Grid-scale batteries: They’re not just lithium

When size and weight don't matter, lots of other battery chemistries can work.

See full article...

 

[David Woodward]

2 other flow battery chemistries.
Allegro Energy microemulsion water based battery is in development, with substantial financial backing.
https://www.allegro.energy/productand Lockheed Martin has it’s GridStar flow battery scheduled for trial with the US Army.
https://www.lockheedmartin.com/en-us/products/gridstar-flow-energy-storage.html

 

[wagnerrp]

Also, batteries have far faster response time than flywheels.

Batteries and flywheels will have exactly the same response time, as it's dictated by the circuitry of the downstream inverter. The amount of time needed to switch between charging, idle, and discharging in order to start replenishing the capacitors of the inverter would likely be measured in hundreds of microseconds.

Or are you talking about flywheels tied directly into the grid just to provide inertia, rather than ones that actually serve as energy storage?

 

[TreeCatKnight]

Hydrogen for cars is extremely stupid. I'm skeptical even for airplanes. But for electricity, it's paradoxically the cheapest path to zero emissions, a path that's currently blocked behind it being incrementally cheaper to pollute for that last few percent of residual unmet demand.

See my prior post near the top of page 2 for a description of the https://model.energy/ simulation that I've been using to model grid mix.

Hydrogen energy storage is currently too expensive to compete with burning natural gas, which explains why there zero of it in our current grid mix, but were we to enact a $50/tCO2 carbon tax, it would be economically efficient to use a mix of hydrogen infrastructure and gas peakers, and were we to demand a zero emissions grid, the solution including hydrogen infrastructure is only +10% more expensive than the lowest-cost polluting solution, versus it being +30% more expensive to sufficiently over-provision solar, wind, and batteries without utilizing any hydrogen storage.

That hydrogen storage can save money may seem paradoxical, given how horribly expensive it is to generate a MWhr of electricity from stored hydrogen that took as many as four MWhrs to produce. And the electrolyzers, compressors, and hydrogen turbines are all capex intensive.

But the alternative to that hydrogen infrastructure, which spends 12% of all capex satisfying the last 3.3% of residual unmet demand (in the model from my prior post), is to spend three times as much building out additional renewables capacity that will be curtailed >95% of the time (in addition to doubling the solution's battery capacity). That overbuilding isn't all that expensive in the big picture, but it's extremely expensive per additional MWhr satisfied. Or alternatively, you could continue using gas turbines, but then pay even more money to sequester the CO2, which is non-economical unless that gets a lot cheaper (turning on the model's CO2 sequestration settings is the same as disabling natural gas). Or a nuclear plant is of a similar cost per MWhr as carbon sequestration, but then you have to pay that higher rate on all of your power, not just the last 3.3%.

I can't tell you when it will make sense to deploy hydrogen infrastructure because that's closer to a political question than to a technical or economic one. As long as CO2 emissions are untaxed, it's cheaper for that last few percent of generation to pollute that to invest in long-term storage. For hydrogen to win in an unsubsidized market, I need to do something aggressive like dividing the model's electrolyzer capex by 4, compressor capex by 2, and making hydrogen turbines free under the assumption that we can burn it in the existing gas peakers. Most gas turbines do support up to 20% hydrogen in their fuel mix, so if electrolyzers really did get a lot cheaper, it's at least conceivable that green hydrogen could outcompete the cost of natural gas and be included in the fuel mix, which even though that's only 20%, it's a huge deal because it allows the hydrogen infrastructure to begin bootstrapping on a "learning curve", and the existence of an affordable large-scale green hydrogen market creates a bigger market for fuel cell development to enter and outcompete the turbines.

I'm glad some model told you it was cheap, but I'm extremely skeptical of your claim. Models often times make assumptions and fill in the blanks that reality can't cope with, which is exactly what it sounds like is going on here (especially since you specifically call for a carbon tax that artificially covers the incredible cost of storing/manufacturing hydrogen.)

In reality, storing the stuff for any period of time is both incredibly wasteful (hydrogen escapes everything that we try to store it in. It is, after all, the smallest molecule.) and incredibly energy intensive. A carbon tax that covers this is both impossible to get past any government and doesn't fix the underlying problem. Not to mention the fact that eventually it will need to expire (and it won't), and then you're left with the same problem as before; big business taking advantage of free government money and entrenched interests making sure it never expires and continuously bleeds money from us normal citizens.

I am however glad you agree that it's unusable for transportation at least!

 

[paulfdietz]

In reality, storing the stuff for any period of time is both incredibly wasteful (hydrogen escapes everything that we try to store it in.

We produce 700 cubic kilometers of hydrogen a year (at STP). But apparently this is impossible, since hydrogen would escape from anything we used to try to contain it.

You are vastly overstating the problem.

