transmission holograms on DCG?

[Thieu]

Hi all,

I've been wondering for a while why there's hardly ever any mention of making *transmission* holograms on DCG. Is there any reason for that besides the fact that it works so well for volume reflection holograms that people rather make those on it?

Thieu

[Joe Farina]

Is there any reason for that besides the fact that it works so well for volume reflection holograms that people rather make those on it?

Hi Thieu. I think that's all there is to it. From what I've read, DCG makes excellent transmission holograms. I've never done it, because I'm only interested in non-image-planed master DCG reflection holograms.

[Dinesh]

We've actually done a lot of transmission work on dcg. The attached picture is a rainbow of a mermaid shot on dcg (designed by Colin, actually)

The problem is that dcg is bright as a reflection because it's broadband. In transmission mode, you're using the entire spectrum of the reconstruction light and so broadband carries no advantage. You can make transmissions on silver which are just as bright, if not brighter.

[Martin]

We've actually done a lot of transmission work on dcg. The attached picture is a rainbow of a mermaid shot on dcg (designed by Colin, actually)

The problem is that dcg is bright as a reflection because it's broadband.

Yes, that's one part of it but not the whole thing (by the way, welcome back, Dinesh and congratulations on your new patent):
DCG reflection holograms tend to produce bright images (brighter than AgX materials most of the time) even when illuminated with a laser only. So they may be very broadband and yet highly efficient under narrow bandwidth illumination.

In transmission mode, you're using the entire spectrum of the reconstruction light and so broadband carries no advantage. You can make transmissions on silver which are just as bright, if not brighter.

Maybe AgX provides better fringe integrity than DCG???

[BobH]

I think it's just a matter of convenience. But DCG should be excellent for transmission holograms because of its stability in light (no printout issues), and its clarity (if processed well). I'd use higher beam ratios to keep the noise low.

[Dinesh]

DCG reflection holograms tend to produce bright images (brighter than AgX materials most of the time) even when illuminated with a laser only. So they may be very broadband and yet highly efficient under narrow bandwidth illumination.

I think this may be because dcg tends to be thicker than commercial AgX (~8 u ?) thus there are more fringes. The problem I find with dcg is differential swelling. As soon as you put the dcg in the water, the top surface starts swelling. As the water penetrates, the lower levels swell, but the top surface is still swelling so that the upper surfaces swell more than the lower surfaces (so causing the higher bandwidths), but there is more uniformity in the lower surfaces unless you put it in the water for a long time. If you harden it so that surface absorption is slowed down, you lose efficiency. So there's a trade-off between differential swelling and efficiency. But, most dcg emulsions are in the order of ~15 microns when the standard coating methods are used. This is twice the thickness of AgX and so you get twice the number of fringes for the same spatial frequency. Thus,depending on the development, the proportion of fringes that are close the laser wavelength is still greater than the proportion of fringes in AgX at the laser lambda even though AgX does not suffer as much from this differential swelling. This may explain why reflection dcg's are brighter under the laser.

Maybe AgX provides better fringe integrity than DCG???

Right. AgX emulsions are harder, so there's less differential swelling.

I'd use higher beam ratios to keep the noise low.

Yes. The fringe contrast is a function of the beam ratio. The old 4:1 was a consequence of the HD curves of Agfa and Kodak. In the linear region, 4:1 gave a good magnification with high gamma in the linear region, even though it didn't give the best fringe contrast. Dcg is by it's nature non-linear and so the rationale for 4:1 no longer holds.

by the way, welcome back, Dinesh and congratulations on your new patent

Thanks!

[Sergio]

Welcome back Dinesh and congratulations for your job-patent. Right today I did test a film photopolymer on my lab that is real time holographic and I setup for a ratio 4:1 for a diffuse object, very hard to distinguish any noise from hologram in situ self developed compared to object, I wonder about the ideal beam ratio for the high linear material, I suppose for low noise. this emulsion was around 30um, guess the matrix plastic helps to maintain the exact thickness or at the end the overall final thickness.

[Martin]

I think this may be because dcg tends to be thicker than commercial AgX (~8 u ?) thus there are more fringes. The problem I find with dcg is differential swelling. As soon as you put the dcg in the water, the top surface starts swelling. As the water penetrates, the lower levels swell, but the top surface is still swelling so that the upper surfaces swell more than the lower surfaces (so causing the higher bandwidths), but there is more uniformity in the lower surfaces unless you put it in the water for a long time.

