Mystery Voltage: Investigating a Strange Phenomenon in MHD Testing

[artis]

So something weird just happened along the way in my experiment, here is a little background of what I'm doing.

I have created a roughly 3x5mm inner diameter rectangular shaped channel for MHD liquid testing for one other project of mine, it's a simple conduction type MHD pump where the sidewalls are made out of Polycarbonate aka plexiglass and the bottom and top are made from platinum plated copper electrodes, now the copper was first nickel plated and then platinum plated but that is just a nuance of chemicals.

So I have this setup as described , now I have wires attached to both electrodes, one wire goes directly to a screw terminal, the other wire goes through a copper sheet of which one side forms the electrode within the MHD channel while the sheet itself just continues out of the channel as a conductor, on the sheet I have some neodymium magnets and a steel cover plate for B field direction.the magnets and the sheet of metal both are electrically connected to the copper plate so they can be though of as essentially one wire.

Now here is the weird part.I used some water (ordinary tap water) to test out my channel for any leaks and how well the seals work. Now I added the water and just out of curiosity used a multimeter to test the resistance of the water, this made some small current flow through the water, nothing much.

Now after that I just left it there, then I somehow probed with a voltmeter the electrodes (just fooling around while some other work piece was drying) now I saw a voltage on the voltmeter, about 60/70 mV DC. The water was already drained from the channel just some water moisture maybe left inside. So as I saw this voltage I was a bit confused, I used a jumper shunt cable with alligator clips to discharge the PD across the channel electrodes, I measured again and the voltage climbed back to it's previous level.

Then I left the jumper cable connected for some 30 minutes as I went to eat, came back took the cable off and measured the voltage again and again it showed about 60 mV DC. Now this was getting really weird.

So I basically just put the jumper cable across the electrodes again, measured with another meter that there is a short circuit across the electrodes and there was, so I left it there for some hours, came back took it off measured the voltage and again I see about 60mV DC!

I switched the meter on before I removed the jumper cable, once the cable was still on the voltmeter showed 0 volts, once I removed the cable the voltage climbed rather fast back up to about 60mV DC, I left the tester on and took another tester switched to amps and measured current , the current was about 10uA, when measuring current voltage dropped to some few mV only, then after it climbed again back.

I tested the voltage with 3 meters all show the same reading, I have now left the electrodes short circuited for many hours and each time I measure there is still at least 40 to 50 mV on DC voltage across, and it slowly climbs up.
Can anyone explain what is going on here? I have never seen a capacitor where when you short it out for more than 10 minutes and do it frequently it keeps on still having voltage.

I attached a simple schematic of my setup.

 

[Cutter Ketch]

Hmmm ... I will speculate and you can see if anything rings true.

First, the conductivity of plexiglass is not zero. It’s a huge resistance but not infinite. Your two plates constitute a very small capacitance. I feel sure that the time constant you see when you change the configuration is the RC time constant of those two contributions.

The larger question is where does the EMF come from? You have several contacts with dissimilar metals with different work functions. I believe you have essentially made a thermocouple, although I’m haven’t exactly worked out why you wouldn’t still see that when they are shunted.

Anyhow, it’s a idea to contemplate.

 

[artis]

I will do some tests putting the device outside where it's some 20 degrees colder to see whether something changes, it is definitely not capacitor action because by shunting the opposite electrodes all PD should vanish as it does but then as soon as I remove the shunt the voltage climbs back up again and rather fast. So something is literally generating this voltage/current

 

[Cutter Ketch]

Thermocouples don’t change if you change the temperature of everything. See the wiki article.

 

[artis]

I know that is why I tried temperature differences, so just to let it out there, the logic as far s I understand it behind a thermocouple is electron diffusion in metals that have different thermal energy along the length of the metal, so in other words the thermocouple cannot have the same temperature or the same conductivity across it's entirety, now in my case my device has the same temperature, and it's not even a real thermocouple because even though each electrode is platinum plated they don't have a junction point , in other words they are not electrically connected , at best they can form a capacitor with both electrodes being the plates and the plexiglass sided 5mm wide channel being the dielectric and yet I cannot discharge the thing , I left it shorted overnight and the voltage is still there, and here is the really weird moment , please don;t take this as an ambush attempt at perpetual motion but I put the device outside where it's about 20 degrees colder than in my room, it is about 5c,
after a while I measured the voltage across the electrodes and guess what, the voltage climbed , in room temp it was about 60mV , now outside it was 150mV, I measured the amperage and it showed about 12uA, while measuring amperage the voltage dropped to about 12mV.
I short circuited the plates again and left them for half an hour , after measuring again the voltage was about 120mV but it did not go lower than that.
This seems really weird because how can decreasing system heat energy increase system electrical energy?

I guess while building one experiment I have now accidentally made a whole new experiment that needs to be investigated, I don't know which way to go now to be honest, whether start a new thread or try to understand more and then seek an explanation.

 

[artis]

Also PS. I apologize I made an error in my first post, the neodymium magnets are electrically isolated from both the steel flux plate and the copper sheet forming the electrode.
The flux plate itself is electrically connected to the beginning of one of the electrodes but only in one spot and the rest of the surface area of the plate is not electrically connected to any other part of the circuit in other words it does not form a closed circuit of any sort so can be disregarded for the sake of understanding what is going on here.

 

[artis]

I think that there might be one other possibility , a chemical reaction , although I can't imagine what kind because the elements I have used there should not react with one another,
As I've said the channel is made out of platinum coated copper electrodes, a ordinary household silicone sealant applied to the sides of the platinum plated electrodes and squeezed from both horizontal sides by plexiglass.
Some minor water vapor might be left in the channel from when I was testing the channel for leaks.Another interesting point, I took a heating fan and heated (modestly) the plexiglass as well as the copper wires leading to the electrodes etc, upon heating the wires the overall voltage dropped, from about 60mV to 30 and less mV, when I stopped heating the voltage climbed back to about 40mV and now stays there, I will put the device outside for night to see whether upon lower temps the voltage would rise again.