E-mail: arnulf@ite.no
arnulf@julo.com
Maeland A. J.
Institute of Energy Technology, Norway
The storage of hydrogen for vehicular use - A review and reality check*
INTRODUCTION
Environmental considerations have given momentum to the search for clean fuels to replace or at least offer an alternative to gasoline or diesel as the primary fuels in vehicular applications. Large amounts of carbon mono- and di-oxide, nitrogen and sulfur oxides, hydrocarbons and particulates produced in the combustion process of gasoline and diesel vehicles continue to pollute our atmosphere, sometimes resulting in dangerous levels of these pollutants in metropolitan areas. Devices to reduce the level of pollutants, e.g. the catalytic converter, have to some extent eased, but by no means eliminated the problem and while further improvements may be forthcoming, complete removal is an unrealistic expectation. However, there is another solution: employ a fuel which does not produce pollutants. Hydrogen with specific energy content of 33.3kWh/kg, nearly three times larger than gasoline or diesel, is such a fuel and it can be used in a conventional internal combustion engine with only minor modifications; the main product of the combustion process is water. (Combustion of hydrogen in air produces small amounts of nitrogen oxide (a pollutant), but the amounts are miniscule in comparison with the pollutants produced by the combustion of gasoline or diesel). Hydrogen may also be used with high efficiency in a fuel cell to generate electrical energy; the estimated efficiency is twice that of present day automobile engines. The fuel cell, however, requires very pure hydrogen for continued use while hydrogen purity is of much less importance for the internal combustion engine. An additional point to keep in mind is that exhaust heat is available from the internal combustion engine to liberate hydrogen, but not from the fuel cell. This feature is important if hydrogen is stored in a hydride. Both alternatives are being pursued for vehicular applications, but the current emphasis is on the use of hydrogen in fuel cells. Among the large automobile makers, the Ford Motor Company is working on a hydrogen powered internal combustion engine as an alternative to the gasoline engine until the automakers perfect the fuel cell power train. BMW plans to put a small fleet of cars on the road in 2000 designed to operate on a fuel cell/hydrogen combustion hybrid concept; the BMW hybrid will run on a hydrogen combustion engine and the fuel cell will power the car's on board electrical system [1].
Major difficulties to making hydrogen the clean fuel of choice in automotive applications exist. Hydrogen is the lightest of the elements with an atomic weight of 1.0079 and is a gas under ordinary conditions, making efficient onboard storage more than a trivial matter. The density of hydrogen gas at STP is 0.08988kg/m3 [2] (only
one-seventh that of natural gas) and onboard storage of quantities needed for practical driving ranges requires large volumes and high compression. Furthermore, hydrogen diffuses in and reacts with many materials, properties which must be taken into account when designing containers and which are reflected in the heavy and bulky steel tanks which have traditionally been used for storing compressed hydrogen. We will review recent work aimed at developing lightweight, high pressure containers for mobile application with hydrogen storage capacity > 10% hydrogen by weight.
Storing hydrogen as a liquid improves the volume efficiency greatly, but the liquefaction process is energy intensive, requiring cooling to 20K with a theoretical energy expenditure of 3.92 kwh/kg; in practice about 10 kwh/kg is needed. Well insulated, expensive, vacuum insulated containers are of course also required, and for prolonged storage of liquid hydrogen there is the additional problem of "boil off".
Chemical storage of hydrogen in the form of metal hydrides represents an attractive alternative which has received much attention in the past 30 years. The advantages of storing hydrogen in the form of metal hydrides include high volume efficiency, relative ease of recovery, indefinite storage capabilities and a high degree of safety. Our review will include a critical evaluation of this form of storage and prospect for further development.
Compressed hydrogen storage capacity can under appropriate conditions of temperature and pressure be augmented by adsoption (physi-sorption) on activated carbon. In addition, recent reports from Northeastern University (NU) in Boston, Massachusetts, U.S.A. have claimed large hydrogen storage capacities (as high as 75% by weight!!) of certain forms of carbon nanofibers. These reports have generated considerable interest, but also much controversy because confirmation by other laboratories are lacking. Our review will address this issue.
HYDROGEN STORAGE TECHNOLOGIES
Compressed Hydrogen Gas
The problem of onboard storage of gaseous hydrogen can be appreciated if we consider the fact that a small fuel cell powered automobile requires approximately 3 kg hydrogen to have a modest range of 500 km. The volume occupied by this amount of gas at room temperature and atmospheric pressure is 36,000 liters! The volumetric density (weight of stored hydrogen/volume of gas) can be improved by compression; e.g. compression to 20 MPa
19
*Presented as report IFE/I-99/012 on the First International Seminar on safety and economy of hydrogen transport, Russia, Sarov, 23-30 july 2000: received 25 july 2000.
(200 bar) reduces the volume to 180 liters. Compressed hydrogen is commonly sold in 50 liter steel cylinders pressurized to about 20 MPa. Four of these provide a little over 3 kg of hydrogen, but they also weigh approximately 270 kg! The challenge for this storage technology is to design and develop lightweight containers with significantly better performance. Early results from research in several laboratories indicate the challenge is being met with noteworthy success. The performance requirements for storing compressed hydrogen are in many respects similar to those of storing compressed natural gas for vehicular use: light weight, low cost and increased safety. Many aspects of the technology developed for natural gas storage have therefore been adopted and modified when neccessary for compressed hydrogen storage. Dynatek Industries of Calgary, Alberta, Canada, for example, makes a light weight cylinder designed for roof top mounting and is used in the Daimler Chrysler Nebus bus recently
Pused in a demonstration project in Oslo. The storage tank is constructed of a thin aluminum liner (about 3mm wall thickness with a burst strength of 7.7 MPa) wrapped with a composite of carbon fiber in an epoxy resin; the cylinder has a burst strength of about 63 MPA [3]. The Nebus has seven 150 liter tanks of this type, pressurized to 30 MPa and delivering 21 kg hydrogen for a 250 km range. In the U.S.A the Lawrence Livermore National Laboratory (LLNL) in California together with their industrial partners, Thiokol and Directed Technologies, Inc., are fabricating high performance prototype tanks which meet the U.S. Department of Energy (DOE) goals for the year 2000. The goals are 4000 Wh/kg, 12% hydrogen by weight, 700 Wh/liter, 34.5 MPa at room temperature, and $20/kWh in high volume production [4]. These pressure vessels use technologies that are easily adopted and instead of four tanks weighing 270 kg as in our example above, one tank weighing 25 kg and with an internal volume of 143 liters would do. The LLNL prototype tanks are made of carbon fiber composites wrapped around a thin metalized plastic liner which provides the permeation barrier for hydrogen. The volume and placement of the storage tank(s) pose major problems in retrofitting passenger cars, but is much less of a problem in a vehicle designed from the ground up. However, a nightmare of safety and regulatory issues remain unresolved. A particular safety concern is in my opinion the potential energy stored in the gas due to the high pressure. The Nebus tank, for example, represents 4.5 million newton meter (4.5 Nm) of force stored as potential energy in the gas pressure alone!! Such a tank, if punctured, could become a dangerous projectile. A frightful thought indeed.
Compression obviously requires energy. To compress from 0.1 MPa to 30 MPA takes approximately 15% of the total energy available in hydrogen. Onboard storage of hydrogen as compressed gas is currently the most common form for storage used in the many vehicular demonstration projects now in progress throughout the world [5]. The technology is here and readily available. While compressed hydrogen storage may be adopted in the public transport sector (buses), I believe it will be hard to sell the general public on the safety of having a high pressure tank of hydrogen in their own cars.
