Reaction of Aluminum with Water to Produce Hydrogen


1 Reaction of Aluminum with Water to Produce Hydrogen A Study of Issues Related to the Use of Aluminum for On-Board Vehicular Hydrogen Storage U.S. Department of Energy Version 1.0 - 2008 Page 1 of 26


3 Reaction of Aluminum with Water to Produce Hydrogen 1 2 John Petrovic and George Thomas Consultants to the DOE Hydrogen Program 1 Los Alamos National Laboratory (retired) 2 Sandia National Laboratories (retired) Executive Summary: The purpose of this White Paper is to describe and evaluate the potential of aluminum-water reactions for the production of hydrogen for on-board hydrogen-powered vehicle applications. Although the concept of reacting aluminum metal with water to produce hydrogen is not new, at such aluminum-water reactions might be there have been a number of recent claims th employed to power fuel cell devices for portable applications such as emergency generators and for possible use as the hydrogen source for fuel laptop computers, and might even be considered cell-powered vehicles. tion between aluminum metal and water to form In the vicinity of room temperature, the reac + 3H O = 2Al(OH) . The aluminum hydroxide and hydrogen is the following: 2Al + 6H 3 2 2 tion is 3.7 wt.% and the volumetric hydrogen gravimetric hydrogen capacity from this reac /L. capacity is 46 g H 2 rable, it does not proceed due to the presence Although this reaction is thermodynamically favo oxide which forms on the surface of aluminum of a coherent and adherent layer of aluminum particles which prevents water from coming into direct contact with the aluminum metal. The uminum with water near room temperature is key to inducing and maintaining the reaction of al nt aluminum oxide layer. the continual removal and/or disruption of this coherent/adhere A number of reaction-promoting approaches have been investigated for the aluminum-water reaction. These include additions of hydroxide promoters such as NaOH, oxide promoters such dditions act to disrupt the aluminum oxide , and salt promoters such as NaCl. These a O as Al 3 2 layer on the aluminum metal. In addition, the reaction of water with molten aluminum alloys such as aluminum-lithium and aluminum-gallium has been studied. In this case, the molten nature of the alloy prevents the development of a coherent and a dherent aluminum oxide layer. proven commercially viable to date. However, none of these approaches have tion to provide onboard hydrogen for hydrogen- The concept of using the aluminum-water reac powered vehicles presents a number of difficulties. First, storage systems using this approach will not be able to meet the 2010 DOE system targets of 6 wt.% hydrogen and 45 grams hydrogen per liter. Second, based on published alu minum-water reaction rate kinetics, it appears Version 1.0 - 2008 Page 3 of 26

4 difficult for this approach to meet the DOE minimum hydrogen flow rate target for fuel-cell powered vehicles. Finally, the cost of producing hydrogen by this approach is dictated by the cost of aluminum metal. The November 2007 commodity price for aluminum is $2.36 per kg. er hydrogen generation approach would cost At this price, hydrogen from an aluminum-wat approximately $21 per kg H . Even assuming high volume production, the DOE target range for 2 would not be met. Additionally, the supply of aluminum hydrogen cost of $2-3 per kg H 2 pplications may be problematic. required for mass market vehicle a r on-board vehicular hydrogen While aluminum-water reaction systems cannot meet the targets fo storage, the use of aluminum as a water splitting agent for generating hydrogen might have utility for non-vehicular applications. Version 1.0 - 2008 Page 4 of 26

5 Introduction: The concept of producing hydrogen by the reactio n of certain metals with water has intrigued in 1972 (1) described an approach using an researchers for many years. A paper by Smith ich was based on even earlier wo amalgamated aluminum surface wh rk by others cited in the hydrogen could be produced via article. In a 1976 U.S. Patent, Gu tbier and Hohne indicated that the reaction of magnesium-aluminum mixtures with sea water (2). flected in a number of publications and patents More recently, there has been renewed activity, re num-based metals and reactions between alumi directed at the production of hydrogen through water. All of the aluminum-based approaches propose methods to circumvent the protective layer of aluminum oxide, thus allowing the reaction with water to proceed. The hydrogen produced via such aluminum-water reactions might be employed to power fuel ergency generators and laptop computers. There cell devices for portable applications such as em is also the suggestion that alu ed for hydrogen storage on fuel minum-water reactions might be us cell-powered vehicles. The purpose of this White Paper is to describe a nd evaluate the potential ly aimed at on-board of aluminum-water reactions for the production of hydrogen, primari first discuss the aluminum-water reaction and hydrogen-powered vehicle applications. We tion. Then we consider the describe the various methods devised to maintain the reac performance of such a system relative to the requirements for on-board vehicular hydrogen storage. Since any hydrogen generation concept, whether it is for on-board storage or not, will need to be regenerable (that is, the reaction produc ts will need to be returned to the initial form of aluminum), the energy and cost requirements for these processes will be discussed. Background: The following are possible reacti ons of aluminum with water: 2Al + 6H (1) + 3H O = 2Al(OH) 3 2 2 O = 2AlO(OH) + 3H (2) 2Al + 4H 2 2 2Al + 3H O = Al O + 3H (3) 2 2 3 2 The first reaction forms the aluminum hydroxide bayerite (Al(OH) ) and hydrogen, the second 3 reaction forms the aluminum hydroxide boehmite (AlO(OH)) and hydrogen, and the third All these reactions are thermodynamically reaction forms aluminum oxide and hydrogen. o C). All are also favorable from room temperature past the melting point of aluminum (660 o highly exothermic. From room temperature to 280 is the most stable product, while C, Al(OH) 3 o o C, Al O is the most stable product (3). from 280-480 C, AlO(OH) is most stable. Above 480 2 3 of the thermodynamics of aluminum-water See the Appendix I for a more thorough review reactions. Version 1.0 - 2008 Page 5 of 26

