2. Hydrogen Economy Systems and Technology Review

This chapter reviews the hydrogen economy as currently envisioned by a variety of researchers and developers, using a systematic approach from supply through end use. Key production and end-use technologies are reviewed, including an assessment of current industry practice and challenges or opportunities within each of those elements.

A simplified system overview of the current and potential hydrogen economy is shown in Figure 2.1. The essential system elements include supply, production, distribution, dispensing, and end use. Although this overview includes a number of potential hydrogen supply and end-use scenarios, some elements may be condensed or abbreviated, depending on the particular application. For example, an early-stage implementation of hydrogen supply for hydrogen-fueled vehicles is through SMR of natural gas at the dispensing station. In that case, natural gas is delivered straight to the retail outlet from the point of supply, and the corresponding hydrogen transportation from the point of production is no more than a short pipe run.

The number of potential feedstock and production process pathways is greater than depicted in Figure 2.1. For example, electricity used in electrolysis could be grid-sourced or provided through a dedicated electric source at the point of production (i.e., wind, solar, biomass, etc.). Some potential feedstocks, such as ethanol, are themselves derived from other feedstocks and can be categorized generally with the primary feedstock source. Finally, hydrogen is also produced as a byproduct of other manufacturing processes, which could provide a hydrogen supply in addition to the hydrogen production technologies shown in Figure 2.1. This chapter considers each system element in turn, with particular emphasis on existing and future considerations with regard to production and end-use applications.

Hydrogen Supply

Hydrogen is the most abundant element in the universe. Yet, there is effectively no natural hydrogen gas resource on Earth. Hydrogen gas is the smallest and lightest of all molecules. When released, it quickly rises to the upper atmosphere and dissipates, leaving virtually no hydrogen gas on the Earth’s surface. Because hydrogen gas must be manufactured from feedstocks that contain hydrogen compounds, it is considered to be an energy carrier, like electricity, rather than a primary energy resource.

Currently, the main sources of hydrogen are hydrocarbon feedstocks such as natural gas, coal, and petroleum; however, some of those feedstocks also produce CO2. Thus, to provide overall emission savings, greenhouse gas (GHG) emissions must be mitigated during hydrogen production through CCS or similar technology, during end use through comparatively greater vehicle efficiency, or at other stages in the life cycle of the hydrogen fuel source.

In terms of fossil fuel supply, the estimated technically recoverable resource base for crude oil, natural gas, and coal in the United States in 2006 was 166 billion barrels,¹³ 1,365 trillion cubic feet,¹⁴ and 264 billion short tons,¹⁵ respectively. Those resource levels amount to 33, 74, and 280 years of supply, respectively, at U.S. production levels in 2006.¹⁶ It is generally recognized, however, that demand is not static and the accessibility of resources may be problematic. Also, the costs for addressing CO2 and other GHG emissions may increase, which could deter the full utilization of fossil fuels as a primary energy source for a hydrogen economy unless suitable mitigation measures are employed.

Hydrogen can also be produced from cellulosic biomass, through a process much like coal gasification, to produce synthesis gas that is a mixture of hydrogen and carbon monoxide, from which the hydrogen can be removed and purified. EIA’s estimate of biomass supply is as much as 10 quadrillion Btu per year in 2030. This estimate was derived in early 2007 using an integrated land and crop competition model known as POLYSYS.¹⁷ Demand for cellulosic biomass is expected to increase as a result of the renewable fuel provisions in the Energy Independence and Security Act of 2007 (EISA2007), including increased production of cellulosic ethanol and other biomass-to-liquid (BTL) fuels.^18>

Another source for hydrogen production is electrolysis of water. For decades, the National Aeronautics and Space Administration (NASA) has used this process in hydrogen fuel cells to produce both power and water for its astronauts in space. However, hydrogen production from conventional grid-based electricity is an expensive process, as discussed below, and at present it is the least carbon-neutral method for hydrogen production, given that more than 49 percent of U.S. electricity generation in 2007 was from coal-fired power plants. Reducing costs and emission impacts may be achievable through the application of CO2 mitigation measures for existing electricity generation technologies or through breakthroughs in advanced electrolysis technologies.

Options for mitigating the CO2 emissions produced when grid-based electricity is used for electrolysis include building new renewable generators (e.g., wind or biomass) and purchasing off-peak (surplus) power, presumably at low wholesale prices, from renewable generators and nuclear power plants to generate hydrogen. Each of these alternatives creates a new set of questions and challenges.

The construction of new renewable generation capacity for the exclusive purpose of producing hydrogen from electrolysis is unlikely to be desirable from an investment perspective if, in order to make the resulting hydrogen competitive, the cost of the electricity is required to be less than the wholesale price at which that electricity could be sold to the grid. The price would include any other tax credits and any Renewable Portfolio Standard (RPS) credits that might accrue if the electricity were sold to the grid. Because the value of wind-generated electricity is likely to be much higher when it is sold to the grid, investments in standalone wind systems to produce hydrogen appear to be unlikely economically. The use of biomass-generated electricity exclusively for hydrogen production would be even less attractive than wind because of higher capital costs and, unlike wind, significant feedstock costs. On the other hand, direct biomass gasification would have much better economic prospects for producing hydrogen than either wind or biomass generation if the engineering challenges of raising the maximum capacity utilization to at least 80 percent were overcome.¹⁹

Under a CO2-constrained scenario, large amounts of existing coal-fired capacity are likely to be retired, and new nuclear and renewable generators are likely to be added, to meet the CO2 emissions target. Because a CO2-constrained scenario is defined by policies that achieve a targeted level of CO2 emission reductions, any grid-based power production would already have those target CO2 emission levels factored into prices, with wind, biomass, and other power sources having been rewarded for their contributions, and higher CO2-emitting technologies having been penalized, as appropriate.

