On development features

RAS-CI implementation scheme

Fortran functions only

_images/rasci_scheme.pdf

RAS-CI Contraction scheme

_images/rasci_contract_even.pdf _images/rasci_contract_odd.pdf

Fragment localization

  • ras_nfrag: Number of fragments

  • ras_nfrag_atoms: Number of atoms in each fragment (atoms should be correlative in input file)

  • ras_frag_st: localization algorithm to use (0: Boys, 1: Sequential [https://aip.scitation.org/doi/10.1063/1.4904292])

  • ras_frag_sets: Define orbitals to localize (default: localize all)

example:

ras_nfrag        4 !NFrag   ! # of fragments
ras_nfrag_atoms  [34,32,34,32]! # atoms in each fragment
RAS_FRAG_ST      1        ! 0: Boys   1: Sequential
ras_frag_sets    [86,191,4,4] ! sets of orbitals to localize

Fractional Orbital Density (FOD)

Generate the fraction orbital density of a RAS-CI excited state.

  • gui 2 : Request generation of .fchk file

  • ras_natorb_state i : Select i excited state for FOD calculation

  • ras_fod: Activate FOD (False: Deactivate(default), True: Activate)

Spin polarization

Perform spin polarization treatment for srDFT using the method described in: Coulsonm C. A. Fisher, I. Notes on the molecular Orbital Tratment of the Hydrogen Molecule. Philos. Mag. 1949, 40, 386-393.

  • ras_srdft_spinpol: Activate spin polarization (False: Deactivate(default), True: Activate)

Sequential Diabatization

To activate the diabatization analysis of excited states in ras-ci method it is necessary to define a section block and 3 REM keywords in the qchem input. The section block defines the excited states number to be included in the diabatization (by increasing order). This is equivalent to the one used in TDDFT diabatization method (check qchem manual for further information).

This block can be placed at the end of the usual input (after the $END of the $REM section) and has an structure like this

$localized_diabatization
adiabatic states
n1 n2 n3 n4 n5 n6
$end

where n1, n2, n3, etc.. are the excited state number (sorted by increasing energy) to be used as reference. Also the use of this block requires a keyword to indicate the number of adiabatic states defined

sts_multi_nroots N_ad

where N_ad are the number of adibatic states defined in $localized_diabatization block.

Once the states to be used in the diabatization analysis are defined it is necessary to request a diabatization calculation using the keyword

cis_diabath_decompose N

where N is the number of sequential steps in the diabatization process (if N=true, this is equivalent to 1). A sequential step is complete diabatization process using a particular set of states and a particular diabatization method. In usual diabatization calculations this is just 1, but in the current implementation it is possible to link sequential diabatization steps using a different number of states and method to obtain more sophisticated results.

Then it is necessary to define the diabatization method and the excited states to use (with respect to the reference states). To do this we have to define the keyword

ras_diab_seq_data    [n1_1,n2_1,n3_1,meth_1,data_1,n1_2,n2_2,meth_2,data_2]

where n1_1, n2_1, n3_1 are the state numbers to be used in the 1st step of the diabatization. These step number are taken respect to the reference states defined in $localized_diabatization section (so the 1st reference state is 1, 2nd reference state is 2, etc..). meth_1 indicates the diabatization method to be used in the 1st step, this method is defined by an integer number where

1  -> ER method   (data ignored)
2  -> Boys method (data ignored)
3  -> DQ method   (data: float between 0.0-1.0, fraction of quadrupole/dipole)

finally, data_1 indicates the additional parameters for the method. Currently only DQ makes use of this parameter, but data value must be defined for all methods (even if it is ignored in the diabatization method).

this structure is repeated for each diabatization step requested. In the above example n1_2,n2_2,meth_2,data_2 are the parameters of the second diabatization step. In order to keep track of which states/data correspond to each diabatization step it is necessary to define another keyword that contains the list of number of states for each diabatization step

ras_diab_seq_list    [sn1 ,sn2]

where sn1 and sn2 are number of states to use in each diabatization state (not counting for method and data, just number of diabatic states)

After each diabatization step the resulting diabatic states are reordered by increasing energy, and this order will be the one used to select the states for the next diabatization step.

This is an example of diabatization input:

$rem
ras_diab_seq_list    [7,4]
ras_diab_seq_data    [1,2,3,4,5,6,7,2,0.0,4,5,6,7,3,1.0]
sts_multi_nroots  7
$end

$localized_diabatization
adiabatic states
2 3 4 5 6 7 8
$end

In this example 2 diabatization steps are performed, the first one uses all 7 diabatic states defined in the $localized_diabatization block (1-7) using method 2 (Boys) and data 0.0 (which is ignored for this method). The second step uses the 4 highest energy diabatic states (4-7) obtained from the previous step and performs a diabatization using method 3 (DQ) with a parameter 1.0 (100% quadrupole).

