Chromatography Forum

Hi,

I have a question concerning the precision (moreover the repeatability) of an LC-MS/MS system.

I read that for LC-UV, repeatability of retention time and area should be CV%<0.5%.

What about LC-MS/MS? I think the retention time should have the same repeatability as LC-UV. What about the peak area? If I run for example a clean standard five times on the same day, what %CV is acceptable?

Thanks!

Just a rough number would be nice...

If I inject a clean standard for example 5 times after each other, what should CV% (=standard deviation/mean) be like? Smaller than 0.5%? 1%? 2%? 5%? 10%?

Thanks

was your system qualified? if so maybe you can find some data in system qualification documents

in hplc-uv injection precision is part of system qualification

I found some values for LC-UV, but our LC is coulped to a MS/MS. I couldn't find any data about the precision of an LC-MS/MS system...

i've searched spome documents and i see that mass calibration is done for MS systems

i think it depends on the purpose of the method how much uncertainity is allowable

this article may be good to start
Anal Bioanal Chem (2006) 386: 38–45
Validation and qualification: the fitness for purpose of mass
spectrometry-based analytical methods and analytical systems

Hi,
Googling horwitz equation may help you to find a rough number.
Just be careful and keep in mind that these are only estimations to help you to find the correct order of magnitude to perform real tests.
Hope this helps,
Bulent

As you can tell, since no one will state "a number" for area repeatability, it varies tremendously depending on the compound, the matrix, and the MS ionization method/conditions.

I'll stick my neck out and opine that the repeatability of LC-MS is generally worse than that of LC-UV, which is why you see so much more use of internal standards in LC-MS.

Thanks for the replies!

@ Tom Jupille: I agree with your statement, my simple tests with caffeïne standard shows a repeatability of area of ~0.1% using LC-UV, while a CV% of 1.5-2% is seen using LC-MS/MS. So I suppose this is normal!

Wow! You must be quite an analytical chemist! I feel lucky to get a CV of 1% by UV/RID and 5% by LC/MS. Nicely done!

I think it is less than 10%

From test I've done to check the performance of our lcms system, I obtained the folowing results:

1) RT between 2.79 and 2.81min

2) Peak area %CV=0.48%

this sample was the same "system precision" mix (of course stored @ 4CDeg) used during the instrument qualification performed around 4months before.

As stated before, the instrument specifications indicated by the vendor play a big part in your results.


I forgot to mention that the results are over 10injections.

So in order to do precision on a lcms system, are you looking at the chromatogram (total ion count) or the spectrum? I have tried using reserpine solution to perform a precision test, but the values are way off. I was hoping to combine our signal to noise test using reserpine into a system IQ (non-validated) if we could get good cv.

Is it the instability of reserpine, or perhaps the suspension that makes reserpine a poor choice for precision testing (eluent is 80%ACN, 20%H2O, .05% formic acid)? Any suggestions?

I cannot see any wrong thing in your eluent.
If you use the same solvent for the reserpine you shouldn't have any problem.
Depending on the concentration level your working at, it could be better to prepare fresh solutions avery time.

Regarding the test, you must check the chromatogram (better an extracted ion chrom than a TIC, and more and more better working in MSMS mode instead of full scan).
The LCMS has a chromatographic system so you work with peaks and not with spectra.
Even for the sensitivity test I prefer to check peaks and not simply the S/N in MS (or MSMS) spectra.

To check for the precision, I looked at peak area of the MRM-traces.

Bear in mind that it will depend on the compound, how well it ionizes under the conditions in your MS source, temperature gas flows and mobile phase composition.
You should also understand that what may be achievable running the manufactures spec may not be what you get in your assay and will not approach what is achievable with LC/UV. For a well ionized compound under optimized conditions at say 100x the LOD, i would say reproducibility should be better that 10% on absolute area and less than 5% referenced with a stable isotope internal standard. The latest generation of ion sources may do better than this.
Also for absolute area counts it will depend on maintenance on the source, some deterioration in sensitivity will occur as the source ion optics become contaminated with use, it is not unusual to loose all sensitivity.
HC