Pegasus 4D GCxGC TOF-MS

[WK]

Dear All,
I am interested in this instrument for analysis of fragrance materials.
Does anyone have this system?
What is your field of work?
Does it perform well?
Thanks in advance for any answers.
WK

[Peter Apps]

Hi WK

We have a Pegasus 4D. I am only peripherally involved in running it. The people who do use it can generate a bewildering array of peaks from any natural mixture - and the operative word is bewildering. For some applications (see the Leco website) the comprehensive GCxGC separation and rapid TOFMS approach is very powerful, in others it just generates more data but no more information. Unless you depend on the data processing software to find, identify and integrate the peaks that you are interested in, you will spend longer on data processing that on running the samples, to the extent that data processing becomes the bottleneck. The MS deconvolutions are very good at pulling target peaks out of the background, which is still necessary with complex mixtures even after GCxGC, but you have to have a target peak, which means that it is no good for finding unknown compounds that might be causing off-odours for e.g. Because the two separations in GCxGC are both dominated by vapour pressure they are not really orthoganol, and the second dimension column is very short and running way above its optimum flow rate. For some compounds this can mean mean a disappointingly weak enhancement of separation compared to one-dimensional GC. You still need to put a substantial effort into optimising column selection and conditions.

You really need to define your requirements very carefully and make up some samples that will test the instrument's performance in those specific areas - then give them to LECO to do a demo with. Please post the results.

Bear in mind that you use a lot of air, liquid nitrogen and nitrogen when you do GCxGC.

Peter

[WK]

Peter Apps,
Thanks for the reply.
Most of the applications here would be targets.
I was very interested in LECOs orange and lemon oil pesticide application which is fairly recent.
Please could you tell me what the air and nitrogen are used for?
I know the liquid nitrogen is for the modulator.
I did go for a demo 3years ago. Roughly how much of each gas is used?
We use liquid withdrawal CO2 cylinders here - I wonder why the 4D needs a special reservoir unit.
Roughly how large are the rawdata files generated? Do you have the 2 monitors?
WK

[Peter Apps]

Hi WK

The air and nitrogen are used for the modulator hot and cold jets respectively. Running GCxGC all day you can get through a cylinder of each in three to four days - it is difficult to be exact because we run other GCs from the same cylinders. If you have any FIDs in the lab you need to run their air and nitrogen from a separate regulator from the modulator or else you get baseline deviations due to the modulator switching the flows on and off.

I do not know if the Leco modulator can work with liquid CO2, certainly the set up we have can only use liquid nitrogen. The liquid nitrogen is used to chill the nitrogen from a cylinder that blows into the modulator.

The raw data files are multiple megabytes, even with the top end computers that come with the instrument, number crunching takes several tens of minutes, and if you set a peak detection threshhold too low it can go on for hours.

Target pesticides in a GC compatible matrix like an essential oil is one area where GCxGC TOFMS does have advantages - Leco would not have published it otherwise of course. As with all manufacturer's application notes you need to be aware that the notes are deigned to sell instruments - check out the concentrations of analytes that produced the peaks that you can see e.g. fig 6 is with 1000 ppb, while the repeatability data in Table 1 is at 25 and 50 ppb, when the target peak is most likely invisible to the human eye and has to be dug out from the background by the deconvolution algorithm. I am an old fashioned kind of chromatographer - I like to see the peaks myself !.

If you are interested in pesticides at low ppb, spike some oils at the appropriate level and see what Leco can do with them in terms of detection, quantitation and repeatability - that is really the only way to know if the machine can do the job that you want it to. You could also give the same samples to other vendors to see what their instruments make of it - Agilent also has deconvolution, but not GCxGC as far as I know, maybe accurate mass SIM would pull the pesticides out of the terpene background. A heartcutting approach with conventional MS might also work, you then maximise the GC separation rather than the MS selectivity. There are several ways to skin this particular cat.


Peter

[WK]

Peter,
Thank you for the postings - very interesting information.
I agree with the plan. We did a similar exercise with another single column GC-MS instrument supplier last year.GCxGC is the way to go I believe for lists of targets in flavours,fragrances and essential oils and the TOF/software is very powerful too.
In this respect you certainly get what you pay for.
WK


PS: is your Pegasus capable of a long unattended run?

[Peter Apps]

Hi WK

As far as I konw the Pegasus does long GC runs without any problem, - apart from the huge data files.

For both GCxGC and deconvolution you have to be looking for molecules that differ in their chromatographic behaviour on at least one stationary phase, and which have clearly different mass spectra. This means that they are not very good at separating molecules with small differences in structure; they will probably work for pesticides in essential oils, but will struggle with trying to detect individual terpenes against a background of other terpenes. I doubt that there is a plug and play system that can do this - each analysis has to be optimised for the particular sample - the closest that you would come would be heartcutting with two long capillary columns. If you really want to separate things run GCxGC off the end of the second dimension column of a classical heart cutting set-up !. Maybe I'll try it myself one day.

Cheers

Peter