Molcas Forum

FedCpp

Member

Registered: 2020-12-07

Posts: 4

Relaxed scan with a dummy atom

Hello,
I'm performing a relaxed scan calculation at SA-CASSCF level using a custom basis set centered on the dummy atom.

The initial guess is in Cartesian coordinates and the dummy has been placed in the barycenter of positive charges.
The relaxed scan performs well and no errors are found as long as the dummy is in the initial position.

Here, my request: I would like to fix the cartesian coordinates of the dummy atom during the scan.

I used the keyword Constraints in the &Seward module:

Constraints
a1 = Angle C3 X C2
a2 = Angle C7 C2 C4
a3 = Angle C6 C3 C5
b1 = Bond C1 X
Values
a1 = 95.74791 Degrees
a2 = $ANGLE Degrees
a3 = $ANGLE Degrees
b1 = 2.02604 Angstrom
End of Constraints

But for some geometries the X dummy changed the position.

Alternatively, I already tried the keyword Fix in the &Seward module:

Constraints
f1 = Cartesian x X
f2 = Cartesian y X
f3 = Cartesian z X
a2 = Angle C7 C2 C4
a3 = Angle C6 C3 C5
Values
f1 = Fix
f2 = Fix
f3 = Fix
a2 = $ANGLE Degrees
a3 = $ANGLE Degrees
End of Constraints

In the &SLAPAF this is the error:

++ Constraints section
                                  CONSTRAINTS

********************************************************************************
F1 = CARTESIAN X X
F2 = CARTESIAN Y X
F3 = CARTESIAN Z X
A2 = ANGLE C7 C2 C4
A3 = ANGLE C6 C3 C5
VALUES
F1 = FIX
F2 = FIX
F3 = FIX
A2 = 37.110 DEGREES
A3 = 37.110 DEGREES
********************************************************************************


 *************************************************************
 * Values of the primitive constraints                       *
 *************************************************************
 F1       : x-component=     -0.0000/ bohr
 ###############################################################################
 ###############################################################################
 ###                                                                         ###
 ###                                                                         ###
 ###    ERROR:  Unrecognizable atom label                                    ###
 ###                                                                         ###
 ###                                                                         ###
 ###############################################################################
 ###############################################################################

Are there any suggestions?
Thanks in advance

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: Relaxed scan with a dummy atom

There were a couple of bugs that would affect your second case (which should be fixed soon: https://gitlab.com/Molcas/OpenMolcas/-/ … quests/419).

But the first case should work properly. At least it should do what the input asks: keep the C3 X C2 angle and the C1 X distance at the requested values. This, however does not keep the X atom fixed because, unless you have some external potential, the system is free to move and rotate. Actually, by default the translation and rotation degrees of freedom are removed, so the molecule as a whole doesn't move, but individual atoms do.

(By the way, you can use "Fix" for your "a1" and "b1" constraints too.)