ReSpect program

xc

Specify details associated with the evaluation of the exchange–correlation (xc) kernel.

Input block

Short variant

xc: 
                                         
                                            
                                               
                                                
                                               
                                               
                                                  
                                                     
[functional]

Extended variant

xc:

functional: [functional]

noncollinearity: [string]

Default

xc:
  functional      : from-scf
  noncollinearity : v2019

Example

xc: XALDA

xc:
     functional      : XALDA
     noncollinearity : v2005

Warning

The use of ALDA or XALDA options is not recommended, because they usually approximate the full xc kernel. Their use is only recommended when comparing the calculated data to implementations in other quantum chemistry programs, where those approximations can not be avoided.

eri

Specify details associated with the evaluation of electron repulsion integrals (ERI) and related two-electron Fock contributions.

Input block

Extended variant

eri:

class: [string]

threshold: [real]

Default

ERI defaults are taken from the SCF section.

Example

eri:
   class:        ssss/abcd 
   threshold:    1.e-15

grid

Specify atomic grids for the numerical evaluation of exchange-correlation DFT contributions.

Input block

Short variant

Extended variant

grid:

all: [string]

[element-symbol]: [string]

[element-index]: [string]

...

Default

Grid defaults are taken from the SCF section.

Example

grid: large

grid:
     C: medium
     7: large

Note

There can be multiple instances of element-symbol and element-index in the grid block.

While lines in the grid block can be mixed, they are always processed in the following order: "all", all "element-symbol" and all "element-index" keywords.

The order of processing the data matters, since the latter lines rewrite the data of the former lines. This way one can easily set the same grid for all Carbons except the Carbon number 7 (see example).

num-eigenvalues

Request number of eigenvalues (excitation energies).

Input line

num-eigenvalues:
                               
                                  
[integer]

Default

num-eigenvalues: 4

Example

num-eigenvalues: 10

convergence

Convergence threshold for excitation energies in [a.u.].

Input line

convergence:
                               
                                  
[real]

Default

convergence: 1.0e-3

Example

convergence: 1.0e-4

maxdim-subspace

Specify the maximum dimension of the reduced subspace into which the true TDDFT operator is projected.

Input line

maxdim-subspace:
                               
                                  
[integer]

Default

maxdim-subspace: 80

Example

maxdim-subspace: 300

tamm-dancoff-approximation

Use Tamm-Dancoff approximation.

Input line

tamm-dancoff-approximation:
                               
                                  
[Boolean]

Default

tamm-dancoff-approximation: False

Example

tamm-dancoff-approximation: True

solver

Specify the type of the eigen-problem solver.

Input line

Default

solver: subspace

Example

solver: direct

Note

The use of direct solver is feasible only for very small number of basis functions, since the full TDDFT matrix is calculated.

pe-rotations

Include positive-energy-state/negative-energy-state rotations in the calculation.

Input line

pe-rotations:
                               
                                  
[Boolean]

Default

pe-rotations: False

Example

pe-rotations: True

print-level

Set the amount of information printed in the output file.

Input line

print-level:
                               
                                  
[string]

Default

print-level: normal

Example

print-level: long

[string]	description
subspace	Iterative subspace eigen-problem solver.
direct	Construct full TDDFT matrix and calculate all eigenvalues using LAPACK routines.

[string]	description
normal	Basic amount of information (recommended).
long	Additional information is printed in the .out file (may clog the output).
debug	New .dbg file will be created with even more information (mostly used for debugging purposes).

[functional]	description
xalda	Uses SVWN5 functional for the exchange-correlation kernel, regardless of the type of the xc functional in the scf step. Full Hartree–Fock kernel is calculated with the percentage of the exchange admixture taken from the scf step.
alda	Uses SVWN5 functional for the exchange-correlation kernel, regardless of the type of the xc functional in the scf step. The Hartree–Fock exchange kernel is omitted even if the scf step was calculated using hybrid functional. Note that the Coulomb part of the Hartree–Fock kernel is not omitted. This is relevant for the open-shell calculation where Coulomb part is nonzero.
from-scf	Functional for calculation of the xc kernel matches the functional from the scf step.

