54 Hydrogen Storage II Chemical Methods

5.4.1 Introduction

None of the methods for storing hydrogen outlined in Section 5.3 is entirely satisfactory. Other approaches that are being developed rely on the use of chemical 'hydrogen carriers'. These could also be described as 'man-made fuels'. There are many compounds that can be manufactured to hold, for their mass, quite large quantities of hydrogen. To be useful these compounds must pass three tests:

1. It must be possible to very easily make these compounds give up their hydrogen, otherwise there is no advantage over using a reformed fuel in one of the ways already outlined in Section 5.2.

2. The manufacturing process must be simple and use little energy; in other words the energy and financial costs of putting the hydrogen into the compound must be low.

3. They must be safe to handle.

A large number of chemicals that show promise have been suggested or tried. Some of these, together with their key properties, are listed in Table 5.6. Some of them do not warrant a great deal of consideration, as they easily fail one or more of the three tests above. Hydrazine is a good example. It passes the first test very well, and it has been used in demonstration fuel cells with some success. However, hydrazine is both highly toxic and very energy-intensive to manufacture, and so fails the second and third tests.

Table 5.6 Liquids that might be used to locally store hydrogen gas for fuel cells



Percent H2

Density, kg.L-1

Vol. (1) to store 1 kg H2


Liquid H2





Cold, —252°C






Toxic, 100 ppm

Liquid methane





Cold, — 175°C
















Highly toxic

30% sodium borohydride

NaBH4 + H2O




Expensive, but works



Nevertheless, several of the compounds of Table 5.6 are being considered for practical applications, and will be described in more detail here.

5.4.2 Methanol

Methanol is the 'man-made' carrier of hydrogen that is attracting the most interest among fuel cell developers. As we saw in Section 5.2, methanol can be reformed to hydrogen by steam reforming, according to the following reaction:

The equipment is much more straightforward, though the process is not so efficient, if the partial oxidation route is used, for which the reaction is:

The former would yield 0.188 kg of hydrogen for each kg of methanol, the latter 0.125 kg of hydrogen for each kg of methanol. We have also seen in Section 5.2 that autothermal reformers use a combination of both these reactions, and this attractive alternative would provide a yield somewhere between these two figures. The key point is that whatever reformation reaction is used (equation (5.9) or (5.10) the reaction takes place at temperatures around 250°C, which is far less than those needed for the reformation of gasoline, as described in Section 5.2 (equations (5.2) or (5.6)). Also, the amount of carbon monoxide produced is far less, which means that far less chemical processing is needed to remove it. All that is needed is one of the four carbon monoxide clean-up systems outlined in Section 5.2.4.

Leading developers of methanol reforming for vehicles at present are Excellsis Fuel cell Engines (DaimlerChrysler), General Motors, Honda, International Fuel Cells, Mitsubishi, Nissan, Toyota, and Johnson Matthey. Most are using steam reforming although some organisations are also working on partial oxidation. DaimlerChrysler developed a methanol processor for the NeCar 3 experimental vehicle. This was demonstrated in September

Table 5.7 Characteristics of the methanol processor for NeCar 3 (Kalhammer et al. 1998)

Maximum unit size Power density


Specific power

CO selective oxidiser 20 L) 0.44kWe.kg-1 (reformer = 34 kg,

Energy efficiency Methanol conversion Efficiency Turn-down ratio Transient response combustor = 20 kg, CO sel. oxidiser = 40 kg) not determined 98-100%

20 to 1

1997 as the world's first methanol-fuelled fuel cell car. It was used in conjunction with a Ballard 50 kW fuel cell stack. Characteristics of the methanol processor are given in

Since the NeCar3 demonstration, DaimlerChrysler and Excellsis have been working with BASF to develop a more advanced catalytic reformer system for their vehicles. In November 2000 DaimlerChrysler launched the NeCar5 which, together with a Jeep Commander vehicle, represents state-of-the-art in methanol fuel cell vehicles. In the past six years, the fuel cell drive system has been shrunk to such an extent that it presently requires no more space than a conventional drive system. The Necar 5 therefore has the full complement of seats and interior space as a conventional gasoline-fuelled internal combustion engine car. The car is based on the A-Class Mercedes design and the methanol reformer is under the passenger compartment, as illustrated in Figure 5.6. The NeCar 5 uses a Ballard 75 kW Mk 9 stack, giving an impressive top speed of over 150km/h.

