[Empower3] How to use Peak Width and Detection Threshold

[DavidHPLC]

I'm new at using EMP3 and I can't figure out how to practically use (in terms of what do settings of these parameters mean and how to use them in analysis).

I've read the Help section however it's a bit too technical for me at the moment so I was wondering if someone has time to explain the two parameters (and why/how use it at analysis)?

Thank you.

[DR]

Peakwidth is fairly critical. Set it using your smallest peak of interest, somewhere between the Empower suggested level and about half that value will work. Then set detection threshold to screen out peaks that are too small to be of any consequence. % liftoff and % touchdown will get integration started and stopped at the right points. This all assumes you use Apex tracking (and you should).

[Hollow]

also use the ProcessingMethod Wizard until you're familiar with these parameters.
It will guide you through every step of defining the parameters. Finally have a look at the created processing method to see where those values were added to.

[DavidHPLC]

If I got it right, peakwidth will set the minimum width for peak recognition? As in, peaks wide less than X seconds will be ignored?

[EmpowersBane]

Set peak width to AutoPeakWidth ie leave the global value blank, then enter a start and end time depending on your chromatogram, ensuring only to include the relevant peak areas and nothing like solvent front or return to gradient artefacts.
When this is done, apply your method set and this way Empower 3 automatically calculates the optimum peak width for each injection, works the same for Autothreshold.

[DR]

If I got it right, peakwidth will set the minimum width for peak recognition? As in, peaks wide less than X seconds will be ignored?

Not quite - Though time plays into it, PW is more of a guide to Empower to let it know how wide peaks should be. Threshold is a bigger part of the start/don't start decision (for traditional integration). For Apex track methods, threshold is not quite as critical except to (along with min. height and min. area) screen out peaks that are too small. Apex track uses the second derivative of slice deltas to establish peak presence, so it's literally a sign change to see if a peak exists.

[DavidHPLC]

DR thanks for the explanation.

Can you give any tips of how to practically use Threshold / Peak Width? I don't know how to connect the visual imagine of the chromatograph with the numbers set for Threshold/Peak Width.

Currently I just enter a random number and then go up/down to get the desired peak integrated (or not).

[DR]

Threshold, in conjunction with height and/or area settings will weed out peaks that you deem to be too small. Peak width, along with liftoff and touchdown, will go a long way toward getting your integration right - no turtlenecks, no excessively upward sloping baselines that ignore the tails of peaks...

Until you get the hang of it (a feel for what reasonable settings are for a given method), just keep tweaking one setting at a time. I typically suggest doubling or halving a given setting to see whether it makes any difference, then homing quickly in on a reasonable value. If a couple of changes have no impact, move on to the next setting and circle back later. It's an iterative process.

The wizard is also good for getting you close, if you are mindful of the clues to be had on the status bars around the edges of the dialog boxes. (I think that's where they are).

[sbashkyrtsev]

If I got it right, peakwidth will set the minimum width for peak recognition? As in, peaks wide less than X seconds will be ignored?

Not really. ApexTrack needs to smooth the signal before detecting peaks - otherwise it'd detect a lot of extra peaks (which are noise). Peak Width determines how much smoothing is applied:

The peak width sets the widths of the digital filters, which are used internally to obtain the smoothed, first and second derivative chromatograms. In ApexTrack, the role of the peak width parameter is solely to determine these filter widths.

But when smoothing is excessive it may smooth out real peaks as well. For tall baseline-resolved peaks it won't matter much. But shoulders and round peaks are more subtle variations of the signal - they can easily be smoothed out to the point that they become undistinguishable. So your 2 poorly resolved peaks would look like 1 big peak to ApexTrack.

See a video tutorial on ApexTrack.