[string]	description
v2005	The 2005 version of the noncollinear xc kernel for tddft section is restricted to open-shell calculations within (X)ALDA approximation. The xc kernel for closed-shell systems have been abandoned because it provides unsatisfactory results. For more details see ref [Komorovsky2019]. The implementation details of the noncollinear xc potential (for scf procedure) can be found in reference ReSpect article [Repisky2020] in eq (68). The noncollinear (X)ALDA xc kernel and its closed-shell limit can be obtained by neglecting gradient variables (g with various subscripts) in ref [Komorovsky2019] in eqs (25) and (45), respectively. The summary of implementation details of the 2005 version of the noncollinear (X)ALDA xc kernel can be found in Table IV in ref [Repisky2020] labeled as ALDA method. The notable disadvantage of the 2005 version of the noncollinear xc kernel is restriction to (X)ALDA approximation. On the other hand this formulation is the most numerically stable noncollinear scheme xc potential and kernel available in the ReSpect for open-shell systems. S. Komorovsky, P. Cherry, and M. Repisky, J. Chem. Phys. 151, 184111 (2019) M. Repisky, S. Komorovsky, M. Kadek, L. Konecny, U. Ekstrom, E. Malkin, M. Kaupp, K. Ruud, O. L. Malkina, and V. G. Malkin, J. Chem. Phys. 152, 184101 (2020)
v2019	The 2019 version of the noncollinear xc kernel for both open- and closed-shell systems have been presented in ref [Komorovsky2019]. This is the most general formulation of the noncollinear scheme and can be used for prediction of all spectroscopic parameters available in the ReSpect. The xc potential and kernel is formulated in the analytic form. The closed-shell limit of the open-shell noncollinear scheme corresponds to the xc kernel for the first time presented in ref [Bast2009]. In the output files we refer to closed-shell and open-shell formulation of the noncollinear xc scheme as WB2009 and WSK2019, respectively. WB2009: [Wang2004] + [Bast2009] WSK2019: [Wullen2002] + [Scalmani2012] + [Komorovsky2019] The implementation details of the 2019 version of the noncollinear xc kernel is summarized in Table I in ref [Komorovsky2019] or in Table IV in reference ReSpect article [Repisky2020] labeled as recommended option. C. Van Wüllen, J. Comput. Chem. 23, 779–785 (2002) F. Wang and T. Ziegler, J. Chem. Phys. 121, 12191 (2004) R. Bast, H. J. A. Jensen, and T. Saue, Int. J. Quant. Chem. 109, 2091–2112 (2009) G. Scalmani and M. J. Frisch, J. Chem. Theory Comput. 8, 2193–2196 (2012) S. Komorovsky, P. Cherry, and M. Repisky, J. Chem. Phys. 151, 184111 (2019) M. Repisky, S. Komorovsky, M. Kadek, L. Konecny, U. Ekstrom, E. Malkin, M. Kaupp, K. Ruud, O. L. Malkina, and V. G. Malkin, J. Chem. Phys. 152, 184101 (2020)

[string]	description
llll	evaluate only the [LL\|LL] class (default for one- and two-component Hamiltonians)
ssll/abcd	evaluate exactly [LL\|LL], [LL\|SS] and [SS\|LL] classes but neglect [SS\|SS]
ssss/aabb	evaluate exactly [LL\|LL], [LL\|SS] and [SS\|LL] classes and use an one-center approximation for [SS\|SS]. In this approximation, both functions in [SS\| and \|SS] should share the same atomic center, otherwise they are excluded from evaluation. This is the default for four-component Hamiltonians.
ssss/abcd	evaluate exactly all four ERI classes

[string]	description
coarse	Adaptive angular part with radial number of points calculated as 30 + (10*p), where p is the number of period of the defined element.
medium	Adaptive angular part with radial number of points calculated as 40 + (10*p), where p is the number of period of the defined element.
large	Adaptive angular part with radial number of points calculated as 50 + (10*p), where p is the number of period of the defined element.
adaptive	Adaptive grid in both the radial and angular part calculated from the input basis.

[string]	description
coarse	Adaptive angular part with radial number of points calculated as 30 + (10*p), where p is the number of period of the defined element.
medium	Adaptive angular part with radial number of points calculated as 40 + (10*p), where p is the number of period of the defined element.
large	Adaptive angular part with radial number of points calculated as 50 + (10*p), where p is the number of period of the defined element.
adaptive	Adaptive grid in both the radial and angular part calculated from the input basis.

[string]	description
coarse	Adaptive angular part with radial number of points calculated as 30 + (10*p), where p is the number of period of the defined element.
medium	Adaptive angular part with radial number of points calculated as 40 + (10*p), where p is the number of period of the defined element.
large	Adaptive angular part with radial number of points calculated as 50 + (10*p), where p is the number of period of the defined element.
adaptive	Adaptive grid in both the radial and angular part calculated from the input basis.

[string]	description
coarse	Adaptive angular part with radial number of points calculated as 30 + (10*p), where p is the number of period of the defined element.
medium	Adaptive angular part with radial number of points calculated as 40 + (10*p), where p is the number of period of the defined element.
large	Adaptive angular part with radial number of points calculated as 50 + (10*p), where p is the number of period of the defined element.
adaptive	Adaptive grid in both the radial and angular part calculated from the input basis.

Latest Posts

Webpage update

July 17, 2024

Program update: ReSpect 5.3.0

July, 2024

Latest Publications

Book chapter on relativistic real-time electron dynamics

May, 2024

Book chapter on relativistic theory of EPR and (p)NMR

May, 2024

Useful Links

Our Contacts

Hylleraas Centre
Department of Chemistry
UiT The Arctic University of Norway
Tromsø, NO-9037 Norway
Email: info@respectprogram.eu

xc

Read more

Warning

eri

Read more

grid

Read more

Note

num-eigenvalues

Read more

convergence

Read more

maxdim-subspace

Read more

tamm-dancoff-approximation

Read more

solver

Read more

Note

pe-rotations

Read more

print-level

Read more

Latest Posts

Webpage update

Program update: ReSpect 5.3.0

Latest Publications

Book chapter on relativistic real-time electron dynamics

Book chapter on relativistic theory of EPR and (p)NMR

Useful Links

Our Contacts