Various expressions for transit times in frustrated total internal reflection are analysed. The incompatibility of evanescent-wave propagation with Einstein causality is established.
Photonic tunnelling experiments have shown that i) the Einstein energy relation is violated, ii) the tunnelling process is non‐local, iii) the signal velocity is faster than light, i.e. superluminal, iv) the tunnelling signal is not observable, since photonic tunnelling is described by virtual photons, and v) according to the experimental results the signal velocity is infinite inside the barriers, implying that tunnelling instantaneously acts at a distance. We think these properties are not compatible with the claims of many text books on Special Relativity, see for instance Refs. [1].
There have been many recent theoretical and experimental reports on the propagation of light pulses at speeds exceeding the speed of light in vacuum c within media with anomalous dispersion, either opaque or with gain. Superluminal propagation has also been reported within vacuum, in the case of inhomogeneous pulses. In this paper we show that the observations of superluminal and non-causal propagation of evanescent pulses under the conditions of frustrated internal reflection are only apparent, and that they can be simply explained employing an explicitly (sub)luminal causal theory. However, the usual one-dimensional approach to the analysis of pulse propagation has to be abandoned and the spatial extent of the incoming pulse along the directions normal to the propagation direction has to be accounted for to correctly interpret the propagation speed of these evanescent waves. We illustrate our theory with animations of the time development of a pulse built upon the Huygen's construction.
The group delay in frustrated total internal reflection (FTIR) is shown to equal the dwell time plus a self-interference delay. The self-interference delay originates from the energy stored in the overlap between incident and reflected waves above the interface. The group delay in FTIR for finite beams is thus shown to be the lifetime of stored energy in the gap between the prisms and in the partial standing wave above the gap. As in other manifestations of tunneling the saturation of group delay with barrier length (Hartman effect) is explained by the saturation of stored energy in the region defined. No superluminal propagation is implied.