ACFDT/RPA calculations - Revision history

Huebsch: Fix Low-scaling GW and RPA link: use piped category link to avoid Category: prefix in display

2026-03-19T08:43:34Z

Fix Low-scaling GW and RPA link: use piped category link to avoid Category: prefix in display

← Older revision		Revision as of 08:43, 19 March 2026
Line 278:		Line 278:
	* [[RPA/ACFDT: Correlation energy in the Random Phase Approximation]] — theory background		* [[RPA/ACFDT: Correlation energy in the Random Phase Approximation]] — theory background
	* [[Many-body perturbation theory]] — parent topic (RPA, GW, BSE, MP2)		* [[Many-body perturbation theory]] — parent topic (RPA, GW, BSE, MP2)
	* [[Low-scaling GW and RPA]] — low-scaling algorithms		* [[:Category:Low-scaling GW and RPA\|Low-scaling GW and RPA]] — low-scaling algorithms
	* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations		* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations
	* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points		* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points

Huebsch: Fix link: use redirect Low-scaling GW and RPA instead of :Category:...

2026-03-19T08:42:30Z

Fix link: use redirect Low-scaling GW and RPA instead of Category:...

← Older revision		Revision as of 08:42, 19 March 2026
Line 278:		Line 278:
	* [[RPA/ACFDT: Correlation energy in the Random Phase Approximation]] — theory background		* [[RPA/ACFDT: Correlation energy in the Random Phase Approximation]] — theory background
	* [[Many-body perturbation theory]] — parent topic (RPA, GW, BSE, MP2)		* [[Many-body perturbation theory]] — parent topic (RPA, GW, BSE, MP2)
	* [[~~:Category:~~Low-scaling GW and RPA]] — low-scaling algorithms		* [[Low-scaling GW and RPA]] — low-scaling algorithms
	* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations		* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations
	* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points		* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points

Huebsch: Rename section to Related tags and articles; add cross-links to theory page, Many-body perturbation theory, Low-scaling GW and RPA

2026-03-19T08:38:29Z

Rename section to Related tags and articles; add cross-links to theory page, Many-body perturbation theory, Low-scaling GW and RPA

← Older revision		Revision as of 08:38, 19 March 2026
Line 274:		Line 274:
	</span>		</span>

	== Related ~~Tags~~ and ~~Sections~~ ==		== Related tags and articles ==

			* [[RPA/ACFDT: Correlation energy in the Random Phase Approximation]] — theory background
			* [[Many-body perturbation theory]] — parent topic (RPA, GW, BSE, MP2)
			* [[:Category:Low-scaling GW and RPA]] — low-scaling algorithms
	* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations		* {{TAG\|ALGO}} for response functions and ''ACFDT'' calculations
	* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points		* {{TAG\|NOMEGA}}, {{TAG\|NOMEGAR}} number of frequency points

Csheldon: /* Exact one-centre density terms */

2026-01-29T09:03:23Z

Exact one-centre density terms

← Older revision		Revision as of 09:03, 29 January 2026
Line 222:		Line 222:

	=== Exact one-centre density terms ===		=== Exact one-centre density terms ===
	Beginning with version 6.5.2, it is possible to include exact one-center terms in the density for the exact exchange (EXX) component of RPA total energies and forces by enabling the {{TAG\|LFOCKSTD}} option. Incorporating these one-center terms significantly enhances the accuracy of both total energies and forces-by at least an order of magnitude-compared to results obtained using finite-difference methods. Therefore, it is strongly recommended to activate {{TAG\|LFOCKSTD}} for GW and RPA calculations starting from version 6.5.2. In earlier versions, the one-center contribution to the density was managed using the {{TAG\|LMAXFOCKAE}} setting.		Beginning with version 6.6.0, it is possible to include exact one-center terms in the density for the exact exchange (EXX) component of RPA total energies and forces by enabling the {{TAG\|LFOCKSTD}} option. Incorporating these one-center terms significantly enhances the accuracy of both total energies and forces-by at least an order of magnitude-compared to results obtained using finite-difference methods. Therefore, it is strongly recommended to activate {{TAG\|LFOCKSTD}} for GW and RPA calculations starting from version 6.6.0. In earlier versions, the one-center contribution to the density was managed using the {{TAG\|LMAXFOCKAE}} setting.