6 Table I: Thermodynamics of the aluminum-water reaction. [The thermodynamic parameters as a function of temperature were calculated usin g HSC Thermodynamic Software, Version 4.1.] (g) + H 2/3Al + 2H O = 2/3Al(OH) 2 2 3 T G Δ S Δ H Δ o C kJ/mol H2 kJ/mol H2 J/K 0 -277 26.2 -284 100 -284 3.29 -285 200 -291 -12.1 -285 indicate that aluminum should spontaneously The reaction thermodynamics shown in Table I react with water. However, in practice a pie ce of aluminum dropped into water will not react under room temperature conditions, or even with boiling water. This is because the aluminum O , on its surface and this alumina has a thin coherent, adhering layer of aluminum oxide, Al 2 3 layer prevents the reaction. So the key to i nducing and maintaining the reaction of aluminum with water at room temperature is the continual removal and/or disruption of the hydrated alumina layer. Organizations that are currently involved with the development of alu minum-water systems for the production of hydrogen are listed in Appendix II. Reaction-Promoting Approaches: Hydroxide Promoters: A number of U.S. patents descri be the use of hydroxides, primarily sodium hydroxide (NaOH), to promote aluminum-water reacti ons (4-8). There are two pertinent technical references on this subject. The first is by Belitskus (9). Here, experiments were described in which aluminum specimens including a cylindrical block, uncompacted powders, and pellets of various densities were reacted with aqueous solutions of sodium hydroxide at different concentrations to produce hydrogen gas at temperatures near room temper ature. The formation of sodium aluminate was observed, as well as the regeneration of sodium hydroxide via the precipitation of aluminum hydroxide. Stockburger (10) described hydrogen generators in which aluminum was reacted with an aqueous solution of 5.75 M sodium hydroxide. The types of reactions found to occur between aluminum, sodium hydroxide, and water are shown below: + 3H Al O (4) O = Na 2Al + 2NaOH + 2H 2 2 2 4 2 (5) O O + 3H + xH 2Al + 6NaOH + xH Al O = Na 6 2 2 6 2 2 Version 1.0 - 2008 Page 6 of 26

7 2Al + 2NaOH + 6H + 3H (6) O = 2NaAl(OH) 2 2 4 2NaAl(OH) (7) = 2NaOH + 2Al(OH) 4 3 One of the problems with the use of aqueous NaOH solutions is the corrosive nature of the pment. An example of a hydrogen-producing liquid, which can lead to corrosion of system equi reactor based on the NaOH approach fro m a 2006 patent is shown in Figure 1. Figure 1: Hydrogen-producing reactor based on the NaOH approach (8). Here, the aluminum powder is fed into the reac tion chamber, where it reacts with the sodium hydroxide solution near room temperature, with the production of hydrogen gas and the e reactor. It is unclear whether this system formation of reaction byproducts at the bottom of th has been commercially developed or utilized to date. Oxide Promoters: O ) powders are reactive aluminum and aluminum oxide (Al It has been shown that mixtures of 2 3 o O C. These Al-Al with water in the pH range of 4-9 (11-13) and at temperatures of 10-90 3 2 ball-milled together in order to produce hydrogen reactions. powder mixtures must be heavily Hydrogen can be evolved at room temperature l water, although the using essentially neutra temperature. An example of this effect is hydrogen evolution rate increases with increasing shown in Figure 2. Version 1.0 - 2008 Page 7 of 26