Liquid hydrogen
The volumetric density of hydrogen can be further improved by liquefaction. Liquid hydrogen is 788 times more dense than the gas at STP, i.e. 70.8kg/m3 at the b.p.(2). 3 kg of hydrogen can be stored in a cryogenically designed
system weighing about 45 kg and having a volume of about 100 liters [5], a bit large, but managable. Liquefaction, however, is energy intensive requiring cooling to 20K and consumes 30% or more of the energy contained in hydrogen. Liquid hydrogen is the choice fuel for launching of space vehicles because it stores 3 times the energy of jet fuel by mass; it can be loaded directly just before take-off and consumed within the short time it takes to launch. Due in part to the space program the liquid hydrogen technology is quite advanced. However, several drawbacks are encountered in use as a fuel for automotive applications. The volumetric density, although better than that of the compressed gas, is still unfavorable as compared to gasoline. It takes a volume 4 times that of gasoline to produce the same power and the auxillary equipment to maintain a storage vessel below 20K is very expensive and continually requires energy to function. Another problem encountered is that of " boil-off" losses. Even with the best of tanks built, hydrogen will need to be vented if dangerous build up of pressure is to be avoided. The venting leads to loss of fuel, typically 1.5-2.0% per day. Furthermore accumulations of hydrogen in closed-in areas such as garages constitute a risk of explosion. A hybrid storage system being tested by LLNL is reported to virtually eliminates boil-off losses [6]. The system uses liquid hydrogen for long trips and ambient temperature pressurized hydrogen for daily driving. Another concept uses magnetic levitation to hold a liquid hydrogen tank magnetically suspended inside an outer tank without the usual struts needed to fix the inside tank [7]. The heat influx into the tank is thus reduced and the "boil off" rate significantly reduced.
Among the automakers, BMW has a four-door sedan using liquid hydrogen as fuel in an internal combustion engine. The car is in use at Munich, Germany, airport ferrying VIPs between the terminal and aircrafts [6]. Renault and Daimler-Chrysler are also testing fuel cell powered vehicles, the Laguna station wagon and the Necar 4 respectively, using liquid hydrogen as fuel [5]. The German busmaker MAN will introduce a fuel cell bus next June in connection with the Expo 2000 World Fair which will be held in Berlin; 350 liters liquid hydrogen will be carried in two tanks mounted on the roof [6].The MAN SL 202, 92 passenger city bus has operated very successfully in Erlangen and Munich since 1996 with an internal combustion engine and liquid hydrogen storage [8].
Slush hydrogen, a mixture of liquid hydrogen and solid hydrogen, is being considered a form of storage for use in aircrafts, but could also be considered for automobiles. Two advantages are apparent: higher density (approximately 15% higher than liquid hydrogen) because of solid constituents and less propensity to evaporate because heat is first used for the phase transition, i.e. melting of solid hydrogen, before supplying heat for the evaporation process.
Hydrogen Storage in Metal Hydrides
Chemical bonding of hydrogen in the form of metal hydrides represents an attractive storage alternative which has received much attention in the past 30 years. Many metals and alloys react with hydrogen according to equation (a)
M + s/2 H2 <===> MH
(a)
where M is a metal, solid solution alloy or an intermetal-lic compound(IC) and s is the atomic ratio of hydrogen to metal. The reaction is exothermic and reversible, i.e. hydrogen may be recovered by heating the hydride. Elements which form solid metal hydrides are shown in Figure 1 [9].
Figure 1. Elements which form solid metal hydrides.
IA HA
Li Be m B
Na Mg Mg IIIA IVA VA VIA VIIA VIII I B II B Al
K Ca Sc Ti V Cr * Mn * Fe * Co * Ni * Cu Zn Ga IV B
Rb Sr Y Zr Nb Mo * Tc * Rh * Pd Cd In Sn
Cs Ba La Hf Ta Tl Pb
Ac
Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Yb Lu
Th Pa U Np Pu Am Cm Bk
Ionic hydrides Covalent hydrides Metallic hydrides
(*) Metals requiring hydrogen pressures greater than 1 atm. (0.1 Mpa) to form hydrides.
They are for convenience classified as covalent, ionic and metallic without implying rigid boundaries between the groups. The covalent hydrides do not form directly and reversibly by reaction (a) and are therefore of no interest as storage materials. Furthermore the ionic and metallic metal hydrides are with few exceptions either too stable (see discussion below), i. e. have too low dissiciation pressures, or too unstable, i. e. have very high dissociation pressures at or near ambient temperature (marked with * in Figure 1) to be useful in automotive applications. The stable binary metal hydrides, e.g.MgH2 andTiH2, need to be heated to several hundred degrees to obtain a 0.1MPA dissociation pressure, and the unstable binary metal hydrides, e. g. CrH and NiH, have dissociation pressures in the kbar range. However, many alloy hydrides do at or near room temperature have dissociation pressures which make them interesting as hydrogen storage compounds [9].
The volumetric density of hydrogen in metal hydrides is very high; higher in fact in some hydrides than in liquid or even solid hydrogen as seen in Table I [9].
Table I. Volumetric and gravimetric hydrogen densities in various media.
Medium Volume density of hydrogen, Hydrogen weight density, wt %
1022 atoms/cm3
H2 gas at 100 atm 0.5 100
H2 liquid (20K) 4.2 100
H2 solid (4.2K) 5.3 100
LiH 5.9 12.6
pdHo6 4.3 0.6
H2O (liquid) 6.7 11.2
MgH2 6.7 7.6
TiH2 9.2 4.0
VH2 11.4 3.8
UH3 4.0 1.3
TiFeH2 6.0 1.9
LaNi5H7 7.6 1.6
LiAlH4 5.7 10.6
Mg2NiH4 5.9 3.6
The number of hydrogen atoms /cm3 in VH2, for example, is 11.4x1022, more than twice that in solid hydrogen at 4.2K, i. e. 5 . 3x1022. For the familiar storage materials TiFeH2 and LaNi5H7 the corresponding numbers are 6.0x1022 and 5.3x1022, respectively.Other compounds, e. g. H2O, also have high volumetric density of hydrogen, but temperatures in excess of 2775K are needed to thermally decompose H2O and recover hydrogen! It is this high volumetric density of hydrogen in metal hydrides and the relative ease of hydrogen recovery that form the basis for their potential use as hydrogen storage materials. The ease of hydrogen recovery is reflected in the dissociation pressure of the hydride and can be understood by considering the pressure-composition isotherm shown in Figure 2. Hydrogen first dissolves in the metal or alloy with increasing hydrogen pressure:
M + y/2 H2 <===>MH metal solution phase
(a-phase)
(b)
CD
a. c
0) <3>
0 >
1
Total Hydrogen to Metal Ratio
Figure 2. Pressure composition isotherm.
Following saturation of hydrogen solution in the metal, the hydride MHx, begins to form. (Note: Substantial deviation from stoichiometry is common in the hydrides). Conversion of the saturated solution phase to hydride phase continues as hydrogen is added while the pressure remains invarient in accordance with the Phase Rule:
MHV + (x - y)/2 H2
=> MH
(c)
The invarient plateau pressure is the equilibrium dissociation pressure of the hydride at the temperature of the isotherm and is a measure of the stability of the hydride. After complete conversion to the hydride phase, the non-stoichi-ometric hydride absorbs hydrogen as the hydrogen pressure increases:
MHx + (s - x)/2 H2
■ MHs hydride solution phase (ß-phase) (d)
Multiple plateaus are possible. In the V-H2 system, for example, two plateaus are observed. The first corresponds to the formation of a monohydride phase, VHx (fi -phase), while the second plateau represents conversion of the hydrogen saturated monohydride phase to the dihydride phase ( /-phase), Figure 3.
