Seawater electrolysis occurs when two electrodes are placed in seawater and an electric current is run between them. Water reacts at one electrode (the anode) to produce chlorine and/or oxygen gas (depending on the local concentration of and other factors), and at the other electrode (the cathode) to produce hydrogen gas:
- Anode: or
The product of interest is the hydroxide ions produced at the cathode: these react with ions to form , alkalizing the water surrounding the cathode. The alkaline water has a higher concentration of carbonate ions (), which makes it easier for corals to precipitate their own calcium carbonate and grow more easily. This alkalization is balanced by the reactions at the anode, which produce and acidify the surrounding water.
Electrolysis forms the basis of the Biorock technique, but it seems like it could be used in other ways to alkalize water and make it easier for corals to grow. In conditions expected in 2100 under RCP 8.5, a 100 watt power source (e.g. a small solar panel) could, in theoretically optimal conditions, alkalize 2.1 liters of water per second (or 7.6 cubic meters per hour) from a pH of 7.76 to 8.2 .
Some laboratory studies of the efficiency of seawater electrolysis have been done. One study  found that the cathode produced hydrogen gas and hydroxide ions at 99% efficiency with respect to the applied current. The main other factor determining the system's efficiency is the voltage necessary to get the desired rate of hydroxide ion production. A minimum theoretical voltage of 1.83V is required to initiate electrolysis , though in practice it seems to be around 2.1V . The higher the voltage, the more current will flow and react on a cathode of a given size, but the less efficient the cell will be (e.g. a 100W power source will provide 20A at 5V, but only 10A at 10V). Using measured rates of hydrogen production , at 6V a 100W power source requires a 592 cm2 cathode, and will alkalize water per the above conditions at 0.65 liters per second, or 2.3 cubic meters per hour. Lower voltages should be possible with larger electrodes.
The cell used in  produces approximately 20 ml h-1 of hydrogen with a 6V power supply (figure 7). The same cell requires 120 mA cm-2 of current to produce 85 ml h-1, giving 20 / 85 * 120 = 28.2 mA cm-2 of current at 6V. To utilize 100W of power we need to produce 100 / 6 = 16.7A of current, which requires a 16.7 / .0282 = 592 cm2 cathodic area.
To alkalize seawater from a pH of 7.76 to 8.2, 2.68e-4 moles of hydroxide are required for each liter of water . Each electron will generate one hydroxide ion (depending on the system's efficiency, but the cell studied here is 99% efficient at generating hydrogen/hydroxide). An amp is one coulomb of electrons per second, and a mole of electrons has a faraday (96485) of coulombs worth of charge, so 16.7A will generate 16.7 / 96485 = 1.73e-4 moles of hydroxide per second, which is enough to alkalize 1.73e-4 / 2.68e-4 = 0.65 liters of water per second.
To cross check some of these calculations, we can check that the volume of hydrogen produced in the cell makes sense. The cell's cathodic area is 1.58 cm2, so the total current is 28.2 * 1.58 = 44.6 mA, or 0.0446 / 96485 = 4.62e-7 mol s-1 of electrons. Two electrons are required for each molecule of produced, so the cell produces 4.62e-7 / 2 * 3600 = 8.32e-4 moles of hydrogen per hour. The molar volume of hydrogen gas is 22.4 l at 0°C, or 22.4 * (273.15 + 25) / 273.15 = 24.5 l at 25°C (it's not clear what the temperature was when measuring the produced hydrogen, so guessing here with room temperature), giving 8.32e-4 * 24.5 = 0.0204 liters of hydrogen per hour and matching up with the earlier estimate of 20 ml.
Other factors reducing the system's efficiency include:
- At the high pH around the cathode, and other minerals will accrete onto the cathode, as exploited by the Biorock technique. This adds electrical resistance and increases voltage requirements. It would be good to find a setup that minimizes the amount of accretion and/or makes it easy to remove.
- The efficiency of electrolysis decreases as the electrodes get further apart. When and are generated at the anode and cathode, other ions need to move around so that the water around the electrodes stays electrically neutral (this flow of ions is an electric current and completes the electrical circuit). The power requirement to move these ions is determined by the water's electrical resistance and means a higher voltage needed to get a certain amount of current flow. Rough calculations suggest that if the cathode and anode are separated by one meter, water resistance could increase voltage requirements by a factor of 10. Efficient electrolysis seems like it will require a way to keep the electrodes close enough that water resistance between them is low, yet separated in a way that the alkalized and acidified water do not simply mix and neutralize each other.
