[[include Navigation (not found)]]
JDFTx is a plane-wave density functional code designed for Joint Density Functional Theory (JDFT), a framework for ab initio calculations of electronic systems in contact with liquid environments. It is distributed under the GPL license (version 3 or higher) and publications resulting from its use must cite the following:
The code itself:
The algebraic framework on which the software is based:
The core algorithm employed in the electronic structure components of the software:
JDFTx evolved from an earlier in-house research code in the Arias research group at Cornell called DFT++, but at this point has been almost entirely rewritten in a modern object oriented framework taking advantage of C++11 for expressive code with advanced memory management, and CUDA for leveraging the computational power of GPUs. See [Compiling] for details on unlocking various features.
Unlike most other electronic structure codes, JDFTx performs total energy minimization using analytically continued energy functionals implemented within the algebraic formulation described in the above references, rather than density-mixing SCF schemes. Hence our motto "Our SCF never diverges, because we don't do SCF". This might be advantageous for vanilla DFT calculations in some cases (try your problematic systems out!), but it is quite important for reliable convergence in the presence of liquids, particularly with charged systems.
The framework of Joint Density functional theory is described in:
which should be cited for work using any of the fluid models. A polarizable continuum-like solvation approach which replaces the fluid by a local dielectric response is also described in that paper. The capability for handling charged systems and absolute voltage references for electrochemistry is added by incorporating ionic screening in the fluid according to:
The above fluid models are accessed using the "Linear" fluid option in the code, and are sometimes internally referred to as JDFT1.
The full power of Joint Density functional theory is unleashed when the electronic density functional is coupled to an explicit classical free energy functional for the fluid, which we sometimes call JDFT3. The coupling is typically achieved using density-only electronic functionals as described in
and available functionals for water include
R. Sundararaman, K. Letchworth-Weaver and T.A. Arias, J. Chem. Phys. 137, 044107 (2012) (dx.doi.org) and arXiv:1112.1442 (arxiv.org)
J. Lischner and T.A. Arias, J Phys Chem B. 114, 1946 (2010) (pubs.acs.org)
JDFTx also incorporates advanced algorithms for converging metallic systems with finite temperature Fermi function fillings, critical for the study of electrochemical systems. Our implementation extends an analytically continued free energy functional version of
C. Freysoldt, S. Boeck, and J. Neugebauer, Phys. Rev. B 79, 241103(R) (2009) (prb.aps.org)
R. Sundararaman and T.A. Arias, (under preparation)
to include minimization at constant chemical potential, and is described in the latter.
Please check back here for updates on citations that are under preparation.