Sciencemadness Discussion Board

Thoughts On Anodes

 Pages:  1  2    4  ..  12

Swede - 23-10-2008 at 15:00

Hi Xenoid... my initial attempts were with a commercial MMO anode offered by Northstar Pyro. The cross section of that Ti strap (gets calculator) is 22.3 mm squared. I have no idea if the strap was Ti or a Ti alloy, and I'm not sure if it makes a difference with regards to conductivity. Anyway, I used a quality Cu connector... everything was pretty optimal for carrying current, and those straps heated to well above 100C at 50 amps. I found out only after I attempted to clean a bit of salt creep with a water spray, and the water flashed to a boil.

Things were improved considerably when I mounted a heat sink on the strap. Interestingly, these were the cathode straps. The anode (MMO) straps were much cooler. I don't pretend to understand the mechanism there, but it was distinct... the all-Ti cathodes were hotter than the Ti + MMO. Anyway, I just wanted to throw that out as a consideration. Sometimes, commercial MMO cells are designed to carry less current than we are considering pushing through these setups, and the "choke point" seems to be the hanger cross section, between the Cu power cables, and the immersed electrode.

Dann2, I'm using a homemade spot welder now. Previously, I had been using a TIG welder on the all Ti cathodes that I had been making, and it was working fine, but I had not tried TIG welding MMO plate or mesh. When my mesh arrived, I dusted off the spot welder to see if it would work. I first tried it on two pieces of Ti and was stunned that it worked as well as it did. For the MMO mesh, I scraped the coating off the mesh junctions, the fat areas where the rows meet, on the areas under the strap; cleaned it a bit with alcohol, and fired away, and it worked very well. Just one weld was secure, and a typical strap to mesh interface is about 15 separate welds. It's clean and fast, no overheating, and appears to be a good way if an experimenter is doing a considerable quantity of these.

Here is the spot welder:

My talents lie more in fabrication than the chemistry behind this process, although I do have a chemistry background, and am trying to catch up on the many years of effort by dedicated hobbyists to produce a simple and cheap anode for perchlorate production.

Xenoid - 23-10-2008 at 15:09

Huh! Tried to edit my post and it was deleted!!!

Anyway the gist of it was;

I have a whole bunch of these 24V water inlet solenoid valves which I got from defunct electronic drive washing machines. They are rated at 60oC and 0.2 - 10 bar. They appear to be made from glass reinforced nylon. They are probably not suitable for concentrated acid but should be OK for dilute dosing, say in the pH range 3 to 7. The outlet (cell inlet) will need to be connected to some narrow bore tubing to restrict flow.

[Edited on 23-10-2008 by Xenoid]

[Edited on 23-10-2008 by Xenoid]

SolenoidValve.jpg - 23kB

dann2 - 23-10-2008 at 15:24


Article on Graphite erosion attached.
It discusses the chemistry involved.


and then I thought why not, A bit of dreamy flower power.
The white things in the pic. are butterflys.
so I post:
It from a Russion republic where their booster rockets/tanks etc fall off their space rockets they are launching. They are scavanged by locals, or not no locals. Worth a fortune as there are made from Ti, Nb, Hf etc etc and big too.
From Magnum photo's.

[Edited on 23-10-2008 by dann2]

Attachment: Graphite wear in Chlorate process.pdf (578kB)
This file has been downloaded 1442 times

dann2 - 23-10-2008 at 15:30

and again

rocket_waste.jpg - 187kB

Xenoid - 23-10-2008 at 15:57

@ Dann2

Geez! - I found one of those in the back paddock a while back! I wondered what it was - Ti, Nb you say! :D

tentacles - 23-10-2008 at 17:19

Swede: I did find a solenoid, an all PTFE job with 1/4 FNPT fittings. On your cathodes, when making the electrical connection between the Ti and wire, did you sand the Ti oxides to clean Ti before securing the connection? I've got some silver (Ag) paint that I plan on using as a conductive interface between the Ti and copper.

From what I've been able to determine, Ti is about 25x more resistive electrically than copper. Thing is, though, at the low voltages we're working with, the resistance can be pretty substantial before much heating occurs. A 1/4" bolt is roughly equivalent to #2 wire (though not at the threads) which in copper is rated to 181 amps chassis. If we derate Ti by a factor of 7, which is probably enough at 3-6V, we get 25 amps or so. A 1/4" conductor has a cross section of 31.75mm^2.

My math could be way off here, but the electrical resistivity of Cu is 16.78 nano-ohm-meters and Ti is .42 micro-ohm-meters. (.42 / .01678 ~ 25)

My opinion on the straps that are of considerably less cross section (see pool chlorinators) is that they are immersed, and so kept cool. The cross section of the bolt (1/4? 5/16?) is likely considerably more than the 10mm^2 of the strap, thus helping to prevent the plastic from overheating. Also on the dry side the resistivity is reduced by the brass nuts and such.

edit: I did find one reference to temperature derating of Ti wire anodes, and it specified (for 1.5mm Cu cored wire) 21A @ 25C and 7A @ 70C. That could have more to do with the coating survival time than anything, though.

[Edited on 23-10-2008 by tentacles]

Swede - 23-10-2008 at 18:46


Swede: I did find a solenoid, an all PTFE job with 1/4 FNPT fittings. On your cathodes, when making the electrical connection between the Ti and wire, did you sand the Ti oxides to clean Ti before securing the connection? I've got some silver (Ag) paint that I plan on using as a conductive interface between the Ti and copper.

Probably a good point. I'm usually pretty conscientious about stuff like that, but I honestly don't remember expending much effort on cleaning the Ti before mounting the Cu cables. I've been using this type of cable mount on most of my cells... this is from a very early cell:

These are nice assemblies from any home supply store, and need be bolted only once to a strap. Connections thereafter happen by sticking the bare Cu wire into the body of the assembly, and tightening a screw. With a bit of care, cleaning, etc before mounting one of these, it should last forever. IIRC these were #10 Cu cables, too light.

I switched to #4 Cu welding cable, which has really fine strands and makes for an awesome connection. They sell lugs sized to fit, requiring a crimp, and it's been this type for me:

#4 barely gets warm at 50 amps, negligible voltage loss.

Anyway, what you are saying makes some sense. Even with Ti being a worse conductor than Cu, with good cleaning, I'm guessing that the heating I saw will mostly disappear.

dann2 - 23-10-2008 at 18:47


There is a device here:
for calculating resistances of wire and strips etc.

It is better to think in terms of the current the conductor is carrying.
The 'voltage we are working at' will not be 'seen' by the conductor in question if you know what I mean.
Only a small voltage (too much if it is getting hot of course) will appear accross the part getting hot/warm.

Taking a strip of Ti going to the anode that is 6 cm long by 1cm wide by 0.2cm thick.
It has a cross section area of 0.2cm squared and length is 6cm.
The (from resistivity of Ti is 0.000000554 = 5.54 X 10^-7 ohm M
Copper is 1.7 X 10^-8 ohm M.
Using the page above gives a resistance of 0.00166 ohms for the Ti strip.
and 0.0000509 ohms if the strip was Copper.

@100amps you get I^2 R watts = 100 X 100 X0.00166 = 16.6 watts for the Ti
and 0.5 watts if the strip was Copper.
@150 amps you will get 37.4 watts with the Ti. (0.249 volts dropped accross strip)
with Copper you would get 1.14 watts.
How hot the thing will get will depend on how well insulated it is.

This assumes connections are perfect etc. Bad connection loads of heat.


hope my calcs. are OK:(

[Edited on 24-10-2008 by dann2]

Swede - 24-10-2008 at 06:06

Haha! Dann2, my ability to do that sort of calculation on paper, while there, would require more effort than I was ready to give. To determine how much Ti cross section I'd need, I simply clamped various test pieces across my big power supply, and dialed in the current. With a small(ish) Ti cross section, it didn't take much current before it became smoking hot. I continued to add Ti until it was able to sustain 50 to 70 amps and not exceed maybe 60 C.

In the end, I settled on 2 straps, 1" wide by 0.062" thick, or 25mm X 1.6mm.

This was a worst-case scenario. The Ti did not have one end immersed in electrolyte, which would draw much of the heat away. But at the same time, the Ti was not encased or potted in plastic, as it would be with a cell.

I will do a test today. I'll mount a Ti strap across my power supply leads, apply current, and measure the temperature of the Ti. Then, I will thoroughly clean any oxides off the contact areas, both on the Ti and the Cu, and repeat the experiment, to see if there is any difference.

simple - 24-10-2008 at 18:59

Thinks to consider with Ti electrode design:
poor thermal conductivity
poor electrical conductivity
shunt current within electrode can corrode titanium pretty fast. normally happens at voltages above 6-7VDC

For calculations on electrical resistance, keep in mind that anodes are made using CP (commercially pure) grade of titanium.
For cathode, CP grade or one-two grades lower titanium plate will do, but may 'buckle" / bend.

Brass / copper connection to titanium is known to give problems if heats up. So must be made such a way so it will not get hot.

dann2 - 24-10-2008 at 20:15

Originally posted by simple
Thinks to consider with Ti electrode design:
shunt current within electrode can corrode titanium pretty fast. normally happens at voltages above 6-7VDC

Hello Simple,
What do you mean by this statement. Can you elaborate.
When you say 'voltages above 6-7 VDC' where are you measuring this voltage from/too.


simple - 24-10-2008 at 21:31

Hi Dann2,
'shunt current' is a leaking current between oppositely charged parts. Normally starts to effect plates by deep etching corners and edges of the plates. It is a leaking path for the energy, which is wasted. Often it it a shortcut for the electrical current between conductors, so instead of plates discharging through the main surface area it will leak to nearby part. A bit difficult concept to explain, how do you post pictures here, I may photograph sample for you.

Voltage measured across electrodes supply terminals. There are three main variables here, voltage, resistance of solution and distance between plates. All three will effect "trigger" voltage for shunt current. Normally can be avoided by design of the electrode.

simple - 24-10-2008 at 21:38

Trying to load the image

IMG_3258.JPG - 44kB

simple - 24-10-2008 at 21:39

Here is corrosion caused by shunt current. You can see how the edge of the place was "chewed".

Edit: I should mention that this happened due to imperfection of the assembly. Error was made on the assembly line and overlooked during test's. In bad cases you can loose significant part it titanium mass and cause structural damage. At the same time you 'dilute" titanium in solution

[Edited on 25-10-2008 by simple]

IMG_3259.JPG - 43kB

tentacles - 25-10-2008 at 05:46

I've heard of this happening in some patent or other - the biggest factors are electrode spacing (the closer together, the more likely it is to happen) and anode voltage. It's very unlikely to happen at lower anode voltages, but I believe around 6.5V you start to overcome the electrical insulation of the Ti oxide layer (of any exposed Ti, like edges). It shouldn't be a big deal if we keep the anode spacing reasonable.

Didn't someone have a paper or something that covered anode spacing / efficiency?

Swede - 25-10-2008 at 06:08

I'd think closer spacing would reduce the voltage needed for a given current. Reduced voltages are desireable, correct? The single MMO anode I had up until recently, I have been running at 5.5 to 7.5 V which gave me anywhere between 40 and 60 amps. I've gone as low as 7% Cl- concentration, but no lower. I haven't seen a hint of damage yet, but that doesn't mean there is none.

I'm thinking if we increase electrode surface area, and strap cross section, and find a spacing that will allow 60 to 80 amps at say 5V for chlorate, then the electrode set will run relatively cool and have a long life. I'm suspecting most cases of damage result from an excess current density, a low ion concentration, or a combination of both.

dann2 - 25-10-2008 at 08:34

Hello Folks,

@ Simple
Welcome to the never ending Chlorate/Perchlorate making + anode making discussions.

