Implementing a hydrogen economy

by James A. Ritter1*, Armin D. Ebner1, Jun Wang1, and Ragaiy Zidan2

1Department of Chemical Engineering,Swearingen Engineering Center,University of South Carolina,Columbia, SC 29208, USAE-mail: [email protected]

2Westinghouse Savannah River CompanySavannah River Technology CenterAiken, SC 29804, USA

*Corresponding author: James A. Ritter

President Bush, during his State of the Union Address

this year, pronounced a $1.2 billion jump-start to the

hydrogen economy. The move would represent not

only freedom from US-dependence on foreign oil,

which is a national security issue, but also a

necessary and gargantuan step toward improving the

environment by reducing the amount of carbon

dioxide released into the atmosphere. However,

hydrogen storage is proving to be one of the most

important issues and potentially biggest roadblock for

the implementation of a hydrogen economy. Of the

three options that exist for storing hydrogen, in a

solid, liquid, or gaseous state, the former is becoming

accepted as the only method potentially able to meet

the gravimetric and volumetric densities of the

recently announced FreedomCar goals; and of all

known hydrogen storage materials, complex hydrides

may be the only hope.

In recent years, months, weeks, and even days, it has

become increasingly clear that hydrogen as an energy

carrier is ‘in’ and carbonaceous fuels are ‘out’1. The

hydrogen economy is coming, with the impetus to

transform our fossil energy-based society, which

inevitably will cease to exist, into a renewable

energy-based one2. However, this transformation will

not occur overnight. It may take several decades to

realize a hydrogen economy. In the meantime,

research and development is necessary to ensure that

the implementation of the hydrogen economy is

completely seamless, with essentially no disruption

of the day-to-day activities of the global economy.

The world has taken on a monumental, but not

insurmountable, task of transforming from

carbonaceous to renewable fuels, with clean burning,

carbon dioxide-free hydrogen as the logical choice.

One of the key roadblocks to the widespread use of

hydrogen as a renewable fuel, especially on-board passenger

vehicles, is hydrogen storage3. Three options exist for storing

hydrogen: as a highly compressed gas, a cryogenic liquid, or

in a solid matrix. However, because hydrogen is the lightest

element known, the gravimetric and volumetric densities of

any storage system will necessarily be very low, especially

compared to carbonaceous fuels. To put these storage

options into perspective, Fig. 1 compares the gravimetric and

volumetric densities, and corresponding specific energies and

energy densities, of a variety of different hydrogen storage

media with two common fossil fuels, gasoline and propane.

Implementinga hydrogen economy

September 200318 ISSN:1369 7021 © Elsevier Ltd 2003

REVIEW FEATURE

Is this as good as it gets?Yes, which is obvious from Fig. 1. Hydrogen as an energy

carrier will never be comparable to carbonaceous materials,

but it does not have to be. For transportation, 4-6 kg of

hydrogen needs to be stored in a mid-sized, fuel cell-powered

vehicle for a range of 500 km7. This does not seem like much,

but at atmospheric pressure (1 atm) and ambient

temperature (298 K), 5 kg of hydrogen occupies 60 m3, a

viable but not practical option for on-board vehicles. In

contrast, 5 kg of gaseous hydrogen at ambient temperature

in a 5000 psi composite tank, which is being commercialized

for on-board vehicles, needs an internal volume of 0.2 m3.

This corresponds to four or five of these high-pressure tanks7,

a feasible but still not desirable option for on-board vehicles.

Cryogenic hydrogen, with a liquid density of 70.8 kg/m3,

seems the best option from a volumetric point of view, as

5 kg occupies an internal volume of only 0.07 m3. A specially-

designed, super-insulated tank for cryogenic hydrogen is

currently under commercialization.

Most automakers are considering either the high-pressure

gaseous or cryogenic liquid hydrogen storage options for

passenger vehicles. However, these two technologies are

fraught with public perception issues on safety. Other issues

also need to be addressed, including compression costs and

safety, liquefaction costs, and dormancy. Fig. 1 shows that

neither of these storage options meet the long-term needs of

automakers and barely those in the near-term. For this

reason, it is becoming increasingly accepted that the solid

matrix method of hydrogen storage is the only option that

has any hope of achieving the gravimetric and volumetric

densities identified in Fig. 1 for the FreedomCar goals.

