Molcas Forum
Support and discussions for Molcas and OpenMolcas users and developers
You are not logged in.
- Topics: Active | Unanswered
Announcement
Welcome to the Molcas forum.
You can choose an avatar and change the default style by going to "Profile" → "Personality" or "Display".
#1 2023-05-10 23:13:53
- rdjiij
- Member
- Registered: 2023-03-15
- Posts: 13
HOP along trajectory for large dE values
Dear Molcas Users,
I am having an hard time understanding why a QM/MM trajectory that I am running is HOPPING from S1 to S0 for a large energy gap (ca. 0.15 a.u.).
The trajectory is propagated at the SA3-CASSCF(10,10)//6-31G* level, from S2.
It is somewhat well-behaved for ~90 fs, when I then get this jump and start to observe an unphysical behavior.
I get this verbose printout which I cannot fully understand:
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
&SURFACEHOP
launched 8 MPI processes, running in PARALLEL mode (work-sharing enabled)
available to each process: 12 GB of memory, 1 thread?
master pid: 110906
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()
------------------------------------------
TULLY ALGORITHM
------------------------------------------
Using CI vector product for WF overlap
Did the density matrix exists? T
Did the Pre-coefficients array exists? T
Did the Pre-Pre-coefficients array exists? T
Density Matrix elements (i=1..#states):
(-4.404987183771590E-002,7.417476883166522E-004)
(1.21535067373835,0.686899466180354) (-25.4300581403399,25.4563425055070)
(1.21636489282894,-0.724701629977835)
(0.375683889552587,-2.691481012497208E-005)
(-0.680507343805885,-0.204158195646083)
(-26.3806754620796,-25.1597184795051) (-0.681771259631409,0.201821529128818)
(0.668365982285127,-7.148328792525604E-004)
Using CI vector products for sign correction/root ordering
Using CI vector product to calculate D Matrix
From input Random is Fixed at: 0.250000000000000
Normalized by Substep number: 1.256281407035176E-003
executing the integration:
Time Step is: 41.3223140000000
Substeps are: 199
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 3
Following Tully, should hop from 2 to 1
this root has an energy lower than the total -742.120777886046
(thus, permitted)
Ediffcheck is: -0.161708711590791
lower than the threshold: 1.797693134862316E+308
temproot set to: 1
Gnuplot: 5.892533827017844E-002 0.334403724788879
0.606670936940940 -742.319445766803 -742.157955839222
-742.117980466388 -742.319445766803
HOP ALLOWED
+----------------------------------------------------------------------------------------------------------------------+
| |
| A HOP event is detected!|
| |
| From state: 2 To state: 1 |
| |
+----------------------------------------------------------------------------------------------------------------------+
--- Stop Module: surfacehop at Wed May 10 21:03:35 2023 /rc=_RC_ALL_IS_WELL_ ---
*** files: xmldump
saved to directory /home/barneschil/BENCHMARK-ASEC-BOX/cinnamic_acid/ASEC_GAFF/Step_0/calculations/optscf/QMMM/SA2/cas_10_10/sa5_correction/fc_trj/from_s2_sa3
--- Module surfacehop spent 6 seconds ---
*** symbolic link created: INPORB -> M2_10_10_neu.RasOrb
--- Start Module: mclr at Wed May 10 21:03:42 2023 ---I checked the active space and it looks fine before and after this HOP. Might this have to do with the NAC?
Any suggestion is highly appreciated.
Kindest regards
Offline
#2 2023-05-11 08:24:14
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,202
Re: HOP along trajectory for large dE values
It might be a limitation/flaw of "using CI vector product for WF overlap". Try with a recent OpenMolcas version (23.02 or later), it should compute more accurate overlaps with RASSI. As far as I remember, SURFACEHOP does not use NACs.
Offline
#3 2023-05-11 08:56:26
- rdjiij
- Member
- Registered: 2023-03-15
- Posts: 13
Re: HOP along trajectory for large dE values
Dear Ignacio,
thank you very much for replying. I have the exact same problem if I use RASSI for wavefunction overlaps (however I am using version 22.10).
Do you think it could be beneficial to install 23.02 or later?
Thanks again!
Offline
#4 2023-05-11 10:00:16
- Ignacio
- Administrator
- From: Uppsala
- Registered: 2015-11-03
- Posts: 1,202
Re: HOP along trajectory for large dE values
I forgot the RASSI overlaps were available earlier... Maybe this has an effect: https://gitlab.com/Molcas/OpenMolcas/-/ … quests/579
Offline
#5 2023-05-11 12:05:09
- rdjiij
- Member
- Registered: 2023-03-15
- Posts: 13
Re: HOP along trajectory for large dE values
Dear Ignacio,
Thank you very much, I will give it a try 
Offline
Last refresh: Today 22:39:06