Downs Cell Project and Related Processes / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Downs Cell Project and Related Processes

1 2 next

Move Thread

LAN_403

Chris

Wed Feb 22 2006, 08:49AM

Registered Member #8 Joined: Thu Feb 02 2006, 04:34AM
Location: Harlowton, MT, United States
Posts: 214

I have now collected just about everything I need for my downs cell, minus a few parts and some real time to work on it.

The cell body itself is complete, little more than a steel box as shown below, measuring 6" x 6" x 5" deep. I'm still waiting on the ceramic top piece which is being glazed by a third party (I don't have a kiln but a representation can be seen below made of cut ceramic tile). A graphite anode will suspended down in the center, and the steel box will be the cathode, minus the bottom which will be covered with refractory material. As with any downs cell, sodium metal is collected from a tube at the cathode and chlorine gas rises up the center, while an iron screen as a membrane divides the cell. This cell is also to be filled with new salt from a hopper through the center. As calcium is electrolyzed out as well as sodium, it simply falls back into the bath as a solid (it is below its melting point) to react with NaCl yielding CaCl2 and Na metal, so calcium chloride it is not consumed at a considerable rate.

The cell holds around 2.6L of molten salts, a mix of 60% CaCl2 and 40% NaCl by weight which melts at 580Â°C. At a density of 1.56g/cc for molten NaCl and 2.1 for CaCl2, this is about 4.9kg of salt, or 1.96kg (33.6 moles) of NaCl and 2.94kg (26.5 moles) of CaCl2. The total lattice energies of these amounts are 26.5MJ and 60MJ respectively, so I guess it will take about 86MJ to pre-melt this cell, not including the energy required to bring the salts up to the melting point and dissapative losses. A 12kW set of electric heating elements will be attached or brazed to the bottom of the cell to accomplish the pre-melt in 2-3 hours, and the cell insulated with 1" ceramic fiber refractory such as used for insulating kilns.

When the melt is complete, the cell will be operated at something like 8V and 500A from a gas powered alternator set which runs rectified 12 phase at approximately 1600Hz. This consists of 4 130A automotive alternators linked to a chain drive, also pictured below. I plan to run 50lb (388 mole) batches of salt through the cell at a stretch, which will require a total of 37.44 MC (388 faradays) or 10.4kAh of charge to electrolyze. At 500A this is roughly 21 hours of elecrtlysis. At 8V this is 83.2kWh (283,890 BTU), which I can estimate (at 20% overall thermal efficiency) as 12 gallons of gasoline for the alternator set.

The sodium metal produced from each batch (roughly 15lbs yield as some remains dissolved in the cell) can be either run directly into the casting trays which are backfilled with argon (incomplete, but the casting tray bed is pictured below, made of angle iron welded together) or collected in a hopper (also with argon filling) to be used for the refining of the more volatile alkali metals. In this process, a steel vessel (again with argon filling) containing molten KCl (alternatively RbCl or CsCl) is heated to 800C or hot enough to easily boil elemental potassium. Molten sodium, still well under boiling, is dripped into the bath from the hopper above through a rudimentary needle valve, displacing the more volatile potassium (or eventually cesium) which boils out. The vapor then rises up a fractionating column (this is complete, though the vessel is not yet finished) similar to any reflux batch still. The potassium after running through a condenser can then run directly into the casting trays. More reactive but lower melting metals such as cesium would alternatively be collected in small jars. This reactor/distillation apparatus will be capable of handling tiny batches of salt, though with the vessel measuring 8" x 8" x 9" large batches (for the potassium) are possible as well. This unit will be heated either with a 12kW electric heater or large natural gas burner, and also insulated with ceramic fiber refractory. The entire reflux column will need insulated as well though to a lesser degree.

