Mobile phase peak in reversed phase gradient

[Fdollet]

I hope someone will help me with the following problem or may I say challenge for the most optimistic ones !

Chromatografic parameters:
X-bridge RP18 shield column
PDA 210 nm
Mobile phase A: Water : Acetonitrile 90 : 10
Mobile phase B: Water : Acetonitrile 20 : 80
Linear Gradient: 0 min : A:90 B:10
30 min : A:10 B:90

Issue : peaks appear from 20 min to 30 min


The peaks remain without injecting nothing so they do not come from the blank
They remain after total cleaning of the HPLC with acetonitrile, cleaning of all the material with acetonitrile, use of HPLC grade water from Merck or use of water filtered through an other X-bridge RP18 column.

I try also to extend the equilibraton time and the peaks areas increased but it can come from water or acetonitrile since the initial conditions are 90 water and 10 acetonitrile.
The last suspect is acetonitrile but it is a HPLC grade from Merck. I try other batches and an other supplier: same problem, same peaks.

Desesperated, I change acetonitrile by methanol keeping the same conditions: no peaks. Does it mean that the water is ok ???

Someone can explain this to me because I do not know what to do...
Thank you

[Uwe Neue]

The experiments that you have described say that the peaks are coming from the solvents, and not from injection, sample, etc.

Coming from the solvent CAN mean a lot of things, and it is not necessarily the quality of the purchased solvent. Think about everthing that your solvent(s) are in contact with: lines between solvent reservoir and pump/instrument, filters in the solvent bottles, stirrers etc. Maybe you degas the solvents or filter them, and the culprit could be in the equipment that you use for this... etc.

The lack of peaks with methanol does not necessarily mean that the water is OK. The elution strength of your methanol gradient could have been lower than that of the acetonitrile gradient. Double the final composisiton for methanol compared to acetonitrile. Then you have a better evidence that the water is not the culprit.

[Bruce Hamilton]

Have you looked at the UV spectra of the peaks on the PDA, to confirm that they aren't related to your samples or known reagents ( which could help idendily possible sources of contamination )?.

If Uwes' suggestions above don't locate the source, it may be remotely possible that your instrument solvent degasser or filter, pump filter, etc. have been contaminated from an earlier mobile phase?

Please keep having fun,

Bruce Hamilton.

[bookoon]

I hope you are using gradient grade acetonitrile! Normal HPLC grade may not give a very good baseline, especially at lower wavelengths.

Are you using water which is stored (overnight) to prepare the mobile phase? MilliQ or equivalent quality water are best used fresh.

Please keep us informed about your trials.

[Anthony Crawshaw]

I have been through a very similar situation and investigated absolutely every posibility. Including using the Empore SPE disc filters to try and clean up our water and solvents.

One thing was for sure though, it definitely was not coming from the sample.

However, one thing that might be worth considering is the proportion of MeCN you have going through the system at that time in the gradient. MeCN is already less viscous than MeOH, but add into the equation temperature and pressure and you may find the pump might start to struggle due to the "fluidity" of the solvent.

Just a thought....

[hookeslaw]

just a thought but could these peaks actually be the noise of the baseline? as your gradient changes so can the noise of the baseline, a problem which can be compounded at low wavelengths so appears a lot worse. What is the area of these peaks in relation to your principle peak?

[hookeslaw]

Hi
I re-read your initial query and it suggests that maybe you have late eluters? ( please advise if i am misleading here), the initial programme runs from 17% organic (at T0) to 43% organic at 20 minutes . Could it be that whatever is injected on to the column has a higher affinity for the column initially and only starts to migrate down the column once a threshold is breached? a quick check would probably be a) use a new column and see if peaks are observed ( before any sample is introduced to the column) or b) flush through your existing column for lets say 1 hour at the final condition (10%a 90%b). then observe your profile ( you say that peaks are observed without injection, are they still observed following a flush?)

