Default purecart inconsistency?

Questions
#1

Hi all,

Q-Chem manual says that default purecart is 1111.

I ran these four tests with CH4:

  • Running a calculation with 6-31G* with purecart 1111 explicitly in $rem gives 22 basis functions (as expected).
  • Running a calculation with 6-31G* with no purecart keyword (default purecart) gives 23 basis functions (and 23 MOs, so not a simple printing error). Must be using 6d functions instead of 5d?
  • Running a calculation with cc-pVDZ with purecart 1111 explicitly in $rem gives 34 basis functions (as expected).
  • Running a calculation with cc-pVDZ with no purecart keyword (default purecart) also gives 34 basis functions.

Conclusion: default purecart is 1111 for Dunning, but 2222 for Pople? Or some other weird thing going on? Any help appreciated!

The four inputs I used are below:

Coordinates:

$molecule
0 1
C 0.0000000 0.0000000 0.0000000
H 1.1094008 0.0000005 -0.0000021
H -0.3697982 0.0206895 1.0457492
H -0.3698007 -0.9159899 -0.5049556
H -0.3698018 0.8952997 -0.5407914
$end

Pople with purecart 1111:

$rem
BASIS = 6-31G*
PURECART = 1111
GUI = 2
METHOD = HF
PRINT_GENERAL_BASIS = true
$end

Pople with default purecart:

$rem
BASIS = 6-31G*
GUI = 2
METHOD = HF
PRINT_GENERAL_BASIS = true
$end

Dunning with purecart 1111:

$rem
BASIS = cc-pVDZ
PURECART = 1111
GUI = 2
METHOD = HF
PRINT_GENERAL_BASIS = true
$end

Dunning with default purecart:

$rem
BASIS = cc-pVDZ
GUI = 2
METHOD = HF
PRINT_GENERAL_BASIS = true
$end

#2

I just realized my exact question is answered a few posts down by Anna! It’s not all Pople vs. Dunning, apparently, just 6-31G* uses 6d functions by default.

#3

Also some of the Dunning basis sets may still be 2111 due to earlier lack of support for h spherical harmonics.

#4

This is only true for Q-Chem prior to 5.2. Since Q-Chem 5.2 release, all functions are pure by default except for Pople’s basis sets, which are 6d.