PaulD's Experiments with Endothermic Potting Compounds

[PaulD]

I did this work almost a year ago following a conversation with Luke (LFP) who mentioned he was working on endothermic fillers that could be useful in preventing thermal runaway and/or runaway propagation. The idea, of course, is that heat would be absorbed from a cell that is overheating to either prevent that cell from entering thermal runaway, or preventing that heat from triggering the neighboring cells into thermal runaway.
I decided to do my own investigation and researched the typical compounds used in endothermic fire retardants such as aluminum hydroxide, magnesium hydroxide, antimony trioxide, etc. Unfortunately, these all begin to react at temperatures much higher than those needed to initiate thermal runaway. After some soul searching, I vaguely recalled that sometimes sodium bicarbonate is used as a fire extinguishing agent. Yes, baking soda. Upon further reading, I found that it starts to react endothermically at 80C, much much lower than the others. Not only that, the byproducts of the reaction are CO2 and water. Both useful things for extinguishing fires. The amount of heat absorbed is quite large, about 1130 Joules per gram of NaHCO3.
My first test experiment involved 2 identical setups with a heater cartridge with 4 thermocouples (T1-T4) spaced between it and a cell (T1 closest to heater, T4 is on the cell). The first setup had ordinary platinum cure silicone potting, the second had 33% NaHC03 mixed in. Ignore the lumpy silicone, I recycled some old chunks of cured silicone to reduce waste .


The results were quite impressive. There was a massive drop in temperature in the filled potting. You can even see the temperature at which it starts reacting.

I think this could be extremely useful as a security measure to prevent TR. Especially for those DIY-ers building sketchy packs with recycled cells. Much much more work is required to understand how much space is needed between the cells for this to be effective. A mix of analysis and real propagation experiments is needed. The next step for me is to complete my thermal runaway chamber where I can (sort of ) safely trigger ebike size packs into TR.

 

[fechter]

Wow, that's interesting. Does the sodium bicarbonate mix OK with the silicone?

 

[PaulD]

Wow, that's interesting. Does the sodium bicarbonate mix OK with the silicone?

Yep. And no observable change in how the silicone cures. I mixed the baking soda in after mixing A and B parts. The viscosity was increased, but still filled the gaps well. I would assume the physical properties of the silicone would be affected.
Not sure if this would work when using a structural potting compound, like in my conversion kit batteries. Perhaps I'll run a test with polyurethane.

 

[A-DamW]

This is really cool, subbed!
I wonder if electrical conductivity will be a problem.

 

[neptronix]

Very interesting!!
I'm also interested in increasing lithium battery safety.

One problem with preventing thermal runway is that the heat is really in the center, not the outside.
Sapping heat from the outside could possibly slow down the reaction, but can we really prevent the thermal event?

The only way to find out is to do a liveforphysics battery kaboom test i think ..?

 

[lgarczyn]

Cool advantage of sodium bicarbonate: it will turn into sodium carbonate if heated too much, but it will slowly turn *back* into sodium bicarbonate if left on its own, by absorbing co2 from the environment. So it's a sort of resettable thermal condensator.

What I would be worried about is if too much humidity gets into the pack, it could act as a conductor?

 

[amberwolf]

One problem with preventing thermal runway is that the heat is really in the center, not the outside.
Sapping heat from the outside could possibly slow down the reaction, but can we really prevent the thermal event?

If something can be done to at least prevent the event from spreading to any other cells, it would at least prevent a fire, even if the pack itself is destroyed (at least that cell group) by the failure of the initial cell.

 

[PaulD]

Sapping heat from the outside could possibly slow down the reaction, but can we really prevent the thermal event?

The only way to find out is to do a liveforphysics battery kaboom test i think ..?

The short answer is always "it depends". All runaway events are different. For example there are many many variables that affect whether an internal short results in anything from a soft short to cell venting, to full TR. My thought is that anything outside a cell that absorbs a lot of heat will reduce the odds of TR, but certainly can't prevent it in all cases.
I think the main value in this filler is to prevent propagation. So that is what I intend to focus the next tests on.

 

[PaulD]

This is really cool, subbed!
I wonder if electrical conductivity will be a problem.

Thanks for your feedback. I will do a bit of research on this. I did a quick 1000V hipot test without issue.
@lgarczyn - I will run a test after soaking in water. I think with the baking soda particles immersed in the silicone, they should be sealed off from water that might get in, but that's an interesting thought that deserves investigation.

 

[neptronix]

The short answer is always "it depends". All runaway events are different. For example there are many many variables that affect whether an internal short results in anything from a soft short to cell venting, to full TR. My thought is that anything outside a cell that absorbs a lot of heat will reduce the odds of TR, but certainly can't prevent it in all cases.
I think the main value in this filler is to prevent propagation. So that is what I intend to focus the next tests on.