The conductivity of seawater in tropical seawater is approximately 5.3 S m-1 , equivalent to a resistivity of 1 / 5.3 = 0.19 Ω m. While the resistance between two electrodes in a large volume of seawater is a much more complicated topic, by the definition of resistivity this will give a resistance of 0.19 Ω between two 1 m2 plates on opposite sides of a 1 m3 cube of seawater.
The cell used in  conducts 28.2 mA cm-2 of current at 6 V per the calculations done above. With a 1 m2 anode and cathode, 0.0282 * 100 * 100 = 282 A of current should be conducted, giving a resistance of 6 / 282 = 0.0213 Ω. Separating the electrodes by one meter (and ignoring their separation in the original experimental setup, which should be small) increases the resistance to 0.0213 + 0.19 = 0.2113, a factor of 0.2113 / 0.0213 = 9.9 times the original resistance.
One appealing and maybe crucially important property of electrolysis is that it has applications independent from any use to alkalize water. The hydrogen gas produced at the cathode is a clean fuel that is becoming increasingly in demand, and seawater electrolysis is being studied for use in producing this fuel . If the hydrogen gas could be captured in a system used for alkalizing water, the hydrogen could either be stored and sold, or be used to generate power onsite and improve the system's efficiency.
Impacts of Chlorine
An important factor is the ecological impact of the chlorine produced at the anode. and related compounds are very toxic, especially to aquatic life. It depends on the setup, but it seems that the reaction taking place at the anode will mostly produce instead of . quickly dissolves in water and disassociates according to:
These compounds are widely used for disinfecting water, whether in drinking water supplies, swimming pools, bleach, or various sanitation and industrial processes. Because of this, their effects on various organisms have been studied. An ECHA assessment of chlorine's toxicity found the lowest No Observed Effect Concentration (NOEC) among several studies on different classes of organism to be 2.1 μg/l. This was divided by 50 to account for uncertainty, giving a Predicted No Effect Concentration (PNEC) of 0.042 μg/l . This is extremely dilute, the equivalent of adding one drop of bleach to 31 m3 of water (or 8200 gallons). Running the above 100W electrolysis cell can potentially generate the amount of chlorine in 4.7 drops of bleach every second.
A metric drop is defined as 0.05 ml , which will weigh about 0.05 g. Bleach is typically 5.25% , and is 35.45 / (22.99 + 16 + 35.45) = 48% chlorine by weight, so one drop of bleach has 0.05 g * 0.0525 * 0.48 = 1.3e-3 g of chlorine. This amount of chlorine reaches the PNEC concentration in 1.3e-3 g / 4.2e-8 g/l = 3.1e4 l of water, which is 3.1e4 / 1000 = 31 m3 or 3.1e4 / 3.79 gal/l = 8.2e3 gallons.
In the worst case, the electrolysis cell only produces at the anode. ions are produced at a 2:1 ratio with molecules. As calculated above, for 100 watts, 1.73e-4 mol of are produced per second, or 1.73e-4 / 2 = 8.65e-5 mol of per second. The molar weight of is 35.45 * 2 = 70.9 g, giving 8.65e-5 * 70.9 = 6.1e-3 g of per second. Compared with the amount of chlorine in a drop of bleach, we get 6.1e-3 / 1.3e-3 = 4.7 drops of bleach per second.
Fortunately, chlorine is very short lived in aqueous environments. It readily reacts with organic matter, but it also breaks down via photolysis, simply by exposure to sunlight. The half life of chlorine in ideal conditions (clear sky, summer sun at noon) at a pH of 8 is 12 minutes . The half life will be longer at a lower pH ( reacts most readily through photolysis, and is the dominant species only at a higher pH), in cloudy conditions, or when the sun is lower in the sky. Photolysis does not occur at all at night.
Another concern is that in seawater, reacts with the common bromide ion to form and . Bromine and chlorine are both halogens and have similar chemical properties: forms a similar equilibrium with and , which are also quite toxic to aquatic life and also break down rapidly in sunlight. It is not clear what effect the presence of bromide ions has on the lifetime of chlorine derived toxic compounds in seawater.
