FID's flame turns off when the main solvent (MTBE) appears

[hmmaps]

any idea why this could be happening??

We have tried in 2 diferents CG (Agilent CG 6890) without success.

The method has been used for a long time now without problems, so we are running out of ideas.

Any idea??

[gserrano_84]

How old are your 6890s? at grad school i was working with old units (mid 90s), and the flow rate values that appear on the front panel didn't match the ones measured with a flow meter at the outlet of the FID, For example, for H2 i had it at 40 mL/min, but in reality it was 30 mL/min, same for air. I had a similar problem with an FID flame, that was solved by increasing my values of H2 and H20.

Now, in your particular case, I guess the MTBE solvent is adding extra oxygen to your flame. In that case, i'll probably increase my H2, and reduce my air flow, that may help, specially, if you indicate that it happens when the MTBE peak appears. by the way, what are your current values of H2 and H20 flow rates?

Gustavo/Supelco

[gserrano_84]

when i put h20...i meant air.... ups....

[dblux_]

any idea why this could be happening??

We have tried in 2 diferents CG (Agilent CG 6890) without success.

The method has been used for a long time now without problems, so we are running out of ideas.

Any idea??

What are your gas flow rates ? (FID incl. makeup, carrier)
Injection volume, split ratio ?

[hmmaps]

How old are your 6890s? at grad school i was working with old units (mid 90s), and the flow rate values that appear on the front panel didn't match the ones measured with a flow meter at the outlet of the FID, For example, for H2 i had it at 40 mL/min, but in reality it was 30 mL/min, same for air. I had a similar problem with an FID flame, that was solved by increasing my values of H2 and H20.

Now, in your particular case, I guess the MTBE solvent is adding extra oxygen to your flame. In that case, i'll probably increase my H2, and reduce my air flow, that may help, specially, if you indicate that it happens when the MTBE peak appears. by the way, what are your current values of H2 and H20 flow rates?

Gustavo/Supelco

Our 68901's are not too old. Maybe 5-7 years or even less. They have a mantainance plan, with verifications twice a year. The conditions of the method should not be a problem since we have analised similar samples before with good results in the same GC.

We work with 45mL/min of H2 and 450mL/min of Air. 30mL He Makeup flow. The flows are steady and the flame ignites without dificulties. The system has a PP inlet, so there is no split involved.

We have tried cleaning the detector, but it still happen the same. We'll try to lower the Air flow and see what happens...

thanks for the answer!

UPDATE: We have tried injecting only 0.5 uL of sample (method volume is 1uL) and it works well, so it seems clear that the detector collapses with the normal injection volume. Maybe the detector is still not clean. Any suggestion??

[Peter Apps]

Let's assume that when you write "ml" you really mean "ul" for microlitres and go on from there.

You are doing a 1 ul splitless injection with a mystery flow rate into a mystery column at mystery temperatures, leaving us to guess wildly.

My wild guess is that your solvent is not what you think it is - did the onset of the problem coincide with opening a new bottle, or some decanting from stock into flasks or something similar ?

Peter

[hmmaps]

Let's assume that when you write "ml" you really mean "ul" for microlitres and go on from there.

You are doing a 1 ul splitless injection with a mystery flow rate into a mystery column at mystery temperatures, leaving us to guess wildly.

My wild guess is that your solvent is not what you think it is - did the onset of the problem coincide with opening a new bottle, or some decanting from stock into flasks or something similar ?

Peter

Yes, sorry, I obviously meant uL.

I didn't post much information about the method because I'm (almost 100%) sure it's not a problem of the method itself. As I said, this method has been used (in the same GC6890) for a lot of time. It's not something new. By the way, I'm certain that the sample is what it is: pure MTBE.

The column is a Porapak Q; 2m; 2mm id 80-100 mesh (it's a really old column). Column flow 30mL/min (steady).

We solved the problem injecting this sample with another GC (this time a 7890), but in the 6890, it continues to turn off the flame at the same point. It must be something related to the detector status (at least is our guess...).

We have contacted our GC supliers but I was courious if someone else have found this problem (or similar).

Thanks for the answers.

[chemist10]

Assuming, you tried two different GC, nothing wrong with your detector. Something happened to your gases (air or hydrogen). Try to change ratio of your detector gases.

[cleh]

Any chance that you changed the jet before this started happening? If so, make sure you're using the 0.018-inch ID jet or 0.030-Inch ID jet for packed columns, not the cap column jet (0.011-inch ID). May want to replace the jet if you just cleaned it when you cleaned the detector.

[dblux_]

... Column flow 30mL/min (steady)...

If column flow is 30 mL/min you may substantialy (!) reduce makeup flow or even turn it off completely.

Hope this will solve your problem without reducing injection volume to 0.5 µL.

Let us know.

[hmmaps]

Thanks for all the answers.

The problem is solved now. After the weekend we have been able to run the sample in the same GC that failed last week... without doing nothing special to correct the problem (apart of what we tried last week).

Our guess is the same that some of you have pointed out. It seems that by mistake (I was not aware of that... until yesterday), the two bottles of sinthetic air went empty on monday, so probably some dirt came into the pipes and ended in both detector. We have in-line gas filters, but appearantly it did not worked very well. We will change them though they are not expired.

Well, again, thanks for the answers!