Strange FPD behavior with H2S, please help

[thiotimoline]

Hello, I looked around the forums and did a few searches for this but I couldn't find anything, so I'm sorry if this is something that's been answered before, but it's been giving me an incredibly hard time, not to mention a splitting headache. I'm also sorry if there's any problem with starting a topic to ask a question with your first post, but I didn't see any rules against it, so here I go.

Alright, so I am using an Agilent 7890A GC equipped with an FPD to run a standard 25ppm sample of hydrogen sulfide in helium. The column I'm using is a GS Gas-pro 30m by .32mm column by J&W Scientific. The carrier gas is helium flowing at 8ml/min, though I've changed that rate several times and it hasn't helped me. My system has been checked thoroughly for leaks, so that isn't the problem. Let me know if you need any more information.

I've been running samples of the H2S in He in an attempt to calibrate the machine, but I haven't even been able to develop a method that gives me good data. I've been using several different methods to measure the concentration, but no matter what I do to the method, the peaks of my runs do not remain constant as I run identical samples with identical methods, instead, they continuously increase.

They (their area, but also height) increase at a decreasing rate, and I think that they may level out after what ends up being pretty much an entire day of doing runs (at least 15, probably closer to 30), though usually I get fed up and try changing the method before that. As you might imagine, running identical tests for an entire day is, well, undesirable. Typically the sample gas will be flowing around 10sccm, though I've tried several different flow rates, all with the same results.

Here's a realistic but fabricated example of the data I am getting.
Run 1: 2000pAs
Run 2: 4000pAs
Run 3: 5500pAs
Run 4: 6750pAs
... after maybe 15 more runs it might (though not necessarily will) stabilize around
Run 19: 1.8E4 pAs
Run 20: 1.79E4pAs
etc.

Typically, my method involves a ramp from a low, say 35C, up to a high temperature, say 200 or 250 C, where it remains for a few minutes in an effort to somewhat bake out the column. Lately I have been trying doing a ramp from 50 to 80C at about 6 degrees/min, with the same results.. I've tried changing the load and inject time, reducing it from .5min to .2min, with strange results. I observed two peaks, right next to each other. I'm unsure if that is relevant to my problem, though I'm really at my wit's end, so I figure I should give as much information as possible.

Initially, I thought that the increasing peaks might be due to sulfur compounds (or really, anything) remaining in the column between runs, and being picked up in the next run. To test this theory, I ran pure helium through the system, using the same method. No peak, so I feel that that's essentially ruled out.

The methods I've used are from literature on the subject, literature that doesn't mention anything along the lines of the problems I've been facing.

Really, I don't care if you have a bizarre suggestion, please, anything at this point is appreciated. If it's my machine, then fine, it can be fixed, if its a foolish error on my part, then that too can be fixed. I am going insane here, and nothing I have read has lead me in the right direction at all. Again, I'll answer any questions about my setup that you have.

[GasMan]

When you ordered your 7890, did you also order it with sulfinert treated lines and inlet? What you are describing is typical of H2S being adsorbed onto active sites on the metal tubing. As you keep on injecting H2S, the number of active sites gets less until all are "filled up", which is when you will start to get reproducible results. If you leave your GC for some time without injecting any samples with H2S, the H2S that has been adsorbed, will slowly desorb and you are back to square one.

Gasman

[thiotimoline]

Tragically, it looks like it was not ordered with sulfinert treated lines or inlet. Thanks so much for your help.

Now to find out if this is something that can be replaced...

[GasMan]

If you can give me the serial number of your 7890, I can probably help you with what parts you will need to replace.

Gasman

[thiotimoline]

Hrm. I mean, I don't know if I think I want to post the serial number of my machine on a public forum... And while you have been helpful and seem trustworthy, at the end of the day, I don't really know who you are nor what giving out the serial number of my machine could do... If I call up Agilent, they'll be able to tell me what needs replacing, right?

[GasMan]

I can understand your concerns. Please send the information to my email address. Although this is my private email address, I will reply from my company email address.

Gasman

[thiotimoline]

Err, your email isn't given on the forums, or at least I can't see it.

I still don't think I want to give out the SN of the machine. I really don't know what could happen if I did, but for most things it's not a good idea.

[GasMan]

Some things have changed with the new forum application that is now being used. My email address was available in the old format. The email address is gasman41@gmx.de. Just contact me, without the serial number, and I will reply with my company address. Then you can make the decision if you wish to go ahead.

Gasman

[Peter Apps]

Having sulfinert gas lines will only make a difference if you are using some kind of valve-based injection (which reading between the lines probably is the case). If you are using a syringe you need a sulfinert needle.

Peter

[chromatographer1]

And while you may wish to flow your std gas mix several seconds through your sampling system (valve) don't forget to pass the waste gas either into a hood or into a potassium hydroxide solution, allowing the gas to bubble through it before exiting into the hood.

Rod

[JI2002]

When I started running GC/FPD for H2S, I experienced the same problem. I tried a couple of things that worked for me:
1) Cut off 20 cm of column (injector end) at the beginning of the day.
2) Inject 1 PPMV STD ten times consecutively in one run.

Hope it helps.

[AICMM]

JI2002, not meant disrespectfully, I have always been bothered by the act of pickling a system. It has always been an analytical question mark for me, even when I was doing it way back when in the environmental business analyzing phenols. If you pickle your system, when does it wear off? Using the example of phenols above, EPA requires GC analysis of phenols to pass initial CCV (which you may have pickled) AND the end CCV which you probably did not. Lots of times you could pass the front but not the back and your time was wasted. (Never mind that the GC/MS people don't have to pass a back standard, which, in my humble opinion, is one of the reasons that GC/MS has taken over much more of that analytical niche....)

When possible, I think it is always better to figure out some way to keep the system in stasis. In the example of phenols above, I always thought the EPA method should use derivatization (yech!) to make the phenols much less reactive and more amenable to chromatography. In the case of H2S, steps like sulfinert are essential.

Having said all that, I am still a realist. Lest you think this is all hypothetical, I am currently faced with the same question with ammonia in air analysis. System reactivity is certainly an issue here.

Best regards,

AICMM

[punt]

Hi thiotimoline,

I have faced with the same problem like you when running a low ppm H2S calibration gas in my GC system. Even though the sampling Loop and the tubing from my calibration gas to the GC inlet has been installed with sulfinert material, it still does not helps. Perhaps somewhere in the system is still absorbing H2S.,

But when i flush the line and sampling Loop and inject with a higher H2S concentration prior to injection for my calibration standard, the area counts for my calibration standard get more consistent.

I have test it with a H2S CRM injection right after the calibration, it manage to produce a repeatable and accurary result. I am not sure if it is a correct way but hope it helps.

Regards

[JI2002]

I am currently faced with the same question with ammonia in air analysis. System reactivity is certainly an issue here.

Testing ammonia in air by GC is very difficult if not impossible. Another approach is to collect the sample in sulfuric acid solution and to analyze the NH4(+) by IC.

[chromatographer1]

JI2002

You are correct about NH3 but it is possible.

Requires a PID and special treatment of metal columns (no silica or silicones are allowed) and a porous polymer packing.

But it has been done.

Just a FYI for those who might wish to buy a dedicated system for that analysis.

best wishes,

Rod