In the first part of my deep dive into the economics of hydrogen, published last week, I looked at the supply side and distribution – in particular the EU’s $550 billion Hydrogen Strategy, which sits at the heart of its Green Deal and Covid recovery plan.
In broad summary, I concluded that ‘green’ hydrogen (i.e. hydrogen produced via electrolysis using renewable energy) will be cost-competitive with ‘blue’ hydrogen (i.e. zero-carbon hydrogen produced via fossil fuels with carbon capture) in around a decade and competitive with ‘gray’ hydrogen (i.e. hydrogen produced from fossil fuels without carbon capture) at around $1/kg by 2050.
In this part two, I’m going to look at the demand side. What if the world throws a hydrogen party and no one shows up?
We know that in theory green hydrogen could be used throughout industry, transport, power and heating. However, it won’t magically happen in sectors that don’t currently use it, just because it is green. Hydrogen is going to have to win, use-case-by-use-case, but it will not be easy. Not only does it have to beat the incumbent technology, it also has to beat every other zero-carbon option for that use-case. This is where hydrogen hype really meets reality.
As a chemical feedstock, of course, hydrogen is irreplaceable. However, as an energy storage medium, it has only a 50% round-trip efficiency – far worse than batteries. As a source of work, fuel cells, turbines and engines are only 60% efficient – far worse than electric motors – and far more complex. As a source of heat, hydrogen costs four times as much as natural gas. As a way of transporting energy, hydrogen pipelines cost three times as much as power lines, and ships and trucks are even worse.
What this means is that hydrogen’s role in the final energy mix of a future net-zero emissions world will be to do things that cannot be done more simply, cheaply and efficiently by the direct use of clean electricity and batteries. To paraphrase the famous Heineken slogan of the 1970s and 1980s, hydrogen decarbonizes parts of the energy system electricity cannot reach.
That does not mean that hydrogen’s future role will be marginal, however, far from it. First of all, that little phrase “as a chemical feedstock” will be doing an awful lot of work. Secondly, the corollary of hydrogen’s limited use to meet final energy demand is that we are going to be enormously dependent on electricity. An electricity system built on a foundation of ‘base-cost’ wind and solar power may be cheap, but it is going to need two things that hydrogen is uniquely positioned to supply: unlimited flexible capacity for reliable backup, and strategic energy storage for resilience to shocks.
The hydrogen economy is dead. Long live the hydrogen economy!
Replacing gray hydrogen
I want to take you on a tour of hard-to-decarbonize sectors, which I began with my piece in March 2018, Beyond Three Thirds – the Road to Deep Decarbonization. Let’s start with industry, which accounts for around 24% of CO2 emissions globally.
The EU Hydrogen Strategy particularly highlights the use of green hydrogen to “reduce and replace the use of carbon-intensive hydrogen in refineries, the production of ammonia, and for new forms of methanol production”. Production of gray hydrogen for these sectors currently accounts for 6% of EU natural gas demand and 2% of its coal demand.
If green hydrogen can be produced at a price competitive with gray, as targeted by the European Hydrogen Strategy, it should be straightforward for it to take over these existing markets, reducing the EU’s CO2 emissions by 2.2%. There is no way for electricity to act as substitute.
Hydrogen is, however, a relatively minor feedstock to the chemicals industry. Globally, the sector accounts for 14% of crude oil, 8% of natural gas and 2% of coal use. As the bulk use of oil, gas and coal as fuels declines, the cost of extracting and refining them to produce chemicals will inevitably increase – as will the cost of offsetting their emissions.
Assembling chemicals from zero-emissions hydrogen and captured carbon will, as a result, become increasingly competitive. We should expect to see ever-more complex molecules being made this way in ever-larger volumes, driving a rapidly increasing market for hydrogen.
A strong start for hydrogen!
Steel is almost as straightforward as chemical feedstocks, but not quite.
The steel industry is, along with cement, the largest industrial emitter of CO2: in Europe it is responsible for 20% of industrial and 8% of total emissions. Setting aside the 28% of output from recycled steel, the most common way of processing raw iron ore is via the humble blast furnace, using coke produced from metallurgical coal. For net-zero steel, you have to either eliminate the use of coking coal or capture the resulting CO2 emissions.
