UPLC Contamination Issue - Wits end

[ConorM]

I have posted before about a contamination issue i am having and need some specfic advice!
last time I posted i was directed to another podt however my investigations into this have resulted in an extremely sore head.

im working on an aquity H-class using an FLR detector

under gradient elution (5-95% over 5mins + reequilibration) on a 1.7um kinetex column at 285/350nm the compound is burprenorphine.

mobile phase: Acn:H2O (0.1% TFA) and working to a nomial calibration line of 0.2ppb-20ppb.

ultimate goal: extraction from plasma

my peak is dwarfed by this interference which starts at approx 2.5mins and increases with the increasing solvent strength.

i have tried everything in my knowledge to remove the interference including a system.wash with 30% phosphoric acid + pH neutralization. this lowered the interference ftom around 500EU to about 50 however my top calibration spike response (in solution is about 20). i have tried running the gradient (with no injection) with bottled lc water, milliq water, different brands of ACN and with/without Tfa but the interference is still there! even flushing in 100% Acn doesnt remove it!

unfortunatly management says chemistry isnt their job so get it sorted. with no managment direction and only 2.5yrs of chromatography experience I feel i am at a loss.

the only thing I can think of is a mixing issue but its just ACN and water!

I have finally achieved the sensitvity required (in solution/unextracted) isocratically but the sensitivity is half of the gradient and extracted samples are resulting in interfereneces of their own, therefore i would prefer gradient.

I have also (with absolutly no success) tried 2 dervitisation techniques

dansyl chloride and hexacyanoferrate) using adapted papers. these used diffefent wavelengths and the interference dissapeared.

With my neck on the line and a resounding NO for the purchase of an MS/MS system i am truely begging for help. not only for my job but for my sleep and sanity!

Hope someone can help an amateur chromatographer at the end of his tether

thanks in advance

CM

[mattmullaney]

Hang in there, ConorM!!

Does the Acquity-H have the Hexane/THF upgrade installed...maybe you can flush the UPLC system, without the detector in the flow path, with a nice strong organic. You've already passivated with 30% phosphoric acid, so maybe this isn't the problem, but we'll keep it in mind anyway.

Burprenorphine fluroresces naturally? Kind of cool...otherwise I'd stick with UV at 210 nm...but that may explain why fluorescent tags may not work...self-quenching. Problem I've seen with AQC-cysteine in the AccQTag method for amino acids.

Okay then...thanks for providing the column...please, if you could, provide dimensions for the column, flow rate, time used to recover from the gradient (conservatively folks use 10 column-volumes...possible to get away with less, but why tempt fate?)...also, what is the solvent composition that your standard is in...we want to make sure that the injection solvent is "weaker or identical in elution strength" to the eluent composition at the beginning of your gradient.

We could try a different Ion Pairing Agent and acidify with phosphoric acid...that would be down the road.

Have you tried to inject a solvent blank under your gradient conditons?

What are you using for wash and purge solvents? Acquity-H is a Flow-Through Needle (FTN), is my recollection correct? The wash solvent can be fairly strong...high in organic content and the purge should be weak...I'd ask Waters to ensure that my recollection is proper on this point, as well. Don't Use the Same Solvent for Both Wash and Purge...this is a Definite No-No.

Waters UPLC FLD manual is quite good...you can flush and/or reverse flush the fluroescence cell with IPA or Methanol.

So then...you've passivated, the Wash and Purge solvents will be attended to if they have to be, and we know how to clean the FL cell.

We can check the solvent you're injecting, oh, almost forgot, retention times of all peaks along with column dimensions and flow, we can ensure you're allowing enough gradient recovery time. We also know we can use IPA to flush the system, if not hexane. Probably wouldn't be any problem to use hot water with a bit of sodium dodecyl sulfate, either, but you passivated already...the plumbing should be clean...must be in one of the other areas.

At least, then, this is a start.

Keep a hold onto your wit...this will be figured out.

matt

[ConorM]

Hi Matt, firstly thank you for the swift reply, much appreciated!

Column dimensions: Kinetex XB-C18 50x2.1mm 1.7um (guards on order). Run time is 6mins 5-95% ACN:H20 (0.1%TFA), 1min requilibration, flow rate is 0.5ml/min. I believe the column volume is 100uL so I think 1min is enough? . Standard is dissolved in initial mobile phase conditions.

