Old School Potassium

I eventually found an OTC bottle of KOH, but never quite got around to trying this.

I've recently gotten the bug again, and, inspired by the work of BromicAcid and others here, decided to take a swing at it. I had a few disadvantages: I'm a bit short of platinum servingware, and significantly more cautious than I was as a teenager. But I also had a few advantages: the postings here, a nice plexiglas shield, that extra caution I mentioned, and a surplus variable-output power supply that can be set to current or voltage regulation. This last seemed like a good thing; with 300V or more compliance, it could start current flowing even through a relatively high resistance, then automatically back down the voltage once a conductive channel was established, minimizing the chances of (a) no result or (b) potash popcorn.

Since I have no platinumware, I used a thin sheet of steel (purchased long ago from a hobby-store metal kiosk) as the base/cathode, and a thin strip cut from that sheet as the probe/anode. I found a bottle of KOH as weird little Hershey-kiss pellets, 3-6mm in diameter and 2-3mm tall. The pellets had whitish rims (hinting that they'd absorbed some CO2?), but weren't caked together at all; apparently the bottle had been well-sealed.

I placed one pellet with its flat base on the steel sheet. I pushed it around a little bit, and noticed that it left a bit of a trail -- already trying to deliquesce? We're in a warm and humid spell here, and the dewpoint was probably well above 10 C, so it wouldn't surprise me. I breathed across the sheet, and it fogged, except for a teardrop-shaped area surrounding and trailing the pellet. (I'd never seen that particular demonstration of aggressive dessication before!) After a few more breaths, I figured it was probably moist enough to start conducting.

I set the power supply to constant-current mode at its minimum power, around 50mA, and touched the probe to the top of the pellet. The voltage quickly dropped to 10V or less, a channel melted/fused from the contact point to the base, and a small amount of gas started to effervesce. The steel discolored slightly, and gray balls began to form around the bottom of the pellet. At this point, I turned up the current to 100mA. After a few minutes, I started to notice some purple streaking (ferrate?), and occasional wisps of white smoke. I saw a few very faint sparks. I saw larger sparks when I pulled away the probe, as the PS shot up to 300V to try to maintain a constant current. The sparks were mostly lavender, but that's not surprising if you think of them as small arcs drawn from a potassium compound -- I don't think they indicated the presence of metal, especially since they seemed to originate from the anode.

I was hoping for Davy's "minute metallic globules", but if they formed, they were too minute for me to spot. But something was certainly happening.

After ten or fifteen minutes, I decided to quit and see what I had. I took the plate to a basin, ran water over it, and was rewarded with a FAT purple spark. So, yep, there was some metal there!

Things to do next:

Gloat -- it's nothing compared to Bromic's performances, but it was a nice step for me, and I didn't damage anything.

Figure out the theoretical yield of potassium in micrograms per milliamp-minute, to avoid unrealistic expectations or unwelcome over-achieving surprises.

Try much higher current levels. This PS only goes up to 300mA, I think, but if I only need 10V or less, I might try something with a 12V supply and a light bulb as ballast.

Think a lot about further steps, possibly including modifications that would let me isolate recognizable quantities of metal.

It's been awhile since I did my in-depth forum search. Have many of you tried this particular method? Any other helpful suggestions or warnings?

Congratulations !

I tried it once before in a u-tube to keep the anode and cathode separate, it was horrible. See the picture below but the oil spattered everywhere and turned brown then black, I guess molten NaOH and hot liquid sodium metal was too much for it.



[Edited on 1/10/2008 by BromicAcid]

I couldn't quite figure out what hot alkali or alkali metal could do to paraffins, but I've got a newfound appreciation for the oxidative power of an anode in molten KOH.

I was definitely making ferrate. After running the power for a few minutes, purple streaks started to build outward from the anode; when they reached the cathode, I think potassium generation ceased.

I ran with a pair of steel nails for electrodes, and higher current, and got jets coming out with purple/yellow flame at the end -- potassium vapor, maybe, or just finely divided potassium riding along on the hydrogen. Again, though, when the purple discoloration got to the cathode, the jets stopped.

When I dumped the purple mass into water, it formed a purple solution, but it also fizzed gently. I thought perhaps it was finely dispersed potassium generating hydrogen, but now I think it was oxygen from the ferrate decomposing. The purple solution fairly quickly went to a ferric-oxide brown.

Tonight, I tried again, this time using carbon electrodes. I found a jar of one-inch carbon planchets (apparently used for EM sample prep) at our local surplus-junk dealer -- I put the KOH pellet on one as the cathode, and broke another in half to use as the anode. I tried running 200mA, and got the jets again; I backed down to 100mA, and got fairly steady current/voltage, with some occasional spitting. At times, a silver-greyish material would bubble out of the molten part, and turn more grey. I suspected this might be potassium, but I would've expected it to burn in the air, not just tarnish.

After five or ten minutes, I stopped, chipped the mass off the planchet, and dropped it into a bucket of water. BIG pop and purple flash. Woohoo!

I repeated the run, this time dropping the mass, planchet and all into the water. Big flash and pop when it hit, with one little purple ball-lightning zipping across the water to the right, and another hitting and sticking to the side of the bucket on the left. The planchet, with most of the unreacted pellet, sank to the bottom, where the pellet dissolved; after a few seconds, a small purple light came chirping up from it.

This is fun. I'm not sure what I'll try next -- probably doing it under paraffin. I'll also think more about setting up to record video.

Sodium with carbon electrodes, or lack thereof

I did see this post. I paid particular attention to the part about Kohlenoxydkalium. If I were to try the procedure, I would probably start with KOH instead of K2CO3, both for safety and because I already have it on hand.

I'm concerned about safety anywhere above the submicro scale, though. KOH tends to attack glass anyhow, and running the vessel up to red heat would surely accelerate the process.

Yeah, and I'm not too concerned about the safety issues of hydrogen coming off the mixture, given that potassium vapor is the desired product.

Given that the KOH and Mg will both be molten at the reaction temperature, I wonder how finely they need to be ground and mixed? I don't have any Mg turnings or powder, but I do have some thin sheet. My KOH is in large prills (4-8mm diameter). I may try this initially with one or two prills and a corresponding amount of Mg, considerably less than a gram total.