Uploaded March 2013 | Updated September 2026, 8 hours ago
Water as a semiconductor
So for those who, like most, have forgotten mots of high school:
A semiconductor is something between an insulator and a conductor, it conducts, but not much. We usually add other substances to these to create "doped" semi-conductors that allow us to build logical gates.
KEY: by applying energy to a semiconductor some can conduct MORE electricity, unlike metals which operate in reverse.(more heat less conductivity)
A transistor is the key building block needed to build computers, it is simple a "gate" or "control valve" that allows one signal to control another.
Background Story:
So I am now working the WK JIA office, and my desk mate is a great fellow from Brazil named Lucio, he is a man of many passions, but his main thing now is designing and building 3D printers.
Lucio and I were spit balling about wether the plastic filament could be replaced with solder or other metals, whether you could print underwater to increase cooling speed, etc..
In playing with this idea I started thinking that metal could be printed in 3D underwater, eliminated the cooling time problem, then I thought "but that would make some disgusting polluted water".. i mulled this over a while and got to the idea.. "freeze it, water doesn't conduct".. but then "oh, there will be metals and ions in it, yes it will"... but.... "will the lattice hold them in place?"
And then the "click"... "if the latice holds the ions in place, then by adding energy, conductivity will decrease"
Failed experiments and keeping it simple
We tried some simple rough designs which didnt work and started trying to find ways to decrese gap size and simplify production to the point where we could test the theory.
The first few tests failed producing some interesting results, but finally it worked. Fails were mostly due to..
heating element too large (and coils which shorted between the loops.. poor planning hehe)
testing for resistance instead of testing a DC current voltage (which just made the results messy)
Finally, back at the office, it worked, using the 6th design. The details of the design are outlined in this guide to DIY reproduce the experiment.
Problem, barriers, concerns..
small variance for large input
Hopefully reducing scale, and tweaking variables such as which ions to use, and maybe even other liquids/crystals)
Electrolysis
We may produce, in this example, HCl and NaOH, but hopefully, because they will be contained in a small bubble of water and heat within the ice, when the bubble collapses(heat dissipates) these will cancel each other out again... hopefully.. they are acid base, so they should "like each other" i would think, But I don't yet have equipment that could test this.
Hydrolysis
We may produce hydrogen and oxygen, this would be harder to "undo"... but why worry about a "byproduct that is a clean fuel".. if this happens perhaps it should be harvested.
I used to much salt, there is way to much conductivty at the baseline, I will now try with lower concentrations of salt which should allow the water to crystalize much more fully and increate the resistance of the baseline.
Conclusion
No conclusions yet but many cool ideas
if a configuration can be found that can lower heat waste of logical gate by maximizing energy waste on electrolysis, then perhaps we finally have a "good reason" to dump power into hydrolysis. compute and hydrolyze a the same time.. two birds with one stone
maybe there are other cool "byproducts of computation" we can find by musing other materials, solvents, and solutions.
A satellite that "eats ice", uses solar panels to power its "ice cube processor", and the burns the "cpu exhaust"(hydrogen and oxygen) for propulsion.
This is essentially the way biological brains work as well. They pass ions back and forth (think of it as packaging up each electron to an atom and then passing it, it creates a whole different set of properties compared to just electrons)
3D logic, printing logical gates with a single material, the possibilities are endless.. at least in my little fantasy world. we will see how far I can take it.
If you think there is a reason why this shouldn't work, or something I have not accounted for, or found a flaw in the experiment, please contact me at vasten@dscript.org
Try it for yourself. I should work at almost any temperature, but if not frozen, then any gases could escape the system. The hydrogen and oxygen should not be a problem, but the chlorine gas would, I suggest you try to keep it frozen so that way if you do make any gases they should be trapped, and the acid-base pairs will hopefully cancel out..
