TiO2 [11]. Here, we define the crystallization temperature (Tc
) as the
temperature at which the TiO2 (101) peak appears in the XRD pattern
after 1 h of annealing. We estimate the Tc
value of TiO2/Cu to be
~220 °C, which is almost the same as that of the TiO2/Ni film.
Next, we discuss the dependence of the MIC process for TiO2 on
the thickness of the Cu layer. We patterned Cu layers with thicknesses
of 2, 8, or 50 nm as bottom layers between the TiO2 thin films and
glass substrates. On bare glass without a Cu bottom layer (left regions
in Fig. 2(a–c)), only a few tiny (3–5 μm) crystallized grains appeared
after annealing at 240 °C for 1 h. In contrast, TiO2 fully crystallized on
the Cu bottom layer of ≥8 nm thickness, as shown in right regions in
Fig. 2(a–b). We detected no clear crystallization for the sample with a
2 nm-thick Cu bottom layer (right region in Fig. 2(c)).
According to the reaction-assisted MIC mechanism[11], we interpret
the crystallization process as a rearrangement of Ti–O bonds induced by
Cu–TiO2 interactions. The intermediate products of metastable Cu–TiO2
compounds easily decompose into Cu and TiO2, accompanied by TiO2
crystallizing at a lower temperature. The released Cu atoms continue
to interact with the remaining amorphous TiO2, resulting in full crystallization of the TiO2 thin film. In other words, the Cu atoms act as catalysts and diffuse from the bottom layer to the TiO2 surface.
To confirm the diffusion of Cu, we performed chemical composition
analyses of the TiO2
film surfaces using XPS, as shown in Fig. 3. The
XPS results from the surface of the as-grown sample before annealing
show only core levels and Auger lines from Ti and O except for the C
1 s peak from surface contamination. After annealing at 240 °C for 1 h,
the Ti peaks are very weak and distinct Cu 2p and 3p core level features
appear, showing that the surface of the crystallized TiO2
film is covered
with Cu. This result suggests that the reaction-assisted MIC mechanism
is also valid for the crystallization of TiO2/Cu.
In reaction-assisted MIC, some of the Cu atoms in the bottom layer
act as catalysts and travel to the surface of the film. To evaluate the
amount of Cu required to crystallize the amorphous TiO2
film, we
analyzed the RBS data before and after MIC. Fig. 4(a) compares RBS
spectra of TiO2/Cu before and after MIC. The Cu main peak was slightly
suppressed with the appearance of a shoulder at high energy side
after MIC. We simulated the RBS spectra using a model in which a part
of the Cu bottom layer (thickness of 18.5 nm before MIC) moved to
thefilm surface during the MIC process, as shown inFig. 4(b). Although
the shapes of experimentally observed spectra were qualitatively