0.8 1.0 H/M Atom Ratio
Figure 3. Pressure - Composition Isotherm for V (schematic).
The dissociation pressures increase with increasing temperatures as shown for selected binary hydrides in Figure 4 and IC hydrides in Figures 5. It is seen from Figure 4 that most binary hydrides are too stable and require high temperatures to liberate hydrogen at usable pressures. MgH2 and TiH2, for example, require heating to nearly 575K and more than 875K, respectively, to recover hydrogen at 0.1 MPa. Only Vanadium dihydride will provide reversible hydrogen at usable pressures at room temperature. However, only half the hydrogen is available, VH2 <=>VH0 95, representing 1.9 wt %; high temperature is required to recover the remainder. The situation is much improved when we consider the IC hydrides, Figure 5. Hydrogen is available at room temperature (or even below) for LaNi5H6, FeTiH2 and MNi5H6 (M = misch metal).
Temperature, °C 600 400 300 200
1000-
§ 100-
° 0.01
0.0001
1.5 2 2.5
1000/T, IC1
Figure 4. Dissociation pressures ks. Temperature (10).
Temperature, °C
300 200 100 50
too
50
£ 20
C 2
o
Œ 1.0
o
° 0.5 b
- 1 1 1 1 1 1 1-
A \
- \i ; \f
Figure
2.0 2.5 3.0 3.5 4.0
1000/t, ic'
5. Dissociation pressures ks. Temperature (10).
Libowitz et al. reported already in 1958 that the IC ZrNi reacted reversibly with hydrogen to form a ternary hydride, ZrNiH3, which had stability (dissociation pressure) intermediate between the very stable ZrH2 and the very unstable NiH [11]. This important discovery opened the door to a new class of hydrides, the intermetallic compound hydrides, which could be considered "pseudo binary" hydrides with properties which could, at least to some extent, be controlled by alloying and substitution. A flurry of activities directed towards discovering new IC hydrides and to "tailormake" hydride stabilities began in the 1970s following the discovery of AB5 (LaNi5 ) [12] and AB (TiFe) [13] types hydrides. The focus of these studies have been IC hydrides whose nature and characteristics place them in the category of metallic hydrides. In order to form an IC hydride the IC must contain at least one hydride forming element.
Several IC hydrides which can provide hydrogen at moderate temperatures (ambient to 423K) and pressures suitable for automotive applications (>0.1 MPa) have been prepared and successfully used as fuel in the internal combustion engine in a number of vehicle demonstration projects since the early 1970s [14]. Exhaust heat was used to liberate hydrogen from the metal hydride. More recently Toyota has employed a metal hydride system to store hydrogen for a fuel cell powered sport-utility vehicle (RAVA4). Mazda has a station wagon (DemicoFCEV) which also utilizes a metal hydride for storing hydrogen [5] and a hybrid hydrogen powered internal combustion/electric bus has been developed for demonstration in the city of Augusta, Georgia, U.S.A., and has been in use since April 1997 [15]. While storing hydrogen as a metal hydride provides a number of desirable advantages such as high volumetric density, ease of hydrogen recovery, indefinite storage time without hydrogen loss, very high degree of safety, etc., it is an unfortunate circumstance that most metal hydrides are heavy in comparison to the hydrogen they carry, i. e. they have low gravimetric densities, and that the light ones, e. g. MgH2 and LiH2 which have high gravimetric densities (see Table I), are quite stable and require high temperatures to release hydrogen. Despite much effort in the past 2530 years it has proven very difficult to design IC hydrides containing more than about 2 wt. % hydrogen and capable of delivering hydrogen at usable pressures (>0.1 MPa) and acceptable temperatures (<425K). This empirical observation is a consequence of the chemistry of the hydrides. The gravimetric density depends on H/M. For the binary metallic hydrides this ratio has a maximum value of 1 for the Groups VIA -VIII elements, 2 for the Groups IVA-VA elements, 3 for the Group IIIA (Thorium is an exception in that a hydride exists for which H/M = 3.75) elements. The maximum H/M ratio for the IC hydrides is generally less, but never more than expected from the constituent elements, e. g. H/M for FeTiH2 is 1 not 1.5 (H/M=2 for TiH2 , and H/M=1 for FeH), H/M for LaNi5H7 is 7/6 not 10/6 (H/M=3 for LaH3, and H/M=1 for NiH), H/M=4/ 3 for Mg2NiH4 not 5/3 (H/M=2 for MgH2 ,and H/M=1 for NiH); in ZrNiH3 the ratio is 1.5, equal to expected the expected value (H/M=2 for ZrH2, and H/M=1 for NiH). Since the H/M ratio is fixed by chemistry and cannot be increased, we must look to the hydrides of light elements for high gravimetric density of hydrogen.The weight penalty associated with using metal hydrides to store hydrogen is the major disadvantage in mobile applications where weight has high priority. Magnesium based alloys ranks high on the list to increase the gravimetric hydrogen density, but efforts so far have had only limited success. Ames Laboratory, Ames, Iowa currently has a project to develop magnesium-based alloys for hydrogen storage [16]. The alloys are processd by high-pressure atomiza-tion to give spherical powder particles of consistently identi-
cal chemical composition. The alloys are developed by Ovonic Battery, Troy, Michigan, U.S.A., and are primarily for use in batteries, but may also be of interest for gaseous storage. Vanadium and titanium based solid solution alloys and possibly IC of vanadium and titanium are also of interest. Masuo Okada, Tohoku University, Sendai, Japan has obtained hydrogen storage capacities of 2.5-2.6 wt. % in V-based (BCC) solid solution alloys, but details are unavailable at this time pending patent disclosure [6, 17]. This is close to the Japanese national goal for effective hydrogen storage which is 3 wt. %, desorption below 373K, operating pressure below 1 MPa, and maintenance of more than 90 % of the initial capacity after 5000 absorption/desorption cycles [6]. The goals of the International Energy Agency (IEA) Program begun in 1995 are: 5 wt. % hydrogen with a desorption temperature of <373K and a desorption pressure of at least 1 bar. To meet the IEA goals it is clear that a new approach to hydrogen storage alloys is needed . Bogdanovic and Schwickardi have recently done just that and bugun a study of complex metal hydrides as possible hydrogen storage compounds [18]. Complex hydrides have not in the past been considered likely candidates for hydrogen storage despite the fact that many of them have high gravimetric hydrogen density. NaAlH4 and Na3AlH6, for example contain 7.5 and 5.9 wt. % hydrogen, respectively. The reason for this is that while hydrogen liberation from NaAlH4 and Na3AlH6 is thermodynamically favorable at moderate temperatures, the process is hampered by slow kinetics [19] and is reversible only under severe conditions [20].
3NaAlH4 <=> Na3AlH6 + 2Al + 3H2; (h)
Na3AlH6 <=> 3NaH + Al + 3/2 H2;
(i)
Recent experiments have shown that Ti doping can significantly improve the kinetics of desorption and make the dehydriding process reversible under moderate conditions; the effect of many hydriding-dehydriding cycles on the capacity are also minimal, a priority requirement of a good storage material [17]. These are truly remarkable results and could signify a whole new class of high gravimetric capacity hydrogen storage materials. Zidan et al [21] has shown that Zr is inferior to titanium as a catalyst for dehydriding NaAlH4 according to (h), but a superior catalyst for dehydriding Na3AlH6 according to (i). By combining the benefits of both catalysts NaAlH4 with greater than 4 wt.% cyclable hydrogen capacity and an onset of rapid hydrogen liberation at temperatures below 373K was reported. These results show promise, but we are still far from viable materials for practical vehicular applications.
Powerball Technologies of Salt Lake City, Utah, U.S.A. [22] has an interesting solution to the gravimetric density problem of hydrides. The company is promoting NaH for producing onboard hydrogen and for that purpose supplies NaH encapsulated in plastic spheres (Powerballs). The Powerballs are sliced open under water to yield hydrogen and NaOH. A small amount of water, stored between the Powerballs, is needed to start the production of hydrogen. Once in progress, however, water becomes available as waste product from the fuel cell. The NaOH produced is recylable and in fact must be returned to the company for feedstock. The system is claimed to be very safe. The new, second generation tank weighs about 23.6 kg, holds 17 liters of powerballs and delivers 1.29 kg hydrogen. The current price is $2000, but the company expect the price to drop below $1000 with large volume sales. A recycling pilot plant is under construction and will demonstrate a unique route to sodium metal using a proprietary process based on natural gas rather than electricity, and will validate the technical viability and overall economics for recycling waste NaOH back into Na metal for NaH production according to
Powerball president Jed Checketts. Zevco is testing the system in alkaline fuel cells in Britain, and Ford, Toyota and General Motors are said to be interested.
The concept of using ionic hydrides to liberate hydrogen by reacting with water as is done by Powerball Technologies is not new, but has been used in the past to generate laboratory hydrogen, fill weather balloons, etc. Kong et al. [23] have recently investigated the hydrolysis reaction in relation to hydrogen storage for fuel cell applications. Rates of reaction, heat evolution and yield of reaction were studied, but many questions relating to the economic and practical aspects of this technology remain unanswered.
Hydrogen Storage in Carbon
Activated carbons, typically obtained by thermochem-ical processing of mineralogical or organic precursors, are capable of adsorbing hydrogen and thus increase the volumetric density of hydrogen. Many different types of carbon structures are formed and provide a variety of environments for binding hydrogen. The adsorption, generally referred to as physisorption, is due to Van der Waals interactions between the solid atoms and the gas molecules resulting in greater concentration of hydrogen at the gas/ solid interface than in the bulk The amount adsorbed initially increases rapidly with increasing hydrogen pressure, but then tails off and eventually reaches "saturation". The interactions are weak by comparison with the chemical interactions (chemisorption) as seen in the metal hydrides and the adsorption is therefore most effective at low temperatures where the kinetic energy of the adsorbed hydrogen is too low to break the carbon-hydrogen attractive interaction. The advantage of this augmented hydrogen storage effect is the potential of storing the same quantity hydrogen gas at a much lower pressure. When saturation has been reached, the hydrogen concentration at the gas/solid interface is the same as in the bulk, and there is nothing further to be gained. The amount of hydrogen stored in an adsorption system is dependent upon the nature of the activated carbon (pore size and surface area), the storage pressure and the temperature. The most meaningful way of describing the amount of gas adsorbed is by the quantity "excess amount" which is "the excess material present in the pores over that which would be present under the normal density at the equilibrium pressure" [24].
While hydrogen adsorption on carbon has been known and studied for a long time, the concept of hydrogen storage on activated carbon is of newer date. Carpetis and Peshka [25] determined hydrogen adsorption isotherms at 65 and 78K and at pressures up to 4.15 MPa on a several different types of high surface area carbon materials and reported a maximum excess amount hydrogen adsorbed of 5.2 wt. % at 65K and 4.15 MPa for a F12/350 carbon material. They also concluded that the process was economically more advantageous than either storing hydrogen as a liquid or FeTi hydride. In a recent study Chahine and Bose [26] identified KOH based activated carbon, produced by the reaction of cokes with KOH, as the best carbons available commercially, but also expressed the need to develop low cost/high capacity solid adsorbants specifically designed for hydrogen storage. This material, (AX-21), has a cage like type of porosity and adsorbs twice as much hydrogen as regular grade activated carbon. The BET surface area is of the order of 3000m2/g, 2-4 times that of regular grade microporous carbon (The surface BET surface area is larger than the maximum possible surface area of both sides of a single isolated graphite sheet, 2620
m2/g, indicating perhaps condensation of N2, the probe used in the BET analysis, in some of the pores). Operation at low temperature is essential to storing hydrogen. Lowering the temperature from ambient to 175K will increase the adsorption by a factor of 5 and by a factor of 15 if the temperature is decreased to 77K. The relative efficiency of the adsorption process is greater at lower pressure. At 3.5 MPa and 77K, the hydrogen density of AX-21 is 25 g/ liter [26] (30% of liquid hydrogen density) and equivalent to 30 MPa of compression at room temperature or 7.8 MPa at 77K resulting in an enhancment factor of 2.2. If the pressure is reduced by 50%, i.e. to 1.75 MPa, the storage capacity is only reduced by 20%. The economics of the technology has been evaluated [27] and reported to compare favorably with the other methods of storage. The the advantage of low pressure operation vs. high pressure compression storage should also be pointed out.
Several new forms of carbon have recently been discovered and/or synthesized. Single walled nanotubes (SWNTs) constitute one such form of carbon with interesting hydrogen storage properties. A SWNT is essentially a sheet of graphite rolled into a seamless tube many microns in length. Van der Waals interactions between the tubes lead to bundles (or ropes) which may contain hundreds of individal tubes. The distance between the walls of nearest neighbor tubes in the bundle is referred to as the van der Waals gap. The nanotubes must have open ends to allow hydrogen adsorption inside the tubes. However, as made, nanotubes are capped with hemispherical fullerene domes made of six pentagons needed for closure. Tube opening can be accomplished by oxidation of the caps which due to their strained conditions, (five-membered rings) are more susceptible to oxidation than the rest of the structure [28]. High yield synthesis of SWNTs by different methods have recently been reported [28-30] and it is even possible to control the synthetic processes to the point of producing specific tube diameters in the range of 1-2 nm [29, 31, 32]. Many of the characteristics of SWNTs are favorable towards hydrogen uptake. They have large theoretical surface area and have very narrow pore size distribution with essentally all their surface area in the micropore range. Activated carbons have by comparison surface areas which are broadly distributed between macropores, mesopores and micropores. Gas adsorption on a porous solid is most effective when the pores are no larger than a few multiples of the diameter of the adsobed molecule [33], i. e. micropores are most effective in hydrogen adsorption. This is due to an enhanced density of hydrogen inside the pore brought about by the attractive potential of the pore walls [33]. It has in fact been suggested that nanotubes should be able to draw up liquids by capillary action [34], and such behavior has indeed been observed [35]. Heben and Dillon [36] have calculated that a single 2 nm SWN filled with hydrogen would have a gravimetric density of 5.1 wt.% assuming a condensed hydrogen phase inside the tube with H2-H2 nearest neighbor distances of 0.351 nm [37, 38] and a H2-C distance of 0.295 nm [39]. Additional increase in storage density would be expected if hydrogen was stored in the interstitial space between the individual SWNT and on the exterior surface of the SWNTs.The recent report by Dillon et al. [40] of 5-10 wt. % uptake of hydrogen at ambient temperature and 0.1 MPa pressure in a sample of crystalline SWNTs, prepared by co-evaporation of cobolt and graphite in an electric arc [41] and consisting of 1.2 nm tubes in bundles of typically 7-14 tubes, appeared to be in line with expectations. It should be noted, however,
that these measurements were made on very small samples (1 mg) and also very dilute samples; the analysis requiered a large correction for more than 99% of the material was assumed inert. The interaction between carbon and hydrogen is according to Dillon et al. an enhanched physical rather than chemical interaction; the strength of the interaction, 19.6kJ/mol, is about 5 times larger the typical hydrogen on planar graphite interaction [35]. Hydrogen adsorption and desorption isotherms on 200 mg high purity SWNTs have just been published [42] and the results are inconsistent with the Dillon et al. report. The hydrogen uptake was reported to be 8.25 wt. %, but only at low temperature (80K) and high pressures (to 12.0 MPa) in contrast to 5-10 wt. % uptake at ambient temperature and pressures below 0.1 MPa as reported by Dillon et al. The SWNTs were prepared by condensation of a laser vaporized carbon/nickel/cobolt mixture at 1473K [28] and the purified SWNTs were in the form of dense bundles (ropes), 6-12 nm in diameter, made up of individual tubes approximately 1.3 nm in diameter [42]. It is difficult to pinpoint a particular reason for the conflicting results other than to note that the two samples were prepared under different conditions and that the Dillon et al. sample was not purified.
