Nonlocal vdW-DF functionals - Revision history

Liebetreu: Reverted edits by Liebetreu (talk) to last revision by Ftran

2026-04-15T12:36:34Z

Reverted edits by Liebetreu (talk) to last revision by Ftran

← Older revision		Revision as of 12:36, 15 April 2026
Line 141:		Line 141:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*'''~~r²SCAN~~+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r<math>^2</math>SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 151:		Line 151:

	=== Kernel file vdw_kernel.bindat ===		=== Kernel file vdw_kernel.bindat ===
	*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and ~~r²SCAN~~+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.		*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and r<math>^2</math>SCAN+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.
	*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.		*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.

Liebetreu: /* List of nonlocal vdW-DF functionals */

2026-04-14T11:53:40Z

List of nonlocal vdW-DF functionals

← Older revision		Revision as of 11:53, 14 April 2026
Line 141:		Line 141:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*'''~~r<math>^2</math>SCAN~~+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r²SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.

Liebetreu: /* Kernel file vdw_kernel.bindat */

2026-04-14T11:53:14Z

Kernel file vdw_kernel.bindat

← Older revision		Revision as of 11:53, 14 April 2026
Line 151:		Line 151:

	=== Kernel file vdw_kernel.bindat ===		=== Kernel file vdw_kernel.bindat ===
	*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and ~~r<math>^2</math>SCAN~~+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.		*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and r²SCAN+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.
	*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.		*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.

Ftran at 15:42, 28 February 2025

2025-02-28T15:42:01Z

← Older revision		Revision as of 15:42, 28 February 2025
Line 19:		Line 19:
	{{NB\|important\|Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}		{{NB\|important\|Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
	{{NB\|mind\|		{{NB\|mind\|
	*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).		*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given [[:Nonlocal_vdW-DF_functionals#Kernel file vdw_kernel.bindat\|below]]. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).
	*In VASP.6.2 (and prior versions) the stress tensor is broken for rVV10 (it is correct for other vdW-DF though). From VASP.6.3.0 onwards, the stress tensor for rVV10 is correct.}}		*In VASP.6.2 (and prior versions) the stress tensor is broken for rVV10 (it is correct for other vdW-DF though). From VASP.6.3.0 onwards, the stress tensor for rVV10 is correct.}}

Ftran at 15:36, 28 February 2025

2025-02-28T15:36:43Z

← Older revision		Revision as of 15:36, 28 February 2025
Line 17:		Line 17:
	An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.		An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.

	{{NB\|important\|		{{NB\|important\|Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
	*Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
	{{NB\|mind\|		{{NB\|mind\|
	*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).		*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).

Ftran at 15:36, 28 February 2025

2025-02-28T15:36:16Z

← Older revision		Revision as of 15:36, 28 February 2025
Line 9:		Line 9:
	d^{3}rd^{3}r',		d^{3}rd^{3}r',
	</math>		</math>
	with ~~the~~ kernel <math>\Phi</math> that depends on the electronic density <math>n</math>, its derivative <math>\nabla n</math> as well as on the interelectronic distance <math>\left\vert\bf{r}-\bf{r}'\right\vert</math>. In VASP, the ~~nonlocal functional~~ is ~~implemented~~ using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).		with a kernel <math>\Phi</math> that depends on the electronic density <math>n</math>, its derivative <math>\nabla n</math> as well as on the interelectronic distance <math>\left\vert\bf{r}-\bf{r}'\right\vert</math>. In VASP, the calculation of <math>E_{\text{c,disp}}</math> is done using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).

	The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.		The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.
Line 18:		Line 18:

	{{NB\|important\|		{{NB\|important\|
	Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}		*Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
			{{NB\|mind\|
			*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).
			*In VASP.6.2 (and prior versions) the stress tensor is broken for rVV10 (it is correct for other vdW-DF though). From VASP.6.3.0 onwards, the stress tensor for rVV10 is correct.}}

	== List of nonlocal vdW-DF functionals ==		== List of nonlocal vdW-DF functionals ==

	*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. (and optionally {{TAG\|IVDW_NL}}) in the {{FILE\|INCAR}} file.		*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal or hybrid exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. (and optionally {{TAG\|IVDW_NL}}) in the {{FILE\|INCAR}} file.

	*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.		*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.


