Why fuel cells are (probably) not the future of mobility
Fuel cells are a great technology, but they have become something of a zombie among the alternatives to conventional gas-powered engines. Why? Because they’ve been around for a long time, they never managed to get on their feet and they just won’t die. For years, people have been trying to establish the fuel cell car as a mass market alternative to traditional gas cars and and battery electric vehicles (BEVs) with little luck.
The reasons for this failure are basically four-fold. The necessary infrastructure is badly underdeveloped and would need a massive boost in terms of subsidies to get off the ground. Efficiency is a pain point for the whole proposed hydrogen economy, but especially if we assume that hydrogen will be made with renewable energy. The added complexity (compared to BEVs) of a fuel cell car means that they will be more expensive than BEVs.
#1 Infrastructure
To talk about hydrogen for mobility, you have to talk about its direct competitor: electricity. Electric sockets are almost everywhere. In fact, most of us will be used to find an electric socket in or around a house with less than 1 min of search.
These electric sockets are fed by a distributor system that taps the regional infrastructure. This in turn is connected to overland lines and finally to electric power plants (skipping some steps for simplicity). My point is that electricity is essentially everywhere. A lack of efficient charging infrastructure for battery cars is only a reflection of a lack of political will. It’s not a universal law, nor is it an impossibly expensive problem. Furthermore, electric infrastructure is not just almost universally accessible, it’s usually planned with redundancy and stability in mind.
Adding charging stations for BEVs is entirely possible without endangering the stability of the grid. If something is lacking, its political will.
Let’s now turn to hydrogen. Almost none of the necessary infrastructure currently exists for a hydrogen economy. That is because hydrogen adds a whole dimension of generation, distribution and storage problems due to its aggregate state (much like oil and gas).
Starting with generation, currently, there is just not enough capacity to produce hydrogen at scale from renewable energy resources to really fuel a revolution. Thus, the first problem we are facing to move to a hydrogen economy is a massive investment in capacity. Another one is the layer of inefficiency that the production of hydrogen adds to the energy generation. But we will go into this a bit later.
The second layer of infrastructure that would have to be built is logistics and distribution. Oil and gas tend to come from regions where water, currently the true raw material for hydrogen, is not abundant. While the most common way to generate hydrogen is to dissolve natural gas into hydrogen and remaining carbon products, we cannot rely on this process to ensure a livable climate in the future.
Therefore and generally speaking, we wouldn’t produce hydrogen in the places where we currently produce oil and gas and we cannot rely on the already existing pipeline infrastructure. Even then, it seems like a massive retrofitting effort would be necessary to make sure that the hydrogen doesn’t escape from these pipelines, as it is a much smaller molecule than even the smallest hydrocarbon.
Generation, storage, distribution and last mile logistics add large challenges that haven’t been solved satisfactorily for a hydrogen economy at scale to succeed.
Even at the level of last mile logistics, i.e. the distribution of the fuel from pipeline to consumer station (“gas station”), we would need to invest in new containers to move the hydrogen. The consumer infrastructure would also have to be extended, i.e. there are very few gas stations currently offering a way to fuel up on hydrogen.
However, none of this infrastructure really exists. That means people espousing the idea of a hydrogen economy must be expecting huge subsidies to build it. Without these, a fast hydrogen revolution is hardly possible. Please note, I’m not saying it’s impossible, it just seems like added baggage compared to the already existing electricity distribution network that we could build on top of. This added baggage produces delays and stops us from pursuing the most important reason for switching to hydrogen: The move away from hydrocarbons.
#2 Efficiency
Generation of hydrogen is lossy. With the general picture of the hydrogen distribution system in mind, let’s imagine where we will have to deal with inefficiencies along the supply chain. Before we launch into the inefficiencies of hydrogen infrastructure, let’s quickly meditate on the following information:
Currently, 95% percent of hydrogen is generated through steam reforming of natural gas, partial oxidation of methane and coal gasification [1] with the rest being generated from renewable sources.
We will not be able to continue this sort of generation if we want to maintain a habitable planet. So, for the sake of the argument here, we should be charitable: We assume that in the future hydrogen generation from hydrocarbons will be phased out and that all generation happens from biomass or electrolysis of water at 100% renewable energy (which won’t be remotely true for a long time).
With this assumption, it’s easy to see why a hydrogen economy is energetically expensive. Electricity generated from a renewable energy source would be used to power a hydrogen plant that uses electrolysis of water as its generation principle. Electrolysis of water has a maximum theoretical efficiency of 82–86% which is a reasonable value for an industrial energy conversion process.
