Electrolysis of Heavy water - Drying the deuterium gas

[Chris Bradley]

Peter, that's nicely done! Where d'you buy these from!?

[Peter Schmelcher]

Thanks Chris for the kind words. The fuel cell is part of a science toy I found on ebay. The company is Horizon and the rather lame toy is called a Hydrocar.

Two new fuel cells with slightly different PEM electrical contacts and possibly a different membrane from mine.
ebay item
280965456352
and item
180756012073

[Mike Veldman]

I thought I'd post a couple of pictures of the rig I built for drying electrolysis derived Hydrogen for my replication of the Scientific American accelerator thirty plus years back. The top half of the board is the Hydrogen dryer, I don't have a clue what I was doing with the two drying columns at the bottom. The gray cylinder is the type, but not the one, I collected the gas in for later use. I dug this out of the warehouse recently to look at building a new electrolysis/dryer system, my old (glass) electrolysis device is long since broken. Maybe add a heated column and some pressure monitoring. I have a couple of PEM modules similar to the ones pictured around here somewhere, a new idea to play with, thanks. Anyway, enjoy the old pictures, sorry about the dust, a hazard of storing things.

mike

[Dennis P Brown]

Peter - brillent! I have one of those (for a small fuel cell to power a toy) and NEVER thought of that. Also, they work really well, too. Just don't take them apart.

[Peter Schmelcher]

Thanks Dennis for the kind thought.

I think I should point out again that the zeolite drying idea is untested with H2 or D2 gas. Zeolite is a manufactured molecular sieve material and allegedly will dry to 100ppb depending on the application. Zeolite drying is not a done deal in my mind and I am not that close to testing it. I believe we need several desiccant methods, and then we need to choose the driest or the driest combination.

-Peter

[aka47]

OK I am going to offer myself up for a kicking by asking a dumb question. Sorry.

Re the concept of drying a gas stream using a desicant in a vacum.

If my understanding is'nt too flawed, is'nt it true that no desicant will work as the low vapour pressure ensures that the desicant just dries out some more into your chamber?

Certainly not one you can dry out. It may temporarily hold some moisture but does'nt it only realy delay it's flow into your chamber.

To my limited understanding only a desicant that permanently binds the wate vaour in a chemical reaction has any hope of working. Something like a water vapour getter.

Vacum stils rely on the vapour pressure diferentials etc as does HVAC and refrigeration. But at much lower diferentials than are being talked about here.

Your best hope is that a mollecular seive or semi permeable membrane is suficiently fine that it will hold back larger heavy water mollecules whilst letting smaller deuterium and hydrogen mollecules pass through.

A filter in fact then, not realy a drier as such, but operating at a mollecular level.

Membrane filters operating at mollecular level are quite common in SCUBA gas seperation for creating oxygen enriched gas streams from compressed air. They do this by filtering certain gasses preferentialy form the higher pressure side of the membrane.

http://www.nitroxmadeeasy.com/Methods.htm

See the section part way down the page re membrane gas seperation. I am pretty sure that this particualr type of product though is not suitable, but it may offer ideas that lead to experimatation with other materials that might (or might not) turn out to be useful.

All in all though I am happy to be completely wrong about this as I am currently still taking things on board to do with the peculiar nature of vacum technology.

[Richard Hull]

Certainly, for those with a dewar, a liquid nitrogen cooling of a long coiled piece of small bore tubing would solve the issue completely and simply.

Flys in this ointment

1. You must own a larger, formal dewar. - 5 liter minimum.
2. Cost, transport
3. You must purchase LN2 every time you want to run your fusor
4. You must heat the coil after the run under vacuum to extract the retained water vapor.

Explanation.

1. Every place I have ever been to buy LN2, absolutely demands that you have a scientific cryogenic dewar. They will not fill or work with any other vessel. This includes large thermos bottles. They will not be liable for idiots wanting to play with "cold stuff".... No dewar......No LN2.

