Flow Battery Research Collective

@danielfp248 said in Alternative Electrolytes:

he Fe3+Gly complex is extremely red while the Fe2+Gly complex is transparent

Got it - I actually didn't know that about Fe-glycine complexes. Thank you for teaching me something new today. Could lend to some cool/inexpensive colorimetric characterization.

The plating makes sense given that Zn plating tends to be much faster than Fe plating. The main other thing I could think of is some invisible amount of oxide depositing over time and fouling the surface (may especially be difficult to see in the felt) since at high Fe2+ concentrations oxides will form (at least thermodynamically) according to its Pourbaix diagram (below). Of course kinetics of that deposition is all affected by complexing with glycine and the like so I'm not certain. Intuitively I would say it could be alleviated by decreasing the pH. I wish I could help more but thank you for being so responsive and detailed. Good luck!

@danielfp248 said in Alternative Electrolytes:

To be honest, I am only interested in systems that use microporous membranes (no ion selective membranes) as selective membranes not only drive the initial cost of systems up a lot but can make them very unreliable, as even slight problems with selectivity can lead to solutions going to waste for non-symmetric electrolytes.

Me too, and that's the route I chose as well (trying to avoid chemistries that required them). Just wanted to ask in case you found an easy/inexpensive solution!

@danielfp248 said in Alternative Electrolytes:

Also note I've seen Fe oxides form at pH values as low as 2.5

Right, that's not too surprising either. Kinetics/rate of formation is difficult to deconvolute from thermodynamic "possibility to form" as it were. Just to clarify, seeing the oxides = orange particulates in the graphite felt?

Some other ideas:

• Does your anolyte yellow over time? I know we're ruling out crossover but that could be another easy way to check - FeCl3 is a deep yellow vs. FeCl2. I'm assuming you've thought of this but I thought I'd suggest it
• You mentioned seeing a lot of undissolved metal on the anode side. Any possibility of dead metal flaking off the felt? In acidic media I wouldn't expect it but who knows. What does the metal on the graphite felt look like usually? any pics?

Thanks for being so responsive. This is interesting

@danielfp248 said in Alternative Electrolytes:

I didn't see any Fe oxides forming

This aligns with what I'd expect - most iron oxides form at pHs above ~ 5 ish. The most stable ones (magnetite is usually the major problem) form in strongly basic conditions, ph >12 ish.

@danielfp248 said in Alternative Electrolytes:

This is 2M FeCl2, 3M ZnCl2, 2M Glycine. At 20mA/cm2. Felt on both sides, daramic membrane. The pH of this is 3.2.

Hmm...your daramic membrane isn't ion-selective, right? Is it just a size exclusion separator? My first thought is ion crossover (or if you already have both electrolytes mixed, just corrosion). Similar to my question in the Fe-Mn post actually! If your separator isn't ion selective (or you already have all active species mixed together), then any Fe2+ and Fe3+ in electrical contact with Zn metal can drive galvanic corrosion because their reaction voltages differ:

2Fe3+  + Zn <-> 2Fe2+ + Zn2+ 

A couple ways you could ad-hoc test this:

• After charging, let the cell rest but track the open circuit potential over time - if the cell voltage drops fairly quickly (over the course of hours), that could suggest corrosion
• If you want to be more quantitative, you could take a look at mixed potential theory - it predicts what the open circuit potential of a system would be if you have two redox couples that are electrically shorted. Here's a nice preprint from a buddy at @quinnale 's lab that explains it nicely (though fair warning, it's still complicated).
• Try having two separate solutions of the FeCl2+Glycine and ZnCl2+Glycine in each half cell. Unless you have an ion selective membrane, you'll still have ion crossover over time, but it will be slower and your capacity loss shouldn't be as significant.

I think this also explains your lower, but very stable coulombic efficiency.

Granted, if that hypothesis is right, you would've seen this problem from the first cycle. So someone else might have a more accurate view of the problem! Do you guys have the equipment for 3 electrode experiments? That could help with diagnostics. I might make a new post with a list of affordable electrochemical equipment like reference electrodes that I've run into.

@danielfp248 I ran into some of your old work on Fe-Mn batteries. I've been interested in Fe-Mn batteries for some time and was wondering if you could share some of your experience.

In particular, if we look at the pourbaix diagram of Fe and Mn overlapped, there's a region in the pH 4-6 range where Mn2+ oxidizes to MnO2 and Fe2+ reduces to Fe on charge.

Let's say we can ignore the fact that there are solid phases - then ion crossover poses a long term concern. Have you guys found any inexpensive or DIY ion selective membrane options (either specific to a particular ion or broad spectrum diy anion/cation exchange membranes?) In my experience screening for inexpensive battery chemistries, crossover of solution phase species is a problem I haven't really seen an easy DIY solution for. Not that I've looked super hard!

Out of curiosity, how are you guys attaching the tubing to the cell stack? are they glued on? hose barbs? Threaded piping?

I ask because leaks are my bane in a lot of my lab setups, even with super tight bolts and threaded fittings. I'm impressed at how well you guys have managed with flexible tubing!

Hey! This is incredible work. I noticed you guys are exploring the shunt current/pressure drop/residence time distribution issue of manifold design. This is well outside my expertise, but I did run into this paper a while back from Kyle Smith, who offered a manifold design (and methodology) to resolve at least the pressure drop/residence time issue. Here's the paper - let me know if you have trouble accessing it and I'm happy to send it along.

They use a tapered header channel with straight diffuser channels and achieve very even flow rate distributions (at Reynolds numbers < 10)

Not sure how well this prevents shunt currents but I hope this helps.

@quinnale said in Towards a working system:

I got excellent prints with high impact polystyrene (HIPS) and have been flowing DI water through the system without leaks. Admittedly, I used a Bambu system for this one while I am trying to get reasonable prints on my Ender3.

Thanks for documenting this! Looks awesome so far. Open question to you and others, especially about alkaline electrolytes. All of the easily printable polymers I've worked with have pretty poor resistance to highly basic solutions. PP has been the best of both worlds but it's hard to say it's easy to print with. Is HIPS working well? Is it easier to print with and still able to handle ~20wt% hydroxide solutions?

@kirk said in Alternative Electrolytes:

I am a big fan of this type of electrolyte engineering! The viscosity can also become an issue with this sort of approach too, no? (You are from now the FBRC viscosity expert ) I see in that paper though they only cycle up to 0.5 M [Fe], which is quite low in terms of energy density (~8 Wh/L if I'm not mistaken). But it seems solid enough to at least test in the development kit as an exercise. Also a huge confidence boost that you were able to reproduce it and make it plate on grafoil!

Haha I certainly don't feel like much of an expert. Viscosity does become an issue, especially with WISE electrolytes. Actually @danielfp248 if you end up trying out the MgCl2 work, you'll notice that the electrolyte is almost oily. The slurries Savinell and Wainright were working with are even more viscous - something like mayo. I don't remember the exact numbers but I'm happy to point you to more papers if you're interested!

@danielfp248 said in Alternative Electrolytes:

I see big increases in ohmic resistance as a function of time, but I don't see any iron oxides forming on the cathode or on the membrane. I don't see full dissolution of the plated metal though, so I bet it has something to do with the oxidation of that metallic deposit (since probably some Fe+Zn alloy is depositing anyway).

Luckily the oxides don't tend to form in acidic conditions (see the Pourbaix diagram - the phase diagram of stable iron species under varying pH and potential). But HER is more prevalent under acidic conditions. I'm interested in the Zn/Fe half cells though - do you notice any galvanic corrosion/passive loss after you charge the cells? Pourbaix diagrams are an excellent tool for designing thermodynamically possible redox couples. Which leads me to...

@kirk said in Alternative Electrolytes:

would love to see it if you're able to share - it's one of my concerns with iodine chemistries at scale (avoiding artisanal iodine production from seaweed).

I'll let you know when I actually churn out the draft! Might be a bit, but in a nutshell, the low cost (redox active materials under ~$3/kWh) and low criticality (generally abundant and decentralized) aren't too surprising: mostly cells using Fe/Mn/Zn and/or air redox couples meet these criteria. Exact redox couples + electrolyte conditions vary but are based on Pourbaix diagrams.

@kirk So my thesis is specifically on slurry/suspension electrodes (instead of using a graphite felt/porous electrode, you suspend conductive carbons with the electrolyte - which also allow you to run solid-phase chemistries in flow) in a way that's sort of chemistry-agnostic. Basically applying chemical reactor design principles to designing slurry electrodes. But here are some salient idiosyncrasies of all-iron cells:

• iron plating in porous electrodes is annoying (acidic = HER, basic = whole host of nasty iron oxides, many of which are quite stable. Plating on carbon substrates is also a pain - most studies I've seen plate onto copper. Also volume expansion is a big pain in static cells. In flow batteries, any time you have plating you end up re-tying power density and energy density through that half cell. Iron plating kinetics are also quite slow, especially in relation to zinc plating.
• Bunch of folks (Savinell and Wainright groups) at case western used slurry electrodes and plated iron onto the slurry particles (ostensibly). They attempted to scale but really struggled with having a performant enough slurry electrode that wasn't too viscous. But plating on suspended particles re-de-couples power and energy density.
• Have you guys looked into water-in-salt electrolytes? They involve dissolving a ton of a supporting electrolyte to the point where they lower the activity of water and suppress HER. I've seen some work using acetate salts and this one using magnesium chloride to support even iron plating - and I've replicated it successfully. The study plates on copper like I mentioned earlier, but I also got it to plate on grafoil.

Solving that performance/viscosity tradeoff in slurry electrodes is part of my thesis! And so is improving the power density of otherwise crappy flow battery chemistries. I'm still working on getting some of it out, but I'd love to share soon or help in any other way.

Material choice from the perspective of mineral centralization and cost is another problem I think regularly about - I did a little study collecting the centralization/governance of various metals and ranked aqueous battery chemistries by cost and "criticality". If that's of interest let me know too!

@kirk @sepi I think Bert's models are free to download off his site, but if you wanted to read the paper, I'm happy to email them to you or you can reach out to him for a copy.

Hey @sepi take a look at Bert Neyhouse's spreadsheet cycling models (https://www.bertrand-neyhouse.com/rfb-models). They might not include stack-scale calculations you're looking for though. If you're interested in more of an economic analysis vs. nitty gritty engineering design let me know.

Hi all, I'm Muntasir, a doctoral student in chemical engineering at Worcester Polytechnic Institute in Massachusetts the USA. I just started perusing your work - I found it on an Autodesk Research lecture you guys did.

I'm a doctoral student in chemical engineering focusing specifically on scaleup and electrode design of flow batteries (the general idea is - if you start from absolutely nothing, can we get a general idea of electrode properties, dimensions, etc. that would lead to specific power/capacity/cost benchmarks, agnostic to chemistry?) Most of my expertise is tying very parsimonious models to practical performance metrics/experimental data, specifically for all-iron and iron-air chemistries. I'm happy to lend some of my experience (if not at least point folks toward useful resources)!