LC-MS Agilent 6420 vs Bruker EVOQ QUBE vs Shimadzu 8040

[bulat]

Dear colleagues,

I need to select LC-MS/MS (triple quad) for my lab. The main goal is to identify compounds in unknown samples, but instrument will be used for quantitative determination, too.

Most our lab instruments (GC-MS, LC) are from Agilent, but it looks that Agilent has the lowest sensitivity. 6460 is more expernsive and we can not afford it.
I heard that Shimadzu improved during last years, but I am still afraid about quality. I saw few Shimadzu instruments at ACS National Meeting, but they did not look solid and reliable.
Bruker looks good and solid, very innovative instrument according to brochures. Did anybody have experience with Bruker LC-MS/MS? Bruker also includes CTC autosampler.
Prices are nearly the sime for all instruments.

What would you suggest? Thank you!

[lmh]

I would query why you want a triple quad if the main goal is identification of compounds in unknown samples. By unknown samples, I assume you'll want to identify compounds where you cannot predict in advance what they might be, even to the extent of having a long list of likely possibilities.

Triple quads are good where you have a list of possibilities. They are best when set up to look for a suite of previously-defined mass transitions, each corresponding to some known compound. They are often very poor at collecting full-scan data, and of course they don't offer accurate mass, which is one of the most useful tools for identification of unknown compounds. They can collect MS2, for searching against libraries (if you have one...), but since they have to choose what parent to select for fragmentation, and they're then scanning the fragment ions (the process they're not so good at), you won't get spectra for every ion in the run (it's a bit random what you do get), and the spectra won't be of particularly good quality. (On the other hand, being CID in a collision cell, the spectra may be more rich than you'd get from many ion-traps using resonant excitation, where you should only see single fragmentation reactions - no fragments of fragments; ion trap MS2 spectra are often dismally dull for a library search).

Triples aren't cheap, so your budget can't be too small. Have you got enough cash to consider a ToF instrument for the accurate mass (particularly, it's worth asking the suppliers whether your budget would stretch to an ex-demo instrument if a new one is out of range)? You could even consider an Orbitrap of some flavour.

Whatever happens, get the manufacturers to demonstrate, and provide you with data from test-samples you supply. It's always reassuring to see that the instrument you propose to buy can generate the sort of data you'd like to see...

[bulat]

lmh, thank you very much for detailed answer!

LC-TOFMS could be a great choice, but I am not sure if we can identify the structure of compounds. Brutto formula can be identified using exact masses of molecular ion, isotopic ions and their ratios (is the resolution and mass accuracy of TOF enough for that?).
For identification, we mostly use the strategy when we more or less know what to look for (e.g., we can set the proper MRM transitions from the literature), but do not have standards to calibrate retention times. In this case what would you propose?

What about sensitivity? As I see from specifications, triple quads are at least 10 times more sensitive than TOFs. But it is for standard samples (reserpine). What about real samples and matrices (food, soil, plants)?

I requested further information from manufacturers, lets see what they respond. Bruker representative advices to get LC-TOFMS.
In principle, we our partners in the city have LC-ITMS and LC-QQQ, so we can ask them to analyze and look for fragments.

[lmh]

Yes QQQ instruments are more sensitive in real life as well as analysing reserpine!

Hm, tricky one to know what to do! I worry slightly about identifying on the basis of a literature MRM alone, on the grounds that I often see multiple peaks in an MRM chromatogram, but if you have several literature MRMs all coeluting, it lends extra weight. It is never easy to be sure of the identity of something without a standard, although we're often obliged to guess when the chemical isn't available.

Mass accuracy of ToF: in my personal view it's generally good enough to be useful but not good enough to be definitive. If you search 1, 2, 3ppm around a particular mass, the number of empirical formulae that are possible tends to increase quite dramatically. If you have a large molecule, even at 1ppm, there will be multiple formulae. Worse, some will differ only very slightly in their isotope pattern. If I had a free choice to specify mass accuracy, I would ask for much better than 1ppm, but that's not realistic at the moment. Even with near-perfect mass accuracy, isomers will weigh the same, and some isomers will fragment almost identically too. But yes, a decent modern ToF will reduce the uncertainty on an identification a lot. It will exclude a lot of near-miss compounds.