- Email: [email protected]
Effect of pre-deformation on martensitic transformation behavior and the microstructure of a Ni–Mn–Ga alloy
Materials Science and Engineering A 438–440 (2006) 974–977
Effect of pre-deformation on martensitic transformation behavior and the microstructure of a Ni–Mn–Ga alloy Z.Y. Gao a,∗ , F. Chen a , W. Cai a , L.C. Zhao a , G.H. Wu b , B.G. Shen b , W.S. Zhan b a
b
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China Institute of Physics and Center for Condensed Matter Physics, Chinese Academy of Science, Beijing 100080, China Received 22 April 2005; received in revised form 4 January 2006; accepted 28 February 2006
Abstract The influence of pre-deformation on the martensitic transformation behavior and its microstructure has been put forward in a Ni55 Mn20 Ga25 (at.%) ferromagnetic shape memory alloy. It is found that the reverse martensitic transformation temperatures (As and Af ) increase linearly with increasing amount of deformation, while the martensitic transformation temperatures (Ms and Mf ) are not changed even when the deformation is up to 6%. However, the effect of the pre-deformation on the reverse martensitic transformation disappears in the following heating process. Transmission electron microscopy shows well-accommodated martensite variants in the undeformed specimen and interfacial boundaries are straight and well-defined. After deformation, the reorientation of martensite variants occurs at the expense of the unfavorable martensite variants, and the interfacial boundary changes from a sharp and straight morphology to a curved and irregular one. Meanwhile, some lattice defects are formed inside the variants even in slightly compressed samples. With increasing deformation, more lattice defects are generated inside the variants as well as in the interfacial boundaries between variants. Based on the analysis of the experimental results, the micromechanism of the effect of pre-deformation on martensitic transformation behavior has been clarified. © 2006 Elsevier B.V. All rights reserved. Keywords: Ni–Mn–Ga alloy; Pre-deformation; Martensitic transformation; Microstructure
1. Introduction In the last decades, the Heusler alloy Ni–Mn–Ga has received more and more attention as a potential smart actuator material because of its large reversible strain and rapid responding frequency, resulting from the twin boundary motion driven by external magnetic field [1,2]. The martensitic and reverse martensitic transformation start temperatures in Ni–Mn–Ga mainly depend on the chemical content and the hydrostatic pressure [3–5]. However, the effect of deformation on the onset of the phase transformation cannot be neglected. Recently, high transformation temperature Ni–Mn–Ga alloys have been found by Xu et al., etc., which promotes the study of Ni–Mn–Ga alloy [6,7]. It has been shown that the deformation may stabilize the martensite in Cu-based and Ni–Ti-based shape memory alloys [8,9]. How the deformation affects the transformation behavior and microstructure of Ni–Mn–Ga alloy including high transforma-
∗
Corresponding author. Tel.: +86 451 86418745; fax: +86 451 86418622. E-mail address: [email protected] (Z.Y. Gao).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.02.129
tion temperature one is still not clear. To clarify the change of transformation behavior and microstructure caused by predeformation, systematic investigations have been carried out in the present work. 2. Experimental A Ni55 Mn20 Ga25 (at.%) ingot was prepared from metal elements Ni, Mn and Ga with a purity of 99.95% by arc melting in an argon atmosphere. The obtained ingot was homogeneously annealed at 1125 K for 24 h in quartz capsules with about 10−4 mbar vacuum, then quenched into ice–water by crushing the capsules. Specimens with a dimension of 3 mm diameter × 5 mm length were cut from the ingot by electrical spark erosion for compression deformation. The deformation was carried out using an Instron-1186-type testing machine at room temperature, and the compression deformation varied from 1.8 to 6%. The transformation behaviors of the specimens after deformation were studied by differential scanning calorimetry (DSC) using a TA2920 differential scanning calorimeter with an argon
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977
975
atmosphere. The DSC measurement was conducted with a cooling/heating rate of 5 K/min. Samples for DSC measurement were cut from the center of each deformed specimens using a lowspeed diamond saw to avoid extra deformation. The microstructure of pre-deformed specimens was investigated by transmission electron microscopy (TEM). Foils for TEM examinations were mechanically polished to 120 m and two-jet electro-polished with an electrolyte consisting of 3% nitric acid and 97% methanol by volume. TEM studies were carried out in a JEOL-2000FXII microscopy operated at 200 kV. 3. Results and discussion Fig. 1 shows the effect of deformation on transformation temperatures Ni55 Mn20 Ga25 alloy. For the pre-deformed specimens, the reverse martensitic transformation temperatures (As and Af ) increased linearly with increasing amount of pre-deformation.
