Dear Asa, 1. You set up magnetism in the DM.InitSpin block for atoms 1 and 2 only; consequently the rest of your atoms, 3 through 18, are initialized non-magnetic. You must be able to see it in the Mulliken populations block since they are printed over all atoms. Looking at it you can find out how large inidividual moment come along with a small total moment (i.e., some spins are inverted, or very small).
2. If you want initial FM setting of ALL spins you don't need the DM.InitSpin block at all. However, if you keep it, you should list all your atoms (which must be magnetic). You don't need any Thetas and Phis in non-collinear case. Just %block DM.InitSpin 1 + 2 + ... 18 + %endblock DM.InitSpin 3. The local moments of 3d atoms on the surface or in thin film are higher than in the bulk; the values of about 4 m_B on Fe or Co are not obviously unreasonable. 4. Do not expect the same values of local moments as those produced by VASP, since they are defined differently in two methods. 5. You don't need the block BandLines (unless you are ready with the calculation and want really study band dispersion), at least not at the beginning. On the contrary, you'd need some k-points for k-space sampling; you defined none meaning that you go ahead with a single k-point, that is not that good. 6. I did not check your structure, but it could be a good idea to make it visible (e.g., transforming the .XV file which will be created in your calculation using my Sies2xsf tools and viewing it with XCrySDen if you have it installed). Best regards Andrei > Dear Andrei Postnikov, > > Thank for your prompt reply. > 1) I have set WriteMullikenPop 1 and it is seen that the individual > magnetic moments are too high for both Co(~4.5) and Fe(~4) . This might be > due to the assignment of high values of magnetic moments as the spin > polarised flag is set true. > Again I ran the calculations with the following DM.InitSpin block setting > > %block DM.InitSpin > 1 + 90. 0. > 2 + 90. 0. > %endblock DM.InitSpin > Still the magnetic moments are coming very high and so the total > magnetisation of the system very small. > 2) Regarding the self consistency I have not seen a trend of decreasing > magnetism, the individual moments are double that of what have been > expected. > I am attaching my input files along with .Can anybody please correct > me > if any other serious error I am making in the input parameters. > Thanks in advance > Asa > > -- > Assa Aravindh.S > SENIOR RESEARCH FELLOW, > THEORY AND COMPUTATIONAL STUDEIS SECTION, > MATERIAL SCIENCE DIVISION, > IGCAR.KALPAKKAM. >
