Re: [Wien] Concerns on the obtained values of momentum matrix elements
The change of sign might simply be a phase change. Since wave functions can change by a phase (exp(i phi)) and are still the same wave functions, also the corresponding momentum matrix elements might show this. About the abrupt change of the matrix elements I can only speculate: There are 2 bands crossing (either in VB or CB) and the wave function changes character ??? Did you check your band structure ? Am 29.11.2016 um 17:16 schrieb Yong Woo Kim: Dear Wien2k users, Hello, I am running wien version 14.2 on linux compiled with gfortran. Right now I am trying to calculate the momentum matrix elements of Al2O3 sapphire. I have managed to get some results but some part of the results worry me that I may have done it wrong. I am particularly interested in the matrix elements between the highest valence and the lowest conduction band. The following is part of the results that I have obtained along the Gamma-A direction that I want to obtain the results for. 36 37 -0.4751700 2.313690e-11 0.46483076000 36 37 -0.4750540 2.336530e-11 0.46490079000 36 37 -0.4749370 2.359330e-11 0.4649715 36 37 0.4748190 -2.382160e-11 0.46504288000 36 37 0.4746990 -2.404930e-11 0.46511493000 36 37 0.4745790 -2.427750e-11 0.46518766000 36 37 -0.4744570 2.450540e-11 0.46526106000 36 37 0.4743340 -2.473300e-11 0.46533514000 36 37 -0.4742100 2.496040e-11 0.46540988000 36 37 0.4740850 -2.518800e-11 0.4654853 36 37 0.4739590 -2.541530e-11 0.46556139000 36 37 0.4738320 -2.564270e-11 0.46563815000 36 37 -0.4737040 2.587000e-11 0.46571559000 36 37 -7.666040e-13 -8.840590e-18 0.46575618000 36 37 1.329390e-11-3.782700e-19 0.46579377000 36 37 -6.203120e-13 -1.299510e-17 0.46583168000 36 37 -6.782450e-12 -2.774380e-17 0.46586991000 36 37 1.133070e-11-2.487130e-17 0.46590846000 36 37 -2.172930e-12 5.122720e-17 0.46594733000 36 37 7.867630e-12-1.914880e-17 0.46598652000 36 and 37 are the band index for my valence and conduction band. Each row refers to a k point along the G-A path and I have 501 rows in total. I also eliminated the x,y components and leaved only the z component plus the energy difference. One minor concern is that the signs of the values change and this doesn't seem to be right. Another concern is that as can be seen from the real part of z above, the value suddenly drops to less than 1e-10 order. Although not shown here, at the same k point, the real part of the x component showed the opposite behaviour, increasing from less that 1e-10 to about 0.4. This abrupt change doesn't seem to be right either. The procedure went like this. run_lapw create case.klist_band x lapw2 -fermi x lapw1 -band x optic I have tried this for different k mesh by using x kgen and run_lapw repeatedly from 1000 to 15000 and the results only had minor differences. Any help would be really appreciated. Thank you very much in advance. Yong Woo Kim ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html -- -- Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna Phone: +43-1-58801-165300 FAX: +43-1-58801-165982 Email: bl...@theochem.tuwien.ac.atWIEN2k: http://www.wien2k.at WWW: http://www.imc.tuwien.ac.at/staff/tc_group_e.php -- ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
[Wien] Concerns on the obtained values of momentum matrix elements
Dear Wien2k users, Hello, I am running wien version 14.2 on linux compiled with gfortran. Right now I am trying to calculate the momentum matrix elements of Al2O3 sapphire. I have managed to get some results but some part of the results worry me that I may have done it wrong. I am particularly interested in the matrix elements between the highest valence and the lowest conduction band. The following is part of the results that I have obtained along the Gamma-A direction that I want to obtain the results for. 36 37 -0.4751700 2.313690e-11 0.46483076000 36 37 -0.4750540 2.336530e-11 0.46490079000 36 37 -0.4749370 2.359330e-11 0.4649715 36 37 0.4748190 -2.382160e-11 0.46504288000 36 37 0.4746990 -2.404930e-11 0.46511493000 36 37 0.4745790 -2.427750e-11 0.46518766000 36 37 -0.4744570 2.450540e-11 0.46526106000 36 37 0.4743340 -2.473300e-11 0.46533514000 36 37 -0.4742100 2.496040e-11 0.46540988000 36 37 0.4740850 -2.518800e-11 0.4654853 36 37 0.4739590 -2.541530e-11 0.46556139000 36 37 0.4738320 -2.564270e-11 0.46563815000 36 37 -0.4737040 2.587000e-11 0.46571559000 36 37 -7.666040e-13 -8.840590e-18 0.46575618000 36 37 1.329390e-11-3.782700e-19 0.46579377000 36 37 -6.203120e-13 -1.299510e-17 0.46583168000 36 37 -6.782450e-12 -2.774380e-17 0.46586991000 36 37 1.133070e-11-2.487130e-17 0.46590846000 36 37 -2.172930e-12 5.122720e-17 0.46594733000 36 37 7.867630e-12-1.914880e-17 0.46598652000 36 and 37 are the band index for my valence and conduction band. Each row refers to a k point along the G-A path and I have 501 rows in total. I also eliminated the x,y components and leaved only the z component plus the energy difference. One minor concern is that the signs of the values change and this doesn't seem to be right. Another concern is that as can be seen from the real part of z above, the value suddenly drops to less than 1e-10 order. Although not shown here, at the same k point, the real part of the x component showed the opposite behaviour, increasing from less that 1e-10 to about 0.4. This abrupt change doesn't seem to be right either. The procedure went like this. run_lapw create case.klist_band x lapw2 -fermi x lapw1 -band x optic I have tried this for different k mesh by using x kgen and run_lapw repeatedly from 1000 to 15000 and the results only had minor differences. Any help would be really appreciated. Thank you very much in advance. Yong Woo Kim ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] How to include the localized d orbitals in the atomic spheres?
On 11/28/2016 06:04 PM, Abderrahmane Reggad wrote: > Sorry for my question No worries! Asking and answering questions is the purpose of this forum, after all. > Wen we use the maximum values for the Rmt such a way the spheres become > touched. Does that guarantee that the 3d electrons are all inside atomic > spheres? To answer the implied question as well: Yes, this means that the U / EECE potentials are applied only to “a part of” the states you specify (or, as Martin wrote: “between the atomic spheres the potentials … are set to zero”). You can view this as a deficiency of the method, but it is standard practice and normally quite good enough. Think about how the target states are defined: as the d states (for example) of some atom, i.e., as the projection of the Kohn-Sham states onto the d manifold around that atom. But to even define this projection, you need to specify a sphere around the atom. In an APW code, the muffin-tin sphere is the natural choice. To go beyond this approach and make sure that you cover the “whole” d states, you would need to provide an alternative definition of those states. One possibility would be Wannier functions, but it would not (normally) make sense to do a Wannier projection during each DFT iteration “only” for DFT+U. Elias -- Elias Assmann Wien2Wannier: maximally localized Wannier functions from linearized augmented plane waves http://wien2wannier.github.io/ https://github.com/wien2wannier/wien2wannier/ signature.asc Description: OpenPGP digital signature ___ Wien mailing list Wien@zeus.theochem.tuwien.ac.at http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien SEARCH the MAILING-LIST at: http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
Re: [Wien] Discrepancy in the simulation of the paramagnetic state
My (and probably Xavier's) concern with Regaard's question was something else. I have no problem whatsoever with you finding an approximation for Pt using wave functions. After all, your ground state model has zero static local moments, as has the Pt you want to model. ;-) However, the approximation seems at least dubious if the ground state model and the low temperature state of the material differ. If the material enters some magnetic state and the spin-polarized(!) DFT model does not one might look for a problem with the structure data, some structural phase transition, ... So I am with Xavier, and I would at least advise to be careful with the idea I understood Regaad did somehow get: Artificially compensate spins (e.g. via LDA instead of LSDA) to find an approximation for the paramagnetic phase at elevated temperature of a low temperature magnet. There