Hydrogen Production

Hydrogen production processes can be classified generally as those using fossil or renewable (biomass) feedstocks and electricity. The technology options for fossil fuels include reforming, primarily of natural gas in “on-purpose” hydrogen production plants,²⁰ and production of hydrogen as a byproduct in the petroleum refining process. Partial oxidation technologies, which can include gasification of solid or liquid feedstocks, are another option for hydrogen production. Electrolysis processes using grid or dedicated energy sources, including some advanced techniques that have not yet been proven, also can be used. Among those advanced techniques are thermochemical processes, including nuclear as an energy source. In addition, hydrogen is produced as a byproduct of some other existing industrial processes.

Significant amounts of hydrogen are produced and consumed in the United States and worldwide, using a number of commercially-proven technologies. For example, EIA estimates that the United States produced about 17 percent of the 53 million metric tons of hydrogen consumed in 2004 throughout the world.²¹ One way to appreciate the scale of the existing hydrogen economy is to consider that the 10.7 million metric tons of U.S. hydrogen production capacity would produce 1.4 quadrillion Btu at full utilization, which is equivalent to 660 thousand barrels of crude oil²² or 1.4 trillion cubic feet of natural gas per day. Appendix C provides an overview of existing hydrogen production capacity in the United States.

On-Purpose Hydrogen Production Technologies

The on-purpose hydrogen production technologies are reforming, partial oxidation (including gasification), and electrolysis. Each process has its own advantages and disadvantages with respect to capital costs, efficiency, life-cycle emissions, and technological progress.

Reforming of hydrocarbon feedstocks can be done using technologies such as the SMR process, which is the most commonly used method to supply large centralized quantities of hydrogen gas to oil refineries, ammonia plants, and methanol plants. The SMR process is popular because its natural gas feedstock has high hydrogen content (four hydrogen atoms per carbon atom) and because a distribution network for the natural gas feedstock already exists.

One benefit of SMR technology is its high degree of scalability. SMR production costs are highly dependent on the scale of production. Large, modern SMR hydrogen plants have been constructed with hydrogen generation capacities exceeding 480,000 kilograms of hydrogen per day, or about 200 million standard cubic feet per day. These large hydrogen plants typically are co-located with the end users in order to reduce hydrogen gas transportation and storage costs. In addition, SMR technology is also scalable to smaller end-use applications. This has the potential advantage, during the early phases of a hydrogen transportation economy, of having hydrogen production located at the dispensing stations, so that the existing natural gas distribution system can be used to have feedstocks delivered close to the point of production and end use. The distributed SMR approach reduces or eliminates the need for a dedicated hydrogen transmission, storage and distribution infrastructure.

Partial oxidation of a hydrogen-rich feedstock (such as natural gas, coal, petroleum coke, or biomass) is another pathway for hydrogen production. With natural gas as a feedstock, the partial oxidation process typically produces hydrogen at a faster rate than SMR, but it produces less hydrogen from the same quantity of feedstock. Moreover, as a result of increasing natural gas prices, the further development of natural gas partial oxidation technology has slowed. The use of solid fuels is also possible, through gasification, to produce a synthetic gas (syngas) that can then be used in a partial oxidation process to obtain a hydrogen product.

Electrolysis, or water splitting, uses energy to split water molecules into their basic constituents of hydrogen and oxygen. The energy for the electrolysis reaction can be supplied in the form of either heat or electricity. Large-scale electrolysis of brine (saltwater) has been commercialized for chemical applications. Some small-scale electrolysis systems also supply hydrogen for high-purity chemical applications, although for most medium- and small-scale applications of hydrogen fuels, electrolysis is cost-prohibitive.

One drawback with all hydrogen production processes is that there is a net energy loss associated with hydrogen production, with the losses from electrolysis technologies being among the largest. The laws of energy conservation dictate that the total amount of energy recovered from the recombination of hydrogen and oxygen must always be less than the amount of energy required to split the original water molecule. For natural gas SMR, the efficiency at which the feedstock is converted into hydrogen ranges from 67 percent to 73 percent. Despite the energy loss resulting from the conversion of natural gas to hydrogen in the SMR process, the fuel costs per mile for compressed natural gas (CNG) vehicles and FCVs are comparable. In fact, assuming current commercial natural gas prices and the current fuel economies of existing FCV and CNG vehicles, operating fuel costs for FCVs are less than those for CNG vehicles. With projected fuel efficiency improvements in both vehicles, however, the comparative operating fuel cost advantage could reverse, making CNG vehicles more competitive with FCVs, if SMR conversion efficiencies do not improve. It should be noted, however, that taking into account the incremental capital costs of these vehicles would result in a much higher cost associated with FCVs, unless there were also dramatic decreases in fuel cell and other production costs.

For electrolysis, the efficiency of converting electricity to hydrogen is 60 to 63 percent.²³ To the extent that electricity production itself involves large transformation losses, however, the efficiency of hydrogen production through electrolysis relative to the primary energy content of the fuel input to generation would be significantly lower. In certain cases, it may be economical to use off-peak electricity if it is priced well below the average electricity price for the day; however, such market applications would have to be balanced with other potential electricity supplies, the cost versus benefits of appropriate metering and rate design, and the implied reduction in utilization of the electrolysis unit, as described above. The development of such an application could also support other technologies, such as PHEVs.

Advanced technologies for hydrogen production are also being explored.²⁴ They include thermochemical reactions, such as those using nuclear fission, photosynthesis, fermentation, landfill gas recovery, and municipal waste reformation. However, the likelihood of the technological and economic success of these advanced technologies is not guaranteed.

Economics of Hydrogen Production Technologies

The economics of hydrogen production depend on the underlying efficiency of the technology employed, the current state of its development (i.e., early stage, developmental, mature, etc.), the scale of the plant, its annual utilization, and the cost of its feedstock. From a systems perspective, as shown in Figure 2.1, other considerations include the physical distance and availability of potential feedstocks from potential end-use markets for hydrogen gas, and whether to use centralized production in order to take advantage of economies of scale in production and incorporate hydrogen transmission and distribution systems from the plant gate, or rely on distributed hydrogen production, where the feedstocks are transported over a greater distances and the hydrogen gas transmission and distribution infrastructure is minimized.

A summary of the economics of select hydrogen production technologies, based on U.S. annual average prices during 2007, is provided in Table 2.1. The values in Table 2.1 are based on a review of existing literature, and many aspects of technology costs and performance have not been independently verified, but some trends in production cost economics can be observed. For example, large SMR plants using natural gas as a feedstock have a clear operating cost advantage over the smaller SMR units designed for distributed hydrogen production applications, mainly as a result of economies of scale and utilization.

For most of the production technologies shown in Table 2.1, plant capital costs are a relatively large portion of the production costs. The capital costs for the distributed wind (electrolysis) and central nuclear thermochemical technologies were obtained from a 2004 study by the National Academies of Sciences, while the other production costs were estimated by the National Renewable Energy Laboratory in 2005, with the exception of central coal gasification with CCS, whose costs were updated this year.²⁵

The degree of sensitivity of total production costs to capital costs will depend on the production method. In the case of a centralized SMR plant, for example, the 2005 overnight capital cost was $181 million. Using the Chemical Engineering Plant Cost Index (CEPCI), the 2008 capital cost is computed to be $209 million (2008 nominal dollars), which would result in an increase in the product cost of $0.21 per kilogram of hydrogen, or 15 percent. If the operating and maintenance costs and feedstock costs are not adjusted, the increase in capital costs results in only a 3-percent increase in the total production cost. Much more important for the centralized SMR production cost is the price of natural gas, which has varied from about $6 per million Btu in 2005 to more than $13 per million Btu in 2008. Similarly, in the case of distributed electrolysis, the capital cost has increased to $1.11 per kilogram of hydrogen in 2008, based on the CEPCI, representing a 2-percent increase in the overall cost due to the relatively high cost of the electricity input.

At the other extreme, the total cost of hydrogen production from the central nuclear thermochemical method is most sensitive to capital costs, which account for 55 percent of the total hydrogen production cost. Using the CEPCI escalator, the capital cost component would have increased to $0.88 per kilogram of hydrogen in 2008, translating to a 9-percent increase in total production cost.

In addition to capital cost disadvantages because of their size, smaller plants tend to have higher feedstock and utility costs, lower conversion efficiencies, and higher per-unit costs for labor and other operations and maintenance costs. In addition, smaller distributed units are likely to have a lower capacity factor over which capital and fixed costs can be amortized, as indicated in Table 2.1. Including consideration of these factors, the distributed SMR production cost of $2.63 per kilogram is 79 percent higher than the central SMR production cost.

The substantially lower feedstock cost for coal drives the relatively lower overall production cost for coal gasification, both with and without CCS. However, coal gasification on this scale is limited,²⁶ and the application of CCS is at an early evaluation and testing stage.²⁷ The large scale of the plant drives unit capital costs down, but at costs approaching or exceeding $600 million, investment risk may be a concern.

Biomass gasification offers some of the same promise and concerns as coal gasification. On the positive side, life-cycle CO2 emissions may be substantially less, with the possibility that CCS combined with biomass gasification might reduce GHG emissions. The energy density of the biomass feedstock is substantially less than that of coal, however, and it may not be practical to build a biomass gasification unit with a hydrogen production capacity of 155,000 kilograms and supply sufficient quantities of biomass to the plant site at a delivered price of $38 per ton.²⁸

Electrolysis technologies suffer from a combination of higher capital costs, lower conversion efficiency, and a generally higher feedstock cost when the required electricity input is considered. A distributed electrolysis unit using grid-supplied electricity is estimated to have a production cost of $6.77 per kilogram of hydrogen when the assumed 70-percent capacity factor is considered. A central electrolysis unit operating at 90-percent capacity factor, with 30 percent of the power requirements coming from wind and 70 percent from the grid, is estimated to have a production cost roughly 15 percent higher than that of a distributed SMR plant.

Advanced nuclear-fueled thermochemical processes are unproven, but they may provide low per-unit production costs in the future, as indicated by the $1.39 per kilogram production cost. The estimated $2.5 billion investment required for a facility with a capacity of 1.2 million kilograms per day leads to capital charges of $0.76 per kilogram. The lower feedstock cost includes the cost of nuclear fuel, net of any co-product benefits from oxygen sales that may be available. Operating and maintenance costs include decommissioning costs in addition to the usual labor, taxes, security, and other costs.

Because feedstock and electricity prices can be expected to vary considerably over time, it is useful to change the assumed values for those prices used in Table 2.1 from point estimates to variables and plot the resulting “breakeven curves,” as shown in Figure 2.2. The figure shows, for a given feedstock price in dollars per million Btu, what the electricity price would be for the cost of hydrogen production to be the same. For electricity prices above each line, the fossil or biomass feedstock in question would be less expensive than electrolysis, with the reverse being true for electricity prices below the line.

Figure 2.2 considers only the carbonaceous feedstock versus the price of electricity used for electrolysis. Other costs, such as capital and operating and maintenance, could be applied to both electrolysis and other processes, with the differential of those costs applied to the lines as shown in Figure 2.3. Because electrolysis technologies generally have higher capital and operating and maintenance costs, the implied price for electricity would have to be lower to achieve cost parity with a fossil or biomass feedstock.

The “breakeven curves” shown in Figures 2.2 and 2.3 illustrate the cost advantage of using fossil or biomass-based feedstocks in comparison with current electrolysis technologies. With coal and biomass prices in the range of $2 to $3 per million Btu, those technologies can be seen to have a significant cost advantage over electrolysis. Even at delivered natural gas prices of $15 per million Btu, delivered electricity prices would have to be no more than 4.9 cents per kilowatthour on a feedstock basis only, as shown in Figure 2.2, or no more than 2.8 cents per kilowatthour when capital and other costs are also considered, as shown in Figure 2.3.

Hydrogen Transmission and Distribution

Centrally produced hydrogen must be transported to markets. The development of a large hydrogen transmission and distribution infrastructure is a key challenge to be faced if the United States is to move toward a hydrogen economy. A variety of hydrogen transmission and distribution methods are likely to be used. Currently, small and mid-sized hydrogen consumers use truck, rail, and barge transportation modes for hydrogen in either liquid or gaseous form. Larger industrial users rely on pipelines and compressors to move the hydrogen gas. In theory, a blend of up to 20 percent hydrogen in natural gas can be transported without modifying the current 180,000-mile natural gas pipeline infrastructure.²⁹ Some States, including Pennsylvania and California, already are examining this option. However, pipelines that carry pure hydrogen will require special construction and materials in order to avoid issues of steel embrittlement and leakage. This analysis provides only a basic overview of hydrogen distribution issues.

A network for the commercial transmission and distribution of hydrogen gas has been developed and used successfully by the industrial gas industry. There are slightly over 1,200 miles of hydrogen gas pipelines today, compared with about 295,000 miles of natural gas transmission lines and approximately 1.9 million miles of natural gas distribution lines to deliver some 23 trillion cubic feet of natural gas per year.³⁰ Delivery methods for hydrogen are determined chiefly by the production volume and the delivery distance. For example, compressed gas pipelines are used to transport large volumes of hydrogen over short distances to industrial users, such as oil refineries and ammonia plants. Cryogenic, over-the-road tank trucks, railcars, and barges are used to transport mid-sized quantities of liquid hydrogen over longer distances.³¹ Very small quantities of gaseous and liquid hydrogen currently are distributed via high-pressure cylinders and tube trailers. For transportation over longer distances, all the distribution options are relatively expensive, and typically they can more than double the delivered cost of the hydrogen.

Because hydrogen is highly volatile, safety is also a necessary enabling requirement for the current and potential future hydrogen economy. Safety issues are not, however, addressed in this report.

Hydrogen Pipeline Systems

Currently, more than 99 percent of all the hydrogen gas transported in the United States is transported by pipeline as a compressed gas. Pipeline transmission of hydrogen dates back to the late 1930s.³² The pipelines that carry hydrogen generally have operated at pressures less than 1,000 pounds per square inch (psi), with a good safety record. As of 2006, the U.S. hydrogen pipeline network totaled over 1,200 miles in length, excluding on-site and in-plant hydrogen piping (Table 2.2). More than 93 percent of the U.S. hydrogen pipeline infrastructure is located in just two States, Texas and Louisiana, where large chemical users of hydrogen, such as refineries and ammonia and methanol plants, are concentrated.

The natural gas supply system provides an interesting example of how a hydrogen supply system might ultimately evolve to support a hydrogen economy. Each day, close to 70 million customers in the United States depend on the natural gas distribution network to deliver fuel to their homes or places of business. Overall, the U.S. network comprises more than 302,000 miles of interstate and intrastate transmission pipelines for natural gas, more than 1,400 compressor stations that maintain pressure on the pipeline network and assure continuous forward movement of supplies, and more than 11,000 delivery points, 5,000 receipt points, and 1,400 interconnection points. Another 29 natural gas hubs or market centers provide additional interconnections, along with 399 underground natural gas storage facilities and 49 locations where natural gas can be imported and exported via pipeline.

In comparison, the existing U.S. hydrogen pipeline network is only one-third of 1 percent of the natural gas network in length and has less than 200 delivery points. Also, because of concerns over potential leakage, the hydrogen pipes tend to be much smaller in diameter and have fewer interconnections. Special positive displacement compressors are also required to move hydrogen through the pipelines. The length of hydrogen gas piping tends to be short, because it is usually less expensive to transport the hydrogen feedstock, such as natural gas, through the existing pipeline network than to move the hydrogen itself through new piping systems. Historically, welded hydrogen pipelines have been relatively expensive to construct (approximately $1.2 million per transmission mile and $0.3 million per distribution mile).³³ Consequently, the pipelines have required a high utilization rate to justify their initial capital costs.³⁴ More recently, polyethylene sleeves and tubing systems have emerged as a possible low-cost alternative solution for new hydrogen distribution systems, with total capital investments for transmission piping potentially dropping to just under $0.5 million per mile (in 2005 dollars) by 2017 and with commensurately lower costs for distribution lines.³⁵

How a centralized hydrogen transmission and distribution system will evolve is unknown, and therefore the costs cannot be estimated with a high degree of confidence. The costs will depend on where the pipelines are sited, rights-of-way, pipeline diameter, quality and nature of the pipeline materials required to address the special properties of hydrogen, operating pressures, contractual arrangements with hydrogen distributors, financing and loan guarantees, the locations of dispensing stations relative to distributors, and how applicable environmental and safety issues in the production, transmission, distribution, and dispensing of hydrogen are addressed. Because all hydrogen gas has to be manufactured, hydrogen production facilities may be located in ways that minimize overall production and delivery costs.

Liquid Hydrogen (Cryogenic) Transport

Hydrogen can be cooled and liquefied in order to increase its storage density and lower its delivery cost. There are currently four liquid hydrogen suppliers and seven production plants in the United States with a total production capacity of about 76,495 metric tons per day. Those facilities support about 10,000 to 20,000 bulk shipments of liquid hydrogen per year to more than 300 locations.³⁶ Most long-distance transfers of hydrogen use large cryogenic barges, tanker trucks, and railcars to transport the liquid hydrogen.³⁷ NASA is the largest consumer of liquid hydrogen. The chief constraints to widespread use of this hydrogen transportation mode relate to the energy losses associated with liquefying hydrogen and the storage losses associated with boil-off.

Compressed Hydrogen Gas Cylinders

Hydrogen is also distributed in high-pressure compressed gas “tube trailer” trucks and cylinder bottles. This delivery method is relatively expensive, and typically it is limited to small quantities and distances of less than 200 miles.

Alternative Chemical Carriers

Hydrogen also can be transported using hydrogen-rich carrier compounds, such as ethanol, methanol, gasoline, and ammonia. Such carriers offer lower transportation costs, because they are liquids at room temperature and usually are easier to handle than cryogenic hydrogen; however, they also require an extra transformation step, with costs that must be weighed against the cost savings associated with transporting low-pressure liquids. Hydrogen carriers such as methanol and ammonia may also present some additional safety and handling challenges.