To go from electricity to hydrogen, we lose between 1/6 and 1/5 of the available energy outright.
For comparison, the loss in usable energy when bringing electricity from power plant to consumer is roughly 8%. All other efficiency losses are associated with the inefficiencies of electrical motors and batteries.
Next comes storage. Here, there are two commercially viable options suitable for later distribution (as we are looking at solutions for mobility, we will not be discussing underground storage and other stationary methods here): liquid hydrogen storage or compressed hydrogen.
Liquid hydrogen needs to be cooled down to very low temperatures to be stored. This cooling costs a massive amount of energy at the order of about 12% of the stored energy. After storage, this leaves 72-76% of the renewable energy used for generation. Compressed hydrogen is much better here, with a loss of only about 2% of the energy stored. However, the drawback is that tanks storing compressed hydrogen will be bigger and heavier (they are pressurized containers) than the tanks we use nowadays to ship hydrocarbons. Additionally, less volume can be shipped per container translating into higher prices per unit. Still, with compressed hydrogen, we are at 80-84% of the energy used to generate the hydrogen.
Whatever way we store the hydrogen (I’m not excluding a new breakthrough for this), we will have to ship it to destination adding the energy cost of logistics. While it’s hard to find numbers, we can estimate the energy cost of shipping crude oil by making some rough assumptions: A tanker moves about 200,000 tons at a fuel consumption of 200 tons of bunker oil per day [3][4]. For a 10 day trip, we are looking at a loss of about 1% of the energy shipped. Tankers can make trips of up to 70 days, leading to an average loss in the 3-5% range just from maritime shipping, without the added consumption of shipping fuel from ports to final destinations (last mile). Liquid hydrogen has a lower energy density by volume than hydrocarbons by a factor of four. Thus we are looking at a much higher loss (up to 20%) per unit of hydrogen shipped if the shipping network is also using hydrogen as a fuel. If we assume the shipping of compressed hydrogen, the numbers are even worse.
We therefore can conclude that there is between 64–74% of the energy used to create the hydrogen left after production, storage and shipping at current technology levels. Note that, regardless of the method, the EROI (energy returned on energy invested) is always below 1.
Energy returned on energy invested is always below 1. The generation itself is not a power generation process but a lossy energy storage process.
Finally, the fuel cell. Assuming no other losses on the way from production to tank, we need to take into account the efficiency of a fuel cell. These are usually in the range of 50–60% conversion efficiency with a theoretical maximum of 85%. That last number however includes cogeneration of heat. In the mobility sector, heat is effectively waste, so we can safely go with 60% as a maximum.
Forgetting about friction and other losses that all cars have in common, we end up with an energy efficiency of about 32–44% (electricity generation to road).
#3 Complexity
The hydrogen car is, due to the fuel cell, a more complex technology than the BEV. For many manufacturers and also the adjacent industry of service providers, this is not necessarily a drawback. One of their main driving forces and selling points in the last century was higher complexity. Take the standard gas engine. It is essentially a containment chamber designed to house miniature explosions of a hydrocarbon-air mixture. The resulting expansion is translated into a lateral movement through pistons. Their movement is transformed into a rotation movement using a crank shaft. Even this most basic function is an incredible engineering feat. The more you think about, the more amazing it becomes.
This transformation of hydrocarbon fuel into rotation is achieved using an impressive array of parts (pistons, crankshafts, gears, …), all of which have to sustain at times massive loads and stresses. Due to temperature development, there is a need for coolant and an air-circulation system (explaining the front grille of conventional cars). A lubricant is needed to ensure that there is as little friction within the engine as possible, as otherwise the engine would grind itself to metal shavings.
It is therefore no wonder that there is an $4 trillion industry connected to cars: Parts manufacturers produce replacement parts, the maintenance and repair sector (MR) in the form of garages replace parts, lubricant, coolant and many more things that ensure the flawless operation of the engine. Not included is of course that incredibly large hydrocarbon industry that provides cars with the fuel they burn.
By contrast, a BEV is a much simpler machine. To begin with, there is no gas tank, just batteries usually sitting in the bottom plate of the car which has the positive side effect of a lower center of gravity. With maximally four electric motors turning electricity directly into the rotation of the wheels without having to go through the additional step of containing the ignition of the fuel-air-mixture, many (MANY!) of the parts of a traditional car just vanish.
