Renewable Domestic Power: the Edison battery

Why use the Edison battery?

A problem with any sort of offline power system that uses accumulators to store power is the environmental impact of battery construction. The truth is, most batteries are loaded with heavy metals and/or organic solvents which are nasty poisonous things.

Another problem is that the batteries themselves are essentially disposable items. Deep cycle batteries only last a few hundred cycles before they are junk. A home constructor cannot expect to refurbish them, so with luck they are recycled but very possibly they are landfill. Which brings me back to the first point: lead and cadmium don't belong in a landfill.

The ideal home power battery doesn't need to be light, since it is not going anywhere. It doesn't need to be efficient, since the energy in it is free. It doesn't even need a low self-discharge, since it will be regularly topped up by sun, wind, or water. All it really needs is to be non-toxic and to last forever. A battery like this was designed around 1910, and some of them are still in service today. It is the Edison or nickel-iron (NiFe) battery.

The Edison battery also has the advantage of being cheap. The materials used are pretty common and so it should be within the power of the home constructor to refurbish - or even construct - these things.

How does it work?

Each cell of this battery consists of a nickel electrode covered in nickel hydroxide, and an iron electrode covered in iron hydroxide, both dipped in a 4M solution of 80% KOH and 20% LiOH.

On charging, hydroxyl ions move from cathode to anode:-

Fe(OH)2 + 2e- -> Fe + 2OH-
Ni(OH)2 + OH- -> NiO2H + H2O + e-

...and on discharging, both these reactions are reversed:-

Fe + 2OH- -> Fe(OH)2
NiO2H + H2O + e- -> Ni(OH)2 + OH-

A naive way to construct a nickel iron battery would be to put a nickel electrode and an iron electrode in a solution of hydroxide, and pass a 3C charge through the cell.

Ni + 3OH- ->NiO2H + H2O + 3e-
2H2O + 2e- ->2OH- + H2

The result is a cell that is fully charged. On discharge the nickel oxyhydroxide is reduced to nickel hydroxide; and the iron is oxidised to iron hydroxide. So when this cell is discharged again, it finishes as two electrodes covered in nickel hydroxide and iron hydroxide.

I believe this process is the reason for the nickel-iron cell's legendary cycle life - if there are reagents missing, then recharging the cell causes them to be reformed from the underlying metals.

From these reactions it can be seen that the electrons are absorbed by the nickel electrode when discharging, so the iron electrode is the - and the nickel electrode is the +. Since the charge carriers in the electrolyte are OH- ions, this is an alkaline cell.

This chemistry is essentially the same as the nickel-cadmium cell, but iron is much cheaper and less poisonous than cadmium.

It is also possible to make this cell using sodium hydroxide, rather than potassium hydroxide: this is what high-temperature NiCd cells do. The cost will be a small increase in internal resistance.

What are the characteristics of the cell?

A trawl online suggests that the characteristics of NiFe cells are a fully-charged voltage of 1.25v per cell, a fully discharged voltage of 1.0v per cell, and a charge endpoint of 1.8v per cell.

They have a rather high rate of self-discharge: 0.33% per hour. That equates to about 8% per day, about 43% per week. about 67% in a fortnight, about 90% in a month.

A trawl online also uncovered a manual for the original Edison product.

It also uncovered the burp charging method. This same document gave typical values for the nickel hydroxide surface area of 17 to 19 m2/g.

Alternatively, typical Ni(OH)2 particle sizes are around the 20μm diameter. Since conductivity is through particle contact, a covering layer of 20μm should be good. The material has a density of 1600kgm-3

Because the nickel compounds that coat the positive electrode are not terribly conductive, the internal resistance is the limiting factor. This problem can be helped by increasing electrode surface area.

NASA have been experimenting with what they call "nickel felt electrodes", which are basically electrodes made of nickel wire wool. These should have a very high surface area.

