by
lmh » Thu Feb 11, 2010 1:47 pm
I only know the original LC-Orbitrap. Its accuracy is certainly better than 0.001u, without internal calibration and it keeps this accuracy for the whole day after a single calibration in the morning (this is in any case a very easy one-click process). It has a nice lock-mass feature using contaminants present in the solvents (phthalates etc.) or bits of silica off the column (your choice) which improves accuracy and the length of time you can keep going on one calibration.
For a better idea of how things have progressed since the early days of Orbitrap, you could try asking Thermo Fisher. They are also the best people to ask about front-ends.
From personal experience, with middle-of-the-range low MWt compounds (400-600Da) I am upset if I get an error greater than 1ppm, and I am a low-skill, very occasional user on someone else's instrument. Note that very weak peaks are measured with just as good an accuracy as intense peaks, and that accuracy does not depend on resolution (i.e. if, for speed of sampling, you use it at lower resolutions, e.g. 15,000, it is just as accurate as at 60,000).
One weakness (which may have been overcome, I don't know) is that positive-negative switching was a nightmare. It takes about 1 hour to warm up after changing over. Also, in any one scan, the mass range is limited to a factor of about 10 (i.e. 100-1000 or 200-2000 etc.). A scan of 100-2000 gives very weak signals above 1000. Again, this may have been improved.