Taking LCO to the Next Level:

Atomic Layer Deposition for Enhanced Performance of Consumer Electronics Materials 

Session Notes:

PRESENTER

Barbara Hughes, Ph.D.

Director of Energy Storage at Forge Nano 

LINKS

ABSTRACT

ALD is a platform technology that has been widely demonstrated throughout the literature to impart significant processing and performance gains on lithium ion battery technologies, having notable impacts on surface stabilization of the positive electrode. At Forge Nano, we have elevated this predominantly research scale technique to production heights, making ALD a viable and affordable technology to meet a variety of materials enhancement needs. Our surface modification process allows for sub-nanometer thickness coatings to be tuned for optimal battery performance for battery chemistries relevant in both consumer electronics as well as vehicle technologies. In this work we will demonstrate the significant gains observed to commercially available LCO by way of increased cycle life, increased energy density and beyond state-of-the art fast charge capabilities. Alongside performance enhancements, we will explore the mechanisms of ALD protection at high voltage.

0:00:02   Hello. My name is Barbara Hughes, and I'm the director of energy storage at for Gina. And today,
0:00:08   rather than talking directly about particle ale de, I'm gonna be talking about some of the
0:00:13   applications or a lady has a significant impact in particular, I'm going to be sharing some of the
0:00:19   work that we've done it. Forge on with him cobalt oxide, or else CEO and the improvements that we've
0:00:26   made on the state of the art of L CIO graphite cells.
0:00:32   So by now, I'm going to assume that most of you have a pretty clear picture about who we are at
0:00:38   Fortune Nano on what we do. And all I'll say is that alongside my 50 or so co workers, it's our
0:00:46   mission to demonstrate a LTE as an innovative material solution.
0:00:53   So there are a number of different fields for a waiter in which a lady has the potential to push the
0:01:02   needle on the state of the art and to disrupt the current convention per se and some showing some of
0:01:08   those on this slide. This is not an exhaustive a list by any means, but the bold it squares
0:01:17   represent those areas that Fortunato is actively exploring in different projects and in particular
0:01:24   today I'm going to talk to you about the green box, which is energy storage and about the impact
0:01:29   that we have seen Teoh lithium ion batteries. Um, using particle a l. D.
0:01:37   So, as I said previously, I'm going to talk today about performance improvements that we've achieved.
0:01:43   It forged Nano to the state of the art of lithium cobalt oxide cells. So why do we care about l CEO?
0:01:53   Why is it interesting for us to enter this space? So in particular, L CEO is interesting because
0:01:59   it's used in a plethora of consumer of electron ICS. So over 90% of consumer electronics batteries
0:02:06   are made from looking cobalt oxide. So that's everything from cell phones and laptops to power tools
0:02:14   and in the entire battery space. L CIO accounts for about. I think it's 15% of batteries utilize
0:02:21   globally. And, um, trends don't show that this is going to change in the next decade or so. So there
0:02:29   are a number of reasons why else CEO is so prolific, and I'm showing some of those on the left. So
0:02:37   first and foremost L CEO is very easy to synthesise. Um, it has, ah, high energy density and high
0:02:44   specific energy. And so, for the small packages that must deliver high energy like cell phones and
0:02:52   laptops, cell CEO is particularly useful for its high energy density. It has stable charge and
0:02:59   discharge voltages, and it has excellent cycling stability as well. However, L CEO does have its
0:03:07   limitations, and fortune now has really been focus on trying to improve some of these limitations
0:03:13   that I'm showing here on the right. So first and foremost, it has a low specific power. So it has
0:03:19   limited load capability when you start to push to higher voltages, um, in and around 4.6 folds, you
0:03:26   do incur some transition Elvis transition, metal dissolution, which means that we start to lose
0:03:33   cobalt in the structure which results in structural instability in the Catholic active material as