Practical Construction

Reagents

An ampere-hour of charge is approximately 37.3mMol of electrons. Here are the weights of various active reagents in this cell:-

Reagent Molar Mass Mass for 1Ah Notes
Hydroxyl Ion 17.0g 630mg  
Nickel 58.7g 2190mg 8900kgm-3
Nickel Hydroxide 92.7g 3460mg 1600kgm-3
Nickel Oxyhydroxide 91.7g 3420mg  
Iron 55.8g 1040mg 7850kgm-3
Iron Hydroxide 89.8g 1675mg  

It is possible to buy nickel and iron wire from wires.co.uk. It is possible to buy sodium and potassium hydroxides from soap making websites.

It is also possible to buy nickel rods as welding rods, and it is possible to electroplate nickel from a solution of nickel sulphamide.

Surface area and dimensions

To obtain the required surface area, each gram of Ni(OH)2 will need to occupy about 20m2 of electrode. That means about 70m2 per Ah. Anything much less than this is going to give excessively high internal resistance. If we use nickel wire where 75% of the wire reacts, then we need 2.92g of wire, which has a volume of 3.28 x 10-7 m3. Modelling wire as a cylinder of area 2πrl and volume πr2l gives a ratio of volume to area of r/2. Substituting our area and volume gives a radius of 9.4nm, or a diameter of 19.8nm. That's silly.

However, the particle sizes of 19μm clearly don't match this area/volume ratio, so the material itself must be porous. By substituting in the densities and molar masses, we can see that to produce 20μm of nickel hydroxide requires 2.3μm of nickel. Given that using 75% of the wire means using 50% of the diameter, a diameter of 10μm would seem about right.

In volume terms, for each Ah of capacity, 0.25cc of nickel combines with 1.26cc of water to form 2.16cc of nickel hydroxide. So the initial charging process that creates the nickel must increase the volume of the cell by 0.65cc. A similar expansion is likely to occur at the iron electrode.

Actual construction musings

Some considerations of actual applications suggest that good sizes for batteries might be 75v/400Ah for an EV, and 75v/3200Ah for remote energy and for EV trucks.

A 400Ah battery at a nominal voltage of 1.25v stores 0.5kWh. 60 of them store 30kWh. 60 3200Ah batteries store 240kWh. These figures equate to 876g/7kg of nickel and 416g/3.33kg of iron. I suspect that with careful design it ought to be possible to make these reagents into a cell weighing less than 3kg/24kg. That gives a Wh/kg figure of 167, which is about the same as lithium ion. Perhaps that is optimistic. But it would be nice.

A possible battery design might be 350mm of 50mm PVC pipe, with a blind end at the bottom and a sleeve with an inspection fitting at the top. The inspection fitting needs a short length of blind-ended narrow (8mm) pipe in it, with 8mm flexible hose to allow venting. A 22mm endcap glued to the centre of the blind end should allow a mild steel rod to go through, with 600g of fine steel wool attached to it and 22mm plastic washers on the end to keep it fixed. At the top another fitting through the side terminates this as the iron electrode.

The 50mm pipe has a 300mm x 157mm strip of thin brass sheet with 100g of very fine steel wool attached to it, fluffed out, rolled into a tube, inserted, and nickel plated in situ with maybe 1200g of nickel. This forms the nickel electrode.

When the whole is assembled, 600ml of 4M NaOH solution is added, and the nickel electrode is connected to a positive supply and 1200Ah of charge passed. The electrolyte is topped with distilled water to keep it covering the electrodes, and the cell is kept well ventilated. During charging, the nickel oxyhydroxide will form on the nickel plated wire wool, and hydrogen will be evolved at the iron electrode. Then the cell can be discharged again, and this discharge will form nickel hydroxide at the nickel electrode and iron hydroxide at the iron electrode.

A block of 60 of these cells arranged as 6x10 should either be 550mm x 212mm x 400mm, or 354mm x 350mm x 400mm, depending on which of the rows gets offset to stack the cylindrical cells together. The weight of the block should be less than 200kg, and the terminal voltage should be between 60 and 75v.


This page is some notes on Domestic Power from Renewable Sources, and is written and maintained by Simon. At this stage these pages are constantly under revision. Thoughts and comments are welcome.