Hydrogen Fuel Distribution

The most economical methods for distributing hydrogen depend on the quantities and distances involved. For distribution of large volumes of hydrogen at high utilization rates, pipeline delivery is almost always cheaper than other methods—except in the case of long-distance transportation, e.g., over an ocean, in which case liquid hydrogen transport is cheaper. Laying a hydrogen distribution system in large, high-density cities can also be very expensive, approaching the cost of transmission systems, because existing roads must be dug up and repaired following practices to minimize disruptions to other co-located systems, such as electricity, natural gas, communication cables, etc.

For smaller quantities of hydrogen, pipeline delivery methods are not as competitive as liquid hydrogen delivery or compressed gas delivery via tube trailer or cylinders. The tube trailers have lower power requirements and slightly lower capital costs, although many more tube trailers may be required to deliver the same quantity of hydrogen. Distance is the chief deciding factor between liquid and gaseous hydrogen. At long distances, costs for the number of trucks needed to deliver a given quantity of compressed hydrogen will be greater than the energy costs associated with liquefaction and fewer trucks.

Hydrogen Storage

Because hydrogen gas has such a low density, and because the energy requirements for hydrogen liquefaction are high, efficient hydrogen storage generally is considered to be among the most challenging issues facing the hydrogen economy. For current chemical applications, storage issues are not so critical, because the large producers of hydrogen both generate and consume the gas simultaneously on site, thereby reducing storage and distribution requirements significantly.

Stationary Storage Systems

There are no official statistics on the locations, designs, and capacities of U.S. hydrogen chemical storage facilities. Some privately published data exist, from which the following estimates were derived:

• Intermediate-Scale Compressed Gas Storage Tanks. About 600 large high-pressure gaseous storage facilities currently exist in the United States.³⁸

• Intermediate-Scale Liquid Hydrogen Storage Tanks. About 459 large liquid hydrogen storage sites exist in 41 States.³⁹ In addition, 4 States (California, Illinois, Michigan, and Nevada) and Washington, DC, currently operate hydrogen vehicle refueling stations that use liquid hydrogen as the storage medium.

• Large-Scale Gaseous Storage in Caverns and Salt Domes. Very large quantities of hydrogen can be stored as a compressed gas in geological formations such as salt caverns or deep saline aquifers. There are two existing underground hydrogen storage sites in the United States.

In addition, the co-storage of hydrogen with natural gas has been proposed. There are 417 locations in the United States where natural gas is currently stored in rock caverns, salt domes, aquifers, abandoned mines, and oil/gas fields, with a total storage capacity exceeding 3,600,000 million cubic feet. Hydrogen stored in salt caverns has the best injection and withdrawal properties.

Small-Scale and Mobile Storage Systems

The largest challenges for hydrogen storage are related to transportation applications, in which constraints on hydrogen vehicle design, weight, volume, and efficiency, limit the amount of the gas that can be stored onboard a vehicle. Currently, about 4 to 10 kilograms of hydrogen are required to power an LDV for 300 miles, which is the driving range that most consumers expect. Neighborhood hydrogen refueling stations also are expected to require small- to medium-scale storage systems compatible with the small footprint of existing gasoline stations. Several small-scale storage options are currently under development, but each has some limitation:

• Compressed Gas Storage Tanks.Compressed gas is currently the preferred method for onboard vehicular storage; however, very high gaseous storage pressures, on the order of 5,000 to 10,000 psi (350 to 700 bar), are required to contain a sufficient driving range of fuel. They are relatively expensive, and the high operating pressures give rise to safety concerns in the event of an accident. In addition, there is significant use of energy to compress the gas. Nevertheless, more than 95 percent of all current hydrogen vehicles use compressed gas storage systems, and driving ranges of 200 to 300 miles are being achieved in the latest U.S. vehicle designs. With production at 500,000 units per year, high-pressure storage tanks for hydrogen (5,000 to 10,000 psi) are estimated to range in cost from about $8 per kilowatthour to $17 per kilowatthour,⁴⁰ depending on the pressure capability. Assuming that the full 5-kilogram and 10-kilogram hydrogen storage capabilities of the 5,000 psi and 10,000 psi rated storage tanks can be utilized, the hydrogen storage costs would range from $1,340 to $1,420 per vehicle at production volumes, which would constitute slightly more than an order of magnitude reduction from current costs

• Liquid Hydrogen Storage Tanks. Liquid hydrogen has the highest energy storage density and lowest vehicular weight of any current method, but it also requires an expensive, insulated storage container (dewar) and an energy-intensive liquefaction process. Several concept vehicles have been developed and placed in service in the United States and Europe with liquid hydrogen storage. The cost of such a storage system is a concern, and if the storage system does not have an active refrigeration unit, approximately 2 percent of the hydrogen will need to be vented every day as it evaporates. The volume capacity required for liquid hydrogen will vary significantly, depending on whether the fuel is used in an HICE vehicle or an FCV. Because liquid hydrogen on a volume basis has approximately 26 percent the energy of a gallon of gasoline, the liquid hydrogen tank must have a capacity 3.8 times that of a gasoline tank to hold the same amount of energy. For conventional internal combustion engine (ICE) vehicles with an efficiency of 30 miles per gallon, a 15-gallon gasoline tank provides approximately the same range as a 60-gallon liquid hydrogen tank. For FCVs with an efficiency equivalent of 62 miles per gallon, a 28-gallon tank containing about 7.3 kilograms of liquid hydrogen will be required.

• Advanced Storage Methods. Other advanced storage methods include metallic and chemical hydrides, amides, alanate storage systems, and carbon nanotubes. Solid metal and chemical systems offer some unique storage solutions for hydrogen, with the main challenges at the current time being their weight and their slow response time during refueling. The interstitial storage of hydrogen in carbon nanotubes is another concept with potential for very lightweight hydrogen storage, but the R&D is still preliminary. In addition, several other storage systems and mechanisms may be promising, including the use of sponge iron and glass microspheres.