Your memory is correct! FTN, purge solvent I have as 5:95 ACN:Water and wash solvent as 100% ACN). The problem I am having is unsual in that I can see the interference by running the gradient through the column without an injection and observing the emission trace. So with injection temporarily ruled out I tried running the gradient with no additives, different sources of LC water and different brands of ACN. All resulting in similar profiles. It has to be said the interference is in no way reproducible over successive gradients. It can get bigger or smaller and move around. It looks like drift with increasing ACN concentration but with defined peaks throughout.


As far as i'm aware there is no hexane/THF upgrade on the system. I will certainly try using different cleaning solutions. What is also interesting is we have another Aquity (bought at the same time), the idea being once methods are validated the studies can be done that system. Both systems brand new at the start of the year but sat as waters left them for a couple of months untill I went on the operators course. The 'study' system therefore hasn't been used so I transferred my column and method to that system a few days ago, the same interference is present but at a much higher level (similar to that of the M.D system prior to phosphoric acid flush). I have tried using a fairly new Waters BEH 1.7um (same dimensions) to see if was down to the column but again the interference was still present. Its probably worth saying all reagents are HPLC grade or better.

I would like to try UV detection (I do have a PDA on the MD system) but experience tells me it would be difficult to detect as low as 0.2ppb I could be wrong? I have some photos of the gradient.

Images;

0.2ppb: LOQ - The BPN peak eluting at ~3.1mins the interference in and around that time are the problem. I must stress again the profile looks nearly identical to the profile seen by running the gradient without and injection.
1.0ppb - Mid spike
20.0ppb - Upper limit of calibration curve.

[mattmullaney]

Hi Conor,

Wow...okay. Pretty crappy, unacceptable chromatography. Let's see into repairing this:

1. Pellicular Stationary Phase, cool...I mean not to say that your re-equilibration time is incorrect, my thought is that maybe it should be increased as follows...maintain 5:95 0.1%TFA in water/ACN post-analytical time for 2.4 minutes, flip the eluent ratio back to 95:5 0.1% TFA in water/ACN for 2.4 minutes, and then all should be ready for the next injection.

This will give a nice long organic flush post-run, ca. 10 column-volumes of strong eluent, and another 10 column-volumes of starting eluent to get you back to the initial conditions for the next injection. I apologize for suggesting to slow your run--just want to ensure that there's lots of flushing between injections. Don't know that this will directly help, but it's easy to try out.

Also...only reason I say 2.4 minutes instead of 2 minutes...I get 120 uL for the column-volume, but this Stationary Phase IS superficially porous, and 100 uL may be correct and I may be incorrect...wouldn't be the first time. If you decided to use two-2 minute flushes as I suggest, my feelings wouldn't be hurt at all. I'll admit freely that I've no clue as to what to use as a void volume marker for FL detection...wonder what Waters would say about that...

Oh, SORRY...not paying close enough attention, just noticed that you are running ACN on the A line and the 0.1% TFA on the B line. This is not a problem, please just reverse in order what I wrote above...and keep the proper idea...Strong ACN flush immediately post-analysis time, then nice starting eluent proportioning after the ACN flush. Ensure that the backpressure and baseline look similar to that at the start of the run after the first trial injection you run after changing things to lengthen the overall run time that has been extended on the back-end.

2. I'm glad to hear that I've not forgotten everything about the Acquity-H...excellent choices for purge and wash solvents. These certainly aren't the trouble. Since the UPLC is a Flow-Through Needle...and we see contamination in injection-less runs (very nice troubleshooting, by the way), whatever we've got here is NOT clearing even the minimal amount of "Injector" with neat ACN.
If you have IPA, please consider running this through as the wash solvent...I may even go as far as to suggest priming both wash and purge with IPA or passivating these lines with 30% phosphoric acid or 10% formic acid first...just to eliminate any possibility that these lines are causing trouble. I know that neither of these lines are supposed to enter the sample flow path...stranger things have happened. This suggestion should be okay, I'd check with Waters particularly for acid solutions for cleaning on the purge line. Problem is, while this acid routine will clean the wash and purge lines, it'll take forever to remove the acid...maybe stick with trying IPA on the wash line.

