modified: Sunday 27 December 2020
Diodes part 4: Glowing ionic diodes recreated & my ionic thesis research
- Diodes part 3: a working diode steering circuit
- Diodes part 2: Homemade aluminium diodes
- Diodes part 1: Homemade copper diodes
In 2019 I did my undergraduate thesis on the home-made ionic diodes that I've previously blogged about, with the hope of finding a way that some form of homemade (low-technology) ionic amplifying device could be devised. I didn't succeed, but I found out about a whole pile of cool stuff.
For those not familiar: if you can make certain types of amplifying device (eg inverters) then you are most of the way to making devices that can do any computation. Traditional IC technologies, like the many families of CMOS, are very performative but are a completely unapproachable technology for most students and experimenters ($$$$$+). You simply can't "make your own IC at home" -- the closest I have seen are people making some individual transistors (after lots of effort, investment in chemicals and furnaces). You can also solder off-the-shelf parts onto a PCB to make an "IC", but this provides minimal educational value for actual IC development and has a wildly different list of tradeoffs.
The devices I looked at have absolutely garbage performance (speeds in single-digit Hz at best, lifetime in minutes) but also had some mild promises of achieving transistor-like functionality. I still have hope that someone will devise a method of making home-technology ionic integrated circuits, but I never cracked the grail of transistor-like devices.
My full thesis PDF goes into great detail of a variety of topics and experiments. In this blog post I'm just going to summarise some more of the interesting things and show pretty pictures.
If any links to PDFs or other resources die (especially the 1920's magazines): please tell me, I have backup copies of everything.
There are lots of errors and informalities in my thesis, my passion isn't in formal writing and I'm a one man band. Nonetheless I will appreciate it if you take the time to read my stuff and provide discussion, criticism or feedback. Keep in mind that many of my peers did their thesis entirely in simulation, I was lucky enough to be able to negotiate a topic that was much wilder and open-ended.
Recreating the glow
I managed to get the aluminium foil and bicarb soda diodes to glow! Homemade kitchen-ingredient photon throwers, no fancy equipment required.
In this first image I have aluminium foil tape glowing inside a pool of electrolyte in a drinking glass. (i) shows with power off, (ii) shows with 100VAC (50Hz) applied:
Very interestingly: my whitish glow colour was completely different to all of the literature which claimed it should be green.
The full circuit is two back-to-back aluminium electrolyte diodes (see full thesis PDF), with the second electrode not within frame of the photos above. Lifetime was measured in hours if operated correctly, however they would slowly dim.
I adapted this from a drinking glass to a planar style with the goal of imitating IC-style construction. Aluminium foil, sticky tape and a clear plastic slide (aka 'overhead transparency slide'):
Note the slightly different colours on the exposed sections of aluminium. These are the oxide/hydroxide layers that grow after you put a drop of electrolyte (water + bicarb soda) on the contacts and slowly apply an AC voltage (starting from 0V, finishing somewhere around 100VAC).
This planar style also glowed reliably once I had some parameters down (mainly purer water). Here is a photo taken through a cardboard tube, with a little bit of light leaking in around the edges. Again (i) is power off and (ii) is with 100VAC+ applied:
I have previously only been able to find a single photo of this phenomena. I was pretty chuffed when it worked, more so when I found it repeatable :)
Mentions of this glow are otherwise mostly limited to 1920's magazines like (page 48 of PDF) Radio News for August 1926 'Chemical Condensers of Large Capacity' by Clyde J Fitch and (page 29 of PDF) QST 1921 September 'Operating Notes on Electrolytic Rectifiers' by Roy Atkinson.
I made (and destroyed) a lot more of these devices than shown here. They have many limitations:
- dim, can't be seen in daylight
- need about 100VAC (some literature claims much lower is possible, perhaps 30VAC?)
- need very clean water ("demineralised" from a store should work well enough)
- need to be formed with AC, but can then be run with DC (temporarily, until one of the aluminium contacts poisons and dims)
- can't be used as part of a larger IC made using similar materials, as they require much greater than the breakdown voltage (of similar individual ionic diodes) to light.
See my full thesis PDF for more chemistry, operations, theory & history of these devices.
Transistor proofs: methodology and misdirection
I was really disappointed by a lot of sources. Not just random websites or magazine articles: published papers didn't seem to be much better.
Ionic devices were mostly abandoned in the mid 1900's due to the advent of more promising solid-state technologies (like germanium & silicon), so prior material on ionic devices is very sparse. In recent years some more research has picked up, but many published papers and websites that initially show promise end up being useless for my goals, eg:
- Things that "transist" (change impedance) by some definition, but in ways that are useless for making chainable logic or even basic amplifiers.
