Area under EPR transitions.

[sg_epr]

Hi, I want to make a rough comparison of the area under the EPR profiles of two samples having same paramagnetic centers but in different relative fractions. Is it possible to determine the area under EPR transitions after least square fitting of spectrum by pepper. Thanks

[joscha_nehrkorn]

You don't need fitting or pepper for that. You should double integrate your data after background correction. Then you can read off a value proportional to the intensity. This is model-free and much more reliable. Anyhow, if you insist on using easyspin for this estimation you can do it the hard way and write yourself a fit function which fit both spectra simultaneously and vary the weight of one of the species. The following might be easier: In simulations (NOT in fitting) the absolute value of Sys.weight matters. So you adjust it in both cases by eye and that's it. Again, the proper way is double-integration.

[Matt Krzyaniak]

Just to add to the above, within matlab you can achieve a basic integral through the use of cumsum that is assuming your increment size is fixed across the range you're integrating. Then you can follow this by a sum to provide a number to compare for your double integral.

[sg_epr]

Dear Matt & Joscha. Thanks for your suggestion. In the process suggested by you, my worry is the background correction. How can I do the background correction from my data.

[Matt Krzyaniak]

This depends on the extent of the background correction needed, if the background isn't particularly bad you could use the ES function basecorr or if the background if pretty bad you could define a set of logical indices and remove a line based on that.
So for you complete problem something such as:

Code: Select all

clear
[field,spc] = eprload;

plot(field,spc)

%%
% If the background isn't too bad
% adjust the second number for the order of polynomial removed as
% background
spc_bkg = basecorr(spc,1,2);

plot(field,spc,field,spc_bkg)

%%
% If the background is bad and the above doesn't cut it you can define a
% set of logical indices through which you want to remove the background

% & and
% | or
% read as field less than 3100 or the field greater than 3800 and less than 4000 
ind = field <3100 | field > 3800 & field <4000;

% adjust the last number in polyfit to specify the order of polynomial for
% the baseline.
[p,S,mu] = polyfit(field(ind),spc(ind),4);
bl = polyval(p,field,S,mu);

plot(field,spc,field,bl)
spc_bkg = spc-bl;

%%
% integrate
spc_int = cumsum(spc_bkg);
plot(field,spc_int)

%%
% often you'll want to basline correct the integral too, the baseline should be zero

ind = field <3100 | field > 3800 & field <4000;

% adjust the last number in polyfit to specify the order of polynomial for
% the baseline.
[p,S,mu] = polyfit(field(ind),spc_int(ind),3);
bl = polyval(p,field,S,mu);

plot(field,spc_int,field,bl)
spc_int_bkg = spc_int-bl;
%%
spc_dint = cumsum(spc_int_bkg);
plot(field,spc_dint)
% for visual purposes I did cumsum again in which case the double integral
% value would be:
spc_dint(end)

% alternatively you could use sum
sum(spc_int_bkg)

[sg_epr]

Dear Matt, How to figure out whether further background correction is required before determining the area. I am attaching the screenshot, can you please give any idea. Thanks