Home

FDMNES User's Guide

1. For the analysis of the cartesian tensors keyword cartesian _car_atoma txt cartesian tensors of the atom a _car_xtal txt Cartesian tensors for the crystal _car_xtal_rxsi txt Cartesian tensors for the crystal for the rxs reflection number 1 FDMNES User s Guide II Basic keywords II 1 Radius of the cluster The final states are calculated inside a sphere whose radius is defined with the keyword radius or rayon Only the atoms inside this sphere are considered Radius gt Obligatory keyword preceding the radius of the cluster 3 5 value in Angstr m of the cluster radius H 2 Cluster or crystal structure Under crystal or molecule stand all the data describing respectively the unit mesh or the molecule If the calculation is done using the flapw output this block is useless because the structure is red in one of the flapw output files Under the keyword come the mesh parameters and the angles degrees Then come all the atoms and not only the non eguivalent ones but when one specify the space group using the keyword spgroup By default and in the absence of the keyword absorbeur the absorbing atom chemical specie corresponds to the first atom in the list Example 1 cfc copper crystal Crystal gt Crystal structure or cristal 3 610 3 610 3 610 90 90 90 gt a b c a B y 29 00 00 0 0 gt Atomic number position 29 05 05 00 29 05 00 0 5 29 00 05 0 5 Example 2 FeOg octahedr
2. If one wants to keep the energy dependant formula but adding 1 2 or more to this formula put a negative sign in from of 1 Imax gt The fmax value is now given by the formula plus 2 2 When the energy is increasing Cmax increases By default the maximum value of Lmax 1S set at 2 for Z 1 at 3 for Z 2 then at 4 for Z lt 18 then at 5 for Z lt 36 then at 6 for Z lt 54 then at 7 for Z lt 86 and 8 for Z over To avoid this limitation put the keyword Imaxfree When working using the finite difference mode the connection to the outer sphere also needs an expansion in spherical waves This one uses the same formula that for the atom but r is now the radius of the outer sphere To modify this max one proceeds exactly in the same way with positive value to fix an energy independent value and negative value to keep the formula but adding to the formula By default the additional value is 5 The key word is now maxso lmaxso 20 IV 13 Muffin tin radius By default the muffin tin radius is calculated to have a jump of potential the smallest as possible and closed between the atoms One can modify it by the use of 3 different keywords If one wants that this radius is calculated using the Norman procedure introduce the keyword Norman If one wants than the muffin tin radius is the half value of the interatomic distance introduce the keyword Raydem 39 Manuel FDMNES By default there is a 10 overlap betwee
3. G where k corresponds to the neuman bessel normalization To avoid eventual convergence problems not seen yet one can prefer to calculate 1 Gt t For this use the keyword Normaltau IV 9 Muffin tin potential If one wants to use the muffin tin approximation in FDM put the keyword muffintin IV 10 Non excited absorbing atom To perform a calculation with a non excited absorbing atom that is with an absorbing atom without a hole in its core level and without screening put Nonexc IV 11 Modification of the grid of point parameters By default the finite difference calculation is performed at order 4 with an inter point distance equal to 0 25 A the radius of the spherically symmetric area is around 0 65 A less for the light elements To modify them write Rmt to modify the muffin tin radius 0 65 Adimp to modify the interpoint distance 0 20 Iord gt To modify the order of the Taylor expansion 2 38 Manuel FDMNES IV 12 Expansion in spherical harmonics An expansion in spherical harmonics is performed in the atoms both in the multiple scattering mode and in the finite difference mode but with a smaller radius The maximum value of is obtained from the formula kr 4l max max 1 where k is the photo electron wave vector and r the muffin tin radius By default to this value one adds 1 To fix a value of Emax independent of energy just write Imax gt The nax Value is fixed at 3 3
4. but when O is along Oz in that case K is substituted by 7 k and 7 are the bases c a vectors of the direct elementary mesh One then gets versus the Bragg angle and the azimuthal angle the incoming and outgoing wave vectors cos cosol cos 8sin g sin BO cos Bcos l cos singl sin BO For the and z polarizations one gets E sin gl cos gJ V sin 3 cos gf sin Bsin gJ cos BO sin Bcos gl sin 8 sin gJ cos BO ll M 1 y N l ll o For example orthogonal mesh beam h 0 0 7 J 0 k i 11 0 gt corresponds to o o with polarization along j 1 1 90 gt corresponds to o o with polarization along k orthogonal mesh beam 0 0 1 7 7 0 j k 11 0 gt corresponds to o o with polarization along j 1 190 gt corresponds to 0 0 with polarization along i If one does not put the angle this implies that one performs a phi scan and all the amplitudes are calculated for all the angles by 2 step There is then a supplementary output files with the extension _scan txt It is also possible to perform a 360 scan by 2 step for the rectilinear incoming or outgoing polarizations For this one has to write 10 in place of the sigma pi notification Dafs 002 101 45 gt incoming polarization is scanned FDMNES User s Guide 002 2 10 45 gt outgoing polarization is scanned The first value angle 0 corresponds to the o polarization 90 correspond
5. Par_phi gt position along for an atom in spherical or cylindrical coordinate Under each parameter must be written the first and last values of the parameter followed by the number of value For the parameters Par_posx Par_posy Par_posz Par_theta Par_phi The number of the atom must also be specified Under the parameter Par _poporb must be also set the number of the orbital found in the order given under the keyword atom Foe example to specify an exchange of charge between two atoms titanium and oxygen one has to write Atom 22 a 2a gt 3d titanium orbital with 2 electrons for the initial occupancy 821 4 gt 2p oxygen orbital with 4 electrons for the initial occupancy 50 Manuel FDMNES Parameter Par_poporb 2 0 3 1 gt The first orbital under atom is the 3d titanium Par_poporb 4 6 3 2 gt The second orbital under atom is the 4p oxygen By default the metric distances are calculated in all the energy range is the intersection between the experimental and calculated spectra It is possible to cut the lower or and the higher energy part of the spectra by the use of the keyword Emin 10 gt Minimum energy for all the spectra Emax 100 gt Maximum energy for all the spectra It is possible to have different values for the different spectra Emin 10 5 20 20 gt Minimum energy for each spectra Emax 45 100 100 100 gt Maximum energy for each spectra If the energy of the experimental spectra is
6. calculation on gets the following scheme Choose of the parameters indata file reading Calculation of the XANES spectra and DAFS amplit Loop over the shift and convolution parameters Loop over the position or charge parameters Convolution and calculation of the DAFS intensities Comparison with the experimental spectra Calculation of the final XANES and DAFS amplit Convolution and calculation of the final DAFS Extraction of DAFS azimuth scan or spectra FDMNES User s Guide Nevertheless it is possible to perform the different steps separately The comparison with the experimental spectra is also not obligatory When used the parameter fit must be performed with care In practice many calculation are limited to the step XANES and DAFS calculation and Convolution and calculation of DAFS intensities These two steps can also be performed together or separately The next chapter treats about the principal indata file for the step XANES and DAFS calculation Generally this file is sufficient to describe all the necessary data for the calculation because the program calculates its atomic bases and the potential Nevertheless the user can prefer use its own atomic bases or uses directly the potential calculated by the band structure program FLAPW WIEN 2k In both cases some other files must be furnished They are described further in the manual The indata necessary for the steps Convolution compari
7. in keV and not in eV put the keyword Kev When there are several spectra to compare simultaneously for the calculation they must be in the same output file One then must give the number of the column contains the calculated spectra and associate it at the corresponding experimental file This is done adding a line after each experimental file contains the number of the column If there are 2 numbers the first one is the number of the column in the experimental file the second one being the number of the column in the calculated file When there is no number this means than in both experimental and calculated file the spectra are in the second column the first one being the energy Example Experiment gt keyword Nom_exp_1 txt gt Name of the file containing the first experimental spectra 2 gt Number of the column in the calculated file containing the corresponding spectra Nom_exp_2 txt gt Name of the file containing the second experimental spectra 3 gt Number of the column in the calculated file containing the corresponding spectra Nom_exp_2 txt gt Name of the file containing the third experimental spectra 4 3 gt Numbers of the column in the experimental file Nom_exp_2 ttt followed by the number of the column in the calculated file 5l Manuel FDMNES By default three confidence factors are used A fourth one Rxg can also been calculated It is eguivalent to Rx but with a unigue normalization factor between experi
