lc/ms user purchasing a gc/ms (eek!)

[matthew]

Hi all,

My boss asked me and another student to choose a gc/ms solution. I originally chose Agilent's 6410 as our LC/MS which seems odd now, because Thermo's TSQ Quantum can plug into a GC. Ugh. In hindsight, I'm not sure if the GC was in anyone's mind back then, but it didn't help that Thermo didn't match Agilent's quote until a few days before Agilent's system was due to arrive.

That said, I have a few queries:
1) it seems that people quantify with a single quad and a gc. Is this at all frowned upon?

2) Hopefully I'll finish up school next year; is it unfair to configure the machine without CI (for downstream students)?

3) as far as triple quads go, do you have good experience with either the TSQ plugged into a GC or Agilent's 7000? Are there other QQQ solutions?

I will agree (ahead of time) to submit samples to the vendors of interest.

[Don_Hilton]

Quantification on a quad and a GC. Before the fancy tripple quads and such, that's the way we did it - and in many cases still do.

On configuring wiht CI - Unless you need it, spend the money elsewhere. You can look for an instrument that will accept a CI source later - if and when it is ever needed. (It requires some extra plumbing)

On your third question - I can't help you much. I'm still doing quantifcation the old way (but with TOFMS so I have full spectra).

[matthew]

Don,

Thanks for your response. The TOFs are snazzy instruments, and the mention of them gives me pause (e.g. why have 2 qqqs in the lab).

Is the TOF MS at all different for a GC? I've seen some literature from Agilent arguing that their new 7000 design qqq is snazzy b/c the first quad is coated in quartz and can withstand 200C. But I know that their QTOF has the same metal quad as the 6410 (and people hook it up to the GC). Is there anything special/laborious about keeping an MS clean when it is attached to a GC?

[jh1]

CI is really useful for a very few specific applications but the increased cost for the extra source, pain involved in obtaining high quality reagent gas and well as issues with swapping the sources back and forth really isn't worth it, unless of course you know future students plan on working on odd projects such as detection of explosive residues at trace levels.

Outiside of thermo and agilent I know Waters makes a GC-QQQ that's been on the market for a while.

[Don_Hilton]

Mathew,

For GC there is TOF and then there is QTOF. The TOF instruments give similar sensitivity to quads running in SIM mode - and TOFs always run full scan. The TOF's available for GC come as integer mass or high resolution instruments. A QTOF places a couple of quadropoe stages ahead of the TOF, allowing you to obtain fragments of a specific mass ion coming out of the first quad and then monitoring a fragment of the fragment in the TOF, like in a QQQ. You get the great sensitivity, but you have no idea what else is coming off the end of the column. -- And while I do know that there are QTOFs out there for LC, I don't remeber if anyone has a QTOF for GC or if there are all QQQs.

But, the important point is The best solution depends entirely on the problem.

Will you be handling mixtures with unknown compunds? - if so the ability to obtain full scan data becomes important. You can get good scan data at lower concentrations with TOF - but you pay more for it.

If you will be doing only analysis of specific, known, compunds the question is sensitivity range requriements. If your injection will contain at least a picogram of analyte (injected into the instrument), a quad or TOF should do fine. If you need to measure below a picogram on column, you may need to look at a QQQ. (Again spending more...)

And, if your mixtures are rather complex, you can increase the GC resolving power by going to heart cut GC or (my favorite) comprehensive two dimensional GC (GCxGC).

There are many options. Figure out the goal - and see who can do the best job meeting it.

Enjoy.

[matthew]

Hi jh1, thanks that's exactly the sort of info I need. I'll order 2 CIs (kidding, kidding). We work with plants (and microbes). I have contacted Waters about the gc-qqq and gc-tof (gct). I have to admit that Waters doesn't seem to be in the private industry; it takes eons to get any info from them. I think they matched Agilent's 6410 quote about 4 months after I bought it.

Hi Don,

I know of Waters' GCT (TOF) and Bruker seems to make a TOF and QTOF ready for use with a GC. Agilent also presented results @ ASMS from a souped-up QTOF plugged into a GC (6890). I see something from LECO, a company I am not familiar with; in its favor the first googling pulls up results from a high profile plant lab @ ucdavis. Hmmm, the unit is built like a washing machine so lab real estate would be impacted. Do you know anything about them? What TOF do you use?