8 r aluminum-aluminum oxide powder mixtures (12). Figure 2: Hydrogen production versus pH fo ), boehmite (AlO(OH)), gamma The aluminum oxide may be in the form of bayerite (Al(OH) 3 -Al γ alumina ( ). Alpha alumina powder was reported to give the O -Al α ), or alpha alumina ( O 3 3 2 2 maximum hydrogen evolution. It has been speculated that the milling of aluminum and aluminum oxide powders together helps rent and coherent oxide layers present on the aluminum powder, to mechanically disrupt the adhe and that this is the reason for the enhanced hy drogen generation in pH neutral water (11-13). However, recent research has suggested that the enhancing effect of aluminum oxide on the reactivity of aluminum with water may also be mechanochemical in nature (14). Aluminum powders that were reacted with fine boehmite powders at elevated temperatures produced a layer of fine-grained, mechanically weak gamma alumina on the surfaces of the aluminum powders. As shown in Figure 3, this gamma alumina could then react with water to produce boehmite Version 1.0 - 2008 Page 8 of 26

9 Figure 3: Reaction mechanism of water with an alumina-coated aluminum particle (14). which would grow in thickness until the boehmite reacted with the underlying aluminum to interface. Under suitable conditions, these produce hydrogen bubbles at the aluminum-boehmite layer, leading to activation of the aluminum for reaction O bubbles would then rupture the Al 2 3 with water. Salt Promoters: Water soluble inorganic salts can produce localiz ed pitting and rupture of the alumina layer on aluminum particles (15,1 6). Such effects have been empl oyed to promote aluminum-water hydrogen production reactions (17,18). The salts NaCl and KCl were found to be the most effective. Powders of these salts were ball-milled together with aluminum powder in a 1:1 Version 1.0 - 2008 Page 9 of 26

10 weight ratio. The resultant hydrogen generation when the ball-milled aluminum-salt mixture o was exposed to water at 55 C is shown in Figure 4. Figure 4: Reaction of aluminum-salt mixtures with water (17). Reaction of the aluminum-salt mixtures with water was observed to be significantly temperature dependent, as shown in Figure 5. Version 1.0 - 2008 Page 10 of 26

11 Figure 5: Effect of temperature on al uminum-salt reactions with water (17). It has also been reported that if the salt is wa shed out of the aluminum powder at a lower water temperature, the aluminum powder will still react to produce hydrogen (18). In this case, the salt o C. Then the aluminum was washed out of the aluminum powder with cold tap water at 12 o C. This might suggest that milling the aluminum powder alone was immersed in water at 55 powder with the salt powder produc es some disruption of the alumina layers on the aluminum particles, thus activating the aluminum for reaction with water. Combined Oxide and Salt Promoters: An international patent appli cation (2006) has indicated that oxide and salt additions may be o C (19). The preferred oxide is combined to promote the reaction of aluminum with water at 20 calcium oxide in the range of 0.5 – 4 % by weight. The preferred salt is sodium chloride in a weight ratio of 1:1 with aluminum. Shown in Fi gure 6 are the hydrogen generation levels as a function of oxide addition. The hydrogen release kinetics indicated are relatively slow, with a maximum rate of only about 0.001 g H /sec under the specific generation conditions (e.g. 2 amount of solution and catalyst) employed for the results in Figure 6. Version 1.0 - 2008 Page 11 of 26

12 Figure 6: Hydrogen generation for different oxide level additions (19). Aluminum Pretreatment: A recent patent application has indicated that aluminum powders can be activated by ball-milling them in water, followed by a rapid heating and cooling thermal shock treatment (20). The improvement in hydrogen generation from the aluminum-water reaction is shown in Figure 7. Figure 7: Effect of powder ac tivation on hydrogen generation (20). Version 1.0 - 2008 Page 12 of 26

13 The hydrogen release rate from the activated alum inum powder is low, only of the order of 8 x -7 g H 10 /sec/g of Al. 2 Molten Aluminum Alloys: In one approach, molten ys has been proposed. The reaction of water with molten aluminum allo water through a nozzle in a reaction chamber in aluminum-lithium alloys are sprayed with liquid this reaction chamber is shown in Figure 8. order to generate hydrogen (21,22). A drawing of Figure 8: Reactor for hydrogen generation from molten aluminum-lithium alloys (21). s also been suggested (23,24). The rationale The reaction of molten aluminum-gallium alloys ha here is that molten aluminum-gallium alloys will not possess a coherent and adherent oxide layer. Because no oxide layer is present on the liquid alloy, it will react readily with liquid water. Additionally, the aluminum-gallium alloys will have a low melting point range, as shown in Figure 9. Version 1.0 - 2008 Page 13 of 26