8. is found along a diagonal For this purpose put the keyword Base_spin FDMNES User s Guide IV 4 Calculation basis By default the connection to the continuum is performed in real base neuman bessel To use a complex base bessel hankel put Basecomp IV 5 Equivalent atoms The program automatically calculates the total signal resulting from the equivalent atoms by symmetry It is nevertheless possible for the expert fdmnes user to impose the symmetry relation between the atoms and thus to impose this summation One must then use the keyword symsite Following this keyword stands the number of non equivalent atoms or number of group of atoms then for each of them the number of eguivalent atoms followed the list of the relative symmetry to the first atom of the list and the atomic position in mesh unit The symmetry are codified by number going from 1 to 64 see the list given bellow Symsite 3 Number of non equivalent atoms or group of atoms 4 Number of equivalent atoms group 1 1 0 2500 0 2500 0 2500 24 0 7500 0 7500 0 2500 23 0 7500 0 2500 0 7500 22 0 2500 0 7500 0 7500 4 Number of equivalent atoms group 2 1 0 0000 0 0000 0 0000 22 0 0000 0 5000 0 5000 23 0 5000 0 0000 0 5000 24 0 5000 0 5000 0 0000 4 Number of equivalent atoms group 3 1 0 5000 0 5000 0 5000 23 0 0000 0 5000 0 0000 22 0 5000 0 0000 0 0000 24 0 0000 0 0000 0 5000 The negative indexes correspond to the same symmetry p
9. metal at the end of the d band the integral must be close to 10 because there are 10 d electrons At the end of the rather unlocalized s and p bands the integrals rarely reach 2 or 6 because the electrons are counted only inside the atomic sphere of radius Rmsta see in the bav file for its value These orbitals having a rather big radius they are not all inside such atomic sphere Calculation being performed in the continuum one also has to recall that n is not anymore a good quantum number Consequently the integral continue to increase indefinitely with energy At the end of a d or f band rather localized one nevertheless reaches more or less at 10 or 14 In this case one can fin the Fermi level energy just looking the energy where the integral reach the supposed number of electron in the corresponding level of the atom Sometimes the integral never reach 10 or 14 or goes far higher One of the reasons can be that the starting energy is too high one thus looses the beginning of the band another reason can be the energy grid is not sufficiently thin One has in this case to take a path smaller down to 0 01 eV or even less for 4f elements In order to avoid such a thin grid it is also possible de broaden a bit these localized states using the keyword Eimag this works only in Green mode III 17 Spherical tensors To get the spherical tensors in number of electron put the keyword Spherical One gets new files with the
10. the indata file with the keywords Jump and Endjump All what is between them is not red For example Jump Quadrupole Spinorbite Endjum gt the keywords Quadrupole and Spinorbite are not considered Jump III 27 Comment It is possible to introduce a line of comment which will be copied in the output files For this purpose write Comment Iron K edge in magnetite gt line of comment IV 28 Atomic spectra To get in last column the atomic absorption spectra without the neighbour atoms put the keyword Xan_ atom IV 29 Non resonant magnetic scattering The non resonant magnetic scattering is taken into account for the RXS This term can be decreased or put to zero using a multiplicative factor No_res_mag 0 8 gt factor An additional factor can be applied to the contribution from the orbital moment This factor corresponds to L 2S and not L S No_res_mom 0 3 factor for the orbital moment By default this factor is calculated using the Hund rules and multiplied by 0 2 FDMNES User s Guide IV Technical keyword Here stand the technical keywords necessary for specific tests and optimization of the code for expert users IV 1 Details on the calculations To get details on the calculations done in every routine put the Keyword Check 33331111 gt Main Lectur Prepar Agregat Etafin Distan Reseau Laplac 11111111 gt Bord Potato Orbval Poisso Pot0 Ylmsrt Ylmatm Potex T1111 gt Phiso
11. the keyword Noncentre If one wants to impose a specific center put Centre 0 0 25 1 gt coordinates in mesh parameter unit of the center III 15 Cut of the potential To get a cut of the potential in the output along a line or a plan put Trace 01 0 0 0 0 0 gt l a b c p g r 1 1 cut along a line with vector a b c crossing the point p g r in unit mesh parameter 1 2 cut along the plane ax by cz p 1 3 gives all the points 29 FDMNES User s Guide III 16 Density of state To get the state density and its integral for any harmonics projected on the central atom put the keyword Density There will get one output file more with the suffix _sd0 txt Note that only the states members of the representations useful to get the asked spectra are calculated Conseguently such produced density of state can be uncomplete To get a complete projection for all the atoms put the keyword sfate all in place of density Calculation is then performed on all the representations what can be longer Then one gets n new output files with the extensions _ sdi txt i being the atom number In these files stand in column first the integral of the total atomic electron density then the density and its integral of each m followed by the sum over m that is the density and its integral for each For magnetic calculation the expansion is split in its up and down components Usually for a transition
12. the program By default the result is given in the output file only for their average corresponding to a powder If one wants to have the XANES along specific orientations of the polarization or the wave vector in quadrupolar orientations or if one wants to make circular dichroism one has to introduce Polarise FDMNES User s Guide Thus one gets the calculations for the different independent polarizations then in the last column the average To get specific polarization orientation the keyword must be followed by the polarizations Polarise 1 0 1 0 0 0 1 0 1 0 0 0 0 0 0 0 1 0 For a guadrupolar calculation one has to specify the wave vector after the polarization Polarise 1 01 0 00 0 0 0 0 1 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 Each line contains the polarization vector the wave vector If the wave vector is zero this polarization is calculated in the dipolar approximation It is possible to perform an average on different polarizations adding a new number at the end of the line is the corresponding weight If at least two weights are non zero in the output there will be a new column with the weighted average of the different polarization Polarise 101 0 0 0 0 00 0 1 0 1 1 0 1 0 00 0 0 0 0 1 0 I 0 0 0 0 10 0 0 0 0 0 0 0 If one wants to have a circular polarization one just has to put this one to zero followed by the value of the wave vector Polarise 0 0 0 0 00 0 0 0 0 1 0 In the output on
13. 0 O 32 Note that for the calculation of the symmetry it is important de define the atom position with a sufficient number of digit say 10 For instance for graphite one has to write Spgroup P63mc Crystal 2 456 2 456 6 696 90 90 120 a b c alpha beta gamma 6 0 0 0 0 0 0 6 0 3333333333 0 6666666667 0 0 Putting in the last line for example 0 3333 would create false atoms Program would stop immediately with an error message FDMNES User s Guide II 3 Atomic electronic densities An electronic configuration is used by default for all the atoms It is possible to modify it by the use of the keyword atom Moreover under crystal or molecule one must not anymore put the atomic number by the atom type number For example in case of a FeO octahedron with the Fe 3d 4s and O 2s 2p configuration Atom gt keyword preceding the atomic electronic densities 26 2 32 6 402 gt atomic number of the chemical specie of type 1 number 8 2202 214 of valence orbital and n l pop of each of these orbitals molecule 1 900 1 900 1 900 90 90 90 gt a b c O B y 1 00 00 00 gt Atom type position 2 10 00 00 2 10 00 00 2 00 10 00 2 00 10 00 2 00 00 1 0 2 00 00 10 Important remark contrary to what one can think the formal charges attributed to the atoms in the ionic compounds are far from the true charge Thus one has to perform exchange of charge between atoms with care and in a moderate way A good techn
14. 19 29 44 46 26 Manuel FDMNES 60 Selec inp Selec out Self absorption Spgroup Sphere all Spherical Spinorbite Supermuf Sym Symsite Table Thomson Trace V0imp Vmax Xalpha Xan atom Zero_azim 53 53 21 15 30 30 22 40 35 36 46 48 29 27 28 27 34 21
15. 5 for inversion III 21 Energy shift of the spectra If one has gotten a reference for the initial orbital it is possible to give it under the keyword epsii This will produce a shift of the output spectra equal to the difference between this energy and the energy calculated in the program It is safer to perform this operation with the shift parameters during the convolution step Epsii 6253 1 gt positive value in eV III 22 Expansion in path In the multiple scattering mode it is possible to avoid the matrix inversion but to make an expansion in path One has to give the number of way Chemin 3 gt number of path III 23 Cartesian tensors To get the atomic cartesian tensors put the keyword Cartesian One gets a new output file with the extension _ car txt Then one can use other keywords that make that from this output file will be calculated f and f for any tensor component For this purpose put the keyword fprim in the indata file III 24 Spectra by atom If one wants to get the signal coming from each atom one has to add the keyword 32 FDMNES User s Guide Allsite One then gets at most of the usual output files new output files of the type atom1 txt atom2 txt etc The suffixes _ atoml atom2 correspond to the number of the atom III 25 Use of densities and potential coming from FLAPW If the potential is imported from a FLAPW calculation one has to introduce the keyword flapw in place of ato
