Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Major findings of the sequestration cost analysis

The results of the sequestration cost analysis are best understood in the context of the relative costs for energy services for transport using alternative fuel/vehicle combinations, without the sequestration of separated CO2.

Consider first a comparison of costs without sequestration on a $/GJ of fuel basis (see fig. 6.2). The first observation one can make about these costs is that gasoline would be far less costly than hydrogen derived from any source. Second, the least costly option for providing hydrogen is from natural gas, for which the cost per GJ to the consumer is likely to be nearly 60 per cent higher than the gasoline price. Third, the cost of hydrogen is about the same for both coal and biomass, despite the assumption that the biomass price is about 50 per cent higher than the coal price - a reflection of the facts that costly sulphur removal is not needed for biomass and that biomass is more reactive than coal and thus easier to gasify (Williams et al. 1995a,b). Fourth, the cost of hydrogen from biomass or coal is about 10 per cent higher than for hydrogen from natural gas. Fifth, all the electrolytic hydrogen options are far more costly than hydrogen derived thermo-chemically from natural gas, coal, or biomass; for the least costly electrolytic option, which represents what might plausibly be realizable for advanced thin-film photovoltaic technologies,18 the cost of hydrogen to the consumer is 60 per cent higher than the indicated cost of hydrogen from biomass.

A more meaningful comparison than the cost of fuel per GJ is the cost of fuel per km of driving the vehicle that the fuel might be used in. The consumer prices and lifecycle CO2 emissions for hydrogen shown in figure 6.2 are converted in figure 6.3 to fuel costs and lifecycle CO2 emissions per km of driving a fuel cell vehicle (FCV), along with a comparison of gasoline costs and lifecycle CO2 emissions per km of driving, for both internal combustion engine vehicle (ICEV) and FCV applications. The reference gasoline ICEV is a year-2000 version of the Ford Taurus automobile with a fuel economy of 11.0 km/litre (25.8 mpg). The hydrogen FCV has performance characteristics that are comparable to those for this ICEV and a gasoline-equivalent fuel economy of 30.4 km/litre (71.6 mpg) (Williams 1995). The cost of hydrogen derived from natural gas and biomass without sequestration of the separated CO2 would be 1.49 and 1.73 cents/km, respectively, compared with 2.65 cents/km for gasoline. Initially fuel cell cars would probably be operated on gasoline converted to a hydrogen-rich gaseous fuel mixture via a process that begins with partial oxidation. The estimated fuel economy of such a fuel cell car having the same performance as the internal combustion engine alternative would be 18.0 km/litre (42.3 mpg). The fuel cost per km for this vehicle would be 1.62 cents/km (see fig. 6.3).19

Consider next the penalties for sequestration associated with hydrogen derived from natural gas for the situations depicted in figure 6.2. These penalties are low - increasing the cost of hydrogen to the consumer by only 0.8 to 2.6 per cent. The low value of the penalty is a result of: (i) being able to sequester the separated CO2 near where it is produced (thus avoiding the costs of long-distance CO2 pipeline transport), (ii) the relatively low penalty for storage in depleted natural gas fields compared with aquifers,20 and (iii) receiving a modest credit for extra natural gas produced as a result of depressurization of the natural gas reservoir - a credit almost large enough to cover the incremental cost of sequestration.

The penalties for sequestration shown in figure 6.2 for natural gas feedstocks are for the case where hydrogen is produced at the natural gas field both without and with sequestration. In this case the carbon tax needed to make equal the costs of hydrogen with and without sequestration, and thereby induce hydrogen producers to sequester the separated CO2, is US$932/tC, depending on the cost of sequestration and the magnitude of the credit for enhanced natural gas recovery (see table 6.5). Such a carbon tax would increase the hydrogen cost by US$0.2-0.7/GJ, some 1.6 to 5.5 per cent of the cost of hydrogen to consumers without sequestration. The magnitude of the carbon tax in this instance is independent of the length of the hydrogen transmission line and thus would apply equally to situations where large potential hydrogen markets are located near the natural gas field.

