Molding a Microlens Array

Microlenses are tiny optical devices used to focus light onto cameras and fiber optics. While debugging a different project, I stumbled on a technique that made hemispherical depressions. So I made some microlenses.

CONSIDER SUBSCRIBING 🥰

==== Laser Parameters ====
Had some requests for more details about the laser parameters. I'm not sure how much this is dependent on my particular machine vs. generally applicable, but in case it helps:

- 50W fiber laser (non-MOPA). Generic out of china, JPT fiber laser inside
- Power ranged from 2-20%, 40kHz frequency, 500mm/s to 3000mm/s, 63mm f-theta lens, single pass
- Most good ablation craters were in the 10-15% power range. Speed from 1500 to 3000mm/s depending on how much spacing you want between craters.
- Power in range from 6-10% tends to create dimples and spikes. Below 6% and there's no visible effect
- 0.05mm stepover on hatching
- Wafer was 111 orientation thick (500um?) wafer

==== Equipment and techniques ====
- Thermo Phenom XL scanning electron microscope

==== Timeline ====
0:00 Intro
0:44 Metal Assisted Chemical Etching
2:25 Strange observation
3:42 Hypothesis #1: heavy oxidation
4:32 Hypothesis #2: Si-Ag "doping"
6:19 Hypothesis #3: indiscriminate etching
7:26 HNA etchant
8:26 Safety Disclaimer
10:22 Etching with HNA
12:16 PDMS molding
14:42 Do they focus light?
16:01 Molding problems
17:47 Other fun phenomena

[BreakingTaps]

**Addendum**
- I meant 4mm^2 (2x2mm) in the intro, although I later measured and it's actually 1x1mm. Oops! Sorry for the confusion, I phrased it poorly :)

[AppliedScience]

I liked the whole journey! Maybe those laser crater spikes would make nice field emission sources. A big array could really pump out some electrons!

[smartereveryday]

This is better than reading a paper. Great work!

[smartereveryday]

I wonder if you you were to tilt the specimen down just a bit before you spin it by placing it on a slight wedge. This may make the mold release to run "down" and flow out of the crater. Like this:

⦣ Angle of plate 🠒 Spin force


Just an idea.

[kaysb]

Anything can be handled in that one shed in Australia ;D

[MaterialsSci]

Hello! In your 'molding problems' section, you mention dimples near the top of your lenses. If indeed the problem is from an over-accumulation of mold release, you could try sputtering a thin film of gold onto the silicon surface and using that as a sort of mold release because the silicon will not stick well to the gold without a proper adhesion layer; additionally, the PDMS will not stick well to the gold without a specific chemical treatment (chemical treatment to make the gold adhere to the PDMS very well can be done really easily with a not very dangerous chemical called MPTMS). If the gold sputtering route does interest you, using the MPTMS surface treatment, you can make static mirror arrays as well as strain-tunable mirror arrays and strain-tunable plasmonic devices because of the metal first-surface on the PDMS.

[thomasesr]

You should shine a cheap laser pointer through the glass slide microlenses and project it on a wall to see if it does anything interesting.

[johnathancorgan]

In Australian sheds, as long as you don't contaminate your working area with anything yellow, HF is fine to handle.

[izzieb]

The oblong micro lenses aren't bad - they're just anamorphic micro lenses.

[HuygensOptics]

Very nice results Zach! I guess if you really dive a little deeper into the process, you could probably make much better lens arrays, maybe it is even possible to make well-defined aspherics. But then you need to be in control in every parameter of the process. Developing new technology into something useful is always a long and bumpy road.

[sakelaine]

6:34 it's a fundamental effect, the reason it's not in papers is because most of the people writing papers would take it for granted that their audience knows that catalysts just speed up a process that would happen by lowering the energy cost through mediation. So if you're in a situation where using a catalyst would help, remember that the desired effect is still happening at a slower rate. For most stuff this doesn't matter at all; at the scales you're working with it's important and I'm glad you figured it out.

[SeanLinsley]

this is awesome, feels like an Applied Science video

[aetius]

Btw, you could try using ammonium fluoride as an etchant instead of HF.
It is much safer to work with and the fact it has a lower etching rate could remove the debris without eating too deep into the creaters.
Also maybe going for an evaporation coating for the mold release agent could be better than spin coating, for instance just heat up gently a silicone oil with your slide just above the surface.

[ChrisTuttle]

Ever since the first demonstration of light field photography, I have been trying to figure out how to make my own micro lens arrays. Your video covers so many aspects of whats involved that “finally” I think I understand the magnitude and scope of this task. This is an awesome video! Thanks! I am looking forward to your follow up videos regarding this project.

[olpizl]

That „Explosions&Fire“ reference. Love it. Great video btw!

[aetius]

Great video (as usual), I found quite funny that you dont have acetic acid but have no problem getting HF :)

[rocketman]

The mold release buildup problem might be improved by -
1. Lowering its viscosity with a compatable solvent.
2. Spinning it with the apparatus upside down after applying mold release with a spray bottle from underneath.
3. Make sure the release agent is properly set in order to prevent it slumping into the hollow. So store it upside-down until certain that it has cured.
° Upside-down application might produce a thickening to occur around the outer rim of each micro bowl, and that might be able to be manipulated so as to result in a superior end product.
Cheers. 👍

[fastindy]

10:21 "There's always a silver lining"
I see what you did there

[edgeeffect]

That's the first time I've heard someone say "do your own research" without hearing a looney conspiracy theory dropping out the other end. ;) ;) ;)

[sublucid]

Can you make super hydrophobic surfaces using the spikes (or some other laser-pattern)?