An analysis of hydrogen adsorption in nanotubes in the low coverage limit using a phenomelogical interaction potential concluded that the adsorption capacity for hydrogen is increased [43], but only by about one half of the value reported by Dillon et al. In another study Darkrim et al. [44] using Monte Carlo calculations determined that SWNs would store about 15% more hydrogen than AX-21 at 298K and 10 MPa. Wang and Johnson studied the adsorption of hydrogen in SWNTs and idealized carbon slit pores by computer simulations [45]. In the model used, hydrogen is treated as structureless spherical particles, an approximation which works well in modelling fluid or solid hydrogen at reasonable pressures. The H2-H2 interaction is represented by the Silvera-Goldman potential (46) which satisfactorily models the properties of fluid hydrogen [47] and the H2-nanotube interaction is represented by the Crowell-Brown potential [48] which models the interaction of H2 with graphite sheets. The Silvera-Goldman and Crow-ell-Brown potentials have previously been used to model hydrogen adsorption on graphite and shown to accurately reproducing experimental adsorption isotherms [49]. Idealized carbon slit pores are represented by single graphite sheets with pore width H, the distance between sheets measured from carbon centers. In the case of the nanotubes the simulations included hydrogen adsorption in the space between the tubes. Adsorption isotherms at 77K and 298K were calculated for idealized slit pores of width 0.6 nm, 0.9 nm, 1.2 nm and 2 nm, corresponding to one, two, three and five layers of adsorbed hydrogen, respectively; the pressure range was 0.1- 10 MPa. The simulated adsorption isotherms for idealized carbon slit pores were in reasonable agreement with experimental data for activated carbon (AX-21), but the simulations for SWNTs did not confirm the large uptake reported by Dillon et al. Simulation at 133K, for example, gave 0.8 wt. % hydrogen at 0.04 MPa and at 10 MPa only 1.9 wt. % [45], far below the experimental value obtained by Dillon et al. A large difference in the isosteric heat of adsorption was also apparent; 19.6 kJ/mole hydrogen reported by Dillon et al. while the value obtained from simulations under the same conditions gave only 6.3 kJ/ mole. In view of the difference in electronic structure between nanotubes and graphite, it is possible that the use of planar graphite-hydrogen potential in modelling nano-
tube-hydrogen interactions is inadequate to describe the real situation and, therefore, lead to wrong conclusions.
At the 1996 fall meeting of the Materials Research Society (MRS), held in Boston, Massacusetts, Rodrigues and Baker of Northeastern University (NU), Boston, presented a paper in which they claimed the development of a "super" hydrogen storage material. The material, graphite nanofibers, discovered by Baker back in 1972 was claimed to be capable of storing 30 liters of hydrogen per gram of graphite [50, 51], i.e. 75 wt. % corresponding to CH36, nearly 5 times the amount calculated to be the theoretical capacity (6.2 liters/gram) by assuming a single layer of hydrogen molecules covering the surface of single crystal graphite. The graphite nanofibers were prepared by reacting hydrocarbons with carbon monoxide on catalytic particles of bi- or tri- metallic nickel or iron. The results reported by Rodriguez and Baker immediately caused controversy. Michael Heben of the National Renewable Energy Laboratory in Denver, Colorado, U.S.A., pointed out that the highest ratio of hydrogen to carbon found in nature is 4/1 (CH4) and corresponds to 25 wt. % and expressed skepticism [51] of the results and attempts to verify the results of Rodrigues and Baker have so far been unsuccessful. Ahn et al. [52] measured hydrogen adsorption and desorption at 77 and 300K on graphite nanofibers, grown by passing ethylene and H2 gases over Fe-Cu catalysts of different composition at 873K and reported that the absolute level of hydrogen desorption from these materials were typically less than 0.01 H/C which is comparable to other forms of carbon. Jarvi et al. [53] reported very low hydrogen storage capacity at 303K, comparable to activated carbon, for carbon nanofibers prepared by catalytic decomposition of ethylene over nickel, iron, copper/nickel and alumina/magnesia catalysts and concluded that they were unlikely storage materials for hydrogen. However, they left the door open by stating that subtle processing effect might convert inactive materials into effective hydrogen sorbents.
The announcement at the MRS meeting did not escape the automakers and Daimler-Chrysler began an evaluation study with the NU group of these "super" hydrogen storage materials. Last fall, however, Daimler-Chrysler ended their participation in the study. Ford Motor Company is currently supporting the NU group [54]. The NU group was also supported by the DOE in U.S.A. [55], but the support was terminated presumably because of the unwillingness of Rodrigues and Baker to share their samples with other DOE laboratories for examination. This unwillingness to submit samples to other investigators continues to fuel the controversy and promted Dr. Gary Sandrock, a well known expert in the field of hydrogen storage materials, to publi-cally call on Rodrigues to submit samples to others for a "Real-World Test" of her nanofiber material [56]. This, however, has not so far as I know been done, despite the fact that Rodrigues and Baker were issued a patent on "Storage of Hydrogen in Layered Nanostructures" August 5, 1997 [57] and have thus secured protection for their process.
The work of the NU group has been presented in a number of talks [58, 59, 60], and in a recent article in the Journal of Physical Chemistry [61]. The group's results may be summarized as follows: Nanofibers, prepared by catalyzed decomposition of carbon containing gases (and mixtures of gases) over selected metal and alloy surfaces in the temperature range of 723-1023K [62-64], yielded structures in which graphite platelets were arranged parallel (tubular), perpendicular (platelet), or at an angle (herringbone) with respect to the fiber axis, Figure 6.
(a) (b) (c)
Figure 6. Schematic diagram of (a) herring-bone, (b) platelet and (c) tubular structures produced by the thermal decomposition of carbon-containing gases over selected metal catalyst particles (58).
The fibers had lengths between 10 and 100 mm with crossections varying from 0.30 to 5 nm2. These structures, made up of platelets with essentially only edges exposed, are said to possess unique properties which are highly desirable for gas sorption applications [61]. The distance between the graphene layers is dependent upon the nature of the catalyst, the gas phase composition as well as the reaction conditions . The minimum layer distance is 0.335 nm, the value of single crystal graphite, which is slightly larger than the kinetic diameter of hydrogen, 0.289 nm. Hydrogen uptakes were determined by exposing, in a system of known volume, a purified nanofiber sample to hydrogen gas at room temperature, and observing the drop in pressure over a 24 hour period from an initial value of 11.2 MPa. The herringbone structures (there are many of them according to Rodriguez) take up hydrogen to give hydrogen densities in excess of 60 wt. %, the platelets take up hydrogen to give hydrogen densities in excess of 50 wt. % and the tubular structures give hydrogen densities in excess of 10 wt. %. [61]. The hydrogen uptake is a form of "intercalation" in which, due to the small difference between the separation between the platelets (the spacing between platelets suitable for hydrogen storage was reported as 0.337 nm [61] ) and the kinetic diameter of hydrogen the hydrogen molecule (0.289 nm), initial insertion of hydrogen is difficult and requires high pressure. Once "opened", however, separation of the graphene sheets occurs and permits additional layers of hydrogen to be added [58]. Most of the stored hydrogen is released at ambient temperature by decreasing the pressure below 3.4 MPa [58], but a fraction remains and requires high temperatures (> 673K) to be removed, suggesting both physisorption and chemisorption occur. It is proposed that hydrogen first form a monolayer on the graphite basel plane (chemisorption) and that
following complete coverage of the graphite planes, hydrogen molecules interact with the monolayer via physisorption [58].