 

[The Lurker Beneath]

12 months of energy in a form factor small enough to be attached to a house. I have the feeling that that involves a fairly high energy density. I wonder what technology can handle that density safely in an urban setting.

Hydrocarbon fuels can. Maybe some battery tech can too. Weren't some mentioned that had separated, pumpable electrolytes, so the energy isn't sitting there in contact year round?

 

[The Lurker Beneath]

I wonder how some of these technologies compare with the giant lead acid batteries that local telcos used so that landline phones remained working even during a power outage. For those of you who may remember, 50 years ago when power outages occurred, landline phones always remained working. They had giant vats of acid, each one a cell in an array to achieve the desired voltage. As I remember, those cells lasted for decades and I don’t remember them having a lot of drawbacks other than size. For all I know, maybe these batteries are still in use.

I guess my main point and a point brought up by others is that when space isn’t a problem, you don’t necessarily have do use a lot of expensive, exotic materials and engineering to store grid power for use when renewable sources aren’t at peak production.

I was thinking about this. The lead-acid accumulator is a well-established, mature technology. While the materials are heavy, they are not too expensive and it can go through many cycles of charge and discharge. It can also deliver a lot of power fast, on demand. I'm thinking that the last bit is less necessary for power grids that have a lot of cells rather than the single battery that starts a car. At the time, maybe the telcos chose the tech that was most accessible, and they might choose otherwise now.

Less exotic materials these days mean iron, sulphur, sodium etc. Cheap with few toxicity issues.

 

[wagnerrp]

Direct air capture of CO2 would be too expensive, I think. If you are going to sequester the CO2, you might as well use that as the source of CO2 for manufacture of the carbon-carrying e-fuel. And you wouldn't make methane, you'd make methanol. This requires a bit more carbon per unit of stored energy but can be made with similar efficiency as methane, and is storable in cheap tanks as a liquid.

Why not both, and various other useful industrial feedstocks like ammonia? Methanol and ammonia are both massive industries, with >100Mt of annual production. Meanwhile, RTE for storage using any of them is a terrible ~30%, so if it's worthwhile for any of them, you've got a bunch of options that already have massive production and storage facilities.

 

[paulfdietz]

Why not both, and various other useful industrial feedstocks like ammonia? Methanol and ammonia are both massive industries, with >100Mt of annual production. Meanwhile, RTE for storage using any of them is a terrible ~30%, so if it's worthwhile for any of them, you've got a bunch of options that already have massive production and storage facilities.

Methane, being a gas, is more difficult to store than methanol. Ammonia requires higher pressure and/or lower temperature than methanol, and is also rather toxic to inhalation. Ammonia does have the advantage of being usable in ordinary combustion turbines.

 

[raxx7]

I was thinking about this. The lead-acid accumulator is a well-established, mature technology. While the materials are heavy, they are not too expensive and it can go through many cycles of charge and discharge. It can also deliver a lot of power fast, on demand. I'm thinking that the last bit is less necessary for power grids that have a lot of cells rather than the single battery that starts a car. At the time, maybe the telcos chose the tech that was most accessible, and they might choose otherwise now.

Lead-acid batteries don't actually last very much if you cycle them often.
So they work OK~ish for backups applications where they're rarely called up but not for applications like grid storage where they're being cycled often.
For these applications Li+ ends up being cheaper once you account for the replacement/maintenance costs.

 

[wagnerrp]

I'm glad some model told you it was cheap, but I'm extremely skeptical of your claim. Models often times make assumptions and fill in the blanks that reality can't cope with, which is exactly what it sounds like is going on here (especially since you specifically call for a carbon tax that artificially covers the incredible cost of storing/manufacturing hydrogen.)

We already manufacture and store hydrogen. We use the shit for damn near everything. Just about any synthetic chemical is going to start with hydrogen as a feedstock, currently cracked out of methane. In the future, coal-consuming processes like cement manufacturing and metals smelting will need to switch over to hydrogen. Manufacturing is already a given, though in many of these cases, it will be produced just-in-time on premises, not requiring storage.

In reality, storing the stuff for any period of time is both incredibly wasteful (hydrogen escapes everything that we try to store it in. It is, after all, the smallest molecule.) and incredibly energy intensive.

Storing it at all is costly, especially if your storage processes are diabatic and don't include heat recovery methods. Once you have it in a storable form, actually storing it isn't really a problem. Yes it does leak, but those leak rates are fractions of a percent per day. It's not like you're going to be storing this stuff for months, or even weeks. The bigger problem with storage is embrittlement, and eventual failure of your containment vessels, but realistically that just puts a (fairly long) upper lifetime on your hardware.

A carbon tax that covers this is both impossible to get past any government and doesn't fix the underlying problem. Not to mention the fact that eventually it will need to expire (and it won't), and then you're left with the same problem as before;

Why would a carbon tax ever need to expire? Atmospheric carbon capture will never be profitable, and will always need funding. If you have some process that unavoidably emits carbon, you should need to pay that funding. Forever. The tax ensures you are motivated to find ways to avoid those "unavoidable" consequences.

 

[bjn]

We produce 700 cubic kilometers of hydrogen a year (at STP). But apparently this is impossible, since hydrogen would escape from anything we used to try to contain it.

You are vastly overstating the problem.

And most of that is used almost straight away and turned into ammonia for fertiliser, not stored in tanks or pipe across continents. Sure you can store H2, but it’s a bitch.

 

[beb01]

Direct air capture of CO2 would be too expensive, I think. If you are going to sequester the CO2, you might as well use that as the source of CO2 for manufacture of the carbon-carrying e-fuel. And you wouldn't make methane, you'd make methanol. This requires a bit more carbon per unit of stored energy but can be made with similar efficiency as methane, and is storable in cheap tanks as a liquid.

I'm not a fan of methanol. While it may be a liquid at room temperature, it evaporates readily creating air pollution. And as a liquid it tends to be toxic. A better solution would be to reform methane into propane which liquefies under moderate pressure and is part of a well established economy.

 

[1Zach1]

Yes it does leak, but those leak rates are fractions of a percent per day. It's not like you're going to be storing this stuff for months, or even weeks.

Isn’t that exactly what is being proposed by using salt domes to store hydrogen that was produced during peak summer months, to be used during winter and long duration weather events? Tough to see how else it would work for seasonal storage. As it stands, seasonal storage with hydrogen seems to involve a lot of hand waving, but as noted above that might change with significant carbon taxes to move to net zero, though there still seem to be significant hurdles to overcome just to make it work, let alone the economics of it.

 

[DDopson]

I'm glad some model told you it was cheap, but I'm extremely skeptical of your claim. Models often times make assumptions and fill in the blanks that reality can't cope with, which is exactly what it sounds like is going on here (especially since you specifically call for a carbon tax that artificially covers the incredible cost of storing/manufacturing hydrogen.)

In reality, storing the stuff for any period of time is both incredibly wasteful (hydrogen escapes everything that we try to store it in. It is, after all, the smallest molecule.) and incredibly energy intensive. A carbon tax that covers this is both impossible to get past any government and doesn't fix the underlying problem. Not to mention the fact that eventually it will need to expire (and it won't), and then you're left with the same problem as before; big business taking advantage of free government money and entrenched interests making sure it never expires and continuously bleeds money from us normal citizens.

I am however glad you agree that it's unusable for transportation at least!

Rather than handwaiving vigorously, you should try to quantify your arguments. The conclusions regarding hydrogen storage were shocking to me, not something that I expected before I started playing with the numbers.

You haven't addressed any of the points that I raised. You mention "models make assumptions", but did you see where I tried adjusting the model's hydrogen cost assumptions to increase storage cost by 100X, electrolyzer capex by 10X, turbine capex by 10X, and none of these pessimizations eliminated hydrogen's role in the optimal cost solution for a zero-emissions grid? If you believe my pessimizations to be insufficient, then the burden is on you to point out which assumption needs to be further adjusted, and we can run the simulation with a number that you believe realistic. I've shared all of my work, so you should be able to replicate my results. I modeled the problem using real-world weather data, not just rhetoric.

Hydrogen is difficult to store, that's true, but we have plenty of experience doing so at industrial scale. Are you aware that the global hydrogen market is already >$250-billion per year, over 100,000,000 tons of hydrogen produced each year, almost all of that being grey hydrogen produced from natural gas. It's used extensively as an input to the chemicals industry. For comparison, the 2024 market for Lithium-ion batteries is only about $80-billion. Thus, the upper bound on hydrogen storage cost isn't some wild unknown; it's the cost of present-day solutions already in widespread use. The capex to build aboveground stainless steel tanks capable of retaining a GWhr worth of hydrogen gas is order-of-magnitude smaller than the capex to purchase a GWhr of Lithium-ion battery capacity. Repurposing underground salt domes is three orders of magnitude cheaper. Hydrogen production is expensive and inefficient, but hydrogen storage scales much better than battery storage. Hydrogen thus can't compete with batteries on managing the diurnal cycle, but batteries can't compete on managing wind fluctuations that last for days to weeks. They have very different roles.

What, exactly, are you proposing that we use as an alternative to hydrogen energy storage?

• A) Emitting the CO2 from peaker plants? This is cheaper, and is the default solution until something changes.
• B) CO2 Sequestration? This is more expensive.
• C) Over-building renewables? This is more expensive.
• D) Nuclear? This is more expensive.
• E) Other? Please enlighten the class with your insights.

It's difficult to otherwise address your post, given that it consists of mostly asserting that hydrogen is "wasteful", without clarifying your proposed alternative.

