FID flame go out

[J_Imi]

My FID flame go out during a run.

GC: 6890, FID gases: mkeup N2: 20 ml/min, air: 300, H2: 40, Temp: 300°C, jet is clear.
Column flow: 2.5 ml/min, injection mode: splitless (purge flow 20 ml/min after 0.1 min). Inlet temp.: 250°C, inlet liner volume ~ 900 ul.
Column: HP-5, 30m*320um*0.25um, calculated holdup time 1.25 min

My sample: water (condensed steam with organic impurities).
The 'flame go out time' ~1,3 min (at every run).

I changed air flow 250-400 ml/min, H2 flow 25-40 ml/min, but nothing happened.

Any idea?

[chromatographer1]

First be certain that the flame actually goes out and it isn't a shorting of the detector due to water or something in your sample.

It appears that something in your sample, be it the water plug or something else puts out the fire !

I would try running the sample split and see if the flame continues.

That it will, is my guess, then you can start the splitless injection for longer and longer periods until you find the point at which the flame goes out.

Running rich usually keeps the flame going. I would run H2 at 40 and decrease the air. It may be that the water plug is enough to put the flame out.

Good luck,

Rod

[J_Imi]

If I use split mode the flame is OK, but the organic impurities concentration in water sample is very low (~ 1-5 ppm), so I need splitless mode.

I set splitless time to 0.05 min. After two injection the flame goes out again.

[chromatographer1]

Well, I have two suggestions.

The first and lesser likely to work is to do service cleaning on the detector including a new flame tip.

The second and more likely to work is to install a packed column flame tip instead of the capillary flame tip. The capillary is about a 0.011 " ID and the packed is about a 0.020" ID. The flame may not go out and it is an easy and quick swap.

If the second doesn't work then do the first. If the first doesn't work, then get another GC, preferably a different brand and see if it will work on it.

Good luck,

Rod

[Don_Hilton]

You might look at using an injection thecnique that would not put large quantities of water into the GC. You can look at methods for purge and trap analysis to see the kinds of analytes you can measure in water.

SPME can also be a nice way to collect organics out of water and get them to a GC inlet. If your compunds of interest are somewhat polar, adding a salt to increase ionic strength could push the partitioning in favor of the organics going into the SPME fiber.

[chromatographer1]

Don's suggestion has great merit.

SPME especially is relatively inexpensive, much less than another GC and less than a purge and trap analyzer.

As long as the analytes are not lacking in volatility from water SPME can see parts per trillion.

Also a multidimensional packed column with a backflush to detector valve design, using a Carbon packing with or without a 1% coating of C-wax can trap, concentrate, and refocus a large injection of water (easily 5-25 µL) of water, but that would require parts of upwards of $1000 to install. (methane through propane might be lost).

Good luck,

Rod

[Peter Apps]

As long as the water is vapour it will probably not put out the flame, but if you have a condensed plug being pushed down the column by the carrier gas it will flash vapourize when it hits hte hot injector and blow the flame out.

What is the column starting temperature ? Can you increase it ? Can you try a megabore column, which is more robust to this sort of abuse ?.

Peter

[J_Imi]

Thanks, my problem was solved:
Oven temperature program: 50°C, 5 min → 20°C/min → 300°C, 7.5 min
When I changed the initial temperature to 100°C the flame burned continuously (no condensing, no plug), so I was able use longer splitless time.

[AICMM]

J_Imi,

I have seen this before. The issue, I believe, is that the detector is too cold. I would recommend external insulation or, even better, external heat applied to the FID. Does not have to be too much extra heat but enough to keep the chimney warm. Moot since you have the problem solved but I did want to add my two cents worth.

Best regards,

AICMM

[J_Imi]

Thanks AICMM

I think that the 300°C are not too cold, but I will try your idea nearest.

[AICMM]

J_Imi,

I apologize, I did not note your operating at 300C. This should be hot enough... Still, I have found that HP/Agilent FID's don't retain heat very will (nature of the configuration) so if you have some available glass wool, or better yet an insulating blanket type material, I would recommend insulating the top portion of your FID. Next time you pull your detector apart for maintenance, it would be interesting to see if you see significant corrosion due to water collection in the chimney.

Best regards,

AICMM