GC Method Development for Low BP Compounds

[GoinBoardin]

Hello, first time posting, I'm very grateful for any help.

I'm trying to quantitate isoprene with a GC/MS, but have been hung up with a tailing issue (so the GC board seemed most appropriate).

I'm using 2-methyl-1-butene as an internal standard, with a boiling point of 31*C.
The boiling point of isoprene is 34*C, but I'm up at 7200' and so the boiling points are further depressed. I don't have any means to cool my GC oven other than fans, and as a result the absolute lowest oven temperature I have achieved is 27*C (and that took hours, 28*C is more reasonable).
My system is working as well as ever: leak checks passed, good spectra of caffeine, good K value (1.67 on a 30m column). I suspect my tailing problem is arising from too warm of a initial oven temperature because all else seems good.

I'm using toluene as a solvent. The internal standard elutes just before isoprene, but both compounds tail, with a time range from about 1.30min to 2.1min. The injection is 1uL of 50ppm isoprene & 50ppm int. std. (about 50ng of each) on a 30m by 0.25mm ID column with 0.25um film thickness, so over loading shouldn't be an issue.

Are there some situations where tailing simply cannot be stopped with the equipment at hand? Is it ever acceptable to do quantitation when tailing is present?

Let me know if I should clarify anything.
Thanks

[Peter Apps]

The stationary phase film in the column is too thin for these low MW compounds without cryogenics in the oven. If you are stuck with the hardware that you have then you need to look at injection conditions - in particular are you doing a split or splitless injection ?, and if splitless for how long is the split closed ?. What is the stationary phase in the column ?

Have you checked that the carrier gas flows and pressures are optimum ?

Checking the spectrum of caffeine is unusual in GC-MS - and it will not help troubleshoot peak tailing.

Peter

[Johnny Rod]

I'm not sure what stationary phase you're using but it sounds like you could use a PLOT column to better effect than a liquid one. There are various ones around from different suppliers. For GCMS you should look for oen which is better bonded and you will need a particle trap on the end (a few metres of DB1, but use a good quality union, not a glass press-fit, as you are using MS). We recently got a Restek Rt-Q-BOND which hasn't given any problems with shedding particles into the MS.

[GoinBoardin]

I have a Phenomenex ZB-5 column in use (5% phenyl), but do have a couple column options. I have a 60mx0.25mmx1.0um SPB-5 column (5% phenyl) from Supelco and a Petrocol DH 50.2 which is 50mx0.2mmx0.5um (also Supelco, 100% dimethylsiloxane). I'm suspecting I should use the Petrocol?

I'm doing splitless injection with a split time of 2.0 minutes (I shut down the detector at 2 minutes as well). Injector temperature is 220*C with the transfer line to the MS at 250*C.

I've been trying to optimize this for a little while. Have used 0.8, 1.0, 1.3, and 1.5 mL/min flow rates with helium carrier gas, some differences but nothing substantial there, except 0.8 looks to increase tailing. Tried many different temperature programs but after further reading realized I should go as low as I can.

Thanks for the help!

[chromatographer1]

You have a packed column FID application which you are trying to do with a capillary column and a mass spec.

Think of trying to fly on a hot air balloon and all you have is a space shuttle. Newer is better but not always more suitable.

A porous polymer, any kind, PLOT column is the best recommendation in your situation.

A thick film (3-5 micron) 100% or 95% methyl silicone would be my second choice.

A 2m 2.1mm ID 25% DC-200 Chromosorb P NAW 80/100 with a FID would be the best choice, but sadly, very inappropriate for your instrument and detector.

best wishes,

Rod

[Peter Apps]

The first thing to try is to reduce the split time to 30 s - the tailing you see is due to slow transfer form the inlet.

Peter

[GoinBoardin]

Thanks for the feedback everyone. I understand my instrumentation is not ideal for this analysis, but it is a rather short side project. If it were something I would do regularly I would certainly look into GC-FID with the PLOT column, or the MS compatible PLOT setup.

I tried the following splitless times: 1.0min, 30sec, 18sec, and 6sec. All with 1.5mL/min flow rates.

Certainly does seem like the compounds are just "hanging out" in the injector and not entering the column in a short plug.

With 6seconds splitless time the magnitude of tailing is reduced and the duration is slightly reduced, but I still have both compounds eluting from ~1.3minutes to ~2.0minutes (with some separation of the initial peaks). I did run isoprene alone and it tails.

