My adventures building a Zinc-Bromine battery

[danielfp248]

Hi Guys,

My name is Daniel, I am a chemist with a passion for battery technology and currently trying to build a highly efficient Zinc-Bromine battery at home using readily available materials. I have a blog where you can follow my progress (https://chemisting.com/). I am using a DIY USB potentiostat/galvanostat (read about how to build it here https://www.sciencedirect.com/science/ar...7217300317) in order to properly characterize the batteries I build and in this way systematically modify/improve my builds.

Right now my electrolyte is a solution containing 0.5M Zinc Bromide + 0.2M Tetrabutylammonium bromide (TBAB)

I am using Swagelok cells for the construction of the test cells (0.5 inch diameter). This is the current configuration I have tested:



So far I have achieved a 96% Coulombic efficiency, although my specific capacity sucks, at around 1.0 mAh/g of cathode material. I am going to change to a Zn anode and to carbon paper cathodes (see my latest blog post), which should help increase the specific capacity by a factor of 10-100x. This is what I hope to reproduce, from recent research papers about this battery type (stationary Zn-Br batteries).

Any suggestions/comments/questions are welcome!

 

[Michael S]

Have you seen the videos by Robert Murray-Smith on this type of battery?
Make Cell

Make Electrolyte

 

[danielfp248]

I have watched all of Robert's videos, I even decided to pay to become a member of his community to get access to several extra videos about this topic not available to the general public.

The videos have not been very useful though, mainly because of the following reasons:

1. While Robert shares a lot of info about fabrication, he shares practically zero information about real results (no charge/discharge curves, coulombic efficiency info, self discharge rates, specific power/capacity, etc) so I have no idea how good or bad the batteries he made are.
2. His choices are sometimes very odd, in particular his battery architecture deviates very substantially from what a modern battery looks like. For example the choice of a very thick phenolic foam separator causes very big internal resistance and heavily reduces ion mobility. These choices are likely to very heavily reduce both energy efficiency and specific power of the battery. The charge/discharge rates are also likely to suffer.

Because of these issues I have instead decided to reproduce the results of well documented research with high specific capacity values. In particular this recent chinese paper on highly efficient static Zinc bromine batteries (https://www.sciencedirect.com/science/article/pii/S2589004220305356). Since TPAB is hard to obtain, my first objective is to see what a battery made using TBAB can achieve within this same architecture.

Note that I am not saying Robert's batteries are "bad" just that because of the lack of information I can't tell and given some odd choices he has made I would rather go with reproducing published research instead.

 

[danielfp248]

Also note that static Zinc bromine batteries without any complexing agents - like the one shown in Robert's zinc bromine battery video outside the members channel - are of no interest to me as the self-discharge rate because of bromine diffusion is way too high, plus having any presence of pure elemental bromine at my house is not acceptable due to safety concerns.

 

[svetz]

Welcome to the forums!

Love the detail in your blog! As you pointed out in your earlier post...it's very hard to find experimental data posted so I applaud you.

I noticed in one of your photos ("c") the red coloring. I'm not sure where I read it, but it reminded me of a paper in which to the battery chemistry being tested where they added something (phenolphthalein?) to get a visual colorization as the charge/discharge was acid/base.

 

[danielfp248]

Welcome to the forums!

Love the detail in your blog! As you pointed out in your earlier post...it's very hard to find experimental data posted so I applaud you.

I noticed in one of your photos ("c") the red coloring. I'm not sure where I read it, but it reminded me of a paper in which to the battery chemistry being tested where they added something (phenolphthalein?) to get a visual colorization as the charge/discharge was acid/base.

Thanks for your welcome message! I hope the community can provide some feedback, comments and suggestions about my experiments!

About the red coloring, I certainly didn't add anything red! Might just be some light effects from the copper electrodes I have been using.

 

[danielfp248]

I have been doing some further experiments but working with the stainless steel electrodes in the Swagelok cells and trying to cover them with conductive HDPE has been a complete pain (https://chemisting.com/2020/09/10/zinc-bromine-batteries-think-about-the-electrodes/). As you can see in the charge/discharge curve I keep on getting weird shapes in discharge curves from side reactions. I am going to build some custom graphite electrodes for the cell from graphite rods so that I can perform these experiments without having to worry about this problem!

 

[danielfp248]

These are some curves I just measured at a higher current (5 mA/ 1.3cm^2) vs the 1 mA I usually use. The resistance of the conductive HDPE is also so high that to increase the current, the potential needs to increase a ton and a lot of energy efficiency is lost. The coulombic efficiency is still pretty high at the start though! Next time I post I hope to share some results using graphite electrodes in the Swagelok cell.

 

[toms]

Following this one, sorry i don’t have the knowledge to provide you with feedback - but i am keen to learn more about zinc bromide cells.