16. 9910 V3 1 0 15620 0 00080 0 70090 V7 1 0 84380 0 50080 0 79910 V4 1 0 15620 0 49920 0 20090 V5 1 0 34380 0 49920 0 79910 V2 2b FDMNES User s Guide l N 0 65620 0 50080 0 20090 VI 0 40700 0 84500 0 65200 O18f 0 09300 0 84500 0 54800 O2 8f 0 59300 0 15500 0 54800 O3 8f 0 90700 0 15500 0 65200 O4 8f 0 90700 0 34500 0 15200 O58f 0 59300 0 34500 0 84800 O6 8f 0 09300 0 65500 0 84800 O7 8f 0 40700 0 65500 0 15200 OS 8f 0 25000 0 19100 0 00000 14e 0 75000 0 80900 0 00000 O2 4e 0 75000 0 69100 0 50000 O3 4e 0 25000 0 30900 0 50000 O4 4e Rod NY NH NHN NHN NN WNW LO II 12 Relativistic calculation To perform a relativistic calculation put the keyword Relativism II 13 Spin axis By default the spin axe is along the c axis but in the trigonal unit mesh where it is along the c axis of the associated hexagonal unit mesh This axis can be orientated along any but uniform direction axe_spin 0 08909 0 0 15025 gt in unit mesh Another way to specify this axis is using the Euler angles The orthogonal basis to do that is such that z is along c but for trigonal system where z is along the hexagonal axis x is along b x c and y is along z x x One then has to write Ang_spin 45 90 0 gt rotation around z then around y then around x In this case spin axis is along 1 1 0 in the internal basis Note that it is possible to define non collinear spin using atomic local basis In this case one has to define the
17. AM under LINUX or Windows The programming language is fortran 90 It uses the LAPACK library There is no graphical output The user must have got a fortran 90 compiler Executable file for Windows XP is also furnished Il The package Several groups of files can be downloaded the fdmnes program itself and a set of examples of indata and output files fdmnes exe is the executable program for Windows XP fdmfile txt is an indata file xsect dat and spacegroup txt are data necessary for the program prog is a directory containing all the subroutines main f general f the file nesparam inc contains the dimension of most of the tables and the file mpif h entree and xanout are directories containing a set of examples of fdmnes indata and output files IHI Differences with previous versions In the versions before 2005 the calculations were necessarily performed in two steps first the main calculation second the convolution Two different programs had to be run Now these two steps can be executed together with the unique package fdmnes Nevertheless the user can separate the different phases of the calculation as previously as is explained in the chapter describing the main indata file For the convolution part the broadening must be now specified is the width of the levels Previously it was the half width Different improvements were also added The program can furnish the different scattering tensors in their
18. Cartesian or spherical basis It is now possible to compare the calculated spectra to experimental ones and to perform calculations on different grid of parameters in order to fit them Some details have also changed the old indata gen file is now called fdmfile txt In the main indata file after the keyword range it is not anymore necessary to put the number of energies the first line is not anymore reserved for a comment and the character makes that what follows in the line is not red From October 2005 the non resonant magnetic scattering is taken into account for the RXS In October 2006 some bugs were corrected The self absorption is now calculated for resonant diffraction The fit of parameters by comparison with experimental spectra is now possible In May 2007 the program symmetrization was improved It takes into account the atoms with non spherical configuration The magnetic moments do not need anymore to be parallel The description of the local atomic base is modified see keyword crystal FDMNES User s Guide Calculations in the finite difference method mode are now faster when three fold axis are present In march 2008 the symmetrization for the multiple scattering mode without spin orbit is improved The program contains also a new module allows the building of new output files extracted from the scan files obtained after the convolution These new files contain spectra or azimuthal scan at specific ener
19. Euler local angles after the keyword crystal or molecule in the line just before the corresponding atoms See non spherical atom Example Crystal 7 7400 7 7400 3 8400 90 90 90 45 90 spin axis along 1 1 0 100 00 0 0 2025 025 0 5 135 90 spin axis along 1 1 0 22 FDMNES User s Guide 105 00 0 0 2 0 75 0 25 0 5 315 90 spin axis along 1 1 0 100 05 0 0 2025 075 05 225 90 spin axis along 1 1 0 105 05 0 0 2075 075 0 5 23 FDMNES User s Guide 24 FDMNES User s Guide HI Sophisticated keywords III 1 Use of atomic electronic densities coming from external files When using electronic densities coming from external files these ones must be specified with the keyword atom One eventually has to specify one atom more because the absorbing atom is a special chemical specie different of the non excited atoms Its files must also contain the electron initial orbital before excitation for instance the 1s one for a K edge For the format of these files see chapter F In the case there are two atoms Atom gt keyword preceding the atomic electronic densities Fe_exc txt gt name of the file containing the excited atom Fe txt gt name of the file containing the type 1 atom Oxygen txt gt name of the file containing the type 2 atom One can also put one or several atoms calculated internally among the exterior files Atom Fe_exc txt gt name of the file containing the excit
20. FDMNES User s Guide Of CENTRE NATIONAL Z o UNIVERSITE BER FEL JOSEPH FOURIER institut FDMNES User s Guide Yves Joly yves joly grenoble cnrs fr Institut N el CNRS Bat F BP 166 38042 Grenoble Cedex 9 France April 2008 FDMNES User s Guide FDMNES User s Guide Introduction The FDMNES program calculates the spectra of different spectroscopies related to the real or virtual absorption of x ray in material It gives the absorption cross sections of photons around the ionization edge that is in the energy range of XANES in the EXAFS The calculation is performed with all conditions of rectilinear or circular polarization In the same way it calculates the structure factors and intensities of anomalous or resonant diffraction spectra DAFS or RXS FDMNES also allows the comparison of the simulated spectra to experimental ones with the help of objective criteria The code uses two techniques of monoelectronic calculations The first one is based on the Finite Difference Method FDM to solve the Schr dinger equation In that way the shape of the potential is free and in particular avoid the muffin tin approximation The second one uses the Green formalism multiple scattering on a muffin tin potential This approach can be less precise but is faster The program is partially symmetrized Symmetry operations are calculated automatically The next section contains the practical informations to run the program and in par
21. arameter pi 3 1415926535897932384 8 parameter quatre_pi 4 pi parameter rydb 13 605698 8 parameter eps4 1 e 4_8 parameter eps6 1 e 6_ 8 parameter eps10 1 e 10_8 parameter epspos l e 4 8 precision on the atom and point positions 55 Manuel FDMNES 56 Manuel FDMNES G File Atomic electronic density When using atomic electronic density furnished by the user these ones must be in files having determined formats One must have one file by type of different atom the absorber being considered as a different type of atom having the same atomic number than non excited atom Example of file fer neutre 001 3d6 4s2 Orbitale n l j _ Popul Energie IS 1 0 0 0 0 0 2 000 6917 713 2S 2 0 0 0 0 0 2 000 804 490 2P 2 0 1 0 0 0 6 000 695 277 3S 3 0 0 0 0 0 2 000 91 497 3P 3 0 1 0 0 0 6 000 59 577 3D 3 0 2 0 0 0 6 000 8 092 4S 4 0 0 0 0 0 2 000 5 482 NUCLEAR CHARGE 26 000000 INTEGRAL OF CHARGE DENSITY 26 000000 ss gt all what is before is not red 26 18 00 2 32 6 000 40 2 000 600 r A rho ua 3 _psi 3D psi 4S 0 00016112638 11678 65000 0 0000000043 0 0023114080 0 00016413798 11675 19000 0 0000000046 0 0023542610 0 00016720588 11671 67000 0 0000000048 0 0023979030 0 00017033115 11668 09000 0 0000000051 0 0024423470 The file must contains a line beginning by All what is before is not red Following line 26 18 00 2 3 2 6 000 40 2 000 atomic number number of heart electrons nu
22. around Oz and inversion 29 rot 27 4 around Ox and inversion 30 rot 27 4 around Oy and inversion 31 rot 27 4 around 0z and inversion 57 plan at 30 gt containing 02 58 rot 27 2 around axe at 30 perp 0z 59 plane at 60 containing 0z 60 rot 27 2 around axe at 60 perp 0z 32 rot 27 3 around 1 1 1 and inversion 33 rot 47 3 around 1 1 1 and inversion 34 rot 27 3 around 1 1 1 and inversion 61 plane at 120 containing 0z 35 rot 47 3 around 1 1 1 and inversion A 36 rot 21 3 d 1 1 1 andj 62 rot 27 2 around axe at 120 perp 0z rof 27 3 around 1 1 1 and inversion 63 plane at 150 containing 0z 37 rot 47 3 around 1 1 1 and inversion 64 rot 27 2 around axe at 150 perp 0z 38 rot 27 3 around 1 1 1 and inversion IV 6 Orientations It is possible to restrict the calculation to the tensor components which are really useful The keywords dipimp and guaimp are used for this purpose respectively for the dipolar and guadrupolar components ldipimp 100 will calculate only the 0x component lguaimp 001 will calculate only the xz component 000 000 37 Manuel FDMNES IV 7 Bounded states If one wants to modify the kinetic energy at are calculated the bounded states beneath the average potential put the keyword efatlie followed by the kinetic energy in eV etatlie 0 05 IV 8 Inversion of the matrix in the multiple scattering mode By default one inverts the matrix i
23. as a constant in all the energy range can be introduced through the keyword Thomson 1 1670478E 02 1 0583769E 02 1 1477827E 02 8 2670689E 01 In the second line stand these terms here for 2 reflections They replace the values calculated by the program and placed in the second line of the output file Photoemission With the keyword photoemission it is the photoemission spectra which is calculated In such calculation the cutting is simply above the Fermi level and not beneath it as in xanes The convolution broadening is the independent of energy and is fixed by default to the hole width This with can nevertheless been modified by the use of the keyword Gamma_hole Note that would have interest having previously performed a calculation in green mode with a minimum energy width keyword eimag at least 0 1 eV all this within an energy range convenient for photoemission 48 Manuel FDMNES D Parameter optimization It is possible to compare the calculated spectra to the experimental ones with the help of metric distances and R factor The metric distances to compare the individual theoretical fe iD and experimental Db spectra are given by exp D YN LU Fy e FO e Ma f i exp cae E max with each time the normalization factors c j f e de Emin The F are the integrals of f The R factor is a conventional one see E Zanazzi and F Jona Surf Sci 62 61 1977 given by Oy a e e
