# Problem with basis set inline format

[ Molcas user's WWWBoard ]

Posted by Matthias Mehring on February 14, 2008 at 14:30:32:

Hi,
I would like to set up calculation of small moleculues on a metaloxide surface, in my case CO on TiO2.
I do have lots of experiences with that, the last project dealt with the adsorption of CO on NiO.
The metaloxide surface is represented by a cluster model which is embedded in a PC field.
The basis set have to be relatively small, due to the huge system. So far, I chose for the Ti a basis set of Wachters + additional f-functions. The basis for the CO is generated of a set of Huzinaga + additional functions.
Hence, the basis has to be input inline and I used following scheme (only part of the SEWARD and system is shown):

#################################
*****************************
&SEWARD &END

Title
Check

*XFIELD
*@/home/mehring/TiO2/XFIELD_C2v_Ti18O18.inp

Verbose

Test

Symmetry
x y

********************************************************************
Basis set
Ti..... / inline
22.0 3
* S-type functions
14 8
206082.00000000
31226.80000000
7199.32000000
2048.75000000
670.79000000
243.65000000
95.92500000
39.81010000
12.22050000
5.00882000
1.28569000
0.51280600
0.08557600
0.03330200
0.00025000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00193000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00988000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.03994000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.12737000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.29981000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.41825000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.23582000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 1.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 1.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 1.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 1.00000000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 1.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 1.00000000
* P-type functions
11 6
1264.70000000
301.23000000
96.97770000
36.37270000
14.78140000
6.27465000
2.47878000
1.01618000
0.39816200
0.15234000
0.05108000
0.00214000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.01725000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.08117000 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.24143000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.42631000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.35714000 0.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 1.00000000 0.00000000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.57709000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.29986000 0.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 1.00000000 0.00000000
0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 1.00000000
* D-type functions
6 4
25.99240000
7.08634000
2.34871000
0.80019800
0.26204900
0.07200000
0.02253000 0.00000000 0.00000000 0.00000000
0.11961000 0.00000000 0.00000000 0.00000000
0.32700000 0.00000000 0.00000000 0.00000000
0.00000000 1.00000000 0.00000000 0.00000000
0.00000000 0.00000000 1.00000000 0.00000000
0.00000000 0.00000000 0.00000000 1.00000000
* F-type functions
3 1
1.39353080
0.49975220
0.21399980
0.17378570
0.59733810
0.39293960
Ti1 0.00000 -5.20581 0.00000
Ti2 0.00000 0.00000 0.00000
End of basis

*********************************************************************************
End of input

#############################

As shown, for the Titanium I used a basis like
(14s11p6d3d) -> [8s6p4d1f].
All the test calculations do finish fine, without any non-zero returncode.
The thing is, that the contraction scheme doesn't appear in the output, i.e., no contraction occurs, the total number of basis (14s11p6d3d)functions is generated by the primitives, in this case (14s11p6d3d) = 294 bf instead of requested 159 contracted bf generated by the [8s6p4d1f] scheme.
The accordingly SEWARD output looks as follows:

############################
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
MOLCAS executing module SEWARD with 1400 MB of memory
at 13:59:36 Thu Feb 14 2008
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

Basis Set 1 Label: Ti.....
Basis set is read from library:inline

*********************************************
* TEST: SEWARD will only process the input! *
*********************************************

****************************************************************************************
* *
* Check *
* *
****************************************************************************************

Integrals are discarded if absolute value <: 0.10E-13
Integral cutoff threshold is set to <: 0.10E-15

--- Group Generators ---
Reflection in the yz-plane
Reflection in the xz-plane

Character Table for C2v

E s(yz) s(xz) C2(z)
a1 1 1 1 1 z
b1 1 -1 1 -1 x, xz, Ry
b2 1 1 -1 -1 y, yz, Rx
a2 1 -1 -1 1 xy, Rz, I

Basis set label:Ti..........

Valence basis set:
==================
Associated Effective Charge 22.000000 au
Associated Actual Charge 22.000000 au
Shell nPrim nBasis Cartesian Spherical Contaminant
s 14 14 X
p 11 11 X
d 6 6 X
f 3 3 X
#######################################

This indicates that no contraction has occured, resulting in huge number of bf, especially in the whole system.
I tried differnt basis sets, and also other atoms in the system, but it didn't change anything.
In my opinion, the input is written correctly, no error was given in the output.
I'm using molcas7.0 on a linux_x86_64 platform and I compiled it in 64bit mode. The verification after compiling was ok.

