quantum mechanics - Why do tunneling photons outrace their non tunneling counterparts in vacuum?

If we describe a photons with a wave packet, moving towards a potential barrier and E smaller than V, there is a finite chance that it will tunnel to the other side. In this process it is likely that it will arrive before a photon that does not tunnel, not because it exceeds the the speed of light, but because the front of a wave packet will contribute most to tunneling. I would understand this if it was concerned with PARTICLES as follows: because each of the different waves that make a wave packet has a different momentum, the faster moving, higher energy, waves will move to the front of the wave packet when it disperses. According to the formula for tunneling probability indeed the higher energy parts will tunnel more often.

BUT, why would a photon wave packet also have the higher energy elements more in front of it's wave packet, since all light-frequencies move at the same speed in vacuum right? Or does the described system indeed not work in vacuum.

At page 14 of this link the mechanism I describe is presented:

http://www.physics.umass.edu/sites/physics/files/admupld/Tunneling-UMass-12Feb10.pdf

Answer

In my opinion the presentation you linked is a weird description that conflates different phenomena in a confusing way.

The following is a totally wrong explanation of the the results of that referenced paper (non-paywall link): "The front part of the pulse has the higher-energy part of the light, which is more likely to tunnel." Both parts of this sentence are wrong:

(1) The frequency profile at the front and back of the pulse is essentially the same in this experiment (the light pulse is not substantially "chirped");

(2) It is not true that higher-energy (higher-frequency) light tunnels with higher probability through this particular structure. (See Fig. 1 in the paper; the photon is around 702nm, where the transmission curve is almost flat. Technically, the transmission minimum is at 692nm, so the lower-frequency parts are transmitted with very slightly higher probability. However, this slight difference is not important for the effect--see the theoretical curves in Fig. 1.)

It is certainly true that the front part of the pulse is reduced a little bit while the back part is reduced a lot, which moves the center of the pulse forward. But the reason for this differential reduction is not that sentence above. I think it's a more subtle kind of wave-interference effect, which is not necessarily easy to explain intuitively.