	~~{{NB\|important\|~~
	*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).}}

	Examples of {{FILE\|INCAR}} files are shown below.		Examples of {{FILE\|INCAR}} files are shown below.
Line 162:		Line 160:
	=== Computational time ===		=== Computational time ===
	*The evaluation of the nonlocal correlation energy requires some additional time. Most of it is spent on performing FFTs to evaluate the energy and potential. Thus the additional time is determined by the number of FFT grid points, basically the size of the simulation cell. It is almost independent on the number of the atoms in the cell, but increases with the amount of vacuum in the cell. The relative increase is high for isolated molecules in large cells, but small for solids in smaller cells with many k-points.		*The evaluation of the nonlocal correlation energy requires some additional time. Most of it is spent on performing FFTs to evaluate the energy and potential. Thus the additional time is determined by the number of FFT grid points, basically the size of the simulation cell. It is almost independent on the number of the atoms in the cell, but increases with the amount of vacuum in the cell. The relative increase is high for isolated molecules in large cells, but small for solids in smaller cells with many k-points.

	~~{{NB\|mind\|~~
	*In VASP.6.2 (and prior versions) the stress tensor is broken for rVV10 (it is correct for other vdW-DF though). From VASP.6.3.0 onwards, the stress tensor for rVV10 is correct.}}

	== Related Tags and Sections ==		== Related Tags and Sections ==

Ftran: /* Important technical remarks */

2025-02-28T15:04:53Z

Important technical remarks

← Older revision		Revision as of 15:04, 28 February 2025
Line 153:		Line 153:
	== Important technical remarks ==		== Important technical remarks ==

	*Kernel file vdw_kernel.bindat:		=== Kernel file vdw_kernel.bindat ===
	**'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and r<math>^2</math>SCAN+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.		*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and r<math>^2</math>SCAN+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.
	**'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.		*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.

			=== POTCAR file ===
	*There are no special {{TAG\|POTCAR}} files for the vdW-DF functionals and the PBE or LDA {{TAG\|POTCAR}} files can be used. Currently the evaluation of the nonlocal correlation functional is not done fully within the PAW method, but the sum of the pseudo-valence density and partial core density is used. This approximation works rather well, as is discussed in {{cite\|klimes:prb:2011}}, and the accuracy generally increases when the number of valence electrons is increased or when harder PAW datasets are used. For example, for adsorption it is recommended to compare the adsorption energy obtained with standard PAW datasets and more-electron {{TAG\|POTCAR}} files for both PBE calculations and vdW-DF calculations to assess the quality of the results.		*There are no special {{TAG\|POTCAR}} files for the vdW-DF functionals and the PBE or LDA {{TAG\|POTCAR}} files can be used. Currently the evaluation of the nonlocal correlation functional is not done fully within the PAW method, but the sum of the pseudo-valence density and partial core density is used. This approximation works rather well, as is discussed in {{cite\|klimes:prb:2011}}, and the accuracy generally increases when the number of valence electrons is increased or when harder PAW datasets are used. For example, for adsorption it is recommended to compare the adsorption energy obtained with standard PAW datasets and more-electron {{TAG\|POTCAR}} files for both PBE calculations and vdW-DF calculations to assess the quality of the results.

			=== Computational time ===
	*The evaluation of the nonlocal correlation energy requires some additional time. Most of it is spent on performing FFTs to evaluate the energy and potential. Thus the additional time is determined by the number of FFT grid points, basically the size of the simulation cell. It is almost independent on the number of the atoms in the cell, but increases with the amount of vacuum in the cell. The relative increase is high for isolated molecules in large cells, but small for solids in smaller cells with many k-points.		*The evaluation of the nonlocal correlation energy requires some additional time. Most of it is spent on performing FFTs to evaluate the energy and potential. Thus the additional time is determined by the number of FFT grid points, basically the size of the simulation cell. It is almost independent on the number of the atoms in the cell, but increases with the amount of vacuum in the cell. The relative increase is high for isolated molecules in large cells, but small for solids in smaller cells with many k-points.

