Has anyone here coated lens elements on silicon wafer CVD thin film deposition equipment? I have a 19” Dagor that I plan to coat at some point, and if there is some way to coat it in a wafer fab lab I’d try to sneak it into our prototype lab this summer. And if so, what materials/coating schedule should I use?
They are ill discoverers that think there is no land, when they can see nothing but sea.
-Francis Bacon
This is a hard thing to do correctly. The first issue is single or multi layers. You need to decide what stack of films and materials you want to use. Issues that need to be considered are finding layers that have the correct optical properties, stick to one another, are hard enough to wear well, are physically and chemically stable over a long period of time, have a coefficient of thermal expansion matching the lenses, are non-toxic, etc. A major issue is getting the lens surfaces really clean so that the films will stick. Usually you do a mild etch to the glass surfaces just before the evaporation to accomplish this. Another issue is that the index of refraction usually depends on the deposition process details such as temperature, rate, etc. Physical thickness is not necessarily the same as optical thickness. Folks who do this for a living monitor the optical thicknesses of the films as they deposit them so as to know how thick to make the films. The normal film thickness monitors that ones uses in a film deposition chamber measure the mass of the film deposited, not its optical thickness and so are of little help here. IMHO, before you mess up a classic lens, you should spend some time googling this. There is a reason the people who do this for a living charge a lot. It is not easy to get right.
Good luck,
Dave B.
As Dave_B mentioned I think this is typically done with an evaporation tool (such as an ebeam evaporator) not CVD.Originally Posted by John Schneider
I would suggest appreciating the character of your Dagor
Hmm. Dagor. Six elements in four groups. Four glass-air interfaces. Won't benefit much from coating. Why bother?
Lens coating is high technology - even single layer single index coatings. I've done it using the common magnesium fluoride material with an electron beam source as mentioned above and shooting for simple 1/4 wave thickness at mid visible spectrum (around 540 nm IIRC). As Dan points out you'll gain little by just coating the front element surface. I don't think you want to disassemble all the elements but I suppose you could if you need a challenge.
Cleanliness is next to godliness in lens coating enterprises. My old sequence:
1. Dust off the glass/Blow off the glass.
2. Wash glass in detergent/rinse in DI water/blow dry using N2 filtered.
3. Degrease in trichloroethylene or equivalent/rinse off in AR grade ethanol/N2 blow dry.
Place the glass in an ebeam evaporator with a source (hearth) to glass distance of at least 18 inches. For best adhesion you need to ion beam etch the glass very lightly then follow with evaporation of the magnesium fluoride using precalibrated thickness monitor like Inficon or equivalent. You'll need to do some test runs for thickness calibration before hand noting the quartz crystal thickness reading then etch a step in the film (or premask for a step) and measure the exact thickness using a profilometer like a Dektac or Tencor or multibeam Tolansky interferometer. The curvature of the lens plus that the MgF source is a point source will cause some radial variation in the coating thickness so the 1/4 reflection will be slightly off radially - don't worry too much - this is a sloppy job. You will need a consistent vacuum, < 10^-6 Torr., for reproducibility.
But using a CVD system - I've never heard of that. I suppose that would be possible using silicon dioxide. But you'd have to find some precursor chemistry for MgF and I would guess that might make the tool master apoplectic - even in a development lab.
It is all possible but heck; if you are really not set up it's a good way to mess up a lens by missing thicknesses and ending up with peeling films.
Just my two cents.
Nate Potter, Austin TX.
Can't someone invent an anti reflection paste that you just apply and then polish the glass with? Wouldn't that simplify things?
The wavelength of light is too small for something like this to work. The films need to be a fraction of a micron thick, accurate to ~2-5% and uniformly applied over the entire lens.Can't someone invent an anti reflection paste that you just apply and then polish the glass with? Wouldn't that simplify things?
There are precursor gases for MgF2, but they are not a normal thing for a semiconductor-lab to have available, and although adding a new gas line is not a major undertaking for a research or development oriented system, the knock-on effects of your residual gases in the reaction chamber would probably make exisiting users reluctant to let you try this for a whim.
An easier way to go would be with a sputtering or evaporation system - this is what professional lens coaters use. Assuming you have an empty boat available, MgF2 targets are readily available from the usual suppliers, although at a price that might make you think again for a hobby project.
If you have an evaporation chamber and an available boat and a source of MgF2, a 1/4 wave single-layer coating won't be a big deal to do. Modern coating machines are reliable and reproducible and you can probably get away with using standard calibration tables to get the thickness right. For a good coating you will have to do some calibrations runs, and measure the thickness of the films with a profilometer, AFM or ellipsometer - whatever your lab has lying around.
If you want to go beyond a simple single coating you will need to learn how to design multilayers. The principles are not mathematically hard, and are well described in many textbooks (see, for example, this one: http://books.google.se/books?id=667iktgtmmkC), but as with any technological process there is a fair bit of hands-on knowledge that is not found in the books.
I have access to all this equipment, and use it in my work (actually, I get other people to do my CVD stuff), but frankly feel that the $300 per surface charged by Focal Point Inc. is something of a bargain. But if you like to tinker, have access to the right machine, and have a competent background in experimentation, it wouldn't take too long to get something workable, although a coating as effective as Pentax's or Cooke's won't come quickly.
PS: J.Patric, you might like to read this paper, it describes easily-prepared polymer anti-reflection coatings. For the moment they are too soft to be used for front elements, but they're working on that:
Nanophase-Separated Polymer Films as High-Performance Antireflection Coatings
Stefan Walheim, Erik Schäffer, Jürgen Mlynek, Ullrich Steiner
Science 283 520 (1999)
http://www.sciencemag.org/cgi/content/full/283/5401/520
Not signed in I could only read the abstract, but it seems to be very innovative. Interesting.
Executive summary: it is now possible to design and structure materials on lengthscales smaller than the wavelength of light. That means that you can make transparent materials with combinations of refractive index, dispersion and mechanical stability which were not possible before. In this case, they made a mix of acrylic and air and could vary the refractive index depending on the proportions in the mix. Because the structure is much finer than the wavelength, light passing through it feels an 'effective', average medium, rather than being refracted and reflected off every tiny interface.
Technologically, these materials can be cast, spun-coated, or printed onto surfaces, and they don't need vacuum deposition onto hot substrates. That too opens up many possibilities, although I haven't seen any proposals for LF lenses :-)
Posting Permissions
Reply With Quote
Bookmarks