	=== Basis set convergence ===		=== Basis set convergence ===

Csheldon: /* Optional: RPA Forces */

2026-01-29T09:03:10Z

Optional: RPA Forces

← Older revision		Revision as of 09:03, 29 January 2026
Line 171:		Line 171:

	The suggested updated positions are also written to the CONTCAR file (if {{TAG\|NSW}}=0). The updated positions are obtained by multiplying the RPA-forces with the parameter {{TAG\|DAMP_NEWTON}} and the inverse of the DFT-Hessian (which is also calculated during the RPA force calculations) and adding the resultant vector to the current positions. {{TAG\|DAMP_NEWTON}} currently defaults to 0.8, but the user might want to change this to 1.0 (no damping), or smaller values if instabilities are observed. The CONTCAR file can be copied to the POSCAR file, and a few such RPA calculations should recover the groundstate structure. Alternatively, standard relaxations ({{TAG\|IBRION}} =1-4) or even MD's can be performed.		The suggested updated positions are also written to the CONTCAR file (if {{TAG\|NSW}}=0). The updated positions are obtained by multiplying the RPA-forces with the parameter {{TAG\|DAMP_NEWTON}} and the inverse of the DFT-Hessian (which is also calculated during the RPA force calculations) and adding the resultant vector to the current positions. {{TAG\|DAMP_NEWTON}} currently defaults to 0.8, but the user might want to change this to 1.0 (no damping), or smaller values if instabilities are observed. The CONTCAR file can be copied to the POSCAR file, and a few such RPA calculations should recover the groundstate structure. Alternatively, standard relaxations ({{TAG\|IBRION}} =1-4) or even MD's can be performed.
	{{NB\|tip\|Use {{TAG\|LFOCKSTD}} to improve total energies and RPA forces as of version 6.5.2.}}		{{NB\|tip\|Use {{TAG\|LFOCKSTD}} to improve total energies and RPA forces as of version 6.6.0.}}
	</span>		</span>

Csheldon: /* Output of low-scaling ACFDT/RPA */

2026-01-29T09:03:01Z

Output of low-scaling ACFDT/RPA

← Older revision		Revision as of 09:03, 29 January 2026
Line 134:		Line 134:

	The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density		The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density
	using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.		using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.6.0, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.
	{{NB\|tip\|Select exact one-centre terms for the electronic density in [[#EXX-inline-ref\|exact exchange energies]] with {{TAG\|LFOCKSTD}} as of version 6.5.2.}}		{{NB\|tip\|Select exact one-centre terms for the electronic density in [[#EXX-inline-ref\|exact exchange energies]] with {{TAG\|LFOCKSTD}} as of version 6.6.0.}}

	=== Singles contribution to the correlation energy ===		=== Singles contribution to the correlation energy ===

Kaltakm: /* Output of low-scaling ACFDT/RPA */

2025-05-02T12:40:39Z

Output of low-scaling ACFDT/RPA

← Older revision		Revision as of 12:40, 2 May 2025
Line 135:		Line 135:
	The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density		The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density
	using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.		using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.
	{{NB\|tip\|~~Improve agreement of <code>FHF</code> and~~ [[#EXX-inline-ref\|exact exchange energies]] ~~by selecting exact one-centre terms for the electronic density~~ with {{TAG\|LFOCKSTD}}.}}		{{NB\|tip\|Select exact one-centre terms for the electronic density in [[#EXX-inline-ref\|exact exchange energies]] with {{TAG\|LFOCKSTD}} as of version 6.5.2.}}

	=== Singles contribution to the correlation energy ===		=== Singles contribution to the correlation energy ===

Kaltakm: /* Output of low-scaling ACFDT/RPA */

2025-05-02T12:39:32Z

Output of low-scaling ACFDT/RPA

← Older revision		Revision as of 12:39, 2 May 2025
Line 135:		Line 135:
	The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density		The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density
	using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.		using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.
			{{NB\|tip\|Improve agreement of <code>FHF</code> and [[#EXX-inline-ref\|exact exchange energies]] by selecting exact one-centre terms for the electronic density with {{TAG\|LFOCKSTD}}.}}

