After more than two decades of hype about the imminent arrival of a transformative “hydrogen economy”, many veteran technology watchers — myself included — had concluded that hype was pretty much all it was. Hydrogen fuel cell vehicles, in particular, looked like a failed dream. Bright innovators like Canada’s Geoff Ballard had attacked the problem and burned through serious investment money trying to develop a product that could stand up to the rigors of the automotive market. All with little success. And beyond the cost and durability issues of fuel cells themselves, the hydrogen storage issue stubbornly resisted commercially practical solutions.

In recent years, hybrids and battery electric vehicles have appeared to hold the inside track for low carbon and zero carbon transportation. Tesla has reshaped perceptions of what is possible for battery electric vehicles. The cost of lithium-ion battery packs has been driven down, while capacity, performance, and reliability have increased dramatically. To be sure, government programs have continued to fund fuel cell R&D. If nothing else, fuel cells still hold broad appeal for military programs. But to those of us who felt we understood the issues, the barriers to broad use of hydrogen as an energy carrier looked pretty fundamental. We — or at least I — didn’t really expect to see them fall anytime soon.

In the face of that expectation, a spate of announcements and news articles over the past year relating to hydrogen have come as a shock. Most prominent have been recent announcements by Toyota, Honda, and Hyundai of new FC vehicles for production release in markets where hydrogen refueling stations are available. Toyota announced the Mirai, Honda announced the Clarity Fuel Cell, and Hyundai announced the Tucson Fuel Cell SUV. But those were just the commercial announcements backed by ad campaigns. When one starts digging, scores of significant news stories and announcements from around the world turn up. The whole idea of the hydrogen economy — which never quite went away — seems to be resurgent.

So what is it that happened while I wasn’t paying attention? A thorough review seems in order. Since battery vs. fuel cell EVs are at the eye of the storm, I’ll start there. Then, time permitting, I’ll go on to look at some of the broader issues of energy storage and hydrogen production.

The EV technology race

Despite disappointing progress in the early years of the Bush “Freedom Car” initiative, fuel cell R&D never dried up. It has been ongoing, and not all of it has been politically driven. There has always been genuine promise in FC technology. The problems have been cost and durability. As far as I can tell, there have been no singular technology breakthroughs behind the resurgence of interest in hydrogen. But persistence and general advances in materials and manufacturing have begun to pay off. A small example: automated machinery able to make reliable gas-tight welds between thin sheets of metal. That turns out to be crucial for fabrication of efficient bipolar plates in PEM fuel cells.

Analysis by DOE’s Fuel Cell Technologies Office puts present cost of automotive FC stacks at $53 per kW for manufacturing volumes of half a million units annually. That’s half of what was projected for the state of the art in 2006.

Ironically, one thing widely seen as needing to change before FCEVs could become practical has stubbornly not changed: technology for carrying hydrogen on-board the vehicle. Despite a plethora of promising lab developments, there seems to have been no practical breakthrough in hydrogen storage. The new FC vehicles all use high pressure gaseous hydrogen stored in polymer-lined, fiber-wound pressure tanks. Similar tanks were made by Quantum in the 1980s. The tanks remain heavy, bulky, and costly. However, with better manufacturing methods and stronger, cheaper carbon fibers, their cost now measures in the low thousands of dollars rather than the high tens of thousands.

FC Advantages: weight, capital cost, refueling time

From Toyota’s product sheet for the Mirai, the fuel cell system delivers 2.0 kW/kg with a power output of 114 kW max. That implies a FC system weight of 57 kg. The hydrogen tanks hold 5 kg H₂ at a weight percentage of 5.7%. That implies a tank weight of 83 kg. So, 145 kg total for tanks + FC system + 5 kg hydrogen, delivering an EPA estimated range of 312 miles. That compares to 540 kg for the battery pack in a Tesla Model S with a rated range of 265 miles.

It appears that despite the heavy and bulky pressure tanks, the Mirai delivers a greater driving range than the Model S, with roughly a 4:1 weight advantage for the energy delivery system. More important for most buyers, however, will be the system cost per kWh to the drive motors. That’s harder to nail down, because manufacturers don’t normally release cost data publically.

