Best separation of PCDD/F on GC/MS/MS

[Hornet]

Hello, i'm writing here hoping someone else has experienced the problems i'm facing by setting up a quantitative method for PCDD/F in enviromental matrices.

We recently bought an Agilent 7890+7000C GC/MS/MS and i'm in charge for the development of this method.
I already figured out how to do extraction and purification.
My method will use isotope dilution and is inspired from EPA1613B, with various modification to enhance the speed of the method.

With the instrument i already ordered a second injector for various purposes. By reading the original method (which uses GC/HRMS in SIM mode) i understood that the 17 toxic 2,3,7,8-substituted congeners can't be perfectly separed from their non-toxic congeners, normally found in some enviromental matrices.
In order to separate all of these 17 congeners, if they are positive on the first run (usually on a DB-5 Type column) there is the need of a second column confirmation; they suggest a highly unstable cyanopropyl column, which will limit my application because of its temperature limit.

Of course the method is of 1994 and columns and instrument developed a lot from there. In fact i will use a 60m Rxi-5sil MS as first column which, from applications found on the internet using HRMS, will separate the 17 congeners a lot better than a classic xx-5 column but it still has some coelutions.

My question is: anyone here achieved a decent separation of all 17 congeners with one injection? On the chart, will MS/MS technique be more effective than SIM to separate isobaric congeners?
I want to avoid the second column confirmation because i don't want to specialize the instrument on PCDD/F only, i would like to use a second high temp column for other purposes rather than place a polar column in there and use it once a month.
I'm also considering GCxGC option, maybe placing a 30m arylene like rxi5 first, and a more polar column last.

Of course i'll try a lot by myself (and i'm already doing it), i'm just asking if someone else has already worked on this problem and has some inputs.

Thanks you

Davide

[James_Ball]

http://www.restek.com/catalog/view/41734

or

http://www.restek.com/catalog/view/32438

as the secondary column.

Use dual injectors with a Y connector like this one

http://www.restek.com/catalog/view/3355

to join the columns before entering into the MS.

I am running this setup with a Rxi5SilMS column on front injector with autosampler and the Rxi624SilMS on the rear injector with a purge and trap setup for volatiles. The two columns (0.18mmID works best) run into the Y connector with 0.53 transfer line that holds the Y connector about a half inch out from the MS inlet nut going into a 7000C QQQ. When running the front column set the rear to about 0.3ml/min and front to 1ml/min and vice versa when running the rear column. It is enough flow to prevent backflow into the unused column and having the short 0.53mm transfer line allows both columns to exit into a vacuum so the GC can accurately set the flow based on head pressures. For your setup you could either use dual autoinjector towers or simply switch from front to back with a single tower depending on which method you are running.

The real trick to the setup is the large diameter transfer line, if you try to use a smaller diameter the instrument will have trouble setting the column flow correctly because there will be backpressure from the exit of the columns instead of full vacuum there. Otherwise for me it has been working great for about two years now.

[Hornet]

Thank you for the input, this means you are still doing double injection for confirmation.
Agilent offered me a two-holes ferrule to install column on the transferline, i'll try both solutions.

[James_Ball]

Yes it would requires a second injection for confirmation but if you run dual columns you would not have to worry about switching out columns and waiting for the instrument to pump down and stabilize again.

I am considering the two hole ferrule but was not sure if two columns will fit into the interface of the MS, if so that would be an even better setup than the Y connector.

[Hornet]

Hi, i wanted to update my findings on the development of this method.
I'm running the two-hole ferrule setup, so far it is working great really, standard flow on injection column, 0.6 ml/min on the other column, quantitation limit is good and no leaks so far.

I'm using a rxi-5 sil MS 60m as primary column for PCDD/F, so it's an silarylene substituted rather than a standard DB-5. I've found that it doesn't separate all of 17 2378 congeners.
The improvement over the db-5 is separation of tcdf from non-toxic congeners which is good, but still doesn't separate 23478 PeCDF and 123478 HxCDF.
I bought a dioxin2 60m column for confirmation, it does separate those two PCDF (but does not separate others which are then separed from Rxi5sil), the sum of the 2 column allows me to perfectly quantitate all 17 congeners without bias due to interferents, so i'm set column wise.

Now i'm asking this to the forum, hoping someone can help me on this preparation step.
I'm VERY new to micro-concentration and i don't understand or missing something on this EPA1613 step:
After purification i end up on toluene solvent (using CAPE tech columns), EPA1613 says to evaporate solvent under a gentle stream of nitrogen and to add 10 ul of nonane (keeper) when the extracts are low of volume and are transferred into a autosampler vial.
The thing i don't understand is when it says "evaporate to the level of nonane".

I mean...i really tried to evaporate under a gentle stream of nitrogen starting from 100 ul of tolueme + 10 ul of nonane, i really didn't see any stop on the evaporation or any slowdown or any advice that made me say "hey you reached nonane level so stop the evaporation". And of course it ended with a perfecly dry vial.

I'm using standard 2ml autosampler vial with 250 ul inserts, there isn't any volume mark so...how you get this done?
I'm not considering any high boiling solvent like dodecane or tetradecane, which will work i guess, but will take other chromatographic problems with them.

Boiling point on toluene and nonane are close, maybe this is the problem; will this "keeper" trick works if i use hexane instead of toluene?

I need to know when the volume is 10ul because the volume is of course used for calculation and needs to be correct, all the methods are very vague on this step. I'm thinking about just going dry very slowly and reddisolve analytes on the volume i need...

Thank you

Davide