Fig 6.2 The estimated to automotive consumers of pressurized hydrogen derived from alternative sources, per unit of delivered energy (Note: the cost of hydrogen derived from natural gas without sequestration is for a plant sited at a depleted natural gas field. The sequestration costs shown for coal and biomass are for sequestration in saline aquifers located 250 km from the hydrogen production facilities. The costs for hydrogen derived electrolytically from nuclear and wind (photovoltaic) sources are for the indicated AC (DC) electricity prices, assuming an 85 per cent efficiency for electrolysis and a 96 per cent efficiency for rectification. The cost of transporting electrolytic hydrogen from production facilities to refuelling stations is assumed to be the same as for hydrogen derived from coal. In all cases, production costs include the costs for pressurization to 75 bar at the production plant. For the intermittent renewable electric sources, costs for hydrogen storage are also included in production costs)

Fig. 6.3 The estimated cost to automotive consumers of pressurized hydrogen derived from alternative sources, per km of driving fuel cell cars. Costs for gasoline used in internal combustion engine cars and fuel cell cars are shown for comparison

If instead hydrogen production without sequestration were to take place 1,100 km from the natural gas field near major remote hydrogen markets (see table 6.8), the hydrogen cost penalty to the consumer for sequestration would increase to the range 3.2-5.1 per cent, for a natural gas wellhead price of US$3/GJ. In this instance the carbon tax required to make equal the costs of hydrogen produced without and with sequestration is US$39-63/tC; with this carbon tax in place the cost of hydrogen to the consumer would be 7-11 per cent higher than without sequestration and with no carbon tax.21 Though the impact of this tax on the cost of hydrogen to the consumer is still modest, it is considerably higher than for the scenario where hydrogen production without sequestration takes place near the natural gas field. Since the latter scenario is most realistic for situations where there are large potential hydrogen markets near the natural gas field, the sequestration option is likely to be pursued first in such regions e.g. in the Netherlands and Texas. If nearby hydrogen markets can be served, there is no need to seek the economies of large-scale hydrogen transmission capacity by building very large hydrogen production plants. This makes it possible to begin sequestration much earlier in the evolution of a hydrogen economy than would be the case if the only potential hydrogen markets were remote from natural gas fields.

For the base case sequestering scenario for coal, in which CO2 is sequestered in aquifers, the cost penalty is higher - giving rise to a 611 per cent increase in the cost of hydrogen to the consumer for the situations indicated in figure 6.2. For the high sequestration cost estimate, the cost of hydrogen to the consumer per km of driving would still be 29 per cent less than the cost of gasoline with an internal combustion engine. In this instance, a carbon tax of US$29-54/tC would be needed to make equal the costs of hydrogen without and with sequestration, a tax that would increase the cost to the consumer of hydrogen derived from coal by 8-15 per cent.

For the base case sequestering scenario for biomass, the cost penalty is somewhat higher but still modest - giving rise to a 10-14 per cent increase in the cost of hydrogen to the consumer for the situations indicated in figure 6.2. But even if the high sequestration cost estimate proves to be valid, the cost of hydrogen to the consumer per km of driving would still be 26 per cent less than the cost of gasoline with an internal combustion engine.

The percentage cost penalties associated with sequestering the separated CO2 would be smaller still if calculated as a contribution to the total cost of owning and operating a car, to which the cost of fuel makes a relatively modest contribution (see fig 6.4).22 Consider, for example, the total lifecycle cost of owning and operating a fuel cell car operated on hydrogen derived from coal without sequestering the separated CO2. The estimated cost is 20.1 cents/km (slightly less than for a gasoline internal combustion engine car of comparable performance) for the assumptions indicated in figure 6.2, of which the price of hydrogen fuel accounts for only 1.71 cents/km. With a carbon tax sufficient to induce producers of hydrogen from coal to sequester, the lifecycle cost of the car would increase only 0.14-0.25 cents/km, or 0.71.3 per cent.

