43
44
spectrums of 100-150 (Mg17.7Cu82.3), 100-100 (Mg23.5Cu76.5), 100-50 (Mg40.4Cu59.6), 100-25 (Mg61.9Cu38.1) and 100-15 (Mg39.9Cu43.5O16.6), the locations of hump center are related to the composition. The centers of the humps shift to the right with increasing Cu content. In addition, 100-150 and 100-25 seem to contain some minor nano-particles.
Figure 4-2 presents the XRD spectrums of the 50 series. In comparing with these specimens, the peaks at 43.4o and 50.6o are referred to the Cu (111) plane and (200) plane, respectively, meaning 50-150 and 50-100 consist of almost pure Cu and no Mg. Next, comparing 50-150 and 50-100, 50-100 exhibits the same Cu crystalline peaks of Cu (111) and (200) as 50-150, but the peaks are more broadened implying that the 50-100 specimen consists of smaller Cu grains. This would be resulted from that a lower Cu power leads to a lower kinetic energy of Cu incident atoms, implying the Mg and Cu could not mix well.
As Cu power decreases, the Cu contents of 50-50 and 50-25 decrease with decreasing Cu power, implying the re-sputtering effect is unapparent. Figure 4-2 shows that the 50-50 and 50-25 specimen exhibit broadened weak Mg peaks consisting of a mixture of (0002) and (1011) planes at 33o to 39o, Cu (111) and (200) planes at 43.4o and 50.6o meaning that the mixing effect is also more and more unapparent during deposition. Due to the occurrence of the specific Mg and Cu structures, it is reasonable to suggest that the two specimens consist of separated nano-grains of Mg and Cu, not like the 100 series. Besides, the nano-domains of Mg and Cu are much more reactive to oxygen in air, meaning the oxide would form on the surface of the films.
4-2-2 XRD analysis of Mg-Cu multilayered thin films
The structures of the as-deposited multilayered thin films are identified by X-ray
45
diffraction, as shown in Figures 4-3 and 4-4. According to the XRD patterns of 20T21 and 20T14, there exhibits a very strong texture of the (0002) plane of Mg at 34.5o, compared with the XRD pattern of the pure Mg film, as shown in Figure 4-5. Meanwhile, the as-deposited Cu film exhibits the similar occurrence with the strong plane textures of the (111) plane at 43.3o and minor (002) at 50.5o, compared with the pure Cu film, as shown in Figure 4-5. In these results, it can be inferred that individual Mg and Cu layers grow along the close-packed crystalline planes, irrespective of the structure of the Si substrate.
In addition, the XRD patterns of 40N32 and 40N14 which are the same texture as those of 20T32 and 20T14, as shown in Figures 4-3 and 4-4, exhibit broadening crystalline peaks of Mg and Cu is due to the thinner thicknesses of individual layers.
4-3 TEM observation of co-sputtered Mg-Cu thin films
4-3-1 Microstructure of co-sputtered 100-150 (Mg
17.7Cu
82.3) and 100-100 (Mg
23.5Cu
76.5) thin films
In order to understand the influence of the microstructure on the mechanical properties, the microstructures of 100-150 and 100-100 are examined using TEM. Figure 4-6, the plane-view bright-field image (BFI) of 100-150, exhibits many particles (~10 nm) dispersed in the matrix. The selected area diffraction patterns (SADP) of 100-150, as shown in Figure 4-7, indicated that the 100-150 film consists of the mixture of the MgCu2 nanocrystalline and Mg-Cu amorphous structure.
Similarly, in another case of 100-100, Figure 4-8 shows that the 100-100 specimen consists of many fine MgCu2 particles, which particle size is about 10 nm. The dark-field
46
image (DFI) of 100-100, as shown in Figure 4-9, exhibits the amorphous-nanocrystalline structure [65]. Moreover, Figure 4-10 shows the microstructure around the MgCu2 particles (about 20 nm) with several nanometer diameter. Supposedly, the structure consist of Mg-Cu amorphous and smaller MgCu2 nanoparticles, as shown in Figure 4-11. It will be discussed in the next section.
4-3-2 High-resolution TEM observation of co-sputtered 100-100 thin films
According to the TEM BFI and DFI, the contrast of the DFI can not provide the sufficient information for the MgCu2, Cu, and Mg-Cu amorphous phases. Hence, it is necessary to identify the phases in the 100-100 specimen using the high-resolution TEM (HRTEM) technique. In addition to the 10~20 nm MgCu2 particles, as shown in Figure 4-12, the specimen seems to exhibit a different structure in the localized area, such as the marked region by the circle. The high-resolution TEM image exhibits many MgCu2 nanocrystals and amorphous particles, as shown in Figures 4-12, 4-13 and 4-14.