Are the anodes in the picture MMO (for Chlorate).
Do the cathodes go in between the anode sheets?

Could you 'repair' an anode (if you knew it was being eaten away by a shunt current problem) while it was still in the cell by shutting off the power supply and then slowly applying an increasing voltage to the cell so that any parts of exposed bare Ti (no Ti Oxide) plate would get a chance to form a protective Oxide coat.

Most cells run in the garage have quite a large spacing between anode and cathode (compared to industrial cells). Therefor I think (guess) that the problem of shunt currents damaging Ti substrates in home cells will probably not occur.
The closer the electrodes the higher the shunt current risk as Tentacles said?
Also the higher the solution resistance the higher the shunt current risk?

Closer spacing of electrodes will give a smaller voltage accross cell (for a given current) but as you decrease the spacing (as Simple said) the risk of shunt current corrosion will increase. Industrial makers are obsessed with voltage accross the cells as this saves power. The home maker will not worry about an extra volt or so accross their cells.

Will keeping all corners on anodes as rounded as possible also reduce the risk of shunt current erosion?

Are modern Perchlorte cells pH controlled?

Sorry about all the questions and thanks for the pict's.


[Edited on 25-10-2008 by dann2]

Swede - 25-10-2008 at 10:47

Great! :P Now we have "shunt current erosion" to worry about... as if there weren't enough issues with this process.

I think what a lot of this boils down to is too much power. Many of us (me included) are taking 2" X 5" mesh anodes and jamming 60+ amps through them. If we either increase the size of our systems, especially electrode area, and/or decrease the current a bit, erosion, overheating, and other problems will be minimized.

Splinky - 25-10-2008 at 11:41

Wow, it's been a while since I've looked around on here...

I just set up my potassium chlorate cell with an MMO anode/cathode set from northstar pyro, I made a run to a hardware store and picked up a 40 lb bag of KCl so I should be set for a while.
My setup is a simple two liter soda bottle with the top cut off and the anode/cathode suspended in the solution with a 6 amp 6 volt battery charger power source. I didn't make any sort of venting apparatus aside from open windows, and in the NaCl test runs I have done I haven't experienced any Cl problems aside from a faint smell for the first half hour or so. After that, all I get is a faint bleachy smell. Is there any problem with that, or could it still be giving off harmful amounts of chlorine without me being able to smell it?

Swede- Your blog on APC is pretty interesting, it's nice to have some sort of running commentary from someone who is putting time into (per)chlorate production. I haven't been able to even get onto the site from my home computer, but it's nice to know you post here too.

kclo4 - 25-10-2008 at 11:47

Well, I'm sure its not good to be exposed to small amounts of atmospheric chlorine, so I guess it depends on how much you care? If you can't smell it, I bet their is very little in the air. Just make sure most of the chlorine gets dissolved so you don't have to worry about it - plus, it will increase your yield.

Splinky - 25-10-2008 at 11:53

I just went out and checked, there is no smell and very little gas production at the anode. I guess that means it's mostly dissolving.

Xenoid - 25-10-2008 at 12:12

Originally posted by dann2

Are modern Perchlorte cells pH controlled?

@ Dann2 - is not 1960 "modern" enough for you! :o

The Djvu version of the Perchlorates Monograph, you yourself posted here a while back has tables (pp 74 - 75) and discussion (pp 85 - 86) on pH control. In general, slightly acid to neutral conditions are preferred, with pH 6.6 - 6.8 being used in commercial operations.

This is only for current efficiency reasons and is not a problem for amateurs using platinized Ti of PbO2 electrodes. Indeed, the stability field of PbO2 is expanded at higher pH. From memory, my assorted perchlorate cells approached pH 12 after extended running. I guess this is why my various baked MnO2 anodes performed very poorly in perchlorate (and less poorly in chlorate) cells. At high pH, MnO4- (permanganate) (recall Purple Haze) and MnO4= (manganate) are the stable (aqueous) species. pH would have to be kept below about 4 for an MnO2 anode to remain stable. I am not sure if a pH this low is feasible in a (per)chlorate cell. In a strongly acid solution an MnO2 anode would rapidly dissolve (Mn++) in the absence of the highly oxidising conditions on the anode surface.

All this assumes I am interpreting the Pb and Mn Pourbaix diagrams correctly! :)

dann2 - 25-10-2008 at 17:16

Hello Folks,

@Xenoid. That clears up the pH thing in Perk. cells.
When I operated a Perk. cell with a LD anode the pH went to 10.8. Lots of Lead was depositing on the cathode. Perhaps this would not have happened if pH was kept at around 6 or 7.

@Splinky. If you are not controlling pH then there is no way Chlorine can escape from the cell after it has started to operate for an hour or so.

Regarding the shunt current problems (the new spanner in the works), I don't think it is a problem for us at all.
There are two types of cells used in Industry. Monopolar and Bipolar. The Bipolar cells use a row of electrodes 'in series' in the same cell. There are no seperate anodes and cathodes as such except the two end electrodes which have the + and - connected to them. All other electrodes are an anode on one side and a cathode on the other side (of the same MMO (or whatever) coated plate) with no actual electrical connection going to them! It is a set up we will never be using. I won't anyways.
All our cells are Monopolar and therefor will be very unlikely to get effected with this problem.
The electrodes pictured by Simple are (I think) from Bipolar set up's.
(I think it would be near impossible for the actual anode or cathode to get effected by shunt currents in a Monopolar cell. Perhaps some metal used as part of the cell construction might get eaten by a shunt current problem)
Bipolar cell pictured below.

Also the linked document explained it better than I can:

There is also a paper discussing the prediction of Shunt Currents in Bipolar cells in the References wanted in the Reference section. It is essential bedtime reading. A absolute must for all garage chlorate making Guru's. :P


[Edited on 26-10-2008 by dann2]

Bipolar.gif - 9kB

simple - 26-10-2008 at 14:20

Dann2, thanks for welcome.
Pictured anodes are RuO based and electrode is designed to generate low concentrations of sodiumhypochlorite in NaCL solution. It is bipolar and revere polarity assembly.

Main rule for shunt current it that resistance between any two oppositely charged and exposed parts of electrode should be significantly more that between working surfaces of the electrodes.

Reading more in to this thread, i can also recommend keeping titanium conductors / assembly made of pretty high grade, say CP or Grade1. pH variations will corrode low grade titanium alloy. If you check grade specs, you will find a lot of aluminum in the lower grades. This makes if fairly unstable.

Rosco Bodine - 27-10-2008 at 23:00

Coaxial electrodes are looking better all the time hmmmm?

12AX7 - 27-10-2008 at 23:07

I don't think so. I have a suspicion that my horribly low current efficiency (~15% even at the best estimate) is due to the horribly large cathode, being the cell itself.


Rosco Bodine - 27-10-2008 at 23:14

Some of the most efficient cells are coaxial cylinder form
cells with 2mm interelectrode gaps having a pumped
electrolyte. The pump rates are so high on some of these
that they actually have to consider flow erosion from
the friction of the moving electrolyte, as much as any chemical wear.

I think to get acceptable cell performance efficencies
along with pH control the working gap is going to have
to be in the 6-10 mm range max.

[Edited on 28-10-2008 by Rosco Bodine]

12AX7 - 28-10-2008 at 01:05

For relatively large pipes, bigger than the gap, that would be reasonably effective, because the areas are then similar.

A continuous pumped cell would pwn. Brine in one end, chlorate out the other; oh, but considerable hydrogen must come out, too. You could pump from the bottom and let the hydrogen rise, although I'm not sure how you prevent it from sticking (passivation). That much flow rate would do it, but inevitably you will produce far more gas than liquid volume if you don't let it go, so I guess it will proceed as any other cell does, the liquor always being well mixed and circulated from a tank, say. Oh well. Just another approach to the process, then, nothing so remarkable as a one-step continuous process.


Swede - 28-10-2008 at 05:57

Those that have read my blog know that I am going for a circulating system, not in the sense that chloride goes in at one end, and chlorate is extracted at the other; rather, it is a closed loop using two cells, one small (containing the electrodes) the other large (being a cooled collection chamber). A variable pump powers the flow. The goal is to create a system that forces crystals to form (and be collected from) the cooler cell.

One of the benefits of this is that the volume is variable. Additional containers can be daisy-chained into the system. The other benefit is that it may be possible to recharge the hot effluent from the small electrode cell with additional chloride ion. I have been putting a bit of thought into how to do this.

One thing that came to mind... the typical 40 lb sack of KCl has large, heavy crystals. I was thinking about a PVC water filter cannister from a typical home improvement store, highly modified. The cannister is charged with large KCl nuggets, and the hot effluent from the small cell (probably at 60 to 80 degrees) is flowed through these crystals, proceeds up and out, and is routed back to the large, cooler collection chamber.

I am using potassium salts, not sodium. Two problems come to mind. Dissolving KCl appears to be slightly endothermic. As the hot, chlorate-laden electrolyte contacts the KCl nuggets, it cools quickly, and chlorate is dumped in the chloride cannister. The second potential problem: assuming it works, you now have recharged electrolyte carrying extra chloride and chlorate flowing to the collection chamber, whereupon it cools even more, and KCl falls out, contaminating the already-crystallized potassium chlorate. With the salts intermingled there, it would be likely that the KCl would NOT redissolve, and your product would be heavily contaminated.

What do you guys think? Chloride cannister = waste of time and resources?

tentacles - 28-10-2008 at 10:04

I think putting the chloride "recharger" after the collection phase would work best, but with a several degrees preheat to keep more chlorate from dropping out. I wonder what the equilibrium of K chloride to chlorate is in this situation. Will more Cl dissolve and precipitate out the chlorate until there is none, or some percentage, left?

There's probably a chart somewhere that's very hard to decipher.

Xenoid - 28-10-2008 at 11:30

Originally posted by tentacles
There's probably a chart somewhere that's very hard to decipher.

The KCl - KClO3 mutual solubility data is available on Dann2's site here;

I think I plotted this up by hand, on graph paper, at some stage but I can't find it now!

Originally posted by 12AX7
I don't think so. I have a suspicion that my horribly low current efficiency (~15% even at the best estimate) is due to the horribly large cathode, being the cell itself.

Or perhaps by your horrible, corroded, twisted wire connections, Tim - :)

12AX7 - 28-10-2008 at 12:40

No. Kirchoff's node law. Bad connections produce a voltage drop, and the voltage drop across connections pales in comparison to the ballast resistor I used. Voltage drop costs power efficiency but not current efficiency. Besides, even if I made that accounting mistake, assuming current efficiency were a typical 50%, the measured voltage drop of 2V would only be 60% loss, dropping power efficiency down to 30%, still more than a factor of two away from the 13% or whatever it was I observed. And the error bars on that 13% are much smaller than a factor of two (and have the statistical support of two measured runs!).

So all that remains in proving this hypothesis is to build a cell with evenly proportioned anodes and cathodes. Maybe next summer I'll pick up some more graphite bars (or some chloronitrosylplatinate or whatever the good stuff is and slap that on titanium) and cook up a five gallon bucket of black-stained goo.

As for process, I'm convinced the best method is a batch/continuous sodium process. Potassium chlorate is low solubility and will crystallize everywhere you don't want it. So don't use it. Sodium chlorate is high solubility, but sodium chloride has no change in solubility vs. temperature, so there is no endothermic dissolution to worry about and it forces sodium chlorate to precipitate on cooling.

