My adventures building a Zinc-Bromine battery

[danielfp248]

First 80 cycles have passed. The last curve's CE is 91.55% and EE is 73.66%. This battery seems to be incredibly more stable than my previous batteries. So far, zinc dendrites do not seem to be an issue.

 

[svetz]

Nice! Any idea what the bump at 45 is about?

 

[danielfp248]

Nice! Any idea what the bump at 45 is about?

Thanks!

I don't know what the bump was about. However, I can speculate a bit. In other cells 45 cycles was around the point where Zinc dendrites started to become a serious problem - which was evident after opening the batteries up - in this case it's likely that dendrites also started to build to a certain point but then were somehow eliminated in subsequent charge/discharge cycles. If this was the case we are likely to see the battery oscillate 85-90% CE as time goes on. I haven't even touched the battery or measuring equipment, so whatever is happening was fully due to the battery chemistry.

I had never run one of the these Zn-Br batteries for so long, so it's really exciting to see we can cycle them 80+ times at 100% depth with no large degradation in their CE or EE

 

[danielfp248]

After 100 cycles the CE and EE remained stable. Final values were CE=89.79%, EE=73.68%. I am now going to try another 100 cycles at a higher current density (these were done at 2mA, I am now going to try see how it changes at 5mA).

 

[Altin961]

i'm really interested in Zn-Br chemistry and believe that the solution is to make the battery no flow setup. I know that some manufactures are working with Zinc Bromine Gel electrolyte and they change the discharge profile of the battery with the thickness and viscosity of the gel layer.
Thanks for sharing your research Daniel, i'm also reading your blog

 

[danielfp248]

i'm really interested in Zn-Br chemistry and believe that the solution is to make the battery no flow setup. I know that some manufactures are working with Zinc Bromine Gel electrolyte and they change the discharge profile of the battery with the thickness and viscosity of the gel layer.
Thanks for sharing your research Daniel, i'm also reading your blog

Thanks for following my progress and reading my blog

The gel batteries by gel-ion are certainly interesting. You can read more about their approach by reading the patent for their technology:

https://patents.google.com/patent/EP3103156A1/en

The approach certainly makes sense, being that the researchers working on this come from the ionic-liquid world.

As a humble solo chemist I am hoping to build an understanding about a much simpler approach and hopefully contribute something useful to those interested in this type of technology.

 

[danielfp248]

Charging at 5mA to 500uAh, then discharging at 5mA to 0.5V. So far so good. Last cycle CE=95.24%, EE=67.51%

 

[danielfp248]

I had to restart the test because of a computer issue (computer restarted in the middle of it for updates ). After 70 cycles of the new test the CE=95.55% and the EE=66.86%. It seems the battery is perfectly stable, after going through 100 cycles at 2mA and 100+ cycles at 5 mA, the CE and EE at each current density remain close to the initial values.

 

[danielfp248]

I just wrote a post about the these results using TMPhABr (https://chemisting.com/2020/10/02/zinc-bromine-batteries-going-for-high-capacity-with-tmphabr/). I am now going to try to use TMPhABr in a solid layer and a ZnBr2 at 2M saturated with TMPhABr as the electrolyte to attempt to get the capacity up to 2000 uAh.

 

[danielfp248]

First curve trying to charge to 2000 uAh using 2M ZnBr2 solution saturated with TMPhABr plus a solid TMPhABr layer. CE=75.8%, EE=56.4%. Let's see if it improves on subsequent curves!

 

[danielfp248]

After 4 cycles we are at CE=79%, EE=58%. Last cycle shown below.

 

[danielfp248]

After 5 cycles dendrites started to become a problem. You can see this in this charge cycle as the voltage drops as dendrites start to generate shorts in the battery. It seems the higher ZnBr2 concentration has reduced the solubility of TMPhABr to the point were dendrite formation is happening yet again and causing big stability issues. We are depositing 4 times more Zn per charge cycle now, so I had some suspicions this might happen! Time to reflect a bit before the next experiment ?

 

[danielfp248]

Next experiment has had a couple of modifications. First, I have added 1% PEG-200 to the electrolyte to reduce Zinc dendrite growth. Second, I have changed the battery structure to go back to 8 continuous layers of fiberglass, followed by a 50mg layer of TMPhABr, right next to the cathode.



These are the results after 7 cycles at 2mA, charging to 2000uAh, discharging to 0.5V:


Last curve:


Last cycle CE=94.43%, EE=63.08%.

There is also a small problem with the addition of the saturated electrolyte because it often carries some undissolved TMPhABr that I end up depositing on top of the CC4 cathode. Filtering the electrolyte is likely going to help with the drop in potential from these solids causing resistance between the CC4 cathode and the graphite electrode of the Swagelok cell.

