My adventures building a Zinc-Bromine battery

svetz said:

The voltage drop above reminds me more of the behavior of a capacitor than a cell. For example with LiFePO4 there's a nominal voltage, similar to how the temperature of water holds steady at 0° C as it freezes. I take it that at 1 mA what you're really measuring is the ion flow through the electrolyte?

Click to expand...

I measured the capacitance of the cells using cells without ZnBr2. Their capacity due to capacitance is just around 1uAh, which is reasonable given the electrode separation and dielectric constant of the solution.The charge extracted here is more than 30x that, so it is due to a chemical process. When I take charged cells apart I can see the plated Zn and deposited TBABr3 so the chemistry is definitely happening.

Some batteries have strong potential plateaus while others have more linear downslopping curves. This depends on how well the battery can sustain the chemical reactions through the discharge phase at the current density demanded by the discharge process.

The CC6 carbon electrode is bad at this, possibly because of a much lower surface area and conductivity, but the process is no doubt based on the chemistry of the battery.

The discharge at 1mA is not limited by ion flow through the battery but most likely by the kinetics of TBABr3 reduction in the CC6 electrode. Self-discharge is likely to also be an important factor limiting efficiency at these charge/discharge currents.

Best results for CC4

Coulombic efficiency = 69.96%
Energy efficiency = 43.43%
Current = 1mA
Charged to 500 uAh

Zinc anode (0.2 mm)
Fiber glass separator (8 layers)
Electrolyte (0.5M ZnBr2 + 0.2M TBABr)
0.5 inch diameter
Measured in a Swagelok cell with graphite electrodes

These are the CC4 electrode results at 10mA. The fact that the Coulombic efficiency and Energy efficiency are better at higher current is a direct indication that there are some serious self-discharge problems with this cell.

This is because both processes charge/discharge the same amount of charge but the fact that the low current process takes longer means there is more time for ion diffusion related discharge processes. Most likely we have very poor formation of the TBABr3 and what we're forming is mostly elemental bromine within the surface of the electrode, which can migrate and self-discharge the battery by reacting with the Zn deposited on the anode.

Coulombic efficiency = 80%
Energy efficiency = 56%

The original 1885 patent for those interested.
And a later one

NMNeil said:

The original 1885 patent for those interested.

https://patentimages.storage.googleapis.com/4f/d7/44/2c780dc892ebed/US312802.pdf

And a later one

https://patentimages.storage.googleapis.com/d0/9b/6b/8beb5ec70ec0b5/US5591538.pdf

Click to expand...

Thanks for posting these! It is worth noting that the reason these Zn-Br batteries never saw commercial use was because of the self-discharge experienced when Br2 migrates across the cells. This is a show-stopper for these batteries as they lose 20%+ of their capacity per day under ideal conditions.

The Princeton publications on "minimal architecture" Zn-Br batteries, points to this problem not being a huge issue if you're interested in commercial scale deployments and have "energy to waste" - only caring about relatively short term storage - but even then, it's hardly easy to sell a battery with strong self-discharge problems, even if the cost per watt is low.

So I opened the CC4 cell when it was fully charged and did see a lot of TBABr3 formation, so it is certainly working, however, since the concentration of TBABr is substantially lower than that of ZnBr2, it might not be able to sequester all the Br2 that is being formed (see my last post - https://chemisting.com/2020/09/15/zinc-bromine-batteries-is-tbab-the-best-complexing-agent/ - as to why I don't prepare more concentrated TBABr solutions).

I decided to try a small experiment and just put 20uL of 1M TBABr solution right on top of the CC4 cathode and reseal the cell and remeasure it. The cell was discharged and then put through some charge/discharge cycles at 1mA to 500 uAh.

Coulombic efficiency > 99%
Energy efficiency = 62%




The internal resistance increased a lot though - probably due to the precipitation of TBABr within the cell - but at least the self-discharge problem seems to have been largely solved. The more experiments, the more I realize TBABr might not be the best sequestering agent to use within these cells because of its solubility limit in the presence of ZnBr2.

I think I will work more on the chemistry before using more of my carbon materials, which I have a small supply of, because without the chemistry properly working, it is not smart to optimize the cathode choice. There are two main roads to try:

1. Increase the solubility of the TBABr in the ZnBr2 solution. This can be done by reducing the polarity of the solvent with additives or reducing the polarity of the Zn ion with the use of Zn complexes. However this can change the solubility of the TBABr3, which we would want to avoid at all costs.
2. Change the complexing agent to something more soluble. This involves changing TBABr to something more soluble in ZnBr2 solution that still forms an insoluble perbromide. The main candidates would be TPABr (the one used in the chinese paper) or TMPhABr (which I think is going to be even more soluble).