 

[afidel]

We already manufacture and store hydrogen. We use the shit for damn near everything. Just about any synthetic chemical is going to start with hydrogen as a feedstock, currently cracked out of methane. In the future, coal-consuming processes like cement manufacturing and metals smelting will need to switch over to hydrogen. Manufacturing is already a given, though in many of these cases, it will be produced just-in-time on premises, not requiring storage.

I just wanted to mention that Hydrogen will NOT be used for metal smelting. We're currently dealing with fallout from unintentionally introduced trace amounts of hydrogen into steel. There is currently a multi-billion dollar campaign by the federal government to inspect and remediate bridges that used T1 steel. The cause for the program was multiple collapses caused by failed critical joints from hydrogen embrittlement of the welds.

 

[leonwid]

Storing it at all is costly, especially if your storage processes are diabatic and don't include heat recovery methods. Once you have it in a storable form, actually storing it isn't really a problem. Yes it does leak, but those leak rates are fractions of a percent per day. It's not like you're going to be storing this stuff for months, or even weeks. The bigger problem with storage is embrittlement, and eventual failure of your containment vessels, but realistically that just puts a (fairly long) upper lifetime on your hardware.

I had the impression from DDopsons comments that hydrogen would probably be used in shifting power over weeks to months, to deal with a prolonged period of bad weather. You think that would be an issue?

I personally don’t think small losses are much of a problem. Hydrogen is fairly safe to let go in the atmosphere in low concentrations, and the process of using renewables to generate hydrogen to generate power later already has so many losses that this may not be that significant.

 

[THT]

Your bigger problem with scaling down a sulpher-sodium battery is that you need to maintain it at a blistering 300C, which is hotter than most home ovens, and the square-cube law says that the insulation required to do that doesn't scale down very well.

Interesting. Does the entire tank need to be at 300 °C or just the fluid at the anode, cathode, membrane junction?

Yeah, if it is the tank, seems rather uneconomical.

 

[DDopson]

Would it be possible to convert abundant green energy not in hydrogen, but in methane? Use carbon-capture to create the methane, and after burning sequester the CO2. The process is more complex, but methane is a lot easier to handle and more energy dense.

You'd probably enjoy reading about a start-up called Terraforming Industries. Their pitch is that solar cells are getting cheaper very quickly, and the problem with current electrolysis solutions is their capex intensity prevents them from economically running at low duty-cycles, so the best path to cheap hydrogen is to pursue low capex, but less energy efficient electrolysis solutions, deploy them in conjunction with the raw power output from solar panels, and then upconvert the hydrogen to methane, which is easier to store.

They're extremely small / early, having only raised about $40M so far, so they may or may not exist in five years, but either way, their written posts are worth checking out because their founder is clearly a technical person who likes to talk in concrete numbers, which makes their posts much more rewarding to read than your typical information-free glossy website full of stock photos of hand models standing near industrial equipment they've clearly never seen before.

For example, see their whitepaper, or their plan for driving green hydrogen to $1/kg, or there's a bunch of other posts on their blog.

 

[zaco]

2 other flow battery chemistries.
Allegro Energy microemulsion water based battery is in development, with substantial financial backing.
https://www.allegro.energy/productand Lockheed Martin has it’s GridStar flow battery scheduled for trial with the US Army.
https://www.lockheedmartin.com/en-us/products/gridstar-flow-energy-storage.html

Any idea what the actual chemistry of these are? They don’t disclose much more than it being aqueous, and mildly alkaline in the case of GridStar.

 

[DDopson]

Interesting. Does the entire tank need to be at 300 °C or just the fluid at the anode, cathode, membrane junction?

Yeah, if it is the tank, seems rather uneconomical.

If it's not the entire tank, then you have a heat exchange problem, and that's even worse.

 

[leonwid]

You'd probably enjoy reading about a start-up called Terraforming Industries. Their pitch is that solar cells are getting cheaper very quickly, and the problem with current electrolysis solutions is their capex intensity prevents them from economically running at low duty-cycles, so the best path to cheap hydrogen is to pursue low capex, but less energy efficient electrolysis solutions, deploy them in conjunction with the raw power output from solar panels, and then upconvert the hydrogen to methane, which is easier to store.

They're extremely small / early, having only raised about $40M so far, so they may or may not exist in five years, but either way, their written posts are worth checking out because their founder is clearly a technical person who likes to talk in concrete numbers, which makes their posts much more rewarding to read than your typical information-free glossy website full of stock photos of hand models standing near industrial equipment they've clearly never seen before.

For example, see their whitepaper, or their plan for driving green hydrogen to $1/kg, or there's a bunch of other posts on their blog.

It’s a great read, thanks. They certainly don’t lack ambition, or a sense of humor.