That's true.
The worst case possible will tell you a lot.

Preventing propagation would be enough!

 

[liveforphysics]

Love to see this work Paul! Thank you for sharing your tests.

With the right endothermic fillers, you can pack feisty cells tightly together and still have 100% success in passing PPR (passive propagation resistance). If you have a Ramset nailer, throw a nail or two into cells in your test cluster (after charging them all to 100% SOC). If you need a space place to do that, we have 1.5inch thick 316 stainless chambers with ceramic lining (and hundreds of isolated data acquisition channels for temps and voltages etc).

 

[amberwolf]

FWIW, silicone is generally hygroscopic, in that water eventually permeates it.

For instance, if you have silicone seals on an aquarium (freshwater or not), they will be perfectly fine as long as the tank is used, kept full of water. Let the tank dry out long enough, and the silicone will shrink and pull away from the glass as the water it had absorbed works it's way out of it. Usually this is RTV-cure silicone using acetic acid.

The same thing can happen to platinum-cure silicone used to make molds for water-based substances, which are used often or continuously (when you stop using them for a long time they can then deform).

I don't know if it's enough water to make a difference in this case, but it is worth checking into.

I have too little experience with polyurethane or other fillers to guess if those would have such problems.

 

[PaulD]

Thanks for the feedback Luke and Amberwolf. Regarding the waterproofness of silicone, it does seem to be used successfully in wet environments, but once again further testing needed. I know Luke is a big fan of polyurea, but I like that the silicone is easy to use and allows disassembly to a certain extent. Urethanes are generally crazy strong and permanent.
After talking to Luke, I decided to try a few more materials that have more hydrates per molecule - so higher potential to release water and absorb heat. I limited my materials to things I can find easily and are non toxic. Both came from the hardware store These samples are curing now. Epson salts (MgSO4*7H2O)and washing soda (Na2C03*10H2O) will be compared directly to baking soda in a simplified test. The little clear 3D printed things are just to hold thermocouples in a (sort of) precise location.
If neither is better than baking soda, at least I will have clean clothes and be free of constipation for a while.

 

[spinningmagnets]

This is awesome! Thanks for posting this PaulD!

edit: I just added a link to this discussion in the "Battery stickie index" in the "potting discussion" thread, so its easier to find in the future.

 

[amberwolf]

Thanks for the feedback Luke and Amberwolf. Regarding the waterproofness of silicone, it does seem to be used successfully in wet environments, but once again further testing needed. I

Yes, it is used, but it depends on the conditions as to whether it will or will not do the job needed without side-effects.

For instance, if the silicone over time absorbs enough moisture from the environment for it to migrate thru the silicone and up against the metals the silicone is bonded to, it could allow corrosion of the metals.

If the silicone is acid-based (like the acetic-acid (vinegar smell) GE Silicone I type), then even without water it may cause corrosion of metals, then if water migrates to the metal it could be worse. (I don't know, as I havent' tested this directly, only seen results in various disassembled devices that I don't know all of the manufacturing and usage circumstances).

 

[CONSIDERABLE SHOUTING]

Watching this thread as well after seeing it mentioned- thought I'd bump it up to see if there were any updates.

 

[PaulD]

Well, what a coincidence, I just finished some testing today. It took way longer than expected, and the test setup had to be changed twice, as it didn't hold the thermocouples precisely enough to get good data. I redesigned the experiment and printed some fixtures on my resin printer to hold the thermocouples in exactly the right locations, then I glued everything in place and potted with the 3 different fillers and 1 control. Thermocouples were secured 1mm and 1.6mm away from the cartridge (T1 and T2). Heating is 20W for 2 minutes. Heater cartridge is 6mm diameter, 20mm long. I recorded the temp for about 15 minutes.

The experiment still didn't go exactly as planned, but the data is much better. I'll explain the issues and my thoughts on the accuracy below. But here are the results:

Thoughts and Conclusions:

• All of the fillers clearly absorb heat better than unfilled silicone
• Heater cartridges in both the washing soda and epsom salts came loose due to gases (water and C02) being generated and heated which forced the heaters up, and thats why you see those blips in the curve - although part of the blip is the endothermic reaction doing it's thing. I pushed the cartridges back into place as quickly as I could. The video shows it took me a while to notice, and this likely had an impact.
• Washing soda looks really good, but I am worried that during decomposition, one water molecule is lost at a very low temperature which is well within the allowable operating temp of a lithium battery. This could lead to corrosion.
• The Epsom salt trial had to be run twice, as it popped the cartridge out quickly, and had to be re-started. I think it would have done better with fresh un-reacted material.
• Both washing soda and Epsom salt are quite a bit denser than the silicone, and therefore settle during curing. Since the test setup places the TCs down near the bottom of the cup, the filler loadings are likely much higher where it counts, skewing the results in their favor.
• The particle size of the washing soda and epsom salts is too large (even with some manual grinding with pestle and mortar). I imagine they might mix better with less settling if they were as fine a particle size as the baking soda.
• Epsom salt decomposition can release sulfur dioxide - not ideal.

