Transparent Semiconducting Oxides Active
Transparent semiconducting oxides (TSOs) like In2O3, ZnO, Ga2O3, and their compounds have unique materials with two opposing material properties of high conductivity and transparency of insulator. In contrast to other conventional semiconductors such as Si and GaAs, the electron mobilities in amorphous phase of TSOs are comparable to those of their crystalline phase due to the spherical symmetry of metal s orbitals of the conduction band showing the significantly large values in comparison with that of amorphous Si (a-Si). In addition, an amorphous material can be easily formed over large area at low temperature on both organic and inorganic substrates, which facilitates the application of TSOs as the active channel layer for the state-of-art flexible and large area display with low cost. Even though many electrical and optical properties of TSOs have been found from experiments, the lack of understanding on the fundamental material properties interrupts to optimize the performance of the device with TSOs. For example, the negative or positive shift of the threshold voltage of TFT using TSOs under light and gate voltage stress is too large affecting the stability of the devices badly, but the microscopic origin for that still remains unclear. Therefore, the theoretical studies on them have been highly required and related researches and projects are actively in progress.
Electronic band structure is the most significant information to grasp the microscopic behavior of electrons in a solid. Even though density functional theory (DFT) with local density approximation (LDA) or generalized gradient approximation (GGA) has been successful to predict the physical properties of materials, it undergoes some difficulties to evaluate the accurate band structure due to the poor description of many-body interaction. Indeed, the experimental band structure would be determined by the photoemission and inverse photoemission experiments. They measure the kinetic energy of the electrons absorbing the photon or the energy of the photons generated by introducing additional electron into a solid. In the microscopic point of view, when an electron comes in or out, the surrounding electrons response to the external perturbation screening the electric field induced by introduced electron or hole. As a result, to evaluate the electronic band structure, the response of electrons to the external perturbation should be considered.
Recently, many-body perturbation theory within GW approximation based on Hedin equation has attracted much attention as the method which provides a way to overcome the problems in DFT. The predictive capabilities of GW approximation originate from involving the many-body effects between electrons by introducing quasiparticle picture. In GW approximation, an electron or a hole combined with surrounding screening electron cloud is dealt with as a particle (not a real particle so we call it quasiparticle) and the self-energy corresponding to the exchange-correlation energy on the quasiparticle is evaluated. The calculation of the self-energy is achieved via multiplying the Green’s function (G) and the screened Coulomb interaction operator (W) with the frequency-dependent dielectric matrix as opposed to LDA or GGA in DFT. Therefore, many studies have reported that DFT based GW yield gaps that are in good agreement with experimental values.
Materials Informatics Active
Through the history of mankind, technical innovations of the time had always accompanied with a discovery of new material. Similarly, in modern science and technology, improving the desired properties by materials design is still the key breakthrough to overcome existing technical barriers. Even though most basic 작은 principles of material properties are well-identified by quantum mechanics during 20 century, there's still no robust way to predict the exact properties of vast unexplored materials. Recently, however, remarkable advances in computing power and first-principles calculation techniques present a new opportunity to explore this unexploited land. As the fast and accurate prediction using first-principles calculation is enabled, the high-throughput calculation is rising up as a strong tool for the data mining of materials properties.
As an example, band gap and permittivity are fundamental properties of materials and both are key properties for high-k or low-k dielectric materials for microelectronic devices. But there are only a little portion of materials which were experimentally identified for these properties due to the costs and difficulties of the measurements. It is said that band gap and permittivity generally has trade-off relation for known materials, but there is no clear explanation about the relation and descriptors of each property are still obscure. To get the massive property database for data mining, we developed the fully-automated code for high-throughput calculation of band gap and static dielectric constant. As a result, we managed to calculate more than 1,000 binary and ternary crystal oxide from ICSD database. From the newly obtained database, we found some interesting candidate materials for next generation high-k dielectrics and now we are trying to find the solid descriptors for each property by analytic and statistic approaches.
Energy materials: (Oxide) catalyst
Figure 1 Oxygen reduction reaction supported by absorption of photo energy or by photocarriers. (Ref.: Workshop of Molecular Photoreactivity on Metal-Oxide Surface from First-Principles, 2009)
Various measures of global security indicate that the world is heading toward an energy crisis. Many efforts are given to find the best solution in the form of the renewable energy. For example, people struggle to find effective catalysts which enable favorable energetics for energy conversions from renewable energy sources such as sunlight.
They are interested in oxide catalysts which show fine stability, but usually there is a trade-off between stability and efficiency of catalysts. Therefore, we approached the issue from two different ways.
