GC/MS Large volume injection

[James_Ball]

Peter,

I am using a cone point syringe that is compatible with a Merlin Microseal.

Ralph,

Currently I am injecting 10ul, the paper I was referencing was using 12.5 because they were using a 25ul syringe and the Agilent autosamplers are limited to 1/2 the syringe volume as maximum injection volume. I have a 50ul syringe so I can inject up to 25ul, but since we are reducing our extraction volume from 1000ml to 100ml I am injecting 10ul instead of 1ul to keep things equal.

If you can talk to Ray I would most appreciate it. With all the concurrent projects I have going now I barely have time to post a few things here. But I do like to throw ideas out here as there are plenty of knowledgeable brains here to help digest all my crazy ideas

[GOM]

Hi James

That is starting to make sense

No worries - I haven't had a reply yet from Ray so I will try to contact him via Linkedin and private email

Please keep up the crazy ideas - a former manager once said to to me "Ralph, you have lots of stupid ideas a year but luckily a couple of them are actually good"

Regards

Ralph

[James_Ball]

My split peak problems turned out to be a bad column. Got a new one in and now I have great peaks all the way from the first NDMA and Pyridine to the late PAHs. I was getting losses of the labile compounds when using wool in a 4mm liner at 250c and 260c injection port temps, so I lowered to 240c and those are performing much better. Now I am optimizing the pressure pulse, pulse time, and splitless time to get the PAHs back while trying not to distort the early eluters. Separation of Benzo(b)fluoranthene and Benzo(k)fluoranthene is still not as great as I would like it, but getting better.

I may go back after I get it optimized with the 4mm liner and try the 2mm liner again to see if there are differences along with the 5.2mm liner I have. Will be interesting to see if or how liner ID affects the peak shapes and early to late eluting peak abundances.

[GOM]

Hi James

I had a chat with Ray.

He hasn't tried this technique. "there are much better ways"

One of the things that leapt out to him was the injection temp of 250°C. That figure is too round and suggests that the injection temp hasn't been optimised.

Since analytes will have a significant vapour pressure even below their bpt you can get away with cooler injection temperatures.

The injection temp is really important and often overlooked. The lower the better. There is better control with cool injection.

For injection port temperature programming the heating rate is not important (something that you were concerned with). However the heating control itself is - and some manufacturers are better than others. Splitless transfer is slow. A slow controlled ramp is actually better. You are getting your narrow peaks from the solvent effect on the column.

A tip on the optimisation of the injector temperature is to use a split injection. This tells you if you have the right temperature and that you have good transfer before going to splitless.

So, might be worth trying the injection temperature optimisation (isothermal and programmed) if you haven't already done so

"Worth sitting down with a glass of wine, visualising what is happening in the injection port etc and figuring out an optimisation plan"

Regards

Ralph

Had to scribble all this down whilst talking so I might have missed something.
https://www.youtube.com/watch?v=FqHH1UQqIMs view from 2mins 15secs



Edit - for PTV consider Gerstel

[James_Ball]

I was able to lower the injection temperature to 240C which improved the recoveries of the labile compounds but caused lower response on the high boilers. Using a longer pressure pulse and longer splitless time brought the high boilers back and gave better peak shapes.

I just have the standard split/splitless injection port currently, so temperature programming would be a little tricky with the high thermal mass of the inlet taking a long time to equilibrate to temperature changes, but I am working on the powers that be to get a PTV soon. If I get this working enough to show the advantages of Large Volume Injection/Low Volume Extraction as a money saver then upgrading injection ports becomes much easier. So far things are looking really good.

[mckrause]

I have several ATAS Optic systems. They were a bit problematic for 8270; getting nitrophenols through the liners was a real issue. Larry Garretson with Scientific Instrument Sales (Round Rock, TX) did a LOT of work on those systems, and got them to work. For the late 90s/early 2000s they were pretty good, but I vastly prefer the Agilent PTV.

So back in the (bad old) days of environmental work for me, I had switched over to using 45 mL extraction volumes (VOA vials) followed by true LVI (not CSR). It worked beautifully with the exception of the previously-cited nitrophenols issue. Big, big difference in cost between VOA vials and 1 liter bottles, especially if you are paying the freight. The State auditors (Texas) were good with it as long as we maintained the sample/solvent ratios in the method, which we did. You could easily shoot 100 uL on the ATAS system and we were getting MDLs in the 1-5 ppb range for most 8270 compounds. Saved on solvent, bottles, internal standards, etc. etc. Pretty easy cost justification for the injectors.

