Peak purity testing with Waters 2996PDA

[dfraser]

Had anyone had difficulties getting the peak purity function to work with 2996 PDA data? I've frequently run into trouble with data from the 2996 where the noise and purity thresholds are not calculated by Empower no matter what I do. Oddly, when I acquire with the older 996 instrument (using a similar method) there isn't a problem. I've tried using different noise intervals, ensured that there are >12 pts across the peak, and using the different threshold choices. I've even used the Processing method wizard to set up the method without success.

Does anyone know why this might happen? Is there something I'm missing?

Thanks.

[Dan]

In the PDA setup screen under the general tab, there is a setting for the digital filter. Ours was always defaulted to a valule of "1" for new methods, but the manual states that the value should be "0" to use the peak purity function. So we always had to remember to reset the value.

Perhaps you can check that parameter.

Regards,
Dan

[dfraser]

Thanks Dan.

I've checked my method, and the digital filter box is off (unchecked). According to the 2996 PDA course notes, this should work. However, I'll play with it and see if it makes a difference.

Regards,
Dave

[bartjoosen]

Did you tried to change the interval for baseline noise?
Probably its a problem with the noise spectrum.

GL

Bart

[DR]

Doublecheck the instrument method used - was it collecting spectra, or just a particular wavelength?

[bartjoosen]

Did you tried to change the interval for baseline noise?
Probably its a problem with the noise spectrum.

GL

Bart

Ooops, please don't read my previous post, I should read more carefully

When you go to review, you start from injections or from channels?
If channels, make sure you take 3D data into review, if injections, start from channels.

Also in review, can you see a 3D plot? If not, there is no 3D data. If so, you normally should be able to see the thresholds and angles.

Is it project related, or method related?

Bart

[dfraser]

Actually, you were right. It definitely seems to be a noise spectrum problem. I just thought to look at the error messages and it is full of "invalid noise spectrum" messages. I have modified the noise interval to longer, shorter, earlier, later, just about anywhere there isn't any other absorbance in the maxplot. I'll do some more investigation into why I've had this trouble in the past, but been successful with the old 996 - less sensitive perhaps?

Thanks for the suggestions.

Dave

[Dan]

Dave,

There have been a number of software revisions from the 996 to the 2996. The basics are all there but enough things that were changed or added can cause some confusion. I didn't use the software all that often (2-4 times/year) and it seems that I relearned it each time to make sure I got the steps correctly (sorry, I don't use the instrument at all now and I am working by memory here, but I can point out problems I remember seeing).

Some things to check (some of this repeats what others said):
- noise interval (looks like you already did this with the maxplot)
- wavelength range
- is the method set to delete the 3D data after extracting the 2D data?
- is the method set for PDA not just LC and does the method have the purity box checked (i.e. is purity enabled)?


Regards,
Dan

[bartjoosen]

Is it a strong absorbing mobile phase?
Sometimes it happens when you use acetate in combination with low wavelengths.

Bart

[dfraser]

My original wavelength range was 210-750nm, as the compounds I'm looking at absorb in the Vis range. I've discovered that if I acquire data from 210-650, the peak purity is calculated. I think this has something to do with the 656nm interpolation, as there seems to be a discontinuity in the spectrum at that wavelength. Selecting/deselecting the "interpolate 656nm" control makes no difference to the spectrum or peak purity calculation.

I've contacted Waters to see if they have any info, ie possible software bugs, etc.