|
The most noticeable capability in ATK is the transport module, which can compute the ballistic coherent tunneling current in a nanoscale device structure, via the use of a non-equillibrium Green's function (NEGF) formalism. Since it is also necessary to have an accurate description of the electronic structure, ATK also comprises methods for this. In fact, two different ones: density-functional theory (DFT) and extended Hückel. In addition, ATK includes a module with classical potentials for geometry optimizations and molecular dynamics calculations. ATK is thus a powerful general-purpose engine for both electronic structure and transport calculations in many kinds of atomic-scale systems.
Below is a detailed list of the basic features of Atomistix ToolKit (ATK). Note that via the Python scripting interface and plug-ins to the graphical user interface, users can also add new features to the platform themselves, or combine existing features into a complex workflow.
Some links in the list below point to additional information in the ATK Reference Manual, which is also a good resource to find out more about the capabilities of the software.
Also, don't forget that ATK contains a graphical user interface, Virtual NanoLab, with its own set of features!
Electronic structure methods
- LCAO-based DFT method
- Numerical atomic orbital basis sets (SIESTA-type), with detailed user control of basis set parameters
- Inclusion of indirect atom pairs for improved accuracy
- Ghost atoms (vacuum basis sets) NEW in 11.2
- High-quality norm-conserving (Troullier-Martins type) Hartwigsen-Goedecker-Hutter (HGH) pseudopotentials are provided for almost all elements of the periodic table, with possibility to use customized pseudopotentials UPDATED in 11.8
- The HGH potentials also include semi-core versions for many elements
- A wide selection of over 300 exchange correlation functionals (see full list), including
- LDA: HL, PW, PZ, RPA, WIGNER, XA
- GGA: BLYP, BP86, BPW91, PBE, PBES, PW91, RPBE, XLYP
- All in restricted and unrestricted (spin-polarized) versions
- Meta-GGA (TB09) for accurate band gaps of semiconductors NEW in 11.8
- Hubbard U term in both LDA and GGA, also in their spin-dependent flavors (LDA+U/GGA+U)
- "Dual" and "on-site" methods NEW in 11.2
- User-defined initial orbital occupations NEW in 11.2
- Semi-empirical method UPDATED in 11.8
- DFTB-type model, 30 different parameter sets are shipped with the product, and more can be downloaded and used directly
- Interface for input of user-defined Slater-Koster parameters; built-in models for group IV semiconductors and III-V binary alloys
- With additional Hartree term for self-consistent response to the electrostatic environment
- Extended Hückel model with over 300 pre-defined basis sets for (almost) every element in period table (Müller and Hoffmann) + Cerda parameters (binary alloys, semiconductors, metals)
- Spin polarization term can be added via internal database of spin-split parameters
Transport calculations
- Improved TranSIESTA-based method (self-energy coupling to semi-infinite leads) for two-probe systems with open boundary conditions
- Non-equillibrium Green's function (NEGF) description of the electron distribution in scattering region
- O(N) Greens function calculation and sparse matrix description of central region
- Fast self-energy calculations through Krylov subspace methods
[H. H. Sørensen et al., Phys. Rev. B 77, 155301 (2008) and Phys. Rev. B 79, 205322 (2009)]
- Scattering states method for fast contour integration in non-equillibrium (finite bias) NEW in 11.2
- Ability to treat heterogeneous systems (different electrodes)
- Use of electronic free energy instead of total energy for open systems
- Fast transmission spectrum calculation for perfect periodic systems NEW in 11.2
|
 |
Electrostatic model
Physical quantities (electronic structure)
- Molecular spectra
- Also projected molecular spectra for periodic systems
- Band structure
- Density of states (DOS)
- Projected DOS (PDOS) onto atoms and angular momenta
- Eigenfunctions (molecular orbitals)
- Bloch functions
- Mulliken populations
- Real-space 3D grid quantities
- Electron density
- Effective potential
- Electrostatic potential
- Orbitals
- Forces (analytic Hellmann-Feynman)
- Stress NEW in 11.2
- Optical properties NEW in 11.8
- Kubo-Greenwood formalism for linear optical properties
- Calculation of optical absorption, dielectric function, refractive index, etc
|
 
  |
Transport analysis
Ion dynamics
General
- Fermi level smearing for improved convergence stability
- Customizable Pulay mixing for self-consistent scheme
- Monkhorst-Pack k-point sampling grids
- Use of Brillouin zone symmetries for improved efficiency
- Initialization of a new calculation via the self-consistent density matrix of a converged one (with automatic spin alignment)
- Kerker preconditioner for improved convergence NEW in 11.2
- Custom initial spin filling schemes NEW in 11.2
- Proprietary sparse matrix library
- Options to prioritize memory conservations vs. performance NEW in 11.2
I/O
- CAR, CIF, XYZ, VASP import filters for geometries UPDATED in 11.8
- XYZ export filter
- Cube file export for 3D grid data
- 2D data export (x/y data)
- Store and restore the state of a calculation for deferred analysis or restart
- Checkpoint files for restarting crashed calculations NEW in 11.2
- NetCDF files with open format and logical structure (IDs, labels, comments) for storing data, geometries, etc
|
 |
Interfaces
- Graphical user interface (GUI) Virtual NanoLab
- Plug-ins, users can develop custom modules
- Python scripting language interface
- Directly coupled to GUI
- Can be used interactively
|
  |
Integrated external calculators
Platform support
- ATK is supplied as a binary installer on all platforms. Nothing to compile or customize, just download and install on
- Windows XP/Vista/7 (GUI and back-end, 32-bit and 64-bit)
- Linux, most distributions (GUI and back-end, 32-bit and 64-bit, GLIBC 2.3 or 2.4)
- MPI parallelization (Windows and Linux), based on MPICH2. ATK has been successfully tested with
- Automatic OpenMP threading support on multicore processors via the Intel Math Kernel Library (MKL)
|
  
|
|
sheldrakes-bath
