Ammonia As An Alternative Fuel?

In the last seven years I've written extensively about a wide variety of alternative fuels, including ethanol, methanol, and higher alcohols like butanol, along with compressed and liquefied natural gas (CNG and LNG), hydrogen, and electricity, but I find I haven't said anything about anhydrous ammonia. It turns out that there is a small but enthusiastic group of people promoting its use as an alternative fuel, going back to at least the 1940s. Much of the recent interest in this stems from the fact that ammonia releases little or no greenhouse gas when burned, and that it's possible to produce it by means that involve minimal GHG emissions throughout its lifecycle. However, when you dig into this a little deeper, you discover that almost all ammonia today is produced by the Haber process, using hydrogen sourced from natural gas. And if that weren't enough of a deterrent, the physical properties of ammonia render it an unattractive candidate for a mass-market fuel.

So-called "green ammonia" would avoid natural gas by substituting hydrogen from electrolysis using wind, solar or other renewable electricity. As long as natural gas remains abundant, it's hard to envision this growing beyond a small niche, because the price of ammonia will ultimately be set by the price of natural gas, which remains a cheaper source of hydrogen than electricity from any source, let alone from expensive renewable power sources. Moreover, electricity is fungible, and the best use of renewable or other low-emission power (e.g., nuclear) is probably in backing out power from higher-emitting sources, rather than diverting it into inefficient production of chemicals. As a result green ammonia, like green power, would require subsidies for at least the near-to-medium term if it is to compete with conventional ammonia, which seems like a crucial prerequisite for competing with conventional fuels. And without green ammonia, the whole rationale for an ammonia fuel-and-vehicle network looks questionable--why not just use the gas as CNG or LNG instead, with a fraction of the headaches?

Even if that weren't the case, ammonia faces serious obstacles as a consumer fuel, compared to either conventional fuels or to many other alternatives. Start with energy density, which is less than half that of gasoline by weight, and about 40% by volume. So a gallon of ammonia would only take you about 40% as far as a gallon of gas, even if you could burn pure ammonia in your engine--and from what I've read it still requires help from another fuel to sustain combustion. (That means two fuel tanks, which constitutes another major hurdle with consumers.)

Then there are the economics. Ammonia itself isn't exactly cheap, if you adjust for its energy content. The price of bulk ammonia for agricultural use appears to be around $550-$600/ton, which equates to $1.55-1.70/gal. But when you factor in its lower energy density, that raises it to at least $3.85/gal. of gasoline equivalent, without any fuel taxes. And while a distribution system exists to supply farms with ammonia, this is a long way from what would be required to fuel anything beyond farm vehicles. Because ammonia boils well below ambient temperature, it must either be refrigerated or stored under pressure, and dispensed through special equipment. And if all that weren't daunting enough for any service station owner considering adding an ammonia pump on the forecourt, the safety aspects of ammonia handling look even worse.

A glance at a typical material safety data sheet (MSDS) for anhydrous ammonia reveals that the recommended exposure limits are very low, under 50 parts per million in air, and the consequences of exposure include caustic burns and much more serious outcomes. Gasoline has its own issues, but spilling some on your hand won't send you to the hospital, and a larger spill or leak doesn't require first responders in hazmat suits. I simply can't imagine any fuel retailer wanting to take on the liabilities that would go along with this, even if there were an attractive margin in it, which there doesn't appear to be.

I concluded long ago that we're heading into a period of much greater fuel diversity, and that certainly seems to be true, with LNG catching on for big-rig trucks and CNG for a few cars but more fleet vehicles and buses, and even hydrogen appearing in a few places for fuel cell vehicles. However, it's very hard to imagine a substance with as many drawbacks as ammonia coming into wide use for consumers or even fleets. Our range of alternative fuel options seems sufficiently broad already, without having to consider a fuel that turns into a poison gas at atmospheric pressure and temperature.

Read More

Fuel Cell Déjà Vu

A small start-up company just launched an amazing new product to much fanfare: a novel fuel cell device capable of running on natural gas and potentially small enough to fit in your basement and power your entire home, replacing the electricity you buy from the grid. That proposition looks so compelling that builders will start incorporating them into new houses, so that the cost of the fuel cell would be absorbed into the mortgage a buyer takes out, making monthly power bills a thing of the past; you just pay your mortgage and your gas bill. Would it surprise you to learn that I'm not describing the "Bloom Box" fuel cell featured on last Sunday's "60 Minutes", but rather a home fuel cell designed by a small company in upstate New York called Plug Power Inc. in the late 1990s? Plug Power is still in business, and they still sell fuel cells of various sizes, including a home model, but the revolution in distributed power that their device was expected to launch hasn't occurred, at least not yet. The reasons why might shed some light on the hype surrounding Bloom Energy.

