In ultra-high resolution electron microscopy, a major limitation is sample damage. Given the propensity of many samples to charge under e-beam irradiation, it is advantageous to use high-voltage e-beams in preference to low-voltage e-beams. Therefore, the only way to limit radiation damage is to lower the electron dose. Low dose unfortunately translates directly into small signals available for detection. This means that noise is increasingly important.
Another implication of increased beam voltage in (S)TEM is the drop-off of image intensity in ordinary phosphorescent detectors, along with reduced high-frequency modulation transfer function (MTF) due to long-range scattering within the phosphor. Industry has responded by introducing direct electron detectors in the form of back-thinned, monolithic active pixel sensors (MAPS).
With these developments in mind, a useful quality measure in addition to the MTF is the detective quantum efficiency (DQE). This is defined as
DQE = ( SNRo / SNRi )2,
where SNRo is the detector output signal-to-noise ratio, and SNRi is the corresponding detector input signal-to-noise ratio. DQE cannot exceed 1.0, but we would like to make it as close to 1.0 as possible.
McMullen et al. have undertaken a careful experimental study comparing the DQE and MTF of three commercial MAPS detectors, emphasizing the importance of DQE in choosing a low-dose detector. Their preprint is available on the arXiv.
Reference: McMullen, Faruqi, Clare, and Henderson, arXiv:1406.1389.