The Great Illusion, Part II - Further Misadventures along the Hydrogen Highway

Our first installment in this series explored some of the less than obvious ambiguities inherent in the strongly expressed support of a hydrogen transition by two entities that seemingly have little to gain from it, namely the oil companies and the automobile manufacturers. This segment is devoted to an examination of the technical challenges that must be met to effect such a transition.

Getting from Here to There

A hydrogen transition will only occur if each of the following is in place: a reliable and cost effective power plant, a comprehensive hydrogen distribution system or, alternately, a practical hydrogen-on-demand system, hydrogen storage systems providing driving ranges equivalent to those of today’s gasoline powered vehicles, and competitive pricing for hydrogen itself.

Hydrogen Power Plants

Most hydrogen proponents appear to favor hydrogen fuel cells as the power plant of choice, but hydrogen burning internal combustion engines are also in contention. Alkali metal thermoelectric devices constitute the dark horse technology.

Hydrogen Fuel Cells

Anyone who attempts to assess the mid and long term prospects of fuel cell technology achieving a real presence in the transportation sector quickly finds himself in deep difficulties. Most of the literature to date has been generated by fierce partisans and much of it can in fact be traced to individuals interested in commercializing the devices. Trustworthy objective analysis is in short supply, but at least one estimable study is available, a D.O.E. document entitled Basic Research Challenges for Fuel Cells and Novel Fuel Cell Materials, which unfortunately, at four years of age, is getting a bit long in the tooth. Still, I recommend it because it has no worthy successor to date and because such changes have occurred in the fuel cell art since its publication have been incremental rather than breakthroughs. In other words, most of the technical obstacles cited in the report remain in force. What follows is based in part on that report and on my own continuous monitoring of the industry in three and half years subsequent.

PEM Fuel Cells, the Industry Choice

Many types of fuel cells exist, but the PEM (polymer electrolyte membrane) type is the only one seriously in contention today as a power source for automobiles and other vehicles. Invented back in 1968, the PEM type remained a laboratory curiosity for almost twenty years when in the mid eighties it excited the interest of one Geoffrey Ballard, the founder of fuel cell company that bears his name but which no longer employs him. Ballard and his engineering team made truly extraordinary improvements in the initially wildly impractical design from the late eighties through the middle nineties, and, for a time, the appearance of a practical moderately priced PEM device appeared to be imminent. But the pace of improvements slackened considerably after 1996 and has never regained its former velocity. Nor have the hundreds of competitors that followed in Ballard’s wake been any more capable of real breakthroughs.

Still, PEMs appear to recommend themselves for automotive transportation to a greater degree than the other designs, and fuel cell supporters have not given up on them. And, in truth, they have much to recommend them.

They operate at low temperatures—well under 200 degrees Fahrenheit—and thus they may be brought to operating temperature quickly. And since they use solid electrolytes, they are simpler in design than the older liquid electrolyte systems. And with over fifty percent efficiency in converting chemical energy into electrical energy, they offer excellent fuel economy.

But their disadvantages and drawbacks remain considerable. They utilize highly complex and rather fragile nanostructures that no one has figured out how to mass produce, they use large amounts of costly platinum catalyst, they require external heating to operate at low ambient temperatures, and they are physically bulky. The DuPont Nafion membranes used as electrolytes in almost all current designs are costly and cannot tolerate high temperatures. Furthermore, PEM fuel cells generally have low tolerance for contaminants and impurities in the hydrogen fuel itself, and typically have operating lives of only a few thousand hours. At a cost of thousands of dollars per kilowatt that’s entirely unacceptable.

Incremental improvements continue to occur in PEM fuel cells, but true breakthroughs remain elusive. Unless those occur, fuel cells are unlikely to establish themselves in any sector of the automotive industry.

Will such breakthroughs occur? Predicting step improvements in an immature technology with any precision is notoriously difficult. If this were not the case, the energy and transportation industries would have made a final decision regarding fuel cells long ago. In general, however, technical progress within any given product category tends to follow an S-shaped curve—slow progress initially, followed by accelerated progress, followed by a further and final slowing. Only when the category is fundamentally changed does this curve no longer obtain, and, in the case of fuel cells, that would entail replacing PEM with something else such as the new and promising solid acid type fuel cell. As it happens, PEM reached the flattened terminal portion of the S curve about a decade ago, and in spite of the surge of investment from 1996 up until about 2001, progress actually slowed. That suggests that the possibilities of the technology may have essentially been realized already, and that significant further progress is rather unlikely.

