Deformation Behavior and Microstructure Evolution of NiTiCu Shape Memory Alloy Subjected to Plastic Deformation at High Temperatures

Deformation behavior and microstructure evolution of NiTiCu shape memory alloy (SMA), which possesses martensite phase at room temperature, were investigated based on a uniaxial compression test at the temperatures of 700~1000 ◦C and at the strain rates of 0.0005~0.5 s−1. The constitutive equation of NiTiCu SMA was established in order to describe the flow characteristic of NiTiCu SMA, which is dominated by dynamic recovery and dynamic recrystallization. Dislocations become the dominant substructures of martensite phase in NiTiCu SMA compressed at 700 ◦C. Martensite twins are dominant in NiTiCu SMA compressed at 800 and 900 ◦C. Martensite twins are not observed in NiTiCu SMA compressed at 1000 ◦C. The microstructures resulting from dynamic recovery or dynamic recrystallization significantly influences the substructures in the martensite phase of NiTiCu SMA at room temperature. Dislocation substructures formed during dynamic recovery, such as dislocation cells and subgrain boundaries, can suppress the formation of twins in the martensite laths of NiTiCu SMA. The size of dynamic recrystallized grains affects the formation of martensite twins. Martensite twins are not easily formed in the larger recrystallized grain, since the constraint of the grain boundaries plays a weak role. However, in the smaller recrystallized grain, martensite twins are induced to accommodate the transformation from austenite to martensite.


Introduction
NiTi shape memory alloys (SMAs) have deserved increasing attention in the engineering field because they have shape memory effect [1].With a view to widening the application of NiTi SMA in the domain of engineering, third elements are added to the binary NiTi SMA in order to change the transformation temperature or hysteresis [2][3][4][5][6].For example, Cu element is added to binary NiTi SMA so as to significantly lower the phase transformation hysteresis, which lays the foundation for the application of an actuator or micro-electro-mechanical system (MEMS) [7,8].As a consequence, over the last few decades, many researchers have paid more attention to NiTiCu SMAs [9][10][11][12].It is well known that plastic deformation substantially influences the microstructures and transformation behavior of NiTi-based SMAs [13,14].In particular, thermomechanical processing, especially plastic working, is an important step in manufacturing NiTi-based SMA products [15].Consequently, it is very important to understand deformation behavior as well as microstructural evolution of NiTi-based SMAs at high temperatures [16][17][18][19][20]. Up to date, many scholars have devoted themselves to studying the flow behavior of NiTi-based SMAs at high temperatures based on the Arrhenius-type constitutive equation [21][22][23][24][25], which lays the foundation for clarifying deformation mechanisms of NiTi-based SMAs.In the present study, the deformation behavior and microstructure evolution of Ni 45 Ti 50 Cu 5 (at %) SMA were investigated based on a uniaxial compression test, where the temperatures range from 700 to 1000 • C and the strain rates range from 0.0005 to 0.5 s −1 .In particular, the Ni 45 Ti 50 Cu 5 SMA of interest possesses martensite phase at room temperature.Therefore, it is of great significance to investigate the deformation behavior and microstructure evolution of Ni 45 Ti 50 Cu 5 SMA.

