GC Method Development for Low BP Compounds

[chromatographer1]

By all means, lower your injection time to 10s then 5s, then even less and see if that makes a big difference. A minute of two of diagnostic tests can be very informative.

At least you may discover if it is the whole injection timing issue, or if it is reactive surfaces in your injector or detector, or BOTH.

I once was able to focus the injection plug of C3-C5s on a thick film megabore capillary column by connecting a 10 meter 0.25mm ID piece of tubing behind it in front of the detector. A thin film narrow bore capillary column would work similarly.

good luck,

Rod

[GoinBoardin]

Is pulsed splitless equivalent to surge splitless? If not, I'm not sure I have the option for pulsed. I'm using a Thermo Trace GC with an old version of Xcalibur (1.2).

I have been pulling up 1uL of air after sample, and yes I have an auto sampler. Did a run with the 2 second injection delay, no post injection delay, increased split flow to 100mL/min, 6 second splitless time, gas saver at 5minutes, and trimmed my column on injector side. This clipped ~3 seconds off the tailing time.

I found an old, but still sealed on the ends J&W column: 15m x 0.53mm x 1.5um that is 5% phenyl. I'm thinking I should swap in this column and see if there is an improvement.

It seems to me the 6 second splitless time with 100mL/min split flow does suggest the problem primarily is not with injection timing. I hope not, but suspect the ion volume could be a source of activity with the compounds.

[CE Instruments]

Should be little difference between 50ml/min and 100 ml/min split flow , both will clean out the injector fast enough. Is the flow in fact to fast ? The principle is that the explosive decompression of the injector on opening the split vent purges the injector and sucks the solvent out of the top of the column leaving compounds that are more soluble in the stationary phase on the column. You will be refocusing the Toluene but will your compounds of interest be more in the stationary phase or in the solvent ?

Yes surge splitless will be the same.

Try an experiment with both running standard split injection with a low split say 5:1 or 10:1 depending on what the GC can manage. Run the same split flow with splitless analysis.

Do you get larger peaks of interest running in split or splitless mode. Is there much difference ? Any difference in the tailing ?


The short thicker film column should help but running with an MS you will need to look at using a narrow bore restrictor in the transfer line of your MS otherwise you will be sucking the gas out of a large proportion of the column and your MS may struggle to handle the gas flow. The 60m 1um film one may be better.
Any chance you can take a jpeg of the chromatogram and get it hosted so you can post up on here ?

[GoinBoardin]

I ran hexanes and pentane for a troubleshooting exercise. Both tailed just as the other low BP compounds did before. After looking more closely at a GC troubleshooting poster on the wall in lab I realized it is more accurate to say my chromatograms have "chair" shaped peaks.

Larger ID liner had little to no effect.
Split injection results in satisfactory peak shapes. Reduced signal intensity but very minimal tail/chair.
I also ran pentane with a 10:1 split ratio that doesn't look very good, though it is an improvement over splitless injection with a splitless time of 30seconds & split ratio of 10:1 (both at 1.2mL/min).

I think CE Instruments was getting at this in the last post: seems what happens is the solvent condenses on the front of the column, and some of the analyte is re-dissolved into the toluene (which is acting as a pseudo stationary phase). A good portion of the sample elutes in a nice sharp peak but the "chair" shape comes from the slow evaporation of the low BP compounds out of the toluene that is still condensed on the front of the column as carrier gas passes over. Basically the same idea of refocusing by solvent effect, but with the BP of solvent higher than that of analyte, things do not work out well.

The PhotoBucket link below has six chromatograms, each labeled in their respective descriptions.

http://s1367.photobucket.com/user/Chemi ... sort=3&o=0

[Peter Apps]

From the chromatograms a split injection should provide the resolution that you need, and as long as the peaks are large enough the problem is solved.

Going from splitless to split has two main effects; the major influence is that the inlet is purged more quickly in split mode so there is no residual vapour in the inlet being progressively diluted by incoming carrier gas and transferred to the column at exponentially declining concentrations that produce a tail on the peaks. Also less sample is transferred to the column which can mean that there is insufficient solvent vapour to form a secondary film, or to cause significant swelling of the stationary phase. If you do get condensed solvent there are nearly always undesired effects on peak shape unless conditions are carefully optimised to exploit solvent effects that sharpen peaks.

Peter

[CE Instruments]

I've had a look at the chromatograms posted and they are tiny ! They look like thumbnails rather than originals , did you upload the right files ? I cannot see the details of the scales but the peaks look awful. The best are the 60:1 split. I think you are massively overloading the system and you are looking at the TIC traces. You need to be looking at selected ion chromatograms for each peak which should then give inteference free peaks if running MS. If this is FID data from Xcalibur what sampling rate are you using as there should never be steps on FID data. 10Hz should give acceptable results for your peaks.

[GoinBoardin]

Peter Apps, I think you're right. A paper I found from Health Canada used split injection and reports LOQ that is acceptable for my work. They used a 60m x 0.32mm x 1.0um column with the same stationary phase I'm using, otherwise basically an equivalent setup.

Overloading is likely not the issue. I've diluted samples down until the S:N~=2 and still see the MS peaks for analytes over a long time window. What I posted was a TIC chromatogram (scan from 50-100m/z). I will be using SIM to reduce LOQ in the actual analysis.

I'm not sure what the issue is with the posted chromatograms. Click on the thumbnail, then the little zoom button on the bottom right of the larger photo, then the zoom button again and they should be plenty large enough to judge peak shape & resolution. Is there a better way?

Thanks for the input.

[GoinBoardin]

Thought I ought to update this. I ended up switching to a different GC/MS instrument that I configured for a headspace auto sampler. This worked out much, much better. No more tailing, with increased sensitivity, and it drastically simplified the sample preparation steps, resulting in a lower method LOD.

I'm pretty well convinced that "CE Instruments" was on the right track; toluene was condensing on the front of the column and acting as a pseudo-stationary phase, retaining some of the isoprene & internal standard that would slowly vaporize as the carrier gas was flowing over the condensed toluene. Almost the opposite of solvent focusing. Eliminating the solvent eliminated the problems while keeping the same column.

Hopefully that update can be searched by someone in the future to help them.

[Peter Apps]

Thanks for the feedback.

Peter