Chromatography Forum

I have a ghost peak that will not go away on my Agilent HPLC. Two Agilent HPLCs, actually. My boss (and the Agilent rep that came out and "helped" fix the problem) mentioned pacifying the system using bases and acids and plenty of water. Does anyone have any experience doing this? What, exactly, do you use? I think they told me 1N NaOH, then water, then 1N nitric, then water, then mobile phase. How much of each should I use?

Any suggestions would be very much appreciated.

Thanks in advance.

-GK

Try a TASER, electric cattle prod, or shotgun. I''ve spent the last 25 years trying to pacify HPLCs, and I think herding cats is easier.

If you want to passivate the system, Agilent offers guidelines at their www site. carefully note the warning about removing the column, detector cell and any fragile modules from other suppliers

http://www.chem.agilent.com/Library/Sup ... a25883.pdf

There also have been extensive discussions on passivation on this site, which adiligent search should reveal.

Please keep having fun,

Bruce Hamilton

Whatever else you do, if you go ahead and follow one of these procedures, don't let any nitric acid waste, even dilute, get mixed with any alcohol (isopropanol) waste. The mixture is liable to explode without warning, possibly hours later.

There are still procedures floating around that fail to mention this fact.

lmh,

while I do agree to be safe and not to mix nitric acid from a passivation procedure with organic solvents I have yet to see some evidence that dilute nitric acid in the presence of rather large quantities of water does form explosive esters (nitrates) with alcohols. Do you have a reference by any chance?

I have done a lot of successful passivations on a variety of LC systems. As Bruce suggested make sure which modules and parts can handle the nitric and which are off limits and have to be taken out of the flow path.

Bretherick's handbook of reactive chemical hazards suggests that solutions of nitric acid in excess of 5% should not come in contact with alcohols (6th edition). Please don't take this as an indication that solutions less than 5% will definitely be safe; I am not a chemist and not qualified to determine that they are OK. Anyone contemplating any procedure like this should get their own risk assessment.

My data are more anecdotal, but very real. An institution close to that where I work had an incident some years ago where an instrument was passivised this way; the procedure used isopropanol as a wash solvent, but failed to indicate that the isopropanol and nitric acid wastes should be collected in separate bottles. The resulting (greatly delayed) explosion caused serious injury to someone who happened to be in the room, as well as extensive damage to the room itself. Since an accident of this size would have to be reported, I can probably find out whether it ever became something in the public domain, if that would help?

The problem with getting information about nitric acid is that "small" accidents at the level of an hplc system hardly register on the history of massive accidents that it's caused over the years.

The other scary feature is that many nitric acid + alcohol incidents seem to happen with long and unpredictable delays.

Thanks for the reference.

I can see that mixing concentrated nitric acid and isopropyl alcohol can cause problems...

Have you verified that the ghost peaks are system related
or column related?

Do you work with proteins / PEGylated compounds?

I resolved a ghost peak issue a few years ago on Agilent 1050 by isolating to autosampler rotor valve switching. I removed and flushed well all the tubing with varying solvents and repalced the rotor valve.

The peaks are definitely coming from the system. I have removed the column, and while running 50:50 ACN:water, all I have to do is open and close the purge valve (which has been replaced) and I get two sharp peaks and one broad peak. They seem to be retained somehow, and they come out almost immediately. The two sharp peaks are at about 0.2min and 0.4min, and the broad one from 1-1.5min. They peaks are larger at lower wavelengths (210nm, 225nm), and almost vanish above 255nm.

We are getting these peaks from two separate instruments, both of which have recently ran a buffered mobile phase. That mobile phase is sodium pyrophoshate, phosphoric acid, and DMF (all adjusted to pH 2.5). Its nasty and I hate it, but we are QC and can't change the methods. I always rinse the system before and after with deionized water.

No proteins or PEGylated (?) compounds.

The autosampler rotor valve has been replaced, so has the purge valve, and the column switching rotor valve. The whole system has been rinsed with water, IPA, 90:10 MeOH:water, and even Agilent's rinsing solution (IPA, methylene chloride, cyclohexane, and something else I can't remember right now).

Any suggestions?

-GK

When you mix 1:1 H2O:CH3CN, is that via the solvent mixing valve?.
Have you tried bypassing the injector to see if the peaks appear from the pump - which would be my first suspect.

I'd try seeing if the system peaks were the same size regardless of which solvent channel you used, eg try 100% A, B,C, and D ( if 4 channel ), each with fresh 1:1 H2O:CH3CN, thyen 100% of H2O or CH3CN.

I'd also suspect a residue of material ( eg DMF ) in the on-line degasser - if you have one, or the solvent proportioning/mixing valve, and trying 100% of each channel you might observe some differences.


I'd also fully flush the whole system ( including each set of solvent lines ) with 100% water with 0.1% TFA, water, and then 100% CH3CN. Then back to either water or CH3CN, and see if the peaks still appear.

I'd not worry about passivating the system until I'd eliminated all the solvent inlet system.

Please keep having fun,

Bruce Hamilton

The mobile phase is premixed. I try to always premix my mobile phases to keep from forming bubbles inline. Our Agilents all have degassers.

I was informed earlier today that Agilent is coming back out to fix the problem, again. Although they didn't really fix it the first time...

We don't have any TFA... Anything similar I can use?

I was actually told to not worry about doing anything with the HPLCs, there are other things to do. So, I don't know how much I can really accomplish in the spare time I have. I just hate that I started to troubleshoot it and can't figure it out, very frustrating.

I'm 90% sure that it doesn't matter which line the MP is coming in, the peaks always appear. Also, the immediate response hinted to me it was closer to the detector...

Thanks again for the help.

-GK