There’s a small cottage industry devoted to guessing and predicting the cost of Tesla’s battery packs. GTM Research projects that by 2020, Tesla’s average cost for packs will be $217 / kWh. Using that figure, the 85 kWh Model S battery pack would come to $18,500. That’s less than some estimates, but more than the $12,000 that Tesla itself is willing to guarantee to Model S owners as the replacement cost after 8 years. Everyone agrees that costs are on the way down as production from new battery “gigafactories” kicks in, so $18,500 is probably a reasonable figure to use for near term comparisons between battery and fuel cell vehicles.

On that basis, fuel cells appear to come out ahead of batteries on cost as well as weight. At $53 per kW, the Mirai’s 114 kW fuel cell system would cost just over $6000. The high pressure storage for 5 kg H₂ is probably around $3000. So the capital cost of the Mirai’s energy delivery system with longer range looks to be roughly half that of the battery pack for the Model S.

Of course, FC vehicles are also much faster to refuel. That’s widely considered their strongest market advantage. But it presumes a network of public hydrogen refueling stations that for the most part does not yet exist.

Normally a “chicken and egg” problem like that would be lethal for a new product introduction. It may prove to be so in this case as well. However, there are some special factors for hydrogen that could potentially enable it to break through. We’ll get to those. First, though, we should look at other issues on the flip side of fuel cells relative to batteries.

FC disadvantages: efficiency, carbon emissions, fuel cost

There are many ways to produce hydrogen. For electrification of transport, the green vision is that it would be by electrolysis of water. That vision is promoted for hydrogen fueling stations. The H₂ to be dispensed each day would be produced on-site the same day or the day before by electrolysis. That reduces on-site H₂ inventory, enhancing safety, and minimizes the capital cost of the station. It also avoids the need for new and costly infrastructure to distribute hydrogen. No need to either dig up the streets to lay hydrogen pipelines, or have liquid hydrogen tanker trucks mixing with city traffic.

In that scenario, the relatively low efficiencies of PEM fuel cells and electrolyzers put fuel cells at a distinct disadvantage relative to batteries. For each kilowatt-hour delivered to the drive motors of the vehicle, the electrolyzer / fuel cell system requires roughly twice the kilowatt-hours of energy input as the battery system.

The rough 2:1 difference in electrical load that FCEVs impose is bad enough, but it also carries over to the indirect carbon emissions of the two classes of vehicles. In terms of what they emit on the road, both BEVs and FCEVs are zero emission vehicles. Both, however, inherit indirect emissions via the power grid. If the grid were supplied entirely from carbon-free power sources, then both BEVs and FCEVs would be carbon-free as well. But that’s far from the case today. A 2:1 difference between FCEVs and BEVs electrical load means that an FCEV will have double the indirect carbon emissions per mile of a BEV.

The actual difference in fuel cost per mile will be quite a bit greater than the 2:1 difference in electrical load suggests. For BEVs, the fuel cost is just the cost of the electricity consumed in charging. There is no capital equipment of any significance between the vehicle and the power grid. For FCEVs, however, there’s the electrolyzer, hydrogen storage, dispensing system, and the commercial property hosting the station. There is also the daily operational overhead of running the station. Those elements raise the retail cost of hydrogen dispensed well beyond the cost of electricity to the electrolyzer.

Solid estimates of what can be expected in the near future are hard to come by. A jumble of subsidies confuse the picture, and estimates for future costs are sensitive to assumptions about rates of adoption, size of refueling stations, and the technology used for supplying H₂. DOE’s aspirational goals for 2020 are a wholesale production cost of $2.00 or less / gge (gallon of gas equivalent; ~1 kg of H₂). The goal for price at the pump, exclusive of taxes, is $4.00 or less (ref. here).