These calculations suggest that deep reductions in CO2 emissions can be achieved in the transport sector at low incremental cost by shifting to hydrogen or a hydrogen carrier derived from chemical fuel feedstocks. Such cost calculations could of course be refined as more knowledge is gained about the various technologies involved. But this basic finding is robust. It is not sensitive to the outlook for the relative prices of the three feedstocks considered. If natural gas prices remain low for the indefinite future, and if high estimates of remaining ultimately recoverable natural gas resources prove to be valid, hydrogen would be predominantly natural gas based in many parts of the world for decades to come, and sequestration could be pursued on large scales at very low incremental cost. If, instead, natural gas prices rise rapidly beyond the year 2010,23 then biomass- and coal based hydrogen production strategies would eventually supplement hydrogen supplies provided by natural gas.

Biomass without sequestration would tend to be favoured over coal with sequestration wherever adequate land resources are available for biomass production,24 because hydrogen production costs would be as low or lower under most conditions (for all the cases considered here - compare tables 6.6 and 6.7), and because lifecycle emissions would be only half as large (see table 6.1). However, in regions where coal is cheap and the availability of land for biomass production is limited, coal-based hydrogen strategies would be favoured.

Biomass-based production strategies with sequestration would be favoured where land-use constraints limit the extent of hydrogen production from biomass,25 or when it is desired to offset emissions from other sectors or parts of the world.

Fig 6.4 The estimated lifecycle cost to consumers of owning and operating a hydrogen fuel cell vehicle, for hydrogen derived from alternative sources, per km of driving. Estimated lifecycle costs for gasoline-powered internal combustion engine and fuel cell cars are shown for comparison (Note: operating lifetimes of 241,000 km and 193,000 km are assumed for fuel cell vehicles and internal combustion engine vehicles, respectively. Sources vehicle performance and cost characteristics based on Ogden et al. 1994; fuel costs from fig. 6.3)

The largest uncertainties underlying this analysis are: (i) the extent to which society will adopt low-temperature fuel cells and their fuels for transportation and distributed combined heat and power applications, (ii) the prospects for converting existing natural gas transmission lines to hydrogen service and building new pipelines dedicated to hydrogen, and (iii) the extent and security of saline aquifers as storage reservoirs for separated CO2.

Major research, development, and demonstration commitments are needed on the hardware for low-temperature fuel cell technology and for the production, storage, and transport of hydrogen and hydrogen carriers. Although there is rapidly growing R&D activity on low temperature fuel cells for transportation applications, the overall level of the R&D effort in this area is still minuscule. Moreover, there is very little ongoing R&D on the production of hydrogen from coal or biomass.

The issues relating to converting existing natural gas transmission and distribution lines to hydrogen service are discussed in Appendix C. Being able to make this conversion would be especially important for the already industrialized world, where a large natural gas pipeline infrastructure is already in place. It appears to be feasible to convert low-pressure distribution lines without great difficulty, but hydrogen embattlement looms as a serious issue for high-pressure transmission lines. Technical fixes might be possible. For example, would mixing trace quantities of another gas with the hydrogen be a suitable strategy for coping with the embattlement problem (see Appendix C)? More research is needed to find out.

More research is also needed on the various issues raised regarding sequestration in depleted oil and gas fields and saline aquifers especially the latter.

Appendix A: The importance of the water-gas shift reaction in fuel decarbonization

It is feasible to convert a carbon-rich fuel to a hydrogen-rich fuel at relatively high overall efficiency by taking advantage of the thermodynamics of the so-called water gas shift reaction.

Suppose it is desired to make hydrogen (H2) out of carbon (C). The first step is to gasify C by partial oxidation to produce carbon monoxide (CO):

C + 1/2 O2 (r) CO, D HR = -110.5MJ/mol.

Note that, with complete oxidation,

C+O2 (r) CO2, D HR = -393.5 MJ/mol,

which shows that the CO produced in partial oxidation retains 71.9 per cent of the higher heating value (HHV) of the original C:

HHVCO = 283.0MJ/mol,

HHVC = 393.5 MJ/mol.

The HHV of H2 is approximately the same as the HHV of CO:

HHVH2 = 285.8 MJ/mol.

The next step is to use the water-gas shift reaction:

CO+H2O(g) (r) CO2 +H2, D HR = -41.2MJ/moI,

a slightly exothermic reaction that makes it possible to "shift" the energy contained in the CO to H2. The needed steam (gaseous water, H2O(g)) can be raised by evaporating liquid water (H2O(l)), a phase change that requires just slightly more heat than is generated in the water gas shift reaction:

H2O(l) (r) H2O(g), D HPC = +44.0MJ/mol.

Note that the extra heat required (44.0 - 41.2 = 2.8 MJ/mol) can be met using just 2.5 per cent of the heat released via the initial partial oxidation of C.

The final step in the production of H2 from C involves separating out the H2 from the gaseous H2/CO2 mixture - for example, using commercial pressure swing adsorption (PSA) technology, which can recover 90 per cent or more of the produced H2 at up to 99.999 per cent purity. (PSA exploits the ability of porous materials to adsorb specific molecules selectively at high pressure and disrobe them at low pressure; the cyclic pressure swing is what gives the process its name.) A by-product of the process is a stream of pure CO2 that is potentially available for sequestration in some secure reservoir.

This discussion shows that it is possible in theory to generate 1 mole of H2 from 1 mole of C at an overall efficiency n = 100*(285.8/393.5) = 72.6 per cent. In the real world, H2 fuel would be made from natural gas, coal, or biomass. H2 can be produced from natural gas at an overall real-world efficiency of 84 per cent via a process that begins by reacting natural gas with steam (Williams et al. 1995a,b). This efficiency is higher than for making H2 from C because natural gas (mainly methane, CH4) already contains a great deal of H2. The real-world efficiencies for making H2 from coal (whose chemical composition can be represented as ~CH0.8O0.08) via oxygen-blown gasification and from biomass (whose chemical composition can be represented as ~CH1.5O0.7) via steam oxidation are each about 64 per cent (Williams et al. 1995a,b).

Appendix B: Biomass CO2 emission offset potential in a world where some coal-rich regions cannot or will not reduce emissions

Consider a world situation in 2100 where the hydrogen fuel cell vehicle is a well established technology and where natural gas resources are in limited supply, so that the primary options for producing hydrogen at low cost are from coal and biomass. Suppose further that some regions with a capacity to produce biomass on large scales agree to seek to help bring about deep reductions in global CO2 emissions, but that other regions must depend on coal and either have no ready access to secure sequestering sites or are unwilling to incur the cost penalties for storage, however modest. In these circumstances it may still be feasible to achieve low global emissions levels because of the large "emissions offset potential" offered by biomass-derived hydrogen (see tables 6.1 and 6.2).

To illustrate the possibilities, suppose that: (i) the world population in 2100 is 10.5 billion; (ii) there are then 0.4 cars per capita in the world (the average for the industrialized market countries in 1985) - some 4.2 billion cars altogether (10 times the present number); (iii) the average car is driven 13,400 km/year (the average for industrialized market countries in 1985); (iv) these cars are all equipped with hydrogen fuel cells; (v) the average automotive fuel economy is 42.5 km/litre (100 mpg) of gasoline equivalent.

With so many cars, would it be possible to have zero net CO2 emissions worldwide from the automotive sector, if some cars were run on hydrogen derived from coal without sequestration of the separated CO2 (so that net lifecycle emissions would be +41.2 kg C/GJ of hydrogen see table 6.1) and if the rest were run on hydrogen derived from biomass with sequestration of the separated CO2 in saline aquifers and/ or depleted natural fields (so that the net lifecycle emissions would be-18.4 kg C/GJ of hydrogen - see table 6.1)?

