Field asymmetric waveform ion mobility spectrometry (FAIMS) holds significant promise for post-ionization separations in conjunction with mass-spectrometric analyses. However, a limited understanding of fundamentals of FAIMS analyzers has made their design and operation largely an empirical exercise. Recently, we developed an a priori simulation of FAIMS that accounts for both ion diffusion (including anisotropic components) and Coulomb repulsion, and validated it by extensive comparisons with FAIMS/MS data. Here it is corroborated further by FAIMS-only measurements, and applied to explore how key instrumental parameters (analytical gap width and length, waveform frequency and profile, the identity and flow speed of buffer gas) affect FAIMS response. We find that the trade-off between resolution and sensitivity can be managed by varying gap width, RF frequency, and (in certain cases) buffer gas, with equivalent outcome. In particular, the resolving power can be approximately doubled compared to "typical" conditions. Throughput may be increased by either accelerating the gas flow (preferable) or shortening the device, but below certain minimum residence times performance deteriorates. Bisinusoidal and clipped-sinusoidal waveforms have comparable merit, but switching to rectangular waveforms would improve resolution and/or sensitivity. For any waveform profile, the ratio of two between voltages in high and low portions of the cycle produces the best performance. (J Am Soc Mass Spectrom 2005, 16, 2-12) © 2004 American Society for Mass Spectrometry S eparation of gas-phase ionic mixtures by field asymmetric waveform ion mobility spectrometry (FAIMS) extends back over a decade in the technical literature [1][2][3], but has attracted a broad interest in the analytical and mass-spectrometry communities only recently [4 -29]. FAIMS separates ions by the difference between mobilities at high (K H ) and low (K L ) electric fields, in other words the average slope of mobility as a function of field, K(E). The form of K(E) may be complex: with increasing E, mobilities (of both cations and anions) may increase (termed species of Type A), decrease (Type C), or initially increase, but decrease at yet higher E (Type B) [4,5]. Usually, small and structurally rigid species belong to Type A and large flexible ones (e.g., proteins) belong to Type C. K(E) can be represented by a polynomial of even powers of E/N:At moderately high E/N relevant to FAIMS (up to ϳ60 -80 Td), eq 1 can usually be truncated after theExperimentally, a gas stream entraining ions passes between two electrodes that carry a time-dependent potential V D (t) that alternates between "high" and "low" voltages, such that the integral value over the cycle is null but time-averaged positive and negative voltages differ [4,5]. Then only a hypothetical ion with constant K(E) would be transmitted, while species exhibiting variable K(E) drift towards one of the electrodes and eventually neutralize upon impact [4,5]. For a particular ion, this may be prevented by applying ...