Ftran at 14:54, 28 February 2025

2025-02-28T14:54:38Z

← Older revision		Revision as of 14:54, 28 February 2025
Line 1:		Line 1:
	The vdW-DF method originally proposed by Dion ''et al.''{{cite\|dion:prl:2004}} consists of a semilocal exchange-correlation functional that is augmented with a nonlocal correlation functional <math>E_{\text{c,disp}}</math> that approximately accounts for dispersion interactions. In VASP, the nonlocal functional is implemented using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).		The vdW-DF method originally proposed by Dion ''et al.''{{cite\|dion:prl:2004}} consists of a semilocal or hybrid exchange-correlation functional <math>E_{\text{xc}}^{\text{SL/hybrid}}</math> that is augmented with a nonlocal correlation functional <math>E_{\text{c,disp}}</math> that approximately accounts for dispersion interactions:
			:<math>
			E_{\text{xc}}^{\text{vdW}} = E_{\text{xc}}^{\text{SL/hybrid}} + E_{\text{c,disp}},
			</math>
			where
			:<math>
			E_{\text{c,disp}} = \frac{1}{2}\int\int n(\textbf{r})
			\Phi\left(\textbf{r},\textbf{r}'\right) n(\textbf{r}')
			d^{3}rd^{3}r',
			</math>
			with the kernel <math>\Phi</math> that depends on the electronic density <math>n</math>, its derivative <math>\nabla n</math> as well as on the interelectronic distance <math>\left\vert\bf{r}-\bf{r}'\right\vert</math>. In VASP, the nonlocal functional is implemented using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).

	The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.		The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.
Line 8:		Line 18:

	{{NB\|important\|		{{NB\|important\|
	Some ~~non-local~~ vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}		Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}


	== List of nonlocal vdW-DF functionals ==		== List of nonlocal vdW-DF functionals ==

	*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. in the {{FILE\|INCAR}} file.		*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. (and optionally {{TAG\|IVDW_NL}}) in the {{FILE\|INCAR}} file.

	*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.		*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.
Line 23:		Line 33:
	Examples of {{FILE\|INCAR}} files are shown below.		Examples of {{FILE\|INCAR}} files are shown below.

	* ~~Original~~ vdW-DF of Dion ''et al.''{{cite\|dion:prl:2004}}:		*'''vdW-DF''' of Dion ''et al.''{{cite\|dion:prl:2004}}:
	{{TAGBL\|GGA}} = RE		{{TAGBL\|GGA}} = RE
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 29:		Line 39:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	* vdW-DF2 of Lee ''et al.'' (2nd version of vdW-DF){{cite\|lee:prb:2010}}:		*'''vdW-DF2''' of Lee ''et al.'' (2nd version of vdW-DF){{cite\|lee:prb:2010}}:
	{{TAGBL\|GGA}} = ML		{{TAGBL\|GGA}} = ML
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 36:		Line 46:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optPBE-vdW of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:		*'''optPBE-vdW''' of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:
	{{TAGBL\|GGA}} = OR		{{TAGBL\|GGA}} = OR
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 42:		Line 52:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optB88-vdW of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:		*'''optB88-vdW''' of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:
	{{TAGBL\|GGA}} = BO		{{TAGBL\|GGA}} = BO
	{{TAGBL\|PARAM1}} = 0.1833333333		{{TAGBL\|PARAM1}} = 0.1833333333
Line 50:		Line 60:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optB86b-vdW of Klimeš ''et al.''{{cite\|klimes:prb:2011}}:		*'''optB86b-vdW''' of Klimeš ''et al.''{{cite\|klimes:prb:2011}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234		{{TAGBL\|PARAM1}} = 0.1234
Line 58:		Line 68:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*BEEF-vdW of Wellendorff ''et al.''{{cite\|beef2012}}:		*'''BEEF-vdW''' of Wellendorff ''et al.''{{cite\|beef2012}}:
	{{TAGBL\|GGA}} = BF		{{TAGBL\|GGA}} = BF
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 72:		Line 82:
	Note that the GGA functional BEEF{{cite\|beef2012}} is available only via an external library, either libbeef ([[Precompiler_options#-Dlibbeef\|-Dlibbeef]]) or Libxc ([[Precompiler_options#-DUSELIBXC\|-DUSELIBXC]]).		Note that the GGA functional BEEF{{cite\|beef2012}} is available only via an external library, either libbeef ([[Precompiler_options#-Dlibbeef\|-Dlibbeef]]) or Libxc ([[Precompiler_options#-DUSELIBXC\|-DUSELIBXC]]).