0:03:40   well as, um, there are some side reactions that can occur with the electrolyte, which lead to S C. I
0:03:47   thickening, um, as compared to its, um, other lithium ion battery counterparts like the in EMC's L
0:03:56   CEO does have a relatively short life span, so commercial targets for L CIO devices are 2 to 500
0:04:04   cycles, which is relatively low if we think about the eight toe 1800 to 1000 cycles that we think of
0:04:11   for, um vehicle technologies L CEO. Also, several suffers from low thermal stability. So at higher
0:04:20   temperatures, we also see this beginning of the S E I layer on this particular chemistry is
0:04:26   susceptible to thermal runaway. So as I mentioned that Forge, our studies really have focused on
0:04:33   trying to address some of these issues of lifetime high voltage stability on and even some of the
0:04:40   thermal testing thermal stability testing. And we've done some of this, but those results are pretty
0:04:45   preliminary at the time. So I'm not gonna including you those at this time. But I'll sort of wet
0:04:50   your appetite for staying tuned to our next talks in the future. So as I mentioned on the previous
0:04:58   line, there are three major areas where L CEO could use improvement on. I'm listing them here so
0:05:06   stabilizing the layer dockside structure specifically as we push the upper cut off voltage to higher
0:05:13   and higher voltages and L CEO the second, which is suppressing growth of a C. I so too remove as
0:05:21   much as possible that detrimental interactions between the Catholic active material and the
0:05:26   electrolyte, and finally, to improve upon safety. Now, a number of approaches have been taken to
0:05:34   achieve these goals, and I'm listing a three of the most promising here. Um, they are to, um, make
0:05:42   the material smaller to improve particle infusion length. So people tried to achieve this using nano
0:05:48   L CIO particles. But as we know, nanoparticles are very much surface dominated. So we increased
0:05:54   surface area, which meant that we actually had more side reactions with the electrolytes. So even
0:05:58   though faster see rates were available, we had Ah, see, I think inning to a large extent, Dough
0:06:07   opens have also been employed. The idea with doping is that they change the activity of service
0:06:12   Adams. So in doing that, you can reduce cobalt, the solution and the structural changes that results.
0:06:20   But doping is really difficult to control, and the electrode is still vulnerable from side reactions
0:06:26   with the electrolyte, especially at high voltages which caused this s e I layer thickening Andi.
0:06:32   That leaves us with the third strategy, which is surface modification. And with service
0:06:39   modifications, you can really You can do a lot of things. You can improve specific capacity rating
0:06:44   capability. Um, but there are various types of surface modifications, and they all kind of have
0:06:50   their benefits and limitations. Of course. Today I'm gonna talk to you about a lady a surface
0:06:55   modification to improve l CEO. There are, of course, a number of different L CEOs on the market. And
0:07:04   as you can see here, they are not all created equally in the plots that I'm showing here. These are
0:07:11   three different L CEOs from three different suppliers that we have tested in house at Forge. Um, and
0:07:18   their performances are vastly different. They have different chemical preparations, different
0:07:23   surface modifications and different dope ins, as most of the suppliers have used, probably many of
0:07:30   the techniques that I showed on the previous slide to try to improve the performance of their
0:07:34   materials. It is worth noting, however, that it forge. We have seen improvement to each of these L
0:07:42   CIO formulations. But the improvements, I would say, are only going to be as good as the base
0:07:49   materials. And so, while we've made some vast improvements, we have chosen to go forward on the data
0:07:54   that I'm going to show you today with L CIO preparation one I'm showing here,
0:08:02   here. I'm just showing some examples from the literature of, um of employing a l d to coat l CEO. We
0:08:11   are not the first to do it at Forge Nano. However, ah, lot of the data that we have seen in the
0:08:17   literature thus far has been overwhelmingly done on the electrode laminate rather than on the raw
0:08:26   cathode materials.
