Crystallography and Microstructure of 7M Martensite in Ni-Mn-Ga Thin Films Epitaxially Grown on (1 1 2¯ 0)-Oriented Al2O3 Substrate

Epitaxial Ni-Mn-Ga thin films have been extensively investigated, due to their potential applications in magnetic micro-electro-mechanical systems. It has been proposed that the martensitic phase in the <1 1 0>A-oriented film is much more stable than that in the <1 0 0>A-oriented film. Nevertheless, the magnetic properties, microstructural features, and crystal structures of martensite in such films have not been fully revealed. In this work, the <1 1 0>A-oriented Ni51.0Mn27.5Ga21.5 films with different thicknesses were prepared by epitaxially growing on Al2O3(1 1 2¯ 0) substrate by magnetron sputtering. The characterization by X-ray diffraction technique and transmission electron microscopy revealed that all the Ni51.0Mn27.5Ga21.5 films are of 7M martensite at the ambient temperature, with their Type-I and Type-II twinning interfaces nearly parallel to the substrate surface.

In the present work, we successfully prepared the <1 1 0> A -oriented Ni 51.0 Mn 27.5 Ga 21.5 films with different thicknesses from 200 nm to 600 nm on Al 2 O 3 (1 1 2 0) substrates by magnetron sputtering. The characterization by X-ray diffraction technique, transmission electron microscopy revealed that all the Ni 51.0 Mn 27.5 Ga 21.5 films are of 7M martensite at the ambient temperature, and with their Type-I and Type-II twinning interfaces nearly parallel to the substrate surface.

Thin Films Preparation
Ni 51.0 Mn 27.5 Ga 21.5 thin films with different thicknesses from 200 nm to 600 nm were grown on Al 2 O 3 (1 1 2 0) mono-crystalline substrates by DC magnetron sputtering and using polycrystalline Ni 48 Mn 30 Ga 22 as the target materials. The target is of 50.8 mm in diameter and 1.5 mm in thickness. Before deposition, the base pressure of the sputtering equipment was vacuumed to below 9.0 × 10 −5 Pa. In order to obtain high-quality films, the substrate was heated to 650 • C. The sputtering process was conducted under a constant Ar working pressure of 0.15 Pa with an applied power of 70 W. Under this condition, the sputtering rate is of 0.1 nm/s roughly estimated through film thicknesses divided by deposition time. Before the actual depositions, a pre-sputtering was performed for 15 min.

