Chromatography Forum

I have Schimadzu GC with millenium software. Basically am running samples for residual heptane,toluene, and dioctyl phthalate. Today started having issues with results from sample runs being about 5%lower than expected. Standards run fine. I have re-run a unknown sample that i have ran before and that also had ran about 5% lower than before. I have changed septum on the injector - it didn't help. Method didn't change either. Column is brand new(ran about 150 injections on it). Any suggestions as to how I could fix this problem?

Next question: what kind of detector? And have you leak checked the column connection in the GC oven?

FID, no leaks in the column- GC is set to give an error message when it happens (has a pressure detector at the end of the column)

When did you last change the inlet liner and have you had a change in gas cylinders for the detector gasses. What pressure is at the back of the GC for each gas? (Has it remained the same as before?)

inlet liner is fairly new (about 160 injections). Gasses were changed but pressure is unchanged. However I had a problem before when one of the detectors wouldn't ignite, and switching gasses off, bleeding the lines, then turning the gasses back on and reigniting helped.

I am not familiar with the Schimadzu instrument, but I suspect the pressure sensor you describe is that in the inlet. (With FID the pressure at the detector end of the column is atmospheric.) And I am familiar with inlets that will sound an alarm if they are not able to maintain pressure. Typically these instruments monitor pressure to maintain flow. If the pressure drops, the electronic flow controller allows more gas to enter the inlet, bringing the pressure back up. This works until the flow controller is wide open -- then it sounds an alarm and shuts down. If there is a small leak, peaks become smaller as some of the injected material is lost through the leak. But pressure is maintained and the instrument drives on. I am aware of one manufacturer who has the flow controller do some things that are even fancier - and have seen that not work well because of failures in the flow controller. So, checking on your flow controller (or column): have the retention times for your peaks changed?

I would strongly suggest checking column connections with a leak sniffer (assuming hydrogen or helium carrier gas) to ensure you do not have a small leak - one that would allow about 5% of your sample to escape and the flow coontroller to maintain pressure.

By the way are you running a split or splitless injection?

Also type of sample and injection method. Liquid sample? Gas sample? Type sampler (or manual)?

Its a split injection.
Thanks for the help. Will look for a leak tomorrow.

Its a split injection. Liquid sample and an auto injection.
Thanks for the help. Will look for a leak tomorrow.

Just to be sure that I understand correctly - when you run standards the peak sizes are the same as they usually are ?, but when you run samples under the same conditions the peak sizes are 5% smaller than usual, and the results for amount of analyte come out 5% too low ?

If the standards are running as usual then the instrument is apparently functioning correctly.

Have you made up a new batch of standard ?

Has anything changed in the sample prep - not just in othe procedure but in any of the individual volumetric hardware (pipettes, flasks etc) or balances that you use ?

Peter

Standard is the same as previously. No changes in sample prep.Hardware and balances are functioning properly. However I have noticed that area counts in both samples and std dropped by about 22% but in standard concentration % change is about half a percent but in sample it is about 5%.

I misread the original posting and took everything to be changing. And the latest post indicates that everything drops by about 22%?

Peter is correct that if the area of the standard does not change - the instrument is fine.

So, let me make sure I understand what is going on here: Integrated areas for everthing have dropped by about 22%. The quantified results against that standard have dropped by 5%. I am not sure what you mean by the "standard concentration percent" changing.

How are you quantifying the analytes against the standard? Are you using an internal standard, percent area of the chromatogram, or ...?

Have a look at some linearity data for Shimadzu GC-2010 FID, it is really poor. viewtopic.php?t=13847
We noticed changes in linearity if you put a new jet in the FID. Linearity also depends on FID gasflow settings.