In August this year, steel-maker SSAB, together with state-owned utility Vattenfall and mining company LKAB, started test operations at the Hybrit pilot plant for fossil-free steel in Lulea, Sweden. SSAB is aiming for the first commercially competitive fossil-free steel in 2026, and to make its entire operations fossil-free by 2045.
In its 2019 analysis of the cost of making fossil-free steel (client link here), BloombergNEF concluded that hydrogen-based steel would become competitive with the most expensive current steel production as soon as it can be made for 2.5 euros per kg, which is any time now. Out-competing the cheapest steel production in the world would require a hydrogen price of 0.6 euros per kg, which it is unlikely even in 2050. A green hydrogen price of $2 per kg by 2030 would require a CO2 price of $125 per metric ton, dropping to $50 per ton by 2050 as hydrogen prices continue to fall.
Hydrogen may not, however, have things all its own way in the steel sector. Another contender is direct reduction using electricity. Boston Metal, which counts oil-major-backed OGCI and Bill Gates-founded Breakthrough Energy Ventures among its investors, is developing a molten oxide electrolysis system that could be used for the production of multiple metals (as well as rare earths), starting with steel, and aiming at a price point even below blast furnaces.
Let’s look next at sectors whose emissions result mainly from their heat demand – for example, glass. In Europe, glass manufacturing accounts for 3.1% of industrial emissions. Float glass plants are wonders of modern engineering: 300 meters long, with temperatures hot enough to melt sand and recycled cullet at one end, and a river of perfectly flat glass emerging at the other. Cold maintenance intervals are measured in decades.
It may seem obvious that the solution is to swap the use of natural gas with hydrogen. But it won’t be easy. The EU Hydrogen Strategy targets a hydrogen price of 1.1 to 2.4 euros by 2030 – which translates into a heat cost of $11.4 to $21.1 per MMBtu. Natural gas in Europe, by contrast, costs $4 per MMBtu.
The European glass industry is covered by the EU-ETS. Assuming hydrogen can reach the low end of the EU’s 2030 target, competing with natural gas at $4 per MMBtu would still require an EU-ETS price of 120 euros ($140) per metric ton of CO2 – over four times today’s price. And if you want to get started using green hydrogen in the European glass industry right now, with the cheapest currently available at 2.5 euros per kg, you would need an EU-ETS price or other support mechanism worth no less than 350 euros per ton.
Looking outside Europe, where natural gas is available at between $2 and $5/MMBtu, hydrogen would need to reach a price of $0.25 to $0.6 per kg in the absence of subsidies or carbon prices, to compete in the market for industrial heat. It will be a long wait.
Green hydrogen versus green electricity
The biggest problem for hydrogen as a provider of industrial heat is not even that it requires a high carbon price to push out natural gas but that it requires a higher carbon price than that required for clean electricity to compete in the heat market.
According to BloombergNEF’s August 2019 Economics of Hydrogen Production from Renewables (client link here), by 2050, green hydrogen may achieve a price of $0.8/kg, dependent on directly connected renewable power being available at $14 to $17/MWh. To compete with $2/MMBtu gas in the heat market, green electricity at those prices would need a $56-per-ton CO2 price. However, the green hydrogen it could produce for $0.8/kg would require a price of $94 per ton to be competitive. In Europe, where natural gas currently sells for $4 per MMBtu, renewable electricity at $17 per MWh would not need a carbon price at all, but the green hydrogen it could produce would still need a CO2 price of $57 per ton. See the problem?
In fact, that’s not even the whole picture. Electricity can, of course, be used to power a heat pump, delivering multiple times its own energy content. Heat pumps are usually associated with low-temperature heat, up to 60 degrees Celsius, but there has been a quiet revolution over the past decade. Industrial heat pumps already work for temperatures up to 160 degrees C with a coefficient of performance of 3, which translates to energy savings of 66%. Add in losses from electrolysis, compression and transport of hydrogen, and you could see an 80% energy saving by using a heat pump.
Research currently under way could push the temperature range of industrial heat pumps to 280 degrees C and beyond. This would allow in theory heat pumps to serve well over a third of industrial heat demand – everything required in pulp, paper, food processing and tobacco, plus pre-heating for higher-temperature processes like glass manufacturing, cement and chemicals.