First, we try to find new oxide materials that support effective catalytic reactions. We can analyze relevant bulk properties of any new materials if we know the structure, like absorption properties of sunlight, effective mass of carrier or band alignment. Usually, the structure is driven by forgone experiments or an educated guess.
Second, we optimize promising candidates for catalyst by surface modification methods, morphology controls or defect controls. We also study the catalytic reactions on well-studied surfaces that we find the mechanisms of the reaction and decisive step among the whole process. It makes possible to scan much more materials by a single simple descriptor, computationally. Either, we study about surface modification effect or effect of defects by calculating the descriptor value at each case.
Being lacking in the atomic scale knowledge for now, studying all of these by atomic level computation will provide useful information in optimizing the properties of catalyst, or in finding new effective materials.
Nanopore Device Simulation
Since the Sanger’s chain-termination sequencing method shed a light to the human DNA sequencing, people have been trying to find more effective and less expensive ways to analyze human DNA. Among a number of attempts, nanopore membrane device has been considered as a promising candidate because of its speed and cost-effectiveness.
Nanopore sequencing system is composed of a biological or synthetic membrane with nanometer-sized pore in electrolytic solution. When external electrostatic field is applied, biomolecules in electrolytic solution are driven through the nanopore changing several properties of the system(e.g., permittivity, ionic current or transverse current). Since these changes contain information about the geometric or electrical characteristics of each biomolecules, it is expected that one can identify biomolecules by simply analyzing these changes.
In our group, nanopore sequencing device is simulated within multi-scale approach.
In atomic scale, we use ab-initio calculation to compute dielectric constant of molecules in nanopore device. Dielectric response of a biomolecule in a theoretical model structure can be estimated using orbital separation approach (OSA). Detailed description of the calculation scheme is given
In molecular scale we perform a classical molecular dynamics (MD) calculation. Classical MD uses classical potentials to approximate interactions between atoms, allow one to simulate a system with several million atoms. Translocation of biomolecules and ionic current profile of nanopore devices are demonstrated and analyzed with classical MD code.
Finally, we utilize finite element method(FEM) to model macroscopic phenomena in nanopore system. By solving Navier-Stokes equation coupled with Nernst-Plank equation, we can simulate macroscale fluidic motion of nanopore system. Several phenomena(e.g., electric double-layer which cannot be understood by
Currently, due to the development of fabrication techniques, the manipulation of various nano-devices becomes feasible. Such devices sometimes have unusual properties that cannot be understood from classical physics. Metal-insulator-metal (MIM) capacitor which is used for DRAMs is one of them. In classical approach, capacitance of MIM capacitor is only depended by dielectric constant of insulator; in nanoscale capacitance, however properties of interface between metal and insulator are also very important.
Figure 1 Schematic representation of an MIM capacitor featuring a high-k HfO2 thin film on a PI substrate (Ref.: Phys. Chem. Chem. Phys., 12, 2582 (2010))
In order to understand nanoscale capacitance and its interface effect, we developed orbital separation approach (OSA) code. It can be used for simulating the effect of bias voltage within the Kohn-Sham formalism of density-functional theory. The method is robust and efficient and is unique in that it provides a well-defined total energy of the charged MIM capacitor. It can also be used for grapheme electrodes system.
Figure 2 Schematic model of the simulation models considered in orbital separation approach (Ref.: Phys. Rev. B., 84, 085120 (2011))
Now we use the OSA code for nanopore sequencing simulation. Biomolecules such as DNA and peptide can have different dielectric constant. Therefore, we try to calculate the dielectric constant of each biomolecules and show the difference between each biomolecules.
4 nodes :
2 * Intel(R) Xeon(R) CPU E5-2650 2.00GHz
4 * 20MB cache memory
(Installed date : 2013.02)
4 nodes :
2 * Intel(R) Xeon(R) CPU E5-2670 2.60GHz
4 * 20MB cache memory
(Installed date : 2013.08)
4 nodes :
2 * Intel(R) Xeon(R) CPU E5-2680V2 2.80GHz
4 * 20MB cache memory
(Installed date : 2014.04)
24 nodes :
2 * Intel(R) Xeon(R) CPU E5520 2.27GHz
4 * 8MB cache memory
(Installed date : 2009)
8 nodes :
2 * Intel(R) Xeon(R) CPU E5-2630 2.30GHz
4 * 16MB cache memory
(Installed date : 2012.5)
24 nodes :
2 * Intel(R) Xeon(R) E5420 2.50GHz
6MB cache memory
(Installed date : 2006)