version: 2010
This document describes my construction
of a compact, high-throughput water electrolysis unit. I do electrochemical
research for my work, so it was fun to "homebrew" something like this. Even so, building this thing was not a light undertaking: it required more skill and
patience than I originally anticipated.
I prefer that you not share this document with other people without my permission. This is to achieve a modicum of access control, as I intend that people email me (dean_wheeler@byu.edu) and give me some kind of inkling that they know what they are doing and will not be endangering themselves and others with this information. Then I will with this document or a link. I know this can be annoying to the impatient, but I just don't want anyone to get hurt. So please use common sense and don't assume accidents will never happen to you, and follow the safety practices I describe at a minimum. Nevertheless, if you construct any devices based on the information here, or share the information with others, you alone are liable for any harm this may cause. The device described here has the capability if used improperly to seriously injure people and damage property. And if you are under age 18, do not use this information.
Most of this document describes my early design, circa 2000. Pay attention to the end of this document where I discuss some of the lessons learned and changes I made to a subsequent design.
In general, I will simply describe my knowledge pertaining to the operation of this device and how I constructed it. If the reader chooses to construct a similar device it can be adapted to the tools and materials available, within the constraints of safety. The materials I use for the electrodes and electrolyte are similar to those used in commercial electrolysis cells. However, my cell is of course much smaller than the industrial units, runs at lower temperature, and does not collect the gases separately. This last fact requires an extra measure of caution in operating my unit: while hydrogen and oxygen gases when mixed are stable at room temperature, they can explode if heated by a spark or flame.
I think it's important to have a rudimentary understanding of a few electrochemical principles. Knowing these things allowed me minimize the amount of trial and error necessary to produce a functional electrolysis unit.
In its most basic sense, an electrochemical cell is composed of two electrodes with liquid electrolyte in-between. Separate chemical reactions take place on each electrode that depend on the passage of electrical current from one electrode to the other. Technically speaking, a "battery" is more than one cell put together in a single package, but in common usage it can be made of a single electrochemical cell. My electrolysis unit is composed of four cells in series (end-to-end), so that the cell voltages add, and the same current flows through each cell.
In electrolysis of water, hydrogen (H2) gas is generated at the negative electrode (called the cathode) while oxygen (O2) gas is generated at the positive electrode (the anode). The respective electrode reactions in alkaline solutions are:
4H2O + 4 e- ---> 2H2 + 4OH-
4OH- ---> 2H2O + O2 + 4e-
This means overall it takes 2 water molecules to make 2 hydrogen and 1 oxygen gas molecules. This requires the passage of 4 electrons through an external circuit (the power source).
While technically the only chemical reactant needed in all this is water, it is necessary that the electrolyte be composed of more than just water to promote its electrical conductivity, otherwise an undo amount of energy is wasted as heat in trying to drive ionic current through the electrolyte. Also the choice of electrode material is important as this determines the ease with which the reactions are catalyzed and the service life of the device. Typically, the rate of reaction is measured as current density, such as the number of amps per square centimeter of electrode surface area.
Let us propose to run an electrolysis cell that is composed of two parallel sheets of nickel metal separated by a constant 0.9-cm gap. In the gap is an electrolyte of 10 wt% NaOH in water. Let's attempt to operate the cell at 25 oC, atmospheric pressure, and with a current density of 0.15 amp/cm2. The total voltage required is composed of two parts:
V = ER + EIR
ER is the "reversible" or equilibrium voltage of the decomposition reaction and is 1.23 V. This is the bare minimum voltage you need across a single cell to get the thing to produce any gas at all. EIR is the "irreversible" voltage or "overpotential" and is the price you pay for being impatient. Electric potential is required to drive the reactions at the electrodes and push ions through the electrolyte. It's like an internal resistance in the cell. The higher the current density you try to run the cell at, the higher the overpotential will be. For our particular example, the overpotential is about 2.2 V after all the complexities are taken into account. Keep in mind that most of this energy is going into heat. In a sense it is wasted energy, but again it is the price you pay for running the cell at a particular current. There are several ways to reduce these losses, but for our purposes the most effective one would be to use larger electrode surfaces so that the current density is lower. This would allow less voltage across the cell to generate the same total current (amps), at the cost of having a physically larger cell.
