Calculation of Magnetic Field Modulation amplitude.

[frapart]

fitting_rat_170112_2D_modamp.m

maim algorithm

(3.37 KiB) Downloaded 280 times

Dear Stefan and contributors,

We are currently working on triarylmethyl data in L-Band in rats. We wanted to determine the evolution in time of the linewidth, double integral, amplitude of the signal with a maximum accuracy. The data are sometimes pretty bad.
To do so we wrote a small algorithm.
1st step we make a rapid fitt of the data using garlic,
2nd step we make a better fit of the data on a smaller range. In this fit we try to implement the calculation of modulation amplitude.
It use a small function mymy :

function y = mymy(Sys,Exp,Opt);
Expfull=Exp;
Expfull.ModAmp=Sys.ModAmp;
[x,y] = garlic(Sys,Expfull,Opt);


We obtain a value of modulation amplitude, which half of the experimental one, is there an explanation, did we make a mistake ?

Thank you very much for your advise.

[Stefan Stoll]

EasySpin interprets Exp.ModAmp as the peak-to-peak modulation amplitude:

Code: Select all

clear

Sys.g = 2;
Sys.lwpp = 0.05; % mT
Exp.ModAmp = 10; % mT
Exp.mwFreq = 9.3775; % GHz
Exp.Range = [328 342]; % mT

garlic(Sys,Exp); % the spectrum is 10 mT wide

Either you are making a conversion error, or the fit is not constrained enough to recover the experimental modulation amplitude from the data.

[frapart]

Thank you very much Stefan,
So ModAmp is not away to take into account modulation amplitude in the case of an over modulated EPR signal ?
What would be the good solution to separate the linewidth from the modulation amplitude effect ?

To try to solve my problem, I worked on a more simple data set.
An experiment using a DPPH sample, which has been recorded as a function of the amplitude of the magnetic field modulation (from 1e-3 to 1.8 mT).
I rewrote a simpler algorithm to study it.
- It is impossible to simulate the 7 first spectrum (mod Amp value <1mT).
- I may artificially solved that multiplying the mod Amp by 10 so that the value for the second scan will be 1, but then it does not work after the 7th spectrum.
- if I multiple it by 1000 (so that the first Mod Amp value will be 1) it doesn't work).
We have to make the calculation by small lot, if not we have the following error message:
----------------------------------
Error using garlic (line 341)
Exp.ModAmp must be either a single positive number or zero.
Error in garlic (line 126)
Error in mymyb (line 5)
[x,y] = garlic(Sys,Expfull,Opt);
Error in esfit>assess (line 808)
Error in esfit_simplex (line 122)
Error in esfit>runFitting (line 660)
Error in esfit (line 535)
Error in fitting_MA_170113_v1 (line 47)
[bestsysti,bestspci]=esfit('mymyb',specDPPH1,SysDPPHi,VaryDPPHi,ExpDPPHi,[],FitOpt);
-----------------------------------------------------------------------------------------------------------------------
It seems the possible value for modAmp are very restricted.

[Stefan Stoll]

ModAmp is restricted in the sense that it cannot be zero or negative. Also, ModAmp in EasySpin is in mT, whereas Bruker software might store it in other units.

In principle, you know the modulation amplitude from experiment, so there should be no need to least-squares fit it.