Under the above conditions, there would be zero net emissions from the automotive sector if a fraction "a" of all cars is operated on coal-derived hydrogen, where "a" is determined from the following equation based on the net lifecycle emission rates for the two hydrogen-producing systems: 41.2*a - (1 - a)*18.4 = 0, or a = 0.31. Thus net global emissions from cars would be zero if 31 per cent of the hydrogen were derived from coal and 69 per cent from biomass. The amount of hydrogen consumed worldwide for cars would be 46 EJ per year. Assuming an efficiency of 64 per cent for making hydrogen from coal or biomass (see Appendix A), some 22 EJ and 51 EJ per year of coal and biomass, respectively, would be needed for fuelling automobiles in 2100. For comparison, total global coal consumption in 1985 was 90 EJ per year; and SO EJ per year is approximately the average rate at which noncommercial biomass is consumed in the world today (Hall et al. 1993). Assume that in 2100 all this biomass would be grown on plantations. It is reasonable to expect an average yield of, say, 20 dry tonnes per hectare per year at that time (Williams 1995). With an energy content of 20 GJ/dry tonne, some 128 million hectares of plantation area would be needed worldwide. There will very likely be far more land available for biomass energy plantations than this (Larson et al. 1995). The required sequestering rate of 0.75 GtC per year is modest in relation to the estimated capacities of aquifers and depleted natural gas and oil fields.

Appendix C: Pipeline transport of hydrogen

Transport costs for new pipelines

The cost of hydrogen pipeline transport is higher than for natural gas. Pottier et al. (1988) estimate that the cost of the pipe would typically be 50 per cent higher than for natural gas transmission lines, largely because embattlement-resistant steels would be specified. Also, the optimized pipe diameter would be perhaps 20 per cent larger for hydrogen to achieve the same energy flow rate (Leeth 1979). The cost of installation would also be higher, because special care would be needed with welds. And much larger compressors would be needed to achieve the same energy flow rates, because the volumetric energy density of hydrogen is just 35 per cent of that for natural gas.

But hydrogen pipeline transport costs would not be prohibitive, for two reasons. First, the diversity of supply options implies that hydrogen would often be available from sources that are relatively close to where the hydrogen is needed (Ogdn and Nitsch 1993). Second, even where long-distance pipelines are needed, the costs involved are relatively modest compared with the total cost of hydrogen to the consumer, and the total cost per unit of service provided would typically still be less than for conventional transport fuels. For example, the transport of hydrogen 1,110 km to consumers for use in fuel cell vehicles from a production facility located at a natural gas field contributes less than one-fifth of the total cost of hydrogen to the consumer (see fig. 6.2 and table 6.8).

Transport costs for pipelines converted to hydrogen from natural gas

In many parts of the industrialized world extensive pipeline networks designed for use with natural gas are already in place. Could these pipelines be converted to hydrogen? Even neglecting considerations of embattlement, these pipelines are not optimized for use with hydrogen, whose energy density and viscosity are very different from the values for natural gas. Nevertheless, from a fluid dynamics perspective, the mismatch would not be severe. For a given pipe diameter and operating pressure, the energy flow rate for hydrogen would be about 85 per cent of that for natural gas with partially turbulent flow and 90 per cent of that for natural gas with fully turbulent flow (Christodoulou 1984). Seals, joints, and metering equipment would probably have to be replaced with conversion. Moreover, almost three times as much compressor power would be needed to obtain the same energy flow rate as for natural gas; and reciprocating rather than centrifugal compressors would be needed. Nevertheless, there could be large savings associated with the sunk costs of the pipelines themselves, if conversion were feasible.

Hydrogen embattlement

A major concern about converting natural gas pipelines to hydrogen service is "hydrogen environment embattlement," which refers to the degradation of mechanical properties that takes place when a metal is exposed to a hydrogen environment.

The available evidence indicates that this is not an issue for existing local natural gas distribution systems, which could therefore be converted to hydrogen with only minor modifications (Ogden et al. 1995). However, because of the higher pressures (up to 70 bar or 1,000 psia) and materials used, hydrogen environment embattlement could be a serious problem for existing long-distance natural gas transmission lines that would transport pure hydrogen. The primary mechanisms are fatigue crack growth under cyclic loading and slow crack growth under stable loads near welds and other "heat-affected" zones in pipes.

