Dr. Wantana Klysubun (Manager of Beamline 8, wantana@slir.or.th) Workshop at Kantary Hills Hotel, Chiang Mai 2 September 2011 2 September 2011 By Monte Carlo Simulation and Molecular Dynamics in physics Research Laboratory, CMU and ThEP () Synchrotron Light Research Institute (Public Organization) Standard procedure for XAS data processing 1. Get to know data file from BL8 : Which columns store photon energy and absorption values? 2. Data processing with Athena 2 1 O d t fil f l 2.1 Open data files of your sample 2.2 Plot all spectra, inspect them, delete the bad ones 2.3 Align spectra 2.4 Merge spectra into one spectrum (from now on, the merged spectrum will be g p p ( , g p processed) 2.5 Normalize spectrum 2.6 Save all your work and a text file of normalized spectrum 3. Linear combination fit by Athena 4. Peak fit by Athena 5. Fitting EXAFS spectrum by Artemis 5.1 Import merged spectrum 5.2 Create atomic model by Atom 5 3 Generate scattering paths by FEFF 5.3 Generate scattering paths by FEFF 5.4 Select important paths and define parameters of each path 5.5 Fit parameters () Synchrotron Light Research Institute (Public Organization) 2 1. Get to know data formats at BL8 At BL8 d t fil di t (1) t f l At BL8 we name a data file according to (1) type of sample, (2) Mode of XAS measurement, (3) type of detector, (4) XANES or EXAFS spectrum and (4) time when the file was Saved. For examples, -Open a file with Notepad or Wordpad to see data stored in each column. What column is photon energy ? File#1 : NiNO3 EXAFS 0623_1 File#2: NiNO3 EXAFS 0623_2
File#6: NiNO3 EXAFS 0623_6
What column is I0? What column is absorption? The file name #1 tells that it is an EXAFS spectrum of NiNO3 (which actually is Ni(NO3)2), measured in transmission mode, saved at 6:23 am on that day. There are six successive scans for this sample. What column is absorption? File#7 : C029 FLGe EXAFS 2233_1 Sample =C029 Mode =Fluorescence yield Detector =13-element Ge detector File contains EXAFS data File#8: xxx FLSi 0000 File#8 : xxx FLSi 0000 Sample =xxx Model =Fluorescent yield Detector =Si drift detetor () Synchrotron Light Research Institute (Public Organization) 3 2. Data processing with Athena 4 Plot multiple spectra EXERCISE 1 : Normalize NiNO3 EXAFS data 1. Run Athena 2. > File> open files select all NiNO3 EXAFS xxxx files 4. Plot multiple spectra select spectra, then click on and select all NiNO3 EXAFS xxxx files mark column 1 for photon energy mark column 6 for numerator 3. Plot one spectrum highlight filename then click highlight filename, then click on plot in mu (E) () Synchrotron Light Research Institute (Public Organization) 4 2. Data processing with Athena 5 Plot all spectra - plot as smoothedderivative 5. Plot all spectra 6. Delete bad spectrum, if any highlight filename > group actions > remove group plot as smoothed derivative - fit as smoothed derivative - Click on Are they consistent? 7. Align spectra Athena will do the followings: > Data > Align scans - select standard spectrum that we think there is no shift of energy - select other spectrum that will be aligned to the standardspectrum () Synchrotron Light Research Institute (Public Organization) 5 standard spectrum 2. Data processing with Athena 8 Align all other spectra 8. Align all other spectra 9. Return to the main window see E shift on each spectrum We can calculate the mean value of E shifts and use it as the uncertainty in photon energy of your experiment ! 