However, whatever reformer is used, full utilisation is not possible; it never is with gas mixtures containing carbon dioxide, as there must still be some hydrogen in the exit gas, as explained in Section 5.2.1. Also, in the case of steam reforming, some of the product hydrogen is needed to provide energy for the reforming reaction. If we assume that the hydrogen utilisation can be 75%, then we can obtain 0.14 kg of hydrogen for each kg of methanol. We can speculate that a 40 litre tank of methanol might be used, with a reformer of about the same size and weight as the tank. Such a system should be possible in the reasonably near term, and would give the figures of Table 5.8.

The potential figures show why methanol systems are looked on with such favour, and why they are receiving a great deal of attention for systems of power above about 10 W right through to tens of kilowatts.

We should note that ethanol, according to the figures of Table 5.6, should be just as promising as methanol as a hydrogen carrier. Its main disadvantage is that the equivalent reformation reactions of equations (5.9) and (5.10) do not proceed nearly so readily, making the reformer markedly larger, more expensive, less efficient and more difficult to control. Ethanol is also usually somewhat more expensive. All these disadvantages more than counter its very slightly higher hydrogen content.

Table 5.7.

Figure 5.6 Packaging of the NeCar 5 methanol fuel processor

Table 5.8 Speculative data for a hydrogen source, storing 40 L (32 kg) of methanol

Mass of reformer and tank 64 kg

Mass of hydrogen storeda 4.4 kg

Storage efficiency (% mass H2) 6.9%

Specific energy 5.5kWh.kg-1

Volume of tank + reformer 0.08 m3

Mass of H2 per litre 0.055 kg.L-1

a Assuming 75% conversion of available H2 to usable H2.

5.4.3 Alkali metal hydrides

An alternative to the reversible metal hydrides (Section 5.3.5) are alkali metal hydrides which react with water to release hydrogen, and produce a metal hydroxide. Bossel (1999) has described a system using calcium hydride which reacts with water to produce calcium hydroxide and release hydrogen:

It could be said that the hydrogen is being released from the water by the hydride.

Another method that is used commercially, under the trade name Powerballs, is based on sodium hydride. These are supplied in the form of polyethylene coated spheres of about 3 cm diameter. They are stored underwater, and cut in half when required to produce hydrogen. An integral unit holds the water, product sodium hydroxide, and a microprocessor controlled cutting mechanism that operates to ensure a continuous supply of hydrogen. In this case the reaction is:

This is a very simple way of producing hydrogen, and its energy density and specific energy can be as good or better than the other methods we have considered so far. Sodium is an abundant element, and so sodium hydride is not expensive. The main problems with these methods are:

• the need to dispose of a corrosive and unpleasant mixture of hydroxide and water; in theory, this could be recycled to produce fresh hydride, but the logistics of this would be difficult;

• the fact that the hydroxide tends to attract and bind water molecules, which means that the volumes of water required tend to be considerably greater than equations (5.11) and (5.12) would imply;

• the energy required to manufacture and transport the hydride is greater than that released in the fuel cell.

A further point is that the method does not stand very good comparison with metal air batteries. If the user is prepared to use quantities of water, and is prepared to dispose of water/metal hydroxide mixtures, then systems such as the aluminium/air or magnesium/air battery are preferable. With a salt water electrolyte, an aluminium/air battery can operate at 0.8 V at quite a high current density, producing three electrons for each aluminium atom. The electrode system is much cheaper and simpler than a fuel cell.

Nevertheless, the method compares quite well with the other systems in several respects. The figures in Table 5.9 are calculated for a self-contained system capable of producing 1 kg of hydrogen, using the sodium hydride system. The equipment for containing the water and gas, and the cutting and control mechanism is assumed to weigh 5 kg. There is three times as much water as equation (5.12) would imply is needed.

The storage efficiency compares well with other systems. This method may well have some niche applications where the disposal of the hydroxide is not a problem, though these are liable to be limited.

