Sample degradation in autosampler

[Mattias]

Hi,

What do you think of this:

I opened 32 ampoules of one of our products and placed them directly in the Acquity autosampler (no transferring to LC-vials). The analysis ran overnight and when I got the results it turns out that all samples were oxidised.

In each row of the autosampler, the first vial is the most oxidised. Then the oxidation decreases for every vial until the last position, no 8, in each row, which is again high. This trend is 100% true for all four rows. The ampoules that are most oxidised are placed closest and furthest away from the cooling fan of the autosampler (10°C).

The problem is that the product is not sensitive to air oxidation in any other place than in the autosampler. If I place opened ampoules on the lab bench for weeks, nothing happens. I have tried to bubble 100% oxygen through the solution, but no oxidation. Any idea why the solution decides to oxidise in the autosampler? Each solution is injected only once, so it has nothing to do with the injection needle.

[unmgvar]

I am guessing the ampoules were opened to the exterior without cap closure

do the experiment again while you have the product in a vial and see if it is happening again,

you say the autosampler was cooling and the fan working, at what cooling temp?
I think your problem is the air humidity that could be entering the ampoules if they are left open

[krickos]

Hi

Pure speculation on my behalf of course but a guess at least:

I am guessing that your fan "spits" out or stir up some catalytic stuff that causes your open sample vials to oxidise as the air in the lab obvously is not the cause. There are a bunch of metals that can catalytically cause oxidation pretty fast in some cases, so if tiny amount of metal debris from the rotor/engine is desposed into the vial compartemen...or some kind of contamination in the specific vial compartment/fan.....

The same thing happens on 2 identical instruments?

[Mattias]

Thanks for your replies,

When the solution is transferred to LC-vials, there is usually no problem.

But I am very interested in the phenomenon for another reason. We have analysed this product for several years now (it is a small peptide), and once in a while we get large oxidation peaks. (This is when the product is transferred to LC-vials). This seems to happen in a about 5% of all sample injections. We have never really found the cause for this problem, but it is not related to the individual ampoule.

I have never analysed the solutions directly from open ampoules before, and this is the first time I see oxidation in all injections. I hope that this is a hint of what is going in the LC-vials from time to time!

I has to be something in the moving air that triggers the oxidation. Are there metal particles floating around? Maybe a dust particle is enough? The solution is sterile filtrated so it is particle free to start with. I will start to add some EDTA to the sample and expose it to the autosampler air.

[HW Mueller]

How do you determine oxidation? Is it forming -SS- ? Do you have a standard for that?

[Mattias]

I have reference standards for the two sulfoxides that is being formed.

[Mattias]

new results: I placed open ampoules overnight in front of a table fan on the lab bench - and all samples were oxidised this morning. So it is not the autosamplers fault, it is the moving air.

I will now ask what the cleaning personel is using in our labs..

[lcrandall]

That is really interesting.
I'd also check what types of solvents that neighboring labs are using, as well as whether any kind of building work is going on (painting/sealing floors/etc).

[HW Mueller]

Mattias, did you keep some of the samples closed, as controls, while doing these "moving air" experiments . If you didn´t do this how do you know that you don´t have an oxidation reagent in a solvent of these samples?

[Mattias]

I forgot to tell that all analyses in this investigation have been perfornmed on the same product batch (the batch was rejected due to high level of oxidation products in one of nine ampoules during QC analysis). In QC, the solutions are always transferred to LC-vials.

I have no evidence of individual ampoules containing any oxidants:

1. I have one run where I have opened the ampoules just before the injection (25 ampoules). None of these solutions showed any oxidation.

2. 100% of the ampoules that have been standing opened in moving air (overnight) have been heavliy oxidised.

3. No opened ampoules that have been standing in a drawer have been oxidised.

Yesterday I spiked in the floor detergent that is used here, but that did not cause any oxidation. I added some dust from the floor to one ampoule - no oxidation. I don't know what else to test for....

The strange thing is that this peptide is not sensitive to air otherwise. It is produced and filled without any special atmosphere. There must be something that is transferred to the solutions, but what?

[Don_Hilton]

Could light be an issue? Flourescent tubes give off a good bit of UV. And if lights are left on in the lab over night that would be an additional difference between in the drawer and open to the air over night.

