LIBXC1
LIBXC1 = [integer] or [string]
Description: LIBXC1 and LIBXC2 specify the functionals from the library of exchange-correlation functionals Libxc[1][2][3] that one wants to use.
LIBXC1 and LIBXC2 can be either integers or strings. The available functionals are listed on the Libxc website[3] along with their associated number and tag (e.g., 101 and GGA_X_PBE for PBE exchange). The tag indicates if this is an exchange (X), correlation (C) or exchange-correlation (XC) functional, and to which family it belongs to, namely LDA (LDA or HYB_LDA), GGA (GGA or HYB_GGA) or meta-GGA (MGGA or HYB_MGGA). Note that calculations with laplacian-dependent meta-GGA functionals are not yet possible with VASP.
Note that Libxc is a separate program that has to be downloaded[3] and installed before VASP is compiled with the corresponding precompiler flag and optionsinstall DFTD4 and add it to the makefile.include
The allowed possibilities for LIBXC1 and LIBXC2 are the following:
- Both LIBXC1 and LIBXC2 are specified and correspond to exchange and correlation functionals, respectively, or vice versa.
- Only LIBXC1 (or LIBXC2) is specified and corresponds to an exchange-correlation, exchange or correlation functional. As a side note, the latter choice would not, from the physical point of view, make sense. Note also that if LIBXC1 (or LIBXC2) is an exchange-correlation functional, then the other tag can not be used.
- LIBXC1 and LIBXC2 can correspond to functionals of different families, e.g., a meta-GGA and a GGA, respectively.
Regarding other tags in INCAR related to Libxc:
- One also has to specify GGA = LIBXC for LDA, GGA and GGA-based hybrid functionals or METAGGA = LIBXC for meta-GGA functionals (calculations with meta-GGA-based hybrid functionals are not yet possible with VASP).
- Many of the functionals implemented in Libxc have parameters that can be modified. This can be done via the tags LIBXC1_Pn and LIBXC2_Pn, where .
- The tag TBOUNDLIBXC, which is .FALSE. by default, allows to enforce the lower bound on the kinetic-energy density () with before is used in a meta-GGA functional from Libxc.
For calculations with hybrid functionals (i.e., when LHFCALC = .TRUE.), the following provides some explanations:
- The Libxc functionals whose tag starts with HYB_LDA_XC, HYB_GGA_XC or HYB_MGGA_XC (not yet usable with VASP) consist directly of the whole (XC) semilocal component of a hybrid functional. For these functionals AGGAX, AGGAC and ALDAC can not be used (more informations on how to modify the mixing and screening parameters can be found here LIBXC1_Pn).
- The Libxc functionals whose tag starts with HYB_GGA_X or HYB_MGGA_X (not yet usable with VASP) consist of the exchange semilocal component of a hybrid functional. For these functionals AGGAX can not be used (more informations on how to modify the mixing and screening parameters can be found here LIBXC1_Pn).
- If the semilocal component of the hybrid functional is constructed using Libxc functionals that do not contain HYB in the tag, then (1-AEXX), AGGAX, ALDAC and AGGAC will be used and multiply
- respectively, where and .
Examples of how to specify a Libxc functional in INCAR are given below for the GGA PBE[4], meta-GGA SCAN[5] and hybrid PBEh (PBE0)[6] functionals.
- PBE
GGA = LIBXC LIBXC1 = 101 LIBXC2 = 130
or
GGA = LIBXC LIBXC1 = GGA_X_PBE LIBXC2 = GGA_C_PBE
- SCAN
METAGGA = LIBXC LIBXC1 = 263 LIBXC2 = 267
or
METAGGA = LIBXC LIBXC1 = MGGA_X_SCAN LIBXC2 = MGGA_C_SCAN
- PBEh (PBE0)
LHFCALC = .TRUE. AEXX = 0.25 GGA = LIBXC LIBXC1 = 406
or
LHFCALC = .TRUE. AEXX = 0.25 GGA = LIBXC LIBXC1 = HYB_GGA_XC_PBEH
Related Tags and Sections
LIBXC1_Pn, TBOUNDLIBXC, GGA, METAGGA, LHFCALC, AEXX, AGGAX, AGGAC, ALDAC, hybrid functionals, settings for specific hybrid functionals
- ↑ M. A. L. Marques, M. J. T. Oliveira, and T. Burnus, Comput. Phys. Commun., 183, 2272 (2012).
- ↑ S. Lehtola, C. Steigemann, M. J. T. Oliveira, and M. A. L. Marques, SoftwareX, 7, 1 (2018).
- ↑ a b c https://libxc.gitlab.io
- ↑ J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev. Lett., 77, 3865 (1996).
- ↑ J. Sun, A. Ruzsinszky, and J. P. Perdew, Phys. Rev. Lett. 115, 036402 (2015).
- ↑ C. Adamo and V. Barone, Phys. Rev. Lett., 110, 6158 (1999).