Hydrogen Dispensing

Currently, only a small number of States and the District of Columbia have announced plans to construct “Hydrogen Highways” with the refueling and maintenance stations needed to support hydrogen LDVs.⁴¹ California has progressed furthest, with 31 installed hydrogen refueling stations (about one-half of the U.S. total) and a few private maintenance facilities.

More recently, an “East Coast Hydrogen Highway” has been proposed by a consortium of automobile manufacturers and hydrogen suppliers.⁴² Initial hydrogen refueling stations have been constructed for public access in Washington, DC, and New York. Also, there is a military hydrogen refueling station in Virginia.

As of 2007, there were a total of 63 hydrogen demonstration refueling stations in the United States (Table 2.3). Two-thirds of the existing refueling stations are capable of self-producing hydrogen, and the remaining one-third are stationary or mobile refueling stations that rely on deliveries of liquid or gaseous hydrogen for their operation. Currently, there are no home refueling stations except those located at manufacturers’ research facilities. California hosts the Nation’s only hydrogen refueling station that is connected to a hydrogen pipeline and a centralized production plant.

Compression costs must be included in any discussion of the operating costs for hydrogen dispensing stations. For example, if hydrogen is produced via distributed SMR, the SMR typically produces hydrogen gas at a pressure of 150 to 200 psi, which then must be compressed to at least 6,000 psi in a storage tank, to be delivered to a vehicle’s 5,000 psi fuel tank. Typically, the energy required for this compression is roughly 3 kilowatthours per kilogram of hydrogen,⁴³ which, at today’s commercial electricity prices (approximately $0.09 per kilowatthour), translates into a compression cost of $0.27 per kilogram.

Hydrogen End Use Applications

A multi-billion-dollar hydrogen industry currently exists in the United States, serving a myriad of hydrogen end-use applications; however, about 99 percent of that hydrogen currently is used in chemical and petrochemical applications. Of the end uses, the largest consumers are oil refineries, ammonia plants, chlor-akali plants, and methanol plants. Some specific examples of hydrogen end use include:

• Petroleum refining—to remove sulfur from crude oil as well as to convert heavy crude oil to lighter products
• Chemical processing—to manufacture ammonia, methanol, chlorine, caustic soda, and hydrogenated non-edible oils for soaps, insulation, plastics, ointments, and other chemicals
• Pharmaceuticals—to produce sorbitol, which is used in cosmetics, adhesives, surfactants, and vitamins
• Metal production and fabrication—to create a protective atmosphere in high-temperature operations, such as stainless steel manufacturing
• Food processing—to hydrogenate oils, such as soybean, fish, cottonseed, and corn oil
• Laboratory research—to conduct research and experimentation
• Electronics—to create a special atmosphere for the production of semiconductor circuits
• Glass manufacturing—to create a protective atmosphere for float glass production
• Power generation—to cool turbo-generators and to protect piping in nuclear reactors.

The transportation sector and stationary power applications are widely viewed as the two critical sectors where there may be an opportunity to expand greatly the future use of hydrogen. These two sectors are the focus of the rest of this section.

Transportation End Uses

A wide variety of transportation end uses have been demonstrated in recent years, including hydrogen-fueled transit buses, ships, submarines, aircraft, bicycles, motorcycles, and scooters. Most of the hydrogen vehicles still are in the conceptual stage, and accurate statistics are difficult to locate.

LDVs are the largest segment of the U.S. vehicle market, the largest consumers of petroleum products, and a large source of GHG emissions in the transportation sector. As a result, fuel switching to hydrogen in LDVs may offer significant potential for oil savings and emissions reductions. Two main types of hydrogen vehicles have been proposed, HICE vehicles, an extension of current vehicle technology, and FCVs. Many analyses of the hydrogen economy consider only FCVs.

Although the discussion below focuses on the future role of hydrogen-powered LDVs, a small number of FCVs and HICE vehicles, including both LDVs and transit buses, already are operational. Appendix D provides a discussion of hydrogen vehicles currently in operation.

HICE Vehicles

Because HICE vehicles typically start from a mass-produced vehicle design and involve relatively low-cost modifications, they could be considered a near-term bridge to the hydrogen economy. In theory, the HICE vehicles can be deployed sooner and in much larger numbers than fuel cell vehicles due to their lower cost. The rapid deployment of HICE vehicles could encourage the construction of hydrogen refueling stations, maintenance facilities, and the development of hydrogen safety codes and standards.

HICE Vehicle Cost

One advantage of HICE vehicles is that their overall cost is only a small fraction of the current cost of an FCV. For example, many conventional vehicles can be converted to run on a mixture of hydrogen and gasoline by adding a small on-board electrolyzer for as little as $1,000. Full HICE vehicle designs that rely on onboard gaseous or liquid hydrogen storage systems to deliver pure hydrogen to the engine require more expensive modifications.

Among the automakers, Ford has demonstrated HICE vehicles and gained insight into current and projected costs versus performance. At production volumes, a vehicle can be designed to optimize the combustion of hydrogen fuel at approximately $5 per kilowatt additional engine cost⁴⁴ to achieve a 12- to 25-percent tank-to-wheels efficiency gain relative to a gasoline LDV, with an engine that is 68 to 83 percent more fuel-efficient being developed at a projected additional incremental cost of $5 per kilowatt.⁴⁵ At an average of 223 horsepower for the 2007 model year LDV, which had an adjusted fuel economy rating of 20.2 miles per gallon,⁴⁶ the additional HICE cost would be $830 to $1,660 per vehicle. At an $830 incremental engine cost, the fuel efficiency would be approximately 22.6 to 25.2 miles per kilogram of hydrogen. The $1,660 incremental engine cost would have lead to a projected fuel efficiency of 34 to 37 miles per kilogram of hydrogen.