3. Interesting...the relatively-unused UPLC is filthier than the used one.

4. You're correct, no need to bother with UV unless you want to see what the impurities look like using different detection. There are reasons to try this, I don't think you need to do this...yet.

5. We'll assume you don't have the hexane/THF upgrade installed...different materials than the standard Acquity-H.

6. Excellent troubleshooting in varying the water and the ACN...is it possible to vary the TFA? Even though it's not that concentrated, if the TFA is "bad" it could mess with the system cleanliness for a long while. Also excellent troubleshooting, we know that it's not Phenomenex' columns that are the culprit...you swapped out, and we have the same problem.

7. My fault--what is the injection volume? Doesn't matter at the moment...we know that we get junk either with or without an injection. If there is any kind of crud stuck in the FTN or immediately after it, we want to know. Again, if you have IPA or MeOH, try injecting these neat with your standard gradient...see what happens to the "peak blob" at @ 3.1 min.

Guess that is it for me, for now.

In order:

Try injecting a strong solvent, neat IPA or MeOH, at twice the normal inj. vol. (if you can) of your method. See what happens. If you get a nice-sized blob at 3.1 min or so, try another injection of the IPA or MeOH, same volume as the first time. If there is no immediate improvement--this is an addition--you may want to try injecting the 30% phosphoric acid...goal is to make sure the "injector" is clean. If the Blob lessens, we're on the proper course...though you still may want to do the next step anyway. If this doesn't help, then Do the Next Step.

Change the gradient program to lengthen the time the system sees ACN and a longer recovery to initial condtions.

If the trial immediately above doesn't work, then flush out the wash and purge lines, using IPA or MeOH, or 30% phosphoric acid, whatever Waters deems acceptable.

Just thought of something else...if you filter the standards, have you injected the solvent blank(s) with and without filtration?

If these all fail, I'll have to return with my thinking-cap on...my thanks for the opportunity to help!

Matt

[HPLCaddict]

First of all, as Matt already wrote, nice systematic troubleshooting !

Let me recall two pivotal things:
- You see this blob in a blank run, too, that is without injecting anything?
- The blob decreased significantly after a system flush with 30% phosphoric acid?
That's both correct?

The second point pretty much rules out all the materials used for the chromatography (column, eluents, ...) and points to a contamination of the system itself. Since you had significant success with flushing the system, why not going on with this? Try 30% phosphoric acid again, probably for a longer time. Does the interference still decrease? I don't know what the Acquity can withstand but if it's possible you might also think of the good old passivation/cleaning using nitric acid.

[Peter Apps]

Part of the old folk lore of HPLC was that going suddenly from end conditions to initial conditions after a run would lock muck onto the column, ready for it to drift the baseline on the next run. This was one of those things that made a kind of intuitive sense and sometimes looked as if it made a practical difference. It might be worth trying a post-run gradient from your high organic wash out to your starting conditions over about the same time period as you take to ramp up to high organic during the run.

Peter

[AA]

Are you sure you don't have a water contamination issue?

Your baseline looks a lot like the one in Tom Jupille's "mini tutorial"

viewtopic.php?f=31&t=19085

[ConorM]

Thanks everyone for the suggestions, its great to pick the minds of you guys! Refreshing!

After consulting with my manager we decided on a longer strong acid flush, its flushing right now (both systems) in 30% phosphoric acid, (sinkers removed from the lines, A-D at 25% 0.5ml/min, lines pre-flushed with ACN followed by Water) flushing period expected to be around 20hrs. A thought here is that the sinkers need to be removed as they are S.S., but the FTN is also S.S.? Its not important now systems are flushing anyway, but just a curious thought! I will also flush the purge, wash lines and injection mechanism with IPA followed by MeOH, then probably ACN to equilibration to method. If all that fails I will do the acid wash on the injector and related lines, only because it would take an age to get it back to a neutral pH after. Lucky I am doing a short 2day analysis I can afford the time for flushing and pH neutralization + the weekend!

I've checked with manuals about the nitric acid flush, the detector says specifically not to use nitric acid on the flow cell, not entirely sure on the other components. I'll check with waters if necessary.