	=== Singles contribution to the correlation energy ===		=== Singles contribution to the correlation energy ===

Kaltakm: /* Output of low-scaling ACFDT/RPA */

2025-05-02T12:37:16Z

Output of low-scaling ACFDT/RPA

← Older revision		Revision as of 12:37, 2 May 2025
Line 132:		Line 132:
	The line <code>HF+RPA corr. energy TOTEN</code> contains the total energy calculated with the largest cutoff {{TAG\|ENCUTGW}}.		The line <code>HF+RPA corr. energy TOTEN</code> contains the total energy calculated with the largest cutoff {{TAG\|ENCUTGW}}.
	The line <code>HF+E_corr(extrapolated)</code> contains the total energy with the extrapolated value for the RPA correlation energy.		The line <code>HF+E_corr(extrapolated)</code> contains the total energy with the extrapolated value for the RPA correlation energy.

			The line <code>FHF</code> denotes the exact exchange contribution to the total energy. This contribution is determined with an electronic density
			using {{TAG\|LMAXFOCKAE}}. For most systems the resulting energy is in very good agreement with the reference EXX energy obtained in [[#EXX-inline-ref\|step 2 described above]]. As of version 6.5.2, exact compatibility with this step can be reached using {{TAG\|LFOCKSTD}}.

	=== Singles contribution to the correlation energy ===		=== Singles contribution to the correlation energy ===
	The low-scaling RPA algorithm also allows for the determination of the so-called singles contribution,{{cite\|klimes:jcp:143}}{{cite\|ren:prb:88}} to the total energy represented by following diagrams:		The low-scaling RPA algorithm also allows for the determination of the so-called singles contribution,{{cite\|klimes:jcp:143}}{{cite\|ren:prb:88}} to the total energy represented by following diagrams:

Kaltakm: /* General recipe to calculate ACFDT-RPA total energies */

2025-05-02T12:22:50Z

General recipe to calculate ACFDT-RPA total energies

← Older revision		Revision as of 12:22, 2 May 2025
Line 28:		Line 28:
	{{NB\|tip\|Use the Perdew-Burke-Ernzerhof (PBE) pseudopotentials with a small {{TAG\|EDIFF}} and avoid {{TAG\|ISMEAR}}>0.\|:}}		{{NB\|tip\|Use the Perdew-Burke-Ernzerhof (PBE) pseudopotentials with a small {{TAG\|EDIFF}} and avoid {{TAG\|ISMEAR}}>0.\|:}}
	:The DFT run can be done with your favorite setup, but we recommend attaining very high precision (small {{TAG\|EDIFF}} tag) and to use a small smearing width ({{TAG\|SIGMA}} tag), and to avoid higher-order Methfessel-Paxton smearing (see also {{TAG\|ISMEAR}}). That is because Methfessel-Paxton smearing can lead to negative one-electron occupancies, which can result in unphysical correlation energies. We suggest using [[GGA\|PBE]] orbitals as input for the ACFDT-RPA run, but other choices, e.g., local-density-approximation (LDA) or [[:Category:Hybrid_functionals\|hybrids]], are possible as well. For hybrid functionals, we suggest carefully considering the caveats mentioned in Ref. {{cite\|paier:2008}}. Specifically, the RPA dielectric matrix yields too weak screening for hybrid functionals, which potentially deteriorates RPA results.		:The DFT run can be done with your favorite setup, but we recommend attaining very high precision (small {{TAG\|EDIFF}} tag) and to use a small smearing width ({{TAG\|SIGMA}} tag), and to avoid higher-order Methfessel-Paxton smearing (see also {{TAG\|ISMEAR}}). That is because Methfessel-Paxton smearing can lead to negative one-electron occupancies, which can result in unphysical correlation energies. We suggest using [[GGA\|PBE]] orbitals as input for the ACFDT-RPA run, but other choices, e.g., local-density-approximation (LDA) or [[:Category:Hybrid_functionals\|hybrids]], are possible as well. For hybrid functionals, we suggest carefully considering the caveats mentioned in Ref. {{cite\|paier:2008}}. Specifically, the RPA dielectric matrix yields too weak screening for hybrid functionals, which potentially deteriorates RPA results.
	'''Step 2:''' <span ~~class~~="EXX-inline-ref">Compute the HF energy <math>E_{\mathrm{EXX}}</math> using the DFT orbitals</span>. That is, e.g.,		'''Step 2:''' <span id="EXX-inline-ref">Compute the HF energy <math>E_{\mathrm{EXX}}</math> using the DFT orbitals</span>. That is, e.g.,
	{{TAG\|ALGO}} = EIGENVAL ; {{TAG\|NELM}} = 1		{{TAG\|ALGO}} = EIGENVAL ; {{TAG\|NELM}} = 1
	{{TAG\|LWAVE}} = .FALSE. ! {{FILE\|WAVECAR}} not written		{{TAG\|LWAVE}} = .FALSE. ! {{FILE\|WAVECAR}} not written

