The battery that bursts the hydrogen bubble

Lithium-ion batteries have changed the world. They’ve upended the automobile industry, are upending the power industry, and coupled with solar panels on your roof are fomenting mass exit from the grid for wealthy consumers.

Most of this upending is due to the cost curve below, which shows electric vehicle battery pack costs per kWh. The historical part of the curve – from 2010 to now – represents a triumph of technological innovation and production scale, while the forecast to 2030 suggests that we are near the floor price of Li-ion battery packs beyond which scale and technology cannot deliver dramatic improvements. Even if the forecast is not quite right, the point is that Li-ion costs plummeted in the last decade and will drop further over this decade.

For all that though, Li-ion is a poor choice for the revolution.

Heavy, not particularly energy dense, at risk of catching fire, and morally challenging due to child labour in the supply chain, Li-ion batteries are our ‘just about good enough’ answer to the pressing need to replace fossil fuels in their entirety. They are just about good enough for electric cars. They are just about good enough to store electricity for a few hours for thousands of customers. And they are just about good enough to power your house for a bit if a storm takes out your electricity.

But just about good enough is not good enough and multi $B is being invested to solve these issues. Many ideas are being explored, and even though most of them will never make it out of the R&D labs they were spawned in, anyone with even a passing interest in the topic will see story after story of exciting new developments in their feeds promising cheaper, denser, and safer Li-ion batteries.

And if you are seeing the battery stories, you are likely also seeing stories about another presumed energy revolutionary…

Hydrogen!

It’s a gas. It’s abundant. It combusts. And we can make it and use it without leaving anywhere near the CO2 footprint that fossil fuels do. If you read the headlines, hydrogen is about to replace natural gas for home heating. It’s going to power trucks, and trains, and even planes. It will be stored in underground reservoirs and used to make electricity when the sun isn’t shining and the wind isn’t blowing. It comes in many colours – with black being bad and green being good! – and it can be generated from ‘free’ renewable energy that would otherwise be wasted.

It sounds terrific and this Washington State University graph is typical of energy density information used to promote hydrogen as the energy carrier answer in a decarbonised society. Looking at the graph, Li-ion batteries sit at just about the 0,0 point, which is essentially non-existent compared to most other energy carriers in the graphic.

By contrast, hydrogen looks terrific, way out there on the right with the highest specific energy by a significant measure. One problem with that thought is that the graph has two scales, and the Y-axis represents energy density. The more energy dense the carrier is, the less space it takes to carry the same charge. Hydrogen is a very diffuse gas, so we need to freeze it or compress it to pack sufficient hydrogen atoms into a small enough space for it to be energetically useful. When frozen, liquid hydrogen at -252.8°C is not much warmer than absolute zero and hydrogen at when compressed, the typical 700 bar is twice to three times the pressure of a SCUBA air tank.

The other problem is that the graph suggests hydrogen occurs naturally as a gas. And it does. But out in space, not so much here on Earth. Hold on to that thought, as it is important and forms the issue with hydrogen’s overall energy efficiency, which is covered further down.

To be clear

I’m not a hydrogen proponent. I wrote about that a few years ago and nothing since has changed my mind.

Hydrogen is not a drop-in replacement for natural gas, yet many people are treating it that way. I would not want a hydrogen space heater in my living room, for example, because burning hydrogen with air creates more NOx than natural gas. This is mostly ignored, even by experts. I heard the ‘burning hydrogen produces only water’ fallacy on the Babbage podcast last week and Australia’s own The Science Show podcast as well. Water is the result only if you burn hydrogen with pure oxygen and that is not and will not be common.

The Babbage podcast also blithely declared that that the National Gas grid can transport hydrogen for heating, another fallacy. NREL found – and others have confirmed – that existing natural gas pipelines can accommodate up to about 20% hydrogen mixed in before the risk of leaks increases unacceptably. The reason is that hydrogen is highly reactive so it readily bonds with many materials, causing embrittlement of the pipeline components. We need different grade steel for hydrogen pipelines than most of the current natural gas networks use. We also need different appliances, another often overlooked impact.