Structure and Microstructure Characterization
The crystal structure and macroscopic crystallographic features were analyzed by temperature-dependent X-ray diffractometer using Cu-K α radiation (λ = 0.15406 nm) (Rigaku Smartlab 9 kW, Tokyo, Japan) and a four-circle X-ray diffractometer (Rigaku Smartlab 3 kW, Tokyo, Japan), respectively. A stylus profiler (Veeco DEKTAK 150, Plainview, NY, USA) was employed to measure the film thickness. The microstructures and chemical composition were examined by scanning electron microscopy (SEM, JEOL-JSM 7001F, Tokyo, Japan) and energy dispersive spectrometry (EDS, Bruker XFlash 4010, Berlin, Germany), respectively. The cross-sectional microstructures at nano and atomic scale were characterized by transmission electron microscopy (JEOL JEM 2100F, Tokyo, Japan) working at 200 kV. The cross-sectional sample for TEM characterization was prepared using the focused-ion beam (FIB, FEI Helios nanolab, Hillsboro, OR, USA) lift-out technique. Temperaturedependent magnetization curves and magnetic hysteresis loops were measured by Versalab (Quantum Design, San Diego, CA, USA).   Figure 1a, it is seen that only one specific diffraction peak of Ni 51.0 Mn 27.5 Ga 21.5 thin films is observed in the patterns, except the peaks from the Al 2 O 3 substrate, indicating that the films possess a strong preferred orientation. However, since the planar distance of (2 2 0) A plane of Austenite and (1 2 10) 7M plane of 7M martensite are roughly equal, we cannot distinguish which specific diffraction peak belongs to the 7M martensite or the Austenite. We performed the temperature-dependent X-ray diffraction measurement, as shown in Figure 1b-d. As can be seen from Figure 1b-d, for all the films, with the increase in temperature, the specific diffraction peak of (1 2 10) 7M gradually shifts to the low angle side. When the temperature is higher than 350 K, there is a sharp shift to the low angle side in the XRD patterns, which indicated that the (1 2 10) 7M of 7M martensite transformed to the (2 2 0) A of Austenite. It should be noted that the coexistence of Austenite and martensite phases in the 600 nm-thick Ni 51.0 Mn 27.5 Ga 21.5 film, which may be due to the restriction from the substrate and the temperature is not high or low enough for the full martensitic transformation. grown on Al2O3 (1 1 2 0) mono-crystalline substrates. As shown in Figure 1a, it is seen that only one specific diffraction peak of Ni51.0Mn27.5Ga21.5 thin films is observed in the patterns, except the peaks from the Al2O3 substrate, indicating that the films possess a strong preferred orientation. However, since the planar distance of (2 2 0)A plane of Austenite and (1 2 10 )7M plane of 7M martensite are roughly equal, we cannot distinguish which specific diffraction peak belongs to the 7M martensite or the Austenite. We performed the temperature-dependent X-ray diffraction measurement, as shown in Figure 1b-d. As can be seen from Figure 1b-d, for all the films, with the increase in temperature, the specific diffraction peak of (1 2 10 )7M gradually shifts to the low angle side. When the temperature is higher than 350 K, there is a sharp shift to the low angle side in the XRD patterns, which indicated that the (1 2 10 )7M of 7M martensite transformed to the (2 2 0)A of Austenite. It should be noted that the coexistence of Austenite and martensite phases in the 600 nmthick Ni51.0Mn27.5Ga21.5 film, which may be due to the restriction from the substrate and the temperature is not high or low enough for the full martensitic transformation. In order to determine the lattice parameters of the martensite, a four-circle X-ray diffractometer was employed to measure more XRD patterns at various azimuth angles (Phi) and tilt angles (Psi), as shown in Figure 2. With the employment of a four-circle XRD diffractometer, the crystal structure of the Ni51.0Mn27.5Ga21.5 thin films is identified as 7M modulated martensite and of monoclinic. The lattice parameters are shown in Table 1. In order to determine the lattice parameters of the martensite, a four-circle X-ray diffractometer was employed to measure more XRD patterns at various azimuth angles (Phi) and tilt angles (Psi), as shown in Figure 2. With the employment of a four-circle XRD diffractometer, the crystal structure of the Ni 51.0 Mn 27.5 Ga 21.5 thin films is identified as 7M modulated martensite and of monoclinic. The lattice parameters are shown in Table 1.  Figure 3a-c shows the microstructure of Ni51.0Mn27.5Ga21.5 thin films with different thicknesses taken by FE-SEM. From these FE-SEM images, no twining microstructure of the martensite variants can be observed in the top-view SEM images. All the films can be identified as continuously grown and the surface of the film becomes coarse with increasing thickness of the films. It is worth noting that white areas on the surface of the films can be identified as precipitates and the pit on the films might be defect pores among the column grains. To further analyze the microstructure and the martensitic configuration of the Ni51.0Mn27.5Ga21.5 films on the Al2O3 substrate, the cross-section TEM characterization was conducted. As displayed in Figure 4a-c, the martensite is visible in all films. The twin interfaces of martensite are parallel to the substrate surface. A selected area electron diffraction confirmed that the martensite is of 7M martensite (insets in Figure 4a). In addition, with the increase in film thickness, the number of both martensite variants and the twin interfaces increased, but the thickness of each martensite variant remains unchanged. Even though the microstructure of all the films can be identified as column grains, a detailed observation in Figure 4d shows that the twin can pass the grain boundaries, which indicated that the column grain boundaries are small-angle grain boundaries. From Figure 4d-f, martensite can be classified as the hierarchical structure in which martensite lath incorporates martensitic variants. The further crystallographic analysis demonstrated that each plate corresponds to one crystallographic variant, there are a maximum of four-oriented variants in each group and Type-I twin, Type-II twin, and Compound twin are present in the film, which is marked as the dotted line.     Figure 3a-c shows the microstructure of Ni51.0Mn27.5Ga21.5 thin films with different thicknesses taken by FE-SEM. From these FE-SEM images, no twining microstructure of the martensite variants can be observed in the top-view SEM images. All the films can be identified as continuously grown and the surface of the film becomes coarse with increasing thickness of the films. It is worth noting that white areas on the surface of the films can be identified as precipitates and the pit on the films might be defect pores among the column grains. To further analyze the microstructure and the martensitic configuration of the Ni51.0Mn27.5Ga21.5 films on the Al2O3 substrate, the cross-section TEM characterization was conducted. As displayed in Figure 4a-c, the martensite is visible in all films. The twin interfaces of martensite are parallel to the substrate surface. A selected area electron diffraction confirmed that the martensite is of 7M martensite (insets in Figure 4a). In addition, with the increase in film thickness, the number of both martensite variants and the twin interfaces increased, but the thickness of each martensite variant remains unchanged. Even though the microstructure of all the films can be identified as column grains, a detailed observation in Figure 4d shows that the twin can pass the grain boundaries, which indicated that the column grain boundaries are small-angle grain boundaries. From Figure 4d-f, martensite can be classified as the hierarchical structure in which martensite lath incorporates martensitic variants. The further crystallographic analysis demonstrated that each plate corresponds to one crystallographic variant, there are a maximum of four-oriented variants in each group and Type-I twin, Type-II twin, and Compound twin are present in the film, which is marked as the dotted line. To further analyze the microstructure and the martensitic configuration of the Ni 51.0 Mn 27.5 Ga 21.5 films on the Al 2 O 3 substrate, the cross-section TEM characterization was conducted. As displayed in Figure 4a-c, the martensite is visible in all films. The twin interfaces of martensite are parallel to the substrate surface. A selected area electron diffraction confirmed that the martensite is of 7M martensite (insets in Figure 4a). In addition, with the increase in film thickness, the number of both martensite variants and the twin interfaces increased, but the thickness of each martensite variant remains unchanged. Even though the microstructure of all the films can be identified as column grains, a detailed observation in Figure 4d shows that the twin can pass the grain boundaries, which indicated that the column grain boundaries are small-angle grain boundaries. From Figure 4d-f, martensite can be classified as the hierarchical structure in which martensite lath incorporates martensitic variants. The further crystallographic analysis demonstrated that each plate corresponds to one crystallographic variant, there are a maximum of fouroriented variants in each group and Type-I twin, Type-II twin, and Compound twin are present in the film, which is marked as the dotted line.

