Eric Darcy - NASA

SPEAKER_01 0:00

Have you ever had a a thermal event in a pack deployed in a mission?

SPEAKER_04 0:04

No. No, we've had all our f all our failures are they've been on the ground. Thank the Lord for that one.

SPEAKER_01 0:27

Welcome to the Battery Potential Podcast. I'm Vincent Chevrier. And I'm Kevin Eberman. Welcome to Battery Potential. Kevin?

SPEAKER_03 0:38

Vincent.

SPEAKER_01 0:40

Today we have a fantastic guest, Eric Darcy, is joining us. Eric had a 38-year career at NASA and worked on all things batteries at NASA. Um Eric, it's a real pleasure to have you join us today. And let's start off with how did you get into the world of batteries?

SPEAKER_04 1:04

Well, Vincent, Kevin, thank you for having me.

Grad School and Modeling Roots

SPEAKER_04 1:08

I basically got into batteries when I went to grad school. After uh going to undergrad and majoring in chemistry, I realized that I wanted to get into engineering because I love math. And I was at a kind of an artsy fartsy school, Pomona College, in uh in the smoggy part of LA back then in the in the 80s. And uh they didn't have engineering. Uh luckily I was able to take some classes at Harvey Mudd College and in engineering, and that set me up for uh getting uh into chemical engineering uh at Texas AM. And so that was a real nice change. And at Texas AM um I had to pick an advisor, and a lot of uh traditional chemical engineering fields were offered, and the one that seemed to be the most novel to me at the time was batteries, and uh, thanks to Dr. Ralph White, and so I joined his group, and he was big into simulation modeling. Very mathy.

SPEAKER_03 2:14

You're interested in math and you got to do math, so yes, yeah, perfect. Yeah.

SPEAKER_01 2:19

So you completed your your master's in 1987 at Texas AM, and you completed it under Ralph White. Now, Ralph White was a student of John Newman, is that correct?

SPEAKER_04 2:32

Correct, yeah, at Berkeley. Yeah, so he uh Ralph really introduced me to how you could learn by simulation of electrochemical systems, so really learned electrochemistry um and being able to use math principles to eliminate a lot of experiments or guide the experiments towards the results that uh we want to get to. And and so yeah, it's uh it's a very powerful technique,

Fortran Simulations to NASA Hire

SPEAKER_04 3:01

and but back then we had limited tools. I mean, uh uh I was using Fortran. Um my code was some four three thousand lines of Fortran. It took me nine months to uh to get it to debug to where it was modeling in two dimensions a lead acid battery or one electrode, the positive electrode of a lead acid battery, um, and to try to see the impact of gradients in in those two dimensions on that. And yeah, and it it was it was fun, it was interesting to do, but it was also a lot of pulling your hair out to to to debug the Fortran code um on that. So it gave me a real appreciation for becoming an experimentalist, which is what I ended up doing in my career. Luckily, the uh research I was doing was partially funded by NASA at the time, so we got a chance to report to NASA uh on our results, and then happened to be at the right time, the right place. Uh, they had an opening in the battery group. Uh, the battery group was one person, so I doubled the size of the battery group uh at NASA Johnson when I joined.

SPEAKER_03 4:17

Who was the other person?

SPEAKER_04 4:19

Yeah, Bob Bragg is uh is my mentor, he's uh a background in physics and just learned batteries uh along the way. Uh so he had been here at least 20 years before and did all the Apollo batteries. Wow. So he had some great

Early NASA Work and Safety Culture

SPEAKER_04 4:35

stories. And I joined in 87 uh after the Challenger accident, and so they were revamping, uh reviewing all the systems for uncovered hazards, and uh particularly looking at the uh fault tolerance of batteries at the time because they wanted the desire was to be too fault tolerant to catastrophic hazard.

SPEAKER_01 5:02

Right. Yeah. So in in your master's the Newman approach is using differential equations to model a system. So your master's was all theoretical or mathematical. And how how did you end up finding the job at NASA? Like how how did you end up in an experimental job at NASA?

SPEAKER_03 5:20

Yeah, how did you go from the simulations into you know building a real system?

SPEAKER_04 5:24

I must have impressed them when I was presenting my mathematical results, my simulation results. They liked the way I was able to answer questions and think on my feet and be able to communicate. I think that's really what they were looking for. Somebody would drive. Um and they took a chance on me. And yeah, uh in the government, you know, when the floodgates open, they've got to have somebody ready to jump in and and and fill those spots. And I uh happened to be at the right spot at the right time.

SPEAKER_01 5:57

Why did you apply to NASA?

SPEAKER_04 5:59

Well, because I wanted I really wanted to push new technology. The petrochemical industry was very mature and established at the time. Refineries and chemical plants. Um, I did several plant trips and visited them. Uh affecting change was a little more difficult there. Uh, NASA was more into innovation.

SPEAKER_03 6:21

Yeah.

SPEAKER_04 6:21

Um, there was a cut in pay, of course, uh working for the government. But that uh I figured, hey, I'd I tried this for a couple years, be a government employee, and then I'd move on. And th 38 years later, you know, I finally moved on and retired. Yeah.

SPEAKER_01 6:40

Right. No, nothing as yeah, nothing's as as permanent as a two-year uh two-year commitment.

SPEAKER_04 6:47

Yeah, and and and so what I did there is I worked uh multiple systems, kind of learned the ropes with Bob Bragg about uh how do you get safe batteries to power the space suit and uh the crew equipment and any uh spacecraft that we were planning. You know, the shuttle at the time was powered by fuel cells, and so we had a separate group that was handling the fuel cells, even though it's electrochemical. Yeah. So I I I worked there for about five years, and then I realized that if I was gonna get the doctorate, I really needed to go back to school before all my math skills would get too rusty. And so I applied for uh a fellowship, a sabbatical, and was fortunate enough to get it. And and then I uh got my doctorate under Richard Pollard at uh University of Houston, and there I was working nickel metalhydride cells. Ah, okay.

SPEAKER_01 7:49

So coming back to

NASA Lab Focus and Primary Cells

SPEAKER_01 7:51

you joined NASA, what does the lab look like? Like how many channels do you have? What are your resources? Like, how are you testing batteries? Like, just walk us through like a day. Like you start at NASA, what does it look like?

SPEAKER_04 8:03

Sure. So we uh we have other centers, uh Glen Research Center, at the time it was Lewis uh Research Center in Cleveland that did a lot of the cycling of nickel hydrogen cells on that. So we didn't do too much of that, we were really focusing more on primary cells: lithium phyyl chloride, lithium bromine chloride complex, uh the BCX uh chemistries. So believe it or not, that we actually flew phyyl chloride cells in the helmet light of the astronauts before we wisened up and realized that that's just way too toxic and zero fault tolerant to a leakage event on that. And uh and so then we uh wisened up and and moved to much safer chemistries, uh lithium MNO uh MNO2, for example. Primaries, you know, the camera type of batteries on that. But yeah, so so what we would do is really uh test these oxyhalide chemistries, the primary chemistries for different applications. We were actually developing at the time a spacesuit battery that was a primary uh lithium BCX composed of double D cells on that. Um and uh the scenario appeared to be feasible for uh space shuttle-based spacewalks, but then when the space station came along, we really need to go rechargeable. And so they reverted back to silver zinc batteries, which is what's was used in Apollo very successfully. But but silver zinc batteries are made by hand, and the variability was very high. Um, they they have all sorts of maintenance issues. Um yeah. And who was making those cells? Uh two companies in Connecticut, uh Yardney Technical Products and the BST. Yeah, and so that was uh you know the first introduction of lithium ion um as the primary life support system to the spacesuit. Uh it took until twenty two thousand and eight to to get that certified on that.