To use a two tank method, I think the best approach would be a circulating pump in the electrolysis cell, drawing or forcing liquor through a basket loaded with rock salt. A small faction of the outgoing liquor is diverted to the crystallizing tank. It must be a small fraction, because if the cell's contents are constantly changing, there will be an extremely small deltaT and crystals will be forced to form everywhere. Imagine it this way: you only want to crystallize as much solution as there has been chlorate added to it. It's okay to have a somewhat higher rate, in which case you're just pumping a little more solution than crystallizes, fine, but not too much that it lowers the cell temperature / raises the crystallizing tank's temperature and not too little that the cell's concentration becomes supersaturated and crystals form in both anyway. So there's an ideal middle range which is good.

I think a small circulating pump in the cell will suffice for mixing, while natural convection into the crystallizing tank will suffice for its flow. Use a relatively low aspect ratio for the tank, so it is able to cool effectively. A narrow rectangular box shape would be excellent.


dann2 - 28-10-2008 at 17:20

Hello Folks,

Originally posted by tentacles
I think putting the chloride "recharger" after the collection phase would work best, but with a several degrees preheat to keep more chlorate from dropping out. I wonder what the equilibrium of K chloride to chlorate is in this situation. Will more Cl dissolve and precipitate out the chlorate until there is none, or some percentage, left?

There's probably a chart somewhere that's very hard to decipher.

Would agree with Tentacles. If you were to (loosly) thermally attach the recharging chamber (which comes after ppt chamber) to the anode/cathode chamber this would keep it warmer than settling chamber. What you want IMHO.
A very small pump is required for the circulation, you may be better to have the pump set up on a timer that pumps every half hour for a few seconds (say).
What about a reaction chamber?? You are going to control pH I presume? If so it will be advantageous to have a large chamber (after anode chamber) that is warm/very warm to allow the chemical formation of Chlorate. You will add acid into this chamber. The anode/cathode chamber will be OK for a reaction chamber IMHO if you do not want to add (yet another chamber).

You could get rid of KCl addition chamber by adding finely powdered KCl to anode/cathode chamber and stirring gently.(magnetic stirrer) every day or so. The amount of KCl will be found out when you have discovered you current efficiency (how much KCl you are destroying) per day.
I presume there will be no evaporation from the system. If there is (say you have a reaction chamber (will be warm 60, 70C) with a fairly large surface area open to the air), then you can easly add KCl via a concentrated solution of the stuff to the anode/cathode chamber.

What will come out/stay/dissolve in the solution is depicted by ye old mutual solubility diagram.
Thanks to Watson.Fawks, who turned up an explanation of how to use (bloody) mutual solubility diagrams, you can educateyourself on how to interpret them here:
See the first few pages of 'How to disign fractional crystalization processes'.
There is a mutual solubility diagram for KCl and KClO3 here:
Unfortunately it is in mass ration, you want mass percent in order to figure out what will/will not stay in solution as per the Watson.Fawks article.
Does anyone know how to convert the graph? The axis are easy but it is difficlult to know where to draw the lines. Would need to sit down and study the thing. The table that Xenoid showed you has the data in mass % (and you could draw a (lovely :D) graph).
There will be no problem getting the KClO3 out as it is less soluble and also has a large slope in it's solubility graph.
Me thinks that if you just set the system up and let it belt away for a day or two it will work as you require. It will take it a day or two to reach its 'working point' where K Chlorate comes out nicely and KCl goes in nicely. You may have to insulate the whole thing (or keep shed reasonably at the same temp.) as wildly varying outside temperatures may upset it.

Regarding efficiencys: The anode to cathode spacing debate has more to do with voltage drop accross cell which is related to power efficiency (very very important to the industrial Chlorate makers). Not very important to you.
The largest increase in current efficency (time efficiency) for the amateur can be got by pH control and (since pH control is being implemented) using a seperate chemical chlorate formation chamber which is hot or use a fairly large anode/cathode chamber (ie. the anode/cathode chamber is also the chemical chlorate formation chamber).

The next biggest step up for to get high CE is to keep the Chlorde level high.

From then on you will just be tweeking at the edges for little CE return.

My two cents worth on Cell chem. is here:

Regarding reduction at the cathode:
From what I can figure out from all my bedtime reading adventures (read: sad fuck) Iron cathode are good at reducing Chlorate and hypochlorite back to Chloride. (What we don't want). Ti is not as good at converting Chlorate and Hypochloride back to Chloride. (that's good for us).
Therefor area of Cathode (large area gives low current density and this increases reduction, read this somewhere, can't remember where I'm afraid) is even less important to us when you use Ti. (I am not trying to say that area of Cathode is a really big deal anyways). Chromates are added to cell to stop reduction in industry. Perhaps you will get away with adding no Chromates if Ti cathode is used?????????
Perhaps if you use a very small Ti Cathode it would really reduce reduction without having to add Chromate etc.
Industry would NEVER use a very small cathode as it would raise voltage accross the cell a bit (they would'nt tolerate that at all).
The less additives the better. Then again Chromate also buffers the solution which is an advantage.

Well, I'm glad I got all that off my chest!!!!!!!!!!


[Edited on 29-10-2008 by dann2]

cutting through the fog

Rosco Bodine - 28-10-2008 at 20:52

Here's some test data on a large coaxial cell .


Attachment: US3385779 High Efficency Coaxial Chlorate Cell.pdf (127kB)
This file has been downloaded 939 times

Swede - 29-10-2008 at 05:15

pH control is important... pH controllers are common and easily available. I picked ip a 1/4 DIN industrial unit off eBay that will hopefully work (He said "Tested and working"). Unfortunately, I don't have a probe yet, and the probe selection has me a bit stumped. Even a "cheap" probe, isn't. I'd rather spend $150 or $200 now than have to buy 4 or 5 $50 probes.

I sent an email to Hanna tech support. In the email, I asked them what they would recommend based upon the species and the environment these electrodes are forced to operate in, continuously.

This is the list of assumptions I've been working with:

- Double junction
- need NOT be fast response
- need NOT be extremely accurate, +/- 0.2 is more than adequate
- PTFE diaphragm? I am not sure if this is needed
- Either glass, CPVC, or PVDF (Kynar) bodied; anything else will suffer
- A threaded body is best, but I can adapt almost anything
- ATC NOT required
- Highly salted electrolyte but few if any solids; flow low or zero
- Temperature from 25 to 80 degrees C.
- Primary range 4.0 to 11.0

I have a few process electrodes in mind, but I'm curious to see what they come up with, if anything. The easiest to adapt would be one with a 1/2" or 3/4" tapered pipe thread, or the metric equivalent.

Swede - 29-10-2008 at 06:19

Hanna emailed right back. They didn't comment on anything, simply pointed me to their HI1000 series of electrodes, which is pretty much what I presumed. Double-junction, PTFE diaphragm, gel-filled, PVDF body. I really hate the thought of coughing up so much money. But I am really concerned that anything less than a high-quality industrial electrode is going to be poisoned quickly or physically die after only a few runs. We're talking DAYS immersed in that stuff.

dann2 - 29-10-2008 at 06:25


It is amazing the near absence of Graphite erosion that industry can achieve.
4kg Graphite per ton Chlorate produced, @30C. That's 4 grams per KG :o, and not all of the four grams will appear as black gunk. Some will have gone off as CO2.
We are doing something(s) very wrong when using Graphite. No pH control I guess is by far the biggest issue.

Would you not need to mention the more aggressive species that the pH probe will be in contact with + the fact that the temperature will be up at 60C. These factors will be a pH probe killer, perhaps not?
It can be hard to extract the info. you need from people selling stuff.
While my suggestion will not be in line with your (to put it mildly) exacting standards, you could set up the system and just have a constant feed of acid into cell and check pH twice per day (or more at the start).
I did this with a cell before and the pH stayed in the wanted region remarkably well.
The rate of acid addition was 0.35ml per amp each hour. Acid was 12% hardware store (used for cleaning cement) HCl. pH stayed in the region of 6.7 to 7.
You would need 17.5ml per houre if your cell was running at 50 amps. Of course you cell might need more/less. 17.5ml per hour sounds alot. 0.42 liters per day.
You still need to get a pH probe and may neet to take a sample out of the cell both to let it cool and to stop voltage gradients in the cell from putting the pH reading haywire.

Which reminds me of another point for your pH probe in the actual tank. Will it operate with the voltage gradient that is in the tank?
Mine would not. Samples had to be removed. If you had a seperate 'chemical chlorate making vessle' (hot), there would be no voltage problem is this vessel and a permanently installed pH probe shoud work OK there.
You may also be able to shield the probel with a metal shield to keep voltage field from probe in anode/cathode chamber.


[Edited on 29-10-2008 by dann2]

Swede - 29-10-2008 at 09:32

Dann2, I love your contributions to this discussion, thank you.

I did mention "heavy concentrations of chlorites, hypochlorites, and chlorates, as well as aqueous chlorine." Hopefully that registered with him. Materials-wise, a PTFE diaphragm and PVDF body would be ideal.

One of my goals for my current project was automation, not because it's needed, but mainly because I enjoy the gadgetry, and the challenge of putting it all together. I agree that periodic check of the pH with a standard (cheaper) system would be much easier, but eventually, I would like to let the controller do its thing. I was going to place the probe in a separate tank from the electrodes, so there should be little or no problem with interference. Placement of the acid injection relative to the probe will be an interesting problem. Too close, and you'll have very rapid cycling of the valve or pump that administers acid. Too far away, and you might overdose with acid. This will depend also upon turbulence in the tank, with higher probably being better. I plan on using aquarium air pumps for stirring.

I am a bit concerned about the volume of acid you mentioned. 1/2 liter a day seems like quite a bit. I'm hoping to use upwards of 75 amps with this thing, and I don't know that I am going to evaporate 1/2 liter per day. Maybe concentrated HCl, to keep the volume low, might help.

Also, for an acid "pump", I was considering an airtight vat, and letting a powerful aquarium air pump pressurize the vat, which would force the acid out a line and into the system. When the controller turns off the air pump, the acid flow should cease. Sort of a poor man's dosing pump.

[Edited on 29-10-2008 by Swede]

dann2 - 29-10-2008 at 12:14


Thank You Swede, it's a pleasure!
<table border=3>
<tr><th colspan=4> pH control on current efficiency and acid consumption</th></tr>

<tr><td align=center>Quality of <BR> pH control</td><td align=center>pH drift</td>
<td align=center>Current efficiency, %</td><td align=center>35% HCl consumed<br>kg/ton - NaClO<SMALL>3</SMALL></td></tr>

<tr><td align=center> Medium </td><td align=center>6.0 to 7.8</td><td align=center>74.4</td><td align=center>87.5</td></tr>

<tr><td align=center> Good </td><td align=center>6.7 to 6.8</td><td align=center>84.5</td><td align=center>62.5</td></tr>
</table><small>From: Electrochemical Technology. 6 (1968) 402. </small>

Data on acid consumption in Chlorate cells has always been hard to find.
Figures from Encylcopedia of Chemical Processing and design gives a figure of 0.996 kg HCl per 185.553 kg Sodium Chlorate produced. (They dont say if they mean 'pure' acid (35% HCl or what)

From the article below:
acid consumption is given as pounds of HCl per ton Na Chlorate . It has wild variation from 5 to 35 (average about 25 lbs) in a 9 amp cell.
In a 144 amp cell they give a figure from 98 to 19 lbs HCl per ton Na Chlorate (100% basis used as Hydrochloric acid in calculations) Electrodes were Pt.

