The Smeagol Project

Hartree potential isosurface of a Benzene dithiol molecule connected to Au-[111] electrodes The need for smaller electronic devices is leading science and technology into a new era: the nanotechnology era. A race has started for constructing new devices, containing only a handfull of atoms, but capable of performing the same tasks of today's integrated circuits. This new generation will be faster, cheaper, smaller and more compact than the present one, and novel functionalities are envisioned.

The possibilities are unlimited and it is difficult to predict a clear roadmap for the future. What is certain is that at the atomic length scale quantistic effects dominate, and the description of a device must be fully quantum mechanical. This is a formidable theoretical task.

Materials specific computer simulations play an important role in engineering new nanoscopic devices. These simulations allow us to make quantitative predictions providing experimentalists with a better understanding and clear directions for designing atomic scale devices.

  Smeagol has been designed to calculate transport properties of atomic scale devices. It is the result of a collaboration between the Computational Spintronics group at Trinity College Dublin, and the condensed matter groups of the University of Lancaster and Oviedo. These three teams form the "Ab inito transport" consortium. This is a multi-disciplinary team aimed to develop efficient and accurate transport methods using state of the art electronic structure calculation schemes.

Smeagol is an ab initio electronic transport code based on a combination of Density Functional Theory (DFT) and Non-Equilibrium Green's function transport methods (NEGF). It has been designed to describe two terminal nanoscale devices, for which the potential drop must be calculated accurately.

Smeagol uses DFT as main electronic structure tool. The Kohn-Sham equations for an open non-periodic system are solved in the NEGF scheme, and the current is then extracted from the Landauer formula. At present we have chosen SIESTA as our DFT platform, although other possible methods are under investigation. SIESTA is particularly convenient since it uses a localized pseudo-atomic basis set, which allows order N scaling.

Smeagol and Spintronics

BDT Molecule connected to Ni leads Besides the ability of downscaling integrated circuits for computer applications, nanotechnology offers the possibility of exploring properties inherent of the quantum world. One of these is the use of the spin degree of freedom as a way of storing, transmiting  and manipulating information. This is the burgeoning field of Spintronics (or spin-electronics).

Spintronics begun in 1988 with the giant magnetoresistance effect (GMR). This is the change of the electrical resistance of a magnetic multilayer, when a magnetic field is applied. Within a decade GMR moved from an academic curiosity to a multibillion reality and at present the read/write component of every hard drives on the market is based on this phenomenon.

However, as with other electronic devices, the fundamental limit of GMR will be reached soon and a new family of devices must be produced to meet the target of storage densities of the order of 1 Tbit/in2 . One possible solution is to combine magnetic materials with molecules, effectively joining the fields of Spin- and Molecular-electronics.

For this reason Smeagol has been specifically created to deal with magnetic systems. It is fully spin polarized and it includes the possibility of performing non-colinear spin calculations.

Smeagol Characteristics and Capabilities

  • Linear combination of atomic orbitals (LCAO) basis set.
  • Uses the Kohn-Sham equations as a single particle equation for electronic structure calculations.
  • Uses the Keldysh Green's function method to obtain the density operator for an open system.
  • Capable of performing calculations on both extended systems as well as molecules (Γ point and k points).
  • Capable of performing calculations of up to 100 atoms.
  • Fully spin polarized including spin non-collinearity.
  • Fully parallelized using the Message Passing Interface (MPI).

Back to top | Home | About | The team | Publications | Downloads | News | Links

Last updated: 9/04/2004.