 

[wagnerrp]

Isn’t that exactly what is being proposed by using salt domes to store hydrogen that was produced during peak summer months, to be used during winter and long duration weather events? Tough to see how else it would work for seasonal storage. As it stands, seasonal storage with hydrogen seems to involve a lot of hand waving, but as noted above that might change with significant carbon taxes to move to net zero, though there still seem to be significant hurdles to overcome just to make it work, let alone the economics of it.

Because we don't need seasonal storage. We need better transmission. Season storage is a concept designed around a grid that uses fossil fuels that can be readily stored. Baseload power is a concept designed around thermal power plants that cannot be rapidly cycled. Both of these are strategies meant to force new technology to adapt to the reality of the existing grid, to the (often intentional) benefit of those existing energy sources for which the existing grid was designed.

You only need weeks of stored energy if you intent to source all your energy locally. We already don't do that, but we distribute that energy in physical form before it's converted to electrical. Yes, building out terawatts of long distance interconnects is a very costly undertaking, but so was building our existing fuel distribution networks, and so would be hundreds of TWh of storage.

 

[wagnerrp]

What, exactly, are you proposing that we use as an alternative to hydrogen energy storage?

• A) Emitting the CO2 from peaker plants? This is cheaper, and is the default solution until something changes.
• B) CO2 Sequestration? This is more expensive.
• C) Over-building renewables? This is more expensive.
• D) Nuclear? This is more expensive.
• E) Other? Please enlighten the class with your insights.

It's difficult to otherwise address your post, given that it consists of mostly asserting that hydrogen is "wasteful", without clarifying your proposed alternative.

To be clear, the use of hydrogen will already require significant overbuilding of renewables, consequence of the poor RTE.

 

[real mikeb_60]

I'm glad some model told you it was cheap, but I'm extremely skeptical of your claim. Models often times make assumptions and fill in the blanks that reality can't cope with, which is exactly what it sounds like is going on here (especially since you specifically call for a carbon tax that artificially covers the incredible cost of storing/manufacturing hydrogen.)

In reality, storing the stuff for any period of time is both incredibly wasteful (hydrogen escapes everything that we try to store it in. It is, after all, the smallest molecule.) and incredibly energy intensive. A carbon tax that covers this is both impossible to get past any government and doesn't fix the underlying problem. Not to mention the fact that eventually it will need to expire (and it won't), and then you're left with the same problem as before; big business taking advantage of free government money and entrenched interests making sure it never expires and continuously bleeds money from us normal citizens.

I am however glad you agree that it's unusable for transportation at least!

H₂'s not actually unusable for transportation, but there are issues that make it a niche thing. If you want to store it densely, for use in relatively long-distance transportation that needs to refuel fairly quickly for instance, you really need to store it liquefied, which has some issues (cryogenics can be hard). If you want to burn it in a combustion engine, you need a spark, which diesels don't have, and you need a little hydrocarbon fuel mixed with it (10-50% range) to start the engine and sometimes to keep it running; using it in an engine still produces NOx (you're burning something in air), and a little GHG from the hydrocarbon fuel component. Leaves me wondering if it might be better overall to skip the hydrogen separation step and just burn the natural gas (Florida East Coast railroad, for instance, runs LNG-powered locomotives in a 3-car set, with 2 locomotives and a "tender" sandwiched between to carry the LNG). Unless of course you have enough surplus electric power and water to just electrolyze hydrogen on-site (possible in low-use situations like a lightly-used HMU transit line).

Apparently there's concern in the railroad business (recent article in a Trains special issue not yet available online) that fuel cells aren't suitable for high-power line-haul use yet. Not powerful enough to replace a diesel in a size that fits a standard road locomotive, though smaller ones can work in switchers. Battery units with comparable output to diesels exist in road locomotive sizes. Wabtec, for instance, offers one with 7 MWH of Li batteries on board (they're not clear on whether those are LC* or LFP) that operates as a hybrid charging during braking, or can be charged at a depot where it has significant dwell time. Those are horrendously expensive, though, and are only a partial diesel replacement (one or two of the several locomotive units in a consist).

Edit: punctuation is hard...

 

[raxx7]

Interesting. Does the entire tank need to be at 300 °C or just the fluid at the anode, cathode, membrane junction?

I am not a chemist but I think it's the beta-alumina electrode you need to keep at 300 C to have enough ion conduction.
But I think you also need to keep the sodium and sulphur above their melting points (98 and 115 ºC) too.

In practice these high temperature sodium-sulphur batteries are built of cells like these (c) and the entire pack is kept at high temperature.

 

[DDopson]

Lead-acid batteries don't actually last very much if you cycle them often.
So they work OK~ish for backups applications where they're rarely called up but not for applications like grid storage where they're being cycled often.
For these applications Li+ ends up being cheaper once you account for the replacement/maintenance costs.

Lead-acid batteries also don't like being fully discharged, as that significantly shortens their lifespan. Thus, you grandfather's advice to never let your car battery go dead. That's the kiss of death for grid storage applications where you want to be able to sell >90% of the battery's discharge capacity every day. A lead-acid battery wouldn't make it six months in that role.