My own cell that I controlled the pH in was an extremly simple jam jar type. Anode was Magnetite. Perhaps I could have added less acid, I am not too sure. If you add too much acid the cell will start to emit more and more Chlorine. I could not detect a smell of Cl. I quoted the figure in ml per amp per time. I achieved a CE of 60%. Magnetite is not a great anode! 3% CE with no pH control. I dont have a figure for the total amount of Chlorate I produces and the total amount of acid I added to get that amount of Chlorate.

Anyhow, the above will give you some guidelines, very very wide one's by the looks of it!!

I attach an artical about adding acid to cells. It's in Japanese. Can anyone decipher.


[Edited on 29-10-2008 by dann2]

Attachment: denki.pdf (395kB)
This file has been downloaded 845 times

tentacles - 29-10-2008 at 18:21

Keep in mind that it isn't just evaporation, but also we're cracking the water for it's delightful oxygen, so there will be water lost there as well.

dann: Your hardware store has 12% HCl rather than regular muriatic (28-31 or whatever %)? Or do you mean you diluted muriatic to 12%?

I'm still waiting on the mail to deliver the anodes Swede made for me to start testing.. Also have to come up with somewhere to *put* a cell. I just recently filled my outside storage area with 2/3 cord of firewood.

Rosco Bodine - 29-10-2008 at 18:36

Originally posted by dann2

It is amazing the near absence of Graphite erosion that industry can achieve.
4kg Graphite per ton Chlorate produced, @30C. That's 4 grams per KG :o, and not all of the four grams will appear as black gunk. Some will have gone off as CO2.
We are doing something(s) very wrong when using Graphite. No pH control I guess is by far the biggest issue.


In the thread "treating gouging rods" is some information that is probably one added factor. Such preparations for
toughening graphite is probably also useful for anodes of the GSLD type which should be workable for perchlorate.
I have always though that a coaxial electrode is the best way to go, as it offers several advantages physically and electrically. It can simplify cell construction also.

dann2 - 29-10-2008 at 19:27


My hardware store has 12% HCl acid only. Cannot obtain any higher except lab supply stuff.

I often wondered if we changed the solvent that we put our Chloride into what would happen. We would still need some water for the O2 so it would have to be miscible with water but would something weird and wonderful happen that might suit our purposes.


12AX7 - 29-10-2008 at 20:19

What solvent? Most will be oxidized by the hypochlorite and radicals, and the rest aren't even miscible.


Splinky - 30-10-2008 at 06:00

I managed to get my first batch of chlorate dry today, I was able to make a red comp yesterday out of some wet chlorate crystals, SrCO3, and powdered sugar, so it seems pretty pure.

It still has a hint of bleach smell, is this normal or should I have washed it more? I just washed it once in a large amount of cold water and kept whatever didn't dissolve.

Swede- Maybe you could have a large tube with the KCl nuggets in it and a smaller tube containing the chlorate solution running through the larger one, so the KCl dissolving would cool down the chlorate and let it fall out, like a condenser. Then push the solution back through the larger tube to restock the KCl and continue the cycle.

Swede - 30-10-2008 at 06:29

Hi Splinky - how did you harvest the chlorate? Did you start with potassium salts, or sodium? Either way, once you had your potassium salts, remember that they are VERY insoluble in cold water. If you used some sort of filter rig, once the potassium chlorate is sitting there in the filter basket, you can pour (wash) significant amounts of ice cold water on them, which will remove much clinging contaminants, and also residual chloride, while dissolving very little of the chlorate. I wash first with ice water, then with an alcohol/water mixture (also cold) as a final wash. The alcohol helps in the wetting of the crystals and also speeds drying. They should not smell like hypochlorites when dry.

It might have been your washing technique. Did you get the ice water to all of the crystals with a bit of mixing/stirring of the mass? If you simply pour cold water over the surface of the crystals, the water will follow a path of easiest flow, and may not access clumps of crystals in some parts of your rig.

I tested some potassium chlorate crystals washed in this manner, and they had a low chloride residual; they were 99.3% pure. Recrystallizing KClO3 is easy and will produce a quality product.

I like your thought of using the KCl crystal mass as some sort of heat sink! Interesting. Along those lines, I picked up what is called an "offline chlorinator" at a pool store today. Thes are cannisters designed to accept 3" chlorine tablets, and have water trickled through them to leach the chlorine, which is then injected back into the pool. This is what it looks like:

This sort of gadget can find all sorts of uses in (per)chlorate electrolysis. The plastic is designed to withstand wicked oxidizers. It is also engineered to be completely sealed. I can see something like this being a stand-alone cell (the volume is impressive, about 4 liters), an electrode cell, or a chloride replenishment cannister; any of these with just a bit of work. They can be had on eBay relatively cheaply. I'll take a look at it today and snap some pics.

@Tentacles: That sucks that the anodes have not arrived yet. Hopefully customs will release them soon. I'll bet they will be opened and inspected. Hopefully they'll repack them properly.

I am going to take my small test anode of this material and set up a perchlorate cell, starting with recrystallized chlorate, and rip some serious amperage/voltage through it to see how it stands up. Another one will do duty as a chlorate-only anode.

[Edited on 30-10-2008 by Swede]

dann2 - 30-10-2008 at 06:29

Originally posted by 12AX7
What solvent? Most will be oxidized by the hypochlorite and radicals, and the rest aren't even miscible.


I guess that's the end of that whacky idea, if solvents that may be suitable would get attacked by Hypochlorite.

Cleanlyness is next to godlyness in the Pyro. (or indeed chemistry business in general) business. It would be no harm to to recyrstallize. Be careful.


Splinky - 30-10-2008 at 06:59

When a little over an inch of chlorate had accumulated on the bottom of the cell (I started with KCl), I shut it off overnight to cool and let more precipitate out. Once it had accumulated on the bottom I poured the water out into another container and filled the cell with cold water and let it sit with occasional stirring for a while. I disposed of that water and did it again with a little less water and agitation, then let it dry by a fan overnight. I should go with more filtering definitely, I think I'll go with your process and run cold water over it for a while.

I would use the offline chlorinator as a cell, that looks to be perfectly suited for it. A three foot or so section of two inch pvc with a section of thin (cpvc maybe) half or three quarter inch running through it would probably work well as the condenser/restocker thingy.

Swede - 30-10-2008 at 09:07

Hmm, that would be one way to wash crystals, and it should have done so thoroughly, Make sure you save the was water for your next run... it'll carry chlorate with it. Replenish that water with chloride, and it'll give you a head start.

"Used" chlorate cell electrolyte is loaded with intermediate species, and significant quantities of chlorine as well. I have about 5 gallons in a poly jug over KCl crystals to replenish for the next run, and with the lid removed, it becomes unbearable in the vicinity of the jug. This is mainly due to the chlorine, but also the bleach.

Anyway, maybe someone else has some thoughts on washing. I used to use coffee filters, but when the batches approached a kilo, that was too slow, so I made a plastic filter/funnel rig. The entire batch gets dumped into this, then the cold water is applied with stirring; the holes allow the water to drain. I have never had a dried chlorate batch smell like anything at all.

Splinky - 30-10-2008 at 10:19

I reused the original liquid from the cell, the immediate production of chlorate is more satisfying.

I found a much larger jar that looks promising for a cell, will most plastics containing chlorine hold up well to cell conditions?

Swede - 30-10-2008 at 11:46

Good plastics: PVC, CPVC, PTFE, PVDF (Kynar), PET (Coke bottle plastic)

Acrylic seems to do OK, but there is definitely some fogging and crazing. HDPE lasts a few batches, but eventually dies. It is economical, though, and it is what most people use. Polypropylene for me has not done well. Nylon is about the worst, and unfortunately that's what most hardware-store hose barbs and fittings are made of. They rot/crack pretty quickly.

PET (polyethylene terepthalate) would be an awesome plastic, except most containers made of PET are really thin and flimsy, like a 1 liter coke bottle.

I made my own PVC/CPVC system, but it wasn't easy. A better bet would be one of these PVC tanks + lid:

They are not cheap but will do some serious production, at about 1 kilo per 6 liters, starting with saturated KCl. The lid will need to be modified to create a seal, and you might have to glue some bolt hold-down strips on the upper rim, but if I ever need to do another tank, I'll buy one of these rather than make one.

Another option is a short section of large-bore PVC pipe. You get a lot of volume per inch when you're talking about 12" dia PVC pipe. Remember PVC is not good above 40 to 50 C, so if you want to run hot, consider CPVC.

All of this you can ignore if you just want to make 400g now and then... use an HDPE food container, slotted for the anode and cathode.

Splinky - 30-10-2008 at 12:04

Well, my current cell is a two liter soda bottle. The other jar didn't look to good on closer inspection. I think I'll try to get a bucket that holds a few gallons so I can just have week long runs over the winter and stock up for seasons more hospitable to pyrotechnics.
The only problem I'm experiencing is the build up of junk on my electrodes and a thin layer on the stuff surrounding the cell, so I should probably put a cover on my next one.

tentacles - 30-10-2008 at 16:20

A possible alternative to your soda bottle might be a PET water cooler jug. The ones we have here are ~5 gallons and are PET of reasonable thickness.

Splinky - 30-10-2008 at 17:19

Aren't those insulated at least partially? Even with temps in the thirties at night my cell has stayed pretty warm, would higher temperatures have any negative effects or just help KCl stay dissolved?

I decided to go with the jar, it's about four liters and if it holds up well it should suit all my needs.

Are there any further precautions I should take if I make barium/strontium/calcium chlorate as well? I seem to remember reading somewhere that some MMO anodes don't do well with barium compounds, but I don't remember where that was.

Swede - 31-10-2008 at 06:07

Warm temps (to a point) are desireable. What sort of current are you seeing? With 50 amps, I had to immerse my 6 liter cell in a vat of water, and direct a fan across the surface of the water. Otherwise, the temperature would climb above 70 or 80 degrees. The evaporative cooling system took the whole assembly below 50.

If you have enough current, I don't think you need to worry about warmth.

One other consideration in large systems... solutions heavy in salts tend to form aggressive temperature gradients. If the electrodes are near the top, as most are, you will heat just the top layer, and the hot liquor will remain near the top, while the bottom remains cool. If you route some stiff polyurethane tubing, or better yet PTFE tubing (don't use nylon) through the lid, to the bottom, and inject air using a strong aquarium airpump, you will break up the layers and form a nice, homogenous liquor, keeping evolved chlorine gas in contact with the solution longer, which is desireable. You'll have a healthier, more efficient system with mixing and agitation.

Splinky - 31-10-2008 at 07:10

I'm only running it with a six amp battery charger. If I'm reading the meter on it right it says it's drawing upwards of eight amps, but it's still nowhere near the 40+ amp cells that people are making. I think I'll look around for a supply more in the area of 20-30 amps if I can find one cheap.

Would it be efficient to just have a plate of titanium/stainless steel on the bottom of the cell to act as a cathode? That would provide some agitation throughout the whole container.

Swede - 31-10-2008 at 13:33

Your electrode spacing is going to determine what current your system will create, among other parameters. For example, if your battery charger is 6V, and capable of 30 amps, but you are seeing 6 on the meter, by moving the electrodes closer together, you are reducing the resistance of the cell. The result... the same 6V will now show 10 or 15 amps.

If you place your cathode on the bottom, you might end up with too much spacing, requiring additional voltage to achieve the same current. If you can get away with it, though, it might not be a bad way to obtain some "free" circulation.

dann2 - 1-11-2008 at 12:43

Originally posted by Splinky
Would it be efficient to just have a plate of titanium/stainless steel on the bottom of the cell to act as a cathode? That would provide some agitation throughout the whole container.