	*rev-vdW-DF2 (also known as vdW-DF2-B86R) of Hamada{{cite\|hamada:prb:14}}:		*'''rev-vdW-DF2''' (also known as vdW-DF2-B86R) of Hamada{{cite\|hamada:prb:14}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234568 # =10/81		{{TAGBL\|PARAM1}} = 0.1234568 # =10/81
Line 83:		Line 93:
	In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.		In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.

	*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:2014}}:		*'''vdW-DF-cx''' of Berland and Hyldgaard{{cite\|berland:prb:2014}}:
	{{TAGBL\|GGA}} = CX		{{TAGBL\|GGA}} = CX
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 89:		Line 99:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*vdW-DF3-opt1 of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:		*'''vdW-DF3-opt1''' of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:
	{{TAGBL\|GGA}} = BO		{{TAGBL\|GGA}} = BO
	{{TAGBL\|PARAM1}} = 0.1122334456		{{TAGBL\|PARAM1}} = 0.1122334456
Line 100:		Line 110:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*vdW-DF3-opt2 of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:		*'''vdW-DF3-opt2''' of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234568 # =10/81		{{TAGBL\|PARAM1}} = 0.1234568 # =10/81
Line 112:		Line 122:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*rVV10 of Sabatini ''et al.'' {{cite\|sabatini:prb:2013}}:		*'''rVV10''' of Sabatini ''et al.'' {{cite\|sabatini:prb:2013}}:
	{{TAGBL\|GGA}} = ML		{{TAGBL\|GGA}} = ML
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 120:		Line 130:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*SCAN+rVV10 of Peng ''et al.'' {{cite\|peng:prx:2016}}:		*'''SCAN+rVV10''' of Peng ''et al.'' {{cite\|peng:prx:2016}}:
	{{TAGBL\|METAGGA}} = SCAN		{{TAGBL\|METAGGA}} = SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 127:		Line 137:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*PBE+rVV10L of Peng and Perdew {{cite\|peng:prb:2017}}:		*'''PBE+rVV10L''' of Peng and Perdew {{cite\|peng:prb:2017}}:
	{{TAGBL\|GGA}} = PE		{{TAGBL\|GGA}} = PE
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 134:		Line 144:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*r<math>^2</math>SCAN+rVV10 of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r<math>^2</math>SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.

Kresse: Recommendation to use PREC=Accurate in combination with Conjugate gradient.

2025-02-27T20:33:28Z

Recommendation to use PREC=Accurate in combination with Conjugate gradient.

← Older revision		Revision as of 20:33, 27 February 2025
Line 6:		Line 6:

	An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.		An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.

			{{NB\|important\|
			Some non-local vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}


	== List of nonlocal vdW-DF functionals ==		== List of nonlocal vdW-DF functionals ==
Line 12:		Line 16:

	*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.		*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.


	{{NB\|important\|		{{NB\|important\|

Ftran at 15:57, 14 October 2024

2024-10-14T15:57:34Z

← Older revision		Revision as of 15:57, 14 October 2024
Line 78:		Line 78:
	In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.		In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.

	*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:~~2015~~}}:		*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:2014}}:
	{{TAGBL\|GGA}} = CX		{{TAGBL\|GGA}} = CX
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0

← Older revision		Revision as of 12:36, 15 April 2026
Line 141:		Line 141:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*'''~~r²SCAN~~+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r<math>^2</math>SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 151:		Line 151:

	=== Kernel file vdw_kernel.bindat ===		=== Kernel file vdw_kernel.bindat ===
	*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and ~~r²SCAN~~+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.		*'''Until VASP.6.4.2''': The calculation of the nonlocal correlation functional of Dion ''et al.'' (used when {{TAG\|IVDW_NL}}=1, which means for all functionals listed above except rVV10, SCAN+rVV10 and r<math>^2</math>SCAN+rVV10) requires a precalculated kernel which is distributed via the VASP download portal (vdw_kernel.bindat.gz has to be decompressed). If VASP does not find this file, the kernel is calculated, which is however extremely demanding (several hours!). Thus, the kernel needs to be either copied to the VASP run directory for each calculation or can be stored in a central location and read from there. The location needs to be set in routine ''PHI_GENERATE''. This does not work on some clusters and the kernel needs to be copied into the working directory in such cases. The distributed file uses little endian convention and can not be read on big endian machines. The big endian version of the file is available also from the VASP portal. In the case of the rVV10 nonlocal correlation functional, no precalculated kernel is required and it is calculated on the fly, which is however much less demanding than in the case of the functional of Dion ''et al.''.
	*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.		*'''Since VASP.6.4.3''': The calculation of the kernel for the functional of Dion ''et al.'' ({{TAG\|IVDW_NL}}=1), as well as for {{TAG\|IVDW_NL}}=3 and 4, is tremendously faster: the default value of a parameter has been reduced (with basically no loss of accuracy) and the calculation has been parallelized (with MPI and OpenACC for GPUs). Therefore, starting a calculation without vdw_kernel.bindat file present in the directory should be no problem for the computational time, and a vdw_kernel.bindat file will be generated rather efficiently during the first iteration. Note that a file vdw_kernel.bindat that was generated for a given kernel ({{TAG\|IVDW_NL}}=1, 3 or 4) can not be used for a calculation using another kernel, and in such a case the incompatibility of the vdw_kernel.bindat file will be detected and a new vdw_kernel.bindat file automatically generated to replace the incompatible one.

← Older revision		Revision as of 11:53, 14 April 2026
Line 141:		Line 141:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*'''~~r<math>^2</math>SCAN~~+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r²SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.

← Older revision		Revision as of 15:36, 28 February 2025
Line 17:		Line 17:
	An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.		An overview of the performance of the vdW-DF functionals can be found for instance in Ref. {{cite\|klimes:prb:2011}}{{cite\|berland:rpp:2015}}{{cite\|tran:prm:19}}.

	{{NB\|important\|		{{NB\|important\|Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
	*Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate.}}
	{{NB\|mind\|		{{NB\|mind\|
	*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).		*For '''VASP.6.4.2''' and prior versions it was necessary to copy the vdw_kernel.bindat file into the working directory for calculations with the van der Waals kernel corresponding to {{TAG\|IVDW_NL}}{{=}}1. Otherwise, vdw_kernel.bindat was generated at the beginning of the calculation, which took several hours. However, since '''VASP.6.4.3''' it is not really necessary to copy vdw_kernel.bindat into the directory, since its calculation has been considerably accelerated (about 2 minutes with 8 MPI ranks). More details are given below. Note that no vdw_kernel.bindat file is needed for calculations with the rVV10 kernel ({{TAG\|IVDW_NL}}{{=}}2).

← Older revision		Revision as of 14:54, 28 February 2025
Line 1:		Line 1:
	The vdW-DF method originally proposed by Dion ''et al.''{{cite\|dion:prl:2004}} consists of a semilocal exchange-correlation functional that is augmented with a nonlocal correlation functional <math>E_{\text{c,disp}}</math> that approximately accounts for dispersion interactions. In VASP, the nonlocal functional is implemented using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).		The vdW-DF method originally proposed by Dion ''et al.''{{cite\|dion:prl:2004}} consists of a semilocal or hybrid exchange-correlation functional <math>E_{\text{xc}}^{\text{SL/hybrid}}</math> that is augmented with a nonlocal correlation functional <math>E_{\text{c,disp}}</math> that approximately accounts for dispersion interactions:
			:<math>
			E_{\text{xc}}^{\text{vdW}} = E_{\text{xc}}^{\text{SL/hybrid}} + E_{\text{c,disp}},
			</math>
			where
			:<math>
			E_{\text{c,disp}} = \frac{1}{2}\int\int n(\textbf{r})
			\Phi\left(\textbf{r},\textbf{r}'\right) n(\textbf{r}')
			d^{3}rd^{3}r',
			</math>
			with the kernel <math>\Phi</math> that depends on the electronic density <math>n</math>, its derivative <math>\nabla n</math> as well as on the interelectronic distance <math>\left\vert\bf{r}-\bf{r}'\right\vert</math>. In VASP, the nonlocal functional is implemented using the algorithm of Román-Pérez and Soler{{cite\|romanperez:prl:09}} that is based on FFTs and the convolution theorem to calculate efficiently the double real-space integral. Several versions of the vdW-DF functionals proposed in the literature can be used (see list below).