At the risk of oversimplification, an electric car is rather similar to a coffee grinder or a sewing machine in its basic complexity — and these machines tend to run for years and years before breaking. Even then it isn’t usually the motor that breaks.
This is the reason why electric cars need much less maintenance. Fewer parts that can break means fewer parts that need monitoring, maintenance, and replacement. This is a worry for the MR sector, as it effectively lives off the complexity of the hydrocarbon economy. Reducing complexity means reducing their revenue. Which in turn reduces the operational costs of a BEV. And of course its potential to create waste. In short, BEVs are a consumer and an environmental win. An unusual combination in today’s economy.
The hydrogen car adds complexity to this simplicity by adding two steps to the generation of rotation compared to the BEV: First there is a fuel tank, which is a pressurized gas container. Then there is the fuel cell itself, that turns the hydrogen into electricity. Finally, there is the motor that uses this electricity. Despite all the advantages of faster fueling and longer distances this is a considerable increase in complexity compared with the simpler battery-electric vehicle.
This added weight also fuels our lowest instincts: While fuel cells and tanks are heavy, economy is not as critical as in BEVs, leading to larger, heavier, bulkier and especially less efficient cars.
The goal should not just be a change in technology for the mobility sector. The goal should be efficiency and this is just not a given with a fuel cell.
#4 Technology
Before finishing, let’s quickly investigate the fuel cell itself. Actually, it isn’t a particularly new technology. It was invented in the first half of the 19th century and subsequently refined and improved on. There is no doubt that it is an interesting technology, but is it applicable to the mobility sector?
First of all, battery vehicles are currently on the fast track to mass market acceptance. Forget range anxiety (a fear that gets alleviated once you actually drive and own an electric car [5]), its an entirely artificial worry that will go away once charging stations are more commonplace. The standard for BEVs should soon become 300 mi (~500 km) of range. Few people drive this distance in one sitting. The argument that a fuel cell delivers faster refueling and alleviates range anxiety is therefore quickly losing its strength. Even discarding this argument, one cannot help but notice that fuel cell vehicles are decidedly behind the curve, with only some manufacturers (e.g. Toyota) truly pushing the technology.
Secondly, while modern catalysts and improvements in the technology have increased efficiency and lowered price, a fuel cell is still a very expensive technology. Not only in production, but also in operation. While BEVs currently have high price tags — a situation that will doubtlessly change in 2020 with many car manufacturers adopting the production of BEVs and therefore generating competition in the market — their cost of operation is very low. Fuel cells, as explained above, come with a lot of logistical baggage, which means that the price of fueling up will be relatively high and definitely higher than charging.
All these costs add up quickly. It’s currently expected that the fuel cell stays especially interesting for the premium sector. On the other hand, BEVs are currently emerging from the premium sector and quickly enter the mass consumer market. At the current rate of development, fuel cells will be left in the dust and adoption might be very low unless markets get unduly shifted by governments toward hydrogen economies.
# Conclusion
The fuel cell technology will find its application in places like stationary electricity generation and maybe in the transportation sector (ships, trains, maybe even planes), but just like the 1950s idea of giving every car a nuclear reactor — it seems just like a very bad idea to use it for personal mobility.
Despite much investment, the technology hasn’t reached mass market maturity with infrastructure and efficiency being two of the main unaddressed problems. BEVs are currently leading the market in the alternative drive sector and fuel cells are a distant second with little chance of catching up. In the long run, one shouldn’t bet the farm on manufacturers pushing for fuel cell car adoption.
# Links
[1] Hydrogen production on Wikipedia
https://en.wikipedia.org/wiki/Hydrogen_production
[2] Tankers on ScienceDirect
https://www.sciencedirect.com/science/article/pii/S0960148117308972
[3] Tankers
https://www.sciencedirect.com/topics/engineering/tankers-ships
[4] Twenty-foot equivalent units on Wikipedia
https://de.wikipedia.org/wiki/Twenty-foot_Equivalent_Unit
[5] Range anxiety
https://newsroom.aaa.com/2020/01/aaa-owning-an-electric-vehicle-is-the-cure-for-most-consumer-concerns/
# Description
This article states the author’s views and does not reflect the views of his employer. Comments (constructive) are welcome.
Dr. Burkhard Schwab is a theoretical physicist with a PhD from ETH Zurich and an MASt. from Cambridge University. He has worked in active research positions at Brown University and Harvard University. Currently he works as a business and strategy consultant.
