PROSPECTS FOR BUILDING A HYDROGEN ENERGY INFRASTRUCTURE

Joan M. Ogden
Center for Energy and Environmental Studies, Princeton University, Princeton, New Jersey 08544; e-mail: [email protected]

Key Words alternative fuels, fuel cells, energy transmission and distribution Abstract About two-thirds of primary energy today is used directly as transportation and heating fuels. Any discussion of energy-related issues, such as air pollution, global climate change, and energy supply security, raises the issue of future use of alternative fuels. Hydrogen offers large potential benefits in terms of reduced emissions of pollutants and greenhouse gases and diversified primary energy supply. Like electricity, hydrogen is a premium-quality energy carrier, which can be used with high efficiency and zero emissions. Hydrogen can be made from a variety of feedstocks, including natural gas, coal, biomass, wastes, solar sources, wind, or nuclear sources. Hydrogen vehicles, heating, and power systems have been technically demonstrated. Key hydrogen end-use technologies such as fuel cells are making rapid progress toward commercialization. If hydrogen were made from renewable or decarbonized fossil sources, it would be possible to have a large-scale energy system with essentially no emissions of pollutants or greenhouse gases. Despite these potential benefits, the development of a large-scale hydrogen energy infrastructure is often seen as an insurmountable technical and economic barrier. Here we review the current status of technologies for hydrogen production, storage, transmission, and distribution; describe likely areas for technological progress; and discuss the implications for developing hydrogen as an energy carrier. CONTENTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 Motivations for Developing Hydrogen as a Fuel . . . . . . . . . . . . . . . . . . . . . . . . . 228 Definitions and Underlying Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229 Development of Markets for Hydrogen Energy . . . . . . . . . . . . . . . . . . . . . . . . . . 230 Units for Hydrogen Production and Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 Hydrogen Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232 Thermochemical Production Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Technologies for Sequestering Carbon During Thermochemical Hydrogen Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236 Electrolysis of Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237 Summary: Economic Comparison of Hydrogen Production Methods . . . . . . . . . . 239 1056-3466/99/1022-0227$12.00 227
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
228 OGDEN
Experimental Methods of Hydrogen Production . . . . . . . . . . . . . . . . . . . . . . . . . 240

Hydrogen Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241 A. Large-Scale Stationary Storage of Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . 241 B. Stationary Storage at Intermediate and Small Scales . . . . . . . . . . . . . . . . . . . 242 C. Storing Hydrogen on Board Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

D. Novel Approaches to Hydrogen Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245 E. Summary of Storage Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246 Hydrogen Transmission, Distribution, and Delivery . . . . . . . . . . . . . . . . . . . . 246 A. Description of the Current Industrial-Hydrogen Transmission and Distribution System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 B. Long-Distance Transmission of Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 C. Local Pipeline Distribution of Hydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249 D. Gaseous-Hydrogen Refueling Stations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Design and Economics of Hydrogen Energy Systems . . . . . . . . . . . . . . . . . . . 253 Estimating the Demand for Hydrogen Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . 254 Selecting the Lowest-Cost Hydrogen Supply Option: General Considerations . . . . 255 Estimating the Delivered Cost of Hydrogen Transportation Fuel: A Southern California Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 Capital Cost of Hydrogen Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 Hydrogen Infrastructure Capital Costs Compared with Those for Methanol, Gasoline, and Synthetic Middle Distillates . . . . . . . . . . . . . . . . . . . . 262 Lifecycle Cost of Automotive Transportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262 Environmental and Safety Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264 Emissions of Greenhouse Gases and Air Pollutants . . . . . . . . . . . . . . . . . . . . . . . 264 Resource, Land, and Water Use for Hydrogen Production . . . . . . . . . . . . . . . . . . 266 Safety Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266 Possible Scenarios for Development of Hydrogen Infrastructure . . . . . . . . . . 268

INTRODUCTION Motivations for Developing Hydrogen as a Fuel

Combustion of fluid fuels for transportation and heating contributes over half of all greenhouse gas emissions and a large fraction of air pollutant emissions. Continued reliance on current fuels and end-use technologies poses significant challenges in air pollution, greenhouse gas emissions, and energy supply security, particularly in the transportation sector. A variety of alternative fuels have been proposed to address these problems, including methanol, ethanol, methane, synthetic liquids from natural gas or coal, and hydrogen. Of these, hydrogen offers perhaps the greatest potential benefits in terms of reduced emissions of pollutants and greenhouse gases and diversified primary energy supply, but it is widely perceived as posing the largest technical and economic challenges. Like electricity, hydrogen is a high-quality energy carrier, which can be used with very high efficiency and zero or near-zero emissions at the point of use. It has
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
HYDROGEN ENERGY INFRASTRUCTURE 229

been technically demonstrated that hydrogen can be used for transportation, heating, and power generation, and it could replace current fuels in all their present uses. Low-temperature fuel cells, which operate on hydrogen or hydrogen-rich gases, are undergoing rapid development worldwide for stationary power and vehicle applications, with commercialization planned within the next 510 years (1).

Eight major automakers have announced plans to commercialize fuel cell vehicles in the 20042005 timeframe. If low-cost fuel cell vehicles were successfully developed, this could encourage greater use of hydrogen. Hydrogen can be made from a variety of widely available feedstocks, such as natural gas, coal, biomass, wastes, solar resources, wind, and nuclear resources. If hydrogen could be made from nonfossil energy sources or decarbonized fossil sources with separation and sequestration of CO2, it would be possible to have a large-scale energy system with essentially no emissions of air pollutants (e.g. nitrogen oxides, sulfur oxides, particulates, and hydrocarbons) or greenhouse gases during fuel production or use. The idea of a hydrogen economy or large-scale hydrogen energy system has been explored several times, first as a complement to a largely nuclear electric energy system (where hydrogen was produced electrolytically from off-peak nuclear power) and later as a storage mechanism for intermittent renewable electricity, such as photovoltaics and wind power (25). More recently, the idea of a hydrogen energy system based on production of hydrogen from fossil fuels with separation and sequestration of byproduct CO2 has been proposed (68). Despite the potential attractions of a zero-emission hydrogen energy economy, the development of hydrogen energy infrastructure is often seen as an insurmountable technical and economic barrier to the use of hydrogen as an energy carrier. The prevailing wisdom is that development of a hydrogen infrastructure will cost many times more than developing such a system for a liquid fuel. In this article, we review the current technical and economic status of technologies for hydrogen production, storage, transmission, and distribution and describe likely areas for technological progress.We estimate the cost of developing a hydrogen infrastructure, as compared with other alternatives. Finally,we discuss possible scenarios for developing hydrogen as an energy carrier. Emphasis is given to use of hydrogen fuel in transportation markets, although hydrogen can be used for electricity generation or cogeneration as well.

Definitions and Underlying Assumptions

A hydrogen energy infrastructure is defined as the system needed to produce hydrogen, store it, and deliver it to users. This includes hydrogen production systems (for converting primary-energy sources or other energy carriers to hydrogen), hydrogen storage capacity (needed to match time-varying fuel demands to production output), long-distance transmission systems (if hydrogen is to be transported long distances from the production site to users), local pipeline distribution systems (analogous to a system of natural gas utility pipes), and equipment for dispensing
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
230 OGDEN hydrogen to users (for example, hydrogen compressors and dispensers at vehicle refueling stations).

Development ofMarkets for Hydrogen Energy

Although hydrogen infrastructure rather than hydrogen end-use systems is the focus of this review, successful commercialization of hydrogen end-use systems,