24. bration amplitude in relation to the interatomic distance This is done writing a second number under Gaussian Gaussian gt keyword 1 gt width of the Gaussian relative amplitude of the vibration d Other keywords To not take into account the Thomson factor fp as for a forbidden beam put Forbidden To get output spectra starting at a lower energy put Estart 8 gt Value of the starting energy eV When there are several indata files the program uses only the intersection of the energy ranges To use the total range put the keyword Nocut To get in the output file the f and f values put the keyword Fprime In the output file for each reflection there are three columns f f then intensity The intensity is for the mesh f and f are for one atom of the mesh f contains also fp To get only f put also the keyword forbidden AAT Manuel FDMNES To get the isolated atomic f and f Fprime_atom The values of the atomic f and f are then in a new output file with the extension _fprime ato txt It can be useful to impose specific value for each RXS reflections for the non resonant Thomson structure factors and the resonant one of the other atoms That is y Palha f atif a p2 Ra with f and f considered only for the other chemical species a and where pa is the site a occupancy rate This is the case for example for occupancy rate non egual to one These complex terms taken
25. e e OR RO Bee with c such that a 0 Jexp Leol i Then for n different spectra one gets the total metric distance or R factor gt Di gt pD and Ry s a 2 yR i l n i l n EG gO where p is the relative weight for the spectra i given by p EE ma min j l n It is possible to vary some parameters in order to optimize the agreement between calculation and experiment Then results are given for a multi dimensionnal grid of parameters The values of the metric distances are given in a special output file By default for a complete calculation the file name is the fdmnes conventional output file name with the suffix _fit txt For a calculation starting calculating the convoluted spectra the default output file name is fdmfit_out txt This name can be modified using a specific keyword To do this task different keyword must be added in the main indata file Experiment gt Keyword preceding the file names containing the experimental Nom_exp spectra Gen_shift gt Minimum and maximum energy shift between calculation and 7108 7114 21 experiment and number of value to test Metric_out gt Keyword preceding the output file name containing the metric File_name_fit txt distances Parameter gt Keyword preceding any group of correlated parameters Par_Gamma_max gt Keyword to specify the convolution broadening I as a parameter 49 Manuel FDMNES 15 25 5 First and last values and number of values Paramete
26. e will have the calculation in right polarization then the left polarization then the difference In this example the calculation corresponds to the polarization x iy and x iy II 10 Anomalous or resonant diffraction In case of anomalous or resonant diffraction DAFS DANES RXS calculation put the keyword dafs or rxs followed by the index of the beams to calculate The orientation of the polarization and wave vector can be described by different way When working in O 0 O 1 T T or Ms or in circular polarization put the number 1 2 3 4 or 5 for the polarization respectively 7 circular right circular left or rectilinear along a general direction in input then in output Then one specify the azimuthal angle between the incidence plane with a 7 J O base such that Dafs FDMNES User s Guide 002 I 2 45 gt reflection indexes 71 azimuth 002 I 1 45 gt reflection indexes 9 azimuth When the polarization is rectilinear but not or 7 but with an angle amp such that a 0 when it is O and 90 when it is 7 one must write Dafs 002 10 5 10 45 gt reflection indexes angle rectilinear angle azimuth 002 10 5 10 45 Note that in this case one must specify both incoming and outgoing polarization angles even when one of them is o 7 or circular When it is circular the angle is not taken into account LJ Q is such that Q is the normalized diffraction vector Ton lk A o J O I
27. ed atom Fe txt gt name of the file containing the type atom 8 220 2 214 type 2 atom calculated internally Once defined the atoms types the molecule or mesh description uses the atom type number and not the atomic number For example for the cfc copper crystal Atom c users joly dirac cu trdrsicu e01 gt excited copper file c users joly dirac cu trdrsIcu 001 gt standard copper file Crystal gt Crystal structure 3 610 3 610 3 610 90 90 90 gt a b c a B y 1 00 00 00 gt Type number position 1505 05 00 1505 00 05 1 00 05 035 The first index is now the atom type number but the excited one in the list under atom III 2 Uses of atomic electronic densities of Clementi and Roetti By default the program calculate the atomic basis using a Hartree Fock Dirac procedure If you prefer the Clementi and Roetti basis use the keyword Clementi 25 FDMNES User s Guide III 3 Non spherical atomic electronic densities When calculating using the finite difference method it is possible to define non spherical atomic electronic densities This is done under the keyword crystal or molecule For each non spherical atom one performs an expansion in spherical harmonics of these non spherical orbitals In practice on the line atom type position one adds an integer giving the number of non spherical orbitals then next lines the expansion in l m of each of these orbitals with at the end the number of electron
28. egrees a b c a B y 290 0 0 0 0 0 gt Atomic number position 29053 05 0 0 29 0 5 0 0 0 5 29 00 0 5 0 5 Arc gt to get convoluted spectra Efermi 6 End gt end of the indata file Two blocks are necessary for any calculations The first one starts with the keyword rayon or radius followed by the value of the radius inside the calculation is performed The second one is necessary to describe the material structure When that one comes from the output of the WIEN 2k package it begins with the keyword flapw In the other cases the molecule structure or the elementary mesh in case of a 3D structure is described in the file This description starts respectively with the keyword molecule or crystal or cristal 1 FDMNES User s Guide All the keywords related to the convolution or to the fit of the parameters are treated in chapter C and D Output file names By default the output file name is fdmnes_out This name can be modified by the use of the keyword filout followed by the name we want without extension There will have several output files by adding the extensions _bav txt output file giving details txt contains only the spectra by column When there are different edge calculated together results are scattered in different output files For example for the threshold L2 and L3 one gets the output files with the extensions _L2 txt et _L3 txt In the same way if a calculation is performed on
29. ence orbital 26 FDMNES User s Guide III 6 Mesh or molecule charge The unit mesh must be neutral A molecule is also often neutral A test is performed in the program to verify this neutrality If one wants to omit it put chlibre III 7 Exchange correlation potential By default the calculations are done using the real Hedin Lundgvist and Von Barth potential If one wants to use the Xalpha potential one must introduce the keyword Xalpha followed by the value of the corresponding parameter Xalpha 0 3333 value of the Xalpha parameter In the case of using the exchange correlation potential coming from flapw if one wants to keep this potential independent of the energy put the keyword xalpha with any value beneath The potential will not be Xalpha but the one calculated by flapw If one prefers to use the Perdew and Wang s potential put the keyword Perdew This potential will be used with the parametrization of Moruzzi Janak and Williams but keeping the energy dependency proposed by Hedin and Lundgvist III 8 Reference of the photoelectron wave vector By default the reference of the electronic wave vector and so of the kinetic energy outside the sphere of calculation and the muffin tin ground in case of muffin tin calculation is taken as the average of the potential between the absorbing atom and the first crown when calculating in green and at the outer sphere frontier in FDM It is possible to impose this grou
30. extension _sph_atomJ txt and _sph_xtal txt for the atom and crystal If one uses such file as indata for the convolution process one gets the f and f of each tensor component of the atom Another file _sph_int_atomJl txt contains the integral of the spherical tensors To get the contributions on the tensor components on each polarization and reflections put in place of spherical the keyword Sphere_all FDMNES User s Guide Then one gets a series of output files with the extensions _sph signal xan txt _sph_ signal poll txt _sph signal rxsl txt for the contributions on each polarization and reflections The number after po or rxs is the number of the polarization or reflection III 18 Calculation area boundary By default in FDM the meshing is performed in a sphere extending up to the last atom inside the sphere of radius given under radius plus the atomic radius by default 0 65 A plus one interpoint distance 0 2 A by default In order to use a bigger sphere put Overad 1 2 gt distance over the last atom its radius to take into account III 19 Displacement of the absorbing atom To move the absorbing atom in reference to its position given under molecule or crystal put dpos 0 20 00 0 gt displacement vector in Angstr m III 20 Getting back the tensor components from a previous calculation It is possible to get back the tensor component from a previous calculation is in the fdmnes_out_bav txt type fi