Adding storage tanks and safety systems⁴⁷ would bring the estimated incremental cost of an HICE LDV in large-scale production to between $2,370 and $3,280 above a comparable average conventional LDV. The range of the incremental costs depends on determining the appropriate tradeoffs among cost, efficiency, and range—while considering consumer preferences—that results in achieving production-level volumes.

Electric Vehicles (EVs) and PHEVs

The chief alternative to today’s ICE vehicle is an electric motor vehicle. The major automakers consider EVs to be the ultimate, emission-free at point-of-use replacement for gasoline and diesel vehicles. The key challenge for EVs has been the development of sufficient onboard electricity supply capacity to satisfy customers’ expectations for vehicle range.⁴⁸ Advanced battery designs and ultra-capacitors are considered to be potential solutions to this challenge. In these vehicle designs, the consumer would plug the vehicle into an electrical outlet, charge the battery or ultra-capacitor, drive the vehicle, and then recharge the battery or capacitor as necessary. Although advanced lithium-ion batteries and ultra-capacitors have been successfully demonstrated, their costs are high, and current storage capacities still are too low.

To create a vehicle with batteries far smaller than required for a full-range electric-only vehicle while retaining an extended driving range, a more modest battery may be supplemented with an on-board liquid-fueled generator to create a PHEV. Given typical driving patterns, a PHEV with a 40-mile range on grid-supplied electricity (PHEV-40) could achieve a 65- to 75-percent reduction in vehicle petroleum consumption compared to a conventional ICE vehicle.⁴⁹ This estimated reduction in petroleum use reflects petroleum savings in charge-depleting operation, when the onboard generator is not running, and in charge-sustaining operation, as in today’s current HEVs, where the onboard generator operates with higher efficiency than a conventional ICE. Compared to an EV with a 220-mile range on grid power (EV-220), a PHEV-40 would reduce initial battery size and cost by a factor of three.⁵⁰ This technology is nearing commercialization and is expected to be offered to consumers in late 2010.⁵¹ Generally speaking, the larger the onboard battery, the less the choice of fuel used for onboard power generation will affect the overall amount of LDV petroleum use and emissions produced.

The relatively small proportion of total travel fueled by power generated onboard a PHEV-40 suggests a large reduction in total petroleum use even if the onboard generator is powered by a petroleum fuel. Generally speaking, the larger the onboard battery, the less the choice of fuel used for onboard power generation will affect the overall amount of LDV petroleum use and emissions produced.

FCVs

Several major automobile manufactures have begun R&D programs to develop hydrogen fuel cells as an onboard electricity generation system, serving as an alternative to a conventional onboard generator in substituting for or supplementing an onboard electricity storage system (see "Hydrogen Fuel Cell Technologies"). Hydrogen fuel cells produce electricity from a chemical reaction much like a battery does. The key difference is that the fuel cell can be recharged continuously with fresh hydrogen from an on-board storage tank, whereas the battery system must be recharged from an electrical outlet. Also, the on-board hydrogen storage tank can be recharged more quickly than batteries.

Much of the industry’s fuel cell R&D information remains proprietary. As of 2005, two major auto manufacturers, GM and Daimler Chrysler, acknowledged expenditures of more than $1 billion in FCV development.⁵² GM has begun market testing of 100 Chevrolet Equinox fuel cell sport utility vehicles.⁵³ Daimler has announced plans to start serial production of its Mercedes Benz B-Class FCV in 2010.⁵⁴ Honda began commercial leasing of its FCX Clarity in 2008.⁵⁵ Other automobile manufacturers, including Toyota, Ford, and Volkswagen, also have developed FCV concept cars.

All FCV concepts currently under development use electric motors to power the wheels, typically accomplished through the combination of an electric battery storage system and an on-board hydrogen fuel cell. Depending on the degree of hybridization, the battery may provide pure “plug-in” electricity to drive the vehicle some distance. The battery system would be complemented by a hydrogen storage system and a fuel cell, with the goal of extending the driving range to 300 miles.

The primary impediments to the deployment of hydrogen FCVs include cost, fuel cell durability, and expanding the operational temperature range of the cell. The costs of current FCVs are prohibitive as a result of high component costs and the fact that the vehicles are either custom-built or produced in limited series. Also of concern is achieving the necessary minimum range for consumer acceptance.

The primary cost component of the FCV is the fuel cell itself, which has a life expectancy about one-half that of an internal combustion engine. Thus, consumers would have to replace the fuel cell twice in order to achieve a vehicle operating lifetime equivalent to that of a traditional engine. Other features of electric/fuel cell vehicles are reasonably well understood at this time and have been commercialized to some extent in the current generation of hybrid vehicles.

Stationary Power Systems

A near-term area of demand for fuel cells includes stationary power applications, such as backup power units, power for remote locations, and distributed generation for hospitals, industrial buildings, and small towns. Stationary fuel cell power systems already are commercially viable in settings where the consumer is willing to pay a small price premium for reliable energy, and in remote areas where fossil fuel transportation costs are prohibitive. To date, approximately 600 stationary power systems, each with 10 kilowatts or more capacity, have been built worldwide; and more than 1,000 smaller stationary fuel cells, less than 10 kilowatts, have been installed in homes and as backup power systems.⁵⁶

Comprehensive data on U.S. stationary fuel cell installations are not available, but the following types of stationary fuel cell applications are under development:

• Large cogeneration (combined heat and power) systems are being manufactured for large commercial buildings or industrial sites that require significant amounts of electricity, water heating, space heating, and/or process heat. Fuel cells combined with a heat recovery system can meet some or all of these needs, as well as providing a source of purified water.

• Small, standalone cogeneration systems currently are viable in some areas where the large cost of transmitting power justifies the added cost of a fuel cell. Currently, U.S. companies (such as Plug Power) manufacture small fuel cell systems that are able to produce up to 5 kilowatts of electricity and 9 kilowatts of thermal energy. The excess heat can be used for water or space heating to further reduce the site’s electrical energy use.