Hydrogen Storage

Hydrogen is difficult and expensive to store. It also suffers from low volumetric energy density even when highly compressed or frozen to a liquid. Breakthroughs in regard to storage do not appear to be imminent.

Proven techniques for hydrogen storage include pressurized tanks, cryogenic tanks, and metal hydrides. None are really practical for automotive transportation today.

High pressure tanks occupying a reasonable volumes—and reasonable in this context means eight times the volume of a normal gas tank—have to be made of carbon fiber or Kevlar to withstand the 10,000 pounds per square inch of internal pressure. Such tanks run about fifteen grand. Cost would decline with high volume production, but no one sees them approaching the price of metal gas tanks.

Hydrogen stored as a liquid has one quarter the energy density by volume of gasoline or twice that of hydrogen gas at maximum safe pressures, which is still pretty marginal. Nevertheless, cryogenic storage is favored many hydrogen advocates, most notably BMW who keep promising to introduce a hydrogen internal combustion engine vehicle.

Cryogenic tanks are double walled with a hard vacuum between the walls for insulation. They must be provided with safety valves to allow the boil off to escape. Cryogenic tanks are quite a bit cheaper than the high pressure type but they’re still costly and they have the added disadvantage of controlled leakage to the extent that most of the fuel will have vaporized after about two weeks of storage.

Metal hydrides constitute a safe, stable, and volumetrically compact method of storing hydrogen, but a stack capable of fueling a vehicle will weigh hundreds of pounds. Steady improvements have been occurring in hydride development, but a practical storage system is at best a long way off and may never appear.

There are many experimental systems for storing hydrogen including clathrates, carbon nanotubes, glass micro-spheres, among others. All appear promising to varying degrees, and one or more may prove out, but this is by no means certain.

Keeping the Fuel Cell Faith

Hydrogen fuel cells are getting decidedly less coverage in the mainstream press than was the case at the beginning of the decade when interest was at a peak. The major auto makers still trot out fuel cell prototype vehicles at the major car shows, but not so many as in the past, and one isn’t hearing so many confident predictions of full commercial rollouts in the near term. Even the staunchest advocates are moving the market introduction horizon out at least a decade and often two.

It is hard to know what to make of such cautious prognostications. My position is that one can’t predict with any precision at all an event that can only occur after multiple scientific breakthroughs have preceded it. And scientific breakthroughs are not, unfortunately, inevitable much as we might believe in progress. Those vaunted cures for cancer, AIDS, baldness, and normal aging that we all looked forward to a quarter century back were never realized. Nor has the fusion reactor, the Holy Grail of renewable energy. And the fact that billions upon billions of dollars and tens of thousands of man hours have been devoted to all of the above made no difference.

And the fact is that hard won breakthroughs do not truly characterize the history of invention—in other words, Edison’s famous aphorism about invention being “99% perspiration and 1% inspiration” is often inaccurate. Many seminal inventions were conceived in an instant and quickly embodied in practical products, among them the Dunlop pneumatic tire, the analog computer, the steel suspension bridge, and the safety razor, to name just a few. Innovations that couldn’t be perfected quickly, such as the transistor and electronic television, have tended to be the province of very large corporations or government labs, simply because speculative private investors are unwilling to fund uncertain research projects extending indefinitely into the future.

Right now, the well funded research projects involving PEM fuel cells are mostly in the hands of the big Japanese and American auto makers. I don’t see that much activity in Europe any more. Given the travails of the surviving American Big Two, it is difficult to envision them continuing to fund fuel cells as lavishly as before, particularly when such efforts are so unlikely to effect speedy improvements in their bottom lines, and so more and more, automotive fuel cell research is likely to fall to the Japanese.

Hydrogen Internal Combustion Engines

If not fuel cells, then why not ordinary piston engines designed to burn hydrogen? Why not indeed?

The technology of hydrogen ICEs (internal combustion engines) is pretty mature, and, properly designed, they compare pretty favorably with ordinary gasoline engines in terms of drivability and power output. Because of issues in storing hydrogen, driving range is obviously pretty limited however.

The engines are prone to certain failings, such as pre-ignition of the gas from contact with hot cylinder walls due to the low combustion point of hydrogen, and the tendency of hydrogen to leak out of the valves and ignite in the crankcase, but careful design can solve both problems.

Currently, a number of companies are doing conversions of existing engines to run on hydrogen, but no one is manufacturing dedicated hydrogen engines with the exception of Mazda which just announced a flex fuel gasoline/hydrogen Wankel for their flagship RX-8 sports car. BMW was supposed to have a hydrogen car on the market by now as well, but it hasn’t happened.

Hydrogen Hybrid Fuel Injection

This technology generates onboard hydrogen and injects it into a conventional internal combustion engine for a cleaner more efficient burn. It may be regarded as a bridge to full blown hydrogen regime in transportation.

Experiments with hydrogen as a fuel additive go back decades, and a number of technical papers on the subject have been published in automotive engineering journals. Essentially what is going on here is that the high flame speed of the hydrogen in the mix accelerates combustion and results in a cleaner burn.

A number of very small companies make equipment for doing this, and most make use of electrolyzers. These hydrogen injection systems work but they represent a very expensive way of accomplishing what can be done just as easily with reasonably priced devices known as water and alcohol injection systems which are primarily designed for racing. In such systems direct cracking of the water takes place under the intense heat and pressure inside the cylinder and the resulting hydrogen is immediately combusted. No outboard electrolyzer or reformer is required.


Hythane is a trademarked name that is the property of Brehon Energy PLC, located in Dublin, Ireland. The name refers to a mixture of hydrogen with natural gas in which hydrogen is 15% by volume. According to company literature the simple addition of hydrogen in this amount will reduce emissions by 50% in a natural gas internal combustion engine without requiring any of the extensive modifications associated with a full hydrogen conversion.

Currently hythane enjoys a similar status to hydrogen itself in transportation, that is, it is available at a few demonstration fueling stations and is utilized by a handful of fleets. Lacking the cachet of pure hydrogen, we’re not sure it will go further than this unless government mandates regarding emissions become increasingly stringent, but should this come to pass, there are other ways of obtaining ultra low greenhouse gas emissions in internal combustion engines that may prove more feasible such as utilizing a combination of dimethyl ether and alcohol.

Hydrogen Internal Combustion Engine Prospects

We are not particularly sanguine as to the immediate or mid term prospects for hydrogen internal combustion engines gaining a large market or even succeeding as a niche product. While we believe that a competitively priced hydrogen internal combustion engine could be produced with current technology, which is not the case with PEM fuel cells, such a power plant would face many of the same challenges and would even be disadvantaged in some respects.

Hydrogen internal combustion engines as presently constituted are likely to offer lower fuel economy than fuel cells which is a particular problem due to the very low energy density by volume of the gas. Very likely BMW is right, and liquid storage is the only practical solution, but that, as we have seen has its own set of problems.

As is the case with fuel cells, hydrogen internal combustion engines fail to provide an enhanced driving experience over conventional internal combustion engines, and since driving range is reduced and the cost of fuel drastically increased, what is their advantage?

Certainly they are greener than gasoline engines, though not quite so green as fuel cells, and at least potentially they free the driving public from dependence on fossil fuels. But here we’re back to the troubling question of how to establish a renewable energy source-based hydrogen economy, a question for which answers remain elusive in our view.

Hydrogen Generation and Distribution for Transportation

A vast published literature exists in this area, most of it consisting of financial projections regarding the cost of putting a fueling infrastructure in place and the logistics of a phased building out of the infrastructure. Most of the studies we’ve seen insist upon the relative ease of setting up a hydrogen infrastructure and suggest that an extremely spotty distribution of fueling facilities would suffice if initial sales were aimed at an urban population.

Apart from a handful of experimental hydrogen filling stations, infrastructure for hydrogen powered vehicles is largely nonexistent today. Many models have been proposed for the development of such an infrastructure, but most, unfortunately, rely on public policy initiatives that may not garner sufficient political support to be realized, or would depend upon major investments on the part of the petroleum industry for operations that cannibalize their existing business and are unlikely to return a profit for many years to come.

