Hologram Exposure Times based on Microjoules

[John Sonley]

During all the years I have been making holograms, the exposure times I have used have been based on trial and error. These produce quite nice bright holograms, but no two are ever quite the same, either in terms of colour or brightness and I have been looking into adding a bit of science to my guesswork.

My home-made light meter uses a cadmium sulphide photocell and the measurement is that of the resistance of the cell. The spectral response is ideal for red light, (I use a 30mW HeNe) and it has the correct sensitivity for my type of holography.
The calibration sheet that comes with the cell is a graph relating Lux to resistance and after searching for a long time on the web I found a conversion between Lux and microwatts per cm. squared, it is 1 Lux = 0.146 microwatts/cm-2 at 555nm

I have now drawn a graph on 4 cycle log/log graph paper of cell resistance versus microwatts/cm-2.

If my sums are right therefore, 1 microwatt/cm-2 for 1 second delivers 1 microjoule/cm-2 of energy and pro-rata

I looked-up the sensitivity of Agfa 8E75HD from my old data sheets - Holotest, and from the graphs I saw that for an :
OD=3 Agfa film requires about 35 micro J/cm-2, for OD = 2, 23micro J/cm-2 and for OD = 1, 12 micro J/cm-2

So my proposed new exposure time in seconds = microjoules required for my chosen optical density / microwatt reading on meter

Using this method, where the time is fixed by beam energy only, to suit the film, leaves the type of developer used out of the equation. For example, Pyrochrome-Saxby and CW-C2 at 20 Deg are fairly slow, whereas ZIP 1 is FFFFast, particularly at 35 Deg. To allow for developer type, I imagine therefore that I should just remove the film from the developer when it looks black enough to the eye, irrespective of how long it takes, whether 10 secs or 4 minutes


I would welcome the Forum's advice about using the above method to calculate exposure time and also any updated information if the Agfa energy figures I am using are wrong
BRgds

John

[Jeffrey Weil]

Hello Everyone,

I never use that method myself. I don't have faith in working out the exposure by the amount of energy falling on the plate. There's too many other variables that are not accounted for to make it work the way you want to. At least in my opinion.

The age of the materials effects their speed. All materials from what I can see. DCG, AGX, change over time. Even resist changes speed as it's solvents out gas.

The temp of the developer is never exactly the same, the developer changes its performance during the day as material goes through it.

The object beam is usually impossible to get a real reading for. At least an exact one.

Many lasers will put out a slightly different amount of power every day.

our chemistry is aging and oxidizing while its still on the shelf, before you even mix it.

reciprocity. Our exposure times are all over the place. 10 sec one time, 2 min the next.

etc....

Even in the photographic biz, if you want your processor to keep the same performance day after day you have to use it all day every day. And use pre-exposed test strips (highly controlled by Kodak) and densitometers to see which way your chem is trending so you can make corrections to it's performance, etc... The point is, even with printing photos if you want the same results all the time it is not easy. Holography is much more difficult. And if you don't use your lab all day every day it'll be very hard to do.

I know people that try, and they make it work too, to do what your doing. But they go through a lot more resist than I do. Lots. And time too. They've worked out an aging index for their resist. The older it is, the more light it needs. They've worked this out with trial and error over time. They would take the readings, shoot the image, change the table and shoot the other images and then develop. They would see what the plate looks like and shoot again with exposure changes. Everyone of those is pretty expensive and time consuming.

If you want consistent results your best bet is to beef up the stability of your table and tighten up the environment of the room.

But, if you still want to use this method, I'm sure someone here has made it work and will say it's the best way. Funny thing is, for them it probably IS the best way. Every lab and every holographer seems to have methods that work better for them then others. This is a very personal science. I can only tell you what works for me.


Jeff Weil
NorthBeach Holography Inc.

[BobH]

I'm one who's lived "by the meter", but like Jeff says, to each his own. Resist is easy compared to AgX. Too much exposure and you bottom out. Very obvious. Too small a beam ratio and the image shows it fast. Silver halide allows much greater range for both, and the holographer has to make more subtle judgement calls.

The sensitivity quoted for AgX materials is chemistry and process dependent. You have to determine a different number every time you change the developer, time in it, temperature of it, (with some) batch of plates (or film), and the material's age. Then you can use the meter for consistent results.