Can limitations be surpassed?Maybe, but this is not obvious from Fig. 1. The gravimetric

and volumetric hydrogen storage densities of many materials

shown meet or exceed the FreedomCar goals. However, the

best hydrogen storage materials from an energy content

point of view are carbon-based fossil fuels, like methane.

These fuels are not an option for a hydrogen-based economy.

Metal hydrides, such as MgH2, are arguably the next best

class of hydrogen storage material. However, despite more

than three decades of research and modification of the

properties of thousands of metal hydrides8, it can be

concluded that, as a class, they are limited to a reversible

hydrogen storage capacity of 2 mass% at conditions suitable

for on-board vehicle storage (<373 K and 2 atm)9. This does

not mean that metal hydrides are not a viable storage

material for vehicle applications; on the contrary, some have

reasonable volumetric densities and other physical and

chemical properties, including sustained cyclability, that make

them an excellent medium for niche transportation markets

able to handle the low gravimetric densities, such as buses10.

Carbon-based materials, such as single- and multiwalled

nanotubes, fibers, and activated carbons have been widely

September 2003 19

Fig. 1 Gravimetric and volumetric densities, and corresponding specific energies and energy densities, of a variety of different hydrogen storage media, including two common fossil fuels,gasoline, and propane. The FreedomCar targets4 for 2005, 2010, and 2015 are indicated within the shaded regions. The gravimetric hydrogen densities for pressurized hydrogen gas5 andcryogenic liquid hydrogen6 (diamond symbols) include the mass of the storage container, whereas the values reported for the metal hydrides, complex hydrides, and other hydrogenstorage materials are based on the absolute (theoretical) amount of hydrogen. [Adapted from Schlapbach and Züttel7. © Nature Publishing Group 2001.]

studied as hydrogen storage materials11. Although fantastic

claims about the storage capacity of some of these materials

have been reported, most are still controversial because they

cannot be easily reproduced12. Recent work by Zhou et al.13

puts most of these materials into perspective and essentially

dismisses their use for on-board vehicle applications because

of their low gravimetric hydrogen storage density.

Chemical or complex hydrides, such as NaAlH4 and NaBH4,

are relatively new to the transportation market as they were

thought to release hydrogen irreversibly at conditions outside

the desirable temperature and pressure ranges. This

‘misconception’ changed abruptly with two technological

breakthroughs. One was a 2003 US patent14 in which

Millennium Cell Inc. claimed that NaBH4, when combined

with a catalyst in an aqueous solution above pH 7, releases

hydrogen at conditions suitable for use in passenger vehicles.

As a result, one major automaker is pursuing limited

commercialization of this technology. However, as exciting as

this new hydrogen storage technology may appear, one major

drawback is the prohibitively high cost of off-board

reprocessing of the resulting borax solution back to NaBH4.

Nor is it not clear whether the gravimetric density of the

Millennium Cell system meets any of the FreedomCar goals,

especially those for 2015.

The second breakthrough with complex hydrides does not

require off-board processing and allows hydrogen to be

stored reversibly in a manner akin to metal hydrides. Based

on the pioneering work of Bogdanovic and Schwickardi15, it

was shown that the addition of a catalyst or dopant to the

complex hydride NaAlH4 makes it reversibly release and take

up 3.7 mass% of hydrogen at conditions close to those

needed for on-board vehicle storage. Since then, a host of

results have appeared in the literature on metal doped

NaAlH4 because it has a theoretical hydrogen capacity of

5.6 mass%. Although it is now accepted that this material

will never meet the FreedomCar goals, exhibiting only about

3 mass% reversible hydrogen capacity at <100°C16, it

represents a model compound that is reversible when doped

with Ti and other additives. In fact, many other complex

hydrides contain far greater amounts of hydrogen, e.g. up to

18.2 mass% in LiBH4. Some of the common complex hydrides

are listed in Table 1. The main question to be addressed is:

Will Ti and other additives make some of these other

complex hydrides reversible? Hopefully, for complex hydrides

are the only known materials that have the potential to meet

the FreedonCar goals, especially those for 2015! Complex

hydrides, the new kid on the block for on-board hydrogen

storage, may be the only kid on the block!