Landing the cathode at the bottom of the tank will give you a uneven current density distribution on the anode.
Greater current density on the bottom of anode than the top.
If you must have a cell that is much taller than the anode/cathode assembly you could place the anode/cathode assemble into a plastic pipe that is about a cm or so off the bottom of the cell (or have holes drilled at the bottom) and a cm or so under the liquid at the top. You will then get circulation from bottom to top inside pipe (opposite outside) of cell (a toroid) via the H2 released.

According to MMO should not be used in electrolytes containing Barium or Cadmium.
According to Encyclopedia of Chemical Processing and Design, Vol 51, 1995 page146, Barium is a poison to Noble Metal or Noble Metal Oxide anodes and may even reduce there lifetimes.
According to US Patent 7,250,144 (July 2007) Fluorine additive can damage Chlorate anodes (probably MMO, since it is a modern patent).


Swede - 1-11-2008 at 13:14

Speaking of patents, I was googling "Industrial chlorate ph control" and found a new patent (Sep 08) claiming the use of phosphate as a buffer allows the industry to do away with chromium in all its forms. The inventors did several experiments, and the best results were obtained with a specially treated carbon steel cathode, electrodeposited with Mo and Fe. Phosphate ions were introduced, and apparently it did exceptionally well. A second experiment used untrreated cathodes, and still did very well.

I can see phosphate doing something to those carbon steel cathodes; "parkerizing" comes to mind. But what interested me was the use of phosphate to help buffer and keep pH under control.

"3g/l sodium acid phosphates (Na2HPO4 & NaH2PO4)" made up the buffer.

Any thoughts on this? Do you think it might be useful? I refuse to touch Chromium in any form, and NaF doesn't excite me either. How could one prepare the sodium acid phosphates? And what might it do to subsequent processing of the product?

The patent number: US20080230381

Swede - 1-11-2008 at 13:23

I'm sorry, I meant to also post this: A doctoral thesis on the electrochemistry of the chlorate process. 119 pages of doctoral yumminess. First glance indicates it will be an excellent resource for the dedicated amateur.

I wish I had written it. Enjoy!:D

dann2 - 1-11-2008 at 15:23

Hello Swede,

A nice piece of bedtime reading.
Would not a sup of Phosphoric acid (rust proofer) be OK.

Persulphates have also been used with Lead Dioxide to stop cathode reduction of products.
I don't know if persulphate buffers the solution.


Swede - 2-11-2008 at 06:44

I was thinking that the easiest way to create the buffer would be to mix phosphoric acid and NaOH and adjust the ratios to achieve a pH of around 6.0, but it would be more economical to have the mono and disodium phosphate salts at hand, determine a correct ratio, and add the salts during preparation. Where one would obtain these salts at an economical price, I have no idea. Anyway, it'd definitely be for a future experiment.

I love the way these patents say things like "current efficiency dropped to 90%; unacceptable." I'd kill to get 90% efficiency! :P

tentacles - 2-11-2008 at 07:43

Chemsavers does have the disodium phosphate at $31/500g.

Reading that hot thesis.. I wonder if that chick is hot?

I see a potential problem for pH control - "The pH of the electrolyte varies from about 4 at the anode to 6-7 in the bulk of the solution and up to high values - in the range of 13-14 - at the hydrogen evolving cathode."

dann2 - 2-11-2008 at 09:32


I don't understand what you mean by the above being a problem.

When we are controlling the pH, it is the pH of the bulk solution we are interested in. This is where 'chemical chlorate formation' will take place.
(We get no 'chemical chlorate formation' in the non pH controlled cell, only 'Chlorate made by electricity'.)
This is why, in pH controlled cell, it is best to have either a fairly large Anode/Cathode chamber (plenty of room away from the anode) for the 'bulk reactions' (chemical Chlorate formation) to take place or to have a second chamber for the 'bulk reactions' to take place.

You will always have a low pH at the anode surface and a high pH at the cathode surface. We can control the bulk solution area.


Swede - 2-11-2008 at 10:51

I would also assume that those varying pH values are in gradients... the closer to the electrode, the higher/lower the pH, and as the electrolyte moves, from either mechanical stirring or heat, it'll all mix together rapidly to an averge pH.

I have done a lot of thinking on pH the last few days. I know the industrial units have a pH electrode immersed constantly. For an electrode to survive any length of time, and not be fatally poisoned by the liquor, it's going to need some excellent (and expensive) engineering, and even then, I think the probable lifetime of a pH electrode can be measured in a batch or two.

Since those guys are making multi-ton batches, they can afford a $300 pH electrode every so often. Their electrode cost per kg is quite low. Whereas OUR cost, with the same $300 electrode, is too high to contemplate.

I'm guessing a single-junction aquarium electrode ($15) will be fatally poisoned in just a few hours. What's the answer? I have no idea. There's no point in having a controller if you have to manually immerse the electrode every hour or two, cycle in some acid, then remove the electrode to prevent contamination. An advanced setup might have some sort of PTFE solenoid, which would open at intervals, flood a pH electrode chamber, acid cycles on, then the chamber is drained... and now you have an electrode which is high and dry, also not a good thing.

I may be completely wrong... maybe a modest double-junction pH electrode will last for years with minor care. I hope so, but I've got a feeling they will die quickly as an accurate measuring device.

Maybe one way would be to monitor a cycle with a cheap electrode, and note what sort of HCL dosing schedule (so many ml/min) would keep things reasonable. Once that is known, you can probably be successful in simply replicating that dosing schedule without any sort of pH monitoring.

I've got the controller; the dosing pump is coming, and an inexpensive electrode is as well. A pH electrode is the ONE thing I am very leery of buying off eBay. They're either used and suspect, or NIB old stock, also suspect, since these things have a shelf life. That is why I am interested in nontoxic buffering like the phosphate salts, to ease the process a little bit.

Sorry for the rambling post. I'd like to get my rig running in a day or two, but I'm not ready to control pH yet.

dann2 - 2-11-2008 at 14:41

Hello Swede,

You might be pleasently surprised by how well the dose + measure strategy may work. It worked OK for me in a very simple cell. The pH started to go astray as the cell became depleated of Chloride but that is not going to happen with your system which will be very stable and predictable when you get it up and running for a few days.
Two measurements per day (I guess) would be easily enough to keep an eye on pH after stability has been achieved.

They sing the praises of Reflex probes here:

Some times in industry the pH is controlled by bubbling in Chlorine.
The pH can go no lower than approx. 4 if a dosing mistake is made.
Not suitable for the garage guru.

I have heard of a guy (long time ago) suggesting (or actually doing, not too sure) that pH could be controlled by bubbling in CO2 to the cell. Your not going to have a CO2 source around though, for to do that. He reckoned that the continous bubbling of CO2 would control pH at around neutral without needing a pH probe. Would there be enough CO2 in air?

Also regarding recharging with KCl. Would KCl dissolve in HCl acid ?.
It will depend on how much acid you have to add but if you were to add KCl to the acid (if it will dissolve), then you could kill two birds with the one stone.


Splinky - 2-11-2008 at 15:51

I don't think the KCl dissolved into hydrochloric acid would work too well as hydrochloric acid is just HCl dissolved in water. Unless the HCl significantly improves the ability of KCl to dissolve in water it would just be like adding a bit of saturated KCl solution.

I think a CO2 welding/beverage tank would hold enough to slowly bubble gas into a cell for several days, if you're willing to buy a tank (in the area of sixty bucks) you could probably manage a few runs in a large cell per charge. Refills are between ten and twenty dollars.

tentacles - 2-11-2008 at 17:04

I read through the first section of that paper today, except the bits I don't understand (about half the charts, anyways). Interestingly, it mentions commercial DSA chlorate anodes as being "worn in" after THREE YEARS of use (continual, or nearly so, I would expect). It also has some interesting info on the coatings, a RuO2/TiO2 blend. Some of them apparently have a RuSn3O4 or something dopant added. Thicker layers prove to wear out faster than a somewhat thinner layer, etc. I'm sure everyone here has read through considerable portions of it. The references will be very useful to look through, I think.

Swede, her diagram of a chlorate cell is very close to what you're building. Her test cell was 100ml cell (chamber?) volume (no idea of solution volume) and she used a flow rate of 2.2L/min through the cell chamber. That's serious flow, but some of the tests she ran were at a staggering 3.8A/cm^2! (38KA/m^2).

dann: I just sent those refex guys an email. Who knows, maybe they're suckers for people who want to make strong oxidizers?

I'm not sure I see the point of adding any acid besides HCl to adjust the pH of solution - all you're going to achieve with CO2 is to make (salt) carbonates, when what we want is chlorides. The pH goes up because the chlorine is evolving away (as we all know). Besides, you really aren't going to get much carbonic acid to dissolve/form in a hot electrolyte, without pressure. Carbonated chlorate/hypochlorite solution, anyone? Don't bump the tank!

[Edited on 2-11-2008 by tentacles]

watson.fawkes - 2-11-2008 at 18:03

Originally posted by Swede
I have done a lot of thinking on pH the last few days. I know the industrial units have a pH electrode immersed constantly. For an electrode to survive any length of time, and not be fatally poisoned by the liquor, it's going to need some excellent (and expensive) engineering, and even then, I think the probable lifetime of a pH electrode can be measured in a batch or two.
Forgive me for asking what may be an obvious question, but do you need to measure actual pH, or will some proxy for pH work? The core of a regular pH electrodes is a proton membrane, which passes only H<sup>+</sup>, and so measures pH directly. But for a known process, where all the species in solution are known, it seems that total conductivity, perhaps without a special membrane, would correlate adequately to the process parameter involved.

Idea Zero: Use ordinary conductivity. If this is adequate, why do anything more complicated?

Idea One: Use a hardier membrane with less selectivity. If you were to use a membrane that passed mono-atomic species but not polyatomic ones, this would seem to give at least some idea of the free chlorine in solution. Surely that's worth something. I can't say now whether it's sufficient.

Idea Two: Use transient current behavior. Ordinary pH electrodes use DC currents, which could be done easily in days past. What was not so easy back then was to get detailed measurement of the transient AC response; that's pretty easy to do nowadays with fast A/D converters and microprocessors. The physical principle is that the depletion region around an electrode acts like a capacitor. The rate of discharge is related to the diffusion constants (read, masses) of the species adsorbed electrically onto the surface of the electrode. So you'd first have a charging cycle that used a potentiostat to bring the electrode to a known state. The next cycle would be to discharge the electrode, <i>i.e.</i> suddenly change the voltage (probably by reference to an internal sample-and-hold, rather than solution). Observing the rate of discharge gives an idea about what species are present. A third stage cleans the electrode with AC before going back to the beginning. Different charge and discharge potentials should give a pretty good sample of the various species present. While the software to run such a device would be pretty extensive to get right, the underlying electronics, while not really simple, are still pretty modest.

Idea Three: Combine with density measurement. Apropos another recent thread, an ultrasonic range finder can be used to measure density. In combination with other conductivity measurements that are not in themselves definitive, density may tip the scales.

Idea Four: Measure electrode capacitance by resonance. Operate an electrode at some DC bias current with a combination potentiostat/galvanostat. Impose upon that bias current an AC signal (likely in the RF range, given the total charge on the surface of a small reference electrode) that uses the electrode capacitance as part of the oscillator. This yields a frequency dependent upon the local electrode conditions. Sweeping over the DC bias by controlling the potentiostat allows sampling over different regimes of ion adsorption.

Idea Five: Use a differential reference electrode which is a tiny loop of platinum wire connected to twisted-pair leads.

I'm making no claim that any of this is any better than a commercial pH probe. I'll consider this posting successful if it jogs a good idea.

