Description

Pescan is used for nonselfconsistent calculations of electron and hole states of large (a few thousands to hundreds of thousands of atoms) nanostructures. The electron wave function is represented by a plane wave basis set. Fast Fourier transform (FFT) is used to transform the wavefunction from reciprocal space to real space to be applied to the potential. Nonlocal pseudopotentials are used in the Hamiltonian and spin-orbit interactions are included.

The code calculates the interior eigenstates of a nanosystem with a given Hamiltonian. The primary approach for the calculation of the interior eigenstates in the code is based on the folded spectrum method. Several numerical algorithms can be used together with this approach, including the conjugate gradient algorithm, (variants of) the Davidson algorithm, and the LOPCG algorithm. The user can select an algorithms based on the problem at hand also on and number of wavefunctions that need to be computed.

To use the Pescan code, one needs to provide:

• a pescan.input file, specifying the calculation parameters
• an atom.config file, describing the atomic position of the system
• a pot.system file, containing the single electron potential in a real space n1*n2*n3 grid
• a vwr.atom file, containing the nonlocal pseudopotentials.

The atom.config file is usually generated by some simple utility program. For a strained system, a valence force field (VFF) calculation is done to relax the atomic positions. The vwr.atom is the usual norm conserving pseudopotential file (in the PEtot style). There are different ways to generate the potential file, pot.system, of the system. For empirical pseudopotential method (EPM), pot.system is generated from atom.config and the corresponding EPM files (to be provided in the other part of the DOE-Nano site). For charge patching method (CPM), charge density of the system is first generated from CPM, then DFT potential is calculated from the charge density, and possible correction is added to vwr.atom to correct the DFT/LDA band gap error. One can of course also use the potential file from the selfconsistent PEtot calculation.

A brief history

An old version of the serial escan code was developed by L.W. Wang in National Renewable Energy Laboratory. After Dr. Wang moved to Lawrence Berkeley National Laboratory in 1999, he rewrote the code to make it parallel using the parallel FFT developed by A. Canning, and the nonlocal part of the pseudopotential was implemented in the standard Klainman-Bylander form. After 2006, new solvers: LOPCG, Davidson methods, and ARPACK are added to the code by C. Voemel, S. Tomov and O. Marques.

References

• L.W. Wang and A. Zunger, Solving Schrodinger's Equation Around a Desired Energy: Application to Silicon Quantum Dots, letter, J. Chem. Phys. 100, 2394 (1994).
  
• A. Canning, L.W. Wang, A. Williamson and A. Zunger, Parallel Empirical Pseudopotential Electronic Structure Calculations for Million Atom Systems, J. Comp. Phys. 160, 29 (2000).
• C. Voemel, S. Tomov, O. Marques, A. Canning, L.W. Wang and J. Dongarra, State-of-the-art Eigensolvers for Electronic Structure Calculations of Large Scale Nano-systems, J. Comp. Phys., to appear.

Availability

• Download Pescan (tar file)

• Example of input file: pescan.input
  
• Example of input file (options for PRIMME): pescan.input.primme
• Example of atomic coordinates: atom.config
  
• Example of atomic pseudopotential: vwr.Cd
  

• Benchmarks

• Guidelines
  
• The data used for the benchmarks are available in the directories 20Cd19Se, 83Cd81Se, 232Cd235Se, 534Cd527Se, dot5 and Qwire. The potential file for Pescan is stored in binary format. Therefore, those directories provide the potential file in big endian (which can be read on IBMs) and little endian (which can be read on Crays).
  

Software Required

The following libraries are used by Pescan and need to be downloaded and installed separately:

• FFTW, used for computing discrete Fourier transforms.

• LAPACK, used for linear algebra calculations.

• PARPACK, used for eigenvalue calculations.

• PRIMME, used for eigenvalue calculations.

• Sample of systems calculated with Pescan

This work was supported by the US Department of Energy,Office of Science, SciDAC, under LAB3-17 initiative, contract nos. DE-FG02-03ER25584 and DE-AC03-76SF00098.