One must also consider the well-to-wheel cost of hydrogen, which is apt to remain high for the foreseeable future. A kilogram of industrial hydrogen is currently several times the cost of a gallon of gasoline, the two being nearly equivalent in the amount energy they can provide. Many projections suggest that a number of distributed generation schemes will become competitive with gasoline within a few years if crude petroleum prices continue to rise at the present rate; however, I believe that other alternative fuels likely to establish themselves in the mid term, and may well be cheaper per BTU than hydrogen from whatever source.

I do not doubt that hydrogen fueling facilities could gradually be added to existing filling stations, and that if an enthusiastic early adopter market for hydrogen fueled vehicles were to develop, such a gradualist approach might work. But all existing evidence suggests that no such early adopter market exists, though, interestingly, markets do exist for pure battery powered electric vehicles, plug-in hybrids, and biofuel. True, a handful of tiny companies performing hydrogen conversions on conventional internal combustion engines are now in operation, but such companies haven’t succeeded in building a real market.

Now one can in theory generate one’s own hydrogen from water using solar powered electrolysis, and then use that hydrogen to run a converted internal combustion engine vehicle, but almost no one is doing so. The cost of converting an existing gasoline vehicle to hydrogen operation runs in excess of $30,000 which is generally much more than the price of a new car and then some, while a home hydrogen generation plant runs tens of thousands of dollars more. In contrast, the decision to use biodiesel involves no modification in most late model diesel engines, and simply obliges the operator to pay somewhat more for fuel if he or she chooses to use a safe, ASTM certified product. Conversely, plug-in hybrids, most of which are modified Toyota Priuses, require much less retro-fitting than gasoline to hydrogen conversions, and impose no recurrent costs on the operator thereafter other than the price of off-peak electricity. Conversion costs are generally under 10k. From an economic perspective hydrogen vehicles are simply not competitive with the other “green” options.

Chicken and the Egg

Yet without a market for hydrogen fueled vehicles a comprehensive hydrogen generation and distribution system will not be built. Hydrogen fuel cells could and probably will play some sort of role in stationary power generation, but it is unlikely to be a role requiring a pervasive network of fueling station and indeed most stationary fuel cells today are being sold with fossil fuel reformers. Transportation is the sole justification for building such a hydrogen distribution network.

Generation itself poses monumental problems if it is to be attempted on a national scale, especially in the light of declining fossil fuel resources. Generation from coal based syngas would probably be the most economical means of producing hydrogen on a massive scale in the near term, but by attempting to build a coal based hydrogen economy one would be throwing away most of the energy value in coal which is locked up in the carbon component. There’s also the problem of emissions. Carbon dioxide wastes could either be used in manufacturing industrial chemicals or in those applications requiring pure carbon dioxide, but, given the volumes generated in a full throttle coal based energy regime, geological sequestration would probably be required for most of the CO2 produced. And as the world grew and more dependent upon coal, a hydrogen transportation system would be vying with the electrical generation industry for the use of coal, and at the same time proposing to use coal in the most energy inefficient manner.

Distributed Generation vs. Centralized Generation

We should note here that most individuals who have given serious thought to the problems of implementing a hydrogen economy feel that distributed generation of hydrogen presents fewer problems than centralized generation. The energy efficiency of generating hydrogen in very large, centralized facilities is greater than that for small scale generation, but such efficiency gains do not compensate for the losses involved in transmitting hydrogen to its ultimate destination. A sizable fraction of the energy carried by the hydrogen will be sacrificed in moving it whether the method is truck delivery in pressurized tanks or cryo-tanks or pumping it through pipelines in either gaseous or liquid form, though pipelines are the more efficient method. Unfortunately, pipelines for either liquid or gaseous hydrogen are very expensive to build. Liquid hydrogen requires that the pipelines be provided with extensive refrigeration facilities, while hydrogen gas pipes must be very carefully sealed and must have much greater diameters than natural gas pipes to minimize frictional losses, which happen to be greater for hydrogen than for natural gas.

It should be noted here that a distribution network of sorts already exists for hydrogen. Industrial uses, as we have seen, are already many and varied, and total volumes generated are immense and growing. While much of the industrial hydrogen used today is generated on the spot by means of steam reformers or electrolysers, a good deal of hydrogen is transported in compressed or cryogenic form. If hydrogen is to become a major energy carrier rather than simply an element in various industrial processes, existing distribution methods will almost certainly have to be replaced, however. The usual mode of hydrogen delivery by tank truck is extremely energy inefficient compared to truck transportation of gasoline simply because volumetric energy density of hydrogen is so poor.

