Comparing Figure 3a (6 h annealed)

and Figure 3b (9 h annealed), the atomic ratio of Si to Al of the microparticle formed through 9 h annealing (50.5%) PXD101 is much larger than that of the microparticle which underwent 6 h annealing (10.5%). Taking into account that the annealing SYN-117 in vitro temperature (550°C) of the present study is lower than the eutectic temperature (577°C) of Al-Si systems and the Si solubility in Al crystal is only about 1.4 at. % at 550°C [22], the measured large Si concentrations reflect solid-state interdiffusion of Al and Si atoms facilitated by compressive stress that is developed by larger expansion of Al film than Si substrate during annealing (see the middle panel of Figure 1) [23, 24]. It is speculated that more mobile Al atoms move first over the surface or through grain boundaries to agglomerate, leaving behind a lot of vacancies. These vacancies in Al film may accelerate outward diffusion

of Si atoms and direct Si atomic flow to Al granules to finally form Al-Si alloys. In addition, since the surface energy of Si (100) plane is relatively high (2.13 J/m2) [25], Si atoms are prone to diffuse into a foreign material at elevated temperatures to reduce the surface energy. This hypoeutectic interdiffusion progresses further as the annealing time is made longer. The atomic ratio of Si/Al rises to 82% for selleck kinase inhibitor a microparticle from the 90-nm-thick film, as shown in Figure 3c. This may be because a larger volume of Al vacancies in Al film absorbs more Si atoms from the substrate. As a consequence of Al-Si microparticle formation, the majority of the original Al film is exhausted as seen in Figure 3b,c. Interestingly, it is found from Figure 3c that the residual Al film resembles the network structures of narrow channels. Figure 3 SEM images of microparticles. SEM images of microparticles transformed through (a) 6 h annealing and (b) 9 h annealing of a 40-nm-thick Al film and (c) 9 h annealing of a 90-nm-thick Al film on Si substrate. Annealing temperature was set at 550°C. Scale bars 2.5 μm. Histone demethylase EDX element analysis results are also presented

for the particle area (notated as ‘1’) and the rest (notated as ‘2’), respectively. The composition and the crystal structure of both untreated and heat-treated Al films on the Si substrate were further analyzed using XRD. Figure 4 shows XRD patterns of 90-nm-thick Al films before and after annealing. For both samples, three major peaks are sharp, representing the samples are crystalline irrespective of heat treatment. The overwhelming peak of 68° to 69° is assigned to Si (400). Al (220) peak that usually appears around 66° is presumed to be superposed with the Si (400) peak. The other two peaks observed at approximately 33° and 62° are related to Al2O3 or Al-Si oxide. The peak intensities of a 9-h annealed sample are far larger than those of the untreated sample at those 2θ angles, particularly at approximately 33°.