14 showing hydrogen-producing alloys (24). Figure 9: Aluminum-gallium phase diagram d be essentially an inert species that could In the hydrogen production process, the gallium woul be reused. As may be seen in the aluminum-gallium phase diagram in Figure 9, the temperature at which the aluminum-gallium alloy is liquid increases with increasing aluminum content. o C only for gallium-rich mixtures where the aluminum However, the alloy is liquid below 100 content is only a few percent by weight. More recent work since the patent application [24] weight alloy could be formed by slow cooling indicates that solid particles of Al 80%:Ga 20% by a Ga –rich mixture (Al 28%:Ga 72%) from the me lt. This aluminum rich alloy was found to react with water, but no information on reaction rates is available. relative to on-board system properties: Properties of the aluminum-water reactions Hydrogen Capacities: orage materials is their available hydrogen One of the important parameters for on-board st targets and capacities in terms of weight and volume. The DOE capacity targets are system include reactants, products lant to contain them (including the tank), but , and all the balance-of-p here we consider the material capacities – a primary consideration to determine the feasibility of achieving system targets. The gravimetric capacity is usually quoted in wt.% of hydrogen and must include all reactants and/or reaction pr oducts. Since the formation of bayerite is o C, we will only consider the reaction shown in Eq. (1): thermodynamically favored below 280 Version 1.0 - 2008 Page 14 of 26

15 O = 2Al(OH) 2Al + 6H + 3H 3 2 2 uminum alone without including the water, 11 The hydrogen gravimetric yield based on the al wt.%, is not a relevant factor. Even if the water could be supplied by, for example, the fuel cell exhaust stream (an unlikely possibility (25)), the hydroxide reaction product remains on-board the vehicle and must be included in the weight. Thus, the theoretically best gravimetric density of hydrogen for this material reaction is 3.7 wt.%. In any practical system, most or all of the water will need to be stored as well and any alloying additions (such as gallium) or other reaction promoters (such as NaOH, NaCl, or Al O ) will add more weight and further reduce the 2 3 hydrogen weight density. Estimated values for some of the promoter schemes described earlier are given in Table II. The material volumetric densities of hydrogen for the reactions can be estimated from a weight fraction. Thus, the left side of Eq. knowledge of the material density and the hydrogen 1.26 g /mL and a weight density of 3.7 wt.% (1) corresponds to an overall material density of hydrogen. The volume density of hydrogen (for the materials only) is then ~ 46 g H /L. 2 Volumetric capacities for the different schemes are also included in Table II. Table II Hydrogen Weight and Volume Capacities for Different Promoter Schemes Gravimetric Material capacity Volumetric capacity /L 46 g H 3.7 wt.% H Pure aluminum 2 2 2.5 wt.% H Hydroxide promoter /L 36 g H 2 2 2.5 wt.% H Oxide promoter /L 40.6 g H 2 2 NaCl salt promoter 2.8 wt.% H 39 g H /L 2 2 3.0 wt.% H Ga 20/Al 80 alloy 37 g H /L 2 2 Kinetic properties (hydrogen delivery rates): The data shown for the various promoter scheme s indicate that the water splitting reaction rate generally has a strong temperature dependence. Also, since the reaction occurs at the metal system is related to the surface area of the surface, the hydrogen generation rate for a given aluminum particles in contact with the water. The experiment al hydrogen production rates are /g of Al per unit time at a given ntity of material, that is, g H usually reported in terms of the qua 2 temperature, but the material para meters (e.g., particle size or surface area) are not always given, making a comparison between some of the promoter schemes less meaningful. There are a few cases that can be considered, how ever. The reaction rate shown in Figure 7 for -7 /sec/g of Al. Deng, et. al. (14) measured g H the pretreated aluminum is quite low, at ~ 8 x 10 2 o -6 g H /sec/g of Al at 50 a hydrogen production rate of ~ 4 x 10 C, using Al O as a promoter. 2 2 3 -4 g H /sec/g zynski (17) measured a rate of ~ 2 x 10 Using KCl and NaCl salts as promoters, Troc 2 o of Al at 55 C. No reaction rate data is available at this time for the Al-rich Ga alloy. Version 1.0 - 2008 Page 15 of 26

