The recent introduction of ion mobility to analytical mass spectrometers had provided an additional method of separation before fragmentation [319, 320]. The technique involves the movement of ions through a buffer gas under the influence of an electric field and separates ions on the basis of charge and shape rather than m/z. This field can be fixed or in the form of a traveling wave (TWIMS). The physical parameter associated with the drift time of ions separated by ion mobility is the rotationally averaged collisional cross section, a parameter specific to the glycan that can be used to aid identification. These cross sections can be measured directly with drift tube instruments, but traveling wave-type instruments require suitable calibration. A table of positive ion cross sections has been published by Fenn and McLean , and Pagel et al. have published extensive tables of positive  and negative  ion cross sections of N-glycans measured in both helium and nitrogen. A third type of ion mobility is high-field asymmetric waveform ion mobility spectrometry (FAIMS), an atmospheric pressure method that separates ions according to differentially mobility in a high electric field. Applications of ion mobility to the analysis of carbohydrates have recently been reviewed .
Although resolutions on current instruments are still comparatively low, they are still sufficient to separate some glycan isomers [321, 325-327]. Two isomers of Man3GlcNAc3 could be separated to baseline as its [M+Na]+ ion , and structures could be assigned by negative ion CID. The arrival time distribution (ATD) of the doubly charged [M+Na2]2+ ion from permethylated Hex5GlcNAc2 from chicken ovalbumin has shown evidence for three isomers . Isailovic et al.  have noted differences in the ATD profiles of sia- lylated biantennary N-glycans from human serum, but specific structures were not identified.
Another use of ion mobility is for group separations of compounds such as carbohydrate, peptides, and lipids [329-331]. N-glycans from ribonuclease could be separated from peptides . The method has proved to be invaluable for extracting N-glycan profiles from contaminated samples ionized by both ESI and MALDI where sometimes no evidence of glycan ions was
Figure 3.9 (a) Driftscope (m/z:drift time, log scale) display of the negative ions from a sample of released gp120 glycans contaminated with PEG. Circled regions are labeled with those of the panels below. (b) Total electrospray spectrum. (c) Extracted singly charged N-glycan ions ([M+H2PO4]- except m/z 2076, which is [M-H]-). (d) Extracted doubly charged N-glycan ions (high-mannose glycans) are [M+(H2PO4]2)2-, glycans with one sialic acid are [M-H+H2PO4]2-, and the disialylated glycans are [M-H2]2-). (e) Extracted triply charged N-glycan ions ([M-H3]3-). (f) Extracted singly charged PEG ions.
apparent in the original spectra [332-334]. Figure 3.9 shows an example from a sample of released N-glycans contaminated with polyethylene glycol (PEG) . Singly and multiply charged glycan ions could also be separated from small amounts of material extracted from recombinant virus samples . When combined with negative ion CID, this method has proved to one of the most powerful for N-linked glycan analysis and is described in detail in .