DerAlte - 2-11-2008 at 23:19

@ Swede

Thanks for that excellent doctoral thesis. It's years since I made any chlorate but that work dispelled a few myths I had accumulated about it. Such as thinking that the Cl- concentration should be as high as possible. The ideas presented probably have meaning for other MMO anodes such as those discussed in this forum. I wish it had a bit more about graphite anodes, though.

The amateur does not have the same concern with power efficiency that commercial makers have, but current efficiency is important to us, because of the time taken for electrolysis. The other factor is electrode erosion. I'd love to get 1kg/ton erosion on graphite! in my efforts it seemed like 50kg/ton!

Regards, Der Alte

dann2 - 3-11-2008 at 06:39


Just in case you folks run out of reading material, I attach another Thesis.
This one was pluged on the board some time ago so you may have seen it before.

Going back up this thread a bit we were talking about having no Chromate in the cell.

On page 4 of the thesis it states:
Reduction of chlorate is highly dependent on cathode material and takes place on iron whereas
cathodes of Co, Ni, Mo, Ti, Hg and C do not show any activity for chlorate reduction [11,12].

So that just leaves Hypochlorite being reduced (destroyed) I guess.
On page 36 it states:
Hypochlorite reduction is limited by mass transport at higher current
densities, and the increase in limiting current density seen in Figure 14 may be a mixed limiting
current due to limited supply of H+ and transport-limited hypochlorite.

What that is saying (I think and hope!) is that if you have a high current density on the cathode (a cathode with a small effective surface area) then Hypochlorite reduction at the cathode will be minimized.

Yittrium is also stated to have a similar effect to Chromate and is less toxic but it would be better to not have any of these additives in the cell (and product) to start with.

About monitoring the pH.
Perhaps another possible way would be to monitor the amount of Chlorine gas that is coming off the cell. I think this would be a good indication of pH value.
Are there any cheap gas monitors that can be purchased?
I have read somewhere that industrial cells have been operated at a slightly higher pH than would be the optimum for max. CE, in order to have less Cl gas to deal with in the exhaust gases from the cells. Going more green, as it were, or literally less green (pun intended).
As pH goes up less Cl comes out.

Don't know if that qualifies as a good idea (joged) or a brain seizure but if it was easy to implement it should work.


Attachment: DSA Thesis.pdf (2MB)
This file has been downloaded 669 times

Swede - 3-11-2008 at 07:56

I like the direction this discussion is taking... altenate methods to do what industry takes for granted. I found yet another patent (I hate reading patents because it's often hard to find good nuggets in them, but they are there) that makes use of a tungsten electrode set and appropriate electronics to measure chlorate concentration, much like watson.fawkes mentions. Further, I am confident that there are ways to determine chloride concentration (end of run) conditions based upon voltage/current plots, even when efficiency may not be known. Logic says that as the species change in the system, the voltage required to create a given current (with a good supply capable of constant current operation) will follow a predictable path. What I've seen with my own system, starting with potassium salts, is a concave voltage curve... it starts high, drops, then begins to rise once more as the chloride is depleted. As WF says, conductivity may be a useful measurement that is overlooked.

My point is, with enough research and some hands-on experimentation, things like quantitative analysis, pH control, and other pressing issues, can and will be overcome.

Dann2, I laughed when you pointed me towards the Reflex electrode. I have literally been researching poisoning-resistant pH electrodes for weeks without finding anything like that. I've come up with "Triple-junction, glass-domed, PTFE membraned, long life electrodes" stuff like that, but nothing that flat-out says "It cannot be poisoned." I am going to email or call them as well and see what they have. I will not hesitate to buy an expensive pH electrode if I know that it will survive. Worth every penny, IMO. Thank you for the resource.

But for setups without a controller, I like the concept of simply adding calculated quantities of HCl, and doing periodic checks with a pH meter. I did not realize that there was a calculable quantity of HCl consumed given a known current. Since we know our current (ammeters are cheap) then we should know how much HCl to trickle in. A dosing pump set up on a timer could administer X ml per hour, and all you have to do is check on it once in a while.

I am interested in efficiency... it tells me that I'm doing it correctly. The runs where I calculated my efficiency, it was at best 61%, maybe a tad bit higher. That was with no pH control at all beyond inadequate shots of HCl using a wash bottle. I think a well-run home chlorate cell can do 75% to 80%, and if I can get there I'd be delighted. The other good thing, if you know your efficiency, and it is reasonably constant batch to batch, you can use math rather than tedious chloride quantitative analysis to determine end-of-run.

I am going to finish plumbing today... I need to get more HCl from the pool store, and rig up some sort of drip system, before I fire it up. I don't want to proceed without at least a crude way to keep the pH in check.

Again, sorry for yet another "stream of consciousness" post, but I am delighted to find a bunch of smart guys here with a common interest. Wish me luck!

Oh yes, I finally figured out a good way to create a swappable anode "cartridge". This was baffling me for some time, but the end result is simple. I created a round slug of CPVC, slotted it 98% of the way through for the anode strap. I drilled 3 or 4 holes through the strap where the CPVC circle would fit, and then glued it in. The holes are critical... PVC cement does not stick at all to Ti. The holes form a PVC "bridge" between the hemispheres, and keep everything secure.

Next, the anode strap + CPVC slug was glued into a rimmed carrier that has a viton o-ring, not installed in the pics. This carrier is installed in the cell lid from below, and an aluminum block is clamped on the topside to fix the strap in place... it cannot move vertically. This will replace the "sculpey" PVC clay plug that I had tried earlier.

dann2 - 3-11-2008 at 12:36

One more figure for acid consumption in Sodium Chlorate process:
Encyclopedia of Chemical Processing and Design
Vol 51 page 184

50kg HCl per ton Sodium Chlorate.
It does not say if they mean 35% HCl acid (or perhaps 100% equivalent).

Thats 50 grams per KG Na Chlorate

What is the No. of the patent showing the set up for measuring Chlorate conc.?

I like that chunk of Al. It will be a great heatsink. If you really have a stubborn heat problem at this point you could cut slots in the block, just like a regular heat sink.

In regards to monitoring Cl2 coming from cell see pic. below. It is from E. of Chem Proc. and Design.
Perhaps in conjunction with this type of device it may do instead of pH monitoring.
Cl coming from cell may/will also be a function of things other that pH so you may have to keep cell parameters constant (eg. temp) when using this method.


cl3.gif - 6kB

Swede - 3-11-2008 at 13:36

dann2, the patent (uElectrodes for chlorate concentration)

The language was beyond my abilities, but I did only glance at it, not study it in detail. It sounds like it will require some advanced equipment and signal processing.

I'm guessing that 50g/kg of HCl is the actual mass of HCl, so we need to take normal dilutions into consideration.

HCl density: Here's an interesting table:

It states at 38% by weight acid (normal pool acid) you get 451.6 g/l, or 0.4516 g/ml, of HCl.

At 100% efficiency, it requires 161 ampere-hours to convert 1 mole of chloride to chlorate. For potassium salts, that converts to 1313 ampere-hours (AH) per kilo; sodium salts, 1511 AH/kg.

Let's determine the acid per AH needed at 100% efficiency. The MW of sodium chlorate = 106.5; there are 9.39 moles/kg. 50g acid is needed for each kg sodium chlorate. 50g HCl / kg sodium chlorate converts to 5.33g HCl per mole.

161 amp-hours creates 1 mole of chlorate. Thus, you need 5.33g acid per 161 AH, or 0.0331 g HCl per AH, undiluted.

With 38% pool acid at 0.4516 g/ml, you need 0.0331 / 0.4516 = 0.0733 ml/AH acid.

I hope my math is correct here. If your cell runs at 50 amps, you should set up a drip or dosing system that delivers 50 X 0.0733 = 3.665 ml per hour of 38% HCl. Not as much acid as I would have thought, frankly. If you use more diluted acids, the table will tell you how much acid per liter is there.


100% acid per amp-hour: 0.0331 g
38% acid per amp-hour: 0.0733 ml

Anyone see anything wrong with my analysis? It assumes 100% efficiency, but since the acid goes in along with the electricity, if you have to run your cell longer due to normal inefficiencies, you'll be applying extra (needed) acid.

Dann2, thanks for bringing this up. I think a rule of thumb like this (3.7ml/hour per 50 amps) will be very helpful, if it in fact works out that way.

dann2 - 3-11-2008 at 15:11

Hello Swede,

I agree with your calculations.
With '100%' acid you need 39.7ml acid per day.
If you used 12% acid you would need 331 ml per day. (hope I got that right).
A figure that I give (from my jamjar type cell that was pH controlled) is not too far (plus 25% :P) from this figure. I calculated on post of 29-0-08 that you would need 420 ml 12% acid (from the jamjar data) for your cell running 50 amps.

If you have room in the cell for the extra water you would be better off using a lower % that 38%, as each time you drop a drop of the 38% stuff into the cell you will get an immediate reaction at the liquid surface and a smell of Cl2.
You could inject into bottom of cell and that would help. Perhaps it is a very minor issue.


[Edited on 3-11-2008 by dann2]

tentacles - 3-11-2008 at 15:47

I finally received the MMO anodes Swede kindly made up for me, and I spent the morning clearing off my work table and getting out the LD plating apparati. I can't seem to find my stir bar, though, I can probably borrow one this weekend but in the mean time I think I will fabricate one with some HDPE tubing and rod magnets.

edit: Got a reply back from the Refex guys. It's not pretty:

Hello Brian

Refex probes can be used in chlorate and perchlorate manufacturing
processes. Their advantage is a unique non-porous reference junction that
prevents poisoning of the probe and extends life to up to 5 times that of
conventional porous junction and direct contact probes.

Refex probes are available direct from South Fork Instruments. Typical
price for a combination probe is approx $950.

Best regards

If anyone wants his contact info, just let me know and I'll U2U it. I assume by combination probe he means pH + ATC

[Edited on 3-11-2008 by tentacles]

Swede - 3-11-2008 at 17:43

Ouch, I had a bad feeling they'd be extraordinarily expensive. That's more than twice the cost of the most exotic process electrode I've seen so far.

I wrote him as well. He's probably wondering why all these people are asking him about the chlorate process. The next thing you know, he's googling "Reflex Electrode" and finding our secret lab here.

Glad you finally got the anodes.

As for pH, I think a calculated dose combined with an occasional measurement will suffice. Unless I am totally wrong about the poisoning, and a modest double-junction electrode costing $70 will last a year or two... that'd be worth it.

Swede - 4-11-2008 at 11:52

Update: I promised in my blog that my "T-Cell" would be running... yesterday. It's still not running, for one primary reason. I had ordered a dosing pump off eBay, but I was not expecting it for several days yet. it arrived early, and now I am cramming, trying to integrate it. I don't think it's going to happen.

The pump: A Hanna BL5.

I made a big mistake. I placed the order before I really thought about the volume of acid to be delivered (low) and the fact that I was planning on automatic pH control, rather than semi-manual. This pump delivers FAR too much with each stroke.

Benefits: The BL pumps can be bought bew off eBay for a hundred bucks. Good dosing pumps are normally $300-$500. Everything wetted is either kynar, PTFE, or glass, good for darned near anything, chemical wise. And the interesting thing, with each stroke of the pump, exactly 1 ml is delivered. As far as I can tell, it is quite precise.

This means for 50 amps, I'll need about 4 strokes per hour. When the pump is set on it's slowest speed, it executes one stroke every 5 seconds. This means that each hour, unless I use a controller, the pump would need to come on for 20 seconds, then shut down. I don't have a timer that will do that. The average consumer timer would come on for at least one minute.