They're also not particularly cheap. Lead-acid is in the ballpark of $50 to $150 per kWh, but as a mature technology, that number isn't improving.

Lithium-ion prices have fallen to around $100 per kWhr, having decreased by a factor of 8X in the trailing decade, and even faster than that in the previous decade. With a few more years of cost learning, or with chemistries like Sodium-ion, we could easily reach the tipping point where modern battery chemistries are cheaper per kWhr than lead-acid.

For car batteries, cranking current matters at least as much as capacity, and lead-acid is pretty strong at this, but there's now hand-held LiPo jump-starters that can source 600 amps from a tiny package that only weighs a few ounces, which is just mind-boggling (eg, video). Full-sized LiFePO4 batteries marketed as an alternative to lead-acid last longer but are still relatively expensive upfront; if their prices continue to drop, we might see lead-acid batteries displaced even from ICE vehicles.

My guess is that by 2040, lead-acid becomes a niche technology without any remaining mass-market use-cases.

 

[TreeCatKnight]

We produce 700 cubic kilometers of hydrogen a year (at STP). But apparently this is impossible, since hydrogen would escape from anything we used to try to contain it.

You are vastly overstating the problem.

And is that green hydrogen? Because I'm willing to bet it's just cracking methane, which runs counter to the entire point of this discussion.

 

[wagnerrp]

I am not a chemist but I think it's the beta-alumina electrode you need to keep at 300 C to have enough ion conduction.
But I think you also need to keep the sodium and sulphur above their melting points (98 and 115 ºC) too.

My understanding was it was more a physical requirement than a chemical/electrical one. While the sodium melts at 98°C, below 300°C, it's just too "thick", and doesn't make good contact with the electrolyte. That does ultimately result in poor ion conduction, because you can't conduct across a void.

 

[wagnerrp]

And is that green hydrogen? Because I'm willing to bet it's just cracking methane, which runs counter to the entire point of this discussion.

Just because we're currently cracking that hydrogen out of methane doesn't mean the demand isn't there, and won't still be there, to be replaced by electrolysis.

 

[TreeCatKnight]

Rather than handwaiving vigorously, you should try to quantify your arguments. The conclusions regarding hydrogen storage were shocking to me, not something that I expected before I started playing with the numbers.

You haven't addressed any of the points that I raised. You mention "models make assumptions", but did you see where I tried adjusting the model's hydrogen cost assumptions to increase storage cost by 100X, electrolyzer capex by 10X, turbine capex by 10X, and none of these pessimizations eliminated hydrogen's role in the optimal cost solution for a zero-emissions grid? If you believe my pessimizations to be insufficient, then the burden is on you to point out which assumption needs to be further adjusted, and we can run the simulation with a number that you believe realistic. I've shared all of my work, so you should be able to replicate my results. I modeled the problem using real-world weather data, not just rhetoric.

Hydrogen is difficult to store, that's true, but we have plenty of experience doing so at industrial scale. Are you aware that the global hydrogen market is already >$250-billion per year, over 100,000,000 tons of hydrogen produced each year, almost all of that being grey hydrogen produced from natural gas. It's used extensively as an input to the chemicals industry. For comparison, the 2024 market for Lithium-ion batteries is only about $80-billion. Thus, the upper bound on hydrogen storage cost isn't some wild unknown; it's the cost of present-day solutions already in widespread use. The capex to build aboveground stainless steel tanks capable of retaining a GWhr worth of hydrogen gas is order-of-magnitude smaller than the capex to purchase a GWhr of Lithium-ion battery capacity. Repurposing underground salt domes is three orders of magnitude cheaper. Hydrogen production is expensive and inefficient, but hydrogen storage scales much better than battery storage. Hydrogen thus can't compete with batteries on managing the diurnal cycle, but batteries can't compete on managing wind fluctuations that last for days to weeks. They have very different roles.

What, exactly, are you proposing that we use as an alternative to hydrogen energy storage?

• A) Emitting the CO2 from peaker plants? This is cheaper, and is the default solution until something changes.
• B) CO2 Sequestration? This is more expensive.
• C) Over-building renewables? This is more expensive.
• D) Nuclear? This is more expensive.
• E) Other? Please enlighten the class with your insights.

It's difficult to otherwise address your post, given that it consists of mostly asserting that hydrogen is "wasteful", without clarifying your proposed alternative.

If you're going to talk about emitting CO2, then we can discuss how your hypothetical hydrogen is being produced, where it's being produced, and then if it's off site, how it's being transported. I already stated in another reply (probably far too short, both in length and attitude) that this is about green hydrogen, as anything else runs contrary to the entire point of reducing CO2.

Here is a nice breakdown of the cost to produce green hydrogen. It's definitely from the point of view of using it for aviation, but many of the arguments (and math) are useful for this discussion.

