Multiwfn forum

mirzanejad

Member

Registered: 2021-09-14

Posts: 4

Electron Spin Density analysis: unrestricted orbitals from molden file

Hi,

I am trying to do spin electron density analysis using Multiwfn 3.8(dev) and it turns out that unrestricted open-shell KS orbitals (al least) through molden format are not appreciated for this purpose (FYI, the molden file is generated by Molpro package for a quintet spin state). However, fchk file with unrestricted open-shell KS orbitals generated by Q-Chem (and molden file containing ROHF orbitals generated by Molpro) does not show any issues in the electron spin density analysis at all. The modification of wavefunction has been tried, but as is mentioned in the manual, only restricted open-shell orbitals are supported in the current version.

Now, I am wondering is there anyway I can do spin electron density analysis using the current version of Multiwfn package by (slightly) modifying the molden file that has UKS orbitals?

Thanks

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,773

Website

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

Multiwfn fully supports analyzing unrestricted open-shell wavefunction recorded in .molden file.
I don't use Molpro very often, but the .molden file produced by UKS calculation of Molpro should be able to normally load into Multiwfn. If you can provide me the file I will look at it.

mirzanejad

Member

Registered: 2021-09-14

Posts: 4

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

Thanks very much for the prompt reply.

The file size is around 27.5 MB and seems this size is out of the acceptable file-size limit defined in "Attachments" interface. A few seconds after pushing the "Upload a file" bottom, the original page loads automatically without any action. I tried a smaller text file and the text file was loaded successfully, but no progress for the molden file. Is it a problem from my side or the file-size limit is exceeded in the molden file?

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,773

Website

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

Please compress the file, the size will be significantly reduced. It is best to share the file via netdisk, because the space of this forum is very limited.

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,773

Website

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

I just did a test, indeed, the molden file exported by Molpro during UKS calculation doesn't contain spin orbitals. I found Molpro automatically generates unrestricted natural orbitals (UNOs) and write them the molden file, and hence the spin information is fully missing. If there is a way to avoid generating UNOs, perhaps the spin orbitals can be finally dumped to the molden file, but I don't know how to realize. I suggest you ask Molpro developer, if there is a way to successfully export spin orbitals to molden, you are welcome to share it here.

mirzanejad

Member

Registered: 2021-09-14

Posts: 4

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

In the input file, the following lines will generate spin orbitals in molden format (in Cartesian form):

put, molden, filename1.molden; orbital, type=alpha
put, molden, filename2.molden; orbital, type=beta

sobereva

Tian Lu (Multiwfn developer)

From: Beijing

Registered: 2017-09-11

Posts: 1,773

Website

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

In the input file, the following lines will generate spin orbitals in molden format (in Cartesian form):

put, molden, filename1.molden; orbital, type=alpha
put, molden, filename2.molden; orbital, type=beta

Well, then you can manually merge the two molden files, analysis of spin density will be feasible.

mirzanejad

Member

Registered: 2021-09-14

Posts: 4

Re: Electron Spin Density analysis: unrestricted orbitals from molden file

In the input file, the following lines will generate spin orbitals in molden format (in Cartesian form):

put, molden, filename1.molden; orbital, type=alpha
put, molden, filename2.molden; orbital, type=beta

Well, then you can manually merge the two molden files, analysis of spin density will be feasible.

Actually, a bit modification was required in my case because there was more basis functions (due to Cartesian format) than total number of orbitals. This inconsistency (I suspect) led to a wrong assignment in number of electrons as well as net charge and multiplicity when Multiwfn loaded the input file. Therefore, I had to convert orbitals from Cartesian to spherical format. After that, I enjoyed doing the analysis!