0:08:30   So why are we interested in currying powders versus coating the electrodes themselves? Well, there
0:08:38   are pros and cons to both strategies. Looking first at the electrodes. One of the benefits of coding
0:08:47   the electrodes is that you're not going to change in your processing parameters for mixing coating
0:08:53   casting the electrode because the alias obviously done post cast. Additionally, the conducted path
0:09:02   with particles should be maintained, so you're coating everything as an aggregate. You will have the
0:09:10   same Ailey coating on all of the components of your electrodes. So that's active material binder
0:09:15   conductive additive. But the one downside to all of this is that any ale de chemistry that you would
0:09:21   like to use your processing window is going to be limited by your least stable component to your
0:09:28   electrode, which is often the binder. So whatever your glass transition temperature is in your
0:09:32   binder, you're gonna be limited to that processing window. The flip side of that is to coat the
0:09:38   particles themselves. So now you're not limited to having a single ale decoding on every single
0:09:45   component of your electrodes, so you can do a different coding for every individual component like
0:09:50   your active, you're buying very conductive additive on, and you're gonna do a complete coating
0:09:54   around every particle. But what that does mean is that you might change, have to change some of your
0:09:59   processing parameters. So we've we've observed in our lab that a nail decoding might have an effect
0:10:07   on such things as the viscosity. And so we might have to change things about the way that we make
0:10:12   our electrode paint or the way that we actually cast are electrodes. So that's one negative there.
0:10:18   But we do have a greater possibility in terms of processing window, so greater possibility and
0:10:24   chemistry's that we can use in terms of a LTE processing when we coat the powders individually. And
0:10:30   then, of course, also posting pre treatments are possible.
0:10:36   So now that we have a little bit of a background on why else CEO is important and why including the
0:10:42   powders is more valuable than coating lamb. It's so we believe we get to get to the more exciting
0:10:48   stuff. So I'm gonna walk you through our optimization of L CIO at Forge Nano. And I'm going to try
0:10:56   to check every one of these boxes. It is my intent to show you the work that we have done in each of
0:11:01   these areas. I'm going to address cycle life, fast charge, increasing the loadings from what are
0:11:08   standard cast looting is for our electrodes, and then we'll also be looking at increasing both
0:11:15   capacity and high voltage performance. So we're going to start here with what I like to call our Jen
0:11:21   one coating and our Jen one coating includes a coating that is just going to be on our cathode.
0:11:29   We're gonna be cycling to moderate upper cut off voltages for Elsie. Oh, so that's 4.4 bolts and
0:11:36   then all of the durability cycling here. So that's all of the cycling done. Post formation, all the
0:11:42   durability cycling will be done at 0.5 c. Charge one. See discharge unless I say otherwise. And so
0:11:50   there's just a couple of cases of obviously, um, fast charge will be going faster than that. And all
0:11:57   of this data was collected using standard gin to electrolytes. So that's 1.2 Mueller elope at six in
0:12:04   3 to 7 weight percent e c M c No additives. Looking at this data is a little bit like stepping back
0:12:13   in time. This was the beginning of our evolution of optimization for L CEO in our very first lesson
0:12:21   in coding L CIO, which is that our Ailey coating is not one size fits all, So there's a lot of
0:12:28   optimization that is required. Thicknesses and coding chemistry have to be optimized as a function
0:12:35   of the base materials and the base material composition, as I was showing you previously, which all
0:12:41   very greatly so once we decided to stick with our initial L CEO or the best performing base material
0:12:48   that we could get our hands on, we moved in. From there, we start with one coding chemistry and in
0:12:54   this particular case, were optimizing the thickness of our coating with a single coding chemistry
0:13:01   and the pristine cell is shown in this blue trace. And then we have a couple of different coating.