[Don_Hilton]

I use the LECO TOF and presented GCxGC-TOFMS data at PittCon this past year. It does take up some floor space - but I guess it is a matter that to get something, sometimes you have to give up something.

I used to work for LECO, so I know a bit about them. To try to give you unbiased guidance: go to the literature and look to see what people are doing that is getting the kinds of information that you are trying to get. Look to see what instruments they are using. Talk with those companies - and get references. The companies will give you names of labs where their instrument is working well. This is good. Find out what it takes to get good results from that instrument. (The labs that do not get good results may have a bad instrument or may be unwilling to learn to use it correctly - either way they are not going to be able to tell you what is involved in getting good results -- they are not doing it.) And the labs that get good results may be able to tell you about the general level of success that others are having.

[matthew]

Hi Don,
Thanks. I feel like this exercise has given me somewhat of a whirlwind (superficial?) overview of the available instruments. Deep down I suspect (from google stalking) that Agilent will have a gc-tof solution pretty soon. The ship has to set sail though. I'm looking at:

a single quad gc-msd (old school quantification)
qqq from Agilent (7000)
gct (waters)
bruker's apci gc source and tof.

For the sake of this exercise, I'm down weighting the importance of a clean qqq spectra and soliciting advice about EI/database searches. Maybe that lowers the relevance of the Bruker instrument. In your opinion, are EI/database searches so mature that I can:

1) ignore TOFs with good chromatography?
2) ignore QQQs with good chromatography?

Sorry if this question is inane, but I do notice you chose a unit resolution machine, which doesn't sound so TOF like(?)...

[Don_Hilton]

starting from the bottom up on your last post.

There are two kinds of TOF's for GC - unit mass and high resolution. Back when I got into this game, the dynamic range on the high resolution TOF's was poor. I gather that it is better now. The integer mass TOF has a high dynamic range, which is required for quantitation of samples with analytes at widely varying concentrations - like you find in metabolomics samples. For GCxGC work, which I do, you need high acquisition rates to make use of cryofocusing. I typically run with acquisition rates of about 150 spectra per second to get adequate data density for deconvolution. As far as I know, the closest a high resolution TOF gets is about 50 spectra per second. This would work with a valve based system or a Dean switch modulator - because the peaks are wider. But the fact that I have to use high acquisition rates for deconvolution with a cryogenic modulator tells you that the wider peaks in the later systems are taking me in the wrong direction.

The question that you ask about choice of system is answered not by the maturity of mass spectral libraries, but rather by the objectives you are trying to meet. If you are asking about library search capabiliites, I would guess that looking at unknowns is important in your work. You have two choices. 1) use full range spectra and library searches to try to identify things or 2) use QQQ for sensitivity and accurate mass to try to identify things.

There are two basic limitations to library searching: 1) If your compund is not in the library, you will not get a correct hit. (you will often get something with a mathematically computed high score - that may be totally unrelated.) 2) Even if the compund is in the library, you are dealing with a match score that can be confounded by noise - giving you something else as the "best" hit. There are two cases that are famous for this - isomers and homologues. Isomers may immediately form the same carbonium ion - and that carbonium ion will fragment the same way, no matter what it came from. Homologs may immeditely fragment with a nutral loss -- that is the side chain in which the difference lies. If there is no molecuar ion, you are left withoout a clue (except from retention index).

With true unknowns you may have to fall back to mass spectral interpretation. Full spectra are handy - particularly if you can get accurate masses. There are techniques that can be used with a QQQ. They are not easy - but interpretation from full spectra is not a given either.

Match up your objectives and skills available with an instrument.

I lean toward the kind of instrument that I run because that is what I've been doing for a number of years. (Input from skill set.) But, in a discussion of a method for a specific series of target analytes at very low levels, I reccomended a tripplequad - for obtaining limits of detection needed for the project.

[matthew]

Hi Don,

You're right, we do often work with unknowns. In the beginning, it wasn't such a problem, because there are/were so many candidates from the literature that we could quantify. Now I think we would be benefitted by moving to more of a metabolomics approach.

I usually work with organic acids and other small things output from roots, so I'm going to try to push for a high mass accuracy TOF if it is financially feasible (even with the small mass ranges for these instruments). There's a tradeoff of course. I expect my boss to want to quantify some things at low concentrations (e.g. plant hormones); we'll see what need wins when push comes to shove. Thanks again.