Running the 50lb salt batch through the cell will also yield 194 moles (13.75kg) of Cl2 gas, or 182mg/sec (0.002567 mol/sec) at the 500A electrolysis rate. Letting this escape into the atmosphere may seem practical at first, but at that rate the toxic gas would build up around the cell operating area rather quickly, plus it would be wasteful to not collect it. For this, the chlorine will be passively condensed into a tank by running it through a coil at well under -35Â°C (the boiling point of Cl2). This will be accomplished by coupling the condenser coil to a phase change heat pump - the evaporator coil and Cl2 condenser assembly is pictured in the attachments below. A check valve will be used near the tank to assure everything flows easily by venting excess pressure to the outside. To condense the chlorine as it passes through at the rate described, the heat pump will need to transfer the heat of vaporization of the gas (52.4W at the 500A rate) as well as cool the gas to it's boiling point according to its specific heat capacity. For this the gas should enter the condenser coil as cool as possible, after passing through a pre-cooler (such as the flat heat exchanger pictured below). A lower temperature also means less corrosion of the condenser coil (which is less replaceable than the pre-cooler) because hot chlorine is more corrosive. If the gas can enter the condenser at 100C, this will require 12W, or 64.4W total assuming perfect heat transfer. In reality the heat pump will be capable of far more power, and lower temperature. The two large 277V single phase compressors pictured below will be used with R-134a refridgerant, both in parallel if they both test out ok. They have not yet been tested but I have an alternative compressor which is also likely capable. The system should pump down to -10psig on the low side to achieve down to -50Â°C. The large heat exchanger (measuring 15" x 9" x 11" but requiring some repair soldering) pictured below is to be used at the condenser for these compressors. A simple expansion valve consisting of at most an adjustable diafragm valve, and at least a carefully crimped tube, will be used. With a continuously operating system such as this there is little need for anything more elaborate. The condenser will be cooled with the large furnace blower I used on my VTTC, along with a water injection. The lower the condenser temperature is the more able the system will be to tranfer large amounts of heat at low temperature. The extra evaporator coil seen coiling out from the Cl2 condenser coil will wrap around the tank into which the liquid chlorine will be collected. This will be insulated by placing in a sealed box with a vacuum pulled inside. The tank is nothing more than an old propane tank, purged and cleaned and fitted with a CGA-660 valve stem which is compatible with corrosive gasses. The liquid volume (at 0Â°C and the corresponding vapor pressure) of the chlorine produced from each 50lb batch of salt (13.75kg) is around 9.4L (2.5gal) and the vapor pressure at 21Â°C is just over 100psi (6.95 bar). The tank can hold 22L (5.5gal), so it would be possible to hold two batches inside.

Rather than storing such large amounts of a hazardous liquified gas away, the tank will be promptly fitted with a regulator (stainless CGA-660) and used in another process. The chlorine will be blown onto heated steel scrap turnings in a copper vessel, suspended inside a vacuum purged plastic drum with a simple overpressure check valve and other filling and draining valves. The anhydrous ferric chloride, FeCl3 powder produced in the exothermic reaction will be collected off the walls of the drum and stored in sealed containers (it's extremely hygroscopic), probably with a consistancy similar to lamp black. The entire 194 mole batch of chlorine is capable of producing rougly 21kg (46lbs) of anhydrous FeCl3, but it is not expected that all this will be able to be collected at least in anhydrous form.

Progress: Overall, the chlorine condenser coil and evaporator assembly is complete, as well as the downs cell body and the reflux column for the potassium distillation. The casting tray bed is welded together, and all the steel panels for the entire project are cut out. A lot of grinding still has to be done since the metal was cut out by me (relatively unskilled) by hand with a carbon arc (not so pretty like a plasma cutter would do). A lot of welding also has to be done for the casting tray and pre-casting hopper assembly, reactor vessel, sodium hopper, and the metal framework that holds it all. The heat pump for the chlorine condenser needs assembled, charged, and tested. Most of the components listed above have been obtained including a large K-type thermocouple for monitoring the cell and reactor, except proper valve fittings and regulators for chlorine gas.