The late-90s' arrival of residential fuel cells that I mentioned above was a development that intrigued me in my professional capacity as a strategist and scenario planner for Texaco, Inc. Small fuel cells looked like a clever way to circumvent grid bottlenecks and reliability problems, using a platform that might eventually allow them to be built more cheaply and require less maintenance than either micro-turbines or the gasoline or diesel generators that dominated the small generator market. They also had the potential to increase the size of that market tremendously. (Rooftop solar was another attractive distributed power option, but without lots of expensive storage it wasn't and still isn't a recipe for 24x7 independence from the grid.)

I'm sure there are many explanations for the failure of home fuel cell sales to take off then or subsequently, including the high cost of the units, which was partly driven by the precious metals requirement of the Proton Exchange Membranes at the heart of these small fuel cells, which were similar to those being developed for cars. Bloom may have cracked this part of the puzzle by using lower-cost raw materials and choosing solid oxide fuel cell technology that can run directly on more complex fuels like methane, rather than requiring the fuel source first to be reformed into pure hydrogen--a step that adds to investment and operating costs and consumes some of the energy in the fuel, reducing overall efficiency.

Another key element of the economics of fuel cells relates to their operating costs, chiefly fuel. This was a problem for Plug and it remains a problem for Bloom, particularly at the residential level. While industrial users and commercial sites can negotiate gas supply contracts at competitive long-term rates that should allow cost-effective power production onsite, residential customers pay somewhat more and are exposed to significant seasonal and annual price volatility--much more than on electricity rates. Through November the average US residential natural gas price last year was $12.86 per thousand cubic feet (MCF). That's close to the weighted average I paid last year of $12.46, which was quite a bit lower than the $15.55 I paid in 2008, thanks to lower gas commodity prices. Based on that price and knowing the unit's "heat rate"--the amount of gas required for each kilowatt-hour (kWh) produced--I can calculate the fuel cost of power. At the stated 6,610 BTU/kWh, and using last year's US average residential gas price, that works out to $0.085/kWh. So even if the device were free, that's the least I'd have paid for electricity coming out of it last year. If you live in California or Long Island, that's pretty cheap power. However, if you live somewhere like Virginia, where my average electricity rate last year was just under $0.12/kWh, all-in, the savings would be much smaller. At just under 10,000 kWh per year of usage, that would have saved me about $340, setting a pretty low upper limit on what I'd be willing to pay for a Bloom Box, even after factoring in the various federal and state tax credits available.

Now, we can argue all of the benefits of producing your own power, particularly if you live in an area subject to power outages during storms or heavy snow. Self-sufficiency is an appealing idea for many. And there's clearly an emissions benefit here; just how large depends on your local generating mix. At 0.77 lb. of CO2 per kWh the Bloom Box beats the national average by about 40%, though it's hardly on par with rooftop solar or residential wind--a singularly expensive distributed energy technology--or indeed with what your regional grid emits if it includes a high proportion of hydro or nuclear power. Potential purchasers of Bloom Boxes will need to assess what such attributes are worth to them.

The enthusiasm that surrounds a new (or at least new-seeming) technology such as this is understandable, and I can't help being infected by it to some degree. At a minimum, it reminds me of how jazzed I was about these possibilities the first time I encountered them more than a decade ago. However, for Bloom and other small fuel cell suppliers to fulfill that potential, a lot of things will have to break their way, including moving rapidly down the cost curve to make these devices as cheap as possible, as well as some good luck concerning the overall economy, and particularly the housing market, especially its new-construction segment. Meanwhile, if the price of rooftop solar continues to fall, fuel cells could face stiff competition, while restrictions on the production of shale gas could boost natural gas prices and thus the net cost of electricity from a home fuel cell. I'll be watching Bloom Energy's progress with great interest as they attempt to develop this market.