Note

In the recent versions of Q-Chem, Mulliken analysis of the diabatic/adiabatic states is not calculated by default (including Attachment/Detachment). To do this analysis use: STATE_ANALYSIS=TRUE

SOC Natural Transition Orbitals (Spinless triplet density matrix NTOs)

The calculation of Natural Transition Orbitals (NTO) is requested using the following keywords:

STATE_ANALYSIS = True
GUI = 2

These keywords will print the NTO’s in the fchk file with titles:

"NTOs occupancies (x,y)”
"NTOs U coefficients (x,y)”
"NTOs V coefficients (x,y)”

Where x and y denote the two states involved in the transition. The format of the NTO’s is the same as for the Natural Orbitals (NO) so they can be visualized using any standard visualization software by changing the title names.

Wave function analysis of RAS-CI states

Analysis of RAS-CI states is requested using STATE_ANALYSIS keyword. This analysis include a calculation of the Natural Orbitals (NO), Transition Natural Orbitals (NTO), Spinless triplet density matrix NTOs (SOC-NTO), Fractional Occupation Density (FOD), Electronic Density, Spin Density and Transition Density. These properties can be plotted in Cube format or Molden format for each RAS-CI state, this is controlled by PLOTS keyword.

Molden format (NO, NTO and SOC-NTO only):

STATE_ANALYSIS = True
PLOTS = 0     ! Note: This can be omitted, default is 0
NTO_PAIRS = 2 ! Note: This can be omitted, default is 2
GUI 0

Cube format

STATE_ANALYSIS = True
NTO_PAIRS = 2 ! Note: This can be omitted, default is 2
PLOTS = 1
GUI 0

$plots
   grid_points                    50 50 50
   grid_range  (-8,8) (-8,8) (-8,8)
$end

Note

PLOTS = 1 requires $plots section to be written in Q-Chem input using new plot format. (https://manual.q-chem.com/5.1/sect-plots.html). grid_range can be omitted to use automatic range adjusted to the molecular size.

Warning

The number of NTO pairs written is controlled by NTO_PAIRS keyword. If not specified the value is set to 2.

Warning

To plot data in both Cube and Molden files GUI should be set to 0. If not a fchk will be generated instead. At this moment is not possible to plot Molden/Cubes/fchk at the same time.

Using STATE_ANALYSIS = 2 will compute and plot interstate properties for all pairs of states. This is required to compute and plot SOC-NTO.

Molden format (NO and NTO only):

STATE_ANALYSIS = 2
PLOTS = 0  ! Note: This can be omitted, default is 0
GUI 0

Cube format

STATE_ANALYSIS = 2
PLOTS = 1
GUI 0

$plots
   grid_points                    50 50 50
$end

The NTO information will be written in the output as

e-/hole pair  1 alpha:  ampl =  0.450353 ( 20.3%)   [3 : 4]
e-/hole pair  2 alpha:  ampl =  0.203136 (  4.1%)   [2 : 5]
e-/hole pair  1 beta :  ampl =  0.311186 (  9.7%)   [3 : 4]
e-/hole pair  2 beta :  ampl =  0.279777 (  7.8%)   [2 : 5]

where the last two numbers “[3: 4]” indicate the cubefile numbers that corresponds of the NTO pair (electron/hole).

Notes about diabatization in TDDFT method

Due to a possible bug in Q-Chem a change of behavior appeared in Qchem v5.x respect to Mulliken analysis of diabatic states. Now in addition to cis_ampl_anal the keyword:

NAMD_NSURFACES 0

is required to analyze the diabatc states. If not set, the adiabatic states are analyzed instead.

Example for ethene dimer:

$molecule
0 1
C     0.0000000   0.0000000   0.6660120
C     0.0000000   0.0000000   -0.6660120
H     0.0000000   0.9228100   1.2279200
H     0.0000000   -0.9228100  1.2279200
H     0.0000000   -0.9228100  -1.2279200
H     0.0000000   0.9228100   -1.2279200
C     4.2000000   0.0000000   0.6660120
C     4.2000000   0.0000000   -0.6660120
H     4.2000000   0.9228100   1.2279200
H     4.2000000   -0.9228100  1.2279200
H     4.2000000   -0.9228100  -1.2279200
H     4.2000000   0.9228100   -1.2279200
$end

$rem
JOBTYPE      sp
EXCHANGE     hf
cis_n_roots  8
cis_singlets true
cis_triplets false
RPA          false
BASIS        6-31G
cis_ampl_anal  true
mem_static   900
!Diabat
NAMD_NSURFACES 0
loc_cis_ov_separate    false
er_cis_numstate        4
cis_diabath_decompose  true
$end

$localized_diabatization
On the next line, list which excited adiabatic states we want to mix.
1 2 7 8
$end