Also, I'd not recommend "living by the meter" with a CdS cell detector. Readings depend on the cleanliness of the cell, and they are generally too large to get accurate object beam readings of "burn" spots in the recording plane of a transfer system (H2). They also don't read out in units that are easily and accurately convertible to microwatts for an exposure calculation.

[Dinesh]

Both Bob and Jeff bring up good points. there is lot of variability. Although I'd beg to differ from the extremes of

10 sec one time, 2 min the next

. It's never so bad that there is a variability of an order of magnitude for any medium. Jeff does bring up a good point that is often overlooked, ie the laser itself "walks" over the day.

Resist is easy compared to AgX. Too much exposure and you bottom out.

I would say silver is just as easy; too much exposure and the plate goes black almost instantly.

So, since you want to bring some "science" into this, let's do that. Bear in mind here that you're taking a picture of a lot of very, very tiny lines- be it as an alteration of the colour of a grain, or the density of a photosensitive medium. Bear in mind also, the lines vary sinusoidally, that is the "intensity" of the lines rise and fall in a sinusoidal manner. They don't go black-white-black, as some books show. The measure of a bright hologram is twofold: how sinusoidal are the lines in the photographic reproduction and what is the ratio of the height of the lines to the depth in the reproduction - the so-called Michelson fringe contrast. Remember, there is a "dc" level; the exposure increases the response of the entire medium, not just the lines themselves. The intensity is (simply stated):

I = I<1> + I<2> + 2*sqrt(I<1>*I<2>)* cos(phase_difference)

The first two terms (I<1> + I<2> ) give an increase in the overall medium response, what I call the "dc" level. Since the ratio is (I<1>/ I<2>), too high a ratio brings up the "dc" level, reduces the fringe contrast and kills the efficiency.

Setting up
Once everything is set up, there are two issues: coherence length mismatch and ratio.

The fringe contrast is a function of the mismatch in coherence length. While the common wisdom has it that so long as you're within the coherence length you're OK, the more extreme the coherence length difference, the lower the fringe contrast. You multiply the above equation by a coherence factor that's between 0 and 1, with 1 being the value when there's no difference between the paths and 0 being the value when you're outside the coherence length - when the fringe contrast is also 0. If you have a really deep object, you don't really have a lot of choice. But sometimes neurology comes to your aid. Distant objects are perceived as being a little dimmer anyway, so the fall-off of the distant parts of an object don't really show up that much. For this reason, it's a good idea to count the coherence length difference from the front of your object rather than the middle. That is, when you're stretch your string, start by doing it from the front of the object to the plate.

As mentioned above, the ratio sets the "dc" level.Too high and the "bottom" of the sinusoid comes up to meet the top of the sinusoid and you have no modulation and no fringe contrast. Too low and the "top" of the sinusoid does not get very far above the "bottom" and you still have no modulation and no contrast (I think I might start calling them "Goldilockograms" from now on!). Again, the common wisdom is the magic 4:1. That's not always true, it depends on the reflectivity of the object and the film response. This is a case where you just have to play a little. One thing I do is to baffle the ref beam and put a white plate in the plateholder. Then I rapidly block and unblock the ref; if there's too low of a contrast the difference is large and the white plate "blinks". If too high, there's no perceptual difference. If you're about right, then the white plate goes from grey to white rapidly (or, at least as fast as you block and unblock).

Expose
Assuming that the path difference is about right and the ratio is not too bad, the exposure now determines the fringe contrast. The exposure creates the latent image - a sort of "proto-fringe" structure - which the developing system amplifies, hopefully linearly. So, the better the latent image, the better the final holo. Usually, I measure just the reference beam, with the object beam turned off, and assume a (rough) 3:1 ratio. Hence, I add about 1/3 to the ref reading and call that the power on the plate. Usually, I start with the stated sensitivity of the medium, then add a "fudge factor" to account for age, local conditions and storage conditions. If the film has been kept in a high temp environment with high humidity it's probably decaying faster than one kept in a cool condition (caution! If you keep it in a fridge, allow for the condensation to dry before exposing!). The exact fudge factor varies but for silver kept in an ordinary cupboard here in San Diego (temp is 70F/20C and humidity is 50% most of year round) I usually add about 1/3. So, Slavich states that the sensitivity is about 150 uJ/sq cm and I use 200. You'd have to work out your own "fudge factor", but starting at 30% is not too bad of an idea.

If the ratio is not optimised, then you increase the exposure to take into account the decrease in light at the plate. Again, a matter of judgement, but increasing the exposure by another 1/3 may give you decent enough results that you can tweak it from there. Remember that the ratio is an average figure, since different parts of the plate will have a different ratio. Luckily, your eye/brain system is very good at averaging over a fairly high range of light levels (your eyes are log detectors, which is why they're very good at averaging)

If the path length difference is not optimised there's not a lot you can do about it. This is where stability becomes important to get as much of the contrast as possible, since it may be precious little. Usually, this is not too bad of a problem. Just remember to undo all those knots on the string from the previous measurements otherwise you're faced with a plethora of knots and it becomes a knotty problem (sorry!).

If you can arrange for a series of strips, independently removable, in front of the plate, you can do exposure tests. Assuming, say, seven strips, start with a nominal exposure and a nominal delta. Remove the first strip and expose for (nominal - 3*delta), then replace the first strip, remove the second and expose for (nominal - 2*delta). Similarly, the next strip is (nominal - delta), nominal, (nominal + delta) etc, ending with (nominal + 3*delta) on the last strip. This should be enough of a range to get you a good idea of exposure.

For DCG, there is no established sensitivity. The material is "analogue", in that the longer you expose, the harder the film becomes. The exact hardness you need depends a lot on the subject and whether it's a technical or display.

Development
This is where you amplify the latent image. This is also where the most uncontrollable variation occurs. Developers are temperature sensitive, concentration sensitive, chemical constituent sensitive, number of plates already developed sensitive and time-in-the-tray sensitive. For silver, I try to get a OD of about 2 in roughly two and a half minutes. Much longer than this and I think I've underdeveloped, and vice versa. Of course, there's a trade-off with exposure and the exposure tests (with the strips) will give a reasonable balance vis a vis exposure/development. The bleach time will also give you an idea on development. If it bleaches too quickly, I've not taken it dark enough. Generally, as a rule of thumb, I'd like it to bleach in about two minutes. Too long in the bleach adds noise, so taking it to a high OD means a long bleach, which adds noise. Of course, all of this depends on the dev/bleach combination, the the numbers of 2 minutes to OD of about 2 and 2 minutes to bleach to clear are a good benchmark to check exposure. The actual OD you take it to depends on the grain density, and varies from film to film. As I remember, for instance, 10E75 needed to be "20% black" for transmission and, with 8E75, "80% black" for reflection, which seems to translate to an OD of about 0.2 or 0.3 for Tx and about 0.8 for Rx. But, it still needed about 2 minutes to go to the prescribed darkness.

For resist, it depends on concentration and how many plates have already gone through the developer. I don't do a lot of resist, but if I don't get an image after about 5 seconds, I've probably undeveloped, or the etch is too dilute.

[Jeffrey Weil]

Hello Dinesh,

Both Bob and Jeff bring up good points. there is lot of variability. Although I'd beg to differ from the extremes of

10 sec one time, 2 min the next

. It's never so bad that there is a variability of an order of magnitude for any medium.

In silver or dcg your right. But in embossed pieces you often have 2-3 or more totally different types of holograms in the same final image. Some one step foregrounds, some transfered backgrounds, no object used gratings, encrypted images, etc...

They are all very different in how much light hits the plate. I've done much larger ranges than 10 sec to 2 min. in one hologram.

But thats only for embossed. Those kinds of things are almost never done with silver and dcg although they could.

I might be going to LA again in a few months. Maybe this time I'll make it down to your place to see the lab and have some beers.

Jeff

[Dinesh]

Hello Jeff
You're right that if there's different images on the same plate, then each one has a different exposure. I was talking about the same image under more-or-less the same conditions. Variations such as the age of the plate and the developer decay will cause a change in exposure by about 1/3 or a 1/2. If the change is much more than this, you either need new plates or fresh developer. In your case, you need to shoot a lot of images on the same plate and each exposure can vary by quite a lot.