Macroscopic Texture
In order to determine the orientation relationship between the variants and the substrate, X-ray diffraction was used to measure the pole figure of the Ni51.0Mn27.5Ga21.5 thin film with a thickness of 600 nm and the Al2O3 (1 1 2 0) substrate. As shown in Figure 5a

Macroscopic Texture
In order to determine the orientation relationship between the variants and the substrate, X-ray diffraction was used to measure the pole figure of the Ni 51.0 Mn 27.5 Ga 21.5 thin film with a thickness of 600 nm and the Al 2 O 3 (1 1 2 0) substrate. As shown in Figure 5a

Macroscopic Texture
In order to determine the orientation relationship between the variants and the substrate, X-ray diffraction was used to measure the pole figure of the Ni51.0Mn27.5Ga21.5 thin film with a thickness of 600 nm and the Al2O3 (1 1 2 0) substrate. As shown in Figure 5a Figure 5d. Since the (1 2 10 )7M plane of 7M martensite is parallel to the (2 2 0)A plane of Austenite, we can conclude that the films were epitaxially grown on the Al2O3 (1 1 2 0) substrate, which justified that there is one crystallographic orientation for the Austenite in the present thin films. The crystallographic orientation of the Austenite can be described using Euler angle (60°, 90°, 45°). The mismatch relationship and the orientation of Austenite can also be de termined, as shown in Figure 6.

Magnetic Properties
Magnetization measurements are employed to identify the magnetic field-induced variant reorientation of the twin related martensitic variants in Ni-Mn-Ga thin films. Be fore the measurement of M-H loops, temperature-dependent magnetization cures were carried out to determine the martensitic transformation temperature as shown in Figure  7a. The M-T curves confirmed that the martensitic transformation temperature of the Ni51.0Mn27.5Ga21.5 thin films are above the ambient temperature, which suggested that these thin films are of martensite state at ambient temperature.

Magnetic Properties
Magnetization measurements are employed to identify the magnetic field-induced variant reorientation of the twin related martensitic variants in Ni-Mn-Ga thin films. Before the measurement of M-H loops, temperature-dependent magnetization cures were carried out to determine the martensitic transformation temperature as shown in Figure 7a. The M-T curves confirmed that the martensitic transformation temperature of the Ni 51.0 Mn 27.5 Ga 21.5 thin films are above the ambient temperature, which suggested that these thin films are of martensite state at ambient temperature.
The magnetization hysteresis (M-H) loop of Ni-Mn-Ga thin films was measured at different temperatures, as shown in Figure 7b-d. "Magnetization jumps" in magnetization hysteresis loops can be a sign of the magnetic field-induced reorientation of martensitic variants. However, no obvious "magnetization jumps" can be found in magnetization hysteresis loops of the present Ni 51.0 Mn 27.5 Ga 21.5 thin films with different thicknesses, from Figure 7b-d, which may be attributed to the high stress from the substrate. The magnetization hysteresis (M-H) loop of Ni-Mn-Ga thin films was measured at different temperatures, as shown in Figure 7b-d. "Magnetization jumps" in magnetization hysteresis loops can be a sign of the magnetic field-induced reorientation of martensitic variants. However, no obvious "magnetization jumps" can be found in magnetization hysteresis loops of the present Ni51.0Mn27.5Ga21.5 thin films with different thicknesses, from Figure 7b-d, which may be attributed to the high stress from the substrate.

Conclusions
In summary, we deposited <1 1 0>A-oriented Ni51.0Mn27.5Ga21.5 thin films from 200 nm to 600 nm which are of monoclinic 7M at room temperature on the Al2O3 (1 1 2 0). X-ray diffraction and transmission electron microscopy characterization revealed that the martensite-related twin interfaces are parallel to the substrate, and each group possesses a maximum of four-oriented variants. Only one crystallographic orientation for the Austenite is in the thin film. The characterization of magnetic properties shows that the magnetic field-induced variant reorientation was not obvious in the present thin films.

Conclusions
In summary, we deposited <1 1 0> A -oriented Ni 51.0 Mn 27.5 Ga 21.5 thin films from 200 nm to 600 nm which are of monoclinic 7M at room temperature on the Al 2 O 3 (1 1 2 0). X-ray diffraction and transmission electron microscopy characterization revealed that the martensite-related twin interfaces are parallel to the substrate, and each group possesses a maximum of four-oriented variants. Only one crystallographic orientation for the Austenite is in the thin film. The characterization of magnetic properties shows that the magnetic field-induced variant reorientation was not obvious in the present thin films.