"Does a Hydrogen Economy Add-up?"

Last week I gave a lecture at a "Cafe' Scientifique" meeting in the charming city of Salisbury - famous for its cathedral - with the title "Does a Hydrogen Economy Add-up?". This is indeed a good question, especially considering the enormous investment of funds sunk into creating a putative "Hydrogen Economy", within which the world is supposed to run most of its transportation on hydrogen "fuel". I have placed the word "fuel" in inverted commas because hydrogen is not in fact a fuel at all, but an energy carrier. This means that hydrogen cannot be simply dug out of the ground, like oil, gas and coal, but it must be artificially synthesised, in general using gas, oil or coal either as a chemical feedstock or a fuel or both. Since the second law of Thermodynamics amounts to the fact that processes whereby one form of energy is transformed into another are never 100% efficient, turning natural gas into hydrogen by "steam-reforming" is less cost effective in terns of energy than simply burning the gas (principally methane) itself, and leads overall to more CO2 being emitted, which rather flies in the face of hydrogen as the ultimate "green fuel".

On paper and said quickly, hydrogen sounds like the perfect solution to all our fossil-fuel related environmental problems. You simply mix it with oxygen (air) in a fuel cell, and this generates electricity to power a "green car". The only product is pure water, which simply drips out of the exhaust-pipe. However, as just noted, if the hydrogen has to be made from natural gas the overall process is anything but "green". It can also be made by steam-reforming coal, but as the following equations show, twice as much CO2 is emitted in the latter case per unit of hydrogen produced:

CH4 + 2H2O ---> CO2 + 4H2

C (coal) + 2H2O ---> CO2 + 2H2.

An alternative means to producing hydrogen is by water electrolysis, but this immediately begs the question of how to make the necessary electricity to do this? A recent analysis by Ulf Bossel has indicated that it would take around one 1 GW power station to produce enough hydrogen to run 20 - 30 filling stations, assuming that the hydrogen will be made in situ, in order to obviate the considerable problems that would be incurred in providing an extensive transportation network of pipelines or a fleet of tankers to move huge amounts of either compressed or liquefied hydrogen around the country from central generating facilities. In the UK there are around 3,000 filling stations ("garages" as we tend to call them over here), which implies that around 100 - 150 new 1 GW power stations would need to be built. If these were coal or gas-fired, that would add considerably to our national CO2 emissions. Indeed, the nation's average amount of electricity generated is around 40 GW (out of a maximum generating capacity of close to 69 GW), around 80% of which is created by burning gas or coal (50% of it comes from coal now). So, as a rough estimate we would increase our CO2 emissions by a factor of three or four from the electricity industry - or about a 50% increase in terms of overall energy, since electricity only accounts for about one quarter of the UK's total "energy".

One alternative is to use nuclear power, an option that is being considered seriously. However, this would mean installing 100 or so new nuclear power stations, and that is on top of replacing the 30 existing ones when they come to the end of their working life in 2024. We need to remember too, that time is fast running out for oil, which is close to the peak of production beyond which world supplies of crude oil will inexorably fall. It appears likely that any new technology will need to be on-stream within 10 - 15 years, in order to take up the slack in oil based fuel. Hence, it appears salient to ask just how quickly could these new nuclear power stations be brought on-line? For the sake of argument, would 5 per year be a reasonable estimate? If so, it would take 20 -30 years to install the lot, by which time oil (and especially the "sweet", light crude) will be long gone.

It is also probably worth noting that there is only one factory in the world (in France) that can produce the necessary new generation of reactors, and it makes 2 per year - in total! So, that would need to be stepped-up considerably. Another important question is how much nuclear fuel does the world have? Current reserves of uranium amount to about 3 million tonnes, and we get through about 75,000 tonnes of it per year (that's the whole world). Ignoring the fact that about 10,000 tonnes of it currently come from recycled nuclear warheads - although all the US, the UK and Russia are revamping their nuclear arsenals, so that supply might well fall - a simple sum suggests that there are 3,000,000/75,000 = 40 years worth. If we install more nuclear generating capacity, we simply get through the stuff faster than we would otherwise, and probably in just a few years if all the world switched-over to hydrogen. I have noted before that there is probably plenty of uranium to be had if people went looking for it hard enough, and especially in low grade deposits. For example, soil contains on average 2.7 parts per million of uranium, and there are extensive deposits of rocks with 10 - 20 ppm of uranium which could be mined with an estimated EROEI of 15 - 30. However, it takes gas and oil to mine, mill and process uranium ores and so these resources might fail first, irrespective of any theoretically comforting EROEI values.

The only feasible means to producing the electricity for hydrogen generation is from renewables, e.g. wind-power. Now, let's assume the best estimate - the lower end of the scale - of 100 GW worth of generating power; that's 100,000 MW. If we used 2 MW turbines, which we need to multiply by a suitable capacity factor, say 0.2, then that means each one delivers an average of 0.4 MW, and we would need 250,000 of them. They would probably need to be placed off-shore, around the coast of the UK mainland. Assuming that the coastal perimeter is 2,560 km, that means if the turbines were placed at the recommended separation of 0.5 km (any closer and each turbine interferes with the wind flow to the next and reduces its efficiency) , a single band around the main island would be occupied by 5,120 turbines. Hence the full band would need to be 250,000/5,120 = 49 turbines deep, stretching to a width of about 25 km. If the French did the same thing, there would be considerable overlap between the English and French wind farms especially in the English Channel, and surely a considerable obstruction to shipping! Agreed, many of them could be placed further up in the North Sea, but they would suffer considerable buffeting there from the elements, which is a great potential problem for off-shore wind farms since it imposes limitations on their durability and indeed "sustainability". The North Sea in particular is notoriously rough.