I am going to think about the experimental design to try 1, since trying 2 is going to be significantly more expensive.

If it were setup as a flow battery, wouldn't that eliminate the self discharge issue?

svetz said:

If it were setup as a flow battery, wouldn't that eliminate the self discharge issue?

Click to expand...

To some extent, but not entirely. It eliminates the self-discharge issue in the sense that you can move the bromine away to store it somewhere - so the self-discharge across long periods of time is eliminated - but when the battery is charged/discharged you can still have bromine diffusion across the flow battery membrane because the zinc-bromine flow battery still has a solid zinc anode where zinc metal is plated that is separated from a bromine solution by a membrane. Bromine can go across that and reduce the efficiency of the battery. This is why flow battery systems also use sequestering agents but in this case they want the sequestering agent to form an immiscible liquid (like tetramethylammoniumbromine does) rather than a solid (like TBABr does).

So your self-discharge on storage is eliminated but your energy efficiency loses due to self-discharge during the actual charging/discharging process are not. The complexity, weight and cost of the system are also increased substantially. A big disadvantage of Zn-Br flow batteries vs Vanadium ones is that your catholyte capacity is determined by a tank (where the bromine is stored) but your anolyte is still going to be limited by how much zinc you can plate onto the actual anode while in Vanadium flow batteries both the catholyte and anolyte are stored in tanks and your capacity is determined by the size of these tanks, not the dimensions of the flow cell electrodes (which determine your power output instead).

If we can find a solution for Zn-Br that requires no flow battery setup, then that will be much cheaper and much more energy dense.

Sandia Labs have already done the hard work.

Zinc bromine flow batteries require a fair bit of maintenance too. It can all be automated with a BMS but with the cycling that is needed you really need two batteries both rated to do the job alone so that when one is doing a maintenance cycle the other is still there to carry your loads.

NMNeil said:

Sandia Labs have already done the hard work.

https://www.sandia.gov/ess-ssl/publications/SAND2000-0893.pdf

Click to expand...

This thread is not about flow-batteries though! It's about static zinc-bromine systems.

gnubie said:

Zinc bromine flow batteries require a fair bit of maintenance too. It can all be automated with a BMS but with the cycling that is needed you really need two batteries both rated to do the job alone so that when one is doing a maintenance cycle the other is still there to carry your loads.

Click to expand...

This is part of the reason why, in practice, zinc-bromine flow batteries ended up being really complicated, hard to maintain and offer low specific energy and power values. A static system would get rid of most of these problems, provided the self-discharge issues can be solved effectively.

danielfp248 said:

This thread is not about flow-batteries though! It's about static zinc-bromine systems.

Click to expand...

Understood. But the link was for those who inquired about the self discharge rate in zinc bromine flow batteries, and if it was the same for non flow batteries.

Wrote a post about the current problem with TBABr and the three main experimental ideas I want to try to increase its solubility in concentrated ZnBr2 solutions (https://chemisting.com/2020/09/17/z...-how-can-we-increase-the-solubility-of-tbabr/)

Just throwing this out as a 4th way (sorry if it's crazy-talk), wouldn't the Zn+2 cation be attracted to something with a strong negative charge? Would it be possible to create that electrically, like a capacitor does with a barrier to where they're attracted such they move through the electrolyte but can't actually electrically interact (no actual power consumed)? Seems they'd push the TBABr out of the area thereby concentrating them where you wanted them to be and of course, nothing new added into the chemistry for side-reactions.

svetz said:

Just throwing this out as a 4th way (sorry if it's crazy-talk), wouldn't the Zn+2 cation be attracted to something with a strong negative charge? Would it be possible to create that electrically, like a capacitor does with a barrier to where they're attracted such they move through the electrolyte but can't actually electrically interact (no actual power consumed)? Seems they'd push the TBABr out of the area thereby concentrating them where you wanted them to be and of course, nothing new added into the chemistry for side-reactions.

Click to expand...

Zinc ions are indeed attracted to negative charges, however in solution things are "neutralized" at a very small scale. A Zinc ion will not just be like an isolated point charge in solution, it's charge will be neutralized by a solvation sphere made by water molecules with their oxygen atoms oriented towards the Zinc, this sphere will grow until the effect of the zinc ion charge is close to zero. Furthermore, there are also solvated negative ions in solution that will further contribute to the net charge of the solution always being zero.

