Carry over test failure.

<p><span>Hi, </span></p><p></p><p><span>After the PE, one of our UPLC sytems failed the carry over test. We've been cleaning the system for a week, and it stilled failed. The preblank was clean, but the postblank have about 0.02% carry over. We used the same column, MP and sample on the other systems, and those system passed with just one trial. Did anyone ever has the similar issue? Any suggestion?</span></p><p></p><p><span>Thanks,</span></p><p><span>Julia</span></p>


  • Dear Julia,

    First we would suggest to ensure the behavior is not just "Contamination". Sometimes when this occurs do not rule out the column as a source of the problem.

    For and existing application that starts showing exhibiting carryover, consider the following in this order.

    Sample Loop- If a sample loop is reversed or improperly swaged it will cause carryover especially if it has just been installed or reinstalled. Additionally any fitting or component which may have been removed or replaced can contribute to carryover.

    Piercing Needle- The piercing needle can get contaminated or clogged over time. Ensure that the piercing needle is not coming in contact with the sample in the vial or wellplate and that new caps, or capmats are being used. Usually if the piercing needle is clogged it can be seen by looking through it.

    Clean BOS pump line- This is the line that goes from the piercing needle into the back of the instrument and eventually connects to the BOS pump. This line pulls liquid through the piercing needle during the wash procedure. If it gets restricted, then the piercing needle will not get washed sufficiently. You can try cleaning this tube and checking for blockages and connectivity to the piercing needle and BOS pump.

    Sample needle- If the needle has been changed recently make sure that the needle is properly swaged in the injection valve. Also over time the needle can become coated with sample to the point where it can't be cleaned by the wash solvents. Also make sure to not fully sealing mechanisms on the vials or plates and ensure there are no gummy substances present.

    Air sensor tube- If the tube has been changed recently make sure that this tube is properly swaged in the injection valve. Also over time this tube can become coated with sample to the point where it can't be cleaned by the wash solvents.

    Wash Station Block- It is possible that the injection port and injection port o-ring may have a build-up, replace if necessary.

    Injection POD- Overtime the fittings and seals can become worn and cause carryover. Recommend that POD is changed at regular PM times.

  • Hello,

    I'm having carryover problems as well, all of a sudden with a previously run method. Nothing has changed instrument wise in the past few weeks.

    There are good suggestions below, but do you have diagrams or videos on how to do these things? I don't have too much experience with maintenance on the Acquity.

    Also, there are options in Empower 2 for the Acquity: purge inj, wash needle, etc...can you give any details on what these do exactly?

    One more...does the condition column command bypass the needle? I am seeing my carryover even when running the condition column.



  • Hi Lisa

    Empower2 options for Acquity in the sample set :

    Purge inj. will do a prime of the strong wash solvent, then the weak wash solvent and lastly the sample syringe will prime.

    Wash needle will wash the needle with first strong wash solvent then weak wash solvent, with amounts of wash solvent as specified in the instrument method chosen.

    Refresh syringe will prime the sample syringe 3 times.

    These 3 functions are not affected by the runtime you specify but will run for the time it takes (a couple of minutes max.)

    Since the needle is not part of the flow path from the pump trough the injection valve to the column, condition column will not go through the needle. As you carry over is seen with condition column you can rule out the needle and wash station. If you do an injection of you sample and the run consecutive condition columns afterward do you see the carry over getting less and less ?

    Have you recently changed the injection loop ?

    A few chromatograms would be helpful in trying to troubleshoot as well as all your analysis conditions.

    Best regards


    best regards

  • Thank you. It would be helpful if Waters defined those terms in the Empower2 glossery!

    Do you know if there is a way to save Condition Column chromatograms?

    I'm running what you suggested below...

    The needle was changed a few weeks ago by our Waters FSE. I just started seeing the carryover this week. But, I'm using a new column this week that has far better sensitivity. So, it's possible that carryover was there all along and I just couldn't see it using the old, trashed column.

    Chromatorgrams to come



  • Hi Lisa

    The only way to save a chromatogram when running condition column is to do a screen dump from the run samples page. Looking forward to seeing your chromatograms.

    best regards


  • Hello:

    Carryover by definition gets lower when you do subsequent blanks, if it does not, then its a memory peak or system contamination and we have to eliminate it. As Rune said chromatogram is good to have, method conditions sample conditions and concentration etc.