A crude 3D rendering of the entire setup is shown below.

edit (2/28/06): When I was home for spring break, I didn't get as much done as I had hoped, but I pretty much got the big condenser ready, I just cut off the damaged section and soldered the main connection back on. It shouldn't take too much more now to have the heat pump/condenser part operational, or at least ready to start fiddling with it. I've more or less settled on using a cascade arrangement with R290 (propane) refridgerant in both circuits, so it will probably take some adjustment (see 'Extreme Phase Change Pump Cooling' here

for more info . For the first time I have put together a complete diagram of the whole assembly, seen below.

Notice how the hot chlorine rises through the salt hopper and through the cold salt. This not only simplifies the design of the apparatus, but allows a significant amount of heat which the chlorine would carry off to be recovered by the salt, which eventually goes into the cell. This pre-heats the salt and cools the chlorine, increasing overall efficiency.

Edit (4/20/06): Instead of doing what most folks do on 4/20, here I am updating my project thread with last weekend's progress. Wee. I just aquired a bunch more minor components for the project, and am in the process of getting more. I recently purchased several stainless steel vacuum/pressure gauges, which will handle all the pressure monitoring for the apparattus itself, and several vacuum - 400psi pressure gauges, 4 of which will be used for the HVAC compontents, high and low side of each circuit in the cascade. I also got the stainless regulator for the chlorine, stainless steel support parts for the anode structure which should be the next thing done, compressor oil, teflon sheet for gasketing, pyrex tube for insulating the anode support posts as they pass through the upper cell wall, various valves, and materials to make oil separators for the HVAC. Those things have all arrived now, it was like Christmas. I also have a large graphite cylinder on the way for the FeCl3 reactor crucible, as well as needle valves for the cascade expansion valves.

I now have a large gas burner (unknown BTU, but a lot) set up with a blower, and running on natural gas, in the area where the cell will be. This burner will be used to heat the FeCl3 reactor, the KCl reactor, and possibly the downs cell itself. I will just use it to braze electric heaters onto the cell if I can find suitably compact and cheap ones totalling about 12kW, but those other things it will heat directly as they are used. I had to turn a tapered graphite fitting to hook up the burner, which the incoming line screws into, and the tapered bottom of the burner sets on and seals to.

I also cut out the anode itself, a 1.5" x 1.5" x ~12" square graphite piece, and have started machining a graphite block to hold it with stainless set screws. It is this which the support rods will screw into, as in the rendering of the cell in this thread. Also I have cut out most of the thinner stainless panels which will comprise the upper part of the cell, outside the molten bath in a hot chlorine atmosphere.

Pictures will come as soon as any of these components are completed, which shouldn't be too long now as there is only 2 weeks of school left.

Renderings of the cell itself, and of the FeCl3 reactor can be seen below.

Edit (5/9/06): The anode and support hardware is now complete. Current will be passed through the 8 stainless steel rods, which pass through pyrex feedthroughs in the wall above the active portion of the cell. I am considering filling the hollow stainless tubes with brass to increase current handling. I don't think I will bother to seal where the rods pass through the glass tubes, but the rods need turned down slightly to fit through with some play. Perhaps I can add some teflon tape or teflon based paste to help the seal, but I'm not sure how hot that portion of the device will get. The anode is pictured below.

The natural gas burner which will heat the KCl reactor/still portion of the setup is also now running more or less as it should with a blower to provide the neccessary airflow. It is running about 257,000BTU/hr (75kW) according to the gas meter readings. The trick will be transferring most of this to the relatively small reactor body. It will need well insulated and fitted with some fins for heat transfer as the hot air passes under it.

Work on the FeCl3 reactor is delated, but the graphite crucible for it was pretty easy to make (drill one hole) and is pictured below. It will fit in the apparatus as shown in the previous rendering, and I now have all the pieces to construct it. The vessel itself will be of two large truck brake drums clamshelled together to form a sort of barrel. At 160lbs this is a little overkill, but that's ok and they were cheap.

