2-Propanol Peak

[ksmb]

I am using a Shimadzu 17A GC, for USP OVI's and Residual solvents. When I run IPA standards using method I or Procedure A the calibration curve is linear and the peaks look fine. When I run a product the IPA peak is small approx. 50ppm. The problem is this same sample ran by two outside labs report an IPA peak above 7000ppm. I do not know why the peak does not show up when I run the sample. My machine is consistant and the standards look fine. Does anyone have any ideas what the problem may be?
(samples are in water)

Thanks

[Okkie]

Do you have leakage somewhere?
Does your injector inject the correct volume?

Try standard addition to your sample. It will rule out matrix effects.

Okkie

[ksmb]

I have checked for leaks, and added the standard to the sample and done spike and recovery on 3 different methods and two columns. I get the same result each time. The sample looks the same since I started running it in June. I need to find out how best to check my injection volume, thank you for your ideas. It is a wierd problem. I would love any more ideas. Thanks again.

[chromatographer1]

IF you have done spiked recovery analysis and still get the same result I can only see a very few options concerning the different results outside labs found.

1. The same these two labs received contained different amounts of IPA than the retained sample you are working with.

This can happen if the finished product is not homogeneous. This is not uncommon especially if the product underwent a drying process, and especially if that process was done in drying pans.

If can also happen if a portion is sampled and stored but is opened repeatedly for analysis and the IPA evaporated away during handling.

OR Something you have done in preparing or handling the product causes it to lose IPA before you test it. Review of these procedures should be done.

2. The outside labs are wrong.

Good luck.

Rod

[chromatographer1]

I wrote:

1. The same these two labs received contained different amounts of IPA than the retained sample you are working with.

I meant to write:

1. The sample these two labs.........

I should have my coffee before I post in the morning. Sorry.

Rod

[ntruong]

1. Have you checked the split ratio of the current method?
2. Is the correct FID jet installed? Partial plug of the FID jet could give you the low area count too.