is at least one difference between the material and the model: the model will NOT be paramagnetic (obtain a positive magnetization in an applied magnetic field). Wether or not this (or any other differences induced by the forced spin compensation) poses a problem will depend on what situation one wants to model. --- Dr. Martin Pieper Karl-Franzens University Institute of Physics Universitätsplatz 5 A-8010 Graz Austria Tel.: +43-(0)316-380-8564 Am 28.11.2016 08:33, schrieb Fecher, Gerhard: I hope you agree that Pt is paramagnetic I did two calculations for Pt, one was spin polarized the other not. The results are identical, no resulting magnetic moment (indeed, I started with one in the spin polarized case), did I play a trick or did Wien2k play a trick ? but may be Wien2k can not be used to calculate the electronic structure of Pt, because it is paramagnetic (Pt, not Wien2k !). I hope you agree that Pt is paramagnetic even at Zero temperature. why do I need to include temperature effects to calculate the ground state of Pt (at 0 K, where else) ? ... and what should MtC calculations tell me about it ? Remark 1: Calculations may be "spin polarized" (LSDA) or not (LDA) or they may be even more sophisticated "non-colinear spin polarized" or they may be for "disordred local moments" or for "spin spirals", or ???, just to name some. Remark 2: Materials may be diamagnetic, paramagnetic (Langevin, Pauli, van Vleck), ferromagnetic (localised moments, itinerant), ferrimagnetic (collinear, non-collinear), etc.. Therefore, I repeat my question: How do you distinguish diamagnetic, paramagnetic, ferromagnetic, and ... states ? The answer is for you, not for me. I tried to calculate for Pt using Hohenberg Kohn DFT, but I could not find the functional, all I found was some approximation using wave functions. Don't worry I will not ask a question about it ;-) Ciao Gerhard DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy: "I think the problem, to be quite honest with you, is that you have never actually known what the question is." Dr. Gerhard H. Fecher Institut of Inorganic and Analytical Chemistry Johannes Gutenberg - University 55099 Mainz and Max Planck Institute for Chemical Physics of Solids 01187 Dresden Von: Wien [wien-boun...@zeus.theochem.tuwien.ac.at] im Auftrag von Xavier Rocquefelte [xavier.rocquefe...@univ-rennes1.fr] Gesendet: Sonntag, 27. November 2016 12:46 An: wien@zeus.theochem.tuwien.ac.at Betreff: Re: [Wien] Discrepancy in the simulation of the paramagnetic state Just to add one more point to this funny discussion, the term "paramagnetic" is sometimes used in the DFT litterature in an improper way. It could clearly lead to misunderstanding for researchers who do not know so much on how magnetic properties could evolve with temperature and applied magnetic field. When you see in a paper "paramagnetic state" simulated using DFT ... it is NOT paramagnetic at all, it is simply a trick which must be considered with care as previously mentionned by Peter, Eliane and Martin. If you want to simulate a paramagnetic state you need to include the temperature effects, i.e. you should consider the spin dynamics and the competition between magnetic exchange interactions and thermal fluctuations. This could be done, at least, using Monte-Carlo calculations based on an effective hamiltonian constructed on top of DFT parameters (including magnetic exchange and anisotropy at least). Best Regards Xavier Le 27/11/2016 à 10:01, Fecher, Gerhard a écrit : How do you distinguish a diamagnetic, a paramagnetic, a ferromagnetic, and an antiferromagnetic state. Think ! This will answer your question, hopefully. Ciao Gerhard DEEP THOUGHT in D. Adams; Hitchhikers Guide to the Galaxy: "I think the problem, to be quite honest with you, is that you have never actually known what the question is." Dr. Gerhard H. Fecher Institut of Inorganic and Analytical Chemistry Johannes Gutenberg - University
Re: [Wien] How to include the localized d orbitals in the atomic spheres?
Look into section 7.3 of the user guide: ORB (Calculate orbital potentials) The very first sentence reads: orb calculates the orbital dependent potentials, i.e. >>>potentials which are nonzero in the atomic spheres only <