Fig. 1. Effect of deformation on transformation temperatures of Ni55 Mn20 Ga25 alloy.
Fig. 2. TEM image of the Ni55 Mn20 Ga25 alloy with different deformation strain: (a) the undeformed sample, (b) 1.8%, (c) 4% and (d) 6%.
976
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977
The reverse transformation start temperature As of the undeformed sample is 508 K. After deformation to 6%, As is raised to 554 K. The relation between deformation and change of As can be described to be: dT(As )/dε = 8 K, where ε represents amount of deformation. While the martensitic transformation temperatures (Ms and Mf ) did not show remarkable change, revealing that the pre-deformation only catalyzes the reverse martensitic transformation and has no obvious effect on martensitic transformation. Fig. 2a shows the TEM image of the thermal martensite in experimental alloy. It can be seen that the thermal martensite variants exhibit typical self-accommodation morphology with the straight and well-defined twin boundaries. When the sample is deformed to about 1.8%, the variant boundaries become curved and irregular, as shown in Fig. 2b. This suggests that the martensite variants in the favorable stress direction accommodate the deformation strain by consuming the neighboring variants in the unfavorable stress direction. Fig. 2c shows typical TEM images for the specimen subjected to 4% pre-deformation. It can be seen that the martensite variants are seriously crushed and crossed in the boundary and the variant-crashed or variantintersected morphologies have been observed. With the increment of the pre-deformation strain to 6%, some areas in the interface between the martensite variants become invisible because of the variants crash. Moreover, a large amount of lattice defects has been generated not only inside the martensite plates but also in the boundary areas. Some martensite plates become narrower and the dislocations are rearranged forming networks as a result of the martensite variants reorientation, as shown in Fig. 2d. Fig. 3 shows the dependence of the transformation temperatures on the number of thermal cycling of the specimens compressed to 4% at room temperature. As mentioned above, As and Af increase remarkably after deformation to 4%. However, As and Af decrease to the original values of the undeformed sample after the first thermal cycling, and keep constant with the further increase of the cycling number. In general, the required driving force for transformation towards austenite upon heating includes three terms: differ-
ence of chemical free energy, energy dissipation and elastic energy. The difference of chemical free energy is due to the phase difference. The energy dissipation term is dominated by the energy required for the movement of fronts of interphase (e.g. austenite/martensite) and interfaced (e.g. between martensite variants), the elastic energy plays a dual role here. The stored elastic energy during A → M transformation helps the reverse martensitic transformation upon heating. Based on the TEM observations, two effects of the pre-deformation on the transformation behavior can be clarified: one is the damage of the interfacial coherence; the other is the change of the elastic energy. While part of the elastic energy generated in detwinning may resist reverse transformation, as some kind of the locked-in microstructure or microquasi-plastic deformation, etc., may be generated which required higher driving force in reverse process, resulting in the increase of the reverse martensitic transformation temperatures. Upon heating to the temperature above Af temperature, the reverse martensitic transformation finished and the detwinned martensite was wholly transformed to the parent phase. Although the dislocations introduced by the compression deformation may be inherited by the parent phase, there is no remarkable influence of the dislocations on the transformation temperatures, which can be attributed to the large grain size of the Ni–Mn–Ga alloy. Then the thermal martensitic with accommodation morphology is formed in the following cooling process and the elastic energy favoring to the A → M transformation is stored in the boundary between martensite variants, resulting of the elimination of the effects of pre-deformation on the energy dissipation and the elastic energy. As a result, the reverse martensitic transformation temperatures show the same value as those of the undeformed specimens. 4. Conclusions (i) With increasing amount of deformation, the reverse martensitic transformation temperatures increase obviously while the martensitic transformation temperatures do not show any remarkable change. (ii) TEM observations of the microstructure of the undeformed sample and the deformed sample reveal the mechanism of the increasing of martensite stability by deformation. It is believed that the damage of interfacial coherence and the change of elastic energy introduced by the pre-compressive deformation resulting in the increase of martensite stability. Acknowledgement This work is financially supported by the National Natural Science Foundation of China (no. 50371022). References
Fig. 3. Effect of the thermal cycling on transformation temperatures of Ni–Mn–Ga alloys after compressed to 4% at room temperature.
