The density of hydrogen at a pressure of 700 bar is 50 kg per cubic metre. This is considerably less than that of petrol and diesel; however, the energy in hydrogen that can be usefully extracted for transportation (when consumed in a fuel cell) is about 6 times greater than for the same mass of diesel, hence the energy per unit volume in hydrogen at this pressure gives a journey range with a full tank that is acceptable.
By liquifying the gas, the density increases to 70 kg per cubic metre. Though the ability to store the fuel at normal pressure is a considerable advantage, it is very challenging to maintain the very low temperature required. Consequently, liquid hydrogen is the preferred solution only in exceptional applications.
Remarkably, by instead using metal hydrides for storage it is possible to exceed the storage density of liquid hydrogen at more-or-less normal pressure. For example, one cubic metre of lithium boron hydride power will absorb 120 kg of hydrogen (adsorption is a more correct description, though the word absorption is commonly used). The penalty in this case is that the mass increases dramatically because the weight of the powder. In addition, the charge and discharge processes can be very slow.
All metals adsorb hydrogen on their surface because electrons and protons are readily incorporated into the electronic structure of the host lattice. The amount and the pressure and temperature effects depend critically on the thermodynamics of the binding process – refer to this website. An ideal substance should adsorb a large amount of hydrogen at high pressure and low temperature, and release hydrogen at low pressure and high temperature as the thermodynamic conditions that favour adsorption become reversed. There is much research in finding materials that maximise the absorbed volume whilst reducing the temperature and pressure differences. In general, some heat is released during the charging process and absorbed during discharge, but moving heat quickly is a major problem because of the low thermal conductivity of many of the powders that otherwise perform well. Filtering may also be required to prevent the metal particles escaping with the gas (it is not known if these metal powders are harmful to health). The ideal discharge temperature would be around 80oC, matching the temperature of a typical fuel cell. There is work also on metal-organic frameworks (MOF) which is producing promising results.
There are of course similarities with the chemistry of batteries, and we need to investigate if hydride bottles can compete with lithium ion batteries in terms of mass, energy capacity and speed of charge (noting that both make huge environmental mining demands). In the simulation above we model the charge-discharge process for a typical metal hydride. The magnifying glass shows how the freely moving rough metal lattice fragments change as more and more hydrogen is trapped. The curve representing the charge level over time is the solution of a first order differential equation (similar to a capacitor charged through a resistor from a fixed voltage source). It quickly charges to 63% but takes much longer to reach 99%.
A tank of volume of internal volume 1 litre containing 4 kg of LaNi5(approximately 50% packing) is modelled. The initial charge temperature is 20oC and the pressure is 10 bar throughout the charge process. The process is exothermic, and a moderate heat transfer rate to the environment is assumed. The charge time is slow (though at 3 bar it will take 10 times as long to charge) but better than for a battery, particularly if 80% charge is acceptable. Discharge is more difficult (dehydrogenation). The metal must be persuaded to give up much of the adsorbed hydrogen, and the inclusion of catalysts can help.
With an appropriate catalyst we can assume a 60% efficient fuel cell producing electrical power 5 kW is consuming gas (0.07 g per second). The initial pressure is 4 bar. The material will provide hydrogen until the pressure drops to 2 bar when it is insufficient to operate the fuel cell (rather unrealistically modelled as a step transition). Overall, the storage system performs very well compared with a battery; at 1 kW it would last 5 times longer.
Investigations