The bottom line is that fuel costs for an FCEV will be at least 5 to 10 times more than for a BEV for some years to come. I doubt that zero-carbon electrolytic hydrogen will ever be less than 4x as expensive. However for context, the fuel costs for a BEV are a fraction of those for a gasoline vehicle and are usually considered negligible. If the cost of hydrogen in an FCEV were 4 times higher than the per mile cost of electricity in a BEV, most drivers would find it acceptable. Fuel would still be a small part of the overall cost of owning and driving a vehicle. Witness to that is the fact that manufacturers of FC vehicles can afford to bundle free hydrogen into the purchase price or lease terms for the vehicles in their California test markets.

Situation in flux

The relative advantages and disadvantages cited above for FCEVs vs. BEVs are mostly soft. They’re subject to changes in technology, design approach, and use patterns. For example, developments in battery technology and manufacturing will almost certainly trim the upfront cost and weight disadvantages of BEVs. At the same time, changes in hydrogen production methods could reduce the per-mile cost disadvantages of FCEVs. There’s also an easy FCEV design change that would substantially reduce their cost of driving and mitigate the H₂ infrastructure challenge. (See below.)

Perhaps most significantly, the arrival and spread of autonomous vehicle capabilities will transform the automotive market in ways that significantly affect the tradeoffs between hydrogen and batteries. I’ll talk about that later.

Universal hybrids?

Controversy over batteries vs. fuel cell aside, there’s consensus on one aspect of future vehicle technology. Electrical motor-generators and solid state power controllers will increasingly be at the heart of drive systems. They make for cheaper, more reliable, and higher performance than mechanical transmissions and engine-coupled drive shafts. Ultimately, all future vehicles will be either pure BEVs or hybrids.

It’s not touted, but the new FC vehicles are, in fact, already hybrids. Toyota’s Mirai is built atop the Prius’ drive system. The two share many components, including traction battery and power controller. That enables regenerative braking and instant throttle response. It also buffers the FC system and reduces its cost. Commonality of components with Prius and a more benign FC environment are key parts of how Toyota limited its costs in fielding a new FC vehicle class.

All it would take to produce a plug-in hybrid version of the Mirai would be addition of a plug-in charging port. The same is likely true of Honda and Hyundai FC offerings as well. But batteries and fuel cells are competing for mindshare in the EV marketplace; it’s understandable that companies backing an FC play don’t want to expose the HEV roots of their flagship FC vehicles. It wouldn’t make marketing sense. A charging port does make technical sense, however. Local miles could be driven mostly in battery electric mode. The cost per mile would be low. Hydrogen consumption for a typical driving profile could be cut by half or more. In Europe, Symbio FCell has in fact taken that approach for a range-extended Nissan e-NV200 van for the taxi market.

A plug-in hybrid capability mitigates the hydrogen infrastructure issue for FC vehicles. They remain drivable even in areas without hydrogen refueling stations. The limited plug-in battery range might be a pain, and drivers would still want to have hydrogen refueling available near their home base. But they wouldn’t be tightly tethered to that base. The plug-in capability would provide flexibility, drawing from on-board hydrogen to extend range between plug-in chargings, or drawing on plug-in charging to extend range between hydrogen fill-ups.

The switch to electric drive changes the tradeoffs between batteries and fuel cells. It’s no longer a stark either-or choice. If electric drive and at least a modicum of battery capacity are givens, then the issues become how much battery capacity to have and what technology to employ for delivering extended range beyond what the hybrid drive battery supports. If the latter is enough to let local miles be driven mostly in battery electric mode, then the optimal solution for extended range is one that minimizes added vehicle cost. That holds even at the expense of higher fuel costs for times when the extended range capability is tapped.

Alternative fuels

It’s possible that neither large batteries nor hydrogen fuel cells are optimum choices for range. With large batteries, capacity above and beyond the needs of local driving may be a costly way to achieve an infrequently tapped range capability. And while future batteries will be lighter and cheaper, that also makes it attractive to offer more capacity for local driving. Increased capacity in the basic battery pack reduces the frequency of resort to the extended range capacity. The conceptual simplicity of having a single large battery may not be worth the cost. Separate subsystems could deliver greater range at lower vehicle cost, while enabling fast fueling as a bonus.