In my experience, peaks observed without injection suggest only a couple of things

1) air intake
2) late eluters
3) increase in noise (see previous post)

this makes sense because if you think about your baseline, what does it actually represent? for me the baseline means the absorbance of the mobile phase i.e background absorbance, when you have a resovoir of homogenous mobile phase, why should it change? i know that this is gradient but to some extent the principle holds,

anyway i am interested in ressolving this issue

one more thing, when you changed to methanol peaks were not observed, this maybe down to an increase in hydrogen bonding so a 43% methanol phase is not ( to my mind at least) the same as 43% acetonitrile phase , although they are organic, more polarity exists in the methanol phase and therefore the late eluters may not of come of the column.

[AdrianF]

Could you post the chromatograms. Blank and sample.

Sometimes you have to live with these problems and work round them.

[DR]

What kind of LC are you using?

[Fdollet]

What kind of LC are you using?

Hello ! Many thanks for all your replies and sorry for not having answered yet but I was out of office. This problem may have killed me ! lol !

The ghost peak are coming from the water: We have runing gradient (Waters system)100% water to 100% acetonitrile (HPLC grade Merck) for 30 min (210 nm) with a Xbridge RP18 shield at 50 ºC without injecting nothing. We varied the equilibration time with 100% water and the peaks increased with this equilibration time. So ok it was our water

So we adquired a fresh milliQ water (TOC below 3ppb) and performed the same gradient. The peaks decreased and desappeared in the noise for this HPLC method.

But we are developping in parallel a UPLC method (Acquity waters system)for which injection volume is very little and then these peaks are coming back annoying me ! So we performed the famous gradient 100% water to 100% acetonitrile for 10 minutes (UPLC scale down:-)) )and we observed a peak at 7 min and others later without injecting nothing. This peak is only 1 mAu but in UPLC with low injection volume, it is enough to ruin your analytical validation.
The column is a Acquity RP18 shield, 50ºC, 210 nm

So we decided to clean this milliq water with various empore disks thinking that maybe this residual peak is linked to the 3 ppb of organic compounds: C18, SDB, ...no one removed this peak. maybe we did it the wrong way, we followed the instructions of the manufacturer. Do you have any tips ?

Unfortunately I can not show chromatograms...

HELPPPPPPPPPPPP !!!

[AdrianF]

You have shown the peaks were from the water in your first system. You solved it by using fresh Milli-Q water.
Now you have changed systems and the problem has come back; the problem must lie with the system - injector connecting lines etc.

[wanda50]

We have seen peaks from MilliQ water in 2 instances. One is storing MilliQ water in nalgene carboys. The second is having a hose attached to the final filter. In both cases, bacteria can grow. The hose caused some nasty stuff to appear in the LC/MS. the chemist running the LC/MS didn't even think about that fact. As soon as i changed out the hose, the baseline settled down.

Whenever i see nasty baselines, I always think - dirty bottles. We have had problems with diswasher soap, including chemsolv, leaving residue in the bottles. We now rinse all our bottles with copious amounts of MilliQ water and none go to the dishwasher.

The next thing is how old is your mobile phase? Do you make up large quantities and use it for months? ACN will start to degrade after a while. (Wish I could elaborate more on that, but I don't recall everything that our solvent guys tell me.) I also don't know if bacteria will grow in 5% ACN.

We also get trouble when our DI water feed is screwed up.

just some thoughts. FWIW

[DR]

Any modifiers being added to the water?
Sometimes TFA can leach stuff from certain glass used for LC reservoirs (Schott, for example - Pyrex seems better). Sometimes, it can be avoided by thoroughly rinsing your reservoir w/ ~10x TFA/water, then not using that bottle for anything else (except your TFA/water).

[Fdollet]

The acetonitrile is HPLC grade and we use glass reservoir washed with milliq water then acetonitrile then milliq water again.

I contact Waters, supplier of our UPLC acquity, in my country and we perform this simple analysis in their lab, they used their owm equipment, column, merck water and acetonitrile and they got the same result: a peak appears at 7 min when gradient reaches 70% of mobile phase B. This component has a max at 250 nm.
I think that they will perform an MS by fragmentation to determine its structure. So far it remains a mystery....