Journal of Scientitic & Industrial Research

Vol. 62, January-February 2003 , pp 46-63

Principles of Hydrogen Energy Production, Storage and Utilization

S A Sherif

University of Florida, Department of Mechanical Engineering, Gainesville, FL 326 1 1 -6300

and

F Barbir

Proton Energy Systems, 50 Inwood Road, Rocky Hi l l , CT 06067

and

T N Veziroglu

Clean Energy Research institute, University of Miami, Box 248294, Coral Gables, FL 33 1 24

This paper presents an overview of the principles of hydrogen energy production, storage, and util ization. Hydrogen production wil l cover a whole array of methods including electrolysis, thermolysis, photolysis, thermochemical cycles, and production from biomass. H ydrogen storage will cover al l modes of gaseous, liquid, slush, and metal hydride storage. Hydrogen uti l ization will focus on a large cross section of applications such as fuel cell s and catalytic combustion of hydrogen, to name a few.

Introduction

Hydrogen is a clean, efficient, and versatile energy carrier, which together with electricity, may satisfy all the energy needs and form an energy system that would be permanent and independent of energy sources l -) . It has unique characteristics that make it an ideal energy carrier4 which include the fact that : (a) i t can be produced from and converted into electricity at relatively h igh efficiencies; (b) its raw material for production is water-available in abundance; (c) i t i s a completely renewable fuel ; (d) it can be stored in gaseous form (convenient for large scale storage), in liquid form (convenient for air and space transportation), or in the form of metal hydrides (convenient for surface vehicles and other relatively small scale storage requirements) ; (e) it can be transported over large distances through pipel ines or via tankers ; (f) it can be converted into other forms of energy in more ways and more efficiently than any other fuel (such as catalytic combustion, electro­chemical conversion, and hydriding); (g) it is environmentally compatible as its production, storage, transportation, and end-use do not produce

any pollutants (except for small amounts of nitrogen oxides), greenhouse gases, or any other harmful effects on the environment.

Figure I shows a global energy system in which electricity and hydrogen are produced from available energy sources and u sed in every application where fossil fuels are being used today-in transportation, residential, commercial, and industrial sectors. In such a system, electricity and hydrogen are produced in large i ndustrial plants as well as in small, decentralized units, wherever the primary energy source (solar, nuclear, and even fossi l ) is available. Electricity is used directly or transformed into hydrogen. For large-scale storage, hydrogen can be stored underground in ex-mines, caverns and/or aquifers. Energy transport to the end-users, depending on distance and overall economics, is either in the form of electricity or in the form of hydrogen . Hydrogen may be transported, by means of pipelines or super tankers. It is then u sed in transportation, industrial, residential, and commercial sectors as a fuel . Some of it may be used to generate electricity via fuel cells, depending on demand, geographical location or time of the day.

S HERIF e/ af.: H YDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 47

conversion & distribution use

electricity

I hydrogen

, .... �

,

II

local conversion to electricity (fuel cel ls)

electrical uses

transportation fuel

, domestic fuel industrial fuel & reducing gas chemicals

Figure 1- Hydrogen/electricity energy system

Hydrogen Production

Hydrogen is the most plentifu l element in the universe, making up about three quarters of all matter. All the stars and many of the planets essential ly consist of hydrogen. However, on earth hydrogen in free from is scarce. The atmosphere contains only trace amounts of it (0.07 per cent), and it is usual ly found in small amounts mixed with natural gas in crustal reservoirs. A few wel ls , however, have been found to contain large amounts of hydrogen, such as some wel l s in Kansas that contain 40 per cent hydrogen, 60 per cent ni trogen and trace amounts of hydrocarbons5. The Earth 's surface contains about 0 . 1 4 per cent hydrogen ( 1 0th most abundant element), most of which resides 111 chemical combination with oxygen as water.

Hydrogen, therefore, must be produced. Logical sources of hydrogen are hydrocarbon (fossi l) fuels (CxHy) and water (H20). Presently, hydrogen is mostly being produced from fossi l fuels (natural gas, oi l , and coal ) . However, except for the space program, hydrogen is not being used directly as a fuel

or an energy carrier. It is being used in refineries to upgrade crude oi l (hydrotreating and hydrocracking), in the chemical industry to synthesize various chemical compounds (such as ammonia. methanol, etc . ) , and in metal lurgical processes (as a reduction or protection gas). The total annual hydrogen production worldwide in 1 996 was about 40 mil lion t6 (5.6 EJ) ; less than 1 0 per cent of i t was suppl ied by industrial gas companies; the rest i s being produced at consumer owned and operated plants (so cal led captive production), such as refineries, and ammonia and methanol producers. Production of hydrogen as an energy carrier would requi re an increase in production rates by several orders of magnitude.

In order to faci li tate an early market introduction of fuel cel l-powered veh icles, car manufacturers in col laboration with oi l and fuel processing companies are intensively working on development of on-board fuel processors . These devices would enable use of conventional fuels such as gasoline and diesel , as wel l as methanol , which, being a l iquid fuel is perceived by some as more convenient than hydrogen. This way, the car

48 J SCI IND RES VOL 62 JANU ARY-FEBRUARY 2003

companies plan to c ircumvent the lack of existence of hydrogen refueling infrastructure and the difficulties of on-board hydrogen storage. However, development of such compact fuel processors poses several engineering chal lenges, such as rapid start-up, high efficiency, and very high purity of the produced gas « 1 00 ppm CO).

Logical source for large-scale hydrogen production is water, which is abundant on Earth. Different methods of hydrogen production from water have been, or are being developed. They include electrolysis, direct thermal decomposition or thermolysis, thermochemical processes, and photo­lysis which are detai led below.

Electrolysis

Electrolysis appears to be the only method developed to date, which can be used for large-scale hydrogen production in a post-fossil fuel era. Production of hydrogen by water electrolysis is a mature technology, based on a fundamentally simple process, it i s very efficient, and does not involve moving parts.

The fol lowing reactions take place at the electrodes of an electrolysis cel l fi l led with a sui table electrolyte (aqueous solution of KOH or NaOH or NaCl) upon the appl ication of a potential :

Cathode reaction :

2 H20 (I) + 2e· --7 Hz (g) + 2 OR" (aq) . . . ( I )

Anode reaction :

2 OH· (aq) --7 1 /2 Oz (g) + HzO (I) . . . (2)

Overal l cel l reaction :

H20 ( I ) --7 H2 (g) + O2 (g) . . . (3)

The reversible decomposition potential (Erev = !JGlnF) of the above reaction is 1 .229 V at standard conditions. The total theoretical water decomposition potential is 1 .480 V corresponding to hydrogen' s enthalpy (since L\H = !JG + T!JS) . Due to irreversible processes occurring at the anode and cathode, inc luding the electrical resistance of the cell , the actual potentials are always h igher, typically 1 .75 -2 .05 V. It corresponds to efficiencies of 72-82 per cent.

Several advanced electrolyzer technologies are being developed such as the advanced alkaline

e lectrolysis (which employs new materials for membranes and electrodes that al low further improvement in efficiencl s, up to 90 per cent) ; the solid polymer electrolytic (SPE) process (which employs a proton-conducting IOn exchange membrane as electrolyte and as membrane that separates the electrolysis cel lS.',) ; and h igh temperature steam electrolysis (which operates between 700- 1 000°C and which employs oxygen ion­conducting ceramics as electrolyte 1 0) . An electrolysis plant can operate over a wide range of capacity factors and is convenient for wide range of operating capacit ies, which makes this process interesting for coupling with renewable energy sources, particularly with photovol taics (PV). Photovoltaics generate low voltage-direct current, which is exactly what is required for the electrolysis process. Theoretical and experimental studies on the performance of photo­voltaic-electrolyzer systems have been performed 1 1 . 1 4 . Examples of experimental PV -electrolysis plants that are currently in operation worldwide inc lude the Solar-Wasserstoff-Bayem pi lot plant l 5 (Neunburg vorm Wald, Germany), the HYSOLAR project l 6

(Saudi Arabia), Schatz Energy Centerl 7 (Humboldt State University, Arcata, Cal ifornia) , Helsinki University of Technology l 8 (Helsinki, Finland), and INTA Energy Laborator/ 9 (Huelva, Spain).

Direct Thermal Decomposition of Water {Thermolysis)

Water can be spl i t thermally at temperatures above 2000 K. The overall thermal dissociation of water can be shown as20:

. . . (4)

The degree of dissociation is a function of temperature : only I per cent at 2000 K. 8 .5 per cent at 2500 K and 34 per cent at 3000 K. The product is a mixture of gases at extremely h igh temperatures. The main problems in connection with this method are related to materials required for extremely high temperatures, recombination of the reaction products at h igh temperatures and separation of hydrogen from the mixture.

Thermochemical Cycles

Thermochemical production of hydrogen involves chemical splitting of water at temperatures lower than those needed for thermolysis, through a

)I

, � .

-..

S HERIF e( at. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 49

series of cycl ical chemical reactions that ult imately release hydrogen. Some of the more thoroughly investigated thermochemical process cycles include the fol lowing9.

2 1 , 22 : sulfuric acid - iodine cycle, hybrid sulfuric acid cycle, hybrid sulfuric acid -hydrogen bromide cycle, calcium bromide - i ron oxide cycle (UT-3), and iron chlorine cycle (Mark 9). Depending on the temperatures at which these processes are occurring, relatively high efficiencies are ach ievable (40-50 per cent) . However the problems related to movement of large mass of materials in chemical reactions, toxicity of some of the chemicals involved, and corrosion at h igh temperatures remain to be solved in order for these methods to become practical .

Photolysis

Photolysis (or direct extraction of hydrogen from water using only sunlight as an energy source) can be accompl i shed by employing photobiological systems, photochemical assembl ies, or photoelectro­chemical cells23,24. Intensive research activ ities are opening new perspectives for photo-conversion, where new redox catalysts, colloidal semiconductors, immobil ized enzymes and selected micro-organisms could provide means of large-scale solar energy harvesting and conversion into hydrogen.

Hydrogen Production From Biomass

Hydrogen can be obtained from biomass by a pyrolysis/gasification process25 . The biomass preparation step invo lves heating of the biomass/water s lurry to high temperatures under pressure in a reactor. This process decomposes and partial ly oxidizes the biomass, producing a gas product consisting of hydrogen, methane, CO2, CO, and nitrogen. M ineral matter is removed from the bottom of the reactor. The gas stream goes to a h igh temperature shift reactor where the hydrogen content is increased. Relatively high purity hydrogen i s produced in subsequent pressure swing adsorption unit. The whole system is very much similar to a coal gasification plant, with the exception of the unit for pretreatment of biomass and the design of the reactor. Because of the lower calorific value per unit mass of biomass as compared to coal , the processing fac i l i ty is larger than that of a comparably sized coal gasification plant.

Gaseous Hydrogen Storage

Hydrogen as an energy carrier must be stored to overcome daily and seasonal discrepancies between energy source avai lab i l i ty and demand. Depending on storage size and appl ication , several types of hydrogen storage systems may be differentiated . This includes stationary large storage systems (which are typically storage devices at the production site or at the start or end of pipel ines and other transportation pathways) ; stationary smal l storage systems at the distribution or final user level (for example, a storage system to meet the demand of an industrial plant) ; mobi le storage systems for transport and distribution including both large-capacity devices (such as a l iquid hydrogen tanker - bulk carrier) and small systems (such as a gaseous or l iquid hydrogen truck trailer) ; and vehicle tanks to store hydrogen used as fuel for road vehicles. Because of hydrogen' s low density, its storage always requires relatively large volumes and is associated with e ither high pressures (thus requiring heavy vessels), or extremely low temperatures, and/or combination with other materials (much heavier than hydrogen i tself) . Table I shows ach ievable storage densities with different types of hydrogen storage methods. Some novel hydrogen storage method may achieve even h igher storage densities, but have yet to be proven in terms of practical ity, cost, and safety.

Table I - Hydrogen storage types and densities

kg H2 /kg Kg Hzlm3

Large volume storage (102 to 104 m3 geometric volume)

Underground storage

Pressurized gas storage(above ground)

Metal hydride

0.0 1 -0.0 1 4

0.0 1 3-0.0 1 5

5- 1 0

2- 1 6

50-55

Liquid hydrogen - I 65-69

Stationary small storage (1 to 100 m3 geometric volume)

Pressurized gas cylinder

Metal hydride

Liquid hydrogen tank

0.0 1 2

0.0 1 2-0.0 1 4

0. 1 5-0.50

- 1 5

50-53

-65

Vehicle tanks (0.1 to 0.5 m3 geometric volume)

Pressurized gas cyl inder

Metal hydride

Liquid hydrogen tank

0.05

0.02

0.09-0. 1 3

1 5

5 5

50-60

so J SCI IND RES VOL 62 JANUARY-FEBRUARY 2003

Large Underground Hydrogen Storage

Future hydrogen supply systems wil l have a structure s imi lar to today' s natural gas supply systems. Underground storage of hydrogen in caverns, aquifers, depleted petroleum and natural gas fields, and man-made caverns resulting from mining and other activities i s l ikely to be technological ly and economical ly feasible

26. Hydrogen storage systems of the same type and the same energy content wi l l be more expensive by approximately a factor of three than natural gas storage systems, due to hydrogen' s lower volumetric heating value. Technical problems, specifical ly for the underground storage of hydrogen other than expected losses of the working gas in the amount of 1 -3 per cent/y are not anticipated. The c ity of Kiel ' s publ ic ut i l ity has been storing town gas with a hydrogen content of 60-65 per cent in a gas cavern with a geometric volume of about 32,000 m' and a pressure of 80 to 1 60 bar at a depth of 1 ,330 m since 1 97 1 (ref. 27). Gaz de France (the French national gas company) has stored hydrogen-rich refinery by­product gases in an aquifer structure near Beynes, France. Imperial Chemical Industries of Great Bri tain stores hydrogen in the salt mlOe caverns near Tees ide, United Kingdom2R.

Above Ground Pressurized Gas Storage Systems

Pressurized gas storage systems are used today in natural gas business in various sizes and pressure ranges from standard pressure cyl inders (SO L, 200

bar) to stationary high-pressure containers (over 200

bar) or low-pressure spherical containers (>30,000

m', 1 2 to 1 6 bar) . This appl ication range wi l l be s imi lar for hydrogen storage.

Vehicular Pressurized Hydrogen Tanks

Development of u l tra-l ight but strong new composite materials has enabled storage of hydrogen in automobi les . Pressure vessels that allow hydrogen storage at pressures >200 bar have been developed and used in automobi les (such as Daimler-Benz NECAR II) . A storage density higher than 0.05 kg of hydrogen per I kg of total weight is easi ly achievable29.

Metal Hydride Storage

Hydrogen can form metal hydrides with some metal s and alloys. During the formation of the metal hydride, hydrogen molecu les are split and hydrogen

atoms are inserted in spaces inside the lattice of suitable metals and/or alloys. In such a way, effect ive �torage comparable to the densi ty of l iquid hydrogen IS created. However, when the mass of the metal or alloy is taken into account, then the metal hydride glravimetric storage densi ty is comparable to storage of pressurized hydrogen. The best ach ievable gravimetric storage density i s about 0.07 kg of H2/kg of metal, for a high temperature hydride'o such as MgHz. During the storage process (charging or absorption) heat is released which must be removed in order to achieve the continuity of the reaction. During the hydrogen release process (discharging or desorption) heat must be suppl ied to the storage tank. An advantage of storing hydrogen in hydriding substances i s the safety aspect. A serious damage to a hydride tank (such as one that could be caused by a col l i s ion) would not pose fire hazard since hydrogen would remain in the metal structure.

Novel Hydrogen Storage Methods

Hydrogen can be physical ly adsorbed on activated carbon and be "packed" on the surface and inside the carbon structure more densely than if it has

·

been just compressed. Amounts of up to 48 g H2 per kg of carbon have been reported, 1 at 6.0 MPa and 87 K. The adsorption capacity i s a function of pressure and temperature, therefore at h igher pressures and/or lower temperatures even larger amounts of hydrogen can be adsorbed. For any practical use, relat ively low temperatures are needed « 1 00 K). Since the adsorption is a surface process, the adsorption capacity of hydrogen on act ivated carbon is largely dlue to the high surface area of the acti vated carbon, although there are some other carbon properties that affect the capabi l i ty of act ivated carbon to adsorb hydrogen.

Researchers from Northeastern University in Boston, MA, have recently announced that they have developed a carbon storage material that can store as high as 75 per cent of hydrogen by weight>2. This material , apparently some kind of carbon nanotubes or carbon whiskers, is currently being researched in $everal laboratories. The best results achieved with carbon nanotubes to date confi rmed by the Nat ional Renewable Energy Laboratory is hydrogen storage density corresponding to about 1 0 per cent of the

b . I 11 nanotu e welg 1t · .

>

!

y

S HERIF et al. : H YDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 5 1

Hydrogen can be stored in glass microspheres of approximately 50 f..lm diam. The microspheres can be fi l led with hydrogen by heating them to increase the glass permeabi l ity to hydrogen. At room temperature, a pressure of approximately 25 MPa is achieved34 resulting in storage density of 1 4 per cent mass fraction and 1 0 kg H2/m3 . At 62 MPa, a bed of glass micro spheres can store 20 kg H2/m3 . The release of hydrogen occurs by reheating the spheres to again increase the permeabi l ity. Researchers at the University of Hawai i are investigating hydrogen storage via polyhydride complexes. Complexes have been found which catalyze the reversible hydrogenation of unsaturated hydrocarbons. This catalytic reaction could be the basis for a low temperature hydrogen storage system with an avai lable hydrogen density greater than 7 per cent's .

Liquid Hydrogen

Liquid hydrogen' s favorable characteristics include its h igh heating value per unit mass and large cool ing capacity due to its h igh specific heat36, 37 . Liquid hydrogen has some important uses such as in the space program, in h igh-energy nuclear physics, and in bubble chambers. The transport of hydrogen is vastly more economical when it is in l iquid form even though cryogenic refrigeration and special dewar vessels are required . Although l iquid hydrogen can provide a lot of advantages, its uses are restricted in part because l iquefying hydrogen by existing conventional methods consumes a large amount of energy (around 30 per cent of its heating value). Liquefying I kg of hydrogen in a medium-size plant requires 1 0 to 1 3 kWh of energy (electrici ty),7 . In addition, boi l -off losses associated with the storage, transportation, and handl ing of l iquid hydrogen can consume up to 40 per cent of its avai lable combustion energy. It is therefore important to search for ways that can improve the effic iency of the l iquefiers and diminish the boi l-off losses.

Hydrogen Liquefaction

The production of l iquid hydrogen requi res the use of l iquefiers that uti l i ze different principles of cool ing. In general, hydrogen l iquefiers may be classified as conventional, magnetic, or hybrid . Many types of conventional l iquefiers exist such as the Linde-Hampson l iquefiers, the Linde Dual-Pressure l iquefiers, the Claude l iquefiers, the Kapitza

l iquefiers, the Heylandt l iquefiers, and the Coll ins l iquefiers, just to name a few. Conventional l iquefiers general ly comprise compressors, expanders, heat exchangers, and Joule-Thomson valves. Magnetic l iquefiers, on the other hand, uti l ize the magnetocaloric effect. This effect is based on the principle that some magnetic materials experience a temperature increase upon the appl ication of a magnetic field and a temperature drop upon l ifting the magnetic field. The magnetic analog of several conventional l iquefiers includes the Brayton l iquefiers, the Stirl ing l iquefiers, and the Active Magnetic Regenerative (AMR) l iquefier. Additional information on l iquid hydrogen production methods can be found in Sherif et al . (ref. 38) .

Liquid Hydrogen Storage

Hydrogen is usual ly transported in large quantities by tnlck tankers of 30 to 60 m3 capacity, by rai l tank cars of 1 1 5 m' capacity, and by barge containers of 950 m3 capacity3,) . Liquid hydrogen storage vessels are usual ly available in sizes ranging from one l iter dewar flasks used in laboratory appl ications to large tanks of 5000 m' capacity. The National Aeronautics and Space Administration (NASA) typical ly uses large tanks of 3,800 m1

capacity (25 m diam)26. The total boi l-off rate from

such dewars is approxi • • lately 600,000 L/y, which is vented to a bum pond.

The contributing mechanisms to boil-off losses in cryogenic hydrogen storage systems are : ( I ) Ortha-para conversion; (2) Heat leak (shape and size effect, thermal stratification, thermal overfi l l , insulation, conduction, radiation, cool-down), (3) S loshing, and (4) Flashing. These mechanisms wi l l be described in what fol lows.

At high temperatures, hydrogen is an equi l i ­brium mixture of 75 per cent ortho-hydrogen and 25

per cent para-hydrogen. Hydrogen of this composition i s termed normal and i s in equi l ibrium at room temperature and above. At zero Kelvin, on the other hand, al l the molecules must be in a rotational ground state at equ i l ibrium, therefore all the hydrogen molecules at

, �qui l ibrium should be in the

para state.

By cool ing hydrogen gas from room temperature to the nbp (normal boi l ing point) the ortha-hydrogen converts to para-hydrogen spontaneously. The conversion of ortho-hydrogen to

52 J S C I INO RES VOL 6 2 JANUARY-FEBRUARY 2003

para-hydrogen is an exothermic reaction. The heat of conversion i s related to the change of momentum of the hydrogen nucleus when the direction-of-spin changes. The amount of heat given off in this conversion process is temperature dependent. The heat of conversion is greater than the latent heat of vaporization of normal and para-hydrogen at the nbp. If the unconverted normal hydrogen is placed in a storage vessel , the heat of conversion wi l l be released within the container, which leads to the evaporation of the l iquid. Because of these pecul iarities in the physical properties of hydrogen, the boi l -off of the stored liquid wi l l be considerably larger than what one would be able to determine from calculations based on ordinary heat leak to the storage tank. In order to minimize the storage boi l-off losses, the conversion rate of ortho- to para-hydrogen should be accelerated with a catalyst that converts the hydrogen d · h I ' f . 40-4, urtng t e (que actIOn process . .

The use of a catalyst usual ly resu lts in a larger refrigeration load and consequently in an effic iency pena lty primari ly because the heat of conversion must be removed. The t ime for which hydrogen is to be stored usual ly determines the optimum amount of conversion. For use within a few hours, no conversion is necessary. For example, large-scale use of l iquid hydrogen as a fuel for jet aircraft i s one of those cases where conversion is not necessary since uti l i zation of the l iquid is almost a continuous process and long-term storage is therefore not needed4"'. For some other uses, a partial conversion might be requ ired to create more favorable conditions. It should be noted that for every initial ortho concentration there exists a unique curve for boi l-off of hydrogen with respect to t ime.

The heat of conversion at higher temperatures is sma l l (at 300 K the heat of conversion is 270 kJ/kg) and increases as the temperature decreases, where it reaches 5 1 9 kJ/kg at 77 K, the nbp for l iquid ni trogen. For temperatures lower than the nbp of nitrogen, the heat of conversion remains constant (523 kJ/kg) . However, the equ i l ibrium percentage of para-hydrogen at the nbp is 50 per cent. Therefore, in order to produce equi l ibrium liquid hydrogen, 50 per cent of hydrogen must be converted from ortha to para at temperatures lower than l iquid nitrogen temperatures. S ince the heat of conversion is almost constant for temperatures lower than those of l iquid nitrogen, the refrigeration load required to remove

the heat of conversion should be a lmost constant at different temperatures. Ideal refrigeration requires that the energy needed for the same amount of refrigeration load be temperature-dependent. This suggests that in order to operate at an optimum power, i t is necessary to have a continuous conversion, especial ly for temperatures lower than those of l iquid nitrogen . By continuous conversion, up to 1 5 per cent energy saving may be achieved.

The heat leakage losses are general ly proportional to the ratio of surface area to the volume of the storage vessel (SIV) . The most favorable shape is therefore spherical since it has the least surface to volume rat io . Spherical shape containers have another advantage. They have good mechan ical strength since stresses and strains are distri buted uniformly. Storage vessels may also be constructed in other shapes such as cylindrical, conical , or any combination of these shapes. Cylindrical vessels are usual ly required for transportation of l iquid hydrogen by trailers or rai lway cars because of l imitations imposed on the maximum allowable diameter of the vessel . For normal highway transportation, the outside diameter of the vessel cannot exceed 2 .44 m. From an economics standpoint, cyl indrical vessels with either dish, e l l iptical or hemispherical heads are very good, and their SlY ratios are only about 1 0

percent greater than that of the sphere44.

S ince boi l-off losses due to heat leak are proportional to the SIV ratio, the evaporation rate wi l l diminish drastical ly as the storage tank size i s increased. For double-walled, vacuum-insulated, spherical dewars, boi l -off losses are typically 0.3 to 0.5 percent per day for containers having a storage volume of 50 m', 0.2 percent for 1 03 m"' tanks, and about 0.06 percent for 1 9000 m' tanks45 . Obvious ly, the larger the size of the dewar, the smal ler the cost per unit volume of storage. Interestingly enough, the rate of evaporation does not substantia l ly decrease with increasing the size of the container for cyl indrical vessels of constant diameter.

Additional boil-off may be caused by thermal stratification in l iquid hydrogen in which case the warmer upper layers evaporate much faster than the bulk l iquid44. One way of decreasing boi l-off losses due to stratification and thermal overfi II is by employing high-conducti vity plates (conductors) instal led vertically in the vessel . The plates produce heat paths of low resistance between the bottom and

y

S HERI F el at. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 53

top of the vessel and can operate most satisfactorily in el iminating temperature gradients and excessive pressures. Another way is to pump the heat out and maintain the l iquid at subcooled or saturated conditions. An ideal refrigeration system to perform this task can be an efficient magnetic refrigerator. The magnetic refrigerator is very suitable for this job because of its relat ively higher efficiency, compactness, lower price, and reli abi l i t/6.53 .

Liquid hydrogen containers are usually of three types: double-jacketed vessels with l iquid nitrogen in the outer jacket, superinsu lated vessels with either a reflecting powder or multi l ayer insulation, and containers with vapor-cooled shields employing super insulation. Although mult i layer insulation eMU) provides for a low boi l-off rate, the addition of a vapor-cooled shield (VeS) wi l l lower the boi l-off losses even further. A ves i s a type of insulation that takes the vapor boi l-off and passes i t past the tank before being re-l iquefied or vented. Publ ished data indicate that a reduction of more than 50 per cent in boi l-off losses may be achieved for a 1 00,000 Ib l iquid hydrogen cryogenic faci l ity with a ves than without one. Brown54 showed that locating the ves at half the distance from the tank to the outer surface of a four-inch MU of a 1 00,000 Ib faci l ity would reduce the boi l-off by ten percent . A dual ves system on the tank would improve the performance by 40 per cent over a single yes . Brown54 also showed that the preferred locations for the inner and outer shields in a dual ves system are 30 and 66 per cent of the distance from the tank to the outer surface of the MLI.

Mechanical supports that connect the inner and outer vessel of double-wal led vacuum tanks are an integral part of the insulation problem, primarily because of the heat influx that can be conducted to the inner vessel . One way to diminish these boil-off losses is to continuously cool the support employing the produced hydrogen vapor in a counter-flow fashion.

Another process that leads to boil-off during l iquid hydrogen transportation by tankers is sloshing. S loshing is the motion of l iquid in a vessel due to acceleration or deceleration. Due to different types of acceleration and deceleration, there exist different types of sloshing. Acceleration causes the l iquid to move to one end and then reflect from that end, thus, producing a hydraul ic jump. The latter then travels to

the other end, thus transforming some of its impact energy to thermal energy. The thermal energy dissipated eventually leads to an i ncrease in the evaporation rate of the l iquid4o•44• The insertion of traverse or anti-slosh baffles not only restrains the motion of the l iquid, thus reducing the impact forces, but also increases the frequency above the natural frequency of the tanker 40. 55 .

Another source of boi l-off i s flashing. Thi s problem occurs when l iquid hydrogen, at a h igh pressure (2.4 to 2.7 atm), is transferred from trucks and rail cars to a low-pressure dewar ( 1 . 1 7 atm). This problem can be reduced if transportation of l iquid hydrogen is carried out at atmospheric pressures. Furthermore, some of the low-pressure hydrogen can be captured and re-l iquefied.

Slush Hydrogen

S lush hydrogen is a mixture of l iquid and frozen hydrogen in equi l ibrium with the gas at the triple point, 1 3 .8 K. The density of the ice-l ike form is about 20 per cent h igher than that of the boi l ing l iquid. To obtain the ice- l ike form, one has to remove the heat content of the l iquid at 20.3 K until the triple point is reached and then remove the latent heat of fusion. The "cold content" of the ice- l ike form of hydrogen i s some 25 per cent h igher than that of the saturated vapor at 20.3 K.

S lush hydrogen i s of great interest for the space program because of its potential in reducing its physical size and sign ificantly cutting the projected gross l i ftoff weight56. Some of the problems related to l iquid hydrogen storage such as, low density, temperature stratification, short holding time due to its low latent heat, hazards associated with h igh vent rates, and unstable fl ight conditi ons caused hy sloshing of the l iquid in the fuel tank, may be e l iminated or reduced i f slush hydrogen is used .

Baker and Matsh57 gave a tabulated comparison among the properties of l iquid and slush hydrogen. According to them, the smal ler enthalpy of slush hydrogen reduces the evaporation losses during storage and transport and as a consequence permits longer storage without venting. The increased density permits ground transport equipment to carry 1 5 per cent greater loads; for space vehic les, a given fuel load can be carried in tankage which are 1 3 per cent smaller in volume.

54 J SCI IND RES VOL 62 JANUARY-FEBRUARY 2003

There are basically two methods for slush hydrogen production, namely : freeze-thaw and auger. Sindt5H summarized the basic principles of the freeze­thaw method and its feasibi l ity. Two of the methods reported by Sindt involve the formation of solid partic les by spraying precooled l iquid into an environment wel l below the triple-point pressure and inducing solid formation by al lowing precooled hel ium gas to flow through a triple-point temperature l iquid. The spray technique produced a very low­density solid particle, which had to be partial ly melted to form a l iquid-sol id mixture. The technique of allowing hel ium to flow through the l iquid produced solid by reducing the partial pressure of hydrogen below its triple-point pressure which resulted in a c lear, rigid tube of solid forming around the helium bubble train. The solid was found not to be convertible to a mixture without further treatment such as cmshing. S indt found that a more desirable preparation method was to evacuate the u l lage over the l iquid, which created a solid layer at the l iquid surface. The sol id layer produced had a texture that was very dependent upon the rate of vacuum pumping. Low pumping rates produced a dense, nearly transparent layer of sol id . Subsequent breaking of the solid left large rigid pieces that formed inhomogeneous mixtures even with vigorous mixing. Higher pumping rates produced a solid layer that was porous and consisted of agglomerates of loosely attached fine solid particles. The solid layer was easi ly broken and mixed with the l iquid to form a homogeneous l iquid-sol id mixture that could be defined as slush. S indt discovered, however, that sol ids could not be continuously generated at the l iquid surface without vigorous mixing.

S indt58 also reported on a continuous preparation method, which he developed, capable of forming solid layers in cycles by periodic vacuum pumping. During the pumping (pressure reduction) portion of the cycle, the solid layer formed as previously described. When the pumping was stopped, the solid layer melted and settled into the l iquid, forming slush. Although mixing was not necessari ly required, the preparation method was accelerated by some mechanical mixing which helped break the layer of solids. Thi s technique has been cal led the "freeze-thaw" process.

Although the freeze-thaw production method was proved to be technical ly feasible and ful ly

developed, i t was shown to have disadvagtages. These disadvantages include the fact that the freeze­tlhaw process is a batch process and that i t operates at the triple point pressure of the cryogen . For hydrogen, this pressure was estimated to be 0.07 bars absolute, a pressure capable of drawing air through inadvertent leaks in the system. The air could be thought of as presenting a potential safety problem in a hydrogen system. Voth59 showed that the freeze­thaw process requi red either costly equipment to recover the generated triple-point vapor or the loss of approximately 1 6 per cent of the nbp l iquid hydrogen and approximately 24 per cent of the nbp l iquid oxygen if the vapor was discarded.

Voth)9 reported on a study for producing I iquid­solid mixtures of oxygen or hydrogen using an auger. He described an auger used to scrape frozen solid from the inside of a refrigerated brass tube in order to produce slush hydrogen . He showed that slush hydrogen could be continuously produced by this method, and s ince it could be immersed in l iquid, sllush was produced at pressures above the triple­point pressure. Voth was also able to produce the increased pressure pneumatical ly or by generating temperature stratification near the surface of the l iquid . He observed that the auger system produced particles in the size realm of the particle produced by the freeze-thaw method so that, l ike the freeze-thaw­produced slush, the auger-produced slush could be readi ly transferred and stored .

According to Voth59, the thermodynamic reversible energy requ ired to produce s lush hydrogen was equal to the thermodynamic avai labi l ity of the s lush. Also using normal hydrogen at a temperature of 300 K and a pressure of 1 .0 1 3 bars as the base fluid, the reversible energy required to produce nbp para-hydrogen was 397 1 .4 W.h/kg whi le the reversible energy required to produce slush with a solid mass fraction of 0.5 was 4372.8 W.h/kg. Voth explained that the energy required by practical systems was higher because of component inefficiencies. The calculated energies for the four cases were based on l iquefier and refrigerator elfficiencies of 40 per cent of that of Carnot. He assumed that the vacuum pumps required by the freeze-thaw production had an efficiency of 50 per cent of the isothermal efficiency. However, he did not include the increased production energy due to heat leak into the containers and transfer l ines

-,.

S HERIF el al. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 55

because it was difficult to estimate without a firm system defin it ion and because the heat leak would have been nearly equal for al l of the systems studied. Because heat leak was not included, the calculated energies were lower than those of an actual system.

It is apparent that the use of slush hydrogen should be considered only for cases in which higher density and greater solid content are real ly needed. This is mainly because the production costs of slush hydrogen now and in the near future are greater than the l iquefaction costs of hydrogen due to the larger energy use involved.

Hydrogen Transport

In the hydrogen energy system it is envisaged that from the production plants and/or storage, hydrogen wi l l be transmitted to consumers by means of underground p ipel ines (gaseous hydrogen) and/or supertankers ( l iquid hydrogen). Presently, hydrogen transportat ion through pipelines is used either in l inks between nearby production and uti l ization sites (up to 1 0 km) or in more extensive networks (roughly 200 km)6o. Future developments wi l l certainly entail greater flow rates and distances. However, it would be possible to use the existing natural gas pipelines with some modifications. For hydrogen pipel ines, it is necessary to use steel less prone to embrittlement by hydrogen under pressure (particularly for very pure hydrogen (>99.5 per cent purity). Reciprocating compressors used for natural gas can be used for hydrogen without major design modifications. However, special attention must be given to sealing ( to avoid hydrogen leaks) and to materials selection for the parts subject to fatigue stress. Use of centrifugal compressors for hydrogen creates more problems due to hydrogen ' s exceptional l ightness .

As a rule, hydrogen transmission through pipel ines requires larger diameter piping and more compression power than natural gas for the same energy throughput. However, due to lower pressure losses in the case of hydrogen, the recompression stations wou ld need to be spaced twice as fart apart. In economic terms, most of the studies found that the cost of large-scale transmission of hydrogen is about 1 .5 to 1 . 8 times that of natural gas transmission. However, transportation of hydrogen over distances greater than 1 ,000 km is more economical than transmission of electricit/ I . Hydrogen in the gas phase IS general ly transported tn pressurized

cylindrical vessels (typical ly at 200 bar) arranged in frames adapted to road transport. The unit capacity of these frames or skids can be as great as 3,000 m'. Hydrogen gas distribution companies also instal l such frames at the user site to serve as a stationary storage.

Hydrogen Conversion

Hydrogen as an energy carrier can be converted into useful forms of energy in several ways including combustion in internal combustion engines and jet and rocket engines, combustion with pure oxygen to generate steam, catalytic combustion to generate heat, electrochemical conversion to e lectrici ty, and metal hydride conversions.

Combustion in Internal Combustion, Jet, and Rocket Engines

Hydrogen powered internal combustion engines are on average about 20 per cent more efficient than comparable gasoline engines. The ideal thermal efficiency of an internal combustion engine is :

. . . ( 5 )

where r is the compression rat io and k is the ratio of specific heats (CriC). The equation shows that thermal efficiency can be improved by increasing either the compression ratio or the specific heat rat io . In hydrogen engines both ratios are higher than in a comparable gasoline engine due to hydrogen ' s lower self-ignition temperature and abi l i ty to burn in lean mixtures.

However, the use of hydrogen in internal combustion engines results in loss of power due to the lower energy content in a stoich iometric mixture in the engine 's cyl inder. A stoichiometric mixture of gasol ine and air, and gaseous hydrogen and air, premixed external ly occupies -2 and 30 per cent of the cylinder volume, respectively . Under these conditions, the energy of the hydrogen mixture is only 85 per cent of the gasol ine mixture, thus resu lt ing in about 1 5 per cent reduction in power. Therefore, the same engine running on hydrogen wi l l have - 1 5 per cent less power than when operated with gasol ine. The power output of a hydrogen engine can be improved by using more advanced fuel injection techniques or l iquid hydrogen. For example, if l iquid hydrogen is premixed with air, the amount of hydrogen that can be introduced into the combustion

56 J SCI IND RES VOL 62 JANUARY-FEBRUARY 2003

chamber, can be increased by approximately one­third62.

One major advantage is the fact that hydrogen engines emit far fewer pol lutants than comparable gasoline engines. Basical ly, the only products of hydrogen combustion in air are water vapor and small amounts of n itrogen oxides. Hydrogen has a wide flammabi lity range in air (4 to 75 per cent vol . ) and therefore h igh excess air can be ut i l ized more effectively. The formation of n itrogen oxides in hydrogen/air combustion can be minimized with excess air. NOx emissions can also be lowered by cooling the combustion environment using techniques such as water injection, exhaust gas re­circulation, or using l iquid hydrogen. The emissions of NOx in hydrogen engines are typical ly one order of magnitude smaller than emissions from comparable gasol ine engines primari ly because the former run leaner than the latter. Smal l amounts of unburned hydrocarbons, CO2, and CO have been detected in hydrogen engines due to lubrication oi 162.

Steam Generation By Hydrogen/Oxygen Combustion

Hydrogen combusted with pure oxygen results in pure steam. This is represented by the fol lowing chemical reaction :

. . . (6)

This creates temperatures in the flame zone above 3 ,000 DC, therefore additional water has to be injected so that the steam temperature can. be regulated at a desired leve l . Both saturated and superheated vapor can be produced.

A compact hydrogen/oxygen steam generator has been commercial ly developed by the German �_�:-ospace Research Establ ishment (DLR)6J . This generator consists of the ignition, combustion, and evaporation chambers. In the ignition chamber, a combustible m ixture of hydrogen and oxygen at a low oxidant/fuel ratio is ignited by means of a spark plug. The rest of the oxygen is added in the combustion chamber to adjust the oxidant/fuel ratio exactly to the stoichiometric one. Water i s also injected in the combustion chamber after i t has passed through the double wal ls of the combustion chamber. The evaporation chamber serves to homogenize the steam. The steam's temperature is monitored and control led. Such a device is close to 1 00 per cent efficient, s ince there are no emissions

other than steam and l ittle or no thermal losses. Hydrogen steam generators can be used to generate steam for spinning reserve in power plants, for peak load electricity generation, for industrial steam supply networks, and as a micro steam generator in medical technology and biotechnolog/J .

Catalytic Combustion of Hydrogen

Hydrogen and oxygen in the presence of a suitable catalyst may be combined at temperatures significantly lower than flame combustion (from ambient to 500 DC). This principle can be used to design catalytic burners and heaters. Catalytic burners require considerably more surface area than conventional flame burners. Therefore, the catalyst is typical ly dispersed in a porous structure. The reaction rare and resulting temperature are easi ly control led, by controll ing the hydrogen flow rate. The reaction takes place in a reaction zone of the porous catalytic s intered metal cyl inders or plates in which hydrogen and oxygen are mixed by diffusion from opposite sides. A combustible mixture is formed only in the reaction zone and assisted with a (platinum) catalyst to bum at low temperatures . The only product of catalytic combustion of hydrogen is water vapor. Due to low temperatures there are no nitrogen oxides fOirmed. The reaction cannot migrate into the hydrogen supply, since there is no flame and hydrogen concentration is above the h igher flammable l imit (75 per cent). Possible appl ications of catalytic burners are in household appl iances such as cooking ranges and space heaters. The same pritnciple is also used in hydrogen sensors .

Electrochemical Conversion (Fuel Cells)

Hydrogen can be combined with oxygen without combustion in an electrochemical reaction (reverse of electrolysis) and produce electric i ty (DC). The device where such a reaction takes place is called the electrochemical fuel cell or just fuel cel l . Depending on the type o f the electrolyte used, there are several types of fuel cel l s :

• Alkaline fuel cells (AFC) use concentrated (85 wt per cent) KOH as the electrolyte for h igh temperature operation (250 DC) and l ess concentrated (35-50 wt per cent) for l ower temperature operation « 1 20 DC) . The e lectro­lyte is . retained in a matrix (usual ly asbestos), and a wide range of e lectrocatalysts can be used

,

S HERIF et al. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION S7

(such as Ni , Ag, metal oxides, and noble metals) . This fuel cell is intolerant to · CO2 present in either the fuel or the oxidant64.

• Polymer electrolyte membrane or proton exchange membrane fuel cel ls (PEMFC) use a thin polymer membrane (such as perfluorosulfonated acid polymer) as the electrolyte. The membranes as thin as 1 2-20 microns have been developed, which are excellent proton conductors. The catalyst i s typical ly platinum with loadings about 0.3 mg/cm

2, or, if the hydrogen feed contains

minute amounts of CO, Pt-Ru al loys are used. Operating temperature is usual ly below 1 00 DC, more typical ly between 60 and 80 DC.

• Phosphoric ac id fuel cel ls (PAFC) , use concentrated phosphoric acid (- 1 00 per cent) as the electrolyte. The matrix used to retain the acid is usually SiC, and the electrocatalyst in both the anode and cathode is Platinum black. Operating temperature is typical ly between 1 50-220 DC64. 65 .

• Molten carbonate fuel cel l s (MCFC) have the electrolyte composed of a combination of alkal i (Li , Na, K) carbonates, which is retained in a ceramic matrix of LiA102. Operating temperatures are between 600-700DC where the carbonates form a highly conductive molten salt, with carbonate ions providing ionic conduction . At such high operating temperatures, noble metal catalysts are typical ly not required64. 65.

• Sol id oxide fue l cells (SOFC) use a solid, nonporous metal oxide, usually Y20rstabi l ized Zr02 as the electrolyte. The cell operates at 900-1 000 DC where ionic conduction by oxygen ions takes place64. 65 .

A typical fuel cell consists of the electrolyte, in contact with porous electrodes, on both s ides. Electrochemical reactions occur at a three-phase interface - porous electrode/electrolyte/reactants. The actual electrochemical reactions that occur in the above l i sted types of fuel cells are different although the overall reaction is the same, i e, H2 + '1202 -7 H20. Low temperature fuel cel ls (AFC, PEMFC, PAFC) require noble electrocatalysts to achieve practical reaction rates at the anode and cathode. High temperature fuel cel ls (MCFC and SOFC) can

also ut il ize CO and CH4 as fuels . The operating temperature is high enough so that CO and CH4 can be converted into hydrogen through the water-gas shift and steam reforming reactions, respectively.

The reversible potential changes with temperature and pressure; in general i t i s lower at h igher temperatures (reaching - 1 .0 V at 1 000 K), and higher at h igher pressures or higher concentrations of reactants. The actual voltage of an operational fue l cel l i s always lower than the reversible potential due to various i rreversible losses (such as activation polarization), concentration polarization, and ohmic resistance. While ohmic resistance is directly proportional to the current, act ivation polarization i s a logarithmic function o f the current, and is thus more pronounced at very low current densities. Concentration polarization is an exponential function of the current and thus becomes a l imi ting factor at high current densities. Figure 2 shows actual fuel cell polarization curves of some representative fuel cel ls . The fuel cel ls are typically operated in a range between 0.6 and 0.8 V . The Space Shuttle fuel cel l (alkal ine) is designed64 to operate at 0.86 V and 4 1 0

1 .0-,-------------------,

0.9

2: OJ 0.8 OJ fl o > Qj 0.7 ()

0.6

MCFC �

PEMFC

0.5-t----t------1i-----t----+----l

o 200 400 600 800 1 000

Current Density (mAlcm2)

Figure 2 - Polarization curves of some representative fuel cel ls . AFC-United Technologies' fuel cel l operating in Space Shuttle, with H2 and O2 at 80-90 "C, 4 1 0 kPA, MCFC-ERC' s atmospheric pressure fuel cel l operating with natural gas ( 1 994) . PAFC-IFC ' s fuel cel l operating in ONSI PC25 200 kW power plant, operating on reformed natural gas and air at atmospheric pressure ( 1 996) (upper l ine is I FC's PC23 operating at 820 kPa), PEMFC-Ballard ' s 25 kW fuel cell operating in Daimler-Benz vehicle (NECAR I I ) with H2 and air ( 1 996), and SOFC­Westinghouse fuel cel l operat ing at atmospheric pressure using natural gas ( 1 99 1 )

58 J SCI IND RES VOL 62 JAN UARY-FEBRUARY 2003

end plate

bus plate bi-polar separator plates

+ water out

Figure 3 - Schematic representation of a typical fuel cel l stack

mA/cm2. PEM fuel cel ls have the highest achievable

current densi ties, between I and 2 mAlcm2

at 0.6 V with pressurized hydrogen and air. In order to get useable vol tages (i e, tens or hundred V), the cells are combined in a stack. The cel l s are physically separated from each other and electrical ly connected in series by a bipolar separator plate. Figure 3 shows a schematic representation of a typical fuel cell stack.

En.ergy Conversions Involving Metal Hydrides

Hydrogen' s property to form metal hydrides may be used not only for hydrogen storage but also for various energy conversions . When a hydride is formed by the chemical combination of hydrogen with a metal, an element or an alloy, heat i s generated, i .e, the process i s exothermic. Conversely, in order to release hydrogen from a metal hydride, heat must be supplied. These processes can be represented by the fol lowing chemical reactions :

Charging or absorption:

M + XH2 � MH2x + heat . . . (7)

Discharging or desorption :

MH2x + heat � M + XH2 . . . (8)

where M represents the hydriding substance, which could be a metal, an e lement, or an al loy. The rate of these reactions increases with an increase in the surface area. Therefore, in general, the hydriding substances are used in powdered form to speed up the reactions. Elements or metals with unfil led shel l s or

subshells are suitable hydriding substances . Metal and hydrogen atoms form chemical compounds by sharing their electrons in the unfi l led subshel l s of the metal atom and the K shells of the hydrogen atoms.

Ideally, for a given temperature, the charging or absorption process and the discharging or desorption process takes place at the same constant pressure. However, actually, there is a hysteresis effect and the pressure is not absolutely constant - for a given temperature charging pressures are higher than the discharging pressures. The heat generated during the charging process and the heat needed for discharging are functions of the hydriding substance, the hydrogen pressure and the temperature at which the heat is supplied or extracted. Using different metals and by forming different alloys, different hydriding characteristics can be obtained. In other words, it is possible to make or find hydriding substances that are more suitable for a given appl ication, such as waste heat storage, electric i ty generation, pumping, hydrogen purification, and isotope separation.

Hydriding substances can be used for electricity storage in two ways. In one of the methods, electricity (direct current) i s used to electrolyze the water, and the hydrogen produced is stored in a hydriding substance. When electricity is needed, the hydrogen is released from the hydriding substance by adding heat and used in a fuel cel l to produce direct current electric ity. Heat from fuel cell can be used to release hydrogen from the metal hydride. In the second method, one electrode is covered with a hydriding substance (e g, t i tanium nickel alloy). During the electrolysis of water, the hydriding substance covering the electrode i mmediately absorbs the hydrogen produced on the surface of the electrode. Then, when electricity is needed, the electrolyzer operates in a reverse mode as a fuel cel l producing electrici ty using the hydrogen released from the metal hydride.

Hydrogen Safety

Like any other fuel or energy carrier hydrogen poses risks if not properly handled or controlled. The rilsk of hydrogen, therefore, must be considered relat ive to the common fuels such as gasol ine, propane or natural gas . The specific physical characteristics of hydrogen are qui te di fferent from those common fuels . Some of those properties make hydrogen potential ly less hazardous, whi le other

r

S HERIF el at. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 59

hydrogen characteristics could theoretically make it more dangerous in certain situations.

S ince hydrogen has the smal lest molecule i t has a greater tendency to escape through smal l openings than other l iquid or gaseous fuels . Based on properties of hydrogen such as density, viscosity and diffusion coefficient in air, the propensity of hydrogen to leak through holes or joints of l ow pressure fue l l ines may be only 1 .26 to 2.8 t imes faster than a natural gas leak through the same hole (and not 3.8 times faster as frequently assumed based solely on diffusion coeffic ients). Experiments have indicated that most leaks from residential natural gas l ines are laminar66. S ince natural gas has over three t imes the energy density per unit volume the natural gas leak would result in more energy release than a hydrogen leak. For very large leaks from h igh­pressure storage tanks, the leak rate is l imited by the sonic speed. Due to higher sonic velocity ( 1 308 m/s) hydrogen would initial ly escape much faster than natural gas ( sonic velocity of natural gas is 449 rnIs) . Again, since natural gas has more than three t imes the energy density than hydrogen, a natural gas leak wi l l always contain more energy.

Some h igh strength steels are prone to hydrogen embrittlement. Prolonged exposure to hydrogen, particularly at high temperatures and pressures, can cause the steel to lose strength, eventual ly leading to fai lure . However, most other construction, tank and pipe material s are not prone to hydrogen embrittlement. Therefore, with proper choice of materials, hydrogen embrittlement should not contribute to hydrogen safety risks.

If a leak should occur for whatever reason, hydrogen wi l l disperse much faster than any other fuel, thus reducing the hazard levels . Hydrogen is both more buoyant and more diffusive than e ither gasoli ne, propane or natural gas . A hydrogen/ai r mixture can bum in relatively wide volume ratios, between 4 and 75 per cent of hydrogen in air. The other fuels have much lower flammabi l ity ranges, viz, natural gas 5 .3- 1 5 per cent, propane 2. 1 - 1 0 per cent, and gasol ine 1 -7.8 per cent. However, the range has a l ittle practical value. In many actual leak situations the key parameter that determines if a leak would ignite i s the lower flammabi l ity l imit, and hydrogen ' s lower flammabi l i ty l imit i s 4 times higher than that of gasol ine, 1 .9 t imes higher than that of propane and sl ightly lower than that of natural gas . It also has a

very low ignition energy (0.02 m1) , about one order of magnitude lower than other fuels . The ignition energy is a function of fue l/air ratio, and for hydrogen it reaches minimum at about 25-30 per cent. At the lower flammabi l ity l imit hydrogen ignition energy is comparable with that of natural gas67 . Its flame velocity is seven times faster than that of natural gas or gasoline. A hydrogen flame would therefore be more l ikely to progress to a deflagration or even a detonation than other fuels . However, the l ikel ihood of a detonation depends in a complex manner on the exact fuel/air ratio, the temperature and particularly the geometry of the confined space. Hydrogen detonation in the open atmosphere is highly unlikely.

The lower detonabi l ity fuel/air ratio for hydrogen i s 1 3- 1 8 per cent, which is two times higher than that of natural gas and 1 2-times higher than that of gasol ine. S ince the lower flammabi lity l imit is 4 per cent an explosion i s possible only under the most unusual scenarios, e g, hydrogen would first have to accumulate and reach 1 3 per cent concentration in a closed space without ign ition, and only then an ignition source would have to be triggered. Should an explosion occur, hydrogen has the lowest explosive energy per unit stored energy in the fuel , and a given volume of hydrogen would have 22-times less explosive energy than the same volume fi l led with gasoline vapor.

Hydrogen flame is nearly invisible, which may be dangerous, because people in the vicinity of a hydrogen flame may not even know there is a fire. Thi s may be remedied by adding some chemicals that wil l provide the necessary luminosity. The low emissivi ty of hydrogen flames means that near-by materials and people wi l l be much less l i kely to ignite and/or hurt by radiant heat transfer. The fumes and soot from a gasoline fire pose a risk to anyone inhaling the smoke, while hydrogen fires produce only water vapor (unless secondary materials begin to burn) .

Liquid hydrogen presents another set of safety i ssues, such as risk of cold bums, and the increased duration of leaked cryogenic fue l . A large spi l l of l iquid hydrogen has some characteristics of a gasoline spi l l , however it wi l l dissipate much faster. Another potential danger is a violent explosion of a boi ling l iquid expanding vapor in case of a pressure relief valve fai lure .

60 J SCI IND RES VOL 62 JANUARY-FEBRUARY 2003

In conclus ion, hydrogen appears to pose risks of the same order of magnitude as other fuels . In spite of publ ic percept ion, in many aspects hydrogen is actually a safer fuel than gasol ine and natural gas . As a matter of fact, hydrogen has a very good safety record, as a consti tuent of the "town gas" widely used in Europe and USA in the 1 9th and early 20th century, as a commercial ly used i ndustrial gas, and as a fuel in space programs. There have been accidents, but noth ing that would characterize hydrogen as more dangerous than other fue ls .

One of the most remembered accidents involving hydrogen is the Hindenburg dirigible disaster in 1 937. However, hydrogen did not cause that accident and hydrogen fire did not directly cause any casualties. The accident appears to be caused by static electricity discharge, and it was the bal l oon ' s l ining that caught fire first6R . Once hydrogen that the bal loon was fi l led with for buoyancy ( instead of he l ium as it was original ly designed to be fi l led with) was ignited i t burned (as any fuel is supposed to). However, hydrogen fire went straight up and it did not radiate heat so the people in the gondola underneath the balloon were not burned or suffocated. As a matter of fact 56 survivors walked out of gondola once it landed after a l l the hydrogen and bal loon structure burned out. Therefore, even in a worst case scenario accident, hydrogen proved to be a safer fuel .

References Bockris, J O'M, Energy: Ihe solar-hydrogell allemQlive ( Halsted Press, New York) 1 975 .

2 Bockris J O' M & Veziroglu, T N, A solar-hydrogen energy system for environmental compatibi l i ty, Enviroll Conserv,

12 (2) ( 1 985) 1 05- 1 1 8 . 3 Bockris J O'M, Veziroglu T N, & Smith 0, Solar hydrogen

e/ll'l'g.\': Ihe power 10 save Ihe earlh (Optima, London) 1 99 1 . 4 Veziroglu TN & Barbir F, Hydrogen: the wonder fuel, Inl J

Hydr Energy, 1 7 (6) ( 1 992) 39 1 -404. 5 Goebel E, Coveney R M (1r), Angino E E, Zeller E J &

DrescholT G A M, Geology, composition, isotopes of natural ly occurring H2/N2 rich gas from wel l s near junction city, Kansas, Oil Gas J, (May 1 994) 2 1 7-222.

6 Heydorn B, SRI CO/lSulling chemical economics handhook ( SRI. Menlo Park, CAl 1 998.

7 Bonner M, Botts T, McBreen J, Mezzina A, Salzano F & Yang C, Status of advanced electrolytic hydrogen production in the united states and abroad, Inl J Hydr EneI', 9 (4) ( 1 984) 269-275.

8 Dulla S, Technology assessment of advanced electrolytic hydrogen production, 1111 J Hydr Energy, 15 (6) ( 1 990) 379-386.

9 Wendt H, Water Splitting Methods, in Hydrogen as an energy carrier, edited by C-J Winter and J Nitsch (Springer­Verlag, Berl in Heidelberg) 1 988, pp 1 66-238 .

10 Liepa M A & Borhan A, High-temperature steam electrolysis : technical and economic evaluation of alternative process designs, lilt J Hydr EneI', 1 1 (7 ) ( 1 986) 435-442.

I I Hancock 0 G (1r), A photovoltaic-powered water electrolyzeI': its performance and economics, in Hydrogell energy progress V, edited by T N Veziroglu and J B Taylor (Pergamon Press, New York) 1 984, pp 335-344.

1 2 Carpetis C, An assessment of electrolytic hydrogen production by means of photovoltaic energy conversion. Int

J Hydr EneI', 9 ( 1 2) , ( 1 984) 969-992. 1 3 Siegel A & Schott T, Optimization or photovoltaic hydrogen

production, lilt J Hydr Eller, 13 ( I I ) ( 1 988) 659-678. 1 4 Steeb H , Brinner A , Bubmann H & Seeger W , Operation

experience of a 1 0 kW PV-electrolysis system in di ffercnt power matching modes, in Hydrogen energy progress VIII.

Vol 2, edited by T N Veziroglu and P K Takahashi (Pergamon Press, New York) 1 990, pp 69 1 -700.

I S Blank H & Szyszka A , Solar Hydrogen. dcmonstration plant in Neunburg vorm Wald, in Hydrogen encrgy progress IX.

Vol 2, edited by T N Veziroglu C Derivc and J Pottier (M C I Paris) 1 992, pp 677-686.

1 6 Grasse W, Oster F & Aba-Oud H, HYSOLAR: The German­Saudi Arabian program on solar hydrogen-5 years or experience, lilt J Hydr EneI', 17 ( I ) ( 1 992) 1 -8.

1 7 Lehman P & Chamberlain C E, Design of a photovoltaic­hydrogen-fuel cel l energy system, Inl J Hydr EneI', 16 ( 5 ) ( 1 99 1 ) 349-352 .

1 8 Lund P 0, Optimization of standalone photovoltaic systcm with hydrogen storagc ror total energy sclf-sufficiency, lilt J Hydr EneI', 16 ( I I ) ( 1 99 1 ) 735-740.

1 9 Garcia-Conde A G & Rosa F, Solar hydrogen production : a Spanish experience, Hydrogen energy progress IX, Vol 2, edited by T N Veziroglu , C Derive and J Pottier (M C I , Paris) 1 992, p p 723-732.

20 Baykara S Z & Bi lgen E, An overall assessment of hydrogen production by solar water thermolysis, Int J Hvdr t'ncr, 14 ( 1 2) ( 1 989) 88 1 -889.

2. 1 Engels H el ai, Thermochemical hydrogen production, Inl J Hydr Eller, 12 (5 ) ( 1 987) 29 1 -295.

22 Yalcin S, A review of nuclear hydrogen production, 1111 .I Hydr EneI', 14 (8) ( 1 989) 55 1 -56 1 .

23 Bul l S R, Hydrogen production by photoprocesses, Proc Inl Renew EneI' Conf (Honolulu, Hawaii) 1 988 , pp 4 1 3-426.

24 Wil lner I & Steinberger-Wil lner B, Solar hydrogen production through photo-biological, photochemical and photoelectrochemical assemblies, Inl J Hvdr EneI', 13 ( 1 0) ( 1 988) 593-604.

25 National Hydrogen Association, The hydrogen lechnology assessment, Phase I, A Report for NASA, Washington, 0 C. 1 99 1 .

26 Taylor J B, Alderson J E A, Kalyanam K M, Lyle A B & Phi l l ips L A, Technical and economic assessment of methods for the storage of large quantities of hydrogen, Inl J Hydr Ener, I I ( I ) ( 1 986) 5-22.

..

S HERIF et £II.: HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 6 1

27 Carpetis C, Storage, Transport & Distribution of hydrogen, in Hydrogen as an energy carrier, edited by C-J Winter and J Nitsch (Springer-Verlag, Berlin Heidelberg) 1 988, pp 249-289.

28 Pottier J D & B londin E, Mass storage of hydrogen, in Hydrogen energy system, utilization of hydrogen and future aspects. edited by Y Yurum, NATO ASI Series E-295 (Kluwer Academic Publishers, Dordrecht, The Netherlands) 1 995, pp 1 67- 1 80.

29 M it l itsky F, Development of an advanced, composite, lightweight, high pressure storage tank for on-board storage of compressed hydrogen, Proc Fuel Cells Transp TOPTEC: Addressing the Fuel Infrastructure Issue ( Alexandria, V A, SAE, Warrendale, PAl 1 996.

30 Veziroglu T N, Hydrogen technology for energy needs of human settlements, fnt J Hydr Ener, 12 (2) ( 1 987) 99- 1 29.

3 1 Schwartz J A & Amankwah K A G, Hydrogen storage systems, in The future of energy gases edited by D G Howell , U S Geological Survey Professional Paper 1 570 ( U S Government Printing Office, Washington) 1 993, pp 725-736.

32 Chambers A, Park C, Baker R T K & Rodriguez N M , Hydrogen storage in graphite nano-fibers. J Pllys Chem B , 102 (22) ( 1 988) 4253-4259.

33 Dil lon A C, Jones K M & Heben M J, Carbon nanotube materials for hydrogen storage, Proc U S DOE Hydr Progr Rev, Vol I I (National Renewable Energy Laboratory, Golden, CO) 1 996, pp 747-763,

34 Rambach G & Hendricks C, Hydrogen transport and storage in engineered glass rnicrospheres, Proc U S DOE Hydr Progr Rev, Vol " (National Renewable Energy Laboratory,. Golden, CO) 1 996, pp 765-772.

35 Jensen C, Hdrogen storage via polyhydride complexes, Proc U S DOE Hydr Progr Rev, Vol " (National Renewable Energy Laboratory, Golden, CO) 1 996, pp pp 787-794.

36 B rewer G D, The prospects for l iquid hydrogen fueled aircraft, fnt J Hydr Ener, 7 ( 1 982) 2 1 -4 1 .

37 Winter C J & Nitsch J , Hydrogen as an energy carrier (Springer-Verlag, Berl in) 1 988.

38 Huston E L, Liquid and sol id storage of hydrogen, Hydrogen energv progress V (Toronto, Ontario, Canada) 1 984.

39 Sherif S A, Lordgooei M & Syed M T, Hydrogen l iquefaction, in Solar hydrogen energy system edited by T N Veziroglu. Final Technical Report (Clean Energy Research Insti tute, University of M iami . Coral Gables, Florida) August 1 989, pp C I -C I 99.

40 Hands B A. Cryogenic egineerillg ( Academic Press, New York) 1 986.

4 1 Newton C L , Hydrogen poduction, l iquefaction and use, Cryogen Eng News, Part I (No 8) ( 1 967) pp 50-60.

42 Newton C L, Hydrogen production, l iquefaction and use, CI)'ogen Eng News. Part II (No 9) ( 1 967) pp 24-29.

43 Baker C R & Shaner R L, A Study of the efliciency of hydrogen liquefaction, Int J Hydr Ener, 3 ( 1 978) 3 2 1 -334.

44 Scott R B , Cryogenic engineering (Van Nostrand Company, Inc, Princeton, New Jersey) 1 962.

45 Ewe H H & Selbach H J , The storage of hydrogen, in A solar hydrogen energy system edited by W E Justi (Plenum Press, London, United Kingdom) 1 987.

46 Barclay J A, Magnetic refrigeration for low-temperature applications, Proc Cryocoo{er ConI. Colorado, September 1 984.

47 Barclay J A, Overton W C (Jr) & Stewart W F, Magnetic reliquefication of LH2 storage tank boi l-off, Phase I, Final Report, Kennedy Space Center, Ref No PT-SPD/6037/25 I I CIOOOOOO/04/82, (NASA-Defense Purchase Req CC-22 1 63B), 1 983.

48 Barclay J A, The theory of an active magnetic regenerative refrigerator, NASA-R-2287, 1 983.

49 Barclay J A & S teyert W A, Materials for magnetic refrigeration between 2 K and 20 K, Cryogell, 22 (February 1 982) 73-79.

50 Barclay J A & Stewart W F, The effect of parasitic refrigeration on the efficiency of magnetic l iquefiers, Proc fECEC '82: Seventeenth flltersociety EneI' COllverv Eng COllI. Los Angeles, Cal ifornia, August 1 982, pp 1 1 66.

5 1 Barclay J A, Use of a ferrofluid as the heat-exchange fluid in a magnetic refrigerator, J App/ Phys, 53 (4) ( 1 982) 2887.

52 B arclay J A & Sarangi S , Selection of regenerator geometry for magnetic refrigerator applications, Proc ASME Winter Alln Meet, New Orleans, LA, December 1 984.

53 Barclay J A, Stewart W F, Overton W C, Candler R C, & Harkleroad 0 D, Experimental results on a low-temperature magnetic refrigerator, Cryogen Eng Conj; Boston, MA, August 1 985.

54 Brown N S, Advanced long term cryogenic storage systems, NASA M arshall Space Flight Center, H untsvi l le, AL, NASA Conf Pub 2465P, Washington, DC, 1 986, pp 7- 1 6.

5 5 B arron R F , Cryogenic systems (Oxford University Press, Oxford, United K ingdom) 1 985.

56 Kandebo S W, Researchers explore slush hydrogen as fuel for national aero-space plane, Aviation Week & Space Technol, (26 June 1 989) pp 37-38.

57 Baker C R & M atsch L C, Production and distribution of l iquid hydrogen. Adv Petr Chem Refin, 10 ( 1 965 ) 37-8 1 .

58 Sindt C F, A summary of the characterization study of slush hydrogen, Cryogen, 10 (October 1 970) 372-3 80.

59 Voth, R 0, Producing liquid-solid mixtures (slushes) of oxygen or hydrogen using an auger, Cryogen, 25 (September 1 985) 5 1 1 -5 1 7.

60 Pottier J D, Hydrogen transmission for future energy systems, in Hydrogen Energy System, Utilizatioll of Hydrogen and Future Aspects edited by Y Yurum, NATO ASI Series E-295 ( Kluwer Academic Publishers, Dordrecht, The Netherlands) 1 995, pp 1 8 1 - 1 94.

6 1 Oney F, The comparison of pipelines transportation of hydrogen and natural gas, M S Thesis, University of Miami , 1 99 1 .

62 Norbeck J M , Heffel J W, Durbin T D, Tabbara B , Bowden J M & Montano M C, Hydrogen fuel for surface transportation ( S AE, Warrendale, PAl 1 996.

62 J SCI IND RES VOL 62 JANUARY-FEBRUARY 2003

63 Sternfeld H J & Heinrich P, A demonstration plant for the hydrogen/oxygen spinning reserve, Int J Hydr Ener, 14

( 1 989) 703-7 1 6.

66 Thomas C E, Preliminary hydrogen vehicle safety report (The Ford Motor Company, Contract No DE-AC02-94CE50389, U S Dept of Energy) 1 996.

64 Kinoshita K. McLarnon F R & Cairns E J, Fuel cells: a handbook (U S Department of Energy, DOE/METC88/6069, Morgantown. WV) 1 988 .

65 B lumen L J M J & Mugerwa M N (eds), FlIel cell systems (Plenum Press, New York) 1 993.

67 Swain M R & Swain M N, A comparison of' H2, CH4, and C3HK fuel leakage in residential settings, II/t .I Hydr Ener, 17 ( 1 0) , ( 1 992) 807-8 1 5 .

68 DiChristina M , What real ly downed the hindenburg, Pop Sci, ( November 1 997).

Dr S A Sherif is a tenured Professor of Mechanical and Aerospace Engineering and is the Associate Director of the Industrial Assessment Center and the Founding Director of the Wayne K and Lyla L Masur HVAC Laboratory at the University of Florida. He served on the faculties of the University of FLorida ( 1 991-presellt), University of Miami ( 1987- 1991), and Northern IlLinois University ( 1 984-1 987). He is a Faculty Fellow of NASA ­Marshall, NASA-Kennedy, Argollne National Laboratory, and the Arnold Engineering Deveiopillent Center. He edited 2 1 bound volumes and published 1 0 book chapters, 100 refereed journal papers, alld l30 conj'erence pa/lers and technicaL reports. He served on over 30 technical and administrative commillees and advisory bOQ/·d.r of ASME, ASHRAE, and AIAA. He served on numerous university commilleeslcollncils, and delivered 40 invited/keynote lectures in international conferences, workshops, or short courses. He is a reviewer to 35 jou1'l1als, 10 publishing companies, 1 8 funding agencies, and more than 150 conference proceedings, books, standards. and

handbook chapters. He also served as a guest editor to several journaLs. He is a Fellow of ASME and ASHRAE, an Associate Fellow of AIAA, alld a member of AIChE, International Association for Hydrogen Energy (IAHE), International Solar Energy Society ( lSES), Ill tenzational lnstitute of Refrigeration (IIR), American Society for Engineering Education (A SEE), and Sigma Xi. He is 11 member of the Executive Commiflee of the ASME Advanced Energy Systems Division (/999-2004) and is the Division Chairman (2002-2003). He is also a member of A SME '.I' Energy Resources Board (2001-2003) and is the current Steering Commiflee Chairman of the Intersociety (now International) Energy Conversion Engineering Conference ( lECEC). He served as a Member-at-Large for Honors and A wards for the Advanced Energy Systems Division (1999-2000), as a Division SecretQ/Y (2000-2001), as a Division Vice Chairman (2001-2002), alld as a member of the ASME's Special Awards COlllmiflee of the Fluids Engineering Division ( 1995-2001). He is a Book Review Editor of ASME's Applied Mechanics Reviews (200l -present). He is a past chairman of the Coordinating Group 011 Fluid Measurements of the Fluids Engineering Division, a past vice chairman of ASHRAE's Thermodynamics and Psychrometrics Commillee, and a past secretary of ASHRAE's Liquid-to-Refrigerant Heat Transfer Commillee. Dr Sherif also chaired the ASHRAE Standards Project COlllllliflee that oversaw revising the standard on measurement of moist air properties ( 1 989-1 994). Dr Sherif serves on K- f9 (Commiflee on fnvirolllnental Heat Transfer) and K-6 (Commillee on Heat Transfer in Energy Systems) of the ASME's Heat Transfer Division. He is the recipient of three Certificates of Recognition for Research Contributions from NASA, a Premium of Academic Excellence (PA CE) award

ji'om Iowa State University, and four research awards from the University of Miami and a "Graduate Research and A rtistl)' A ward" from Northern Illinois University. He is also the recipient of the E K Campbell A ward of Merit from ASHRAE in f997for "outstanding service qlld achievement in teaching " and a Teaching Incentive Program (TIP) award from the University of Florida in 1998 for sustained excellence in teaching. In 2000, he was selected by the UF College of A rts and Sciences as an Anderson/CLAS Scholar Faculty Honoree for excellence in teaching and in 2001 was awarded the Kuwait Prize in Applied Sciences, the highest award bestowed on any individual hy Kuwait. He is listed in //lore than 70 who 's who publications. He can be reached at: Dr S A Sherif, Professor of Mechanical and Aerospace Engineering, Director, Wayne K. and Lyla L. Masur HVA C Laboratory, Associate Director, Industrial Assessment Center, Department of Mechanical and Aerospace Engineering, University of Florida, P.O. Box 1 16300, 228 MEB, Gainesville, Florida 326 1 1 -630(}, Tel (352) 392-7821/Fax (352) 392-107 I . e-mail: sasheri(@ufl.edu

Dr Frano Barbir is Director of Fuel Cell Technology and Chief Scientist at Protoll Energy Systems, Rod)' Hill, Connecticut. In the last ten years he has been actively invoLved in development of the PEM fuel cells and fuel cell systems. Concurrently with working in industry he was aLso associated with the University of Miami as an adjunct professor at the Department of Mechanical Engineering and as a project direclOr with the Clean Energy Research Institute. He is a member of the International Association for Hydrogen Energy, the Electrochemical Society, and the American Solar Energy Society. He has more than 120 publications on hydrogen and fuel cells in technical and scientific journals, books, encyclopedias, and conference proceedings to his credit. He is an associate editor of the International Journal of Hydrogen Energy and serves on the editorial boards or as a reviewer for severaL technical and scientific joumals. He has a Ph D in Mechanical Engineering

from the University of Miami, specializing in hydrogen energy, and an M Sc in Chemical/Environmental Engineering alld a Dipl -Ing in

Mechanical Engineering, both from the University of Zagreb, Croatia. He call be reached at: Dr Frano Barbir, ProlOn Energy System.\',

50 Inwood Road, Rocky Hill, CT 06067, USA, Tel. 860-571 ,6533 ext. 427, Fax 860-571 -6505, e-mail: fharbir@/Jrotonenerl.!\,.col/l

S HERIF et al. : HYDROGEN ENERGY PRODUCTION, STORAGE AND UTILIZATION 63

Dr T Nejat Veziroglu is Professor and Director, Clean Energy Research Institute (CERI) of the University of Miami, Coral Gables, Florida, USA. He organised the first major conference on hydrogen energy: The Hydrogen Economy Miami Energy (THEME) Conference, Miami Beach, March 1974. Subsequently, several conferences and symposia have been organised by CERI on energy and environment related subjects such as Hydrogen Energy, Remote Sensing Applied to Energy Problems, Ocean Energy Systems, Multi-phase Flows, Biosphere: Problems and Solutions, and Energy and Environment. His research interests are renewable energy sources in general and hydrogen energy in particular. He has published some 300 scientific reports and papers, edited 200 proceedings volumes, and is the Editor-in-Chief of International Journal of Hydrogen Energy and International Journal of Energy Environment Economics. He has memberships in 18 scientific organizations, elected to the grade of Fellow

in the British Institution of Mechanical Engineers, American Society of Mechanical Engineers, and is the Founding President of the International Association for Hydrogen Energy.