[carls]

To summarize:

"If I place opened ampoules on the lab bench for weeks, nothing happens."

If you repeat this experiment with a fan blowing on the opened ampoules you see considerable oxidation.

"I have tried to bubble 100% oxygen through the solution, but no oxidation."

Not oxygen sensitive.

"We have analysed this product for several years now (it is a small peptide), and once in a while we get large oxidation peaks. (This is when the product is transferred to LC-vials). This seems to happen in a about 5% of all sample injections."

Oxidation normally only happens occasionally in (capped?) autosampler vials.

"I placed open ampoules overnight in front of a table fan on the lab bench - and all samples were oxidised this morning. So it is not the autosamplers fault, it is the moving air."

Seems consistent with first observation. But why would exchanging the air slightly more efficeintly increase oxidation? Is the formulation buffered? If so to what pH, buffer concentration and using what buffer?

" I have one run where I have opened the ampoules just before the injection (25 ampoules). None of these solutions showed any oxidation."

I assume you injected directly from the open ampoule?

"No opened ampoules that have been standing in a drawer have been oxidised."

I assume this was done at the same time as one of the above "moving air" experiments?

Could reduced pH cause the accelerated oxidation? Air contains ~0.04% CO2 which dissolves in aqueous solutions causing the pH to drop in unbuffered/inadequately buffered solutions. The moving air and other vibration/mixing sources will increase the adsorption rate of CO2 from air.

This doesn't necessarily agree with your previous experiment where opened ampoules were left on the bench (without moving/forced air) for several weeks without oxidation. Perhaps you can run a control where opened ampoules are exposed overnight in moving air and others are as close as possible (maybe to the right or left and slightly behind the fan) but not in the moving air stream.

Another experiment to try would be exposing the samples to high purity compressed air such as that used for GC. You could bubble it thru like you did for 100% O2. The idea being to expose the sample to an air source different than your labs. If you dont see oxidation with this gas/air then you definately have something (a reducing agent?) circulating in your lab air.

Perhaps the occasional oxidation in autosampler vials could be remedied by ensuring they are tightly sealed?

[Peter Apps]

Is there an chance that the oxidation is autocatalytic. If it is you will see dramatic differences in oxidation with small changes in conditions. I suspect that Don is on to something with the UV light from fluorescent tubes; say the ampules are brown glass, have UV opaque labels on them, or are in a deep tray - if they are under a light the UV shines directly down onto the solution through the open top. Over to one side and the solution is exposed only to reflected and scattered UV. This might also explain why the vials on the edges of the autosampler tray were more oxidised.

Peter

[HW Mueller]

There is something wrong here. First Mattias forgets to tell us results, without which one could not possibly come to a meaningful conclusion. Then we have no idea on how well these samples were randomized, actually they were not, as he tells us that the experiments were done on rejected samples which had oxidation in one out of nine. He then proceeds to say that no opened ampoules in a drawer had oxidation. He missed any of the "one out of nine". We have no idea whether all of this faulty batch was treated the same until the experiments were done.
My principle doctorate advisor always told his students: If you get a spectacular result check whether you made a mistake before you alarm the world.

[Mattias]

Don Hilton+Peter Apps> We have done light stress (10*ICH) on the product, and the impurities that were (slightly) increased were not the oxidation products. We have clear glass ampoules with a transparent plastic label on the outside, since the product was proven not light sensitive.

Carls> It is correct that we rarely see oxidation when the product is transferred to capped LC-vials. But it happens (as in the rejected batch), and it happens in two different labs (the QC lab is in US, I am working in Denmark). I have not observed any oxidised samples in capped LC-vials on this batch so far, but on several other occasions. I have put not capped LC-vials in the air flow as well, same oxidation occurs (so it is not the ampoules themselves).

The ampoules there were injected directly after opening were not transferred to LC-vials.

I have measured pH on the oxidised solutions but there is no pH change (the solutions also contain an acetate buffer).

It is possible that I have two different phenomena. 1. About 5% of the ampoules do contain some oxidant, that we see in the "sealed LC-vial" case. 2. The moving air oxidation phenomenon that affects all solutions. I just hope that this last phenomenon is the explaination to everything.

HW Mueller>I would appreciate if you stop commenting my posts on Chromforum. I am fed up with your aggressive attitude and everything but constructive comments.