Sticking to our stated conditions, we see that 3.4 V is required to run the cell. In actuality, because of the heat realeased during operation, the solution will tend to warm up, and the higher temperatures will tend to reduce the overpotential and thus let the cell run more efficiently. Also, if the electrode surface area is increased due to a rough finish, this will improve things. On the other hand, the cell will demand more voltage if we attempt to produce gas at elevated pressures, use a less concentrated electrolyte, or produce so much froth in the electrolyte that the bubbles act as a significant insulating barrier to current flow.
What happens if the voltage is applied at a level higher than the 3.4 V we calculated? The reactions will run a bit faster, of course. However, there are limits to how much current you can push through a cell. Moreover, undesirable side reactions will also run faster at higher applied voltages. These side reactions generally will destroy your electrodes and engage in weird chemistries that could be dangerous. The problem is mitigated if the electrolyte is at very low concentration (nearly pure water) since in this case most of the voltage will go to the ohmic voltage drop rather than the reactions. Nevertheless, I recommend that you never run an electrolysis cell above 3.5 V. If your cells are designed and sized properly, you should get 0.1-0.2 amp/cm2 current density from your cell at a cell voltage of 3 V.
As stated before, the electrodes need to be of a material that will allow the respective reactions to proceed quickly with minimal overpotential (electrical encouragement). From what I have read, standard commerial electrolysis units use either nickel or steel for the hydrogen-producing positive eletrode and use nickel for the oxygen-producing negative electrode. For this reason I used nickel-plated steel for all the electrodes on my prototype (now defunct) electrolysis unit. The electrodes worked wonderfully, but the exterior container had problems, and I decided I wanted somewhat greater throughput. For my second-generation unit, I use all stainless steel electrodes, also with good success. I think that commercial operations prefer nickel for its slightly better efficiencies and longer service life (it is more corrosion resistant), but obviously it is harder to obtain for the average person unless your cousin works at a metal-plating shop.
All my electrodes are stainless-steel
plates. I made five plates, 1.5mm thick, cut as shown in the
picture using a band saw. Notice that two of them are taller than
the other three--they will form part of the exterior walls, as
will be explained later. Notice the small extensions on the lower-right
corner of each plate. These tabs are designed to protrude through
the protective enclosure of the electrolysis unit to allow for
electrical contacts. I roughened all the surfaces with emory
cloth, giving them a grain in the vertical direction, so they
would have higher surface area but still be able to shed bubbles
effectively. This also had the benefit of cleaning off any
undesirable oxides accumulated on the surfaces. In fact, it is
important to make sure all surface impurities and organic
contaminants are removed through abrasion and scrubbing with soap
and water.
The areas of the plates in the picture outlined with dotted green lines are the final working areas of the electrodes after the unit has been assembled, and amount to 45.5cm2 on each plate. The regions outside the dotted green rectangle will be covered with epoxy. Based on those working areas I estimated my unit would draw 8 to 10 amps current. It is important that the power supply or battery that is intended to run this thing be capable of delivering the required voltage (12v) at that current level. I use a small sealed lead acid battery (7amp-hr, purchased online for US$25), which can deliver this current without straining too much. Fortunately, if the electrodes end up being oversized in terms of current flow after the unit is finished, this can be easily remedied by putting less electrolyte in (so that electrolyte doesn't contact the full surface area of the electrodes), or by lowering the concentration of electrolyte (so that ohmic losses are greater). Of the two options the latter is safer, because less electrolyte means more gas space within the unit. This gas mixture of hydrogen and oxygen is explosive, so the less of it in the unit the better.
If stainless steel plate and a band saw to cut it are not available, other options for electrodes include using thin stainless-steel sheet cut from the side of metal containers. However, do NOT use any metals for the electrodes other than nickel or steel or possibly chrome. Aluminum and copper-containing alloys are not acceptable because they will corrode. A few other exotic metals such as cobalt are acceptable, but are not readily available anyway. Technically, carbon electrodes will work, however they tend to flake and erode under the vigorous bubbling action. It is preferable to use only one kind of metal for all the electrodes, to eliminate the possibility for unintended reactions when the unit is not in use.