I could dilute the acid by three, OR, I could have it come on for a minute, overdose slightly, then shut down for three hours.

Anyway, I need to figure out how to do the acid additions. Maybe I should set up a drip while I figure out how to integrate the pump.

Lesson learned... if you shop for a dosing or chemical pump for a home chlorate system, get one that delivers very small dosages; or, one that can be minutely controlled.

I hope to get the system running tonight or tomorrow morning. I'll need at least 24 to 48 hours of nearly cotinuous monitoring and tweaking to ensure there are no leaks, the temps are good, and the system is behaving as it should.

It has really gotten more complicated than I had anticipated. Oh well! :P It's also been a lot of fun! I just hope all the work and materials are not going to go to waste.

My old system with no pH control was 60% efficient. I think I can get this to 80% or better, meaning at 80 amps (if it'll handle that without overheating, the projected yield would be 1.46 kilos per 24 hour period at 100%, and 1.10 kilos at 75%, with an estimated capacity of 25 liters.

My old system of 6 liters was doing 1 kg per batch from 14% down to 7% chloride, so this rig should do at least 4 kilos per run. If I can take it to 3% chloride, it should do 6.8 kilos. Since I am hopelessly optomistic with estimates, a more realistic value is in fact 3 to 4 kilos max. Can't wait to fire it up!

Swede - 4-11-2008 at 14:02

Sorry for the string of posts... it's running, and everything so far is perfect. No problems at all. 50 amps created a delta T (temperature differential) of 20 degrees within 15 minutes. The pump is barely ticking over right now. If the DT gets too large, I'll crank the pump up a bit.

pH within minutes was at 8.4. Until I can rig up a drip (probably tomorrow) I'm simply ging to inject some HCl using a syringe.

Electrode spacing and area looks good... 5V on the supply created a current of 55 amps. If everything goes well, and the system doesn't overheat, I may try 70 or 80 amps.

I'll get some pics up later. :D

tentacles - 4-11-2008 at 15:04

Nice news! Why not setup the dosing pump with a tee and a couple of kinks/knots to divide the flow, and divert most of the flow back to the HCl container? The pump will deliver sufficient pressure to keep the flow rate reliable (you certainly aren't going to get crap growing in the solution).

Pretty f'in sweet that your 5v delivers just the right level of current.

[Edited on 4-11-2008 by tentacles]

dann2 - 4-11-2008 at 15:16


Good to see machine up and cranking.
One small note. If you happen to pump far too much acid into the electrolyte and put the pH way down, the cell will start to produce Cl<small>2 </small>gas. This may be a serious problem if you have it in a place where the gas cannot vent off safely, so be careful with the acid.

I presume you are using KCl?


Swede - 4-11-2008 at 18:54

Tentacles - great idea! I didn't even think of diverting the flow. I'll try to set something up tomorrow.

Dann2, the system pipes to the outside, no fears about gas. Everything looks tight and dry. Just a little bit of salt creep on a couple of threaded joints, otherwise, all the seals are doing well. Yes, I'm using KCl, plus some used electrolyte from previous runs, recharged with chloride.

The pump and temperatures are behaving exactly as I wanted them to. The actual flow through the EC is probably no more than 1 liter/hour or less, it's just a thin stream. As the pump rate goes up, the DT goes down, and of course if the pump is backed off, within a few minutes the DT rises. The CPVC electrode cell is at 60 degrees, and the big chamber right now is at 35.

The cathode pair is barely warm. The anode, a bit hotter. I'll probably take a few minutes tomorrow and bolt a heat sink onto it. Nothing extreme, it's probably 80 degrees C or so.

The cheap pH electrode is working well so far. I'm doing manual checks, drawing the liquor hot from the EC, and it's tending to hang around 8.0. I'm limiting contact of the pH probe to the nasty electrolyte, and washing it clean immediately. I've been light with the acid so far; I'll add a bit more.

[Edited on 4-11-2008 by Swede]

tentacles - 4-11-2008 at 19:47

Swede, is that the same size shank as you installed on mine (it looks like it)?

Did you get just one cheap probe? I'm really curious to see how long they'll hold up in continuous use. At $15 a pop, even if they only last a month or two of use, that's really not too bad at 2lbs a day. Hopefully we'll be surprised by their longevitiy - I know I've already been surprised at low long DSA last in commercial use (more than 3 years).

Swede - 5-11-2008 at 07:00

Tentacles - The shank is identical, 0.041" thick by maybe 0.950" wide. It's doing 50 amps with ease. The aluminum block connector helps a bit. I may clamp a heat sink to the block, and later attach it permanently, but it's really not needed.

pH - when I bought the dosing pump from an eBay outfit called eSeasonGear

I picked up the electrode as well. I'd say it is not a super-cheapie, but it is definitely low on the totem pole. $37. It is a double-junction electrode from Milwaukee instruments...

It is the yellow one in the picture. The first thing I did was prepare a liter of electrode storage solution using KCl, and I soaked it overnight. The next day, I calibrated the Extech controller using pH 7.00 and 10.00 buffers. I was really happy to see 7.01 and 10.01 on the meter without having to touch the set screw.

The way I am measuring pH right now is simple. I draw off a few ml of liquor from the ball valve on the exit port of the EC... comes out quite hot.

The electrode is rinsed, goes into the liquor for maybe 2 seconds, I note the reading, and out it comes immediately into distilled water for a rinse. Then, right back into storage solution. With such limited contact to the liquor, I am betting it will last a long time.

One thing I've noticed in the 16 hours it's been running... the pH is not responding as well as I'd like to additions of HCl. I started by dosing 4ml conc. HCl per hour. The pH would go from 8.2 to 7.9, but fairly quickly go back up to 8.2. It seems to want to hang there at 8.2, as if it is buffered there.

Today I am going to force the issue with much larger HCl additions. If I can get it below 7.0, I am going to see how long it will stay there. If it climbs back quickly to 8.2, then I think I can conclude that manual, periodic acid additions, while helpful, probably won't be able to keep the pH at 6.6 or 6.8. That would require true control; or, a continuous drip or dosing setup.

This begs the question again of buffering with chromates, or the phosphate ions. I'd really like to give the latter a try. I will conclude this run with no buffers, but for the next one, I'd definitely like to find some sodium phosphate salts and try the buffering as described in the patent.

[Edited on 5-11-2008 by Swede]

dann2 - 5-11-2008 at 13:26

Just a though. Would the sample you are taking for pH measurement be coming from close to the Cathode. It may show too high a pH if so.
You should be taking a sample from the solution bulk. The bulk pH should be at 6.6 to 6.8 for the 'Chemical Chlorate formation' reactions(s) to happen in the bulk area.
Perhaps you could measure the pH of the crystallization chamber to see what that is. It may be a better indication of the cell bulk pH.
Keep checking calibration of probe.
Make sure the sample has cooled before you measure (or is probe ATC?).

Remember that ALL stuff that you have read about pH control is in relation to Sodium Chlorate cells (not K). Perhaps pH control of K cell is not so simple or perhaps it is inclined to run at a higher pH. I would be inclined to think that a K cell would be similar to a Na cell but it's possible that they are different.


Swede - 5-11-2008 at 14:13

Good news, with some moderate HCl doing it's down to a more realistic 7.5. And good comments too re: bulk cell pH. I'll pull some samples from there and check them. I was not aware that the bulk of the reactions take place away from the electrodes.

I recieved an overnight order from MSCdirect, enough plumbing stuff to hook up the dosing pump. I have a hardware store timer, and with no bypass, it'll administer 12ml in one minute. From what I've seen so far, that is a reasonable dose every three hours. I'm close to having a hands-off setup!

The initial chloride content BTW was 145 g/l. Unless something odd happens, I'm going to take it to 50 g/l this run, perhaps lower in future runs.

dann2 - 5-11-2008 at 16:00

Originally posted by Swede
Good news, with some moderate HCl doing it's down to a more realistic 7.5. And good comments too re: bulk cell pH. I'll pull some samples from there and check them. I was not aware that the bulk of the reactions take place away from the electrodes.

The initial chloride content BTW was 145 g/l. Unless something odd happens, I'm going to take it to 50 g/l this run, perhaps lower in future runs.


It's not really 'the bulk of the reactions' but rather 'the bulk' reactions. I know it sounds ridiculous to bring it up but there is a difference.
This is why, when we are controllig pH, it is good to have a seperate chamber where the 'chemical Chlorate formation reactions' (also called the 'bulk reactions') can take place, away from the electrodes. (Or you can have a single, large, chamber which contains the electrodes + this chamber will have volume away from the electrodes, known as the 'bulk').
When controlling pH is not great to have a small 'bulk' area in the electrodes chamber. Large chamber in relation to electrode size is best (or a seperate, very warm chamber).

When we are not controlling pH, it does not matter a toss about the size of electrode chamber in relation to electrode size (within reason) as there is NO 'Chemical Chlorate formation reactions' taking place. Chlorate is make by a reaction called 'Anodic Chlorate formation'. The vast majority of garage Chlorate cells (no pH control) make there Chlorate by 'Anodic Chlorate formation'.

'Anodic Chlorate formation' is avoided like the plague in industry (pH controlled used!) as it is considered an electricity robbing reaction.
Since you are doing pH control you should perhaps aim for this lofty ideal! (seperate 'chemical Chlorate formation' chamber)

I think you will agree that the above explanation of cell chemistry is an exquisitely clear and terse explanation of what is going on :o:P::P ..........perhaps not!
The picture below explaines it better, I think.
The how, why and where of reactions A, B and C are really all you need to know about, in order to understand the vast majority (from our simple point of view anyways) of what is going on in the cell. (you may also like to look at the graph of distribution of species (HOCl & ClO-) in the cell at different (bulk of solution rem.) pH values, next post)


[Edited on 6-11-2008 by dann2]

rea2.gif - 36kB

This diagram just about sums up what pH controll is all about.

dann2 - 5-11-2008 at 16:03

perhaps straying off topic a bit but what the heck.

species.gif - 11kB

Swede - 5-11-2008 at 16:14

Dann2, before I absorb your last post, I want to say - WOW - you saved my ass. As I mentioned before, the pump turnover is exquisitely slow, the EC cell contents probably swap out only every 2 to 3 hours or so. I was adding, adding, adding HCl to the collection vat, which gets trickled into the EC. I noticed that the EC pH was in fact coming down, around 7.5, but I wanted to check the vat pH as you suggested.

It was around 3! Way too low. The highly acidic vat contents were being injected into the EC, keeping that around 7.8 to 8.2 most of the time.

I turned the pump way up in an attempt to stabilize the system. Hopefully it'll calm down around 6 or 7.

Now I know why all my HCl dosing wasn't showing up. I was measuring in the wrong place! I probably administered 150 ml concentrated HCl into this 25 liter system in the last 30 hours, far more than my previous calculations showed was needed, probably 2X as much.

Would it be best just to let the system recover on it's own, or should I add a bit of NaOH? I don't want to screw up the reaction. I think I'll add just a touch of NaOH and remeasure. I should be using KOH, but NaOH is all I have.

Reading your last post, it does make sense and is enlightening. My best efforts earlier yielded 60% efficiency, which is in-line with a non pH controlled system. What I've learned in the last 24 hours will help. I think, even without a true continuous pH controller, I'll be capable of maintaining pH between 5.5 and 7.5 with little trouble, perhaps even better. I just needed to learn where tto measure!

I hope I didn't screw things up too badly.

dann2 - 5-11-2008 at 18:28

Hello Swede,

Looking at the diagram below, I think it is fair to assume that I would not cut it as a pilot. Moving swiftly along..................