0:13:08   Thickness is represented in Red and Gray's Elegy Wanna Nail D to. What's most striking is that in
0:13:15   both cases we significantly improved the lifetime of the material. But we do it at a cost, which is
0:13:22   a bummer. So we lose about 6% capacity at the beginning of life. And so my team has begun calling
0:13:30   this the triangle of death. And so that 6% capacity lost at the beginning of life is the toll that
0:13:38   we've had to pay initially to getting that higher cycle life
0:13:44   taking the best of the two coatings that we saw previously. So the red that only had 6% capacity
0:13:52   loss at the beginning of life, we can dig a little deeper. We look at the differential capacity
0:13:57   plots here as a function of voltage, and you can see some pretty striking qualitative trends that
0:14:05   jumped right out at you. So, first and foremost, we look at the pristine evolution of the pristine
0:14:10   material from the first durability cycle, which is cycle five all the way through to cycle 600. And
0:14:16   what you notice is that the onset potential we see a shift over time a pretty significant shift that
0:14:23   over potential caused by an increased in internal resistance in the cell. And we see it in both the
0:14:29   cathartic and in the Antarctic currents. So we see this shifting behavior for the onset potential in
0:14:37   the ail decoded material. We don't see that significant shifts way. Still see some light shifting
0:14:44   and the onset potential is a little higher and a lady material at the beginning of life, and it is
0:14:50   for the pristine. But if we take a look at each cycle independently and look at the pristine and ale
0:14:57   decoded, you'll see that there is a little bit of an offset for the onset voltage. So the pristine
0:15:03   has a lower onset voltage, but quickly it surpasses the ale. Decoding the other thing that you'll
0:15:09   notice is that for the ale decoding, we have this what would call ah, break in period? And I'll keep
0:15:15   referring to this break in period throughout this presentation. So this dark red line is actually
0:15:23   the beginning of life for the ale decoded. So we see that way have very non, um um resolved peaks in
0:15:34   the Dick Utd spectrum which become more resolved over time and maintain that sharpness throughout
0:15:41   life, while the Christine Material is is starting to lose. Ah, lot of that definition. And so we do
0:15:47   have a break in period for the ale decoded material, but then it's very stable over its lifetime as
0:15:52   compared to the pristine material. That breaking period is also reflected in the voltage versus
0:15:59   capacity plot. So in this particular plot, the, um ailed you coded material is represented by the
0:16:07   dashed wines. The pristine materials are represented by the solid lines and you can see at the very
0:16:13   beginning of life, the ale de material has this again break in period, where after 100 cycles air.
0:16:21   So we see much smoother curves in the voltages, this capacity plot and then over the life of time of
0:16:28   the material. It's a much tighter distributions of less capacity loss. In the vaulted versus
0:16:35   capacity plot. You see a much tighter distribution here for the ale decoded material on been a
0:16:42   larger loss of capacity for the pristine material. If we look at the plot on the right, I'm zoomed
0:16:49   into that initial portion of the world's versus capacity plot that shows us the initial IR drop at
0:16:57   the beginning of discharge and you'll see that that I are drop is pretty significant over the
0:17:02   lifetime of the pristine material. Whereas the ale decoded material, it's much less significant. And
0:17:09   again you get a tighter distribution over the cycle life of the ale decoded interior.
0:17:17   Given what we've seen thus far, we now wanted to see whether or not the improvements to the ale
0:17:23   decoded material would still be observed under fast charge conditions. So we're keeping the upper
0:17:29   cut off voltage at 4.4 volts, but we're now is changing the charge rate to three C still with a one
0:17:36   C discharge. And these are the plots that result under fast charge conditions both the pristine and
0:17:43   the ale de coded L CEO. They both achieve slightly fewer cycles, and they did that lower Syrians,
0:17:50   but not appreciably lower. But what has changed significantly are the capacities reached or the
0:17:57   capacity values reached due to the lower charge efficiency at high sea rates in both cases, but
0:18:04   otherwise the overall trends do right do remain the same taking a really quick look at their
0:18:12   differential capacity plots as well as the voltage versus capacity plots. We see some very similar
0:18:18   trends, So on the left we have the differential capacity plots in the blue. We have the pristine at
0:18:23   the beginning of life. The solid line, the dashed line after 300 cycles, which is the end of life of
0:18:29   this particular material again, with the Red Me are ailed, decoded, the solid being beginning of
0:18:35   life and the dashed the end of life are at 300 cycles. We see a significant shift in the, um,
0:18:44   initial voltage onset voltage for the pristine material. And hardly anything changes for the Ailey
0:18:52   coded material over the lifetime of over 300 cycles. So that's pretty significant. That's really
0:18:57   that was really interesting to us. You can also see in the voltage versus capacity plots on the
0:19:03   right that we do see a more significant IR drop a t beginning of discharge for the pristine material
0:19:11   than we do in the ail decoded. So we really see a ah a large growth and internal resistance for the
0:19:18   pristine cells on, and we see a lot of stability and the ale decoding. Finally, we can look at the
0:19:26   efficiency of charging during fast charge. So here I'm showing a state of charge overtime plot, and
0:19:35   you can see that starting at cycle four for both the pristine and ale decoding, which is our first
0:19:41   durability cycle and then going out to cycle 300. So at Cycle four, for both of the materials, they
0:19:47   can hit 80% state of charge and under 15 minutes. So they're both hitting at about 12 minutes at the
0:19:53   beginning of durability cycling. But by 300 cycles, the pristine is is going to take it at least 20
0:19:59   minutes to reach that same state of charge. And this is presumably due to that greater growth and
0:20:07   resistance that was alluded to in our differential capacity plots as well as in our capacity. Verses,
0:20:13   vultures, plots, um, Walters versus capacity plots. Excuse me that we saw previously for the
0:20:19   pristine materials, the pristine material at the end of life. It's just not as efficient at fast
0:20:25   charge as our ale. Decoding is
0:20:30   so now, having demonstrated improved cycle life and better fast charge capability, we're gonna be
0:20:36   moving on to higher electrode loadings so we're gonna keep everything else the same. And we're just
0:20:42   going to transition to electro loadings that are higher than our standard, which is 10 milligrams
0:20:48   per centimeters squared. And the reason for doing this is that 10 milligrams per centimeters squared.
