Input file for NTO calculation

lilirass · 2022-02-11 14:39:16

I want to write a plain text file containing configuration coefficients of excited states as the input file for NTO calculation. I use CP2K for TDDFT calculation.
I checked the Gaussian output file and tried to create an identical format. But I received this error:

forrtl: severe (59): list-directed I/O syntax error.
Image PC Routine Line Source
Multiwfn.exe 00007FF625A4BC7F Unknown Unknown Unknown
Multiwfn.exe 00007FF625A06F1C Unknown Unknown Unknown
Multiwfn.exe 00007FF62532BAC4 Unknown Unknown Unknown
Multiwfn.exe 00007FF62534BB63 Unknown Unknown Unknown
Multiwfn.exe 00007FF6253135C8 Unknown Unknown Unknown
Multiwfn.exe 00007FF6255E8808 Unknown Unknown Unknown
Multiwfn.exe 00007FF6263C2B4E Unknown Unknown Unknown
Multiwfn.exe 00007FF62668FCD4 Unknown Unknown Unknown
KERNEL32.DLL 00007FF92EF07034 Unknown Unknown Unknown
ntdll.dll 00007FF930242651 Unknown Unknown Unknown

You can find some versions of my input file in the following. Can you please help me to understand the problem?
I tried to find a sample of a plain text file containing configuration coefficients of excited states in the examples folder, but I could not find it. Can you please guide me to the right path to find a sample of this input file?
Best regards,
Lili

Version 1:
Excited State 1: Singlet-A 2.07774 eV 596.72 nm f=0.000 <S**2>=0.000
598A -> 602A 0.454805
600A -> 603A 0.416875
598B -> 602B -0.298488
599B -> 603B -0.250645
597B -> 603B -0.070004
600B -> 601B -0.676085

Excited State 2: Singlet-A 2.07792 eV 596.67 nm f=0.000 <S**2>=0.000
598A -> 601A 0.496762
599A -> 602A 0.435484
596A -> 602A -0.079478
599B -> 601B 0.561757
600B -> 603B 0.468381
597B -> 601B 0.120320

Version 2:
Excited State 1: Singlet-A 2.07774 eV 596.72 nm f=0.000 <S**2>=0.000
600B -> 601B -0.676085
598A -> 602A 0.454805
600A -> 603A 0.416875
598B -> 602B -0.298488
599B -> 603B -0.250645
597B -> 603B -0.070004

Excited State 2: Singlet-A 2.07792 eV 596.67 nm f=0.000 <S**2>=0.000
599B -> 601B 0.561757
598A -> 601A 0.496762
600B -> 603B 0.468381
599A -> 602A 0.435484
597B -> 601B 0.120320
596A -> 602A -0.079478

Version 3:
Excited State 1 1 2.07774 // Label, index, multiplicity and excitation energy (eV)
600 -> 601 -0.676085 // MO pairs and configuration coefficients
598 -> 602 0.454805
600 -> 603 0.416875
598 -> 602 -0.298488
599 -> 603 -0.250645
597 -> 603 -0.070004
// Use a blank line to separate each excited state
Excited State 1 1 2.07792
599 -> 601 0.561757
598 -> 601 0.496762
600 -> 603 0.468381
599 -> 602 0.435484
597 -> 601 0.120320
596 -> 602 -0.079478

sobereva · 2022-02-11 17:22:37

Please provide me your .molden file of CP2K, I will check. You can compress it and then send to me via E-mail, or paste download link here.
If possible, please also provide me your CP2K input and output file.

In your version 3, please remove // and all text after it, this content may cause trouble.

lilirass · 2022-02-11 18:56:39

Thank you, I sent the files via email.

sobereva · 2022-02-12 06:33:54

Your file is incorrect, it should be

 Excited State   1      0     2.07774
       600B -> 601B        -0.676085
       598A -> 602A         0.454805
       600A -> 603A         0.416875
       598B -> 602B        -0.298488
       599B -> 603B        -0.250645
       597B -> 603B        -0.070004

 Excited State   2      0     2.07792
       599B -> 601B         0.561757
       598A -> 601A         0.496762
       600B -> 603B         0.468381
       599A -> 602A         0.435484
       597B -> 601B         0.120320
       596A -> 602A        -0.079478

Note that if you just want to obtain eigenvalues of NTO pairs, using the Multiwfn_DFT.molden file you sent me is OK. However, if you want to obtain NTO orbitals, your .molden file cannot be used, because this file only records occupied orbitals. In order to generate NTO orbital wavefunctions, all unoccupied orbitals are also needed.

sobereva · 2022-02-14 04:41:23

Now, Multiwfn has supported directly loading electronic excitation information from CP2K TDDFT output file, you no longer need to manually prepare the plain text file, see http://sobereva.com/wfnbbs/viewtopic.php?pid=2254

lilirass · 2022-02-14 16:32:39

Thank you, It is awesome.
I had some questions, but I think my first and second questions are solved with this new function.
1- Can I give the information of a specific number of occupied and unoccupied orbitals to be considered in NTO calculation? Or should the molden file have the information of all occupied and all unoccupied orbitals?
2- Is there any function in Multiwfn to convert a .MOLog file (a file with the information of the occupied and unoccupied orbitals in Cp2k) to a .molden file?
3- Why did you assign the multiplicity equal to 0 instead of 1 in the input file?
I really appreciate your help.
Best regards,
Lili

sobereva · 2022-02-14 21:36:58

Dear Lili,

1 This is not supported. I suggest always let .molden file record all orbitals, simply add "ADDED_MOS 99999" in $SCF section.

