Phil. Trans. R. Soc. A (2010) 368, 3329–3342doi:10.1098/rsta.2010.0113

REVIEW

Hydrogen: the future energy carrierBY ANDREAS ZÜTTEL*, ARNDT REMHOF, ANDREAS BORGSCHULTE

AND OLIVER FRIEDRICHS

Empa Materials Sciences and Technology, Department of Environment,Energy and Mobility, Division of Hydrogen & Energy,

CH-8600 Dübendorf, Switzerland

Since the beginning of the twenty-first century the limitations of the fossil age with regardto the continuing growth of energy demand, the peaking mining rate of oil, the growingimpact of CO2 emissions on the environment and the dependency of the economy inthe industrialized world on the availability of fossil fuels became very obvious. A majorchange in the energy economy from fossil energy carriers to renewable energy fluxes isnecessary. The main challenge is to efficiently convert renewable energy into electricityand the storage of electricity or the production of a synthetic fuel. Hydrogen is producedfrom water by electricity through an electrolyser. The storage of hydrogen in its molecularor atomic form is a materials challenge. Some hydrides are known to exhibit a hydrogendensity comparable to oil; however, these hydrides require a sophisticated storage system.The system energy density is significantly smaller than the energy density of fossil fuels.An interesting alternative to the direct storage of hydrogen are synthetic hydrocarbonsproduced from hydrogen and CO2 extracted from the atmosphere. They are CO2 neutraland stored like fossil fuels. Conventional combustion engines and turbines can be used inorder to convert the stored energy into work and heat.

Keywords: energy carriers; hydrogen; synthetic fuels; efficiency

1. Energy and development

(a) Growing energy demand

Human beings developed on Earth based on plants, i.e. biomass, as the onlyenergy carrier. The consumption of plants did not change the environmentbecause the carbon dioxide that was liberated by the humans and animalswas reabsorbed by the plants during the process of photosynthesis. In 1698,Thomas Savery built a simple steam engine able to pump water. ThomasNewcomen in 1712 developed an improved steam engine (Wilson 1981) ableto deliver mechanical work. For the first time, a non-living machine worked

*Author for correspondence (andreas.zuettel@empa.ch).

One contribution of 13 to a Discussion Meeting Issue ‘Energy materials to combat climate change’.

This journal is © 2010 The Royal Society3329

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3330 A. Züttel et al.

120

steamengine coal

oil

USAoil peak

VW

End2nd World

War

wor

ld e

nerg

y de

man

d (1

012 k

Wh

yr–1

)100

80

60

40

20

01800 1850 1900 1950 2000

year

Figure 1. Energy demand according to the energy carriers over the last 200 years (J.-M. Martin-Amouroux, private communication 2002). Orange, renewables; red, nuclear fission; blue, hydropower; grey, natural gas; dark grey, crude oil; black, coal; green, biomass.

for human beings. However, the steam engines of Savery and Newcomer wereneither very powerful nor energy efficient. James Watt significantly improved thesteam engine and, because the energy efficiency was increased by a factor of 4, thesteam engine was widely introduced for the conversion of heat into mechanicalwork and enabled the start of industrialization (figure 1).

The energy for the steam engine was found in the form of mineral coal.The world energy consumption increased from 5 × 1012 kWh yr−1 in 1860 to1.2 × 1014 kWh yr−1 today. Approximately 1.0 × 1014 kWh yr−1 (80%) is based onfossil fuels (coal, oil and gas). The population of human beings increased duringthe twentieth century by a factor of 6, but the energy consumption increasedby a factor of 80 (Eidgenössischen Technischen Hochschule Zürich 2000). Theworldwide average continuous power consumption today is 2 kW per capita.

(b) The efficiency paradox

The invention and development of the steam engine is an excellent historicalexample of the phenomenon that an increase in the efficiency of an engineleads to an increase in the overall energy consumption. In economics, theJevons Paradox is an observation made by William Stanley Jevons that, astechnological improvements increase the efficiency with which a resource isused, the total consumption of that resource tends to increase, rather thandecrease. Increased energy efficiency tends to increase energy consumption bytwo means. First, increased energy efficiency makes the use of energy relativelycheaper, thus encouraging increased use (the rebound effect). Second, increasedenergy efficiency leads to increased economic growth, which increases energy usefor the whole economy (Saunders 1992).