[1] K. Ullakko, J.K. Huang, V.V. Kokorin, R.C. O’Handley, Scripta Mater. 10 (1997) 1133–1138. [2] A.A. Likhachev, K. Ullakko, Phys. Lett. A 275 (2000) 142–151.
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977 [3] M. Matsumoto, T. Takagi, J. Tani, T. Kanomata, N. Muramatsu, A.N. Vasil’ev, Mater. Sci. Eng. A 273–275 (1999) 326–328. [4] V.A. Chernenko, O. Babii, V. L’vov, P.G. Mccormick, Mater. Sci. Forum 327–328 (2000) 485–488. [5] H.C. Lin, S.K. Wu, T.S. Chou, Acta Metall. Mater. 39 (1991) 2069–2080.
977
[6] H. Xu, Y. Ma, C. Jiang, Appl. Phys. Lett. 82 (2003) 3206–3208. [7] C. Jiang, G. Feng, S. Gong, H. Xu, Mater. Sci. Eng. A 342 (2003) 231– 235. [8] M. Piao, K. Otsuka, S. Miyazaki, Mater. Trans. JIM 34 (1993) 919–922. [9] M. Piao, S. Miyazaki, K. Otsuka, Mater. Trans. JIM 33 (1992) 346–351.
Effect of pre-deformation on martensitic transformation behavior and the microstructure of a Ni–Mn–Ga alloy Z.Y. Gao a,∗ , F. Chen a , W. Cai a , L.C. Zhao a , G.H. Wu b , B.G. Shen b , W.S. Zhan b a
b
School of Materials Science and Engineering, Harbin Institute of Technology, Harbin 150001, China Institute of Physics and Center for Condensed Matter Physics, Chinese Academy of Science, Beijing 100080, China Received 22 April 2005; received in revised form 4 January 2006; accepted 28 February 2006
Abstract The influence of pre-deformation on the martensitic transformation behavior and its microstructure has been put forward in a Ni55 Mn20 Ga25 (at.%) ferromagnetic shape memory alloy. It is found that the reverse martensitic transformation temperatures (As and Af ) increase linearly with increasing amount of deformation, while the martensitic transformation temperatures (Ms and Mf ) are not changed even when the deformation is up to 6%. However, the effect of the pre-deformation on the reverse martensitic transformation disappears in the following heating process. Transmission electron microscopy shows well-accommodated martensite variants in the undeformed specimen and interfacial boundaries are straight and well-defined. After deformation, the reorientation of martensite variants occurs at the expense of the unfavorable martensite variants, and the interfacial boundary changes from a sharp and straight morphology to a curved and irregular one. Meanwhile, some lattice defects are formed inside the variants even in slightly compressed samples. With increasing deformation, more lattice defects are generated inside the variants as well as in the interfacial boundaries between variants. Based on the analysis of the experimental results, the micromechanism of the effect of pre-deformation on martensitic transformation behavior has been clarified. © 2006 Elsevier B.V. All rights reserved. Keywords: Ni–Mn–Ga alloy; Pre-deformation; Martensitic transformation; Microstructure
1. Introduction In the last decades, the Heusler alloy Ni–Mn–Ga has received more and more attention as a potential smart actuator material because of its large reversible strain and rapid responding frequency, resulting from the twin boundary motion driven by external magnetic field [1,2]. The martensitic and reverse martensitic transformation start temperatures in Ni–Mn–Ga mainly depend on the chemical content and the hydrostatic pressure [3–5]. However, the effect of deformation on the onset of the phase transformation cannot be neglected. Recently, high transformation temperature Ni–Mn–Ga alloys have been found by Xu et al., etc., which promotes the study of Ni–Mn–Ga alloy [6,7]. It has been shown that the deformation may stabilize the martensite in Cu-based and Ni–Ti-based shape memory alloys [8,9]. How the deformation affects the transformation behavior and microstructure of Ni–Mn–Ga alloy including high transforma-
∗
Corresponding author. Tel.: +86 451 86418745; fax: +86 451 86418622. E-mail address: [email protected] (Z.Y. Gao).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.02.129