And consider that Y-axis again. Like watered down whiskey, if you mix hydrogen with natural gas, you are lowering the energy content of the final product. So, you need to burn more of the combined gas to achieve the same heating value, which undermines the objective of burning less CO2-spewing natural gas.

In terms of user convenience, hydrogen does not come close to liquid fossil fuels such as kerosene, gasoline, or diesel. These are reasonably energy dense and easy to handle. They do not need to be compressed or frozen, and do not so aggressively react with other materials. They are also harder to ignite than hydrogen. That sounds like a good problem, but it isn’t. Hydrogen’s ignition energy is one-tenth that of gasoline as this chart from The Pacific Northwest National Laboratory illustrates. Being too trigger happy when it comes to combusting is more nuisance than benefit and to be fair, while hydrogen ignites easily it tends to burn fast and upwards because it is lighter than air and risk of use is generally overblown[1].

These issues are not insurmountable, but they are inconvenient and addressing them will require infrastructure investment. It does not take much research to find thoughtful analysis on this aspect and the numbers are large – $T large – and probably underestimated.

Energy efficiency

I would advocate that we spend those $T if I thought it would be effective.

However, as I noted above, we do not find much hydrogen as a gas here on Earth, so we need to liberate it. Being chemically reactive, hydrogen is more likely to bind with other elements than remain aloof and it forms strong chemical bonds. Water – H2O – is the most obvious and most cited source of hydrogen. We can use electricity to separate water into hydrogen and oxygen atoms in a process called electrolysis, and a perfectly efficient electrolyser requires 39.4 kWh / kg hydrogen. Of course, real-world electrolysers are not as efficient and typically consume between 45 – 50 kWh / kg.

I have seen this aspect dismissed as immaterial on the assumption we can generate hydrogen using ‘free’ electricity. Depending on the weather, renewable energy generation frequently creates more electricity than there is demand for at the time. It is known as curtailed electricity and is assumed to be freely available for hydrogen production. The premise is contestable[2] and will be less likely as we electrify more sectors of our economies.

There is also a slight wrinkle with electrolysis in that we should use seawater, not potable water, and this requires the additional step of desalinisation. That requires more energy and adds the complexity of bolting-on another industrial process plant to the hydrogen generation facility. The largest desalination plants can generate more than one million cubic meters of fresh water a day, which is considerably more than the largest electrolysers can currently process[3] so desalination will not be a limiting factor[4].

If generation was the only inefficient aspect of hydrogen, I would not be writing this article. But as noted on the graph, to impart useful energy density to hydrogen we need to liquefy it or compress it. Both of these operations use energy, with liquification being more energy intensive.

The lower heating value (LHV) of hydrogen is 33.3 kWh / kg H2, while the minimum theoretical energy to liquefy hydrogen from room temperature and pressure is 3.3 kWh / kg liquid hydrogen (LH2). In practice, liquification can consume almost 30% of the LHV. Compression to 700 bar requires less energy, from 5% to 20% of LHV.

Once hydrogen has been liquified or compressed, storage remains a problem. LH2 needs to be stored in very well insulated containers, or the hydrogen heats up, returns to its gaseous state, and boils away. And hydrogen under pressure needs to be stored in very strong containers. Compare a simple gasoline can at ambient pressure, BBQ gas cylinders pressurised to 9 bar, and hydrogen pressurised to 700 bar and the design parameters of hydrogen tanks is apparent. Toyota’s latest Mirai has carbon-fiber-wrapped, polymer-lined hydrogen fuel tanks that will absorb five times the crash energy of steel to minimise the risk to occupants in the event of an accident. Given the pressure in the tanks, it is a prudent design[5].

We have gotten away with inefficient internal combustion engines and gas space heaters in our leaky living rooms because the energy density of fossil fuels has allowed us to. We dig and pump the stuff out of the ground as an energy rich fuel and while it needs processing, we are starting up the energy curve and coming down. With hydrogen, we’re starting behind the energy curve because we have to liberate those promiscuous hydrogen atoms from their chemical companions and that takes energy, it does not generate energy.