16 A delivery rate of 1.6 g H r fuel cell. At a kinetic reaction /sec is required for an 80 kW vehicula 2 -4 rate of 2 x 10 g H /sec/g of Al for the aluminum-water reaction (the maximum value in the 2 literature), one must react 8000 g of aluminum with the necessary amount of water in order to achieve the hydrogen delivery rate of 1.6 g H /sec needed to supply the 80 kW fuel cell. This 2 makes it clear that the kinetics of least those presently documented aluminum-water reactions (at in the literature) are likely to be problematic for vehicular applications. These kinetic requirements are discussed further in the next section where system issues are considered. System considerations In addition to the material properties discussed above, th ere remain a number of issues related to vehicle to generate hydrogen for a fuel cell actual use of the aluminum/water reaction onboard a or ICE. These include: (1) containment of fresh reactants and the separation and storage of ms to promote and control the generation of reaction byproducts; (2) components or subsyste hydrogen in response to a highly transient fuel demand profile; (3) th ermal management to control temperatures and energy, (4) loading of fresh material and unloading of reaction byproducts. An important advantage of the proposed process is that containment of the materials is relatively straight forward, although the caustic nature of NaOH may require special consideration. for the liquids and solids would most likely not Overall, however, it would appear that containers be prohibitively heavy or expensive. The packing density of the solid metal particles would result in a volumetric density which typically is about 60% of the material density. Thus, the volume densities shown in Table I would be slightly lower (the density of the water remains the same regardless of the aluminum density). Fo r example, the volumetri c density for the pure aluminum case would be about 42 g H /L rather than 46 g H / L. 2 2 The second issue, however, has a greater impact on the system parameters. In order to control the reaction (which runs to completion once star ted), either some metal or some water would need to be transported to a reactor. This is apparent in the patent proposals described in an h include schemes either for liquid or particle flow into a earlier section of this paper whic control the rate of water flow to a container reaction chamber. The simplest case might be to which holds all of the metal reactant. Even then, the transient nature of the fuel demand on a vehicle would necessitate a buffer for retaining some quantity of hydrogen. The generation of 18 liters (STP) of hydrogen in one ) would require a container sufficiently large to second (the volume equivalent of 1.6 gH 2 tisfy safety concerns. This fuel supply accommodate the gas and a pressure rating to sa requirement, furthermore, would typically be required for more than just one second. Thus, an engineering approach might be a continuous wate r stream to maintain a roughly steady state hydrogen generation rate that is supplied to a bu ffer tank sufficiently large to satisfy the expected fuel rate requirements of the system. The bu ffer tank and the reaction chamber (and/or metal ng hydrogen at some reasonable pressure. Thus, container) would need to be capable of retaini icantly larger than the material weights and the system weight and volume would be signif volumes alone. Version 1.0 - 2008 Page 16 of 26

17 Thermal management is a nother system issue that cannot be ignored. The Al/water reaction is at ~50-100 C (see highly exothermic with an enthal py of reaction of about 280 kJ/mol H 2 Appendix I). At the peak hydrogen rate example cited above for an 80 kW fuel cell, the reaction would generate a heating power of about 225 kW . For the hydrogen buffer tank approach tion rate of ~20% of full power, described above with an assumed steady state hydrogen genera the average heat generation rate would be about 45 kW. This level of heat generation could be manageable, but would have to be considered in the system engineering design and would result that heat rejection and radiator size are already in additional weight and volume. Due to the fact issues in typical fuel cell system designs, the additional heat rejection may be considered a key temperature dependent, some temperature control disadvantage. Also, since the reaction rate is the reaction bed at a preferred operating temperature, e.g., ~50 C. would be desirable to maintain Except for the occasional case of inadvertently runn ing out of fuel, vehicles are always refueled before the fuel tank is fully emptied. Hence, an on-board hydrogen generation system might utilize most, but not all, of the metal reactant. Refueling in this case entails not only supplying more aluminum or alloy, but also the removal of the aluminum hydroxide reaction product. The straightforward approach proposed above, where th e liquid is supplied to the metal container, results in a mixture of fresh and spent products. In this case, some method for separating out the reaction products from the unreacted aluminum on-board the vehicle will be required, so that the same time that it is refueled with fresh these products can be discharged from the vehicle at quantities of aluminum and water . Regeneration of Aluminum -Water Reaction Products: All of the promoter approaches result in the fo rmation of aluminum hydroxide and it is this material that will need to be refined back to pure aluminum with very high efficiency for the process to be viable for hydrogen generation. Furthermore, any alloying species (such as gallium) or materials used to promote the reaction (such as NaOH or NaCl) must also be fully recoverable (or nearly so). An excellent discussion of the process for prima ry aluminum production, as well as world-wide values for the energy requirements for alumi num smelting, can be found on a website produced by the International Aluminium Institute ( Briefly, the process is as follows: ore uses the Bayer process chemistry which Aluminum refining from aluminum-bearing bauxite forms a hydrate which is essentially the same as the reaction product in the proposed aluminum- water reactions described above. The hydrate is then calcined to remove the water to form o tallic aluminum at about 900 alumina. The alumina is electrolytically reduced into me C using ing a metal with 99.7% purity (see Figure 10). The smelting the Hall-Heroult Process, produc process requires continuous operation to be efficient. For 2005 (the latest figures reported by the Institute), the North American average energy used to reduce the oxide to the metal (smelting) is 15.552 kWh per kg of Al. This number does not include the energy used in mining and transporting the ore, the energy for processing the ore to the oxide, or the energy used in casting or carbon plants. Version 1.0 - 2008 Page 17 of 26

