Hydrogen and fuel cells
Tom Standing furnished an excellent rebuttal to Michael Wang's letter, "Hydrogen as a Transportation Fuel" (OGJ, Aug 25, 2003, p. 10). Some additional data confirm Standing's thesis that fuel cells constitute a transportation energy source whose time has not arrived.
Although Mr. Wang acknowledges that "... electrolysis hydrogen is not a preferable way to deliver hydrogen for fuel cell vehicles," he then states, "... one needs to go beyond energy efficiency calculations. With electrolysis hydrogen, energy feedstocks other than petroleum [such as natural gas, coal, hydropower] could be used for hydrogen production. Thus, even with relatively poor overall energy efficiency, electrolysis hydrogen fuel cell vehicles will still be able to help reduce petroleum use in the US transportation sector ...."
I would suggest that efficiency is important, and because the overall efficiency of the electrolysis hydrogen route is so dismal, there is no economic, environmental, or resource management justification for using electrolysis hydrogen fuel cells.
Let's look at the numbers. The theoretical energy required to dissociate liquid water into hydrogen and oxygen is 68.32 kcal/g-mol, or using Wang's convention, 39.4 kw-hr/kg of hydrogen. If we also use Wang's assumption that the efficiency of the electrolysis process is 70%, then the actual electrical energy required to produce a kg of hydrogen by electrolysis is 56.3 kw-hr.
Let's suppose the electrical energy for electrolysis is provided by a coal-fired power plant with an overall conversion efficiency of 40%. Then the energy input to the boiler of the power plant required to produce the kg of hydrogen is 140.8 kw-hr.
Now, let's turn our attention to the fuel cell. Since the reaction that occurs in the fuel cell produces water vapor, the theoretical energy obtainable from the fuel cell is –57.8 kcal/g-mol of hydrogen, or 33.4 kw-hr/kg of hydrogen. If the fuel cell operates at an efficiency of 70%, then the energy actually obtained from the reaction of a kg of hydrogen in the fuel cell is 23.3 kw-hr.
If hydrogen is to be used as a transportation fuel, it must be compressed to at least 4,000 psi. Wang and Standing offer slightly different estimates of the energy loss in compressing the gas. Even if we use Wang's more-conservative estimate (15% of the energy content of the hydrogen), the energy loss in compressing a kg of hydrogen is 5.9 kw-hr.
So, an energy input to the electrical power generating plant of 140.8 kw-hr produces 1 kg of hydrogen by electrolysis. This kg of hydrogen requires 5.9 kw-hr of energy for compression, so the net output from the fuel cell is 17.4 kw-hr. The overall conversion efficiency of this whole process (coal to electrical output from the fuel cell) is a dismal 12%.
This efficiency does not even include diffusion losses of hydrogen during transport and storage at high pressure.
Now, let's estimate the electrical output from fuel cells that would be needed to power the US vehicle fleet. According to the most recent data from the Bureau of Transportation Statistics, in the year 2000 US passenger car travel totalled 1,602 billion miles, and other two-axle, four-tire vehicle (pick-up trucks, vans, SUVs) travel totalled an additional 924 billion miles.1
Southern California Edison has tested some electric vehicles (Evs) on the Pomona Loop in Southern California. This test loop, which includes both city and freeway driving as well as hills, is designed to simulate real-world driving conditions. The test vehicles included small electric vehicles (Toyoto RAV4, GM EV1, and Honda EV Plus) as well as larger EVs (Ford Ranger, Chevy S10, Nissan Altra, and Chrysler EPIC). The average energy consumption for all EVs tested was 0.422 kw-hr/mile, while the average for the larger EVs was 0.460 kw-hr/mile.2
Note that even the larger EVs included in this test were much smaller and lighter than the full-sized pick-up trucks, vans, or SUVs in operation today. Furthermore, the vehicles tested were EVs, not vehicles powered by fuel cells. In order for a vehicle powered by a fuel cell to achieve the results shown above, it would have to carry a large battery pack (in addition to the steel tank containing the compressed hydrogen) which could store the electrical energy generated in braking—an economic cost and weight penalty not usually considered in fuel cell vehicles.
For this calculation, we shall use the value of 0.46 kw-hr/mile obtained for the larger EVs tested. Applying this figure to the total 2,526 billion vehicle-miles traveled in 2000, we find the fuel cells would need to produce 1.16 trillion kw-hr. Earlier we found that the overall efficiency for converting the energy input to the boiler of a coal-fired power plant into electrical output from a fuel cell was 12%. Therefore, in order to obtain 1.16 trillion kw-hr from fuel cells, we need an energy input to electrical generating plants of 32.0 quadrillion btu (quads). It is interesting to compare this with the energy content of the US gasoline supply.