DIY Insructions : dscript.org/thermal-water-ion-transistor
Full article : dscript.org/ice-cube-transistor
Water as a semiconductor
So for those who, like most, have forgotten mots of high school:
A semiconductor is something between an insulator and a conductor, it conducts, but not much. We usually add other substances to these to create "doped" semi-conductors that allow us to build logical gates.
KEY: by applying energy to a semiconductor some can conduct MORE electricity, unlike metals which operate in reverse.(more heat less conductivity)
A transistor is the key building block needed to build computers, it is simple a "gate" or "control valve" that allows one signal to control another.
Background Story:
So I am now working the WK JIA office, and my desk mate is a great fellow from Brazil named Lucio, he is a man of many passions, but his main thing now is designing and building 3D printers.
Lucio and I were spit balling about wether the plastic filament could be replaced with solder or other metals, whether you could print underwater to increase cooling speed, etc..
In playing with this idea I started thinking that metal could be printed in 3D underwater, eliminated the cooling time problem, then I thought "but that would make some disgusting polluted water".. i mulled this over a while and got to the idea.. "freeze it, water doesn't conduct".. but then "oh, there will be metals and ions in it, yes it will"... but.... "will the lattice hold them in place?"
And then the "click"... "if the latice holds the ions in place, then by adding energy, conductivity will decrease"
Failed experiments and keeping it simple
We tried some simple rough designs which didnt work and started trying to find ways to decrese gap size and simplify production to the point where we could test the theory.
The first few tests failed producing some interesting results, but finally it worked. Fails were mostly due to..
heating element too large (and coils which shorted between the loops.. poor planning hehe)
testing for resistance instead of testing a DC current voltage (which just made the results messy)
Finally, back at the office, it worked, using the 6th design. The details of the design are outlined in this guide to DIY reproduce the experiment.
Problem, barriers, concerns..
small variance for large input
Hopefully reducing scale, and tweaking variables such as which ions to use, and maybe even other liquids/crystals)
Electrolysis
We may produce, in this example, HCl and NaOH, but hopefully, because they will be contained in a small bubble of water and heat within the ice, when the bubble collapses(heat dissipates) these will cancel each other out again... hopefully.. they are acid base, so they should "like each other" i would think, But I don't yet have equipment that could test this.
Hydrolysis
We may produce hydrogen and oxygen, this would be harder to "undo"... but why worry about a "byproduct that is a clean fuel".. if this happens perhaps it should be harvested.
I used to much salt, there is way to much conductivty at the baseline, I will now try with lower concentrations of salt which should allow the water to crystalize much more fully and increate the resistance of the baseline.
Conclusion
No conclusions yet but many cool ideas
if a configuration can be found that can lower heat waste of logical gate by maximizing energy waste on electrolysis, then perhaps we finally have a "good reason" to dump power into hydrolysis. compute and hydrolyze a the same time.. two birds with one stone
maybe there are other cool "byproducts of computation" we can find by musing other materials, solvents, and solutions.
A satellite that "eats ice", uses solar panels to power its "ice cube processor", and the burns the "cpu exhaust"(hydrogen and oxygen) for propulsion.
This is essentially the way biological brains work as well. They pass ions back and forth (think of it as packaging up each electron to an atom and then passing it, it creates a whole different set of properties compared to just electrons)
3D logic, printing logical gates with a single material, the possibilities are endless.. at least in my little fantasy world. we will see how far I can take it.
If you think there is a reason why this shouldn't work, or something I have not accounted for, or found a flaw in the experiment, please contact me at vasten@dscript.org
Try it for yourself. I should work at almost any temperature, but if not frozen, then any gases could escape the system. The hydrogen and oxygen should not be a problem, but the chlorine gas would, I suggest you try to keep it frozen so that way if you do make any gases they should be trapped, and the acid-base pairs will hopefully cancel out..
DIY Insructions : dscript.org/thermal-water-ion-transistor
Full article : dscript.org/ice-cube-transistor