10 Plot all spectra 10. Plot all spectra 11. Merge all spectra > Merge> merge marked data in mu(E) 12. Rename merged spectrum double click on the old name (merge) andreplace with NiNO3 mergedEXAFS and replace with NiNO3 merged EXAFS 13. Save all your work > File > Save entire project () Synchrotron Light Research Institute (Public Organization) 6 2. Data processing with Athena 14 Normalized merged spectrum 14 . Normalized merged spectrum -highlight NiNO3 merged EXAFS - mark to plot pre-edge line and post edge line -Select good region that can be fit by a linear function Pre-edge range from eV to eV relative to E0 - What is E0? () Synchrotron Light Research Institute (Public Organization) 7 2. Data processing with Athena - What is E0? - What is E0? - Plot the 1 st derivative of the spectrum, select E0 at the center of the first peak to get () Synchrotron Light Research Institute (Public Organization) 8 2. Data processing with Athena -- select regionafter the edge which -- select region after the edge which will be your normalization range. Arthena will fit data in this range with a polynomial function (purple line) of order 1, 2 or 3 Normalization range from . eV to eV - Click on to select the order of the polynomial function - Recall that normalized (E) =(E)/A A is the edge step =post(E0) - Never flatten the normalized spectrum, so unclick it g p p ( ) - What is your edge step ? () Synchrotron Light Research Institute (Public Organization) 9 2. Data processing with Athena 15 Plot normalized spectrum 15 . Plot normalized spectrum 19. Plot all normalized spectra 16 . Save your work 17 S t t fil f th li d t 20 E t t EXAFS ill ti Chi(k) 17. Save a text file of the normalized spectrum > File > Save norm(E) 18. Normalize other spectra with the same parameters 20. Extract EXAFS oscillation or Chi(k) - Highlight NiNO3 merged EXAFS, plot mu(E) and background Highlight NiNO3 merged EXAFS, mark all other spectra > Value > Set all marked group to the current () Synchrotron Light Research Institute (Public Organization) 10 2. Data processing with Athena 21. Plot EXAFS in k space - click on - select region for background absorption, Athena will use a spline function (red line) to fit background of the data in this region _(E) = {(E)- 0(E) }/ A 0(E) =background absorption(red line) () Synchrotron Light Research Institute (Public Organization) 11 0(E) =background absorption (red line) 2. Data processing with Athena 22 Pl t ll t i k 24 S hi(k) 22. Plot all spectra in k space 24. Save chi(k) > File > Save chi(k) > k^2*chi(k) 25. Fourier transform - set good k range for the Fourier transform -mark to show the Fourier transform function that you have selected If you have more scans, you will get less noise, especially i hi h k i ! that you have selected. in high k region ! 23. Plot EXAFS in k space with other k-weight - click on What different do you see? () Synchrotron Light Research Institute (Public Organization) 12 2. Data processing with Athena Pl hi(R) li k - Plot chi(R), click on -What are k range, dk of the window and window type 26 S hi(R) g yp suitable for this chi(k) data ? k range from .. to dk = window type =.. 26. Save chi(R) > File > Save chi(R) 27. Save your entire project End of Athena Tutorial ! () Synchrotron Light Research Institute (Public Organization) 13 5. Fitting the EXAFS spectrum by Artemis EXERCISE 2 Fi i h NiNO3 d EXAFS 2 R A t i EXERCISE 2 : Fitting the NiNO3, merged EXAFS spectrum 1. Search information on the atomic arrangement for Ni(NO3)2.6H2O that will help to model this molecule. 2. Run Artemis > File > open files - select NiNO3 EXAFS. prj, hightlight on NiNO3 merged EXAFS, click on ( ) p - Ni is surrounded by six oxygen atoms in octahedral geometry. We will start with this simple model. g y p +y +z R Ni +x R =2 Ni O () Synchrotron Light Research Institute (Public Organization) 14 5. Fitting the EXAFS spectrum by Artemis 2 C t t th t i d l > f ff i 2. Construct the atomic model > Theory > Quick first shell theory - set Absorbing atom to Ni Edge to K Scattering atom to O > feff.inp - see that Artemis creates one atom of Ni and six atoms of oxygen g Distance to 2.0 A Coordination to octahedral - click do it +y +z R +x R () Synchrotron Light Research Institute (Public Organization) 15 +x 5. Fitting the EXAFS spectrum by Artemis 2 P iti Ni d t f f ff i - see that Feff generates five scatteringpaths 2. Position Ni and oxygen atoms from feff.inp - see that Feff generates five scattering paths 4. Lets fit the data with path#1 - click on path 1 +z +y 3. Click Run Feff , then > interpretation +x - Draw scattering route of photoelectron