Materials and Methods
As-rolled Ni 45 Ti 50 Cu 5 (at %) SMA bar with a diameter of 30 mm was commercially received from Xi'an Saite Metal Materials Development Co., Ltd.(Xi'an, China).The phase transformation of the as-rolled NiTiCu SMA was measured using Pyris Diamond type differential scanning calorimetry (DSC, Perkin Elmer Inc., Waltham, MA, USA).The DSC test was carried out in the range of −150~150 • C, where the heating and cooling steps were 10 • C/min.The DSC curve of the as-rolled NiTiCu SMA can be found in Reference [26].The phase transformation temperatures of the as-rolled NiTiCu SMA were as follows: NiTiCu SMA samples, which possess a height of 9 mm and a diameter of 6 mm, were electro-discharge machined from the as-rolled NiTiCu SMA bar.Subsequently, they were used for the compression test.An INSTRON-5500R universal material testing machine (Instron Corporation, Norwood, MA, USA) was used for implementing the compression tests.The NiTiCu SMA samples were compressed by the deformation extent of 60%, where the temperatures range from 700 to 1000 • C and the strain rates range from 0.0005 to 0.5 s −1 .Subsequently, all of the compressed NiTiCu SMA specimens were put into ice water for the purpose of guaranteeing complete martensite phase transformation.
As for the as-rolled and compressed NiTiCu SMA samples, the microstructures were captured by transmission electron microscopy (TEM).The NiTiCu SMA samples used for TEM observation were made into foils with the thickness of 70 µm by means of mechanical grinding.Subsequently, the foils were thinned by twin-jet polishing in an electrolyte which is composed of 6% HClO 4 , 34% CH 3 (CH 2 ) 3 OH, and 60% CH 3 OH by volume fraction.Finally, the NiTiCu SMA samples for TEM observation were characterized by virtue of an FEI TECNAI G2 F30 microscope (FEI Corporation, Hillsboro, OR, USA).TEM observation results indicate that the substructure of martensite phase contains martensite laths and martensite twins, as shown in Figure 1.
In the present study, the deformation behavior and microstructure evolution of Ni45Ti50Cu5 (at %) SMA were investigated based on a uniaxial compression test, where the temperatures range from 700 to 1000 °C and the strain rates range from 0.0005 to 0.5 s −1 .In particular, the Ni45Ti50Cu5 SMA of interest possesses martensite phase at room temperature.Therefore, it is of great significance to investigate the deformation behavior and microstructure evolution of Ni45Ti50Cu5 SMA.

Materials and Methods
As-rolled Ni45Ti50Cu5 (at %) SMA bar with a diameter of 30 mm was commercially received from Xi'an Saite Metal Materials Development Co., Ltd.The phase transformation of the as-rolled NiTiCu SMA was measured using Pyris Diamond type differential scanning calorimetry (DSC, Perkin Elmer Inc., Waltham, MA, USA).The DSC test was carried out in the range of −150~150 °C, where the heating and cooling steps were 10 °C/min.The DSC curve of the as-rolled NiTiCu SMA can be found in Reference [26].The phase transformation temperatures of the as-rolled NiTiCu SMA were as follows: NiTiCu SMA samples, which possess a height of 9 mm and a diameter of 6 mm, were electro-discharge machined from the as-rolled NiTiCu SMA bar.Subsequently, they were used for the compression test.An INSTRON-5500R universal material testing machine (Instron Corporation, Norwood, MA, USA) was used for implementing the compression tests.The NiTiCu SMA samples were compressed by the deformation extent of 60%, where the temperatures range from 700 to 1000 °C and the strain rates range from 0.0005 to 0.5 s −1 .Subsequently, all of the compressed NiTiCu SMA specimens were put into ice water for the purpose of guaranteeing complete martensite phase transformation.
As for the as-rolled and compressed NiTiCu SMA samples, the microstructures were captured by transmission electron microscopy (TEM).The NiTiCu SMA samples used for TEM observation were made into foils with the thickness of 70 μm by means of mechanical grinding.Subsequently, the foils were thinned by twin-jet polishing in an electrolyte which is composed of 6% HClO4, 34% CH3(CH2)3OH, and 60% CH3OH by volume fraction.Finally, the NiTiCu SMA samples for TEM observation were characterized by virtue of an FEI TECNAI G2 F30 microscope (FEI Corporation, Hillsboro, OR, USA).TEM observation results indicate that the substructure of martensite phase contains martensite laths and martensite twins, as shown in Figure 1.