I will definitely add the longer flush and equilibration. Unfortunately there is pressure on this from higher above, so I managed to get a nice 99+ R2 calibration line in solution by isocratic elution (30:70 ACN/H20 0.1% TFA) there was no interference, I've done a few extraction attempts and the best (surprisingly) was the liquid liquid nHexane:IPA 99:1 extraction although there was bad matrix peaks and recovery approx 88%, reconstitution of the sample in 100uL has proven difficult even to filter with 0.2um filters. I think I'm rambling, basically i would prefer gradient as the sensitivity is twice as high as isocratic elution, so that's my selling point for now.

I would love to get this sorted, I got a promotion last week to Senior Analyst, so I need to produce some workable method soon. We have been pushing for a mass spec too, its difficult to even find research papers specifically on Buprenorphine using hplc, its all MS/MS. But its not likely, I can hope can't I?

Thanks again for the suggestions (and compliments!), I will keep you updated!

Conor

[mattmullaney]

Hi Conor,

My congratulations on your promotion!

Sinker/filters shouldn't be a chemistry problem (they've likely been passivated with nitric acid already), but the 30% phosphoric acid could be a pain to remove from them, good move to take them out for the flushing.

On the Acquity-H, No Nitric Acid. Funny...holds to high pressures, but not as "tough" as an Alliance in many ways. Can use 10% formic acid, I believe.

Remember...if you cannot filter, centrifugation and decanting can be quite good, too...if you have the proper centrifuge and tubes...agreed also on the advantages of the gradient program for quantitation.

Last Item...http://www.ncbi.nlm.nih.gov/pubmed/ (Learn it Live it Love it! My apology is in order if this site you already use.)

J Anal Toxicol. 2012 Oct;36(8):541-7. doi: 10.1093/jat/bks063. Epub 2012 Jul 24.

Improved detection of opioid use in chronic pain patients through monitoring of opioid glucuronides in urine.

Dickerson JA, Laha TJ, Pagano MB, O'Donnell BR, Hoofnagle AN.


Source

Department of Laboratory Medicine, University of Washington, Seattle, WA 98195, USA.


Abstract


When chronic pain patients are suspected of being non-compliant, their therapy can be withdrawn. Therefore, sensitive and specific confirmatory testing is important for identifying diversion and adherence. This work aimed to develop a novel liquid chromatography tandem mass spectrometry (LC-MS-MS) method to detect 14 opioids and six opioid glucuronide metabolites in urine with minimal sample preparation. Analytes included were morphine, oxymorphone, hydromorphone, oxycodone, hydrocodone, codeine, fentanyl, norfentanyl, 6-monoacetylmorphine, meperidine, normeperidine, propoxyphene, methadone, buprenorphine, morphine-3-glucuronide, morphine-6-glucuronide, oxymorphone glucuronide, hydromorphone glucuronide, codeine-6-glucuronide and norbuprenorphine glucuronide. Samples were processed by centrifugation and diluted in equal volume with a deuterated internal standard containing 14 opioids and four opioid glucuronides. The separation of all compounds was complete in nine minutes. The assay was linear between 10 and 1,000 ng/mL (fentanyl 0.25-25 ng/mL). Intra-assay imprecision (500 ng/mL, fentanyl 12.5 ng/mL) ranged from 1.0 to 8.4% coefficient of variation. Inter-assay precision ranged from 2.9 to 6.0%. Recovery was determined by spiking five patient specimens with opioid and opioid glucuronide standards at 100 ng/mL (fentanyl 2.5 ng/mL). Recoveries ranged from 82 to 107% (median 98.9%). The method correlated with our current quantitative LC-MS-MS assay for opioids, which employs different chromatography. Internal standards were not available for every analyte to critically evaluate for ion suppression. Instead, a novel approach was designed to achieve the most rigorous quality control possible, in which the recovery of each analyte was evaluated in each negative sample.

You will be pleasantly surprised!! Between PubMed and ScienceDirect (Elseveir), there is very little that cannot be found!

[sarakuhl]

So…. What was the answer?? I’ve got almost the exact same problem on H-class UPLC and waters has no idea what is going on. Again, column, mobile phase, everything ruled out.