Due to the strain gradient near each surface of a BaTiO3 nanocube in their ordered assembly, electric polarization appears due to flexoelectric effect. The magnitude of the flexoelectric polarization could be one order of magnitude larger than that of ferroelectric spontaneous polarization of BaTiO3. Thus, dielectric response of an assembly could be dominated by that of the flexoelectric polarization if there is no ferroelectric domain-wall motion. Numerical simulations of the dielectric response of a BaTiO3 nanocube in an ordered assembly are performed from kHz to MHz range based on a dynamic model of flexoelectric polarization assuming anharmonic potential. The calculated temperature dependence of the dielectric constant is consistent with the experimental data of high dielectric constant with nearly-flat temperature dependence. It is suggested that high dielectric constant with nearly-flat temperature dependence is not originated in ferroelectric nature of BaTiO3 nanocubes but originated in flexoelectric polarization in nanocubes which is also seen in non-ferroelectric materials.

It has been reported that the flexoelectric effect could be dominant in the nanoscale. The discrepancy between theory and experiments on the frequency dependence of the dielectric constant of an ordered assembly of BaTiO3 nanocubes is nearly resolved by assuming the coexistence of flexo- and ferro-electric effects. Although flexoelectric polarizations perpendicular to the applied alternating electric field contribute to the dielectric constant, those parallel to the electric field do not contribute because the magnitude of the flexoelectric polarization does not change due to the mismatch of strain at the interface of the nanocubes. On the other hand, some dielectric response is possible for the ferroelectric component of the polarization parallel to the electric field.

Locally selective formation of gold nanostructures on a silicon wafer without photolithography, electrolysis, or hydrofluoric acid was reported previously. With our previous method, gold nanostructures grow selectively on the locally amorphized surface on a silicon wafer when the silicon wafer is exposed to a customized Au ion-containing solution. The locally amorphized surface is produced by, for example, irradiation of a focused ion beam (FIB) or a femtosecond laser. The Au ion-containing solution is prepared by mixing (3-mercaptopropyl)trimethoxysilane (MPTMS) with an aqueous solution of HAuCl 4 . Here, the mechanism of selective growth of gold nanostructures on the amorphized silicon is explained. Au ions in the solution placed on the FIB-or laserprocessed silicon are reduced by the amorphized silicon. Silicon-dangling bond defects, which exist in the amorphized silicon, seem to reduce the Au ions. Finally, we find that the use of MPTMS is unnecessary and a simple HAuCl 4 aqueous solution produces desired results.

The direct growth of gold nanostructures at the interface between a focused ion beam (FIB)-amorphized silicon surface and a HAuCl4-dissolved solution is demonstrated, and its mechanism is discussed in terms of the solvent for HAuCl4. Gold nanostructures successfully grew when a protic solvent such as water or alcohol, which has a hydroxyl group, was used as the solvent for HAuCl4. In contrast, no gold nanostructures were observed in the case of dimethyl sulfoxide (DMSO), an aprotic solvent having no hydroxyl group. Each protic solvent deactivates the FIB-amorphized silicon surface in the absence of Au ions. The growth of gold was inhibited when the FIB-amorphized silicon was exposed to the solvent before exposure to HAuCl4(aq). Silicon dangling bonds, which commonly exist in amorphized silicon, seem to reduce Au ions with the participation of protic solvent molecules in the redox reaction, initiating the growth of gold. The results of time-of-flight secondary ion mass spectrometry (TOF-SIMS) also support our suggestion.

We describe and experimentally verify the mechanism of area-selective electroless deposition of gold nanostructures at interfaces between locally induced naked silicon surfaces and pure HAuCl 4 solution. We previously demonstrated that gold nanostructures selectively grow on focused ion beam (FIB)-irradiated areas of a silicon substrate when the substrate is exposed to pure HAuCl 4 solution. We revealed that silicon dangling bonds formed by FIB irradiation reduce Au(III) ions and initiate the nucleation of gold. Here, we focus on the growth stage. Quantitative estimate of the amount of gold grown reveals that silicon dangling bonds explain only a fraction of the total amount of gold. Thus, the growth mechanism is different from the nucleation mechanism. In the growth stage, soon after silicon/gold interfaces form via gold nucleation, electrons in the crystalline silicon near the interfaces reduce Au(III) ions through the interfaces and the growing gold. This mechanism is consistent with the Fermi level (chemical potential of electrons), which decreases in the order silicon > gold > HAuCl 4 solution in the isolated form. Double exposure to the same and different HAuCl 4 solutions shows the results that support the mechanism, and allows further control of gold morphology.