Deformation Behavior of NiTiCu SMA
Figure 2 indicates the true stress-strain curves of NiTiCu SMA undergoing uniaxial compression at the temperatures ranging from 700 to 1000 • C and the strain rates ranging from 0.0005 to 0.5 s −1 .As can be seen in Figure 3, the flow stress of NiTiCu SMA increases with increasing strain rate.Furthermore, elevating the temperature contributes to lowering the flow stress of NiTiCu SMA.It can be noted from the stress-strain curves of NiTiCu SMA that dynamic recovery (DRV) or dynamic recrystallization (DRX) can occur during compression deformation.

Deformation Behavior of NiTiCu SMA
Figure 2 indicates the true stress-strain curves of NiTiCu SMA undergoing uniaxial compression at the temperatures ranging from 700 to 1000 °C and the strain rates ranging from 0.0005 to 0.5 s −1 .As can be seen in Figure 3, the flow stress of NiTiCu SMA increases with increasing strain rate.Furthermore, elevating the temperature contributes to lowering the flow stress of NiTiCu SMA.It can be noted from the stress-strain curves of NiTiCu SMA that dynamic recovery (DRV) or dynamic recrystallization (DRX) can occur during compression deformation.

Deformation Behavior of NiTiCu SMA
Figure 2 indicates the true stress-strain curves of NiTiCu SMA undergoing uniaxial compression at the temperatures ranging from 700 to 1000 °C and the strain rates ranging from 0.0005 to 0.5 s −1 .As can be seen in Figure 3, the flow stress of NiTiCu SMA increases with increasing strain rate.Furthermore, elevating the temperature contributes to lowering the flow stress of NiTiCu SMA.It can be noted from the stress-strain curves of NiTiCu SMA that dynamic recovery (DRV) or dynamic recrystallization (DRX) can occur during compression deformation.Based on the stress-strain curves of NiTiCu SMA at high temperatures, it can be found that the plastic flow of NiTiCu SMA at high temperatures is dependent on the strain rates as well as the deformation temperatures.As a consequence, the constitutive equation of NiTiCu SMA at high temperatures is established according to the Arrhenius type equation [22][23][24][25], namely: .
ε is the strain rate, σ the flow stress, T the absolute temperature, Q the activation energy, R the universal gas constant (8.314J•mol −1 •K −1 ), and A, α and n the material constants.
At a low stress level, Equation ( 1) is approximately expressed by: .
where A 1 is still a material constant and At a high stress level, Equation ( 1) is approximately simplified as: .
where A 2 and β remain the material constants and It is generally accepted that the Zener-Hollomon parameter Z, which is viewed as the function of strain rate and temperature, can be expressed as follows: Substituting the parameter Z into Equation (1) leads to: The following equation can be obtained by transforming Equation (5): As a consequence, Equation ( 6) is the constitutive equation which is expressed by the parameter Z.
The constitutive equation of NiTiCu SMA can be obtained after the values of A, α, n, and Q can be determined on the basis of the experimental data.In the present work, the peak stress is used to determine the corresponding parameters, as shown in Table 1.
