CHGCAR - Revision history

Huebsch: /* Related tags and articles */

2025-02-18T12:39:34Z

Related tags and articles

← Older revision		Revision as of 12:39, 18 February 2025
Line 68:		Line 68:
	{{TAG\|ICHARG}},		{{TAG\|ICHARG}},
	{{TAG\|LMAXMIX}},		{{TAG\|LMAXMIX}},
	FFT-grid dimensions: {{TAG\|ENCUT}},{{TAG\|NGXF}},{{TAG\|NGYF}},{{TAG\|NGZF}}		FFT-grid dimensions: {{TAG\|ENCUT}}, {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}


	~~----~~
	[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Electronic ground-state properties]][[Category:Charge density]]		[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Electronic ground-state properties]][[Category:Charge density]]

Schlipf: /* Related tags and articles */

2024-06-11T12:00:14Z

Related tags and articles

← Older revision		Revision as of 12:00, 11 June 2024
Line 71:		Line 71:

	----		----
	[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Charge density]]		[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Electronic ground-state properties]][[Category:Charge density]]

Schlipf: /* Related tags and articles */

2024-06-10T15:47:23Z

Related tags and articles

← Older revision		Revision as of 15:47, 10 June 2024
Line 71:		Line 71:

	----		----
	[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:~~Electronic ground-state properties~~]]		[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Charge density]]

Schlipf: /* Related tags and articles */

2024-06-10T15:37:26Z

Related tags and articles

← Older revision		Revision as of 15:37, 10 June 2024
Line 71:		Line 71:

	----		----
	[[Category:Files]][[Category:Input files]][[Category:Output files]]		[[Category:Files]][[Category:Input files]][[Category:Output files]][[Category:Electronic ground-state properties]]

Karsai at 22:44, 16 February 2024

2024-02-16T22:44:48Z

← Older revision		Revision as of 22:44, 16 February 2024
Line 8:		Line 8:
	* Structure in {{FILE\|POSCAR}} format		* Structure in {{FILE\|POSCAR}} format
	* FFT-grid dimensions {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}		* FFT-grid dimensions {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}
	* Charge ~~density~~ times FFT-grid volume is written with multiple real numbers per line until all {{TAG\|NGXF}}{{TAG\|NGYF}}{{TAG\|NGZF}} values of the block are written.		* Charge times FFT-grid volume is written with multiple real numbers per line until all {{TAG\|NGXF}}{{TAG\|NGYF}}{{TAG\|NGZF}} values of the block are written.
	* Augmentation occupancies		* Augmentation occupancies

Huebsch: /* Format */

2024-02-07T14:38:55Z

Format

← Older revision		Revision as of 14:38, 7 February 2024
Line 21:		Line 21:
	{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math> in units 1/Å<math>^3</math>.}}		{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math> in units 1/Å<math>^3</math>.}}
	Hence,		Hence,
	::<math>n(r)=data(r)/(V_{grid}V_{cell}), V_{grid} = N_{GXF}N_{GYF}*N_{GZF}, V_{cell} = \|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\| </math>,		::<math>n(r)=data(r)/(V_{grid}*V_{cell}), </math>
			::<math>V_{grid} = N_{GXF}N_{GYF}N_{GZF}, </math>
			::<math>V_{cell} = \|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\| </math>,
	where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),		where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),

Huebsch: /* Magnetic calculations */

2024-02-07T10:44:16Z

Magnetic calculations

← Older revision		Revision as of 10:44, 7 February 2024
Line 35:		Line 35:
	* Charge density times FFT-grid volume (spin up + spin down)		* Charge density times FFT-grid volume (spin up + spin down)
	* Augmentation occupancies		* Augmentation occupancies
			* FFT-grid dimensions
	* Magnetization density (spin up - spin down)		* Magnetization density (spin up - spin down)
	* Augmentation occupancies		* Augmentation occupancies
Line 43:		Line 44:
	* Augmentation occupancies		* Augmentation occupancies
	* Augmentation occupancies (imaginary part)		* Augmentation occupancies (imaginary part)
			* FFT-grid dimensions
	* Magnetization density times FFT-grid volume in <math>\sigma_1</math> direction		* Magnetization density times FFT-grid volume in <math>\sigma_1</math> direction
	* Augmentation occupancies		* Augmentation occupancies
	* Augmentation occupancies (imaginary part)		* Augmentation occupancies (imaginary part)
			* FFT-grid dimensions
	* Magnetization density times FFT-grid volume in <math>\sigma_2</math> direction		* Magnetization density times FFT-grid volume in <math>\sigma_2</math> direction
	* ...		* ...
			* FFT-grid dimensions
	* Magnetization density times FFT-grid volume in <math>\sigma_3</math> direction		* Magnetization density times FFT-grid volume in <math>\sigma_3</math> direction
	* ....		* ....