← Older revision		Revision as of 12:22, 2 May 2025
Line 28:		Line 28:
	{{NB\|tip\|Use the Perdew-Burke-Ernzerhof (PBE) pseudopotentials with a small {{TAG\|EDIFF}} and avoid {{TAG\|ISMEAR}}>0.\|:}}		{{NB\|tip\|Use the Perdew-Burke-Ernzerhof (PBE) pseudopotentials with a small {{TAG\|EDIFF}} and avoid {{TAG\|ISMEAR}}>0.\|:}}
	:The DFT run can be done with your favorite setup, but we recommend attaining very high precision (small {{TAG\|EDIFF}} tag) and to use a small smearing width ({{TAG\|SIGMA}} tag), and to avoid higher-order Methfessel-Paxton smearing (see also {{TAG\|ISMEAR}}). That is because Methfessel-Paxton smearing can lead to negative one-electron occupancies, which can result in unphysical correlation energies. We suggest using [[GGA\|PBE]] orbitals as input for the ACFDT-RPA run, but other choices, e.g., local-density-approximation (LDA) or [[:Category:Hybrid_functionals\|hybrids]], are possible as well. For hybrid functionals, we suggest carefully considering the caveats mentioned in Ref. {{cite\|paier:2008}}. Specifically, the RPA dielectric matrix yields too weak screening for hybrid functionals, which potentially deteriorates RPA results.		:The DFT run can be done with your favorite setup, but we recommend attaining very high precision (small {{TAG\|EDIFF}} tag) and to use a small smearing width ({{TAG\|SIGMA}} tag), and to avoid higher-order Methfessel-Paxton smearing (see also {{TAG\|ISMEAR}}). That is because Methfessel-Paxton smearing can lead to negative one-electron occupancies, which can result in unphysical correlation energies. We suggest using [[GGA\|PBE]] orbitals as input for the ACFDT-RPA run, but other choices, e.g., local-density-approximation (LDA) or [[:Category:Hybrid_functionals\|hybrids]], are possible as well. For hybrid functionals, we suggest carefully considering the caveats mentioned in Ref. {{cite\|paier:2008}}. Specifically, the RPA dielectric matrix yields too weak screening for hybrid functionals, which potentially deteriorates RPA results.
	'''Step 2:''' <span ~~class~~="EXX-inline-ref">Compute the HF energy <math>E_{\mathrm{EXX}}</math> using the DFT orbitals</span>. That is, e.g.,		'''Step 2:''' <span id="EXX-inline-ref">Compute the HF energy <math>E_{\mathrm{EXX}}</math> using the DFT orbitals</span>. That is, e.g.,
	{{TAG\|ALGO}} = EIGENVAL ; {{TAG\|NELM}} = 1		{{TAG\|ALGO}} = EIGENVAL ; {{TAG\|NELM}} = 1
	{{TAG\|LWAVE}} = .FALSE. ! {{FILE\|WAVECAR}} not written		{{TAG\|LWAVE}} = .FALSE. ! {{FILE\|WAVECAR}} not written