SPEAKER_01 10:37

Oh so that's interesting. Yeah, so you you join in '87. You're using primary cells, so single discharge cells.

SPEAKER_04 10:47

Along with some along with some NICADs and some silver zincs at the time. And then yeah, and then uh the satellites are using NICADs or nickel hydrogen. The Hubble was making the conversion from NICAD to nickel hydrogen at the time. This was in uh in 1990s.

SPEAKER_01 11:08

And all of these cells were provided by companies like Yarni or BSC. Eagle Pitcher. Eagle Pitcher. Eagle Pitcher as well. Yeah. Yardney eventually got purchased by Eagle Pitcher. Eagle Pitcher to this day exists, still makes specialty cells. Um

Lithium-ion Considerations

SPEAKER_01 11:27

when when the lithium ion battery gets commercialized, 91 by Sony, um, sort of what's your perspective as a professional working in the realm of batteries at NASA? What do you think?

SPEAKER_04 11:40

Yeah, and we so so we look at these one amp hour uh 18,650 cells. At the time we were using uh nickel metal hydride four-thirds A cells, so 17 millimeter diameter, uh a little taller, I think 67 tall, a little a little bit taller, and yeah, and metal hydrides you know didn't have the flambo electrolyte. The capacity of the cells, I think, was about three, three and a half amp hours, but the voltage was three times lower.

SPEAKER_01 12:16

Right.

SPEAKER_04 12:17

Yeah, yeah.

SPEAKER_01 12:18

So like a one point two, like a one point two, one point three-ish volts. Right. Um energy is gonna be the voltage times the capacity, which is what you know determines the usable life of one.

SPEAKER_04 12:34

So in the nineties, yeah, we were looking at uh the lithium ion, but uh uh from a specific energy standpoint, it didn't provide a compelling reason to put up with the hazard of the flammable electrolyte. Right. So in the 90s, we pretty much ignored it and just went and we developed helmet lights with nickel metahydride. We developed a pistol grip tool. Yeah, the pistol grip tool is a fancy brushless motor type of uh power tool that uh was suitable for astronauts to use with a gloved hand during a spacewalk. Like a drill? Yeah, exactly. It's a power drill. Okay, yeah, and it's got a lot of fancy settings. I think Goddard was the uh space flight center that brought that bought the tool. It came with no batteries, and so we had to develop, our group had to develop a metal hydride or solution, and metal hydride was the quickest way to do that. And at the time we put 30 of these uh four-thirds A cells in a that sounds heavy. Yeah, oh yeah, yeah, it was a good fist uh size battery uh on that, and it it it fit into the holster. Basically, the battery case was already designed, so we had an envelope volume that we had to uh fit as many cells as we could cram into it. I see, okay. And so we did that. We also did a rechargeable EVA battery that was for accessories that would be in the backpack outside kind of the backpack into the the thermal garment flap that would power the video camera on the helmet, and also power uh the glove heaters at the fingertips of the astronauts' gloves. Because uh right opposite their fingernails, they'd have these little heaters, and when they get into their EVAs, there's 30 minutes of darkness, 60 minutes of sunshine. In those 30 minutes of darkness, their hands can get cold, their fingertips can get cold, particularly because they want to have a lot of tactile feeling when they're handling stuff.

SPEAKER_01 14:51

You know, what comes to mind for me listening to you describe these applications is you had to design battery systems for very different requirements. I imagine that the requirement for a power tool is completely different from a headlight from you know a glove heater.

SPEAKER_03 15:08

Um it's also kind of weird to me, is because I I worked for a while in this uh professional safety group, you know, making battery packs of different kinds. And the people that had that role uh were not uh chemical engineers who were into math and studied uh uh differential equations. They're they're electrical engineers that became battery pack people. So it's kind of uh interesting because that's a lot of like uh battery pack work and electronics work and things like that.

SPEAKER_04 15:35

Yeah, and and so at the time uh we wanted to optimize the charging. Um, the problem with all the nickel chemistries near the top end of charge, they generate a lot of oxygen, and so the efficiency goes way down. But by then, now we're talking year 2000, lithium ion is starting to get to beat nickel metalhydride. Uh you know, now we're getting to two amp hour 18650 cells on that, and so now it becomes it behooves us that hey, we gotta pay attention to these cells, we gotta figure out a way to make these uh into safe batteries.

First Li-ion in Space

SPEAKER_01 16:15

Do you know what the first application of lithium ion was within NASA?

SPEAKER_04 16:20

In space, in a crewed mission, yes. Um uh it uh it was with the Canon Camcorder battery. The basically our astronauts wanted to use the latest and greatest stuff, and we were the ones to say, no, you can't use this because it's just not safe. It could flame up, you know, particularly when you're doing in a space suit. You got 100% oxygen in the suit, you're you know, you're dealing in uh in an airlock with lower pressure, so um we have to pay really good attention to make sure there are no ignition sources and uh well eliminate the uh possibility of fires. Yeah, right. And and so and like I mentioned earlier, we wanted to be two-fault tolerant um in our battery systems. And what is that? That's the desire.

SPEAKER_03 17:14

What is that what does that mean? Two-fault tolerant.

SPEAKER_04 17:16

That means uh you can impose any two failures and it's not going to be catastrophic.

SPEAKER_03 17:22

Like at the same two two bad things could happen to this battery at the same time.

SPEAKER_04 17:25

Two independent bad things can happen. I see. And you've got to uh gracefully survive uh survive. Not it it doesn't have to work, but it doesn't you can't have a catastrophic fire, catastrophic event uh on that. And so with lithium ion, well, we're zero fault tolerant to an internal short. You only got one layer separator, right? You you can't design uh lithium ion cells with three separators in order to be too fault tolerant, they won't work. Yeah, right. So at the time we started, okay, can we screen out defects in cells because it was too new, you know, we're talking early 2000. The thought of designing uh passively propagation resistant packs was just something, uh, there's no way we can do that. And it wasn't until a decade later that we realized that yes, we could do it. And then that became one of my you know fields of expertise.

SPEAKER_01 18:28

Yeah, so the the first lithium-on batteries to go to space with NASA were actually part of a camcorder, 18650s, consumer electronics.

SPEAKER_04 18:39

Right.

SPEAKER_03 18:40

Okay.

SPEAKER_04 18:40

Yeah.

SPEAKER_03 18:41

Yeah, you said you can't take it, it's too dangerous. So how did they end up taking it?

SPEAKER_04 18:45

Yeah, so uh we we went through a whole bunch of testing uh with the camcorder. And so after about a year's worth of testing and being able to control the lot size and test the lot, make sure that the quantity that we have bought are all from the same uh similar date codes so that we could qualify the design and not worry that the manufacturer has changed that design.

SPEAKER_00 19:14

I see.

SPEAKER_04 19:15

In the commercial electronics field, you don't have configuration control. Uh every six months, every year, they change the design, they don't change the model number, and yeah, all of a sudden anodes have all sorts of silicon in them. Yeah. Did you get separators are much thinner?

unknown 19:32

Yeah.

SPEAKER_03 19:33

Did you get information directly from the cell makers? Like, I want to know that these are all from the same cathode coding on.

SPEAKER_04 19:40

Oh, no, no, you we couldn't get that. Oh, best we could do is control the date code. Oh, okay. Okay, got it.