Now, thankfully, I don't have to play in the USEPA world. I can truthfully say that I don't miss it at all.

[James_Ball]

Been experimenting with the new Multi Mode Inlet and I really am liking the cold splitless injection mode.

I have started moving the TCLP SV extractions over to smaller volumes, 100ml extracted with proper ratio of solvent in the sample bottles, 3x each for acid and Base/Neutral.

Everything is recovering great except the hexa chlorinated compounds (hexachlorobutadiene 26.4%, hexachlorobenzene 24.2% and hexachloroethane 50.5%), Pyridine and 1,4-Dichlorobenzene are in the upper 60s recoveries and all others are above 70%. This is actually better than the 1L extraction using liquid/liquid setup.

The first two shakes are done on acid fraction with diethyl ether/methylene chloride then only methylene chloride on the third shake. Add 10g sodium chloride and adjust to >12pH and do three more shakes with methylene chloride.

I am wondering if adding the salt at the beginning or possibly adding the diethyl ether to all shakes will improve the recoveries of the hexa chlorinated compounds? Anyone have any thoughts on this?

The Pyridine and Dichlorobenzene are probably being lost in blow down so that should be easy to fix.

[sav]

I am wondering if adding the salt at the beginning or possibly adding the diethyl ether to all shakes will improve the recoveries of the hexa chlorinated compounds? Anyone have any thoughts on this?

Yes, adding NaCl at first step must improve recovery. You can also try to change extragent composition or extraction sequence (DEE > DEE/DCM > DCM; DEE/DCM > DCM > DCM etc.). Also take care about evaporation step.

[Rndirk]

Everything is recovering great except the hexa chlorinated compounds (hexachlorobutadiene 26.4%, hexachlorobenzene 24.2% and hexachloroethane 50.5%), Pyridine and 1,4-Dichlorobenzene are in the upper 60s recoveries and all others are above 70%. This is actually better than the 1L extraction using liquid/liquid setup.

From my experience, For these kind of compounds, also consider instrumental issues. If you perform LVI, you might lose some low boilers during your solvent vent step.

Moreover, in SIM or MRM, small molecules with relatively lots of chlorine atoms are annoying due to the distribution of isotopes and cumbersome fragmentation.

EDIT: The issues above are not responsible for bad recoveries if they occur reproducibly troughout your calibration and samples... Not sure if they do because my calibration curve and recoveries are often not what they should be with especially the compound hexachloroethane

[James_Ball]

Everything is recovering great except the hexa chlorinated compounds (hexachlorobutadiene 26.4%, hexachlorobenzene 24.2% and hexachloroethane 50.5%), Pyridine and 1,4-Dichlorobenzene are in the upper 60s recoveries and all others are above 70%. This is actually better than the 1L extraction using liquid/liquid setup.

From my experience, For these kind of compounds, also consider instrumental issues. If you perform LVI, you might lose some low boilers during your solvent vent step.

Moreover, in SIM or MRM, small molecules with relatively lots of chlorine atoms are annoying due to the distribution of isotopes and cumbersome fragmentation.

EDIT: The issues above are not responsible for bad recoveries if they occur reproducibly troughout your calibration and samples... Not sure if they do because my calibration curve and recoveries are often not what they should be with especially the compound hexachloroethane

The low standard for hexachloroethane was 0.05ppm and though it was a small peak I did still get 0.999 linear fit calibrating up through 5ppm. The cold splitless injection seems to work great for these compounds. Pentachlorophenol was the only one I could not see at 0.05, but it did show up at 0.1ppm. With a cold splitless injection at 5ul I will be able to hit the lowest TCLP limits doing a 100ml initial sample volume and 10ml final extract volume, which will save a lot of time and solvent versus the full overnight liquid/liquid extraction.

This run was in full scan mode, with SIM I could really knock the detection levels down if needed.

[James_Ball]

I am wondering if adding the salt at the beginning or possibly adding the diethyl ether to all shakes will improve the recoveries of the hexa chlorinated compounds? Anyone have any thoughts on this?

Yes, adding NaCl at first step must improve recovery. You can also try to change extragent composition or extraction sequence (DEE > DEE/DCM > DCM; DEE/DCM > DCM > DCM etc.). Also take care about evaporation step.

I tried to not bias the test too much in my favor so I sent the extracts over to our extraction lab to have the solvent evaporated and switched to DCM. They use the Zymark evaporators and I am not certain those are optimized as well as they could be since they are normally trying to remove 300+ ml of solvent as quickly as possible. If we can transition to only needing to evaporate 60-80ml of solvent we should be able to optimize that process as well.