In the year 2000, US motor gasoline consumption averaged 8.472 million b/d.3 Using an energy content of 5.150 million btu/bbl for reformulated gasoline, we find the energy content of this gasoline is 16.0 quads, or exactly half the energy required for the fuel cell route using coal to generate the electricity for hydrolysis.4
The effect of this energy substitution on carbon emissions would be enormous. Using EIA's carbon emission coefficients of 19.38 million metric tons (tonnes) of carbon per quad for reformulated gasoline and 25.74 million tonnes of carbon per quad for coal, we find the 16.0 quads of energy from gasoline would produce 309 million tonnes of carbon while the 32.0 quads of coal energy would produce 824 million tonnes of carbon or 2.7 times as much.5 This hardly qualifies as an environmental benefit.
Earlier we found that with an efficiency of 70% for the electrolysis process, we needed 56.3 kw-hr of electrical energy in order to obtain 17.4 kw-hr of output from fuel cells. Therefore, 3.75 trillion kw-hr of electrical energy would be needed to produce the 1.16 trillion kw-hr of output from fuel cells necessary to power the US vehicle fleet.
Wang alleges that use of renewable electricity, such as hydropower, for hydrogen production, would reduce fossil fuel use. Let's put this electrical energy requirement into some perspective. The 3.75 trillion kw-hr is 14 times the total hydroelectric energy generated in the US last year.6 Furthermore, mounting public pressure to remove some existing dams effectively precludes construction of any significant hydroelectric generating capacity.
Wang correctly notes that hydrogen production by steam reforming of natural gas is a commercial technology and that almost all hydrogen is produced with this technology. However, the overall efficiency of this process (including both electrical and thermal energy inputs) is less than the 70% figure used by Wang.
One large nitrogen fertilizer producer requires an energy input of 1 g-mol of methane to produce 1.85 g-mol of hydrogen. Earlier we found that the net output from 1 kg of hydrogen (regardless of how it is produced) in a fuel cell is 17.4 kw-hr. In order to provide the 1.16 trillion kw-hr of output needed to power the US vehicle fleet, 66.7 billion kg of hydrogen would be needed.
About 15 tcf of natural gas would be required to produce this hydrogen by the steam reforming process. Last year, total natural gas consumption was about 22.5 tcf. Thus, powering the US vehicle fleet with hydrogen derived from natural gas would increase natural gas consumption by almost 70%—and at a time when natural gas imports from Canada are declining and domestic production is barely able to keep up with demand. Indeed, there is mounting evidence that natural gas prices are destined to remain high and that rising gas demand will be satisfied with imports of liquefied natural gas (LNG)—also from the Middle East.
Is there any merit in replacing oil imports with LNG imports from the same region at a greater cost? Furthermore, as Standing noted, the overall conversion efficiency for this whole process is only about 30%—much less than if the natural gas were burned in an electrical power generating plant.
Contrary to Wang's assertion, efficiency does matter. If the goal is simply to reduce oil use, then a better strategy would be to convert coal into electrical energy and use the electricity directly, such as for home heating, rather than using it in a convoluted process to produce hydrogen with the attendant efficiency losses and capital costs.
Finally, before Wang derides the "simplistic calculations" of Tom Standing (and others) may I suggest he read the account of the first atomic pile experiment conducted beneath the University of Chicago stadium on Dec. 2, 1942.
As the pile approached criticality, Nobel physicist Enrico Fermi checked the progress of this historic and dangerous experiment with a slide rule. Furthermore, all of the thermodynamic data we need to conclude that fuel cells constitute a hopelessly inefficient transportation power source were obtained without any assistance from Argonne's super computers.
References
1. "U.S. Vehicle Miles," Tables 1-32, National Transportation Statistics 2002, Bureau of Transportation Statistics, U.S. Dept. of Transportation, Washington, D.C.
2. "Pomona Loop Test Data," Southern California Edison, California Energy Commission, Sacramento, CA 95814.
3. "Finished Motor Gasoline Supply and Disposition," Table 3.4, Monthly Energy Review, DOE/EIA-0035(2003/07), July 2003.
4. "Approximate Heat Content of Petroleum Products," Table A1, Monthly Energy Review, DOE/EIA-0035(2003/07), July 2003.
5. "Carbon Emission Coefficients at Full Combustion," Table B1, Emissions of Greenhouse Gases in the U.S. 1987-94, DOE/EIA-0573(87-94), October 1995.
6. "Electricity Net Generation (All Sectors)," Table 7.2a, Monthly Energy Review, DOE/EIA-0035(2003/07), July 2003.
Donald F. Anthrop
Professor
Department of Environmental Studies
San Jose State University
San Jose, Calif.