Various countermeasures have been suggested, including adding gases to inhibit embattlement, pre-loading with an inert gas, coating pipelines, and selective replacement of steels susceptible to embattlement. The most promising approach seems to be gas additives. Available research indicates strong inhibition of fatigue crack growth with oxygen, at concentrations ranging from 100 ppm to 1 per cent (Ogden et al. 1995). A high priority for hydrogen research should be given to gas additives and other countermeasures for inhibiting embattlement in existing natural gas transmission lines that might be converted to hydrogen.


Financial support for this research was provided by the Geraldine R. Dodge Foundation, the Energy Foundation, the W. Alton Jones Foundation, and the Merck Fund.


1. The use of coal for power generation in the IS92a scenario accounts for 19 per cent of total CO2 emissions from fossil fuels in 2100, the same percentage as in 1990.

2. If there were no coal and if those concerned about climate change had only conventional oil and gas resources to worry about, it might be feasible to stabilize the atmospheric concentration of CO2 near the present level. According to the US Geological Survey (Masters et al. 1994), remaining ultimately recoverable conventional oil and gas resources (proved reserves plus estimated recoverable undiscovered resources) amount (as of 1992) to 11,300 exajoules (EJ) of oil and 12,500 EJ of natural gas. Assuming CO2 emission coefficients of 19.5 million tonnes of carbon per EJ (MtC/EJ) for oil and 13.5 MtC/EJ for natural gas, cumulative emissions associated with burning all conventional oil and gas resources amount to 390 GtC, which is within the range of cumulative emissions consistent with stabilizing the atmosphere at the present concentration of CO2 (IPCC 1996).

3. Hydrogen produced electrolytically is more costly than hydrogen derived from fuels largely because electricity is far more expensive than fuel per unit of contained energy. Averaged over all users in the United States in 1993 the electricity price was three times the price of all petroleum products, five times the price of natural gas, and 14 times the price of coal (EIA 1995). Hydrogen derived from electricity will be even more expensive than the electricity "feed" used in its manufacture, both because of inefficiencies in electrolytic conversion and because of the capital needed for electrolytic equipment.

Although hydrogen derived from carbon-rich fuels via thermochemical processes will be more expensive than these feedstocks per unit of contained energy, it will not be nearly so expensive as electrolytic hydrogen, because the basic processes involved in "turning carbon into hydrogen" (see Appendix A) are relatively simple and not nearly so capital intensive as the process of making electricity.

4. When hydrogen is derived electrolytically from off-peak power sources that have low running costs, it can be economically attractive, because the capital charges can be avoided. However, the total quantities of hydrogen potentially available via this route are tiny in relation to the demand for fluid fuel. Nevertheless, hydrogen derived electrolytically from off-peak hydroelectric power will play important roles in providing hydrogen for various demonstration projects and other niche applications in helping launch a hydrogen economy.

5. In late 1995, Ballard introduced a 60 passenger, 275 hp. hydrogen PEM fuel cell "commercial prototype" bus having a 400 km (250 mile) range. It has sold three hydrogen fuel cell buses each to the Chicago Transit Authority and to BC Transit in Vancouver, British Columbia; the first bus was delivered to Chicago in September 1996. In 1998 Ballard expects to be producing commercially 75 passenger, 275 hp. hydrogen fuel cell buses having a 560 km (350 mile) range. In collaboration with Ballard and using Ballard fuel cells, Daimler-Benz has introduced three experimental fuel cell vehicles: a proof-of-concept hydrogen PEM fuel cell van (NECAR I) in April 1994, a prototype hydrogen powered fuel cell passenger van (NECAR II) in April 1996, and a small prototype methanol-fuelled fuel cell passenger car (NCAR III) with an onboard fuel processor in September 1997. In joint ventures with Ballard, Daimler-Benz hopes to produce 100,000 engines per year for fuel cell vehicles by 2005. Also Toyota has introduced two experimental fuel cell vehicles: a prototype hydrogen fuel cell passenger car using metal hydride storage in October 1996, and a prototype methanol fuel cell passenger car with onboard fuel processing in September 1997.

