Bringing a 6890N back into operation

[rb6banjo]

What's the temperature of your FID? If I'm you, I'd try to get that up pretty high (at the isothermal limit for the column installed). Get rid of as many fingerprints, etc. as you can. I'd also run the columns up as high as you can. Are you using zero-grade air to support your flame? If you're using breathing air, it will be louder. I use zero air and my baselines (Agilent 6890's) are more like 3-4 pA.

Be conscious of the fact that the rotors of the injection valve will limit the oven temperature. If you go too hot, you'll melt them and ruin that valve. I have a GC that is primarily designed for analysis of impurities in CO2 and the max. oven temp. is 150 °C because of the rotors in the valves in the oven.

[Gaston]

What's the temperature of your FID? If I'm you, I'd try to get that up pretty high (at the isothermal limit for the column installed). Get rid of as many fingerprints, etc. as you can. I'd also run the columns up as high as you can. Are you using zero-grade air to support your flame? If you're using breathing air, it will be louder. I use zero air and my baselines (Agilent 6890's) are more like 3-4 pA.

I'm running it at 300°C right now. The column (GS-CarbonPLOT) has a limit of 360°C (and indeed, I have not done a bake out of the column so far, as most likely this will not be the column that I will use for real). But it should be good to do this, I suppose.

I think the air might not be of the best quality (it's connected to a "utility" compressed air line), so that might certainly be an issue. I checked one of our other GC's in operation, and its FID has a baseline of 18 pA or so (which I suppose it's also rather high).

Be conscious of the fact that the rotors of the injection valve will limit the oven temperature. If you go too hot, you'll melt them and ruin that valve. I have a GC that is primarily designed for analysis of impurities in CO2 and the max. oven temp. is 150 °C because of the rotors in the valves in the oven.

Can this be a problem even if the six-way valve is in the valve box on top of the GC? However, inside the oven there's the tee which divides the column outflow to the two detectors (such as this one: http://www.vici.com/vfit/microvol-tees.php). Could that part get ruined by running the oven at very high temperature?

[rb6banjo]

If your valves are in an external box, you're ok with oven temperature. The box is probably heated to a temperature warm enough to inhibit condensation of your analytes but not so hot as to melt your rotors. 300 °C detector temperature is plenty to get rid of fingerprints/oils, etc.

House air has a lot of junk in it. It's not designed for chromatographic applications. It's only good for actuating pneumatic valves, etc. Your baselines with get much more quiet when/if you switch to zero grade air. If you are not looking for "microsniffs of stuff" in your samples, your air might be fine. Just know that you won't have optimum detection limits with the house air. Also, the house air might fluctuate some as the compressors cycle and age.

The tee is probably fine for whatever temperature in the oven. No moving parts. The rotors in the valve have to be flexible to make good seals.

Seeing peaks when you inject?

[Gaston]

If your valves are in an external box, you're ok with oven temperature. The box is probably heated to a temperature warm enough to inhibit condensation of your analytes but not so hot as to melt your rotors. 300 °C detector temperature is plenty to get rid of fingerprints/oils, etc.

House air has a lot of junk in it. It's not designed for chromatographic applications. It's only good for actuating pneumatic valves, etc. Your baselines with get much more quiet when/if you switch to zero grade air. If you are not looking for "microsniffs of stuff" in your samples, your air might be fine. Just know that you won't have optimum detection limits with the house air. Also, the house air might fluctuate some as the compressors cycle and age.

Thanks for the tips. If I end up using the FID, I think I would have to change to better quality air, as I think that the amount of analytes in the sample will be rather low. But I hope that I will be able to do my analysis using the TCD only, hopefully!

I have noticed that the FID baseline has gotten a bit less "driftey" in the past days, though, and now it's mostly stabilized at 21-22 pA (chromatogram coming up).

The tee is probably fine for whatever temperature in the oven. No moving parts. The rotors in the valve have to be flexible to make good seals.

Seeing peaks when you inject?

I injected the contents of the sampling loop (i.e. air) through the inlet in splitless mode. In contrast to the chromatogram I posted earlier, I get a huge broad peak instead. The two "jolts" in the FID, which occur simultaneously with valve switching on and off, still remain. I also installed ChemStation MSD to control the GC and acquire data, seeing as this is the software I will most likely end up using. With the same settings in the two programs, the obtained chromatograms look, well, similar.

Using OpenLAB CDS ChemStation Edition


Using ChemStation MSD D.01.00


In ChemStation MSD, the y-axis is "response" instead of microvolts or picoamps. Is there a way to change the settings so that the "real" signal is displayed instead? (although I can imagine this will be rather irrelevant in the end for the integration and quantitation).

I feel like a bit of a monkey here hitting random keys in the GC, but from this very blunt method, I suppose it is fair to say that the equipment works mostly alright I'd like to try something more methodical, though. As far as I know, this column is for separating gaseous hydrocarbons (GS-CarbonPLOT), but I think I don't have anything in the lab at the moment that could work as a suitable "problem sample". Any suggestions for that? Maybe inyecting a mixture of pentane and hexane?

[rb6banjo]

It can be tough to evaluate a column like this on only air.

I would agree. It seems you're on the right path. Do you have a natural gas source in your lab (for Bunsen burners, etc.)? You could try to fill your loop with that and see what you get. There's a bunch of junk (besides methane) in natural gas. It might give you a better idea of the quality of your column.

The other thing you could try is to open the valve of a butane lighter (don't strike the flint), stick the needle of a gas-tight syringe down into that, suck up enough of what's coming out of that to fill the loop and shoot that into your chromatograph. It's not just butane. It'll be a mixture of stuff similar to butane. A GS-CarbonPlot that is working well should give you some nice peak shapes for those mixtures.

