Make Your Own Optical Lenses

Today we're making lenses with epoxy, using a replication molding technique. It... mostly works 😇

CONSIDER SUBSCRIBING 🥰

==== Details ====
We can replicate lenses using silicone molds and optical epoxy resin. But there are a lot of caveats to this technique, and little hiccups that will ruin the optical finish. To achieve an optical finish, you need surface roughness and figure deviation in the nanometer range. This means curing temperature and shrinkage become major sources of error.

Residual stress in the molded parts can be assessed with polarized light and a polarizer in front of the camera.

==== Prior Videos ====

==== Materials ====
Mold material: Smooth-On Mold Max XLS ii
Optical resin: Smooth-On Epoxacast 690 clear

==== Analysis Equipment ====
Phenom XL SEM from Thermo
Gwyddion for AFM post-processing
Blender for 3D AFM rendering

==== Timeline ====
0:00 Intro
1:09 Replication Molding
3:04 Alternative mandrel material
3:37 Molding and casting technique
4:30 Fabricated lens examples
7:26 Molding priorities
9:38 Molding materials and considerations
10:49 Mold release difficulties
14:30 Shrinkage difficulties
16:04 Effect of Temperature
17:20 Internal stress and polarized light
18:51 Mechanical difficulties
19:40 Alternatives to silicone?
20:40 Direct molding off mandrel?
21:45 Refinement for future work
23:45 Concluding remarks

[BreakingTaps]

*Note about the audio* 🚨 Sorry for the noise gate issue! The SEM is in the same room and has two pumps running continuously (diaphragm pump + turbomolecular pump). I cover it all in acoustic material, but there is still a lot of noise to clean up. And apparently I did a bad job this time (combination of noise removal in Izotope and gating in Resolve). Apologies for the audio discomfort! I edit with headphones on, but probably went numb to it after a while and didn't notice. Luckily, it should improve soon... the SEM is moving to a new room in the near future 🙂

*Addendum* It seems I overstated how hard it is to grind and polish polymers! Several folks have informed me that e.g. eyeglasses are mostly polymers these days. I'm not sure what kind of tolerance they hold relative to glass lenses, but clearly it's a mass-produced and probably polished on-demand for each prescription. I suspect it's still a lot harder to DIY good polymer lens grinding than glass though.

[PeuSHINIGA]

Me, seated on a brick as chair into a 3rd world country watching this carefully in order to make no mistakes when I have to cast precision lens with resin somehow.

[codemakeshare]

The definition of "home shop" seems to a bit fluid, but maybe that's just me ;)
"making optical-quality surfaces [...] is really hard to do in a home shop, I don't have a single-point diamond lathe."
"just sputter some silver onto your glass in your sputtering chamber..."
"let's look at this under the SEM / atomic force microscope..."
Just joking of course, awesome content, very interesting.

[AlphaPhoenixChannel]

Very cool! The surface finish and actual shape (I forget your actual word for that) look amazing! I have to say I’m drooling at the thought of a diamond lathe for making aspheric curvatures but I’m afraid to look up the cost…

[SarahKchannel]

hmmm... i did a few things but improved the surface by spin coating the optical elements in a second step with resin. The layer is very thing to the spin coating but finishes the surface to much higher perfection.

[seanmcelwee]

Would be curious if you could use a non silica glass, like CaF2 or MgF2 to avoid to silicone/silica bonding issues with the mold and not require sputter coating

[squidcaps]

Hmm... i should try this. I've been making deformed lenses from cyanoacrylate. The whole point of those is to have optical quality "blobs". They give uneven ghostly images when used with RGB leds, deliberately breaking the smooth beam. As the RGB elements are slightly off set from each other, each individual led gives its own distorted image. But because they are so close to each other, you get similar shapes, for ex arc where one side is blue, one is red with a green band in the middle. When you slowly fade them out, it animates the image.. Very, very beautiful.