0:20:53   While it's our standard for most electrode materials, kind of a generic standard for commercial
0:20:59   applications, it's on the low side, so this particular study is still in progress. But the first
0:21:06   step is to move to we moved to, like 11.5, but our trajectory is is really something more like 17
0:21:13   milligrams per centimeters squared. But we're going to be stepping it up incrementally to see if
0:21:19   these trends hold, and for the most part, they dio at this particular loading. Maybe the step wasn't
0:21:26   quite big enough, but we do see some changes. Not much has changed in terms of general trends. At
0:21:33   higher loadings, we are reaching slightly higher capacities, but in general, the trends. So I'm
0:21:41   showing the lower loadings and the two plots above and in the to lower plots. I'm showing the 11.5
0:21:48   milligrams per centimeters squared, nothing too significant changed in terms of the trends between
0:21:54   the pristine and the coded. Now, looking at the differential capacity plots, we also see that we see
0:22:00   significant shifting in the onset Voltage, um, at higher loading as well. Is that low loading? So
0:22:07   again, the top is showing the 10 milligram per centimeter square initial loading that I've already
0:22:12   shown. And then the plots below are showing the higher loadings. We see incredible stability again
0:22:19   with our ale de coded L CEO material and a lot of shifting in that onset voltage for our pristine.
0:22:31   So in this case, we does a little bit deeper. As I mentioned, we suspect that all of the over
0:22:39   potentials that we see in the Spectra for the pristine material correlates directly to of growth and
0:22:46   internal resistance in the cell. So we do a quick a quick look at some rate dependent DCR
0:22:55   measurements on and you can see again. The blue material is the pristine the red is the ale decoded.
0:23:03   The beginning off life is the solid line. The end of life or the 1st 100 cycles is represented by
0:23:11   the dash wine. And what you can see is that the tone resistance increases over time or is increasing
0:23:19   over time for the pristine material while it's dropping for the ale decoded material. And we fully
0:23:24   expect to see crossover in these cells as they reach higher cycle lives. But for now, we're only at
0:23:32   about 100 cycles. So we're continuing testing here. And as I said before, we will be incrementally
0:23:38   increasing the electoral loadings to see if this trend does hold. Um, and if we do have a lower
0:23:44   resistance overall in our ale decoded cells, then as we move to higher loadings, this lowered
0:23:49   resistance will work in our favor.
0:23:54   So moving right along, we have thus far seen that we can make significant improvements to cycle life.