Next steps:

I think I will have to repeat these tests with a better setup (again). Unfortunately the gas generation causes everything to move around making it hard to keep sensors and heaters located precisely.
The impatient part of me just wants jump to the part where I pot a pack with the baking soda filler, and do a full on nail penetration propagation test. I'll be working on my nail pen test chamber in the coming weeks if time allows.

Edit adding another setup photo:

 

[amberwolf]

You can use something like rotomolding/casting's methodology to ensure better filler distribution and help prevent settling. Those use high speeds which you won't want, because it would force everything to the outer edges. But if you can make the filled mold rotate "gently" on multiple axes (either one at a time or all at once), it would help suspend the filler in the silicone.

For some reason google will not serve me up an image of one, but there are exercise devices that rotate on 3 axes. If you've ever seen the terrible movie "Event Horizon", their gateway drive is built this way.

 

[DogDipstick]

View attachment 331677

That sure is a very nice, 2017 hole, anodized, high grade aluminum, fixture plate. In its 23" x 13" dimension. Quite nice.

Thank you for the sale.

So far I have sold about 500 of those.

-jeepsells2 from eBay

Lol.

 

[PaulD]

You can use something like rotomolding/casting's methodology to ensure better filler distribution and help prevent settling.

Yes, I think I might have to do this to prevent settling. But, the end goal of this research is to provide a way for the average user to buy some cheap materials, mix them in their garage (or kitchen) and fill their pack to make it much much safer. Perhaps if wouldn't be that big of a deal to rotate the pack once or twice during curing at specific times to decent (but far from perfect) distribution. For my testing, I have open molds, but I'll have to figure out a closed container with TC and heater wires exiting through seals to be able to rotate it.

That sure is a very nice, 2017 hole, anodized, high grade aluminum, fixture plate. In its 23" x 13" dimension. Quite nice.
Thank you for the sale.

They make great photo backdrops too!

 

[amberwolf]

Sounds like doing more fine-particle-ization of the materials before mixing is a better solution for the end-goal, then. The finer they are the better suspended they will stay (based on my experiences with casting and molding platinum silicone with various additives).

 

[jonescg]

Great research mate!
I'm going to be doing a PCM (phase change material) battery and testing it in a very similar way. Not to prevent thermal runaway propagation, but to absorb the heat produced by a discharging/charging cell by exploiting the latent heat of fusion of paraffin/beeswax or something. Finding a suitable melting point will be the challenge, and getting the battery to an appropriate temperature for filling it with liquid PCM will be hard too. Still, worth a try.

 

[amberwolf]

If it's helpful, back when I worked at Honeywell CFSG, there was a project in engineering (where I was visiting a coworker; I worked on the factory floor in final testing/QC) where they were developing a PCM (I don't know for what, it was "secret") out of a paraffin / graphite compound blended with paper fibers to physically stabilize it so it wouldn't "leak" everywhere. It resembled shiny wet cardstock, but was not actually "wet", just slick-feeling.

Since that was in the late 1980s, I don't remember the testing results for heat capacity and such....

 

[PaulD]

Sounds like doing more fine-particle-ization of the materials before mixing is a better solution for the end-goal, then. The finer they are the better suspended they will stay (based on my experiences with casting and molding platinum silicone with various additives).

Agreed, I'll grind as much as my forearms can handle! I might borrow the Vitamix if I think I can do it without my wife noticing.

Great research mate!
I'm going to be doing a PCM (phase change material) battery and testing it in a very similar way. Not to prevent thermal runaway propagation, but to absorb the heat produced by a discharging/charging cell by exploiting the latent heat of fusion of paraffin/beeswax or something. Finding a suitable melting point will be the challenge, and getting the battery to an appropriate temperature for filling it with liquid PCM will be hard too. Still, worth a try.

I had a client that was building batteries at the same factory in Taiwan as Gogoro about 6 years ago. I saw the machine that injected PCM it into their packs at the end of the assembly line. I can't remember a whole lot, I never knew what material it was, but the rumor was that it was wax that was actually flammable. Hope they changed that.
Allcell is the other company that comes to mind when I think of PCM. Maybe check out their patents. I think it was some kind of graphite matrix with wax in it, similar to what Amberwolf mentioned.
The downside to PCM is that long cool-down time. I was surprised that a battery share service like Gogoro would use it for that reason, but maybe it just never got so hot as to need cool down time before recharging for the next user.

 

[99t4]

Agreed, I'll grind as much as my forearms can handle! I might borrow the Vitamix if I think I can do it without my wife noticing.

Try a coffee grinder.