Imperfect Waveforms Noise and Ripple

So far, we have assumed FAIMS waveforms to follow the ordained U(t) dependence. The waveforms produced by real electronics always carry unwanted oscillations (electronic noise) of diverse physical origin, including thermal, shot, and inductive coupling of instmmental and environmental RF. Thermal ( white ) noise has uniform power spectrum that allows unbiased gauging of the impact of noise on FAIMS performance. Other noises appear over limited fiequency ranges ( pink noise) or at specific frequencies such as overtones of harmonics comprising Ed(0 or the industrial AC power frequency (60 Hz in USA) and its overtones. The spectra of those noises are sensitive to specific FAIMS hardware and lab environment. The effect of noise ceases above some frequency because ions have not enough time to respond. Based on the relaxation time estimates (3.2.1), 

A kind of noise is ripple —a harmonic oscillation (of some frequency wr) that is much slower than FAIMS waveform but periodic on the separation timescale (i.e., 1/ res Wr C Wc). For example, a common FAIMS design where 0.2 s calls for 

Wr 20 Hz. Ripple may be spontaneous (e.g., a noise at 60 Hz from AC coupling) or be added to E t) on purpose for resolution control as shown below. In Simula-tion, ripple is introduced as  

if the imperfections of E t) he in that ballpark, the resolution would be near-independent of the ion current the postulated redistribution of ions in the gap (4.3.5), though continuous rather than periodic, may be caused by noise on the FAIMS waveform. However, other randomizing phenomena may be as or more important (4.3.5) and the origins of peak smearing in FAIMS remain uncertain. 

What is the merit of controlling FAIMS resolution using ripple versus adjusting the gap width or waveform frequency (4.3.4) This is clarified by resolution/sensitivity diagrams (4.3.7). 

---