6. In one study carried out for the US Department of Energy by the Allison Gas Turbine Division of General Motors (AGTD 1994), it is estimated that, in mass production, the cost for a 60 kWe continuous output (~90 kWe peak output) automotive electrochemical engine system based on use of the PEM fuel cell operated on methanol would be US$3,899 - consisting of USS1,752 for the fuel cell stack, plus US$1,077 for the fuel processor, US$195 for the heat rejection and water management system, and US$875 for system auxiliaries - so that the total unit installed cost would be USS65/kWe continuous (US$46/kWe peak).

In another study carried out for the US Department of Energy by Directed Technologies, Inc. (James et al. 1994), the cost of mass-produced automotive hydrogen/air PEM fuel cells (for production at a rate of 106 units per year in the year 2004) is estimated to be US$31/ kWe, and the cost of an 85 kWe hydrogen/air PEM fuel-cell-based automotive power system (inducing the cost of the fuel cell, the heat management system, the power conditioning, an ultra capacitor for peak power, an electric motor, and storage tanks for compressed hydrogen) is estimated to be US$4,400-5,100, compared with US$3,000-4,000 for the cost of the internal combustion engine equipment that would be displaced.

7. At the site or, in transport applications, onboard the vehicle, methanol is "re-formed" to produce a gaseous mixture of hydrogen and carbon dioxide via reactions that are summarized as:

CH3OH + H2O(g) (r) CO2 + 3 H2

This gaseous fuel mixture can be utilized directly by proton-exchange-membrane fuel cells, which (unlike the alkaline fuel cells used in the space programme) are not poisoned by CO2.

8. Hydrogen and methanol can be produced from natural gas with commercially available technology. These energy carriers can also be produced from coal using commercially ready oxygen-blown coal gasifies plus commercially available technologies for the needed further processing. In the case of biomass, the fuel-processing technologies downstream of the gasifier are also commercially available. Although suitable gasifies tailored to biomass are not commercially available, such gasifies could be commercially available by 2000 with a relatively modest R&D effort (Williams et al. 1995a,b).

9. When hydrogen is produced from natural gas, the carbon content of the natural gas feedstock amounts to 15.2 kg C per GJ of produced hydrogen, and the CO2 stream separated at the PSA unit has a carbon content of 10.7 kg C per GJ of produced hydrogen. When hydrogen is produced from coal, the carbon content of the coal feedstock amounts to 31.8 kg C per GJ of produced hydrogen, and the CO2 stream separated at the PSA unit has a carbon content of 29.8 kg C per GJ of produced hydrogen.

10. When hydrogen is produced from biomass, the carbon content of the biomass feedstock amounts to 34.4 kg C per GJ of produced hydrogen, and the CO2 stream separated at the PSA unit has a carbon content of 23.8 kg C per GJ of produced hydrogen.

11. In this paper, all costs are presented in 1991US$ and lifecycle costs are evaluated using a 10 per cent real (inflation-corrected) discount rate. Corporate income, property, and sales taxes are neglected. Also, the energy content of fuels is given in terms of the higher heating value.

12. The present analysis is limited to hydrogen because the sequestration potential is much larger than for methanol production (see table 6.1). However, the general finding that sequestration costs are low holds for methanol as well. In fact, the costs for CO2 compression would be less with methanol production. In this case, the CO2 is generally released at higher pressures using Selexol CO2 separation technology (e.g. 13 bar for methanol produced from biomass using the Battelle Columbus Laboratory biomass gasifier and 21.8 bar for methanol produced from coal using the Shell gasifier Katofsky 1993) than is the case for the PSA technology used in hydrogen production (1.3 bar). If the CO2 must be compressed to 110 bar, the required compression ratio is just 8.5 for biomass- and 5.0 for coal-derived methanol, compared with 85 for hydrogen produced from natural gas, coal, or biomass.

13. The costs for drying and compressing CO2 to the pressures needed for pipeline transport and reservoir injection are based on Farla et al. (1992) at Utrecht University- see note d, table 6.5.

14. The costs of CO2 transport by pipeline are based on analyses carried out at the Statoil R&D Centre in Trondheim, Norway (Skovholt 1993) - see note e, tables 6.6 and 6.7.

