Do you mean the UV Impurity check??
If yes - then please have a look at the reference guide in help, this explains the sensitivity value.
And yes curves can be printed and seen in the peak details view - again see help.
For sure I took a look on the help, which is quite good, by the way!
This sw has a new approach on the spectral settings, with basically the wavelength range as the only parameter that would be able to change the Purity value. Is that right?
As I see it, the Sensitivity value would only change the color flag, green/red, for the purity. I can obtain a red even with a purity of > 999, which cause can some confusion. I’d like to have a better explanation on how to setup the sensitivity correctly and what to look for during the peak evaluation.
On the ChemStation there is a plotted Threshold and Similarity curves, but on OLCDS 2.1, there is just the ratio of them, right?
Here some more details from - maybe this helps.
By the way: We are still "tuning" the algorithm - so if you have some good example data where you think we are doing it wrong - please send it to me. To the curve: Correct we only show the ratio curve since it's the most useful one.
I think people coming from ChemStation always expect they can enter a fixed value above which a peak is pure and below is impure. But what is the correct value here 980? 981? 990 ? – nobody can tell you this because it’s highly depended on noise, column, solvent, matrix etc.
That’s why we have chosen a slightly different approach in OpenLAB 2.
- a) We do the calculation fully automatic – similar to the ChemStation automatic mode
- b) This is an check for impurities – not that a peak is pure. You can actually never say if a peak is really pure – you can only detect impurities with significant spectral difference.
- c) We have simplified the set of parameters – talking to ChemStation customers I found out that people use the Purity without really understanding all the parameters.
The main parameter we have is the “Sensitivity”. The way you should use it is as follows (in my opinion):
- 1) You should do a run of your pure standard which contains the compounds of interest. You might even choose to spike a blank with the pure standard compounds – all that depends on the type of samples and your application. This requires the chemists know-how.
- 2) You do an impurity check calculation for that standard. Adapt the sensitivity (globally or per compound) so that the peaks of interest are shown as pure.
- 3) Now you can use this method to analyze your samples – and hopefully detect if there unexpected peaks co-eluting with your compounds of interest.
Users should be aware that UV Impurity checking has it limits and does require that the chemists reviews the results.
I like the automatic approach of OLCDS 2.1. Some considerations:
- The Customer said to me they would like to see the similarity and threshold curve. If possible, that could be an option of the graphical window.
The UV spectra is very broad in general and have a lot of limitations as you mentioned, but the auditors got really freak about these impurities check nowadays. Here they are complaining why the Customer choose a limit (990) different from default. What is not comfortable is having a high purity (> 999) and a red flag.
I need to double check, but I think the traditional DEMODAD and DEMODADNEW should more impurities than ChemStation, all at default.
I just wanted to follow-up to see if the previous replies in this thread answered your question. If so, please come back to the post and click the "Correct Answer" button on the response that helped so it will make the solution more visible. If you still need help, just let us know and we would be happy to continue working with you.
I went ahead and flagged the "Correct Answer" on this post. If you still need help though, please let us know and we would be happy to continue working with you.