That said, when close to a charged plate, the ions closest will indeed be attracted to it - as a function of how strong the electric field is - and this is why they will get close enough to be reduced. If you put a field that is really high but provides no current, you can indeed create a small amount of ions that get "stuck to it" because of the potential and no power is consumed, but this effects falls very rapidly as you move further into the solution and the ions are "neutralized' enough by the water molecules not to "feel" anything from the plate. You indeed create a super small capacitor - nanometers wide - but the effect is not useful to capture any large number of ions. Furthermore, this capturing is somewhat selective due to the diffusion coefficient of the ions, but you are bound to capture a distribution of everything in solution with the same charge sign.

These are some efficiency measurements as a function of the number of cycles for a cell with CC4 that was initially prepared with 100uL of ZnBr 0.5M + TBABr 0.2M electrolyte with 50uL of TBABr 1M added on top of the cathode after assembly. The cell was charged to 500 uAh at 1mA and discharged to 0.5V at 1mA. Second plot shows all the charge/discharge cycles.

I found this paper (https://pubs.rsc.org/en/content/articlelanding/2017/SE/C7SE00350A#!divAbstract) on energy efficiency in batteries very interesting.

Also, you can get a 1.0M TBABr + 1.0M ZnBr2 solution if you use 50% Isopropyl alcohol / 50% water mixture. Sadly the TBABr3 is soluble in this mixture as well, so the battery strongly self-discharges This is evident in the terrible drop in Coulombic and energy efficiencies at lower charging currents. The next test will be using PEG200, which I will do next week (shipping takes a while!).

I am also ordering trimethylphenylammonium bromide (TMPhABr) and tetrapropylammonium bromide (TPABr) so we should have some very interesting experiments coming up within the next couple of weeks

In the meantime - while I get the chemicals for the next round of experiments - I created a battery using carbon felt as the cathode, where the electrolyte was prepared by dissolving 0.720g of ZnBr2 in 1mL of water, then taking that to 10 mL in a volumetric flask using a 1M TBABr solution. As expected a lot of the TBABr precipitated out but it became suspended whenever the solution was agitated. The 100uL of the electrolyte were put into the cell right after agitating the volumetric flask, just putting the suspension on top of the cathode and letting it wick through. I want to test the long cycling properties of this device. I am running these cycles charging to 500 uAh, discharging to 0.5V.

So far these are the results:




As you can see both the CE and the EE of the cell have been increasing as a function of the cycle number. My hypothesis is that the precipitated TBABr is initially randomly distributed within the battery but is then reorganized within the cell as the cell precipitates and dissolves TBABr3 as it goes through charge/discharge cycles. When this happens some TBABr can go into solution and then when the cell discharges this TBABr precipitates selectively at the cathode, where the concentration of TBABr increases locally when the TBABr3 is dissolved.

This process seems to slowly increase the cell's internal resistance - probably because it's forming a layer of solid TBABr at the cathode after discharge - but the overall CE and EE of the cell increase as the formation of TBABr3 from TBABr becomes more efficient. I want to see when this process reaches a steady-state and how stable the CE and EE of the final cell are. It will be at least 4-5 days till I get new materials so I will cycle this cell continuously till then

So the discharge curves reaching towards 500 are higher cycle numbers? How temperature sensitive is the precipate?

Why not add some sodium like you were talking about in your blog to see what happens?

svetz said:

So the discharge curves reaching towards 500 are higher cycle numbers? How temperature sensitive is the precipate?

Click to expand...

Yes, these ones are the higher cycle numbers. I have no idea how sensitive to temperature the precipitate is (TPABr3). From a review of the literature it seems stable - melts at 71C without decomposing - but I don't know how its solubility in water changes as a function of temperature (could be an important factor). The battery tests have been done inside my house, which is kept at 69-75F but I have done no efforts to actively measure or control temperatures.

svetz said:

Why not add some sodium like you were talking about in your blog to see what happens?

Click to expand...

Since I need to buy all the reagents and perform all the experiments (and my time is sadly limited!) I limit myself to the experiments I consider have the highest chances of success, for this reason I have decided to pursue the PEG200 addition and the TMPhABr experiments next and leave aside any experiments dealing with ZnEDTA2Na2 or NaBr. This is because these experiments require me to purchase new reagents and complicate the chemistry more than I would like, given the limited characterization tools at my disposal. I might get to them one day though!

These are all the charge/discharge plots for the last described battery up until now with the CE and EE information, in case you guys want a better look into the evolution of the charge/discharge curves as a function of time.

I just published a blog post discussing this experiment (https://chemisting.com/2020/09/19/zinc-bromine-batteries-can-we-just-put-solid-tbabr-in-there/)