An open-source flow battery kit
dualpower.supply
dualpower.supply
Currently reading through these posts, feel free to ask questions. Great to see interest.
FYI, I am quitting my postdoc job in two months to work on this full-time, which should help the rate of progress, but my main source of income will stop. We hav ea small but it will only cover a few months of full-time work.
If you want to support the project financially we have an Open Collective here: https://opencollective.com/fbrc/donate
We'd really appreciate any support you're able to give, which we'll use to push this open technology as far as we can! We are planning to start work on a much bigger stack after the kit.
Question: What happens to the liquids once they have been used up and depleted? Is there a "recharge" procedure? If they can be reused, how many times before they become disposal waste?
The liquids are reusable, and are charged and discharged repeatedly without needing to replace the fluids. In other words the system is closed with respect to mass, only electrical energy (and minor amounts of thermal) are transferred in and out, reversibly. Flow batteries are similar to so-called reversible or regenerative fuel cells for this reason.
The answer to how long it lasts depends on many factors, and we hope to provide a clearer picture of that in our work.
In a well-designed system, they can last extremely long in comparison to, say, lithium-ion batteries. This is because flow batteries have different degradation pathways that are less severe and, if present, can usually be overcome through other solutions (e.g. electrolyte rebalancing, see ESS's "proton pump").
I was at a tiny house competition, and we were using golf cart batteries, and the winner: The University of Santa Clara, CA:
"The house stores its energy using saltwater batteries, the only batteries in the world to be Cradle to Cradle certified."
They are still busy patent trolling today.
The technology looks great, but they seem annoyingly incompetent at marketing/selling their product...or are just holding out for "whale" customers, refusing to work with anyone except microgrid (ie college campus) and utility scale customers.
So many promising products and technologies die because the inventors/developers hold out for huge customers while ignoring the huge demand from retail/small/medium corp customers.
"We won't talk to anyone except corporations with deep pockets. Once we find a couple of those, we'll be filthy stinking rich!" instead of "if we sell the components at a price that undercuts LiFePO4, we'll have as many customers as we can handle, and there's plenty of margin for distributors and retailers, so we don't have to be B2C."
but that's what happens when you don't vent the hydrogen to the atmosphere
For example by just working with a handful of customers. Or by building a tiny factory that merely prove technical feasibility instead of a bigger one that would prove the business case as well as the technology benefits in terms of cost. Big battery factories are expensive and risky. And most battery tech doesn't really get profitable until you have a big factory and a few years of optimization. The price of the first batteries is generally really high and it can take years to get there.
This has been a problem with battery tech in particular where mostly companies in the US or Europe don't get much funding for new tech and where China actually seems to be doing a bit better. Which is why sodium batteries are shipping in China (CATL) and sort of stuck in R&D for many years elsewhere.
With flow batteries, the technology is kind of proven at this point. They work, they exist, etc. What's left to prove is the price point: can it be done cheaply? Everything seems to suggest yes. But proving that is going to require scaling production and a massive amount of investment to get that going. Think billions, not millions.
Small retail customers are not helpful here because those don't pay until after the factory gets built and starts shipping product. Demand is never the issue. It's doing the leg work to get to the promised price point at which that demand exists. That requires investment.
We are going to start a work coop soon to build and sell the educational kits, as soon as we finish with the initial development. We also have a Hardware X paper coming, showing the initial results of the DIY kit. Things are slowly coming together, but it takes a while since we're literally doing these experiments in our houses.
Looks like a cool project!
They are also have high upfront costs and poor energy density, so there have not been much application outside of grid-scale deployments. Getting something practical for onsite commercial, residential, and vehicular applications have been something aggressively pursued. (Solid-state batteries being another battery tech that is also pursued).
So for someone to make a open-source DIY flow battery that can scale well can change a lot of things.
https://www.wevolver.com/article/what-is-a-flow-battery-a-co...