I would have expected even more non-linearities to occur from the the alcohol dehydrating baths. Putting a wet DCG layer into an alcohol bath, dehydration will start on top of the layer to proceed into deeper areas. However, the way I "understand" it, is that it cannot be a linear process since the freshly dehydrated section of the gelatin layer won't make it easier for the alcohol to penetrate the layer throughout its depth. So I guess that's one reason why it becomes difficult to process thick DCG layers above a certain limit.

Another subject I've never seen treated in classic DCG literature is the local heating that takes place when the water soaked layer gets in contact with the alcohol. Locally and probably for a very short period of time, the rise in temperature might be considerable.

[Dinesh]

I would have expected even more non-linearities to occur from the the alcohol dehydrating baths. Putting a wet DCG layer into an alcohol bath, dehydration will start on top of the layer to proceed into deeper areas. However, the way I "understand" it, is that it cannot be a linear process since the freshly dehydrated section of the gelatin layer won't make it easier for the alcohol to penetrate the layer throughout its depth. So I guess that's one reason why it becomes difficult to process thick DCG layers above a certain limit.

It's probably a fairly complicated process. Dehydration starts from the top of the plate, which will harden the top layer. At the same time, the alcohol is being diluted by the water, which slows down the dehydration process. So there seems to be two equilibrium states:
alcohol dehydrating the gelatin -> top layer hardening -> resistance to further penetration of alcohol
alcohol dilution by diffusion of water from surface into the alcohol layer above it -> penetration of diluted alcohol into gelatin -> dehydration of water out of gelatin

And then there's the hardness, or age, of the gelatin and the original bloom strength. Having said all this, I think the effects will depend on the spatial frequency and so will become more and more important as we push into deep blue and UV.

The heating effect is interesting. I had not thought of it before. I wonder if the local heating will cause a local swelling of the softer gelatin. This may cause the dcg to have surface characteristics - and so increase dispersion - due to surface deformities exactly between the fringes, or rather, between where the fringes cross the surface.

[Dinesh]

very hard to distinguish any noise from hologram in situ self developed compared to object

If the emulsion is smooth at the start I would think that in situ development would avoid local inhomogeneous areas in the emulsion, so that there would be no scattering centers. In AgX, it's hard to avoid scattering because of the crystals embedded in the gelatin matrix. But, if the crystal size is small, Rayleigh scattering would apply in the blue and may not be seen much for wavelengths in red - green. Dcg is usually pretty clean because, again, it's smooth. So, if the preparation of the emulsion is done at the proper temperature etc to avoid any local "clumping" of the gelatin before applying to the glass, there's very little noise. There may be effects due to speckle, which will create low frequency gratings at random throughout the emulsion, but I'd think the effect will be small if the primary grating is strong.

I wonder about the ideal beam ratio for the high linear material

Well, the fringe visibility is given by:
V =( R^(0.5)/(1+R))*(factors due to coherence and polarisation)
Where R = (r/o)^2 (beam ratio squared)

So for a perfectly coherent, s polarised beams, R=1 gives the maximum visibility. But for a high index material, the visibility also depends on the density modulation,by Kramers-Kronig. If the index is high, the modulation must center around the density of the material. This then depends on the mechanism of fringe creation. If this mechanism is slow and non-linear, then achieving a visibility less than 1 quickly may be better than achieving the best visibility with R=1 over a (much) longer exposure time in order to avoid motion problems.

For a diffuse object, there's a variability in the reflected light so usually R > 1, otherwise you get intermodulation noise creating spurious, low frequency gratings. There's also the effect of speckle, but I think this is small. I think there must be an effect depending on the spatial frequency also. The low frequency spurious (noise) gratings will have frequencies centered around some average noise frequency range f_n, while the frequencies due to the object beam will center around f. I suspect that noise is proportional to |<f_n> - <f>| .

So, the "best" ratio is probably close to 1:1 and depends on the size of the object relative to the size of the recording medium, the exposure wavelength and the mechanism off ringe creation. Also, if recorded with one wavelength and reconstructed with another, then it may be clean at one wavelength and noisier at the other. I've often wondered whether the noise in a rainbow is uniform over all the slits.