According to Equation (7), the value of n can be obtained by means of the linear fitting method on the basis of experimental data, as shown in Figure 3a.The value of n is determined as 5.7988.
Prior to obtaining the value of α, the value of β needs to be determined by employing the natural logarithm of Equation ( 3), and consequently the following equation is obtained: In the same manner, the value of n can be obtained by virtue of the linear fitting method on the basis of experimental data, as shown in Figure 3b.As a consequence, the value of β is calculated as 0.07059.Finally, according to the values of n and β, the value of α can be determined as In order to further modify the value of n, the following equation can be obtained by employing the natural logarithm of Equation (1): Similarly, the value of n can be modified as 4.1318 according to the linear fitting method on the basis of experimental data, as shown in Figure 3c.
For the purpose of obtaining the value of Q, in the case of the given strain rates, the following equation can be obtained by differentiating T −1 in Equation ( 9): The value of Q can be determined as 198.84 × 103 J•mol −1 by combining the modified value of n and the value of R with the fitting value derived from Figure 3d.
For the purpose of obtaining the value of A, the following equation can be obtained by employing the natural logarithm of Equation (5): According to Equation ( 11), the value of ln A can be determined as 16.3994 based on the experimental data, as shown in Figure 4, and consequently the value of A is further calculated to be 1.3249 × 10 7 .
According to Equation (7), the value of n can be obtained by means of the linear fitting method on the basis of experimental data, as shown in Figure 3a.The value of n is determined as 5.7988.
Prior to obtaining the value of α , the value of β needs to be determined by employing the natural logarithm of Equation ( 3), and consequently the following equation is obtained: In the same manner, the value of n can be obtained by virtue of the linear fitting method on the basis of experimental data, as shown in Figure 3b.As a consequence, the value of β is calculated as 0.07059.Finally, according to the values of n and β , the value of α can be determined as In order to further modify the value of n , the following equation can be obtained by employing the natural logarithm of Equation ( 1): Similarly, the value of n can be modified as 4.1318 according to the linear fitting method on the basis of experimental data, as shown in Figure 3c.
For the purpose of obtaining the value of Q , in the case of the given strain rates, the following equation can be obtained by differentiating 1 − T in Equation ( 9): The value of Q can be determined as 198.84 × 103 J•mol −1 by combining the modified value of n and the value of R with the fitting value derived from Figure 3d.
For the purpose of obtaining the value of A , the following equation can be obtained by employing the natural logarithm of Equation (5): )] ln[sinh( ln ln ασ n A Z + = (11) According to Equation ( 11), the value of A ln can be determined as 16.3994 based on the experimental data, as shown in Figure 4, and consequently the value of A is further calculated to be 1.3249 × 10 7 .