So, in this case both are a bit underwhelming obviously. But the main point is that you can increase the kwh by simply using bigger tanks and the kw by using multiple cells in parallel or by improving the anode/cathode somehow.
Once we have materials and electrolytes validated with the kit, we plan to move to a much larger cell size which will be part of a flow battery stack, which would actually function as a battery for useful storage.
From the data it appears a battery with 1L of electrolyte provides about 18Wh of energy. Mind you this is at ~1.2V, which isn't especially useful without a boost converter. With a boost converter though you would need a low internal impedance from the battery, which I highly doubt is any good with a paper membrane (from what I understand it already isn't great for flow batteries).
Meanwhile a pair of 18650 lithium ion batteries can be had for $5 and can provide 24Wh at a very usable 7V with no power conditioning or a range of voltages with more than enough ability to source current. And it is a fraction the size, weight, and complexity.
I don't mean to tear apart the project, perhaps there is a key detail I am missing, but I just don't see what this is trying to do outside being a learning experience for students.
This demo cell isn't super interesting on its own but to validate the chemistry it's super helpful. Once you got that done you'd then work on a stack of cells, say 10 or 20 or 40 to get up to normal system voltages.
Once you have that working it's just a matter of making the tank as big as you want for your storage. Provided the initial chemistry is reasonable you could probably use a pair of IBC totes and really go somewhere.
Ie the small electrodes cost something but the big bag of fluid might be cheap.
Say, vitamin- like substances consisting of extremely common elements like hydrogen, oxygen, nitrogen, carbon etc could be used to store energy in a flow battery. Even with quite low performance, they could be very cheap compared to things like cobalt, nickel, manganese or lithium.
Or what about quinones? And sodium, sulphur, sodium are cheap too. There are a lot of very cheap chemistries that could be explored!
Obviously it's a research project not a commercial product. What do you expect?
Back of envelope stuff:
1liter for 18Wh.
1k liter 18KWh (this is an average hot tub).
10k litre for 180Kwh. This is a ~$1000 farming tank.
~100KWh lithium batteries are around the $20-30k. (Used Tesla pack for reference)
Quick google shows flow electrolyte in the neighbourhood of $100 per KWh. Or $10k for a ~100KWh battery.
All this is nothing definitive, but it’s not showing any 10x or 100x differences that would rule out an interesting idea.
18kWh becomes near useless if it can only source enough current to power your TV at any given time. Or to put that another way: 18kWh doesn't do you much good if you can only draw 200W from it at a time.
Given that flow batteries are known for their virtually zero self-discharge, and this project is aiming for a cheap/easy membrane, it seems very likely that internal impedance will kill most use cases here.
Mind you I don't think flow batteries themselves are useless to pursue. It's just that I believe a viable flow battery is almost certainly going to be something that requires complex chemistries and advanced manufacturing. In the same way you can build an open source EV from scratch, but you really wouldn't want to ever take that thing on the street.
This is a very simplified project to prove the concept and provide a test bed for further exploration, not an end-product by any stretch. This seems like the perfect project to test various membranes and electrolyte solutions.
Because you don't want to snap your spine in a minor fender bender.
Scratch built is not the same thing as an EV conversion kit, where all the hard stuff (like a frame and body panels) was already made by commercial manufacturers.
This flow battery is from scratch (well except for the pumps and electronics, but the cell itself is). They are not using off the shelf electrolyte and electrochemical cells like a flow battery kit would.
It's a neat project and would teach a lot, but I just cannot find a scenario in my head where I would want this (even a scaled up version) over another solution.
https://www.solarcarchallenge.org/challenge/about.shtml
https://www.solarcarchallenge.org/challenge/docs/rules2023.p...
You would be ok getting into an accident with one the same as you would get in an accident driving a civic?
Or maybe those are just learning vehicles meant to teach, and not meant to be car replacements? Just like this flow battery project?
Ironically your post validates my whole point: This is a student project at best, and likely worthless as a "democratization of energy storage". The same way those EV's would never be considered "democratization of EV cars".
C'mon...
