Increasing background, poor standard reproducibility

[walkerd2]

Hi,
Currently we have been analyzing a specific set of PCB congeners in biological samples on our Agilent 5975C/7890A GC-MS. During our method development stage, we have been finding that when we run 1 set of standards, our extracted samples and then re-run a set of standards, the peak areas for the second set of standards (all though they are the exact same batch as the first) increase by as great as 50% along with the baseline, which slowly goes back to the initial value after running a number of standards. Our recoveries for spiked samples are almost always greater then 100%. However when we complete repeat injections for the same standard, peak areas are approximately the same (%RSD <5)
We have tried the following:
Clean inlet, change liner types (shape, w and w/out glass wool), different syringes, re-installing column after trimming about 15 cm, using internal standards, changing injection mode (splitless, pulsed splitless etc).
The MS has been tuning fine and we do not suspect that is where the problem lies.
Also we have ran the same extraction procedure and analyzed by uECD and the results were very reproducible, therefore we don't think it is sample related

GC MS method is: 2 uL pulsed splitless injection inlet liner 275C. Column: 30m DB-5ms 0.25x250. Inital temp of 100C ramp to 180 @25; ramp to 250 at 3; ramp to 280@20 hold 5 mins. Flow is 1 mL/min. MS is EI in SIM mode with gain factor of 5. Source temp of 250, Quad 150

Extraction procedure is:
Vortex with 1:1 acetone:hexane 3x. concentrate to 1mL; elute from 15 mL florisil SPE cartridge, concentration in multivap to ~200 uL.
Final extract is hexane, all standards prepared in hexane (from stocks in isooctane)

Attached is a chromatogram with the 0.5 ppb standard before (Black) and after(Blue) the extract injections
Thank you for your help

[chromatonewbie]

How many solvent blanks do you run in between each standard/sample injection?
Do they contain peaks that match any of those in your samples/standards?

[willnatalie]

What m/z changes do you see under these base line changes? What is the age of you column?

[walkerd2]

I typically run one solvent blank following the samples. My sequence usually is: Standards with blank->Samples->Standards with blank. My blank almost never has analyte peaks, and when it does it is usually a mistake on my part. The only thing that does increase between the first blank and second blank is the areas for interferences, which are identified as different siloxanes. I'm assuming this is due to the septa since these peaks do not occur when I run a method blank.

As far as baseline changes go, there is not much. We are operating in sim mode with only a few ions, so for te baseline the ratio stays the same but the abundances increase. The column itself is about 2.5 years old, but was used infrequently for the first 1.5 years (but still had carrier gas flowing through it)

Thanks again for you help

[chromatonewbie]

Do you have an inert source?

[walkerd2]

Yes our source is an EI 350 inert source

[willnatalie]

Run this same method under a scan range. I bet what you will find is some additional peaks cause from your column. I would suggest running a blank first in your daily sequence and ramp it close to 325 for a brief time; see if that alleviates the issue. If not you might just need to swap out the column.


Let me know.

[nschwartz]

Are you using a deactivated inlet liner? That may make a difference. Also you might try without the glass wool. A cyclosplitter design is a good substitute for the wool.

[walkerd2]

Hi thanks for all the responses-
WillNatalie I will give that a shot, although I've tried baking it out at 300C and it didn't help much. I do get a few interference peaks when running in full scan, I throught it was due to the septa but maybe it is the column. I will try swapping it with another one.

I have tried a number of different liners, from w/ wool and w/out, and I just recently tried one of the agilent ultra inert deactivated liners, all with no difference.

[JI2002]

Hi,
Currently we have been analyzing a specific set of PCB congeners in biological samples on our Agilent 5975C/7890A GC-MS. During our method development stage, we have been finding that when we run 1 set of standards, our extracted samples and then re-run a set of standards, the peak areas for the second set of standards (all though they are the exact same batch as the first) increase by as great as 50% along with the baseline, which slowly goes back to the initial value after running a number of standards. Our recoveries for spiked samples are almost always greater then 100%. However when we complete repeat injections for the same standard, peak areas are approximately the same (%RSD <5)

This is typical of ion enhancement in the ion source caused by matrix interferences. I would bet the peak areas will jump if you rerun the samples and then come down again by running the standards. Don't you use IS for the analyses? If you do, then the responses for IS should increase too and the the quantitation results shouldn't be affected much.

[walkerd2]

Hi all,
Thank you for all of your help. I think I was able to pinpoint the problem and eliminate it. I am currently suspecting that JI2002 was correct, and the peak jumps were due to ion enhancement due to interference from the florisil SPE cartridges we were using. When I run hexane through the cartridges, collect different fractions, concentrate them and run them on the GC-MS that was found to be the primary cause of the noisy background in the initial 15 minutes of the run. I've tried using more preconditioning steps and am experimenting with different cartridge types and manufacturers. We are currently using Grace Extract clean Florisil 2000mg/15 mL cartridges. If anyone has any advice I would appreciate it.
Thank you again