Peak Fit - Voigt - Negative Lorentz Width

Share icon

Hi,

Using MPF to look at some X-ray diffraction data. I am using the Voigt model because the literature suggests that as an appropriate model. The literature does need to be taken with sufficient quantities of NaCl.  That being said, the fits look good from a Chi-square and residual perspective, but the concern is that the Lorentzian component of Voigt fit has a negative width.

Fit completed: 10:50 AM Friday, September 2, 2022
Y data wave: root:xrd_norm
Chi square: 0.32001
Total fitted points: 4001
Multipeak fit version: 3.00
Total Peak Area = 8.0122 +/- 0.17831

Baseline    Type: Linear

    a =     0.0100005     +/-     0.0018435
    b =     -4.49194e-05     +/-     2.3951e-05

Peak 0    Type: Voigt

    Location =     18.5914     +/-     0.086334
    Height =     0.331744     +/-     0.0043902
    Area =     1.83043     +/-     0.04968
    FWHM =     6.09601     +/-     0.094873
    Gauss Width =     7.31523     +/-     0.1953
    Lorentz Width =     -2.44521     +/-     -0.27189

    Fit function parameters
    Location =    18.5914    +/-    0.086334
    Width =    0.227622    +/-    0.006077
    Height =    0.235068    +/-    0.0075406
    Shape =    -0.278292    +/-    0.024403
 

How should I interpret this?

Andy

Which Igor and MPF version you are using? A negative shape value means likely that your fit is not working, since negative shape values are not properly defined (i.e., the Shape should be between 0 and inf and is not symmetric around zero). If you look closely, the Voigt shape will likely look strange, dipping partly into negative values. It is difficult to suggest something without seeing the data or fit result, but I guess you may want to check your background function to improve things. I don't have much experience with XRD data, but I imagine the background can get a bit 'wavy', which may lead to this behavior. Since your Gaussian contribution seems relatively large, you may also try to simply use a Gaussian shape. This depends on what you want to do (do you just want to get positions, area, width?).

September 2, 2022 at 01:58 pm - Permalink

Hi,

I amusing MPF 3.00 on IP 9.01.

I think you are correct about being an indicator of poor fit and something involving the baseline.  I am working on getting a good baseline on the data.

 

Andy

September 6, 2022 at 01:21 pm - Permalink