Neutron scattering studies made on physisorbed mon-olayers of hydrogen on graphite [37, 38], which showed that at low temperature and low coverage the adsorbate adopted a commensurate,J3x*J3 triangular structure with a lattice parameter of 0.426 nm and a more condensed arrangement as the coverage increased and approached monolayer coverage, are sited in support of the remarkable hydrogen concentrations in these materials [58]. In this more condensed arrangement and as the second layer formed the lattice contracted and the lattice parameter of 0.351 nm, considerably smaller than observed in solid hydrogen at 4.2K, i. e. 0.376 nm, was obtained [37, 38]. This, according to the views of the NU group, proves that "the presence of graphite generates perturbations in the hydrogen packing characteristics to create a phase that is much more condensed than that of the bulk hexagonal closed packed structure typically present in the solid state" [58].
It is clear that it is impossible for hydrogen to adsorb into a 0.337 nm wide graphite pore because that is the distance between the carbon centers leaving no space available for adsorption. The spacing must be increased considerably to allow for large amounts of hydrogen to be adsorbed. The spacings have unfortunately not been measured while the nanofibers are exposed to hydrogen and the in situ spacings are therefore unknown. Wang and Johnson [45, 65] reported that based on Monte Carlo simulations of hydrogen adsorption in graphite nanofibers at 298K and 100 bar, a slit pore of 0.9 nm gave the largest hydrogen adsorption; the simulated weight fraction of hydrogen adsorbed for graphite nanofiber platelets were 0.46 wt. %. This is less by a factor of 100 or more from the val-
ues reported by the NU group, 46 and 54 wt. % [61]. Because the pore width is held constant in the simulations, additional simulations under the same conditions with pore widths of 1.5, 3.0 and 6.0 nm were performed in order to model the effect of expansion on the equilibrium adsorption. It was found that in platelets with pore widths larger than 0.9 nm, the weight fraction decreased an eventually became constant as the width increased [65]. The reason for this is that the carbon-hydrogen potential is very weak in the center of large pores and consequently has very little effect on the gas density near the pore center in the absence of capillary condensation; increasing the pore size will therefore not cause an increase in the adsorption in the pore. Wang and Johnson concluded [65] that standard solid-fluid potential models for carbon-hydrogen interactions cannot explain the very large values reported by the NU group and suggested that perhaps much stronger solid-fluid potentials could reproduce the NU results. However, when this was done they were led to the result that unrealistically strong solid-liquid potentials were needed to reproduce the experimental data and they therefore concluded that, if the solid-fluid potential used in their calculations was not in serious error, "no slit pore geometries are capable of adsorbing the amount of H2 reported by Rodrigues et al." and they stressed the need for independent experimental verification of the data [65].
In the July 2, 1999 issue of Science Chen et al. [66] of the National University of Singapore reported high hydrogen uptake in litium- and potassium-doped multiwalled na-notubes, about 20 and 14 wt. % respectively, under rather mild conditions of atmospheric hydrogen pressure and ambient temperature in case of potassium-doped nanotubes and moderate temperatures (473-673K) in case of lithium-doped nanotubes. Lithium doped graphite and potassium graphite by comparison retained up to 14.0, and 5.0 wt. % hydrogen, respectively, under similar conditions. The carbon nanotubes were made by catalytic decomposition of methane [68, 69], purified to remove the catalyst, and doped by solid state reactions with lithium or potassium containing compounds, eg. carbonates and nitrates. More than 90 % of the samples were multiwalled nano-tubes; the structure was formed by multilayers of graph-ene sheets rolled up in the shape of a hollow cicular cone. Compared to graphite, the structure has much more open edge and greater interplanar spacing, 0.347 nm vs. 0.335 nm in graphite [28, 30], features that favor high hydrogen uptake according to Chen et al. Without doping, however, hydrogen uptake was only 0.4 wt. % suggesting the role of the alkali metals was that of a catalyst. In support of this view in situ infrared data of the lithium doped material were presented. Hydrogen uptake by this sample did not occur at temperatures above 773K, but when cooled to 653K and held for two hours the uptake reached 20 wt. % at 0.1MPA pressure. At the higher temperature the infrared spectrum showed only a weak Li-H vibration. At the lower temperature the Li-H band became more intense and a new band, due to C-H stretching, appeared and grew with the time exposed to hydrogen while the Li-H band did not change with time. Spill over of dissociated atoms from lithium sites to the carbon network of graphene sheets were assumed to eventually bond to carbon atoms; thus the catalytic role of the alkali metal. Hydrogen was recovered from the doped carbon nanotubes by heating and a number of sorption-desorption cycles were in fact reported with little loss of capacity.
The questions raised by Wang and Johnson [45, 65] and discussed above about hydrogen uptake by nanofib-ers reported by the NU group apply of course to this work as well. At the same time, however, the appropriateness of using planar graphite-hydrogen interaction potentials in modelling nanotube-hydrogen interactions may be questioned. The reported infrared data also raise concern about this work , e.g. the band at 1420cm is assigned to the Li-H vibration, but this is not in agreement with the previously reported value of 1280cm [70]. Conflicting results have been obtained in this case also. Professor Fultz of California Institute of Technology, Pasadena, California, U.S.A., reports [71] no hydrogen uptake in room temperature adsorption runs on lithiated KS-44 graphite which Chen et al. claim take up to 14.0 wt. % [66].
Carbon cones, another form of carbon, may also have interesting hydrogen adsorption properties. They were first prepared by condensing carbon vapor on a highly-oriented pyrolytic graphite surface in ultrahigh vacuum [72]. The cones measured up to 24 nm in length, 8 nm in diameter at the base and all had the same cone angle of about 19 degrees although four additional ones are possible [72]. Krishnan et al. [73] more recently described all five types of cones observed in samples prepared by the pyrolysis of hydrogen in a carbon arc using the Kvwrner CB & H industrial process [74-77]. The cone angles observed by Krishnan et al. were 19.2, 38.9, 60.0, 84.6 and 112.9 degrees in agreement with predictions arrived at by considering the symmetry of the graphite sheet and application of Euler's theorem. A Norwegian patent application dealing with the storage of hydrogen in carbon cones, entitled "Lagring av hydrogen i karbonmateriale", has been filed [78]. An international patent application, entitled "Hydrogen storage in carbon material", has also been filed [79].
