Incorrect Fitting Basis for ACFDT2

[Connor Fawcett]

Hey everyone,

I've noticed while trying to perform ACFDT2 calculations that my selected auxiliary basis set is not being used when the ACFDT2 program runs. I'm trying to use the aug-cc-PVDZ-PP basis set for Cu atoms in my system, but it seems like the def2-QZVPP fitting set is selected instead of the appropriate mp2fit for aug-cc-PVDZ. To the best of my knowledge, my input file is in line with the example provided in the molpro manual (https://www.molpro.net/manual/doku.php?id=kohn-sham_random-phase_approximation&s[]=rpa)

Please let me know if you have any suspicions why this error may be occurring or if you've experienced similar bugs!

Here is my molpro input file:
geometry=coords.xyz
symmetry,nosym

basis={
set,orbital; default,aug-cc-pVDZ-pp
set,mp2fit; default,aug-cc-pVDZ-pp/mp2fit}

{uks,pbe,maxit=10000
wf,280,1,0
save,2100.2
}

start,2100.2
{rks,pbe,maxit=10000
wf,280,1,0
save,2100.2
}

{ksrpa; acfdt2,orb=2100.2,auxbas=mp2fit,core=0,thren=1.d-8}

Here is the portion of the output file related to the ACFDT2 program:

Basis set: MP2FIT
Default basis aVDZ-PP_MP2F/MP2FIT selected for CU
Default basis aug-cc-pVDZ/MP2FIT selected for O
Default basis aug-cc-pVDZ/MP2FIT selected for C

Basis size: 3042

PROGRAM * ACFDT2 (Fluctuation-dissipation density functional theory)

Authors: A. Hesselmann

Number of occupied orbitals = 140
Number of virtual orbitals = 662
Number of occ-virt elements = 92680
Number of auxiliary functions = 3042
Local exchange factor = 1.00000000
Local correlation factor = 1.00000000
Kernel density threshold = 0.10000000E-11

Number of coupling strength integration points = 7
Number of frequency integration points = 20

Molecular orbitals read from record 2100.2 Type=RKS/CANONICAL
Density matrix read from record 2100.2 Type=RKS/CHARGE
Orbital energies read from record 2100.2 Type=RKS/CANONICAL

Generating basis AVDZ-PP-JKFIT
Atom=CU basis=def2-QZVPP/JKFIT
Atom=O basis=aug-cc-pVDZ/JKFIT
Atom=C basis=aug-cc-pVDZ/JKFIT

Number of basis functions: 3868
Number of fitting functions: 3868

Coulomb fitting:T
Exchange fitting:T

Reference energy = -2858.45625912

Compute (ia|P) integrals... CPU time: 247.38
Compute (ia|fxc|P) integrals...
CPU time: 7548.90

Number of grid points: 7

ngrid alpha weight

1 0.974554 0.064742
2 0.870766 0.139853
3 0.702923 0.190915
4 0.500000 0.208980
5 0.297077 0.190915
6 0.129234 0.139853
7 0.025446 0.064742

Gauss-Legendre omega quadrature

Number of grid points: 20

ngrid omega weight

1 580.122934 1492.201923
2 109.024534 125.117846
3 43.575958 32.545092
4 22.862784 12.869547
5 13.768637 6.336229
6 8.990632 3.569297
7 6.177735 2.201681
8 4.386778 1.449058
9 3.179910 1.000633
10 2.331473 0.716477
11 1.715654 0.527232
12 1.257897 0.395827
13 0.911831 0.301200
14 0.647487 0.230758
15 0.444908 0.176629
16 0.290515 0.133693
17 0.174957 0.098484
18 0.091794 0.068557
19 0.036689 0.042105
20 0.006895 0.017736

Calculate X0(omega)... CPU time: 857.57

ncoup alpha Ecorr(ACFDT) Ecorr(RPA) time

1 0.974554 -19.80862732 -23.14278853 1070.66
2 0.870766 -18.51594724 -21.61752446 1069.54
3 0.702923 -16.22018353 -18.91227054 1069.64
4 0.500000 -12.99634183 -15.12251973 1070.03
5 0.297077 -9.02209274 -10.46887438 1069.55
6 0.129234 -4.71521539 -5.45391861 1069.63
7 0.025446 -1.10027031 -1.26962112 1070.06

Ecorr(ACFDT) = -12.13773469
Ecorr(RPA) = -14.13613434
Total ACFDT energy = -2870.59399381

[peterso...@gmail.com]

Hi Connor,

 

From what I can tell from the testjob, acfdt also uses a jkfit auxiliary basis.  You can either specify JKFIT in your basis block or take the default, which is evidently what is echoed below in your output.

 

Regards,  -Kirk

--
You received this message because you are subscribed to the Google Groups "molpro-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/molpro-user/d6e5258c-db19-455d-8204-04c67e58cc88n%40googlegroups.com.