The long and short of it is that sure, we can use green hydrogen for many things, but it's incredibly expensive and difficult to do it with the quantities we would need for energy generation (or in the context of the article I linked, aviation. Sure, you can make the argument that it's different enough that there's room for a slice of hydrogen, but it's not as good as you're making it out to be). One of the primary reasons for this is that it takes SO MUCH ENERGY just to generate it (again, not cracking methane), let alone transport it and store it. It's so much cheaper to just build out renewables/storage.

There's a reason why that's what is already being done.

 

[wagnerrp]

If you want to burn it in a combustion engine, you need a spark, which diesels don't have, and you need a little hydrocarbon fuel mixed with it (10-50% range) to start the engine and sometimes to keep it running; using it in an engine still produces NOx (you're burning something in air), and a little GHG from the hydrocarbon fuel component.

Glow plugs are a thing. You don't need a spark, because ignition timing is controlled by the injection of fuel, but cold diesels still need help.

Leaves me wondering if it might be better overall to skip the hydrogen separation step and just burn the natural gas (Florida East Coast railroad, for instance, runs LNG-powered locomotives in a 3-car set, with 2 locomotives and a "tender" sandwiched between to carry the LNG).

If you're doing combustion, and starting with natural gas, then absolutely just burn the natural gas. The only reason use hydrogen is so that you don't foul up a fuel cell, and the only reason to use a hydrogen fuel cell is so that you can continue to use natural gas while claiming to have a pathway to a green future.

 

[TreeCatKnight]

Just because we're currently cracking that hydrogen out of methane doesn't mean the demand isn't there, and won't still be there, to be replaced by electrolysis.

The difference (as I'm sure you're aware, since you seem to be far deeper into the industry than I) is the cost to generate via green methods VS cracking methane.

All I'm saying (and I'm willing to admit to being far too shallow in my original response) is that it's nowhere near as easy/cost effective as the person I replied to made it out to be.

 

[DasSchu]

You overlooked nickel-zinc batteries: https://zincfive.com/

 

[TreeCatKnight]

We already manufacture and store hydrogen. We use the shit for damn near everything. Just about any synthetic chemical is going to start with hydrogen as a feedstock, currently cracked out of methane. In the future, coal-consuming processes like cement manufacturing and metals smelting will need to switch over to hydrogen. Manufacturing is already a given, though in many of these cases, it will be produced just-in-time on premises, not requiring storage.


Storing it at all is costly, especially if your storage processes are diabatic and don't include heat recovery methods. Once you have it in a storable form, actually storing it isn't really a problem. Yes it does leak, but those leak rates are fractions of a percent per day. It's not like you're going to be storing this stuff for months, or even weeks. The bigger problem with storage is embrittlement, and eventual failure of your containment vessels, but realistically that just puts a (fairly long) upper lifetime on your hardware.


Why would a carbon tax ever need to expire? Atmospheric carbon capture will never be profitable, and will always need funding. If you have some process that unavoidably emits carbon, you should need to pay that funding. Forever. The tax ensures you are motivated to find ways to avoid those "unavoidable" consequences.

To your last point, sometime in the far future, we will eventually reach a happy medium on the CO2 content of our atmosphere and will need to adjust accordingly. Not planning for this now, while we have the chance, will result in the same big business interests holding us back from change.

 

[wagnerrp]

The difference (as I'm sure you're aware, since you seem to be far deeper into the industry than I) is the cost to generate via green methods VS cracking methane.

All I'm saying (and I'm willing to admit to being far too shallow in my original response) is that it's nowhere near as easy/cost effective as the person I replied to made it out to be.

The problem is that the industrial hydrogen we're discussing isn't being used as a fuel, but as a chemical reagent. We need the hydrogen because we actually just need the hydrogen. Grey or green, the cost is irrelevant unless someone can come up with replacement products that don't need the hydrogen in their synthesis.

 

[paulfdietz]

And is that green hydrogen? Because I'm willing to bet it's just cracking methane, which runs counter to the entire point of this discussion.

What the fuck? Are you imagining hydrogen molecules somehow behave differently depending on their source?

 

[leonwid]

And is that green hydrogen? Because I'm willing to bet it's just cracking methane, which runs counter to the entire point of this discussion.

Uhm, you expect green hydrogen to be more difficult to store long-term than gray hydrogen?

 

[wagnerrp]

I just wanted to mention that Hydrogen will NOT be used for metal smelting. We're currently dealing with fallout from unintentionally introduced trace amounts of hydrogen into steel. There is currently a multi-billion dollar campaign by the federal government to inspect and remediate bridges that used T1 steel. The cause for the program was multiple collapses caused by failed critical joints from hydrogen embrittlement of the welds.

That's a big bummer. So what's the next alternative? Go closed cycle and recapture and reuse the carbon out of your own flue gas?