0:24:01   So we've checked that box and these improvements are preserved at high sea rates. So we check that
0:24:06   box and we were able to increase the electoral bloating, albeit only slightly for now. But we can
0:24:13   also check that box and it's still in progress. So you're probably thinking you can't really check
0:24:21   the capacity box because you took a 6% capacity loss at the beginning of like you still have that
0:24:27   triangle of death and we we understand that so we have to go back and we need to address capacities,
0:24:34   capacity issues that we have created with our Jen one coating. And so that's what we set out to do
0:24:42   next. So our first thoughts when thinking about negating that 6% capacity loss at the beginning of
0:24:50   life was to change the coding chemistry. But what we actually found worked better was to coat both
0:24:58   the ANA and the cathode. So what I'm showing here in the top left quadrant is the cycling data for
0:25:08   the pristine in blue and the ale, the dual ale de coated with cathode and an ode in green. And what
0:25:18   you see is that we no longer have that 6% capacity drop at the beginning of life. But we also
0:25:26   haven't improved the lifetime of the material and we haven't changed the capacity. Now, what we were
0:25:33   interested in doing here was just figuring out if we could count her balance the loss of lithium
0:25:41   inventory that you see in a pristine sell duty s EI formation by putting a coating on the Yano. So
0:25:48   by putting a pseudo S c I on the Anna, we didn't see the same thickening of the s e I. We didn't see
0:25:54   the same loss of lithium inventory due to S C I formation. And so we didn't see the same
0:26:01   irreversible capacity drop in the ail decoded material as we do in the pristine. And in that way, we
0:26:08   recouped that 6% capacity lost that we see from just coating the cathode. So nothing too interesting
0:26:16   about the cycling data. If we look at it in the top left quadrant, it's like a big deal. We create.
0:26:22   We, uh, um fixed a mistake that we created. But we were very interested because of all of the
0:26:30   resistance related issues that we've seen in the pristine material previously to see how the
0:26:36   resistance fares when we now have a coating on both the note and the Catholic. And so we looked into
0:26:41   the electoral impedance spectrum that I'm showing in the bottom left quadrant. And so the again blue
0:26:48   dots are the pristine. Ah, the green is representing the ale D, and this is over the course of 350
0:26:55   or so cycles. And I've quantified the change in charge transfer resistance here in this bar chart,
0:27:03   and what's very interesting is again we go back to this breaking in period that we expect in the ail
0:27:08   decoded material. So at the beginning of durability, cycling and cycling in cycle four, we see that
0:27:15   the resistance a charge transfer resistance is actually slightly higher for the Ailey coded material.
0:27:21   But over time it goes down, and then resistance growth over the lifetime of the ale decoded material
0:27:26   is lower than it is in the pristine. So we were pretty excited about that. But as I said previously,
0:27:34   we hadn't done anything inherently interesting. We hadn't improved the cycle life. We hadn't
0:27:40   improved the capacity. So our optimization journey continues, and we tried to recoup that long cycle
0:27:49   life that we had before that we had observed before in our ale decoded material. So I'm showing you
0:27:56   the next iteration in four Gino's L CIO saga. And on the left, I'm showing at the beginning of life.
0:28:04   Um, the blue traces there just replicants of one another are the pristine. The red is our general
0:28:11   one ale decoding where we do see that 6% capacity life. But then over the lifetime of material, we
0:28:18   actually get a longer a lifetime, and then the green is now our jin to ale de coding our dual ale
0:28:26   decoding and those replicates that you can now see on the right, Um, get the same cycle life as our,
0:28:33   uh, singularly coated cathodes, which is improved over our pristine material. And so you can see
0:28:41   that now, not only have we negated that 6% capacity left capacity, loss of the beginning of life,
0:28:47   but now we also maintain a higher capacity and a longer lifetime with our pristine with our new
0:28:54   optimized ale decoding. So now that we've obliterated the triangle of death, I'm gonna move the
0:29:02   conversation towards high voltage. And we're also gonna put what I like to say is the cherry on top
0:29:08   of the capacity portion of this talk because we all know that the easiest way to increase available
0:29:13   capacity is by raising the upper cut off voltage. Now we're raising the upper cut off voltage from
0:29:20   4.42448 And in doing so, our formation capacity increases from approximately 180 million power
0:29:29   program to right around 2 10 and so increasing epic off voltage, we've also increased our capacity.