Around the MgCu2 particles, there exhibits an irregular amorphous particle about 15 nm surrounded by the {110} MgCu2 crystalline phases as shown in Figure 4-12. In a localized area, Figures 4-13 and 4-14 show different structures: one consists of about 5-nm amorphous particle in the MgCu2 matrix; another exhibits 2-nm amorphous particles in the MgCu2 {110}
matrix.
4-4 Thermal analysis
4-4-1 DSC analysis of Mg-Cu co-sputtered thin films
47
For the 100 and 50 series, the thermal properties must be compared under the same condition, such as the same substrate and background pressure, etc. However, only the specimens of 100-150 and 100-100 can be peeled off from the Si substrates. Hence, the thermal properties of other specimens can not be measured, and the thermal properties are evaluated based on the composition difference. The modified heating DSC curve of the 100-150 (Mg17.7Cu82.3) specimen is shown in Figure 4-15 with a heating rate of 5 K/min. The glass-transition temperature (Tg) is located at 428 K, and the onset of crystallization temperature is located at 460 K.
4-4-2 Structural transformation of the 100 series thin films at 423 K
According to the heating DSC curve of 100-150, the partial amorphous Mg-Cu thin film, the glass transition temperature is located at 428 K. Due to the partial amorphous structure, existed nuclei would make the Mg-Cu crystallization occurs at a lower temperature. Hence, the temperature of the isothermal heat-treatment for 100-150 (Mg17.7Cu82.3), 100-100 (Mg23.5Cu76.5), 100-50 (Mg40.4Cu59.6) and 100-25 (Mg61.9Cu38.1) is evaluated at 423 K, about 40 K lower than the onset of the crystallization at 460 K.
Figure 4-16 shows the change of the XRD patterns of 100-150, showing the structural transformation due to crystallization. The smooth hump gradually transfers to a sharp peak with increasing time of the isothermal heat-treatment. According to the Mg-Cu phase diagram, the MgCu2 phase, with the FCC structure and a lattice constant a = 7.034 A。, would form in the Cu-MgCu2 region. Compared with JCPDS, the large peak appears to reflect the formation of the compound phase, suggesting the Mg-Cu amorphous phase transfers to the MgCu2
phase. The small humps at around 21o and 50o correspond to the formation of MgCu2 in (111) and (400) planes, respectively. In the case of 100-100, it exhibits the similar occurrence but
48
another peak forms at about 34o possibly owning to the contribution of the MgCu2 (220) plane, as shown in Figure 4-17.
While the Cu content decreases, a different structural transformation occurs in the 100-50 and 100-25 specimens. Figure 4-18 shows the transformation of 100-50, showing that the hump between 36o to 48o, suggesting a mixture of Mg2Cu (260), (331), (080), (351), (440), (191) planes and MgCu2 (220), (311), (222), (400) planes gradually sharpens, and the range of sharpening hump increases with time. In the case of 100-25, the structure of as-deposited film already exhibited a minor dispersed nano-crystalline Mg2Cu phase, as shown in Figure 4-19. As the heat treatment time increases, the peak shifts to the left compared to the original structure, and the strong peaks at 24.1o, 37.1o, 39.6o, and 44.5o formed which correspond to the Mg2Cu (131), (331), (080), and (440), respectively. Besides, during isothermal annealing, the small intermediate hump from 33o to 40o, such as the ellipses shown in Figure 4-19, appears in a short period, and disappears. In the meantime the peak of Mg2Cu (331) appears, suggesting that the as-mentioned hump seems to correspond the transitional phase induced by thermal diffusion.
4-4-3 Formation of intermediate phase of Mg-Cu multilayered thin films at 413 K and 363 K
According to the research of Arcot et al. [62], diffusion would occur at temperatures higher than 363 K. In the case of specimens consisting of thicker individual layers, no evidence of the Mg-Cu metastable phase, such as Mg-Cu amorphous phase, was formed in the Mg-Cu multilayered system as isothermally annealed at 413 K and 363 K, implying the thickness of Mg and Cu individual layers would lead to the variation of the overall interface energy in the mass-conservation system. The structural transformation of the 20T32
49
specimens annealed at 413 K is shown in Figure 4-20. The peaks of Mg2Cu at 19.5o, 37.4o, 39.6o, 44.6o, and 49.1o, indentified as the planes of (040), (331), (080), (440), and (191), gradually form, and the intensites of the peaks at 19.5o and 39.6o become very strong. In other words, the peak intensities of Mg and Cu decrease with increasing time.
