Re-collecting injected sample using flow splitter setup?

[feejee]

Hi,
I am using a GC which has a Y flow splitter near the end of the column, the one part of column then goes to the FID, the other part to outside the machine for use with an electroantennogram (EAG) setup. The column leaves the GC through a heated arm, which I have set to 300C.

What I am wanting to do is re-collect half of my injected sample from the part of the column which comes out of the GC. Has anyone ever tried to do this? I have tried chilling a pasteur / or vial right down and placed this over the end of the column, with the thought being the hot gas flow containing sample will condense against the side of the chilled glass. I have also attempted to collect it by bubbling the end of the column straight into solvent. Occasionally, i have obtained about 0.02% of my sample back which is useless for what I need to use it for next, but most of the time I get nothing back.

Has anyone any successfully tried this before?

My main aim is to find a way to separate out a mixture of about 40 hydrocarbons (alkane, alkene, diene mix) into much smaller fractions. Ive used TLC/ Ag-TLC to reduce the components in my mixtures but still need to separate out these further to test either componenets individually or mixes of 2/3 compounds. Im a biologist, so my chemistry is still at beginners stage shall we say - I would be incredibly grateful for any advice.

Thanks

[AICMM]

feejee,

I set something like this up for a customer but I don't think they ever implemented it. I took the exhaust of a TCD and hooked it to a three way valve under the GC's control. Most of the time the effluent from the detector passed to exhaust but they could flip the valve to a collection vessel at discrete times. In their case, they were shooting a sample of 40% analyte in a diluent gas.

The bigger question is what do you hope to do with the sample after you collect it? If you want to do something like NMR where you would need a lot of material, this is a difficult task unless your analyte is at large concentrations. Figure this, if you are shooting 1 uL of a liquid with a component that is at 500 ppm, then you are only putting about 500 ng of the component on column. It is going to take a lot of injections to collect a significant amount, right?

However, if you are looking to do something like secondary GC-MS, then 500 ng is plenty of material to collect. For TLC, I will admit my complete and utter ignorance and say that I don't know how much you need to collect. (You have to understand.... TLC is not GC......)

So, how would I approach this problem? Two things I would look at. First, sorbent tubes would allow you to do secondary analysis like GC or GC/MS even with only a small amount of component collected. Or, in your case, a capillary tube cooled in something like dry ice acetone (assuming you are a grad student and don't have access to LN2) that you could subsequently elute with 20 uL (or so) of solvent onto a TLC plate.

How's that for free advice (beware, you get what you pay for.... ;>) )?

Feel free to contact me if you wish to discuss further aicmm at flash.net

Best regards,

AICMM

[Peter Apps]

Hi Feejee

I suspect that after the Y-splitter the antennogram line might have a much higer flow resistance than the line to the FID. The effect will be to direct nearly all the sample to the FID. How do the two lines compare for length and internal diameter ?. Are you adding make-up or purge flow anywhere on the antennogram side ? Increasing the temperature of the antennogram line will actually make the problem worse by inceasing the viscosity of the gas.

Another possibility is that the antennagram side has become adsorptive for some reason and so the sample is sticking to the walls.

If there is a difference in gas flows through the two lines the appearance of the peak at the FID might not coincide with the analyte entering your trap.

Have you actually done antennography with this set up ?

Peter (another biologist)

[feejee]

My downstream application is electrophysiology, so we dont need large amounts of sample, but more than what I can get at the moment. Unfortunately, we cant use the EAG set-up as the hydrocarbons of interest to us have low volatility, and as such wouldnt make it through the glass tubing that carries the effluent from the heated arm of the GC to the antenna. Hence I need to collect small fractions of my mix and test each individually. I believe the EAG setup has previously been used successfully on more volatile components. Ive never used it for its purpose myself, but I did run some ethyl acetate through the GC to check the timing of the FID detecting it and the smell coming out the heated arm, so definitely for the smelly high volatile compounds the system works.

We actually replaced the flow splitter and put new guard column on it a few weeks ago, as we thought maybe the sample was getting stuck in the column. The lengths and internal diameter are the same -we used new guard column on both the FID and the EAG arm.

Ive checked the flow, and it is very close to 50% each way at the split, so 1ml/min comes out of the EAG arm (from a 2ml/min flow).

So far ive attempted to collect:
1) into a glass pasteur, inserting the guard column from the heated arm into the tip of this and tilting slightly upwards. I have chilled the pasteur with just ice and also currently insert the pasteur into an aluminum block the length of the pasteur which is chilled in liquid nitrogen.
2) into a small vial containing hexane (2ml). Inserted the column from the heated arm through the lid of the vial (also inserted into the lid a glass microtube as a chimney to allow flow).

Im wondering a few things:

Whether the sample effluent actually comes out of the EAG arm, or again gets stuck inside the column. But given that the FID detects plenty of sample, and the columns each side of the split are the same length I assume it gets through the EAG side too.

Am i trying to collect into much too short a tube? is 1ml/min flow rate too fast to possibly collect into something as short as a pasteur?

Too cold? I am concerned with the use of the metal block chilled in liq N2 that the end of the guard column is getting too cold and the compounds could be accumulating at the tip. I could try to find a longer tube to collect into, chill with ice the end closest to the heated arm, and then further along have the liq N2 cooled section.

[Peter Apps]

It sounds as if the basic set up is OK. One more question - how are you measuring how much you catch ?, presumably you run it back through the GC somehow ?

Peter

[feejee]

After I try to collect into the pasteur, I rinse the pasteur through with 300ul hexane, and run through the GC again to look for the desired peak. At the minute im just running through a C26 standard (5ug is injected initially) to try to recollect this before using my real samples.

[Peter Apps]

After I try to collect into the pasteur, I rinse the pasteur through with 300ul hexane, and run through the GC again to look for the desired peak. At the minute im just running through a C26 standard (5ug is injected initially) to try to recollect this before using my real samples.

Are you calculating the concentration in the rinsings using an external calibration ?, or are you just comapring the size of the peak from the rinsings with the size of the peak from the original injection ? If the latter then you have to take into account that you are not injecting all that you catch, and if you are doing a split injection then even less of it gets onto the column and ends up in the FID the second time around.

Say you inject 1 ul with a split of 10:1 - what goes onto the column is 1/3000 of what was in the trap.

What column are you using ? 5ug is a lot to be putting through a capillary column.

Peter

[Johnny Rod]

Is it possible to heat the cold parts before the antennae? You can get heating tape that you just wrap around stuff and switch it on, from people like Sigma.

[CE Instruments]

You could buy a commercial system
Prep GC
I used a very basic GC and collection device at college with packed columns, however with capillary volumes you need to be certain when your peak is eluting from the split (not necessarily the same time the peak is seen on the FID). Your Hydrocarbons will be vapours as they exit the column so just hoping they condense and not get swept away with the carrier gas may also be optimistic.