Huebsch at 10:43, 7 February 2024

2024-02-07T10:43:09Z

Huebsch at 09:39, 7 February 2024

2024-02-07T09:39:50Z

← Older revision		Revision as of 09:39, 7 February 2024
Line 13:		Line 13:
	The real-space mesh (NX,NY,NZ) is uniform and is spanned by the lattice vectors <math>\vec{a}, \vec{b}, \vec{c}</math> defined in the structure block. The coordinates of the mesh points can be restored via		The real-space mesh (NX,NY,NZ) is uniform and is spanned by the lattice vectors <math>\vec{a}, \vec{b}, \vec{c}</math> defined in the structure block. The coordinates of the mesh points can be restored via

	::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\~~vec~~{a}+\frac{N_y-1}{N_{GYF}}\~~vec~~{b}+\frac{N_z-1}{N_{GZF}}\~~vec~~{c}</math>.		::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\mathbf{a}+\frac{N_y-1}{N_{GYF}}\mathbf{b}+\frac{N_z-1}{N_{GZF}}\mathbf{c}</math>.
	The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}.		The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}.

Line 19:		Line 19:

	::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.		::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.
	{{NB\|important\|Remember that the values must be divided by the FFT-grid volume to obtain the charge density.}}		{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math>.}}
	::<math>n(r)=data(r)/(~~NGXF~~~~NGYF~~~~NGZF~~)</math>		Hence,
			::<math>n(r)=data(r)/(N_{GXF}N_{GYF}N_{GZF})/\|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\|</math>,
			where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),

			::<math>\text{NELECT}=\int_{V_{cell}} n(\mathbf{r}) d^3\mathbf{r}= \sum_{N_X,N_Y,N_Z} data(N_X,N_Y,N_Z)/(N_{GXF}N_{GYF}N_{GZF})</math>.
			By convention, the charge density <math>n(r)</math> is not assumed to have the unit e/Å<math>^3</math> because the potential (e.g. {{FILE\|LOCPOT}}, {{TAG\|WRT_POTENTIAL}}) is assumed to be in eV. However, e<math>=1</math>, so while this convention makes the sign of <math>n(r)</math> less ambiguous, it has no effect on the numerical values.
	{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}		{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}
	Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.		Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.

Huebsch at 08:56, 7 February 2024

2024-02-07T08:56:14Z

Show changes

← Older revision		Revision as of 09:39, 7 February 2024
Line 13:		Line 13:
	The real-space mesh (NX,NY,NZ) is uniform and is spanned by the lattice vectors <math>\vec{a}, \vec{b}, \vec{c}</math> defined in the structure block. The coordinates of the mesh points can be restored via		The real-space mesh (NX,NY,NZ) is uniform and is spanned by the lattice vectors <math>\vec{a}, \vec{b}, \vec{c}</math> defined in the structure block. The coordinates of the mesh points can be restored via

	::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\~~vec~~{a}+\frac{N_y-1}{N_{GYF}}\~~vec~~{b}+\frac{N_z-1}{N_{GZF}}\~~vec~~{c}</math>.		::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\mathbf{a}+\frac{N_y-1}{N_{GYF}}\mathbf{b}+\frac{N_z-1}{N_{GZF}}\mathbf{c}</math>.
	The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}.		The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}.

Line 19:		Line 19:

	::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.		::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.
	{{NB\|important\|Remember that the values must be divided by the FFT-grid volume to obtain the charge density.}}		{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math>.}}
	::<math>n(r)=data(r)/(~~NGXF~~~~NGYF~~~~NGZF~~)</math>		Hence,
			::<math>n(r)=data(r)/(N_{GXF}N_{GYF}N_{GZF})/\|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\|</math>,
			where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),

			::<math>\text{NELECT}=\int_{V_{cell}} n(\mathbf{r}) d^3\mathbf{r}= \sum_{N_X,N_Y,N_Z} data(N_X,N_Y,N_Z)/(N_{GXF}N_{GYF}N_{GZF})</math>.
			By convention, the charge density <math>n(r)</math> is not assumed to have the unit e/Å<math>^3</math> because the potential (e.g. {{FILE\|LOCPOT}}, {{TAG\|WRT_POTENTIAL}}) is assumed to be in eV. However, e<math>=1</math>, so while this convention makes the sign of <math>n(r)</math> less ambiguous, it has no effect on the numerical values.
	{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}		{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}
	Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.		Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.

