The hydrogen economy revisited

Hydrogen is not a primary energy source. It does not occur naturally on earth and can only be extracted from water or other substances by the application of large amounts of energy, some of which can be used to perform useful work when the hydrogen is oxidised to form water. Viewed in this light, hydrogen is just a way to store large amounts of energy in a transportable form. Hydrogen can be generated by using any convenient form of energy to extract it from the source materials. If renewable energy is used it is carbon neutral. It can, of course, be generated using nuclear energy from fission, but this seems unlikely when you consider how little uranium is available (see Fission power). Introducing a fission driven hydrogen economy would certainly be dependent on large-scale use of fast-breeder reactors.

The 'traditional' hydrogen economy

Renewable hydrogen

You generate hydrogen renewably by splitting water into hydrogen gas and oxygen, usually by electrolysis, though biogenic hydrogen may be a possibility. The oxygen is released into the atmosphere while the hydrogen is cooled and compressed until it liquifies. The liquid hydrogen would be taken to where it is needed and used to fly a plane or fuel a car. Here it releases energy by combining with oxygen to regenerate the water that was split to generate it in the first place. Electrolysing water to obtain hydrogen requires a great deal of energy: you put as least as much energy into electrolysing water as you get back later from a fuel cell or by burning it. Liquifying hydrogen also requires a lot of energy: possibly as much as was used in the electrolysis process. I haven't seen any firm figures on the overall efficiency of hydrogen use, so here's a best guess:

Another contender, direct conversion of solar energy to hydrogen, emerged in 2012. This process uses a thin (20-30 micrometer) layer of ferric oxide (rust) which is deposited on a transparent insulating substrate so the ferric oxide forms a fractal structure with a very high surface to volume ratio. This is critical: the thickness of the layer needs to be at least 30 uM or the solar absorption efficiency is too low (under 18%), though this can be improved if the substrate is grooved so that unabsorbed photons are reflected or refracted back onto the rust layer. Up to 72% absorption is claimed for this technique. The surface to volume ratio within the rust layer must also be as high as possible so that the electrons liberated when ferric oxide absorbs solar photons can escape to split water molecules rather than being reabsorbed by the ferric oxide: this limits the thickness of the rust layer to a few tens of micrometers. Ferric oxide is inherently more efficient at splitting water than silicon: the electrons need to carry energy equivalent to 1.23eV to split water. Silicon emits electrons of 1.11eV or higher energy while ferric oxide emits electrons carrying 2.1eV or more. This means that all electrons emitted by ferric oxide have the energy to split water molecules, which is not the case for silicon. The end result is that lab systems have already shown energy conversion efficiencies of 4-5% with the prospect of matching the best silicon PV systems generating hydrogen by electrolysis (10%). If this can be achieved, the ferric oxide process will win on cost because ferric oxide is much cheaper to make than silicon PV cells and needs no toxic reagents to make it. The remaining problem is separating the oxygen and hydrogen: while electrolysis provides physically separated sources of these gasses the ferric oxide process does not. However, it seem that, unlike electrolysis, the ferric oxide process will work with low-quality waste water that contains organic materials and that these will be preferentially oxidised, thus capturing the oxygen and allowing relatively pure hydrogen to be collected. The research on using ferric oxide for solar hydrogen production is described in New Scientist, 26 January 2013 page 34.

Current hydrogen sources

However, at present hydrogen isn't produced renewably: it is at least as big an environmental nightmare as fossil fuel because it is manufactured by splitting the methane in natural gas to release hydrogen. This process not only requires a substantial energy input, probably produced by burning still more fossil fuel, but releases a carbon-containing exhaust stream, almost certainly in the form of carbon monoxide and dioxide.

Don't expect hydrogen to appear soon at a gas pump near you, ready and waiting to be poured into your shiny new eco-friendly hydrogen fuelled car. The technology is not ready yet. This is particularly true of the high output fuel cells needed to replace internal combustion engines. Despite the hype, large fuel cells have not advanced significantly past the 40 kW produced by the units in the Apollo spacecraft. Thats 53 horsepower in old money, or about 43 hp at the wheels of an electric car. So, if the Hydrogen Economy does appear, expect to drive a modified fossil fuel engine long before you have an electric vehicle with a fuel cell. Even if the technology was ready to be rolled out, economists consider that at least another 10 years would be required to build enough non-fossil hydrogen plants to meet demand and to design and install a hydrogen distribution system. Of top of that there are a few inherent disadvantages that no amount of technology can overcome:

You get a lot of water produced when you burn hydrogen or run a fuel cell on it. This table shows the tonnes of carbon dioxide (CO2) and water (H2O) produced when one tonne of hydrogen or jet fuel is burnt:

Fuel Composition CO2 H2O
Hydrogen H2 0 9.00
Jet Fuel A C12H26 3.11 1.38

Hydrogen produces three times as much energy per tonne as Jet Fuel A when its used in an engine or fuel cell (New Scientist, 9 July 2005 page 23). After you factor that in, that hydrogen-powered jet flight will still dump 2.2 times as much water in the atmosphere as the same flight powered with fossil fuel. The relative water production figures for petrol and oil versus hydrogen are unlikely to differ much from jet fuel, so the same calculation applies to hydrogen powered road vehicles. While water in the exhausts of hydrogen powered vehicles is just replacing the amount that was split to obtain the hydrogen, its now in a very different place. This has knock-on effects:

In 2005 aviation contributed 3% of the human-generated carbon dioxide input to the atmosphere. This is projected to rise to 6% by 2050. Commercial jet travel's input is thus just a small part of the human-generated green-house effect. However, water vapour from jet travel has already been shown to have a direct, deleterious and measurable effect on the weather. I think on balance I'd prefer to stick to fossil fuel for civil aviation if the only alternative is liquid hydrogen.

