Certain passive devices make use of a evanescent waves, and the fact that they operate consistently with that theory I think is the most direct proof there is. An example would be a coupling iris in waveguide -- that is, two waveguides separated by an "iris" (a smaller waveguide that is cutoff in the frequency of operation). You could try different iris thicknesses and verify that the coupling between the input and output waveguide decays exponentially with length and wavenumber, as is expected for evanescent waves.

For fiber, you could do an experiment where one fiber is slowly brought closer to another without touching until some light from one is coupled into the adjacent one (or equivalently light is lost from the first one) due to the evanescent waves outside the glass. Until the second fiber is brought close enough, these waves are totally reactive and carry no energy, but once the two are coupled there is a path for the light to escape by jumping across the gap. (One might call this a kind of "tunneling".) I'm not a fiber optics expert, but I do believe this is how optical couplers are made.

Thanks Morgan, one thing that I am concerned about is the evanescent depth (i.e.  the penetration depth) is in order of 400 to 800 nm depending on geometry of the optical fiber. What I do in my experiments is that I declad the fiber (remove the clad region by chemical treatment and expose the core region to surrounding medium)  and use it as an evanescent wave sensor. Yes, I will try to get the surface optical fiber as close as possible and observe what would happen. I will try to adjust the distances using an optical microscope.

Evanescent modes are specific due to propagation constant which is real. Propagating modes have propagation constant imaginary (in lossless case). What is the propagation constant of light in air or vacuum? I guess it is imaginary.

Dear Sir,

I think you can use optical sensors, an electromagnetic sensor is only the second possibility.

All the best

Thanks Adam and frank, I am using clad-less optical fibers (without the clad portion), light propagates by total internal reflection. At the point of reflection there is an exponential decay on the incident field (depth of penetration). My question is how to measure it experimentally, since the depth of evanescent field is in order of few hundreds of nanometers.