[Gaston]

Hey rb6banjo, thanks for your suggestions. Unfortunately I don't have a natural gas source here, but I could try the lighter idea. In the meantime, though, I made a [very qualitative] 1:1:1:1 mixture of diethyl ether, acetone, pentane and hexane and injected it into the S/SL inlet in split mode. I suppose this wasn't too bad, as there are four decent peaks in the chromatograms.


Column: GS-CarbonPLOT 30 m x 320 um x 3 um
Carrier: Helium at 30 cm/s
Injection: 10 uL in S/SL inlet in split mode 1:200 at 250°C
Oven: 125°C for 2 min, 125-350°C at 20°C/min
Detectors: TCD at 150°C (ref. 40 mL/min, make-up 10 mL/min) and FID 300°C (make-up 40 mL/min)

I think there is an issue with the sampling valve, though. Sometimes, when the valve is on the OFF position, I get a inlet pressure shutdown. This doesn't happen when the valve is left on the ON position, but it doesn't always happen when it is OFF, so it's kind of a random occurrence. I thought that after tightening the nuts of the valve it was fixed, but apparently not. Any ideas of what could this be?

[rb6banjo]

I looked up the GS-CarbonPlot in an Agilent catalog and they say it's really for C1-C5 hydrocarbons. They don't call out much else. Ultimately, what will be your analytes? Overall though, I'd say your chromatography is looking pretty good and it seems that your system is working pretty well.

Can you tell if the "inlet pressure error" is because of high pressure or low pressure?

[Gaston]

I looked up the GS-CarbonPlot in an Agilent catalog and they say it's really for C1-C5 hydrocarbons. They don't call out much else. Ultimately, what will be your analytes? Overall though, I'd say your chromatography is looking pretty good and it seems that your system is working pretty well.

Yeah, that's what I had read as well. I found an Agilent presentation (http://bit.ly/14mPVTd) where they used it for solvents, though, so that's why I tried that out.

In the end, what I will be analyzing is the amount of CO and H2 (and possibly methane and ethylene, if present) in a CO2 stream. From what I've read, I think that this GS-CarbonPLOT column will not help me much with that. It seems that either a MolSieve 5A or Carboxen 1010 column would be better for that (I am leaning more towards the second option).

Can you tell if the "inlet pressure error" is because of high pressure or low pressure?

It's because of low pressure. Oddly enough, today I've been operating the GC with the valve in the 'off' position and I have not gotten any shutdowns. It kind of comes and goes.

[rb6banjo]

Intermittent problems are the worst ones. I'm trying to think of a way you can test the integrity of the rotor in your valve. Sometimes they can be scored. My experience has been that once scored and leaking, they leak all the time but I guess it could be intermittent.

[muGC]

I agree with rb6banjo. The issue might be scoring on the valve rotor, I've had this issue before. I would just rip out the entire valve and install a new one.

[Gaston]

Hey guys... I have done some more tests with the valve and the conclusion is this: with the valve in the ON position, there is no leak; with the valve in the OFF position, there are leaks UNLESS the valve is hot enough.

Here I did a test based on this Agilent YouTube video ( http://www.youtube.com/watch?v=fBy2JJbqV4A ). I set the inlet in splitless mode and the valve box temperature at 180°C with the valve in the OFF position, and then turned the valve box heating off at the start of the run. Then I recorded the total flow and the temperature of the valve box. According to the Agilent guy from the video, the total flow should not be much larger than 10 ml/min. If it is, there is a leak.



The orange line is the 10 ml/min line. You can see that when the valve is warm it works okay, but as it cools down, it starts leaking. When the valve is ON, then it works fine.

So... well, I can imagine that, with the heating, something in the valve expands and creates a seal which is not working when the valve is cool. Maybe I could try disassembling it? If the rotor is scored, maybe can only the rotor be replaced, or is that not worth it?

Again, many thanks for your help!

[rb6banjo]

Good detective work. Honestly, I've never replaced a valve rotor before. Thankfully/luckily, I've always been in the situation where we could just replace the valve. Maybe someone else can offer up some suggestions?

[Gaston]

Well, I removed the valve from the valve box and took it apart while being reasonably careful. Two things drew my attention:

- Lots of black powder (burnt isolating material from the valve box, perhaps?)
- The rotor looked OK (no scoring as far as I could tell) but one of the channels seems to have a red hue (as if it were a thin layer of rust or something).

But anyways, I cleaned the valve and the rotor with acetone and dried it with a compressed air gun and then put it back together again. Afterwards, I repeated the same test from my previous post (starting from 150°C, though).

And the result is...



...no leaks in either the ON or OFF positions, at least so far (after some three days of use).

So I am hopeful that the valve can stay, but if the problem reappears, then I think I would go on with buying a new one.

Is there a way to know from the outside what is the volume of the loop that is installed? When I turned on the GC it was configured as 1 mL, but I'd like to have confirmation of that. Sample loops are not too expensive, though, right?

Well, I think that this pretty much brings the 'testing phase' to an end. Now comes the part of selecting the column and configuring the GC for its actual intended application (which is a whole new topic) Thanks a lot for the tips and help so far!

[rb6banjo]

Good deal. You can make loops using your own tubing. If you're looking for analytes that might adsorb to surfaces, I'd recommend using treated tubing (Silco-steel, etc.). Just treat it like an empty cylinder and use V = pi x r^2 x h (r = the inner radius and h = length of the tube) to calculate the volume.

You're in business!