Issue with excitation energies (TDDFT) when using the wB97XD functional

Raphaël Rullan

unread,

Jan 13, 2025, 12:15:16 PM1/13/25

to cp2k

Dear CP2K users,

Sorry for bothering you again. I observed a particularly strange issue while running a TDDFT calculation. The calculation seems to be running just fine, the SCF part converged and the energies of the orbitals seem fine. The TDDFT part start, the guess seems consistent with the HOMO-LUMO gap and it starts to converge. At some point a new excitation appears with a negative energy (see below and .out). I don't understand why, and how I could fix this. I guess maybe the GS wavefunction is not fully converged but the code says so. Any idea ?

The example below is from a TDDFT calculation using the wB97-XD functional with the 2023 CP2K release. However, I also observed this issue with the PBE0 functional with a different version of CP2K (2022).

State Exc. energy (eV) Convergence (eV)

1 2.6610598 -3.1548E-01

2 2.6610618 -3.1548E-01

3 2.6610647 -3.1548E-01

4 3.9977385 -3.6060E-01

5 4.8783993 -5.6683E-01

6 4.8784003 -5.6683E-01

7 5.0410401 -5.2693E-01

8 5.0410451 -5.2693E-01

9 5.0410482 -5.2693E-01

10 5.4661866 -3.0007E-01

2 10573.8 2.0831E-02 0

State Exc. energy (eV) Convergence (eV)

1 2.6372645 -2.3795E-02

2 2.6376604 -2.3401E-02

3 2.6378627 -2.3202E-02

4 3.9549216 -4.2817E-02

5 4.6556179 -2.2278E-01

6 4.6556197 -2.2278E-01

7 4.9364090 -1.0463E-01

8 4.9401741 -1.0087E-01

9 4.9418617 -9.9187E-02

10 5.3915938 -7.4593E-02

3 9342.6 8.1871E-03 0

State Exc. energy (eV) Convergence (eV)

1 2.6348592 -2.4052E-03

2 2.6362756 -1.3847E-03

3 2.6365731 -1.2896E-03

4 3.9509578 -3.9639E-03

5 4.5793122 -7.6306E-02

6 4.5793141 -7.6306E-02

7 4.9300251 -6.3839E-03

8 4.9328265 -7.3477E-03

9 4.9344891 -7.3726E-03

10 5.3743071 -1.7287E-02

4 9335.0 2.8042E-03 0

State Exc. energy (eV) Convergence (eV)

1 -68.5095683 -7.1144E+01

2 2.6342572 -2.0185E-03

3 2.6345047 -2.0684E-03

4 2.6362695 -1.3147E+00

5 3.9507303 -6.2858E-01

6 4.5663465 -1.2968E-02

7 4.5663508 -3.6367E-01

8 4.9201774 -1.2649E-02

9 4.9229758 -1.1513E-02

10 4.9321820 -4.4213E-01

Thank you very much,

Have a good day,

Best regards

Raphael Rullan

TDDFT_wB97XD.out

sulfur.xyz

TDDFT_wB97XD.inp

Message has been deleted

Raphaël Rullan

unread,

Feb 20, 2025, 4:45:33 AM2/20/25

to cp2k

Dear CP2K users,

Sorry for bothering you again with this topic but I did not find any answers yet despite the many attempts. Does anyone have an idea of what could be the problem?

I also would like to raise another issue I encountered while running RKS_TRIPLETS calculations. For PBE and PBE0 I have no issue but as soon as I try B3LYP or CAM-B3LYP it stopped with a scary error message (presented below). Does anyone encountered this issue and maybe have an idea about how to deal with it ? I don't think that the memory issue comes from my computer.

Thank you very much,

Best regards,

Raphael Rullan

==== backtrace (tid: 253750) ====

0 0x0000000000013140 __funlockfile() :0

1 0x0000000002e7078c __xc_MOD_xc_calc_2nd_deriv_analytical._omp_fn.90() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/xc.F90:2766

2 0x0000000000015076 GOMP_parallel() /applis/PSMN/debian11/Lake/build/GCCcore/12.3.0/system-system/gcc-12.3.0/stage3_obj/x86_64-pc-linux-gnu/libgomp/../../../libgomp/parallel.c:178

3 0x0000000000015076 GOMP_parallel_end() /applis/PSMN/debian11/Lake/build/GCCcore/12.3.0/system-system/gcc-12.3.0/stage3_obj/x86_64-pc-linux-gnu/libgomp/../../../libgomp/parallel.c:140

