Update: 11/15/2007.
For the past, hmm how long has it been -- four months I suppose, I've been running the cell as usual, but even lazier. I realized that the diagram has too many heating steps. It takes a lot of fuel to heat water! What I've been skipping is the hot salination, realizing that it accomplishes little. I've looked at the phase diagram (sort of; a mutual solubility plot of NaCl vs. NaClO3 vs. temperature curves) and seen that some more chlorate should precipitate when NaCl is dissolved in heated solution. But not much seems to drop out after. I also noticed likewise that little salt dissolves, so not only am I not putting much additional chloride back into the system, it also implies that the liquor, after the chlorate crystallizes out, is already rich with chloride! This seems amazing since it's been in the cell, but it must be that chloride is converted at about the same rate as water is converted and evaporated, such that the concentration remains about constant.
Meanwhile, the chlorate concentration rises, and I usually end up with crystals in the cell. In fact, I've been running for about two or three days, topping off (with liquor from the Tank) about every 12 hours, and stopping when the top (despite the moisture-conserving wadding in the top of the cell) crusts over.
To improve the quality of crude NaClO3, I've been doing this: the fresh liquor sits for 24 hours, at which point most of the chlorate has crystallized. (Disturbing it now results in fine grains, because it's still supersaturated. 48 hours really are needed to do a good job of crystallization, but I can't spare that much time, I need to crystallize more by then!) Since the particulates have settled over those 24 hours, the solution on top is mostly clear. I pour that off, then give the solids a stir. This loosens the caked-together crystals and stirs up the settled graphite oxide. Then I pour back portions of the clear supernatant, stir it in and pour off the goo into the filter, taking with it the fine particulates: mostly graphite. Some fine chlorate grains do go along for the ride, so I'll dissolve them out of the filter later. The drained crude sodium chlorate has a dark green and black color, much lighter of course than the pitch black stuff I started with.
Periodically (every week or two), I've been recrystallizing the crude, moist material. This involves estimating the amount of water needed (not much, about half its weight: the stuff is pretty damn soluble at 100°C), boiling, scooping in the material and then hot filtering it. It is something of a chore, as it takes a long time to heat up, especially since sodium chlorate gets so cold as it dissolves (it would make an excellent ingredient for cold packs). I want to add the material to the vessel slowly, so it doesn't build up on the bottom and cause a hotspot. Then I have to pour it into the filter and wait, which means it cools down and starts crystallizing, so I have to gather some of the liquid and reheat it and keep filtering. The whole process takes a few hours of attention.
That's an awfully ominous header. But why? It's dead, Jim! I have since worn out both of my 1.5" square x 18" long graphite anodes. Here's a pic of the gore:
The first anode I used is on the right. Since I didn't control the level as well, its wear zone is more gently tapered. Not only that, but it suffered from a vein of porosity on the right corner (as oriented in the picture), corroding the copper plating from beneath. I had to replace it once during its lifetime.
The second anode, on the other hand, suffered from no such problems and its plating remained in excellent condition throughout its life -- only tarnish on the surface, the best you could hope for in such a corrosive environment. It broke at a somewhat larger diameter, my fault. I could've otherwise gotten another week or two out of it. At the time, the first anode was about the same dimension, and had the notable quality of being full length, so I put it in for another week until it got eroded to the thickness seen here. It spontaneously broke up when I went to check the water level, so it was pretty thoroughly done with.
The copper plating, by the way, went very nicely with a crude bath of copper sulfate, a dash sulfuric acid (putting pH about 1 or 2) and a few drops of Tarn-X (which contains thiourea, a known brightener), operating at a mild current flow. 24-36 hours built a good layer with ample thickness (0.010" or so? -- I'll peel it off and measure it eventually) and excellent adhesion.
Here's the cell casing / cathode. It got pretty gnarly after all that time, but much of it washed off: just caked-on salts. I gave it a washing and it looks pretty good now. The steel plate I welded for the bottom is also in good condition, considering it's been basted in salt for eight months. It's rusty, but not pitted or eaten through -- certainly owing to its cathodic protection on the inside.
The inside, around the top, does happen to be quite rusty. It's stainless, not stainproof steel, and it's been exposed to moist chlorine and salt spray. Understandable, and the cathodic protection, though it doesn't do much above the liquid level, does protect the rest quite well.
Here's all the chlorate (minus about a pound elsewhere, minus a few more -- read on) I've collected from this cell. The crude sodium chlorate (bottom right) weighs 8.025 kg, including bucket and water. It's probably 80% solids, so more like 6.5-7 kg actual solids. Above it, recrystallized sodium chlorate in bags, totalling 7.830 kg. To the left, recrystallized potassium chlorate, 1.637 kg. This is popular stuff, and I've heard it gives clean potassium purple in a flame test -- free of sodium. Finally in the lower left, crude potassium chlorate totalling 3.82 kg. I've also sold (via mail: both sodium and potassium chlorate are shippable as ORM-D for less than 25 pounds/package) 5 kilos (total) of both. That's a grand total of 26 kg, or 58 lbs mixed chlorates.
If I assume I've been using an average of 200W to produce this, then over 8 months = 240 days = 5760 hours, I've spent 1152 kWh, or roughly $115 worth of electricity. If I sell all 58 lbs for $7/lb, I could turn $406 in revenue, or $291 in profit (if I were paying the electric bill, neener neener!). 1152 kWh at about 4V is 288 kAh = 1.0G A*s = C, or about 10.7 kmol. This is electrons, and six of them go into one mol of chlorate, or 1790 mol. One mol of chlorate weighs 83.5 g, so I would've produced 149.5 kg chlorate ion, or with cation, around 205 kg of sodium and potassium chlorates. But I actually produced 26 kg, so efficiency is again around 13%, as I estimated earlier. Humph, that's still pretty shitty, and matches my earlier estimate.
Well, I still have the power supply and cell and all that. A generous friend is supplying titanium strips and potassium hexachloroplatinate (K+2[PtCl6]2-, man, the complete ionic formula takes a lot of tags to write in HTML!) to plate them with. Ooh, platinum. I've done trials already and titanium can be plated with platinum dumbfoundingly easy. With some small platinized titanium strips I've already made about a hundred grams worth of perchlorate. As soon as he delivers the stuff, I can get to work. I'll be back posting about it here as soon as I get some results! *Wrings hands madly*