Using a half life of two hours seems a reasonable approximation for tropical locations during the day, though this is not necessarily conservative and will need to be carefully calculated and measured before deployment at anything other than a very small scale.
With this information about the toxicity and approximate lifetime of chlorine, it is possible to study and evaluate designs for managing the chlorine generated by an electrolysis cell. Several schematics are shown below:
- Design #1 simply places the anode directly into the environment.
- Design #2 places the anode into a compartment, which has an inlet and an outlet to the environment with constant flow and whose contents are well mixed. By keeping the inlet/outlet flow low, the chlorine will mostly break down while still in the compartment, and less will be released into the environment. The flow could potentially be extremely low, though some cycling of water is needed to keep the chemistry (particularly the pH) of the compartment stable.
- Design #3 modifies design #2 by placing a second compartment at the outlet, which has its own outlet to the environment. The second compartment is not mixed, and water has to flow through it from the first compartment to the final outlet. By the time the water finally reaches the environment, most of the chlorine it had when leaving the primary compartment will have broken down.
For the chlorine produced by the 100W cell and a half life of 2 hours, design #1 is expected to add a steady state amount of 63.4 g of chlorine to the environment. By using a 3m x 3m x 3m compartment with 0.2 l/s of inflow/outflow, design #2 is expected to add 4.9 g of chlorine, a 92% reduction over design #1. By adding a 3m x 3m x 1m secondary compartment, design #3 is expected to add 0.064 g of chlorine, a 98.7% reduction over design #2 and 99.9% over design #1. These are equivalent, respectively, to 2.5 l, 0.19 l, and 2.5 ml of bleach. The latter amount is equivalent to the PNEC concentration in 1500 m3 of water, which is a 23m x 23m x 3m volume and easily diluted in a lagoon environment .
These calculations are all very rough and preliminary. However, they do illustrate that by leveraging photolysis it should be possible to manage the release of chlorine produced by electrolysis into the environment and avoid any adverse environmental impact.
Compartments containing the anode and cathode need to be connected to each other to complete the electrical circuit and allow electrolysis to occur. This junction gives ions an opportunity to move between the compartments. This movement can have the following negative effects:
- The pH around the cathode can be lowered by moving ions, undoing the work done to alkalize the water there. For example, moving from the anode to the cathode or moving from the cathode to the anode can both have this effect.
- Chlorine compounds can move from the area around the anode to the area around the cathode, potentially building to concentrations harmful to the life (e.g. corals) around the cathode.
Movement can happen in any of the following ways :
- Migration occurs when ions move between compartments due to a gradient in electric potential. Migrating ions form an electric current and complete the circuit.
- Diffusion occurs when ions move between compartments due to a gradient in concentration. If an ion's concentration differs between the compartments, it will diffuse from the more concentrated compartment to the less concentrated one.
- Convection occurs when water moves between compartments, bringing any ions it contains with it.
It should be possible to largely eliminate convection with careful design of the junction and the flow of water across it. Both migration and diffusion can still occur, though. Both of these processes can be modeled. Consider the extension of design #3 below, with a compartment containing the cathode (and corals) which is connected to the compartment containing the anode. Propellers are used to generate a current which flushes the water around the electrodes and the junction between the compartments.
The new parameters to consider here are the speed of the current produced by the propellers, and the size of the junction between the anode and cathode compartments. If we take the parameters for design #3 above and use a 4 cm2 junction with a 25 cm/s current (about half a knot), migration and diffusion are expected to undo about 0.3% of the alkalization work done, and diffusion of chlorine to the cathode will reach PNEC concentration in 42 cubic meters of water .
As above, these calculations are pretty rough, but they suggest that ion migration and diffusion will be small concerns and that cell design can focus on the effects of convection within and across compartments.
Atmospheric CO2 Influx
When a body of water is alkalized, its concentration decreases. concentrations in the air and water are normally in equilibrium, so this decrease will cause to enter the water from the air and restore the equilibrium. This influx of acidifies the water, and needs to be taken into account. The rate at which atmospheric will enter the water to correct the imbalance depends on the sea state. The windier it is, the more turbulent the water surface will be, and the higher the rate. After raising a body of water from a pH of 7.76 to 8.2, at a wind speed of 6 m/s (probably above average for many areas of French Polynesia) the entering the water in a 366 m2 area is estimated to acidify the water at the same rate that the above 100W cell will alkalize it.