 

[THT]

To be fair, the article was talking about grid-scale storage, not residential on a grand scale. Residential (usually with solar) storage currently uses some flavor of Li-based battery; LFP is fairly common in recent installations.

Your last paragraph illustrates "grand scale" for me. I have 3.5 kw of solar on my roof. Perhaps double that might fit if I use all reasonably available space. I probably can spare space in a side yard for 2-3 standard Powerwall modules or something similar - perhaps as much as 30 kwh but closer to 20 is more likely. I would not be able to interconnect with that much of either solar or storage, though, because interconnection is limited to 110% of recent (usually the past 3 years) of usage, and the solar I have now (without storage, but with Net Metering 1.0 that expires in 2030) covers 80-90% most years. V2G would be interesting if I could find a (edit: affordable for me) EV that does it (other than the Leaf); so far, not many, and an expensive option for the few that do. Granted, that's for only the 2 of us in a decently (but not up to current code) insulated small (by current standards, under 2000 sq ft) suburban house with a recent (about 5 years ago; gas heat and "Gold" rated a/c) HVAC system, gas cooking, and gas water heat.

I see numbers like you quoted, though, and see a retail system cost, installed, in the $100Ks in California especially if that includes the EV with V2G (the EV being around $100K all by itself); my simple little system (before tax credit) was about $14K. Your system seems to be scoped for a large rural property designed for off-the-grid operation most of the time. My own utility would refuse to allow that property to interconnect with their system, except under some kind of commercial distributed-generation tariff based on unlimited utility dispatchability of the production. I just don't see a more commonly-sized suburban or urban property needing that much power.

Yes, it is being pitched as a grid battery system. I was just replying to lnlgnem's question about seasonal storage for a house, who said they use about 16.5 MWH per year. Flow batteries looks like it is just finicky for home usage, especially is it needs yearly maintenance service or some other characteristic that is inappropriate for a home.

A typical solar PV, stationary home battery (LFP), and V2G gets you 99.9% of the way there, and it should have the output power characteristics to drive everything you use in your home. Not so sure a flow battery can do that. V2G is a waiting game now. It will become more common eventually, and when you need to get a new vehicle in future, you can get an EV with V2G.

Not sure I understand why your utility would not allow you to connect a solar+PV system. As long as it has a proper switch to stop feeding the grid during an outage, there aren't any grounds for them to say no, other than malice. They may say no to net-metering for a system that generates more than your usage?

 

[DDopson]

H₂'s not actually unusable for transportation, but there are issues that make it a niche thing. If you want to store it densely, for use in relatively long-distance transportation that needs to refuel fairly quickly for instance, you really need to store it liquefied, which has some issues (cryogenics can be hard). If you want to burn it in a combustion engine, you need a spark, which diesels don't have, and you need a little hydrocarbon fuel mixed with it (10-50% range) to start the engine and sometimes to keep it running; using it in an engine still produces NOx (you're burning something in air), and a little GHG from the hydrocarbon fuel component. Leaves me wondering if it might be better overall to skip the hydrogen separation step and just burn the natural gas (Florida East Coast railroad, for instance, runs LNG-powered locomotives in a 3-car set, with 2 locomotives and a "tender" sandwiched between to carry the LNG). Unless of course you have enough surplus electric power and water to just electrolyze hydrogen on-site (possible in low-use situations like a lightly-used HMU transit line).

Apparently there's concern in the railroad business (recent article in a Trains special issue not yet available online) that fuel cells aren't suitable for high-power line-haul use yet. Not powerful enough to replace a diesel in a size that fits a standard road locomotive, though smaller ones can work in switchers. Battery units with comparable output to diesels exist in road locomotive sizes. Wabtec, for instance, offers one with 7 MWH of Li batteries on board (they're not clear on whether those are LC* or LFP) that operates as a hybrid charging during braking, or can be charged at a depot where it has significant dwell time. Those are horrendously expensive, though, and are only a partial diesel replacement (one or two of the several locomotive units in a consist).

Edit: punctuation is hard...

LH2 is only three times the density of gaseous hydrogen stored at 3000 psi, but the energy inputs make LH2 several times more expensive, and at 20 degrees above absolute zero, the insulation requirements are onerous. It's difficult to make a car-scale LH2 tank that doesn't boil itself dry within a week or two. That boiling also means that you have to deal with hydrogen offgassing, and while hydrogen isn't as much of a fire or explosion risk as is popularly imagined -- hydrogen needs 4% concentration to burn, 18% to detonate, and if outdoors, any leaks rapidly rise away from the ground -- you can't just assume that it's safe to release hydrogen gas inside confined parking garages. Even for larger-scale vehicles like trucks and trains, where the square-cube law reduces the insulation challenges, the huge increase in fuel cost is going to be prohibitive.

These problems make LH2 an even harder sell than regular hydrogen.