Read More

Fuel Cell Trains

I often use my gym time to catch up on interesting podcasts, and NPR's excellent Science Friday series is one of my favorite sources. I just caught up with a recent segment on the development of hydrogen-powered trains, which seem like a particularly clever use of a promising technology that must still overcome serious obstacles in its automotive applications. But while I give the host, Ira Flatow, credit for pursuing the question of where the hydrogen for trains would come from, his guests' answers left something to be desired. That's not just because they tended to downplay the emissions associated with producing hydrogen, but because this omission might result in ignoring what could be an even better, more efficient fuel-cell configuration for trains and other large vehicles.

The basic idea of powering trains with fuel cells offers several important advantages--and one very serious disadvantage--for rail companies and their stakeholders. It also represents a less revolutionary change for rail than for automobiles, since trains are already partially or wholly-electrified, and a fuel cell is just another way to generate that electricity. Even the diesel locomotives that fuel-cell locos would be intended to replace are really diesel-electric hybrids. The key benefits of using fuel cells instead of big diesels for this application include substantial reductions in local pollutants, including soot, along with much quieter operation. Unfortunately, even if fuel cell trains could circumvent many of the infrastructure hurdles that have impeded automotive fuel cells, they still look prohibitively expensive. Diesels are pretty cheap on the basis of $ per kilowatt of generating capacity, while fuel cells are still much pricier, by at least a factor of 10.

Ignoring cost, fuel cell trains would face fewer obstacles to wide-scale deployment than fuel cell cars. As one of the program's guests pointed out, hydrogen storage, the Achilles heel of fuel cell cars, is not a problem in this situation. If necessary, a fuel cell train could carry an entire tank-car of compressed hydrogen behind the locomotive, and it wouldn't alter the train's performance or cost appreciably. That would also reduce the need for a widely-dispersed refueling infrastructure. For that matter, a train could carry along its own refueling set-up, in the form of an electrolyzer and compressor. It would require only fresh water--reminiscent of the coal-burning locos of yore--and a place to plug in. However, when you follow that plug back to its ultimate source, you find that the CO2 emissions of a hydrogen train could be quite a bit higher than zero, and possibly even higher than those of the diesel train it would replace, because our power generating mix is still dominated by fossil fuels.

So whether the H2 for a fuel cell train would be produced from natural gas, as most of the substantial quantity of industrial H2 in the US is, or from grid electricity, it results in CO2 emissions somewhere. In fact, because electrolysis of water into H2 is only about 80% efficient, the associated emissions of electrolytic H2 used to fuel a train would be 25% higher than the average of the grid power used to produce it. And although it's theoretically possible to generate H2 solely from off-peak renewable electricity when the latter is not being used to back out higher-emitting power sources, the capital cost of that route is much higher, because it would only operate a small fraction of the time. At least for the near-to-medium term, most H2 will likely be generated from natural gas, and that argues for a very different configuration for the fuel cell train than the one considered in this episode of Science Friday. Instead of using low-temperature automotive-design fuel cells, which require a source of pure H2, a high-temperature fuel cell of the type used for stationary power generation might make more sense. Not only do these operate more efficiently, resulting in lower overall emissions, but they can also run directly on natural gas and other light hydrocarbons, producing the H2 they require internally, rather than externally. In that case, the fuel tank for a fuel cell locomotive might just be an ordinary propane tank car, for which the entire supply chain is already well-developed.

If you've ever waited for a train in an underground or partially-enclosed station with several diesel locomotives idling away, you'll probably join me in wishing the hydrogen train test project team good luck with this initiative. The benefits of converting trains to fuel cells seem obvious, assuming this can ever be done at a competitive cost. At the same time, I hope the developers will take a broader view of hydrogen as not just another fuel, but as part of our overall energy ecology. That might lead them to an even more viable, beneficial result, with a better chance of showing up in real train yards, and eventually even passenger trains.

Read More

How Many Miracles?

Over the weekend I was catching up on articles, and one from Technology Review last week caught my attention. It was a brief interview with the new Secretary of Energy, Dr. Chu, covering nuclear power and fuel cells. In the back half, Secretary Chu explained why the DOE has cut funding for fuel cell R&D, suggesting that fuel cell cars were always a long shot, because they required "four miracles" to happen. Although I haven't mentioned fuel cells very frequently here in the last few years, I must say it's hard for me to consider the commercialization of something that I've already driven as requiring quite so many miracles as that. At the same time, I don't trivialize the obstacles that explain why fuel cell cars are still not available in large numbers, despite previous expectations--including my own--that they would be by now.