Estimate that at a wind speed of 6 m/s, 1.7e-3 moles of hydroxide are needed to neutralize the ions added by the hourly flux of through each square meter of water surface . With the 16.7A current of a 100W cell, 16.7 * 3600 / 96485 = 0.623 moles of hydroxide are generated per hour, enough to neutralize the ions added over a 0.623 / 1.7e-3 = 366 m2 area.
The model used for this estimate  is very approximate and it would be nice to verify the rate of this flux experimentally. In particular, sea state on a reef depends on many factors besides wind speed (fetch, water depth, current), so fluxes can vary widely even at the same wind speed.
If the hydrogen gas produced at the cathode is captured, it has its own economic value. A kilogram of hydrogen has the same energy as a gallon of gasoline, 33.7 kWh . The 100W cell considered above will produce 1 kg of hydrogen after running continuously for 66 days and consuming 159 kWh of energy. Efficiency here can be improved by lowering the voltage required to drive the process: the amount of hydrogen produced is related stoichiometrically to the amount of current run through the cell, and the lower the voltage the less energy is consumed by a given amount of current. At the theoretical minimum voltage, the cell would only consume 49 kWh. Other losses can occur due to not collecting all produced hydrogen, and to power requirements for managing hydrogen (collecting, transporting, storing) and running any ancillary equipment (e.g. propellers for circulating water).
The molar weight of is 1.008 * 2 = 2.016 g/mol. One kilogram contains 1000 / 2.016 = 496 mol of . At 16.7A, the cell transfers 16.7 * 3600 / 96485 = 0.623 mol/h of electrons, producing 0.623 / 2 = 0.312 mol/h of . The cell needs to run for 496 / 0.312 = 1590 hours, taking 1590 / 24 = 66 days and consuming 1590 / 10 = 159 kWh of energy to produce 1 kg of . At the theoretical minimum voltage of 1.83V, the cell will be 6 / 1.83 = 3.3 times as efficient and use 159 / 3.3 = 49 kWh of energy.
Selling the produced hydrogen will defray the costs of an electrolysis installation. If a design is sufficiently cost effective that most of its costs are covered by this hydrogen, or if it could even turn a profit, economic barriers to building large scale installations will be greatly reduced. Maximizing the cost effectiveness of a design is not simple: for example, lowering voltage requirements will lower energy requirements but could also increase capital costs due to larger electrodes or more expensive electrode materials. Still, this process is helped by ongoing work to reduce the cost of renewable power and improve the design of other types of electrolysis installations.
It's an open question as to how best this technology could be deployed. While experimental installations could mostly enclose a small part of a reef in barriers to restrict water circulation, this doesn't seem an ideal general purpose strategy. The natural topography of a reef could be leveraged to provide something similar in certain locations. For example, many atolls have no navigable passes to the ocean, with circulation in their lagoons consequently restricted. Deployment on such a reef-wide scale would require a solid understanding of the circulation patterns and residence times for waters on the reef.
- H.K. Abdel-Aal, K.M. Zohdy, and M. Abdel Kareem, Hydrogen Production Using Sea Water Electrolysis, The Open Fuel Cells Journal, 2010, 3, 1-7
- Jonathan T. Davis, Ji Qi, Xinran Fan, Justin C. Bui, Daniel V. Esposito, Floating membraneless PV-electrolyzer based on buoyancy-driven product separation, International Journal of Hydrogen Energy, Volume 43, Issue 3, 18 January 2018
- Assessment Report, Active chlorine released from sodium hypochlorite, Regulation (EU) No 528/2012 concerning the making available on the market and use of biocidal products, January 2017
- R. L. Jolley, J. H. Carpenter, Aqueous Chemistry of Chlorine: Chemistry, Analysis, and Environmental Fate of Reactive Oxidant Species, OKNL/TM- 7788, January 1982
- Allen J. Bard, Larry R. Faulkner, Electrochemical methods: fundamentals and applications, 2nd ed., John Wiley & Sons Inc, 2001, page 28
- Peter S. Liss, Liliane Merlivat, Air-Sea Gas Exchange Rates: Introduction and Synthesis, The Role of Air-Sea Exchange in Geochemical Cycling 113-127, 1986 (D Reidel, Dordrecht)