Table 5.9 Figures for a self-contained system producing 1 kg of hydrogen using water and sodium hydride

Mass of container and all materials

45 kg

Mass of hydrogen stored

1.0 kg

Storage efficiency (% mass H2)


Specific energy


Volume of tank (approx.)

50 L

Mass of H2 per litre

0.020 kg.L-1

5.4.4 Sodium borohydride

A good deal of interest has recently been shown in the use of sodium tetrahydridoborate, or sodium borohydride as it is usually called, as a chemical hydrogen carrier. This reacts with water to form hydrogen according to the reaction:

NaBH4 + 2H2O-> 4H2 + NaBO2 (AH = —218kJ.mol—1) (5.13)

This reaction does not normally proceed spontaneously, and solutions of NaBH4 in water are quite stable. Some form of catalyst is usually needed. The result is one of the great advantages of this system: it is highly controllable. Millennium Cell Corp. in the USA has been actively promoting this system and has built demonstration vehicles running on both fuel cells and internal combustion engines using hydrogen made in this way. Companies in Europe, notably NovArs GmbH in Germany (Koschany 2001), have also made smaller demonstrators. Notable features of equation (5.13) are:

• it is exothermic, at the rate of 54.5 kJ per mole of hydrogen;

• hydrogen is the only gas produced, it is not diluted with carbon dioxide;

• if the system is warm, then water vapour will be mixed with the hydrogen, which is highly desirable for PEM fuel cell systems.

Although rather overlooked in recent years, NaBH4 has been known as a viable hydrogen generator since 1943. The compound was discovered by the Nobel laureate Herbert C. Brown, and the story is full of interest and charm, but is well told by Prof. Brown himself (Brown 1992). Suffice to say that shortly before the end of the 1939-1945 war plans were well advanced to bulk-manufacture the compound for use in hydrogen generators by the US Army Signals Corps, when peace rendered this unnecessary. However, in the following years many other uses of sodium borohydride, notably in the paper processing industries, were discovered, and it is produced at the rate of about 5000 tonnes per year,3 mostly using Brown's method, by Morton International (merged with Rohm and Haas in 1999).

If mixed with a suitable catalyst, NaBH can be used in solid form, and water added to make hydrogen. The disadvantage of this method is that the material to be transported is a flammable solid, which spontaneously gives off H2 gas if it comes into contact with water. This is obviously a safety hazard. It is possible to purchase sodium borohydride mixed with 7% cobalt chloride for this purpose. However, this is not the most practical way to use the compound.

Current work centres on the use of solutions. This has several advantages. Firstly the hydrogen source becomes a single liquid, no separate water supply is needed. Secondly this liquid is not flammable, and only mildly corrosive, unlike the solid form. The hydrogen releasing reaction of equation (5.13) is made to happen by bringing the solution into contact with a suitable catalyst. Removing the catalyst stops the reaction. The gas generation is thus very easily controlled, a major advantage in fuel cell applications.

The maximum practical solution strength used is about 30%. Higher concentrations are possible, but take too long to prepare, and are subject to loss of solid at lower temperatures.

3 Kirk-Othmer Encyclopedia of Chemical Technology, Wiley.

The solution is made alkaline by the addition of about 3% sodium hydroxide, otherwise the hydrogen evolution occurs spontaneously. The 30% solution is quite thick, and so weaker solutions are sometimes used, even though their effectiveness as a hydrogen carrier is worse. One litre of a 30% solution will give 67 g of hydrogen, which equates to about 800 NL. This is a very good volumetric storage efficiency.

Generators using these solutions can take several forms. The principle is that to generate hydrogen the solution is brought into contact with a suitable catalyst, and that generation ceases when the solution is removed from the catalyst. Suitable catalysts include platinum and ruthenium, but other less expensive materials are effective, including iron oxide. Fuel cell electrodes make very good reactors for this type of generator.

A practical sodium borohydride system is shown in Figure 5.7. The solution is pumped over the reactor, releasing hydrogen. The motor driving the pump is turned on and off by a simple controller that senses the pressure of the hydrogen, and which turns it on when more is required. The solution is forced through the reactor, and so fresh solution is continually brought in contact with the catalyst. The rate of production is simply controlled by the duty cycle of the pump. The reaction takes place at room temperature, and the whole system is extremely simple when compared to any of the other generators that have been outlined in this section.

