. Metallk. 70, 113 (1979). 7. K. Ando, T. Omori, I. Ohnuma, R. Kainuma, K. Ishida, Ferromagnetic to weakmagnetic transition accompanied by bcc to fcc transformation in Fe-Mn-Al alloy. Appl. Phys. Lett. 95, 212504 (2009). doi:10.1063/1.3266848 8. T. Maki, K. Kobayashi, M. Minato, I. Tamura, Thermoelastic martensite in an ausaged Fe-Ni-Ti-Co alloy. Scr. Metall. 18, 1105 (1984). doi:10.1016/00369748(84)90187-X 9. Y. Tanaka et al., Ferrous polycrystalline shape-memory alloy showing huge superelasticity. Science 327, 1488 (2010). doi:10.1126/science.1183169 Medline 10. R. Kainuma, M. Ise, K. Ishikawa, I. Ohnuma, K. Ishida, Phase equilibria and stability of the B2 phase in Ni-Mn-Al and Co-Mn-Al systems. J. Alloy. Comp. 269, 173 (1998). doi:10.1016/S0925-8388(98)00127-3 11. S. M. Hao, T. Takayama, K. Ishida, T. Nishizawa, Miscibility gap in Fe-Ni-Al and Fe-Ni-Co system. Metall. Trans. 15, 1819 (1984). doi:10.1007/BF02664895 12. Materials and methods are available as supporting material on Science Online. 13. T. Maki, in Shape Memory Materials, K. Otsuka, C. M. Wayman, Eds. (Cambridge Univ. Press, Cambridge, 1998), pp. 117–132. 14. L. Muldawer, F. de Bergevin, Antiferromagnetic-ferromagnetic transformation in FeRh. J. Chem. Phys. 35, 1904 (1961). doi:10.1063/1.1732175 15. Y. Sutou, N. Koeda, T. Omori, R. Kainuma, K. Ishida, Effect of aging on stressinduced martensitic transformation in ductile Cu-Al-Mn-based shape memory alloys. Acta Mater. 57, 5759 (2009). doi:10.1016/j.actamat.2009.08.011 16. H. Y. Kim, S. Hashimoto, J. I. Kim, H. Hosoda, S. Miyazaki, Mechanical properties and shape memory behaviour of Ti-Nb alloys. Mater. Trans. 45, 2443 (2004). doi:10.2320/matertrans.45.2443

17. J. Ma, I. Karaman, R. D. Noebe, High temperature shape memory alloys. Int. Mater. Rev. 55, 257 (2010). doi:10.1179/095066010X12646898728363 18. R. Umino et al., Experimental determination and thermodynamic calculation of phase equilibria in the Fe-Mn-Al system. J. Phase Equilibria Diffus. 27, 54 (2006). doi:10.1361/105497106X92817 19. H. C. Tong, C. M. Wayman, Characteristic temperatures and other properties of thermoelastic martensites. Acta Metall. 22, 887 (1974). doi:10.1016/00016160(74)90055-8 20. C. Zener, Impact of magnetism upon metallurgy. J. Met. 7, 619 (1955). 21. S. S. Yan et al., Anomalous behaviour of electrical resistivity in FeMnAl alloys. Solid State Commun. 54, 831 (1985). doi:10.1016/0038-1098(85)90296-0 22. G. A. Pérez Alcazar, E. Galvão da Silva, C. Paduani, Magnetic properties of Fe-MnAl alloy system in the FCC disordered phase. Hyperfine Int. 66, 221 (1991). doi:10.1007/BF02395869 23. W. Tang, R. Sandström, Z. G. Wei, S. Miyazaki, Experimental investigation and thermodynamic calculation of the Ti-Ni-Cu shape memory alloys. Metall. Mater. Trans., A Phys. Metall. Mater. Sci. 31, 2423 (2000). doi:10.1007/s11661-0000187-y 24. C. L. Magee, R. G. Davies, On the expansion accompanying the f.c.c. to b.c.c. transformation in ferrous alloys. Acta Metall. 