Sudden decline in resolution in Headspace test

[RoseCity]

We have a Shimadzu GCMS-QP2010 with ZB-5MS column (30 x .25 x .5), which I began using for headspace tests that were formerly on an SRI instrument. When I first started using the method, I was very pleased to see its high resolution. Argon, butane, and its isomers all separated into nice peaks. This worked for a few weeks then suddenly the resolution took a dive and everything started coming out in one large tailing peak. I included pictures for what it used to look like and now looks like. I don't think anything was changed; it just started suddenly.

http://postimg.org/image/64rwrjwl3/
http://postimg.org/image/tuhcg8uyf/

Interestingly, only the early peaks were affected. In this sample, D-Limonene still has fairly good peak shape:

http://postimg.org/image/3yfp3di8r/

The method is as follows:

With the exception that I am doing manual injection because of not having the headspace autosampler.

Things I have tried so far:

Changing liner/septum
Increasing split ratio from 25 up to 300 (no change)
Lowering temperature and flow rate (this separates the peaks somewhat, but does nothing to reduce their tailing)
Reinstalling column
Cleaning split vent, exchanging split vent filter

[Morrison]

I'm sorry if I missed something in your reasoning but from the image you posted it seems that the injector temperature is 340 °C and the maximum oven temperature is 180°C. It is so? This would be a wrong practice that over time may create some of the problems you say. It is always advisable to maintain the injector temperature below the maximum oven temperature.

When you reinstalled the column have you cut some piece of the column itself or not? In your case a cut of about 20 cm at the injector side could be decisive. Try this together with a changing in temperatures according to the above suggested.



Hope this helps. Regards.

[RoseCity]

I had put troubleshooting this on hold for a while, because we got a new column in the meantime. I am now booting up the new column and I see that nothing seems to have changed. Also a new purge filter was installed which seems to have fixed the issues we were having with flow control. Regardless, this same loss of resolution is still here. I did change the injection temperature like you suggested, now 160C.

New column: 30x.25x.25 Rxi-5ms, with 5mx.25 guard column.

[Peter Apps]

Can you clarify what you are doing ?; you say headspace but the method that you post has syringe solvent washes and looks as if it is for a 10 ul syringe. It has no sample temperature or equilibration times. This is not typical for headspace.

How do you know that the lump on the bad chromatogram contains the peaks that were separated before ? It could be from something else entirely, in which case you are looking at a loss of sensitivity and a source of contamination, not for a loss of resolution.

Did the drop in performance coincide with starting to use manual injections ?

Peter

[RoseCity]

The method has that part because it was copied off a method for liquid injections, but there was never a time that we weren't doing manual injection for this method. So the manual injections worked well at first but are now failing.

The reason I know the peaks are spread, is that the ion spectra are the same as before. If I look at the left part of the blob, it is ion 40 = argon, then in the middle portion is matches for butane.

As far as contamination goes, I thought that was the cause as well, but I just changed the whole column and liner so I don't know what could be contaminated. Now the first step after installing the new column, according to the instructions, is to inject an unretained compound and check for symmetrical peak shape. So I can't even get past that part right now due to this issue.

[Peter Apps]

"The method has that part because it was copied off a method for liquid injections, but there was never a time that we weren't doing manual injection for this method. So the manual injections worked well at first but are now failing."

It would be really, really helpful if you told us what you are doing instead of what you are not doing. Sample temperature, equilibration time, sample matrix, injection volume, injection time etc.

What unretained substance are you injecting, and what is its retnetion time ?

Peter

[RoseCity]

A 40 mL vile sealed with a septum had 0.02g of liquid butane added to it, then was capped with the septum and equilibrated 30 min at room temperature until all the butane had evaporated. 5 uL of the gas phase was then taken into a syringe and injected in the GC. Butane peak appears at ~1.4min on the new column (flow rate 1.5 mL/min). Unretained compound is argon, which appears slightly before.

[Peter Apps]

A 40 mL vile sealed with a septum had 0.02g of liquid butane added to it, then was capped with the septum and equilibrated 30 min at room temperature until all the butane had evaporated. 5 uL of the gas phase was then taken into a syringe and injected in the GC. Butane peak appears at ~1.4min on the new column (flow rate 1.5 mL/min). Unretained compound is argon, which appears slightly before.

If the retention of butane is now 1.4 min it has increased compared to what it was on the old column - what has changed that might account for that, is it just the addition of the guard column ?, did you increase the inlet pressure to compensate for the extra length ?

Is the butane peak now sharp or broad - -you mentioned lack of resolution with the new column, but unresolved peaks can be sharp. If broad, how broad is it at baseline ?