Figure 5.7 Example reactor for releasing hydrogen from a solution of sodium borohydride in water, stabilised with sodium hydroxide. The rate of production of hydrogen is controlled by varying the rate at which the solution is pumped over the reactor

Hydrogen gas to fuel cell



Figure 5.7 Example reactor for releasing hydrogen from a solution of sodium borohydride in water, stabilised with sodium hydroxide. The rate of production of hydrogen is controlled by varying the rate at which the solution is pumped over the reactor

However, when the solution is weak, the reaction rate will be much slower, and the system will behave differently. It is likely that it will not be practical to obtain highly efficient solution usage. Also, the solution cannot be renewed at the user's convenience; it must be completely replaced when the NaBH4 has been used up, and at no other time.

Another method that is used is that of the Hydrogen on Demand® system promoted by Millennium Cell. This uses a single pass catalyst, rather than the re-circulation system of Figure 5.7. A major advantage of this is that the tank of fresh solution can be topped up at any time. A disadvantage is that two tanks are needed, the second one for the spent solution that has passed over the catalyst.

In terms of the general figure of merit, 'volume required to produce 1 kg of hydrogen', the 30% NaBH4 solution is the worst of the liquid carriers shown in Table 5.6. However, it is competitive, and is only very slightly worse than pure liquid hydrogen. However, it has many advantages over the other technologies.

• it is arguably the safest of all the liquids to transport;

• apart from cryogenic hydrogen it is the only liquid that gives pure hydrogen as the product. This is very important, as it means this the only one where the product gas can be 100% utilised within the fuel cell;

• the reactor needed to release the hydrogen requires no energy, and can operate at ambient temperature and pressure;

• the rate of production of hydrogen can be simply controlled;

• the reactor needed to promote the hydrogen production reaction is very simple, far simpler than that needed for any of the other liquids;

• if desired, the product hydrogen gas can contain large quantities of water vapour, which is highly desirable for PEM fuel cells.

In order to compare a complete system, and produce comparative figures for gravimetric and volumetric storage efficiency, we need to speculate what a complete hydrogen generation system would be like. Systems have been built where the mass of the unit is about the same as the mass of the solution stored, and about twice the volume of solution held. So, a system that holds 1 litre of solution has a volume of about 2 litres, and weighs about 2 kg. Such a system would yield the figures shown in Table 5.10.

These figures are very competitive with all other systems. So what are the disadvantages? There are three main problems, the second two being related. The first is the problem of disposing of the borate solution. This is not unduly difficult, as it is not a

Table 5.10 Speculative data for a hydrogen source, storing 1.01 of 30% NaBH4, 3% NaOH and 67% H2O solution

Mass of reformer, tank, and solution 2.0 kg

Mass of hydrogen stored 0.067 kg

Storage efficiency (% mass H2) 3.35%

Specific energy 1.34kWh.kg-1

Volume of system (approx.) 2.0 L

Mass of H2 per litre 0.036kg.L-1

hazardous substance. However, the other disadvantages are far more severe. The first is the cost. Sodium borohydride is an expensive compound. By simple calculation and reference to catalogues it can be shown that the cost of producing hydrogen this way is about $630 per kilogram.4 This is over 100 times more expensive than using an electrolyser driven by grid-supplied electricity (Larminie 2002). At this sort of cost the system is not at all viable.

Linked to this problem of cost is the energy required to manufacture sodium boro-hydride. Using current methods this far exceeds the requirements of compounds such as methanol. Currently sodium borohydride is made from borax (NaO.2B2O3.10H2O), a naturally occurring mineral with many uses that is mined in large quantities, and methanol. The aim is that the sodium metaborate produced by the hydrolysis reaction of equation (5.13) is recycled back to sodium borohydride. Table 5.11 shows the molar enthalpies of formation of the key compounds. It can be seen that such recycling will be a formidable challenge, requiring at least 788kJ.mol-1. However, the prize is four moles of hydrogen, so that is at least 197kJ.mol-1, which is not quite so daunting. Nevertheless, there are many problems to be overcome before such recycling is viable.