The final leg of this inefficiency tripod is that renewable energy electricity generation is not as energy dense as fossil fuel electricity generation[6]. Wind farms and solar plants occupy a larger land use footprint than fossil fuel power plants (ignoring all the mining to provide the fossil fuels, of course). And they are sadly intermittent, which is why we need storage to cover non-generation periods. Li-ion batteries are being deployed for short-term storage, but that’s minutes and hours, not days and weeks. This is seen as hydrogen’s superpower and it is the reason governments are throwing taxpayer funds at hydrogen projects with what seems to be very little probity.

Indeed, the last time we saw this degree of hydrogen hype was twenty years ago. Economics burst that bubble and from what I can see, economics will burst this one, too.

Hydrogen efficiency example

Fuel cells are how we make use of hydrogen without burning it. A fuel cell converts the chemical energy of hydrogen with an oxidizing agent – usually oxygen – into electricity. That Toyota Mirai I referenced is a Fuel Cell Electric Vehicle (FCEV) and Toyota has been pushing such technology for years, yet sales of their FCEV has been negligible[7].

There are many reasons for this, but primarily it is economic. Hydrogen made from fossil fuels – black hydrogen – is more expensive than the original fossil fuel. Without a seriously tilted playing field by government (taxes, subsidies, mandates, etc.), oil and gas companies have little incentive to price hydrogen lower or innovate ways to generate it. Low demand makes the hydrogen expensive – about US $15 / kg at the pump in California – which makes consumers avoid it, which means there is no market, so vehicle manufacturers do not invest, the price stays high, and FCEV adoption stalls. And green hydrogen struggles to compete because it is generated in such small volumes that it is even more expensive, with no market demand to scale production and reduce cost.

All that aside, my issue with FCEV is efficiency. This graph is representative of the electricity supply chain for three types of passenger vehicles and what it highlights is that you can drive three BEVs the same distance as one FCEV for the same amount of renewable energy[8].

If renewable energy is intermittent and takes more land use than fossil fuels to produce the electricity, then we need to be using it as efficiently as possible. Other hydrogen use cases suffer similarly from poor energy equivalence. For almost every application you can think of, direct use of electricity is considerably more energy efficient than producing, transporting, and using hydrogen.

Now, about that battery

Look at the energy density graph again. At the top of the Y-axis, and sitting close to fossil fuels on the X-axis, is aluminium.

If Li-ion is the just about good enough battery, then a battery made of aluminium is going to be well and truly good enough. I’ve written about aluminium-air before, but that’s more like using gasoline because it is a fuel, not a rechargeable battery. In that article, I noted “we are a long way from building a robust, consumer-ready, Al-ion battery.”

That was late 2018 and I may have been pessimistic because recently published research is showing promising signs that workable Al-ion batteries are years, not decades, away.

Scientists at Dalian University of Technology and the University of Nebraska collaborated to demonstrate a fast-charge battery that acts more like a supercapacitor than the Li-ion rechargeable batteries we’re familiar with[9]. One of the main advantages cited for hydrogen is that a car can be refuelled in minutes, where batteries take hours to recharge. I’ve been bullish on what I call the “5-minute recharge” with Li-ion for a few years, but for an Al-ion design to come out of the gate with such a feature really is a game changer[10].

And Cornell researchers built on this to develop a battery with 99.8% reversibility over cycle times of around five months. That suggests multiple-thousands of charge / discharge cycles with little fade, which would be superior to current Li-ion architectures.

The anode and cathode elements are aluminium and carbon, both of which are inexpensive and environmentally friendlier than lithium. Interestingly, the architecture is a three-dimensional network of graphene as the cathode to promote charge capacity, rather than the planar layout typically used in rechargeable batteries currently. The 3D technique creates a deeper, more consistent layering of aluminium that can be finely controlled and provides better endurance than Li-ion.

Insightfully, Jingxu (Kent) Zheng, the lead author on the paper which describes this research[11], observed, “They also have a very long cycle life. When we calculate the cost of energy storage, we need to amortize it over the overall energy throughput, meaning that the battery is rechargeable, so we can use it many, many times. So, if we have a longer service life, then this cost will be further reduced.”