18 Figure 10: Schematic of the Hall-Heroult aluminum smelting process ( A regeneration loop for the spent material in the proposed hydrogen generation schemes would ned to remove the water and action product would first be calci be essentially similar. The re . This would then be reduced electrolytically to metallic aluminum. Thus, O form alumina, Al 3 2 the energy requirement quoted above would also apply to the present application, assuming a large facility that would be opera ted continuously. This energy would not include that needed for collecting, separating, and shipping the reac tion products of the water reaction, calcining the material or for returning the metal to the point of use. It can provide an indicator, however, for other methods and to the DOE targets. comparing hydrogen generation costs to eaction generates one kg of , the aluminum water splitting r At a weight fraction of 3.7 wt.% H 2 hydrogen through the consumption of 9 kg of Al (assuming 100% yield). Using a value of 15.552 kWh per kg Al, the energy required to produce 1 kg of hydrogen would then be 140 kWh, or 500 MJ. Since the energy content of 1 kg of hydrogen is 120 MJ (lower heating value of drogen obtained using aluminum produced by the hydrogen), the overall energy efficiency for hy Hall-Heroult electrolytic process is 24%. st of electricity. lt process depends on the co The cost of hydrogen generated by the Hall-Herou ice for electricity in the U.S. in 2005 was 5.05 cents per kWh (26). The average wholesale pr Using this value, the electricity cost alone to pr oduce 1 kg of hydrogen from 9 kg of aluminum is $7/kg H . The actual cost of hydrogen produced from aluminum water splitting will 2 undoubtedly be higher than this, due to additional costs associated with transporting material for reprocessing and back again for refueling, as well as for preparing the hydroxide for the e November 2007 commodity price for aluminum aluminum smelting operation. For example, th if the aluminum is $2.36/kg, which would translate into a hydrogen cost of $21/kg H 2 commodity price were employed. Version 1.0 - 2008 Page 18 of 26

19 Another factor that should be considered is the amount of aluminum that would be required to produce hydrogen for large numbers of hydrogen-fueled vehicles. It is estimated that the fueling drogen per year. Since it of 300 million vehicles would requi re 64 million metric tons of hy requires 9 tons of aluminum to produce 1 ton of hydrogen through the aluminum-water reaction, this means that the fueling of 300 million vehicles would require 575 million metric tons of aluminum per year. To put this number in perspective, the world-wide production of aluminum in the year 2006 was 24 million metric tons ( Thus, the hydrogen a the aluminum-water re fueling of very large numbers of vehicles vi action would require an approximately a factor of 25. In addition to expansion of world-wide aluminum production by the capital cost of new aluminum smelting facilities, the electricity consumption for aluminum production would have to increase by a similar factor. Summary: The key aspects associated with the production of hydrogen using the aluminum-water reaction are: Aluminum Required: 9 kg Al per kg H assuming 100% yield 2 Gravimetric Hydrogen Capacity: 3.7 wt.% (materials only) Volumetric Hydrogen Capacity: 36-46 kg H /L (materials only) 2 -4 /sec/g of Al – from published data to date g H Reaction Kinetics: 2 x 10 2 (based on the cost of electricity for aluminum production considering only Cost: $7 per kg H 2 the reduction of alumina to aluminum step) It should be emphasized that the hydrogen capacity values given in this paper for the aluminum- water reactions are for the materials only. Ther e are, in addition, a number of on-board system requirements that would add more weight and volume. Some ex amples are containers for the ism for unloading spent materials and loading fresh materials and the reaction products, a mechan ntities of materials to would allow controlled qua fresh materials, a reactor that react, devices for materials between the different components, water recovery sub- transporting solid and/or liquid systems (if used), heat exchangers, pressure control valves, etc. Some of these system gn and fabricate for reliability and longevity. components may prove to be very difficult to desi The highly transient behavior of the fuel requireme nts for vehicles would be particularly difficult e started, runs to completion. This is very to accommodate with a chemical system that, onc different from a metal hydride, for example, where the thermodynamics of the material both supply and limit the equilibrium hydrogen pressure at a given temperature. The current DOE hydrogen storage system capacity targets are a hydrogen gravimetric capacity /L (27). It is clear from the analysis of 6 wt.% and a hydrogen volumetric capacity of 45 g H 2 presented in this White Paper that no aluminum-w ater reaction system can meet these targets. Additional negative factors are the high cost of hydrogen from this process, and the amount of ale vehicular applications. aluminum required for large-sc Version 1.0 - 2008 Page 19 of 26