SPEAKER_01 19:46

So just look at the wrapper basically on the look at the wrapper on the cell. And look at the serial number and say, okay.

SPEAKER_04 19:56

When they're making 30,000 cells a day, and all you want is a thousand or two thousand or ten thousand, yeah, yeah, yeah. NASA support is good in in a half a day's work.

SPEAKER_01 20:09

But in practice, how how does that work? Do you just go and you buy a hundred camcorders?

SPEAKER_04 20:14

Oh, yeah. At the time you would buy the camcorders with a lot of batteries, and then we would uh tear apart a few of the batteries and look at the date codes of the cell, uh, and then be able to uh realize that they were similar, they were within the same consecutive days or consecutive weeks, and that would give us some reassurance. And then, of course, the teardown of the cells uh was really the the best proof that they were high quality cells, that we weren't getting counterfeit type of cells. We we really got to learn to be good at uh doing uh DPAs, and at the time, then we CT scanning wasn't really uh it was too novel. So we would do cross-sections of cells. Wow. Oh yeah, um exponent was really good at doing cross-sections of cells.

SPEAKER_01 21:10

So when you say that, do you mean potting the cell in epoxy, cutting it, and then looking at the cross-section?

SPEAKER_04 21:17

Yeah, so first they would take a cell, puncture a hole in it, put it into a centrifuge, get the electrolyte out of the cell, and then they would be able to backfill with an epoxy that would then solidify everything, and then they get the s the band saw in and to cut it, and then do lots of polishing and lots of polishing, and then taking immediate pictures after the polishing because corrosion would start setting in hours in. So your cross section would only be good for a few hours on that. But it gave you beautiful color and beautiful morphology of the uh uh of the electrodes.

SPEAKER_01 22:00

You want to do that to convince yourself, okay, the separator is this thickness, or you know, there's no winding defects, or like what what are you looking for in that cross-section?

SPEAKER_04 22:10

Yeah, yeah. So we are looking for uh absence of defects, absence of uh um metallic inclusions inside the electrodes. Um it's really hard to tell if there's a a tear or a hole in the separator when you're doing a cross-section, but you can definitely see if a bright spot in a in an anode that shouldn't be there. A little more difficult to see it in the cathode, but uh you would also you know you'd look for weld splatter um at the tabs uh on that, or any type of occlusions in the insulators at the top and the bottom of the jelly roll.

SPEAKER_03 22:51

Wow, yeah, and you saw those sometimes?

SPEAKER_04 22:54

Yes, mm-hmm. Yeah, we have we have rejected lots from top-tier manufacturers. Wow. Um, because of uh weld splatter and that interesting, and this is what like what year are we talking about here? This is year you know 2000 to 2010, but then we kept keep keep going on and just perfecting that, and then after that, those are supported with CT scans. Yeah,

Screening Limits and Pack Safety

SPEAKER_04 23:19

so in the year 2000, lithium ion is really starting to be compelling from a specific energy and also from energy density standpoint. Uh, we've got to look at it. Um, the consumer electronic industry, the astronauts are wanting their camcorders. Pretty soon they're gonna want uh tablets and so forth. So we have to get ahead of the game and learn, you know, with the charge control, um, how well um they keep us out of trouble from overcharging the balancing of multiple cells in series. Um, but the real issue was the internal short and being comfortable with that and to be able to look at uh OCV screening. So at the time uh we made a rule that you bought cells, 10,000 cells, you'd put them on a shelf for a year, and then you would do the OCV test, and then you'd be convinced that you've got the cells that have the uh charge retention issues that are due to some type of bridging defects in the cells. And we we took apart those cells and confirmed that yes, they had bridging defects um in there. Just basically bridging defects would be as simple as uh active material from the cathode going through the pores and showing up on the anode side of the separator.

SPEAKER_03 24:51

I just want to explain for any listener, you know, for a long time in the early days of lithiumion, the standard at a factory was to charge up the cells and and wait uh about about a month and to see if the uh open circuit voltage dropped a little bit. More like two weeks. And then about 10 years later, it dropped to two weeks. I don't know what they're at now, um, but um even at a month, it it was known that there's what we call escape, which is what Eric is talking about.

SPEAKER_04 25:22

Yeah, yeah. And so then it wasn't until Brian Barnett at TIAX started uh implanting nickel particles in just the right spot in 18650 cells, basically right at the edge of the cathode near the aluminum current collector, and he was able to show the latency of these defects. In other words, they would pass our screening, even our best screening that we would do, wow, and cycle them, and within 25, 50, 100 cycles, they'd go into thermal runaway. Wow. Right. So you're like, okay, even our one-year test is not. Exactly. And and then it told us that uh we could not rely on just screening. We had to assume that our screening was going to fail, and that we had to be passively propagation resistant in our battery packs.

SPEAKER_01 26:23

One of the topics of this podcast is exploring the 80s, 90s, and 2000s, the battery landscape in North America, how battery technology was developed, how there were commercialization attempts. And our previous guests were all sort of on the production side, if you will, people who were trying to make cells. And you're on the, in a sense, on the consumer side, right? You're you're sourcing cells. So by the early 2000s, there's there's players in Asia and in North America. In North America, it's more specialty makers like Yardney, Saft, Eagle Pitcher. Um, and in Asia, you have more consumer electronics producers, you have Sony, GSUasa, Panasonic, right.

First Li-ion NASA Pack - Moli

SPEAKER_01 27:09

So NASA has this challenge of we have to source lithium-ion cells. And I'm really curious about those initial moments of where do we get cells? Who do we go to?

SPEAKER_04 27:22

Yeah, so we we wanted to go to the Asian manufacturers, mainly Japan in and South Korea. Um, the what we've felt the top-tier manufacturers, so Panasonic, Sony, uh, LG, and Samsung, and MOLI in Taiwan.

SPEAKER_01 27:42

Right. And we actually did an episode with with Jeff Don. So we sort of heard the history of Molly and how it went through bankruptcy, became Taiwanese owned. And so in sourcing cells with MOLLE, did you interact with the Canadian center?

SPEAKER_04 28:00

Yeah, the timing was just perfect because we realized the um ICR 18650 J cell from MOLI Energy, 2.4 amp hours. At the time, it was leading in capacity-specific energy. And there they had Canadian well, uh engineers that would you'd call up on the phone and they spoke English and you you could ask uh detailed questions. We would never get that from an LG or a Samsung at the time, or Sony, or Panasonic. So um we went and did testing on their cell, you know, and we looked at this was a way of producing our spacesuit battery. Uh at the time we had 80 cells, so it was uh 16 P5s, and it would fit in the volume of the silver zinc battery and become our rechargeable battery, long life battery for the space station base spacesuit. That was the first critical application for lithium ion.

SPEAKER_01 29:05

Um, was that the first pack that you guys made with lithium ion? Because, like, there were you know, there were conservative products that went up, but the first pack that you guys made.

SPEAKER_04 29:15

Okay, that was the first custom pack for a critical application, it had to work. So we designed the simplest battery we could. We put all the smarts in the charger, all the active components, because we really needed to have it work for seven hours, the length of a spacewalk. And so we just had a a TC and we had some poly uh switches, and we had a charge connector and a discharge connector, and then the battery had a fuse, and so very simple for our astronauts.

SPEAKER_03 29:49

Did you have a battery management system, a chip that would do things like make sure the cells are at the right voltage? Did you have a TI chip in there or something like that?

SPEAKER_04 29:57

Not in the battery. Interesting. Because the suit was designed for a silver zinc battery, it was designed for a two-terminal battery, silver zinc. So there's nothing that the suit could do about it. So all we got for during the spacewalk was battery voltage, current, temperature.