[chromatographer1]

Adjacent double bonds and bare metal do not mix well, as you have discovered. That is the problem with MS, the entire flow path is not inert.

About the only hope is an expensive one, a long thck film column, 60m 0.32mm ID 3 micron methyl silicone. But even there, higher pressure is required to get the C5s to focus their plugs on the column and when you are depressurizing one end of your column at high vacuum, just how much pressure can you get at the head of the column and keep your flow at a low rate?

Your analysis may not be possible with the hardware you have. Perhaps the boss will OK sending the samples out. I am certain there are many labs who could do this standing on their heads, and do it quickly and well.

best wishes,

Rod

[GoinBoardin]

I appreciate the explanation. Makes sense, I've seen the concerns of ss interactions mentioned in a few MS papers. I'll have to explore other options.

[Peter Apps]

One more thing to check - is the inlet liner clean (and what type is it ?) ?

Peter

[GoinBoardin]

The liner is a Thermo Scientific, 105mm long, 3mm ID splitless liner. It has been previously used but I cleaned it thoroughly with NoChromix and deactivated it with dichlorodimethylsilane (5% in toluene)...I'm in an academic research lab where time is more abundant than funds thus the recycling of liners. I treated several liners in the same batch and have not had issues with previous analysis, so I don't think the problem stems from liner treatment.

I replaced the liner with a freshly cleaned & deactivated one, with a new septa. The liner I removed was dirty (not filthy but dirty), so maybe I just had a total rookie move and that was the problem all along (I am a rookie). I'm re-running samples now and will know soon. When I started on this the liner was clean, I've just done many many injections trying out all sorts of variations on the method, and becoming more familiar with the instrument.

Yesterday I reran isoprene but with reduced injector temperature: 150C & 130C to see if it had any effect on the likely ss-solute interaction but I observed no difference in tailing. Sort of a "hey why not try" type idea.

[chromatographer1]

There is no GC with a TCD or FID anywhere available?

$50,000 for a mass spec and not $5000 for a TCD !

Oh well, play the cards as they are dealt to you.

Rod

[GoinBoardin]

I have a 1970's model GC with a TCD sitting behind me, ss column & the entire thing fits in a metal suitecase to run on 120V. The lab I'm in is the only separations analytical lab on campus, with the focus more on MS. I agree it is odd we don't have either of the cheaper instruments, not even for the undergraduate teaching labs. I have access to a MALDI, two LCQs, the GC w/ TSQ that I'm using now, and a GC w/ DSQ but no TCD or FID.

With the fresh liner, leak tests passed, column eval looked good, there is still tailing. The ss-isoprene interaction seems to be the limiting factor in this separation.

Thanks for the help everyone. I look forward to reading more of your posts and learning as much as I can! Hopefully I can answer others' questions later on once I know a thing or two.

[CE Instruments]

Doing a quick read through of all the suggestions I can add the following.

Try doing a pulsed splitless injection with the splitless time at 30 seconds.
You do not state the Split flow rate that you are using to purge the injector after the splitless time. Try setting this to 100ml/min and gas save after say 5 minutes. The Thermo injector is very good , your compound will not be interacting with any metal in the injector.
Autosampler injection ? Can you draw 2ul of air into the syringe after the sample ? If so set a delay before injection of 2 seconds prior to a fast injection of the sample into the injector and remove the syringe without delay. (hot empty needle technique)
Injection depth ? This should be as low as possible, with some Thermo samplers you can use a 7cm needle for greater depth.

As stated your column has too thin a film , you will get some tailing. Provided it is reproducable there is no reason you should not use it for quantitation. Note the splitless time will define the minimum peak width, any sharpening is thanks to the column. I was not surprised to see 2 minute peak widths with early eluting peaks when you had a two minute splitless time.

[GoinBoardin]

My internal standard lacks the adjacent carbon carbon double bond and still tails. I realized the significance of this last Friday. My advisor (I'm a grad student) says that it looks like an injection issue, which is what many of you have also suggested and what my texts point to. I will try your suggestions. If I'm still having trouble I intend to switch to the 60m DB5 with 1um film. My 30m column may be in need of trimming anyway (could be degraded/active near injector and transfer line), and if the film is too thin it seems to make sense to switch out.

I was using 50mL/min split flow without gas saver. The tailing is reproducible, but hurts LOD and my samples will be low level. That will be next, sampling.

Seems like there are probably a few causes of the problem left while I started with several others. I'll keep chiseling away at it.
Thanks again.