I am not familiar with the Shimadzu inlets, but on at east three makes of instrument that I can think of it is possible for bits of septum or graphite ferrule to fall to the bottom of the inlet, and remain there when you change the liner. Graphite bits especially will give broad early eluting peaks. Another possibility is that bits of ferrule lodge din the column when you installed it - did you trim the column end after you threaded it through the ferrules ?

Peter

[RoseCity]

The change in retention time was likely due to both the guard column, and the fact that the old column had been trimmed many times, so the new guard plus column, is a good deal longer then the old one.

The butane peak looks like the ones I posted earlier - broad and tailing. Base covers about 0.3 min.

i did cut the column after putting it through the ferrule which was good thinking. As far as something stuck in the inlet, this could be the case. I did take apart the inlet recently but I didn't look for anything there. But the inlet liner has glass wool in it so a piece of septum would be unlikely to make it to the bottom.

[Peter Apps]

Your split line may be blocked - so you are actually doing splitless injections. This is difficult to troubleshoot with back-pressure regulated inlets. You mentioned earlier that increasing the split to 300:1 made no difference -was this no difference in peak width or no difference to peak area ?

If you deliberatley do a splitless injection, does it look different in width or area to a split injection ?

Peter

[RoseCity]

I think you may have hit the nail on the head with this suggestion. I have often wondered whether the split option is malfunctioning, because the peak area is not affected by changing the split, outside of normal variance. I have tried cleaning the split line but maybe i'm not doing it effectively. Do you have any advice on this?

[Peter Apps]

I think you may have hit the nail on the head with this suggestion. I have often wondered whether the split option is malfunctioning, because the peak area is not affected by changing the split, outside of normal variance. I have tried cleaning the split line but maybe i'm not doing it effectively. Do you have any advice on this?

If area does not change when you change the split setting then the line is definitely blocked.

The last Shimadzu GC I worked on was a 9A in the early 1990s, so I have no idea how your plumbing looks.

Depending what you can disconnect from where the two approaches I would try would be to remove the filter and heat up the tube leading from the inlet to the filter with a hot air gun to try to melt the gunk, then rinse with a series of solvents, or just try forcing solvent through. Best to remove the whole inlet from the GC first of course.

[RoseCity]

OK So I looked into this a little more to follow up on your suggestion. Using an independent flow meter, I found that the split vent does have flow, and correctly sets the flow according to the split ratio. So it is not just plugged, but nonetheless, the sample does not seem to split with the split ratio.

Also, a similar sort of effect seems to be occurring with liquid injections test, for example this run of terpenes. The most early eluting peaks have bad peak shape, either fronting or just coming out as blobs. Over the course of the chromatogram, the peak shape gradually improves. The split ratio also has not affected the shape or the area. Here's the method and image. This time it was actually with the autosampler.

http://s7.postimg.org/e20qofzq3/5_PPM_Terpenes_02121603_DS_2_Page_1.jpg
http://s18.postimg.org/3ls0pkpd5/Terpenes021516_Split5_Page_1.jpg

So it seems like there is an issue with the splitting, but its not a plugged line. And seems to hit early eluting compounds the hardest. I tried a second brand of split liner but there was no difference.

[Peter Apps]

OK So I looked into this a little more to follow up on your suggestion. Using an independent flow meter, I found that the split vent does have flow, and correctly sets the flow according to the split ratio. So it is not just plugged, but nonetheless, the sample does not seem to split with the split ratio.

Also, a similar sort of effect seems to be occurring with liquid injections test, for example this run of terpenes. The most early eluting peaks have bad peak shape, either fronting or just coming out as blobs. Over the course of the chromatogram, the peak shape gradually improves. The split ratio also has not affected the shape or the area. Here's the method and image. This time it was actually with the autosampler.

http://s7.postimg.org/e20qofzq3/5_PPM_Terpenes_02121603_DS_2_Page_1.jpg
http://s18.postimg.org/3ls0pkpd5/Terpenes021516_Split5_Page_1.jpg

So it seems like there is an issue with the splitting, but its not a plugged line. And seems to hit early eluting compounds the hardest. I tried a second brand of split liner but there was no difference.

If your Shimadzu is similar to other back-pressure regulated inlets then the change from split to splitless is accomplished by re-routing gas flows in the plumbing - the total gas flow and the flow out of the vent stays the same, so measuring split vent flow does not necessarily diagnose blockages. If there is a blockage it would have the same effect on liquid as on gas samples - the inlet converts liquids to gasses by flash vaporization anyway.

Find a schematic of the inlet and gas control plumbing, if there is a blockage it will be between the exit of the split line form the inlet body and the two-way valve that re-routs the gas flow.

It is also possible that the two-way valve is stuck, or is not getting the proper signals from the electronics.

Peter

[RoseCity]

I figured out what the issue was today. Purge line was attached to split vent and vice versa during an earlier maintenance. Thus why the flow was setting right but not actually carrying away sample. Thanks for the help!