20, 1031 (1972). doi:10.1016/00016160(72)90137-X 25. E. J. Graesser, F. A. Cozzarelli, Shape-memory alloys as new materials for aseismic isolation. J. Eng. Mech. 117, 2590 (1991). doi:10.1061/(ASCE)07339399(1991)117:11(2590) 26. M. Dolce, D. Cardone, R. Marnetto, Implementation and testing of passive control devices based on shape memory alloys. Earthquake Eng. Struct. Dynam. 29, 945 (2000). doi:10.1002/1096-9845(200007)29:7<945::AID-EQE958>3.0.CO;2-# 27. A. Abbott, Rescuers of Europe’s cultural heritage struggle for funding. Nature 414, 572 (2001). doi:10.1038/414572b Medline 28. T. H. Nam, T. Saburi, K. Shimizu, Cu-content dependence of shape memory characteristics in Ti-Ni-Cu alloys. Mater. Trans. JIM 31, 959 (1990). 29. K. Otsuka, C. M. Wayman, K. Nakai, H. Sakamoto, K. Shimizu, Superelasticity effects and stress-induced martensitic transformations in Cu-Al-Ni alloys. Acta Metall. 24, 207 (1976). doi:10.1016/0001-6160(76)90071-7 30. T. Saburi, Y. Inada, S. Nenno, N. Hori, Stress-induced martensitic transformations in Cu-Zn-Al and Cu-Zn-Ga alloys. J. Phys. 43, 633 (1982). 31. Y. Sutou, T. Omori, R. Kainuma, K. Ishida, Ductile Cu-Al-Mn based shape memory alloys: General properties and applications. Mater. Sci. Technol. 24, 896 (2008). doi:10.1179/174328408X302567

32. V. A. Chernenko, J. Pons, E. Cesari, K. Ishikawa, Stress-temperature phase diagram of a ferromagnetic Ni-Mn-Ga shape memory alloy. Acta Mater. 53, 5071 (2005). doi:10.1016/j.actamat.2005.07.018 33. V. V. Martynov, X-ray diffraction study of thermally and stress-induced phase transformations in single crystalline Ni-Mn-Ga alloys. J. Phys. 5, 91 (1995). doi:10.1051/jp4:1995810 34. K. Oikawa et al., Phase equilibria and phase transition on the Ni-Fe-Ga ferromagnetic shape memory alloy system. Metall. Mater. Trans., A Phys. Metall. Mater. Sci. 38, 767 (2007). doi:10.1007/s11661-007-9095-8 35. Y. Sutou et al., Stress-strain characteristics in Ni-Ga-Fe ferromagnetic shape memory alloys. Appl. Phys. Lett. 84, 1275 (2004). doi:10.1063/1.1642277 36. H. Y. Kim, Y. Ikehara, J. I. Kim, H. Hosoda, S. Miyazaki, Martensitic transformation, shape memory effect and superelasticity of Ti-Ni binary alloys. Acta Mater. 54, 2419 (2006). doi:10.1016/j.actamat.2006.01.019 37. J. Y. Guedou, M. Paliard, J. Rieu, Pseudo-elasticity in ordered Fe-Al alloys. Scr. Metall. 10, 631 (1976). doi:10.1016/0036-9748(76)90194-0 38. L. P. Kubin, A. Fourdeux, J. Y. Guedou, J. Rieu, Pseudoelasticity and slip reversibility in D03-ordered Fe-Al single-crystals by in situ experiments. Philos. Mag. A 46, 357 (1982). doi:10.1080/01418618208239564 39. H. Y. Yasuda, K. Nakano, T. Nakajima, M. Ueda, Y. Umakoshi, Effect of ordering process on giant pseudoelasticity in Fe3Al single crystals. Acta Mater. 51, 5101 (2003). doi:10.1016/S1359-6454(03)00360-4 40. H. Y. Yasuda, M. Aoki, A. Takaoka, Y. Umakoshi, Pseudoelasticity in Fe3Ga single crystals. Scr. Mater. 53, 253 (2005). doi:10.1016/j.scriptamat.2005.03.032 41. Y. Sutou et al., Effect of grain size and texture on pseudoelasticity in Cu-Al-Mnbased shape memory wire. Acta Mater. 53, 4121 (2005). doi:10.1016/j.actamat.2005.05.013