Who is doing the research?Many, mainly in five countries at various universities, national

laboratories, and private companies. Bogdanovic15-18

continues to research NaAlH4 at the Max-Planck-Institut für

Kohlenforschung; while Fichtner and coworkers at the

Forshcungszentrum Karlsruhe19,20 are studying Mg(AlH4)2 in

collaboration with Ritter and Zidan at the University of South

Carolina (USC) and the Savannah River Technology Center

(SRTC)21. NaAlH4 is also under intense study at the

University of Hawaii (UH) by Jensen22-24, in collaboration

with Takara at UH, Gross and Thomas at Sandia National

Laboratories25-28, and Sandrock at SunaTech Inc.29-31. Meisner

and Balogh at the General Motors Research and Development

Center32,33 are also studying NaAlH4, as are Anton and

coworkers at the United Technologies Research Center34 and

Ritter35,36. At McGill University, Zaluska is investigating

Li3AlH6, Na3AlH6, and combinations of Li and Na with the

AlH6-3 complex37,38, in addition to NaAlH4. Pecharsky at the

Ames Laboratory and Iowa State University39-42 is exploring

LiAlH4 and Li3AlH6, as is Ritter36. Chen is studying LiAlH4 and

Li3AlH6 at the National Institute of Advanced Industrial

Science and Technology43,44 in Japan. LiBH4 is the subject of

Züttel’s work45,46 at the University of Fribourg, while

Morioka47 at Sony in Japan is studying the KAlH4.

Clearly, only a paucity of the complex hydrides listed in

Table 1 are currently being studied, essentially those that are

REVIEW FEATURE

September 200320

Table 1 Common complex hydrides for hydrogen storage applications8.

Hydride Mass% hydrogena Availability

KAlH4 5.8 J. Alloys Compd. (2003) 335533, 310

LiAlH4 10.6 Commercially available

LiBH4 18.5 Commercially available

Al(BH4)3 16.9 J. Am. Chem. Soc. (1953) 7755, 209

LiAlH2(BH4)2 15.3 British Patents 840 572, 863 491

Mg(AlH4)2 9.3 Inorg. Chem. (1970) 99, 325

Mg(BH4)2 14.9 Inorg. Chem. (1972) 1111, 929

Ca(AlH4)2 7.9 J. Inorg. Nucl. Chem. (1955) 11, 317

NaAlH4 7.5 Commercially available

NaBH4 10.6 Commercially available

Ti(BH4)3 13.1 J. Am. Chem. Soc. (1949) 7711, 2488

Zr(BH4)3 8.9 J. Am. Chem. Soc. (1949) 7711, 2488

a Mass% of hydrogen in each molecule is based on theoretical maximum.

REVIEW FEATURE

commercially available19-21. It is disconcerting that so far

only NaAlH4 has been shown to be reversible; few attempts

have been made to determine the conditions for reversibility

with other complex hydrides, except the LiAlH4 system43.

What progress is being made?Much, based on Bogdanovic’s work15 on metal doped NaAlH4,

which showed that Ti-doped NaAlH4 can decompose at

80-85°C lower than the pure material:

NaAlH4 ↔ 1/3Na3AlH6 + 2/3Al + H2 (1)

Na3AlH6 ↔ 3NaH + Al + 3/2H2 (2)

Although the first reaction generates only 3.7 mass%

hydrogen at most and the second is limited to 1.9 mass%

(a total of 5.6 mass% compared with 7.5 mass% in Table 1),

this generated a great deal of excitement in the hydrogen

storage community. For the first time, a complex hydride

exhibited reversible hydrogen storage properties and

reasonable capacities (similar to those of metal hydrides) at

conditions commensurate with on-board storage and

regeneration. Most researchers thought this was impossible.