Serious proposals have been made for combining a new electrical distribution system with a network of liquid hydrogen pipelines. This concept, known as the Super Grid, would use liquid hydrogen to cool superconducting or cryoconducting electrical cables that would greatly increase the efficiency of the transmission grid while supplying hydrogen to vehicles or generating plants. It’s an intriguing notion, but would acquire an extraordinary degree of consensus among political and business leaders to be realized.

More on Distributed Hydrogen Generation

Many methods of hydrogen generation such as concentrating biological processes, biomass reforming, direct cracking, and thermal chemical processes using nuclear energy or concentrating solar seem better suited to large centralized production facilities. By contrast, distributed generation favors only two methods, steam reforming of natural gas and electrolysis.

A highly distributed generation model, particularly one based upon electrolysis and relying on renewable energy sources as the primary energy resources, would require a vast increase in electrical generation and transmission capacity. That this could be done incrementally on a pay-as-you-go basis is doubtful. More likely, a broad national consensus supporting such an action would have to occur.

Generating hydrogen locally by reforming hydrocarbons at or near the point of use avoids most of these problems while giving rise to others. Steam reforming of natural gas remains the least expensive means of generating hydrogen, though that will almost certainly change in the future, and I might add that natural gas has a very extensive pipeline infrastructure of its own. Local steam reforming also fosters a gradualist approach which of course allows the new market to define itself. But distributed steam reforming shifts the burden of carbon sequestration to countless small operators, virtually ensuring that it will fail to be implemented, unless perhaps if a separate infrastructure for piping away carbon dioxide is extended to every hydrogen refueling facility—at what cost is difficult to determine, but it would surely be very high.

One faces even greater problems if coal derived syngas is substituted for natural gas. No one is currently manufacturing a small scale reformer for coal or syngas, though natural gas steam reformers could be modified for use with syngas fairly easily. Furthermore, no current infrastructure exists for distributing syngas, though one did a century and more ago when syngas was used for the gaslights. On the other hand, if raw coal rather than syngas were the feedstock, then that would have to be trucked to destinations all over the country, necessitating the manufacture of new types of vehicles for conveying it and new handling procedures for facilitating mass usage.

Generation by Means of Electrolysis

Most advocates of a hydrogen transition do not favor the reforming of fossil fuels as a long term strategy since it results in greenhouse gas emissions that are difficult to control. And, in any case, the process is scarcely sustainable in the long run. At best they see reforming as a bridge technology. Renewable energy sources-based water electrolysis is their favored method for generating hydrogen, with the renewable energy of choice being wind since the economics of other renewable energy sources are largely unproven.

In this scheme, electrolysis would take place in small facilities at the local level. Some advocates see fuel cell powered vehicles becoming highly integrated parts of the grid and generating electricity which is returned to the grid when they are not actually involved in travel.

Unfortunately, electrolytic hydrogen wants a cheap, superabundant source of electricity, and the only such source known to exist was fossil fuel prior to recent price increases. No inexpensive replacement technology currently exists, including nuclear.

On Board Hydrogen-on-Demand Generation

When the automobile industry began to cultivate a mass market in the first decade of the twentieth century the manufacturers gave little thought to fueling. Such was their faith in the attractiveness of their products that they assumed that car owners would go to any lengths to obtain gasoline to run their vehicles, and that they the manufacturers need not concern themselves with providing infrastructure.

As it happened, the expanding oil industry, sensing a vast new market in the making, went ahead and built the infrastructure in lieu of the auto makers doing so, and both benefited immeasurably. Things sorted themselves out, as they say.

Auto makers today evince no such faith in the inherent attractiveness of fuel cell vehicles and have assiduously cultivated alliances with petroleum companies and industrial gases companies to construct hydrogen fueling stations. Yet, at the same time, they have never quite abandoned an alternative approach, namely on-board reforming.

