Chromatography Forum

Hello,

our company is thinking about buying a GC (mainly, but not entirely, for solvent purity determination). I dont have much previous experience with GCs, but from what I know, I think the best detector for such application is FID. With regard to the rest of applications we are considering also MS. I have tried to figure out what are specifications of a GC, so it can be equipped with both FID and MS, however I have found no specific parameters.

Thanks.

Juraj

There are various splitter kits for FID and MS, look up SGE or the instrument vendors. There are simple tee pieces and also Deans splitter type of things.

Why have simultaneous FID and MSD? The MSD is more sensitive and gives compound identification information. I'd pick an FID to save money, and have easier operation and maintenance if you are quantitating a known compound over and over again on a routine basis. If you are needing to indentify unique impurities or need higher sensitivity I'd pick the MSD.

The FID is a great detector and can give LODs for many compounds below that which you can obtain with an MSD. However, the FID is blind to some compounds as well. So, if you are looking at chromatograms for specific target compounds that contain carbon (and are not one of the compounds that is nondetectable by FID), then FID may be the perfect option. If you are looking at unknowns or potential unknowns (ID confirmation required) then a MSD can be very helpful. The MSD is the nice toy in the toy box - but it has a cost for purchase, requires more time consuming maintenance that the FID and (unlike the machines on CSI) does not necessarily lead to definitive identifications of unknown compounds.

If you have a requirement for an instrument that will have an MSD, consider whether you need the FID at all. An MSD is best left up and running. Thus you will want to leave a column installed and carrier flow – in which case you have a column that is ready to perform analyses. Years ago I worked in a lab that had a number of GC-MSD systems and each was purchased with second split/splitless inlet and a FID installed in the instrument “Just in case.” I recall instruments being retired with neither the second inlet nor the FID ever having been used.

Where I mentioned that you can get lower LOD's on many compounds with the FID, take note: This becomes a bit of a problem because at very low levels, you have a greater risk of confusion with low level interfering compounds - and selectivity and/or compound confirmation becomes very important - driving you to an MSD where you have detection problems with it. The solution is changing sample preparation rather than to using the less expensive detector - and because you are concentrating background garbage as well as analytes, the MSD may still be required for selectivity and compund confirmation.

In that lab we did have instruments with the gas stream from a single inlet split to two (or more) detectors. While this approach allows for flexibility in use – the sample is split (decreasing sensitivity) and the additional plumbing becomes an additional source of maintenance issues in the instrument. You can install columns into each of two inlets, with one column running to the FID and the other to the MSD. Either both columns are of the same type or you run into the issue that the temperature limits for both of the columns must be observed to prevent damaging the more fragile column.

If you purchase the instrument with both MSD and FID, plumb in the MSD only until you determine that you have a need for the FID as well. It keeps life simple – and you have the FID, just in case.

If you have an FID you have to know what you are looking at as the only identification information is retention time and peak area. You really need a reference standard.

However, if you know what you are looking at you can set the MS in SIM mode and have much greater sensitivity than the FID. Also modern MSD's have SIM/scan giving you the highest sensitivity of the compounds you know while enabling you to simultaneous have a TIC so you aren't missing anything.

Other than cost and simplicity, I believe the MSD is a superior analytical tool. The only question is if you need it.

To add to what MSchemist has already said - a small advantage of an FID over an MS is that you can calibrate the response of an FID using a standard of a compound that is similar to your target. With an MS you really need the target compound itself since MS response is more variable among compounds - especially if you are only looking at a subset of the total ions. Also an FID's response is very stable with time, so you can get away with calibrating less often.


Peter

The original idea why FID + MSD, was that even though the usual impurities in organic solvents are known, we still need to confirm their identity (that could be done on MSD) to prove, its not something else. Quantification would be then done on FID, because we though its more simple/reliable then on MSD (since MS response is variable among compounds). Is such deliberation correct?

BTW: Thank you all for your responses.

If you have sufficient signal to confirm identity on the MSD, you have adequate signal to quantify on the MSD. You do need a calibration curve built with that particular compund. While you can use a response factor from some other compund in the FID - it is best to use a calibration curve with the particular compund in question. This accounts, not only for the response of the detector, but also for things like inlet discrimination. It becomes a question of how good a number you need.

We always found a TCD to be a better choice for solvent purity, since one of the major impurities in solvents is water. The MSD in full scan mode is going to give you similar sensitivity to the FID, but you will have an issue trying to scan through the solvent peak. We time programmed ours to look for peaks prior to and following the solvent peak - even so, the tail on the solvent peak gave us trouble on the MS.

Typically you only have a handful of contaminant peaks in high purity solvents - you can easily make a run on a standalone MS to identify the compounds, and then run the GC-TCD as your routine system.

Thank you all for your advice.

Juraj

There is one more question I would like to ask:

What do you think of GC-IR/FID (is there such a combination available?) + standalone MS(we have got). IR may provide additional qualitative information which is complementary to the MS (as long as some isomers may give a same MS spectrum). IR spectrum should be enough to identify the solvent, so standalone MS would not be necessary at all.

Juraj

It's been a while since I looked at GC-IR. And while it is a great idea, what I came away with was that it was good only for compunds where you have a large quantity of material passing through the column. For impurities in solvent, you would get a good spectrum o fhte big peak (solvent) but beyond that, perhaps not much. If you really think you want to go that way, get a coupe of typical contaminated solvent sample and see if the vendor can run a chromatogram and show you what is in the sample. Be sure that you have some "big" peaks, some "small" peaks, and some really small peaks - and that you know 1) what they are and 2) that they are really there. If you ahve run a chromatogram, it might help to send a column along with the samples so the chromatograpy is what you expect.

Well, there is no need for us to identify impurities... so basically identification of the main peak would be sufficient. Anyway, because there are only limited number of solvents we would like to analyze and most of them have simple structure with no isomers, we will probably use for identification standalone MS and for quantification only FID. Another identification may be retention time on FID compared with standard.

Thanks a lot.

Juraj

Another issue to consider is if the impurities coelute with the main peak on the GC. Using the MSD you can quantitiate both using distinct fragment ions. If you are using FID you are SOL.