Finally, if you want to run your fusor many times in a day, 5 liters is a minimum as the stuff boils off fast, especially in small dewars. A 10 liter dewar should carry you through a weekend. Dewars can be found at Techno fleamarkets (HEAS) or on E-bay, labX, etc., for between $100.00 and $1000.00. (better "good luck deals" can be had on rare occasion) I will be offering a 25-30 liter Linde dewar at HEAS 2012 for $200.00

2. LN2 is about the same price as beer. (~2.00-$3.00/liter) Not horribly expensive but it does boil away real fast. As you move about, it is like leaving a trail of nickels behind you. Every time you wish to operate your electrolyzer you will have to get in your car and drive to your LN2 dealer (usually a specialty welder supply place) and back. Unless you have a stand to support the dewar, it is best to take a friend to place the dewar in the seat with him and roll the windows down. (bad in winter)

3. All the above expense and effort is needed every time you operate your system within a full week separation of experimentation/operation. A giant rental dewar of 100+ liters can be had and delivered for about $300.00 per month plus LN2 cost that would allow a period of continuous, un-interrupted experimentation if you have the money.

4. You must keep the condensing coil under LN2 until you can seal it off or isolate it from the fusor chamber. Only then can you reconnect or valve the line to a vacuum pump and heat the coil to near red heat to expell the moisture and pump it out of the coil.

This method is mostly a lot of hassle if you have a dewar, but if you don't and need the valving to work and evacuate the line after use then it could also involve some major costs, as well.

All said, this is probably the best method of absolute drying of deuterium made by electrolyzing heavy water.


Ricahrd Hull

[rkhertz]

A trap temperature of 0 C does not yield very dry gas: about 6000 ppm H2O by volume. Dry ice (-78 C) is much better: about 0.75 ppm H2O by volume. Liquid N2 (-196 C) will yield essentially completely dry gas. (calculated from data at http://www.spscientific.com/Tech-Center ... Chart.aspx ).

A dry ice bath should be adequate for fusor applications. Dry ice can be purchased reasonably at some large grocery stores. It comes in blocks of about ten pounds, and can be stored in a good, thick Styrofoam ice chest for several days. Wrap the block in a piece of canvas and break it up with a mallet. Make up a slush by dropping fragments of dry ice into 100% isopropyl alcohol in a dewar. Wear safety goggles and gloves, as the mixture will bubble vigorously. (Don't use rubbing alcohol. It typically contains 30% water and will yield a viscous, gummy slush). Do not store the slush in a tightly sealed container. Remember that isopropyl alcohol is flammable. Any suggestions for a non-flammable, low toxicity liquid to use instead?

I would not recommend using a copper coil for the trap, especially if a liquid nitrogen bath is used. Due to the high thermal conductivity of copper, it will conduct much heat into the bath and evaporate the liquid N2 rapidly. Stainless steel, with less than a tenth the thermal conductivity of copper, is a better choice. Pyrex glass, with much lower thermal conductivity than metals, is also a good choice; but, of course, is fragile. (See http://www.engineeringtoolbox.com/therm ... d_429.html for some representative thermal conductivities). Don't use soda-lime glass, as it won't take the thermal shock. In grad school we used a Pyrex trap packed with glass helices to trap moisture from hydrogen gas prior to measurement of its pressure and volume in a Toepler system. A Toepler pump is a good way to move hydrogen in a vacuum system if you don't mind slinging fifty pounds of mercury around.

We also used sodium-potassium alloy (NaK) in grad school for drying and removing oxygen from gases and some solvents. It is liquid at room temperature and a very effective drying agent. A disadvantage: it explodes on contact with air and halogenated solvents.

My choice for a desiccant for deuterium would be 3A molecular sieves. 4-8 mesh is recommended. (See http://www.sigmaaldrich.com/chemistry/c ... ieves.html ). It is easy to handle and regenerate. I may have some around, if I can find it.

CaCl2 and silica gel are other desiccants worth trying.

I don't like to use P2O5, as its surface gets gummy and loses its effectiveness as it absorbs water. Magnesium perchlorate is often used to dry gases, but I don't recommend it because it reacts with organics and some metals to form explosive compounds.

Merck has published a brochure with a number of useful chemical tables, including a list of desiccants on page 77: www.emdmillipore.com/showBrochure/17691.ProNet.pdf .

[Roberto Ferrari]

Hi Bruce

Any knews about your H2 generator?
In the worst case the cell is poisoned/damaged, you can recover some Pd tube from the cell and make a highly efficient D2 dryer after an electrolysis of D2O as Richard suggested.
Roberto

[Peter Schmelcher]

Great links Ray! I’ve searched unsuccessfully many nights for that information, thanks. I guess using the liquid line dryer and achieving 100 parts per billion is not very probable.

A pair of Peltier coolers sandwiching a spiral SS capillary tube might get a 60C to 70C temperature drop. The hot side could be an ice water bath with a pinch of salt for flavoring.

You obviously know substantially more chemistry than me. By chance does a reactive metal or other low vapor pressure compound come to mind that would capture the O from the water and liberate the D or H?

Another esoteric possibility I started pondering today are vacuum tube getter materials.

[Jack Puntawong]

Ray,

This is my design I am creating to dry deuterium. After the electrolysis process, deuterium is flow through soft copper tubing ( cheap and commonly found in refrigerator) to the silica gel to absorb some water vapor. Then the deuterium will flow down into a thermos where there are dry ice and alcohol mixture. This design can be improve by using liquid nitrogen (which is hard to get a hold of out side the USA) instead of dry ice. I think that this design will produce deuterium dry enough for a cash poor fusioneer. What do you think ?

The first picture is the design of the deuterium drying system

The second one is the electrolysis tube made up of copperwire, kimax biology glass and hallow tube. The cathode is inside of the hallow tube and the anode is coiled around it.

Jack Puntawong

[kcdodd]

This is a stupid question, but why can you not use a compressor to drive the pressure way up and then cool with a refrigerant line or something, then back through a pressure regulator to the fusor. Does that just not condense enough water?

[rkhertz]

Jack,

Your system as designed is an atmospheric pressure system, so you would have to use great care in opening the valve to the fusor. If opened too far, you could suck the contents of your deuterium generator into the drying train, or (even worse) the fusor. To prevent this, I would suggest that you include a trap to prevent this. One type is shown in the drawing in the attached file.

I would suggest putting a valve between the safety trap and the drying train so you can evacuate the drying train. The most reliable valves are probably metal vacuum-rated valves. I think the best greaseless glass valves are the Kontes Telon-plug vacuum valves. The plug should be given a very light coating of grease for best reliability and long life. Do not let a Teflon-plug valve get cold. It will leak if it gets too cold. The coefficient of thermal expansion of Teflon is much higher than glass. Greased vacuum stopcocks are also good, but they are messy, harder to operate, and liable to need regreasing at the most inopportune times. In general, don't use non-vacuum-rated stopcocks or valves. They have a high probability of leaking.

I would also suggest sealing the top of the generator to prevent dilution of your heavy water with moisture from the air. This would require incorporation of a one-way valve into the design.

A very inexpensive one-way valve can be constructed by making a small lengthwise slit in a small piece of rubber tubing. Attach one end to a piece of glass tubing extending into the D2 generator, and close off the other end with a piece of glass rod or a glass bead. This simple design will not withstand a very high back pressure and is not suitable for long-term storage of your valuable D2O, but you might find it useful for short runs.

You may not need both a desiccant trap and a cold trap. One or the other might do. The real fusioneers here could give you a better answer than I could.

A means of storing your D2 would be a desirable addition.

I hope this helps, and good luck in your efforts.

Ray

[rkhertz]

Peter,

The reactive metal route offers real possibilities. The metals that come immediately to mind are alkali metals such as sodium:

2Na + 2D2O → 2NaOD + D2

Alkali metals react readily at room temperature; however, they are flammable and you lose half the deuterium.

Other metals such as Mg, Zn, Cr, Ti, and U have been used to quantitatively liberate D2 from D2O. Using Mg as an example,

Mg + D2O → MgO + D2

U is not a realistic choice due to radioactivity, cost, and unavailability. Zn might be a poor choice because of volatility at elevated temperatures. Mg, Cr, Ti, or some other metal might be a good choice. Elevated temperatures are required for the reactions. A carefully designed reactor should give D2 with very low levels of residual water.

