Dear Mehrzad, Yes, a molecule doesn't have a bandstructure, only a few discrete energy levels. That's why I said a possible COOP should be quite easy to interpret. Possible COOP because the website I referred you to explicitly mentions the theory is applicable to crystals and molecules, nevertheless, I don't know about the Siesta implementation. Why don't you give it a try? I never did it on molecules myself. If you have the WFSX and HSX files from your siesta run, it should be just fine, I guess.
Good luck, Nicolas PS : Please don't send the same question 3 times in different format. And allow for people some time to answer... On Tue, Sep 25, 2012 at 7:37 AM, Mehrzad Sasanpoor <[email protected]>wrote: > > Dear Mr. Nicolas > thank you so much for your best reply,I think I got it. but isolated > molecule doesn't have the band structure! I did all of my programs for one > gama point before. > Do I say correct? do you think this method is appropriate for this case? > > thank you so much for your help > > > ------------------------------ > *From:* Nicolas Leconte <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* siesta-l <[email protected]> > *Sent:* Monday, September 24, 2012 5:53 PM > *Subject:* Re: Fw: [SIESTA-L] read H matrix from .hsx file > > Dear Mehrzad, > > Don't get me wrong. The COOP analysis won't give you exact numerical > values. For that, you'd have to get the information out of the Hamiltonian, > as mentioned before by Nick as well. Methodology we're not familiar with. > > However, a COOP analysis, as well as other (postprocessing) tools such as > the DOS, LDOS, charge density differences, etc can give you more > information about which orbitals participate in a specific bond. Once you > have such understanding, you can construct a minimal TB model to construct > a electronic bandstructure which you fit to the DFT electronic > bandstructure. > > As for your question on the use of the mprop utility, check the archive of > the mailing list. This has been discussed several times. > > Good luck! > Nicolas > > On Mon, Sep 24, 2012 at 4:13 PM, Mehrzad Sasanpoor <[email protected]>wrote: > > dear Nicolas, I will be very thankful to explain me how to relate between > a band structure of molecule and the coop curve of molecule to get the > quatitative value of site and hopping energy of molecule. in the article > you referencedPRB 84, 235420 (2011), I really get confused how the TB curve > achieve and where the values extract? > > > thanks a lot for your attention > I really need > > > > > ------------------------------ > *From:* Nicolas Leconte <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Sent:* Saturday, September 22, 2012 12:53 PM > *Subject:* Re: Fw: [SIESTA-L] read H matrix from .hsx file > > Yes, you can. For theory and examples, check out http://www.cohp.de/ and > references within. > I haven't done it on molecules yet, but I guess interpretation of the > curves will be much more straightforward than for crystals. > > Nicolas > > On Sat, Sep 22, 2012 at 6:49 AM, Mehrzad Sasanpoor <[email protected]>wrote: > > > > > dear Nicolas thanks a lot for your useful reply, > my system is isolated molecule, in fact it is benzene molecule which is > functionalizd by methyl group, and I want to find the site and hopping > energy for every atom. > is it possible to do this task according to your suggestion or coop /cooh > is related to crystal system only? > > thanks in advance > ------------------------------ > *From:* Nicolas Leconte <[email protected]> > *To:* [email protected] > *Sent:* Wednesday, September 19, 2012 12:22 PM > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > Dear Sf Kshp, > > Indeed, all the information you need to get onsite energies and hopping > parameters is available in the Hamiltonian. But Siesta won't give them to > you just like that. You'll have to do some extra coding, and similar to > Nick, I'm not familiar enough with this to help you. > > However, the way we did it recently, was exactly by combining different > output utilities in the siesta directory, including the COOP/COHP one. > These different tools can help you finding which orbitals are important in > the energy window you're interested in. In the end, a fit of the electronic > bandstructure was used to get the quantitative values. You can find more > information on this methodology in PRB 84, 235420 (2011). > > Good luck, > Nicolas > > On Wed, Sep 19, 2012 at 10:26 AM, Nick Papior Andersen < > [email protected]> wrote: > > I would rather not take on this task, there is a substantial amount of > subtleties in obtaining these quantities (many of which I am not too > familiar with). > As such I do not feel that I am the right person to help you with this > task. > > I am not too experienced with the COOP code either. I cannot say whether > it will or will not help you. > > Kind regards Nick > > 2012/9/19 sf kshp <[email protected]> > > Dear Nick, > Now I get it.I 'm very thankful for you attention, > I want to open this file only to see how the Hamiltonian and S matrices > are wirtten but i can't, would you please instruct me what should I do? > and do you have any suggestion to find these parameters (hopping and > onsit) from siesta? do you think COOP can help me? > > thanks for your time. > Sf Kshp > > > On Wed, Sep 19, 2012 at 10:45 AM, Nick Papior Andersen < > [email protected]> wrote: > > I am not sure what you mean by "anything helpful". > You have access to the Hamiltonian and overlap matrices that is used in > siesta. Whatever, you want to gather from the Hamiltonian and the overlap > matrix you in principle can. > > Kind regards Nick > PS. It is no problem to open the file, it is more the interpretation of > the records in the HS file which is the key. > > 2012/9/19 sf kshp <[email protected]> > > Dear Nick, > thank you for your comprehensive help, > let me know if I understand your