I’m quite confident that hydrogen is a dead end for cars. My reasons aren’t personal- they’re thermodynamic and rather difficult to argue with. They’re presented here, in detail, with references:


In that paper, I point out that even if you’re very generous toward hydrogen, you end up with an inescapable conclusion: a hydrogen fuelcell vehicle will use, at minimum, 2.4 times as much energy as a battery electric vehicle of similar size and features. That’s the best case, and the reasons for that are thermodynamic and hence “difficult”- in the Japanese sense of that word- to overcome in a meaningful way. We’re not getting that down to 2x, for instance, ever, irrespective of what we invent or how clever we are.

The other obvious problem is the absence of any “green” hydrogen to fuel such a vehicle, even if it were a good idea:


Electrical grids aren’t all renewable or non emitting yet, but they’re doing a darned sight better than H2, which is still 99% produced from fossils without carbon capture.

You’ll see some papers were written that talk about how the viability of a fuelcell vehicle changes if we extend the range of the vehicle- at very long ranges, EV batteries get so heavy that hydrogen in comparison starts to look better. Are they right?

We can do better than that! We can compare two EXISTING cars: the Toyota Mirai Mark 1 hydrogen fuelcell vehicle, and the Tesla Model 3 long range version. You can buy both of them- at least in California. They’re both real cars, not theoretical constructs.

So let’s compare them!

The Mirai Mark 1 and the Model 3 are similar cars in terms of size, passenger and cargo capacity etc. Surprisingly the Mirai tips in at 200 pounds heavier than the Model 3 even with its comparatively huge 75 kWh battery pack. But in terms of performance, the Model 3 has the edge, by a long shot. Its motor delivers 211 kW and a 0-60 of 5.1 seconds, versus the Mirai’s 113 kW peak power and 9 second 0-60 mph. Yes, the Mirai IS an electric car- but its smaller hybrid battery pack isn’t up to the power draw which comes effortlessly (i.e. at a very low C rate) from the Model 3’s giant pack.

The Mirai is more expensive to buy, despite reportedly still being subsidized by Toyota, whereas the Model 3 is reportedly already making Tesla a profit. Both cars are WELL outside MY price range! Further drops in price of both FC and EV technology are expected in future of course.

The ranges of the two cars per the EPA “mixed” driving cycle are similar- and both are similar in range to that of a comparable gasoline car.

Both cars will lose range in cold climates. The Mirai apparently uses waste heat from its fuelcell for cabin heating, but it has another weird feature- it needs to use a heated blower to “dry out” its fuelcell in cold weather to keep it from freezing up.  So both cars will have some winter range loss, for sure. Note that all cars lose 10-20% of effective range on a tank (i.e. fuel efficiency) in winter, even run of the mill ICE cars. Note also that both cars will need some standby heating: the Model 3 to keep its pack warm enough so its electrolyte doesn’t get viscous, so it can accept and deliver charge when needed, and the Mirai so that its fuelcell performs properly and its product (water) doesn’t freeze.

The Mirai refuels faster- in 3-5 minutes once connected to the pump. Of course, that assumes you have a hydrogen filling station on the route between home and work. In contrast, the Model 3 can be recharged at home, leaving the use of Tesla’s supercharger network for longer trips. But on longer trips, you can expect to spend some time refilling the Model 3’s 75 kWh battery pack. Tesla’s superchargers are about 120 kW peak and hence should refill the Model 3 to about 80% SOC in about 30 minutes. Honestly, a 30-minute break after 4.5 hours of highway driving is likely a good idea for safety anyway.

As you know, my primary interest in this is energy efficiency and GHG emissions efficiency, and here’s where the comparison becomes rather stark!

The Mirai, as my other paper shows, is likely about as source-to-wheels energy efficient as a Toyota Prius gasoline hybrid. If driven with fossil-derived hydrogen, its GHG emissions from source to wheels per mile driven will be nearly the same as a Prius running on gasoline- modestly better in the best case. If the hydrogen has to be transported a long distance or is distributed using an energy-inefficient means such as a tube trailer, results for the Mirai will rapidly worsen. So if the source of hydrogen is fossil natural gas, it’s safe to say that the only environmental benefit you’re getting from the Mirai is the absence of a local tailpipe. That’s a toxic emissions benefit, especially in a densely populated, emissions-sensitive region like California. At least the SMR at the hydrogen plant has a tall stack and emissions monitoring equipment on it.

The same is true if you were to refill the Model 3 with electricity derived from a 30% efficient last-generation coal-fired power plant. Then you’re likely better off with the Prius running on gasoline.