Dr. Chu helpfully breaks down the challenges facing fuel cells into four categories. Start with his concern about the principal source of hydrogen (H2) today, via extraction from natural gas. This route certainly undermines the "zero emissions" claim often attached to fuel cells. In practice, that means zero tailpipe emissions, but hardly zero emissions overall. Still, it's worth considering what else we could do with the natural gas in question. We could compress it and burn it in a modified internal combustion engine (ICE). T. Boone Pickens is quite fond of that idea, and it's not as foolish as some suggest, since it can displace a lot of petroleum and reduce emissions by 15-20% compared to a conventional car, on a well-to-wheels lifecycle basis. We could also use the Fischer-Tropsch process to convert natural gas to top-quality synthetic diesel at a somewhat smaller emissions savings, because the higher efficiency of a diesel engine is largely offset by the energy lost in fuel synthesis. Or we could produce H2, which as Dr. Chu notes involves throwing away about a third of the original energy in the gas by the time we've compressed the resulting H2. Yet the latter is the only one of these pathways that, despite the high energy price paid in producing H2, affords the opportunity to cut our overall lifecycle vehicle emissions in half, because producing electricity in a fuel cell is inherently so much more efficient than burning a fuel in an internal combustion engine. It's not perfect, but it's far from awful--unless you put the H2 into an ICE--and no miracles at all are required to make the H2.

Miracle number two involves H2 storage, and this looks a bit tougher. I am not keen on carrying around compressed gases at 5,000 or 10,000 psi in the same vehicle with my family, and that is no irrational fear. In my refinery days I saw examples of how much mechanical energy even 2,000 psi held, and I will never trust a Kevlar-wrapped tank enough to be fully comfortable with this option. Moreover, I don't think Dr. Chu is entirely correct that "compressed hydrogen is the best mechanism." ECD, a company that my former employer once invested in, has a technology for storing H2 via chemical absorption in metal hydrides, and you can buy canisters that use their technology today. The advantage of this system is that it doesn't involve high pressure. The disadvantage is that these hydrides are heavy, a drawback shared by the nickel-metal-hydride batteries (same basic technology) used in the Toyota Prius and other non-plug-in hybrids. None of these systems yet stores energy at the equivalent density (and thus driving range) of gasoline, but then neither do Lithium-ion batteries.

The third challenge concerns distribution, and this is a doozy. Transporting H2 in tube-trailers is fine for servicing demonstration refueling stations, but can't be scaled up to handle millions of cars. That may not be necessary, because the "reformers" that extract H2 from natural gas can be built on a scale that fits into a service station dispenser, drawing feedstock from local gas lines and delivering fuel without any need to transport it as H2, other than in the car. Installing such devices in thousands of locations would be a massive undertaking, but frankly the same can be said for the goal of installing E85 pumps at 10% of service stations, compared to about 2000 today. To me, cost-effective H2 distribution is a matter of engineering and economics, not scientific breakthroughs.

That leaves us with the one item on Dr. Chu's list that might qualify as requiring a genuine miracle: bringing the cost of a fuel cell stack down to a level comparable to an internal combustion engine, or at least to a point not so much more expensive as to render a fuel cell car inherently unaffordable. Fuel cells still cost over $1,000/kW--implying that just the fuel cell stack for a real car would cost more than an entire luxury car today. Forecasts that this would fall to anywhere near the roughly $35/kW of today's car engines remain theoretical, relying mostly on learning-curve effects analogized from other industries. Given the current state of the car industry, manufacturers will struggle enough just absorbing the initial high cost of low-volume plug-in hybrid car production, without taking on tens of thousands of dollars in losses per car for vehicles like the Honda FCX Clarity. Absent a breakthrough, an investment like that might truly require a miracle.

Whether making fuel cell cars a practical reality requires four miracles or only one, I have to agree with Dr. Chu's conclusion that their commercialization looks neither imminent nor assured. It's important to recall that hydrogen is merely another energy carrier, like electricity, rather than an energy source like petroleum or biofuels. The smart money today is on battery-electric cars, including plug-in hybrids. In order to beat them an automotive fuel cell stack must cost less than the battery pack required to give drivers the 250-300 mile range they seem to want, because in every other respect that matters a fuel cell vehicle is an electric car. However, we must keep in mind that the smart money is not always right. Cutting back federal R&D on fuel cells to a level that puts a higher priority on other options that can deliver meaningful results sooner seems prudent, as long as we don't abandon this option entirely, or cede our competitive position to others.

Read More