Industrial electrolysis units use either highly concentrated KOH or NaOH. The reason is you need a high-conductivity electrolyte solution. But you also don't want the electrolyte to corrode the electrodes. Steel and nickel survive much better in alkaline solutions than acidic ones. In fact, steel will corrode less in strong alkaline solutions than in pure water, due to the formation of a dense protective oxide film on the surface. (Standard rust is also an oxide film, but it is not so dense). I do not recommend using substances other than relatively pure KOH or NaOH to make the electrolyte. For instance, if you use common table salt (NaCl) in water, when you run your cell your anode could potentially produce poisonous chlorine gas rather than oxygen. Furthermore, the chlorine in solution makes the steel electrodes corrode like crazy. I hope you get the idea that common household stuff can do unexpected stuff and even be dangerous when put in an electrochemical environment.
I purchased NaOH ("lye") in pellet form at the hardware store (it is used as a drain-opening agent). This is nasty stuff--wear eye protection and rubber gloves when working with it. After all, it is used to dissolve hair, grease, and the like in your drain, so it will chew up your bare skin in a short time. A caustic or alkaline solution feels "slippery" to the touch because it turns your skin oils into soap (saponification), so if you ever get that slippery feeling, rapidly rinse it off.
Even though my unit requires only 85mL of electrolyte, I made up a 1-liter batch of 10 wt% NaOH in water, so that I wouldn't need to be constantly making smaller batches. For those with a chemistry background, 10 wt% is around 2.5 molar. Always add the pellets to the water and not vice-versa. Do not add the pellets all at once: the dissolving process releases heat and you don't want the solution to get too hot. I store the solution in a HDPE plastic container with a screw-tight lid, and with a warning label. It is better to use distilled water in the mixture (available in the 'bottled water' section at the grocery store) because the impurities in tap water may decrease performance of the electrolysis.
You must choose inert construction materials for the electrolysis unit that can survive in a caustic environment, with joints that can maintain a water- and air-tight seal in spite of thermal expansion, and are tough enough to withstand possible internal pressure and bumps and jolts to the unit without cracking. Even with the right materials, it still requires considerable skill and patience to build a container for the electrodes that meets these requirements.
The following materials are suitable for operation in contact with a caustic solution at or slightly above room temperature.
I used plexiglas (acrylate), HDPE, and epoxy in my design. I recommend the use of a high-quality epoxy or a solvent-based glue.
Materials which are not suitable for caustic service include:
Do not assume a material is suitable just because it is not on this list!
Now, some words about cell configuration. We want to put all our cells into a single container, since this is more convenient than trying to make several identical, but physically separate cells. Recall that a cell is composed of two electrode surfaces separated by electrolyte. The voltage across each cell, as stated above, should be about 3v. If our power source put out 3v then we could arrange a bunch of cells in parallel like this:
__________________________
where the red and green lines are the electrode plates in the electrolyte. Generally, however, we want to put our cells in series so we can use higher voltages and lower currents to get the same production rate. A naive attempt at a series configuration might be:
However, the current will take the path of least resistance and bypass the electrodes in the center, rendering them inert. Our stack of electrodes is effectively only a single cell, and applying a large voltage across this single cell would be a waste, if not detrimental. The solution is to eliminate the electrolyte path that allows the "shunt" current to take place:
This improved series configuration is called a bipolar stack because the plates in the middle are bipolar, each with one side acting as anode and the other side acting as cathode. This is the cell configuration I use in my unit. Note that very small amounts of electrolyte fluid contact between the cells are acceptable, as the amount of current that will leak through is proportionately small. And in fact, the leaks may be desirable in equilizing the electrolyte fluid levels between cells. (Of course, we do not want any electrolyte to leak out of the exterior enclosure.) The alternative, if any leaks are not sufficient to equalize the fluid levels, is to do it by tipping the unit (with the cap on, so electrolyte does not spill out).
Note that with the plates arranged as above, it is possible to power the electrolysis unit with other than 12v, provide there are electrical connections to the interior plates, such as
We want the electrodes to be close together to reduce the voltage penalty of the electrolyte. However, if the electrodes are too close, the gas generated on the electrodes will push too much liquid out the top of the cells as foam/froth. This is just like a opening a shaken bottle of soda pop. In the extreme case, the froth from the different cells will come in contact and cause undesirable shunt currents, and the froth can possibly get pushed out of the gas outlet at the top of the electrolysis unit. This froth-effect can be reduced by rendering inert the upper portion of the electrode plates. This keeps the cells isolated from each other in this region while at the same time does not contribute further to the amount of gas generated. I accomplished this by gluing a strip of plexiglas along the top of each of the interior plates and also by painting epoxy along a narrow upper portion of all the plates, using masking tape to keep the epoxy confined to where I wanted it. This modification to the bare electrodes was performed prior to final unit assembly.
