Diethylamine by Headspace GC

[muggibar]

Hi All,
I am wondering if anyone has any experience or seen anything like this in their vast years of experience any help would be greatly appreciated.

We have a new residual solvent HS-GC method for the determination of know solvents in an peptide API.

An unknown peak was obtained in comparative testing analysis and this was eventually identified as diethylamine (used in the process) at approximately 150ppm. This has been confirmed with GC-MS and 2 separate HS-GC methods.
The levels of diethylamine have also been determined using IC and the level is much lower at approximately 10ppm.
In trying to determine the true results we have seen that heating the sample for longer periods of time in the headspace oven has led to a large increase in the levels of diethylamine observed in our sample. We have passed this onto our chemists but they are unable to come up with a mechanism that would lead to the formation of diethylamine in the molecule and the impurities.

I am wondering if anyone may have seen anything like this previously.
Is there any experiments that we could try and perform to determine the true result?

Also would I be correct in assuming that for headspace analysis 30minutes equilibration should be more than sufficient for equilibration or could there be other mechanisms in place that would mean we need to give the sample much more time to reach equilibration?

The 3 main methods used are summarized below
IC: Ran on CS-17 column with Methane Sulfonica Acid and sample preppred in 5mMol Naoh
GC-MS: Ran on 624. Headspace oven at 120C for 30mins. Sample prepped in dimethylimidazolidione + 100uL NaOH
GC-FID: Ran on 624. Headspace oven at 80C for 60mins. Sample prepped in DMSO/NaOH.

Any help/input would be greatly appreciated.
Thanks in advance.

[rb6banjo]

The pH of the mobile phase in IC and sample in HSGC will have a huge impact on how much diethylamine you can detect. What is the pH of your mobile phase and your HSGC sample?

[muggibar]

Thanks for the reply I will be back in the lab again tomorrow so I will try and get a definite answer for you.

Very roughly for the HSGC the sample is extremely alkaline approx pH 12 or
The 30mMol Methane Sulfonic Acid has a pH of approximately pH 4.5 with the sample being prepared in 5mMol NaOH

[rb6banjo]

I'd expect that those to extremes are basic enough to keep the free amine for HSGC and acidic enough keep it in the protonated form for IC. It seems that you should be able to get the right answer either way. Assuming that your standards are prepared properly, perhaps there's a coeluter in the GC analysis? How good is the MS hit on the DEA in the sample?

[muggibar]

We have checked the standards thoroughly and we are happy that they are correct in both methods (hopefully ).

Spiking experiments on the GC methods have not accounted for the large differences we have seen. The MS hit is between 70-80% for the GC-MS, not as good as we would like but manual examination of the MS spectra across the peak have not indicated the presence of a co-eluter in any of the work we have done thus far.

Thanks again

[RogerBardsley]

Have you tested spiked samples on the LC? Did the spiked samples confirm to amount of diethylamine that was calculated?

[muggibar]

Hi Roger,
Spiking experiments on the IC showed excellent recovery (e.g. spiking 50ppm to sample gave a results of 60ppm)

We carried out testing leaving samples in the Headspace oven for different equilibration times of 1 hour, 2 hours and, 4 hours. An increase in the level of diethylamine from 100 for the 1 hour to 230ppm for the 4 hours was observed. This clearly indicates that the GC result is caused by some sort of degradation/reaction but proving this definitively is proving rather difficult.

Thanks once again

[krickos]

dimethylimidazolidione (DMI)is a "shitty'" solvent in HS-GC in my experience though limited.

We have seen as it degrade slightly over time, were degradation products sometimes cause reactions with sample (+99% pure drug substance) or cause extra peaks. My best example is that we had a about a 100ppm limit for acrolein based on UPHLC and dervatisation, no problem, but when DMI got "old" it caused extra peak(ie acrolein) in residual solvents analysis way higher than in LC analysis.


As I stay away from this solvent my experience is limited, but we noted that in new bottles of DMI may be OK but it appears that it may degrade over time.
You may by DMI filled under inert gas (HS special quality) and compare different qualities old-open bottles/new ones to pinpoint it more, but it sure looks like DMI is the problem.

/Chris