Problems with &surfacehop - need help

angeloraymondrossi · 2024-06-23 16:08:23

Hello Everyone.

I need help in determining the problem for the following input file for the dication of ethanol (i.e. H3C-CH2-OH 2+) which includes the use of the SURFACEHOP module.

&Gateway

COORD
 9
 angstrom
 C         0.436000        0.311000        0.000000
 O         1.305000       -0.814000        0.000000
 H         2.197000       -0.459000       -0.000000
 H         0.646000        0.923000        0.903000
 H         0.646000        0.923000       -0.903000
 C        -1.027000       -0.138000       -0.000000
 H        -1.699000        0.747000        0.000000
 H        -1.251000       -0.747000        0.903000
 H        -1.251000       -0.747000       -0.903000


basis= 6-31G*
group=nosym

/*
basis= def2-TZVP
basis= aug-cc-pVTZ
basis= 6-31G*
*/

>> EXPORT MOLCAS_MAXITER=500

>> DOWHILE

&Seward

>> IF ( ITER = 1 )

&SCF
 TITLE= EtOH
 OCCUPIED= 13
 ITERATIONS= 100

&RASSCF
/* LUMORB
 FileOrb= $Project.GssOrb */
 Symmetry= 1
 Spin= 1
 Inactive= 9
 nActEl= 6 0 0
 RAS2=8
 CIroot= 3 3 1

>> COPY $Project.JobIph $Project.JobOld

>> ENDIF
&CASSCF
 JOBIPH; CIRESTART
 Symmetry= 1
 Spin= 1
 Inactive= 9
 nActEl= 6 0 0
 RAS2=8
 CIroot= 3 3 1
 MDRLXR= 2
 XMULTISTATE= 1

>> COPY $Project.JobIph $Project.JobOld

&surfacehop
 TULLY
 DECOHERENCE = 0.1
 PSUB

&ALASKA

&Dynamix
 VELVer
 DT= 20.65
 VELO= 3
 TEMP= 300.0
/*  HOP= 9 */

>> END DO

OpenMolcas stops after a few interations:

  time          Epot                 Ekin                 Etot
    0.00  -1.530508137100E+02   2.760151566428E-02  -1.530232121943E+02  -1.530885597700E+02  -1.530720790499E+02  -1.530508137100E+02
   20.65  -1.530544650115E+02   3.121152103916E-02  -1.530232534904E+02  -1.530867116841E+02  -1.530730103514E+02  -1.530544650115E+02
   41.30  -1.530593079243E+02   3.597315678409E-02  -1.530233347675E+02  -1.530835458465E+02  -1.530739831665E+02  -1.530593079243E+02
   61.95  -1.530642603968E+02   4.082733391938E-02  -1.530234330629E+02  -1.530790879258E+02  -1.530742708292E+02  -1.530642603968E+02
   82.60  -1.530674639400E+02   4.421715396274E-02  -1.530232467861E+02  -1.530762453912E+02  -1.530707378189E+02  -1.530674639400E+02

and this persists after many attempts. There are two types of errors:

                                         RASSCF iterations: Energy and convergence statistics
                                         ----------------------------------------------------

      Iter CI   SX   CI       RASSCF       Energy    max ROT     max BLB   max BLB  Level Ln srch  Step   QN   Walltime
          iter iter root      energy       change     param      element    value   shift minimum  type update hh:mm:ss
 reading initial CI vectors from JOBOLD
--- Stop Module: rasscf at Sun Jun 23 10:42:22 2024 /rc=_RC_FLOATING_EXCEPTION_ ---
*** files: test.rasscf.h5 xmldump
    saved to directory /home/rossi/Downloads/workshop/OpenMolcas-projects/test2
--- Module rasscf spent 3 seconds ---

>>> END DO

or "there is an inconsistency found in the program"

I don't understand the convergence error! The program runs only a few iterations, and the output molden file has the same geometry as the input file.
Most of the molecules for surface hopping in the very recent literature involve neutral molecules, whereas I am calculating a dication with vastly different electron density, and that may influence the hopping algorithm. However, I understand that the sample provided with OpenMolcas involves a cation: H2N-CH2 +

I have tried many different modification of the input file, and nothing works. A few basis sets have been tried, but the 6-31G* is used as a test for obtaining results quickly.

Is the SURFACEHOP code robust?

The same results are obtained both with my laptop and university Linux cluster.

Here is the OpenMolcas info:

           *************************************************
           * pymolcas version py2.29                       *
           *   build c0481229b6f3f65ed2c83878be511d3d      *
           *   (after the EMIL interpreter by V. Veryazov) *
           *************************************************

configuration info

------------------
Host name: debian12 (Linux)
C Compiler ID: GNU
C flags: -std=gnu99 -fopenmp
Fortran Compiler ID: GNU
Fortran flags: -fno-aggressive-loop-optimizations -cpp -fdefault-integer-8 -fopenmp
Definitions: _MOLCAS_;_I8_;_LINUX_;_MOLCAS_MPP_;_GA_
Parallel: ON (GA=ON)

   ------------------------------------------------------------------------------
  |
  |             Project: test
  |      Submitted from: /home/rossi/Downloads/workshop/OpenMolcas-projects/test2
  |        Scratch area: /home/rossi/scratch/Molcas/test
  |     Save outputs to: /home/rossi/Downloads/workshop/OpenMolcas-projects/test2
  |              Molcas: /home/rossi/Downloads/workshop/OpenMolcas/build
  |
  | Scratch area is NOT empty
  |
  |      MOLCAS_DRIVER = /home/rossi/Downloads/workshop/OpenMolcas/build/pymolcas
  |      MOLCAS_MOLDEN = ON
  |      MOLCAS_NEW_WORKDIR = YES
  |      MOLCAS_NPROCS = 6
  |      MOLCAS_REDUCE_PRT = NO
  |      MOLCAS_SOURCE = /home/rossi/Downloads/workshop/OpenMolcas
  |      MOLCAS_STRUCTURE = 0
  |
   ------------------------------------------------------------------------------

Can someone please help me with this. I am sure that if you run the input file above, it will yield similar results and stop execution suddenly.

Thank you in advance for any hints, pointers, help you can provide.

Kind regards,

Angelo

angeloraymondrossi · 2024-06-23 18:58:05

Using RAS2= 6 in the above input still yielded the same problem. The job stops executing!

Here is a other specific error that can happen which I mentioned. Both errors don't happen in the same run.

.########################################################. 
.# Some internal inconsistency of the code was detected                     #. 
.########################################################.

I think there is a problem with the code, but I am not sure.

There are several very recent publications using the Molcas surfacehop code for neutral molecules.

I want it to work for me too!! Thanks again.

angeloraymondrossi · 2024-06-24 12:09:09

Good Morning.

It appears that

&Dynamix
VELVer
DT= 10.00
VELO= 3
TEMP= 300.0

will work. But there is one example in Molcas for a surfacehop/dynamix value of dt= 40. However, the default value is dt= 10.0

It is somewhat confusing.

Ignacio · 2024-06-24 14:05:16

If you have a large timestep, there may be large changes in the wavefunction/orbitals between timesteps and that will cause convergence problems.

Molcas Forum

Announcement

#1 2024-06-23 16:08:23

Problems with &surfacehop - need help

#2 2024-06-23 18:58:05

Re: Problems with &surfacehop - need help

#3 2024-06-24 12:09:09

Re: Problems with &surfacehop - need help

#4 2024-06-24 14:05:16

Re: Problems with &surfacehop - need help

Board footer