Is that any way to discover new physics without doing theory and experiments? Yes, by computational physics. Purpose of computational physics is to achieve fundamental understanding of physics by computer modeling. Computational physics now becomes so powerful in discovering science, and it has evolved into the third branch of physics (i.e., experimental, theoretical, and computational). Statistics shows that results in more than half of published journal papers are obtained by numerical approaches.

Then, what can computational physics do? Nowadays computational physics is so advanced that one needs only very minimum information such as chemical species and atomic charges to predict all sorts of materials properties. These properties include, for instances, where atoms will be exactly located, how chemical bonds are formed, how hard is the material, whether the material will be insulator or metal, what color the material samples will be… The list will just go on and on.

What are we doing here? Three subjects:

• Electronic structure calculations, by which many materials properties can be ready to understand.
• Optical properties of materials, in which interactions of matter with lights are considered.
• Electrical and electromechanical properties, where atomic structure, strain, and spontaneous polarization play roles

We care particularly about those materials of technological importance. We mainly focus, among others, on three classes of materials:

1. Semiconductor nanomaterials, especially quantum dots, wires, and films. Applications include nanolasers, nano solar cells, single-electron transistors, nano mechanical devices, and DNA-sequence labeling.
2. Ferroelectric and piezoelectric oxides, a special class of materials that can efficiently mechanical energy into electricity and vice versa. Applications include telecommunication, medical imaging, sonar listening devices, and military sensors.
3. Organic/inorganic hybrid materials, which combine the advantages of polymers and semiconductors. By these novel materials, "plastic semiconductors" can be made, and new age of electronic devices is possible.

We also develop computational codes targeted on new methods and new properties. Our current codes include (i) First-principles pseudopotential method with numerical atomic orbitals and plane waves as basis, (ii) Berry’s phase code within mixed-basis pseudopotential method to calculate polarization, (iii) Second-principles pseudopotential method, in which the screened atomic potentials are derived from first-principles LDA calculations.