31. gy IV Compilation All the fortran routines must be compiled and linked with nesparam inc and mpif h in the same directory because these files are called during the compilation sub_util f is a file containing different routines from the Lapack and Blas library It can be advantageous to replace it by the calling to these libraries The dimensions are checked when running For most use the declared dimensions in nesparam inc are sufficient Nevertheless if one of them is too small a warning is given in the different output files and on the screen giving the parameter must be increased in nesparam inc After modifying it one must compile again When using a Cray one has to modify the calling to lapack routines have a different name zgetrf zsytrf zgetri zsytri in routine invcomp Lines to modify are spotted by the comment cray modify in the procedure timesec is in the file main f the calling to time changing the two lines tps4 secnds 0 _ 4 tps dble tps4 by the only line tps second V Parallelization Thanks to Keisuke Hatada Kuniko Hayakawa and Rainer Wilcke the user shaving the access to a cluster of computer can using the MPI library run the program in parallel mode For this one has to delete the files mpif h and not_mpi f when compiling and make the call to the corresponding library VI Running After compilation the program can be run following the usual procedure available on your syste
32. he new output file This is set after the keyword scan Scan Fe_rs64_01_scan txt gt name of the indata dafs file 1 Fe_rs64_02_scan txt gt name of the indata dafs file 2 Fe_rs64_scan_conv txt gt name of the new output file with the convoluted scan To specify a working directory put the keyword directory followed by the directory name with at the end the separator or Directory C Documents and Settings joly Mes documents xanes xanout v203 b Fermi level The states beneath the Fermi level are occupied Thus the cross section is set to zero before convolution By default this level is set to 5 eV Nevertheless most often this value must be specified with the keyword Efermi gt keyword to specify the Fermi energy Er relative to the calculation 5 5 The Fermi level is applied before the energy shift because in principle it is used to simulate the core level shift Anyway it is possible to apply it after the shift adding the keyword Dec It is also possible to have different values of the Fermi level for the different files This is written in third column after each file name 44 Manuel FDMNES Calculation gt To give the file names resulting from the previous step Fe_rs64_01 txt gt name of the file 1 1 0 0 2 5 gt weight shift Fermi level Fe_rs64_02 txt gt name of the file 2 1 0 0 2 5 9 gt weight shift Fermi level The Fermi energy written under Efermi when specified is then not con
33. idth This procedure seems improve the agreement with experiment especially in the pre edge range where the other procedure increase the background The author does not understand why To do this use the keyword Gamma_fix It is also possible to use the Seah Dench formula for the calculation of the broadening In this case one gets ALE 1 JE po y A ae Da ao with E E Ep In AE Ades Ta This is performed with the keyword Seah 1 20 gt Alm One can also use a simple table with the keyword table Table 0 0 1 gt Energy E Er and broadening I E Er 10 0 15 20 0 2 30 0 3 5 100 0 5 46 Manuel FDMNES When the keyword Efermi is used without the keywords Arc or Seah the width of the convolution is constant and egual to the width of the core hole In all cases the curve shape can be obtained in a specific output file named gamma_conv txt when using the keyword Check_conv To simulate the experimental resolution it is also possible to convolute by a Gaussian This convolution is performed after the lorentzian For this write Gaussian gt keyword 1 0 05 gt width of the gaussian This same convolution can be used as a first approximation to simulate the thermal vibration effects This is valid only in the extended part of the spectra In this case on apply the principal kr constant and the convolution width becomes proportional to the energy The parameter to write is the relative vi
34. ile copper_out_1 and copper_out 5 The summation over these different sites and the eventual energy shifts is to be performed during a following step II 5 Energy range The energy range E that one defines in the indata is the energy of the photoelectron relative to the Fermi level The kinetic energy Ec of the photoelectrons when they leave the cluster must be positive Ec is related to the input energy by Ec E W Vm where W is the work function and Vm the average potential eguivalent to the muffin tin ground in the cluster In the case of a calculation using the flapw the coulombian part of the average potential Vm is zero So Vm is greater than zero Thus the work function is not anymore subtracted and the input energy is defined relatively to the vacuum By default the energy range is 5 to 60 eV by 0 5 eV step One can change the range the step or even have a variable step using Range gt keyword for the energy range or gamme 1 0 5 60 gt Emin step Emax Other example with variable step Range 1 0 1 10 0 5 20 1 60 00 gt E min step E intermediate step To get a continuously increasing step k step constant put Rangel 1 0 1 200 gt Emin step at the Fermi level E max By default the output energy range is relatively to the Fermi level If one wants that the output energy is the photon energy put the keyword energpho FDMNES User s Guide II 6 Multiple scattering mode If one want
35. ine 4000 gt maximum number of character in one line 4 Manuel FDMNES 42 Manuel FDMNES C Convolution The fdmnes program allows performing 1 the convolution by a lorentzian of absorption spectra eliminating the occupied states For the anomalous diffraction it makes the integration over energy of the unoccupied states then calculates the intensity of the diffracted peaks 2 a weighted summation over different outputs of the fdmnes program This summation can include a relative shift between the spectra This step of the calculation can be done together with the previous one just adding the corresponding keywords It can also be done independently writing another input file contains only the keywords related to the convolution Here comes an example Calculation To give the file name resulting from the previous step g rs43 txt gt name of the file Conv_out gt To give the file name of the convoluted spectra g_rs43_conv txt gt name of the file Arc gt keyword to specify an arctangeant shape for the broadening Efermi gt keyword to specify the Fermi energy Er relative to the calculation 5 5 Keywords for the convolution a File names The indata files for the convolution step are the output files of the previous step whose names are defined under filout When the convolution is performed together with the main calculation it is not necessary to specify again this name In the other cases o
36. ique is to set the good number of d electron following the formal charge but keeping the neutral atom putting electrons in the large radius 4s or 4p orbitals for a transition metal II 4 Absorbing atoms All the atoms present in the structure participate to the scattering By default the calculated spectra correspond to the sum of the scattering produced by all the atoms of the same atomic number than the first one in the list under crystal or molecule When using the keyword extract or one wants to select some of the absorbing atoms one has to use the keyword absorbeur For example if one wants that the absorbing atom is the n of the list and only this one under crystal or molecule put Absorbeur 3 gt absorbing atom number here the 3 in the list If this example goes with the FeOs structure defined just above the result in the output file would correspond to a calculation of an oxygen atom summed over the 6 atoms because they are equivalent by symmetry If one wants the result from a unique site use the keyword symsite or allsite see further on FDMNES User s Guide If there are non equivalent sites one has to perform independent calculations For this one must specify each atom number under absorbeur The name of each of the corresponding output file get the suffix _n where n it the atom number Filout copper_out Absorbeur 15 atom numbers from the results will be put in the output f
37. it contains Optionnaly a rotation of the local atomic basis can be performed using the Euler angles see keyword spin axis for the Euler angle definition In this case the Euler angles must be given in the line in front of the atom Example molecule 1 9 1 9 1 9 90 90 90 gt a b c a B y 0 45 0 gt 45 rotation around the Oy axis 1 0 0 0 1 there is one non spherical orbital 0 0 0 0 0 0 1 0 0 1 gt lm in the natural order here l 2 m 0 has 1 2 10 00 00 electron 2 10 00 0 0 2 00 10 0 0 2 00 10 0 0 III 4 Orbital dilatation It is possible to modify the valence orbitals defined above dilating or contracting them This can be very useful for ionic material for instance oxygen 2 where the atomic bases are calculated for neutral atoms For this purpose introduce the dilatorb then for each orbital one wants to dilate the atomic type number in the list atom the valence orbital number and the expansion coefficient Dilatorb 3103 3203 III 5 Screening By default the screening is one electron on the first non full valence orbital of the absorber If one wants to modify this value or the number of the orbital put the keyword screening or ecrantage followed by the quantum number of the valence orbital and the value in number of electron better inferior to 1 of the screening In this case the cluster is not anymore necessary neutral Screening 32 0 2 gt n l and screening on the val
