James_Ball wrote:
Besides, with modern computer integrator software, even the area of tailing peaks is calculated quite accurately.
Was tailing ever an issue for the area integration? As far as I know there are 2 types of integrating algorithms:
1. Determine the boundaries of the peak and use Trapezoid integration or Simpson's Rule. These are probably the most popular choices and they shouldn't have issues with the tailing.
2. Model based: come up with a mathematical model for peak shape (e.g. Gaussian), then determine the parameters for the formula (e.g. amplitude, width, SD) that would minimize the difference between mathematical model and the real data. These work badly if the peak doesn't follow the mathematical model - tailing doesn't allow to accurately describe the peak with a Gaussian model.
My understanding was that the 2nd option isn't very popular. And we don't like tailing/fronting because it's impossible to tell apart a true tail and a well-hidden shoulder.
But maybe those were different times and this Peak Gaussian Factor was introduced simply because model-based approaches were more popular. And the guidelines were written for this type of algorithms. But that's just a guess.
James_Ball wrote:
What really doesn't make sense is that most of the tailing that would interfere with resolution or integration occurs below 5% peak height and that is ignored in all the calculations dealing with peak shapes that evaluate chromatographic performance.
But does it really matter at which height we measure tailing factor? If we measure it at 1%, we'll get a bigger value. But then the acceptable range for this value would be different.