GC-MS / GC-FID - Verification / Validation

Hello everybody,

I operate an Agilent 7890B-GC/5977A-MSD and a 7820A-GC-FID, both with autosampler.

Those systems are yearly maintained by Agilent.

So far, GC-MS is for qualitative analysis (compounds identification through library search for normal injection and HS-SPME) and GC-FID is used for quantitative analysis through a multiple point internal standard method.

We are certified ISO-9001 but not ISO-17025. The analysis performed are for internal use only.

I have been asked to propose a routine check of the systems for QC. Basically to show that they are working fine.

Since I am not sure what kind of tests should be performed for this purpose, I would like to ask your opinion.

At first, it does not have to be too fancy. It will improve with time.

As for GC-MS, we were thinking about providing regular autotune / tune evaluation reports.

For GC-FID, I have no idea if a report can be generated for this purpose.

Thanks a lot for the help.

Parents
  • It sounds like what you need to implement is some type of system suitability test for each instrument.  If you don't have a starting point I recommend looking up the guidelines for system suitability in the USP (US Pharmacopeia).  The basic idea involves injecting a standard multiple times to demonstrate detection sensitivity, resolution, and reproducibility.  Ideally the standard will represent the testing being done by the instrument by both type of compounds and the method parameters.

    A tune is fine and may give an idea of the sensitivity of the detector, but a tune won't let you know the phase of your column is worn out and needs to be replaced, or the inlet liner is contaminated and needs replacing.

Reply
  • It sounds like what you need to implement is some type of system suitability test for each instrument.  If you don't have a starting point I recommend looking up the guidelines for system suitability in the USP (US Pharmacopeia).  The basic idea involves injecting a standard multiple times to demonstrate detection sensitivity, resolution, and reproducibility.  Ideally the standard will represent the testing being done by the instrument by both type of compounds and the method parameters.

    A tune is fine and may give an idea of the sensitivity of the detector, but a tune won't let you know the phase of your column is worn out and needs to be replaced, or the inlet liner is contaminated and needs replacing.

Children
  • Thank you.

    I wish there was a recipe to follow...

    An Agilent document to consult for this purpose.

    Oh well, I guess more reading is needed.

    Cheers!

  • As Wcody suggested , system suitability maybe all that is needed. If the need goes beyond that, Agilent does offer an Operational Qualification that could be performed on a yearly bases. The Qualification comes with documentation that can be reviewed by auditors.

    Regards

    James

  • I have looked into that and I think I got to a solution for the 7820A-GC-FID (using OpenLab CDS Chemstation Ed.)

    I will inject several times the less concentrated of my calibration solutions for my analytical method.

    After editing the Noises Ranges from the System Suitability menu and the Integration Events Table,  I can determine the RSD for each peak's area and I can print a performance+Noise report and get values for: peaks symmetry, resolution and signal/noise.

    The system should be ok if: RSD is ≤ 2%, Peak symmetry is between 0.80 and 1.20, resolution is ≥ 2 and signal/noise ≥ 10.

    It seems those values can be calculated differently in the report. If formulas from USP or ASTM are used. I did not find yet how to define one or another...

    For the GC-MS, and because I don't use quantification methods yet, I would probably inject column performance mixture (or Grob mixture) if I had any. I'll probably make one. Any ideas on what kind of compounds, concentration and solvent?

    Thanks.

Was this helpful?