Barium is the active component in most getter materials and might be a good choice; however, barium is quite flammable, toxic, and reacts violently with water.

Some low oxidation state metal oxides are capable of generating D2:

3FeO + D2O → Fe3O4 + D2

Ce2O3 + D2O → 2CeO2 + D2

I will do a literature search over the next few weeks and report back with more details and some ideas for a system.

I think Peltier cooling has the possibility of reaching the temperatures required to trap water vapor. It would be worth a try.

Temperatures attainable with salt-ice-water baths:

-10°C can be achieved with a 1 to 2.5 ratio of calcium chloride hexahydrate to ice.
-20°C can be achieved with a 1 to 3 ratio of sodium chloride to ice.
-40°C can be achieved with a 1 to 0.8 ratio of calcium chloride hexahydrate to ice.

From http://en.wikipedia.org/wiki/Cooling_bath .

Ray

[Jack Puntawong]

Thank you Ray,

I will add one safety trap and a valve as you suggested. Thanks for all the advise and useful links you provided.

Jack Puntawong

[ab0032]

As a kid, when we made ice cream, we took ice, salt and alcohol, and if I remember correctly we got below -20C. Any cheap salt and any non-beverage alcohol for grilling or whatever will do.
I cant explain what the alcohol did, but it did something, maybe someone here can tell me?

>
> Temperatures attainable with salt-ice-water baths:
>
> -10°C can be achieved with a 1 to 2.5 ratio of calcium chloride hexahydrate to ice.
> -20°C can be achieved with a 1 to 3 ratio of sodium chloride to ice.
> -40°C can be achieved with a 1 to 0.8 ratio of calcium chloride hexahydrate to ice.
>
> From http://en.wikipedia.org/wiki/Cooling_bath .
>

[Nick Peskosky]

Alex,

The addition of NaCl (salt) to ice results in a depression of the freezing point of the resultant melt solution. As the ice changes phase (melts) it will continually remove heat from its surroundings. The alcohol bath is used to increase heat conduction from a thermally "hot" body (ie. the ice cream) while maintaining its liquid state. The coefficient of heat transfer for ice/salt alone is relatively poor. High purity isopropyl alcohol or acetone can be used as both have freezing points which are far lower then saline solutions. In the past I have used denatured alcohol and crushed dry ice to achieve cooling bath temperatures < -70*C.

[rkhertz]

To elaborate on Nick's reply, all phase changes from solid to liquid, from liquid to gas, or from a solid directly to a gas, require heat from their surroundings: that is, they are endothermic.

In the case of the alcohol-dry ice bath, the endothermic process that removes heat from the surroundings is sublimation (conversion of solid dry ice directly to gaseous CO2). The alcohol serves a single purpose - heat transfer agent - since dry ice and alcohol do not dissolve to an appreciable extent in each other. The temperature of the bath is controlled by the sublimation temperature of dry ice (-78 C at atmospheric pressure). Therefore, you can pretty much "nail" a temperature of -78 C with a dry ice-alcohol bath.

In the case of the salt-water bath, the water serves three purposes. First, the liquid water serves as the heat transfer agent. Second, the melting of the ice is an endothermic process that removes heat from the system. Third, the water participates in another endothermic process: the dissolution of the salt. If you dissolve salt in water at room temperature, you will observe that the temperature of the resulting solution is lower than the original temperature of water. This, along with the freezing point depression that results from dissolution of salt in water, is the explanation for the "ice cream freezer effect."

As more salt is added, a point is reached at which the entire mixture freezes solid. This is the "eutectic point," and represents the minimum temperature at which salt, ice, and liquid water can be in equilibrium. For the salt-water system , the eutectic point is at -21 C and a composition of 23% NaCl and 77% H2O (http://en.wikipedia.org/wiki/Eutectic_system ).