However, remember that the reason people are interested in both hydrogen and battery EVs is the fact that we CAN make electricity from low-emission sources- thereby eliminating BOTH fossil GHG AND toxic tailpipe emissions, making personal transport greener and more sustainable. Electricity in my own province of Ontario, Canada, is less than 40 g CO2/kWh and is only 8% fossil fired- all natural gas, no coal. Electricity is generated largely from nuclear, hydro and wind. Grids in BC, Manitoba and Quebec are even greener and lower in CO2 emissions, being principally hydro- and those four provinces contain over 75% of Canada’s population. Even California’s grid is greener than the average grid in the US. Grids worldwide are getting greener daily.

When you’re collecting a “free” resource like wind or solar energy, efficiency matters less than the average cost of electricity in dollars per kWh delivered over the project life. But once you’ve made electricity, the efficiency with which you use it ALWAYS MATTERS. And that’s where the Model 3 comes out ahead- way, way ahead.

As my previous paper makes clear, hydrogen’s 3-step process- hydrogen generation, storage at high pressure and then generation of electricity using a PEM fuelcell- means that its energy storage cycle efficiency is terrible. Best case, it means hydrogen will require 2.4x as much energy as if you used a battery instead. Does that hold when comparing the Mirai and the Model 3, or did I get it wrong in my calculations?

It does get a little tricky to make these comparisons fair and properly, but I’ve tried my best- and as always, will revise my paper if I receive credible, referenced alternative information.

First of all, I HATE the use of the MPGe figures. Per the EPA website and this handy table:


the MPGe figures are calculated using the following equivalences: 1 gallon of gasoline equals 1 kg of hydrogen equals 33.7 kWh of electricity. But these figures commit the 1st Sin of Thermodynamics- they compare a joule of chemical energy or heat with a joule of thermodynamic work or electricity (which can be converted to work with very high efficiency) as if they are worth the same, just because they have the same units! And that is just simply WRONG- it leads to incorrect conclusions.

Looking at the MPGe figures, you’d conclude that the Mirai’s 66 MPGe is better than the Prius’s 52 MPG figure on gasoline- and it is. You’d also conclude that it’s worse than the 130 MPGe figure for the Model 3- and it is. But does the Mirai use 130/66 = 1.96 times as much energy per mile driven? Nope- it’s way, way worse than that!

You have to be careful doing the calculations. The EPA figure of 260 Wh/mile given for the Model 3 long range checks out against the 130 MPGe at the 33.7 kWh/gallon figure they use (which again confuses J of chemical energy with J of work). But when you divide the pack’s capacity of 75 kWh by the range estimate of 334 miles, you get a lower figure- 224 Wh/mile. The latter figure is an “ex-pack” figure, i.e. out of the battery, whereas the former is a “from the wall” figure, including the efficiency of the charger and the charge/discharge cycle efficiency of the battery. These figures are fairly close to what I managed with my E-Fire, i.e. 260 Wh/mile out of the wall and 235 Wh/mile out of the pack- impressive for a car which weighs almost twice as much as my E-Fire did.

The Mirai’s figure similarly checks out- 66 MPGe is similar to 5 kg of H2 LHV consumed over 312 miles. However, it includes NOTHING for the inefficiency of making or storing hydrogen.

When you include the North American grid average loss of 6% to both cars, and then use a world class 70% for the efficiency of an electrolyzer (or SMR, if you want to use methane) in terms of LHV of hydrogen per unit energy fed, and 90% for a large, efficient hydrogen compression train, you soon discover that the Mirai uses 3.2 TIMES as much energy from source per mile driven as the Model 3.

When you look at costs, it’s even worse. Retail hydrogen in California sold, at the time this was written, for $15/kg (today’s price, March 2023, is $26/kg at the hydrogen pumps in California…), and that’s hydrogen which is only 33% renewable- the remainder is much cheaper fossil-sourced hydrogen, burdened with a modest carbon tax (California has a major emitter cap and trade program which it shares with Quebec and, until recently, with Ontario). Refilling the Mirai cost $75 US- to drive 312 miles. In contrast, you can completely refill a Model 3 on Tesla’s supercharging network for $18, at $0.24/kWh (measured presumably onboard the car’s battery, not from the wall). Using the average retail price of electricity to recharge at home in California ($0.15/kWh), it’s even a little cheaper- $14. Off-peak, at night, presumably it’s even cheaper still. The Mirai is on the order of 5.4 times as expensive per mile drove, relative to the Model 3.

And that, folks, is why I think hydrogen is a dead end for cars. Who is going to pay 5 times as much per mile driven, just to be able to refill their car faster? In a world where people will buy filthy diesel to save 20% on their refuelling cost?

UPDATE:  Toyota introduced a Mark 2 version of the Mirai in 2021.  Hydrogen capacity increased, by adding a 3rd hydrogen tank, at the cost of some interior room and about another 200 pounds of curb weight.  Range per the EPA test is now 402 miles- more than the Model 3, but not by enough to be jaw-dropping.