	The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.		The vdW-DF functionals are available since the 5.2.12.26May2011 version of VASP for the calculation of total energies and forces. The stress tensor calculation for the cell optimization ({{TAG\|ISIF}}=3) is available since the VASP 5.2.12.11Nov2011 version for spin-unpolarized systems and VASP 5.3.1 for spin-polarized systems. They have been implemented by J. Klimeš. If you make use of the vdW-DF functionals presented in this section, we ask you to cite Ref. {{cite\|klimes:prb:2011}}. Please also cite the original vdW-DF paper of Dion ''et al.''{{cite\|dion:prl:2004}} and the paper of Román-Pérez and Soler{{cite\|romanperez:prl:09}}.
Line 8:		Line 18:

	{{NB\|important\|		{{NB\|important\|
	Some ~~non-local~~ vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}		Some nonlocal vdW-DF result in very noisy energies, which can degrade the convergence to the electronic groundstate. Conjugate-gradient algorithms are particularly prone to show issues, such as a sudden increase in the energy. If {{TAG \| ALGO}} {{=}} all (conjugate gradient algorithm) fails to converge, try to use denser FFT grids, for instance by setting {{TAG \|PREC}} {{=}} Accurate. }}


	== List of nonlocal vdW-DF functionals ==		== List of nonlocal vdW-DF functionals ==

	*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. in the {{FILE\|INCAR}} file.		*To add a nonlocal correlation energy <math>E_{\text{c,disp}}</math> to the semilocal exchange-correlation energy (selected with the {{TAG\|GGA}}, {{TAG\|METAGGA}} or {{TAG\|XC}} tag) one needs to set {{TAG\|LUSE_VDW}}=.TRUE. (and optionally {{TAG\|IVDW_NL}}) in the {{FILE\|INCAR}} file.

	*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.		*Since vdW-DF functionals tend to yield less spherical densities than standard GGA functionals, it is recommended to set {{TAG\|LASPH}}=.TRUE. to get reasonably accurate contributions from the spheres around the atoms.
Line 23:		Line 33:
	Examples of {{FILE\|INCAR}} files are shown below.		Examples of {{FILE\|INCAR}} files are shown below.

	* ~~Original~~ vdW-DF of Dion ''et al.''{{cite\|dion:prl:2004}}:		*'''vdW-DF''' of Dion ''et al.''{{cite\|dion:prl:2004}}:
	{{TAGBL\|GGA}} = RE		{{TAGBL\|GGA}} = RE
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 29:		Line 39:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	* vdW-DF2 of Lee ''et al.'' (2nd version of vdW-DF){{cite\|lee:prb:2010}}:		*'''vdW-DF2''' of Lee ''et al.'' (2nd version of vdW-DF){{cite\|lee:prb:2010}}:
	{{TAGBL\|GGA}} = ML		{{TAGBL\|GGA}} = ML
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 36:		Line 46:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optPBE-vdW of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:		*'''optPBE-vdW''' of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:
	{{TAGBL\|GGA}} = OR		{{TAGBL\|GGA}} = OR
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 42:		Line 52:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optB88-vdW of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:		*'''optB88-vdW''' of Klimeš ''et al.''{{cite\|klimes:jpcm:2010}}:
	{{TAGBL\|GGA}} = BO		{{TAGBL\|GGA}} = BO
	{{TAGBL\|PARAM1}} = 0.1833333333		{{TAGBL\|PARAM1}} = 0.1833333333
Line 50:		Line 60:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*optB86b-vdW of Klimeš ''et al.''{{cite\|klimes:prb:2011}}:		*'''optB86b-vdW''' of Klimeš ''et al.''{{cite\|klimes:prb:2011}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234		{{TAGBL\|PARAM1}} = 0.1234
Line 58:		Line 68:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*BEEF-vdW of Wellendorff ''et al.''{{cite\|beef2012}}:		*'''BEEF-vdW''' of Wellendorff ''et al.''{{cite\|beef2012}}:
	{{TAGBL\|GGA}} = BF		{{TAGBL\|GGA}} = BF
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 72:		Line 82:
	Note that the GGA functional BEEF{{cite\|beef2012}} is available only via an external library, either libbeef ([[Precompiler_options#-Dlibbeef\|-Dlibbeef]]) or Libxc ([[Precompiler_options#-DUSELIBXC\|-DUSELIBXC]]).		Note that the GGA functional BEEF{{cite\|beef2012}} is available only via an external library, either libbeef ([[Precompiler_options#-Dlibbeef\|-Dlibbeef]]) or Libxc ([[Precompiler_options#-DUSELIBXC\|-DUSELIBXC]]).