If you introduce a third component (e.g., ethanol) into the system, as Alex observed, you can further depress the freezing point, although I do not know how much. The ethanol-water eutectic freezes at -119 C; however, the dissolution of ethanol in water is not nearly as endothermic as the dissolution of NaCl in water. Furthermore, the addition of ethanol will suppress the solubility of NaCl, so the two effects will counteract each other to some extent.

[rkhertz]

A refrigeration-type liquid line dryer does contain molecular sieves; however, a used one may be contaminated or depleted. You might be able to regenerate it by heating while flowing clean, dry air through it. 100 ppb is pretty rigorous, though.

In grad school we had a glove box set up with a purification train consisting of about ten pounds of molecular sieves followed by about ten pounds of Dow Q1 catalyst. Q1 consists of finely divided Cu (with small amounts of other metals) supported on aluminum oxide. It reacts with oxygen at room temperature. The glove box gas (nitrogen) was circulated through the purification train by a Roots blower. With care, we could keep the water + oxygen concentration below 10 ppm for months. At that level a broken light bulb would stay lit for about a day before burning out. Regeneration of the purification train involved heating the molecular sieves and Q1 in a stream of 5% H2 in N2.

I have retrieved some articles on generating hydrogen by reacting water with active metals. I will start a new thread, as this one is getting long and active metals are off-subject (not that that has stopped me before).

[Peter Schmelcher]

I came across some teaching sales literature by Merck. The information includes the residual water content resulting from several different descant materials.

I am now leaning towards a simple drying method. Assume that a fuel cell produces D2 gas saturated with heavy water (I think this is very pessimistic) then about 3% of the heavy water will get wasted by drying. IMHO using an alkali metal or??? to grab the O2 out the water is just too much trouble. The exception would of course be if parts per billion or trillion contamination results in a big neutron increase.

[ibex44]

what about this method:
http://revroum.getion.ro/wp-content/upl ... t%2002.pdf ?
regards
Lothar Steinbock


> Ideas?

[Richard Hull]

Lothar, please do not reply and change the title of post. The original title was fine.

In future if you wish to reply, leave the title line alone. Just hit "reply" and in the text box, only, advance your thoughts in reply to the previous postings.

Richard Hull

[ed]

I don't know how low a temperature is really needed for adequate D2 drying, but I do know that R-12 type refrigerators with piston compressors typically can reach -40 deg C (or F - nearly the same there) when allowed to run open-loop. I have done a number of experiments with a mini-frig by filling the interior with styrofoam, bypassing the thermostat, and monitoring the evaporator temperature with a T/C. The line power required drops off to a minimal amount once it's done working the temperature down to the limit. This one took about 150-200W early on, then dropped off to less than 100W after chill-down, with no heat load other than insulation loss.

This operation range is not good for long term compressor life since the oil recirculation rate will be very low, but so far the unit has survived for many days of continuous operation. I also have a large industrial freezer that I used for food storage, that had its PID controller fail, causing it go open-loop full-on for several weeks. The food was still fine - in deep freeze at -40 deg. This unit's compressor seems to still be OK.

That temperature seems to be about the limit for the compression ratio of piston type units driving R-12. I have plans for future experiments with rotary compressors to see how low they can go in a single-stage refrigerant system with R-22, R-12, and R-134a.

For D2 drying, I'm picturing a chiller/cold trap that would get most of the D2O right at the electrolyzer, and then allow it to be returned - like a frig defrost cycle - to the supply after the run, to conserve as much of it as possible. If further drying is needed, it could be followed by absorption or chemical final stages to get the residual.

I'll be posting a related discussion in the vacuum forum in context of diffusion pump cooling.

Ed

[Roberto Ferrari]

Fellows

In the recent past I suggested to recovery palladium tubes from electrolysis cell from a damaged H2 generator.

See a couple of images: the complete cell and a disassembled one.
There you see the 8 palladium tubes (1/8" x 8").

Be careful in case you open one because may still be filled with NaOH solution (caustic).

Pay attention to one of the chronical failure modes: the corrosion and puncture of the tube ends (capped and brazed). The damaged tube will leak so you cut the good ones or you seal/isolate the damaged one, using the rest of the cell.

Regards

Roberto