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3331

(c) Limited resources of fossil fuels

Most geologists view crude oil, like coal and natural gas, as the productof compression and heating of ancient vegetation over geological time scales(biogenic theory). According to this theory, it is formed from the decayedremains of prehistoric marine animals and plants. Three conditions must bepresent for oil reservoirs to form: a rich source rock, a migration conduitand a trap (seal) that forms the reservoir. The oil reserves on Earth arefinite and will, therefore, not last for ever. The Hubbert peak theory (Hubbert1956), also known as peak oil, is a theory concerning the long-term rate ofproduction of conventional oil and other fossil fuels. It assumes that oil reservesare not replenishable (i.e. that abiogenic replenishment, if it exists at all,is negligible), and predicts that future world oil production must inevitablyreach a peak and then decline as these reserves are exhausted. For variousreasons (perhaps most importantly, the lack of transparency in accounting ofglobal oil reserves), it is difficult to predict the oil peak in any given region.On the basis of available production data, proponents have previously (andincorrectly) predicted the peak for the world to be in years 1989, 1995, or1995–2000. However, these predictions date from before the recession of theearly 1980s, and the consequent reduction in global consumption, the effect ofwhich was to delay the date of any peak by several years. A new prediction byGoldman Sachs (Engdahl 2008) gives 2007 for oil and some time later for naturalgas. Just as the 1971 US peak in oil production was only clearly recognizedafter the fact, a peak in world production will be difficult to discern untilproduction clearly drops off.

(d) Environmental impact

The consumption of fossil fuels together with deforestation leads to theliberation of 7 × 1012 kg yr−1 of carbon in the form of CO2. The two major naturalsinks of carbon dioxide are plants, which absorb an additional 2 × 1012 kg yr−1

carbon by the photosynthesis process, and oceans, in which the same amount ofcarbon is dissolved. Therefore, the net increase in carbon in the form of CO2 inthe atmosphere owing to human activities is approximately 3 × 1012 kg yr−1. Thiscorresponds to an annual increase of 0.4 per cent of the CO2 concentration inthe atmosphere.

Svante Arrhenius (1896) determined that a doubling of CO2 in the atmospherewould lead to an average temperature increase of 5–6◦C. According to basicscience, observational sensitivity studies (Petit et al. 1999) and the climatemodels referenced by the IPCC (Working Group I 2007), temperatures mayincrease by 1.4–5.8◦C between 1990 and 2100. This is expected to resultin other climate changes, including rises in sea level and changes in theamount and pattern of precipitation. Such changes may increase the frequencyand intensity of extreme weather events such as floods, droughts, heatwaves and hurricanes; these changes themselves cause changes in agriculturalyield, glacier retreat, reduced summer streamflows or contribute to biologicalextinctions.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3332 A. Züttel et al.

150

125

100

75

50

25

0

1945

1950

1955

1960

1965

1970

1975

1980

1985

date (yr)

1990

1995

2000

2005

2010

pric

e ($

bar

rel–

1 )

End2nd World

War

VW

founded

1864

oildiscovered

Iranrevolution Iraq war

Persiangulf war

economic crisis

limited reserves

USoil peak

Figure 2. Development of the oil price without correction for inflation (http://www.ioga.com/Special/crudeoil_Hist.htm).

(e) Economic dependency

The demand for fossil fuels has a strong impact on social, political and economicinteractions between the various countries. For example, two-thirds of the crudeoil reserves are located in the Middle East region, but most of the petrol isconsumed in the USA, Europe and Japan. Fossil energy has a very stronginfluence on the development of a society and the standard of living. There isa dependency of the gross national product on the average amount of energyconsumed per capita. This leads to the conclusion that today the economic gainof the industrialized world is, to a significant extent, correlated to the energyconsumption from fossil fuels. Furthermore, economic growth directly dependson the availability of fossil fuels, and the recent steep increase in the oil pricecoincided with a worldwide economic crisis (figure 2).

In summary, the four major challenges for the future energy economy are:

— Energy demand: increase in the world energy demand owing to theindustrialization of the largest countries (J.-M. Martin-Amouroux, privatecommunication 2002; Schilling et al. 1977).

— Limited resources: the rate of discovered oil has already decreased since1970 but the mining rate has increased. Peak oil is expected to take placebetween 2008 and 2012 (Campbell 2003).

— Climate change: the average temperature on Earth has increased owingto the increase in the CO2 concentration in the atmosphere (Mannet al. 1999).

— Economic dependency: energy per capita versus gross domestic product percapita (http://muller.lbl.gov/teaching/physics10/PffP_textbook/PffP-10-climate_files/image022.gif) grows linearly in the beginning of industriali-zation and then saturates.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3333

Fossil fuels are naturally occurring energy carriers and are, therefore, in aneconomic context called primary energy sources. In the physical context, the onlyprimary energy resources are solar light (heat, precipitation, wind), geothermalheat and the tides. These sources deliver energy fluxes (power flow) that varyaccording to the day/night cycle and the seasonal cycles and have an underlyingspatial distribution on Earth. Therefore, natural energy fluxes have to beprocessed for residential, industrial and mobile energy needs by:

— conversion into a usable form of energy, i.e. heat and work (electricity),— storage in a synthetic energy carrier, e.g. hydrogen, and— transport to the consumer and for mobility.

This involves a technical process and, therefore, an investment. The economicsystem that is based on fossil energy has to be changed while fossil fuels arestill available. Otherwise, the investments necessary for the change to renewableenergy supply will no longer be possible. In order to change from fossil torenewable energy, the world economy has to be converted to a sustainableeconomy based on investments instead of mining resources. Renewable energyrequires, in general, an investment of approximately the energy produced in5 years. This means an investment of about five times the annual energycost. While the investment in energy is defined by technology and is thereforeconstant, the financial investment directly depends on the cost of the fossilfuels. The investment corresponds today to approximately 25 t of oil in a highlyindustrialized country. Each person, e.g. in Europe, has to invest about 250 barrelsof crude oil. This depends on the price of fossil fuels but is between 20 and 50per cent of the annual income in Europe.

2. Hydrogen

(a) Hydrogen cycle

The natural cycles in the atmosphere are able to transport only a few compounds,namely oxygen, nitrogen, water and CO2. Therefore, the combustion product ofany kind of fuel has to be water, CO2 or nitrogen. Hydrogen is a fuel that producesonly water as a combustion product.

The hydrogen cycle is analogous to the natural carbon cycle, except that nocarbon is involved. The advantage is that, in contrast to biomass, no living matteris necessary for the production of hydrogen, but the disadvantage is that hydrogenis a gas at ambient conditions.

Electricity from a renewable energy source is used for the electrolysis of water.Hydrogen as a gas occupies a large volume (11 m3 kg−1) under ambient conditions,for storage. The main challenge in hydrogen storage is to reduce the volume ofgas in equilibrium with the environment. Finally, the hydrogen is reacted withoxygen to produce water in an internal combustion engine or a fuel cell and torelease the stored energy (figure 3).

The production of hydrogen requires renewable energy, i.e. solar energy orgeothermal energy. The solar constant (intensity at the top of the atmosphere)is 1369 W m−2. Approximately 50 per cent reaches the surface of the Earth, and

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3334 A. Züttel et al.

sun

39 kWh kg–1

energy

energy

H2

H2O

H2

H2

H2

O2

O2

O2

O2

25%

82%

85%

50%

electrolysis

storage

combustion

2H2O → 2H2 + O2

M + H2 → MH21

Figure 3. The hydrogen cycle: the energy from sunlight is converted into electricity by means ofphotovoltaic cells. The electricity is used to dissociate water into hydrogen and oxygen. The latter isreleased in the atmosphere and hydrogen is stored, transported and distributed. Finally, hydrogentogether with the oxygen is combusted and the energy is released as work and heat, releasing wateror steam into the atmosphere. Therefore, the hydrogen cycle is closed.

50 per cent of the time it is night. Assuming a conversion efficiency of 10 per cent(e.g. photovoltaic cells), a surface area of 500 000 km2 or 80 m2 per capita (240 m2

per capita for Europeans) is necessary to cover the current world energy demand.

(b) Hydrogen production

Hydrogen is a renewable fuel only if it is produced directly from solar light orindirectly via electricity from a renewable source, e.g. wind power or hydro power.The direct thermal dissociation of H2O requires temperatures greater than 4000 K(Steinfeld & Palumbo 2001). Furthermore, the thermal dissociation of water notonly produces H2 but also atomic hydrogen, oxygen and hydroxyl ions. Thisproduction method is therefore the subject of research activities and variouschallenges, e.g. the separation of hydrogen from oxygen at high temperature,have to be solved before the thermal dissociation becomes an applicable methodfor hydrogen production. Electricity from a renewable energy source can beused for the electrolysis of water. Electrolysis at ambient temperature andambient pressure requires a minimum voltage of 1.481 V and therefore a minimumenergy of 39.7 kWh kg−1 (34.7 kWh kg−1 electricity) hydrogen. Large scale alkalineelectrolyser systems today consume approximately 47 kWh kg−1 hydrogen andwork at 90◦C, i.e. the efficiency is approximately 82 per cent (E. Burkholter 2001,personal communication). High-temperature electrolysis is based on oxygen ion-conducting ceramics. The electrical energy needed to split water at 1000◦C isconsiderably less than electrolysis at 100◦C, i.e. the minimum energy required is34.7 kWh and 25 kWh electricity, respectively.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3335

Table 1. Gravimetric and volumetric hydrogen density in various materials and systems. Thedensity is estimated for systems and in parenthesis is the theoretical limit of the material given.

gravimetric hydrogen volumetric hydrogendensity density(kg H2 kg−1system) (%) (kg H2 m−3 system)

pressure cylinder (500 bar, 25◦C) 4 27liquid hydrogen (1 bar, −253◦C) 3 (100%) 40(71)physisorbed hydrogen (70 bar, −200◦C) 2(4%) 30metal hydrides (1 bar, 25◦C) 1.2 (1.85%) 50(110)complex hydrides (1 bar, 150◦C) 4 (13.5%) 50(120)metals (1 bar, 25◦C) Zn (H2O) 3 (3.8%) 90ammonia (1 bar, −33◦C) (17.6%) (15)hydrocarbons (1 bar, 25◦C) 14 (14%) 10water (1 bar, 25◦C) 11 11

(c) Hydrogen storage

Hydrogen storage basically implies the reduction of the volume of the hydrogengas; 1 kg of hydrogen at a temperature of 25◦C and a pressure of 1 bar requires avolume of 11 m3. There are six methods to store hydrogen with a high volumetricand gravimetric hydrogen density (table 1).

The most common storage systems are high-pressure gas cylinders with amaximum pressure of 20 MPa. New lightweight composite cylinders have beendeveloped that support pressures up to 80 MPa; therefore, the hydrogen reachesa volumetric density of 36 kg m−3, which is approximately half as much as inits liquid form at the normal boiling point. The gravimetric hydrogen densitydecreases with increasing pressure owing to the increasing thickness of thewalls of the pressure cylinder. The volumetric density of hydrogen increaseswith pressure and reaches a maximum, depending on the tensile strength ofthe material, above 100 MPa. However, the gravimetric density decreases withincreasing pressure and the maximum gravimetric density is found for zerooverpressure! Therefore, the increase in volumetric storage density is sacrificedby the reduction in the gravimetric density in pressurized gas systems. Therelatively low hydrogen density together with the very high gas pressures, leakageby diffusion and the cyclic stability of the cylinders are important drawbacks ofthe technically simple and on the laboratory scale well-established high-pressurestorage method.

Liquid hydrogen is stored in cryogenic tanks at 21.2 K at ambient pressure.Owing to the low critical temperature of hydrogen (33 K), liquid hydrogen canonly be stored in open systems, because there is no liquid phase above the criticaltemperature. The pressure in a closed storage system at room temperature couldincrease to about 104 bar before hydrogen solidifies. The volumetric density ofliquid hydrogen at the boiling point (21 K) is 70.8 kg m−3. The challenges forliquid hydrogen storage are improving the energy efficiency of the liquefactionprocess (currently 10 kWh kg−1 H2, theoretically 4 kWh kg−1 H2) and the thermalinsulation of the cryogenic storage vessel in order to reduce the boil-off of

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3336 A. Züttel et al.

hydrogen (typically 0.4% d−1 for tanks with a storage volume of 50 m3, 0.2% for100 m3). The large amount of energy necessary for liquefaction and the continuousboil-off of hydrogen limits the possible applications for liquid hydrogen storagesystems. Applications in which the cost of hydrogen is not an important issueand the hydrogen is consumed within a rather short period of time, e.g. air andspace, are possible.

The origins of the physisorption of gas molecules on the surface of a solidare resonant fluctuations of the charge distributions and are therefore calleddispersive interactions or van der Waals interactions. In the physisorption process,a gas molecule interacts with several atoms at the surface of the solid. Thepotential energy of the molecule shows a minimum at a distance of approximatelyone molecular radius of the adsorbate and the energy minimum (London 1930a,b)is of the order of 0.01–0.1 eV (1–10 kJ mol−1). Owing to the weak interaction,significant physisorption is observed only at low temperatures (less than 100 K).Once a monolayer of adsorbate molecules is formed, the gaseous molecules interactwith a surface of the liquid or solid adsorbate. Therefore, the binding energy ofthe second layer of adsorbate molecules is similar to the latent heat of sublimationor vaporization of the adsorbate. Consequently, the adsorption at a temperatureequal to or greater than the boiling point of the adsorbate at a given pressureleads to the adsorption of one single monolayer (Brunauer et al. 1938). In thecase of carbon as the substrate and hydrogen as the adsorbate, the maximumspecific surface area of carbon is Sspec = 1315 m2 g−1 (single side graphene sheet)and the maximum amount of adsorbed hydrogen (monolayer liquid hydrogen atthe surface) is mads = 3.0 mass%. From this theoretical approximation, we mayconclude that the amount of adsorbed hydrogen is proportional to the specificsurface area of the adsorbent with mads/Sspec = 2.27 10−3 mass% m−2 g and canonly be observed at low temperatures, i.e. close to the critical temperature ofthe gas, such as hydrogen. Despite the numerous publications during the last10 years about hydogen adsorption in nanoporous materials, carbon nanotubes,metal organic frameworks, polymers and spillover from metal catalysts to carbon,no scientific evidence for a new phenomenon beyond physisorption has been found.

Thomas Graham (1867) described the absorption of hydrogen by the metalpalladium. Many metals, intermetallic compounds and alloys in general react withhydrogen and form mainly solid metal–hydrogen compounds. Hydrides exist asionic, polymeric covalent, volatile covalent and metallic hydrides according to theAllred–Rochow electronegativity (Huheey 1983). Hydrogen reacts at an elevatedtemperature with many transition metals and their alloys to form hydrides.The electropositive elements are the most reactive, i.e. scandium, yttrium, thelanthanides, the actinides and the members of the titanium and vanadium groups.The lattice structure is that of a typical metal with atoms of hydrogen onthe interstitial sites; for this reason, they are also called interstitial hydrides.This type of structure has the limiting compositions MH, MH2 and MH3; thehydrogen atoms fit into octahedral or tetrahedral holes in the metal lattice, or acombination of the two types. Especially interesting are the metallic hydridesof intermetallic compounds, in the simplest case the ternary system ABxHn ,because the variation in the elements allows the properties of the hydrides tobe tailored. The A element is usually a rare-earth or an alkaline earth metaland tends to form a stable hydride. The B element is often a transition metaland forms only unstable hydrides. Some well-defined ratios of B to A in the

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3337

(a) (b)

PdH0.6EF

D(E

)

4d-band

energy (eV)

x MH2/x → x M + H2 ABH4 → AH + B + H2

–8 –6 –4 –2 0 2energy (eV)

–8

40

+ –

30

20 s p s

LiB Htotal

10

0

DO

S (s

tate

s pe

r eV

uni

t-ce

ll)

–6 –4 –2 0 2 4 6 8 10

s-ba

nd (

H)

5s-b

and

(Pd)

Tdec(p = p0) = ΔH0

ΔS0⎞⎠

⎛⎝

ln = – ΔH0

R

pp0

⎞⎠

⎛⎝

⎞⎠

⎛⎝

⎞⎠

⎛⎝

ΔS0

R⎞⎠

⎛⎝

1 + T

23

23

23

Figure 4. Comparison of (a) metal and (b) complex hydrides. Metal hydrides retain their structureupon hydrogen sorption, while complex hydrides only exist with hydrogen and form several phasesupon hydrogen desorption.

intermetallic compound x = 0.5, 1, 2, 5 have been found to form hydrides witha hydrogen to metal ratio of up to 2. The metal lattice expands proportional tothe hydrogen concentration by approximately 2–3 Å3 per hydrogen atom (Fukai1989). At greater hydrogen concentrations in the host metal (H/M > 0.1), astrong H–H interaction owing to the lattice expansion becomes important andthe hydride phase (b-phase) nucleates and grows. Metal hydrides have, owingto the phase transition upon hydrogen absorption, the very useful property ofabsorbing large amounts of hydrogen at a constant pressure, i.e. the pressure doesnot increase with the amount of hydrogen absorbed as long as the phase transitiontakes place. The characteristics of the hydrogen absorption and desorption canbe tailored by partial substitution of the constituent elements in the host lattice.Some metal hydrides absorb and desorb hydrogen at ambient temperature andclose to atmospheric pressure. A special feature of the metallic hydrides is the veryhigh volumetric density of the hydrogen atoms present in the host lattice. Thehighest volumetric hydrogen density known today is 150 kg m−3 found in Mg2FeH6and Al(BH4)3. Both hydrides belong to the complex hydrides and will be discussedbelow. Metal hydrides are very effective for storing large amounts of hydrogenin a safe and compact way. All the metallic hydrides that work around ambienttemperature and atmospheric pressure consist of transition metals and, therefore,the gravimetric hydrogen density is limited to less than 2 mass%. Exploring theproperties of the lightweight metal hydrides is still a challenge (figure 4).

Aluminium and boron build complex hydrides of the type M+x [AlH−4 ]x and

M+x [BH−4 ]x . One electron is almost completely transferred from the cation to

each [AlH−4 ] and [BH−

4 ] anion, while the hydrogen is covalently bound to the

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3338 A. Züttel et al.

00 5

500

2013

200120

8050

2013

pressurized H2gas

(composit material)p (MPa)

pres. H2gas

(steel)p (MPa)

5080

density 5 g cm–3 2 g cm–3 1 g cm–3 0.7 g cm–3

H2 chemisorbedon carbon

H2 physisorbedon carbon

10 15 20 25gravimetric H2 density (mass%)

volu

met

ric

H2

dens

ity (

kg H

2 m

–3)

20

40

60

80

100

120

140BaReH9 Mg9FeH6

Mg2NiH4

MgH2

AlH3Al(BH4)3

NaBH4

NaAlH4LiAlH4

LiNH2

C8H18

C4H10

CH4

KBH4

LiH

LiBH4LaNi5H6

FeTiH1.7

<373 K, 1 bar620 K, 1 bar

dec. <273 Kdec. 373 K

dec. 553 K

m.p. 208 K

b.p. 239.7 K

b.p. 272 K

b.p. 112 K

20.3 K

dec. 680 K

dec. 650 K

dec. 400 Kdec. 580 K

dec. >520 K

620 K, 5 bar300 K, 2 bar

300 K, 1.5 bar550 K, 4 bar

160 H± Hcov

H2

NH3liq.

liq.

liq.

liq.

liq.

H2

Figure 5. Volumetric and gravimetric hydrogen density of some selected hydrides. Mg2FeH6 showsthe highest known volumetric hydrogen density of 150 kg m−3, which is more than double that ofliquid hydrogen. BaReH9 has the largest H/M ratio of 4.5, i.e. 4.5 hydrogen atoms per metal atom.LiBH4 exhibits the highest gravimetric hydrogen density of 18 mass%. Pressurized gas storage isshown for steel (tensile strength sv = 460 MPa, density 6500 kg m−3) and a hypothetical compositematerial (sv = 1500 MPa, density 3000 kg m−3).

aluminium or boron. The alkali, alkali earth and many of the transition metalsbuild together with boron and aluminium a large variety of lightweight metal–hydrogen complexes. They exhibit a gravimetric hydrogen density up to an orderof magnitude greater than that of metal hydrides. The number of hydrogen atomsper metal atom is in many cases 2. The hydrogen in the complex hydrides is oftenlocated in the corners of a tetrahedron with boron or aluminium in the centre.The negative charge of the anion, [BH4]− and [AlH4]−, is compensated by a cation,e.g. Li or Na. The hydride complexes of borane, the tetrahydroborates (boranate)M[BH4], and of alane, the tetrahydroaluminate (alanate) M[AlH4], are interestingstorage materials; however, they are known to be stable and decompose onlyat elevated temperatures and often above the melting point of the complex.One of the compounds with the highest gravimetric hydrogen density at roomtemperature known today is Li[BH4] (18 mass%), compared with Be[BH4]2 (28.9mass%), because beryllium is not a feasible hydride (figure 5).

The stability of the tetrahydroboranates and the tetrahydroalanates is relatedto the Pauling electronegativity of the cation element. The empirical equation forthe enthalpy of formation (Nakamori et al. 2006) of the tetrahydroboranates isDH (kJ mol−1 BH4) = 247.4 En − 421.2. The borohydride Al[BH4]3 is one of thefew liquid hydrides at room temperature and releases hydrogen spontaneouslybelow room temperature. The application of the complex hydrides is a challenge,

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3339

because complex hydrides decompose into several phases upon hydrogendesorption according to the equation: M+x [BH−

4 ]x → MHx +x /2 B2H6 → MHx +xB +3x /2H2. According to Pauling (1929), the stability of the elemental hydrideMHx is given by the empirical equation: DH = DHMM + 194 [En(M) − En(H)](kJ mol−1). The stability of the alkali and alkali earth alanates and boranatesvaries over a wide range. Most of the physical properties of the complex hydrideshave not yet been investigated; some of the transition metal boranates may evenbe liquids at room temperature.

3. Energy density

The photosynthesis of plants is inefficient because it is based on a two-photonprocess. The process converts CO2 that is absorbed from the atmosphere bythe leaves and water collected by the roots into carbohydrates (-HCOH-) andoxygen and is therefore binding the same amount of CO2 as is released by thecombustion of the carbohydrates (CO2 neutral). The maximum annual energyconverted in the most efficient plants (C4-plants) is 10 kWh m−2. Furthermore, thecultivation and harvesting is energy intense, especially in heavily industrializedregions of the world, which leads to an additional reduction in the efficiency.The reason for the low efficiency of the energy conversion in plants is thetwo-photon process in the photosynthesis and the large surface area requiredfor the absorption of CO2. The conversion of solar energy to electricity byphotovoltaics is one order of magnitude more efficient and delivers 100 kWh m−2

electricity per year in central Europe and twice that value in the Sahara (NorthAfrica). The storage of electricity in today’s batteries is limited to 0.2 kWh kg−1

and 200 kWh m−3. An average European car today consumes 8.6 l petroleum100 km−1, which corresponds to about 20 kWh of mechanical energy for 100 km.Therefore, the battery required for a 500 km driving range would have a weightof 500 kg, a volume of 2.5 m3 and would cost ¤100 000 (¤1000 kWh−1; K. Hirose,private communication). Alternatively, renewable energy is stored in a syntheticenergy carrier, e.g. hydrogen, hydrocarbons or ammonia, and the energy ofthe synthetic fuel is converted in an internal combustion engine, a turbine ora fuel cell.

(a) Synthetic fuels

According to the empirical models available today the maximum energy densityin hydrides is limited to 7 kWh kg−1 and 5800 kWh m−3, which is half of theenergy density of hydrocarbons like octane (C8H18). Therefore, the combinationof hydrogen from an ideal hydrogen storage material with a fuel cell is able todeliver the same amount of work as the combustion of petroleum in an internalcombustion engine. The best of the recent hydrogen storage systems allowsbetween 0.3 and 1.3 kWh kg−1 and 660 and 1320 kWh m−3 to be stored. Thisis approximately 20 per cent of the theoretical maximum energy density of theideal hydrogen storage material. The main reason for the limited energy densityreached in technical systems is that the best storage materials are known toexhibit 50 per cent of the storage density of the ideal material and the storagesystem contains a maximum of 50 per cent of the storage material.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3340 A. Züttel et al.

sun

energy

energy

synthesis

H2O

H2O

H2O

CO2

CO2O2

O2–(CH2)n–

H2

electrolysis

H2O → H2 + O2

n CO2 + (3n+1) H2 → CH3–(CH2)n–2 – CH3 + 2n H2O

combustion

12

Figure 6. The synthetic fuel cycle: hydrogen is produced from solar energy and water. The CO2 isadsorbed from the atmosphere and reacted with the hydrogen to form hydrocarbons. The cycle isa purely technical representation of natural photosynthesis.

Renewable electricity is used to dissociate water into hydrogen and oxygen inan electrolyser, which is an efficient (approx. 82%) process at large scale. Thehydrogen can be transported in pipelines or absorbed in hydrides. Furthermore,atomic hydrogen at the surface of a metal hydride may react with CO2 toform methane and water. The methane may subsequently be reacted to longerhydrocarbon chains in order to produce a liquid fuel. This represents an efficientroute to synthesize hydrocarbons from hydrogen and CO2. If the CO2 is extractedfrom the atmosphere by an adsorber, then the hydrocarbons from such a processrepresent a synthetic fuel and are CO2 neutral (figure 6). Furthermore, thesynthetic hydrocarbons contain 75 per cent of the energy in the hydrogen.

Synthetic fuels are stored and used just like the fossil fuels and allow an efficientsustainable energy economy based on renewable energy. Besides the advantagethat the well-established technology for the storage, transport and distributionof fuels can be directly used for synthetic hydrocarbons, the process also offers areal CO2 sink. Independent of where the CO2 would be produced and released, itcan be extracted anywhere from the atmosphere. This represents a big advantageover the not sustainable and local sequestration of carbon in power plants.

(b) Energy carriers

The origin of primary energy worldwide is more than 75 per cent fossilfuels (hydrocarbons and coal), complemented by 6 per cent nuclear power(fission), 7 per cent hydropower and 12 per cent biomass. The importance of thehydrocarbons is due to the fact that these energy carriers were abundant a centuryago, easy and safe to store and they exhibit an especially high energy density.Only hydrides and ammonia exhibit an energy density close to that of fossil fuels,

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

Review. Hydrogen: future energy carrier 3341

capacitor

EDLC

mag. coil Pb-acidbattery

hotwater

flywheel

hydrogenstrorage

Li ionbattery

ultimatebattery

biomass oilNH3

hydrides

coal

fusion

fission

comp. air

natural gas

hydrogen

10

100

1000

10 000

100 000

gravitation

electrostatic

electrochemicaland inertia

chemical

nuclear

hydro-power

0.01E

0.1 1 10 100

energy density (kWh kg–1)

ener

gy d

ensi

ty (

kWh

m–3

)

Figure 7. Volumetric versus gravimetric energy density of the most important energy carriers.

as shown in figure 7; all other energy carriers are at least an order of magnitudeless dense in energy except for nuclear reactions, i.e. fission and fusion. However,nuclear reactions are stationary, i.e. very system demanding for the applicationand continuous delivery of power. In addition, fusion is still not carried out ina controlled efficient way in a reactor. The energy density in batteries is verylimited (less than 0.2 kWh kg−1 in a Li ion battery today). Furthermore, batteriesexhibit time constants in the range of several minutes to hours. The potential forthe improvement of current battery technology is estimated to be approximately1 kWh kg−1. Such an ultimate battery could be derived from the Zn–air batteryby replacing the Zn with lithium in a set-up with a non-aqueous electrolyte. Newtypes of batteries, e.g. metal–air, where the energy is stored in metal clusterssuspended in a non-aqueous electrolyte combined with an oxygen (air) electrode,have the potential for a significantly increased energy density and the clustersuspension would allow the battery to be refuelled.

4. Conclusion

In order to combat limited resources, primary energy has to be converted intoheat and electricity. Electricity is partially used directly and also used to produceCO2-neutral synthetic fuels, which do not increase climate change. The scienceand technology for the conversion of renewable energy as well as the synthesis offuels from electricity, water and CO2 is of great economic value because developingcountries in Asia are becoming large market areas. Therefore, the new technologywould allow the challenge of the increasing energy demand to be solved andwould liberate heavily industrialized countries in the West from their economicdependence on fossil fuels.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from

3342 A. Züttel et al.

References

Arrhenius, S. 1896 On the influence of carbonic acid in the air upon the temperature of the ground.Phil. Mag. 41, 237–276.

Brunauer, S., Emmett, P. H. & Teller, E. 1938 Adsorption of gases in multimolecular layers.J. Am. Chem. Soc. 60, 309–319. (doi:10.1021/ja01269a023)

Campbell, C. J. 2003 The essence of oil & gas depletion. Brentwood, UK: Multi-science Publishing.Eidgenössischen Technischen Hochschule Zürich. 2000 Magazine bulletin no. 276. Zurich,

Switzerland: ETH.Engdahl, F. W. 2008 Perhaps 60% of today’s oil price is pure speculation. See

http://www.financialsense.com/editorials/engdahl/2008/0502.html.Fukai, Y. 1989 Site occupancy and phase stability of some metal hydrides. Z. Phys. Chem. 164,

165–174.Graham, T. 1867 On the occlusion of hydrogen gas by metals. Proc. R. Soc. Lond. 16, 422–427.

(doi:10.1098/rspl.1867.0090)Hubbert, M. K. 1956 Nuclear energy and the fossil fuels. Drilling and Production Practice,

Publication No. 95. Washington, DC: American Petroleum Institute & Shell Development Co.Huheey, J. E. 1983 Inorganic chemistry. New York, NY: Harper & Row.London, F. 1930a Zur Theorie und Systematik der Molekularkräfte. Z. Phys. 63, 245–271.

(doi:10.1007/BF01421741)London, F. 1930b Über einige Eigenschaften und Anwendungen der Molekularkräfte. Z. Phys.

Chem. 11, 222–251.Mann, M. E., Bradley, R. S. & Hughes, M. K. 1999 Northern hemisphere temperatures during the

past millennium: interferences, uncertainties, and limitations. Geophys. Res. Lett. 26, 759–762.Nakamori, Y., Miwa, K., Ninomiya, A., Li, H., Ohba, N., Towata, S.-I., Züttel, A. & Orimo,

S.-I. 2006 Correlation between thermodynamical stabilities of metal borohydrides and cationelectronegatives: first-principles calculations and experiments. Phys. Rev. B 74, 045126.(doi:10.1103/PhysRevB.74.045126)

Pauling, L. 1929 The principles determining the structure of complex ionic crystals. J. Am. Chem.Soc. 51, 1010–1026. (doi:10.1021/ja01379a006)

Petit, J. R. et al. 1999 Climate and atmospheric history of the past 420,000 years from the Vostokice core, Antarctica. Nature 399, 429–436. (doi:10.1038/20859)

Saunders, H. D. 1992 The Khazzoom-Brookes postulate and neoclassical growth. Energy J. 13,131–148.

Steinfeld, A. & Palumbo, R. 2001 Solar thermochemical process technology. In Encyclopedia ofphysical science and technology, vol. 15 (ed. R. A. Meyers), pp. 237–256. New York, NY:Academic Press.

Wilson, S. S. 1981 Sadi Carnot: Technik und Theorie der Dampfmaschine. Spektrum Wiss. 99–109.Working Group I. 2007 The physical science basis. Fourth Assessment Report of the IPCC:

Climate Change 2007. Geneva, Switzerland: Intergovernmental Panel on Climate Change. Seehttp://www.ipcc.ch/.

Phil. Trans. R. Soc. A (2010)

on May 24, 2018http://rsta.royalsocietypublishing.org/Downloaded from