SPEAKER_03 30:15

Okay.

SPEAKER_04 30:16

But when it came back in, and then we uh connected it to the charger to recharge the battery, then we were controlling the five cell banks to the precise voltage. We

Flying EVA Packs and Charge Limits

SPEAKER_04 30:28

did some 25 spacewalks with that battery system, never had the voltage go out of balance by more than 20 millivolts between the five cell banks. Oh, that's great. But we managed the state of charge really well. They would never charge the battery more than two weeks prior to the spacewalk. And if the spacewalk ended early, they would always discharge the battery within two weeks to get it back to 30% state of charge. Nice. Yeah, interesting.

SPEAKER_01 30:59

Do you remember the chemistry of that cell? So you you know the first cell to go up in a package LCO, yeah, lithium cobalt oxide on there. So an LCO graphite cell.

SPEAKER_04 31:10

And that's the cell, that cell has the longest maturity. I think it lasted for was in production for 20 years. Yeah. The wow, this 2.4 amp hour cell. Molly kept it going for quite a while.

SPEAKER_01 31:24

So just before we move on from that cell, I'm just curious. I assume the cell was balanced to 4.2 volts, like the spec sheet was a 4.2 volts cutoff.

SPEAKER_04 31:33

We we went to 4.1. So just for yeah, so that's that's always keeping margin. Yeah, we only went to 4.1.

SPEAKER_01 31:42

All right, so you backed off.

SPEAKER_04 31:43

We would cut off at uh 3.2 volts, and that was mainly because the suit couldn't handle uh the voltage and voltage window at bottom was 16 volts, and so yeah, but cobalt oxide is pretty funny. But cobalt offside is higher, so it doesn't matter. It didn't matter, yeah. Yep, yeah, right.

SPEAKER_01 32:01

Okay, so yeah, so that 4.1 volt cutoff gave you a little bit of of margin of error for the cell safety.

SPEAKER_04 32:08

Yeah, so this is now we're flying this, we're doing it in spacewalks. I think 2010 was the very first spacewalk, and at the same time, we were looking at going lithium ion for the space station.

SPEAKER_03 32:21

Much bigger battery.

Abuse Testing and 787 Wake Up Call

SPEAKER_03 32:23

I just want to ask on that on that pack, did you do like abuse tests on the pack? Oh, yeah. Try to say, like, what would are there any stories from that? Because it seems like it could be uh the an interesting experience because often the abuse leads to fires and all that sort of stuff.

SPEAKER_04 32:39

Well, we we did at the cell level, we did the abuse test, and we knew that we could not tolerate overcharge at the bank level. We did do some testing where we ended up doing a smart short at a 16 cells in parallel, and it worked out okay. So every failure mode was credible failure modes that we didn't do like nail penetration because we had astronauts using this, they weren't going to abuse the battery. Yeah, well no, but we were still haunted by the spontaneous thermal runway due to an internal short. Right. And we can't at the time, 2008, 2009, we figured we did good screening, we could take credit for that. And it wasn't until the 787 in 2013 uh incidents came that we finally got the attention uh it deserved so that we could get our management to look, hey, let's look at redesigning packs to be passively propagation resistant.

SPEAKER_01 33:52

Yeah, so let's talk a little bit about that Boeing incident because both Kevin and I were at 3M and definitely Tob Brass of 3M talked about that incident. You know, it really shook the the battery world, I think, for sure.

SPEAKER_03 34:05

Um that and the Samsung Galaxy Note 7, of course, where every passenger would be notified again and again again to be worried about their Galaxy Note 7.

SPEAKER_01 34:14

Yeah, so let's talk a little bit about that Boeing incident.

ISS Cell Specs

SPEAKER_01 34:17

So the cells were GSUASA cells, they were big prismatic cells.

SPEAKER_04 34:21

Um 75 amp hours.

SPEAKER_01 34:24

75 amp hours. Um and from what I could read up on it, it was a prismatic cell, you know, it's a rectangular metal can. Um, and there were three jelly rolls in this big can, and it was an LCO graphite chemistry. Interestingly, what I could find online was that it had a 4.025 upper cutoff voltage, which is quite conservative. Maybe you can give a bit more color about that Boeing incidental learnings that you you got from it.

SPEAKER_04 34:51

So stepping back to 2010, the nickel hydrogen cells were getting old. Um, a space station, we needed to have a long-term plan uh and move to lithium ion. And so we went out and surveyed five different cell manufacturers for who could make a safe cell for a space station battery. Now we're talking about you know 75 kilowatt hour type of battery, uh, much, much bigger battery system divided into four segments, but far away from the crew. Um, and we wanted to use the existing box that the nickel hydrogen cells were in and just populate that with Lifumat.

SPEAKER_03 35:36

I just want to let any listener know that's about the size of like a of a long-range electric vehicle, 75 kilowatt hour.

SPEAKER_04 35:43

Yeah. Yeah. And so the big requirement we had was 60,000 cycles over 10 years. 60,000 Leo cycles. These are 90 minute cycles. So 60 minutes of charging, 30 minutes of discharge. Never stops. Wow.

SPEAKER_01 36:04

And what's your SOC swing on that cycle? Yeah. How much does your state of charge change on your on your Leo cycle?

SPEAKER_04 36:11

I think we started at 3.95 volts top of charge, and we went to 20% depth. Ah so very shallow uh depth of discharge. And then we would reserve being able to bump up that top of charge as the battery uh degraded to get the extra capacity. That was the strategy. Okay. So we we looked at the the what we called the boutique manufacturers, the yardneys, the uh eagle pitchers. Um and we also looked at GSUasa because and they were also what I call a boutique manufacturer. By boutique, I mean they're talking for them to make a production run of a thousand cells, it's a big deal. Uh, it takes weeks to do that. And we also compared it to Sony's commercial cell uh at the time, they made a hard carbon cell um on the LCO, uh lithium cobalt oxide, and that had some good cycle life, but we didn't have enough data on it. Um, and so we ended up going with GSUasa, even though that would be a 134 amp hour GS USA cell in a wow uh jelly roll that is in a uh elliptical cylindrical format, an aluminum can that was a floating can with two terminals, um a very challenging cell to make, and um but we did cell production line audits of every manufacturer. We did CT scans and DPAs of these cells to look for who had the best control of minimizing defects uh in those cells. It was a great learning experience uh about the uh cell manufacturing, um, particularly with the high volume manufacturers. I was very impressed uh back then. And at the time, you know, Sony was uh well, I guess they they were kind of one of the leaders, but they've been surpassed um by the uh the other cell manufacturers.

SPEAKER_01 38:30

Uh yeah,

Auditing Cell Factories

SPEAKER_01 38:31

I mean I would love to hear a little bit more about you know your experience of auditing cell lines.

SPEAKER_04 38:36

Yeah, yeah. What year is this again? So this is 2010. Um I I I hired Exponent, uh Selena Mikola Whitechik and Troy Hayes, and they led the production line audits um at the uh five different cell manufacturers. So we went to two trips to uh Japan, one at Sony and one at uh GS Uasa. Uh and then we uh went and looked at uh Saft in uh Kakusville and uh Yardney in uh in Connecticut and Qualion in north of LA. Yeah.

SPEAKER_01 39:14

Yeah, so that's fascinating. You had the the perspective of seeing of auditing the cell manufacturing, three cell manufacturers were in North America, two were in Japan. Um how how did they compare? Like what what was your experience?