The separate subsystem for extended range might or might not be hydrogen. The added vehicle cost of the hydrogen approach looks like it would be about $9,000; that’s not small, but it’s not all that far above the cost of an IC engine and the various subsystems around it. The question is, what would it be buying?

With a fully decarbonized electricity grid and electrolytic hydrogen, the HFC approach would be buying carbon-free transportation. Yet if addition of easy and ubiquitous plug-in capability with larger hybrid drive batteries has already enabled most local miles to be driven in battery electric mode, then carbon emissions have already been slashed. If average fuel consumption for new plug-in vehicles is already 150 mpg or better, then the incentive to use hydrogen will be weak.

Barring a major breakthrough in hydrogen storage technology and further reductions in fuel cell cost, the default competitor to both batteries and fuel cells for extended range driving will likely be gasoline or compressed natural gas. Perhaps, if the price of fossil carbon emissions gets high enough, a carbon-neutral synthetic fuel might prove cheaper and more competitive. The energy cost of producing synthetic fuels from CO₂ and H₂ isn’t much greater than that of H₂.

Heavy Transport

The discussion so far has been about passenger cars. For a broader view of the hydrogen economy, we need to consider heavy transport as well: trucks, buses, trains, ships, and airplanes. Not to mention farm and heavy construction machinery. For the sake of brevity, I won’t cover any of the latter here. But trucks and buses play big roles and warrant comment.

For trucks and buses, the factors favoring hybrid electric drive systems are at least as strong as they are for passenger vehicles. The ability to deliver smoothly controlled torque for acceleration and uphill driving across the full speed range, with attendant capacity for regenerative braking, are attractive. Electric drive can deliver performance and safety advantages, along with fuel economy, clean air, and quiet operation. Low production volume for the heavy duty batteries, power control units, and motor-generators have hampered widespread adoption so far, but things are changing.

For energy supply to the electric drive system, there are different tradeoffs and different options that may be favored, depending on the application sector. All-battery approaches are attractive for metropolitan buses and utility trucks. Metro buses spend hours parked each day, either in their barns at low service times or at route ends while drivers change or take rest breaks before starting their next scheduled runs. It should be relatively easy to provide fast recharging at those points. The on-board batteries should never have to deliver more than about 25 miles in regular service.

For long-haul trucking and inter-city buses, all-battery approaches are currently impractical — and likely to remain so. Hydrogen has potential opportunities there. The recent unveiling of the prototype for the Nikola One electric semi (pictured below) has, in fact, caused quite a stir.

Credit: Nikola Motors

The truck is a hydrogen FC model, and its specs are quite impressive. 1000 horsepower (twice that of a diesel semi), 2000 ft. lbs torque, range of 1,200 miles, … If Nikola Motors can deliver on its promises, it will have a winner. Production deliveries aren’t scheduled to start until 2020, but truckers have already been plunking down $1500 deposits for reservations.

The high cost of electrolytic hydrogen will still make the per-mile fuel cost for a Nikola One relatively high — assuming that Nikola Motors is even able to deliver on ambitious plans to build solar farms for supplying its trucks with zero-carbon hydrogen fuel. The financial case for the vehicles would probably be stronger if they ran on compressed natural gas rather than H₂. They would still be hybrids — the Nikola One is planned to carry a 315 kWh battery that will give it the power to maintain 65 mph up a 6% highway grade and soak up the energy of descent from a mountain pass without touching the brakes — but it would lose its cachet as a hydrogen fuel cell vehicle.

It could retain some of that cachet if the Nicola One used high temperature SOFC fuel cells that run directly on methane. That’s an approach recently demo’d by an alliance between Ascend Energy and Atrex Energy. High temperature SOFCs are at least as efficient as PEMFCs, and if their high temperature waste heat is used to power a Brayton cycle turbine, they are a lot more efficient. The combination would certainly make for a low-carbon vehicle. To be zero-carbon, however, the methane burned would need to be from a carbon-neutral source.

Further topics

I haven’t yet covered the likely impact from autonomous vehicle developments, nor have I talked about different technologies for hydrogen production, or the use of hydrogen for energy storage and backing of intermittent renewables. Those are important topics, but I’ll leave discussion of them for part 2, next week.