2 There is no such a function.

3 Your system is an open-shell system, and thus the reference state of TDDFT is UKS wavefunction, in this case the solved excited states are not pure spin states, in other words, the spin multiplicity of the solved excited states is not defined. In this case the spin multiplicity should be set to 0 to let Multiwfn know this fact.

Best regards,

Tian

lilirass · 2022-02-14 22:53:20

Dear Tian,
Thank you for sending me this information. It was exactly what I needed.

lilirass · 2022-02-27 17:05:39

Dear Tian,
I could create a .mwfn file for the first excited state. As you see in the pictures, the NTO eigenvalues are consistent with CP2K results.
But when I try to visualize them, nothing is shown, specifically for LUMO-beta (-601 orbital(2785)). I changed isosurface quality to 1500k points (perfect quality), but still, there is no electron density (even with 0.002 isovalue)
After Booting up Multiwfn, I upload S1_log.mwfn file, and then I choose 0 (Show molecular structure and view orbitals) from the main menu. But I cannot see the electron density as I expect.
When I visualize molecular orbital cube files of the first excited state:
600B -> 601B
598A -> 602A
600A -> 603A
598B -> 602B
599B -> 603B
597B -> 603B
All follow the same pattern (Electron density is around oxygen atoms in HOMO, and it goes to Fe atoms in LUMO). But I cannot see the same pattern in NTOs. And there is even no electron density on some NTOs.
Can you please help me to visualize NTOs correctly?
Sincerely,
Lili

sobereva · 2022-02-27 17:56:12

Dear Lili,

In order to visualize NTO generated by Multiwfn for CP2K, please note two points:
(1) The .molden file for NTO analysis must contains all virtual orbitals involved in the printed configurations.
(2) After loading the .mwfn file containing NTOs generated by Multiwfn, in main function 0 you should use "Orbital info." - "Show all" option in the menu bar to check eigenvalue of orbitals (NTO eigenvalues correspond to energy in a.u. in this case), and then select the orbitals corresponding to NTOs with large contribution to visualize them.

If you are sure that you have paid attention to the above two points, please send relevant files to my E-mail, I will check reason.

Best regards,

Tian

sobereva · 2022-03-01 18:22:21

Dear Lili,

I have received your files.

There should be no problem in your .mwfn file. It is important to note that currently Multiwfn is unable to correctly plot isosurface map for non-orthogonal cell. In order to correctly visualize the NTOs generated by Multiwfn for your system, after loading the .mwfn file containing NTOs, please go to subfunction 3 of main function 200. Then you can input index of an orbital to export its cube file of orbital wavefunction. Then you can use e.g. VESTA and VMD to visualize the isosurface map of NTO based on the cube file.

It is worth to note that I frequently found NTO is basically useless for periodic systems, and this is also true for your system. As you can see from the eigenvalues of NTOs, there is no dominant NTO pair, therefore you still need to visualize many NTOs to examine the character of the excitation.

I strongly suggest you using hole-electron analysis to characterize the excitations. Hole-electron analysis works well for any kind of system, in all cases the excitation can be represented as hole → electron transition. Very recently, I wrote a blog article carefully illustrated how to use Multiwfn to study electron excitation for CP2K TDDFT calculation, including creating input file, plotting spectrum, visualizing MOs, visualizing hole and electron distributions, visualizing transition density and performing NTO analysis, see http://sobereva.com/634 (written in Chinese, please use Google translator). I hope this article is useful for you.

Best regards,

Tian

lilirass · 2022-03-04 13:43:59

Thanks for your response and guidance. I am currently working on hole-electron analysis and its theory. I will let you know if I have questions about this analysis.
Thank you again,
Lili

Multiwfn forum

#1 2022-02-11 14:39:16

Input file for NTO calculation

#2 2022-02-11 17:22:37

Re: Input file for NTO calculation

#3 2022-02-11 18:56:39

Re: Input file for NTO calculation

#4 2022-02-12 06:33:54

Re: Input file for NTO calculation

#5 2022-02-14 04:41:23

Re: Input file for NTO calculation

#6 2022-02-14 16:32:39

Re: Input file for NTO calculation

#7 2022-02-14 21:36:58

Re: Input file for NTO calculation

#8 2022-02-14 22:53:20

Re: Input file for NTO calculation

#9 2022-02-27 17:05:39

Re: Input file for NTO calculation

#10 2022-02-27 17:56:12

Re: Input file for NTO calculation

#11 2022-03-01 18:22:21

Re: Input file for NTO calculation

#12 2022-03-04 13:43:59

Re: Input file for NTO calculation

Board footer