mean or not: > > suppose we can open hs file, we can see anything helpful. right? > > Best, > Sf Kshp > > > On Wed, Sep 19, 2012 at 12:35 AM, Nick Papior Andersen < > [email protected]> wrote: > > Yes, in the subfolder Utils, from where you gathered hsx2hs. :) > > Whether it is impossible to get using siesta code, yes and no. > To my knowledge you cannot ask siesta to pop out the value. However, you > can program your way out of it. > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > do you mean the utilities in siesta? > > ------------------------------ > *From:* Nick Papior Andersen <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* siesta-l <[email protected]> > *Sent:* Wednesday, September 19, 2012 12:28 AM > > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > No, they are still in sparse format. Same problem as before. > > Probably some of the utilities can be utilized for this. However, I do not > have experience with any utils of that sort. > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > > > > dear Nick, how about the writeDMHS.History.NetCDF parameter of siesta > that If true, a series of netCDF files with names of the > form DMHS-NNNN.nc <http://dmhs-nnnn.nc/> is created to hold the complete > history of the input and output density > matrix, and the Hamiltonian.Each file corresponds to a geometry step. The > overlap matrix is stored only once per SCF cycl > > is it useful for this goal? > > thank you so much for your reply > ------------------------------ > *From:* Nick Papior Andersen <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* "[email protected]" <[email protected]> > *Sent:* Wednesday, September 19, 2012 12:12 AM > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > Yes, it is not easier in HS or HSX. Nor is it in fact a "simple" path to > get these values. > > For you to get to these values it is important that you look into the > setup of the sparse matrix form of H and S. Try and search for this and see > if you can find anything. > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > dear Nick, I want to find the value of site energy and hopping energy of > atomic orbitals for typical molecules, according to manual and mailing list > I thought by opening the HS file I can achieve these values, > do I made a mistake? > I get really confused > > thanks a lot for your attention > > > ------------------------------ > *From:* Nick Papior Andersen <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* siesta-l <[email protected]> > *Sent:* Tuesday, September 18, 2012 7:13 PM > > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > Read through the mailing list there should be many examples of this. > If still in doubt read the code for hs2hsx about reading a HSX file. > Within this you can see what it contains and a little about the formation > of H and S. > > It seems as if you had no apparent reason for changing the format to HS? > HSX and HS contain "the same" information, although HSX is more compact. > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > dear Nick thanks a lot , I could achive mol.HS > > I really thank you very much > but 1 thing , how can I open mol.HS file to read its information? > > ------------------------------ > *From:* Nick Papior Andersen <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Sent:* Tuesday, September 18, 2012 5:46 PM > > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > It could mean that you have a corrupted HSX file. It is very hard to tell > without having the file. > It successfully read the first records without errors and then errors at a > new read. This, to me suggests an erroneous file. > However, just for a test you can try to copy the file instead of linking... > So try this: > rm -f HSX > cp mol.hsx HSX > hsx2hs > > If this does not work try a simple system and do the same. If converting > the hsx file from another run works, then you have a corrupt file. > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > > dear Sir > thank you very much > after doing ln -s mol.hsx HSX > when I type hsx2hs > again, I have this error > > Using unit: 10 > At line 80 of file hsx_m.f90 (unit = 10, file = 'HSX') > please help me > thanks a lot > > ------------------------------ > *From:* Nick Papior Andersen <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* "[email protected]" <[email protected]> > *Sent:* Tuesday, September 18, 2012 5:12 PM > > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > Dear Mehrzad > That function is not intented to read in from stdin. > > It reads the HSX file from a file called "HSX". > Thus you should do this: > ln -s mol.hsx HSX > hsx2hs > mv HS mol.HS > rm HSX > > And that should be it... > > Kind regards Nick > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > dear Sir > thank a lot for your response, I understood that mprop.f90 read only hs > file; > however when I convert the mol.HSX to HS file, > hsx2hs <mol.hsx > I face with this error > > Using unit: 10 > At line 67 of file hsx_m.f90 (unit = 10, file = '') > Fortran runtime error: File 'HSX' does not exist > > what should I do? > thanks a lot > > ------------------------------ > *From:* Huang Bing <[email protected]> > *To:* Mehrzad Sasanpoor <[email protected]> > *Cc:* siesta-l <[email protected]> > *Sent:* Tuesday, September 18, 2012 3:53 PM > *Subject:* Re: [SIESTA-L] read H matrix from .hsx file > > Hi, > Having a look at the mprop.f90 (siesta-3.1/Util/COOP/mprop.f90) > could perhaps gives you some hint. > > Best Regards, > Bing > > > 2012/9/18 Mehrzad Sasanpoor <[email protected]> > > hi siesta users, > I want to open and read the *.HSX file which have the site and hopping > energy of my typical molecule, > but I couldn't, > can anybody tell me how I open and read my informations from it, > > I really need these informations soon > > thanks a lot > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
