Well, you as software for experimental canes or exafs data analysis does: in your experimental example, it seems to me the the first energy with normalized absorption equal to „1“ is at about 20 eV above the edge. So go to your calculated spectrum, and take all energies above 20 eV, built the average, and calculate the conversion factor to unity. With this factor, you then have to multiply the entire spectrum.

You may also import the data in software packages such as Athena, WinXas or others and use those software, originally meant for treating experimental data, for your purposes. Good luck, Dirk

Dirk Luetzenkirchen-Hecht

Thanks very very very much for your answer!

I am confused with your second suggestion. As you can see, I just compute a small range of energy by fdmnes, if I normalize the spectrum with Athena, it's difficult to determine the value of "pre-edge range" and "normalization range" in Athena. And the values of the two parameters will affect the comparison between the calculated spectrum( by fdmnes) and the experimental spectrum. It means that the calculated spectrum still can not fit the experimental spectrum well.

Thank you again for your unselfish answer~

Dear Naichi Zhang ,

you are right, the 'pre-edge range' and 'normalization range' will affect the normalized XANES spectrum a bit. But that is the situation we have always to deal with.

My suggestion:

please take the same 'pre-edge range' as well as the same 'normalization range' for 'pre-processing' your calculated 'fdmnes' spectrum, as you have done it for your experimental spectrum.

If these ranges do not fit, please make them fit;

a) either by re-calculation via fdmnes to be able to choose the same ranges as in the experimental case,

or

b) by repeating the pre-processing of the experimental data, taking the 'pre-edge range' and the 'normalization range' from your fdmnes calculation.

But as it is suggested by Dirk Luetzenkirchen-Hecht , the lower limit of the normalization range should be around (close to) 20eV.

You should avoid the XANES peak minimum around 35eV as lower limit of the 'normalization range'.

Unfortunately there is a bit 'arbitrariness' involved in this procedure, but you have to 'tame' XANES structures somehow, when preparing the normalization.

With respect to the choice of the pre-edge range; please take not a too short range. This range is used to extrapolate the energy dependence of pre-edge absorption coefficient up to the upper limit of the 'normalization range'. Any errors in the extrapolation will affect the preprocessing result, because the extrapolated pre-edge µ(E) will be subtracted from the µ data prior to the normalization.

So, it seems to be better to repeate the fdmnes calculation to have a sufficient pre-edge range.

Additional remark:

the energy positions of your 'white lines'*) (i.e. the first strong peak on top of the edge) do not fit.

The peak of the 'yellow' curve is at about 12eV.

The peak(s) of the blue and red one are at about 14,5eV.

So please check your energy scale(s).

*) please scroll a bit down the page of the following link in order to see why this line is called 'white line':

https://chem.libretexts.org/Bookshelves/Physical_and_Theoretical_Chemistry_Textbook_Maps/Supplemental_Modules_(Physical_and_Theoretical_Chemistry)/Spectroscopy/X-ray_Spectroscopy/XANES_-_Theory

"White Line

In certain XANES spectra, the rising absorption edge might lead to a sharp intense peak referred as “white line”. The reason is that in the past, X-ray absorption spectra were recorded using photographic plates, and the strong absorption of certain wavelength leads to an unexposed band on the photographic plates which after develops in negative, appear as a white vertical stripe, hence the term “white line” ".

Good luck and

best regards

G.M.

Gerhard Martens

Thank for your detailed and patient answer!!

Follow your instructions, I got a relatively good fitting results! Yeah!

As for the position of white line peaks you mentioned, it is another problem that bothers me a lot. As far as I know (I can just understand a few fdmnes parameters...), fdmnes code can not align experimental and computed spectra automatically, we should do it manually.

Could I ask you about if you have ever used FDMNES code?Do you know any useful parameters for alignment? What do you mean by "check your energy scale(s)"?

Best wishes for you and your family~

Naichi Zhang ,

thanks for feed back. You results sound good.

I have not used FDMNES yet.

With 'check your energy scale' I mean to look at the calculation of your energy scale(s) for both the experimental as well as the calculated spectra.

With respect to the experimental spectra the monochromator settings are essential, e.g. the angular alignment as well as the starting angle, and considering the backlash of the gears. Performing an energy calibration just before and/or after the monochromator scans could help to impove the reliability of the energy scale.

You see, there is a lot to check.

But I think, as you already suggesed above, in your case, you should simply manually shift the energy scale to have a matching of the white lines.

By the way; which x-ray energy regime your are dealing with your XANES?

Best regards

G.M.

Gerhard Martens ,

Forgive my inexperience, do you mean energy region by "energy regime"?

The picture I plotted above is just an example in FDMNES manual, it's Fe K-edge for FeO6 molecular.

In the future, I will do some job in the analysis of heavy metal with XAS, and with the help of mechine learning...

It is a huge challenge for me...

So sincerely appreciate your help~~~

✧*｡٩(ˊωˋ*)و✧*｡

Dirk Luetzenkirchen-Hecht

I have solve my question perfectly with your selfless help!

Good luck with your research and your life!

ｸﾞｯ!(๑•̀ㅂ•́)و✧

Dear Naichi Zhang ,

thanks for feed back...

You are right, 'energy regime' is another term for 'energy region'.

Thanks for the details

and best regards

G.M.

Naichi Zhang you are welcome. Good that the given answers have helped you on your way. Best regards, Dirk