tion temperature one is still not clear. To clarify the change of transformation behavior and microstructure caused by predeformation, systematic investigations have been carried out in the present work. 2. Experimental A Ni55 Mn20 Ga25 (at.%) ingot was prepared from metal elements Ni, Mn and Ga with a purity of 99.95% by arc melting in an argon atmosphere. The obtained ingot was homogeneously annealed at 1125 K for 24 h in quartz capsules with about 10−4 mbar vacuum, then quenched into ice–water by crushing the capsules. Specimens with a dimension of 3 mm diameter × 5 mm length were cut from the ingot by electrical spark erosion for compression deformation. The deformation was carried out using an Instron-1186-type testing machine at room temperature, and the compression deformation varied from 1.8 to 6%. The transformation behaviors of the specimens after deformation were studied by differential scanning calorimetry (DSC) using a TA2920 differential scanning calorimeter with an argon
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977
975
atmosphere. The DSC measurement was conducted with a cooling/heating rate of 5 K/min. Samples for DSC measurement were cut from the center of each deformed specimens using a lowspeed diamond saw to avoid extra deformation. The microstructure of pre-deformed specimens was investigated by transmission electron microscopy (TEM). Foils for TEM examinations were mechanically polished to 120 m and two-jet electro-polished with an electrolyte consisting of 3% nitric acid and 97% methanol by volume. TEM studies were carried out in a JEOL-2000FXII microscopy operated at 200 kV. 3. Results and discussion Fig. 1 shows the effect of deformation on transformation temperatures Ni55 Mn20 Ga25 alloy. For the pre-deformed specimens, the reverse martensitic transformation temperatures (As and Af ) increased linearly with increasing amount of pre-deformation.
Fig. 1. Effect of deformation on transformation temperatures of Ni55 Mn20 Ga25 alloy.
Fig. 2. TEM image of the Ni55 Mn20 Ga25 alloy with different deformation strain: (a) the undeformed sample, (b) 1.8%, (c) 4% and (d) 6%.
976
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977
The reverse transformation start temperature As of the undeformed sample is 508 K. After deformation to 6%, As is raised to 554 K. The relation between deformation and change of As can be described to be: dT(As )/dε = 8 K, where ε represents amount of deformation. While the martensitic transformation temperatures (Ms and Mf ) did not show remarkable change, revealing that the pre-deformation only catalyzes the reverse martensitic transformation and has no obvious effect on martensitic transformation. Fig. 2a shows the TEM image of the thermal martensite in experimental alloy. It can be seen that the thermal martensite variants exhibit typical self-accommodation morphology with the straight and well-defined twin boundaries. When the sample is deformed to about 1.8%, the variant boundaries become curved and irregular, as shown in Fig. 2b. This suggests that the martensite variants in the favorable stress direction accommodate the deformation strain by consuming the neighboring variants in the unfavorable stress direction. Fig. 2c shows typical TEM images for the specimen subjected to 4% pre-deformation. It can be seen that the martensite variants are seriously crushed and crossed in the boundary and the variant-crashed or variantintersected morphologies have been observed. With the increment of the pre-deformation strain to 6%, some areas in the interface between the martensite variants become invisible because of the variants crash. Moreover, a large amount of lattice defects has been generated not only inside the martensite plates but also in the boundary areas. Some martensite plates become narrower and the dislocations are rearranged forming networks as a result of the martensite variants reorientation, as shown in Fig. 2d. Fig. 3 shows the dependence of the transformation temperatures on the number of thermal cycling of the specimens compressed to 4% at room temperature. As mentioned above, As and Af increase remarkably after deformation to 4%. However, As and Af decrease to the original values of the undeformed sample after the first thermal cycling, and keep constant with the further increase of the cycling number. In general, the required driving force for transformation towards austenite upon heating includes three terms: differ-