0:29:36   Unfortunately, we've also significantly shortened the cycle life in both materials. So, um, the
0:29:44   cycle life is credited on the right as being to 80% of, um, the initial capacity and so you can see
0:29:52   for the pristine material we only get about 100 cycles, but we're still getting twice that in the
0:29:58   ale de coded material. What I'm also showing here a the very bottom is resistance over a cycle life
0:30:06   for both the pristine and the ale decoded. And you can see that during durability Cycling the ale
0:30:13   decoded maintains a lower resistance over time as compared Teoh the pristine material. So there's
0:30:21   not a lot mechanistic lee that we can get from just a generic resistance plot over the cycle life of
0:30:28   the material on. And so we've also investigated this particular system using E i s or electro
0:30:34   impedance spectroscopy.
0:30:37   But before we dive into the spectra for the 448 system, I just wanted to take a quick minute to give
0:30:45   a little refresher on what we expect to see from the Nyquist plot of just pristine L CEO as a
0:30:52   function of two different voltages. So what I'm showing here on the left is the pristine at 4.4, um,
0:30:59   and on the right is the pristine and 4.48 volts. Both of these air taken at 40% state of charge, and
0:31:07   they're both taken in cycle 56. So what you'll notice is that as a function of voltage, the
0:31:16   contributions of the semi circles changes a little bit. So if we look at the semi circles for L CEO,
0:31:23   they're sort of three qualities of things that we can pull from a Nyquist pot of L CIO, the first of
0:31:29   which we can only really quantify if we fit it. Um, and that would be where that first semi circle
0:31:37   would cross the X axis. So that's our contact resistance. The second is the width of that first
0:31:42   semicircle, which is commonly accepted to be, um, an attributes, um of the S e i resistance on both
0:31:52   the honored and the Catholic. So the CSC I as well and then the second semi circle is attributed to
0:31:59   the charge transfer resistance, which is the resistance. Um attributed Teoh charges moving from the
0:32:07   electrolyte through that first, um, Montel layer, or through the first layer of the cathode active
0:32:15   material So what? You'll see as you go from 4.4 volts to 4.48 volts, where I should again remind
0:32:23   everybody that we're just using standard electrolyte no additives. This is not a high voltage
0:32:27   electrolyte. So obviously we're going to have more electrolyte breakdown at higher voltage, which
0:32:33   you can see very clearly from the contributions to the, um, Nyquist plot of the S E i portion. So
0:32:41   that first a semicircle for 448 So at high bowl, did you get more electrolyte breakdown? We get a
0:32:47   higher contribution from the S E I layer as compared to 4.4
0:32:53   Now, we're gonna look at each component of the Nyquist plot. As I mentioned previously, the best way
0:32:59   to do this would actually be to fit that first semi circle to get the contact resistance, which we
0:33:04   have done. But I'm not showing here. So in each of these plots, I'm just quantifying are in each of
0:33:10   these charge on quantifying each of these, uh, resistance is that you can get from the Nyquist plot
0:33:17   and I'm showing here that the contact resistance is lower over the lifetime of the ale. The material
0:33:23   it starts lower and maintains lower growth throughout the lifetime. Barely. Um, it's barely getting
0:33:31   larger in the A league coded material as compared to the pristine. Now, if we look at that s c I
0:33:39   layer resistance. So this one is very interesting, Which is why I took a minute to, uh, take a look
0:33:44   at just the pristine material as a function of the voltage. And so if we look at the S e, I lay
0:33:51   resistance over time from cycle five to cycle 107 you'll see that the pristine, which is in the blue,
0:33:57   has that large contribution from S C I layer resistance, however, are ailed. De, um, uh, Spectrum
0:34:07   shows that we have nearly equal contributions. We have a much smaller contribution from the S E I
0:34:12   layer resistance. And so we're not seeing the same resistance growth a. T beginning of life are
0:34:18   ailed de coded material. And then over time we actually see that resistance growth, um is still
0:34:25   lower, and they'll decoded betrayal over time. Now where we do where we do finally see the
0:34:34   resistance in our ale decoded material start out higher than our pristine is in the charge transfer
0:34:40   existence. And as you'll recall, the charge, transfer resistance is the charge transfer from the
0:34:48   electrolyte into the cathode active material. And we've now put this additional barrier this ale de
0:34:55   coding on our Catholic active material. So at the beginning of life, we have a higher charge
0:34:59   transfer resistance observed for the Ailey coded material. But then, after this break in period, we
0:35:06   see that the resistance goes down and then is maintained at a low value over the lifetime of the
0:35:12   material as compared to the pristine. Now, as I mentioned before with raising the upper con off
0:35:19   voltage, we really are putting the cherry on top of our endeavours to increase the capacity as well.
