Almost two years ago I published my first hydrogen decade-by-decade demand projection through 2100. My projection has changed over time, most recently with my assessment of iron and steel demand through 2100, but I hadn’t republished it formally for a while. And now it’s time to do that.
The trigger for this was Michael Liebreich publishing his first update to his useful hydrogen ladder in two years, version 5.0. I recommend readers dig through that as well, as Liebreich’s position on hydrogen has hardened quite substantially, although I remain more bearish on the molecule. There are now eight use cases in the ladder which are in the row of doom where there is no hope for any hydrogen demand.
As a frame for this discussion, it’s important to remember that hydrogen manufacturing today is a greenhouse gas emissions problem, and hence climate change problem, in the same range as all of aviation everywhere in the world, around 1.5 billion tons of greenhouse gases or equivalents. Decarbonizing hydrogen manufacturing is a requirement.
With decarbonizing hydrogen manufacturing comes significant added costs — realistically 3-5 times the cost per ton of hydrogen. Blue hydrogen, manufactured from natural gas with carbon capture and sequestration bolted onto the process, has significant capital and operational cost adders, as well as more energy requirements that typically double the cost or more, and still leaks methane and CO2 along the way, adding to carbon pricing. Green hydrogen has capital and electricity cost curves that mean it’s unlikely to get below US$6 per kilogram at the point of manufacturing. Any distribution of hydrogen as hydrogen is very expensive.
Those costs become a significant factor, and there’s a kicker in any use of hydrogen for energy which I’ll discuss below.
Steel: A Rare Growth Area
Going through each demand category, let’s start with steel. My original projection had roughly 40 million tons of new demand from steel manufacturing. Iron ore has too much oxygen in the mix and must be reduced in some way. That’s done with blast furnaces, direct reduction with synthetic gases, and now with green hydrogen in the HYBRIT process. That latter process uses about 55 kilograms of hydrogen per ton of new steel.
My projection for steel demand growth flattens considerably in the coming decades as China is coming to the end of its massive infrastructure and urban build-out, and other developing regions will be slower. Demand still rises from current levels, but a lot more of steel demand will be met by scrapping existing steel in electric arc furnaces, something the USA has been doing for 70% of demand for 20 years or so. Most of the fossil fuel infrastructure is going to be scrapped in the coming decades, so there will be a lot of scrap steel to play with.
The combination means that it goes from zero tons today to 30 million tons in 2100. My original guesstimate was 40 million tons and a slower growth, but the steel demand projection both reduced the total requirement and brought it forward.
Transportation: Biofuels Require A Little Hydrogen For Processing
Hydrogen for transportation, either directly or via synthetic fuels manufactured from electrolyzed or blue hydrogen combined with captured CO2, is a non-starter. I spent nine years looking at transportation repowering and analyzing biofuels, batteries, hydrogen, and synthetic fuels, and it’s going to be grid-tied and batteries on the ground, and batteries and biofuels in the air and water.
The kicker for hydrogen for energy is related to blue hydrogen. The carbon atom that’s part of the carbon and hydrogen that make up natural gas is 45% of the energy carrier in the mix. Blue hydrogen takes natural gas, removes 45% of the energy with more capital and operating costs for capturing and sequestering the carbon dioxide, and ends up being much more expensive per unit of energy as a result. Not only is the per unit cost of hydrogen doubling, the energy embodied in the resulting hydrogen is about half of what was in the natural gas. Not really a great tradeoff.
In the coming decades, all inland shipping and most short sea shipping will be battery-electric. Bulk shipping will decline radically as bulk fossil fuels represent 40% of volume, and we won’t be extracting, processing, refining, and distributing the roughly 18 billion tons of them that we do today. Similarly, the raw iron ore which represents 15% of bulk shipping will diminish as more processing is done near mines with electricity and hydrogen instead of coal. Only transoceanic and the longer coastal routes will require liquid fuels, and they’ll be biofuels.
Most aviation inside of continents will become battery-electric too, in my projection. At least trans-Atlantic aviation and likely trans-Pacific aviation will become viable as well. My aviation projection has liquid fuels required to at least the end of the century in diminishing amounts after 2060 or so, and they’ll be sustainable aviation biofuels.
The total liquid biofuel requirement is in the range of 200 million tons annually, and we already manufacture 100 million tons of the stuff, with about 70 million tons of it being biodiesel. We waste the vast majority of that on ground transportation, and all of that will electrify, so we’ll have capacity.
Waste biomass has all the energy required to manufacture diesel and jet fuels, which are just different combinations of the hydrogen, oxygen and carbons in the biomass. And there are tens of billions of tons of biomass waste generated annually by the forestry industry, agriculture, livestock, and the food processing industry. Every ton of dried biomass turns into about 0.4 tons of biofuel.
And all of that waste biomass creates a lot of methane as it decomposes, so diverting biomass waste to biofuels is a very climate virtuous circle, and with carbon pricing and carbon border adjustment mechanisms, a very economic sensible one.
The dominant use of hydrogen in refineries is hydrocracking to separate lighter distillates from heavy crude. The lower volume uses are hydrotreating and desulfurization. Hydrotreating is the dominant one required for biofuels. That probably means about 5 kg hydrogen per ton of biofuel, very roughly.
With 200 million tons of biofuels, that means about a million tons of hydrogen for hydrotreating at five kilograms per ton, turns into about a million tons a year of mostly new demand. There are some biofuels processes that add some hydrogen, but with the increased cost of hydrogen, I suspect they’ll prove less competitive than processes which don’t require. I give another million tons for that.
I’d set aside 4 million tons for biofuel processing as a guess in my initial hydrogen projection, and apparently I was generous.
Desulphurization & Hydrotreating: A Plummeting Demand Segment
Oil refineries remain the single biggest demand segment for hydrogen. About 40 million tons, about a third of total global hydrogen demand, is used to process and refine oil. The biggest portion of that, about 75%, is for hydrocracking which separates crude into lighter and heavier fractions. Hydrotreating removes excess water. Desulphurization removes excess sulfur, which is a serious pollutant and a big air pollution and acid rain concern.
The crude oil that generates the biggest demand for hydrogen is heavy, high-sulphur crude. That’s the stuff from Alberta, Venezuela, and Mexico. The requirement for about 7.7 kilograms per barrel of those products means that the quality discount is going to multiply as the price of hydrogen multiplies as it is required to be decarbonized.
Blue or green hydrogen would add US$25 to $77 to the cost of processing and refining a barrel of heavy, sour oil.
We’re in the decade of peak fossil fuel demand. The International Energy Agency, founded almost entirely as a fossil fuel agency 50 years ago, recently declared that all fossil fuel demand, not just oil and coal, would peak this decade. China’s oil and gas refining and distribution giant, Sinopec, declared that peak gasoline demand as this year, 2023, two years ahead of expectations. Asia’s Lantau Group projects that China’s coal demand will peak in 2024, a year ahead of expectations. Etc. Etc. Etc.
With the peak comes the decline. With declining demand comes major shifts in market economics. In a declining market, the cheapest oil will win customers, and that oil is light, low sulphur, and close to water. The first crude oils off the market as a result will be the ones that are getting more expensive to process and refine, which is to say the ones that require the most hydrogen.
This balancing act of global market economics means that my original projection of declining hydrogen demand in refineries wasn’t aggressive enough. As a result, I’ve steepened the curve of decline for hydrogen for those segments.
Fertilizer: Feeding 8 Billion People Doesn’t Require As Much Of It
Fertilizer is multiple products, but the important one for hydrogen is ammonia. That substance is one nitrogen atom and three hydrogen atoms. Plants are made up of up to 5% nitrogen by mass. It’s the single biggest element by mass in their structures.
Getting nitrogen to plants is a huge part of agriculture. We used to leave fields fallow and planted with clover one out of four years to do that. Now we manufacture hydrogen from natural gas and coal and bind it with nitrogen from the air in industrial chemical plants to make ammonia, and then further process the ammonia in various ways for applications.
We use about 30 million tons of hydrogen annually for this process. It’s been a huge part of why we’ve been able to keep producing enough calories to match population growth, multiplying crop yield per hectare.
We waste a lot of the agricultural nutrients we use today. We over-apply them. We spray them from helicopters and planes over fields, blowing a lot of it outside of the boundaries of the fields and into waterways where it causes significant downstream problems.
And when ammonia comes in contact with moisture in the soil, some of it turns into nitrous oxides with global warming potentials about 265 times as high as carbon dioxide.
Fertilizer is agriculture’s big climate change problem, and that problem is going to be priced. Canada’s carbon price already includes all greenhouse gases, including nitrous oxide and methane. The EU’s emissions trading scheme and carbon border adjustment mechanism will include them starting in 2026. The projected price per ton of CO2 or equivalent is adequately high for the severity of the problem, reaching about US$203 in 2030 in 2022 dollar values.
That means that ammonia is going to get more expensive coming and going. The roughly six tons of carbon debt per ton of ammonia for manufacturing will be priced, or the expense of blue or green hydrogen will be added. The nitrous oxide emissions of use will be priced.
The combination will mean significant downward pressure on ammonia demand and a hunt for alternatives. And we already have levers that we can pull there. Pivot Bio has genetically engineered normal nitrogen fixing microbes in the soil to remove their off switch so that they keep working even in the presence of fertilizer, and brew them in what are basically beer vats. They were reducing ammonia fertilizer demand on a million acres of US corn by 25% a couple of years ago, and have a stretch goal of 100% for corn, wheat and rice by 2030. Many others are working on the same problem, so good results will be achieved. And precision agriculture, including with heavy spray drones, is cutting product application volumes for the same crop yields.
Criticism of my first projection from Paul Martin was apt. I had reduced demand in 2030 by a couple of million tons, and that was too aggressive of a timeframe. Other than that, the projection remains unchanged with a reduction but not elimination of the demand area.
Heating: Zero Hydrogen Demand
There are now something like 45 serious studies which all say that hydrogen has no play in commercial or residential water or air heating and that heat pumps are the answer. 45% of industrial heat is below 200° Celsius, and modern heat pumps are delivering that, so it’s not a market either.
That leaves industrial heat above 200° Celsius. I’ve looked at innumerable uses cases and technologies for industrial heating and haven’t found a single one where there aren’t electric solutions. There is no temperature range that electricity and its heat technologies like resistance heating, induction heating, EMF, microwaves, electric arcs, and electric gas plasmas can’t achieve. Electric arc furnaces already run up to 3,000° Celsius.
The only reason we don’t electrify higher temperature heating today is because fossil fuels are a dirt cheap source of high temperatures, especially natural gas for the past 20 or so years. But in 2030, the EU’s ETS and CBAM mean that every gigajoule of natural gas heat is going to cost about US$14 more. Natural gas averaged about US$5 per gigajoule from 2009 to 2021.
A realistic carbon price, one that the EU is applying to all goods crossing itss border, all goods manufactured inside its borders, and one which both US EPA and Canadian social cost of carbon is closely aligned with, means that the price of heat from natural gas will be multiplied by a factor of four in 2030, and a factor of six in 2040. Canada’s carbon price, California’s cap and trade price, and China’s cap and trade price are going to trend to the same points, in part because that’s the economic reality, and in part to align trading value across economies to avoid being uncompetitive.
The carbon pricing alone would mean massive shifts to electric heating solutions, which are a lot cheaper and more efficient in most jurisdictions. But hydrogen energy will always be a lot more expensive than natural gas energy.
Remember, blue hydrogen uses energy to throw away 45% of the energy in the natural gas and costs more. Green hydrogen is going to be five to ten times the cost per gigajoule of liquid natural gas, the most expensive form of energy most economies use today.
Economics mean that every attempt to use hydrogen for heat will be uncompetitive with just using electricity directly. This is no place where hydrogen will pencil out in my opinion, or if there is it’s such a rounding error that I’m comfortable ignoring it.
Long-Term Storage: Nothing About The Requirement Screams Hydrogen
One of my projections is the requirement and technologies which will be used for grid storage. My projection has pumped hydro — specifically closed-loop, off-river pumped hydro with high head heights and small reservoirs — continuing to dominate grid storage as it has since 1907 when the first site went live. After that, redox flow batteries and cell-based batteries battle for second place, with redox flow taking the lead because of the decoupling of power and energy that they provide. Then there’s a 100 GW of also-rans, including compressed air, liquid air, thermal storage for electricity, and hydrogen will fight over scraps.
That takes care of everything except 10- to 100-year events when very large geographic regions go weeks without sunshine or wind. Ten years is the island problem, and that’s what Sir Chris Llewellyn-Smith’s modeling shows for the UK. On continents with more room and more connected grids, it’s longer than that. In a lot of the tropics, there’s always sunshine, so it’s a non-issue.
Those rare occurrences will need some strategic reserves. A lot of people seem to think that putting hydrogen in salt caverns sluiced out for the purpose is the solution. I think that’s presupposing that hydrogen is the answer and going looking for use cases.
There’s nothing about hydrogen that makes it the obvious choice for very long duration storage. Its energy density by volume is still the poorest among the alternatives. The cost of manufacturing it from energy or fossil fuels is still higher than the alternatives. It’s still the leakiest molecule of the alternatives, and one of the ones that’s hardest on metals and electronics.
Personally, my bet is on capturing and diverting anthropogenic biomethane from the waste biomass we don’t eliminate or otherwise diffuse, and don’t use for other purposes, and shoving that it into strategic reserves to burn in bog standard combined cycle gas turbines. That turns a current major climate problem into a solution for a specific transition problem and containing and monitoring methane is relatively trivial compared to hydrogen.
I’m also not particularly concerned about this segment and think arguing about it a lot is a distraction. We have an awful lot of heavy lifting to do before we get to the last couple of percent of decarbonization. It’s great that there are people thinking about it, but it doesn’t make a lick of difference to what we have to do in the next 10 to 15 years.
The combination means that I think some hydrogen will be used in grid storage, but only about a million tons a year starting in 2040. Nothing I’ve seen in the past two years has changed my mind on that one.
Hydrogenation: Margarine & Trans-Fats Will Decline A Bit
We use a lot of hydrogen to make unhealthy food products right now. Partial hydrogenation makes trans fats, pretty much the least healthy thing we can eat that isn’t actively a poison. With significant increases in hydrogen costs and growing health consciousness globally, I see a slight decline by 2100 in hydrogen demand for food hydrogenation.
Unchanged since the first projection, going from about 8 million tons a year now to 5 million tons in 2100.
Methanol: Despite Industry Bait & Switch, A Small Decline
My initial projection left demand unchanged at about nine million tons of hydrogen a year for the roughly 170 million tons of methanol in the global market. I didn’t see a reason for that to go up or down, but I’ve spent a lot of time looking at methanol in the past two years.
The methanol industry is a major climate problem. Methanol manufacturing, with the hydrogen challenges that feed it, adds more greenhouse gas emissions and ends up in the range of 500 to 700 million tons of greenhouse gases or equivalent a year. That’s a very big number, about 1.5% of global greenhouse gas emissions.
Methanol is a climate change problem which depends on hydrogen. Methanol is manufactured from natural gas, coal, or some waste synthetic gases. It depends on its feedstocks being dirt cheap, and being able to treat the atmosphere as an open sewer to maintain its commodity prices.
It’s used for a variety of industrial purposes, including a bunch of stuff that is going away, like being a gasoline additive, a diesel additive, and a burnable fuel call DME. It’s used to make formaldehyde, plastics, and solvents.
Cleaning up methanol is a major requirement, and carbon pricing alone is going to make its costs shoot up. As its price to end consumers goes way up, doubling, tripling, or more, substitutes will be found in many cases. There are other solvents. There are other chemicals which can do the same job. Methanol gets used because it’s useful and cheap. When it stops being cheap, other things which aren’t as expensive will end up in the mix.
This means that the natural pathway for methanol is actually a decline in demand, something I hadn’t projected. That’s what my current model shows, with a decline from nine to seven million tons by 2100. Not a lot, but certainly not a major increase.
I’ve been paying attention to methanol because the industry is trying really hard to convince the maritime shipping industry that it’s the replacement for the fossil fuels that they currently bunker. As I’ve noted, they are running a bait and switch operation, selling black methanol today at likely discounted prices and promising green methanol in the future at high prices. Maersk has bought into the story and is spending tens of millions on dual-fuel ships and low-carbon methanol contracts that won’t decarbonize current methanol and likely won’t be used much at all.