← Older revision		Revision as of 10:43, 7 February 2024
Line 8:		Line 8:
	* Structure in {{FILE\|POSCAR}} format		* Structure in {{FILE\|POSCAR}} format
	* FFT-grid dimensions {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}		* FFT-grid dimensions {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}
	* Charge density times FFT-grid volume is written with ~~max five~~ real numbers per line until all {{TAG\|NGXF}}{{TAG\|NGYF}}{{TAG\|NGZF}} values of the block are written.		* Charge density times FFT-grid volume is written with multiple real numbers per line until all {{TAG\|NGXF}}{{TAG\|NGYF}}{{TAG\|NGZF}} values of the block are written.
	* Augmentation occupancies		* Augmentation occupancies

Line 14:		Line 14:

	::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\mathbf{a}+\frac{N_y-1}{N_{GYF}}\mathbf{b}+\frac{N_z-1}{N_{GZF}}\mathbf{c}</math>.		::<math>(N_x,N_y,N_z) \hat{=} \frac{N_x-1}{N_{GXF}}\mathbf{a}+\frac{N_y-1}{N_{GYF}}\mathbf{b}+\frac{N_z-1}{N_{GZF}}\mathbf{c}</math>.
	The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting {{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}.		The dimensions can be increased by increasing the cutoff energy ({{TAG\|ENCUT}}) or explicitly by setting the fine FFT-grid dimensions ({{TAG\|NGXF}}, {{TAG\|NGYF}}, {{TAG\|NGZF}}).

	To arrange the data on the real-space grid in the unit cell, mind that the data runs fastest over NX and slowest over NZ. To be more explicit, the density is written using the following command in Fortran		To arrange the data on the real-space grid in the unit cell, mind that the data runs fastest over NX and slowest over NZ. To be more explicit, the density is written using the following command in Fortran

	::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.		::<code> WRITE(IU,FORM) (((C(NX,NY,NZ),NX=1,{{TAGBL\|NGXF}}),NY=1,{{TAGBL\|NGYF}}),NZ=1,{{TAGBL\|NGZF}}) </code>.
	{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math>.}}		{{NB\|important\|Remember that the values must be divided by the FFT-grid volume and the cell volume to obtain the charge density <math>n(r)</math> in units 1/Å<math>^3</math>.}}
	Hence,		Hence,
	::<math>n(r)=data(r)/(N_{GXF}N_{GYF}N_{GZF})/\|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\|</math>,		::<math>n(r)=data(r)/(V_{grid}V_{cell}), V_{grid} = N_{GXF}N_{GYF}*N_{GZF}, V_{cell} = \|\mathbf{a}\cdot(\mathbf{b}\times\mathbf{c})\| </math>,
	where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),		where <math>n(r)</math> is the charge density in units 1/Å<math>^3</math>. Sanity check: The integral of <math>n(r)</math> over the unit cell yields the number of valence electrons ({{TAG\|NELECT}}),

	::<math>\text{NELECT}=\int_{V_{cell}} n(\mathbf{r}) d^3\mathbf{r}= \sum_{N_X,N_Y,N_Z} data(N_X,N_Y,N_Z)/(N_{GXF}N_{GYF}N_{GZF})</math>.		::<math>\text{NELECT}=\int_{V_{cell}} n(\mathbf{r}) d^3\mathbf{r}= \sum_{N_X,N_Y,N_Z} data(N_X,N_Y,N_Z)/(N_{GXF}N_{GYF}N_{GZF})</math>.
	By convention, the charge density <math>n(r)</math> is not ~~assumed to have the unit~~ e/Å<math>^3</math> because the potential (e.g. {{FILE\|LOCPOT}}, {{TAG\|WRT_POTENTIAL}}) is assumed to be in eV. However, e<math>=1</math>, so while this convention makes the sign of <math>n(r)</math> less ambiguous, it has no effect on the numerical values.		By our convention, the charge density <math>n(r)</math> is in units 1/Å<math>^3</math> and not e/Å<math>^3</math> because the potential (e.g. {{FILE\|LOCPOT}}, {{TAG\|WRT_POTENTIAL}}) is assumed to be in eV. However, e<math>=1</math>, so while this convention makes the sign of <math>n(r)</math> less ambiguous, it has no effect on the numerical values.
	{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}		{{NB\|warning\|The augmentation occupancies are written up to the ''l''-quantum number set by the {{TAG\|LMAXMIX}}.}}
	Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.		Restarting calculations without one-center PAW occupancy matrices up to the appropriate ''l''-quantum number leads to loss of information. This is particularly problematic for calculations with fixed charge density, e.g., band-structure calculations. See {{TAG\|LMAXMIX}} for more details.