SPEAKER_04 39:28

Yeah, I have to be guarded on here, but because all this the stuff is the results are all confidential and sensitive, but it really gave us some good insight, and the consistency of GSUASA and defect-free DPA results uh led us to that. But the cycle life data uh was much more mature at the time uh with GSU Awesome, and so that's how we ended up picking them.

SPEAKER_03 39:58

Well, let me just say one thing, then you don't have to point to anyone, but you know, I've done some auditing at cell factories, and um I just this well, there was an incident I saw where I'm getting a tour with a line manager and then a marketing person who marketing person can't be expected to understand what's going on, and we're looking through windows at different parts of the cell production, and I'm at the point where electrolyte is being filled, and electrolyte is sensitive to air, so they've got they're doing it through a glove box. And uh, while I'm there watching, one worker, I can see that they're they have open cells and they're filling. One worker opens, simultaneously opens both sides of the airlock, which you should never do, it should be one side and evacuate, and then the other side. Simultaneously opens both sides, they move some finished cells out, close it back up, and they continue working. And I pointed out to the uh line manager who's with me, and I say, Oh, did you just see what happened? He said, Oh, yeah, I did. And I said, What do you think of that? And he didn't he didn't even know that it was a problem. Oh, and he's the line manager, yeah. So, so there can be situations like that.

SPEAKER_04 41:02

Maybe you never saw anything quite like that, but no, no, I was uh overall pretty darn impressed, particularly with uh GSU WASA and being able to make such a large cell with such high surface area, um, and us us being able to unwind that cell in a fume hood. I mean, you had to unwind the two electrodes and reel them into two separate rollers, because otherwise it would just be a mess. So it was a very challenging DPA uh to do. Much easier to do a DPA and inspect quality with small cells. All right, so that's 2010.

Dreamliner Wake Up Call

SPEAKER_04 41:41

Then 2013, GSU ASA has the big problem with uh 787, the Dreamliner. Right. And the Dreamliner cells are made in the same plant uh production line that the space station cells are made in. Oh, but it's a totally different design, uh, we hear from uh our prime um on this. And so, no, we so we get pretty alarmed by this, and so I was uh able to convince our NASA Engineering Safety Center, kind of our independent watchdog group that collects uh the experts from the agency, both in the government and in industry, to solve technical problems. Uh, about well, can we build passively propagation resistant battery packs? And I have to give credit to SpaceX um they having started that with their uh batteries for the Dragon vehicle with 18650s. So I knew that it was already possible. The question is, could we do it with not too much of a mass and volume penalty? Yeah.

SPEAKER_01 42:54

So the the passive propagation resistance, I just want to come back to that for a second. Um because what was happening in the in the Boeing incident was one cell failed. Now, the reason it failed presumably was an internal short. Was that ever a hundred percent determined what was causing the GSUasa cells to fail in the Boeing incidents?

SPEAKER_04 43:18

No, but I I think the uh National Transportation Safety Board, the NTSB made it the most likely root cause.

SPEAKER_03 43:26

Okay, I got it. After all the reports came back from places like Explonate, they they that's what they decided.

SPEAKER_04 43:33

So then the challenge was you know being able to build a pack that wasn't so heavy and so massive to protect the adjacent cells.

SPEAKER_01 43:43

Right. Because what was happening in the Boeing situation was that one cell would fail, it would go into thermal runaway, heat up, and then cause the adjacent cell to go into thermal runaway.

SPEAKER_04 43:54

Yeah, there's just one layer of capton between the two cells.

SPEAKER_01 43:57

And then the next one would go, and the next one would go, and then all of a And you have this huge fire on your hands.

SPEAKER_04 44:02

Exactly. Yeah. Yeah. And that's catastrophic. Yeah. And even one cell going off at 75 amp hours is catastrophic. Even uh not inside the crew cabin, in the uh unpressurized volume of a capsule, uh 75 amp hour cell, I don't think we could tolerate that kind of failure because of the damage it would do to the other systems nearby.

Focus on Small Cells

SPEAKER_01 44:25

So did that cause you to reconsider the balance between cell size and pack size?

SPEAKER_04 44:32

Yes. Oh yeah. So this is why I uh from early on became a small cell guy. Yeah. Right. Because you uh it's a little firecracker going off as opposed to just uh you know uh a huge catastrophic event. Right. And being able to protect the adjacent cells is just so much more feasible. And so then the the challenge after that was we found that uh the volume for the spacesuit, for example, was what drove uh the design. So we need to squeeze the cells really close together, half a millimeter apart from each other. Ooh, that's close, yeah. On that and the only way to protect cells that are that way is to put an interstitial aluminum heat sink in the cells. And when you do that, then how do you drive a cell into thermal runaway? Your trigger cell to prove you're protecting the adjacent cells. Well, uh it's really hard, you don't have any surface area to drive any heat to put a heater on, and most likely the heat is gonna go into the heat sink and then heat up all the other adjacent cells. Yep. Or you could try to do a nail, but then you've gotta uh get through the battery structure to do that.

SPEAKER_01 45:50

Right, you might compromise your

Internal Short Trigger

SPEAKER_01 45:52

pack.

SPEAKER_04 45:53

So I was fortunate at the time um in 2010, I had my second sabbatical at the National Renewable Energy Labs. And so I worked with their battery group with Matt Kaiser and Dirk Long. Uh, we came up with the internal short circuit device, and so I brought the idea of using wax layer, a paraffin type of wax, and then but that wax was brittle, it was not flexible, and so Matt introduced a uh flexible well, women's hairspray wax, and mixed that with the paraffin wax to make that wax flexible. So now our little device, so it'd be compatible with a wound cell because with spin coating that wax layer onto the aluminum foil, it'd be on the order of 10 to 20 microns thick, and could be the insulator, and we designed the wax to melt at 57 degrees Celsius. So you just had to heat the cell above 57. The winding tension would then squeeze the wax, the molten wax, out of the way, and you'd have a hard short.

SPEAKER_01 47:07

Right. So just for the for the listener, the device looks like a you know a little a little disc, and so the the sandwich ends up being Yeah.

SPEAKER_04 47:17

So we have uh two an aluminum disc and a copper disc. The copper goes against the anode, the aluminum goes for the cathode, and in between that we have a separator with a hole in it, and inside that hole there's a copper puck, as we call it, but it's the same thickness foil as the copper foil. And the only thing that's insulating the copper puck from the aluminum pad or aluminum foil is the wax layer, the flexible wax layer. And so you give that device to a cell manufacturer, and he can introduce it into his dry jelly rolls. Or better yet, during the winding process, he can stop the winder and then uh make a hole in the separator. Uh, and then the best way to get what we learned is the most vicious short is when you go from the graphite anode to the aluminum current collector on the positive side, bypassing the cathode active material. So you have to remove the cathode active material to expose the aluminum, seat the device against the aluminum current collector, and then you would get uh a 95%, 97% chance of getting a hard internal short that would lead to thermal runaway.

SPEAKER_01 48:38

So you would clean a little spot on the cathode?

SPEAKER_04 48:41

Yeah.

unknown 48:42

Yeah.

SPEAKER_01 48:43

So remove the active material, cut a hole in the separator, and then place that little device.

SPEAKER_04 48:49

The manufacturer, high volume manufacturers couldn't stop their winders, so they would have to take the jelly rolls off the production line, manually unwind them, which is why they never really wanted to go more than about three winds in, implant the device, and then manually rewind the jelly roll. And the skill of the uh technician in doing that uh was directly proportional to how many formation failures they would have when they do the first charge of that jelly roll. Yeah, right. But we got consistent enough after three, four years, and I have to give lots of credit to Molly, uh, particularly Mark Shoesmith, for being the only manufacturer that was willing to implant to be the first to implant the device in their cells.

