4.7 Supported methods in autosr

Currently, the keywords of all supported methods in autosr are:

MP2, CCD, CCSD, CCSD(T), CCSDT, CCSDT(Q), CCSDTQ,

CCSD-F12, CCSD(T)-F12,

EOM-CCSD, EOM-IP-CCSD, EOM-EA-CCSD,

CC2, CC3,

ADC(2), ADC(3),

QCISD, QCISD(T)

4.7.1 MP2

Second-order Møller-Plesset perturbation theory. An example is shown below

%mem=4GB
%nprocshared=2
#p MP2/cc-pVTZ

mokit{noRI}

0 1
O     0.000000    0.000000    0.062007
H     0.000000   -0.783976   -0.492052
H     0.000000    0.783976   -0.492052

Different keywords in the Title Card Line can be applied:

(TBD: to be determined)

The local correlation method DLPNO-MP2 is supported in ORCA/PSI4, and it can be invoked by

#p MP2/def2TZVPP

mokit{DLPNO,MP2_prog=ORCA}

or by

#p DLPNO-MP2/def2TZVPP

mokit{MP2_prog=ORCA}

Please read related keyword MP2_prog.

4.7.2 CCD

Coupled cluster doubles.

Please read related keyword CC_prog.

4.7.3 CCSD

Coupled cluster singles and doubles.

The CCSD nonlinear equations will be solved to obtain the t1 and t2 amplitudes. The T1-diagnostic value will be printed as well.

By default, the MP2 t2 amplitudes are used as the initial guess of a CCSD calculation. For CC_prog=PySCF, if the STC-CC package is installed, the STC-RI-CCSD amplitudes will be used as the initial guess of CCSD; and if there are converged t1, t2 amplitudes generated from a previous calculation (saved in .npy files), they will be read and the CCSD will be converged in 1 cycle.

The local correlation method DLPNO-CCSD is supported in ORCA, and it can be invoked by

#p CCSD/def2TZVPP

mokit{DLPNO,CC_prog=ORCA}

or by

#p DLPNO-CCSD/def2TZVPP

mokit{CC_prog=ORCA}

Please read related keyword CC_prog.

4.7.4 CCSD(T)

Coupled cluster singles, doubles and perturbative triples.

The CCSD nonlinear equations will be solved to obtain the t1 and t2 amplitudes, then t3 amplitudes are directly calculated via perturbation theory (using electron integrals, t1 and t2 amplitudes). The CCSD T1-diagnostic value will be printed as well.

The local correlation method DLPNO-CCSD(T) is supported in ORCA, and it can be invoked by

#p CCSD(T)/def2TZVPP

mokit{DLPNO,CC_prog=ORCA}

or by

#p DLPNO-CCSD(T)/def2TZVPP

mokit{CC_prog=ORCA}

DLPNO-CCSD(T) employs localized MOs in correlation energy calculation, rather than canonical MOs. So the perturbative triples (i.e. t3 amplitudes) cannot be directly calculated using t1 and t2. In principle, t3 amplitudes must be solved in a iterative way (by solving a set of linear equations). This “iterative” way is different from the iteration in the CCSDT method. For the former, the t1/t2 amplutides are solved in (DLPNO-)CCSD and fixed in t3 computation; for the latter, the t1/t2/t3 amplutides are solved together in a set of nonlinear equations.

By default, when the user requests DLPNO-CCSD(T) using two input formats above, the DLPNO-CCSD(T0) method is used, i.e. the perturbative triples are directly calculated as an approximation. If more accurate results are desired, it is strongly recommended to use DLPNO-CCSD(T1), where the perturbative t3 amplitudes are solved iteratively due to the non-canonical MO basis.

Please read related keyword CC_prog.

4.7.5 CCSDT

Coupled cluster singles, doubles and (full) triples.

The CCSDT nonlinear equations will be solved to obtain t1, t2 and t3 amplitudes, i.e. three sets of amplitudes are solved together and iteratively. The final numerical values of t1 and t2 amplitudes will be different from those values of t1 and t2 in CCSD. So usually we do not obtain the CCSD T1-diagnostic value in a CCSDT calculation. But when CC_prog=PySCF, autosr will firstly perform a CCSD calculation and then use obtained t1, t2 amplitudes as the initial guess of CCSDT, so you can see a CCSD T1-diagnostic value in such case.

Please read related keyword CC_prog.

(TBD: to be determined)

4.7.6 CCSDT(Q)

Coupled cluster singles, doubles, triples and perturbative quadruples.

The CCSDT nonlinear equations will be solved to obtain t1, t2 and t3 amplitudes, i.e. three sets of amplitudes are solved together and iteratively. t4 amplitudes are directly calculated via perturbation theory (using electron integrals, t1, t2 and t3 amplitudes). When CC_prog=PySCF, autosr will firstly perform a CCSD calculation and then use obtained t1, t2 amplitudes as the initial guess of CCSDT, so you can see a CCSD T1-diagnostic value in such case.