such as fuel cell vehicles or fuel cell heat and power systems, is a key precondition for the development of a hydrogen infrastructure. What will drive adoption of hydrogen rather than other alternative transportation fuels? There are a variety of advanced vehicle/alternative-fuel combinations that have the potential to significantly improve fuel economy and reduce fuel cycle emissions, as compared with current vehicles. These include fuel cells, hybrid internal combustion engine-/battery-powered vehicles, and electric battery-powered vehicles. Recent studies comparing future transportation alternatives suggest that fuel cell vehicles are a promising technology for meeting future goals for zero tailpipe emissions, high efficiency, good performance, and lowcost in mass production (9). Advanced diesel/battery hybrids might also achieve high efficiency and acceptable cost, but currently emissions remain an issue, especially regarding particulates (9). (Development of new, low-sulfur fuels for diesel engines may help ameliorate this problem.) Although it is too early to pick a winner among emerging advanced transportation technologies, fuel cells are regarded as a leading contender. It is possible that fuel cellpowered vehicles will be commercialized first with onboard fuel processors (to produce hydrogen for the fuel cell from other more easily handled fuels such as methanol or gasoline), rather than with hydrogen stored directly onboard. However, hydrogen fuel cellpowered vehicles are likely to be lower cost and more efficient than those with onboard fuel processors (10, 11). The lower first cost of the hydrogen fuel cellpowered vehicle and its higher efficiency combine to give a lower lifecycle cost than fuel cellpowered vehicles run on liquid fuels. If fuel cell vehicles become widely used, there is reason to believe that the market will move toward hydrogen as a fuel. Another possible market for hydrogen is in combined heat and power applications in buildings. It has been suggested that fuel cells could become competitive in these markets first, where cost barriers are less stringent than for automotive markets. Initially, hydrogen for fuel cells would be made in natural gas reformers coupled to the fuel cells. Eventually, hydrogen might be made centrally and distributed to users in buildings. Fuel cell power systems could be made in small sizes, making them potentially attractive for distributed generation. The possibility of separating and sequestering carbon during production of hydrogen from fossil fuels is another unique potential benefit of a hydrogen energy system. Other fossil-based synthetic fuels, such as methanol or synthetic middle distillates, carry fossil carbon in the fuel, and it is eventually emitted into the atmosphere from the vehicle, which would limit the extent to which greenhouse
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
HYDROGEN ENERGY INFRASTRUCTURE 231

gas emissions could be reduced. Hydrogen production with sequestration of CO 2 gives a lever to reduce carbon emissions in the transportation sector, without curtailing the future use of fossil resources.

Units for Hydrogen Production and Use

Hydrogen production capacity is usually given in units of standard cubic feet (scf)

produced per day, normal cubic meters (Nm 3) per day, gigajoules per day, or kilowatts of hydrogen output (on a continuous basis). In this paper, specific capital costs for production plants are expressed as dollars ($) per kilowatt of hydrogen output capacity. All energy and power units are given, based on the higher heating value (HHV) of hydrogen. Hydrogen storage capacity is given in volume units (scf or Nm 3), in tons, or in energy stored (gigajoules). Capital costs for storage are given in $ per ton of hydrogen stored or $ per gigajoule stored. In this paper, the levelized cost of hydrogen production, transmission, or storage is given in $ per gigajoule of hydrogen on a higher heating value basis. (1 GJ = 109 J = 0.95 million Btu.) Table 1 contains useful conversion factors for relating these units to others and also contains physical properties of hydrogen and other fuels. To relate rather unfamiliar hydrogen production units (millions of scf/day) to more familiar quantities, we show, in Table 2, typical energy demands expressed in scf of H2 per day and gigajoules per day, ranging from hydrogen required for one
TABLE 1 Conversion factors and economic assumptionsa 1 GJ = 109 J = 0.95 million Btu 1 EJ = 1018 J = 0.95 quadrillion (1015) Btus 1 million standard cubic feet (scf) = 26,850 normal cubic meters (Nm3) = 343 GJ (HHV) 1 million scf/day = 2.66 tons/day = 3.97 MW of H2/day (based on the HHV of hydrogen) 1 scf of H2 = 343 kJ (HHV) = 325 Btu (HHV); 1 pound of H2 = 64.4 MJ (HHV) = 61.4 kBtu (HHV) = 187.8 scf 1-Nm3 = 12.8 MJ (HHV); 1 kg of H2 = 141.9 MJ (HHV) = 414 scf 1 gallon of gasoline = 130.8 MJ (HHV) = 115,400 Btu/gallon (LHV) Gasoline heating value = 45.9 MJ/kg (HHV) = 43.0 MJ/kg (LHV) $1/gallon of gasoline = $7.67/GJ (HHV) 1 gallon of methanol = 64,600 Btu/gallon (HHV) = 56,560 Btu/gallon (LHV) Methanol heating value = 22.7 MJ/kg (HHV) = 19.9 MJ/kg (LHV) $1/gallon of methanol = $15.4/GJ (HHV)
aAll

costs are given in constant 1995 dollars. The capital recovery factor for hydrogen production systems, distribution systems, and refueling stations = 15%.

Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
232 OGDEN
TABLE 2 Hydrogen demand and supply: scales of interesta H2 flow Standard cubic Demand feet/day (GJ/day) 1 fuel cell car (driven 11,000 miles/year) 109 0.038 1 fuel cell bus (driven 50,000 miles/year) 8000 2.7

10 fuel cell buses 80,000 27 100 fuel cell buses or 7000 fuel cell cars 800,000 270 1% of cars in the Los Angeles Basin 9 million 3200 H2 production at large refinery 100 million 34,300 10% of cars in the Los Angeles Basin 90 million 32,000 100% of cars in the Los Angeles Basin 900 million 320,000 Energy flow = NG flow in the Los Angeles Basin 9 billion 3,000,000
aValues

in the table have been rounded. It is assumed that a hydrogen fuel cell car has an average fuel economy of 106 miles/gallon of gasoline equivalent.

fuel cell-powered car to that required for a modest-sized fleet to full-scale use of hydrogen in transportation markets. Hydrogen production systems are also shown. Atypical refinery-scale steam methane reformer (SMR) producing 25100 million scf/day could fuel a fleet of "225,000900,000 hydrogen fuel cell-powered cars. A small-scale SMR or electrolyzer rated at 0.11.0 million scf/day could fuel a fleet of 9009000 hydrogen fuel cell cars or 14140 buses.

HYDROGEN PRODUCTION
Hydrogen is made at large scale today (mostly from natural gas) for use in chemical processes such as oil refining and ammonia production. About 1% of U.S. primary energy use ("5% of U.S. natural gas use) goes to hydrogen production for chemical applications. A variety of hydrogen production processes are commercially available today, including thermochemical methods, which are used to derive hydrogen from hydrocarbons, and electrolysis of water, during which electricity is used to split water into its constituent elements, hydrogen and oxygen. Future potential methods of hydrogen production involving direct conversion of sunlight to hydrogen in electrochemical cells or biological hydrogen production are being researched at a fundamental-science level. In this section, we describe methods of hydrogen production, the current status of and projections for technical progress, and its economics, including capital costs for production equipment, conversion efficiency, and the levelized cost of hydrogen production.
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
HYDROGEN ENERGY INFRASTRUCTURE 233

Thermochemical Production Methods

Hydrogen is made thermochemically by processing hydrocarbons (such as natural gas, coal, biomass, or wastes) in high-temperature chemical reactors to make a synthetic gas or syngas, composed of H2, CO, CO2, H2O, and CH4. The syngas is further processed to increase the hydrogen content, and hydrogen is separated

out of the mixture at the desired purity. Figure 1 shows process steps for typical hydrogen production plants based on thermochemical methods. 1. Steam Reforming of Methane Catalytic steam reforming of methane (the main component of natural gas) is a well-known, commercially available process Figure 1 Thermochemical processes for producing hydrogen.
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

234 OGDEN for hydrogen production (12, 13). In the United States, most hydrogen today (>90%) is manufactured via steam reforming of natural gas (14). Hydrogen production is accomplished in several steps: steam reforming, water gas shift reaction, and hydrogen purification (see Figure 1). The steam reforming reaction CH4 + H2O # CO + 3H2 is endothermic and requires external heat input. Economics favor reactor operation at pressures of 325 atm and temperatures of 700$C850$C. The external heat needed to drive the reaction is often provided by the combustion of a fraction of the incoming natural gas feedstock (%25%) or from burning waste gases, such as purge gas from the hydrogen purification system. After reforming, the resulting syngas is sent to one or more shift reactors, where the hydrogen concentration is increased via the water-gas shift reaction CO + H2O & CO2 + H2. The gas exiting the shift reactor contains mostly H 2 (70%80%) plus CO2, CH4, and small quantities of H2O and CO. Hydrogen is then purified. The degree of purification depends on the application. For industrial hydrogen, pressure swing adsorption (PSA) systems or palladium membranes are used to produce hydrogen at %99.999% purity. The energy conversion efficiency [=hydrogen out (HHV)/energy input (HHV)] of large-scale SMRs is perhaps 75%80%, although 85% efficiencies might be achieved with good waste heat recovery and use (15). SMRs have been built over a wide range of sizes. For large-scale chemical processes such as oil refining, steam reformers produce 25 million to 100 million scf of hydrogen/day. (In energy terms, this is enough hydrogen to power a fleet of "225,000900,000 hydrogen fuel cell cars, each driven 11,000 miles/year.) These systems consist of long (12-meter), catalyst-filled tubes, and they operate at temperatures of 850$C and pressures of 1525 atm, which necessitates use of expensive alloy steels. Capital costs for a 20-million-scf-of-H2/day steam reformer plant (including the reformer, shift reactor, and PSA) are about $200/kW of H 2 output; for a 200-million-scf/day plant, capital costs are estimated to be about $80/kW of H2 (16). Refinery type (high-pressure, high-temperature) reformers can be scaled down to as small as 0.11.0 million scf/day (the scale needed for producing hydrogen at refueling stations), but scale economies in the capital cost are significant (the capital cost is "$750/kW of H2 at 1 million scf/day and $4000/kW of H2 at 0.1 million scf/day). At small sizes, a more cost-effective approach is to use a lower pressure and temperature reformer, with lower-cost materials. SMRs in the range of 2000 to 120,000 scf of H2/day have been developed for use with fuel cells, and these have recently been adapted for stand-alone hydrogen production (17). In these systems the heat transfer path is curved, to make the device more
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
HYDROGEN ENERGY INFRASTRUCTURE 235

compact, and the reformer operates at a lower temperature and pressure (700$C and 3 atm), which relaxes materials requirements. Estimates of mass production costs for small fuel cell type SMRs indicate that the capital cost for hydrogen production plants in the 0.1-million1.0-million scf/day range would be $150 $180/kW of H2, assuming that 1000 units were produced (16). (Costs are given

on a higher-heating-value basis, and for comparison, do not include hydrogen compression, storage, or dispensing to vehicles.) The capital costs in dollars per kilowatt of hydrogen production are similar for fuel cell type small reformers and conventional, one-of-a-kind large reformers, assuming that many small units are built. Energy conversion efficiencies of 70% 80% are possible for these units. 2. Partial Oxidation of Hydrocarbons Another commercially available method for deriving hydrogen from hydrocarbons is partial oxidation (POX). Here methane (or some other hydrocarbon feedstock such as oil) is oxidized to produce carbon monoxide and hydrogen as shown by the reaction CH4 + 1/2O2 & CO + 2H2. This reaction is exothermic, and no indirect heat exchanger is needed. Catalysts are not required because of the high temperature. However, the hydrogen yield per mole of methane input (and the system efficiency) can be significantly enhanced by use of catalysts (18). A hydrogen plant based on partial oxidation includes a partial-oxidation reactor, followed by a shift reactor and hydrogen purification equipment. Large-scale, partial-oxidation systems have been used commercially to produce hydrogen from hydrocarbons such as residual oil, for applications such as those in refineries. Large systems generally incorporate an oxygen plant, because operation with pure oxygen rather than air reduces the size and cost of the reactors. Small-scale, partial-oxidation systems that use oxygen in air have recently become commercially available, but these systems are still undergoing intensive research and development (1820). Partial oxidation systems are under development by Arthur D. Little, Inc., and its spin-off company Epyx (18, 2022) and by Hydrogen Burner Technology (19). Small-scale, partial-oxidation systems have a fast response time, making them attractive for following rapidly varying loads, and these systems can handle a variety of fuels, including methane, ethanol, methanol, and gasoline. Because POX systems are more compact and do not require indirect heat exchange (as in steam reforming), it has been suggested that the costs of small partialoxidation systems could be less than those of small steam reformers. Although the partial-oxidation reactor is likely to be less expensive than a steam reformer vessel, the downstream shift and purification stages are likely to be more expensive (23). Developing low-cost purification technologies is key if POX systems are to be used for small-scale, stationary hydrogen production. Another approach is to use pure oxygen feed to the POX, which incurs high capital costs for small-scale oxygen production, but eliminates the need to deal with nitrogen downstream.
Annu. Rev. Energy. Environ. 1999.24:227-279. Downloaded from arjournals.annualreviews.org by MASSACHUSETTS INSTITUTE OF TECHNOLOGY on 04/01/07. For personal use only.

?
236 OGDEN Oxygen enrichment of incoming air is another way of reducing, but not eliminating, the amount of nitrogen. Innovative membrane technologies may allow lower-cost oxygen for POX reactors (