SPEAKER_01 49:41

Yeah, Mark Shoesmith is still at Molly. I was chatting with him. Yes, the at the battery seminar in in Florida. Fantastic.

SPEAKER_04 49:48

Yeah, so now we had a trigger cell that we could use to verify our batteries were passively propagation resistant with that. Right. Because you could use a small heater at the bottom of the cell was just enough, and you only needed a few watts to get the cell above 60 degrees and to trigger thermal runaway without dumping that heat into uh the heatsink uh or and the adjacent cells.

unknown 50:20

Right.

SPEAKER_01 50:20

The pack can remain at a temperature that is still representative of the application.

SPEAKER_04 50:26

And the other thing is, is we are able with the device to test all four different types of internal shorts. So we could have an active material to active material short where there's just a defect in the separator. We could have a collector to collector short by removing the active material on both sides, or you could do the the hybrid of both sides, and there you could then characterize the vulnerability of a cell design to an internal short. And what we found with the that MOLI 2.4 amp hour cell, it was only vulnerable to the anode graphite short to the aluminum current collector. You had to bypass the cathode active material, and that and that's really valuable because now that means you need a defect that is bigger than the coating of the active material on the cathode to cause a bridging problem. And so production lines could then you know what they're looking for, what they need to filter out uh in terms of defects.

SPEAKER_03 51:32

So I'm curious if I go back to what where we started, you learned a lot about simulations and this kind of question of the four types of defects that you could have, the four types of shorts you could have, very simulatable sort of a question. So I'm wondering, did you did you have a project there at NASA at the same time to say, does this make sense? We can simulate it in these four types of shorts and to see if the data come out similar to what we're what we predict by simulation or or not?

SPEAKER_04 51:58

Yeah, uh Guillaume Kim at uh at the time the National Renewable Energy Labs, which is now the National Lab of the Rockies, did that simulation um on that to to help us with the design of the device? Because how did we come up with that geometry of the device? Um, yeah, so we uh he he did the simulation, but it was more experimentally led uh in order to get something that works, and then the modeling came back and helped guide the final configuration.

SPEAKER_01 52:33

So, can you explain why that's the modality that leads to the catastrophic failure?

SPEAKER_04 52:40

Yeah, why is that the worst? Yeah, yeah, it's it's really interesting. Um, because the collector, the collector short is such low impedance short, it tends to burn out um on that. Well, in fact, when we did the test with the the Molly J cell, the uh trilayer separator that they were using in that cell uh would shut down and safe to sell, but it would not shut down for the anode to aluminum current collector short. And it's just because it just had the right sweet spot of impedance, something that may have matched the maximum power capability of that cell to lead to the most vicious short, a more sustained vicious short. And it was interesting, we did the active to active material short and only got soft short on that. So, in other words, a hole in the separator uh isn't what's going to drive that cell into thermal runaway. Now, this is all true with a 2.4 amp hour cell. When you get to today's nearly 4 amp hour 18650s, um, we haven't done the tests, but I would bet that the active the active short would be much more vicious.

SPEAKER_01 53:59

Yeah, that's fascinating.

Chemistry And Ruptures

SPEAKER_01 54:01

So one thing that I was wondering about was, you know, the only chemistry we've mentioned up to now is LCO graphite. And by 2013, you know, LFP exists, uh, NMC exists, NCA.

SPEAKER_04 54:16

NCA, yeah.

SPEAKER_01 54:17

Yeah. How how do you go about choosing your chemistry? Like what informs your choice of chemistry? Because your set of requirements are pretty unique. Like you, I assume, have no cost constraint, essentially. Um, and your highest priority is safety and then energy density, right, and performance. Um, so how how does that inform your your choice of chemistry?

SPEAKER_04 54:41

Yeah, so the 2.4 amp hour cell never improved or uh and got bypassed by uh NCA chemistries from Panasonic when they came out with a 2.9 and a 3.1 uh and then a 3.35 uh amp hour NCA cell. And then after that, uh NMC cells came along uh and uh the LG MJ1 cell, the 3.5 amp hour cell. And so we started testing both the Panasonic and the uh LG cell, and we discovered sidewall rupturing, really vicious failure modes that we didn't see before. And basically, if you stayed under 2.8 amp hours per cell, you would not get the vicious sidewall rupture. Uh and get the bang, the the really kinetic type of uh thermal runway. So that introduced a whole new challenge in being able to protect adjacent cells. But the the challenge then, if you know now we need to we need to introduce the device into those bigger, you know, higher capacity cells.

SPEAKER_03 55:52

Yeah, these cylinders normally would look kind of like a road flare where the there'd be flame jetting out the the circular end at the top, more or less. But at some point when the cells got more and more energy packed into them, the whole round side edge would would give and it would be like a grenade sound, like a bam, because all the pressure would be released instantaneously and shrapnel would fly everywhere.

SPEAKER_04 56:14

Exactly. Very good. Yeah, yeah, Kevin. You and so instead of instead of venting out the top and sometimes having a bottom rupture, uh the cylindrical wall would fail, and then that's a blowtorch for adjacent cells. Instant thermal runaway propagation occurs when that happens. So at the time, that's when Donald Finnegan at the University College of London contacted Matt Kaiser and myself because he was uh doing synchrotron uh experiments with really high-speed X-rays and wanted to be able to be the first to do tomography on a cell that was going into the thermal runaway. But he was very intrigued that we had an internal short device that could be the instigator thermal runaway as opposed to driving a nail. So we managed to get him six cells, the uh uh the MOLLE cells, and he took some beautiful images at uh 2,000 frames per second of an internal short developing inside a cell. I'd love to share it with you and to see it. And once I presented that at conferences, then LG and Samsung and the other manufacturers uh were interested, particularly LG, and were willing to implant it in their MJ1 cell. So I got invited to a battery conference in Seoul and made the contacts there. But basically, we were able to show them high-speed radiography, data that they never had on their cell, and they were able to learn with us. So to counterbalance the fact that I would only buy very small quantities of their cells, they were willing to put up with me and listen to me because of the unique data I was able to present them. And so we got trigger cells of DMJ1, and then later we improved it with the M36 cell that had a thicker can wall.

SPEAKER_03 58:23

Oh, yeah, that's what I was kind of wondering: like, okay, so what did they do to stop the sidewall rupture? It sounds like they did thicker can wall, or were there other things, or was that yeah?

SPEAKER_04 58:30

The M36 cell is the best cell for against sidewall rupture because it's a 250 micron can wall as opposed to 160 that the the other manufacturers are using. Okay.

SPEAKER_03 58:43

When you make a cylindrical can like that, are there different process choices? Some that lead to kind of line defects uh that where you could have a rupture and others that don't.

SPEAKER_04 58:54

Yeah, the design of the spin groove is very important. For those who don't know, the header of the cell has to be crimped. And in order to do that, there is a neck in the cell that what I call the spin groove, um, and then it folds over at the top in order to seal the cell in a gasket uh seal. It's a very good seal. It's we've had 20 years uh plus uh in space Leo cycling cells, cylindrical 18650s uh working, so it's a very good seal, but that spin groove um is one that's vulnerable, particularly for an internal short that's in the outer winds of the jelly roll, because the fluidization of the electrodes is lined up with that spin groove and it ruptures that, and then that ends up being a a blowtorch path of a high momentum thermal runway ejected towards adjacent cells.