Good to know there was a easily found problem.
Consider using your crystallization chamber as a 'chemical Chlorate formation' chamber. You will then have to add another chamber. I reckon your electrode chamber is a bit too small (in relation to electrode size).
That's not to say this device will not work. You may get CE that are high enough, and it may not be worth the extra effort. The crystallization chambe could be a bucket with a cloth in it just to see how things went. There would be litttle need for venting etc.
You could also consider drilling lots of holes in the cathodes to reduce surface area and putting plastic on the back sides of them where CD is probably very low. (will stop/minimize cathode reduction).

Regarding your system at the moment I would be inclined to meter in what ever acid you think is needed and leave for 24 hours. (if your pumping system allows that).

Why do you want to lower Chloride concentration down to 50 gpl or less? If you want to keep CE high you should keep it at 100 gpl or more.

MMO always seems to give good CE. You should be able to get 80 or 90% with pH control.

It is hard to give a figure on what the circulation rate of the whole system should be. I would guess if you pump the contents of the electrode cell once every hour. (Wild assed guess).

Keep it up!!

[Edited on 6-11-2008 by dann2]

Chambers.GIF - 13kB

12AX7 - 5-11-2008 at 18:31

Why not change the pH after cooling? I always added HCl until a bit too much chlorine was produced, after crystallizing the NaClO3.

Actively controlling pH (by rate of HCl addition or by actively monitoring pH) at the cell seems like a better idea to me, as its pH will gradually rise no matter what. Adding HCl later sounds half assed.


Rosco Bodine - 5-11-2008 at 19:15

Something I have been thinking may be worth considering is closing the loop for recombinant chlorine as HCl recycle.
A small H2O electrolyzer of slightly excess of needed capacity could supply a steady stream of H2 into a reaction chamber iluminated by a UV lamp and into which
the venting chlorine from the main cell is also supplied,
maybe something like a graham coil condenser for a tube reactor where the H2 and Cl2 are mixed and reacted to form HCl, and a small sprayer nozzle would disperse some small diverted flow of the electrolyte from the electrolyte
recirculation loop , which would absorb the HCl and discharge back into the main cell. Such a scheme would
tend to keep the pH constant without attention or constant monitoring, by virtue of returning any evolved
Cl as HCl , closing the loop.

kclo4 - 5-11-2008 at 19:58

That seems like it would be very difficult to do, Possibly Impractical for how much is being produce, as well as how cheap acids are. How would you get it so the Cl2 + H2 can combine with out destroying whatever device it is in?
Also, Isn't a significant amount of H2 produced compared to the Cl2 that is released? You'd need to either let that off, or react that with Air and capture it.

Rosco Bodine - 5-11-2008 at 20:39

Originally posted by kclo4
That seems like it would be very difficult to do, Possibly Impractical for how much is being produce, as well as how cheap acids are. How would you get it so the Cl2 + H2 can combine with out destroying whatever device it is in?
Also, Isn't a significant amount of H2 produced compared to the Cl2 that is released? You'd need to either let that off, or react that with Air and capture it.

Glassware and teflon tubing I have aplenty. A glass graham condenser with a three way adapter on top and a small
spray nozzle for the electrolyte entering the center of the adapter , a UV rich (halogen) lamp shining on the gases mix coming in through a two hole fitting on the sidearm of the three-way adapter, and the bottom discharge opening as return to the main cell or electrolyte reservoir. H2 and Cl2
combine spontaneously in the presence of light and the HCl
produced would immediately be absorbed by the aqueous
electrolyte as the mixture spiraled down through the glass
coil of the condenser.

Something like this would be far less expensive than pH monitoring and regulation by other means ..I think. If a cell is already using a pumped electrolyte and a coaxial electrode,
along with a reservoir tank , this would be fairly simple to implement and it has no additional moving parts or sensors.
Anyway for a high capacity cell that is what I would do.

[Edited on 5-11-2008 by Rosco Bodine]

12AX7 - 5-11-2008 at 21:48

Originally posted by Rosco Bodine
Something I have been thinking may be worth considering is closing the loop for recombinant chlorine as HCl recycle.
A small H2O electrolyzer

Why another bit of hardware? You've got a rather large H2O (and other stuff) electrolyzer going on just inches below. Toss in a mesh platinum filter and be done with it. (Oooh, do they make platinum wool? That would be awesome.)


Rosco Bodine - 5-11-2008 at 22:37

I'll have to review this one to be certain ...
but I believe you are right

I see what you are saying is that there may already be a mix of H2 and Cl2 in exact proportion coming off the cell similar as "Brown's Gas" coming off the electrolysis of H2O as 2 H2 + O2 ????

Possibly all one has to do is react the cell evolved gases in an empty headspace inside the cell or in a column above , equipped with a UV source and a spray nozzle for electrolyte at the top to provide an HCl absorption / precipitation tower...yeah that should work ...good idea.

You could just use a "gurgle column" filled with Berl saddles or beads, broken cubes of automobile safety glass
mixed with glass beads, short pieces of broken or intact
pieces of glass tubing and rod, hit it with a UV lamp and
pump a bit of electrolyte in at the top to keep all the
"plates" in the column wet without flooding it , so it
would work as a counter current absorption tower.

Platinum on alumina or shards of porcelain bisque might work, even scrap catalytic converter pieces might do the
trick as a column packing even better. I'm not sure and
would have to see what is specifically catalytic for HCl formation from the elements. Palladium on charcoal chunks
maybe ?

Interesting resource links, don't recall having seen this one

concerning the photocatalytic reaction of hydrogen and chlorine

[Edited on 6-11-2008 by Rosco Bodine]

Swede - 6-11-2008 at 05:03

Before I can digest the evening input, I have a small situation to take care of... ;) Why does bad stuff always happen when you are asleep?

I went to bed with everything looking good. This morning, the DT (Temp Diff) between the two had gone from 30 to 60. The temp of the EC was 90 Celsius! I found out why...

When I built the thing, I created two ports on the big vat, one at 1/3 height, the other at 2/3 height. I plugged the pump input into the 1/3 height port... which jammed solid with crystals in the evening chill. The EC flow ceased, and at 50 amps, the temperature soared. Hopefully no damage was done. I dialed the current down to 2 amp just to keep the electrodes charged. It's cooling right now.

Two things I did wrong. I placed the pump inlet source too low on the collection vat, and I selected tubing with way too small a bore for the pump. AND I should have installed some sort of prefilter for the pump. I'm lucky the pump didn't chew through the tubing and spring a leak. I should have known better on all of these issues.

I am convinced this concept is workable. Everything was going as planned with the exception of the pH. I am going to tear down what I have this morning, harvest what crystals exist, replumb, and fire it back up. It's going to be a noxious task.

My HCl dosing pump came with a beautiful mesh PVDF prefilter which I am going to cannibalize and use to protect the system pump; AND I'm going to plug into the port at 2/3 height. There's no way crystals are going to reach that port.

When I'm done with this chore, I'll come back and read the latest inputs. But it's going to be a busy morning! :D

[Edited on 6-11-2008 by Swede]

[Edited on 6-11-2008 by Swede]

Swede - 6-11-2008 at 08:14

Well, I've got some good news, plenty of bad news. Here's what happened...

The line leading to the pump was JAMMED with fine crystals. The crystals had not reached the port level in the collection vat; the liquor had simply cooled, and once the solution was saturated with chlorate, fine seed crystals formed in the tubing, and from there, they grew.

With the pump in a no-flow state, the temperature in the EC soared. I measured 90 degrees C in the thermowell!

I took it outside and opened it up. The bottom had a layer of very fine crystals about 2" deep. Good news: the materials (CPVC and PVDF) held up perfectly, which is remarkeable given those temperatures. The bad news - the cathodes. I have never sen anything like it, and I am at an utter loss. They are both warped HUGELY in a symmetrical pattern away from the anode, which looked fine.

Check this out:

What the f***? All I can think of is that perhaps the TIG welds set up some sort of stress, which was released by the current? More bad news - the cathodes are permanently installed, and the lid is ruined unless I don't mind warped cathodes. Really, really bizarre.

Continued; next post

12AX7 - 6-11-2008 at 08:28

Hmm... interesting...

Hydride implanation? Hydride is lower density = expands on the near face (which carries most of the current). It's not weld stress, because the curvature is along the length of the plate.

Heat it to release the stress and hydrogen. I suppose red hot would do it, which isn't a problem that far from the lid (titanium conducts poorly). Though do keep a wet rag on hand just in case. Probably worth wet-ragging the anode too, so it doesn't get toasted in the process.


[Edited on 11-6-2008 by 12AX7]

Swede - 6-11-2008 at 08:32

I wheeled the collection vat outside... the cart IS the way to go for a heavy setup. I attempted to drain it from the side ports, but they too were clogged hopelessly.

I popped the lid. There was a small amount of chlorate in the bottom, some nice, fat crystals, but not much. I'm estimating the chloride level in the liquor went from 145 g/l to no less than 130, to deliver such a paltry yield. I expected this. the rig had not been running long enough for a big yield.

With the exception of the clogged ports, everything was perfect. The lid, with all it's vent ports and gadgets, was likewise good.

I had installed those CP Ti cathodes earlier. I had a hunch that something catastrophic might happen, and I want to be able to run the vat as a stand-alone cell. That's probably what's going to happen now.

I was forced to scoop out much of the liquor by hand, with a cup, to salvage what little chlorate there was.

Thoughts: Everything was going perfectly so long as the entire system was liquid. Sodium chlorate would not have given me this problem. Someone needs to talk me into using sodium salts.

I was surprised at the amount of crystals formed in the EC. BUT, and that's a big but, I suspect everything I saw in there happened AFTER the circulation stopped. I had 50 amps running in a relatively small cell, and it simply brute-forced a certain yield. If the liquor had been circulating, I would have expected crystallization in the big vat.

I'm not sure this concept is sound with potassium salts. If it can be done, it'll require careful monitoring of the pump lines, which need to be fairly wide-bore, and screening of the pump inlet. A thermostat needs to be set up to remove power if the temperature rises too high in the EC, which would indicate a stopped pump.

A pump capable of carrying a slurry, and the use of hard CPVC pipe instead of tubing might also help.

For the moment, I am going to shelve the EC, and execute a carefully-controlled run with just the collection vat. With my old, crude setup, I was getting a kilogram of fat, clean potassium chlorate crystals, and this vat is at least 4X the size. With no external plumbing beyond venting, the system is dead-simple.

I'm down but not out. I'll find a way to boost efficiency. And the big question, what in the HELL hapened to my cathodes? :(

No biggie. It's part of the research process. And Tentacles, you'll be happy to know that anode material looks like a winner for chlorate, so far.

[Edited on 6-11-2008 by Swede]

tentacles - 6-11-2008 at 12:29

Swede: Perhaps the answer is to spot weld the cathodes, or reinforce them somehow. Maybe even pre-hydride them? A couple hours in a weak H2SO4 solution hydriding (use SS anodes) might make the stress more even. The strange thing is that when I've heard of the hydride warping phenomenon, it's been said to take at least a year to become a significant problem. I think it's more likely a combination of weld stress and hydride warping, maybe. I'm looking to get my hands on a MOT to make a spot welder soon. What kind of clamps did you use for yours?

At least the anode material works well! Maybe you should try it with sodium salt for a run, just to try the system as designed? Not like you couldn't reassemble and try again later, but it could be fun. Think of the yield! HUGE! The problem, of course, is washing out the Na, but it's probably not all that difficult once you've got perchlorate.

 Pages:  1  2    4  ..  12