The next breakthrough with complex hydrides was made

by Jensen, Zidan, and coworkers22,23. They showed that the

decomposition temperature of Ti-doped NaAlH4 can be

further lowered using a high energy ball milling process to

prepare the doped sample. Not only did this bring the

conditions closer to those desired for on-board hydrogen

storage, it also ignited the metal hydride research community

to look at this novel, reversible complex hydride.

Another important discovery was the effect of doping

NaAlH4 with Ti-Fe17,35,36 or Ti-Zr23,35,36, which lowers

decomposition temperatures compared with using Ti, Zr, or

Fe alone, as shown in Fig. 235,36. Bogdanovic et al.16 and,

most recently, we have shown that capacity losses from

cycling can be alleviated by the addition of Al powder to the

Ti-doped sample36. We have also shown an improvement in

the decomposition rate and temperature using a proprietary

additive with Ti-doped NaAlH4; the additive works even when

Al is added to minimize cycling losses, as shown in Fig. 336.

Other results are piecing together the mechanistic role of

the metal dopant in promoting the thermodynamic and

kinetic uptake and release of hydrogen by NaAlH4. For

example, X-ray diffraction (XRD) and solid-state NMR studies

show that the loss of hydrogen capacity upon cycling is the

result of unreacted Al preventing reaction 1 to reverse

completely16. In addition, studies using in situ XRD, scanning

electron microscopy, energy dispersive spectroscopy, and

Auger spectra analysis show that most of the metal dopant

remains at the sample surface24-28,30. Long-range transport

of the metal species during the dehydriding process has also

been discovered, where Al is possibly being transported to

and segregating at the surface during rehydrogenation, with

speculation that the transported species may be a more

mobile hydride, such as AlH325-27. In situ XRD studies reveal

that the doping process alters certain lattice parameters,

inducing distortion that reaches a maximum when NaAlH4 is

doped with 2 mol% Ti24. This not only provides crucial

September 2003 21

Fig. 2 Constant temperature (90, 110, and 130°C) and varying temperature 0th cyclethermogravimetric (2°C/min) dehydrogenations of NaAlH4 doped with 4 mol% Ti, Ti-Zr,

Ti-Fe, Ti-Zr-Fe, and Ti-Fe, showing the synergistic behavior of Ti-Fe and Ti-Zr over Tialone17,23,35,36.

Fig. 3 Varying temperature thermogravimetric (2°C/min) dehydrogenations of NaAlH4

doped with 2 mol% Ti and 5 wt% Al powder with and without 10 wt% proprietary additive.Initially (0th cycle) and after five rehydrogenation cycles (at 150°C and 100 atm for 2 hrs),the Al powder eliminates cycle losses, while the proprietary additive improves thedehydrogenation conditions36. The insert reveals the effect of the proprietary additivewhen added to NaAlH4 doped with 2 mol% Ti in 5 wt% and 10 wt% amounts during 0th

cycle constant temperature (90°C) thermogravimetric dehydrogenations36.

evidence on the role of Ti in improving hydrogen transport

via lattice distortion, but also indicates that 2 mol% Ti may

be close to the optimum dopant level. X-ray photoelectron

spectroscopy studies of Ti-doped Li3AlH6 reveal that the

surface Ti contains Ti0/Ti2+/Ti3+ species43. This indicates that

Ti0 ↔ Ti3+(Ti0/Ti2+/Ti3+) defect site chemistry may play a

role in enhancing the system’s reversibility. Whether this

applies to NaAlH4 is not known, however. To assist in the

search for the elusive mechanistic role of the metal dopant,

we are carrying out molecular orbital calculations for the

LiAlH4 and NaAlH4 systems; the total electron density map of

LiAlH4 and the structure of NaAlH4 are depicted in Fig. 436.

What are the limitations and issues?Ample, for although exciting discoveries are being made with

some complex hydrides, especially Ti-doped NaAlH4, and

light is being shed on the role of Ti, a detailed mechanism

based on definitive experimental evidence, is still lacking.