The nanotubes and carbon cones are derived from the fullerene or "Buckey-ball" structure which itself reacts with hydrogen at high hydrogen pressures (50 - 85 MPa) and temperatures in the range of 573 - 623K to form fullerene hydride containing more than 7 wt. % hydrogen [80]; hydrogen release occurs at 800K indicating relatively strong hydrogen-carbon bonds (chemisorption). One might say that the fullerene - hydrogen system in some respects qualitatively resembles the magnesium - hydrogen system; like magnesium hydride fullerene hydride is too stable to provide reversible hydrogen at usable pressures and room temperature. Attempts to catalyze the reaction and reduce the hydriding/dehydriding temperature has been made using metal- and IC hydrides [81] and a catalyst/solvent combination [82]. These efforts have met with some success. Hydrogen absorption occurs under less severe conditions in the presence of IC metal hydrides, but carbide formation in the metals and ICs may cause decomposition of fullerene at temperatures slighly above 800K [81]. Fullerene in a solvent charged at 453K and 2.4 - 2.7 MPa pressure absorbed hydrogen up to more than 6 wt.% and released hydrogen below 498K in the presence of an Ir-based P-C-P pincer complex catalyst in a solvent [82]. Another approach to destabilizing, i.e. weaken the carbon-hydrogen bonds and reduce the energies required for the absorption/desorption process, might be to modify the electronic structure of fullerene by substitution and/or addition ( cf. "tailormaking" dissociation pressures in metal hydrides by alloying), e.g. fulleride compounds with metals. The presence of the metal may also have a catalytic effect.
SUMMARY AND CONCLUSIONS
A comparative summary of the current state of hydrogen storage technologies, their advantages and disadvantages, and future prospects are shown in Table 2.
Table 2. Comparison of Hydrogen Storage Technologies.
Compressed Gas Liquid Hydrogen
28
Advantages:
Disadvantages:
Good technology base Fair gravimetric and volumetric densities
Cost of compression Bulky
Safety concerns
Good technology base Good gravimetric and volumetric densities
Cost of liquefaction Boil-off losses Safety concerns
Metal Hydride
Excellent volumetric density. High degree of safety
Alloy cost Poor gravimetric densities
Storage in Carbon
Potentially high
gravimetric
densities
Unknown
Future prospects: Goal of 12 wt. % H2
will be met or exceeded
Mature technology
Complex hydrides could improve gravimetric density
Uncertain High risk reseach Possibly high pay-off
It is interesting to note that three hydrogen refuelling stations have opened recently, one at the Munich Airport, Germany [6], another in Hamburg, Germany, and the very latest one at Dearborn, Michigan, USA [54]. All three stations supply compressed and liquid hydrogen are opened to the public. This is a clear sign of anticipation of an expanding automobile market for hydrogen.
The compressed gas option appears to be today's choice among the carmakers, especially for buses, although liquid hydrogen storage and metal hydride storage are all in the running at this time. The DOE"s goal for compressed gas storage will certainly be met and surpassed as the technology matures. The question of safety and public acceptance, however, is in my opinion not settled. Liquid hydrogen storage is a mature technology and any future improvements will be marginal. Safety is perhaps of less concern, but the energy cost of liquefaction is a factor on which little improvement can be expected. Storage of hydrogen in metal hydrides is, from a safety point of view, the preferred storage method. Leak from a punctured storage tank will contain itself, when hydrogen is evolved from the hydride, the storage system cools due to the heat required in the dissociation process and as the system cools, the dissociation pressure decreases and hydrogen evolution stops. The weight penalty is difficult to overcome, but could be partially reduced if research on the complex hydrides pay off. The cost of metal hydride storage alloys is also a factor of concern. Hydrogen storage in carbon materials is an open question with few definite answers at this time. The conflicting experimental reports suggest suttle differences in samples and preparations. To resolve the conflict, synthetic routes to making carbon materials need to be explored and materials characterization must be improved. Theoretical descriptions of hydrogen adsorption in these new forms of carbon must be developed from firm physical concepts and not on risk in terms of results, but at the same time must be seen as having the possibility of high pay-off if viable hydrogen storage materials are discovered.
REFERENCES
[1] Hydrogen & Fuel Cell Letter, March 1999, Vol.XIV/No. 3
[2] Elmsey J., The Elements, 3rd Edition, Oxford University Press, Oxford, UK, 1998.
[3] Webster C., Hydrogen Power Now, 9th Canadian Hydrogen Conference, February 7-10, 1999, Vancouver, Canada.
[4] Mitlitsky F., Vehicular Hydrogen Storage Using Lightweight Tanks, Report UCRL-JC-134540,Lawrence Livermore National Laboratory, Livermore, CA.
[5] Appleby J. A., The Electrochemical Engine for Vehicles, Scientific American, July 1999, pp. 58- 63
[6] Hydrogen & Fuel Cell Letter, June 1999, Vol. XIV/No. 6.
[7] Ibid., September 1998, Vol XIII/No. 9.
[8] Gretz J., The Euro-Quebec Hydro-Hydrogen Pilot Project (EQHHPP), in Hydrogen Power: Theoretical and Engineering Solutions, T. O. Saetre, ed., Kluwer Academic Press, 1998, p. 433.
[9] Maeland A. J. , Hydrogen Storage in Metal Hydrides, Symposium on Hydrogen-Electrochemistry and Energetics, NTNU, Trondheim, Norway, June 6-8, 1999 (Proceedings in Press).
[10] Sandrock G. , Hydrogen - Metal Systems, in Hydrogen Energy Systems, Production and Utilization of Hydrogen and Future Aspects, Y. Yurum, ed. NATO ASI Series, 1995, vol.295, pp. 135 - 166, Kluwer Publishing, Netherland.
[11] Libowitz G. G., Hayes H. F. and Gibb T. R. P., Jr., J. Phys. Chem. 62, (1958) 76
[12] Van VuchtJ. H. N., Kuijpers F. A. and Bruning H. C. A. M., Philips Res. Rep. 25, (1970) 133
[13] Reilly J. J. and Wiswall R. H., Inorg. Chem. 13, (1974) 218
[14] Sandrock G., Suda S. and Schlapbach L., Applications of Hydrides, in Intermetallic Compounds II, in Appl. Phys., 67,(1992) 197
[15] Heung L. K., On-board Hydrogen Storage System Using Metal Hydride, in Hydrogen Power: Theoretical and En-
[16 [17 [18
[19
[20
[21
[22 [23
[24
[25 [26
[27
[28 [29 [30
[31
[32 [33
[34
[35
[36
[37
[38 [39 [40 [41 [42
[43 [44
[45
[46
[47
[48 [49
gineering Solutions, T. O. Saetre, ed., Kluwer Academic [50] Press, 1998, p. 251.
Ames report. [51]
Private com., August 1999. [52]
B. Bogdanowic and M. Schwiickardi, J. Alloys Comp., 1999, vol. 253, p. 119. [53]
Dymova T. N., Dergachev Y. M., Sokolov V. A. and Grecha-naya N. A., Dokl. Akad. Nauk USSR, 1975, vol. 224, p. 591. Dymova T. N., Eliseeva N. G., Bakum S. I. and Dergachev Y. M., Dokl.Akad.Nauk USSR, 1975, vol. 225, p. 1369. [54] Zidan R. A., Takara S., Hee A. G. and Jensen C. M., J. [55] Alloys Comp. 1999, vol. 285, p. 119.
http://www.powerball.net/inside/products/ [56]
Kong V. C. Y., Foulkes F. R., Kirk D. W. and Hinatsu J. T., [57] Int. J. Hydrogen Energy, 1999, vol. 24, p. 665. [58]
Kidnay A. J. and Hiza M. J., Adv. Cryogenic Eng. 1967, vol. 12, p. 730.
Carpetis C. and Peschka W., ibid., 1980, vol. 5, p. 539. Chahine R. and Bose T. K., Int. J. Hydrogen Energy, 1994, vol. 19, p. 161; [59]
Amankwah K. A. G., Noh J. S. and Schwarz J. A., ibid., 1989, vol. 14, p. 437.