Our small team is fully qualified to work at any flow battery company. Just give us some time and let us work on it full-time for a bit (which will happen soon). The linked post on top was a blogpost I cranked out in a few minutes one night, not something I ever expected to be on HN.
Nitpick: that seems high, and probably very specific to high capacity Model S packs. A brand new 75kWh Tesla pack for a Model 3 is around $10K installed these days.
How much does it cost to store 10m^3 of water? And hos much does it cost to store the same energy in 18650 batteries?
Also, the internal resistance depends entirely on how many cells you have. But a practical battery wouldn't use paper.
Right, from the one study I can find, commercial flow batteries have about 10-20x the internal resistance of a lithium ion battery, so the match the power and energy capabilities of a single li-ion cell you would need a liter of electrolyte and about 30 (!) cells (3 for voltage x 10 for power).
And that is for a commercial quality flow battery. And lithium ion batteries are wholesale in the $2 a piece range.
I'm not trying to say flow batteries are stupid or dumb, but their use cases are going to be very limited without some huge breakthroughs that will probably dramatically increase the complexity too.
So that's something. Learning how to build one from scratch seems worthwhile, much like learning to build a radio from scratch
On top of that, you don't have to match the internal resistance to match power. If you have plenty of material to absorb the heat, then you can tolerate more percentage points of loss.
In particular, while lithium ion batteries can be built to sacrifice discharge rate for a bit of extra capacity, something like a 3C discharge rate is easy enough to reach. And if your use case is powering a building for several hours, you might only need a .2C discharge rate. That would mean lithium ion as a technology is 15x overqualified, and a flow battery that gives you 10x less power would still be overqualified.
Many elements of being the "best choice" are thresholds. Excess performance doesn't make it better. Price is extremely important, but power density is not so important for most use cases. So if it's even slightly cheaper, expect to see a lot of it.
Christmas present for a battery enthusiast? :D
I think the argument against that is that we've spent tens of thousands of person-hours a millions of dollars more researching, developing, improving, and refining lithium-ion batteries than we have flow-batteries, and the purpose of the project in OP is to make it easier to tinker with flow-batteries.
Think of how much better lithium-ion batteries have gotten since we first started using them. We went from "You simply cannot achieve the energy density to propel (tethered) cars without fossil fuels" to "Oh wait we have electric cars now." in like a decade, and in the decade since them, lithium ion batteries have improved in every metric by at least an order of magnitude. We simply don't know how good flow-batteries can get, because we haven't tried, and it's silly to say "why bother trying, they're not very good right now" when we've just seen how much of a difference it makes to invest into this sort of technology.
I'll take a basement (or garage) with a flow battery over lithium ion ANY day of the week if I want battery backup for my house.
At some point I expect we will look back at the era when we used really flammable electrolytes and laugh about how wild that was. I bet we are not that far from it being just a memory.
I've seen occasional 18650 builds for DIY powerwalls but the vast majority are using prismatic LFP cells. Way easier to wire up, cheaper (unless you have a secret source for surplus 18650s), and minimal fire risk.
The largest use case is going to be grid scale storage, and for that one a bunch of dumb tanks and a bank of cells are far easier to handle and less risky than a bunch of li-ion cells that can go into runaway for whatever reason.
Flow batteries are cool because the storage element is extremely easy to scale. But its not even that great because you also need to scale the amount of cells dramatically to make it useful outside of edge cases. At which point it probably makes more sense to just use another storage mechanism.
Start taking GWh of storage and lithium ion technology gets really expensive and has a lot of associated risks. Flow batteries on the other hand don’t need to worry about a single cell failure resulting in a fire which then spreads.
Even hydrogen is competitive here. IMO, more competitive than that battery chemistry on the article.
I guess there's still aircraft.
While they both need to scale the amount of DC<>AC inverters based on peak power demand. If you want to discharge over 16 hours you’re using 2% of lithium ion’s peak power output and need a huge mess of wiring to move power from each internal cell to that inverter + complex battery packs with individual electronics cooling etc.