However, don't just take my word for this. Robert Zubrin, who is an aerospace engineer, is even more scathing about the hydrogen economy than I am. The Hydrogen Hoax is his analysis. He is president of the R&D form Pioneer Astronautics.

Hydrogen on demand

A "hydrogen on demand" system, also known as a hydrogen battery, generates the hydrogen as it is needed at the place where it will be used, extracting the energy for this from a precursor substance. The resulting hydrogen is then used to run an engine or cause a fuel cell to generate electricity. Such a system has major advantages over the 'traditional' hydrogen economy:

Hydrogen gas is light and bulky, so must be liquified for bulk storage and transport. Liquifying it uses 40% of the energy content of the stored hydrogen, so almost anything that can avoid this requirment is worthwhile. Storage tanks for liquid hydrogen are heavy, bulky, and consume energy running the cryogenic systems that are needed to keep the hydrogen in a liquid form.

Possible "Hydrogen on demand" systems

These differ in precursor material and hence in the chemistry involved.


Engineuity is working on a system using the reaction between aluminium and water to liberate hydrogen. The end of a roll of aluminium wire is lit and dipped into water. The resulting mixture of steam and hydrogen is burnt in a conventional engine. The system may also use magnesium or boron instead of aluminium. The metal oxide would be returned to a reprocessing plant where the metal wire would be regenerated via electrolysis. No efficiency estimates are available. The system is claimed to be a zero emission technology with the possibility of eliminating nitrogen oxides completely if the water is replaced with a solution of hydrogen peroxide.

The company expects to have a prototype running in 2009.


This system has been proposed by a team from the University of Minnesota and Weizmann Institute at Rehovot led by Tareq Abu-Hamed. Unlike some of the other schemes, the complete fuel cycle's efficiency has been evaluated and it has been designed from the outset as a zero emission system.

Pure powdered boron is reacted with water vapour at 800°C. The boron combines with the oxygen in the water molecules to form boron oxide and hydrogen gas. The hydrogen is used in a conventional engine or fuel cell to produce energy and water, which can be recycled into the boron reactor. The boron oxide is periodically swapped for a fresh supply of boron. The oxide is returned to a solar-powered plant that reduces it to pure boron. This is achieved by reacting it with magnesium. The resulting magnesium oxide is in turn reacted with chlorine, which forms magnesium chloride and returns oxygen to the air. Solar energy is used to melt and electrolyse the magnesium chloride at 700°C. Both the magnesium and the chlorine are continuously recycled within the plant. The result is zero pollution power because the engine emits only water and the boron reduction process emits only oxygen.

The substances needed to power a vehicle, boron and water, are lighter, safer, and easier to handle than hydrogen in high pressure cylinders. Boron, a black powder is highly flammable. It does not spontaneously combust or react with liquid water at normal temparatures, so it may be safer to store and transport than petrol. It is three times denser than petrol, so occupies proportionately less space.

27 kg of boron and 68 kg of water are needed to produce 7.5 kg of hydrogen, which contains the same energy as 60 litres (48 kg) of petrol and would occupy 30% more space. On the assumption that the solar cells driving the electrolysis process is 35%, the overall efficiency of this system is 11% - about the same as a current petrol engine.

Magnesium hydride

Safe Hydrogen is developing a hydrogen storage and generation technology using magnesium hydride slurry as a pumpable hydrogen fuel. Hydrogen is generated as needed by mixing the slurry with water in a special mixing device. A pilot project used lithium hydride slurry to fuel a converted internal combustion engine in a Ford truck.

The hydrogen density of magnesium slurry is twice that of liquid hydrogen. It is claimed to be safe to transport and store at normal temperature and pressure. The slurry can be distributed using the existing fossil fuel infrastructure of tanks, trucks and pipelines. Depleted slurry can be completely recycled into fresh hydrogen-dense slurry. The overall thermal efficiency is said to be similar to liquid hydrogen, i.e. somewhere between 8% and 18% depending on the source of the hydrogen.

Sodium boro-hydride

Daimler-Chrysler's Natrium project demonstrated a car using sodium borohydride as the precursor substance. The vehicle was a standard Town & Country petrol minivan fitted with the replacement drive train.

When a 25% aqueous solution of sodium borohydride is passed over a catalyst of the precious metal ruthenium, it reacts with water to produce hydrogen and a solution of sodium metaborate. The hydrogen was used in a fuel cell to generate electricity which drove a 35 kW electric motor and recharged a 40 kWh lithium-ion battery, which was also charged from regenerative braking. The hydrogen generator and fuel cell were supplied by Millennium Cell, which went into voluntary liquidation in October, 2008. The car had a top speed of 130 kph and a range of 500 km. The drive train is heavy: the Natrium was about 226 kg heavier than the original petrol car.

The cost, in both energy and financial terms of reprocessing sodium metaborate to regenerate the sodium borohydride is currently high. In 2003, sodium borohydride cost about 50 times as much as the energy-equivalent amount of petrol. The current Brown-Schlesinger process is 50 years old and is chemically quite inelegant, using methane, processed boric acid and metallic sodium as raw materials. The main source, Rohm and Haas's specialty chemical division only produced about 15 tons a year. In 2005 Millennium Cell claimed to have made a breakthrough in producing sodium borohydride but details were not available. They were predicting a 4 fold drop in cost, but while this might have made their technology suitable for military use, it would have been too expensive to replace fossil fuel.

Despite showing respectable range and performance, the Natrium project was terminated in 2003 because of the difficulties of providing the fuel transport infrastructure and making it efficient and environmentally friendly. The 25% sodium borohydride solution is non-flammable and non-explosive and the waste sodium metaborate solution is safe to handle.