 

[svetz]

Why is the starting voltage on these under 1.4? These experiments were ~1.6V so I was bit surprised?
Also interesting... the other experiment had that same inflection, but earlier... around 1.52V and 20 µAh. Could the pH have been off?

... I keep on getting weird shapes...

My first thought was that an exothermic reaction might be slightly warming the electrode initially keeping the reaction rate higher, then around 100 µAh it hit steady state heat transference and therefore the rate declined.

But from the Sandia paper, it doesn't look like the reaction is overly sensitive to temperature, efficiencies only varying < ~6%:

​

So, some sort of secondary reaction occurring initially that falls off as the voltage drops? Water is around 1.2V... possibly the electrolyte is having a secondary reaction?

...i don’t have the knowledge .. but i am keen to learn more about zinc bromide cells.

Same here... I just remember enough to be dangerous and hopefully won't offend the OP with dumb questions. In the Sandia paper, they built a Zn-Br flow-battery out of low-cost plastic, I could see printing those on a 3D printer...but if it were so easy to make a "real" flow battery you'd be able to buy them commercially. So, it'll be interesting to follow the OP on their journey.

 

[danielfp248]

Following this one, sorry i don’t have the knowledge to provide you with feedback - but i am keen to learn more about zinc bromide cells.

I appreciate the support! Feel free to ask any questions you might have

 

[danielfp248]

Why is the starting voltage on these under 1.4? These experiments were ~1.6V so I was bit surprised?
Also interesting... the other experiment had that same inflection, but earlier... around 1.52V and 20 µAh. Could the pH have been off?

The first experiment charged/discharged the batteries at 1mA, these last experiments did so at 5 mA. The higher current means that a higher voltage is needed to charge the battery - due to its high internal resistance - which means more side reactions are likely to happen, so more energy is wasted in things like water splitting. This also means that the voltage on discharge drops more, because more current needs to go through the same internal resistance so, per Ohm's law, the discharge voltage needs to drop to compensate. This is why the energy efficiency, coulombic efficiency and starting voltage are lower.

My first thought was that an exothermic reaction might be slightly warming the electrode initially keeping the reaction rate higher, then around 100 µAh it hit steady state heat transference and therefore the rate declined.

It's certainly a good initial guess! But I doubt this has to do with temperature, because it hasn't happened on other cells I've built with virtually the same construction (https://chemisting.com/2020/09/05/zinc-bromine-batteries-first-success/), there is also only a 1mA draw from an area of 0.5 square inches, so that's hardly enough power density to significantly heat up the cell (power dissipation of just the electrolyte can easily absorb that).

About water splitting, that certainly does happen but it only happens when you charge the battery, it cannot happen when you discharge, because the discharge can only include reactions that are spontaneous without an external source of power. The pH of this electrolyte is also only slightly acidic, so I wouldn't think that Zn oxidation with water to produce hydrogen would be a significant contributor. My bet is that the side reactions are iron from the stainless steel being oxidized, this reaction has a significantly lower potential than the intended Zn oxidation - so it shouldn't happen first - but the stainless steel electrode provides a direct path for current leaving while flow from the Zn the current needs to go through the conductive HDPE. This means that - because of the voltage drop getting across the conductive HDPE - the Zn and Fe oxidations are probably in similar terms regarding how spontaneous they can be.

Of course these are all hypothesis, I haven't verified what is actually happening. However - from my past experiments - I do know that exposure of the stainless steel has to do with this happening, so I'm changing to graphite electrodes to get rid of all of this (hopefully!). I hope to share these results with you as soon as I get all the materials to build those cells!

 

[danielfp248]

Here you can see the exact same battery charged/discharged at 0.5 mA. With a low current I can charge it to an even higher capacity (500 uAh) with a higher Coulombic efficiency (94%). The lower current means that only the most favored reactions at that potential will happen and ions will be able to better accommodate at the anode/cathode, leading to better energy and Coulombic efficiency values. You can see that the "shoulder" at the start of the discharge curve is also now significantly smaller, because whatever side reaction is happening here probably happens to a smaller extent when the current draw is smaller. You can see that the voltage drop between charge and discharge is also significantly smaller.




That said, the batteries are still pretty terrible, due to the heavy cathode, the use of that conductive HDPE that is causing me to have huge internal resistance in the cells and the interference from reactions of the stainless steel electrodes that although largely covered, are still probably microscopically exposed at some places. I hope these can get much better as I improve their construction!

 

[danielfp248]

I just published a blog post (https://chemisting.com/2020/09/12/zinc-bromine-batteries-can-they-really-be-that-good/) discussing the results in this paper (https://www.sciencedirect.com/science/article/pii/S2589004220305356#mmc1) and how the published specific capacity/energy/power values for the studied ZnBr2 + TPAB system can be misleading if the way they calculated the weight of the "battery" is not well understood. Long story short, a real-life Zn-Br battery is bound to have specific power/energy densities significantly lower than what they estimate from accounting only for a very limited set of components within the battery.