This experimentation started from using crumpled mylar as a reflector. It just loses a lot of strength and it is difficult to get it at optimal distance. But distorting the image at the source, i get must clearer image. I add layers of cyanoacrylate, let it drip using accelerator in the other hand to freeze the shapes in place.

[wolpumba]

Can you use a lens with anti-reflection coating (MgF) to prevent sticking of silicone?

[danielpirone]

I am a noob and I know this sounds a bit bonkers - but I wonder about some out of box approaches - thinking about how fresnel work - I wonder if building a lens or mirror - micron by micron, so you can control refractive index and chromatic aberration and internal reflection
Is possible in a well appointed home shop. You and Applied Science channel could have a fun go at it !

[ebrewste]

You have a great presence in your videos. The lack of BS and depth of technical combined with an accessible persona is great. I look forward to every video, often in topics I didn’t know I was interested in .

[wyattb]

I did my own lens with epoxy by using sandwich bags and air.

[ikocheratcr]

Regarding stiffness issue, maybe use a CNCed part the is "pretty" close to the lens you want, so the silicone pour you need is just a thin layer. I know you will hit the issue of silicone not adhering to this stiff base, so maybe CNC also some "holes" (|like a mesh), so silicone has a lot of mechanical grab.
Really nice that you share the issues you found, I think that is the most interesting part. When things work, those are more boring ;)

[RamiRouhana]

A small technical note, epoxy is not exactly a plastic, it is a polymer but it falls under the thermoset group. It actually doe not melt nor smears if polished. It is possible to polish it. Check any example of polished carbon fiber composite.

[HuygensOptics]

Hi Zach, nice work. I was wondering what your ultimate goal is exactly. If you want to make spherical surfaces, you should not start with a spherical mold, because then you will end up with an oblate elliptical surface, due to the volume shrink. So for making spherical surfaces, you need to start out with some kind of parabola/hyperbola. By the way, if you want to make epoxy/glass fiber mirrors of high quality (for example for telescopes), maybe reconsider: it is technically virtually impossible. I can mail you a few references if you want.

[CraigAndera]

Anyway, I've always really enjoyed your videos. I particularly love the ego-free way you approach the tasks, where failure is an acceptable outcome and the learning is the main objective. Sciencing! Great stuff as always.

[manyirons]

Since silicone bonds to glass, why not use a glass plate as a backer when you're casting the upper half of the mold? Then it would support against sag when casting the lens.

[excitedbox]

You could polish the lenses before hand to remove some of the imperfections. Then you will want to spin coat a low viscosity mold release to get a smooth surface. degassing the resins premixing will let you degas it better/more gently once it is mixed. If you sputter coat the glass lens with silver as a mold release you could get a perfect surface. Since it has no silicon in it, it wont stick.

[JonS]

Instead of sputtering metal on to the glass lenses, you could try an oleophobic coating that once applied and then polished off leaves a nanometer-scale fluorine-based layer on the glass. Try talking to Aculon in San Diego.

As you alluded to, you probably want to minimize the thickness of the silicone molds and bond them to a steel plate with cooling channels so you can actively cool the parts while curing.
For the single-step molding, you can use an optical grade silicone (yes, you can make silicone lenses, I've seen a few very nice examples).

A longer runner might help reduce the stress at the gate location.

The chromatic aberrations you are seeing are likely caused by dispersion. The epoxy you are using likely has a high dispersion (low Abbe number). The common method of canceling dispersion is to combine high and low dispersion elements (classically this was a flint and crown glass element).


Do note that your lens copies won't match the glass originals, even if you perfectly copy the form and eliminate all stresses, as the refractive index and Abbe numbers will be different from that of the glass.

[titter]

You have a injection molding machine? More videos on that please.

[uwezimmermann]

That's a really interesting project!
I guess the biggest step is the stress-free curing of the epoxy so that you get rid of the birefringence.
(just a small comment: the noise gate on your audio is a bit disturbing when listening with headphones. Don't be afraid of a bit of background noise...)