	*rev-vdW-DF2 (also known as vdW-DF2-B86R) of Hamada{{cite\|hamada:prb:14}}:		*'''rev-vdW-DF2''' (also known as vdW-DF2-B86R) of Hamada{{cite\|hamada:prb:14}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234568 # =10/81		{{TAGBL\|PARAM1}} = 0.1234568 # =10/81
Line 83:		Line 93:
	In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.		In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.

	*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:2014}}:		*'''vdW-DF-cx''' of Berland and Hyldgaard{{cite\|berland:prb:2014}}:
	{{TAGBL\|GGA}} = CX		{{TAGBL\|GGA}} = CX
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0
Line 89:		Line 99:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*vdW-DF3-opt1 of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:		*'''vdW-DF3-opt1''' of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:
	{{TAGBL\|GGA}} = BO		{{TAGBL\|GGA}} = BO
	{{TAGBL\|PARAM1}} = 0.1122334456		{{TAGBL\|PARAM1}} = 0.1122334456
Line 100:		Line 110:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*vdW-DF3-opt2 of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:		*'''vdW-DF3-opt2''' of Chakraborty ''et al.'' {{cite\|chakraborty:jctc:2020}}:
	{{TAGBL\|GGA}} = MK		{{TAGBL\|GGA}} = MK
	{{TAGBL\|PARAM1}} = 0.1234568 # =10/81		{{TAGBL\|PARAM1}} = 0.1234568 # =10/81
Line 112:		Line 122:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*rVV10 of Sabatini ''et al.'' {{cite\|sabatini:prb:2013}}:		*'''rVV10''' of Sabatini ''et al.'' {{cite\|sabatini:prb:2013}}:
	{{TAGBL\|GGA}} = ML		{{TAGBL\|GGA}} = ML
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 120:		Line 130:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*SCAN+rVV10 of Peng ''et al.'' {{cite\|peng:prx:2016}}:		*'''SCAN+rVV10''' of Peng ''et al.'' {{cite\|peng:prx:2016}}:
	{{TAGBL\|METAGGA}} = SCAN		{{TAGBL\|METAGGA}} = SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 127:		Line 137:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*PBE+rVV10L of Peng and Perdew {{cite\|peng:prb:2017}}:		*'''PBE+rVV10L''' of Peng and Perdew {{cite\|peng:prb:2017}}:
	{{TAGBL\|GGA}} = PE		{{TAGBL\|GGA}} = PE
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.
Line 134:		Line 144:
	{{TAGBL\|LASPH}} = .TRUE.		{{TAGBL\|LASPH}} = .TRUE.

	*r<math>^2</math>SCAN+rVV10 of Ning ''et al.'' {{cite\|ning:prb:2022}}:		*'''r<math>^2</math>SCAN+rVV10''' of Ning ''et al.'' {{cite\|ning:prb:2022}}:
	{{TAGBL\|METAGGA}} = R2SCAN		{{TAGBL\|METAGGA}} = R2SCAN
	{{TAGBL\|LUSE_VDW}} = .TRUE.		{{TAGBL\|LUSE_VDW}} = .TRUE.

← Older revision		Revision as of 15:57, 14 October 2024
Line 78:		Line 78:
	In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.		In the vdW-DF2, BEEF-vdW and rev-vdW-DF2 functionals, the nonlocal correlation consists of the Dion ''et al.'' functional, but with the parameter <math>Z_{ab}</math> that is changed from -0.8491 (the default value in VASP) to -1.8867 by setting {{TAG\|ZAB_VDW}}=-1.8867.

	*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:~~2015~~}}:		*vdW-DF-cx of Berland and Hyldgaard{{cite\|berland:prb:2014}}:
	{{TAGBL\|GGA}} = CX		{{TAGBL\|GGA}} = CX
	{{TAGBL\|AGGAC}} = 0.0		{{TAGBL\|AGGAC}} = 0.0