Falsly inflated results...

[lnaylor]

I'm running ethanol fermentation samples on a Shimadzu HPLC. Our standards and check standards look fine, the areas are as they should be. The samples however, are greatly inflated. The areas are up to 40% higher than they should be. Does anyone have an idea as to what may be going on here? We're also running the samples on an FTIR. It reads the samples at what we would expect them to be. We can't find a reason for our standards to be fine, but our samples to read high.
Thanks.

[danko]

The first thing I would think of and consequently check in your situation is the injection volumes.
Otherwise, is there any dilution procedure involved in the sample preparation step?

Best Regards

[lnaylor]

No dilution, but we do suspect an injection volume problem. It just never seems to happen to the standards. Odd.

[danko]

Viscosity differences?
Btw, did you check the actual values that were entered for the sample injection volumes?

Best Regards

[tom jupille]

Any chance that something else is coeluting?

[grzesiek]

"Any chance that something else is coeluting?"
- peak purity PDA, MS, rechromatography, did you try any of these?

"The areas are up to 40% higher than they should be" - is it possible that results are just higher?? more of your analyte present?

"we do suspect an injection volume problem" - maybe some error in software, If analysis is fast then injecting your sample few times and standards and quantifying the difference (weighting) can tell if this is the case

[lnaylor]

There's always a chance something is coeluting. It's never happened before in the EtOH peak, but I won't rule it out. We have it in the sugar peaks every now and then.

It's physically impossible for a corn EtOH fermenter to have 27% EtOH in it. Theoretically it tops out around 15%. I know the FTIR is right, and it's saying around 11 to 13%. All other paramenters from the FTIR show the same percentage higher on the HPLC too.

I strongly suspect it's injecting odd amounts of sample. I have no real way to check this. We inject 10ul and I don't have anything that will accurately weigh the difference.

I'm running standards only on it all day today in hopes that I can catch what it's doing on a standard. I think I'm also going to request a new metering pump just on spec. That should fix the problem unless it's a hidden software issue, and the method has been reloaded and the parmeters checked.

[HW Mueller]

Have you done a FTIR of a collected "inflated" peak?

[lnaylor]

The FTIR shows the percentages to be where we would expect them to be and shows no extra "stuff". Our FTIR isn't connected in any way to our HPLC, so the samples are analyzed seperately. Our FTIR data looks as it should.

[grzesiek]

"We inject 10ul and I don't have anything that will accurately weigh the difference." - inject it many times with shortened method

[Bruce Hamilton]

What sort of detection?. Has this problem suddenly appeared or is this a new method for this product?. What are the mobile phase/column conditions. Are you diluting with mobile phase, and filtering or centrifuging?.

My suggestion would be to dilute the samples 1:1 with mobile phase and see if you get 50% area. I would then spike the diluted sample with some standard, and see if the peak size recovery is as expected.

I'd favour co-elution of another compound as most likely cause, followed by viscosity of sample being much higher than standards.

Please keep having fun,

Bruce Hamilton

[Uwe Neue]

Is it possible that somewhere there was a 2x calculation error: 2*13 ~ 27?

[HW Mueller]

Inaylor, what I asked is whether your FTIR still looks as it should after the stuff has gone through the LC.

[lnaylor]

FTIR still looks as it should after the HPLC. No change from running the sample in the FTIR by itself.

[lnaylor]

I guess I should clarify- our HPLC and FTIR are stand alone instruments. They're not connected to each other. The FTIR is a BioFoss optomized for fermentation use. The HPLC is set up stricly for fermentation as well. We use a Phenomenex ROA Organic acid H+ column (300mm). We're running the same filtered sample side by side on both instruments.