The 40N32 specimens with the thicknesses of the individual layers ten times lower than those in 20T32, are annealed under the same condition. In addition to a slight negative heat of mixing between Mg and Cu (-3 kJ/mol), the thickness effect also induces the greater driving force for the Mg and Cu atoms to mix together. According to Figure 4-21, Mg2Cu rapid forms within 30 minutes during the isothermal annealing at 413 K owning to a large global interface energy. At 413 K, Mg2Cu formed easily since the Mg atoms diffuse into the Cu layers and react with Cu atoms rapidly.
For the case at a lower temperature, 363 K, the diffusion rate of Mg at this temperature is very slow [62]. Considering the case of 20T14, the specimens annealed at 363 K, as shown in Figure 4-22, the peaks of Mg2Cu still grow gradually with the time of the heat treatment, without the trace to support the formation of an amorphous phase. In the Cu-rich multilayered thin films, pure Cu peaks are left after annealing. Then, the phase transformation of 40N14 at 363 K, the formation of Mg2Cu with the increasing annealing time is similar to that in 40N32 annealed at 413 K, as shown in Figure 4-23.
Compared with the multilayered film consisting of thick and thin individual layers, Mg and Cu layers in the thinner-layer multilayered films would react with each other at a lower temperature and at a faster reaction rate than that in the thicker-layer multilayered films.
40N32 and 40N14 exhibit the same occurrence. Comparatively speaking, the current experiment and previous researches all express the missing metastable amorphous phase
50
during the thermal-induced diffusion between the Mg and Cu layers.
4-4-4 Structural transformation of 20T32 by TEM observation
According to section 4-4-3, the XRD results shows Mg2Cu rapidly forms during annealing. Then, from the XTEM observation, as shown in Figure 4-24 (a), for a representative image for the Mg-Cu multilayer, the white layers and black layers represent Mg layers with nominal 150 nm in thickness and Cu layers with nominal 50 nm in thickness, respectively. After annealing in vacuum at 413 K for 2 hours, Figure 4-24 (b) shows the thickness of Mg and Cu individual layers decrease, and the intermediate gray layers, Mg2Cu layers, rapidly formed within Mg and Cu layers due to a negative heat of mixing (-3 kJ/mol) and an increase of interfacial energy in the overall system. Moreover, according to the XRD result, as shown in Figure 4-22, the result in the 20T14 specimens annealed at 363 K is similar to that in the 20T32 specimen annealed at 413 K, implying that the Mg2Cu formation in not related to the thickness of the individual layer (greater than approximately 50 nm).
Also, the TEM observation proved the intermediate layers formed at the Mg-Cu interfaces
4-5 Mechanical analysis of co-sputtered 100-150 (Mg
17.7Cu
82.3), 100-100 (Mg
23.5Cu
76.5), and 100-50 (Mg
40.4Cu
59.6) using nanoindenter
Summarized sections 4-1 to 4-4, Table 4-2 shows the overall Mg-Cu specimens and their structural characteristics. Only three specimens, 100-150, 100-100, and 100-50, exhibit a structure consisting of the major Mg-Cu amorphous phase and the minor MgCu2 phase.
Hence, the mechanical properties would be focused on as-mentioned three specimens. For the cases of thin films, it is found that the detected hardness and Young’s modulus increase with increasing indentation depth, called as substrate effect, when the film is too thin, we would
51
not get the true properties of the films. However, the substrate effect is different case by case for different films on the same substrate. It is generally acknowledged that the nano-mechanical properties tested within 10% film depth of the film is free from the substrate effect [66].
The hardness-displacement and modulus-displacement curves, with a penetration depth of 400 nm at the strain rate of 5×10-3 s-1, are shown in Figure 4-25. It is obvious that the values of hardness almost maintain a constant over the displacement, and the value of Young’s modulus, too. In the initial stage (<50 nm in displacement), the obtained value is very irregular due to the geometry of the Berkevich tip.
Furthermore, Figure 4-26 exhibits the Young’s modulus and hardness of 100-150, 100-100, and 100-50 from the unloading curves. According to these results, the Young’s modulus and hardness of the 100-100 specimen are higher than those of 100-150 and 100-50.
Figures 4-27 to 4-29 are the load-displacement curves of 100-150, 100-100, and 100-50. In these curves, the “pop-in” effect appears during the indentation, showing the formation of many shear bands in the specimens.
52