 

[svetz]

ROFL: 3 mg electrode! The per cycle capacity drop wasn't pretty. I noticed they had some inflections in the voltage at various mA/g -->

 

[danielfp248]

ROFL: 3 mg electrode!

hahah, exactly! Consider the cathode is actually 80% binder: 10% porous carbon: 10% highly conductive carbon on a titanium foil. Counting all those things, the weight is probably 10x higher, at least (50x if you count the Ti foil). The carbon paper I'm going to try is actually 11mg/cathode, we'll see if I can get 500 mAh/g!

The per cycle capacity drop wasn't pretty. I noticed they had some inflections in the voltage at various mA/g -->

They did comment on those in the paper. They say those shoulders are caused by the solid tetrapropylammonium tribromide solid dissolving when the discharge reaction starts. I haven't observed such behavior in my TBAB batteries, but then my potential losses due to my conductive HDPE to cover my stainless steel electrodes are horrible. Once I change to graphite electrodes next week we'll see what I get.

 

[danielfp248]

I ran another test on the same architecture at 0.5 mA but this time charging to 1000 uAh. I was able to get a Coulombic efficiency of 88.2% with an energy efficiency of 74%. This gives me a lot of hope for the upcoming graphite electrode plus carbon paper tests I'll be carrying out next week!

 

[danielfp248]

If I try to take this to 2000 uAh at 0.5mA the Coulombic efficiency drops to 69%, with an energy efficiency of 33.6%. Seems most of the energy is wasted when I go significantly above 1000 uAh. Given the amount of volume of electrolyte (0.1mL - 0.5M ZnBr2 + 0.2M TBAB), this implies that I am able to efficiently harvest as much Br2 as there are TBAB molecules to capture it, but when I go beyond that most of the energy is just wasted, probably due to the formation of elemental Br2 and the cross-difussion of it across the thin cell.



I am therefore going to do my experiments with carbon paper and graphite electrodes for the swagelok cell next week at 1000 uAh at 0.5mA. If I am able to keep the capacity using the thin cathode the capacity of the battery will be very significantly enhanced.

 

[danielfp248]

Just published a post about some thoughts in building an actual practical battery (https://chemisting.com/2020/09/13/zinc-bromine-batteries-initial-thoughts-about-a-practical-battery/). Still very far away - I need to understand TBAB cells, other carbon cathode choices, different ZnBr2 concentrations, etc - but worth thinking about.

 

[danielfp248]

I just built my first cell changing my Swagelok cell stainless steel electrodes covered with conductive HDPE to graphite electrodes. The results are pretty awesome. This eliminated all shoulders in the charge/discharge curves (they were side reactions from the stainless steel!) and the internal resistance dropped so much that I can now use much higher current densities with even better efficiencies. This is using the traditional carbon felt electrode - haven't started testing my carbon papers yet - using the new graphite electrodes, charging/discharging to 100 uAh at 5mA. The electrolyte is 0.5M ZnBr2 + 0.2M TBAB.



Note that this is only 1/10th the capacity I've charged the previous cells to (1000 uAh), but at this capacity I am getting a Coulombic efficiency of 89% with an energy efficiency of 72%.

With the previous electrodes the resistance was so high that my potential needed to be at 2.2V to charge at 5mA! Now I'm getting the potential I got with the previous cells at 0.5mA.

 

[danielfp248]

I just got the sample kit of carbon cloth electrodes I will be testing within my cells. It contains the following carbon cloth choices, all from AvCarb (https://www.fuelcellearth.com/fuel-cell-products/carbon-cloth-variety-kit/):

CCP – Carbon Cloth Plain 15 mil thick
CC4P – Carbon Cloth Plain 25 mil thick
CC6P – Carbon Cloth Plain 35 mil thick
HCB-1071 – Carbon Cloth Plain 15 mil thick

Carbon felt they are being compared to (https://www.ceramaterials.com/product/gfe-1-pan-graphite-felt/)

These are all untreated and I will use them as-is.

Cell structure will be as follows:



Electrolyte will be 100 uL of ZnBr2 0.5M + TBAB 0.2M, all drawn from the exact same solution.

The basic test will be to do 20 cycles at 1mA, charging to 500uAh.

 

[danielfp248]

A picture of the first test cell I made for these tests, now evaluating the CC6 carbon cathode. Note this cell goes into the Swagelok cell for testing.

 

[svetz]

That'll be interesting! Sidebar: Just ran across a WSU/PNNL article about O3 & salt ions...

 

[danielfp248]

Best results for CC6

Coulombic efficiency = 73.54%
Energy efficiency = 45.26%
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

 

[svetz]

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?