Uncertainty of Peak Area
Variance of the area is computed by a standard error propagation method. This method is well known, and provides the overall uncertainty of arithmetic expressions, which in turn depend on other uncertain values themselves.
The calculation incorporates
the uncertainty of the fitted parameters, and
the dependence of the peak area on the individual uncertain parameters
where
| var( Area )
| variance of peak area
|
|
| partial derivative of the peak area by fitted parameter pk
|
According to our Monte-Carlo simulations, this approximates the real peak area uncertainties very well.
You are able to check this calculation using
the covariance matrix of the fitted parameters, which may be displayed in HyperLab for each region by the “File / General info” menu in HyperLab’s Peak evaluator module,
and calculating the partial derivatives of the Peak Area using the following expressions:
| Fitted parameter (pk)
|
|
| GAmpl
| Gsig ( 1 + GSig ( LSAmpl LSSlope + RSAmpl RSSlope ) ) √π
|
| GWidth
| GAmpl ( 1 + 2 GSig ( LSAmpl LSSlope + RSAmpl RSSlope ) ) √π
|
| LSAmpl
| GAmpl GSig² LSSlope √π
|
| LSSlope
| GAmpl GSig² LSAmpl √π
|
| RSAmpl
| GAmpl GSig² RSSlope √π
|
| RSSlope
| GAmpl GSig² RSAmpl √π
|