wave leaving from the Ni atom and coming back to the Ni atom +y +z Path #1 +x -How many routes are equivalent to this one? () Synchrotron Light Research Institute (Public Organization) 16 Degeneracy (degen) =Number of equivalent routes 5. Fitting the EXAFS spectrum by Artemis - Set which parameters should be varied or fixed. > Guess, Def, Set - What is Reff? Effitive view of multiple scattering R 2 R 3 R =(R 1 +R 2 +R 3 )/2 scattering is R 1 4. Define parameters of path#1 - Edit the 1 st row to rename amp as S02 Set S02 =1 andGuess it (to be varied) - Set S02 =1 and Guess it (to be varied) - EXAFS equation of path#1 is - Set enot =0 and Guess it - Set delr =0 and Guess it - Set ss =0.003 and Guess it - By default, Artemis defines amp =S0^2 and use N =1 - Rename amp to S02 and set N =6 - R1 =Reff +delr -Note that enot =E0 is the energy different between initial 1s state and the final p state, i.e. the edge energy . -At E =E0, k =0 because () Synchrotron Light Research Institute (Public Organization) 17 5. Fitting the EXAFS spectrum by Artemis 6 Define the fitting range to fit the data with path#1 6. Define the fitting range to fit the data with path#1 - Click on NiNO3 merged EXAFS, set R-range =1 to 3 5. Plot chi(R) from path#1 - Highlight the data and path 1 -try setting S02 =0.7, What different dose it make? () Synchrotron Light Research Institute (Public Organization) 18 5. Fitting the EXAFS spectrum by Artemis -To see fitting result, >highlight fit 1 and click on parameter report - Keep in mind that number of variables (parameters that you set them to Guess must be lower than independent points. -Are the values of variables determined by Artemis y reasonable? () Synchrotron Light Research Institute (Public Organization) 19 5. Fitting the EXAFS spectrum by Artemis -Always check how good the fit is to the real part and imaginary part of _(R) -Always check how good the fit is to the original _(k) - What do you think ? () Synchrotron Light Research Institute (Public Organization) 20 5. Fitting the EXAFS spectrum by Artemis fi h d i h h#1 d h#2 7. Lets fit the data with path#1 and path#2 -Highlight Ni-O, then path #2, right click on it, add and jump to feff0002.dat +y +z Path #2 +x -Study the atom coordinates and draw scattering route from the Ni atomto oxygen atoms and on the way back to the Ni - Can you identify all of 24 equivalent routes? -How many routes are equivalent to this one? the Ni atom to oxygen atoms and on the way back.to the Ni Degeneracy (degen) =Number of equivalent routes - What is Reff of path #2 ? () Synchrotron Light Research Institute (Public Organization) 21 5. Fitting the EXAFS spectrum by Artemis 8. Define parameters of path#2 - ss2 is different from ss1 because each oxygen atom contribute mean square variation of path lenght. - Right click on ss2 >make ss2 a Guess and jump This is a short cut to Guess, Def, Set window Edit ss2 0 003 then hit Enter - S02 is the scaling factor, it should be equal for all paths generated from the same atomic model. However, if there are more than one structure in the sample, S02 can be different. -Edit ss2 =0.003, then hit Enter -Plot data, path#1 and path#2 - E0 is equal for all paths by the same reason as above. - Why set delR =delr*1.707? () Synchrotron Light Research Institute (Public Organization) 22 5. Fitting the EXAFS spectrum by Artemis L t t ib ti f th#1 d#2 i th k -Extend R range for fitting, from R =1 to 4 - Click on Fit - See fitting results and plot contribution from each path -Lets see contribution of path#1 and #2 in the k space. -Because ss2 is very large, the sine wave damps quickly in k, as e -2k^2*ss2 () Synchrotron Light Research Institute (Public Organization) 23 5. Fitting the EXAFS spectrum by Artemis 9 fi h d i h h#1 #2 d #3 C id tif ll f 6 i l t t ? 9. Lets fit the data with path#1, #2 and #3 -Highlight Ni-O, then path #3, right click on it, add and jump to feff0003.dat -Study the atom coordinates and draw scattering route. - Can you identify all of 6 equivalent routes? Wh t i R ff f th #3 ? - What is Reff of path #3 ? 