Changing areas in multiple injection

[cynthia zapatel]

In a series of 15 inyections/lot and 2 or three lots in a continous analisis usually during night, we are used to find
2 or 3 inyections that inform 80 or 120 % the spected value. That invalidate de whole job. We have confirmed that the preparation is ok, for re inyection and new preparation of the sample. We have 5 HPLC and this problem occurs whith any of them but preferible in this numerous inyections.
The equipement is mantain and certified by Shimadzu frequently. And we are being carefully whith degassing.

[haiedc]

In a series of 15 inyections/lot and 2 or three lots in a continous analisis usually during night, we are used to find
2 or 3 inyections that inform 80 or 120 % the spected value. That invalidate de whole job. We have confirmed that the preparation is ok, for re inyection and new preparation of the sample. We have 5 HPLC and this problem occurs whith any of them but preferible in this numerous inyections.
The equipement is mantain and certified by Shimadzu frequently. And we are being carefully whith degassing.

How about the frequency of fault samples, I mean they appear periodically or just random. You can try this by running larger lots (more sample) , say 30 injections and see.

[tom jupille]

Intermittent problems are the most difficult to diagnose!

You have already explored the most probable causes. If this were my problem, I would set up a series of vials containing the same standard and run 2 or 3 replicate injections per vial. That should tell you if the problem is associated with particular vials (misaligned? cored septa?). If it is not (i.e., if only one injection from a given vial has the problem and the others are OK, I would look at the possibility that you are entraining air bubbles in you injector.

Also, look at the k' value of the peak that is having the problem. If the k' is too low (say, below 1), then you are more likely to see random errors due to incomplete re-equilibration of the baseline.

[cynthia zapatel]

Thank you for your advice. The situation appears ramdomly. Never in the first lot (20 first inyections), preferible in the 2nd or 3rd lot.
I will try whith multiple vials and inyection.

[unmgvar]

if the samples are all injected in the same serie of the same sequence then maybe it is related to vial position in the autosampler or "time"
is it occurring for all your applications, can you find a common relation to those applications where it does occur?

are the sample cooled?
is the problem occurring more frequently to samples in place 20 to 40 in the autosampler?
or is it to the 20-40 sample injected located in what ever place in the autosampler?

these differences especially if the samples are cooled can make a difference