Various properties of oxides can be designed by controlling the microscopic structures.
We have proposed a promising way to enhance the dielectric properties by tuning the composition of multi-element oxides.

In SSDM 2004 (K. Kita et al.) we have demonstrated, a dramatic enhancement of dielectric constant of HfO₂ approaching ~30 by a crystal-structure modification of HfO₂ for the first time in the world. This is realized by doping of different oxides, such as Y₂O₃, into HfO₂ [K. Kita et al., APL 2005.]. This concept enables us to design several types of novel higher-k ternary oxides including Si-doped HfO₂[K. Tomida et al., APL 2006], LaYO_x [Y. Zhao et al., APL 2009], and LaTaO_x [Y. Zhao et al., JAP 2009].

Rapid thermal annealing

We have started to investigate the structural control of High-k films without doping, but with an ultrafast thermal processes. Since HfO₂ has a tendency to start to crystallize in a higher-symmetric structure followed by a transformation to the monoclinic structure [Y. Nakajima et al., ECS Trans. 2010], the dielectric properties of HfO₂ is largely controllable by tuning the time and temperature. Now we are extending the phase control to Ferro-HfO₂.

Tuning the threshold voltage is one of the critical issues in high-k/metal gate (HK/MG) technologies. It is because the flatband voltage (VFB) of HK/MG stack shows a significant shift depending on the employed high-k materials. In 2006 (Y. Yamamoto et al., SSDM 2006, Yokohama), we have clarified that the one of the most important factor to determine the unexpected VFB shift is due to the dipole layer formation at high-k/SiO₂ interface [Y. Yamamoto et al., JJAP 2007].  

In 2008 (K. Kita and A. Toriumi, IEDM 2008, San Francisco) we have also proposed a possible physical origin of the dipole formation at high-k/SiO₂ interface. The structural difference between two oxides will be the key factor to determine the direction / strength of the driving force to form a dipole [K. Kita and A. Toriumi, APL 2009]

Thermally unstable GeO₂ electrically degrade channel interface of Ge MOSFET with GeO desorption. At first, we focused on substrate dependence of GeO desorption. The figure shows temperature desorption spectrometer (TDS) signal of GeO[K. Kita et al., JJAP 2008]. Low temperature GeO desorption is caused by direct GeO₂/Ge interfacing, and it is caused by the reaction of Ge + GeO₂ -> 2GeO.

 GeO desorption is limited by diffusion process in GeO₂[S. K. Wang et al., JAP 2010]. Now, we are focusing on the diffusing species by isotope marker. The figure is the GeO signal measured by thermal desorption spectrometer (TDS). Oxygen of desorbed GeO derives from top-side GeO₂ and that GeO itself is not diffusion species[K. Kita et al., IEDM 2009].

 Extinction coefficients of various GeO₂ estimated by spectroscopic ellipsometry. To obtain electrically excellent GeO₂/Ge interface, high pressure oxidation which thermodynamically and kinetically suppresses GeO desorption from GeO₂/Ge stack [C. H. Lee et al., APEX 2009]. Additionally, high pressure oxygen annealing affects bulk GeO₂ film quality. Sub-gap absorption (<6 eV) level is expected to be formed by GeO desorption [K. Kita et a l., IEDM 2009].

Electron inversion layer mobility of Ge (100) and (111) nMOSFET with GeO₂ gate stacks formed by high pressure oxidation [C. H. Lee APEX 2009]. Ge (111) should be better in terms of the lower effective mass in two dimensional inversion layer* and higher process stability [M. Toyama et al., SSDM 2004] . The peak mobility on Ge (111) is about 1100cm²/Vs at 300K and it is 1.5 times improved compared to that of Si Universal curve.
*S. Takagi, 2003 Symposium on VLSI Technology Digest of Technical Papers

C-V characteristics of LaLuO(rare-earth oxide)/Ge MIS capacitor with high pressure oxidation. High-k gate dielectric is required for enhancement of Ge MOSFET performance by scaling. We reported advantage of rare-earth oxides on Ge from the viewpoint of interfacial electrical characteristics [K. Kita et al., Appl. Surf. Sci. 2008]. Furthermore, combining this Ge friendly rare-earth oxide with high pressure oxidation (through rare-earth oxide), almost ideal C-V characteristics are obtained [T. Tabata et al., ECS Trans. 2008]. Now we are focusing on a role of rare-earth atom in interfacial passivation.

 At metal/Ge contact, band alignment is hardly determined by work function of metal. Fermi level of metal is strongly pinned to the valence band edge of Ge, hence Schottky characteristics is observed on n-type Ge and ohmic ones on p-type Ge irrespective to metal [T. Nishimura et al., APL 2007]. The origin of this strong pinning has not been clarified.
*H. B. Michaelson, J. Appl. Phys. 48, 4729 (1977).

Complete junction characteristics conversion (Schottky-ohmic conversion) by ultra-thin insulator insertion at metal/Ge interface. We found out that the strong Fermi level pinning at metal/Ge interface can be easily modulated only by ultra-thin insulator insertion[T. Nishimura et al., APEX 2008]. This is a key to understanding a guiding principle of band alignment formation at hetero junction. 

Graphene research has moved to Nagashio’s Lab. since April, 2014. Then, we have just launched TMD(transition metal dichalcogenide) research. Now, MoS2 is under investigation. The result seems to be quite interesting.