4 0x0000000000015076 GOMP_parallel() /applis/PSMN/debian11/Lake/build/GCCcore/12.3.0/system-system/gcc-12.3.0/stage3_obj/x86_64-pc-linux-gnu/libgomp/../../../libgomp/parallel.c:179

5 0x0000000002ea830d __xc_MOD_xc_calc_2nd_deriv_analytical() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/xc.F90:2766

6 0x000000000274b877 __qs_tddfpt2_operators_MOD_tddfpt_apply_xc_analytic() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_operators.F90:378

7 0x000000000274d82a __qs_tddfpt2_operators_MOD_tddfpt_apply_xc() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_operators.F90:248

8 0x0000000002615a53 __qs_tddfpt2_fhxc_MOD_fhxc_kernel() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_fhxc.F90:261

9 0x00000000024a9f4e __qs_tddfpt2_eigensolver_MOD_tddfpt_davidson_solver() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_eigensolver.F90:378

10 0x00000000024a9f4e __qs_tddfpt2_eigensolver_MOD_tddfpt_davidson_solver() /Xnfs/chimie/debian11/cp2k/2023.2/src/qs_tddfpt2_eigensolver.F:769

11 0x00000000021b3a80 __qs_tddfpt2_methods_MOD_tddfpt_energies() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_methods.F90:952

12 0x00000000021b7ea2 __qs_tddfpt2_methods_MOD_tddfpt() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_tddfpt2_methods.F90:339

13 0x00000000016748b3 __qs_energy_utils_MOD_qs_energies_properties() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_energy_utils.F90:157

14 0x0000000000d78a53 __qs_energy_MOD_qs_energies() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/qs_energy.F90:136

15 0x0000000000bc66ff __force_env_methods_MOD_force_env_calc_energy_force() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/force_env_methods.F90:256

16 0x0000000000803cfb __cp2k_runs_MOD_cp2k_run.constprop.0() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/cp2k_runs.F90:355

17 0x0000000000806793 __cp2k_runs_MOD_run_input() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/cp2k_runs.F90:997

18 0x00000000007f9023 MAIN__() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/cp2k.F90:379

19 0x0000000000491cad main() /Xnfs/chimie/debian11/cp2k/2023.2/obj/local/psmp/cp2k.F90:44

20 0x0000000000023d7a __libc_start_main() ./csu/../csu/libc-start.c:308

21 0x00000000007f750a _start() ???:0

=================================

==== backtrace (tid: 253743) ====

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Jürg Hutter

unread,

Feb 21, 2025, 10:14:49 AM2/21/25

to cp...@googlegroups.com

Hi

can you provide a sample input file that triggers the error?

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Raphaël Rullan <raphael...@gmail.com>
Sent: Thursday, February 20, 2025 10:45 AM
To: cp2k
Subject: [CP2K:21180] Re: Issue with excitation energies (TDDFT) when using the wB97XD functional

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/a3ba12d6-731d-4fc3-8cdc-fcbe7555430bn%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/a3ba12d6-731d-4fc3-8cdc-fcbe7555430bn%40googlegroups.com?utm_medium=email&utm_source=footer>.

Raphaël Rullan

unread,

Feb 21, 2025, 10:31:14 AM2/21/25

to cp2k

Hi,

For the first issue about the negative values the input is provided in the first post.

For the RKS_TRIPLETS issue please find attach the input and the necessary files (I can't upload the wfn file as it is too big, it is just a PBE calculation).

Thank you very much,

Best regards.
Raphael Rullan

TDDFT_wB97XD.inp

coord.xyz

BASIS_MOLOPT

GTH_POTENTIALS

Jürg Hutter

unread,

Feb 21, 2025, 11:09:00 AM2/21/25

to cp...@googlegroups.com

Hi
Works with the latest version of CP2K (I replaced your system with a small molecule
for test purposes).
Check if your setup works for a small test molecule. If not, switch to the latest version.

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Raphaël Rullan <raphael...@gmail.com>

Sent: Friday, February 21, 2025 4:31 PM
To: cp2k
Subject: Re: [CP2K:21190] Re: Issue with excitation energies (TDDFT) when using the wB97XD functional

To view this discussion visit https://groups.google.com/d/msgid/cp2k/bc6d92eb-63d6-442a-a046-a99defb0ad7bn%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/bc6d92eb-63d6-442a-a046-a99defb0ad7bn%40googlegroups.com?utm_medium=email&utm_source=footer>.

Reply all

Reply to author

Forward