In retrospect, though, this "masking" step on the top of the electrodes is probably unnecessary as long as you don't fill the electrolyte up to the top of the plates. In other words, plan on the electrolyte volume expanding considerably once the cell gets going (see video below). You want the frothy liquid level to just reach the top of the plates. This can take some adjustment to find the proper level.
Also, if the unit is run at several psi of positive pressure (created by making the gas outlet from the unit somewhat constricted), this will reduce gas bubble sizes and thus reduce the amount of froth. I tried to accomplish this by putting a small wad of cotton in the outlet line where it attaches to the cap, however I doubt this did much. Anyway, the cotton has the additional advantage of partially filtering out the micro-droplets of electrolyte that are inevitably suspended in the gas leaving the electrolysis unit.
(What follows is a description of my 2nd generation unit. At the end of this document you will see that I would do a few things differently if/when I make a third generation unit.)
If you think of the enclosure to my unit as a rectangular solid shape, it has six outside surfaces. The two outside (larger) electrode plates serve as two of those surfaces; the remaining four surfaces are made of 6mm-thick plexiglas (acrylic) pieces each cut to dimensions 88mm x 52mm. In addition, there is a screw-top and cap (from a shampoo bottle) made of HDPE. Grooves--1mm deep and 1.5mm wide--are cut into the plexiglas pieces lengthwise. The grooves are spaced 7.5mm apart (on center). The metal electrode plates fit into the grooves, thus ensuring a uniform 6mm gap between electrode plates, after taking into account the plate thickness. The interelectrode gap distance could have been increased to as much as 10mm with satisfactory results. Also, the four plexiglas pieces have 1-mm-depth rabit cuts on the ends to allow better joints between them. (I realize that the rabit cuts are only needed on the ends of the top and bottom pieces rather than on all four pieces, but I got a bit overzealous while cutting them.)
You may notice that all five of my electrodes have tabs which protrude outside the enclosure. I did this so I could measure the voltage difference between each pair of adjacent plates to make sure each cell in the unit was running fine. However, this introduced additional complication into the design--tabs on only the outside two plates are necessary. Notice the vertical slots along the bottom edge of one of the plexiglas side-pieces to allow the metal connection tabs to stick out of the enclosure. The tabs are about 6mm tall and 14mm long, and permit a female quarter-inch electrical coupler to be attached.
Everything was glued together with epoxy.
I use a high-quality low-volatility marine-grade epoxy that comes
in quart sizes. One of my complaints about the small two-barrel-syringe-applicator
epoxies available in all the hardware and autoparts stores is
that these epoxies are usually quite thick/viscous. This makes
them convenient to work with in most applications, but wouldn't
have worked as well here. I needed the epoxy to be able to
penetrate into cracks well and to form strong, airtight bonds. I
find that when I mix the two components of those thick viscous
epoxies a lot of bubbles get permanently trapped in the mixture,
and I felt that would compromise this design. This enclosure has
to keep a highly caustic solution contained under somewhat
positive internal pressure, with zero gas or liquid leaks.
Frankly, this is the toughest part of building this electrolysis
unit. I found that the areas most likely to have leaks were the
corners where three pieces are coming together in one place.
I also believe that the epoxy I use has higher chemical resistance than the generic hardware-store type. I purchased mine at TAP Plastics (a company with a few store locations on the U.S. west coast) for US$25 for 1.25 quarts (total volume of both components). Marine supply stores might also be a good source of epoxy since fiberglassing with epoxy is extensively used in boatbuilding. I use two well-cleaned-out (with acetone) liquid-hand-soap pump dispensers to dispense the two components for making small batches. It is also nice to have a supply of latex gloves on hand to wear while working with the stuff.
My epoxy has a 15-30 minute pot life (if
it is thinly spread it takes longer to harden due to more heat
dissipation). After mixing the epoxy, I placed it into a 10cc
syringe fitted with a fine plastic tip. This was invaluable in
allowing me to squirt the epoxy where I wanted (in the grooves,
for instance) without dripping it all over the place and making a
mess of things. If the epoxy were to have dripped all over the
electrode plates while I was installing them, this would have
created inert spots where no reaction could take place. In order
to preserve the syringe, I immediately clean it out with acetone
before the epoxy has hardened inside.
