Seeking suggestions to fit EPR spectrum of a Fe(III) complex having both high-spin and low-spin components

[smandal]

Hello,
We conducted an Electron Paramagnetic Resonance (EPR) spectroscopy analysis on a Fe(III) complex in frozen methanol at 6K, revealing a combination of high-spin and low-spin components associated with the Fe3+ complex ion. While we satisfactorily modelled the low-spin component (g ~ 2), the high-spin portions (g ~ 4.3) fitting proved challenging, with the intensity of the fitted spectral line not aligning well with the experimental one.
Our simulation utilized the pepper function, incorporating zero-field splitting (D) and E/D parameters to simulate the high-spin components. Additionally, we employed a two-component system to fit the overall spectrum for the iron complex. Attached herewith are the Matlab script and the spectrum fitted thus far. We welcome any suggestions to improve and refine the fitting process.
................................................................
[B,spc] = textread('FeLHH2O.txt','%f %f');
Exp.mwFreq = 9.64; % GHz
Exp.Range = [min(B) max(B)]; % mT
Exp.nPoints = numel(B);
Exp.Temperature = 6; % temperetaure in Kelvin

Opt.nKnots = 10;
Opt.Transition = [1 2]; % level 1 <-> level 2

Sys1.S = 5/2;
Sys1.Nucs = 'Fe';
Sys1.g = [2.05 2.03 2.01];
Sys1.A = [1]; % MHz
Sys1.D = [9650 9650*0.28]; % MHz
Sys1.lwpp = [3 13]; % mT
Sys1.weight = 90; % percentage

Vary1.g = [0.01 0.01 0.01];
Vary1.D = [500 9650*0.05];
Vary1.lwpp = [1 2]; % mT

Sys2.S = 1/2;
Sys2.Nucs = 'Fe';
Sys2.g = [2.13];
Sys2.A = [1]; % MHz
Sys2.D = 0; % MHz
Sys2.lwpp = [2 9]; %mT
Sys2.weight = 10; % percentage

Vary2.g = [0.02];
Vary2.D = 0;
Vary2.lwpp = [1 2]; %mT

plot(B,(spc/max(spc)));

pepper({Sys1,Sys2},Exp,Opt);

esfit(@pepper,spc,{Sys1,Sys2},{Vary1,Vary2},Exp,Opt);

---

[christiansen]

It looks like your fit has at a quite hefty baseline correction. Does the problem persist without baseline, or if the baseline is restricted to a constant?

[smandal]

@christiansen: No baseline corrections were performed on the raw data. Data was used as received from the instrument. In fitting, scaling option was chosen: scale and quad baseline (isq2)

[christiansen]

What I mean is that Easyspin clearly adds a significant baseline correction to the fitted spectrum. The curvature of the green line at the largest B-fields is from the baseline, not from the spectrum.

I think the baseline added to the calculated spectrum may distort the relative intensities of the calculated spectrum. You could try to use "scale and lin. baseline" as an option. You could also manually correct for the baseline in the experimental data and not add a baseline to the calculated spectrum.