----------------------------------------------- ---------------------------------
The operating procedure will follow something like this: First the cell will be heated and the salt in it melted, and filled to a normal level. The separator valve at the cathode riser pipe will be closed off and both sides of the cell evacuated by the vacuum pump. The anode side of the cell, including the salt hopper, chlorine condenser and tank, will be backfilled with chlorine, which has been produced by chemical reaction and then dried by bubbling through sulfuric acid. The cathode side of the cell, including sodium hopper, KCl reactor, reflux column, pre-casting hopper, and casting tray bed, will be backfilled with argon. The pressure on both sides of the cell will then be equalized to 1 atmosphere, and the separator valve opened so that now the molten salt keeps the atmosphere on each side separated hydraulically. It will then be time to power up the refridgeration system for the chlorine condenser. As the chlorine starts to cool and condense, the pressure on the anode side of the cell will drop and the salt level in the center of the cell will rise. When this starts to occur, it will be time to power up the cell and electrolysis will begin, replacing the chlorine on the anode side as it condenses. The argon atmosphere in the cathode side of the apparatus is allowed to bleed out through a valve as the sodium hopper fills with liquid. When all the salt stored in the hopper is depleted, the system will be allowed to cool naturally (a very slow process).

While this is going on, the chlorine tank will be closed off and slowly brought back to ambient temperature by letting air into the evacuated enclosure that kept it insulated. The tank will then be fitted with a stainless or monel regulator and the gas will be injected over a vessel of heated iron turnings, inside an evacuated drum with a pressure relief/check valve. The anhydrous FeCl3 produced will then be collected from inside the drum.

It will now be time to extract potassium or other more volatile metals. The salt in the KCl reactor will be heated and melted, as well as the sodium in the hopper. Sodium will then be dripped into the molten KCl or CsCl bath, which will reduce the salt and drive out the more volatile metal (the molten salt is well above the boiling point of potassium). Potassium will then rise through the reflux column to be purified, and collected in the pre-casting hopper where it can finally be run into the casting trays.

The anode side of the cell will then be purged slowly with dry air, removing the excess chlorine so it can be opened and refilled.
------------------------------------------------- -------------------------------

ragnar

Wed Feb 22 2006, 10:14AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Chris, this endeavour of yours has never ceased to amaze me. Congratulations on getting this far, its great to see all the equipment lined up almost ready for the fun part =)

Take care, keep us informed

Steve Conner

Mon Feb 27 2006, 07:19PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Have you got the fan shroud for your engine? It'll probably overheat without it.

Chris

Mon Feb 27 2006, 07:28PM

Registered Member #8 Joined: Thu Feb 02 2006, 04:34AM
Location: Harlowton, MT, United States
Posts: 214

No, It was gone when I got the engine. I will need to rig up a seperate electric blower to cool the cylinder, and I may eventually make a shroud to direct the flywheel fan air to the intake like a supercharger. The operating cost of this thing will be cut to about 1/3 if I can run the engine on natural gas instead of gasoline, too. I'm still worried about reliability of the engine and alternators for the long run times that will be required of it, but we will see how it holds up. I plan to shorten the chain by raising the alternators, and doing away with the auxilliary belt driven alternator which was going to power the engine. This way the chain drive will be far happier running at the extreme speed it needs to (the alternators run at a 4:1 ratio with the engine, which in turn runs at around 3000rpm). The battery can be maintained from a line powered charger, and a variac and low voltage transformer/rectifier can control the alternators' excitor current.

I'm going to need to get a large muffler for the engine too. Normally I wouldn't bother, but it's a serious necessity when running this thing all day long. From what little I've run the engine with my sketchy homemade breaker points, it's LOUD with no exhaust (duh

). Speaking of those sketchy homemade breaker points, they will be soon replaced by a simple optical sensor and IGBT to drive the coil.