• Uninterruptible power supply (UPS) systems, in which fuel cells are used as backup power supplies if the primary power system fails, are one of the fastest growth areas for stationary fuel cell technologies. UPS systems often are used in important services, such as telecommunications, banking, hospitals, and military applications. Battery systems have been used for many years to provide backup power to essential services; however, the battery output time is relatively short. In contrast, fuel cells with refillable fuel storage systems can provide power for as long as required during a blackout.

• Home energy stations are another variant of small, standalone cogeneration systems. They use either reformers or electrolyzers to produce hydrogen fuel for personal vehicles, and they also incorporate a hydrogen fuel cell that can provide heat and electricity for the home. One advantage of the stations is that they offer enhanced utilization of the hydrogen gas, i.e., higher capacity factors for the hydrogen production unit, and therefore help to defray some of the overall cost of the hydrogen refueling station. Appliance-sized home energy stations are undergoing development by several automobile manufacturers as a potential alternative to commercial refueling stations.

Market Potential for Hydrogen in Distributed Generation

The market for distributed generation could be significant if selected goals of the U.S. Department of Energy (DOE) hydrogen program are met. The appropriate match between a fuel cell technology and the intended application depends on the magnitude and duration of the power needed, the cost, performance, and durability of the fuel cells, and the operating temperature range.

All fuel cells produce some byproduct heat, but the temperature of the byproduct heat can vary dramatically, from about 180 degrees Fahrenheit for PEM fuels to more than 1,200 degrees for molten carbonate fuel cells. Fuel cells that produce high-temperature byproduct heat with over 250 kilowatts of capacity are suitable for combined heat and power applications in industrial and large commercial settings; those that produce low-temperature byproduct heat are suitable for both mobile uses (e.g., LDVs and forklifts, 80 to 130 kilowatts) and residential applications (e.g., providing electricity, space, and water heating, up to 10 kilowatts). Most fuel cell designs, including PEM and molten carbonate technologies, use different electrolytes, stack designs, and balance of plant.

Consequently, technology learning achieved for one of the technologies is not entirely transferable to other fuel cell technologies, with a few exceptions.⁵⁷

The installed capital cost of phosphoric acid fuel cells in the commercial sector varies according to size. For 200-kilowatt systems, the cost quoted by United Technologies Corporation (UTC) for the PureCell 200 ranges from $6,000 to $7,750 per kilowatt, and for the PureCell 400 system the installed cost ranges from $3,625 to $4,500 per kilowatt in 2008.⁵⁸ The first generation of commercial molten carbonate fuel cells in 2010 is estimated to cost about $6,200.⁵⁹ Molten carbonate fuel cells use the high operating temperatures of the fuel cell to reform methane and steam to produce hydrogen. The CO2 produced is recycled to restore the molten chemical used to generate electricity. Efficiencies to produce only electricity can approach 50 percent, and overall efficiencies (electricity plus byproduct heat) are approximately 70 percent when both products are fully used. A U.S. DOE program supports R&D to develop and implement a molten carbonate fuel cell design that uses some of the lost heat to mechanically turn a turbine to increase generation efficiency by another 10 percent.

If the R&D succeeds in lowering the installed capital costs of molten carbonate fuel cells below $2,500 per kilowatt, the technology could satisfy a significant percentage of new demand for combined heat and power in the industrial and commercial markets. The resulting market penetration, once the cost reductions are achieved, may be slow due to the fact that industrial and commercial boilers are long-lived. According to a 2005 study by Energy and Environmental Analysis, Inc. (EEA),⁶⁰ at least 47 percent of all boilers were at least 40 years old. Because boiler equipment rarely is replaced before it fails, the fuel cell technology is unlikely to replace existing boilers or existing cogeneration equipment before it fails. Also, because energy-intensive industries are in decline in the United States, the market potential for molten carbonate fuel cells in the industrial sector is limited, unless significant economic benefits could be realized by replacing current equipment. For example, in the Annual Energy Outlook 2008 (AEO2008) reference case, demand for boiler steam (heat) applications in the industrial sector⁶¹ is projected to decline by 360 trillion Btu, or 9.5 percent, while the demand for electricity⁶² is projected to increase by 170 trillion Btu, or 4.3 percent, between 2010 and 2030.

The market potential in the commercial sector is better but does not promise rapid growth. Commercial electricity and heat demands are expected to grow more quickly than in the industrial sector, including space and water heating by 358 trillion Btu (16 percent) and purchased electricity by 1,896 trillion Btu (40 percent). Nevertheless, it appears unlikely that the capital costs and performance of molten carbonate fuel cells will improve to the levels needed for substantial penetration of the new market. Any technology learning from PEM fuel cell successes may not be readily transferable to molten carbonate fuel cell production due to the difference in technologies.

With only about 1,350 hours between stack and catalyst replacement, the PEM fuel cell currently is not sufficiently durable to penetrate most markets in large numbers. The electricity generation efficiency of a PEM fuel cell is projected to rise to 36 percent by 2030, while the combined efficiency for electricity and byproduct heat is expected to range between 50 percent and 65 percent if all of the electricity and heat are used. At a delivered hydrogen cost of $2 to $3 per kilogram ($17.54 to $26.32 per million Btu), the fuel component of the cost of electricity generation is expected to range between 14 cents and 21 cents per kilowatthour, which would not be competitive with projected central station delivered electricity prices of 10.5 cents per kilowatthour in 2030. Because the construction costs for hydrogen pipelines to all homes would be extremely expensive, a more likely option might use the existing natural gas infrastructure and on-site natural gas steam reforming. The cost of that option is currently too high, at up to $40 per million Btu according to DOE’s Office of Fossil Energy, and additional R&D on small-scale SMR will be required to bring the delivered fuel cost under $2 per kilogram of hydrogen.

Figure 2.4 illustrates the fuel-related costs of electricity generation as a function of the cost of hydrogen, excluding the capital plus operating costs of the PEM units. The ability to also satisfy space and water heating demand allows the range to increase, depending on how well the end-use demands match the PEM supply and whether backup space and water heating equipment has to be purchased to satisfy any unmet heating demand.

Notes