Even Ti-doped NaAlH4 with Al and the proprietary additive

only provides 3 mass% reversible hydrogen at reasonable

rates when discharged at >100°C. These limitations of

Ti-doped NaAlH4 are shown in Fig. 528,36. Although the

additive improves the kinetics, the capacity is still too low for

on-board passenger vehicles. Moreover, as shown in Fig. 6,

pressures of 1500 psig are required to achieve reasonable

rates during rehydrogenation. Even so, this is still not fast

enough for on-board refueling, unless temperatures exceed

110°C36. Despite these limitations, Ti-doped NaAlH4 is one of

the best hydrogen storage materials known, surpassing all

metal hydrides in release capacity8,9.

Have the kinetic and thermodynamic limitations of

Ti-doped NaAlH4 been reached? Maybe! But all the evidence

is not in, especially concerning the roles of the metal and

other additives. More research is needed, even on metal

doped NaAlH4. In the meantime, the knowledge gained from

this system must be applied to other complex hydrides with

higher hydrogen capacities to determine if reversibility at

reasonable conditions is possible.

One key issue is the lack of availability of some of the

most attractive complex hydrides. Except for a select few

(see Table 1), most complex hydrides have to be synthesized

in-house. Some progress is being made, however, by

ourselves21,36 and others43 in the painstaking search for

conditions that make other complex hydrides reversible.

What may complex hydrides bring?Excitement, for if the recent past can be used to judge the

future, then the outlook for complex hydrides is very bright.

In the past few years, studies with various complex hydrides,

REVIEW FEATURE

September 200322

Fig. 5 Thermodynamic and kinetic limitations of the metal doped NaAlH4 system: (a)

thermodynamic Van't Hoff plots (upper graph) show the equilibrium desorption plateaupressure as a function of temperature for reactions 1 and 2 with NaAlH4 doped with

2 mol% each of Ti and Zr28; and (b) kinetic Arrhenius plots (lower graph) show thedesorption rate as a function of temperature for reactions 1 and 2 with NaAlH4 undoped,

doped with 4 mol% Ti, and doped with 4 mol% Ti and 10 wt% proprietary additive31,36. [© Elsevier Ltd. 2002.]

Fig. 4 Depictions of the CASTEP total electron density map (iso-surface = 0.1451) ofLiAlH4 (upper image) and the first CASTEP optimized solution-space group I41/a of

NaAlH4 (lower image) obtained from molecular orbital calculations36.

REVIEW FEATURE

especially NaAlH4, have generated important insights into

the hydrogen storage properties of these materials. Ti as a

dopant or catalyst and high energy ball milling, as well as the

synergistic effect of Ti and Fe or Ti and Zr, have been shown

to play key roles in enhancing the conditions for reversibility.

This is the good news for the viability of metal doped NaAlH4

as a storage material. It clearly has a place among the

technologies being considered for stationary hydrogen

storage applications. However, its reversible hydrogen

capacity of 3 mass% and relatively slow release/uptake rates

are simply not good enough for passenger vehicles.

In the next few years it will become clear whether the

exciting results from metal doped NaAlH4 carry over to other

complex hydrides. Limited evidence exists that suggests this

may be the case, with Ti-doped LiAlH436,43 and Mg(AlH4)2

21

both showing promise. It will be disappointing if complex

hydrides do not play a key role in the future of on-board

hydrogen storage. But more research needs to be done to

solve the problem in a reasonable amount of time. Only time

will tell if complex hydrides prove to be the only kid on the

block, if they are even on the block at all! MT

September 2003 23

Fig. 6 Constant temperature (100°C and 110°C) 5th cycle volumetric discharge (ordehydrogenation, upper graph) and charge (or rehydrogenation, lower graph) ratelimitations for NaAlH4 doped with 4 mol% Ti, 5 wt% Al powder, and 10 wt% proprietary

additive36.

