Full configuration interaction (FCI)

Description

FCI exactly diagonalizes the molecular Hamiltonian within the space spanned by all of the Slater determinants. This FCI section is closely related to that for the Complete active space self-consistent field (CASSCF) method.

Prerequisites

A reference wavefunction generated by a HF or CASSCF calculation.

Keywords

maxiter (or maxiter_fci)

Description: Maximum number of iterations in the FCI algorithm.
Datatype: int
Default: 100

thresh (or thresh_fci)

Description: Threshold for the convergence of the selected CI algorithm.
Datatype: double
Default: 1.0e-10

algorithm

Description: FCI algorithm to be used.
Datatype: string
Values:
KH, Knowles, Handy: Use the Knowles–Handy algorithm.
HZ, Harrison, Zarrabian: Use the Harrison–Zarrabian algorithm.
Dist, parallel: Use a parallel algorithm based on HZ.
Default: KH or Dist depending on the size of the calculation and the number of MPI processes.

frozen

Description: Use of the freeze core approximation.
Datatype: bool
Default: false

ncore

Description: Number of frozen core orbitals. ncore overides frozen.
Datatype: int
Default: If frozen is true, subvalence orbitals are frozen. If false, zero.

norb

Description: Number of correlated orbitals. Any high-energy orbitals in excess of this number will be excluded.
Datatype: int
Default: All molecular orbitals except those excluded using ncore.

active

Description: Orbital indices for the orbitals to be included. Note that orbital numbers start from 1.
Datatype: vector<int>
Default: Frontier orbitals are used.

charge

Description: The total charge of the system.
Datatype: int
Default: 0

nspin

Description: The spin multiplicity of the wavefunction (2S+1).
Datatype: int
Default: 0

nstate

Description: Number of states to be calculated.
Datatype: int
Default: 1

davidson_subspace

Description: Number of vectors retained in the limited-memory Davidson algorithm.
Datatype: int
Default: 20
Recommendation: Reduce if an insufficient amount of memory is available (do not reduce to a value lower than 3).

nguess

Description: Number of guess configurations
Datatype: int
Default: nstate

only_ints

Description: If true, calculates integrals and dumps to a file. This is used to interface BAGEL to an external program.
Datatype: bool
Default: false

dipoles

Description: Calculate dipoles
Datatype: bool
Default: false

restart

Description: Generate binary archive files that can be used to restart a calculation.
Datatype: bool
Default: false

print_thresh

Description: Threshold below which CI coefficients are not printed
Datatype: double
Default: 0.05

Example

A Hartree–Fock and FCI calculation on an oxygen molecule with an elongated bond. The HF calculation is a references for the FCI calculation. Please refer to MOLDEN for more information on the print input block.

Sample input

{ "bagel" : [

{
  "title" : "molecule",
  "basis" : "sto-3g",
  "df_basis" : "cc-pvdz-jkfit",
  "angstrom" : true,
  "geometry" : [
    { "atom" : "H",  "xyz" : [ 0.0, 0.0, 1.5 ]},
    { "atom" : "F",  "xyz" : [ 0.0, 0.0, 0.0 ]}
  ]
},

{
  "title" : "hf"
},

{
  "title" : "fci",
  "thresh" : 1.0e-6,
  "nstate" : 2
}

]}

from which one obtains

---------------------------
      FCI calculation
---------------------------

Performs exactly the same way as Knowles & Handy 1984 CPL

o alpha-beta strings
    length:            12
o size of the space
    determinant space:  36
    spin-adapted space: 21

o single displacement lists (alpha)
    length:            60
o single displacement lists (beta)
    length:            60
  * Integral transformation done. Elapsed time: 0.00

   guess   0:   closed 11111.               open ......
   guess   1:   closed 111.1.               open ...1.1

=== FCI iteration ===

    0  0       -98.41480561     4.76e-02      0.00
    0  1       -98.26379090     4.50e-02      0.00

    1  0       -98.51664721     1.09e-02      0.00
    1  1       -98.41322803     1.00e-02      0.00

    2  0       -98.51931832     1.71e-03      0.00
    2  1       -98.41524119     1.91e-04      0.00

    3  0       -98.51940404     3.13e-04      0.00
    3  1 *     -98.41524124     2.55e-14      0.00

    4  0       -98.51940654     4.91e-05      0.00
    4  1 *     -98.41524124     2.80e-14      0.00

    5  0       -98.51940660     2.62e-06      0.00
    5  1 *     -98.41524124     1.86e-14      0.00

    6  0 *     -98.51940660     1.96e-07      0.00
    6  1 *     -98.41524124     2.02e-14      0.00

   * ci vector, state   0, <S^2> = 0.0000
     22222.    -0.9320921105
     2222.2     0.3574718122

   * ci vector, state   1, <S^2> = -0.0000
     222a2b     0.6167681811
     222b2a     0.6167681811
     222ab2    -0.3441653374
     222ba2    -0.3441653374

  * METHOD: FCI                                  0.00

References

Description of Reference Reference
Efficient calculation of sigma vectors (KH) P. J. Knowles and N. C. Handy, Chem. Phys. Lett. 111, 315 (1984).
Efficient calculation of sigma vectors (HZ) R. J. Harrison and S. Zarrabian, Chem. Phys. Lett. 158, 393 (1989).