Re the beers, give me a few weeks notice and I'll brew a batch! If you want to play with dcg let me know and I'll coat a few plates

[Ed Wesly]

Jeffrey, Jeffrey, Jeffrey, why hast thou forsaken me? Of course it is possible to keep everything under control photographically and holographically. Haven’t you ever had to go through the photographic rite of passage of making the same print several times on different days and seeing how close they match? Not easy, but do-able. It’s just part of the discipline and how far you are willing to embrace it. “Leave nothing to chance” is a maxim I’ve adopted from the can-do cowboy himself, Steve Provence.

Usually there’s enough of the important ingredients in photographic developers etc. to last at least a day in a tray without much change in activity. It helps to keep the tray covered to prevent oxidation. I used to do plenty of B&W and color developing in my home lab. Temperature is critical for consistency, especially in the latter, so you need to have a scheme for temperature control of the water, like with a temperature regulator. Some people may not be as anal as I am, but it is a key to success and a minimal amount of waste. It took me a while, but I have built my dream home lab.

I leave developing time constant, as that is what I grew up with doing the afore-mentioned prints. You can’t get a good black in the shadows if you don’t develop for a minute and a half in B&W prints in developer at least 20C. I employed that philosophy in the holographic darkroom until I took TJ’s advice and discovered that 30” immersion does work well with the PFG-03M and GEO-3 plates, and on top of that at 18C! But now I religiously follow that time and temp, and vary the expo time, at least for those particular batches of plates.

I do have my own speeds for holographic materials, which are off by an order of magnitude from John’s, namely I would use 100 to 400 uJ/cm^2 for Agfa 8E75HD plates with 2’ in CWC2. The ultra-fine grained plates mentioned above I shoot at an insane 3200 to 6400 uJ/cm^2 with 30” @ 18C in JD-4. These numbers are what I use in determining my brackets for my test exposures before I commit to a full plate. Usually it works!

When I run the test series, I usually use like a photographic stop system, doubling or halving the times, like 1, 2, 4, and 8 in honor of the logarithmic response of the human eye. If I want finer precision, I use half stops, which look like the f/number system of photography: 1, 1.4, 2, 2.8, etc. See my grad school web page, nlutie.com/ewesly for details.

So I guess that I am the guy you mentioned earlier whose way IS the best!

[Jeffrey Weil]

These numbers are what I use in determining my brackets for my test exposures before I commit to a full plate.

Hello Ed,

I was laughing pretty hard there combined with some guilt that I had offended one of my mentors sensibilities....until I saw that You do test strips too. We were talking about using the meter to go all the way to a final plate without tests.

I feel so much better now knowing I've not gone astray from your excellent teachings.

I want to see your dream lab but I'm not headed your way. Maybe we can have a skype tour of each others labs. I'm really proud of mine.

When are you coming down here again?

Jeff

[Dinesh]

I would use 100 to 400 uJ/cm^2 for Agfa 8E75HD plates with 2’ in CWC2.

I seem to remember that the sensitivity of Agfa was stated at 25 uJ/sq cm, while I usually found them to be around 75 or more, sometimes going up to 100. Of course this depends on the wavelength you're shooting with and I believe that Agfa stated a figure appropriate for 633. At ABNH, we used about 100 and went 2' with D19/FeNitrate. It's interesting that you go up to 400 for CWC2, but I was wondering if you get reciprocilty problems?

like 1, 2, 4, and 8 in honor of the logarithmic response of the human eye.

Actually, the f stops are not a factor of the log sensitivity of the human eye, but the illuminance at the film plane, whcih is assumed linear, due to the area of the aperture and the relative to the focal length of the lens. The question is: if you wanted to halve the illuminance at the film plane, by how much would you reduce the apreture? Since the area of the aperture is proportional to r^2, for a given lens, every time you halved the amount of light, you reduced the aperture stop by 1/sqrt(2) = 1/1.414.

The logarithmic response of the human eye - known as the Weber-Fechner Law - is a relationship between physical illuminance and subjectively perceived brightness.

[Jeffrey Weil]

Re the beers, give me a few weeks notice and I'll brew a batch! If you want to play with dcg let me know and I'll coat a few plates

Those both sound great, I'll keep you informed on my plans. I've never had home brewed beer!

Jeff