On-board reforming is the production of hydrogen from a hydrocarbon feedstock, be it gasoline, diesel fuel, methanol, or ammonia, the candidates most frequently advanced. Clearly the presence of an on-board reformer further complicates an already complex and physically massive power plant, but it does offer a couple of advantages in that it substitutes an easily stored, energy dense, liquid hydrocarbon fuel for gaseous or liquid hydrogen, and in that it would appear to require less additional fueling infrastructure for its realization, though this claim may be considered somewhat deceptive. Indeed if gasoline were the choice of feed stock, then no additional infrastructure at all would be required, but neither methanol nor ammonia can share the same transport facilities as hydrogen and would require an entirely different infrastructure.

When the auto manufacturers were conducting early prototyping efforts with fuel cell cars in the middle and late nineties, reformers figured significantly in their efforts, but the basic difficulties inherent in this approach soon manifested itself. If, after all, the purpose of adopting fuel cell propulsion were to reduce dependence upon fossil fuel, achieve excellent fuel economy, and eliminate greenhouse gas emissions, why in the world would one want to install a device that depended upon fossil fuel, emitted carbon dioxide, and required a power input sufficient to nullify in large measure any efficiency advantage conferred by the fuel cell itself? Not to mention the bulk, mass, and considerable expense of a reformer.

And yet on-board reformers continue to lurk in the background of fuel cell vehicle research simply because of the problem of storing elemental hydrogen.

Most on-board hydrogen reformers make use of the partial oxidation method for converting hydrocarbons into syngas, though some have used steam reforming, or even autothermal reforming. In general they are not terribly different in general design from the giant reformers used in the chemical and petroleum industries, though of course they face stringent space constraints not relevant to the latter, and, which, incidentally, are quite difficult to address since reformer technology does not readily scale downward in dimensions. Because they must produce hydrogen of very high purity, on-board automotive reformers generally make use of costly membrane purification systems utilizing noble metal palladium catalysts. In some cases these membranes are sufficiently effective as to allow the designer to dispense with the second water gas shift stage of reformation.

Obviously, the high operating temperatures of reformers present problems since a PEM cell itself cannot endure high heat, and for that reason many of the advocates of this approach favor methanol since it can be reformed at temperatures of under 300 Centigrade. Methanol, however, presents other problems, including fairly high cost, toxicity, fire hazard, and lack of any considerable existing distribution infrastructure.

Here I would point out that several years ago ExxonMobil generated a very interesting position paper on the subject of on-board reformers. The authors argued rather persuasively that a fuel cell powered vehicle using a gasoline reformer would be the best choice for a successor technology in personal transportation because it would offer significantly lowered emissions over the current internal combustion engine art, though nothing approaching zero emissions, could utilize the existing refueling infrastructure, would be the least expensive option for the consumer in respect to fuel, and, even with the losses imposed by the reformer, would offer superior mileage to what was likely to be achieved with internal combustion engines. As an added benefit it would offer a cruising range equal or better to a conventional internal combustion engine motorcar.

Obviously, ExxonMobil is not a disinterested observer since they are in the business of selling petroleum products, but I found the paper persuasive as I’m sure would any open minded individual. Nevertheless, interest in automotive reformers appears to be at low ebb. Prototype fuel cell vehicles employing such reformers have been built, but the additional cost of the reformer on top of what is already a frightfully expensive power plant seems wholly insupportable. If PEM fuel cells were cheap, practical, and available, the reformer option might be appealing. Lacking that, it would seem little more than a thought exercise.

Except that at least one reformer for automotive use is already on the market and serious research continues on another under the aegis of Renault.

The Collier Reformer

The production unit in question has been developed by Collier Technologies, Inc. of Reno, Nevada. It is designed for a single niche application, for use on buses already powered by natural gas. A portion of the natural gas is reformed into hydrogen using the steam reforming process and the resulting hydrogen is mixed with natural gas to achieve a 30% mixture, promoting a very lean burn, increased fuel efficiency, and reduced emissions. The Collier reformer can be retrofitted onto existing units, but the company is more interested in selling engines that have been designed from the ground up to utilize this mixture.

Meanwhile, Nuvera, a startup whose genesis was a reformer project sponsored by Arthur Little, is seriously engaged in developing a reformer for Renault. The anticipated completion date for the project is 2010.

In the next issue,I'll consider the use of hydrogen fuel in other forms of transportation such as aircraft, boats, and heavy equipment.