Atmospheric methane measurement

[kltowery]

Hi all
I'm new to this and not sure how much info is needed, so pardon if this is extra!


I'm trying to measure changes in methane levels and have no problems with high levels (1-10% CH4 in air) but have some difficulties with atmospheric, so in the range of ~2ppm. I am able to get some peaks that look good, but then others will have a huge tail, and occasionally the baseline will start to oscillate for 15-20 minutes and I have nothing to do but wait it out. I'm also seeing around 35% RSD at these levels

What I would like:
1) any recommendations on temps/flow rates/any modifications I need to make

2) what to do about the tailing effect

3) how to stop the oscillating baseline from ruining my day!

4) any ideas to decrease the error



I'm using a shimadzu GC 2014 with 80/100 porapak Q and 80/100 porapak N columns. We have a FID and TCD detector, though I have not yet used the TCD so I'm not super familiar with it.
Current method settings: Nitrogen carrier, flow rate is at 50. Injector 150C, column 100C, FID 150C.

I'm injecting manually, 500ul, and it may be as simple a problem as me being a human and not an autosampler?

[chromatographer1]

It sounds like you have some real problems.

Are you using 1/4 inch glass columns or 1/8 inch metal columns.

Have your columns been conditioned PROPERLY ?

Do you have a reference chromatogram so you can compare your results with a reference?

good luck,

Rodney George

[AICMM]

kltowery,

Oscillate: with width or very sharp? Like spiky or wavy? If spikes, could be electronic problems with FID or electrometer connection/card. If wavy, then could easily be junk coming off the columns and, thus, Rodney's comments about proper conditioning. In addition, might be something you are adding by syringe injection or from the nitrogen you are using, especially if the peaks are very broad. Do you have filters on the nitrogen cylinder?

The tailing question might be resolved in a different manner. I would first try to get to a point where you can get a reproducible clean chromatogram from a clean matrix. To do this, I would shoot something like UHP helium for a number of shots to see if you can get to a stable chromatographic condition. Once that is established go back to room air and see if it starts mis-behaving again.

0.5 mL is a lot to shoot by syringe, not the best way to do things but if they are packed columns, you will probably be okay.

Probably not a bad idea to get the TCD up an running just to have a second channel and look at stability of overall system.

Best regards.

[kltowery]

Rodney:

that is NOT what I wanted to hear! haha... The columns are metal and have been stagnant for at least a year or two before I picked up the project. They could be several years old. We did have a Shimadzu tech come out to teach me the ropes of running the GC and using their software, but obviously he doesn't care much about columns as he doesn't sell them. He didn't mention anything about the columns needing work or replacement, so I've been using them as is.

As to reference chromatogram, what I have is that 60% of the time I get pretty peaks at the same RT and they're making a clean standard curve (as long as I weed out the strange peaks)

AICMM:

The baseline oscillates like a sine wave with a low amplitude and long period... maybe 2000-3000mv and .6s wide? (from memory). This only happens sometimes and can go on for 20-30 min. I believe we do have filters on the N2 cylinder, but I can double check that. There is sometimes liquid in the bottles from which I'm sampling the headspace, so maybe I'm getting some of the liquid into the syringe?

What volume would you recommend for injecting? I have tried 100-500ul with no noticeable difference in peaks.


Thanks so much for your help thus far! Any comments on temperatures or flow rates I'm currently using? I'm mainly going by previous user's methods for these and I'm not sure how much 'play' I have.

[chromatographer1]

Your columns are probably ok. As long as no one injected junk onto them. 1/4 inch metal columns: 50cc/min is an appropriate flow. It is a bit fast for 1/8 inch columns:20-40 cc/min would be more appropriate. But faster flow won't hurt anything.

I would keep your detector and oven temperatures below 170°C.

Make sure your septum and your syringe are not leaking. Inject your 500µL over a period of 2-3 seconds, not an instantaneous "burst"

Condition overnight at 150°C if necessary. Keep dry, oxygen free gas flowing during the conditioning.

Porous polymer columns can be packed poorly from vendors. I suggest that you buy new columns from Supelco as I have taught my former employer to pack these difficult packings well (especially N) with minimal voids. There is a science to it, not just hammering the packing with vibration which doesn't work.