- Devices that look like they might amplify, but upon further inspection are actually something simpler (eg a resistive voltage divider).
- Under-described experiments with miraculous results that no-one has managed to recreate, eg transistors made of nothing but ice and metal contacts.
- Explanations that use words like 'cold fusion'. Yes.
Chapter 5 of my report is dedicated to questioning research and claims around ionic transistors. If you have a good grasp of discrete transistor logic and discrete amplifiers then this may be amusing to read:
I go into a lot of detail of what is and isn't useful in a transistor-like-device if you want to make logic circuits or amplifiers:
"This device is likely a transformer" is about as passive-aggressive insulting as I think I've ever been to a totally imaginary device. Nonetheless it's very important to properly contemplate & criticise all of these fictional examples, otherwise you end up accidentally convincing yourself that your have created a useful amplifier (like I did at one point with one of my experiments).
I was really eager to try and recreate an experiment by Letaw and Bardeen (minus the massive quantities of mercury, which for some reason my supervisor didn't want me to have). Sadly in the end I realised that their own data didn't support their device actually being a useful or functional amplifier in any way (section 5.7.1). I was pretty miffed.
Alternative technologies for home-technology ICs
It's always worth considering that your current path isn't the best path.
A lot of traditional transistor designs simply don't work at a macro-manageable scale, they need micron-thin insulators or active regions. I explore scaling several of these ideas up to macro ('kitchen-technology') scale in section 6.2 of my report. Interestingly: ionic transistor research (at the nano scale) has yielded working devices over the past few decades, so there are some less mainstream ideas worth trying to scale up from there too.
Chapter 8 of my thesis report outlines some crazier ideas to try. Resonant tunnel-diode logic concepts could potentially make logic circuits out of the ionic diodes I already have (they have some funky behaviour when driven to higher voltages, including some hysteresis). Bubble-control transistors are an idea where gas-bubble positions are abused to modulate fluid conductivity, ie turn the 'bad' electrolysis I get on my contacts into something useful):
Ideally I would avoid this whole wet-state-electronics concept, but it was the most promising path I had thanks to the working ionic diodes I had previously developed.
Pretty construction methods and various oddities
Making a single functional diode was possible with a (relatively unreactive) coin, sticky-tape, aluminium foil, bicarb soda and water:
The above shows an over-saturated solution (white). My later tests used much more controlled concentrations, so the fluids appear as transparent water instead:
Interdigitated gold electrodes were made on clear slide transparency (ie the clear A4-sized pieces of plastic you print onto for overhead projectors). The fingers are made of gold-leaf foil (very cheap) and double-sided tape. Cheap kapton clone tape was used as an insulator. Ordinary clear sticky-tape works just as well, but I found photo examples were easier to make with the coloured kapton:
Note that the two sets of gold contacts shown above are slightly different colours. The photographed specimen had been used under DC conditions, which affected each electrode differently.
I didn't have automated sampling equipment in my primary lab (ie my home), so the actual test jigs looked a lot more primitive then you might expect:
I also followed some tangents into making the other components of ICs, such as resistors, using easily available materials and methods that were somewhat analogous to actual IC masking processes. Here is an example using a mixture of tube-silicone (used on gutters & bathrooms) and powdered graphite (used as lock lubricant) to make resistors:
These resistors were pressure sensitive, so I imagine they could make interesting sensors. Depending on graphite concentration and track geometry I was able to get anything from hundreds of ohms to megaohms:
Mal-forming the oxide/hydroxide layers on aluminium contacts would lead to poisoned contacts:
Finally my attempts to recreate the glow were not initially successful. I instead recreated what some of the early 1920's authors described as a failure mode called 'fireworks' (from bad/impure ingredients):
A few weeks ago I woke up in the morning and realised with a bit of shock "Hey I never put this stuff on my website, did I?". Woops.
Eventually I might get back to doing some experiments in this area, but for now I'm snowed under with projects and other things I want to write about.
Hopefully someone else tries to recreate the glowing effect and shares their results. If someone can achieve a different colour again or success with lower driving voltages then it would give more clues as to the method of light production.
I still dream of one day being able to lay down some materials by hand (or similar) on a flat sheet of plastic to create a collection of functioning transistor/amplifier/logic devices. So far I've had a lot more success than I expected with my diode steering circuits and I was quite shocked to find non-linear resistors were dead easy to make ; but a 'transistor-like-device' is a still a very big step.