38. le and to compute other polarizations putting the key word extract followed by the corresponding name Extract c users joly xanout calcul_prec_bav txt In case of calculations for different absorber position whose number is the number of output files one has to extract the tensors from these n files if there are n non equivalent absorbing atoms The use of the keyword absorbeur is then absolutely necessary The list of the files under extract must correspond in number and in order to the list of atoms under absorbeur Extract c users joly xanout calcul_prec_1_bav txt c users joly xanout calcul_prec_2_bav txt Absorbeur 12 It is possible to make rotations of the local repair It can be useful when using tensor calculated in a different base It is possible de make 1 2 or 3 rotations around the axes 0z then Ox then Oy Each rotation defines a new local base starting point of the next rotation For this write FDMNES User s Guide Rotsup 30 0 45 gt 30 rotation around Oz 0 around Ox then 45 around Oy For the same purpose on can apply symmetry operation on the extracted tensors They are codified as explained in the manual with the symsite keyword They must be given in the same order than the extracted files These codes are given after the keyword extractsym extract xanout fe3o4 w_g ch015_3d55_rs80_1_bav txt xanout fe3o4 w_g ch015_3d55_rs80_1_bav txt Extractsym 125 gt code 1 for identity 2
39. lus time reversal in case of magnetism Symmetry code 1 identity 8 rot 27 3 around 1 1 1 9 rot 47 3 around 1 1 1 2 rot 27 3 around 1 1 1 3 rot 47 3 around 1 1 1 10 rot 27 2 around 1 1 0 4 rot 27 3 around 1 1 1 11 rot 27 2 around 1 1 0 5 rot 47 3 around 1 1 1 12 rot 27 2 around 1 0 1 6 rot 27 3 around 1 1 1 13 rot 27 2 around 1 0 1 7 rot 47 3 around 1 1 1 14 rot 27 2 around 0 1 1 36 FDMNES User s Guide 15 rot 27 2 around 0 1 1 39 rot 47 3 around 1 1 1 and inversion 40 plane perpendicular with 0x 16 rot 27 4 around 0x 41 plane perpendicular with 0y 17 rot 27 4 around Oy 42 plane perpendicular with 0z 18 rot 27 4 around 0z 19 rot 27 4 around 0x 43 diagonal plane y z containing Ox 20 rot 27 4 around Oy 44 diagonal plane x z containing Oy 45 diagonal plane x y containing Oz 46 diagonal plane y z containing Ox 47 diagonal plane x z containing Oy 48 diagonal plane x y containing Oz 21 rot 27 4 around 0z 22 rot 27 2 around 0x 23 rot 27 2 around Oy 24 rot 27 2 around 0z 49 rot 27 3 around 0z 50 rot 47 3 around 0z 51 rot 27 6 around 0z 52 rot 107 6 around 0z 53 rot 27 3 around Oz negative axe 54 rot 47 3 around Oz negative axe 55 rot 27 6 around Oz negative axe 56 rot 107 6 around Oz negative axe 25 inversion 26 rot 27 4 around Ox and inversion 27 rot 27 4 around Oy and inversion 28 rot 27 4
40. m The keyword crystal or molecule becomes also unnecessary Then must stand the names of the 5 FLAPW output files in case of spin unpolarized calculation and the 7 files in case of spin polarized calculation Flapw gt names of the output FLAPW files tio2 struct structure and symmetry tio2 vcoul coulombian potential tio2 r2v exchange correlation potential tio2 clmsum electronic density tio2 tils initial wave function In case of polarized calculation the keyword magnetism must be before the keyword flapw Magnetism Flapw tio2 struct structure and symmetry tio2 vcoul coulombian potential tio2 r2v exchange correlation potential spin up and spin down tio2 clmsum total electronic density tio2 clmup valence electron electronic density spin up tio2 clmdn valence electron electronic density spin down tio2 tils initial wave function The last file contains the wave function of the initial core orbital If one is ok with the initial wave function calculated internally one can avoid it One must in this case substitute the keyword flapw by flapw_psi If one wants to have an energy dependant exchange correlation potential one must add the keyword hedin By default the absorbing atom is the first one in the struct file list If one wants that it is the n put the keyword absorbeur n FDMNES User s Guide III 26 Jump in the indata file It is possible to jump over a part of the information written in
41. m The files xsect dat and spacegroup txt must be set in the same directory than the executable file because they are used in some cases As soon as the program is running it calls the file fdmfile txt This file must be in the same directory than the executable file It only contains the number of independent calculation to perform followed the name of the indata file of each of these calculation For example Indata file for fdmnes 1 gt number of indata files example cu cu_inp txt gt name of the indata file FDMNES User s Guide When an error stops the program execution a file called fdmnes_error txt is in principal created This file contains a message explaining the error If when starting to run the program immediately stops one has to check several points 1 The called files are in the good directory 2 Their names are OK under linux upper and lower cases must be respected 3 When using downloaded indata files some problems of compatibility between systems can occur It can be better to write again completely these files 4 Avoid the tabulation VII Structure of the calculation The program allows the calculation of spectra from a grid of parameters position charge to convolute them with other parameters convolution width energy shift then to compare them to experimental spectra with the help of objective criteria These different steps can be performed together or separately For a complete
42. mber of valence orbitals then quantum numbers n 1 and population in number of electron in each of these orbitals Following line 600 number of radius where are calculated the electronic densities and wave function Following line r A rho ua 3 psi 3D _ psi 4S gt comment Following line By column column 1 radius in angstrom column 2 total electronic density ua V4nry r column 3 4 wave function of the valence orbitals multiplied by 447r 57 The absorbing atom type 1 must include one column more containing the value of the initial wave function for example ls for a K threshold before the absorption of the photon Manuel FDMNES fer excite neutre E01 3d7 4s2 Orbitale n l j _ Popul Energie IS 1 0 0 0 0 0 1 000 7204 473 2S 2 0 0 0 0 0 2 000 859 982 2P 2 0 1 0 0 0 6 000 763 986 3S 3 0 0 0 0 0 2 000 96 485 3P 3 0 1 0 0 0 6 000 64 835 3D 3 0 2 0 0 0 7 000 9 151 4S 4 0 0 0 0 0 2 000 5 544 NUCLEAR CHARGE 26 000000 INTEGRAL OF CHARGE DENSITY 26 000000 26 17 00 2 32 7 000 40 2 000 600 r A 0 00016112638 0 00016413798 0 00016720588 0 00017033115 rho ua 3 _psi 3D psi 4S 6572 94900 0 0000000052 0 0023377680 6571 00800 0 0000000055 0 0023811100 6569 02300 0 0000000058 0 0024252490 6567 00900 0 0000000061 0 0024701990 58 pssi 0 0784658600 0 079920653500 0 0814021800 0 0829109400 Absorbeur Adimp Allsite Ang spin Arc Atom Axe spin Azimuth Base_spin Basecomp Calc
43. ment and calculation When there is a single spectra Rxg is eguivalent to Rx To get this confidence factor put the keyword Rxg 52 Manuel FDMNES E Extraction of DAFS scans and spectra When realizing a DAFS simulation with azimuth dependence scan the output scan file after convolution contains the intensity of all the reflections at all energies and all azimuth angles Often it is useful to extract from this big file some spectra at specific azimuth angle or some scan at specific energy It is what is done in this part with the following indata file Indata file for FDMNES Selection part Selec_inp gt keyword for the input file name output of the convolution part xanout fe304 2008_bland cc_1221 1221 orig Iml0_scan_conv txt Selec_out gt keyword for the output file name xanout test fe304_scan_selec_conv txt Energy 4 Reflection 2569 Azimuth End gt keyword for the selected energy for the scan gt value of the energy for the scan in eV gt keyword for the selected reflections gt number of the selected reflections gt keyword for the selected azimuth angle If no number is given beneath keyword Azimuth it is azimuth scan are extracted at the energy in eV given after keyword Energy On the contrary if no number is given after Energy this is spectra are extracted at the azimuth given after keyword Azimuth It is possible to have several reflections in the o
44. n polarized calculation put the keyword Magnetism 20 FDMNES User s Guide By default this calculation is done neglecting the spin orbit coupling If one wants to take into account this last put the keyword Spinorbite The keyword magnetism is thus useless If the polarized potential comes from LAPW spinorbite or magnetisme must be before flapw With the spin orbit interaction the calculation is automatically relativistic To get a non relativistic calculation put Nonrelat When the calculation does not use the flapw output the use of the keyword atom is necessary to specify the electronic configuration different for the spin up and spin down parts The orbital occupancy must be given for each spin in doubling the corresponding columns Example of metal fcc nickel Magnetism Atom 28 2 325 4 40 0 5 0 5 Crystal 3 52387 3 52387 3 52387 90 90 90 1 0 0 0 0 0 0 1 0 5 0 5 0 0 1 0 5 0 0 0 5 1 0 0 0 5 0 5 The configuration is 3d 4s with 5 electrons 3d up and 4 electrons 3d down 0 5 electron 4s up and 0 5 electron 4s down In case of an antiferromagnetic structure one must put a minus sign in front of the atom type number to specify the atom with the reverse spin For example in V203 monoclinic Atom 23 3322 0 401 1 410 505 8 2201 1 212 2 Crystal 7 255 5 002 5 548 90 0 96 75 90 0 a b c alfa beta gamma 1 0 34380 0 00080 0 29910 VS 1 0 65620 0 99920 0 70090 V6 1 0 84380 0 99920 0 2