8. Define parameters of path#3 +y +z Path #3 +x - make ss3 a Guess and jump, edit ss3 =0.003, click Enter. () Synchrotron Light Research Institute (Public Organization) 24 5. Fitting the EXAFS spectrum by Artemis Pl d h#1 #2 d #3 Pl t d t fit d h th - Plot data, path#1, #2 and #3. - Plot data, fit, and each path - Click Fit - See fitting results () Synchrotron Light Research Institute (Public Organization) 25 5. Fitting the EXAFS spectrum by Artemis 10 C f th i th fit? Wh t i i 10. Can we further improve the fit? What are we missing in the model? 12. Draw all 12 hydrogen atoms in the model + +z +y 11. Lets add hydrogen atoms ( from water molecules) in the model Water molecule : O-H bond =0.957 A +x u =104.5 deg x1 =0.957cos(u/2)=0.586 y1 =0.957sin(u/2) =0.757 () Synchrotron Light Research Institute (Public Organization) 26 y ( ) 5. Fitting the EXAFS spectrum by Artemis 13 Add H atoms in the model 14 R n Feff and see the scattering paths generated 13. Add H atoms in the model - Click on feff.inp , Edit it to add 12 hydrogen atoms in the model. 14. Run Feff and see the scattering paths generated. - Note that paths #1,#5,#7,#8 and #9 are what we had before (see p. 16), but now Feff makes new paths #2, #3, #4 and #6 from scattering with the hydrogen atoms !. Th i l tt i Ni H ( th #2) d lti l The single scattering Ni-H (path #2) and multiple scattering Ni-O-H (path #3) are comparably strong as the multiple scattering from Ni-O-O (#7& #5) () Synchrotron Light Research Institute (Public Organization) 27 5. Fitting the EXAFS spectrum by Artemis 15 To start the new fit we must delete paths 1 3 in the 15. To start the new fit, we must delete paths 1-3 in the last fit. 17 I l d Ni H th #2 d d fi it t 17. Include Ni-H path #2 and define its parameter as follows: 16. Include oxygen paths #1, #5 and #7 and define their parameters the same way as you did before. () Synchrotron Light Research Institute (Public Organization) 28 5. Fitting the EXAFS spectrum by Artemis Cli k Fi b d l k h fi i l if - Click on Fit button and look at the fitting results, see if they have sensible values. - Note that R factor is much smaller, so the fit including Ni-H is better. -Its recommended to check the fit by plotting the imaginary part in R space and k space. () Synchrotron Light Research Institute (Public Organization) 29 5. Fitting the EXAFS spectrum by Artemis 19. Add Ni-O-H scattering, path #3, define its paramaters and fit. From the atomic coordinates in Feff.inp, the program calculated Reff of the path #3 from (R1+R2+R3)/2 where R1 =2.0 A, R2 = 2 695A R3 =0 957A Th fit l k b tt d 2.695 A, R3 =0.957 A. Lets allow the program to vary the position of hydrogen atom freely, so we add a variable (a) in R3 and define delR of the path #3 as (delr+delrH+a)/2 The fit looks better and R factor is smaller ! R1+AR () Synchrotron Light Research Institute (Public Organization) 30 5. Fitting the EXAFS spectrum by Artemis S 19. If you think you this is the best fit that you can get, then make a table to report the fitting parameters. -Calculate R from Reff+delR for each path, e.g. Reff of Path #1 (Ni-O) =2.0 A and delR =0.0413 A from the fit., So R =2.0413 A with Recommended XAS websites for more tutorials by XASFS ld id iti Coordination N R (A) o 2 uncertainty from that of delR (0.00064 A) XASFS worldwide communities (www.xafs.org) Thailand XAS Society Ni-O 6 2.041+/-0.006 0.007+/-0.001 Ni-H 12 2.903+/-0.093 0.007+/-0.012 Thailand XAS Society (www. slri.or.th) - Additionally, we can also calculate the bond length of O-H and the angle of O-H bond from the obtained parameters of path#3 (Ni-O-H). -Bond length of O-H =0.957+a =0.957+0.143 =1.10 A - Angle between two O-H bonds =47.2x2 =94.4 deg. Contact me: wantana@slri.or.th Office phone: 044-217-040 2.041 47.2 O Office phone: 044-217-040 Cell phone: 087-239-0026 Synchrotron Light Research Insititute P.O. Box 93, Nakhorn Ratchasima 30000, Thailand End of Artemis Tutorial ! () Synchrotron Light Research Institute (Public Organization) 31