There is substantial internal resistance showing up in this entire process, reason why the EE has dropped substantially from the first TMPhABr experiments at lower ZnBr2 concentrations. I am going to continue cycling the above cell and possibly run a couple of new experiments filtering the electrolyte before I start experimenting with my other cathode materials to see what difference they can make.

 

[danielfp248]

The above cell remained stable for another 5 cycles, ending at a CE=93.03% and an EE=62.43%.


I have opened up this cell and built a new one, making sure to filter the electrolyte before applying it on top of the cathode (we'll see if that makes a difference or not)

 

[danielfp248]

I have made the new cell with filtered electrolyte and have also decided to introduce a waiting time of 5 minutes after adding the electrolyte - before closing the Swagelok cell - to ensure the electrolyte is able to properly soak through the entire device. These are the results of this cell so far, charged/discharged at 2mA to 2000uAh and discharged to 0.5V:



Last curve had a CE=88% and an EE=71%. The device wants to continue to improve so I'll cycle it to see how it behaves and to evaluate if Zinc dendrites will become an issue. This is the last curve:


Remember this is an electrolyte with 2M ZnBr2 that has been saturated with TMPhABr and with 1% PEG-200 added, with a 50mg layer of TMPhABr between the cathode and the first layer of fiberglass. The device uses a CC4 carbon cloth cathode.

It will likely take a week to cycle this cell. Time after which I believe it will be time to start testing differences between cathode materials.

This is a big step! A 4x improvement in specific energy ? moving way closer to something that would have specific energy greater than that of lead acid

 

[danielfp248]

So I decided to take apart this battery after 11 cycles. Here's the final result:


Final curve has a CE=86.41% and an EE=68.74%.

The reason why I took it apart is because there started to be some deterioration of the cell. The coulombic efficiency wasn't sufficiently high and the cycles were taking too long. Plus I have only one Swagelok cell and I wanted to perform an interesting test today since I had the time to put together another cell before the start of the week

Given how the solubility of quaternary ammonium bromides drops as a function of ZnBr2 concentration, it is likely unnecessary to pre-saturate the ZnBr2 solution with TMPhABr before putting it in the cell, especially if we include a high enough amount of the solid within the cell. The idea is then to move to what the ideal concentration of ZnBr2 in a high density battery would be (around 3M + 1% PEG-200) and see if we can just add that to the battery if we include a 50mg TMPhABr layer between the CC4 cathode and the fiberglass layers.

I noticed that after closing the Swagelok cell the TMPhABr solid layer is compacted quite nicely between the cathode and the fiberglass layer, so I decided to close the cell before adding the electrolyte - to compact it - then open it up again to add the electrolyte, waiting until 100uL soaked entirely through the cathode and the fiberglass layers (around 2 minutes).

I have built a battery per the above in my Swagelok cell now. I am going to try to charge this to 5000 uAh at 5mA, 10x the specific energy of the initial cells I was testing. If this is achieved, the battery will be above the specific energy of lead acid and close to the upper part of the range of commercial flow Zn-Br batteries. Will we achieve this amazing feat?? (probably not on the first try )

 

[danielfp248]

So, going to higher capacity/current density, zinc dendrites become a huge problem again, really fast. After 5-10 cycles the separators are fully pierced and the battery is effectively killed by dendrites.

This a classic problem when trying to increase the charge speed of Zinc containing systems, as higher currents tend to create more dendrites and less homogeneous deposits. Going back to the drawing board for a bit, to think how I might address this issue more effectively ?

 

[NMNeil]

You may have already seen this patent application from Princeton regarding dendrite formation.

https://patentimages.storage.googleapis.com/63/98/49/d5851cb53c827e/US20200036046A1.pdf

 

[danielfp248]

You may have already seen this patent application from Princeton regarding dendrite formation.

https://patentimages.storage.googleapis.com/63/98/49/d5851cb53c827e/US20200036046A1.pdf

Thanks! I have read these patents. Their batteries are very different though - they form elemental Bromine - and they therefore have very particular configurations, materials and very bad self-discharge problems (due to the migration of Br2 through the cell). Zinc dendrites for them are not so problematic - as they fall and basically react with the elemental bromine in this configuration - but for the battery I'm trying to build (where Br is sequestered as solid perbromides and a fiberglass separator is present) the problem needs to be fully avoided.

 

[danielfp248]

I just published a blog post about the results at higher capacities, the Zn dendrite issue and some tests I will be carrying out to try to address it (https://chemisting.com/2020/10/06/zinc-bromine-batteries-problems-at-higher-capacities-with-tmphabr/)