45. n the spheres If one wants modify this put the keyword overlap followed by the value of the overlap Overlap 0 15 gt 15 overlap If one wants that the muffin tin radius be determined by the value it gets when the atomic potential is equal to the interstitial potential put the Keyword rmtv0 followed by the value of this potential takes place the one define by the keyword v0imp Rmtv0 12 If one wants to impose the radius for each chemical species put rmtimp followed by the radius in Angstr m for each of them Rmtimp 0 65 1 1 The program gives in the output file the atomic charge integrated up to the muffin tin radius If one wants to get this value for a different radius put Rchimp gt corresponding keyword 1 2 gt radius value for each type of atom 0 7 To limit potential jump between the interstitial area and the muffin tin sphere a linear interpolation on the potential is performed in the last 10 of the radius For this put Supermuf IV 14 Number format In the output files the numbers are written with 15 characters with 7 after the point To change this put Length_word 16 gt number of characters between 11 and 17 IV 15 Line length Lines are red over a specific length When reading files containing a large number of columns for example calculation of more than 100 RXS spectra the default line length 3020 characters can be not sufficient To modify this put 40 Manuel FDMNES Length_l
46. nd potential at the Fermi energy The dependence versus the kinetic photoelectron energy is then added automatically to this term Put then the keyword V0imp 11 5 value of the ground potential at the Fermi energy III 9 Maximum potential For a calculation performed on a molecule the potential increases when going away from the atoms When this molecule is not in the gaz phase but in solution or in any relatively dense surrounding this incease is artificial In case of calculation under the finite difference mode this reach to faulse bounded states very thin in energy It can be usefull to give a maximum value to this potential to avois this phenomena with the use of the keyword vmax followed by the value of this potential oy FDMNES User s Guide Vmax 6 III 10 Complex energy It is possible to use a complex energy when working in the multiple scattering mode green For this one has to specify the imaginary part positive of this energy in a table versus the photoelectron energy under the keyword eimag Eimag 0 0 5 10 0 7 30 3 50 5 100 6 When a uniform broadening is sufficient it is not necessary to specify the photoelectron energy Eimag 0 1 gt value of the uniform width eV The use of a small width 0 1 eV is sometimes useful for the calculations at low kinetic energy of the photoelectron because the localized level 3d or 4f can be too thin in energy to be correctly evaluated It i
47. ne has to introduce it or them with the keyword Calculation Calculation gt To give the file names resulting from the previous step g rs43 txt gt name of the file Another example with 2 indata files and with different shifts Calculation gt To give the file names resulting from the previous step Fe_rs64_01 txt gt name of the file 1 1 0 0 2 gt weight shift Fe_rs64_02 txt gt name of the file 2 1 0 0 2 gt weight shift The weight is applied on the XANES and the anomalous scattered amplitudes For the RXS calculations the outpout files before convolution contain in second line the Thomson AA Manuel FDMNES factors plus the anomalous par of the non resonant atoms The weighting applied on these terms is the weight divided by the sum over the files of the different weights In this way is taken into account the average configuration for substitutions of atoms of different chemical specie on the same sites but when it is the resonant atom has not an occupancy rate equal to one In this last case one also has to use the keyword Thomson The name of the convolution output file is by default the input file name with the added suffix conv txt Anyway it is possible to impose another name with the keyword Conv_out Fe_rs64_sum_ conv txt gt name of the convoluted spectra file If there is an azimuthal scan in anomalous diffraction one has also to give the corresponding indata file names type _scan txt and t
48. on molecule 1 900 1 900 1 900 90 90 90 a b c a B y 26 00 00 0 0 gt Atomic number position 8 10 00 00 8 10 00 0 0 8 00 10 0 0 8 00 10 0 0 8 00 00 10 8 00 00 1 0 The atomic structure can also be given in cylindrical or spherical coordinates To use cylindrical coordinates it is sufficient to give only two numbers under molecule The program will understand they are a and c and that the positions of the atoms are given by r and z Thus to describe the same octahedron as previously molecule 1 900 1 900 gt a c FDMNES User s Guide 26 0 0 0 0 0 0 gt Atomic number position 8 1 0 00 0 0 8 10 180 0 00 8 1 0 90 0 0 0 8 10 90 0 0 0 8 0 0 00 1 0 8 0 0 0 0 1 0 To use spherical coordinates only one number a must be set after molecule The position of the atoms is then given byr 9 For the same octahedron molecule 1 900 gt a 26 0 0 0 0 0 0 gt Atomic number position 8 1 0 90 0 0 0 8 1 0 90 0 90 0 8 1 0 90 0 180 0 8 1 0 90 0 270 0 8 1 0 0 0 0 0 8 1 0 180 0 0 0 It is possible in the non magnetic case to specify only the non equivalent atoms Then one has to give the space group under the keyword Spgroup The complete name as in the international table must be given For example for the magnetite one gets Spgroup Fd 3m 1 it is also possible to write 227 1 Crystal 8 3940 8 3940 8 3940 90 0 90 0 90 0 26 6250 6250 6250 Fel6d 26 0000 0000 0000 Fe 8a 8 3800 3800 380
49. r Par_Posx 0 05 0 05 3 Par_Posy 0 05 0 05 3 In the previous example the parameters Posx and Posy of the atom 1 are completely correlated Thus the atom is displaced along the diagonal The parameters can be fitted are For the convolution Par_ecent gt Central energy for the arctangente Par_elarg gt Energy width for the arctangente Par_efermi gt Fermi energy Par_gamma_hole Par_gamma_max gt Hole width gt Maximum width for the final states Par_gauss gt Gaussian width or resolution Par_shift gt Energy shift Par_aseah gt First parameter of the Seah Dench formula For the spectra calculation Par_a gt Contraction or expansion of the mesh parameter a in Par_b gt Contraction or expansion of the mesh parameter b in Parc gt Contraction or expansion of the mesh parameter c in Par_abc gt General contraction or expansion in Par_anga gt Value of the unit mesh angle o Par_angb gt Value of the unit mesh angle B Par_ange gt Value of the unit mesh angle y Par_poporb gt Orbital occupancy Par_posx gt Atom position along x Par_posy gt Atom position along y Par_posz gt Atom position along z Par_dposx gt shift of the atom position along x from the original position Par_dposy gt shift of the atom position along y from the original position Par_dposz gt shift of the atom position along z from the original position Par_theta gt position along 9 for an atom in spherical coordinate
50. rt Sphere Newind Mat MSM Coabs Values can go from 0 to 4 giving more and more for each routine By default there is 1 for all the routines One can also write check_all gt equivalent to icheck 3 for all the subroutines no_check gt equivalent to icheck 0 for all the subroutines check_pot gt equivalent to icheck 3 for all the subroutines concerning the potential check mat gt equivalent to icheck 3 for the matrices MSM or FDM check_sph gt equivalent to icheck 3 for the subroutine sphere check_coabs gt equivalent to icheck 3 for the subroutine coabs IV 2 Symmetry The point symmetry is calculated automatically Anyway the expert users can impose it by the keyword Sym 2 m gt Schoenflies coefficient or international table of crystallography coefficient Be careful that these symmetries are possible only for the mesh axis defined in the indata IV 3 Tensor basis By default the atomic tensors given in the output files are expressed in the orthonormal basis R where z is the crystal c axis but for the trigonal structure where z is the c axis of the associated hexagonal mesh x is along bxc and y along zxx Nevertheless it is possible to express it in the basis where it is in fact calculated the so called basis R2 From R another rotation is performed for magnetic calculation when the spin axis is not along z Then a another 45 rotation is sometimes performed when a 2 fold axis or a symmetry plane
51. s even truer for photoemission calculation see keyword photoemission If this broadening is small it has no effect on the forthcoming convolution to take into account the widths of the hole and final states HI 11 Radius of the cluster for the superposition of the potential The potential inside the sphere of calculation with a radius R set under the keyword radius is calculated by superposition To avoid frontier problems the atoms taken into account for the superposition are all the atoms inside the calculation sphere plus an outer shell By default this outer shell is 2 5 A thick giving thus a new sphere with a R 2 5 radius If one wants a bigger radius useful for the oxides put the keyword Rpotmax 15 radius of the cluster for superposition in angstr m 28 FDMNES User s Guide III 12 1 approximation for the selection rule If one wants to make the l71 approximation where only the transition Al 1 is authorized so one neglects AL 1 for example for the threshold Ly and Lyy the transitions are only toward the d states put the keyword lplus I In the same way if you want only the l 1 states put Iminus1 III 13 Rydberg series One can add an outer sphere having a 1 r potential to analyze Rydberg series In this case put the keyword Rydberg Rydberg III 14 Cluster origin By default the cluster origin is set on the absorbing atom If one want that this origin is as in the cluster indata put
52. s to calculate in the multiple scattering mode use the keyword Green Then the potential is automatically a muffin tin one The mode is faster than the finite difference method so one has to use it first II 7 Threshold type By default the threshold is the K one For other threshold put the keyword Edge gt keyword preceding the threshold type or seui Ll gt threshold K L1 L2 L3 M1 It is possible in a single run to calculate 2 edges with the same initial n l that is the edges L2 and L3 or M2 and M3 or M4 and MS For this write Edge gt keyword preceding the threshold type L23 gt threshold L2 and L3 or M23 M45 N23 N45 In this case one gets two output files with the suffixes _L2 and _L3 II 8 Multipolar expansion By default only the dipolar component is calculated This is modified by the keywords quadrupole gt quadrupolar calculation octupole gt octupolar calculation nondipole gt No calculation of the dipolar dipolar component nonquadrupole gt No calculation of the quadrupolar quadrupolar component noninterf gt No calculation of the dipolar quadrupolar component nonoctupole gt No calculation of the dipolar octupolar component II 9 Polarization and dichroism By default the calculation is performed along 1 2 or 3 orthogonal polarizations in the dipole mode and up to six in quadrupolar depending of the symmetry The polarizations are chosen along the axis of an internal basis in