  • Hello,

    Ok, I've attached some chromatograms, in order of running them today.

    The sequence is:



    condition column 1

    condition column 2

    condition column 3


    The first diluent injection has two relatively small peaks in it (but still a problem at that level). The sample injection doesn't seem to increase the contamination in the subsequent condition column functions. But, the amount in the 2nd diluent injection is much greater than the first.

    Do I have a dual carryover/system contamination mechanism occurring?

    I hope this wasn't too hard to follow.

    System is running a gradient from 90:10 to 45:55

    Mobile phase A: pH 3.39 KHPO4 buffer with 0.5% TEA, with 1% IPA

    Mobile phase B: 40:60:0.5 ACN:MeOH:TEA

    WNW 90:10:1 water:MPB:IPA acidified to ~3.5

    SNW 50:50 water:ACN

    Injections are Partial loop with overfill.



  • Hi Lisa

    First of all did you calculate the % carry over from the first condition column run ? I cannot judge how high your two main peaks are and therefore not calculate the % carry over, but it seems pretty low.

    Also it seems as if the retention time of the contaminant peaks varies from the diluent injections to the condition column runs. Have you confirmed that the contaminant peaks really are your two main peaks from the sample (the ones that elute around 12.4 and 12.7 min) ? I see that you are using a PDA so maybe the UV spectra can confirm that the contaminant peaks really are the two main peaks.

    Could it be that the contaminant peaks instead are contaminants coming from the A eluent and just elutes at the same time as your main peaks ? TEA is a know source of contaminant peaks if it is not stored properly. To judge this I need you gradient time. It would also be helpfull to know the injection volume and the loop size

    If it is indeed your main peaks that are the contaminants I suggest the following :

    Since the contaminants are present in the condition runs we can rule out the sample needle, the wash station and the Volume Detection Device (bubble sensor). This leaves us with the loop itself or a memory effect on the column or system contamination as Liz suggests. Did you try to change the column to a new one ? Did you try and make completely fresh eluents ?

    Another thing to try is to run a wash step after the gradient finishes where you go to 100 % B for a short time, since your gradient stops at 55% organic. This would wash both the column and the loop with 100% organic.

    Try and use the partial loop pressure assist injection mode instead to see if this is different. When using this injection mode in combination with the load ahead function and loop off line time, the loop will be washed with both SNW and WNW. Increase the % organic in the SNW to 80%.

    I hope this information will help in isolating the problem and eventually get rid of it.

    Best regards

  • I agree with everything Rune says..just make one change at a time or we shall get lost.

    As Rune says this feels like memory to me, becasue you can see in condition column and I also like the idea that we may need a longer gradient hold.

    Also again using organic IPA is a good choice to rinse all the lines.

    Swapping injection mode is also a useful tool.

    The attached is a big hammer to throw at this, but it is an aggressive cleaning procedure if these steps do not work and we have established that it is indeed memory issue.

    One other tip, in PLNO mode, if you know your sample is very concentrated ppm levels etc, then try to keep the sample from dispersing out of the loop into places it should not get to by keeping the injection below 50% of the loop volume.

    Best regards,


  • Hello,

    Thanks for the suggestions.

    I haven't spent much time trying to resolve the problem recently because we've been focused on other parts of the project.

    But, as per Grace, the column manufacturer, I shouldn't go above 55% organic with this mobile phase combination because of buffer ppt problems. Long story.

    It's not a column memory problem, because it appeared in the first diluent injections on a brand new column.

    The carryover/contamination is actually >1% of my smallest sample concentration, so it's way too high.

    Sorry about the 12min vs 14 min confusion Rune, I accidently posted chromatograms using different gradients.

    I will try the IPA wash.

    Thank you,


  • Lisa,

    Instead of doing a condition column you can do a zero injection. You just do an injection as normal, but put zero for the injection volume. That way you have a way to save the injection data.



  • This is a good suggestion, but be aaware that the needle dips and the washes wash - so the needle, puncture needle, wash station cannot be ruled out for contamination, as per Rune's diagnosis. If you see in condition column, there is either a contamination/memory effect or the peaks are drifting.


  • If you wish to save the column condition chromatogram, then you can run the injection in a different way. In the sample set there is a function called "inject immediate sample". This will run the method (like condition column) without injecting the sample and will save the channel data in the empower project. We use it all the time to monitor issues without having to sit and watch the run samples window during a "condition column" function.

Sign In or Register to comment.