20 While such systems cannot meet the requirements for on-board vehicular hydrogen storage, the use of aluminum as a water splitting agent for generating hydrogen may have utility for non- vehicular applications, such as fixed-site electrical generators and electronic devices. The critical issues in these cases will be the modular ity, the cost of such al uminum-water systems, and the economics of the delivered energy content relative to other fuel systems. References: 1. I. E. Smith, Hydrogen generation by means of the aluminum/water reaction; Journal of Hydronautics (1972), vol. 6, #2, 106-109 U.S. Patent 3,932,600; Process for the generati 2. on of hydrogen; January 13, 1976; Inventors: Heinric Gutbier, Karl Hohne; Assignee: Siemens Aktiengesellschaft. M. Digne, P. Sautet, P. Raybaud, H. Toulhoat, E. Artacho, Structure and Stability of 3. Aluminum Hydroxides: A Theoretical Study”, J. Phys. Chem. B, 106 , 5155-5162 (2002). rial and method to dissociate water; December 29, 1981; U.S. Patent 4,308,248; Mate 4. Inventor: Eugene R. Anderson; Assi gnee: Horizon Manufacturing Corporation. 5. U.S. Patent 6,506,360; Method for producing hy drogen; January 14, 2003; Inventors: Erling Reidar Andersen, Erling Jim Andersen; Assignee: None listed drogen; October 28, 2003; Inventors: 6. U.S. Patent 6,638,493; Method for producing hy Erling Reidar Andersen, Erling Jim Andersen; Assignee: None listed U.S. Patent 6,800,258; Apparatus for producing hydrogen; October 5, 2004; Inventors: 7. Erling Reidar Andersen, Erling Jim Andersen; Assignee: None listed U.S. Patent 7,144,567; Renewable energy carri er system and method; December 5, 2006; 8. Inventor: Erling Jim Andersen; Assignee: None listed D. Belitskus, “Reaction of Aluminum With Sodium Hydroxide Solution as a Source of 9. Hydrogen”, J. Electrochem. Soc., 117 , 1097-1099 (1970). Generation from Aluminum in an Alkaline D. Stockburger,, “On-Line Hydrogen 10. Solution”, Proc. Symp. Hydrogen Storage, Electrochem. Soc., 43, 1-44 (1992). 11. International Patent A pplication PCT/CA2001/001115; Hydrogen generation from water split reaction; February 21, 2002; Inventors: As oke Chaklader, Das Chandra; Assignee: The University of British Columbia. from water split reaction; August 27, 2002; U.S. Patent 6,440,385; Hydrogen generation 12. Inventor: Asok C.D. Chaklader; Assignee: The University of British Columbia. Version 1.0 - 2008 Page 20 of 26

21 13. U.S. Patent 6,582,676; Hydrogen generati on from water split reaction; June 24, 2003; Inventor: Asoke Chandra Das Chaklader; Assignee: The University of British Columbia. Z-Y Deng, J.M.F. Ferreira, Y. Tanaka, a 14. nd J. Ye, “Physicochemical Mechanism for the -Al2O3-Modified Aluminum Powder with Water”, J. Am. Ceram. Soc., γ Continuous Reaction of 90, 1521–1526 (2007). A.G. Munoz and J.B. Bessone, "Pitting of 15. aluminum in non-aqueous chloride media", Corrosion Science, 41 , 1447- 1463 (1999). 16. itting of aluminum by chloride ions", Corrosion E. McCafferty, "Sequence of steps in the p , 1421-1438 (2003). Science, 45 International Patent A 17. pplication PCT/CA2005/000546; Compositions and methods for generating hydrogen from water; October 20, 2005; Inventors: Tomasz Troczynski, Edith Czech; Assignee: The University of British Columbia. croporous metals and methods 18. International Patent A pplication PCT/CA2006/001300; Mi February 15, 2007; Inventors: Tomasz for hydrogen generation from water split reaction; Troczynski, Edith Czech; Assignee: The University of British Columbia. Application PCT/US2006/000180; 19. Method and composition for International Patent Jasbir Kaur Anand; Assignee: Hydrogen production of hydrogen; July 6, 2006; Inventor: Power, Inc. hod for generating hydrogen gas utilizing U.S. Patent Application 20060034756; Met 20. ors: Maseo Wa les; February 16, 2006; Invent activated aluminum fine partic tanabe, Ximeng Jiang, Ryuichi Saito; Assignee: Dynax Corporation. 21. U.S. Patent 5,634,341; System for generating hydrogen; June 3, 1997; Inventors: Martin Klanchar, Thomas G. Hughes; Assignee: The Penn State Research Foundation. U.S. Patent 5,867,978; System for generating hydrogen; February 9, 1999; Inventors: 22. gnee: The Penn State Research Foundation. Martin Klanchar, Thomas G. Hughes; Assi U.S. Patent 4,358,291; Solid state renewable energy supply; November 9, 1982; Inventors: 23. Jerome J. Cuomo, Jerry M. Woodall; Assignee: International Business Machines Corporation. J.M. Woodall, “The Science and Technology of Aluminum-Gallium Alloys as a Material for 24. Hydrogen Storage, Transport and Splitting of Water”, Keynote Address, ECHI-2 Conference, April 12, 2007, Purdue University. 25. Tarek-Abdel Baset, “Systems Approach to On-Board Hydrogen Storage Systems”, Chrysler Focus Session on High-Density Hydrogen Storage for Corporation presentation in the Automotive Applications: Materials and Methods, Materials Science & Technology 2007 Conference, September 16-20, 2007, Detroit, Michigan. Version 1.0 - 2008 Page 21 of 26