Before, I was using molcas64.sp1. I tried the SAME input file using, the 'old' binaries of molcas64.sp1, and with this version, the basisfunctions were generated correctly, as shown here:
####################################

^^^^^ M O L C A S
^^^^^^^ version 6.4 patchlevel 142
^^^^^ ^^^^^^^
^^^^^^^ ^^^ ^^^
^^^^^^^ ^^^^ ^^^
^^^ ^^^ ^^^^ ^^^
^^^ ^^^^ ^^^ ^^^
^^^ ^^ ^^^^^ ^^ ^^^^
^^^^ ^^^^ ^ ^^^
^ ^^^ ^^^^ ^^^^ ^
^ ^^^^ ^^ ^^^^ ^
^^^^^ ^^^^^ ^^ ^^^^^ ^^^^^
^^^^^^^^ ^^^^^ ^^^^^ ^^^^^^^
^^^^^^^^^^^ ^^^^^^ ^^^^^^^ ^^^^^^^^^^^
^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^^^
^^^^^^^^^^^^^ ^^^^ ^^^ ^^^ ^^^^
^^^^^^^^^^^^^ ^^^ ^^^^
^^^^^^^^^^^^^ ^^^^^^^^^^^^ ^^ ^^^^
^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^ ^^ ^^^^
^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ ^^ ^^^^
^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ ^^ ^^^^ ^^^^^ ^^^ ^^^^^^
^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^^^ ^^ ^^^^ ^^^^^^^ ^^^ ^^^ ^^^
^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^^^ ^ ^^^^ ^^^ ^^ ^^^ ^^ ^^
^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ ^^^^ ^^ ^^ ^^ ^^^^^
^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^ ^ ^^^^ ^^ ^^ ^^ ^^^^^^
^^^^^^^^ ^^^^^^^^^^^^^^^^ ^ ^^^^ ^^ ^^^^^^^ ^^^^
^^^^^^ ^^^ ^^^^^^^^^^^^ ^ ^ ^^^^ ^^^ ^^^ ^^^^^^^ ^^ ^^
^^^^^^ ^^ ^ ^^^^ ^^^^^^^ ^^ ^^ ^^^ ^^^
^^^^^^^^^^^^ ^^^^^^ ^ ^^^^^ ^^ ^^ ^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^
^^^^^^

Permission is hereby granted to use
but not to reproduce or distribute any part of this
program. The use is restricted to research purposes only.
Lund University Sweden, 2006.

For the author list and the recommended citation consult section 1.5 of the MOLCAS user's guide.

*********************************** L I N K I N F O R M A T I O N ***********************************

--- Start Module: seward at Thu Feb 14 14:13:09 CET 2008
Input File: autoinput.seward.1 ---

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
MOLCAS executing module SEWARD with 1400 MB of memory
at 14:13:09 Thu Feb 14 2008
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

Basis Set 1 Label: Ti.....
Basis set is read from library:inline

*********************************************
* TEST: SEWARD will only process the input! *
*********************************************

****************************************************************************************
* *
* Check *
* *
****************************************************************************************

--- Group Generators ---
Reflection in the yz-plane
Reflection in the xz-plane

Integrals are discarded if absolute value <: 0.10E-13
Integral cutoff threshold is set to <: 0.10E-15

Character Table for C2v

E s(yz) s(xz) C2(z)
a1 1 1 1 1 z
b1 1 -1 1 -1 x, xz, Ry
b2 1 1 -1 -1 y, yz, Rx
a2 1 -1 -1 1 xy, Rz, I

Basis set label:Ti.....

Valence basis set:
==================
Associated Effective Charge 22.000000 au
Associated Actual Charge 22.000000 au
Shell nPrim nBasis Cartesian Spherical Contaminant
s 14 8 X
p 11 6 X
d 6 4 X
f 3 1 X

##############################################

Here nBasis is different from nPrim, as the normal procedure.
As mentioned, I tried this with all atoms in the system using different basis sets. Even if I don't input the basis Inline, but from the basis library, say Ti.ano-s.Pierloot.17s12p9d4f.8s7p7d4f., the contraction doesn't occur but the nBasis is identical to the number of primitves.

May it be a bug in the program ?
Can you give a hint to solve the problem ?!
Thanks a lot!

Best regards,
Matthias

Name:
E-Mail:

Subject:

if B is 1s22s22p1, what is Li?

Passfield: