In 2017, I started this science YouTube channel. Folks were pretty interested in batteries. I thought that it would be fun to make something accessible – maybe even a do-it-yourself style battery. There’s the classic lemon cell demonstration, but I wanted to make something cheap and rechargeable.

I was inspired by Donald Sadoway’s TED talk about the Ambri liquid metal battery. He started with the highest abundance materials he could find. Then he worked backward from there to build a high performance energy storage system. He ended up making a battery where the positive and negative electrode materials were made of molten metal at about 500 degrees Celsius (900 degrees Fahrenheit). That’s definitely not garage friendly, and it requires special care even in a university laboratory.

So I tried to come up with a battery chemistry made of highly abundant chemicals, that was DIY/garage friendly and with more modest performance requirements. Iron is very abundant. Iron battery electrodes have existed for a hundred years in the Edison cell, which uses iron at the negative electrode and nickel at the positive.

But nickel is relatively expensive and it violates the “use abundant materials” rule. Iron can exist as a metal, ferrous ions or ferric ions (0, 2+ and 3+ ions respectively). I figured it should be possible to make a battery with iron for the positive and negative electrode. Iron is fantastically cheap, not especially toxic, and fairly stable. But that stability means it doesn’t store a lot of energy and it tends to transfer energy quite slowly.

But! Could that work for a stationary backup battery? Answering that question turned into a crowd-funded project in the summer of 2017 where a talented undergraduate student named Nico Yensen put together the first working version of the chemistry. As expected, the performance was… limited. It would take something the size of a car battery to power a decent flashlight. But it proved the principle that you can use iron chemistry at the positive and negative side of the battery and that it could be recharged.

So where are we now?

We’ve been through many iterations since then and I’m happy to announce that it’s much more usable now. My former student, Dr. Dipak Koirala, had a fantastic insight: we need to find mobile, soluble molecules to carry charge into and out of that stable, slow iron. He found two cheap chemicals – ABTS and Viologen – that can do the job at the positive and negative electrode respectively. We call those mediators.

With those added mediators and some optimization of the electrolyte and a commercial separator membrane, we are now in the range of a practical battery. It’s still not as powerful as a commercial cell, but a big battery can deliver all its energy in hours rather than days. So, if you built a household backup battery, you could actually use it to run your house.

The energy density of the new batteries is 11 wH/L. For reference, that would require a 2000 L battery (two of these) to run a typical household for a day. In terms of power, a battery that large could supply a peak of 50 kW, so it would be adequate for a house. That’s an improvement of almost 100x over iron battery 2.0 and more than 1000x better than AIB 1.0. It’s still about 30 times worse than a LFP battery, but it has some advantages for stationary storage (less toxic, non-flammable).

I wrote a blog post many years ago speculating that solar/storage could match a typical baseload power plant (coal or nuclear or hydroelectric) if the price of storage got low enough. Baseload power sells for about 11 cents per kilowatt hour in cheap regions. You can buy electricity from big solar farms for about 2 cents per kilowatt hour. So if your battery can store it and deliver it again for less than 9 cents, you can compete.  Our back-of-the envelope math says that we are getting close.

This is not a total-cost-accounting, just a quick estimate, but it gives you the idea. The active materials and membrane cost $50/kwh storage capacity. If they last 1000 cycles, then you can divide that cost by 1000 to get an idea that the cost of the storage is about 5 cents per kwh delivered. If they only last 250 cycles, then it’s more like 20 cents per kwh delivered. The actual battery looked like it could handle a number of deep cycles between 250-1000 depending on how you test, so we are in the right ballpark for this kind of energy arbitrage.

That leaves out lots of details – how big are the parasitic losses, how much do assembly, maintenance and recycling cost, etc. But this is evidence that solar and batteries are on the right order of magnitude for providing baseload, 24/7 renewable electricity at competitive prices.

I think that’s pretty exciting.

I started this project because I wanted to contribute a little bit to the big global project of sustainable energy. This publication is the end of the line for this project for me. I’ve moved on from the University, Dr. Koirala has graduated. I’m delighted to get to a good stopping point. The technology is ready for others to pick it up and use if they want.

We did our best to document all the details so that anyone can duplicate and expand the work without having to reinvent anything. That’s why I’m so happy to get this published in Hardware X, the open source hardware journal. It’s open access, so no subscriptions are required to read it, and the instructions are as comprehensive as we could make them.

If you’ve followed the iron battery here on YouTube over the years, thank you. It means a lot to me that you were interested and involved. If you supported the 2017 crowdfunding effort, you have my gratitude forever. It was a real honor that people cared enough about this project to share their hard-earned resources. And it meant that a student got to have a full summer of uninterrupted work in the lab, which I know made a difference in his career.

Going forward, this channel is changing. I have updated very infrequently because my responsibilities at the startup where I work have expanded. To be clear, all content of this YouTube channel represents my own effort and does not represent the views or opinion of my employer. When I do have time to update, it will probably be more about my interest in aging biology and longevity research, scientific funding, and maybe even (ugh) the American politics surrounding science and scientific progress.

If it’s time for you to unsubscribe, I understand. No hard feelings.

Thank you again for your time and attention.