| T O P I C R E V I E W |
| grassian |
Posted - 08/12/2004 : 5:49:47 PM
Hi all,
I am working with IR spectroscopy data. It's rather complicated: lots of peaks overlaping. I use PFM to fit those and in general I do poorly... Lack of experience.
Now the general problem is the location of my peaks. I determine original location of my peak from 2nd derivative. Then in PFM I show those values. The problem is, that when doing fitting procedure the peak location travels from their original positions and don't resemble values of 2nd derivative. If I try to fix those values before doing fit, I get poor overall fit.
Second annoying problem is that sometimes when peaks are really small they go to negative direction when performing fitting.
Could You guys give me some valuable advices with those to problems?
thank You
Jonas |
| 3 L A T E S T R E P L I E S (Newest First) |
| rlewis |
Posted - 08/13/2004 : 01:21:57 AM
In my experience IR absorbance bands are usually better fit by functions which essentially are a blend of Gaussian and Lorentzian characteristics. The Voigt function for example. Another possibility is using a linear combination of Gaussian and Lorentzian functions ... something like ....
Y =(a*Lorentzian)+((1-a)*Gaussian) ... 0 <=a <=1 |
| grassian |
Posted - 08/12/2004 : 9:35:23 PM
Lorentzian
thank You
Jonas |
| rlewis |
Posted - 08/12/2004 : 8:31:23 PM
What function(s) are you using ? |
