Can' t get rid of the carryover

Dear colleagues,

I do not have a lot of experience with UPLC, but recently started using one for my research project on LEVOFLOXACIN detection.

Acquity HClass equipped with FLR detector.
Column Acquity UPLC BEH C18 1.7mcm 2.1x75 mm
Pre-column: Acquity UPLC BEH C18 1.7mcm Van Guard 2.1x5 mm
Mobile phase: 0.02M KH2PO4 with 0.012 Tetraethylamine pH=4 and ACN 85:15 v/v

After injection of 1 mcl of 5mcg/ml (diluted in mobile phase buffer:ACN 85:15) LEVOFLOXACIN standard solution (highest standard concentration according to the literature method, I've found) there was a peak found in blank mobile phase sample.

Strong wash: 100% methanol
Purge solution: 50% methanol

I have performed a lot of FRESH blank mobile phase injections (around 50). The peak at the retention time of LEVOFLOXACIN is still visible. The peak area changes proportionally to the injection volume (very little at 1 mcl, but visible; up to 1.25 EU high at 20 mcl injection).

Washed needle several times for up to 90 seconds.

Changed column to Acquity UPLC BEH C18 1.7mcm 2.1x50 mm - Peak still visible, but retention time is shorter. I suspect that analyte is somewhere in the system not in the column.

Since LEVOFLOXACIN is solubile in acids, according to my colleagues advice, I have changed the wash buffer, purge solution and mobile phase (prepared in same bottle) to 
Water with H3PO4 pH=3.8 (needed 4, but it was 3.8) and ACN (85:15 v/v) and washed the column and performed 30 injections of the same blank sample (same solution) to clean everything out.

After I have returned to the same settings Line A potassium dihydrogenphosphate buffer Line B ACN (85:15) and performed an injection of blank sample (10 mcl injection volume and 20 mcl injection volume). Still the peak at the retention time of LEVOFLOXACIN is detectable (about 1.2 EU height and 55000 mcV*sec) is visible. And it is clearly differentiable from the baseline.

That is why I do not understand, whether the injection syringe is still contaminated or maybe the column. 
I would be very grateful for some ideas of how to clean it up and avoid this carryover in the future. And I understand, that i should not inject so huge concentrations of standard to such sophisticated column.

I hope the Waters community colleagues could help me resolve this problem.



  • MJS
    So, it seems like you have a lot going on and contamination/carryover sucks.  You seem really worried about your injection volume on the column, but it isn't clear what sort of peak shape you are seeing.  You'd be more likely to see terrible peak broadening or a simple overloading (topping out the scale of the chromatogram) if the injection were too much volume or analyte.

    It would probably be good to start by assessing what the level of the contaminant/carryover is in relation to your LOD/LOQ and accuracy needs.  A small amount of carryover isn't uncommon and the Waters specs exists with some level that is tolerated because you can't guarantee 100% removal during the cleaning cycle.  So, while you see it clearly, is it even of practical significance to your analysis (e.g. 8% of your nominal standard which is problematic or 0.1% which is likely to not be an issue for your assay)?  If it really is sufficiently small, then just stick with the wash cycle you have and move on.  If it really is too much to live with, then start a systematic approach by eliminating the risk of contaminated solvents/vials and work your way through the system.  Below is a start...

    So, start with what the solution prep and cleaning solution composition:
    While a low to mid level pH may provide for great solubility, how does ACN impact that?  I'm not familiar with the molecule, but could it be reducing the overall solubility and therefore hindering the effectiveness of the wash cycle?  You indicate you use mobile phase as a diluent, but could there be a tiny amount of precipitation you just aren't seeing which would then complicate the washing?  Is the problem still there if you make and inject a less concentrated solution (e.g. 5 µL of a 1 µg/mL being equivalent to your 1 µL of 5 µg/mL)?  Your wash solution change from methanol to a low pH solution probably improved things, but could you go a little stronger yet in the wash solution?

    Eliminate solvents/vials:
    What happens if you perform a 0 µL injection after the standard vs the blank?  Just thought I'd ask as that would also eliminate any contamination concern in the blank.  I suppose my thought is that you inject the std, the wash cycle is insufficient, the contamination then gets into that blank solution which is then confounding everything after that.  If a 0 µL injection looks better, clearly that would point to the solution contamination.  If it is worse, you probably need to go back up to the wash solution evaluation.  Also, how consistent is the peak in the blanks with each blank injection? Is it getting smaller or staying about the same?  Can you get a clean blank before the standard is injected?
  • Could also be something as simple as the needle seal is defective or not functioning properly.  I have resolved carry-over issues in the past by replacing and re-characterizing the needle seal (this can be performed in the sample manager diagnostics page)
  • Dear colleagues,

    Thank you for your suggestions. I have tested different options and it appeared to be a problem with the pre column. 

    Without the mentioned precolumn there is no problem with the carryover.

    However, since i am about to analyze biological samples (plasma) i think i will need the precolumn. Otherwise it is easy to contaminate the column.

    My questions are then: can it be that Column Acquity UPLC BEH C18 1.7mcm 2.1x75 mm is not matching the Pre-column Acquity UPLC BEH C18 1.7mcm Van Guard 2.1x5 mm?

    Are there any tricks to put precolumn properly? I have used video on waters website as a guide to attach the precolumn.

    Is precolumn usable after i have removed it from the column? I have put it back, but the peak shape became very ugly (double peaks; very broad)

    Thanks in advance

  • Pflan
    Hi Andrejs,
    I spoke to an expert at Waters and this was their suggestion:

    "The precolumn will not cause carry over.  I suspect that there is a void in the connection due to incorrect assembly.  Unfortunately, once the ferrule is set there is no going back.  They will need to use a new one."

    It seems you are pretty confident it is the pre-column - but this may be another piece of data to try out.

  • Use peek ferrules. Every time you can have the right position IN&OUT of columns.
    Wash externaly the needle. Clean all system without column with 5% acetic acid. With purge closed flow on the system (disconet detector) 4 mL/min doing some reset injection. wash pre and column with 100% methanol some hour e run 100 uL of methanol. wash vial without septum (methanol better acid or water). I do this for agilent HPLC. You can invert the direction of column maybe the head of fhase and filter are dirty.