Problem with Mn value on GPC

[alambic]

I've build a calibration curve for GPC analysis with 6 narrow standards ranging from 6 120 000 to 1050 as Mp value. My R2 coefficient is 0.999944 and my highest % residual is 4% obtained with Waters Breeze software on a 3rd order fit curve. So I think my curve is pretty good for analysis. To verify my curve, I've injected the 6 120 000 standard as a broad unknown. The results are good for Mw (about 5% deviation) and Mp (about 2% deviation) but are really off for Mn. The deviation is 40-50% below expected value. The supplier standard Mn value should be
5 585 000 but I obtained about 3 000 000 on each run I did. Does somebody have an idea on how to correct that and have Mn value near standard supplier value?

[Rolandas Plausinaitis]

Hi,

I don't have a solution, but want to draw your attention to the fact that half of the peak is in "uncalibrated zone", so your calculation is based on extrapolation, not calibration - this might create a significant bias. Second consideration - what is your column's exclusion limit? Your peak might be distorted if some part of polymer is above exclusion limit.

Good luck.

[ivanvins]

The Mn values are known to be prone to high errors - they are extremely dependent where you place the peak ending point. In your case you are integrating probably too far after the end of the peak.

[alambic]

Rolandas, my exclusion limit is 10 000 000 on my Styragel HT6 so it should be OK with that standard. I've tried recently with Mp = 49 170 standard and got same results. Standard supplier's Mn should be 47 650 and I got about 30 000 so it' still a 36% deviation.

ivanvins, I tried to manually adjust peak ending point in order to correct error due to automatic integration but the Mn didn't vary so much and still far away from what I expected.

I have done before a calibration with my standards processed as broad standards (I wrote Mw, Mp and Mn from supplier's values) but calibration curve was bad even in 1st order fit, my % residual on some value was 30%. I reprocessed my standard as narrow standard with supplier's Mp value only (the curve I use now) so the software calculate for the curve Mn, Mp, Mw, Mz and Mz+1 and get good curve except for Mn value.

[ivanvins]

If you can send me your chromatograms exported as AIA files (uniform sampling) with the standard Mn, Mw and Mp details, I can try to evaluate them in our software Clarity. Also, a picture of the baseline used may help - is the peak end on the chromatogram line or does the baseline continue bellow the chromatogram?

vins@dataapex.com

Ivan Vins

[alambic]

Sorry Ivan, our GPC computer is not on network. All I can do is print chromatogram, scan them and send it to you in PDF format.

[ivanvins]

The picture of chromatogram including the baseline with peak start/end will help.

You do not need to be on network to send files - just copy the exported chromatograms in AIA format (*.cdf files) to floppy or flash memory and attach them to an e-mail send from another computer. (you may need to check Help for the AIA export, in Millennium this was not a straightforward task)


Ivan