0:35:25   So what I'm showing here is both the 4.4 and the 448 data plotted on the same plot. The 448 data. It
0:35:34   has the black fill. So the 448 gin to a L. D. Is in green the gin, two at four fours, also in green.
0:35:44   The pristine at 44 is blue with a black fill. And then at four, our Stephen it for 48 Blue of the
0:35:52   black fill And at 44 it's just the blue Trace. And so you can see here that, um if we look at the
0:36:01   lifetime incapacity gain that we get at 448 we can hit the same lifetime that we're hitting at 44
0:36:09   with about a 50% increase in available capacity by increasing the upper Khanna voltage before for
0:36:18   eight. And the Ailey coating here really is allowing us to push the limits of this material and its
0:36:24   allowing us to do it under standard cycling conditions with standard electrolyte. There's nothing
0:36:29   special about electrolyte and were significantly increasing a lifetime and the available capacity
0:36:37   before I wrap this up. I just wanted Teoh. Look at some directional data will say so. Um, the state
0:36:44   of the art is quickly increasing Teoh and will soon be at 46 Right now there are extreme stretch
0:36:53   bulls. Um and this material is not optimized. This is our jen one that just for kicks, we decided to
0:37:02   see how it would perform at 46 and I would say it's directional and that it shows the real
0:37:10   importance of having an ale de coding on the Catholic material to stabilize the cathode material. So
0:37:17   while our pristine material only gets about 20 cycles, um, rld coded gets, you know, almost three
0:37:25   times that at 46 neither of them performed extremely well. But you can see that there is a
0:37:33   significant stabilization effect that are Ailey coding is having, um, on this layered material again.
0:37:43   Just a quick look at the Nyquist plots and quantifying the values for a charge transfer and contact
0:37:50   resistance. You see that at 46 the charge transfer resistance where the ale decoding is actually on
0:37:56   par with the pristine on ben ale de the building material. The growth of resistance in the building
0:38:04   material is a lot less over the lifetime is compared to the pristine. The same can be seen for the
0:38:10   contact resistance. And so, overall, um, ale de Arjun one ale de material, um, stands up firmly at
0:38:19   46 against the pristine. Um, and so we're excited to see, um, how much further we can take these
0:38:25   optimization. And, um, we're pretty confident that will be able. Teoh hit significant lifetimes
0:38:33   between two and 500 at 4.5 volts on and maybe even hit our stretch goals. at some point. So in the
0:38:42   end, I think we have set out to do. We've done what we set out to Dio, in some cases Onley
0:38:49   preliminarily, and there's more work to be done. But we've demonstrated increased cycle life with
0:38:55   more efficient Bashar capability. Um, we've shown that through improved resistance characteristics
0:39:01   of the ale de material that we do maintain the benefits of a lady at a higher electrode loadings. Um,
0:39:08   and we've been addressed some of the inherent pitfalls and improve capacity and high voltage
0:39:14   operation of L CIO graphite cells.
0:39:20   So with that, I'd like to take just a moment to recognize all of my teammates and all of their hard
0:39:27   work. Aziz. Well, A. Some of our collaborators at Carlos will minds to do a lot of the visible
0:39:34   characterizations for us. Um, if, by chance I am not able Teoh, answer any questions that you may
0:39:41   have during the Q and A section here. I am very happy to be contacted personally. Email address is
0:39:48   here. I will get back with you as Justus soon as I possibly can. Um, if you decide that you just
0:39:55   cannot wait to get your hands on some of this amazing l ceo graphite. Um, definitely direct any of
0:40:01   your business enquiries towards Daniel Hicks, our, um, account manager in the interview storage
0:40:08   realm. And with that, I would just like to thank all of you for your attention. And I hope that you
0:40:14   are staying safe and staying healthy. Thank you.