That really does hit the nail on the head. An energy dense rechargeable battery made from abundant, inexpensive materials that lasts for multi-thousands of charges and which can be quickly recharged kills hydrogen for most of the use cases it is being sold into. Certainly, we still need green hydrogen as it is an important chemical for industrial processes such as ammonia production. But that requirement is a pittance compared to powering the planet.

Conclusion

This recent Al-ion news is merely the tip of a very large battery research iceberg that is seeking the successor to the lithium-ion batteries powering our laptops, mobiles, and battery electric vehicles, along with everything else. Numerous chemical compounds are being explored including sulphur, zinc, and iron, and long-life flow batteries are already in the market that may challenge Li-ion for utility storage.

Still, the only substantial battery change I am expecting before 2025 is solid state Li-ion because introducing new architectures at scale takes time. Many green hydrogen projects will also kick off during this period but without carbon taxes or other substantial government intervention, profitability is questionable. And below the waterline, that battery R&D will be working its way to the surface as companies seek a better battery. I think Al-ion is a hot contender, but it will only take one energy-dense option to displace Li-ion and usher in cheaper, lighter, more environmentally friendly batteries.

And when that happens, the hydrogen bubble bursts. Again.

Two random thoughts

Looking at the energy density graph, you may be wondering why I am not advocating that we use lithium borohydride (LiBH4). Look the stuff up and you’ll find it is spectacularly hard to work with, but you will probably also find a flurry of ‘rock salt’ battery news from eight or so years ago along with some more recent work on solid-state versions. Given LiBH4’s physical characteristics, I put it in my outlier group of likely battery chemistries and will be pleasantly surprised if any company makes a viable contender from it.

On the fun side, I am occasionally challenged on LinkedIn by hydrogen internal combustion engine (H-ICE) proponents. Generally, they claim to have a working engine and are extolling the significant benefits of H-ICE. There is no doubt that you can build H-ICE. But should you build H-ICE is the question. I say ‘no’ (and suggest that so does Toyota and you’d think they’d know) but Jason Fenske from Engineering Explained details the shortcomings of the concept far better than I can over on YouTube. After watching his video again, it’s still a no from me!

Notes

[1] Jacob Leachman from Washington State University’s Hydrogen Properties for Energy Research (HYPER) Laboratory offers an informed view of hydrogen safety.

[2] I contend that as soon as the ‘free’ electricity is found to be useful, the generators will charge for it. And that production facilities like electrolysis plants are more efficient if run 24/7, so operators will build in electricity contracts for 24/7 supply. Free electricity is not free if the lack of it stops your facility!

[3] The world’s largest proton exchange membrane (PEM) electrolyser was commissioned earlier this year and produces up to 8.2 tonnes of hydrogen per day.

[4] The resulting brine may be however, depending on where the plant is situated.

[5] Though I cannot commend Toyota for their ‘How safe is hydrogen?’ marketing, which states, “any leaked hydrogen will rapidly escape safely back into the atmosphere.” It may…but it may not, that depends on the accident!

[6] “But what about nuclear?” is a question that might be asked. I do not include nuclear fission in my definition of renewable energy. Nuclear is low carbon and low land use, but it is not renewable. It is also a diminishing source of electricity that I do not consider will see a resurgence.

[7] It is an interesting thought experiment to wonder what might have happened if Toyota had invested in hydrogen fueling stations the way Tesla invested in their Supercharger network.

[8] As an aside, it seems doubtful that ‘Power to liquid’ fuels can be produced in sufficient quantities to run our societies, even if we wanted them to.

[9] Shen, X., Sun, T., Yang, L. et al. Ultra-fast charging in aluminum-ion batteries: electric double layers on active anode. Nat Commun 12, 820 (2021). https://doi.org/10.1038/s41467-021-21108-4

[10] Al-ion will likely also work better at lower temperatures than Li-ion because the electrolyte has a lower freezing point.

[11] Zheng, J., Bock, D.C., Tang, T. et al. Regulating electrodeposition morphology in high-capacity aluminium and zinc battery anodes using interfacial metal–substrate bonding. Nat Energy (2021). https://doi.org/10.1038/s41560-021-00797-7