ence of chemical free energy, energy dissipation and elastic energy. The difference of chemical free energy is due to the phase difference. The energy dissipation term is dominated by the energy required for the movement of fronts of interphase (e.g. austenite/martensite) and interfaced (e.g. between martensite variants), the elastic energy plays a dual role here. The stored elastic energy during A → M transformation helps the reverse martensitic transformation upon heating. Based on the TEM observations, two effects of the pre-deformation on the transformation behavior can be clarified: one is the damage of the interfacial coherence; the other is the change of the elastic energy. While part of the elastic energy generated in detwinning may resist reverse transformation, as some kind of the locked-in microstructure or microquasi-plastic deformation, etc., may be generated which required higher driving force in reverse process, resulting in the increase of the reverse martensitic transformation temperatures. Upon heating to the temperature above Af temperature, the reverse martensitic transformation finished and the detwinned martensite was wholly transformed to the parent phase. Although the dislocations introduced by the compression deformation may be inherited by the parent phase, there is no remarkable influence of the dislocations on the transformation temperatures, which can be attributed to the large grain size of the Ni–Mn–Ga alloy. Then the thermal martensitic with accommodation morphology is formed in the following cooling process and the elastic energy favoring to the A → M transformation is stored in the boundary between martensite variants, resulting of the elimination of the effects of pre-deformation on the energy dissipation and the elastic energy. As a result, the reverse martensitic transformation temperatures show the same value as those of the undeformed specimens. 4. Conclusions (i) With increasing amount of deformation, the reverse martensitic transformation temperatures increase obviously while the martensitic transformation temperatures do not show any remarkable change. (ii) TEM observations of the microstructure of the undeformed sample and the deformed sample reveal the mechanism of the increasing of martensite stability by deformation. It is believed that the damage of interfacial coherence and the change of elastic energy introduced by the pre-compressive deformation resulting in the increase of martensite stability. Acknowledgement This work is financially supported by the National Natural Science Foundation of China (no. 50371022). References
Fig. 3. Effect of the thermal cycling on transformation temperatures of Ni–Mn–Ga alloys after compressed to 4% at room temperature.
[1] K. Ullakko, J.K. Huang, V.V. Kokorin, R.C. O’Handley, Scripta Mater. 10 (1997) 1133–1138. [2] A.A. Likhachev, K. Ullakko, Phys. Lett. A 275 (2000) 142–151.
Z.Y. Gao et al. / Materials Science and Engineering A 438–440 (2006) 974–977 [3] M. Matsumoto, T. Takagi, J. Tani, T. Kanomata, N. Muramatsu, A.N. Vasil’ev, Mater. Sci. Eng. A 273–275 (1999) 326–328. [4] V.A. Chernenko, O. Babii, V. L’vov, P.G. Mccormick, Mater. Sci. Forum 327–328 (2000) 485–488. [5] H.C. Lin, S.K. Wu, T.S. Chou, Acta Metall. Mater. 39 (1991) 2069–2080.
977
[6] H. Xu, Y. Ma, C. Jiang, Appl. Phys. Lett. 82 (2003) 3206–3208. [7] C. Jiang, G. Feng, S. Gong, H. Xu, Mater. Sci. Eng. A 342 (2003) 231– 235. [8] M. Piao, K. Otsuka, S. Miyazaki, Mater. Trans. JIM 34 (1993) 919–922. [9] M. Piao, S. Miyazaki, K. Otsuka, Mater. Trans. JIM 33 (1992) 346–351.