Ghost peak in chromatograph

[fall4p]

Hello

The image below is from HPLC. The "ghost" peak is the broad one at 9.39min
So far I tried changing the mobile phase,the inlet filter, the tubing between cell and column, the column itself, the filter on the purge valve and the AIV cartridge but this ghost peak keeps appearing with various retention times. Any suggestions would be appreciated.

Thank you

[Gerhard Kratz]

And what is your Mobile phase composition, your column, flow rate, wavelength etc.???????
And what do you inject???????
More details are needed.

[fall4p]

Mr Kratz thank you for your response

I was hoping for general directions in altering parameters to make the peak go.
Here's the details you ask

MP: acetonitrile:buffer (ph=2) 50:50
column: RP18, 5µm, 150x4.6
flow rate: 1 ml/min
wavelength: 250nm

I inject amino-nitrothiazole dissolved in acetonitrile at 40°C

[Gerhard Kratz]

OK, thanks a lot for that information, they are very helpful.
If you will inject 100% buffer you will get a negative peak!
What you see is an refractive index effect in your detector cell. You inject 100% ACN and your mobile phase is 50:50 ACN:Buffer.
If you would dilute your sample in mobile phase you will not see this peak.
I had this effect many years ago and I checked everything in the lab, washed all glass flasks twice etc.!

Good luck.

[tom jupille]

Another possibility (especially since the retention time of that wide peak is variable) is that it represents a (very) late eluting compound from a prevoius injection. If Gerhardt's suggestion doesn't work, try flushing the column with a higher %ACN, then inject a sample and monitor the baseline for an hour or so.

[M_Farooq]

If the retention time of this peak is variable, then it cannot be a system peak. It is most likely a highly adsorbed component, which is probably coming out from previous runs as stated above. Do you still see this peak if you inject the pure mobile phase as a sample? Also, see if it is possible to monitor several wavelengths.

[Hollow]

would also think of a late eluter

by using the formula for the plate count (N=16*(tr/w)^2) and assuming a peak width of about 2 min for the peak at 9.3min,
and also estimating the plate count of your column about 6000-10000 plates, the peak would have a real retention time of somewhere between 38-50 min.
So it can come from the samples "n-3" to "n-5" from the acutal one...

so I would go with Tom's sugestion to clean the column with high organic, equilibrate to your method again and inject your sample(s) but record the chromatogramm up to 60min.
Maybe space different samples with solvent blanks if the peak is not seen during the first 60 min.
If there is such a peak, do replicates and see if the retention times are reproducible yet.

[Hollow]

btw: the k of your main compound is quite low, about 0.6.
If revision needs to be done, try to get it k>1 (tr >2.6 min) for more intraction with the stationary phase.

[fall4p]

Thank you all for your suggestions.

The situation is this: I cannot change the parameters of the test to make the peak go, but even so it is still very persistent. If it is an impurity that contaminated the system would it possibly stay on the detector cell? Could it also be a salt? And if so how could it be washed off? I tried hot water, IPA, mobile phase-acetonitrile 10-90, but no success.

Thanks again for your help

[Hollow]

why should something stucked to the detector cell have a retention time and show up as peak...?
-> I would excpect to give you decreased sensitivity and maybe eratic signals throughout your chrom.

So have done the suggested tests, column flush (e.g. 10-15 ml of water, 10-15ml of ACN, 5ml of water, then equilibrate with mobile phase),
then inject a blank (e.g. pure water or pure organic solv) and record it up to 60min, then, if blank doesn't show any peak, inject a sample and record it again up to 60min?

If peak is already present in blank chromatogram, then suggest having a look at your injector and/or chemicals.
If blank is ok, then it seems to be a real impurity/degradation of your sample.

Also try those tests of Gerhard (inj 100% buffer, buffer50/50 and 100% ACN) to see if he is correct.

[fall4p]

Thank you all.

I tried all of your suggestions but the peak will not go away. In the 1 hour tests you suggested the ghost peak appears randomly (not in every injection) but always retained at approximately 45 minutes. In a ten minute test appears after approx. 90 minutes.. So the disturbance is periodic.

Chemicals, solvents all look fine.
Also the detector is not refractive index but VWD with UV.

It is something in the system, but I can't find where. I can't get rid of it.

[Hollow]

In the 1 hour tests you suggested the ghost peak appears randomly (not in every injection) but always retained at approximately 45 minutes.

What do you mean by "not in every injection"). Were those inj from the same vial or different samples?
Did it occur in the blank runs as well?
How precise were these "approx. 45 min"? About same precision as the retention times of your "good" peaks?

In a ten minute test appears after approx. 90 minutes.. So the disturbance is periodic.

What do you mean by this? Please specifiy what you've done with more details.

It is something in the system, but I can't find where. I can't get rid of it.

According to your decriptions on the 60 min runs, I don't think so.

I guess it is a real compound which is present in only some of your actual samples.
Have you done repeated injections from the same vial with the 60min runs?

Maybe you already know in which sample(s) the peak is present and in which it's not. (from the 60min runs)
I would select one of each (present/absent) and do some triplicate injections of each vial, recording up to 60 min, also inject a blank at the beginning and ending
of the sequence.
-> if the peak is always present in only one sample but not in the blank nor the second sample and shows up always at the same retention time
= it's a real substance within your sample(s), so it's not a "ghost" peak comming from the system!

Then you have to stop troubleshooting your system but to find out what that peak is...
If you could switch to a system with a PDA on it, you could check the UV-Spectra of the peak

[Gerhard Kratz]

Can you please bring your mobile phase to the same temperature as your column?
What autosampler system do you have? Syringe or a loop? What injection volume do you inject?

[fall4p]

What do you mean by "not in every injection").

I'll give you an example: I flush with IPA. I do 1hr blank test and there is no peak. I do a sample test, again no peak, then a second test and there you have it. I go on testing a third, a fourth sample and the peak appears at 45 min. Then a fifth sample has the normal profile, thus no peak at 45imn

Were those inj from the same vial or different samples?

Apart from the blank I use the same vial, same sample

Did it occur in the blank runs as well?

Yes, some of the tests have it and some don't

[/quote]How precise were these "approx. 45 min"? About same precision as the retention times of your "good" peaks?[/quote]

Rsd's for normal peaks are 0.04%,-0.10% and for the ghost peak 0.94%

Then you have to stop troubleshooting your system but to find out what that peak is...

Already in the process of geting an LC-MS

Hollow: I will come back with more details of my tests
Regards

[Hollow]

hm, some of your statments don't fit the picture... "Mr. Murphy where are you?"

still I think that it's something that is coming to your column during your injection (because of the stable retention time and the
good correlation of peak width with plate count).
But why not with every injection and also with the blank?

-> I would inspect the injection system (needle, -seal, rheodyne, needle wash etc.) more closely.
What about the septums used? Anything different that in history? (just brainstorming...)


Got a busy week coming, so I prob don't have much time so response could take longer.
but keep us updated about your "quest"