22 DOE Energy Information Administration, Form EIA-861, Annual Electric Power Industry 26. ricity/wholesale/w holesalet2.xls). Report ( c, Carole Read, George Thomas, and Grace Ordaz, “The U.S. 27. Sunita Satyapal, John Petrovi Department of Energy’s Nati onal Hydrogen Storage Projec t: Progress towards meeting lysis Today, vol. 120, 246-256 (2007). Also see hydrogen-powered vehicle requirements”, Cata Version 1.0 - 2008 Page 22 of 26

23 Appendix I Thermodynamics of Alum inum-Water Reactions The possible reactions of aluminum with water are the following (3): + 3H O = 2Al(OH) 2Al + 6H (1) 2 2 3 O = 2AlO(OH) + 3H 2Al + 4H (2) 2 2 2Al + 3H O = Al O + 3H (3) 3 2 2 2 (bayerite). The second possible reaction product The first possible reaction product is Al(OH) 3 (alumina). These reaction O is AlO(OH) (boehmite). The third possible reaction product is Al 2 3 All three of them produce the same amount of products differ in their level of hydration. hydrogen with respect to the amount of aluminum reacted, but they differ in the amount of water that is required for the reaction. These reacti ons are all thermodynamically favorable over a wide temperature range from room temperature to temperatures far in excess of the melting point o C). In addition, all of these re actions are highly exothermic. of aluminum (660 Table A1: Thermodynamic data for the aluminum-water reaction to form bayerite. 2/3Al + 2H2O = 2/3Al(OH)3 + H2(g) T Δ H Δ S Δ G kJ/mol H2 J/K kJ/mol H2 C 0 -277 26.2 -284 100 -284 3.29 -286 200 -291 -12.1 -285 300 -298 -25.1 -283 400 -306 -38.0 -280 500 -316 -51.8 -276 600 -328 -66.8 -270 700 -350 -90.9 -262 800 -369 -109 -252 900 -391 -128 -240 1000 -417 -149 -232 Version 1.0 - 2008 Page 23 of 26

24 Table A2: Thermodynamic data for the aluminum-water reaction to form alumina. 2/3Al + H2O = 2/3Al2O3 + H2(g) T Δ G Δ S H Δ C kJ/mol h2 J/K kJ/mol h2 0 -272 62.1 -289 100 -275 51.1 -294 200 -279 43.1 -299 300 -283 35.5 -303 400 -288 27.3 -306 500 -294 18.1 -308 600 -303 7.80 -310 700 -320 -11.3 -309 800 -333 -23.7 -308 900 -348 -37.1 -305 1000 -366 -51.6 -304 O and Al The thermodynamic parameters for the Al(OH) reactions are given in Tables A1 and 3 3 2 C Thermodynamic Software, Version 4.1. The A2 as a function of temperature, using HS . Thermodynamic data for AlO(OH) was not available in HSC tabulated values are per mol H 2 es for this species are intermediate between 4.1, but it is presumed that the thermodynamic valu O and Al . Al(OH) 3 3 2 For both of these reactions, the enthalpy is high ly exothermic, with an average value of -280 o o at 700 C. One may also see that Al C and -335 kJ/mol H at 100 becomes more O kJ/mol H 3 2 2 2 thermodynamically favorable than Al(OH) at elevated temperatures. 3 Version 1.0 - 2008 Page 24 of 26

25 ared to aluminum oxide is shown in Figure The free energy of the aluminum hydroxides comp A1 below (3): Figure A1: Thermodynamic stability of aluminum hydroxides versus temperature (Figure 6 and Table 8 from Reference 3). o C, Al(OH) is the most stable product, while As may be seen, from room temperature to 280 3 o o from 280-480 C, AlO(OH) is most stable. Above 480 C, Al O is the most stable product. 2 3 This means that the stable product of the reacti on of aluminum with water at room temperature will be the aluminum hydroxide Al(OH) . 3 Version 1.0 - 2008 Page 25 of 26

26 Appendix II volved With Hydrogen Generation From Examples of Organizations Presently In Aluminum-Water Reactions Website Product Organization AlumiFuel Cartridge Hydrogen Power Inc. Altek Fuel Group Inc. Hydrogen Fuel Cartridge Aluminum-Gallium Purdue University Alloys No website available HydPo Ltd. Hydrogen from aluminum-water Email contact: [email protected] reaction Version 1.0 - 2008 Page 26 of 26

Related documents