Chris

Carbon_Rod

Wed Mar 01 2006, 01:41AM

Registered Member #65 Joined: Thu Feb 09 2006, 06:43AM
Location:
Posts: 1155

I do not know very much about this set-up, but I must say putting that tank in a deep vented-exhaust sandbag lined pit would be a good idea. Remote start and shut off may be the feature you really appreciate.

;D

Chris

Wed Mar 01 2006, 02:47AM

Registered Member #8 Joined: Thu Feb 02 2006, 04:34AM
Location: Harlowton, MT, United States
Posts: 214

The tank will be well below the boiling point of chlorine during the time the cell is running so it can condense and flow in at low pressure, as covered before. When it comes time to close it off and bring it back up to room temperature is the moment of truth/concern. Fortunately the vapor pressure is relatively low even then. Attached here is a vapor pressure vs. temperature graph for chlorine. With a steel tank and chlorine rated fittings there is little risk, and I'm going to pressure test everything to 200psi.

1141181239 8 FT1987 Chlorine Vapor Pressure

One thing that must be assured to insure the safety of this part is that the cell be evacuated and totally dry before electrolysis begins, as moist chlorine is more corrosive and not compatible with ordinary CGA-660 hardware. Any CCl4 formed from the reaction of chlorine with the graphite anode will be condensed out and collected in a small tube in the condenser coil, just above it's freezing point (-23Â°C). This toxic solvent is an undesireable contaminant and would freeze if it got to the tank, but is otherwise not a major hazard.

Remote start and shut off may be the feature you really appreciate.

The cell has to be constantly filled as the salt is electrolyzed away (the cell could explode if the liquid level fell too low allowing the products to combine). All the systems, like cell temperature, electrolyte level, cell voltage and current, the engine, the condenser system, and the vacuum systems, must be monitored carefully when operating to assure things remain within specifications and can be shut down promptly if anything fails. Attempting to control it remotely would probably be asking for trouble. It would be like trying to operate an old steam locomotive via remote, there's just too many things happening that need watched and I lack sophisticated computerized controls.

Hazmatt_(The Underdog)

Wed Mar 01 2006, 07:07AM

Registered Member #135 Joined: Sat Feb 11 2006, 12:06AM
Location: Anywhere is fine
Posts: 1735

This project is really starting to get interesting. I would want to operate with a friend close by for 'accidents' just in case, and full protective gear like respirator (for chlorine) and kevlar (for hot stuff coming your way).

Hope it works!

I would buy some Potassium off of you when you're done, but that's breaking interstate Hazmat laws. Oh well. I'll just have to pay through the nose for it I guess.

Chris

Fri Mar 03 2006, 03:02AM

Registered Member #8 Joined: Thu Feb 02 2006, 04:34AM
Location: Harlowton, MT, United States
Posts: 214

I would buy some Potassium off of you when you're done, but that's breaking interstate Hazmat laws. Oh well. I'll just have to pay through the nose for it I guess.

Shipping potassium and sodium is perfectly legal, but DOT hazmat charges, labelling and packaging regulations apply. It is classed as 4.3 (dangerous when wet) packing group II; you can easily look up the details in 49 CFR 173 of the DOT regulations, and exemptions for this class in part 173.13. In any case that has little to do with my project itself now, and is something to worry about when everything is done, but yes it may be made available on 4hv. Anyways lets try to keep logistics to a minimum on the project board folks, that stuff will be dealt with seperately.

Chris

Wed Mar 29 2006, 04:45AM

Registered Member #8 Joined: Thu Feb 02 2006, 04:34AM
Location: Harlowton, MT, United States
Posts: 214

Minor update: I have more or less completed the refridgerant condenser, and have posted a detailed diagram.

Jim

Tue Apr 11 2006, 08:43AM

Dunce.
Registered Member #28 Joined: Fri Feb 03 2006, 03:29AM
Location:
Posts: 76

We want updates! We want updates!

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob