Computational Platform @ IIT-CBN

Research Activity:

Computational approaches are of nowadays of utmost importance in the field of nanoscience and nanotechnology, to understand the physics of new materials and to design more efficient devices.
Despite recent advances in specific fields, such as quantum chemistry and solid-state physics, much less attention  have been paid to the multiscale integration of different computational techniques. Such integration is of fundamental importance to tackle several challenging applicative problems in nanoscience.
The computational activity of the Research unit will be to develop new computational methods for the Integrated Multiscale Computational Technology project of the IIT Computational Platform and to  apply them to applications of large interest for the Energy and Smart Material platforms, such as semiconductor nanocrystals and organic-inorganic interfaces.
The activity of the computational team  will be based on three interacting main activities

A1) Modeling of nanocrystals: Inorganic nanocrystals are at the forefront of the present revolution in nanotechnology for their unique applications optoelectronics and biology. Typical inorganic nanocrystals are composed by a semiconductor core surrounded by organic surfactants. The inorganic core can be of complex shapes (rod,  tetrapods, core-shell systems ,.. ), with different dimensions (from 5 nm to 100 nm). For a theoretical description of such complex nanosystems  modeling at different scale is required. We aim to develop a new code  to study the electronic and optical properties of nanocrystals of arbitrary shapes, dimensions and materials. This code  is based on the envelop-function-approximation (EFA) including both single-band (effective mass) and multi-band (k.p) schemes.. The wavefunctions will be calculated on an real-space adaptive-grid with iterative diagonalization schemes. Excitonic properties will be described using configuration-interaction(CI) schemes including the description of the external (e.g. biological) polarizable environment.

A2) Modeling of organic-inorganic interfaces; Organic-inorganic interfaces play a key role in different nanoscience research areas such as molecular electronics and photovoltaic cells. When organic molecules are physi- or chemi-sorbed on a metal (e.g. gold, silver, … ) or semiconductor (silicon, TiO2, ...)  substrate, their electronic and optical properties change significantly. For an accurate theoretical description of interface properties first-principles approaches are required. Density-functional-theory (DFT) methods using conventional (i.e. the local density approximation or the generalized-gradient approximation) exchange-correlation (XC) functionals have been applied to different properties such as adsorption geometries and surfaces reconstruction. However conventional XC functional cannot describe correctly charge-transfer effects, one of the main issue in nanodevices.
We will develop and apply new advanced theoretical methods to correctly describe these effects, namely: Hybrid DFT functionals, i.e. containing a constant fraction of the non-local Hartree-Fock Exchange; Orbital dependent DFT method, based on the optimized effective potential (OEP); Many-body perturbation theory(MBPT) technique, such the the GW approach.
We will investigate different hybrid interfaces, namely: Organic dyes chemisorbed on semiconductor substrate to optimized photovoltaic cells; Surfactants/semiconductor interfaces; Organic molecules chemi-/physi-sorbed on noble metal substrates, to study the electronic level alignment;

A3) Developing of embedding methods.
While periodic systems can be well described using plane-wave (PW) pseudopotential approaches, isolated systems, such as nanosystems in solution or the low-coverage regime of organic molecules on inorganic substrates are better described using atomic-orbital(AO) based methods. Moreover AO method can treat the non-local exchange (required to described charge-transfer effects) more efficiently than PW methods. However the main limitation of conventional AO schemes is that extended substrates are described using finite clusters.  
We aim to develop efficient embedding schemes based on the so called frozen density orbital free (FDOF). The FDOF method can be applied both to solute molecules and to semiconductor or metal extended environment. The main aim is to describe complex systems using a full ab-initio (i.e. without empirical parameters) multiscale approach: a smaller, energy relevant part of the systems will be described using high accuracy correlated wavefunction approaches  (or accurate OEP DFT methods), while the rest of the systems will be described using local DFT methods.

Labs: Computational Center
Facilities : 2TeraFlops HPC Cluster  with dual Intel-Nehalem quad-core cpu,, 24GB ram; Infiniband interconnection network; cluster-file-system storage.