Another question is how quickly might the turbines be installed? 10 a week, say? 500 a year? Then that would take 500 years to install the lot. 100 a week? 50 years - and oil will be long-gone by then!

The other problem rests with the use of hydrogen itself. It is intended that it will be "burned" in fuel-cells, which use 50 - 100 grammes of platinum in each as the working electrode. Platinum is a very rare metal, and currently 150 tonnes of new platinum are produced each year, 40% of which is used in catalytic converters, and coincidentally is almost exactly the amount used to make jewelry. Since current demand for platinum already outstrips its supply, we would need to produce more "new" platinum. For the sake of argument, let's suppose another 150 tonnes could be made per year (and there is no reason to be sure it could) - double the present quantity. Therefore, 150 tonnes x 1000 kg x 1000 g/50 g = 3 million.

So, 3 million new fuel-cell powered vehicles might be brought on-stream per year, a figure which can be compared with around 700 million vehicles worldwide. Hence, 45 million could be available in 15 years (about 6% of the current total) when at least half of all the remaining oil will be used up. 90 million might have been produced (about 13%) in 30 years by when the conventional oil will certainly have all gone - to all intents and purposes at any rate - and we will be making it from tar sands and coal-liquefaction instead!

How would you store the hydrogen in vehicles? One way is as a compressed gas in a steel tank at 5,000 psi (pounds per square inch pressure) - that's just over 300 atmospheres - but the tank would weigh-in at 65x the weight of the hydrogen it contained. For a car containing 20 kilograms of hydrogen (which is the energy equivalent of 20 US gallons of gasoline), the tank would weigh about 1.3 tonnes, which is about twice the weight of the car itself. The tank would need to be fabricated in the form of a sphere about 5 feet in diameter - not so conveniently located in a car, unlike a normal "gas-tank" which can be made to fit unused space. This baby would be pretty evident! Tanks could be made lighter and wrapped with carbon fibre, which would get the weight down to about 200 kg, but they offer poor crash-resistance (particularly the connections, even if the tank itself survives), and so the car would become a mobile bomb.

In any case, compressing the hydrogen takes about 20% of the energy that might be recovered from it. Another possibility is to store it as liquid hydrogen; however, hydrogen is extremely difficult to liquefy, since it boils at -253 degrees C (20 degrees above absolute zero), and the process takes 40% of the energy that the hydrogen actually contains! Hydrogen could be piped around from central generating stations, but these pipes would need to be very big since hydrogen contains only around one third the energy of an equal volume of natural gas, and a lot of energy would be required to move the gas down the pipeline.

Hydrogen is a marvelous "escape artist" too. It can get through the minutest of cracks and can even diffuse through solid steel. It also makes metals brittle over time and so cracks would keep appearing in the pipes etc. requiring constant maintenance, and very likely leading to fires and explosions. If on the other hand the hydrogen were employed in liquid form, because of the risk of explosion (very cold liquid suddenly becoming a gas at room temperature leading to a huge pressure increase) the system would need to be open, so that the gas could be vented. Consequently, there would be multi-story car-parks full of cars potentially venting hydrogen into the air. Since hydrogen has the greatest explosive range of any gas (any mixture containing between about 14% and 75% of hydrogen in air will explode) regular catastrophes could be expected.

Another problem is that the oxygen is drawn into the fuel-cell in terms of air, which contains nitrogen oxides and sulphur compounds etc. which are well known to poison (inactivate) catalysts and so the fuel-cell would probably not work for too long (a few months maybe) before needing to be replaced. I could go on at far greater length about why the "hydrogen economy" in its present generation of design will never work, but probably I have made my point by now. I am not saying there will never be any use for hydrogen, for example in stand-alone applications, for "storing" energy in remote locations generated from solar-power say, hydrogen could be just the job. But on the scale required to replace petroleum based transportation at an equivalent of 20 billion barrels of oil per year (from a toal of 30 billion recovered in total) the whole idea is a complete non-starter.

Ulf Bossel also points out that three times the efficiency could be gained in storing electricity as electrons rather than as hydrogen, which would require a considerable installation in effective "battery" technology to run electric cars. This would mean that only one third of the new generating capacity is needed! [It's crazy really, the "hydrogen option" means taking electricity from a power station, using it to make hydrogen and then recombining the hydrogen with oxygen from air to make electricity again!]. There is still a pressure on resources in this option too, and probably there is not enough lithium in known reserves (or even resources) to implement the full transformation to 700 million vehicles using lithium or lithium ion batteries. You would need an awful lot of batteries. If half the world's nickel production were turned over to making nickel-cadmium batteries instead (5 million tonnes annually) we could bring around 10 million vehicles onto the roads per year, arriving at 20% of current numbers in 15 years. Either way, transportation will be curbed considerably, forcing us to live in small communities...



Related Reading.
Ulf Bossel, Proceedings of the IEEE, Vol. 94, No. 10, 2006, pages 1826 - 1837.