Flow batteries on the other hand can use a single pump (+ redundancy) and fat pipe to supply a huge array of ion-exchange membranes which then sit next to the inverters.
Obviously we’d need a real ion exchange membrane and put 40 of them in series, but it looks pretty scalable even in its present form. This looks very practical to me, once a few more years of tinkering is done.
I’d love to have more information about electrode fluid cost, life and reconditioning/reprocessing, as well as power densities for membrane area.
I’d love to be able to add capacity just by adding tanks and electrode fluid! For microgrids like ours, this is a longstanding goal.
A pair of 55 gallon drums equals 7.4kWh, and I'm guessing a lot of us could easily find that much space in our basements. That's enough to power 300W of load 24x7 (a modern fridge is about 60W. 100W will get you really far in terms of LED lighting given that most "60W" bulbs are well under 10W these days.)
One "car battery" sized LiFePO4 battery is about 1400Wh, and costs anywhere from $100 to $500+ depending on the manufacturer/reseller.
I'm a little mystified why they didn't go with a simpler iron-flow design as it is very cheap, and can be nearly completely non-toxic.
1) Its scalable to dishwasher size, ( enough to power a tiny house )
2) If you shot it with a bullet, it would just leak salt water. That is all. Lithium Ion will explode:
Now here is the quiz: If you have a cell phone that is inflating, do you A) Dunk it in water? or B) Toss it in a full document safe? or c) Quickly empty your document safe, and toss it in?
If a flow battery leaks, you can toss in a chicken into the delightful brine.
Since you cannot scale this easily to Utility sized batteries easily, the D.O.E. is not interested. i.e. if you are looking to scale this to a couple of hundred megawatts, just stop reading and thinking about this now. This is NOT mobile. Its not useful for cars or cities. Its right sized for homes.
… if you are looking to scale this to a couple of hundred megawatts, just stop reading and thinking about this now
https://www.pv-magazine.com/2022/09/29/china-connects-worlds...
It doesn't seem like it would take up that much space to have 200 shipping containers sitting somewhere, i'm pretty sure the Home Depot distrubution center in our town already is close to that in their parking lot (yes, you would want them not on wheels, and farther apart)
I only wish they made one that were barrel sized and fit for consumers. Worst case, you have a leak vs a home battery fire.
Battery fires are an EV (and then, only certain models) and laptop/phone thing. Only a few home batteries use the really flammable electrolytes (mostly Tesla Powerwalls, I'd bet). Most people are using LFP for home, which is less expensive and doesn't have the fire problem.
The Chinese company didn't steal this technology. It was given to them — by the U.S. Department of Energy. First in 2017, as part of a sublicense, and later, in 2021, as part of a license transfer."
https://www.npr.org/2022/08/03/1114964240/new-battery-techno...
Not all Lithium Ion chemistries react this way. LFP does not explode:
Perhaps one of those students will figure out how to make a useful large scale flow battery? I have solar, and the missing piece is being able to store electricity for the winter.
Perhaps the person who figures it out learned something from a project like this?
There are much cheaper membranes; ESS for example uses a membrane that is used by lithium ion batteries (I think) and thus is commonly available and very inexpensive.
the blog is listed in the first paragraph of the linked article: https://chemisting.com/2024/03/15/an-open-source-diy-flow-ba...
The Coulomb efficiency of a device with a microporous, non-selective membrane depends fundamentally on how fast you charge/discharge it, as the device self-discharges while it runs. The faster you charge/discharge, the higher the CE will be.
In the case of the photopaper device, it will be in the 85-90% range when charging to high SOC values at 20mA/cm2. The big advantage is that microporous membranes are really cheap and they still work even if dendrites pierce them. I must me clear that photopaper is meant as a DIY demonstration, a commercial unit would never use that but a polyethylene microporous separator - such as Daramic - with these memrbanes the CE and EE tend to be higher.
Perhaps report this issue to your malware provider rather than the site owner?