SPEAKER_03 59:58

I see.

SPEAKER_04 59:59

Yeah.

SPEAKER_01 1:00:00

So this internal short circuit device allowed you to identify and understand all of these failure modes, and then you had these collaborations where you could image them in high speed and high detail.

Designing PPR Packs

SPEAKER_01 1:00:12

So, how did you end up modifying your packs to make them propagation resistant?

SPEAKER_04 1:00:17

Very good question. So, first we needed to know the amount of thermal runaway energy uh for each cell design. And unfortunately, it's not uh linear uh with the capacity or the energy, uh the electrical energy of the cell. Certain electrochemistries are just more violent than others. So we needed to develop a calorimeter, what we call our fractional thermal runaway calorimeter, and so we developed that in about 2016, um, and we used that started in 2016 and then perfected it, and then we used that calorimeter in order to test the cells at the synchrotron and to do it safely. Nowadays, when we go to the synchrotron, we can test uh over 130 cells in a 96-hour window uh at a synchrotron. So we're able to test dozens runs of a cell design in a certain condition, so we can really characterize the stochastic nature of the response.

SPEAKER_03 1:01:24

And so then you're able to say, okay, these chemistries have this much energy, these chemistries have this much energy. Now you know how much energy there is in the thermal runaways. Do you then do you say, like, well, we want to use these cells, we don't want to use those, or do you say, okay, this is how much we've got to deal with, let's figure out how to engineer the pack. How did how did you do that?

SPEAKER_04 1:01:42

A little bit of both. Yeah, the really high silicon percentages are super violent. So we want to back off from that. Um, the question, I mean, I give you an example. Uh, we recently showed that those results. Uh, Rin cells 4.1 amp hour cell uh would give us about 30% more thermal runaway thermal heat for only about 10% more extra electrical energy. So the trade-off is not good there on that. So that's that's the engineering decisions that we would make. Let's back off uh on this yeah, the violence of thermal runaway, but still get the highest performance we can get. So back off to more like the 7% silicon doping that's in these anodes. We also studied the bottom vents. We were fortunate enough to have a manufacturer make cells without the bottom vent and with the bottom vent. And uh we found that the cells were less violent if they had a bottom vent.

SPEAKER_03 1:02:46

Get rid of more of the gas before it burns.

SPEAKER_04 1:02:49

Yeah, what happens? It it the uh thermal runaway escape of the gases would happen at a lower pressure because it'd blow out both ends, and the post-test mass was actually higher on those cells when you have two vents than when you only have one. In other words, the residence time of the thermal runaway reactants was shorter with the two vents on that. Yeah, interesting. But unfortunately, the the manufacturer uh didn't want to make the cell because his main customers couldn't accept the bottom vent.

SPEAKER_01 1:03:20

So does the pack end up having material between the cells to prevent propagation?

SPEAKER_04 1:03:25

Or yes, yeah. Uh bread and butter most volumetrically efficient design is to use the interstitial aluminum. Um, it's the one we have the most maturity on, but it's heavy. We've been looking at using uh interstitial foams. Did you ever use air gap? Yeah, and and that's unsuccessful because then there's no protection against sidewall rupture. Sidewall rupture. There you go. Exactly.

Five PPR Rules

SPEAKER_04 1:03:50

Yeah, so so then uh one of my interns, Alexander Quinn at uh MIT, wrote up this paper along with a lot of my uh history over the last nine years or of developing PPR packs, and we developed five rules, five design guidelines for battery achieving PPR. And rule number one is to control sidewall rupture, protect adjacent cells against sidewall rupture. So you need to characterize. So we test uh 200, two, 300 cells to get defendable statistics on that. Then the next rule is heat rejection. You can't just isolate, thermally isolate cells with an air gap, like you mentioned, because then the heat is eventually going to only hit the most adjacent cells, and it's gonna be a slow dissipation. And you'll be on razor's edge for tens of minutes to find out whether you pass or not. So you need to have a heat path for the troubled cell to get away, get that heat away from the adjacent cell. And then rule number three, you need to electrically isolate the troubled cell from the parallel cells, the cells that's that are in parallel with it, because those parallel cells are now feeding an internal short in that troubled cell. And so they're getting hot, feeding that short while they're exposed to a cell that's in thermal runaway, and that reduces your margins. And then the fourth rule is probably the most challenging rule, is a high temperature path for the thermal runaway ejector. From our calorimetry, we know that over 80% of the heat is ejected from the cell, particularly with the high-energy cells nowadays, it's nearly 90% of the heat. So you've got to protect the adjacent cells, the tops and the bottoms of the adjacent cells, from that raining ejector. And you want that ejector to smear out and not agglomerate next to the adjacent cells. And have a have a blast plate that doesn't perforate. So that the ejecta hits that blast plate and doesn't perforate to hit the uh other cells that are on the other side.

SPEAKER_03 1:06:07

I at first I would have thought you'd want like almost like a ceramic insulator between the cells, but no, you need a heat path. So that's why you have metal. But probably at that top area, you probably do have some kind of a ceramic.

SPEAKER_04 1:06:17

Yeah, the bottoms of the of the cells can be a good place to make some contact through some gap pad materials for the electrical isolation.

SPEAKER_03 1:06:27

Have you seen packs made for electrical energy storage today? Like battery energy storage for power plants has taken off very recently, and it's predicted within I think it's like two or three more years will surpass electric vehicles in demands for batteries, something like that. It's unbelievable. Um but so they care a lot about this, and there are major incidents happening, unfortunately, like practically every month in the world at power plants. Have you know, have you ever have you met with people working on those and designing those and shared your theories? Because it seems like they could learn a lot from all that hard work you did.

SPEAKER_04 1:07:03

Unfortunately, the answer is no. Uh they're typically using LFP, heavier chemistry that I'm can't fly because of the weight penalty. Um on that. But yeah, they're they could definitely learn from some of the uh lessons. Yeah, the hard knocks that uh all the PPR failures that we've had, and boy, you learn a lot from those failures.

Current NASA Chemistries

SPEAKER_01 1:07:28

Yeah, you just mentioned that LFP wasn't a viable choice because of the you know, the watt hours per kg essentially. What is the chemistry that is currently flown today?

SPEAKER_04 1:07:38

Yeah, so right now it's LCO um is in the spacesuit, LCO is in the space station, um uh LCO is also in the pistol grip tool. Uh it's also in the uh emergency jet pack uh for EVA rescue of the astronaut. Now, Artemis 2 just flew with uh NMC, it flew with the LG MJ1 cells, and those are cells that I bought 10 years ago, and so Artemis II landed with 10-year-old cells.

SPEAKER_03 1:08:15

And you kept them at like 30% state of charge and uh refrigerated or an air conditioned warehouse for 10 years.

SPEAKER_04 1:08:21

Yes, the batteries were kept that low, yeah. And uh and unfortunately, we got alarmed when we found uh cells from that lot that we had bought uh that were left on the shelf and their charge retention was pretty darn poor. So luckily refrigerator storage uh really helped. So so why the 10 years because of project delays, because uh numerous project delays, and the fact that we uh Artemis III and beyond are using the LG M36 cell with the cell with the thicker can wall, and that cell is being uh tested for calendar life, and we're seeing it track a lot better than the MJ1, at least the MJ lot that I bought 10 years ago.