Methanol is already 1.6 times the cost of diesel per unit of energy on average around the world. When that cost multiplies as methanol decarbonizes, biofuels and batteries will be vastly more competitive. I think Maersk’s dual fuel ships will bunker a lot more biodiesel than methanol.
I don’t see an upside for methanol in maritime shipping. The economics don’t pencil out. But as I’ve said, it’s the best of the also-rans, so if the industry and governments and the IMAO for some head-scratching reason force the square peg of green methanol into the round hole of marine power, I will shrug and move on. At least maritime shipping will be decarbonized, if unnecessarily expensive.
But the 1,000 km regular route of the 700-container ship that’s now plying the Yangtze with containerized batteries and its sibling tell me I’m more likely right.
Mixed Other: Hydrogen Is Used For A Lot Of Stuff
One of the most common metaphors for hydrogen is that it’s a Swiss Army knife. It can do a lot of things, but it can’t do most of them particularly well. When it’s cheap and the atmosphere can be used as an open sewer without cost, it’s frequently the ‘best’ economic choice for a bunch of things. It’s cheaper than alternatives.
One of the interesting use cases I found in the past couple of years is in big electricity generating turbines in dams, coal plants, and nuclear plants, where it’s used because its properties make it a good coolant that doesn’t add friction. Gigawatt-scale nuclear plants like the Nine Mile Point reactors use a few hundred kilograms of the stuff a day. Those big turbines aren’t increasing in number, but decreasing in number and usage as coal goes away, of course. And the world isn’t going to be building thousands of new nuclear plants when wind and solar are so cheap, fast, low-risk, and reliable.
When the price of hydrogen to end consumers goes up significantly for all of these myriad incidental use cases, they’ll mostly look around and see if there are cheaper alternatives. And they’ll find them in many cases. Hydrogen isn’t magic or irreplaceable in a lot of processes, it was just the economically correct choice at its black hydrogen price point.
Originally, I’d left this demand segment alone, leaving it flat at 30 million tons a year. But the same economic pressure to find replacements that impact fertilizer and methanol impacts this segment. I now have it slowly declining to 25 million tons a year.
Hydrogen Demand: Off A Third By 2100
The result of all these puts and takes is a lower demand in the end than I’d originally projected in 2100. My original projection was around 90 million tons, and now it’s slightly under 80 million tons.
Am I right? Of course not. This is a logical scenario based on empirical data and major global trends. It’s defensible, but it’s going to be wrong in multiple ways. The error bars are big, and at 77 years, I’m projecting about 76 years past what we are actually capable of doing with any accuracy.
Am I less wrong than most projections? I think so, but I would, wouldn’t I. Time will tell.
But there are some emerging data points that make it clear that the hype cycle for hydrogen for energy is disappearing. One of them is Michael Liebreich’s ladder, where use cases he had in play have dropped completely out of play, like off-road vehicle energy demand. Another one is a major BCG report that recently dropped where the illusory consensus of US$3.20 per kilogram of green hydrogen was found to be wrong, and that US$5.30 green hydrogen was more likely.
It’s not like a lot of people haven’t been saying that for years, including me. Hydrogen can be green, but it won’t be cheap. My projection is that hydrogen will come in at $6-8 per kilogram to manufacture at point of consumption, and more when delivered of course. The consensus of $3.20 or even the fatuous $1 per kilogram was among STEM-illiterate financial and policy types and boosters of hydrogen for energy who in many cases knew better and were lying intentionally.
And so, version whatever this is of my hydrogen demand scenario through 2100 is now published. Please tear it apart and challenge me. Tell me where else I’m wrong, or how these changes are wrong. I live by post-publication expert review, with all of the slings and arrows of occasional humiliation that brings.
As has been attributed to Keynes and many others, “When the facts change, I change my mind. What do you do, sir?”
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
CleanTechnica Holiday Wish Book
Our Latest EVObsession Video
CleanTechnica uses affiliate links. See our policy here.