This is supported since MOKIT >= 1.2.9rc3, remember to check the version you used and update MOKIT if necessary. Please read related keyword CC_prog.

4.7.7 CCSDTQ

Coupled cluster singles, doubles, triples and quadruples.

The CCSDTQ nonlinear equations will be solved to obtain t1, t2, t3 and t4 amplitudes, i.e. four sets of amplitudes are solved together and iteratively. When CC_prog=PySCF, autosr will firstly perform a CCSD calculation and then use obtained t1, t2 amplitudes as the initial guess of CCSDT. Further, CCSDT t1, t2 and t3 converged amplitudes will be used as the initial guess of CCSDTQ, so you can see a CCSD T1-diagnostic value in such case.

This is supported since MOKIT >= 1.2.9rc3, remember to check the version you used and update MOKIT if necessary. Please read related keyword CC_prog.

Note 1: Currently UHF-based CCSDTQ is not supported in PySCF. You can use CC_prog=MRCC/ORCA.

Note 2: Currently RHF-based CCSDTQ is not supported in ORCA 6.1. UHF-based CCSDTQ is supported and you can use that method.

4.7.8 CCSD-F12

CCSD with F12 correction.

4.7.9 CCSD(T)-F12

CCSD(T) with F12 correction.

CCSD(T)-F12A and CCSD(T)-F12B methods (i.e. the A/B approximations) are also supported when CC_prog=Molpro.

4.7.10 EOM-CCSD

Equation-of-motion based on the CCSD ground state wave function. This is a frequently used wave function method for excited state calculations. EOMCCSD is a synonym for this option.

Please read related keyword EOM_prog.

4.7.11 EOM-IP-CCSD

Ionized potential (IP) EOM framework based on the CCSD ground state wave function. This is a frequently used wave function method for IP calculations. IP-EOM-CCSD is a synonym for this option.

Please read related keyword EOM_prog.

4.7.12 EOM-EA-CCSD

Electron attachment (EA) EOM framework based on the CCSD ground state wave function. This is a frequently used wave function method for EA calculations. EA-EOM-CCSD is a synonym for this option.

Please read related keyword EOM_prog.

4.7.13 CC2

Second-order approximate coupled cluster singles and doubles.

Please read related keyword CC_prog.

4.7.14 CC3

Third-order approximate coupled cluster singles, doubles, and triples.

Please read related keyword CC_prog.

4.7.15 ADC(2)

Second-order algebraic diagrammatic construction method.

The SOS-ADC(2) and SCS-ADC(2) variants are also supported. The IP-ADC(2) and EA-ADC(2) methods for IP/EA calculations are also supported.

Please read related keyword ADC_prog.

RI-ADC(2) for electronic excited states

%mem=6GB
%nprocshared=4
#p ADC(2)/def2TZVP

mokit{}

0 1
O         0.00000000     0.00000000     0.06200700
H         0.00000000    -0.78397600    -0.49205200
H         0.00000000     0.78397600    -0.49205200

RI-accelerated IP-ADC(2) for ionized states

%mem=6GB
%nprocshared=4
#p IP-ADC(2)/def2TZVP

mokit{}

0 1
O         0.00000000     0.00000000     0.06200700
H         0.00000000    -0.78397600    -0.49205200
H         0.00000000     0.78397600    -0.49205200

RI-accelerated EA-ADC(2) for electron attachment

%mem=6GB
%nprocshared=4
#p EA-ADC(2)/def2TZVP

mokit{}

0 1
O         0.00000000     0.00000000     0.06200700
H         0.00000000    -0.78397600    -0.49205200
H         0.00000000     0.78397600    -0.49205200

ADC(2) (with RI approximation turned off) for electronic excited states

%mem=6GB
%nprocshared=4
#p ADC(2)/def2TZVP

mokit{ADC_prog=PSI4,noRI}

0 1
O         0.00000000     0.00000000     0.06200700
H         0.00000000    -0.78397600    -0.49205200
H         0.00000000     0.78397600    -0.49205200

4.7.16 ADC(3)

Third-order algebraic diagrammatic construction method.

Please read related keyword ADC_prog.

RI-ADC(3) for electronic excited states

%mem=6GB
%nprocshared=4
#p ADC(3)/def2TZVP

mokit{ADC_prog=PSI4}

0 1
O         0.00000000     0.00000000     0.06200700
H         0.00000000    -0.78397600    -0.49205200
H         0.00000000     0.78397600    -0.49205200

4.7.17 QCISD

Quadratic CI calculation including single and double substitutions.

4.7.18 QCISD(T)

Quadratic CI calculation including single and double substitutions with a triples contribution to the energy.