Microstructural Evolution of NiTiCu SMA
The microstructural evolution of several representative NiTiCu SMA samples, which undergo compression deformation, is captured in order to reveal the corresponding plastic deformation mechanisms at high temperatures.It can be generally accepted that DRV or DRX can take place when metallic alloy experiences plastic deformation at elevated temperatures.In the case of DRX, in particular, the size of the recrystallized grains increases with increasing the deformation temperature or decreasing the strain rate.According to the phase transformation temperatures of NiTiCu SMA, it can be noted that when NiTiCu SMA samples are subjected to compression deformation at high temperatures and are quenched into the ice water, they are transformed from B2 austenite into B19 martensite because of complete martensite phase transformation.
For the purpose of better revealing the martensite structure of NiTiCu SMA, TEM micrographs of the NiTiCu SMA samples, which are subjected to compression at the temperatures of 700~1000 • C at the strain rate of 0.005 s −1 , were captured, as shown in Figures 5-8.
It can be found that dislocation substructures, such as dislocation cells and subgrain boundaries, appear in the martensite phase of NiTiCu SMA subjected to compression deformation at 700 • C, as shown in Figure 5a.It can thus be concluded that dislocations become the dominant substructures of martensite phase in the NiTiCu SMA specimen undergoing compression at 700 • C. Martensite twins are dominant in the NiTiCu SMA specimen undergoing compression at 800 and 900 • C, as shown in Figures 6 and 7.However, martensite twins are not observed in the NiTiCu SMA specimen undergoing compression at 1000 • C, as shown in Figure 8.According to the aforementioned analysis, it is noted that DRV or DRX can occur in NiTiCu SMA samples subjected to compression deformation at high temperatures, which depends on the deformation temperatures.It can be deduced that DRV or DRX microstructures have an influence on the substructures of martensite phase of NiTiCu SMA.In general, DRV is the dominant mechanism when NiTi-based SMAs are subjected to compression deformation at 700 • C [21].As a consequence, DRV leads to the deformation of grains and thus dislocation substructures, such as dislocation cells and subgrain boundaries, are formed in the grain interior.These dislocation substructures are kept in the martensite phase of the NiTiCu SMA sample during subsequent martensite phase transformations.These retained dislocation substructures suppress the formation of martensite twins.It can be generally accepted that complete DRX can take place in NiTi-based SMAs suffering from compression at temperatures above 800 • C [18,21].It is proposed that the size of the dynamic recrystallized grains influences the formation of martensite twins in NiTiCu SMAs.In general, the size of the dynamic recrystallized grains increases with increasing the deformation temperature [18].Furthermore, it is generally accepted that the occurrence of martensite twins aims to accommodate the formation of martensite phase during the transformation from B2 austenite to B19 martensite.It can be inferred that when the dynamic recrystallized grains possess a smaller size, the grain boundaries are able to suppress the formation of martensite phase.Consequently, the occurrence of martensite twins contributes to the formation of martensite phase.When the dynamic recrystallized grains possess a larger size, there is sufficient space to guarantee the formation of martensite phase in the grain interior.Therefore, martensite twins are not easily formed in NiTiCu SMA samples subjected to compression deformation at 1000 • C.
temperature or decreasing the strain rate.According to the phase transformation temperatures of NiTiCu SMA, it can be noted that when NiTiCu SMA samples are subjected to compression deformation at high temperatures and are quenched into the ice water, they are transformed from B2 austenite into B19′ martensite because of complete martensite phase transformation.
For the purpose of better revealing the martensite structure of NiTiCu SMA, TEM micrographs of the NiTiCu SMA samples, which are subjected to compression at the temperatures of 700~1000 °C at the strain rate of 0.005 s −1 , were captured, as shown in Figures 5-8.It can be found that dislocation substructures, such as dislocation cells and subgrain boundaries, appear in the martensite phase of NiTiCu SMA subjected to compression deformation at 700 °C, as shown in Figure 5a.It can thus be concluded that dislocations become the dominant substructures of martensite phase in the NiTiCu SMA specimen undergoing compression at 700 °C.Martensite twins are dominant in the NiTiCu SMA specimen undergoing compression at 800 and 900 °C, as shown in Figures 6 and 7.However, martensite twins are not observed in the NiTiCu SMA specimen undergoing compression at 1000 °C, as shown in Figure 8.According to the aforementioned analysis, it is noted that DRV or DRX can occur in NiTiCu SMA samples subjected to compression deformation at high temperatures, which depends on the deformation temperatures.It can be deduced that DRV or DRX microstructures have an influence on the substructures of martensite phase of NiTiCu SMA.In general, DRV is the dominant mechanism when NiTi-based SMAs are subjected to compression deformation at 700 °C [21].As a consequence, DRV leads to the deformation of grains and thus dislocation substructures, such as dislocation cells and subgrain boundaries, are formed in the grain interior.These dislocation substructures are kept in the martensite phase of the NiTiCu SMA sample during subsequent martensite phase transformations.These retained dislocation substructures suppress the formation of martensite twins.It can be generally accepted that complete DRX can take place in NiTi-based SMAs suffering from compression at temperatures above 800 °C [18,21].It is proposed that the size of the dynamic recrystallized grains influences the formation of martensite twins in NiTiCu SMAs.In general, the size of the dynamic recrystallized grains increases with increasing the deformation temperature [18].

