Ghost Peak every injection, help !

[dmeul]

Hi,

I'm desperate to measure my samples but have the following problem with ghost peaks:

A large peak occurs after every injection (of water or mobile fase, or any other sample) about 1.5 times the dead time. Its a large peak, sharp front but long tail, the area is rather constant.

Changing the temperature or mobile phase does not really produce a reliable change retention.

Injecting straight into the detector also causes a peak.

tried two detectors, both show the peak.

The spectrum of the peaks shows absorption only at low wave length (up to 240 nm).

Is this a contamination of some sort? What can it be and where?

A have a thermo scientific "Ultimate 3000" system with autosampler en dionex detector.

Thanks for any help!!

[mattmullaney]

Please, more details about the method...isocratic, gradient, ion-pairing...is the Thermo Dionex Ultimate binary or quaternary gradient, does it use a loop or flow-through needle, does the LC system have needle and/or seal wash solutions or both, what solutions are you using for these washes, if present?

Sounds as if you've possible contamination within the LC system somewhere, agreed. The eluent(s) haven't necessarily been eliminated--when the mobile phase is swapped, are you replacing the source(s) of buffers or the buffers entirely (say replace phosphate with acetate, see same peak that doesn't belong)? Have you changed the column out for another column and still see this extra peak?

If you think based on what you've done for troubleshooting isolates the LC system, try flushing the system...see what Thermo Dionex recommends for this. When you flush the system, don't include the column or detector in the flow path (!)...with the Waters Alliance or Agilent 1100/1200 6N nitric acid or 30% phosphoric acid could be used, followed by a system flush with DI water back to neutral pH (this can take overnight). A similar flush for the injector could also work, check with the system manual for a procedure.

See what you think and please write back.

[dmeul]

Hi, many thanks for your reply.

So it's a quaternary system. Its connected to an external reservoir for rear seal wash, this solution is 70% methanol in water. No separate solution for needle wash.

I use 1% MeOH, 10 mM phosphate buffer at ph 3.0 for A, and an 70% MeOH phosphate buffer solution for flushing between samples. I inject extractions from biological (plasma) samples. Proteins have been precipitated from these samples.
I have not tried connecting another column but the column is new. I have also not tried different bufffers, what could this tell me about the origin of the contamination?

So my best option would be to disconnect the column and detectors, and then flush with nitric acid and water according to a cleaning protocol? What else could I try?

Kind regards

[mattmullaney]

Hi Demeul,

The terse answer seems to be yes, it's probably not the new column or the eluents you've been using, signs point to the UHPLC system--you've tried different preparations of the same buffer I take it, hopefully you were able to try different bottles of the same buffer salts and pH adjusting acid or base...I hoped to eliminate the buffered eluent as a source of contamination altogether. Maybe that is Not So, but the early-eluting nature of the apparent contaminant speaks to a different source than the buffer salt and other additives to the eluent.

I don't know much about the Ulitmate 3000--if it uses a stainless steel sampling loop, perhaps the first thing to try then actually would be to clean it or swap it out. I've done some work with proteins...junk can adsorb to the surface inside the loop. See if Thermo Dionex recommends any procedures for cleaning the loop. If it is safe as per Dionex, things like peroxide or more dilute nitric acid than 6N can be brought to bear on the injector, or progressively hitting the injector with solvents stronger in elution strength in RP than water (MeOH, ACN, IPA, THF...)...repeated injections. If you use full-loop injection, then do these injections in that fashion, partial-loop, then use one-half the loop volume.

For this, you will need no column or detector in the flow path also, just a capillary directed to waste, 50:50 MeOH/water as eluent, flow rate set high enough just to allow proper pump behavior. After the injector cleaning, try a test injection...see if that extra peak goes away. No need to rush to passivate the entire system (if Dionex allows for this) at the first blow.

Good Luck...if you need more specifics of what I'd try to clean the loop, write back. I've these children, you see..

[dmeul]

Hi,

Very good suggestions. Ill see if it is the injector loop first. It is stainless steel.

Considering the extraction method that I use on the samples, I have the feeling that the contaminant is probably best removed using apolar solvent. I tried injecting isopropanol repeatedly, that hasnt reduced the peak size of the contaminant at all.

Is it possible to go to something like hexane (via isopropanol)? I will run isopropanol trough the system then.

How many injections will I need per solvent when going to hexane and back?

Thanks !

Best wishes,

Demeul

[mattmullaney]

Short Answer is again Yes, you'll do no harm injecting hexane after IPA.

I'm Really Surpised that IPA did NOTHING. If hexane also does nothing (after going back through IPA first so as not to upset things with the injector) to clean the injector, I'm afraid the next move would be to something like 30% H3PO4 or 6N HNO3 provided that the injector can tolerate this. It's still a common way (the 6N HNO3) to passivate Stainless Steel prior to use in LC systems.

As to a number, a minimum of seven to switch between IPA to hexane and back to IPA prior to 4-6 injections of LC-grade water--pretty much what your samples are in--and then see if the peak is finally gone. Will be a longer procedure if phosphoric or nitric acid are employed...these are generally effective, but rougher and harder to clean up afterward.

Let's see how it goes....

[ACrawshaw]

Hi Demeul,

From my experience a "ghost peak" is more than likely going to have 2 sources
1) The injection system
2) The reagents

Try these couple of suggestions and it may help you nail thinsg down:

1) Run a gradient profile i.e. run the gradient without an injection. If you happen to be running Chromeleon as your CDS, just set the injection Type to 'Blank' within the sequence table.

If you still get a peak at this stage, then you can effectively rule out the injection system as the point of your contamination.


2) Try preparing mobile phases (specifically A) that either a) contain a different phosphate or b) no phosphate at all i.e. Water and run gradient profiles again. In both cases, if you either see a large difference or complete removal of the peak then you can most certainly attribute it to the reagents. If you try this however, you will either need to flush your lines first or switch to alternative lines that have not had the mobile phase through them. In this instance you can then be sure of no cross contamination if it is indeed your reagents.


Some other points to tackle...

You probably want to change your seal wash from 70% MeOH to about 10% MeOH. There is no need for such a high organic proportion as this is being used to flush aqueous soluble matierals such as buffer salts. Such a high percentage can actually have a detrimental effect.

The injection system is a split loop design and as such carry over is substantially reduced. However, an external needle wash can be employed by either using the small bottle on the front of the autosampler or by T-ing off one of the mobile phase lines (usually Line D). The instrument method can then be amended as necessary to include a wash cycle.

[HippyLabRat]

What about the well-plates or sample vials?

[LCbob]

I would also check you have not got a problem with the injection system and are injecting air.

GL

[nick22]

Its connected to an external reservoir for rear seal wash, this solution is 70% methanol in water. No separate solution for needle wash.

Does this mean you use the same solution for both seal and needle wash? If so, you might try a needle wash that is more compatible with your mobile phase. The peak you are seeing could be due to 70% methanol retained in the needle.

I use 1% MeOH, 10 mM phosphate buffer at ph 3.0 for A, and an 70% MeOH phosphate buffer solution for flushing between samples.

Do you allow enough time for the column to re-equilibrate after switching from the 70% MeOH wash back to 1% MeOH? If not, I would expect a baseline issue, not an early eluting peak, but stranger things have happened.