1) To my knowledge Darran Green is the current leader in home preparing Lippmann emulsions from scratch, and making images too. But also see the "ultimate" website of Yves Gentet, who is in France. He makes great holographic emulsions, and was offering Lippmann emulsions which have the same grain structure, but are thinner than the holographic ones (see below), though they are not advertised at present.
2) There are tradeoffs to be made in this game: for simplicity use no reflector, emulsion facing away from the camera lens, as stated in a post above. But then the reflected wave is only about 4% as bright as the incoming wave, and so the interference fringe contrast is cut severely. The result is an image of lower contrast and saturation. That having been said Darran's results and Hans Bjelkhagen's results, at their best are about 80-85% as good as Lippmann's or Neuhauss's from the early 1900s. In part this has to do with careful processing post exposure.
3) Coherence length is a major issue for at least 2 reasons. First, the greater the coherence length the deeper the interference pattern. The deeper the pattern the more layers, and in general the stronger the resulting reflection and image. Secondly, if the reflector you use is farther than the coherence length from the emulsion, due to an air gap or index-matching layer thickness, then no pattern will record in the emulsion, and you will merely get a black and white negative with no color.
Coherence length is given (to a good approximation) by: c.l. = (λ)^2/Δλ = (wavelength squared)/(bandwidth) in consistent length units. Two examples: for the visible spectrum, take the wavelength as 5500 angstroms (550 nanometers) and the bandwidth as from 3500 up to 6500 angstroms= 3000 angstroms. The coherence length is then about 10^-4 cm, or 1 micron. Hence, as stated above, for white or general broadband light an emulsion deeper than about 1-1.5 microns is not needed, and in fact to the extent that photons not in the interference pattern expose the emulsion at greater depths, they constitute noise and decrease the quality of the image. Thus the usual holographic emulsions, which are typically 6-7 microns thick, are not ideal for Lippmanns. A second example, for a typical gas discharge spectrum line the linewidth becomes the bandwidth, and these vary quite a bit, but take the line to be 1 angstrom wide. From the equation above you can see that the coherence length will be about 3000 times longer than in the first example, which was 10^-4cm = 10^-3mm, so the 1 angstrom line will have a coherence length of about 3 mm.
4) In my experiments, using aluminized mylar film of an ordinary sort as a mirror, adhered to Agfa 8E-75 with
with various evaporants, and using as a light source a sodium vapor safelight, which has the 2 bright, fat (several angstroms wide) yellow D lines at about 5900 angstroms, I could see that not only did I get color, but I got a 3-D image of the bumps and crimps in the mylar film, which had occurred due to uneven handling and adherence. I recorded a very shallow Lippmann reflection hologram! A 1 mm deep hologram is not very pleasing for ordinary 3-D, but using a non-laser to make an even quite shallow hologram has interesting possibilities. Without the mylar and using the sodium vapor lamp for illumination of objects in a darkened room you still get quite strong color because the fringe pattern is still quite deep, and it gives interesting images. You also see that a single color source may result in multi-colored images, in a not easily controlled way, due to variations in density (i.e., exposure) in the negative.
It should also be noted, from the equation above that longer wavelengths have longer coherence lengths, as noted in a post above, so you get deeper patterns in the red than in the blue. Further, the interference layers are thinner (λ/2 spacing!) in the blue so that the required resolution of the emulsion is set by the blue if you are doing broad band work and it is the shortest wavelength.
5) I have used Kodak D-19 as a developer, with only modest success; better is one of the older formulas, especially one by the Lumiere brothers, which has storable parts A and B which are mixed at time of use.
Lippmann and others succeeded with many types of developers, tanning and non-tanning, and with both unbleached and bleached (usually mercury based) images. Generally these images are not fixed,as fixing tends to shift the color to shorter wavelengths due to emulsion shrinkage. I also found that bleaching shifted the color of narrow band images, using ferric EDTA bleach as for holograms. For wide band images bleaching tended to destroy the images.
6) Besides a narrow band source such as the sodium safelight, an excellent early trial subject is a spectrum. It is relatively easy because since the colors are spread out, at each point along the spectrum in effect you have only a narrow band width, which means, as above, that the coherence length is deeper than if you had the whole visible spectrum at each location in the image. This will also show you, by way of the intensity variation along the spectrum, just how panchromatic the emulsion sensitivity is. By the way, Lippmann showed, first using Fourier synthesis theory and then in practice, in contrast to his doubters, that he could record a good white color. In that case the interference patterns add up to one thick layer, with a density (intensity) maximum 1/4 wavelength in from or at the surface of the emulsion, depending on whether there was a mirror in perfect contact or just the air interface.
7) You really should look at the Wiki when it comes back up.
8)Finally, where are you Layladies from and who are you?
I hope this all helps, and I am Yours,