53. s to m polarization It is possible to write the exact polarization directions This can be useful for peculiar uses as in photoemission In case of rectilinear polarization write Dafs 000 gt Reflection index 0 0 1 0 1 0 gt Ee Ve 0 0 1 0 7071 0 7071 0 gt E Vs For circular polarization it is complex Dafs 000 gt reflection index 0 7071 0 7071 0 0 0 0 0 1 0 gt Ee Ve Exr Exi Eyr Eyi Ezr Ezi Vex Vey Vez 0 7071 0 7071 0 0 0 0 0 1 0 gt Es Vs When one does not want phase term between atoms for example to simulate photoemission just write 0 0 0 as reflection index Note that it is possible to choose any other origin for the azimuth just using another vector than k or i to define the basis vector J and J For this just write Zero_azim 0 1 1 gt vector in the direct crystal base It is now possible to calculate the self absorption corresponding to the incoming and outgoing photon polarizations For this put the keyword Self_absorption Then one gets in the output files after each reflection two new columns containing these data The unit like for XANES is the Megabarn They contain after the convolution also the absorption coming from the other atoms and the other edges That is that the absorption before the edge is not zero The new data allows the correction due to the self absorption in order to compare with experimental spectra II 11 Spin polarized calculation If one wants to make a spi
54. several crystallographic sites one gets the extensions _i txt _j txt where i and j are the index of the sites see keyword absorbeur In option or depending on the type of calculation one can also get the files _conv txt convoluted spectra scan keyword arc _scan txt dafs versus angles for azimuthal scan keyword rxs _sda txt state density for the atom number a keyword density _atoma txt results for one atom at position number a keyword allsite _atoma _ scan txt rxs scan results pour for the atom a keyword allsite and rxs For the analysis of the spherical tensors keyword spherical and sphere_all _sph atoma txt spherical tensors of the atom a _sph_atoma_int txt integral of the spherical tensors of the atom a sph signal atoma xan txt contribution of each atomic spherical tensor on the average xanes signal sph signal atoma poli txt contribution of each atomic spherical tensor on the xanes polarisation number i sph signal atoma rxsi txt contribution of each atomic spherical tensor on the rxs reflection number i sph signal xtal xan txt contribution of the crystal spherical tensor on the average xanes signal sph signal xtal rxsi txt contribution of the crystal spherical tensor on the rxs reflexion number i _sph_xtal txt spherical tensors of the crystal _sph_xtal_int txt integral of the spherical tensors of the crystal _sph_xtal_rxsi txt spherical tensor of the crystal for the rxs reflexion number i
55. sidered c Convolution width The convolution to apply depends on the core level width Myoie and the final state width I woie is automatically determined Nevertheless it is possible to modify it with the keyword Gamma_hole 2 05 gt Core level width I Hoe in eV There are different ways to calculate the energy dependant broadening The first one uses an arctangent formula I T L Hole E arctan e Et arg e E E l with e where I n Ecent and Eiarg are respectively the maximum width at high cent energy of the final state the center and the width of the arctangent function The depth at the center of the arctangent is I m Eiarg A typical curve is given bellow Width eV lambda_ A 40 60 Energy eV 45 Manuel FDMNES Such a convolution is introduced with the keyword Arc When no number is specified under arc default values for Eiarg Ecent and Im are used respectively 30 30 and 15 eV It is possible to modify them given them in order under arc Arc 30 30 20 gt Ecents Elarg I m It is also possible to modify these parameters individually Ecent 3 0 gt Ecent Elarg 30 gt Elarg Gamma_max 20 2 I m In the convolution along the integration it is the width of the running energy which is taken It is possible to use the width of the final state energy corresponding to the energy of the elastic photon One then makes the integration with a constant w
56. son with the experimental spectra and Extraction of azimuth scan or spectra can be set in the same indata file but they are explained separately in the chapter C D and E FDMNES User s Guide B Main indata file 1 General Structure It contains most of the inputs necessary for the calculation All the data in input and output are in Angstr m and electron Volt Many parameters are chosen by default One can modify those using keywords Text can be in upper or lower case Blank lines or beginning by are not taken into account Between number one must put at least one blank Tabulations are forbidden When getting problem when opening these indata files one has to check if their name is correct Moreover some compilers do not like files written under other system MAC DOS LINUX In case of difficulties when the program wants to open one of these downloaded files it can be useful to completely write them again The indata file contains several blocks of data each ones starting with a specific keyword The end of the indata file is noted by the keyword end Whatever is after is not red Here comes an example of indata file Fdmnes indata file Calculation for the copper K edge in copper cfc Filout example cu cu_out gt Name of the output files without extension Range 10 0 2 0 0 5 10 1 40 Radius gt Cluster radius 3 0 Crystal crystal structure 3 610 3 610 3 610 90 90 90 gt Mesh parameters A and d
57. ticular the description of the indata files An introduction to x ray spectroscopies is available in French The FDMNES program can be freely downloaded at the web address http www neel cnrs fr fdmnes In case of publication related to the use of the program thanks to cite Y Joly X ray absorption near edge structure calculations beyond the muffin tin approximation Phys Rev B 63 125120 2001 The FDMNES program highly benefited from the scientific contribution of Calogero Natoli who has provided a constant and essential help He is in particular at the origin of all the developments using the multiple scattering theory and the extensions to resonant diffraction and magnetism The program also benefited from the expertise of Delphine Cabaret Hubert Renevier Sergio Di Matteo Christian Brouder and Emilio Lorenzo without whom different advances would have not been realized Finally this work has been made greatly easier with the support of Denis Raoux FDMNES User s Guide FDMNES User s Guide Running the program A General presentation B Main indata file C Convolution D Parameter optimization E Extraction of DAFS scan and spectra F File nesparam inc G File atomic electronic densities List of the keywords of fdmnes 11 43 49 53 55 57 59 FDMNES User s Guide FDMNES User s Guide A General presentation I Computer configuration FDMNES run on all the computers having at least 256 Mo of R
58. ulation Cartesian Centre Check Check all Check coabs Check conv Check mat Check pot Check sph Chemin Chlibre Clementi Comment Conv_out Crystal Dafs Density Dec Dilatorb Directory Dpos Ecent Edge Efermi Eimag Elarg Emax Emin Endjump Energpho Energy Epsii Estart Etatlie Experiment Extract Extractsym 16 38 33 23 46 16 23 53 35 36 43 32 29 35 35 35 45 35 35 35 32 27 25 34 44 14 19 30 44 26 44 31 46 18 44 28 46 51 51 34 17 53 32 47 38 49 31 32 Manuel FDMNES List of the keywords of fdmnes 59 Filout Flapw Flapw_psi Forbidden Fprime Fprime_atom Gamma fix Gamma hole Gamma max Gaussian Gen shift Green Hedin Iord Jump Kev Ldipimp Length line Length word Lmax Lmaxfree Lmaxso Lminusl Lplus1 Lguaimp Magnetism Metric out Molecule Muffintin No_check No res mag No_res mom Nocut Noncentre Nondipole Nonexc Noninterf Nonoctupole Nonguadrupole Nonrelat Normaltau Norman Octupole Overad Overlap Parameter Perdew Photoemission 12 33 33 47 47 48 46 45 46 47 51 18 35 38 34 51 37 41 40 39 39 39 29 29 37 21 49 14 38 30 34 34 47 29 18 38 18 18 18 22 38 39 18 31 40 49 27 48 Polarise Quadrupole Range Rangel Raydem Radius Rchimp Relativism Reflection Rmtimp Rmt Rmtv0 Rotsup Rpotmax Rxg Rxs Rydberg Scan Seah Screening 18 18 17 17 39 14 40 23 53 40 38 40 32 28 52
59. utput file at one energy or one azimuth It is also possible to have several energies but only one reflection for the scan or several azimuth angles but one reflection for the spectra Energy 4 7 5 Reflection 2 Azimuth or Energy Reflection 2 gt selected energy for the scan gt reflection number 2 selected gt No azimuth given one gets scan gt no energy is given one gets spectra gt reflection number 2 selected 53 Manuel FDMNES Azimuth 30 60 90 gt Azimuth selected 54 Manuel FDMNES F File nesparam inc This file contains dimensions of tables used by fdmnes It s introduced in each routine by the instruction include If one of these dimensions is overpasses during a run an error message is given with the name of the parameter to increase nesparam inc file File containing the table dimensions of the fdmnes program When these dimensions are modified the program must be compiled again Dimensions can be modified parameter nlfm 52 Max number of l m for the selection rules Dimensions cannot be modified parameter nnlm 34 Max number of orbitals in the excited atom parameter nopsm 64 Number of symmetry operation parameter nrepm 12 Max number of representation parameter nslapwm 48 Max number of symmetry matrix for the FLAPW data Constants complex kind 8 img parameter img 0 _8 1 8 parameter bohr 0 529177249 8 p

FDMNES User's Guide

Contents

Download Pdf Manuals

Related Search

Related Contents