REFERENCES

1. Goltsov, V. A., and Veziroglu, T. N., Int. J. Hydrogen Energy (2001) 2244, 909

2. Ogden, J. M., Annu. Rev. Energy Environ. (1999) 2244, 227

3. National Hydrogen Energy Roadmap, US Department of Energy (2002)

4. Milliken, J., US Department of Energy (2003) private communication

5. Thomas, G., and Keller, J., Hydrogen Delivery and Infrastructure Workshop(2003)

6. Pettersson, J., and Hjortsberg, O., KFB-Meddelande (1999), 27

7. Schlapbach, L., and Züttel, A., Nature (2001) 441144, 353

8. Sandrock, G., and Thomas, G., Appl. Phys. A (2001) 7722, 153

9. Sandrock, G., J. Alloys Compd. (1999) 229933--229955, 877

10. Heung, L. K., Hypothesis II (1997) 1

11. Dillon, A. C., and Heben, M. J., Appl. Phys. A (2001) 7722, 133

12. Dagani, R., Chem. Eng. News (2002) 8800 (2) 25

13. Zhou, L., et al., Int. J. Hydrogen Energy (2003) in press

14. Amendola, S. C., et al., Process for synthesizing borohydride compounds, US Patent 6,524,542 (2003)

15. Bogdanovic, B., and Schwickardi, M., J. Alloys Compd. (1997) 225533, 1

16. Bogdanovic, B., et al., J. Alloys Compd. (2003) 335500, 246

17. Bogdanovic, B., et al., J. Alloys Compd. (2000) 330022, 36

18. Bogdanovic, B., and Schwickardi, M., Appl. Phys. A (2001) 7722, 221

19. Fichtner, M., and Fuhr, O., J. Alloys Compd. (2002) 334455, 286

20. Fichtner, M., et al., J. Alloys Compd. (2003) in press

21. Fichtner, M., et al., University of South Carolina (2003), unpublished results

22. Jensen, C. M., et al., Int. J. Hydrogen Energy (1999) 2244, 461

23. Zidan, R., et al., J. Alloys Compd. (1999) 228855, 119

24. Sun, D., et al., J. Alloys Compd. (2002) 333377, L8

25. Gross, K. J., et al., J. Alloys Compd. (2000) 229977, 270

26. Jensen, C., and Gross, K., Appl. Phys. A (2001) 7722, 213

27. Thomas, G., et al., J. Alloys Compd. (2002) 333300--333322, 702

28. Gross, K. J, et al., J. Alloys Compd. (2002) 333300--333322, 683

29. Sandrock, G., et al., J. Alloys Compd. (2002) 333300--333322, 696

30. Gross, K., et al., J. Alloys Compd. (2002) 333300--333322, 691

31. Sandrock, G., et al., J. Alloys Compd. (2002) 333399, 299

32. Meisner, G. P., et al., J. Alloys Compd. (2002) 333377, 254

33. Balogh, M. P., et al., J. Alloys Compd. (2003) 335500, 136

34. Anton, D. L., J. Alloys Compd. (2003) in press

35. Wang, J., et al., Adsorption Science and Technology, Lee, C. H., (ed.), WorldScientific, Korea (2003)

36. Wang, J., et al., University of South Carolina (2003) unpublished results

37. Zaluski, L., et al., J. Alloys Compd. (1999) 229900, 71

38. Zaluska, A., et al., J. Alloys Compd. (2000) 229988, 125

39. Balema, V. P., et al., Chem. Commun. (2000) 1177, 1665

40. Balema, V. P., et al., J. Alloys Compd. (2000) 331133, 69

41. Balema, V. P., et al., J. Alloys Compd. (2001) 332299, 108

42. Pecharsky, V. K., and Balema, V. P., Fundamentals of Advanced Materials forEnergy Conversion. In Proceedings of EPD Congress 2002, Taylor, P. R., et al.(eds.) TMS, Warrendale, PA (2002)

43. Chen, J., et al., J. Phys. Chem. B (2001) 110055, 11214

44. Chen, J., et al., Adv. Eng. Mater. (2001) 33 (9), 695

45. Züttel, A., et al., J. Power Sources (2003) 111188, 1

46. Züttel, A., et al., J. Alloys Compd. (2003) in press

47. Morioka, H., et al., J. Alloys Compd. (2003) 335533, 310