15. The costs for disposal in both depleted natural gas fields and saline aquifers are based on the work of Hendriks (1994) at Utrecht University - see note e, table 6.5, and note f, tables 6.6 and 6.7.

16. Drawing on commercial plantation experience in Brazil (Carpentieri et al. 1993), biomass supply curves (potential production vs. long-run marginal production cost) have been generated on a country-by country basis for Africa, Latin America, and Asia, for the year 2025, taking into account land requirements for food production at that time (Larson et al. 1995). Marginal costs were related to prospective yields, and prospective yields were estimated via a correlation with rainfall. It was found that 70 EJ/year (35 EJ/year) of biomass could be produced on 10 per cent (5 per cent) of "available" land in those countries where biomass can be produced at or below this cost. Available land is defined here as non-forest, non wilderness land that is not needed for producing food crops. To put these energy quantities into perspective, consumption of coal, oil, and natural gas in 1990 was 35.8 EJ, 40.5 EJ, and 12.2 EJ, respectively, for developing countries.

17. In a major assessment carried out by a team with participants from the Oak Ridge National Laboratory and the US Department of Agriculture, it is estimated that if R&D goals for plantation biomass in United States can be realized, 5 EJ/year of plantation biomass could be produced in 2020 on 17 million hectares at a long-run marginal cost of US$1.5/GJ (Graham et al. 1995). For comparison, about 3 EJ/year of biomass would be required to support a fleet of 120 million cars (the total number in the United States in 1992) powered by hydrogen fuel cells, assuming the gasoline-equivalent fuel economy of these fuel cell cars is 34 km/litre (80 mpg).

18. The least costly electrolytic option is for a photovoltaic (PV) module efficiency of 18 per cent, a PV installed system cost of US$1,030/kW, and a plant siting in an area of high insolation (270 W/m2). These PV performance and cost parameters are optimistic but plausible for advanced thin-film PV technologies (Ogden and Nitsch 1993).

19. With a cost of service companson, even costs for electrolytic hydrogen used in fuel cell vehicles are not much different from the 2.65 cents/km cost of gasoline for an internal combustion engine vehicle ranging from 2.75 to 3.90 cents/km for the electrolytic options shown in figure 6.3.

20. Storage costs estimated by Hendriks (1994) range from US$3/tC to US$13/tC for depleted natural gas fields, compared with US$9/tC to US$34/tC for saline aquifers.

21. For comparison, a carbon tax in the range US$39-63/tC would increase the retail price of gasoline shown in figure 6.2 by 11-17 per cent.

22. It is assumed that hydrogen is stored onboard vehicles in carbon-fibre-wrapped aluminium tanks at high pressure (550 bar). Because of the bulkiness of gaseous hydrogen storage, the hydrogen FCV is designed for a range between refuellings of 400 km, compared with 640 km for a gasoline ICEV. The weight of the hydrogen FCV is estimated to be 1.3 tonnes, compared with 1.4 tonnes for the ICEV. Initial costs are estimated to be US$17,800 for an ICEV and US$25,100 for a hydrogen FCV (in mass production). The initial cost for a gasoline FCV is assumed to be US$21,700, the same as the estimated cost for a methanol FCV (Ogden et al. 1994). Retail fuel taxes are included under "other non fuel operating costs" at the average US rate for gasoline used in ICEVs; to ensure that road tax revenues are the same for all options, it is assumed that retail taxes are 0.75 cents/km for all options (equivalent to 8.2 cents/litre or 31 cents/gallon for gasoline used in ICEVs).

23. The US Department of Energy has projected that the wellhead price of natural gas in the United States will increase at an annual average rate of 3.1 per cent per year. 1993-2010 (EIA 1995).

24. Considerable hydrogen production from biomass is likely to be possible before land scarcity becomes a major limiting factor for the growing of biomass. See, for example, the calculation presented in Appendix B. and discussions of land-use availability for industrialized countries in Williams (1994a) and for developing countries in Larson et al. (1995).

25. With sequestration of the separated CO2, the amount of biomass grown on a given land area could make a much larger contribution in reducing global emissions than without sequestration.

Contents - Previous - Next