SPEAKER_01 1:09:12

So that speaks to a sort of an interesting question. If you have to go to the commercial cell manufacturers, the life cycle of a product will be a certain number of years, and for you, the the time frame of a project can be much longer. Yep. So how how do you mitigate that risk that the cell that you want is just no longer available? Is it simply you just buy a warehouse of them and you put them in the fridge?

SPEAKER_04 1:09:40

We buy a 60,000 cells, we qualify that lot, and then we store them in controlled condition, and we do calendar life testing. And as the calendar life data supports extending that life, uh, we're doing spacewalks now with uh 12-year-old Samsung cells, 26F. Um there's you know it's supported by the data. We feel comfortable with that. Yeah, lithium ions is just an awesome chemistry. I mean, you just you you couldn't really do that with NICADs or any uh the nickel chemistries, uh leave them on a shelf like that.

SPEAKER_03 1:10:20

Um yeah, if they don't get warm and you keep them at lower voltage, it's amazing because it's it's exponential with voltage

Plastic Collectors Future

SPEAKER_03 1:10:27

and temperature.

SPEAKER_04 1:10:27

Yeah, so so in the in the last seven years, my focus has been a lot on plastic current collectors, the metalized polymer current collectors. Oh yeah, as a safety device, as a safety feature, um in in order to tolerate localized internal shorts. And what'd you find out? So we started out with 2.1 amp hour cells, 18650s, and we proved the concept that yes, it can tolerate the nail, but and it can do it consistently once you have good materials, but it was 2.1 amp hour 18650s. We went to then 5 amp hour 21700s, and all of a sudden we're overwhelming the plastic, and we're failing uh 70% of the time with nail penetration, and we found out that the separator plays a key role in there. We need a separator that has isotropic mechanical strength, in other words, uh doesn't tear in one direction preferentially compared to the other direction, so that when the nail goes in, it doesn't induce a big tear in that weak direction. I see, which then leads to a wide area short. So as long as we can keep the short into a small point location defect oriented, then the plastic current collector has huge promise to uh be able to isolate that part of the cell that is involved in the short from the rest of the jelly roll by having its plastic current collector vaporize, if you will, and that induces then a delamination of the active material that's involved in the short from the rest of the jelly roll or the rest of the electrode stack. Yeah, well, it's a high impedance path, and so it turns a hard short into a soft short, and it it acts and it acts in in milliseconds. It's just amazing, right? Yeah, yeah. So I think that's a really promising technology, and that's why I've uh working part-time now with uh Satiri, I'm uh the co-CTO uh of that of that organization uh to help promote the plastic current collector.

SPEAKER_01 1:12:51

You know, something

Safety Versus Density

SPEAKER_01 1:12:52

I've often wondered, is safety at the pack level simply uh a question of the total energy density of the pack? Another way of saying this is if I had low energy density cells like LFP cells, I could, you know, just pack them together and get a certain energy density, or I could take some higher energy density cells like LCO cells and add some material between the cells to space them out, and I'd have two packs for the same energy density and therefore the same safety in the end. Um but what you're pointing out is is actually you're sort of searching for all of these changes where you can get a benefit without getting uh a hit on energy density. And you were pointing that out with the the calorimeter where you said if I add this much silicon, that's not worth it. And with the you know, with the plastic current collector, well, maybe I don't really get it a hit on energy density, but I can increase the the safety. Um can you speak a little bit about that? About like the what is the cost of introducing safety in a pack?

SPEAKER_04 1:13:56

Well, we uh evolved from the 18650 and got to design parameters to protect adjacent cells, and then decided, well, let's try moving up to the 21700. And after we did all our testing, we were able to get about a 9% advantage by going to the bigger 21700 cell. So will that follow when we go to the 46 millimeter cell? That's TBD. I I doubt it, but NASA, my group is still uh answering that question. Uh, in other words, the amount of mass and volume you have to add to protect the adjacent cells uh would not overcome the energy and specific energy advantage of the cell, such that you would come up with a net benefit. And like like you said, Vincent, really the 20 or 30 percent silicon anode cells were giving us a 30% more violent, more thermal runaway heat cell, or only giving us a 10% capacity energy improvement. So that was not a good trade.

SPEAKER_01 1:15:08

Have you ever had a a thermal event in a pack deployed in a mission?

SPEAKER_04 1:15:14

No. Thank goodness. No, we've had all our all our failures are been on the ground.

SPEAKER_00 1:15:20

Yeah. That's fantastic. Yes, yes.

SPEAKER_04 1:15:24

Thank the Lord on that one.

SPEAKER_00 1:15:26

Yeah, absolutely. Yeah. I mean, it speaks to all the yeah, exactly.

SPEAKER_01 1:15:30

It speaks to all the hard work that that you guys do and how successful

Domestic Supplier Criteria

SPEAKER_01 1:15:33

you've been. And as we wrap up, I'm thinking, what would a North American cell manufacturer have to show to make NASA consider them as a supplier?

SPEAKER_04 1:15:44

Great question. Great question. If they could show us some insight on configuration control, that would be a huge advantage over the overseas supplier. Um, in other words, give us insight as to you know when are you changing something in that model cell, and so that we can assess that change and make sure it's of uh a no impact or of moderate impact or something that we can live put up with.

SPEAKER_03 1:16:15

You could be read into every change that came up, change in materials, change in equipment, etc.

SPEAKER_04 1:16:21

Yeah, and to be able to be there for production line audits. And if they would be uh willing to welcome that and use that uh and see the benefit of how they can improve their processes. Because when we do a production line audit, uh we bring a magnet and hold it over the coating line, hold it over the the calendaring line, and over the mixing area. Yeah, what kind of iron particles are their processes generating? And are they filtering those particles out successfully through mechanical and magnetic filtering? Yeah, the introduction of the plastic current collector, for example, would be really good with a domestic manufacturer.

SPEAKER_01 1:17:07

And do you think they would have to be at a certain scale so that you can have some confidence that you know their failure rates per million or parts per billion?

SPEAKER_04 1:17:15

Yeah, yeah. It would we would be giving up some of that. Uh, we wouldn't be able to expect them to make a million cells a day, uh, even uh a thousand cell. We would hope to get to a thousand cells a day, and then we'd have to go and show the consistency day to day. But yeah, so the the key is I mean, the uniformity that we get from these high-rate production lines are just phenomenal. The consistency in the capacities, the consistency um in the OCV uh charge retention is just beautiful. So that's why I've I've been a small cell guy all these years. What we do at NASA is we spin in commercial technology into our applications to be successful, as opposed to inventing new electrochemistry.

Closing And Credits

SPEAKER_01 1:18:08

Well, thank you so much, Eric, for giving us this insight into the world of evaluating cells and designing safe packs at NASA and in the aerospace applications.

SPEAKER_03 1:18:20

So many years of trying to make the perfect pack that is as safe as humanly possible with with so little constraints to explore it. You've got to be the world's expert on the on the safe pack at this point. Like what it takes to do it. And uh that's just so valuable. I I hope that your uh expertise gets uh lots of good use yet, and that you have fun sharing it.

SPEAKER_04 1:18:47

Well, thank you. I mean it's it I do have to say I had a very blessed career at NASA where they let me dig in to the problems and and be able to really find out the nature of internal shorts so that we could understand them and then see if, for example, the plastic current collector could be a good defense against that. Um, we're not there yet, but we think now the high-volume manufacturers are ready to adopt it on that, so stay tuned for that. It's batteries, there's just an exciting time to be in batteries, and so thank you for your great questions.

SPEAKER_01 1:19:28

Battery potential is produced by Cyclical, a battery consulting and services company headed by Vincent Chevry and Larry Krauss. Music by Big Flame.