Your baseline problem is probably a detector hardware maintenance problem, not a column issue.

best wishes,

Rodney George

[AICMM]

kltowery,

I am a bit confused. 0.6 seconds wide is very narrow, noise like. If truly that narrow, start looking at how well the electrometer card is seated, any cabling issues, etc....

Wider than that, and based on what you describe you are shooting, I would wonder if you don't have late eluting constituents that came from 1, 2 or 4 runs before....

Possible for you to post a chromatogram?

Best regards.

[kltowery]

Please pardon my ignorance, but I'm not exactly sure what is entailed in baking out the column over night. I was told I should keep the Inj and FID temps above the column temperature, I assume to keep things from condensing in the system? Should these just be off while the column is at 150C? Should all 3 gasses (Air, N2, H2) be on during this time?


Thanks so much!


A few pictures of what I'm dealing with:
I couldn't get any shots of the meandering baseline because I only collect data to 2 minutes and the problem decided not to come back today. Instead it was replaced with this tailing/extra peak business that I have not seen before today. I swear a new problem shows up every time I think an old one is solved!
And sorry about the links. Embedding images is not cooperating either. Sigh... This is going to turn me Amish.



http://picasaweb.google.com/lh/photo/04 ... directlink
This is as good as it gets, 4ppm, injected a week ago with the previously described parameters. Peaks look like this about once in 4 injections.

http://picasaweb.google.com/lh/photo/o- ... directlink
this is injection of a 10ppm calibration mix with air balance (purchased)


http://picasaweb.google.com/lh/photo/zM ... directlink
500ul of supposedly pure N2. Things got even weirder after this and

[chromatographer1]

Your chromatograms were informative.

They state that you are shooting into one of two columns, each column is 2m in length and 1/8 inch in internal diameter.

Your flow is fast but acceptable.

Since you are using FID I suspect that you are injecting the sample with a leaking septum, or the septum is leaking after you inject your sample.

Your standard may be contaminated. But with your high oven temperature your peaks could be a great many things. (methanol, acetone, propane, butane, methylene chloride, acetonitrile, etc) You should run your oven at 60°C and you should run blanks until you have a flat baseline.

Running nitrogen carrier if you have a leak you can change your baseline as the flame mix changes at the detector with the change in flow.

I suspect that you are contaminating your syringe with vapors in your lab. I also suspect from your tailing peaks and variable baseline, you have leaks, probably from the septum.

Thanks for posting the 'grams.

Good luck in fixing your problems.

Too bad you don't have a valve to inject your samples.

Rodney George
consultant

[kltowery]

Hi all,


Baked column overnight at 150C, changed out septums, new syringe/needle, and different standards:

I tried setting the column temp at 60C, but for some reason the FID would not light or it would not remain lit. I doubt it has to do with the column temp, but they did both happen at the same time. I rolled back to old methods and checked the filament (it was fine) then started up again and it lit as normal.

http://picasaweb.google.com/lh/photo/q ... directlink
Here's a 0.2%CH4 standard where I injected 200ul. Do you think I can do anything about the tailing or just live with it here?

http://picasaweb.google.com/lh/photo/Di ... directlink
Here's the same sample, injected 500ul just 3 min later and this happens. Could it be explained by a pressure pulse from the syringe or is something more seriously wrong here?

http://picasaweb.google.com/lh/photo/RC ... directlink

Finally, here's a pic of the same 4ppm sample that I injected yesterday. STill tailing but much less crazy. Am I safe in assuming that the negative starting point is because I didnt' wait long enough for the previous sample to clear out and the system to re zero?

[AICMM]

kltowery,

I would second Rodney's contamination theory. On your last post it appears most apparent. The broad peaks (not the tailing) that appear in your chromatograms look like components that are late eluting. To test this theory, set up a run for 30 minutes or so, make an injection of you 4 ppm methane and let it go. If nothing, just hit start again, no injection, and keep watching. Should not take more than 2 runs if that is the culprit.

Second, regarding #3 where you shot pure nitrogen your scale factor is X100 so this may be starting to get down near the noise. In addition, you may be getting near (at low ppm) the certification limit for nitrogen from the plant so it could still be present. Five 9's purity can still have ppm levels of impurities.

Finally, Rodney is absolutely right about the need for a valve. I have visited with quite a few people who have had very poor luck with syringes and fixed gases. Some of them switched to valves and the chromatography improved significantly.

Best regards.