Chromatography Forum

Hi, I have a question about some method development and instrumentation but I don't have much experience in method development. We are a small lab so we have limited resources and instrumentation. The answer might be completely obvious but I wanted assurance before I jumped in.
We are looking to assay zinc phosphide, a rodenticide, in various biological matrices. Our current method involves adding an organic solvent (we use p-xylene) and sulfuric acid which reacts with the analyte and gives off phosphine gas. This takes place in a closed vial. The samples are then tumbled for 2 hours during which the phosphine gas is dissolved into the xylene layer. The xylene layer is then analyzed on a GC with an NPD and the phosphine is detected. We make a calibration curve beforehand with purified zinc phosphide and back calculate the amount of zinc phosphide in the starting matrix.

We have a lot of reproducibility problems with precision. We believe the instrument is operating correctly, and we have checked the homogeneity of the samples, which we don't think is the issue either. The samples sometimes degas when filling the GC vials.

Would it make more sense to just do this all by static headspace analysis? Skip the xylene, let the reaction occur in the headspace vial, and inject the gas? Phosphine is not water soluble so it would all enter the headspace. If so, what's an older model of HSA to use? We've found Varian Genesis and Tekmar 7000 units for sale.. are these identical systems? How about the HP 7694? Any experiences? We had issues at an older lab with the 7694 not dropping the vials into the oven correctly.
Thanks for your help.

Several years ago I did exactly what you suggest on aluminium phosphide tablets, using an NPD detector. Details elude me I am afraid, but I recall that the main problem was the need to use a very high split ratio to get the peaks within the detector's linear range. Definitely worth a try, good luck.

Peter

Perhaps some sort of internal standard would help with the precision?

We have a lot of reproducibility problems with precision. We believe the instrument is operating correctly, and we have checked the homogeneity of the samples, which we don't think is the issue either. The samples sometimes degas when filling the GC vials.


Thanks for your help.

As an alternative to going with headspace - try increasing the amount of xylene to reduce the outgassing.

A problem that you might find with an internal standard is that technical grade phosphides produce a whole range of impurities in the gas phase that give extra peaks that might interfere with the IS peak.

Peter

You can also consider SPME.

A quick update -

We added an internal standard to the xylene which showed us that there was indeed sample degassing occuring when we would handle or fill GC vials (the ISTD would remain steady while the phosphine peak area would drop.)

There was also instrument drift that was unaccounted for. I had been turning off the NPD at night and cooling down the heated zones to 110-150C. We're keeping it online at night for the time being to avoid this.

This screams to me that headspace with the NPD is the best option but we don't want to spend the money until all other possibilities are exhausted.

Is there another solvent that you can use instead of xylene to reduce the outgassing ?

Is it bait formulations that you are working on, or residue levels ?. If it is bait (i.e. high levels of phosphide) try reducing the quantity of sample - the outgassing suggests that you are saturating the solvent with too much phosphine.

You can do a quick and dirty experiment on headpsace with a gas-tight syringe - phosphine is a gas at room temp so you will not have serious problems with adsorption to surfaces.

Peter

Correct me if I am wrong, but isn't phosphine pyrophoric? Therefore, air in the headspace of the vial would consume the phosphine? Might also account for the variance as well since the equilibrium between the xylene and the headspace concentration would be changing as the phosphine is consumed?

Best regards,

AICMM

We've found Varian Genesis and Tekmar 7000 units for sale.. are these identical systems?

They are the same machine.

I have found them reliable and durable.

I recommend getting the deactivated sample path component replacement parts with your reactive chemical analysis.

best wishes,

Rod

For what it's worth I did my headspace phosphine analyses on a Varian Genesis. I put in a smaller loop, threaded deactivated fused silica through the transfer line and connected it to the 6-port valve and direct to the column with a press-fit connector (this was in the days before Silcosteel).

Peter

Thank you all for your help.
We got around the problem of reproducibility of results for now, at least, by running in triplicate, pouring instead of pipetting the xylene from the sample prep vial to the GC vial, and using an internal standard to verify injection volumes.
It looks like the volume of samples will increase in the near future, and as we are still using a 5890 with NPD for this (and beads are hard to come by) we will probably be upgrading to a new GC/HSA in the near future.