Molding Complex Optics In A Completely Fluid System

Molding complex lenses

Traditional lensmaking is a grind — literally. One starts with a piece of glass, rubs it against an abrasive surface to wear away the excess bits, and eventually gets it to just the right shape and size for the job. Whether done by machine or by hand, it’s a time-consuming process, and it sure seems like there’s got to be a better way.

Thanks to [Moran Bercovici] at Technion: Israel Institute of Technology, there is. He leads a team that uses fluids to create complex optics quickly and cheaply, and the process looks remarkably simple. It’s something akin to the injection-molded lenses that are common in mass-produced optical equipment, but with a twist — there’s no mold per se. Instead, a UV-curable resin is injected into a 3D printed constraining ring that’s sitting inside a tank of fluid. The resin takes a shape determined by the geometry of the constraining ring and gravitational forces, hydrostatic forces, and surface tension forces acting on the resin. Once the resin archives the right shape, a blast of UV light cures it. Presto, instant lenses!

The interface between the resin and the restraining fluid makes for incredibly smooth lenses; they quote surface roughness in the range of one nanometer. The use of the fluid bed to constrain the lens also means that this method can be scaled up to lenses 200-mm in diameter or more. The paper is not entirely clear on what fluids are being used, but when we pinged our friend [Zachary Tong] about this, he said he’s heard that the resin is an optical-grade UV adhesive, while the restraining fluid is a mix of glycerol and water.

We’re keen to see [Zach] give this a try — after all, he did something similar lately, albeit on a much smaller scale.

25 thoughts on “Molding Complex Optics In A Completely Fluid System

  1. Is it just me, or isn’t there a dramatic change in the shape of the surface as soon as curing starts? (Around 0:30 in the video). AFAIK most UV-cured resins shrink 10-20% when cured. This would surely mess up any fine calculations of the shape of the bounding perimeter.

    However, it’s an extremely neat way to get a smooth surface. Since surface tension can be influenced by electric field (think: Eidophor video projector), I wonder if it would be possible to control the surface shape on-the-fly with a grid of electrodes somehow…..

      1. It definitely does change shape, it starts with a bulge in the middle, that turns to a straight line between the ridges on the constraining ring.
        As long as you know the properties though, you could control it.
        Might make a simple lens difficult though, which may be why they demonstrated with a complex shape.

        Might be able to pre-cast a central rod to maintain a simple lens shape?

    1. As long as the surface finish and optical clarity of the final part are good, then the rest is just modifying the frame and process parameters to converge on any particular profile that may be desired. And given the extreme speed of the technique, it’s also cheap to make test articles.

    2. UV cure optical epoxy have very low shrinkage as otherwise they wouldn’t be useful. Glueing two optical surfaces together which then strains them both massively would not be good.

  2. Are the optical properties of these resins good enough for optical applications though? I believe the poor properties of plastic lenses used in some ultra-cheap optics are due to the plastic itself, not the smoothness of the surface. Same issue with SLA clear resin that can be sanded and polished, or clear PU resin injected into a silicone mold (which you can mold from an existing high-quality sens) : you can make lenses with them, but not great ones. I guess it probably has to do with the complex molecular structure of those resins, compared to the crystalline structure of glass.

    I somewhat recently did something a bit similar to the trick used in the article : in order to produce a custom clear flat part I poured some silicone resin into a flat tray and let it cure under pressure (with a plastic sheet above to avoid dust). Since the top surface is not in contact with anything, it cures perfectly flat. I then cut it into 2 parts and put them front-to-front with a spacer and a sprue and I injected some clear PU resin inside. I had a few issues with shrinkage and defects on the mold but I managed to make decent clear parts like this. Though in the end it was too much of a hassle and I ended up using laser-cut acrylic sheets which did the job well enough.

    1. “Are the optical properties of these resins good enough for optical applications though?”


      These optical UV resins are used both to bond glass lens elements together (a replacement for Canada Balsam) in all modern lenses with cemented elements, and to form Aspherical lens elements by casting a shaped layer of resin into the surface of a spherical ground glass element of the same refractive index. Both techniques are used in the highest-of-the-high-end lenses for photography, videography, machine vision, etc.

    2. You know I had the same question (about optical properties) but even if it’s not with some practice to account for quirks, this would be a great way to make the backings for first surface mirrors. Either way, now that someone’s proven the basic concept tweaking optical properties might be a much simpler task. Repetitive sure, but I’d imagine as long as the relative densities between the restraining fluid and the resin were maintained, you could fiddle with the resin itself.

      1. I too would like a nice telescope, but I wonder if this could provide a low cost 14 inch reflector for a Dobsonian.

        BTW, I apparently have a secret admirer, they are now using my latest handle when commenting i. e. The person you responded to.

  3. I really like the approach, however I’m not quite convinced that there are many use cases for such a free form. An aspheric lens is a rotationally symmetric lens with a sag function that differs from a sphere. A free form is the more generalized version of an asphere where the rotational symmetry is not met any more. Via the surface tension in combination with the form of the surrounding housing it is possible to control the non-rotational properties of the lens. However the rotationally symmetric part can only be changed by one parameter (amount of resin that is filled is). Therefor it is not possible to control the aspheric terms, or in general when describes e.g. as a Zernike polynomial you can control things like astigmatism and trefoil, but not spherical aberration and also not coma. Still I like the idea.

    1. That’s what I was wondering. One illustration shows three different shapes for n=0 (solid of revolution, no azimuthal variation) but with a straight-walled cylindrical frame they can only control the “Bond number” (B₀) and the total volume.

      Maybe by shaping the wall profile r(z), say, like a barrel, torus, etc? If the wall has to invade the light path to achieve a shape, the void could be filled in afterwards with the same material.

    2. It seems like there could be a few ways to vary the shape –
      Use different density liquids, spin the container, apply an electric field, magnetic field, vary the intensity of the curing light… Selectively curing the resin while varying things might be useful. Standing waves. There was a paper some time age (Disney research maybe?) on controlling water in a fountain with actuators on the perimeter to control the depth vs position sort of like a video screen.

    3. For some surface shapes, the optical resin could instead of free-floating within the form ring instead be placed within a cylinder. The ‘support’ fluid on either side could be pressurised to change the surface geometry before curing. Might be a good fit for parabolas or hyperbolas, probably not so much for more arbitrary aspherical elements.

  4. So, I think this kind of settles what materials that University used to 3-D print “In Gel”. You’ve got UV curing resin, your suspension, and your curing leds. That solves that mystery.

  5. Electric field mod wiould certainly work.
    If it wasn’t for the ludicrous cost of UV resin this would potentially be a solution for Third World countries.
    Modify existing glass lens for the patient, with a thin layer on top of the non curved side.
    Perhaps start with a bulk sheet of glass of the right size, and make a multilayer resin-glass-resin structure.
    The variable refractive index would actually be an advantage here!!!

  6. This puts me in mind of a 1970s paper in Scientific American where they were making lenses in the fluid state using three liquids of different densities: I /think/ they were managing at least one type of aspheric profile. The UV curing technology obviously didn’t exist (at least in public) in those days, and I think it was presented more as an interesting curiosity than one with immediate application.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.