How realistic is the provision of industrial heat via electricity? A strong candidate for the most important academic paper of the year has just been accepted for publication in Environmental Research Letters. Sylvia Madeddu of the Potsdam Institute and her co-authors looked at 11 industry sectors, making up 92% of Europe’s industrial CO2 emissions. The answer? Some 78% of heat demand could be met electrically using technologies already in existence; the figure increases to 99% if you include technologies that are under development.
It looks like even the cement industry, responsible for 8% of global CO2 emissions, can meet its heat needs directly using renewable electricity. Some 68% of its footprint consist of process emissions, which have to be captured and stored (unless an entirely different chemistry can be used). But the remaining 32% could potentially be eliminated through electrification with clean power, perhaps via a thermal plasma process, with far higher efficiency than via hydrogen.
There are two further reasons why electricity, and not hydrogen, should be the economically preferred source for industrial heat. The first is that last-mile distribution of electricity is much cheaper and safer than last-mile distribution of hydrogen. The second arises from a fascinating corollary of Madeddu’s work: fully one third of industry’s energy use is wasted between ‘final energy’ (e.g. energy input into a boiler) and ‘useful energy’ (e.g. energy output from that boiler).
In just the same way that an induction hob is twice as good at transferring energy into your cooking – not to mention better for air quality – electrical industrial heating can simply be more efficiently targeted and managed than combustion.
There may, of course, be local reasons why hydrogen beats green electricity in industry. In particular, many industries – including non-ferrous metals, ceramics, chemicals and food – use batch processes requiring large amounts of energy in short bursts. These can cause voltage or frequency problems, necessitating upgrading of the power grid. Battery costs are falling, but storing sufficient electrical power locally to meet bursts of demand may prove prohibitively expensive.
Hydrogen could have the advantage here, since it can be delivered by pipeline at very high rates, or stored on site, in compressed gas or liquid form. However, it is by no means a slam dunk. By their nature, batch processes can be used to help balance the grid.
What we are going to see, therefore, is a fight, plant-by-plant, between the value of demand response and the efficiency difference between direct use of green electricity and green hydrogen on the one hand, and the extra investment required in the distribution grid on the other.
It’s a fight that green electricity should generally win. It is cheaper per unit of heat, cheaper to deliver, and better at enabling end-use efficiency. Imagine the potential prize for countries or companies that don’t waste resources on hydrogen, but focus instead on developing electrical solutions for heat in different industrial processes.
Water and space heating
Let’s turn to water and space heating – which together account for 6% of global emissions.
We have already covered the basic economics. Because of efficiency losses between renewable power and green hydrogen, if electricity can be used as a heat source, it should be. Even if green hydrogen is available at BloombergNEF’s lowest 2050 price of $0.8/kg, it would still need a CO2 price of 47 euros ($57) per ton to compete with European gas at $4/MMBtu, and $94 per ton to compete with gas at $2/MMBtu.
That is at the wholesale level. However, where green hydrogen and electricity will be competing in space heating is at the user location, in commercial and residential property. First of all, you need to take into consideration the wholesale-retail mark-up, then sales tax – which in many countries (incredibly foolishly, given the climate crisis) currently favors natural gas – as well as capital cost and operating cost at the customer’s premises.
On all but the latter, heat pumps are the killer app, producing around four time more heat per unit of wind or solar power than could be delivered via a hydrogen boiler or furnace. Basic thermodynamics says that heat pump efficiency falls as the outside temperatures fall – just when you need heat the most – but modern heat pumps continue to work down to minus 22 degrees C, which most parts of the world never reach. Hot water can be stored before a cold snap much more cheaply than storing hydrogen or electricity, and the thermal batteries are seeing rapid innovation. Where there is a risk of very low temperatures, hybrid heat pumps look like an elegant solution – switching some proportion of demand back to gas for just a few very cold days to spare the grid.
Capital cost is a real concern. Upgrading hundreds of millions of homes and offices worldwide to heat pumps – at a global cost of trillions of dollars – is not something that can happen overnight. Ramping up the provision of biogas will reduce the scale of the problem. Increasing combined heat-and-power (CHP) capacity will help. Blending some hydrogen into the natural gas system might buy some time.
Sooner or later, however, you have to switch every building to either clean electric heat pumps or pure green hydrogen. Four times more efficient, using a fraction of the land footprint to generate their power supply, heat pumps should win. And of course they can be reversed in summer to act as air-conditioners – unlike hydrogen boilers, furnaces or fuel cells – great for those climate-change-driven heat waves.
The late Frank Zappa once said that stupidity, rather than hydrogen, was the most common element on the planet. What he did not tell us is that, in the minds of car buffs, the two combine to create the alloy Hopeium, which has an extraordinary ability to absorb public and private money.
In 2003, the EU-backed European Hydrogen and Fuel Cell Technology Platform forecast up to 5 million hydrogen cars on the streets by 2020, replacing 5% of transport fuel with hydrogen. Japan targeted 5 million fuel cell vehicles by 2020. George W Bush said fuel cell cars would be competitive with internal combustion ones by 2010, and would eliminate over 11 million barrels per day of U.S. oil demand by 2040.
Back in the real world, there are still fewer than 20,000 heavily subsidized hydrogen fuel cell (H2FC) vehicles on the roads globally, served by around 400 almost exclusively publicly funded hydrogen filling stations.
There are three commercial hydrogen models on the market: the Toyota Mirai, the Hyundai Nexo and the Honda Clarity. They have no more range than comparable sized battery electric vehicles (BEVs). They are no lighter. They have less luggage space (those pressurized hydrogen tanks). They have half the acceleration and a lower top speed. And they have more moving parts, meaning higher maintenance costs. If none of that has not dissuaded you from buying one, there’s the price: up to 20% higher than an equivalent BEV; if you want a funny-looking two-seater for the same monthly lease cost as a Tesla 3, there’s always the Riversimple Rasa.
When it comes to refueling, most BEV drivers do it at home or work, eliminating regular trips to the gas station. On the odd longer trip, rapid-chargers can add 200 miles of range – about as far as most people want to drive between bathroom breaks – in 20 minutes. Concerns about lack of on-street and motorway charging are like turn-of-the-millennium concerns about internet bandwidth: would video-on-demand ever work? Short answer, yes.
As for those who want to use hydrogen in the form of “drop-in” transport fuels like dimethyl ether (DME) or methanol, their economics are even worse. You are adding production stages and – because they all require a source of zero-emissions carbon – much more cost.
Most carmakers have abandoned their H2FC programs, including Mercedes-Benz, Ford and GM; VW has come down firmly on the side of BEVs (though it keeps a token fuel cell program going, with one Audi SUV in the pipeline for 2023). Even Japanese EV laggards Honda and Toyota – for decades the standard-bearer for H2FC cars – are now fighting to fast-track the launch of dozens of plug-in hybrid and pure electric vehicles.
From a public policy perspective, the real killer for H2FC cars is their wind-to-wheel (or solar-to-wheel) inefficiency. Driving a small family car 100km, whether H2FC or BEV, uses 15kWh of motive energy at the wheels. For the BEV, taking into account losses on the grid and in the battery cycle and drive train, that translates into a need to generate 25kWh at the plant where the electricity is generated. The equivalent for the H2FC car, given losses in electrolysis, compression, transport, storage and reconversion of hydrogen, is at least 50kWh. Put simply, hydrogen cars are half as efficient as BEVs – and there is no reason in physics to think that will change. There is reason why Elon Musk calls them “fool cell” cars.
Undeterred by all this, the Hydrogen Council, a fast-growing association of the largest industrial names promoting hydrogen, still sees one in 12 cars sold in California, Germany, Japan, and South Korea by 2030 powered by hydrogen, cumulative sales of 13 million, and 10,000 fillings stations. In their dreams.
It’s not just cars, either. The underlying physics and economics are no different when it comes to urban buses, delivery vans, commercial vehicles, service and industrial vehicles. In fact, pretty much anything that does not regularly drive over 300 miles without stopping is better as a BEV than as an H2FC. Forklift trucks, which operate 24/7 in an enclosed space, often automated, are so far the only vehicles to have bucked the trend. Weight makes no difference. The fundamental energy economics of an 18-wheeler undertaking 150 miles per day of drayage are the same as a delivery van or a private car driving the same distance. You just need a bigger battery and a higher-powered charger.
Surely, when it comes to longer distances, hydrogen finally comes into its own? South Korea’s Hyundai has delivered its first 10 fuel cell trucks for testing by commercial fleet customers in Switzerland. Kenworth has been testing a fuel cell truck since 2017. Toyota and Hino are collaborating on a fuel cell truck, as are Daimler and Volvo. Hydrogen fuel cell trucks must work, otherwise how did Nikola Motor raise so much money, right?
Not necessarily, says Auke Hoekstra, a researcher at Eindhoven Technical University. First, he found that 90% of 40-ton trucks entering and leaving the Port of Rotterdam drive less than 750km per day. After optimizing the truck’s drag, that means a 1MWh battery – weighing 6 tons based on today’s technology, but on track to decrease to 3 or 4 tons within the decade. What he also found was that replacing the internal combustion drive train with electric motors would save up to 3 tons, meaning that in terms of payload, between BEVs and traditional trucks, it’s almost a wash. As for fueling, as long as trucks have drivers, they need to stop for breaks – 45 minutes per 4.5 hours in the EU.
If you are worried about big rigs spending too long at charging stations, David Cebon, a professor of mechanical engineering at Cambridge University, estimates that a catenary (overhead) charging network for the U.K.’s major road system could be delivered at a cost of just 19.3 billion pounds ($25 billion). That is far cheaper than switching the U.K.’s commercial vehicle fleet to hydrogen and providing the associated fueling infrastructure. Intercontinental road systems could have intermittent sections offering catenary charging – perfect for platooning when we eventually have self-driving trucks.
As it goes for trucks, so it goes for long-distance coaches, as well as for trains. It’s fun to read about Alstom’s hydrogen iLint train, running a 62-mile (100km) route in northern Germany since 2018, or the Hydroflex train, which is currently on trials in the U.K. But, as so often where hydrogen is involved, you have to ask: what problem does it solve? A hydrogen train eliminates the need to electrify the track, but at the cost of locking in forever an option that is more complex, higher-maintenance and less than half the efficiency.
Trains carrying batteries to get them between electrified stretches of track – as developed by Bombardier and Alstom, as well as U.K. startup Vivarail – feel like a cheaper and more elegant solution.
Aviation and shipping
When it comes to aviation and shipping, the prospects look much more promising for hydrogen.
We are seeing a rush of electrified ferries around the world (as predicted in my 2018 piece Planes, Trains and Automobiles – the Electric Remake), but the poor energy density of batteries today makes it difficult to serve routes over 40 nautical miles (75km). Even with a breakthrough in battery chemistry, it’s hard to see the 150 nautical-mile (280km) barrier being breached electrically.
For longer routes, and certainly for ocean-going vessels, zero-carbon shipping means zero-carbon fuel. Biogas might meet some of the demand, if it is not all snapped up for heating. Far more likely, however, will be either hydrogen or a derivative molecule like methanol. Maersk Shipping is looking at a range of alcohols and ammonia. The latter, according to the International Energy Agency and many others, looks like the leading contender.
Then there is aviation. Prior to Covid-19, it consumed around 8 million barrels of jet fuel per day. Biojet is certainly an option, but global biofuels production is stalled at around 2 million barrels per day. Assuming the demand for air travel bounces back after the pandemic, and keeps growing thereafter, a breakthrough is needed.
Electrification looks highly promising for general aviation and short haul, up to around 500 miles; maybe that will push out to 1,000 miles over the coming couple of decades, with solid state battery technology. Beyond that, however, even though the propulsion system will almost certainly be electric, aircraft will be hybrids. The fuel of choice will either hydrogen – Airbus recently revealed a range of hydrogen-powered concepts – or ammonia, or a synthetic liquid fuel made by combining green hydrogen with carbon either captured from the air or produced via biomass.
I have deliberately left the question of hydrogen’s role in the power system until last. Until you get a sense of how competition between hydrogen and electricity is going to play out in the rest of the economy, it is hard to speculate about the architecture of a future net-zero emissions power system.
So far, we have seen that zero-carbon hydrogen will become the mainstay of the chemicals industry: it will replace gray hydrogen, force out fossil-based chemical feedstocks, and form either the fuel or the building blocks for fuels in aviation and shipping. But its direct use on the demand side of the economy will be far from pervasive.
What will be pervasive, by contrast, is electricity. Electricity currently amounts to just over 20% of global final energy. As we move toward net zero, we are likely to see this proportion quadruple. It is this all-encompassing use of electricity that is going to provide green hydrogen with its most substantial opportunity – making sure, quite simply, that the lights stay on in absolutely all circumstances.
As we saw in part one, it is perfectly possible to envisage a power system reaching 80% capacity factor based on reasonably priced renewables plus interconnections, demand response and batteries. You might even get to 90%. The remaining 10% to 20%, however, will be much tougher to deliver. In particular, you have to cover long renewable energy lulls – rainy periods near the equator, windless weeks further North and South – and sudden demand surges – the Beast from the East, polar vortices and the like.
There a number of technologies that can provide energy even during these periods, which we can put in a cost merit order. The cheapest will be demand response – ask energy-intensive users to turn down if necessary. Batteries will be good for a few hours, maybe a few days. We can do a bit more with pumped storage and biomass or biogas. Then there is overcapacity – build your generation fleet to meet peak demand at the worst time of year, as grid designers have done since time immemorial. HVDC will find its place: SunCable and X-Links – designed to bring dispatchable solar power from Australia to Singapore and North Africa to the U.K. respectively – sound outlandish today, but the concept of a motorway would have sounded outlandish in 1920. Of course you can build nuclear plants, but running them only as intermittent backup destroys whatever hope they might have had of being economically viable.
Zero-carbon hydrogen (or ammonia), and natural gas with CCS can provide unlimited amounts of flexible power. Conventional gas generation with CCS makes no sense: as gas is used increasingly as backup for intermittent renewables, the capital cost of fitting it with CCS, only to run it at a low capacity factor, will be prohibitive. Split the process into two, however, and the capital-intensive stage, separating out and reinjecting the CO2, can be run at a very high capacity factor. The second stage, using the hydrogen, can then be intermittent. I am, of course, just describing blue hydrogen.
The extra value of zero-emissions hydrogen – be it green, blue, turquoise or whatever – over and above all the other flexible power options listed above, is that it can be stored in unlimited quantities. Hydrogen is therefore the only solution that can provide deep resilience to the highly electrified net-zero economy of the future. To do so, it will need to be pervasively available: stored in salt caverns, in pressure vessels, as a liquid in insulated tanks, or as ammonia. It will be moved around, cheaply via pipelines, or at a higher cost by ship, train or truck. And it will need to be strategically positioned to cover the risk of supply shocks, whether they be the result of normal weather patterns, extreme weather events and natural disasters, conflict, terrorism or any other cause.
Will the resulting energy system be prohibitively expensive? Assume that 80-90% of power is super-cheap wind and solar at $20 per MWh or less; perhaps it will be $30 per MWh once you have added some storage and interconnections. If the remaining 10-20% of flexible power delivered from net-zero hydrogen and providing 100% network uptime costs $150/MWh, that gives a blended wholesale power price of around $50/MWh. That’s not so far from where most industrialized countries are today – and seems a small price to pay for a high-performing, resilient net-zero economy.
So this is what the future will look like. This is the energy system that politicians need to communicate, policy-makers need to design, and that regulators need to usher into life.
Coda – hydrogen hubs
I want to leave you with one final thought.
As you look at role of clean hydrogen in this future net-zero energy system, you should notice something striking. None of the compelling use cases for hydrogen are widely distributed. No massive demand for hydrogen filling stations, nor hydrogen boilers, no hydrogen-based heat in most industries. The overwhelming bulk of its use will be in the chemicals industry and the power system.
What this means is that we should forget “hydrogen homesteads” – homes and small communities trying to get off-grid using hydrogen – and focus resources instead on “hydrogen hubs”. These are areas where heavy industry, particularly chemicals, fertilizer, refineries and steel come together with shipping, freight transport, pipeline and power infrastructure. This is also where we will see providers of what I call CCS-as-a-service plying their trade – dealing with process emissions from the cement industry and perhaps from the production of blue hydrogen. And maybe – depending on their cost and ability to manage risk – small modular nuclear plants.
HyNet and H2H Saltend in the U.K.; H-Vision Rotterdam and NortH2 in the Netherlands; H2Cluster Norway; Fukushima in Japan; Ulsan in South Korea; Les Hauts de France; San Pedro Bay Ports in California; Pilbara in Western Australia; Hebei, Hunan and Shandong in China; and their like around the world. These will be the real future of the hydrogen economy.
Michael Liebreich is founder and senior contributor to BloombergNEF. He is on the international advisory board of Equinor.