Before I even considered putting caustic electrolyte in my soon-to-be-finished cell I wanted to be sure it could hold air to a reasonably high pressure. After the leakproofing in step 9 above, I pressure tested my unit to 50psi and found no leaks (I could have gone higher, but didn't want to push my luck). Even with that margin of safety, I keep the unit inside a secondary container--a smallish plastic 'tupperware' box--while it is busy electrolyzing, in case a leak or more serious problem were to develop. I don't want concentrated NaOH going all over the place. period. It is also a good idea to keep a bottle of water nearby whenever operating the unit, so that in case of an emergency you have a means of washing the NaOH off. I always wear safety glasses when I approach the operating unit. I also don't use the unit for long-term storage of electrolyte: after I have finished using the unit for the day, I transfer the electrolyte back into another container and rinse out the electrolysis unit with distilled water and let it drip-dry while inverted.
The presence of the explosive mixture of
H2 and O2 gases in the electrolysis unit is
unavoidable without going to great lengths to keep the two gases
separated from the outset. Nevertheless, in my design I tried to
minimize the amount of gas space in the unit. If this gas mixture
were to combust, it would generate pressures of at least 240psi.
I think my current electrolysis unit could actually handle that
pressure without fracturing, but I don't want to find out. In
fact, this very thing happened to my much flimsier prototype unit.
The whole top-piece popped off and several mL of NaOH splattered
inside the secondary container. One idea I'd like to try it to purposefully weakened the burst strength of the cap that threads
onto the top of the unit so that the cap will be the first (and
hopefully only) thing to fail in the event of a catastrophe. This
is the same principle as having a "pressure relief valve" on a home water heater. I also now ensure that the point where the combustible gas enters the rocket is below the water line at all times so that the combustion can't get back to the electrolysis unit.
When connected to my fully charged 12v battery (which actually
puts out more like 12.5v at these currents) I measured a current
of 9.4 amps running through the unit. This means that I am
generating 430 mL of gas mixture (67% hydrogen, 33% oxygen) per
minute. In addition, the unit is generating about 60 watts of
heat in steady-state operation. The conversion factor between
amps and mL-gas-per-min is 45.6 for a four-cell unit. Incidently,
9.4 amps is a lot of current, so it is important that the
electrical connections be rated to take that kind of current. I
use a household-style light switch between the battery and my
unit so I can switch the current flow on and off in a hurry.
The bulk of this document was written in October 2000. This section is written in 2010. After having used my second-generation electrolysis cell several times I discovered a problem in its design which I remedied in a successive design iteration.
Basically this is the problem: if I ran the unit for more than a couple minutes it heats up due to resistive heating (basically, about half the power applied to the unit is converted to heat; the other half causes the electrolysis). Upon heating, the steel electrode plates and the acrylic and epoxy all expand, but unfortunately in different amounts. This mismatch in expansion caused stresses which opened a hairline crack where the metal tabs protrude from the enclosure. I noticed a small amount of electrolyte seaping through this opening.
What is the solution? When I made a third-generation unit, I had the tabs protrude from the enclosure near the top, above the liquid line. This way if an opening develops the electrolyte will not leak out. Next, I only have two tabs protrude--only those for the outer-most plates across which the 12-13v is applied. Third, I tried not to use the steel electrodes as such an integral part of the structure, that is, I had them sort of loosely fit within the grooves on the interrior of the enclosure, which is still constructed of acrylic sheets, now on all 6 sides. I more carefully sealed the opening in the acrylic where the two steel tabs protrude. Fourth, I made the plates about twice as large in area and use six stainless steel plates (instead of 5) making 5 cells in series (instead of 4). This means there will be less voltage per cell, but the larger size allows the plates to operate more efficiently (less current density). The net result is at least 25% more gas for the same electrical power input. Fifth, for safety I cut the concentration of electrolyte in half. And I found a source of KOH instead of NaOH to use becuase KOH has a bit more conductivity, again making things more efficient so there is less heating during operation.
All contents copyright ©2000-2017 Dean Wheeler
Disclaimer: This document is for informational purposes only and is provided free of charge, obligations, and liability to the author. The author expressly disclaims any warranties on the accuracy and fitness of the information provided on this page for any particular purpose.