Tess A. et al., Science 1996, vol. 273, p. 483. [60]
Cheng H. M., Appl. Phys. Lett., 1998, vol. 72. Journet C. , Maser W. K., Bernier P., Loiseau A., Lamy M. de la Chappelle, Lefrant S., Deniard P., Lee R. and Fischer J. E., Nature, 1997, vol. 388, p. 756. [61]
Bandow S., Asaka S., Sato Y., Rao A. M., Grigorian L., Richter E. and Eklund P. C., Phys. Rev. Lett. 1998, vol. 80, p. 3779. [62] Dillon Mat. Res 1998, vol. 9.
Gregg S. J. and Sing K. S. W., Adsorption, Surface Area [63] and Porosity, Academic Press, London, 1982. [64]
Pedersen M. R. and Broughton J. Q., Phys. Rev. Lett., 1992, vol. 69, p. 2692. [65]
Dujardin E., Ebbesen E., Hiura T. W. and Tanigaki H., Science, 1994, vol. 265, p. 1850. [66] Heben M. J. and Dillon A. C., Hydrogen Storage using Carbon Adsorbents: Past, Present, and Future, 1997 An- [67] nual Report, International Energy Agency (IEA), 1. Nielsen M., McTaque J. P. and Passell L., Neutron Scat- [68] tering of PhysisorbedMonolayers, in Dash J. G. and Ru-valds J., eds., NASIS 1980, vol. 51, Series B: Physics, Ple- [69] num Press, New York, N. Y.
Nielsen M., McTague J. P. and Ellenson W., J. de Phys. , [70] 1997, vol.38, C4-10.
Pace E. L. and Siebert A. R., J. Phys. Chem., 1959, [71] vol. 63, p. 1398. [72]
Dillon A. C., Jones K. M., Bekkedahl T. M., Kiang C. H., Bet- [73] hune D. S. and Heben M. J., Nature, 1997, vol. 386, p. 377. Bethune D. S., Klang C. H., de Vries M. S., Gorman G., Savoy [74] R., Vazquez J. and Beyers R., Nature, 1993, vol. 373, p. 605. Ye Y., Ahn C. C., Witham C., Fulz B., Liu J., Rinzler A. G., [75] Colbert D., Smith K. A. and Smalley R. E., Applied Phys. [76] Lett., 1999, vol. 74, p. 2307.
Stan G. and Cole M. W., J. Low Temp. Phys., 1998, vol. 110, p. 539. [77]
Darkrim F., Malbrunot P., and Levesque D., Hydrogen Storage in Carbon Nanotubes, in Bolcich J. C. and Veziro- [78] glu T. N. eds., Hydrogen Energy Progress XII, International Association for Hydrogen Energy, 1998. [79] Wang Q. and Johnson J. K., J. Chem. Phys., 1999, vol. 110, p. 577. [80] Silvera I. F. and Goldman V. V., J. Chem. Phys., 1978, [81] vol. 69, p. 4209
Wang Q., Johnson J. K. and Boughton J., Q. Mol. Phys., [82] 1996, vol. 89, p. 1105.
Crowell A. D. and Brown J. S., Surf. Sci., 1982, vol. 123, p. 296. [83] Wang Q. and Johnson J. K., Mol. Phys., 1998, vol. 95, p. 299.
Hill S. , Green cars go farther with graphite, New Scientist, 1996, December, vol. 21/28, p. 20. Hydrogen & Fuel Cell Letter, 1997, February, vol. 12, #2. Ahn C. C.,Ye Y. , Ratnakumar B. V., Withman C., Bowman R. C., Jr. and Fultz B., App.Phys. Lett., 1998, vol. 73, p. 3378. Jarvi T. D., Sun J., Conopask L. F. and Satyapal S., Interaction of Hydrogen with Nanoporous Carbon Materials, abstract no. 542, The Electrochemical Society Meeting, Seattle, WA, 1999, May 2-6.
Hydrogen & Fuel Cell Letter, September 1999, vol. XIV, # 9. Baker T. T. K., Storage of Hydrogen in Nanostructures, DOE Hydrogen Program 1998 Annual Peer Review. Hydrogen & Fuel Cell Letter, 1998, July, vol. XIII, #9. United States Patent No. 5,653,951. Anderson P. E., Tan C. D., Crowe A. and Rodriguez N. M., Hydrogen Storage in Carbon Nanostructures, Symposium on Hydrogen-Electrochemistry and Energetics, NTNU, Trondheim, Norway, 1999, June 16-18 (Proceedings in Press).
Ridrigues N.M., Hydrogen Storage of Carbon Nanofibres, 9th Canadian Hydrogen Conference. Vancouver, British Colombia, Canada, 1999, Feb 7-10. Tan C. D., Anderson P. E. and Rodriguez N. M., Interactions of Hydrogen with GraphiteNanofibers, abstract no. 541, The Electrochemical Society Meeting, Seattle, WA, 1999, May 2-6.
Chambers A., Park C., Baker R. T. K. and Rodriguez N. M., J. Phys. Chem., 1998, vol. 102, p. 4253. Kim M. S., Rodriguez N. M. and Baker R. T. K., J. Catal., 1991, vol. 131, p. 60.
Rodriguez N. M., Mater J. Res., 1993, vol. 8, p. 3233. Krishnankutty N., Rodriguez N. M. and Baker R. T. K., J. Catal. Today, 1997, vol. 37, p. 295. Wang Q. and Johnson J. K., J. Phys. Chem. B, 1999, vol. 1103, p. 276.
Chen P. , Wu X., Lin J. and Tan K. L., Science, 1999, vol. 285, p. 91.
Wu X.B., Chen P., Lin J. and Tau K.L., Int. J Hydrogen Energy, in press.
Chen P., Zhang H. B., Lin G. D., Hong Q. and Tsai K. R., Carbon, 1997, vol. 35, p. 1495.
Chen P., Wu X., Sun X., Lin J. and Tan K. L., Phys. Rev. Lett., 1999, vol. 82, p. 2548.
Smyrl N. R., Fuller E. L., Jr. and Powell G. L., Appl. Spec-trosc. 1983, vol. 37, p. 38. Fultz B., private com. August 8, 1999. Ge and K. Sattler, Chem. Phys. Lett., 1994, vol. 220, p. 192. Krishnan A., Dujardin E., Treacy M. J., Hugdal J.,Lynum S. and Ebbesen T.W., Nature, 1997, vol. 388, p. 451. Lynum S., 5th Annual US Hydrogen Meeting, National Hydrogen Assocociation, Washington, D. C. 1994. Gaudernack B. and Lynum S., in Proc. 11th WHEC, vol. 1, p. 511. Lynum S., Hugdahl J. and Hildrum R., The Kvrnner Process, paper presented at Carbon Black World, Nice, France, March 4-6. 1996.
Hildrum R. and Norvik, Norwegian patent application no. 1997, 1413, March 25, 1997. Maeland A. J. and Skjeltorp A. T., Norwegian patent application no. 980530, February 12, 1998. Maeland A. J. and Skjeltorp A. T., International patent application no. PCT/NO99/00040. Baum R. M., Chem. Eng. News, Nov. 22, 1993, p. 8. Tarasov B. P., Fokin V. N., Moravsky A. P., Shul'ga Yu. M. and Yartys V. A., J. Alloys Comp., 1997, vol. 253-254, p. 25. Murphy R. W., Hydrogen Storage: the Fullerene Option, DOE Hydrogen Program 1999 Annual Peer Review. Kowalska E., Byszewski P., Dluzewski P., Diduszko R. and Kucharski Z., J. Alloys Comp, 1999, vol. 286, p. 297.
29