06-JUL-2018   
      
   My simple experiment measures the frequency of an evanescent wave   
   that appears to form at the surface of digital detectors during   
   image formation.  Wave pattern is recoverable from the image   
   by Fourier transform.  Maybe.   
      
   Returning again from the drawing board, I will try to   
   briefly articulate the phenomenon I'm observing.  I find   
   certain insights borne out in the arithmetic, but my under-   
   standing falls short, perhaps far short, of what the maths   
   might be saying.   
      
   By the same token, my lack of fundamental grounding prevents   
   my recognizing what might be obviously flawed in my formulations   
   and interpretation, and why the whole enterprise should be of no   
   interest to anyone but me.   
      
   It's one thing to think about nature, deeply, as an informally   
   educated person; a respectful pedestrian in this forum trying not   
   to exhaust the patience of resident artisans and students.   
      
   It's a whole other thing to actually observe something deeply,   
   and then work out a way of making valid measurements with which to   
   substantiate and communicate those observations, on the odd chance   
   it has significance to others.   
      
   The detector is represented by any well-formed image produced by the   
   device; and defined computationally as a square, N-dimensional array   
   of square pixels with dimension d.  The observed fringes are then   
   a normal spatial image as a result of a computed image transform   
   applied to an input in the Fourier domain.   
      
   A histogram of measured frequencies, each sample denoted nu_n, is   
   produced, for arrays having pixel size d of the order of 10 microns.   
   The plot shows a fairly narrow, normal distribution of frequencies   
   centered on about 1.2THz, using a constant N = 512 sub-field.   
      
   By centering a source image sub-field on a different pixel, even   
   adjacent, a unique set of fringes is obtained by transform.   
      
   Using data from multiple, independent instruments, with pixel sizes in   
   the range of 6.8 to 21 microns square, I get consistent results, +/-   
   maybe 300GHz.   
      
   >From these data I've worked out the general formula for the quantized   
   frequencies observed:   
      
   nu_n = n(nu_o), E_n = h(nu_n)   
      
   where h = 4.135 x 10^-15 eV sec, n = (Ndp_n / h) is quantum mode,   
   p_n = E_n / c = n(h / Nd) is total momentum and nu_o is frequency   
   in the limit n=1; nu_o = (c / d) / N, with c equal to light velocity   
   in vacuum.   
      
   This was my point of departure some time back.  Can anyone comment on   
   the validity of this configuration, and what one might expect from a   
   detailed model to account for not just the frequency, but all the other   
   things visible in the images of the fringes; things even a pedestrian would   
   see.  I can make only limited sense of standard texts on diffraction   
   theory, quantum mechanics etc.   
      
   My analysis, for what it's worth, leads to a model that suggests the   
   fringes are interference of the not quite simultaneous emission of two   
   photons from each pixel, due to the recoil of the confined photocharge   
   packet from the scattering of a background photon.  The photocharge mass   
   is proportional to the flux of the primary (imaging) photons putting the   
   charges in the pixel during the exposure.  Thus, the last scattering event-   
   induced secondary emissions are must be encoded in the image data read off   
   the detector after the exposure.   
      
   I was beginning to think this might make sense for inelastic scattering   
   of mid-infrared photons during image formation, but then I applied the   
   procedure to the output of a bolometer array, where each bolometer is a   
   pixel with a diameter of 1.5mm.  The above formula still works; I'm getting   
   fringes with a mean frequency of about 14GHz.   
      
   CCDs and bolometers are fundamentally different.  What is confined in a   
   bolometer that could recoil and emit?  It can't be a photocharge. . .   
   In this bolometer array (photometer in SPIRE instrument on Herschel   
   Space Observatory), near as I can tell from reading, there is a measure of   
   power by a fantastic little contraption called a spider web that looks   
   like a Brancusi.  But a lot smaller.  Interestingly, the photometer images   
   in three passbands, 250, 350 and 500 microns, and the fringes generally   
   disappear in the midrange passband.  No phase information?  Very confusing.   
      
   More to the point:   
   How can power, confined in a pixel, scatter a photon and subsequently   
   lead to a correlated emission at proportionately lower energy?   
      
   Can evanescent waves be recorded digitally?  It seems that this would allow   
   a measurement of the wave without disturbing it.  The input exposures I'm   
   working with ended ten years ago.   
      
      
   Confused pedestrian grateful for any directions;   
      
   cheers,   
   mark jonathan horn   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)