The companies, such as Millennium Cell Inc., who are hoping to commercialise this process are working hard on this problem of production cost, finance and energy. If they succeed there will be a useful hydrogen carrier, but until the costs come down by a factor of at least 10 then this method will only be suitable for special niche applications.

5.4.5 Ammonia

Ammonia is a colourless gas with a pungent choking smell that is easy to recognise. It is highly toxic. The molecular formula is NH3, which immediately indicates its potential as a hydrogen carrier. It has many uses in the chemical industry, the most important being in the manufacture of fertiliser, which accounts for about 80% of the use of ammonia. It is also used in the manufacture of explosives. Ammonia is produced in huge quantities. The annual production is estimated at about 100 million tonnes, of which a little over 16 million is produced in the USA.5

Ammonia liquefies at — 33°C, not an unduly low temperature, and can be kept in liquid form at normal temperature under its own vapour pressure of about 8 bar, not an unduly high pressure. Bulk ammonia is normally transported and stored in this form. However, it also readily dissolves in water - in fact it is the most water-soluble of all gases. The solution (ammonium hydroxide) is strongly alkaline, and is sometimes known as ammonia water or ammonia liquor. Some workers have built hydrogen generators using this as the

Table 5.11 Key thermodynamic data for sodium borohydride and borate



Molar enthalpy of formation

— 189kJ.mol-1


4 2003 prices.

5 Information provided by the Lousiana Ammonia Association, www.lammonia.com.

4 2003 prices.

5 Information provided by the Lousiana Ammonia Association, www.lammonia.com.

form of ammonia supplied, but this negates the main advantage of ammonia, which is its high hydrogen density, as well as adding complexity to the process.

Liquid ammonia is one of the most compact ways of storing hydrogen. In terms of volume needed to store 1 kg of hydrogen, it is better than almost all competing materials; see Table 5.6. Counter-intuitively, it is approximately 1.7 times as effective as liquid hydrogen. (This is because, even in liquid form, hydrogen molecules are very widely spaced, and LH2 has a very low density.)

Table 5.6 shows ammonia to be the best liquid carrier, in terms of space to store 1 kg of hydrogen, apart from hydrazine, which is so toxic and carcinogenic that it is definitely not a candidate for regular use. However, the margin between the leading candidates is not very large. The figures ignore the large size of container that would be needed, especially in the case of ammonia, liquid methane and LH2, though not for the key rival compound methanol.

Two other features of ammonia lie behind the interest in using it as a hydrogen carrier. The first is that large stockpiles are usually available, due to the seasonal nature of fertiliser use. The second is that ammonia prices are sometimes somewhat depressed due to an excess of supply over demand. However, when the details of the manufacture of ammonia, and its conversion back to hydrogen are considered, it becomes much less attractive.

Using ammonia as a hydrogen carrier involves the manufacture of the compound from natural gas and atmospheric nitrogen, the compression of the product gas into liquid form, and then, at the point of use, the dissociation of the ammonia back into nitrogen and hydrogen.

The production of ammonia involves the steam reformation of methane (natural gas), as outlined in Section 5.2. The reaction has to take place at high temperature, and the resulting hydrogen has to be compressed to very high pressure (typically 100 bar) to react with nitrogen in the Haber process. According to the Lousiana Ammonia Producers Association, who make about 40% of the ammonia produced in the USA, the efficiency of this process is about 60%. By this they mean that 60% of the gas used goes to provide hydrogen, and 40% is used to provide energy for the process. This must be considered a 'best case' figure, since there will no doubt be considerable use of electrical energy to drive pumps and compressors that is not considered here. The process is inherently very similar to methanol production; hydrogen is made from fuel, and is them reacted with another gas. In this case it is nitrogen instead of carbon dioxide. The process efficiencies and costs are probably similar.

The recovery of hydrogen from ammonia involves the simple dissociation reaction:

For this reaction to occur at a useful rate the ammonia has to be heated to between 600 and 800° C, and passed over a catalyst. Higher temperatures of about 900° C are needed if the output from the converter is to have remnant ammonia levels down to the ppm level. On the other hand the catalysts need not be expensive: iron, copper, cobalt and nickel are among many materials that work well. Systems doing this have been described in the literature (Kaye et al. 1997, Faleschini et al. 2000). The later paper has a good review of the catalysts that can be used.