Conclusions
The deformation behavior and microstructure evolution of NiTiCu SMA, which possesses martensite phase at room temperature, were investigated based on a uniaxial compression test at the temperatures of 700~1000 • C and at the strain rates of 0.0005~0.5 s −1 .The following conclusions can be drawn: The size of dynamic recrystallized grains has an effect on the formation of martensite twins.Martensite twins are not easily formed in larger dynamic recrystallized grain, since the constraint of the grain boundaries plays a weak role.However, in smaller dynamic recrystallized grain, martensite twins are induced in order to accommodate the occurrence of the transformation from austenite phase to martensite phase.

Figure 3 .
Figure 3. Determination of material constants by means of fitting method based on experimental data: (a) Solving the value of n ; (b) Solving the value of β ; (c) Modifying the value of n ; (d) Solving the value of Q .

Figure 3 .
Figure 3. Determination of material constants by means of fitting method based on experimental data: (a) Solving the value of n ; (b) Solving the value of β ; (c) Modifying the value of n ; (d) Solving the value of Q .

Figure 3 .
Figure 3. Determination of material constants by means of fitting method based on experimental data: (a) Solving the value of n; (b) Solving the value of β; (c) Modifying the value of n; (d) Solving the value of Q.

Figure 4 .
Figure 4. Determination of material constant A based on the relationship between lnZ and

Figure 4 .
Figure 4. Determination of material constant A based on the relationship between ln Z and ln[sinh(ασ)].

Figure 5 .
Figure 5. TEM micrographs of NiTiCu SMA subjected to compression deformation at 700 °C: (a) Bright field image showing the existence of dislocation substructures in martensite laths; (b) Diffraction pattern of (a).

Figure 5 .
Figure 5. TEM micrographs of NiTiCu SMA subjected to compression deformation at 700 • C: (a) Bright field image showing the existence of dislocation substructures in martensite laths; (b) Diffraction pattern of (a).

Figure 6 .
Figure 6.TEM micrographs of NiTiCu SMA subjected to compression deformation at 800 °C: (a) Bright field image showing the existence of martensite twins; (b) Diffraction pattern of (a).

Figure 7 .
Figure 7. TEM micrographs of NiTiCu SMA subjected to compression deformation at 900 °C: (a) Bright field image showing the existence of martensite twins; (b) Diffraction pattern of (a).

Figure 6 . 10 Figure 6 .
Figure 6.TEM micrographs of NiTiCu SMA subjected to compression deformation at 800 • C: (a) Bright field image showing the existence of martensite twins; (b) Diffraction pattern of (a).

Figure 7 .
Figure 7. TEM micrographs of NiTiCu SMA subjected to compression deformation at 900 °C: (a) Bright field image showing the existence of martensite twins; (b) Diffraction pattern of (a).

Figure 7 .
Figure 7. TEM micrographs of NiTiCu SMA subjected to compression deformation at 900 • C: (a) Bright field image showing the existence of martensite twins; (b) Diffraction pattern of (a).

Figure 8 .
Figure 8. TEM micrographs of NiTiCu SMA subjected to compression deformation at 1000 °C: (a) Bright field image showing no martensite laths; (b) Bright field image showing the existence of martensite laths; (c) Diffraction pattern of (b).

Figure 8 .
Figure 8. TEM micrographs of NiTiCu SMA subjected to compression deformation at 1000 • C: (a) Bright field image showing no martensite laths; (b) Bright field image showing the existence of martensite laths; (c) Diffraction pattern of (b).

( 1 )
The constitutive equation of NiTiCu SMA based on the Zener-Hollomon parameter was established in order to describe the flow characteristic of NiTiCu SMA.The results show that the flow stress of NiTiCu SMA depends on the strain rates.Depending on temperatures, DRV or DRX are the main mechanisms for the plastic deformation of NiTiCu SMA at elevated temperatures.(2) The microstructures resulting from DRV or DRX have a significant influence on the substructures in the martensite phase of the NiTiCu SMA sample at room temperature.Dislocations become the dominant substructures of martensite in the NiTiCu SMA specimen undergoing compression at 700 • C. Martensite twins are dominant in the NiTiCu SMA specimen undergoing compression at 800 and 900 • C. Martensite twins are not observed in the NiTiCu SMA specimen undergoing compression at 1000 • C. (3) Dislocation substructures resulting from DRV, such as dislocation cells and subgrain boundaries, are able to suppress the formation of martensite twins in the martensite laths of NiTiCu SMA.

Table 1 .
Peak stress of NiTiCu SMA at the various strain rates and deformation temperatures (MPa).