The reaction is endothermic, as shown. However, this is not the only energy input required. The liquid ammonia absorbs large amounts of energy as it vaporises into a gas, which is why it is still quite extensively used as a refrigerant.

Once a gas at normal temperature, it then has to be heated, because the dissociation reaction only takes place satisfactorily at temperatures of around 800 to 900°C. For simplicity we will assume an 800°C temperature rise. The molar specific heat of ammonia is 36.4 J.mol-1.kg-1. So:

This process results in the production of 1.5 moles of hydrogen, for which the molar enthalpy of formation (HHV) is —285.84 kJ.mol-1. The best possible efficiency of this stage of the process is thus:

This should be considered an upper limit of efficiency, as we have not considered the fact that the reformation process will involve heat losses to the surroundings. However, systems should be able to get quite close to this figure, since there is scope for using heat recovery, as the product gases would need to be cooled to about 80° C before entering the fuel cell. The vaporisation might also take place below ambient temperature, allowing some heat to be taken from the surroundings.

The corrosive nature of ammonia and ammonium hydroxide is another major problem. Water is bound to be present in a fuel cell. Any traces of ammonia left in the hydrogen and nitrogen product gas stream will dissolve in this water, and thus form an alkali (ammonium hydroxide) inside the cell. In small quantities, in an alkaline electrolyte fuel cell, this is tolerable. However, in the PEM fuel it would be fatal. This point is admitted by some proponents of ammonia, and is used by them as an advantage for alkaline fuel cells (Kordesch et al. 1999). Hydrogen from other hydrogen carriers such as methanol and methane also contains poisons, notably carbon monoxide. However, these can be removed, and do not permanently harm the cell, they just temporarily degrade performance. Ammonia on the other hand, will do permanent damage, and this damage will steadily get worse and worse.

Ammonia as a hydrogen carrier can easily be compared to methanol. If it were, the following points would be made:

• the production methods and costs are similar;

• the product hydrogen per litre of carrier is slightly better;

• ammonia is far harder to store, handle and transport;

• ammonia is more dangerous and toxic;

• the process of extracting the hydrogen is more complex;

• the reformer operates at very high temperatures, making integration into small fuel cell systems much more difficult than for methanol;

• the product gas is difficult to use with any type of fuel cell other than alkaline.

The conclusion must be that the use of ammonia as a hydrogen carrier is going to be confined to only the most unusual circumstances.

5.4.6 Storage methods compared

We have looked at a range of hydrogen storage methods. In Section 5.3 we looked at fairly simply 'hydrogen in, hydrogen out' systems. In Section 5.4 we looked at some more complex systems involving the use of hydrogen-rich chemicals that can be used as carriers.

None of the methods is without major problems. Table 5.12 compares the systems that are currently feasible in relation to gravimetric and volumetric effectiveness. Together with the summary comments, this should enable the designer to choose the least difficult alternative. It is worth noting that the method with the worst figures (storage in high pressure cylinders) is actually the most widely used. This is because it is so simple and straightforward. The figures also show why methanol is such a promising candidate for the future.

Table 5.12 Data for comparing methods of storing hydrogen fuel. The figures include the associated equipment, e.g. tanks for liquid hydrogen, or reformers for methanol


Gravimetric storage efficiency, % mass H2

Volumetric mass (in kg) of H2 per litre


High pressure in cylinders



'Cheap and cheerful', widely


Metal hydride



Suitable for small systems

Cryogenic liquid



Widely used for bulk storage




Low cost chemical, potentially

useful in a wide range of


Sodium hydride pellets



Problem of disposing of spent


NaBH4 solution in water



Very expensive to run

Hybrid Cars The Whole Truth Revealed

Hybrid Cars The Whole Truth Revealed

Hybrid Cars! Man! Is that a HOT topic right now! There are some good reasons why hybrids are so hot. If you’ve pulled your present car or SUV or truck up next to a gas pumpand inserted the nozzle, you know exactly what I mean! I written this book to give you some basic information on some things<br />you may have been wondering about.

Get My Free Ebook

Post a comment